Reflection-absorption-based near infrared spectroscopy for predicting water content in bio-oil

A novel fiber-optic NIR system based on reflection-absorption phenomenon for predicting water content in bio-oil was developed. System successfully overcomes from the sampling limitations posed by dark and sticky nature of bio-oil. Spectral information in near infrared region (900-2100 nm) was obtained for bio-oils with varied levels of water content. The feasibility of designing an NIR system for estimating the water content in bio-oil was tested by applying multivariate analysis to the spectral data. Two important techniques of multivariate data analysis, principal component regression (PCR) and partial least squares (PLS) were independently applied to experimental results to construct theoretical models. The calibration results demonstrated that spectral information can successfully predict the bio-oil water content (from 16% to 36%) with less than 2% of error and with a high coefficient of determination (R-2 > 0.85). The reliability of a fiber-optic NIR system for predicting the bio-oil water content indicates that there is potential for utilization of NIR spectroscopy in monitoring bio-oil quality during production. (C) 2008 Elsevier B.V. All rights reserved.