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 239Pu 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 239Pu, 241Am, 244Cm or uranium isotopes between 4-5 MeV (238U, 236U, 235U, 234U) 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 (E2/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.

Liquid Scintillation Counting

a p p l i c a t i o n n o t e

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

A Rapid Procedure for Screening Transuranium Nuclides in Urine Using Actinide Resin and Low Level a/b-LSC

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 234U and 235U, or anthro-pogenic 239Pu and 241Am 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

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 com-posed 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 100-150 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 trans-ferred into a 1 L glass beaker and 100 mL 65% HNO3 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.

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 pre-concentration chemistry with actinide coprecipitation by Ca3(PO4)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 239Pu, 241Am or 244Cm 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 238U, 234U, 210Po or 226Ra 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 210Po and 226Ra 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 radio-nuclides are automatically eliminated using extraction chromatography without the need of further purification steps.

Table 1. Typical range of the activities of natural a emitting radionuclides in human urine.

2

as 238U, 234U, 210Po or 226Ra from the U and Th decayseries in a blank sample. In human urine, the activi-ties of these natural α emitters typically rangebetween 0.1 and 20 mBq/L, mainly caused by 210Poand 226Ra uptake with plant diet.9

Table 1 summarizes the typical range of activities(and corresponding count rates for α/β-LSC) forthe most important natural α emitting nuclides inurine. Since typical blank count rates are as low as0.1 CPM with the Packard Tri-Carb 2550TR/AB,the contribution from natural components signifi-cantly exceeds reagent blank values. This reducesthe counting sensitivity for screening anthropogenicactinides in bioassay studies. In this work, Ra and Poradionuclides are automatically eliminated usingextraction chromatography without the need offurther purification steps.

Methods and Materials

Composition of the Actinide ResinThe separation procedure described here is based onthe extraordinarily strong affinity of the resin foractinides (particularly in the tri-, tetra- and hexavalentoxidation states), even from strongly acidic solu-tions.10 Actinide resin is composed of a liquid extrac-tant (containing a diphosphonic acid functional group)coated onto a chromatographic support. The extrac-tant, trade named DIPEX™ (Eichrom Industries,

Inc., Darien, Illinois, USA), is shown in Figure 1,where R = 2-ethylhexyl. The chromatographic sup-port used in this study has a nominal particle sizedistribution of 100-150 microns.

The Separation ProcedureThe schematic radiochemical procedure is shown inFigure 2. The individual steps will be explained indetail below.

Partial oxidation of the organic matter: 0.5 Lurine is transferred into a 1 L glass beaker and100 mL 65% HNO3 is added. The beaker is thencovered with watch glass and gently boiled for twohours under infrared light. Subsequently, the solu-tion has to cool down to room temperature.

Sorption on actinide resin: 200 mg of actinide resinis added and the solution is stirred for four hours toensure sorption equilibrium (see Figure 3).

Separation of the resin from the solution: Theseparation of the resin containing the actinides fromthe bulk of the solution is obtained via filtration on0.3 µm (25 mm diameter) WCN cellulose nitratemembrane filters (Whatman Inc., Ann Arbor, Michi-gan, USA) mounted on a 25 mm glass frit membraneholder. To increase the filtration rate, the filtration isperformed under a vacuum using a water pump.

Table 1.Typical range of the activities of natural α emitting radionuclides in human urine.

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.

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% HNO3 and 1 mL of concentrated H2SO4 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 H2O2 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

3

After filtration, the resin on the filter will show ayellow color due to adsorption of some organicsubstances during the exposure process. The filtrate(solution) can then be removed.

Stripping of the reagent from the resin: Strippingof the reagent from the inert support (polymericsubstrate) is performed in three consecutive stepsusing 5 mL isopropanol in each step. As in thepreceding step, the filtration is performed under avacuum using a water pump. The organic solutionwill be yellow and (after dissolution of the reagentfrom its bed) the substrate will be white in color.The support can then be discarded.

Preparation of the solution for LSC: The organicsolution is transferred into a 100 mL wideneck quartzglass flask and is taken to dryness on a heatingplate using additional infrared. Five mL 65% HNO3and 1 mL of concentrated H2SO4 is then added foroxidation of the reagent. The solution is gentlyboiled and slowly evaporated to dryness until athoroughly clear, transparent residue is obtained. Ifthe color is not transparent, addition of 30% of H2O2to the cooled residue is helpful. The transparentresidue is then dissolved in 2 mL 0.5 M HCl.

