A Rapid Procedure for Screening Transuranium Nuclides in Urine Using Actinide Resin and Low

Transcription

1 application Note A Rapid Procedure for Screening Transuranium Nuclides in Urine Using Actinide Resin and Low Level a/b-lsc Liquid Scintillation Counting Authors J. Eikenberg, I. Zumsteg, M. Rüthi S. Bajo, Paul Scherrer Institute CH-5232 Villigen (PSI) Switzerland C. J. Passo PerkinElmer, Inc. Waltham, MA, USA M. J. Fern Eichrom Industries, Inc. Darien Illinois, USA Abstract A fast and simple radiochemical procedure for determining a-emitting nuclides in urine is presented. The method is based on a/b-lsc using the Tri-Carb with alpha/beta discrimination as well as on the high selectivity of Eichrom s actinide resin for heavy isotopes with atomic numbers above 90 (i.e., Th, U, transuranium nuclides). Under optimized pulse shape discriminator settings, a very efficient a/b discrimination of only 0.1% spill-over of b into a was obtained at 95% counting efficiency for a-pulses. In addition, using a mixture of the scintillation liquids, Gold AB and Ultima Gold F, the a-peak resolution turned out to be rather high (40 kev FWHM for 239 Pu on LSC scale). This allows the Tri-Carb with alpha/beta discrimination to be used as a spectrometer for screening either transuranium nuclides with energies exceeding 5 MeV such as 239 Pu, 241 Am, 244 Cm or uranium isotopes between 4-5 MeV ( 238 U, 236 U, 235 U, 234 U) in small counting windows of 120 kev each. Under these conditions, very low background count rates of 0.05 CPM are obtained in each window, resulting in a high figure of merit (E 2 /B) of» 180,000 and a detection limit as low as 1.5 mbq/l (or 0.04 pci/l) in a 500 minute count interval. Introduction One of the most extensive tasks in the field of bioassay analysis is the determination of pure a- (and b-) emitting radionuclides from the nuclear fuel cycle such as 234 U and 235 U, or anthropogenic 239 Pu and 241 Am in urine samples. However, any radiochemical method, which is applied to perform such analyses, has to be highly sensitive since even small amounts of incorporated radionuclides decaying by a emission may contribute to harmful doses to human organs. 1 Since a radiation has an extremely short penetration length in water and solid substances, direct counting of a salt residue of dry ashed urine is not possible. Therefore, complex radiochemical techniques have been developed for efficient separation of the transuranium elements from the bulk matrix. 2,3 However, in addition to several purification steps, these methods require the production of almost weightless planar sources (e.g., via electrolytic deposition) in order to perform radioassays with proportional or surface barrier detectors. In contrast to the extensive preparative techniques, fast methods using a/b-lsc are of increasing interest. 4,5 Due to the efficient detection of a emitters in a liquid scintillation cocktail, extensive radiochemical purification procedures are not necessary provided the sample is homogeneous in the liquid scintillation cocktail. Although for a counting, liquid

2 scintillation detectors are mainly used only as gross analyzers, they are highly suitable for screening alpha activities in bioassay samples since low detection limits of a few mbq/l can be obtained. This could be shown using a simple preconcentration chemistry with actinide coprecipitation by Ca 3(PO 4) 2. 6 A radioassay was performed using PerkinElmer s low level model Tri-Carb with alpha/beta discrimination equipped with highly efficient discrimination between a and b radiation. 