Advanced Process Monitoring Techniques for Safeguarding Reprocessing Facilities

Transcription

1 1 IAEA-CN-184/119 Advanced Process Monitoring Techniques for Safeguarding Reprocessing Facilities C.R. Orton a*, S.A. Bryan a, J.M. Schwantes a, T.G. Levitskaia a, C.G. Fraga a, S.M. Peper a a Pacific Northwest National Laboratory Richland, WA 9935 U.S.A. * corresponding author christopher.orton@pnl.gov Abstract The International Atomic Energy Agency (IAEA) has established international safeguards standards for fissionable material at spent fuel reprocessing plants to ensure that significant quantities of weapons-useable nuclear material are not diverted from these facilities. For large throughput nuclear facilities, it is difficult to satisfy the IAEA safeguards accountancy goal for detection of abrupt diversion. Currently, methods to verify material control and accountancy (MC&A) at these facilities require time-consuming and resourceintensive destructive assay (DA). Leveraging new on-line non destructive assay (NDA) process monitoring techniques in conjunction with the traditional and highly precise DA methods may provide an additional measure to nuclear material accountancy which would potentially result in a more timely, cost-effective and resource efficient means for safeguards verification at such facilities. By monitoring process control measurements (e.g. flowrates, temperatures, or concentrations of reagents, products or wastes), abnormal plant operations can be detected. Pacific Northwest National Laboratory (PNNL) is developing on-line NDA process monitoring technologies, including both the Multi-Isotope Process (MIP) Monitor and a spectroscopy-based monitoring system, to potentially reduce the time and resource burden associated with current techniques. The MIP Monitor uses gamma spectroscopy and multivariate analysis to identify offnormal conditions in process streams. The spectroscopic monitor continuously measures chemical compositions of the process streams including actinide metal ions (U, Pu, Np), selected fission products, and major stable flowsheet reagents using UV-Vis, Near IR and Raman spectroscopy. This paper will provide an overview of our methods and report our on-going efforts to develop and demonstrate the technologies. 1. Introduction Conventional nuclear safeguards rely, in large part, on destructive analyses to quantify nuclear material within bulk handling facilities and verify the location of all material within a significant quantity. Though their accuracy is superb, destructive analyses are extremely resource intensive, have limited sampling rates, and are associated with a significant time lag from sampling to final reporting. In addition, the error associated with these analyses scales with the size of the facility [1]. As such, highly precise destructive measurements of special nuclear material alone at large facilities may not be adequate to ensure diversions have not occurred. While it is not likely more precise measurement techniques will be developed in the near future, intelligent integration of a variety of on-line process monitoring tools capable of near-real-time, nondestructive measurements may be successful in adequately safeguarding even the largest envisioned facility. A combination of these techniques could provide a more robust framework that would utilize both material accountancy and augmented material control via continuous process flow sheet verification.

2 Pacific Northwest National Laboratory (PNNL) is currently developing and demonstrating two technologies capable of monitoring conditions at a nuclear reprocessing plant online, non-destructively and in near-realtime They include the Multi-Isotope Process (MIP) Monitor [-7] and the spectroscopy-based monitor [8] (UV-vis-NIR and Raman spectrometers). The MIP monitor is designed to track numerous radioactive constituents within process streams to indirectly monitor conditions in that stream. The monitor would employ small, relatively high resolution gamma detectors that do not require cryogenic cooling and that can be easily placed throughout an existing facility. Multivariate analysis would be used to obtain near-realtime, automated processing of the spectra to continuously monitor the gamma-ray response for changes in stream conditions (e.g., acid concentration, TBP concentration, temperature, etc.). Initial simulations and experiments have illustrated the MIP Monitor s ability to monitor for changes in process streams through gamma spectra. An overview of the MIP Monitor technique will be discussed here. Spectroscopic