Journal of Analytical Atomic Spectrometry Characterization of Spent Nuclear Fuels by Ion Chromatog raphy-Inductively Coupled Plasma Mass Spectrometry* JOSEFA M. BARRERO MORENO J. IGNACIO GARCIA ALONSOt PHILIPPE ARBORE GEORGOS NICOLAOU AND LOTAR KOCH European Commission JRC Institute for Transuranium Elements Postfach 2340 761 25 Karlsruhe Germany Ion chromatography (IC) coupled to ICP-MS was applied to the characterization of two different spent fuel samples [uranium oxide ( U02) and mixed uranium-plutonium oxide (MOX)] and the results were compared with those obtained by other techniques. Isotope dilution analysis with ion chromatographic separation was applied to the determination of the fission products (Rb Sr Cs Ce Nd Sm Eu and Gd). The standard additions method was employed for the determination of monoisotopic fission products and actinides (Y La Pr 147Pm 237Np "'Am 243Am and 244Cm).Total U and Pu were determined only by ID-TIMS. Nd Am and Cm isotope concentrations were determined also by ID-TIMS. gamma spectrometry. Based on the ID-TIMS results for U 9 134Cs and I3'Cs were determined in parallel by 144Ce 154EU Pu and 148Nd fuel burn-up was calculated and the value introduced into the computer code KORIGEN in order to calculate the complete fuel inventory based on the known irradiation parameters for the two fuels. The agreement between the experimental IC-ICP-MS results and the theoretical calculations was within 15% for most isotopes. Keywords Inductively coupled plasma mass spectrometry; chromatography; isotope dilution; spent nuclear fuel ion The inventory of fission products and actinides in spent nuclear fuel is often required when new types of fuels and/or reactor operating conditions are being investigated.Fuels containing minor actinides (Np Am Cm) for transmutation purposes and fuels subjected to very high burn-ups are examples where post-irradiation examination of the spent fuel has to be performed. Current techniques used in the Institute for Transuranium Elements for the determination of fuel param- eters include mainly ID-TIMS (for U Pu Am Cm and Nd) and gamma spectrometry (for 134Cs 137Cs '44Ce ls4Eu and lo6Ru). Based on the heavy metal and Nd isotope concen- trations determined the fuel burn-up can be calculated; this is then applied in the computer code KORIGEN,' together with the known irradiation history of the fuel and the reactor operating conditions. The complete inventory of the fuel is then calculated and the results are compared for the other measured isotopes.However the determination of Am. Cm and Nd by TIMS is very time-consuming because of the difficult chemical separations that have to be performed. The determination of gamma-emitting nuclides is simpler but suffers from poor accuracy and precision in spent fuel samples for some nuclides. In order to obtain a larger amount of data and simplify the handling of the samples we have applied ICP-MS to the characterization of spent fuel ICP-MS is a sensitive multi-elemental technique that offers very low detec- * Presented at the 1995 European Winter Conference on Plasma Spectrochemistry Cambridge UK January 1995.7 To whom correspondence should be addressed. Present address Department of Physical and Analytical Chemistry Faculty of Chemistry Julian Claveria 8 33006 Oviedo Spain. tion limits and broad elemental coverage with the additional capabilities of providing isotopic abundance information and absolute concentrations by applying the technique of isotope dilution. However the complete determination of fission prod- ucts and actinide isotopes in spent nuclear fuels by ICP-MS is hindered by the presence of some isobaric interferences which cannot be corrected f ~ r . ~ ~ In order to overcome these interferences chemical separation is needed. The coupling of liquid chromatography to ICP-MS is well documented in the literature for trace metal speciation and environmental analy- S ~ S .