Chemometric software for the automated detection of ignitable liquids in casework fire debris, ongoing developments
James J Harynuk
University of Alberta, Canada
: Forensic Toxicol Pharmacol 2015, 4:4
Abstract
Ongoing research in our group is focused on the development of tools for the rapid, automated identification of ignitable liquids in fire debris. We have previously demonstrated the automated classification of simulated and casework debris samples on the basis of gasoline content. Herein we present current research into the development of a suitable method to simulate fire debris to generate a background matrix that is suitable for the training of chemometric models for ignitable liquids other than gasoline. The simulation method is based on batch pyrolysis of materials under controlled temperatures and atmospheres in a tube furnace. Apart from hopefully generating a more realistic debris matrix without resorting to more expensive options such as the use of large burn cells, this method is also shedding light on the combustion/pyrolysis chemistry of household materials. We are also presenting a prototype classification software client for forensic laboratories developed in collaboration with the Royal Canadian Mounted Police. The software client provides a simple user interface for loading GC- MS data files performing chromatographic alignment and application of models to identify ignitable liquid signatures in the data. The software then reports what (if any) ignitable liquids were found in the debris along with a probability score for the identification and QA/QC reports for the alignment and data quality.
Biography