Cocktail preparation and liquid scintillationcounting: The sample solution is transferred into a20 mL plastic scintillation vial containing a mixtureof Ultima Gold AB (12 mL) and Ultima Gold F(5 mL). This mixture yields optimal α/β pulseshape discrimination and peak resolution of the αpulses (see Figure 4). The vial is shaken until aque-ous and organic phases are mixed completely andthe cocktail solution is cooled to a temperature ofabout 10 °C in a refrigerator. Prior to measurement,the cocktail must be checked for phase separation(solution must be homogeneous, transparent andcolorless); finally liquid scintillation counting isperformed using the Tri-Carb 2550TR/AB in theα/β-mode.

Results and Discussion

A detailed description of various experiments tostudy the sensitivity of different chemical param-eters, uptake capacity of the resin, and countingconditions can be obtained from Eikenberg, et al.19

Figure 1.Structure of the actinide resin.

Figure 2.Schematic illustration of the fast procedure forseparation of actinides from urine.

Figure 1. Structure of the actinide resin.

Figure 2. Schematic illustration of the fast procedure for separation of actinides from urine.

3

After filtration, the resin on the filter will show ayellow color due to adsorption of some organicsubstances during the exposure process. The filtrate(solution) can then be removed.

Stripping of the reagent from the resin: Strippingof the reagent from the inert support (polymericsubstrate) is performed in three consecutive stepsusing 5 mL isopropanol in each step. As in thepreceding step, the filtration is performed under avacuum using a water pump. The organic solutionwill be yellow and (after dissolution of the reagentfrom its bed) the substrate will be white in color.The support can then be discarded.

Preparation of the solution for LSC: The organicsolution is transferred into a 100 mL wideneck quartzglass flask and is taken to dryness on a heatingplate using additional infrared. Five mL 65% HNO3and 1 mL of concentrated H2SO4 is then added foroxidation of the reagent. The solution is gentlyboiled and slowly evaporated to dryness until athoroughly clear, transparent residue is obtained. Ifthe color is not transparent, addition of 30% of H2O2to the cooled residue is helpful. The transparentresidue is then dissolved in 2 mL 0.5 M HCl.

Cocktail preparation and liquid scintillationcounting: The sample solution is transferred into a20 mL plastic scintillation vial containing a mixtureof Ultima Gold AB (12 mL) and Ultima Gold F(5 mL). This mixture yields optimal α/β pulseshape discrimination and peak resolution of the αpulses (see Figure 4). The vial is shaken until aque-ous and organic phases are mixed completely andthe cocktail solution is cooled to a temperature ofabout 10 °C in a refrigerator. Prior to measurement,the cocktail must be checked for phase separation(solution must be homogeneous, transparent andcolorless); finally liquid scintillation counting isperformed using the Tri-Carb 2550TR/AB in theα/β-mode.

Results and Discussion

A detailed description of various experiments tostudy the sensitivity of different chemical param-eters, uptake capacity of the resin, and countingconditions can be obtained from Eikenberg, et al.19

Figure 1.Structure of the actinide resin.

Figure 2.Schematic illustration of the fast procedure forseparation of actinides from urine.

4

Chemical Yield Investigations

Direct spike experiments: The counting efficiencies were determined with radiolabeled spiked solutions added to cocktail mixtures to simulate routine chemical analysis. The cocktails were measured each under two different discrimi-nator settings, i.e., at the crossover point (125 ns) and at a value of 140 ns. While the counting efficiencies were 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).

4

Chemical Yield Investigations

Direct spike experiments: The counting efficien-cies were determined with radiolabeled spikedsolutions added to cocktail mixtures to simulateroutine chemical analysis. The cocktails were mea-sured each under two different discriminator set-tings, i.e., at the crossover point (125 ns) and at a

value of 140 ns. While the counting efficiencies wereclose to 100% at the crossover setting, high valuesof 95% were still obtained at the higher (140 ns)discriminator setting (see the “Analysis of CountingSensitivities” section).

Recoveries of Th, Pa, U, Pu, Am, Cm: The chemi-cal yield or recovery following a complete radio-chemical analysis was determined from the additionof radiospikes of known activity to a blank urinesample (Table 2). In particular, two steps were stud-ied to check on chemical recovery:

• The adsorption yield on the resin• The overall (total) chemical recovery

To obtain the reproducibility of the results, all spikeexperiments were repeated at least four times foreach radionuclide. The sorption yield on the resinwas by means of γ-spectrometry of the nuclide itselfor via decay or ingrowth of daughter nuclides.

