Anal. Chem. 1997, 69, 2275-2282
Determination of Uranium, Thorium, Plutonium, Americium, and Curium Ultratraces by Photon Electron Rejecting r Liquid Scintillation Nicolas Dacheux* and Jean Aupiais
Commissariat a` l’Energie Atomique, Centre d’Etudes de Bruye` res-le-chaˆ tel, 91680 Bruye` res-le-chaˆ tel, France
The determination of activities of thorium, uranium, plutonium, americium, and curium at very low levels has been performed by a new r liquid scintillation system (PERALS, name registered to Ordela, Inc.). The limit of detection has been determined for these nuclides with calculated values often lower than those obtained by other methods, like ICPMS/HP/Mistral, time-resolved laserinduced spectrofluorometry, and r spectrometry. All the results obtained show that the PERALS system is a promising method for the determination of these activities at very low levels. However, its energy resolution is inferior in comparison to that obtained by r spectrometry. For this reason, we have developed a process for separation of the five actinides as quickly and easily as possible. For each actinide, the conditions required to obtain optimal extraction yields and a complete separation have been determined. It is possible to perform the separation in only six extraction steps and to measure activities as low as a few millibecquerels per liter independently. This process has been applied with success to French granitic mineral or doped water and to complex media (biological samples like urines). In this latter case, the extraction recoveries are not quantitative, and it is necessary to determine the recovery yields by labeling with spikes like 230Th, 232U, 236Pu, 248Cm, and 148Gd. Currently, the measurement of low concentrations of actinides has become very important in the field of environmental management.1-10 The detection of very low activities of actinides like thorium, uranium, plutonium, americium, and curium has * Corresponding author. Current address: Groupe de Radiochimie, Institut de Physique Nucleaire, Universite´ Paris-Sud, 91406 Orsay Cedex, France. E-mail:
[email protected]. (1) Cadieux, J. R.; Reboul, S. H. Radioact. Radiochem. 1996, 7, 30-34. (2) Cadieux, J. R.; Clark, S.; Fjeld, R. A.; Reboul, S.; Sowder, A. Nucl. Instrum. Methods Phys. Res. A 1994, 353, 534-538. (3) Zhu, Y. J.; Yang, D. Z. J. Radioanal. Nucl. Chem., Art. 1995, 194-I, 173175. (4) Yang, D. Z.; Gou, Y. F.; Zhu, Y. J. J. Radioanal. Nucl. Chem., Art. 1995, 194-I, 177-183. (5) Rao, R. R.; Cooper, E. L. J. Radioanal. Nucl. Chem., Art. 1995, 197-I, 133148. (6) Leyba, J. D.; Vollmar, H. S.; Fjeld, R. A.; Devol, T. A.; Brown, D. D.; Cadieux, J. R. J. Radioanal. Nucl. Chem., Art. 1995, 194-2, 337-344. (7) Bourlat, Y. ; Millies-Lacroix, J. C. ; Nazard, R. J. Radioanal. Nucl. Chem., Art. 1995, 197-2, 387-408. (8) Cadieux, J. R. Presented at the 7th symposium on radiation measurements and applications, University of Michigan, Ann Arbor, MI, May 21-24, 1990. (9) McDowell, W. J. NAS-NS 3116, Nuclear Science Series on Radiochemical Techniques; U.S. Government Printing Office: Washington, DC, 1986; pp 88-89. S0003-2700(96)01209-7 CCC: $14.00
© 1997 American Chemical Society
been performed in various samples by several methods (radioanalysis, electroanalysis, mass spectrometry, etc.). Among them, R spectrometry using grid ionization chambers and, now, semiconductor detectors is one of the most important and sensitive techniques for the determination of these very low activities. Unfortunately, it is often necessary to measure the activity in solid samples. Various complicated procedures including steps such as precipitations, evaporations, elutions, and filtrations are often required for the separation of the R emitters contained in the matrix. Moreover, the self-absorption of R emitters in the sample matrix causes additional problems. Low counting efficiency due to this self absorption results in poor counting statistics, which necessitates very long counting times for the determination of very low activities. Furthermore, gradual contamination of the R detectors by sample particles deposited during air removal by vacuum pumping is often observed and decreases the