Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Pretreatment and fermentation of Douglas-fir whitewood and bark feedstocks for ethanol production Robinson, Jamie 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_2003-860701.pdf [ 10.82MB ]
Metadata
JSON: 831-1.0075111.json
JSON-LD: 831-1.0075111-ld.json
RDF/XML (Pretty): 831-1.0075111-rdf.xml
RDF/JSON: 831-1.0075111-rdf.json
Turtle: 831-1.0075111-turtle.txt
N-Triples: 831-1.0075111-rdf-ntriples.txt
Original Record: 831-1.0075111-source.json
Full Text
831-1.0075111-fulltext.txt
Citation
831-1.0075111.ris

Full Text

P R E T R E A T M E N T A N D FERMENTATION OF DOUGLAS-FIR WHITEWOOD A N D B A R K FEEDSTOCKS FOR E T H A N O L PRODUCTION.  by  JAMIE ROBINSON B . S c , University of British Columbia, 1996  A THESIS SUBMITTED IN PARTIAL F U L F I L L M E N T OF T H E REQUIREMENTS FOR THE D E G R E E OF DOCTOR OF PHILOSOPHY  in  THE F A C U L T Y OF G R A D U A T E STUDIES Department of Wood Science  We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRITISH C O L U M B I A September, 2003 © Jamie Robinson  In presenting this thesis in partial fulfillment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. 1 further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head o f my department or by his or her representatives. It is understood that copying or publication o f this thesis for financial gain shall not be allowed without my written permission.  Department of  hJVQi'*  S c f  The University of British Columbia Vancouver, Canada  Date  3cr /0  lOOS  Abstract A s part o f ongoing research to evaluate the bioconversion o f softwood feedstocks to ethanol, the feasibility of using Douglas-fir residues  was investigated.  Whitewood feedstocks  were  pretreated using SCVcatalyzed steam explosion under three pretreatment severity conditions (low, medium and  high)  with  a goal of demonstrating  efficient  fermentation  of  the  hemicellulose-rich water-soluble (WS) fraction. The chosen severity had a pronounced effect on the recovery of the hemicellulose sugars from the feedstock, as well as the proportion of monomeric sugars. Low-severity pretreatment resulted in the greatest recovery of hemicellulose sugars, with yields decreasing significantly for an increase in severity.  Reduced-severity  pretreatments also resulted in the recovery of fewer sugar decomposition products in the W S fraction,  which  Saccharomyces  translated cerevisiae.  into improved fermentation  using an  SSL-adapted strain of  In contrast, the W S fraction obtained under high-severity conditions  could not be fermented.  Douglas-fir feedstocks containing bark (10, 20, 30 and 100% w/w) were pretreated  under  medium-severity to evaluate bark's impact on bioconversion. Bark had only a minor impact on the yield of hemicellulose sugars, with no negative effect on the monomeric sugar recovery. However, bark caused a significant decrease in the soluble sugar concentration. Process-derived fermentation inhibitors (e.g., furfural, H M F ) also decreased, while naturally occurring inhibitors (e.g., lipophilic compounds) increased with greater bark loading. Despite this increase, bark had no detrimental impact on the rate of fermentation, and all hydrolysates were fermentable to high ethanol yield.  Efforts to increase the low sugar concentration in the W S fractions were evaluated, in an attempt to increase the ethanol concentration recovered following fermentation.  Increasing the sugar ii  concentration in the W S fraction by physical means resulted in decreased rates of fermentation and reduced ethanol yields, even at low concentration factors (2- to 3-fold). A s an alternative strategy, the sugar concentration in the W S fraction was augmented with carbohydrate derived from the water-insoluble cellulose component. Enzymatic hydrolysis of the cellulose component directly in the W S fraction proved unsatisfactory, due to inhibition by both carbohydrate and non-carbohydrate components.  However, supplementation (1:1) of the W S fraction with the  cellulose hydrolysate obtained separately in buffer provided improved sugar concentration, and significantly faster fermentation due to the effective dilution of inhibitors in the W S fraction. Using this approach, the initial hexose sugar concentration in the whitewood W S fraction was increased by 56%, and a final ethanol concentration of 23.4 g L " was obtained. 1  iii  Table of Contents Abstract Table of contents List of figures List of tables List of abbreviations Acknowledgments  Chapter 1: Introduction 1.1 Introduction and background 1.1.1 Advantages to fuel ethanol 1.1.2 Disadvantages to fuel ethanol 1.1.3 Alternative transportation fuels 1.1.4 Ethanol production 1.2 Choice of substrate 1.2.1 Structure and composition of wood 1.2.1.1 Cellulose 1.2.1.2 Hemicellulose 1.2.1.3 Lignin 1.2.1.4 Additional components 1.2.2 Bark 1.3 Pretreatment 1.3.1 Pretreatment strategies 1.3.2 Acid-catalysed steam explosion 1.3.3 Choice of acid catalyst 1.3.4 Severity of pretreatment 1.4 Fermentation 1.4.1 Selection of microorganism 1.4.2 Inhibition of fermentation 1.4.2.1 Process-derived inhibitors 1.4.2.2 Naturally-occurring inhibitors 1.5 Inhibition of enzymatic hydrolysis 1.5.1 Inhibition by carbohydrates 1.5.2 Inhibition by non-carbohydrates components 1.6 Research approach and objectives  Chapter 2: Materials and Methods 2.1  Feedstock and pretreatment 2.1.1 Feedstock 2.1.2 Feedstock preparation 2.1.3 Pretreatment and fractionation 2.1.4 Delignification of the water-insoluble component 2.2 Enzymatic hydrolysis 2.2.1 Enzymatic hydrolysis of the cellulose component 2.2.2 Hydrolysis of the delignified water-insoluble fraction 2.2.3 High consistency hydrolysis 2.2.4 Enzymatic hydrolysis in the water-soluble fraction 2.3 Improving the enzymatic hydrolysis in the water-soluble fraction  u  i v  v  m  xii x  x  v  v  a  1 1 3 4 6 6 8 8 9 10 11 11 11 12 14 15 16 17 19 19 21 21 24 24 24 25 26  31 31 31 31 31 32 33 33 34 35 35 36 iv  2.3.1 2.3.2 2.3.3  Influence of monomeric sugars on the enzymatic hydrolysis 36 Enzyme stability in the water-soluble fraction 36 Treatments used on the water-soluble fraction prior to hydrolysis 37 2.3.3.1 Treatment with anion exchange resin (ANEX) 37 2.3.3.2 Detoxification by overliming with calcium hydroxide (OL) 37 2.3.3.3 Extraction with ethyl acetate (EX) 38 2.3.3.4 Combined overliming and extraction (OLEX) 38 2.4 Improving the starting sugar concentration in the water-soluble fraction 38 2.4.1 Concentrating the water-soluble fractions 38 2.4.2 Supplementing water-soluble fractions with the cellulose hydrolysates .... 39 2.5 Fermentation 39 2.5.1 Culturing yeast for fermentation 39 2.5.2 Fermentation of the water-soluble fractions and cellulose hydrolysates .... 40 2.6 Analytical methodology 41 2.6.1 Analysis of the water-soluble fraction 41 2.6.1.1 Monomeric sugars 41 2.6.1.2 Furfural, H M F and Organic Acids 41 2.6.1.3 Glycerol 42 2.6.1.4 Ethanol 42 2.6.1.5 Oligomeric sugars 42 2.6.1.6 Lipophilic compounds 43 2.6.1.7 Total phenolics 43 2.6.2 Feedstock analysis 44 2.6.3 Assays 45 2.6.3.1 Total cellulase activity 45 2.6.3.2 B-glucosidase enzyme activity 45 2.6.3.3 Xylanase activity 46 2.6.3.4 Mannanase activity 46 2.6.3.5 Hydrogen peroxide assay 46  Chapter 3: Pretreatment of Douglas-fir Whitewood and Bark Chips by Steam Explosion 3.1 Background 3.2 Results and discussion 3.2.1 Pretreatment of Douglas-fir whitewood 3.2.1.1 Feedstock composition 3.2.1.2 Solids recovery (shot yield) 3.2.1.3 Soluble sugar recovery 3.2.1.4 Monomeric sugar recovery 3.2.1.5 Composition of the water-soluble fraction 3.2.2 Pretreatment of Douglas-fir whitewood feedstocks containing bark 3.2.2.1 Feedstock composition 3.2.2.2 Solids recovery (shot yield) 3.2.2.3 Soluble sugar recovery 3.2.2.4 Composition of the water-soluble fraction 3.2.2.5 Monomeric sugar recovery 3.2.2.6 Inhibitor recovery 3.3 Conclusions  48 48 50 50 50 53 55 58 60 63 63 64 68 71 75 77 81  v  C h a p t e r 4: Fermentation of the Water-Soluble F r a c t i o n to E t h a n o l  84  4.1 Background 84 4.2 Results and discussion 85 4.2.1 Yeast strain selection 85 4.2.1.1 Hexose utilization by the T l and W T yeast strains 86 4.2.1.2 Ethanol production 91 4.2.2 Pretreatment severity and fermentation 92 4.2.3 Bark content and fermentation 97 4.2.3.1 Influence of bark on hexose sugar consumption 99 4.2.3.2 Influence of bark on ethanol production 104 4.2.4 Pretreatment severity and fermentability of bark-derived WS fractions ..107 4.2.5 Metabolism of H M F and furfural during fermentation 109 4.3 Conclusions 112 C h a p t e r 5: Increasing the Sugar Concentration i n the W a t e r - S o l u b l e F r a c t i o n  5.1 Background 5.2 Results and discussion 5.2.1 Concentrating the water-soluble fractions 5.2.1.1 Feedstock composition and pretreatment 5.2.1.2 Fermentation of the 2-fold water-soluble concentrates 5.2.1.3 Fermentation of the 2- to 3- fold concentrates (WS-30) 5.2.1.4 Final observations 5.2.2 Supplementation of the water-soluble fraction with carbohydrate 5.2.2.1 Delignification and enzymatic hydrolysis of the waterinsoluble component 5.2.2.2 Loss of cellulose by delignification 5.2.2.3 Enzymatic hydrolysis of the original and delignified water-insoluble components 5.2.2.4 Hydrolysis of the water-insoluble component directly in the water-soluble fraction 5.2.2.5 Inhibition of enzymatic hydrolysis a. Simulated water-soluble fraction b. Monomeric sugars c. Non-carbohydrate components 5.2.2.6 Detoxification of the water-soluble fraction 5.2.3 Supplementation of the water-soluble fraction with the cellulose hydrolysate 5.2.3.1 Hydrolysis of the water-insoluble component at 5% consistency 5.2.3.2 Fermentation of the cellulose hydrolysate 5.2.3.3 Combining the two streams 5.3 Conclusions Chapter 6: Conclusions a n d F u t u r e W o r k  6.1 General conclusions 6.2 Future work 6.2.1 Improved recovery of soluble sugars during fractionation 6.2.2 Galactose fermentation  114  114 116 116 116 120 128 134 136 136 137 139 141 148 150 154 159 163 169 170 173 173 180 183  183 185 186 186  vi  6.2.3 6.2.4  Selection of cellulolytic enzymes exhibiting reduced end-product inhibition Simultaneous saccharification and fermentation (SSF)  186 187  References  188  Appendices  205  List of Figures Figure 3-1: Monomeric sugar concentration (g L" ) in the water washes (performed at 20% w/v consistency) obtained from the water-insoluble components WI-0 and WI-30 after an initial fractionation  72  Figure 3-2: Monomeric sugar concentration (g L" ) in the water-soluble fractions derived from Douglas-fir whitewood/bark chip mixtures (containing 0, 10, 20, 30 or 100% bark by weight) pretreated under medium-severity conditions  74  Figure 3-3: Extractive composition of the Douglas-fir bark feedstock (100% bark) before and after steam explosion under medium-severity conditions. Samples were Soxhlet-extracted sequentially with hexane, dichloromethane ( D C M ) , ethyl acetate (EtOAc), acetone, methanol ( M e O H ) and water  80  Figure 4-1: Comparison of (A) glucose uptake, (B) mannose uptake, (C) galactose uptake and (D) combined hexose uptake by the Tembec SSL-adapted Saccharomyces cerevisiae strain T l and a wild-type S. cerevisiae strain, during fermentation of the water-soluble fraction derived from Douglas-fir whitewood prepared under medium-severity pretreatment, and synthetic medium (1% yeast extract, 1% peptone containing comparable sugar concentrations to the water-soluble fraction)  87  Figure 4-2: Hexose sugar consumption and ethanol production during fermentation of the Douglas-fir whitewood water-soluble fractions obtained after pretreatment at various severity conditions. (A) Low-severity; (B) Medium-severity; (C) High-severity; (D) Ethanol production  93  Figure 4-3: Sugar consumption during fermentation of the water-soluble fractions derived from Douglas-fir whitewood feedstocks containing 0-100% supplemental bark. (A) W S - 0 ; (B) WS-10; (C) WS-20; (D) W S - 3 0 ; (E) WS-100. Note: due to the very low sugar concentration in WS-100, a different scale is used (Figure E )  100  Figure 4-4: Sugar consumption and ethanol production during fermentation of the watersoluble fractions derived from Douglas-fir whitewood/bark feedstocks pretreated under medium-severity conditions. ( A ) Combined hexose sugar consumption; (B) Ethanol production. Legend applies to both figures  101  Figure 4-5: Fermentation of the water-soluble fractions W S - 0 , W S - 1 0 0 and WS-100S (WS-100 supplemented with hexose sugars to achieve a sugar concentration comparable to WS-0). A reference fermentation (medium containing glucose as the sole fermentable sugar) was also used. ( A ) Hexose sugar consumption; (B) Ethanol production  103  1  1  viii  Figure 4-6: Fermentation of the water-soluble fractions ( W S - 3 0 L o w and WS-30) derived from steam-exploded whitewood/bark feedstock (30% w/w bark) under lowand medium-severity pretreatment conditions. The initial sugar concentration in W S - 3 0 L o w was adjusted with galactose, glucose and mannose to achieve a comparable concentration to fraction W S - 3 0 M e d . (A) Ffexose consumption; (B) Ethanol production  110  Figure 4-7: Metabolism of H M F during fermentation of the water-soluble fractions derived from Douglas-fir whitewood/bark feedstocks (0, 10, 20, 30 or 100% bark by weight) pretreated under medium-severity  Ill  Figure 5-1: Process flow diagram for concentrating the water-soluble fractions derived from Douglas-fir whitewood (DF-0) and mixed whitewood/bark (DF-30), illustrating the key steps of pretreatment, fractionation, and concentration by rotary-evaporation ( R V ) or freeze-drying (FD), as required  118  Figure 5-2: Fermentation of the original and 2-fold concentrated water-soluble fractions derived from (A) Douglas-fir whitewood chips and (B) whitewood chips supplemented with 30% bark  122  Figure 5-3: Fermentation of the concentrates obtained by rotary-evaporation of the watersoluble fraction W S - 3 0 to 3-fold, followed by dilution to 2.75-, 2.5-, 2.25-, 2-, and 1-fold. Monomeric sugars were supplemented to each fraction to achieve a comparable starting sugar concentration. A simulated water-soluble fraction (Sim. W S , containing sugars in water) and reference fermentation ( G M Y P medium) were also used. Legend applies to both figures  129  Figure 5-4: Enzymatic hydrolysis of the water-insoluble component (2% consistency in A C buffer) derived from steam exploded Douglas-fir whitewood (WI-0) and whitewood/bark feedstocks (WI-30) pretreated under medium-severity conditions, with or without delignification by alkaline peroxide ( A P ) treatment  140  Figure 5-5: Process flow diagram for supplementing the water-soluble fractions derived from Douglas-fir whitewood (DF-0) with the delignified cellulose component (WI-AP-0) prior to enzymatic hydrolysis and fermentation. The key steps of pretreatment, fractionation, delignification, enzymatic hydrolysis and fermentation are illustrated  143  Figure 5-6: Hydrolysis of oligomeric sugars derived from the water-soluble fraction W S - 0 during enzymatic hydrolysis in the water-soluble fraction. Reaction flasks contained no water-insoluble material, but were supplemented with the equivalent enzyme loading  146  ix  Figure 5-7: Enzymatic hydrolysis of the delignified, water-insoluble component ( W I - A P 0) at 5% (w/v) consistency. (A) Hydrolysis in the water-soluble fraction ( W S 0, non-buffered, p H 4.8), acetate buffer ( A C , 50 m M , p H 4.8), or a simulated water-soluble fraction ( S - A C , 50 m M , p H 4.8) at 10 F P U g-cellulose. (B) Hydrolysis in the W S - 0 or A C buffer at 10, 20 or 30 F P U g-cellulose-1. A l l hydrolyses were supplemented with 6-glucosidase to achieve a 3:1 I U to F P U ratio  147  Figure 5-8: The influence of 6-glucosidase loading (0, 30, 60 or 90 I U g-cellulose" ) on the enzymatic hydrolysis of the delignified, water-insoluble component ( W I - A P ) in S - A C buffer. Celluclast was added to an activity of 10 F P U g-cellulose"' in all reactions. Control hydrolyses were conducted in A C buffer...  152  Figure 5-9: Enzymatic hydrolysis of the Hemlock kraft pulp (2% w/v consistency) in acetate buffer supplemented with individual monomeric sugars (arabinose, galactose, glucose, mannose or xylose) to achieve a concentration of (A) 10 g L " , (B) 20 g L " , or (C) 40 g L ' at the start of hydrolysis  157  Figure 5-10: The effect of extended hydrolysis time on the glucose yield from hemlock pulp enzymatically hydrolysed in the water-soluble fraction (WS-0)  161  Figure 5-11: Decrease in the hydrolytic power of the enzyme preparation (Celluclast and Novozym-188) after prolonged pre-incubation in the water-soluble fraction (WS-0), acetate buffer ( A C ) , or a simulated water-soluble fraction ( S - A C ) prior to hydrolysis of the Hemlock kraft pulp in A C buffer. (A) Hydrolysis yield. (B) Hydrolysis yield expressed as a percentage of the hydrolysis in AC-buffer  162  Figure 5-12: The influence of ethyl acetate extraction on the enzymatic hydrolysis of the hemlock kraft pulp in the water-soluble fraction. Hydrolyses were performed in the untreated water-soluble fraction (WS-0), the extracted water-soluble fraction ( W S - 0 - E X ) , sugar-acetate buffer ( S - A C ) , extracted sugar-acetate buffer ( S - A C - E X ) , acetate buffer ( A C ) and acetate buffer supplemented with the ethyl acetate extracted component  167  Figure 5-13: The influence of overliming and combined overliming/extraction on the enzymatic hydrolysis of the hemlock kraft pulp in the water-soluble fraction. Hydrolyses were performed in the untreated water-soluble fraction (WS-0), the overlimed water-soluble fraction (OL), the overlimed/extracted watersoluble fraction ( O L E X ) , sugar-acetate buffer ( S - A C ) , and acetate buffer (AC)  168  Figure 5-14: Process flow diagram for supplementing the water-soluble fractions derived from Douglas-fir whitewood (DF-0) and mixed whitewood/bark (DF-30) (with and without concentration by rotary-evaporation) with the cellulose hydrolysate ( C H ) derived from the delignified water-insoluble component (WI-AP)  171  1  1  1  1  x  Figure 5-15: Enzymatic hydrolysis (5% consistency in A C buffer) of the alkaline peroxide (AP) treated water-insoluble component derived from Douglas-fir whitewood (WI-AP-0) and whitewood/bark feedstock (WI-AP-30) pretreated under medium-severity conditions. Closed symbols, glucose yield; Open symbols, glucose concentration  172  Figure 5-16: Fermentation of the cellulose hydrolysates, C H - 0 and C H - 3 0 , derived from the enzymatically hydrolysed whitewood (WI-AP-0) and whitewood/bark water-insoluble components (WI-AP-30), respectively. Monomeric sugars were supplemented to C H - 3 0 to achieve a comparable sugar concentration to C H - 0 . Solid symbols, hexose sugars; Open symbols, ethanol  174  Figure 5-17: Fermentation of the water-soluble fractions after supplementation with the cellulose hydrolysate. (A) Whitewood-derived water-soluble fraction and cellulose hydrolysate. (B) Whitewood/bark-derived water-soluble fraction and cellulose hydrolysate. A l l bark-derived hydrolysates were supplemented with monomeric sugars to be comparable to the equivalent whitewood hydrolysate. Closed symbols, hexose sugars; Open symbols, ethanol  176  Figure 5-18: Fermentation of the water-soluble fraction (WS-0), the cellulose hydrolysate (CH-0) obtained from a 10% (w/v) consistency hydrolysis, and the watersoluble fraction after supplementation with the cellulose hydrolysate  179  xi  List of Tables Table 3-1: Chemical composition (%) of the original Douglas-fir whitewood feedstock used in the pretreatment severity experiments  51  Table 3-2: Shot yield recovery of the feedstock (%, based on oven-dried weight) following pretreatment of the whitewood chip feedstock under low-, medium-, and high-severity  54  Table 3-3: Recovery of original hemicellulose sugars and glucose in the water-soluble fractions (expressed as a percentage of the available sugars in the original wood feedstock) under low-, medium- and high-severity pretreatment  56  Table 3.4: Percentage of monomeric hemicellulose sugar content contained in the Douglas-fir water-soluble fractions prepared under low-, medium- and highseverity pretreatment conditions  59  Table 3.5: Composition of the water-soluble fractions generated from steam-exploded Douglas-fir under low-, medium-, and high-severity pretreatment, in terms of sugars and sugar-degradation products (g L" )  62  Table 3-6: Chemical composition of the Douglas-fir whitewood and bark feedstocks (%)...  