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

Artificial intelligence and thermodynamics help solving arson cases

In arson cases, evidence such as DNA or fingerprints is often destroyed. One of the most important evidence modalities left is relating fire accelerants to a suspect. When gasoline is used as accelerant, the aim is to find a strong indication that a gasoline sample from a fire scene is related to a sample of a suspect. Gasoline samples from a fire scene are weathered, which prohibits a straightforward comparison. We combine machine learning, thermodynamic modeling, and quantum mechanics to predict the composition of unweathered gasoline samples starting from weathered ones. Our approach predicts the initial (unweathered) composition of the sixty main components in a weathered gasoline sample, with error bars of ca. 4% when weathered up to 80% w/w. This shows that machine learning is a valuable tool for predicting the initial composition of a weathered gasoline, and thereby relating samples to suspects.

Femtosecond covariance spectroscopy

The success of nonlinear optics relies largely on pulse-to-pulse consistency. In contrast, covariance-based techniques used in photoionization electron spectroscopy and mass spectrometry have shown that a wealth of information can be extracted from noise that is lost when averaging multiple measurements. Here, we apply covariance-based detection to nonlinear optical spectroscopy, and show that noise in a femtosecond laser is not necessarily a liability to be mitigated, but can act as a unique and powerful asset. As a proof of principle we apply this approach to the process of stimulated Raman scattering in α-quartz. Our results demonstrate how nonlinear processes in the sample can encode correlations between the spectral components of ultrashort pulses with uncorrelated stochastic fluctuations. This in turn provides richer information compared with the standard nonlinear optics techniques that are based on averages over many repetitions with well-behaved laser pulses. These proof-of-principle results suggest that covariance-based nonlinear spectroscopy will improve the applicability of fs nonlinear spectroscopy in wavelength ranges where stable, transform-limited pulses are not available, such as X-ray free-electron lasers which naturally have spectrally noisy pulses ideally suited for this approach.

Application of Self-Organizing Maps to the Analysis of Ignitable Liquid and Substrate Pyrolysis Samples

Classification of un-weathered ignitable liquids is a problem that is currently addressed by visual pattern recognition under the guidelines of Standard Test Method for Ignitable Liquid Residues in Extracts from Fire Debris Samples by Gas Chromatography-Mass Spectrometry, ASTM E1618-14. This standard method does not separately address the identification of substrate pyrolysis patterns. This report details the use of a Kohonen self-organizing map coupled with extracted ion spectra to organize ignitable liquids and substrate pyrolysis samples on a two-dimensional map with groupings that correspond to the ASTM-classifications and separate the substrate pyrolysis samples from the ignitable liquids. The component planes give important information regarding the ions from the extracted ion spectra that contribute to the different classes. Some additional insight is gained into grouping of substrate pyrolysis samples based on the nature of the unburned material as a wood or non-wood material. Further subclassification was not apparent from the self-organizing maps (SOM) results.

Characterization and Differentiation of Petroleum-Derived Products by E-Nose Fingerprints

Characterization of petroleum-derived products is an area of continuing importance in environmental science, mainly related to fuel spills. In this study, a non-separative analytical method based on E-Nose (Electronic Nose) is presented as a rapid alternative for the characterization of several different petroleum-derived products including gasoline, diesel, aromatic solvents, and ethanol samples, which were poured onto different surfaces (wood, cork, and cotton). The working conditions about the headspace generation were 145 °C and 10 min. Mass spectroscopic data (45–200 m/z) combined with chemometric tools such as hierarchical cluster analysis (HCA), later principal component analysis (PCA), and finally linear discriminant analysis (LDA) allowed for a full discrimination of the samples. A characteristic fingerprint for each product can be used for discrimination or identification. The E-Nose can be considered as a green technique, and it is rapid and easy to use in routine analysis, thus providing a good alternative to currently used methods.

