The information inherent in cigarette ash in the form of trace-metal concentrations may be of use in a forensic context as it can indicate the brand from which the ash originated. This knowledge might help place suspects at crime scenes or determine how many people may have been present. To develop and test statistical models capable of classifying ash samples according to brand, commercial cigarettes procured in the U.S. and overseas were "smoked" using a peristaltic pump, mimicking the range of human smoking habits. Ash samples were digested in a mixture of nitric and hydrochloric acid applying microwave digestion and analyzed using inductively coupled plasma-mass spectrometry. Principal component analysis of the elemental data showed intrinsic differences between brands. Partial least squares-discriminant analysis demonstrated that brand classification yields good sensitivity and specificity for a number of models tested. Varying smoking parameters did not impact the classification of ash samples.
Keywords: cigarette ash; forensic science; inductively coupled plasma-mass spectrometry; partial least squares-discriminant analysis; principal component analysis; trace evidence.
© 2016 American Academy of Forensic Sciences.