The Use of Near-Infrared (NIR) Spectroscopy and Principal Component Analysis (PCA) To Discriminate Bark and Wood of the Most Common Species of the Pellet Sector

Abstract

The pellet energy market is expanding rapidly in Europe and also at the global level, in response to the continuously growing energy demand and because of the high degree of reliability, the easy handling, and the cheap and simple logistics, in comparison to other solid biomasses. The fast growth of this market has highlighted the problem of product quality, which has strong repercussions for technical, environmental, and economic aspects. The biomass quality is defined by several chemical–physical parameters that are directly measurable in the laboratory. In addition, there are quality attributes related to origin and source, difficult to investigate through traditional analyses, such as the type of wood (hardwood/softwood) and the presence of bark. The development of a rapid technique able to provide this information could be an advantageous tool for the energy sector proving indications on biofuel traceability and sustainability. More than 90 samples belonging to three of the most common European species for the energy pellet sector (fir, pine, and beech) have been collected and analyzed by means of Fourier transform near-infrared (FT-NIR) coupled with Principal Component Analysis (PCA). The results show that the main variation in the data is due to both differences between bark and wood, as well as differences between hardwood and softwood. The differences are mainly linked to changes in composition and amounts of cellulose, hemicellulose, lignin, and extractives. Moreover, 30 wood/bark blends (2%–20% (w/w)) were analyzed, indicating the ability of the system to recognize blends from pure material. This study has shown that spectroscopy coupled with multivariate data analysis is a useful tool verifying the compliance of producer declarations and assisting experts in evaluation of the biomass quality.

Wood Fuel Procurement to Bioenergy Facilities: Analysis of Moisture Content Variability and Optimal Sampling Strategy

“…This is probably due to the fact that biomass producers use intermediate storage-generally uncovered-to create a buffer of feedstock, securing the provision to the end user in adverse weather conditions. This is confirmed by several studies on the effect of storage of wood chips [18][19][20][21][22][23][24][25][26][27][28][29][30][31][32], which leads to variations of the fuel properties, particularly if prolonged for several months. Additionally, wood chips from different forest operations are delivered and mixed at the intermediate storage, thus including several species and diverse tree sections.…”

“…Nystrom and Dahlquist [23] compared several alternative technologies with the standard oven dry method, concluding that Near Infrared (NIR) spectroscopy and Radio Frequency (RF, also known as dielectric) technologies were the best-suited for MC measurement in flow and bulk fuels, respectively. More recently, further studies confirmed the potential of NIR instruments for MC determination due to their real-time and intuitive approach and the possibility to measure samples without any specific preparation or alteration of the biomass [24], allowing reiterate analysis on the same sample, if required. However, fuel-specific calibration is necessary for correct NIR deployment, making it a less flexible option in conditions of high variability of the biomass characteristics [19].…”

Chemical characterization of cork, phloem and wood from different Quercus suber provenances and trees

“…In this work we tentatively assigned this band to lignin because it appears in the lignin Klason spectra from the same samples (not shown). Toscano et al. (2017) already highlighted this peak as one of the most relevant wavelengths for the discrimination between bark and wood samples.…”

Visible-Near Infrared Spectroscopy and Chemometric Methods for Wood Density Prediction and Origin/Species Identification

“…When infrared energy falls on the surface of wood samples, the energy of the C-H, N-H, and O-H bonds is excited above their ground state and, depending on inner structure, composition, and surface feature, the Vis-NIR spectra can be translated into fiber morphology and chemistry information through multivariate modeling and removal of noise and irrelevant information [19,20]. Therefore, advanced spectral optimizing techniques should be explored to extract key information and improve model performance.…”