CHARACTERIZATION OF KEY PARAMETERS FOR BIOTECHNOLOGICAL LIGNOCELLULOSE CONVERSION ASSESSED BY FT-NIR SPECTROSCOPY. PART II: QUANTITATIVE ANALYSIS BY PARTIAL LEAST SQUARES REGRESSION

• Main Authors: Chularat Krongtaew, Kurt Messner, Thomas Ters, Karin Fackler
• Format: Article
• Language: English
• Published: North Carolina State University 2010-08-01
• Series: BioResources
• Subjects: Wheat and oat straw pre-treatment; Fourier transform near-infrared spectroscopy (FT-NIR); Biogas potential test; Anaerobic fermentation; Multivariate data analysis; Partial least-squares (PLS) regression
• Online Access: http://ojs.cnr.ncsu.edu/index.php/BioRes/article/viewFile/BioRes_05_4_2081_Krongtaew_MRF_Pt2_Char_Param_Ligno_Conver_Quantitative_NIR_LMS/700

Wheat straw (Triticum aestivum L.) and oat straw (Avena sativa L.) were chemically pretreated at different severities with the purpose of delignification, which in turn leads to a better accessibility of plant cell wall polysaccharides for further biotechnological conversion. Key parameters of these samples, i.e. weight loss, residual lignin content, and hydrolysable sugars serving as precursors for biofuel production were monitored by wet-chemistry analyses. Fourier transform near infrared (FT-NIR) spectra were correlated to these data by means of partial least-squares (PLS) regression. Weight loss (4.0 – 33.5%) of the wheat straw could be predicted (RMSEP = 3.5%, R²test = 0.75) from the entire FT-NIR spectra (10000 – 4000 cm-1). Residual lignin content (7.9 – 20.7%, RMSEP = 0.9%, R²test = 0.94) and amount of reducing sugars based on pretreated wheat straw (128 – 1000 mg g-1, RMSEP = 83 mg g-1, R²test = 0.89) were powerfully evaluated between 6900 and 5510 cm-1, a spectral region where polysaccharides and lignin absorb. All these parameters could be equally predicted with even higher accuracy from pre-treated oat straw samples. Furthermore, some important parameters for anaerobic conversion of wheat straw to biogas – biogas production, total solids, and volatile solids content – could be estimated.