monitoring using UV-vis-NIR and Raman spectroscopy allows for continuous monitoring of chemical compositions within the process streams including actinide metal ions (U, Pu, Np), selected fission products, and major reagent chemicals. Availability of on-line, real-time techniques that directly measure process concentrations of nuclear materials will enhance performance and proliferation resistance of the solvent extraction processes. Further, on-line monitoring of radiochemical streams will also improve reprocessing plant operation and safety. This manuscript describes initial results and the current state of development of both the MIP Monitor and spectroscopic on-line monitoring techniques for safeguards applications.. MIP Monitor Modern industrial reprocessing techniques, including the PUREX and UREX+ family of separations technologies, employ liquid-liquid extraction techniques to recycle spent nuclear fuel. The fuel is dissolved into an aqueous solution and the products (e.g. actinides) are extracted by contacting the fuel solution with an organic solution that preferentially (thermodynamically) removes them from contaminants (e.g. fission and activation products) [9]. In these bi-phase systems small amounts of both product and contaminants remain in the other phase. The distribution of each element between the organic and aqueous phases is determined by major process variables such as acid concentration, organic ligand concentration, reduction potential, and temperature. Hence for a consistent industrial process (i.e., steady-state and reproducible conditions), the distribution of each element between the organic and aqueous phases should be relatively constant. Consequently, the pattern of elements distributing into product and waste streams at each stage in the facility should be reproducible, within normal industrial variations, resulting in a signature indicative of normal process conditions. Under abnormal conditions (such as those expected under some protracted diversion scenarios), patterns of elements within the various streams would be expected to change measurably. The MIP monitor is designed to track changes in the distribution of gamma-emitting elements as evidence that process conditions are changing. In-process surveillance by the MIP monitor is accomplished by coupling the gamma spectra recorded from constituent streams with multivariate analysis, such as Principal Component Analysis (PCA). PCA is a chemometrics tool that finds combinations of variables (principal components or PCs) that best describe the variance between differing datasets [1]. The MIP technique evaluates patterns of the gamma-emitting nuclides in near-real-time for statistically relevant signs of significant changes to the process. By assuring the process is operating as declared, the MIP monitor warns of possible process migration to an undeclared operation. Because the pattern comparison is automatic and autonomous, proprietary operational or fuel information can be protected while assuring process integrity. However, if desired, process conditions can also be quantified using a multivariate calibration technique such as Partial Least Squares [11]. Another advantage of the MIP monitor is the ability to place gamma detectors throughout a reprocessing facility. Gamma detectors can be small and easily collimated, facilitating deployment to locations in the process line where other monitoring systems might be restricted do to their size, maintenance, or technical limitations. While High Purity Germanium (HPGe) gamma detectors give superior resolution, their need for cryogenic cooling and relative fragility make them less desirable for deployment behind a biological shield where maintenance would be difficult. In addition, their resolution may not be necessary for the MIP

3 Scores on PC (7.8%) Monitor analysis techniques. Multivariate analysis can be tailored to consider overall shifts in the spectra, as opposed to focusing solely on isotopic analysis, to monitor changes even when peaks are not completely resolved. Medium resolution gamma detectors require less maintenance and no cooling appendages, maximizing the ease of their deployment and maintenance. Initial experimental and modeling efforts have suggested the MIP Monitor is capable of the described operations. A few selected results are discussed below. A more detailed overview of the MIP Monitor can be found in the following references [-7]..1. Experimental Experiments were performed to test the MIP Monitor concept and to develop its effectiveness. A segment of commercial