Achieving a Near-Theoretical Maximum in Peak Capacity Gain for the Forensic Analysis of Ignitable Liquids Using GC×GC-TOFMS

Assessment of artificial intelligence to detect gasoline in fire debris using HS-SPME-GC/MS and transfer learning

Due to the complex nature of the chemical compositions of ignitable liquids (IL) and the interferences from fire debris matrices, interpreting chromatographic data poses challenges to analysts. In this work, artificial intelligence (AI) was developed by transfer learning in a convolutional neural network (CNN), GoogLeNet. The image classification AI was fine-tuned to create intelligent classification systems to discriminate samples containing gasoline residues from burned substrates. All ground truth samples were analyzed by headspace solid-phase microextraction (HS-SPME) coupled with a gas chromatograph and mass spectrometer (GC/MS). The HS-SPME-GC/MS data were transformed into three types of image presentations, that is, heatmaps, extracted ion heatmaps, and total ion chromatograms. The abundance and mass-to-charge ratios of each scan were converted into image patterns that are characteristic of the chemical profiles of gasoline. The transfer learning data were labeled as "gasoline present" and "gasoline absent" classes. The assessment results demonstrated that all AI models achieved 100 ± 0% accuracy in identifying neat gasoline. When the models were assessed using the spiked samples, the AI model developed using the extracted ion heatmap obtained the highest accuracy rate (95.9 ± 0.4%), which was greater than those obtained by other machine learning models, ranging from 17.3 ± 0.7% to 78.7 ± 0.7%. The proposed work demonstrated that the heatmaps created from GC/MS data can represent chemical features from the samples. Additionally, the pretrained CNN models are readily available in the transfer learning workflow to develop AI for GC/MS data interpretation in fire debris analysis.

Review: Recent advancements and moving trends in chemical analysis of fire debris

This review describes recent advances and current trends in fire debris analysis from 2014 to 2021. Onsite analytical techniques used for fire scene investigation, identifying samples of interest for later analysis as well as onsite confirmatory techniques are examined. Laboratory techniques are reviewed both from a perspective of instrumentation and data analysis. Advances in analytical techniques include GC x GC-TOFMS, DART-MS, HS-GC-IMS. New and emerging methods of data analysis including those using machine learning are assessed. Each aspect is essential for forensic scientists to obtain the correct conclusion when collecting, examining, analysing, and interpreting fire debris. This review concludes that there is a need for the validity and certainty of all methods to be assessed if they are to be used to generate reports or draw conclusions.

Simulating comprehensive two-dimensional gas chromatography mass spectrometry data with realistic run-to-run shifting to evaluate the robustness of tile-based Fisher ratio analysis

Untargeted analysis of comprehensive two-dimensional (2D) gas chromatography time-of-flight mass spectrometry (GC×GC-TOFMS) data has the potential to be hindered by run-to-run retention time shifting. To address this challenge, tile-based Fisher ratio (F-ratio) analysis (FRA) has been developed, which utilizes a supervised, untargeted approach involving a chromatographic segmentation routine termed "tiling" combined with the ANOVA F-ratio statistic to discover class-distinguishing analytes while minimizing false positives arising from shifting. The tiling algorithm is designed to account for retention shifting in both separation dimensions. Although applications of FRA have been reported, there remains a need to thoroughly evaluate the robustness of FRA for different levels of run-to-run retention shifting in order to broaden the scope of its application. To this end, a novel method of simulating GC×GC-TOFMS chromatograms with realistic run-to-run shifting is presented by random generation of low-frequency "shift functions". The dimensionless retention-time precision, <δ_r>, which is four times the standard deviation in retention time normalized to the peak width-at-base is used as a key modeling variable along with the 2D chromatographic saturation, α_e,2D, and within-class relative standard deviation in peak area, RSD_wc. We demonstrate that all three of these variables operate together to impact true positive discovery. To quantify the "success" of true positive discovery, GC×GC-TOFMS datasets for various combinations of <δ_r>, α_e,2D, and RSD_wc were simulated and then analyzed by FRA using a wide range of relative tile areas (RTA), which is a dimensionless measure of tile size. Since each hit in the FRA hit list was known a priori as either a true or false positive based on the simulation inputs, receiver operating characteristic (ROC) curves were readily constructed. Then, the area under the ROC curve (AUROC) was used as a metric for discovery "success" for various combinations of the modeling variables. Based on the results of this study, recommendations for tile size selection and experimental design are provided, and further supported by comparison to previous tile-based FRA applications. For instance, values for <δ_r>, α_e,2D, and RSD_wc obtained from a GC×GC-TOFMS dataset of yeast metabolites suggested an optimum RTA of 6.25, corresponding closely to the RTA of 4.00 employed in the study, implying the simulation results obtained here can be generalized to real datasets.