~ Also the separation of the lanthanide elements in order to eliminate isobaric interferences has been described." We have used also ion chromatography (IC) coupled to ICP-MS for the elimination of isobaric interferences in the analysis of spent fuel samples." Fission products that can be determined in spent fuel solutions by IC-ICP-MS include Rb Sr Y Cs Ba La Ce Pr Nd Pm Sm Eu and Gd. Other fission products including Zr Mo Tc Ru Rh Pd and Te are not completely dissolved in 7 mol 1-1 nitric acid and remain partially as insoluble residues.' For this reason we have selected only the first series of elements together with the actinides Np Am and Cm for a comparison with the standard techniques and with the theoretical calculations. EXPERIMENTAL Instrumentation An ELAN 5000 ICP-MS instrument (Perkin-Elmer SCIEX Thornhill Ontario Canada) was modified to handle radioac- tive samples in a glove box and coupled with a 45001 high- pressure chromatographic pump (Dionex Sunnyvale CA USA).Details of the instrumental set-up have been published previously." The Dionex high-pressure pump was located outside the glove-box while the injection valve (Dionex Part No. 42766 200 pl loop) guard column and analytical column were placed inside the glove-box. The effluent from the chroma- tographic column was directed to the cross-flow nebulizer. A peristaltic pump was used to carry the waste to a vessel inside the glove-box. Standard plasma and ion lens operating con- ditions are summarized in Table 1.They were optimized as described previously." Reagents and Materials Lanthanide separations were performed using Dionex CG5 and CS5 mixed bed columns using a 0.1 mol 1-' oxalic acid (Suprapur Merck Darmstadt Germany)-0.19 moll - ' lithium hydroxide (analytical-reagent grade Merck) eluent. For the separation of the actinides and Rb Sr Y Cs and Ba Dionex CGlO and CSlO cation-xchange columns were used with a 1 mol 1-' nitric acid (Suprapur Merck) eluent. Natural element standards (1000 mg ml-') of Rb Sr Y Cs Ba La Ce Pr Nd Sm Eu and Gd were obtained from SPEX (Grasbrunn Germany). The 237Np standard was obtained from Los Alamos National Laboratory (Los Alamos NM USA) Journal of Analytical Atomic Spectrometry October 1996 Vol. 11 (929-935) 929Table 1 Operating conditions Rf power/W Argon outer gas flow rate/ 1 min-' Argon intermediate gas flow rate/l min-' Argon nebulizer gas flow rate/ 1 min-') Nebulizer type Spray chamber Load coil-sampler distancelmm Quadrupole working pressure/Pa Sampler and skimmer cones Ion lens 1050 15 0.8 1 .o Cross-flow Scott-type double-pass Ryton 15 (fixed) 2.13 10-3 Platinum Setting (% of range) Ions lens Fission products Actinides Bessel box B 45 55 Bessel box P 50 50 Einzel lens El 30 30 Photon stop S2 40 35 as the metal and dissolved in 7 moll-' nitric acid.Its concen- tration was standardized by titration.12 All standard solutions spikes and samples were prepared by dilution by mass in polyethylene bottles. Nitric acid (Merck Suprapur) and Milli-Q water (Millipore Eschborn Germany) were used for all dilutions.Experimental Procedures Sample preparation One pellet (about 30 g) of irradiated U 0 2 fuel and another of mixed uranium-plutonium oxide (MOX) fuel were dissolved (in duplicate) in 7 mol I-' nitric acid in a hot cell facility by reflux and diluted with 4 moll-' nitric acid. A second dilution by mass was performed using 1 moll-' nitric acid and 5 ml of these solutions containing about 100 pg g-' of fuel solution these conditions and Rb Sr and Y could also be eluted with 1 moll-' nitric acid from the CSlO column. Additionally Rb eluted separated from Sr (possible interference between "Rb and 87Sr in the natural elements) and "Sr was separated from 90Zr which was also present in the samples. Quantification procedures The sample spiked sample and spike solution were injected into the chromatograph consecutively.Mass ranges used were 85-90 133-138 234-237 and 239-244 for the separations performed with the CSlO column and 