sensivity of the technique. Five years ago, a new analytical method, PERALS spectrometry (acronym for photon electron rejecting alpha liquid scintillation, PERALS name registered to Ordela, Inc.) appeared. This analysis of R nuclides combines chemical separation by liquid-liquid extraction with measurement of R activity by liquid scintillation in the same procedure. After extraction from aqueous solution, actinides are presented into an organic phase which contains a complexing extractant, an energy transfer reagent, and a lightemitting fluor.2,8,9,11 The organic and aqueous phases are separated, and oxygen is purged from the organic phase to improve the resolution of the PERALS spectrometer. This technique uses a nearly total rejection of β emission (99.95%) by means of a pulse shape discrimination (PSD). Ambient γ ray activity produces photoelectrons, which creates a background. PSD reduces this background. It also eliminates interferences from β emitters which are coextracted with R emitters or produced by decay. The theory of liquid-liquid extraction leads to the definition of a percentage of recovery P, which depends on the coefficient of distribution D and the volumes of organic (V h ) and aqueous phases (V). Furthermore, the D value is very sensitive to the acidity, to the aqueous solution, and to the ions present. Thus, for five actinides (Th, U, Pu, Am, and Cm), we have studied the variation of the percentage of recovery in terms of pH in various acidic media (HNO3, HCl, and H2SO4) using several extractant(10) Boatner, L. A.; Sales, B. C. In Radioactive waste forms for the future; Lutze, W., Ewing, R. C., Eds.; Elsevier Science Publishing New York, 1988; pp 556-558. (11) McDowell, W. J. NAS-NS-3116; Technical Information Center, U.S. Department of Energy: Oak Ridge, TN 1986.
Analytical Chemistry, Vol. 69, No. 13, July 1, 1997 2275
scintillator mixtures (URAEXR, THOREXR, and ALPHAEXR). As noted previously in the literature, the R energy resolution of the PERALS spectrometer is rather poor in comparison to that of R spectrometry. Thus, we have been compelled to separate the nuclides before R liquid scintillation counting. The conditions required to obtain optimal extraction for each actinide have been determined. We have applied these results to develop a separation process that is as easy and fast as possible using only a few liquidliquid extraction steps. This separation process has been applied to distilled water, French granitic mineral water (like Badoit and Volvic), and complex media (urines, soils, etc.) doped with nuclides. This process allows, after only six liquid-liquid extraction steps, the complete separation of uranium, thorium, plutonium, americium, and curium and the independent measurement of the R activity of each nuclide in the solution down to a few millibecquerels per liter. THEORY Theory of Liquid-Liquid Extraction. The extraction of a cation into the organic phase depends on the nature of the extractant, the solvation of the cation, and all the anions present in solution and involved in the formation of complexes.12,13 The theory of liquid-liquid extraction defines a distribution coefficient D, which represents the performance of the extraction step. As an example, we report here the description of the extraction of a cation Mn+ by the HDEHP molecule (which is the complexing extractant in the ALPHAEXR extractant-scintillator mixture). It can be described as follows:14
Mn+ + n(HA)2 T M(HA2)n + nH+
(1)
The constant of equilibrium K and the coefficient of distribution D can be written + n
K)
[M(HA2)n] [H ] [Mn+]
[(HA)2]n (2)
D)
[M(HA2)n] [Mn+]
Then, the relationship between K and D is
D)K
[(HA)2]n [H+]n
(3)
Moreover, the coefficient of distribution can be written
D)
Av A V ) Av A V
(4)
(12) Qureshi, I. H. ; McClendon, L. T. ; LaFleur, P. D. Radiochemical separations in modern trends in activation analysis; NBS Special Publication 312; U.S. National Bureau of Standards: Washington, DC, 1969, Vol. I, pp 666-671. (13) Baes, C. F.; Zingaro, R. A.; Coleman, C. F. J. Phys. Chem. 1958, 62, 129. (14) Adloff, J. P. ; Guillaumont, R. Fundamentals of Radiochemistry; CRC Press: Boca Raton, FL, 1993; p 358.