65  Table 3-7: Shot yield recovery of the feedstock (%, based on oven-dried weight) following pretreatment of whitewood/bark chip feedstocks under mediumseverity conditions  66  Table 3-8: Carbohydrate (combined oligomeric and monomeric) recovered from the 0%bark and 30%-bark feedstocks following pretreatment and fractionation (% of available sugar in the feedstock)  69  Table 3-9: Monomeric sugar concentration (g L ' ) in the water-soluble fractions generated from steam-exploded Douglas-fir whitewood/bark feedstocks under mediumseverity conditions  73  Table 3-10: Oligomeric sugar content (% total soluble carbohydrate) in the Douglas-fir water-soluble fractions prepared from whitewood and whitewood/bark feedstocks pretreated under medium-severity conditions  76  Table 3-11: Inhibitor concentration (g L" ) in the water-soluble fractions prepared from steam-exploded Douglas-fir whitewood/bark feedstocks under mediumseverity pretreatment, in terms of sugar-degradation products, lipophilic compounds, and phenolic compounds  78  Table 4-1: Sugar uptake rate by the wild-type (WT) or Tembec SSL-adapted ( T l ) strains of Saccharomyces cerevisiae, during fermentation of either the Douglas-fir water-soluble fraction derived from whitewood pretreated under mediumseverity, or synthetic medium containing comparable sugar concentrations  89  1  1  1  xii  Table 4-2: Ethanol and glycerol production during fermentation of the three water-soluble fractions obtained after pretreatment under low-, medium-, and high-severity conditions  96  Table 4-3: Fermentation of the water-soluble fractions derived from Douglas-fir whitewood/bark feedstocks (containing 0, 10, 20, 30, or 100% bark by weight) pretreated under medium-severity conditions  102  Table 4-4: Ethanol and glycerol production during fermentation of the water-soluble fractions derived from Douglas-fir whitewood/bark feedstocks (containing 0, 10, 20, 30, or 100% bark by weight) pretreated under medium-severity conditions  105  Table 4-5: Composition of the water-soluble fraction W S - 3 0 L o w derived from a whitewood feedstock containing 30% bark by weight, pretreated under lowseverity conditions  108  Table 5-1: Composition (%) of the original Douglas-fir wood and bark feedstocks used in concentration experiments  117  Table 5-2: Chemical composition of the water-soluble fractions (g L" ) after steam explosion of Douglas-fir whitewood chips (WS-0) and whitewood chips containing 30% bark (WS-30)  120  Table 5-3: Summary of the fermentations employing the original and 2-fold concentrated water-soluble fractions (obtained by rotary-evaporation and freeze-drying)  123  1  Table 5-4: Concentration of inhibitors in a synthetic cocktail following 2-fold concentration by rotary-evaporation or freeze-drying. Inhibitors were present in the original cocktail at a concentration of 1 g L " 1  125  Table 5-5: Concentration of inhibitors (g L" ) in the original water-soluble fractions (WS-0 and WS-30) and the 2-fold concentrates obtained by rotary-evaporation ( R V ) and freeze-drying (FD)  127  Table 5-6: Ethanol yield from the water-soluble fraction W S - 3 0 concentrated 1-, 2-, 2.25-, 2.5-, 2.75-, and 3-fold by rotary-evaporation prior to fermentation  133  Table 5-7: Composition (%) of the Douglas-fir whitewood- and whitewood/bark-derived water-insoluble components following steam explosion only (WI-0 and W I 30), or following steam explosion and delignification by alkaline peroxide (AP) treatment ( W I - A P - 0 and WI-AP-30)  138  Table 5-8: Composition (%) of the 0 -delignified Hemlock kraft pulp used for selected enzymatic hydrolysis experiments  156  1  2  xiii  Table 5-9: Reduction in the hydrolysis yield after 24 hours (%, relative to control hydrolyses performed in A C buffer) during hydrolysis of the CVdelignified Hemlock kraft pulp in A C buffer supplemented with individual monomeric sugars (Arabinose, Galactose, Glucose, Mannose or Xylose)  158  Table 5-10: Summary of the fermentations of the original water-soluble fractions (WS-0 and WS-30) and 2-fold concentrates ( R V - 0 and R V - 3 0 ) with and without supplementation with the cellulose hydrolysate ( C H - 0 and CH-30). A l l barkderived hydrolysates were supplemented with monomeric sugars to be comparable to the equivalent whitewood hydrolysate  177  xiv  List of Abbreviations °C AC AIL ANEX AP ASL BTEX CH CH-0 CH-30 CH-30S DCM DF-0 DF-30 dH20 DNS DOE EtOAc EX FD FD-0 FD-30 FPU g g GC GMYP h H2SO4 HemSug HMF HPLC IU kg L LCC M MeOH mg min mL mm mM MMT Mt MTBE N  degrees Celsius sodium acetate buffer (50mM, p H 4.8) acid insoluble lignin anion exchange treatment (for water-soluble fraction) alkaline peroxide treatment (delignification) acid soluble lignin Benzene, Toluene, Ethylbenzene, Xylene cellulose hydrolysate cellulose hydrolysate derived from W I - A P - 0 cellulose hydrolysate derived from W I - A P - 3 0 cellulose hydrolysate C H - 3 0 supplemented with monomeric sugars dichloromethane Douglas-fir feedstock containing whitewood only Douglas-fir feedstock containing 30% bark by weight distilled water dinitrosalicylic acid Department of Environment ( U S A ) ethyl acetate ethyl acetate extraction treatment (for water-soluble fraction) freeze-drying 2-fold concentrate of W S - 0 by freeze-drying 2-fold concentrate of W S - 3 0 by freeze-drying filter paper unit grams acceleration due to gravity (i.e., 5000 x g) gas chromatography culture medium (1% glucose, 0.5% malt extract, 0.3% yeast extract, 0.5% peptone) hour sulphuric acid hemicellulose sugars (i.e., arabinose, galactose, mannose and xylose) 5-hydroxymethylfurfural high performance (pressure) liquid chromatography international unit kilogram litre lignin carbohydrate complexes molar concentration (moles per L ) methanol milligram minute millilitre millimetre millimolar concentration methylcyclopentadienyl manganese tricarbonyl megatonnes methyl tertiary-butyl ether normal (concentration) xv  NAD+ nm NO NREL OD ODW  nicotinamide adenine dinucleotide (oxidized state) nanometre Nitrogen oxides National Renewable Energy Laboratory ( U S A ) optical density oven-dried weight  OL OLEX PAD RO RPM RV RV-0 RV-30 s S-AC Sim-WS SO2 SSF SSL t T ug uL um v/v v/w VOC w/v w/w  overliming treatment with calcium hydroxide (for water-soluble fraction) combined overliming and extraction treatment (for water-soluble fraction) pulsed amperometric detector pretreatment severity factor revolutions per minute rotary-evaporation 2-fold concentrate of W S - 0 by rotary evaporation 2-fold concentrate of W S - 3 0 by rotary evaporation second sugar acetate buffer ( 5 0 m M , p H 4.8, containing monomeric sugars) simulated water-soluble fraction, containing sugars in water sulphur dioxide simultaneous saccharification and fermentation spent-sulphite liquor time temperature microgram microlitre micrometre volume per volume volume per weight volatile organic compounds weight per volume weight per weight  WI WI-0 WI-30 WI-AP-0 WI-AP-30 WS WS-0 WS-30 WT  water-insoluble component following fractionation (i.e., residual cellulose/lignin) water-insoluble component, 0% bark (i.e., whitewood) water-insoluble component, 30% bark delignified (alkaline-peroxide) Douglas-fir water-insoluble component, 0% bark delignified (alkaline peroxide) Douglas-fir water-insoluble component, 30% bark water-soluble component following fractionation water-soluble component following fractionation, 0% bark water-soluble component following fractionation, 30% bark wild-type, referring to the wild-type designated Saccharomyces cerevisiae  x  xvi  Acknowledgments  There are countless individuals with whom I have had the privilege of working over the course of this degree, and I am indebted to them all. First and foremost, I would like to thank my supervisor, D r . Jack Saddler, for giving me the opportunity to pursue research in a field I find so both challenging and worthwhile, and for teaching me how to prioritize my research goals. I am certain I would still be working down yet another tangent were it not for his objective advice. I was also very fortunate to have worked closely with Dr. Shawn Mansfield. He was, in essence, a second supervisor. H i s unique motivational methods (i.e., choke-holds and threats) were compelling enough to convince me to sacrifice evenings and weekends to science. The discipline and work ethic he taught me w i l l prove invaluable. H e was always available to address my many research concerns, to critically review a manuscript on short notice, or to fix a broken H P L C , even when his own deadlines were imminent. I am further grateful for the support of my colleagues, the students and post-doctoral fellows of Forest Products Biotechnology, past and present, for making the lab such an interesting "home away from home". They were there to discuss research, to sympathize with over numerous issues in the lab, and I enjoyed our camaraderie. Special thanks must be made to Dr. Abdel Boussaid, who mentored me in the methods of bioconversion during his time at F P B , and helped me define the objectives of my research project. M y initial immersion in bioconversion would not have been as rewarding without his positive attitude and encouragement. Last, but certainly not least, I must acknowledge the encouragement and love of family and friends alike, who remained supportive despite my spending too much time in the lab to see them as often as I should have. I would like to express my sincere gratitude to Barbra Guiao, without whom I could not have completed this process. She was there, day in and day out, listening patiently to my problems, helping me through the stressful times, and celebrating during the good. She provided undying support for me over the years. Barb, you have taught me to keep things in perspective, and that there is life outside the lab. For this, I am eternally grateful. Thank you. Personal funding and funding for the research project was provided by the Science Council of British Columbia, and Natural Resources Canada. Their contributions are greatly appreciated.  xvii  Chapter 1. Introduction 1.1 Introduction and Background Arguably, since the initial discovery in the early 19 century that lignocellulosic biomass could th  be broken down into its component  carbohydrate  and lignin fractions  through chemical  treatments, there has been an interest in developing efficient and economical commercial processes exploiting this fact.  Lignocellulosic materials were recognized for their potential to  serve as inexpensive and plentiful resources for the production of chemicals, and as feedstocks for chemical and biochemical conversions to valuable specialty and commodity products such as ethanol. Although a concentrated sulphuric acid process for the hydrolysis of cellulose was first described in 1819 by Braconnot (Scherrard and Kressman, 1945), and a dilute acid process in 1888 (Parisi, 1989), it was not until 1898 that the commercial hydrolysis of lignocellulosics to sugar was first realized (Scherrard and Kressman, 1945). However, the German-designed dilute acid process achieved only low ethanol yields (76 L ethanol per tonne of biomass).  The  collective research efforts of numerous laboratories culminated in the development of the vastly improved Scholler and Madison processes for wood hydrolysis in the 1930s and respectively.  1940s,  These processes employed high temperature, dilute acid percolation for the  hydrolysis of cellulose to glucose. Both processes, and similar ones based on concentrated acid hydrolysis, were not without their problems, but the growing demand for chemicals and ethanol during W o r l d W a r II made bioconversion technology a commercially attractive proposition.  Following the war, the majority of the lignocellulosic bioconversion processes fell out of favour due to their relatively poor process economics.  It was not until the much later, during the oil  crisis of the 1970s, that the bioconversion "renaissance" was experienced.  Lowered global oil  production meant dramatically increased costs of crude o i l , and ultimately the cost of petroleum skyrocketed. Consequently, interest in alternative fuel sources was renewed.  Economically, it 1  made sense to investigate alternative fuels such as ethanol, propane and natural gas, as these products could reduce the cost of expensive o i l imports.  In the United States, political  motivation was also instrumental in renewing interest in alternative energies.  A n alternative,  domestic fuel source could alleviate dependence on foreign o i l reserves, which are typically obtained from politically unstable regions in the world (e.g., the Middle East, Africa and South America) (Rowland, 1997; Sheehan, 2001).  Currently, the major impetus for establishing lignocellulose-to-ethanol conversion processes stems from environmental concerns over the pollution and emissions from the combustion of fossil fuels. Perhaps the greatest environmental "threat" imposed by the burning of fossil fuels is the emission of carbon dioxide (CO2) and CGVequivalents. Carbon dioxide is a greenhouse gas, and rampant CO2 production from the burning of fossil fuels over the last century has been implicated in global warming. Although not nearly as toxic as other chemicals derived from fossil fuels, increased atmospheric CO2 concentrations stand to have a dramatic effect on the global climate. W i t h the 1997 adoption of the Kyoto protocol to reduce global greenhouse gas emissions, and Canada's recent ratification of the protocol (December 2002), it has grown increasingly apparent that the use of renewable energy sources such as ethanol derived from biomass could aid in reducing the net output of CO2. The combustion of lignocellulosic ethanol results in only very low net production of CO2 when the total fuel (Bailey, 1996), owing to the fact that the produced CO2 is reused in the generation of new biomass. In contrast, fossil fuels such as oil and coal are ordinarily sequestered from and do not participate in the global carbon cycle.  This is true until they are recovered and burned as a fuel source, whereupon they  constitute a significant net gain in environmentally available atmospheric carbon.  2  For Canada to meet its target for C O 2 and CCVequivalent emissions reductions under the Kyoto protocol (6% below the 1990 CCVequivalents production level of 601 M t ) , current consumption patterns must be altered.  fuel  Based on the 1997 projections for greenhouse gas  production in Canada in 2010, meeting this target w i l l actually require a reduction in C O 2 emissions of 20-25% (Environment Canada, 2001). One of the greatest opportunities in Canada for reducing emissions of C02-equivalents is the transportation sector, which accounted for 26.2%  of Canada's greenhouse gas emissions in 2000 (Olsen et al,  2002).  B y completely  replacing gasoline with renewable ethanol, a significant reduction in greenhouse gas production could be realized. However, this goal would be difficult to realize. A more modest ethanolgasoline blend (e.g., 10-20% ethanol) would still contribute significantly to improved air quality, while helping Canada meet its Kyoto protocol commitments.  Fuel ethanol has additional  benefits over fossil fuels, as w i l l be discussed in the following section.  1.1.1  A d v a n t a g e s to Fuel E t h a n o l  Ethanol is generally recognised to be a cleaner burning fuel than gasoline (Bailey, 1996). Consequently, the use of ethanol as a supplement to, or replacement for, gasoline in urban centres could dramatically reduce air pollution from various emissions. This is largely due to ethanol's increased oxygen content, which results in the more complete combustion of the fuel. This in turn contributes to a sizeable reduction in carbon monoxide (CO) production, and exhaust volatile organic compounds ( V O C ) (Bailey, 1996).  Furthermore, the combustion of  ethanol-blended fuels can result i n decreased nitrogen oxide ( N O ) production, due to the x  reduced use of aromatics, olefins and sulphur-containing compounds within the gasoline portion of the fuel ( M c C l o y and O'Connor, 1999). Reductions in N O emissions are important to air x  quality, as these emissions are involved in the production of ground-level ozone, the main component of urban smog.  Ground-level ozone is a powerful irritant that has been linked to 3  respiratory disorders such as asthma, and is formed through the chemical reaction of N O with x  V O C , which themselves may be toxic.  Although V O C may also come from natural sources,  non-combusted hydrocarbons in vehicle exhaust are normally a greater concern in urban centres.  Health hazards related to petroleum use are posed not only by the fuel itself, but from potentially carcinogenic oxygenates such as benzene, toluene, ethylbenzene and xylene [i.e., B T E X compounds), which may be supplemented to a given fuel to enhance its octane rating (McCloy and O'Connor, 1998; Poltak and Grumet, 2001).  Additional supplements include M M T  (methylcyclopentadienyl manganese tricarbonyl), used extensively in Canada since 1977 when unleaded gasoline was first introduced, and M T B E (methyl tertiary butyl ether), which is the most common oxygenate used in the US for the production of reformulated gasoline (RFG). These additives also have associated health risks. Manganese is a recognized neurotoxin, and although the relative environmental toxicity and health risk posed by inhaled manganese oxides generated by the combustion of M M T appears to be small (Wood and Egyed, 1994), the full impact of M M T on human health requires further study. M T B E has been implicated in ground and surface water contamination in the US (Hathaway and Hawkins, 1999), and is classified as a potential carcinogen by the Environmental Protection Agency.  These compounds could be  replaced by ethanol, resulting in a cleaner burning fuel with an improved octane rating.  1.1.2  Disadvantages to Fuel Ethanol  Although there are definite advantages to using ethanol over gasoline, fuel ethanol is not perfect, nor is it necessarily the ideal fuel for all situations. As far as transportation is concerned, most vehicles can operate on 10-15% ethanol-blends (Bailey, 1996). Newer vehicles may run on up to 20% ethanol-blended fuel without requiring any modification to the engine or exhaust system. However, using fuel with a higher ethanol concentration (which may be required to achieve 4  Canada's Kyoto commitments) may require costly engine modifications and upgrades  to  optimize compression and emission control systems (Galbe and Zacchi, 2002). Ethanol also has a one-third lower caloric content compared to gasoline, meaning that the vehicle driving range is decreased when using pure ethanol (Bailey, 1996). This reduction i n energy per unit volume is partly offset by ethanol's higher octane rating, which provides a 15% increase in energy efficiency in comparison to gasoline when used in optimized spark-ignition engines (Bailey, 1996). However, only a very small decrease in driving range should be expected for ethanol blends in the 10-15% range.  Although ethanol-blended fuel may improve combustion and exhaust emissions, it can result in increased evaporative V O C emissions. This occurs because, although ethanol has a low vapour pressure, when mixed with gasoline it forms azeotropes with greater vapour pressure than either the ethanol or the gasoline/fuel additives (Poltak and Grumet, 2001).  However, it has been  asserted that, for most areas in Canada, the reduction in exhaust V O C emissions w i l l more than offset the increase in evaporative V O C emissions ( M c C l o y and O'Connor, 1998).  The  combustion of ethanol can also result in the formation of reactive aldehydes such as the carcinogens acetaldehyde and formaldehyde (Galbe and Zacchi, 2002; Hathaway and Hawkins, 1999;  Sheehan, 2001), which can cause lung irritation and contribute to breathing disorders.  This problem can be largely countered through the use of appropriate exhaust control systems, but remains a consideration.  