Characterization of paper finishes by use of infrared spectroscopy in combination with canonical variate analysis

T. Canals, J. R. Riba, R. Cantero, J. Cansino, D. Domingo, H. Iturriaga

Research output: Contribution to journalArticleResearchpeer-review

17 Citations (Scopus)


The finishing process used by the paper industry involves subjecting the paper surface to the action of chemicals and physical treatments in a series of operations intended to provide an end-product suitable for its intended use. In this work, we studied various paper finishes by using infrared spectra processed with appropriate chemometric techniques. To this end, we used a wide range of paper samples supplied in various finishes (coated, offset and cast-coated) by several paper manufacturers. Fourier transform middle-infrared (FTIR) spectra for the paper samples were recorded by using an ATR module, and reflectance near-infrared (NIR) spectra with the aid of a fibre-optic probe. Both techniques are fast and require no sample pretreatment. The primary aim of this work was to develop a new methodology affording the accurate classification and identification of paper finishes in samples other than those used to construct the calibration model. To this end, we used the discriminant chemometric techniques principal component analysis (PCA) and canonical variate analysis (CVA), application of which was followed by that of the k-nearest neighbour algorithm to the samples in the prediction set. This procedure was also used to classify the coated samples into three subgroups. Both FTIR and NIR spectroscopy allowed most of the samples in the prediction sets to be accurately classified and identified. © 2008 Elsevier B.V. All rights reserved.
Original languageEnglish
Pages (from-to)751-757
Issue number2
Publication statusPublished - 15 Dec 2008


  • CVA
  • FTIR
  • NIR
  • Paper coating
  • Paper finish
  • PCA


Dive into the research topics of 'Characterization of paper finishes by use of infrared spectroscopy in combination with canonical variate analysis'. Together they form a unique fingerprint.

Cite this