A novel, non-invasive, multi-purpose and comprehensive method to date inks in real handwritten documents based on the monitoring of the dye ageing processes

Abstract

Blue ink strokes belonging to 11 types of writing tools from 7 different brands were naturally aged in the darkness for 2 years and characterised by Vis-microspectrophotometry (Vis-MSP). Ink clustering and classification was performed by applying principal component analysis (PCA) and hierarchical cluster analysis (HCA) in order to subsequently assign an optimal orthogonal partial least squares (OPLS) model capable to predict the age of different inks. This method was found applicable to the exact dating of forged documents with ink strokes of up to 2 years old. The method was tested with blind samples supplied by the Forensic Science Unit of the Basque Country Police Department. Age predictions (p ​< ​0.05) with an accuracy error below 25% were obtained whenever: (i) the two ink replicas were placed inside the ellipse of the predicted score graph, (ii) the age of the ink was within the temporary application range of the OPLS model, and (iii) the ink was found to fit correctly in any of the classifications of the pen brands studied. The OPLS model was also able to detect those inks out of its temporary application range, if the ink samples were correctly clustered in the PCA model and/or classified in the HCA model with one of the pen brands studied, but one of the two ink replicas was placed outside the ellipse of the predicted score graph. Thus, these inks were temporarily above or below the application limit of the OPLS model of the corresponding pen brand.