Experimental Validation of Nuclear Forensics Methodologies for Reactor-type Attribution,
Burnup Determination, and Time Since Irradiation Estimation.

Nuclear forensics (NF) is a critical field of nuclear security and nonproliferation that must continually grow and improve to combat potential threats. Two NF methodologies developed at Texas A&M University (TAMU), Maximum Likelihood Methodology (MLM) and Machine Learning Technique (MLT), are capable of determining three key parameters of interdicted irradiated nuclear material. The three parameters are reactor-type, fuel burnup, and the time since fuel irradiation was completed (TSI). The current database at TAMU for the NF methodologies contain information from eight reactor types: pressured water reactor (PWR); pressurized heavy water reactor (PHWR); fast breeder reactor (FBR); fast flux test facility (FFTF); Canadian national research experimental (NRX); MAGNOX; high flux isotope research reactor (HFIR); and the University of Missouri Research Reactor (MURR). Additionally, data for each reactor spans burnup values up to 5 GWd/MTU (to simulate weapons-grade plutonium) and radioactive decay calculations up to 5000 days post-irradiation. All the data points for these parameters were generated by creating Monte Carlo N-Particle (MCNP) radiation transport models of these nuclear reactors and by performing fuel burnup simulations. To determine the above three parameters of interest, the MLM and MLT methodologies require intra-element isotopic ratios of plutonium and fission products, namely: 137/ 133Cs, 134/ 137Cs, 135/ 137Cs, 154/ 153Eu, 150/ 149Sm, 152/ 149Sm, 240/ 239Pu, and 241/ 239Pu. Mass spectrometry is the best tool to measure the required isotopic fractions and, in this study, inductively coupled plasma mass spectrometry (ICP-MS) was used. The NF (MLM and MLT) methodologies had previously been validated using experimental data from post-irradiation examinations (PIE). One PIE was completed for depleted UO2 (DUO2) irradiated in HFIR to ~5 GWd/MTU and another for natural UO2 (NatUO2) irradiated in MURR to ~1 GWd/MTU. A third validation dataset for low enriched uranium dioxide (LEUO2) has been done for the work presented here. The results showed that both methodologies can accurately predict reactor-type and burnup. Additional steps such as adding another fission product ratio was required to improve the accuracy of the TSI calculations. The results support that MLM and MLT methodologies are powerful and beneficial tools in the NF repertoire.