Development of pretreatment technology and enzymatic hydrolysis for biorefineries: Dissertation

The growing demand for energy, materials and food, depletion of fossil raw material reservoirs and increasing environmental concerns have all increased interest in renewable resources. Lignocellulosic biomass is an alternative for replacing fossil raw materials in the production of fuels, materials and various chemicals. Lignocellulose present in plant cell walls consists mainly of polysaccharides, cellulose and hemicellulose, and aromatic lignin. These major components form a complex structure that is resistant to microbial and enzymatic activity. Due to the recalcitrant structure of plant cell walls, lignocellulosic raw materials must be pretreated before their enzymatic hydrolysis to monosaccharides. Various pretreatment methods; chemical, physical, biological or their combinations, have been developed. After pretreatment, polysaccharides can be hydrolysed enzymatically to monosaccharides, which in turn can be fermented to different products such as ethanol. Currently the first commercial scale lignocellulosic ethanol plants have started production. A secure supply of biomass is one of the key factors for a feasible biorefinery, and new alternative feedstocks are still required especially in northern climates in order to fulfil the raw material demands of biorefineries in a sustainable way. In addition, development of new pretreatment technologies and more efficient enzymatic hydrolysis are needed. New lignocellulosic feedstocks and improved pretreatment methods were studied in the work described in this thesis. Reed canary grass and barley straw were found to be interesting carbohydrate-rich raw materials that could be pretreated by steam explosion and hydrolysed enzymatically with yields comparable to those obtained from wheat straw. Selection of the most favourable harvest time for reed canary grass, autumn or spring, was studied in relation to pretreatment and hydrolysis yields. Spring harvested reed canary grass was found to be the more suitable raw material as it had a higher cellulose content and the pretreated fibre was hydrolysed more efficiently compared to autumn harvested material. A new pretreatment method using sodium carbonate and oxygen pressure was developed. The alkaline oxidation method fractionated biomass into a carbohydrate-rich fibre and a dissolved fraction containing most of the lignin. The produced carbohydrate-rich fibre could be efficiently hydrolysed by enzymes and the hydrolysis was also efficient at 12% dry matter content. Compared to the 52% total glucose yield obtained in enzyme hydrolysis of spruce after pretreatment by steam explosion, a significantly higher glucose yield of 84% was obtained in hydrolysis after alkaline oxidation. Different kinds of raw materials, such as spruce, birch and sugar cane bagasse, could be efficiently pretreated by alkaline oxidation. The main effects of alkaline oxidation pretreatment were dissolution and partial degradation of lignin and hemicellulose. Some galactoglucomannan and xylan was solubilised and further oxidised to other products, and therefore relatively low yields of hemicellulose were obtained. Organic acids were formed as degradation products of lignin and carbohydrates. Process conditions were partially optimized using spruce as raw material in order to improve the efficiency of alkaline oxidation. The pretreatment could be accelerated by increasing the treatment temperature, by the use of copper-phenanthroline catalyst, and by decreasing the particle size of the raw material. Further optimization of e.g. alkali dosage and the solid to liquid ratio is still required to improve hemicellulose yield and economical feasibility. Fibre fractions of alkaline oxidation could be hydrolysed by low enzyme dosages, 2-4 FPU/g dry matter. Significantly higher enzyme dosages were required in the hydrolysis of steam exploded materials, probably due to the inhibitory effect of the high residual lignin content after the pretreatment. The efficient hydrolysis of alkaline oxidised materials by low enzyme dosages can decrease enzyme costs or enable shorter hydrolysis time. In order to further improve the hydrolysis efficiency and decrease the required enzyme dosage, enzyme mixtures were optimized regarding the major enzymes needed in biomass hydrolysis. Optimized mixtures of thermostable enzymes were found to have significantly different proportions of cellobiohydrolases, endoglucanases and xylanase than the optimized mixtures of Trichoderma reesei enzymes. Although different, the significant role of cellobiohydrolases was demonstrated in both types of mixtures. The results also indicated that high xylanase activity was required in the hydrolysis of pretreated materials having decreased enzyme accessibility to cellulose due to high xylan content or possibly due to drying of the substrate. The hydrolysis performance of optimized enzyme mixtures of five thermostable enzyme components was shown to be close to that of state-of-the-art commercial mixtures.