liquid counting organic - CiteSeerX

Loading...
LIQUID SCINTILLATION COUNTING and ORGANIC SCINTILLATORS

Harley Ross John E. Noakes Jim D. Spaulding

LEWIS

Library of Congress Cataloging-in-Publication Data Liquid scintillation counting and organic scintillators/ edited by Harley Ross, John E. Noakes, and Jim D. Spaulding.

p. cm. 'Proceedings of the International Conference on New Trends in Liquid

Scintillation Counting and Organic Scintillators, 1989P. Includes bibliographical referencesand index.

1. Liquid scintillation countingCongresses. 2. Liquid scintillatorsCongresses. 3. Organic scintillatorsCongresses. 1. Ross Harley. II. Noakes, John E. III Spaulding, Jim D. IV. International Conferenceon New Trends in Liquid Scintillation Counting and Organic Scintillators (19989: University of Georgia) QC787.S34L57 1991 ISBN 0-87371-246-3

539.7'75dc2O

90-26005

COPYRIGHT © 1991 by LEWIS PUBLISHERS, INC. ALL RIGHTS RESERVED Neither this book nor any part may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, microfilming, and recording, or by any information storage and retrieval system, without permission in writing from the publisher. LEWIS PUBLISHERS, INC. 121 South Main Street, Chelsea, Michigan 48118 PRINTED IN THE UNITED STATES OF AMERICA

Preface It has now been over 30 years since th.e first conference dedicated exclusively

to liquid scintillation counting was held at Northwestern University (Aug. 20-22, 1957). One might expect that over this time period, something as routine as radionuclide assay via the liquid scintillation process would have entered the realm of the prosaic or perhaps even stodgy. Fortunately, this is not the case. Exciting, new research initiatives in the areas of organic scintillators, electronic detection systems, and advanced information processing methodology have come together to create completely new measurement concepts that further expand liquid scintillation applications. Today, we are able to make measurements that were unthinkable just a few short years ago. It is probably fair to say that the driving force in this continuing technology development is the growing need for improved sensitivity, selectivity, and accuracy in nuclide assay. Primarily this arises from demands in biomedical and environmental assessment efforts. Only liquid scintillation counting appears to meet the wide range of application that is required in many of these projects. The broad expanse of problems investigated with liquid scintillation techniques often complicates the reporting, making it difficult for workers in fields

outside some specific area of interest. For example, a cursory search of the literature on liquid scintillation in 1988 showed that there were articles of general interest in at least 33 different technical journals! Even the most dedicated researcher would have difficulty keeping up with the latest developments. In response to this diversity of interest, the periodic liquid scintillation conferences, and their proceedings, have served the important role of bringing investigators with different concerns into a common forum. Such conferences have been held continuously every three to five years since 1957. The international conference on New Trends in Liquid Scintillation Counting and Organic Scintillators, held in October, 1989 and sponsored by the Oak Ridge National Laboratory and The University of Georgia, is a continuation of this essential

tradition. Our hope is that both the conference and these proceedings have fulfilled the promise of our eminent predecessors. No conference can take place, nor proceedings published, without the help of a great many people and organizations. Thus, the organizers would like to recognize Beckman Instruments, Inc., Packard Instrument Co., and Pharmacia LKB Nuclear, Inc. for the generous financial support that made the confer-

ence possible; B. M. Coursey, C. Ediss, A. Grau Malonda, S. Guan,

C-T. Peng, H. Polach, and F. Schoenhofer, our international Scientific Advi-

sory Committee that served to unify the various research interests into a superb technical program; Carolyn Morrill (UGa) and Linda Plemmons (ORNL) for handling the mountains of paperwork and the massive logistics associated with the conference; Arnold Harrod our conference coordinator for his efforts before and during the conference; the staff of Lewis Publishers, and particularly Vivian Collier, who made the publication of these proceedings as painless as possible. Finally, we would like to acknowledge our conference participants and authors, without whom the above would have meant very little. Our sincerest thanks to all of you! Harley Ross John Noakes Jim Spaulding

Harley H. Ross is a Senior Research Staff Member at the Oak Ridge National Laboratory (ORNL), Oak Ridge, Tennessee and also serves as Professor of Chemistry at the University of Tennessee, Knoxville. His formal training as an analytical-radiochemist has guided his research work into a broad range of

theoretical and applied nuclear-related problems, primarily in the area of radionuclide characterization and measurement. His professional interests also include the development of enhanced instrumentation concepts for analytical assay, the role of computer simulation in radiochemistry, and the use of radioisotopes in making a variety of chemical measurements. He has published over 70 papers, has contributed chapters to several books, including two encyclopedias, and has worked as a consultant, both for other national laboratories and private industry. In 1967, he was awarded an IR-100 award, the first at ORNL, for his development of a radioisotope-excited light source spectroscopy system. Since his first publication in liquid scintillation counting almost

30 years ago, he has maintained his interest in the field and continues to explore and publish novel ideas for improved scintillation measurements.

John E. Noakes holds a dual position as Director of the Center for Applied Isotope Studies and Professor of Geology at the University of Georgia in Athens. His formal training has been in the fields of chemistry, geology, and oceanography, in which he has published over 80 scientific papers and holds 9 patents. His scientific interests are directed to applied and basic research with a strong emphasis in the use of stable and radioactive isotopes. He has over 30 years experience in liquid scintillation counting and has consulted for most of the major instrument manufacturers, lectured at the national and international levels, and assisted in the set-up of 28 benzene radiocarbon laboratories world-

wide. His current interests in liquid scintillation counting are directed to instrument design and the application of low-level counting to environmental studies, hydrogeology, and radiocarbon dating.

Jim D. Spaulding is a member of the research staff at the Center for Applied Isotope Studies, University of Georgia, Athens, where he is Associate Director and manager of the environmental monitoring program. His formal training in physics and mathematics, and an academic year fellowship at the Special Training Division of the Oak Ridge Associated Universities, Oak Ridge, Tennessee, gave him the background for continued research in low background nuclear instrumentation. He has co-authored in excess of 30 scientific papers, several of which are on the design and use of low-level liquid scintillation counting, especially as applied to environmental monitoring. He has also served as a consultant to a major liquid scintillation counter manufacturer and as an Expert in Indonesia for the Department of Technical Assistance and Cooperation, International Atomic Energy Agency, Vienna, Austria.

In memory of Donald L. Horrocks 1 929-1985

A Tribute to Donald L. Horrocks

C.T. Peng

It is fitting for Harley Ross, John Noakes, and Jim Spaulding, the organizers of the International Conference on New Trends in Liquid Scintillation Counting and Organic Scintillators, 1989 to dedicate this Conference to the memory of Donald Leonard Horrocks, who contributed so much to the advancement of liquid scintillation science and technology. Donald L. Horrocks was born on July 14, 1929 in Dearborn, Michigan. He attended Reed College, Portland, Oregon and graduated in 1951. He continued his study at Iowa State College, Ames, Iowa and graduated in 1955 with a Ph.D. in Physical Chemistry. From there, after two brief research positions, he went to work in Argonne National Laboratory as a research scientist for 15 years. He moved to Beckman Instruments, Inc. in 1971. He passed away on May 22, 1985. He is survived by his wife Margaret and two daughters, Andrea and Cynthia. Don began his scientific career in Nuclear Chemistry, measuring the tritium yield in the fission of 242Cf and 233U and in fissile elements proposed for future

reactor use. He measured tritium, alpha emitters, and fission fragments by liquid scintillation counting, using single phototube instruments and pulse height analysis. He recognized the importance of pulse height analysis and energy resolution in radiation measufement by liquid scintillation techniques, and he was the first to report the use of the relative pulse height of Compton edges in the presence of sample beta continuum for quench correction in liquid scintillation systems) This was an important observation and was the basis for the "H" (or "Hor-

rocks") number concept for quench correction, which was introduced at a much later date. His interest in improving detection efficiency led him to the study of scintillation processes, especially the pulse height energy relationship of liquid scintillators for alpha particles, fission fragments, and electrons of less than 100 keY in energy. He also studied the effect internal conversion and intersystem crossing of organic scintillator molecules had on their fluorescence and scintillation efficiency. He studied the self- and concentration-quenching

of scintillation solutes as a result of molecular coplanarity; he studied H-

bonding, proton transfer and excimer emissions, the aromatic solvent transfer efficiencies for different excitation energies, and potential organic scintillator synthesis. He demonstrated the superiority of 1 ,2,4-trimethylbenzene over pxylene as a scintillation solvent; it not only increased the transfer efficiency but also decreases solute excimer emission. He also reported the use of pulse-shape discrimination as a particle counting method with high specific ionization. He reported the use of liquid scintillation counting to measure the disintegration rate of beta-emitting nuclides and the use of H3SnuSmIn for energy calibration.

He measured radioactive noble gases by liquid scintillation counting. His research in organic scintillators and the esteem that his peers had for him lead him to organize and convene a successful international conference on organic scintillators at Argonne, Illinois, in 1966.2 Its success also afforded the impetus

for him to co-organize two additional conferences on liquid scintillation counting held in 1970 and 1979 in San Francisco.3'4

Don moved from ANL to Beckman Industries, Inc. shortly after the first San Francisco conference. This change of employment did not interrupt his research career. In fact, it provided him with an opportunity to put many of his theories and observations on liquid scintillation efficiency and quenching into practical use. He completed his book, Applications of Liquid Scintillation Counting, which was comprehensive and thorough in both theory and practical applications and is a classic of the scientific literature.5 Don introduced the X-ray-X-ray coincidence method for measuring the disintegration rate of 25I by gamma-ray scintillation spectrometry.6 He introduced the concept of the H number for quench correction in liquid scintillation systems. The H number is a measure of the shift of the Compton edge between the spectra of quenched and unquenched samples.7 Both of these inventions allowed new counters based on these principles to be designed, manufactured, and successfully marketed. Don had over 100 scientific publications and held 20 patients related to liquid scintillation science and technology. Many of his contributions, such as quench measure by the H number method, detection of sample-scintillator phase separation by the two phase method, determination of the radionuclide disintegration rate by the extrapolation-H number method, determination of a complete quench curve from a single sample by optical absorption method, development of micro volume counting methods, understanding of system background sources and pulse-height distribution, plastic shift, and detection of measurement of chemiluminescence in samples, are the mainstay of modern liquid scintillation instrumentation. Don was a quiet, thorough, and persuasive man. He approached problems systemically and logically. He was a trusted source for authoritative and reliable information on every aspect of liquid scintillation science and theory. He will be long remembered.

REFERENCES

Horrocks, D.L. "Measurement of Sample Quenching of Liquid Scintillator Solutions with X-ray and Gamma Ray Sources," Nature 202(4927):78-9 (1964). D.L. Horrocks, Ed. Organic Scintillators, New York: (Gordon and Breach, 1968), p. 413. D.L. Horrocks and C.T. Peng, Eds. Organic Scintillators and Liquid Scintillation Counting, Proc. Intl. Conf., (New York: Academic Press, 1971), P. 1078. C.T. Peng, D.L. Horrocks, and E.L. Alpen, Eds. Liquid Scintillation Counting. Recent Application and Development. Vol. I and II, New York: Academic Press, 1980), p.414 and 538. Horrocks, D.L. Applications of Liquid Scintillation Counting, New York: Academic Press, 1974). p. 346. Horrocks, D.L. "Measurements of Disintegration Rates of 1251 sources with a Single Well-Type Thallium Activated Sodium Iodide Detector," Nuci. Instrum. Methods 125(l):105-l1 (1975).

Horrocks, D.L. "A New Method of Quench Monitoring in Liquid Scintillation Counting: The H Number Concept," J. Radioanal. Che,n. 43(2):489-521 (1978).

Contents PMP, a Novel Scintillation Solute with a Largte Stokes' Shift, Hans Güsten and Jeffrey Mirsky Scintillation Counting of Harvested Biological Samples with Low-Energy Beta Emitters, Using Solid Scintillant Filters, C.G. Potter and G.T. Warner

9

DI-ISOPROPYLNAPHTHALENEA New Solvent for Liquid Scintillation Counting, J. Thomson

19

Safe Scintillation Chemicals for High Efficiency, High Throughput Counting, Kenneth E. Neumaann, Norbert Roessler, Ph.D., and Jan ter Wiel

35

New Organic Scintillators, Stepen W. Wunderly and Joel M. Kauffman

43

Advances in Scintillation Cocktails, Jan ter Wiel and Theo Hegge

51

Solidifying Scintillator for Solid Support Samples, Haruo Fujii, Ph.D. and Norbert Roessler, Ph.D.

69

New Red-Emitting Liquid Scintillators with Decay Times Near One Nanosecond, f.M. Flournoy and C.B. Ashford

83

New Developements in X-ray Sensitive Liquid Scintillators at EGG/EM, C.B. Ashford, J.M. Flournoy, S.S. Lutz, and I.B. Berlman

93

Liquid Scintillation Alpha Spectrometry: A Method for Today and Tomorrow, W. Jack McDowell and Betty L. McDowell

105

Application of High Purity Synthetic Quartz Vials to Liquid Scintillation Low-Level 14C Counting of Benzene, A. Hogg, H. Polach, S. Robertson, and J. Noakes

123

An Introduction to Flat-Bed LSC: The Betaplate Counter, G.T. Warner, C.G. Potter, and T. Yrjonen

133

A Plastic Scintillation Detector with Internal Sample Counting and Its Applications to Measuring 3H-Labeled Cultured Cells, Shou-li Yang, Ming Hu, Jia-chang Yue, Xiu-ming Wang, Xu Yue, Jie Li, Zi-ang Pan, and Zhong-hou Liu

143

A New, Rapid Analysis Technique for Quantitation of Radioactive Samples Isolated on a Solid Support, Michael J. Kessler, Ph.D.

155

Dynodic Efficiency of Beta Emitters, F. Ortiz, A. Grau, andJ.M. Los Arcos

167

Solid Scintillation Counting: A new TechniqueTheory and Applications, Stephen W. Wunderly

and Graham J. Threadgill

185

Photon Scattering Effects in Heterogeneous Scintillator Systems, Harley H. Ross

195

Some Factors Affecting Alpha Particle Detection in Liquid Scintillation Spectrometry, Lauri Kaihola and Timo Oikari

211

Modem Techniqes for Quench Correction and dpm Determination in Windowless Liquid Scintillation Counting: A Critical Review, Staf van Cauter and Norbert Roessler, Ph.D.

219

Multilabel Counting Using Digital Overlay Technique, Heikki Kouru and Kenneth Rundt

239

A New Quench Curve Fitting Procedure: Fine Tuning of a Spectrum Library, Heikki Kouru

247

The Effect on Quench Curve Shape of the Solvent and Quencher in a Liquid Scintillation Counter, Kenneth Rundt

257

67Ga Double Spectral Index Plots and Their Applications to Quench Correction of Mixed Quench Samples in Liquid Scintillation Counting, Shou-li Yang, ltsuo Yamamoto, Satoko Yamamura, Junji Konishi, Ming Hu, and Kanji Torizuka

269

Modem Applications in Liquid Scintillation Counting, Yutaka Kobayashi

289

A New Procedure for Multiple Isotope Analysis in Liquid Scintillation Counting, A. Grau Caries and A. Grau Malonda

295

On the Standardization of Bete-Gamma-Emitting Nuclides by Liquid Scintillation Counting, E. Garcia-Toraño, M.T. Martin Casallo, L. Rodriguez, A. Grau, and f.M. Los Arcos

307

The Standardization of 355 Methionine by Liquid Scintillation Efficiency Tracing with 3H, f.M. Calhoun, B.M. Coursey,

D. Gray, andL. Karam

317

A Review and Experimental Evaluation of Commercial Radiocarbon Dating Software, S. Dc Filippis and J. Noakes

325

Efficiency Extrapolation Adapted to Liquid Scintillation Counters, Charles Dodson

335

Applicaationss of Quench Monitoring Using Transformed External Standard Spectrum (tSTE), Michael J. Kessler, Ph. D.

343

Scintillation Proximity Assay: Instrumentation Requirements and Performance, Kenneth E. Neumann and Staf van Cauter

365

The LSC Approach to Radon Countingin Air and Water, Charles J. Passo, .Jr. and James M. Floeckher

375

Liquid Scintillation Screening Method for Isotopic Uranium in Drinking Water, Howard M. Prichard and Anamaria Cox

385

Assessment and Assurance of the Quality in the Determination of Low Contents of Tritium in Ground Water, C. Vestergaard and Chr. Ursin

399

Use of Liquid Scintillation in the Appraisal of Non-Radioactive Waste Shipments from Nuclear Facilities, William L. McDowell

407

Development of Aqueous Tritium Effluent Monitor, K.f. Hofstetter

421

Rapid Determination of Pu Content on Filters and Smears Using Alpha Liquid Scintillation, P.G. Shaw

435

Vagaries of Wipe Testing Data, Jill Eveloff, Howard Tisdale, and Ara Tahmassian

457

The Determination of 234Th in Water Column Studies by Liquid Scintillation Counting, R. Anderson, G.T. Cook, A.B. Mackenzie, and D.D. Harkness

461

Perormance of Small Quartz Vials in a Low-Level, High Resolution Liquid Scintillation Spectrometer, Robert M. Kahn, James M. Devine, and Austin Long

471

The Optimization of Scintillation Counters Employing Burst Counting Circiutry, G.T. Cook, R. Anderson, D.D. Harkness, and P. Naysmith

481

Statistical Considerations of Very Low Background Count Rates in Liquid Scintillation Spectrometry with Applications to Radiocarbon Dating, Robert M. Kahn, Larry Wright, James M. Devine, and Austin Long

489

Liquid Scintillation Counting Performance Using Glass Vials in the Wallac 1220 QuantalusM, Lauri Kaihola

495

Time Resolved-Liquid Scintillation Counting, Norhert Roessler, Robert J. Valenta, and Staf van Cauter

501

Compaarison of Various Anticoincident Shields in Liquid Scintillation Counters, Jiang Han-ying, Lu Shao-wan, Zhang Ting-kui, Zhang Wen-xin, and Wang Shu-xian

513

The Effect of Altitude on the Background of Low-Level Liquid Scintillation Counter, Jiang Han-ying and Lu Shao-wan

521

Calculational Mathod for the Resolution of 90Sr and 89Sr from Cerenkov and Liquid Scintillation Counting, Thomaas L. Rucker

529

Determination of 222Rn and 226Ra in Drinking Water by Low-Level

Liquid Scintillation CountingSurveys in Austria and Arizona, Franz Schonhofer, J. Matthew Barnet, and John McKlveen

537

A New Simplified Determination Method for Environmental 90Sr by Ultra Low-Level Liquid Scintillation Counting, Franz Schanhofer, Manfred Friedrich, and Karl Buchtela

547

Mixed WasteA Review from a Generator's Perspective, John Hsu and Jeanne K. Krieger

557

Disposal of Scintillation Cocktails,.John W. McCormick

561

Liquid Scintillation Waste, Ara Tahmassian, Jill Eveloff and Howard Tisdale

573

Liquid Scintillation Counting of Radon and its Daughters, Daniel Penman

577

History and Present Status of Liquid Scintillation Counting in China, Shou-li Yang

583

LSC Standardization of 32P in Inorganic and Organic Samples by the Efficiency Tracing Method, L. Rodriguez, J.M. Los Arcos, and A. Grau

593

LSC Standardization of Multiganima Electron-Capture Radionuclides by the Efficiency Tracing Method, f.M. Los Arcos, A. Grau, and E. Garcia-Toraño

611

ANSI Standards for Liquid Scintillation Counting, Yutaka Kobayashi

623

A General Method for Determining Sample Efficiencies in Liquid Scintillaation Counting, Willard G. Winn

629

Absolute Activity Liquid Scintillation Counting: An Attractive Alternative to Quench-Corrected dpm for Higher Energy Isotopes, Michael J. Kessler, Ph.D

647

Chemical and Color Quench Curves Over Extended Quench Ranges, Charles Dodson

655

The Characteristics of the Cl-I Number Method in Liquid Scintillation Counting, Wu Xue Zhou

665

Low-Level Scintillation Counting with a LKB Quantalus Counter Establishing Optimal Parameter Settings. Herbert Haas and Veronica Trigg

669

Quench Correction of Colored Samples in LSC, Kenneth Rundt, Timo Oikari, and Heikki Kouru

677

64.

Low-Level Liquid Scintillation Counting Workshop, Gordon Cook and Henry Polach

691

List of Attendees

695

Index

705

CHAPTER 1

PMP, a Novel Scintillation Solute with a Large Stokes' Shift

Hans GUsten and Jeffrey Mirsky

ABSTRACT The excellent fluorescence properties of PMP (1-phenyl-e-mesityl-2-pyrazoline) such as a long wavelength emission of 425 nm, a high fluorescence quantum yield, and a short fluorescence decay time make this compound a promising solute for scintillation counting. Due to a very large Stokes' shift, which is more than twice as large as that of commonly used organic scintillators, PMP is suited as a primary solute requiring no secondary solute as wavelength shifter. The exceptionally large Stokes' shift results in a small overlap between the absorption and emission spectra of the solute, even at high concentrations; hence a much longer light attenuation path length compared to common scintillators. The structural features of the new solute also give rise to a unique self-quenching behavior of PMP. Unlike other scintillators, a shallow maximum of light output vs solute concentration curve is exhibited between 0.01 and 1 M in common solvents or solvent mixtures used in large-volume scintillation chambers. As to other scintillation characteristics such as the scintillation efficiency for 14C and 3H, the performance of PMP is similar to that of good commercial liquid or plastic scintillators. PMP is now used successfully in an 80,000 L large-volume scintillation detector applied in neutrino physics. The unique spectroscopic scintillation properties of PMP also suggest its use in plastic scintillating fibers.

INTRODUCTION

We have shown recently that the class of 1 ,3-diphenyl-2-pyrazolines (DPs) exhibit exceptionally large Stokes' shifts when properly substituted by bulky substituents in the ortho' positions of the phenyl rings.' These sterically hindered DPs have promising potential as novel primary solutes for liquid scintillation counting.2'3 Among the sterically hindered DPs tested for liquid scintillation counting, 1-phenyl-3-mesityl-2-pyrazoline, henceforth called PMP,4 is the best compromise in terms of good photophysical and scintillation properties and ease of synthesis. In this chapter the results obtained on the application of PMP in scintillation counting are summarized.

2

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 1. Photophysical Properties of PPO, DP, PMP, and POPOP in Benzene at Room Temperature Sa(gII) Scintiliator QbF i(s) St(cm) XA(nm) XF(nm) PPO OP PMP

414 49 335

POPOP

1

0.79 0.92 0.88 0.85

1.3

3.2 3.0 1.5

308 362 295 362

365 444 425 418

5,240 4,980 10,370 3,540

toluene at 20CC. bin degassed solution. am

EXPERIMENTAL

A detailed description has ben given56 of the techniques used to measure the photophysical data such as the absolute fluorescence spectra, the fluorescence quantum yields, and fluorescence decay times, as well as the scintillation characteristic.4 The method used to evaluate and optimize the attenuation length of various scintillators consists in measuring the intensity of light collected by a

photomultiplier placed at one end of a 2 m long quartz rod when this is orthogonally crossed by the ionizing radiation of a 60Co or '"Cs source at various distances from the photomultiplier. The light output of the different scintillators has been measured relative to NE 235 H using the techniques described by Kowaiski et al.7 With 60Co, the 70% value of the Compton maximum was used. RESULTS AND DISCUSSION

The main photophysical properties of PMP, DP, the parent compound of PMP, and the two widely used organic scintillators, PPO and POPOP, are compiled in Table 1. PMP combines very good photophysical properties such as a high fluorescence quantum yield QF, a short decay time (r), and the long wavelength emission of the fluorescence maximum (XF), with a high solubility (S). The most striking difference of PMP in comparison to DP and the com-

mon scintillators PPO and POPOP is the exceptionally large Stokes' shift (St) of more than 10,000 cm-'. Since there is no fine structure in the absorption and fluorescence spectra of these large organic molecules at room temperature, the 0-0 transition is difficult to determine, and the energy different of

the absorption and emission maxima is therefore taken as a measure of the Stokes' shift. A comparison of the parent compound DP with PMP, where three methyl groups are substituted in the 2,4,6-position of the phenyl ring, reveals that the large Stokes' shift is introduced only by the sterical hindrance of the bulky methyl groups in 2- and 6-positions of the phenyl ring. The result of this sterical hindrance is that in a large hypochromic shift the absorption is shifted from 362 to 295 nm, while the fluorescence spectrum of PMP in comparison to DP remains nearly unaffected energetically."2 Thus, the Stokes'

shift of PMP is more than twice as large as that of commonly used organic

PMP, A NOVEL SCINTILLATION SOLUTE

3

scintillators. A comparison of the absorption and fluorescence spectra of PMP and POPOP is shown in Figure 1. The advantage of the large Stokes' shift and the broad absorption spectrum of PMP is that PMP can be used as a single wavelength shifter in scintillation counting. Thus, contrary to the classical combinations of primary and secondary solutes used in scintillation counting for more than 35 years, with the use of PMP, secondary solutes are no longer necessary. In several standard tests in

'C and 3H scintillation counting, the results obtained with PMP and with common organic scintillators were intercompared.4 The results will not be repeated here. In general, PMP resembles good commercial liquid or plastic scintillators in its scintillation characteristics. Better results are obtained in samples exploiting the large Stokes' shift of PMP, as in the case of samples where color quenching reduces the counting efficiency.4

Large-Volume Scintillation Since all common organic scintillators show some overlap of their absorption and emission spectra (see Figure 1), self-transfer can occur at high concen-

trations by reabsorption of light emitted from the scintillator molecule. The importance of reducing the reabsorption effects in scintillators by employing solutes with large Stokes' shifts was stressed recently by Renschler and Harrah.8 Due to the large Stokes' shift of PMP, the self-absorption of its own scintillation light is lower than in conventional scintillators. This is of great importance in large-volume applications such as whole-body counters, neutrino detectors, as well as scintillating fibers of tracking detectors for the new generation of particle accelerators. Since the intensity of the transmitted photons is related to the solute concentration and the length the emitted photons have to travel to reach the photomultiplier, the light attenuation length should be greater with solutes having a larger Stokes' shift. Otherwise, each time reabsorption occurs some quanta of energy are lost due to the deactivation processes inherently associated with the fate of the electronically excited singlet state of the solute molecule. A comparative study of five commercially available large-volume scintillators revealed that the attenuation length of scintillators with PMP is greater by 23-48% while the light output is in the range of 57-62% of an anthracene crystal (= 100%). These light output values are obtained with all scintillators tested. A solvent combination called PPP 225 with PMP as the only solute proved to be the best choice for large-volume application. This scintillator had an attenuation length of 320 cm with 57% light output. This scintillator is now operating successfully in the 80,000 L KARMEN detector (Karlsruhe-Rutherford Medium Energy Neutrino experiment) at the Rutherford-Appleton Laboratory.9 An even greater attenuation length of nearly 4 m has been obtained with PMP in a new solvent combination called PMP ND 380. There is another benefit when PMP is used as a large-volume scintillator. Unlike the common organic scintillators, PMP shows little tendency to self-

(4

a

0.6

1.2

0

0

1.0

I

250

I

POPUP

ABSORPTION SPECTRUM

I

300

I

i I

SPECTRUM

EMISSION

400

450

500

550

-j U

I-

>

U

z

I-

U

0

-0.2

-0.4

-0.6

-0.8 Cr

-n

- 0.4

0.8

-1.0

jul i'Ii,uu IiuuuI)'.,...iluui. 350

WAVELENGTH mm)

i

ABSORPTION SPECTRUM

EMISSION SPECTRUM

Figure 1. Absorption and flourescence spectra of PMP and POPOP in cyclohexane at room temperature.

a

Cr

U 3.0

a

a Li

U

U.-

LL

LI

LsJ

i-- 0.4

-

1.4

PMP, A NOVEL SCINTILLATION SOLUTE

5

quenching. Most scintillators exhibit a narrow maximum in their light output vs. solute concentration curves. This phenomenon is partly due to the reabsorption processes described, partly to concentration quenching of the fluorescent state of the fluorophore by its own molecule in the ground state. In Figure 2 the relative pulse heights of PMP and PPO, in a 14C standard toluene solution at 10°C, is plotted vs the solute concentration. While PPO shows a fairly narrow maximum at 0.025 M, PMP displays a shallow maximum between 0.01 and 0.1 M. Even at a concentration of I MPMP (264 g/l) the relative pulse height has dropped only by about 10%. Consequently, by use of PMP the doping concentration can be raised to values at which only non-

radiative energy transfer is effective. We explain this low tendency to selfquenching of PMP by the different geometries of PMP in the ground and in the electronically excited singlet state. Scintillating Fibers The new generation of tracking detectors use parallel bundles of scintillating

optical fibers of extremely small diameters)° Theoretical reasoning about plastic-based scintillating fibers has pointed to the need for improving the attenuation length of common scintillators)' It was shown recently that PMP in polyvinyl toluene (PVT) or polystyrene reach nearly the scintillation efficiency of NE 110 while the transparency of PMP is better by at least one order of magnitude in the concentration range of 0.025 to 0.050 M over the entire

wavelength range of its emission spectrum from 400-450 nm)°'2 Since the

Figure 2.

Dependence on concentration of the relative pulse height of PMP and PPO in a standard toluene solution at 100.

6

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

PVT emission is effectively absorbed at a distance of <10 m, plastic scintillators that contain PMP can now be used for the production of efficient small diameter plastic scintillating fibers.'°"2 Furthermore, PMP in polysiloxane showed good radiation resistance in an argon atmosphere)3 This is another

important feature because a radiation hardness of 106 rads is required to ensure a reasonable operational lifetime of the scintillating fiber. ACKNOWLEDGEMENTS

We thank Drs. K. F. Schmidt and R. Maschuw of the Institute of Nuclear Physics of the Kernforschungszentrum Karlsruhe for the measurements of the attenuation lengths. REFERENCES

Strähle, H., W. Seitz, and H. GUsten. "Struktur- und Losungsmittelabhangigkeit der Fluoreszenz von 1 ,3-Diphenyl-2-pyrazolinen," Ber. Bunsenges. Phys. Chem. 80: 228-294 (1976). GUsten, H., P. Schuster, and W. Seitz. "Organic Scintillators with Unusually Large

Stokes' Shifts," J. Phys. Chem. 82: 459-463 (1978). GUsten, H. and W. Seitz, "Novel Primary Solutes for Liquid Scintillation Counting," in Liquid Scintillation Counting: Recent Applications and Development, Vol.

1, C. T. Peng, D. L. Horrocks, and E. L. Alpen, Eds. (New York, Academic Press, 1980), pp. 51-57. GUsten, H. "PMP, A Novel Solute for Liquid and Plastic Scintillation Counting," in Advances in Scintillation Counting, S. A. McQuarrie, C. Ediss and L. I. Wiebe, Eds. (Edmonton, Canada: University of Alberta, 1983), pp. 330-339. Rinke, M., H. Güsten and H.J. Ache. "Photophysical Properties and Laser Per-

formance of Photostable UV Laser Dyes. 1. Substituted p-Quaterphenyls," J. Phys. Chem. 90: 2661-2665 (1986). Blume, H. and H. Güsten. In Ultraviolette Strahien, J. Kiefer, Ed., (Berlin: W. de Verlag, 1977), Chapter 6. Kowalski, E., R. Anliker and K. Schmid. "Criteria for the Selection of Solutes in Liquid Scintillation Counting. New Efficient Solutes with High Solubility," mt. J. Appi. Radiat. Isotopes 18: 307-323 (1967). Renschler, C.L., and L.A. Harrah. "Reduction of Reabsorption Effects in Scintillators by Employing Solutes with Large Stokes' Shifts," Nucl. Instr. Meth. Phys. Res. A 235: 41-45 (1985). Drexlin, G., H. Gemmeke, G. Giorginis, W. Grandegger, J. Kleinfeller, R. Maschuw, P. Plischke, F. Raupp, F.K. Schmidt, J. Wochele, B. Zeitnitz, E. Finckh, W. Kretschmer, D. Vötisch, J.A. Edgington, T. Corringe, N.E. Booth, and A. Dodd. "KARMEN: Neutrino Physics at ISIS," in Neutrino Physics, Proc. Intern. Work-

shop, Heidelberg, Oct. 20-22, 1987, H. V. Klapdor and B. Povh, Eds. (BerlinHeidelberg: Springer Verlag, 1988), pp. 147-152. D'Ambrosio, C., C. DaVia, J.P. Fabre, T. Gys, J. Kirby, H. Leutz, M. Primout, P. Destruel, M. Taufer, L. Van Hamme, and G. Wilguet, "Supercollider SCIFI Trackers," Proceedings of the Workshop on Scintillating Fiber Detector Development

PMP, A NOVEL SCINTILLATION SOLUTE

7

for the SSC, in press. Fermi National Accelerator Laboratory, Batavia, IL, (1988), pp. 14-16. Leutz, H., private communication. Destruel, P., M. Taufer, C. D'Ambrosio, C. DaVia, J. P. Fabre, J. Kirby, and H. Leutz, "A New Plastic Scintillator with Large Stokes' Shift," Nuci. Inst. Meth. in Phys. Res. A 276: 69-77 (1989). Bowen, M., S. Majewski, D. Pettey, J. Walker, R. Wojcik, and C. Zorn, "A New Radiation-Hard Plastic Scintillator," Nuci. Inst. Meth. in Phys. Res. A 276:391-393 (1989).

CHAPTER 2

Scintillation Counting of Harvested Biological Samples with Low-Energy Beta Emitters, Using Solid Scintillant Filters

C.G. Potter and G.T. Warner

INTRODUCTION

Liquid scintillation counting (LSC) of low-energy beta-emitting isotopes has an important role in the quantitation of biological samples. Although the level of radioactivity is low, the use of liquid scintillant gives rise to disposal problems because of the volume of organic solvents involved. A significant proportion of LSC work employs samples deposited on filter discs to measure the uptake of labeled compounds into cells and the binding of ligands to receptors. Placing many samples into vials for counting by LSC is a tedious procedure open to identification errors. Recently methods have been developed whereby a complete filter, bearing many samples, is placed in a thin plastic bag with only a small volume of scintillant (typically 10 mL/96-sample filter) and counted using a flat-bed scintillation counter.'2 This counter has a low background count rate and good counting efficiency, and the commercial version has a high sample capacity

and rate of throughput. This makes it particularly useful for lymphocyte assays using 3H-thymidine uptake to monitor proliferation, or to measure 51Cr

release in cytotoxic assays.3 The very small volume of scintillant required makes sample disposal easy as well as economical.4 Despite these advantageous features, there are other areas of potential application where the flat-bed counter is at some disadvantage. In particular, where

samples are soluble in scintillant, diffusion may result in the movement of activity away from the sample area. Movement may also occur if the sample is

not firmly bound to the filter and material is released by the addition and spreading of the scintillant while preparing the filter for counting. Another disadvantage not seen in a vial-counting liquid scintillation counter, is that of inter-sample interference, whereby activity in one sample is detected in an adjacent sample area. The degree of this "cross talk" can be reduced to less 9

10

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

than 0.01 °Zo for low energy isotopes by printing black lines on the filters. These absorb light emitted laterally as well as help identify and index during harvesting and placement of the filter on the plate."2 A further reduction in cross talk,

however, is desired when using thick filters, higher energy isotopes, and the most stringent requirements. The possibility that filters could be composed of scintillant has been previously described.5 This paper presents preliminary data on a technique using

special filters made of fibers coated with a thin layer of solid scintillant. Radioactive samples are deposited on the filter, dried in the usual way and counted without addition of a liquid scintillant. The beta emission excites the

solid scintillant and the events are counted using the flat-bed scintillation counter. Cross talk and background are reduced using this method and the samples have little tendency to migrate. Disposal of samples is further simplified because there is no liquid component. The properties of this type of filter are characterized for different solid scintillants as well as describe a heating method to increase the rather low tritium counting efficiency to a level useable for many applications. MATERIALS AND METHODS

Preparation of Filters

Conventional harvesting uses glass fiber depth filters of a mean porosity suitable for the samples required, e.g., Whatman GF/A or GF/C for cells, the thicker GF/B for membrane preparations, and GF/F for microorganisms. For the flat-bed scintillation counter (Betaplate, Pharmacia-Wallac OY, Finland), special reinforced versions of these filters were available, printed to facilitate orientation of the filter and reduce cross talk. As well as coating these filters with solid scintillant, the counter uses polypropylene depth filters in a variety of thicknesses and porosities. The solid scintillant was prepared from PPO (2,5-diphenyloxazole) dissolved in toluene (Analar grade) at concentrations up to 2.75''o, together with a spectral shifter (Bis-MSB, [l,4-di(2-methylsteryl)benzinel), added at 10 w/w of the PPO. Tetramethylbenzine (durene, from Aldrich, U.K.), which has a low melting point (55°C), was also added at levels of up to w/v of toluene. The filter sheets were immersed in the scintillant solution, drained and hung up to dry in a fume hood having a fast flow of air. After an hour. the filters were completely dry and virtually odorless. This method ensured that a thin coating of scintillant covered each fiber of the filter. 'When durene was used, it formed a solid solvent as a vehicle for the scintillant proper. Even for the finest porosity filters, there was little reduction in filtration capability with the addition of solid scintillant so that suitably cut sheets could be used in a cell harvester or filtration manifold. Similar filter sheets were prepared for tests using polyvinyl toluene/PPO, butyl-PBD, or anthracene in place of PPO as the scintillant.

SOLID SCINTILLANT FILTRATION

11

Cell Harvesting

The cells used for harvesting tests were K562 cells grown in RPMI and l2.5''o fetal calf serum. They were labeled overnight in the flask with 5Ci/mL 3H-thymidine or 0. ljCi/mL 35S-methionine. Microtitration plates were plated

with 0.2 mL aliquots at cell densities of 1-5 x l0 cells/mL and the aliquots filtered using a cell harvester (Skaatron, Norway) with a 10-sec water wash, followed by a flow of air for 5-10 sec. The filters were generally dried for 2 or 3 hours in an incubator at 370, or on a hot plate, or in a microwave oven. A

methanol wash could not be used as it washed away some of the solid scintillant.

Heat Treatment and Counting Preparation

After drying, filters were placed in thin plastic bags having a low-meltingpoint inner layer, laminated to an outer heat resistant plastic. Control filters had liquid scintillant added (Betaplatescint, Pharmacia Wallac OY, Finland) and the bag heat-sealed in the usual way. Samples with solid scintillant were counted directly or after heating to melt the scintillant, which could thus permeate the radioactive samples and be better coupled to the electron emisSiOn.6 Uniformity of heating was achieved in two ways, one by use of a microwave oven where the dielectric losses in the filter and solid scintillant produced heat enough to melt the scintillant after a few minutes exposure. The second method was to pass the dried filter in its bag through the heated rollers of an

office laminator (Coated Specialties Ltd., Basildon, U.K.). Thus the filter became compacted and the bag sealed, producing a thin, rugged set of samples ready for counting. The temperature required should not be high enough to produce shrinkage of the composite during lamination, otherwise registration of the samples with the counter support plate will be lost. Too low a temperature will not melt the scintillant. For polypropylene filters the temperature may be made high enough to melt the plastic also producing a transparent set of samples. These may have a pleasing, solid appearance, but the melted plastic can move while under the rollers giving unreliable results. RESULTS

Glass Fiber Filters

Cells were harvested onto glass fiber filters (Pharmacia-Wallac) with and without solid scintillant. In the first experiment, the amount of PPO in toluene for the coating solution varied between 5.5 and 27.5 g/L also with or without

5% durene. Table 1 shows that the highest amount of PPO provided the highest efficiency, which was enhanced by heating and passing the filter and bag through heated rollers. Addition of durene generally increased efficiency a

12

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 1. Relative Counting Efficiencies of Solid Scintillant, with and without Durene, for

Different Amounts of PPO + BisMSB (9:1 wlw), Using Samples (n = 12) of 3H-Labeled Cells Filtered Onto Scintillant Coated Glass Fiber Sheets Heatingb %PPO + BisMSB Rel. Effc (% ± SD) ± Durene 0.55 5.1 ± 0.4

-

+ +

1.10

1.65

-

2.75

-

5.6 ± 0.7

+

+ +

-

-

+ -

+ +

2.20

+

4.3±0.5 9.4±1.5 12.3±1.3

+

-

+

+

-

7.2±0.5 12.7±2.4 17.9±1.2 5.9 ± 0.9

6.7±0.6

18.9±1.5 23.1±1.9 7.0 ± 0.9

+

8.5±0.6 20.5±2.0

+ +

-

+

26.6 ± 2.0

-

+ -

10.3 ± 1.7

+ +

+

10.7±0.7

26.6 ± 3.5 26.1 ± 1.8

appQ + B1sMSB dissolved in toluene, with or without 5% durene and used to soak filters dried and used in a filtration assay of 3H-thymidine-labeled K562 cells. bHeating by the heated rollers of oflice laminator.

CRelative efficiency, compared with parallel samples counted in the flat-bed scintillation counter using BetaplateScint cocktail.

little and could partly replace PPO. At the higher quantities of PPO or with

durene present, the filters became stiffer and filtration became a little impaired. In another experiment (Table 2), butyl-PBD showed a similar effi-

ciency to PPO before heating, but it was reduced after passing the filters through the heated rollers. Addition of durene did not improve efficiency either before or after heating. Table 3 shows the results of an experiment comparing 2% PVT/toluene with 2.75% of either butyl-PBD or PPO, both with and without durene, (increased to 10%) in toluene. Again there was no difference or reduction in counting efficiency for butyl-PBD and PVT, even when using microwave heating, which would not produce the compaction made by heated rollers. Anthracene with or without durene was also tried and

heating efficiency improved slightly. One noteworthy aspect was that the energy spectrum for anthracene was shifted to higher levels compared with the others scintillants tested. 35S-Iabeled Samples

For cells labeled with 35S-methionine and harvested onto plain or printed glass fiber filters with PPO/BisMSB/durene solid scintillant, the efficiency

SOLID SCINTILLANT FILTRATION

13

Table 2. Relative Counting Efficiencies of Solid Scintillant, with and without Durene and for Different Amounts of Butyl-PBD, Using Samples (n = 12) of 3H-Labaled Cells Filtered Onto Scintillant-Coated Glass Fiber Sheets, with and without Heating Relative Counting Efficiency of BetaplateScint (% ± SD) Heatingb % of Butyl-PBD8 - Durene + Durene 5.4 ± 1.8 0.5 3.8 ± 1.3 3.4 ± 1.4 4.0 ± 1.2 +

-

1.0

+

-

1.5

+

-

2.0

+

-

2.5

+

8.5 ± 2.6 6.9 ± 2.1

7.8 ± 2.9 7.0 ± 2.3

8.1 ± 2.4 6.7 ± 2.2

9.0 ± 2.5 8.1 ± 2.2

13.0 ± 3.5 10.7 ± 3.0

7.2 ± 1.8 7.0 ± 2.0

10.8 ± 2.6 8.0 ± 1.8

8.3 ± 2.0 7.3 ± 1.9

aBUl.pBD dissolved in toluene, with or without 5% durene, and used to soak glass fiber filter, which was dried and used in a filtration assay of 3H-thymidine-labeled cells. bHeating by the heated rollers of an office laminator.

Table 3. Relative Counting Efficiencies of Solid Scintillants with Different Compositions, Coated on Glass Fiber Filters and Used to Harvest 3H-Labeled Cells ± Dureneb

Fluor

Heating8

Liquid scintillant control

None

-

Polyvinyltoluene plastic scintillant

None

-

2% in toluene Butyl-PBC

+ roll None

2.5% in toluene

+ roll None

+ roll PPO + 10% w/w BisMSB; 2.75% in toluene

None

+ roll None

+ roll a

-

+ + +

+ + +

Rel. Eff.c (% ± SD) 100.0 ± 8.8

N

20

8.4 ± 0.5 8.5 ± 0.6 4.3 ± 0.5

20

8.9 ± 0.7 8.9 ± 0.4 7.6 ± 0.4 9.0 ± 0.4

10

9.1 ±0.7 7.3 ± 0.6

7.7 ± 0.7 10.8 ± 1.4 19.2 ± 1.7 7.5 ± 1.3

20

10.3±0.8

22.7 ± 2.5

= microwave oven heating (400W for 120 sac); roll = microwaved samples passed through office laminator (>90°C). bDurene as 10% of the solution used to make the solid scintillant filters. cRelative efficiency compared with parallel samples counted in the flat-bed scintillation counter using BetaplateScint cocktail.

14

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 4. Relative Counting Efficiencies, Cross Talk and Background Count Rates for Glass Fiber Filters Coated with Solid Scintillant (PPO/BisMSB/durene) and Used to Harvest 35S-methionine-Labeled Cells, with and without Subsequent Heating

Plain Glass Fiber

Treatment Solid scintillant

set 1 (n = 12) Set 1 + liquid

bg

cpm

4.3

42181

11.1 128590

9.2 5.0

0.024 (10.0 cpm)

Printed Glass Fiber cpm xtW CV% <0.001 3.0 50065 6.8

bg

(0.3 cpm)

0.020 14.8 126234

4.5

43746

8.8

5.1

by laminator Relative efficiency Increase in relative efficiency with heat

53160

8.0

0.011

(14.3 cpm) 39.7

32.8 3.9

set2(n = 12) Set 2 + heating

xt%

(25.7 cpm)

scintillant Relative efficiency Solid scintillant

CV%

0.007 (3.0 cpm)

2.2

0.002 (1.1 cpm)

2.6

53463

4.8

<0.001 (0.4 cpm)

68628

4.9

40.0

50.1

22

28

0.002 (1.6 cpm)

across talk-mean of 12 adjacent blanks less backgrounds.

was only slightly increased by heating. Table 4 shows that the efficiency reached up to 50°/s of glass fiber using liquid scintillant. Background count rates were also low, but in addition, these filters showed good inter-sample cross-talk characteristics even without printing, although with the addition of printed lines the cross talk fell to very low levels which are not easily measured, i.e., <0.001 °/o. Similar experiments with 3H-labeled cells gave no measurable cross talk, with or without printing. Polypropylene Filters Samples of several types of polypropylene depth filters were obtained from three suppliers. Each was tested by harvesting cells labeled with 3H-thymidine onto the polypropylene filters with solid scintillant and comparing them with controls harvested on glass fiber or polypropylene filters with liquid scintillant. Any reduction in the effectiveness of filtration, together with any self-

absorption or quenching, should produce a reduced count rate. As before, filters bearing solid scintillant were counted directly or after heating by microwave or passing through heated rollers. Some filters were found to have too coarse a mesh to filter out the majority of cells and others were too flimsy to handle satisfactorily. The properties of some suitable filters are illustrated by data on three examples, designated A, B, and C. Table 5 shows that the count rate for the polypropylene filters with liquid scintillant was slightly reduced compared to glass fiber. Counting solid scintillant filters directly gave relative efficiencies between 7.4 and 17.9°/o, which was increased by microwave radiation. A further increase was produced by hot rolling, to a maximum of 33.5°/a for sample B. At temperatures high enough to melt the polypropylene, counting efficiency was reduced and variation increased. Background count rates

SOLID SCINTILLANT FILTRATION

15

Table 5. Relative Efficiencies (% ± CV) for 3H-Labeled Cells Filtered onto Solid Scintillant Polypropylene Filters Treatment Type of Filter (Porosity) A(2.5) B(1O) C (?) Liquid scintillanta Solid scintillantb

91.8 ± 5.5

+ microwave + microwave and heated rollers

23.6 ± 7.8

13.3 ± 8.1 16.6 ± 8.4

78.5 ± 6.8 17.9 ± 4.9 24.8 ± 15.7 33.5 ± 8.4

89.7 ± 7.5 7.4 ± 10.6 9.0 ± 12.8 21.3 ± 11.5

aBetaplatescint CV for glass fiber was 5.9%. b275010 PPO/BisMSB, 10% durene/toluene.

were very low unless the samples were melted, which probably increased optical cross talk between the pairs of photomultipliers.

Linearity with Cell Density Using glass fiber filters and liquid scintillant, linearity of count rate with cell density is possible over a wide range.3 In order to test whether this was also true for solid scintillant filters, microtration plates were set up with 6 replicate wells at cell densities ranging from approximately l0 to 4 X 1O cells/well and harvested on glass fiber filters with or without solid scintillant. Table 6 shows that linearity was maintained over this range, before and after passing the samples through heated rollers.

Uniformity of Samples A complete microtitre plate was set up with approximately 2.5 x l0 cells/ well and harvested using a glass fiber filter, coated with solid scintillant, giving a matrix of 6 x 16 samples. Twelve more samples in another plate were also harvested using a standard glass fiber filter with liquid scintillant and gave a mean of 107,800 cpm with a 2.2Wo coefficient of variation (CV). The solid scintillant filter was counted directly (mean of 10,995 cpm) and after passing through heated rollers (mean of 18,203 cpm). Table 7 shows that the normalized means for each row exhibit a significant positive regression, either when Table 6. Linearity of Relative Counting Efficiency for Different Cell Densities Harvested

onto Glass Fiber Filters Coated with Solid Sclntlllant.a Expressed as Solid Scintillant Filter Count Rate/Liquid Scintillant Control Count Rate (% ± SD) Cell Density x 103/Well Before Heating After Heating 380 6.49 ± 0.70 11.19 ± 1.15 114 6.14 ± 0.83 10.97 ± 2.03 38 6.73 ± 1.46 11.82 ± 2.62 11.4 6.18 ± 0.66 10.31 ± 1.71 3.8 5.79 ± 1.15 9.47 ± 2.16 1.1 6.98 ± 1.65 12.25 ± 3.00 a275% PPO/Bi5MSB/5% durene in toluene.

16

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 7. Regression for Normalized Data (y = a + bx) of Means of Rows (n = 16, 6 In

each row) for Cells Filtered onto Glass Fiber, and for Glass Fiber or Polypropylene Filters Coated with Solid Scintillant Spiked with 3H-hexadecane, Before and After Heating.

Type of Filter Glass fiber

+ cells

± Heating

+

a ±SD 90.5 84.8

± 2.6 ± 2.2

94.4 100.4

± 1.1 ± 1.1

± 2.6 ± 5.7

Glass fiber +

-

3H-hexadecane

+

Polypropylene + 3H-hexadecane

-

86.1

+

85.5

b ±SD

t

p

r 0.74 0.90

± 0.27 ± 0.23

4.1

7.8

<0.01 <0.001

0.72 ± 0.12 -0.05 ± 0.11

6.1

<0.001

0.5

NS

0.85 0.13

6.1

<0.001

0.85

2.9

<0.02

0.61

1.12 1.78

1.64 1.70

± 0.27 ± 0.59

counting the filters directly, or after heating. The average CV for the rows, which allows for the effects of this regression, was 7.9% for direct counting and 9.5% after heating. These values and the regression when using glass fiber filters, show that the performance falls short of standard methods using liquid scintillant. In contrast, in another test, using polypropylene filters with solid scintillant, a single set of samples from 12 wells was harvested and the sample

was dried and passed through the hot rollers. These gave a CV of 4.5%, compared with controls on a standard glass fiber filter with liquid scintillant, which gave a CV of 5.6°lo. Therefore, these polypropylene filters were capable of as good a consistency in any row as for the standard glass fiber filters using liquid scintillant, although any regression according to the sample row detracts from the technique.

It is possible that these problems of variation could be explained if the deposition of solid scintillant on the filter was not homogeneous. To test this, a standard glass fiber filter was prepared with solid scintillant spiked with 3Hhexadecane. After preparation, the filter was dried and counted, before and after heating. The results, also in Table 7, showed a clear positive regression according to the sample row when counted direct, but the regression became insignificant after heating. The count rate also decreased, indicating that some quenching or self absorption of the activity had occurred which was associated with compaction of the glass fiber filter while melting the solid scintillant. A similar experiment with a thin grade of polypropylene showed no decrease in

count rate after heating, and the regression was also unchanged (Table 7). These results are in contrast to the samples derived from cells which showed an increase in count rate after heating. Presumably the increased counting efficiency due to scintillant permeation of the samples is greater than any reduction due to compaction-produced self-absorption. The variation in count rate for harvested cells therefore appears to be associated with different amounts of deposited solid scintillant and is probably due to the solid-scintillant solu-

tion pooling at the lower end of the filters during the drying process. It is therefore evident that any production of these filters should minimize this variation.

SOLID SCINTILLANT FILTRATION

17

DISCUSSION

The experiments described, indicate that a useful efficiency can be obtained for solid scintillant filters especially when the scintillant is melted using the hot rollers of an office laminator. Further improvements occur when polypropylene filters are substituted for glass fiber. In general, the higher the efficiency the smaller the coefficient of variation obtained, and satisfactory results were

obtained with a PPO/durene/polyropylene combination. There is a rather complex interaction between the increased efficiency obtained by filter compaction, which may bring particles of the sample closer to the solid scintillant, and any possible losses of light output, due to the increased absorption by the filter/scintillant matrix. Reduced optical efficiency produced by compaction

was most apparent where the scintillant did not melt at the temperatures achieved during the hot rolling. This reduction was greatest for the PVT plastic coating and less for crystaline butyl-PBD. For anthracene, however, there was an increase in counting efficiency. Perhaps, because the energy output is higher for this scintillant, fewer events are lost through quenching, while with the proximity of the compacted sample to the solid scintillant, more of the sample may be exposed to the scintillant. Any reduction in efficiency produced by compaction appeared to be less deleterious using polypropylene. When polypropylene filters bearing cells were used with liquid scintillant the polypropylene is a slightly less effective filter. Rather than being due to a poor optical performance, preliminary experiments show that these filters, with or without solid scintillant, trap a smaller proportion of material harvested than do glass fiber filters, even though the porosity of 10 should retain whole cells entirely. It is likely however, that cells are rapidly lysed when washed from the microtitre plate with water. Reduced retention of this partial lysate is due to differences in the surface properties of the polypropylene or the solid scintillant (which are both hydrophobic), as compared with the (hydrophilic) surface of glass fiber. It is possible that the surfaces of other filters and solid scintillants will be developed for other porosities and may retain nearly 100% of the sample. Indeed, other scintillants and solid solvents that increase the efficiency still further may also be developed, although the level described here would be adequate for many applications. The technique is useful in that it reduces background count rate and cross talk as well as eases disposal. Furthermore, the lack of sample movement during preparation or its diffusion afterwards, makes the counting of many filtered particulate samples possible using the flat-bed scintillation counter. ACKNOWLEDGEMENTS

We wish to thank Mrs. A. C. Potter for technical assistance and Mr. P. Harrison for suggesting durene as a solid solvent. We also thank PharmaciaWallac for financial support.

18

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

REFERENCES

Warner, G.T. and C.G. Potter. "A New Design for a Liquid Scintillation Counter for Microsamples Using a Flat-Bed Geometry." mt. J. App!. Isot. 36:819-821 (1985).

Potter, C.G., p.T. Warner, T. Yrjonen, and E. Soini. "A Liquid Scintillation Counter Specifically Designed for Samples Deposited on a Flat Matrix," Phys. Med. Biol., 31:361-369 (1986).

Potter, C.G., F. Gotch, G.T. Warner, and J. Oestrup. "Lymphocyte Proliferation and Cytotoxic Assays Using Flat-Bed Scintillation Counting," J. Immunol. Methods, 105:171-177 (1987).

Warner, G.T. and C.G. Potter. "A New Scintillation Counter Design Eases Vial Disposal Problems," Health Phys. 5 1:385 (1986). Warner, G.T. and C.G. Potter. "Method of, and Apparatus for, Monitoring Radioactivity," U.S. Patent Specification 4,298,796 (1981).

Potter, C.G. and G.T. Warner. "Carrier for at Least One Beta-Particle-Emitting Sample to be Measured in a Scintillation Counter," Swedish Patent Specification 8801879-1 (1988).

CHAPTER 3

Dl-ISOPROPYLNAPHTHALENEA New Solvent for Liquid Scintillation Counting

J. Thomson

INTRODUCTION

Although there have recently been a few significant developments in the field of solvents suitable for use in liquid scintillation counting (LSC), the traditional solvents xylene and toluene are still used. A review of the evolution of LSC Solvents (Figure 1) shows that after an initial period of development in the 1950s and '60s very little has happened until quite recently. Toluene and xylene were the initial solvents of choice due to their commercial availability and suitability.' Dioxan, on its own and in conjunction with naphthalene, was also widely used at this time due to the ability of this system to accept quantities of aqueous sample, thus expanding the application.

After a brief period of inactivity, the '60s saw the emergence of some alternative solvents of which decalin and aromatic petroleum distillates (C9 to C12) were the most notable. These partly supplanted toluene and xylene because of their lower cost. However, the development of emulsifying cocktails in the '60s further established toluene and xylene as the preferred LSC solvents. Then, in the 70s changes in market trends saw the introduction of a new type of emulsion cocktailfor Radio-immunoassaybased on pseudocumene. This change coincided with an increased awareness of safety brought about by the restrictions imposed on the handling and use of certain hazardous solvents. The search was now on for safer alternatives to toluene and xylene that had equivalent or better characteristics with respect to their LSC performance. Another factor which became important at this time was the adverse effect on the environment of waste chemicals. Formulators gradually became aware of these factors and realized that a very different type of solvent would

be required to satisfy the changing demands of both the market and the environment. Thus the main driving force behind recent developments in sol19

Figure 1.

30-

40-

50 -

60 -

1950

(C9C12)

Mixed Alkyl Benzenes

1955

1960

1965

1970

Dioxan / Naphthalene

Xylene

Evolution of LSC solvents.

1945

Toluerie

Tritium efficiency, %

1975

1980

1985

1990

Polyalkylbenzenes

Pseudocumene

Phenyl-o-xylylethane

Di-isopropyl naphthalene

DI-ISOPROPYLNAPHTHALENE FOR LIQUID SCINTILLATION COUNTING

21

vents for LSC has not been the search for more efficient solvents, but rather

the requirement to substitute less hazardous and more environmentally friendly solvents for those previously used. In the early '80s the first attempts to resolve these problems involved the use of polyalkylbenzenes which met some but not all of the requirements. The major failings were their low counting efficiency, lack of biodegradability and their toxicity. At Fisons we were active in this field of development. Indeed we used polyalkylbenzene solvents in emulsion cocktails simply because they were the best compromise available. During the course of our investigation about 30 solvents were evaluated for their potential in LSC, but for one or more of the above reasons they all proved to be unsuitable. In due course we became aware of the existence of Di-isopropylnaphthalene (DIN) and quickly realized that here was a solvent that could be ideally suited to LSC.2 DIN is a mixture of positionisomeric di-isopropylnaphthalenes and is a product of Rutgers-Kureha-Solvents Ltd. In order to fully appreciate the benefits afforded by DIN consider an ideal LSC solvent.

DEFINING THE IDEAL

In evaluating solvents for use in LSC our start point was to define the characteristics of an ideal LSC solvent. Based on our own experience and comments by other workers in the field we listed the following essential attributes: high flash point low vapor pressure odorlessness low toxicity and irritancy

no permeation through plastics biodegradability good fluor solubility low photo- and chemiluminescence high counting efficiency (tritium) good colour and chemical quench resistance

The ideal LSC solvent would obviously satisfy all these requirements but hitherto no solvent has. The evaluation of DIN against these criteria will indicate how it equates to the ideal LSC solvent. Pseudocumene is used in the comparison because it is typical of the solvents currently in wide use in LSC cocktails.

22

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

EVALUATION OF DIN VS THE IDEAL

Flash Point The 148°C (298°F) flash point of DIN makes it safe to use on the laboratory bench since the flammability risk is minimal. DIN is classified as a nondangerous product in accordance with national and international traffic regulations because the flash point exceeds 100°C. By comparison pseudocumene with a flash point of 52°C (125°F) is classified as flammable and should be handled accordingly. Vapor Pressure

DIN has a very low vapor pressure (1 mm Hg at 25°C) thus ensuring that there is a low vapor concentration at 25°C. This means that there would be very little build up of DIN vapor in the event of a spillage in an enclosed space. By comparison pseudocumene has a vapor pressure twice (2 x) that of DIN.

Odor

DIN is virtually odorless and as such makes it pleasant to work with. Pseudocumene, on the other hand, has a highly aromatic penetrating odor, which, if inhaled in appreciable amounts, can cause headaches and narcosis. Toxicity and Irritancy Based on the acute data (LD50 = 5600 mg/kg oral rat). DIN is not classified as toxic and imposes no acute health hazard to man. DIN is widely used in the manufacture of carbonless copy paper (NCR paper) and consequently its toxicological properties have been extensively studied. A considerable amount of work (to EEC directives) has been completed on DIN and a summary of the toxicological data shows DIN to have the following reactions: acute toxicity (oral - rat) nontoxic acute toxicity (dermal - rat) = nontoxic skin irritation = nonirritating eye irritation = nonirritating skin sensitization = nonsensitizing toxicokinetics nonaccumulative mutagenicity = nonmutagenic cancerogenicity = noncarcinogenic teratogenicity = nonteratogenic acute toxicity to fish = nonharmful to fish bioaccumulation (fish) = nonbioaccumulative

A copy of "Toxicological and Physico-Chemical Studies on Diisopropylnaphthalene" is available from the author upon request. The toxicity

of pseudocumene is well known and its principal hazards are that it is a

DI-ISOPAOPYLNAPHTHALENE FOR LIQUID SCINTILLATION COUNTING

23

primary skin irritant, is irritating to the eyes, and on inhalation causes respiratory irritation together with central nervous system depression. Plastic Permeation

In both short- and long-term studies DIN has been shown not to permeate through polyethylene counting vials. Indeed, in a long-term study there was no loss of DIN from a polyethylene vial during a 12 month storage test at room temperature. Pseudocumene, although better than toluene and xylene, is nevertheless steadily lost from polyethylene counting vials. In practice, perme-

ation of the solvent through the vial wall has two effects. Firstly, loss of solvent increases the effect of quench in the vial. Secondly, the solvent and fluors penetrate the vial wall causing the vial to become an additional solid scintillator. This solid scintillator in conjunction with the external standard produces an additional spectrum. This solid scintillator spectrum adds to the "Compton" spectrum produced by the interaction of the external standard with the cocktail solvent. Therefore, there is a gradual change in shape of the overall spectrum. This particularly affects the external standard channels ratio (ESCR) and is manifest as a continual change in the ESCR value. This phenomenon is most evident with ESCR, but the newer parameters (special quench parameter and H-number) are not significantly affected. With ESCR this leads to an underestimate of the efficiency and hence an overestimate of the dpm. The effect is known as the "plastic vial" effect and is shown in Figures 2 and 3. Biodegradability By comparison with other aromatic solvents DIN solvent has the remarkable

property of being biodegradable in its own right. DIN is described as being greater than 80% biodegraded after 28 d at 4 ppm available oxygen and is therefore classified as biodegradable according to EEC directive 79/831. Annex VII. An independent evaluation by the Severn Trent Water Authority (U.K.) found Optiphase Hi-Safe II (DIN-based emulsion cocktail) to be readily biodegradable by the ISO 7827-1984 (E) method, achieving a degradation level of greater than 80% in 2 d. Greater than 80% degradation in 28 d is the standard necessary for classification as readily biodegradable. A copy of this evaluation is available from the author upon request. However, it is essential that your radiological safety officer and local water authority are consulted to obtain permission for drain disposal before embarking upon any particular course of action. Fluor Solubility All the commonly used fluors are soluble in DIN solvent with PPO having particularly good solubility. Figure 4 shows the solubilities of selected fluors in

Figure 2.

0

40 -

50 -

60 -

70 -

80 -

90 -

100

2

3

4 5

6

7

Permeation of various solvents (at 20°C in 20 mL polyethylene vials).

1

% of original weight remaining

8

9

10

Time (weeks)

Pseudocumene

Di-isopropylnaphthalene

Figure 3.

0

35 -

40-

45 -

50

55 -

60 -

10

15

20 25

30 35

40

45

50

55

60

65

Pseudocumene

Time (hours)

Di-isopropylnaphthalene

Plastic vial" effect: variation in tritium efficiency with time for various solvents in 20 mL polyethylene vials.

5

Tritium efficiency, /0

26

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

FLUORS

Max. concentration, gil

PPO

100

PBD

9.9

Butyl-PBD

40

Bis-MSB

1.8

TP

3.6

POPOP

0.5

Dimethyl-POPOP

1.0

Figure 4.

SolublIlty of various fluors in di-isopropylnapthalene at 15°C.

DIN at 15°C, and easily exceeds the optimum concentration as determined by a simplex optimization. Pseudocumene is also a good solvent for all the common fluors. Luminescence

DIN solvent is compatible with all the conventional toluene-based alkaline tissue solubilizers and any chemiluminescence should decay within 30 mm. Any induced photoluminescence will also decay within 30 mm under normal counting conditions. Pseudocumene from some sources will require purification to remove those impurities which can give a color reaction with alkaline tissue solubilizers.

Counting Efficiency (Tritium) When DIN is used as the solvent in both lipophilic and hydrophilic cocktails then the resulting cocktails exhibit a superior counting efficiency when compared with other types of cocktail. This is especially important when counting to statistical limits since those limits will be reached quicker with DIN based

cocktails thus reducing instrument time. The following graphs, Figures 5 through 7, illustrate the superior detection efficiency obtained with tritiumlabeled water samples. The DIN based cocktail, Optiphase Hi-Safe II, used in this comparison is a product of Pharmacia-Wallac

200

Figure 5.

28

29-

30 -

31 -

33 32 -

34 -

35-

36 -

37 -

38 -

39 -

40-

41 -

42

1'

400

500

600 700

Efficiency vs L water added (sample added to 5 rnL cocktail).

300

H3 efficiency, %

800

900

Traditional xylene based high efficiency emulsion gel cocktail

OPT/PHASE HI-SAFE II

1000

WateraddedjI

300

400

500

600

700

Figure 6. Efficiency vs L water added (sample added to 5 mL cocktail).

200

I

H3 efficiency, %

800

900

1000

Long chain alkyl benzene based cocktail

OPTIPHASE HI-SAFE II

*Wateradded,pi

200

300

400

500 600

700

Figure 7. Efficiency vs L water added (sample added to 5 mL Cocktail).

28

29 -

30 -

31 -

32 -

33 -

34 -

35 -

36 -

37 -

38 -

39 -

40 -

41 -

42 -

I

H3 efficiency, %

800

900

1000

Wateradded,.tI

30

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Quench Resistance

DIN exhibits an exceptional resistance to chemical quench as shown in Figure 8. This is important where samples are prepared with acids, solvents etc., which introduce unacceptably high levels of chemical quench with other solvents. DIN also has resistance to color quenching (Figure 9), which is necessary when working with colored samples, e.g., urine and serum. Pseudocumene has good resistance to both chemical and color quench, but has been found to be not as good as DIN.

DIN-BASED EMULSION COCKTAILS

DIN can be formulated into emulsion cocktails which achieve a greater degree of safety without compromising performance. This is aptly demonstrated in the Optiphase Hi-Safe range of cocktails which have seen wide acceptance in the U.K. and Europe over the last few years. Taking a closer look at two of these products: Optiphase Hi-Safe H©

This cocktail possesses all of the safety features previously cited about DIN solvent together with the expected very high counting efficiency and the ability to incorporate a diversity of sample types (Figure 10). Optiphase Hi-Safe III©

This cocktail combines the capability of accepting large volumes of normal samples with the ability to accept high-ionic strength solutes in moderate to good volume (Figure 11). This makes Optiphase Hi-Safe III almost unique among currently available cocktails. As expected Optiphase Hi-Safe 3 has all the inherent advantages afforded by being a DIN-based cocktail. Optiphase Hi-Safe III has the additional capability of accepting sample types which other cocktails find difficult or impossible to accommodate.

CONCLUSION

Di-isopropylnaphthalene is a significant improvement over existing LSC solvents. Its low vapor pressure, low flammability, low plastic permeability, and low toxicity make it safe and pleasant to handle. It is a more environmentally acceptable product because of its rapid and extensive biodegradability. Lastly and by no means least it is unexcelled in its LSC performance. Therefore DIN comes closest to matching the ideal LSC solvent than any other currently known solvent.

H3 Efficiency, %

Figure 8.

25

Pseudocumene

50 75

Di-isopropylnaphthalene

100

microlitres Carbon tetrachloride added

-*

Chemical quench comparison for di-isopropylnapthalene and pseudocumene (JLL carbon tetrachloride added to 10 mL cocktail, cocktails = solvent + 4 gIL PPO/0.1 g/L bis-MSB).

0

50-

55

1'

C)

Figure 9.

100

150

200

Pseudocumene

250

300

350

Di-isopropylnaphthalefle

400

500 -* 450 microtitres Dimethyl yellow added

Color quench comparisonf or di-isopropytnapthalene and pseudocumene (L of 0.001% dimethyl yellow solution added to 10 mL cocktail, cocktails = solvent + gIL PPOIO.l g/L bis-MSB).

0

20 -

25-

30 -

35 -

40-

45 -

50

H3 Efficiency, %

50 -

55

I

DI-ISOPROPYLNAPHTHALENE FOR LIQUID SCINTILLATION COUNTING

Temperature, C

SAMPLE

L

33

10

15

20

25

Water

2.5

2.5

2.6

2.6

0.15M Sodium chloride

3.4

3.4

3.8

3.8

Phosphate saline buffer 0.01 M

3.4

3.4

3.4

3.4

0.05M Tris-HCI

3.4

3.4

3.4

3.4

0.1M Hydrochloric acid

9.0

10.0

10.0

10.0

10% Sucrose

3.8

3.8

3.8

3.8

0.2M Sodium hydroxide

4.0

4.6

5.0

5.0

0.1M Sodium hydroxide

4.0

4.0

4.0

4.4

Urine

2.2

2.2

2.2

2.2

Serum-canine

1.0

1.0

1.0

1.0

20% Sucrose

6.6

6.6

5.8

5.8

0.lMAmmonium sulphate 0.05M Sodium acetate

3.4

3.4

3.4

3.6

5.8

5.6

5.0

5.0

0.04 Disodium phosphate

8.0

8.0

8.6

8.6

5mM Hepes

2.8

2.8

2.8

2.8

0.lMTris-5OmM EDTA

5.0

5.2

5.8

5.8

0.2M Ammonium acetate

2.4

2.4

2.6

2.8

i.OM Sodium hydroxide

2.0

2.0

2.2

2.4

10% Trichloroacetic acid

2.2

2.3

2.3

2.3

8aM Urea

1.1

1.1

1.1

1.1

Optisolve

1.0

1.0

1.0

1.0

Figure 10. Optiphase Hi-Safe II sample acceptance (values ImLI determined by addition of sample to 10 mL of cocktail).

ACKNOWLEDGEMENTS

The author would like to thank the following staff at FSA for their help and support: Mr. H.S. Wagstaff, Mr. B. Illingworth, Mr. R.W.R. Carss, and Mr. G.A. Giblin.

34

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

SAMPLE

Temperature, C 10

15

20

25

Water

10.0

10.0

9.0

9.0

Hepes

10.0

10.0

9.5

7.5

Tris HCI 50mM

10.0

10.0

10.0

10.0

Ringers

10.0

10.0

8.5

8.0

0.5M Phosphate saline buffer

9.0

7.0

6.5

5.5

0.1 M Phosphate saline buffer

3.0

3.0

8.0

8.0

Sea water

3.0

3.0

7.5

7.5

1 .OM Sodium acetate

7.0

6.5

6.0

5.5

4.OM Sodium hydroxide

1.9

1.9

1.9

2.0

40% Caesium chloride

6.5

6.0

5.5

5.0

10% Trichloro acetic acid

3.5

6.5

6.5

6.5

Urine

3.0

3.5

8.0

8.0

4.OM Ammonium formate

2.5

2.5

2.0

1.75

1 .OM Ammonium formate

7.5

6.0

5.5

5.0

1.OM Sodium chloride

3.0

7.5

7.5

7.5

2.OM Nitric acid

7.5

6.0

4.0

4.0

0.5M Potassium di-hydrogen orthophosphate 8.OM Urea

7.5

6.0

5.0

5.0

3.25

7.0

7.0

7.5

0.5M Ammonium sulphate

5.0

3.75

3.5

3.25

Figure 11.

I

I

Optiphase Hi-Safe III sample acceptance (values [mU determined by addition of sample to 10 mL of cocktail).

REFERENCES Birks, J.B. The Theory

of Practice of Scintillation Counting

(Pergamon Press,

1964), pp.272-279.

Thomson, J. "Scintillation Counting Medium and Counting Method", U.S. Patent 4,657,696, April 14, 1987; European Patent 0176281, September 11, 1985."

CHAPTER 4

Safe Scintillation Chemicals for High Efficiency, High Throughput Counting

Kenneth E. Neumann, Norbert Roessler, Ph.D., and Jan ter Wiel

ABSTRACT For the last 35 years, the choice of solvent, fluor, and emulsifier for scintillation cocktails has been dictated by the need for efficient energy transfer between the beta electron and the final photon-emitting species. The modern scintillation counting laboratory has several additional requirements which make the design of a scintillation cocktail more critical than ever before: a wide variety of samples, the chemical environment of the radiolabeled sample, and the trend toward microvolurne analysis. Furthermore, new environmental regulations are dramatically increasing disposal Costs, making the use of high flash point, environmentally safe cocktails quite attractive. These factors combine to create the need for new types of scintillation chemicals capable of safe, high efficiency, low-volume counting. Additionally, the development of many new beta counting assays requires scintillators that are directly suitable for these applications. These requirements have lead to the development of a new generation of liquid cocktail. Data are presented indicating that this formulation provides superior counting performance, while offering a high degree of safety. Additionally, the novel characteristics of this cocktail are discussed with regard to new instrument technologies.

INTRODUCTION

The design of commercial liquid scintillation cocktails has, for many years,

been based on the technical requirements of those investigators using the chemicals. These include high radionuclide counting efficiencies, large sample holding capacity, and superior resistance to quench.' Issues of chemical safety, storage, and disposability were assigned a relatively minor importance. The

solvent used as the cocktail base plays a critical role in determining overall cocktail performance with regard to the above parameters. Acceptable results have classically been achieved with lower order benzene derivatives such as toluene, xylene, and 1 ,2,4-trimethylbenzene.2 In recent years, occupational health and safety and environmental issues have attracted much attention, both in the media and the scientific community. State and federal regulations regarding shipping, use, storage, and dis35

36

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

posal of hazardous chemicals are increasingly stringent. Scintillation cocktails, because of the solvents cited above, are recognized as potential environmental and health hazards As a result, considerable effort has been made to develop scintillation chemicals which offer greater safety.3 One of the first significant breakthroughs involved the use of a higher order benzene derivative as the solvent. Optifluor, manufactured by Packard Instrument Company, is based on a long-chain phenylalkane. The use of this solvent permitted the cocktail to be designated as nonflammable, nonhazardous, and

drain disposable. However, because of the solvent low scintillation yield, counting efficiencies are less than those of conventional cocktails. The cocktail is also somewhat less quench resistant. The most recent development is a cocktail possessing all of the characteris-

tics necessary for superior counting performance. Ultima Gold, developed by Packard, is based on a new solvent known as diisopropylnaphthalene.4 This solvent also has a very high flash point. As a result, Ultima Gold has been classified as nonflammable. The cocktail has been approved for drain

disposal, thus reducing many of the costs associated with LS counting. Finally, Ultima Gold presents no unusual, chronic, or severe health hazards to laboratory personnel, and has no noxious or unpleasant odors associated with it. Of greater significance is that diisopropylnaphthalene, being a derivative of naphthalene, retains the properties of a primary scintillator, while existing in a liquid form. Additional fluors and emulsifiers are readily miscible in the solvent, creating a high performance universal cocktail. The resulting formulation provides high sample capacities, with excellent quench resistance. Due to the scintillating nature of the solvent, excellent counting efficiencies are

maintained even with high sample loads or the presence of tissue solubilizers. The scintillating properties of diisopropylnaphthalene have broad implications for cocktails based on this solvent and, more importantly, liquid scintillation counting equipment. Experimental evidence suggests that the pulse shapes and fluorescent decay characteristics of cocktails based on diisopropylnaphthalene differ markedly from those of conventional cocktails and from ther-

mal noise generated within the PMT. These differences can be used with patented time resolved pulse discrimination circuitry to provide single PMT counting systems capable of high efficiency and low background.5 One exist-

ing application for such a combination is the use of radiochromatography systems, where a radiolabeled HPLC effluent is mixed with scintillation cocktail and presented in front of a single-PMT detector for quantitation. This chapter summarizes accumulated data on each of the above classes of liquid scintillation cocktails. Experimental results are described which confirm the performance of these new cocktails. Furthermore, data are presented suggesting that Ultima Gold can be effectively used as a liquid cocktail in a single PMT detection system.

37

SAFE SCINTILLATION CHEMICALS FOR THROUGHPUT COUNTING

Table 1. General Solvent Data

Solvent

Boiling Point°C

Flash-

Vapor TLV Pressure

CLASSIFICATION

Intema-

U.S.A. Point°C ppm 25°C mmHg tional Flammable Flamm. 11 40 50

Scintillaboa Yield

Dioxane

101

liquid

class. lB

Toluene

110

4

100

28

Same

Same

100

Xylene

137-139

25

100

8

Same

Flamm. class. IC

110

152

31

50

5

Same

Combustible class II

100

Same Same

Same

112

Same

100

Cumene Pseudo-cumene

168

50

25

2

Tn-methylbenzenes

166-178

50

100

2

Phenyl cyclohexane

235-236

100

na.

Phenyl alkane 290-310 130-150 na. (alkylbenzene) Phenylxylyl 302-309 150 na. ethane Diisopropyl naphthalene

290-299 132-140 na.

Harmless Combustible chemical class. IIIB

103

0.76

Same

Same

94

<1

Same

Same

110

1,1

Same

Same

112

EXPERIMENTAL

Physical constants, classifications, and characteristic data of solvents typically used in scintillation cocktails were compi1ed.9 These are summarized in Table 1. To evaluate the counting efficiency obtainable from a variety of LS cocktails, 50 L of 3H-labeled thymidine was spiked into 10 mL samples of each cocktail. Triplicate samples were made. A second set of samples was prepared without any label, to evaluate background countrates. All samples were then assayed for 3H CPM in a Tri-Carb 2250C A liquid scintillation analyzer. From these count data efficiencies and figures of merit (E2/B) were calculated for each cocktail (Table 2). Next, to observe the effect of increasing quench on efficiency, a series of samples were prepared in each cocktail. Samples were quenched with increasing volumes of 0.01 M PBS, and spiked with a known and constant amount of Table 2. LSC Efficiencies and Figures of Merit Classified Cocktail Insta-Gel XF Optitluor Ultima gold

Solvent Pseudocumene Phenylalkane Dilsopropylnaphthaiene

Percent

as safe

tritium efficiency

Bkgrd cpm

E'2/B

No Yes Yes

55.3 48.4 56.4

14.4 13.4 13.8

212 175 231

FOM

38

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

3H-labeled water. Each was assayed for count rate in the aforementioned LSC. The results of this assay are illustrated in Figure 1.

To explain the performance of the Ultima Gold formulation, electronic signals produced at the anode of the PMT were observed and recorded on a digitizing oscilloscope. Figure 2 displays typical pulses for both Ultima Gold and PMT thermal noise. The significant differences between these pulses suggests that time-resolved techniques can be used to discriminate against PMT noise, while retaining adequate counting efficiency. Therefore, another set of samples, similar to the ones prepared for the first set of experiments, were prepared. The assay was repeated in an experimental counting system. This system is composed of a single PMT and a pulse discrimination circuit based on time resolution. Table 3 summarizes the results of this experiment.

RESULTS AND CONCLUSIONS

While LS cocktails based on benzene derivatives offer generally excellent performance, most are characterized by some degree of environmental or safety hazard. A comparison of pertinent physical constants and data for these solvents indicates that, in some cases, these. hazards and associated costs are significant. This has lead to the development of cocktails based on safe sol-

vents. However, the data presented above suggest that the performance of these cocktails is inferior to xylene or pseudocumene based formulations. Counting efficiencies can be 10-15% lower, leading to a corresponding decrease in figure of merit. Additionally, the poor scintillation energy transfer

qualities of the solvents result in cocktails which are easily and severely quenched.

The use of a new class of solvent diisopropylnaphthalene - results in a cocktail whose counting performance meets or exceeds that of earlier safe cocktails. In fact, the results presented above suggest that this new cocktail (Ultima Gold) performs better than even cocktails based on benzene derivatives. Counting efficiencies can be somewhat higher, resulting in a 5-10% increase in figure of merit. Because diisopropylnaphthalene has scintillation properties of its own, efficient energy transfer is maintained even at extreme sample loading conditions, resulting in a cocktail with excellent quench resistance. Furthermore, this cocktail meets or exceeds all current environmental and safety requirements. Drain disposability also aids in reducing laboratory costs. Experimental evidence gathered during these studies indicates that Ultima Gold has unique scintillation properties. The width of a typical electronic pulse produced by the interaction of a beta decay event with Ultima Gold is signifi-

cantly longer than noise pulses produced in the PMT itself. We have seen differences of a factor of four. This is apparently due to the scintillating nature

Figure 1.

0.5

mL FE

1

1.5

2

per 10 mL CCO
Cocktail quench resistance (3H, 0.01 M PBS quencher).

45

25

3

UThA GW

TFLLX

1\iSTAGEL

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

40

a)

Typical EMT Dark Noise

'_

-.

Threshold

-

10 mV/div 5 ns/cIiV

3

H-

b)

Ulti ma C id

Thesho id 10 my/di 5 ir /di

Figure 2.

I

V

Electric pulses, a) PMT thermal noise, b) Ultima Gold (3H).

of the solvent, which effectively prevents quenching of the slow component of the fluorescent decay. As has been demonstrated, these differences in pulse width can be exploited in the use of time-resolved noise discrimination techniques. This technology Table 3. Single PMT Counting Results

Cocktail Insta-Gel XF Optifluor Ultima gold

Classified

Percent

as

Solvent

safe

tritium efficiency

Pseudocumene Phenylalkane Diisopropylnaphthalene

No Yes Yes

20.0 14.0 32.0

Bkgrd cpm 114.0 47.0 54.0

FOM

E'2/B 3

4 19

SAFE SCINTILLATION CHEMICALS FOR THROUGHPUT COUNTING

41

offers the possibility of single PMT scintillation counting systems. These systems were not previously feasible with conventional liquid cocktails because of the insignificant differences between decay and noise pulses. Results in an experimental counting system based on these principles show that improvements in efficiency range from 30-60% when Ultima Gold is used. This results

in a factor of five improvement in figure of merit. While efficiencies and sensitivities using this technology are somewhat less than in conventional LSC, further developments in both electronics and chemicals are expected to lead to

improved performance. Furthermore, the use of time-resolved single PMT technology can lead to the development of high performance multidetector liquid scintillation counting systems. REFERENCES

Bray, G.A. "Determination of Radioactivity in Aqueous Samples," in The Current Status of Liquid Scintillation Counting, E.D. Bransome, Ed. (New York: Grune and Stratton, 1970), p. 170. Bray, G.A. "Determination of Radioactivity in Aqueous Samples," in The Current Status of Liquid Scintillation Counting, E.D. Bransome (New York: Grune and Stratton, 1970), pp 13-24. Kalbhen, D.A. and V.J. Takkanen. "Review of the Evolution of Safety, Ecological and Economical Aspects of Liquid Scintillation Counting Materials and Techniques," in Advances in Scintillation Counting, S.A. McQuarrie, E. Ediss, and L. I. Wiebe, Eds. (Edmonton, Alberta: University of Alberta, 1983), pp. 66-70. U.S. Patent Number 4,657,696, "Scintillation Counting Medium and Counting Method." U.S. Patent Number 4,528,450, "Method and Apparatus for Measuring Radioactive Decay."

Weast, R.C., Ed. CRC Handbook of Chemistry and Physics, 60th ed. (Boca Raton, FL: CRC Press, Inc., 1979). Data supplied by vendor. Birks, J. B. The Theory and Practice of Scintillation Counting, (New York: The Macmillan Company, 1964), pp. 272-278. Dangerous Goods Regulations, 30th ed. (Montreal: International Air Transport Association, I 988).

CHAPTER 5

New Organic Scintillators

Stephen W. Wunderly and Joel M. Kauffman

ABSTRACT Several new organic scintillators have been synthesized based on quaterphenyl and sexiphenyl structures. Physical properties of the new fluors are given. The fluors were compared with commercial scintillators in a variety of applications. The sexiphenyl structure was the most efficient fluor as well as the most resistant to chemical quenching from nitromethane. Both the sexiphenyl and quaterphenyl structures were much more efficient than either PPO or b-PBD based systems for scintillation solutions capable of emulsifying aqueous samples.

INTRODUCTION

Since the development of liquid scintillation counting, those involved in improvements of the technique have sought better and more efficient scintillators. In a paper presented by Birks' the focus was minimizing quench rather than compensating for quench. In the same spirit we would like to report on new primary scintillators that lead to improved scintillation efficiency. In Birks' report, he ranked a variety of scintillator systems with respect to

two counting conditions minimal quench and strong quench due to carbon tetrachioride. He found that while some systems performed well in minimal quench circumstance (the quaterphenyl BBQ or BIBUQ was 2% more efficient than PBD systems), these same systems performed quite poorly in the presence of strong quenchers (BBQ was 49% less efficient than PBD systems with added carbon tetrachloride). It is obvious from these results that sample conditions causing quench affect the ranking of scintillator systems. Since most liquid scintillation cocktail is used in bioresearch, primarily with aqueous samples, we will rank scintillator systems for aqueous quench as well as minimum quench and strong chemical quench. 43

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

44

Table 1. Fluor Systems Investigated Systema Number

1° Fluorb PPO PPO b-PBD b-PBD PPF

1

2 3 4 5 6

7 8

PPF d-CH3O-PPF d-amyl-PPF

9

(PF)2

2° FIuorc

-

Bis-MSB Bis-MSB

Bis-MSB

aSolvent_pseudocumene bprtmary fluor at 2.26x 10-2 M. CSecondary fluor at 1.6x i0 M.

EXPERIMENTAL

Synthesis

The new primary fluors based on quaterphenyl and sexiphenyl ring structures are pictured in Figure 1. Detailed synthetic procedures will be submitted for publication at a later date. Synthetic details for PPF and d-CH3O-PPF are also found in the Ph.D. thesis of Alem Ghiorghis.2 New Fluors: Physical Properties PPF 2,7-diphenyl-9,9-dipropylfluorene m.p. 195.5-197.5 NMR: 7.8-7.2 (m, 16H, aromatic), 2.0 (m, 4H, CH2CH2CH3) 0.7 (m, 1OH, CH2CH2CH3). UV: 330 (4.76) Emission: 357 and 375 (excitation 326)

d-amyl-PPF 2,7-bis(4-t-amylphenyl)-9,9-dipropylfluorene m.p. 202-204.5 NMR: 7.8-7.4 (m, 14H, aromatic), 2.05 (m, 4H, CH2CH2CH3) 1.7 (q, 4H, C(CH3)2CH2CH3), 1.35 (s, 12H, C(CH3)2CH2CH3) 0.75 (m, 16H, C(CH3)2CH2-CH3 and CH2-CH,-CH3>) UV: 212 (4.785), 328 (4.716) Emission: 365, 383, 401 (excitation 344)

d-CH3O-PPF 2,7-bis(4-methoxyphenyl)-9,9-dipropylfluorene m.p. 197.5-201.5

NMR: 7.8-6.9 (m, 14H, aromatic), 3.9 (s, 6H, OCH3), 2.0 (m, 4H, CH,-CH2CH3), 0.65 (m, 1OH, CH2CH2CH3) UV: 335 (4.748) Emission: 371, 390 (excitation 337)

NEW ORGANIC SCINTILLATORS

CHCH2CH2

CH2CH2CH3

PPF: R = HcI-CH3O-PPF: R

CH3OCH3

d-8myl-PPF: R: CH3-CH2-CIH3

(Pr)2 Figure 1.

(PF)2

Pr: Propyl = -CH2CH2CH3

Structures of new oligophenylene fluors.

7,7'-diphenyl-9,9,9',9'-tetrapropyl-2,2'-bifluorene

m.p. 241.5-242.5 NMR: 7.9-7.3 (m, 20 H, aromatic), 2.15 (m, 8H, CH2-CH2CH3, 0.8 (bs, 20 H, CH2CH2CH3) UV: 204 (4.98) 347 (4.93) Emission: 386, 408, 430 (excitation 347)

45

46

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Other Material pseudocumene, PPO, b-PBD and bis-MSB purchased from Beckman Instruments, used as received.

nonylphenolpolyethoxylate purchased from Dorsett and Jackson, Long Beach, CA. scintillation measurements made with Model 3801 liquid scintillation counter from Beckman Instruments.

Procedure 5 mg of secondary fluor (when required) were placed in a maxi glass scintillation vial. Pseudocumene, 10 ml, was added to the vial, capped, and shaken, and an average of 5 H-numbers was determined. This was repeated for each sample. Then each vial was opened and 30 iL of nitromethane were added. The vials were again capped and shaken, and an average of 5 H-numbers was determined for each vial. Primary fluor, 2.26 x i0 mol, and 5 mg of secondary fluor (if required) were added to a maxi-glass scintillation vial. Pseudocumene, 6 g, was added to each

Primary fluor, 2.26 x l0 mol, and

vial. The vial was capped and shaken until the fluors dissolved. Then 4 g of nonylphenolpolyethoxylate emulsifier and 1 g of water were added to each vial and then capped and shaken. The average of 5 H-numbers was determined for each vial.

CALCULATIONS

Horrocks3 established a relationship between pulse height (PH) and energy (E) on a Beckman Model 9800 liquid scintillation counter that has been maintained through subsequent models. That relationship is

PH = 76 + 280 log E.

(1)

The pulse height for the Compton edge inflection point for '37Cs, energy 478 KeV, is by definition 826. The amount of quench of a sample is the difference between the inflection point of the Compton edge of that sample compared with the inflection point of the Compton edge of the unquenched standard. This difference is quantified as the H-number and defined as

H# = PH0 - PH1

(2)

where PH0 = the pulse height of the unquenched standard at the Compton edge inflection point PH1 = pulse height of the sample in question at the Compton edge inflection point.

By combining Equations 1 and 2 the relationship of Equation 3 may be established

H# = PH0 - PH1 = (76 + 280 log 478) - (76 + 280 log E,)

(3)

47

NEW ORGANIC SCINTILLATORS

Table 2. Relative Light Output Air-Quenched Cocktail System Systema

Relative Light Output

H-Number

100 98 95 95 95

1.5

(PF)2

4.5 7.0 8.0 8.0 9.5 11.5 18.8 30.0

PPF PPFFb1s

d-CH3O-PPF d-amyl-PPF b-PBD/bis b-PBD PPO/bis PPO

94 92 86 78

a10 mL of cocktailair quenched only.

Solving for log E1 we arrive at Equation 4 log E1 = 2.6794 - H#/280

(4)

E is the apparent energy of the quenched sample, i. However, since the decay energy of the '37Cs has not changed, it must be the amount of light from the scintillator per unit of energy input that has changed. Horrocks3 has also established a relationship between energy (Ei) and number of photoelectrons generated at the photocathode of the PMT, (Equation 5), where #e1 is the number of photoelectrons = 20 + (0.666) #e

(5)

The number of photoelectrons generated at the photocathode is directly proportional to the scintillator light impinging on the PMT face. Therefore, the ratio of photoelectrons between two scintillators expresses the relative light output of the two scintillators (Equation 6). Using this relationship we may

compare and rank the new scintillators with standard commercial scintillators.

Relative light output = RLO eJx 100 =[(E -20) RLO = (e,

(E - 20)lx 100

(6)

RESULTS

The traditional fluor systems of PPO or b-PBD, with or without secondary

fluor, performed in our tests as would be predicted from Birks' work for minimal quench and heavy chemical quench (Tables 2 and 3). It is suprising to see in Table 4 the equivalence of the two systems in formulations for aqueous

samples (one of the most common applications of scintillation cocktails). Therefore, the more expensive fluor, b-PBD, offers no advantage in scintillation performance over the less expensive fluor, PPO, for the measurement of aqueous samples.

Three of the new fluors based on fluorene are structurally related to

48

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 3. Relative Light Output Nitromethane Quenched Cocktail System Systema (PF)2

b-PBD/bis b-PBD d-CH3O-PPF PPF/bis PPO/bis PPF d-amyl PPF PPO

H-Number 124 150 152 176 188 199 208 212 220

Relative Light Output 100 77.8 76.6 60.5 53.7 47.8 43.6 41.8 38.2

a10 mL of cocktail systemquenched with 30 L of nitromethane.

quaterphenyls. Quaterphenyls are known to be effective fluors. One example,

BBQ, is reported by Birks'. BBQ was the result of an exhaustive study by Wirth4 of substitution on oligophenylenes to improve solubility and maximize scintillation pulse height. The latter was achieved only at much higher concentrations than were required with PPO. Barnett et al.5 showed that the o,o'-methylene-bridged quaterphenyls, 2,2'bifluorene, and 2,7-diphenylfluorene (PPF without propyl groups) gave superior pulse heights to quaterphenyl in liquid scintillation counting, a process related to lasing, at least in that the S1-S0 transition of the fluor or dye occurs mainly by fluorescence.

Pavlopoulos and Hammond6 suggested that methylene bridged quaterphenyls such as 2,2'-bifluorene and 2,7-diphenylfluorene might prove to be superior laser dyes. All of these quaterphenyls have very low solubility, which severely limits their utility. Both 2,2' bifluorene and 2,7-diphenylfluorene have recently been reported by Rinke7 as effective eximer-pumped laser dyes. Both the solubility and photochemical stability of these laser dyes were dramatically improved when nonaromatic hydrogens were replaced by propy groups as reported by Kauffman.8 PPF and d-CH3O-PPF were among the quaterphenyls studied as laser dyes. Recently PPF as well as other fluorenes were examined by Kauffman9 as Table 4. Relative Light Output Aqueous Cocktail Systems Systema (PF)2

PPF d-CH3O-PPF PPF/bis d-amyl-PPF PPO/bis b-PBD/bis b.PBD

pPO

H-Number 75.0 80.0 81,5 82.2 82.5 100.5 100.8 116.8 118.5

Relative Light Output 100 95.6 94.3 93.7

93.4 79.4 79.2 68.4 67.3

a6 g of solvent fluor system, 4 g of nonylphenolpolyethoxylate emulsifier and 1 g of 0.1. water constitute the aqueous cocktail system.

NEW ORGANIC SCINTILLATORS

49

scintillation fluors in a polysiloxane matrix for detection of gamma rays, mainly in quest of radiation hardness. Despite the low concentrations of fluor

obtainable in the polymer, the light output of PPF was surprisingly high, which encouraged us to examine PPF as a fluor for liquid scintillation solutions. In the current study all three quatraphenyls structures, PPF, di-t-amyl PPF and di-methoxy PPF, have similar scintillation properties. All three have simi-

lar performance to b-PBD systems and are much better than PPO based systems for experiments with only air quenching present (see Table 2). All three have poorer performance than b-PBD systems, but better than PPO systems when a strong organic chemical quencher, such as nitromethane, is present (see Table 3). The dimethoxy derivative was more quench resistant to nitromethane than either the parent or di-t-amyl quaterphenyl Finally, all three diphenyl fluorene systems perform much better than either b-PBD or PPO systems when aqueous samples are a cause of the quench (see Table 4).

The final new fluor, (PF)2, which is named here as a dimer of phenylfluorene, is structurally a sexiphenyl. Its scintillation properties are outstand-

ing compared to any of the other new or traditional fluors. It is the most efficient fluor for all three test systems; air quench, strong chemical quench, and aqueous quench.

The sexiphenyl structure would appear to be a promising direction for research into more efficient and more quench resistant scintillation solutions.

REFERENCES

1. Birks, J.B. "Impurity Quenching of Organic Liquid Scintillators," in Liquid Scmlillation Counting: Proceedings of a Symposium on Liquid Scintillation Counting; M.A. Cook and P. Johnsons, Eds. (New York: Heyden, 1977), pp. 3-14. Ghiorghis, A. "Synthesis of Oligophenylene Laser Dyes," PhD Thesis, Massachusetts College of Pharmacy and Science, Boston, MA (1988). Horrocks, D.L. "Energy per Photoelectron in a Liquid Scintillation Counter as a Function of Electron Energy," in Advances in Scintillation Counting, S.A. McQuar-

ne, C. Ediss, and L.I. Wiebe, Eds. (Edmonton, Alberta: University of Alberta, 1983), pp. 16-29. Wirth, H.O., F.U. Herrman, G. Herrman, and W. Kern. Mo!. Cryst. 4:321 (1968). Barnett, M.D., G.H. Daub, F.N. Hayes, and D.G. Ott, J. Am. Chem. Soc., 81:4583 (1959).

Pavlopoulos, T.G. and P.R. Hammond. J. Am. Chem. Soc., 96:6568 (1974). Rinke, M., H. Gusten, and J.J. Ache. J. Phys. Chem., 90:2666 (1986). Kauffman, J.M., C.J. Kelley, A. Ghiorghis, E. Neister, L. Armstrong, and P.R. Prause. Laser Che,n., 7:343 (1987). Kauffman, J .M. Proceedings of Workshop on Scintillating Fiber Detector Development for the SSC, Vol. II, (Batavia, IL: Fermilab, 1988), pp. 677-740.

CHAPTER 6

Advances In Scintillation Cocktails

Jan ter Wiel and Theo Hegge

Liquid scintillation counting techniques have become a wide-spread method to obtain quantitative data for , fi, and ' emitting radionuclides. The detection sensitivity and efficiency for measuring soft fl-emitters, tritium, and '4C made liquid scintillation counting the accepted analysis technique. Ever since the very beginning of liquid scintillation counting, the improvement in instrumentation is clearly demonstrated. The continuous increasing

demand for multipurpose scintillation cocktails contributed to the development of a series of investigations leading to a better understanding of the scintillation process and the development of new cocktails. Early requirements not taken into consideration, liquid scintillation cocktails have been improved in the 70's, and progress was made concerning increased sensitivity improved compatibility with samples volume reduction safety of handling and storage disposal

The present situation is a result of a development which started in the early 80's. A growing concern of health risk handling scintillation cocktails led to the introduction of new liquid scintillators. Besides the above mentioned, economical factors and continuous development of new surfactant systems contributed considerably to the desires of scientific and routine laboratories using liquid scintillation counting as an analytical tool. The introduction of new cocktails based on solvents with a high flashpoint with a minimum of safety restrictions was a step forward. The improvement of surfactants applied in LS cocktails contributed sigmficantly to the simplification of sample preparation procedures. The well known classical sol-gel cocktails have been replaced by cocktails exhibiting a continuous clear liquid phase in applications with aqueous samples. An early example of new types of cocktail construction was discussed by O'Conner and Bransome.'

In the newest generation of liquid scintillators, a series of improvements 51

52

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

combine a high degree of aqueous sample compatibility with progressive safety characteristics. COCKTAILS, ASPECTS AND REQUIREMENTS

The major part of samples for radioactivity determination by liquid scintillation counting is of aqueous origin. The requirements to apply the technique in a proper way have been discussed in review articles and conference proceedings (Peng).2 The general theory is assumed to be known.

Constituents of Cocktails:

The original construction of a cocktail remained the same throughout the years: solvent surfactant system (emulsifiers) scintillators

It is obvious that organic samples or samples soluble in an organic solvent can be quantified in a system without surfactants. The advances of liquid scintillators are mainly based on progress in applica-

tion of new solvents and surfactant systems, which will be discussed in general.

Scintillators The solid primary and secondary scintillators did not change in fact.The most common primary scintillator is still PPO (diphenyl oxazole). Bis-MSB, together with POPOP, are the secondary scintillators in most cocktails. Solvents

Toluene, dioxane, and xylene were applied initially in liqiud scintillation cocktails, followed by pseudo-cumene (1 ,2,4-trimethylbenzene). The newest

generation of aromatic solvents all have in common the high flashpoint, accompanied by much more preferred safety characteristics. Some influences and requirements have been important in improving the quality of solvents. Economical:

availability of aromatic solvents in large quantities availability of aromatic solvents economically priced improvement of quality broadening of range available

Safety/ Economical:

demand for less volatile products application in polyethylene vials less restricted storage, handling and transport

ADVANCES IN SCINTILLATION COCKTAILS

Safety:

53

reduction of volatility of organic solvents resulting in reduced risk to personnel reduction of toxicity reduced fire and hazard risks growing concern of environment

Surtactants

The incorporation of water or aqueous samples in aromatic solvents was initially solved by using, e.g., dioxane and methyiglycol. A solution was proposed by Bray.3 Application of Triton-X-1004 for example produces results comparable to many commercially available sol-gel scintillation cocktails. Triton-X-200 is a trademark of Rohm and Haas. As the acceptance of the LSC technique became more important and widespread, its uses demanded improvements. Some criteria of influence were: increased detection sensitivity compatibility with various types of aqueous samples increased sample load in the fluid region higher resistance to quench economical reasons (counting in small vials)

A combination of investigations on solvents and surfactants created a new generation of cocktails for liquid scintillation counting. EXPERIMENTAL

Counting efficiencies were determined on Packard Tri-Carb liquid scintillation counters model 2250CA or 1900CA as indicated in the tables. Instrument settings for tritium resulted in 64.1% efficiency for the 1900CA and 67-68% efficiency for the 2250CA for a sealed tritium standard. Instrument efficiency

for 'C is 95.8%. All efficiency determinations were performed at room temperature.

Counting efficiency of aqueous samples was obtained by spiking 10 L tritiated water corresponding to 19,384 dpm in a glass vial containing the sample of interest, all in triplicate. Nonaqueous sample efficiency was obtained by spiking 10 L tritiated toluene corresponding to 14,000 dpm into the samples, all in triplicate. For determining the counting efficiency of '4C the same procedure as for tritium was followed, except for the amount of '4C, a aliquot of 10 L '4C corresponding to 3814 dpm was spiked. All efficiency determinations were obtained by using 20 ml low potassium borosilicate vials. Efficiency (abs.) was obtained by calculating: cpm x 100 = Wo efficiency dpm

54

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Samples were visually and instrumentally checked for homogeneity

sample load -

volume of sample volume of sample + volume of cocktail. x 100

Phase diagrams were obtained by adding the aqueous buffer or salt solutions

to 10 mL of cocktail with increments of 0.5 mL. Near phase separations, increments of 0.1 mL were taken. Counting of samples was started after equilibration for light and temperature, all samples were counted for 4 mm. Cocktails

Ultima-Gold, Opti-Fluor, Pico-Aqua, Pico-Fluor 40, Emulsifier Safe, and Insta-Gel are commercially available products from Packard Instrument. The cocktail phenylxylyl ethane was prepared by adding 40% w/w nonionic emulsifier blend to phenylxylyl ethane including 5 g PPO and 0.2 g of bis-MSB (PXE N) per liter cocktail. SOLVENTS

Physical Constants, Classification:

Some characteristic data of common scintillation solvents are summarized in Table 1. The data show that a considerable improvement on safety has been made: Increased boiling point and flashpoint, and decreased vapor pressure indicate a

reduced risk in the release of solvent vapors, thus reducing the danger to personnel and increasing laboratory safety. Lower flammability not only contributes to less restricted transport and storage regulations, but also lessens the chance of a possible fire hazard Stopping the penetration of solvents through plastic vials is significant to all safety aspects. High flashpoint solvents received no TLV values; it is unlikely that vapors will be released in the working atmosphere under normal conditions.

