Detecting Chemical Vapor Diffusion through Firefighter Turnout Gear.

Firefighters are exposed to burning materials that may release toxic partial combustion and pyrolysis products into the environment, including compounds listed as priority pollutants by the United States Environmental Protection Agency (EPA). A novel passive sampling dosimeter device containing firefighter turnout gear as a diffusion membrane and an activated charcoal strip (ACS) for volatile analyte collection was designed and used to monitor potential exposures of firefighters to volatile organic compounds. Solvent extracts from the ACS and turnout gear diffusion layer were analyzed using Gas Chromatography-Mass Spectrometry (GC-MS) to determine the diffusion of compounds from burned substrates through firefighter turnout gear and compound adsorption to the turnout gear. The compounds in these samples were identified using target factor analysis (TFA). An activated carbon layer (ACL) was added to the dosimeter between the turnout gear and the ACS. The presence of combustion and pyrolysis compounds identified on the ACS in the dosimeter was reduced.

Assessing evidentiary value in fire debris analysis by chemometric and likelihood ratio approaches.

Results are presented from support vector machine (SVM), linear and quadratic discriminant analysis (LDA and QDA) and k-nearest neighbors (kNN) methods of binary classification of fire debris samples as positive or negative for ignitable liquid residue. Training samples were prepared by computationally mixing data from ignitable liquid and substrate pyrolysis databases. Validation was performed on an unseen set of computationally mixed (in silico) data and on fire debris from large-scale research burns. The probabilities of class membership were calculated using an uninformative (equal) prior and a likelihood ratio was calculated from the resulting class membership probabilities. The SVM method demonstrated a high discrimination, low error rate and good calibration for the in silico validation data; however, the performance decreased significantly for the fire debris validation data, as indicated by a significant increase in the error rate and decrease in the calibration. The QDA and kNN methods showed similar performance trends. The LDA method gave poorer discrimination, higher error rates and slightly poorer calibration for the in silico validation data; however the performance did not deteriorate for the fire debris validation data.

Assessing the evidentiary value of smokeless powder comparisons.

Gas chromatography-electron ionization-mass spectrometry (GC-EI-MS) and physical characteristics data for 726 smokeless reloading powders were analyzed by pairwise comparisons of samples comprising the same product and different products. Pairwise comparisons were restricted to samples having matching kernel shape, color, presence or absence of a perforation and measurements. Discrete results were analyzed for same and different products having matching chemical composition determined from a list of 13 organic components. A continuous score-based likelihood ratio was determined for same and different product comparisons using the Fisher transform of the Pearson correlation between the total ion spectra of the compared samples. Probability distributions for same product and different product comparisons appeared bimodal and were modeled with kernel density distributions. In the discrete and continuous data comparisons, the likelihood ratios for probabilities conditioned on same shape, color, presence/absence of perforation and size were found to provide relatively limited support for either the proposition of same product or different product. Further restricting the pairwise comparisons to samples belonging to the same cluster, as determined by agglomerative hierarchical cluster analysis, provided probability distributions for same product and different product comparisons that were more normal, but did not improve the resulting likelihood ratios. These results inform the forensic analyst regarding the evidentiary value of database search results and direct comparisons of recovered and control samples of smokeless powders.

Class-conditional feature modeling for ignitable liquid classification with substantial substrate contribution in fire debris analysis.

Forensic chemical analysis of fire debris addresses the question of whether ignitable liquid residue is present in a sample and, if so, what type. Evidence evaluation regarding this question is complicated by interference from pyrolysis products of the substrate materials present in a fire. A method is developed to derive a set of class-conditional features for the evaluation of such complex samples. The use of a forensic reference collection allows characterization of the variation in complex mixtures of substrate materials and ignitable liquids even when the dominant feature is not specific to an ignitable liquid. Making use of a novel method for data imputation under complex mixing conditions, a distribution is modeled for the variation between pairs of samples containing similar ignitable liquid residues. Examining the covariance of variables within the different classes allows different weights to be placed on features more important in discerning the presence of a particular ignitable liquid residue. Performance of the method is evaluated using a database of total ion spectrum (TIS) measurements of ignitable liquid and fire debris samples. These measurements include 119 nominal masses measured by GC-MS and averaged across a chromatographic profile. Ignitable liquids are labeled using the American Society for Testing and Materials (ASTM) E1618 standard class definitions. Statistical analysis is performed in the class-conditional feature space wherein new forensic traces are represented based on their likeness to known samples contained in a forensic reference collection. The demonstrated method uses forensic reference data as the basis of probabilistic statements concerning the likelihood of the obtained analytical results given the presence of ignitable liquid residue of each of the ASTM classes (including a substrate only class). When prior probabilities of these classes can be assumed, these likelihoods can be connected to class probabilities. In order to compare the performance of this method to previous work, a uniform prior was assumed, resulting in an 81% accuracy for an independent test of 129 real burn samples.

