Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The effect of steam explosion pretreatment parameters on softwood delignification efficiency Cullis, Ian Frank 2003

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2003-0135.pdf [ 6.28MB ]
Metadata
JSON: 831-1.0075019.json
JSON-LD: 831-1.0075019-ld.json
RDF/XML (Pretty): 831-1.0075019-rdf.xml
RDF/JSON: 831-1.0075019-rdf.json
Turtle: 831-1.0075019-turtle.txt
N-Triples: 831-1.0075019-rdf-ntriples.txt
Original Record: 831-1.0075019-source.json
Full Text
831-1.0075019-fulltext.txt
Citation
831-1.0075019.ris

Full Text

THE EFFECT OF STEAM EXPLOSION PRETREATMENT PARAMETERS ON SOFTWOOD DELIGNIFICATION EFFICIENCY by Ian Frank Cullis B.Sc, University of British Columbia, 2000 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (Faculty of Forestry) We accept this thesis as conforming J Q the required standard THE UNIVERSITY OF BRITISH COLUMBIA January 2003 © Ian Frank Cullis, 2003 I n p r e s e n t i n g t h i s t h e s i s i n p a r t i a l f u l f i l m e n t o f t h e r e q u i r e m e n t s f o r a n a d v a n c e d d e g r e e a t t h e U n i v e r s i t y o f B r i t i s h C o l u m b i a , I a g r e e t h a t t h e L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r r e f e r e n c e -a n d s t u d y . I f u r t h e r a g r e e t h a t p e r m i s s i o n f o r e x t e n s i v e c o p y i n g o f t h i s t h e s i s f o r s c h o l a r l y p u r p o s e s may be g r a n t e d b y t h e h e a d o f my d e p a r t m e n t o r b y h i s o r h e r r e p r e s e n t a t i v e s . I t i s u n d e r s t o o d t h a t c o p y i n g o r p u b l i c a t i o n o f t h i s t h e s i s f o r f i n a n c i a l g a i n s h a l l n o t be a l l o w e d w i t h o u t my w r i t t e n p e r m i s s i o n . D e p a r t m e n t The U n i v e r s i t y • o f B r i t i s h C o l u m b i a V a n c o u v e r , C a n a d a A B S T R A C T Current environmental problems have led to concerns about the condition of the environment and the effect of automobile emissions on global warming. Recent work has determined that there are other alternate forms of energy, which are cleaner burning, and can be produced from renewable resources. Some alternative sources of energy include the combustion of hydrogen in the fuel cell and ethanol. Ethanol fuel has shown significant promise due largely to its ability to be used in internal combustion engines that are currently running gasoline or gasoline ethanol blends. Previous work has shown that steam explosion is a viable pretreatment technique for the conversion of biomass to ethanol, and that medium severity conditions are the best compromise between hemicellulose recovery and cellulose digestibility. To further develop this pretreatment strategy the effect of feedstock variation (moisture content and chip size) on the steam explosion process was examined, and the effect of the substrate on subsequent fractionation and hydrolysis were evaluated. Additionally, the effect of a secondary treatment regime, post steam explosion particle size reduction, was examined for its effect on fractionation and hydrolysis efficiency. In the next series of experiments the effect of EDTA chelation, stabilization (DTMPA, Sodium Silicate, and Magnesium Sulphate), consistency, and alternative delignification techniques (oxygen and wet oxidation) were examined. The techniques, which showed an improvement in delignification efficiency, were then used to optimize chemical loading and temperature during peroxide fractionation. The manipulation of feedstock conditions (moisture content and chip size) caused noticeable variations in the properties and effectiveness of further stages of the bioconversion process. Increased chip size caused an increase in the solid recovery, increasing from 62 to 82 %, with concurrent increases in the prehydrolysate sugar recovery (7.5 %). Increased recovery is the result of decreased relative severity of steam treatment as chip size increases. Decreased severity affects the overall process by decreasing the recalcitrance of lignin and therefore increasing the efficacy of peroxide fractionation, which removed 16 %-more lignin from the largest chip size. Similarly n increased initial moisture content appeared to reduce the relative severity of the treatment, prehydrolysate sugars (mainly glucose and mannose), and solid recovery. Both increasing chip size and moisture content results in a substrate that performs better in peroxide delignification and enzymatic hydrolysis. Furthermore, post steam-explosion refining solubilized more of the glucose and mannose present in the prehydrolysate, resulting in a decrease in the solid recovery, while concurrently increasing prehydrolysate sugar recovery. The resulting solid substrate was more effectively delignified, resulting in a decrease in the residual lignin of 4.6 % in the largest chip size. Improvements in the peroxide delignification process previously optimized by Yang et al. (2002) were achieved by altering substrate consistency, and the addition of peroxide stabilizers and chelants. Increasing consistency from 2 to 10 % resulted in a decrease in the residual lignin content from 5.4 % to less than 3 % respectively. Additionally, stabilization (DTMPA) reduced the residual lignin content, after optimization a 40 % decrease in lignin content is achievable, while maintaining glucose yield from enzymatic hydrolysis. Therefore, by optimizing the conditions it was possible to reduce the chemical (peroxide) loading by greater than 40 %. in T A B L E O F C O N T E N T S A B S T R A C T ii T A B L E O F C O N T E N T S iv L I S T O F F I G U R E S viii L I S T O F T A B L E S xi L I S T O F A B B R E V I A T I O N S xii A C K N O W L E D G M E N T S xiii 1 I N T R O D U C T I O N 1 1.1 Sources of biomass 5 1.1.1 Agricultural residues 6 1.1.2 Municipal waste 7 1.1.3 Industrial waste 8 1.2 Biomass chemistry 9 1.3 Treatment methods 14 1.3.1 Mechanical treatment and hydrolysis 15 1.3.2 Acid hydrolysis 16 1.3.3 Enzymatic hydrolysis 17 1.3.3.1 AFEX 19 1.3.3.2 Steam explosion 20 1.4 Delignification 25 1.4.1 Peroxide delignification 26 1.4.2 Oxygen delignification 27 1.4.3 Wet oxidation ..28 iv 1.5 Research objectives 29 2 M A T E R I A L S A N D M E T H O D S 31 2.1 Pretreatment stage of the bioconversion process 31 2.2 Substrate 33 2.2.1 Chip fractionation 33 2.2.2 Chip conditioning 35 2.3 Steam explosion 35 2.4 Post steam explosion refining 36 2.5 Water wash 37 2.6 Delignification 37 2.6.1 Hot alkali peroxide delignification 38 2.6.2 Chelation 39 2.6.3 Peroxide delignification consistency 40 2.6.4 Peroxide loading 41 2.6.5 Factorial matrix 41 2.6.6 Oxygen delignification .42 2.6.7 Wet oxidation 42 2.6.8 Alkali wash 43 2.7 Hydrolysis 43 2.8 Substrate composition 44 2.8.1 Acid soluble and insoluble lignin 44 2.8.2 Sugar analysis 44 2.8.3 Monomer/ oligomer analysis 45 v 2.8.4 Glucose analysis 45 2.8.5 Furan analysis 45 2.8.6 Mass balance determination 46 2.8.7 Statistical analysis 46 3 CHIP SIZE 47 3.1 Introduction 47 3.2 Steam explosion sugar recovery 49 3.3 Fractionation 56 3.4 Hydrolysis 60 3.5 Conclusion 63 4 MOISTURE CONTENT 64 4.1 Introduction 64 4.2 Steam explosion 65 4.3 Fractionation 69 4.4 Hydrolysis 72 4.5 Conclusion 77 5 REFINING 80 5.1 Introduction 80 5.2 Steam explosion/refining 81 5.3 Fractionation 83 5.4 Hydrolysis 83 5.5 Conclusion 90 v i 6 DELIGNIFIC ATION 91 6.1 Introduction , 91 6.2 Alternative delignification techniques 92 6.3 Peroxide delignification 95 6.4 Chelation and stabilization 98 6.5 Optimization 103 6.6 Conclusion 112 CONCLUDING REMARKS 113 FUTURE WORK 116 BIBLIOGRAPHY 117 APPENDIX A 123 APPENDIX B 124 APPENDIX C 125 APPENDIX D 126 APPENDIX E 127 vii LIST OF FIGURES Figure 1. Lignin content across the cell wall 12 Figure 2. Wood lignin monomers (1) coniferyl alcohol and (2) sinapyl alcohol 13 Figure 3. Flow diagram illustrating the generation of substrate from Douglas-fir wood chips for steam explosion, subsequent refining of the substrate, peroxide fractionation, and hydrolysis during the pretreatment stage of the bioconversion process 32 Figure 4. Categories of Douglas-fir chips used in the steam explosion process. A) 40-mesh, B) 1.5 x 1.5, C) 5 x 5 34 Figure 5. Sugar content recovered (g) in the water-soluble fraction after steam explosion of wood chips of different initial sizes at 30 % moisture content 50 Figure 6. Concentration of individual neutral sugars recovered in the water-soluble fraction after steam explosion of wood chips of different initial particle size at 30 % moisture content 51 Figure 7. Oligomeric sugars recovered in the water-soluble fraction after steam explosion of woods chips of different initial size at 30 % moisture content 53 Figure 8. Monomeric sugar recovered in the water-soluble fraction after steam explosion of wood chips with different initial size at 30 % moisture content'. 54 Figure 9. Furans (furfural and hydroxymethylfurfural) recovered in the water-soluble fraction after steam explosion of chips with different initial sizes at 30 % moisture content 55 Figure 10. Solids recovered from the steam explosion process of wood chips with different initial moisture contents at 30 % moisture content 57 Figure 11. Concentration of residual lignin in the solids fraction recovered after steam explosion of wood chips with different initial sizes at 30 % moisture content 58 Figure 12. Total residual lignin retained in the solid fraction after hot alkali peroxide (1 % peroxide, 2 % substrate, 195°C, pH 11.5) fractionation 59 Figure 13. Rate and yield of enzymatic hydrolysis of the cellulosic fractions resulting from the steam explosion of different wood chip classes followed by HAP 62 viii Figure 14. Total carbohydrates recovered in the water-soluble fraction recovered from steam explosion of chips of different initial size (Very fine, Medium, Very coarse) and 12 % and 30 % moisture contents 66 Figure 15. Concentration of carbohydrates recovered in the water-soluble fraction derived from wood chips with different initial size at (A) 12 % and (B) 30 % moisture content 68 Figure 16. Total recovery of solid substrate after steam explosion of chips of different initial size and moisture contents 70 Figure 17. Total lignin retained in the solid portion after steam explosion of substrate derived from different chip sizes and moisture contents 71 Figure 18. Residual lignin after steam explosion and fractionation (HAP) of wood chips derived from different chip sizes and moisture contents 73 Figure 19. Hydrolysis (20 FPU/g, pH 4.8, 50 mM sodium acetate) rates and yield of substrate derived from steam explosion and fractionation of wood chips with different size at (A) 12 % and (B) 30 % initial moisture content 74 Figure 20. Comparison of hydrolysis rates from chips derived from very coarse chips with different moisture contents 76 Figure 21. Total carbohydrate recovered in the water-soluble fraction derived from steam explosion of chips of different initial size and moisture content subject to refining. 82 Figure 22. Total recovery of solid substrate after steam explosion of chips of different initial size and moisture content subjected to refining 84 Figure 23. Total lignin retained in the solid fraction after steam explosion of substrate derived from different chip sizes and moisture contents subjected to refining 85 Figure 24. Residual lignin after steam explosion, fractionation (1 % peroxide, 2 % substrate, 195°C, pH 11.5) and refining of wood chips derived from different chip sizes and moisture content subjected to refining 86 Figure 25. Comparison of the hydrolysis rates of substrates derived from wood chips of different size: (A) 12 % moisture content; (B) 12 % initial moisture content and refined prior to fractionation; (C) 30 % initial moisture content; (D) 30 % initial moisture content and refined prior to fractionation 87 Figure 26. Comparison of hydrolysis rates from chips derived from very coarse chips with different moisture contents subjected to refining 88 ix Figure 27. The effect of different fractionation chemicals on residual lignin in steam exploded Douglas-fir very coarse wood chips, which were refined 93 Figure 28. The effect of substrate concentration during delignification on residual lignin in Douglas-fir feedstock derived from coarse wood chips that were steam exploded, refined and fractionated 96 Figure 29. Rate of decompostion of residual peroxide during delignification under three conditions, standard: 1 % H2O2, 2 % substrate, 80°C, pH 11.5; Double peroxide: 2 % H 2 0 2 , 2 % substrate, 80°C, pH 11.5; Substrate Blank: 1 % H 2 0 2 , 80°C, pH 11.5 97 Figure 30. The effect of stabilizers, ethylene-diamine-tetraaccetic acid (EDTA), diethylene-triamine-pentamethylene-phosphonic acid (DTMPA), and sodium silicate and magnesium sulphate in combination with ethylene-diamine-tetraaccetic acid (EDTA), on residual peroxide at optimized peroxide 1 % peroxide, 2 % substrate, 80°C, pH 11.5) delignification conditions 100 Figure 31. The effect of stabilizers on residual lignin in Douglas-fir substrate derived from coarse wood chips that have been steam exploded, refined and fractionated Figure 32. The effect of varying peroxide concentration on residual lignin content of steam exploded Douglas-fir chelated with EDTA and treated with peroxide stabilized with DTMPA 104 Figure 33. Rate of hydrolysis of steam-exploded Douglas-fir substrate subjected to different peroxide loadings during fractionation 105 Figure 34. Three-dimensional surface plot showing the relationship between temperature (°C), peroxide concentration (% w/v) and residual lignin content 108 Figure 35. Three-dimensional surface plot showing the relationship between temperature (°C), DTMPA loading (% w/v) and residual lignin content 109 Figure 36. Three-dimensional surface plot showing the relationship between peroxide concentration (% w/v), DTMPA loading (% w/v) and residual lignin content 110 Figure 37. Hydrolysis of optimized (factorial design) delignification conditions (Very Low- 0.45 % peroxide, 65°C, and 0.18 % DTMPA; Low- 0.465 % peroxide, 75°C, and 0.18 % DTMPA; High- 0.480 % peroxide, 85°C, and 0.18 % DTMPA; Very High- 0.50 % peroxide, 95°C, and 0.18 % DTMPA), determined by statistical analysis, of Douglas-fir wood chips that have been steam exploded, refined and fractionated I l l x L I S T O F T A B L E S Table 1. Legislation affecting the use of ethanol fuel (IEA Bioenergy task, 2000) 2 Table 2. Greenhouse gas emissions from ethanol fuels 4 Table 3. Lignocellulosic feedstock availability in the United States 5 Table 4. Recovery after steam explosion, fractionation and hydrolysis of the original components of the chips entered into steam explosion (% of the original component weight) 61 Table 5. Recovery after steam explosion, fractionation and hydrolysis of the original components of the chips entered into steam explosion (g/lOOg of original wood)...75 Table 6. Recovery after steam explosion, fractionation and hydrolysis of original components of the chips entered into steam explosion (% of original weight) 77 Table 7. Carbohydrate content in the prehydrolysate recovered from steam explosion and refining of the original chips (g/lOOg of original weight) 81 Table 8. Recovery of wood constituents after steam explosion, refining, fractionation and hydrolysis of original components of wood chips entered into steam explosion (% of original weight) 89 xi LIST OF ABBREVIATIONS CBU cellobiase units CBU/mL cellobiase units per millilitre cm centimetres DTMPA diethylene-triamine-pentamethylene-phosphonic acid °C degrees Celsius EDTA ethylene-diamine-tetraacetic acid FPU filter paper units FPU/mL filter paper units per millilitre g grams HAP hot alkali peroxide H 2 0 2 hydrogen peroxide HPLC high performance liquid chromatography IU/g international units per gram M molar mM millimolar min minutes mL millilitres mL/min millilitres per minute Na 2C0 3 sodium carbonate NaOH sodium hydroxide nm nanometers psi pounds per square inch % percent rpm revolutions per minute S0 2 sulphur dioxide UBC University of British Columbia VC very coarse VF very fine w/v weight per volume w/w weight per weight xii A C K N O W L E D G M E N T S First, I would like to thank my supervisor, Dr. Shawn Mansfield, for his dedication, guidance and friendship during my time spent in the Department of Wood Science at UBC. I have been lucky to work in a lab filled with talented scientists, who have helped facilitate co-operation and a learning environment within the lab. I would also like to thank Dr. Jack Saddler for his help and the conditions he create in the lab. I would also like to acknowledge the support of my parents, Frank and Janet, sisters, Jessica and Trisha, and my best friend, Terra. I would also like to thank my friends, old and new. xiii 1 INTRODUCTION Research investigating the conversion and use of biomass to produce fuel-grade ethanol has been occurring since the introduction of the internal combustion engine. For example, one of the first vehicles, the model T Ford, ran on a fuel ethanol that was derived from plant biomass. Unfortunately, the availability of cheap foreign oil imports severely hampered progress of the technology by undercutting the prices of ethanol in the original industry. More recent biomass to ethanol systems began to develop with the oil crisis of the 1970's. During this decade the threat of depleting global oil reserves led to rampantly increasing oil prices, and increased interest in alternatives to traditional large internal combustion engine vehicles. A few viable alternatives presented themselves, including the development of fuel-efficient vehicles, which have evolved into the present generation of vehicles, such as the Honda Insight that currently gets 72 miles per gallon (3.9 L/100 km) (oee.nrcan.gc.ca/autosmart/mostfuel/index.cfm). Concurrent research saw the advent of vehicles utilizing alternative fuels, including electric cars and vehicles employing fuel cells. However, the popularity of electric vehicles is poor due to the relatively high cost, limited power supply, and the short distance that can be travelled on a single charge. Similarly, the fuel cell based system has a number of drawbacks including a significantly different motor and fuel distribution system: the Ballard fuel cell uses the combustion of hydrogen and oxygen for the generation of electricity, which is then used to power the car. In contrast, the advantage of ethanol-powered vehicles is that they can retain the current fuel distribution network, and run in an internal combustion engine with only 1 minor modification. Additionally, ethanol from biomass is one of the cleanest sources of fuel for the future. The fuel itself has limited emissions and thus decreases the overall levels of pollution released into the environment, it also employs a feedstock that is renewable, and would otherwise be land filled or considered waste. Government incentives established in 1967 have increased the interest in the biofuels programs, including tax breaks and legislation. Currently, three different taxes are applied at the sale of gasoline: 10 % federal tax, 7 % GST, and provincial tax (11 % in B.C.). In comparison, companies that produce ethanol are only subjected to 9 % federal tax, and the 7 % GST. Some of the legislation passed by government for the benefit of producers of ethanol is shown in Table 1. The belief is that through government incentives the reliance on foreign fuel will be reduced, and pollution levels will be decreased. A decrease in the reliance on foreign fuel should also lead to increased domestic security, and higher employment in North America. Since, the incentive program began the production of ethanol in the USA has increased from 38 million litres in 1979 to a level of 5.3 billion litres in 1999 (IEA Bioenergy task, 2000). Table 1. Legislation affecting the use of ethanol fuel (IEA Bioenergy task, 2000). Year Legislation 1967 Air quality act of 1967 1970 Clean air amendments 1988 Alternative motor fuels act 1989 Reid vapour pressure regulations 1990 Clean air act amendments of 1990 1992 Energy policy act of 1992 1994 EPA mandatory thirty percent minimum renewable oxygenate content 1998 Transportation equity act 2 Locally, the major incentive for British Columbia to entertain alternative fuels programs is to curtail the vast amount of wood waste produced by the Forest Industry. Annually, the lumber industry in B.C. generates some 69,640 cubic metres of wood, 40 metric tons of pulp, 178,071 cubic metres of round wood, and 176,572 cubic metres of industrial round wood (www.fao.org). The extensive dependence in B.C. on the forest industry results in 8,624 cubic metres of wood residue generated annually over and above the 1,499 cubic metres of wood fuel (www.fao.org). Should a viable bioconversion program be successfully established a significant amount of revenue, which is spent by the mills on waste residue disposal, could be turned into a profit-generating commodity. The benefits of ethanol generation would be three fold; land filling costs would be reduced, generation of a value-added product from waste, and emissions generated from the transportation industry through fossil fuel consumption would be reduced. Generally, ethanol is considered a clean fuel because the growth of plants offsets the emissions of CO2 that is produced through the combustion of ethanol generated from the biomass. Since transportation is one of the major sources of global pollution, conversion of biomass to fuel-grade ethanol would substantially reduce global pollution. The burning of ethanol would also reduce energy demands placed on the environment; in the United States 3.3 billion tons of CO2 are released annually, of this vehicles account for 27 %. Of the total global carbon dioxide emissions ground transportation in the United States alone accounted for 2.5 %. In addition to CO2, other pollutants are also released into the atmosphere, including carbon monoxide, volatile organic carbon, and nitrogen oxides. In general, it has been estimated that vehicles account for 27 % of the greenhouse 3 gas, 28 % of carbon monoxide, 57 % of volatile organic carbon, and 67 % of the nitrogen oxide emissions globally (IEA Bioenergy task, 2000). It has clearly been shown that ethanol fuels have the potential to significantly reduce emissions of all of these major chemicals (Table 2) (IEA Bioenergy task, 2000). The two main producers of ethanol in the world today are Brazil and the USA, producing in total over 5.3 billion litres of ethanol for the transportation industry (IEA Bioenergy task, 2000). In contrast, Canada currently produces 210 million litres per year, primarily from wheat starch by Mohawk. Table 2. Greenhouse gas emissions from ethanol fuels United States Canada % Reduction compared to % Reduction compared to gasoline gasoline 10% ethanol blend 0.8-1.6 3.9-4.6 85 % ethanol blend 13.7-18.8 37.1-44.5 Although the generation of ethanol from tree derived biomass has not yet become a reality, other feedstocks have been employed for a number of years. For example, Brazil is one of the few countries that developed a biofuels program relatively early on in the twentieth century making use of their vast sugar cane bagasse resources. Sugar cane is easily grown, harvested, and treated for the production of ethanol due to the ideal growing conditions in Brazil, and the fact that sugar cane has a very high carbohydrate content. In 1975 the government of Brazil employed a program to ensure that the country would have a national source of fuel (Himmel, 1994), and as a result many laws were passed to make the production of fuel ethanol feasible, and more economical than gasoline. By 1985 the sale of pure ethanol fuelled cars exceeded conventional and ethanol blended cars (Himmel, 1994). 