Ethanol is also hygroscopic, and can result in increased water  contamination in vehicle fuel systems.  Related to this, spilled ethanol-blended gasoline may  contaminate ground- and surface-water more readily than gasoline-only fuel, due to increased solubility o f the ethanol, and its high solvency for many of the compounds present in gasoline (e.g., B T E X compounds) (Poltak and Grumet, 2001).  5  1.1.3  Alternative transportation fuels  There are numerous alternatives to ethanol for use as a transportation fuel.  Briefly, two of the  common commercial alternatives include propane and compressed natural gas. These fuels are cleaner burning than petroleum-based fuels, but require specific modifications to the engine, and would not replace ethanol as an oxygenate for petroleum-based fuels.  Furthermore, because  these gases are fossil fuels, they offer no benefit for greenhouse gas reductions.  1.1.4  Ethanol Production  The ethanol required for transportation fuel could be obtained from two unique processes: chemical synthesis and fermentation. The commercial synthesis route employs either the direct or indirect hydration of ethylene (Kniel et al,  1980).  This is a relatively efficient and  economical procedure, depending on the cost of ethylene, and as recently as the 1980s accounted for more than 95% of the industrial ethanol production in the U S (Kniel et al, 1980). However, ethanol derived in this manner does not constitute a CCVneutral fuel, as the ethylene is derived from fossil fuel sources. T o enact a reduction in CO2 emissions, the ethanol must be obtained from a renewable resource, harvested in a sustainable manner. Fermentation of biomass-derived carbohydrates could satisfy this requirement.  However, most ethanologenic fermentative  processes utilize only the starch component of biomass, failing to make use of the non-storage, structural component of the substrate. This is true of the potable alcohol industry, which uses starch and sugar derived from crops such as corn, wheat and barley for the production of ethanol, but also of the current fuel ethanol production in Canada, the U S and Brazil.  Although this  conversion process is relatively cheap and easy, such fermentations utilize only a minor fraction of the total carbohydrate available from the plant. The employment of the structural components of lignocellulosics would provide significantly increased quantities of sugar for fermentation.  6  The potential to use lignocellulosic wastes for the production o f fuel ethanol is particularly attractive.  Agricultural residues (e.g., wheat/rice straw, corn fibre/stover, sugar cane bagasse)  (Asghari et al, 1996; Bura et al, 2002; Wayman et al., 1984), municipal solid wastes (e.g., waste paper, building materials) (Clausen and Gaddy, 1987; Shleser, 1994) and forestry wastes (e.g., sawdust, mill shavings, slabs, reject chips and boards) have been suggested as potential lignocellulosic feedstocks for bioethanol production. Forest-derived wastes and residues are a particularly attractive feedstock for fuel ethanol production, due to their plentiful supply. In British Columbia, there is a sizeable quantity of forestry wastes generated yearly, with current estimates indicating approximately 2,249,000 bone-dry tonnes of residues available ( M c C l o y and Associates, 2003), o f which approximately one-half is estimated to be whitewood (Hatton, 1999). Use of these residues could alleviate disposal problems associated with landfilling (e.g., toxic leachate) and incineration (e.g., air pollution), while at the same time producing a valueadded product from a waste.  The potential to use forest "thinnings", removed either as a fire-  prevention technique (Nguyen et ai, 1998, 1999) or from intensively managed forests ( M c C l o y and O'Connor, 1998), has also been recognized.  These sources o f carbohydrate may have  limited application to other forestry applications (e.g., pulping, engineered wood products) due to chemical and structural properties.  Although British Columbia has a proven supply of fibre for bioconversion, the commercial realization of ethanol from wood is burdened by technical difficulties. While there has been a significant amount of work dedicated to improving product yields and the overall economics of ethanol production from lignocellulosics over the last 30 years, considerable research is still needed.  The production of ethanol from lignocellulosics involves numerous, costly steps,  including pretreatment, fractionation, hydrolysis and fermentation (Gregg and Saddler, 1995; Gregg et ai,  1998).  This fact, combined with the relatively high contribution of the 7  lignocellulosic feedstock to overall production costs (Nguyen and Saddler, 1991; von Si vers and Zacchi, 1995; Zacchi et al, 1988), and the relatively low price that a commodity product such as fuel ethanol demands, make for narrow profit margins.  Although government subsidies and  incentives may make bioethanol more attractive to commercial ventures, only through a refining of the process w i l l ethanol's competitiveness with fossil fuels be realized.  1.2 Choice of substrate A s mentioned earlier, a number of lignocellulosics have been explored for use as feedstocks for bioethanol production. The current research focuses on wood residues, and specifically, chips derived from the softwood, Douglas-fir (Pseudotsuga menziesii). recalcitrant  to  In general, softwoods are more  bioconversion strategies including pretreatment and  hydrolysis than  are  hardwoods and agricultural residues, and considerable research is required to advance the technology for using softwoods residues of all species.  Douglas-fir was selected as a model  softwood substrate, due in part to its regional availability. It is a commercially relevant species, representing 6% of B C ' s coastal forest standing volume, but 17.6% o f the lumber production (COFI, 2000). Because of its high density and strength properties, Douglas-fir is used mainly as dimensional lumber, and thus w i l l be present in mill residues predominantly as sawdust, shavings and reject lumber.  1.2.1  Structure and composition of wood  The structural arrangement and chemistry of the feedstock directly influences its susceptibility to the processing involved with bioconversion. A s such, an understanding of the composition of the resource is important.  W o o d is a composite material composed of three main polymers:  cellulose, hemicellulose and lignin. These polymers are arranged in such a manner to provide structure and support for the growing tree.  However, by no means is wood a homogeneous 8  material. Variation in terms of chemical composition and morphology is considerable between tree species, even within each of the two broad divisions of hardwood and softwood species. Furthermore, within a given species, factors such as climate/environmental stresses, nutrient availability and age can be expected to introduce significant variability (Browning, 1963; Jozsa and Middleton, 1994).  Cellulose, the principle structural component of biomass, is the most  abundant of the polymers in wood, generally comprising 40-45% of the available material in softwoods (Sjostrom, 1993).  Hemicellulose, a second structural polymer of carbohydrate,  accounts for 20-30% of softwood biomass.  Lignin, which binds together and  stiffens  microfibrils, normally accounts for 25-30%. In contrast, hardwoods generally contain slightly higher cellulose and hemicellulose content, and lower lignin content than do softwood species (Sjostrom, 1993).  1.2.1.1  Cellulose  Cellulose is a non-branched homopolymer of p-D-glucopyranose  units, linked via (1-4)-  glycosidic bonds (Sjostrom, 1993). Each glucose residue is rotated 180° relative to neighbouring glucose residues, making the actual repeating unit a glucose dimer, cellobiose. The cellulose chains vary in the number of glucose molecules (i.e., degree of polymerization), ranging from as little as 15 to 10000-14000 molecules in size (Cowling and K i r k , 1975). Due to the particular orientation of the glucose units, combined with the equatorial placement of larger substituents (i.e., O H and CH2OH), cellulose forms linear chains capable of extensive hydrogen bonding and van der Waals interactions with parallel chains of cellulose (Haygreen and Bowyer, 1996; Sjostrom, 1993).  Structurally, the chains of cellulose aggregate to form elementary fibrils,  which contain both crystalline and amorphous regions. These fibrils are the building blocks of microfibrils, which in turn form macrofibrils and are incorporated into the cell wall (Fengel and Wegener, 1989; H i l l i s , 1982).  The tightly packed, largely crystalline nature of cellulose 9  accounts for its improved resistance to the chemical and physical treatments often employed during bioconversion. Of the carbohydrate polymers in wood, cellulose is the most recalcitrant to acid-catalyzed hydrolysis.  1.2.1.2 Hemicellulose In contrast to cellulose, hemicellulose is a heteropolymer of carbohydrates.  It functions as a  structural polymer, closely associated with both cellulose (via hydrogen bonds) and lignin (via covalent bonds) (Haygreen and Bowyer, 1996). In wood, the principle sugars of hemicelluloses are D-glucose, D-mannose, D-galactose, D-xylose, L-arabinose and L-rhamnose (Sjostrom, 1993). In addition, uronic acids (e.g., D-glucuronic acid, D-galacturonic acid and 4-O-methylD-glucuronic acid) may be found in small quantities. Hemicelluloses differ from cellulose in that they generally have a much lower degree of polymerization (less than 200 units), may be branched, may form linear or helical chains, and may be substituted (e.g., with acetate or other sugars). Branching and substitution can affect susceptibility to acid hydrolysis (Sjostrom, 1993). In softwoods, the principle hemicellulose is galactoglucomannan. It is composed of linear (1-4)linked B-D-glucopyranose and B-D-mannose, and is partially substituted with both a-Dgalactose bound via the C6-position (forming a single-sugar branch), and acetate at the C2- and C3-positions, although the degree of substitution can vary (Sjostrom, 1993; Theander, 1982). In addition, softwoods contain arabinoglucuronoxylan and arabinogalactan, although at much lower concentrations than glucomannan-based hemicelluloses. Arabinoglucuronoxylan consists of a (l-4)-linked B-D-xylopyranose backbone, partially substituted with 4-O-methyl-D-glucuronic acid and L-arabinofuranose at the C2 and C3 positions, respectively. Arabinogalactan is a highly branched, water-soluble polymer built from (l-3)-linked B-D-galactopyranose, and substituted with B-D-galactopyranose and L-arabinose. Additional carbohydrates in the form of starch and  10  pectic materials (e.g., amylose, amylopectin) may represent a minor component of woody biomass.  1.2.1.3  Lignin  In contrast to the carbohydrate polymers, lignin is a high molecular weight phenolic compound present in biomass to strengthen and stiffen the wood.  Lignin is closely associated with the  carbohydrate fraction, and encrusts microfibrils during cell wall synthesis and individual cells via the middle lamella. Chemically, lignin is composed of phenylpropane units, with softwood lignin comprised principally o f guaiacyl units polymerized v i a free radical reactions (Glasser, 1982; Sjostrom, 1993).  Structurally, lignin possesses a complex and largely ill-defined three-  dimensional structure. The highly polymerized matrix and its close association with the cell wall carbohydrate  (via lignin-carbohydrate  complexes,  LCCs)  ( L a i , 1991) are particularly  problematic for bioconversion, as access to the carbohydrate fraction can be restricted.  1.2.1.4 Additional c o m p o n e n t s In addition to carbohydrates  and lignin, wood contains additional chemical  constituents,  including extractives and inorganic compounds (Theander, 1982). Extractives are non-structural components, and may serve either nutritional (e.g., fats) or defensive roles (e.g., terpenoids, resin acids and phenolics) within the tree (Sjostrom, 1993). Inorganic components account for only a minor component of the whitewood of softwood species (usually less than 1%) (Sjostrom, 1993).  1.2.2 Bark Although bark is not normally considered as a feedstock for bioethanol production, some researchers have investigated its suitability due to the large available quantities, and limited commercial value (David and Atarhouch, 1987; Torget et al, 1991; Vasquez et al,  1988). 11  However, bark is more likely to be a relatively minor component in whitewood feedstocks due to the utilization of contaminated chips or whole tree feedstocks. Although bark contains many of the same chemical compounds present in the whitewood, their distribution and quantity differs enormously (Jensen et al,  1963). In particular, the carbohydrate content of the bark is lower  than in whitewood, while extractive content, lignin and inorganic compounds (ash) are generally much higher (Sjostrom, 1993).  Additional compounds such as alkali soluble suberins and  carbohydrates (e.g., callose) are also present in the bark. Structurally, bark is composed of inner (i.e., phloem) and outer (i.e., rhytidome) fractions (Jensen et al, 1963). The inner bark is living, and its principle function is conduction of photosynthates from the tree crown, but also storage of food reserves. The outer bark offers a protective barrier to the elements, preventing moisture loss, and offering resistance to microbial and insect attack.  1.3 Pretreatment Although both chemical and enzymatic process have been studied extensively for the bioconversion of lignocellulosics, enzymatic-based processes have grown in favour over the last 20 years. This may be attributed to a number of factors, including improved understanding of the catalytic enzymes combined with decreased production costs, but also the high specificity that enzymatic processes afford a conversion process.  However, the enzymatic conversion of  native lignocellulosics can be difficult. Because the lignocellulosic matrix is largely an insoluble composite of lignin and carbohydrates, direct physical contact between hydrolytic enzymes and the matrix is required to achieve hydrolysis of the carbohydrate fraction.  Consequently,  substrate accessibility is of paramount importance to bioconversion via enzymatic processes, and any physical or chemical barrier restricting access for the enzyme can result in decreased hydrolysis (Cowling and K i r k , 1975). Accordingly, various pretreatment processes have been  12  investigated in an effort to improve accessibility and susceptibility of the lignocellulosic and, consequently, improve the rates and yields of hydrolysis.  A number of factors are thought to contribute to the observed recalcitrance of lignocellulosics to enzymatic hydrolysis. Pretreatment strategies, in general, attempt to address these factors, which include porosity/accessible surface area of the substrate, its crystallinity, and the degree and type of lignification. Accessible surface area, and the degree/type of lignification restricts the number of places that an enzyme can physically bind the substrate, and many researchers believe this to be the critical factor governing the initial rate of hydrolysis (King, 1964; Thompson et al, 1992). Consequently, factors  such as fibre coarseness and particle size (Mooney et al,  hemicellulose content (Grethlein, 1985; Grohmann et al, Clark and Mackie, 1987; Converse et al,  1999),  1985), lignin content (Baker, 1973;  1990; Mooney et al,  1998; Vinzant et al,  1997),  which can effectively limit substrate surface area and porosity, have been shown to restrict hydrolysis. In contrast, Lee and Fan (1982) showed that, although available surface area appears to govern the adsorption of enzyme, crystallinity determines the effectiveness of hydrolysis, and a correlation between cellulose crystallinity and digestibility was demonstrated (Fan et al, 1987; Sinitsyn et al,  1991). This claim has been refuted by researchers arguing that the correlation  between crystallinity index and digestibility is not causal (Caulfield and Moore, 1974), and others that have shown improved hydrolysis despite increased substrate crystallinity following pretreatment (Carrasco et al, 1994; Tanahashi et al, 1983; Thompson et al, 1992).  Beyond merely improving the susceptibility of the cellulose component to hydrolysis, it should be recognized that pretreatment is also a means by which the lignocellulosic feedstock can be fractionated into its three main constituents: hemicellulose, cellulose and lignin. The utilization of the non-cellulose components is critical to the economic success of bioconversion, and one of 13  the major focuses o f the current research was to ensure high recovery of the hemicellulose sugars to provide a readily fermentable stream.  However, in the interest o f maximizing total sugar  recovery from the substrate, it should be clear that high hemicellulose sugar recovery cannot come at the expense o f cellulose digestibility.  Ideally, a pretreatment process should allow  complete and efficient fractionation, with minimal material losses (especially carbohydrate), and provide high-value lignin with minimal condensation or modification. Process-wise, it should be cheap and fast with a limited number of processing steps, and be environmentally benign. Moreover, the same procedure should be amenable to a wide range o f substrates, so that different sources of lignocellulosics could be used without requiring substantial modifications to the overall process.  Clearly, these requirements are unlikely to be met i n a single process, and a  compromise w i l l be required.  1.3.1  Pretreatment strategies  Methods to pretreat biomass are almost as numerous as the feedstocks themselves—each with its own particular advantage—and have been reviewed extensively (Fan et al, 1982; Hsu, 1996; McMillan,  1994; Millett et al,  comminution  1974).  Most processes rely on some form of physical  {e.g., chipping, refining, ball milling, explosive decompression) or chemical  treatments (e.g., swelling agents, solvents), or a combination o f both. Although by no means an exhaustive list, researchers have investigated numerous lignocellulosic pretreatment processes including milling, steaming, hydrothermolysis, irradiation, treatment with concentrated  acids  such as sulphuric and hydrochloric acid, swelling with sodium hydroxide, calcium hydroxide and ammonia, extraction with sodium hydroxide, ammonia fibre explosion and carbon dioxide explosion, organic solvent pulping (i.e., organosolv pulping), super-critical fluid extraction, biological  treatments,  dilute  acid pretreatment  using  sulphuric,  nitric, phosphoric and  hydrochloric acids, and acid-catalyzed steam explosion using sulphur dioxide and sulphuric acid. 14  1.3.2 A c i d - c a t a l y z e d steam e x p l o s i o n The suitability of a particular pretreatment strategy depends largely on the feedstock of choice (with many processes not suited to softwood feedstocks), but beyond this, a major determinant is the cost and relative complexity of the process. Many of the aforementioned pretreatments are energy intensive and oftentimes poorly effective at achieving their goal, and thus have not progressed much beyond bench-scale testing (Hsu, 1996).  For the bioconversion of wood  lignocellulosics, dilute acid hydrolysis and acid-catalyzed steam explosion are, arguably, the most commonly employed.  Dilute acid hydrolysis as a pretreatment is used extensively by  N R E L ( M c M i l l a n , 1994).  However, steam explosion is the most commercialized of the  pretreatment processes ( M c M i l l a n , 1994; Saddler et al, 1993). The process, which is based on the Masonite pulping process, is characterized by high temperature/pressure steaming (170280°C) for relatively short residence times (10 seconds to as long as 20 minutes) (Saddler et al, 1993). This results in physical and chemical changes to the lignocellulosic, due to the hydrolysis and solubilization of the hemicellulose component, partial solubilization of the cellulose, and modification of lignin.  