MONTE CARLO FITTING OF THE AMLI NEUTRON SPECTRUM

LA-UR-16-27725 MONTE CARLO FITTING OF THE AMLI NEUTRON SPECTRUM Robert Weinmann-Smith University of Florida, Los Alamos National Laboratory David Beddingfield Los Alamos National Laboratory Andreas Enqvist University of Florida Martyn Swinhoe Los Alamos National Laboratory

NOVEL OUTCOMES Characterization of the differences between AmLi sources AmLi R1 0.505 0.500 0.495 0.490 0.485 0.480 R1 R1 average 0.475 C-470 C-476 C-477 - N-160 N-161 AmLi N-165 N-262 Source N-263 N-432 N-436 N-437 10 Spectra specific to different series C-series N-series of AmLi sources, and to individual sources Probability (MeV -1 ) 1 0.1 0.01 0 1 2 3 4 Energy (MeV) 2

OUTLINE AmLi applications Active neutron interrogation Physics considerations Current AmLi spectra Measurements Source variations Simulations Spectra fitting Limitations 3

APPLICATIONS The IAEA uses an AmLi source in the Uranium Neutron Coincidence Collar (UNCL) to verify compliance with nonproliferation treaties The UNCL requires calibration with known uranium samples. The AmLi spectrum isn t known well enough to allow simulated calibrations AmLi source UNCL UNCL with fuel 4

PHYSICS CONSIDERATIONS Alphas lose energy traveling through AmO 2 particle of unknown size Energy reduction below Li threshold enhances O contribution Isotope Energy (MeV) LiO Am-241 (α) 5.486, 5.443 2 AmO 2 Li-7 (α,n) 4.38 AmO 2 α O-18 (α,n) 0.85 U-238 (n,f) ~1 Unknown Li matrix material affects LiOH neutron production and thermalization LiH LiH 2 Be Intensity Large variation in spectra from each element 1E0 1E-5 0 5 10 Energy (MeV) 5 Li O Be

CURRENT SPECTRA Measured single sources Tagziria 2003 2004 Obninsk 1998 Birch 1984 Ing 1981 Owen 1982 and more Probability (MeV -1 ) 10 1 0.1 Measured: Tagziria 2003 Obninsk 1998 Calculated: GVZ 1971 Beddingfield 2015 Calculated initial spectra Geiger and van der Zwan 1971 Tagziria 2012 Beddingfield 1999 2015 0.01 0 1 2 3 4 Energy (MeV) Selected AmLi spectra. The high-energy tail is key for a good fit 6

MEASUREMENT OVERVIEW Sources Cf-252 Type 3 A7-series 3 Cf-series AmLi 3 Gammatron C-series 8 Gammatron N-series 12345 5-Ring Multiplicity Counter (5RMC) gives singles rate by ring Higher neutron energy is required to reach the outer rings 7

MEASUREMENT RESULTS VARIATION BETWEEN SOURCES 0.360 0.505 0.355 R1 R1 average 0.500 R1 R1 average 0.350 0.495 Cf-252 R1 0.345 0.340 AmLi R1 0.490 0.485 0.335 0.480 0.330 A7-866 A7-867 A7-869 Cf-9 Cf-11 Cf-12 Cf-252 Source Cf-252 inner ring ratios Statistical uncert. 0.11% Standard deviation 0.12% Chi-squared 1.41 E-6 0.475 C-470 C-476 C-477 - N-160 N-161 AmLi N-165 N-262 Source N-263 N-432 N-436 N-437 AmLi inner ring ratios Statistical uncert. 0.01% Standard deviation 0.63% Chi-squared (C) 1.73 E-5 Chi-squared (N) 6.22 E-5 8

MEASUREMENT RESULTS TARGETS FOR SPECTRA FITTING 0.505 0.500 R1 R1 average 0.0085 R5 R5 average 0.0080 0.495 AmLi R1 0.490 AmLi R5 0.0075 0.485 0.0070 0.480 0.475 C-470 C-476 C-477 - N-160 N-161 AmLi N-165 N-262 Source AmLi inner ratios N-263 N-432 N-436 N-437 0.0065 Ratio C-series N-series R1 0.495414 0.490165 C-470 C-476 C-477 - N-160 N-161 AmLi N-165 N-262 source N-263 N-432 N-436 N-437 AmLi outer ratios R2 0.353878 0.354726 R3 0.114653 0.117043 R4 0.028918 0.03027 R5 0.007147 0.007798 Ratios also exist for individual sources 9