Hierarchical cluster analysis of ignitable liquids based on the total ion spectrum.

Gas chromatography-mass spectrometry (GC-MS) data of ignitable liquids in the Ignitable Liquids Reference Collection (ILRC) database were processed to obtain 445 total ion spectra (TIS), that is, average mass spectra across the chromatographic profile. Hierarchical cluster analysis, an unsupervised learning technique, was applied to find features useful for classification of ignitable liquids. A combination of the correlation distance and average linkage was utilized for grouping ignitable liquids with similar chemical composition. This study evaluated whether hierarchical cluster analysis of the TIS would cluster together ignitable liquids of the same ASTM class assignment, as designated in the ILRC database. The ignitable liquids clustered based on their chemical composition, and the ignitable liquids within each cluster were predominantly from one ASTM E1618-11 class. These results reinforce use of the TIS as a tool to aid in forensic fire debris analysis.

Application of self-organizing feature maps to analyze the relationships between ignitable liquids and selected mass spectral ions.

The unsupervised artificial neural networks method of self-organizing feature maps (SOFMs) is applied to spectral data of ignitable liquids to visualize the grouping of similar ignitable liquids with respect to their American Society for Testing and Materials (ASTM) class designations and to determine the ions associated with each group. The spectral data consists of extracted ion spectra (EIS), defined as the time-averaged mass spectrum across the chromatographic profile for select ions, where the selected ions are a subset of ions from Table 2 of the ASTM standard E1618-11. Utilization of the EIS allows for inter-laboratory comparisons without the concern of retention time shifts. The trained SOFM demonstrates clustering of the ignitable liquid samples according to designated ASTM classes. The EIS of select samples designated as miscellaneous or oxygenated as well as ignitable liquid residues from fire debris samples are projected onto the SOFM. The results indicate the similarities and differences between the variables of the newly projected data compared to those of the data used to train the SOFM.

Progress toward the determination of correct classification rates in fire debris analysis II: utilizing soft independent modeling of class analogy (SIMCA).

A multistep classification scheme was used to detect and classify ignitable liquid residues in fire debris into the classes defined by the ASTM E1618-10 standard method. The total ion spectra (TIS) of the samples were classified by soft independent modeling of class analogy (SIMCA) with cross-validation and tested on fire debris. For detection of ignitable liquid residue, the true-positive rate was 94.2% for cross-validation and 79.1% for fire debris, with false-positive rates of 5.1% and 8.9%, respectively. Evaluation of SIMCA classifications for fire debris relative to a reviewer's examination led to an increase in the true-positive rate to 95.1%; however, the false-positive rate also increased to 15.0%. The correct classification rates for assigning ignitable liquid residues into ASTM E1618-10 classes were generally in the range of 80-90%, with the exception of gasoline samples, which were incorrectly classified as aromatic solvents following evaporative weathering in fire debris.

Progress toward the determination of correct classification rates in fire debris analysis.

Principal components analysis (PCA), linear discriminant analysis (LDA), and quadratic discriminant analysis (QDA) were used to develop a multistep classification procedure for determining the presence of ignitable liquid residue in fire debris and assigning any ignitable liquid residue present into the classes defined under the American Society for Testing and Materials (ASTM) E 1618-10 standard method. A multistep classification procedure was tested by cross-validation based on model data sets comprised of the time-averaged mass spectra (also referred to as total ion spectra) of commercial ignitable liquids and pyrolysis products from common building materials and household furnishings (referred to simply as substrates). Fire debris samples from laboratory-scale and field test burns were also used to test the model. The optimal model's true-positive rate was 81.3% for cross-validation samples and 70.9% for fire debris samples. The false-positive rate was 9.9% for cross-validation samples and 8.9% for fire debris samples.

Discriminant analysis in the presence of interferences: combined application of target factor analysis and a Bayesian soft-classifier.

