Monitoramento em tempo real da hidrólise enzimática do bagaço da casca de coco verde por espectroscopia no infravermelho próximo (NIRS)

• Ano de defesa: 2016
• Autor(a) principal: Nascimento, Ruthinéia Jéssica Alves do
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Brasil; UFRN; PROGRAMA DE PÓS-GRADUAÇÃO EM ENGENHARIA QUÍMICA
• Palavras-chave em Português: Monitoramento; NIRS; Hidrólise enzimática; Etanol de 2ª geração; CNPQ::ENGENHARIAS::ENGENHARIA QUIMICA
• Link de acesso: https://repositorio.ufrn.br/jspui/handle/123456789/21611

This study investigates the application of near infrared spectroscopy for rapid qualitative characterization of coconut husk fiber and evaluates the potential of this lignocellulosic material for producing second-generation ethanol. The near infrared spectroscopy associated with methods of data mathematical pretreatment and multivariate calibration methods have been used to their potential for real time monitoring of coconut husk fiber enzymatic hydrolysis process. The coconut husk was subjected to four types of physico-chemical pretreatments: Pretreatment with dilute sulfuric acid (2% w/v), pretreatment with dilute phosphoric acid (0.2% w/v), alkaline pretreatment (NaOH 0.5% w/v) and hydrothermal pretreatment. Morphological changes, chemical composition and crystallinity index variations caused by the physical-chemical pretreatments were evaluated. Also, the influence of pretreatments in enzymatic hydrolysis yield and the theoretical ethanol yield was investigated. The changes in lignin, hemicellulose and crystalline cellulose were characterized by traditional methods such as SEM and XRD as well as using an alternative analytical method such as NIR spectroscopy showing good results. Pretreatment with dilute phosphoric acid proved to be more efficient aiming the production of bioethanol, resulting in aproximately 39.16% of enzymatic hydrolysis yield and 83.68 L/t of theoretical ethanol yield. The monitoring calibration models were built and optimized with the cross-validation process and the models predictive ability were analyzed by the external validation. The mathematical pretreatments performed on spectral data were Standard normal variation (SNV), Multiplicative scatter correction (MSC) and Smoothing moving average (SMA) coupled with 1st and 2nd derivative of Savitzky-Golay. The calibration intervals for ART, glucose, cellulose conversion and theoretical ethanol yield were: 0.12 – 7.87 g.L-1; 0.00 – 5.87 g.L-1; 0.00 – 28.85% and 0,00 – 61,68 L/t of fiber. The best calibration models for quantification of ART, glucose, cellulose conversion and the theoretical ethanol yield were obtained for the mathematical pretreatment SMA + 2nd derivative having R2 and RMSEP values equal to, 0.98 and 0.1994 g.L-1, for ART; 0.99 and 0,0266 g.L-1 for glucose; 0.99 and 0.1839% for cellulose conversion; 0.99 and 0.2279 L/t of theoretical ethanol yield. The obtained calibration models were used for monitoring an enzymatic hydrolysis process of coconut husk with satisfactory results.