Table 2 clearly indicates that actinide resin has anextremely strong affinity for all tested actinides evenfrom a very strong acidic urine solution with high saltcontent (average salt content 30 g/L). Table 2 alsoshows that there is almost no difference between thechemical yields for a complete analysis and theadsorption yield. This means that additional chemi-cal losses from (i) stripping, (ii) digestion and (iii)

Figure 3.Kinetic uptake experiments: determination of sorption half-lives for U and Am.

Figure 4.Smoothed liquid scintillation spectrum of 239Pu and 244Cmobtained with Packard Tri-Carb 2550TR/AB.

Table 2.Chemical recoveries obtained from a complete analyticalprocedure.

4

Chemical Yield Investigations

Direct spike experiments: The counting efficien-cies were determined with radiolabeled spikedsolutions added to cocktail mixtures to simulateroutine chemical analysis. The cocktails were mea-sured each under two different discriminator set-tings, i.e., at the crossover point (125 ns) and at a

value of 140 ns. While the counting efficiencies wereclose to 100% at the crossover setting, high valuesof 95% were still obtained at the higher (140 ns)discriminator setting (see the “Analysis of CountingSensitivities” section).

Recoveries of Th, Pa, U, Pu, Am, Cm: The chemi-cal yield or recovery following a complete radio-chemical analysis was determined from the additionof radiospikes of known activity to a blank urinesample (Table 2). In particular, two steps were stud-ied to check on chemical recovery:

• The adsorption yield on the resin• The overall (total) chemical recovery

To obtain the reproducibility of the results, all spikeexperiments were repeated at least four times foreach radionuclide. The sorption yield on the resinwas by means of γ-spectrometry of the nuclide itselfor via decay or ingrowth of daughter nuclides.

Table 2 clearly indicates that actinide resin has anextremely strong affinity for all tested actinides evenfrom a very strong acidic urine solution with high saltcontent (average salt content 30 g/L). Table 2 alsoshows that there is almost no difference between thechemical yields for a complete analysis and theadsorption yield. This means that additional chemi-cal losses from (i) stripping, (ii) digestion and (iii)

Figure 3.Kinetic uptake experiments: determination of sorption half-lives for U and Am.

Figure 4.Smoothed liquid scintillation spectrum of 239Pu and 244Cmobtained with Packard Tri-Carb 2550TR/AB.

Table 2.Chemical recoveries obtained from a complete analyticalprocedure.4

Chemical Yield Investigations

Direct spike experiments: The counting efficien-cies were determined with radiolabeled spikedsolutions added to cocktail mixtures to simulateroutine chemical analysis. The cocktails were mea-sured each under two different discriminator set-tings, i.e., at the crossover point (125 ns) and at a

value of 140 ns. While the counting efficiencies wereclose to 100% at the crossover setting, high valuesof 95% were still obtained at the higher (140 ns)discriminator setting (see the “Analysis of CountingSensitivities” section).

Recoveries of Th, Pa, U, Pu, Am, Cm: The chemi-cal yield or recovery following a complete radio-chemical analysis was determined from the additionof radiospikes of known activity to a blank urinesample (Table 2). In particular, two steps were stud-ied to check on chemical recovery:

• The adsorption yield on the resin• The overall (total) chemical recovery

To obtain the reproducibility of the results, all spikeexperiments were repeated at least four times foreach radionuclide. The sorption yield on the resinwas by means of γ-spectrometry of the nuclide itselfor via decay or ingrowth of daughter nuclides.

Table 2 clearly indicates that actinide resin has anextremely strong affinity for all tested actinides evenfrom a very strong acidic urine solution with high saltcontent (average salt content 30 g/L). Table 2 alsoshows that there is almost no difference between thechemical yields for a complete analysis and theadsorption yield. This means that additional chemi-cal losses from (i) stripping, (ii) digestion and (iii)

Figure 3.Kinetic uptake experiments: determination of sorption half-lives for U and Am.

Figure 4.Smoothed liquid scintillation spectrum of 239Pu and 244Cmobtained with Packard Tri-Carb 2550TR/AB.

Table 2.Chemical recoveries obtained from a complete analyticalprocedure.

Figure 3. Kinetic uptake experiments: determination of sorption halflives for U and Am.

Figure 4. Smoothed liquid scintillation spectrum of 239Pu and 244Cmobtained with PerkinElmer Tri-Carb 2550TR/AB.

