Hydrothermal fractionation is a well-known process for lignocellulosic biomass upgrading. It is based on the continuous treatment of biomass with hot pressurized water, converting the biomass major components (hemicellulose, cellulose and lignin) into soluble compounds. Thus, it is one of the most promising options to produce chemicals and energy from biomass since it only requires water and temperature. However, it is a highly complex process that involves a great deal of physical phenomena: biopolymer and oligomer cleavage, sugar production and degradation, acid releasing, homogeneous acid catalysis, solid-liquid mass transfers and porosity changes. For this reason, the development of a comprehensive kinetic model for biomass hydrothermal fractionation is a difficult matter. Despite these burdens, several modelling options that go from the simplest option, a first order kinetic model, to deeper and complete studies, can be found in the literature. These models are able to reproduce the experimental data of different biomasses in both, packed bed and batch reactors. Nonetheless, they are generally used for just one reactor or for one specific biomass, like spruce or wheat bran. Thereby, this work was aimed at developing an overall model for hemicellulose extraction, validating it with data from different packed bed reactors and for completely different biomasses. Additionally, this model included the whole set of physical phenomena. So, a novel reaction path way with all the physico-chemical process involved in the both phases, the liquid and the solid, was proposed. The model was obtained applying a mass balance for each compound in this mechanism and it was solved applying advanced numerical methods (orthogonal collocation over finite elements and Runge-Kutta with 8th order of convergence). Regarding kinetics, auto-catalytic expressions were selected since they have been demonstrated to be a suitable option to simulate quick changes in concentration profiles. On the other hand, the data of the hydrothermal fractionation were taken from 4 different reactors and using 3 biomasses (holm oak, wheat straw and catalpa) were used (0.1, 3, 6 and 40L). The temperature range was between 140 ºC and 215 ºC to focus the study on the hemicellulose extraction and the residence time was fixed around 5 min to promote extraction without degrading the hydrolysate. The model reproduced the experimental behaviour with average deviations between 6 and 50%. It is worth mentioning that the higher deviations were obtained when also the uncertainty of the experimental measure of the compounds was high. The simulated profiles included: oligomers, hexoses, pentoses, acetic acid, degradation products and pH. Furthermore, it is worth mentioning that the fractionation was mainly controlled by the soluble oligomers production and their hydrolysis kinetics. Therefore, it could be concluded that, for the cases studied, if temperature and pH profiles inside two different reactors are similar, their global behaviour will be also the same, independently of the biomass studied and reactor properties. Finally, it should be remarked that all the calculated parameters had physical meaning.
La licencia del ítem se describe como Attribution-NonCommercial-NoDerivatives 4.0 International