Thesis (MSc)--Stellenbosch University, 2011. / ENGLISH ABSTRACT: The need to mitigate the twin crises of peak oil and climate change has driven a headlong rush to biofuels. This study was aimed at the development of a process to efficiently convert steam explosion pretreated (STEX) sugarcane bagasse into ethanol by using combinations of commercial enzyme cocktails and recombinant Saccharomyces cerevisiae strains. Though enzymatic saccharification is promising in obtaining sugars from lignocellulosics, the low enzymatic accessibility of the cellulose and hemicellulose is a key impediment thus necessitating development of an effective pretreatment scheme and optimized enzyme mixtures with essential accessory activities. In this context, the effect of uncatalysed and SO2 catalysed STEX pretreatment of sugarcane bagasse on the composition of pretreated material, digestibility of the water insoluble solids (WIS) fraction and overall sugar recovery was investigated. STEX pretreatment with water impregnation was found to result in a higher glucose recovery (28.1 g/ 100 bagasse) and produced WIS with a higher enzymatic digestibility, thus was used in the optimization of saccharification and fermentation. Response surface methodology (RSM) based on the 33 factorial design was used to optimize the composition of the saccharolytic enzyme mixture so as to maximize glucose and xylose production from steam exploded bagasse. It was established that a combination of 20 FPU cellulase/ g WIS and 30 IU -glucosidases/ g WIS produced the highest desirability for glucose yield. Subsequently the optimal enzyme mixture was used to supplement enzyme activities of recombinant yeast strains co-expressing several cellulases and xylanases in simultaneous saccharification and fermentations SSFs. In the SSFs, ethanol yield was found to be inversely proportional to substrate concentration with the lowest ethanol yield of 70% being achieved in the SSF at a WIS concentration of 10% (w/v). The ultimate process would however be a one-step “consolidated” bio-processing (CBP) of lignocellulose to ethanol, where hydrolysis and fermentation of polysaccharides would be mediated by a single microorganism or microbial consortium without added saccharolytic enzymes. The cellulolytic yeast strains were able to autonomously multiply on sugarcane bagasse and concomitantly produce ethanol, though at very low titres (0.4 g/L). This study therefore confirms that saccharolytic enzymes exhibit synergism and that bagasse is a potential substrate for bioethanol production. Furthermore the concept of CBP was proven to be feasible. / AFRIKAANSE OPSOMMING: Die behoefte om die twee krisisse van piek-olie en klimaatsverandering te versag, het veroorsaak dat mense na biobrandstof as alternatiewe energiebron begin kyk het. Hierdie studie is gemik op die ontwikkeling van 'n proses om stoomontplofde voorafbehandelde (STEX) suikerriet bagasse doeltreffend te omskep in etanol deur die gebruik van kombinasies van kommersiële ensiem mengsels en rekombinante Saccharomyces cerevisiae stamme. Alhoewel ensiematiese versuikering belowend is vir die verkryging van suikers vanaf lignosellulose, skep die lae ensiematiese toeganklikheid van die sellulose en hemisellulose 'n hindernis en dus is die ontwikkeling van' n effektiewe behandelingskema en optimiseerde ensiemmengsels met essensiële bykomstige aktiwiteite noodsaaklik. In hierdie konteks, was die effek van ongekataliseerde en SO2 gekataliseerde stoomontploffing voorafbehandeling van suikerriet bagasse op die samestelling van voorafbehandelde materiaal, die verteerbaarheid van die (WIS) breuk van onoplosbare vastestowwe in water (WIS), en die algehele suikerherstel ondersoek. Daar was bevind dat stoomontploffing behandeling (STEX) met water versadiging lei tot 'n hoër suikerherstel (21.8 g/ 100g bagasse) en dit het WIS met ‘n hoër ensimatiese verteerbaarheid vervaardig en was dus gebruik in die optimalisering van versuikering en fermentasie. Reaksie oppervlak metodologie (RSM), gebasseer op die 33 faktoriële ontwerp, was gebruik om die samestelling van die ‘saccharolytic’ ensiemmengsel te optimaliseer om sodoende die maksimering van glukose en ‘xylose’ produksie van stoomontplofde bagasse te optimaliseer. Daar was bevestig dat ‘n kombinasie van 20 FPU sellulase/ g WIS en 30 IU ‘ -glucosidases/ g’ WIS die hoogste wenslikheid vir glukose-opbrengs produseer het. Daarna was die optimale ensiemmengsel gebruik om ensiemaktiwiteit van rekombinante gisstamme aan te vul, wat gelei het tot die medeuitdrukking van verskillende ‘cellulases’ en ‘xylanases’ in gelyktydige versuikering en fermentasie SSFs. In die SSFs was daar bevind dat die etanol-produksie omgekeerd proporsioneel is tot substraat konsentrasie, met die laagste etanolopbrengs van 70% wat bereik was in die SSF by ‘n WIS konsentrasie van 10% (w/v). Die uiteindelike proses sal egter 'n eenmalige "gekonsolideerde" bioprosessering (CBP) van lignosellulose na etanol behels, waar die hidrolise en fermentasie van polisakkariede deur' n enkele mikroorganisme of mikrobiese konsortium sonder bygevoegde ‘saccharolytic’ ensieme bemiddel sal word. Die ‘cellulolytic’ gisstamme was in staat om vanself te vermeerder op suikerriet bagasse en gelyktydig alkohol te produseer, al was dit by baie lae titres (0.4 g/L). Hierdie studie bevestig dus dat ‘saccharolytic’ ensieme sinergisme vertoon en dat bagasse 'n potensiële substraat is vir bio-etanol produksie. Daar was ook onder meer bewys dat die konsep van CBP uitvoerbaar is. / The National Research Foundation (NRF) for financial support