Table 2. Chemical recoveries obtained from a complete analytical procedure.

Recoveries of Th, Pa, U, Pu, Am, Cm: The chemical yield or recovery following a complete radiochemical analysis was determined from the addition of radiospikes of known activity to a blank urine sample (Table 2). In particular, two steps were studied to check on chemical recovery:

• The adsorption yield on the resin • The overall (total) chemical recovery

To obtain the reproducibility of the results, all spike experi-ments were repeated at least four times for each radio-nuclide. The sorption yield on the resin was by means of g-spectrometry of the nuclide itself or via decay or ingrowth of daughter nuclides.

Table 2 clearly indicates that actinide resin has an extremely strong affinity for all tested actinides even from a very strong acidic urine solution with high salt content (average salt content 30 g/L). Table 2 also shows that there is almost no difference between the chemical yields for a complete analysis and the adsorption yield. This means that additional chemical losses from (i) stripping, (ii) digestion and (iii) transfer into the liquid scintillation vial are insignificant. This method can hence be easily adopted to routine laboratory use.

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 (Ca2+ and Ra2+) in strong acidic medium as obtained by Horwitz, et al.10 Even in the presence of 2 M HCl solutions containing 1 M CaCl2, the uptake of the least efficiently sorbed species Am(III) remained considerably high (k' = 103). Since k' is lower for Ra2+ than Ca2+ (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 HNO3/H2SO4 mixtures (boiling point of sulfuric acid = 338 ˚C), the second naturally occurring component 210Po is efficiently eliminated since under acidic conditions at elevated temperatures, Po (probably present as Po-oxide in the ash) is volatile. Tracer experiments with 209Po(NO3)4 spike solutions indeed revealed repeatedly no detectable activity in the liquid scintillation cocktail.

Uptake studies with different resin additions: As dis-cussed above, the uptake coefficient of Am(III) decreases rapidly with acidity. Therefore, slight neutralization of the aqueous samples with NH4OH (following the oxidation step with HNO3) was attempted. However, when adding NH4OH to reduce the acidity of the solution from ≈ 2 M to 1 M HNO3, 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.

5

transfer into the liquid scintillation vial are insignifi-cant. This method can hence be easily adopted toroutine laboratory use.

All experiments were performed with 200 mg resinper 0.5 L sample and an extraction time of at leastfour hours (for kinetic studies see the followingsection entitled “Experiments on Uptake Kinetics”).Under these conditions only about 75% Am wasconsistently recovered, whereas the recoveries ofmost of the investigated actinides exceeded 90%.This discrepancy is most likely due to the fact that theresin uptake coefficient for Am(III) more rapidlydecreases with acidity compared to those actinidespresent in the tetra- or hexavalent state such asTh(IV), Pu(IV) or U(VI).10

Recovery of Ra: In contrast to the actinides, theuptake of Ra on actinide resin was found to be lessthan 5% (Table 2). This result is consistent with thelow sorption coefficients (k'-values) for the alkalineearth elements (Ca2+ and Ra2+) in strong acidicmedium as obtained by Horwitz, et al.10 Even in thepresence of 2 M HCl solutions containing 1 M CaCl2,the uptake of the least efficiently sorbed speciesAm(III) remained considerably high (k' = 103).Since k' is lower for Ra2+ than Ca2+ (k' <1), andaverage urine Ca/Ra ratios are extremely high, noadditions of Ca or Ba carrier are required for aroutine analysis.

Recovery of Po: Because oxidation of the strippedreagent fraction is performed under high tempera-tures using HNO3/H2SO4 mixtures (boiling point ofsulfuric acid = 338 °C), the second naturally occur-ring component 210Po is efficiently eliminated sinceunder acidic conditions at elevated temperatures, Po(probably present as Po-oxide in the ash) is volatile.Tracer experiments with 209Po(NO3)4 spike solutionsindeed revealed repeatedly no detectable activity inthe liquid scintillation cocktail.

Uptake studies with different resin additions: Asdiscussed above, the uptake coefficient of Am(III)decreases rapidly with acidity. Therefore, slight neu-tralization of the aqueous samples with NH4OH(following the oxidation step with HNO3) was at-tempted. However, when adding NH4OH to reducethe acidity of the solution from ≈ 2 M to 1 M HNO3,the solutions became black and opaque. An im-proved technique to obtain a higher extraction yieldis simply to increase the amount of actinide resin persame sample volume. The results for additions of0.4 g/L and 1 g/L are depicted in Figure 5. Almostquantitative extraction for all the actinides wereobtained when taking 1 g/L actinide resin.

