Concentrating hemicellulosic hydrolysates with different membrane technologies

Abstract

This thesis constitutes a part of the "IMPRESS" project which is aiming at the production of bio-plastics by utilizing hemicellulosic hydrolysates. The hydrolysate stream, after following the steps of acid hydrolysis and neutralization, consists of monosaccharides (C5/C6 sugars), salt and sugar degradation products. The C5/C6 sugars need to be concentrated by a factor of 10 to be purified afterwards in downstream processes. Evaporation has been the most dominant technology in the sugar treatment sector, however the current study aimed to investigate tubular nanofiltration (NF) and vacuum membrane distillation (VMD) units for concentrating the C5/C6 sugars, as more sustainable alternatives. Synthetic solutions with C5/C6 sugars and salt were used to assess the performance of the NF and MD membranes in terms of sugar and salt rejection before the experiments with the real hydrolysate solution. Among the different NF and MD membranes, the one that met the C5/C6 sugar rejection requirements (>90 %) during the synthetic solution experiments was tested for concentrating the C5/C6 sugars. The most suitable technology for concentrating the C5/C6 sugars was also assessed for its energy consumption. Additionally, single filtration tests were conducted to characterize the NF membranes, while surface tension and contact angle measurements were performed to predict the membrane wetting phenomena during the VMD operation. Based on the C5/C6 sugar rejection values (range of 22-81 %) obtained with ceramic and polymeric NF membranes, it was deduced that the tubular NF membranes were not capable of meeting the C5/C6 sugar concentration goals. The polymeric membranes performed better than the ceramic ones in terms of C5/C6 sugar rejection, with a maximum sugar rejection of 81 %. However, the polymeric membranes were found to be more prone to fouling, showing a maximum decrease of 34 % in the water permeability after 4 hours of operation with the hydrolysate solution. Additionally, the characterization of the NF membranes based on the Donnan Steric Pore Model (DSPM) revealed membrane pore sizes larger than those provided by the manufacturer. On the contrary, the 0.2 μm MD membrane showed >99 % of C5/C6 sugar rejection and was used for the sugar concentration tests. The maximum C5/C6 sugar concentration factor (CF) achieved with the VMD was 8 with a negligible sugar loss (<1 %) in the permeate, indicating that concentrating the C5/C6 sugars by a factor of 10 is feasible with this technology. Surface tension and contact angle measurements disclosed inconsistencies in the hydrophobicity of the MD membrane, which can be linked to the membrane wetting phenomena occurred during the VMD operation. The fouling of the MD membrane (50 % of flux decrease after 39 hours of operation) was found to be reversible with complete flux recovery after chemical cleaning and drying of the membrane. The energy assessment of the VMD technology showed that the main energy consumer was the cooler, contributing to 96 % of the total consumed energy. Based on the cooler's efficiency, the energy consumption of the VMD unit was calculated to be in the range of 207-736 KWh/m³ of distillate. When compared with multi-effect evaporators with up to three effects, the VMD was found to be more energy efficient. Investigating less energy-intensive alternatives for concentrating the C5/C6 sugars, experiments with electrodialysis (ED) showed 90 % removal of both acid and salt from the raw and neutralized hydrolysate water, respectively, making it feasible for reverse osmosis (RO) membranes to be further tested for concentrating the C5/C6 sugars. Therefore, a treatment scheme of ED and tubular RO membranes is proposed for further research on concentrating the C5/C6 sugars. Overall, a pre-treatment step with the most permeable ceramic NF membrane is suggested, as evinced by the high color removal achieved (elimination of big foulants) and the high permeation of the C5/C6 sugars and salt.