Discrimination of Ignitable Liquid Residues in Burned Petroleum-Derived Substrates by Using HS-MS eNose and Chemometrics

Implementation of SPME and Rapid GC-MS as a Screening Approach for Forensic Fire Debris Applications

Analysis of ignitable liquids in fire debris samples can be a time-consuming process, from extraction of volatile compounds to instrumental analysis. Rapid gas chromatography-mass spectrometry (GC-MS) is a screening technique that can be utilized prior to confirmatory GC-MS analysis to provide an informative screening approach and possibly reduce the need to further analyze negative samples. Though rapid GC-MS is fast (less than two minutes), extraction techniques such as passive headspace extraction remain a bottleneck for decreasing overall workflow times. In this work, solid phase microextraction (SPME) was implemented with rapid GC-MS for ignitable liquid analysis for a faster, more sensitive screening approach compared to extraction with passive headspace. Using optimized inlet conditions, limits of detection as low as 27 ng/mL per compound were achieved. Gasoline and diesel fuel were extracted and analyzed, and major compounds in each liquid were identified in the resulting chromatograms. Extracted ion profiles (EIPs) and deconvolution methods were useful for additional compound identifications. Lastly, the SPME-rapid GC-MS workflow was extended to the analysis of gasoline and diesel fuel in mock burn samples using carpet and wood substrates. From SPME sample extraction to rapid GC-MS instrumental analysis and data processing, the total workflow for a single sample was reduced to under 20 min. These results indicate that SPME is a suitable injection technique for rapid GC-MS to provide a fast and sensitive screening approach for fire debris applications.

Rapid GC-MS as a Screening Tool for Forensic Fire Debris Analysis

Techniques developed for the screening of forensic samples can be useful for increasing sample throughput and decreasing backlog in forensic laboratories. One such technique, rapid gas chromatography mass spectrometry (GC-MS), allows for fast sample screening (≈1 min) and has gained interest in recent years for forensic applications. This work focuses on the development of a method for ignitable liquid analysis using rapid GC-MS. A sampling protocol and temperature program were developed for the analysis of these volatile samples. Using the optimized method for analysis, the limits of detection for compounds commonly found in ignitable liquids ranged from 0.012 mg/mL to 0.018 mg/mL. Once the method was developed, neat ignitable liquids (i.e., gasoline and diesel fuel) were analyzed, and major components in each liquid were identified. The identification of major compounds in gasoline and diesel fuel in the presence of substrate interferences was then assessed through the analysis of simulated fire debris samples. Three different substrates were spiked with each ignitable liquid, burned, and analyzed. Major compounds in both liquids were identified using the total ion chromatograms, relevant extracted ion profiles, and deconvolution methods.

Recognition of gasoline in fire debris using machine learning: Part I, application of random forest, gradient boosting, support vector machine, and naïve bayes

The detection and identification of ignitable liquid (IL) residues in fire debris are two very challenging tasks in a fire investigation. To this day, the recognition of IL in fire debris includes the chemical analysis of the fire debris composition, followed by the examination and interpretation of the analysis result by a trained forensic examiner. Throughout the last decade, chemometrics and artificial intelligence have become increasingly important. In the present study, machine learning algorithms capable of recognizing gasoline residues in fire debris based on GC-MS data have been developed. Four methods, including random forest, gradient boosting, support vector machine, and naïve bayes are applied and used to classify fire debris samples into the two categories "with gasoline" or "without gasoline". A fifth method (logistic regression) did not converge due to well separated classes. A database comprising 360 measurements, including fire debris samples of real cases as well as fire debris samples spiked with known amounts of weathered gasoline (up to 99.6%), was available to train the machine learning algorithms (using 85% of the data) and to subsequently test the performance of the methods when classifying unknown samples (using 15% of the data). In general, the methods perform very well, as three of it succeeded to classify all test samples correctly without any false positive or false negative allocations. One (naïve bayes) was not trained enough to classify other (non-gasoline) IL correctly as "no gasoline". Furthermore, the random forest method reveals which chemical compounds are most relevant for the algorithm to classify the samples. In general, the presented approach is highly promising and could easily be extended or adapted to other types of IL. Similar to the neural network presented in the accompanying paper, such methods have the potential to serve as a fast screening technique for fire debris samples, thus supporting the forensic examiner by providing an additional independent opinion. Nonetheless, the definite identification of IL residues in fire debris always has to be accomplished by a forensic examiner.