Figure 5.Bar chart showing the chemical recoveries obtained fromtwo different resin additions.

Experiments on uptake kinetics: To obtain thetimes required for sorption equilibrium at a steadystate, the uptake kinetics were studied for tri- andhexavalent species using Am(III) and U(VI) tracersolutions. For these investigations, aliquots wereprepared as explained in the previous section. Thistime, however, aliquots were spiked with identicalactivity concentrations and the extraction wasinterrupted at times given in Figure 3. Very rapiduptake was observed and in about two hours steadystate conditions were obtained independently of theamount of added resin.

If the sorption process follows first order kinetics,the data should plot on a straight line in a semi-logdiagram with the remaining activity in solution plot-ted versus the exposure time (Figure 3). In this casethe sorption exponent ksorp can be extracted from therelation:

asolution = e-ksorp . t

(with asolution = activity in solution) and via regressionanalysis of the data. A more comprehensiveapproach is the use of the sorption half-lives(i.e., T1/2 = ln2/k). Very short half-lives of only eightand 20 minutes were calculated for U and Am,respectively using this approach.

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 investiga-tions, aliquots were prepared as explained in the previous section. This time, however, aliquots were spiked with identical activity concentrations and the extraction was inter-rupted 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 ksorp can be extracted from the relation:

asolution = e – ksorp • t

(with asolution = activity in solution) and via regression analysis of the data. A more comprehensive approach is the use of the sorption half-lives (i.e., T1/2 = ln2/k). Very short half-lives of only eight and 20 minutes were calculated for U and Am, respectively using this approach.

6

Analysis of Counting Sensitivities

Optimum a/b discriminator settings: For gross a/b counting systems, two parameters are essential to determine the sensitivity of a radioassay: background count rate (b) and counting efficiency (E), which can be expressed as figure of merit or E2/B.20 To reduce background scatter, misclas-sification of b pulses counted as a has been minimized by optimizing the pulse decay and afterpulse analysis features of the Tri-Carb with alpha/beta discrimination.7,8 As shown in Figure 6, low a and b misclassification (0.6%) resulted at the optimum pulse decay discriminator (PDD) setting of 125. Optimum E2/B values were, however, obtained for a slightly higher PDD setting of 140. At this value, the b spill (0.1%) is extremely low (hence significantly reducing the a back-ground count rate), while the loss in counting efficiency due to counting some a pulses as b is minimal.

Comparison of detection limits: The lower limit of detection (LLD) at the 95% confidence probability level was calculated from analysis of blank samples using the equation as given by Seymour, et al.21

where (K) = 1.64 = statistical value for a confidence interval of 95%; (I0) = total background counts in time t; (t) = counting time; (Yi) = chemical recovery; (E) = counting or detector efficiency; (Vs) = sample volume; and (μ) = attenuation coefficient.

6

Analysis of Counting Sensitivities

Optimum α/βα/βα/βα/βα/β discriminator settings: For grossα/β counting systems, two parameters are essentialto determine the sensitivity of a radioassay: back-ground count rate (B) and counting efficiency (E),which can be expressed as figure of merit or E2/B.20

To reduce background scatter, misclassification of βpulses counted as α has been minimized by optimiz-ing the pulse decay and afterpulse analysis featuresof the Packard Tri-Carb 2500TR/AB.7,8 As shownin Figure 6, low α and β misclassification (0.6%)resulted at the optimum pulse decay discriminator(PDD) setting of 125. Optimum E2/B values were,however, obtained for a slightly higher PDD settingof 140. At this value, the β spill (0.1%) is extremelylow (hence significantly reducing the α backgroundcount rate), while the loss in counting efficiency dueto counting some α pulses as β is minimal.

Comparison of detection limits: The lower limit ofdetection (LLD) at the 95% confidence probabilitylevel was calculated from analysis of blank samplesusing the equation as given by Seymour, et al.:21

where (K) = 1.64 = statistical value for a confidenceinterval of 95%; (I0) = total background counts intime t; (t) = counting time; (Yi) = chemical recovery;(E) = counting or detector efficiency; (Vs) = samplevolume; and (µ) = attenuation coefficient.

Table 3.Set of values used for the calculation of the LLDs for 239Pu.

