Prediction of Hardwood and Softwood Contents in Blends of Wood Powders Using Mid-Infrared Spectroscopy

Biofuel quality control plays an important role considering the recent European policy about renewable energy source promotion. Origin and source of the raw material are often required to be stated by biofuel chain operators for traceability and sustainability issues. Being fast, non-destructive, and low-cost, infrared spectroscopy coupled with chemometrics is already applied to several sectors and could also be employed in the solid biomass sector. The result of this work is a tool for the prediction of hardwood and softwood contents in blend samples by means of Fourier transform infrared spectroscopy coupled with partial least squares regression. A total of 61 samples of fir, pine, sessile oak, and beech and four series of binary blends (28 samples) from wood powders of one hardwood species and one softwood species were analyzed. The infrared prediction model was full cross-validated. The results of this work showed the good performance of the model with a standard error of a few percentage points (3.8%). As a consequence, the development of an analytical instrument based on such techniques could be useful to support the bioenergy chain stakeholders, such as solid biofuel producers, traders, and customers, for traceability, process tuning, and quality control issues.