Identification and Discrimination of Brands of Fuels by Gas Chromatography and Neural Networks Algorithm in Forensic Research

The detection of adulteration of fuels and its use in criminal scenes like arson has a high interest in forensic investigations. In this work, a method based on gas chromatography (GC) and neural networks (NN) has been developed and applied to the identification and discrimination of brands of fuels such as gasoline and diesel without the necessity to determine the composition of the samples. The study included five main brands of fuels from Spain, collected from fifteen different local petrol stations. The methodology allowed the identification of the gasoline and diesel brands with a high accuracy close to 100%, without any false positives or false negatives. A success rate of three blind samples was obtained as 73.3%, 80%, and 100%, respectively. The results obtained demonstrate the potential of this methodology to help in resolving criminal situations.

Determination of Ignitable Liquids in Fire Debris: Direct Analysis by Electronic Nose

Arsonists usually use an accelerant in order to start or accelerate a fire. The most widely used analytical method to determine the presence of such accelerants consists of a pre-concentration step of the ignitable liquid residues followed by chromatographic analysis. A rapid analytical method based on headspace-mass spectrometry electronic nose (E-Nose) has been developed for the analysis of Ignitable Liquid Residues (ILRs). The working conditions for the E-Nose analytical procedure were optimized by studying different fire debris samples. The optimized experimental variables were related to headspace generation, specifically, incubation temperature and incubation time. The optimal conditions were 115 °C and 10 min for these two parameters. Chemometric tools such as hierarchical cluster analysis (HCA) and linear discriminant analysis (LDA) were applied to the MS data (45–200 m/z) to establish the most suitable spectroscopic signals for the discrimination of several ignitable liquids. The optimized method was applied to a set of fire debris samples. In order to simulate post-burn samples several ignitable liquids (gasoline, diesel, citronella, kerosene, paraffin) were used to ignite different substrates (wood, cotton, cork, paper and paperboard). A full discrimination was obtained on using discriminant analysis. This method reported here can be considered as a green technique for fire debris analyses.

Forensic Mass Spectrometry

Developments in forensic mass spectrometry tend to follow, rather than lead, the developments in other disciplines. Examples of techniques having forensic potential born independently of forensic applications include ambient ionization, imaging mass spectrometry, isotope ratio mass spectrometry, portable mass spectrometers, and hyphenated chromatography-mass spectrometry instruments, to name a few. Forensic science has the potential to benefit enormously from developments that are funded by other means, if only the infrastructure and personnel existed to adopt, validate, and implement the new technologies into casework. Perhaps one unique area in which forensic science is at the cutting edge is in the area of chemometrics and the determination of likelihood ratios for the evaluation of the weight of evidence. Such statistical techniques have been developed most extensively for ignitable-liquid residue analyses and isotope ratio analysis. This review attempts to capture the trends, motivating forces, and likely impact of developing areas of forensic mass spectrometry, with the caveat that none of this research is likely to have any real impact in the forensic community unless: (a) The instruments developed are turned into robust black boxes with red and green lights for positives and negatives, respectively, or (b) there are PhD graduates in the workforce who can help adopt these sophisticated techniques.

Application of an HS-MS for the detection of ignitable liquids from fire debris

In arson attacks, accelerants such as ignitable liquids are commonly used to initiate or accelerate a fire. The detection of ignitable liquid residues at fire scenes is therefore a key step in fire investigations. The most widely used analytical technique for the analysis of accelerants is GC-MS. However, pre-concentration of the ignitable liquid residues is required prior to the chromatographic analysis. The standard method, ASTM E1412, involves passive headspace concentration with activated charcoal strips as a method to isolate the ignitable liquid residues from fire debris and these residues are subsequently desorbed from the carbon strip with solvents such as carbon disulfide. In the work described here, an alternative analytical technique based on an HS-MS (headspace mass spectrometry) has been developed for the thermal desorption of the carbon strips and analysis of different ignitable liquid residues in fire debris. The working conditions for the HS-MS analytical procedure were optimized using different types of fire debris (pine wood burned with gasoline and diesel). The optimized variables were desorption temperature and desorption time. The optimal conditions were 145°C and 15 min. The optimized method was applied to a set of fire debris samples. In order to simulate post burn samples several accelerants (gasoline, diesel, citronella, kerosene, paraffin, and alcohol) were used to ignite different substrates (wood, cotton, cork, paper, and paperboard). chemometric methods (cluster analysis and discriminant analysis) were applied to the total ion spectrum obtained from the MS (45-200 m/z) to discriminate between the burned samples according to the accelerant used. The method was validated by analyzing all samples by GC-MS according to the standard methods ASTM E1412 and ASTM E1618. The results obtained on using the method developed in this study were comparable to those obtained with the reference method. However, the newly developed HS-MS method is faster, safer, and more environmental friendly than the standard method.