spent nuclear fuel was dissolved and a liquid extraction performed at the hot cell facilities at PNNL. The fuel, known as ATM-19, was from a boiling water reactor (BWR), had an initial enrichment of 3% 35 U, a total irradiation of approximately 7 MWd/kgU, and a cooling time of approximately 16 years. The characterization of the segment can be found in the following references [1,13]. The fuel was dissolved in nitric acid and samples were prepared at acid concentrations from.3 to 5.1 M. The prepared concentrations represented low, normal (.5 M) and high acid concentrations for the feed entering a PUREX extraction. Liquid extractions were performed on each of the feed samples by contacting a portion of them with an equal volume of 3 v% TBP in dodecane. A.1 ml sample of the organic extract from the separation performed at the lowest, highest and normal acid concentration was taken and diluted to 1 ml and counted on a 74.3% relative efficiency HPGe detector for two hours. The counts were repeated ten times to provide a sample set with some stochastic variability. While replicate extraction samples would have provided a better representation of the normal variability expected in a continuous process, this was not possible due to resource constraints. PCA was then applied to 9 of the 3 spectra after the spectra were normalized by area and mean-centered. One of the high acid concentration spectra was excluded as an outlier because it deviated substantially and inexplicably from the other spectra. Eight of the spectra from the normal organic acid set were used as a training set to build a PCA model. The model condensed the variation within the training set to principal components while capturing 96.83% of the variance. The remaining 1 spectra were then projected onto the model to determine if the model could separate the normal from the off normal conditions. Figure 1 illustrates the result of the analysis. x Normal Training Set Low Acid Normal High Acid 95% Conf. Level Scores on PC 1 (89.75%) Figure 1. Principal Component Analysis of experimental HPGe spectra from the organic phase of a PUREX liquid extraction performed on spent fuel (BWR, ~7 MWd/kgU) at different acid concentrations. The model was built around samples of extractions performed at.5 M nitric acid and samples of normal,

4 Predicted MWd/kgU high and low acid concentrations were projected onto the model. As can be observed in Figure 1, the PCA model correctly characterized the projected spectra using principal components. The two normal spectra fall within the confidence interval established by the training set and the high and low acid concentration samples fall outside the confidence limit. Though the experiment is a simplification of the challenge expected at a functioning reprocessing facility, it proves that the principle of the MIP monitor is sound, which is that gamma spectral patterns emitted from reprocessing streams can be used to monitor process conditions... Simulations Recent modeling efforts have focused on whether the MIP Monitor can make a valid assessment of spent fuel burnup. Spent BWR fuel was modeled using ORIGEN-ARP [14]. The base case was a General Electric 8x8 assembly with 3% initial 35 U enrichment placed in a reactor at a power density of MW/MTU for three consecutive cycles and then cooled for three years. The total irradiation was varied and included 14, 1, 8, 35, 4, 49, 56, 63, and 7 MWd/kgU. The output from ORIGEN-ARP consisted of the fuel composition, including each nuclide s activity, and it was used as input to PNNL s SYNTH computer code. SYNTH is a 1-D radiation transport code designed to mimic the response of a selected detector type among several choices, including sodium iodide, germanium and cadmium zinc telluride [15]. A cadmium zinc telluride (CZT) detector was modeled with a.1 cm 3 crystal, a collection time of one minute, and a point source located 15 cm from the detector. Each irradiation and decay case from ORIGEN-ARP was used a source list in SYNTH. The output consisted of a CZT gamma spectrum of each case. Partial Least Squares multivariate analysis was applied to the spectra after they were normalized by area and meancentered. The 14, 8, 35, 49, 56, and 7 MWd/kgU cases were used as a calibration set. The irradiation levels of the 1, 4, and 63 MWd/kgU cases were then predicted using their simulated gamma spectra. The result can be seen in Figure. 7 6 Predictions Model (6 Spectra Calibration) MWd/kgU Figure. Partial Least Squares (PLS) predictions (using 3 Latent Variables) of the burnup of simulated spent nuclear fuel using its simulated.1 cc CZT gamma spectrum. The model was calibrated with 6 spectra, and 3 predictions were made. The model was able to predict the burnup of the test samples within 1% of the actual value (< ±3 MWd/kgU). The normalization before analysis of the spectra removes the intensity component, which removes artificial variation due to shifts in the geometry of the sample and detector or changes in the count