Lignin is only poorly solubilized (Clark et al,  1989), although it may  undergo extensive modification including condensation (Brownell, 1988; Shevchenko et  al,  1999) and redistribution following liquefaction and coalescence under the heat of steaming (Brownell and Saddler, 1987; Donaldson et al, 1988; Michalowicz et al, 1991). This process of "lignin melting" may contribute to opening up the lignocellulosic, although condensation and localization of the lignin to the matrix surface may have negative consequences by blocking pores ( M c M i l l a n , 1994).  It is imperative that reaction conditions are fully optimized, as  pretreatment can result in the extensive decomposition of the hemicellulose sugars (Brownell and Saddler, 1984), which should be avoided. Typically, the steam treatment is followed by rapid decompression of the reactor. This is generally thought to improve enzymatic hydrolysis of the resulting pulp, due to mechanical shearing and defibrillation of the fibre, and the 15  associated increase in accessible surface area.  However, some research suggests that, at least  under the relatively severe pretreatment conditions used in their work (240°C for 90 s + bleeddown time), explosive decompression is not required to improve the subsequent enzymatic hydrolysis (Brownell and Saddler, 1987; Brownell et al, 1986).  1.3.3  Choice of acid catalyst  Pretreatment of softwoods is much harder than hardwoods and agricultural residues, but can be greatly improved through the use of an acid catalyst such as sulphur dioxide (SCh) or sulphuric acid  (H2SO4).  The original steam explosion processes  autohydrolysis reactions, and relied on the presence  designed for hardwoods  were  of acetic acid liberated from  the  hemicellulose fraction to catalyze the hydrolysis of the feedstock. In softwoods, hemicellulose is acetylated to a much lower extent, necessitating supplementation of acid.  In general, the  acidifying agents permit the use of lower reaction temperatures. Furthermore, the addition of acid during steam explosion helps to improve hydrolysis and solubilization of the hemicellulose, but also partly stabilizes the sugar from decomposition once it has been solubilized (Brownell et al, 1986; Clark et al, 1989). In the SCVcatalyzed process, SO2 is not the true catalyst. Rather, it undergoes oxidation and/or disproportionation yielding H2SO4 within the wood (Brownell et al, 1988). Only limited sulphonation of the lignin results during S02-catalyzed pretreatment (in contrast with acid-sulphite pulping of wood), suggesting that the mechanism of SO2 for improving enzymatic hydrolysis is largely hydrolytic, and not swelling or dissolving the lignin (Clark et al,  1989; Mamers and Menz, 1984).  Both acids appear similarly effective for  hemicellulose sugar recovery and improving cellulose hydrolysis (Eklund et al, 1995; Mackie et al,  1985; Tengborg et al,  1998), although S 0  2  affords  a generally easier and  faster  impregnation, results in decreased consumption of thermal energy during steaming (due to decreased  moisture content compared to dilute acid-impregnated feedstocks), and is less 16  corrosive to the reactor. Sulphur dioxide is also purported to be more economical than sulphuric acid (Wayman et al, 1984). In work comparing the two acid catalysts, H2SO4 was reported to improve recovery of the hemicellulose sugars from softwood feedstocks during the pretreatment step compared to S O 2 , although the enzymatic hydrolysis step was better following S O 2 impregnation (Tengborg et al, 1998; Stenberg et al, 1998). Fermentation of the hemicellulose rich water-soluble fraction was also improved when SO2 was used during pretreatment instead of H2SO4 (Tengborg et al,  1998).  Similar trends to these were also observed with a willow  feedstock (Eklund et al, 1995).  1.3.4  Severity of Pretreatment  The pretreatment of biomass by steam explosion is in essence a high temperature/pressure cooking step for a predetermined time period. Increasing the variables of time or temperature will result in more severe processing of the feedstock, with greater hemicellulose and cellulose solubilization.  T o aid in the comparison of results obtained under various experimental  pretreatment conditions, Overend and Chornet (1987) proposed the use of the severity factor. This factor, log R , which is based on the concepts of the " F T factor used in kraft pulping 0  (Vroom, 1957) and the " P " factor used in pre-hydrolysis kraft pulping (Brasch and Free, 1965), combines the effects  of time and temperature into a single variable, R , according to 0  Equation 1-1:  R = K 1  | e  r  ' -  r  '  W  H  (Eq.1-1)  where t is the residence time (min), T is the reaction temperature (°C), and Tb is the reference r  temperature (100°C).  This factor allows researchers to manipulate time and temperature  parameters while achieving effectively comparable pretreatment of the feedstock.  Although  imperfect (it fails to account for a number of additional factors that have been shown to influence pretreatment, such as chip size, moisture content, and acid concentration), it provides a useful 17  method for comparison of results obtained under different processing conditions. A combined severity factor has been described by Chum et al, (1990), which incorporates reaction pH arising from the added acid: Combined severity factor = log R - pH 0  (Eq. 1-2)  It should be noted that while the combined severity factor is easily applied to pretreatments using H SC»4 impregnation, it is difficult to monitor the reaction pH of SCh-impregnated chips prior to 2  explosion.  Under conditions of very mild pretreatment severity, only limited physical and chemical changes occur to the feedstock. As severity increases, greater hydrolysis of the carbohydrate component will occur, beginning with the easily hydrolysed hemicellulose component.  Ultimately, under  sufficiently severe pretreatment, the cellulose component will be significantly hydrolysed. However, it is important that the severity of pretreatment be carefully controlled. Achieving high sugar yields with acid-catalyzed steam pretreatment is a delicate balance between the hydrolysis of polymeric and oligomeric sugars to monomers, and the degradation of soluble monomeric sugars to decomposition products, which can have a pronounced impact on fermentation.  Complicating this is the fact that the optimum conditions for hemicellulose and  cellulose recovery do not occur at the same severity conditions. This is related to the structure and arrangement of the two polysaccharides, as described previously. Under conditions that provide optimal pretreatment of the cellulose component for enzymatic hydrolysis, substantial degradation of the hemicellulose component can occur. Tailoring reaction conditions to provide maximum recovery of the hemicellulose sugar component is potentially advantageous. This can provide an increase in the total carbohydrate recovered from the feedstock, thus allowing for increased ethanol production. A second advantage is that the hemicellulose hydrolysate will be more readily fermentable due to the decreased production of inhibitory compounds, as will be 18  addressed subsequently.  A drawback to this approach is that, under the milder pretreatment  conditions required to ensure the increased survival of the monomeric hemicellulose sugars, the residual fibre (cellulose and lignin) may remain largely recalcitrant to enzymatic hydrolysis (Boussaid et al, 2000; W u et al, 1999). Researchers have attempted to overcome this through the use of multi-stage pretreatments (Boussaid et al, 2000; Eklund et al, 1988; Nguyen et al, 1999; Nguyen et al,  2000; Soderstrom et al,  2002).  However, problems associated with  additional handling and fractionation of the material combined with increased energy costs may make this approach economically unattractive, despite the increased sugar yields (Boussaid et al, 2000).  1.4 Fermentation 1.4.1 Selection of m i c r o o r g a n i s m A number of different microorganisms have been employed in the fermentation of the watersoluble hemicellulose-derived hexose sugars to ethanol. Saccharomyces  For the most part, the  cerevisiae has been used for these conversions.  yeast  This is a particularly robust  organism, selected for over literally thousands of years for its fermentation abilities (Wills, 1990). Strains of S. cerevisiae are capable of high rates of ethanol production from a range of carbon sources, display high osmotolerance and ethanol tolerance (Ingram, 1986), and often demonstrate increased resistance to potentially inhibitory compounds present in the watersoluble fraction, although individual performances have been shown to vary (Martin and Jonsson, 2003). Arguably, the principle drawback for bioconversion is the yeast's inability to metabolize pentose sugars to ethanol.  This is particularly important for the bioconversion of  hardwood and agricultural residues, but is of reduced concern for softwood residues, which contain a significantly reduced proportion of pentosan sugars (-5-10% of the available carbohydrate). 19  To address the concerns with pentose sugar fermentation, researchers have explored the use of alternative organisms such as Pichia  stipitis, Pachysolen  tannophilus  and Candida shehatae  (Hahn-Hagerdal et al, 1993; Jeffries, 1983). These yeasts can ferment xylose to ethanol, but often with low productivity and ethanol yield (du Preez et al, 1986, 1989; Hahn-Hagerdal et al, 1994a), and with more complex culturing demands (e.g., increased nutritional and aeration requirements) (Dellweg et al, 1984; du Preez et al, 1985; Skoog and Hahn-Hagerdal, 1990). These yeasts may also show increased sensitivity to ethanol, resulting i n the inhibition of both growth and fermentation (du Preez et al, 1987, 1989). Furthermore, they do not metabolize other sugars such as glucose as rapidly as 5. cerevisiae (du Preez et al, 1986; Hahn-Hagerdal et al, 1994b), and may be more sensitive to potential inhibitors (e.g., acetic acid, vanillin) present in the water-soluble fraction (Delgenes et al, 1996; Linden and Hahn-Hagerdal, 1989). Bacteria have also been investigated, with the principle naturally-occurring, ethanologenic bacterium being Zymomonas mobilis.  This organism can produce greater quantities of ethanol per mole of  biomass produced, due to a metabolic pathway (i.e., the Entner-Doudoroff pathway) that yields less energy in the form of A T P than does the Embden-Mayerhof pathway i n S. cerevisiae (Swings and De L e y , 1977). However, the natural substrate range o f Z. mobilis is restricted to the fermentation  of glucose, fructose  transformation (Lindsay et al, 1995). Escherichia  and sucrose,  limiting its usefulness  for biomass  Engineered microorganisms, including ethanologenic  coli (Ingram et al, 1987; Lindsay et al, 1995; Neale et al, 1988; Ohta et al, 1991),  and pentose-fermenting S. cerevisiae (Ho et al, 1998; H o et al, 1999) and Z. mobilis (Zhang et al,  1995) have been developed to address these "shortcomings".  These organisms are likely  destined to be the workhorses of lignocellulosics conversions, but currently do not have a proven track record in industrial bioconversion. However, concerns regarding the stability of the cloned constructs (Dien et al, 1998; Dumsday et al, 1999; Ohta et al, 1991), and related tolerance to  20  water-soluble inhibitors (Asghari et al, 1996; Hahn-Hagerdal et al, 1994a) are being addressed, and commercial processes are currently being designed with these recombinants in mind.  1.4.2 Inhibition of fermentation The fermentability of the water-soluble fraction is dependent on its composition, both in terms of carbohydrates  (monomers and oligomers), but also compounds inhibitory to fermentation.  Inhibitors belong to numerous chemical functional groups, including acids, aldehydes, and phenolics.  F o r the ease of discussion, these compounds can be grouped into one of two  categories: process-derived inhibitors arising from the pretreatment step (e.g., sugar and lignin degradation products), or naturally-occurring inhibitors liberated from the feedstock and recovered in the water-soluble fraction (e.g., resin/fatty acids, sterols, organic acids such as acetic and uronic acids). Inhibitors may play a significant role during the fermentation of the sugars in the water-soluble fraction to ethanol.  1.4.2.1  P r o c e s s - d e r i v e d inhibitors  Sugar decomposition during acid-catalyzed pretreatment under increased severity can result i n the inhibition of microbial growth and ethanol production. Although pentose sugars are more labile  than  hexose  sugars  (BeMiller,  1967),  degradation  products  derived from  both  carbohydrates contribute to the inhibition of fermentation. Hexose sugars may decompose under first-order  reactions  to yield  furan-type  compounds  (e.g., 5-hydroxymethylfurfural, 5-  methylfurfural), acids (e.g., levulinic and formic acid) and polymeric phenolic materials, while pentose sugars predominantly yield 2-furfural and condensation products (Lai, 1991; Saeman, 1945). Both H M F and furfural have been reported to cause a reduction in C 0  2  evolution and  growth rate (Banerjee et al, 1981a; Delgenes et al, 1996; Nishikawa et al, 1987; Sanchez and Bautista, 1988; Taherzadeh et al, 2000). In addition, they may result in a reduction in ethanol 21  yield (Chung and Lee, 1985; Delgenes et al, 1996), although other researchers have indicated only decreased ethanol productivity (Larsson et al, 1999a). Furfural has been shown to inhibit enzymes involved in glycolysis (Banerjee et al,  1981b), and based on the similarity of the  compounds and the yeast's physiological response to H M F , it is suspected that H M F operates in a similar manner (Taherzadeh et al, 2000). Hydroxymethylfurfural is generally reported to be less inhibitory than furfural (Nilvebrant et al, 1997; Sanchez and Bautista, 1988), although its metabolism is slower (Taherzadeh et al, 2000) meaning a prolonged period of inhibition is possible. Fermentative microorganisms may be acclimated to these inhibitors by prior exposure at non-toxic levels (Amartey and Jeffries, 1996; Banerjee et al, 1981a; Keller et al, 1998; Sene et al, 2001), although the mechanism by which this occurs is not entirely understood (Boyer et al, 1992). However, even without prior acclimation, these furan aldehydes may be metabolized to less inhibitory alcohols and acids during fermentation (Palmqvist et al, 1999a; Taherzadeh et al, 1999; Taherzadeh et al, 2000), resulting in improved fermentation.  The weak acids such as levulinic, formic, and acetic are known to influence cell growth and ethanol production (Larsson et al, Hagerdal, 2000; Taherzadeh  1999a; Palmqvist et al,  et al,  1997b).  1999b; Palmqvist and Hahn-  One proposed mechanism is intracellular  acidification of the cytoplasm, and increased energy requirements for cell maintenance (i.e., uncoupled growth) (Verduyn et al,  1990). The undissociated form of these acids can diffuse  through the cell membrane to the cytoplasm, whereupon they dissociate due to the higher intracellular p H and cause a drop in cytoplasmic p H . The excess hydrogen ions must be actively pumped out of the cell to maintain a neutral cytoplasmic p H , at a cost of cellular A T P ; consequently, biomass yield decreases.  To satisfy the increased demand for A T P , cells may  produce more ethanol (Maiorella et al,  1983). Thus, at low concentrations, these acids may  actually stimulate ethanol yield due to the associated reduction in biomass production. Larsson 22  et al, (1999a) reported an increase in ethanol yield up to an acetic acid concentration of 100 m M , while Taherzadeh et al, (1997b) showed increased ethanol yields up to 150 m M . Higher concentrations were shown to result in decreased ethanol yields and cell death (due to dissipation of the proton motive force). However, the intracellular anion accumulation may also contribute to direct inhibition of cellular metabolic pathways (Russel, 1992).  In addition to carbohydrate decomposition and inhibition, lignin can degrade under acidic conditions, primarily by the cleavage of the aryl ether bonds at the a - and [3-positions (Lai, 1991). Limited solubilization of lignin via sulphonation has also been reported (Clark et al, 1989), and presumably these lignin compounds contribute to inhibition as well.  The lignin-  derived phenolics have been shown to inhibit lignocellulosic fermentations, both for the production of ethanol by S. cerevisiae and other microorganisms (Ando et al, 1986; Clark and Mackie, 1984; Delgenes et al,  1996), but also 2,3-butanediol by Klebsiella  pneumoniae  (Nishikawa et al, 1987, 1988; Tran and Chambers, 1986). It is believed that inhibition is related to disruption of the plasma membrane. Inhibitory effects are related to the molecular weight of the compound, which affects its permeability, but also the functional group attached to the benzene ring (Ando et al,  1986).  Removal of these compounds by extraction with solvents  (Clark and Mackie, 1984; Frazer and McCaskey, 1989), overliming and treatment with laccases (Jonsson et al,  1998; Larsson et al,  1999b) has been shown to improve the subsequent  fermentation step. Although these components are generally reported to be the most toxic of inhibitors present in wood hydrolysates, the concentration of individual components is normally quite low, due to their limited solubility.  23  1.4.2.2  Naturally-occurring inhibitors  As mentioned earlier, the naturally-occurring inhibitors are primarily of extractives origin. Biologically, many of these extractable components play a defensive role against microbes and insects in protecting the wood from decay (Haygreen and Bowyer, 1996), thus the accumulation of these materials i n the water-soluble fraction should also be expected to be detrimental to fermentation efficiency.  Tran and Chambers (1986) reported on the toxicity of various  extractive model compounds, but also on the inhibition resulting from the fermentation of the water-soluble fraction derived from southern  pine, observing a significant reduction in  fermentability. The use of feedstocks with notably higher concentrations of extractives, such as bark, might be expected to cause problems for fermentation. However, as for the lignin-derived phenolics, the concentration o f these compounds recovered i n the water-soluble fraction is often low, owing to their generally hydrophobic nature, and reduced solubility (Lomax et al, 1994).  1.5 Inhibition of enzymatic hydrolysis As addressed earlier, the enzymatic hydrolysis of lignocellulosics is complicated for a number of reasons, largely pertaining to accessibility issues. However, beyond these concerns is the issue of inhibition of enzymatic hydrolysis by compounds present in and liberated from the lignocellulosic substrate.  These include the carbohydrate and lignin components themselves.  Much of this inhibition has been originally reported in the context of ruminant nutrition (Jung and Fahey, 1983).  However, the same effects apply to the bioconversion of pretreated  lignocellulosics feedstocks.  1.5.1  Inhibition by carbohydrates  The accumulation of carbohydrates in the hydrolysis medium is known to slow the conversion of cellulose to glucose.  Such end-product inhibition caused by the liberation of cellobiose and 24  glucose from the cellulose polymer can result in substantial inhibition of the major components of the cellulase system. This has been shown for both the endo- and exo-acting cellulases and (3glucosidase (Breuil et al, 1992; Dekker, 1986; Holtzapple et al, 1990). The exact mechanism of inhibition by cellobiose and glucose remains contentious (Gusakov and Sinitsyn, 1992), and has been modeled as competitive, non-competitive, and uncompetitive in nature by various researchers (for a detailed summary, see Holtzapple, 1990). Cellobiose is generally reported to be a stronger inhibitor of the cellulases than is glucose (Holtzapple et al,  1990; Lee and Fan,  1983). This is not to say that monomeric glucose is not a capable inhibitor. Hadj-Taieb et al, (1992) reported a 30% decrease in total cellulase activity of a Penicillium  occitanis cellulase  preparation in the presence of only 5% glucose, while Marsden et al, (1983) reported a 50% reduction in activity using a 4% glucose medium with a Trichoderma preparation.  