2276
Analytical Chemistry, Vol. 69, No. 13, July 1, 1997
where Av ) A/V and Av ) A/V are respectively the activities per unit volume in aqueous and organic phases (A and A h are the corresponding absolute activities; V and V h are the volumes of aqueous and organic phases). From these relations, it is clear that, at constant extractant concentration, the distribution coefficient depends on the acidity of the aqueous phase (eq 3) and on the volume ratio V h /V (eq 4). For this reason, this ratio has been kept constant in the first series of experiments (equal to 1/5) for all the extraction curves presented in Figures 1-5. For the separation summary scheme (Figure 6), the ratio of volumes used depends on the initial volume of solution treated: they are indicated for each step of extraction. The associated percentage of recovery P is
P)
A A × 100 × 100 ) Atotal A+A
(5)
which corresponds to
P)
D(V/V) 1 + D(V/V)
× 100
(6)
Determination of the Limit of Detection by the PERALS Spectrometer. For R and β risks equal to 2.5%, the limit of detection γd can be defined as follows:15,16
4 γd ) 2(∆c)s ) (1 + x1 + 2Xbackground) t
(7)
where (∆c)s is the threshold of detection (the signal is not detected if it is below (∆c)s), Xbackground is the background count number, and t is the counting time (identical for the background and the sample). EXPERIMENTAL SECTION Extraction Process. The R liquid scintillation apparatus (PERALS spectrometer) is produced by Ordela Inc. Two kinds of extractive scintillator mixtures supplied by Etrac were used to perform the liquid-liquid extraction: ALPHAEXR, in which the extractive molecule is bis(2-ethylhexyl)phosphoric acid (HDEHP), and THOREXR and URAEXR, containing the primary amine 1-nonyldecylamine and the tertiary amine tri-n-octylamine as extractive molecules, respectively. All the other reagents used during the extraction procedure were of analytical grade from Prolabo, Aldrich, and Merck. Extraction processes were performed in Pyrex tubes and bottles because of their mechanical stability in the shaking and centrifugation steps. Each extraction step is performed according to the following procedure. The pH of the aqueous phase is first adjusted. A gentle shaking of a known volume of aqueous solution (up to 250 mL) containing the nuclide with 1.2-3.0 mL of an extractantscintillator mixture is then performed for 5-30 min in a rotative mixer (Turbula). The separation of both phases is performed by centrifugation at 1600 rpm in a Jouan G 4.11 centrifuge for 1530 min. The organic phase is taken off. One milliliter is (15) Limite de de´tection d’un signal dans le bruit de fond, application aux mesures de radioactivite´ par comptage Report CEM R-5201; CEN: Grenoble, France, 1983. (16) Neuilly, M. Statistique applique´ e a` l’exploitation des mesures Masson: Paris, 1986; pp 113-117.
Table 1. Limits of Detection of the PERALS Spectrometer limit of detectiona (M) radionuclide
3 days
10 days
232Th
1× 1 × 10-14 2 × 10-10 1 × 10-13
3× 5 × 10-15 9 × 10-11 4 × 10-14
234U 238U 237Np a
limit of detectiona (M)
10-9
radionuclide
10-10
238Pu
3 days
10 days
5× 1 × 10-15 2 × 10-17 9 × 10-19
2 × 10-18 5 × 10-16 9 × 10-18 4 × 10-19
10-18
239Pu 241Am 244Cm
Obtained for 250 mL of aqueous solution.
