Thesis

Multivariate profiling of gel inks : a novel tool for the discrimination of within and between brand variation

Downloadable Content

Download PDF
Creator
Awarding institution
  • University of Strathclyde
Date of award
  • 2013
Thesis identifier
  • T13745
Qualification Level
Qualification Name
Department, School or Faculty
Abstract
  • The gel ink pen is the fastest growing pen class available on the modern market. Consequently, its prevalence in forensic casework is expected to increase. This poses a challenge to forensic scientists, since the chemistry of gel inks differ to other commonly encountered inks, thus discrimination by traditional methods like Thin Layer Chromatography (TLC) is limited and a new analytical methodology is required for distinguishing different formulations effectively. A study to evaluate discriminating potential of modern spectroscopic techniques alongside traditional methods of ink examination for within and between brand variations of two commonly encountered colours was undertaken. Within brand variation was detected between multiple samples of a limited number of inks using filtered light, FTIR-ATR and specifically Raman Spectroscopy. For 31 blue gel inks, a discriminating power of at least 0.93 was achieved using a combined analytical sequence of TLC, filtered light (IR absorption and IR fluorescence), Vis-MSP (transmittance), FTIR-ATR and Raman Spectroscopy (514.5 nm, 685 nm, 785 nm and 830 nm combined). A discriminating power of 0.98 was achieved for 25 red gel inks using a combination of TLC, filtered light (IR fluorescence only), FTIR-ATR and Raman Spectroscopy (785 nm only). Hybrid gel inks were readily identified using solubility testing, TLC and Raman Spectroscopy combined, with 92% of red inks tested found to contain both dyes and pigments. A library of Raman spectra from 200 known pigments was compiled and CI Pigment Blue 15:1 and 15:3, and Pigment Red 112 and 254 identified as likely colour components of several blue and red gel inks respectively. An objective multivariate statistical methodology incorporating peak selection, data binning and normalisation was developed and successfully applied to both IR and Raman Spectroscopic data. Hierarchical Cluster Analysis (HCA) and Self-Organising Feature Maps (SOFM) were more effective than Principal Component Analysis (PCA) for successful clustering.
Resource Type
DOI
Date Created
  • 2013
Former identifier
  • 1032699

Relations

Items

Thumbnail Title Date Uploaded Visibility Actions
file details PDF of T13745 2021-07-02 Public Download