Dans un contexte de transition énergétique et de lutte contre le réchauffement climatique, la production d’éthanol de seconde génération semble une voie très prometteuse afin de réduire notre dépendance aux énergies fossiles. Il existe 3 étapes clés pour la production de ce nouveau biocarburant : le prétraitement qui permet de déstructurer la matrice lignocellulosique afin de rendre la cellulose plus accessible aux enzymes, l’hydrolyse enzymatique qui a pour but de produire des sucres fermentescibles et la fermentation qui permet de transformer ces sucres en éthanol. Actuellement, le prétraitement considéré comme le plus efficace, et principalement retenu par les industriels, est le prétraitement par explosion à la vapeur. Cependant, certains aspects comme les effets physicochimiques induits par le prétraitement ainsi que leurs impacts sur les caractéristiques de la biomasse prétraitée restent encore mal compris.Schématiquement, le prétraitement par explosion vapeur peut se décomposer en deux étapes : la première se rapproche d’une cuisson acide réalisée à 150-200°C durant 5 30 min et permet principalement l’hydrolyse des hémicelluloses, alors que la seconde est une détente explosive qui permet un éclatement mécanique du substrat rendant potentiellement la cellulose plus réactive à l’hydrolyse enzymatique. Globalement les effets de ce type de prétraitement sur la biomasse lignocellulosique sont aujourd’hui bien connus mais la compréhension des différents phénomènes physico-chimiques ayant lieu en son sein reste limitée. En effet le découplage de l’étape de cuisson et de l’étape de détente est délicat car, la température du réacteur (qui impacte principalement les réactions de cuisson) est directement liée à sa pression (qui impacte principalement la détente) par la thermodynamique des phases.Ce travail de thèse se propose donc de mieux appréhender l’ensemble des phénomènes physico-chimiques ayant lieu durant le prétraitement par explosion à la vapeur en s’appuyant notamment sur une discrimination expérimentale des phénomènes chimiques (réactions de dépolymérisation) et des phénomènes physiques (détente explosive) ainsi que sur une caractérisation multi-techniques et multi-échelles de la biomasse lignocellulosique obtenue après ce type de prétraitement. L’objectif est aussi de comprendre quelles sont les principales caractéristiques de la biomasse qui expliquent les différences de réactivité observées lors de l’étape d’hydrolyse enzymatique et d’expliquer l’impact du prétraitement par explosion à la vapeur sur les propriétés physicochimiques et donc sur la réactivité. / In a context of energy transition and climate change challenge, the production of second generation ethanol seems to be a very promising way to reduce our dependence on fossil fuels. There are 3 key steps for producing this new biofuel: pretreatment to decompose the lignocellulosic biomass and to make cellulose more accessible to enzyme attacks, enzymatic hydrolysis to produce fermentable sugars and fermentation to convert these sugars into ethanol. Currently, the pretreatment considered to be the most efficient, and mainly retained for industrialization, is the steam explosion pretreatment. However, some aspects such as the physicochemical effects induced by pretreatment and their impacts on the characteristics of pretreated biomass remain misunderstood.Schematically, the steam explosion pretreatment can be separated into two stages: the first is similar to an acid cooking carried out at 150-200°C during 5-30 min and allows mainly the hydrolysis of hemicelluloses, while the second is an explosive release which allows a mechanical bursting of the substrate potentially making the cellulose more reactive to enzymatic hydrolysis. As a whole, the effects of this type of pretreatment on lignocellulosic biomass are now well known, but the understanding of the different physicochemical phenomena occurring within it remains limited. Indeed, decoupling the cooking stage and the expansion stage is complicated because the reactor temperature (which mainly impacts the cooking reactions) is directly related to its pressure (which mainly impacts the explosive release) by the phase thermodynamics.This thesis work aims to better understand all the physicochemical phenomena occurring during a steam explosion pretreatment, based in particular on experimental discrimination of chemical phenomena (depolymerization reactions) and physical phenomena (explosive release) as well as on a multi-technical and multi-scale characterization of the lignocellulosic biomass obtained after this type of pretreatment. The objective is also to understand what are the main characteristics of biomass that explain the differences in reactivity observed during the enzymatic hydrolysis step and to explain the impact of the steam explosion pretreatment on the physicochemical properties and therefore the reactivity.

