This paper describes a research project in which Dr. Michael Sigman and Mary Williams of the University of Central Florida developed a method for analyzing fire debris for ignitable liquids in samples that are highly contaminated with pyrolysis interferences.
The method averages mass spectra across an entire chromatographic run. It can effectively classify ignitable liquids according to the widely accepted ASTM International classification scheme. Fire-debris analysts currently render opinions as to whether or not a fire debris sample tested positive for ignitable liquid residue without stating a known error rate for the assessment. Use of the new method, however, assists analysts in estimating an objective measure of confidence based on established statistical models. The research team developed a chemometric data analysis method to facilitate the identification of gas chromatography-mass spectral patterns associated with ignitable liquid classes. A novel method was developed based on target factor analysis (TFA) and Bayesian decision theory that classifies ignitable liquid residue in the presence of background interferences found in the debris. The testing of this new method of fire debris analysis correctly identified the presence of ignitable fluid 80-85 percent of the time.