Profiling Olefins in Gasoline by Bromination Using GC×GC-TOFMS Followed by Discovery-Based Comparative Analysis

An analytical workflow for the analysis of olefins in gasoline that combines selective bromination and comprehensive two-dimensional (2D) gas chromatography time-of-flight mass spectrometry (GC×GC-TOFMS) with discovery-based analysis is reported. First, a standard mix containing n-alkanes, 1-alkenes, and aromatic species was brominated and quantified using % reacted as a metric for each compound class, defined as the difference in the total peak area between the brominated and original samples normalized to the original sample. The average % reacted (1 s.d.) values were -1.45% (2.8%) for the alkanes, 99.5% (0.4%) for the alkenes, and 6.7% (11.6%) for the aromatics, demonstrating excellent selectivity toward the alkenes with only minor aromatic bromination. The bromination chemistry was then applied to gasoline, followed by GC×GC-TOFMS analysis of the original and brominated gasoline. This GC×GC-TOFMS data set was then submitted to the supervised discovery tool tile-based F-ratio analysis (FRA), which reduced the large data set to only the chromatographic regions which distinguish between the original and brominated gasoline samples. FRA discovered 314 hits, 56 of which were determined statistically significant using combinatorial null distribution analysis (CNDA), a permutation-based significance test. Since the brominated olefins elute in an uncrowded region of the 2D chromatogram and have no signal in the original sample, their discoverability was greatly increased relative to the original olefins. By combining the information gained from brominated olefin standards and the structured patterns of the GC×GC separations, the top hits were identified as the dibromoalkane products of mono-olefins, with five C5 mono-olefins identified on a species level. The analytical workflow has broad implications for using selective reaction chemistries to facilitate supervised discovery by targeting desired compound classes.

Method development for optimizing analysis of ignitable liquid residues using flow-modulated comprehensive two-dimensional gas chromatography

The abundance and composition of matrix compounds in fire debris samples undergoing ignitable liquid residue analysis frequently leads to inconclusive results, which can be diminished by applying comprehensive two-dimensional gas chromatography (GC × GC). Method development must be undertaken to fully utilize the potential of GC × GC by maximizing separation space and resolution.. The three main areas to consider for method development are column selection, modulator settings and parameter optimization. Seven column combinations with different stationary phase chemistry, column dimensions and orthogonality were assessed for suitability based on target compound selectivity, retention, resolution, and peak shapes, as well as overall peak capacity and area use. Using Box-Behnken design of experimentation (DoE), the effect of modulator settings such as flow ratio and loop fill capacity were evaluated using carbon loading potential, dilution effect, as well as target peak amplitude and skewing effect. The run parameters explored for parameter optimization were oven programming, inlet pressure (column flow rate), and modulation period. Comparing DoE approaches, Box-Behnken and Doehlert designs assessed sensitivity, selectivity, peak capacity, and wraparound; alongside target peak retention, resolution, and shape evaluation. Certified reference standards and simulated wildfire debris were used for method development and verification, and wildfire debris case samples scrutinized for method validation. The final method employed a low polarity column (5% diphenyl) coupled to a semi-polar column (50% diphenyl) and resulted in an average Separation Number (SN) exceeding 1 in both dimensions after optimization. Separation Numbers of 18.16 for first and 1.46 for second dimension without wraparound for compounds with at least four aromatic rings signified successful separation of all target compounds from varied matrix compositions and allowed for easy visual comparison of extracted ion profiles. Mass spectrometry (MS) was required during validation to differentiate ions where no baseline separation between target compounds and extraneous matrix compounds was possible. The resulting method was evaluated against ASTM E1618 and found to be an ideal routine analysis method providing great resolution of target compounds from interferences and excellent potential for ILR classification within a complex sample matrix.

Strategies for the identification and sensory evaluation of volatile constituents in wine

Wine aroma, which stems from complex perceptual and cognitive processes, is initially driven by a multitude of naturally occurring volatile constituents. Its interpretation depends on the characterization of relevant volatile constituents. With large numbers of volatile constituents already identified, the search for unknown volatiles in wine has become increasingly challenging. However, the opportunities to discover unknown volatile compounds contributing to the wine volatilome are still of great interest, as demonstrated by the recent identification of highly odorous trace (µg/L) to ultra-trace (ng/L) volatile compounds in wine. This review provides an overview of both existing strategies and future directions on identifying unknown volatile constituents in wine. Chemical identification, including sample extraction, fractionation, gas chromatography, olfactometry, and mass spectrometry, is comprehensively covered. In addition, this review also focuses on aspects related to sensory-guided wine selection, authentic reference standards, artifacts and interferences, and the evaluation of the sensory significance of discovered wine volatiles. Powerful key volatile odorants present at ultra-trace levels, for which these analytical approaches have been successfully applied, are discussed. Research areas where novel wine volatiles are likely to be identified are pointed out. The importance of perceptual interaction phenomena is emphasized. Finally, future avenues for the exploration of yet unknown wine volatiles by coupling analytical approaches and sensory evaluation are suggested.

