Rapid identification of tissue paper made from blended recycled fibre by Fourier transform near infrared spectroscopy

This paper presents a Fourier transform near infrared spectroscopic method, coupled with principal-component analysis (PCA) and partial least-squares discriminate analysis (PLS-DA) techniques, for discriminating between paper products made of virgin fibre only and those made of virgin fibres blended with recycled fibres. The PLS-DA method was used to construct the discrimination models based on PCA. The study showed that the effects of the number of layers of samples, texture and moisture content can be reduced to acceptable levels by using >72 layers of papers, pressing them against a glass plate and subjecting the spectral data to preprocessing algorithms using standard normal variate analysis, multiplicative scattering correction and first-derivative calculations (FOCI. The PLS-DA model based on an FDC transformation provided the best discrimination between the virgin-fibre samples and the samples blended with recycled fibre. The present method is non-destructive and enables a particularly fast classification response, without sample pretreatment. Above all, it does not consume chemicals and reagents or require a qualified laboratory technician and laboratory-grade facilities. Therefore, it appears to be suitable for use in identifying blended recycled-fibre tissue paper samples both at the manufacturer stage and in point-of-sale samples from commercial markets.