Date of Award

Summer 8-2026

Document Type

Thesis

Degree Name

Master of Science (MS)

Department/Program

Forensic Science

Language

English

First Advisor or Mentor

Jack Hietpas

Second Reader

Nicholas D. K. Petraco

Third Advisor

Libby Stern

Abstract

Soil is a common form of trace evidence that can associate people, objects, and locations. While forensic soil comparisons are commonly based on particle characteristics, elemental profiling may provide additional information. XRF offers a rapid, nondestructive method for bulk qualitative and quantitative elemental analysis and may complement other methods such as color, PLM, SEM-EDS, and XRD. Traditional bulk XRF analysis often homogenizes samples, potentially losing valuable information about mineralogy, anthropogenic components, particle-size distribution, morphology, and other properties. This study analyzed the fine silt fraction of previously collected soil samples from three physiographic regions in North Carolina, USA. This fraction was targeted because it can be difficult to characterize using PLM. The goals of this study were to investigate the discriminatory potential of the fine silt elemental profiles and to explore a method accessible to practitioners without extensive geological expertise, while preserving the ability to analyze samples using traditional methods. Soil samples were processed using a modified fractionation procedure outlined by Palenik (2007), and XRF spectra were compared using correlation scores and ROC analysis. The results showed strong reproducibility and high within-site similarity; however, overlap between same-site and different-site samples limited overall discriminatory capability. When combined with previously collected color and mineralogy data, XRF provided additional discriminatory information for some sample pairs previously classified as indistinguishable. These findings suggest that XRF analysis of the fine silt fraction can provide useful complementary information in forensic soil examinations while preserving material for subsequent analyses.

Available for download on Wednesday, July 19, 2028

Share

COinS