Development of retention time indices for comprehensive multidimensional gas chromatography and application to ignitable liquid residue mapping in wildfire investigations

In this study, we introduce a simple three-step workflow for a universally applicable RI system, to be used in GC×GC analysis of ignitable liquid residue (ILR) for arson investigations. The proposed RI system applies a combination of two well-established GC RI systems: non-isothermal Kovats (K) index in the first dimension and Lee (L) index in the second dimension. The proposed KLI RI system showed very good correlations when compared with predicted values and existing RI systems (r² = 0.97 in first dimension, r² = 0.99 in second dimension) and was valid for a wide range of analyte concentrations and operational settings (coefficient of variance (CV) < 1% in first dimension, < 10% in second dimension). Using the KLI RI, an ILR classification contour map was created to assist with the identification of ILR types within ASTM E1618. The contour map was successfully applied to neat fuels and a fire scene sample, highlighting the application to wildfire investigation. Standardizing the RI process and establishing acceptable error margins allows the exploration and comparison of comprehensive data generated from GC×GC analysis of ILRs regardless of location, time, or system, further enhancing comprehensive and tenable chemometric analyses of samples. Overall, the KLI workflow was inexpensive, quick to apply, and user-friendly with its simple 3-step design.

Translation of a One-Dimensional to a Comprehensive Two-Dimensional Gas Chromatography Method with Dual-Channel Detection for Volatile Organic Compound Measurement in Forensic Applications

After its introduction in the early 1990s, comprehensive two-dimensional gas chromatography (GC×GC) has evolved from a separation science research tool to the central component of many industries. Despite the maturity of the technique, some fields remain reluctant to its use in routine applications. In the case of forensic science, some constraints are the strict requirements enforced in forensic laboratories and the time and effort that must be invested for intralaboratory method validation. Concerns may also arise about whether information could be lost when transitioning to a new technique. This study reports on a method translation from conventional one-dimensional (1D) GC to GC×GC, ensuring the integrity of data as conversion is made. The GC was retrofitted with a reverse fill/flush (RFF) flow modulator and equipped with dual-channel detection using a quadrupole mass spectrometer (qMS) and a flame ionization detector (FID). The parallel use of two detectors, where qMS was applied for qualitative identification and FID for quantification, allowed higher flows and slightly wider peaks to be exploited for the analysis of a volatile organic compound (VOC) reference mixture relevant to forensic VOC profiling. Peak quality assessment and calibration curves using GC-qMS and GC×GC-qMS/FID document the transfer and adaptation of the original method without a loss in data quality. Furthermore, the preprocessing and the data analysis processing steps, including calibration and peak quality assessment for each of the three data sets, are explained in detail. This information provides benchmark data for routine laboratories that want to implement a GC×GC approach into routine workflows.

Developing a Method for the Collection and Analysis of Burnt Remains for the Detection and Identification of Ignitable Liquid Residues Using Body Bags, Dynamic Headspace Sampling, and TD-GC×GC-TOFMS

In cases of suspected arson, a body may be intentionally burnt to cause loss of life, dispose of remains, or conceal identification. A primary focus of a fire investigation, particularly involving human remains, is to establish the cause of the fire; this often includes the forensic analysis of fire debris for the detection of ignitable liquid residues (ILRs). Commercial containers for the collection of fire debris evidence include metal cans, glass jars, and polymer/nylon bags of limited size. This presents a complication in cases where the fire debris consists of an intact, or partially intact, human cadaver. This study proposed the use of a body bag as an alternative sampling container. A method was developed and tested for the collection and analysis of ILRs from burnt porcine remains contained within a body bag using dynamic headspace sampling (using an Easy-VOC™ hand-held manually operated grab-sampler and stainless steel sorbent tubes containing Tenax TA) followed by thermal desorption comprehensive two-dimensional gas chromatography–time-of-flight mass spectrometry (TD-GC×GC-TOFMS). The results demonstrated that a body bag containing remains burnt with gasoline tested positive for the presence of gasoline, while blank body bag controls and a body bag containing remains burnt without gasoline tested negative. The proposed method permits the collection of headspace samples from burnt remains before the remains are removed from the crime scene, limiting the potential for contamination and the loss of volatiles during transit and storage.

Detection and Characterization of Ignitable Liquid Residues in Forensic Fire Debris Samples by Comprehensive Two-Dimensional Gas Chromatography

This study covers an extensive experimental design that was developed for creating simulated fire debris samples under controlled conditions for the detection and identification of ignitable liquids (IL) residues. This design included 19 different substrates, 45 substrate combinations with and without ignitable liquids, and 45 different ILs from three classes (i.e., white spirit, gasoline, and lamp oil). Chemical analysis was performed with comprehensive two-dimensional gas chromatography coupled to time-of-flight mass spectrometry (GC×GC-TOFMS) for improved separation and compound identification. The enhanced peak capacity offered by GC×GC-TOFMS allowed the use of a target compound list in combination with a simple binary decision model to arrive at quite acceptable results with respect to IL detection (89% true positive and 7% false positive rate) and classification (100% correct white spirit, 79% correct gasoline, and 77% correct lamp oil assignment). Although these results were obtained in a limited set of laboratory controlled fire experiments including only three IL classes, this study confirms the conclusions of other studies that GC×GC-TOFMS can be a powerful tool in the challenging task of forensic fire debris analysis.