Semester

Summer

Date of Graduation

2022

Document Type

Thesis

Degree Type

MS

College

Eberly College of Arts and Sciences

Department

Forensic and Investigative Science

Committee Chair

Tatiana Trejos

Committee Co-Chair

Ruthmara Corzo

Committee Member

Ruthmara Corzo

Committee Member

Glen P Jackson

Abstract

Broken glass is a trace material frequently found at crime scenes such as hit-and-runs, burglaries, assaults, and homicides. Existing research encompassing the forensic analysis of glass evidence is vast. Published studies cover the analysis and interpretation of various types of glass. However, organizations such as the National Institute of Justice (NIJ), National Institute of Standards and Technology Organization of Scientific Area Committees (NIST-OSAC), and American Society of Trace Evidence Examiners (ASTEE) continue to identify glass-specific research needs to strengthen the scientific foundations of the field. Current gaps within the forensic glass community involve understanding modern glass formulations, re-evaluating the performance of technologically advanced instrumentation, assessing the effect of small glass fragments applied to standard practices, and developing new methods to advance data analysis and interpretation. This thesis aims to address these knowledge gaps by identifying, understanding, and proposing ways to minimize sources of variability within the elemental analysis of glass using laser-induced breakdown spectroscopy (LIBS) and micro X-ray fluorescence spectrometry (m-XRF).

In the first objective of this thesis, LIBS and m-XRF were used to analyze small and irregular glass fragments between 0.4 mm and 1.0 mm in length from same-source and different-source sets. The analysis of the smaller glass resulted in poorer precision than full-thickness glass using both techniques. Using an increased number of known glass fragments resulted in better sample characterization and reduced the false exclusion rates for both small and full-thickness glass fragments analyzed with both instrumental techniques. When analyzing small and irregular samples, new protocols are recommended for sampling, analysis, and interpretation compared to current practice.

The second objective of this thesis examined the elemental composition of modern glass from 30 portable electronic devices (PED), 15 tempered glass screen protectors (SP), and three liquid glass (LG) formulations using m-XRF and LIBS. Using spectral overlay of the m-XRF spectra resulted in five major PED groups and four SP groups based on their elemental composition. LIBS analysis corroborated the PED classes but severe cracking occurred during LIBS analysis. m-XRF comparisons of glass within PED and SP subgroupings achieved high discrimination powers and low false exclusion rates. The application of a newly proposed method based on spectral contrast angle ratios to m-XRF spectra of PED and SP glass provided an additional metric that complemented the spectral overlay results. The application of liquid glass to the surface of a PED screen did not significantly affect the m-XRF analysis of the PED glass. These findings provide the community with a preliminary assessment of the elemental composition of PED glass, SP glass, and LG and the forensic capabilities of m-XRF and LIBS applied to this glass type.

The third objective of this thesis explores the multivariate quantitative analysis of glass using LIBS. Eight commonly used glass standards and characterized reference materials were used to construct calibration models to predict the concentrations of eight major and minor elements within soda-lime glass. Multivariate and univariate methods were compared using the coefficient of determination (R2) and bias to determine the best performing method. Multivariate methods outperformed univariate linear regression. An MLR calibration model using the entire LIBS spectrum provided bias values less than 10-20% while maintaining R2 values greater than 0.9 for all quantified elements. Inputting selected regions of the LIBS spectrum improved the performance of the MLR model. This thesis provides the recommended standards to quantify each of the eight target elements, demonstrating the feasibility of quantitative examinations of glass by LIBS.

The culmination of this thesis addresses several current research gaps in the forensic glass community. The proposed methods involving the analysis of small glass fragments simulating casework-size items, the forensic comparisons of modern PED-related glass, and the quantification of elements within glass standards are anticipated to offer newer knowledge on sources of variability within forensic glass comparisons and approaches to minimize error rates. This research provides glass examiners with additional support when testifying about glass evidence within a court of law.

Share

COinS