139-160 for the separa- tions using the CS5 column. Detection was carried out by scanning at 1 point per u with a 0.1 s integration time per point with approximately 2.5 s per scan. The data were trans- ferred to a personal computer for further treatment. Integration of the chromatographic peaks was performed using the software GRAMS/386 (Galactic Industries Salem NH USA). The element concentrations in the measured solutions were adjusted so that no detector dead-time correction needed to be applied. The mass discrimination factor was determined using the natural elements spike solution.' The computer code KORIGEN KORIGEN is a dimensionless simulation program i.e.it does not take into account the position of the fuel element in the reactor which calculates the fuel inventory after irradiation based on different data sets which include fuel burn-up (U Pu I4'Nd data) original fuel inventory (235U enrichment initial Pu concentration and isotopic composition) reactor type (neutron flux and energy) and irradiation history (power generated versus time). Also databases on decay characteristics of different nuclides and fission and neutron capture cross- sections are used in the calculations. The accuracy of the predictions is considered to be between 1 and 3% for major actinide isotopes (U Pu) and up to 10% for fission products and minor actinides.' RESULTS Semi-quantitativeData were transferred into a glove-box for further dilution and spiking.A solution containing natural Rb Sr Y Cs Ba La Ce Pr Nd Sm Eu and Gd was prepared by taking aliquots by mass in the glove-box in order t o spike the sample for analysis. For the determination of Np Am and Cm only a Np standard was used. The mass spectrum obtained during the semi-quantitative evaluation of the MOX sample for the analysis of fission products is shown in Fig. l(a) and that for the actinides in Optimization of the isotope dilution procedure In order to optimize the spike addition the concentrations of the isotopes present in the fuels were determined semi- quantitatively based on the response curve of the in~trument.~ For this purpose solutions containing In Tb and Th were added at concentration levels of about 50 ng ml- '.Mass spectra were measured for all samples in the range 80-160 u for fission products and 230-25Ou for actinides using the standard operating conditions. Separation ofjission products and actinides by ion chromatography The separation of the lanthanides was performed using the experimental conditions previously described'' with 0.1 mol 1-1 oxalic acid-0.19 moll-' lithium hydroxide eluent. Under these conditions the lanthanides elute in order of increasing atomic number. The separation of the actinides was carried out using 1 moll-' nitric acid eluent as described previously." The samples were treated with solid silver@) oxide in order to oxidize all plutonium valencies to Puv' and after 5 min the solutions were injected into the chromatograph.It was observed that Cs could be separated from Ba under Fig. l(b). The peaks corresponding to added In and Tb in Fig. l ( a ) and Th in Fig. l(b) were used as internal standards for the evaluation of the concentrations. Indium was used for Rb Sr and Y terbium for Cs Ba La Ce Pr Nd Pm Sm Eu and Gd and thorium for the actinides. The assignment of concentrations to the different fission and actinide isotopes was carried out based on the main fission isotope for the various fission chains.' Similar experiments were performed for the UOz fuel. No matrix interferences were expected at about 100 pg g-' levels." Based on the results obtained the concentration of the multielemental spike solution was adjusted so as to have the same total elemental concentration as the sample for the isotope dilution analysis and standard additions procedures.Determination of the Lanthanides Lanthanides (La to Gd) are eluted in about 25 min from the CS5 column. The separation obtained for the uranium oxide fuel is shown