The set of parameters used for the calculation of theLLD is given in Table 3. The other parameters werekept either constant (i.e., Vs = 0.5 L) or were notrelevant (µ). However, it has to be noted that incontrast to procedures based on LSC, µ can only beomitted when almost weightless sample discs areproduced. If that is not the case, absorption of αradiation in the sample source itself has to beconsidered seriously.

In Figure 7, three methods are compared; two proce-dures based on α/β-LSC (previous work of Eikenberg,et al.)6 and a method2 developed for low level gas-flow proportional counting (GPC). Although a verylow background count rate of only 0.04 CPM was

Figure 7.Evolution of the lower limit of detection for three methodsbased on GPC and LSC.

Figure 6.α/β crossover curves as function of PDD setting obtainedwith 241Am and 36Cl.

6

Analysis of Counting Sensitivities

Optimum α/βα/βα/βα/βα/β discriminator settings: For grossα/β counting systems, two parameters are essentialto determine the sensitivity of a radioassay: back-ground count rate (B) and counting efficiency (E),which can be expressed as figure of merit or E2/B.20

To reduce background scatter, misclassification of βpulses counted as α has been minimized by optimiz-ing the pulse decay and afterpulse analysis featuresof the Packard Tri-Carb 2500TR/AB.7,8 As shownin Figure 6, low α and β misclassification (0.6%)resulted at the optimum pulse decay discriminator(PDD) setting of 125. Optimum E2/B values were,however, obtained for a slightly higher PDD settingof 140. At this value, the β spill (0.1%) is extremelylow (hence significantly reducing the α backgroundcount rate), while the loss in counting efficiency dueto counting some α pulses as β is minimal.

Comparison of detection limits: The lower limit ofdetection (LLD) at the 95% confidence probabilitylevel was calculated from analysis of blank samplesusing the equation as given by Seymour, et al.:21

where (K) = 1.64 = statistical value for a confidenceinterval of 95%; (I0) = total background counts intime t; (t) = counting time; (Yi) = chemical recovery;(E) = counting or detector efficiency; (Vs) = samplevolume; and (µ) = attenuation coefficient.

Table 3.Set of values used for the calculation of the LLDs for 239Pu.

The set of parameters used for the calculation of theLLD is given in Table 3. The other parameters werekept either constant (i.e., Vs = 0.5 L) or were notrelevant (µ). However, it has to be noted that incontrast to procedures based on LSC, µ can only beomitted when almost weightless sample discs areproduced. If that is not the case, absorption of αradiation in the sample source itself has to beconsidered seriously.

In Figure 7, three methods are compared; two proce-dures based on α/β-LSC (previous work of Eikenberg,et al.)6 and a method2 developed for low level gas-flow proportional counting (GPC). Although a verylow background count rate of only 0.04 CPM was

Figure 7.Evolution of the lower limit of detection for three methodsbased on GPC and LSC.

Figure 6.α/β crossover curves as function of PDD setting obtainedwith 241Am and 36Cl.

6

Analysis of Counting Sensitivities

Optimum α/βα/βα/βα/βα/β discriminator settings: For grossα/β counting systems, two parameters are essentialto determine the sensitivity of a radioassay: back-ground count rate (B) and counting efficiency (E),which can be expressed as figure of merit or E2/B.20

To reduce background scatter, misclassification of βpulses counted as α has been minimized by optimiz-ing the pulse decay and afterpulse analysis featuresof the Packard Tri-Carb 2500TR/AB.7,8 As shownin Figure 6, low α and β misclassification (0.6%)resulted at the optimum pulse decay discriminator(PDD) setting of 125. Optimum E2/B values were,however, obtained for a slightly higher PDD settingof 140. At this value, the β spill (0.1%) is extremelylow (hence significantly reducing the α backgroundcount rate), while the loss in counting efficiency dueto counting some α pulses as β is minimal.

Comparison of detection limits: The lower limit ofdetection (LLD) at the 95% confidence probabilitylevel was calculated from analysis of blank samplesusing the equation as given by Seymour, et al.:21

where (K) = 1.64 = statistical value for a confidenceinterval of 95%; (I0) = total background counts intime t; (t) = counting time; (Yi) = chemical recovery;(E) = counting or detector efficiency; (Vs) = samplevolume; and (µ) = attenuation coefficient.

Table 3.Set of values used for the calculation of the LLDs for 239Pu.

The set of parameters used for the calculation of theLLD is given in Table 3. The other parameters werekept either constant (i.e., Vs = 0.5 L) or were notrelevant (µ). However, it has to be noted that incontrast to procedures based on LSC, µ can only beomitted when almost weightless sample discs areproduced. If that is not the case, absorption of αradiation in the sample source itself has to beconsidered seriously.