reesei cellulase  It is important to consider that separating the influence of glucose from that of  cellobiose can be difficult, owing to the fact that glucose accumulation leads to the inhibition of (3-glucosidase, and consequently cellobiose will accumulate. Cellobiose, in turn, w i l l inhibit the cellulase enzymes. In addition to glucose and cellobiose, other carbohydrates have been shown to inhibit cellulose hydrolysis.  Xylose has been implicated (Dekker, 1986; Marsden et  al,  1983), as well as oligomeric xylan (Mes-Hartree and Saddler, 1983).  1.5.2  Inhibition by non-carbohydrate c o m p o u n d s  The presence of compounds within the water-insoluble cellulose matrix following pretreatment have long been recognized to adversely affect enzymatic hydrolysis. The hydrolysis of steampretreated feedstocks may be especially problematic due to the accumulation of degradation compounds. Researchers have suggested that sugar degradation products such as furfural may adversely affect the cellulose hydrolysis (Sinitsyn et al,  1982), although not normally at the  concentrations typically found in wood hydrolysates (Mes-Hartree and Saddler, 1983; Tengborg 25  et al, 2001). O f potentially greater impact are the phenolic compounds present in the watersoluble and insoluble fractions, which may arise during pretreatment, but also during enzymatic hydrolysis (Dekker, 1988). W h i l e lignin associated with the cellulose matrix has been shown to mediate access to the cellulose component (Converse et al,  1990), the aromatic compounds  liberated from the degradation of lignin contribute greatly to the inhibition of a range of extracellular enzymes, including cellulases (Sinsabaugh and Linkins, 1987).  Monoaromatic  compounds (e.g., vanillin, vanillic acid, ferulic acid, p-coumaric acid) (Asiegbu et al,  1996;  Heredia et al, 1990; Jecu, 1998; Mandels and Reese, 1963; Martin and Blake, 1989), polymeric tannins and extractives (Juntheikki and Julkunen-Tiitto, 2000; Mandels and Reese, 1963; Sinsabaugh and Linkins, 1987; Walch et al, 1992), and lignin itself (Sewalt et al, 1997) have been implicated in the inhibition of hydrolytic enzymes, through reversible and irreversible binding, precipitation and deactivation of the enzymes (Jung and Fahey, 1983). Early research addressing this issue found that a simple water-wash of the water-insoluble component could remove much of this inhibition (Mes-Hartree and Saddler, 1983; Sinitsyn et al, 1982), and in general, this approach has been exploited in order to reduce this inhibition.  Limiting the  production of these degradation compounds should also enable improved enzymatic hydrolysis.  1.6 R e s e a r c h a p p r o a c h a n d objectives Since the 1970s, there has been significant research and development brought about by numerous research groups for the improvement of lignocellulosic pretreatment, fractionation, enzymatic hydrolysis and fermentation.  A n d yet, even now, more than a quarter of a century  later, there has been only limited commercialization of a process that could produce the "potential transportation fuel of the future" (Lynd, 1996). transportation remains a worthwhile pursuit.  Clean, renewable energy for  However, its commercial realization requires  continued breakthroughs and advances in technology. It is hoped that the work described in this 26  thesis will contribute to the general understanding and practical application of bioconversion technology, ultimately bringing the bioconversion of woody lignocellulosics closer to reality.  The current project was initiated as part of ongoing research designed to evaluate and improve acid-catalyzed steam explosion as a strategy for the pretreatment of softwood feedstocks. Specifically, the project was to address the issues of feedstock pretreatment, with the major focus on the fermentability of the water-soluble fraction derived from steam exploded Douglas-fir feedstocks. The goal of the project was to demonstrate high yield recovery of the carbohydrates from the feedstock, and improved fermentability of the hemicellulose-rich water-soluble fraction to high ethanol yield and concentration, through the tailoring of pretreatment conditions.  As mentioned previously, one of the general requisites of bioconversion processes is to maximize total carbohydrate recovery from the lignocellulosic feedstock. Due to the recognized decrease in hemicellulose sugar yields with increased pretreatment severity, it quickly became apparent that a mild pretreatment step (i.e., reduced severity of pretreatment) would be vital for achieving maximum hemicellulose sugar recovery from the starting feedstock to the watersoluble fraction.  In addition, the selection of relatively mild conditions was anticipated to  minimize both the decomposition of the feedstock and the production of compounds inhibitory to the fermentation process. In theory, use of the above approach could permit a greater overall recovery of carbohydrate, provided that the carbohydrate from the water-insoluble cellulose component remained hydrolysable. However, the pretreatment step is recognized to serve a dual purpose. Not only is it necessary to actuate hemicellulose sugar recovery, but also to improve the accessibility of the cellulose component to hydrolytic enzymes. Pretreatment at conditions optimized for the recovery of hemicellulose sugars, for example, might not necessarily provide sufficient modification to the water-insoluble component  to allow effective hydrolysis. 27  A l t h o u g h the c e l l u l o s e h y d r o l y s i s step was not i n i t i a l l y a focus o f this research project, efficient e n z y m a t i c h y d r o l y s i s is i m p e r a t i v e for the v i a b i l i t y o f the b i o c o n v e r s i o n process.  In anticipation  that the g l u c o s e y i e l d s f o l l o w i n g e n z y m a t i c h y d r o l y s i s w o u l d be c o m p r o m i s e d b y the m i l d pretreatment c o n d i t i o n s , t w o a d d i t i o n a l pretreatment c o n d i t i o n s were e x p l o r e d : a high-severity pretreatment, o p t i m i z e d f o r d i g e s t i b i l i t y o f the w a t e r - i n s o l u b l e c e l l u l o s e component,  and a  c o m p r o m i s e c o n d i t i o n r o u g h l y m i d p o i n t between the h i g h and l o w - s e v e r i t y pretreatments.  Both  h e m i c e l l u l o s e sugar r e c o v e r y and fermentability o f the r e s u l t i n g water-soluble fractions  are  addressed i n the current w o r k .  W h i l e v i r t u a l l y a l l research into the b i o c o n v e r s i o n o f s o f t w o o d and h a r d w o o d species has r e l i e d on w h i t e w o o d feedstocks, it was thought l i k e l y that bark w o u l d be a c o m p o n e n t o f c o m m e r c i a l l y relevant feedstocks. c o n t a i n i n g feedstocks,  O n l y l i m i t e d research has l o o k e d at the b i o c o n v e r s i o n o f bark or barkand it was not entirely clear h o w D o u g l a s - f i r b a r k m i g h t impact the  b i o c o n v e r s i o n o f w h i t e w o o d c h i p s . F o r e x a m p l e , w o u l d the i n c l u s i o n o f bark i n the w h i t e w o o d feedstock i m p a c t the pretreatment step, due to difference i n the p h y s i c a l structure and c h e m i c a l c o m p o s i t i o n o f the b a r k ? A l t h o u g h the process was thought robust e n o u g h to tolerate l o w levels o f contaminant bark, w o u l d it still be p o s s i b l e to recover the h e m i c e l l u l o s e sugar w i t h h i g h y i e l d i f bark was a s i g n i f i c a n t c o m p o n e n t o f the feedstock?  In an effort to answer these and related  questions, the i n f l u e n c e o f b a r k o n the recovery o f h e m i c e l l u l o s e sugars f r o m the a v a i l a b l e starting substrate was e x a m i n e d , u s i n g a feedstock c o n t a i n i n g 3 0 % b a r k b y weight. T h i s quantity o f bark was thought to be the upper l i m i t for bark " c o n t a m i n a t i o n " that m i g h t be r e a l i z e d through the use o f whole-tree forest thinnings or contaminated w o o d c h i p s .  Representative  feedstocks  were pretreated and the r e c o v e r y o f carbohydrates was assessed.  28  In addition to carbohydrate recovery issues, it was anticipated that the presence of bark in the feedstock would greatly affect the composition of the water-soluble fraction. In general, bark differs substantially from whitewood, in terms of the quantities of chemical constituents such as extractives and lignin.  W o u l d an increase in these components in the feedstock contribute to  their increased recovery in the water-soluble fraction? If so, would this translate into increased inhibition and decreased  fermentability of the water-soluble fraction?  Whitewood chip  feedstocks containing various concentrations of supplemental bark (10%, 20%, 30% and 100%) were pretreated under medium-severity conditions, and the resulting water-soluble fractions were analysed for inhibitors. In addition, the resulting water-soluble fractions were fermented to evaluate the impact of bark on the rate and yield of ethanol production.  A n additional concern for many bioconversion processes is the dilute nature of the water-soluble fractions. Complete recovery of the hemicellulose sugars following steam explosion can result in very dilute process streams.  In the current work, for example, the maximum sugar  concentration was less than 50 g L " . Such dilute process streams can cause economic problems 1  due to the associated equipment and energy costs.  In particular, the economics of ethanol  recovery is greatly affected. A low fermentable sugar concentration gives rise to a low ethanol concentration, which in turn contributes greatly to the costs of conventional distillation. Given the low sugar concentration in the Douglas-fir water-soluble fractions, an effort was made to improve them.  The application of physical concentration techniques such as rotary-evaporation and freezedrying to improve the starting sugar concentration of the water-soluble fractions was addressed. However, while the sugar concentration would undoubtedly be improved, it was thought likely that potential fermentation inhibitors present in the water-soluble fraction might also accumulate 29  to toxic levels. This might realistically impact the production of ethanol. In recognition of this fact, two additional alternatives were also investigated.  Carbohydrate derived from the water-  insoluble cellulose component was used to augment the sugar concentration in the water-soluble fraction. This was accomplished by performing the enzymatic hydrolysis directly in the watersoluble fraction, or by hydrolysing the water-insoluble fraction in buffer, and combining the resulting sugar stream with the water-soluble fraction.  The fermentability of the resulting  mixtures, in terms of sugar consumption and ethanol production, is reported.  30  Chapter 2: Materials and Methods 2.1 Feedstock and pretreatment 2.1.1  Feedstock  Douglas-fir (Pseudotsuga  menziesii)  whitewood and bark feedstocks were used to generate  substrate for all fermentations and hydrolyses, except where specifically noted. The Douglas-fir wood bolts were obtained from the M a l c o l m Knapp Research Forest. Over the course of the research, two different Douglas-fir wood bolts were employed. Initial work made use of bolts from a 129-year old tree; subsequent experiments utilized wood and bark from a 145-year-old tree.  The composition of these feedstocks differed slightly, and is described in detail in  subsequent chapters of the thesis.  2.1.2  Feedstock preparation  Freshly felled bolts of Douglas-fir were debarked by hand.  The whitewood was then either  chipped as-is, or was first separated into heartwood and sapwood fractions prior to chipping. In both cases, the whitewood was chipped using a mechanical chipper to an average size of 25 x 25 x 2 mm. The bark, which had been peeled from the wood at the cambial layer, consisted of outer bark, inner bark and the cambial zone. It was chipped by hand to minimize the generation of small particles. The average chip size was approximately 30 x 30 x 5 mm. W o o d and bark chips were stored at - 2 0 ° C until required.  2.1.3  Pretreatment and fractionation  Whitewood and bark chips were pretreated by S02-catalyzed steam explosion using a 2-litre Stake-Tech II steam gun (Stake Technologies, Norval, O N , C A N ) .  The severity of the steam  explosion pretreatment was represented by the severity parameter, L o g R , defined by Overend 0  31  and Chornet (1987) (Equation 1.1).  Three pretreatment severity conditions were used in this  work: low-severity (175°C, 7.5 min, 4.5% (w/w) S 0 , L o g R =3.08), medium-severity (195°C, 2  o  4.5 min, 4.5% (w/w) S 0 , L o g R =3.45), and high-severity (215°C, 2.38 min, 2.38% (w/w) S 0 , 2  L o g R =3.73). 0  D  2  These conditions were selected based on a previously defined conditions for  Douglas-fir whitewood (Boussaid et al, 1997), established using a hybrid experimental design (Roquemore, 1976) and second-order surface response analysis. The whitewood and bark chips were either pretreated separately, or in whitewood/bark mixtures containing 10%, 20% or 30% (w/w) bark based on oven-dried weight ( O D W ) .  The frozen chips were thawed and left in  plastic bags to ensure uniform moisture content prior to use. Aliquots containing 300 g O D W chips were placed in freezer bags (1.3 L capacity) and impregnated with S 0  2  (total weight  monitored). The bags were sealed and left overnight, and then steam pretreated in 50 g batches. Following steam explosion, the pretreated material was washed from the steam gun's receiving vessel with distilled water. The recovered slurries were adjusted to a dry matter consistency of 15-20% (w/w) and were filtered through Whatman glass microfibre G F / C filters under vacuum. Both the resulting water-insoluble component (WI) and water-soluble filtrate (WS) were named according to their bark content (e.g., WI-0, WI-10, WI-20; W S - 0 , W S - 1 0 , W S - 2 0 , etc.) and were frozen at - 2 0 ° C until required.  2.1.4 Delignification of the water-insoluble c o m p o n e n t Following steam explosion, the WI-0 and WI-30 component were delignified using a hot, alkaliperoxide protocol based on the method of Yang et al, (2002). The steam exploded pulp was washed with 30 volumes of d H 0 , then delignified at 2% (w/v) O D W consistency in 1% (w/v) 2  H 0 2  2  (Fisher Scientific Ltd., Nepean, O N , C A N ) .  described in section 2.6.3.5.  The peroxide was assayed for activity as  The mixture was adjusted to p H 11.5 using 50% N a O H ( V W R  International, Edmonton, A B , C A N ) . The delignification reactions were conducted in 1 L 32  Erlenmeyer flasks containing 450 m L total liquid, immersed in a hot water bath with gyratory shaking (80°C, 250 R P M ) for 60 min. Immediately following the delignification step, the pulp was vacuum filtered through Whatman #1 filter paper and washed with 40 volumes of warm, distilled water (50°C).  The residual fibre was named according to the bark content of the  feedstock (i.e., W I - A P - 0 , WI-AP-30) and was stored moist in sealed freezer bags at 4 ° C until required.  2.2 2.2.1  Enzymatic hydrolysis Enzymatic hydrolysis of the cellulose component  A l l enzymatic hydrolyses performed in this work adhered to the general parameters described in this section.  However, the specifics of particular experiments are described in detail in the  subsequent sections.  Hydrolyses were conducted using commercial enzymes supplied by  Novozymes North America, Inc., (Franklinton, N C , U S A ) .  Never-dried fibre was hydrolysed  using a complete cellulase preparation (Celluclast 1.5L) derived from the fungi,  Trichoderma  reesei. T o augment the naturally low (3-glucosidase activity of this preparation, the hydrolyses were supplemented  with additional (3-glucosidase enzyme  (Novozym-188) derived from  Aspergillus niger. The cellulase and (3-glucosidase preparations were assayed for total cellulase activity and P-glucosidase activity as described in section 2.6.3.1 and 2.6.3.2, respectively, with results presented in Appendix 1. Hydrolyses were conducted in one of three aqueous buffers: Acetate buffer  ( " A C " , 5 0 m M , p H 4.8), the non-buffered  WS-fraction derived from the  whitewood feedstock ("WS-0", adjusted to p H 4.8 with 50% N a O H ) , or a simulated watersoluble fraction consisting of A C buffer supplemented with anhydrous, monomeric sugars (arabinose, galactose, glucose, mannose and xylose) to the same concentration present in the WS-fraction, as required by the particular experiment ( " S - A C " ) . Celluclast was loaded to each  33  hydrolysis reaction to achieve a final activity of 10 filter paper units (FPU) g-cellulose" . 1  Novozym-188 was loaded to achieve an I U (international units) to F P U ratio of 3:1, unless indicated otherwise.  T o minimize the risk of microbial contamination, all hydrolyses except  those intended for fermentation to ethanol were supplemented with the antibiotics tetracycline (Sigma Chemical C o . , St. Louis, M O , U S A ) and cyclohexamide (Sigma) to a final concentration of 40 u,g mL" and 30 u.g mL" , respectively. Hydrolyses were performed in a shaking water bath 1  1  (45°C and 150 R P M ) . Samples (0.5 m L ) were taken periodically during the hydrolysis, boiled for 10 min, and then centrifuged for 5 min at 15000 x g and 4 ° C .  The supernatant was  transferred to a fresh micro-centrifuge tube and frozen at - 2 0 ° C prior to analysis for monomeric sugar by H P L C .  Control hydrolyses, containing either no enzyme or no fibre, were run in  parallel to correct for sugar in the buffer or derived from the enzyme preparations, respectively.  The extent or yield of hydrolysis is expressed as the percentage of the theoretical glucose content in the feedstock at the start of hydrolysis that was recovered as monomeric glucose (i.e., the glucose yield). Other monomeric sugars such as xylose and mannose, which increased slightly during hydrolysis, were not included in the calculation. The determination of the theoretical glucose content of the feedstocks was based on Klason analysis of the feedstock, and assumed all available glucose was present as cellulose (due to prior solubilization of the hemicellulose during pretreatment).  A factor was included in the calculation to account for the hydration of the  cellulose during cleavage. A l l hydrolyses were performed in duplicate, and the results presented are the average of the two values.  2.2.2  Hydrolysis of the delignified water-insoluble fraction  Preliminary hydrolyses conducted to determine the hydrolysability of the delignified Douglas-fir substrate ( W I - A P - 0 , W I - A P - 3 0 ) were conducted in A C buffer. The hydrolyses were conducted 34  at either 2% or 5% (w/v) consistency based on equivalent oven-dried pulp weight, in 150 m L serum bottles containing 50 m L total liquid. In order to assess the fermentability of the cellulose hydrolysate, larger volumes of the cellulosic hydrolysate were required. These hydrolyses were performed in 500 m L Erlenmeyer flasks containing 5% (w/v) pulp consistency in 250 m L total liquid for 72 h. After 72 h, the flasks were placed in a boiling water bath for 10 min to denature the enzymes. The cellulosic hydrolysates ( C H ) were named according to the bark content of the feedstock used to prepare them ( C H - 0 or CH-30), and were frozen at - 2 0 ° C until required for fermentation.  2.2.3  High consistency hydrolysis  In order to generate a cellulose hydrolysate for fermentation with greater glucose concentration than could be obtained at 5% (w/v) consistency, an enzymatic hydrolysis was also performed at 10% (w/v) consistency in A C buffer.  The hydrolysis was performed in 500 m L Erlenmeyer  flasks containing 250 m L total liquid. T o help overcome the problems of poor mixing and endproduct inhibition while ensuring a rapid hydrolysis, a higher cellulase loading (60 F P U ) was used. 6-glucosidase was still supplemented at an I U F P U ratio of 3:1.  2.2.4  Enzymatic hydrolysis in the water-soluble fraction  To increase the initial sugar concentration of the WS-fraction, enzymatic hydrolysis of the delignified Douglas-fir water-insoluble fraction (i.e., W I - A P - 0 ) was performed in the watersoluble fraction, W S - 0 (non-buffered, p H 4.8).  