| AVS 60th International Symposium and Exhibition | |
| Applied Surface Science | Tuesday Sessions |
| Session AS+BI-TuA |
| Session: | Forensic Science, Art and Archaeology (2:00-3:20 pm)/Quasicrystals and Complex Metal Alloys (4:00-6:00 pm) |
| Presenter: | P.K. Weber, Lawrence Livermore National Laboratory |
| Authors: | P.K. Weber, Lawrence Livermore National Laboratory M.L. Davisson, Lawrence Livermore National Laboratory C.E. Ramon, Lawrence Livermore National Laboratory S.P. Velsko, Lawrence Livermore National Laboratory |
| Correspondent: | Click to Email |
The threat associated with the potential use of radiological, nuclear, chemical and biological materials in terrorist acts has resulted in new fields of forensic science using state-of-the-science analytical techniques. One such method is high spatial resolution secondary ion mass spectrometry (SIMS) performed with a Cameca NanoSIMS 50. This instrument allows us to extract quantitative trace element and isotopic information for forensic purposes at a resolution of ~100 nanometers. With this capability, target particles with contaminated or ultra-trace samples can be analyzed. Here we present a general validation scheme and extend it to the analysis of small numbers (3-10) of anthrax spores. The scheme tests the hypothesis that two samples were produced in the same laboratory by the same process. This test is generally known as “sample matching”, though the term “match” is not used. Instead, the test uses receiver-operating characteristics (ROC) curves derived from test samples to generate a likelihood ratio that is combined with other data relating to the hypothesis. For our work, we are using the elemental composition of single anthrax spores as evidence.
To evaluate the sample matching test we used a well-defined statistical design to generate B. anthracis samples that are representative of agents made by benchtop scale processes that might be encountered in terrorism events. We used the NanoSIMS to profile the elemental composition of individual bacterial spores. Between 8 and 20 elements were monitored. The dynamic range of the analyses was on the order of one million. An objective metric for the “closeness” of two samples was defined in terms of the differences between elemental concentrations. A ROC curve for the same lab-same process and same batch hypothesis tests were calculated based on the difference metric. The ROC curves for averages of small numbers of spores (3-10) were determined, and were compared to bulk elemental analysis.
Our results support the following conclusions: The average elemental composition of two samples containing a small number of spores (3 - 10) can have reasonable inferential power for determining if the samples were made in the same laboratory using the same process. The ROC curve for determining if two samples originate from the same batch offer less inferential power than those for the same lab - same process test. ROC curves for NanoSIMS-based comparisons are optimized by using a particular set of elements, but the dependence on element set is relatively weak.