4 1.1 Sources of biomass Currently, governments and research groups in Canada, United States, Sweden, and a few other countries are searching for economical and sustainable ways to generate fuel, other than the current rapid consumption of fossil fuel sources. One of the largest sources of sustainable and relatively clean fuel is ethanol generated from plant biomass. By definition, biomass includes plants, dead plant material and products that have been produced from plant material, and emphasises that all of the plant biomass is required to meet the increasing energy needs of the planet (Robinson, 1980). At present, a variety of sources of biomass are available to the biofuels industry, including; agricultural residue, municipal solid waste, industrial solid waste, bioenergy crops, or forest residues. However, it is highly unlikely that only one feedstock will be able to provide the total ethanol needs of a country that uses large quantities of fuel such as the United States. In most countries it is likely a combination of feedstocks will be required to support and/or replace the demand for fossil fuels. Table 3 shows the availability of some lignocellulosic feedstocks present in the United States (IEA Bioenergy task, 2000). Table 3. Lignocellulosic feec stock availability in the United States Feedstock Total quantity Total available Forest residue Mill waste Urban wood waste Agricultural residues Bioenergy crops 41 million dry tons 82 million dry tons 33 million dry tons 300 million dry tons 165 million dry tons 41 million dry tons 1.5 million dry tons 22 million dry tons 140 million dry tons 165 million dry tons 5 1.1.1 Agricultural residues Agricultural residues are currently a major feedstock source for ethanol production. For example over 9 billion kilograms of corn are processed annually in Canada, while in the USA over 229 billion kilograms of corn are processed (IEA Bioenergy task, 2000). Corn is processed through the wet milling process, which does not utilize the corn fibre accounting for over 11 % of the dry weight of the corn kernel (Anderson and Watson, 1982). Corn fibre is an attractive substrate for bioconversion due to the high carbohydrate content, low cost and high quantities. Currently, wheat is also used to produce ethanol for companies such as Mohawk, which blends the ethanol into the higher-grade fuels. Mohawk currently has a plant, which produces 10 million litres of ethanol per year. However, this ethanol is simply a by-product of the protein generation by Mohawk. In the European Union it has been previously estimated that agricultural crops could sustain 7 % of their annual energy use (Robinson, 1980). This is achievable with little or no change in the current agricultural practices. In contrast, at present in North America ethanol production accounts for only 1 % of the gasoline market and this could easily be increased by 25 % with the current agricultural supplies (IEA Bioenergy task, 2000). Furthermore, as the productivity of farmland in North America increases due to the development of different strains of wheat and other agricultural products, less farmland could possibly be required for the production of food for local markets. This would release the lower grade agricultural land for the production of potential energy crops such as bagasse or switchgrass, which are able to generate large amounts of biomass in relatively short periods of time. 6 Improvements in agricultural practices are necessary to protect the soil against the loss of nutrients and therefore loss of productivity. The loss of nutrients from fertilized land is 65 % higher than leaching losses from unfertilized land (www.fao.org). This contributes to the pollution of rivers of which 72 % show nitrate pollution problems in the United States (www.fao.org). Therefore, improvements in agricultural practices could circumvent the loss of soil nutrients, and increase plant growth potentially becoming a reservoir of material for the bioconversion process. 1.1.2 Municipal waste Another potential source of residue for the bioconversion process is solid waste from municipalities, which is one of the largest sources of pollution present in the world (Robinson, 1980). In addition, the decomposition of the land filled residues generates significant quantities of gases, which contribute to the greenhouse gas effect that is affecting the global temperatures. Consequently, if this residue were to be employed for ethanol production a significant reduction in greenhouse gas emission and pollution would likely occur. Currently, over 300 million tons of this residue are produced annually (Himmel, 1994), and of this less than half is presently being sent to processing plants, leaving a significant amount of solid residue that could be employed (Robinson, 1980). Another source of waste in municipalities is sewage, 13 % of municipal sewage waste is collected annually. The generation of sewage waste is quite variable, but has been reported as being as high as 0.223 kg per person per day (Robinson, 1980). There are many potential energy uses for this residue such as anaerobic methane production or incineration energy due to the high content of organic material. 7 1.1.3 Industr ia l w a s t e Industries are also a plentiful source of solid waste that could be better used for the production of fuel grade ethanol. For example, residues generated from paper, office work and recovered material from demolition sites, and of course the value added wood products industry, could serve as sources of lignocellulosics. Another valuable source of residue included in the list of industrial sources is the forestry sector: the British Columbia softwood lumber industry is one of the leading industries in the provincial economy, and is also one of the main generators of pollution in the province. If this waste resource could be tapped and used for power generation, the goal of energy self-sustainability would be one step closer to reality. Sources of softwood waste include thinnings, sawdust, mill rejects, and dead trees left in the forest. A substantial amount of wood is left behind in the forest after cutting operations. Although a portion of this wood is required by the ecosystem to maintain nutrient levels and buffer the land against loss of nutrients, a large portion of this material could be used as a resource for a value-added industry (energy generation). Thinnings represent an additional source of biomass. Thinning trees reduces competition for resources, such as sunlight and nutrients, and subsequently increases the growth rate of the remaining trees. However, thinning may simulate the effect of surface fires in areas such as the Rocky Mountains where the occurrence of smaller fires occur frequently (Aber and Melillo, 1991). Without the effects of surface fires or thinning the build up of biomass allows fire to jump from the underbrush into the crowns of trees and become a much more deadly and destructive fire. Crown fires kill dominant trees by 8 burning foliage and branches of trees, and tend to burn hotter than the surface fires (Aber andMelillo, 1991) Sawmills produce 82 million dry tonnes of waste annually in the United States (IEA Bioenergy task, 2000). Sawmills generate waste residue consisting of chips, reject boards and sawdust. In B.C. dimensional lumber and plywood are two of the main products generated by sawmills that contribute to the waste residue problem. Although, some of this material is a good feedstock for pulp and paper production, a large portion of this material is used for the generation of hog fuel or land filled due to the composition of the material, which may have a high bark concentration. However, pulp mills also dispose of sawmill waste. For example, Harmac on Vancouver Island are given a supply of hog fuel with their chip supply to reduce the land filling and energy costs of producing pulp. Similarly, a significant amount of wood waste is generated from pulp and paper industry. The majority of the residue from pulp and paper consists of chips that are contaminated with bark, rejects, and knots. Bark and knots are not efficient for the pulping of wood and therefore are discarded. 1.2 Biomass chemistry As suggested in the previous section, there are a variety of potential sources of biomass, such as agricultural residue, municipal solid waste, industrial solid waste, biofuels feedstock, or forest residue that could be employed in the bioconversion process. Each comes with its own treatment issues. Treatment limitations become increasingly problematic as the complexity of the substrate is increased. Sugar cane or sugar beets are the easiest substrate to employ because the sugar can be directly fermented to ethanol, and as such it is the main crop used currently for ethanol production in countries such as 9 Brazil. In contrast, starch based or in the worst-case lignocellulose based feedstock treatment regimes must convert the starch or lignocellulose into its monomelic constituents before fermentation, and the subsequent generation of ethanol can then occur. Lignocellulosic-based feedstocks, such as Douglas-fir, are the main softwood feedstock available in B.C., and they are also the hardest feedstock to treat prior to fermentation. The difficulty in converting this residue to ethanol is related to the complex nature and interaction of the three main components of wood. To understand process treatment problems a general understanding of the organization and composition of the substrate must be demonstrated. The ultrastructure and chemistry of softwood resources is complex, and is a formidable barrier to the bioconversion of the lignocellulose matrix. Softwoods are mainly composed of two different types of cells: tracheids and parenchyma cells. About 90-95 % of the cells in softwoods are tracheids so these are by far the dominant cell type in softwood trees (Haygreen and Bowyer, 1996). The fact that >90 % of the cells in softwood trees are tracheids does not demonstrate the variation that occurs within this cell type. A variety of environmental factors such gravity, wind, and/or snow can exert forces on trees and cause modifications in cell wall development, forming compression wood. Compression wood contains cells that exhibit shorter length, thicker cell wall and a less lignified compound middle lamella, than normal wood (Hon and Shiraishi, 2001). Within what would be considered a normal tree cell, development occurs in the cambial zone. This is an area a few cells thick that surrounds the tree. These cells give rise to the xylem and phloem, also referred to as wood and bark, respectively. Within the 10 xylem, the cells are composed primarily of the secondary walls, which contain three main layers: Si, S2, and S3. The main ultrastructural difference between the layers of the secondary and primary cell wall is the orientation of the microfibrils. The primary wall has a high microfibril angle to allow for growth of the expanding cell. In contrast, the secondary wall (S2 layer) has a low microfibril angle. A microfibril is composed primarily of cellulose, with hemicellulose and lignin as the minor components. The cellulose forms a crystalline core, while hemicellulose is then organized around the cellulose as matrix support. Although, lignin is a minor component found in the secondary wall where it encrusts the hemicellulose and cellulose gluing the microfibril to the surrounding fibrils (Hon and Shiraishi, 2001), it is highly concentrated in the compound middle lamella (Figure 1). However, due to the volume of the secondary cell wall there is a significantly larger total amount of lignin in the secondary cell wall. Elementary microfibrils are the smallest division of wood that can be visualized in the cell wall. The length of the microfibrils depends on the crystallinity of the cellulose chain. Cellulose in softwood may contain upwards of 10,000 glucose units, but the entire chain is not believed to bond properly with the neighbouring chains, generating amorphous regions. Amorphous regions form where irregular cellulose chain organization occurs, and have the ability to absorb water due available hydrogen bonding. The chemical composition of cell walls is quite variable. Within a tree there are two main types of variation between cells: sapwood and heartwood. Generally, this change is caused by the death of parenchyma cells, as they undergo the formation of heartwood. Heartwood cells are considered dead, due to the fact that the movement of water and sap ceases to occur, and the pits generally become aspirated. The composition of the cell 11 wall does not change, instead the heartwood cells are used for the storage of chemicals inside the lumen, including polyphenols and flavonoids. These chemicals improve the durability of the wood and limit decay of the heartwood (Hon and Shiraishi, 2001). The most noticeable difference caused by the change to heartwood is a change in the colour of the wood. Cellulose is highly concentrated in the secondary cell wall, as the secondary wall constitutes >85% of the total wood cell and is primarily composed of microfibrils, and consequently contributes most of the structural integrity to the cell wall. In contrast, the lower concentration of cellulose in the compound middle lamella (less cellulose microfibrils) facilitates cell growth and expansion before the organized layers of cellulose are deposited sealing the dimensions of the fibers. Additionally, the hemicellulose is highly concentrated in the primary cell wall, while the concentration of cellulose is low. Figure 1. Lignin content across the cell wall Lignin constitutes 25-35 % of the cellular material. Two types of lignin may occur in softwoods with the majority being of the guaiacyl variety (95 %) and a minor amount of syringyl lignin (5 %) (Hon and Shiraishi, 2001). The structures of the precursors to the 12 major types of lignin monomers are shown in figure 2. In contrast, hardwoods have a significantly higher proportion of syringyl lignin, but also contain high concentrations of the guaiacyl variety. The highest proportion of guaiacyl lignin is located in the compound middle lamella and decreases towards the secondary cell wall where guaiacyl lignin is replaced by syringyl lignin. The variation of syringyl and guaiacyl lignin is due to the timing of the expression of ferulic 5-hydroxylase (F5H), an enzyme required for the generation of syringyl lignin (Beatson, 1986). OH OH (1) (2) Figure 2. Wood lignin monomers (1) coniferyl alcohol and (2) sinapyl alcohol Hemicellulose is another of the main components of wood. There are two main types of hemicellulose in softwoods: partially acetylated glucuronoarabinoxylan, and galactoglucomannan (Hon and Shiraishi, 2001). The xylan-based sugar occupies 10 % of the total composition of the tree, while the mannan-based sugar occupies roughly 18 % (Hon and Shiraishi, 2001). One of the main reasons that softwoods are harder to pretreat than hardwoods is a decrease in the amount of acetylated hemicellulose sugars. The 13 breakdown of the acetyl groups leads to the formation of acetic acid, which aids in the hydrolysis of the hemicellulose sugars. Cellulose is the main constituent of wood, occupying up to 64 % of the total wood volume (Hon and Shiraishi, 2001). B-D-glycopyranose units are linked via glucosidic linkages between the Ci and C 4 carbons of the individual units to form cellulose chains. The chains are then linked via intermolecular bonds, which attach adjacent cellulose chains to form microfibrils, and as the microfibrils aggregate fibrils are formed (Sjostrom, 1981). Due to the complex structure and crystallinity of cellulose, it has a high tensile strength and is insoluble in most solvents For effective and economical production of ethanol, all of the components of wood must be separated. Separation of the individual components is difficult due to close interaction between individual constituents, and the complexity of the substrate. Usually, the separation of the components is a balance between the amount removed and the degradation of the other components. 1.3 Treatment methods The goal of biomass to ethanol programs is to generate as much ethanol and value-added by-products as possible, while minimizing the degradation of substrate. As has just been suggested previously, softwood lignocellulosics are an extremely difficult substrate to treat due to their highly complex ultrastructure, and therefore effective pretreatment is required to be a compromise between substrate loss and effective separation of components (Nguyen et al, 1998; Wu et al, 1999). A variety of techniques exist for the hydrolysis of cellulose and hemicellulose to produce sugar streams for fermentation, including mechanical, acidic, or enzymatic treatments. All of these 14 techniques are quite costly and require optimization for feedstock composition, size, moisture content, and severity of treatment. Each of these techniques has a variety of benefits and limitations. Furthermore, some of the techniques are impractical due to the problems with decomposition and overall cost. Unfortunately, none of the techniques are able to fully account for the actual variation that occurs in actual biomass substrate. Ultimately, the successful technique must have low capital and operating costs, and be able to withstand variations in substrate chemical composition, without severe loss of productivity. 1.3.1 Mechanical treatment and hydrolysis Mechanical degradation of the lignocellulosic matrix can be preformed using a vibratory ball milling techniques (Millett et al, 1976; Himrnel, 1994). Ball milling breaks down the crystallinity of the cellulose (Fan et al, 1987), and in the most severe cases completely digests the biomass. This form of treatment utilizes a substantial amount of energy and as a result is an uneconomical pretreatment for the production of ethanol. However, all of the treatment types for lignocellulosic material require some form of mechanical pretreatment to reduce the inherent size of the substrate. This may be in the form of chipping, refining or milling of the original wood. Although some of the wood may arrive at a facility as a mechanically pretreated (i.e. chipped) substrate, some of the facilities may need to further mechanically treat the wood, before other techniques such as steam explosion may be initiated. Generally, it is agreed that mechanical hydrolysis alone is an uneconomical treatment, due to high costs associated with the energy consumption during mechanical treatments 15 (Himmel, 1994), and low value of the ethanol produced. However, refining may be used as a mild pretreatment for other processes such as steam explosion or enzymatic saccharification to enhance the effectiveness of these processes. Clearly, a balance must be attained between energy consumed and the benefits that a mechanical pretreatment could have on the overall bioconversion process. 1.3.2 Acid hydrolysis Acid hydrolysis has been used for a long time as an effective means of cellulose saccharification (Himmel, 1994). This method utilizes an acid, such as sulphuric, hydrochloric, or phosphonic, to hydrolyse the glycosidic bonds of the cellulose chain. There are, however, two types of acid hydrolysis; these are dilute acid and concentrated acid hydrolysis. Concentrated acid hydrolysis utilizes a strong acid, such as hydrochloric at a concentration of 72 %, and operates at or near room temperature. If the temperature is increased, degradation of the carbohydrate moieties into HMF and furfural will result and, consequently, decrease the yield of the overall conversion process. A room temperature process also has the added advantage of minimal energy costs. Unfortunately, the equipment and chemicals needed for concentrated acid hydrolysis are very costly, due to corrosion and the consumption of acid during the process. Research evaluating acid recycling and optimizing acid levels has reduced the cost of the process, but is still a major consideration if this process were to be commercially employed. The advantage of the concentrated acid process is the high yield of glucose, usually 90 % of theoretical. 16 Dilute acid hydrolysis is the preferred technique for the acid treatments of lignocellulosics. The advantage of the dilute acid pretreatment is the significantly reduced cost of the acids required (Galbe and Zacchi, 2002). Acid concentrations typically range from 4-10 %, with the reaction temperature elevated to 60°C to increase the rate of hydrolysis (Nguyen et al, 1999). These conditions lead to substantial energy consumption, and due to low acid loadings a recycling regime is not required, although the effluent may need to be treated. Unfortunately, the elevated temperature imparts substantial equipment corrosion and sugar degradation, thereby offsetting the advantages gained by reducing acid concentrations (Galbe and Zacchi, 2002). An alternative acid system that could potentially operate is a pretreatment for the enzymatic hydrolysis of cellulose. This pretreatment would reduce the decomposition of hemicellulose as a result of shorter residence times required with the acids, and concurrently improve the porosity of the lignocellulosic matrix (by removal of hemicellulose) and ultimately improve the enzymatic digestion of the cellulose. 