7,8 However, the presence of anthropogenic nuclides such as 239 Pu, 241 Am or 244 Cm in urine using gross counting methods can be justified only if the a net count rate clearly exceeds those produced by decay of the naturally occurring radionuclides such as 238 U, 234 U, 210 Po or 226 Ra from the U and Th decay series in a blank sample. In human urine, the activities of these natural a emitters typically range between 0.1 and 20 mbq/l, mainly caused by 210 Po and 226 Ra uptake with plant diet. 9 Table 1 summarizes the typical range of activities (and corresponding count rates for a/b-lsc) for the most important natural a emitting nuclides in urine. Since typical blank count rates are as low as 0.1 CPM with the Tri-Carb with alpha/beta discrimination, the contribution from natural components significantly exceeds reagent blank values. This reduces the counting sensitivity for screening anthropogenic actinides in bioassay studies. In this work, Ra and Po radionuclides are automatically eliminated using extraction chromatography without the need of further purification steps. Methods and Materials Composition of the Actinide Resin The separation procedure described here is based on the extraordinarily strong affinity of the resin for actinides (particularly in the tri-, tetra- and hexavalent oxidation states), even from strongly acidic solutions. 10 Actinide resin is composed of a liquid extractant (containing a diphosphonic acid functional group) coated onto a chromatographic support. The extractant, trade named DIPEX (Eichrom Industries, Inc., Darien, Illinois, USA), is shown in Figure 1, where R = 2-ethylhexyl. The chromatographic support used in this study has a nominal particle size distribution of microns. The Separation Procedure The schematic radiochemical procedure is shown in Figure 2. The individual steps will be explained in detail below. Partial oxidation of the organic matter: 0.5 L urine is transferred into a 1 L glass beaker and 100 ml 65% HNO 3 is added. The beaker is then covered with watch glass and gently boiled for two hours under infrared light. Subsequently, the solution has to cool down to room temperature. Sorption on actinide resin: 200 mg of actinide resin is added and the solution is stirred for four hours to ensure sorption equilibrium (see Figure 3). Separation of the resin from the solution: The separation of the resin containing the actinides from the bulk of the solution is obtained via filtration on 0.3 μm (25 mm diameter) WCN cellulose nitrate membrane filters (Whatman Inc., Ann Arbor, Michigan, USA) mounted on a 25 mm glass frit membrane holder. To increase the filtration rate, the filtration is performed under a vacuum using a water pump. Table 1. Typical range of the activities of natural a emitting radionuclides in human urine. 2

3 After filtration, the resin on the filter will show a yellow color due to adsorption of some organic substances during the exposure process. The filtrate (solution) can then be removed. Figure 1. Structure of the actinide resin. Stripping of the reagent from the resin: Stripping of the reagent from the inert support (polymeric substrate) is performed in three consecutive steps using 5 ml isopropanol in each step. As in the preceding step, the filtration is performed under a vacuum using a water pump. The organic solution will be yellow and (after dissolution of the reagent from its bed) the substrate will be white in color. The support can then be discarded. Preparation of the solution for LSC: The organic solution is transferred into a 100 ml wideneck quartz glass flask and is taken to dryness on a heating plate using additional infrared. Five ml 65% HNO 3 and 1 ml of concentrated H 2SO 4 is then added for oxidation of the reagent. The solution is gently boiled and slowly evaporated to dryness until a thoroughly clear, transparent residue is obtained. If the color is not transparent, addition of 30% of H 2O 2 to the cooled residue is helpful. The transparent residue is then dissolved in 2 ml 0.5 M HCl. Cocktail preparation and liquid scintillation counting: The sample solution is transferred into a 20 ml plastic scintillation vial containing a mixture of Ultima Gold AB (12 ml) and Gold F (5 ml). This mixture yields optimal a/b pulse shape discrimination and peak resolution of the a pulses (see Figure 4). The vial is shaken until aqueous and organic phases are mixed completely and the cocktail solution is cooled to a temperature of about 10 C in a refrigerator. Prior to measurement, the cocktail must be checked for phase separation (solution must be homogeneous, transparent and colorless); finally liquid scintillation counting is performed using the Tri-Carb Alpha/beta discrimination option in the a/b-mode. Results and Discussion A detailed description of various experiments to study the sensitivity of different chemical parameters, uptake capacity of the resin, and counting conditions can be obtained from Eikenberg, et al. 19 Figure 2. Schematic illustration of the fast procedure for separation of actinides from urine. 3