The hydrolyses were performed at 5% (w/v)  consistency in serum bottles containing 50 m L total liquid.  35  2.3  Improving the enzymatic hydrolysis in the water-soluble fraction  To address issues pertaining to the poor enzymatic hydrolysis in the WS-fraction, an O2delignified Hemlock kraft pulp supplied by Howe Sound Pulp and Paper (Port Mellon, B C , C A N ) was used for a series of hydrolysis studies.  A l l subsequent enzymatic hydrolyses  described in section 2.3 utilized this pulp, and were conducted at 2% (w/v) consistency in serum bottles containing 25 m L total liquid, but were otherwise performed as described previously.  2.3.1  Influence of monomeric sugar on the enzymatic hydrolysis  To evaluate the inhibition caused by sugars present in the WS-fraction, hydrolyses were performed in A C buffer supplemented with a single neutral sugar present in the water-soluble fraction (arabinose, galactose, glucose, mannose, or xylose) to achieve a final concentration of 10, 20 or 40 g L " prior to hydrolysis. Control hydrolyses were performed in sugar-free A C 1  buffer. A l l other conditions were consistent with the methodology described previously.  2.3.2  Enzyme stability in the water-soluble fraction  The stability of the cellulase/B-glucosidase preparation  was evaluated  during prolonged  incubation in the WS-fraction. A n enzyme preparation consisting of Celluclast and Novozym188 was added to pulp-free W S - 0 (pH 4.8), A C buffer or S - A C buffer to achieve a final activity of 0.33 F P U mL" and 0.96 I U mL" for Celluclast and Novozym-188, respectively. Aliquots 1  1  (1 m L ) were taken from each enzyme/buffer mixture at t = 0, 1,6, 12, 24, and 48 h, and added to test tubes (in duplicate) containing 1 m L of a 4% (w/v) consistency pulp (hemlock kraft pulp) in A C buffer. The resulting dilution provided a final cellulase and B-glucosidase activity in the test tube of approximately 10 F P U g-cellulose" and 30 I U g-cellulose" , respectively. 1  1  A 1 mL  aliquot was also added to a reaction tube containing no pulp, to correct for sugars derived from  36  the enzyme preparation and the buffer. After 2 h hydrolysis, the reactions tubes were placed in a boiling water bath for 10 min to stop the hydrolysis, and sampled for analysis.  2.3.3 Treatments used on the water-soluble fraction prior to hydrolysis The water-soluble fraction W S - 0 was subjected to a number of treatments, as described below, in an effort  to improve glucose yield during enzymatic hydrolysis of the water-insoluble  component.  Because several treatments had an appreciable impact on the sugar concentration  present in the WS-fraction, anhydrous monomeric sugars were supplemented, as appropriate, to the treated or untreated fractions, to maintain consistent sugar concentration. concentration in the S - A C buffer was also adjusted to be comparable.  The sugar  Experimental controls  were performed in A C buffer.  2.3.3.1  Treatment with anion exchange resin (ANEX)  The WS-fraction was treated with an anion exchange resin ( A G 1-X8, Bio-Rad, Hercules, California) as described by Larsson et al., (1999b). The resin was first washed thoroughly with nanopure water (final resin moisture content = 72.4%), and then added to the W S - 0 until p H 10 was obtained, which required approximately 5.2 g O D W resin per litre of hydrolysate.  The  water-soluble fraction, W S - 0 , was treated for 1 h at room temperature with gentle mixing on a magnetic stir-plate. Following the treatment, the resin was removed by gravity filtration through Whatman #1 filter paper. The treated fraction was adjusted to p H 4.8 with concentrated H2SO4 (Fisher) and centrifuged for 10 min at 3000 x g to remove a precipitate that had formed.  2.3.3.2  Detoxification by overliming with calcium hydroxide (OL)  The water-soluble fraction was "overtimed" based on a method described by Olsson et al., (1995). The water-soluble fraction, W S - 0 , was supplemented with C a ( O H ) ( B D H Ltd., Poole, 2  37  England) to achieve a p H of 10.5, which required approximately 5 g per litre of wood hydrolysate. The mixture was heated to 90°C and maintained at temperature for 30 min with mixing on a magnetic stir-plate. The mixture was subsequently cooled to 20°C in a water bath, and was adjusted to p H 4.8 with the addition of concentrated H2SO4. The treated W S - 0 fraction was centrifuged (10 min at 10000 x g, 4°C), and the pellet was discarded.  2.3.3.3  Extraction with ethyl acetate (EX)  A volume of the water-soluble fraction, W S - 0 , was extracted three times with an equal volume of ethyl acetate (Fisher) in a separatory funnel. Because residual ethyl acetate proved detrimental to the hydrolysis, it was removed via an evaporation step. A measured volume of extracted W S fraction was heated to ~80°C for 10 min with stirring on a magnetic stir-plate to volatilize the ethyl acetate. Distilled water was added back to replace the volume removed by evaporation. A n identical extraction procedure was conducted on the S - A C buffer.  2.3.3.4  Combined overliming and extraction (OLEX)  The water-soluble fraction, W S - 0 , was first overlimed, and then extracted with ethyl acetate, as described previously. Prior to hydrolysis, the treated WS-fraction was centrifuged a second time to remove an additional white precipitate that formed after volatilization of the residual ethyl acetate.  2.4 2.4.1  Improving the starting sugar concentration in the water-soluble fraction Concentrating the water-soluble fractions  The water-soluble fractions W S - 0 and WS-30 were concentrated either by rotary-evaporation (RV) or freeze-drying (FD). R V was performed at 50°C under vacuum on a Biichi Rotavapor  38  (model R E 1 1 1 , Biichi Labortechnik A G , Flawil, Switzerland), while F D was performed by first freezing at - 8 0 ° C , followed by sublimation under vacuum in an Edwards Modulyo freeze-dryer ( B O C Edwards, West Sussex, England). Samples were concentrated slightly more than desired, and were diluted with distilled water to achieve the desired final concentration (2- or 3-fold w/w increase).  2.4.2  Supplementing water-soluble fractions with the cellulose hydrolysates  The water-soluble fractions W S - 0 and WS-30 were supplemented hydrolysates C H - 0 and C H - 3 0 , respectively.  1:1 (v/v) with cellulose  The cellulose hydrolysates were not filtered to  remove insoluble wood and lignin prior to mixing.  In addition to the non-concentrated W S -  fractions, the 2-fold R V concentrates ( R V - 0 , R V - 3 0 ) were also supplemented with the cellulose hydrolysates.  Anhydrous galactose, glucose and mannose were added to the WS-fraction and  cellulose hydrolysate mixtures derived from feedstocks containing bark to bring them to the same starting sugar concentration as their corresponding bark-free mixtures.  2.5 Fermentation 2.5.1  Culturing yeast for fermentation  A spent-sulphite liquor ( S S L ) adapted strain of Saccharomyces  cerevisiae kindly provided by  Tembec L t d (Temiscaming, Q C , C A N ) was used for all fermentations.  This yeast has the  capacity to ferment hexose sugars, but does not ferment pentose sugars.  The yeast was  maintained at 4 ° C on solid G M Y P medium containing 1% glucose, 0.5% malt extract (Difco Laboratories, Detroit, M I , U S A ) , 0.3% yeast extract (Difco), 0.5% peptone (Difco) and 1.5% agarose (Difco).  Cells were cultured in 500 m L Erlenmeyer flasks containing 100 m L liquid  G M Y P at 30°C with constant shaking (250 R P M ) .  Cells were harvested after 24 h by  centrifugation (10 min at 3000 x g), and were transferred to fresh G M Y P for a second culturing 39  stage. Following the second harvest, the cell pellet was washed 3 times with an equal volume of sterile, distilled water, to remove residual ethanol and media components.  A standard curve  relating the inoculum's optical density ( O D , determined spectrophotometrically at 600 nm) and oven-dried weight ( O D W ) was established (Appendix 2), and the inoculum concentration was adjusted with sterile d H 0 such that 3 m L of culture inoculated into 50 m L of the WS-fraction 2  would provide a final cell concentration of 6 g O D W L " . The final dry weight of the inoculum 1  was confirmed by overnight drying of a culture sample at 105°C.  2.5.2 Fermentation of the water-soluble fractions a n d c e l l u l o s e hydrolysates The WS-factions were adjusted to p H 6.0 using 50% N a O H . Fermentations were performed in 125 m L serum bottles containing 50 m L medium supplemented with ammonium phosphate dibasic (Fisher) to a final concentration of 20 m M . The serum bottles were vented using a syringe needle fitted with a 0.22 urn filter.  Control fermentations were run in parallel using  glucose-based media (variable glucose concentration, 0.3% yeast extract, 0.5% peptone).  Malt  extract was not included, due to its high sugar content and probable influence on ethanol production.  Aliquots (0.5 m L ) were sampled periodically for determination of fermentation products and residual sugar concentrations.  Samples were withdrawn aseptically by syringe and were  centrifuged for 4 min at 15000 x g and 4°C. The culture supernatants were frozen at - 2 0 ° C until they were analysed.  A l l fermentations were performed in duplicate and the mean value is  reported.  40  2.6 Analytical methodology 2.6.1 Analysis of the water-soluble fraction The concentrations of monomeric sugars, 2-furaldehyde (furfural), 5-hydroxymethylfurfural ( H M F ) and organic acids were determined using a Dionex D X - 5 0 0 H P L C equipped with an E D 4 0 electrochemical detector (gold electrode), A D 2 0 absorbance detector and autosampler (Dionex Corp., Sunnyvale, C A , U S A ) , as described below.  2.6.1.1 Monomeric sugars Monomeric neutral sugar concentrations (arabinose, galactose, glucose, mannose and xylose) were analysed electrochemically. Sugars were separated using a Carbopac P A - 1 column (4 x 250 mm) equilibrated with 250 m M N a O H and eluted with nanopure water at a flow rate of 1 m L min" (Dionex Corp.). Sodium hydroxide (0.2 M ) was added post-column (for detection) at a 1  flow rate of 0.6 m L min" . Standards prepared from analytical-grade L-arabinose, D-galactose, D1  glucose, D-mannose and D-xylose (Sigma) were used to quantify the sugar concentrations. The sugar L-fucose (Sigma) was used as an internal standard.  2.6.1.2 Furfural, HMF and Organic Acids Furfural and H M F concentrations were determined by absorbance at 280 nm. A n Aminex H P X 87H column (7.8 x 300 mm) was used for separation (Bio-Rad). The mobile-phase was 5 m M H2SO4 at a flow rate of 0.6 m L min" . The column was heated to 65°C. Standards were prepared 1  from analytical-grade furfural (Aldrich Chemical Company, Inc., Milwaukee, W I , U S A ) , and H M F (Sigma). Organic acid concentrations (acetic, formic and levulinic acid) were determined as above, but absorbance was measured at 205 nm. Standards were prepared from analyticalgrade formic acid (J.T. Baker, Phillipsburg, N J , U S A ) , glacial acetic acid (Fisher), and levulinic acid (Sigma). 41  2.6.1.3 G l y c e r o l Glycerol concentration was determined on a Dionex D X - 3 0 0 H P L C equipped with a pulsedamperometric detector ( P A D , gold electrode) and autosampler.  Glycerol was separated on a  Carbopac M A - 1 column (4 x 250 mm) equilibrated with 480 m M N a O H and eluted with 200 m M N a O H at a flow rate of 0.4 m L min" (Dionex Corp.). Standards were prepared from H P L C 1  grade glycerol (Fisher). Erythritol (Sigma) was used as an internal standard.  2.6.1.4 Ethanol Ethanol concentration was determined using a Hewlett-Packard 5890 gas chromatograph equipped with a 6890 autoinjector, splitless injection and flame ionization detector (Agilent Technologies, Palo Alto, C A , U S A ) . Ethanol was separated using a Stabilwax-DA column (30 m, inner diameter 0.53 mm) fitted with a 5 m deactivated guard column (Restek Corp., Bellefonte, P A , U S A ) . For detection, the following parameters were used: injector temperature, 90°C; detector temperature 250°C; carrier gas (helium) flow rate, 1 m l min" . The column-oven 1  temperature was controlled as follows: 45°C for 6 min, ramped to 230°C at a rate of 20°C min" , 1  and held for a duration of 16 min. The ethanol concentration was quantified using standards prepared from HPLC-grade anhydrous ethanol (Riedel-de Haen A G , Seelze, Germany). For the internal standard, 1-Butanol (Fisher) was used.  2.6.1.5 Oligomeric s u g a r s The oligomeric sugar content in the WS-fractions was estimated by post hydrolysis treatment (Browning, 1967). Duplicate samples were supplemented with concentrated sulphuric acid to obtain a final concentration of 3% sulphuric acid. The reaction mixture was heated at 121°C for 1 h in an autoclave.  The concentration of oligomeric sugars was estimated by the change in  42  monomeric sugar concentration following post-hydrolysis, with a correction for sugar losses due to degradation, as determined by H P L C .  2.6.1.6 Lipophilic compounds Total lipophilic compounds present in the WS-fractions were determined by liquid-liquid extraction using methyl tertiary-butyl ether ( M T B E ) .  The WS-fraction (30 m L ) was extracted  with 3 x 30 m L M T B E . The M T B E fractions were pooled, dried by rotary-evaporation (Biichi Rotavapor) and weighed. The calculated concentration of lipophilic compounds was corrected for the inclusion of H M F , which is soluble in M T B E .  N o correction was applied for furfural,  which was volatilized and lost during rotary-evaporation.  2.6.1.7 Total Phenolics The concentration of total phenolics in the WS-fractions was estimated by reaction with FolinCiocalteu reagent (Sigma) (Singleton and Rossi, 1965). The fractions were initially diluted 1:25. A 1 m L aliquot of diluted sample was mixed with 30 m L of nanopure H2O in a 50 m L volumetric flask.  T o this mixture, 2.5 m L of Folin-Ciocalteu reagent was added and mixed.  After 5 min, 7.5 m L of 20% (w/v) NaaCC^ was added, and nanopure water was added to bring the volume to 50 m L exactly. The flasks were incubated for 2 h at 22°C with constant mixing on a magnetic stir-plate. The absorbance of each reaction was measured spectrophotometrically at 760 nm.  Reaction blanks containing distilled water in place of the WS-fraction were run in  parallel, and were used to zero the spectrophotometer.  A standard curve prepared using vanillin  (a common lignin-derived phenolic present in wood hydrolysates)  was used to convert  absorbance values to concentration and estimate total phenolics (Appendix 3). Reactions were performed in duplicate, and the average value is reported.  43  2.6.2 Feedstock analysis The  starting compositions of the wood and bark feedstocks  were determined using a  modification of Tappi Standard T-222 O M - 9 8 . In brief, Douglas-fir whitewood and bark were ground to pass a 40-mesh screen using a W i l e y m i l l .  The ground whitewood was Soxhlet-  extracted with acetone for 7 h to remove extractable components and to minimize the formation of "pseudolignin".  The ground Douglas-fir bark was Soxhlet-extracted using a sequential  treatment of hexane, dichloromethane, acetone, methanol and water for 7 h each. Extracts were dried by rotary-evaporation and weighed. The extractive-free bark and wood was air-dried and were analysed in triplicate for sugar and lignin composition as follows.  Extracted wood or bark (0.2 g) was transferred to a 15 ml test tube in an ice bath. To each tube, 3 m L of 72% (w/w) H 2 S O 4 was added and mixed for 1 min. The test tube was transferred to a water bath (20°C), and the reaction was mixed for 1 min every 10 min. After 2 h, the contents of each test tube were transferred to a 125 m L serum bottle, and diluted with 112 m L nanopure H 0. 2  The serum bottles (containing 115 m L H2SO4 at 4% (w/w) plus wood or bark) were sealed  and autoclaved at 121 °C for 60 min. The hydrolysates were cooled and vacuum-filtered through medium coarseness sintered-glass crucibles. determined by H P L C .  The sugar concentration in the filtrate was  Acid-soluble lignin in the filtrate was determined by absorbance at 205  nm according to T A P P I Useful Method U M 2 5 0 .  A c i d insoluble material (Klason lignin) was  determined gravimetrically after rinsing the solids in the crucibles with 200 m L warm (~50°C) nanopure H 0 , and drying overnight at 105°C. 2  The ash contents of the feedstocks were determined by ignition at 575°C, according to T A P P I standard T-211.  44  2.6.3 Assays 2.6.3.1 Total cellulase activity The total cellulase activity of the Celluclast preparation was determined by the filter paper assay reported by Ghose (1987), and is expressed in filter paper units (FPU). Several dilutions of the stock enzyme were prepared in sodium acetate buffer (0.05 M , p H 4.8), to find a dilution that would liberate slightly more and slightly less than 2.0 mg of glucose during the assay.  An  aliquot (0.5 m L ) of the diluted enzyme was added to 25 m L test tube containing a 0.050 g strip of Whatman #1 filter paper in 1 m L sodium acetate buffer.  The mixture was incubated in a  shaking water bath under conditions used in the experimental hydrolyses (45°C, 150 R P M ) . After 1 h, the reaction was stopped with the addition of 3 m L dinitrosalicylic acid reagent (DNS), and the tubes were transferred to a boiling water-bath for 5 min. The test tubes were mixed vigorously, the reaction was cooled in water, and diluted with 20 m L dH^O. Cellulase activity was determined spectrophotometrically at 540 nm after allowing the pulp to settle, and was corrected with enzyme and reagent blanks. Standards prepared from a glucose stock were used to convert the absorbance into concentration. The assay was preformed in triplicate, and the results averaged.  2.6.3.2  B-glucosidase enzyme assay  The activity of the Novozym-188 preparation was determined colourimetrically using pnitrophenyl-6-D-glucoside (Wood and Bhat, 1988). Briefly, a stock of 5 m M p-nitrophenyl-BD-glucoside (Sigma) was prepared in sodium acetate buffer (50 m M , p H 4.8). From this stock, a 1 m L aliquot was transferred to a 10 m L test tube. Sodium acetate buffer (1.8 m L , 50 m M , p H 4.8) and 200 u L of a diluted enzyme preparation (1:1000 up to 1:5000) were added to the test tube and mixed vigorously on a vortex. The mixture was incubated for 30 m i n at 45°C. Glycine  45  buffer (4 m L , 0.4 M , p H 10.8) was added to stop the reaction. p-Nitrophenol was used as standard. The enzyme activity was monitored by absorbance at 430 nm, and was expressed in international units (IU), where 1 I U = l u m o l p-nitrophenol liberated per minute per m L of enzyme.  2.6.3.3  Xylanase activity  Xylanase activity in the enzyme preparation was determined based on a procedure described by Bailey et al, (1992).  Birch wood xylan (Sigma) was used to prepare a 1% (w/v) solution in  sodium acetate buffer (pH 4.8). A volume (1.8 m L ) was added to a 15 m L test tube, to which 200 u L of diluted enzyme was added. The mixture was incubated for 5 min at 50°C, then 3.0 m L D N S was added to the test tube and mixed. The reaction tube was placed in a boiling water bath for 5 min, cooled in cold water, and then monitored for absorbance at 540 nm. D-xylose (stock concentration 0.01 M ) was used to standardize the reaction. Activity was expressed in international units, I U , where 1 I U = 1 umol reducing sugar (xylose) liberated per minute per m L of enzyme.  2.6.3.4  Mannanase activity  Mannanase activity in the enzyme preparation was determined as for xylanase activity, substituting ivory-nut mannan (Megazyme International, Ireland) as substrate, and D-mannose for use in standards.  2.6.3.5  Hydrogen peroxide assay  Hydrogen peroxide used for delignification was titrated with sodium thiosulphate (Fisher) prior to use, to confirm its concentration. The concentration was calculated based on titration of two known amounts (-0.1 g and 0.15 g) of H 0 2  2  (30% w/w). The H 0 2  2  sample was added to a 46  beaker containing 50 m L distilled water, 10 m L  H2SO4  (10% v/v), and -0.5 m L saturated  ammonium molybdate solution ( B D H ) . Immediately following the addition of H2O2, an aliquot (2 m L ) of K I (10% solution) was added.  