1.3.3 Enzymatic hydrolysis The use of enzymes for the saccharification of cellulose has been the subject of a great deal of investigation recently (Converse et al, 1990; Vallander and Eriksson, 1991; Nidetzky et al, 1994; Yu and Saddler, 1995; Converti et al, 2000; Tengborg et al, 2001; 2001). Currently, enzyme-based biomass to ethanol processes are proving to be cheaper than acid or mechanical hydrolysis due to the continually decreasing cost of enzymes, lack of equipment corrosion, and the potential for enzyme recycling. Unfortunately, the complexity of lignocellulosic matrixes limits the ability of enzymes to access the substrate. Basically, enzymes must be able to penetrate the 17 lignocellulose matrix if effective hydrolysis of the polymers is to occur (Mansfield et al, 1999). Recent investigations have suggested that it is the distribution of the lignin on the substrate, which limits the accessibility of the substrate to enzymes (Mooney et al, 1998). The redistribution of lignin during pretreatment decreases the pore spaces present in the substrates and thereby decreases the accessibility of enzymes to the substrate (Wong et al, 1988). Pretreatment processes generally impart increased pore volume and surface area, and therefore increased available area for enzymes to interact with the substrate. Much work has focused on substrate factors such as fibre porosity, crystallinity, lignin content, degree of polymerization, and surface area (Lee et al, 1982; Saddler et al, 1982; Wong et al, 1988; Sinitsyn et al, 1991; Mooney et al, 1999; Chang and Holzapple, 2000), and has found that these factors affect the rates of hydrolysis. Unfortunately, the contribution of these individual factors on the rate of hydrolysis of a complex substrate is questionable. However, the factors exerting the most influence over the efficacy of enzymatic hydrolysis of softwood based lignocellulosic substrate are residual lignin content, and inhibition of enzymes (Schwald et al, 1988; Excoffier et al, 1991; Philippidis et al, 1993). The inhibitory compounds formed during the pretreatment of the substrate come in the form of carboxylic acids, phenols, vanillin, and />coumaric acid (Excoffier et al, 1991). However, other forms of enzyme inhibition may also occur during the saccharification reaction, such as end product inhibition due to the sugars generated (Philippidis et al, 1993; Teixeira et al, 2000). Lignin has been shown to acts as an inhibitor during enzymatic hydrolysis of lignocellulosic substrates (Excoffier et al, 1991). It has also been suggested that lignin 18. redistribution during steam explosion (Ramos et al, 1999) can block the pores formed during steam explosion if the treatment severity is high, and therefore inhibits effective hydrolysis (Wong et al, 1988; Ooshima, 1990). Consequently, the removal of lignin, prior to enzymatic hydrolysis, substantially improves the rate and yield of hydrolysis (Wong et al, 1988; Excoffier et al, 1991). Lignin removal is also important for recycling of enzymes since lignin irreversibly binds with hydrolytic enzymes (Gregg and Saddler, 1996). Currently, a few different techniques are being evaluated as pretreatment methods for enzymatic hydrolysis of lignocellulosic residues. These techniques include AFEX, or ammonia fibre explosion, and steam explosion. Previous work also investigated carbon dioxide explosion, but found few benefits (Himmel, 1994). The advantage of using one of these pretreatment techniques is that a substantial portion of the hemicellulose-derived sugars are removed, increasing the cellular porosity, and concurrently decreasing cell wall thickness, surface area and length (Wong et al, 1988; Toussaint et al, 1991). 1.3.3.1 A F E X Ammonia fibre explosion is a process that attempts to open up the fibre cell wall, through the removal of hemicellulose and explosive decompression. The technique, which is used, is similar to steam explosion, but the pressure is created by the vapour pressure of NH3. Substrates subjected to the AFEX procedure are ground (dimensions of 1-2 mm) and possess relatively high moisture contents. This substrate is then submersed in a pressure vessel with liquid NH3. The partial pressure of NH3 is quite high at ambient temperatures so the pressure in the chamber rises, reaching as high as 180 psi. 19 The major effect of APEX is increased accessible surface area of the wood fibres. This happens primarily due to the explosive decompression of the wood chips. When the pressure of the NH3 within the wood chips is released, a rapid expansion of the NH3 occurs and the surface area of the wood chips increase. As well, increases in surface area occur due to the hydrolysis of hemicellulose, by a base formed by the reaction of NH3 with the inherent moisture in the chips. A secondary effect of the AFEX process is a decrease in the crystallinity of the wood cellulose. The crystallinity is reduced due to the swelling of the lignocellulose matrix. As ammonia penetrates into the chips it causes a decrease in the relative crystallinity, and therefore an increase in the reactivity of the cellulose to enzymatic hydrolysis. 1.3.3.2 Steam explosion Steam explosion is a technique that has proven to be an effective method for the pretreatment of wood prior to the enzymatic hydrolysis of the cellulose (Avellar and Glasser, 1998; Boussaid et al, 2000). It is the most promising pretreatment technique for an enzymatic-based biomass to ethanol process, due to the low capital and production costs, and the efficient nature of the process. However, further improvements such as enzyme recycle could improve the economic viability of this process (Gregg et al, 1998). At present steam explosion operates at conditions that are a compromise between hydrolysis efficiency and the overall substrate recovery after steam explosion. The conditions presently employed were optimized for both hydrolysis efficiency and recovery of softwood substrates such as Douglas-fir are 195°C, 4.5 minutes, and 4.5 % SO2 (Boussaid et al, 2000). These conditions have shown to give a high recovery, in the range of 80-90 %, and a hydrolysis yield as high as 95 %. 20 The pretreatment conditions and their, relationship to" the recovery are governed by a severity parameter (Overend and Chornet, 1987), defined by the equation: Ro=t-exp[Tr-Tb/14.75] [1] t = residence time (min) T r = reaction temperature (°C) Tb = reference temperature (100°C) This equation takes into account the main parameters of the steam explosion process, but if an acid catalyst is applied to the substrate the effects of the acid must also be taken into account. To take into account the effect of acid the combined severity parameter was defined (Chum et al, 1990). CS = l o g R o - p H [2] The combined severity parameter takes into account the ability of an acid catalyst to increase the effectiveness of steam explosion to liberate the sugars within the lignocellulosic matrix (Chum et al, 1990). At present, a severity factor (log R o ) of 3.45 is used in most steam pre-treatment regimes for softwoods substrates, as this severity is a good compromise between cellulose digestibility and hemicellulose sugar degradation. At severities above log R o 3.45 excessive hemicellulose degradation occurs, while at lower severities cellulose digestibility is limited (Boussaid et al, 1998). Currently, the acid catalyst that has proven to be the most effective is SO2 (Nguyen et al, 1998; Tengborg et al, 1998). It is believed that the moisture in the chips and steam generated during processing will react with the SO2 through disappropriation and oxidation to form sulphuric acid, which can then effectively cleave hemicellulose and cellulose glycosidic bonds during the steam explosion process (Tengborg et al, 1998). 21 Previous work has looked at the effectiveness of SO2 and compared it to other treatments such as CO2, and H2SO4, and found that yields were the highest with the sulphur dioxide gas (Tengborg et al, 1998; Teixeira et al, 1999). The improvement in recovery with SO2 could be the result of an even and rapid treatment, compared to H2SO4, and the strength of the acid catalyst formed, when compared to CO2. Steam explosion has a variety of effects on the wood chips. The first and most obvious of these is the decompression effect due to the explosion process (Garrote et al, 1999). This force causes the chips to be torn apart and imparts large reductions in the fibre size and length. The effect of this treatment is important for subsequent downstream processes in the overall application, such as fractionation and hydrolysis. Also, the shearing forces cause fractures in the interior of the S2 layer of the cell wall, and further increase the surface area for chemicals and enzymes to attach during fractionation and hydrolysis, respectively (Toussaint et al, 1991; Kallavus and Gravitis, 1995). Reducing fibre size could lead to increased delignification and hydrolysis rates (Saddler et al, 1982). The reason for an increase in the hydrolysis rates would be the creation of more sites to which the enzymes can bind. A much more important factor which must be considered in a lignocellulosic substrate is the effect of fibre coarseness, which affects the total surface area of the fibres and may mean that a small fibre could still be quite thick and therefore have low cellulose accessibility. A secondary benefit of the steam explosion process is the hydrolysis and subsequent liberation of hemicellulose (Waymann et al, 1984). The hydrolysis of these components causes a substantial increase in the porosity of the substrate, which also benefits fractionation and hydrolysis by increasing the surface area available for chemical 22 interactions (Wong et al, 1988). There are many complex reactions occurring during the steam explosion process, some of which lead to the degradation of cellulose and hemicellulose, and the formation of furfural, from pentose sugars, and hydroxymethylfurfural, from hexose sugars (Larsson et al, 1999). Further breakdown of these compounds leads to the generation of acids such as levulinic acid and acetic acid (Larsson et al, 1999). These compounds have been shown to be inhibitory to the fermentative organisms. Unfortunately, the combined effects of these inhibitors do not account for all of the inhibition of fermentation at higher severities. Other inhibitors such as catechol and vanillin have been shown to be inhibitory to fermentation and the synergistic interaction of these two chemicals further increases inhibition. The effective removal of lignin would generally make the overall biomass to ethanol process economical. Unfortunately, only a small amount of lignin is removed during steam explosion, and the high temperature and pressures employed modify the lignin significantly. During the steam explosion process the high temperatures cause the hydrolysis of B-O-4 bonds and liquification of lignin, which permits the lignin to become elastic (Kallavus and Gravitis, 1995). In this state the lignin liberated from the cell wall where it agglomerates (due to the hydrophobic nature) on the surface of the wood fibre, as the temperature and pressure are reduced, the residual lignin condenses to form 5-5' biphenyl bonds (Michalowicz et al, 1991). This condensed lignin is much harder to remove than native lignin (Ramos et al, 1999), and severely inhibits enzymatic hydrolysis by limiting access to the cellulose. Lignin tends to accumulate in the areas that have been opened up during steam explosion, as well as inside the lumen, and substantially decreases the pore volume. The condensed lignin must be removed prior to 23 enzymatic hydrolysis, if effective solubilization of the cellulose is to occur. There are many techniques that have been shown to remove the lignaceous materials, such as alkali washing, peroxide, oxygen, and wet oxidation delignification. One way to avoid the problems associated with the condensed lignin to reduce its generation. Previous work has shown that chip size and moisture content have a substantial effect on the steam explosion conditions (Brownell et al, 1985; Ballesteros et al, 2000), suggesting that longer chips (8-12 mm) provided a much better recovery of hemicellulose and cellulose. During steam explosion chips are heated through the pores and lumen of the cells, generally in the longitudinal direction of the chips. The speed of heating, in the longitudinal direction, is many times faster than in the tangential or radial directions (Brownell et al, 1985). Since most chips have the largest dimension in the longitudinal direction it takes longer to heat larger chips. In contrast, shorter chips then require less heating time, and the steam pretreatment is more consistent and uniform throughout the entire chip (Ballesteros et al, 2000). Additionally, higher relative severity treatments may result in an increase in the generation of fermentation inhibitors, and such as sugar degradation products (furfural and hydroxymethylfurfural). Moisture content also plays an influential role in the severity of the pretreatment. Moisture content changes alter the ability of heat and chemicals to penetrate wood. The major effect of increased moisture content is a reduction in the heating rate of the chips due to the high specific heat of water, which therefore causes an uneven treatment for the lignocellulosic substrate (Brownell et al, 1985). An uneven treatment causes the degradation of the outer portion of the chips while at the same time the interior is not 24 affected by the treatment. An uneven treatment could lead to increased generation of inhibitors due-to the severe treatment that the exterior of the chip is subjected to, while at the same time not disrupting the interior of the substrate enough for efficient saccharification. Due to the concentration of lignin in the solid substrate post steam explosion a second step must be added to remove the lignin. This is difficult due to the recalcitrant nature of the lignin after steam explosion. Previous work has shown that peroxide delignification is able to remove the recalcitrant lignin and produce a substrate, which has a high hydrolysis rate and yield (Yang et al, 2002). 1.4 Delignification The benefits of the steam explosion process are the "opening up" of the cell walls through the removal of hemicellulose and the formation of fractures during explosion of the substrate. However, a side effect of the high temperatures used during steam explosion is the condensation of lignin. During the condensation reaction a dramatic change in the structure of lignin occurs (formation of 5-5' biphenyl structures) which produce a more recalcitrant form of lignin. To improve the rate of hydrolysis a substantial decrease in the lignin content of the substrate must occur (Yang et al, 2002), but it appears that the removal of condensed lignin must be facilitated by some means other than alkaline delignification (Schell et al, 1998). Currently, in the pulp and paper industry, mills are switching to more environmentally friendly forms of bleaching, such as peroxide, oxygen, and ozone. Delignification in the bioconversion process involves a modification of the bleaching protocol used in the mechanical pulp industry so that the end result is enhanced removal 25 of lignin, whereas bleaching (brightening) of mechanical pulps targets lignin chromophores and only brightens the residual mechanical pulp lignin. However, the bleaching of Kraft pulp targets the removal of lignin and can be considered a form of delignification. The main mechanism of peroxide brightening of mechanical pulps is facilitated by the perhydroxyl ion (HOO) attacking the lignin chromophores. In the bioconversion process the lignin must be removed to increase the cellulose accessibility, which means that the substrate cannot be bleached by the perhydroxyl ion, it must be delignified which occurs through the action of the superoxide and hydroxyl radicals. All of the treatments employed for the effective removal of lignin add a substantial cost to the overall process. Since pretreatment is proportionally one of the most expensive step in the bioconversion process, care must be taken to select a highly effective delignification process that is economical at the same time. Oxygen and wet oxidation are two new techniques that are gaining in popularity, as both of these techniques offer the advantage of a relatively cheap delignification agent that are also environmentally friendly. 1.4.1 Peroxide delignification Previous work has shown that a hot alkali peroxide technique is quite effective at the removal of recalcitrant softwood lignin (Yang et al, 2002). The optimization of this technique showed a substantial peroxide consumption, which caused the price of delignification to become excessive. The optimized peroxide delignification stage did not take into account the fact that stabilizers may be needed to reduce peroxide loadings and therefore the cost. 26 Peroxide bleaching, or delignification, has been one of the most studied techniques for a total chlorine free delignification process. Peroxide has proven that it can remove lignin under certain conditions. The optimum pKa of peroxide delignification is a pH of 11.5-11.6 (Gould, 1985). This illustrates that under optimized conditions the system must use a significant amount of alkaline. Unfortunately, alkaline extractions alone do not remove enough of the recalcitrant lignin (Schell et al, 1998; Ramos et al, 1999). Also, as pH increases the cost rises and the degradation of sugars reaches unacceptable levels. The combination of peroxide and alkali provide substantial delignification with a minimal loss of the carbohydrate components of the substrate. Previous work has shown that 1 % (w/w) peroxide is needed to achieve significant delignification to 10 % residual lignin content (Yang et al, 2001). These peroxide loadings are extremely high, suggesting that although the peroxide delignification is effective the cost is too high for the overall process to be viable. 1.4.2 Oxygen delignification The tendency in current pulp mills is to use ECF, or elemental chlorine free, bleaching regimes. This initiative has lead to a substantial amount of research into the use of alternative chemicals for the delignification of pulp. One of these systems is oxygen delignification. It has been repeatedly shown that oxygen delignification is capable of removing up to 67 % of the residual lignin (Bennington and Pinealt, 1999). The majority of these systems currently employ medium consistency (10-25 %) reaction conditions to effectively attain this level of lignin removal. Currently, mills that are using oxygen delignification treatments regularly demonstrate efficiencies in lignin removal ranging from 25-55 %. Even though many 27 mills use oxygen delignification, each has had to optimize the process for its substrate (Bennington and Pinealt, 1999), and has been suggested that many of the problems stem from improper use of the technology. The main problems point to improper chemical loadings, mixing, void space, and oxygen pressure. Other problems with oxygen delignification include the formation of complex lignin structures, such as the 5-5' biphenyl structure, during steam explosion that are inert towards oxidation (Akim et al, 2001). The reaction by which delignification occurs is quite rapid and is similar to the process which peroxide delignifies lignocellulosic residues. Oxygen in the presence of NaOH reacts to form -OH, and 0 2 - - . Once the hydroxyl radicals have been formed delignification occurs in the same way as peroxide delignification. The alkaline conditions solubilize the carboxylic acids formed by the reaction of superoxide and hydroxyl radicals. The use of oxygen delignification has been shown to improve hydrolysis rates by 2.1 times with the removal of 67 % of the residual lignin (Draude et al, 2001). 1.4.3 Wet oxidation Wet oxidation is a modification of the oxygen delignification technique. The advantage of the wet oxidation technique is the fact that the temperature of the pulping vessel is higher than oxygen delignification. Also, wet oxidation has the added advantage of other reactions. Lignin oxidation is the main reaction, but alkaline hydrolysis, and extraction of hemicellulose also occur (Schmidt and Thomsen, 1998). The oxidation of lignin occurs in a similar manner that of the oxygen delignification technique: oxygen is introduced to the pressure vessel and reacts with the alkali to form 28 superoxide and hydroxyl radicals. These radicals then penetrate the substrate and react with the lignin. Wet oxidation also has the benefit of supplemented Na2C03, which acts as an alkaline buffer in the wet oxidation process. This means that wet oxidation could be used with a modified pretreatment technique to maximize the recovery of the polysaccharide portion of the original wood. 1.5 Research objectives Recent work has shown that the bioconversion process has the potential to produce significant amounts of ethanol from lignocellulosic substrate. To date, the process has had limited success at removing and disintegrating the substrate for rapid enzymatic hydrolysis, while at the same time maintaining high recoveries of the individual components of the original substrate. Previous work has not looked at the effect of chip size and moisture content and their effect on the overall recovery of the presently optimized conditions for steam explosion (195°C, 4.5 min., and 4.5 % S O 2 ) . Moisture content has been shown to exert a large influence on the extent and severity of the pretreatment. Increased moisture content may also demonstrate increased recoveries and hydrolysis rates, and thereby increasing the effectiveness of the bioconversion process. Initial chip size may also have a significant effect on the severity of pretreatment, and may lead to a breakthrough on increasing the economics and production of glucose from the bioconversion process. With the addition of a refining step, which may liberate increased levels of carbohydrates into the prehydrolysate, the recovery of total sugars may be increased enough to finally ensure that the overall bioconversion process is economically viable. 29 Finally, in an attempt to improve the efficiency of delignification, and reduce the loadings of chemicals required for suitable delignification, the effect(s) of altering the currently employed delignification conditions (consistency, chelation, stabilization) were evaluated. The goal of this work is to increase the recovery of the Douglas-fir wood components throughout the bioconversion process, while concurrently decreasing the loading of chemicals required to efficiently remove recalcitrant lignin, and maintain hydrolysis rates. This thesis aims to investigate the effect of: 1) Chip size and moisture content on the relative severity of steam treatment, efficacy of fractionation and subsequent enzymatic hydrolysis. 2) Evaluate the possibility of refining post steam explosion and the effect of refining on the fractionation and hydrolysis stages. 3) Determine the effect of stabilizers and other delignification techniques on the efficacy of peroxide fractionation. 4) Optimize the best performing delignification technique based on the temperature and chemical loading while maintaining hydrolysis rates. The following chapters aim to elucidate how these changes will affect the overall bioconversion process and illustrate the impact of substrate conditions throughout the process. 