In Figure 7, three methods are compared; two proce-dures based on α/β-LSC (previous work of Eikenberg,et al.)6 and a method2 developed for low level gas-flow proportional counting (GPC). Although a verylow background count rate of only 0.04 CPM was

Figure 7.Evolution of the lower limit of detection for three methodsbased on GPC and LSC.

Figure 6.α/β crossover curves as function of PDD setting obtainedwith 241Am and 36Cl.

6

Analysis of Counting Sensitivities

Optimum α/βα/βα/βα/βα/β discriminator settings: For grossα/β counting systems, two parameters are essentialto determine the sensitivity of a radioassay: back-ground count rate (B) and counting efficiency (E),which can be expressed as figure of merit or E2/B.20

To reduce background scatter, misclassification of βpulses counted as α has been minimized by optimiz-ing the pulse decay and afterpulse analysis featuresof the Packard Tri-Carb 2500TR/AB.7,8 As shownin Figure 6, low α and β misclassification (0.6%)resulted at the optimum pulse decay discriminator(PDD) setting of 125. Optimum E2/B values were,however, obtained for a slightly higher PDD settingof 140. At this value, the β spill (0.1%) is extremelylow (hence significantly reducing the α backgroundcount rate), while the loss in counting efficiency dueto counting some α pulses as β is minimal.

Comparison of detection limits: The lower limit ofdetection (LLD) at the 95% confidence probabilitylevel was calculated from analysis of blank samplesusing the equation as given by Seymour, et al.:21

where (K) = 1.64 = statistical value for a confidenceinterval of 95%; (I0) = total background counts intime t; (t) = counting time; (Yi) = chemical recovery;(E) = counting or detector efficiency; (Vs) = samplevolume; and (µ) = attenuation coefficient.

Table 3.Set of values used for the calculation of the LLDs for 239Pu.

The set of parameters used for the calculation of theLLD is given in Table 3. The other parameters werekept either constant (i.e., Vs = 0.5 L) or were notrelevant (µ). However, it has to be noted that incontrast to procedures based on LSC, µ can only beomitted when almost weightless sample discs areproduced. If that is not the case, absorption of αradiation in the sample source itself has to beconsidered seriously.

In Figure 7, three methods are compared; two proce-dures based on α/β-LSC (previous work of Eikenberg,et al.)6 and a method2 developed for low level gas-flow proportional counting (GPC). Although a verylow background count rate of only 0.04 CPM was

Figure 7.Evolution of the lower limit of detection for three methodsbased on GPC and LSC.

Figure 6.α/β crossover curves as function of PDD setting obtainedwith 241Am and 36Cl.

Figure 6. a/b crossover curves as function of PDD setting obtained with 241Am and 36Cl.

Table 3. Set of values used for the calculation of the LLDs for 239Pu.

Figure 7. Evolution of the lower limit of detection for three methodsbased on GPC and LSC.

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 239Pu 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., 234U, 235U, 238U). 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 239Pu, 240Pu, 241Am and 244Cm. 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 (160-280 keV) as well as for the transuranium nuclides (260-380 keV).

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., Vs = 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 method2 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 (E2/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 300-400 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 234U and 238U or, as shown in Figure 4, between transuranium nuclides such as 239Pu and 244Cm (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

7

taken to calculate the LLD using low level GPCcounters, the new procedure based on • •/• •-LSC yieldsconsiderably lower LLD values. This result is theconsequence of the very high counting efficienciesand chemical recoveries. If, in particular, LSCis carried out using a small window for analysisof a special group of actinides (see the “• • peakresolution and liquid scintillation quench” paragraphbelow), the background decreases to values about0.05 CPM. This yields an extraordinary high figureof merit (E2/B) of 180,000 or a detection limit of1.5 mBq/L in a 500 minute counting interval.

• •• •• •• •• • peak resolution and liquid scintillation quench:It is well known that • • peak resolution using LSC ispoor in comparison to • • spectrometry and hence• •••• •-LSC systems are used mainly as gross counters.Nevertheless, if • • pulse stretching scintillatorsare used, FWHM values of 300-400 keV can beobtained.22 Since, for the current procedure, a nearorganic cocktail mixture is prepared, the • • peakresolution becomes fairly high (400 keV or 40 keVon a liquid scintillation scale). This allows peakseparation between 234U and 238U or, as shown inFigure 4, between transuranium nuclides such as239Pu and 244Cm (• •E = 600 keV). Two observationsare of interest. First, the peaks are symmetricallyshaped and simple Gaussian-based fitting proce-dures (without additional terms for peak tailing) aresufficient for fitting of overlapping peaks. Second,there is a significant shift of the • • energy using LSCwith respect to the real emission energy.