The high flashpoint solvents, e.g., phenylalkane and diisopropylnaphthalene have been well investigated concerning their toxicological properties.5 Extensive investigations for diisopropylnaphthalene have been performed on

bioaccumulation and biodegradability and the outcome is very positive. According to the EEC-directive 79/831 EEC Annex VII, diisopropylnaphthalene is considered a non dangerous substance, having a biodegradability of

more than 80 28 days after it is determined.6 Scintillation Characteristics

The question remains whether high flashpoint solvents can compete with established solvents concerning counting efficiency. Tables 2 and 3 summarize

55

ADVANCES IN SCINTILLATION COCKTAILS

Table 1. Characteristics of Common Scintillation Solvents CLASSIFICATION

Vapor

Solvent

Boiling Point°C

Dioxane

101

flV Pressure

lntemaFlashPoint°C ppm 25°C mmHg tlonal 11

50

40

Toluene

110

4

100

28

Xylene

137-139

25

100

8

Cumene

152

31

Flammable Flamm. liquid class. lB Same Same Flamm.

110

class. IC Combustible class. II

100

50

5

Same

168

50

25

2

Same

166-178

50

100

2

Same

Phenyl cyclohexane

235-236

100

n.a.

290-310 130-150 n.a. Phenyl alkane (alkylbenzene) 150 n.a. Phenylxylyl 302-309 ethane Diisopropyl naphthalene TLV

290-299 132-140 n.a.

100

Same

Tn-methylbenzenes

Pseudo-cumene

ScintIlladon Yield

U.S.A.

Same Same

112

100

Harmless Combustible chemical class. IIIB

103

0,76

Same

Same

94

<1

Same

Same

110

1,1

Same

Same

112

threshold limit value and na. = not assigned.

some experimental data for tritium counting efficiency of some commercially available solvents containing PPO and bis-MSB as scintillators. These data show clearly a very important fact: without suffering in tritium counting efficiency, the safety aspects can be increased significantly.

Table 2. Percentage Tritium Counting Efficiency (TRI CARB 1900 CA) DiisopropylnaphPhenylxylylthalene (DIPN) Phenylalkane ethane (PXE) Pseudo-cumene 63.6

55.2

62.2

62.8

Table 3. Quench Characteristic with Carbon Tetrachloride; Tritium Counting Efficiency (TRI CARB 1900 CA)

I Carbon Tetrachloride 0 10 20 30 40 50

PseudoCumene 63.3 50.8 44.5 38.5 33.9 30.8

PXE

62.2 51.8

433 36.6 29.5 23.8

Diisopropyl Naphthalene 62.8

51.7 47.5 41.9 38.8 34.4

56

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

SURFACTANTS

Surfactant Systems

Surfactants have to be added to organic solvents when, in particular, aqueous samples are the subject of analysis in activity determination The very well known sol-gel scintillators are most widely used because of their wide applicability. The surfactants, nonionics, applied to obtain these types of cocktails are discussed by Benson7 and Lieberman and Moghissi.8 The increasing demand of applications led to the introduction of cocktails based on nonionic and anionic surfactants, an example is given by O'Conner and Bransome) An important advantage is obtained; with aqueous samples give a clear non viscous homogeneous sample, important when small (e.g., 6 mL) counting vials are used. The diversification of sample types, increase of sample amounts, and when-

ever possible, minimum of sample preparation procedures influenced the development of new scintillation cocktails, in which in a combination with anionic and nonionics, additional anionic surfactants are applied, e.g., Hegge and Wiel,9 Sena et al.,'° Mallik," and references cited therein. The additional surfactants all have improved characteristics of cocktails in common Some general formulas of surfactants are given in Figure 1. Alkyiphenolethoxylates were the first applied nonionic surfactants, followed by succinates. Some general structures of additional surfactants show that these all have an anionic character. The characteristics of products differ considerably, as can be seen in Figure 2. Instead of a gel area when a nonionic surfactant is applied, a clear homoge-

neous fluid results when a combination of non ionics and anionics in an aromatic solvent are the constituents of a cocktail. A general advantage shows that in the application of a nonionic/anionic combination, whatever the sol-

vent is, the relative independence of sample load capacity of temperature. Introduction of combined surfactants results in different sample load capacities for water and O.OlM Pbs (phosphate buffered saline), as shown in Figure 2. This is a general behavior although a wide range of samples is applicable. To each sample type belongs a unique sample load capacity.

Counting Efficiency for Liquid Scintillation Types The counting efficiency for tritium of cocktails depends on solvent, surfac-

tant system, and amount of surfactants dissolved in the original solvent, assuming the concentration of scintillators is at optimum. Counting efficiency of aforementioned products is given in Table 4. The counting efficiency with sample of Pico-Aqua is comparable to the solgel scintillator, but the surfactant system applied has the advantage that compatible samples form a clear homogeneous mixture.

The other two products contain an anionic/nonionic surfactant system about 25¼ w/w in different solvents. The lower efficiency of Opti-Fluor is a consequence of the solvent applied.

ADVANCES IN SCINTILLATION COCKTAILS

57

NONIONIC

alkyl phenol ethoxylates

CH21

o

(OH2 - CH2

e.g. n=9

0)

m

H

N=4 to 20

ANIONIC SUCOINATES 0 C

I

e.g. n=6

H2C

__.

__

SO3Na

/ C

CH21

0

0' ANIONIC

H = e.g. alkyl ethoxyi.ated alkyl

additional

ethoxylated aryl

R - 000Na

carboxylates

R - SO3Na

sulfonates

S03t1

o ).sP0M

suiphates

phosphates

0

S1, 2

Figure 1.

General formulas of surfactants.

Figure 2.

0 50

2-phase

28 33,3 44,4

clear

16,7

10

gel

OPTI -FLUOR

PICO-AQUA

% sample load

SOC

27°C

nil sample to 10 nil cocktail

Insta-Gelclassical sol-gel liquid scintillatorpseudocumenelxylene based. Pico-Aqua--modern liquid scintillatorpseudocumene based. Pico-Flour 1 5nonioniclanionic scintillatorpsuedocumene based. Opti-Flournonionic/anionic scintillatorphenylalkane based.

gel area

a

Sample capacities for different types of cocktail:

0,O1M Pbs

deionised water

8

PICO-FLUOR 15

(gel)

0,O1M Pbs

I

6

INSTA-GEL

4

(gel)

2

deionised water

0

ADVANCES IN SCINTILLATION COCKTAILS

59

Table 4. counting Efficiency for Tritium for Cocktails from Figure 1 (TRI CARB 2250 CA at 67% EfficIency)

ml of 0.10M Pbs to 10 mL of Cocktail 0 1

3 10

Note: n.c.

InstaGel

PicoAqua

PicoFluor 15

OptiFluor

54.5 47.3 38.4 27.2

48.8 42.3 36.5 25.3

56.9 48.2

46.7 42.2 38.3

n.c.

nc.

52.1

Sample Load 0

9.7 23.0 50.0

no capacity; two-phase sample.

Surfactant Types, Sample Load Capacity

Another illustration showing surfactant influence on sample load capacity of a cocktail for aqueous samples is given in Figure 3. The cocktails differ in composition concerning percentage nonionic surfactants and anionic surfactants as mentioned in, e.g., Benson, Hegge, and O'Conner. The result is extraordinary. The sample load capacity of a product depends on concentration of buffer (salts) present in the aqueous sample. COMPARISON OF COCKTAILS

Counting Efficiency General

For a comparison of tritium counting efficiency between some different types of cocktails, data are put together in Figures 4-9, using some typical samples.

The pseudo-cumene based products have the advantage of large sample holding capacities in the fluid region, while a very fast mixing of sample and cocktail is observed. Very high counting efficiency is achievable too, but consequently sample load capacity is smaller. On the other hand, when applying high flashpoint solvents, reasonable to excellent counting efficiencies are obtained compared to pseudo-cumene based cocktails. A comparison for a common sample type is given in Figures 4 and 5.

It is obvious that the difference observed in tritium counting efficiency is much less pronounced when '4C labeled samples have to be analyzed. In Table 5 some typical examples are given. Difference in Quenching Agents A striking difference is observed when tricholoracetic acid solutions are the

subject of analysis. The application of aforementioned surfactant systems shows a remarkable difference, as shown in Figures 6 and 7. The decrease in counting efficiency is much smaller in Pico-Aqua, PicoFluor 40, and Ultima-Gold compared to nonionic surfactant based products. The example given in Figures 8 and 9 of dark yellow colored urine, a big

Figure 3.

0

V

r

4,

10

28

37,5 44,5 50

clear homogeneous fluid

167

?

Sample load capacities.

1,0 M SodiumChloride

0,01 M Pbs Ph 7.4

deionised water

aqueous sampie9

PICO-AQUA

two phase area

PICO-FLIJOR 40

I

HIONIC-FLUOR nil of sample to 10 niL

I sample Lod

5°C

27°C

of cocktail

ADVANCES IN SCINTILLATION COCKTAILS

61

Table 5. Counting Efficiency for 14C (TRI CARB 1900 CA)

Capacity

% C.E.

Moderate/high High Moderate/high Moderate/high

95.6

Cocktail Type Pseudo-cumene based Pseudo-cumene based Phenylalkane based Diisopropylnaphthalene based

92.1

92.3 92.5

Table 6. Counting Efficiency for Tritium Modern Cocktail (TRI CARB 2250 CA at 68% cH), Example Ultima-Gold 1 mL of Sample to 10 mL of Cocktail

Cocktail

None

Water

1M

Pbs 0,01 M

TCA 10%

Sucrose 30%

Ultima-Gold

57.6

54.4

54.5

53.8

53.9

53.0

HC1O4

color quencher besides chemical quencher, shows good quench characteristics for modern cocktails compared to established examples. Background Decay

Another typical aspect of the application of different types of surfactants is background decay when alkaline samples are applied. Sol-gel scintillators, with quaternary ammonium hydroxide solutions, exhibit quite large background levels. Acidification is an answer to this problem, but it needs an extra 60

58 56 54 S

52

50 48 46 1!

44 42 40 38 36 34 32 30

0.4

0.8

1.2

ml INS TAGEL

+

PICOA QUA

A

1.6

2

2.4

2.8

mpe0.O1 M PBS PJCOFL(JOR 15 PICOFLUOR 40

Figure 4. Counting efficiency cocktails. pseudocumene (or xylene) based solvents.

62

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS 60

55 50

1 40 35 30

25 20

I

0.4

0.8

I

I

1.2

1.6

I

I

2

2.8

2.4

m smpe 0.01 M PBS 0 PXEN OP TIFL UOR

Figure 5.

*

ULTIMAGOLD

EMULSIFIER SAFE

Counting efficiency cocktails, high flashpoint solvents.

60

55 50

40

35 30

25

20 0

0.2

0.4

0.6

0.8

0 0

INS TA-GEL PICOA QUA

I

1.2

1.4

1.6

7.8

mp..10

mt +

/w rCA PICaFL UOR 15 PICO-FLUOR 40

Figure 6. Counting efficiency cocktails, pseudocumene (Or xylene) based solvents.

2

ADVANCES IN SCINTILLATION COCKTAILS

63

60

55 -

20 0.2

0

0.4

0.8

0.6

1

O

PXEN

*

L4

1.2

ml smp(e.10 % o/

1.6

2

TCA

ULTIMAGOLD

OP TIFL UOR

EMULSIFIER SAFE

Figure 7. Counting efficiency cocktails, high flashpoint solvents.

60

55

50

I 40

0

35

C

30

25

20 0

0.2

0.4

0.6

0.8

I

1.2

1.4

1.6

1.8

mt sa)npIe;unnc o

INS TAGEL PICOA QUA

Figure 8.

+

PICOFLUOR 15 PICOFLUOR 40

Counting efficiency cocktails, pseudocumene (Or xylene) based solvents.

2

64

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS 60 55 50

1

30

25 20 15

0.2

0

0.4

0.8

0.6

I

(.2

(.4

16

(.8

2

mr 0 0

PXEN

OPTIFLUOR

*

SUfl% e:urune ULTIMAGOLD EMULSJFJL'R SAFE

Figure 9. Counting efficiency cocktails, high flashpoint solvents.

processing step. Some modern cocktails, however, make acidification superfluous. An example is given in Figure 10. Newest Generation Cocktail

An example of the newest generation cocktails is described below, UltimaGold is based on diisopropylnaphthalene: high flashpoint solvent, e.g., diisopropylnaphthalene safety, no vapor release, very low toxicity easy transport and storage EPA classified as nonhazardous no bioaccumulation single phase sample accommodation (some examples are given in Figure 11) high quench resistance and counting efficiency for tritium (see Table 6) compatibility with alkaline samples

CONCLUSION

Advances in liquid scintillation counting are clearly demonstrated. The application of both new solvents and surfactant systems contributed to the progress. Pseudo-cumene was a first step forward, and it is still important. The introduction of new high flashpoint solvents such as alkylbenzenes and

ADVANCES IN SCINTILLATION COCKTAILS

65

0

0

Sal-gel scintillator

0

Acidified aol-gel scintillator

Instrument Background Level

cocktail

30

60

-. time - minutes

Figure 10.

Some modern cocktails make acidification superfluous.

diisopropylnapththalene provided a significant increase in all safety, toxicity, and economic aspects. The newest generation of liquid scintillation cocktails made improvements in many areas. A high tritium counting efficiency, a fast sample incorporation,

Figure 11.

0

4

clear

16,7 23

2

r

8

I

50

10

2-phase

33,3 44,4

6

5°C

z7°C

I

0

2

4

Phase diagram for aqueous samples of a modern high flashpoint liquid scintillation cocktail.

cetate 0,25M

Jnmonium

Pbs 0,O1M

Deionised water

0

6

8

-

C

ample load

40% (w/w) sucrose

NaH2PO4 0,1N

HC1 0,1M

10 ml sample to 10 ml cocktail

ADVANCES IN SCINTILLATION COCKTAILS

67

and a high and expanded sample load capacity are achieved. Until now a major drawback to cocktails based on a high flashpoint solvent was the higher viscosity, making sample handling a little more difficult. These developments all contributed to the safer and easier use of the liquid scintillation technique while enhancing the quality of results.

REFERENCES

O'Conner, J.L. and E.D. Bransome. "Difficulties in Counting Emulsions of 3H and '4C Labeled Molecules," in Liquid Scintillation Counting Recent Application and Developments, Vol. 2, C.T. Peng, D. Horrocks, and E.L. Alpen, Eds. (New York: Academic Press, Inc., 1980). Peng, C.T. "Sample Preparation in Liquid Scintillation Counting," in Advances in Scintillation Counting, S.A. McQuarrie, C. Ediss, and L.I. Wiebe, Eds. (1983), p. 279.

Bray, G.A. Analytical Biochemistry, I (4-5):279-285 (1960). Meade, R.C. and R.A. Stiglitz. mt. J. Appi. Radiat. Isotopes 13:11 (1962). Thomson, J. Scintillation Counting Medium and Counting Method, U.S. Patent 4,657,696.

"Toxicological and Physicochemical Studies on KMC," (Rutgers Kureha Solvents GmbH). Benson, R.H. "The Importance of Phase Contact in Sol-Gel Scintillator, Aqueous Sample Systems," in Liquid Scintillation Counting, Vol. 2, C.T. Peng, D.L. Hor-

rocks, and E.L. Alpen, Eds. (New York: Academic Press, Inc., 1980), pp. 237-244.

Liebermann, R. and A.A. Moghissi. mt. J. App. Radiat. Isotopes 21:319-327 (1970).

Hegge, Th.C.J.M. and J. ter Wiel. Mixture for Use in the LSC Analysis Technique in U.S. Patent 4,624,799, priority 1983, Netherlands 8,303,213. Sena, E.A., B.M. Tolber, and C.L. Sutula. Liquid Scintillation, Counting and Compositions, U.S. Patent 3,928,227. Mallik, A., and H. Edelstein. Liquid Scintillation Composition for Low Volume Specimens, U.S. Patent 4,443,356.

CHAPTER 7

Solidifying Scintillator for Solid Support Samples Haruo Fujii, Ph.D. and Norbert Roessler, Ph.D.

ABSTRACT Although solid support counting suffers from some disadvantages such as self-absorption and relatively poor reproducibility. it is widely employed for screening assays because it makes sample preparation easy. However, as the number of assays increases, the amount of scintilla-

tor consumed makes waste disposal costs problematical in the light of environmental restrictions.

A solidifying scintillator formulation will be described which allows for dispensing of the scintillator in liquid form with subsequent solidification at room temperature. Issues of sample handling, counting performance, and configuration of the counting equipment, will be discussed in comparison to liquid cocktails.

INTRODUCTION

Solid support counting of weak beta emitters is an established radioassay technique. It is widely employed because the advantages obtained by easy sample preparation outweigh the disadvantages resulting from poor reproducibility due to self-absorption. In practice, radioactive material is spotted onto chromatography paper or isolated on a filter which is then dried and immersed in a cocktail for counting. ' A common element of many of these assays is that no solubilizer is used and the sample remains on the solid support rather than dispersing into the bulk of the cocktail. This adds the effect of nonreproducible counting geometry to the self-absorption problem. Another drawback of this technique is that excessive amounts of cocktail are used. This leads to a vexing disposal problem when large numbers of samples are processed. These problems can be overcome by using a paraffin based solidifying scmtillator. The solidifying scintillator was formulated to allow the experimenter to reverse the phase of the sample from solid to liquid scintillator. The melting point is chosen to allow for good counting efficiency in the solid phase at room

temperature and for easily dispensing the material at moderately elevated temperatures. Since the sample is solid it can be handled easily to obtain good 69

70

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

counting geometry. Also, the smaller volume solid samples are more easily disposed of. Paraffin Scintillator Formulation and Measurement Geometry Common to all conventional scintillators, whether solid or liquid, is the fact that the phase does not change after sample preparation. To obtain a cocktail which can be dispensed in liquid form and counted as a solid, a conventional organic scintillator is modified by replacing a part of the solvent (p-xylene, diisopropylnaphthalene) with pure paraffin having an appropriate melting point. The melting point can be adjusted within limits by changing the molecu-

lar weight distribution of the linear hydrocarbons which make up the paraffin. The paraffin scintillator consists of PPO, bis-MSB, paraffin, and solvent. The paraffin can be homogeneously mixed with solvent and liquid scintillation fluors without giving rise to quenching. PPO and bis-MSB were used because they are the most popular liquid scintillation fluors.4 The solvent dissolves the fluors and works effectively as an energy transfer medium to the primary fluor, which is essential for obtaining a high counting efficiency.56 In use, the scintillator is heated to about 40°C to melt it, and 0.3 mL of the liquid is dropped on a solid support sample. After 10 sec, it solidifies, and a rigid and translucent solid support sample can be obtained. The appearance of the prepared samples is somewhat cloudy, but it is not necessary for them to be fully transparent. Once the solid sample is formed, it is stable and will remain rigid at room temperature. The solid support samples thus impregnated with the paraffin scintillator can be counted in a liquid scintillation counter with each sample suspended in a polyethylene bag which is vertically supported in a plastic holder as shown in Figure 1. It is well known that the counting efficiency of the solid support sample depends on its orientation in a counter.79 Best results are obtained when the plane of the solid support sample is parallel to the photomultiplier faces, due to improved light dispersion from the sample.9 EXPERIMENTAL

Solid Support Samples Aqueous solutions of 3H leucine and '4C uridine with known activity were employed; 0.1 mL of these solutions were individually deposited on solid supports, and then dried overnight at room temperature. The solid supports used here are glass fiber filters, membrane filter chromatography paper, and thin layer chromatography plates. The samples were counted on a liquid scintillation counter, Packard TriCarb Model 4000 Series, coupled with a multichannel pulse height analyzer.

I

Photoinultiplier tube

I"

Solid support sample impregnated with paraffin scintillator

Figure 1. Measurement geometry for solid support sample in LSC.

(tube

Photomultiplier

I

Polyethylene bag (disposable)

Plastic holder (re-usable 25mm4X5Onim)

72

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

500

400C

Liquid scintillation sample

300

I-)

200 a-

100 C-)

0

icco

2 cOJ

3 ca

4000

Channel number

500

400'C

Solid support sample

U

(Class fiber filter)

; 300 aU.'

200-

100 0 0

1000

3003

4003

Channel number Figure 2.

Pulse height spectrum of a 3H solid support sample: in a conventional cocktail (top), treated with the solidifying scintillator.

Homogeneous Samples

Several formulations were also prepared as homogeneous mixtures to test the relative counting performance. In these formulations, 5 g PPO and 0.8 g bis-MSB were dissolved in 250 mL of solvent and mixed with 500 mg of paraffin. After thorough mixing at an elevated temperature, 15 mL aliquots were dispensed into 20 mL scintillation vials. The solvents tested were xylene

SOLIDIFYING SCINTILLATOR FOR SOLID SUPPORT SAMPLES

73

-c

U

U 2000

Channel number 500

-

C

400-

14C Solid support sample

(Glass fiber filter)

-C

0 .-

300-

0 U,

0 U

200-

100

0-0

i obo

2000

300

4000

Channel number Figure 3.

Pulse height spectrum of a 14C solid support sample: in a conventional cocktail (top), tested with the solidiyfying scintillator.

(X), pseudo-cumene (P), and diisopropylnaphthalene (KO, KS). Different batches of paraffin were used for formulations KO and K5. The formulations were evaluated by spiking them with 3H or 14C labeled toluene. After thorough mixing, each sample was allowed to solidify. The samples were assayed in a Packard TriCarb Model 2000CA as well as an experimental single tube time-resolved counting apparatus.'°

74

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

-o

-+ Liquid scintillation sample 0

Glass fiber filter Membrane filter

15

/ /

Chromatography paper

/

n Thin layer chromatography plate,,,' '4

0 ,(

,:1-'

E

10

,

U U

0

0

I

0

10

20

3H Activity (dpmxl0') Figure 4.

I

30

Count rate as a function of 3H activity for various solid support samples.

Results

The scintillation pulse height spectrum was investigated in order to better

interpret the counting data. The pulse height distribution of the 3H solid support sample, prepared in the technique described in Figure 1, is very similar

to that of the conventional 3H sample, although the counting rate for each channel may be reduced (Figure 2). As for a '4C sample, however, the volume and dispersion area of the paraffin scintillator are not large enough to absorb the 3-ray energy of the 14C completely, which seems to cause a shifted pulse height distribution (Figure 3). The optimum fluor concentration and volume ratio of paraffin to p-xylene were also investigated. The formulation for best counting efficiency was found to be; PPO: 10 g; bis-MSB: 1.0 g; paraffin: 670 mL; and p-xylene: 330 mL. The relationship between the measured activity and the count rate is shown

75

SOLIDIFYING SCINTILLATOR FOR SOLID SUPPORT SAMPLES

20

Liquid

scintillation sample

0 Glass fiber filter

Membrane filter '4

0

I

Chroinat ography paper

p

-o

15

Thin layer chromatography plate

i":

'

/ p

I

C

+

V

0

0 0

t

5

I

10

15

20

Activity (dpmxlo4) Figure 5. Count rate as a function of 14C activity tor various solid support samples.

in Figures 4 and 5 for various kinds of solid support samples. The relationship is linear over a wide range of activity for each nuclide. Freshly prepared samples were monitored for several days (Figure 6). The count rate of the sample does not decrease with the time elapsed for more than two weeks. This indicates the absence of any significant vaporization of the paraffin scintillator from the prepared sample.

The counting efficiencies for each support material determined are displayed in Table 1; 3H and 14C can be measured with counting efficiencies of 6-30°/n and 70-87°/n respectively. The difference in the counting efficiency for each support material can be attributed to that of the 3-ray self-absorption and photon reduction inside the solid support. Figure 7 displays the 3H and 14C efficiency values of homogeneous paraffin

solutions measured on a two tube coincidence counter. The formulations tested used the solvents xylene (X), pseudo-cumene (P), and diisopropylnaph-

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

76

+ Liquid scintillation sample

+

'

20 -

°

-° Glass fiber filter

'-

Menbrane filter

C) 'C

E

a. U

-

Chromatography paper

-D

Thin layer chromatography pla

+ ----- +

.

e

---1--------------+ .

,-

. 14c

-

)_0C'

C

?

?

0

0

0

0

C

0

L)

5

.

0

.

S

5

10

15

Elapsed time(days) Figure 6. Count rate as a function of elapsed time for various solid support samples.

thalene (K5, KO). We see that excellent efficiencies are obtained for the solid samples.

In Figure 8, 3H efficiency values for the formulations are displayed as obtained from the experimental time-resolved single tube counter (Valenta). The apparatus is able to operate as a two pulse coincidence counterrequiring that a second pulse be registered in the tube after the first - or as a three pulse triple coincidence counter. While cocktail formulation had little effect on efficiency in the two-tube counter, we see a dramatic improvement for diisopropylnaphthalene in the single tube counter. The results for '4C (see Figure 9) are similar.

15.6 ± 0.2 80.1

30.5 ± 0.2

85.6 86.5 86.8 86.5

2 3 4 5

Mean ± S.D.0

87.1

86.5 ± 0.3

2 3 4 5

Mean ± S.D.

1

81.6 ± 0.4

82.2 82.5 81.9 81.5

15.5 15.7 15.3 15.3 16.2

(14 mg/cm2)a

29.9 30.9 31.0 30.7 30.0

1

Membrane Filter (5.5 mg/cm2)

Glass Fiber Filter

Sample No.

bThickness of thin layer. cStandard deviation.

aSuI.face density.

14c

3H

Nuclide

70.7 ± 0.2

71.4 70.2 70.4 70.6

71.1

5.61 ± 0.04

5.77 5.60 5.56 5.60 5.54

Chromatography Paper (8.5 mg/cm2)

Table 1. Comparison of Counting Efficiencies Obtained from this Study for some Kinds of Solid Support Samples

70.8 ± 0.6

69.5 69.6

71.1

72.8

71.1

7.40 ± 0.10

7.60 7.02 7.43 7.38 7.58

(0.1 mm)b

Thin Layer Chromato. graphy Plate

78

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Li FE

0 U, a)

C

E

=

9 0 0 = a)

G0

0 0 a,

C a,

a

x C

0 0)

-p

0)

=

0 LU

00)

(a

x

(a

0

0. Co

8

0 (0

,JNII3WI

0 N

0

0) a,

CC

>-. = a. 020.

Figure 8.

P

FORMULA110N

K5

KU

2-PULSE

3-PULSE

7////I/II/I/i

Time resolved single tube counting efficiencies for 3H paraffin scintillator samples using various solvents: (X) xylene, (P) pseudo-cumene, (K5, KO) diisopropylnaphthalene.

5

10

15

20

25

30

35

40

0

20

40

60

x

Figure 9. Time resolved single tube counting efficiencies for (K5, KO) diisopropylnaphthalene.

3

z

0 z

±

C)

(J

>-

80

100

FORMULATiON

K5

KO

C paraffin scintillator samples using various solvents: (X) xylene, (P) pseudo-cumene,

P

2-PULSE

3-PULSE

WI/Il//I//h

SOLIDIFYING SCINTILLATOR FOR SOLID SUPPORT SAMPLES

81

Discussion The results described above indicate that a solidifying scintillator with good

counting performance can be formulated using paraffin as a base material. The scintillator is applied to samples isolated on a solid support while melted. This allows it to impregnate the sample while it is in the liquid state. Because it is solid on cooling, an approximate sample geometry for different counting geometries can be assured. Chemical quenching is not a concern in solid support counting, since the paraffin scintillator becomes rigid before the quenching substance is eluted into the scintillator. Color quenching, on the other hand, may be present when

a colored compound is counted. However, considering the thinness of the sample and the possibility of spreading it over a large area, the color quenching effect can be minimized. These considerations mean that samples prepared using a paraffin scintillator will have uniform counting efficiency, although that efficiency will depend, to a large extent, on the kind of solid support. In summary, the main advantages of this technique are: Decreased cocktail consumption Reduced disposal costs because each sample generates a small volume of solid radioactive waste Good reproducibility due to improved measurement geometry Easy and rapid sample preparation

REFERENCES

Peng, C.T. Sample Preparation in Liquid Scintillation Counting, RCC Review 17, (The Radiochemical Centre Ltd., England, 1977), p. 64. Horrocks, D.L. Applications of Liquid Scintillation Counting, (Academic Press, New York, 1970), p. 156.

Furlong, N.B. The Current Status of Liquid Scintillation Counting, (Grune & Stratton, New York, 1970), p. 201. L'Annunziata, M.F. Radionuclide Tracers, Their Detection and Measurement, (Academic Press, London, 1987), p. 182. Birks, J.B. and G.C. Poullis. Liquid Scintillation Counting, Vol. 2, (Heyden, London, 1972), P. 327. Laustriat, G., R. Voltz, and J. Klein. The Current Status of Liquid Scintillation Counting, (Grune & Stratton, New York, 1970), p. 13. Winder, F.G. and G.R. Campbell. Anal. Bioche,n., 57:477 (1974). Vanderheiden, B.S. Anal. Biochem. 49:459 (1972). Nakshbandi, M.M. mt. J. Appl. Radial. Isot. 16:157 (1965).

Valenta, R.J. Method and Apparatus for Measuring Radioactive Decay, U.S. Patent 4,528,450 (1985).

CHAPTER 8

New Red-Emitting Liquid Scintillators with Decay Times Near One Nanosecond*

J.M. Flournoy and C.B. Ashford

ABSTRACT Several fast liquid scintillators with peak emission wavelengths in the 600 to 750 nm range are described, which are useful for transmitting fast plasma diagnostic information through long fiber optic cables. These fluors all use styryl laser dyes as the final emitters: DCM in the 650 nm fluors, and LDS-722 in the 735 nm fluors. The solvents are either straight benzyl alcohol (BA) or BA mixed with 1-methylnaphthalene. Coumarin-480 is the intermediate wavelength shifter for the 0CM fluors and rhodamine-610 perchiorate (Rh-610P) for the LDS-722 fluors. Some systems also contain added tetramethyltin for enhanced X-ray sensitivity. It was found that very high rhodamine concentrations cause a large increase in the solubility of LDS722 in BA, which leads to a decrease in scintillation decay time. The fastest 735 nm fluor has a WHM of 370 psec and a decay time of about 320 psec. The fastest 600 nm emitter has a FWHM of 1.1 nsec and a decay time of 0.87 nsec.

INTRODUCTION

Several fast, red-emitting liquid scintillator formulations have been described previously)'2 They all consist of relatively polar laser dyes dissolved

in polar solvents, either benzyl alcohol or benzonitrile. Intermediate wavelength shifters are used in all cases to bridge the gap between the solvent emission spectrum and the absorption spectrum of the final emitter. The new, faster fluors are the result of changes in the solvent and/or the intermediate shifter. *Thjs work was performed under the auspices of the U.S. Department of Energy under Contract No. DE-ACO8-88-NV10617.

NOTE: By acceptance of this article, the publisher and/or recipient acknowledges the U.S. Government right to retain a nonexciusive royalty-free license in and to any copyright covering this paper. Reference to a company or product name does not imply approval or recommendation of the product by the U.S. Department of Energy to the exclusion of others that may be suitable. 83

84

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

The efficiencies of all the red-emitting fluors are relatively low, only about 3% or of that of anthracene.2 However, that is the price one pays for very fast time response, which is usually a result of quenching of the excited states and/or nonradiative internal molecular deactivation processes, both of which reduce the fraction of the excited molecules that actually emit light. It should be noted, though, that the peak efficiencies are affected less than the integral efficiencies when the scintillator is quenched. This comes about because the area of a scintillation pulse can be approximated by the product of the FWHM and the peak height. Thus, when the decay time and the integral light output both become smaller, the peak intensity tends to remain more constant. The extent to which the peak height remains constant depends on the pulse rise time; the faster the rise time, the less quenching affects the peak efficiency. The peak efficiency that is most important when the time response of the fluor is slower than the overall time response of the measurement system. On the other hand, it is actually disadvantageous to employ a fluor whose response time is significantly faster than the overall response of the measurement system, because the peak intensity will be integrated to a lower value by the slower system. It would, therefore, be better to employ a slower but brighter fluor whose time response is appropriate for the bandwidth of the detection system. The terms "scintillator" and "fluor" are used interchangeably in this paper. For the sake of brevity, abbreviations are used for most of the individual fluor ingredients and some formulations, as shown in the Appendix. EXPERIMENTAL

All the materials were used as received from the manufacturer, except for 1methylnaphthalene (AMN), which was purified on a spinning-tape column to remove a distinct yellow color. The fluorescence emission spectra were taken with a SPEX Fluorolog 2 spectrofluorimeter. The exciting wavelength was either 300 or 350 nm. Emission spectra were observed from the same face as the incident excitation light, as well as through a 1 cm thick solution. The latter measurement gives an indication of the extent of self-absorption of the emitted light. Absorption spectra were taken with a Beckman Model UV 5270 absorption spectrophotometer. The dye concentrations are too high to easily determine the complete absorption spectra of the actual fluors, although spectra of the individual ingredients could be taken in dilute solution. As a rough crite-

rion for estimating solutions transparency to their own emission, we have chosen to tabulate the wavelength beyond which the transmittance of a 1-cm thick fluor is greater than 50. The time-response data were obtained with a time-correlated single photon counting system that has been described elsewhere,3 but some of the data were obtained with excitation by a 90Sr beta source, which gave a much faster data rate than the 60Co source mentioned in Reference 3. The time response FWHM

of the system was determined by the beta particles response to flashes of

NEW RED-EMITrING LIQUID SCINTILLATORS

85

Cerenkov light generated in a piece of high-purity quartz or by Compton electrons produced by the t0Co gamma rays. The system FWHM was less than 200 psec; therefore, the scintillation pulses were not significantly broadened by the instrument response. Start-timing signals were obtained from a disc of the bright, commercial (Bicron) scintillator, BC-422. Relative scintillation efficiencies of the various fluors were determined from

the relative single-photon count rates for t0Co gamma ray excitation. The comparisons were made through a 570-nm long-pass filter, because the emission peaks were not the same. The observed count rates were corrected for

background dark counts and for Cerenkov light produced by 1 cm of pure solvent, observed through the same filter. The relative peak efficiencies were estimated by dividing the integral efficiencies by the respective FWHM values.

RESU LTS

735-nm Fluors A fast, 735 nm liquid scintillator, consisting of 0.02 M LDS-722 and 0.10 M C-540A in benzonitrile (BN), has been described previously.2 This fluor has an impulse response FWHM of about 1.7 nsec, and has been named L-735. When 10% by volume of tetramethyltin (TMSN) was added to improve the absorption cross section for soft X-rays, the solubility of the LDS-722 decreased to just over 0.01 M. A fluor with 10% TMSN, 0.01 M LDS-722, and 0.10 M C540A is called L-735A-1OT, and its pulse parameters appear at the bottom of Table 1. The methyl groups in the TMSN effectively shield the tin atoms from quenching the excited singlet states,4 as can be seen from the relative efficiencies and decay times of L-735A (without TMSN) and L-735A-1OT. A number of experiments with changes in solvent and intermediate wave-

length shifters have now been carried out, partly in an effort to reduce the slight photochemical degradation that was observed with L-735 under ultraviolet illumination. One such combination, benzyl alcohol (BA) as the solvent and rhodamine 610 perchiorate (Rh-610P) as the shifter,* actually decreased the photochemical stability somewhat but resulted in a series of fluors with exceptionally fast time response.

In the course of these studies, it was found that Rh-610P is much more soluble in BA than had been realized: to approximately 0.45 M/L. This amounts to about 25 wt% of the rhodamine salt, so the medium can no longer be expected to have the same solvent properties as benzyl alcohol. One unexpected consequence of this is that the solubility of LDS-722 in BA increases might be mentioned that rhodamine 610 (rhodamine "B") has an exceptionally small Stokes' shift, and therefore would not be expected to be useful as an intermediate shifter. We attribute its efficiency in this application to excitation into S2 or S3 by energy transfer from the solvent, since these bands occur near 300 nm.

-

1003

101

-

-

4

:'

S.

4

S

5

I

7

6

8

I

I

10

I

9

Time (ns)

I

I

11

I

....%%_.

I

12

I

I

13

-

Figure 1. Time-correlated photon counting data for L-735 and L-735XF, along with the system response function. The slowest decay time shown, 1.27 nsec, is for L-735.

C)

0

C

4-.

C,)

ir2 - C) ILl

a -c

£

C

a)

1

I

NEW RED-EMITTING LIQUID SCINTILLATORS

87

dramatically as the rhodamine concentration is increased, from just over 0.005 M LDS-722 in pure BA to at least 0.02 M in BA containing 0.40 M Rh-610P. The combined effects of increasing the concentration of the LDS-722 and the Rh-610P result in the fastest red-emitting scintillators we have observed to date. Figure 1 contains semilogarithmic plots of data for the fastest of these, L735XF, as well as for the original L-735 and the pertinent system response function. Table 1 contains pulse parameters, relative brightness with gamma excitation, and fluorescence emission maxima for a series of these BA-based fluors. The fluorescence emission maxima were determined with an excitation wavelength of 300 nm, and the fluorescence was observed from the opposite face of

the 1 cm spectrophotometer cells. The 50% absorption edges for the 1 cm optical path appear in the last column of the table. The results in Table 1 are presented in order of increasing rhodamine concentration. The fluors described in the last two lines have FWHM values near 500 psec and peak emission wavelengths near 740 nm. The 50% absorption edges for these scintillators increase rather smoothly from 675 nm for the 0.01 M to 0.03 M solutions to 698 nm for 0.20 M Rh-610P. The solution with no

rhodamine is the only one out of line. The l/e self-absorption lengths are estimated to be in the range of 30 to 50 cm at 740 nm, which is of considerable interest for possible use in systems involving fluor-filled capillaries. Incidentally, the decay time of Rh-610P itself in BA was found to decrease from almost 9.0 nsec in very dilute solution (5.0 x 10-6 M) to 2.1 nsec at 0.10 M and about 1.35 nsec at 0.20 M. This may be either an energy-transfer effect or possible evidence of concentration quenching.

600 to 650 nm Fluors A fast, 650 nm liquid scintillator consisting of 0.03 M DCM and 0.10 M C480 in BA has also been described previously.2 Its impulse response FWHM is 1.8 nsec, and it has been named L-650A. The original motivation for trying

AMN as a solvent for DCM-based fluors was twofold: to enhance the efficiency, since BA is not a particularly efficient scintillator solvent, and to take advantage of the high refractive index of AMN, 1.60, for possible use in scintillating capillaries. A number of DCM-containing fluor formulations in BA, AMN, and mixtures of the two have been examined, and the principal results are shown in Table 2, in order of increasing proportion of AMN. It can be seen that the addition of AMN generally causes the integral efficiency to increase, as had been hoped. What was not anticipated was that the AMN-rich fluors were also faster. As a result, the peak efficiencies increase even more than the integral efficiencies as the solvent is changed from BA to AMN. The effect of coumarin-480 as an intermediate wavelength shifter can be seen in the first two lines and the last two lines of Table 2. When the solvent is BA, there is only fair spectral overlap between the solvent emission and the DCM absorption, resulting in the fairly slow pulse rise time of about 1 .05 nsec.

0020C 0.010 0.010

L-735 L-735A L-735A-1OT

-

0.35 0.54 0.60 0.134

1.18 0.73 0.47 0.33 0.37 0.25 0.15

(1.0)

0.193

1.67 2.12 1.89

0.97 0.55 0.37

6.44 (6.6) 3.26 2.08 1.39 1.52

1.27 1.48 1.42 0.085

0.63 0.58 0.32

1.25

5.25 2.03 1.37 0.99

6.11

2.25 3.4 3.4 0.36

0.91

2.1

1.75

4.6 3.3 2.4 2.85

16.1 12.1

-

-

[1.00]

1.1

1.2 [1.001

-

1.0

0.25 0.14 0.06

0.45

-

0.81

0.69 0.60 0.98

1.1

0.5 0.45 0.3

-

0.55 0.75 0.6

0.2

aLDS722 was supersaturated; some crystallized out over several days. bcontajns 10 volume percent tetramethyltin. cshlfter is 0.10 M C-540A; solvent is benzonitrile. See Appendix for composition. dResponse to Cerenkov light flashed from Compton electrons generated by the 60Co gamma rays (see Reference 4).

System responsed

0.020

0005b

o.oio'

0020a 0010b o.oioab 0010b 0.005

0.00 0.00 0.02 0.05 0.10 0.20 0.20 0.30 0.40

0010a,b

-

735 735 732

740

736

-

720 722 730

-

720

698

711

698

--

678 690

675

685

-

Table 1. Pulse Parameters for 60Co Gamma-excited LDS-722 in BA with and without 10% Tetramethyltin (TMSN), using Rhodamine 610 Perchlorate (Rh-610P) as the Intermediate Wavelength Shifter. A 670-nm Long-pass Filter was used in all the Runs. The Peak and Integral Values are Referenced to L-735-A-1OT. AT is the 10% to 90% Pulse Rise Times, TAU is the lie Decay Time (taken from 0.7 to 0.7/e of the peak height), and IRT Is the 10% to 90% Rise Time of the Integral of the Pulse Wavelength (nm) Absorption Emission IRT Peak TAU FWHM AT LDS-722 Rh-610P edge (50%) maximum Integral (est.) (nsec) (nsec) (nsec) (nsec) (Mil) (M/l)

l00%

100%

50% 67%

0.0 0.0

0.0 0.03 0.10 0.03 (L-650-A reference) 0.10 0.05 0.20 0.05 0.10 0.05 0.0 0.05

(M/l)

AMN %

DCM

C-480 (MIt)

0.59 0.49 0.49 0.57 1.11 1.61

1.62 1.37

1.79

0.69 1.49 1.13 0.87 1.34

1.58

(nsec) 2.93

(nsec) 3.58

TAU

FWHM

1.06

AT (nsec)

4.11

3.90 3.22 2.85

6.03 4.05

IRT (nsec)

1.55

1.7

1.25

1.4

0.9 [1.00]

(est.)

Peak

1.06 1.38

1.26 0.96

1.77 [1.00]

Integral

640 632 600 600

648

-

624 622 605 600

629

-

Wavelength (nm) Emission Absorption maximum edge (50%)

Table 2. Pulse Parameters for DCM Fluors in BA, AMN, and Mixtures of the Two. A 650-nm Band-pass Filter was used in all the Runs. The Peak and Integral Values are Referenced to L-650A. AT is the 10% to 90% Pulse Rise Times, TAU is the l/e Decay Time (taken from 0.7 to 0.7/e of the peak height), and IRT is the 10% to 90% Rise Time of the Integral of the Pulse.

z-1

C)

Cl)

-

z

90

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

When 0.10 M C-480 is added, the pulse rise time decreases to about 0.7 nsec,

due to improved energy transfer. On the other hand, the pulse rise time of DCM in AMN, without C-480, is nearly as short as that of L-650A, because there is much better spectral overlap between the broad emission spectrum of AMN and the absorption band of DCM than there is between BA and DCM. This is why the addition of C-480 to DCM in AMN has less effect on the pulse rise time than it does in BA. In addition, there is quenching of DCM by C-480 in either solvent, as shown by the decreases in decay time and integral brightness when C-480 is added. Quenching by C-480 accounts for the lower integral efficiency of the fluor with 67% AMN than the fluor with only 50% AMN, since there is twice as much C-480 in the 67 fluor. The fact that there is quenching by C-480 was confirmed by laser-excitation experiments with DCM in BA, in which a similar effect was observed. Unfortunately, the fluors with AMN show such strong self-absorption that their use in capillaries seems unlikely. This is due to the fact that the emission

spectra shift faster towards the blue than do the absorption spectra as the proportion of AMN is increased. When the solvent is changed from pure BA to pure AMN, the emission peak wavelength for DCM moves from 650 to 600

mn, but the absorption peak moves only from 490 to 470 nm. The 50 absorption edge through a 1 cm path of 0.05 M DCM moves from about 630 to about 600 nm, while the peak emission moves from 648 to about 600 nm.

CONCLUSIONS

The time response of scintillator formulations can be strongly dependent on the nature of the solvent, as well as any intermediate wavelength shifter. In each of the systems described above, there is quenching of the final emitter by the intermediate shifter, which was completely unanticipated, and which is largely responsible for the faster time response of the modified formulations. These results also demonstrate that a considerable fraction of the peak efficiency can be retained in the quenched scintillator, provided the pulse rise time is fast enough.

ACKNOWLEDGEMENTS

The authors gratefully acknowledge the contributions of Mr. Steven Lutz, who developed the original L-735 formulation, Dr. Isadore Berlman, who originated the idea of tetramethyltin as a nonquenching, heavy-atom additive, and Prof. Bruce Rickborn for purification of the AMN.

NEW RED-EMI fliNG LIQUID SCINTILLATORS

91

REFERENCES

I. Lutz, S.S., L.A. Franks, J.M. Flournoy, and P.B. Lyons. "New Liquid Scintillators for Fiber Optic Applications," in Proceedings of Los Alanos Conference on Optics, SPIE 288, pp. 322-328 (1981). Ogle, J.W., D. Thayer, L. Looney, F. Cverna, 0. Yates, C.E. Iverson, S.S. Lutz, M.A. Nelson, and B. Whitcomb. "Radiation-Induced Imaging System Over Long Fiber-Optic Bundles," SPIE 720, pp 24-30 (1986). Flournoy, J.M. "Measurement of Subnanosecond Scintillationn Decay Times by Time-Correlated Single Photon Counting," Radiat. Phys. Chem., 32, pp. 265-268 (1988).

Ashford, C.B., I.B. Beriman, J.M. Flournoy, L.A. Franks, S.G. Iversen, and S.S. Lutz. "High-Z Liquid Scintillators Containing Tin," Nuci. Instr. Meth. Res. Sect. A (Netherlands), A243, pp. 131-136 (1986).

APPENDIX

alpha-methylnaphthalene, or 1 -methylnaphthalene benzyl alcohol (alpha-hydroxy toluene) benzonitrile (phenyl cyanide, cyanobenzene) Coumarin 480 (or Coumarin 102) Coumarin 540A (or Coumarin 153) red-emitting styryl laser dye, also Kodak dye No. 14567 BA with 0.1 M C-480 and 0.03 M DCM L-735 BN with 0.10 M C-540A and 0.02 M LDS-722 L-735A BN with 0.10 M C-540 and 0.O1M LDS-722 L-735A-1OT BN with 0.10 M C-540A, 0.01 M LDS-722, and 10o TMSN L-735XF BA with 0.40 M Rh-610P and 0.02 M LDS-722 LDS-722 far-red-emitting styryl laser dye (Exciton) Rh-6 lOP Rhodamine 610 (or Rhodamine B) perchlorate salt TMSN tetramethyltin AMN BA BN C-480 C-540A DCM L-650A

CHAPTER 9

New Developments in X-ray Sensitive Liquid Scintillators at EGG/EM* C.B. Ashford, J.M. Flournoy, S.S. Lutz, and I.B. Beriman

ABSTRACT A summary of the liquid scintillators developed at EGG/EM SBO is presented. Among the characteristics presented are the emission spectra and efficiency values. All of the scintillation solutions mentioned herein have the potential for sensitivity enhancement to X-rays. One such successful red-emitting solution is L-735A-1 01 which shows a severalfold enhancement for a sample thickness of 6 mm, excited by 17-keV X-rays. Other compounds containing heavy atoms such as tetramethylgermanium, tetraethyllead, and tetramethyllead were tested for possible heavy-atom quenching effects using high-energy electrons from a linac. The germanium compound showed no quenching effects, but both lead compounds exhibited considerable quenching of the fluorescence.

INTRODUCTION

At EG&G/Energy Measurements Inc., Santa Barbara Operations, work is in progress in the development of fast and efficient scintillators whose peak emissions are in the range from blue (350 nm) to red (840 nm). Some of the more successful formulations' are summarized in Table 1, along with the two commercially available plastic scintillators, BC-422 and BC-400.2 Their spectral efficiencies are illustrated in Figure 1 - The data in both table and figure were obtained by exciting the sample with a pulse of 16-MeV electrons whose FWHM is 50 psec. In our work incorporating compounds with heavy atoms into scintillation solutions in order to increase their sensitivity to X-rays, we found that tetrabu*This work was performed under the auspices of the U.S. Department of Energy under Contract No. DE-ACO8-88-NV1O6I7.

NOTE: By acceptance of this article, the publisher and/or recipient acknowledges the U.S. Government right to retain a nonexclusive royalty-free license in and to any copyright covering this paper. Reference to a company or product name does not imply approval or recommendation of the product by the U. S. Department of Energy to the exclusion of others that may be suitable.

93

SG-180 BHTP TPB DCM SR-MO LDS-722 LDS-821

--

0.035 0.14 0.022 0.03 0.002 0.02 0.005

--

Conc. (Mil)

PBD C-480 C-540A C-540A Rh-610P

---

Shifter

0.027 0.10 0.02 0.10 0.10

---

Conc. (MIl)

PC PC PC BA BA BN BA

--

Solvent

8Refers to the ratio of the optical energy out vs the energy absorbed.

L-841

BC-400 BC-422 L-360 L-370 Liquid-A L-650A L-660 L-735

Scintillator

Final Emitter

0.52

0.21

2.66

0.21

0.9 0.24 0.45 0.18 0.23

(nsec)

PRT

1.56

0.24 0.84 0.13 0.27 0.10 0.04 1.26 3.23 3.53 26.30 2.77 3.99

1.72 1.56 0.40 0.90 1.52 12.09 1.27 1.37

1.20 1.65 14.45 1.54 1.88

0.45

1.91

2.31

9.4 4.44 4.17

2.3

1.58 2.21

% Ef1.

3.1

(nsec)

IRT (nsec)

Decay (nsec)

FWHM

420 390 360 370 460 650 640 735 850

Peak Lambda (nm)

Table 1. Properties of Liquid Scintillators Developed at EGG/EM Santa Barbara Operations. Excitation by 50 Ps FWHM Pulses of 16-MeV Electrons

95

X-RAY SENSITIVE LIQUID SCINTILLATOR DEVELOPMENTS

101

__ io

z iouJ

0 UU-

i

1 o-

300

I

I

340

I

I

I

I

I

I

580

540

500 460 WAVELENGTH (NM) 420

380

Figure la. Absolute scintillation efficiencies vs wavelength for liquid scintillators developed at EGG/EM Santa Barbara Operations.

101

I

I

I

I

I

I

I

I

I

I

I

I

BC-422 LIQUID A

io

z

L-650A

L-660

10-°

L-84 1

UJ

0 UUJ

-

i r

-5

300

FIgure lb.

I

400

500

I

700 600 WAVELENGTH (NM)

I

800

I

I

900

1000

96

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

tyltin (TBSN) and tetramethyltin (TMSN) would not affect the scintillation properties except for a dilution factor.3'4 The butyl and methyl groups are effective in isolating the heavy atoms from quenching the excited scintillator molecules. Thus TMSN, because of its higher percentage of tin, is deemed the more useful of the two compounds. Tin-loaded modifications of all the liquid scintillators listed in Table 1 have been successfully prepared. Tin additives are particularly effective for X-rays in the energy range from about 10 to 100 keV; the K absorption edge for tin is at 29 keY. For lower energies, a lighter additive such as germanium, with a K edge at 11 keV, should be a more efficient absorber; for higher energies something like lead with a K

edge of 88 keV would be preferable. Three compounds, tetramethylgermanium (TMGE), tetraethyllead (TEPB), and tetramethyllead (TMPB) were investigated for possible fluorescence quenching based on emission intensity and decay time measurements at the linac. A new, tin-loaded, 735 nm liquid, L-735A-1OT, has been developed, and its enhanced sensitivity to 17 keV X-rays has been measured. This new solution has a measured index of refraction of 1.53 at 740 nm and an optical attenuation of about 50 cm at 740 nm which make it attractive for use in capillaries. EXPERIMENTAL

The scintillation parameters, including light output, were measured by exciting the solutions with pulses from a linac. The experimental arrangement is shown in Figure 2. These parameters include: 10 to 90% pulse rise time, FWHM, decay time from 0.7 of maximum to 0.7/e, 10 to 90% integral rise time (IRT), and the relative peak and integral of the fluorescence pulse to a reference scintillator. The solutions were tested in 1 cm-thick Suprasil spectrophotometer cells and were bubbled with argon for two minutes before use. The samples were excited by 6 MeY electron pulses whose FWHM was 50 psec. The emitted light was viewed at an angle of 135° to the direction of the linac beam, as seen in Figure 2. This was done to reduce the Cerenkov contribution to the signal. Light from the samples was focused onto a Varian VPM 221D microchannel plate (MCP) photomultiplier tube (PMT), whose gain was l0. The light passed through a 335 to 427 nm FWHM band-pass filter. These signals were then sampled by a Hewlett-Packard sampling system Model 141A, with a remote sampling head. The signals were digitized with a CAMAC ADC and averaged and recorded with a DEC PDP-1l/34. A modification was made on the red-emitting liquid scintillator, L-735, listed in Table 1. The new scintillation solution, L-735A-1OT consists of 0.01 M

LDS-722 as final emitter and 0.10 M C-540A as intermediate wavelength shifter in a solvent consisting of 90 vol% benzonitrile (BN) and 10 vol% TMSN. The LDS-722 concentration was lowered from 0.02 Mm L-735 to 0.01

M in L-735A-lOT, because the solubility of LDS-722 decreased when the

Figure 2.

SAMPLE

TO SAMPLING OSCILLOSCOPE HEWLETT-PACKARD 140A

1430A

REMOTE SAMPLING HEAD HEWLETT-PACKARD

RADIATION SHIELD

LINAC BEAM TUBE

Schematic diagram of test configuration for the linac measurements.

(ORTEC 493)

CURRENT DIGITIZER

HIGH VOLTAGE

MPC DETECTOR

OPTICAL FILTER

FARADAY CUP

98

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

TMSN was added. L-735A has the same dye concentrations as L-735A-1OT but no TMSN. The relative sensitivity of L-735A-1OT to L-735A was measured by exposing

samples to pulsed X-rays. These pulses were generated by 50 keY electrons, with a FWHM of 850 psec, incident on a molybdenum target as shown in Figure 3. Characteristic X-rays are superimposed on a bremsstrahlung spectrum that extends out to the 50 keY endpoint. The X-rays pass from the vacuum through a 5 rn-thick beryllium window and are then filtered by a 3 m molybdenum foil which selectively passes the characteristic X-rays. The X-ray spectrum was determined by placing a 2 atm Xe proportional counter at the same position as the cell. A sample of the resulting spectrum, with the system response unfolded, is shown in Figure 4. The samples were contained in a cell 12.7 cm in diameter by 0.6 cm in thickness. The window facing the X-ray source was 0 2 mm-thick beryllium, and the one facing the detector. was 3.2 mm-thick Spectrasil-B. During mea-

surements the liquids were heated to 43°C, a temperature typical of that encountered under field conditions. The light was filtered with a 630 nm longpass filter and was detected with an ITT F4129 MCP PMT. The photocathode has an S-20 extended red response. The tube was as close as possible to the Spectrasil window for efficient light collection, and the output was sampled and averaged using a Tektronix 7854 digitizing oscilloscope with an S-12 sampling plug-in and an S-4 sampling head. The sampled pulses were then inte-

grated, and the relative sensitivities were determined from the ratio of the integrals.

All chemicals were used as received. The BN, PC, and TMSN were from Aldrich Chemical Co.. The chemical TMGE was supplied by Alfa Products. The laser dyes LDS-722 and Coumarin 540A were from Exciton Chemical Company, Inc.. The compounds TMPB and TEPB were contributed by Ethyl Corp.. RESULTS

Tetramethylgermanium

As can be seen from the data in Table 2, the addition of TMGE had no measurable effect on the decay time of the 4,4"-di-(5-tridecyl)-p-terphenyl, SG-180,5 in pseudocumene (PC). The relative pulse integral values actually decreased somewhat less than would have been expected from simple dilution of the PC by addition of the TMGE. Thus there is no evidence of any quenching by the TMGE. It was found previously that when TMSN or TBSN was added, the emission intensity also decreased in proportion only to dilution of the PC.3'4

One problem with TMGE is its high volatility: it boils at 43°C. Tetraethylgermanium has a highe. boiling point, 163°C, but it contains a lower percent-

Figure 3.

3 MIL THICK

MOLYBDENUM FOIL

XRAYS

TEST SAMPLE

SCOPE

TO TEKTRONIX 7854 SAMPLING

ITT TUBE

5 MIL THICK Be WINDOW

Experimental setup for the relative X-ray sensitivity measurements.

MOLYBDENUM TARGET

3 MIL THICK

TARGET LADDER

;

Figure 4.

01

2

I

15

I

I

20

I

25

'"-'-....I

35

I

I

40

ENERGY (KEy)

30

I

45

50

I

55

60

Measured X-ray spectrum for a 3-mil-thick Molybdenum target through a 3-mu-thick Molybdenum filter. The response of the system has been unfolded.

C-)

LU

F-

D 10 LU

z12

LU

z 14

(I-)

>-

18

20

22

0 0

101

X-RAY SENSITIVE LIQUID SCINTILLATOR DEVELOPMENTS

Table 2; Results of Linac Studies of 0.035 M SG-180 in PC (L-360) with Added Organometallic Compounds RT

Scintillator

(nsec)

FWIfM (nsec)

DECAY

IRT

(nsec)

(nsec)

System Response (Suprasil)

0.15

0.22

0.08

0.54

100%TMGE 0.035 M SG-180 (100% Pc) 0.035 M SG-180 10% TMGE 0.035 M SG-180 25% TMGE 0.035 MSG-180 35% TMGE 0.035 M SG-180 70% TMGE 0.035 M SG-180 10°/a TMPB 0.035 MSG-180 20% TMPB 0.035 M SG-180 10% TEPB 0.035 M SG-180 20% TEPB

0.16 0.55 0.55 0.60 0.55 0.55 0.26 0.18 0.20 0.17

0.25 2.56 2.56

0.11 1.80 1.75

2.61 2.61 2.71

1.80 1.80 1.80 1.38

1.25 5.26 5.34 5.45 5.69 5.49 3.86

1.31

3.28

0.99 0.46

2.71 1.72

1.92 1.74 0.89 0,41

Peaka

lnt.a

1.00 1.03 0.90 0.76 0.65

1.00 1.04 0.91

-

--

042

-

0.78 0.68 0.33 0.22

--

apeak and mt. values referenced to L-360.

age of germanium by weight, 38.5% vs 55% for the tetramethyl compound. There is also a question of long term chemical stability of solutions containing TMGE. Further studies are contemplated. Tetramethyllead and Tetraethyllead emission The lead compounds were disappointing in that they quenched the seriously. Addition of 10 vol% of TMPB reduced the light output to less than 50%. When 20% TMPB was added the signal was dominated by the Cerenkov contribution, precluding any conclusive information. It was hoped that the somewhat larger alkyl groups in TEPB would reduce the extent of quenching, but it appeared to be somewhat worse. In addition, there was some decomposition, possibly photochemical, of the TEPB as the solutions stood in the hood before testing. A small amount of a clear, apparently crystalline, precipitate accumulated on the bottom of the sample vials. Table 3 shows the relative sensitivities of L-735A and L-735A-1OT to 17-keV X-rays. Included in the table are predicted relative sensitivities when 10 vol% Table 3. Effects of the Addition of 10 Volume Percent TMGE, TMSN, or TMPB or BN Containing 0.01 M LDS-722 and 0.10 M C-540A. Sensitivities are Relative to the

Scintillator with no Organometallics. Excitation was by Molybdenum-filtered Molybdenum X-rays.

Fluor Candidate

Relative Energy Absorption

Predicated Relative Brightness

Observed Relative Brightness

1.00 3.04 3.15 4.02

1.00

1.0

2.74 2.84 133a

2.8

0.01 M LDS-722; 0.10 M C-540A in: 100°/a BN

10% TMGE / 90% BN 10% TMSN / 90% BN 10% TMPB / 90% BN aincludes quenching by the TMPB.

-

102

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

TMGE, TMSN, or TMPB is added to BN. Energy absorption coefficients, as tabulated in Reference 6, were used to generate curves for the BN-based scmtillators. These curves were folded with the X-ray spectrum shown in Figure 4. The relative energy absorbed, the predicted brightness, and the observed brightness are shown in Table 3. The predicted relative sensitivities include a factor for the dilution of the BN, and in the case of the lead compound, consideration for the quenching by the lead compound. The agreement between the predicted and observed values for TMSN is excellent. CONCLUSIONS

The compound TMSN still appears to be the best compound that we have tested to enhance the sensitivity of an organic scintillation solution to low energy X-rays. Although the addition of lead would greatly increase the energy deposited, the severity of the quenching more than offsets the gain made, and no further work is anticipated with these lead compounds. BBTMGE shows promise as an additive for lower energy X-rays. There may be more work in this direction once the question of long-term chemical stability is resolved.

ACKNOWLEDGEMENTS

The authors wish to thank our colleagues, Mr. K. Moy and Dr. S. Iverson, for many helpful discussions concerning X-ray dosimetry. REFERENCES

Ogle, J.W., D. Thayer, L.D. Looney, F. Cverna, G. Yates, C. Iverson, S.S. Lutz, M.A. Nelson, and B. Whitcomb. "Radiation-Induced Imaging System Over Long Fiber-Optic Bundles," in High Bandwidth Analog Applications of Photonics, SPIE 720, pp. 24-30 (1986). Bicron Corporation, 12345 Kinsman Road, Newbury, OH 44065.

Berlman, I.B., L.A. Franks, S.S. Lutz, J.M. Flournoy, C.B. Ashford, and P.B. Lyons. "A High-Z Organic Scintillation Solution," in Advances in Scintillation Counting, (Banff, Alberta, Canada: Faculty of Pharmacy and Pharmaceutical Sciences, University of Alberta, 1983), pp. 365-375. Ashford, C.B., I.B. Berlman, J.M. Flournoy, L.A. Franks, S.G. Iversen, and S.S. Lutz. "High-Z Liquid Scintillators Containing Tin," Nuci. Instruments and Methods in Phys. Res., A243:131-136 (1986). Gershuni, S., M. Rabinovitz, I. Agranat, and I.B. Berlman. "Effect of Substituents on The Melting Points and Spectroscopic Characteristics of Some Popular Scintillators," J. Phys. Chem., 84(5):517-520 (1980).

Storm, E. and H.I. Israel, "Photon Cross Sections From 1 key to 100 MeV for Elements Z = 1 to Z = 100," Nucl. Data Tables, A7(6):565-681 (1970).

X-RAY SENSITIVE LIQUID SCINTILLATOR DEVELOPMENTS

103

APPENDIX

BA BN

BHTP (4-BHTP) C-480 C-540A DCM LDS-722 LDS-821 PC PBD Rh-610P SG-180 SR-640 TMSN

TPB

benzyl alcohol (alpha-hydroxy toluene) benzonitrile (phenyl cyanide, cyanobenzene) 4-bromo-4' '-(5-hexadecyl)-p-terphenyl Coumarin 480 (or Coumarin 102) Coumarin 540A (or Coumarin 153) red-emitting styryl laser dye, also Kodak dye No. 14567 far-red-emitting styryl laser dye (Exciton) near-infrared-emitting styryl laser dye (Exciton) pseudocumene (1, 2, 4-trimethylbenzene) 2-phenyl-5-(4-biphenylyl)-1 ,3,4-oxadiazole Rhodamine 610 perchiorate 4,4"-di-(5-tridecyl)-p-terphenyl Sulforhodamine 640 tetramethyltin 1,1 ,4,4-tetraphenylbutadiene

CHAPTER 10

Liquid Scintillation Alpha Spectrometry: A Method for Today and Tomorrow W. Jack McDowell and Betty L. McDowell

ABSTRACT Alpha spectrometry using liquid scintillation methods has matured into a technology showing great potential. The nonpenetrating properties shared by beta and alpha radiation make them both candidates for liquid scintillation counting/spectrometry. However, applying liquid scintillation to alpha spectrometry has been difficult, because inefficient light production by alpha particles led to poor alpha energy resolution and alpha-produced scintillation interference of beta- and gamma-produced scintillations. Recent developments in alpha liquid scintillation spectrometers and pulse-shape discrimination have removed the beta-gamma interference problem and greatly improved alpha energy resolution. In addition, selective solvent (liquidliquid) extraction methods (phase-transfers), setting the nuclide of interest into the scintillator, simplifies sample preparation and provides important additional information for nuclide identification. Such rapid, selective procedures have been developed for radium, uranium, thorium, plutonium, polonium, and the trivalent transplutonium elements. Presently, available energy resolution (230 keV FWHM) allows identification of many of the isotopes of these nuclides. Chemical separations methods and typical alpha-spectral results are presented in this chapter.

Promising methods of improving both energy resolution and chemical separations are suggested.

INTRODUCTION

The maturing of alpha liquid scintillation spectrometry into a useful radiometric tool is a fascinating story. It has been known since 19501.2 that alpha particles could be counted by liquid scintillation methods, but little practical use has been made of this knowledge. The problems of poor alpha energy resolution, quenching and variable scintillator response, and beta and gamma radiation interference all combined to make useful applications of liquid scintillation to alpha assay extremely limited. In 1964, DL. Horrocks and co-workers3'4 demonstrated that it was possible to obtain useful alpha energy resolution in liquid scintillation systems. They found it necessary to use a detector quite different from that used for beta liquid scintillation. It consisted of a single phototube facing a reflector, with 105

106

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

the sample between the phototube and reflector, and efficient optical coupling between the sample, reflector, and phototube. Both Horrocks and Hanschke5 working independently, demonstrated that a highly-reflective, diffuse-white, hemispheric reflector cavity, oil-coupled to a single phototube provided the optimum alpha energy resolution in a liquid scintillation detector. Hanschke produced both an experimental and a mathematical demonstration of this. A cross section of such a detector may be seen in Figure 1. Even though the detector described above was able to achieve useful alpha

energy resolution in a liquid scintillation system, the problems of variable energy response/quenching and beta-gamma interference with the alpha spectra still prevented useful application of liquid scintillation to alpha counting and spectrometry. The development of methods to overcome the latter two problems is the subject of this chapter. EXPERIMENTAL NARRATIVE

The experimental work reported here extended over 20 years. Some of it has never been reported, and none of it has been reported in this context, i.e., an overview of the development of Photon Electron Rejecting Alpha Liquid Scintillation (PERALS) spectrometry. This development began in 1967 with a need to determine aqueous/organic phase distribution of the alpha-emitting trivalent actinides to an accuracy of ± or better. It was found impossible to do this with plate-counting methods. Horrocks6 suggested that liquid scintillation might be a possible answer to

our problem. With well-characterized samples of trivalent actinides, we achieved the needed analytical accuracy using a commercial beta liquid scintil-

lation counter. However, problems with this approach soon appeared. The two-phase distribution data for some nuclides did not fit the pattern we expected. In order to determine that what we were counting was indeed the alpha from the trivalent actinide under investigation, we needed to see an energy spectrum

of the sample. The 1960 model Packard Tricarb liquid scintillation counter was modified to allow the energy signal to be displayed on a multichannel analyzer (MCA). The pulses carrying the energy information did not exit through the window produced by the upper and lower discriminators but instead were converted to logic pulses that were sent to the scaler. In order to see a spectrum and to determine what portion was being counted, the scaler pulses were picked off, modified, and sent to the gate input of the MCA while the energy analog pulses were sent to the energy input. This allowed the discriminators effect on the energy spectrum to be viewed on the MCA. With this arrangement, we were able to observe that some nuclides had interfering beta/gamma-emitting daughters and/or beta- , gamma- , or alpha-emitting impurities, and in all cases the energy resolution was poor, on the order of 800

Figure 1.

A cross section of a PERALS detector assembly.

RTV SILICONE SEAL

SLOT OR HOLE TO RELEASE TRAPPED AIR

CAVITY FILLED WITH SILICONE OIL

SURFACE

WHITE REFLECTING

PAIR OF POSITIONING PINS IN REFLECTOR

CULTURE TUBE

75-mm x 10-mm

SAMPLE IN

OIL OVERFLOW

- RESERVOIR FOR

108

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

to 900 keV full peak width at half maximum height (FWHM). In some cases the additional spectral information allowed a reasonably accurate determination of the test nuclide concentration, while in others it still did not yield the required analytical accuracy due to variations in the energy response of the scintillator, among other problems. The next step taken was to build a detector that would give improved energy resolution following the work of Horrocks34 and Hanschke.5 An experimental enclosure (Figure 2) that would accomodate a variety of phototubes and reflectors was built for testing various detector optical arrangements. In true scientific manner, we tested somewhere in the neighborhood of 25 different optical arrangements of our own before we conceded that Horrocks and Hanschke were essentially correct. The contributions to detector construction that we

have made are of the nature of refinements and compromises designed to enhance the practical usefulness of the detector. We added an improved reflecting surface, a light-coupling oil reservoir, and an insertion method for a small sample container (a 10 x 75 mm culture tube). The accurate positioning was an improvement suggested by Steve Musolino of Brookhaven National Laboratory.7 One of the crucial requirements uncovered by work on the detector was that the reflective surface must be highly reflective but nonspecular, i.e., diffusing. Magnesium oxide laid down in several layers and bonded by sodium silicate

solution (water glass) was effective, but barium sulfate with a very small amount of binder was equally good and was commercially available (Eastman high reflectance coating). The need for a diffusing reflector appears to arise from the nonuniform response of the photocathodes in multiplier phototubes. Figure 3 illustrates typical sensitivity profiles of 2 in phototubes. Spectra collected with the new detector gave much better energy resolution than those obtainable with the beta liquid scintillation spectrometer (see Figure 4), but we still had the problem of sample nonreproducibility and beta/gamma interference, either one of which alone could make alpha liquid scintillation of

very limited practical use. Progress at this point in the development was reported in Organic Scintillators and Liquid Scintillation Counting.8 Variable energy response and variable quenching are problems associated with the scintillation cocktail. Most of those commonly used for beta assay were, and are, aqueous-phase accepting. These cocktails contain detergents and aqueous/organic coupling solvents. Usually an aqueous sample is added to the scintillation cocktail, where it is incorporated by the constituents of the cocktail into a reasonably clear and homogeneous solution. Variations in the matrix cause variable quenching, but because beta-emitting nuclides produce a continuous spectrum from zero to maximum, this quenching variation can usually be corrected by measuring the radiation produced in that sample by a known external or internal source. Quench corrections of this type cannot be used with alpha spectra. Initial small amounts of quenching do not reduce the counts under an alpha peak but simply shift the alpha peak to a lower voltage/ energy scale position. More severe quenching can push the alpha peak out of

LIQUID SCINTILLATION ALPHA SPECTROMETRY

Figure 2.

109

Experimental light-tight enclosure in which various reflector arrangements were tested.

the detectable region, but the reduction in count is not a simple function of the amount of quencher added. First the count reduction is small, then, with more

quencher, large, and then, with still more quencher, small again, as the left edge, median and right edge of the bell-shaped curve pass below the detection threshold.

Figure 3.

Typical sensitivity profiles of 2-in multiplier phototubes.

LIQUID SCINTILLATION ALPHA SPECTROMETRY

111

232Th. -J

w

zz

I

U

239Pu

Ui

3(I)

z 0 0

(a)_

:

CHANNEL NUMBER Figure 4.

Comparison of the spectra of 232Th(4.O1 MeV) and 239Pu(5.t9 MeV) obtained on a beta liquid scintillation spectrometer (a) and on a PERALS spectrometer.

Thus, it became clear that a method of incorporating the alpha-emitting nuclide into the scintillation cocktail in some reproducible manner was necessary so that one could expect the peak for a given alpha energy to appear at the same position on the amplifier voltage output; thus at the same position on a

spectrum as that displayed on a multichannel analyzer. Application of the extensive knowledge of liquid-liquid extraction existing in the Chemical Separations Research group at Oak Ridge National Laboratory9 provided a solu-

tion to the cocktail reproducibility problem. Progress at this point was reported at the 1974 International Solvent Extraction Conference)° Developing a method of placing the alpha-emitting nuclide into the scintillator so that each time the scintillator and its response to alpha energies would be the same required the production of an identical organic-phase-soluble com-

pound for the nuclide in each sample. Variable quenching, either color quenching or chemical quenching, could not be tolerated. The first approach in many cases has been to transfer the nuclide into an organic phase by using liquid-liquid extraction or "solvent" extraction techniques and then add a portion of the organic phase to a scintillator. Our solution to this problem was to incorporate an extractant into the scintillator thereby producing an "extractive scintillator". There are several extractants one may choose for this purpose,

112

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 1. Properties of Selected Solvent Extraction Reagents Useful in Extractive Scintillators Type Cation Exchange

Example Bis (2-ethylhexyl) Phosphoric acid

Neocarboxylic acids

Ion Pair Coordinators From sulfate systemsa

Primary Alkyl Amines.

MW>250.

Metals Extracted Alkali and alkaline earths weakly extracted; pH 4-14. Lanthanides/actinides (Ill) strongly extracted; pH 3-5. Actinides (lVVl) strongly extracted to pH 0.5. Weak extractant for most ions pH 4-7. Strong Ext. for some ions in combination with appropriate crown ethers. From Sulfate; Fe (Ill), Y, Zr, Nb(V), Tc(Vll), Pd(lI), ln(lll) Sn(lI), Eu, Hf, Ta(V), Re(Vll) Os(IV), Mo(Vl), Pu(IV), U(IV) Th.

Secondary alkyl

amine, MW>250 (extraction varies

From Sulfate: Zr, Nb(V), Mo(Vl) Tc(Vll), Pd(Il), Ta(V), Re(Vll), Os(IV).

with Structure.)

Neutral Coordinators

Tertiary alkyl

From Sulfate; U(Vl).

amine, MW>250. Quat. ammonium MW>300.

From Sulfate; Mo(Vl), Tc(Vll) Ta(V), Re(VlI), Os(VI), Pd(Il).

Trioctyl phosphine oxide.

From Nitrate; U(Vl), Zr, Tc(Vll) Au(lIl), Th, Np(IV)(Vl), Pu(lV)(Vl), Pa, Hf. From Chloride; Au(llI), Zn, Zr, Sn(IV), Sb(llI), Cr(IV), Mo(lV), Fe(lll), Th, U(lV)(Vl) Pu(IV)(VI), Ga(llI), Nb, Bi.

aExtraction from nitrate and chloride systems are not listed because the amine salts of these ions are highly quenching and therefore not useful in scintillators. Note: Amine perchlorates are not extractants. Perchloric acid can be used to strip metals from any of the ion-pair-coordinator extractants.

each having different metal-ion selectivities. Table 1 lists some extractants that are currently used and shows their selectivities from various aqueous systems. Figure 5 shows the quenching (peak shifting) effects of some organic phase compounds. The use of an extractive scintillator involves a simple two-phase equilibration of the scintillator with an appropriate aqueous phase. Then the scintillator is placed in a small culture tube, the tube placed in the PERALS detector, and the spectrum collected. Thus, we found that in most cases the resolution of the sample nonreprodu-

cibility problem was to prepare a water-immiscible scintillator containing a liquid-liquid extraction reagent (in addition to the fluor and an energy transfer agent) and to extract (phase transfer) the nuclide-of-interest into such a scintil-

85

65

70

75

0

Figure 5.

3-

w

0

zUJ 80

I-

0

U-

x

90

95

too

0.02

0.04

0.16

AND I N DA S

0.12 0. 4 0.08 0. 0 EXTRACTANT CONCENTRATION CM)

0.06

The quenching effects of various reagents.

\

0.18

TBP

0.20

114

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

lation cocktail. The material extracted into the scintillator solution had little or no effect on its energy response characteristics. This development resulted in a reproducible, calibratible system where the peak for 4.0 MeY 232Th or 5.15 MeV 239Pu, for example, would always appear at the same place on an energy scale. A number of applications of alpha liquid scintillation were made at this point and some were reported at the 1979 International Conference on Liquid Scintillation Counting)"2 One additional, serious problem that impeded the use of liquid scintillation as a useful, practical method of alpha spectrometry, was the interference from beta and gamma radiation. Beta and gamma radiation are about 10 times more

effective, keY for keV, in producing light in a liquid scintillator than alpha particles)3 Thus, if a sample contains beta- and/or gamma-emitting nuclides, any alpha peak will be underlain with a beta/gamma continuum. This makes the alpha count background high, variable, and difficult to determine, and it sometimes even makes the alpha peak difficult to locate. In the following we describe how this problem was dramatically and elegantly solved. The work we were doing with alpha liquid scintillation attracted numerous visitors to our laboratory. Among them was an electrical engineer named John Thorngate. We were explaining to him the terrible problem that the beta and gamma pulses presented when he said, "I can get rid of those pulses." The next day John returned with an armload of NIM modules and within an hour we were seeing liquid scintillation alpha spectra free of beta and gamma interfer-

ence. By afternoon we had collected sufficient information to publish a paper. ' There is about a 40 nanosecond (nsec) difference in the length of beta- or gamma-produced and alpha-produced pulses in a liquid scintillation system.'3 It is possible to sort these pulses electronically and send only the longer, alphaproduced pulses to the multichannel analyzer. This technique is extensively used for separating gamma- and neutron-produced pulses. Similar electronic

circuitry was adapted and improved for beta/gamma - alpha separation. In currently available alpha liquid scintillation detectors the rejection of beta/ gamma pulses is > 99.95 efficient,'5"6 thus it is possible to collect an alpha spectrum that is essentially free of beta/gamma interference. At this point, we had the means of removing or minimizing the poor alpha energy resolution, the quenching and variable scintillator response, and the beta and gamma radiation interference. This provided the technology for the development of a practical method of producing liquid scintillation alpha spectra. We called this new development Photon Electron Rejecting Alpha Liquid Scintillation (PERALS) spectrometry. The development won an IR-lOO award in 1981 and the PERALS spectrometer is now in commercial production.'7 A number of laboratories are having excellent success with the method.

The special chemicals and extractive scintillators needed for the PERALS system are also commercially available.'8 Time will tell if the method is to fulfill its promise of a new and useful addition to the existing alpha assay methods.

LIQUID SCINTILLATION ALPHA SPECTROMETRY

115

RESULTS

As in virtually all cases the advance in alpha liquid scintillation methods described above resulted from a search for a suitable radiometric method for our own work, which was primarily a study of the chemistry of solvent extraction separations of metal ions. There was very little separate funding for this work. However, the development of a useful radiometric method for alphaemitting nuclides seems to fill a gap in the methods already available. Alpha energy resolution was not as good as surface-barrier or Frische-grid methods,

but nuclide quantification was better and sample preparation was easier. Sample preparation steps for PERALS alpha spectrometry usually were selec-

tive for one ion or a group of ions, and this selectivity supplemented the identification of nuclides by peak position. Typical PERALS Analysis

Describing a typical analytical procedure may summarize the descriptive material above and bring it to focus on its end use: let us begin with a solid sample presented for thorium and uranium analysis that appears to be soil mixed with leaves and other organic material. One- to two-gram portions of the sample are weighed into crucibles and ignited at 600°C until all the organic

material is destroyed. Then the material is placed in a solution of nitric, hydrochloric, and hydrofluoric acids; pressure vessels of organic fluorpolymers are best for this step)9-22 Sulfuric acid is added and the other acids are boiled away (fusion methods can also be used).23 Then the acidity of the sulfate system is adjusted to pH 1 to 2 in a 10 mL volume, and a two-phase equilibration extraction is made into a measured 1 .2 to 1.5 mL water-immiscible extractive scintillator containing 0.3 Mhigh-molecular-weight tertiary amine sulfate.

This uranium in the sample now will have been quantitatively transferred to the scintillator phase. A 1 mL portion is pipetted into a small culture tube, and the spectrum is collected on a PERALS spectrometer. The remaining aqueous phase is contacted with a 1.2 to 1.5 mL portion of an extractive scintillator containing 0.3 M high-molecular-weight primary amine Again 1 mL of the extract is counted. Multiplying the integral of the peak counts obtained in the first case by 1.2 or 1.5 and dividing by 0.9968 gives the disintegrations per minute (dpm) of uranium, and so calculating the integral of the peak obtained

in the second instance gives the dpm of thorium. The accuracy of such a procedure depends on two factors: (1) how carefully the sample is prepared and (2) how many counts are accumulated. Standard mineral samples have been run giving consistent results within 0.1

of the given value.

Solvent Extraction

This operation plays an important part in PERALS spectrometry in that it allows the radionuclide to be selectively and quantitatively transferred to the

116

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

(extractive) scintillator with little change in the scintillator composition. This makes it possible to obtain a reproducible calibration of the energy scale for a given extractive scintillator/aqueous phase composition. A solvent extraction system is composed of an organic extractant phase (one or more extraction reagents dissolved in an organic diluent) and an aqueous phase. These two phases are immiscible, but under proper conditions phase transfer of one or more species can occur across the organic/aqueous interface. Expansion of the interface by gentle agitation of the two phases allows phase equilibrium to be attained quickly, usually in one to two minutes.

In an aqueous/organic solvent extraction system, the distribution of a metal, M, between the two phases at equilibrium is defined by a distribution coefficient DM. D1 = (Conc. M)org/(COflC. M)aq

This is true at any phase ratio, Vorg/Vaq. However, it should be remembered that DN is a concentration ratio, and the amount of metal recovered depends also on the phase ratio. If we call the recovery factor, FR, the relationship FR

= Di.i(Vorg/Vq)

describes the total org/aq metal ratio. Thus if DM is 1000 (not an unusual value) and Vorg/Vaq is 1 then FR is also 1000, but if Vorg/Vq = 1/100 or 1/1000

the recovery is much less. The percent recovery in the first case (1:1 phase ratio) is lOOx 10001(1000 + 1) = 99.9%; while in the second case (phase ratio 1:100), the percent recovery is 100 x 10/(10 + 1) = 90.9%, and in a case where the phase ratio is 1:1000 the recovery will be only 50%. These simple calculations make obvious the importance of phase ratio in analytical applications of solvent extraction. It should also be emphasized that a distribution coefficient

depends on the composition of both the organic and aqueous phases. If the

composition of either phase is changed, the distribution coefficient can change. It is obvious that one should be aware of distribution coefficients and phase ratios in following, and more particularly in modifying, procedures for PERALS.

OUTLINES OF PROCEDURES

Some procedures that have been tested in our laboratory and shown to work well are briefly described below. Certain requirements that apply to all samples

should be noted. Samples to be counted should be optically clear and as colorless as possible. The objective is to get as much light as possible per alpha

event to the phototube. All samples should be sparged with a dry, toluenesaturated oxygen-free gas. Dissolved oxygen in the sample impairs both energy

and pulse-shape resolution. Any aqueous/organic system that transfers a

LIQUID SCINTILLATION ALPHA SPECTROMETRY

117

Table 2. QuenchIng and Nonquenching Ions and Molecules Relative Degree of Quenching SpecIes In Organic Phase Any yellow colored material. Color quenching of varying severity depending on the intensity of the color.

Chloride salts of amines and most chlorine substituted organics and dissolved HCI. Nitrate salts of amines, nitrate substituted organics and dissolved HNO3.

Severe chemical quenching.

Severe chemical quenching plus color quenching.

Alcohols and ketones.

Moderate quenching.

Ethers.

Slight quenching

Sulfate salts of amines.

Minimal quenching.

Phosphates, phosphate esters and phosphine oxides.

Minimal quenching.

quenching ion or molecule to the organic phase must be avoided for best results (see Figure 5 and Table 2). Counting on Glass-Fiber Filters24

For these two procedures to succeed the filter must be relatively clean, i.e., free of material of a dark color. Place the filter (up to 2 in diam) in a lOx 75 mm culture tube and add 1 mL of a PERALS aqueous-immiscible scintillator. Use either extractant-free or HDEHP [bis-(2-ethylhexyl)phosphoric acid] - containing scintillator. Momentarily evacuate the tube to water-aspirator pressures and release to argon. Do this twice to remove entrapped air and then count on a PERALS spectrometer. This allows separate determination of alpha and beta events. The spectrum of alpha energies is usually sufficiently resolved to allow identification of the alpha-active material. Uranium Activity on Cellulose Filters25

Cellulose air filters, clean or dirty, can be placed easily in solution and assayed as follows: the filter paper (up to 2 in diam) is placed in a 2 dram, screw-cap borosilicate glass vial and heated, open, in an oven or furnace at 500°C for 2 hr. After removing and cooling, add 2 to 5 drops of concentrated nitric acid, 3 drops of 30% hydrogen peroxide, and 1 drop of a saturated aluminum sulfate solution to the vial. The vial is then heated to 200°C to remove the nitric acid. The solids remaining are redissolved in a solution 1 M in Na,SO4 and 0.01 Mm H2SO4. A measured quantity of an extractive scintilla-

118

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

tor containing a high molecular weight tertiary amine sulfate is then added to the vial, and the two phases are equilibrated for 2 to 3 mm. After the phases have separated, 1 mL of the organic phase is counted on a PERALS spectrometer. Alpha spectra thus obtained usually allow identification of the uranium isotopes present. The results are uniformly more accurate than direct counting of the air filter. Gross Alpha in Environmental Materials26

With the sample in solution in nitric or hydrochloric acid, add

0.5

g of

LiC1O4 and 1 to 1.5 mL of 0.1 M HC1O4. Evaporate under heat lamps or in an

aluminum block at 160°C until boiling of the sample ceases and the first perchloric acid fumes appear. Cool the beaker and add 5 to 7 mL of water to dissolve the viscous residue. Measure the pH of the solution; it must be between 2 and 3.5. Transfer the solution to a small equilibration vessel and add a measured quantity, 1.2 to 1.5 mL, of an extractive scintillator that is 0.2 M

in HDEHP. Equilibration will transfer most of the alpha-emitting nuclides, with the exception of radium and radon, to the scintillator. Lanthanides are also extracted. Counting on a PERALS spectrometer will usually allow identification of the nuclides as well as an accurate quantification of them. Uranium and Thorium27

With the sample in solution as above, add sulfuric acid and sodium sulfate

and convert to a sulfate system at pH 1 to 2. An initial extraction into a scintillator containing 0.3 M high molecular weight tertiary amine sulfate such as trioctylamine sulfate will quantitatively remove uranium. A second

extraction from the same aqueous using a scintillator containing a high molecular weight primary amine sulfate such as 1 -nonyldecylamine sulfate will remove thorium. Each organic extractive scintillator solution is then

sampled and counted on the PERALS spectrometer. Normal uranium in secular equilibrium will show the double peaks due to 238U and 234U. How-

ever, natural processes can concentrate 234U in water, and sediments and artificial processes have concentrated 234235U; thus, not all uranium now in the environment is "normal" uranium. If the uranium is highly enriched in 235U, the spectrum will show primarily 2U. Spectra of tailings from the enrichment processes will show a predominance of 238U. Uranium and Thorium in Phosphates28

These elements can be separated from a variety of phosphate-containing materials, e.g., fertilizers, bones, teeth, animal tissues, and wastes. The sample is dissolved and placed in a nitrate or nitrate perchlorate solution. Sufficient

aluminum nitrate is added to this solution to complex the phosphate. The solution is then contacted with a toluene solution of trioctyl phosphine oxide

LIQUID SCINTILLATION ALPHA SPECTROMETRY

119

(TOPO). Both uranium and thorium are transferred to the organic phase. The organic phase is then stripped with an equal volume of 0.5 M ammonium carbonate solution, the ammonium carbonate is evaporated, the sample is converted to nitrate or nitrate/perchlorate, and any entrained organics are destroyed. The clear solution is then converted to a sulfate system and treated as above. In many nonphosphate samples uranium can be coprecipitated with magnesium hydroxide or otherwise concentrated and extracted from a sulfate system into a scintillator containing tertiary amine sulfate. Polonium29 The radioisotopes of polonium (usually 210Po) have been difficult to analyze with accuracy using the conventional methods. Using PERALS, however, the procedure is simple, rapid, and accurate. With the sample in solution, add 3 to 5 mL of concentrated phosphoric acid and evaporate to remove other acids. Transfer this phosphoric acid solution to a small equilibration vessel using 3 to 5 mL of water. Add 1 mL of 0.1 M HC1. Add a measured volume, 1.2 to 1.5 mL, of an extractive scintillator that contains 0.1 M TOPO. Equilibrate and count 1 mL on a PERALS spectrometer. Because of the minimal chemical manipulations required, the accuracy of this determination easily can be better than ±lWo.

Plutonium30

Plutonium can be chemically separated from all other elements except neptunium and counted quantitatively by this procedure, and the 237Np peak can be resolved by its energy difference from 239240Pu. The initial extractant in this procedure is 0.3 M high-molecular-weight tertiary amine nitrate (TANO3) in toluene. The sample should be in solution in 3 to 4 Mtotal NO3 and 0.5 to 1 M HNO3. The plutonium is reduced with ferrous sulfate and reoxidized to Pu(IV) with sodium nitrite. This solution is contacted with not less than /4 its volume of TANO3 solution. Equilibrate and separate the aqueous phase. Wash the organic phase with two '/4-vol portions of 0.7 MHNO1. The aqueous from the first equilibration and the washes can be combined and analyzed for uranium,

if desired. Plutonium can be stripped from the organic phase either with perchloric acid or, after diluting the organic phase with 2-ethylhexanol, with 1 N H2SO4. The plutonium is reextracted into a scintillator containing HDEHP in the first instance and into a scintillator containing 1 -nonyldecylamine sulfate in the second instance. Radium31

A new extraction reagent allows radium to be separated from other alkaline earth elements and the following procedure allows separation from many, if not all, other cations. The sample solution is spiked first with '33Ba, and 10mg

120

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

of barium carrier is added. Then radium is precipitated from the sample solution as barium/radium sulfate. Next the precipitate is converted to the carbonate by heating it with a saturated potassium carbonate solution, and the separated barium/radium carbonate is dissolved in dilute acid. The pH of this solution is then adjusted to between 9 and 10, and the radium extracted selectively into an extractive scintillator containing a synergistic reagent mixture of a high-molecular-weight-branched carboxylic acid such as 2-methyl-2heptylnonanoic acid (HMHN), 0.1 M, and dicyclohexano-21-crown-7 (DC21C7), 0.05 M. The extraction is quantitative, and any chemical losses in the coprecipitation and metathesis steps can be corrected by gamma counting the '33Ba. Radium is counted on a PERALS spectrometer, initially showing a single alpha peak each for both 226Ra and 224Ra. All the alpha peaks from both

radium daughters can be seen if one waits for their ingrowth. The percent relative error in this determination, with appropriate counting statistics, is usually less than ±5% and can be much less. CONCLUSIONS

The basic steps leading to the development of the PERALS system of sample assaying for alpha-emitting nuclides have been presented. The many blind alleys and dead ends explored were not. This is not the end of the story, however. More selective and perhaps simpler procedures are certainly possible. Although pulse-shape resolution (the separation of betagamma pulses) hardly needs improvement, energy resolution does. We believe that this is possible. There seems to be no fundamental reason why resolution should be limited in such a system. We hope to be able to pursue this quest in the future. ACKNOWLEDGEMENTS

The Department of Energy sponsored this work under contract No. DEACO5-840R21400 with Union Carbide Nuclear Corp. and with Martin Marietta Energy Systems.

G.N. Case is responsible for most of the laboratory development work described above.

C.F. Coleman, A.P. Malinauskas, and R.G. Wymer provided much-needed encouragement and shielding from many negative comments by higher supervision and our sponsors. C.F. Coleman invented the name and acronym PERALS.

LIQUID SCINTILLATION ALPHA SPECTROMETRY

121

REFERENCES

Broser, I. and H.P. Kallmann. "Uber den Elementarprozess der Lichtanregung in Leuchtstoffen durch alpha-Tielschen, Schnelle Elektronen und Gamma -Quanten II," Z. Naturforschg., 2A:642-650 (1947). Horrocks, D.L. Applications of Liquid Scintillation Counting, (New York: Academic Press, Inc., 1974), P. 2. Horrocks, D.L. "Alpha Particle Energy Resolution in a Liquid Scintillator," Rev. Sci. Instru,n., 35:334-340 (1964). Horrocks, D.L. and M.H. Studier. "Low Level Plutonium -241 Analysis by Liquid Scintillation Techniques," Analytical Chem., 30:1747-1750 (1958).

Hanschke, T. "High Resolution Alpha Spectroscopy by Liquid Scintillation Through Optimization of Geometry," PhD Dissertation, Hannover Technical University, Hannover, FRG (1972). Horrocks, D.L., Personal Communication 1967. Musolino, S., Personal Communication, Ca. 1985. McDowell, W.J. "Liquid Scintillation Counting Techniques for the Higher Actinides," in OrganicScintillators and Liquid Scintillation Counting, D.L. Horrocks and C.T. Peng, Eds. (New York: Academic Press, 1971), PP. 937-950. Coleman, C.F., C.A. Blake, and K.B. Brown. "Analytical Potential of Separations by Liquid Ion Exchange," Talanta, 9:297 (1962). McDowell, W.J. and C.F. Coleman. "Combined Solvent Extraction-Liquid Scintillation Method for Radioassay of Alpha Emitters," Proceedings of International

Solvent Extraction Conference 1974, (London: Society of Chemical Industry, 1974), pp. 2123-2135.

McDowell, W.J., E.J. Bouwer, and J.W. McKlveen. "Application of the Combined Solvent Extraction-High Resolution Liquid Scintillation Method to the Determination of 230Th and 234238U in Phosphate Materials," in Liquid Scintillation Counting: Recent Applications and Developments, Vol. 1, C.T. Peng, D.L. Horrocks and E.L. Alpen, Eds. (New York: Academic Press, 1980), pp. 333-346. McDowell, W.J. "Alpha Liquid Scintillation Counting; Past, Present and Future," in Liquid Scintillation Counting: Recent Applications and Developments, Vol. 1,

C.T. Peng, D.L. Horrocks and E.L. Alpen, Eds. (New York: Academic Press, 1980), pp. 315-332.

Horrocks, D.L. Applications of Liquid Scintillation Counting, (New York: Academic Press, Inc. 1974), PP. 29-33. Thorngate, J.H., W.J. McDowell, and D.J. Christian. Health Physics, 27:123 (1974).

McDowell, W.J. "Alpha Counting and Spectrometry Using Liquid Scintillation Methods," NAS-NS -3116, (Technical Information Center, U.S. Dept. of Energy P.O. Box 62, Oak Ridge, TN 37831, 1986), PP. 53-58, 66. Kopp, M.K., Oak Ridge Detector Laboratory, Personal communication 1988. Manufactured by Oak Ridge Detector Laboratory, Inc., 139 Valley Court, Oak Ridge, TN 37830. ETRAC (East Tennessee Radiometric/Analytical Chemicals) Inc., 10903 Melton View Lane, Knoxville, TN 37931.

Farrell, R.F., S.A. Matthes, and A.J. Mackie. "A Simple Low-Cost Method for the Dissolution of Metal and Mineral Samples in Plastic Pressure Vessels," Bureau of Mines Report of investigations; No. 8480, (1980).

122

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Matthes, S.A., R.F. Farrell, and A.J. Mackie. "A Microwave System for the Acid Dissolution of Metal and Mineral Samples," Bureau of Mines Technical Progress Report-120, (1983).

For heating in a furnace or oven, Parr Inst. Co., Teflon-lined pressure vessels numbered 243AC, T303, and 012880 have been used. For use in microwave ovens Parr lists No's 4781 and 4782 as suitable. CEM Corporation in Matthews, NC supplies a microwave dissolution apparatus complete with oven and vessels of their own design. Chiu, N.W., J.R. Dean, and C.W. Sill. "Techniques of Sample Attack Used in Soil

Analysis," A Research Report Prepared for the Atomic Energy Control Board, Ottowa, Canada, INFO-0128- 1 (1984). McDowell, W.J. and G.N. Case. Unpublished Data. McDowell, W.J. and G.N. Case. "A Procedure for the Determination of Uranium on Cellulose Air-Sampling Filters by Photon-Electron-Rejecting Alpha Liquid Scintillation Spectrometry," ORNL/TM 10175 (Aug 1986). McDowell, W.J. "Alpha Counting and Spectrometry Using Liquid Scintillation Methods," NAS-NS-31 16, Technical Information Center, U.S. Dept. of Energy P.O. Box 62, Oak Ridge, TN 37831 (1986), pp 88-89. Bouwer, E.J., J.W. McKlveen, and W.J. McDowell, "Uranium Assay of Phosphate Fertilizers and other Phosphatic Materials," Health Phys., 34:345-352 (1978).

Bouwer, E.J., J.W. McKlveen, and W.J. McDowell. "A Solvent Extraction Liquid Scintillation Method for Assay of Uranium and Thorium in Phosphate-Containing Material," NucI. Tech., 42:102-1 10 (1979). Case, G.N. and W.J. McDowell. "An Improved Sensitive Assay for Polonium-2l0

by Use of a Background-Rejecting Extractive Liquid Scintillation Method," Talanta, 29:845-848 (1982). McDowell, W.J., D.T. Farrar, and M.R. Billings. "Plutonium and Uranium Determination in Environmental Samples: Combined Solvent Extraction-Liquid Scintillation Method," Talanta, 21:1231-1245 (1974). Case, G.N. and W.J. McDowell. "Separation of Radium and Its Determination by Photon-Electron-Rejecting Alpha Liquid Scintillation (PERALS) Spectrometry," in Proceedings of the 33rd Annual Conference on Bioassay, Analytical & Environmental Radiochemistry, 6-8 Oct. 1987. Berkeley, California. (No editors, no page numbers given. Proceedings are Xerox copies of papers in ring binder.)

CHAPTER 11

Application of High Purity Synthetic Quartz Vials to Liquid Scintillation Low-Level 14C Counting of Benzene

A. Hogg, H. Polach, S. Robertson, and J. Noakes

ABSTRACT High purity synthetic quartz is evaluated for low-level 4C detection through liquid scintillation (LS) counting of benzene. A simple cylinder-cell vial design is presented, which incorporates a Teflon® stopper and Deirin shield. The counting characteristics (counting efficiency and background) of the quartz vials are compared to the Wallac Teflon and low-K glass vials, in new technology LS spectrometers (Pharmacia-Wallac, Quantulus, at ANU and Waikato, and the Packard 1050 CAILL, modified at UGA). The effect of the vial counting characteristics upon maximum and minimum detectable 14C age and the magnitude of the counting error, for both

low and high count rate samples, is examined. Synthetic quartz is shown to have counting characteristics superior to low-K glass and is equal to Teflon vials for most applications. Further, quartz does not require the extensive cleaning procedures necessary for Teflon.*

INTRODUCTION

The application of commercially available liquid scintillation (LS) counters for radiocarbon ('4C) dating has evolved over the last decade from the utilization of general purpose instruments, to new technology low-level LS spectrometers. The old technology counters, even with extensive in-house modifications, yield at best, between 60 to 75% 14C counting efficiency, characterized by a high background, which at best is > 10% of the '4C Modern reference signal (Polach et aL, 1983). Two modern low-level LS spectrometers, such as the modified Packard 1050 CA/LL and Pharmacia-Wallac, Quantulus, use electronic optimization (e.g., pulse shape, duration and ratio analyses), and Quantulus uses active and enhanced passive shield, to generally further reduce the background. The performance of these LS spectrometers in relation to '4C dating has been described.1'2 The best performance is at >80% '4C efficiency characterised by an ultralow background at 0.8% of the '4C reference signal. *TefIon is a Registered Trademark of E. I. DuPont de Nemours. 123

124

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Such improved performance gives higher precision, extended detectable age limit, or significantly smaller sample 4C LS radiometry.3 The critical application to low-level 14C spectrometry not only requires evaluation of LS counter performances, but it necessitates consideration of errors and assurances4 and counter unrelated parameters such as counting vial design, materials, benzene purity, and scintillant performance. This paper deals with different counting vial materials and their significance to low-level 14C detection characteristics and performance.

COUNTING VIALS

A great variety of counting vial materials are available and were applied to '4C isotope detection in benzene. Glass, quartz, Teflon, and Delrin were found suitable, albeit different in performance characteristics. Polyethylene (PE) and high density PE were not suitable due to their permeability of benzene. Shapes and volumes of vials include special purpose large volume cylinders (50 to 100 mL), square or cylindrical mini-vials within special holders (0.3 to 3 mL), and those based on the standard 20 mL LS counting vial design. High precision '4C low count rate determinations require calibration of each vial independently

for efficiency and background; therefore, it is general practice to reuse the same set of vials over many years. The glass and quartz vials have excellent physical counting properties. Their 14C signal and background detection efficiency remains constant over many years, and memory effects are nil with only minimal washing between samples. Quartz vials give inherently lower backgrounds than K-free glass counting vials.

Teflon and Delrin have excellent counting properties (high efficiency and inherently low backgrounds) but many researchers experience difficulties in their usage due to deformation over a period of time. Memory effects require rigorous cleaning procedures between samples, and alter efficiency over time (Figure 1). EXPERIMENTAL PROCEDURES AND VIAL DESIGN

Using available LS spectrometers, the authors tested vials of various designs

and materials in their laboratories. The University of Georgia, Packard counter, used low-K glass vials in all their measurements. The Australian National University and the University of Waikato used Wallac counters with Wallac Teflon and quartz vials of various origins and sample sizes. The Teflon vials were used as supplied by their manufacturer (Wallac Oy).5 One has a wall thickness of 0.9 mm the other of 1.1 mm. This affects their performance (cf. Table 1). The synthetic quartz vials were manufactured using material from three differ-

QUARTZ VIAL APPLICATIONS TO BENZENE COUNTING

125

10

4-

z2 Q)

i

i

I

60

30

90

Cumulative (%) of samples Figure 1.

Time series plot of SQPE (external standard endpoint) values for Teflon vials compared to low-K glass vials. Changes in endpoint in Teflon with time correspond to a related change in Count rate of the modern reference standard, from 24.2 to

23.7 ± 0.1 CPM, while the background remained the same (within statistical limits). Such changes in performance must be allowed for in high resolution 14C Counting.

ent sources: (1) Thermal Syndicate, UK (TS silica), (2) Mikro-Glasstechnik, Germany (MG silica), and (3) GM Associates, USA (GM silica). The silica vial consists of a flat bottomed cylindrical cell, 34 mm high and 25 mm in diam, sealed by a Teflon stopper containing a Viton 'O'ring (Figure 2).

The Teflon stopper contains a partially threaded central opening which is sealed, after the stopper is inserted, by a close fitting tapered stainless steel pin. Table 1. Counting Characteristics of Teflon, Silica and Low-K vials of 3 mL and 5 mL Benzene Samples at ANU and Waikato, Wallac Quantulus Counters VIAL

1OLa

B'

N0°

Ed

fMe

FM

tMAX9

tMIN"

3 3 3

0.29 0.24 0.32 0.43 0.36 0.50 0.45 0.58

27.6 24.3 27.5 28.0 27.5 22.8 47.5 47.7

83.7 73.5 83.3 84.8 83.3 69.2 84.4 84.7

51.3 49.5 48.6 42.7 45.8 32.2 70.8 62.7

24,200 22,500 21,700 16,700 19,300 9,600 15,800 12,400

55,400 55,100 55,000 54,000 54,500 51,700 58,000 57,000

39 42 40 39 40 44 30 30

0.9mm Teflon 1.1mm Teflon MG silica TS silica GM silica Low-K 0.9mm Teflon MG silica Note: a_isee Table 2.

3 3 3 5 5

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

126

Black deirin shield \\

0) a)

U)

a)

0.

EE U)

a) 0

I4)

CO

a)

Stainless

steel pin

E E

0 (0

Viton "0" ring Teflon stopper

E

-

° %- °E

EEE

CO CO

E E (Y)

Sample::. .:benzene.:.