30 2 MATERIALS AND METHODS 2.1 Pretreatment stage of the bioconversion process Pretreatment consists of the primary stage of the bioconversion process. There are two parts to pretreatment; steam explosion and fractionation. Steam explosion separates the volatile extractives from the substrate and solubilizes hemicellulose through autohydrolysis. Fractionation consists of two steps: a water wash to remove residual hemicellulose and any other water-soluble compounds, such as phenolics, and a peroxide delignification treatment to remove lignin. Steam explosion has been previously optimized for optimum yield of glucose and hemicellulose sugars after hydrolysis (Boussaid et ai, 2000). Previous work by Boussaid et al (2000) did not consider the effects of initial chip size, moisture content, or post steam explosion particle size reduction through refining, and thus the present work modified the pretreatment parameters to include these variables. In an attempt to evaluate these features, prior to steam explosion the chip size, and moisture content were adjusted to quantify their effect on SO2 catalyzed steam explosion. Following explosion, substrates were refined to determine the effect of particle size reduction on the fractionation, and subsequent hydrolysis of the substrate. The overall approach is outlined in Figure 3. The substrate, which showed the best performance, based on delignification efficiency, recovery after steam explosion, and glucose and hemicellulose yield after enzymatic hydrolysis, was later used for the optimization of the delignification conditions. This optimization included an evaluation of the effect of stabilizers and 31 • Douglas-fir wood chips Chip Fractionation Three Sizes (VF, M , VC) M oisture Content Acclimation (12 and 30%) Steam Explosion Water Soluble Sugar Recovery Non-Refined Refine Solid Substrate I Water Soluble Sugar Recovery Fractionation Lignin Recovery Hydrolysis Cellulose Recovery Figure 3. Flow diagram illustrating the generation of substrate from Douglas-fir wood chips for steam explosion, subsequent refining of the substrate, peroxide fractionation, and hydrolysis during the pretreatment stage of the bioconversion process 32 chelants to determine their effect on peroxide delignification. Concurrently, alternative delignification techniques were evaluated, including wet oxidation and oxygen. Finally, a three factorial matrix was employed to determine the best conditions for economic delignification of the steam-exploded substrate. 2.2 Substrate The substrate utilized for all experiments was obtained from a 150 year old Douglas-fir tree obtained from the UBC research forest. The chips were stored at -20°C and defrosted for 24 hours prior to use. 2.2.1 Chip fractionation Four different size classes were created using a chip size classifier designed by Macmillan Bloedel research group. The screen sizes employed were used to retain and collect chips of homogeneous size, were 5 x 5, 2 x 2, 1.5 x 1.5, and 1 x 1 cm square screens. Twenty kilograms of chips were separated into individual size classes so that each class contained about 2,000 g (dry) chips, which were required for steam explosion and refining. As well, knots, bark, and splinters were removed from the samples. Finally, a Wiley mill was used to generate a fifth chip size, by grinding the chips until they past through a 40-mesh screen on the Wiley mill. The chip sizes utilized for the majority of experiments were the very fine (passing the 40-mesh), medium (1.5 x 1.5) and very coarse (5 x 5) size classes (Figure 4). Once the size classes were created the chips were sealed in plastic bags to stop moisture loss, and stored at -20°C in a freezer. 33 (A) (C) Figure 4. Categories of Douglas-fir chips used in the steam explosion process. A) 40-mesh, B) 1.5 x 1.5, C) 5 x 5. 34 2.2.2 Chip conditioning The goal of conditioning the chips was to obtain two separate moisture contents: 12 % and 30 %. Chips were conditioned according to a modified Tappi T402om-93 method. All chips were conditioned at the same time in a climate-controlled room. The chips were spread out on separate plastic bags in a shaded coiner of the room to ensure that the drying treatment was even for all of the chips. The moisture content of the chips was tested every six hours to determine when equilibrium moisture content was reached, which generally took three days. Once, an equilibrium moisture content of 10-12 % was obtained the chips were then sealed in airtight plastic bags to prevent moisture gain or loss, and frozen at -20?C. To reach the second moisture content the chips were immersed in water. Due to the effect hysterisis has on the wetting and drying of chips the second moisture content was reached by drying the chips to thirty percent moisture content. Once the chips were saturated with water, they were conditioned by drying down to 30 % in a climate-controlled room with testing every six hours to determine when an equilibrium moisture content of 30 % had been reached. Again the chips were sealed in plastic bags and frozen at -20°C. 2.3 Steam explosion Six different substrates were generated for steam explosion, consisting of two different moisture contents, one at twelve percent (three chip sizes) and the other at thirty (three chip sizes). One kilogram of each of the substrates generated at all different conditions were prepared for steam explosion in a similar manner as described by Boussaid et al. (1997) (Boussaid et al., 1997). 35 After defrosting, 1 kg portions of each substrate were separated into individual 250 g (dry) sample bags and injected with a 4.5 % (w/w) loading of SO2 obtained from Praxair Canada, and sealed. The weight of the bags and substrate were monitored overnight to determine the volume of SO2 that was absorbed by the chips. Samples were then divided into 50 (dry) gram samples for explosion in the steam reactor. Samples were treated in the Stake Tech II steam reactor (Stake Technology, Norval, Ontario) for 4.5 minutes, at a temperature of 195°C. After treatment, the steam reactor was explosively decompressed and the substrate collected in a cyclone. Following the pretreatment of each separate feedstock, the substrates were collected and transferred to plastic buckets. The solid portion was then separated from the liquid portion (prehydrolysate) through filtration in a Buchner funnel. The solid portion was analyzed for Klason lignin, acid soluble lignin, and sugars. The liquid portion was analyzed for the hemicellulose sugars and inhibitors generated during steam explosion. A mass balance of the substrate was preformed on the substrate before and after the steam explosion process. 2.4 Post steam explosion refining After steam explosion, 500 (dry) grams of substrate (very fine, medium, very coarse chip size classes only for both moisture contents) was transferred to a mechanical refiner were the particle size of the substrate was reduced through refining. Prior to refining the moisture content was determined, and water was added to the substrate to bring all of the samples to the same consistency (12 %) prior to refining. This substrate was then run through a Sprout Waldron (Mundy, PA) mechanical refiner, preset with a plate gap of 0.007" and a plate speed of 2500 rpm. Samples were 36 refined for approximately 1 minute, and run through the refiner four times to ensure that each sample was treated equally and that the post refining particle size was consistent. The substrate was then filtered with the solid portion stored in the freezer in plastic bags, and a sample of the liquid portion stored for analysis. A sample of the solid portion was analyzed for Klason lignin, acid soluble lignin, and sugar content, while the liquid portion was analyzed for hemicellulose and cellulose derived sugars, and inhibitor production. A mass balance was also preformed to determine the total recovery of substrate after steam explosion and refining. 2.5 Water wash The twelve samples recovered from steam explosion and subsequent refining were washed with water. This was performed in a Buchner funnel attached to a four-litre vacuum Erlenmeyer flask. A volume of warm water twenty times the dry weight of the substrate was used to wash the substrate and remove water-soluble compounds, such as hemicellulose and water-soluble phenolic compounds. The water was added two litres at a time, and was constantly observed to ensure that the substrate did not dry out. Following the water wash the moisture content and mass balance were preformed. The solid substrate was collected in plastic bags and stored frozen for later use. A sample of the solid and liquid portion was analyzed for sugar, klason lignin, and acid soluble lignin. 2.6 Delignification Delignification was preformed on the twelve washed samples. Initially, the substrates were Hot Alkali Peroxide (HAP) treated as previously optimized (Yang et al, 2001) to 37 evaluate how the pretreatment variables affected the efficacy of peroxide delignification. Different forms of delignification were examined to determine the best way to remove the high levels of condensed lignin. 2.6.1 Hot alkali peroxide delignification Previous work has shown that the optimum conditions for a Hot Alkali Peroxide (HAP) treatment are 2 % (w/v) substrate, 1 % (w/v) peroxide, at 80°C for one hour, at an initial pH of 11.5. For the peroxide delignification 3 g (dry) substrate was immersed in 135 mL of nanopure water in a 500 mL Erlenmeyer flask to achieve a final consistency of 2 % (w/v). 5 g of a 30 % peroxide solution (Fisher) was added, to achieve the peroxide content of 1 % (w/v), and the pH was adjusted to 11.5, using NaOH (VWR). The reaction was allowed to proceed at 80°C using a polystat shaker bath operating at 150 rpm. A mass balance of the HAP washed substrate was then determined, so that the recovery could be determined. The HAP washed substrate was frozen in plastic bags for further experimentation. Prior to freezing a sample was analyzed for klason lignin, acid soluble lignin, and sugar composition. The substrate, which had the most effective lignin removal and the lowest residual lignin, during Hot Alkali Peroxide delignification experiments, was employed in the subsequent studies evaluating chelation, consistency and delignification optimization. 38 2.6.2 Chelation Four types of stabilizers/chelants were tested: ethylene-diamine-tetraaccetic acid (EDTA), diethylene-triamine-pentamethylene-phosphonic acid (DTMPA), sodium silicate, and a combination of sodium silicate and magnesium sulphate. EDTA (Sigma) was applied before the HAP treatment. A consistency of 2 % (w/v) was maintained for EDTA chelation by adding 12 g (dry) substrate to 568 mL of distilled water. Once the consistency was adjusted, a 38 % solution of the sodium salt of EDTA (Sigma) was applied to the substrate to obtain an EDTA concentration of 0.06 % (w/v) in solution, and the pH adjusted to 5 with H2SO4 (Fisher). The chelation was allowed to proceed at 50°C for 30 min., with constant stirring (150 rpm). After the thirty minute reaction the solution was washed with distilled water and prepared for HAP delignification. DTMPA chelation occurred during the HAP treatment. As mentioned earlier the HAP conditions are 2 % substrate, 1 % peroxide, 80°C for one hour. For the DTMPA chelation, a 0.2 % DTMPA loading was added to the standard HAP conditions. Once DTMPA was added to the substrate, water and peroxide were added and the pH adjusted to 11.5 using 50 % NaOH. Substrate for the DTMPA stabilized delignification came from two sources: (1) EDTA (Sigma) chelation and (2) directly from water washed steam exploded substrate. Sodium silicate (Fisher) was also tested as a stabilizer. In this situation the HAP delignification was preceded by the addition of sodium silicate (Fisher) at a concentration of 5 % (w/w), then the peroxide was added and pH was adjusted to 11.5. The reaction was maintained at 80°C and constantly stirred at 150 rpm for 1 hour. Sodium silicate 39 chelation was tested on two substrates: (1) EDTA chelated substrate and (2) substrate obtained directly from the water washing of the steam-exploded substrate. Magnesium sulphate (Fisher) was also added in combination with sodium silicate (Fisher) during peroxide stabilization. The magnesium sulphate (Fisher) was diluted to 2 % consistency (w/v) simultaneously with the 5 % (w/w) loading of sodium silicate (Fisher). Once the magnesium sulphate (Fisher) and sodium silicate (Fisher) were in solution, a 1 % (w/v) loading of peroxide was added and the pH adjusted to 11.5. The temperature was maintained at 80°C with continuous shaking at 150 rpm for 1 hour. Substrate was obtained for the combined sodium silicate and magnesium sulphate stabilized delignification directly from the water washed steam-exploded substrate and from the EDTA chelation prior to delignification. Each of the substrates was collected after delignification and stored at -20°C in sealed plastic bags. The performance of the delignification was assessed and based on the recovery of cellulose and the efficiency of lignin removal. Cellulose and lignin recovery were calculated by analyzing the solids recovered for sugar, klason lignin, and acid soluble lignin. 2.6.3 Peroxide delignification consistency The effect of substrate consistency on delignification was assessed by adjusting the percentage of solids from 2 % to 12 % (w/v). The peroxide conditions employed in this experiment were the original conditions (50 %, 80°C, and manual mixing). However, at higher consistencies, the reactions were conducted in plastic bags and mixed by hand every 5 minutes for the entire one-hour reaction. The actual substrate concentrations employed were 2, 5, 7, 10, and 12 % (w/v). As the substrate loadings increased the 40 peroxide loadings increased from 1, 2.5, 3.5, 5, and 6 % to maintain 50 % peroxide to substrate ratio. The substrate was dried after delignification and analysed for sugar content, klason lignin, and acid soluble lignin. 2.6.4 Peroxide loading Variation in peroxide loading (0.2, 0.3, 0.4, 0.5, 0.6, 0.7, and 0.8 % (w/v) peroxide loadings) in combination with DTMPA stabilization and prior EDTA (Sigma) chelation, were investigated. In this series of experiments the conditions of peroxide delignification were maintained (2 % consistency, 80°C, 1 hour, 150 rpm shaking) as in the optimized delignification conditions (Yang et al, 2002), however EDTA (Sigma) chelated substrate was added to the solution. A DTMPA loading of 0.2 % (w/w) was added to the HAP delignification. Prior to the DTMPA chelated HAP delignification, EDTA chelation was preformed using a previously described protocol (Section 2.6.2). After, the HAP washed a mass balance and klason lignin was preformed so that the substrate could be enzymatically hydrolysed. 2.6.5 Factorial matrix The optimum conditions for delignification were investigated by factorial matrix, using (13 samples) temperature, peroxide loading and DTMPA concentration as design variables. Three of the samples were designated as a center point for the factorial design. The peroxide loading was adjusted from 0.3 to 1.10 %; the temperature was increased from 70 to 90°C, and the DTMPA loading was varied from 0.15 to 0.25 %. The substrate used for this experiment was derived from the very coarse chips prior to steam explosion that had been refined, water washed, and chelated prior to this series of experiments. 41 The samples were analyzed for recovery and therefore mass balance, and the composition of the substrate recovered. The composition of the substrate was quantified by klason lignin analysis, acid soluble lignin, and sugar analysis. After the generation of a three dimensional matrix, four conditions were determined to be the optimum for each of the three dimensional matrices. The samples were then delignified according to the conditions outlined by the matrix and subjected to enzymatic hydrolysis. 2.6.6 Oxygen delignification Oxygen delignification was preformed at 3 % (w/v) consistency (15 g dry substrate in 449.1 g of nanopure water). A 3 % (w/w) loading of NaOH was added to the substrate. This solution was then heated to 125°C, once the solution reached temperature, and 100 psi of oxygen (Praxair Canada) was added to the mixture and maintained at temperature for 30 min. After 30 min., the pressure vessel was removed and cooled rapidly in a water bath to stop the delignification reaction. Once cooled the pulp was filtered and washed with a large volume of distilled water. Recovered substrate was analyzed for the sugar composition and klason lignin content. 2.6.7 Wet oxidation Pulping was done in a 0.65 % (w/v) solution of Na2C03 (J.T. Baker). The consistency of the pulp was 6.5 % (w/v) during cooking. The pH of this solution was adjusted to 11 with a solution of NaOH (VWR). The pulping solution was heated to 195°C and 175-psi oxygen (Praxair Canada) was applied to the pulping solution. The pulp was maintained at 195°C for 30 minutes, and then cooled in a water bath and filtered 42 with large volumes of distilled water. Samples were analysed for the sugar and klason lignin analysis. 2.6.8 Alkali wash Alkali washing was preformed on 10 g (dry) pulp samples, which were obtained from wet oxidation or oxygen delignification. The pulps were dissolved in a solution to a consistency of 5 % and the pH adjusted to 11 with NaOH (VWR, West Chester, PA). This solution was heated to 80°C and held at this temperature for 2 hours, and then filtered and washed with excess water. 2.7 Hydrolysis All substrates prepared from the various delignification reactions and unbleached material were subjected to enzymatic hydrolysis Enzymatic hydrolysis was preformed in 50 mM acetate buffer solution (pH of 4.8) at 2 % consistency in a 100 mL Erlenmeyer flask. Enzymes were added to the solutions based on the predetermined sugar concentration in the substrate (RJ/g cellulose). The enzymes were loaded at 20 FPU with a FPU to CBU ratio of 2:1. This involved adding on average 0.17 mL of celluclast (Novozymes North America) and 0.069 mL of Novozym-188 (Novozymes NA), which had activities of 111 FPU/mL and 572 CBU/mL, respectively. The enzyme, substrate, and buffer mixture was incubated at 45 °C in a shaker bath with constant mixing (150 rpm), for 0-72 hours. 43 2.8 Substrate composition 2.8.1 Acid soluble and insoluble lignin The substrate used for klason analysis was initially ground to pass a 40-mesh size screen in a Wiley mill. A sample of 0.2 g (dry) was then weighed into a reaction flask, to which 3 mL of 72 % sulphuric acid (Fisher) was added and stirred vigorously. The sample was then stirred every 10 min. for two hours, after which time the sample was diluted with 112 mL of nanopure water. This solution was then transferred to an autoclave were it was heated to 121°C for one hour. After one hour the sample was removed and cooled. The sample was then filtered through a medium coarseness sintered glass crucible that had been oven dried and pre-weighed, and washed with 200 mL of nanopure water to determine acid insoluble (klason) lignin gravimetrically. A sample of the filtrate was removed and stored in a falcon tube for analysis of carbohydrates and acid soluble lignin Acid soluble lignin was calculated according to Tappi method UM-250, in which the lignin concentration in solution is calculated using the absorbance at 205 nm, as measured on a Milton Roy spectronic spectrophotometer. The lignin content is calculated using an expression of Beer's law Lignin=absorbance (A) / b (light path in cm) x a (absorptivity in 1/gcm) 2.8.2 Sugar analysis The sugar concentration was determined by HPLC analysis on a Dionex DX-600 HPLC equipped with a Carbopac PA1 column and a pulsed electrochemical detector. The dionex was conditioned with 0.2 M NaOH (VWR) and the mobile phase was 44 nanopure water at flow rate of 1 mL/min. Integration of the peak area determined the concentration of the sugars based on arabinose, galactose, glucose, mannose, and xylose, standards, which were run concurrently. 2.8.3 Monomer/ oligomer analysis Samples for the monomer and oligomer analysis were obtained from the prehydrolysate after steam explosion. The sugar content of the initial samples was analyzed by HPLC (previous section). A second sample was hydrolysed in a 4 % solution of sulphuric acid to hydrolyse any oligomers present in the prehydrolysate. This hydrolysis occurred in an autoclave at 121°C for one hour. Once the samples had cooled they were analysed on the HPLC. The difference in sugar concentration was used to determine the concentration of monomer to oligomeric sugars. 2.8.4 Glucose analysis Aliquots from the enzymatic hydrolysis were removed at set time intervals (0, 1, 3, 5, 7, 10, 12, 24, and 48 hours). At each time points 0.5 mL of hydrolysate was removed, and boiled for 10 min to denature the enzymes. The samples were then analyzed on a Dionex DX-600 HPLC, as previously described. 2.8.5 Furan analysis Samples of prehydrolysate collected from the steam explosion pretreatment of the wood chips were obtained and analysed on a HPLC. The HPLC (Dionex DX-300) equipped with a PAD (pulsed amperometric detector) and a PA1 column. The samples were eluted with 0.25 mM NaOH. The samples were calibrated with 5-hydroxymethylfurfural (Sigma) using fucose (Sigma) as an internal standard. 45 Furfural was integrated using a similar technique, but employed a Dionex MAI column and was eluted with 250 mM NaOH at a flow rate of 0.4 mL/min which had been equilibrated with 1.0 M NaOH. The calibration standards contained furfural (Sigma) 2.8.6 Mass balance determination Percent recovery was determined by dividing the total amount of dry substrate entered into steam explosion and fractionation by the dry weight of substrate recovered. The dry weight before and after each process was determined on three individual samples of each substrate. The moisture content was then assumed to be true for the total substrate. To obtain the mass balance of cellulose, hemicellulose, or lignin the weight of each sample was multiplied by the concentration of each component determined by klason and sugar analysis. By tracking the volume of each component of the wood the loss of cellulose, hemicellulose, or lignin was estimated. The addition of the individual components generated the mass balance for each individual step. The mass balances for all substrates are located in appendices A through E at the end of this thesis. 2.8.7 Statistical analysis Statistical analysis was preformed in excel, using the regression analysis function. To generate the linear regression equation the parameters were used which had a p-value of less than 0.05. A 95 % confidence limit was imposed on the accuracy of the regression equation. 