Application of Laser-Induced Breakdown Spectroscopy to Forensic Science: Analysis of Paint Samples
Michael E. Sigman, Ph.D., Erin M. McIntee, M.S., Candice Bridge, Ph.D.
Abstract:
A comparison has been made between laser induced breakdown spectroscopy (LIBS), X-ray fluorescence (XRF), scanning electron microscopy – energy dispersive X-ray spectroscopy (SEM/EDS) and laser ablation – inductively coupled plasma – mass spectrometry (LA-ICP-MS) for the discrimination of automotive paint samples. All discriminations were performed by hypothesis testing at the a = 0.05 significance level using both parametric and nonparametric statistical tests. Discrimination was tested across all paint samples, irrespective of paint color or other features, and in a more forensically relevant fashion, discrimination was determined for only those samples of the same color group, number of paint layers and the presence or absence of effect pigments in the paint. The paint samples came from automobiles manufactured in years 1985 – 2006, representing both original equipment manufacturers paint and repaint samples. The samples came from color groups that included black, blue, green, red, silver, tan and white. A total of 200 paint samples, comprising one group of 110 samples and one group of 90 samples were examined by the different analytical methods. Not all samples were analyzed by each method.
LIBS was determined to have a discrimination power of 90% (10% Type II errors) at a verified 5% Type I error rate. Discrimination was found to be slightly lower (86.6%) among the white color group. Variations in the LIBS signal over time led to same sample discriminations and an artificially high Type I error rate, which was overcome by attention to the sampling protocol and confining spectra collection on samples that were to be compared to a narrow time window. LA-ICP-MS was determined to give the best sample discrimination (100%), with XRF and SEM/EDS giving the lowest discriminations, 85% and 73% total discrimination respectively for each method. The results of this study suggest that LIBS may provide an important screening tool in the analysis of automotive paint samples; however, careful attention to sampling protocols and statistical comparison of samples is recommended. In cases where two samples can not be discriminated, a more accurate comparison method or multiple comparison methods should be sought.