Assessing Biochar Stability Indices Using near Infrared Spectroscopy

Abstract

New and rapid techniques for estimating the stable fraction of organic carbon (Corg) in biochar are needed for carbon (C) sequestration accounting. In this study, 25 biochar samples produced from different feedstocks and pyrolysis temperatures were scanned using visible near infrared (NIR) spectroscopy and analysed using standard laboratory methods for reference data. Principal component analysis and linear discriminant analysis of preprocessed spectral data were used to extract relevant information and discriminate among biochars, respectively. Partial least squares regression was used to build calibration models between preprocessed spectral data and reference data, and the accuracy of calibration models in predicting biochar properties was then tested using a cross-validation procedure. Biochar indices related to C stability, such as aromatic C, fixed C, atomic H/Corg ratio and the fraction of total C that is aromatic(fa) were successfully predicted ( R2CV 0.92–0.94, RPDCV 3.26–4.22) using the NIR spectroscopy technique (NIR bands, 780–2500 nm). Aromatic C and fixed C could be predicted independently from fa. Other biochar properties, such as Corg, H and O content, and atomic O/Corg ratio, were predicted with an accuracy ranging from moderate to very high ( R2CV 0.80–0.96, RPDCV 2.24–4.66). The study illustrates the potential of this rapid and low-cost technique for measuring biochar stability indices in routine analysis if accurate calibration models for each index are available.