5 Absorbance Pu(IV), mm predicted [Pu(IV)], mm, model time. However, in the case of burnup, it removes an important indicator. If the geometry and count time can be held constant (which is easily done in simulations), the burnup analysis can be much improved by keeping the intensity data. When the model was built without normalizing the spectra by area before analysis, the predictions fall within.1% of the actual value, proving to be a much more accurate approach. Though it requires more research, the MIP Monitor s ability to quantitatively measure the burnup of spent fuels and other process conditions is promising. 3. Spectroscopic Methods Raman [16] and UV-Vis [17-] spectroscopy are analytical techniques that have been used extensively for measuring the concentrations of various organic and inorganic compounds, including actinides. Additionally, measurement of dielectric properties has also been proposed for on-line monitoring of fuel reprocessing systems [3]. The feasibility of on-line control of nuclear fuel reprocessing streams by using analytical techniques has been investigated as early as the 197 s [4]. Proof-of-principal experiments for design of on-line, real-time process spectroscopic instrumentation for use in monitoring, controlling, and safeguarding fuel reprocessing flowsheets was conducted at PNNL. The preliminary investigations included evaluation of the PUREX flowsheet for component concentrations and process parameters; preparation of aqueous PUREX matrix simulants for feed, raffinate, and organic solvent; measuring Raman and UV-vis-NIR spectroscopic responses of U, Pu, and Np in matrix simulants; and evaluation of sensitivity and detection limits of U, Pu, Np, and NO 3 - in each simulant for available inhouse spectroscopy instrumentation. The following scheme (Figure 3) depicts the flow and use of information building from initial spectra acquired for model building (top-left) to initial chemometric model development (top-center) to application of models to real-time data collected (top-right) to final deployment as an integrated real-time data processing/analysis/storage/archive system for real-time display and interpretation of spectroscopic data used by PNNL for spectroscopic process monitoring Static measurements: Model training database Chemometric model development 1 1 Pu chemometric model PLS model for Pu(IV) using UV-vis data y =.9997x +.6 R =.9999 On-line model verification and translation wavelength, nm Pu(IV), mm Real-time on-line concentration data display Integrated software for data collection, processing, storage and archiving Figure 3. Schematic of how information was used toward building an integrated on-line, real-time process monitoring system Demonstration of spectroscopic methods The spectroscopic methods being developed are amenable to use for measuring components in commercial fuel and fuel simulant feeds. The section below discusses work performed using actual spent commercial fuel, including the preparation and dissolution of the fuel and the spectroscopic measurements of these fuel

6 Raman response Raman response samples. The results were compared to ICP-MS analysis and ORIGEN-ARP calculations. Section 3.1. contains a description of the demonstration of real-time spectroscopic monitoring using a centrifugal contactor system deployed at PNNL which includes measurements of fuel feed simulant solutions containing relevant concentrations of UO (NO 3 ), Np(V), and Np(VI) in nitric acid Spectroscopic Measurements of Commercial Fuel To further examine applicability of optical spectroscopy for monitoring reprocessing solutions, two segments of spent nuclear fuel were dissolved in concentrated nitric acid and prepared as described in the MIP Monitor section (.1). The same ATM-19 fuel solutions were measured using Raman and vis-nir spectroscopic techniques. An additional fuel, ATM 15, was also prepared in the same manner as the ATM-19 fuel and considered. ATM-15 consists of spent BWR fuel with a lower burnup ( ~16 MWd/kgU) and longer cooling time (6 years) [5]. The goal of the initial measurements was to confirm the utility of optical process monitoring methods employing Raman