EE CO L()

E \ Black delrin base 28 mm Figure 2.

Cross section of the experimental silica counting vial. A Teflon stopper with a single 0-ring is shown. Later experiments gave a significant reduction in vapor loss when a double 0-ring was used. The Deirin masks are made in two lengths, one to reach the 5 mL and the other the 3 mL sample meniscus.

The opening prevents the sample from becoming pressurized when the stopper is inserted. A black Deirin base, 10 mm high and 28 mm diam centers the vial along the optical axis of the photomultiplier tubes. A black Deirin shield is fitted over the vial, up to the meniscus of the sample benzene, to reduce optical cross talk. The practical counting volume range is 3 mL to 5 mL of benzene.

The ANU and Waikato experiments with quartz and Teflon used 15 g/L concentrations of Butyl PBD as scintillant, while UGA used 6 g/L PPO with 0.2 g/L secondary pulse length shifter POPOP in benzene. In all cases the counters were set up to give optimal '4C detection performance, e.g., highest obtainable '4C reference sample signal to background ratio.

127

QUARTZ VIAL APPLICATIONS TO BENZENE COUNTING

Table 2. Counting Characteristics of Low-K vial of 3 mL and 5 mL Benzene Samples at UGA Packard Counter VIAL

VOLa

Bb

N0c

Ed

fMe

FM

tMAX9

tMIN"

3

0.46

21.08

63.09

31.0

8653

51,400

45

0.75

35.17

64.74

41.0

5588

52,400

35

Low-K

(1050)

5

avolume of benzene (3 mL = 2.637g; 5 mL = 4.5g). bB = background, cpm. CN = 95% Oxalic Acid Modern reference standard, cpm. dE = % counting efficiency. = factor of Merit, N01..JB. Figure of Merit, E2/B. 9Age limit, tMAX-3000 mm for B, N0, and S (2 SD detection criterion, reference 7, p 96), years. hMinimum age, tMIN-3000 mm for B, N0, and S (1 SD detection criterion, Reference 7, p 97), FM

years. '7 mL low 40K borosilicate glass vial with Teflon cap liner. Counted using Packard 1050 modified with scintillating plastic detector guard.

As two important sources of sample 14C count rate variations are known, their effect on performance was tested: Handling the counting vials, exposing them to light, stirring the counting cocktail, or allowing an electrostatic discharge can cause spurious Counts at the beginning of the counting cycle.6 Both Teflon and MG silica were tested for this effect in the Wallac counters. Vials were: (1) rubbed by a Nylon cloth to induce an electrostatic charge, (2) agitated for 30 minutes, (3) exposed for 30 minutes to fluorescent light 15 cm distance, and (4) irradiated by gamma

and beta particles from an external high energy and countrate source. A sample of ca., 150% Modern (ANU-Sucrose '4C standard) was counted for 5 mm intervals for 30 minutes (6 repeats). All tests were negative. No induced count rate variation could be detected. Loss of sample benzene during prolonged low-level counting times (1 to 3

days) will also cause variation in the observed count rate with time. The evaporative loss in the Teflon vials were nil per day at room temperature. The evaporative loss of the experimental silica vials (Figure 1) was 0.7 mg/d. To minimize this a double 'O'ring stopper was later tested and achieved 0.1 mg/d losses.

Results The '4C low-level counting characteristics of Teflon, silica, and low-K glass in the Wallac counters are presented in Table 1. Similar data for the low-K glass vial from UGA using the Packard counter is presented in Table 2. The

performance of the vials with 3 mL and 5 mL of sample beazene in the various counters is characterized, by their background (B) and net 14C countrate (N0), at 95% of the Oxalic Acid International Reference Standard. To assist evaluation of data, the radiocarbon dating "factor of Merit" (fM = N0/'B), the conventional "Figure of Merit" (FM = E2/B), and the oldest (tMAX)7 and youngest (tMIN)7 detectable '4C age were included. In the Wallac counters the

128

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

best performance was obtained with Teflon at

5

mL followed by the MG

silica.

DISCUSSION AND CONCLUSIONS

The background and modern reference sample count rates determine the '4C dating performance of both the counter and counting vial: the magnitude of the counting error limits the minimum and maximum determinable ages as well as the minimum countable sample size. Sample size is another variable which affects '4C age resolution. While counted benzene volumes are usually fixed (in our study to 3 mL or 5 mL), the sample component within the benzene varies in practice. Indeed, all radiometric laboratories dilute the sample, CO2 or C6H6, with gas or benzene containing no '4C. A case study of sample sizes dated at the University of Waikato is given in Figure 3 (count rates are plotted against age). The full size (undiluted) samples must lie on the solid line. The vast majority are diluted and lie below the line. In Figure 4, the samples shown in Figure 3 are expressed as a cumulative plot of sample for a given count rate (Figure 4). The plot

indicates that the count rate rather than sample age is therefore, in practical terms, the limiting variable in '4C age determinations.

5

10

15

20

25

Net count rate (cpm) FIgure 3.

Scatter plot of routine 14C age determination samples counted in the Wallac Quantulus at Waikato over a 12 months period. The solid line gives age and count rate limits for undiluted 3 mL samples. Most samples were diluted with 14C free material to make up the required counting volume.

129

QUARTZ VIAL APPLICATIONS TO BENZENE COUNTING

866

739

737

863

Glass

C

0

-

735 860 w

w

a 733

a

0

(I)

U)

857 731

72921

25

May'88 Figure 4.

30

35

Set number

40

45

50

854

May'89

Cumulative % of samples shown on Figure 3, plotted against the net sample count rate. The plot demonstrates that in routine 14C dating the majority of samples have significantly lower count rates than the modern reference standard, primarily due to the high degree of sample dilution rather than age. High resolution, close to the limit of detection is therefore an important factor in routine 140 dating applications.

To assess the merit of vials, we have taken the 3 mL data (Table 1) for the Quantulus and related each type of counting vial tested to the best vial (0 9 mm thick Teflon). The difference in calculated counting error (1 Standard Deviation, SD), expressed in years, with respect to the net sample count rate is the merit of vials (Figures 5A and 5B). We can conclude the following: The majority of samples '4C dated at ANU and Waikato give low count rates, predominantly due to sample size, rather than the age of material submitted for dating. Whenever this is the case, high resolution '4C counting at the lowest attainable background and highest efficiency is very desirable. When dealing with low count rates samples (old or highly diluted, where "C signal approaches noise), a reduction in counting efficiency (E) is tolerable

only if accompanied by a compensating reduction in background (B). For example, a reduction of 84 to 74% E and almost proportional reduction in B only marginally affects tMAX (e.g., 0.9 and 1.1 Teflon, 3 mL, Table 1) and slightly increases the counting error by 15 years at 30 K years (Figure 5A). A reduction of 84 to 69% E without compensating reduction in B has a significant effect on tMAX (e.g., 0.9 Teflon and low-K, 3 mL, Table 1) and significant increase in the counting error by 90 years at 30 K years (Figure SA). When dealing with high Count rate samples (signal approaches reference stan-

130

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

100 0)

A 0)

60 A

0 40 0) 0)

... A AM

20 C-)

S.. DD.

0

0.66

1.38

AA

UU

UU

2.11

2.83

A

3.55

4.27

15

A Low-K glass

1.1 mm teflon

0

o MG silica

110 A

0) 0)

A

Co

0)

0 0

hull..

4 27

IIUI1U 8.61

III U II 12.96

U

U

. AA

A

--

17.30

21.64

25.89

Net count rate (cpm) Figure 5.

The difference in the standard deviation (SD) between the best performing vial (0.9 mm wall Teflon) and 3 other vials is given as "increase in counting error years." It is plotted against net count rate (CPM). Figure 5A covers the age range equivalent to 30 to 15K years for undiluted sample. Figure 5B covers the age range equivalent to

15 K years to 500 years for undiluted samples. All counting times were fixed at 3000 minutes.

QUARTZ VIAL APPLICATIONS TO BENZENE COUNTING

131

dard count rate) an increase in B is tolerable if 'Vo E becomes higher (e.g., 0.9 Teflon, B = 0.29 cpm, and TS silica, B = 0.43 cpm, both give the same tMIN, Table 1). However, should the E be significantly decreased and or B signifi-

cantly increased (e.g., low-K, Table 1 and Figure 5) the tMIN, tMAX, and error deteriorate significantly. The performance of the MG silica is closely related to that of the 0.9 Teflon in the Wallac counters. The tMAX and tMIN are not significantly different nor is there a significant difference in the error of determination at the low and high end of count rates (Figure 5). Silica varies in performance both in terms of efficiency and background. When suitable material selection is made, and the high resolution and ultralow background characteristics of the Quantalus are used to the utmost, silica will be an effective substitute for Teflon. The low-K glass vials in both the Quantulus and Packard counters have similar

counting performances (Table 1 and 2). It is hoped the improved silica vial counting characteristics exhibited by the Quantulus measurements can also improve the counting capability for the Packard counters. Since the silica vial configuration, as shown in Figure 2, could not be accommodated in the Packard 1050 under optimum 7 mL vial configuration, measurements will be delayed until suitable vials can be fabricated.

REFERENCES

I. Polach, H., G. Calf, D. Harkness, A. Hogg, L. Kaihola, and S. Robertson. "Performance of New Technology Liquid Scintillation Counters for '4C Dating," Nuci. Geophys. 2(2):75-79 (1988).

Noakes, J. and R. Valenta. "Low Background in Liquid Scintillation Counting Using an Active Vial Holder and Pulse Discrimination Electrons," paper presented at the 13th International Conference on Radiocarbon Dating, Dubrovnik, Yugoslavia, June 20-25, 1988. Otlet, R.L. and H.A. Polach. "Improvements in the Precision of Radiocarbon Dating through the Recent Developments in Liquid Scintillation Counters," paper presented at the 2nd International Symposium in Archaeology and '4C, Groningen, Holland, Sept 18, 1987. Polach, H. "Liquid Scintillation '4C Spectrometry: Errors and Assurances," paper presented at the 13th International Conference on Radiocarbon Dating, Dubrovnik, Yugoslavia, June 20-25, 1988. Polach, H.A., J. Gower, H. Kojola, and A. Heinonen. "An Ideal Vial and Cocktail for Low-Level Scintillation Counting," in Proceedings of the International Conference on Advances in Scintillation Counting (Banff, Canada: University of Alberta Press, 1983), pp. 508-525. Haas, H. "Specific Problems with Liquid Scintillation Counting of Small Benzene Volumes and Background Count Rates Estimation," in Proceedings of the 9th International Conference on Radiocarbon Dating (Los Angeles: University of California Press, 1979), pp 246-255. Gupta, S.K. and H.A. Polach. Radiocarbon Dating Practices at ANU (Canberra, Australia: Radiocarbon Dating Research, 1985), pp 175.

CHAPTER 12

An Introduction to Flat-Bed LSC: The Betaplate Counter

G.T. Warner, C.G. Potter and T. Yrjonen

INTRODUCTION

The use of liquid scintillators was first described in 1950.1,2 Neither paper

refers to the use of liquid scintillators for internal sample counting, but describe their use for alpha, gamma, and neutron counting. Reynolds shows a block diagram of a two PM tube coincidence circuit, and thus was the first to suggest this arrangement for liquid scintillation counting reduction of what he calls, "the well known inconvenience of PM tube background." Both papers refer to the use of benzene, xylene, and toluene, and to the addition of anthracene. Kallman reports the effect of napthalene and the use of terphenyl. The

paper usually cited as the first reference to internal sample counting was published in 1953. However a paper in l952, described a method of counting 14CO2 in toluene-PPO solution by condensing the gas at -80°C, and thus may have a prior claim. The first commercial liquid scintillation counter (LSC) is quoted as being marketed in 1954, by Packard, and the earliest reference found to the evaluation of the standard vial was 1957,6 although this may not have been the first. In this paper, the Wheaton 5 dram vial was compared for background and

CPM against 85 mL weighing bottles containing between 10 to 50 mL of scintillant. The Wheaton vial, having a lower background, was shown to be the container of choice. By 1957, Packard seemed to have adopted the Wheaton vial as standard, and their first model in 1954 may also have used a small sized vial. In the U.K. as late as 1968, equipments using the large weighing bottles were still being marketed on the basis of a larger sample volume incorporation. To reduce thermal noise from the PM tubes, the Packard LSC of 1957, in addition to counting coincidence, used a domestic deep-freeze to house the counting head. Even so, this particular counter had a background for 3H of 32 cpm and an efficency of 20o, giving a figure of merit of 12.5. However, these 133

134

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

figures were very good as a counter marketed in 1968 by a British company, 9 years after the Packard, it had a figure of merit of less than one (0.67). While

the Packard counter represents an early version of a liquid scintillation counter, modern equipment is of course based on the same use of vials to contain the scintillant and sample. Up until recently, cooling of the PM tube

was a standard feature. The only significant change has been in the instrumentation. THE FLAT-BED GEOMETRY COUNTER (THE BETAPLATE)

The Betaplate concept was designed in 1976, when it was recognized that the procedure for counting the 3H labeled DNA of cells filtered onto glass fiber discs could be improved. The principle use of this technique was in the mixed lymphocyte assay where the sample preparation proceedure is based on a cell harvester. Suspensions of 3H labeled cells are aspirated from a microtitre plate, filtered onto glass fiber sheets and washed using the harvester. The glass fiber filter is 10 cm wide by 25.5 cm long, and with one particular harvester (Skaatron Inc.), the samples are aspirated 12 at a time, each sample occupying an area 1 cm in diam, a whole filter holding 96 samples in a 6 x 16 matrix.

Conventionally, after the sheet is dried, the discs bearing the samples are individually removed from the sheet with forceps, placed in a standard or minivial and combined with a few milliliters of scintillant. It takes about 20 minutes to prepare 96 samples and load them into a standard LSC, and it is a very tedious operation. Recently, some degree of automation has been available from some manufacturers. The proposed design (Figure 1) did away with the vial, replacing it with a flat plastic container in which the whole filter sheet could be placed with the 96 samples intact. Scintillant would be added to the whole container and the PM tubes brought up close to each sample in turn; a flat-bed geometry. In the original drawing, the plate was to be scanned in a raster fashion. The

problem of cross talk (that is adjacent sample interference caused by light conduction) was considered and some black material in the form of lines or circles printed on the glass fiber filter between the samples was thought might be an effective method of reducing this problem. It has now been shown 89 that the printing reduces cross talk by two orders of magnitude for Tritium, and this forms an important part of the Betaplate concept.

RESULTS AND DISCUSSION

Some basic experiments designed to test the concept were performed using a metal jig made of two sandwiched together aluminium plates, each with a 1 cm diam hole at one end. This device is fitted into a standard vial while a small piece of glass fiber filter bearing 3H labeled cells could be placed between the

INTRODUCTION TO FLAT-BED LSC

135

/2 H

Figure 1. Layout drawing of the flat-bed counter. The photomultiplier tubes (1), are placed either side of the sample container (2) which could be moved in a raster fashion.

metal plates, either opposite the holes to measure sensitivity, or not opposite, and thus masked, to test for cross talk. With just a wetting of scintillant, the background countrate in the 3H chan-

nel was <10 cpm, and the efficiency was about 50o of that obtained in a standard vial with 2mL of scintillant. After some experimentation, the cross talk was reduced to <0. 1 °lo; thus, this crude representation of the principle, showed that cross talk and sensitivity were certainly good enough for work with 3H labeled cells, although at that time cross talk was considered to be too high with '4C. A simple, manually operated prototype constructed by Wallac Oy of Fin-

136

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

land gave efficiency values of 28% for 3H and 78% for '4C, showing that reasonable efficiences could be obtained. Their design staff suggested a sealed plastic bag, rather than a flat plastic box, as the sample container and a metal cassette as a carrier and a means of further cross talk reduction. A second fully automatic prototype (Figure 2) having a 2016 sample capacity was constructed. In our laboratory, the background is low, but not untypi-

cal of sites away from high radiation sources, or high enviromental back-

ground; however, the prototype was fitted with quartz tubes so the background figures are lower than the standard production Betaplate, although that is also available with quartz tubes if required. The production 1205 Betaplate is shown in Figure 3. With the Betaplate system, harvesting is the same save the glass fiber filters. They are especially formulated to be strong enough to prevent the sample areas

from easily breaking away, and they are printed with a grid to reduce cross talk. Having harvested the samples and dried the filter, it is then slid into the special plastic bag, 10 mL of scintillant added which, for standard glass fiber sheets is sufficent for all 96 samples. The scintillant is gently encouraged to permeate the filter by a hand roller, and the remaining edge is heat sealed. It is then placed in a metal cassette for counting. Time taken about 2 minutes. Glass fiber filters and the special heat-sealable plastic bags are shown in Figure 4.

In addition to glass fiber filters, nylon membranes are available for DNA work and the figures for background, efficiency, and cross talk, include data on nylon membranes (Tables 1-3). The low values of cross talk (Table 3), show that filter printing not only

Figure 2.

The layout of the MK II prototype.

INTRODUCTION TO FLAT-BED LSC

Figure 3.

137

Production model Betaplate (Model 1205 Pharamacia/Wallac). The counter has six counting heads and a loading capacity of 1920 samples.

Figure 4. The filter mats with a grid pattern are the standard type used for filtration and the "tile pattern" are the "Spot on" mats taking volumes up to 30 L.

138

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 1. Background Countrates (CPM) Scintillant Vol. Material lOmL Glassfiber Nylon

4mL

32p

3H

<2 <1

<7 <6

<6 <5

Note: These values were obtained using quartz photomultiplier tubes.

Table 2. Efficiency and Figure of Merit (F) Flatbed Counter Conventional LSC F Eff. Isotope Eff. F Material >2000 54% 110 Glassfiber 3H 47% 180 97% >1750 96% Glassfiber 32p >1000 84% 90% 324 Nylon The samples used were labeled cells and therefore show working" efficiencies rather Note: than theoretical maximums.

reduces cross talk for low-energy isotopes, but also works well with higher It has also been found that good results energy emmitters such as 14C and are obtainable with gamma emmitters that have low-energy electron emmission, such as 51Cr and 125j Specific cytotoxicity and receptor binding assays, therefore, can be performed using the Betaplate. While the Betaplate was developed for use with filtered cells, plastic backed filters have been developed for spotting small volumes of up to 30 L, which then can be dried onto the filter. This technique is used for the cytotoxic assay ("Spot-on" mats, Pharmacia-Wallac). A more recent development is a 400 L liquid-holding tray (Figure 5), containing 96 wells in the standard Betaplate format ('T' tray Pharmacia-Wallac) It was designed for use with scintillation proximity assay (SPA) kits requiring a 400 L sample container (Amersham International). The background count and efficiency with a liquid sample are similar to a standard vial geometry LSC. Correlations between standard LSC and Betaplate counting methods are shown in Figures 6 & 7. The advantages and disadvantages of the Betaplate system are summarised in Table 4. Another aspect of the Betaplate is the low bulk of disposables (Figure 8). The bags containing filtered samples save 95°lo of the plastic bulk and a similar reduction in the volume of scintillant used.'° The 'T' trays, while not having Table 3. Cross Talk (Nearest Sample) Material Isotope Glassfiber Glassfiber Nylon

Glassfiber Glassfiber

3H

32P 51Cr 1251

No lines 0.600% 3.25% 0.16% 0.36% 2.00/s

Printed lines 0.006% 0.080% 0.015% 0.01 5%

0.3%

INTRODUCTION TO FLAT-BED LSC

Figure 5.

139

The 'T' tray for liquid samples. (Well capacity 400 ML.)

LKB LSC

BETAPLATE

*

+

%BIBo 100

*

* 80

60

4

40

* 20 100

200

400

800

1600

3200

6400

FMOLE/TUBE Figure 6. Correlation between a standard LSC and a Betaplate counter: 1251 cAMP Scintillation Proximity Assay. (Data kindly supplied by Dr. Nigel Bosworth, Amersham International.)

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

140

LKB LSC

BETAPLATE

*

+

%BIBo 100

80

60

40

20

0 6

13

40

110

300

PG/TUBE Figure 7.

Correlation between a standard LSC and a Betaplate counter: 3H Thromboxane B2 Scintillation Proximity Assay. (Data kindly supplied by Dr. Nigel Bosworth, Amersham International.

quite the same reduction in volume, still show a considerable saving over conventional vials. In summary, the flat-bed geometry Betaplate has many advantages over the vial based geometry counter when filtered cells or volumes up to 400 L are being measured; the concept represents the first radical change in LSC design for over 35 years. Table 4. The Advantages and Disadvantages of the Betaplate System Disadvantages CPM only, no automatic quench correction possible, no external standard Some cross talk between samples, but much reduced by printing and sensible sample layout

Sample size limited to 400 L Advantages Ease and speed of sample preparation Economy of disposables Sealed samples reduce biohazards Low background Count rate together with good efficiency gives high Figures of Merit Lightweight, compact machine High sample loading capacity Six counting heads gives high sample throughout Low volume of waste for disposal

INTRODUCTION TO FLAT-BED LSC

Figure 8.

141

There is a 95% reduction in the volume of disposables when using filter mats. A 96 sample filter can be contained in a single standard vial.

REFERENCES

Reynolds, G.T., F.B Harrison, and G. Salvini. "Liquid Scintillation Counters." Phys. Rev. 78:488.(1950). Kallman, H. "Scintillation Counting with Solutions." Phys. Rev. 78:621(1950). Haynes, F.N and R.G Gould. "Liquid Scintillation Counting of Tritium Labeled Water and Organic Compounds," Science, 117:480-482(1953). Williams, D.L. "U.S.A.E.0 Document" LA-1484. (1952). Rapkin, E. in The Current State of Liquid Scintillation Counting, (New York: Gene & Straton New York. 1970), p. 45

Davidson, J.D. in Liquid Scintillation Counting, (New York: Pergamon Press. 1958), p 88.

Warner, G.T. and C.G. Potter. "A Method of, and Apparatus for, the Monitoring of a Plurality of Samples Incorporating Low Energy Beta-emitting Raioisotopes," Brit. Patent 1586966 (London: HM Patent Office 1980). Warner, G.T. and C.G. Potter. "Sorption Sheet for Sorbing a Plurality of Discrete Samples and a Method of Producing Such a Sheet," US Patent 4,728,792 (Washington: US Patent Office 1988).

Potter, C.G., G.T. Warner, T. Yrjonen, and E. Soini. "A Liquid Scintillation Counter Specfically Designed for Samples Deposited on a Flat Matrix," Phys. Med. Biology, 31(4):361-369(1986).

Warner. G.T., and C.G. Potter. "New Liquid Scintillation Counter Eases Vial Disposal Problems," Health Phys. 51 (3):385-386(1986).

CHAPTER 13

A Plastic Scintillation Detector with Internal Sample Counting, and Its Applications to Measuring 3H-Labeled Cultured Cells

Shou-li Yang, Ming Hu, Jia-chang Yue, Xiu-ming Wang, Xu Yue, Jie Li, Zi-ang Pan, and Zhong-hou Liu

INTRODUCTION

Liquid scintillation counting has the advantage of homogenous 4ir counting

geometry and resultant high counting efficiency, but it frequently requires complex sample preparation and produces excessive amounts of radioactive organic waste.' As a result, Cerenkov counting and other counting methods have been used.2 The plastic scintillation detector for determining soft beta nuclides, which is used without cocktail, has about a 30 year history.3-7 Inter-

nal sample counting with gas detectors has been used to measure soft beta emitting nuclides, but internal sample counting with plastic scintillation detectors has been less frequently reported in the literature. This paper describes a simple and convenient plastic scintillation method for routine internal sample counting.89 The application of this method, in measuring the amount of 3H incorporated into cultured cells, was tested for its homogeneity, counting efficiency, reproducibility, and spectrum analysis. MATERIALS AND INSTRUMENTS

The radioactive isotopes used as reference samples in this work were 3H and '4C labeled hexadecane. Also used were 3H labeled lysine and thymidine and as Na2SO4 solution diluted by an emulsion cocktail. All radio isotopes were made by the Chinese Academy of Atomic Energy. The plastic scintillator sheet (PSS) was made from styrene monomers puri-

fied by vacuum distillation. Fluors of PPO and POPOP at concentration levels of 10 g/L and 0.8 g/L, respectively were added to this monomer. The polymerization procedure was conducted in an oil bath at 110°C for 5 days. The PSS was pressure shaped into rectangular blocks 32 mm long, 15 mm 143

144

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

wide, and 0.25 mm thick. Two different detectors were made from the initial plastic sheet. One PSS detector was rectangular with base area dimensions of 24mm x 8 mm and a volume capacity of 0.996 mL. A second PSS detector was

circular with a diameter of 12 mm. A glass filter paper disc used for cell harvesting measured 12 mm in diameter. Cocktails were made using a fluor concentration of 6 g/L of butyl-PBD in a solvent of pure toluene or 2:1 ratio of toluene/Triton X-l0O. Radioactivity measurements were made using a Beckman Model LS-9800 liquid scintillation counter. The General Program and the Special Spectrum Analysis Program of the instrument were used in the counting procedures. A Hitachi Model S-520 scanning electron microscope was used for microscopic analyses. METHODOLOGY AND RESULTS

The plastic detector used in this study sandwiches the radioactive sample between two PSS, and seals it with a gluing optical coupling solvent (GOCS). The radioactive sample may be in solution or on solid support. The GOCS must have good qualities for both optical coupling and gluing the sample to the PSS. In general, aromatic solvents are superior for gluing. Solvents such as toluene, xylene, anisole, dioxane, ethyl acetate, chloroform,

or dichloromethane can be used to soften or dissolve the PSS. Additional emulsifiers such as Triton X-100 in GOCS are necessary for aqueous sample counting. In fact, the toluene cocktail was found to be an excellent GOCS. To prepare samples on filter paper support, they were dipped in cocktail. A radioactive sample, either dissolved or on support, is then fixed between two PSS by GOCS. It was found the GOCS should not be used in excess, otherwise the plastic sheets would become deformed. Finally, after gluing the two PSS together, they were placed in a 20 mL standard glass vial to be counted. The sample prepared for counting is a composite of radioactive material, plastic scintillator, and GOCS. Figures 1 and 2 are scanning electron microscope photographs, shown amplified at 5000 and 2500 times. Figure 1 shows a cut section of fiberglass paper, and Figure 2 shows a torn section of the glass filter paper support sandwiched between the plastic scintillator. Figure 2 also shows some broken granules of wrinkled and cracked plastic scintillator folded around and among the glass fibers. Harvested cells in a ruptured broken state due to the treatment by the organic solvents contained in the GOCS are also shown. These photographs reveal that the distance between the glass fiber and plastic scintillator is 0.3 to 2.4 m. COUNTING EFFICIENCY AND REPRODUCIBILITY

As known radioactive reference standards, 10 L of 3H n-hexadecane and '4C n-hexadecane were sealed between 2 plastic sheets and placed in an empty,

PLASTIC SCINTILLATION DETECTOR

145

Figure 1. A cut section of filter glass paper.

standard 20 mL glass liquid scintillation vial. The counting efficiencies of 3H

and '4C were 30.6 ± 1.5% (mean + S.D.) and 85.2 ± 0.5% using 100 L of GOCS with a toluene b-PBD-6 g/L cocktail. The counting efficiencies, without the use of GOCS, were reduced to 5.9 ± 1.3% and 43.1 ± 6.2%, respectively. These data show that GOCS not only increases 3H and '4C counting efficiencies but also decreases the relative standard deviation. For 35S, as a Na2504 aqueous sample, a 2:1 toluene/Triton X-100 b-PBD-6 g/L cocktail was used as the GOCS. The relative counting efficiency of 35S by PSS method is 82% compared to the emulsive cocktail counting method.

146

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Figure 2.

A torn section of the glass filter paper sandwiched between the plastic scintillator (broken granules of wrinkled and cracked plastic scintillator are folded among the glass fibers).

The shape of the PSS can be rectangular or circular. In the liquid scintillation counting procedure the rectangular PSS was leaned obliquely against the inner wall of the counting vial so that the radioactive sample was located in an elevated position relative to the counter phototubes. As a result, higher count-

ing efficiencies could be obtained with the rectangular PSS than with the circular PSS, which was laid flat on the bottom of the vial during counting (Table 1).

The counting reproducibilities (coefficient of variation) of three measure-

PLASTIC SCINTILLATION DETECTOR

147

Table 1. Comparison of the Spectral Parameters of the Plastic Scintillator Method with Homogeneous Counting H-3 Hexadecane

Compton Electron Spetron

Toluene Cocktail 10 mL

+ GOCS

Main Peak Channel

790

End

C-14 Hexadecane

Plastic Scm.

Toluene Cocktail 10 mL

Plastic Scm. + GOCS

Plastic Scm.

640

550

780

640

550

850

760

750

840

760

750

270

240

180

510

510

510

395 363 343

347 312 295

339 289 262

650 623 604

581

572 543 524

Plastic Scm

Channel

Sample Spectrum

Main Peak Channel

End Channel 1% 5%

l0%

554 536

ments for the rectangular and circular P55 are 1.9 ± 1.3% and 2.3 ± 1.2%, respectively. The ratio of cpm of circular to rectangular PSS is 0.862 ± 0.056, which shows that the variation of height of the rectangular PSS position may cause larger counting error (coefficient of variation 6.7%). Because of less variation in the position of the circular PSS at the bottom of the vial, a smaller counting error (SD/mean) of 1.3 ± 1.2% was measured (Table 1). If GOCS is used in excess, the count of the rectangular PSS decreases progressively as its shape changes (Figure 3). The angle in the vial of the rectangular PSS to the PM tubes has little or no effect on counting (Figure 4). From these data, we can see that the rectangular PSS has higher counting efficiency than the circular PSS, but the circular PSS has better reproducibility than the rectangular PSS. ENERGY SPECTRUM ANALYSIS

Table 2 shows the spectrum analysis from the external standard Compton electron spectrum of an empty 20 mL vial, a circular and rectangular PSS in a vial, and a 10 mL toluene cocktail in a vial. The data reveal that the main peak and end channel for the rectangular PSS, with and without GOCS, are significantly lower than the toluene cocktail sample. They also show the main peak

channel for the plastic scintillator with GOCS is significantly higher than without GOCS, though the end channel is approximately the same. Sample spectrum analyses of the main peak channels for '4C homogeneous

counting and rectangular PSS, with and without GOCS, had no obvious change. The 3H main peak channels did decrease in the order mentioned previously. The end channels of the sample spectrum, which were eliminated at

148

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Time After Adding GOCS (hours) Figure 3.

GOCS mount effect on counting efficiency.

Plastic scintillation

Suppoter method

Counting Efficiency I

+

I

90% 100%110%90%100% 110%

Rectangular PSS 89846 + 1172 cts (1.3%) Circular PSS 7715 + 87 cts (1.2%)

Figure 4.

AngIe effect on rectangular PSS with PM tube vs counting efficiency.

PLASTIC SCI NTILLATION DETECTOR

Table 2. Comparison of Background Spectral Parameters 20 mL Empty Two Circle

End Channel Eliminated High energy

BackGround (CPM)

Glass

PSS9 in a

Vial"

Vial" 720

149

Two Rectangle PSS° in a Vial"

10 mL Toluene Cocktail in a Vialb 1000 983

5%

650 588 488

10%

451

515 472

890 803 675 618

Integral

28.5 ± 0.6

25.7 ± 2.1

28.8 ± 0.6

58.8 ± 2.8

28.5 ± 0.6

25.6 ± 0.5

28.7 ± 0.6

43.3 ± 0.7

0 1%

611

963 905

C-14

Channel

(50-670) apSS: Plastic Scintillator Sheet. bviaI Without Cap.

the 1%, 5 Wo, and 10% high energy counts of the plastic scintillator with GOCS, increased slightly in comparison to that without GOCS. The end channels of homogeneous counting also increased compared to plastic scintillator with and without GOCS. The results for 3H are similar to that for '4C. BACKGROUND SPECTRUM

The background spectrum was measured with a 20 mL glass standard vial (no cap), two rectangular PSS in a vial, two circular PSS in a vial, and 10 mL of a toluene-6 g/L of b-PBD cocktail in a vial. The main parameters of the background spectrum are listed in Table 3. The counting time was 90 minutes. When high energy counts were eliminated at the 10% level, the end channel for the 20 mL empty glass vial decreased from 650 to 451. The rectangular and circular PSS in the glass vial decreased from 720 to 472 and from 890 to 618, respectively. The 10 mL cocktail in the vial decreased from 1000 to 905. These data indicate that the range of the background spectrum decreased about one Table 3. Comparison of the Relative Counting Efficiency and Reproducibility of the Rectangular PSS with the Circular PSS Rectangle PSS

No. of Sample 1

2 3 4 5

Mean + SD

Mean ± SD SD/Mean Mean ± SD SD/Mean (cpm)a

10495 ± 205 11209 ± 26 15736 ± 186 17114 ± 395 18033 ± 651

1.9 ± 1.3

Effect of Circle Position

Circle PSS

(%)

(cpm)

(%)

Error (%)b

(%)

2.0

8335 ± 101 10274 ± 134 13060 ± 557 15066 ± 356 14187 ± 365

1.2 1.3 4.3 2.4 2.6

79.4 91.7 83.0 88.0 78.7

+2.9 +0.03 +0.3

0.02 1.2

2.3 3.6

2.3 ± 1.2

= 3. bReiative Counting Efficiency of a Circle to Rectangle PSS.

-1.8 -1.5

0.842 ± 1.056 1.3 ± 1.2

150

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

third for the two kinds of PSS in their high energy background content. This

decrease is greater than the homogeneous counting. This effect is clearly shown in Figure 5 and Table 3. STABILITY

The radioactivities and backgrounds of PSS have been counted repeatly over a period exceeding one year. The data show that the counts have not significantly changed. It is from these results that PSS are said to be stable and

Empty Standard Vial (20m1)

15.0

10.0

5.0

0.0 100

200

300

400

500

600

700

800

900

Empty Vial + Two Plastic Scintillator Sheets

5.0

0 0,0 100

200

300

400

500

600

700

800

900

800

900

Toluene Cocktail (b-PBD: 6g/L) lOmi 15.0

10.0

5.0

0.0

I

I

100

200

300

400

500

600

700

1000

Relative Pulse Height Figure 5.

Background spectrum comparison of an empty 20 mL standard glass vial, two rectangular PSS in a vial, and 10 mL of toluene cocktail in a vial.

PLASTIC SCINTILLATION DETECTOR

151

capable of being stored for long periods of time without undergoing any change.

APPLICATION TO MEASURING OF 3H LABELED CULTURED CELLS

Screening of Chinese medicinal herbs for immunopotentiators was made with the aid of mitogen-induced lymphocyte transformation Human peripheral blood lymphocytes were cultured in a flat-bottomed 96-well microliter plate at a concentration of 1 x 106 cells/well. A final 0.2 mL contained PRMI 1640 medium, supplemented with 10% fetal calf serum, and 100 units/mL of penicillin G, consisting of 100 jig/mL of streptomycin sulphates. Mitogen and/or various concentrations of herbal extracts were added, and the cultures were kept under 37°C at a 5% CO2 atmosphere. After 48 hours of incubation 0.5 or 1.0 Ci of 3H-TdR was added to each well and incubation was continued for 6 to 12 hr. Cells were harvested with a Titertek multiharvester and collected on glass filter discs, after which the radioactivities were counted by the meth-

ods of supporter and rectangular PSS. The stimulation index (S.I.) was calculated as follows: cpm of test well

- cpm of control well An S.I. value larger than 1.2 indicates a significant enhancement of lymphocyte transformation induced by the tested herbs. The results shown on Table 4 indicated that herbs No. 162 and No. 520 exhibited significant enhancement of lymphocyte transformation. The S.I. values obtained in the supporter method were quite similar to that in PSS. Also the relative figure of merit of the PSS (290.0 ± 41.5) was higher than that

in the supporter method (151.9 ± 8.1). The ratio of the PSS to supporter method was calculated at 1.91. The macrophage active factor (MAF) was evaluated by means of macrophage tumor cell cytostasis (MTC). The self-prepared MAF was diluted by 1:25, 1:50, 1:100, and 1:200 and incubated with macrophages (Ms) at a 1 x 105/well concentration for 12 hr at 37°C with a 5% CO, humidified air atmosphere. After two vigorous washings with RPMI 1640, p815 mouse macrophage-tumor cells were introduced at a concentration of 1 x 104/well and incubated together with Ms under the same conditions for 30 hr. During the last 6 hr of incubation an alliquot of 3H-TdR, at a 0.5 Ci concentration was added to each well. The degree of incorporation of 3H-TdR by p815 cells was measured by both solid support and PSS counting methods.

The rate of cytostasis at different dilutions of MAF activated Ms were determined by the two counting methods listed in Table 5. The rate of cytostasis was calculated as follows:

152

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 4. Plastic Scintillator Method Compared with LS Counting in Screening of Chinese Medicinal Herbs for Immunopotentiators

Supporter Method Group

S.l.

cpm (net) ± SD Cell ref. PHA: 2r

PHA: lr PHA: O.05r No. 162A No. 520a No. 251a No. 162Aa

1831 ±

B (cpm)

E2/B

65 66 68 63 66 60 69

153.8 151.5

346

63213 ± 7465

34.5 27.3

50032 ± 19042 341 1954 ± 4662 ± 222 6821 ± 1269 52 1671 ± 3527 ± 852

2.54 3.49 0.85 1.80

147.1

158.7 151.5 166.7 144.9 140.8

71

151.9 ± 8.1

Mean ± S.D.

Plastic Scintillation Counting Group Cell ref. PHA: 2r

PHA: ir PHA:0.05r No. 162A No. 520a No. 251a No. 162Aa

cpm(net) ± SD

E (%)b

79 1360 ± 48400 ± 2289 37285 ± 14607 192 1406 ± 308 3300 ± 535 5094 ± 83 1185 ± 937 3059 ±

74.3 76.6 74.5 71.9 70.8 74.7 70.9 86.7

SI. 35.5 27.4 2.42 3.62 0.87 2.17

B(cpm)

E2IB

19.5 19.5 19.5 19.5 19.5 19.5 19.5 19.5

282.8 300.9 284.8 265.3 257.0 286.0 257.9 385.6

290 ± 41.4

75.1 ± 5.1

Mean ± S.D.

°Plus 0.05r PHA. bRelative Counting Efficiency of Plastic Scintillation to Supporter Method.

MTC rate = (1

(p815 + M) cpm p815 cpm

X 100%

The result shows that even under a relatively high diluting situation (1:200), the MTC rate of MAF activated Ms was still as high as 72.1%, indicating the high activity of the self-prepared MAF. Also, the ratio of the relative figure of merit of the rectangular PSS to that of the supporter method was found to be 1.94.

Monitoring of self-immunity ability in humans was carried out through the assay for interleukin 2 (IL-2) activity. The peripheral blood lymphocytes of a renal transplant patient were stimulated by phytohemagglutinin (PHA) for 24 Table 5. Comparison of the Two Counting Methods in Evaluation of MAF by Means of Anti-Cancer Cell Proliferation MAF

Dilution

1:25

1:50

1:100

1:200

0

Supporter Method Plastic Scint.

1196

3037

10131

19601

70238

B ± S.D. 59 ± 17

169.9

977

2392

8142

15322

57072

19.5 ± 7.8

329.0

Relative Eff. (%)

81.7

78.8

80.4

78.2

81.3

80.1 ± 1.4

E2/B

PLASTIC SCINTILLATION DETECTOR

153

Table 6. Comparison of the Two Counting Methods in Monitoring of Self-immunity Ability

in Renal Transplanting Patient Titer % 50 25 12.5 Supporter 219360 202110 18927 Method Plastic Scint. Relative

6.25

3.125

0

2132 3345

730 855

225700

192170

17542

5609 735

161010 158500

150460 134370

14450 12679

4237 520

1661

521

2436

667

81.7

72.2

74.3

73.1

76.9

74.5

B(cpm) 65 + 19

153.8

19 + 1.4

286.7

E218

73.8 + 3.0

Eff. (%)

hr. The supernant, which was diluted by 1:2, 1:4, 1:8, 1:16, and 1:32, etc., was incubated with C57B1 mouse thymus cells which had been stimulated for 12 hr by Con A. Sixteen hours prior to harvesting 3H-TdR was added. The radioactivities in the cells collected on glass filter paper were compared by PSS count-

ing and the supporter method. Under proper dilution of the supernant, the thymus cells grew well where the activity of IL-2 was high, which demonstrates the renal transplant patientability for cell immunity. By this method, it is possible to predict the body rejection capability of a renal transplant patient. This procedure is at best a quasiquantitative assay method, since it may be influenced by many factors. Its primary usefulness is its ability to compare the relative activity of different measurements for one person. As shown in Table 6, the two counting methods obtained comparable results. The ratio of E2/B of PSS (286.7) to the supporter counting method (153.8) is 1.86. The PSS method applied in the above three experiments gives the same biological conclusions as the supporter method of counting. The PSS method has the advantage of a higher figure of merit, long storage, and no liquid

waste. It is because of these attributes, we strongly recommend the PSS method, especially where small sample size or volume is to be prepared and counted and liquid scintillation waste disposal represents a problem. Note: This manuscript has undergone extensive revision and condensation in order to meet the style requirements and space limitations of the proceedings. We sincerely hope that we have not changed the substance or connotation of any information intended to be conveyed by the authors. Eds. REFERENCES Kalbhen, D.A. and V.J. Tarkkanen. "Review of the Evolution of Safety, Ecological

and Economical Aspects of Liquid Scintillation Counting Materials and Techniques," in Advance in Scintillation Counting, S.A. McQuarrie, C. Ediss, and L.I. Wiele, Eds., (Edmonton, Alberta: University of Alberta Press, 1984), p.66. Yang Shou-li. "The Comment on Literature (1976-1986) for Liquid Scintillation Counting and Prospecting its Future," NucI. Electronics & Detection Tech., 7(4):222(1978).

154

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Schorr, M.G. and F.L. Torney. "Solid Non-crystalline Scintillation Phophors," Phys. Rev., 80:474(1950). Schram, E. Organic Scintillation Detector, (Elsevier Pub. Co. Amsterdam,1963), p.66. Basile, L.J. "Progress in Plastic Scintillators," U.S. Government Printing Office Washington, 1961.

Carlsson, S. "The Use of a Plastic Scintillator to Determine the Activity of Solid

Samples in a Liquid Scintillation Counter," mt. J. App!. Radiat. Isotopes, 28(7):670-671(1979).

Simonnet, F., J. Combe, and G. Simonnet. "Detection of P-32 Scintillation Plastic Vials," App!. Radiat. Isotopes, 38(4):3 11-312(1987). Yue Jia-chang and Xio Jing-cheng, "A Method in Using Plastic Scintillator Film

instead of the Liquid Scintillator Solution for the Low Energy Beta Nuclides," Annual Report of Chinese Acad. of Agr. Sci., (1983), p.141.

Yang Shou-li, H. B. Ma, J. H. Sun and Z. H. Liu. "The Plastic Scintillation Detector with Internal Sample of Soft Beta Nuclides," NucI. Electronics & Detection Tech., 7(6),32l-327(l987).

CHAPTER 14

A New, Rapid Analysis Technique for Quantitation of Radioactive Samples Isolated on a Solid Support Michael J. Kessler, Ph.D.

INTRODUCTION

The quantitation of radioactivity on solid supports has grown extremely rapidly in the past five years due to the increased use of microplate assays. These assays use a special microplate capable of holding 96 samples of volumes up to 300 JLL in the 8 x 12 format. A typical microplate is 3 in x 5 in. The incubation of cells, DNA, tissues, or other substrate takes place directly in the

microplate. Once the incubation has been completed in the presence of a radioactive labeled substrate, the unreacted/unincorporated components must be separated from the incorporated/bound substrate. This is normally done with a cell harvester, which deposits the cell particulate bound material on a filter media, usually glass fiber filter. Alternatively, the DNA can be spotted with a dot blot apparatus in the 8 x 12 format and hybridization can be performed. A wash solution (buffer or water) removes any unincorporated/ unbound radioactive material on the filter. Each of the filters is punched out of the filter mat into individual scintillation vials, 1 to 12 samples at a time. Five mL of scintillation cocktail is added to each of the individual scintillation vials containing the filters. The 96 vials are capped, shaken, and placed in the counting cassettes of a liquid scintillation counter. Each of the 96 samples is

counted for 5 to 10 minutes. The data are printed out and transferred to a computer for final data reduction and graphic presentation. As can be clearly seen from this description of the steps involved in harvesting and quantitation, this procedure is not only time consuming but expensive. Packard has developed an alternative to this procedure that reduces both the time and cost of harvesting and quantifying each sample by a factor of over ten. This procedure uses two new instruments. The first is a special 96 sample cell harvester (Micromate 196). It can harvest all 96 samples from the microplate simultaneously and is as efficient as the manual harvesters. The second is a 155

156

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

special 96 sample radioactivity reader. This Packard Matrix 96 can analyze 96 samples simultaneous using 96 individual detectors using no vial, no cocktail, no bags, and no special filters and causing no destruction to the samples. A description of this harvester and reader will be presented in this article along with applications for which this equipment can be used. HARVESTING/BLOTTING

Three different methods exist for preparing samples on a solid support. The

first is the direct method. This involves spotting the samples in the 8 x 12 format. The best example of this application is dot blots. For this application the DNA is spotted on a membrane in the microplate format, and 32P or DNA probes are used. A typical DNA hybridization device is shown in Figure 1. The radioisotope, which becomes hybridized to the DNA, is then quanti-

tated using radioautography/densitometry or liquid scintillation counting. This DNA probe hybridization assay has become increasingly popular because of the strong interest in DNA probes and DNA sequencing. The second method is the use of a cell harvester. This method involves using a special device which is able to aspirate the sample (cell/particulate material) from the wells of the microplate onto a filter media, and wash the wells and the filter to remove any unincorporated/unbound radioactive material. These filter bound samples are punched into liquid scintillation vials, processed, and

quantitated in a liquid scintillation counter. With the present manual harvesters, only 6 samples can be harvested, washed or punched out at a time. The new Packard Micromate 196 cell harvester is capable of harvesting and wash-

ing all 96 samples simultaneously from a microplate. This is done directly, without any excess tubing in the cell harvester that could become contaminated. This harvester is shown in Figure 2. In order to assess the Micromate 196 cell harvester performance compared to the manual harvester, two separate microplates were prepared using the 3H-Thymidine cell proliferation assay. These separate plates were analyzed by the manual (6 samples) and the Micromate 196 sample cell harvester. The correlation of the data found an R2 of 0.949 (Matrix Application Note). From this data it is clear that the new Micromate 196 cell harvester performs as well as the manual harvester but 16 times faster. These samples from the Micromate harvester can either be manually punched out and analyzed by liquid scintillation counter or they can be quantitated directly by the Matrix 96, direct beta counter. The third method is the use of special filter bottom plates. These filter bottom plates offer the advantage of being able to incubate, wash, and harvest all in a single microplate. Two different types of microplates are available, those with and those without removable filters. Once the cell harvesting and washing is complete each of the 96 filters can either be punched out into liquid scintillation counting or gamma counting vials. If the membrane can be removed from the microplate, the complete membrane (96 samples) can be

NEW TECHNIQUE FOR QUANTITATION OF SAMPLES

I.

=

1M

fM

I

e Il

.

sample template with attached sealing screws

.1 S

membrane

sealing gasket

/ - .ei C -OS lOt 1,-el '0,0.

I.-'h.I.l CI (Cl let /(J(JtI £1(0IIeIt-Ae!-. /071071011It-I I.le2 IeIA.

j&1eILeiiej IeIIelIe,tote,e

J',LOPZ.P (Cl tOilet lOt I.IteIt

feI JO! IOU IOU CI (CI t011-i i I&V go. IC! CI (.1Iea 101102105101-

-

vacuum manifold

_1J .

I

gasket suppo plate

ii Figure 1. Typical dot blot apparatus.

Ii1

3,,

iii,.

tubing and flow valve

157

Figure 2.

Packard MicromateTM 196, 96 sample microplate harvester.

NEW TECHNIQUE FOR QUANTITATION OF SAMPLES

159

analyzed on the new Packard Matrix 96 radioactivity reader. This stripping of

the membrane from Microplate is illustrated in Figure 3. At present these strippable membrane bottom plates are manufactured by Pall Biosupports. In summary, three separate methods exist for preparing samples on a solid support media in the microplate format. The first is the direct method of spotting as exemplified by the dot blot application. The second is the harvester

applications which aspirate the cell/particulate from the microplate onto a filter media and remove the unincorporated/unbound radioactivity by extensive washing steps. The third is the use of membrane bottom microplates which allow the incubation, harvesting and washing of the assay components in a single microplate. All three methods can be used to prepare samples for quantitation using the Matrix 96 radioactivity reader. (Figure 4).

Figure 3.

Strippable bottom microplates pall biosupports.

160

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Figure 4. Packard MatrixTM 96, 96 sample direct beta counter.

QUANTITATIVE TECHNIQUES

Four separate techniques for quantitating samples on a solid support exist in the microplate format. These include standard liquid scintillation counting, multi detector liquid scintillation counting, position sensitive proportional counter scanning, and multi-detector (96) avalanche gas ionization detector quantitation. Each technique will be evaluated in detail with the number of steps, time, and cost of each presented. The first technique is liquid scintillation counting. This technique is the gold standard by which all others are measured. This technique involves over 500 steps from the sample harvesting to the actual quantitation. The first step is harvesting the samples from the microplate 6 samples at a time. This procedure is repeated 16 times in order to harvest all of the samples from the 8 x 12 microplate. Each of the individual samples, which are harvested and washed onto a filter, is punched out into one of the 96 individual scintillation vials which were previously labeled to prevent

a sample mix up. Scintillation fluid (4-7 mL) is added to each of the 96 scintillation vials one sample at a time. Each of the 96 samples is manually capped, transferred to the samples cassettes for the counter, placed in the liquid scintillation counter, and analyzed for 1 mm each. The data is removed from the counter and processed using an external computer system. The over 500 sample handling steps from harvesting to sample analysis, takes over 3 hr to complete. The second technique is harvesting the sample on a specially prepared filter

NEW TECHNIQUE FOR QUANTITATION OF SAMPLES

161

mat using a multidetector (6) liquid scintillation counter. This technique harvests 12 samples at a time using a manual harvester and a special filter mat in

the 6 x 16 format. Thus the harvesting time is similar to that of the liquid scintillation counter technique. Once the samples have been harvested, the entire filter mat is dried and placed into a special chemically resistant plastic bag. To this bag is added 10 to 15 mL of a hydrophobic scintillation cocktail. The bag is sealed to prevent the cocktail from escaping and rolled to insure the cocktail completely saturates the filter mat. This sealed bag is placed into a special holder, and 6 samples are analyzed at a time on the multidetector (6) liquid scintillation counter. This reduces the counting time by one sixth that of conventional liquid scintillation counting. Once the entire plate has been quantitated the data must be converted from the 6 x 16 format to the 8 x 12. This technique reduces the total time of harvesting and analysis to one third that of liquid scintillation counting. It takes over 60 minutes for a microplate of 96 samples and involves 18-20 steps. The third technique is position sensitive proportional counter scanning. The scanners were originally designed to detect radioactivity on a flat surface for such thing as TLC or paper c.hromatography. The technique uses a position sensitive wire detector to count 12 samples (in a single row) simultaneously and scans lane by lane over the 8 x 12 matrix (8 separate scans). This method has three major disadvantages. First, the detectors are not highly sensitive to low energy beta emitter (i.e., 3H). Second, the detectors are subject to high amounts of cross talk when high energy isotopes (i.e., 32P) are analyzed. Third, the detectors are not uniform across the entire length of the wire. This is critical because the wire detectors

must be able to locate and quantitate the radioactivity on the solid filter

matrix. These factors make the technique unsuitable for quantitative applications. The fourth technique is quantitating with a specially designed 96 detector

quantitation system. It uses avalanche gas amplification detectors with collector/cathode voltage bias operating in the Geiger-Muller voltage region. This system is capable of analyzing 96 samples simultaneously with 96 individual detectors in the 8 x 12 format. The technique uses an open ended avalanche gas ionization detector and is capable of quantitating 3H, 32P, 35S, '4C, 125J and many other isotopes which produce ionizing radiation. The only steps involved in using this technique is harvesting 96 samples simultaneously with the Micromate 196, drying the sample filter, and quantitating all 96 samples simultane-

ously in the Matrix 96. This entire process from harvesting to quantitation, requires less than 12 min/microplate of 96 samples, and there is no liquid radioactive waste to dispose of at the end of the experiment, only a filter membrane. This technique does not destroy the sample either, so the sample can be removed from the filter mat and analyzed further (NMR, Mass spectrometry, DNA sequencing, etc.). The filter mat can then be analyzed by the Matrix 96 and placed in a plastic bag for storage or liquid scintillation counting for quantitation at a later date.

162

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 1. Cost of Analysis of 500 Microplate/Vear for Mixed Lymphocyte Cultures Assays by Various Methods MDLSCb MDAGDc LSC 1.

2. 3. 4. 5. 6.

Scintillation Vials Scintillation Cocktail Glass Fiber Filters Special Sample Filter Bags Cost/Technician Time (sample preparation, harvesting, counting) Waste Disposal Costs TOTAL

$2500 $2500 $350 $0 $25,000

0

$500

$100

$10

$30,850

$10,250

$2060

$250 $1000 $350 $8500

0 0

$350 $0 $1700

LSC = liquid scintillation counting. bMDLSC = multi-detector LSC. CMDAGD = multi-detector avalanche gas detector.

Now that the steps involved in each technique have been shown, what about the cost involved in analyzing a series of microplates? If 10 microplates/week were analyzed over the period of a year then approximately 500 plates would be quantitated in one year. The cost for each of these three methods is based on the time required to prepare and analyze the samples and the cost of chemicals

and supplies required. The first technique of liquid scintillation counting requires scintillation vials, scintillation cocktail, filter mats, and time to prepare and dispose of the 96 samples obtained from each plate. If 500 plates were analyzed by this method it would require over $30,000/year as shown in Table 1. For the second technique of multidetector liquid scintillation counting, the scintillation vials have been eliminated, but more expensive special filter mats are required. The cost of the filter bags, cocktail, special filter mats, and labor to prepare and dispose of these samples is over $10,000 for the same 500 plates (Table 1). For the third technique which uses the Micromate 196 cell harvester, to harvest all 96 samples simultaneously, and the Matrix 96 reader, to quantitate all 96 samples simultaneously, the only cost is the filter mat (the same type

used on the manual harvester) and the labor costs. The cost of labor and materials is approximately $2000 for the 500 microplates or 15 times less than the standard liquid scintillation counting technique. APPLICATIONS

Several applications exist involving radioactivity quantitation on a solid support in a microplate format. These include dot blots, 3H-thymidine cell proliferation assays, receptor binding assays, radioimmunoassays, DNA polymerase spot assays, broken cell assays (fungi), and many others. Three specific assay types will be evaluated in detail with reference to the three quantitation techniques described earlier.

The first application is dot blots. This technique specifically identifies a sequence of DNA or RNA of interest in a specific disease or DNA/RNA fragment. The technique involves the following basic steps. The DNA or RNA

NEW TECHNIQUE FOR QUANTITATION OF SAMPLES

163

of interest is bound to a special membrane in a dot blot device (Figure 1). A special radiolabeled DNA/RNA probe, complementary to the region of DNA! RNA of interest is prepared. The radiolabeled probe is added to various DNAs to locate the sequence of interest. The noncomplementary DNA is removed from the membrane by washing, and the radioactivity on the membrane is quantitated. This quantitation is normally accomplished by using X-ray film exposure, because the radionuclide is high energy 35S or 32P. The radioactivity on the film is determined by densitometry. The alternative method is the use of

the Matrix 96 direct beta counter for these samples. A comparison of the quantitation of the densitometry and the Matrix 96 is shown in Figure 5. As can be clearly seen, the densitometry has a small dynamic range because the X-

ray film becomes saturated (50 to 100 fold range). On the other hand the Matrix 96 has a dynamic range of over 1O for the 32P dot blots. In addition to

MATRIX 96 VS DENSITOMETRY 32P DOT BLOT 25

20

15

J

10

4

1

DNA CONCENTRATION MATRIX 9 Figure 5.

DENSITO

Comparison of Matrix 96 and densitrometry for blot applications.

164

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

the larger dynamic range the Matrix 96 is able to analyze the dot blots in 1/50

to 1/100th the time of the X-ray film method. A correlation between the Matrix 96 and liquid scintillation counting on the same dot blot samples was determined to be an R2 of 0.949 (Matrix Application Note). The second application is 3H-Thymidine cell proliferation assays. These assays use the tritiated thymidine which becomes specifically incorporated into DNA as a measure of cell growth or proliferation in culture. This proliferation assay is used to test toxic substances, potential cancer drugs, AIDS drugs, and other important naturally occurring and synthetic substances. The conventional method of analysis is cell harvesting with a manual harvester and analysis by liquid scintillation counting. A series of samples with various radioactivity incorporated into the cellular DNA are analyzed using both the Matrix 96 reader and the conventional liquid scintillation counting techniques. The data from this experiment is shown in Figure 6. The correlation of the data was performed and an R2 of 0.999 was calculated (Matrix Application Note). This clearly demonstrates that this technique provides results which are as accurate as the liquid scintillation counting technique in one tenth the time and at one tenth the cost. The third application is the radiolabeled receptor binding assays. This technique involves using either a specific type of cell culture or a tissue homogenate preparation of a specific animal tissue which contains the receptors of interest. The substrate for the receptor is radiolabeled and a competitive binding assay is performed with unlabeled substrate. The number of receptors and the binding constant can be determined using this technique. The application can be analyzed by one of the three quantitative methods described earlier. If the multidetector liquid scintillation counter is compared to the Matrix 96 radioactivity reader, the results of the two techniques can be correlated. The R2 for these two techniques was calculated to be 0.993 (Figure 7). A similar correlation for liquid scintillation counting and the Matrix 96 was also obtained. In addition to performing receptor binding assays with 3H, the radionuclide 125J

can be performed and quantitated on the Matrix 96 (Matrix Application Note).

Several other applications (Matrix Application Notes) can be performed which use quantitation on a solid support in the microplate format. These include radioimmunoassays with either 3H or 1251. Special DNA polymerase

reactions which using spotting in a microplate format can be used. Initial experiments indicate that chromium release studies can be performed with the Matrix 96 detectors.

SUMMARY

This article presents three different methods for preparing samples on a solid support in the microplate format. These include dot blotting or direct spotting, cell harvesting in a manual or Micromate 196 96 sample microplate

$

Liquid Scintillation Counter

I i 0.0 0 0 2.0E4 4.0E4 6.0E4 8.0E4 1 .0E5 1 .2E5 I .4E.5 I .6E5 1 .E5

5000.0-

1.0E4-

1.5E4-

Figure 6. Comparison of MLC samples analyzed on Matrix 96 and liquid scintillation counter.

0

a

2.0E4-

3.042.E4-

4.OE

166

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS 120

100

60 cm

40

20

0

18

3

64

I

16

I

4

0.25

1

REL CONC. BETAP LATE

Figure 7.

MATX1

Receptor binding application comparison of Matrix 96 and betaplate.

harvester, and the use of a membrane bottom strippable microplate which allows incubation, washing, and harvesting in a single microplate. The strippa-

ble bottom allows the membrane to be analyzed directly on the Matrix 96 radioactivity reader.

Four separate methods of quantitating these harvested samples are presented and compared. The liquid scintillation counting technique requires about 3 hours of sample preparation/microplate and cost about $60/plate. The second technique is multidetector liquid scintillation counting which requires about 60 min/microplate and cost about $20/plate. The third technique is scanning with position sensitive proportional counter. This method improves the throughput but suffers from the disadvantages of cross talk, low 3H-efficiency, and wire detector nonuniformity. The fourth technique, and the most efficient, is the use of the Micromate 19696 sample cell harvester and the Matrix 96 96 detector direct beta counter. This technique requires about 12 min/microplate and cost $4/microplate. In summary, this new technique provides a method of rapidly analyzing 96 samples simultaneously from microplates into a solid support, and quantitating the radioactivity directly, without any liquid scintillation counting waste. The Matrix technique uses No Vials,

No Cocktail, No Bag, No Sample Destruction, No Special Filters, and No Waste (liquid) to analyze the 96 samples on the 8 x 12 Microplate.

CHAPTER 15

Dynodic Efficiency of Beta Emitters F. Ortiz, A. Grau, and J.M. Los Arcos

INTRODUCTION

In liquid scintillation counting measurements'-3 of beta emitters, contrasted models give us efficient detection through a study of internal processes in the vial and the response of the photocathode. Many of these models use a figure of merit of the system as a parameter.48 Other models explain the dynodes response in a photomultiplier as a set of electrons reaching the first dynode. In this way we can get the electron distribution at the output of any dynodic stage in the photomultiplier.92 For each electron set that enters the first dynode, however, there is always a nonzero probability of nondetection in the dynodes. This leads to a lower efficiency at the dynode output than at the photocathode. The present paper has studied the efficiency loss through the dynodic stages for 3H and '4C and applied it to the standardization of '4C via the efficiency

tracing method.78 First, a model for the electron multiplication loss is described, then the efficiency loss through dynodes is computed for 3H and '4C, for a single tube and for two tubes in coincidence. Experimental quench curves for '4C have been compared to the theoretical predictions. NON DETECTION PROBABILITY MODEL

It is assummed that the electron production process, at the output of the photocathode, is governed by the Poisson statistics. If N electrons are expected in the average, the probability of obtaining n electrons is:

P(n) =

NnN n!

(I)

The number n of electrons that leaves the photocathode is multiplied through sucessive dynodes, the amplification being typified by means, of a free 167

168

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

parameter, G, the dynodic gain. This parameter is defined as the number of electrons that come out from a dynode when a single electron comes into the dynode.

Statistically, there is a nonzero probability of having no electrons at the output of the Kth dynodic stage when n electrons arrive at the first dynode. This nonzero probability follows this iterative expression:'2 PK(fl,O) = exp(-G n) exp[G n PKI(l,O)]; K > 1

(2)

K=1

(3)

PK(n,O) = exp(-G n); where

n = the number of electrons that leaves the photocathode K = the order of dynodic stages G = the dynodic gain

The nondetection probability leads to a dynodic detection efficiency lower than the photocathode efficiency, because when n electrons leave the photocathode with probability 1, the total sum of the emission probabilities at the Kth dynodic output is 1 - PK(fl,O).

DETECTION EFFICIENCY

Single photomultiplier tube Beta emitters have an emission spectra given by the Fermi distribution:

N,(E) = C F(Z,E) . (E, - E)2 (E + l)(E2 + 2E)"2

(4)

where N(E)

= the number of particle beta with energies between E, E

+ dE E. = the maximum kinetic energy of beta particles in the ith energy band C = the shape factor13 F(Z,E) = the relativistic expression of the Coulomb factor'4 (all in m0c2 units)

When a radionuclide is dissolved in a scintillation liquid, its beta particles transfer almost all their energy to the scintillator, that is, convert to light. This light emission is partially detected by the photocathode of the photomultiplier tube. The number of electrons that leave the photocathode is a function of the beta particle energy.

DYNODIC EFFICIENCY OF BETA EMIUERS

169

The average number of electrons that can be expected when a beta particle is

emitted in the energy band (E, E + dE), is:

E-Q(E)

E

where

(5)

E = the beta particle energy Q(E) = the ionization quench factor M = the figure of merit, or energy necessary to produce an electron at the photocathode

A good aproximation for Q(E) is Equation 7: Q(E) = 1 - 0.9624 e°5457; E > 10 keY

(6)

Q(E) = 0.1253 Ln(E) + 0.4339; 0.1 keY < E < 10 keY

(7)

According to that, mean value r electrons could be obtained with probability: P(r,E)

-

N(E)r eN(E)

r!

(8)

The detection efficiency at the output of the photocathode for a single tube and for a unique beta particle in the ith band is: P(r,E)

Ef1 =

(9)

r

For all the particles in the ith energy band, the probability for having electrons is:

P1(r) = P(r,E) N(E)

(10)

and the detection efficiency for the ith energy band at the output of the photocathode is: =

P(r)

(11)

r

Therefore the efficiency for the whole spectrum at the output of the photocathode for a single photomultiplier tube is:

Ef =

Efi

(12)

170

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Now the detection efficiency for a beta emitter of the Kth dynodic stage output of a single photomultiplier tube can be analyzed. When a beta particle in the

ith energy band is emitted, the r electrons that leave the photocathode with probability P(r,E) will have a nonemission probability at the Kth dynode given by: P1(r,O) = PK(r,O) P(r,E)

(13)

and summing up for all the electrons that leave the photocathode when a beta particle in the ith energy band is emitted, we arrive at: pKfr,O) P(r,E)

PKI(0) =

(14)

r

Therefore, the efficiency at the output of the Kth dynode for this single beta particle in the ith band is: EKI = (E P(r,E)) - PKI(0) r

(15)

For the N(E) particles in the ith band, the nondetection probability at the output of the Kth dynode is: PKI(0) = PKI(0) N(E)

(16)

and the efficiency of detection for this band will be: EK = Ef1 - PK(0)

(17)

Finally the detection efficiency of the Kth dynodic stage output of a single photomultiplier tube for the whole beta emitter spectrum is:

E,

EK =

(18)

Two Photomultiplier Tubes in Coincidence

The previous procedure can be easily extended for a system composed of two photomultiplier tubes in coincidence. Both tubes are assummed to be identical so the detection efficiency at the output of the photocathodes for a beta particle emitted in the ith-energy band for that coincidence system is: = (E P(r,E))2

(19)

r

and for N(E) particles in the energy band we will have: E'f1 = E'f1

N,(E)

(20)

DYNODIC EFFICIENCY OF BETA EMITTERS

171

Therefore the photocathode efficiency for the whole electron spectrum at the output of the system of two tubes in coincidence is:

E'f =

E'

(21)

Now the detection efficiency at the output of the Kth dynode for two tubes in coincidence can be analyzed. For a single beta particle in the ith energy band the detection efficiency is: E'Kl = (EKI)2

(22)

and for all electrons in the ith energy band we will have: E'K, =

N(E)

(23)

Finally, the detection efficiency for a beta spectrum at the output of the Kth dynodic stages for two tubes in coincidence will be: (24)

RESULTS

Non detection probability

The nondetection probability has been studied for the following parameters: The dynodic gain varying, from 1 to 10 with step 1, the number of dynodic stages 12, and the number of electrons injected into the first dynode, from 1 to 19 with step increment 2. Table 1 shows the nondetection probability for the 12 dynodic stages and for the 1, 3, 5, 7, 11, and 19 electrons injected into the first dynode. The dynodic gain is 1 or 2. In Figure la, the nondetection probability is plotted vs the dynodic stage. For 1, 5, 11, or 19 electrons injected and a dynodic gain 1, in Figure lb. the nondetection probability is plotted vs dynodics gain

Detection Efficiency for Beta Emitters Two different photomultiplier tubes with 12 dynodes have been simulated: low gain, G has a value of 7 in the first dynode and a value of 2 in the others.

high gain, G has a value of 30 in the first dynode and a value of 3 in the others.

Table 2 gives the efficiency at the output of the 12th dynode for 3H, for high gain and two photomultiplier tubes in coincidence.

172

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 1. Nondetection Probability in Successive Dynodic Stages and Different Number of Electrons Injected into the First Dynode STAGE 01

02 03 04 05 06 07 08 09 10 11

12

1 ELECT.

3 ELECT.

5 ELECT.

7 ELECT.

11 ELECT.

19 ELECT.

0.37 0.53 0.63 0.69 0.73 0.76 0.79

0.05 0.15 0.24 0.32 0.39 0.45 0.49 0.53 0.57 0.60 0.62 0.64

0.01

0.00

0.04 0.10 0.15

0.01

0.00 0.00 0.00

0.21

0.11

0.00 0.00 0.00 0.00 0.00

0.26

0.15 0.19 0.23 0.26 0.30 0.33 0.36

0.81

0.83 0.84 0.85 0.86

0.31

0.35 0.39 0.42 0.45 0.48

0.04 0.07

0.01

0.02 0.03 0.05 0.07 0.10 0.13 0.15 0.17

0.01 0.01

0.02 0.03 0.04 0.05 0.06

DYNODIC GAIN: 1 01

02 03 04 05 06 07 08 09 10 11

12

0.13 0.17 0.19 0.20 0.20 0.20 0.20 0.20 0.20 0.20 0.20 0.20

0.00 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01

0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Nondetection probability < 1.E-4

Figures 2a, 2b, 2c, and 2d show the plot of the efficiency vs figure of merit at the photocathode, at the 12th dynode with high gain and with low gain. Figures 2a and 2b are for 3H and 2c and 2d are for '4C. Figures 3a, 3b, 3c, and 3d plot the difference between the efficiency at the output of the photocathode and the efficiency at the output of the 12th dynode as a function of the figure of merit for high and low gain. Table 3 shows the efficiency variation from the photocathode to the 12th dynode for different values of the figure of merit at low and high gain, for 3H, for a single photomultiplier tube, and for two photomultiplier tubes working in coincidence. Figure 4 plots the composed and the experimental efficiency vs quenching

for 3H and 'C. Table 4 contains the numerical values shown in the Figure 4. Figure 5 shows the efficiency for low quantum yield at the photocathode for The points are computed and the line corresponds to the experimental data. The experimental values data have been obtained with a set of quenched standards for 3H and '4C, in a L.K.B. 12 19 Rackbeta spectral liquid scintillation spectrometer.

DYNODIC EFFICIENCY OF BETA EMITTERS

173

0.90 .4-,

c

0.80 - 1 electron injected

Dynodic gain

0.70 -

:

1

0.60 0.50 -

0 0.40 0

5 electrons injected

0.30 C)

11 electrons injected

0.20 -

z

9 electrons injected 0.00

0.00

Figure la.

2.00

4.00

8.00

6.00

10.00

Dynodic stage

12.00

14.00

Nondetection probability vs dynodic gain.

0.90 0.80 0.70 0.60 0.50

0.30 0.20

0.10

0.00

0.'

1.0

2.00

3.0

4.0

5.''

Dynodic gain

Figure lb. Nondetection probability vs dynodic stage.

6.00

0

0.2500 0.5000 0.7500 1.0000 1.2500 1.5000 1.7500 2.0000 2.2500 2.5000 2.7500 3.0000 3.2500 3.5000 3.7500 4.0000 4.2500 4.5000 4.7500 5.0000 5.2500 5.5000 5.7500 6.0000 6.2500

16.883 15.535 14.339 13.273 12.320 11.465 10.695 9.9984 9.3673 8.7936 8.2704

18.4 10

20.146

24.411 22.131

27.044

30. 100

63.529 55.439 48.576 42.766 37.845 33.665

73.011

84.093

*

*

* *

* *

* *

* *

*

* *

*

* *

*

*

* *

*

* * *

12.7500 13.0000 13.2500 13.5000 13.7500 14.0000 14.2500 14.5000 14.7500 15.0000 15.2500 15.5000 15.7500 16.0000 16.2500 16.5000 16.7500 17.0000 17.2500 17.5000 17.7500 18.0000 18.2500 18.5000 18.7500 1.2992

1.33 12

1.4720 1.4348 1.3990 1.3645

1. 51 07

1. 5510

1.6364 1.5928

1.7291 1.68 18

2.1238 2.0596 1.9983 1.9397 1.8836 1.8299 1.7784

2.1911

2.3355 2.2616

2.41 31

2.5803 2.4946

*

* * * * * *

* *

* *

*

*

* *

* *

* *

* *

*

* *

*

31 .0000 31 .2500

25.2500 25.5000 25.7500 26.0000 26.2500 26.5000 26.7500 27.0000 27.2500 27.5000 27.7500 28.0000 28.2500 28.5000 28.7500 29.0000 29.2500 29.5000 29.7500 30.0000 30.2500 30.5000 30.7500 0.50343

0.51 898 0.51111

0.75114 0.73746 0.72417 0.71121 0.69862 0.68637 0.67441 0.66278 0.65142 0.64038 0.62958 0.61909 0.60886 0.59885 0.58911 0.57958 0.57031 0.56125 0.55238 0.54375 0.53529 0.52706 42.0000 42.2500 42.5000 42.7500 43.0000 43.2500 43.5000 43.7500

41 .0000 41 .2500 41 .5000 41 .7500

37.7500 38.0000 38.2500 38.5000 38.7500 39.0000 39.2500 39.5000 39.7500 40.0000 40.2500 40.5000 40.7500

Table 2. Total LSC Efficiency Calculation for H-3 at the 12th Dynode, High Gain Two Photomultiplier Tubes in Coincidence Fig. Merit Efficiency (%) Fig. Merit Fig. Merit Efficiency (%) Efficiency (%) Fig. Merit

0.30726 0.30363 0.30010 0.29660 0.29320 0.28985 0.28652 0.28329 0.28007 0.27694 0.27387 0.27081 0.26784 0.26488

0.31 092

0.31464

0.31 846

0.34278 0.33855 0.33436 0.33028 0.32628 0.32232

0.35151 0.34712

Efficiency (%)

z-I

0 C 0

I-

6.5000 6.7500 7.0000 7.2500 7.5000 7.7500 8.0000 8.2500 8.5000 8.7500 9.0000 9.2500 9.5000 9.7500 10.0000 10.2500 10.5000 10.7500 11.0000 11.2500 11.5000 11.7500 12.0000 12.2500 12.5000

3.5983 3.4580 3.3257 3.2009 3.0829 2.9713 2.8657 2.7655 2.6705

37473

6.9515 6.5809 6.2390 5.9228 5.6300 5.3562 5.1056 4.8703 4.6508 4.4458 4.2540 4.0743 3.9057

7.3541

7.7923

*

* *

* *

* *

*

*

*

*

*

*

*

*

*

*

*

*

* * *

*

*

22.0000 22.2500 22.5000 22.7500 23.0000 23.2500 23.5000 23.7500 24.0000 24.2500 24.5000 24.7500 25.0000

21 .7500

21.2500 21.5000

21 .0000

19.0000 19.2500 19.5000 19.7500 20.0000 20.2500 20.5000 20.7500

1.0336 1.0116 0.99030 0.96964 0.94966 0.93026 0.91148 0.89326 0.87555 0.85839 0.84169 0.82551 0.80976 0.79449 0.77965 0.76518

1.1291 1.1040 1.0798 1.0563

1.2683 1.2384 1.2097 1.1818 1.1550

* * * *

*

*

*

* *

*

*

*

* *

* *

* * *

*

*

*

*

*

*

31.5000 31.7500 32.0000 32.2500 32.5000 32.7500 33.0000 33.2500 33.5000 33.7500 34.0000 34.2500 34.5000 34.7500 35.0000 35.2500 35.5000 35.7500 36.0000 36.2500 36.5000 36.7500 37.0000 37.2500 37.5000

0.41163 0.40603 0.40059 0.39525 0.38998 0.38485 0.37978 0.37485 0.37002 0.36524 0.36059 0.35599

0.41731

0.49588 0.48853 0.48132 0.47429 0.46742 0.46066 0.45408 0.44760 0.44130 0.43510 0.42906 0.42314

* * * * * * *

*

* *

*

*

*

* *

* * *

* * *

*

* *

*

44.0000 44.2500 44.5000 44.7500 45.0000 45.2500 45.5000 45.7500 46.0000 46.2500 46.5000 46.7500 47.0000 47.2500 47.5000 47.7500 48.0000 48.2500 48.5000 48.7500 49.0000 49.2500 49.5000 49.7500 50.0000

0.21495 0.21284 0.21077 0.20869 0.20668 0.20466

0.21711

0.26200 0.25917 0.25636 0.25362 0.25089 0.24823 0.24562 0.24302 0.24049 0.23797 0.23552 0.23310 0.23070 0.22836 0.22602 0.22375 0.22152 0.21928

176

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

100a b

90-

c

PHOTOCATHODE

12TH DYNODE, HIGH GAIN 12Th DYNODE. LOW GAIN

8070C.

80-

40-

3020100 0

20

115

I'O

Figure of merit

25

Figure 2a. 3H: Efficiency vs Figure of merit, a single photomultipller tube. 100

a

90-

b c

PHOTOCATHODE

12TH DYNODE. HIGH GAIN 12TH DYNODE. LOW GAIN

80

70-

O 60..

.-0

50

1-4

3020105

I'O

l's

Figure of merit

20

25

Figure 2b. 3H: Efficiency vs Figure of merit, two photomultiplier tubes in coincidence.

DYNODIC EFFICIENCY OF BETA EMITTERS

177

100

a b 90

PHOTOC&THODE

12TH DYNODE. HIGH GAIN 12TH DYNODE. LOW GAIN

80

60

50

40

I

0

I

I

5

15

10

Figure of merit

20

25

Figure 2c. 14C: Efficiency vs Figure of merit, a single photomultiplier tube.

100

a b

90-

PHOTOCATHODE

12TH DYNODE. HIGH GAIN 12TH DYNODE. LOW GAIN

80-

70Q)

...4

60-

C)

40-

3020

0

i

5

I

10

I

15

Figure of merit

I

20

25

Figure 2d. 14C: Efficiency vs Figure of merit, two photomultiplier tubes in coincidence.

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

178 8

*

a.)

0

Low gain

b High gain

.-4

b

10

15

Figure of merit Figure 3a.

20

25

Difference of efficiencies at the photocathode and at the 12th dynode for 3H, a single photomultiplier tube.

8

a.) C.)

10

1'S

Figure of merit Figure 3b.

2

25

Difference of efficiencies at the photocathode and at the 12th dynode for 3H, two photomultiplier tube.

179

DYNODIC EFFICIENCY OF BETA EMITTERS 8

6C)

a)

a)

Low gain

a

a)

b High gain

4-

2-

0

0 Figure 3c.

I

5

I

10

I

15

Figure of merit

I

20

25

Difference of efficiencies at the photocathode and at the 12th dynode for 14C, a single photomultiplier tube.

8

a Lowgain

6-

b High gain

a)

0

a)

;.4 4a)

'4-

'-I

0 5

Figure 3d.

10

15

Figure of merit

20

25

Difference of efficiencies at the photocathode and at the 12th dynode for 14C, two photomultiplier tube.

180

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 3. Efficiency for 3H in Photocathode and Successfive Dynodes Two Photomultiplier Tubes in Coincidence STAGE

FOT.

2TH DY.

4TH DY.

6TH CY.

12TH DV.

84.81

84.21

84.09*

55.74

84.10 55.45 11.47

84.09

3.91

3.91

20.0 25.0

57.25 12.49 4.32 2.16 1.29 0.86

1.16 1.16 0.58

0.25

84.81

100

57.25 12.49 4.32 2.16

83.04 52.89 10.16 3.39 1.66 0.99 0.65

82.13 50.77 9.19

F.M. 0.25 1.00 5.00 10.0 15.0

5.00 10.0 15.0 20.0 25.0

1.29

0.86

11.63 3.97 1.97 1.18 0.65

3.01

1.47 0.87 0.58

5544

5544*

11.46 1.15 1.15 0.56

11.46* 3.91* 1.15* 1.15* 0.56*

81.99 50.44 9.04 2.96 1.45 0.86 0.56

81.96* 50.38* 9.01* 2.95* 1.44* 0.85* 0.56*

HIGH GAIN

LOW GAIN

A Single Photomultiplier Tube STAGE F.M. 0.25 1.00 5.00 10.0 15.0 20.0 25.0

0.25 1.00 5.00

100 15.0

20.0 25.0

FOT.

2TH DY.

4TH DY.

6TH CV.

12TH DY.

89.03 69.44 29.54 16.93 11.85

88.61

88.53

88.53

68.33 28.44

68.11

68.11

28.23 16.07

28.22 16.06 11.21

9.11

8.69 7.05

11.21 8.61

88.53 68.10* 28.22* 16.06* 11.21*

8.60 6.98

8.60* 6.98*

87.06 64.35 24.89 13.92 9.65 7.38 5.98

87.04* 64.30 24.85* 13.89* 9.63*

16.21 11.31

7.39

89.03 69.44 29.54 16.93 11.85 9.11

7.39

87.80 66.20 26.47 14.23 10.38 7.95 6.44

6.98

87.16 64.60 25.09 14.05 9.75 7.46 6.04

HIGH GAIN

LOW GAIN

737* 597*

Table 4. Quenching Curve for '4C Theoretical Values FIGURE OF MERIT

0(E) 451.61 434.61

394.12 351.88 309.94 250.72 206.16

EFFICIENCY FOR 14C

EFFICIENCY 3H

FOT.

H.G.

L.G.

FOT.

H.G.

L.W.

58.33 54.18 47.15 38.05

0.97

0.91 1.04 1.31

0.77 0.89

95.23 94.67 93.54 91.79 88.88 82.95 75.47

95.24 94.69 93.57 91.79 88.82 82.95 75.45

95.25 94.66 93.57 91.82 88.79 82.99 75.49

1.11

27.41

1.40 1.85 2.61

15.86 09.03

4.19 6.27

1.74 2.47 3.94 5.90

1.11

1.47

2.10 3.33 4.99

PHO: Photocathode; HG.: 12th Dynode high gain; L.W.: 12th Dynode low gain

DYNODIC EFFICIENCY OF BETA EMITTERS

181

70

60

50

0 -4

0

30

'420

10

20

0

350

Quenching

450

400

500

Figure 4a. Quenching curve for 3H.

100

90-

Kxpermenta1 value.

60-

50-

40

100

150

200

250

300

350

Quenching

400

Figure 4b. Quenching curve for 14C, experimental and calculated.

450

500

182

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS 100

a

90

b c

photocathode

low quantum efficiency in the photocathode 12th dynode. low gain

80 70 C.)

C.)

..-

60

50 40

30

20 10

20

15

10

Figure of merit Figure 5a.

25

3H efficiency vs Figure of merit, two photomultiplier tubes in coincidence, lo quantum efficiency at the photocathode.

100

90-

60* Experimental va1uo

C.)

70-

. C.)

'.4

"-4

80

50-

40

Figure 5b.

100

I

150

200

I

250

I

300

I

350

Quenching

I

400

I

450

500

Quenching curve for '4C, experimental and calculated, low quantum efficiency at the photocathode.

DYNODIC EFFICIENCY OF BETA EMITTERS

183

DISCUSSION

The results of the previous section show that the nondetection probability falls down very quickly when the number of electrons reaching the first dynode grows. It also falls with the increase of the dynodic gain. However, this nondetection probability is greater for the latest dynodic stages. The model developed allows efficiency computation in any dynodic stage

that is lower than the efficiency at the photocathode, as a function of the figure of merit for a single photomultiplier or for a two photomultiplier system working in coincidence. The computed efficiencies do not vary significantly after the 4th or 5th dynodic stage.

The model permits obtaining the quenching curve of a nuclide from the experimental quenching curve of another nuclide. It does this by using the figure of merit efficiency function for both nuclides, at the photocathode or at the output of any dynodic stage. The computed values are the same regardless of the case considered. REFERENCES

Horrocks, D.L. Applications of Liquid Scintillation Counting, (New York: Academic Press, 1965). Gibson, J.A.B. "Modern Techniques for Measuring the Quenching Correction in a Liquid Scintillation Counter", The ml. Conf. on Liquid Scintillation Counting (San Francisco: Aug. 1979) P. 21. Horrocks, D.L. and M.H. Studier. "Determination of the Absolute Disintigration Rates of Low Energy Beta Emitters in a Liquid Scintillation Counting," Analyt. Chem. 33: 615 (1961).

Gibson, J.A. and H.J. Gale. "Absolute Standardization with Liquid Scintillation Counters," J. Scint. Instrurn. ser. 21 1 199(1968).

Grau, A. and E. Garca-Torano. "Evaluation of Counting Efficiency in Liquid Scintillation Counting," mt. J. Appl. Radial. Isot. 33:249 (1981). Garca-Toraflo, E. and A. Grau. "EFFY, a New Program to Compute the Counting

Efficiency of Beta Particles in Liquid Scintillators," A Comp. Phys Commun 36:307 (1985).

Houtermans H. "Probability of Non Detection in Liquid Scintillation Counting," Nuci. !nstr. and Methods 112:121 (1973). Jordan, P. "On Statistics of Coincidence Detection Efficiency in Liquid Scintillation Spectrometry," Nuci. Instr. and Methods 97 1 07 (1971).

Lombard, F.J. and F. Martin. "Statistics of Electron Multiplication," Rev. Sci. Instr. 32:200 (1971).

Gale, H.J. and J.A. Gibson. "Methods for Calculating the Pulse Height Distribution at the Output of a Scintillation Counter," J. Sci. Instr. 43:225 (1965). Ortiz, J.F. and A. Grau. "Estadistica de Ia Multiplicación de Electrones en un Fotomultiplicador: Métodos Iterativos," JEN report 574 (Madrid, Spain: Junta de EnergIa Nuclear). Prescott, J.R. "A Statistical Model for Photomultiplier Single-Electron Statistics," Nuci. Instr. and Methods 39:173 (1966).

184

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Konopinski, E.J. The Theory of Beta Radioactivity, (Oxford U.K.: Oxford University Press 1966), p. 280. Rose, M.E. Beta and Gamma-Ray Spectrometry (Amsterdam, North-Holland: Ed K. Siegbahm, 1055), p. 280.

CHAPTER 16

Solid Scintillation Counting: A New TechniqueTheory and Applications

Stephen W. Wunderly and Graham J. Threadgill

ABSTRACT Solid scintillation counting isa new technique for measuring radioactive samples that previously had to be counted in liquid scintillators. The solid scintillator, XtalScintTM, produces a light output brighter than standard organic scintillators. It is nonhazardous, nonodorous, and insusceptible to impurity (chemical) quench. It can be attached to either porous or nonporous surfaces for solution and filtration applications. Under optimum conditions, tritium efficiency approaches 60%. This report will discuss the theory and applications of XtalScint to solid scintillation counting techniques.

INTRODUCTION

Recent developments in liquid scintillation counting have focused on improving liquid scintillation solution safety by decreasing the volatility of the

solvent (Kellogg,' Kalbhen and Tarkkanen,2 Reed,3 and Lin and Mei4). A solid, fine, scintillator powder that replaces the solvent would be the ultimate in low volatility. The powder, supported on either a porous or nonporous carrier, could be placed in a vial and counted in a liquid scintillation counter. The scintillation performance of the solid scintillation system would be equiva-

lent to or superior to counting with liquid scintillation solutions. Also, in contrast to liquid scintillation solutions, the counting vials could be reused; the labeled samples could be recovered from the solid scintillators. Solids also have health and disposal advantages that make this new technique potentially revolutionary to the industry. This report will discuss, in depth, the theory and applications of such a solid scintillation counting system using Beckman's Ready Cap and Ready Filter.

185

186

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

THEORY

There are five important properties of a scintillator. First, it must efficiently convert the energy of the radiation decay into measurable light. Second, it must be chemically inert to the conditions of measurement. Third, it must have low noise, or background. Fourth, it must be able to interact with low energy betas and augers Finally, it should present minimal hazards to the user. XtalScint by Beckman meets all these criteria.

Conversion Efficiency

Figure 1 demonstrates that the light output for XtalScint is much greater than even an unquenched liquid scintillation calibration standard. Since these emission spectra were generated by the same isotope, this implies that XtalScint has a much greater light output per KeV of excitation energy than the best liquid scintillators. The wavelength of the emission maxima for XtalScint is 395nm, ideal for maximum sensitivity of current photomultiplier technology. The decay time of emission is 80 to 120 nsec. While this is slower than most traditional liquid scintillators, it is well below the minimum microsecond resolving power of current liquid scintillation counters and therefore very acceptable. Chemical Properties XtalScint is a solid with a melting point above 1000°C. It is chemically inert to aqueous buffers, aqueous bases, and organic solvents. It partially dissolves in concentrated acid, however, it can withstand moderate exposure to 1 M hydrochloric acid and 0.1 Msulfuric acid. Since XtalScint is completely impervious to organic reagents it is immune to impurity (chemical) quenching interferences. See Table 1 for data. It is also immune to chemiluminescence caused by common chemical reagents, such as base and peroxide.

Low Noise (Background) XtalScint as used in Ready Cap and Ready Filter applications has essentially the background of an empty vial. Background generated from cosmic particle interaction with 10 mL of liquid scintillator is greatly reduced with Ready Cap or Ready Filter (see Table 2) because of the small scintillator target. Gillespie5

has reported an 8- to 10-fold improvement in the signal to noise ratio for measurements of 32P and 125J with Ready Cap compared to liquid scintillator cocktail.

Detection Capability While plastic scintillators and crystal scintillators are not new solid scintillators, they have found very little acceptance for measuring biological samples, most of which are labeled with tritium and '4C. Although effective as scintilla-

Figure 1.

0

0.2

p-

I I

I

0.6

(Thousands) Channels

0.4

I

0.8

(Log Energy)

I

XTALSCINT SOLE SCINTILLATOR

I

-s UNQUENCHED COCKTAIL STANDARD

READY SAFE COCKTAIL WITh 20% WATER

Light output (Tritium Spectra) with various scintillator forms.

Greater Brightness

0

10-

30-

40-

50-

60-

70-

80-

90-

100-

110

1

188

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 1. Effect of Chemical Quenching on Tritium Efficiency, at 20 nsec, of inulin in Liquid and Solid Scintillators Scintillator Quench Agent

1O0L

Nitromethane

1OOL

Xtalscint

Liquid 455,0a

Volume

None Isopropanolc

39% (DRY)b 28.5% (WET) 28.5% (WET)

45.0% 2.9%

aTritiated inulin in water added to 5 mL of ready protein. bTritiated inulin in water added to ready cap, dried and counted. clsopropanol, 100L, added to both cocktail and ready cap preparations. Samples were counted without removing isopropanol. dNitromethane, 10OiL, added to both cocktail and ready cap preparations. Samples were counted without removing nitromethane.

tors, their poor contact with the labeled substrate has meant poor detection capability. Two important developments make the XtalScint scintillator system effec-

tive. First, XtalScint is a fine (3 to 8 micron) powder providing enormous surface area for the intimate interaction between the labeled substrate and the scintillator. Second, Seltzer and Berger6'7 and Unak8 have calculated that by replacing the liquid media between scintillator particles with air, one increases the distance an electron can travel by a factor of 1000. Thus, removing water from an aqueous solution of 3H 5-fluorouracil on ready cap resulted increased the tritium count rate from 13,700 to 75,900 cpm. This is a fivefold increase. The performance of XtalScint may also be enhanced by altering the coincidence gate on the LS instrument. By increasing the coincidence gate time one

can improve the tritium count rate by 25 to 40%. Enhancement of more energetic isotopes is not as significant (see Table 3). Table 2. Background for XTALSCINT and Ready Safe Window Sample

0-400

401-670

>14C 671-1000

Empty Plastic Vial 10 mL Ready Safe in Plastic Vial XTALSCINT, 35 mg, in Plastic Vial

14.43

14.17

0.33

26.10

13.90

17.77

14.55

14.06

0.53

3H

Table 3. Radioisotope Efficiencies of XTALSCINT vs Ready Protein + Radioisotope

1251

Substrate

Palmitic Acid Glutamic Acid Triodothyroxane Phosphoric Acid

Solvent

2-Propanol Water 2-Propanol Water

XtalScint 20 nsec

XtalScint

Ready

XTAL

PROTEIN

GATE

GATE

4l.5%

57%

93.4% 75% 94%

97.7Vo

78% 98.5%

44% 94% 78% 100%

SOLID SCINTILLATION COUNTING

189

Health and Safety Aspects XtalScint has certain inherent laboratory safety advantages over liquid scm-

tillators. XtalScint has no vapor, protecting the user from exposure to bad odors or toxic vapor inhalation. XtalScint is not absorbed through the skin, saving the user from possible internal organ damage. XtalScint does not need to be used in a hood, providing additional hood space previously occupied by liquid scintillators. Spilled XtalScint, should it come free from its carrier, can be cleaned up with a broom and dust pan rather than spill pillows and absorbent. XtalScint will not burn, in contrast to most liquid scintillators which are classified as flammable or combustible. XtalScint with carrier is smaller than liquid scintillators, making storage more efficient. XtalScint with carrier may

be stored in any quantity; storage of hazardous liquid scintillators is proscribed by regulations.

Disposal Aspects

Solid scintillators such as XtalScint are much easier to dispose of than are liquid scintillators. Liquid scintillator waste must be packaged with an absorb-

ent adequate to hold the spilled liquid in the event the scintillation vials or other waste containers should leak or break. As a result, a drum of liquid scintillator waste contains about 2/3 absorbent and only 1/3 actual vials. XtalScint, in contrast, is classified as dry waste and requires no supplemental absorbent. Because of its compactness (small size) and dry form, one drum of XtalScint samples on its carrier is equal to 30 to 300 drums of the same number of liquid scintillator samples in vials. This represents tremendous cost savings as well as stress relief for disposal sites and the environment in general. SAMPLE PREPARATION AND COUNTING PROCEDURE

Ready Cap

Ready Cap is the cap of a small plastic vial coated with a layer of XtalScint9 (see Figure 2). It is designed for counting nonvolatile, radiolabeled substrate in a volatile solvent. The liquid sample, less than 200 /LL, is pipetted onto the

scintillator surface in the Ready Cap. The volatile liquid is evaporated by numerous methods (evaporation in hood, heat lamp, hot air blow dryer, microwave oven, or vacuum centrifuge), leaving the labeled, nonvolatile substrate deposited on the surface of the scintillator. The Ready Cap is placed in an LS vial and counted on an LS counter. The Ready Cap can subsequently be removed from the vial and the vial reused. In many cases the radiolabeled

substrate may be extracted and recovered for further analysis; this is not possible with liquid scintillator solutions. The position of the cap in the vial makes very little difference to the measured count rate (see Table 4). The labeled substrate may be any nonvolatile material, such as proteins,

190

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Figure 2.

READY CAP: A small plastic container coated on the bottom with XtalScint.

peptides, amino acids, oligionucletides, steroids, etc., dissolved in a volatile liquid. There are three limitations on the use of Ready Cap. First, the volume of the

sample is limited to 200 giL. Second, because of the small target area of scintillator, it is difficult to generate a spectrum from an external gamma source for quench determination. Finally, noncorrectable quench is introduced

by codeposition of nonradioactive, nonvolatile substances such as salts Table 4. Ready Cap Position in Vial: Effect on Efficiencya Count Rate (CPM) Ready Cap Position Bottom Down Bottom Up

239,000

100%

239,000

100%

0

245,500

102.7%

1

236,300

98.9%

..

Side Up Bottom Parallel to PMT Face Side Up Bottom Perpendicular to PMT Face

Relative to Bottom Down Configuration

aSingle ready cap labeled with 3H-5-Fluorouracil counted on four possible positions resting on bottom of plastic 20 mL Poly Q Vial.

SOLID SCINTILLATION COUNTING

191

Table 5A. Effect of Buffers on Ready Cap 3H Counting Efficiency; Values Reported are Counting Efficiencies (%) at 20 nsec Liquid Carrier for 3H-Palmitic Acid 0.1. Water

0.1 MNaCI 1.0 M NaCI

Volume of 3H-Palmitic Acid Added

25,L

5OL

1O0.L

200,L

41.8 42.7 41.7

39.6 40.3 38.0

35.8

33.0 33.2 23.3

37.1

32.8

(greater than 1 M) or high boiling solvents like glycerol, which absorb the beta decay before it reaches the scintillator (see Table 5a and 5b).

Ready Filter

Ready Filters are glass fiber filters coated on one side with XtalScint and designed to be used in a manner similar to conventional glass fiber filters. Ready Filter comes in two formats: filter mats for automated cell harvesters, and 25 mm filter circles for manual filtration experiments. Particulate suspensions are filtered onto the XtalScint side of the filter. To inactivate nonspecific binding, it may be occasionally necessary to presoak the filter in unlabeled substrate. The flow rate of water through the Ready Filter is about 70 the flow rate through S&S #32 glass fiber filter. The retention of BSA precipitate for the two

filters is greater than 99. Counting efficiency comparisons were made between Ready Filter and S & S #32 glass fiber filter counted in Ready Organic. The samples used contained 250 g of BSA or DNA, labeled with different isotopes. The results of this comparison, reported in Table 6, show that the efficiencies of the two filters are similar to each other for 3H, '4C, 1251, and 32P labeled precipitates when counted on conventional liquid scintillation counters. Using a counter with an optimized coincidence gate yields superior tritium efficiency for the Ready Filter. Other isotopes are not as dramatically affected. Both filter types show similar patterns of efficiency dependence on substrate identity (Table 7) and similar weight of precipitate deposited (Table 8). Efficiency for Ready Filter was not affected by the vacuum caused increased flow rate through the filter (Table 9). After drying Ready Filter, it may be counted in any vial, but the orientation of the filter in the vial makes a difference in counting efficiency (Table 10). Table 5B. Effect of Buffers on Ready Cap Counting Efficiency; Values Reported AE Counting Efficiencies (%) at 20 nsec Liquid Carrier Volume of 14C-L-Amino Acid Mixture Added for Amino Acid Mixture D.I. Water

0.1 MNaCI 1.0 M

50 L

100 jiL

10 4

200 4

83.0 83.2 80.5

78.2 78.8 76.2

78.1

77.4 77.4 67.0

78.2 72.4

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

192

Table 6. Counting Efficiency of Common Isotopes: Ready Filter vs Standard Filter in Cocktail

Efficiency Isotope 3H

l4

1251

32p

Substrate

Glass Filtera

BSA BSA BSA LYSATE

32.6 96.9 57.8 101.6

Ready Filter"

Ready Filterc 36.6 92.0 70.8 95.5

30.1

87.3 62.2 87.35

as & S #32 glass fiber filter counted in Ready Organic. bReady filter counted with 20 nsec gate. cReady filter counted with Xtal gate.

Table 7. Counting Efficiency as a Function of Substrate: Ready Filter vs Standard Filter in Cocktail

Efficiency Substrate

ii BSA 3H Lysozyme 3H DNA

Glass Filtera

Ready FiIter

Ready Filterc

30.1

36.6 30.7 36.3

32.2 27.4 29.2

24.9 30.2

as & 5 #32 glass fiber filter counted in Ready Organic. bReady filter counted with 20 nsec gate. cReady filter counted with Xtal gate.

Table 8. Efficiency vs Weight of Precipitate: Ready Filter vs Standard Filter in Cocktail Efficiency Weight of 3H BSA (g) Ready Filterb Glass Filtera 5 50 100 250 500 750 1000

36.5 41.0 37.2 35.5 32.0 29.6 30.5

35.1 43.1

37.4 35.8 32.0 31.5 31.0

as & S #32 glass fiber filter counted in Ready Organic. bneady filter counted with 20 nsec gate.

Table 9. Efficiency vs Flow Rate: Ready Filter vs. Standard Filter in Cocktail Vacuum Measured in Inches of Hg Drop 6.8 10.0 15.0

20.0 25.0

Glass Filtera 30.6 29.3 30.3 32.9 32.2

a5 & S #32 glass fiber filter counted in Ready Organic. bReady filter counted with 20 nsec gate.

Efficiency of 3H BSA Ready Filterb 33.1

30.0 30.4 32.9 32.2

193

SOLID SCINTILLATION COUNTING

Table 10. Ready Filter Position in Vial Effect on Efficiency0 Count Rate Position of (CPM) Vial XTAL in Vial

Relative to Maxi Up Configuration

Up Down

Maxi Maxi

52,500 21,000

100% 40%

In

Maxi Maxi

41,500 39,000

79% 75%

Mini Mini

44,500 41,500

84% 79%

Blo Bio

38,000 39,500

72% 75%

Out In

Out In

Out

aSinglo ready filter with 3H BSA precipitate counted in different orientations in three different vial types.

The best position for counting Ready Filter is flat on the bottom of the vial with the XtalScint side up. As with Ready Cap, the analyte can be recovered by extraction from the XtalScint and used for further analysis.

Ready Filter should not be used for counting nonvolatile substrate in true solution. The glass fiber support for Ready Filter is capable of strong, nonspe-

cific binding. Upon drying, the nonvolatile substrate can be bound to the nonscintillating support and give reduced scintillations. An example of this is reported in Table 11. Precipitates on the other hand, are captured in the solid scintillator layer and give the high counting efficiencies reported. External

quench monitors are not recommended with Ready Filter due to the time required to obtain the Compton spectra from such a small target. CONCLUSIONS

XtalScint is an effective solid scintillator ideally suited for measuring ioniz-

ing radiation from radioactive decay. It may be measured on conventional liquid scintillation instrumentation, performing similar to conventional liquid Table 11. Risks of Counting Labeled Substrate in Solution on Porous XtalScint Base (Ready Filter) Efficiency0

Solution 3H Uracil 3H Histamine

Ready Filter

Ready Cap

38.8% 6.1%

39.1% 39.5%

Scrap XtalScint from Ready Filter surface Extract Filter Base and XtalSclnt with Ready Protein + % Total DPM Extracted Uracil

XtalScint from Ready Filter Filter Base from Ready Filter 0Eniciency measured at 20 nsec gate.

52.4 47.6

Histamine 7.4

92.6

194

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

scintillators, or measured on instruments with optimized coincidence circuitry (such as the Beckman LS 6000 series), performing better than conventional

liquid scintillators. Its low background and high signal make it especially attractive for those experiments requiring high sensitivity. REFERENCES Kellogg, T.F. "Progress in the Development of Water-Miscible Non-Hazardous Liq-

uid Scintillation Solvents," in Advances in Liquid Scintillation Counting, S.A. McQuarrie, C. Ediss, and L.I. Wiebe, Eds. (Edmonton, Alberta: University of Alberta, 1983) pp. 387-393. Kalbhen, .D.A. and V.J. Tarkanen. "Review on the Evolution of Safety, Ecology and Economical Aspects in Liquid Scintillation Counting Materials and Techniques," in Advances in Liquid Scintillation Counting: S.A. McQuarrie, C. Ediss, and L.I. Wiebe, Eds. (Edmonton, Alberta: University of Alberta, 1983) pp. 66-70. Reed, D.W. "Triton x-lOO as a Complete Liquid Scintillation Cocktail for Counting

Aqueous Solutions and Ionic Nutrient Salts," mt. J. App!. Radiat. Isot. 35(5): 367-370 (1984).

Lin, C.Y. and T.Y.C. Mei, "The BAM Scintillators for the Measurement of Radionucides," mt. J. Appl. Radiat. Isot. 35(1): 25-38 (1984).

Solomon, R., J. Thompson, and D. Gillespie, "Low Background Counting with Ready Cap," Technical Information Bulletin, T-l688-NUC-89-26 (Beckman Instruments: Fullerton, CA, 1989).

Seltzer, S.M. and M.J. Berger. "Evaluation of the Collision Stopping Power of Elements and Compounds for Electrons and Positrons," mt. J. AppI. Radiat. Isot. 33: 1189-1218 (1982).

Seltzer, S.M. and M.J. Berger, "Improved Procedure for Calculating the Collision Stopping Power of Elements and Compounds for Electrons and Positrons," mt. J. App!. Radiat. Isot. 35(7): 665-676 (1984). Unak, T. "A Practical Method for the Calculation of the Linear Energy Transfers

and Ranges of Low Energy Electrons in Different Chemical Systems," Nuci. Instrum. Methods Phys. Res. A255, 274-280 (1987). Wunderly, S.W. "Solid Scintillation Counting: A New Technique for Measuring Radiolabeled Compounds," mt. J. App!. Radiat. Isot. 40(7): 569-573 (1989).

