This article reports on a research project in which a unified kinetic and thermodynamic model was derived to predict evaporation for forensic applications, such as fire debris analysis.
In this model, a reversible first-order reaction served as the foundation, with rate constants (kinetic regime) and standard vapor pressures (thermodynamic regime) as input parameters. The rate constants and standard vapor pressures for normal (n-) alkanes were fit by linear regression to the retention index, with correlation coefficients of 0.9969 and 0.9998, respectively. These regression equations were used to calculate fraction-remaining curves as a function of the retention index. From these curves, the kinetic and thermodynamic models were able to predict the total fraction remaining of the fuel, either as a bulk quantity or as a chromatogram, as well as the fractions remaining of individual compounds. To evaluate the kinetic and thermodynamic models, gasoline samples were experimentally evaporated to nominal fractions remaining of 0.7, 0.5, 0.3, and 0.1 (30 percent, 50 percent, 70 percent, and 90 percent evaporated). The experimental chromatograms were compared to predicted chromatograms, with Pearson product-moment correlation coefficients of 0.9913 – 0.9068 for the kinetic model and 0.9903 – 0.8921 for the thermodynamic model. The experimental and predicted fractions remaining for individual compounds were compared for n-alkanes and alkylbenzenes spanning a wide range of retention indices and abundances. For the n-alkanes, the mean average percent error was 2.8 – 13.9 percent for the kinetic model and 3.2 – 20.2 percent for the thermodynamic model. This approach provides a unified basis for the comparison of the models and demonstrates the accurate performance of each model. (publisher abstract modified)
- Vibrational Spectroscopy: Recent Developments to Revolutionize Forensic Science
- Expert Algorithm for Substance Identification (EASI) using Mass Spectrometry: Application to the Identification of Cocaine on Different Instruments using Binary Classification Models
- Comparison of Novel Raman Techniques for Fiber Forensics