Quantitative characterisation of historic paper using THz spectroscopy and multivariate data analysis

THz spectroscopy in the time domain was explored in combination with multivariate data analysis, for quantitative determination of chemical and mechanical properties of historic paper, such as lignin content, tensile strength, and ash content. Using partial least squares (PLS) regression, it was shown that quantitative prediction of the material properties is possible, which indicates the potential of THz spectroscopy for chemical characterisation of complex organic materials of natural origin. In addition, the results demonstrate that THz spectra and PLS loading weights for lignin content differ significantly, which leads to the conclusion that THz spectra of composite macromolecular materials do not represent sums of spectra of the individual components. This supports the premise that THz spectra reflect intermolecular interactions. The study was carried out using 250 historical paper samples from the sixteenth century to present. Although the measurements were performed in vacuum to improve the quality of spectra, THz spectroscopy is in principle non-destructive. This research therefore reinforces the role of THz spectroscopy in characterisation of valuable historic materials, where invasive analysis is often not possible.