46 3 C H I P S I Z E 3.1 Introduction Previous work by Boussaid et al. (2000) has shown that medium severity (195°C, 4.5 minutes, and 4.5 % SO2) SC>2-catalyzed steam explosion is an effective pretreatment regime for Douglas-fir wood chips. These conditions have shown recoveries greater than 88 % of the original material, with 76 % recovery in the solid portion. The process, which has been optimized, is a balance between good hydrolyzability of the cellulose and high recovery of hemicellulose-derived sugars. It is known that as severity is increased the hydrolyzability of cellulose is improved, while the hemicellulose sugars are degraded. Additionally, it has been suggested that for the bioconversion process to be economically viable as much of the hemicellulose and cellulose derived sugars (Wu et al., 1999) must be recovered in a usable form, which means that the severity of the steam pretreatment must be lowered to ensure high recoveries. However, as the severity is reduced an additional step, delignification, must be added to the bioconversion process to facilitate the full recovery of the cellulosic sugar stream. Delignification adds another step to the pretreatment of the wood. The pretreatment of wood for the bioconversion process has been suggested to occupy 60 to 80 % of the expense of producing ethanol (Gregg and Saddler, 1995). The cost of delignification is largely due to the high chemical loadings that are required to remove the condensed and recalcitrant lignin formed during the steam explosion of softwoods. It has been suggested that the condensation of lignin could occur due to irregularities in the treatment of the chips caused by substrate differences, such as moisture content and chip size (Foody, 1982; Ballesteros et al., 2000). Similarly, the high chemical costs associated with 47 delignification may be related to the delignification process itself and may be improved through post steam explosion modification (particle size), different delignification techniques, chelation of the substrate prior to delignification, and stabilization of the delignification agents. Currently, the chip size entering steam pretreatment ranges from sawdust to chips greater than 5 centimetres squared. Chip size has been shown to change the overall severity of steam pretreatment due to the rate at which the chips heat up during the steam explosion processes (Ballesteros et al, 2000). Generally, it is believed that the heating of chips occurs tangentially, with the heat travelling down the pore network within the chips. As chip size increases the steam takes longer to penetrate the pore network due to aspirated pits and the inherent complexity of the pore network. This means that the innermost portion of the chips is not treated the same as the exterior, and that the relative severity of the pretreatment of the larger chips may be reduced and non-uniform. In the current study the effects of modified chip size (30 % moisture content) was examined by analysing the sugar and lignin concentration in the solid and liquid portions recovered from the steam explosion process. Additionally, the efficacy of the fractionation process and chemicals was examined with the goal of removing the maximum amount of lignin using the current delignification technique. The generated material was then assessed for its hydrolability using the enzyme loadings outlined previously (Section 2.7). Finally, the total recovered carbohydrate, were compared to determine the effect of chip size on the steam explosion process. 48 3.2 Steam explosion sugar recovery It is apparent that chip size (30 % moisture content) has a dramatic effect on the efficacy of the steam explosion process. Changes in the recovery of sugars, lignin, and furans in the solid and liquid portions were observed. The starting chip size affected the overall concentration of sugars recovered in the prehydrolysate (Figure 5). The observed increases in prehydrolysate sugar concentration can likely be accounted for by concurrent decreases in the decomposition rates of the sugars hydrolysed during steam explosion, as has been suggested by Wu et al. (1999). Furthermore, the elevated sugar concentration agrees with the theory that increases in initial substrate chip size leads to a reduction in the severity of steam explosion; as chip size increases the sugar concentration in the prehydrolysate increases due to a decrease in the relative severity of the steam pretreatment of the larger chips. A closer inspection of the individual sugars comprising the prehydrolysate, demonstrate a very interesting trend (Figure 6); the concentration of all neutral wood sugars do not all follow a similar trend, as the concentration of glucose and mannose increased, while arabinose and xylose decreased and the concentration of galactose was stable. Clearly, glucomannan-derived hemicelluloses are accumulating and not being degraded as the chip size was increased, which could be the result of a decrease in relative severity. Previous work has shown that a decrease in severity results in an increase in hemicellulose sugar recovery (Boussaid et al, 1998). In contrast, these findings suggest that arabinoxylan liberation is limited by the increasing size of the initial substrate. 49 22 Very Fine Fine Medium Coarse Very Coarse Figure 5. Sugar content recovered (g) in the water-soluble fraction after steam explosion of wood chips of different initial sizes at 30 % moisture content (error bars represent the range based on the analysis of three samples recovered from the liquid portion). 50 Very Fine Fine Medium Coarse Very Coarse Figure 6. Concentration of individual neutral sugars recovered in the water-soluble fraction after steam explosion of wood chips of different initial particle size at 30 % moisture content (error bars represent the range based on the analysis of three samples recovered from the liquid portion). 51 The concentration of oligomeric (Figure 7) and monomeric (Figure 8) sugars in the prehydrolysate followed a similar trend to that observed with the total sugar liberated into the prehydrolysate. The concentration of both oligomeric and monomeric sugars demonstrates higher levels of solubilization with increasing chip size. These results are likely due to a combination of lower rates of decomposition and incomplete autohydrolysis of the sugars. Therefore, as the chip size increases the relative severity decreases, and the recovery of monomeric and oligomeric sugars increases in the prehydrolysate due primarily to decreased carbohydrate decomposition (Shevchenko et al, 2000). Another indication of the buffering ability of the larger chips is illustrated by the concentration of furans present in the liquid portion recovered after steam explosion, as the sugar degradation products, such as furfural and hydroxymethylfurfural, showed a substantial decrease as the chip size increased (Figure 9). The sugar degradation products generated from glucose and mannose degradation (hydroxymethylfurfural) had the lowest concentration compared to furfural, which is generated from arabinose and xylose, indicating that they were less susceptible to decomposition. These results support the hypothesis that relative severity is reduced by the size of the substrate entering into the steam explosion process. Solid substrate recovery from the steam explosion process also showed an upward trend as the size of the substrate was increased (Figure 10), and support previous findings that suggest that severity is linked to the recovery of solid substrate (Boussaid et al, 2000). Therefore, since recovery increases as chip size increases, the relative severity of the treatment of the chips must, in some way, decrease over the same period. 52 Arabinose V777A Galactose Very Fine Fine Medium Coarse Very Coarse Figure 7. Oligomeric sugars recovered in the water-soluble fraction after steam explosion of woods chips of different initial size at 30 % moisture content (error bars represent the range based on the analysis of three samples recovered from the liquid portion). 5 3 Very Fine Fine Medium Coarse Very Fine Figure 8. Monomeric sugar recovered in the water-soluble fraction after steam explosion of wood chips with different initial size at 30 % moisture content (error bars represent range based on the analysis of three samples recovered from the liquid portion). 54 "O o o j> c o 3.0 2.5 k £ 2.0 CD O O c o ' ro i_ -t—• c CD O c o Q o 1.5 1.0 0.5 £ 0.0 Furfural HMF rn Very Fine Fine Medium Coarse Very Coarse Figure 9. Furans (furfural and hydroxymethylfurfural) recovered in the water-soluble fraction after steam explosion of chips with different initial sizes at 30 % moisture content (error bars represent the range based on the analysis of three mass balances). 55 The proportion of lignin in the substrate also showed trends indicating a decrease in severity due to increased initial chip size. Total lignin concentration in the solid substrate increased as the chip size increased, which is likely due to a decrease in the decomposition and solubilization of lignin during steam treatment (Figure 11). An increase in lignin content has previously been associated with reductions in the severity of steam treatments (Schwald et al, 1989; Shevchenko et al, 1999), therefore it is believed that the relative severity of steam treatment is reduced as the chip size increases, and in the smaller chips more of the lignin is solubilized. 3.3 Fractionation The effect of size on the relative severity of steam treatments was evident in the efficiency of peroxide fractionation of the steam-exploded substrate. Lignin was much easier to remove from the largest chip sizes where the largest decrease in relative severity appeared to occur (Figure 12). It has been suggested that the ease of lignin removal is limited by two main reasons: condensation and fibre size (Shevchenko et al, 1999). As the relative severity increases, the extent of lignin modification also increases. Therefore, the main reason for the improved efficacy of lignin removal in the largest chip size was likely a decrease in the chemical modification occurring in the lignin through condensation reactions. As lignin condenses it has been shown to restrict the swelling of the fibres (Schwald et al, 1989; Toussaint et al, 1991; Kallavus and Gravitis, 1995; Shevchenko et al, 1999), and therefore blocks pores and cracks within the cell wall and restricts the removal of lignin from the lignocellulosic matrix. Additionally, the melting and chemical condensation of lignin lowers the reactivity of the lignin, leaving a product that will not react with the chemicals used during removal (Hemmingson, 1987). 56 Figure 10. Solids recovered from the steam explosion process of wood chips with different initial moisture contents at 30 % moisture content (error bars represent the range based on the analysis of three mass balances). 57 Figure 11. Concentration of residual lignin in the solids fraction recovered after steam explosion of wood chips with different initial sizes at 30 % moisture content (error bars represent the range based on the results of three klason lignin analyses). 58 Very Fine Fine Medium Coarse Very Coarse Figure 12. Total residual lignin retained in the solid fraction after hot alkali peroxide (1 % peroxide, 2 % substrate, 195°C, pH 11.5) fractionation (error bars represent the range based on the results of three klason lignin analyses) 59 It has previously been shown (Converse et al, 1990; Boussaid et al, 2000) that increased severity generally decreases the overall size of residual solids after steam explosion, this would imply that the very fine initial chips size would have the smallest starting fibres and residual fibres after explosion, and therefore the greatest accessibility of lignin for peroxide. Unfortunately, the reactivity and restricted swelling due to steam explosion appear to offset the effects of fibre size changes resulting from pre-selection of smaller chips. Although this hot alkali peroxide fractionation is selective for lignin, it also removes a portion of the carbohydrates. For example, the remaining hemicellulose-derived sugars were removed from all the chip sizes, and the residual glucose was only slightly reduced (ranging from 0.4 - 4 % of the total glucose content of the chips). 3.4 Hydrolysis While increasing the initial chip size has been shown to have a desirable effect on the recovery of glucose during steam explosion, and subsequently from fractionation, it also demonstrates an increase in the rate and yield of enzymatic hydrolysis (Figure 13). The improved hydrolysis is largely due to improved removal of lignin in the solids residue. Lignin content has a two-fold effect on the rate of hydrolysis: the first effect is a reduction in the accessibility of the enzymes to the cellulose and a decrease in the swelling of the substrate (Mooney et al, 1998). As mentioned earlier, lignin condenses on and in the pores and cracks generated during the steam explosion of the substrate. Since the lignin was not effectively removed from all substrates (Figure 12), the residual lignin can impede the accessibility of the cellulose substrate to the hydrolytic enzymes. The second effect of lignin on hydrolysis rates is the inactivation of the enzymes. 60 Enzymes irreversibly bind to the lignin (Converse et al, 1990), and if the elevated concentrations of the lignin are retained on the fibres, the enzymes will bind to the lignin and limit their action on the cellulosic polymer. Since there are fewer enzymes available for hydrolysis, the rate of the hydrolysis is depressed and the total yield of hydrolysis is reduced (Figure 13). Generally, the rate of hydrolysis for the largest chips was the greatest (Figure 13), and the overall recovery of glucose, hemicellulose, and lignin was the highest for the very coarse initial substrate (Table 4). Decreasing recovery of the hemicellulose and glucose derived components could be due to the degradation and therefore loss of this substrate from the process. Meanwhile the decrease in lignin recovery as chip size decreased could be the solubilization of the lignin, which may be recovered in the liquid portion after steam explosion, and degradation of the lignin into products such as catechol and vanillin. Table 4. Recovery after steam explosion, fractionation and hydrolysis of the original components of the chips entered into steam explosion (% of the original component weight). ^ Recovery (%) Substrate Hemicellulose Glucose Lignin Very Fine 49.5 56.0 79.1 Fine 55.6 70.0 94.8 Medium 55.2 74.8 94.8 Coarse 57.7 73.0 92.0 Very Coarse 56.7 84.0 100.0 61 100 Time (Hours) Figure 13. Rate and yield of enzymatic hydrolysis of the cellulosic fractions resulting from the steam explosion of different wood chip classes followed by HAP (error bars represent the range based on the results of two separate hydrolyses). 62 3.5 Conclusion Clearly chip size affects every part of the bioconversion process from steam explosion to the hydrolysis of the steam exploded fractionated substrate. Increased recovery of glucose, hemicellulose derived sugars and lignin occurs when the chip size is increased. The reason for this is most likely changes in the relative severity of steam treatment. As the chip size increases heat takes longer to penetrate the chips, and in the case of the largest chips the limited heating time reduced the calculated severity of the treatment that the core of the chip is exposed too. With optimization of the steam explosion conditions for the very fine material, results similar to those obtained with the largest chip sizes may be achievable. In fact a steam treatment optimized for the very fine chips may have better results due to more even heating of the chips, which may result in a decrease in the degradation of sugars and therefore higher recovery of sugars. 63 4 MOISTURE CONTENT 4.1 Introduction Like chip size, moisture content also influences the rate and extent of heat penetrating chips, and therefore affects the overall severity of the pretreatment. It has been shown that heat is transferred into chips faster in the tangential direction due to the porous nature of the substrate (Brownell et al, 1985), and is dependant on the inherent porosity of the wood, complexity of the pore network, and most importantly on the moisture content of the substrate. Moisture content, itself, has three major effects on the treatment of chips in the bioconversion process. The first is due to the two types of heat transfer that occur as the substrate is heated in the steam explosion reactor: conduction and convection. Convection is the faster of the two, occurring through the movement of currents in the air. The transfer of heat through convection requires open pore space and connections between the lumen of the woody substrate (Brownell et al, 1985). As the moisture content increases the pore volume will not change until the substrate reaches the fibre saturation point, around 25-30 % moisture content, and the fibres become saturated and water (free water) is loaded into the lumen of the fibres. The storage of water in the lumen decreases the space that convection has to transport heat into the chips (Himmel, 1994). As the convection of heat into the chips slows, heat must be transferred by conduction. Conduction transports heat through solid materials and is much slower than heat transfer via convection. Therefore, the chips with higher initial moisture contents are not subject to as severe heating as lower moisture content chips (Himmel, 1994). As the fibres dry the pits between the cells become aspirated and seal against the transport of moisture between the cells. 64 The second effect of an increase in the moisture content is an increase in the specific heat. Water has a significantly higher specific heat than wood so therefore an increase in the moisture content from 12 % to 30 % will result in a substantial increase in the specific heat and therefore the amount of steam required to heat the substrate will increase and the rate of heating will decrease. The final affect of an increase in the moisture content of the substrate prior to steam explosion is an increase in the effectiveness of the S O 2 pretreatment of the chips (Brownell et al, 1985). S O 2 requires water to react with, through disappropriation and oxidation, in order to form sulphuric acid. The sulphuric acid pretreatment improves the efficacy of steam explosion and fractionation, and therefore improved S O 2 treatment would enhance the overall pretreatment of the wood chips. 4.2 Steam explosion The effect of an increase in the moisture content from 12 % to 30 %, was a substantial decrease in the amount of hemicellulose derived sugars recovered in the prehydrolysate (Figure 14). The decreased concentration of sugar is related to a decrease in the severity of treatment with an increase in the moisture content, similar to that which occurred as a result of the increased initial chip size. The higher moisture content buffers against sugar solubilization, decomposition and autohydrolysis due to slower rate of heating in the interior of the chips (Foody, 1982). Similar trends were observed in the water-soluble sugars derived from substrate acclimated to both initial moisture contents (12 and 30 %), thereby, the total concentration of sugars solubilized increased as with increasing chip size. However, more sugar was liberated from chips with lower initial moisture content (Figure 14). 65 Figure 14. Total carbohydrates recovered in the water-soluble fraction recovered from steam explosion of chips of different initial size (Very fine, Medium, Very coarse) and 12 % and 30 % moisture contents (error bars represent the range based on the analysis of three samples recovered from the liquid portion). 66 Generally, the increased levels of total sugar in the prehydrolysate are related to the liberation of glucose and mannose, and to a lesser extent galactose (Figure 15). The observed change in the concentration of the sugar in the prehydrolysate is likely due to increased autohydrolysis of the lower moisture content chips, which is a result of the higher relative severity of the treatment because of faster temperature rise in the drier substrate. Overall the same trends are observed in the 12 and 30 % moisture content substrate, whereby the degradation of sugars decreases as chip size increases. The 12 % initial moisture content substrate also has a substantially higher volume of monomelic sugar in the prehydrolysate, which is on average 6-7g/100g of original material higher than the 30 % moisture content substrate. The majority of this increase can be attributed to increased glucose and mannose solubilized. Similarly, the increased relative severity of the treatment hydrolysed more of the glucose and mannose, while concurrently degrading the other sugars (arabinose and xylose) present in the prehydrolysate. The concentration of furans again illustrates that reduced severity due to increased moisture content protects against the degradation of sugars released into the prehydrolysate. Furan generation was higher for the 12 % moisture content substrate. The fine initial chip size substrate, which has the highest furan concentration for both moisture contents, had 0.55g/100g greater concentration of furfural and a 0.95g/100g greater concentration of hydroxymethylfurfural concentration. This indicates that the increase in the moisture content restricts the quantity of sugar, released into the liquid portion and then subsequently degraded. 67 Very Fine Medium Very Coarse 14 12 o o ro to XJ CD Q. o o •Bin Arabinose Galactose I I Glucose (B) Xylose •F^xl Mannose J i 1 I Very Fine Medium Very Coarse Figure 15. Concentration of carbohydrates recovered in the water-soluble fraction derived from wood chips with different initial size at (A) 12 % and (B) 30 % moisture content (error bars represent that range based on the analysis of three samples recovered from the liquid portion). 68 Solid substrate recovered also shows a trend towards the increased severity of pretreatment of the lower initial moisture content substrate. Solid recovery was decreased substantially by lowering of the moisture content (Figure 16) for all chip sizes. Previous work has shown that a drop in the solid recovery, from 77 % to 61 % of the original material, is due to an increase in the severity of the treatment from low (175°C, 4.5 % S0 2, 7.5 min.) too high severity (215°C, 2.38 % S0 2, 2.38 min.) (Boussaid et al., 2000). The recovery of this substrate also supports this theory, in that as the moisture content of the initial solid substrate is decreased the treatment severity increases. Additionally, the recovery of glucose was significantly lower for the 12 % moisture content substrate, likely due to an increase in the autohydrolysis of cellulose and solubilization of lignin occurring during steam explosion. A similar trend was also observed in hemicellulose-derived sugars. A higher proportion of lignin was recovered from the 30 % moisture content substrate (Figure 17), which could be a function of decreased lignin solubilization during a less severe pretreatment for the higher moisture content chips. Boussaid et al. (1998) have shown previously that as treatment severity is increased the solubilization of lignin is increased. Although there is increased solubilization of lignin in the 12 % moisture content substrate, the recalcitrant lignin remaining on the solid substrate is also harder to remove than the lignin remaining on the substrate with higher initial moisture content. 4.3 Fractionation The substrate with lower initial moisture content demonstrated decreased total lignin content compared to higher (30 %) moisture content prior to the fractionation process although the proportion of lignin in the substrate was higher for the 12 % moisture 69 Figure 16. Total recovery of solid substrate after steam explosion of chips of different initial size and moisture contents (error bars represent the range based on the results of three mass balances). 70 Figure 17. Total lignin retained in the solid portion after steam explosion of substrate derived from different chip sizes and moisture contents (error bars represent the range based on the results of three klason lignin analyses). 