and vis-nir for the direct measurement of dissolved fuel feed, TBP-dodecane extraction, and raffinate phases Direct Raman measurements of the aqueous/nitric acid feed and raffinate solutions performed on ATM-19 samples are shown in Figure 4A. This figure also contains a measurement of the simple feed fuel simulant (containing 1.3 M UO (NO 3 ) /.8M HNO 3 ) for comparison of the Raman response between simulants and actual fuel samples. The spectral features responsible for the UO + and NO - 3 bands (87 and 147 cm -1 respectively) in the fuel feed and raffinate samples are in agreement with those contained within the simple feed simulant. The TBP-dodecane extract phase of ATM-19 fuel was also measured by Raman spectroscopy. Figure 4B contains the Raman spectrum of the ATM-19 TBP-dodecane phase and the extract phase of simple seed simulant prepared for comparison. The comparison of the Raman band locations for the UO + and NO - 3 bands (858 and 19 cm -1 respectively) between the actual commercial fuel and simulant are in agreement. Other bands observed in the extractant phase Raman spectra are assigned to the solvent (TBP and dodecane) A UO + Simple Feed fuel simulant ATM-19 Feed, commercial fuel ATM-19 Raffinate, commercial fuel - NO UO + TBP B NO3 - ATM-19, commercial fuel Simple Feed, fuel simulant dodecane 6 4 offset added for clairity wavenumber, cm wavenumber, cm -1 Figure 4. (A) Raman spectra of ATM-19, high burn up BWR commercial fuel feed and raffinate solutions, and simple feed simulant solution. ATM-19 Feed solution contains.3 M HNO3, simple feed contains 1.33 M uranyl nitrate in.8 M HNO3. (B) Raman spectra of TBP-dodecane extraction solutions of ATM-19 commercial fuel and simple feed simulant solution. Spectrophotometric measurements of the aqueous feed solutions of the high (ATM-19) and low (ATM- 15) burn up commercial fuel samples were performed using the vis-nir spectral region. Plutonium in both Pu(IV) and Pu(VI) oxidation states was observed in the dissolved fuel feed solution, with varying concentrations depending on final HNO 3 concentration, as is apparent in Figure 5. Although more prevalent in the high burn up fuel(atm-19, Figure 5A), neptunium as Np(V) was also evident in the low burn up (ATM-15, Figure 5B) fuel.

7 absorbance absorbance absorbance absorbance A A Pu(IV) Pu(IV) Pu(VI) Pu(VI) M HNO3 5.1 M HNO3 3.8 M HNO3.5 M HNO3 1.3 M HNO3.3 B B Pu(IV) Pu(VI) Pu(VI) M HNO3 5.1 M HNO3 3.8 M HNO3.5 M HNO3 1.3 M HNO Np(V).3.6 Nd(III) wavelength, nm wavelength, nm wavelength, nm Figure 5. Vis-NIR spectra of commercial fuel A) high burn up ATM-19 and B) low burn up ATM-15). Aqueous feed acid concentrations range.3m 5.1 M HNO3. As discussed previously, the commercial fuel solutions were contacted with an organic phase extractant containing 3 v% TBP in dodecane mimicking the PUREX/UREX extraction. Figure 6A shows the vis- NIR spectra of the resulting organic product phase from a) a low burnup, 1.3 M HNO 3, b) a low burnup, 5.1 M HNO 3 c) a high burnup,.3 M HNO 3, and d) a high burn up in 5.1 M HNO 3 fuel solution. The spectral bands observed in Figure 6 are those diagnostic for Pu(VI). For comparison, a fuel simulant containing a feed composition of 1.33 M UO (NO 3 ) in.8 M HNO 3 with a variable Pu(IV) concentration ranging from.1 to mm was contacted with the 3% TBP/dodecane PUREX solvent followed by spectroscopic measurement by vis-nir spectroscopy, with the resulting spectra shown in Figure 6B. This figure illustrates the vis-nir spectral region for the variable Pu(VI) concentrations in the organic solvent. There is excellent agreement in comparing the Pu(VI) bands between the actual commercial fuel extract (Figure 6A) and fuel simulant extract (Figure 6B). Absorbance Units A d4 c3 b 1 a B Wavelength, nm wavelength, nm Figure 6. (A) Vis-NIR spectra of TBP/dodecane extraction phase of low and high burn up commercial fuel (+.5,.1, and.15od offset added to b, c, and d for clarity). (B) vis-nir Spectra of TBP/dodecane extraction phase of fuel simulant. The Raman and vis-nir spectra of the ATM-19 feeds were subjected to chemometric and standard Beers Law spectral analysis to determine the concentrations of U, Pu, Np, and Nd present in solution. The resulting concentrations are contained in Table 1. For comparison, the ICP-MS results are also displayed as well as the ORIGEN code calculations for these fuel samples. From this table, it is evident that the spectroscopic method is in relatively good agreement with the standard ICP-MS analysis. Table 1. Analytical results for ATM-19 commercial fuel samples (Mole/L).