Quantitative classification of two-dimensional correlation spectra

Two-dimensional (2D) correlation spectroscopy, which takes advantage of the apparent enhancement of spectral resolution, is known to be useful in qualitative discrimination of seemingly similar samples. The possibility of quantitative classification of 2D correlation spectra is even more desirable. Two useful parameters, namely Euclidian distance and correlation coefficient between 2D correlation spectra, are introduced for this purpose. Dry and sweet red wine samples are used to demonstrate the utility of these parameters. The distances between the 2D infrared (IR) spectra of sweet and dry red wines are roughly proportional to the differences of sugar contents in them. The result shows that the two parameters are useful measures for the quantitative evaluation of the similarity among the samples and their corresponding 2D correlation spectra.

Comparison of differential mobility spectrometry and mass spectrometry for gas chromatographic detection of ignitable liquids from fire debris using projected difference resolution

The significance of forensic arson analysis accelerates the applications of new technologies in this area. Based on the previously reported application of differential mobility spectrometry (DMS) as a detection method for gas chromatography (GC) in arson analysis, the performances of DMS and mass spectrometry (MS) were compared using a novel chemometric tool, projected difference resolutions (PDRs). The PDR results show that one-way mass spectra data exhibit higher resolution than DMS data, while total ion chromatograms from GC-DMS show higher resolution than that from GC/MS for differentiating seven kinds of ignitable liquids. Combining the information from both chromatography and spectra, two-way data always have higher resolution than one-way data for these two detection methods, and GC/MS would exhibit better performance than GC-DMS according to the minimum resolution value. To verify the PDR results, a fuzzy rule-building expert system was applied for classifying these seven kinds of ignitable liquids from fire debris based on GC-DMS and GC/MS data, respectively. The prediction accuracies were consistent with PDR results, which proved that PDR is a powerful tool in comparing the performances of different analysis methods for pattern recognition.

Ion formation mechanism in laser desorption ionization of individual nanoparticles

Covariance mapping is used to study ion formation mechanisms in laser desorption ionization of individual 50 or 220 nm diameter particles having compositions similar to ambient aerosol. Single particle mass spectra are found to vary substantially from particle to particle. This variation is systematic--the energetically preferred ions (e.g., lowest ionization energy, highest electron affinity) are positively correlated with each other and negatively correlated with less preferred ions. For the compositions studied, the average positive ion yield is two to five times greater than the negative ion yield, indicating that free electrons are the main negatively charged species. For many particles, typically 20% to 40% of those analyzed, only positive ions are detected. Smaller particles give fewer negative ions, presumably because the plume is less dense and electron capture is less likely. The results suggest that ion formation occurs by a two stage process. In the first stage, photoionization of laser desorbed neutrals gives cations and free electrons. In the second stage, collisions in the plume cause electron capture and competitive charge transfer. When the particle ablates in a manner giving a dense plume with many collisions, the energetically preferred positive and negative ions are dominant. When the particle ablates in a manner giving a less dense plume with fewer collisions, the less preferred ions are able to survive and the energetically preferred ions constitute a lower fraction of the total ion signal. Systematic particle to particle variations of relative signal intensities can complicate ambient particle classification efforts by spreading a single particle composition over several classes.

Multivariate pattern recognition of petroleum-based accelerants by solid-phase microextraction gas chromatography with flame ionization detection

A novel method has been developed for the extraction, analysis and identification of petroleum-based fuels using solid-phase microextraction with analysis by GC-FID. Multivariate data analysis is employed to simplify these data allowing for more accurate classification. Principal component analysis (PCA) and soft independent modeling of class analogy (SIMCA) are explored for their effectiveness in establishing accelerant groupings based on the current and previous ASTM International guidelines. The SIMCA models developed for the previous and current ASTM system were 98.5% and 97.2% accurate in unknown sample class prediction. SPME in conjunction with multivariate data analysis is a new approach in accelerant sampling and classification.

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.