The two main sources of biomass for energy generation are purpose-grown energy crops and waste materials. Energy crops, such as Miscanthus and short rotation woody crops (coppice), are cultivated mainly for energy purposes and are associated with the food vs. fuels debate, which is concerned with whether land should be used for fuel rather than food production. The use of residues from agriculture, such as barley, canola, oat and wheat straw, for energy generation circumvents the food vs. fuel dilemma and adds value to existing crops. In fact, these residues represent an abundant, inexpensive and readily available source of renewable lignocellulosic biomass. In order to reduce industrys operational cost as well as to meet the requirement of raw material for biofuel production, biomass must be processed and handled in an efficient manner. Due to its high moisture content, irregular shape and size, and low bulk density, biomass is very difficult to handle, transport, store, and utilize in its original form. Densification of biomass into durable compacts is an effective solution to these problems and it can reduce material waste. Upon densification, many agricultural biomass materials, especially those from straw and stover, result in a poorly formed pellets or compacts that are more often dusty, difficult to handle and costly to manufacture. This is caused by lack of complete understanding on the natural binding characteristics of the components that make up biomass. An integrated approach to postharvest processing (chopping, grinding and steam explosion), and feasibility study on lab-scale and pilot scale densification of non-treated and steam exploded barley, canola, oat and wheat straw was successfully established to develop baseline data and correlations, that assisted in performing overall specific energy analysis. A new procedure was developed to rapidly characterize the lignocellulosic composition of agricultural biomass using the Fourier Transform Infrared (FTIR) spectroscopy. In addition, baseline knowledge was created to determine the physical and frictional properties of non-treated and steam exploded agricultural biomass grinds. Particle size reduction of agricultural biomass was performed to increase the total surface area, pore size of the material and the number of contact points for inter-particle bonding in the compaction process. Predictive regression equations having higher R2 values were developed that could be used by biorefineries to perform economic feasibility of establishing a processing plant. Specific energy required by a hammer mill to grind non-treated and steam exploded barley, canola, oat and wheat straw showed a negative power correlation with hammer mill screen sizes. Rapid and cost effective quantification of lignocellulosic components (cellulose, hemicelluloses and lignin) of agricultural biomass (barley, canola, oat and wheat) is essential to determine the effect of various pre-treatments (such as steam explosion) on biomass used as feedstock for the biofuel industry. A novel procedure to quantitatively predict lignocellulosic components of non-treated and steam exploded barley, canola, oat and wheat straw was developed using Fourier Transformed Infrared (FTIR) spectroscopy. Regression equations having R2 values of 0.89, 0.99 and 0.98 were developed to predict the cellulose, hemicelluloses and lignin compounds of biomass, respectively. The average absolute difference in predicted and measured cellulose, hemicellulose and lignin in agricultural biomass was 7.5%, 2.5%, and 3.8%, respectively. Application of steam explosion pre-treatment on agricultural straw significantly altered the physical and frictional properties, which has direct significance on designing new and modifying existing bins, hoppers and feeders for handling and storage of straw for biofuel industry. As a result, regression equations were developed to enhance process efficiency by eliminating the need for experimental procedure while designing and manufacturing of new handling equipment. Compaction of low bulk density agricultural biomass is a critical and desirable operation for sustainable and economic availability of feedstock for the biofuel industry. A comprehensive study of the compression characteristics (density of pellet and total specific energy required for compression) of ground non-treated and steam exploded barley, canola, oat and wheat straw obtained from three hammer mill screen sizes of 6.4, 3.2 and 1.6 mm at 10% moisture content (wb) was conducted. Four preset pressures of 31.6, 63.2, 94.7 and 138.9 MPa, were applied using an Instron testing machine to compress samples in a cylindrical die. It was determined that the applied pressure (60.4%) was the most significant factor affecting pellet density followed by the application of steam explosion pre-treatment (39.4%). Similarly, the type of biomass (47.1%) is the most significant factor affecting durability followed by the application of pre-treatment (38.2%) and grind size (14.6%). Also, the applied pressure (58.3%) was the most significant factor affecting specific energy required