8 ATM-19 Nd Np U Pu ORIGEN 1.1E- 4.6E-4 7.E-1 7.5E-3 ICP-MS 8.4E-3 4.7E-4 7.E-1 9.E-3 Spectroscopic a 5.4E-3 3.E-4 7.3E-1 9.6E-3 ORIGEN / ICP ratio ORIGEN / Spectroscopic ratio a) Spectroscopic values are preliminary estimate based on combination of chemometric analysis and traditional Beers Law analysis. Concentration values in Molar (M) units Process Monitoring Demonstration using Centrifugal Contactors: Hot Testing A series of feed solutions containing UO (NO 3 ) in nitric acid were introduced into the hot centrifugal contactor system to test the functionality of the contactor system with the instrumented spectroscopic equipment. A series of solutions, sequentially increasing in nitric acid and then increasing in UO (NO 3 ), were introduced into the centrifugal contactor flow loop system. Raman spectra were collected concurrent with the additions of HNO 3 and UO (NO 3 ) into the feed. Figure 7 shows the accumulated Raman spectra taken over the time frame of the nitric acid and U additions. Several spectral features are apparent within this figure: the water region at 3-4 cm -1 ; the nitrate band at 15 cm -1 ; and the UO + at 871 cm -1. By using a chemometric model formed from spectra containing known quantities of UO (NO 3 ) /HNO 3, a successful translation of the model based on static measurements to on-line measurements for on-line monitoring was achieved. Figure 8 contains the expected and predicted concentrations of the Raman online measurements and shows excellent agreement between values. It is worth noting that the model is capable of not only predicting the UO + and nitrate concentrations but is also capable of differentiating between total nitrate and nitric acid. The distinction between nitrate and nitric acid is due to the inclusion of all the spectral data within the Raman spectrum, including the water region (3-4 cm -1 ) and multiple nitrate bands (of which 15 cm -1 is the largest), which show subtle but reproducible changes based on acid content and the ionic strength of the solution. Nitrate band at 15 cm -1 O-H region at 3-4 cm -1 HNO 3-1 wash HNO 3-3 HNO 3 - HNO 3-5 HNO 3-4 HNO 3-6 UO -1 UO -6 UO -5 UO -4 UO -3 UO - UO + band at 871 cm -1 Fuel Feed Simulant, M ID HNO 3 UO (NO 3 ) NO 3 total HNO HNO HNO 3-3. HNO HNO HNO UO UO UO UO UO UO Figure 7. Real-time Raman monitoring of the fuel simulant extraction solution.

9 concentration, concentration, M (HNO M 3, nitrate) concentration, M (UO (NO 3 ) ) Predicted Concentrations from Process Raman Spectroscopy total nitrate nitric acid uranyl nitrate Expected values HNO 3-1 HNO 3 - wash HNO 3-3 HNO 3-4 HNO 3-5 HNO 3-6 UO -1 UO - UO -3 UO -4 UO -5 UO time, min Figure 8. Measured and predicted Raman on-line measurements showing agreement between values. The light blue lines are the expected concentration of analyte in solution; the red, green, and dark blue lines are the predicted concentration of HNO 3, total nitrate, and UO (NO 3 ) respectively. One of the many potential benefits of the discussed process monitor is its utility for safeguards purposes. Because the monitor acts on-line and in near-real-time, it provides a unique capability to rapidly identify unwanted/suspect deviations from normal operation conditions. This multi-parametric monitoring includes measurements of chemical compositions and physicochemical parameters of the radiochemical streams and allows identification of several chemical and physicochemical signatures of a separations process, such as concentration and speciation of the key components, stream ph, flow rate, temperature, etc. Elimination of a grab sample technique provides the capability for immediate alarming/conformation of off-normal operations. Coupling the results of complementary analytical techniques with the ability to search chemical libraries of spectroscopic, chemical, and physicochemical properties provides a path for detecting modified or unwanted chemical agents in the various process streams. Our approach is to explore the potential of chemical diagnostics to detect possible facility misuse. Large deviations can readily be detected and measured with standard monitoring equipment. Additional focus will be made on identifying small-scale deviations in large facilities that are currently not well-detected. 