CHAPTER 17

Photon Scattering Effects in Heterogeneous Scintillator Systems Harley H. Ross

SUMMARY

This paper describes a new experimental approach that reveals the individual contributions of sample geometry and scattering phenomena in heterogeneous flow-cell detectors. The experimental detector responses obtained using scintillating polystyrene beads with optically smooth surfaces are compared with those obtained using similar beads with highly diffuse surfaces. These comparisons are carried out for both alpha- and beta-emitting nuclides. The experimental detection efficiencies are compared to Monte Carlo simulations

of the detection process. Also, a new technique will be described for the fabrication of scintillating beads. INTRODUCTION

Flow cell scintillation detectors are used extensively for monitoring alpha-

and beta-emitting nuclides in flowing aqueous or organic streams. Major applications of such cells include liquid chromatography detectors, in-line process monitors, and a variety of environmental measurement and control devices. For low-energy emitters, two quite different approaches are used: the heter-

ogeneous and the homogeneous flow cell. Both, however, use a scintillation process. Homogeneous flow cell systems operate by mixing a suitable liquid scintillator with all or a portion of the flowing stream. Several mixing and flow control units are used to establish stable, fixed measurement conditions. The *Research sponsored by Office of Energy Research, U.S. Department of Energy under contract DE-ACO5-840R21400 with Martin Marietta Energy Systems. "The submitted manuscript has been authored by a contractor of the U.S. Government under contract No. DE-ACO5-840R21400. Accordingly, the U.S. Government retains a nonexclusive, royalty-free license to publish or reproduce the published form of this contribution, or allow others to do so, for U.S. Government purposes." 195

196

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

cell is optically coupled to one or two photomultiplier tubes which are used to detect the scintillation output. Although this type of device can exhibit high

sensitivity, it suffers from a number of problems: it uses large amounts of expensive liquid scintillator, it must provide safe disposal of its scintillator effluent, its quenching effects are unstable and unpredictable, it is difficult to maintain (mechanical and reagents), and its resolution is often poor. For these reasons the heterogeneous detector cell is usually preferred for most on-line measurement tasks. Heterogeneous detector cells4 are characterized by contacting the flowing liquid (to be measured) with a solid phase scintillator, most often in powder form. Such a flow cell must fulfill certain physical requirements for efficient

operation. Obviously, detection geometry must be high in relation to the energy of the nuclide(s) to be measured. Cell volume must be as large as possible (for maximum sensitivity), consistent with the required temporal resolution of the application. The cell must be designed for minimal mixing and

virtually no dead space to prevent any hold up. There may also be requirements for the shape of the cell, the surface to volume ratio, and the optical demands of coupling photodetectors. The most efficient heterogeneous units have a tube or column that contains a solid scintillator (organic, plastic, or glass); the sample flows within the interstices of the scintillator particles or fibers. While the geometry of this device is not as good as that of the homogeneous detector, commercially available cells that use finely divided organic or glass scintillators show that it can count '4C at moderate efficiency. The tritium efficiency is, however, quite low. Other powdered materials used in cells of this type include europiumactivated calcium fluoride and scintillating plastics. Powdered scintillator detector cells also exhibit deficiencies characteristic of

their design. For example, radiation generated by weak emitters degrades significantly in its passage from the sample liquid phase to the solid fluor. If the sample absorption path length is sufficiently large, and the radiation energy sufficiently low, total absorption takes place, and the event is lost to the system. Attempts to use even finer particles to improve the geometry have not improved detection efficiency as expected and routinely have made other cell

characteristics worse. For example, back pressure of fine powder cells can exceed the working limits of conventional flow systems. Also, such cells with their high surface to volume ratio exhibit strong and sometimes irreversible memory effects. Rucker et al.56 have attempted to overcome these difficulties by designing a

cell that uses aligned scintillator fibers rather than the usual powder. Back pressure and memory effects improved significantly, but detection efficiency increased only slightly over previous designs. However, the results of this work prompted the idea that photon scattering within the detector, in addition to the geometry parameter, was crucial in establishing the ultimate efficiency of these devices. While smaller fluor particles lead to improved geometry, they also

generate increased light scatter. The reason for this is that most of the fine-

EFFECTS OF PHOTON SCATTERING

197

particle scintillators have very poor surface characteristics and simultaneously exhibit extensive scattering within the particles themselves. This scatter leads to elevated absorption of the generated photons with subsequent loss of pulse height, detection efficiency, and spectral resolution. Although it appears clear that both geometry and photon scatter can diminish detector performance, it is not obvious which parameter is most important in a specific application, nor

has it been demonstrated how the effects of these parameters change with different cell designs. What is clear, however, is that attempts to optimize one of the parameters often result in a degradation of the other; small particle size scintillators usually exhibit maximum scattering. Practical flow-cell implementations are often a compromise between these two, conflicting optimization constraints. On the basis of the above cited work and other work, it is clear that scintillation efficiency, geometry, and light collection are the major parameters that determine flow cell radiation response to low-energy emitters. It also appears that using conventional powdered materials for the solid scintillator cell phase cannot improve detector geometry and reduce photon scatter simultaneously.

This apparent stalemate is unfortunate since improved flow cell detectors would have a broad range of important applications. Thus, while numerous different flow radiation detectors are used daily, throughout the world, for monitoring energetic nuclides, virtually none exist for low-energy emitters. This is particularly disturbing when one notes that many of these weak emitters have significant biological application. In order to design better flow cell radiation detectors, it became clear that more fundamental information was needed about the effects of geometry and light scatter, and the relationships between the two. This investigation was initiated to hopefully develop such information. EXPERIMENTAL

Equipment and Reagents

All scintillation measurements were carried out using a Packard Tri-Carb liquid scintillation counter, model 4530. This unit contains a low-resolution multichannel capability used solely for a visual display of the pulse-height spectrum. Two of the spectra shown in this paper are simply photos of this display. When detailed spectral information was needed, appropriate linear signals were taken from the Packard counter and fed to a Nuclear Data multichannel analyzer system. That system included all of the standard MCA input/ output and spectrum storage features. As high resolution was not needed for these studies, data were collected in a minimal 128 channels. The clear, polystyrene spheres (three size distributions) were synthesized in the Department of Chemistry, University of Tennessee, Knoxville. These spheres were processed with conventional scintillation fluors and organic sol-

198

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

vents to produce the scintillating beads used here. The '4C test solution (as aqueous carbonate) was standardized by liquid scintillation counting using the NBS hexadecane as an internal standard. 241Am (as nitrate, in dilute nitric acid,

pH 2 to 2.5) was assayed via 2r gas counting and confirmed in a liquid scintillation counter. All other reagents and fluors used in this study were reagent or scintillation grade. Monte Carlo calculations were carried out on three different personal computers that all used the conventional MS-DOS operating system. Each computer was fitted with a numeric coprocessor chip (8087 or 80287) to speed the large number of floating point calculations simulation requires. Source code was written in Turbo Pascal (Borland International) that was compiled to executable files using the Turbo compiler, versions 3.01 and 4.0. Preparation of Scintillating Beads The untreated polystyrene beads had certain physical and optical properties

that were crucial to the success of this project. First, they exhibited an extremely clear internal structure that was virtually fracture free. Also, the surface of each bead looked mirror smooth. Finally, the majority of beads assumed an almost perfect spherical shape. The visual effect through an optical microscope was similar to looking at drops of water. Although there were some spheres that were undeniably poor in optical quality, their number was small. The removal of such beads did not seem to be a viable task. The beads were received already divided into three size distributions. There were many more beads of the larger size; these were selected for the preliminary tests directed toward endowing the beads with efficient scintillating properties. It was known that the polystyrene beads, when placed in toluene, would swell to several times their dry size. An obvious first approach was to dissolve scintillation fluors in toluene, place the beads in this solution, let them swell, filter the expanded beads, rinse them with ethanol, and air dry them so they shrink to their original size. The idea was that the solid fluors trapped within the beads would serve as efficient scintillating centers when excited by ionizing

radiation. After some experimentation with different fluors and different fluor concentrations, the soak/dry technique did yield efficient scintillating beads. A 1 g portion of dimethyl-POPOP was added to 100 mL of toluene and was allowed to mix for 24 hr at room temperature. About 3 g PPO were added to

the saturated solution and the mixing was continued for 12 hr. The doubly saturated solution was separated from the excess fluors by filtration. The polystyrene beads were allowed to soak and swell in this solution for about 12 hr. The swelled beads were separated by filtration, washed with several portions of absolute ethanol, and air dried. This was the procedure used to treat all of the beads used in this study. (Some tests carried out after the start of this

work have indicated that scintillation efficiency of the beads could be improved slightly by incorporating naphthalene in the soak solution. However,

EFFECTS OF PHOTON SCATTERING

199

Table 1. Average Bead Sizes and Size Distributions Size

Ave. Radius (sm)

Small Medium

Std. Dev. (tm)

28.6 68 295

Largea

2.8 7 36

aThis bead size was not used for any of the nuclide measurements. It was used during the bead processing phase and is reported here for completeness.

to keep results consistent, this was not done for any of the beads studied here.)

Samples of each size distribution were examined by scanning electron microscopy (SEM) to evaluate the range of sizes in each distribution and to reveal clearly the surface characteristics and bulk structure of the beads after treatment. The size data are shown in Table 1. Figure 1 is a 300 x photo of the 28.6 micron cut; the highly damaged bead is obvious. Some of the measurements carried out in this study required surface alteration of the scintillating beads from very smooth to thoroughly diffuse. A piece of fine sandpaper (8 x 10 in) was glued to a piece of plate glass, rough side exposed. The sample of beads to be etched was placed on the sandpaper surface along with 1 to 2 mL of distilled water. A second piece of plate glass was placed on top of the beads and gently rotated by hand to work the beads

20KV X300 Figure 1.

8120

Photo (300X) of 28.6 micron radius beads.

[email protected] M

200

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

against the abrasive surface. As the beads are relatively soft, only two to three minutes of grinding were needed. The beads were washed into a small gas bubbler and separated from the sandpaper debris by flotation. The beads were inspected with an optical microscope; the surfaces were sufficiently etched such that it was not possiole to see inside the beads. Also, no changes in the

bead size could be discerned, although this was not rechecked with SEM photo. Radionuclide Measurements

All activity measurements using the scintillating beads were carried out using an assembly similar to that shown in Figure 2. A glass tube of approximately 5 mm. diameter was cut to an appropriate length and flame sealed at one end. A small amount of white, room-temperature curing silicone rubber was placed on the inside bottom of a standard LSC glass vial. The sealed end of the tube was placed in the center of the vial and pressed into the rubber. The silicone was allowed to cure for 24 hr. For use, a selected sample of scintillating beads was transferred into the central glass rod and vibrated to aid packing. The activity to be measured was pipetted onto the beads. Water was added to

the space between the vial and rod (to act as a light coupler), the vial was capped, and the assembly was gently centrifuged to draw the sample into the

beads and eliminate air voids. This device was transferred to the Packard instrument for counting. The Monte Carlo Simulations Spheres of uniform size can be loaded into a container in two different close

packed arrangements of identical efficiency: hexagonal and face centered cubic. In both cases, the spheres fill about 74 of the container, which leaves

26 void volume. It is interesting to note that 74 is not the maximum volume of space which can be occupied in the packing of spheres, although it is the maximum for symmetrical periodic arrangements. For irregular packing it

can be shown that the maximum must be less than 78%. Although crystallographic studies tend to focus on the arrangements of packed spheres, this investigation gives major importance to void geometry. Spheres that are hexagonally close packed create a "unit" void that is illustrated in Figure 3. The shaded portion is the void created by four spheres in two layers having mutual contact. When another layer of spheres is added (three total layers), the total void created is made up of two unit voids as seen through a single plane. All of the total voids in hexagonal close packing are of this type. Cubic close packing exhibits a symmetry that requires four layers

rather than three. The effect on the void structure is that two different arrangements of the unit void are created. These are usually referred to in the literature as tetrahedral and octahedral interstitial holes. Both hole structures must be considered in the cubic pack simulation.

EFFECTS OF PHOTON SCATTERING

Figure 2.

201

Counting assembly using scintillator beads.

The Monte Carlo simulation is designed to answer the following question: if scintillating spheres are close packed and if a liquid containing a radioactive emitter fills the void volume, what fraction of the decay particles are expected to reach the solid spheres? This is, of course, the geometry. Several factors must be considered in designing the simulation. These include:

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

202

Figure 3.

Diagram of void volume in hexagonal close packed spheres.

the emitted particle (alpha or beta) the decay energy range-energy relationships and absorption characteristics the absorber composition the sphere size the detailed geometry of the voids

Four major simulations (two different but related programs) were used to develop geometry information as a function of sphere size. These were for 241Am aiphas and '4C betas for both the hexagonal and cubic close packing arrangements. After the appropriate program is loaded, the user is asked to supply: (1) the decay energy (an average of 5.46 MeV was used for 241Am, (2)

the absorber density and effective atomic mass, (3) the hexagonal or cubic close packing, and (4) the number of decay events to evaluate (2000 to 5000).

At this point the simulation starts and continues until the final result is obtained. The geometrical aspects for both alpha and beta simulations are very similar. Using a fixed coordinate system, a random point is selected and tested to verify its position within the liquid phase (the void volume). This point is used as the origin of the decay event. Next, a random direction is selected to create a

linear displacement from the origin point. Although these two operations appear to be relatively straightforward, much of the program run time is used

EFFECTS OF PHOTON SCATTERING

203

up here. The complex geometry and extensive spacial testing cause this. Also, for cubic packing, even more verification must be carried out. Processing from this point is quite different for alphas and betas. Simulations of alphas are simplified by two important factors. Alpha radiation is monoenergetic and the emitted alpha particles travel in virtually straight lines. After the emission origin and random direction are determined, an alpha range corresponding to decay energy helps to create a vector from the event origin. The alpha range is calculated from experimental air measured alpha ranges combined with the Bragg-Kleeman7 rule (to correct for absorbers other

than air). Finally, the vector is tested for intersection with any of the voidsurrounding spheres. If intersection is not observed, and the alpha range could go to a second layer of surrounding spheres, than another algorithm is used to evaluate intersection. All events that intersect any sphere are totaled; the result is divided by the total of events tested. The outcome is the predicted geometry. A convenient simplification can also be employed for evaluations of beta simulation response. Although betas are not monoenergetic and can travel in very convoluted paths, an empirical relationship that exits between Em,, and linear distance can be used to evaluate the response of an entire beta spectrum. This requires that the maximum beta range of the simulated nuclide be sufficiently long to allow at least some intersection of the beta energy distribution (from the origin of the decay event) with the solid scintillator. The technique used is described in detail in Rucker et al. ;6 the geometry studied, however, is quite different from that examined here. After the degree of intersection of each beta distribution is calculated, the individual percent values are averaged to predict the geometry. RESULTS AND DISCUSSION

The first series of tests were designed to evaluate the basic quality of the scintillating beads. The two major factors involved are photon yield and photon transmission. The baseline values of these parameters are important. If the bead emission and transmission characteristics proved substantially inferior to a quality liquid scintillator, it would have been necessary to devise an alternate procedure for bead preparation. Aqueous samples of the 241Am tracer were mixed with different 15 mL. aliquots of Insta-Gel and Opti-Fluor liquid scintillation cocktails contained in standard size vials. The tracer was also added to bead scintillator samples

contained in the assembly described above. A pulse-height spectrum was obtained for each sample. Figure 4 shows such a spectrum obtained in InstaGel and Figure 5 a spectrum using 68 micron beads. The marker is at the same position in both spectra. Although the bead spectrum shows considerable energy degradation due to absorption in the liquid phase, it is quite clear that the full energy peak is both

sharper and at a higher pulse height then that seen in the liquid scintillator

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

204

r)

Li) 1ti C\J-4W

Li) Li) Li)

Li)

U

C\J

j _J

-4 c.j CJ ('j -4 c L4J

L

c

c!, (r-1 4

Figure 5.

Pulse-height spectrum of 241Am using 68 micron beads.

REGION C

Et(RY,KEV/CFi :i-iz: -$øøø] REGIOt A REGION P i-2ØØ] 9382

1ø12

206

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

spectrum. This was a most surprising result. It is not really possible to determine the individual quantitative effects of photon yield and transmission from these data, but I believe it is fair to conclude that the bead scintillators do extremely well in both of these parameters. One might argue that the bead results could arise from an outstanding photon emission combined with a somewhat attenuated photon transmission characteristic. While such a conclusion might be suggested, the observed optical properties of the beads clearly argue against it. Similar tests were made with '4C tracer. Again bead results were very good, although they were more difficult to observe because of the continuous nature of the beta spectrum. Figures 6 and 7 illustrate the results obtained with the four Monte Carlo simulations. The second order regression lines in both figures are only plotted to show point continuity. It is not suggested that they represent an analytical solution to the simulation. As might be expected, all four of the simulations tend to a geometric efficiency of one as the sphere radius goes to zero. The responses of the alpha regressions show less sensitivity to sphere radius when the radius is small; the opposite is true for the 'C beta simulations. The values obtained for cubic packing are less than those obtained for hexagonal packing for both nuclides. The Monte Carlo simulations only give an expected geometric efficiency for a given sphere size. This efficiency can be thought of as the maximum detection efficiency that could possibly be observed in the system. Of course, in real counters, results are always less than the geometry because of factors such as the minimum energy required for photon generation, presence of nonradiative

1.000 a HEX CLOSE PACK o CUBIC CLOSE PACK

Mi-241 EXP. VALUE

0 0.800

a U

2 0.600 0 0 IJ

SaUD UNES SHOW

0.400

0.200

0.000

2nd ORDER REGRESSION.

0.005

0.010

0.015

RADIUS (cm.) Figure 6.

Monte Carlo simulation of 241Am and experimental data.

0.020

0.025

EFFECTS OF PHOTON SCAUERING

207

1.0 A HEX CLOSE PACK

0 CUBIC CLOSE PACK

0.8

0.6

0.4

SOUD UNES SHOW 2nd ORDER REGRESSION.

0.2 0.000

0.005

0.010

0.015

0.020

0.025

RADIUS (cm.) Figure 7.

Monte Carlo simulation (linear plot) of 14C.

absorption processes, photon collection efficiency, and electronic thresholds in

the counter. Also, it would be unreasonable to assume that close packing would be of one type only or that close packing could even be achieved throughout the majority of detector volume. However, if all of these factors remain essentially constant within a given system, the validity of the simulation can be verified by comparing an experimental curve with the appropriate simulation. A shift to lower efficiency would be expected but the shape of the curves should match. Figure 6 shows two experimental points for 241Am obtained with two different size smooth-bead samples. While it is not possible to verify the curve with only two points, the experimental values do exhibit the trend of the simulation. Figure 8 shows the '4C simulations replotted in a semilog format along with four experimental points. The average experimental values for the smooth beads lend additional support for the simulation results. Figure 6 also includes data obtained with etched beads that exhibit a high degree of light scattering. Along with the drastic drop in counting efficiency, one sees a surprising reversal in the response as a function of bead size. Table 2 summarizes the counting results for data shown in Figures 7 and 8. CONCLUSIONS

Monte Carlo simulations for heterogeneous flow cell radiation detectors (fabricated with spherical scintillator beads) show somewhat different geometric response functions for alpha and beta emissions. For alpha radiation, the

208

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

1.0

0.9 0.8 0.7

A HEX CLOSE PACK O CUBIC CLOSE PACK

C-14 EXP. VALUE (SMOOTH)

* C-14 EXP. VALUE (ETCH)

0.6

0.5

0.4

I I

I

0.3

I

0.2 0.000

0.005

0.010

0.015

0.020

0.025

RADIUS (cm.) Figure 8.

Monte Carlo simulation (semi-log plot) of 14C and experimental data.

response is relatively flat up to about a 40 bead radius. This implies that particles smaller than this value probably will not materially improve alpha detection efficiency in practical implementations. The experimental data observed with 241Am support the shape of the observed simulation function which further strengthens the idea that ultrasmall particles are not needed. The overall maximum pulse height and spectral resolution (not counting efficiency) exceed that obtained with a typical homogeneous cocktail for aqueous samples. Conversely, the '4C response functions exhibit an increasing slope as bead radius drops. Obviously, much more attention must be directed toward bead size for low-energy beta emitters. The experimental data with smooth and rough bead surfaces demonstrates, however, that significant photon scattering is an extremely important limitation in heterogeneous cells. The minimal conditions investigated in this study show that these cells may be 3 to 4 times more sensitive to photon scatter and absorption than to geometry. The conclusion is Table 2. Counting Efficiency of 241Am and 14C With Scintillating Beads Bead Radius

Counting Eff.

Nuclide

(am)

Bead Surface

(ave.)

Am-241 Am-241 C.14 C.14 C-14 C-14

28.6 68 28.6 68 28.6 68

Smooth Smooth Smooth Smooth Etched Etched

0.79 0.75 0.70 0.55 0.30 0.39

EFFECTS OF PHOTON SCATTERING

209

clear that the development of high-sensitivity heterogeneous flow cell detectors may ultimately be determined more by the optical rather than physical properties of the scintillator cell.

ACKNOWLEDGEMENTS

The author would like to thank Professor Spiro Alexandratos, Department of Chemistry, University of Tennessee, for providing the polystyrene beads used in this study; Ms. Lisa Rozevink, Science Alliance Summer Participant, University of Tennessee, for writing major pieces of the Pascal source code used in the simulations, executing some of the alpha simulations, and making several of the alpha response measurements with the beads; and Dr. R. G. Haire, Transuranium Research Laboratory, Oak Ridge National Laboratory, for supplying the 241Am tracer.

REFERENCES

Mackey, L.N., P.A. Rodriguez, and F.B. Schroeder. J. Chromatogr., 208, 1 (1982). Everett, L.J. Chro,natographia, 15, 445 (1982).

Harding, N.G.L., Y. Farid, M.J. Stewart, J. Sheperd, and D. Nicoll. Chromatographia, 15, 468 (1982).

Frey, B.M. and F.J. Frey. Clin. Chem., 28, 689 (1982). Rucker, T.L., H.H. Ross, and G.K. Schweitzer. Chromatographia, 25, 31(1988). Rucker, T.L., H.H. Ross, and O.K. Schweitzer. Nuci. Insir. Methods in Physics Research, A267, 511 (1988).

Evans, R.D. "The Atomic Nucleus," McGraw-Hill, 1955, p.652-3.

CHAPTER 18

Some Factors Affecting Alpha Particle Detection In Liquid Scintillation Spectrometry

Laurl Kaihola and Timo Oikari

ABSTRACT Pulse shape analyzers have been introduced in commercial liquid scintillation counters, such as the WaIlac 1220 QuantulusTM and 1219 SM, enabling separation of alpha particle spectra from other types of radiation in a single measurement and making ultra sensitive alpha counting possible. Background count rates as low as one count per day for 214Po in a teflon vial

are achievable while still retaining nearly 100% counting efficiency. The number of photons per decay event however, is relatively low in liquid scintillation counting, leading to energy resolution of the order of 10%. Energy resolution has been found to improve with low water to cocktail ratios and opaque vial types. Easy and fast sample preparation together with the ultra low background make liquid scintillation spectrometry a competitive method in environmental counting of alpha particles.

INTRODUCTION

Liquid scintillation counting (LSC) is a well established method for beta particle counting. Its use in alpha particle counting has suffered from large background count rates in energy range due to the cosmic ray and environmen-

tal radiation caused events in the cocktail. Background reduction for alpha counting requires identification of the origin of the background. It has long been known that the relative amounts of the prompt and delayed components of fluorescent decay depend on the specific ionization, hence the type of particle causing it. Typically, the heavily ionizing particles, such as neutrons and alpha particles, produce more delayed component than beta particles, thus longer pulses. This phenomenon is the basis for the pulse shape analysis (PSA)

as particle identifier, and it makes possible rejection of the beta-like background component in alpha counting. The method has been applied in nuclear physics for neutron counting in the presence of gamma radiation since the '50s. In alpha counting it has also been applied with good success and is well reviewed)'2 Experimental liquid scintillation counters, developed in the '70s, used zero 211

212

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

crossing techniques for pulse shape discrimination.34 These devices contained silicon oil as a light guide, improving the coupling from the sample to a single photomultiplier tube, thus giving an excellent energy resolution. Wallac 1220 Quantulus, an ultra low background liquid scintillation spectrometer, has contained a pulse shape analyzer since 1987. Being a general purpose alpha-beta counter with an automatic sample changing mechanism, the optical coupling described above was not applied, but a standard detector configuration with two PMTs was maintained.

The Wallac pulse shape analyzer uses a concept quite similar to the one adopted by, e.g., Brooks.5 It initiates the integration of the pulse tail after 50 nsec from the leading edge and compares it with the integral of the total pulse. Surface mounted electronic components manufacture the high speed analog circuitry that produces the pulse shape related signal without amplitude dependence. By setting the PSA level through software commands it is possible to route alpha-like events into one half of the MCA and beta-like events (due to beta particles, Compton recoil electrons, Cerenkov radiation, and X-rays) into the other half in a single measurement. Maximum resolution of 1024 channels is available with a logarithmic energy scale. The Quantulus is equipped with a guard counter, set in anticoincidence with the sample detector PMTs, to minimize the background component caused by environmental radiation, cosmic flux, and gamma flux from the bedrock and building materials. Further, a user programmable pulse amplitude comparator is available for reduction of cross talk events.6 Pulse shape separation characteristics depend on cocktail. Adding naphthalene to conventional toluene or xylene based cocktails enhances their pulse shape separation. Modern Hisafe cocktails by Wallac contain naphthalene derivatives as solvent and offer excellent pulse shape characteristics without any further naphthalene addition.

PULSE SHAPE ANALYSIS

Since the pulse shape characteristics vary between different cocktails, the pulse shape analyzer has to be set individually for each one to achieve the optimum separation of alpha and beta particle spectra. This can be done by preparing a pure alpha and beta sample whose amplitude spectra overlap, e.g., 36Cl and 241Am. The PSA level is then scanned for acceptable rejection of the radiation not of interest. With conventional cocktails one may allow say 5% loss of alpha particle counting efficiency to remove beta-like background, if alpha particle spectrum is only needed. If both particle types are to be measured, new cocktails containing naphthalene derivatives will give good separation (Table 1).

FACTORS AFFECTING ALPHA PARTICLE DETECTION IN SPECTROMETRY

213

Table 1. Beta Residual, the Percentage of Betas Remaining Among Alpha-like Events. Normalized to 100% when PSA is off. The Residual is Given for 36C1 Beta Particles in 241Am Alpha Window for Corresponding Alpha Counting Efficiencies. The Sample is Contained in 1 mL Water with 10 mL Optiphase Hisafe 3 in Equilibrium with Air. Beta Residual Alpha

Eff%

%

25 50 75

100 100 100 100

100 100

100 120 130 140 150 160

99.3 98.9 95.9 88.7 73.4 54.4

PSA

o = off

84.7 36.1

2.75 0.098 0.025 0.02

<0.02

ENERGY RESOLUTION IN LIQUID SCINTILLATION SPECTROMETRY

Energy resolution is proportional to the number of photons created in the decay event.7 With LSC one may not expect to achieve the energy resolution of

solid state detectors, because number of liquid scintillator charge carriers is considerably greater than that of photons detected in a typical liquid scintillation event. Therefore, considerable improvement in energy resolution will only be achieved with a profound change in the light production properties of the cocktail. Current cocktails have about 10% resolution in opaque vials at 5 MeV alpha energy (Figure 1). Corresponding resolution in standard glass vials is poorer, 20% (see item 2). There are ways to improve the energy resolution: High light output cocktails give better results. Opaque vials transmit light with fewer losses than standard glass vials where photons can get trapped through total reflections in the vial wall. Teflon and etched glass have the best resolutions so narrower windows can be used to obtain smaller backgrounds with the same counting efficiency. Alignment of the liquid meniscus with the center axis of the PMT gives the best sensitivity and optimal resolution.

The lower the quench level, the better the photon emission; this leads to improved energy resolution. In practice this means having the smallest amount of water, contradicting the need to have as good a lower limit of detection as possible. In extractive methods, using two phase samples, radioactivity from water is transferred into the cocktail phase thereby eliminating the aqueous quench. Nitrogen bubbling improves the performance further, due to the removal of quenching by dissolved oxygen.

214

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

LA)

4.000 CPIVch

(B]

4.088 CIIVch

180

59.4? .in 59.4? .i

200

1

A:\DT15781L

400

IMTECB( 598- 777) J$Q

A:Dm!A\M6'MI15781L800 SPill

* 35 *

LA]

688

788 *

SF112

1800

71.562 CIII (B] 311.28? 3.291 ( 4.6 ) 6.%3 ( 2.2 x)

Figure 1. 226Ra spectrum from 2 mL water sample with 18 mL Optiphase Hisafe 3 in 20 mL teflon vial. Alpha emission peaks from left to right are 22"Ra (4.78 MeV), 222Rn (5.49 MeV), 218Po (6.00 MeV), and 214Po (7.69 MeV).

In the logarithmic amplitude scale, the alpha peak shape is quite close to being symmetric (Figure 2). At very high alpha particle energies, an asymmetric shape may result due to increased losses close to the vial walls, c.f. 214Po, 7.7 MeV.

PERFORMANCE IN ALPHA COUNTING

Glass vials contain U and Th isotopes, their contribution results in a wide background spectrum under the alpha peaks (Figures 3 and 4). These alpha particles are emitted in a 2 r geometry and are subject to losses due to their

need to penetrate through a layer of glass in order to reach the cocktail. Background also contains slow fluorescence from the glass excited by the environmental radiation. Attenuation of the beta-like decay events among alpha-like ones by pulse shape analyzer is given in Table 1. Typical background count rates and lower

FACTORS AFFECTING ALPHA PARTICLE DETECTION IN SPECTROMETRY

(Al 18.8 (B) Z388

CPVch CW.cb

(C)

CPPIcl

688 INTECE(

JI(i FIgure 2.

1

09.17 mm

AOQRN811581L881 SF112

2.84 89.17

A:CIm18181L1 51112

mm

in

A:'JIOQ

650

215

1681L1 SF812

750

700

- 750) [A) 47158.711 dO (BI 73346.578 CPO [C) 47.875 CIII * 8.1 ) 68,904 ( 8.1 c) 482.513 ( 8.7 x) 69. C

241Am alpha emission spectrum (5.5 MeV) in teflon vial (A), etched glass vial (B), and in standard glass vial (C).

limits of detection are given for 241Am (5.5 MeV) and 2'4Po (7.7 MeY) in two instruments.

CONCLUSIONS

There are several advantages to alpha particle counting with the LS method compared to solid state detection, gridded ion chambers, or the nuclear track

method, especially when high energy resolution is not required. The LSC performance does not involve a variable geometrical factor, variable counting efficiency, self absorption, and detector contamination. A counting efficiency of at least 98 with a background of less than 0.1 cpm can easily be achieved. Sample preparation is simple; saving time in low level counting. Sample sizes considerably larger than in solid state alpha spectrometry can be analyzed, and enrichment is a simple procedure. Spectrum analysis allows early recognition of alpha-emitting nuclei in the samples and thus time for decision making if any further analyses are required.

216

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

16? 13

N /cli 19.6? nm N /di 428833 mm

A:\Lau-9-l4281La

SPII2

A:\Lg11-9-?5781N.8

SF112

p

188

1

*

INTECR(

1-1024)

WJ4QI: 4

* 3S *

Figure 3.

?88

680

[Al 9fj.UJ0 N [ill 2868 ( 8.3

)

1888

1293.88 N 1.08 ( 8.3 )

241Am alpha emission and background spectra in a standard glass vial (1 mL water + 9 mL Lumagel + 20% wlv naphthalene). Background count rate in full window is 0.38 cpm and 0.002 cpm in 214Po window.

Table 2. Alpha Peak Background (cpm) and Lower Limit for Detection (LLD, Bq/L) for 100 mm Counting Time in the Wallac 1220 Quantulus and 1219 SM (which contains

no guard counter). Sample is 2 mL water + 18 mL Lumagel + 20% w/v Naphthalene in Equilibrium with Air Instrument Vial bkg

LLD

Comment

1219 SM 1219 SM 1219 SM 1219 SM

glass glass, op glass, op teflon

0.23

1.2 0.8 0.08 0.05

Am-241 Am-241 P0.214 Am.241

1220 1220 1220

glass teflon teflon

0.03 0.005

0.08

0.001

0.002

Am-241 Am-241 P0-214

0.11

0.03 0.02

0.01

FACTORS AFFECTING ALPHA PARTICLE DETECTION IN SPECTROMETRY

LA]

19918

N /ch

1475

4 N /ch 4288.

[B]

488

188

1

INTEGR(

In

1-1824)

]JQ:4 *3S*

217

A:'LgI1-9-Na343a1L

112

A:\L1+3\Qa881L

$12

680

[A] 65998.80 N

700 * (B]

2437( 8.4)

1000

.80 H

(40.5x)

Figure 4. 241Am alpha emission and background spectra in a teflon vial (1 mL water + 9 mL Lumagel ± 20% w/v naphthalene). Background count rate in full window is 0.013 cpm and 0.001 cpm in 214Po window.

REFERENCES

Horrocks, D.L. "Pulse Shape Discrimination with Organic Liquid Scintillator Solutions," Applied Spectroscopy, 24:397 (1970).

Brooks, RD. "Developments of Organic Scintillators," Nucl. Instr. Methods, 162:477 (1979).

McDowell, W.J. "Alpha Counting and Spectrometry Using Liquid Scintillatio Methods," Monograph, NAS-NS-31 16 (DE86 007601) (publ. by Technical Informa-

tion Center, Office of Scientific and Technical Information, US DOE), 1 January 1986.

McKlveen, J.W. and W.J. McDowell, "Liquid Scintillation Alpha Spectrometry Techniques," Nuci. Instr. Methods Phys. Res. 223:372 (1984). Brooks, F.D. "A Scintillation Counter with Neutron and Gamma-Ray Discriminators," Nuci. Instr. Methods 4:151 (1959). Kaihola, L. "Liquid Scintillation Counting Performance Using Glass Vials with the Wallac 1220 QuantulusTM," paper presented at tis conference.

218

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Oikari, T., H. Kojola, J. Nurmi, and L. Kaihola, "Simultaneous Counting of Low Alpha-and Beta-Particle Activities with Liquid-Scintillation Spectrometry and Pulse-Shape Analysis," App!. Radiat. Isot. 38A(lO):875-878 (1987). Salonen, L. "Simultaneous Determination of Gross Alpha and Beta in Water by Low Level Liquid Scintillation Counting," paper presented at the 2nd Karlsruhe International Conference on Analytical Chemistry in Nuclear Technology, Karlsruhe, FRG, June 5-9, 1989.

CHAPTER 19

Modern Techniques for Quench Correction and dpm Determination in Windowless Liquid Scintillation Counting: A Critical Review

Stat van Cauter and Norbert Roessler, Ph.D.

ABSTRACT With the advent of readily available computing power, it seems reasonable to describe the probability of the liquid scintillation process replacing quench curves. The pulse height spectrum and the counting efficiency can be expressed as functions of the scintillator photon yield as measured by the external standard Compton spectrum. The energy dependence of the photon yield can be accounted for by combining information from the sample spectrum and the Compton spectrum. A variety of numerical techniques can then be applied to describe the spectral response and to calculate dpm values of an unknown from appropriate reference spectra. These methods for determining dpm values have recently been applied to liquid scintillation counters equipped with multichannel analyzers. Although these "windowless" dpm determinations are very convenient to the user, many inherent characteristics of the scintillation process may render results invalid under certain experimental conditions. Indeed, the photon yield at a particular energy is significantly affected by the microscopic environment of the nuclide and the composition of the sample, expressed by the concept of ionization quenching. As a further complication, the prompt and delayed components of the fluorescence depend on these parameters as well. As a consequence, theoretical expressions proposed to account for ionization quenching must remain approximations: their validity needs to be verified for each cocktail-sample combination. Due to the unstructured nature of LSC pulse height spectra, it is difficult to accomplish this verification using spectral overlay and comparison techniques. Pulse height discrimination is often required after the fact to make results less dependent on varying experimental conditions. Numerical methods applied to windowless dpm counting will be reviewed, and the results will be compared to methods based on quench compensated discrete window settings.

INTRODUCTION

In the field of optical spectroscopy, energy dispersion is obtained using filters or monochromators that can be calibrated independently of the sample using absolute physical methods. In liquid scintillation counting, this is not the case. Although the response of the PMT and its related pulse shaping and 219

220

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

aplification circuits can be determined independently of the sample, quench compensation is required to correlate the pulse height measured in the multichannel analyzer with the electron energy of the nuclear event. Most commercial systems monitor the quench level using the mean sample pulse height (SIS)

and the Compton edge from an external source of gamma radiation (tSIE or H#). Horrocks has presented the results of extensive studies determining the energy needed to produce a photoelectron in a liquid scintillation counter.' This energy depends both on electron energy and on quench level. These functions are not linear. Rundt et al., investigated theoretical expressions for the effect of ionization quenching and found them all dependent on adjustable parameters which vary with the solvent.2 The theoretical functions are valid for chemically quenched homogeneous samples only. The reference method of compensating for quench effects, the method of standard addition, is too labor intensive for routine use. Usually, "quench curves" are constructed using a series of standards at various quench levels. Ideally, the standards are prepared using the same cocktail and quench material present in the sample. A variety of mathematical methods is then used to interpolate counting efficiency values for the measured value of the quench indicating parameter. In practice, commercially prepared sets of quench standards in a toluene based cocktail, yield satisfactory results for 3H and '4C. In no small part, this is due to the large body of experimental information available on the use of these two nuclides. Indeed, hardware, software, and cocktail formulations have been optimized to ensure reliable results for the broadest possible range of applications. Especially important are experiments using emulsifier cocktails to count aqueous samples. The use of liquid scintillation counters equipped with multichannel analyzers and the computing power to automatically analyze the results offers considerable convenience for the user. Channel selection in the MCA has replaced adjustment of analog threshold values, and spectral storage features even allow for post run analysis. This has allowed the development of windowless protocols for dpm determinations that are very general and almost as easy to use as "absolute" instruments such as monochromator based spectrophotometers. Nevertheless, the physics of scintillation counting has not changed. This presentation will focus on quench correction and dpm determination of dual labeled samples containing 3H and '4C because these are the two nuclides most commonly counted. The discussion can easily be generalized to other nuclides or to the triple label case. Figure 1 is a schematic drawing of the beta spectra of 3H and '4C, as well as a composite of the two. The implicit assumption made in the drawing is that the

spectra are from a sample at the same quench level. As is well known, the shape of a spectrum is distorted as quench changes. Consequently, accurate quench correction is a prerequisite for accurate dpm calculation, whether of single or multilabel samples.

QUENCH CORRECTION AND DPM DETERMINATION

Figure 1.

221

Schematic spectra for 3H, 14C, and dual label sample.

FIXED WINDOW SETTINGS

When using fixed window settings (see Figure 2), the multichannel analyzer

is used in a quasianalog fashion. The energy scale is divided up into two regions, A and B, and the counts from these channels are summed up as cpmA and cpmB. According to the principle of radionuclide exclusion, the lower limit of region B is chosen above the endpoint energy of the low energy nuclide. This allows direct calculation of the high energy nuclide dpmH without interference from the low energy nuclide. The efficiency for the high energy nuclide in channel B, EHB, is interpolated from a quench curve. dpmH =

cpmB

(1)

When calculating dpmL of the low energy nuclide, the counts need a spill down correction due to the high energy nuclide in the low energy channel, EHA. The efficiency for the low energy nuclide in the A channel, ELA, is also needed. dpmL

cpmA - dpmH EHA

=

ELA

(2)

The relative precision for the low energy nuclide in the dpm value can be expressed as proportional to factor analogous to the figure of merit (E2/B) where B is the spill down (= dpmH EHA).

222

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

COMPOSITE SPECTRUM

dN dE

I ke0n Figure 2.

Region B

Key

Fixed window region settings.

'7o Precision

E2 LA

/

E2\

dpmH EllA

1

B)

The relative precision is highly quench dependent because as quench increases,

the 3H efficiency decreases and the spill down from the '4C increases. The dotted lines in Figure 3 represent quench curves obtained using the fixed window method, while the solid lines represent quench curves obtained by automatically adjusting the windows as a function of quench. In our case the quench level is determined by the quench indicating parameter of the external standard tSIE. This method is also called automatic efficiency control (AEC) because the efficiency of the high energy nuclide in the low energy channel is kept constant. As a result, the relative precision is independent of quench level.

The absolute precision of dpmL, however, still depends on the relative activity of the high energy nuclide, R = dpmH/dpmL. Precision

E2LA

LA

dpm dpmL

EllA

R EHA

As R increases the precision decreases due to increasing spill down, which leads to an activity ratio dependency.

QUENCH CORRECTION AND DPM DETERMINATION

223

% EFFICIENCY (EHB)

14C IN '4C CHANNEL

/ / /

00

dS

(E1) 3H IN 3H CHANNEL

' / 40

I' (EHA)

20

0

200

!4CIN3HCHANNEL

II

I

3 SOC

000

000

tS1E

Figure 3.

Quench curves: automatic window tracking (Solid line), fixed window setting (dashed line).

FULL SPECTRUM DPM

One method of calculating dpm without windows and using the information in the full spectrum is to separate out the counts due the two nuclides with the spectral index of the sample (SIS) (see Figure 4). Since the spectral index of the composite, SIST, is a cpm weighted average of the spectral indices of the low energy nuclide, SISL, and the high energy nuclide, SISH, it is possible to assign the correct proportion of the total count rate, cpmT, to the low and high energy nuclide, cpmL and cpmH. cpmL

CPfflH

=

SIS11 - SIS.. - SISL

SIST - SISL

= SISH - SISL

cpmT

(5a)

cpmT

(5b)

While SIST is measured, the SISH and SISL values are interpolated from a quench curve (see Figure 5), as are the full spectrum efficiencies used to calculate dpm values.

224

Figure 4.

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Full spectrum dpm.

dpmL

SIS11 - SIST

= SISH - SISL

SIST - SISL dpm H SISH - SISL

cpm1 EL

cpmT EH

We see from the curve that both the efficiency and the SIS interpolations depend on the accuracy of the external standard quench indicating parameter since the curves have a slope. Furthermore, the leveraging effect in Equation 6 on the difference between SIST and SLS (and to a lesser extent SISL) magnifies the effect of any inaccuracies on calculating dpmL. The dpm calculation is now quite sensitive to experimental conditions affecting the QIPES differently from SIS.

Figure 6 contains the results of SIS measurements of tritium standards made up in a variety of cocktails. We see that the relationship between the quench indicating parameters of the external standard and the internal standard is similar for all cocktails, but not identical. Using the tSIE (see Figure 7) brings some improvement, but a discrepancy remains. Since the early work on homogeneous solutions as presented in Birks's3 monograph reviewed at the International Symposium on Organic Scintillators at Argonne,4 the photochemical literature has expanded into the area of colbid solutions such as those used in emulsifier based cocktails. Interest in laser dyes, which include some common scintillators, has led to a better understanding of the prompt and delayed components of scintillator emissions. It turns

QUENCH CORRECTION AND DPM DETERMINATION

225

SISA SISH

SIST

SISL

-

LQIP8)

tSIE

EFFA EFFH EFFL

tSIE

LtSIE Figure 5.

Full spectrum dpm quench curves.

out that the microscopic environment of the nuclide in the sample varies with sample load and sample composition. Furthermore, this effect is different for the prompt and delayed component of the scintillation pulse. Tracks from high energy electrons pass through the different phase regions of emulsifier cocktails so that their effect tends to be averaged out. Low energy electrons, such as those emitted by tritium, do not. This makes it difficult to correlate scintillation efficiencies of high energy electrons with the scintillation efficiency of low energy electrons. Unfortunately, it is precisely this correlation that is at the heart of all spectral interpolation schemes, whether they use the sample spectrum alone or a combination of sample and external standard spectra. This

Figure 6.

(IS)

QIP

150

300

alp (ES)

250

Effect of various cocktails on the relationship QIP(1s) =

1CC

4

5

6

7

8

400

450

500

-A-

To tue ri e

IG - H20

LLT - H20

:

iritiurn Standr'5I

r')

Figure 7.

100

tSIE

300

Effect of various cocktails on the relationship SIS = f(tSIE).

200

400

500

H20 To I ue n e

10

--(c-)--

LLT - H2(

Tritium Sorid

228

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

problem becomes even more critical when quench standard and sample are counted in different cocktails. As a consequence, theoretical expressions accounting for ionization quenching break down in the critical low energy region. They are approximations whose validity needs to be verified for each cocktail/sample combination. SPECTRUM OVERLAY TECHNIQUES

An alternative method of windowless counting that does not reduce the spectrum to a single parameter is the spectral overlay technique (see Figure 8).

In this technique, the QIP of the external standard is measured and used to interpolate the spectra of quench standards channel by channel to the quench level of the sample. The spectra are normalized to separate the shape function from the counting efficiency. The cpm values of the nuclides are adjusted by a method of fitting the least square of the standard spectra to the composite spectrum of the sample. When the best fit is obtained, the cpm values for the nuclides are known, and the dpm values can be calculated directly using the efficiencies. dpmL

-

dpmH

-

cpmL EL

cpmH EH

(7a)

(7b)

This numerical method is derived from fields such as optical spectroscopy, which allow for absolute energy calibration. The error in the energy value is at least constant if not negligible. This allows for curve fitting routines to evaluate the goodness of fit based only on the variability of the photon count rate.

The energy is assumed to be known exactly. In LSC count spectra, this assumption is not valid. As a result, these algorithms tend to give spurious results when challenged by errors in the energy value.

One way to overcome this problem is to allow the fitting procedure to "adjust" the quench level at which the standard spectra are calculated. Now the quench level is no longer a measured value; it becomes the third parameter in a

least square fit of the interpolated spectra to the measured composite spectrum. The added parameter improves the fit. However, the algorithm assumes that the physics of the sample is identical to that of the standards. That is, the effect of sample preparation is not accounted for. Figure 9 illustrates the consequences of adjusting QIPES. Two single label tritium standards are shown in toluene and a gel phase cocktail (Insta-Gel). The Insta-Gel spectrum is slightly deformed due to beta self-absorption in the gel phase. In order to improve the least square fit, an overlay technique will

raise the external standard QIP to broaden, and thus deform, the whole spectrum.

9pect1'

I

Figure 9.

10

11

4

2

I

9

I

8

I

7

F

6

I

5

Consequences of adjusting 01ES on spectrum overlay technique.

I

I

a.

I

It,

I

11

1

12

13

IF 14

16

1

2

I

3

I I

a

I

+

I

+

-r v v-- -

4 5 6 7 Quenche stanciaards

I

- -- -- -r -

4--

10

Figure 10. dpm Recovery of toluene samples using OptiFluor Standards. + added dpm, A DOT-dpm with OptiFluor library, V AEC.

240

260

b 280

300

320

340

232

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Test Tellers (3H)quenchestrds:246.5 +/-3 1e3 dpm -, - pdwd -, 11(8 - - Ib sm 1

cal

360

A"

A

330 310

1J290

_*_

270 260 A

230 1

2

3

&

5

4

6

7

8

9

10

tWIW

Test Tellers (3H)quenchestrds:2 46.5 +1-3 1 e3 dpm

- Ib

- V - .cks,d -0-

sm

2c&

lc.I

280 270

240 2 Figure 11.

3

4

6

6

7

8

9

10

O,*'d. stvi.wds

Effect of fine tuning" the spectrum library on dpm recovery using the spectrum overlay technique: with one quench standard (top), with two quench standards.

QUENCH CORRECTION AND DPM DETERMINATION

233

160 140 120

100 % DPM

80

Recovery

60

40

20

'k

*Iø II' A

C

B

0

E

C

A= 2OmL glass Ultlma Gold B= 2OmL plastic Ultima Gold C= 2OmL glass OptI-Fluor

D 2OmL glass OptI-Fluor 0 E= 2OmL glass Insta-Gel F= 2OrnL glass Totuene

G 6mL glass Ultima Gold

Figure 12. Effect of sample preparation on % dpm recovery of single label 3H samples using AEC.

Figure 10 contains the test results of toluene single label samples using an

Optifluor standard. While the automatic window racking method (AEC) results in values close to the dpm added, the spectral overlay technique leads to spuriously high dpm values. Clearly this technique is not reliable because of its sensitive to sample composition. One way to improve the performance of the technique is to "fine tune" it by adding a standard prepared in the same cocktail as the sample. This ameliorates the problem (see Figure 1 la), but even using a second quench standard (see Figure 1 ib) does not eliminate it. These results show that most reliable results are still obtained using several standards on a quench curve.

Figure 12 contains the data from a test of tritium samples prepared using different vial types and cocktail compositions. We see that the dpm recovery of

Insta-Gel is under 100% due to self-absorption. The values from all other samples are quite close to 100%. In Figure 13, we see the performance of the

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

234

160

140 120

100 DPM

80

Recovery

60 40

20

0

C

A B

C

E

0

F

A= 2OmL glass UltIma Gold B 2OmL plasLic UlUma Gold C= 2OmL glass Opti-Fluor 0= 2OmL glass OptI.Fluor 0 E= 2OmL glass lnsta-Gel F= 2OmL glass Toluene G= 6mL glass Ultima Gold

Figure 13.

Effect of sample preparation on % dpm recovery o single label "C samples using spectrum overlay technique.

spectrum overlay technique. The error in the Insta-Gel sample is particularly

large due to the effect of self-absorption, but by almost any standard of comparison, the overlay technique does not perform as well as automatic window tracking. POST COUNT dpm WITH AUTOMATIC WINDOW TRACKING

In order to obtain the most reliable performance and maximum precision it

is possible to combine automatic window tracking with automatic region adjustment. In this method, the full spectra of all quench standards are stored in the computer. This allows the software to create quench curves with any region settings, saving the user from having to recount the samples. When the sample is counted, the energy limits of region A are adjusted to obtain the best precision for the activity ratio of that particular sample (see Figure 14a). Once

QUENCH CORRECTION AND DPM DETERMINATION

235

COMPOSITE SPECTRUM

dN dE

1

R.ion

Region

A

B

K.V

% EFFICIENCY (EH&

"C IN "C CHANNEL

IELAj

'H IN 'H CHANNEL

'0-

"C IN 'N CHANNEL. I

0 200

400

I

4

t'l

000

OIP 1000

Figure 14. Automatic optimization of region settings for quench and activity ratio using Uniquench-dpm.

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

236

1E+04

H-3 DPM

Recovery &XXJ

7000

50

100

150

200

250

300

350

tSIE Figure 15.

dpm recovery of AEC and Uniquench-dpm for 3H:14C (1:15) sample.

the region is determined, the stored quench spectra are used to create a full quench curve under the optimum conditions for that sample (see Figure 14b). The improved performance obtained by adjusting the region settings is shown in Figure 15. We see that the regions set for the unquenched standards no longer give good results at higher quench levels (lower tSIE). Adjustment of the regions extends the range of good performance considerably.

CONCLUSION

In this review of quench correction methods, we see that addition of computing power to a state of the art liquid scintillation counter will result in improved performance and more convenience for the user. However, the physical problems connected with the relationship between external standard and sample spectra still remain. Software techniques based on a probabilistic anal-

ysis of instrument performance and sample composition now allow us to extract the maximum possible amount of information from the sample spectrum. However, for general purpose instruments, they are no substitute for an appropriately chosen set of quench standards and algorithms that respect the limitations on precision set by counting statistics and our limited knowledge of nuclear track photochemistry.

QUENCH CORRECTION AND DPM DETERMINATION

237

REFERENCES

Horrocks, D.L. "Energy per Photoelectron in a Liquid Scintillation Counter as a function of Electron Energy," in Advances in Liquid Scintillation Counting, S.A. McQuarrie, C. Ediss, and L.I. Wiebe, Eds., (Edmonton, Alberta, Canada: University of Alberta, 1983), P. 16. Rundt, K., H. Kouru, and T. Oikari. "Theoretical Expressions for the Counting Efficiency and Pulse Height Distribution Obtained for a Beta-Emitting Isotope by a Liquid Scintillation Counter," in Advances in Liquid Scintillation Counting, S.A. McQuarrie, C. Ediss, and L.I. Wiebe, Eds., (Edmonton, Alberta, Canada: University of Alberta, 1983), p. 30. Birks, J.B. The Theory and Practice of Liquid Scintillation Counting (New York: Macmillan, 1964). Horrocks, D.L., Ed. Organic Scintillators (New York: Gordon and Breach, 1966).

CHAPTER 20

Multilabel Counting Using Digital Overlay Technique Heikki Kouru and Kenneth Rundt

ABSTRACT The development of electronics has brought multichannel analyzers (MCA's) into liquid scintillation counting. However, the data analysis methods are not on the same level. The currently used window-based calculation methods cannot use all the information measured from the sample. In this presentation a new method called digital overlay technique (DOT) is introduced. This method efficiently uses the information available from the MCA. In DOT, the shape of the Sample spectrum is used for resolving multicomponent samples, by fitting the spectrum of each component to the measured composite spectrum. Furthermore, the fitting of a reference spectrum to the measured spectrum enables the monitoring of sample quality. In this report the basic principles of DOT are illustrated. Measurements made by a liquid scintillation counter, the Wallac 1410, which uses this principle, are presented. Comparisons are made with a conventional window-based method. The results show that DOT gives the best precision irrespective of the activity ratio of the isotopes.

INTRODUCTION

Dual label liquid scintillation counting is currently a relatively common discipline for which almost all commercial instruments provide some kind of solution. Traditionally, liquid scintillation counting of dual-label samples is performed in two counting windows. The windows may either be fixed, moving,' or automatic. Fixed window counting is a fairly uncomplicated procedure accurate enough for samples of moderate quench and activity ratios. If the quench level of the samples within one batch varies immensely, fixed windows have to be exchanged for moving or automatic windows. In the method with

moving windows the windows are first set by the user for an unquenched sample. The counter thereafter automatically moves the window limits for each sample prior to counting, according to the quench level of the sample. In the method with automatic windows, there are no counting windows during counting, but only a multichannel analyzer (MCA). The user need not specify 239

240

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

any window limits, only isotopes. After counting, the windows are selected from a table of window limits as a function of an external standard quench index for different isotope combinations. Moving windows and automatic windows have the advantage of providing quite robust methods for a wide quench range. However, window methods are not particularly suitable for counting multilabel samples containing more than two isotopes, as the windows have to be defined either by the user (moving windows) or by the manufacturer (automatic windows) through a tedious trial-and-error procedure. Furthermore, separate quench curves for different labeling situations are needed as the windows may be set differently. This means that ordinary single label quench curves just cannot be combined for multilabel counting when using moving or automatic windows. During the last few years, three new methods for dual label counting have been introduced in commercial counters: Full Spectrum DPM by Packard,2 Three-Over-Two introduced by Wallac Oy in the RackBeta LS counters in 1987, and Digital Overlay Technique introduced in 1988 by Wallac Oy in the new LS counter 1410. Three-Over-Two counts dual labeled samples with three fixed windows. The count rate of each isotope is determined by solving a set of three equations in two unknowns by using, e.g., the method of least squares. For dual label counting the number of windows with this method is three, or

one more than the number of isotopes. This method can, in analogy, be extended to multilabel counting of n isotopes as well by always having the number of windows greater than the number of isotopes n. The number of windows may naturally be much higher than n + 1. The ultimate solution is when the number of windows approaches the resolution of the MCA; this is then equal to Digital Overlay Technique (DOT).

DOT is based on a methodology similar to spectral analysis in other branches of spectroscopy and analytical sciences. 011er and Plato3 were the

first to mention spectrum analysis in connection with liquid scintillation counting. In their original paper they did not take quench into consideration though, nor did they describe how the composite spectrum was resolved mathematically. DOT makes use of a multichannel analyzer having 1024 channels. Mathematically, DOT can be described as follows. Let S1 and S2 denote the normalized spectra of isotopes 1 and 2 respectively, and s the normalized

count rate of isotope i (i = 1 or 2) in channel j. A normalized spectrum is a spectrum for which the total count rate is equal to 1 cpm. For a dual-label test sample, let U denote the predicted spectrum and u the predicted count rate in channel j. Finally, let c1 and c2 denote the unknown, true count rates of the two

isotopes in the sample. Then the following linear equation is valid for each channel:

U=

C1

S1, +

2

=1

.

.m

(1)

MULTILABEL COUNTING USING DIGITAL OVERLAY TECHNIQUE

241

or, for n isotopes n

j = 1 .. m

S

(2)

The number of equations that describe the whole spectrum of m channels is equal to m. For the general case of n isotopes it is most convenient to introduce matrix notation. Let S denote an n by m matrix comprising the n spectra S1 to

S0 and let C denote a vector comprising the n count rates c1 to c. For the spectrum U the equation

U=SC

(3)

is then valid. Let Y denote the measured composite spectrum and, y the measured count rate in channel j. In the least squares solution the count rates c are found by minimizing the squared difference between the measured and the predicted spectrum: m

=

(yj

c

-

s1)2

(4)

j=1

which is the least squares solution without weights. In matrix notation, the solution is

C = (ST S)' ST Y

(5)

The count rates c, can thereafter be converted to activities a, (or dpm1) by using the familiar relation a, = c1/e1

(6)

where e, is equal to the counting efficiency of isotope i, which must separately be determined from a quench curve. Although Equations 4 and 5 would be

accurate enough in favorable situations, they should not be used when the count rates are low and the spectra overlap considerably. The next step is to introduce weighted least squares fit. Let W denote a weight vector comprising the, so far unknown, weight values. Equation 4 then becomes

w(y

-

s)2

(7)

C = (ST.W.S)l.ST.W1.Y

(8)

X2 = j=1

C,

i-I

and the matrix solution becomes

242

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

W now denotes a weight matrix in which all elements are equal to zero except the diagonal ones, which are equal to the weight values in vector W. The weight values can be chosen in many ways, but the optimal values can only be

found through an iterative procedure. A better procedure in this case is to make full use of the maximum likelihood technique,4'5'6 of which least squares fit is a subclass. The maximum likelihood technique gives the minimum vari-

ance solution. The principle of MLT applied to LS counting is shortly as follows: radioactive decay is governed by Poisson statistics and the counts y in channel j are Poisson distributed (y3 is the mean value). The probability p for detecting k counts in channel j is:

p=

(y1)kj

ei / k!

(9)

For a spectrum S the joint probability 1 is the product of p1 for all m channels: m

1= llp

j=1

Combining Equations 2 and 10 leads to the conclusion that the probability I is a function of the component spectra S1 and the count rates c1, or 1 = l(S,c1). In DOT, the shape of the spectra Si are fixed, but the count rates are iteratively determined so that 1 achieves a maximum value. The spectra S, are retrieved from the spectrum library as soon as the quench level of the sample has been

determined with the external standard. In the LS counter 1410 there is a spectrum library comprising spectra that cover a large quench region of both chemical and color quench for six common isotopes. This library can be used as such or extended with the users own calibrations or "fine-tunings."7 The influence of color on counting efficiencies and spectrum shape, and the Wallac color quench monitor, has been dealt with elsewhere by the present authors.8 EXPERIMENTAL

This report is an account of a comparison between the traditional two fixed window method and DOT, under similar conditions and in the same instrument. Organic tritium and '4C were used in this study. Two calibration standards with the pure isotopes were prepared for fine tuning the Wallac library. Unknown samples were prepared with the activities shown in Table 1. The scintillation liquid was OptiScint HiSafe, a high flashpoint cocktail. The samples were quenched with carbon tetrachloride so that the counting efficiency of tritium was around 23% and the counting efficiency of '4C around 86%. The samples were measured repeatedly 120 times and the counting time was one minute. All measurements were made on a typical 1410 LS counter. As this

instrument is not equipped with any traditional window methods, the cpm

MULTILABEL COUNTING USING DIGITAL OVERLAY TECHNIQUE

243

Table 1. Activity of the Two Isotopes of the Samples used in this Work. "dpmexc" is the

Activity of the Isotope in Excess and "dpmie" Is the Activity of the Other Isotope. "Ratio" is the Ratio Between These Two

dpmie

dpm.0

Ratio

95000 95000 95000 95000 95000 95000 95000

1500

64/1 32/1

3000 6000 12000 23800 47500 95000

16/1 8/1

4/1 2/1 1/1

values in five different windows for each isotope were printed out and converted to dpm off-line. RESULTS AND DISCUSSION

The aim of this test was to find experimental evidence for our theoretical assumption that DOT is at least as good as or even better than any window method. For this reason, the standard deviation of the 120 dpm values was computed. The results are presented in Figures 1 to 4 as relative standard deviation (rDev) as a function of the activity ratio. rDev is defined according to the equation:

rDev = Dev

/ DevDOT

2.2 2

5-100 1.8

U 5-150 5-200 - 5-250 -A-

64

32

16

8

4

2

5-300

1

Isotope ratio (C14:H3) Figure 1.

The relative standard deviation of dpm for 14C as a function of the activity ratio for a

number of low window limits. rDev is defined in the text.

244

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

3.50

3.00

5-100

D 5-150

2.50

4

rDev

5-200

2.00

- 5-250 1.50

-A-

5-300

1.00 64

32

16

8

4

2

Isotope ratio (C14:H3) Figure 2.

The relative standard deviation of dpm for tritium as a function of the activity ratio for a number of low window limits. rDev is defined in the text.

wherein Dev and DevDOT are equal to the standard deviations of the 120 dpm values for the two isotopes computed by using a windows method and DOT, respectively. In Figure 1, '4C is in excess, and rDev is computed for '4C while in Figure 2, '4C is also in excess, but rDev is now computed for tritium. Accordingly, in Figures 3 and 4, tritium is in excess and rDev is computed for '4C and tritium, respectively. These figures also mentioned the window limits of the 2.8 2.6

rDev

2.4

5-100

2.2

D 5-150

2

4 5-200

1.8

5-250

1.6 1.4

-A-

5-300

1 .2

1/64

1/32

1/16

1/8

1/4

1/2

Isotope ratio (C14:H3) Figure 3.

The relative standard deviation of dpm for 14C as a function of the activity ratio for a number of low window limits. rDev is defined in the text.

245

MULTILABEL COUNTING USING DIGITAL OVERLAY TECHNIQUE

2.00 1.80 -

.

5-100

1.60 -

o 5-150

rDev 1.40 -

5-200

1.20 -

5-250

-k 5-300

1.00 0.80 1/64

1/32

1/16

1/8

1/4

1/2

Isotope ratio (C14:H3) Figure 4.

The relative standard deviation of dpm for tritium as a function of the activity ratio for a number of low window limits. rDev is defined in the text.

lower (A) of the two windows A and B. Window B extends from the upper limit of A up to channel 650. The data in Figures 1, 2, and 4 states that, at the quench level assessed here, the narrow A window 5-150 gives a standard deviation almost as low standard deviation as DOT. However, Figure 3 shows that the window 5-150 is not the best when tritium is in excess. The data in these figures disclose one of the limitations of the traditional two-window method: there is no single window setting that gives as low standard deviations ad DOT for a wide range of activity ratios. One important conclusion that can be drawn from these figures is that selecting the correct limits for window A is very important when using two windows. If using an automatic method like moving windows, the unquenched windows must be set correctly and the user can do no more than hope that the program always chooses the best windows settings at different quench levels. With DOT, the user has no such worries, as the method always gives lower or equally low deviations as any window method. SUMMARY

The advantages of DOT are the following: Statistically, DOT is always better than or equally good as any combination of moving or automatic windows. The calibrations are of general characters and may be used in any connection, single label or multilabel.

246

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Counting is optimized for all conditions, such as varying quench levels or activity ratios. The method is similar in the general multilabel case as in the special case of dual label. The reference spectra are also used in a quality control manner to ensure that the unknown sample does indeed contain the assumed radioisotopes; they are also used to determine, in a qualitative manner, what radioisotopes are present in completely unknown samples.

REFERENCES

Nather, R.E., U.S. Pat. No. 4,029,401 (1977). van Cauter, G.C., L.J. Everett, and S.J. DeFilippis, European Patent AppI. No. 0 202 185 (1986).

OIler, W.L. and P. Plato, "Beta Spectrum Analysis: A new Method to Analyze Mixtures of Beta-Emitting Radionuclides by Liquid Scintillation Techniques," mt. J. App!. Rad. isot., 23:481-485 (1972). Orth, P.H.R., W.R. Falk, and G. Jones. "Use of the Maximum Likelihood Technique, for Fitting Counting Distributions," Nucl. Inst. and Meth., 65:301-306 (1968).

Ciampi, M., L. Daddi, and V. D'Angelo. "Fitting of Gaussians to Peaks by a Maximum Probability Method," Nuci. Inst. and Meth., 66:102-104 (1968). Awaya, T. "A New Method for Curve Fitting to the Data with Low statistics Not Using the x2-Method," Nuci. Inst. and Meth., 165:317-323 (1979). Kouru, H. "A New Quench Curve Fitting Procedure: Fine-Tuning of Spectrum Library," in New Trends in Liquid Scintillation Counting and Organic Scintillators, (1989).

Rundt, K., T. Oikari, and H. Kouru. "Quench Correction of Colored Samples in LSC," in New Trends in Liquid Scintillation Counting and Organic Scintillators, (1989).

CHAPTER 21

A New Quench Curve Fitting Procedure: Fine Tuning of a Spectrum Library

Heikki Kouru

ABSTRACT In liquid scintillation counting the counting efficiency and the spectrum of the sample depend on the quench level; quench calibration is needed to establish the true activity of the sample. The quench curve relates the counting efficiency to an appropriate quench indicating parameter. The quench curve is formed by measuring a number of reference samples with known activity and varying quench levels. Typically six to ten reference samples are required in quench calibration. This presentation shows that the number of reference samples can be reduced by using model based quench curve fitting. The model uses a large set of reference samples, which form the spectrum library of the isotope. The fine tuning of the spectrum library is the model adjustment procedure in which the spectrum library data is modified to fit to the data of the new calibration. The principles of the fine tuning are illustrated and a theoretical example is calculated. Measurements made by a liquid scintillation counter that uses the described principle, the Wallac 1410, are presented and a comparison with a conventional counter is made.

INTRODUCTION

The quench level of liquid scintillation counting samples is often variable. Quenching decreases the number of photons detected after one radioactive decay. In instruments the variation of quench level is seen as variations of the counting efficiency and shifts of the pulse height energy distribution from sample to sample of the same isotope. The reduction in the amount of photons shifts the pulse height energy distribution or the spectrum to the smaller pulses

and decreases the counting efficiency. Figure 1 shows three '4C spectra of different quench levels. To relate the disintegration rate of a sample to the count rate it is necessary to determine the quench level of the sample and correlate it to the counting

efficiency. There are different methods to estimate the quench level of a sample, but most of them rely on the effect of quenching on the position of the 247

248

Figure 1.

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Three 14C spectra of different quench levels.

spectrum on the amplitude scale. A common method uses an external gamma ray source or an external standard and measures the shift in pulse amplitudes of the spectrum resulting from Comptom electrons scattered by this external standard. The end point of an external standard spectrum (Sqp(E)) can be used as a quench indicating parameter. The relationship between a quench indicating parameter and the counting efficiency of a radionuclide is established by measuring reference or standard samples of known activity and expressing the counting efficiency of these samples as a function of the quench indicating parameter. The whole procedure is called quench calibration and the resulting function, which correlates counting efficiency and the quench indicating parameter, is called a quench curve. Figure 2 shows tritium quench curves for different scintillation solutions. Because typical quench curves are not linear but curved, normally six to ten reference samples have to be prepared and measured to make a quench calibration that covers the desired quench level range. Figure 2 shows that even if the quench curves differ from each other, the basic shape of the curves is the same. This is especially true for one radionuclide, but the same shape can be used as a good approximation for isotopes with energy of the same order of magnitude. Traditionally quench curves have been formed by fitting a polynomial or a spline function to the quench calibration data or by making a linear interpola-

FINE-TUNING OF A SPECTRUM LIBRARY

249

60

50

40

30

20

10

250

300

350

400

450

580

550

600

Figure 2. Tritium efficiency quench curves for different commercial scintillation solutions1.

tion between the data points. In this presentation a quench curve producing method, which uses the known form of the quench curve is introduced; the method is called model based quench curve fitting. This introduction is concerned with the efficiency quench curve, but the spectrum can be treated in an analogous way. The spectrum can be expressed as parametric form, and the parameter values can be expressed as a function of the quench indicating parameter. Another possible way to treat spectrum data is to normalize the spectrum and follow the variation of the spectrum proportion in each channel The preservation of the quench curve form can be realized in many ways. It is possible to develop a theoretical quench curve model with a few parameters and find the solution of these parameters for the particular quench data. This kind of procedure can preserve the right form of the curve if the number of free parameters is low enough. Another more straightforward possibility is to use an experimental model. The model adjustment should fulfill two demands: flexibility and stiffness. The procedure should be flexible enough to follow the natural variation and stiff enough to preserve the right shape of the curve. One

250

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

possible solution is to make a transformation of the variables, make the fit on them, and then return to the real values by inverse transformation. THEORETICAL EXAMPLE

As a theoretical example, we may consider a situation where the true efficiency quench curve is assumed to be a second degree polynomial. The quench

curve model is then also a second degree polynomial as shown in Figure 3. Figure 3 also shows the measured counting efficiency and sample quench index value pairs for two reference samples, which are assumed to have 4 and 5%

higher counting efficiency values than predicted from the model. Figure 4 shows the same model curve and reference data points after a transformation on the efficiency scale. Here the curve fitting is not made on efficiency values but on the relative deviations of efficiency from the model. A linear interpola-

tion is used as a fitting method. Figure 5 shows the corresponding fit in untransformed coordinates; it also shows a normal linear interpolation without transformation. The comparison between model based fitting and normal fitting is shown in Figure 6, where the relative dpm errors of both systems are given as a function of counting efficiency. The error in normal fitting is three times greater than the one using model based estimation. EXPERIMENTS

The Wallac 1410 counter uses a model based fitting principle for quench calibration. As a model, the 1410 has built in spectrum libraries for the most common isotopes. The spectrum library includes counting efficiencies, whole sample spectra and external standard spectra, external standard based total quench indication parameters, and external standard based color quench indication parameters of reference samples. The data is retrieved from the spectrum library using both quench parameter as keys. The quench calibration in the 1410 effects both counting efficiency and spectra, and it is called fine tuning of the spectrum library. A tritium quench series of nine samples was made using Hisafe liquid scintillation cocktail and nitro methane as quenching agent. Using two samples of the series as references, a quench calibration has been made in the 1410 and a RackBeta 1219. In all measurements 0.5°lo one sigma counting statistics were used. RackBeta 1219 uses linear interpolation, as only two reference samples are used. Figure 7 shows the efficiency quench curve made in the 1410. Figure 8 shows the relative error of activities in the quench series samples, which have been measured as unknown samples. The systematic errors are one order of magnitude lower in System 1410 when using the fine tuning principle. To illustrate the robustness of the 1410 fine tuning principle an experiment was conducted, where the spectrum library of another isotope is used as a

FINE-TUNING OF A SPECTRUM LIBRARY

251

60 50 40

-_ True

Eff[%] 30-

- Mod Ref

20 10 0

500

600

700

800

900

Quench parameter Figure 3.

Mod is a tritium efficiency quench curve model, Ref is used for measured quench parametercounting efficiency values of reference samplesand True is the real efficiency quench curve corresponding to the reference samples.

1.7 1.6

1.5 -

True

trEff 1.4 -

- MFit Ref

1.3 -

1.2 1.1

500

600

700

800

900

Quench parameter Figure 4.

Transformed efficiency quench curves. The values of trEff-axes are calculated by dividing the efficiency with the efficiency of the model curve of the corresponding quench parameter value. MFit is the model based quench curve in this coordinate system. This quench curve was made using a linear interpolation of the efficiency quench curve in this coordinate system. Symbols True and Ref are defined in Figure 3.

252

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

60

/

55 50 45

- MFit

40 35 Eff[%} 30 25 20

- LFu

- True Ref

15

10 5

I

500

550

600

650

I

700

750

800

I

850

900

Quench parameter Figure 5. LFit is a quench curve made using a linear interpolation without the model. Symbols MFit, True and Ref are defined in Figures 3 and 4.

model in quench calibration. A quench series of nine '4C samples was made using Hisafe cocktail and nitro methane as a quenching agent. Quench was calibrated in the 1410 using three "C samples, having the 45Ca spectrum library as a model. Figure 9 shows the efficiency quench curve. All nine samples were measured as unknowns, and the dpm errors are shown as a function of sample 30 20

DpmE[%]

10

- MFit

0

- LFit

Ref

-10 -20 -30 0

10

20

30

40

50

60

Counting efficiency [%] Figure 6.

Relative DPM errors, DpmE as a function of counting efficiency for the model based quench curve and for the quench curve not based on a model.

FINE-TUNING OF A SPECTRUM LIBRARY

253

o.o -

-fi

C

4-

I

I

I

I

I

I

I

I

I

I

700

600

800

SQP(E

Figure 7. Tritium efficiency quench curve of the 1410 using two reference samples.

15.00 10.00 5.00 0.00 DpmE [%]

-5.00

Cony

-10.00 --15.00 --20.00 -

Li

1410

-25.00 -30.00

0.00 10.00 20.00 30.00 40.00 50.00 60.00 H-3 efficiency [%] Figure 8. DPM error, DpmE for 3H samples as a function of the counting efficiency of the samples when two reference samples were used in quench calibration. The counting efficiencies of the reference samples were 6.7% and 47.3%. 1410 is the 1410 and Cony is the 1219.

254

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

I

I

?00

600

I

800

SQP(E)

Figure 9. Carbon efficiency quench curve of the 1410 using three reference samples and 45Ca spectrum library.

2 1

0

DpmE[%il -2 -3

-4 -5 6

I

55

60

65

70

I

75

I

80

I

I

85

90

95

C-14 efficiency [%] Figure 10.

DPM error, DpmE for 14C samples as a function of the counting efficiency of the samples when three reference samples and 45Ca library were used in quench calibration. The counting efficiencies of the reference samples were 93%, 84%, and 65%.

FINE-TUNING OF A SPECTRUM LIBRARY

255

quench level in Figure 10. In every case the interpolation error is under 2, hence a model could also be selected from a spectrum library of some other isotope. Of course, the library of a radionuclide with a similar energy would be preferred. CONCLUSIONS

Using the model based quench calibration or spectrum library fine tuning considerably increases the accuracy of the results when only a few reference samples are used. REFERENCES

1. Rundt, K. "On the Determination and Compensation of Quench in Liquid Scintillation Counting," PhD Thesis, Abo Akademi, Turku, Finland (1989).

CHAPTER 22

The Effect on Quench Curve Shape of the Solvent and Quencher in a Liquid Scintillation Counter

Kenneth Rundt

ABSTRACT When preparing quench calibration curves, commercial sealed standards are often used, although it is advisable to use the same composition, scintillation liquid, and sample as the unknowns. Commercial sealed standards usually contain a toluene or xylene based cocktail and carbon tetrachloride or nitromethane as quencher. However, the unknown samples may contain another solvent and other chemical quenchers. The solvent has become more significant with the introduction of new "safe" solvents that behave differently from the traditional toluene solvent. As reported in earlier literature, different solvents and chemical quenchers lead to different quench curves, but no physical explanation has been given. In this work, quench curves for several commercial cocktails with chemical quenchers and one color quencher have been measured. The counting efficiency and the mean pulse height of the isotope (3H) were recorded for different amounts of quencher and with varying coincidence resolving time of the counter. The differences in quenching characteristics can be explained as differences in the scintillation pulse shapes. The shape of the scintillation pulse is primarily dependent on the way the excitation energy is deposited in the solvent and on how the chemical quenchers act on the excited solvent molecules. The data indicate that there are mainly two different chemical quenching mechanisms: quenching of the higher excited states, which are precursors of the triplet states, and quenching of the lower singlet states.

INTRODUCTION

As liquid scintillation counting is frequently used for quantitative analytical measurements of low energy isotopes, such as tritium and 4C, it is extremely important to know the counting efficiency of the samples. For that purpose, most counters today have automatic methods for determining the quench level and the counting efficiency. Generally, an empirical relationship between the counting efficiency and the quench index must be established at first, e.g., by measuring a set of quenched standards. A question that often arises is how to prepare this set of quenched standards; which scintillation cocktail and which 257

258

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

quencher substance should be used in the standards? The usual advise to users who want to suppress the risk of systematic errors is to use the same composi-

tion in the standards as in the unknowns. This means using the same vials, same scintillation cocktail (same volume), and same quencher. Using the same

vials is mostly not a problem, but the two other factors are worth some consideration. During the last few years the number of new solvents used in liquid scintilla-

tion cocktails has increased remarkably through the introduction of high flashpoint solvents. Typically these solvents have a flashpoint above 60° C. The purpose of these solvents is to make scintillation counting less hazardous.

The most common scintillation liquids today are still based on toluene or xylene, both of which are rendered as toxic and hazardous chemicals. The new solvents not only have a low vapor pressure, they are generally considered less toxic than toluene. In a laboratory, cocktails based on both traditional solvents

and on the new safe solvents may be in use at the same time. Commercial sealed calibration standards usually contain a traditional solvent. Because of this, a natural question that arises is whether or not quench curves based on traditional cocktails may be used together with these new safe cocktails. In

current literature, there is very little or no information available on the quenching behavior of these new solvents as compared to the traditional ones. There are also remarkably few reports in current literature comparing quench

curves from different quenchers. Wunderly' has investigated a number of substances in relation to the sample channels ratio (SCR) and the H-number used in Beckman Instruments counters as quench level index. The H-number is equal to the shift of the '37Cs compton edge as compared to the unquenched position. Wunderly found differences between the assessed quenchers in terms

of counting efficiency vs. SCR or H-number. He could not explain this nor could he propose any solution on how to avoid it. It is also well known at Wallac Oy that users of the Wallac Rackbeta LS counters have encountered this problem without being able to explain it or reduce the errors involved. Furthermore, the new safe cocktails introduce one more source of uncertainty. The present work has been undertaken in order to cast some light on these problems and show how the errors can be minimized by improving the LS counter. The aim of this work has also been to increase the general knowledge of the physics behind liquid scintillation and quenching. EXPERIMENTAL

Two different measurements were performed. At first, quench curves for five commercial cocktails were recorded. The cocktails were: xylene based Lipoluma from Lumac AG, Netherlands, toluene based OptiScint 'T', pseudo cumene based OptiScint Safe, di-isopropylnaphthalene based OptiScint

HiSafe, and OptiPhase HiSafe II (containing emulsifiers, but no added water).The last four are trademarks of Wallac Oy. The isotope spectra were

EFFECT ON QUENCH CURVE SHAPE OF SOLVENT AND QUENCHER

259

recorded at three coincidence resolving times (15, 72, and 800 nsec). The isotope mean pulse height of the isotope spectrum, SQP(I), was used as the quench index. The quenching agent was carbon tetrachloride. In the second set of measurements, the quench behavior of eight quenchers was recorded using the toluene based cocktail (OptiScint 'T', Wallac Oy). The uncolored substances were: nitromethane, carbon tetrachloride, methanol, acetone, acetophenone, dibutylamine, and methyl benzoate. One colored substance, the commercial yellow dye Sudan 1, was also used. All quenchers were of pro analysi quality. A dilution series was also made by diluting the OptiScint 'T' with pure toluene, resulting in a decreased concentration of the fluors. Normal 20 mL glass vials where used, with 10 mL of scintillation liquid. All

samples were in equilibrium with air; thus containing the same amount of oxygen-quench. Wallac Internal Standard Capsules 1210-120, containing tntium labeled cholesterol, were used for dispensing the activity. All measurements where done on a prototype of the new 1410 LS counter from Wallac Oy, which has an adjustable coincidence resolving time. RESULTS

The results of the solvent measurement are shown in Figures 1-3. Figure la shows the quench curves for all the cocktails at normal coincidence resolving time (15 nsec), and Figure lb shows the relative error in DPM when the toluene based calibration curve was used as a reference for all cocktails. The DPM error is defined according to the equation DPMErr = 100

(DPMO - DPMJ) / DPM1

wherein DPM1 is the activity predicted by the toluene calibration qurve and DPM0 is the true activity. Figures 2 and 3 show the DPM errors at longer coincidence resolving times. A quencher other than CCI4 may lead to quench curve behavior different than that described above. The next figures show the influence of quencher on the quench curve shape in a toluene based cocktail. Figure 4a shows a plot of all nine quench curves for tritium in the case of normal coincidence resolving time, while Figure 4b shows the same data plotted as relative error in DPM when a CC14-based calibration curve is used as a reference. The curves in Figures 4a and 4b show that at normal coincidence resolving time, and in the case of tnitium, the nine quenchers can be divided into two groups. One group comprises methanol, carbon tetrachloride, nitromethane, and dibutylamine situated above, and another group, comprises the toluene dilution, acetone, methyl benzoate, acetophenone, and Sudan 1 situated below. The toluene dilution shows the greatest deviation. From these curves one can see that

nitromethane or carbon tetrachloride, which are often used for preparing quenched standards, may not be the correct choice when the unknown samples contain ketones like acetone or acetophenone.

260

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

60 50 .

40 30

o

20

t10 0

I

90

100

110

120

130

I

140

150

160

170

180

SQP(I) Figure la. Tritium quench curves (counting efficiency as a function of quench index) for the 5 different cocktails and a coincidence resolving time of 15 nsec.

The relative error for some of the quenchers depends very much on the coincidence resolving time. Figures 5 and 6 show the relative DPM error as a function of SQP(L) for two more resolving times, 40 and 72 nsec. At 72 nsec all of the curves join except the color curve and the toluene dilution. The color curve does not in fact depend very much on the coincidence resolving time. In normal commercial instruments the coincidence resolving time has to be quite short ( 15 nsec) in order to decrease the number of random coinci-

dences from, e.g., chemiluminescence. As can be seen from Figure ib, in single label counting, the relative error in the DPM will be a constant value independent of quench level when measuring unknown samples containing other solvents than toluene against a toluene based calibration curve in a normal instrument using SQP(I) as the quench index. This can be satisfactory if the absolute activity is of minor interest and the results are used for comparison only. With different quenchers this is not so as the error varies with the quench level from 0 to -40% for a sample quenched with a ketone measured against a CC14 calibration curve. When measuring dual labeled samples, the errors in both cases will be even bigger, as the shape of the isotopic spectra differ and the spill over curves look completely different. Figures 7a and 7b

EFFECT ON QUENCH CURVE SHAPE OF SOLVENT AND QUENCHER

-20

10

0

20

30

40

50

60

261

70

Counting Efficiency 1%] Figure lb. The DPM error as a function of the tritium counting efficiency when using the toluene based quench curve (OptiScint 'T') for computing DPM values. The sym-

bols are: 0 = OptiScint T', = Lipoluma, A = OptiScint Safe, V = OptiScint HiSafe, V = OptiPhase HiSafe II.

shows the influence of coincidence resolving time on spectral shape for 1. an OptiScint 'T' sample and 2. an OptiScint HiSafe sample. DISCUSSION

From the curves in Figures 1-6 one general conclusion can be drawn: the shorter the coincidence resolving time, the bigger the difference between the various solvents and quenchers. Why then is the coincidence resolving time critical to the quench curve behavior? The lowest excited singlet state of the aromatic solvent and most fluors have quite short lifetimes and will result in prompt light emission, a few nanoseconds after the disintegration. Triplet states, superexcited states, and ionized states have, in comparison to the singlet states, long lifetimes. These states may either relaxate, through internal conversion, to the ground singlet state and liberate energy in the form of heat, or they may end up in an excited singlet state which can lead to light emission several hundreds of nanoseconds after the disintegration. The first light burst

is usually called the prompt pulse, and the second emission is called the

262

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

10

20

30

40

50

60

70

Counting Efficiency [%J

Figure 2. The DPM error as a function of the tritium counting efficiency when using the toluene based quench curve and a coincidence resolving time of 72 nsec. The symbols are the same as in Figure 1

p. Counting Efficiency 1%] Figure 3.

The DPM error as a function of the tritium counting efficiency when using the toluene based quench curve and a coincidence resolving time of 800 nsec. The symbols are the same as in Figure 1.

EFFECT ON QUENCH CURVE SHAPE OF SOLVENT AND QUENCHER

263

60

50

40

30

0

o

20

t10 0

100

110

120

130

140

150

160

170

SQP(I)

Figure 4a.

Tritium quench curves for OptiScint 'T' and 9 different quenchers and a coincidence resolving time of 15 nsec.

20

0

-20

-40

-60

0

10

20

30

40

50

60

* Counting Efficiency 1%) Figure 4b. The DPM error as a function of the tritium counting efficiency when using the carbon tetrachioride quench curve for computing DPM values. The symbols are: o methanol, 0 = carbon tetrachloride, A = nitromethane, V dibutylamine, = toluene dilution,S = acetone, V = methyl benzoate, = acetophenone and 0 = Sudan 1 (a yellow dye).

264

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

20

10

30

40

50

60

-+ Counting Efficiency 1%] Figure 5. The DPM error as a function of the tritium counting efficiency when using the carbon tetrachloride quench curve and a coincidence resolving time of 40 nsec. Symbols as in Figure 4.

20

10

-P Figure 6.

30

40

50

60

Counting Efficiency 1%1

The DPM error as a function of the tritium counting efficiency when using the carbon tetrachloride quench curve and a coincidence resolving time of 72 nsec. Symbols as in Figure 4.

EFFECT ON QUENCH CURVE SHAPE OF SOLVENT AND QUENCHER

265

1000

800

600

400

200

0 0

50

100

150

200

250

300

350

-# Channel # Figure 7a.

The tritium spectrum of an air-quenched OptiScint T' sample at two different coincidence resolving times.

50

100

150

200

250

300

350

-P Channel # Figure 7b.

The tritium spectrum of an air quenched OptiScint HiSafe sample at two different coincidence resolving times.

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

266

delayed pulse (see, e.g., Fuchs, et al.2'3). When the coincidence resolving time is so short that a noticeable part of the delayed component is situated after the coincidence analyzer closes, one would expect to find mainly a reduced number of small pulses comprising only a few photoelectrons. The reason for this

is as follows: when the number of photoelectrons produced decreases, the spacing in time between the photoelectrons increases if the overall pulse length remains constant (compare randomly dividing a distance into ten parts or two parts). Thus the chance increases that the time delay between the first electron,

starting the coincidence analyzer, and the second electron, causing a valid coincidence pulse, will be longer than the coincidence resolving time. Hence, for a sample at fixed quench level, as the number of small pulses is decreased, the mean pulse height is shifted upwards and the counting efficiency downwards. This behavior can be observed in all the data presented here and especially in Figure 7. Solvents As all investigated cocktails may be expected to contain principally the same fluors, like PPO and bis-MSB, the reason for the difference in quench curve shape lies in the behavior of the solvents. It is known that naphthalene crystals have very long scintillation pulses4 and that solvents based on naphthalene derivatives have longer pulses than mono-aromatics.5 There may be three reaSons for the longer pulse: The electron may cause more extensive ionization in the naphthalene based solvent than in xylene. Ionization leads to formation of triplets through ion and electron recombination. The triplet states may eventually form singlet states through the triplet-triplet annihilation process. Longer lifetimes of the excited singlet states may cause longer pulse decay times. The bulkier molecule may slow the diffusion rates of the excited molecules.

Even at the longest coincidence resolving time (800 nsec), there is still some difference between the curves. But this difference may be due to other factors than the pulse shape. In general, with "infinitely" long coincidence resolving time, both the counting efficiency (E) and the mean pulse height (F) may be considered as functions of three parameters:6

E = E(a,b,X) and F = F(a,b,X) The parameter a is equal to the scintillation efficiency of the liquid, and thus depends on both the solvent and the concentration of quenching agent, while

the parameters b and X are dependent on solvent only. The parameter b reflects the degree excited states are quenched by ionized states in the vicinity,

while X, the mean excitation potential, has a direct effect on the stopping power of the solvent, and thus also on the density of ionized and excited states.

EFFECT ON QUENCH CURVE SHAPE OF SOLVENT AND QUENCHER

267

Variations in the mean excitation potential X and the ionization quenching parameter b result in variations of the quench curve shape. Quenchers

What then is the explanation for the behaviors depicted in Figures 4-6? The answer might lie in the relation between the energy levels of a certain quencher and the solvent (toluene). This phenomenon has been extensively studied by,

e.g., Fuchs, et. al.,23 who have proposed the theory that certain quenchers, which can be classified as "electron scavengers" (like oxygen and carbon tetrachloride) and have their excited levels above the energy levels of the solvent, may interact with the higher energy states of the solvent, and thus cause a reduced relative contribution of the delayed component. Compounds having their singlet energy levels just below the corresponding energy levels of the solvent may interact by accepting the excitation energy from the singlet states of the solvent and decrease the intensity of both the prompt component

and the delayed component. The different forms of "impurity" or chemical quenching has also been discussed extensively by Birks.7

The fact that the color quench curve deviates from all the others, at all coincidence resolving times, is related to the special characteristics of color quench.8 Apart from the other quench modes, color quench depends on the spatial coordinates of the disintegrations. Sudan 1 also has a small amount of chemical quench, otherwise it would behave exactly like the toluene dilution, but the main amount of quenching comes from the color. The results show that color quench can not be handled the same as other forms, just by prolonging the coincidence resolving time. CONCLUSION

Different quench curves are produced by different solvents and quenchers, when using a quench index reflecting the position of the pulse height spectrum of the dissolved isotope (as the mean pulse height or samples channels ratio), this can be explained on the basis of the shape of the scintillation pulse. The pulse shape is mainly dependent on two factors: 1. the distribution of different excited states of the solvent molecules when a fl-electron traverses the solution and 2. the probabilty of quenching certain excited states in the solution. If the probabilty of forming higher excited states (e.g., Rydberg states) or ionized states is large, then the scintillation pulse will be streched out in time. Quenchers capable of quenching the triplet and higher excited states, combining with

ionized states, or acting as electron scavengers very effectively quench the delayed component of the pulse. Substances capable of quenching only the lowest singlet levels however, quench both the prompt and the delayed component to an equal degree. The differences in the quench curves caused by this

effect can be reduced by increasing the coincidence resolving time of the counter to above 40 nsec.

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

268

REFERENCES

Wunderly, S. W. "Evaluation of Liquid Scintillation Counting Accuracy: Effect of Various Chemical Quenching Agents and Effect of Milky Sample Preparations," in International Conference on Advances in Scintillation Counting, S. A. McQuarrie, C. Ediss, and L. I. Wiebe, Eds. (Edmonton, Canada University of Alberta, 1983), pp. 376-386. Fuchs, C., F. Heisel, R. Voltz, and A. Coche. "Scintillation Decay and Absolute Efficiencies in Organic Liquid Scintillators," in Organic Scintillators and Liquid Scintillation Counting, D. L. Horrocks and C.-T. Peng, Eds. (New York: Academic Press, 1971), pp. 171-186. Fuchs, C., F. Heisel, and R. Voltz. "Formation of Excited Singlet States in Irradiated Aromatic Liquids," J. Phys. Chem., 76(25):3867-3875 (1972). Birks, J. B. and R. W. Pringle. "Organic Scintillators with Improved Timing Characteristics," Proc. Roy. Soc. Edinburg (A), 70:233-244 (1971/72). Koike, Y. "Measurement of Decay Times of Loaded Liquid Scintillators," Nuci. Instr. Meth., 97:443-447 (1971).

Rundt, K., H. Kouru, and T. Oikari. "Theoretical Expressions for the Counting Efficiency and Pulse Height Distribution Obtained for a Beta-Emitting Isotope by a Liquid Scintillation Counter," in International Conference on Advances in Scintilla-

tion Counting, S. A. McQuarrie, C. Ediss, and L. I. Wiebe, Eds. (Edmonton, Canada: University of Alberta, 1983), pp. 30-42. Birks, J. "Impurity Quenching of Organic Liquid Scintillators," in Liquid Scintillation Counting, Vol 4, M. A. Crook and P. Johnson, Eds. (London: Heyden & Son Ltd, 1977), pp 3-14. Rundt, K., T. Oikari, and H. Kouru. "Quench Correction of Colored Samples in LSC," in New Trends in Liquid Scintillation Counting and Organic Scintillators, 1989.

CHAPTER 23

67Ga Double Spectral Index Plots and Their Applications to Quench Correction of Mixed Quench Samples

in Liquid Scintillation Counting Shou-Ii Yang, Itsuo Yamamoto, Satoko Yamamura, Junji Konishi, Ming Hu, and Kanji Torizuka

INTRODUCTION

To prepare samples suitable for liquid scintillation counting various materials are added to scintillator solutions. These materials and the sample itself will alter the counting efficiency to some degree; this effect is known as quenching.

Chemical quenching arises in an energy transfer process prior to creating luminescence, whereas color quenching is a phenomenon occurring after the luminescence production. Color quenching effect diminishes the mean free path of the fluorescence photons. It is emphasized that chemical quenching is really a physical effect rather than a chemical effect.' The pulse height distribution of a color quenched sample differs from that of a chemically quenched sample, even if both samples have identical radioactivity and counting effi-

ciencies. Therefore, chemical and color quench correction curves are also different. For strongly quenched samples there is a large difference between the two kinds of quench correction curves.2 Significant errors can be encountered if a quench correction curve obtained from chemically quenched standards is applied to color quenched samples. Thus, it is advantageous to determine the type of quench in a sample before quench correction is applied. Let us discuss some mixed quench samples where quenching is produced by the mixed solutions of a given chemical quencher (Cd4) and a color quenching solution at different ratios. Since the counting efficiency of the sample with known radioactivity (dpm) can be calculated, its quenching type is easily determined according to the location on the chemical and color quench correction curves. This method cannot be used, however, if sample activity is unknown. Therefore, a new parameter, quench type contribution factor (QTCF), was

established to denote the degree of contribution from chemical or color 269

270

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

quenching in a mixed quenched sample.34 The mixed quenched sample could

be corrected automatically by means of the chemical and color quenched standard samples, and QTCF value, and an on-line computer. Some authors have used the double ratio technique to determine the quench type.5 This method was first used by Bush in 1968 to detect adsorption or precipitation of radioactive material from liquid scintillation solutions.6 Earlier, an isolated

internal standard method7 and four curve method8 were also tried for this purpose. This paper describes a method called 67Ga Double Spectral Index Plots, which can determine quench type of unknown samples and correct the quenching of mixed quench samples. A novel report about using 67Ga as an internal standard in liquid scintillation counting was presented by McQuarrie and Noujaim in 1983.° Gallium is an element in Group lila of the periodic table. The

physical half life of the radionuclide 67Ga is 78 hours with major gamma emissions at 93 (40%), 184 (24%), and 296(22%) keV. A variety of radiophar-

maceuticals have been developed in the past two decades in an attempt to obtain tracers for tumor imaging, among these 67Ga-citrate has found widespread use'° thus, 67Ga can be obtained easily from clinical departments for use in liquid scintillation counting. MATERIALS AND COUNTING CONDITIONS scintillator solution: aquasol-2, a commercially available xylene based cocktail (New England Nuclear) chemical and color quenchers: carbon tetrachloride (CC!4), saturated solution of methyl red (reagent grade) in acetone (reagent grade), a specially prepared reagent. All quenchers and their solvents, products of Nakarai Chemicals Ltd. (Japan) labeled compounds: 67Ga-citrate injection (Medi-physics Co. Japan) as 67Ga

source, Thymidine (6-3H), > 15 Ci/mmol, aqueous solution (New England Nuclear), n-(1,2(n)-3H)hexadecane (TRR-6) and n-(1-14C)hexadecane (CFR-6),

manufactured by Amersham International, used as standards in preparations of 3H and 14C quenched series

instrument and counting conditions: Packard Tri-Carb (Model No.460CD) liquid scintillation spectrometer; spectral index (SIS and SIE), sample channels ratio (SCR), and spectra, as counting conditions used as listed in Table 1

All counting was done at 16°C and all samples were contained in 20 mL standard glass vials in which 10 mL cocktail was added. EXPERIMENTS AND RESULTS

Quenched 67Ga Spectra

A series of quenched samples containing the same amount of 67Ga-citrate (5 L aqueous solution) and varying amounts of carbon tetrachloride (chemical

APPLICATIONS OF 67Ga DOUBLE SPECTRAL INDEX PLOTS

271

Table 1. The Selection of Counting Conditions Spectra (key) Energy Range (key) 3H

0-19

Window Width

Range

1

0-20

counts (1 0-1 56) key)

1

0-20

counts(0l56keV)

5

21-160

counts (0-19 key)

1

0-20

counts(0l56keV)

5

21-156

SCR

counts (4-19) key) counts (0-19 key)

0-156

67Ga

0-3,5,7,10,19,156 0-variable upper limit 10-100,15-100

quencher) were prepared and tested. Increasing the amounts of the quencher caused a shift of the observed 67Ga pulse height spectrum in the direction of lower energy (Figure 1). These spectra of 67Ga are characterized by two peaks, i.e., lower and higher energy peaks corresponding to Auger electrons at 8 keY and conversion electrons from the 90 keY level, respectively. Though the spectral shift of both lower and higher energy peaks is concentration dependent, the counts at the lower energy peak were sequentially reduced by increasing the

quencher. In contrast, the counts at the higher peak were sequentially increased. This response is similar to the spectral shift of 3H and '4C quenched series (Figure 2). The energies of two peaks of 67Ga are closely related to the average beta decay energies of 3H and '4C, thus, a single or dual labeled sample containing 3H and/or '4C would behave similarly to the corresponding peaks of 67Ga at the same level of sample quench. McQuarrie and Noujaim9 developed a technique using parameters related to the pulse height and number of events in each peak of 67Ga. With it they may obtain a mathematical relationship that monitors the degree of sample quench.

Spectral Index of 67Ga Spectral index of samples (SIS) and spectral index of external standard (SIE) are two of the current quench monitoring techniques. The spectral index is a unique number for the given total spectrum and is related to the radionuclide and its level of quenching. If only a part of the spectrum is analyzed, however, this number will not be unique and will depend on the counting conditions. This last point has not attracted sufficient attention until now. Variations of SIS in an unquenched 67Ga sample for various pulse height ranges (i.e., energy range) are shown in Figure 3. With a region of interest set between a lower limit of 0 keV and an upper limit of 10 to 30 keY, a plateau is seen on the curve; if the upper range is greater than 60 keY, a second plateau is seen. Both lower and higher energy plateaus are related to two peaks of 67Ga pulse height spectrum. The 67Ga SIS values determined in the same region of interest will vary as the quenching increases (see Figure 4).

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

272

3.0 2.5

2.0

1.5

E

0 1.0

0.5

2

3

5

10

2030 50

100

Pulse height (keV) Figure 1. The unquenched and quenched (by carbon tetrachloride) spectra of 67Ga.

Double Spectral Index Plots for 67Ga The plot of SIS (0 to iOO keY) vs SLE for chemically quenched 67Ga is shown

in Figure 5. The differences between the shape of chemical vs color quenched spectra of 67Ga were also obtained (Figure 6); thus, an isolative curve from a chemical quenching curve was obtained for color quenched sample series that contained the same amounts of 67Ga-citrate (5 iL aqueous solution) and varying amounts of color quenching solution (from 0 to 20 jzL). A plot of SIS in the lower energy region vs that of SIS in the higher energy region can be obtained when two regions of interest are used to determine 67Ga SIS values for both chemical and color quenched series, the series must include the lower and higher energy peaks of 67Ga (Figure 7a). Because the peaks of the color quenched sample were lower and broader in the higher energy side than

APPLICATIONS OF 67Ga DOUBLE SPECTRAL INDEX PLOTS

273

CCI4Q.iI)

6000

0

10

5000

20

4000

120

F g- 3000 2000 1000

1

Figure 2.

23 5

203050 Pulse height (keV) 10

100

The unquenched and quenched (by carbon tetrachloride) beta spectra of 3H and

the peaks of a chemically quenched sample for 67Ga (Figure 4), in the same way as in Figure 5, two separate curves were given for chemical and color quenched

series (Figure 7a). There is an inflection point on each curve and a crossover point at moderate quenching. When the regions of interest were changed to 7 to 100 keY and 0 to 19 keY, new curves were obtained (Figure 7b) in which there are two inflection points on each curve and a crossover point at a higher level of quench. Mixed Quench Zone Let us analyze the region between the chemical and color quenched curves in

Figures 5, 7a, and 7b. Obviously, these are transitional regions in which samples will simultaneously contain both chemical and color quench; they are referred to as a mixed quench zone. In fact, we prepared samples containing

274

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

60

50

0

30

(0

20

0

0

2

3

5

10

20 30

50

100 200300 500

Upper limit of counting channel (keV) Figure 3. The variation of SIS of unquenched 67Ga sample in various pulse height ranges.

varying amounts of chemical quenching agent (carbon tetrachioride) and color quenching agent (saturated solution of methyl red in acetone) (Table 2) and measured them at the counting conditions of Figures 5, 7a, and 7b. These points of mixed quench samples fall in a zone; the positions in the zone are generally near chemical or color quenching curves, according to their ratio of chemical to color quenching. Similar phenomena have been observed and analyzed in the quench correction curves of sample channels ratio (SCR) and SIS, i.e., the curves of counting efficiency vs SCR or SIS. Thus, we can identify the quench type of an unknown sample and calculate the degree of contribution of chemical or color quenching according to its location on the double spectral index plot (Figures 5, 7a, and 7b.) For example, if the quenching agent of a sample is not known, 67Ga is added to the sample and measured at the above counting condition. A point in double spectral index plots can be obtained from the observed SIS and/or SIE values of the unknown sample.

APPLICATIONS OF 67Ga DOUBLE SPECTRAL INDEX PLOTS

275

60

10-100, 50 15-1 00,

0-156 keV

0 (0

30

2020 0-19 keV

0-5

0

0-3 keV

0

0

I

I

I

I

I

I

50

100

150

200

250

300

-

Quantity of chemical quencher CCI4(p.l) Figure 4. The variations of 67Ga SIS with increasing quench.

The sample was quenched by chemical quenching when the point is on the chemical quench curve and, in contrast, by color quenching when the point was on the color quench curve. Those samples corresponding to points lying in

a mixed quench zone can be considered to contain both chemical and color quenchers simultaneously. The mixed quench zone was a function of counting conditions in the same way as double spectral index plots. Attention should be paid to the points located near the crossover. Two points [(40,2) and (50,2)] fall outside of the mixed quench zone in Figure 7a, and one point (30,6) is on the color quenched curve in Figure 7b. When counting conditions were changed the mixed quench zone also changed, and these points fell in the new mixed quench zone. Generally speaking, these double spectral index plots are not sensitive to the degree of quench near their crossover point.

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

276

(0,0)

60 (10,0)

>

0 0 (0,2)

(500)

150

(20,2)

(0,4)

(30,2) (40,2)

(50,2) (0,7)

(100,0) (50,2)

(30,4) 'j' (60,2) (0,10)

(20,7)

N60,4 -'

Chemical quenched Colour quenched

(30,6) k40'4'

(50,5) (50,6) (0,15)

*

20,5)

(50,4)



!

Mix-quenched Mix-quenched (added 3H)

' (100,2)

(60,6)

(200,0)

(0,20) (50,7)

20 100

(100,5)

400

500

600

700

800

SIE Figure 5.

Double spectral index plots of 67Ga SIS in the channel of 15 to 100 keV vs SIE. The

figures of the first and second groups in the parentheses represent the amounts (pL) of chemical quencher (CCI4) and color quencher (saturated solution of methyl red in acetone).