to manufacture pellets followed by the biomass (15.3%), pre-treatment (13.3%) and grind size (13.2%), which had lower but similar effect affect on specific energy. In addition, correlations for pellet density and specific energy with applied pressure and hammer mill screen sizes having highest R2 values were developed. Higher grind sizes and lower applied pressures resulted in higher relaxations (lower pellet densities) during storage of pellets. Three compression models, namely: Jones model, Cooper-Eaton model, and Kawakita-Ludde model were considered to determine the pressure-volume and pressure-density relationship of non-treated and steam exploded straws. Kawakita-Ludde model provided the best fit to the experimental data having R2 values of 0.99 for non-treated straw and 1.00 for steam exploded biomass samples. The steam exploded straw had higher porosity than non-treated straw. In addition, the steam exploded straw was easier to compress since it had lower yield strength or failure stress values compared to non-treated straw. Pilot scale pelleting experiments were performed on non-treated, steam exploded and customized (adding steam exploded straw grinds in increments of 25% to non-treated straw) barley, canola, oat and wheat straw grinds obtained from 6.4, 3.2, 1.6 and 0.8 mm hammer mill screen sizes at 10% moisture content (wb). The pilot scale pellet mill produced pellets from ground non-treated straw at hammer mill screen sizes of 0.8 and 1.6 mm and customized samples having 25% steam exploded straw at 0.8 mm. It was observed that the pellet bulk density and particle density are positively correlated. The density and durability of agricultural straw pellets significantly increased with a decrease in hammer mill screen size from 1.6 mm to 0.8 mm. Interestingly, customization of agricultural straw by adding 25% of steam exploded straw by weight resulted in higher durability (> 80%) pellets but did not improve durability compared to non-treated straw pellets. In addition, durability of pellets was negatively correlated to pellet mill throughput and was positively correlated to specific energy consumption. Total specific energy required to form pellets increased with a decrease in hammer mill screen size from 1.6 to 0.8 mm and also the total specific energy significantly increased with customization of straw at 0.8 mm screen size. It has been determined that the net specific energy available for production of biofuel is a significant portion of original agricultural biomass energy (89-94%) for all agricultural biomass.

The two main sources of biomass for energy generation are purpose-grown energy crops and waste materials. Energy crops, such as Miscanthus and short rotation woody crops (coppice), are cultivated mainly for energy purposes and are associated with the food vs. fuels debate, which is concerned with whether land should be used for fuel rather than food production. The use of residues from agriculture, such as barley, canola, oat and wheat straw, for energy generation circumvents the food vs. fuel dilemma and adds value to existing crops. In fact, these residues represent an abundant, inexpensive and readily available source of renewable lignocellulosic biomass. In order to reduce industrys operational cost as well as to meet the requirement of raw material for biofuel production, biomass must be processed and handled in an efficient manner. Due to its high moisture content, irregular shape and size, and low bulk density, biomass is very difficult to handle, transport, store, and utilize in its original form. Densification of biomass into durable compacts is an effective solution to these problems and it can reduce material waste. Upon densification, many agricultural biomass materials, especially those from straw and stover, result in a poorly formed pellets or compacts that are more often dusty, difficult to handle and costly to manufacture. This is caused by lack of complete understanding on the natural binding characteristics of the components that make up biomass. An integrated approach to postharvest processing (chopping, grinding and steam explosion), and feasibility study on lab-scale and pilot scale densification of non-treated and steam exploded barley, canola, oat and wheat straw was successfully established to develop baseline data and correlations, that assisted in performing overall specific energy analysis. A new procedure was developed to rapidly characterize the lignocellulosic composition of agricultural biomass using the Fourier Transform Infrared (FTIR) spectroscopy. In addition, baseline knowledge was created to determine the physical and frictional properties of non-treated and steam exploded agricultural biomass grinds. Particle size reduction of agricultural biomass was performed to increase the total surface area, pore size of the material and the number of contact points for inter-particle bonding in the compaction process. Predictive regression equations having higher R2 values were developed that could be used by biorefineries to perform economic feasibility of establishing a processing plant. Specific