4 close to 100% at the crossover setting, high values of 95% were still obtained at the higher (140 ns) discriminator setting (see the Analysis of Counting Sensitivities section). Figure 3. Kinetic uptake experiments: determination of sorption halflives for U and Am. Figure 3. Kinetic uptake experiments: determination of sorption halflives for U and Am. Recoveries of Th, Pa, U, Pu, Am, Cm: The chemical yield or Recoveries of Th, Pa, U, Pu, Am, Cm: The chemical yield or recovery following a complete radio- recovery following a complete radiochemical analysis was determined from the addition of radiospikes of known chemical analysis was determined from the addition activity to a blank urine sample (Table 2). In particular, two of radiospikes of known activity to a blank urine steps were studied to check on chemical recovery: sample (Table 2). In particular, two steps were studied to check The on adsorption chemical yield recovery: on the resin The overall (total) chemical recovery The adsorption yield on the resin To The obtain overall the reproducibility (total) chemical of the recovery results, all spike experiments were repeated at least four times for each radionuclide. To obtain The the sorption reproducibility yield on the of resin the results, was by all means spikeof g-spectrometry experiments were of the repeated nuclide itself at least or via four decay times or ingrowth for of each daughter radionuclide. nuclides. The sorption yield on the resin was by means of γ-spectrometry of the nuclide itself Table or via 2 decay clearly or indicates ingrowth that of actinide daughter resin nuclides. has an extremely strong affinity for all tested actinides even from a very strong Table 2 acidic clearly urine indicates solution with that actinide high salt content resin has (average an salt extremely content strong 30 g/l). affinity Table 2 for also all shows tested actinides that there even is almost no from difference a very strong between acidic the urine chemical solution yields with for a high complete salt analysis content and (average the adsorption salt content yield. 30 This g/l). means Table that 2 additional also chemical shows that losses there from is almost (i) stripping, no difference (ii) digestion between and (iii) the transfer chemical into yields the liquid for scintillation a complete vial analysis are insignificant. and the This adsorption method yield. can hence This be means easily that adopted additional to routine chemical losses use. from (i) stripping, (ii) digestion and laboratory (iii) Figure 4. Smoothed liquid scintillation spectrum of 239 Pu and 244 Cm obtained with Packard Tri-Carb 2550TR/AB. Figure 4. Smoothed liquid scintillation spectrum of 239 Pu and 244 Cm obtained with PerkinElmer Tri-Carb 2550TR/AB. Chemical Yield Investigations Chemical Direct Yield spike Investigations experiments: The counting efficiencies spike were experiments: determined The with counting radiolabeled efficiencies spiked were Direct solutions added to cocktail mixtures to simulate determined with radiolabeled spiked solutions added to routine chemical analysis. The cocktails were measured each under two different discriminator set- cocktail mixtures to simulate routine chemical analysis. The cocktails were measured each under two different discriminator settings, i.e., at the crossover point (125 ns) and at value of 140 ns. While the counting efficiencies were close tings, i.e., at the crossover point (125 ns) and at a to 100% at the crossover setting, high values of 95% were still obtained at the higher (140 ns) discriminator setting (see the Analysis of Counting Sensitivities section). 4 Table 2. Chemical recoveries obtained from a complete analytical Table 2. procedure. Chemical recoveries obtained from a complete analytical procedure. 4

