Title
Rapid determination method of methoxyl content in lignin by chemometric analysis of FT-IR spectroscopic data
Authors
MOHAMMAD NASHIR UDDIN, M. NUR ALAM LIKHON, M. MOSTAFIZUR RAHMAN, M. TUSHAR UDDIN, M. KAMAL HOSSAIN and M. SARWAR JAHAN

Received August 12, 2025
Published Volume 60 Issue 1-2 January-February
Keywords lignin, methoxyl group, FT-IR spectral data, partial least squares regression

Abstract
Lignin is the second largest biopolymer on earth, and an important feedstock of biomaterials and biochemicals. The methoxyl group (-OCH₃) content in lignin controls its properties. To elucidate lignin’s structure, quantification of –OCH3 a critical step in lignin characterization. With FT-IR spectral data of 30 lignin samples, chemometric models – Support Vector Regression (SVR) and Partial Least Squares Regression (PLSR) – have been established. These SVR and PLSR were calibrated with the full FT-IR data and its different segments, and the range 3600-3100 cm-1 was found the most informative. Both models have been developed based on raw and pretreated spectroscopic data with Moving Average (MA), Mean Normalization (MN), Standard Normal Variate (SNV) and Multiplicative Scatter Correction (MSC). The predictive performance of the PLSR model is better for predicting methoxyl contents (R2=89.3%), but the model is unstable in the validation stage (R2=26.3%) with the data pretreated by MSC. Through this exercise, the best model for predicting methoxyl group in lignin is PLSR, when it is calibrated with the raw spectral data of FT-IR spectroscopic data for the range 3600-3100 cm-1 (R2=82.5% for calibration and 53.5% for validation dataset). Finally, the predicted values of the methoxyl content by the developed model are very close to the values calculated by the tedious and time-consuming wet method.