Study on the discrimination of tires using chemical profiles obtained by Py-GC/MS

- Major compounds produced by the pyrolysis of tire treads were identified.
- Repeatability of the methodology over time was established on reference materials.
- Discrimination of tires was studied using multivariate statistics (PCA, QDA).
- Tire chemical profiles were differentiated according to brand and model.

Tire traces are very important for traffic accident reconstruction. Chemical analysis by Py-GC/MS proved to be useful to help determine which tire is the source of a tire trace. This is done through statistical comparisons of the chemical profiles of tire traces and tires suspected to be the source of the traces. By analysing polymers and copolymers known to be used for tread formulation, this study could have identified 21 compounds on the 86 used for comparison purposes. These reference materials were analysed in two series remote in time, to assess the repeatability of the methodology over time. The Relative Standard Deviations (RSDs) obtained for pyrolysates of each reference material were low (max. 6.04% for pre-treated data) supporting that the methodology is repeatable. The second aim of the present research was to investigate the discrimination of a random sample of 60 tires divided into 11 brands and 22 models. The discrimination was evaluated on chemical profiles of tires obtained by Py-GC/MS. Principal Component Analysis and Quadratic Discriminant Analysis were applied to the multivariate data. The observation of the scores of the Principal Component Analysis showed that the 180 replicates are well grouped by model of tires and tendencies per brand were observed. Performances of Quadratic Discriminant Analysis for training and test sets were respectively 97.60 ± 1.28% and 93.00 ± 3.92% for a classification by brand. For a classification by model of tires, the performances were 95.30 ± 1.57% for the training set and 85.10 ± 4.40% for the test set. Obtained performances are judged high and the results support that the chemical profiles of tires vary between brands and between models of tires in such an extent that they can be well discriminated according to these two parameters.