5 All experiments were performed with 200 mg resin per 0.5 L sample and an extraction time of at least four hours (for kinetic studies see the following section entitled Experiments on Uptake Kinetics ). Under these conditions only about 75% Am was consistently recovered, whereas the recoveries of most of the investigated actinides exceeded 90%. This discrepancy is most likely due to the fact that the resin uptake coefficient for Am(III) more rapidly decreases with acidity compared to those actinides present in the tetra- or hexavalent state such as Th(IV), Pu(IV) or U(VI). 10 Recovery of Ra: In contrast to the actinides, the uptake of Ra on actinide resin was found to be less than 5% (Table 2). This result is consistent with the low sorption coefficients (k'-values) for the alkaline earth elements (Ca 2+ and Ra 2+ ) in strong acidic medium as obtained by Horwitz, et al. 10 Even in the presence of 2 M HCl solutions containing 1 M CaCl 2, the uptake of the least efficiently sorbed species Am(III) remained considerably high (k' = 103). Since k' is lower for Ra 2+ than Ca 2+ (k' <1), and average urine Ca/Ra ratios are extremely high, no additions of Ca or Ba carrier are required for a routine analysis. Recovery of Po: Because oxidation of the stripped reagent fraction is performed under high temperatures using HNO 3/ H 2SO 4 mixtures (boiling point of sulfuric acid = 338 C), the second naturally occurring component 210 Po is efficiently eliminated since under acidic conditions at elevated temperatures, Po (probably present as Po-oxide in the ash) is volatile. Tracer experiments with 209 Po(NO 3) 4 spike solutions indeed revealed repeatedly no detectable activity in the liquid scintillation cocktail. Uptake studies with different resin additions: As discussed above, the uptake coefficient of Am(III) decreases rapidly with acidity. Therefore, slight neutralization of the aqueous samples with NH 4OH (following the oxidation step with HNO 3) was attempted. However, when adding NH 4OH to reduce the acidity of the solution from 2 M to 1 M HNO 3, the solutions became black and opaque. An improved technique to obtain a higher extraction yield is simply to increase the amount of actinide resin per same sample volume. The results for additions of 0.4 g/l and 1 g/l are depicted in Figure 5. Almost quantitative extraction for all the actinides were obtained when taking 1 g/l actinide resin. Figure 5. Bar chart showing the chemical recoveries obtained from two different resin additions. Experiments on uptake kinetics: To obtain the times required for sorption equilibrium at a steady state, the uptake kinetics were studied for tri- and hexavalent species using Am(III) and U(VI) tracer solutions. For these investigations, aliquots were prepared as explained in the previous section. This time, however, aliquots were spiked with identical activity concentrations and the extraction was interrupted at times given in Figure 3. Very rapid uptake was observed and in about two hours steady state conditions were obtained independently of the amount of added resin. If the sorption process follows first order kinetics, the data should plot on a straight line in a semi-log diagram with the remaining activity in solution plotted versus the exposure time (Figure 3). In this case the sorption exponent k sorp can be extracted from the relation: a solution = e k sorp t (with a solution = activity in solution) and via regression analysis of the data. A more comprehensive approach is the use of the sorption half-lives (i.e., T 1/2 = ln2/k). Very short half-lives of only eight and 20 minutes were calculated for U and Am, respectively using this approach. 5