in Fig. 2 for all the isotopes monitored (139- 160u). As can be observed the order of elution was La Ce Pr Nd Pm Sm Eu and Gd as expected. Maximum peak intensities were always lower than 250 000 counts s-'; hence no dead-time correction procedures needed to be applied. All isobaric interferences present in the original sample are elimin- 930 Journal of Analytical Atomic Spectrometry October 1996 Vol. 11v) 80 90 100 110 B 120 130 140 150 160 1 0 230 235 240 245 Mass Fig.1 Semi-quantitative analysis of fission products and actinides in spent nuclear fuel. (a) Mass spectrum of fission products in a MOX fuel with In and Tb (50 ng g-I) added as internal standards. (b! Mass spectrum of the actinides in the same fuel using Th (50 ng g -') as internal standard 0 La I 0 200 400 600 800 lo00 1200 1400 1600 Time/s Fig. 2 IC-ICP-MS trace obtained for the lanthanide elements present in UO fuel using an oxalic acid-LiOH eluent. Mass range monitored was 139-160 ated. The oxide polyatomic ions from '39La 14'Ce 14'Pr and '42Ce which would interfere with most Gd isotopes measured are observed in the chromatogram at the retention time corresponding to the metals. Direct isobaric interferences between Ce and Nd Nd and Sm Pm and Sm and Sm Eu and Gd are also resolved.Chromatograms obtained at selected masses where some isobaric interferences occur are shown in Fig. 3. The separation between 144Ce (half-life 285 d) and 144Nd (decay product) is shown in Fig. 3 ( 4 and similarly Fig. 3(b) shows the separation between 14'Nd and 14%m. This separation is important as it allows the fuel burn-up to be determined by measuring the 148Nd concentration (the fission yield of this isotope is similar for both 235U and 239Pu fission thus reducing uncertainties arising from estimating the relative number of fissions from these two nuclides). Isotope dilution analysis for Ce Nd Sm Eu and Gd was performed as described previously'' using natural elements as spikes. The isotopes 14'Ce 142Nd lS2Sm '"Eu and ls8Gd were used as reference isotopes in the spike while '42Ce 144Nd "'Sm lS3Eu and lS6Gd were the reference isotopes in the sample.The mass discrimination factor',' was determined using the natural elements for the range 140-160 u and it was found to show no long-term variations. Applying linear regression to four independent determinations of the mass 2500 2000 I500 lo00 500 -. I m o IuNd L 0 500 lo00 1500 2 .d . 0 500 lo00 1500 Time/s Fig.3 Single-ion monitoring of the lanthanides in the U02 fuel (cooling time about 3 years) at selected masses where isobaric inter- ferences occur due to neutron capture or parent-daughter pairs. (a) Separation between '44Ce (half-life 285 d) and 144Nd. (b) Separation between 14'Nd (burn-up indicator) and 148Sm Journal of Analytical Atomic Spectrometry October 1996 Vol.11 931discrimination factor on four different days (76 data points) gave a slope the mass discrimination factor per unit mass of -0.0066 (a=0.0015). This value was applied to correct all isotope ratios measured. The method of standard additions was applied to the determination of monoisotopic La and Pr. The concentration of '47Pm was determined assuming an equal response to '47Sm. The results obtained for the uranium oxide fuel are summarized in Table 2 and those for the MOX fuel in Table 3. The results are compared with those obtained by ID-TIMS (some Nd isotopes) gamma spectrometry (154E~ and 144Ce) and with the theoretical calculations using the computer code KORIGEN. Determination of Rb Sr Y Cs Np Am and Cm The separation was performed using the CSlO column and 1 moll-' nitric acid as eluent as described previously.'' Under these conditions the elution order was Rb Cs Sr and Y for the fission products and Np Pu U Am and Cm for the actinides.The resolution between Rb and Sr and Cs and Ba and the differences in isotope abundances from the natural elements allowed the use of natural elements as spikes for the isotope dilution procedure. The