4. Conclusion Pacific Northwest National laboratory has developed new technologies with potential to enhance nuclear safeguards at reprocessing facilities. Though in its infancy, the Multi-Isotope Process Monitor has shown great promise for monitoring spent fuel solutions. The MIP monitor has proved to be sensitive to small changes in process conditions and fuel specifications. Its sensitivity to burnup, though not yet well characterized, shows promise as a potential tool for verifying burnup in the initial dissolver or accountancy tank at the front end of a reprocessing facility. The MIP Monitor has also shown promise as an indicator of off normal process conditions. With more robust preprocessing development and experimental efforts to delineate cooling time and burnup effects on spectral patterns, the MIP monitor might be employed throughout a facility to monitor process conditions (like acid concentration). Further development is ongoing and focused on improving data processing methodologies for analyzing spectra in near-real-time. The spectroscopic methods approach provides online, real-time analysis of actinide concentrations in process streams. The potential applicability of optical spectroscopic techniques for on-line monitoring of reprocessing streams has been demonstrated using spent nuclear fuel solutions, comprised of high burn-up (7 MWd/kgM), ceramic commercial fuel, dissolved in.3 to 5.1 M HNO 3, measured spectroscopically without further modification (e.g., dilution, separation of fuel components, etc.). The spent fuel feed

10 solutions, equilibrium aqueous raffinates, and loaded TBP/n-dodecane organic fractions were subjected to Raman and Visible/Near Infrared (vis-nir) spectroscopic measurements. Concentration values obtained using these methods for U(VI), Pu(IV), Pu(VI), Np(V), and Nd(III) were compared to ICP-MS analytical results of the same solutions, and to ORIGEN code calculations for fuel of similar burn-up and composition. The spectroscopic measurements were in good agreement with both the ICP-MS and ORIGEN calculations. 5. Acknowledgments This research is sponsored by the U.S. Department of Energy s NA-4 (NNSA) and Fuel Cycle Research and Development (NE). It was conducted by Pacific Northwest National Laboratory under DOE contract number DE-AC6-76RLO-183. REFERENCES [1] U.S. Congress, Office of Technology Assessment. Nuclear Safeguards and the International Atomic Energy Agency, Appendix A OTA-ISS-615. Washington, D.C.: U.S. Government Printing Office, June (1995). [] L.E. Smith, J.M. Schwantes, J.J. Ressler, M. Douglas, K.A. Anderson, C.G. Fraga, P.C. Durst, C.R. Orton, R.N. Christensen. Next Generation On-line MC&A Technologies for Reprocessing Plants Proceedings of Global 7 Conference on Future Nuclear Energy System, Boise, ID (7). [3] J.M. Schwantes, M. Douglas, C.R. Orton, C. Fraga and R.N. Christensen, Multi-Isotope Process (MIP) Monitor: a Near-Real-Time Monitor for Reprocessing Facilities, ANS Transactions from the Annual Meeting, Anaheim, CA (8). [4] C.R. Orton, J.M. Schwantes, S. Bryan, T. Levitskaia, D. Duckworth, M. Douglas, O.T. Farmer, C. Fraga, S. Lehn, M. Liezers, S. Peper, R.N. Christensen. Advanced Safeguards Technology Demonstration at Pacific Northwest