6 Table 3. Set of values used for the calculation of the LLDs for 239 Pu. The set of parameters used for the calculation of the Figure 6. LLD is given in Table 3. The other parameters were α/β crossover curves as function of PDD setting obtained with kept either constant (i.e., V s = 0.5 L) or were not 241 Am and 36 Cl. relevant (µ). However, it has to be noted that in contrast to procedures based on LSC, µ can only be Analysis of Counting Sensitivities omitted when almost weightless sample discs are produced. If that is not the case, absorption of α Optimum α/β discriminator settings: For gross radiation in the sample source itself has to be α/β counting systems, two parameters are essential considered seriously. to determine the sensitivity of a radioassay: background count rate (B) and counting efficiency (E), Set of In values Figure used 7, three for the methods calculation are of the compared; LLDs for 239 two Pu. proce- Table 3. which can be expressed as figure of merit or E 2 /B. 20 dures based on α/β-lsc (previous work of Eikenberg, To reduce background scatter, misclassification of β et al.) 6 and a method 2 developed for low level gasflow Table of 3. proportional parameters Set of values used counting for for the calculation the (GPC). calculation of the Although LLDs of for the 239 a Pu. very pulses counted as α has been minimized by optimiz-thing the pulse decay Figure and afterpulse 6. analysis featureslld low is given background in Table count 3. The rate other of parameters only 0.04 CPM were was set α/β crossover curves as function of PDD setting obtained of the Packard Tri-Carb 2500TR/AB. 7,8 As shown with kept either constant (i.e., V s = 0.5 L) or were not in 241 Am and Figure 36 Cl. 6, low α and β misclassification (0.6%) relevant (µ). However, it has to be noted that in Figure 6. a/b crossover curves as function of PDD setting obtained resulted at the optimum pulse decay discriminatorcontrast to procedures based on LSC, µ can only be with Analysis (PDD) 241 Am and of setting Counting 36 Cl. of 125. Sensitivities Optimum E 2 /B values were, omitted when almost weightless sample discs are however, obtained for a slightly higher PDD settingproduced. If that is not the case, absorption of α Optimum Analysis of 140. of α/β At Counting this discriminator value, Sensitivities the β spill settings: (0.1%) For is gross extremelyradiation in the sample source itself has to be α/β Optimum counting low (hence a/b systems, discriminator significantly two parameters settings: reducing For the are gross α essential background a/b considered seriously. to counting determine count systems, rate), the while sensitivity two the parameters loss of in a counting radioassay: are essential efficiency background the to sensitivity counting rate of some a (B) radioassay: and α pulses counting background as β is efficiency minimal. count (E), rate (b) In Figure 7, three methods are compared; two proce- to determine due which and counting can be expressed efficiency (E), as figure which of can merit be expressed or E 2 /B. 20 as figure dures based on α/β-lsc (previous work of Eikenberg, To of reduce merit Comparison or background E 2 /B. 20 of To detection reduce scatter, background misclassification limits: The scatter, lower of misclassification detection counted of b (LLD) pulses as α has counted at been the 95% minimized as a confidence has been by optimiz- minimized probability by flow proportional counting (GPC). Although a very limit β ofet al.) 6 and a method 2 developed for low level gas- pulses ing optimizing the level pulse was the decay calculated pulse and decay afterpulse from and analysis afterpulse analysis of analysis blank features samples features low background count rate of only 0.04 CPM was of the of the using Packard Tri-Carb the equation Tri-Carb with alpha/beta as 2500TR/AB. given discrimination. by Seymour, 7,8 As shown et 7,8 As al.: shown 21 in Figure 6, low α and β misclassification (0.6%) in Figure 6, low a and b misclassification (0.6%) resulted at resulted at the optimum pulse decay discriminator the optimum pulse decay discriminator (PDD) setting of 125. (PDD) setting of 125. Optimum E 2 /B values were, Optimum E 2 /B values were, however, obtained for a slightly however, obtained for a slightly higher PDD setting higher PDD setting of 140. At this value, the b spill (0.1%) of 140. where At this (K) value, = 1.64 the = statistical β spill (0.1%) is for extremely a confidence is extremely low (hence significantly reducing the a background time rate), low (hence interval significantly of 95%; (I 0 reducing ) = total the background α background counts in count count t; (t) while = rate), counting the