A method is described for performing discriminant analysis in the presence of interfering background signal. The method is based on performing target factor analysis on a data set comprised of contributions from analyte(s) and interfering components. A library of data from representative analyte classes is tested for possible contributing factors by performing oblique rotations of the principal factors to obtain the best match, in a least-squares sense, between test and predicted vectors. The degree of match between the test and predicted vectors is measured by the Pearson correlation coefficient, r, and the distribution of r for each class is determined. A Bayesian soft classifier is used to calculate the posterior probability based on the distributions of r for each class, which assist the analyst in assessing the presence of one or more analytes. The method is demonstrated by analyses performed on spectra obtained by laser induced breakdown spectroscopy (LIBS). Single and multiple bullet jacketing transfers to steel and porcelain substrates were analyzed to identify the jacketing materials. Additionally, the metal surrounding bullet holes was analyzed to identify the class of bullet jacketing that passed through a stainless steel plate. Of 36 single sample transfers, the copper jacketed (CJ) and non-jacketed (NJ) class on porcelain had an average posterior probability of the metal deposited on the substrate of 1.0. Metal jacketed (MJ) bullet transfers to steel and porcelain were not detected as successfully. Multiple transfers of CJ/NJ and CJ/MJ on the two substrates resulted in posterior probabilities that reflected the presence of both jacketing materials. The MJ/NJ transfers gave posterior probabilities that reflected the presence of the NJ material, but the MJ component was mistaken for CJ on steel, while non-zero probabilities were obtained for both CJ and MJ on porcelain. Jacketing transfer from a bullet to steel as the projectile passed through the steel also proved difficult to analyze. Over 50% of the samples left insufficient transfer to be identified. Transfer from NJ and CJ jacketing was successfully identified by posterior probabilities greater than 0.8.

Combined target factor analysis and Bayesian soft-classification of interference-contaminated samples: forensic fire debris analysis.

A bayesian soft classification method combined with target factor analysis (TFA) is described and tested for the analysis of fire debris data. The method relies on analysis of the average mass spectrum across the chromatographic profile (i.e., the total ion spectrum, TIS) from multiple samples taken from a single fire scene. A library of TIS from reference ignitable liquids with assigned ASTM classification is used as the target factors in TFA. The class-conditional distributions of correlations between the target and predicted factors for each ASTM class are represented by kernel functions and analyzed by bayesian decision theory. The soft classification approach assists in assessing the probability that ignitable liquid residue from a specific ASTM E1618 class, is present in a set of samples from a single fire scene, even in the presence of unspecified background contributions from pyrolysis products. The method is demonstrated with sample data sets and then tested on laboratory-scale burn data and large-scale field test burns. The overall performance achieved in laboratory and field test of the method is approximately 80% correct classification of fire debris samples.

Analysis of black writing ink by electrospray ionization mass spectrometry.

A method for the analysis of dyes and vehicles within writing inks was developed. The method was tested on a set of 18 black ink pens comprised of 6 ballpoint, 6 gel, and 6 rollerball pens. The sampling procedure utilized a small number of ink-coated paper fibers collected surreptitiously from the document, causing minimal damage and providing a sufficient quantity of ink for analysis. Methanol proved suitable for the extraction of ink components from ballpoint, gel and rollerball pens. Three separate electrospray ionisation-mass spectrometry (ESI-MS) methods were required to detect the dyes and vehicles from the inks. The ions present in the ESI-MS spectra at a signal-to-noise ratio of greater than 3:1 provided sufficient information to allow differentiation between the inks of each type. A tentative identification of the ink components was made based on a comparison of the ions present in the ink extract ESI-MS spectra with the ions present in a series of standards. The limits of detection for the standards were generally in the 2.5-10 ppm range. The method reported here extends the ESI-MS method of ink analysis to include gel and rollerball pens, includes the analysis of vehicles as well as dyes in the inks and demonstrates a minimally destructive sampling method that does not require a "biopsy" of the document.

Performance testing of commercial containers for collection and storage of fire debris evidence.

Fire debris evidence may contain ignitable liquid residues valuable in the investigation of a potential arson scene. The ability to obtain evidence containers that are contaminant-free and vapor-tight is essential to the analysis and storage of fire debris evidence. Commercial containers such as metal "paint" cans, glass mason jars, and polymer bags are often employed as fire debris evidence containers. The purpose of this research was to determine which of these three types of containers provided the most vapor-tight seal for the prevention of ignitable liquid vapor loss and to assess the potential for cross-contamination. Leak rates for each type of container were measured under controlled conditions. Simple mixtures of hydrocarbons were utilized in these experiments. Leak rates were determined based on the amounts of hydrocarbon recovered from activated charcoal located outside the test container and within a secondary container. Quantitation of the hydrocarbons recovered from activated charcoal was calculated using external standard calibration curves following analysis by gas chromatography-mass spectrometry. The results demonstrated that glass jars had the fastest leak rate followed by metal paint cans and properly heat-sealed polymer bags with the slowest leak rate. Each container exhibited a different leak mechanism, which resulted in an observable effect on the composition of hydrocarbons lost from the container. Hydrocarbon transfer from one container to another is also demonstrated. This study presents results that reveal the most vapor-tight container to be a properly heat-sealed copolymer bag.

