Date of Submission

4-2025

Document Type

Thesis

Degree Name

Master of Science in Forensic Science

Department

Forensic Science

Advisor

Brooke W. Kammrath, Ph.D.

Committee Member

Marisia Fikiet, Ph.D.

Committee Member

Pauline Leary, Ph.D.

Committee Member

Peter Harrington, Ph.D.

Keywords

Forensic Soil Analysis, Particle Correlated Raman Spectroscopy (PCRS), Raman Spectroscopy, Particle Size and Shape Distributions, Morphological Characteristics, Soil Classification

MeSH

Soil, Spectrum Analysis, Raman

LCSH

Soils--Analysis, Raman spectroscopy, Minerals--Classification

Abstract

Forensic soil analysis has had a rich history of providing valuable information for investigating criminal events. However, many modern forensic laboratories have stopped performing this analysis due to the perception that it is either too time-consuming or labor­ intensive. Further complicating the issue is the view that many forensic soil analytical methods are subjective due to their reliance on feature comparisons. Particle Correlated Raman Spectroscopy (PCRS) is a new method that was proposed to address both concerns by providing automated and objective analysis and comparison of sample mixtures. PCRS is an integrated technique that combines automated image analysis with Raman spectroscopy. Particle imaging determines particle size and shape distributions for each component in a sample, yielding detailed morphological information (e.g., circularity, area). At the same time, Raman spectroscopy can probe the molecular chemistry of specific particles of interest. Particle size distributions can be generated for the entire sample or for each mineral present, along with quantitative information on the relative amount of each type of particle.

Mineral counts and morphological properties are used as the basis for the classification and comparison of Raman-identified particles. The discrimination potential of PCRS was explored using various statistical methods from data collected from topsoil samples collected in triplicate from 30 different locations in the Northeast United States. Following analysis of the particle sets within each sample, applicable mineral sets were subject to statistical analysis via Analysis of Variance (ANOVA) and Kruskal-Wallis. While the process was plagued with technological obstacles during data collection, physical and chemical data were still collected for all samples and statistical analysis showed promising results. Ultimately, this research provides statistical evidence for the discriminatory power of minerals and their morphologies for the classification and source identification of soil samples.

Download

Share

COinS