while loss time; in the counting loss (Y counting efficiency i ) = chemical efficiency recovery; due Figure 7. Evolution of the lower limit of detection for three methods due (E) to = counting some or detector a pulses as efficiency; b is minimal. Figure 7. to counting some α pulses as β is minimal. (V s ) = sample based on GPC and LSC. Evolution of the lower limit of detection for three methods volume; and (µ) = attenuation coefficient. based on GPC and LSC. Comparison of detection limits: The lower limit of detection Comparison (LLD) at the 95% of detection confidence limits: probability The lower level limit was calculated of detection (LLD) at the 95% confidence probability from analysis of blank samples using the equation as given level was calculated from analysis of blank samples 6 by Seymour, et al. using the equation 21 as given by Seymour, et al.: 21 where (K) = 1.64 = statistical value for a confidence interval of 95%; (I 0 ) = total background counts in time where t; (t) (K) = = counting 1.64 = statistical time; (Yvalue for a confidence interval of i ) = chemical recovery; 95%; (I 0 ) = total background counts in time t; (t) = counting Figure 7. (E) = counting or detector efficiency; (V s ) = sample Evolution of the lower limit of detection for three methods volume; time; (Yand i ) = chemical (µ) = attenuation recovery; (E) coefficient. = counting or detector based on GPC and LSC. efficiency; (V s ) = sample volume; and (μ) = attenuation coefficient. 6 6

7 The set of parameters used for the calculation of the LLD is given in Table 3. The other parameters were kept either constant (i.e., V s = 0.5 L) or were not relevant (μ). However, it has to be noted that in contrast to procedures based on LSC, μ can only be omitted when almost weightless sample discs are produced. If that is not the case, absorption of a radiation in the sample source itself has to be considered seriously. In Figure 7, three methods are compared; two procedures based on a/b-lsc (previous work of Eikenberg, et al.) 6 and a method 2 developed for low level gasflow proportional counting (GPC). Although a very low background count rate of only 0.04 CPM was taken to calculate the LLD using low level GPC counters, the new procedure based on a/b-lsc yields considerably lower LLD values. This result is the consequence of the very high counting efficiencies and chemical recoveries. If, in particular, LSC is carried out using a small window for analysis of a special group of actinides (see the a peak resolution and liquid scintillation quench paragraph below), the background decreases to values about 0.05 CPM. This yields an extraordinary high figure of merit (E 2 /B) of 180,000 or a detection limit of 1.5 mbq/l in a 500 minute counting interval. a peak resolution and liquid scintillation quench: It is well known that a peak resolution using LSC is poor in comparison to a spectrometry and hence a/b-lsc systems are used mainly as gross counters. Nevertheless, if a pulse stretching scintillators are used, FWHM values of kev can be obtained. 22 Since, for the current procedure, a near organic cocktail mixture is prepared, the a peak resolution becomes fairly high (400 kev or 40 kev on a liquid scintillation scale). This allows peak separation between 234 U and 238 U or, as shown in Figure 4, between transuranium nuclides such as 239 Pu and 244 Cm (DE = 600 kev). Two observations are of interest. First, the peaks are symmetrically shaped and simple Gaussian-based fitting procedures (without additional terms for peak tailing) are sufficient for fitting of overlapping peaks. Second, there is a significant shift of the a energy using LS with respect to the real emission energy. This shifting phenomenon is due to ionization quench because the a particles dissipate their energy over a very small distance causing less excited niveaus in the orbitals of the scintillator targets. 23 However, the shift in energy from quenching remains constant for a fixed cocktail mixture. The relation between the true a emission energy and the liquid scintillation quenched output is demonstrated in Figure 8 for the procedure given here and an aqueous cocktail. 