This shifting phenomenon is due to ionization quenchbecause the • • particles dissipate their energy over avery small distance causing less excited niveaus inthe orbitals of the scintillator targets.23 However, theshift in energy from quenching remains constant fora fixed cocktail mixture. The relation between thetrue • • emission energy and the liquid scintillationquenched output is demonstrated in Figure 8 for theprocedure given here and an aqueous cocktail.6 It isinteresting to note, that for a set of • • emitters withdifferent energies, the ratio between both energyscales is highly linear. Although the aqueous cocktailis quenched to a higher degree, regression analysesyielded 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 mul-tichannel analyzer (MCA). Indeed stability testsusing 239Pu spiked cocktails revealed identical peakpositions which scattered less then 15 keV on theliquid scintillation scale for samples produced withinone year. This fact is helpful to distinguish betweentwo major groups of actinides which are of interestfor in vitro measurements. As shown in Figure 8, allanthropogenic transuranium nuclides are character-ized by emission of higher • • energies compared toall natural uranium isotopes (i.e., 234U, 235U, 238U).For in vitro screening, a distinction between thesegroups is often reasonable. For instance, monitoringof employees involved in uranium mining should belimited to uranium isotopes, whereas in nuclearreprocessing plants or hot laboratories handling spentfuel elements, radiation hazards may arise predomi-nantly from incorporation of 239Pu, 240Pu, 241Am and244Cm. Therefore, instead of counting gross • • overa wide range of energy, a small window of only120 keV can be taken for the group of uraniumisotopes (160-280 keV) as well as for the transura-nium nuclides (260-380 keV).

Figure 8.Relations between the true emission energy and liquidscintillation quenched • • energies of actinides.

Figure 8. Relations between the true emission energy and liquid scintillation quenched a energies of actinides.

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, 461-472.

3. Horwitz, E., Ph Dietz, M.L., Nelson, D.M., LaRosa, J.J. and Fairman, W.D. (1990) Anal. Chim. Acta Vol. 238, 263-271.

4. Salonen, L. (1993) Sci. Tot. Environ. Vol. 130/131, 23-35.

5. Bickel, M., Möbius, S., Kilian, F. and Becker, H. (1992) Radiochim. Acta Vol. 57, 141-151.

6. Eikenberg, J., Fiechtner, A., Ruethi, M. and Zumsteg, I. Liquid Scintillation Spectrometry 1994, Radiocarbon 1996, 283-292.

7. Passo, C.J. and Kessler, M.J. (1992) Packard Instrument Company Publication, Report PBR0012, 8.

8. Passo, C.J. and Kessler, M.J. Liquid Scintillation Spectrometry 1992, Radiocarbon 1993, 51-57.

9. Shiraishi, K., Yamamoto, M., Yoshimizu, K., Igarashi, Y. and Ueno, K. (1994) Health Phys. Vol. 66, 30-35.

10. 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) 407-412.

12. Karpas, et al., Health Phys. 71 (1996) 879-885.

13. Dang, et al., Health Phys. 57 (1989) 393-396.

14. Dahlheimer and Henrichs, Rad. Prot. Dos. 53 (1994) 207-209.

15. Fisenne, et al., Health Phys. 53 (1987) 357-363.

16. ICRP Publication. 23 (1975).

17. Shiraishi, et al., Health Phys. 66 (1994) 30-35.

18. Fellmann, et al., Health Phys. 57 (1989) 615-621.

19. 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, 586-593.

21. Seymour, R., Sergent, F., Knight, K. and Kyker, B. (1992) Radioact. Radiochem. 3, 14-27.

22. Yu, Y.F., Salbu, B., Bjornstad, H.E. and Lien, H.J. (1990) Radioanal. Nucl. Chem. Lett. 145, 345-353.

23. Horrocks, D. (1974) Academic Press. New York-London 346.

For a complete listing of our global offices, visit www.perkinelmer.com/ContactUs

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. 009599A_01 Printed in USA April 2011

PerkinElmer, Inc. 940 Winter Street Waltham, MA 02451 USA P: (800) 762-4000 or (+1) 203-925-4602www.perkinelmer.com