Table 2. Limits of Detection of the PERALS Spectrometer, ICPMS/HP/Mistral, TRLIF, and r Spectrometry limit of detection (M) radionuclide
PERALSa
TRLIF20-23
ICP/MS24
R spectrometry 24,b
232Th
1 × 10-9 1 × 10-14 2 × 10-10 1 × 10-13 5 × 10-18 1 × 10-15 2 × 10-17 4 × 10-16 9 × 10-19 2 × 10-14
c 5 × 10-12 5 × 10-12 c c c 5 × 10-9 5 × 10-9 4 × 10-13 4 × 10-13
2 × 10-13 5 × 10-15 2 × 10-13 5 × 10-15 d 5 × 10-15 5 × 10-15 5 × 10-15 5 × 10-15 5 × 10-15
2 × 10-9 2 × 10-14 4 × 10-10 2 × 10-13 7 × 10-18 2 × 10-15 3 × 10-17 6 × 10-16 2 × 10-18 3 × 10-14
234U 238U 237Np 238Pu 239Pu 241Am 243Am 244Cm 248Cm
a Obtained for 250 mL of aqueous solution and 3 days of counting time. b Obtained for 100 mL of aqueous solution and 3 days of counting time during the first utilization. c Not mentioned or ion not fluorescent in the FLUO 2001 range.20 d The presence of 238U in the reagents does not allow the measure of low concentrations of 238Pu.24
transferred into a 10 mm × 75 mm culture tube and sparged with toluene-saturated argon (purity 99.9999%) for 5 min. Determination of Percentages of Recovery. Standardized actinide solutions in 4 M HNO3 were used. The specific activity of each solution was determined by R spectrometry and mass spectrometry. The obtained values were 4.64, 5.78, 1.85, 1.62, and 1.84 Bq/g for 232Th, 234U/238U, 239Pu, 241Am, and 244Cm, respectively. Diluted solutions were used for the determination of very low activities (1-100 mBq/L). For all the extraction curves presented in this paper (Figures 1-5), the determination is carried out using 6 mL of aqueous solution and 1.2 mL of extractive scintillator mixture (corresponding to a volume ratio V h /V equal to 1/5). In contrast, all the recovery yields mentioned during the separation process (Figure 6) have been obtained for real conditions of extraction (e.g., volumes of aqueous solution up to 250 mL). RESULTS AND DISCUSSION Limit of Detection of the PERALS Spectrometer. We have studied the evolution of the activity measured in solution in terms of that introduced for 241Am. The results obtained have shown that it is perfectly linear in the 10-3-103 Bq/L range, even for large volumes of aqueous solution (up to 250 mL) and after only one or two steps of extraction.19 According to the eq 7, the limits of detection of the PERALS spectrometer were determined for counting times equal to 3 and 10 days taking into account volumes of organic and aqueous phases equal to 1.5 and 250 mL, respectively. The values obtained for several nuclides are gathered in Table 1. They are in good agreement with those mentioned in the literature.2,11,17-19 Moreover, they are very low, (17) Brown, D. D.; Fjeld, R. A.; Cadieux, J. R. Presented at the 39th Annual Conference on Bioassay, Colorado Springs, CO, 1993. (18) Aupiais, J. J. Radioanal. Nucl. Chem. Art. Accepted. (19) Dacheux, N.; Aupiais, J. To be published.
