Analysis and discrimination of architectural paint samples via a population study

Abstract This study involved the comparative analysis and discrimination of 964 architectural paint samples collected in the United States and Canada. The samples were evaluated to determine the extent to which randomly collected architectural finishes can be discriminated following standard operating protocols for paint analysis. The study also provides a basis for assessing the significance of a result in which a pair or group are undifferentiated. The techniques utilized were stereomicroscopy, Fourier transform infrared spectroscopy (FTIR), scanning electron microscopy (SEM) with both backscatter electron imaging (BSI) and energy dispersive spectroscopy (EDS), and pyrolysis-gas chromatography/mass spectrometry (Py-GC/MS). To assess the layer characteristics (e.g., number, color, sequence, thickness), stereomicroscopy was used with FTIR and/or SEM as needed. The use of visual and microscopic evaluations in concert with FTIR resulted in 42 undifferentiated pairs, a discrimination of 99.99%. Adding SEM and Py-GC/MS analysis resulted in further discrimination of 31 pairs, resulting in eleven undiscriminated pairs of samples. At the conclusion of the study, samples that remained undifferentiated proved to originate from the same source. Therefore, no random pairs were indistinguishable at the conclusion of all examinations.