Reproducibility of the Point Positions for Mixed Quench Samples We have given two kinds of double spectral index plots for 67Ga, i.e., SIS vs

SIE, and SIS in a lower energy region of interest vs SIS in a higher energy region of interest. Generally, these two kinds of curves also can be used as a tool to identify the quenching type of an unknown sample. A more important

APPLICATIONS OF 67Ga DOUBLE SPECTRAL INDEX PLOTS

277

2.5

2.0

Colour quenched

Chemical quenched

©1.5 E

0 0.5

Chemical quenched Colour quenched 1

23

5

10

203050

100

Pulse height (keV) Figure 6.

The differences of the shapes of 67Ga pulse height spectra for chemical and color quenching samples containing the same degree of quenching (relative counting efficiency was 68%).

problem for mixed quench samples is the reproducibility of the position in the mixed quench zone, because the zone is smaller than the total region of quench variation. The main factors affecting the reproducibility of this point are: (1) variation of 67Ga activity, (2) interference from 3H and/or '4C, and (3) selection and reproducibility of counting conditions. The selection of counting conditions

has been discussed under "Spectral Index of 67Ga." The reproducibility of counting conditions in this instrument are satisfactory as they are set automatically. To observe the effect of 67Ga radioactivity on the reproducibility of SIS, a series of quenched samples with three different quantities of 67Ga (5,10, and 20 L) were prepared and measured repeatedly. The standard deviations of measurements caused by the variation of 67Ga activity were (0.60 ± 0.20)% (mean ± S.D.) for the region of interest of 0 to 7 keY and (1.14 ± 0.66)% for the region of interest 15 to 100 keV which were similar to the standard devia-

tion for 10 measurements of a single sample were similar (0.80 ± 0.26)%. There was no significant difference between either. Changing the quantity of 67Ga from 5 to 20 L affected the reproducibility of SIE 2.1 ± 0.8%. If unknown samples containing 3H and/or '4C only make these points shift a

278

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

*

Chemical quenched

(0,0)

Colour quenched

Mix-quenched

a

(10,0)

50

(0,2)

(50,0) (0,4)

(20,2)

(30,2)

£ (40,2) (100.0)

(50,2) (60,2) (30,4)

a (50,2) (0,7) (20,5)

(50,4)

(40,4)

(0,10)

(30,6)

(40,6)

(150,0>

a(50,6)

(100,2)

30

(200,0)

(62,6) a (20,7)

(0,15)

(50,7)

£

a

(100,5)

(300,0)

20 10

11

12

13

14

15

67Ga SIS in the channel of 0-7 keV Figure 7a.

67Ga double spectral index plots of SIS in the channels of 15-100 keV to 0-7 keV. The illustration of the figures in the parentheses is shown in Figure 5.

APPLICATIONS OF 67Ga DOUBLE SPECTRAL INDEX PLOTS

*

(0,0)

60

279

Chemical quenched Colour quenched

(100)

Mix-quenched 1

Mix-quenched (added

-1)

(0,2) (50,0) (20,2)

0,4)

(0,20)

11111

20 15

(300,0)

20

II

liii

(100,5) (50,7)

25

I

I

II

0

67Ga SIS in the channel of 0-19 keV Figure 7b. 67Ga double spectral index plots of SIS in the channels of 7-100 key to 0i 9 key. The illustration of the figures in the parentheses is shown in Figure 5.

little (Figures 5, 7a, and 7b), their effect on the identification of quenching type and on quench correction can be ignored, because 670a has a sufficiently high counting rate. If 3H and/or '4C in the samples have higher counting rates, usual double ratio techniques5 and SCR quench correction methods3'4 can also be used to identify and correct the quench.

280

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 2. The Components and Parameters of Mixed Quench Samples

Batch

Cd4 (DL)

(FL)

3H

"C

QTCFb

2 5 7

63.5 34.2 22.7

95.0 85.6 77.7

0.615 0.313

0.407 0.219 0.185

2 5

47.9 25.8

7

17.1

92.6 83.5 75.8

0.706 0.353 0.278

0.607 0.385 0.337

2 5

28.1 15.2

7

10.1

85.9 77.4 70.3

0.892 0.695 0.497

0.724 0.516 0.464

2 4 6

57.8 38.4 25.8

94.6 88.7 80.8

0.308 0.273 0.188

0.500 0.327 0.267

2 4 6

52.8 34.8 23.5

93.6 87.8 80.0

0.759 0.523 0.417

0.564 0.360 0.320

2 4 6

47.9 31.8 21.4

92.6 86.9

0.531

79.1

0.445 0.333

0.607 0.428 0.359

2

43.7 29.0

91.2 86.0 78.3

0.625 0.488 0.387

0.639 0.462 0.392

20

1

Equivalent Counting Efficiency (%)

Color Solutiona

50

100

30

40

0.381

Fc

2

50

60 F

4 6

19.6

0.1296 ± 0.6280 QTCF

r = 0.7967

asaturated solution of methyl red in acetone. bExperimental QTCF of chemical quench from Figure 3. cCalculated contribution factor (F) of chemical quench from 3H.

QUANTITATIVE QUENCHING CORRECTION OF MIXED QUENCHED SAMPLES

The color quencher used in this study was a solution of methyl red in acetone (Table 2), but the acetone itself also is a chemical quenching agent. In order to decrease the effect of chemical quenching of acetone, we prepared a

saturated solution of methyl red in acetone. The maximum quantity of the color solution used in the work was 20 L, which decreased the relative counting efficiency of 3H to 3.5o, and that of '4C to 32o (Figure 8), relative to the unquenched sample. Under this condition the relative quenching contribution factor (Ft,) of chemical quenching of acetone was only about 0.04 for 3H. The formula for calculating F factor is as follows: 1 - Erc

= (1 - Erc) + (1 - Erci)

(1)

APPLICATIONS OF 67Ga DOUBLE SPECTRAL INDEX PLOTS

281

100 80

0 60 ci)

40 ci)

cr

20 0

2030 50

10

100 200300500 1000

Quantity of quancher (iiI) Figure 8. Chemical quench curves for 3H and '4C (quenching agents, acetone and carbon tetrachloride).

where Erc and Erci are relative counting efficiencies caused by chemical quencher and color quencher in the mixed quench sample. The chemical quenching effect of methyl red was not considered. The decrease in relative counting efficiencies of 3H and '4C with increasing amounts of carbon tetrachioride and acetone are shown in Figure 8. Likewise, changes in relative counting efficiencies due to the addition of increased quantities of saturated methyl red solution in acetone are shown in Figure 9. The equivalent relative counting efficiency of mixed quench samples can be represented by the product of the relative counting efficiencies of chemical and

colour quenching in these samples (Table 2). In addition, F of chemical quenching can be calculated according to Equation 1 (Table 2). The quenching contribution factor for color quenching (F1) equals 1 - F. Obviously, Equation 1 is usable only for the samples with known quenching components. When the quenching components are unknown, a new parameter, quenching type contribution factor (QTCF), has been defined by Yang4 to represent the relative position of mixed quench samples in a mixed quench zone: QTCFCI =

QTCF

E, - E1 E - En

(2)

282

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

100

>,

0C

80

360 ci)

C)

> 1

ci)

1

23

5

10

203050

100

Quantity of the saturated solution of methyl red in acetone (tl) Figure 9.

Color quench curves for 3H and 14C (quenching agent saturated solution of methyl red in acetone).

where E and E1 are the counting efficiencies on the chemical and color quenching correction curve corresponding to the QIP (quenching indicating parameter) of the mixed quench sample. E, is the counting efficiency of a mixed quench sample. In contrast, when QTCF is known we can calculate the counting efficiency of unknown mixed quench samples according to its QIP and QTCF:

E, = E - QTCFCI (E - E1) E = E1 + QTCFC (E - E1)

(3)

The QTCF value of the sample with unknown quenching components cannot be determined by Equation 2 as the counting efficiency E is unknown. The double vial method4 has been used to solve this difficulty: l/m of total volume

is taken from an unknown sample after the first counting and is diluted to original volume with the same cocktail, then it is recounted with the same conditions. Suppose QIP1 and QIP2 were obtained from the first and the second measurements, respectively; thus,

APPLICATIONS OF 67Ga DOUBLE SPECTRAL INDEX PLOTS

283

0.9

0.8

0 1

0.5

4-I

!0.3 0 o

F=0.1 2960.62800TCF

.

0.2

(r=0.7967) 0.1

0

I

0.1

I

0.2

0.3

I

I

0.4

0.5

I

0.6

I

I

I

0.7

0.8

0.9

Experimental QTCF Figure 10.

The relation of calculated F factor to experimental QTCF. These parameters were obtained from mixed quench samples.

QTCF

mrE " = mr E "- E CC

(4) CC

where E' and E" are the counting efficiencies from the chemical quenching correction curve corresponding to QIP1 and QIP2, respectively. Therefore, E'= EC' - ECu' and E" = E" - E1"; and E1' and ECI" are the counting efficiencies on the color quenching correction curve corresponding to QIP1 and QIP,, respectively; where r = n2/n1, and where n1 and n2 are the counting rates of the first and the second measurements, repectively. If the amount of the second sample was m times that of the first sample, then QTCF

=

mE" I

(5)

284

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

The reason for using the double vial method to determine QTCF is that the position of an unknown sample in a mixed quench zone cannot be determined

by a single count, as in the usual quench correction method. Now, in this experiment the difficulty was resolved by the addition of 67Ga to the unknown sample. The quench correction curves mentioned above were replaced by dou-

ble spectral index plots. The point distributive character of mixed quench samples in the mixed quench zone of double spectral index plots are the same

as in quench correction curves. Therefore, QTCF can also be used, for example, when an unknown sample was measured and its SIE was determined. The chemical and color quenched SIS values in the energy range of 15 to 100 keV, corresponding to SIE in Figure 5, were SIS and SISI respectively, while the SIS value of the sample itself was SIS. We can then obtain experimental QTCF values from Formulas 6 and 7 listed below:

QTCF i

TCF

- sIs

- sIs - sIs -

- sis,

where SIS, - SIS,1, SIS, - SIS, and SIS - SIS are negative values because the color quenched curve was above the chemical quenched curve. However, the value of QTCF is always positive. QTCFCI and QTCFC characterize the degree of the contribution of color and chemical quench to mixed quench samples, respectively. The values of QTCF taken from Figure 7 are always the same as those taken from Figure 5, as only 67Ga SIS and SIS1 in the region of interest of 15 to 100 keY (corresponding to the SIE of a mixed quench sample) will change correspondingly to 67Ga SIS. in the region of interest of 0 to 7 keV. Thus, when we substitute QTCF into Equation 3, the counting efficiency of an unknown sample is resolved. The 3H corrected results, which are obtained by using QTCF (from Figure 5) and SIS quench correction curves, are given in Table 3. Except for one sample, the QTCF method gave more accurate values than those obtained by using either the chemical or color quenching correction curve alone. The mixed quench zone is wider in the range of strong quenching; therefore, the employment of QTCF for quench correction in that range has particularly significance. DISCUSSION

The experimental QTCF values of 21 mixed quench samples were obtained from Figure 3 and Equation 7 (Table 2). The correlation of the calculated F factor with experimental QTCF values for mixed quench samples was given in Fig. 10, in which F = 0.1296 + 0.6280 QTCF (r = 0.7967). It is illustrated that F factor is proportionally related to QTCF; in fact, this is the basis of the

11.8 10.8

12.1 11.1

12.8 12.4 11.2

6290 5610 3204 5328 2917 4444 2445 9.1

15.3 8.5 13.2 6.3

13.2

20.0

12.6 17.8 10.7

17.4

20.4

24.5 21.9 22.9 18.8 10.2 18.9 10.6 15.6 8.3

aEach sample contained 20 ,L aqueous solution of 3H-thymidine (30014 dpm). actual activity - corrected activity bError (%) 100 actual activity cTaken from curves in Figure 7 (left) dcounting efficiencies were obtained by QTCF from Figure 3.

7

2 3 4 5 6

1

24966 22850

23151

25673 25616 24273 26640

-17.6±4.3

-11.2 -22.9 -16.8 -23.9

-19.1

-14.5 -14.6 35209 34824 34318 33667 38810

32241

30833

14.2±7.8

2.7 7.4 17.3 16.0 14.3 12.2 29.3

27467 29840 31412 28190 27519 28487 29458

-1.9 -3.7±4.4

-5.1

-8.3

-6.1

4.7

-8.5 -0.6

Table 3. The Comparison Between Activities Obtained Using QTCF and Activities Obtained Using the Usual Quench Correcting Curves for Strongly Mixed Quench Samplesa Counting Efficiency Activity (dpm) SIS of Sample Sample Counts Chem. Color 0-19 OTCE Chem. Error Color Error Using Error (%)C (%)C (%)d (%)b (%)b Number keV (cpm) Curve Curve QTCF (%)'

286

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

quench correction method for mixed quench samples in this chapter. If we call

the usual quench correction "first order quench correction," then QTFC quench correction in mixed quench zone can be called "second order quench correction." Obviously, the magnitude of the second order quench correction is less than the first order quench correction; therefore, it must have greater relative variability. Based on this point, in principle we can divide the mixed

quench zone into subzones by the measurement of two parameters of the curve. All points falling in the same subzone were considered to have the same QTCF value which is more practical for application. If a liquid scintillation counting system can automatically give two spectral indexes for two different parts of a sample (or external standard) spectrum at the same time, it is not only convenient for measurement of double spectral index plots, but also more usable tovarious new and existing applications. The authors had hoped to prepare double spectral index plots of external standards

to correct mixed quench samples, but the commercial instrument used was unable to measure the STE of different energy ranges. The main advantages of using -7Ga or fl3mIn as an internal standard include, (1) the elimination of accurate pipetting of the internal standard, (2) the ability to recover the sample after the spiked activity has decayed, and (3) the high count rate.' Among these advantages the one with the greatest importance is probably the fact that the ctntent of "standard" was changed from quantity of

activity to energy or its spectrum shape. The mixed quench samples in this paper consisted of simple chemical and color quenchers. Practical samples are very complex, therefore, it is necessary to do further work. In addition, further work with the double spectral index plots or ratio of spectral index may yield some new applications.

ACKNOWLEDGMENTS The authors wish to express their gratitude to Mr. Toru Fujita and Mr. Hideo Miyatade for kindly supplying 67Ga-citrate and labeled hexadecane respectively.

Note: This manuscript has undergone extensive revision and condensation in order to meet the style requirements and space limitations of the proceedings. We sincerely hope that we have not changed the substance or connotation of any information intended to be conveyed by the authors. Eds.

REFERENCES Dyer, A. An Introduction to Liquid Scintillation Counting (London: Heyden & Son, 1974), p. 57. Parmentier, J.H. and F.E.L. Ten Haaf. "Developments in Liquid Scintillation Counting Since 1963," mt. J. AppI. Radiat. Isot. 20:305 (1969). Yang, S.L. "Effect of Chemical and Color Quenching Ratios on Calibration Curve

APPLICATIONS OF 67Ga DOUBLE SPECTRAL INDEX PLOTS

287

in Liquid Scintillation Counting," Chinese .1. of Nuc. Med. 1:92 (1981); C.A. 96, 170718 (1982).

Yang, S.L. "Automatic Quenching Correction Method for Samples with Both Chemical and Color Quenching in Liquid Scintillation Counting," Acta A cad. Med. Sinicae 3:114 (1981); C.A. 96, 42804 (1982). Takiue, M., T. Natake, and M. Hayashi. "Double Ratio Technique for Determin-

ing the Type of Quenching in Liquid Scintillation Measurement," ml. J. App!. Radial. Isot., 34:1483 (1983). Bush, E.T. "A Double Ratio Technique as an Aid to Selection of Sample Preparation Procedure in Liquid Scintillation Counting," mt. J. App!. Radial. Isol., 19:447 (1968).

Ross, H. H. "Color Quench Correction in Liquid Scintillator Systems Using an Isolated Internal Standard," Anal. Chem., 37:621-623 (1965). Lang, J.F. "Chemical vs. Color Quenching in Automatic External Standard Cali-

bration. Application of Empirical Observation in a Computer Program," in Organic Scintillators and Liquid Scintillation Counting, D.L. Horrocks and C.T. Peng, Eds. (New York: Academic Press, 1971), p. 823. McQuarrie, S.A. and A.A. Noujaim. "67Ga: A Novel Internal Standard for LSC,"

in Advances in Scintillation Counting, McQuarrie, S. A., C. Ediss, and L.I. Wiebe, Eds. (Edmonton: University of Alberta Press, 1983), pp. 57-65. Ell, J. and E.S. William. Nuclear Medicine: An Introductory Text. (Oxford, London: Blackwell Scientific Publications, 1981), p. 159. Oldendorf, W.H. "Liquid Scintillation Quench Correction Using 1 13m-Indium Conversion Electrons as an Internal Standard," Anal. Biochem., 44:154 (1971).

CHAPTER 24

Modem Applications in Liquid Scintillation Counting

Yutaka Kobayashi

New counters today are all controlled by sophisticated software which means that new liquid scintillation counters are now married to computers. Therefore, it seemed appropriate about two years ago to explore the applications of a modem to liquid scintillation counting technology. The person who aroused my interest in this study was Stan Dc Filippis, senior marketing applications specialist, who was then at the Packard Instrument Company. He envisioned that the modem could provide both the instrument manufacturer and the user with a very powerful tool. The major concern with this idea was reliability because of the well known problems associated with early attempts at data transmission with modems and personal computers. To explore this concept, a modem was placed into my counter, a Packard Model 2200CA, located at the Worcester Foundation for Experimental Biology in Shrewsbury, MA. The Packard 2200CA is controlled by an IBM model PS/2-30 computer equiped with a 20 MB hard disk. The modem was a Leading Edge Model "L" set to run at 2400 baud/sec. An identical modem was used in a computer at the Packard Instrument Company located in Downers Grove, IL. Later in this study a Hayes Smartmodem 2400B modem was installed in my home computer in Wellesley, MA, about 25 miles from Shrewsbury.

Once the modems were in place, we needed to choose a communications package to make the system work. After trying a few commercial software packages, we settled on one called CLOSE-UP from Norton-Lambert Corporation. The CLOSE-UP system consists of two separate software packages: one called Support, which was used at the remote receiving stations located in Downers Grove and Wellesley and another called Customer, which was installed in the counter. At the Worcester Foundation all telephone calls are routed through a switchboard. Although the communication software is designed to operate under these conditions, we decided to install an outside line directly to the modem in the counter to avoid this high potential source of problems. 289

290

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 1. Modem Applications in Liquid Scintillation Counting from a Remote Site Investigated in this Study Monitor counter Control/Operate counter Trouble-shooting aid Update software Transfer data Automatic data transfer

When the counter is called from Downers Grove, IL and contact is established, the screen of the counter appears on the computer screen in Illinois. From that moment, the counter can be controlled from the computer keyboard in Illinois just as if the operator were standing at the counter in Massachusetts. We have found the system to be very reliable and useful. In Table 1, all the applications successfully executed during this study are listed. This system allows you to monitor the counter from a remote site. For example, if you wanted to check on the state of your counter from home, you can call your counter and examine the Status Page. The Status Page in a Packard 2200 counter tells you what program is being used and which sample is currently being counted. On the other hand, if you were scanning a group of samples to see the distribution of counts among them and wanted to change the counting time from, say, one minute to ten minutes, you can also do that from home without going to the laboratory. This has proven to be a convenient option. Perhaps the most intriguing potential application of the modem is its use as a trouble shooting aid. This aspect of the modem, however, was not really tested because the counter did not have any serious problems during the study. In the case of the Packard 2000 series counters, the potential usefulness of the modem is there because of the diagnostic software contained within the counting system software. An engineer in Illinois can actually look at all the operating parameters of the counter in Massachusetts, such as the history of the high voltage applied to the photomultiplier tubes for the last 10 instrument normalizations. He can also check the performance history in terms of the check source counting efficiency for the last 90 normalization cycles; all of this is stored in memory. In terms of diagnosing problems, the only occasion in which

the modem was used for this purpose was to send the puzzling data for a particular spectrum to the engineers in Illinois for analysis. During the course of this evaluation, a software update for the Packard 2200 counter was released. The installation of the new software was done using the modem. This meant erasing the old software and installing the revised version. The entire process went without any problems. This demonstrated one operation which could save both the instrument manufacturer and the user a lot of time. In the case of the manufacturer, the need to schedule a service call for the installation of the new software is eliminated, and, in the case of the user, the entire software update can be scheduled after working hours when the instrument is not in use.

MODEN APPLICATIONS IN SCINTILLATION COUNTING

291

Table 2. Example of a Task File used in CLOSE-UP for Automatic Data Acquisition using a Modem File Name: DATAOUT.TSK

CONFIG DIAL REDIAL = 3 WAIT = 3 WAIT UNTIL 23:45 BAUD 2400 PRINT LOG ON DIAL 1 5558425555a FETCH C:SDATA1O.DAT TO A: HANGUP PRINT LOG OFF aFjCtitiOUs telephone number.

The most useful feature of a modem-equipped counter is its ability to transfer data to a remote site. In the Packard 2200CA, it is possible to store any counting data onto the hard disk. This data can then be transferred to any remote site via modem. During this study, this was achieved as a routine option. The convenience of this option is that the researcher can now transfer his counting data to his home computer and analyze the data without going to the laboratory. Perhaps the most unique feature of the communication software package, CLOSE-UP, is its ability to transfer counting data automatically. For example, the computer at home can be programmed to call the counter, say at midnight when the phone rates are low, and transfer counting data while I sleep. This is accomplished by using a simple routine called a task file. An example of a task file is shown in Table 2. This simple, eight line test program was written using EDLIN which is a resident DOS text editor. This task file was called DATAOUT and all task files are designated by the extension, TSK. DATAOUT is stored in the subdirectory

which contains CLOSE-UP. The first line instructs CLOSE-UP to make a total of three attempts to connect with the counter. If the first try is unsuccessful, it will wait three minutes before trying again. The next line instructs the program to make the first call at 11:45 PM; military time is used here. The next line sets the baud rate. The next line turns on the printer to type a log of all activities. The next line gives the phone number (the phone number listed is a

fictitious number). After contact is made with the counter, the next line instructs the computer operating the counter to fetch a file from the C drive called sdatalO.dat and copy it to the A drive of the home computer. When the task is completed, the program hangs up the phone and turns off the printer. Table 3 shows the log generated by the task file. The program was initiated on September 14th and the name of the task file in control was DATAOUT.TSK. The program was activated at 11:45:07 PM; connection was made on the first try at 11:46:32. The data file, sdatalO.dat, was located on the C drive and copied to the A drive of the home computer. The counting data for 28

samples were counted under low-level conditions. The transferred data received and stored in the home computer is shown in Table 4. The log shows

292

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 3. Example of a Log for Task File, DATAOUT.TSK Sept. 14 1989 DATAOUT.TSK 5dial 15558425555a 23:45:07 Connect: 1 attempt 23:46:32

6fetch c:sdatal0.dat to a: Complete - 00:00:06

23:46:47 23:46:51

7hangup 8print log off

aFICtItIOUS telephone number.

that the call was terminated at 11:46:47 for a total connect time of only 15 seconds and 4 seconds later, the printer was turned off. All of this was done without anyone being present. This sample test file was run to check out the system. Also, in actual practice, the file transfer system can be protected by requiring a password to allow only authorized access to the counter computer system. In setting up this automated data collection system, it was reasoned that the

counting data should be collected in the A drive of the computer controlling the counter rather than in the C drive. The removable diskette would be a backup system for the automated data processing system; should any problem occur with data transfer, the original data will always be saved on a removable Table 4. Data File, SDATA1O.DAT, Transferred by Modem from a Packard 2200CA Counter. Data Transferred, in Sequence, are: Sample Number, Time in Mm.: Region A cpm, Region B cpm, Region C cpm, SIS and tSIE. 1 ,10.O0,10.9000,4.2000,9.8000,47.29,0.00 2,10.00,8.8000,2.1000,7.7000,54.20,0.00 3,10.00,8.4000,3.9000,7.7000,60.330.0O 4,10.00,10.4000,4.5000,8.8000,62.50,0.00 5,10.00,10.2000,4.3000,9.2000,66.46,0.00 6,10.00,11 .1000,3.5000,9.7000,58.33,0.00 7,10.00,10.8000,3.8000,9.3000,55.33,0.00 8,10.00,14.1000,3.9000,12.2000,49.20,0.00 9,10.00,10.4000,4.7000,9.0000,69.42,0.00 10,10.00,9.3000,3.7000,7.6000,56.67,0.00 13,10.00,12.5000,3.8000,11.7000,54.37,0.00 14,10.00,12.2000,3.6000,9.4000,51.17,0.00 15,10.00,9.6000,3.5000,7.9000,57.22,0.00 16,10.00,9.9000,3.2000,8.9000,53.73,0.00 17,10.00,11.7000,3.7000,9.9000,51.33,0.00 18,10.00,8.8000,3.3000,7.8000,59.44,0.00 19,10.00,9.7000,4.4000,8.2000,61.98,0.00 20,10.00,10.9000,3.8000,9.9000,60.83,0.00 21,10.00,9.3000,3.0000,8.4000,60.77,0.00 22,10.00,12.5000,3.2000,10.0000,44.14,0.00 25,10.00,8.0000,0.7000,6.2000,21 .22,0.00 26,10.00,31.7000,12.300,30.2000,46.69,0.00 27,10.00,30.9000,14.1000,29.1000,49.59,0.00 28,10.00,30.5000,15.9000,29.1000,52.94,0.00 30,10.00,76.0000,45.0000,74.2000,56.67,0.00 31,10.00,168.600,89.0000,164.9000,53.49,0.00 32,10.00,77.4000,46.4000,75.6000,57.51,0.00 33,10.00,15.200,0.4000,13.4000,15.63,0.00

MODEN APPLICATIONS IN SCINTILLATION COUNTING

293

diskette. This means that the Packard 2200 counter will always have a diskette in the A drive. This can be a problem in the event of a power failure, because the counting system is programmed to automatically reboot from the C drive

after power is restored. The problem is that, in the IBM PS/2-30 computer used in the counter, the system always first looks at the A drive, for rebooting. This situation can be accommodated by putting all the counting system software on the A drive but this would reduce the data storage capacity of the data diskette. This problem was avoided simply. First, the A diskette is formatted using the FORMAT A:/S command. The slash S calls for two hidden DOS system files to be copied onto the formatted diskette. Without these two hidden files, the system will not boot up properly

from the A drive. Next, using EDLIN, a DOS text editor, write a one-line autoexec.bat file as follows: C: \KO BY. BAT

All data diskettes used in the A drive must be formatted this way and contain the one line autoexec.bat file. Then, on the C drive, that is, on the hard disk in the DOS directory, write another batch file called KOBY.BAT using EDLIN. This file contains only two instructions: C:

autoexec.bat

This little routine ensures that the counter will reboot normally after power failure when there is a data diskette prepared as described in the A drive. In considering the practical aspects of modems in counting technology, I believe they have limited, but very useful attributes. From an instrument manufacturer's point of view, the possibility of updating software via modem directly from the plant should be an attractive prospect. This system would assure timely distribution of software to all customers and save a service call as well. Also, the ability to inspect a customer's counter performance remotely can be advantageous, especially after a service call was made or a new procedure is being evaluated. I can envision the time when a manufacturer can offer a preventive maintenance program which involves scheduled, invisible inspec-

tions of a working counter through a modem. The fact that a log of these invisible inspections can be generated would assure the customer that the service was indeed provided. For the user, the modem does have limited applications at this time. However, in certain situations I believe the modem can be very useful. For example, any diagnostic laboratory running assays can have the completed results transferred to its various customers via modem automatically every night by using a task file. The customers will then have the results available before the start of the working day. In an industrial plant where the counting is done in a central analytical laboratory, the completed counting data can be sent directly to any laboratory on site provided the remote site, has a modem-equipped computer.

294

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

In my limited experience, I have found the modem to be a very convenient and useful option. Even though it has limited applications, the data transfer capability alone justifies its presence in my counter . What is attractive about this technique is that it is a relatively inexpensive option to own. Modems can be purchased for under $100 and good communications software is available in the public domain. For example, a highly regarded public domain software is PROCOMM. This software has been commercialized and is sold as PROCOMM PLUS at nominal cost, especially at software discount houses. However, the original PROCOMM is available and should be more than adequate for this application. The advantage of using a commercially produced software is that these packages are easy to use and well documented.

CHAPTER 25

A New Procedure for Multiple Isotope Analysis in Liquid Scintillation Counting A. Grau Caries and A. Grau Maionda

INTRODUCTION

Double and triple labeled radiative samples are interesting in biological tracing work. The procedures to obtain the activity of each one of the components in a multilabeled sample are: double or triple window measurementes,' spectral shift by quench variation,2 rate counting at different times,3 different counting methods,4 different decay modes.5 All these procedures are based on the counting rate of the whole or part of the pulse height spectra. It seems interesting to look for new counting procedures in order to overcome some of the very known restrictions. This procedure could be based on using the differential shape of the pulse height spectra. The aim of this chapter is to develop a procedure, based on spectra fitting, that obtains the activity of each one of the radionuclides in the sample. The procedure will be applied to mixtures of pure beta-ray emitters, 3H + '4C, and 35S + 14C. The application to other kinds of decay schemes is possible, but they will not be studied here. A code named DILATA has been developed in order to compute the spectral components and activities of radionuclides mixtures. In this chapter, only the mathematical procedure will be described. FITTING PROCEDURE

The spectral issue of a liquid scintillation counter for beta-ray emission depends on the quenching degree of the sample. Two of the eight '4C spectra obtained with a logarithmic amplifier have been plotted in Figure 1. The shape and pulse height of the spectra depend very strongly on the quenching of the samples. The first problem to be solved is how to obtain the spectrum which corre-

sponds to a given quench level. A direct interpolation can not be applied 295

296

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Figure 1.

Experimental beta-ray spectra of 14C for two different quenching values.

because of the dispersion of the spectra. One possible interpolation method could be the interpolation by spectral dilatation. The general procedure has the following steps: A set of spectra with different quench is obtained for each nuclide. The interpolation procedure is applied to each nuclide, to obtain the spectra corresponding to the quench level of the experimental double labeled sample. A least square fitting is done with the interpolated spectra and the double

labeled spectrum in order to compute the activities of each one of the nucides.

The procedure can be applied to multilabeled samples, but for simplicity and clarity we only discuss double labeled samples. Special Functions

The experimental spectra, see Figure 1, are histograms and have the incon

venience of being defined for discrete values of the channel spectra. It is interesting, from a mathematical point of view, to have a continuous function defining the spectra.

The spectral function has been obtained by fitting Fourier series to the experimental spectra. The spectral function f(w) is obtained from:

A NEW PROCEDURE FOR MULTIPLE ISOTOPE ANALYSIS

f(w)= fa+ bw +

krw

Ck

M-1

o

w

297

(1)

where N is the number of Fourier harmonies and M is the number of experimental points of the spectrum. The first and last spectral points are w = 0 and w = M - 1. w' is the end of the spectrum. The constants a, b and Ck can be obtained from the following equations: a = b

2

Ck= M-1

y0

YM-1Y0

M-1

M-1

j=0

. irkw YJS1flM_hl

where y the number of counts in channel j, and

= y - (a + bw1)

w=j=O,l,2,...M-l. The calculation of the spectral function f(w) is interesting for two reasons: it can eliminate very properly the statiscal fluctuations of the spectra and permits a continuous function that allows application of the spectral dilatation interpolation method.

The first problem is defining the number of harmonics to be taken in the Fourier series. It is known that if the number of harmonics is equal or close to

the number of experimental points of the spectrum, the function f(w) will follow all the experimental points. In this case the statistical fluctuations will be included. We are interested in having a certain smoothing of the spectrum in order to attenuate the effects of statistical fluctuations. However, if we take a

very low value for N, the function f(w) can loose interesting or essential structures of data. That will produce a modification of the real shape of the spectrum. Different values of N have been tried. N = M/2 seems to be a good compromise to obtain a moderate smoothing without loosing essential structures.

Spectral Dilatation-interpolation Method

The spectral dilatation-interpolation method is interesting in the case of logarithmic spectra, nevertheless it is also applicable to linear spectra.

The spectral function f(w) can be separated into two regions. Region 1 ranges from zero to the value that corresponds to the maximum of the spectrum. Region 2 is the remainder of the function. The spectral dilatation method consists of applying a shift to region 1 of all

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

298

the spectra that makes all the maxima coincide in the same point. The same procedure can be applied to region 2, however, the common point is w' in this case.

The transformation applied is a linear dilatation defined by:

w' = --w Wk

(7)

where w1 is the value of the common point after doing the dilatation and Wk is the position of this common point for each spectral function before carrying out the dilatation (k = 2,3 . . . v), where v is the number of spectra for each nuclide. We see in Figure 2 that the common point for region 1 is the maximun

position of the less quenched spectrum. The w value of the less quenched spectrum is the common point for region 2. It is proved that the choose of the position of the inflexion in the spectra or the point y/r as the common point for region 2 also gives good results. Once the dilatation is computed it is not difficult to obtain the interpolated spectrum which corresponds to a given value of the quench parameter. First of

all, we carry out a channel by channel interpolation between the dilated regions of the spectral functions. Then we calculate a new spectral function for 25000

20000

15000

0 10000

5000

25

Figure 2.

Fourier spectra after shifting.

75

Channel

100

125

150

A NEW PROCEDURE FOR MULTIPLE ISOTOPE ANALYSIS

299

these points as we have done before for the experimental spectra. Finally, we apply the contraction: w2

w' =

WI

(8)

w

2 are the positions of the common point before and after the contraction is applied respectively. The next section shows how to calculate where w1 and w2.

Spectral Parameters Depending on the Quench

In the previous section we associated the maxima and the ends of the spectral functions with the common points that correspond to the maximum and the end of the least quenched spectrum. It is interesting to see the variation of these quantities as a function of the quench parameter. A straight line can be obtained by least square fitting of the data. The maximum and the end of the spectral function for a given value of the quench is calculated from this fitting. Special Fitting

Once the radionuclide spectra, corresponding to the quench parameter of the experimental mixture have been interpolated, obtaining the activity of each component of a radionuclide mixture is a problem of least square fitting. Let y be the number of counts in the channel i of the mixture spectrum and y be the number of counts in the channel i for the spectral component j. The

intensities p for each nuclide are related to the spectral components by the equation:

yl -

(9)

is the number of counts of channel i in the computed spectrum and n is the number of spectral components. The values for p are obtained from the condition of minimum for the quantity: where

G (Pi.....PN)

=

i=M W [y, -

(p)]2

(10)

i=1

where w is the weight for channel i. These weights have been taken inversely porportional to the statistical variance of the count number in each channel: WI =

1

y

300

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

As the values of y depend linearly on the p, these can be computed by solving a system of n simultaneous linear equations. The necessary and sufficient condition for the minimum of G is that the following equations hold,

ÔG_G_

öG_0

P-

&Pi -

ôPn -

these are equivalent to the system of equations: p1

w1y21+

p1

w y11 y

+p

w1yy1 =

wyy11

w, V2 , ,,, --

w, Y y,,

+

or in matricial form: PQ = Y

(14)

where P =

(p1

..

. p)

wy21

= -

...

wy1y

=(

wyy1

wY

wyy,,)

The index i has been omitted in order to lighten the expressions. The p values are obtained from the equation: P = YQ

(18)

The activity A1 of the i-radionuclide is given by the formula: A, =

(19)

where Si is the counting rate and e, the counting efficiency obtained from a calibration curve functioning as a quench parameter. EXPERIMENTAL

The procedure described above can be applied to any mixture of radionuclides independently of the decay scheme. The limitations of the procedure will be given by the measurement data. In principle the procedure can be applied to beta- and alpha-ray emitters and

to electron capture nuclides. The multiple mixture of these nuclides and the

A NEW PROCEDURE FOR MULTIPLE ISOTOPE ANALYSIS

301

Table 1. Activity Discrepancies for '4C + 3H mixtures

a

'4C/3H

326.6 216.2 326.0

6.3 14.2 0.6

z14C (%)

0.16 0.55

7.0

3H (%)

0.57 7.1 2.2

different quench and proportions give a great number of possible combinations to be checked. In this chapter the procedure will be applied only to some particular mixture in order to show its possibilities. Three radionuclides have been used in our experiment: 3H and '4C (both as n-hexadecane), and 35S (dioctyl suiphide). All the solutions were standardized by LMRI (France) and dispersed homo-

geneously into 10 mL of toluene or dioxane based scintillator. The standard solutions were added gravimetrically to each vial using a Sartorious microbalance. Carbon tetrachioride was used as a quencher. A LKB 1219 Rackbeta Spectral liquid scintillation counter has been used to obtain the spectra. RESULTS AND DISCUSSION

Four different situations have been considered: Mixture of '4C and 3H with middle quench and 6.3 times more counting rate of '4C than 3H Mixture of '4C and 3H with high quench and 14.2 times more counting rate of '4C than 3H Mixture of '4C and 3H with moderate level of quench and very poor counting statistics. Mixture of '4C and 35S with moderate level of quench and sim ilar counting rates for both radionuclides.

Figure 3a shows the experimental spectrum of the '4C and 3H mixture. The quench paramenter for this sample is Q = 326.6, this corresponds to a counting efficiency for 3H of about 0.29%. The counting ratio relationship between '4C and 3H was about 6. The spectrum of residuals Figure 3b shows a balanced dispersion. Table 1 shows that the discrepancies between experimental and computed activities are 0.16% for '4C and 0.57% for 3H. Figure 4a shows the experimental spectrum of a '4C and 3H mixture with a quench parameter Q = 216.2, the corresponding counting efficiency is 0.11%. The counting ratio relationship between '4C and 3H is 14.2. The spectrum of residuals, Figure 4b, shows some predominant positive values due to the oscillations of the experimental spectrum close to the maximum. Table 1 shows that the discrepancies between computed and experimental activities are 0.55% for '4C and 7.1 Wo for 3H. A better stability of the experimental spectrum would improve the discrepancies.

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

302

25000

EperInntcI

C+ H

20000

rn

15000

0

0

10000

5000

75

50

Figure 3a.

Channel

100

125

150

Experimental spectrum of a mixture of 1C + 3H and spectral decomposition for middle quenching.

1000750-

500

250

I

C)

'I a)

'I

250-

1

-500-

1

-750-

1000 0

Figure 3b.

25

50

75

Channel

Residual spectrum for the fitting of Figure 3a.

10

15

10

A NEW PROCEDURE FOR MULTIPLE ISOTOPE ANALYSIS

/

35000

+H

EperirnentoI

Cornpted

30000

303

C

25000

20000

0 0 isoio

10000

5000

25

Figure 4a.

50

Channel

Experimental spectrum for a high quenched mixture of 14C +

10001

Figure 4b. Residual spectrum for the fitting of Figure 4a.

100

75

l-l.

304

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS 70

60

50

EperentoI 'C + H

.40 0

030

Cornptod

H

20 cornpted

C

10

25

Figure 5a.

50

75

Channel

100

125

Experimental spectrum of a mixture of 14C and 3H with very poor statistics and spectral components.

10-

5-

0-

10 0

Figure 5b.

25

5b

15

Channel

Residual spectrum for the filling of Figure 5a.

100

125

305

A NEW PROCEDURE FOR MULTIPLE ISOTOPE ANALYSIS 40000

35000

+

EperirnntoI

30000

25000

Cornpted 'C

rj

0

20000

C-)

co,npded

s

15000

10000

5000

50

25

75

100

125

Channel

Experimental spectrum and spectral components for a mixture of 14C and 35S.

Figure 6a.

1000-

800-

I

600-

400 'I)

I

200

200 400

600 800 1000 0

Figure 6b.

25

50

75

Channel

Residual spectrum for the fitting of Figure 6a.

100

125

306

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Figure 5a shows the experimental spectrum of '4C and 3H mixture with a quench parameter Q = 326. The measurement time was very short (1.5 seconds). That means that the counting rate is very low, like a low level activity sample, but without background. The counting rate between '4C and 3H is about 0.6. Figure 5b shows the residuals of the fitting. Systematic positive or negative values are clearly obtained, but the balance between both is quite acceptable. Table 1 shows that the discrepancies between computed and experimental activities are 7.0% for '4C and 2.2% for 3H. Figure 6a shows the experimental spectrum of '4C and 35S. The quench is Q

= 330, and the counting ratio relationship is C/S = 1.5. The discrepancies for 35S and 'C are 3.4% and 2.2%, respectively. Figura 6b shows the residuals of the fitting. It can be seen that for the low and high region of the spectrun there are several residuals unbalanced. The analysis of 35S and 4C is a very difficult problem because the maximum beta-ray energies for both radionu-

clides are very close, but the differences in the spectral shape permit the activities of each one of the radionuclides.

CONCLUSION

A computer code DILATA has been developed which is applicable to the analysis of radionuclide mixtures. The method is based on a new procedure to interpolate spectra and has been applied to obtain the activity of 14C and 3H mixtures in different conditions. The application to a mixture of '4C and 35S has shown the power of the method. REFERENCES Simonnet, G. and M. Oria. Les Mesures de Radioactivité a l'Aide des Co,npteurs a Scintillateur Liquide, (Paris: Eyrolles, 1977), PP. 89-95. Martin-Casallo, M.T. and A. Grau Malonda. "Un Nuevo Procedimiento de Calibra-

cion de Muestras Doblemente Marcadas Basado en el Método del Trazador," CIEMAT Report (in press). Horrocks, D.L. Applications of Liquid Scintillation Counting, (New York: Academic Press, 1974), p. 229, Brown, L.C. "Determination of Phosphorus-32 and 33 in Aqueous Solution," Anal. Chern., 43(10): 1326-1328 (1971). Horrocks, D.L. and M.H. Studier. "Determination of Radioactive Noble Gases with a Liquid Scintillator," Anal. Che,n., 36(11): 2077-2079 (1964).

CHAPTER 26

On the Standardization of Beta-Gamma-Emitting Nuclides by Liquid Scintillation Counting

E. Garcia-Toraho, M.T. Martin Casatlo, L. Rodriguez, A. Grau, and J.M. Los Arcos

INTRODUCTION

The 4ir3(LS) efficiency tracing method with 3H has been successfully applied to the standardization of pure beta emitters.1-5 In this method, a 3H

standard is used to calibrate the measuring system, and a combination of experimental measurements and theoretical computations standardizes the unknown nuclide. We present here the extension of this method to the case of any ,3--emitting nuclide. The basic diagram of the efficiency tracing method is shown in Figure 1 and a complete description can be found in previous papers)'2 A set of vials containing a 3H standard is measured and a distribution of efficiency vs a quench parameter is obtained. On the other hand, the theoretical efficiency can be computed as a function of the figure of merit M, that we will define here as the energy in keY required to produce one photoelectron at the first dynode of the phototube.4 The next step is obtaining the distribution of the figure of merit vs

the quench parameter, which is independent of the radionuclide. The last distribution characterizes the measuring system. If we can compute the theoretical overall efficiency of the unknown nuclide vs the figure of merit, then we can calculate its activity from the precedent distributions. COMPUTATION OF THE THEORETICAL COUNTING EFFICIENCY

In order to extend the method to any beta-gamma emitter, the counting efficiency must be computed. This implies the computation of all the possible

ways the nuclide can decay to the ground state. There are three different processes that must be taken into account: beta emission, gamma emission, and electron conversion. First we will study the efficiency computation for these individual processes and then we will show how they can be combined to obtain the overall efficiency. 307

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

308

EFFICIENCY

3H

3H

U N

EFFICIENCY

COUNTING

K N

IUENCII PARAM.

0

FIG. OF MERIT

w

QUENCH PARAM.

N

FIG. OF MERIT

i U N

QUENCH PARAM.

COUNTING

K N

EFFICIENCY N

ACTIVITY

Figure 1.

Block diagram of the efficiency tracing method with 3H by LSC. Boxes drawn in thick lines correspond to experimental measurements.

Beta Efficiency

Taking into account shape factors, the beta spectra are calculated from th Fermi theory of beta decay. According to this theory, the number of particles of energy between E and E + dE is given by the equation: dN(E) - C(E) FO(Z,E) L0(Z,E) (Em - E)2 (E + 1) p, dE

(1)

where Em is the maximum kinetic beta particle energy in m0c2 units, and F0 and L0 are given in reference.4 C(E) is the shape factor, and its value6 is given in the

Table 1. In the case E = 0, the equation is: dN(E) dE

- 8ir

C(o) (2R)2

(aZ)2'

E2 Lm(ZöO)

F(2

+ 12

(2)

where R is the nuclear radius of the residual nuclide and a is the fine structure constant. A more detailed description of the spectrum calculation can be found in References 1 and 4. Table 1. Values of the Prohibition Parameter (q is the Neutrino Momentum in m0c2 Units) Prohibition Value Allowed and first forbidden Unique-first forbidden and second forbidden

+ q2

Unique-second forbidden and third forbidden

p4 + 19p2 . q2 + q4

Unique-third forbidden and fourth forbidden

p6 + 7p2q2 (p2 + q2) + q6

STANDARDIZATION OF BETA-GAMMA-EMITrING NUCLIDES

If the beta spectrum

309

normalized to be unity, the counting efficiency is

is

given by the expression: Em

N(E)

(i

- exp [

=

E.Q(E)i )2dE 2.M J

(3)

Here, Q(E) is the ionization quenching correction factor given by: Em

Q(E) =

dE

(4)

dE 1 + k.B (dx) 0

In this work, we used the following approximate equation7: Q(E)

-

.357478 + 459577*t + .159905*t2 1 + .0977557*t + .215882*t2

with t = log10E

(5)

The expression E.Q(E) in Equation 3 represents the amount of energy converted in photons and if the figure of merit (M) is considered, the probability of nondetection will be given by:

exP( E.Q(E) M

)

Hence, for a system with two tubes working in coincidence mode, the probability of detection will be given by: { 1 _exP(

E.Q(E))}2 2.M

Gamma Efficiency

To compute the y counting efficiency, the interaction probability and the Compton spectrum distribution must be obtained. This is carried out by means

of a Monte Carlo simulation. The emission point of the photon is drawn according to the vial dimensions, then three direction cosines and a pathlength are drawn to define the arrival point for this step. The type of interaction is decided by a random process, taking into account the cross sections for photo-

electric and Compton processes; the trajectory of the photon is modified accordingly. The process finishes when the photon escapes out of the vial, its energy drops below 10 eV, or if the interaction is photoelectric. In any case, the amount of energy lost by the photon in the scintillator is considered to build up the electron distribution. The interaction probability is also obtained from the spectrum. A complete description of the process is given in GarcIa-Toraflo and

310

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Grau (1987), and a typical spectrum obtained in the simulation is shown in Figure 2.8 If N(E) is the distribution of energy of the electrons, the counting efficiency

can be obtained from the expression:

=

r

E.Q(E)

IN(E)(1_exP 12.MJ

dE

(8)

where E is the gamma-ray energy. Internal Conversion

Internal conversion transitions produce vacancies in the atomic shells. To compute the effective energy converted into light, one must first calculate the probabilities and effective energies for the different processes involved. Only K,L, and M internal transitions are considered. The complete expressions for all the processes have been detailed in Grau (1982). The internal conversion processes contributes to the efficiency as a function

F (, n1, E3, E), where 4 and n1 are the probability and number of emitted particles in the i-atomic rearrangement, E, is the energy corresponding to the j-th particle and E is the converted electron energy. Overall Efficiency Consider the case of a radionuclide which decays by beta emission and is followed by a cascade of n gamma rays to the ground state of the daughter nuclide. Although this is a very simplified model of 3-y emitter, the expression that gives the counting efficiency is as complex as: E

E1 E, N(E)S(E). . S(E) (l_e_

= 0

0

f 0

...

_EjQ(E))2dE

(9)

.dE

0

where

Efl = Maximum energy of the beta emitter E7, = Energy of the i-th gamma ray N(E) = Beta spectrum S1(E1) = Spectrum of Compton and photoelectric electrons In a real case, some of the gamma transitions could be converted, and the appropriate conversion coefficients would affect Equation 9, which should be split in all the possible combinations of gamma and electron conversion processes. The argument of the exponential should also be modified in accordance

0-

60-

Figure 2.

j20

r/3

80-

800

rrr-1r1,Il-T,I.lI 1000

Energy (keV)

600

1200

Spectrum of Compton electrons obtained in a typical simulation of the interaction between the 1.17 MeV .y ray of the °°Co and the scintillator (dioxane-naphthalene). The volume of the vial is 10 ml and the radius is 1.25 cm.

400

r1rrir1-zrI-1,rr.1 ullj

200

312

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

with the internal conversion processes, and Equation 9 would become a sum of similar expressions, with the required normalizing factors. If we adopt a general scheme, other 3 emissions can feed the intermediate levels, and some additional gamma transitions will appear. The general expression for the efficiency then becomes a combination of expressions like Equation 9; for instance, in the case of 60Co, 15 expressions must be computed. Its

numerical evaluation is not possible without some kind of factorization; a typical Compton spectrum has about 2000 points and the multiple integral could take excessive time. It may be seen that the integral (Equation 9) can be factorized in terms of some factors which only depend on the individual processes. We have devel-

oped a FORTRAN program which builds up all the possible decay ways, factorizes the integrals, and takes into account the normalizing factors (branching ratios, conversion coefficients, interaction probabilities, etc.).1° It has been used in the efficiency calculations needed in this work. EXPERIMENTAL RESULTS

The method described in this paper has been applied to the standardization of 60Co. This nuclide, which decays by 3-y emission to 60Ni is usually standardized by the method of 4irf3-'y coincidence-anticoincidence. A simplified decay

scheme is shown in Figure 3, and some characteristic nuclear data are presented in Table 2.11 Although there are other transitions, only the most significant have been considered in this study. Materials A standard solution of n-hexadecane 3H from Amersham, U.K., was used as the reference for the efficiency tracing method. The radioactive concentration was 50 kBq/mL and the uncertainty, taken as the addition of both systematic

and random components, was certified to 3%. The 60Co was also from Amersham, and its chemical form was C12Co, 0.1 M. The certified uncertainty was 0.5%.

The scintillation solution was formed by naphthalene 60 gr, PPO 4 gr, Dimetil POPOP 0.1 gr, methanol 100 mL, etilenglicol 20 mL and Dioxane until 1 L. Sample Preparation and Equipment Two sets of seven identical vials were prepared, one for the 3H, the other for the 60Co. Each vial contained 10 mL of the cocktail solution. In each set C14C

was added in 5 L increments to obtain different quench parameters. The radionuclide solutions were added gravimetrically to the vials. The stability was studied over a period of one week and proved to be good.

STANDARDIZATION OF BETA-GAMMA-EMI1TING NUCLIDES

60 27CO

5,272 years

2505,75

1332,52

0

60 28

Figure 3. A simplified decay scheme of the 60Co.1°

Ni

313

2.40

I

3.20

*

Figure of merit

2.80

111111 tliiil lilt,, i

2 00

*

I

3.60

*

I

4.00

*

I

4 40

Figure 4. Measured (single points) and computed (full line) efficiencies in the standardization of 60Co, as a function of the figure of merit.

1.60

85.0

100.0 -

315

STANDARDIZATION OF BETA-GAMMA-EMITrING NUCLIDES

Table 2. Selected Nuclear Data'° from 60Co

Transition

317.9

132

1491.1

71

1173.22 (0.05) 1332.51 (0.05)

72

1

344

2

327

3

306 289 275 262 249

5 6

7

2.1492 2.4262 2.7979 3.1562 3.4985 3.7898 4.2260

99.89 (0.06) 0.10 (0.02)

(0.3) (0.3)

Table 3. Measured and Computed Efficiencies for 60Co Measured Figure Quench Efficiency Parameter of Merit Sample

4

Intensity (%)

Energy (key)

131

99.89 (0.06) 99.993 (0.002)

Computed Efficiency

94.61 94

94.81 44

94.1109 93.1427 92.5480 92.1405 91.1859 90.4283

94.2893 93.5946 92.9340 92.3104 91.7846

Diff (%)

-0.20 -0.19 -0.48 -0.41

-0.18 -0.66 91 .0053 -0.63 Average -0.4%

Measurements were carried out with a LKB RackBeta liquid scintillation counter, which was connected on-line to a personal computer. Results

We present in Table 3 and Figure 4 the results for the measured and computed efficiencies. In the same table, values are also given for the quench parameter and the figure of merit. The differences between experimental and computed" efficiencies, also shown in the table, vary between 0.18 and 0.66%, with an average value of 0.4%. The uncertainties estimated for the method are shown in Table 4. The most important are due to the counting statistics and the quench parameter determination; the contribution of the 3% uncertainty in the 3H standard leads to only a 0.1% in the efficiency of 60Co. We also considered the component due to the nuclear data, and finally, we numerically estimated the influence of the Monte Table 4. Estimated Uncertainties in the Standardization of 60Co by the LSC Efficiency Tracing Method with 3H Uncertainty (%) Source of Uncertainty Liquid scintillation counting of 3H Liquid scintillation counting of 60Co Quench parameter determination Sample preparation Nuclear data Monte Carlo simulation 3H standard Combined uncertainty Overall uncertainty (three times the combined uncertainty)

0.2

0.17 0.1 0.1

0.09 0.06 0.1

0.33% 0.99%

316

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Carlo simulation as 0.06. The combined uncertainty resulted in 0.332Wo and

for the overall uncertainty, taken as three times the combined, we found 0.99o. These values are in good agreement with the average differences found in Table 3, taking into account the uncertainties in the 60Co standard. In conclusion, a method based on the 4rLS efficiency tracing with 3H has been developed which allows the standardization of -y emitters; the application to the case of 60Co has given results which agree well with the values obtained for other methods.

REFERENCES

Grau, A. and E. GarcIa-Toraño. "Evaluation of Counting Efficiency in Liquid

Scintillation Counting of Pure 3-ray emitters," mt. J. App!. Radial. Isot., 33:249-253 (1982).

Coursey, B.M., W.B. Mann, A. Grau, E. GarcIa-Toraño, J.M. Los Arcos, and D. Reher. "Standardization of Carbon-14 by 4ir3 Liquid Scintillation Efficiency Tracing with Hydrogen-3," App!. Radiat. Isot., 37(5):403-408 (1986). Coursey, B.M., L.L. Lucas, A. Grau, and E. Garcia-Toraflo. "The Standardization of Plutonium-24l and Nickel-63," Nuci. fnstr. and Methods, A279:603-610 (1989).

Garcia-Toraño, E. and A. Grau. "EFFY, a New Program to Compute The Counting Efficiency of Beta Particles in Liquid Scintillators," Cornp. Phys. Comm. 36:307-312 (1985).

Coursey, B.M., J.A. Gibson, M.W. Heitzman, and J.C. Leak. mt. J. App!. Radiat. Isot., 35:1103-1107 (1982). Daniell, H. "Shapes of Beta-Ray Spectra," Rev, of Modern Phys. 40:659 (1968). Los Arcos, J.M., A. Grau, and A. Fernández. "VIASKL: A Computer Program to Evaluate the LSC Efficiency and Its Associated Uncertainty for K-L-Atomic Shell Electron-Capture Nuclides," Comp. Phys. Co,nm., 44:209 (1987). GarcIa-Toraño, E. and A. Grau. "EFYGA, A Monte Carlo Program To Compute The Interaction Probability and The Counting Efficiency of Gamma Rays in Liquid Scintillators," Comp. Phys. Comm. 47:341-347 (1987). Grau, A. "Counting Efficiency for Electron-Capturing Nuclides in Liquid Scintillator Solutions," mt. J. App!. Radial. Isot. 33:249-253 (1982). Garcia-Toraflo, E., J.M. Los Arcos, and A. Grau. "MULBEGA: Counting Efficiency for 13-'y Emitters in Liquid Scintillation Counting," to be published in Comp. Phys. Com,n. Lagoutine, F., N. Coursol, and J. Legrand. Table de Radionucleides, (Giff Sur Ivette, France: Laboratoire de Metrologie des Rayonnements lonisants, 1983).

CHAPTER 27

The Standardization of 35S Methionine by Liquid Scintillation Efficiency Tracing with 3H

J.M. Calhoun, B.M. Coursey, D. Gray, and L. Karam

INTRODUCTION

35S (T112 = 87.44 days) decays by beta-particle emission of maximum energy 166.74 keY and average energy 48.60 keY) 35S has been long used for labeling organic compounds for in vitro measurements.2 It is now increasingly used for labeling sulfur-containing amino acids methionine (Met) and cystine (Cys) for protein sequencing studies, as well as in following the progress of amino acid incorporation during protein synthesis. Further potential applications of 35S labeled compounds are in studies of the interactions of sulfur-containing car-

cinogens and mutagens (e.g., sulfur mustards, methyl methane sulfonate, ethicnine, etc.) with DNA, particularly with respect to the nature of the binding of such compounds to the macromolecule. Because of the low energy of the beta particles, 5S is usually assayed by liquid scintillation counting.3 Very Bryant et al., used the few papers are available on the standardization of method of 4irf3-y efficiency tracing with 60Co.4 In previous work we have shown that efficiency tracing with 3H can be used to standardize pure beta-particle emitters of low energy, e.g., nickel 63Ni and 241P1,5'6 and intermediate energy, e.g., 14c.5,6 These measurements were made,

hwoever, under low quenching conditions. The present work was undertaken to see if the method could be used for 35S labeled amino acids under high quench conditions such as are routinely encountered in assaying biological samples. The basic principle of the efficiency tracing technique is to explicitly account for the beta-particle spectra of the radionuclide to be assayed (35S) and the standard nuclide (3H). The counting efficiency for a liquid scintillation system with two phototubes operating in coincidence, , is given by Coursey et al.6

317

318

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

(çE =

mx

P(Z,E) x [1 - exp(-E,Q(E)W(E))]2 dE

x

fE maxP(ZE)dE}

(1)

0

where P(Z,E) dE is the Fermi distribution function, is the figure of merit, photoelectrons per keV, Q(E) is the ionization quenching function to account for the differences in light yield for electrons as a function of energy,7 and W(E) is a wall-loss function, taken here as unity for these low-energy 3 particles.

The liquid scintillation counter is the first efficiency calibrated with 3H

standards and the optimum value for the Q(E) term in Equation 1

is

obtained. The system may then be used to standardize for activity any other radionuclide for which the Fermi spectrum is known. EXPERIMENTAL

Materials and Methods

35S labeled Met (5 mCi/0.5 mL (185 MBq/0.5 mL)) was obtained from Nordion (Ottawa, Canada).* A carrier solution was prepared for all dilutions of the 35S consisting of 10 mM beta-mercaptoethanol, 50-mM N-[tris (hydroxymethyl) methyl] glycine (tricine), and 0.1 mil'1 stable Met. The 3H water was a dilution of NIST Standard Reference Material (SRM) 4927C. Each scintillation vial contained 10 mL Beckman Readysolv HPb scintillator in a polyconeseal glass vial (Kimble). Two sets of identical vials were prepared; one for the 35S and one for the 3H standards. For each set, chloroform was added in 50 tL increments (0 to 200 jL) to simulate the quenching expected in biological samples. The radionuclide samples (27 to 82 mg) were added gravimetrically to the vials containing scintillation cocktail. Equipment

Measurements were made with a Beckman LS7800 liquid scintillation counter on line to a Charles River Data Systems supermicro data aquisition system.8 Data were downloaded to an IBM XT PC for processing. Efficiencies *Mentjon of commercial products does not imply recommendation or endorsement by the National Institute of Standards, nor does it imply that the products identified are necessarily the best available for the purpose.

319

THE STANDARDIZATION OF 35S METHIONINE

and spectra were computed using the program EFFY2 implemented on the NIST PC by Eduardo GarcIa-Toraflo in l988. RESULTS AND DISCUSSION

The samples were measured over a period of 30 days. The 35S-methionine proved to be very stable (less than 0.1 Wo change), while for the 3H samples, the

count rate decreased by approximately 0.16o per day.5 The tritiated water separates from the scintillator with time, while the 35S labeled Met is apparently in the organic phase. Figure 1 shows the Fermi spectra of 35S (Figure la) and 3H(Figure ib). The measured LS spectra for the two nuclides for low quenching (Horrocks' H# around 70) and high quenching (H# around 150) are shown in Figures ic and id for the 35S and 3H, respectively. The low quenching samples were used to standardize the 35S by the efficiency tracing technique. Figure 2 shows the relationship between the figure of merit, the quench parameter H#, and the counting efficiencies of 3H and 35S. Although the 3H efficiency decreases with quenching from 49 to 24Wo, the 35S efficiency only

3 =

(a)

Ec)

2 =

2

I 157 key

-

0

0

3

(b)

=

(d) 3

2

I

H

18.6keV

0

0

E (keV)

Fermi Spectra Figure 1.

500

0 1000

Channel No. Beckman LS 7800

Beta particle spectra and liquid scintillation spectra for S and 3H. The experimental spectra was obtained on a system with a logarithmic amplifier.

Figure 2.

0.86

0.25

I

I

i

0.87

160

i

0.30

I

0.20

I

0.88

i

140

i

120

I

0.89 0.90

i

0.35

I

0.25

0.40

I

0.91

100

I

0.30

0.45

I

0.92

80

I

0.35

0.50

0.40

F

Relationships between the figure of merit (tj), the quench paremeter (H#), and the two-phototube coincidence counting efficiencies for 3H and S. The shaded area is the unquenched region used for the most accurate standardization of the S. This nomograph is strictly valid for only one scintillator and counting system.

c(35S)

(3H)

H#

11

60

THE STANDARDIZATION OF 35S METHIONINE

321

Table 1. Results for Nordlon 35S Methionine as of 1200 EST February 24, 1989 (uncertainties are given as one standard deviation)

Predicted Activity Vial No. 1

2 3 4 5

(Bq/mg)

H# 71.3 70.5 71.4 72.5 71.6

69.43 69.15 69.14 69.34 69.23

0.9256 0.9261

0.9255 0.9245 0.9253

69.26 ± 0.12 6 7 8 9 10 11

104.7 118.1

131.4 141.0 152.8 171.3

69.50

0.9028 0.8940 0.8854 0.8793 0.8719 0.8604

69.81 69.51

69.45 69.25 68.75

69.38 ± 0.36

drops from 92 to 87Wo over the same chloroform concentration range. The predicted values of the activity for all samples are shown in Table 1. The average for the quenched samples, 69.38 ± 0.36, compares favorably with the value obtained for the unquenched samples 69.26 ± 0.12 Bg/mg (uncertainties are one standard deviation). The estimated uncertainty in the 35S activity concentration, which includes systematic as well as random components, is given in Table 2. Since this standardization method can be extended to high quenching for 35S, it should work as well for 14C and 32P. Figure 3 shows computed counting efficiency for all three radionuclides and 3H as a function of quenching. Using

the "measured 3H efficiency" one can compute the efficiency of any of the other three to within ito 2%. Table 2. Estimated Uncertainties in the Standardization of 35S Methionine Percent (%) Liquid scintillation measurements 3H reference beta-particle standard Quenching in the liquid scintillation measurements Source preparation Uncertainty in efficiency curve fit Scintillator stability Uncertainty in numerical spectra integration Dead time

0.18 0.21

0.05 0.10 0.07 0.05 0.05 0.03

Combined in quadrature

0.32

Overall uncertainty (x 3)

1.0

Figure 3.

H

Counting efficiency as a function of quenching, with experimental points from this work. Curves computed using code EEFY2 (Grau Malonda and Garcia-Torao (1985)).

H#

3

35s

140

32p

THE STANDARDIZATION OF 35S METHIONINE

323

CONCLUSIONS

35S labeled methionine has been standardized by the method of 4irfl liquid scintillation efficiency tracing with 3H. The method appears to work well for chemically quenched samples up to at least H# 171. Since this program is now

implemented on a personal computer, it should be possible to adapt most commercial liquid scintillation counters to directly compute 35S activities, providing a set of 3H quenched standards is available to establish a quench curve vs a quenching parameter.

REFERENCES

Handbook of Radioactivity Measurements Procedure 58, Second ed. (Bethesda, MD: National Council on Radiation Protection and Measurements, 1985). Gordon, B.E., H.R. Lukens, and W. ten Hove. mt. J. App!. Radiat. Isotp. 12: 145 (1961).

Horrocks, D.L. Applications of Liquid Scintillation Counting, (New York: Academic Press, 1974).

Bryant, J., D.G. Jones, and A. McNair. in "Standardization of Radionuclides," proceedings of a symposium, International Atomic Energy Agency, Vienna, (1967).

Coursey, B.M., L.L. Lucas, A. Grau Malonda, and E. GarcIa-Toraño. Nuci. Instrum. Meth. Part A. in press, (1989). Coursey, B.M., W.B. Mann, A. Grau Malonda, E. GarcIa-Toraflo, J.M. Los Arcos, J.A.B. Gibson, and D. Reher. mt. J. App!. Radiat. Isopt. 37, 403 (1986).

Grau Malonda, A. and E. GarcIa-Torano. mt. J. App!. Radiat. Isopt. 33: 249 (1982).

Coursey, B.M., M.P. Unterweger, L.L. Lucas, and E. GarcIa-Toraflo. Trans. Amer. Nuci. Soc. 55: 54 (1982). Grau Malonda, A. and E. GarcIa-Toraño. Comp. Phys. Co,nm. 36:307 (1985).

CHAPTER 28

A Review and Experimental Evaluation of Commercial Radiocarbon Dating Software

S. De Flllppis and J. Noakes

ABSTRACT New commercial radiocarbon dating computer software has been developed to provide online data analysis for Packard computer assisted counting systems and off-line data analysis for other instruments. Because different radiocarbon laboratories may have slightly different methods of calibration and data presentation, the software was designed based on the routine methods and calculations performed at the Center for Applied Isotope Studies (CAIS). Input from several other radiocarbon laboratories was also incorporated in the design of the computer software. This chapter presents a complete product critique, from a users' point of view, covering ease of use, calculations, and future applications of radiocarbon application software. Comparative data will be presented to show the differences between this radiocarbon software data analysis and the routine data analysis that is performed at the Center for Applied Isotope Studies. The software is menu driven with on-screen programming and editing for a professional user interface environment. It also has the Capability to save and analyze counting data directly from the Packard liquid scintillation system. The software can be programmed to accept oxalic calibration information and count data. The software data disk management develops count history files of a carbon reference standard and background over time. The user can archive data into a unique data subdirectory. Data can be selected to export collected and stored information for incorporation into other computer software such as LotusRi ,2,3 or database programs. The software offers several correction features, including an account of benzene evapora-

tion and Delta 13C isotope value, and it can be programmed for the scintillation counter radionuclide efficiency.

INTRODUCTION

Laboratories that specialize in low level radiocarbon measurements as a part of their own research programs or as a commercial service to the scientific community have had to develop in-house computer software to automate these routine calculations. A question often raised by new laboratories that for the first time enter into low level radiocarbon measurements concerns the analysis and interpretation of the counting data. In many cases this question remains 325

326

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

unanswered until a literature search through the annals of RADIOCARBON uncovers the commonly used calculations and methods of correction for radio-

carbon age dating. It is not uncommon to discover a startup radiocarbon laboratory that has limited knowledge of these calculations and how to apply the correction schemes. The application of laboratory computers used in routine data analysis for a

specialized task like radiocarbon age dating has, in the past, required a programmer/scientist to develop and test the software used by the laboratory. In the absence of computer software, hand calculations would suffice. When we learned that specialized application commercial software was being developed for radiocarbon age dating analysis, we became interested in evaluating and comparing the results of the new software package to our own radiocarbon analysis program. BACKGROUND

The radiocarbon software was developed by Packard Instrument Company

of Downers Grove, IL, for use with the computer assisted, as well as the noncomputer automated, line of liquid scintillation counters that the company manufactures. At the time of this independent product evaluation, the software has not been released for commercial distribution. The software provided

for our studies was a prerelease version. The software as we know it was intended for on-line data analysis using the Packard TriCarb® series models 2260, 2250, 2200, 1050, 1000 liquid scintillation analyzers. Off-line data entry

and subsequent analysis does not seem at this time to be an option of this software. The liquid scintillation analyzer used in our evaluation was a Packard Tri-Carb model 2050, predecessor of the model 2250, both are specially designed low background counting systems. EVALUATION CRITERIA

This product evaluation was conducted solely on the basis of an individual end user. The software was evaluated in five areas of interest; (1) installation and documentation, (2) ease of use, (3) correctness of calculations, (4) user benefit features, and (5) future applications. In order to compare results obtained from the new Packard software to the data analysis performed at CAIS, several wood samples from the radiocarbon laboratory at the University of Waikato in New Zealand were dated using both computer programs. The samples of wood ranged in age between 2,000 years before present (YBP) to approximately 50,000 YBP. All sample dates were corrected for carbon isotope fractionation. The opinions expressed here are those of the authors and do not represent those of Packard or the University of Georgia.

COMMERCIAL RADIOCARBON DATING SOFTWARE EVALUATION

327

MATERIALS AND METHODS

Each wood sample was prepared for liquid scintillation counting by first cutting away the surface related wood material and then successively pretreating it with chemicals to remove possible contamination from modern carbon components. The treated wood sample was processed to pure benzene using a benzene synthesizer. A 3 cc sample of benzene was derived from wood combustion and mixed with scintillators; counting followed. The samples were first counted using the CATS scintillation counters, nor-

mally used to process routine samples. Each sample was counted for 2700 minutes in a Picker/Nuclear Liquimat 220 with a specially designed low background copper/Teflon®* shielded counting vial, and then the age date was calculated. Originally, the counting experiment was designed for the CATS system and the Packard software to be similar. It was later discovered that the Packard software was designed to count each sample once, unload, and cycle it around the sample changer before counting for a repeat time. To minimize statistical variations due to cycling, each sample was counted for a single 999 mm interval, and then the date was computed using the Packard software. Normally at CATS, each sample, background, and oxalic standard is counted in the same counting vial because each liquid scintillation counter is calibrated with its own vial to minimize experimental variations. Using the Packard system, the software is designed to count the sample in a multi-user environment. This means

that the investigator should place the samples in a cassette in a desired sequence. This sequence can be programmed by the user into the software, so the software can recognize the difference between a standard, background, oxalic, and sample. However, this multi-user environment does not lend itself

to counting each radiocarbon sample, oxalic, or background in the same counting vial. To accomplish this, the investigator must manually intervene to stop the counting system, prepare a new sample in the same vial, and start the counting again. This system works best when using a different vial for each sample. The counting vial used with the Packard counting system was a standard low 40K borosilicate glass vial with an internal volume of 7 cc. Teflon cap liners were used to prevent solvent loss during the counting interval. HARDWARE AND SOFTWARE REQUIREMENTS

The software was developed in the "C" programming environment and runs

on an IBM Personal Computer (PC) or Professional Series (PS) microcomputer system and compatible. The computer in any case must be compatible with the interfacing to the Packard scintillation counter and the executing of the scintillation software. The radiocarbon application software is an addon program that runs within the scintillation software operating environment *Teflon is a Registrated Trademark of E.I. DuPont de Nemours

328

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

on computer assisted counting systems. Using the Packard DatalinkTM stand alone counting systems like the Tri-Carb model 1050 the investigator can compute age dates with the radiocarbon software. No additional computer equipment, other than what is typically provided with each Packard counting system, is required for operation. During the software installation, the color option can be activated for use with top of the line Packard couting systems, which are equipped with a color monitor. Installation and Documentation

Since the software provided for evaluation was a prerelease version, the documentation supplied was not the information that would accompany the commercially released product. At this time, we cannot describe the quality or completeness of this material. The program installation was complete. The original disk, provided in the 5.25 in. floppy format, contained an installation routine that prompted the user for the different Packard counting system models. Since the Packard system used was an earlier vintage, the use of subdirectories was not supported. The radiocarbon software was then installed to the root directory on the hard disk, which contained all other files used to operate and communicate with the scintillation counter. Newer models of Packard counters support the use of subdirectories, making the house keeping of different files of a given software package much easier. Ease of Use

Once installed, the software was very easy to access through the Packard scintillation software user interface. The radiocarbon program was developed

as a menu driven system with preprogrammed function keys for different operations and modules. The Packard operating environment allows application software, like the radiocarbon program to execute and run while the scintillation counter was performing the most recent instruction from the scintillation software. If the scintillation system was counting a sample when the user executed the radiocarbon software, the sample would continue to count

until its time or statistical termination was satisfied. At that point, if the scintillation program was not returned to the main portion of memory, the counting system would wait until the user terminated conversation with the application program before the counter would go on to the next instructed task. A safety feature of the system is that it can sense keyboard activity. If the user executes the application program and is called away from the computer, the program will automatically terminate, saving the most recent information and returning to the operation of the scintillation counter within a preset period of time. The use of eight preprogrammed function keys enables the user to navigate through the system as shown in the flow chart (Figure 1). There is a help line at

Weight Final

Archive Protocol

4,

CPM C14 Check Source

Quench Parameter (tSl E,CPMA/CPMb)

Reference Year

Modern C14 CPM

Modern C14 DPM/Gram C

C-14 %Efficiency

Sysgen

Weight Initial

Protocol

Delete

Evaporation Correction

Restore

Sample l.D.

Process

4,,

On-Line Data Analysis

Delta C-13

Delete

Edit Count Data

Sample #

Edit Sample Data

Figure 1. Radiocarbon software flow chart.

Exit

4,

Main Menu

Acess from LSC Software

Restore Protocol

330

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

the bottom of each screen to provide expected responses for software prompts. Once the counting protocol in the radiocarbon software has been predefined by the user, all that is required is counting the sample. The subsequent data analysis is automatic, a printed report follows. The protocol defined in the radiocarbon software must be identical to the counting protocol on the scintillation counter. The radiocarbon software can be user programmed through the scintillation software to process the sample data after each sample is counted (Sample Mode) or after a contiguous group of samples is counted (Batch Mode). However, in Batch Mode, the radiocarbon data analysis and report are not generated until the logical end of the group. So, if each sample is programmed to count for 1000 minutes, it could be some time before the radiocarbon age of the first sample is reported. Overall, the software can automatically calibrate, based on the user providing the correct oxalic reference information, weight of the sample in terms of benzene, and Delta 'C data. The user can edit previously counted data onscreen, should there be a need to delete or restore data. In addition, the user

can achieve logical groups of raw sample count data and also export that information to a selected disk drive as an ASCII file for later import into other software packages. SOFTWARE FEATURES

System Generation

One of the initialization functions the software provides the user is the ability to enter the externally computed '4C counting efficiency, oxalic acid specific activity, and the count rate of a check source. Other information regarding the correction of benzene evaporation is activated within Sysgen. The Sysgen screen allows the user to select or enter their own specific activity for the oxalic acid standard. Presets were available for oxalic acid I (SRM 49) of 14.27 DPM/gr C and oxalic acid IL (SRM 4990-C) of 18.46 DPM/gr C. The software was designed to automatically compute updated counting efficiency information, check source CPM, and oxalic acid CPM when these samples were identified by the user and counted. This screen appears to be simple in design, but the prompts were somewhat vague and should be explained in the final documentation. The software does, however, allow the user to enter a nondefined specific activity for the oxalic acid in the event that the user is not calibrating on oxalic acid, but possibly some other reference material. Edit Sample Data

The radiocarbon software provides lets user program each sample identification with a sample number corresponding to the position of the sample in

COMMERCIAL RADIOCARBON DATING SOFTWARE EVALUATION

331

the cassette, the Delta 'C value, and the weight of sample benzene. When the evaporation question is activated through the Sysgen menu, another column of data appears on this screen, allowing the user to enter the final sample benzene weight. In order to correct the previous data for the loss of benzene during counting, the original count data must be manually reprocessed. The sample specific activity is corrected for the sample benzene loss by taking half of the difference between initial and final benzene weights. As a result of benzene loss,

the percent difference between the initial and final weights are

computed. Once the sample data have been entered, the user can initiate counting and

wait for the final results. This module of the software allows the user to program on what given date the calculation should begin. Edit Count Data

Once the samples have been counted, a count history file is generated for each sample. When a particular sample is counted repeatedly, the total counts for each cycle are combined and a correlation value computed. If the sample has counted for more than 5 cycles, an additional correlation value over the most recent 5 cycles is computed. This provides the user with information regarding the reproducibility of the count results. If there is any bias in the data, the counts/day are computed and printed. In this module the user can selectively delete one or more count repeats. The program flags the deleted point(s), but does not discard the information. If the user wishes to restore the deleted point(s), the Restore function can remove the

delete flag so that the sample data can be incorporated into the final calculations.

COMPARATIVE RESULTS

In order to compare the radiocarbon age calculations between the CATS routine method and the new Packard software, six different wood samples were processed to benzene and counted. These samples ranged in ages between approximately 2,000 YBP and 50,000 YBP. Also a modern oxalic standard was measured to see how the software would handle the 136% of modern sample. The benzene samples were measured in the standard CATS counting vial when

the sample was counted using the CATS LSC spectrometers. The Packard software was designed to count each sample once and then cycle the sample, enabling the instrument to count another sample before repeating a given measurement. Because of this limitation, the same counting vial could not be used for the standard, sample, and background as was accomplished in the CATS routine. Each benzene sample was transferred from the CATS counting vial into a common 7 cc low 40K glass vial. The comparative results are presented in Table 1. Tt was apparent that on the

332

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 1. Radiocarbon Age Dates-Charcoal Samples CAIS vs Packard Software Sample 1.0. UGA 5894 5897 5891

5903 5900 5888

Oxalic-4990C

13C

-25.81

25.85 24.96 24.67 24.83 25.56 17.68

CAIS

PACKARD

YBP ± la

YBP ± icy

1,974 ± 50 2,036 ± 52 3,232 ± 419 5,058 ± 65 49,033 ± 7569 >50,000 Modern

% DIFFERENCE

2,019 ± 50 1,894 ± 50 2,894 ± 50 4,695 ± 50 38,635 ± 50

6E38 2393 ± 50 (Modern)

2.3

7.0

10.5

7.2

21.2

-

basis of counting these radiocarbon samples there were differences between the CATS method and the Packard software. Most of the age dates computed automatically with the Packard software demonstrated a negative bias ranging from approximately 2 to 21% from the CATS corresponding values. The software was executed, with no correction schemes activated, to yield computed age dates that were not influenced by any possible errors in the code. Upon discussing the results, Packard is still in a program developmental stage. CATS endeavors to assist the program developers at Packard in identifying the program errors so this commercial software package will carry the full merit of radiocarbon age dating capability. GENERAL COMMENTS

At present, the Packard software is in need of further debugging and testing before it is ready for commercial distribution. We believe that new startup radiocarbon laboratories will probably benefit the most from this on-line software analysis. Established laboratories such as CAIS and others around the world, most likely will contine to utilize their own developed radiocarbon software for computation of age dates. Since this software is closely related to the equipment of one particular manufacturer, it does provide the capability for new laboratories to obtain the counting equipment and analysis software from a single source and it is ready to use. During our test and evaluation period, we had discovered several bugs in the program, some which prevented us from continuing with our evaluation. Packard responded quickly with programmer support to correct the apparent problems and provided us with another version. The program source code was not made availabe to us for modification, but we were pleased with the prompt attention from Packard. FUTURE APPLICATIONS

During the recent year, other radiocarbon laboratories have developed radiocarbon age date and radiocarbon correction type application software.

COMMERCIAL RADIOCARBON DATING SOFTWARE EVALUATION

333

The Centrum Voor Isotopen Onderzoek in Groningen, the Netherlands, under Professor Van de Plicht, has made available to the radiocarbon community computer software that will provide tree ring corrections to radiocarbon dates. This software was developed in the Pascal language. Recently, a computer program named "CAL!" was published by the Laboratorio de Datacion por Carbono-14 at the University of Granada in Spain. This program was developed similarily, to provide a calibration of radiocarbon age dates. Although at CATS, we have not had an opportunity to fully evaluate and compare these newest radiocarbon software packages, there appears to be a growing interest among different laboratories in the development of computer software for this specific purpose. We believe that the direction many radiocarbon laboratories are considering is one that will completely automate the radiocarbon sample calibration pro-

cess. Ultimately, once the sample is loaded into the counter, the computer software that controls the scintillation counter or acquires the data from a stand alone counting system will capture the counting data from the equipment, compute the radiocarbon age of the sample, and provide the different types of correction schemes. At present, the Packard computer-assisted count-

ing systems can execute user-programmed software to further analyze the counting data. The on-line radiocarbon software, once fully operational, will provide radiocarbon calibration of samples counted, but it does not completely include all of the different correction schemes. The next logical step in the sequence of software development will be the linking of on-line radiocarbon software to the other programs developed to correct the calibration dates for a complete data presentation.

CHAPTER 29

Efficiency Extrapolation Adapted to Liquid Scintillation Counters

Charles Dodson

ABSTRACT Efficiency extrapolation applied to fi emittors has been adapted previously to conventional liquid scintillation counters. This concept has been extended to cover broad quench ranges, it has also been extended to cover tritium with a small decrease in accuracy compared with a priori approaches.

INTRODUCTION

The disintegration rate of an unknown 13-emitting nuclide may be obtained by using a standard 13-emitting nuclide to determine the overall system effi-

ciency of a counter. Initial work by Gibson and Gal&'2 was followed by Malonda and colleagues,3-5 and used by Coursey et al.6 who reviewed it, as did

NCRP Report Number 58. This approach computes the detector efficiency for a two-phototube coincident counting system. An efficiency curve of a standard is measured experimentally in terms of a selected quench parameter (Q). That permits calculation of the corresponding figure of merit (M) of the system via the Fermi 13-ray spectral distribution for the standard. The distribution is corrected for two energy exchange processes that do not contribute to photon production: ionization quenching and the wall effect. These corrections provide a figure of merit independent of the detectable energy of the counting system. Therefore

the figure of merit can be expressed as a function of the selected quench parameter:

M = f(Q)

(1)

With that independence, the efficiency, (unk), of an unknown beta emitter can be computed from the Fermi spectral distribution corrected for ionization quenching and the wall effect: 335

336

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 1. Summary of Variables Studied Isotopes: 3H, 63N1, 14C,

Na, 36C1, 32P, 241Am

Cocktails: Toluene, Xylene and Pseudocumeme based gel, Non-emulsifier, Biodegradable Volume: 0.4-16 ML. Count rates: 1000-6 Million Quench range: 0-475 H# Reference standards: 3H, 14C 3H, 63N1, 14C, CI, and 241Am (NIST - SRMs) 35S, 2Na and 32P Not as well characterized monitored by quench curves

e(unk) = f(M)

(2)

e(unk) = f(Q)

(3)

From Equations 1 and 2,

is available. The quench curve of the unknown has been calculated from the

experimentally measured quench curve of the standard. Details of this approach are available in References 3 through 7 and their bibliographies. An experimental approach initiated by Ishikawa and collegues8-1° calibrates the counting system efficiency by determining the efficiencies of a standard for six defined pulse height regions. The count rates of an unknown are obtained for the same spectral regions as the standard. This provides integrated (or cumulatively summed) count rates of the unknown as a function of the system integrated efficiency determined by the standard. The resulting linear or quadratic functionality may be extrapolated to 100% efficiency to provide disintegrations of the unknown. Good results have been obtained for reasonable quench ranges. EXTENDING THE QUENCH RANGE

First examine the use of the complete spectrum of the standard for system calibration and the complete spectra of unknown nuclides for count rate inte-

gration. Figure 1 illustrates one result based upon '4C as the standard and quenched samples of 35S as the unknowns. The counting efficiency range of the

four 35S samples taken from the same quench set is 60 to 95%. Figure 2 provides analogous data for 32P and 36Cl. Depending upon the quench level of the sample, the specific nuclide, and the resolution of the multichannel analyzer (MCA), tens to hundreds of channel data are available. A 4096 channel MCA provides about 300 channels for an unquenched '4C spectrum.

Several hundred samples have been monitored. The variables examined (nuclides, cocktails, sample volume, count rates, quench levels, and quenching agents) are summarized in Table 1. Various curve fitting and extrapolation algorithms led to an iterative procedure which produces a least squares linear fit. It is subsequently extrapolated as illustrated in Figure 3. Criteria defining linearity are important but not hypercritical. Results reported here are based

Figure 1.

z

F-

CJ

ILO

UJO

o

-

0

02

1E

04

Integrated 35S Spectra vs Integrated Standard 14C Counting Efficiency.

0

10

20

30

40

50

60

70

80

90

100

110

120

06

08

240

0

02

1E

04

06

Integrated 32P and 36C1 Spectra vs Integrated Standard 14C Counting Efficiency.

40

60

80

100

120

140

160

180

200

220

Figure 2.

F-

U0

260

280

300

320

340

360

380

400

08

0

0.04

0.08

0.16

(1E) OF STANDARD

0.12

02

Integrated 14C Spectra vs Integrated Standard 14C Counting Efficiency.

0

10

20

30

40

50

60

70

80

90

100

110

120

130

Figure 3.

z

IdE-

U

140

150

160

170

0.24

0.28

0.32

0

02

04

1E

06

Integrated 3H Spectra vs Integrated Standard 3H Counting Efficiency.

0

100

200

300

Figure 4.

LUH

Lo

Ido

a,. Ov,

400

500

08

EFFICIENCY EXTRAPOLATION

341

Table 2. Results For H# ranges of 0-350:

0.1 to 3.5% (except H3)

For H# range of 0-150 for H3 (65_30% counting efficiency)

0.1 to8.1%

upon the square of the correlation coefficient (or coefficient of determination) exceeding 0.99. The three samples shown in Figure 3 were taken from a standard quench set containing 2470 Bq (148000 DPM). The raw data are represented by the thick line, the segment satisfying the linear constraint and subse-

quent extrapolation is the narrow line. These samples cover a counting efficiency range of 55 to 95% with an error in DPM recovery less than 2%. APPLICATION TO TRITIUM

Direct application of the procedure to tritium provides useful results if the quench level remains small. For example, % error in DPM recovery for NIST SRMs with counting efficiency above 45% (or Horrocks' numbers less than 75) is less than 4%. However, if tritium unknowns are measured after calibration

by tritium standards, recoveries better than 8% are obtained for counting efficiencies of 30%. A summary of the results is provided by Table 2. The range of the error in DPM recovery is 0.1 to 3.5% for all nuclides except 3H for the volume, quench ranges, cocktails, and count rates presented in Table 1. For 3H monitored for all the same variables but over a quench range of 65 to 30%, the error range is 0.1 to 8.1%.

Lastly, ongoing work has established that recoveries better than 4% are possible for 14C with efficiencies down to 20% (or Horrock's numbers of 450) if the system is calibrated by standard 3H. REFERENCES

Gale, H.J. and J.A.B. Gibson. J. Sci. Instrument., 43:224 (1966). Gibson, J.A.B. and H.J. Gale. J. Phys., El:99 (1968). Malonda, A.G. and E. Garcia-Toraño. mt. .1. App!. Radial. Isot., 33:249 (1982). Malonda, A.G., E. Garcia-Toraflo, and J.M. Los Arcos. Int. J. App!. Radial. Isot., 36:157 (1985). Malonda, A.G. Int. J. App!. Radial. Isot. 34:763 (1983). Coursey, B.M., W.B. Mann, A.G. Malonda, E. Garcia-Toraflo, J.M. Los Arcos, J.A.B. Gibson, and D. Reher. App!. Radial. Isot. 37:403 (1986). Mann, W.B., Ed. NCRP Report 58, A Handbook of Radioactivity Measurements Procedures, 2nd edition, (Washington, DC: NCRP Publications, 1985) p. 199. Ishikawa, H., M. Takiue, and T. Aburai. mt. J. App!. Radial. Isot. 35:463 (1984). Takiue, M. and H. Ishikawa. Nuci. Inst. Meth. 148:157 (1978). Fujii, H., M. Takiue, and H. Ishikawa. App!. Radiat. Isot. 37:1147 (1986).

CHAPTER 30

Applications of Quench Monitoring Using Transformed External Standard Spectrum (tSIE)

Michael J. Kessler, Ph.D.

ABSTRACT The transformed external standard spectrum is an effective quench monitor in quantitating samples in liquid scintillation analysis. It uses the Compton spectrum from the gamma source, 133Ba, to monitor sample quench. This technique uses tSIE as an accurate method of quantitating the DPM in radiolabeled samples; it is shown to produce accurate DPM. The use of tSIE is an accurate method of monitoring quench in a sample even under the following conditions: a large dynamic range of tSlE (quench) a large range of sample volumes independence of wall effect, cocktail type and quenching agent DPM determination without substantially increasing the count time of the sample

The use of tSIE parameter provides an accurate method for quench correction in the liquid scintillation analyzer. In addition, an automatic method (AEC) for adjusting the counting regions assists in DPM determination for both dual and triple labeled radioisotope samples.

INTRODUCTION

To understand the use of external standards for quantitating DPM in a radiolabeled sample, it is important that the basic principles of liquid scintillation process be understood. The entire process is to convert the energy from the beta particle into photons of light. This conversion must be efficient, and the intensity of the photons of light must be directly proportional to the energy of the beta particle. This process is illustrated in Figure 1. The process of liquid scintillation involves the transfer of the energy of the beta particle to the solvent. Most all solvents used today for the liquid scintillation process contain an aromatic ring structure with ir electrons. The solvent is chosen because it efficiently transfers the energy from the beta particles to the electrons of the solvent. The second step in the process is to transfer the energy from the activated solvent to a scintillator. This scintillator molecule becomes 343

344

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

xl

hv

SOLVENT'SCINTILLATOR

RADIONUCLIDE

CHEMICAL QUENCH

- (PMTIJ

COLOR QUENCH

Figure 1. The basic liquid scintillation process.

excited, and when it returns to the ground state, it emits photons of light which are directly proportional to the energy of the beta particle. The photon (intensity of light) is then converted to a voltage pulse by the photomultiplier. The intensity or pulse height will be dependent on the number of photons entering the photomultiplier. These voltage pulses (analog in nature) are then converted to a digital pulse for direct analysis by a multichannel analyzer (4096 channels) over the analysis time of the sample. The entire energy range which is analyzed, is 0-2000 keY. This is the energy range of all beta-emitting radionuclides. The beta particle decay is a process in which a neutron from the nucleus is converted into a photon, an electron (beta particle), and a neutrino (mass-like particle). The energy loss in the decay is distributed between the beta particle and

the neutrino. Thus, the beta particle can have any energy from zero to the maximum energy of the beta particle. For example, for '4C, an energy from 0-156 keV can be seen in the spectrum. This is illustrated in Figure 2; it is a typical energy spectrum of an unquenched '4C spectrum using the Spectralyzer of the Packard 2500TR liquid scintillation analyzer. The x-axis is a linear plot of the energy or light intensity of the beta particles, and the y-axis is the number of counts/unit time that a pulse of that intensity has been seen by the photomultiplier tube. If the area under the spectrum is integrated, then the total amount of

radioactivity present in the sample, as measured by the counter (CPM) is obtained. The CPM value measures of how efficiently the energy transfers from the beta particle to the solvent, and how effeciently it transfers to the scintillator. It measures how efficiently the scintillator produces light, and how efficiently the instrument electronics convert the photons to voltage pulse. All of the steps

together determine the sample CPM seen by the system. The CPM value is measured and compared to the actual DPM in the sample (obtained from the Ci in the sample). The efficiency of counting the sample is then calculated by the following equation: cpm

dpm

< l00o = counting efficiency

The difference between DPM and the CPM measured by the liquid scintillation analyzer, is a result of a process termed quenching. This quenching phenomenon is involved in the scintillation process at two points: energy transfer and light quantitation. The first process of quenching (the energy transfer) is called chem-

APPLICATIONS OF QUENCH MONITORING USING tSIE

Figure 2.

345

Energy spectrum of unquenched 14C sample on Packard 2500TR liquid scintillation analyzer.

ical quenching. It involves the energy transfer of the beta particle to the solvent, or the transfer of energy from the solvent to the scintillator. Common chemical quenching agents (H20, nitromethane, CHCI3, CCI4, etc.) reduce the transfer of energy from the beta particle to the solvent. The quenching agent absorbs or reduces the energy of the beta particles transferred to the scintillator. The second process of quenching (photon reduction) is called color quenching. This phenomenon reduces the intensity of the scintillator produced photons seen by the

photomultiplier. This phenomenon is similar to the use of a color filter on a photographic camera, which filters out certain colors of light on the photographs. The color in the scintillation process reduces the intensity of the photons seen by the photomultiplier. A typical spectrum for a '4C sample, quenched by chemical and color quenching agents, is demonstrated in Figure 3.

The results of these quenching processes on the CPM can be extremely variable depending upon the quenching in the sample. For example, a sample prepared from a binding study may contain 2000 CPM, but a similar sample counted in solution might give 5000 CPM. In order to be able to compare these two samples, it is necessary to compare the DPM values (CPM compensates

346

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Figure 3. Chemical and color quenched spectrum.

for quenching). Three methods of quench correction can be used to obtain accurate DPM values: internal standard, sample spectrum, and external standard spectrum analysis. Transformed External Standard Spectrum-tSIE Most of methods for DPM determination involve the preparation of a quench curve (o efficiency vs quenching level) for a set of standard quenched vials. The most accurate method of determining DPM presence in the sample is by using the external standard spectrum technique. This technique involves exposing each sample to an external gamma source ('33Ba). The gamma radionuclide creates a

Compton spectrum by way of the Compton scattering phenomenon. The gamma ray interacts with an electron to create a new gamma ray with less energy and a Compton electron. This Compton electron is similar to a beta particle and creates a Compton energy spectrum, Figure 4, in the LSA. The Compton spectral distribution of 133Ba can be used to monitor quenching

in the sample. The Compton spectrum is stored in the Spectralyzer so that various features of this spectrum can be monitored. After close evaluation of the

Compton spectrum of various quenched samples, it was apparent that this spectrum could not be used directly as a measure of quench; volume, wall effect,

APPLICATIONS OF QUENCH MONITORING USING ISlE

Figure 4. Compton spectrum of

347

Ba-Iinear.

and vial size changed the quenched indicating parameters in the Compton spectrum. After further investigation, it was found that if the integral spectrum of the '33Ba external standard spectrum was computed, the end point of this trans-

formed external standard spectrum (tSIE) could be used to measure sample quenching. The tSIE decreases as the quenching of the sample increases. This change of the tSIE with increasing quenching is illustrated clearly in Figure 5. This shows the Compton spectrum of two samples of different quench levels which have been transformed. The external standard end point of each sample is shown. This external standard end point is multiplied by a constant, such that the tSIE for an unquenched, argon purged '4C sample is equal to 1000. The tSIE

(quench indicating parameterQIP) can be used as a method of relating the o efficiency of a standard quench set as a function of quenching the sample. This plot of We efficiency can be used as an accurate method of obtaining DPM for samples with unknown DPM. This first requires the preparation of a standard quench curve (Wo efficiency vs tSIE) with a sealed set of standards containing the same radionuclide that is present in the unknown. Once each efficiency is detersample in the standard quench set has been counted, the

348

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

100

200

300

400

500

00

700

800

900 13

tSIE (Transformed Spectruti) Figure 5. tSIE quench curves for tritium and 14C.

mined. This is done by counting each sample and obtaining its CPM. Since the quenched set was made with a known amount (DPM) of radionuclides of interest, the Wo efficiency can be calculated by the following equation: cpm dpm

lOOWo = Wo efficiency

This is the y-axis number for each sample. The x-axis values are determined next. This is done by exposing each standard in the quenched set to the external standard ('33Ba). The Compton spectrum of each sample is then recorded in the Spectralyzer spectrum analyzer. This is converted to an integral spectrum. The tSIE value is determined from the extrapolated end point of this transformed spectrum. The tSIE value is then used as a measure of the quenching in the sample (QIP). Once the tSIE and We efficiency for each of the standard quench sets has been determined, it is plotted on a graph. The curve fit used to connect the points is a fixed point least square quadratic (FPLSQ). A typical quench curve for '4C and 3H is shown in Figure 5. In summary, the following steps are used to prepare a quench curve for determining DPM: prepare standards (two different methods can be used): (1) purchase sealed standards containing radioisotope of interest, (2) prepare and analyze set of standards of known DPM with different levels of quenching agent, and use a cocktail similar to that used in unknown sample

count standards in defined counting region determined by evaluation of sample spectrum of least quenched sample count sample to determine CPM for each standard

APPLICATIONS OF QUENCH MONITORING USING tSIE

349

compute Wo efficiency for each sample by using the following equation:

% efficiency =

cpm x 100 dpm

expose each sample to an external standard, '33Ba, and obtain Compton spectrum. Transform spectrum to obtain the end point of this spectrum, and then determine tSIE value plot Wo efficiency of each standard vs tSIE value use fixed point least square quadratic curve fit (FPLSQ) to connect points and obtain quench curve for radionuclide of interest

Now that this quench curve has been obtained, how can it be used to obtain the DPM in an unknown sample? The first step in determining the DPM in the

unknown sample, containing the same radionuclide as the nuclide used to prepare the standard curve, is to count the unknown sample under the same conditions used to create the standard curve. Then the CPM in the sample is determined The sample is then exposed to the external standard, and the tSIE (quench indicating parameter) is determined. From the standard curve and the tSIE, the counting efficiency for the sample is extrapolated. The DPM for the unknown is then calculated from the equation:

DPM-

-

cpm 'o efficiency

In order to obtain accurate DPM values, it is necessary to have a quench indicating parameter (tSIE value) which has the following characteristics: high dynamic range reproducibly and accuracy at normal and high quench conditions independence of sample volume independence of "wall effect" independence of vial type independence of vial size independence of cocktail density

The external standard used to determine the tSIE must have the following characteristics: I. low radiation hazard not subject to stringent radiation safety regulation the Em low enough so as not to require excessive shielding which could increase background maximum energy of Compton spectrum close to energy of 3H and '4C, most frequently used radionuclides in dual label counting energy of external standard source sufficiently low to reduce or eliminate spectral distortions caused by gamma ray interaction with material in counting chamber environment

timely determination such that counting time of sample is not greatly increased in determining DPM

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

350

This chapter will evaluate each characterstic and present the data as to the accuracy of the DPM determination. This can be illustrated best by defining a term called Wo recovery. The Wo recovery of an experiment is the DPM determined by the liquid scintillation analyzer compared to the actual DPM in the sample times lOOWo.

DPM counter x 100 DPM actual For all characteristics of the tSLE, the o recovery is calculated and plotted. The first characteristic of the tSIE is its volume independence. This can be illustrated by preparing a series of samples: standard 20 mL vial 19, 15, 12, 8 mL 7 mL vial-6, 4, 3, 2, 1, 0.5 mL microfuge tube-0.4, 0.2, 0.1, 0.05, 0.025 mL miniature

The samples were counted and a tSIE determined for each sample with a subsequent DPM printed out. The o recovery was then determined from the DPM of the counter, and the DPM actually present in the sample. A graph of the Wo recovery vs sample volume for 3H is shown in Figure 6. For the 15 sample volumes, a straight line was obtained near 100 recovery. The exact statistical numbers of 3H were a mean of 98. 19, a standard deviation of (SD) 1.58, and a coefficient of variation of (%CV) 1.61. For '4C, the numbers 120

110-

M100 90800 MICROFUGE TUBES

60-

SMALL VIALS

50-

® STANDARD VIALS

40-

p3020 -

100 .01

.10

1.00

VOLUME (mL) Figure 6.

3H DPM % recovery as a function of sample volume.

10.00

100.00

APPLICATIONS OF QUENCH MONITORING USING tSIE

351

Table 1. Quantitation of 3H and '4C and Dual Label Samples at Various Volumes Using tSIE as a Quench Monitor tSIE Nuclide % Recovery Volume

400 L 200 ,L

100 L 50 tL

25 L 400 L 200 L

3H 3H 3H 3H 3H

14C

100 L

50 L 25 L

14C

400 L 200 L 100 L

3H/14C 3H/14C

50 1L

3H/14C 3H1'4C

25 L

98.9 98.5 97.6 96.8 96.4

663 617 555 528 453

100.6 101.5 100.0 98.3 99.6

698

93.3/99.6 95.1/99.2 95.3/104.5 96.9/104.5 99.2/98.5

677

630 600 538

494 629

610 543 475

were even closer to 100 with a mean of 99.5lWo, a SD of 0.917 and WoCV of 0.922. This data clearly indicates that the DPM values obtained using the tSIE is volume independent. Because of the increasing use of microvolume counting, a special series of samples were prepared with volumes of 400, 200, 100, 50, and 25 L containing 3H, '4C, or dual label 3HP4C. These volumes were chosen because they are commonly used in the microvolume counting proce-

dure. The dual label samples were incorporated to determine if at low volumes, the DPM of a dual label sample could be determined accurately. Both recoveries (Table 1) show excellent recoveries. single label 3H and '4C DPM recoveries are 97.6% (3H) and 100 ('4C) with a small SD of 1.1 to The 1.2% for these isotopes. For the 3H/'4C dual label samples, the r/o recoveries

are 96.56% for 3H, and 101.2% for '4C. The standard deviation for 3H was 1.70, and for '4C it was 2.94. The data in Table 1 clearly indicate that the tSIE is decreasing as the sample volume decreases. If the tSIE and r/o efficiency decreases, then the CPM/r/o efficiency increases, and the corrected DPM is determined accurately. As mentioned in the requirements for a DPM (quenching method), the counting time should not be substantially increased. Despite these low sample volumes, the time required to calculate the tSIE (external standard quench indicating parameter) is short (0.5 to 1.0 minutes). The optimal geometry of the external source positioned directly below the sample and the ease of measurement of the Compton spectrum from the 'tm3Ba gamma source requires only a short time period for external source quench correction. The sample throughput, therefore, is not substantially affected by counting samples containing small volumes. These results are illustrated in Table 2. As can be clearly seen, the tSIE is very reproducible over the large quench range, with the efficiency for 3H decreasing from 58.47 to 3.32%. The data shows that the Wo recovery is very close to 100, with a maximum of %CV of 0.640 for the sample, with a tSIE = 83.59. The tSIE value is also very repro-

352

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 2. Reproducibility of % Efficiency, % Recovery and tSIE for 3H Quenched Samples, Counting Each Sample Ten Times Sample 1

2 3 4 5 6

7 8

% EFF

% Rec.

% CV

tSlE

% CV/tSIE

58.47 55.78 44.49 24.77 18.07

99.94 99.77 100.23 99.82 99.42 99.6 99.29 99.00

0.362 0.200 0.224 0.304 0.303 0.318 0.640

815.2 725.9 472.2

0.412 0.334 0.240 0.246 0.179 0.352 0.480 0.230

10.81

6.15 3.32

0.381

231.1 177.1 119.8

83.59 58.75

ducible even for the most heavily quenched sample of the sample set tSIE = 58.75 (%CV = 0.230). The maximum %CV for any tSIE is 0.480. Now that the tSIE has been shown to be reproducible at samples containing normal quenching values, what about samples that are severely quenched? Does the tSIE value become less reproducible as the quench becomes more severe? In order to assess this, a special set of ten samples was prepared. The tSIE values varied from 39 to 10, and the % efficiency varied from 1.388 to 0.011 Wo. Each of these samples was counted 10 times with the statistical data presented in Table 3. The samples (3H-labeled) are severely quenched with efficiency decreasing to

0.011%. Even at this level, the o efficiency has a %CV of 3.31 which is extremely good for these heavily quenched samples. The tSIE values decrease to 9.956 with a %CV of 0.296. The table also shows the three tSIE values; the tSIE was identical each of the 10 times that the tSIE was determined. This data clearly indicates that the tSIE method of determining quench is very reproducible and has a large dynamic range. These experiments were also conducted with '4C at

similar quench level. Almost identical results for tSIE reproducibility were obtained. A graphic summary of the type of quench curve which could be obtained with these 18 standard quenched samples is shown in Figure 7. The major portion of the quench curve (% efficiency vs tSIE) is shown in the graph. The heavily quenched sample region of the curve has been exploded in order to assess the nature of these curves at the extremely heavily quenched levels. Table 3. Reproducibility of % Efficiency, % Recovery and tSlE For Extended 3H Quench Samples, Counting Each Sample Ten Times Sample % EFF % CV/EFF tSIE % CV/tSIE 1

2

3 4 5 6

7 8

9 10

1.388 0.617 0.311

0.166 0.096 0.060 0.038 0.023 0.016 0.011

0.922 0.868 0.476 1.072 2.876 1.587 3.990 2.410 2.740 3.310

39.00 26.99 21.08 18.15 15.41

14.88 12.70 12.30 10.20 9.956

0.209 0.324 0.300 0.290 0.205 0.283 0.000 0.000 0.000 0.296

APPLICATIONS OF QUENCH MONITORING USING tSIE

353

60

5550-

45-

403530252015-

10-

50

0

100

200

400

300

500

600

700

800

900

800

90

tSIE

100 PLOT 90 80

70. 60

"C - exrENDEO OOENCH SEE EFFICIENCY vs. ISIS

20

50.

