Qualitative and quantitative analysis of wood samples by Fourier transform infrared spectroscopy and multivariate analysis

Abstract

Fourier transform infrared (FTIR) spectroscopy, in combination with multivariate analysis, enable the analysis of wood samples without time-consuming sample preparation. The aim of our work was to analysis the wood samples qualitatively and quantitatively by FTIR spectroscopy. A Van Soest method to determine the lignin, cellulose and hemicellulose content, was used as reference method. Multivariate calibration was performed based on first derivative of the FTIR spectra in the wave number range from 1900 to 800 cm −1, using principal component analysis (PCA), hierarchical cluster analysis (HCA) and partial least-squares (PLS) chemometric methods. Multivariate calibration models for FTIR spectroscopy have been developed. Using PCA and HCA approach, wood samples were classified as softwoods and hardwoods while wood samples with and without treatments were labeled as wood, neutral detergent solution fiber (NDSF), acid detergent solution fiber (ADSF) and strong acid solution fiber (SASF). Furthermore, PLS regression method was applied to correlate lignin, cellulose and hemicellulose contents in wood samples with the FTIR spectral data. The models’ refinement procedure and validation was performed by cross-validation. Although a wide range of input parameters (i.e., various wood species) was used, highly satisfactory results were obtained with the root-mean-square errors for the contents of lignin, cellulose and hemicellulose are 1.51, 0.96 and 0.62%, respectively. This study showed that FTIR spectra have the potential to be an important source of information for a quick evaluation of the chemical composition of wood samples.