71 content substrate, due to lower solid recovery after steam explosion (Figure 17). However, fractionation revealed much better results with the higher initial moisture content substrate, except in the case of the medium chip size. The results of the medium chip size are the opposite of the other two samples, which could be due to an overestimation of the lignin content of the medium 30 % moisture content substrate. The proportion of lignin remaining in the best performing chip size (very large) in the higher initial moisture content had 0.5-1 % less lignin than the that remaining in the lower initial moisture content substrate (Figure 18). It appears that the steam explosion conditions for the lower initial moisture content substrate were much more severe than the higher initial moisture content, as illustrated by the increased recalcitrance of the residual lignin. Previous work has shown that an increase in the recalcitrance of the lignin can be attributed to increased cross-linking of the lignin moieties and therefore a decrease in the reactivity of the lignin (Shevchenko et al, 1999). 4.4 Hydrolysis The rate and extent of enzymatic hydrolysis showed similar trends at all three-chip sizes evaluated (Figure 19). Furthermore, the two categories of acclimated chips (12 and 30 % moisture content) showed a trend towards increasing chip sizes having the highest rate and yield during hydrolysis. The coarse chips for both of the moisture contents demonstrated similarly high hydrolysis rates and yields (Figure 20). In general, decreasing the initial moisture content shows the same effect as decreasing chip size on hydrolysis rate; the hydrolysis rate of the lower initial moisture content (12 %) substrate was inhibited. The reason for an inhibited hydrolysis rate and yield is likely related 72 Figure 18. Residual lignin after steam explosion and fractionation (HAP) of wood chips derived from different chip sizes and moisture contents (error bars represent the range based on the results of three klason lignin analyses). 7 3 Figure 19. Hydrolysis (20 FPU/g, pH 4.8, 50 mM sodium acetate) rates and yield of substrate derived from steam explosion and fractionation of wood chips with different size at (A) 12 % and (B) 30 % initial moisture content (error bars represent the range based on the results of two separate hydrolyses). 74 to the recalcitrance of the lignin present in recovered solid steam-exploded substrate. The initial moisture present in the substrate appears to have a limit the severity of pretreatment and decreased the modification of lignin present during steam explosion, and therefore even though more lignin is solubilized during the steam treatment in the lower initial moisture content substrate there is less lignin modification in the higher moisture content substrate. The residual lignin present in the substrate has a two-fold effect on the hydrolysis rates: increased lignin impedes access to the cellulosic fraction and increases the adsorption of the enzymes onto the lignin. Basically lignin acts as a shield preventing enzymes from accessing the substrate, and therefore decreasing the rate and yield of hydrolysis. Overall the yield from hydrolysis was higher for the higher initial moisture content (30 %) substrate based on higher hydrolysis yield (89 % versus 93 % for Very coarse 12 % and 30 % moisture content chips respectively). The higher recovery of solid cellulose after steam explosion meant that more cellulose was available for hydrolysis, and that a 93 % yield from hydrolysis would yield overall more glucose to be used downstream in the bioconversion process for fermentation (Table 5). Table 5. Recovery after steam explosion, fractionation and hydrolysis of the original components of the chips entered into steam explosion (g/lOOg of original wood) 30 % Moisture Content Hemicellulose Glucose Lignin Very Fine 11.4 26.1 14.2 Medium 12.7 34.9 6.2 Very Coarse 13.0 39.3 4.8 12 % Moisture Content Hemicellulose Glucose Lignin Very Fine 12.8 20.2 14.8 Medium 14.4 35.3 4.8 Very Coarse 12.9 33.9 3.9 75 100 1 2 % Moisture Content 30 % Moisture Content i i i i i 30 40 50 Time (hours) 60 70 Figure 20. Comparison of hydrolysis rates from chips derived from very coarse chips with different moisture contents (error bars represent the range based on the results of two separate hydrolyses). 76 Table 6. Recovery after steam explosion, fractionation and hydrolysis of original components of the chips entered into steam explosion (% of original weight) Percent recovery 12 % Moisture Content Hemicellulose Glucose Lignin Very Fine 55.7 48.2 78.6 Medium 62.5 75.6 82.6 Very Coarse 56.1 72.5 89.4 30 % Moisture Content Hemicellulose Glucose Lignin Very Fine 49.5 56.0 79.1 Medium 55.2 74.8 94.8 Very Coarse 56.7 84.1 100.0 4.5 Conclusion This work has shown that increased chip size and moisture content result in a decrease in the relative severity of the steam explosion process. The previous chapter showed that chip size influences the recovery of substrate from steam explosion, and the efficacy of fractionation chemicals to remove lignin and therefore the hydrolysis rates. These trends carried over to the chips with lower moisture content. The larger chip sizes had a substantial improvement in recovery, at lower moisture content, due to a decrease in the relative severity of the treatment which the chips were exposed to. The decrease in relative severity is attributed to a decrease in the rate of heating of the chips, and therefore a decrease in the overall treatment severity. Previous work by Boussaid et al. (2000) showed that as the severity of steam explosion is reduced the recovery in the solid portion increases. The increase in recovery is caused by an increase in the amount of lignin, cellulose, and hemicellulose remaining 77 in a water insoluble form after the steam explosion event, which is a function of reduced autohydrolysis and solubilization of the components. Lignin, which is recovered, from the larger chip sizes at 12 % moisture content has been shown to be easier to remove than from the smaller chip sizes at the same moisture content, due to decreased modification through chemical condensation. In contrast, the very coarse 12 % moisture content substrate was not as easily fractionated as the very coarse 30 % moisture content substrate. Following lignin removal, the rate and extent of hydrolysis is improved due to a decrease in the inhibition caused by limitations in accessibility and adsorption sites related to recalcitrant lignin. Therefore, at both initial moisture contents, increased chip size acts to reduce the severity of steam explosion. The hemicellulose derived sugar concentrations recovered in the prehydrolysate were substantially increased for the 12 % moisture content substrate, when compared to the same chips acclimated to 30 % moisture content. This indicates that the sugars in the 30 % moisture content substrate are not being as readily solubilized. Clearly, in the case of the 12 % moisture content substrate more degradation of the sugars to furans appears to be occurring. The monomer to oligomer ratio in the prehydrolysate was elevated, signifying an increase in the autohydrolysis of the oligomeric sugars to monomers, and the recovery of hemicellulose sugars in the solid substrate recovered was decreased. Another indication of an increase in the severity of the treatment was a substantial drop in recovery in the solid portion as the moisture content is decreased (Figure 16). As mentioned earlier, decreased recovery in the solids is a strong indication of increased relative treatment severity. Within the 12 % moisture content substrate the lignin concentrations were higher than the lignin concentrations in the 30 % moisture content 78 substrate. Since the recovery of the 12 % moisture content substrate was low there was actually more lignin recovered from the 30 % moisture content substrate, indicating that more of the lignin was solubilized with the 12 % moisture content substrate, and that likely the recalcitrant lignin was substantially modified during the steam treatment of the 12 % moisture content substrate. Fractionation of the steam treated substrate proved that an increase in the amount of recalcitrant lignin had occurred during the steam explosion of the substrate. The substrate with initial moisture content of 12 %, although it contained elevated levels of lignin post steam explosion, was limited in lignin removal during the peroxide treatment. This indicates that an increase in the amount of recalcitrant lignin had occurred during a more severe steam pretreatment. Finally, the lower initial moisture content substrate showed slightly inhibited hydrolysis rates, which amounts to a large decrease in the overall conversion of cellulose into glucose. The reason for this is an increase in the amount of condensed lignin, which blocks pores and also adsorbs hydrolytic enzymes disabling their ability to hydrolyse cellulose. The results of these experiments show that a decrease in the moisture content of the substrate substantially inhibits the overall conversion of cellulose, and hemicellulose to their requisite sugars (Table 6). If the overall bioconversion process is to be successful, the conversion of these sugars must be maximized, and therefore the steam treatment must be optimized to account for variability within the substrate. 79 5 R E F I N I N G 5.1 Introduction Previous chapters have shown the importance of initial substrate chip size and moisture content on the steam explosion process. However, substrate particle size also plays a significant role on the overall effectiveness of the bioconversion process (Toussaint et al, 1991; Mooney et al, 1999). Further to these two parameters, one way to improve the subsequent steps of the bioconversion process is to reduce the particle size, through mechanical refining of the substrate. Refining can initiate the solubilization of hemicellulose-derived sugars, as well as a portion of the cellulose sugars. The solubilization of these sugars increases the pore space, and thereby increases the total surface area for peroxide and enzymatic hydrolysis (Himmel, 1994). Although peroxide is a very small molecule and can penetrate fibres, regardless of size, lignin extraction is limited by the pore volume of fibres. A reduction in particle size results in a decrease in the length that lignin fragments must travel in order to be removed from the substrate. During peroxide delignification the interaction between peroxide and the lignin results in the generation of soluble by-products, if the pore space is small or not connected the soluble lignin by-products may not be able to be removed from the substrate resulting in higher residual lignin. Fibre size also affects the rate and extent of enzymatic hydrolysis in two ways: indirectly fibre size influences peroxide delignification, and therefore affects overall lignin concentration of the fibres which has been shown to have a negative correlation with the rate and extent of hydrolysis (Yang et al, 2002). Fibre size also affects the hydrolysis rates directly, by influencing the availability of the cellulose fibres to the 80 hydrolytic action of the enzymes, as smaller fibres have a greater surface area to weight ratio than larger fibre (i.e. more contact points). 5.2 Steam explosion/refining The most obvious effect of refining the steam-exploded substrate was a change in the proportion of material recovered in the solid and liquid fractions. These changes are largely due to the solubilization of a small amount of the hemicellulose and cellulose based sugars that result from the elevated temperatures that the substrate is subject to within the confines of the refiner. Comparing the total sugars recovered in the refined prehydrolysate with the unrefined, the former showed a substantial increase, at all initial chip sizes (Figure 21). Clearly, the combined effect of the heat generated during refining, and the defibration event solubilized sugars. The majority of the sugars released appear to be derived from mannose and glucose (Table 7). A comparison of monomeric and oligomeric sugars released during refining demonstrates an accumulation of oligomeric carbohydrates. Table 7. Carbohydrate content in the prehydrolysate recovered from steam explosion and refining of the original chips (g/lOOg of original weight) Chip size Arabinose Galactose Glucose Xylose Mannose 12% Very Fine 1.0 2.1 4.9 3.4 6.3 Moisture Medium 0.8 2.5 9.7 2.7 8.3 Content Very Coarse 0.7 2.4 9.8 2.5 7.4 12% Very Fine 1.0 2.7 7.0 3.7 9.0 Moisture Medium 1.0 2.8 10.0 3.5 10.4 Content Very Coarse 0.8 2.8 8.3 2.7 7.3 Refined 30% Very Fine 0.8 1.5 4.1 3.0 6.1 Moisture Medium 0.7 1.9 6.3 2.5 7.6 Content Very Coarse 0.5 2.2 8.0 1.9 8.4 30% Very Fine 0.8 2.2 4.5 3.2 9.2 Moisture Medium 0.7 2.3 8.7 3.2 10.8 Content Very Coarse 0.6 2.5 11.6 3.1 11.4 Refined 81 36 32 28 24 20 16 12 Y777A 12 % Moisture Content 12 % Moisture Content Refined 30 % Moisture Content 30 % Moisture Content Refined Very Fine Medium Very Coarse Figure 21. Total carbohydrate recovered in the water-soluble fraction derived from steam explosion of chips of different initial size and moisture content subjected to refining (error bars represent the range based on the analysis of three samples removed from the liquid portion). 82 In addition to the increased liberation of oligomeric sugars, the generation of furans increased slightly, which is likely related to the heat generated during refining. As a result of the elevated levels of solubilization a decrease in the amount of solid material (lignin, cellulose and hemicellulose) recovered was apparent (Figure 22). Furthermore, during the solubilization initiated by refining process a concurrent slight decrease in the proportion of lignin remaining in the solid substrate was observed (Figure 23). Suggesting a greater proportion of lignin was also solubilized during refining, or was directly associated with the solubilized hemicellulose derived sugars (LCC). 5.3 Fractionation Refining the substrate has many benefits, but the two major anatomical features that have been modified include an increase in the pore volume, due to the removal of hemicellulose and lignin, and a decrease in the overall fibre size. Both of these factors are beneficial to the efficacy of the peroxide fractionation process, which ultimately, amplifies the lignin removal. Not only was the amount of lignin removed from the refined substrate slightly improved, the residual lignin after peroxide fractionation was decreased (Figure 24). 5.4 Hydrolysis The residual lignin content plays the most important role in the rate and extent of enzymatic hydrolysis by controlling the availability of cellulosic substrate and the amount of enzyme-substrate interaction. The results of this work agree with previous work, whereby a decrease in the fibre size increased the rate of enzymatic hydrolysis (Mooney et al., 1999). It was apparent that the best substrate for enzymatic hydrolysis 83 95 90 85 80 I V77Z\ 12 % Moisture Content 12 % Moisture Content Refined 30 % Moisture Content 30 % Mositure Content Refined 75 h-70 CD > o o CD DC T3 = 65 00 60 55 50 Very Fine Medium Very Coarse Figure 22. Total recovery of solid substrate after steam explosion of chips of different initial size and moisture content subjected to refining (error bars represent the range based on the results of three mass balances). 8 4 Very Fine Medium Very Coarse Figure 23. Total lignin retained in the solid fraction after steam explosion of substrate derived from different chip sizes and moisture contents subjected to refining (error bars represent the range based on the results of three klason lignin analyses). 8 5 V777A 12 % Moisture Content 12 % Moisture Content Refined Very Fine Medium Very Coarse Figure 24. Residual lignin after steam explosion, fractionation (1 % peroxide, 2 % substrate, 195°C, pH 11.5) and refining of wood chips derived from different chip sizes and moisture content subjected to refining (error bars represent the range based on the results of three klason lignin analyses). 86 Figure 25. Comparison of the hydrolysis rates of substrates derived from wood chips of different size: (A) 12 % moisture content; (B) 12 % initial moisture content and refined prior to fractionation; (C) 30 % initial moisture content; (D) 30 % initial moisture content and refined prior to fractionation (error bars represent the range based on the results of two separate hydrolyses). 87 100 Time (hours) Figure 26. Comparison of hydrolysis rates from chips derived from very coarse chips with different moisture contents subjected to refining (error bars represent the range based on the results of two separate hydrolyses). 88 was the refined substrate derived from the very coarse, 30 % initial moisture content chips (Figure 25). This substrate showed an 84 % and a 97 % cellulose conversion after 12 and 48 hours, respectively. The initial rate of hydrolysis appears to be quite similar up to 12 hours, however, after this point the very coarse refined 30 % moisture content substrate proceeded more rapidly than the other substrates (Figure 26). Overall, the very coarse refined substrate had a 97 % yield after 48 hours, which was 4 % higher that the 12 % moisture content refined sample, and 10-11 % higher than the samples which were not refined. Even though refining improved hydrolysis efficiency, other refined substrates only reached 70 % yield after 72 hours. Table 8. Recovery of wood constituents after steam explosion, refining, fractionation and hydrolysis of original components of wood chips entered into steam explosion (% of original weight) Percent recovery (%) 12 % Moisture Content Unret ined Hemicellulose Glucose Lignin Very Fine Medium Very Coarse 55.7 62.5 56.1 48.2 75.6 72.5 78.6 82.6 89.4 30 % Moisture Content Unrefined Hemicellulose Glucose Lignin Very Fine Medium Very Coarse 49.5 55.2 56.7 56.0 74.8 84.1 79.1 94.8 100.0 12 % Moisture Content Refined Hemicellulose Glucose Lignin Very Fine Medium Very Coarse 70.9 77.4 51.5 62.2 82.6 84.6 74.3 77.6 82.3 30 % Moisture Content Refined Hemicellulose Glucose Lignin Very Fine Medium Very Coarse 66.2 73.5 76.7 55.0 82.2 90.2 68.3 81.3 91.0 89 5.5 Conclusion Clearly, refining has been shown to be an effective pretreatment technique for improving the process of steam-exploding substrates. The refined substrate releases a larger amount of sugar into the prehydrolysate where it is readily available for fermentative organisms. However, the solubilization of sugar from the substrate decreases the residual sugar content of the solid substrate. The remaining lignin, although in a lower concentration, is much easier to remove from the substrate. Lignin removal seems to be facilitated through a decrease in the fibre size, which increases the available surface area, and the elevated hemicellulose removal increases the amount of pore volume of the substrate. The combination of these two factors improves the efficiency of the peroxide wash, and therefore leaves less lignin, which could potentially inhibit enzymatic hydrolysis. The residual solid substrate has therefore substantial improved hydrolysis rates. Generally, the glucose yield was 5 % to 7 % higher for the substrate refined after steam explosion, which shows the importance of fibre size to the overall process. Generally, the glucose stream generated from enzymatic hydrolysis was lower, but the total glucose released was higher due to a greater concentration of glucose in the prehydrolysate of the very coarse refined substrate (Table 8). Additionally, the concentration of hemicellulose sugars in the prehydrolysate was higher due to solubilization of sugars during the refining, with minimal furan production. 90 6 D E L I G N I F I C A T I O N 6.1 Introduction At present, the most effective means of delignifying recalcitrant softwood substrates generated by the steam explosion process is a hot alkali peroxide treatment developed by Yang et al. (2002). This system employs a chemical loading of 1% (w/v) peroxide, 2 % (w/v) substrate consistency, at 80°C for 1 hour (pH 11.5). Although effective, the peroxide loading is excessive, when considering the economic implications of such a large chemical loading. In order for the bioconversion process to become truly viable the cost of chemicals (peroxide) must be either reduced or replaced with another delignification technique. It has been suggested that the pretreatment, fractionation and hydrolysis of lignocellulosic substrates comprises 60 % of the overall cost of producing ethanol (Nguyen, 1991), of which peroxide is one of the major contributors. It is believed that peroxide delignification could be more effective than present, if the consistency and reaction times are optimized. As well, control of the rate and extent of peroxide decomposition plays a significant role in the overall effectiveness of the delignification event. Currently, an array of chemicals are available for minimizing the decomposition of peroxide, and the role of each of these chemicals is slightly different. For example DTMPA, an organic chelating agent, is used during peroxide bleaching to surround and stop metal ions from decomposing peroxide (Xo, 1994). Another chemical, sodium silicate, has also been used to stabilize peroxide. Sodium silicate forms silicic acid, which complexes with the hydrogen peroxide in solution resulting in a very stable peroxide solution (Harder et al, 1960). In combination, sodium silicate and magnesium sulphate react to form magnesium silicate, which inactivates metal ions through 91 adsorption (Singh, 1979). The current research project evaluated four chemicals individually in an attempt to improve the efficacy of the current peroxide fractionation step. 6.2 Alternative delignification techniques Four types of delignification techniques that have shown promise in the bleaching or delignification of lignocellulosics include: peroxide, oxygen, ozone, and wet oxidation. Although very different chemically, all four of these delignification systems utilize the same general mechanism for the cleavage and removal of phenolic groups within the lignin moieties. Three of the four techniques were examined for their effectiveness at delignifying steam exploded Douglas-fir wood chips. Oxygen delignification was evaluated with little success (Figure 27), as was only shown to remove on average only 5 % of the initial lignin. An additional alkali wash was also performed to test whether the oxygen delignification increased the reactivity of the lignin. The alkali wash reduced the lignin content on average an additional 3 %. However, the lignin remaining after the alkali wash was roughly 39 % of the dry weight of the samples, and without further delignification this amount of lignin would negatively influence the rate and extent hydrolysis and ultimately limit saccharification. The reason for the low delignification efficacy of oxygen could be related to the transfer of oxygen through the suspending solution into the substrate, as was noticed in the pulp mills observed by Bennington et al. (1999). At a 2 % substrate consistency the mass transfer of oxygen through the alkali solution and onto the substrate is limited, and therefore a substantial amount of oxygen moves through the reactor without ever reacting 92 Stock Oxygen Alkali Wet Alkali Wet Standard Substrate Oxygen Oxidation Oxidation Peroxide Figure 27. The effect of different fractionation chemicals on residual lignin in steam exploded Douglas-fir very coarse wood chips, which were refined (error bars represent the range based on the results of three klason lignin analyses). 