Raman Spectroscopic Characterisation and Chemometric Analysis of Facial Cosmetics as Associative Trace Evidence

Decorative cosmetics such as facial foundations or finishing powders are widely used and easily transferred upon physical contact. As such, they may be used as associative trace evidence to link people to each other or to places in criminal investigations. To maximise their probative value, it is important to understand the variability amongst representative market products and the degree of sample discrimination that can be achieved. Additionally, it is

required that analysis techniques be non-destructive, readily available and relatively inexpensive. Raman spectroscopy is a powerful tool for probing the chemistry of facial cosmetics. As well as fitting the criteria above, it offers the capability of studying a wide range of sample types with minimal prior preparation. The information derived from Raman spectra can help analysts to understand and visualise spectral variability, potentially enabling discrimination between samples.

This thesis presents the novel application of Raman microspectroscopy to the analysis of 297 facial cosmetic samples with subsequent chemometric methods for objective spectral interpretation. The analysis of 177 newly purchased cosmetic products revealed the most important chemical components for sample discrimination and highlighted the issue of spectral heterogeneity, leading to the separation of 126 spectrally homogeneous from 44 spectrally heterogeneous samples. Microscopic examination of samples allowed for the assessment of visual homogeneity but revealed that this was not a reliable indicator of spectral homogeneity. Subjective assessment of Raman spectra did not always correlate with the principal component analysis (PCA) models as the dimensionality reduction technique applied different variable weightings. Assessment of PCA loadings showed the primary distinction between samples to be their titanium dioxide polymorph, followed by iron oxide and lecithin content.

The effects of ageing were more noticeable visually among the water-based products yet nearly undetectable via Raman spectra. Drastic colour and texture changes were evident after 15 months of passive ageing, as were the detachment of borosilicate glass pigments and silica microspheres from their respective matrices, which may have implications for casework. The spectral changes were more pronounced amongst the aged samples that contained a higher organic composition, indicated by the loss of organic components or their degradation to other

species. Furthermore, inter-batch comparisons of some products showed a change of titanium dioxide polymorph used (from anatase to rutile) showing a lack of formulation consistency.

Donations of 120 used samples, old discarded products and expired shop testers allowed for the assessment of a set of "real world" samples. Raman analysis of these donated samples and comparisons with their newly purchased counterparts revealed an increase in baseline fluorescence, increased spectral heterogeneity, and extra peaks in the spectral mid-range. These spectral alterations suggest chemical changes associated with ageing, and/or the contamination of these samples, most likely from a biological source.

This study describes the first microscopic and Raman spectroscopic characterisation of facial cosmetics within a forensic context, further enhanced by the addition of multivariate data analysis methods, which addresses the need for objective, unbiased interpretation of spectral data. These cosmetic traces may be exploited in criminal investigations in questioned (Q) versus known (K) comparisons, or sample eliminations, or to provide investigative leads. The discovery of a counterfeit product demonstrated the utility of these analyses not just in a crime scene investigation context but also for the verification of genuine consumer goods. The assessment of cosmetic traces on a purely visual basis is not recommended, owing to subjective descriptions and the difficulties associated with analysing interference pigments; the addition of Raman spectroscopy offers much needed enhanced discrimination potential. Raman microscopy is well suited for the analysis of microtraces such as those that might be encountered in casework, and involves less sample preparation than many other analytical techniques such as surface enhanced Raman spectroscopy (SERS).