National Laboratory Proceedings of the 49th INMM Annual Meeting, Nashville, TN, (8). [5] J.M. Schwantes, C.R. Orton, C.G. Fraga, M. Douglas and R.N. Christensen, The Multi-Isotope Process (MIP) Monitor: A Near-Real-Time, Nondestructive, Indicator of Spent Nuclear Fuel Reprocessing Conditions, Proceedings of the 5 th INMM Annual Meeting, Tucson, AZ, (9). [6] C.R. Orton, J.M. Schwantes, C.G. Fraga, M. Douglas and R.N. Christensen, Experimental Validation of the Multi-Isotope Process Monitor Concept, Proceedings of Global 9, Paris, France, (9). [7] C.R. Orton, C.G. Fraga, M. Douglas, R.N. Christensen, and J.M. Schwantes, Monitoring Spent Nuclear Fuel Reprocessing Conditions Non-Destructively and in Near-Real-Time Using The Multi- Isotope Process (Mip) Monitor, Proceedings of the nd JAPAN-IAEA Workshop on Advanced Safeguards Technology for the Future Nuclear Fuel Cycle, Tokai-mura, Japan, (9). [8] S.A Bryan, and T. G Levitskaia. Monitoring and Control of UREX Radiochemical Processes Proceedings of Global 7 Conference on Future Nuclear Energy Systems (7). [9] M. Benedict, T.M. Pigford, and H.W. Levi, Nuclear Chemical Engineering, nd Ed. McGraw-Hill, New York, NY (1981). [1] E. Malinowski, Factor analysis in chemistry, John Wiley & Sons, New York, NY, 45p (8). [11] K.R. Beebe, R.J. Pell, M.B. Seasholtz, Chemometrics: A Practical Guide, John Wiley & Sons, Inc., New York (1998). [1] S.F. Wolf, D.L. Bowers, J.C. Cunnane. Analysis of high burnup spent nuclear fuel by ICP-MS. Journal of Radioanalytical and Nuclear Chemistry, 63, 3, (5). [13] S. Vaidyanathan, R. D. Reager, R. W. Warner, et al., High Burnup BWR Fuel Pellet Performance. Proceedings of the International Topical Meeting on Light Water Reactor Fuel Performance, American Nuclear Society, p. 471, Portland, OR, March -6 (1997). [14] A.G. Croff. ORIGEN: A versatile computer code for calculating the nuclide compositions and characteristics of nuclear materials, Nucl. Technol., 6, 3 (1983). [15] W.K. Hensley, A.D. McKinnon, H.S. Miley, M.E. Panisko, and R.M. Savard, SYNTH: a spectrum synthesizer, J. Radioanal. Nucl. Chem., 193, 9, (5). [16] C. Madic, G. M. Begun, D. E. Hobart, and R. L. Hahn Raman Spectroscopy of Neptunyl and Plutonyl Ions in Aqueous Solution: Hydrolysis of Np(VI) and Pu(VI) and Disproportionation of Pu(V). Inorg.

11 Chem., 3, (1984). [17] H. Schmieder and E. Kuhn Automatic Measurement and Control of Nuclear Fuel Reprocessing by Spectrophotometry and Conductivity Measurements. Chemie Ingenieur Technik, 44(3): 14-& (197). [18] D. Ertel Analytical Methods in The Purex Process Atomkernenergie-kerntechnik, 46(), (1985). [19] D. Ertel, P. Groll, G. Knittel, and W. Thessis Process Analysis in Purex Method J. Radioanal. Chem., 3(), (1976). [] F. Baumgartner and D. Ertel The Modern Purex Process and Its Analytical Requirements. J. Radioanal. Chem., 58(1-), 11-8 (198). [1] J. Burck Spectrophotometric Determination of Uranium and Nitric Acid by Applying Partial Least Squares Regression to Uranium(VI) Absorption Spectra. Anal. Chim. Acta, 54(1-), (1991). [] B. J. Colston and G. R. Choppin Evaluating The Performance of a Stopped-Flow Near-Infrared Spectrophotometer For Studying Fast Kinetics Of Actinide Reactions. J. Radioanal. Nuc. Chem., 51(1), 1-6 (1). [3] M. Yamamoto Determination of Nitric Acid Concentration in 3 Vol Percent TBP-n-dodecane by Measuring Dielectric Properties. J. Nucl. Sci. Tech., 5(6), (1988). [4] I. Parus, J. Kierzek, and T. Zoltowski On-line Control of Nuclear Fuel Reprocessing Nukleonika, (9), (1977). [5] R.J. Guenther, D.E. Blahnik, T.K.Campbell, U.P. Jenquin, J.E. Mendel, L.E. Thomas, C.K. Thornhill, Characterization of Spent Fuel Approved Testing Material ATM-15, Pacific Northwest Laboratory, PNL , June (1989), accessed at (October 8).