93 with the substrate. An increase in the consistency may increase the mass transfer of oxygen and therefore improve the ability of the system to remove lignin (Bennington and Pinealt, 1999). This is related to an improved transfer of superoxide and hydroxyl radicals from the water to the substrate. In a low consistency solution these radicals react to form oxygen and are lost during venting. Wet oxidation has also been shown to be an effective means of removing lignin from wheat (Schmidt and Thomsen, 1998), however, was relatively ineffective at removing the recalcitrant lignin from the Douglas-fir substrate (Figure 27). Generally, wet oxidation reduced the lignin content by 50 %, and thus produced a substrate containing 21 % lignin. Following a subsequent alkali wash the lignin content was reduced to 17 %. This residual lignin concentration is also far too high for the efficient hydrolysis of the substrate. These results again suggest that a mass transfer problem existed in the delignification reactor. The consistency of this solution was also 2 %, and similar to the oxygen delignification system the substrate loadings may be adjusted higher to improve the transfer of the delignification radicals into the substrate. Both wet oxidation and oxygen delignification were not able to remove as much lignin as the current peroxide delignification process as outlined by Yang et al. (2002), which employed a substrate concentration of 2 %. The peroxide technique has been shown to generate substrate with a 10 % residual lignin content. However, on a close inspection of the current reaction conditions it was evident that there is room for improvement in this system, including increased reaction consistency and the addition of peroxide stabilizers. 94 6.3 Peroxide delignification Peroxide delignification, as was previously shown, demonstrated the greatest ability to remove the recalcitrant lignin remaining after steam explosion. However, the current level of peroxide addition required for the effective removal lignin is far too high. This suggests that a huge volume of the peroxide applied is lost during peroxide fractionation reactions. Currently, high consistency peroxide delignification in the pulp and paper industry is showing great promise (Bennington and Pinealt, 1999), whereby, increasing the substrate consistency during bleaching/delignification improves the contact between peroxide in solution and the substrate (Gould, 1985). Similar results were also observed with the steam-exploded substrate, which showed a substantial improvement in the lignin removal at elevated substrate concentration (Figure 28). Delignification was 50 % better at a 10 % substrate concentration than at a 2 % substrate concentration. This is a substantial improvement in the capacity of peroxide to remove lignin from the cellulosic polymer without further increasing the loading of chemicals and therefore the cost. Although it has been suggested that even higher substrate concentrations are more effective (Bennington and Pinealt, 1999), no further improvements were observed above 10 % substrate consistency in the current system (Figure 28). In fact, the residual lignin in the 12 % consistency trials showed a slight increase in residual lignin composition compared to 10 % consistency delignification, which could be due to decomposition of the peroxide. It became apparent during the investigations evaluating consistency that the rate of peroxide consumption is very rapid. On average, the 2 % substrate consistency showed 95 c CD -4—' c o o c 'c 0 I WHS * * i rf - i f r ' T H H H H H ¥ . if* r .. . . . . s 10% 12% 2% 5% 7% Substrate Consistency Figure 28. The effect of substrate concentration during delignification on residual lignin in Douglas-fir feedstock derived from coarse wood chips that were steam exploded, refined and fractionated (error bars represent the range based on the results of three klason lignin analyses). 96 2.0 - • — Standard • Double ^ - Blank 30 40 Time (min) 50 60 Figure 29. Rate of decomposition of residual peroxide during delignification under three conditions, standard: 1 % H 2 0 2 , 2 % substrate, 80°C, pH 11.5; Double peroxide: 2 % H 2 0 2 , 2 % substrate, 80°C, pH 11.5; Substrate Blank: 1 % H 2 0 2 , 80°C, pH 11.5 (error bars represent the range based on the results of two separate delignification events). 97 total peroxide decomposition within 20 minutes (Figure 29), suggesting that the delignification event in the currently "optimized" delignification stage occurs in the first half of the entire reaction. Furthermore, doubling the peroxide loading did not improve the lifetime of peroxide in solution (Figure 29). Actually, peroxide decomposition during the double loading was faster than the standard peroxide delignification, which is likely due to the increased concentration of the perhydroxyl (OOH') ions, which would increase the decomposition of peroxide (Gould, 1985). Furthermore, peroxide decomposes within 20 minutes in a blank solution, which contained all of the previously optimized loadings (1 % (w/v) peroxide, pH of 11.5, at 80°) except there was no substrate added. Therefore, the reaction time of the current peroxide delignification stage could be decreased by at least half, or more importantly, the peroxide could be stabilized to reduce the decomposition of peroxide and consequently improve the efficiency of delignification. 6.4 Chelation and stabilization To address the question of peroxide stabilization three different stabilizers were evaluated individually and with prior EDTA chelation combination, in an attempt to minimize the effects of peroxide decomposition, and therefore improve delignification of the steam-exploded substrate. In addition, the substrate was first chelated prior to the addition of stabilizers to evaluate the effects of metal remaining in solution. The stabilizers evaluated were diethylene-triamine-pentamethylene-phosphonic acid (DTMPA), sodium silicate and magnesium sulphate with half of the samples subjected to chelation with ethylene-diamine-tetraaccetic acid (EDTA) prior to stabilization. In the previous work (chapters 3-5) the coarse chip sizes after steam explosion and refining had 98 the highest level of lignin removal during delignification, and was therefore the substrate used for all subsequent delignification experiments. As was shown earlier, the consumption of peroxide is rapid during the previously optimized peroxide delignification the peroxide is consumed within the first twenty minutes of the reaction (Figure 29). The rate of peroxide decomposition and the amount of gas generated by visual inspection (bubbles in reaction flasks) indicates that the peroxide is decomposed before reacting with the substrate present in the solution. The addition of stabilizers such as sodium silicate and magnesium sulphate slowed the decomposition of the peroxide to the extent that residual peroxide was detected after a 40-minute reaction time, which was better than the previous standard delignification treatment. With the addition of diethylene-triamine-pentamethylene-phosphonic acid the decomposition of peroxide was slowed by 40 minutes, leaving residual peroxide after 1 hour (Figure 30). A klason analysis of the substrate indicates that stabilization of the peroxide was not an effective means to enhance lignin removal (Figure 31). Comparing the standard delignification with the samples stabilized with sodium silicate alone or with sodium silicate and magnesium sulfate resulted in a lower degree of lignin removal (Figure 31). Diethylene-triamine-pentamethylene-phosphonic acid, on the other hand, reduced the residual lignin content by an additional 1.5 % over that attainable under standard conditions. The residual lignin content and the peroxide consumed during delignification indicate that there may still be a source of peroxide decomposition in the system. Peroxide decomposition may be due to metals contained within the substrate, which could be removed by chelation of the substrate prior to peroxide delignification. 99 1.4 | • EDTA chelated - • — DTMPA stabilized 1 2 " A EDTA and DTMPA • • EDTA, sodium silicate and magnesium sulphate 1 .0 - EDTA and sodium silicate 0 10 20 30 40 50 60 Time (min) Figure 30. The effect of stabilizers, ethylene-diamine-tetraaccetic acid (EDTA), diethylene-triamine-pentamethylene-phosphonic acid (DTMPA), and sodium silicate and magnesium sulphate in combination with ethylene-diamine-tetraaccetic acid (EDTA), on residual peroxide at optimized peroxide 1 % peroxide, 2 % substrate, 80°C, pH 11.5) delignification conditions. 100 Nonchelated Substrate Y777A Chelated Substrate Figure 31. The effect of stabilizers on residual lignin in Douglas-fir substrate derived from coarse wood chips that have been steam exploded, refined and fractionated (error bars represent the range based on the results of three klason lignin analyses). 101 Chelation of the substrate prior to delignification substantially slowed the consumption of peroxide (Figure 30). The subsequent stabilization with sodium silicate and chelation slowed the consumption of peroxide by 40 minutes compared to the standard solution. This is 20 minutes longer than the sample that was treated with sodium silicate alone. In contrast, the addition of magnesium sulfate, sodium silicate and chelation after 1 hour of delignification demonstrated that 20 % of the initial peroxide loading remained, and consequently was identified as the best means of stabilizing the peroxide (Figure 30). Diethylene-triamine-pentamethylene-phosphonic acid combined with chelation resulted in 7.5 % of the initial peroxide remaining in solution after 1 hour. However, interestingly the rate of peroxide consumption was different for this sample when compared to the other samples (Figure 30). Instead of rapid peroxide consumption from 10 to 15 minutes reaction time and then gradual consumption, the peroxide was shown to gradually decompose until the 40-minute mark after which the rate of peroxide consumption slowed. The difference in the rate of peroxide consumption may indicate a difference in how peroxide is reacting with the substrate. The performances of the chelated peroxide samples were better than the samples without prior chelation (Figure 31). The samples chelated prior to peroxide treatment with sodium silicate and sodium silicate and magnesium sulfate showed on average 1-1.25 % decreases in lignin removal, which still resulted in lignin contents higher than the previously optimized conditions. The chelation with EDTA alone improved the residual lignin content more than when used in combination with sodium silicate or magnesium sulfate. However, the diethylene-triamine-pentamethylene-phosphonic acid and EDTA 102 chelated samples proved to be the best overall delignifying agent, resulting in a residual lignin content of only 2.5 %. 6.5 Optimization Prior to the optimization of delignification the effects of residual lignin on the rate and extent of hydrolysis were evaluated. This was accomplished by subjecting substrate containing different amounts of residual lignin to similar enzymatic hydrolysis reactions (based on FPU/g cellulose) maintained at a single condition, and thereby establishing a target lignin content that needs to be achieved to facilitate effective hydrolysis of the substrate. To understand the effect of lignin content on the subsequent hydrolysis rates the loading of peroxide was reduced in 0.1 % increments, resulting in an increase in the residual lignin content (Figure 32). The effect of the increase in lignin content was a reduction in the rate of hydrolysis, as expected. The samples, which showed the most promise, had peroxide loadings down to 0.5 % while maintain high hydrolysis rates (Figure 33), which correspond to a residual lignin content ranging anywhere from 2.5 to 15 %. The samples below 7.5 % residual lignin showed 97 % conversion after 24 hours, while samples with lignin contents ranging from 7.5 to 15 % residual lignin demonstrated 97 % conversion after 48 hours and had yield ranging from 84 to 90 % after 24 hours. It appears that if the hydrolysis was allowed to proceed for 72 hours or longer all of the samples would reach 100 % of theoretical yield. 103 30 25 \-20 h c -£ 15 o o & 10 0.8 % 0.7 % 0.6 % 0.5 % 0.4 % Peroxide loading (%) 0.3 % 0.2 % Figure 32. The effect of varying peroxide concentration on residual lignin content of steam exploded Douglas-fir chelated with EDTA and treated with peroxide stabilized with DTMPA (error bars represent the range based on the results of three klason lignin analyses). 104 100 0.8 % Peroxide 0.7 % Peroxide 0.6 % Peroxide 0.5 % Peroxide 0.4 % Peroxide 0.3 % Peroxide 0.2 % Peroxide 50 Time (hours) Figure 33. Rate of hydrolysis of steam-exploded Douglas-fir substrate subjected to different peroxide loadings during fractionation (error bars represent the range based on the results of two separate hydrolyses). 105 This illustrates that the peroxide loading could be minimized, decreasing the chemical loadings, while at the same time maintaining hydrolysis yields. The effect of EDTA was also tested on the subsequent lignin concentrations. Without EDTA chelation the peroxide loadings for adequate hydrolysis efficiency were 0.8 % (w/v), which is 0.3 % (w/v) higher than the peroxide loadings required for samples chelated with EDTA prior to delignification. Due to the fact that peroxide loading of 0.8 % with the addition of DTMPA and prior chelation was rapidly hydrolysed, this sample was utilized as a centre point for the optimization of the peroxide loading. The goal of this optimization was to determine the peroxide loading, DTMPA loading and temperature required to obtain a lignin content, which would not hinder the saccharification of the substrate. Also, a secondary goal was to determine how the variation in delignification conditions affected lignin removal. The relationship between peroxide loading and temperature appears to be a fairly simple (Figure 34), and indicates that the optimum peroxide loading and temperature are 65° and a peroxide loading of 0.9 %, which results in a lignin content of less than 2 %. In order to achieve a residual lignin content of 14 % in the delignified substrate, peroxide loading can be reduced to 0.45-0.5 % peroxide with a temperature range of 65-95 °C. It appears that temperature is not as important as the peroxide loading, as optimum residual lignin concentration fall in the full range of temperatures tested, from 65-95°C. The relationship between temperature and DTMPA loading is quite simple as well (Figure 35). The relationship appears to depend mostly on DTMPA loading, and have only a minor temperature effect. The highest DTMPA loading appeared to achieve the best results with a high temperature (95°C). The relationship between peroxide, DTMPA, 106 and lignin content is fairly complex (Figure 36), however. It appears that peroxide loading and DTMPA concentrations both have strong effects on the removal of lignin from the substrate. There appears to be a plateau in the ability to delignify the chelated substrates that can be achieved with the use of DTMPA and peroxide, suggesting that the best conditions for delignification of the previously chelated substrate are 0.18 % (w/w) DTMPA, 0.45 to 0.5 % (w/v) peroxide, increasing as temperature is increased, and 65 to 95°C temperatures to achieve a 10-14 % residual lignin content which has been shown to result in a high hydrolysis rate and yield. Prior to the acceptance of these conditions they were tested to determine if the results are true for the delignification and hydrolysis. The peroxide delignification of samples at four conditions, determined to be optimum from the statistical analysis, show that the optimum conditions may require slightly higher peroxide loadings (0.1 % higher). The optimum conditions utilized were: 0.45 % peroxide, 65°C, and 0.18 % DTMPA; 0.465 % peroxide, 75°C, and 0.18 % DTMPA; 0.480 % peroxide, 85°C, and 0.18 % DTMPA; 0.50 % peroxide, 95°C, and 0.18 % DTMPA. The residual lignin after this treatment was in the range of 13-15 %, as targeted. The hydrolysis of these samples was quite rapid with full hydrolysis occurring after 72 hours (Figure 37). Three of the samples had high initial rates of hydrolysis as seen by the 90 % conversion after 48 hours. The fourth sample had the highest lignin content, 14.8 % compared to 13.4 % for the best-delignified sample, and also showed a substantially decreased performance in enzymatic hydrolysis. Clearly, this indicates that below 14 % lignin is a feasible target for lignin concentration to obtain a substrate that efficiently can be hydrolysed. 107 Figure 34. Three-dimensional surface plot showing the relationship between temperature (°C), peroxide concentration (% w/v) and residual lignin content. 108 Figure 35. Three-dimensional surface plot showing the relationship between temperature (°C), DTMPA loading (% w/v) and residual lignin content. 109 Figure 36. Three-dimensional surface plot showing the relationship between peroxide concentration (% w/v), DTMPA loading (% w/v) and residual lignin content. 110 o I i I i I i I i I i I i I i I I 0 10 20 30 40 50 60 70 Time (hours) Figure 37. Hydrolysis of optimized (factorial design) delignification conditions (Very Low- 0.45 % peroxide, 65°C, and 0.18 % DTMPA; Low- 0.465 % peroxide, 75°C, and 0.18 % DTMPA; High- 0.480 % peroxide, 85°C, and 0.18 % DTMPA; Very High- 0.50 % peroxide, 95°C, and 0.18 % DTMPA), determined by statistical analysis, of Douglas-fir wood chips that have been steam exploded, refined and fractionated (error bars represent the range based on the results of two separate hydrolyses). I l l 6.6 Conclusion To date peroxide delignification has proven itself to be the best of the three types of delignification techniques evaluated (oxygen, wet oxidation, and peroxide). The peroxide delignification technique left 34.5 % less lignin than oxygen delignification and 15.5 % less lignin than wet oxidation. This is without further improvement in the techniques used for peroxide delignification (stabilization/chelation, and higher consistency). With improvements such as increased consistency and chelation, the amount of residual lignin could be further reduced. The performance of peroxide delignification at high substrate consistency and prior chelation was increased to the point were hydrolysis rates were not improved. Therefore, an attempt to reduce the loading of chemicals was preformed through a statistical analysis, which showed that a significant reduction in peroxide could be achieved with a substantial cost benefit, in the range of 40 % saving in chemical costs. Also, a reduction in the time taken for delignification could probably be preformed with little loss in overall performance. 112 C O N C L U D I N G R E M A R K S The primary goal of this work was to elucidate the effect of chip size, moisture content and post explosion refining on the recovery of carbohydrate sugars in the prehydrolysate and glucose from enzymatic hydrolysis. Previous optimization of the steam explosion conditions determined that medium severity conditions were the best compromise between hemicellulose recovery and cellulose digestibility (Boussaid et al., 2000). This fact remains true for certain wood chip conditions, but the steaming conditions determining medium "relative" severity vary depending on the chip size and moisture content. The effect of increasing the chip size is a decrease in the severity of treatment due to a reduction in the rate of heat transfer during steam treatment, which is illustrated by dramatic changes in the composition of the residual substrate. The first effect of an increase in the chip size from very fine to very coarse is a decrease in the sugars solubilized and a concurrent decrease in the degradation of the sugars into furfural and hydroxymethylfurfural. The decrease in solubilized sugars led to an increase in the water insoluble substrate recovered after steam explosion. Due to decreased severity as chip size increased the lignin is easier to remove during peroxide fractionation. Thus, there is an overall decrease in the residual lignin in the substrate and the hydrolysis rates are much faster. The results of chip size show that the coarsest chips used are the best chips to employ under the current conditions. An increase in the moisture content also dramatically decreased the severity of steam treatment, causing a decrease in the recovery of water-soluble sugars and an increase in 113 the solid recovery. An increase in the moisture content resulted in decreased condensation of the lignin and therefore a substrate, which performed better during delignification. Lower lignin contents seem to result in increased hydrolysis rates. Overall, the recovery of sugars, in the 30 % initial moisture content substrate, was increased due to higher recoveries of solid sugars after the steam explosion process. Chip size and moisture content manipulation indicate that the recovery of sugars is higher for the coarse and high moisture content substrate, implying that the severity of treatment may be too high for fine low moisture content substrates. Optimization of steam pretreatment conditions to account for the rate of heat transfer into the fine low moisture content chips may result in a substrate, which has improved performance during the steam treatment. Post steam explosion refining results in increased solubilization of sugars from the steam-exploded material, which increases the pore volume available for the removal of lignin degradation products after interacting with peroxide radicals. The increase in pore volume combined with an increase in the surface area of the individual fibres resulted in improved delignification (improved lignin removal and decreased carbohydrate degradation). Lignin removal during delignification was 5 % higher after the refining of the substrate, which demonstrated the best hydrolysis conditions, with the highest rate and yield. Generally, refining resulted in the highest total recovery of sugars, due to better performance during peroxide delignification and hydrolysis. This work suggests that the largest chip size with 30 % moisture content and subsequent refining is the optimum substrate for the currently optimized steam explosion bioconversion process. This substrate provides the highest yield and decreases the 114 chemical requirements to achieve effective removal. Furthermore, if the severity of steam explosion is reduced the performance of the finer substrates could be optimized, increasing the performance of the medium and very fine chips sizes and possibly increasing the yield of the bioconversion process. Due to the performance of the very coarse 30 % moisture content refined substrate it was employed during investigations attempting to reduce the chemical loadings during delignification. Previous optimization of the peroxide delignification conditions were 2 % substrate loading, 1 % peroxide and a pH of 11.5 for a 1-hour reaction period (Yang et al, 2001). These conditions result in a substrate, which has a rapid hydrolysis rate and yield. Unfortunately, this peroxide loading is far too high. Alternative delignification techniques were tested, including wet oxidation and oxygen, and proved not to be useful for the removal of recalcitrant lignin, since a minimum of 20 % lignin remained. Increasing the consistency of the hot alkaline peroxide delignification process showed promise, through increasing the substrate to peroxide interactions, as a lignin content of 2.5 % was achieved after fractionation. Chelation followed by DTMPA stabilized delignification proved to be a substantial improvement in the delignification process. A residual lignin content of 2.4 % was achieved with the stabilization of peroxide. With the use of a factorial matrix an attempt was made to reduce the peroxide required for fractionation. A lignin content of 13 % was shown to be the highest lignin content that maintained near theoretical glucose yields after 48-hour enzymatic hydrolysis, therefore it was used as a target for peroxide delignification. With the 13 % residual lignin target a 40 % reduction in peroxide loading is achievable with no loss in the performance in the substrate during delignification. 