Validation and expansion of sex determination method through analysis of human hair using electrothermal vaporization coupled to inductively coupled plasma optical emission spectrometry

Abstract

Within forensic science, the analysis of trace evidence is paramount to criminal investigations. Hair is a stable tissue that contains a record of elemental exposure that can be valuable to investigators. A previous study established a method for discriminating biological male and female hair samples through multi-elemental analysis via electrothermal vaporization inductively coupled plasma optical emission spectrometry (ETV-ICPOES), in combination with multivariate statistics; Mg, S, Sr, and Zn were used as predictors of sex, and both principal component analysis and linear discriminant analysis (LDA) were used for dimensionality reduction. This preliminary proof-of-concept work sought to test two common variables in real life samples – dyed hair, and samples originating from closely related family members – against the preliminary model. Hair samples were washed in portions of hexanes and doubly deionized water, following which they were dried, and ground into a fine powder prior to analysis via ETV-ICPOES. Point-by-point internal standardization with an argon emission line compensated for sample loading effects on the plasma. Peak areas were mass-corrected prior to multivariate analysis; no quantification is performed, eliminating the need for a matrix-matched solid certified reference material. The method was revalidated using CF₄ as an ETV reaction gas due to restricted usage of CCl₂F₂; despite using a different chemical modifier, LDA remained effective at accurately predicting the sex of male and female samples, indicating method robustness. Hair dye caused a significant increase in the levels of the predictor elements in both male and female samples; this resulted in several mispredictions. The selection of new predictor elements that are not impacted by hair dye was necessary for dyed samples. The use of Cd, Ce, Fe, and Sn as predictor elements with a model based on undyed hair samples allowed for accurate classification of hair samples, whether dyed or not. The inclusion of closely related family members did not appear to influence the predictive abilities of the LDA model, however the inclusion of individuals sharing living and dietary habits resulted in mispredictions, highlighting the impact of diet and lifestyle on the elemental composition of hair.