The reaction was titrated with 0.1 N sodium  thiosulphate to a reddish-brown colour, whereupon 1 m L of starch indicator (5 g soluble starch boiled in 100 m L dH^O) was added. Titration was continued until no blue colour could be seen. The procedure was repeated for a blank (water instead of peroxide). The peroxide concentration was then calculated according to Equation 2.1: [ H 0 ] (% w/w) = ( A - B ) x N x 1.7007 _ ., . , Peroxide weight 2  2  (£<Q> •^•1  J  where " A " is the required volume of titrant for the sample, " B " is the required volume of titrant for the blank, and N is the normality of the sodium thiosulphate solution (determined by standardization with 0.1 N potassium iodate solution).  47  Chapter 3: Pretreatment of Douglas-fir Whitewood and Bark Chips by Steam Explosion 3.1 Background Historically, pretreatment strategies for the bioconversion of lignocellulosic to ethanol have primarily focused on optimizing conditions for the recovery of the cellulose component in a readily digestible form, with little regard for hemicellulose or lignin recovery. For softwoods, it has previously been shown that relatively severe pretreatment conditions are required, due to the recalcitrant nature of the substrate (Clark and Mackie, 1987; Schwald et al,  1989).  However,  there are major drawbacks to this approach. Not only are significant quantities of the more labile hemicellulose  sugars degraded  under severe pretreatment, but  the  production  of  both  hemicellulose- and lignin-derived decomposition products and their localization to the watersoluble fraction can severely impact its fermentation to ethanol.  W i t h the recognition that  significant sugar losses and poor fermentability of the hemicellulose-derived sugars after highseverity pretreatment would be unacceptable to a commercial process, there has been a shift towards the use of less destructive pretreatment conditions. In general, by employing less severe pretreatment conditions, it should be possible to recover greater quantities of the available hemicellulose sugars. However, it is recognized that the conditions that permit optimal recovery of the hemicellulose sugars may also result in a cellulose component that is poorly digested during enzymatic hydrolysis. Thus, a compromise between hemicellulose sugar recovery, watersoluble fraction fermentability and cellulose digestibility is probably required.  Most research into the bioconversion of wood residues to ethanol has explored the use of whitewood (i.e., non-bark) residues only. However, the use of "clean" whitewood feedstocks might not represent either a realistic or an economically viable source of carbohydrate for industrial bioconversion.  More suitable residues such as low-cost m i l l wastes (e.g., sawdust,  48  shavings, and slabs) and whole-tree forest "thinnings" (Nguyen et al, 1998; Nguyen et al, 1999) can be anticipated to contain some bark. However, there is currently little reported about how the bark content in these feedstocks might influence the overall bioconversion process.  The chemical composition of bark is radically different from the whitewood derived from the same tree species, and the high lignin, ash and extractive contents, combined with reduced carbohydrate content (Fengel and Wegener, 1989; Kurth, 1949) were anticipated to have a detrimental impact on the various sub-processes of bioconversion. O f concern was the impact that bark might have on the pretreatment of whitewood feedstocks.  Increased inorganic  compounds might possibly neutralize a portion of the acid catalyst used during pretreatment as has been suggested previously (Saeman, 1945; Torget et al, 1991), while increased quantities of non-carbohydrate  components  such as lignin and cork might restrict hydrolysis of the  carbohydrate fraction resulting in decreased hemicellulose solubilization and monomeric sugar recovery.  Furthermore, increased lignin content in the water-insoluble fraction might have a  detrimental impact on enzymatic hydrolysis ( M c M i l l a n , 1994; Sewalt et al,  1997).  The  inclusion of bark in the feedstock was also expected to render fermentation of the water-soluble fraction more difficult, due to the accumulation of naturally-occurring inhibitors such as extractives (e.g., resin/fatty acids, esters, waxes, etc.) and phenolic compounds (Ando et  al,  1986; Clark and Mackie, 1984; Tran and Chambers, 1986).  The work described in this chapter of the thesis addresses issues pertaining to pretreatment and sugar recovery from Douglas-fir whitewood and mixed whitewood/bark feedstocks, and is divided into two parts. The first part of the work reports on the sugar recovery from a Douglasfir whitewood feedstock, steam exploded under three pretreatment conditions: low-, mediumand high-severity. The second part reports on the sugar recovery and composition of the water49  soluble fraction derived from Douglas-fir whitewood feedstocks supplemented with Douglas-fir bark to a final concentration of 10%, 20% or 30% (w/w). For the purposes of comparison, a 100%-bark feedstock and 0%-bark feedstock (i.e., 100% whitewood) were also prepared.  The  impact of bark on the recovery of naturally-occurring inhibitors in the water-soluble fraction is also reported.  3.2 3.2.1  Results and Discussion Pretreatment of Douglas-fir Whitewood  3.2.1.1 Feedstock Composition Douglas-fir whitewood chips were prepared for steam explosion as a mixture of sapwood and heartwood chips (2:1).  A 2:1 ratio of sapwood to heartwood was selected to reflect the  anticipated proportion of sapwood and heartwood residues available for a putative commercial process using softwood residues, and not the expected composition of whole-tree feedstocks for bioconversion, which might vary considerably with tree age and size. Previously, differences were reported for the conversion of Douglas-fir heartwood and sapwood feedstocks, both in terms of susceptibility to pretreatment and enzymatic hydrolysis (Boussaid et al, 2000). Based on this work, it appeared that sapwood would be an easier feedstock for bioconversion. Realistically, however, it would be impractical i f not impossible to fractionate heartwood and sapwood chip residues prior to use. In order to evaluate the recovery of sugars following steam explosion, it was necessary to first establish the chemical composition of the mixed heartwood and sapwood feedstock (Table 3-1).  To simplify the interpretation of the results, the term  "hemicellulose sugar" in the subsequent sections refers collectively to arabinose, galactose, mannose, and xylose.  Although glucose is a constituent of Douglas-fir hemicellulose (i.e.,  galactoglucomannan), glucose recovered from the hemicellulose component is indistinguishable from cellulose-derived glucose.  Thus, the glucose recovered in the water-soluble fraction is 50  Table 3-1: Chemical composition (%) of the original Douglas-fir whitewood feedstock used in the pretreatment severity experiments. Values in parentheses are standard deviation (n=3).  Arabinan  Galactan  Glucan  Mannan  Xylan  Total Carbohydrates  Klason Lignin  Total Yield  1.2 (0.1)  2.5 (0.1)  48.8 (0.3)  13.3 (0.2)  3.7 (0.1)  69.5 (0.4)  29.7 (0.3)  99.2 (0.5)  51  considered within one recovery term for the whole feedstock, and is not included as a hemicellulose sugar.  The composition of the feedstock emphasized the fact that tailoring pretreatment conditions towards the optimal recovery of hemicellulose sugars would be beneficial, due to the high hexose sugar content. M o r e than 76% of the available hemicellulose sugars were hexose sugars, which can be readily fermented to ethanol using conventional strains of yeast (Table 3-1). Combined, galactose and mannose accounted for almost 16% of the original  feedstock.  Pretreating the feedstock under conditions optimized for efficient cellulose hydrolysis would result in the loss of a significant portion of these hexose sugars, and ultimately a reduction in the quantity of ethanol that could be produced.  Optimizing the pretreatment conditions for  hemicellulose sugar recovery would also provide increased recovery of pentose sugars, as these sugars are especially labile and easily degraded during acid hydrolysis and steam explosion pretreatments. In the case of the Douglas-fir whitewood feedstock, optimum recovery of xylose and arabinose should provide approximately 49 kg of carbohydrate per tonne of Douglas-fir whitewood processed. However, it should be noted that the current research is concerned only with the fermentation of the hexose sugars to ethanol. It was thought that the low total pentosan content (less than 5% of the feedstock) did not warrant the use of a xylose-fermenting organism, as xylose fermentation remains technologically more challenging than hexose fermentation. However, there has been significant progress recently in the development of robust and efficient xylose-fermenting microorganisms (Ho et al, 1999; Jeffries and Shi, 1999). It may ultimately prove worthwhile to also consider this carbohydrate fraction in addition to the hexose sugars, even though the increased ethanol production would not be considerable. Work was conducted under the assumption that i f efficient fermentation of the hexose sugars was possible, then xylose fermentation by a robust recombinant organism would likely be possible as well. 52  3.2.1.2  Solids Recovery (Shot Yield)  The Douglas-fir whitewood feedstock was pretreated under three conditions, selected based on the previously determined optima for hemicellulose sugar recovery (low-severity, log R = 3.08) 0  and cellulose hydrolysis (high-severity, log R = 3.76) of a Douglas-fir whitewood feedstock 0  (Boussaid et al,  1997).  In addition, a compromise condition halfway between the low- and  high-severity treatments was included (medium-severity, log R  Q  explosion, the whitewood slurries were recovered and analysed.  = 3.45).  Following steam  The selected pretreatment  parameters had a significant impact on the recovery of the starting material. The shot yield, which is calculated by dividing the dry weight of recovered material by the dry weight of the feedstock  (300 g), was greatest for the low-severity pretreatment (Table 3-2).  Higher  pretreatment severity resulted in the reduced recovery of the original materials, with a reduction to 86% and 75% under medium- and high-severity pretreatment, respectively. Shot yields lower than 100% reflect decomposition of the original feedstock to volatile compounds such as furfural and formic acid, which arise due to acid-catalyzed degradation reactions (Brownell and Saddler, 1987; Brownell et al,  1986; Clark and Mackie, 1987).  In addition, water liberated from  carbohydrates during decomposition to H M F and/or furfural would also be unaccounted for, resulting in additional mass losses.  Many of these compounds may be lost during the  pretreatment step because the receiving vessel is open to the atmosphere.  Furthermore, volatiles  retained in the solids and liquid fraction may be lost during oven-drying of the material to determine dry weight recovery of the feedstock, and consequently a lower calculation of the shot yield would result.  In addition to acid-catalyzed decomposition, pyrolysis reactions during  steam explosion have also been reported to decrease feedstock recovery. These reactions yield mainly water and carbon dioxide from the feedstock, but do not contribute greatly to yield losses at temperatures less than 200°C (Brownell, 1985; Brownell and Saddler, 1987).  Thus, it is  unlikely that pyrolysis reactions contributed significantly to the reduction in shot yield for the 53  Table 3-2:  Shot yield recovery of the feedstock (%, based on oven-dried weight) following pretreatment of the whitewood chip feedstock under low-, medium-, and high-severity. Severity Factor  ws  a  WI Shot Y i e l d b  a  b  Low  Medium  High  11.3 81.0 92.3  12.3 73.7 86.0  16.0 59.0 75.0  W S : Water-soluble fraction W I : Water-insoluble fraction  54  low- and medium-severity pretreatments.  A t high-severity, pretreatment, however, these  reactions may have played a role in reducing the shot yield.  A small percentage of the losses  may also be accounted for as material baked on to the receiving vessel and thus not accounted for.  However, the vessel was rinsed with distilled water during collection of the pretreated  material, in an effort to minimize these losses.  As can be seen from Table 3-2, when greater pretreatment severity was employed, an increased solubilization of the starting feedstock was observed. In fact, - 2 7 % less material was recovered in the water-insoluble component following pretreatment under high-severity conditions than was obtained after low-severity treatment. This trend was expected, as it is generally recognized that steam pretreatment can effectively solubilize the hemicellulose, as well as the cellulose component, depending on the pretreatment conditions employed. Increased solubilization of the cellulose component did not translate into a significant increase in material recovered to the water-soluble component, owing to significant decomposition and losses of material, due to the mechanisms described previously.  3.2.1.3  S o l u b l e S u g a r Recovery  Having established the original composition of the feedstock and the shot yield, it was possible to determine the sugar recovery from the feedstock following pretreatment under each of the three severity conditions. The slurry recovered following steam explosion was adjusted to 5% (w/w) consistency prior to fractionation, to aid in recovery of the soluble carbohydrate.  The  sugar content in the water-soluble fractions was determined and is presented as combined oligomeric and monomeric sugars in Table 3-3.  55  VO  8 .2? CO  73 -C CJ  OO  T-J  O  CJ  G  3  3  CO  O  a  x S  03 N  73 CO i-  c  CO  <+-!  <D  O  o CJ '  S  G  RT  .2  S 05  >  ,- II  u  o  3  I  fc ^ ~ ^  °  oo  X  .5  ^03 3  "  §  .co  oo  73  CO  ° £ c-o •a '5b -S  5  'C <" CD oo CO CJ  O  (-i  6 3  X  c  « 2CO oo *r  G -C CD  a,  00 .  <u CJ  X  oo  CJ VH CO oo  o o  73 CJ  OJ  C  CJ =3  'G  .  G  11  ON  G  <0 ^  .5  cd  o  O  r—H  vo oo  X  CO 03  co  CJ  CJ oo  VO CO  VO  » G CO O 73 O<'o3  oo O  IX  S3  G 03  CJ  G  X  o  "c3  c 'Eb 'C  G C 03  o  VO  cn  oo m'  >+H  O  G 13 >> G G o3  "ST  g - g  JS *s  vo CJ oo  cd  •4—»  O  G  O  o l-l  »  03  *03  CO  a  OH  oo oo cn O oo m vo VO  G G  o  03 73 CO  O  O  CN 00  m  cn  03  oo  .3  3  co •a oo O  CJ  CJ  GO  oo T3  cj  fc o£ O a s *a  O  >  O  o  CO oo  i-  CJ  ^  oo  CN  03  03  S -S  £^  VO  PL,  SH  CJ 03  •S 503 t- "  CJ  G X 03  U  O  O  CJ O CO G > , ISO &  cj co o 53  2&  w  CJ 00  CJ  CJ 4 —» > CJ G O <L> CJ  O  CO CQ CJ Cij,  O  O  * cc!  ON  OH  X! G  00  •a  H  cn  'Hb  03 G  in  O  VO  =j  CJ  co  r-o  00  *  3  i—l  co  X 00  IX 00  G  '1  'T3  x  co  a * O  1  <+5 oo  oo CO  .3 >  43 00 3  Q  Pretreatment under low-severity conditions resulted in high recovery of hemicellulose sugar from the feedstock to the water-soluble fraction.  Each o f the hemicellulose-derived sugars  (arabinose, galactose, mannose and xylose) was recovered with yields i n excess of 85%, with a cumulative hemicellulose sugar recovery of 87%. A s pretreatment severity increased from low to high, sizeable decreases in the recovery for all hemicellulose sugars were noted.  The  percentage of recovered hemicellulose sugars decreased from 87% to 64% and 43% for the low, medium- and high-severity samples, respectively. This observation is best explained by the balance between hydrolysis o f oligomeric/polymeric carbohydrate and secondary decomposition of the soluble monomeric sugars. Under higher severity pretreatment, destructive side reactions (e.g., dehydration/condensation) begin to play a greater role in determining sugar recovery than do hydrolytic reactions, and consequently, a decrease i n sugar recovery is observed (Clark and Mackie, 1987; L a i , 1991). Under medium-severity pretreatment, hemicellulose degradation was reduced, but losses were still significant compared to the low- severity pretreatment. This is understandable as, based on the calculation of pretreatment severity ( R ) we should expect 0  comparable pretreatment severity (and thus sugar recovery) to that obtained with low-severity pretreatment after only 2 min at medium-severity pretreatment temperature (195°C).  Limited  polymeric hemicellulose sugar should remain after this time, based on the high yield of sugars in the low-severity fraction (-87%). With minimal new hemicellulose sugar being liberated, but an additional 2.5 minutes reaction time remaining (total 4.5 min), it is not surprising that acidcatalyzed decomposition results in a net loss of hemicellulose sugar. These reactions may yield furans (e.g., H M F , furfural), acids (e.g., levulinic, formic), reversion products, anhydro sugars, and lignin-like products due to condensation (Lai, 1991), and likely account for the decrease in both sugar and solids recovery ("shot yield") following steam explosion.  57  In contrast to the hemicellulose sugars, the quantity of glucose recovered in the water-soluble fraction actually improved as pretreatment seventy increased from low to high. This trend was due to the increased hydrolysis and solubilization of the cellulose component under more severe processing conditions (Table 3-2), and should not be confused with increased survival of glucose from the hemicellulose component. Decomposition of glucose liberated from the hemicellulose fraction was more than offset by the increased solubilization of the cellulose component. Consequently, a net increase in glucose was observed.  3.2.1.4  M o n o m e r i c S u g a r Recovery  The proportion of monomeric hemicellulose sugars in the water-soluble fractions was dependent on pretreatment severity (Table 3-4).  This trend was expected as, in the same way that the  hydrolysis of polymeric carbohydrates  is governed by treatment time and temperature  parameters, so too is the secondary depolymerization of oligomeric sugars to monomers (Adams, 1965; Browning, 1967).  Thus, as pretreatment severity increases, a greater proportion of the  oligomeric sugars will be depolymerized to monomers (Chum et al,  1990; Clark et al, 1989;  San Martin et al, 1995; Zhuang and V i d a l , 1997). The percentage of individual hemicellulose sugars present as monomers in the low-severity water-soluble fraction ranged from 83% to 89%. Mannan is generally more resistant to acid hydrolysis than are other hemicellulose polymers, which likely explains the lower monomeric sugar content in the low-severity water-soluble fraction. In contrast, the pentosans are hydrolysed much more readily, and a higher monomeric sugar content was observed for arabinose and xylose at all pretreatment conditions.  A s the  pretreatment severity increased, the percent monomer content for each hemicellulose sugar increased slightly.  Collectively, the percentage of hemicellulose sugar monomers increased  from 85% to 94% over the conditions tested. H i g h monomeric sugar concentration is important for the bioconversion process, as the microorganisms commonly used for fermentation typically 58  Table 3-4: Percentage of monomeric hemicellulose sugar content contained in the Douglas-fir water-soluble fractions prepared under low-, medium- and high-severity pretreatment conditions.  a  Severity Factor  Arabinose  Galactose  Mannose  Xylose  Combined HemSug  Low  89  85  83  81  85  Medium  91  87  85  90  87  High  96  92  95  92  94  3  HemSug: hemicellulose sugars (Arabinose, Galactose, Mannose and Xylose)  59  lack the metabolic pathways required for metabolism of oligomeric sugars.  The lower  monomeric sugar content of the low-severity water-soluble fraction could be improved slightly through the application of a post-hydrolysis step with dilute sulphuric acid (Shevchenko et al, 2000). However, any improvement would be.small, owing to the fact that almost 85% of the available sugars were already in a monomeric form. Consequently, such a treatment might not be worthwhile for the current samples.  Although the monomer content in the high-severity  water-soluble fraction was much better, it is important to consider that under these pretreatment conditions, significant sugar decomposition has already occurred.  A s w i l l be discussed later,  these decomposition reactions can have a pronounced and detrimental effect on fermentation.  