Rapid and Nondestructive Approach for Examination of Coloured Printed Documents Using Micro‐Raman Spectroscopy and Chemometrics

ABSTRACT The widespread use of multifunction printers (MFPs) has led to the evolution of crimes involving printed documents. Considering the escalating need and complexity of investigation pertaining to the authenticity and source of such documents, we present the first ever study that characterizes and discriminates both coloured and black printed documents generated using ink‐based and toner‐based MFPs. The study explores the potential of micro‐Raman spectroscopy to characterize the printed samples based on their pigment composition. The Raman spectra revealed the presence of various pigments such as copper phthalocyanine in the colour cyan, pigment yellow (PY) 16, PY 55, PY 73, 81, 110, PY 111 and PY 155 in the yellow colour printed samples, iron oxide red, lithol rubine and barium lithol red in Magenta samples and carbon‐based pigments, manganese dioxide and nigrosine dye in the black colour printed samples. The study is further supported by two pattern recognition techniques viz. PCA and PLS‐DA. PCA has been used as exploratory technique to study the variance in the dataset, which resulted in a variance of 98% for cyan and magenta, 92% for yellow and 99% for black printed samples. PLS‐DA has been further used to discriminate and classify the samples into the predefined class of samples. A trained PLS‐DA model was thus developed showing the highest discrimination efficiency for the colour Cyan with an accuracy of 96%. To conduct a validation study and assess the efficacy of the model, the trained model was further used to predict four unknown test samples. The model gave 100% classification results with no misclassification.