SIMULATIONS 5RMC MCNP model benchmarked to Cf-252 measurements Chi-squared ring ratio agreement of 9.0E-6 5RMC neutron energy 100-bin monoenergetic response function for AmLi matrix MCNP6 Alpha energy for different particle sizes MCNP6 Li (α,n) neutron energy spectra for different particle sizes SOURCES4C Energy spectra * response function = ring ratios 10

SIMULATIONS 0.6 Arb. 0.1 0.01 0.001 1E-4 1E-5 min (calculated) max (calculated) min (extended) max (extended Efficiency 0.5 0.4 0.3 0.2 T R1 R2 R3 R4 R5 1E-6 1E-7 0.1 1E-8 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 Energy (MeV) 0.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 Energy (MeV) Li (α,n) neutron energy spectrum range between 1 and 7 microns 5RMC response function 11

SPECTRA FITTING Minimize chi-squared of simulated and measured ring ratios - adjust oxygen contribution and spectra within expanded limits 10 10 C-series spectrum N-series spectrum Probability (MeV -1 ) 1 0.1 Probability (MeV -1 ) 1 0.1 0.01 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 Energy (MeV) 0.01 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 Energy (MeV) Parameter C-series N-series Oxygen contribution 6.7% 8.1% Chi-squared fit 1.3 E-6 1.4 E-6 12

LIMITATIONS Chi-squared value only relates to Cf-252 agreement, no physical meaning MCNP model bias of chi-squared 9.0 E-6, with 1.4 E-6 disagreement between Cf-252 sources We assume the spectra fitting accounts for the unknown AmO 2 particle size, matrix composition, other effects Need to benchmark simulations to UNCL measurements and characterize spectrum effects on U-235 mass in a UNCL 13

CONCLUSIONS Characterized the variation in AmLi sources compared to Cf-252 Generation of spectra that are precise enough to distinguish between sources Improved accuracy in modeling AmLi active interrogation systems Measurement of non-calibrated samples Future work find effects on UNCL measurement results 14

ACKNOWLEDGEMENTS This work was funded in-part by the Consortium for Verification Technology under Department of Energy National Nuclear Security Administration award number DE- NA0002534 This work was funded in-part by the Department of Energy National Nuclear Security Administration Office of Nonproliferation and International Security 15

LA-UR-16-27725 MONTE CARLO FITTING OF THE AMLI NEUTRON SPECTRUM Robert Weinmann-Smith University of Florida, Los Alamos National Laboratory David Beddingfield Los Alamos National Laboratory Andreas Enqvist University of Florida Martyn Swinhoe Los Alamos National Laboratory

COMPARISON OF SPECTRA 10 10 Probability (MeV -1 ) 1 0.1 Tagziria 2003 Obninsk 1998 GVZ 1971 Beddingfield 2015 Probability (MeV -1 ) 1 0.1 C-series N-series 0.01 0 1 2 3 4 Energy (MeV) 0.01 0 1 2 3 4 Energy (MeV) 17

SUMMARY SLIDES FOLLOW: 18

MONTE CARLO FITTING OF THE AMLI NEUTRON SPECTRUM AmLi sources are used for active interrogation of fresh uranium fuel for treaty verification Characterization of the differences between AmLi sources Cf-252 R1 0.360 0.355 0.350 0.345 0.340 R1 (error bars are Cf-252 counting statistics) R1 average (error bars are AmLi Rel. Std. Dev.) AmLi R1 0.505 0.500 0.495 0.490 0.485 R1 (error bars are AmLi counting statistics) R1 average (error bars are Cf-252 Rel. Std. Dev.) 0.335 0.330 A7-866 A7-867 A7-869 Cf-9 Cf-11 Cf-12 Cf-252 Source Source 0.480 0.475 C-470 C-476 C-477 Chi-squared agreement - N-160 N-161 N-165 AmLi Source N-262 N-263 N-432 N-436 N-437 Cf-252 1.41 E-6 AmLi C-series 1.73 E-5 AmLi N-series 6.22 E-5 19

MONTE CARLO FITTING OF THE AMLI NEUTRON SPECTRUM Spectra specific to different series of AmLi sources, and to individual sources Improved accuracy in modeling AmLi Probability (MeV -1 ) 10 1 0.1 C-series N-series active interrogation systems 0.01 0 1 2 3 4 Energy (MeV) Parameter C-series N-series Oxygen contribution 6.7% 8.1% Chi-squared fit 1.3 E-6 1.4 E-6 20