intensity at mass 237 during elution for the MOX fuel solutions is shown in Fig.4(a). As can be observed the peak corresponding to 237Np can be resolved from the tailing of the 238U peak although in the mass spectrum it only appears as a shoulder on the 238U peak [see Fig. l(b)]. The separation between Pu and Am for the same fuel is demonstrated in Fig.4(b) showing the response at mass 241. Separation between Np and Pu was not necessary as there is no mass spectral overlap. The separation between U and Pu was observed to degrade with increasing number of injections which was assumed to be due to unspecific adsorp- tion and reduction of Pu on the column as the retention time for Np and U did not change. This could be restored by cleaning the column with 6 moll-' nitric acid for 30 min. Rb Sr and Cs were determined by the method of isotope dilution analysis using the natural elements as spikes as described above for the lanthanides whereas the determination of Y and Np which have only one stable isotope each was performed by the method of standard additions. Am and Cm isotopes were determined assuming the same response as for 237Np.The results obtained for these elements in both samples are summarized in Tables 4 and 5 and compared with those obtained using the standard procedures. DISCUSSION Precision of Duplicate Analyses Based on the data presented in Tables 2-5 the differences between IC-ICP-MS values and those from the other methods observed when analysing the spent fuel samples by ICP-MS can be assessed. For this purpose the mean values were calculated and corrected for the small differences in the concen- tration of the dissolver solutions. The values of the relative deviation for the duplicate analyses showed typical precisions under 10% except for isotopes at very low concentrations and for La and Pr values which were determined by the method of standard additions.The poor results obtained for monoiso- topic lanthanides could not be explained. In general terms the precision seems to be better for the MOX fuel than for the UOz fuel. Typical precision values for the MOX fuel were under 5% which we believe could be a practical value for these types of analyses using chromatographic separation and ICP-MS detection. The precision of ID-TIMS for the determi- nation of Nd isotopes was better than 1% in all instances (duplicate analyses) which is adequate for burn-up measure- ments using 148Nd. Accuracy of the Concentrations Measured It is difficult to check the accuracy of a new procedure when there is no reference material available for the type of analysis Table 2 Inventory of the lanthanide elements in uranium oxide spent fuel dissolver solutions (atoms x 10I6 g-l) and comparison with TIMS gamma spectrometry and KORIGEN Dissolver solution 1 Dissolver solution 2 Element La Ce Pr Nd Pm Sm Eu Gd Mass 139 140 142 144 141 142 143 144 145 146 148 150 147 147 148 149 150 151 152 154 153 154 155 154 155 156 158 160 Gamma ICP-MS TIMS spectrometry KORIGEN 4.82 - - 6.27 - 6.66 6.77 - 5.71 5.92 0.04 - 0.06 0.05 - 5.63 5.81 - 0.26 0.20 - 2.46 2.64 7.54 - - 8.05 2.87 2.92 - 2.89 3.76 3.77 - 4.01 1.81 1.87 - 1.84 0.92 0.95 - 0.93 - - 0.18 0.20 0.61 - - 0.51 0.99 - - 0.87 0.002 - - 0.01 1.44 - - 1.48 - 0.05 0.03 - 0.50 0.41 - 0.21 0.19 - 0.44 0.54 0.09 - 0.16 0.20 - 0.06 0.03 - 0.to 0.05 0.02 - - 0.04 1.03 0.97 - 0.22 0.13 0.01 - - 0.007 - - - - - - - - - - - - - - Gamma ICP-MS TIMS spectrometry KORIGEN 6.24 4.69 6.62 6.77 5.69 6.72 0.04 - 0.05 0.05 - 5.61 4.45 - 0.26 0.24 2.45 3.04 8.01 8.53 3.30 2.88 - 2.87 4.34 3.71 - 3.99 2.08 1.86 - 1.83 1.05 0.96 - 0.93 - 0.18 0.16 - 0.5 1 0.76 - 0.87 1.16 - - 0.01 0.007 - - 1.48 1.64 - - 0.05 0.04 - 0.49 - 0.21 0.48 0.22 0.44 0.52 0.09 - 0.07 0.20 - 0.06 0.02 - 0.10 0.05 0.04 0.02 1.03 0.99 - 0.22 0.14 - - 0.007 0.01 - - - - - - - - - - - - - - - - - - - - - - - - - - 932 Journal of Analytical Atomic Spectrometry October 1996 Vol.11Table 3 Inventory of the lanthanide