Solubilidade da madeira de Eucalyptus por meio da espectroscopia no infravermelho próximo

• Ano de defesa: 2016
• Autor(a) principal: Ribeiro, Yllian Banchieri
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Lavras; Programa de Pós-Graduação em Ciência e Tecnologia da Madeira; UFLA; brasil; Departamento de Ciências Florestais
• Palavras-chave em Português: Madeira – Extratos vegetais; Solução (Química); Radiação infravermelha; Wood – Plant extracts; Solution (Chemistry); Infrared radiation; Resinas de Madeiras; Química da Madeira
• Link de acesso: http://repositorio.ufla.br/jspui/handle/1/11916

This research aimed to evaluate the extractive content present in wood of Eucalyptus hybrid clones by different solvents of distinct polarities, and adjust prediction models for wood solubility in organic solvents by near infrared spectroscopy technic (NIRS) using solid and powdered wood. Twelve clones, ten at six years old and two at 3.75 years old, were used in this study. These genotypes were coming from municipality of Itamarandiba, located in North of Minas Gerais State, Brazil. Five sample-trees were used from each clone. A composed sample was prepared at each clone, considering a sampling of seven longitudinal positions to determine the extractive content as well as obtaining the NIR spectrum. In addition to this sampling, short logs were removed at breast height diameter (BHD) to obtain NIR spectrum in the solid wood at tangential plane (heartwood and sapwood position) and radial plane (transition position between sapwood and heartwood). The following solvents were used for quantifying extractives content: cold water, acetone, ethanol and extraction sequence in toluene:ethanol (2:1 v/v), ethanol and hot water. A Bruker spectrometer, MPA model, was used to obtain NIR spectrum in the wood. The wavelength used by equipment ranges from 12,500 to 3,500 cm -1 , with spectral resolution of 8 cm -1 , and operating in diffuse reflection mode. The NIR spectrum was processed by multivariate statistical of partial least square regression (PLS regression) for setting the prediction models. The results showed that extraction sequence in toluene: ethanol, ethanol, and hot water were the procedure removed most quantity of wood extractives, ranging from 3.76% to 6.27%, with average value of 4.93%. The models based on NIR to predict the wood extractives content calibrated by spectrum obtained in wood powder had better statistics compared with those obtained in solid wood. The best multivariate model to estimate the solubility of the timber was obtained based on measured NIR spectrum of wood powder and the ethanol solvent, with R 2 of 0.88 and 0.37, respectively, in calibration and cross validation.