PLOT

IS

40 30

10

20

'U

0 12245170 9

10

10

% EFFICIENCY I

100

I

200

I

300

400

500

600

700

tSIE Figure 7.

Quench curve for 3H tSIE = 815-9.956 (top). Quench curve for 14C for tSIE = 8159.956.

354

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

The dynamic range for the tSIE method of determining efficiency can be defined by the following equation: Dynamic Range*

Wo eff. of unquenched sample = Wo eff. of most highly quenched

Using this definition, the dynamic ranges are as follows:

63.57 (tSIE = 1000)

0.011 (tSIE = 9.956) -96.54 (tSIE = 1000)

- 85.43 1.13 (tSIE = 9.956) -

These results clearly indicate that the tSIE method of determining quenching in a sample have a very broad dynamic range with very reproducible values even at a heavily quenched sample. The third characteristic of the tSIE is its ability to eliminate the "wall effect" problem, which may be present in samples prepared in liquid scintillation vials. The common organic solvent present in liquid scintillation solutions (toluene, xylene, pseudocumene) can readily penetrate the walls of plastic liquid scintillation vials. Two problems can be observed when using plastic liquid scintillation vials. First, the solvent can permeate the plastic wall of the scintillation vial. This

gives in inside of the liquid scintillation counter a strong organic odor. The penetration of the solvent into and through the scintillation vials results in actual

swelling of the vials; this makes it impossible to store and recount samples prepared in plastic vials. Second, not only the solvent but the scintillator can penetrate into the plastic wall. This can result in the plastic vial wall acting as a plastic scintillator. This "plastic scintillator wall" has a lower efficiency than the

sample in solution. Thus, if the plastic vial with plastic scintillator wall is exposed to the external gamma source, extra low energy photons will be emitted which could affect the low energy external standard spectrum. This could thus affect the external standard quench indicating parameter and result in incorrect

DPM values. The wall effect on the Wo recovery of a sample can easily be assessed by counting a sample initially at 12 hour periods. If the wall effect is affecting the DPM, then the o recovery would change as a function of time. This experiment was conducted for both the plastic standard (15 mL) and miniature vials (5 mL) over an 84 hour period.

The results in Table 4 clearly show that the °1 recovery of neither the standard vial nor the miniature vials change as a function of time. The oCV

for this entire time period is 0.71 and 0.36 for the 5 and 15 mL samples respectively. This indicates that the wall effect commonly seen in some liquid scintillation counters has been completely eliminated by the Packard quench indicating parameter (tSIE). The fourth characteristic of tSIE is its ability to give accurate DPM values *% recovery of all samples must be 100 ± 5o.

355

APPLICATIONS OF QUENCH MONITORING USING tSIE

Table 4. Effect of Possible "Wall Effect" on DPM Recovery in Plastic Vials Using 3H-Alanine in Instagel

Time (hr)

Volume (mL)

% DPM recovery

o 0 12 12

5 15 5 15 5 15 5 15 5 15

100.2 100.2

5

100.1

15

100.5 98.8 100.0

24 24

48 48 60 60 72 72

5

84 84

15

101.1 100.9 100.3 100.1 100.4

99.9 100.2 99.9

for different scintillation vial types (glass or plastic), different sample sizes (standard, miniature, and microfuge tubes), different scintillation solution types, and different quenching agents types (chemical and color). Let us divide these into two sections. The first will address the vial size and the vial type question. The second will cover the different types of scintillation solution, as well as various types of quenching agents. The results for different vial size and vial type is shown in Table 5. The results in the table indicate the % recovery using glass or plastic, and any vial size is 100% for all of the samples assayed. This shows that the DPM using tSIE is independent of the type or size of vial used to hold the sample. The second aspect to be investigated was the effect of various scintillation solutions and various quenching agents (color or chemical). The results are shown in Table 6. The first four samples indicate the result of the various quenching agents (color and chemical) present in the sample. The results in Table 6 indicate that the three chemical quenching agents (CHC13 and Cd3, and CH3NO2) do not affect the Vo recovery of the 3H-toluene in the sample. The addition of a red colored organic material, eosin, does not affect the % recovery with a 100.2% value obtained. This is further shown for various quench levels for 3H and '4C in Figure 8. The next three samples show that the sample density in g/cc does not Table 5. Assessment of DPM/% RecoveryVarious Vial Sizes and Types Using 3H-Toluene

Sample 1

2 3 4 5 6

7

Vial Type

% Recovery

20 mL glass 7 mL glass 6 x 50 mm Sealed 20 mL plastic 7 mL plastic 1.5 mL microfuge tube 0.5 mL microfuge tube

99.9 99.6 98.9 100.2 100.2 99.8 98.9

356

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 6. Effect of Type of Quenching Agent on DPM Recovery for3H Treatment

x 1023/mL Elect Density

Density (glcc)

Eosin Red (10 4.) Toluene + PPO Xylene + PPO Pseudocume/PPO

% Recovery

-

-

CHCI3 (10 L) CCI4 (10 L) CH3NO3 (10 L)

99.2 98.5 99.6 100.2 99.9 100.2 100.8

-

0.863 0.877 0.860

2.82 2.89 2.84

affect the recovery of the sample. In addition, the cocktail electron density of the samples was calculated. The cocktail electron density is the sum of the weight

of each component, times the electrons per molecule, times 6.0238 x 10 molecules/mo!. The first units for the electron density is electrons/mL. The effect of this electron density on the tSIE/DPM and 'o recovery was evaluated and found not to affect the recovery for the samples tested. Now, the tSIE method of evaluating sample quenching has been shown to be independent of most of the factors which affect DPM determination in simple single radiolabeled analysis. What about its accuracy in counting the more complicated dual label and the most complicated triple label samples? First, let us evaluate the method used for quantitating dual labeled samples. The problem is how to obtain the true DPM for each of the two single labels in a dual label sample. This problem is shown in Figure 9 for dual labeled 3H/14C samples. In order to obtain accurate DPM values for both radionuclides, the following steps should be taken: 120 110

icC 90

70

100

200

300

4CC

500

tsl Figure 8.

3H and 14C recovery for colored samples at various tSIEs.

357

APPLICATIONS OF QUENCH MONITORING USING tSIE

COMPOSITE SPECTRUM

dN

dE

Region

A I*e9io

Figure 9.

KeV

B

Dual label 3H/14C region settings for dual label counting region A = 0-12 key, region B = 12-156 key.

Prepare a set of H samples, each containing a constant number of DPMs, with zero to a maximum amount of quenching agent. A commercial sealed set can also be obtained for each radionuclide. Prepare a set of '4C samples similar to the 3H samples with a constant DPM in each and different quench levels. Next, count each sample in the 3H quench set to obtain DPM. m equation is for For each sample, calculate the 'o efficiency p both regions A and B.

Expose each sample to the built-in external gamma source, '33Ba, with the

tSIE being determined for each sample. This tSIE is independent of the radionuclide present or the region settings. The tSIE is an indicator of the quench level in the sample.

Obtain two plots of obtained(Figure 10) of % efficiency, 3H in region A and 3H in region B, as a function of tSIE. From the plots, the Vo efficiency in both the A and B regions can be obtained.

Assay the '4C standard in the same manner as the 3H standards, with two additional plots being obtained (Figure 11). The plots are of Wo efficiency of '4C in region A and '4C in region B. From these plots, the Wo efficiencycA and efficiencycB can be obtained using fixed point, at least square quadratic curve fit.

358

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS V. EFFICIENCY

0.00

00

20.0

40.0

1.0

2.0

60.0

80.0

100.0

220.00

440.00 QUENCH

660.00

000.00

1100.0

X EFFICIENCY

0. 00

0.0

30

4.0

2. C

220.00

440.00 CUENCI-I

(0.00

eBO. 00

1 100. 0

Figure 10.

Dual label 3H/14C quench curves for 3H using ISlE for region A (top). Dual label quench curve for 3H using tSIE for region B.

APPLICATIONS OF QUENCH MONITORING USING tSIE

359

EFFICLE.iCy

0.0

0.

3.0

10.0

15.0

20.0

220.00

440.00 QUENCH

860.00

090.00

1100.0 V. EFFICIENCY

0.00

00

20. 0

40.0

60 0

00.0

220. 00

440.00 QUENCH

660.00

090. 00

1100.0

Figure 11.

Dual label 3H/ 4C quench curves for 14C using tSlE for region A (top). Dual label quench curve for 14C using tSIE for region B.

360

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

8. Now the system is ready to determine DPM values for unknown samples. First the tSIE value is determined, and by using this value on the four plots above, four 0/o efficiencies can be obtained (% efficiencyHA, Wo efficiencyHB,

°lo efficiencyÃ, and /o efficiency). The sample is then counted in each of the two regions with the CPMA and CPMB obtained. Then the following equation can be used: CPMA = CPMB =

Eff. x DPM) + Eff.AB

x DPMH) +

Eff.CA x DPMC)

Eff.CB x DPM)

By rearranging and substituting into these equations, the DPMH and DPMC can be obtained.

DPM - (CPMH x

Eff.HA x

DPM=

(CPM x Eff.HA x

Eff.cD) - (CPMC x Eff.CB) - (CPMC x

Eff.CA) Eff.CA)

Eff.HA) - (CPMH x EffHB X Eff.CB) -

Eff.HB) Eff.CA)

As can be seen from the equations and plots, two of the most important efficiency and carbon in region A. The factors are the efficiencyCA (crossover of '4C into 3H) increases drastically when the sample becomes more quenched. This could cause considerable problems when a large ratio of '4C to 3H exists in the sample. Therefore, a special feature termed AEC (Automatic

Efficiency Control) was implemented in the system to help overcome this problem. This feature automatically tracks the theoretical end point of both the 3H and 'C spectrum in the sample. By doing this, the efficiency of '4C in the A region is kept relatively constant, as shown in Figure 12. This enables

accurate DPM determinations for 3H/14C dual labeled samples at various quench levels and at various radionuclide ratios. This technique can be used for any dual labeled sample separated by at least 175 keV of energy. In order to assess the performance of the tSIE as a quench indicating parameter, a DPM for dual labeled samples and a series of samples were prepared. These samples contained 3H/'4C ratios of 1:1, 1:5, 1:10, 1:20, and 1:50. Each of these samples were prepared at various quenched levels. The DPM for each of the dual labeled samples was determined using the previously described procedure with AEC activated (Figure 13). The results, shown in Table 7, indicate that for the 14 samples evaluated, the 3H recovery was 100.31 and the recoveries are extremely stable, even for samples at These '4C was 100.85 high '4C/3H ratios of 20: 1 and 50:1. It is also clear from the table that a quench recovery of both range from very small to moderate to severe, that the radionuclides is close to 100. The results are also very reproducible with a WoCV of 0.887 for 3H, and 0.886 for 'AC.

In order to evaluate the DPM determination for a dual labeled sample of 3H/4C samples of a 1:20 ratio were prepared and quantitated. A plot of

APPLICATIONS OF QUENCH MONITORING USING tSIE

361

% EFFICIENCY 100

(EHB)

"C IN "C CHANNEL

(ELAJ

H IN 3H CHANNEL

40

(EHA.)

20

14

IN H CHANNEL

H IN "C CHANNEL(ELB) QIP 400

800

600

1000

tSIE

Figure 12.

Dual label with AEC 3H and 14C in regions A and B.

recovery of both 3H and '4C is shown in Figure 10. The quenched levels range from a tSIE of 800 to 83.9% to 62.1% for '4C. The results in the graph indicate a stable 0/o recovery of 3H and '4C, even at a high ratio of 20:1. Now that the dual label samples have been shown to use tSLE to determine accurate DPM values, what about the ultimate analysis of a triple label sample containing 3H/14C/32P? This triple label sample analysis is very similar to the dual labeled analysis except that instead of 'Io efficiency vs tSLE curves, 9 are required. Three for each radionuclide in each of the three regions of interest. From these 9 Wo efficiencies, the three CPM values for regions A, B, and C,

and the DPM of each of the three radionuclides can be quantitated by an inverse matrix analysis procedure. In order to completely evaluate the triple label DPM procedure, four different ratios of 3H/'4C/32P were used at various quenched levels. The DPM was determined for each radionuclide. The statis-

tics, mean, SD and CV were also calculated. The results clearly indicate that the ¼ recoveries are very close to 100¼ (99,65, 101.46, and 100.15), and all samples have a WoCV of less than 1.36 over

the entire quenched range. This would be appropriate if all experiments were

done at a ratio of 1:1:1, but most scientists use various ratios of the three radionuclides present. How do the tSIE perform under conditions when one of

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

362

11n.

1O.

t 11114C 1:2131

A 311/14C 1:21 14C IECOI3Y

COVUY

10I. 1W.

11-

lot

'I w

8 w

HXJ I

I

411

I

II

S.I

II

1

Il

t

II

f

I0

1.1

tUE Figure 13. % recovery of 3H and 14C for dual label with AEC 3H1'4C ratio 1:20.

Table 7. % Recovery of Dual Label 3H1'4C at Various Quench Levels and Isotope Ratios Ratio 3H114C 1:1 1:1 1:1

1:5 1:5 1:5 1:10 1:10 1:10 1:20 1:20 1:20 1:50 1:50

3H % REC

14C ¼ REC

tSIE

100.6 100.6 100.2 99.9 102.3 100.5

101.0 101.3 98.3 101.8 100.5 102.0 101.3

808

100.1

99.7 100.8 99.7 99.8 99.5 100.9 101.5

101.1

101.5 100.9 100.5 100.9 100.6 100.2

473 161

828 463 174

786 464 173

780 448 166 786 681

APPLICATIONS OF QUENCH MONITORING USING tSIE

363

Table 8. 0,4 Recovery of 3H/'4C/32P at Ratios of 1:1:1 at Various Quench Levels tSIE

3H

98.90 98.12 98.79 98.98 99.23 99.80 101.9 101.50

101.89 100.78 100.57 101.78 101.14 101.04 101.94 101.78

100.04 99.16

Mean

99.65

101.15

SD % CV

1.35 1.36

101.46 0.56 0.55

811

724 622 499 385

100.01 100.51 100.35 101.13 100.10

329 185

99.87

117

0.62 0.62

Table 9. % Recovery of 3H114C/32P at Ratios of 10:1:1 at Various Quench Levels

Mean SD

% CV

3H

"C

100.5

99.05

97.50

817

100.1

94.71

100.2

95.54

98.90 100.9

753 639

100.1 101.1

98.96

100.5

98.23 98.54 97.80

100.62 99.2

572 498 420

102.3

96.50

99.33

262

102.5

97.40

102.50

178

100.91

97.24

99.68

0.97 0.96

1.52 1.57

1.41 1.41

tSIE

the radioisotopes is in a tenfold excess over the other two radioisotopes present in the sample. This can be demonstrated by analyzing the data in Tables 9 to 11. Table 9 shows the results with an excess (10 x) of the lower energy 3H in the sample.

The results show that the recoveries are very close to 100. The mean recoveries are 100.91 (3H), 97.24 ('4C), and 99.68 (32C). All show a standard deviation of less than 1.52. Table 10 displays the results for a sample containTable 10. % Recovery of 3H/'4C132P at Ratios of 1:10:1 at Various Quench Levels 3H

"C

100.24 103.70 98.71 105.17 99.81

99.88 99.54 101.56 100.14 100.28

101.43

103.40 104.03 103.21

100.48 100.45 100.78

100.73 99.98 104.67

97.89

100.75

99.53

Mean SD

102.44 2.56

% CV

2.51

100.42 0.59 0.59

100.81 1.58 1.57

tSIE 99.70 100.02

813

101.11 100.18

646

738 572 496 435 361

265

179

364

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 11. % Recovery of 3H114C132P at Ratios of 1:1:10 at Various Quench Levels tSIE

3H

96.72

103.73 103.55

99.89 108.28 95.53 99.99 106.95 102.09 91.66 107.47

100.49 99.67 101.38 102.04 99.97 101.36 101.24 100.29

99.45 3.20 3.22

101.45 5.90 5.82

100.83 0.80 0.80

98.51 102.19

98.40 95.92 96.66

Mean SD

% CV

817 751

655 605 504 443 269 183

ing a 3H/14C/32P ratio of 1:10:1. The results show a mean of 102.44 (3H), 100.42 (14C), and 100.81 (32P). These results are over the entire quench range analyzed by the liquid scintillation analyzer. The WoCV are all less than 2.51. Finally, the most difficult triple label sample ratio to analyze is at a 3H/'4C/ 32p ratio of 1:1:10. The reason that this is the most difficult is that both '4C and

large amounts of 32p must be compensated for within the 3H region. The results for this ratio are in Table 11. The statistical analysis of all of the recoveries (99.45 [3H], 101.45 quenched samples indicates extremely good [14C], and 100.83 [32Pj. Because of the high ratio of 32P in the sample, and consequently the high crossover of 32P into both 3H and '4C, the WoCV for these isotopes increased to 3.22 for 3H, and 5.82 for 14C. These numbers are extremely accurate considering that Wo efficiencies and three CPM values are used to calculate the DPM of each of the three radionuclides. Thus, the tSIE is an accurate method of determining the DPM in a triple label sample at various quench levels and with various isotope ratios.

SUMMARY

This paper definitely demonstrates that the use of the quenched indicating

parameter, tSIE, is a rapid and accurate method of determining DPM of single, dual, and triple labeled samples. The DPM obtained with this technique is independent of sample volume, extent of quenching, wall effect, vial size, vial type, cocktail density, radionuclides number being analyzed (maximum = 3), and quenching agent type. Accurate DPM can be obtained for dual label samples at 100:1 to 1:50 ratios

and at various quenched levels. The DPM for triple label samples can be determined accurately, and with a small coefficient of variation for various ratios of radionuclides. In conclusion, the use of tSIE as a quench indicating parameter results in accurate DPM values independent of most interferences found in liquid scintillation counting.

CHAPTER 31

Scintillation Proximity Assay: Instrumentation Requirements and Performance Kenneth E. Neumann and Stat van Cauter

ABSTRACT Recently, a new immunoassay and receptor binding analysis technology has been introduced. This methodology, scintillation proximity assay (SPA), belongs to a family of ligand binding techniques known as sandwich assays" SPA technology employs a scintillation microsphere as the solid support, but requires no separation step. In addition, no conventional liquid cocktail is needed, minimizing safety hazards and disposal costs. The unusual demands this technology places on currently available liquid scintillation instrumentation will be discussed. We will concentrate on the ability to accurately quantitate ow levels of a low energy isotope, such as tritium, with adequate sensitivity. Since the recommended total sample volume is no more than 0.4 mL, acceptable sensitivity requires excellent light collection efficiency. Additionally, this technology is intended for high volume applications. Throughput, to a significant degree, is influenced by counting efficiency, and hence, instrument performance. Finally, the unique scintillation properties of the fluor material may affect instrument counting efficiencies, leading to poor quality data. Because assay sensitivity may be affected by the instrument employed, it is important that the technology is evaluated on several types of liquid scintillation counters. The results of such a study are presented. Alternative instrumental methods for counting SPA samples, providing increased sample throughput, are discussed.

INTRODUCTION

Classical immunodiagnostic and ligand-binding methodology using betaand/or gamma-emitting radionuclides, is widely recognized in the industry for its sensitivity and specificity for the analyte of interest."2 Several general techniques, such as direct and competitive assays, exist for a variety of applications. A third category is typically referred to as a sandwich assay.3 Here, the analyte (usually an antigen) is bound to two different antibodies. One antibody is labeled with a radioactive tracer (typically 3H or 1251), while the other is

permanently bound to a solid support structure such as a polymeric microsphere. When incubated together, these three elements form an Ab-Ag-Ab complex bound to the microsphere. The fraction of radiotracer bound to the 365

366

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

ligand can be separated from the tracer in free solution via a number of techniques including centrifugation, filtration, charcoal adsorption, or magnetic separation. One significant disadvantage in these RIA methods is the time and effort required for this separation step.4 A key challenge over the past 20 years has been to simplify the sample preparation steps, especially separation, required prior to assaying the tubes or vials. A new technique, scintillation proximity assay (SPA), has recently been introduced for immunodiagnostic and ligand-binding assays. This method exploits the very low radioactive energy of the common beta emitting radionuclide tritium (3H) or the Auger electrons emitted by 125j Because tritium decays energy on the average of 6 keV, beta electrons formed during the decay process

travel only a short distance (4 Jzm in water). Therefore, in order to create a scintillation event, a tritium label must be in intimate contact with the scintillating medium.5

SPA technology employs a scintillating microsphere as the solid support structure. Formation of the Ab-Ag-Ab complex, which requires the presence of Ag in free solution, brings the radiolabeled antibody close to the scintillating particle, excites it, and causes photon emission. If there is little or no Ag in

free solution, the complex will not be formed to any large degree. In this situation, radiolabeled Ab will not be bound close to the scintillating microsphere, and therefore, will not cause appreciable excitation of the fluor,6'7 thus, the number of detected scintillation events is directly related to the amount of Ag in the sample. This effectively separates the bound fraction from the free fraction, without any need for manual separation steps.8 Samples can be assayed directly in a conventional liquid scintillation counting system, following an appropriate incubation period. As a result, assay precision is ultimately improved due to fewer sample processing steps. While SPA technology has obvious advantages in the areas of sample prepa-

ration, incubation, and separation, it also places unusual demands on commercially available liquid scintillation counting instrumentation. Most important is accurate quantitation of low levels of tritium with adequate sensitivity for routine immunoassay and receptor binding applications. Acceptable sensitivity requires both excellent tritium counting efficiency and low background count rate. A primary factor influencing these requirements is the fact that typical SPA samples have a total volume of 0.4 mL and are prepared in 7 mL LS vials. A compounding factor is that the bound (scintillating) fraction consists of only a few milligrams of material which rapidly fall to the bottom of the vial. Sample

geometry, as presented to the liquid scintillation counter, is therefore quite poor. This will tend to limit the photon collection efficiency, and thus the effective radionuclide counting efficiency. A low overall counting efficiency reduces the statistical accuracy with which one can measure a sample at a given count time. In order to compensate, the investigator or clinician must increase the sample count time or the sample volume to achieve better statistics. This either decreases sample throughput or increases assay cost. Furthermore, over-

INSTRUMENTATION REQUIREMENTS AND PERFORMANCE

367

all counting efficiency is critical in the determination of the signal to noise (SI N) ratio for the instrument employed in the assay. Poor efficiencies will limit the S/N ratio. As a result, assay sensitivity might be adversely affected.

A third factor influencing the net performance of this technology is the ability of a liquid scintillation counter to directly assay samples contained in 7 mL vials. Most currently available instrumentation is capable of loading and counting these vials directly. However, older counters may only be capable of counting samples in 20 mL vials. This necessitates placing the small SPA vial within an adaptor and assaying it as a large vial. This significantly reduces sample throughput. Another factor influencing SPA performance is the fluor employed as the solid support medium. Experimental evidence indicates that the fluorescent emission of the scintillating microsphere occurs over a relatively long period of time compared to the emission from conventional LS samples. In addition, pulse height distributions obtained from typical SPA samples show that the fluorescent emissions average a slightly higher energy level than those usually encountered from conventional scintillation fluors. Currently available liquid scintillation systems incorporate various pulse discrimination schemes, based

on pulse height and width, in order to optimize instrument performance. These systems have been optimized for use with existing scintillation chemicals. The novel pulse decay characteristics of the scintillator used in SPA technology has broad implications for the signal processing techniques that discrimi-

nate true beta decay events from PMT thermal noise. Conventional liquid scintillation counters accomplish this via coincidence circuitry in conjunction with narrow pulse width scintillators. The design of conventional LSCs has been based on, and limited by, these requirements. By using a scintillator with a wider decay pulse, as shown below, patented time-resolved pulse discrimination techniques, using only a single PMT per detector, can discriminate against noise events. Radiochromatography counting systems based on this technology are well known. The experiments detailed in the following section describe recent work done to better characterize SPA technology performance on existing liquid scintillation counting equipment. In addition, the evaluation has been carried out on experimental time-resolved counting equipment, using single PMT detectors. The results of these experiments are also presented.

EXPERIMENTAL

All experiments described herein were performed using either the Thromboxane B2 (code II TRK.951) or 6-Keto-prostaglandin Fla (Code # TRK.952) SPA kits available from Amersham Corporation. These kits contain all of the

reagents necessary to prepare SPA samples. The components are listed in Table 1.

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINT$LLATORS

368

Table 1. Contents of SPA Reagent Kit Item

Component Assay buffer (PBS + gelatin +

1

Volume (DL) 100

thimerosal) 2

Tracer (H-3 labeled antibody)

100

3

Antiserum

100

4

SPA protein a reagent (coupled to scintillating microspheres)

100

5

Standards (a through E, with varying amounts of antibody)

100

each

The instructions included with the reagents describe two sample preparation protocols - one day and one overnight. A preliminary study was done in which these two protocols were directly compared, to find the optimum preparation protocol. Triplicate sets of standard, NSB, and Bo tubes were prepared per the instructions provided for each protocol. All samples were prepared in 7 mL polyethylene LS vials (Packard, #6000192), and incubated with mixing for the

appropriate period using a commercially available multi tube vortexer. All tubes were then assayed in a Packard Tri-Carb 2500TR liquid scintillation analyzer using a region setting of 0 to 200 keV. The results of this experiment,

illustrated in Figure 1, indicated that few significant differences in sample count rate exist between the two protocols. For this reason, and for reasons of convenience, the overnight procedure was chosen as the preparation protocol to be used in all further experiments. Because most current-generation liquid scintillation counting systems are

10 lLD

0

A

B

C

0

STPNR Figure 1.

Comparison of 6 hr and overnight SPA incubation protocols.

INSTRUMENTATION REQUIREMENTS AND PERFORMANCE

369

equipped with MCA technology, counting regions can be optimized for the particular radionuclide/fluor combination being assayed. SPA technology employs a unique combination of radiolabeled tracer and solid scintillator. This can be illustrated by comparing an oscilloscope trace for a typical SPA sample to a conventional LS sample, as illustrated in Figure 2. Therefore, an initial experiment was performed to evaluate the spectral characteristics of a typical SPA sample. A Bo tube was prepared, incubated, and assayed in the Tri-Carb 2500TR LSA. A representative sample spectrum was collected and plotted (Figure 3). This plot indicates that preset 3H counting regions will not capture the entire SPA sample spectrum. It is necessary to manually set the counting window to a region encompassing 0 to 30 keV.

a) Conventional LS Cocktail 1.4 mV/div 10 ns/div

:.:b) 'F :.

Figure 2.

SP Scintillator ..14...mV/.d4.v. 1Q ns/div

Pulse shape characteristics: a) conventional LS cocktail, b) typical SPA sample.

370

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS 2.4 2.2 2 1_I 1.6

tao

00

1.4 1.2

0.I 0.6

0.4 0.2 0

0

1*

12

0

20

24

28

ENERGY (kaY) SPA

+

UNOSTD

Figure 3. Normalized 3H spectra.

The first set of experiments was designed to determine the long-term stability of typical SPA samples after preparation. Triplicate sets of standard, NSB, and blank tubes were prepared per the previously qualified overnight protocol. Following incubation, all tubes were repeatedly assayed in the aforementioned liquid scintillation analyzer for a period of 60 hr. The data obtained from these assays indicated that the count rates of the samples, as a group, decreased by a significant amount over the 60 hr period. Figure 4 illustrates this for standards

A.E. Furthermore, it was noted that the rate and function of this decrease varied with the initial count rate of the sample. Because of the time dependency of the sample count rates, it becomes necessary to normalize the results of any one assay to an appropriate point in time relative to the incubation period. Therefore, a mathematical function was developed for each sample type (Stds A-O, and NSB) that relates the sample CPM to the time after incubation, when the sample was assayed. These functions were found to be fourth-order polynomials. These functions were used to calculate normalization factors based on sample type and time of assay. A counting system using the aforementioned single-PMT discrimination scheme has been developed and constructed.9 Experiments performed during development have demonstrated that efficiencies and backgrounds approaching those of conventional LSC can be achieved using a fluor with a relatively long decay constant. SPA technology employs such a scintillator. Therefore, a series of experiments was performed to evaluate SPA technology using this type of counter. Triplicate sets of Thromboxane B2 standards were prepared

INSTRUMENTATION REQUIREMENTS AND PERFORMANCE

371

ci SD E S1D 0

SID C

SID B S1D A

ELPPD 11E (rs) Figure 4.

SPA sample stability after overnight incubation.

per the overnight protocol and assayed in the Tri-Carb 2500TR LSA. All samples were then assayed in the experimental counting system. Raw count results were normalized to time zero using the relationships described above. The normalized results for each test instrument were then compared to the reference instrument (the Tri-Carb 2500TR) by dividing each result by the

count rate obtained in the reference system. The resulting CPM ratios are illustrated in Figure 5. The data obtained from the above trials were also used to calculate figures 1.2

1.C 2X1R

Lt2

0

A

B

C

0

E

SRE Figure 5.

Normalized CPM ratios for reference and experiemental counters.

372

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 2. Figures of Merit for SPA Counting Systems Parameter TRI-CARB 2500TR CPM, Bo tube CPM, blank tube FOM

Experimental System

1911.3

1160.0

20.1

3.1

94.1

373.2

BLK ) (Bo.BLK

of merit for each instrument. To evaluate the sensitivity of the counter, the following equation was employed: FOM

-

CPM(Bo) - CPM(blank) CPM(blank)

The value of this parameter is directly proportional to the dynamic range of the instrument for this application; that is, a maximum value is indicative of excellent performance. The values for each of the instruments employed in this study are displayed in Table 2. Finally, standard curves were plotted for each of the trials, according to the format recommended in the SPA instruction booklet (Figure 6). Note that while count rates for the experimental system were generally about 40% lower, the resulting standard curves are almost identical. The count rates obtained for each of the runs were then plotted against each other in order to directly compare instrument performance. Figure 7 illustrates these results. Here, we observe a high degree of correlation between the experimental system and the liquid scintillation counter. This suggests that a commercial system of the type described above would be ideally suited for SPA applications.

TC 2AXJ1R

I bg(og

1h,xntoxar

Figure 6. Thromboxane B2 standard curves.

B2)

INSTRUMENTATION REQUIREMENTS AND PERFORMANCE

2f+03

IE+03

cM 1}C Figure 7.

373

2UJIR

SPA Count rate correlation.

RESULTS AND CONCLUSION

SPA technology represents a significant development in the field of radiometric assay. Because of the lack of a separation step in the sample preparation, labor costs for laboratories running routine ligand binding assays can be markedly reduced. Although most current assay kits are based on gammaemitting radionuclides such as 1251, the use of a beta-emitter such as 3H is attractive due to its longer shelf life and lesser radiological hazard. Previous 3H assays required separation steps and conventional liquid scintillation cocktails. These too represent both a cost and a hazard to many laboratories. Scintillation proximity assay addresses these issues by eliminating both of them. In doing so, however, the technology forces a critical evaluation of the instrumentation required to perform these assays. The nature of the samples, and particularly the scintillator employed in them, is unique in comparison to classical LS samples and cocktails. While most commercially available liquid scintillation counting systems are capable of counting samples of this type, it is important to realize that SPA technology creates somewhat unusual counting requirements. The data presented above clearly indicate that a typical current-generation liquid scintillation counter can effectively analyze SPA samples with adequate efficiency and sensitivity. Commercially available immunoassay data reduction packages such as RiaSmart (Packard Instrument Company) can readily be joined to the LSC to provide an integrated immunoassay environment. One critical drawback, however, is the net sample throughput. Scintillation proximity assay is intended for high volume applications such as receptor binding and drug screening. In addition, the homogeneous nature of SPA technology theo-

374

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

retically allows the user to perform kinetic measurements rather than endpoint determinations, effectively reducing incubation times. Conventional LSCs, possessing only a single detector, limit the number of samples that can be

analyzed in a given period and make it impossible to perform kinetic measurements.

It has also been demonstrated that unique instrument technologies can count SPA samples. While classical coincidence counting, using two PMTs, is quite effective, new techniques, based on the unique scintillation properties of the fluor, are equally if not more effective for this type of sample. Single-PMT noise discrimination schemes can offer comparable counting efficiencies with

reasonable background levels. One such scheme, used in the above experiments, resulted in an approximate quadrupling of the figure of merit. This ultimately results in greater assay sensitivity. The effectiveness of time-resolved single PMT designs has major implications for the instrument technologies suited for scintillation proximity assay. Because the samples are of limited volume, large detectors are not required. In addition, the noise discrimination circuitry employed in the above experiments does not require massive lead shielding. These factors can lead to the development of extremely compact dedicated instrumentation. In addition, multiple detector instrumentation, which uses this technology, can address the problem

of sample throughput and enhance assay performance through kinetic measurements.

REFERENCES

Yallow, R.S. and S.A. Berson. .1. C/in. Invest. 39:1157 (1960). Roitt, I.M., J. Brostoff, and D.K. Male. Immunology, (New York: Gower Medical Publishing Company, 1989), pp. 25.5-25.6.

Odell, W.D. and P. Franchimont. Principles of Competitive Protein-Binding Assays, (New York: John Wiley & Sons, 1983), pp. 243-254. Chard, T. An Introduction to Radioimmunoassay and Related Techniques (1982). Kirk, G. and S.M. Gruner. IEEE Trans. NucI. Sci., NS-29, p. 769 (1982). Udenfriend, S., L. Gerber, and N. Nelson. "Scintillation Proximity Assay: A Sensi-

tive and Continuous Isotopic Method for Monitoring Ligand/Receptor and Antigen/Antibody Interactions," Anal. Biochem., 161:494-500 (1987).

Udenfriend, S., L.D. Gerber, L. Brink, and S. Spector. "Scintillation Proximity Radioimmunoassay Utilizing '251-Labeled Ligands," Proc. Nat!. Acad. Sci. USA, 82:8672-8676 (1985).

U. S. Patent Number 4,568,649, Immediate Ligand Detection Assay. U. S. Patent Number 4,528,450, Method and Apparatus for Measuring Radioactive Decay.

CHAPTER 32

The LSC Approach to Radon Counting in Air and Water

Charles J. Passo, Jr. and James M. Floeckher

INTRODUCTION

Various methods exist to monitor 222Rn in air. There are seven commonly used types of detection: alpha track, activated charcoal adsorption, continuous radon monitoring, continuous working level monitoring, grab radon sampling, grab working level sampling, and radon progeny integrated sampling. Of these methods, the passive diffusion activated charcoal canister technique has been the method of choice when screening for radon. The liquid scintillation method is an improvement of the activated charcoal canister technique. Specially designed air detectors containing a small amount of activated charcoal and desiccant serve as the collection and counting vial. Because the liquid scintillation method offers the advantages of higher sensitivity, higher throughput, and shorter exposure times, interest in the method of measuring radon in air has increased.

Radon in water has been measured by both the Lucas celP2 and liquid scintillation methods.3 The liquid scintillation technique combines the advantages of minimal sample preparation time (1 mm/sample), small sample sizes (10 mL), and high sensitivity (200 pCi/L). This chapter will describe a liquid scintillation method for neasuring radon

in both air and waterthe PICORADTI system from Packard Instruments. Data comparing air measurements made with typical charcoal canisters and the liquid scintillation detectors will be presented.

375

376

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

RADON IN AIR USING LIQUID SCINTILLATION

The Detector The PICORAD detectors are passive devices requiring no power. They are integrating detectors that determine the average radon concentration in the air they are placed. The detector consists of a porous canister welded securely near the top of a plastic liquid scintillation vial. The porous canister contains a bed of a controlled weight of charcoal (1.3 g) and silica gel desiccant (0.9 g). The vial has a removable cap and an 0-ring seal to prevent moisture or radon from entering the vial during storage and after exposure. Exposure Procedure

To initiate the exposure, the vial is uncapped to allow radon laden air to diffuse into the charcoal. The nominal measurement time is 48 hr, at which point the radon accumulation has reached 95% saturation. A testing time of 24 hr is recommended in the summer when moisture problems can be severe. A

2 day exposure at 100% humidity will result in a 5% weight gain and a subsequent 10% loss of maximum counts. The radon accumulation reaches 80% of its saturation value in 24 hr. The rate of radon accumulation in the PICORAD detectors, empirically determined over a span of 9 hr to 5 days at room temperature, is shown in Figure 1. At the end of the exposure time, the detector is capped and sent to the laboratory for analysis. Elution Procedure

In the laboratory, 10 cc of xylene based cocktail (InstaFlourTM Packard Instruments, Meriden, CT) is syringed or pipetted into the bottom of the detector, which is then recapped. With the cocktail below the charcoal canister, the desorption takes place through the vapor phase. Desorption is more than 80% complete after 3 hr (the time for full equilibrium of the decay product). The count rate reaches its maximum value in about 8 hr. The elution curve vs time, generated at room temperature, is shown in Figure 2. This curve is in good agreement with the elution curve determined by other investigators using a similar airborne radon detector of their own design.4

Analysis of Results

A calibration constant of 37 cpm (pCiL) has been empirically determined at room temperature. Counts per minute corrected for background, are multiplied by the calibration constant and by correction factors for the decay of radon, for adsorption time, and for elution time. The calculations are part of a computer program licensed through Niton Corporation in Bedford, Massa-

1

Figure 1. Adsorption of radon in air in PICO-RAD detectors.

10

20

30

40

Percent Adsorption 50

60

70

80

90

100

3

Time of Exposure in Days

2

Figure 2.

1

2

4 5

6

7

8

9

Time after start of elution in hours

3

Desorption of radon from PICO-RAD detectors using a xylene based cocktail.

10

20 -

30 -

40

Percent of 60 maximum cqunts per 50 minute

70-

80'

100 90 -

10

11

12

THE LSC APPROACH TO RADON COUNTING

379

chusetts. The general form of the equation programmed into the computer is as follows: I (pCi/i) = K (N1 - N0 (e(T4T2)o)' (1 - ep_(T2 - Ti)/r1Y' (1 - e_(T4 - T3/r211

where

I = radon concentration in pCi/i K = normalization constant that converts net cpm to pCi/I N1 = cpm obtained in the appropriate energy window of the LSC

N0 = mean value of the background counts obtained from many background samples r0 = mean life of 222Rn r1

= mean time for the adsorption of radon into the acti-

vated charcoal = mean time for desorption of the radon out of the activated charcoal and into the liquid scintillation eluant T1 = exposure start time T2 = exposure end time T3 = time the elutant cocktail was placed in the detector T4 = analysis time r2

Comparisons with the Charcoal Canister Technique

A side by side study was performed to investigate the correlation of results

obtained with the activated charcoal canister and the liquid scintillation method. The charcoal canisters were obtained from Canberra Industries, Meriden, CT. The basic design and analysis of the canisters is the same as described by

George.5 An individual container is 10.2 cm by 3.3 cm in dimension and contains 150-200 g of activated charcoal. The container has no diffusion barrier. The canisters were weighed before and after exposure to determine the moisture content and correct for humidity. The liquid scintillation detectors are the PICORAD detectors described ear-

lier. The study consisted of exposing both types of detectors at the same location in the basement of eight private homes. Results

Figure 3 is a bar graph illustrating the results of side by side exposures done in eight private homes. From the graph, it is apparent that the liquid scintillation results compare favorably to those obtained by the activated charcoal canister method over a range of radon levels. This relationship is shown in Figure 4. The data plotted are the mean values of 2 to 6 individual determinations for each test. The individual results for tests number 7 and 8 are pre-

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

3

5

4

2

6

8

# of ILTS Figure 3.

Method comparison of radon levels in eight private homes.

sented in Table 1. Six determinations with each method were made for these tests, because these were the homes found to have significant radon levels from an earlier screening. Student t tests were done to determine if the difference between the mean exposures for each method is significant. As reported in Table 1, the p values for test 7 and 8 indicate that in test 7, the mean values did not achieve a level of significance, while in test 8 they did. Even though the mean values for test 8 seem close, the standard errors around the means are

ion

+ LSC

8

REERB'ICE = 0.983

0

0

2

4

6

8

10

12

A C. CM'ISTERS pCi/L

Figure 4.

Relationship of radon measurements made with the activated charcoal and liquid scintillation methods.

THE LSC APPROACH TO RADON COUNTING

Table 1. Individual Data and Statistics for Tests Numbered 7 and 8a Test 7 Sample

A.C.

1

7.0 6.8 6.3

LSC 8.9 6.4 6.8

7.1

8.1

6.5 6.2

7.3 6.8

2 3 4 5 6 Mean S.D.

%CV

6.65 0.37 5.6

7.38 0.94 12.8

Student t

Test 8

A.C.

LSC

1.9

1.8 1.6 1.2 1.5 1.5 1.6

2.3 1.7 2.0 2.0 1.9

1.97

1.53

0.197

0.197

10.0

12.8

1.77

3.88

lOdf

lOdf

p>0.1

381

p<0.01

aRadon Levels pCi/L.

small, accounting for the level of significance achieved. Ideally, a t test run on a larger sample size would be more appropriate. RADON IN WATER

The Collection Technique

The method of collection is important to insure against loss of radon from the water sample. The water collection procedure used here is based on EPA documents and individual test results. The collection technique is as follows: If taking water from a faucet, remove the aerator and draw the water for five to ten minutes for the sample to be fresh. The water drawn should be a cold water sample since radon is less soluble in warm water. If taking water from a lake or stream, a direct sample can be taken. The recommended container is a standard EPA type water collection bottle with a volume of at least 20 mL. These vials have rubber-teflon plenums and prevent leakage of radon from the vial. Teste of radon leakage from water in such bottles indicate that the mean leakage time exceeds 3000 hr. The bottle should be filled such that a minimum amount of air is retained, preferably less than 1 cc, since radon prefers to be in air rather than water in the ratio of 4:1. As a result, a I cc bubble of air in a 20 mL bottle will

yield an apparent radon concentration in the water which is 9W0 lower than the true Concentration.

The Elution Procedure The elution procedure for radon in water begins with a 20 mL syringe filled with 10 mL of Opti-Flour® 0 (Packard Instruments, Meriden, CT.). The cap of the water collection bottle is removed, and 10 mL of water is drawn into the syringe from the bottom of the bottle. The 20 mL mixture is then transferred to a liquid scintillation vial for counting. The syringe should be rinsed with

"radon free" water between each test. The liquid scintillation vial is then shaken for approximately 5 sec (not critical) and set aside for equilibration of

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

382

the radon into the scintillation cocktail. A typical count rate vs time curve such as the one reported by Prichard3 is shown in Figure 5. As can be seen from the figure, a 3 hr equilibration time is appropriate. The elution will then be more

than 95o complete. Counting the Water Sample

Each decay of radon in the water will result in five detected counts: three alpha particles and two beta particles. The efficiency of counting the aiphas and betas is close to l00o, with the lower level discriminator on the liquid scintillation counter set to 25 keY and upper value set to 900 keV. This value has been determined using a Bureau of Standards source of radium in water. The number of pCi of radon in the water is calculated using the following formula: pCi/I = (N - N1) * 100 / (5 * 2.22 * 0.964) (D) where

N = cpm obtained from the liquid scintillation counter N1 = background cpm 100 = multiplication factor to give the concentration in pCi/i 5 = number of counts per decay of radon 2.22 = number of decays of radon per minute per pCi

0.964 = correction factor applied for the radon in air vs water in the collection bottle

D = radon decay correction factor

lOrnL water

3

4

5

liME (hrs.)

Figure 5.

Count rate vs time for a

10

mL water sample in 10 mL of Opti.Fluor 0.

THE LSC APPROACH TO RADON COUNTING

383

A background sample can be prepared by from water drawn from a municipal source and either aerated for a number of hours, heated to 80°C for several hours, or allowed to stand idle for several months.

Comparisons with the Lucas Cell Method A good correlation exists between the results obtained with the liquid scintillation and Lucas cell method. In early experiments, Prichard3 reported a 9.34 counts/mm/pCi calibration factor from a plot of the relationship of the liquid scintillation count rate vs the Lucas cell results.

The predicted value, based on the relative solubilities of radon and the volumes of water, scintillation fluid, and air in the collection bottle, was 9.75 counts/mm/pCi. This relationship is shown in Figure 6.

250 E

200

-

W

c_) -

slope 0.0934 cpm/p Ci// U)

0

500 222

FIgure 6.

1000

1500

2000 2500

Rn Content of Water (pCi/I)

Count rate by liquid scintillation method vs 222Rn content of water as determined by

radon bubbier and Lucas cell. (This figure reproduced with the permission of Dr. Howard M. Prichard.).

384

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Summary

The liquid scintillation method offers many desirable advantages for the measurement of radon: (1) Only one counting system is required to measure radon in both air and water. (2) The method is fast. A short exposure time of

24-48 hr for air measurements is ideal for short term screening. (3) The method is simple. Water samples can be counted directly while special air detectors serve both as a collection device and counting vial. (4) The method offers high throughput capabilities. Up to 200 samples can be counted and analyzed in a 24 hr period. (5) The method is accurate. It has passed EPA proficiency testing in rounds 4 and 5 and is currently a method listed by that agency for radon monitoring. (6) Measurements made with the liquid scintillation method show good correlation with established procedures for quantitating radon in air and water.

REFERENCES

Lucas, H.F. "Improved Low Level Alpha Scintillation Counter for Radon," Rev. Sci. Instrum., 28:680 (1957). Lucas, H.F. "A Fast and Accurate Survey Technique for Both Radon-222 and Radium-226," in The Natural Radiation Environment J.A.S. Adams and W.M. Lowder, Eds. (Chicago: University of Chicago Press, 1964), p. 315. Prichard, H.M. and T.F. Gesell. "Rapid Measurements of 222Rn Concentrations in Water with a Commercial Liquid Scintillation Counter," Health Physics, 33:577-58 1 (1977).

Schroeder, M.C., U. Vanags, and C.T. Hess. "An Activated Charcoal-Based, Liquid Scintillation Analyzed Airborne Rn Detector," Health Physics, 57(1):43-49 (1989).

George, A.C. "Passive, Integrated Measurement of Indoor Radon Using Activated Carbon," Health Physics, 46(4):867-872 (1984).

CHAPTER 33

Liquid Scintillation Screening Method For Isotopic Uranium in Drinking Water

Howard M. Prichard and Anamaria Cox

ABSTRACT The radiation dose resulting from the ingestion of uranium is determined by activity rather can contribute significantly to the total uranium activity, than mass; isotopes such as without being detected by mass-based fluorometric techniques. As 2U I 238U ratios signif icantly in excess of 1 are frequently noted in groundwater, the determination of uranium on a mass basis and the assumption of isotopic equilibrium can lead to serious dose underestimation. Techniques for specific uranium isotope determination are well developed, but involve considerably more effort than fluorimetry. The relative ease with which uranium can be extracted from water into a liquid scintillation compatible solution provides an alternative approach for screening drinking water supplies. A technique originally developed for PERALS (Photon, Electron Rejecting Alpha Liquid Scintillation) systems has been modified for use in conventional LS systems as a screening test for isotopic uranium and other actinides in drinking water. While lacking the degree of alphaibeta discrimination and energy resolution available with PERALS systems, conventional units offer the advantages of large sample volume, high throughput, and widespread availability. This chapter demonstrates that with appropriate sample preparation, the resolution and background attainable with conventional systems is more than adequate to provide a simple, effective screening technique for uranium concentrations of significance to public health.

INTRODUCTION

Because of its long half-life and correspondingly low specific activity, 28U is

by far the most abundant uranium isotope in water on a mass basis, but the radiation dose resulting from the consumption of the water is determined on an activity basis. Other uranium isotopes, notably 234U, can contribute significantly to the total uranium activity without being detected by existing massbased fluorometric techniques."2 The geochemical mobility of 234u is often enhanced by the alpha recoil energy accompanying its formation, and 234U/ 238 ratios significantly in excess of 1 are frequently noted in groundwater."2 The determination of uranium on a mass basis and the assumption of isotopic

equilibrium can therefore lead to a serious underestimation of the radiation 385

386

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

dose from drinking water consumption. Techniques for specific uranium isotope determination are well-developed but involve considerably more effort than fluorimetry. The high uranium extractibility from aqueous solutions into organic solvents is well known and has long been exploited in uranium recovery processes. Uranium in natural waters is nearly always in the form of the uranyl ion (UO2) , and is often complexed with carbonate or OH ions.34 The complexes are destroyed below pH 4, leaving the simple uranyl ion in solution. At low pH, this ion is very readily extracted from water into toluene containing 60 g/L of HDEHP (bis[2-ethylhexyl] phosphoric acid), a scintillation compatible compound.5 Liquid scintillation is therefore an attractive option which offers the possibility of a single step process: directly concentrating uranium into the counting medium. The procedure described in this chapter is derived from a solvent extraction, alpha liquid scintillation technique developed at Oak Ridge National Laboratory and published in "Alpha Counting and Spectrometry Using Liquid Scintillation Methods."5 The set of techniques outlined in that document rely on high resolution energy discrimination and pulse-shape discrimination available with the PERALS system, which uses small (ca. I mL) volumes of highly purified organic scintillators.

The high energy resolution and low backgrounds achievable with the PERALS system make this technology extremely attractive for many applica-

tions. However, the necessarily small liquid scintillator volumes and the amount of sample processing involved potentially limit the application of this approach to large scale, low specific activity drinking water screening programs. To meet this particular need, we have "scaled up" the extraction meth-

ods for use with conventional liquid scintillation systems, which have the advantages of large (20 mL) sample volume, automatic operation, and widespread distribution. The disadvantages of the commercial systems are low energy resolution, relatively high backgrounds, and less effective alpha/beta discrimination. However, in the restricted context of screening potable water for uranium, these disadvantages prove to be relatively unimportant. Furthermore, the energy resolution of commercial systems can be enhanced by a few simple steps to the extent that more definitive spectral analyses of positive samples can be performed.

SCREENING WATER SAMPLES FOR TOTAL URANIUM ACTIVITY

The main purpose of a screening program is to show that water is suitable for consumption, in other words, to demonstrate the absence of uranium above a particular concentration. This approach greatly simplifies analytical requirements in the (presumably) great majority of cases where gross extractable radioactivity is in fact below an established level.

SCREENING METHOD FOR URANIUM IN WATER

387

Method

To screen for total uranium activity, a 1 L sample of water is treated by aeration or boiling to remove radon, which is highly soluble in toluene. The sample is then brought to approximately pH 1.5 with HNO3 and agitated with 20 mL toluene containing 60 g/L of HDEHP and an appropriate fluor concentration. After a few minutes of contact, the phases are allowed to separate and

the extractant scintillator drawn off for counting. If the initial solution had not been purged of radon three or more hours prior to extraction, the extracts should be held at least 3 hr before counting to allow for radon daughter decay. A screening count is then taken in a region of interest empirically set around the uranium alpha peaks (4.2 and 4.8 MeV).

Extraction and Counting Efficiency Figure 1 shows the spectra of a gels incorporating 2 mL aliquots of a spiked

sample containing 412 Bq/L of natural uranium alpha activity (A) and a second aliquot taken after liquid-liquid extraction (B). The net alpha count rate in the second gel indicated an activity of 21 Bq/L in the aqueous phase, or a deficit of 391 Bq/L. The net alpha activity of the extract was 376 Bq, with

another 6 Bq recovered after a secondary organic stripping of the aqueous phase. The recovered activity is in good agreement with the deficit, indicating

that the alpha counting is virtually quantitative in the extract, as might be expected from general principles. The ratio of activities in the aqueous phases after and before extraction is 21:412, or 5. 1 o, indicating a 95 Wo recovery.

Background and Detection Limits

In a narrow region of interest around the uranium alpha peaks, the blank counting rate was found to be 3.2 cpm on our system. While this is still higher

than that of the PERALS unit, the twentyfold increase in sample volume available with conventional LS systems produces adequate statistical reliability at concentrations of regulatory interest. (Some users may prefer a wider window with higher background but more tolerance for quench variation.) Under these conditions, the minimum true detectable activity (MDTA),6 is given by:

MDTA = g(ka + kb)(CB)°5

g (K + Kb) C5 where

g = calibration factor (Bq/count), CB = total measured background count,

and ka

and kb = number of normal standard deviations associated with the probabilities a and b of making Type I (false indication that activity is present) and Type II (false indication that activity is absent) errors.

0475

HET:

Prior to Extraction

CPUS'S

'JIIlII H

B:

t1 El:

After Extraction

Figure 1. Aliquots of natural uranium spike solution in gel prior to liquid-liquid extraction (A) and after extraction (B). Counting times, scales, and regions of interest are identical.

RUSS:

A:

SCREENING METHOD FOR URANIUM IN WATER

389

The approximation holds if the background is stable and the square root of the total observed background counts is much greater than ka. For screening measurements, it is more important to avoid a Type II error than it is to avoid a Type I error, so we might choose to set a = .05 and b = .025, with ka = 1.645

and kb = 1.960. If the sample volume is 1 L and combined extraction and counting efficiency is 90%, then a 10 mm screening count would be expected

to produce (10 x 60 x 0.9) or 540 counts per Bq/L, and g = 1/540 = 0.00185. The background count in 10 mm, CO3 would on the average be (3.2 x 10) = 32 counts, and ka + kb = 3.605. The minimum detectable true activity would then be:

MDTA = 0.00185 x (3.605) x (64)0.5 = 0.0377 Bq/L (1.02 pCi/L) SPECTROSCOPIC ANALYSIS OF POSITIVE SAMPLES

As noted above, the demonstration of the absence of activity poses less of an

analytical challenge than the identification of the nuclide(s) producing an excess count rate. Several options are available for samples that exceed a predetermined screening level. One is to process the extract for high resolution alpha spectroscopy with either the PERALS system or solid state detectors. However, a relatively modest improvement in the energy resolution of conventional LS counting, combined with chemical and radiological considerations, is sufficient to identify and quantify the excess activity in many cases.

Resolution Enhancement Because most liquid scintillation systems are multi-user devices, our efforts to further improve resolution have been directed to the vial and its contents, rather than on modifications to the instrument itself. As noted by McDowel,'5 a number of simple modifications can be made to improve the energy resolution potential of a liquid scintillator. Sparging the organic scintillator with inert gas reduces oxygen quenching, thus increasing both absolute light output

and energy resolution. Total light yield and energy resolution are also increased by adding of 200 g/L of naphthalene to the primary solvent and by using PBBO (2-(4'-biphenylyl)-6-phenylbenzoxazole) instead of PPO as a fluor.

The substitution of a translucent for a transparent counting vial also has a marked effect on the alpha spectrum. (This was brought to my attention by Bernard Cohen7 in the context of our long-standing interest in detecting radon in air by activated carbon adsorption and subsequent liquid scintillator desorption. A similar effect is noted in Donald Horrocks's 1972 text on liquid scintillation.) The diffuse transmission of light through the translucent material decreases spectral degradation due to photocathode nonuniformity, thus improving resolution. Figures 2 and 3 show the spectra of 238U and 234U in

Figure 2.

?h-234

Th-234,

P.-234.:

2.29 N.Y I.t.

Alpha spectra of 238U (4.2 MeV) and 234U (4.7 MeV) obtained with conventional liquid scintillation (18 mL). The beta spectra of the short-lived intermediates are also shown.

0.103 N.Y I.t. 0.193 N.Y I.t.

Co.bid Aiph. P..k

U-234

4.2 N.Y A1pb

4.77 N.Y Alpb.

F.-234.:

2.29 N.Y a.t.

Spectrum of the solution shown in Figure 2 after transfer to resolution enhancing vial.

0.193 N.Y l.t.

Th-234

Figure 3.

0.103 N.Y $.t.

Tb-234

V-238

392

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

NATURAL URANIUM, REAGENT GRADE 1.88 Counts

1,28

5ESOLUTION (GLASS VIAL) 0.60

0.00 3.50

2.48

4.00

4.50

5.00

5.50

lIeU

t

Counts

1.60 +

10% RESOLUTION (PLASTIC VIAL) 0.80

0.00 3.60

4.00

4.48

4.88

5,0

5.68

Net

1 XX

Counts

38 X X

1.5+

3.60

5% RESOLUTION (PERALS SYSTEM)

x

X

x X

X

X

x

4.00

4.40

4.80

5.28

5.60

Met

Figure 4. Simulated alpha spectra at three levels of resolution.

SCREENING METI-IOD FOR URANIUM IN WATER

AruRAL URANIUM, REAGENT GRADE

3,0t Counts

2.0 RESOLUTION 1.0

9.0 3.50

4.99

4.59

5,99

5.5

MeU

3,0+ Counts

2.9 +

8% RESOLUTION 1.0

4.00

4.40

4.89

5.20

5.60

HeU

3,6 Counts

2.4

7% RESOLUTION

)4eU

Figure 5.

Simulated alpha spectra at three levels of resolution.

393

394

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS U-238 & (J-234, 7,5

1

:

3

.

Counts

5.0

9 % RESOLUTION 2.5

0,0 NeU

7.5 Counts

5.0 +

8% RESOLUTION 2.5

0.0 + MeU

,0 Counts

60+ 7% RESOLUTI 3.0

00f 3.50

4.90

4.59

5,90

550

MeU

Figure 6. Simulated alpha spectra of 2°U and 2U in disequilibrium at three levels of resolution.

SCREENING METHOD FOR URANIUM IN WATER

395

argon-sparged extractant scintillator obtained in a glass and in a slightly modified polyethylene vial respectively. The beta spectra are largely unchanged, but the alpha resolution in the polyethylene vial is considerably improved. The alpha resolution is approximately 10 of the peak energy, just about midway between the resolution obtained in clear glass on a conventional system and the resolution available with a PERALS system, as shown in the simulated alpha spectra in Figure 4. It is not yet known how much more resolution can be obtained on existing conventional LS systems by modifying the solvent and counting vial, but even if little additional resolution can be achieved, the level shown in Figure 3 is useful. While 238U and 234U are not completely resolved, the total alpha activity is obtained, and the degree of disequilibrium can be estimated from the relative peak areas. The importance of relatively small gains in energy resolution is illustrated in Figures 4 through 6. These figures depict synthetic alpha spectra generated by summing normal distributions with means corresponding to the 238J, 235U, and 234U alpha peaks, full widths at half maximum at the stated percent of mean energy, and amplitude at the stated equilibrium ratios. 235U is normalized to its natural abundance with respect to 238U; while not resolved at any level shown, the presence of its alpha emission between that of the other two isotopes diminishes the resolution to some extent. Figure 4 shows the two main uranium alpha peaks unresolved at 15% percent resolution, which may be compared with Figure 2, the actual alpha plus beta spectrum of our uranium extract in a glass vial. Also shown is a simulation of the 5% resolution

spectrum obtainable with the PERALS system and the 10% resolution obtained with a slightly modified high density plastic vial in our conventional LS system. Figure 5 shows the projected effects of improving resolution to 9, 8, and 7%. Each incremental gain is seen to produce noticeable effects. The importance of gains of this magnitude is further illustrated in Figure 6, in which uranium in a state of disequilibrium is viewed with 9, 8, and 7% resolution.

Nuclide Identification The 10 resolution currently available permits discrimination between uranium isotopes and many of the other alpha emitters of interest. Table 1 lists the nuclides of the 238U and the 232Th decay series that emit alpha particles with

energies between 3.75 and 5.25 MeV. These nuclides could not be readily distinguished from 238U - 234U at the present level of resolution. However, the

screening extractant HDEHP does not extract radium from the aqueous phase. Furthermore, except in one unusual well, 232Th and 23Th concentrations over 0.1 pCi/L have not been found in potable water.2 In short, natural alpha activity from 3.75 to 5.25 MeV extracted from drinking water with HDEHP is

quite likely to be due to uranium. Table 2 lists common naturally occurring nuclides with alpha energies greater than 5.25 MeV. These nuclides are thus readily distinguishable from

396

Table

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS 1.

Natural Radlonuclides with Alpha Emissions Between 3.75 and 5.25 MeV Alpha Energy Abundance

Nuclide 232Th

Half-Life

(Mev)

(%)

1.4x1010Y

3.95 4.01

24 76

238ua

4.5x109Y

4.15 4.20

25 75

234ua

2.5x 10 Y

4.72 4.77

28 72

8.OxlO4Y

4.62 4.68

24 76

1.6x1O3Y

4.62 4.78

95

226Raa

5

3U series nuclide.

uranium at the 1Oo energy resolution currently available. 222Rn and its shortlived progeny can be easily excluded by appropriate sample handling. 224Ra would not be extracted into HDEHP, but would build up with a 3.6 day halflife into any 228Th extracted. A month after isolation, the 228Th alpha emissions will be accompanied by several high energy alphas from its progeny, including one at 8.78 MeV. As noted above, however, it is quite unlikely that significant levels of 228Th would be present in a drinking water sample, except through ingrowth (half-life = 1.9 years) into older samples containing 228Ra. The presence of significant amounts of 228Ra in even a fresh sample would cause a "false positive" in the initial screening test due to the rapid ingrowth of 228Ac, a complex beta emitter with a major portion of its beta emission overlapping the alpha window. Subsequent medium resolution spectrometry would clearly disTable 2. Natural Radionuclides with Alpha Emissions Above 5.25 MeV

Abundance Nuclide 2l0p0a

224flab

222fla,b

Half-Life

Alpha Energy

(%)

138.4 D

5.30

100

1.9 Y

5.34 5.43

28

3.6 D

3.8 D

a238(J series member. blncludes radiations from short-lived daughters.

71

5.45 5.68 6.05 6.09 6.29 6.78 8.78

100 100 36

5.49 6.00 7.69

100 100 100

6

94 25 10

SCREENING METHOD FOR URANIUM IN WATER

397

Table 3. Important Artificial or Enriched Alpha Emitters Nuclide

Half-Life

Alpha Energy

7.lxlO8Y

4.37 4.40 4.42 4.56

Abundance (%) 18

57 5 5

237Np

2.1 x 106 Y

4.8ca

86

242PU

3.8x105Y

4.86 4.90

24 76

2.4x104Y

5.10 5.14 5.16

15 73

5.44 5.49

13 86

458Y

12

acomplex peaks near 4.8 MeV.

tinguish between uranium isotopes and actinium. This, in effect, leaves 210Po as the only common natural alpha emitter in this energy range likely to be present in a timely HDEHP extraction from potable water. If artificial transuranic nuclides might be present in a sample, the situation becomes more complicated, as illustrated in Table 3. 241Am alphas are energetic enough to be distinguished from 234U at lOve resolution, but 239Pu is only marginally distinguishable. 242Pu and 237Np would be interpreted as uranium unless higher resolution spectroscopic techniques or chemical separations are employed.

REFERENCES

Aieta, E.M., J.E. Singley, A.R. Trussel, K.W. Thorbjarnson, and M.J. McGuire. "Radionuclides in Drinking Water: An Overview," J. Am. Water Works Assn., 79: 144-152 (1987).

Hess, C.T., J. Michael, T.R. Horton, H.M. Prichard, and W.A. Coniglio. "The Occurrence of Radioactivity in Public Water Supplies in the United States," Health Phys. 48: 53-58 (1985). Cordfunke, E.H.P. "The Chemistry of Uranium," (Amsterdam: Elsevier Publishing Co., 1969). Sorg, T.J. "Methods for Removing Uranium from Drinking Water," J. Am. Water WorksAssn., 80L:105-11l (1988).

McDowell, W.J. "Alpha Counting and Spectrometry Using Liquid Scintillation Methods," Nuci. Sci. Series, NAS-NS-3116, (U.S. Dept. of Energy, January, 1986.) Altshuler, B. and B. Pasternack. "Statistical Measures of the Lower Limit of Detection of a Radioactivity Counter," Health Phys. 9:293-298 (1963). Cohen, Bernard, University of Pittsburg, Personal Communication, 1988.

CHAPTER 34

Assessment and Assurance of the Quality in the Determination of Low Contents of Tritium in Groundwater

C. Vestergaard and Chr. Ursin

INTRODUCTION

In the last few years it has become clear that groundwater resources are in

danger of becoming unfit sources of drinking water as a result of human activities, e.g., waste dumps for hazardous materials and agricultural activities involving extensive use of fertilizers. In Denmark, the environmental protection authorities have launched a major program for monitoring and ensuring the quality of groundwater. The program involves periodic measurements of some 80 parameters of interest, e.g., nutrients, organic pollutants, trace and ultratrace metals, and tritium, in approximately 1000 positions covering the types of aquifers seen in the country.' Tritium released into the atmosphere by the testing of thermonuclear weap-

ons in the 1950s and early 1960s, and subsequently incorporated in atmospheric water and precipitation, has proven very useful in hydrological research over the last decades. In hydrological research the tritium content is often expressed in tritium units (TU), 1 TU being equivalent to 1 tritium nucleus per 1018 hydrogen nuclei and approx. 0.118 Bq/L. In many cases a profile of the tritium content in the groundwater, as shown in Figure 1, gives a

quite precise dating of the groundwater and furthermore can give valuable hydrological information. Figure 1 shows tritium profiles from two geologically different locations common in Denmark. The vulnerability of a well in case of the threat of polution may be assessed by a single tritium measurement. A content of tritium approaching 1 TU, would indicate that the water originates from the time before the atmospheric bomb testing; consequently, that

resource is relatively well protected. The interpretation of tritium data is beyound the scope of this chapter, but examples of such work may be found in the IAEA series of publications on isotope hydrology.2 Awareness of the need to ensure and document the quality of environmental 399

400

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

TRITIUM PROFILES 0 0

50

Tritium 100 [TU]

SBV 5

10

50

0 15

Tritium 100 [TU]

15

Ti 20

20

-

[m] depth 25

30

-

35

[ml depth Figure 1. Tritium in groundwater profiles from an area by Syv Baek, Zealand (SB V), a glacial clay deposit (moraine), and an area by Rabis Baek, Jutland (T I) a sandy aquifer, 1988.

measurements is growing as the consequences of descisions based upon such measuremnets for human health, quality of life, and economy are becoming more widespread and complex. This chapter reports our attempt at assessing and assuring the quality of measurements we believe will be of growing importance in the work done to understand the mechanisms of groundwater pollution in the years to come.

QUALITY ASSESSMENT AND ASSURANCE

401

Table 1. Summary of Method Electrolytical Enrichment: Distillation of water sample. Addition of Na202. Electrolysis in IAEA cell (Mild steel cathode, stainless steel anode), volume: 300 mL, 850

Ah, t = 0-5°C. Enrichment Factory (EF): approximately 14. Distillation of enriched sample, volume: approximately 20 mL. Liquid Scintillation Spectrometry: - 12 mL sample in 10 mL Pico-Fluor LLT (Packard). - Low potassium glass vials. - Spectrometer: Packard 2000 CNLL.

- Counting time: 300-500 mm. t = 15°C. - Efficiency (): approx. 0.18. - Background (tritium free water): 2.4 CPM.

METHOD

The method for determining the content of tritium in groundwater is well established and commonly used. It is based on IAEA's method for the electrolytic enrichment of tritium in water,3 followed by liquid scintillation counting. This method is regarded as applicable to the determination of tritium contents in water samples down to 1 TU.4 Table 1 gives an outline of the method as it is routinely applied. It should be noted that multiple calibration standards are used and that the enrichment of samples with regard to tritium in the electrolysis is determined experimentally in each run by five identical samples of tritiated water. The liquid scintillation counting is done on uniformly quenched

samples so that errors introduced by incorrect or poorly defined efficiency correction are avoided. QUALITY ASSURANCE

Probably the most important aspect of any measurement quality is the reliability of its associated statement of uncertainty, which must be an integral part of reporting any result. This aspect has been discussed thoroughly in a book on neutron activation analysis in which it is shown that " . . . the precision of a single analytical result is determined by the method with which it is found. When all sources of variability are properly taken into account, the estimated precision will account for all the observed variability of analytical results."5 This would seem to hold for most types of measurements and certainly for the measurements discussed here. As can be seen in Table 1 several processes are involved in a measurement; that may contribute to the observed variability. The processes are the calibration of the spectrometer, the handling of samples, including the electrolytical enrichment, the determination of the enrichment factor, and finally the count-

ing of the sample. This is also indicated by the expression used to convert counting results into tritium content (A), in the original sample:

402

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

A = (CCb) *

* EF- * K

Where C is the observed count rate in the sample, Cb is the background count rate, is the the efficiency of the spectrometer, EF is the degree of enrichment by the electrolysis, and K is a factor for unit conversion, decay time correction etc. In our method EF is determined by the actual enrichment of five identical samples of tritiated water in each run; thus: EF = Ce * C

Where C0 is the mean count rate of the sample of tritiated water before electrolysis and Ce is the mean count rate of the same sample after electrolysis.

An estimate of the standard deviation of an individual result, SA may be calculated from this expression: SA

= A*((s + s)/(C - Cb)2 +

(s/)2

+ (s0/c0)2

(s/C)2)112

Contributions from the counting process, s and s(.b, may be calculated from the counting statistics assuming identity between the total number of counts observed and variance. In estimating the contributions from the enrichment factor, and s0, and calibration, s, Shewart control charts of standard deviations are used. Figure 2 shows such a control chart for the relative standard deviation of .

It can be seen how a minor modification of the spectrometer, involving improved grounding, has improved the precision of the calibration; over the last 10 runs the precision has been homogeneous with a relative standard deviation of 0.4%. In a similar way the contributions from the enrichment may be found. The result is summarized in Table 2. When the standard deviation of the single

result has been estimated, the ability of this estimate to account for the observed variability may be tested by the analysis of precision, using the T statistic. In the case of independent duplicate measurements of M different samples, the statistic T is determined as:5 M

((A1 - A,)2 / (sf12 +

T

s122))

i= I

Where A1 and Al2 are the two independent results and s and s, are the estimated standard deviations of the two results. T is approximated by an X2 distribution with M degrees of freedom. Figure 3 shows T as a function of the number of duplicate determinations made over a 10 month period, in which approximately 10% of the determinations were done in duplicate. The individual results range from below 1 TU to 45 TU. As T is not significantly different from an X2 distribution, it may be assumed that the estimated standard devia-

Figure 2.

0

0.18

0.81

1.41

10

20

I

0.10

run number

T

0.42

0.73

Control chart for the relative standard deviation of the counter efficiency, , as calculated from three indepent calibration standards per run. The vertical dotted line coincides with the time of modification of the spectrometer. Control limits are given as 0.05 .

0

1.0 -

2.0 -

404

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 2. Individual Error Contributions Rel. Std. Dev. I%1

Efficiency Tritiated water

0.4 1.2

after enrichment

4

tions of individual results do account for the observed variability in the duplicate results; consequently, no other significant source of variability is present. In Figure 4 the calculated standard deviation for the samples is shown as a

function of tritium content. sA/A varies between approximately 0.5 for a content of 1 TU to approximately 0.05 at 45 TU. When the random error has been determined, what remains to be shown is that systematic errors or bias do not affect the results. An important source of bias would be the presence of a significant blank value in the determinations. This is checked by routinely measuring samples of supposedly tritium free groundwater. These measurements give results well below 1 TU and it is assumed that the blank can be neglected. The absence of systematic error can be demonstrated by measuring reference samples with certified tritium contents, similar to those usually measured in the laboratory, or by the participation in interlaboratory comparison exercises. The laboratory has participated in an intercomparision organized by the IAEA in which no systematic error was detected.6

T

40 A 0.95

30 20 0.05

10 0

I

0 Figure 3.

10

I

I

I

I

I

I

20

M

The statistic T, calculated from duplicate determinations plotted as a function of M, the number of duplicate determinations. Dotted lines represent the 5% and 95% fractiles in the X2 distribution with M degrees of freedom.

QUALITY ASSESSMENT AND ASSURANCE

0

10

20

30

40

405

50

Tritium content [TU] Figure 4. Calculated standard deviations as a function of tritium content.

CONCLUSION

The measurement quality of low tritium contents in groundwater can be controlled by comparing expected and observed variability in duplicate determinations of randomly selected samples. The analysis of precision shows that the identified sources of random error account for the observed variability, and the method may be said to be in statistical control.

406

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

ACKNOWLEDGMENTS

The authors wish to thank Ms. Lisbeth Just and Ms. Helle Schmidt Pedersen from the Danish Isotope Centre for performing the tritium determinations. The groundwater samples shown in Figure 1 were made available by the Geological Survey of Denmark, as part of a project sponsored by the National Agency of Environmental Protection, Denmark.

REFERENCES

"Monitoring Programme of the Aquatic Environment Plan," (in Danish) Environ-

mental Project 115, National Agency of Environmental Protection, Denmark (1989).

Andersen, L.J. and T. Sevel. "Six Years Environmental Tritium Profiles in the Unsaturated and Saturated Zones, Gronhoj, Denmark," in Isotope Techniques in Groundwater Hydrology 1974, Vol. I, (Vienna: International Atomic Energy Agency, 1974).

"Enrichment of Tritium in Water by Electrolysis," IAEA Isotope Hydrology Laboratory Vienna (1972). Florkowski, T. "Low-Level Assay in Water Samples by Electrolytical Enrichment and Liquid-Scintillation Counting in the IAEA Laboratory" in Methods of LowLevel Counting and Spectrometry, (Vienna: International Atomic Energy Agency, 1981).

Heydorn, K. Neutron Activation Analysis for Clinical Trace Element Research, Volu,ne I (Boca Raton, FL: CRC Press, 1984), pp. 167-205. Hut, G. "Intercomparison of Low-Level Tritium Measurements in Water," (Vienna: International Atomic Energy Agency, September 1986.

CHAPTER 35

Use of Liquid Scintillation in the Appraisal of Non-Radioactive Waste Shipments from Nuclear Facilities* William L. McDowell

ABSTRACT The generation of non-nuclear waste at the Savannah River Site (SRS) presents a special problem for analysis. As all waste must be suspected to contain radioactive materials prior to appraisal, large volumes of waste are retained on site prior to off-site shipment for incineration or burial. The amount of flammable waste solvent stored on site in late 1987 threatened to place the Savannah River Site in violation of EPA and OSHA regulations. One of the major contributors to this burden was some 2000 55-gallon drums of used paint solvent and waste paint. The Environmental Technology Section at the Savannah River Laboratory was charged with development of quick, reliable, and simple method for measurement of the maximum possible activity of paint waste samples. The <2 nCi/g guideline set by the Department of Transportation* is the upper limit of the radioisotope contamination for removal of the waste from the site.

Owing to the possible presence of tritium and the desire for a relatively quick analysis technique, liquid scintillation was used as the method of choice. Novel methods are presented for dealing with various phases (organic, aqueous, and solid) present in the samples and for determining the optimum dilution for severely quenched samples. Arguments are presented for the fragmentation of solid particles by ultrasonic vibration, and mathematical corrections are given for the escape fraction from solid particles.

INTRODUCTION

Nuclear facilities generate waste materials that must be removed from the site or disposed of on site. The characterization, isolation, and confinement of high level nuclear waste is currently the subject of much study. By contrast, *This paper was prepared in Connection with work done under Contract No. DEACO9-88SR18035 with the U.S. Department of Energy. By acceptance of this paper the publisher and/or recipient acknowledges the U.S. Government's right to retain a nonexclusive, royality-free license in and to

any copyright covering this paper along with the right to reproduce and to authorize others to reproduce all or part of the copyrighted paper. *Code of Federal Regulations, 49CF173.403(g), June 1986. 407

408

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

relatively little attention has been given to non-nuclear waste from nuclear facilities.

On the Savannah River Site (SRS), nonradioactive or clean waste is normally assayed with a survey meter to appraise the extent of surface contamination. Because of the nature of the operations conducted at SRS, the majority of the waste generated is treated as contaminated waste until proven otherwise. Before such waste can be shipped off site, more rigorous assay must demonstrate compliance with government standards. In particular, it is necessary to appraise the samples with regard to the Department of Transportation (DOT) guideline for a total activity of 2 nCi/g of waste material.' Furthermore, the assay methods must be sufficient to comply with Environmental Protection Agency (EPA) regulations to prevent overcrowding of waste storage locations;

thus, the present work developed efficient liquid scintillation methods for meeting these requirements.

BACKGROUND

The waste material at SRS is accumulated in 55-gallon drums and stored in interim storage facilities. These drums must be assayed in a relatively short period of time so that off-plant shipment can proceed in an orderly fashion, and the storage inventory complies with EPA regulations. All of this material was considered incinerateable nonradioactive hazardous waste. The present work examines 900 55-gallon drums containing paint solvent/waste which were to be shipped off plant for incineration. These drums contained up to three distinct phases of material with widely differing components. The materials were primarily mixed hydrocarbon solvents. A significant fraction of the material included paint solids, dust, and dirt. Additionally, many of the materials contained a measurable amount of water. All the samples contained the liquid organic material, but the other two phases were not always present. For these samples, neither the specific composition nor the contaminating isotopes could be assumed. The radiometric methods used for these waste assays had to measure the total activity of waste samples in a relatively short turn around time. Sample preparation schemes such as fractional distillation for solvent cleanup, ashing of solid residue, digestion, solvent extraction, and either electroplating of a purified sample or pelletization of the original ash would have been simply too time consuming. Methods that have been used for total low-level analysis of waste streams from other facilities also have such limitations. Traditional electroplating and alpha counting by semiconductor detector is a sufficiently sensitive method; however, this often requires lengthy chemical separations, and occasionally, sample ashing prior to plating of the sample. Gross plating of the original sample rarely proves satisfactory for such situations owing to the mixed character of the material and the time required. Because a large number of samples

NON-RADIOACTIVE WASTE SHIPMENT APPRAISAL

409

must be processed on a continuing basis, this was not considered a viable alternative. Tritium was considered one of the most likely contaminants of liquid waste materials resulting from on-site, processes. On-line methods for tritium detection in organic matrices do exist,2 however, because much of the waste also contained a significant fraction of solid material, neither the on-line method

previously mentioned nor simple liquid scintillation of untreated samples could be relied upon. The total radioactive concentrations of the various waste materials had to be measured; thus, analysis for individual isotopes was not required. Based on the SRS process history, the major radionuclides were expected to be alpha, beta, and gamma emitters, although minor contributions from electron capture and positron decay could not be ruled out. Primary manmade radionuclides could be fission products, tritium, transuranics including plutonium, and neutron activation products. Liquid scintillation is very efficient in detecting the charged particle emitters and thus yields the total activity for samples devoid of radionuclides that decay by electron capture. Some of the neutron activation products do decay by electron capture, and their activities are best appraised by gamma spectroscopy due to their weak charged-particle emissions. The percentage of activity due to electron capture was judged to be fairly low and therefore liquid scintillation was enlisted as the primary technique for determination of "total activity." High purity, germanium diode, gamma spectrometry (HPGe spectrometry) was used as a consistency check to ensure that assumptions concerning cc-gamma activity were correct. HPGe spectrometry of the samples used routine methods; however, liquid scintillation required some method development to overcome the various problems associated with these samples. Liquid scintillation can provide a measure of the total activity of a sample. All ionizing radiation can be detected by liquid scintillation; thus all isotopes could be expected to yield some degree of signal in a liquid scintillation sample. Of course, the detection limit varies for counting time and isotope of interest. As an example, tritium is one of the lowest energy beta emitters and its detection limit on a modern beta liquid scintillation spectrometer is about 11 picoCuries per sample given that no quenching is present and that the sample is colorless. HPGe spectrometry is ideal for assay of gamma emitting nuclides, especially for gamma energies above 100 keY. The detection limits of HPGe spectrometry depend on the isotope and the counting geometry. In the Savannah River Site/Environmental Technology Section (SRS/ETS) Underground Counting Facility (UCF) the absolute sensitivity for 60C in soil/water may be as low as 0.007 mBq/cm3, depending on sample configuration. This value is for a 24 hr counting time, and lower levels are attainable with longer counting periods. The samples must conform to a limited number of fixed geometries for the accurate use of HPGe spectrometry. For many situations, low-level HPGe spectrometry is adequate for determi-

410

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

nation of total activity. In situations where the primary emission of the contaminate is a strong gamma-ray and the material can be made to conform to a known geometry, this technique is quite useful. Generally, little or no sample preparation is required for HPGe spectrometry. However, for appraising total activity, a significant amount of sample information is required prior to measurement. Also there must be a reasonable assurance that the primary contaminate is not tritium or some isotope such as 129J which is not suitable for HPGe spectrometry. ANALYSIS

Sample Preparation Because of the large number of drums, it was decided to define an individual drum as a sampling unit and composite samples from these units were put in groups of N = 2 drums each. By using powers of two for the composite, a sample that was subdivided to trace a possible source of contamination (and minimize the amount of material that would have to be treated as low-level waste) could be subdivided down to a single barrel if necessary without the necessity of testing every single barrel in the original composite. This compositing allowed for much faster processing of the samples, but it did lead to increased background values as the dilution factor of N was now included in the detection limits. That made the acceptable maximum value for a single composite sample 2/N nCi/g. This smaller limit arose from the consideration that, in a single composite, there could be a single contaminated drum while all

the other drums were clean. If N were too large, the analysis limit of 2/N would be too close to the detection limit of the instrument. Because of this, the maximum number of sampling units (55-gallon drums) in a single composite

was determined by the background level of the instrument to be used. The number of sampling units in a single composite had to be selected so that 2/N nCi/g was higher than the background of the instrument. For most applications 16 or 20 was found to be a usable value for N. To create a composite sample, the drum contents were agitated, and then a long cylinder was used to retrieve a column of material extending from the top to the bottom of the drum. Uniform 25 mL aliquots were taken from each of these increments, and these subsamples were combined to give the analysis composite. For the samples that consisted of multiple layers or phases of material, the components were separated into "homogeneous" samples for analysis; the entire sample could not be counted as a single entity or the heterogeneity would effect the results. After the phases were analyzed separately, the total activity would be found by factoring the relative masses of the phases into the summation of the individual activities. The relative masses of the phases were determined by measuring the volume (as a cylinder) of a single phase and using the empirically determined average density to calculate the mass of the phase in question.

NON-RADIOACTIVE WASTE SHIPMENT APPRAISAL.

411

For the organic liquid layer of the paint samples, quenching was dependent on the amount of the sample present in the cocktail. With the proper dilution of such a sample, it is possible to achieve an optimum count rate since both quenching and activity level would increase with concentration. Conceptually, this is shown in Figure 1. The resultant, optimum count rate, is the result of the increasing efficiency and the decreasing count rate with dilution. Assuming that the count rate (C), is the result of both the efficiency and the activity,

C=EA=Enm

(1)

where n is the specific activity and m is the mass.

3m

= En +3mnm = 0

(2)

E

A

Dilution

A: Activity

E: Efficiency C: Observed Count Rate Figure 1.

Graphic representation of dilution effects.

412

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Substituting Equation 2 for the optimum mass we get,

m = m0

-E =

(3)

To estimate an expression for E, it is assumed that the mass M of the sample in the cocktail reduces the scintillation light collection efficiency according to the

Beer-Lambert law. A differential form of this law, as applied to the present work is

= -LaM

(4)

where L is a constant. This type of fractional decrease in efficiency with increasing sample mass is probably reasonable for a variety of quenching mechanisms; thus, Equation 4 was used for quenching in general as well as its integrated form

E = EoM

(5)

where Eo is a constant. From Equations 3 and 4 it is seen that Mo = i/L. Thus Mo could be determined by solving Equation 5 for L, using Eo = 1 for the infinitely dilute (M - 0) case and E = 0.001 from tritium spike measurements. The resulting value for L is 2.24 gm-'. Using this value for L a result of

0.4 g was arrived at as the optimum sample size. With this as a starting point, a few empirical trials determined that the optimum dilution for the majority of the cases was 250 L of organic material to a single cocktail. Each sample was combined with 20 mL of Opti-Fluor® scintillation cocktail and sufficient water to give a total volume of 23 mL. Any aqueous phase was collected separately. This was achieved by reaching through the top organic layer with a disposable polyethylene pipette and transferring an aliquot of the aqueous material into a glass vial. The cocktail was then prepared out of this vial. The final sample preparation was identical to that for the organic layer with the exception that ito 1.5 mL of the sample was placed in the cocktail. In each case, the aqueous layer was less colored and exhibited less quenching than the organic layer. Suspended solid materials presented the greatest problem. After examining several of the samples, it was discovered that most of the solids were chips of paint from 1 to 2 mm in "diameter." There were two primary concerns with this material: (1) most alpha events that occur in the chips themselves could be buried and never reach the scintillation cocktail, and (2) light resulting from alpha events in solution could be blocked or absorbed by the relatively large paint chips. Furthermore, the material was heterogeneous and not all particles were soluble in the OptiFluor® cocktail. Thus, it was necessary to reduce the size of the particles to the point where they would: (1) allow the majority of the alpha events to escape into the cocktail, (2) disperse homogeneously, and (3) remain suspended long enough to allow effective counting.

NON-RADIOACTIVE WASTE SHIPMENT APPRAISAL

413

This dispersion and size reduction of the particles was accomplished by ultrasonically shattering the solids while they were immersed in a fairly viscous water/surfactant solution. For samples with a visible layer of solids, a polyethylene pipette was used to transfer 1.5 g of these solids to a tared glass vial. This material was diluted to a total volume of 10 mL with a 50/50 (v/v) mixture of Joy® detergent and deionized water. The vial was closed and immersed in an ultrasonic bath for a period of 15 mm. One mL of this mixture was then used to prepare the liquid scintillation cocktail for evaluation of solids in the given sample. In previous applications at SRS laboratories, Joy® had been used for

radiological cleanups and was found to harbor no abnormally high levels of activity. Also, blanks using tritium and Joy® showed no signs of quenching. All samples were made up to a uniform volume of 23 mL. In all cases, both a normal sample and a spiked sample were prepared. For this work, the spike used was a dilution of a NBS standard tritiated water solution. A single spike consisted of 100 L of tritiated water having a disintegration rate of 32,000 dpm/mL. All samples were counted for three to nine periods of ten minute durations. The number of count periods depended on the time for any luminescence or chemiluminescence of the samples to decay. All samples were counted on a Packard Tri-CARB 2000 CA/LL with the output divided into three windows, 0 to 2 keY, 2 to 18 keY, and 18 to 2000 keV. Sample preparation for the HPGe counting was much simpler. This preparation consisted of merely weighing the total sample and measuring the thicknesses of the individual layers in the bottle. In order to assure that any material on the outside of the bottles would not contaminate the SRS ultra low-level counting facilities, each bottle was placed in a tight-fitting polyethylene bag which was taped closed. The mass and height of material in each bottle was incorporated into the efficiency calculation for the HPGe detection.

ANALYSIS

Activity Calculations As previously stated, the purpose of this work was to generate upper limit values which could be used as the basis for a decision for off-site shipment of

hazardous waste. To this end, the liquid scintillation calculations were designed to yield the tritium and total radioactive concentrations. A tritium window of 2.0 to 18.6 keV and a total activity window of 0 to 2000 keV were selected for the basic analysis. Each sample and corresponding tritium-spiked sample yielded a tritium concentration (At), and a psudo-total concentration (Ar). The value of A is larger that the actual total (Aa), because the tritium calibration spike is strongly quenched relative to other nuclides. Assuming that

all other radionuclides are detected with an efficiency that is a factor of f greater than the tritium efficiency, the total activity concentration is given by:

414

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Aa = A, +

(A-A)

(6)

f

In the present work, f is defined as the ratio of the '4C and 3H efficiencies3 as

shown in Figure 2. Here the plotted quench parameter (tSIE) is defined for each sample from its spectrum with an external 133Ba source.3 The magnitude of each efficiency curve is strongly correlated with the beta energy. Accordingly, the curve for 3H betas, which range from 0 to 18 keV, is considerably lower than that for '4C betas which range from 0 to 156 keV. Effectively, all other beta-emitting radionuclides have energies comparable to or greater than that for '4C. Alpha particles are less efficient by a factor of 10 in converting their energies to scintillations; however, their higher energies (- 5000 key) result in scintillation responses that are well above those of the 14C spectrum. Consequently, all radionuclides other than tritium are treated as having the

same efficiency as

'4C;

thus making the overall calculation somewhat

conservative. The individual region of interest calculations of A1 and A used the following formula: A

-

)*IC_o B-B0

C-C BS-B)

(7)

A = A1 or A for corresponding spectral window (dpm/g) S = Tritium spike activity (dpm/g) M = Mass of Sample (g) C/B = Sample/blank count rates C0/BØ = Sample/blank non-scintillant constant background

where

rates

CS/BS = Sample/blank tritium spike count rates Calculation of Conservative Estimate and Best Estimate For each composite sample, a conservative estimate (CE) and a best estimate

(BE) were made for the total concentration of radioactivity in the entire sample. Both liquid and solid phases were treated and incorporated into a general formula:

E = F1 x A1 + C, x F, x A, where

E = Either CE or BE, as appropriate F1 = Liquid fraction of sample A1 = Activity for liquid fraction (nCi/g) C, = Solids correction factor F, = Solids suspension fraction of sample A, = Activity of solids suspension of sample (nCi/g)

(8)

NON-RADIOACTIVE WASTE SHIPMENT APPRAISAL 100

Q) C.)

uJ

40

%Efficie cy C-14 %Efficie: cy H-3

200

800

600

400

1000

Quench Parameter (tSIE)

1000

.1.

.. aiuuiii

aiuiia

100

___$1111111...$.u1Iui _______e -________

10

*uaiiii 111111111._RIllilli

saau

....-- -- s_a

_______ -

I

10

I

I.

..iiii

..

uuuu

Rililill

100

Quench Parameter (tSIE) Figure 2. Individual isotope quench curves and ratio curve.

RR111

.,*IIII 1000

415

416

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Here, only two phases are represented in the formula. In some cases, the formula was extended to handle an additional liquid phase (water). The conservative estimate uses pseudo total A for activities A1 and A,, and a conservative correction (C,) of four. As discussed earlier, A would be the total activity if all the count rate were assumed to be tritium. Because tritium is detected least efficiently, A is noticeably greater than the actual total activity Aa. The solids correction factor of 4 corresponds to a 250/o escape probability for 4.5 MeV alphas uniformly distributed within a sphere of density 1 g/cm3 and a diameter of 200 M. Actually, the escape probability for tritium betas would be even lower, but tritium is not expected to be trapped in these solids after disintegration in the ultrasonic bath. Most other betas would escape more readily than the alphas. The correction of 4 is also conservatively high as the actual diameters of the particles are at or below 100 jim, many alphas have higher energies than the values mentioned above, and no credit is taken for the distributional effects between solids and the suspending liquids. The best estimate uses actual total Aa for activities A1 and A,, and a correction factor of C, = 1. The use of Aa is an obvious selection; however, arguments for choosing C, = 1 need to be reviewed. Here a 53% escape probability is calculated for 5 MeV alphas uniformly distributed within a 100gm sphere of density 1. However, any contamination for paint solids is likely to be distributed on the surface of the spheres, because the ultrasonic breakup of the paint solids should cause fractures along regions which have been exposed to poten-

tial contamination sources. If all activity is on the surface of the l00/Lm spheres, the escape probability increases to 69%. The solids suspension contains > 60% liquid; thus, if the alpha activity were distributed proportionately to the solid and liquid fraction, the effective alpha escape probability would be >88%. The estimate should be even higher because (1) it is unlikely that all the activity is due to alphas (which have the shortest range in solids), (2) the typical liquid/solids ratio of > 100 in the scintillation cocktail favors activity in the liquid, (3) many of the particles are likely to be smaller than 100gm in diameter, and (4) any internal solids activities are not likely to be from SRS sources and thus should contribute only a minor component from natural radiation. Adoption of C, = 1 assumes that, taken together, each of the four additional factors contribute an average increase of only 3%. RESULTS

The results of the individual samples are graphically shown in Figure 3 and listed numerically in Table I. All values contain the average of at least three trials for liquid scintillation and the activity detected by HPGe analysis. Not only are the conservative and best estimates shown, but two additional values, the highest likely and highest possible activities, are shown. In these two additional cases, the assumption was made that all potential activity came from a single sample in the composite (highest possible), or that the original

NON-RADIOACTIVE WASTE SHIPMENT APPRAISAL

417

Liquid Scintillation Results 2

0

-2

-3

44

-3

-2

-1

0

2

Conservative Estimate Values -log (nCi/g) Figure 3.

Plot of all composite and sub-composite samples (darker lines indicate limits for acceptance [log of 2 nCi/gfl.

distribution of possible activity in the samples was based on a log-normal distribution with 99'o confidence (highest likely). The highest possible estimate is the product of the best estimate activity and the number of samples in the composite:

HPE = N x BE

(9)

This represented the highest possible activity for one single drum in the composite. Only in the cases where both this value and the conservative estimate were both less than 2 nCi/g were the samples deemed acceptable for off-site release.

The highest likely estimate used a factor of 501 in the place of the number of samples, N. This value results from a log-normal probabilistic distribution of the material among the drums which make up the composite.5 These values are felt to be more realistic than the highest possible value above.

HLE = 5.01 x B

(10)

418

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 1. Numeric Results for all Composite and Sub-Composite Samples BEa Composite # Samples CE8 Highest Likely Highest Possible 1

11

2

20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 10 10 20 20 10

3 4 5

6 4 8 9 10 11

12 13 14 15 16

17 18 19

20 21

22 23 24 25 26 27 28 29 30 31

32A 32B 33

34 35A 35BA 35BB 36

37 38 39 40 41

42 43

44 45 46

47 48

5 5

20 20 20 20 9 20 20 20 20 20 4 20 5

0.0007 0.7600 0.0280 0.0510 0.1070 0.1250 0.0340 0.1580 0.4370 0.1520 0.1340 0.1590 0.0620 0.0830 0.1660 0.1750 0.0860 0.6230 0.2930 0.1750 0.1880 0.1960 0.8430 0.2010 0.0870 0.1440 0.4160 0.4160 0.2650 0.31 80

0.2620 0.0620 0.8160 0.6930 0.1600 0.2710 0.1360 1.2880 0.3200 0.0760 0.2780 0.6140 0.3630 0.8600 0.2590 0.5270 0.4120 0.0540 1.0180 0.0710 0.0660

0.0006 0.0820 0.0075 0.0240 0.0690 0.0860 0.0080 0.0650 0.0900 0.0560 0.0460 0.0610 0.0180 0.0180 0.0770 0.0800 0.0210 0.0500 0.0800 0.0380 0.0590 0.0720 0.0720 0.0440 0.0280 0.0450 0.0360 0.0360 0.0970 0.0460 0.0430 0.0180 0.1300 0.0880 0.0490 0.0820 0.0670 0.1670 0.0750 0.0310 0.0770 0.0750 0.1110 0.0570 0.0760 0.0840 0.0700 0.0110 0.1180 0.0210 0.0150

0.0030 4.1080 0.0376 0.1202 0.3457 0.4309 0.0401

0.3257 0.4509 0.2806 0.2305 0.3056 0.0902 0.0902 0.3858 0.4008 0.1052 0.2505 0.4008 0.1904 0.2956 0.3607 0.3607 0.2204 0.1403 0.2254 0.1804 0.1804 0.4860 0.2305 0.2154 0.0902 0.6513 0.4409 0.2455 0.4108 0.3357 0.8367 0.3757 0.1553 0.3858 0.3757 0.5561 0.2856 0.3808 0.4208 0.3507 0.0551 0.5912 0.1052 0.0751

arid BE stand for conservation estimate and best estimate respectively.

0.0066 1.6400 0.1500 0.4800 1.3800 1.7200 0.1600 1.3000 1.8000 1.1200 0.9200 1.2200 0.3600 0.3600 1.5400 1.6000 0.4200 1.0000 1.6000 0.7600 1.1800 1.4400 1.4400 0.8800 0.5600 0.9000 0.7200 0.7200 1.9400 0.9200 0.8600 0.1800 1.3000 1.7600 0.9800 0.8200 0.3350 0.8350 1.5000 0.6200 1.5400 1.5000 0.9990 1.1400 1.5200 1.6800 1.4000 0.2200 0.4720 0.4200 0.0750

NON-RADIOACTIVE WASTE SHIPMENT APPRAISAL

419

The grounds for acceptance of a sample are illustrated by the diagram in Figure 3. If both the conservative estimate (CE) and the highest possible estimate (HPE) were below the 2 nCi/g guideline, the sample was accepted as below DOT guidelines. If the sample fell outside the DOT guidelines on either of the two above parameters, the samples in the composite were subdivided to create two new samples (NSCOd = N/2) and two new composites made up for analysis. This was required for two composites and one subeomposite. In fact,

none of the samples which were outside the acceptance limits for the first composite were ever traced to a single "contaminated" drum. This fact supports the argument that the samples, which did initially fail to meet criteria, did so because of mathematical multiplication of background levels. CONCLUSIONS The method of phase separation, sample processing, and liquid scintillation analysis proved effective for the analysis of mixed phase, hydrocarbon based waste samples. The described method was readily applicable to a variety of circumstances and was straightforward enough to be used by staff analysts after a minimum of training. The results were obtained in far shorter time (usually less than 2 hr for a single sample) than would be required for other typical methods of analysis. All samples analyzed by the methods presented above were found to have activities below the 2 nCi/g DOT guideline for shipment of non-nuclear waste. The effects of quenching and luminescence deserve further attention in the future. Methods for luminescence approximation through graphic interpretation have been considered, as well as the use of chemical agents to decrease luminescence effects. Overall, the compositing scheme reduced the number of samples almost by a factor of 20. The ultrasonic fracturing of the solid particles provides a method for rapid analysis of this particularly difficult material. In the future, a more effective approach will be developed. Trial use of a cellular disintegration probe, as opposed to a ultrasonic bath, is underway at Waste Management Technology. Also, other materials such as Triton X-lOO would probably provide better suspension characteristics than Joy.®

REFERENCES Code of Federal Regulations, 49CFRI 73.403(g), June 1986.

Gevirts, V.B., A.A. Palladiev, and T. Yu Pautova. Soviet Radiochem. 28 (4): 462-464 (1978). Operation Manual: Model 2000CA TRi-CARB Liquid Scintillation Analyzer: Publi-

cation No. 169-3023, (Downer's Grove, IL: Packard Instrument Company, 1987), p. 5-79. Winn, W.G., W.W. Bowman, and A.L. Boni. "Ultra-Clean Underground Counting

420

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Facility for Low-Level Environmental Samples," The Science of The Total Environment, 69: 107-144 (1988). 5. Bury, Karl V. Statistical Models in Applied Sciences, (New York: John Wiley and Sons, 1975), p. 295.

CHAPTER 36

Development of Aqueous Tritium Effluent Monitor

K.J. Hoistetter

ABSTRACT A variety of techniques are being evaluated and tested in an attempt to develop a real-time monitor for low-level tritium in aqueous streams. One system being tested is a commercially available HPLC radioactivity monitor. This system uses crushed yttrium silicate as the scintillator and employs standard fast coincidence electronics to measure tritium. Laboratory tests of this unit indicate that the monitor can sense tritium at concentrations above 600 pCi/cc using a two minute counting interval. Pooling the count rate data over a longer interval, e.g., 24 hr results in a detection limit of 20 pCi/cc under constant background conditions. Unfortunately, the cells are easily plugged with debris even under laboratory conditions. To overcome this problem, a prototype system using unclad fibers of plastic scintillator as the detection medium was designed and is being tested in the laboratory. Approximately 500 1-mm in diam fibers were assembled into a flow cell, with two 51-mm in diam photomultipliers (PMT5) coupled to the ends of the fiber bundles, to detect the scintillations. Fast coincidence, pulse shaping electronics are used to minimize the single photon and dark current backgrounds. The tritium counting efficiency, background, and sensitivity will be determined in the laboratory followed by field reliability testing. The results of laboratory tests and a comparison of other types of scintillators (liquids, plastic beads and fibers, crushed inorganic, etc.) are presented. The sample preconditioning requirements (e.g., filtration, ion exchange, etc.) and known interferences (e.g., chemical and biological luminescence, natural radioactivity, etc.) for continuous monitoring of tritium in surface waters are discussed.

INTRODUCTION

Tritium is one of the most significant radioactive isotopes released to the environment by the nuclear industry. It is produced in light water reactors by ternary fission, in heavy water reactors by neutron capture in the moderator, in tritium production facilities at our defense facilities, in fallout from nuclear

weapons testing, and as a byproduct of nuclear fusion development. The releases are predominantly in the form of tritiated water (HTO), its most active 421

422

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

biologically form. While tritium is produced by cosmic ray protons in the atmosphere, the manmade quantities dwarf the natural levels)'2 The measurement of tritium is difficult due to its low energy beta emissions (maximum energy = 18.6 keV) and long half life (12.33 years).3 Most monitoring methods for tritium have relied on grab sampling and laboratory analysis by liquid scintillation counting methods. The main interferences in the labora-

tory analysis are due to color and chemical quenching in the sample. These laboratory analysis techniques have been summarized in a variety of publications.4 Real-time monitoring for tritium in the environment is not presently performed. Liquid scintillation techniques do not adapt readily to the measurement of tritium in flowing streams. A semicontinuous flow system has been tested using rapid mixing of the sample with the scintillation cocktail followed by injection into the counting chamber for a short count and then out to waste. The entire process is then repeated at short intervals to provide near real-time monitoring.5

The use of solid scintillators for tritiated water measurements in flowing streams has also been reported.5 While these detectors are less sensitive than liquid scintillation techniques, they offer three main advantages: the use of solid scintillators permit continuous monitoring, are not affected by chemical

and color quenching, and are less expensive to operate (the cost of liquid scintillation cocktail is high and disposal of the contaminated cocktail is difficult). Background

Management at the Department of Energy Savannah River Site (SRS) requested the development of an on-line monitor to detect tritium in various aqueous effluent streams. The Environmental Technology Section of the Savannah River Laboratory is conducting laboratory and field testing of various systems; they are looking for environmental levels of tritium with response time suitable for corrective action. After reviewing various flow-through scintillation systems, counting cells loaded with solid scintillators were recommended over water/liquid scintillation cocktail mixtures since: no potentially mixed waste was generated, an improved response time was achieved, the sensitivity was sufficient, and the operation was much simpler. The specific objective was to develop a monitoring system capable of detecting tritium at a concentration of 2000 pCi/cc within 2 mm, detect changes at the 2 pCi/cc level over a 24 hr interval, and to measure concentrations within l0o at the normal reactor discharge levels over a 24 hr period. The continuous monitor was not necessarily intended to replace the existing tritium grab samplers, with subsequent laboratory analyses as a means of quantifying a release. The primary purpose was to detect an abnormal release and provide early warning for emergency response. To efficiently detect tritium, one needs a detector with high surface area as the most probable beta energy is about 5.7 keY; this corresponds to an average

DEVELOPMENT OF AQUEOUS TRITIUM EFFLUENT MONITOR

423

range of only 6 m in water. Most of the beta energy is lost in the water before reaching the scintillator, and relatively few photons are produced when there is an interaction. Maximizing the surface area increases the probability of a light

producing event, while minimizing the detector volume reduces the background from cosmic rays and other beta emitters. Efficient light collection is mandatory. Beads of plastic scintillator have been suggested as a tritium detector, but

have large spacings relative to the short beta range. Only tritium in a thin sheath over the surface of the beads would be detected, resulting in a low counting efficiency and low sensitivity. Large sheets of plastic scintillator6 and thin fibers coated with anthracene7 have also been evaluated for tritium sensitivity. Crushed scintillators offer the highest surface to volume ratio and even with the decreased light collection efficiency, provides a sensitive tritium detection method.