elements in mixed oxide spent fuel dissolver solutions (atoms x 10l6 g-I) and comparison with TIMS gamma spectrometry and KORIGEN Dissolver solution 1 Dissolver solution 2 Gamma spectrometry Gamma spect rome t ry - - - 0.04 - - Element La Ce Mass 139 140 142 144 141 142 143 144 145 146 148 150 147 147 148 149 150 151 152 154 153 154 155 154 155 156 158 160 ICP-MS 6.79 8.27 7.41 0.02 6.80 0.12 5.62 7.70 4.30 4.58 2.59 1.51 0.28 1.31 1.22 0.02 2.14 0.13 0.81 0.39 1.01 0.19 0.01 0.12 0.03 0.86 0.20 0.006 TIMS KORIGEN 7.98 7.93 6.86 0.04 7.12 0.11 5.15 6.79 3.93 4.16 2.41 1.40 0.33 1.14 0.96 0.03 2.01 0.15 0.88 0.38 0.77 0.33 0.11 0.17 0.09 0.67 0.22 0.02 ICP-MS 8.43 8.24 7.41 0.03 8.54 0.14 5.86 7.85 4.47 4.77 2.74 1.59 0.3 1 1.27 1.17 0.03 2.1 1 0.12 0.78 0.39 1.08 0.22 0.04 0.13 0.04 0.93 0.22 0.006 TIMS KORIGEN 7.96 7.91 6.84 0.04 7.11 0.11 5.14 6.77 3.92 4.14 2.40 1.40 0.33 1.14 0.96 0.03 2.00 0.15 0.88 0.38 0.77 0.33 0.1 1 0.17 0.09 0.66 0.22 0.02 - 0.03 Pr Nd - 3.94 4.23 2.48 1.43 - 3.89 4.18 2.42 0.32 Pm Sm Eu Gd - 0.52 - Table 4 Inventory of Rb Sr Y Cs Np Am and Cm isotopes in uranium oxide spent fuel dissolver solutions (atoms x loi6 g-I) and comparison with TIMS gamma spectrometry and KORIGEN Dissolver solution 1 Dissolver solution 2 Gamma spectrometry Gamma TIMS spectrometry Element Rb Sr Mass 85 87 86 88 90 89 133 134 135 137 237 24 1 243 244 ICP-MS 1.18 2.45 0.03 3.16 4.01 3.84 6.08 0.34 2.17 6.77 1.16 0.37 0.57 0.37 TIMS KORIGEN 0.83 1.87 0.007 2.53 3.50 3.34 5.35 0.32 1.64 6.05 1.47 0.55 0.74 0.40 ICP-MS 0.92 2.05 0.03 3.25 4.15 3.61 5.56 0.36 2.07 6.48 1.43 0.38 0.65 0.41 KORIGEN 0.82 1.86 0.007 2.52 3.48 3.32 5.33 0.32 1.63 6.02 1.46 0.55 0.73 0.40 Y c s - 0.30 6.25 - - NP Am - 0.49 0.71 0.48 Cm performed.However in this instance the results can be com- pared with those obtained by two independent methods for some isotopes and with the computer predictions for 21.11 the isotopes measured which can be considered to be accurate to about 10% for well-characterized reactors.' It must be borne in mind that the computer predictions are based on measure- ments of the total U total Pu total Am and Cm and I4*Nd concentrations which in our case were taken from the ID-TIMS results.Figs. 5 and 6 show the comparison of ICP-MS measurement data and KORIGEN predictions for the UOz and MOX fuels respectively. The data for the duplicate analyses are included in the graphs. As cxi be observed for both fuels there is a systematic difference between the obtained and calculated concentrations of about 10 -15% for virtually all elements. Hence the ICP-MS determined fuel burn-ups are about 10-15% higher than the values introduced into the computer code in order to calculate the fuel inventories (the fission product inventory increases linearly with the fuel burn-up). If we compare the results presented for Nd isotopes in Tables 2 and 3 by ICP-MS and ID-TIMS it can be seen that except for one solution the results are always 5-1570 higher by ICP-MS.Because the 14*Nd concentrations measured by ID-TIMS were used to determine the fuel burn-up the differences between ICP-MS and ID-TIMS for Nd will be reflected throughout the comparison of ICP-MS measurements with KORIGEN values. The results obtained by gamma spectrometry for 144Ce and IS4Eu are typically higher than those obtained by ICP-MS and for 144Ce are in closer agreement with the KORIGEN predictions.However for ls4Eu there is a large difference between ICP-MS and gamma spectrometry with the computer predictions far from both results. The results obtained for 134Cs and '37Cs presented in Journal of Analytical Atomic Spectrometry October 1996 Vo!. 11 933v) 4 20000 B 15000 iij 10000 SO00 - I G o B 0 100 200 300 400 500 600 16000 14000 12000 10000 8000 6000 4000 2000 0 0 200 400 600 800 1000 1200 Time/s Fig.4 Single-ion monitoring of the actinides in the U02 fuel at selected masses where isobaric interferences occur. (a) Separation between 237Np and 238U