Cracking of Sugars for Production of Chemicals

Abstract

The sugar cracking process is a promising technology, which may be useful for renewable chem-ical production of glycolaldehyde and other oxygenate products including pyruvaldehyde, ace-tol, glyoxal and formaldehyde. Glycolaldehyde is a renewable platform molecule that can be used for further conversion of other useful chemicals such as ethylene glycol, methyl vinyl glycolate (MVG) and ethanol amines. Ethylene glycol is a large commodity chemical mainly produced from petrochemical ethylene and is primarily used in the synthesis of polyester fibers and PET bottles. Producing ethylene glycol from sugars through the sugar cracking process would allow for a renewable production of ethylene glycol contributing to the green transition. In the sugar cracking process, aqueous solutions of sugars are sprayed into a fluidized bed reactor in a fast pyrolysis type process. The sugar cracking process has been investigated in this thesis with the objective to optimize the yield of glycolaldehyde and to improve the mechanistic un-derstanding of the process. This has been achieved through experimental investigation the effects of operating conditions, and by developing a kinetic model for the process. The process has been investigated for a wide of range operating parameters, including time on stream, bed temperature, feed concentration, feed substrate, liquid feed flowrate, gas feed flowrate, spray atomization parameters, downstream temperature, type of bed material and bed material loading. The process generally exhibited robustness with respect to variation in several parameters. However, some parameters were found to strongly influence the performance of the process. These include bed temperature, type of bed material and liquid feed flowrate. The highest yield of glycolaldehyde obtained from glucose in this thesis was 74%, which is also the highest yield reported in the open literature. The optimum temperature for glycolaldehyde production was found to be 515–525 °C. At higher temperature, the yield was reduced due to formation of permanent gases. At lower temperature, the yield was reduced and the process was more difficult to operate due to increased defluidization tendencies, likely due to formation char and other products. The type of bed material was an important parameter influencing the yield of glycolaldehyde. Silibeads Type S 90–150 μm (non-porous sodium silicate) showed the highest yield of glycolal-dehyde among the bed materials tested in this thesis. From experimental observation, the best performing materials were found to be non-porous particles of silicate or metal oxides. It is proposed that the role of the bed material is to be an inert heat carrier providing high heat transfer, facilitating rapid evaporation and heating in the fluidized bed reactor, minimizing undesired reactions rather than to have a catalytic role in the sugar cracking process. Investigation of the effects of the liquid feed flowrate showed that operation at >2 g/min resulted in little to no formation of anhydrosugars, whereas reducing the liquid feed flowrate significantly increased formation of anhydrosugars and reduced formation of glycolaldehyde. At 0.5 g/min, the yield of glycolaldehyde was only 17%, while the yield of anhydrosugars (levoglucosan and 1,6-anhydroglucofuranose) was 20%. Various mono-, di- and polysaccharides were fed and tested in the sugar cracking process. From experiments feeding monosaccharides, the yield of glycolaldehyde was found to decrease in the order glucose > xylose > fructose ≈ 1,3-dihydroxyacetone > erythrulose. A reaction network was developed based primarily on isomerization and retro-aldol reactions that can account forthe relative distribution of oxygenate products feeding different monosaccharides. Di- and poly-saccharides showed lower carbon balance yields (likely due to increased char formation) compared to monosaccharides, although similar condensed products were formed. Cellulose and starch were fed from stirred suspensions and not aqueous solutions, and yielded 10–12% gly-colaldehyde. These experiments show the importance of the initial ring-opening step that is nec-essary for the retro-aldol reactions to occur. A kinetic model has been developed for the sugar cracking process. It is based on literature data of computational kinetic parameters and on empirical kinetic parameters based on experiments with glycolaldehyde and 1,3-dihydroxyacetone that were carried out and modelled in this thesis. The kinetic model was validated against sugar cracking experiments of glucose, fructose, xylose and erythrulose at different temperatures. The kinetic model captured important observations of sugar cracking of the various sugars at different temperatures, including the optimum temperature range for maximum glycolaldehyde yield. The kinetic model showed some limitations in that glycolaldehyde yields were generally overpredicted by 5–15%. Similarly, formaldehyde was also overpredicted, while pyruvaldehyde was underpredicted. This suggests that there may be limitations in the reaction network and room for further improvement of the model. The reactions modelled in this work were proposed based on experimental observations, but may occur from different pathways. Further work to improve the model could be to further investigate the decomposition reactions to ensure that the correct reactions are modelled. Furthermore, the ki-netic model is an isothermal gas-phase model that neglects important aspects such as bed mate-rial and its interaction with droplets. This thesis contributes with new insights to the sugar cracking process, improving the mechanistic understanding and optimizing the yield of glycolaldehyde. This may help further development of the mechanistic understanding and further improvements useful during process scale-up.