Glass microtraces originated from mobile phones – Interpretation of their elemental composition for forensic purposes

Mobile devices have become an integral part of our lives. Therefore, those devices could be present and be damaged during various incidents and crimes as a consequence. However, the glass microtraces from broken smartphone displays have limited utility in forensic cases so far due to the narrow state of the art in analyzing this type of potential evidence. Therefore, this research aimed to create a model to verify the possibility of classifying glass microfragments into smartphone screens (PED; 23 samples) or other types of glasses (window sheets or vehicle windows (CW; 30 samples) and glass containers (P; 30 samples)). The elemental composition of collected samples was determined by the SEM-EDS technique. The investigated classification problem was solved using two different likelihood ratio (LR) models. In both cases, the collected dataset was divided into training (60 % of samples) and test sets (40 % of samples). The results received for the LR_LDA and LR_rar models indicate that the variables based on the levels of signals for Al and Ca are sufficient for the correct classification of the analyzed dataset into PED or CWP categories. Therefore, the proposed SEM-EDS method for distinguishing PED glass samples with the employment of LR models may have a high potential for implementation in forensic purposes.