Face to the increasing demand for energy, the declining availability of fossil fuels and the periodical tensions in the geopolitical context of oil-producing countries have raised serious concerns in the governments worldwide as well as in the public opinion. Similarly, awareness has been increasing of the environmental harms caused by burning fossil fuels and of the associated adverse effects on human health. A number of different potential sources of renewable energy have been investigated over the last decades, including wind power, solar energy and biomass, among others. In particular, concerning the use of biomass, the production of ethanol as a transportation fuel has been considered a particularly attractive and feasible option. Production of ethanol from feedstocks containing easily fermentable sugars such as corn or sugar cane has been actually implemented, especially in the US and in Brazil, but it has been harshly debated especially on the ground of the induced competition between energy and food uses. Hence, a so-called second generation production has been envisaged, which makes use of lignocellulosic materials, namely biomass residues from different agricultural and industrial processes, which appear particularly interesting due to their abundance and the lack of possible use for human consumption. Unfortunately, the production of ethanol from cellulosic feedstocks still presents technical challenges to be overcome before production at the industrial scale could become profitable. In a number of tropical countries in Africa, Asia and South America, important amounts of agricultural residues are generated during the process of oil extraction from the fruits of the palm tree Elaeis guineensis. Among these residues, the empty bunches that carry the fruitlets are at present severely underused, despite their high content of cellulose (of roughly 30%) and the presence of other saccharides. In the scientific literature, relatively limited attention has been given to this feedstock, when compared to other residues such as, for instance, corn stover. Moreover, different studies can be difficult to compare because of the many dissimilar factors involved, as the diverse geographical origin of the residues, and different techniques and analytical protocols used. This thesis deals with the comparison of possible alternatives for processing empty fruit bunches to produce ethanol. The attention has been focused on the pretreatment step, the first operation performed on the feedstock, which is crucial because of its cost and impact on downstream processing (enzymatic saccharification and fermentation). Thus, three pretreatment techniques generally accepted as among the most promising have been chosen: dilute acid, dilute alkali and liquid hot water treatment. They have been performed on the same feedstock under controlled laboratory conditions. For each pretreatment, experimental parameters (such as temperature, time, pH, etc.) have been varied according to a statistical experiment design. Pretreated samples have been enzymatically hydrolyzed to assess the degree of conversion of cellulose into fermentable glucose. The influence of the conditions employed during pretreatment have been evaluated using surface response analysis, which also enabled to define optimal conditions with respect to the total amount of glucose liberated upon enzymatic saccharification. Rigorous compositional analyses using wet chemistry protocols have been carried out on the samples before and after pretreatment, allowing their complete characterization. The fate of each type of compound of the biomass could thus be followed during the processing including pretreatment and enzymatic saccharification. Additionally, IR spectroscopy has been used to retrieve precious information about the nature of the solid substrate. The liquids produced during pretreatment have been similarly analyzed to detect the presence of solubilized sugars as well as that of a number of degradation products which could be inhibitors for microorganisms used in a successive fermentation stage. Taken together, all these elements allowed to investigate the mode of action of the pretreatment techniques employed and establishing a meaningful comparison among the options when empty fruit bunches are used as a feedstock. An in-depth discussion is presented concerning the effectiveness of each treatment with respect to the disruption of the recalcitrant lignocellulosic structure, the improvements in glucose conversions which are attainable upon pretreatment, and other significant criteria such as the production of inhibitors and the possible use of the liquid streams. All these factors, in fact, may have a strong impact on a possible extension from the laboratory to the pilot-plant or industrial scale. Treatments using either a dilute sulfuric acid solution or only pure water have proved overall more efficient than the dilute alkali treatment in attacking the lignocellulosic matrix and relevant amounts of sugars could be liberated under optimized conditions. Oppositely, the alkali treatment under the conditions explored shows limitations probably attributable to the feedstock considered, especially given its relevant lignin content. Each pretreatment, however, has characteristic strengths and weaknesses that must be considered carefully before adopting a choice, especially in relation with the possible products of interest: an approach including the production of different chemicals along with ethanol could probably be a realistic option for possible future scale-up. One of the major contributions of this thesis is hence to provide a comparative study of different techniques performed under standardized conditions, and to investigate a feedstock of interest for many countries for which however information about its response to pretreatment techniques is still fragmented.
EPFL2012
