Semester

Summer

Date of Graduation

2023

Document Type

Thesis

Degree Type

MS

College

Eberly College of Arts and Sciences

Department

Forensic and Investigative Science

Committee Chair

Glen Jackson

Committee Member

Jacqueline Speir

Committee Member

Shikha Sharma

Abstract

To determine whether a questioned fire is the result of arson or not, forensic chemists will often attempt to extract and identify any ignitable liquid residues present in the fire debris. Forensic chemists use gas chromatography-mass spectrometry (GC-MS) to identify ignitable liquids based on similarities between chromatograms of fire debris extracts and those of reference ignitable liquids that have been evaporated (weathered) to different extents. Casework samples often appear to be relatively unweathered (e.g., 40-60%) and conventional wisdom is that the apparent lack of weathering is the result of entrapment; a process by which liquid residues are absorbed into the pores of substrates, shielding them from the heat of a fire and limiting evaporation. In contrast to this conventional wisdom, the current work shows that ignitable liquids can be extensively weathered (e.g., 95% by mass) at elevated temperatures (e.g., 210 ℃) in the presence of porous wooden substrates and still provide residues that are almost indistinguishable from the unweathered liquid.

In the present work, all experimental evaporations were performed at 210 ℃ to provide somewhat realistic casework conditions without posing a fire hazard. To simplify quantitation and calculations, a nine-component artificial gasoline mixture was used for all evaporations instead of real gasoline. Evaporations were conducted in the presence of four household substrates—cotton fabric, nylon carpet, pine wood, and plywood—with varying absorption times prior to evaporation: e.g., 0 seconds, 30 seconds, 10 minutes, and 30 minutes. The extent of evaporation was determined by spiking known quantities of liquid on each substrate and measuring the mass of residue remaining in the substrate after a designated evaporation time. Before use, each substrate was baked to a constant mass at 210 ℃ to ensure accurate quantitation of the extent of evaporation. Quality control experiments verified that liquid extracts of the un-spiked, baked-out substrates were devoid of organic residues and that liquid-liquid extractions of unweathered residues from the spiked substrates were quantitatively reliable under the chosen extraction conditions, which was a 30-minute extraction time in pentane. At least 8 different evaporations were carried out to generate ~45-98% weathered samples for each set of substrate and absorption time conditions.

In the absence of a substrate, and in the presence of wicking substrates like cotton fabric and nylon carpet, the thermodynamic model developed previously by Willis et al. provided accurate predictions of the experimentally determined extent of evaporation, with weathering errors as small as 8.2% and a root mean square error of predictions (RMSEP) of chromatographic peak areas of ~4.0%. As reported previously, the predicted weathering errors increased in the presence of substrates that restricted mass transfer, like the wood samples.

A correction factor was developed for the thermodynamic model to account for porous substrates that inhibit the evaporation process. The correction factor is a single term that varies between 0-1 that is designed to model the resistance to mass transfer that the liquid-phase components of the artificial gasoline face in the solid substrates. The resistance phenomena could include the combined effects of adsorption, absorption, and diffusion. A correction term of 1 predicts that the liquid components evaporate in proportion to their liquid phase molar ratios rather than when the correction term is 0, in which case they are modeled to evaporate in proportion to their partial pressures. Additional experiments demonstrated that the initial surface area:volume ratio of ignitable liquids as they evaporate replicated the effect of resistive substrates on the measured peak areas of the recovered liquid residues. Smaller initial surface area:volume ratios acted like the porous substrates and provided a greater proportion of volatiles in the residues, even when extensively weathered (e.g., to 90%). The correction factor was able to substantially decrease the weathering errors and RMSEP (reductions by as much as 40% and 8% respectively for wooden substrates) by accounting for resistance to mass transfer in the substrates. The model and experiments both support the finding that chromatograms of ignitable liquids weathered more than 90% at 210 ℃ on wooden substrates can be almost indistinguishable from the original unweathered liquid. These experiments, and the supporting thermodynamic model, support the premise that gasoline residues may appear to be unweathered in fire debris, even after extensive

evaporation at elevated temperatures, as encountered in structure fires.

Embargo Reason

Publication Pending

Share

COinS