which indicates that PERALS spectrometry is a very promising technique for the detection of very low activities. Several techniques are usually used for the determination of actinides in solution at the trace level. Among them, time-resolved laser-induced spectrofluorometry (TRLIF) can be applied for fluorescent ions, e.g., for three actinides uranium (in the UO22+ form),20,21 americium,20,22 and curium.20,23 The limits of detection usually mentioned in the literature for this method are gathered in Table 2 as well as those obtained by ICPMS (HP/Mistral apparatus) and R spectrometry. The detection limits achieved with the ICPMS/HP/Mistral and TRLIF are obtained for 1 mL of solution: they are independent to the volume of the solution sample used. The values reported for PERALS and R spectrometries have been calculated for 3 days of counting time. From Table 2, it is clear that the use of ICPMS is an advantage for the measurement of long decay time nuclides (232Th, 234U, 238U, 237Np): the limits of detection published are lower than those obtained by all other methods. For other nuclides with shorter decay times (like plutonium, americium, and curium isotopes), it seems better to use the PERALS system as well as R spectrometry. Indeed, the values reported in the Table 2 are always of the same order of magnitude or lower than those calculated for TRLIF and ICPMS. Nevertheless, the limits of detection calculated for R spectrometry correspond to performances during the first utilization. Indeed, a contamination of the detector is always observed (20) Moulin, C.; Decambox, P.; Mauchien, P. J. Phys. IV, Collect. C7, Suppl. J. Phys. III 1991, 1, 677-680. (21) Moulin, C.; Beaucaire, C.; Decambox, P.; Mauchien, P. Anal. Chim. Acta 1990, 238, 291-296. (22) Thouvenot, P.; Hubert, S.; Moulin, C.; Decambox, P.; Mauchien, P. Radiochim. Acta 1993, 61, 15-21. (23) Moulin, C.; Decambox, P.; Mauchien, P. Anal. Chim. Acta 1991, 254, 145151. (24) Chiappini, R.; Taillade, J. M.; Bre´bion, S. J. Anal. At. Spectrom. In press.
Analytical Chemistry, Vol. 69, No. 13, July 1, 1997
2277
Figure 1. Yields of extraction of (9) thorium, (1) uranium, (O) plutonium, (×) americium, and ([) curium by ALPHAEXR in nitric acid.
Figure 2. Yields of extraction of (9) thorium, (1) uranium, (O) plutonium, (×) americium, and ([) curium by ALPHAEXR in hydrochloric acid.
during the vacuum process, which leads to an increase of the background and then of the detection limit of this technique over the time of utilization. This problem does not occur for the PERALS system. Furthermore, the PERALS procedure is easier and faster than all the chemical processes required during the preparation of the solid samples for R spectrometry (precipitation, filtration, separation, evaporation, etc.). Thus, it is often better and faster to use PERALS spectrometry, even if its energy resolution (in the range of 200-350 keV) is worse than that of R spectrometry (about 10-30 keV). Study of the Extraction of Thorium, Uranium, Plutonium, Americium, and Curium. Taking into account a volume ratio V h /V equal to 1/5 (e.g., 1.2 mL of organic phase/6 mL of aqueous phase), we have determined the percentage of recovery P as a function of pH for each actinide (thorium, uranium, plutonium, americium, and curium) in various acidic media (hydrochloric, nitric, and sulfuric acids) and using several extractant-scintillator mixtures (URAEXR, THOREXR, and ALPHAEXR). Main results are reported in Figures 1-5. The extraction of all actinides by ALPHAEXR in nitric and hydrochloric acids is presented in Figures 1 and 2, respectively. For trivalent actinides, the recovery yield is near 100% in both media for pH higher than 2.4. These results are in good 2278 Analytical Chemistry, Vol. 69, No. 13, July 1, 1997
Figure 3. Yields of extraction of (9) thorium, (1) uranium, and (O) plutonium by ALPHAEXR in sulfuric acid.
Figure 4. Yields of extraction of (9) thorium, (1) uranium, and (O) plutonium by URAEXR in sulfuric acid.
Figure 5. Yields of extraction of (9) thorium, (1) uranium, and (O) plutonium by THOREXR in sulfuric acid.
agreement with those already published.1,25,26 When the aqueous solution is more acidic, we can observe a very quick decrease of the recovery yield (americium and curium are not extracted for pH lower than 1.3). For americium and curium, no difference is observed between nitric and hydrochloric acid media. To the contrary, the extraction (25) Braun, T.; Ghersini, G. Extraction Chromatography; Elsevier: Amsterdam, 1975; p 89. (26) Colemen, G. H. NAS-NS 3058, Nuclear Science Series on Radiochemical Techniques; U.S. Government Printing Office: Washington, DC, 1965; p 43.