detected at mass 237. (b) Separation between 2 4 r P ~ and 241Am Tables 4 and 5 show a much better agreement.For the UOz fuel (Table 4) the 13'Cs values obtained by ICP-MS are 4 and 8% higher than by gamma spectrometry. However for the MOX fuel in Table 5 the results are 5% higher in one instance and 11% lower in the other. For both techniques the 13'Cs t a 35 30 25 20 15 10 5 0 0 10 20 30 40 KORIGEN data Fig. 5 Comparison between the computer predictions and the exper- imental values for the U02 fuel. Duplicate analyses are shown with open and filled circles. The line corresponds to a slope of 1 30 25 20 E l5 h 10 5 0 10 I5 20 25 30 0 5 KORIGEN data Fig. 6 Comparison between the computer predictions and the exper- imental values for the MOX fuel. Duplicate analyses are shown with open and filled circles. The line corresponds to a slope of 1 Table 5 Inventory of Rb Sr Y Cs Np Am and Cm isotopes in mixed oxide spent fuel dissolver solutions (atoms x 10l6 g-') and comparison with TIMS gamma spectrometry and KORIGEN Dissolver solution 1 Dissolver solution 2 Element Mass Rb 85 87 Sr 86 88 90 Y 89 c s 133 134 135 137 NP 23 7 Am 24 1 243 Cm 244 ICP-MS 0.84 1 .a0 0.03 2.77 3.35 2.80 8.81 0.28 5.25 9.10 0.87 3.86 3.30 2.00 Gamma TIMS spectrometry KORIGEN 0.72 1.54 0.003 2.03 2.76 2.63 8.20 0.23 4.40 7.94 0.58 4.56 3.88 1.43 ICP-MS 0.79 1.67 n.d.2.78 3.40 2.94 8.90 0.3 1 5.12 8.98 0.85 4.2 1 3.26 2.01 Gamma TIMS spectrometry KORIGEN 0.72 1.53 0.003 2.02 2.75 2.62 8.18 0.23 4.39 7.92 0.57 4.53 3.87 1.42 934 Journal of Analytical Atomic Spectrometry October 1996 Vol. 1 1results are always higher than the computer predictions.On the other hand the data for 134Cs obtained by gamma spec- trometry show better agreement with the computer predictions than those obtained by ICP-MS. From the data presented in Tables 4 and 5 for Np Am and Cm it can be observed that the values obtained by ID-TIMS for Am and Cm are about 10-30% higher than those otbtained by ICP-MS in contrast to Nd which showed the opposite effect for three of the solutions analysed. The agreement with the predicted concentrations for Am isotopes is better for TIMS in the UO fuel but worse for the MOX fuel. For 244Cm ICP-MS values are closer to the predicted concentrations. The Np concentrations found are lower than predicted for the U 0 2 fuel but higher for the MOX fuel. From the data shown in Tables 2-5 it is difficult t o draw clear conclusions about the accuracy of the proposed rnethod for the determination of fission products in spent fuel s:,mples.The data obtained by gamma spectrometry are not conclusive but seem to indicate a higher burn-up than that calculated based on 14*Nd by TIMS. In summary the accuracj of the proposed method seems to be on average better thaln 15% (assuming that the TIMS results for Nd are correct) which for the purposes of high burn-up studies or new fuel dcvelop- ments for transmutation can be considered to be acccptable considering the large amount of data that can be plovided and the reduced radiation hazard for the operator in ccimpari- son with existing techniques. Accuracy of the Isotope Ratios For most elements the isotope ratios measured by I(:P-MS and those predicted by KORIGEN are in close agreement (data can be extracted from Tables 2-5).Minor differences can be explained by small changes in the neutron capture cross-sections of the fission nuclides with neutron entsrgy or by experimental errors. However significant differences were found for Eu and Gd isotopes in all samples which carinot be explained by measurement errors. These differences [can be attributed to errors in the computer-assigned neutron apture cross-section for lS3Eu. If the neutron capture cross-sect ion for lS3Eu is lower than assumed the formation of 154E~. which does not occur by direct fission and lssEu will be rt'aduced which was observed experimentally. Similarly the c oncen- trations of 154Gd and "'Gd which are decay productil of the Eu isotopes will be lower than predicted thus explainjmg the differences observed also for Gd.It is clear that the direct measurement of isotope ratios of fission elements by I(:P-MS can help in the refining of neutron capture cross-sectioiis. CONCLUSIONS ICP-MS has been applied to the characterization 01' spent nuclear fuels with IC separation and the method of isotope dilution analysis o r that of standard additions. The results obtained for duplicate measurements of two different fut:l types showed that the precision attainable was better than I.