Individualization of gasoline samples by covariance mapping and gas chromatography/mass spectrometry.

A set of 10 fresh (unevaporated) gasoline samples from a single metropolitan area were differentiated based on a covariance mapping method combined with a t-test statistic. The covariance matrix for each sample was calculated from the retention time-ion abundance data set obtained by gas chromatography/mass spectrometry analysis. Distance metrics were calculated between the covariance matrices from replicate analyses of the same sample and between the replicate analyses of different samples. The distance metric for the same-sample comparisons were shown to constitute a population significantly different from the distance metric for the different-sample comparisons. A power analysis was performed to estimate the number of analyses needed to discriminate between two samples while maintaining a probability of type II error, beta, below 1%, e.g., a test power greater than 99%. Triplicate analyses of two gasoline samples was shown to be sufficient to discriminate between the two using a t-test, while keeping beta<0.01 at a significance level, alpha, of 0.05. Analysis of the 45 possible pairwise comparisons between samples found that 100% of the samples were statistically distinguishable, and no type II errors occurred. Blind tests were conducted wherein 2 of the 10 gasoline samples where presented as unknowns. One of the unknowns was found to be indistinguishable from the original source, and one unknown was determined to be statistically different from the original source, constituting a type I error. The effects of evaporation on sample comparison are not addressed in this paper. The results from this study demonstrate a statistically acceptable method of physical evidence comparison in forensic casework.

Covariance mapping in the analysis of ignitable liquids by gas chromatography/mass spectrometry.

The covariance matrix computed from the retention time-ion abundance data matrix from gas chromatography/mass spectrometry analysis of ignitable liquids is shown to be a useful tool for automated identification of ignitable liquids in a database. The absolute value of the element-by-element difference between two normalized covariance matrices is shown to quantitatively differentiate between ignitable liquids composed of complex mixtures of hydrocarbons and is amenable to automated searching of ignitable liquid databases. The covariance mapping method is applied to a matrix-contaminated postburn sample, allowing the determination of a high degree of similarity between the ignitable liquid and a heavily evaporated gasoline.

Equilibrium partitioning model applied to RDX-halide adduct formation in electrospray ionization mass spectrometry.

An equilibrium partitioning model is applied for the first time to the sequential formation of 1:1 and then 2:1 adducts between the high explosive cyclo-1,3,5-trimethylene-2,4,6-trinitramine (RDX) and halide anions fluoride, chloride, bromide, and iodide in electrospray ionization interface (ESI) mass spectrometry. The equilibrium partitioning model is developed and model calculations are presented to demonstrate the generic behavior of the system, which is in qualitative agreement with the observed changes in 1:1 (RDX-halide) and 2:1 (RDX-halide) responses in ESI-MS. The model is successfully applied to the experimental data with the use of octanol-water partitioning coefficients to predict interior-to-surface partitioning behavior of the complexes in droplets formed in the ESI. The data and model suggest that the significantly more hydrophobic 2:1 complexes are readily observed in ESI-MS, even though their formation constants may be several orders of magnitude less than that of the 1:1 complex. Structures for RDX-halide 1:1 and 2:1 complexes are proposed based on ion-dipole attractions and destabilizing dipole-dipole interactions.

Adsorption saturation and chromatographic distortion effects on passive headspace sampling with activated charcoal in fire debris analysis.

Distortion of the chromatographic profile obtained for hydrocarbons that have been sampled by adsorption onto activated charcoal is a well-known phenomenon. The work reported here helps to better define the causes of chromatographic profile distortion and offers a potential method to avoid chromatographic distortion in some cases through a subsampling technique. The recovery of hydrocarbons from an equimolar mixture was investigated to determine the influence of hydrocarbon concentration on the molar ratios of recovered components. In a one-quart container, hydrocarbon volumes as small as 24 microL (liquid) were sufficient to saturate the surface area available for adsorption on a 99.0 mm2 square of activated charcoal, resulting in significant distortions in the molar ratio and the chromatographic profile of the recovered hydrocarbons. Passive headspace sampling of a similarly small volume of unweathered gasoline spiked onto carpet padding resulted in a significant distortion of the chromatographic profile. The chromatographic profile of the recovered hydrocarbons closely resembled 75% weathered gasoline. Heating the container spiked with unweathered gasoline to evenly distribute the components and then removing a subsample of the carpet padding to a second container for passive headspace analysis greatly reduced the amount of distortion in the resulting chromatogram.