Figure 6. Synoptic scheme of the separation process.
in HCl is always lower than that obtained in HNO3 for the other actinides studied (U, Th, and Pu). The aim of this study is to develop a fast and easy separation of actinides: all the recovery yields must be near 100%. For this reason, we prefer the use of extractions by ALPHAEXR in nitric acid, as shown in the synoptic scheme of the separation process (Figure 6). For tetravalent actinides like Pu and Th, the extraction is rather good from pH 3 to 5M HNO3 (recovery is in the range 95-100%). Contrarily, uranium is hexavalent in our conditions of extraction: the extraction is complete in the range 10-5-0.5 M HNO3 and decreases slightly at higher acid concentrations; only 70% of uranium is extracted when aqueous solution becomes more concentrated than 1 M HNO3. We have noticed every time a third phase between the aqueous and organic phases for pH beyond 5 and the decrease of the extraction yield down to 40-50%. The separation of uranium from tetravalent actinides is generally performed by means of several extractions by THOREXR or URAEXR in sulfuric media because ALPHAEXR is not selective enough. Thus, the extraction of thorium, uranium, and plutonium by ALPHAEXR (Figure 3), URAEXR (Figure 4), and THOREXR (Figure 5) has been also studied in sulfuric acid. As shown in Figure 3, the extraction by ALPHAEXR in H2SO4 media is not good: the recovery yields are often low and do not allow an efficient and quick separation. The recovery yields of thorium and uranium quickly decrease in the range 10-2-5 M H2SO4. Moreover, the yield of extraction is lower than 80% in the range 10-5-10-2 M H2SO4 for plutonium. Taking into account the results obtained, a good separation of uranium, thorium, and plutonium cannot occur under these conditions. Nevertheless, it will be possible to use ALPHAEXR in 3 M H2SO4 as a washing step to recover quantitatively thorium and uranium in aqueous solution after only two extraction steps (Figure 6, step 1).
In return, the behaviors of URAEXR and THOREXR scintillation cocktails appear better in H2SO4 media. As shown in Figure 4, uranium is well extracted by URAEXR in the range 10-5-0.5 M H2SO4, while the recovery yields of plutonium and thorium are nearly 70% and 5%, respectively. It is the reverse when extractions are performed in THOREXR: thorium and plutonium are well recovered in the range 0.5-5 M H2SO4, while the recovery yield decreases for acidic media more concentrated than 0.1 M. As noted in the following section, the use of THOREXR and URAEXR is preferred to perform the separation process in sulfuric media. Actinides Separation. (a) Separation of Thorium and Uranium from Plutonium, Americium, and Curium. As shown in Figures 1 and 2, the extraction of thorium, uranium, and plutonium with ALPHAEXR is almost quantitative for nitric solutions below 0.1 M while trivalent actinides like americium and curium are not extracted. As noted previously, the separation of plutonium in the tetravalent state from uranium and thorium appears very difficult in sulfuric media. Indeed, the extraction of plutonium is seldom quantitative (even after two steps of extraction). For this reason, the extraction of plutonium in the trivalent state has been chosen. Thus, the separatation of americium, curium, and plutonium(III) from thorium and uranium has been preferred during the first step of extraction. Several methods of reduction of plutonium(IV) to the trivalent state have been reported in the literature. Many of these techniques involve using the divalent ion Fe2+ as a reducing reactant.27 Unfortunately, iron ions in the form Fe3+ are often coextracted into the organic phase, leading to yellow organic solutions which absorb energy in the light-transfer process in the (27) Yang, D. Z.; Zhu, Y.; Jiao, R. J. Radioanal. Nucl. Chem., Art. 1994, 183-2, 245-260.
Analytical Chemistry, Vol. 69, No. 13, July 1, 1997
2279
Table 3. Recovery Yields of Nuclides for Each Step of the Extraction Process recovery yield of actinide in organic phase (%) nuclide 232Th 234U/238U 236Pu/239Pu 241Am, 244Cm a
step 1 1st ext 2nd ext 100 88 -
0 9 -
step 2 1st ext 2nd ext 0