6 Figure 8. Relations between the true emission energy and liquid scintillation quenched a energies of actinides. It is interesting to note, that for a set of a emitters with different energies, the ratio between both energy scales is highly linear. Although the aqueous cocktail is quenched to a higher degree, regression analyses yielded identical slopes of exactly 10 (see Figure 8). This also implies that there is almost no drift for the electronic assignment of a given energy to the multichannel analyzer (MCA). Indeed stability tests using 239 Pu spiked cocktails revealed identical peak positions which scattered less then 15 kev on the liquid scintillation scale for samples produced within one year. This fact is helpful to distinguish between two major groups of actinides which are of interest for in vitro measurements. As shown in Figure 8, all anthropogenic transuranium nuclides are characterized by emission of higher a energies compared to all natural uranium isotopes (i.e., 234 U, 235 U, 238 U). For in vitro screening, a distinction between these groups is often reasonable. For instance, monitoring of employees involved in uranium mining should be limited to uranium isotopes, whereas in nuclear reprocessing plants or hot laboratories handling spent fuel elements, radiation hazards may arise predominantly from incorporation of 239 Pu, 240 Pu, 241 Am and 244 Cm. Therefore, instead of counting gross a over a wide range of energy, a small window of only 120 kev can be taken for the group of uranium isotopes ( kev) as well as for the transuranium nuclides ( kev). 7

8 Conclusions A fast and very efficient radiochemical procedure for screening a activities in urine was developed based on sorption using an actinide extractive resin. A high figure of merit (180,000) is obtained by performing the radioassay with a Tri-Carb with the Alpha/beta discrimination option. A low detection limit of 1.5 mbq/l (0.04 pci/l) can be obtained in about eight hours counting time. This allows an annual throughput of about 1,000 samples for screening a activities in urine. A complete analysis requires the use of only 0.5 L samples, additions of only 0.2 g resin and can be terminated within one day. References 1. Int. Com. Radiological. Protection. ICRP Publication 54, Ann. ICRP Vol. 19 (1988). 2. Eakins, J.D. and Gomm, P.J. (1968) Health Phys. 14, Horwitz, E., Ph Dietz, M.L., Nelson, D.M., LaRosa, J.J. and Fairman, W.D. (1990) Anal. Chim. Acta Vol. 238, Salonen, L. (1993) Sci. Tot. Environ. Vol. 130/131, Bickel, M., Möbius, S., Kilian, F. and Becker, H. (1992) Radiochim. Acta Vol. 57, Eikenberg, J., Fiechtner, A., Ruethi, M. and Zumsteg, I. Liquid Scintillation Spectrometry 1994, Radiocarbon 1996, Passo, C.J. and Kessler, M.J. (1992) Packard Instrument Company Publication, Report PBR0012, Shiraishi, K., Yamamoto, M., Yoshimizu, K., Igarashi, Y. and Ueno, K. (1994) Health Phys. Vol. 66, Horwitz, E.P., Chiarizia, R. and Diez, M.L. React. Funct. Polymers (in press). 11. Wrenn, et al., J. Rad. Nuc. Chem. Art. 156 (1992) Karpas, et al., Health Phys. 71 (1996) Dang, et al., Health Phys. 57 (1989) Dahlheimer and Henrichs, Rad. Prot. Dos. 53 (1994) Fisenne, et al., Health Phys. 53 (1987) ICRP Publication. 23 (1975). 17. Shiraishi, et al., Health Phys. 66 (1994) Fellmann, et al., Health Phys. 57 (1989) Eikenberg, J., Zumsteg, I., Ruethi, M., Bajo, S., Fern, M. J. and Passo, C.J., J. Radact. Radiochem. (in press). 20. Currie, L. A. (1968) Anal. Chem. 40, Seymour, R., Sergent, F., Knight, K. and Kyker, B. (1992) Radioact. Radiochem. 3, Yu, Y.F., Salbu, B., Bjornstad, H.E. and Lien, H.J. (1990) Radioanal. Nucl. Chem. Lett. 145, Horrocks, D. (1974) Academic Press. New York-London Passo, C.J. and Kessler, M.J. Liquid Scintillation Spectrometry 1992, Radiocarbon 1993, PerkinElmer, Inc. 940 Winter Street Waltham, MA USA P: (800) or (+1) For a complete listing of our global offices, visit Copyright 2011, PerkinElmer, Inc. All rights reserved. PerkinElmer is a registered trademark of PerkinElmer, Inc. All other trademarks are the property of their respective owners. DIPEX is a trademark of Eichrom Industries, Inc A_01 Printed in USA April 2011