115 FUTURE WORK The bioconversion process is still in a development stage, which leaves much room for improvement, particularly in pretreatment and delignification stages. As this project has highlighted the influence of feedstock parameters have on the efficacy of treatment, further work should examine the actual feedstock, which is likely going to be utilized by the bioconversion process for the production of cheap fuel grade ethanol. This work must test other softwood species to examine the effect of the current steam explosion conditions and examine the robustness of the current conditions. Once the feedstock properties have been elucidated the steam explosion conditions can then be optimized to obtain high hydrolysis rates while at the same time maintaining hemicellulose recovery in the prehydrolysate. Delignification conditions and technique also require further work. The current conditions have been shown in this work to remove enough lignin to maintain high hydrolysis rates and yield, but at the same time there may be more efficient and cheaper delignification techniques available. This work showed that wet oxidation and oxygen delignification were not as effective as peroxide, but further work has shown that increased consistency may increase the effectiveness of the technique. Furthermore, other techniques for delignification need to be tested, such as ozone, so that the most economically viable delignification technique is used. If peroxide maintains the position as the most effective delignification agent for recalcitrant lignin removal, other stabilizers should be tested to reduce the cost and loadings of chemicals. 116 BIBLIOGRAPHY Aber, J. D. and Melillo, J. M. (1991). Terrestrial Ecosystems. Toronto, Saunders College Publishing. Akim, L. G., Colodette, J. L. and Argyropoulos, D. S. (2001). "Factors limiting oxygen delignification of kraft pulp." Canadian Journal of Chemistry 79: 201-210. Anderson, R. A. and Watson, S. A. (1982). Handbook of Processing and Utilization in Agriculture. Boca Raton. Avellar, B. and Glasser, W. G. (1998). "Steam-assisted biomass fractionation. I process considerations and economic evaluation." Biomass and Bioenergy 14(3): 205-218. Ballesteros, I., Oliva, J. M., Navarro, A. A., Gonzalez, A., Carrasco, J. and Ballestros, M. (2000). "Effect of chip size on steam explosion pretreatment of softwood." Applied Biochemistry and biotechnology 84-86: 97-110. Beatson, R. P. (1986). "The topochemistry of aspen sulphonation." Holzforschung 40(supplement): 11-15. Bennington, C. P. J. and Pinealt, I. (1999). "Mass transfer in oxygen delignification systems: mill survey results, analysis, and interpretation." Pulp and Paper Canada 100(12): 123-131. Boussaid, A., Esteghlalian, A., Gregg, D. J., Lee, K. H. and Saddler, J. N. (2000). "Steam pretreatment of Douglas-fir wood chips." Applied Biochemistry and biotechnology 84-86: 693-705. Boussaid, A., Jarvis, J., Gregg, D. J. and Saddler, J. N. (1997). "Optimization of hemicellulose sugar recovery from a steam-exploded softwood (Douglas-fir)." The Third Biomass Conference of The Americas: 873-880. Boussaid, A., Robinson, J., Cai, Y., Gregg, D. J., Saddler, J. N. and Robinson, J. (1998). "Fermentability of the hemicellulose-derived sugars from steam-exploded softwood." Biotechnology and Bioengineering 64(3): 284-289. Brownell, H. H., Yu, A. H. C. and Saddler, J. N. (1985). "Steam-Explosion Pretreatment of Wood: Effect of Chip Size, Acid, Moisture Content, and Pressure Drop." Biotechnology and Bioengineering 28: 792-801. Chang, V. S. and Holzapple, M. T. (2000). "Fundamental factors affecting biomass enzymatic reactivity." Applied Biochemistry and biotechnology 84-86: 5-37. 117 Chum, H. L., Johnson, D., Black, S. and Overend, R. P. (1990). "Pretreatment catalysts^  effects and the combined severity parameter." Applied Biochemistry and biotechnology 25(5): 1-14. Converse, A. O., Ooshima, H. and Burns, D. S. (1990). "Kinetics of enzymatic hydrolysis of lignocellulosic materials based on surface area of cellulose accessible to enzyme and enzyme adsorption on lignin and cellulose." Applied Biochemistry and biotechnology 24/25: 67-73. Converti, A., Dominguez, J. M., Perego, P., Silverio da Silva, S. and Zilli, M. (2000). "Wood hydrolysis and hydrolysate detoxification for subsequent xylitol production." Chemical engineering technology 23(11): 1013-1020. Draude, K. M., Kurniawan, C. B. and Duff, S. J. B. (2001). "Effect of oxygen delignification on the rate and extent of enzymatic hydrolysis of lignocellulosic material." Bioresource Technology 79: 113-120. Excoffier, G., Toussaint, B. and Vignon, M. R. (1991). "Saccharification of steam exploded poplar wood." Biotechnology and Bioengineering 38: 1308-1317. Fan, L. T., Gharpuray, M. M. and Lee, Y. H. (1987). Cellulose hydrolysis. NewYork, Springer-Verlage. Foody, P. (1982). Steam Explosion as a Pretreatment for Biomass Conversion. Final report to MidwestResearch Institute. Galbe, M. and Zacchi, Z. (2002). "A review of the production of ethanol from softwood." Applied Microbial Biotechnology 59: 618-628. Garrote, G., Dominguez, H. and Parjo, J. C. (1999). "Hydrothermal processing of lignocellulosic materials." Holz als Roh- und Werkstoff 57: 191-202. Gould, J. M. (1985). "Studies on the mechanism of alkaline peroxide delignification of agricultural residues." Biotechnology and Bioengineering 27: 225-231. Gregg, D. J., Boussaid, A. and Saddler, J. N. (1998). "Techno-economic evaluations of a generic wood to ethanol process: effect of increased cellulose yields and enzyme recycle." Bioresource Technology 63: 7-12. Gregg, D. J. and Saddler, J. N. (1995). "Bioconversion of lignocellulosic residue to ethanol: process flowsheet development." Biomass and Bioenergy 9(1-5): 287-302. Gregg, D. J. and Saddler, J. N. (1996). "Factors affecting cellulose hydrolysis and the potential of enzyme recycle to enhance the efficiency of an integrated wood to ethanol process." Biotechnology and Bioengineering 51: 375-383. 118 Hartler, N., Lindahl, N., Moberg, C. and Stockman, L. (1960). Tappi 43(10): 806-813. Haygreen, J. G. and Bowyer, J. L., Eds. (1996). Forest Products and Wood Science. Ames, Iowa State University Press. Hemmingson, J. A. (1987). "Steam explosion lignins: fractionation, composition, structure, and extractives." Journal of Wood Chemistry and Technology 7(4): 527-553. Himmel, M. (1994). Enzymatic Conversion of Biomass for Fuels Production. Colorado, American Chemical Society. Hon, D. and Shiraishi, N. (2001). Wood and Cellulosic Chemistry. New York, Marcel Dekker Inc. IEA Bioenergy Task (2000). Liquid Fuels From Biomass: North America Impact of Non-Technical Barriers on Implementation. Delta. Kallavus, U. and Gravitis, J. (1995). "A comparative investigation of the ultrastructure of steam exploded wood with light, scanning and transmission electron microscopy." Holzforschung 49: 182-188. Larsson, S., Palmquvist, E., Hahn-Hagerdal, B., Tengborg, C , Stenberg, K., Zacchi, G. and Nilvebrant, N. (1999). "The generation of fermentation inhibitors during dilute acid hydrolysis of softwood." Enzyme Microbial Technology 24: 151-159. Lee, S. B., Shin, H. S. and Ryu, D. D. Y. (1982). "Adsorption of cellulase on cellulose: Effect of physiochemical properties of cellulose on adsorption and rate of hydrolysis." Biotechnology and Bioengineering 24: 2137-2153. Mansfield, S. D., Mooney, C. A. and Saddler, J. N. (1999). "Substrate and enzyme characteristics that limit cellulose hydrolysis." Biotechnology Progress 15(5): 804-816. Michalowicz, G., Toussaint, B. and Vignon, M. R. (1991). "Ultrastructural changes in poplar cell wall during steam explosion treatment." Holzforschung 45: 175-179. Millett, M., Baker, A. and Satter, L. (1976). Biotechnology and Bioengineering Symposium. 6: 125-153. Mooney, C. A., Mansfield, S. D., Beatson, R. P. and Saddler, J. N. (1999). "The effect of fiber characteristics on hydrolysis and cellulase accessibility to softwood substrates." Enzyme Microbial Technology 25: 644-650. 119 Mooney, C. A., Mansfield, S. D., Touhy, M. J. and Saddler, J. N. (1998). "The effect of initial pore volume and lignin content on the enzymatic hydrolysis of softwoods." Bioresource Technology 64: 113-119. Nguyen, Q. A., Tucker, M. P., Boynton, B. L., Keller, F. A. and Schell, D. J. (1998). "Dilute acid pretreatment of softwoods." Applied Biochemistry and biotechnology 70-72: 77-87. Nguyen, Q. A., Tucker, M. P., Keller, F. A., Beaty, D. A., Connors, K. M. and Eddy, F. P. (1999). "Dilute acid hydrolysis of softwoods." Applied Biochemistry and biotechnology 77-79: 133-142. Nidetzky, B., Steiner, W., Hayn, M. and Claeysses, M. (1994). "Cellulose hydrolysis by the cellulases from Trichoderma reesei: a new model for synergistic interaction." Journal of Biochemistry 298: 705-710. oee.mcan.gc.ca/autosmart/mostfuel/index.cfm. Ooshima, H. B., D.S.; Converse, A.O. (1990). "Adsorption of cellulase from Trichoderma reesei on cellulose and lignaceous residue in wood pretreated by dilute sulfuric acid with explosive decompression." Biotechnology and Bioengineering 36: 446-452. Overend, R. P. and Chornet, E. (1987). "Fractionation of lignocellulosics by steam-aqueous pretreatments." Phil. Trans. R. Soc. Lond. A321: 523-536. Philippidis, G. P., Smith, T. K. and Wyman, C. E. (1993). "Study of the enzymatic hydrolysis of cellulose for production of fuel ethanol by the simultaneous saccharification and fermentation process." Biotechnology and Bioengineering 41(9): 846-853. Ramos, L. P., Mathias, A. L., Silva, F. T., Cotrim, A. R., Ferraz, A. L. and Chen, C. L. (1999). "Characterization of residual lignin after SC -^catalyzed steam explosion and enzymatic hydrolysis of Eucalyptus viminalis wood chips." Journal of Agriculture and Food Chemistry 47: 2295-2302. Robinson, J. (1980). Fuels from Biomass Technology and Feasibility. New Jersey, No yes Data Corporation. Saddler, J. N., Brownell, H. H., Clermont, L. P. and Levitin, N. (1982). "Enzymatic hydrolysis of cellulose and various pretreated wood fractions." Biotechnology and Bioengineering 24: 1389-1402. Schell, D., Nguten, Q. A., Tucker, M. P. and Boynton, B. (1998). "Pretreatment of softwood by acid-catalyzed steam explosion followed by alkali extraction." Applied Biochemistry and biotechnology 70-72: 17-24. 120 Schrnidt, A. S. and Thomsen, A. B. (1998). "Optimization of wet oxidation pretreatment of wheat straw." Bioresource Technology 64: 139-151. Schwald, W., Breuil, C , Brownell, H. H., Chan, M. and Saddler, J. N. (1989). "Assessment of pretreatment conditions to obtain fast complete hydrolysis on high substrate concentrations." Applied Biochemistry and biotechnology 20/21: 29-44. Schwald, W., Brownell, H. H. and Saddler, J. N. (1988). "Assessment of pretreatment conditions for fast and complete hydrolysis of biomass." Applied Biochemistry and biotechnology 29: 71. Shevchenko, S. M., Beatson, R. P. and Saddler, J. N. (1999). "The nature of lignin from steam explosion/enzymatic hydrolysis of softwood." Applied Biochemistry and biotechnology 77-79: 867-876. Shevchenko, S. M., Chang, K., Robinson, J. and Saddler, J. N. (2000). "Optimization of monosaccharide recovery by post-hydrolysis of the water-soluble hemicellulose components after steam explosion of softwood chips." Bioresource Technology 72: 207-211. Singh, R. P. (1979). The bleaching of pulp. Atlanta, Tappi. Sinitsyn, A., Gusakov, A. and Vlasenko, E. (1991). "Effect of structural and physiochemical features of cellulosic substrates on the efficiency of enzymatic hydrolysis." Applied Biochemistry and biotechnology 30(1): 43-59. Sjostrdm, E. (1981). Wood Chemistry Fundamentals and Applications. Toronto, Academic Press. Teixeira, L. C , Linden, J. C. and Schroder, Ff. A. (1999). "Alkaline and peracetic acid pretreatments of biomass for ethanol production." Applied Biochemistry and biotechnology 77-79: 19-34. Teixeira, L. C , Linden, J. C. and Schroder, Ft. A. (2000). "Simultaneous saccharification and cofermentation of peracetic acid pretreated biomass." Applied Biochemistry and biotechnology 84-86: 111-127. Tengborg, C , Galbe, M. and Zacchi, Z. (2001). "Influence of enzyme loading and physical parameters on the enzymatic hydrolysis of steam-pretreated softwood." Biotechnology Progress 17(1): 110-117. Tengborg, C , Galbe, M. and Zacchi, Z. (2001). "Reduced inhibition of enzymatic hydrolysis of steam-pretreated softwood." Enzyme Microbial Technology 28: 835-844. 121 Tengborg, C , Stenberg, K., Galbe, M., Zacchi, Z., Larsson, S., Palmquvist, E. and Hahn-Hagerdal, B. (1998). "Comparison of SO2 and H2SO4 impregnation of softwood prior to steam pretreatment on ethanol production." Applied Biochemistry and biotechnology 70-72: 3-15. Toussaint, B., Excoffier, G. and Vignon, M. R. (1991). "Effect of steam explosion treatment on the physio-chemical characteristics and enzymatic hydrolysis of poplar cell wall components." Animal Feed Science and Technology 32: 235-242. Vallander, L. and Eriksson, K. (1991). "Enzymatic hydrolysis of lignocellulosic materials: II. Experimental Investigations of theoretical hydrolysis-process models for an increased enzyme recovery." Biotechnology and Bioengineering 38: 139-144. Waymann, M., Tallevi, A. and Winsborrow, B. (1984). "Hydrolysis of biomass by sulphur dioxide." Biomass: 183-191. Wong, K. K. Y., Deverell, K. F., Mackie, K. L., Clark, T. A. and Donaldson, L. A. (1988). "The relationship between fibre porosity and cellulose digestibility in steam-exploded Pinus radiata." Biotechnology and Bioengineering 31: 447-456. Wu, M. M., Chang, K., Gregg, D. J., Boussaid, A., Beatson, R. P. and Saddler, J. N. (1999). "Optimization of steam explosion to enhance hemicellulose recovery and enzymatic hydrolysis of cellulose in softwoods." Applied Biochemistry and biotechnology 77-79: 47-54. www.fao.org. Xo, C.-b. (1994). "An Improved peroxide bleaching process and its use for ECF and TCF sequences." Tappi Journal 92:88-96. Yang, B., Boussaid, A., Mansfield, S. D. and Saddler, J. N. (2002). "A fast and efficient alkaline peroxide treatment to enhance the enzymatic digestability of steam exploded softwood substrates." Biotechnology and Bioengineering 77:678-684.. Yu, A. H. C. and Saddler, J. N. (1995). "Identification of essential cellulase components in the hydrolysis of a steam-exploded birch substrate." Applied Biochemistry and biotechnology 21: 185-202. 122 APPENDIX A Composition of initial Douglas-fir chips used for the bioconversion process (expressed in grams per 100 grams substrate). Substrate Composition (g/lOOg) Arabinose 1.7 Galactose 2.7 Glucose 46.7 Xylose 4.9 Mannose 13.7 Lignin 34.0 Total 103.7 123 APPENDIX B All tables in Appendix B represent the mass balances for the various stages of the bioconversion process with substrates having 12 % initial moisture content. Mass balance following steam explosion (all values are expressed in grams per 100 grams of starting substrate). Arabinose Galactose Glucose Xyl ose Mannose Lignin WS WI WS WI WS WI WS WI WS WI WS WI VF 1.0 0.2 2.1 0.5 4.9 31.0 3.4 0.3 6.3 4.3 6.3 27.7 M 0.8 0.2 2.5 0.6 9.7 32.8 2.7 0.8 8.3 4.6 4.6 29.4 VC 0.7 0.2 2.4 0.5 9.8 33.3 2.5 0.9 7.4 4.6 2.1 31.9 Mass balance following Hot Alkali Peroxide Fractionation (all values are expressed in grams per 100 grams of starting substrate). Water-soluble sugar concentrations were calculated based on the sugar concentration removed from the water insoluble sugars, Arabinose Galactose Glucose Xyl ose Mannose Lignin WS WI WS WI WS WI WS WI WS WI WS WI VF 0.18 0.02 0.48 0.02 1.0 30.0 0.1 0.2 3.8 0.5 12.9 14.8 M 0.18 0.02 0.58 0.02 3.3 29.5 0.4 0.4 4.3 0.3 24.6 4.8 VC 0.18 0.02 0.48 0.02 6.6 26.7 0.4 0.5 4.3 0.3 28.0 3.9 Mass balance following hydrolysis (all values are expressed in grams per 100 grams of starting substrate). Arabinose Galactose Glucose Xy] ose Mannose Lignin WS WI WS WI WS WI WS WI WS WI WS WI VF 0 0.02 0 0.02 15.3 14.7 0 0.2 0 0.5 0 14.8 M 0 0.02 0 0.02 25.6 3.9 0 0.4 0 0.3 0 4.8 VC 0 0.02 0 0.02 24.1 2.6 0 0.5 0 0.3 0 3.9 Total mass balance (all values are expressed in grams per 100 grams of starting substrate). The total water soluble sugars recovered are the sum of the soluble sugars from steam explosion and hydrolysis, meanwhile the water insoluble sugars are the Arabinose Galactose Glucose Xyl ose Mannose Lignin WS WI WS WI WS WI WS WI WS WI WS WI VF 1.0 0.02 2.1 0.02 20.2 14.7 3.4 0.2 6.3 0.5 12.9 14.8 M 0.8 0.02 2.5 0.02 35.3 3.9 2.7 0.4 8.3 0.3 24.6 4.8 VC 0.7 0.02 2.4 0.02 33.9 2.6 2.5 0.5 7.4 0.3 28.0 3.9 124 APPENDIX C All tables in Appendix C represent the mass balances for the various stages of the bioconversion process with substrates having 12 % initial moisture content and refining post steam explosion. Mass balance following steam explosion (all values are expressed in grams per 100 grams of starting substrate). Arabinose Galactose Glucose Xyl ose Mannose Lignin WS WI WS WI WS WI WS WI WS WI WS WI VF 1.0 0.2 2.7 0 7.0 30.2 3.7 0 9.0 2.0 8.6 25.4 M 1.0 0.05 2.8 0.3 10.0 30.9 3.5 0 10.4 2.4 8.0 26.0 VC 0.8 0.05 2.8 0.2 8.3 31.8 2.7 0.7 7.3 2.5 4.0 30.0 Mass balance following Hot Alkali Peroxide Fractionation (all values are expressed in grams per 100 grams of starting substrate). Water-soluble sugar concentrations were calculated based on the sugar concentration removed from the water insoluble sugars, they may be present as sugar degradation products Arabinose Galactose Glucose Xy ose Mannose Lignin WS WI WS WI WS WI WS WI WS WI WS WI VF 0.18 0.02 0 0 0.3 29.9 0 0 1.5 0.5 12.7 14.1 M 0.03 0.02 0.28 0.02 0.9 30.0 0 0 2.2 0.2 22.1 3.9 VC 0.03 0.02 0.18 0.02 4.0 26.8 0.3 0.4 2.3 0.2 27.8 2.2 Mass balance following hydrolysis (all values are expressed in grams per 100 grams of starting substrate). Arabinose Galactose Glucose Xyl ose Mannose Lignin WS WI WS WI WS WI WS WI WS WI WS WI VF 0 0.02 0 0 19.9 10.0 0 0 0 0.5 0 12.7 M 0 0.02 0 0.02 25.8 4.2 0 0 0 0.2 0 3.9 VC 0 0.02 0 0.02 25.3 1.5 0 0.4 0 0.2 0 2.2 Total mass balance (all values are expressed in grams per 100 grams of starting substrate). The total water soluble sugars recovered are the sum of the soluble sugars from steam explosion and hydrolysis, meanwhile the water insoluble sugars are the remnant left after hydrolysis. Arabinose Galactose Glucose Xylose Mannose Lignin WS WI WS WI WS WI WS WI WS WI WS WI VF 1.0 0.02 2.7 0 26.9 10.0 3.7 0 9.0 0.5 12.7 12.7 M 1.0 0.02 2.8 0.02 35.8 4.2 3.5 0 10.4 0.2 22.1 3.9 VC 0.8 0.02 2.8 0.02 33.6 1.5 2.7 0.4 7.3 0.2 27.8 2.2 125 APPENDIX D All tables in Appendix D represent mass balances for the various stages of the bioconversion process with substrates having 30 % initial moisture content. Mass balance following steam explosion (all values are expressed in grams per 100 grams of starting substrate). Arabinose Galactose Glucose Xyl ose Mannose Lignin WS WI WS WI WS WI WS WI WS WI WS WI VF 0.8 0.2 1.5 0.5 4.1 39.6 3.0 0.4 6.1 5.4 2.4 31.6 M 0.7 0.2 1.9 0.5 6.3 40.4 2.5 0.9 7.6 4.9 1.1 32.9 VC 0.5 0.2 2.2 0.5 8.0 40.5 1.9 1.4 8.4 4.4 0.1 33.9 Mass balance following Hot Alkali Peroxide Fractionation (all values are expressed in grams per 100 grams of starting substrate). Water-soluble sugar concentrations were calculated based on the sugar concentration removed from the water insoluble sugars, Arabinose Galactose Glucose Xyl ose Mannose Lignin WS WI WS WI WS WI WS WI WS WI WS WI VF 0.2 0 0.44 0.06 8.8 30.8 0.3 0.1 4.5 0.9 17.4 14.2 M 0.19 0.01 0.44 0.06 5.6 34.8 0.5 0.4 4.1 0.8 23.7 6.2 VC 0.19 0.01 0.44 0.06 6.2 34.3 0.8 0.6 3.7 0.7 29.1 4.8 Mass balance following hydrolysis (all values are expressed in grams per 100 grams of starting substrate). Arabinose Galactose Glucose Xy] ose Mannose Lignin WS WI WS WI WS WI WS WI WS WI WS WI VF 0 0 0 0.06 22.0 8.8 0 0.1 0 0.9 0 14.2 M 0 0.01 0 0.06 28.6 6.2 0 0.4 0 0.8 0 6.2 VC 0 0.01 0 0.06 31.3 3.0 0 0.6 0 0.7 0 4.8 Total mass balance (all values are expressed in grams per 100 grams of starting substrate). The total water soluble sugars recovered are the sum of the soluble sugars from steam explosion and hydrolysis, meanwhile the water insoluble sugars are the remnant left after hydrolysis. Arabinose Galactose Glucose Xyl ose Mannose Lignin WS WI WS WI WS WI WS WI WS WI WS WI VF 0.8 0 1.5 0.06 26.1 8.8 3.0 0.1 6.1 0.9 17.4 14.2 M 0.7 0.01 1.9 0.06 34.9 6.2 2.5 0.4 7.6 0.8 23.7 9.2 VC 0.5 0.01 2.2 0.06 39.3 3.0 1.9 0.6 8.4 0.7 29.1 5.6 126 APPENDIX E All tables in Appendix E represent mass balances for the various stages of the bioconversion process with substrates having 30 % initial moisture content and refining post steam explosion. Mass balance following steam explosion (all values are expressed in grams per 100 grams of starting substrate). Arabinose Galactose Glucose X y l ose Mannose Lignin WS WI WS WI WS WI WS WI WS WI WS WI VF 0.8 0.1 2.2 0 4.5 31.1 3.2 0.2 9.2 2.4 10.1 23.9 M 0.7 0.1 2.3 0.2 8.7 33.0 3.2 0.2 10.8 1.7 5.2 28.8 VC 0.6 0.1 2.5 0.2 11.6 35.5 3.1 0.2 11.4 1.4 0.8 33.2 Mass balance following Hot Alkali Peroxide Fractionation (all values are expressed in grams per 100 grams of starting substrate). Water-soluble sugar concentrations were calculated based on the sugar concentration removed from the water insoluble sugars, Arabinose Galactose Glucose X y l ose Mannose Lignin WS WI WS WI WS WI WS WI WS WI WS WI VF 0.1 0 0 0 3.0 28.1 0.1 0.1 1.8 0.8 11.8 12.1 M 0.1 0 0.15 0.05 1.6 31.4 0.17 0.03 1.1 0.6 24.8 4.0 VC 0.1 0 0.15 0.05 4.3 31.2 0.17 0.03 0.6 0.6 30.7 2.5 Mass balance following hydrolysis (all values are expressed in grams per 100 grams of Arabinose Galactose Glucose X y l ose Mannose Lignin WS WI WS WI WS WI WS WI WS WI WS WI VF 0 0 0 0 21.1 7.0 0 0.1 0 0.8 0 12.1 M 0 0 0 0.05 29.7 1.7 0 0.03 0 0.6 0 4.0 VC 0 0 0 0.05 30.5 0.7 0 0.03 0 0.6 0 2.5 Total mass balance (all values are expressed in grams per 100 grams of starting substrate). The total water soluble sugars recovered are the sum of the soluble sugars from steam explosion and hydrolysis, meanwhile the water insoluble sugars are the Arabinose Galactose Glucose Xy] ose Mannose Lignin WS WI WS WI WS WI WS WI WS WI WS WI VF 0.8 0 2.2 0 25.6 7.0 3.2 0.1 9.2 0.8 11.8 12.1 M 0.7 0 2.3 0.05 38.4 1.7 3.2 0.03 10.8 0.6 24.8 4.0 VC 0.6 0 2.5 0.05 42.1 0.7 3.1 0.03 11.4 0.6 30.7 2.5 127 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0075019/manifest

Comment

Related Items