EXPERIMENTAL

A commercially available tritium detection system was purchased for evalu-

ation at SRS. The system was originally designed to detect tritium labeled compounds as they are eluted from a high-pressure liquid chromatography column. * The detector consisted of a Teflon measurement cell filled with crushed yttrium silicate solid scintillator interposed between two PMTs. Special pulse shaping and timing electronics are provided to minimize the background and maximize the counting efficiency for tritium. A computer-based data acquisition and analysis system is also provided as part of the system. Several solid scintillation materials are offered by the vendor for use as the tritium detector: cerium-activated lithium glass, calcium fluoride, yttrium glass, and yttrium silicate. The latter was chosen due to its high light output.8 The inorganic scintillator was crushed and then suitable grain sizes were selected to optimize the counting efficiency while minimizing back pressure. A recirculation system was set up in the laboratory to circulate aqueous solutions containing tritium and various contaminants. The detector system uses a HPLC pump capable of reaching system pressures of 100 psi at wellcontrolled flow rates. Due to the burst pressure of the Teflon cell, the system pressures were carefully monitored and did not exceed this value during the testing. Several cells of differing void volumes and scintillator loadings were evaluated. A simple diagram of the monitoring system is shown in Figure 1. Backgrounds were taken by circulating demineralized water through the system. The efficiency of the system was determined using demineralized water spiked with various quantities of tritium and by comparing the monitor output with the concentrations determined from aliquots (analyzed by standard liquid

scintillation counting techniques in our laboratory). System stability and Distributed by Berthold Analytical, Nashua, NH 03063

Filters

Figure 1. Simplified diagram of a continuous aqueous effluent tritium monitor using crushed scintillator.

H20 lfl

Alarm Circuit

Counting Electronics Data Logger

Shield

Photomultiplier Tubes

Counting Cell (Solid Scintillant)

DEVELOPMENT OF AQUEOUS TRITIUM EFFLUENT MONITOR

425

Table 1. Performance Characteristics of Cells Containing Crushed Yttrium Silicate Inorganic Scintillator Cell Void Volume (cc) 0.4 0.5 1.0

Background (cpm)

Efficiency (%)

12.5 14 18

0.4 0.35

1.77

reproducibility were determined by recirculating aqueous solutions through the system for extended periods (30 to 40 days).

The output of the tritium monitoring system was recorded using the computer-based data acquisition and analysis system. Two counting channels

were set up on the system, the low-energy channel, set to encompass the tritium beta spectrum, and the high-energy channel, set to detect events above the tritium beta spectrum endpoint. Data were output at three different count-

ing intervals, a short time interval (usually 10 mm), an intermediate time interval (usually 10 hr), and a long time interval (usually 24 hr). The times were

selected as typical for a field installation of the monitor, consistent with the overall goals of the project. Three different counting cells were tested with three different cell configurations, each containing crushed yttrium silicate scintillator mesh sized to 35 microns. The largest cell had a void volume of 1 mL and was configured in a simple U-tube geometry using 7 mm OD x 1 mm wall tubing. The smallest cell

was a compact spiral of 5 mm OD scintillator filled tubing and with a void volume of 0.4 mL. The intermediate cell was a single loop of 6 mm tubing with a void volume of 0.5 mL. In all cases, the cells were placed between the two photomultipliers and maintained at a fixed distance relative to each other. The cells were sealed from stray light and the entire assembly placed inside a lead background shield.

The results of the testing of these cells are summarized in Table 1. The backgrounds are the minimum attainable for each cell. It was observed that the background would increase monotonically over a period of several weeks. A cell wash procedure restored the background to its minimum value. DISCUSSION

Sensitivity

The sensitivity for measuring tritium is a function of the efficiency, the background, and the counting time. The detection limits were calculated using the methodology given by Currie.9 For completeness, the formalism used to evaluate the data are summarized below. The minimum detectable concentration (pCi/cc) is defined as that concen-

426

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 2. Minimum Detectable and Quantifiable Concentrations of Tritium for Cells Tested Void MDC (2 min)a MDC (24 hr) MQC (24 hr)b (cc) (pCi/cc) (pCi/cc) (pCi/cc) 0.4 0.5 1.0

600 2300 1450

20 73 48

60 230 150

aMinimum detectable concentration for 2 mm count (See text). bMinimum quantifiable concentration for 24 hr count (see text).

tration which has 95% probability of being above a threshold that is exceeded by only 5% of the background counts. It is given by:

MDC = DL/(V*t*e*2.22) where DL (counts) is the detection limit, V (cc) is the void volume of water in the active region of the cell, t (mm) is the count time, e (c/d) is the counting efficiency in the tritium channel, and 2.22 (dpm/pCi) is the rate conversion factor. The detection limit is given by:

DL = 2.71 + 3.29 * sqrt (B) where the total background counts (B) is given by the product of the background count rate and the count time. The minimum quantifiable concentration, as defined by Currie, is that concentration that will permit the quantification at a precision of 10% one sigma relative standard deviation as computed by the following:

MQC = QL/(V*t**2.22) where QL is the quantification limit (counts) as defined as

QL = 50 * [1 + sqrt[ 1 + ( B / 25) ] ]. Using the data obtained from the cells tested in our laboratory and the above formalisms, the MDC and MQC for these cells for a short counting interval of 2 mm and a long counting time of 24 hr can be calculated. The results are shown in Table 2. In all cases, it is obvious that the detection of a 2000 pCi/cc concentration is possible in a short period of time. At a 2 mL/min

flow rate through the cells, careful engineering is required to minimize the holdup of the solution from the sampling point to the counting cell. It is also obvious that none of the cells meet the criteria for detecting or quantifying a release at the 2 to 10 pCi/cc concentration. A combination of a larger cell, with higher efficiency and lower backgrounds, is required. Figure 2 shows how the sensitivities could improve with reasonable advances in the cell and system design. A counting cell with a void volume of 3 mL a 7% counting efficiency, and a background of only 15 cpm would satisfy our requirements. The vendors

claim that such improvements are possible, but to date, none have been

DEVELOPMENT OF AQUEOUS TRITIUM EFFLUENT MONITOR

427

10,000

Observed

1,000

0.5 ml Cell 0.4% Efficiency 15 cpm Background Modeled 3.0 ml Cell 0.4% Efficiency 5 cpm Background

100

Modeled 3.0 ml Cell

10

7.0% Efficiency 15 cpm Background

1

0.1 10

100

1,000

10,000

1,000

10,000

Counting Time (minutes)

10,000 Observed

0.5 ml Cell 0.4% Efficiency

1,000

Modeled 3.0 ml Cell

0.4% Efficiency 100

Modeled 3.0 ml Cell 7.0% Efficiency

10

1

10

100

Counting Time (minutes)

Figure 2.

Minimum detectable and quantifiable activities for continuous aqueous effluent tritium monitor.

428

LIQUID SCINTILLATION COUNTING AND ORGAN!C SCINTILLATORS

Stability

The variation of the efficiency or background of the detector will have a significant effect on the overall performance of the monitor. As stated previously, the background was observed to change by a factor of two over a several week period. This may be caused by the buildup of material on the surface of the scintillant or the presence of contaminants in the light-emitting water. Bioluminescence is a known interference with scintillation counting techniques. To test the long term performance in a laboratory environment, and validate the calculated sensitivities, a 3-week-long experiment was performed using the 0.4 mL cell. First the background was established by recirculating demineralized water through the system, periodically cleaning with NaOC1 or dilute nitric acid. This was continued until a constant background was attained (12.5 ± 0.8 cpm). After adequate background stability had been reached, a small quantity of tritium was introduced into the water recirculation system, and the monitor response noted. The initial aliqout increased the concentration of tritium to about 140 pCi/cc; an aliquot was removed for laboratory analysis. The monitor count rate increased by about 20% in the tritium counting channel while the high-energy beta channel showed no change in count rate (see Figure 3 for the hourly observations). After a suitable recirculation time (at least 24 hr) another aliqout of tritium standard was spiked into the system, and the response was measured. This cycle was continued until the concentration in the recirculation system reached 1000 pCi/cc. At this point the monitor was recording a system count rate of twice the background in the tritium counting channel and no change in the gross beta-counting channel. By removing aliquots of each solution for laboratory analysis, the efficiency of the monitor at each concentration could be calculated (see Figure 4 for a plot of the efficiency as a function of concentration). It was found that over the concentration range tested, the efficiency (e = 0.0177 ± laR.S.D.) appears to be independent of tritium concentration (correlation coefficient = -0.073). At this point in the experiment, the system was flushed with repeated washings of fresh demineralized water, and the detector response returned to background suggesting no memory effects. To begin testing aqueous solutions typical of surface streams, a sample of secondary cooling water from one of the reactors was obtained for testing and analysis. The cooling water is taken from the streams in the SRS area, passed through the reactor primary heat exchanger, and then discharged into a cooling pond before it is released back into the surface streams. A composite sample of the water discharged from one of the reactors was circulated through the monitor in an attempt to determine the tritium concentration. The calculated tritium concentration, from the increase in count rate observed by the monitor, was 200 pCi/cc, while laboratory analysis of an aliqout showed a tritium concentration of only 50 pCi/cc. The sample was highly quenched. The monitor response to the reactor effluent is shown on the right hand side of Figure 3. It is noteworthy that the gross beta

Figure 3.

10 20 30 40

0

0

I

Tritium

40

I

80

I

160

I

I

200

Observation Number + GrossBeta(>20 keV)

120

Monitor response to aqueous solutions containing tritium.

0

I

100

110

120

130

140

150

I

240

0

280

100

I

300

I

500

I

Tritium Conceritretiori (pCi/cc)

I

R2 - 0.073

Figure 4. Variation of tritium counting efficiency with concentrations tested.

0

0.2 -

0.4 -

0.6 -

0.8

1.2

1.4

1.6

1.8

2

2.2

2.4

2.6

2.8

3

700

I

*

900

DEVELOPMENT OF AQUEOUS TRITIUM EFFLUENT MONITOR

431

channel count rate decreased by 35°Io during the experiment with reactor effluent. During this run, the differential pressure rose from 20 psi (its equilibrium value during the demineralized water tests) to 100 psi after about 2 days of run time. The explanation of this phenomenon can be attributed to the microorganisms in the sample. Bioluminescence is occurring in the tritium channel, resulting in an apparent increase in count rate. Meanwhile the microorganisms are creating a slime on the scintillator which reduces the light output

in the gross beta channel. Plugging of the scintillator is consistent with this speculation Care will have to be taken to precondition the solution prior to reaching the measurement cell. Sterilization of the sample stream using biocides, ultraviolet light, ultrafiltration, etc. is mandatory if solid scintillators are to be used in the monitor. Other conditioning steps that will be required include the filtration of

debris to prevent cell plugging, the removal of organics (e.g., by activated charcoal) that might produce light, and the removal of inorganic impurities (e.g., by ion exchange). These technologies are available using standard swimming pool water purification processes.

DEVELOPMENT OF ALTERNATE SENSORS

Since the plugging of tritium measurement cells containing crushed inorganic scintillators was apparently weak environmental applications, development of alternate sensors is underway at SRS. To minimize the plugging problem, the spacing between the light detecting media should be made larger for less restriction to flow. To provide a sensitive tritium monitor, a large surface area with efficient light detection is required. Borrowing on the design of the anthracene coated fiber detector7 and recent developments on the fabrication of plastic scintillator fibers, a flow cell was constructed with overall dimensions of 51-mm in diam by 51-mm in length, containing 500 fibers of 1-mm diam plastic scintillator* spaced on 2-mm centers. The ends of the fibers are cemented into a pair of light pipes which are, in turn, optically coupled to a pair of PMTs. The aqueous sample flows perpendicular to the fibers and the light is transmitted down the fibers to the PMTs. The high index of refraction of the plastic scintillator (l.58),12 relative to that of typical aqueous solutions (1.3), should limit the loss of scattered light. The plastic scintillator is nearly transparent to the light produced by the tritium interactions (bulk light attenuation length = 250 cm'2), so detectors containing long lengths of fibers could be used. The fabrication of the detector is underway and will be tested in the laboratory when the construction is complete. Pulse shaping and fast coincidence electronics will be used to discriminate single photon events and dark current

noise. The background of the cell should depend only on the amount of scintillator and the lead shielding from cosmic rays. A comparison of the *Manufactured by Bicron Inc., Newberry, OH

44065

432

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 3. ComparIson of Measured and Expected Responses for Various Types of Solid Scintillator Detectors Background Volume Efficiency MDC (10 mm) Scintillator (cpm) (cc) (c/d) (pCi/cc) Crushed Crushed Crushed

12.5 14 18

Plates (Ref. 6)

40

Anthr. (Ref. 7)

30

1mm fibers 0.5mm fibers 0.25mm fibers

42

0.1mm fibersd 0.25mm fiber&' 0.5mm fibersd 1.0mm fibersd

185a

0.4 0.5 1.0

028b

6.8

09b 065b

43a 39a 39a

0029b 0012b 00057b 00053b

0.01 77

251

0.004 0.0035

938 603

0.006

286

0.016

600

304 228 199

oo1o4c oo1o4c oo1oz1

3551 8761 17671 19005

aBased on mass of scintillator relative to mass of plates and background observed in Ref. 6. bBased on active surface area and 6 m range of 3H beta. cBased on 65% light output of anthracene. dAssumes 50% void volume in U-cell 8.5-cm long.

expected response to that of already tested systems is shown in Table 3. The basis for the comparison is the intrinsic light production of the scintillator, the

surface area of the active material, and an estimate of the light collection efficiency. The data presented in Table

3 show that the detectors using plastic scintillator fibers offer the best chance of reaching the lower detection sensitivities, even with the conservative background estimates. If modern pulse shaping and timing electronics and photomultipliers can improve the signal to background ratio, the MDC can be reduced even further. Additional shielding may also increase the sensitivity for making tritium measurements. All these enhancements will be evaluated by further laboratory and field testing.

REFERENCES

Jacobs, D.J. "Sources of Tritium and its Behavior Upon Release to the Environment," USAEC Report TID-24635, (1968). Eisenbud, Merril. Environmental Radioactivity, 3rd Edition, (New York: Academic Press, 1987). Lederer, C.M. and V.S. Shirley, Ed. Table of Isotopes, 7th Edition, (New York: John Wiley & Sons, 1978). Crook, M.A., P. Johnson, and B. Scales, Ed. Liquid Scintillation Counting, Volume 2, (New York: Heyden & Son Inc., 1972).

Budnitz, R.J. "Tritium Instrumentation for Environmental and Occupational Monitoring - A Review," Health Phys., 26:165 (1974). Osborne, R.V. "Detector for Tritium in Water," NucI. Instrum. Meth., 77:170 (1970).

DEVELOPMENT OF AQUEOUS TRITIUM EFFLUENT MONITOR

433

Moghissi, A.A., H.L. Kelly, C.R. Phillips, and J.E. Regnier. "Tritium Monitor Based on Scintillation," Nuci. Instrum. Meth., 68:159 (1969). Berthold Analytical, "Measuring Cells" Unpublished Results (1989). Currie, L.A. "Limits for Qualitative and Quantitative Detection," A nal. Chem., 40:586 (1968).

Reeve, D.R. "HPLC Radioactivity MonitorsFact and Fiction," Laboratory Practice, 3 (1983). Schram, Eric. "Bioluminescence Measurements: Fundamental Aspects, Analytical

Applications and Prospects," in Liquid Scintillation CountingRecent Developments, P.E. Stanley and B.A. Scoggins, Ed. (New York: Academic Press 1974), pp. 383-402.

Hurlbut, C.R. "Plastic Scintillators A Survey," Trans. Am. Nuci. Soc., 50:20 (1985).

CHAPTER 37

Rapid Determination of Pu Content on Filters and Smears Using Alpha Liquid Scintillation

P.G. Shaw

ABSTRACT This chapter discusses a technique for rapidly determining plutonium content on filters and smears using alpha liquid scintillation. Filter and smear samples will be analyzed daily for plutonium (239Pu) content during projected waste retrieval operations at the Radioactive Waste Management Complex (RWMC) of the Idaho National Engineering Laboratory. Daily monitoring will allow for trending of airborne and surface contamination. Present analysis techniques

are time consuming, as both numerous naturally occurring isotopes, such as uranium and thorium daughters, and inert solids must be removed prior to counting to avoid interference with Pu detection. Alpha liquid scintillation (ALS) in conjunction with microwave digestion was investigated as a technique for rapid Pu analyses. Advantages offered by ALS are short turnaround time and field use with acceptable accuracy. A state of the art Photon Electron Rejecting Alpha Liquid Scintillation (PERALS) Spectrometer with pulse shape discrimination (PSD), and an oil filled photomultiplier tube counting chamber with 99.7% counting efficiency and 99.95% rejection of beta and gamma pulses, was used. Relatively clean filter samples could be directly counted in an all purpose scintillant, bis 2-ethylhexyl phosphoric acid (HDEHP), 4-biphenyl-6-phenylbenzoxazole (PBBO), toluene, and naphthalene. Laboratory preparation of soil samples and smears with high inert solids content was accomplished by dissolution of the sample in nitric and hydrofluoric acids using a microwave digestion system in

teflon pressure vessels. The Pu in the dissolved sample was extracted into tertiary amine nitrate and counted in a HDEHP or 1-nonyldecylamine sulfate (NDAS) containing extractive scintillant. This method is applicable to the determination of total plutonium in air filters, smears, and soils. The minimum detectable activity (MDA) for direct counting of air filters is about 100 pCi/g (3.7 Bq/g) for an hour count. If the sample is dissolved and Pu extracted, activities near 1 pCi/g (0.037 BqIg) can be seen with a 20 mm count.

INTRODUCTION

Filter and smear samples will be regularly analyzed over the course of the day for 239Pu content during projected waste retrieval operations at the Radioactive Waste Management Complex (RWMC) of the Idaho National Engineer-

ing Laboratory. Daily monitoring will allow for trending of airborne and 435

436

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

surface contamination. Airborne contamination collected on filters and surface contamination collected on smears are the two normal forms of samples expected from the contamination monitoring system. Alpha liquid scintillation (ALS) techniques developed with beta and originally used the same equipment as beta scintillation counting.1-3 A specifically designed alpha liquid spectrometer (PERALS)4 with 99.7% counting efficiency, 99.95% rejection of beta and gamma pulses, 5''o (250 keV) resolution for energies in the Pu region of 5162 keV and 0.02 cpm electronic background was tested. Relatively clean filter samples were directly counted in an all purpose scintilbis 2-ethyihexyl phosphoric acid (HDEHP), 4-biphenyl-6phenylbenzoxazole (PBBO), toluene, and naphthalene. This formulation was

lant,

found to be the optimum for ALS when compared to common beta cocktails.3'4 Routine treatment for soil samples and smears with high inert solids content, however, requires laboratory dissolution and two extractions, first into a tertiary amine nitrate, and second into a HDEHP or 1 -nonyldecylamine sulfate (NDAS) containing extractive scintillant.4'5 Special sample preparation techniques, such as microwave digestion, are important in ALS. Microwave digestion quickly reduces solid samples to an extractable form6 within a nitrate system. Microwave digestion in specially designed, closed, raised-pressure, teflon vessels has been used for dissolution of a variety of materials and is currently being considered as an alternate EPA method for standard open vessel wet ashing techniques.7 With the proper organic extractants, filters can be directly counted and soil samples extracted and counted by PERALS. This should be possible in a short amount of time (about 1 hr), in the field and with acceptable accuracy and peak discrimination against natural background components in the soil.

This chapter is a scoping study to test procedures needed in preparing samples, standards, and blanks for the PERALS system and gives preliminary results of filter and soil Pu analysis. Screening samples in less than an hour was tested rather than routine analytical analysis. APPARATUS

Until recently adaptation of existing beta detectors was the only way to perform alpha liquid scintillation spectrometry)'2 However, these detector systems did not provide good spectrometric results. To obtain good results, we needed an alpha only detector with (1) improved counting chamber design with no air gap, (2) improved pulse shape discriminator (PSD) to separate alpha

from beta and gamma pulse continuum, (3) multichannel analysis (MCA) rather than the pulse height analysis (PHA) counting of energy regions, (4) elimination of most quenching, (5) lower background through the rejection of

afterpulses characteristic of cosmic radiation, and (6) improved efficiency through rejection of unwanted luminescence.

RAPID DETERMINATION OF PU CONTENT

437

McDowell4 has designed a workable ALS instrument called a PhotonElectron Rejecting Alpha Liquid Scintillation Spectrometer (PERALS). Currently an improved alpha liquid scintillation design is manufactured exclusively by Oak Ridge Detector Labs (ORDELA). The ORDELA PERALS Model 8200B detector uses pulse shape discrimination (PSD) and an oil filled photomultiplier tube counting chamber giving: 99.7% counting efficiency,

99.95% rejection of beta and gamma pulses, 5% (200 key) resolution at energies in the Pu region (5162 key), 0.02 cpm background. A PERALS spectrum of 226R and its daughters is overlaid on a typical one using silicon surface barrier alpha spectrometry (Figure 1). This comparison shows the radium, radon and polonium peaks resolved in both systems. The 250 keY for Pu resolution of PERALS, noted in the first peak when compared to about 20 keV in the surface barrier spectrum, does not allow resolution of the two 226R peaks at 4.6 and 4.78; thus 241Am, always found with 239Pu, are also seen as one peak on the PERALS system. Analysis equipment includes: the CEM 81D microwave digestion unit, assorted sizes of separatory funnels (30 to 500 mL), adjustable hot plates, heat lamps, a balance capable of weighing to 0.1 g, a dry argon sparging apparatus,

(Figure 1), 10 x 75 mm culture test-tubes, cork stoppers, parafilm, and lambda pipettes. The ALS spectrometer was powered by a Canberra Model 3120 High Voltage Power Supply. Data is transferred through a Canberra Model 18075 A to D Converter to a Canberra MCA system 100 operating an IBM PC, for peak detection and analysis. Four types of reagents are used in the procedure: (1) mineral acids for sample dissolution and organic stripping, (2) inorganic salts for oxidation state adjustment, (3) large organic amines or phosphates for sample extraction, and (4) scintillation grade organic reagents and fluors for cocktail preparations. High purity reagents decrease the probability of various unwanted reactions, minimize introduction of undesirable quenching species, and reduce background. PROCEDURE

This section describes the ALS analytical scheme. The results give a preliminary performance evaluation of the ALS system, microwave digestion, and

organic extraction as an analytical tool for both directly analyzed and digested/extracted filter and soil samples.

Certain air filters and smears may be counted directly in the extractive scintillator. Swipe or smear samples and lightly coated air filters must be relatively clean, contain mostly alpha activity, have low beta-gamma activity and have low inert solid content. If the sample contains a large amount of soil or matter, it will require digestion and extraction. Sample dissolution was accomplished by wet ashing in nitric and hydrochloric acids using hydrofluoric acid to break down silicates. A microwave

5

Energy (MeV)

6

Figure 1. Surface barrier and PERALS 226Ra spectrum.

4

7

8

RAPID DETERMINATION OF PU CONTENT

439

digestion system with teflon pressure vessels contained the HF and reduced dissolution time. The sample (soil, smear, or filter) is weighed, placed in a 120 mL high temperature teflon vessel, and a 7:3 mixture of aqua regia and hydrofluoric acid is added. The cap is tightened on the vessel to a prescribed torque. This allows the acid mixture to become pressurized to 120 psi. At this pressure, the temperature of the solution reaches 150°C and the HF remains in solution longer than in an open system. The microwave is operated for 5 mm at full power (650 watts) then 10 mm at 50 power. Vessels are uncapped and 5 mL of nitric and 5 mL of 30 hydrogen peroxide are added to drive off chloride and fluoride ions and oxidize minute traces of organic matter. The sample is placed in a 30 mL beaker, and 3 mL aluminum nitrate is added to tie up free fluoride, prevent calcium fluoride precipitation, and adjust the ionic strength. Volume is reduced to 5 mL by evaporation to remove high acidity, chloride, and fluoride ions, giving a pure nitrate system. The solution should be about 1 M nitric acid and 3 M total nitrate. The dissolved sample is converted to a suitable oxidation state for extraction into an an aqueous immiscible scintillator. Pu is primarily in the + 6 state after nitric acid digestion. It is brought to + 4 by reduction with ferrous sulfate or potassium metabisulfite (if the system contains appreciable iron). Any Pu +3 present is raised to + 4 with sodium nitrite. After addition of these reagents the solution must be contacted immediately with the tertiary amine or Pu will disproportionate back into multiple oxidation states. After ionic strength and oxidation state adjustments, Pu is extracted into the high molecular weight (>300) tertiary amine such as tri-octyl amine or, as used here, the proprietary formulation, Adogen-364. The amine is nitrated by con-

tact with 0.7 M nitric acid before extraction. The purity of this amine is critical. Any primary or secondary amines present may keep the Pu from being extracted, bind it to the amine so it can not be stripped, or extract unwanted ions. Due to time constraints the purity of the amine could not be assayed and purifying procedures could not be undertaken. The main concern in the extraction step is the removal of Th, U, and any colorant such as iron. Removal of other elements in the extraction procedure is not as crucial in this method as the sample will not be plated out on a surface as a solid.

The acid solution is transferred to a 30 mL separatory funnel and shaken with the amine for several minutes by an automatic shaker. The Pu is stripped from the amine nitrate, which is highly quenched, and put into an organic with less quenching. It is stripped with a solution of 1 N sulfuric or perchloric acids and a small amount of an associated salt of each acid, lithium perchlorate or sodium sulfate. Salt provides a surface for Pu to adhere to upon evaporation and prevents Pu from plating on the container sidewall. Acidity and volume of the stripped solution is reduced by heating. This also destroys any residual amine. The type of acid determines the final extractive scintillant. An extractive scintillant containing HDEHP was used for perchlonc acid, NDAS for sulfuric. A dilutent, 2-ethylhexanol, was to the triocty-

440

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

lamine nitrate (TANO3) as an aid to stripping into 1 N sulfuric acid. The pH of the perchioric solution for HDEHP extraction should be greater than one, as

the extractant is a weak acid. The pH for the sulfate system can be between zero and two. K2S208 is added to ensure the Pu + 4 for the perchloric system. Other than the extractants HDEHP or NDAS, the scintillant cocktails are formulated the same for either system with solvent, enhancer, and fluor. A matrix standard soil was specially prepared to a 100 pCi/g concentration of Pu homogeneously distributed, and chemically and physically bound to the soil. Soil for the standard was obtained from the RWMC at an 8 ft depth near the location of the proposed retrieval effort. Soil was dried at 105°C in a laboratory oven and sieved through a 35 mesh screen (525 m) followed by a 200 mesh screen (75 /Lm). Both fractions were weighed with more screened soil

added until a total of 1 kg of screened soil was obtained. The 1 kg of soil was then spiked by taking approximately 100 g from each fraction, wetting completely to a slurry consistency, and adding an accurately weighed aliquot of 239Pu stock solution. These slurries were mixed thoroughly,

dried, ground, resieved, and then mixed in a special dual cylinder mixer. Enough 239Pu spike was added to each fraction so that the activity concentration was approximately 100 pCi/g. The final blended product had an activity

concentration of 102.3 pCi/g. Sieving of the soil was done to enhance the particle size fraction that was less than 10 m. This is the respirable fraction and should be the same size as found on most of the smears and filters to be analyzed.

An above background sample of Pu contaminated soil that had been environmentally aged and may have been "high fired" was obtained from the same

source as most of the waste, Rocky Flats Plant, and used as another matrix standard. This soil has a high silica and low clay content, a more refractory (hard to dissolve) Pu oxide, and Pu that is more intimately bound to the soil. The RFP soil was obtained by RFP personnel, downwind from a former drum storage area. The Pu originated 20 years ago from leaking drums that con-

tained contaminated cutting oil. The oil held a suspension of <3 m Pu particles. This area was decontaminated in 1969 and covered with an asphalt pad.9 The estimated 239Pu concentration was 1000 pCi/g. The RFP soil was dried and sieved to determine particle size distribution. The finest particle size range, less than 45 jm was used for most of the tests as this approximates air deposition or filter samples more closely than the bulk soil. A similar sized sample from the RWMC standard was also used for this reason. Both matrix standards, RWMC and RFP soil, were used to test the sample dissolution and extraction efficiencies and verify elimination of background interferences. Matrix blanks were used to test the method, particularly the contribution of natural background aiphas and reagent impurities to the peak. For a blank soil known to be Pu free, a subsurface soil sample from a basement excavation, presumably having no Pu fallout, was used. This soil was counted directly or was dissolved, extracted, and counted. Differences between the matrix stan-

RAPID DETERMINATION OF PU CONTENT

441

dards and the matrix blank were used to determine the minimum detectable activity, as well as how the ALS system and chemistry is distinguishing Pu alpha from those alpha naturally in the soil. The Pu and Am concentration of the standard and blank soil was determined by the Rocky Flats Plant, Golden, CO, UNC Geotech at Grand Junction, CO; and the INEL. They used dissolution and extraction, followed by conventional alpha spectrometry and whole sample counting of Am by gamma ray spectrometry. RFP and INEL used gamma ray spectrometry of 241Am and assumed a 10:1 ratio of Am to Pu. UNC used total dissolution, organic extraction, ion exchange, precipitation, and alpha spec for Pu and Am. The results averaged 1010 pCi/g Pu and 110 pCi/g Am for the suspendable (less than 100 m) portion of the soil. The blank had less than the detectable (about 0.1 pCi/ g) for both isoptopes. Following sample preparation, samples were purged of dissolved oxygen and dried before insertion in the oil filled counting chamber. In direct filter analysis, occluded air on the filter was also removed. A filter disc, smear, or portions there are folded and pushed with a lambda pipette into the counting test tube so it takes up no more space than the one mL volume scintillation

solution. Air is removed by bubbling dry argon, saturated with toluene, through a lambda pipette while probing with the pipette to remove all bubbles from the test-tube. The paper becomes transparent and seems to disappear. A transfer lambda pipette tip was used as a sparging lance. Water in the gas is removed with molecular sieve and metallic sodium. Bubbling through toluene saturates the gas with the scintillant solvent so the sample volume remains constant. The test tube is corked and sealed with parafilm, wiped clean, and placed in the sample holder oil bath. The light tight cap is replaced, the high voltage activated, and counting on the ALS spectrometer initiated. A directly prepared

sample will count with near 100% efficiency if the alpha activity is on the surface of the fibers in a very thin layer. This method works best on samples with ultrafine particulates, such as air filter samples with only respirable size particulates. Sparged samples sealed with parafilm may last for several weeks, but slow evaporation of the scintillant still occurs. Only glass sealing and dark storage will ensure long term stability. Two types of calibration are necessary in the ALS system: energy and amplitude. Energy calibration is done by extracting Pu from the standard aqueous solution into the same scintillant in the same manner as the sample. The peak

energy of the sample should match that of the standard. Instrument and extraction efficiency are verified with check standards and spikes, assuming near 100% counting efficiency. When directly counting samples the calibration for both energy and amplitude is difficult, as color and chemical quenching shifts the spectrum, and inclusion of alphas in some particles lowers the efficiency. Results for direct counting of various types of blanks, prepared calibration standards, and standard spiked soil are given below.

442

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

RESULTS

This section discusses results of analysis of soils and filters on the PERALS system, both directly without treatment and following microwave digestion and extraction of the sample. Three experimental parameters will be reviewed as they apply to both direct analysis and analysis after dissolution and extraction: efficiency, resolution, and background. Optimum is 99.7% efficiency, 5 10 resolution, and <0.02 cpm background for the entire process (dissolution,

extraction, and counting). Precision for these parameters, for direct and extracted soil standards, blanks, soil blanks, and soil spikes, for problems encountered, and for further work necessary are discussed. Background includes electronic noise, cosmic rays, and chemical impurity contributions to the final count rate. Efficiency is the percentage of the analyte extracted and pulses successfully detected by the instrument and converted to the count rate. Resolution is the degree of energy separation of one group of pulses from another and the location on the energy scale of a specific peak. Precision is the stability of the instrument and reproducibility of the other parameters. Resolution

Energy resolution, the separation of two alpha peaks by energy and the energy location on the spectrum, has both instrument and chemistry contributions. Resolution depends on the amount of light per pulse received; thus an efficient and stable scintillator and diffuse reflector are necessary. The characteristics of a PM tube and its physical relationship to a sample are critical to maximizing this parameter. The detector uses an oil-filled cavity, eliminating the air gap between sample and phototube. This prevents spectrum distortion caused by refractive index discontinuity and improves resolution. Figure 1 shows a PERALS alpha spectrum for 226R and its daughters overlaid with a surface barrier spectrum.9 The radium, radon, and polonium peaks

are resolved in both systems. The weak 4.6 MeV radium, however, is not resolved from the strong 4.78 MeV radium, illustrating the the practical limitations 5% resolution. The resolution of the major peak is about 20 keV in the surface barrier spectrum and 250 keV in the PERALS spectrum. Resolution is needed to separate background nuclide activities from those activities of Pu in the soil. Energy stability is important in identifying the peak. The separation of the Pu peak from that of background Th depends on energy stability and resolution. A spectrum of a naturally occurring alpha emitter, Th, is shown with that of Pu in Figure 2 after one week of ingrowth. The separa-

tion of the single Pu peak (5.1 MeV) from the major Th (4.2 MeV) and daughter peaks (6.0 and 7.7 MeV) can be seen. Table 1 gives resolution for prepared Pu standards. The full width at half maximum (FWHM) and the region of interest (ROL) width were both used as

measurements of resolution. Most peaks had resolutions under 10%. We

50

4.2 MeV Th-232

Plutonium and thorium standard spectrum.

0

50

100

Figure 2.

Ci

0

C

U)

150

200

150

200

Th-daughters

Channel Number

100

Pu239

5.1 MeV

256

444

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 1. Plutonium Peak Resolution on Prepared Standards Standard FWHM PCI Key

Resolution Percent

60

290 502 546 490

74.4 110.2 142.9

557 638 585

5.2 9.7 10.6 9.5 10.8 12.4 11.3

Average Standard Deviation

515

10

0.1

0.7 5.3

111

2.2

achieved a 5.6% resolution on a low-level sample (0.1 pCi) in the the HDEHP extractive scintillant. This is the scintillant used for Pu extraction in the per-

chloric system and direct filter counting. Peak location is also a factor in energy calibration and peak identification. The standard deviation (1 sigma) of peak energies (location) was less than 10% that of resolution was 2.2%. The effect of the soil on resolution was tested by using direct soil laboratory

spikes. Two soils were spiked with a standard Pu solution directly in the counting tube. The Pu peak was shifted by about 10% from that of the standard without soil and broadened by 30 to 50%. A slight increase in the beta continuum region channel 5-10 was also noted. Spike recovery was 90%. Energy shifts and loss of resolution occur when adding soil to the scintillant in direct analysis. Resolution also changes with various soil types - blank soil, spiked RWMC soil, and RFP soil. The direct soil sample method often gives a highly colored sample, which gives an energy shift toward the beta continuum

region, channel 1-10. This type of quenching may also decrease counting efficiency with the loss of resolution. Energy shifting and quenching is discussed in the background and precision sections.

Efficiency Total efficiency is the sum of counting efficiency (the percentage of pulses successfully detected by the instrument) and the efficiency of dissolution and extraction, sometimes called chemical yield. Several types of samples were analyzed to measure these operational parameters and thus assess the quality of data achievable with the ALS system (instrument and chemistry). Efficiency

of various samples and control standards in both the direct and extracted mode are given in the tables and discussion that follows. Problems and interferences that affect efficiency are given in the discussion of precision. Table 2 gives combined counting and extraction efficiency for various soils both directly counted and digested-extracted before counting. At the current efficiency of about 20 to 25% in the direct mode, 100 pCi/g of Pu on RFP soil should be detectable on relatively clean filter samples. At this concentration (one tenth the RFP soil concentration 1000 pCi/g) and efficiency, the count

RAPID DETERMINATION OF PU CONTENT

445

Table 2. ALS PERALS Analysis of RWMC Standard Soil, Rocky Flats Soil, Rocky Flats Soil

on Filters Peak

Soil Type 100 pCi/g Standard 1000 pCi/g RFP Soil AFP Soil on Filter Standard RFP Filter

Method

Centroid

Direct Direct Direct Extraction Extraction

94 12 19

72 85

Percent Efficiency

Percent Resolution

45 8.6 28 85 22

14 12 28 9 8

rate of 50mg of sample is 2.2 dpm, or over ten times the background of 0.15 dpin. The direct filter analysis had a higher efficiency than direct RFP soil analysis, 9 vs 28%. The suspension of the soil in the PMT viewing area and the limited settling of contents could account for this higher efficiency. The filter samples here were highly quenched, as evidenced by the peak shifting seen in Figures 3 and 4. Some of the RFP Pu was seen in the higher channel regions, with less quenching and count rates about 3 times the soil background. Count rates for RFP and blank soil (Tables 2, 4, and 5) indicate the Pu could be distinguished from the background activities present in non-Pu-containing blank soil in the direct analysis mode. When the spiked soil standard was counted directly, the overall efficiency approached 45% (Table 2). The efficiency for the RFP soil with a much more refractory (hard to dissolve) form of Pu was only 9%. The count rate of 0.93 cpm (Table 5) for the spiked soil in the Pu region of interest (channels 90 to 110) is over 5 times (0.16 cpm) that of the blank soil. HDEHP is the most effective extractive scintillant for direct counting in actually leaching some of the Pu. The NDAS does not leach the Pu as well as the HDEHP and has a lower blank and standard spiked soil count rate. RFP and standard soil samples that were digested and multiply extracted have the same type of relationship between spiked soil and RFP soil held in direct counting. Extraction of Pu from the RFP soil is more difficult than the RWMC standard soil. Efficiencies for the RWMC standard soil approached 85% and the RFP soil 22%. Some RFP soil samples were digested and extracted directly into the extractive scintillator, HDEHP. Recovery efficiencies of 25% were achieved for single extractions, 22% for multiples giving a detectable peak in 1 hr of counting for 1 g of a 1 pCi/g sample. This is about 0.55 cpm, about 5 times greater than the background of 0.1 cpm (Table I). The peak location on the average is somewhat lower for the standard soil extract than the RFP soil. The RFP soil had a higher counting uncertainty and gave a wider peak (Figure 5). Time did not permit full development in this area, but further work should bring this extraction process up to 99% efficiency. Table 3 lists overall efficiencies for prepared standards. The extraction and counting efficiency of 95% approach the optimum 99.7% (counting wall

U)

0

50

Figure 3.

C-)

0

C

0

100

I

I

50

Direct Pu soil spectrum.

0

118 I

150

IChannel Number

100

I

200

-- .AI

.

256

100

200

300

Figure 4.

0

0

U)

400

500

Direct RFP soil on filter spectrum.

25

75

Channel Number

50

100

125

448

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 3. Plutonium Efficiency on Prepared Standards Standard PCI

Efficiency Percent

0.1

98

5.3 74.0 88.0 100.9 142.9 177.0

92 93 95 89 101

95

Average Standard Deviation

95 3.9

losses only). The main problem with efficiency was the extraction rather than the counting. Some standards were not in either the proper oxidation state or pH range for efficient extraction. The reproducibility of extraction was about

4o. Improvement in extraction efficiencies and consistent recoveries is desirable.

Background

There are two types of background radiations of concern in ALS, those from the beta and gamma emissions caused by naturally occurring substances in the soil such as 40K, and those from other alpha emitters in the same region as Pu, such as the naturally occurring radionuclide daughters of the uranium and thorium chains Total background is all counts in the region of interest (ROT) not from the desired element (Pu). This includes contributions from the reagents, scintillants, and other nuclides in the sample, including that of the instrument. Elimination of beta and gamma background is achieved by pulse

shape discrimination (PSD). Decreasing alpha background requires clean Table 4. Summary of Background Total

Background Type

Energy Shift Percent

Extractive Scintillant

DPM

AOl

in ROl

Centroid

Standard Soil Method

HDEHP HDEHP HDEHP

1.69 ± 0.90 1.34 ± 0.74 1.58 ± 0.6

9.4 8.7 5.7

39

Standard Soil Method

NDAS NDAS NDAS

0.92 ± 0.90 1.16 ± 0.71 0.71 ± 0.12

4.9 4.7 5.8

21

Standard Soil Method

NDEHP HDEHP NDEHP

0.08 ± 0.04 0.15 ± 0.02 0.29 ± 0.21

105 92

12

81

23

Standard Soil Method

NDAS NDAS NDAS

0.20 ± 0.07 0.04 ± 0.01 0.09 ± 0.07

75 58 55

24 28

7

4

449

RAPID DETERMINATION OF PU CONTENT

Table 5. ALS Duplicate Analysis DPM

Method Standard blank

in ROl

0.14 ± 0.02 0.15 ± 0.02

Relative Percent Deviation

Primary Peak Channel 108 109

12

62 77

22

7

0.16 ± 0.10 0.18 ± 0.18

RFP soil direct

2.10 ± 0.38 1.14 ± 0.21

39

13 14

RFP filter direct

4.30 ± 0.30 3.63 ± 0.33

17

9 17

Standard soil direct

0.97 ± 0.12 0.73 ± 0.04

28

102

Standard extraction

192.6 ± 9.4 193.8 ± 12.8 83 ± 2.1 88 ± 1.1

Standard soil extraction

28 ± 0.84 39 ± 0.4

Location

7

Soil blank direct

RFP soil extraction

RPD

Peak

0.9

61

98

0.6

103 92

6

72 98

31

82 83

4 11

15

0.6

reagents and a low radon working area. Currently the method blanks give a background 4 to 10 times higher than the optimum electronic background. Various blanks (method, standard) were prepared for both extracted and directly analyzed samples. The background contributions can also indicate the efficiency of the scintillant and show energy shifting. The method blank is blank soil directly prepared in scintillant or digested and extracted into the scintillant. The standard blank is pure extractive scintillant. Table 4 lists the average count rates for backgrounds in two different extrac-

tive scintillants. A typical background spectrum for the HDEHP extractive scintillator is shown in Figure 6. Peak location and width for two different regions of interest (ROT) are given, one region is the area of highest background near the beta continuum, containing possible thorium peaks, the other is further down field where Pu peaks should be located. The primary channel is the center of the peak where the largest number of counts are clustered. The peak width is given by the number of channels in the region of interest. The counts per channel gives the actual background that can be used for correction of sample peaks falling in those ROTs. Background varies more at different regions of interest than with different types of blanks. This seems to indicate that the background contribution of natural alpha emitters in soil is negligible. Near the edge of the beta continuum (peak channel 5-9) the background is about 10 times higher, 0.9 to 1.7 dpm, than in the region of interest for Pu in a clean sample, 0.04 to 0.2 dpm. The HDEHP, which is the scintillant of choice for direct soil counting, has a higher background than the NDAS. The NDAS is perhaps easier to use for the final

Figure 5.

Extracted RFP soil spectra.

50

150

Channel Number

100

200

256

Figure 6.

Background spectrum.

0

50

150

Channel Number

100

200

452

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

secondary extraction than the HDEHP but does not partially extract Pu in the direct soil mode. Soil added to to the scintillant does not increase the background significantly but does shift the ROl. From the background count rate of blank soil (Table 5, 0.16 dpm) the approximate detection limit of 100 pCi/g can be estimated. Assuming a detection limit of 5 (4.66 sigma) times background, the activity of the sample would have to be about 0.8 dpm. At 20¼ efficiency this is 4 dpm in the sample, or

about 0.02 g of sample. The count rates for about 10 mg of the 100 pCi/g standard, 1000 pCi/g RFP soil, and soil on filter are also at least 5 times the background of the blank soil. Precision

The precision associated with background, resolution, and efficiency for direct and extracted soil analysis can be seen in Tables 1, 3, and 4 with a summation in Table 5. Standard deviations (1 sigma) in Tables 1 and 3 give an idea of the stability of the instrument and standards and the reproducibility of the final extraction procedure. The combined instrument and sample stability, as expressed by reproducible counts of the same standard or blank over 2 weeks time, is about 1.4%. Some degradation of standard was noted after 2 weeks. The extraction procedure is more of a factor in efficiency reproducibility than the instrument instability. Lack of temperature control in the basement lab must also be considered. The

standard deviation of efficiency for multiple counts of different standards, covering a wide range of concentrations (0.1 to 200 pCi) on different days, was

about 4%. The peak location or energy stability varied by about 10¼ for the extraction and counting of standards that varied by over the three orders of magnitude in concentration. The percent standard deviation (1 sigma) in resolution, as expressed by FWHM, was 2.2%. The energy stability for a well sealed standard over time was about 2%. Variations in extraction procedure are more of a factor in any energy shifts than the instrument instability. Energy shifts between soil types for both extracted and direct analysis can be seen in comparing ROT reproducibility in Table 5 and the spectra of the standard soil and the RFP soil on a filter Figures

3 and 4. The location of the peak for the spiked RWMC soil is not shifted noticeably from that of the standard, but there is a shift when comparing the direct RFP soil on filter sample. The primary peak activity for the RFP soil on filter was located in the 10-20 channel ROI, whereas the standard and standard soil peak is around channel 100. Background stability can be seen when comparing the ROL of either extractant in the standard blank (scintillant only) to soil blank (soil and scintillant) in Tables 4 and 5. In HDEHP the shift is 7% from 9.4 to 8.7 in the low channels near the continuum and 12% from 105 to 92 in the Pu ROL (5100 keV). For

RAPID DETERMINATION OF PU CONTENT

453

NDAS the shift is 4% in the higher background beta continuum region and 24% in the Pu ROl. Considerable shifting of the peak is apparent when comparing direct soil and extracted soil analysis (Table 3). Rocky Flats soil and filters caused a greater energy shift than the directly prepared RWMC soil standard. The Pu on the standard soil is in a more extractable form than that of the aged RFP soil; thus some Pu was detected in true solution rather than from a soil particle. The clarity of the sample in direct analysis was critical and made a great difference in the peak location and the overall counting efficiency. After sparging, some samples settled more rapidly than others. Their clearing up gave rise to much of the variation between replicate counts. A general idea of background stability can be seen in Tables 4 and 5 with the

use of replicate background counts and uncertainties within a single count. Background can be influenced by instrument stability and chemical stability, and it is a factor in analysis precision. The uncertainties for the replicate blanks are the standard deviations (1 sigma) of multiple counts. The background spectrum lacked well defined peaks (Figure 6). The width of the ROIs for most of the background counts, the low count rates, and the energy shifting do not allow identification of specific contaminant isotope contributions to background activity. The counting uncertainties are higher than when activity is actually present (Table 5). High standard deviation of replicate runs and high blank count rates could have been from the poor location of the spectrometer in a known high-radon-background basement lab. Digestion and a single extraction into the final scintillator gives most of the advantages not found in a directly counted sample, such as improved resolution and efficiency and background elimination, but it saves time by eliminating subsequent stripping and extraction steps. This was tried with several dissolution systems and extractive scintillators with some success. The primary problem was incomplete extraction and extraction of iron, making the solution highly colored. Getting the aqueous sample into the proper state, by adjusting pH and ionic strength without precipitating out calcium, was also a problem without the preliminary extraction. The time saved is significant and for some applications may be feasible, especially if some other organic extractant could be found. The ideal extractant needs to be selective for removing Pu from a

nitrate or sulphate system, rejecting thorium, uranium, and iron, and not containing chloride, nitrate, or other quenchant groups. The tertiary amine nitrate now used is selective but is itself highly quenched due to the nitrate group. Multiple extractions should give lower backgrounds, less energy shifting, and increased resolution. Possible interferences while extracting the aqueous solution from the digested sample into the organic amine or extractive scintillator, in order of importance are: (1) incomplete extraction of Pu, (2) extraction of unwanted ions, and (3) inability to strip the Pu from the first extractant

into the aqueous solution. Some of the causes of these interferences are: incorrect Pu oxidation state, incorrect pH, insufficient ionic strength, impure

454

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

extractants, and nonextractable Pu complexes in the aqueous solution. The spectral separation of Pu from an interfering alpha emitter such as thorium is shown in Figure 2. Thorium should not be a problem in direct analysis unless present in great amount and allowed to ingrow, as some thorium daughters will add to the Pu peak. Multiple extractions chemically remove thorium and uranium and eliminate alpha interference problems. CONCLUSIONS

An ALS system called PERALS has been investigated for use in obtaining alpha radiation levels during the TRU waste retrieval. By directly analyzing lightly soiled filters, the ALS system can detect concentrations as low as 100 pCi/g Pu within 1 hr of receiving the sample. For heavily soiled filters or soil samples, a minimum detectable activity of 1 pCi/g can be obtained in 2 hr including the time to dissolve and extract the sample. Direct counting of filters gave efficiencies of about 2Oo. Alpha resolution of about was sufficient to separate the Pu peak from the primary Th peaks. Areas of future work include: improving extraction efficiency and reproducibility, determining detection limits in the presence of uranium and varying amounts of thorium, and decreasing the preparation time and complexities. Instrumental improvements such as better photomultiplier tubes, better light couplings, and better electronics are possible, though interferences in the

instrumentation and counting area have been reduced to a considerable extent. The largest gains can be made by improving the overall efficiency for the entire analytical scheme. Future work would involve spiking at various points

during dissolution and extracting to pinpoint critical procedural parameters where losses are occurring. Noninstrumental interferences can hinder the separation of alpha and beta pulses and the resolution of alpha peaks. Chemistry improvements in the selection of fluors and scintillants, and elimination of quenching and interferences through organic extraction, are currently under investigation. The inherent advantages of near 100 counting efficiency make ALS a viable method for rapid Pu analysis and screening in conjunction with traditional analysis methods. REFERENCES

Horrocks, D.L. Applications of Liquid Scintillation Counting, (New York: Academic Press, 1974), pp. 28-80. Birks, J.B. The Theory and Practice of Scintillation Counting, (New York: MacMillan, 1964).

Polack, H. et al. Advances in Scintillation Counting, (Edmonton: University of Alberta Press, 1983). pp. 508. McDowell, W.J. Alpha Counting and Spectrometry using Liquid Scintillation Methods, Nuclear Science Series, ORNL, NAS-NS-31 16 (1986).

RAPID DETERMINATION OF PU CONTENT

455

McDowell, W.J. and G.N. Case. "Methods for Radium, Uranium and Plutonium by Alpha Liquid Scintillation," ORNL TM 8531 (1982). Matthes, S.A., R.F. Farrel, and A.J. Mackie. "A Microwave System for the Acid Dissolution of Metal and Mineral Samples," NTIS PB83-225391, U. S. Bureau of Mines Albany OR. (1983).

Joshi, B.M., L.C. Butler, and G. Lebanc. "Application of Microwave Oven to Digest Environmental Samples for Inorganic Analysis," Preliminary Investigation, Lockheed-EMSCOM, U. S. EPA, Las Vegas NV, (1988). Langer, G. "Wind Resuspension of Trace Amounts of Plutonium Particles from Soil in a Semi-Arid Climate," l International Aerosol Conference, Minneapolis, Minn. (1984) pp. 484-487. Chanda, R.N.and R.A. Deal. Catalogue of Semiconductor Alpha-Particle Spectra, Idaho Nuclear Co., USAEC, IN-l261 (1970).

CHAPTER 38

Vagaries of Wipe Testing Data

Jill Eveloff, Howard Tisdale, and Ara Tahmassian

At the University of California at San Francisco, the common radioisotopes and 32P constitute over 8Oo of all radioisotope usage. An effective technique for determining contamination by these isotopes is wipe testing and liquid scintillation counting (LSC). The State of California requires all radioactive materials users at UCSF to maintain wipe test data of their laboratory to demonstrate contamination control. The Radiation Safety Department of the Office of Environmental Health and Safety at UCSF provides a special service to its research staff and radioactive materials users. This service performs wipe tests of their laboratories in accordance to their required monitoring frequency. This includes the analysis of the wipes taken. Because of the large volume of wipe test analyses performed by our office, it is important to understand some of the vagaries associated with wipe testing data. This knowledge helps our office provide better service to the university community. The vagaries investigated include the following: 3H, 14C,

effects on wipe test data of wiping two different surface materials effects on wipe test data using three different wipe media effects on wipe test data using varying amounts of cocktail effects on wipe test data of self-absorption in the wipe media frequent complications to tritium monitoring from fluorescence

INVESTIGATIONS OF SURFACE MATERIALS AND WIPE MEDIA and 32P was The efficiency of detecting low level contamination of 3H, tested using two lab bench surfaces, transite and wood. These materials were chosen because they are most common in laboratories at UCSF. Wipes of each isotope from each surface were made using three different wipe media; a twoinch diameter dry wipe, a two-inch diameter wet wipe, and a 0.25 x 4.0 inch piece of scotch tape. These wipes were conducted using controlled activities of 4000 cpm, per isotope, per wipe, and on controlled 3 x 3 inch surface areas. cpm removed by wipe, are shown in Table 1. The results, 457

458

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 1. % CPM Removed by Wipe

Transite Surface

Wood Surface

32p

32P

Dry wipe Wet wipe Scotch tape

0.7 0.2

2.1

4,5

9.3

18.7

1.7

1.7

5.1

0.0 2.8 4.3

2.2 37.3 11.7

5.6 33.9 15.8

INVESTIGATIONS OF VARYNG COCKTAIL VOLUME AND SELF ABSORPTION-EFFECTS

The efficiency of detecting low level contamination of 3H, "S, and 32P was tested using different volumes of scintillation cocktail: 10 mL, 5 mL, and 2.5

mL. In addition, the of self-absorption effects of the isotope by the wipe media was investigated. Controlled amounts of isotope activities were placed directly on two-inch diam dry wipes and inserted in vials with specific volumes of scintillation cocktail. The results, 'o cpm recovery, are shown in Table 2. FLUORESCENT COMPLICATIONS IN TRITIUM ANALYSIS

The detection of 3H contamination was found to be complicated frequently

by the interference of fluorescent compounds. With the wide variety of research done on the UCSF campus, there exists the probability of encountering chemical compounds with fluorescent components during typical wipe testing procedures. Fluorescence may frequently produce significant cpm in the 3H counting channel (up to 18.6 key). It is sometimes impossible to know the difference between fluorescence and 3H contamination unless spectrum analysis is performed. Most present day LSCs have spectral analysis capabilities; however, some have associated computer software which will generate graphic representations of the spectral analysis. These representations can show, most obviously, the differences between fluorescence and 3H contamination. A typical 3H spectrum is shown in Figure 1 and a typical fluorescence spectrum in Figure 2; however, fluorescence spectra, are not restricted to the type seen in Figure 2. The differences between the two spectra can be seen with close observation. Table 2. % Recovery of CPM ISOTOPE 3H

32p

ACTIVITY (CPM)

COCKTAIL VOLUME

10 mL

5mL

2.5mL

456 4748

20.6

18.2

16.2

19.7

16.1

13.1

766 6525

56.8 62.5

63.7 63.0

65.9 63.6

519 5098

94.4 99.5

91.3 100.0

91.0 100.0

VAGARIES OF WIPE TESTING DATA 12

459

0,50

Time:

XeV Full Scale? LL

1JL

Region A Region B

2.0

18.6

18,6

167,

670.0

0.74 10.92

Region C

16?.

2000

4.0

1414

CPN 143086.

2S'/

3900 C 0 II

N I S

Figure 1.

Typical 3H spectrum.

The differences include both the distribution of energies (i.e., up to Emax) and the average energy; therefore, it is always prudent to take one step further in 3H contamination analysis by evaluating whether the source of counts could be attributed to fluorescence.

9

TiNe:

0.50

eV Full Scale? Region A

Region B Region C

LL 5.0

UL 18.6

CPN 4346.0

0.0

0.0 0.0

0,0 0.0

0.8

2S/ 4,29

0.00 0.80

XEV Figure 2. Typical flourescence spectrum.

I

15

460

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

RESULTS

A summary of the results of these experiments is as follows: Contamination can be more easily removed by wipe from transite as opposed to wood surfaces. Low level 3H was difficult to detect under all conditions studied. Wet wipes provide the more efficient contamination removal. Scotch tape also increased the efficiency of contamination detection.

Dry wipes produce up to 80% self absorption of 3H samples, 40% selfabsorption of 35S samples, and very little self absorption of 32P recovery of cpm was seen in reduction of scintillaVery slight decreases in

tion cocktail used per sample (The implications here for cost savings are incredible.)

The results of this somewhat shallow study produce thought trends that can be used in wipe test data manipulations and/or analysis. All samples used in this study were counted in a Packard LSC model 2000CA.

CHAPTER 39

The Determination of 234Th in Water Column

Studies by Liquid Scintillation Counting

R. Anderson, G.T. Cook, A.B. Mackenzie, and D.D. Harkness

INTRODUCTION

As a result of the pronounced differences in their geochemical properties, uranium and thorium exhibit markedly different solubilities in seawater under oxidizing conditions. Uranium is relatively soluble in seawater, with an average open ocean 238U concentration of 0.04 Bq 1-' (3.2 pg-') and it exists as an anionic carbonate complex of the uranyl ion. In contrast, 232Th is highly insol-

uble in seawater, with a concentration of about 2 x l0 Bq 1' (6 x l0 jg 11) and it is highly susceptible to removal from solution by hydrolysis or by sorption on, and incorporation in, particulate material. Similarly, thorium isotopes produced in situ in seawater by decay of soluble parent radionuclides also have a very short residence time in solution. The resulting situations of

radioactive disequilibrium which develop in the marine environment, eg., between 230Th and 234U, (members of the 238U decay series shown in Figure 1),

have long been recognized as a means of investigating the rates and mechanisms of a variety of oceanographic processes. For example, the systematics of 230Th/234u disequilibria have been used in the development of oceanic sediment chronologies,'2 investigation of uranium diagenetic chemistry and mobility in sediments,3'4 and study of the efficiency of particulate scavenging of thorium

and protactinium in different areas of the oceans.57 In recent years, however, it has become apparent that the short-half-life 234Th (Figure 1) also has considerable potential for investigating the rates of a range of important marine processes; 234Th/238U disequilibrium has been used to investigate open ocean euphotic zone productivity rates,8 nearshore scavenging processes,9"°

and particle fluxes and reworking rates in deep ocean sediments)' 234Th is therefore an extremely useful natural radio-tracer. There are, however, difficulties associated with its analysis; it is a weak 3emitter (E j3 = 1 9OkeV), it gives rise to a daughter nuclide, 234mpa, which is also a 3 emitter (F /3MAX = 2.33 MeV), and it has a half life of 1.17 minutes. 234'Pa is generally 461

Th

234Pa

S

/

2JOp0

MeV

3,11mm

'3

22.3y

210Pb

MeV

7,687

P

/5.Old

a tab Ia

aoapb

5,304 May

13 5.4 d

Natural decay series (major branching schemes).

26.SrnnIn

2"Pb

/

19.9 rn

p

5.4 90 Me V

6.623

/

atoblo

4.77nIn

no a11 '13

36,Innln MeV

Mo V

7,386

a

ISa 10

2l5p

May

6.619

Rn

3.965

22 pfl

MeV

5.716

3,825d

4.784MeV

2.14 raIn

6.038

a

Il.4d

I.6l0

°°'Th ,18.7d

I600y

2I.8y

2274c

MaV

.om3

3.28xI0Y

223R0

I,OSd

oaTh

j

j a 4 .395 Mc V

226p a

4.6 68 Me V

a

4776Mev

7.54nl0y

y

6.003 Me V

1.17 mm

'3

Figure 1.

24.Id

4.!0Ot CV

.47l0'y

?IBU

U

May

4,010

P

0.6h

00011

MeV

6.051

/ 13

aloble

200Pb

May

33.7%

8.784

aI0S

°t2Po

Ma V 1.0th

'3

66,3%

6.779 p

0.155

ajaR0

6.288 May

55.6S

220Rn

5.686 May

S

5,423 May

DETERMINATION OF 2TH IN WATER COLUMN

463

present in secular equilibrium with 234Th during its analysis. The conventional approach to the analysis of 234Th concentrations in seawater involves spiking the filtered sample (typically of volume - 40L) with Fh as a yield tracer, scavenging the thorium from solution by coprecipitation with ferric hydroxide, redissolution and extraction of the thorium by anion exchange separation, electrodeposition of the thorium on a planchette, and measurement of the combined 2Th plus 2Pa activity using a G.M. or proportional counter. The °Th activity is subsequently determined by alpha spectroscopy using a surface barrier detector. The limitation on detection efficiency using these 2r counting systems is one of the main factors affecting the sensitivity of this analysis, and the method also requires intercalibration of the alpha and beta counting systems. Gamma spectroscopy analysis of 2Th using its gamma photon of energy (63.2 key) offers an alternative method of determining 234Th concentrations, but the intensity of 2.5''o for this decay mode, in conjunction with the low efficiency of gamma photon detectors, gives rise to a relatively low sensitivity for this method (e.g. typical detection efficiency using a 130 cc Ge (Li) detector = 0.5°!o). Modern liquid scintillation counters offer a highly attractive alternative for the analysis of 234Th in that much higher detection efficiencies can be obtained

and simpler chemical separation and source preparation techniques can be employed. The low-level analysis capability now routinely available with mod-

ern liquid scintillation counters provides suitably low backgrounds for the determination of 234Th, and the provision of spectroscopy capability simplifies simultaneous counting of alpha and beta particles so that 234Th and the 230Th yield tracer can be analyzed in a single count. A description is provided below

of the development of a liquid scintillation method for analysis of 234Th in seawater. The work is being undertaken as part of the U.K. BOFS (Biogeochemical Ocean Flux Study) program which is associated with the international JGOFS (Joint Geochemical Ocean Flux Study) program. Development of the method was performed using a Packard 2000CA/LL liquid scintillation counter. Comparison was made between the low level count mode which employs pulse shape analysis to discriminate true f3 events from background events'2 and the normal counting mode. As a result of this pulse shape discrimination, a small percentage of the 3- efficiency is lost, together with a proportionately higher percentage of the a events. The latter are well documented as having much

broader pulse widths and therefore being more liable to discrimination. Despite this, it was envisaged that the achievable reductions in background would more than compensate for the loss of efficiency.

EXPERIMENTAL Isolations of 230Th and 234Th

The 230Th spike was prepared by anion exchange isolation of Th, from a mineralized uranium nodule in which the 28U decay chain was in equilibrium,

464

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

at least to 23&Th. The concentration of 232Th was below the limit of detection as

well. The 230Th fraction was retained in dilute nitric acid with aluminium carrier. Aging for approximately 12 months ensures decay of 234Th, 231Th, 227Th, and the corresponding decay products to a negligible level (Figure 1). The uranium fraction was also retained and similar aging allows 234Th to come to secular equilibrium with the 238U. This provides a source of 234Th free from 230Th for method development. Following isolation of 234Th from 2U, a period of a few days should be allowed for decay of 231Th (tl/2 = 1 .06d) formed by decay of 235U which was also present in the solution. The purities of both the 230Th and the U/234Th aged solutions were confirmed by electrodeposition and alpha spectroscopy. The work presented here was performed with spikes which were not completely aged, however, so an allowance was made for this in calculations. Liquid Scintillation Counting of 230Th and 234Th To obtain representative spectra, suitable aliquots of 234Th, 230Th, the combination of 234Th and 230Th, and background were prepared as follows: Aliquots were added to 7 mL Packard low potassium glass vials and gently taken to dryness. The Th was redissolved in 0.5 mL of I mol dm-3 HC1, and 5 g Packard Hionic Fluor was added.

Each sample was counted on a Packard 2000 CA/LL liquid scintillation counter for 300 minutes, first in the low level mode and then in the normal mode.

Spectra were saved for subsequent analysis using the Packard Spectragraph software package.

Following measurement of the 234U activity, the 234Th activity was calculated

based on the ingrowth time since U/Th separation. To determine the counting efficiency of 234Th (234mPa) the following steps were carried out: a known aliquot of 238U/234Th was spiked with the 230Th yield tracer, and the Th was isolated again by anion exchange. The volume was reduced to 1 to 2 mL and quantitatively transferred to a 7 mL scintillation vial, taken to dryness, and prepared for counting as previously described. The same activity of 230Th yield

tracer was added directly to a further vial, taken to dryness, and similarly prepared. Seawater analysis

Replicate 20 L seawater samples were collected from the River Clyde estuary. The 238U concentration was determined by anion exchange separation and

conventional alpha spectroscopy. The samples were filtered through 0.45 m millipore membrane filters to remove particulate material, and the pH was lowered to 2 with HC1 to maintain Th solubility. Three of these samples were then analyzed for 234Th almost immediately, while a further four were stored to

DETERMINATION OF 2TH IN WATER COLUMN

465

allow 234Th ingrowth (23 to 25 days). Samples were first spiked with 23&Th and thereafter 500 mg of FeC11. 6H20 was added to scavenge out Fe/U/Th etc. by

pH adjustment to 9. To maximize recovery, the 20 L samples were stirred for approximately 2 hr and left overnight for precipitate settling. Th separation and vialing/scintillation counting was as previously described.

RESULTS

The spectra obtained from counting 234Th, 230Th, the combination of 234Th and 230Th, and a background, in both counting modes, are presented in Figure 2. Although there is a reasonable degree of spectral separation, it is not complete. There is some interference in the 230Th region from the high energy 234mPa 3 emissions. Conversely, residual 234Th in the 230Th yield tracer brings about a

small interference in the 234Th region. Provided that the degree of sample quenching remains constant, the extent of interference will remain constant, and indeed this has been the case throughout the study. Using Spectragraph, windows of 0 to 80 keY and 100 to 220 keY were selected for 234Th and 230Th,

respectively. For the 234Th spectrum obtained under low-level conditions, a fully optimized window of 10 to 70 keV was determined using software developed at SURRC. The respective crossovers were then calculated in both counting modes and for both 0 to 80 and 10 to 70 keY counting windows. The 230Th contribution into the 234Th region will decrease with aging, i.e., as the residual 234Th decays. The contribution of 234Th/234mPa will remain a constant percent-

age of the net count rates in the 0 to 80 and 10 to 70 keY windows. As an example of this, under low-level conditions and using a 0 to 80 keY window, the contribution 234Th/234mPa under the 230Th peak was estimated at 16.5 ± 0.002%. The contribution, at the time of counting of 230Th in the 234Th region was 3.9 ± 0.93%. Using the low level option there is an approximate 10% loss in 234Th counting efficiency and a 45% loss in 230Th efficiency. 234Th efficiencies were determined as follows: (1) gross counts in appropriate regions were obtained for all vials from stored spectra using Spectragraph. (2) Appropriate backgrounds were subtracted from each region and corrections for crossover interferences were made, thus yielding net 234Th/234mPa and 230Th

count rates. (3) Yields were determined from the ratio of sample 230Th count rates to those of 230Th added directly to scintillation vials. (4) The count rates in the 234Th/234mPa region were corrected for decay and the yield factor was applied. (5) The activity of 234Th was calculated from the 238U activity as 470.5 ± 7.6 dpm ml-'. Since 234mPa is in equilibrium with 234Th, the gross activity will

be 941 dpm mL'. Table 1 presents triplicate values of the overall counting efficiency for 234Th/234mPa in the two counting windows and in both counting modes together with yields representing the efficiency of the chemistry.

Table 2 indicates background count rates and E2/B values for relevant counting windows and in both counting modes. The results show that E2/B is maximized using a 10 to 70 keY window in the low level mode. However, this

466

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS 200, 0CPNY.7j,. //A0tj0FiI

(a)

RKC00007 .1.0.

\\N7lA0t00Ft

Low level Count Mode Background Spectrum

500

//Ati

(b)

0ldp01!lt02]7k7V Fl1: 23000027,LL

\\N,A010jl7Fj

Low Level Count Mode

"Th Spectrum

200.0

IIAti.. Pi1

(c)

23000007.0.1.

\\N77A0t0,/.

Low Level Count Mode °Th Spectrum

500

200. 0CP77V7j (d)

Low Level Count Mode

//A0j1' P11: 002001127.0.0.

Z!ld200,1501370,V

\\N7A01tlFj0

Th+°Th Spectrum

070

Figure 2a.

Low-level Count mode spectra for background, 2Th, 2°Th, and 2Th + 230Th.

467

DETERMINATION OF 2TH IN WATER COLUMN 200. 0CPMY4,0j 020:

I I

(e)

220000024.20CM

ECOpOICt lOSS kfl

0004ttjCE FlOE:

Normal Count Mode Background Spectrum

ECI000

//aOtjOE Eli.: 25440024.4CM

(I)

40 N2008C2010E

450 SEV 2.:

Normal Count Mode

2'th Spectrum

CpMtE&tiE0t 200.0 /1802040. ill.: 22000C22,I30M

(g)

0dp02MO 392 \\H0030tlOE FIX.:

8200

Normal Count Mode °Th Spectrum

SOS

//AC204O.

(h)

pilE:

C0M011024.CCM

44200204001

flI,:

Normal Count Mode

Th+°Th Spectrum

Figure 2b.

Normal count mode spectra for background, 2Th, 230Th, and 2Th + 230Th.

468

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 1. 3Th Chemical Yields and Counting Efficiencies Using Both the Low-Level and Normal Counting Modes and 0 to 80 and 10 to 70 keV Counting Windows Chemical Normal Mode % Eff. Low Level Mode % Eff. Yield Replicate (%) 0-80 key 10-70 keV 0-80 keV 10-70 keV 1 136.9 80.0 100.3 115.5 85.0 79.1

2 3

138.7 137.8

73.0

102.0 101.3

118.2 120.0

85.7 88.2

Table 2. Background Count Rates and E2IB Factors for 234Th and Background Count Rates for 230Th Using Both the Normal and Low-Level Counting Modes Nomial Count Mode

Low Level Count Mode

Th Th °Th Th 2Th 230Th (0-80 key) (10-70 key) (100-220 key) (0-80 key) (10-70 keV) (100-220 key) Background (cpm) E2/B

16.75 1134

6.73 1522

-

2.27

6.46 2152

2.77 2689

-

0.93

improvement is partly negated at present by larger errors in the final net count rate calculations because the 230Th spike is not fully aged. With a fully aged spike, overspiking with 230Th, followed by counting in the low level mode, should be the optimum method. Table 3 represents the measured 234Th concentrations and the predicted values based upon the assumption of insignificant 234Th activity initially present in these high suspended particulate content, shallow waters.'° The results for the short term ingrowth of 234Th are generally consistent with the predicted values,

whereas systematically lower values of 80 to 90 of predicted values are observed for longer ingrowth periods. This is probably a result of partial scavenging during the longer ingrowth periods. This would not represent a problem in practical applications of the technique in which 234Th separation is effected as quickly as possible after sample collection.

Table 3. Predicted and Measured 234Th Concentrations in Replicate 20 L Estuarine Seawater Samples (Errors Quoted at the ± 1 Sigma Level of Confidence) Measured 234Th (dpm L) Predicted Replicate lngrowth 234Th Normal Mode Low Level Mode Number time (days) (dpm L1) 0-80 key 10-70 keV 0-80 keV 10-70 keV 1

2 3 4 5 6 7

2.0 4.7 4.7 23.8 23.8 25.7 25.7

0.11(0.01) 0.25(0.01) 0.25(0.01) 0.98(0.02) 0.98(0.02) 1.03(0.02) 1.03(0.02)

0.12(0.03) 0.25(0.03) 0.35(0.03) 0.87(0.02) 0.86(0.04) 0.82(0.03) 0.82(0.03)

0.15(0.03) 0.28(0.03) 0.36(0.03) 0.89(0.04) 0.88(0.04) 0.84(0.03) 0.88(0.03)

0.16(0.02) 0.26(0.02) 0.37(0.03) 0.90(0.03) 0.85(0.03) 0.81(0.03) 0.82(0.03)

0.16(0.02) 0.26(0.02) 0.36(0.03) 0.93(0.03) 0.84(0.03) 0.82(0.03) 0.83(0.03)

DETERMINATION OF 234TH IN WATER COLUMN

469

DISCUSSION AND CONCLUSIONS

The results to date indicate that 234Th measurements from seawater samples are indeed feasible by this technique. The advantages of the method are that the preparation chemistry is simpler, since no electrodeposition is required, both isotopes may be measured in a single count on equipment employing an automatic sample changing facility, and (3) counting efficiencies are much higher for both isotopes of the order of 100% for 230Th and 137% relative to 234Th activity for combined z34Th/234mPa. As a result of these, a higher through-

put of samples can be achieved with greater precision. The possibility also therefore exists of carrying out the determinations on much smaller samples. This is a major advantage for seawater analysis, where sample size is often a limiting factor. Five litre samples should be well within the capabilities of this method. In order to validate the method yield factors can and will be confirmed when all 234Th has decayed. Where a fast turnaround time of results is not required, this should enable higher precision to be obtained. The possibility also exists of using scintillation counters employing simultaneous a/li separation and count-

ing which should in theory virtually eliminate the need for crossover calculations. ACKNOWLEDGEMENTS

This work is being undertaken as part of the Biogeochemical Ocean Flux Study Program for which the financial assistance of the U.K. Natural Environment Research Council under grant GST/02/302 is gratefully acknowledged. REFERENCES

Turekian, K.K. and J.K. Cochran. "Determination of Marine Chronologies Using Natural Radionuclides," in Chem. Oceanography, vol 7, 2nd ed. J.P. Riley and R. Chester, Eds. (London: Academic Press, 1978). Lalou, C. "Sediments and sedimentation processes," in Uranium Series Disequilibrium. Applications to Environmental Problems, M. Ivanovich and R.S. Harmon, Eds. (Oxford: Clarenden Press, 1982). Colley, S. and J. Thomson. "Recurrent Uranium Relocations in Distal Turbidites Emplaced in Pelagic Conditions." Geochim. Cosmochi,n. Acta, 49:2339-2348 (19851).

Colley, S., J. Thomson, and J. Toole. "Uranium Relocations and Derivation of Quasi-isochrons for a Turbidite/Pelagic Sequence in the Northeast Atlantic," Geochini. Cosmochim. Acta, 53:1223-1234 (1989). Anderson, R.F., M.P. Bacon, and P.G. Brewer. "Removal of 23°Th and 231Pa from the Open Ocean," Earth Planet. Sci. Lelt., 62:7-23 (1983).

Anderson, R.F., M.P. Bacon, and P.G. Brewer. "Removal of 230Th and 231Pa at Ocean Margins," Earth Planet. Sc Lett., 66:73-90 (1983).

470

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Shimmield, G.B., J.W. Murray, J. Thomson, M.P. Bacon., R.F. Anderson, and N.B. Price. "The Distribution and Behaviour of 23Th and 231Pa at an Ocean Margin, Baja, California, Mexico," Geochim. Cosmochim. Acta, 50:2499-2507 (1986).

Coale, K.H. and K.W. Bruland. "Oceanic Stratified Euphotic Zone as Elucidated by 234Th:238U Disequilibria," Limnol. Oceanogr., 32:189-200 (1987). Tanaka, N., Y. Tanaka, and S. Tsunogai. "Biological Effect on Removal of 234Th, 210Po and 210Pb from Surface Water in Funka Bay, Japan," Geochim. Cosmochirn. Acta, 47:1783-1790 (1983). McKee, B.A., D.J. DeMaster and C.A. Nittrocier. "The Use of 234U/238U Disequilibrium to Examine the Fate of Particle Reactive Species in the Yangtze Continen-

tal Shelf," Earth Planet. Sci. Lett., 68:431-442 (1984). Aller, R.C. and D.J. DeMaster. "Estimates of Particle Flux and Reworking at the Deep Sea Floor Using 234Th/238U Disequilibrium," Earth Planet. Sci. Lett., 67:308-318 (1984).

CHAPTER 40

Performance of Small Quartz Vials in a Low-Level, High Resolution Liquid Scintillation Spectrometer

Robert M. Kahn, James M. Devine, and Austin Long

ABSTRACT To meet an increased demand for the radiocarbon dating of small samples via liquid scintillation counting. Haas1 introduced a technique which employed square, optically flat, quartz vials,

with optimal UV transmission characteristics, as counting solution containers. Benzene samples as small as 0.3 mL (240 mg carbon) were dated using these vials. With proper installation and handling, these vials were shown to provide excellent long-term stability and precision. We now report results on small quartz vials used in a modern low-level LS counter, the LKBWallac 1220 Quantulus, in the underground counting laboratory at the University of Arizona. Counting solutions comprise sample benzene diluted by the addition of dead" benzene to a counting volume of 0.3 mL, with approximately 0.95 weight % butyl-PBD scintillant. A lowbackground copper vial holder was designed to accommodate the vials; constant orientation and alignment during loading is maintained by the heavy weight of the holder. Samples and standards are counted for 1200 mm (60 x 20 mm count periods). Background cpm, based on running average, is 0.0454 ± 0.006 cpm at a counter efficiency of 67.4%. Figure of merit E2/B = 100,130; signal/noise (net modern standard B cpm/background cpm) = 53.5. Maximum determinable age, based on the Stuiver and Polach2 criterion, is 39,200 years before present. These performance characteristics compare favorably with data presented for 0.3 mL Teflon vials and 3.0 mL glass vials.4

INTRODUCTION

In the past decade there has been an effort to push the limits of nuclear counting techniques. The introduction of improved liquid scintillation (LS) instruments5 has reduced background noise dramatically. This decrease in noise enables the user to measure smaller samples with improved accuracy. Small sample 14C dating techniques use miniature gas proportional counting systems to measurement as little as 5 mg of carbon)2'5 Haas1 introduced the use of 0.3 mL quartz spectrophotometric vials for '4C LS counting. Devine'6 reported a figure of merit value of 2477 when using these vials in an Intertech471

472

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

nique LS2O counter. The comparison of 0.3 ml quartz and teflon vials in various counters3 shows that small sample counting has benefited greatly from developments in LS technology. This chapter presents the results of small sample (0.3 mL of benzene) counting using quartz spectrophotometric vials in an LKB-Wallac 1220 Quantulus LS counter at the University of Arizona, Radiocarbon Dating Lab. METHODS

Two quartz spectrophotometric vials of 1 mm path length, like those used by Haas,' were used (Figure 1). A special vial holder was manufactured from low-

background, low-oxygen copper (Figure 2) such that the maximum area of sample can be seen by the photomultiplier tubes, and the vial holder blocks photomuitiplier crosstalk in the rest of the sample chamber. The diameter and height of this vial holder is identical to that of a 20 mL glass LS vial. Benzene samples are synthesized from Oxalic Acid I (OXI) primary standard and from Mississippian limestone background material. Spectrophotometric benzene is also used as background. Samples to be dated, yielding

'-

Teft.on Stopper SuppLIec with vial

Figure 1.

One millimeter pathlength quartz spectrophotometric vial.

PERFORMANCE OF SMALL QUARTZ VIALS

Top View

S;c

473

g( )P

Side Vw

View

Made of Low

Oxygen

Copper

Bottom View Figure 2. Low-background, low-oxygen copper vial holder.

benzene and weighing between 80 and 350 mg, are diluted to 350 mg carbon as benzene. Butyl-PBD, 0.95 wt-percent, is added to the benzene. The sample is

introduced into the quartz vial via syringe, weighed, and placed into the LS counter, counted 60 x 20 mm intervals, weighed after counting, and removed from the vial with a syringe. The quartz vials are flushed with spectrophotometric benzene repeatedly, then dried. Time series data on OXI and background are compiled and averaged for age

calculations 2 Laboratory quality assurance - quality control (QA/QC) samples of well known age are analyzed as unknowns and used to check the validity of the time series data. The energy spectrum of a synthesized OXI '4C benzene and background (BKG) benzene samples for the LKB-Wallac 1220 Quantulus are shown in Figure 3. The energy windows for this study were set to maximize efficiency while avoiding the low energy background peak.

474

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS L

R

Ox1ic Acid Primary Standard

yl

Background

108

288

300

Channels 175 - 450

480

588

600

708

808

80

1008

Efficiency = 61.22 %

Figure 3. Spectrum output from LKB-WaIIac 1220 Quantulus of OXI primary 14C standard and background benzene.

RESULT'S AND DISCUSSION

Efficiency of photon measurement is related to the alignment of the quartz vial in the LS counter. The faces of the vial must be parallel to the faces of the photomultiplier tube. The vial holder was placed in the sample tray so that the optimal alignment occurs when the elevator lifts the sample into the chamber. The degree of rotation of the copper vial holder during elevator operation was determined by raising and lowering the vial holder. Some misalignment was found over 60 repeat events, but the weight of the copper vial holder produces no noticeable misalignment for one loading event. Therefore, no special alignment mechanism was manufactured, rather, samples are loaded into the cham-

ber, and 60 x 20 mm repeat counts are taken without unloading the sample vial. Performance characteristics of the 0.3 mL quartz vials in the LKB-Wallac 1220 Quantulus counter (Table 1) compare favorably to those of the 0.3 mL Teflon vials,3 and to 3 mL glass vials.4 The difference in efficiency between the 0.3 mL Teflon vials and the 0.3 mL quartz vials is due to the exclusion of the low-energy end of the energy spectrum for this study. This area of low-energy background noise seen in the spectrum (Figure 3) is instrument dependent. A second LKB-Wallac 1220 Quantulus, which is next to this instrument, does not have this noise peak.4 Differences in photomultiplier tube composition most

475

PERFORMANCE OF SMALL QUARTZ VIALS

Table 1. Small Sample '4C Results in 0.3 mL Quartz Vials, 0.3 mL Teflon Vials3 and 3.0 mL Glass Vials4 0.3 ml. Quartz Vial 0.3 mL Teflon Vial 3.0 mL Glass Vials 14C efficiency (E) Average background (B) Average modern3 signal (N0)

Figure of merit E2/B Factor of merit (fM) N0/B SIN (modemibackground) Maximum determinable age Error modem sample (yrs)

74.0% 0.49 27.5

80.8% 0.05 2.66 130,570 11.9 53.2 43,230'

67.4% 0.04537 CPM 2.5286 CPM 100,130 11.9 53.46 39,20O' 140

11,090 39.1

55.7 53,400 50

130

aModem signal = 0.95 Oxalic Acid I primary standard. bSample Counted 1200 minutes. CSample Counted 3000 minutes.

likely leads to this variation among instruments (P. Makinen personal communication).

The distribution of measured events in each 20 mm time period for OXI (Figure 4) and BKG (Figure 5) fit a Poisson distribution. This technique was used for samples containing between 80 and 350 mg of carbon. The statistical uncertainty for samples less than 125 mg is usually

Oxdlc Std

30

35

40

45

50

55

60

65

70

75

80

N&rrer of Enth i 20 mh Figure 4.

Distribution of the number of events measured during each 20 mm time interval for OXI primary standard.

476

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

3503(X)

Bockg-oirid 250

2(X)

150

1(X)

in

50

0

2

0 1

3

4

5

6

Nuiter of ets i-i Figure 5.

7

8

10 9

20 m

Distribution of the number of events measured during each 20 mm time interval for BKG samples.

unacceptable (> 500 years), but for certain applications this information may be useful. The results of three analyses of our in-lab QA/QC program, using the small quartz vials, is presented with the 1988 to 1989 results of the same sample for 3.0 mL vials from three different counters (Figure 6). These three samples were less than 350 mg carbon, thus they were diluted before measurement. At one standard deviation, the error bars overlap with the known age range of this sample. The performance of small sample counting with the quartz vials makes LS counting an option for those who need radiocarbon dates on small samples but do not have access to, or funds for, AMS determinations. The disadvantage of small sample counting, compared to our routine determinations,4 is the loss of some accuracy and a reduction of the age limit which we can attain. Techniques of low-level counting, including small sample techniques, need not be limited to '4C dating. Application of low-level counting to environmental testing and biomedical research would allow a decrease in dosimetry while maintaining a high signal to noise ratio. This could reduce costs of materials and may ease some of the problems of radioactive waste disposal.

* 4

a

Same sample no Rn * Rn in sample

0

radiocarbon years

I

D

Figure 6. Small sample QA/QC results with the July 1988 to June 1989 QNQC results of the Arizona fladiocarbon Laboratory.

10000

F

F Ill liii 1 11111111111111111 I' I Jill! ''I' 11000 12000 13000 14000 Radiocarbon Age

p

210 mg sample a 212 mg sample 214 mg sample

Known range 11.700 to 12.200 ybp

478

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

CONCLUSION

The advent of new LS technology has allowed our lab to decrease the minimum sample size for 14C dating to less than 300 mg of carbon. Special alignment mechanisms are not required to keep the vial in the optimal position, provided the vial is loaded only once into the chamber. The performance of the 0.3 mL quartz vials compares favorably with that of 0.3 mL Teflon vials and 3.0 rnL glass vials. Signal and background events fit a Poisson distribution. Saniples as small as 80 mg have been measured, but they lose accuracy. QA/QC samples determined with this method are within an acceptable range. Applications beyond '4C dating of milligram size samples should be explored using techniques for small sample LS determinations. REFERENCES Haas, H. "Radiocarbon Dating," in Proceedings of the 9th International '4C Conference, R. Berger and H. Suess, Eds. (1979), pp. 246-255. Stuiver, M. and H.A. Polach. Radiocarbon, 19:355-363 (1977). Polach, H.A., L. Kaihola, S. Robertson, and H. Haas. Radiocarbon, 30:153-155 (1988).

Kahn, R.M. and A. Long. in Proceedings of the 13th International Radiocarbon Conference A. Long and R. Kra, Eds. (Dubrovnic, Yugoslavia: Radiocarbon Vol. 31, 1989).

Noakes, J.E. and J.D. Spaulding. in Liquid Scintillation Counting: Recent Applications and Developments Vol. 1, Physical Aspects, C.T. Peng, D.L. Horrocks, and E.L. Alten, Eds. (New York: Academic Press, 1980), pp. 160-170. Rajamae, R. and J.M. Punning. in Proceedings of the 10th International Radiocarbon Conference, M. Stuiver and R. Kra, Eds. (Bern: Radiocarbon 22, no. 2, 1980), pp. 435-441. Kojola, H., H. Polach, J. Nurmi, T. Oikari, and B. Soini. in mt. J. AppI. Radial. Isotopes, 30:452-454 (1984).

Polach, H.A., J. Nurmi, H. Kojola, and E. Soini. in Advances in Scintillation Counting, S.A. McQuarrie et al., Eds. (Edmonton: University of Alberta Press, 1984). pp. 420-441.

Oikari, T.H. Kojola, J. Nurmi, and L. Kaihola. in mt. J. Appl. Radiat. isotopes, 38:9 (1987).

Noakes, J.E. and R.J. Valenta. in Proceedings of the 13th International Radiocarbon Conference. A. Long and R. Kra, Eds. (Dubrovnic, Yugoslavia: Radiocarbon Vol. 31, 1989), pp. 332-341. Kessler, M. in Proceedings of the International Workshop and Inter-C'omparison of '4C' Laboratories. A Long and R. Kra, Eds. (Glasgow: 1989). Srdoc, D., B. Obelic, and N. Horvatincic. in Proceedings of the 11th International Radiocarbon Conference, M. Stuiver and R. Kra, Eds. (Seattle: Radiocarbon 25, no. 2, 1983), pp. 485-492. Sheppard, J.C., J.F. Hopper, and Y. Welter, in Proceedings of the 11th International Radiocarbon Conference, M. Stuiver and R. Kra, Eds. (Seattle: Radiocarbon 25, no. 2, 1983), pp. 493-499.

PERFORMANCE OF SMALL QUARTZ VIALS

479

Hut, G., J. Keyser, and S. Winjima. in Proceedings of the 11th International Radiocarbon Conference, M. Stuiver and R. Kra, Eds. (Seattle: Radiocarbon 25, no. 2, 1983), pp. 547-552.

Otlet, R.L., G. Huxtable, G.V. Evans, D.G. Humphreys, T.D. Short, and S.J. Conchie. in Proceedings of the 11th International Radiocarbon Conference, M. Stuiver and R. Kra, Eds. (Seattle: Radiocarbon 25, no. 2, 1983), pp. 565-575. Devine, J.M. and H. Haas. Radiocarbon, 29,1:12-17 (1987).

CHAPTER 41

The Optimization of Scintillation Counters Employing Burst Counting Circuitry

G.T. Cook, R. Anderson, D.D. Harkness, and P. Naysmith

INTRODUCTION

Recognition of the versatility of liquid scintillation counters in environmental radioactivity, together with the advent of small gas proportional counters and the accelerator mass spectrometry technique (AMS) for small sample radiocarbon dating, has undoubtedly led to the development of a new generation of scintillation counters with much reduced background count rates and enhanced E2/B values. This consequently enables much lower detection limits. The two instruments most commonly employed in U.K. laboratories are the LKB-Pharmacia Quantulus and the Packard 2000CA/LL (and subsequent models). The Quantulus has much enhanced passive shielding in addition to an active coincidence guard counter; the counter contains a mineral oil based liquid scintillant enabling the detection of cosmic and environmental gamma radiations. Further features to optimize performance include a user-set pulse amplitude comparator and high purity Teflon/copper counting vials, the latter enabling a 30% reduction in background compared to glass vials."2 The Packard 2000CA/LL does not require an active or enhanced passive shielding. Its ability to reduce background count rates via burst counting circuitry enables the characteristics of sample scintillation pulses to be differentiated from those of background pulses through pulse shape/duration analysis.3 Most background pulses have a number of afterpulses in addition to the fast prompt pulse common to all events. Although true f3 events, particularly at higher energies, may have a certain number of afterpulses, a good degree of discrimination is certainly achievable. The detector assembly incorporating the slow scintillating plastic, a feature of the Packard-2260XL, enhances this time resolved burst discrimination since its long decay constant increases the number of photons in the afterpulses of background radiation, therefore acting as a quasiactive guard. One criticism of the Quantulus has been that a lack of uniformity in the 481

482

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Teflon vial characteristics has meant characterizing each vial individually. This necessarily reduces the element of routineness obtainable from the more uni-

form glass vials, and hence caused their abandonment for t4C dating at SURRC except with samples whose expected ages exceed 25,000 years. Here,

the improvement in background can be critical in terms of age resolution. Nevertheless, for 2 mL (1.75 g) and 4 mL (3.5 g) geometries in 7 mL slimline glass vials, efficiencies of 65 and 67.6% and backgrounds of 0.44 and 0.70 cpm, respectively, are obtainable. A criticism of the Packard 2000CA/LL has been that although background count rates are considerably reduced, there is also a considerable reduction in overall counting efficiency.4 This implies that the burst counting circuit is not 100% effective and that a certain percentage of true events are being discriminated out, presumably because their prompt pulse widths are sufficiently

broad or they have an uncharacteristically large number of afterpulses. A logical extension to this argument is that by sharpening these pulses, much of this loss in efficiency could be regained. The results presented here are a résumé of a considerable data set built up while assessing counters employing burst counting circuitry.

EXPERIMENTAL

This study can, in essence, be divided into three component stages. First, the influence of a range of scintillants, their concentrations, and their combinations were assessed to determine influences on counting efficiency. This was followed by an assessment of two modifications to the original system: (1) the detector assembly incorporating a slow scintillating plastic and (2) vial holders produced from the same material; studies were carried out on both 3H and '4C.

Finally, for any radiocarbon dating laboratory the ultimate test must be the production of accurate age measurements. Naturally, the sample counting is only one aspect of the entire dating process; however, recent studies indicate that it is particularly important in terms of overall accuracy.5 This th:ird section can be subdivided as follows: A sample of benzene, synthesized from the peat used in stage 3 of the recent International Collaborative Study, was re-vialed into two 7 mL slimline glass vials. This sample had previously been dated using an "old technology" Packard 4530, with butyl-PBD/bis-MSB as the scintillant, giving an age of 3360 ± 50 years BP. The mean result for the 28 laboratories who dated this material was 3388 BP.6

One aspect which arose from the international study and which caused some

considerable concern was that the NERC Laboratory dated two duplicate samples of marine shell material, each separated into inner and outer fractions, thus yielding four age measurements. When using old technology counters, all four results were well within the overall spread, however, when using the burst counting circuit, one of the four was enriched relative to

OPTIMIZATION OF COUNTERS EMPLOYING BURST COUNTING CIRCUITRY

483

modern. In this instance, butyl-PBD alone (13 mg mL') was the scintillant employed. To determine whether this effect was specific to the scintillant, the sample was re-vialed into three separate vials with appropriate dilutions and additions of toluene and bis-MSB to exactly duplicate the scintillant used in the University of Glasgow Laboratory.

Throughout the optimization studies involving different scintillant combinations and concentrations, standard 20 mL low potassium glass vials were employed; 4.5 g of a '4C benzene standard (178.5 ± 2.0 dpm/g-1) were used throughout. The essential criteria for optimization were (1) constancy of efficiency with variation in scintillant concentration as measured in a 0 to 156 keV window and (2) constancy of quenching as determined by t-SIE value. t-SIE is the spectral index of the transformed Compton spectrum of the external standard. Values can in theory range between 0 and 1000; a value of 1000 indicating no quenching. The 2000CA/LL employs a 20 Ci '33Ba standard, the 2260XL an 8 Ci '33Ba standard. The optimization studies on the burst counting circuit, slow scintillant detector assembly, and vial holders were carried out in 7 mL glass vials using 2 g of a

high activity '4C standard (2340 ± 8 dpm/g benzene) and 0.42 g of a butylPBD/bis-MSB (12 and 6 g/L respectively) scintillant combination dissolved in toluene. The same geometry was also adopted for all dating studies. For 3H assessments, tritiated water (410 dpm/g-1) was used throughout. Dating of the benzene sample synthesized from peat was carried out in the 2000CA/LL using the burst counting circuit, in the 2260XL using the burst

counting circuit, and finally with the addition of the vial holders to the 2260XL system. Dating of the shell-derived sample was carried out on the 2000CA/LL.

RESULTS AND DISCUSSION

Table 1 indicates that with the burst counting circuit activated, open window

counting efficiency decreases from 84.1 to 73.907o as the concentration of butyl-PBD increases from 2 to 14 mg/g-' benzene. In parallel, t-SIE values increase to a plateau value at 10 mg/g and above. With the burst circuit off, the same trend in t-SIE values is observed; however, efficiency remains constant between 4 to 20 mg/g at approximately 93.5. These features are found to be common to virtually all scintillants and scintillant combinations. That is: t-SIE value trends are virtually identical whether the burst counter is on or off. Throughout virtually all the concentration ranges, efficiency is constant at approximately 93.5% with the burst Counter off, but variable, tending to a plateau value with the burst counter on.

484

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 1. Effect to Butyl-PBD Concentration '4C Spectral Stability and Counting Efficiency with and without the Burst Counter Circuit in Operation Butyl-PBD Eff. (%) EU. (%) t-SIE Difference t-SIEb BC ona (mgIg1 of C6H6) BC on BC off BC off in % Eff. 2 4 6 8 10 12 14 16 18 20 BC

84.1

80.7 78.9 76.6 75.7 75.2 74.8 73.9 73.5 73.7

557 683 730 749 757

91.8 93.5 92.8 93.5 93.2 93.6

761 761

93.1

760 758 755

93.4 93.0 93.8

557

7.7

681

12.8 13.9 16.9 17.5 18.4 18.3 19.5 19.5 20.1

732 751

755 762 760 759 759 755

Burst counter.

btSIE = spectral index of hte transformed Compton spectrum of the external standard.

Table 2 represents the optima in efficiency for these plateau values using a range of scintillants and scintillant combinations. It should be noted that the highest efficiencies are always observed in the presence of bis-MSB, ie. bisMSB alone, butyl-PBD/bis-MSB and PPO/bis-MSB, at about 89 to 90% compared. with about 93.5% with the burst counter off. This suggests that bis-

MSB differs in some respect from the other scintillants, a factor which undoubtedly merits further investigation in terms of the pulse shaping rather than the ultimate influence on efficiency. Since the introduction of the Packard 2000CA/LL, several modifications have been introduced as previously described. Table 3 indicates the performance enhancements these can bring about. With the burst counting circuit off, i.e., used as a conventional liquid scintillation counter, optimum window efficiency is relatively poor (57.8%). This is due to the shape of the background spectrum. To maximize the E2/B parameter, the lower discriminator had to be set to 18 keY. With the burst counter on, although open window efficiency is slightly reduced, optimum window efficiency is greatly enhanced (65.0%). This is due to a lower discriminator setting of 12.5 keV, enabling Table 2. Optimum Concentrations of a Range of Scintillants/Scintillant Combinations for Achieving Maximum '4C Efficiency with Stability of Quenching on the Packard 2000CAILL

Scintillant (mg/g' of C6H6) Butyl-PBD (16) PPO (5.6) PPO (5.6) + POPOP (0.7) PPO (5.6) -F- Me2 POPOP (0.7) PPO (5.6) -F. bis-MSB (2.7) Butyl-PBD (12) + POPOP (0.7) Butyl-PBD (12) + Me2 POPOP (0.7) Butyl-PBD (12) + bis-MSB (4.0) Bis-MSB (5.3)

Eff. (%) (Burst Circuit on) 73.9 83.6 81.9 82.6 89.6 78.7 77.2 88.5 88.5

Eff. (%)

(Burst circuit off) 93.4 93.5 93.8 93.6 93.7 93.2 93.5 93.1

93.0

Difference in % Eff. 19.5 9.9 11.9 11.0 4.1

14.5 16.3 4.6 4.5

OPTIMIZATION OF COUNTERS EMPLOYING BURST COUNTING CIRCUITRY

485

Table 3. Progressive Optimization for 14C of the Packard Low-Level Counting System,

Employing Burst Counting Circuitry (Packard 2000CA/LL) and Using (1) a Detector Assembly with a Slow Scintillating Plastic (Packard 2260XL) and (2) Vial Holders Made from the Same Material (Scintillant = Butyl-PBD/Bis-MSB in Toluene) Open window Opt. window Opt. Window E2,B Background (cpm) % Eff. Counting Conditions % Eff. 2000CA burst circuit off 2000CA burst circuit on 2000CA burst circuit on + vial holders 2260XL burst Circuit Ofl 2260XL burst circuit on + vial holders

92.1 88.1

57.8 65.0

2.87

1166

1.31

3226

87.5 86.9

66.0 70.2

0.85 0.94

5134 5269

88.6

71.4

0.69

7383

E2/B to be maximized while excluding all 3H. A general feature of the modifi-

cations (slow scintillant detector assembly and vial holders) is that they enhance performance not only by an overall reduction in the background count rate, but by changing the background spectra shape and enabling reduced lower discriminator settings. E2/B is enhanced from 3226 (Packard 2000CA/LL burst counter on) to 7383 (Packard 2260XL plus vial holder). Similar trends are also observed for 3H (Table 4). With a standard 20 mL glass

vial, using 10 mL of tritiated water and 10 mL Picofluor LLT, E2/B is enhanced from 62 (2000CA/LL burst counter off) to 161 (2260XL) Similarly,

with 3.5 mL tritiated water and 3.5 mL Picofluor LLT in a standard 7 mL slimline glass vial, E2/B is enhanced from 155 to 400. The holders do not appear to enhance performance beyond that of the 2260XL. With the data currently available it is not possible to determine whether the differences between the 2260XL, 2260XL plus vial holders, and 2000CA/LL plus vial Table 4. Progressive Optimisation for 3H of the Packard Low-Level Counting System Employing Burst Counting Circuitry (Packard 2000CA/LL) and Using (1) a Detector Assembly with a Slow Scintillating Plastic (Packard 2260XL) and (2) Vial Holders Made from the Same Material Opt. Window Open Window Opt. Window Background Counting Conditions (cpm) % Eff. % Eff. E2/B (a) 10 mL H20 + 10 mL Picofluor LLT 2000CA burst circuit off (glass vials) 2000CA burst circuit on (glass vials) 2260XL glass vials 2000CA burst circuit on (plastic vials) 260XL (plastic vials)

(b) 3.5 mL H20 + 3.5 mL Picofluor LLT 2000CA burst circuit off 2000CA burst circuit on 2000CA burst circuit on + vial holders 2260XL 2260XL + vial holders

22.6 24.3

19.7 21.5 23.2 18.5 21.3

6.30 3.77 3.33 1.93 2.12

26.9 25.5

19.7

2.51

155

22.6

2.06

248

22.3 25.5 23.7

19.1

23.5 22.0

1.05 1.38 1.32

350 400 367

25.1

23.5 25.1

62 123 161

178

215

486

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 5. 14C Age Determinations on Benzene Synthesized from Peat and Shell Material Used in the International Collaborative Study

Counting conditions Peat sample 2000CA/LL burst circuit on 2260XL

2260XL + vial holders Shell sample 2000CA\LL burst circuit on

Mean '4C Age

3410 ± 40 3440 ± 60 3410 ± 70

750 ± 70

holders are significant. Table 5 gives the age measurements made on the benzene syntlhesized from both the peat and shell samples. The peat results show excellent agreement with those of the international study regardless of the counting conditions employed. Results for the shell sample seem to confirm that the anomalous measurement on this sample is in some respect tied to the scintillant employed. The mean age derived from the three replicate subsamples was calculated to be 750 ± 70, and considering the extra manipulations involved in re-vialing etc., this compares quite favorably with the mean result of 637 BP from 31 laboratories, and it is obviously significantly older than the previously derived modern age. It would appear that under certain circumstances a combination of vial plus contents and the butyl-PBD must enhance efficiency beyond the norm. The most obvious possibility would seem to be some impurity in the

benzene, perhaps altering pulse shapes or variations in dissolved oxygen, which will influence the triplet state and reducing the delayed component.2 For '4C dating with the 2260XL, the counting time of the external standard

was increased from 15 to 60 sec to reduce the greater variability in t-SIE observed. Determining a t-SIE value in the 2260XL requires an additional step. Along with the Compton spectrum produced by the vial-content interaction, there is also a spectrum produced by the interaction of the y radiations with the slow scintillating plastic in the detector assembly. The combination of these two spectra will not yield a true t-SIE value. To compensate for this, an additiona.l spectrum must be measured, i.e., that of the external standard with

an empty vial which, in effect, is brought about by interactions with the plastic. This second measurement subtracted from the first produces the "true" t-SIE value. A one minute counting time for the external standard plus sample

reduces the variability in t-SIE to the same magnitude as found with the 2000CA/LL. Using the 2260XL with vial holders produces much greater variations in t-SIE which can not be overcome by increasing the external standard counting time, even to 4 mm. To circumvent this problem, the following steps

were taken. First, a uniform response of the vial holders has to be verified. This was carried out as follows. 1. Each vial holder was loaded with a vial containing 4.5 g of scintillation grade benzene and 0.95 g of butyl-PBD (12 g/L) and bis-MSB (6 g/L) in toluene.

OPTIMIZATION OF COUNTERS EMPLOYING BURST COUNTING CIRCUITRY

487

Each vial was then counted for 5 x 100 mm. The results were entirely consistent with a uniform background response. 2. A single high activity standard (2340 dpm/g benzene) was counted sequentially in each of the vial holders (3 x 25 mm counts). The vial holders with the lowest and highest count rates then underwent further counting (4 x 50 mm).