()% for most nuclides measured. In comparison with other techniques ICP-MS is capable of providing data on a larger number of isotopes both short medium long-lived and stable which will help in the extrapolation of the current prediction techniques based on KORIGEN or similar computer codes to higher burn-ups or to experimental fuels for transmutation purposes.ICP-MS yields more precise measurements than gamma spec- trometry and the ICP-MS data fit better to the general trend observed for other isotopes. However the precision of ICP-MS measurements after IC is worse than that attainable by TIMS. Also discrepancies of up to 15% between ICP-MS and TIMS for Nd isotopes could not be explained. These discrepancies resulted in differences between the observed and computer- predicted fuel inventory of a similar magnitude. Similar discrep- ancies were observed between TCP-MS and TIMS for Am and Cm isotopes but in opposite directions which rules out the existence of dilution errors in the ICP-MS methodology.The isotope ratios determined by ICP-MS showed good agreement with the predicted values. This allows the correction of estimated neutron capture cross-sections for those nuclides where the discrepancies between both techniques were too large to be due only to experimental errors. This could help in establishing better databases for burn-up-dependent neutron capture cross-sections and to extrapolate current data for different reactor operating conditions or for non-commercial nuclear fuels. REFERENCES 1 2 3 4 5 6 7 8 9 10 11 12 Fischer U. and Wiese H. W. Verbesserte konsistente Berechnung des nuklearen Inventars abgehrannter D WR-Brennstofe auf der Basis von Zell-Abbrand- Verfahren mit KORIGEN Kernfors- chungszentrum Karlsruhe Report 3014 1983.Garcia Alonso J. I. Babelot J.-F. Glatz J.-P. Cromboom O. and Koch L. Radiochim. Acta 1993 62 71. Garcia Alonso J. I. Thoby-Schultzendorff D. Giovanonne B. Koch L. and Wiesmann H. J . Anal. At. Spectrom. 1993 8 673. Garcia Alonso J. I. Thoby-Schultzendorff D. and Koch L. Proceedings of the 15th Annual Symposium on Safeguards and Nuclear Material Management European Safeguards Research and Development Association (ESARDA) Report 26 EUR 15214 EN Ispra 1993 pp. 485-489. Garcia Alonso J. I. Thoby-Schultzendorff D. Giovanonne B. Glatz J.-P. Pagliosa G. and Koch L. J. Anal. At. Spectrom. 1994 9 1209. Garcia Alonso J. I. Garcia Serrano J. Babelot J.-F. Closset J.-C. Nicolaou G. and Koch L. in Applications of Plasma Source Mass Spectrometry II eds. Holland G. and Eaton A. N. Royal Society of Chemistry Cambridge 1993 p. 193. Betti M. Garcia Alonso J. I. Arbore Ph. and Koch L. in Applications of Plasma Source Mass Spectrometry I I ed. Holland G. and Eaton A. N. Royal Society of Chemistry Cambridge 1993 p. 205. Garcia Alonso J. I. Anal. Chim. Acta 1995 312 57. Byrdy F. A. and Caruso J. A. Environ. Sci. Technol. 1994 28 528A. Braverman D. S. J. Anal. At. Spectrom. 1992 7 43. Garcia Alonso J. I. Sena F. Arbore Ph. Betti M. and Koch L. J . Anal. At. Spectrom. 1595,10 381. Cromboom O. Garcia Alonso J. I. Koch L. Goerten J. Roesgen E. Wagner H. G. Ottmar H. and Eberle H. Proceedings of the 4th International Conference on Facility Operations-Safeguards Interface American Nuclear Society IL 1992 pp. 431. Paper 6/01 290F Received February 22 1996 Accepted June 13 1996 Journal of Analytical Atomic Spectrometry October 1996 Vol. 11 935