energy required by a hammer mill to grind non-treated and steam exploded barley, canola, oat and wheat straw showed a negative power correlation with hammer mill screen sizes. Rapid and cost effective quantification of lignocellulosic components (cellulose, hemicelluloses and lignin) of agricultural biomass (barley, canola, oat and wheat) is essential to determine the effect of various pre-treatments (such as steam explosion) on biomass used as feedstock for the biofuel industry. A novel procedure to quantitatively predict lignocellulosic components of non-treated and steam exploded barley, canola, oat and wheat straw was developed using Fourier Transformed Infrared (FTIR) spectroscopy. Regression equations having R2 values of 0.89, 0.99 and 0.98 were developed to predict the cellulose, hemicelluloses and lignin compounds of biomass, respectively. The average absolute difference in predicted and measured cellulose, hemicellulose and lignin in agricultural biomass was 7.5%, 2.5%, and 3.8%, respectively. Application of steam explosion pre-treatment on agricultural straw significantly altered the physical and frictional properties, which has direct significance on designing new and modifying existing bins, hoppers and feeders for handling and storage of straw for biofuel industry. As a result, regression equations were developed to enhance process efficiency by eliminating the need for experimental procedure while designing and manufacturing of new handling equipment. Compaction of low bulk density agricultural biomass is a critical and desirable operation for sustainable and economic availability of feedstock for the biofuel industry. A comprehensive study of the compression characteristics (density of pellet and total specific energy required for compression) of ground non-treated and steam exploded barley, canola, oat and wheat straw obtained from three hammer mill screen sizes of 6.4, 3.2 and 1.6 mm at 10% moisture content (wb) was conducted. Four preset pressures of 31.6, 63.2, 94.7 and 138.9 MPa, were applied using an Instron testing machine to compress samples in a cylindrical die. It was determined that the applied pressure (60.4%) was the most significant factor affecting pellet density followed by the application of steam explosion pre-treatment (39.4%). Similarly, the type of biomass (47.1%) is the most significant factor affecting durability followed by the application of pre-treatment (38.2%) and grind size (14.6%). Also, the applied pressure (58.3%) was the most significant factor affecting specific energy required to manufacture pellets followed by the biomass (15.3%), pre-treatment (13.3%) and grind size (13.2%), which had lower but similar effect affect on specific energy. In addition, correlations for pellet density and specific energy with applied pressure and hammer mill screen sizes having highest R2 values were developed. Higher grind sizes and lower applied pressures resulted in higher relaxations (lower pellet densities) during storage of pellets. Three compression models, namely: Jones model, Cooper-Eaton model, and Kawakita-Ludde model were considered to determine the pressure-volume and pressure-density relationship of non-treated and steam exploded straws. Kawakita-Ludde model provided the best fit to the experimental data having R2 values of 0.99 for non-treated straw and 1.00 for steam exploded biomass samples. The steam exploded straw had higher porosity than non-treated straw. In addition, the steam exploded straw was easier to compress since it had lower yield strength or failure stress values compared to non-treated straw. Pilot scale pelleting experiments were performed on non-treated, steam exploded and customized (adding steam exploded straw grinds in increments of 25% to non-treated straw) barley, canola, oat and wheat straw grinds obtained from 6.4, 3.2, 1.6 and 0.8 mm hammer mill screen sizes at 10% moisture content (wb). The pilot scale pellet mill produced pellets from ground non-treated straw at hammer mill screen sizes of 0.8 and 1.6 mm and customized samples having 25% steam exploded straw at 0.8 mm. It was observed that the pellet bulk density and particle density are positively correlated. The density and durability of agricultural straw pellets significantly increased with a decrease in hammer mill screen size from 1.6 mm to 0.8 mm. Interestingly, customization of agricultural straw by adding 25% of steam exploded straw by weight resulted in higher durability (> 80%) pellets but did not improve durability compared to non-treated straw pellets. In addition, durability of pellets was negatively correlated to pellet mill throughput and was positively correlated to specific energy consumption. Total specific energy required to form pellets increased with a decrease in hammer mill screen size from 1.6 to 0.8 mm and also the total specific energy significantly increased with customization of straw at 0.8 mm screen size. It has been determined that the net specific energy available for production of biofuel is a significant portion of original agricultural biomass energy (89-94%) for all agricultural biomass.