At the end of this counting period, open and optimum window counting efficiencies were identical (87.4 and 71.0o, respectively). It appears that despite the variability in t-SIE values, between successive counts from a single

vial holder and values from different vial holders, the background and efficiency responses are essentially constant within the 20 individuals tested.

Having established this constancy of response, a quench curve was constructed using 16 vials of 2 g high activity benzene standard quenched to differing degrees with acetone. This was carried Out first without the vial holders and subsequently with them. A second degree polynomial regression analysis of the former yielded an R2 value of 99.8%, the latter yielded a value of 79.9%. Finally, the count rate using the vial holders was regressed against the t-SIE values without the vial holders yielding a value of 99.6%. The same system was then employed to determine the degree of quenching in samples, backgrounds, and modern reference standards, i.e., a large number (50) of very short (0.1 mm) sample counts, with 1 mm counts of the external standard, were undertaken without the vial holders. The samples were then counted using the vial holders for a minimum of 2000 mm via the quasisimultaneous batch counting method as would normally be adopted.7 CONCLUSIONS

The results of these studies confirm that the burst counting circuit makes the efficiency of the Packard 2000CA/LL highly sensitive to scintillant concentra-

tion and type. Decreases of up to 20 in counting efficiency are observed compared to conventional counting. These can largely be overcome by careful manipulation of scintillant concentration and type. Best overall performance is always obtained in association with the secondary scintillant bis-MSB. Efficiencies approaching 90% are routinely observed, i.e., within approximately of conventional levels. It is proposed that bis-MSB sharpens pulse widths thereby bringing many more true /3 events within the cut-off threshold of the burst counter. In some instances it has been observed that efficiency increases as the degree of quenching increases which is completely contrary to normal theory. This obviously reflects the complexity of the pulse shape analyses and must represent a complex balance of sufficient scintillant, self absorption, and the energy transfer process. The use of the detector assembly incorporating both the slow scintillating plastic (2260XL) and the slow scintillating plastic vial holders enable signifi-

cant enhancements in performance for '4C. The burst counter and plastic detector assembly both enhance 3H performance, although it is questionable

488

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

whether the vial holders with the plastic detector assembly represents any further enhancement. With standard 20 mL plastic vials containing 10 mL 10 mL Picofluor LLT a limit of detection of 1.4 Bq/L is tritiated water achievab][e for a 500 mm count. In terms of actual '4C age measurements, the limited results so far indicate that accuracy is possible with burst counting circuitry and subsequent modifications to this system using butyl-PBD/bis-MSB as the scintillant. Many more

measurements, however, are required to confirm this, particularly when the use of vital holders and the associated modifications of normal '4C dating practices are envisaged. The use of butyl-PBD alone as a scintillant for '4C dating obviously has associated problems, and until these can be identified and eliminated, its use should not be recommended. REFERENCES

Kojola, H., H. Polach, J. Nurmi, T. Oikari, and E. Soini. "High Resolution LowLevel Liquid Scintillation f3 Spectrometer," ml. J. App!. Rad. Isotopes, 35: 949-952 (1984).

Gupta, S.K. and H. Polach. Radiocarbon Dating Practices at ANU-Handbook, (Canberra: ANU, 1985). Van Cauter, S. "Three Dimension Spectrum Analysis: A New Approach to Reduce Background of Liquid Scintillation Counters," Packard Applications Bulletin, No. 006 (1986).

Polach, H., G. Calf, D.D. Harkness, A. Hogg, L. Kaihola, and S. Robertson. "Performance of New Technology Liquid Scintillation Counters for '4C Dating." NucI. Geophys. 2(2):75-79 (1988).

Scott, E.M., T.C. Aitchison, D.D. Harkness, M.S. Baxter, and G.T. Cook. "An Interim Progress Report on Stages 1 and 2 of the International Collaborative Programme," Radiocarbon, 31(1989). Aitchison, T.C., E.M. Scott, D.D. Harkness, M.S. Baxter, and G.T. Cook. "Report on Stage 3 of the International Collaborative Programme," in International Workshop on InterComparison of '4C Laboratories, 12-15 September, 1989, East Kilbride.

Stenhcuse, M.J. and M.S. Baxter. "4C Dating Reproducibility: Evidence from Routine Dating of Archaeological Samples," in '4C and Archaeology, Groningen, 1981 (1983), pp. 147-161.

CHAPTER 42

Statistical Considerations of Very Low Background Count Rates in Liquid Scintillation Spectrometry with Applications to Radiocarbon Dating

Robert M. Kahn, Larry Wright, James M. Devine, and Austin Long

ABSTRACT The new generation of low-level liquid scintillation spectrophotometers have significantly reduced background count rates. These very low rates require non-Gaussian statistical treatment of data. An LKB Quantulus, employed for radiocarbon dating of samples of between 80 to 300 mg, received no events in 41% of the 780 20-mm intervals on background benzene. The distribution of the background events is Poisson. For very small count rates, which are nonbackground (measurable 14C content) benzene, Poisson statistical treatment of data is preferable to Gaussian treatment.

INTRODUCTION

This chapter is not meant to supercede previous discussions of counting statistics. The authors agree on strict adherence to the recommendations put forth by Stuiver and Polach1 for use in the general case of radiocarbon date calculation. We would like to submit that very low background count rates may allow for the non-Gaussian treatment of data. In the last decade, many advances in liquid scintillation (LS) technology have taken place. The incorporation of anticoincidence shielding reduced background count rates.28 Recently, pulse shape discrimination techniques within active shielding9 and novel energy discrimination - after pulse analysis techniques without active shielding'°" have decreased background count rate further. The level of background noise in LS counters without these techniques can be approximated with Gaussian statistics. This assumes that the background

events are stochastic, without a periodic noise from electronics or other source. As background events are reduced over two magnitudes of order, these 489

490

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

events approach the predicted Poisson distribution for small numbers of stochastic events. The statistical interpretation of '4C data collected using decay counting has been discussed,12-17 and for '4C date calculations, Stuiver and Polach1 published the guidelines which should be followed by all radiocarbon labs. Yet, when the number of measured events approaches zero, careful consideration of the distribution of the events is required. The very low background count rates, and subsequent low sample count rates measured with LS counters, can be compared to the statistical analysis of microcoliLtamination in ultraclean environments)8 VERY LOW BACKGROUND COUNT RATES Measu:rements of very low background count rates (less than 0.1 cpm) have

been measured both with gas proportional counting systems'7"9-2' and LS counting.23'23

Figure 1 shows background data collected using 0.3 mL vials in an LKBWallac 1220 Quantulus LS counter. These data compromise 13 different background samples counted for 60 to 20 mm intervals, totaling 1200 mm each. A chi-square test of this data confirms that they fit a Poisson distribution. The arithmetic mean of these data is 0.04539 cpm. The distribution of events from

11 separate 0.95 Oxalic acid '4C standard samples (Figure 2), has 4 values which do not fit a Poisson distribution. TREATMENT OF DATA COLLECTED

A Poisson distribution is quite asymmetrical when the number of counts is near zero, but it quickly approaches a Gaussian distribution by approximation as the count number increases. Tsoulfanidis states that the Gaussian distribution is almost identical with the Poisson distribution at a mean value of 25 counts/time.24 Therefore, if an unknown were counted, such that an average

25 counts/time were measured, Gaussian statistics would be a close approximation. Data from Otlet Ct al., on small gas proportional counting systems, reported background count rates ranging between 21 and 400 counts per day.2° Increasing the time period is pne method of treating the results as a Gaussian distribu-

tion. Important information on the stability and performance of an LS counter can be lost if repeat time series data is not collected. There are many factors which can affect the fit of time series data to an expected Poisson distribution. Periodic external noise, i.e., electronic spikes, 222Rn and 220Rn contamination, and periodic internal noise will alter the fit of measured data. The signal and background will be a Poisson distribution if the. LS counter is only recording random events.

STATISTICAL CONSIDERATIONS OF COUNT RATES

491

350 300

-.250

j200 150

1Q0

I

50

0

0

2

1

4 3

5

6

7

8

10 9

F&rter of ents ii 20 m Figure 1.

Distribution of background events, 780-20 mm

intervals.

The background data collected follow a Poisson distribution. There is not any evidence to suggest a nonrandom component of these data. The '4C data have four values (of 660) which do not follow a Poisson distribution Small amounts of Radon gas were detected in 2 of the 11 samples. We hypothesize

that the non-Poisson values are the result of Radon contamination or an artifact of slight sample evaporation during the collection of data. RADIOCARBON DATING AND POISSON STATISTICS

Questions arise when determining the age of a sample from a small number of events. How certain are the investigators that the events measured are from the original sample material and not from some source of contamination? An assessment of this can be obtained by analyzing background material which undergoes identical treatment along with samples.

A chi-square test of the time series background count rates can alert the investigator to Radon contamination or periodic noise. We also suspect that an analysis of the length of time between background events may be useful in

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

492

50

40

I

30

I

20

10

25

30

35

45

40

50

65

55 60

70

75

80

Fthr of E,ts i-i 20 ml,. Figure 2. Distribution of 0.95 oxalic acid 14C standard, 660-20 mm intervals.

determining a component of periodic noise, and we will investigate this possibility further. The introduction of systematic errors by lab personnel can be determined by

measuring repeat samples of known '4C activity, such as lab standards or known values quality assurance - quality control samples. The authors would like to suggest the use of Poisson distribution tables like that of Crow and Gardner for determining acceptance criterion when calculating radiccarbon ages.2' For the background count rate in this study, an average of 55 events is expected during the 1200 mm of counting. The two criteria of Stuiver and Polach for 55 events would calculate a "greater than" age using 17 events above background.' The Poisson tables of Crow and Gardener would calculate the "greater than" age, 95o confidence, using 16 events.2' Although one event may be considered trivial, it represents 2o of the total events measured. Assuming that the laboratory researchers have very carefully studied all sources of error within the lab and those associated with the LS equipment, there is a probability that one event can be considered discrete. Rejecting that one event is a loss of information. As the background count rates approach zero, this discrepancy between Gaussian and Poisson distributions becomes a greater portion of the total signal. It may be best to measure a sample for a period of time sufficient to collect enough events so that this discrepancy is negligible. But, this may not be cost

STATISTICAL CONSIDERATIONS OF COUNT RATES

493

efficient, or LS technology may advance to the point that the discrete nature of single events must be taken into account. CONCLUSION

The background events measured follow a Poisson distribution. Four of the 660 events for the 'C standard did not fall within the chi-square goodness of fit criteria. These events are hypothesized to be radon contamination or artifacts of evaporation during counting. For very low background count rates, the goodness of fit to a Poisson distribution should be calculated, and an attempt to identify any nonrandom component of the background should be made. Each lab should have a degree of confidence for samples prepared in the lab based on reproduceablility of backgrounds and known activity samples. And finally, by using Poisson statistics, slightly more information may be

obtained than when using the recommendations of Stuiver and Polach (1977).'

REFERENCES

Stuiver, M. and H.A. Polach. Radiocarbon. 25, 2:458-492 (1977). Peitig, von F. and H.W. Scharpenseel. Atompraxis. 7:1-3 (1964). Punning, J.M. and R. Rajarmae in Low-Radioactivity Measurements and Applications. P. Povinec and S. Usacev, Eds. (Bratislava: Slov Pedagog Nakladatelstov, 1975), pp. 169-171. Allesio, M., L. Allegri, F. Bella, and S. Improta. Nuclear Instruments and Methods. 137:537-543. Noakes, J.E. in Liquid Scintillation Counting, Vol. 4, M.A. Crook and P. Johnson, Ed. (London: Heyden and Sons, 1977), pp. 189-206. Broda, R. and T. Radoszewski. International Conference on Low Radioactivities P. Povinec, Ed. (Bratistav: VEDA, 1982), pp. 189-206. Jaing, H., S. Lu, S. Fu, W. Zhang, T. Zhang, Y. Ye, M. Li, P. Fu, C. Peng, and P. Jaing, in Advances in Scintillation Counting, S.A. McQuarrie et at., Eds. (Edmonton: University of Alberta Press, 1983), pp. 478-493.

Kojola, H., H. Polach, J. Nurmi, T. Oikari, and E. Soini. mt. J. AppI. Radiat. isotopes. 35:949-952 (1984).

Oikari, T., H. Kojola, J. Nurmi, and L. Kaihola. mt. J. Appl. Radiat. Isotopes. 38:9 (1987).

Noakes, J.E. and R. Valenta. Proceedings of the 13th International Radiocarbon Conference, A. Long and R. Kra, Eds. (Dubrovnic: Radiocarbon 31, 1989), in press.

Kessler, M. Proceedings of the International Workshiop on inter-Comparison of 14C Laboratories, A. Long and R. Kra, Eds. (1989), in preparation. Currie, L.A. Anal. Letters, 4:777-784 (1971). Currie, L.A. International Conference on Radiocarbon Dating, 8th Proceeding,

T.A. Rafter and T. Grant-Taylor, Eds. (Wellington: Royal Society of New Zealand, 1973) pp. 598-611.

494

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Pazdur, M.F. ml. J. App!. Isotopes. 20:179-184 (1976). Walanus, A. and M.F. Pazdur. Radiocarbon. 22, 4: 1021-1027 (1980). Sheppard, J.C., J.F. Hopper, and Y. Welter. Radiocarbon, 25,2:493-500. (1983).

Currie, L.A., R.W. Gerlach, G.A. Kiouda, F.C. Ruegg, and G.B. Tompkins. Proceedings of the 11th International Conference, M. Stuiver and R, Kra, Eds. (Seattle: Radiocarbon, 24,2 1983), Pp. 553-564. Bzik, T.J. Microcontatnination, May 1986, pp. 59-99. Loosi, H.H., M. Heimann, and H. Oeschger. Radiocarbon, 22,2:461-469.

Otlet, R.L., G. Huxtable, G.V. Evans, D.G. Humphreys, T.D. Short, and S.J. Conchie. Radiocarbon 25,2:565-575. Crow, E.L. and R.S. Gardener. Biometrika, 46:441-453 (1964).

Polach, H.A., L. Kaihola,

S.

Robertson, and H. Haas. Radiocarbon,

30,2:153-155 (1988).

Kahn, R.M., J.M. Devine, and A. Long, personal observation. Tsou!fanidis, N. In Measurement and Detection of Radiation. (Washington, D.C.: Hemisphere Publishing Corp., 1983), p. 42.

CHAPTER 43

Liquid Scintillation Counting Performance Using Glass Vials in the Wallac 1220 QuantulusTM

Lauri Kaihola

ABSTRACT Low-potassium glass vials can obtain reduced background count rates for beta particles in

liquid scintillation counting with the aid of pulse amplitude comparisons and pulse shape analysis; they also retain relatively high counting efficiencies. Adding secondary fluor to the solvent, to shorten the pulses from sample decay events, helps the effectively separate these from the slow glass fluorescences. The power of this method in improving counting performance was examined for: 14C in benzene in normal and low background environments using the Wallac lowlevel liquid scintillation counter, 1220 QuantulusTM, with an active anticoincidence guard detector. The primary fluor was 15 mg/mL butyl-PBD, and the secondary fluor was 1 mg/mL bis-MSB, 3H in aqueous solution with commercial cocktails. In a normal environment, using Wheaton 20 mL low-40K vials filled with 5 mL of benzene, the best 14C figures of merit were achieved in a window of 65% counting efficiency. It gave a 0.78 cpm background count rate using low bias with the pulse shape analyzer (PSA) alone and a 0.51 cpm background count rate using the pulse amplitude comparator (PAC) in conjunction with PSA. The corresponding figures in a low background environment were 0.57 and 0.27 cpm. Tritium background count rates were 2.8 and 2.9 cpm at 27 and 22% counting efficiency using QuickzintTM 400 and Optiphase HisafelM, respectively, with PSA + PAC.

INTRODUCTION

Using glass vials in low-level liquid scintillation counting offers some advantages over plastic and teflon vials. The glass is inert and easy to clean (cleanliness can be checked by visual inspection). It is dimensionally stable and imper-

meable to aromatic and volatile solvents such as the benzene used in

'4C

dating. For long term low-level counting, the accompanying plastic caps can be replaced with metal caps lined with indium gaskets to prevent vapor loss of the volatile sample.' Teflon vials were specially designed for counting '4C using 495

496

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

benzene and still reach the best performance, but they require very careful calibration and cleaning.2 The search for an ideal vial material, in terms of low-level counting performance, is in progress; quartz might be the one.3'4'5 However,, the cost of inherently low radioactivity synthetic quartz is very high. The ability to use standard low-40K glass vials would be a good choice for low-

level counting if the background can be reduced to acceptable levels at a relatively high counting efficiency.

METHOD

Glass vials exhibit unquenchable and quenchable background radiation components. The unquenchable background is caused by 1.3 MeV beta particles emitted in the 40K decays in the glass and by high energy particles interacting with i:he vial and PMTs (approximately 10 cpm in the QuantulusTM; this is seen with empty glass vials as well). Further, fluorescence events are also created in the glass, mainly in the low energy region of the spectrum. These pulses are slower than the Cerenkov radiation created within the glass. This part of the background severly interferes with 3H beta spectrum. The quenchable, mainly high-energy background, is caused by photons from 40K decays in the vial and PMTs; the result is a Compton continuum in separable from a be.ta particle spectrum. Beta particles from 40K decays and alpha particles (U,Th) from the inner surface of the vial reach the cocktail, contributing to the high energy background.6 Bias Threshold

The Quantulus has a user controlled bias threshold which can be set either low or high. The acceptance levels of the pulse amplitudes are tested for both PMT signals individually, before the coincidence summing of pulse heights.7 At low bias, maximum 3H detection efficiency is retained, while high bias effectively rejects cross talk, Cerenkov, chemiluminescence, and low energy pulses (e.g., part of the 40K spectrum in glass and most of the 3H spectrum). It is very effective in reducing background for '4C, but it cannot be used for this purpose with low energy radiations. Pulse Amplitude Comparator The Wallac Quantulus electronics contain a user adjustable pulse amplitude compara1or (PAC) that can to reject those events in the PMT tubes which vary more in their amplitudes than specified by their selected PAC setting.8'9 PAC is designed to discriminate differently between low energy (3H and others) and high energy (e.g., 14C) pulses; the lower the energy of the ionizing event, the milder the PAC effect. The 14C background cross talk events at low bias are more likely to be rejected by PAC due to their greater pulse amplitude dispar-

COUNTING PERFORMANCE USING GLASS VIALS

497

ity than the sample decay events. In this way one is also able to improve the '4C S/N ratio without significant efficiency losses.

Pulse Shape Analysis

The pulse shape analyzer in the Quantulus is an analog device which integrates the tail of each pulse)° This integral is then compared with the total pulse integral to produce an amplitude independent parameter that relates to the pulse shape. Typically a true beta pulse decays exponentially in a few nanoseconds, while an alpha pulse decays non-exponentially in a few hundred nanoseconds. The application of this pulse shape parameter is user selectable and variable. Pulse shape analysis (PSA) is primarily intended for alpha/beta particle spectrum separation. It can also be used for background reduction, provided that the pulse shape of the sample and background signals differ.

Glass Vial Backround Reduction In '4C counting of benzene, butyl-PBD is widely accepted as the fluor due to its high efficiency and quench resistance in the presence of impurities in the sample solute.2 It is, however, a quite slow fluor requiring the addition of a secondary fluor, e.g., bis-MSB, to improve pulse shape contrast between the fluorescent glass vial backround events and the fast '4C sample pulses. The

secondary fluor shifts the emission peak wavelength. This is essential to matching the peak transmission of the guard in the new Packard low-level counters, where it aids in achieving higher counting efficiency at lower The scheme was tested in an ultralow background liquid scintillation spectrometer, the Quantulus, inside and outside our low-level laboratory at Wallac. The Quantulus is fitted with a cosmic guard counter whose performance is totally independent of sample characteristics. There cannot be any beta particle energy transfer out of the actual sample spectrum into the guard. Therefore user selectable PSA enable free discrimination of long fluorescent pulses from the fast sample events in the glass, thus reducing the background without losing any efficiency through erroneous interpretation of their origin. Both the accepted and rejected pulse spectra can be analyzed and recorded to evaluate PSA and PAC performance at different settings and under different cocktail and environmental conditions.

RESULTS

Radiocarbon Standard 20 mL low-40K glass vials (Wheaton) were used for the experiments at low-bias settings. The PSA settings were scanned from ito 25 in steps of 5,

498

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 1. 14C Efficiency and Background Variations, UsIng 20 mL Low-40K Glass Vials with a 5 mL Benzene Sample, in a Normal and Low Radiation Environment with PSA

and PSA + PAC, at low bias PSA + PAC

PSA

Mode

Efficiency 14C %

77%

Bcpm

Bcpm

B reduction %

B reduction % LR

N

LR

B = 1.34

1.00

4400

5900 59%

0.78 7600

0.49 12100 82%

N

E2/B

Bred = 46% 65%

B = 0.78

E2IB = 5400 Bred = 31% Note: B = background cpm

0.57 7400 38%

72% 0.51

8450

56%

0.27 15650 67%

Bred = background reduction % in the selected window with respect to no PSA or PSA + PAC electronic discrimination

N = normal environment LR = low radiation environment

and the PAC was scanned from 180 to 250 in steps of 10. These setting ranges have perceptible effects on the counting performance in the cocktail, 15 mgI mL butyll-PBD and 1 mg/mL bis-MSB. The sample 5 mL volume was placed in an unmasked vial. Spectroscopic grade benzene was used as the background sample. '4C labeled fatty acid was used as the reference standard. Performance figures are given at optimum windows and at balance point, with the highest

figure of merit (E2/B) for the specified vial type, size, and sample volume (Table 1). Typically, a reduction of 31 to 72% in the background is observed at

65 and 77% efficiency windows when only PSA or PAC and PSA are activated, respectively, in a normal environment. The background is reduced by some 38 to 82% from the original when PAC also is activated in a low radiation envi:ronment. There is some 40% reduction in backround count rate when the instrument is taken from a normal environment into the low-level laboratory. More significantly, however, when using PAC + PSA, there is a background reduction in low-40K glass vials of 82% at a counting efficiency of 77%, and a 67% reduction at 65% efficiency. Tritium

The 3H beta emission in aqueous solutions and glass fluorescence pulse amplitude spectra are very similar and overlap. The highest figure of merit is

therefore achieved in the widest window. Water samples were tested as a mixture of 8 mL H20 and 12 mL scintillation cocktail, QuickszintTM 400 from Zinsser and Optiphase HisafeTM 3 from Pharmacia-Wallac. In a low-level radiation environment the background count rate in a full window was 2.8 and 2.9 cpm respectively, at 27 and 22% counting efficiencies, using PSA at maximum effect and PAC at 255. In a normal environment the increase is marginal, on the order of 0.5 cpm, leading to background count rates less than 3 cpm at efficiencies up to 26% for aqueous solutions.

COUNTING PERFORMANCE USING GLASS VIALS

499

In Teflon vials the best reported performance gave a background of 0.42 cpm at 27.90 with Quickszint 400.' In a normal environmental gamma flux the background would be below 0.8 cpm. In plastic vials the background is marginally higher (0.05 cpm). Modern biogradeable cocktails do not cause problems with Teflon and plastic vials, therefore there is no need to run low background water samples in glass vials; plastic vials are equally cheap and give better performance. Counting in Teflon and plastic vials in the Quantulus

allows direct measurement of most environmental 3H samples without enrichment. DISCUSSION AND CONCLUSIONS

The use of standard low-40K glass vials is of merit, and it is recommended for low-level '4C determinations when maximum resolution of weak signals (close to background) is not essential. This is the case with close to Modern samples, while for very old samples, Teflon retains its merit due to very high counting

efficiencies and ultralow background count rates. Application of PAC and PSA at high bias does not improve the performance of Teflon vials. In glass vials background reductions are achieved at the expense of some 10 to 20'o loss in counting efficiency.

Optimum performance is not critical in all applications. Moreover, it depends on environment, vial, sample purity (lack of quench), cocktail selection, and concentration. The inherent stability of the Quantulus ensures reproducible performance under optimal counting parameter settings. Performance in the 3H energy region is affected by the residual radioactivity of the Iow-40K glass, hence true, low background counting is not possible. Plastic and Teflon vials make use of the full power of low-level liquid scintillation spectrometers. ACKNOWLEDGEMENTS

Hannu Kojola, Wallac Oy, Turku, and Henry Polach, ANU, Canberra, critically read this text. Their comments were appreciated. REFERENCES

Pearson, G. University of Belfast (personal communication). Polach, H.A., J. Gower, H. Kojola, and A. Heinonen. "An Ideal Vial and Cocktail for Low-Level Scintillation Counting," in Proceedings, of the International Conference on Advances in Scintillation Counting (Banff, Canada: University of Alberta Press, 1983), pp. 508-525. Hogg, A., H. Polach, S. Robertson, and J. Noakes. "Applications of High Purity Synthetic Quartz Vials to Liquid Scintillation Low-Level '4C Counting of Benzene," paper presented at this conference.

500

LiQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Devine, J.M., R.M. Kahn, and A. Long. "Performance of Small Quartz Vials in a Low-Level, High Resolution Liquid Scintillation Spectrometer," paper presented at this conference. Haas, H. "Low Level Scintillation Counting with a Wallac Quantulus: Established Optimal Parameter Settings," paper presented at this conference. Kaihola, L. and T. Oikari. "Some Factors Affecting Alpha Particle Detection in Liquid Scintillation Spectrometry," paper presented at this conference.

Polach, H.A., J. Nurmi, H. Kojola, and E. Soini. "Electronic Optimization of Scintillation Counters for Detection of Low-Level 3H and '4C,"in Proceedings, of the International Conference on Advances in Scintillation Counting (Banff, Canada: Univesity of Alberta Press, 1983), pp. 420-441. Laney, B.H. "Electronic Rejection of Optical Crosstalk in a Twin Phototube Scintillation Counter," in Organic Scintillator and Liquid Scintillation Counting, (New York: Academic Press, Inc., 1971), pp. 991-1003.

Soini, E. "Rejection of Optical Cross-Talk in Photomultiplier Tubes in Liquid Scintillation Counters," Wallac Report, (Turku, Finland: Wallac Oy, 1975), p. 9. Oikarii, T., H. Kojola, J. Nurmi, and L. Kaihola. "Simultaneous Counting of Low Alpha-and Beta-Particle Activities with Liquid-Scintillation Spectrometry and Pulse-Shape Analysis," AppI. Radiat. Isot. 38(l0):875-878 (1987). Cook, G.T., D.D. Harkness, and R. Anderson. "Performance of the Packard 2000 CA/LL and 2250 CA/XL Liquid Scintillation Counters for '4C Dating," paper presented at the 13th International Conference on Radiocarbon Dating, Dubrovnik, Yugoslavia, June 20-25, 1988. Harkness, D.D. and B. Miller. "Recent Developments in LS Counting: Too far, too fast," paper presented at the International Workshop on Inter-Comparison of '4C Laboratories, East Kilbride, Scotland, September 12-15, 1989.

Polach, H., G. Calf, D. Harkness, A. Hogg, L. Kaihola, and S. Robertson. "Performance of New Technology Liquid Scintillation Counters for '4C Dating," Nuc. Geophysics, 2(2):75-79 (1988). Schönhofer, F. and E. Henrich. "Trace Analysis of Radionuclides by Liquid Scmtillaticn Counting," Report UBA-STS-85-02, Vienna, 1985, 25 pp.

CHAPTER 44

Time-Resolved Liquid Scintillation Counting

Norbert Roessler, Robert J. Valenta, and Stat van Cauter

ABSTRACT A comparison is made between standard, two-tube coincidence liquid scintillation counting and the newly developed technique of time resolved-LSC (TR-LSC). In conventional LSC, coincidence requirements are fulfilled only by the fluorescence from excited primary singlets. Any delayed component that results from triplet-triplet annihilation is ignored by the circuit. TRLSC, however, makes use of slow decaying pulse components. Certain scintillators, such as calcium fluoride, and scintillating glasses have much longer fluorescence decay times. Others are characterized by a prompt scintillation pulse followed by after pulses. These different pulse characteristics are analyzed by TR-LSC and used as discrimination criteria to validate scintillation counts. TR-LSC recognizes and discriminates background in low-level LS counting. It uses the long pulse duration due to the interaction of cosmic rays with the glass of the photomultiplier tubes, the scintillation vials,and any other material surrounding the sample. This results in a two to fourfold improvement of the figure of merit (E2IB) as compared to conventional LSC. The issues of cocktail composition and quench correction are addressed. TR-LSC can also be used in dedicated liquid scintillation counters. The long pulse duration is used as a criterion to accept counts while discriminating pulses from thermionic emission. This allows single photomultiplier detection with efficiencies and backgrounds comparable to twotube coincidence counting. The effect of cocktail and sample composition on performance will be correlated with lifetime data.

INTRODUCTION

Two-tube coincidence counting has remained the state of the art in liquid scintillation counter design. Instruments and cocktails were optimized for this configuration because it gives the best overall counting performance. Cock-

tails using scintillators with short lifetimes allow for maximum tube noise reduction using coincidence circuits with short coincidence resolving times. High efficiency is maintained because short lifetime scintillators release all of their energy within a few nanoseconds and give the maximum possible pulse height at the photomultiplier anode. When cocktails containing long lifetime scintillators are used, the photons emitted during the tail end of the scintillation pulse arrive too late and are ignored by the coincidence counting circuit. 501

502

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

This leads to an appreciable loss in efficiency. In specialized applications, such as radioactivity flow monitors based on heterogeneous counting, solid scintillators (e.g., calcium fluoride) and doped glasses are used. These have much longer luminescence decay times than dissolved organic fluors. Consequently, coincidence resolving times in the microsecond range must be used to obtain

an acceptable efficiency. This in turn increases the background due to tube noise significantly. Time resolved LSC' makes use of the fact that scintillators do not give off their light energy instantaneously. TIME RESOLVED COUNTING

The scintillation pulse originating from a scintillator is a burst of photons lasting from a few nanoseconds to several hundred microseconds. In Figure 1 the photoluminescence decay curve of a typical cocktail is shown in schematic

form. This is not the pulse shape of a single decay, rather it is a function describing the probability of a photon being emitted as a result of that decay. Immediately after the decay, we have the prompt pulse or fast component. In

liquids this time lasts from 2 to 8 nsec and represents the fact that most scintillation energy is emitted as direct fluorescence from excited singlet states of the secondary scintillator. The delayed pulse or slow component is attributed to the delayed fluorescence emission from a process called triplet-triplet annihilation. In this process, two scintillator molecules in the electronically excited triplet state collide to form one excited singlet state from which fluorescence occurs. In Figure 2, the average pulse shape due to a high energy decay is shown for a conventional scintillation cocktail and a solid scintillating particle. The relative amount of light originating from prompt singlet emission and the delayed

/

PROMPT PULSE OR FAST COMPONENT

PULSE HEIGHT

DELAYED PULSE ON SLOW COMPONENT

2ns

UP TO 900 ne

Figure 1. Typical cocktail photoluminescence decay curve.

503

TIME-RESOLVED LIQUID SCINTILLATION COUNTING

Conventiorial Scintil1atin Coc1tail Prónpt Pulse

5 nanoseconds (.

20., n.ascPfld!

.

SPA Solid scintiflator Prompt Puse

5nanoseonds

De.Iayed...Piilse....).. $0...nanoeconds

Figure 2.

Averaged pulse shape due to high energy decay: conventional cocktail (top), solid scintillator.

component depends on the specific ionization of the sample. Since beta particles, cosmic rays, and Compton electrons have lower specific ionization than alpha particles, the concentration of triplets formed in the track is lower. As a result, triplet-triplet annihilation is less likely to occur and the delayed component is diminished. This is the basis for alpha particle discrimination using pulse height analysis. Figure 3 is a schematic which emphasizes the fact that a typical afterpulse

pattern consists of individual photoelectrons. The prompt pulse is usually much larger than the afterpulses since it contains a large number of photons emitted in a time too short for the dynode circuit of the photomultiplier to resolve. The photons causing afterpulses, due to the slow component, are

504

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Figure 3. Typical pulse pattern due to beta decay.

fewer in number and spread out over time so as to be distinguishable as single photoelectrons. The triplet concentration in the track does not depend solely on the specific ionization; however, it also depends on the composition of the cocktail. The

concentration of oxygen in a cocktail exposed to air, for instance, is high enough to scavenge most triplets and practically eliminate afterpulses due to the delayed component. In deoxygenated solutions, triplets survive to recombine and cause afterpulsing. Solid scintillators, such as glass, exhibit significant afterpulsing for a different reason. Oxygen quenching and other diffusional mechanisms do not occur in solids so the triplet states formed from the nuclear decay are not quenched, but they can emit light with their characteristically long lifetime. This results in afterpulsing due to an "unquenchable" delayed component. LOW-LEVEL COUNTING

Time resolved techniques can be used in low-level counting to recognize background pulses from the natural radioactivity of the glass vial and the envelope of the PMT. The most likely number of afterpulses for a given energy is greater for glass than for liquid scintillation cocktails; thus, when a cosmic particle passes through the system, it can be discriminated against by counting the number of afterpulses - the burst counting technique. In this technique, each coincidence opens a burst counting window which counts the number of afterpulses occurring to about five microseconds after the event. The total number of afterpulses is defined as the pulse index. Using the pulse index it is possible to create a 3-D spectral plot containing time resolved information on the delayed component (see Figure 4). We see that a background sample gives of an appreciable number of afterpulses at the low energy end of the spectrum. An unquenched 3H sample (see Figure 5) gives

TIME-RESOLVED LIQUID SCINTILLATION COUNTING

505

y

SIXTEEN SPECTRAL PLANES CONSTITUTE THE TOTAL BACKGROUND SPECTRUM. pulse height

counts pulse Index

x

Figure 4.

Three-dimensional plot of pulse height spectrum of background sample. The pulse index is the third dimension.

off few afterpulses and only at the high energy end of the spectrum. An air quenched 3H sample (see Figure 6) gives off almost no afterpulses. By accepting only counts with a low pulse index, a spectrum free from glass

pulse height counts pulse Index

Figure 5. Three-dimensional spectrum of an unquenched 3H sample.

506

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

TRITIUM.

pulse height counts pulse index

x

Figure 6.

Three-dimensional spectrum of an air quenched 3H Sample.

scintillation counts is obtained. This background reduction comes at the price of a reduction in efficiency, since there is a significant probability that sample events will also result in a pulse index, leading to rejection of the count. This is especially true for deoxygenated samples and long lifetime scintillators which have a significant delayed component. Nevertheless, the effect of using the pulse index on low-level performance can be dramatic (see Table 1) because a large part of the background is due to glass scintillations. When air quenched samples are analyzed, the figure of merit (E2/B) can be improved significantly, because the reduction of efficiency due to the time resolved circuit is overshadowed by the background reduction. The data in Table 1 contains the results of a 14C benzene sample prepared with a benzene synthesizer. A small glass vial was used to analyze 3.5 mL sample volume. On optimization, the counting region for this sample was Table 1. Effect of Pulse Index Discrimination on Background, Efficiency, and E2/B for 14C-Benzene

Degree of Pulse Index Discrimination None Normal High sensitivity Low level

14C

Background

E2/B

Efficieny (%)

(CPM)

(Figure of Merit)

83.45 81.87 78.50 70.70

9.67 7.07 4.74

720.15 948.05 1300.05 3560.00

1.38

Note: 3.5 mL Benzene with 4 g/L in small glass vial, 02 quenched, 10 to 100 key.

TIME-RESOLVED LIQUID SCINTILLATION COUNTING

507

Table 2. Effect of Pulse Index Discrimination on Background, Efficiency and E2/B for 3H Degree of Pulse Index 3H Background E2IB Discrimination (CPM) (Figure of Merit) Efficieny (%) 26.50 26.24 24.68 22.59

None Normal

High sensitivity Low level

18.45 12.75 9.25 3.33

38.06 54.08 65.85 153.25

Note: 10 mL InstaGel and 10 mL H20 in large glass vial, 0.5 to 5.0 keV, °2 quenched.

found to be 10 to 102 keY. As pulse discrimination is applied, the efficiency drops from 84% to 70'o, but the background drops from 9.67 to 1.38 counts. This results in an almost fivefold increase in figure of merit. Similar results were obtained on a tritiated water sample (see Table 2). The sample consisted of 10 mL tritiated water mixed with 10 mL of Insta-Gel in a large glass vial. The optimal counting region was determined to be 0.5 to 5 keY. While the efficiency is reduced only 15%, the background is reduced to 18% of its original value. This results in a fourfold increase in the figure of merit. ANTI-COINCIDENCE GUARD SHIELDING2

The time resolved technique can be used to implement anticoincidence guard

shielding in a low-level counter. For this application the normal reflector is replaced by a guard shield that has a scintillating plastic with a long lifetime. For small volume samples in small glass vials, a special adapter fitting into large vial cassettes can fulfill the same function. When cosmic or environmental radiation excite the slow fluor in the plastic, they produce afterpulses that lead to a high pulse index and a rejection of the count. In Table 3, data showing the effect of using various degrees of pulse rejection are displayed for a '4C benzene sample in a small vial. The scintillator used was 6 g/L PPO and 0.2 g/L POPOP. We see that the time resolved counting

coupled with the slow fluor coincidence shielding raises the figure of merit from 1167 to 9520. The data for tritium (see Table 4) also shows an improvement, from 272 to 1745. Here, however, the combination of vial holder with guard shield does not give a significant advantage over either used alone. Table 3. '4C Benzene in PPO (6 gIL) and POPOP (0.2 g/L) Degree of Pulse Index Discrimination None Maximum Maximum Maximum Maximum

Background

E2/B

Sample

3C Efficiency (%)

(CPM)

(Figure of Merit)

Sample only Sample only + vial holder + guard elevator + vial holder & guard detector

64.99 54.05 60.99 66.20 63.98

3.62 0.76 0.63

1167 3844 5904 8593 9520

0.51

0.43

508

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 4. 3H Benzene in PPO (6 gIL) and POPOP (0.2 gIL)

Degree of Pulse Index Discrimination None Maximum Maximum Maximum Maximum

Sample

Background

E2IB

3H Efficiency (36)

(CPM)

(Figure of Merit)

54.71

10.99 3.05 1.37 0.86 0.79

272

Sample only Sample only

50.94 49.38 38.44 37.13

+ vial holder + guard holder + vial holder & guard detector

851

1778 1718 1745

Table 5 contains results from a study of large volume tritiated water samples in Insta-Gel. The use of the guard detector significantly improves the figure of merit for counting environmental water samples. SINGLE TUBE COUNTING3

Time resolved LSC can also be applied to single tube counting of high activity samples. In this case, the pulse index alone is used as a time resolved coincidence criterion. Only those pulses followed by one or more afterpulses are accepted. The physical basis for the operation of the circuit is shown in Figure 7. We see that the typical pulse due to PMT noise has a short width, determined exclusively by the characteristics of the dynode string. The Ultima Gold pulse, on the other hand, has a more complex shape even with a distinguishable afterpulse. A burst count circuit was modified to test this concept. In two pulse-mode, the circuit counts those events which result in two distinguishable pulses within the coincidence resolving time. In three-pulse mode, a triple coincidence is required. The data in Figure 8 was obtained with this modified burst count circuit.

The fact that efficiency varies with the cocktail used can be attributed to variations in the proportion of the delayed component. The lower efficiency in the three-pulse mode is due to the fact that the emission has decayed by the time the circuit has recovered from the first and second pulse. The samples were also counted in a regular 2000CA to show the performance of the cocktails. Because the geometry of the breadboard is different from the production Table 5. 3H Water Analysis in Large Glass Vial (8 mL sample, 12 mL Pico.Fluor LLT) Degree of Pulse Index Discrimination None

Minimum Maximum Maximum

Sample

Background

E2IB

3H Efficiency (36)

(CPM)

(Figure of Merit)

22.85 25.39 25.63 23.77

6.12 6.63 3.87 2.29

5460 6223 10863

Sample only Sample Only Sample only

+ guard holder

Note: 8 mL sample, 12 mL Pico-Fluor LLT.

15791

TIME-RESOLVED LIQUID SCINTILLATION COUNTING

509

lOmV 5ns



lOmV 5ns

Figure 7. Oscil oscope traces of typical pulses: top) PMT dark noise, bottom) 3H in Ultima Gold

counter, it is difficult to assess the reason for the lower efficiency of the single tube circuit.

CONCLUSION

The results presented above demonstrate that the addition of digital time resolved techniques to liquid scintillation counting has already resulted in technical innovations that advance the state of the art. Further research into this area, combined with a better understanding of the mechanisms involved in creating the delayed component of the scintillation pulse in solid and liquid

phase detectors, promises to widen the field of applications for soft beta counting.

Figure 8.

CPU

HlIC

Coddn ii Fctrrnulatjon

CO

LJ.G..

[PA

Tritium efficiency for different cocktail formulations. LSC: liquid scintillation counter, 2PC: single tube counter with time resolved double coincidence, 3PC: single tube counter with time resolved triple coincidence.

I-0

U

I I

2P(

0

TIME-RESOLVED LIQUID SCINTILLATION COUNTING

511

REFERENCES

Valenta, Robert J. "Reduced Background Scintillation Counting," U.S. Patent 4,651,006 March 17, 1987. Valenta, Robert J. and John E. Noakes. "Improved Liquid Scintillation System with Active Guard Shield," U.S. Patent 4,833,326 S/N 07/167,407, March 14, 1988.

Valenta, Robert J. "Method and Apparatus for Measuring Radioactive Decay," U.S. Patent 4,528,450, July 9, 1985.

CHAPTER 45

Comparison of Various Anticoincident Shields in Liquid Scintillation Counters

Jiang Han-ying, Lu Shao-wan, Zhang Ting-kui, Zhang Wen-xin, and Wang Shu-xian

ABSTRACT In low level liquid scintillation counters, most of the background can be eliminated by an anticoincident detector. Nal(Tl) crystal is well-known as an effective shielding detector in gamma ray spectrometers or in liquid scintillation counters. The plastic scintillator and liquid scintillator have also been used in commercial instruments, but all of these detectors are large and expensive. Four types of anticoincident detector have been tested in our laboratory. As a result of the experiments, G-M tubes, made of stainless steel or glass, are very effective in reducing the background from cosmic rays, in particular for 14C. Background in 14C measuring channel is lower than 1 cpm, with the efficiency over 75% for 5 mL benzene samples. So, G-M tubes provide a cheap, compact, and effective anticoincident shield. Bismuth germanate(BGO) crystal, in pieces as an anticoincident detector, is Only about 2 kg. The effect is as well as Nal(Tl) crystal which weighs about 14 kg. Furthermore, different from Nal(Tl), BGO is not hygroscopic and damagable. The effect of the plastic scintillator as an anticoincident detector is between the effect of solid crystals and G-M tubes.

The background spectrum for all four types of detector and the analysis are given and analyzed.

INTRODUCTION

As is known, an anticoincident shielding can effectively decrease the background of the main detector. J. E. Noakes' first introduced NaI(Tl) crystal as an anticoincident shielding detector (AD) to the low background liquid scintillation counter. Then plastic scintillator and liquid scintillation solution were

introduced to the liquid scintillation counter and good results have been obtained. Since the '80s, counting tubes and multiwire proportional counters have been tested,24 but people have not got the anticipated results. Therefore low background liquid scintillation counters are always associated with being large, heavy, and expensive. A lot of experiments about various AD for liquid scintillation counters have 513

514

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

been completed in our laboratory during the past few years. They are NaI(Tl) crystal, plastic scintillator, bismuth germanate(BGO), and common or special G-M counter. From the results, we can see that even when the common G-M counter is used as an AD, better results can be obtained. G-M counting tubes provide the cheapest AD. The 'y ray efficiency of BGO crystal is higher than NaI(Tl) crystal, so a better result can be expected.

INSTRUMENT AND METHOD

Installations and Instruments of Liquid Scintillation Counters with AD First, thirty-two MC-6 G-M counting tubes surrounding the main detector are arranged as a rectangle (Figure la). The matter shielding is 5-cm-thick lead. Second, an annular multiwire G-M halogen counting tube (Figure 1 b) is specially designed (technology and manufacture completed by Beijing Nuclear Instrument Factory). Different from common commercial counting tubes, the shell is made of stainless steel. Because the product rate is too low and the cost is too much, it will not be used in commercial, low background liquid scintillation counters. The matter shielding is 5- or 10-cm-thick lead. Third, a kind of stainless steel cylindrical G-M counting tube is manufactured with shape and function similar to the common G-M y counting tube. Fifteen counting tubes are arranged between two metal cylinders. The main detector is surrounded by the inner cylinder; the outer cylinder is surrounded by 10-cm-thick lead (Figure lc). Fourth, the main detector is in 5-cm-thick lead annular and encircled by twenty-six common commercial glass counting tubes. Another 5-cm-thick lead annular encircles it (Figure id). The glass is replaced by stainless steel in order to study the effect of 40K in glass on the background. The anticoincident annular of plastic scintillator consists of seven detectors. Each of them includes a 4 76 x 150 mm plastic scintillator cylinder and a GDB-52 PMT. The distance between the center of each detector and the main

detector is 105 mm. The seven detectors are arranged in various types as shown in Figure 2. The anticoincident effect of various arrangements are studied. In some experiments, only a few of the seven plastic scintillator detectors are operated. The NaI(Tl) crystal annular (200mm OD, 80mm I.D., and 150mm in length) is not purified for potassium. Six GDB-52 PMTs are used for collecting the photons from the NaI(Tl) crystal. This NaI(Tl) crystal is used in a DYS-15 low-level liquid scintillation counter.

The BGO AD consists of several pieces of BGO and a pair of GDB-52L PMTs. The weight of the crystals is about 2 kg. A 100 x 100 x 100 mm cubic case is made to contain these crystals. There is a 60mm hollow in the middle of the case for placing the main detector. Please note that 2, 3, and 4 all have a 10-cm-thick lead shielding.

COMPARISON OF VARIOUS ANTICOINCIDENT SHIELDS

515

Pb Pb

Pb

Pb

Pb

Pb

Ph

Ph

(03

(d) Figure 1. Four types of counting tube AD.

Electronics

Most of the electronics in these experiments are the same as the DYS-1. Some experiments are completed with a model DYS-8411 automatic, low level

liquid scintillation counter which has a capacity of twenty-four samples, is controlled by a Zijin II microcomputer, and has a software multichannel analyzer.

Standard Sample and Blank Sample All of these samples are contained in quartz vials. The OD of the vial is about 27.5 mm. Different height vials are made for different volumes of samples. The

516

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

coD

Figure 2.

(b)

(c)

The arrangement of plastic scintillator detectors.

5 mL '4C sample contains 5 mL benzene and 30 mg B-PBD. The 10 mL and 20 mL tritium water samples contain 60o emulsion scintillation solution (7 g PPO + 0.5 g POPOP/l toluene:triton X - 100 = 2:1) and 40o water. 20 mL tritium standard contains 20 mL benzene and 0.1 g PPO + 0.001 g POPOP. Standard

samples are made up by standard solution. Blank samples are free from the radioactivity.

Main detector

The main detector consists of two EMI 9635QB PMTs and a type DYS sample chamber which can hold a 5 to 20 mL sample vial. The PMT is selected from many PMTs. Unfortunately, all of these experiments do not use the same pair of PMTs. E. Experimental condition

Experimental Condition Counts per minute in total energy channel were measured for 20 mL benzene sample, in the operation points for 20 mL 3H benzene sample, 10 mL tritium water sample, and 5 mL '4C sample. When the AD operates, the counting rate

of blank samples is abbreviated as AB. If the AD does not operate, the counting rate of blank samples is abbreviated as CB. The energy spectra of 20 mL blank and standard samples are measured. When there is no sample in the sample chamber, the measured cpm is named as EAC and ECC separately, corresponding to operation or non operation of AD. RESULTS

Table 1 shows the results from G-M counting tube as AD in various types and conditions. It is shown that whatever the G-M counting tube is, commercial or specially made, all of them can effectively decrease the background, particularly for the '4C channel. For 5 mL '4C sample, when the efficiency is

COMPARISON OF VARIOUS ANTICOINCIDENT SHIELDS

517

Table 1. The Effect of AD of Counting Tubes

la Operating State No

The type of AD Annular tube 32 MC-6 tube

Yes

A.E.

E (%)

B (cpm)

E2IB

E (%)

B (cpm)

E2/B

(1)

77.0 81.9

3.81 ± 0.06 4.15 ± 0.16

1560 1616

76.8

1.32 ± 0.03 2.02 ± 0.04

4470 3260

65%

81.1

5l%

Note: 5 cm-thick lead

lb 5 mL 14C Operating State No

Yes

Type

E (%)

B (cpm)

E2IB

E (%)

B (cpm)

E2IB

Figure 1(d) Figure 1(C)

76.5 76.7

3.42 ± 0.18 2.91 ± 0.05

1711

2022

76.5 76.7

0.88 ± 0.09 0.73 ± 0.03

6650 8059

ic Operating State Yes

No

Sample E (%) 20 mL

3H 14C

E2IB

E (%)

B (cpm)

E2IB

(1)

54.3 79.3

4.70 ± 0.07 6.10 ± 0.09 24.89 ± 0.17 3.27 ± 0.05 3.28 ± 0.05

600 1030

53.7 79.3

1100 2450

340 1830

33.4 77.5

2.59 ± 0.05 2.56 ± 0.04 7.83 ± 0.10 1.97 ± 0.04 0.83 ± 0.03

total 10 mL 5 mL

3H20 14C

A.E.

B (cpm)

33.5 77.6

560 7250

45% 58% 68% 40% 75%

Note: (1).A.E.: the efficiency of anticoincident.

more than 75%, the background can decrease below 1 cpm. In the '4C channel, although 5cm-thick lead is used to absorb 40K, which comes from the shell of

counting tubes, the background is still a bit higher, but it is clear that in general, the result is acceptable. The anticoincident efficiency for tritium is approximately 40%. Table 2 shows the results from the plastic scintillator AD in different conditions. When all seven detectors are operated (Table 2[A1]), the anticoincident effect is the best. For 20 mL benzene sample, the anticoincident efficiency for blank samples in the total energy channel is about 75%. For empty chambers it is about 87%. The EAC is 2.08 ± 0.03 cpm and the AB is 5.99 ± 0.07 cpm. When only four detectors are operated (Table 2 fBi]), the EAC is 4.18 cpm and the AB is 8.70 cpm. If the same number of detectors operate and there is a gap in the AD ring (Table 2 [A2 and B2, A3 and B3]), with one is on the top of the main detector, and the other below it, the results illustrate that there is no obvious difference between the two conditions. It is proved that the contribu-

tion of cosmic radiation is isotroic; the larger the gap, the higher the background. The results of NaI(Tl) and BGO crystal detectors are shown in Table 3. In

518

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 2. The Comparison of Resuts in Various Arrangements of 7 Plastic Scintillator Detectors

AD Operation (Yes or No)

A

2

3

1,2,3, 4,5,6, 7

Yes

5.99 ± 0.07

No

2,3,4 5,6,7 1,2,3 6,7

Yes No

23.76 ± 0.14 7.32 ± 0.05 23.75 ± 0.09 8.35 ± 0.07 23.64 ± 0.11

15.06 ± 0.08 2.90 ± 0.03 14.81 ± 0.07 3.96 ± 0.04 14.98 ± 0.08

1,2

Yes No

4.18 ± 0.1 14.87 ± 0.09 2.92 ± 0.05 14.82 ± 0.12 3.81 ± 0.04 15.0 ± 0.07

Yes No

1,2,3 5,6,7 2,3,4 5,6

Yes No Yes No

8.70 ± 0.06 23.78 ± 0.10 7.16 ± 0.05 23.72 ± 0.10 8.50 ± 0.07 23.92 ± 0.11

1,2,3 4,5,6 7

Yes

7.52 ± 0.06

3.08 ± 0.05

No

23.82 ± 0.10

14.78 ± 0.10

6,7 B

2 3

C

Background (cpm)

cpm of Empty Chamber 2.08 ± 0.03

Figure 3

order to make a comparison, the main experiment results of G-M counting tube and plastic scintillator are also shown in Table 3.

Figure 3 shows all spectra of four anticoincident liquid scintillation counters. There are spectra of anticoincident and coincident background, spectra of empty chamber, and spectra of tritium and '4C.

Table 3. Comparison of Four Types of ADs 2OmL3H B Total (cpm) A

Y

6.31

N

31.56 80%

AE

V

6.01

N

24.17 75%

48.2 48.4

AE Y

D

53.4 53.8

50.6 51.4

N

AE

C

(%)

5.92 23.66 75%

Y B

E

N

AE

6.95 23.96 71%

52.8 52.9

B (cpm)

5mL'4C

1OmL3H2O E

B

E2/B

(%)

(cpm)

E2/B

(%)

(cpm)

E2/B

2.16 6.03 64%

1320 480

29.7 30.0

1.23 3.14 61%

720 290

80.0

0.54 4.09 87%

11852 1569

1.83

1400 640

28.5 29.4

1.07 2.38 55%

760 360

74.5 74.8

0.40 2.98 87%

13870 1880

28.5 28.7

1.31

626 335

76.2 76.4

0.55 2.43 77%

10557 2402

655 410

74.6 74.6

0.74 3.39 78%

7520

4.15 56% 1.73

1343

4.13 58%

567

2.32 4.04 42%

1200 690

34.1

34.2

2.46 47% 1.77 2.84 37%

E

Note: A:Nal(T1); B:BGO; C:Plastic scintillator; D:G-M counting tube Y:operating AD; N:not operating AD; AE:anticoincident efficiency.

80.1

B

1641

COMPARISON OF VARIOUS ANTICOINCIDENT SHIELDS

519

(C-14)

CAB.)

(C,B.)

(H-3) (E.C,)

CA.B.

F19. 3<-2)

Figure 3.

(a). The spectra of Table 3a. (b). The spectra of Table 3b. (C-i). The spectra of Table 3c. (c-2). The spectra of Table 3c. (d). The spectra of Table 3d.

DISCUSSION

From Table 3, except for NaI(Tl) crystal, the other three types of anticoincident detectors all have approximately the same CB, about 24 cpm in the total energy channel. That is to say, these ADs, as matter shielding, not evidently decrease background. For NaI(Tl) crystal, this background is more than 30

520

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

cpm. The reason is explained elsewhere.6 Except for G-M counting tubes, the other types of anticoincident detectors all have almost the same anticoincident background in total energy channel, about 6 cpm. For G-M counting tubes,

this background is 7 cpm or so, slightly more than others. From the background spectra, it is shown that meson peak is basically elimited. Although the y efficiency is very low, the main detector is not very sensitive for 7-rays,

and the heavy shielding absorbs most environmental radiation. This is the reason why the G-M counting tube can get better effects. The requirement for electronics is very simple and the cost is less than others. In particular, it can be made longer without the price rising too much. So if large vials are used it can provide the cheapest large anticoincident shielding. We think if the longer G-M

counting tube is combined with the BOO, perhaps the best result can be obtained. The plastic scintillator anticoincident detector gives a better result, particularly for measuring tritium. The advantage is clear when comparing with a GM counting tube, but it is similar to NaI(Tl) crystal, in that more PMTs and complex electronics are necessary and the volume is large. Therefore the matter shielding is larger than others. The NaI(Tl) and BGO crystals all have high absorption for 7-ray and good data are attained if using one of them as an anticoincident detector. The BGO crystal can more effectively absorb 7-ray and cosmic radiation than NaI(Tl), and it is not hygroscopic and damageable. It is expensive, but only 2 kg are used. The real cost is largely lower than NaI(Tl) because NaI(Tl) crystal weighs 14.5 kg; therefore, a more compact and effective AD can be made of BGO.

The disadvantage of BGO is the high cost, therefore large crystals are too expensive to use.

REFERENCE

Noakes, J.E., M.P. Neary, and J.D. Spaulding. "Tritium Measurements with a New Liquid Scintillation Counter," Nuci. Instruments Methods, 109:177 (1973). Iwakura, T. et al. "Behaviour of Tritium in the Environment," IAEA-sm-232:63 (1979).

Rajamae, R., et al. Radiocarbon, 22(2):435 (1980). 1220 Quantulus A multi-parameter low level LSC system from LKB-Wallac. Jiang Hanying et al. "Model DYS Low-Level Liquid Scintillation Counter," in Advances in Scintillation Counting, (Edmonton, Canada: University of Alberta, Edmonton, 1983), pp.478-493. Jiang Hanying and Lu Shao-wan. "The Effect of Altitute on the Background of Low-Level Liquid Scintillation Counter," in An International Conference on New Trends in Liquid Scintillation Counting and Organic Scintillators (1989).

CHAPTER 46

The Effect of Altitude on the Background of Low-Level Liquid Scintillation Counter Jiang Han-ying and Lu Shao-wan

ABSTRACT The background counting rate increases more than 100% in Xining over Beijing for the same low-level liquid scintillation counter with Nal(Tl) anticoincident shield. Another experimental counter with bismuth germanate(BGO) anticoincident shield has been transported to Xining, and a series of experiments have been performed. The source of the increasing background is analyzed and the methods of decreasing the background are provided. The altitude of Xining is 2000 meters higher than Beijing, so the earth surface intensity and the constituent of cosmic rays is different in the two cities. Based on our experiments, it seems that the increasing background rate in Xining is not only due to the increased cosmic ray intensity, but also due to the high atomic number of shielding material, e.g., Nal(Tl), BGO, and lead. It is possible that most of the increasing background comes from the interaction between cosmic rays and shielding material. Appropriate selection of shielding material can weaken the effect of the interaction and decrease the background.

INTRODUCTION

The first low-background liquid scintillation counter with an anticoincident shielding was developed by J.E. Noakes) Following a growing application and

requirement of the radioisotopes, such as '4C and tritium, in the fields of geohydrology, geography, geology, archaeology, and others, many types of low-level liquid scintillation counters2 with various anticoincident shielding

have been rapidly developed. Most of them are used at elevations in the hundreds of meters. It is interesting to find what will happen when an instrument is operated at an elevation of more than one thousand meters. In 1985, a model DYS-3 low-background liquid scintillation counter with a NaI(Tl) crystal as an anticoincident shielding was transported to the Institute of Qinghai

Salt-Lake. We found that the background in Xining increased more than twofold. We completed a series of experiments to find out how environmental radiation affects background. From the experimental results, it is thought that the increasing background in Xining is mainly due to the interaction between 521

522

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

cosmic rays and shielding materials (lead and NaI(Tl)). If the shielding material is appropriately readjusted, the background can be obviously decreased. INSTRUMENTS

The Model DYS-3 automatic low-level liquid scintillation counter was made

at the Institute of Biophysics, Academia Sinica, Beijing. A sketch of the detector is shown in Figure 1. An annular NaI(Tl) crystal(< 200 mm OD, 80 mm ID and 150 mm in length) with a 4) 50 mm is used as an anticoincident detector. The sample is elevated from the changing machinery to the sample chamber through the hollow. The matter shielding is 10-cm-thick lead. The changing of sample, data acquisition, and processing are automatically completed by an APPLE II microcomputer. An experimental installation with bismuth germanate(BGO) crystal is the anticoincident shielding. Several pieces of BGO are in a 100 x 100 x 100 mm cubic case. A 4) 60 mm hollow in the middle of the cuboid is for placing the sample chamber and PMTs and a 4) 30 mm hollow of the case over the sample

Pb

NT1) Pb

Pb S

PMT

Pb

P MT

No.1

Pb

Figure 1. DYS-3 detector made at the Institute of Biophysics, Academia Sinica.

EFFECT OF ALTITUDE ON BACKGROUND OF COUNTER

523

chamber is for changing the sample. The weight of the BGO crystal is about 2 kg. Two windows ( 50 mm) on the surface of the cuboid allow for two PMTs to collect the photons from the BGO (Figure 2). The matter shielding is 10-cm-thick lead. An experimental installation with plastic scintillator is an anticoincident shielding. Most of it is the same as the DYS-1 except for the anticoincident shielding. It consists of seven detectors in annular. Each one has a 76 x 150 mm cylindrical plastic scintillator and a PMT. EXPERIMENT AND RESULT

The blank sample is 20 mL benzene scintillation cocktails (5 g PPO + 0.05 g POPOP/l benzene) in a 20 mL quartz vial. The vial is sealed by 914 adhesive agent. All of the counters are operated at high voltage for tritium measuring. The background counts per minute are measured in total energy channel. If

Pb

Pb

Pb

012 PMT1 Pb

Pb

Pb

Figure 2. BGO experimental installation.

524

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 1. The Comparison of the Background in Beijing and Xinlng BGO Installation

DYS-3

Beijing Xining

CB

AB

EB

EA

CB

AB

EC

EA

30.19 71.47

5.64 13.75

18.07 29.95

1.44 1.75

24.40 55.88

5.39

15.63

1.32

17.11

Note: E.C.: empty sample chamber C.B.; E.A.: empty sample chamber A.B.

the anticoincident detector does not operate, the background measured is named by coincident background (CB). If the anticoincident detector operates, the background measured is named by anticoincident background (AB). When there is no sample in the sample chamber, the counting rate is named by empty chamber background (separately ECB and EAB). The experiments are done separately in Beijing (the elevation is about 100 meters, 400 N) and in Xining (the elevation is about 2300 meters, 36.5° N ). The background and the empty chamber counting rate are measured first in Beijing and then in Xining. The results (Table 1) show the AB of two counters all increase about threefold. The shielding experiments are completed in a BGO installation in Xining.

All of the experimental results are listed in Table 2. We find, if the lead shielding is increased from 10 cm to 20 cm, the background increases. The CB increases from 55.88 cpm to 59.44 cpm and the AB increases from 17.11 cpm to 19.49 cpm. If we append the lead with plastic material (polystyrenee) and water, the CB decreases to 50.37 cpm and AB decreases to 14.20 cpm. The 10 cm lead chamber is enlarged to test the radon effects on the background. There Table 2. The Data of the Shielding Experiments in Two Places Xining

Place

Shielding Condition CB BGO installation A B

C D E F

G H I

Only 10 cm lead under the detector Only 20 cm lead under the detector Entire 10cm lead shielding C + appending 10cm lead on top C + appending 10cm lead around C + many polystyrene pellets fills the empty space of the lead chamber and 30 cm thick polystyrene on top of and 601 water around the lead chamber Enlarged 10 cm lead chamber G + 30 kg. paraffin wax and polystyrene fills the empty of the lead chamber H + 60 L water + 60 kg. paraffin wax around and on the lead chamber DYS-3

10 cm lead shielding The empty of the lead chamber are filled with 30 kg. of paraffin wax

295.5 338.9 55.88 57.07 59.44

Beijing AB

CB

AB

171.7 198.6

300.8

163.8

17.11

24.40

5.39

30.19

5.64

17.68 19.49

50.37 54.90

14.20 17.93

44.77

10.57

44.65

9.67

71.47

13.75

58.00

9.41

EFFECT OF ALTITUDE ON BACKGROUND OF COUNTER

Table

3.

525

The Estimated Intensity of the Cosmic Rays at 50 N 2000 meters Sea Level Elevation 800 g/cm2 Atmosphere

Total intensity (particles /sec.cm2, in all directions Hard constituent (particles /sec.cm2, in all directions) Soft constituent (particles /sec.cm2, in all directions) E:electron and positron with energy more than 10 MeV :ji mesons with various energy P: proton with energy more than 400 MeV

0.020

0.035

0.013

0.015

0.007

0.017

x1031cm2.sec.D

6x103/cm2.sec.0

8x 103/cm2.sec.tI

1.5x 102/cm2.sec.0

3x 105/cm2.sec.D

2x 104/cm2.sec.0

1

are 10 cm empty spaces above and under the detector. The background obviously does not change, but if the empty spaces of the lead chamber are filled by 30 kg paraffin wax, the background decreases. The CB decreases from 55.88 cpm to 44.77 cpm; the AB decreases from 17.11 cpm to 10.57 cpm. If the empty space of the lead chamber of DYS-3 is filled with paraffin wax, the CB of DYS-3 decreases from 71.47 cpm to 58.00 cpm; the AB of DYS-3 decreases from 13.75 cpm to 9.41 cpm. DISCUSSION

Comparing the empty chamber background in two places (Table 1), we find the CB of the empty chamber of DYS-3 in Xining is 1.66 times higher than that

in Beijing, but the AB of the empty chamber in Xining is only 1.22 times higher than in Beijing. Table 3 shows that the total intensity of various energy i mesons is 1.88 times higher at the elevation with 800 g/cm atmosphere (or elevation of 2000 meters or so) than at sea level. According to our past experiments, most contributions of the mesons to the background can be elimi-

nated by any type of anticoincident detector; thus it is suggested that the background of the empty chamber in coincident counters at sea level is mainly from the mesons of cosmic rays. The increasing background of the blank sample in Xining mainly does not come from the increasing mesons. In Beijing, most of the low-level liquid scintillation counters with different anticoincident detectors all have a CB value of about 24 cpm. The counter with NaI(Tl) crystal as an anticoincident detector, however, has a CB of about 30 cpm; it is 6 cpm higher than the others. In the past we thought this came from the K in NaI(Tl). Now, from Table 2, under the same experimental condition,

the CB of DYS-3 is 15.6 cpm higher than the CB of BGO installation in Xining. Obviously, it can not be assigned the 40K of the NaI(Tl). Maybe it comes from the interaction between NaI(Tl) crystal and some high energy

526

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 4. The Anticoincident Effect of Two Detectors for Neutron Coincident Anticoincident Counting Countingrate Rate for for Neutron Neutron (cpm) (cpm) 214.2 ± 1.5 BGO installation 158.1 ± 1.3 Plastic scintillator installation

186.9 ± 2.7

81.6 ± 1.4

The Anticoincident Effect for Neutron 19.5%

53.3%

particles in cosmic radiation; new radiation is created, and it increases the background. In Xining, the greater the lead thickness, the greater the background. In Beijing, the background decreases if the lead thickens.2 When the lead is only placed under the detector and there is no lead above and around the detector,

the background is obviously no different in two places, but the difference rapidly increases as the thickness of the lead increases. From this, we think the

increasing background is partly from the interaction between the lead and high-energy cosmic radiation. When a particular cosmic ray shoots the lead, it produces heavy radiation. If the empty space between the detector and lead is filled with paraffin wax, the background decreases. We feel the interaction between cosmic radiation and lead produces a nuclear reaction, and a lot of neutrons are produced. The difference between CB and AB, after and before being filled with paraffin wax for DYS-3, is respectively 13.47 cpm and 4.34 cpm. It means the effect of anticoindence of NaI(Tl) for this part of the background is about 68%. For

BGO installation, the difference of the CB and AB after and before being filled with paraffin wax is 10.13 cpm and 7.37 cpm, respectively; the effect of anticoincidence is only about 27.2%. In Beijing, 5 x 105/sec Am-Be neutron (in the state for transportation) is placed out of the anticoincident counters of BGO and plastic scintillator, and the counting rate is measured with the same condition as above. The results are shown in Table 4. From Table 4, the anticoincident effect for BGO and plastic scintillator is 20% or so separately and more than 50% combined. We think it probably depends on the volume, the atomic number and geometric condition of the anticoincident detector, and also the energy spectrum of the neutron and the interaction between neutron and matter. According to the experimental results, the increasing background can be decreased by paraffin wax, and there is a difference of anticoincident effect between two counters in Xining. We consider that the increasing background is due to the neutrons produced by the lead interacting with the cosmic radiation. The other part of the increasing background for DYS-3 is difficult

to decrease. It comes from the interaction between the NaI(Tl) and cosmic radiation. To sum up, plastic scintillators have better anticoincident effect. When interacting with cosmic radiation, it does not produce much second radiation; it does not lead to increased background for liquid scintillation counters as

EFFECT OF ALTITUDE ON BACKGROUND OF COUNTER

527

NaI(Tl) crystal does. In a high elevation area, as an anticoincident shielding, it is better than NaI(Tl). A complex matter shielding of lead and paraffin wax for a liquid scintillation counter is necessary.

REFERENCES

Noakes, J.E., M.P. Neary, and J.D. Spaulding, J.D. "Tritium Measurement with a New Liquid Scintillation Counter," Nuc. Instrum. Methods, 109:177 (1973). Yang Shouli, Jiang Peidong, and Lin Han. "The Advance and Application in Liquid Scintillation Measurement Technique," p.88 (1987). Jiang Hanying et al. "Model DYS Low-Level Liquid Scintillation Counter," in Advance in Scintillation Counting (S.A. McQuarrie, et al.: (Ed. Edmonton, Canada: University of Alberta, 1983).

CHAPTER 47

Calculational Method for the Resolution of 90Sr and 89Sr Counts from Cerenkov

and Liquid Scintillation Counting*

Thomas L. Rucker

INTRODUCTION

Environmental samples taken around reactor facilities potentially contain both 50.4 day half-life strontium-89 (89Sr) and 29 year half-life strontium-90 (90Sr). The health risk associated with exposure to 9°Sr is considered to be a factor of 20 greater than that associated with 89Sr due to its longer half-life and its energetic daughter yttrium-90 (90Y). It is therefore often important to distinguish between these two isotopes when monitoring for health and environment protection. Most common methods for determining both 89Sr and 90Sr are based on time allowed ingrowth of 90Y after the total strontium is chemically purified and counted. This usually requires from 7 to 14 days to allow sufficient 90y ingrowth for adequate statistics. A method has been developed that determines both 89Sr and 90Sr without requiring an ingrowth period for 90Y; therefore, the results can be obtained much sooner. The method is similar to that developed more than a decade ago by Randolph' which uses sequential Cerenkov and liquid scintillation counting after chemical purification. A new calculational method has been employed, however, which corrects for the fraction of the Cerenkov counts due to 90Sr and the fraction of both counts due to the early ingrowth of 90Y. The method does not rely on the spectral resolution of 89Sr, 90Sr, and 90Y through the use of channel ratios, as has previously been required with the combined liquid scintillation of a Cerenkov technique. Therefore, higher detection efficiencies are obtained which result in better sensitivity.

*Research performed under the auspices of the U.S. Department of Energy under Contract DE-ACO9-76SR00001 with F. I. DuPont De Nemours & Co. and Subcontract AX-7 15305 with Science Applications International Corporation. 529

530

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 1. Nucllde Characteristics Nuclide

89Sr

90Sr

Half-life

50 d

Max. beta energy (key)

1491

29 V 546

2284

Avg. beta energy (key)

583

196

934

1

20

-

80

8

Rel. health risk M.P.C. (pCIIL)

64 h

drinking water

THEORY

Cerenkov radiation is produced when a charged particle passes through a

transparent medium at a velocity greater than the speed of light in that medium. A threshold condition must be met which is defined by

vn/c = 1,

(1)

where v is the velocity of the particle, c is the speed of light, and n is the refractive index of the transparent medium. Furthermore, the Cerenkov photon yield increases proportionally as the energy of the particle increases above the threshold. The nuclear decay characteristics for 89Sr, 90Sr, and 90Y are shown in Table 1.

The combined Cerenkov and liquid scintillation counting technique takes advantage of the large difference in average beta energies of 89Sr and 90Sr to discriminate between the two after the 90Y is separated chemically. The threshold energy for the beta particle stimulation of Cerenkov radiation in water is 263 keY; therefore, in a modern liquid scintillation counter, Cerenkov counting provides over 40% efficiency for the 583 keV average beta energy 89Sr, while providing less than 1 'o counting efficiency for the 196 keY average beta energy 90Sr. The total strontium activity is determined by liquid scintillation counting, which provides over 90% counting efficiency for both isotopes. The combined information obtained from these two counts yields the activity of each isotope. OPERATIONAL METHOD

Chemical Separation A standard procedure, similar to that used in EPA Method 905.0,2 is used to chemically purify the sample. Calcium and magnesium are removed by precip-

itation of the strontium as the nitrate. Barium and radium are removed by a

chromate precipitation. Yttrium and other impurities are removed with hydroxide scavenging. A final strontium precipitation is made as the carbonate

STRONTIUM-90, STRONTIUM-89 COUNTS USING CALCULATION METHOD

Table 2. Experimental Efficiencies and Backgrounds Cerenkov

531

Eff. (%)

Liquid Scint Eff. (%)

89SI 90Sr

40

95

1

91

90Y

61

95

(cpm)

(cpm)

15

62

Nuclide

Background

in a tared liquid scintillation vial. The precipitate is washed, dried, and weighed to calculate the recovery of the standardized strontium carrier. The average chemical recovery is approximately 80%. Counting The precipitate is dissolved in 5 mL of 0.3 M HC1, and the aqueous solution is counted on a liquid scintillation counter without scintillator for 20 mm for Cerenkov radiation, primarily from 89Sr. Seventeen mL of "Atomlight" liquid scintillator is then added and mixed with the sample. The solution is again counted for 20 mm on the liquid scintillation counter for the measurement of both 895r and 90Sr.

Randolph used optimized discriminator channel settings to provide energy

windows for each count. He then used the count ratios in each widow to calculate factors for correcting the fraction of the Cerenkov counts due to 90Sr and the fraction of both counts due to the early ingrowth of 9°Y. However, due to the severe energy spectrum overlap of each of the isotopes, the optimized windows only contain a fraction of the total counts from the isotope of interest. Therefore, the sensitivity is reduced. In the present experiment, the novel calculational method is used to correct for the fraction of the Cerenkov counts due to 90Sr and the fraction of both counts due to the early ingrowth of 90Y. The discriminators were set for each count such that the energy window contained the entire peak of the isotope or isotopes of interest, thus improving the sensitivity of the method. Beckman LS-9800 liquid scintillation counters were used in this study. The experimental efficiencies for each isotope and the backgrounds obtained for each counting method are shown in Table 2. CALCULATIONAL METHOD

The Cerenkov counts are primarily from 89Sr, with a small contribution from 90Sr and the early ingrowth of 90Y. The net Cerenkov count rate (cpm), C1, is given by

C1 = 89Sr cpm + 905r cpm + 90Y cpm,,

(2)

532

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

where cpm represents the Cerenkov counts per minute obtained from each nuclide. The liquid scintillation counts are due primarily to both 89Sr and 90Sr, with a small contribution from the early ingrowth of 90Y. The net liquid scintillation count rate (cpm), C2, is given by C2 = 89Sr cpm1 + 90Sr cpm1 + 90Y cpm1,

(3)

where cpm represents the liquid scintillation counts per minute obtained from each nuclide. The count rate due to early ingrowth of 90Y can be defined in terms of the 90Sr count rate by knowing the elapsed time from chemical separation of 90Y to counting. Equations 2 and 3 then become C1

= 89Sr cpm + ('°Sr cpm) (A1)

(4)

and C, = 9Sr cpm

(90Sr cpm) (A2),

(5)

where A1 =

1 + (D1)(l - etc)

(6)

A2 =

1 + (D,)(1 - e'),

(7)

and

where D1 is the 9°Y Cerenkov efficiency to 90Sr Cerenkov efficiency ratio, t is the time from chemical separation to Cerenkov counting, D2 is the 90Y liquid scintillation efficiency to 90Sr liquid scintillation efficiency ratio, t1 is the time from chemical separation to liquid scintillation counting, and X is the 90Y decay constant.

The Cerenkov and liquid scintillation count rates for each nuclide are related by their counting efficiency ratio of each for the two counting methods such that 89Sr cpm = (89Sr cpm) (B1)

(8)

90Sr cpm = (90Sr cpm) (B2),

(9)

and

where B1 is the 89Sr Cerenkov efficiency to 89Sr liquid scintillation efficiency ratio and B2 is the 90Sr Cerenkov efficiency to 90Sr liquid scintillation efficiency ratio. By replacing these terms in Equation 4, C1 =

(8Sr cpm) (B1) + (90Sr cpm1) (B2) (A1).

(10)

If the ratios of counting efficiencies are determined by counting pure standards, Equations 5 and 10 contain two unknowns and can be solved simultaneously to yield

STRONTIUM-90, STRONTIUM-89 COUNTS USING CALCULATION METHOD

Table 3. Results from Measuring Mixtures of Isotopes (Lower Level) 2 Sigma Measured Known Sample! CL (pCUL) (pCi!L) Nuclide 1 89Sr

190Sr

120 0

130

0.7

+1+1-

11

5

533

Ratio

-

1.08

289Sr 290Sr

120 220

125

234

+1+1-

15 11

1.04 1.06

389Sr 390Sr

240

219 259

+1+1-

16 13

0.91 1.18

120

130 491

+1+1-

19 16

1.08

440

+1+1-

12

489Sr 490Sr

589Sr 590Sr

220

0

8

220

260

10

1.11

-

1.18

90Sr cpm1 = [(B1)(C2) - C1] / [(B1)(A2) - (B2)(A1)]

(11)

89Sr cpm1 = C2 - (A2)(90Sr cpm1).

(12)

and

The activity concentration of each nuclide can then be determined by Equations 11 and 12 and by 90Sr (pCi/L) = (90Sr cpm1) / [(E1)(V)(R)(2.22)]

(13)

89Sr (pCi/L) = (89Sr cpm1) / [(E2)(V)(R)(2.22)]

(14)

and

where

E1 = 90Sr liquid scintillation efficiency, E2 = 89Sr liquid scintillation efficiency,

V = Sample Volume (L), and R = Chemical Recovery. RESULTS

Known mixtures of 89Sr and 90Sr, ranging to 100% of each, have been measured by this technique at both low and moderate activity levels. The results are shown in Tables 3 and 4. While it appears from the lower level data that a slight positive bias may exist for 90Sr, it is likely that the apparent bias is due to counting uncertainty since the bias disappears at the higher activity level. The error due to uncertainty in correction for the early ingrowth of 90Y is minimized by counting the samples as soon as possible after separation. Since this method requires two counts, the uncertainty must be considered when estimating the total measurement due to counting statistics. An estimate of the standard deviation, s, due to counting uncertainty can be made by

534

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

Table 4. Results from Measuring Mixtures of Isotopes (Higher Level) Sample/ Known Measured 2 Sigma Nuclide (pCi/L) (pCIIL) CL 2042

690Sr

2044 0

789Sr 790Sr

2044 2199

2041

8 89Sr

6 89Sr

890Sr 989Sr

9 °Sr

Ratio

-

+ I+1-

47

2045

+1+1-

53 44

1.00 0.93

4088 2199

3938 2145

+1+1-

71

60

0.98 0.98

2044 4398

2074

+1+1-

66 52

0.99

+1+1-

36 26

0.99

2

4371

1089Sr

0

38

1O90Sr

2199

2177

41

1.00

1.01

-

90Sr cpm, s = {[(B1)2(G2/T2 + F2/T2) + G1/T1 + F1/T1] /[(B1)(A2) - (B2)(A1)]2} 1/2

(15)

and 89Sr cpm, S = [(G2/T2 + F2/T2) + (A2)2(90Sr cpm, where

s)2]"2,

(16)

G1 = gross Cerenkov count rate (cpm), G2 = gross liquid scintillation count rate (cpm), F1 = background Cerenkov count rate (cpm), F2 = background liquid scintillation count rate (cpm), T1 = Cerenkov counting time (mm), and T2 = liquid scintillation counting time (mm)

when the background and sample counting times are equal. Similarly, the lower limit of detection (LLD), considering only counting statistics, can be estimated by

LLD (cpm) = 2.71/T2 + 4.65(sb)

(17)

where sb is the standard deviation of the blank as defined by 90Sr sb = {[(B1)2(F2/T2) + F1/T1] / [(B1)(A2) - (B2)(A1)]2}"2

(18)

and 89Sr sb = [F2/T2 + (A2)2(90Sr

1/2

(19)

The liquid scintillation counters have a higher background than the lowbackground gas-proportional counters typically used for counting environmental strontium samples. This results in reduced sensitivity. However, the higher efficiencies obtained with liquid scintillation counters allow lower limits of detection which are only a factor of 2 to 3 higher than those obtained using

STRONTIUM-go, STRONTIUM-89 COUNTS USING CALCULATION METHOD

535

Table 5. Calculated Lower Limits of Detection

Cerenkov and Liquid Scintillation Count 9°Sr = 12.8 cpm = 7.9 pCi/L 89Sr 15.9 cpm = 9.4 pCi/L Traditional 14-Day Two-Count Method

90Sr = 1.4 cpm = 2.4 pCi/L 89Sr = 2.0 cpm = 3.6 pCi/L Note: All LLDs assume 20 mm Counts on 1 L samples and 80% chemical recovery.

the same count time on gas-proportional counters. The lower limits of detection obtained in this study are shown in Table 5 with the LLDs obtained with the traditional 14-day two-count method (assuming 1-L samples counted for 20 mm and 80 chemical recovery for both). If low-background liquid scintillation counters are used, the lower limits of detection may be even lower. CONCLUSION

The described method provides a useful way to analyze both 89Sr and 90Sr

without a 14 day wait for 90Y ingrowth. The use of the derived equations provides a correction for early 90Y ingrowth and the 90Sr Cerenkov counts, without the use of energy window ratios. Better sensitivity has been provided than has previously been possible with this technique. The sensitivity of this method is adequate for monitoring plant effluents and many environmental level samples. Lower limits of detection, below the maximum contaminant concentration allowed for drinking water, can be achieved when a 1-L sample and a 20 mm count time are used. If monitoring for lower levels of contamination are necessary, a larger sample or a longer count time may be used. REFERENCES

Randolph, R.B. "Determination of Strontium-90 and Strontium-89 by Cerenkov and Liquid-Scintillation Counting," In!. J. Appi. Radiat. Iso., 26:9-16 (1975). Krieger, H.L. and E.L. Whittaker. "Prescribed Procedures for Measurement of Radioactivity in Drinking Water," EPA-600/4-80-032 (1980), pp. 58-74.

CHAPTER 48

Determination of 222Rn and 226Ra in Drinking Water

by Low-Level Liquid Scintillation CountingSurveys in Austria and Arizona

Franz Schönhofer, J. Matthew Barnet, and John McKlveen

ABSTRACT One important source to man for both 222Rn and 226Ra is drinking water. 226Ra may be ingested and 222Rn is liberated by household activities; thus adding to the radon concentration already present in indoor air. Very simple yet specific methods, avoiding any chemical separation, are presented for the determination of both radionuclides, using a commercially available ultra low-level liquid scintillation counter. The lower limit of detection is 40 mBq/L (1.1 pCi/L)based on 500 mm count and 3 a of the background. These methods have been applied for surveys both in Vienna and Arizona, the results of which are presented.

HEALTH HAZARDS OF 226RA AND 222RN

Most countries have strict regulations about the maximum permissible concentration (MPC) of 226Ra in drinking water (224Ra is a bone seeker). MPCs

range in different countries from 0.11 to 1.85 Bq/L. In Austria it is 0.122 Bq/L. 222Rn is regarded as hazardous, since it is abundant in all indoor air in enhanced concentrations. It emanates from ground or building material, and it is liberated from water by household activities like cooking, washing, toilet flushing, etc. It is estimated' that tap water containing 100 kBq/m3 gives off an additional effective dose equivalent to about 0.4 to 0.7 mSv/a from the liberated radon and its daughters. An air change rate of 0.5 changes per hour is

assumed. Modern isolation techniques used for energy conservation have altered the air change rate by approximately 0.1 in most cases, enhancing radon concentration in houses. The above mentioned additional dose corresponds to the average additional dose the Austrian population received in the first year after Chernobyl. (Austria was one of the most contaminated European countries outside the U.S.S.R.) 537

538

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

No MPCs for 222Rn are known in any country, but in some countries, for instance Sweden, national health boards give recommendations.2 Below 100 Bq/L no restrictions or counter-measures are considered necessary. Between 100 and 1000 Bq/L cases have to be considered individually with regard to other sources of radon and ventilation. If water contains more than 1000 Bq/L it is likely that the concentration of radon daughters in the air exceeds 400 Bq/ m3 - which is the action level - and remedial measures should be undertaken. It is expected that the U.S. Environmental Protection Agency will set recommendation levels for radon in water soon, and they will probably be lower than the Swedish ones at present. In principle, it is possible to apply the recommen-

dation of 4 pCi/L air (0.15 Bq/L), taking the liberation from water into account. Doses from ingestion of radon are negiglible compared to the lung doses from inhalation. Determination of 222Rn and 226Ra in water is therefore of great importance in health physics.

DETERMINATION OF 222RN AND 226RA IN WATER

222Ra is highly soluble in toluene and other organic solvents frequently used in cocktails for LSC. Since a-particles are counted with approximately 100% efficiency, LSC is nowadays widely used for measurement of radon.3-5 One simple method is to mix water with a gel-forming scintillation cocktail and wait about three hours until equilibrium with the daughters is established. Normally 222Rn is present in such high concentrations that any contribution of other naturally occurring radionuclides is negiglible. Figure 1 shows the spectrum of an actual water sample, 8 mL of which were mixed with 12 mL of Quickszint 400 (obtained from Zinsser Analytic, Frankfurt). The vial used was a PTFE coated polyethylene vial (from Zinsser). The left peak is the unresolved sum of the peaks of 222Rn (5.49 MeV) and 218Po (6.00 MeV), the right, well-resolved one is due to 214Po (7.69 MeV). The resolution is estimated to be about 300 to 400 keV. The x-axis shows the logarithmic pulse height and not the energy. Efficiency is between 300 and 400%, depending on vial and cocktail. It is determined with a 226Ra standard treated the same way and taking the 100% efficiency of the Ra-a-particles into account. Another method uses radon extraction from the water phase into a waterimmiscible, mineral oil-based scintillation cocktail.5 The cocktail employed in this research was PSS-007H from NEN. Ten mL of water is mixed with 10 mL cocktail, shaken vigorously, and counted after about three hours. In this time, equilibrium is established, the sample is cooled, and the phases are separated. The resolution is better (Figure 2) than in the case of the gel-forming cocktail. The efficiency is equally high and is determined with a 226Ra standard. 226Ra is

not extracted into the organic phase. 226Ra is usually enriched and isolated chemically. Measurement is frequently done with LSC.67 We have been able to develop a method which completely

I

* 3$ *

I

i, 11 CI1fch

Figure 1.

S

[Al

?iiEl

[email protected]

3

1

'ETiB.??

'TTi

4

'.

C 11.8 >)

6

L376 CPM

[email protected]ø )

1H'11I1N

tJN]MJ1\Q4241M .

rJmf

@.158 C 11.5

[email protected]

E : \3A i\

655 CPM [B]

i-

in 11

PT1?

SPUIZ

Pulse height spectrum of a water sample containing 222Rn, 8 mL mixed with 12 mL Quick szint 400, and PTFE Coated polyethylene

ffliNCH

JNTEGE( 515- ?2)

Ti

vial.

10'

Figure 2.

1

3

3

C

6Ei

700

T

J

800

\TT THk\RN\ 4\Oft411N

[email protected] 500 [A] 439.823 CPN 11.543 2.6 :-<)

ZEIO

2i/1

900

:P*?2

1060

Pulse height spectrum of a water sample containing 222Rn radon extracted from 10 mL water into 10 mL of NEN cocktail, and PTFE coated polyethylene vial.

iliNCH

JNTE( 670 830)

I

?M1'H CPM/ch

222RN 226RA DETERMINATION IN DRINKING WATER

541

avoids the chemical isolation by using the ultra low-level LS counter "Quantulus" (PharmaciaWallac, Turku, Finland). This instrument exhibits extremely

low background due to heavy passive and active shielding consisting of an anticoincidence unit based on a liquid scintillator. Moreover activity of the selected construction material is very low. Pulse height spectra are recorded automatically and can be used for control of possible interferences. Tests with standards of 226Ra showed that the above mentioned methods for 222Rn (simple mixing of water with the above mentioned cocktails) were well applicable for 226Ra and they gave a lower limit of detection of about 37 to 48 mBq/L (1.0 to 1.3 pCi/L) (based on 3 of the background and 500 mm counting time), which is well below the maximum permissible concentration in Austria of 0.122 Bq/L

(3.3 pCi/L). The mixture of water with the gel-forming cocktail is not recommended for 226Ra, since in such low concentration ranges heavy interferences from other naturally occurring radionuclides, especially 40K, may occur. Quench effects from the water may also cause heavy interference. The extraction with the mineral oil-based cocktail, however, is specific for 226Ra. After ingrowth of 222Rn from 226Ra (or decay of excess and unsupported 222Rn) only 222Rn in equilibrium with 226Ra is extracted. 220Rn and 219Rn would not interfere because

each has a very short half-life. In practice, water was analyzed for 222Rn immediately after it received the sample by the mineral oil cocktail method. Then it was stored for the appropriate time to allow for decay of unsupported 222Rn. The same sample was then measured again for 226Ra, thus the manual labor involved for determination of both radionuclides was simply the pipetting of 10 mL of sample and 10 mL of the cocktailnot to forget the shaking. For 222Rn the measurement time is usually 60 mm. For 226Ra 500 mm is was chosen, which means automatic determination of three samples per day, also during weekends and holidays. The high costs of our instruments have been compensated for by long time savings of personnel and labor. RESULTS

The afore described methods were applied for surveillance purposes in Austria, and the method for 222Rn has been used to survey risk area in Carefree/ Cave Creek Basin in Arizona. In Carefree, 26 wells were tested. The frequency of the concentrations found can be seen in Figure 3. Eleven wells (42%) show concentrations above the 100 Bq/L, which the Swedish authorities regard as the concentration below which no countermeasures need to be considered. They result in an additional effective dose equivalent of 0.4 to 0.7 mSv/a, which is much higher than the impact of nuclear power plants on people living in the surrounding area. Concerning 222Rn in Austria, the results from a risk area in the northeastern part of the country are presented. After a nationwide survey, work was con-

542

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

FREQUENCY

0 LLD - 10

NJ

0

0

// //

10 - 20

20 - 30 30 - 40

// // // //

40 - 50

50 - 60

// //

60 - 70

70 - 80

80 - 90 90 - 100 100 - 150

150-200 200 - 250

/

250 - 300 -

Figure 3.

Distribution of concentrations of 222Rn in water, Carefree/Cave Creek Basin, Arizona.

0

Gn,Und

x

x

0

10km

0

Heidenre(chsfejn

00

00

000

0

500

0200-500

0 100-200

o 50-100

o <50

Sq/I

°X /aidho ten

00

0

00

Cs

00

Roabs 0q

Figure 4. Geographical distribution of 222Rn in water in a risk area in northeastern Austria.

00

0

0 k/eifra"oc

Cs

oQ

00

0

544

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

centrated on an area where the ground consists partly of granite and partly of metamorphicum (mostly from sediments). The regional distribution of 222Rn concentrations in household water is shown in Figure 4. Clearly, small areas with extreme concentrations can be distinguished from areas with rather high concentrations. The former ones coincide well with granite areas and the latter ones with metamorphicum. In spite of the extreme values for 222Rn, even in the risk area, only a few samples exceeded the lower limit of detection of 40 mBq/L for 226Ra. By far, the highest value observed was 490 mBq/L, followed by 180 mBq/L. CONCLUSIONS

The simple methods for both 222Rn and 226Ra determination have made it possible to survey a large number of samples within a short time. The results show that both Arizona and Austria have risk areas for high 222Rn concentrations in water, and analysis has to be extended to other risk areas. The concentrations found are significant health concerns, and after limits are defined, countermeasures have to be undertaken. Regarding the instrumentation, it is attempted to test the a-3-discrimination unit of "Quantulus" thoroughly for its application in 226Ra determination. Since the background for a emitters can be reduced practically to zero, it is expected that a much lower LLD for 226Ra can be achieved. Careful investigations of vials and cocktails will be necessary. ACKNOWLEDGEMENT

The authors thank Christian Tschapka and Jeffrey Brown for performance of measurements and evaluation of data used in this chapter, Robert Mecl, Franz Oesterreicher, Kurt Hierath (water department of the Federal Institute), and Oswald Ruttner (BBSUA, Vienna) for collecting and providing most of the Austrian water samples. The research in Arizona was financially supported by the Arizona Department of Environmental Quality, which is gratefully acknowledged. REFERENCES

Naturally Occurring Radiation in the Nordic Countries - Recommendations, (Reykjavik: The Radiation Protection Institutes in Denmark, Finland, Iceland, Norway, and Sweden, 1986). Kulich, J., H. More, and G.A. Swedjemark. Radon och radium i hush?illsvatten, (Radon and Radium in Household Water), Swedish Radiation Protection Institute, Report SSI-rapport 88-11, 1988 (in Swedish). Parks, N.J. and K.K. Tsuboi. "Emulsion Scintillation Counting of Radium and Radon," Inst. J. App!. Radiat. Isotopes, 29:77 (1978).

222RN, 226RA DETERMINATION IN DRINKING WATER

545

Asikainen, M. and H. Kahios. "State of Disequilibrium Between 238U, 234U, 226Ra,

and 222Rn in Groundwater from Bedrock," Geochimica et Cosmochinica Acta, 45:201 (1981).

Pritchard, H.M. and T.S. Gesell. "Radon-222 in Municipal Water Supplies in the Central United States," Health Phys. 45:991 (1983). Prichard, H.M., T.F. Gesell and C.R. Meyer. "Liquid Scintillation Analyses from Radium-226 and Radon-222 in Potable Waters," in Liquid Scintillation Counting, Recent Applications and Developments, Vol. I, Physical Aspects. (New York: Academic Press Inc. 1980). Cooper, M.B. and M.J. Wilks. An Analytical Met hod for 226Ra in Environmental Samples by the Use of Liquid Scintillation Counting, (Yallambie, Victoria: Australian Radiation Laboratory, ARL/TRO4O, 1981).

CHAPTER 49

A New Simplified Determination Method for Environmental 90Sr by Ultra Low-Level Liquid Scintillation Counting

Franz Schonhofer, Manf red Friedrich, and Karl Buchtela

ABSTRACT 90Sr is a bone seeker and when in equilibrium with its short-lived, hard-beta-emitting daughter gay, it may cause damage to the bone marrow. There is considerable interest in determining 90Sr in a wide variety of environmental material, particularly food. Normally, time-consuming chemical separation steps for isolation and decontamination from interfering radionuclides are necessary. Using a commercially available ultra low-level liquid scintillation counter, the required time for analysis could be reduced to a minimum. Extremely low LLDs allow for very small samples, and the entire separations can be performed in centrifugation tubes. The LLD achieved depends on the measurement method (Cerenkov or LS count-

ing), and is between 50 and 90 fCi/sample (1.7 to 3.0 mBq/sample), based on 3 a of the background and 500 mm counting time. The method has been tested with certified reference materials, and some results from analysis of food contaminated by Chernobyl fallout will be given.

INTRODUCTION

'°Sr is of considerable interest in radiation protection. It is produced by nuclear fission, and its fission yield is approximately the same as for 131j or '37Cs (57°/o). 90Sr, originating from atmospheric bomb tests, has been spread all over the world and found its way to man through the food chain; there is direct deposition on plants and root uptake from the soil. The main source is milk. Transfer from soil to plants is normally about four times higher for 90Sr as compared to '37Cs; transfer in the cow to milk is lower- 90Sr is incorporated into bone tissue and therefore has an extremely long biological half-life. 90Sr is a pure 3-emitter with a maximum energy of 0.546 MeV and a half-life of 28.5 years. Its daughter 90Y is a pure $-emitter as well, with a maximum energy of 2.27 MeV and a half-life of 64.1 hours, so it is very fast in equilibrium with 90Sr. The high energetic 3-partic1es may damage bone marrow. Since 90Sr is a 547

548

LIQUID SCINTILLATION COUNTING AND ORGANIC SCINTILLATORS

highly radiotoxic nuclide maximum permissible levels in food are very low, and sensitive methods are needed for its determination. Since the ban of atmospheric nuclear weapons testing in 1962 the 90Sr concentration in food has decreased considerably. In the case of the nuclear accident at Chernobyl, 90Sr was emitted to a much lower fraction than '37Cs because it is less volatile. In Western Europe the total 90Sr deposition was much lower than the one originating from the atmospheric bomb tests.' Nevertheless measurement of 90Sr is necessary both for radiation protection purposes and public interest. It has been shown for Austria, by measurements with a conventional method, that in 1986 the additional intake of 90Sr was equal to the intake of remaining 90Sr from atmospheric bomb tests.2 This paper will deal only with 90Sr, in the absence of short lived fission products, and 89Sr (half life 50.5 d). Analysis of 90Sr and 89Sr will be the next step in development of fast measurement methods for the case of a nuclear accident.

STRATEGY FOR THE DEVELOPMENT OF THE METHOD

The traditional "nitrate method"3 has some serious disadvantages, the most important is handling the large amounts of fuming nitric acid for repeated precipitations. It was therefore attempted to avoid or at least reduce the number of purification steps with fuming nitric acid. Liquid scintillation counting has been used for 90Sr determinations since at least 1965 (an overview is given in Dehos4). The determination of low levels of

90Sr has been considered by Salonen,5 who also gave a first comparison between LSC and gas flow counters. In LSC the sample has a 4 r geometry, a very high efficiency compared to gas flow counters.

Before choosing a separation method it was necessary to determine the lower limits of detection of the liquid scintillation counters for 90Sr. The equipment used was the commercially available ultra low-level liquid scintillation counter "Quantulus" (Pharmacia - Wallac, Turku, Finland). This instrument exhibits extremely low background due to heavy passive and an active shielding. The concentration of 90Sr can be determined by several methods. In samples where 90Y is in equilibrium with 90Sr, the former can be extracted by, e.g., HDEHP, and measured either by Cerenkov counting or after mixture with a suitable scintillation cocktail. (This method was first applied to cheese, but interference by an unidentified radionuclide with a half-life of approximately 6 hours was found.) 90Sr can be measured directly after separation and mixture with a suitable cocktail and it can be remeasured after ingrowth of 90Y for

control. 90Y also can be measured in the final solution after ingrowth by Cerenkov counting followed by scintillation counting. Data from Cerenkov counting can be obtained only after some waiting time, but then possible

A NEW SIMPLIFIED DETERMINATION METHOD

549

interferences from lower energy beta emitters and gamma emitters can be excluded.

First tests were conducted at the low-level laboratory ("LOLA") of Wallac OY in Turku, Finland. On one hand the environment shows high background

radiation due to granite rock, on the other hand the laboratory has heavy concrete shielding. For all measurements, PTFE coated polyethylene vials (Zinsser, Frankfurt) were used. Cerenkov measurements were done in 10 ml 0.5 M hydrochloric acid; in LSC measurements 4 ml of 0.5 M hydrochloric acid solution were mixed with 6 ml Quickszint 400 (Zinsser). Figure 1 shows the obtained pulse height spectra from Cerenkov counting both for background and standard. The window was optimized for highest

figure of merit, using standard software for Quantulus. Efficiency was 69.25 07, background 0.753 cpm. In Figure 2 90Sr and 90Y are in equilibrium and mixed with cocktail. The window is adjusted for the sum of the radionuclides, giving an efficiency of 184.5707o and a background of 1.988 cpm. If the window is adjusted for the part of the 90 spectrum above the maximum energy of 90Sr (Figure 3), then an efficiency of 58. 1207o and a background of 0.603 cpm results.

Table 1 is a summary of the results obtained in terms of figure of merit (E2/B) and lower limit of detection both for counting times of 100 and 500 mm. The LLD is based on 3 a of the background. DETERMINATION OF 90SR

The extremely low LLDs achievable with our equipment makes it possible to use considerably lower amounts of material than in conventional methods. In the conventional nitrate method described, 20 g milk or cheese ash are used.

We have used for our analysis between 1 and 2 g of ash. The following procedure was tested and is now used in routine work: The ash of dairy products is dissolved in dilute nitric acid; carrier solutions of strontium, barium, iron, and chromate are added. The resulting volume in the first step is about 120 mL. Upon addition of ammonium acetate and adjustment to pH 5, barium chromate, iron phosphate, and basic iron acetates are precipitated and centrifugated. Phosphate is thus removed from the sample. Radium is also removed and it can be recovered and isolated from the precipitation. After precipitation with ammonium carbonate, work can be continued in small volumes of 10 to 20 mL. After dissolution in nitric acid, hydrogen peroxide, and yttrium carrier, the hydroxides of chromium and yttrium are removed. The final solution can be either acidified and measured directly or have one carbonate precipitation step added.

The use of fuming nitric acid is avoided completely. The chemical recovery,

which is determined by atomic absorption spectroscopy, is in the range of 7007o, which is regarded as sufficient compared with the ease of the procedure. In three days 20 samples of cheese ash have been processed. The limiting and

figure 1.

Eecsy ratio

01 . 07.,, 79 00:00 20 . 04 Ei9 00 : 00

(ci/Ao) = 0.7813

mE :Euremen t d te

Ha1f L...i.+.?

69.26 '.)

600

8P1112

SPItI2

2841&, .0) ...... 22207 .42 DPN ( decay : or r

206 300 400 580 [A1153771.641 CPM [8] 6.753 CPu (222626.42 DPIlco) FM {[A],[B1} 6367 (E

160

pr8p-.r tior dt

* FM *

58- 390)

H

A:\SR\BUCHT-Fg\cER1\[email protected]@1NH A:\SB\BUCHT-FS\CEBI\Q040401H 886

Pulse height spectra of 90Sr standard and background, and Cerenkov counting in 0.5 M hydrochloric acid, with window optimized.

sroN):.)4c.o

QtcI::)E

iN TEGH C [email protected]

1

[email protected] CPII/c}i

L

3.01 mii

394.07 nhi

756.UHH CPM/ch

Figure 2.

ie

2

3

3.39 miii 394.8? miii

4

[email protected] [email protected]

8fl

,.-

rnLArEm'1nt date cIay (c/P'o)

Half prEpa ....ti.D.... Y

222c:)26. 42 DFM (decay

:,::::.Q4.89 0000

78l

284160, c'::'

c:1 .o;.79 cx:c:,o

i .



ccrr.

and 90Y.

Pulse height spectra of 90Sr standard and background, and mixture with Quickszint 400, with window optimized for sum of 9°Sr

STNDñID

SPfl12

SP*12

(E184.57 >.)

[email protected]

R

A:\SR\BIJCHT-FS\LSC1\Q561NMU A:"SH\BUCHT-FS\LSC1\QO4Ø4jNB

1NTER( 324- 875) [[email protected] CPM [B] 1.988 CPM BUNCH 5 * FM (22226.42 DPMco) FM {[A],[B]} 17136

1.

@[email protected] CPtI/ch

13MO CPM/ch

3.39 mm 394.87 mm L

ASR\BUCHT-F8\LSC1\Q056081M.000 :\8H\[email protected] 8Pi12

8Pfl12

Figure 3.

Iiir? ........

,:Lio

. (A/;c .:

..:....c. o7si.:::

90Y alone.

Pulse height spectra of 90Sr standard and background, and mixture with Quickszint 400, with window optimized for

J.:::t..::': ...r:t)

200 308 408 580 608 700 888 988 1800 1 188 [A]129032.133 CPM [II] 0.683 CPM 1NTE6R( 665- 875) FM {[],[B]} 5604 (E= 58.12 z) &JNCH= 5 FM (222026.42 DPMco)

[!] 1300.000 CPM/ch [B] 0.030 CPM/ch

A NEW SIMPLIFIED DETERMINATION METHOD

Table 1. Performance of Quantulus for 90Sr for Different Conditions Figure LLD (mBq) Measurement Condition of Merit (100 mm) Cerenkov counting LS counting, open window LS counting, 90Y window

553

LLD (mBq) (500 mm)

6.3 3.8

2.8

17,136 5,604

6.7

3.0

6,367

1.7

most time consuming step is ashing the samples. In the case of fresh milk, ion excha