Rapid Estimation of the chemical component composition of mangium (Acacia mangium Willd.) with Near Infrared (NIR) Spectroscopy

Abstract

Mangium (Acacia mangium Willd.) is a fast growing species with the characteristics of high soil tolerance and high wood quality. Its characteristics have led mangium to be a preferred species for pulpwood. Information on chemical components of the wood cell wall determines the preference of wood utilization. An accurate and rapid method of chemical components measurement is thus paramount. Accuracy and promptness of near infrared (NIR) spectroscopy method in the measurement of chemical components of various organic materials has been reported. The present study was aimed to develop a fast and accurate measurement method of the chemical components of mangium based on NIR spectroscopy techniques. The study was conducted by measuring the NIR spectra of mangium wood both in the form of solid wood and wood meal. In the present works, wood samples were procured from 5, 6, and 7 years old trees. NIR spectroscopy measurements were carried out on both solid and wood meal samples. Chemical methods of measurements on chemical components of wood were also carried out. These included the measurements of holocellulose, α-cellulose, hemicellulose, lignin, and alcohol-benzene soluble extractives. Modeling analysis for the calibration and validation of the NIR spectra was done by Partial Least Square (PLS). It was found that the content of α-cellulose, hemicellulose, and extractives of the mangium wood was significantly influenced by the age of wood. However, only the lignin content of the 5 years old wood was significantly different from the others. The spectra of NIR absorbance of the solid wood samples were of a wider range compared to these of the wood meal samples. The average of solid wood absorbance was higher than that of the wood meal. The best calibration model was obtained for the estimation of α-cellulose and hemicellulose content both for solid wood and wood meal samples. R2 cal value for α-cellulose of solid wood and wood meal samples were respectively of 0.7773 and 0.7779, and the value of RPD was 1.68 and 2.00, respectively. R2 cal value for hemicelluloses of solid wood and wood meal samples were respectively of 0.7756 and 0.8277, and the value of RPD was 2.08 and 2.21, respectively. Estimation of the α-cellulose and hemicellulose content of mangium using NIR spectroscopy is better to be done with wood meal sample than with solid wood sample. In contrary, for the estimation of lignin and alcohol-benzene soluble extractive content, solid wood samples resulted in a better measurement results.