Investigating Gas Chromatography Instrumentation Performance and Developing Data Analysis Strategies for Chemical Analysis Applications

Abstract

Gas chromatography (GC) is a widely used analytical technique for the separation, detection, and identification of volatile and semi-volatile organic compounds. This thesis investigates two complementary analytical applications of GC-based instrumentation. The first chapter evaluates the analytical performance of the Agilent Intuvo 9000 gas chromatograph equipped with a flame ionization detector (GC-FID), while the second chapter explores the feasibility of using activated carbon strips to recover, preserve, and analyze fragrance residues for forensic fingerprinting applications. In Chapter 1, a complex test mixture was analyzed to determine the limit of detection (LOD) of the Intuvo 9000 GC-FID and to investigate compound identification using GC-quadrupole mass spectrometry (GC-qMS). Chromatographic data were preprocessed using baseline correction and analyzed using principal component analysis (PCA) to evaluate analytical reproducibility across multiple concentration levels. The Intuvo 9000 GC demonstrated excellent reproducibility, demonstrating the instrument's ability to detect compounds at low concentrations. Because FID does not provide structural information, GC-qMS was employed for compound identification for the test mixture. A mass spectrum averaging approach combined with intensity thresholding significantly improved NIST library matching performance, increasing representative match values (MV) from the complex mixture. Overall, this work demonstrates the analytical capabilities of the Agilent Intuvo 9000 GC for sensitive chromatographic detection and highlights the advantages of GC-qMS for compound identification. Chapter 2 investigates the use of activated carbon strips for the recovery, storage, and extraction of fragrance residues as a sample preparation method for GC-MS analysis in forensic fragrance fingerprinting applications. Thirteen commercial fragrances were examined as neat samples, initial carbon strip extracts (I-CSE), and stored carbon strip extracts (S-CSE). Total ion current (TIC) chromatogram-based PCA demonstrated good reproducibility for neat samples but failed to reliably classify I-CSE and S-CSE samples because of analyte loss and increased background noise. To overcome these limitations, a mass spectrum-based workflow incorporating TIC signal intensity thresholding, MV filtering, and concatenated spectral features were developed. Applying a neat TIC peak threshold, mass spectral intensity threshold and a MV criteria substantially improved spectral quality and multivariate classification. The resulting PCA provided improved clustering of fragrance samples, while a target-based peak-counting strategy could distinguish fragrances using persistent chemical fingerprint markers retained after 2.5 months of freezer storage. Therefore, the developed activated carbon strip workflow, combined with mass spectrum-based chemometric analysis, provides a practical approach for preserving and comparing fragrance chemical fingerprints. These findings support the potential forensic application of fragrance residue analysis for sample comparison and chemical fingerprinting following storage and environmental exposure.