Chemical recognition based on high-accuracy matching factors as per time-of-flight-secondary-ion mass spectrometry: Application to trace cosmetic residues in human forensics

Abstract

Recognizing the chemical composition of samples or specimens from time-of-flight secondary-ion mass spectrometry (ToF-SIMS) is particularly challenging. Although ToF-SIMS is a powerful tool for obtaining high-quality mass spectra, and produces a large raw dataset with rich chemical information, analysis and interpretation of the results are rather complex. In particular, chemical identification of multi-component mass spectra is difficult because of the large number of peaks. Applications of chemometric and statistical methods are essential. In this work, we used a probability-based matching factor (MF) in combination with principal component analysis (PCA) to recognize the most probable compound from various forensic specimens. We used PCA to classify cosmetics as a function of category and then highlighted the possible origins of various residues. In combination with PCA, we compared the fragmentation pattern of unknown residues with reference mass spectra in a library by using an MF algorithm, and produced a list of the most probable identities based on the best-matching values. We focused our investigation on popular Korean cosmetics and the most common types of forensic. We established a large database of references from different brands among various categories of cosmetic products. We collected both positive and negative ion mass spectra to resolve the chemical compositions, depending on the most appropriate polarity present in ToF-SIMS datasets. This new method offers a unique approach to identify the cosmetic residues present in forensic specimens, and compliments conventional methods for exploring the origin of residues or validating forensic evidence. The algorithm combined with PCA will benefit diverse chemical characterization fields.(1)