3.2.1.5  C o m p o s i t i o n of the water-soluble fraction  In previously described experiments, the slurry of steam-exploded wood was adjusted to 5% (w/w) consistency prior to fractionation. This was done to improve recovery of the solubilized sugars to the water-soluble fraction while still permitting accurate analysis of the soluble sugars. When a greater consistency was used, a significant portion of the soluble sugars was retained in the water-insoluble component (for example, at 15-20% consistency, about 30% of the total solubilized sugars remained in the solids). This arose partly because complete recovery of the liquid from the solids fraction was not possible by vacuum filtration, but also due to the interactions between the soluble sugars and the cellulose/lignin matrix.  Although the use of  dilute slurries was advantageous for the quantification of sugar recovery, the resulting watersoluble fraction was consequently very dilute and was not suitable for fermentation. The sugar concentration in the low-severity water-soluble component fractionated at 5% (w/w) consistency was less than 10 g L " combined sugars. 1  In order to generate a liquid stream with a sugar  concentration appropriate for fermentation, three new wood feedstocks were steam exploded under low-, medium- and high-severity conditions. The recovered slurries were adjusted to a 60  20% (w/w) consistency prior to fractionation. While this enabled the recovery of most of the soluble sugars, some soluble sugar remained in the water-insoluble fraction. This illustrates a key dilemma for bioconversion: it is difficult to simultaneously achieve high yield recovery of sugars and high concentration of sugars in the water-soluble fraction.  This dilemma w i l l be  addressed in more detail later.  When the water-soluble fractions were analysed, and their compositions determined (Table 3-5), it was apparent that the pretreatment severity greatly impacted the resulting sugar concentration in the water-soluble fraction. The concentration of monomeric hemicellulose sugars (arabinose, galactose, mannose and xylose) decreased notably as pretreatment severity increased, while glucose concentration improved, following the same trend described previously for sugar recovery. The net effect was an increase in the sugar concentration of the water-soluble fraction, as severity increased. However, the maximum sugar concentration attained was still less than 50 g L " in the high-severity fraction, in terms of total sugars, and less that 46 g L " in terms of 1  1  readily fermentable hexose sugars.  This sugar concentration is many times more dilute than  would be desirable for commercial ethanol fermentations. For example, in a typical commercial batch fermentation, starting hexose sugar concentration may exceed 200 g L " , although most are 1  in the range of 100-150 g L " . 1  The loss of greater quantities of hemicellulose sugars under more severe pretreatment generally translated into increased concentrations of the furans, H M F and furfural, after treatment at medium- and high-severity conditions.  It has been  shown that the decomposition of  carbohydrates is directly related to processing temperature and duration of exposure (Saeman, 1945). Thus, for an increase in reaction severity, a greater production and recovery of furans in the water-soluble component should be anticipated. Although these compounds are not the sole 61  CN VO  o  3 T3 O  PL,  O  O  O  o  o  o  as  00  ©  o  o o  oo o  CO CN  T t  T t  CO  o  o  ©  r-  T t  o  3  VO CO  o  -1—I •*—»  cd  T3 cd  cd  OJO  co  T3  cd bo  3  PH  1—1  GO  cd  •*—»  O  co oo  O  X  "o  CD  co  O 3 3  00 I  4—»  ON CO  CO  ON  T t  T t  o o  o o  o  o vd  T t  in  *—(  in CN  oo  CN o  o  cd  ON  CM  in  —<  ©  ©  in  o  00  CN CO  OO  cd  co  B u c "o  3 U  • 1s CD 00  4—> cd <U  cd CD  .3  o  e o 00  o  OH  s o  U  JS 3  CD  3  O  o © o o  3  co  Cd*  >>  O  CD CD  o o  OH  00  3  •_  4—>  CD  •-4o—r—I 1 cd  CD  oo O  £50  -u  CD  3  in Id  CO  co  o o vq co  — i  o  o o  CO  CN  o ©  o o  o ©  ON  T f  O  *- fa  o  •c 2 CD tj  I—I  3  a  > CD 00  CD  oo  JS  bo  O 3  • i—1  JS T3 3  -O  cd  cd  < j  > cd  CD  [L  3  B  JS  .5? K  sugar degradation products isolated from lignocellulosic hydrolysates (common compounds include levulinic acid, formic acid and additional furan-based compounds), they are two of the major products derived from hexose and pentose sugars (Lai, 1991).  H i g h concentrations of  H M F and furfural have been implicated in the inhibition of fermentation (Banerjee et al, 1981a; Nilvebrant et al, 1997; Taherzadeh et al, 2000), and it has been suggested that these compounds are good indicators of how fermentable a given lignocellulose hydrolysate may be (Taherzadeh et al,  1997a).  feedstock,  It was apparent that, in addition to providing greater sugar recovery from the  a lower severity pretreatment had the additional benefit of reducing inhibitor  production.  This was expected to translate into significantly improved yields and rates of  fermentation.  3.2.2  Pretreatment of Douglas-fir Whitewood Feedstocks Containing Bark  A s was shown in the previous section, Douglas-fir whitewood feedstocks could be effectively pretreated by SC>2-catalyzed steam explosion. However, it was recognized that the low-value wood wastes appropriate for commercial bioconversion would probably contain some amount of bark.  A s described earlier, compositional differences between bark and whitewood were  anticipated to affect pretreatment of the feedstock, and the composition of the resulting watersoluble and water-insoluble fractions. Although low concentrations of bark might be tolerable, appreciable quantities (e.g., greater than 10%) were anticipated to hinder the conversion process. Experiments were thus designed to evaluate the impact of bark on the bioconversion of whitewood.  3.2.2.1 Feedstock Composition Feedstocks for pretreatment were prepared from chipped Douglas-fir whitewood and Douglas-fir bark.  The chemical compositions of both chip stocks were determined prior to steam 63  pretreatment (Table 3-6), and it can clearly be seen how significant the differences are between the whitewood and bark feedstocks are.  A l l sugars, with the exception of arabinose, were  present at higher concentration in the wood feedstock.  The bark feedstock  contained  significantly more acid insoluble material and ash than were detected in the wood, while the extractive content of the bark was about 17 times greater than that found in the whitewood. Based on these results, it was clear that, even i f bark played a negligible role in influencing pretreatment conditions, there could still be a sizeable discrepancy in the chemical composition of the water-soluble fractions.  Following impregnation with SO2, the mixed whitewood and bark feedstocks were pretreated under medium-severity conditions. Previous work with Douglas-fir whitewood revealed that the hemicellulose sugar recovery was greater after low-severity pretreatment. Despite the desire to maximize hemicellulose sugar recovery from the feedstock, parallel work with Douglas-fir whitewood revealed that low-severity pretreatment conditions resulted in the  ineffective  enzymatic hydrolysis of the water-insoluble component, and low overall sugar yield from the feedstock (Boussaid et al., 2000; W u et al, 1999). Accordingly, a compromise in pretreatment was deemed  necessary.  Thus, more severe  conditions were selected to affect  greater  pretreatment of the feedstock, and to allow improved enzymatic hydrolysis of the waterinsoluble component in subsequent experiments.  3.2.2.2  S o l i d s R e c o v e r y (Shot Yield)  Following steam explosion, the wood and bark slurries were recovered and analysed.  The  selected pretreatment parameters for the wood/bark mixtures resulted in a high recovery of the starting material (Table 3-7).  A s reported earlier, shot yields less than 100% reflect the  decomposition and loss of the original feedstock due to the formation of volatile compounds 64  m vo  o  VO  10 o  © o* GO  00  ON  ON  ^ ©  o o  1-- —1  ^  o  o © © VO  >H  CO  t>0  4-»  03  ON  ^  CN  2,  00  m  j-:  c  2  03  "o _>  *-— t»  o o  o SH  C  i n VO  03  — -t »  ON CN  s  x  co"  W  c o  c  aa 03 03  CN  2  o  C  ON  X CS JH Td  >,  O -Q  SH 03  u  c  o3 O  cn O  10  1  CN  ©  00  ~H  ^  o3  co" C 03 X  o  O  >H  o  . cn  T3  co"  3  c  c  03  03  o  CN  O  C  HO CS  ^  ©  r-5  g  2  c  03  X <D  o  VO  o  T3  CO  3  4H "o3  -4 o o +->  0)  o o  o 03  00  _  O X  CN'  g -3  O 03 H  M co X ; 4H  3 3  •4—> 4-J  'S x  o o  c  00 C  * •X:  03  PQ  cp co o 03 o co „  X  PP  3  CO  00  T3  -S 0 <  VO VO  o o  Tt  CO  > <  o  oo  co oo  <-< ON  ©  oo  oo  i-H  f«  ON  >n  o  in  in tr] ©  Tt'  co  ON  1—1  I/O  vo  in  Tt  _ d ON  o o  »  -a  in  CN  CO  d ON  —i  CN  CO  oo  O  CO  co o  vd  CN  *—i  a  co  -I—»  c o  U  00 oo  CQ bJO >-  co >  ro od f~- 00  ro oo  © ©  VO  Tt  o  i-H  ON  l>  00  Ii o  l>  ON  00  1—1  VO  1> 1>  oo  •3 cd  cd ^  CO — ' oo  2 00  t-n  >. o  si  GO  O 00  C  fe fe cd cd GO  M  during the steam explosion step, and also lost during oven drying for the determination of dry weight (Clark and Mackie, 1987). A s can be seen from the data in Table 3-7, shot yields were greater than 87% for all feedstocks, and exhibited a slight upward trend with a greater percentage of bark in the feedstock. Based on the above explanation regarding material losses, these results suggest that bark may have had an influence on the pretreatment step.  A reduction in the  formation of volatile decomposition products such as furfural, formic acid and lignin-derived phenolics could explain the observed increase in shot yield. Although this effect might be due to the mitigation of pretreatment severity by the inclusion of bark, it was more likely related to the decreased quantities of carbohydrate in the bark component that might give rise to volatile decomposition products.  Alternatively, the loss of these volatiles may have been reduced in  feedstocks containing bark due to physical or chemical associations (e.g., condensation reactions) with the increased quantities of lignin in the insoluble matrix. Regardless of the actual cause, it was apparent that the observed effect was relatively minor, particularly at low bark contents.  As the percentage of bark in the feedstock increased, a generally smaller percentage of the feedstock was recovered in the water-soluble fraction (Table 3-7).  This observation can be  explained by the decreased carbohydrate content of the bark feedstock, which was less than 1/3 of that available in the whitewood feedstock. A s the bark content increased in each feedstock, the concentration of available sugars decreased accordingly.  Steam pretreatment has been  shown to primarily liberate carbohydrate from the feedstock and, under milder pretreatment, mainly hemicellulose sugars.  In contrast, lignin has been reported to be poorly solubilized  (Clark and Mackie, 1987), and does not usually account for a significant portion of the watersoluble fraction's dry weight.  However, other compounds in addition to carbohydrate (and  carbohydrate degradation products) account for some of the material present in the water-soluble 67  fraction.  This non-carbohydrate material includes lignin-derived phenolic compounds, water-  soluble tannins and extractives derived from the bark during pretreatment.  3.2.2.3  S o l u b l e S u g a r Recovery  To determine whether or not the inclusion of bark had a detrimental impact on the recovery of carbohydrate from the feedstock, soluble sugars (oligomeric and monomeric) were quantified in the water-soluble stream. Previously, the whitewood slurries obtained under low-, medium- and high-severity had been adjusted to 5 % (w/w) consistency prior to fractionation, to ensure virtually all of the soluble sugars were recovered from the water-insoluble component.  In this  case, the recovered wood and bark slurries were adjusted to 1 5 % (w/w) consistency (to allow fermentation of the water-soluble component). The use of a relatively concentrated slurry meant that complete recovery of the soluble sugars from the cellulose/lignin matrix would not be possible in a single-stage fractionation.  Consequently, the solids were repeatedly washed with  fresh water at 20% (w/w) consistency until only trace quantities of sugar were detected in the wash. This may have improved sugar recovery compared to the single fractionation at 5 % (w/w) consistency, but the difference was not expected to be significant. The influence of bark on the recovery of sugars following pretreatment was determined using a whitewood-only feedstock (0%-bark), and a feedstock containing 30%-bark by weight (Table 3-8).  The recovery of soluble sugars from the whitewood feedstock was comparable to the results obtained previously after medium-severity pretreatment (Table 3-3).  Differences in percent  recovery were noted for individual sugars and likely reflect subtle differences in the composition of the Douglas-fir feedstock.  In addition, factors such as moisture content and chip size/density  would be expected to impact stream pretreatment (Brownell and Saddler, 1984; Brownell et al, 1986), and these factors could not be completely controlled.  Overall, slightly lower soluble 68  J2 . S  ^£ |  B  ,  03  CN  i#  Tf  VO C O  o o  d  d  d  ON  VO  *—<  O  CN d d  vd  g <k  VO  CJ  ON  o O  d r-H  o3  H  CO  oo  ©v. ©  00  ON  ON  CJN  co co co co co  d  d  d  d  00  00  VO  CN  r-H  H  d  in  CN  CJ  d  •ac  or; rt  d  d  in r-H  in 00  CN  ^  o '  '  "O  CJ  x  CN  O  d o d o CN  O  O CJ  (N  o 03 03 d rH  CO  O C O r-H f_ r-! O O  in 2-  Tf  ^  O  & 8  CJ rH  ^H  Ol © ©  Tf  vd  ON  ON  ON  CO  C O C O C O CJ  d> d> d d  CJ  CJ CJ  rH  SH  03 03  in r - H C O •> ' d co —  H H  ON  in  VO  00  VO  O oo  •rd  I, 3  CJ  o  r-H  .5 -3 -9  -5 cu  S <u o -c ^ c  p  h > O  oo  p  CO Tf  |  t/3  ^ CJ oo  o > 03-  £  o  03  "d  oi O 03 CJ 6 0 <JH oo • rH H-H CJ  O  O  -a  T3  o3  C  £  S  s S  SH  03 ft  g G.S o3  S IH  +-  03  ^  1  t>0  •e ^ G3 5=3  a 00 I  ©  00  CN  00  d  vo i n C N ^ - i d o d o  CO  Tf  v©  r-H  d  CO  _  o CO Tf  CJ O0  o  4—» CJ  B 13 O CJ OO  C X  F H  rH  — i  03  !H  *-H  ©  vo vd  ©  r-H  d  Tf  CO  vd  rt  CJ  CO  o d  2 ^' ^ o ©  oo  CO  in  o co  CN CO  O ON  O o T f  O  ©,  o CN  CO  O  CO  r-J  CO  vo  CO  CO  O  ©  00 Tf  d  co CN CN  < -°  VO  R—  d vq  O O O O 00  - H  CC)  r H  CO  r-I  O  O  d  vd  ir]  Tf  Tf  d vo  vo ©  O  d  CO  oo  T f  r-1  d  d  Tf  d  co  CO  . t r-H  CJ CJ  03  ^ h  O  Tf  o3  o  CN  GO  C O T f l O VO  o 00 d d i n CN r - H p  H O O O  CO  S S S S S CN  ON ON C O  ON  CN  03 C-H  O O O  si si si oo  03  oo oo  03 03  si si oo  03  03  oo ,  03  £££££ £  =3 CO  CN  OO  © ON VO  Tf  Tf  d  ©  in co"  IJO  CN CO T f  -2 o  H  si sioo si oo oo  •s  CO  pq  3 H  o col  03  £  oo od  03 od  £  £  £  in vo X ! si oo  £  c o o o3  Tf  'o  oo  CU  si  oo  ©  CN IT)  oo CJ  CU  Tf  d  O  CN CN CN  TT  Tf  m Tf o o d d d d  Tf  i n  +H  OH  CN  CN r - H r - l O O O O O  ON  ©  >n  CN  o d  oo  d  d  £ ^ -d ^ >  <^ Q CN ^ O O O O O O  ©  . H 9 o o o o  rt  CJ  CO  Tf  O  CJ  3  CN in  S  in o d 03  o  OH >  X!  u  d o d o CO  & o  " -2 .g  W  CO  —H  • Si  U  CO  X  3  c o -a  °  CN O  Tf  ° ^ s  CJ  o d C N ^ ~ £ in oo  SH  2  S  i—i  r3  5^ ^ •> x cu  o3 B  vo co  o o o o  oo  03 03  £  i  rC  rH  O0  CU  cd  4->  O  o  6 00  03  00  glucose was detected (Table 3-8) when compared to previous work (Table 3-3), while combined hemicellulose sugar was comparable. The inclusion of bark (30%) in the feedstock resulted in slightly decreased sugar recoveries compared to whitewood (Table 3-8).  However, the effect  was less pronounced than originally anticipated. It was thought that the high quantities of ash and lignin might hinder the pretreatment process, resulting in significantly reduced sugar recoveries.  On the basis of available hemicellulose sugars in the feedstock, just over 62% of  these carbohydrates were recovered from the whitewood feedstock.  In contrast, only 55.5% of  the hemicellulose was recovered from the 30%-bark feedstock. The results for individual sugars revealed slightly greater variation. While glucose recovery actually improved slightly with the 30% bark feedstock, the recovery for galactose, mannose and xylose decreased by 6%, 5% and 8%, respectively. Arabinose suffered the greatest decrease in recovery, at about 15%. A s stated earlier, arabinan is known to be very reactive towards acid hydrolysis compared to other hemicellulose polymers, and hydrolyses readily (BeMiller, 1967; L a i , 1991).  Its increased  reactivity can contribute to greater decomposition during pretreatment and low yields likely reflect increased losses due to degradation reactions.  It is worth noting that the chosen pretreatment whitewood-only feedstocks.  conditions were initially defined using  Consequently, adjusting the pretreatment variables slightly might  have resulted in improved sugar recoveries. However, this work sought to evaluate how bark might influence a whitewood-based process, and thus the optimization of bark pretreatment conditions was not conducted.  The data in Table 3-8 reveal an important consideration for the bioconversion process.  While  the recovery of carbohydrate from the starting feedstock can be high, complete recovery of the solubilized sugars can require numerous washes. This is due largely to the physical interactions 70  between the soluble sugars and the cellulose/lignin matrix. It is thus difficult to achieve high sugar yield concurrently with high sugar concentration in the water-soluble fraction. Based on the current results, only - 7 0 % of the soluble sugars present in the slurry following steam explosion could be recovered in the first filtrate (i.e., the water-soluble fraction). This was true for both the 0% and 30% bark feedstocks. Although the washes could have been added to the water-soluble fraction to maximize the total quantity of soluble sugar recovered from the feedstock, the sugar concentration of the resulting liquid stream would be too dilute for cost effective fermentation and distillation, due to the low sugar concentration in the washes (Figure 3-1). Commercially, this issue may be partly resolved through the application of counter-current extraction ( K i m et al, 2001, 2002), but sugar losses w i l l be inevitable i f the goal is to maximize sugar concentration in the water-soluble fraction.  3.2.2.4  C o m p o s i t i o n of the water-soluble fraction  The compositions of the water-soluble fractions derived from whitewood/bark mixtures were evaluated in terms of sugars and several classes of inhibitors. decreased  with increased bark loading (Table 3-9).  The sugar concentrations  The combined monomeric  sugar  concentration decreased by 65%, from approximately 38 g L " in the 0% bark sample (WS-0) to 1  13 g L " in the 100% bark sample (WS-100). Glucose and mannose were most significantly 1  affected, and decreased by - 7 5 % and - 8 5 % , respectively. It is probable that the decrease in sugar concentration was related to the decreased polysaccharide content in the bark feedstock, and not a reduction in pretreatment efficacy. With the exception of galactose, the decreased sugar concentration in each fraction followed a near-linear trend (Figure 3-2), and the concentration could be predicted (within 0.5 g L" ) by a weighted average of the sugar 1  concentration in the W S - 0 and WS-100 fractions. O f the neutral sugars derived from Douglas-  71  15  Wash 1  Wash 2  Wash 3  Wash 4 Wash 5 W a s h 6  W a s h e s at 20% (w/v) Consistency  Figure 3-1: Monomeric sugar concentration (g L" ) in the water washes (performed at 20% w/v consistency) obtained from the water-insoluble components WI-0 and WI-30 after an initial fractionation. Triplicate pulp samples were washed; error bars denote standard deviation. 1  72  Table 3-9:  Monomeric sugar concentration (g L " ) in the water-soluble fractions generated from steam-exploded Douglas-fir whitewood/bark feedstocks under medium-severity conditions. Values in parentheses indicate the range for analysis of samples in duplicate. 1  Arabinose  Galactose  Glucose  Mannose  Xylose  0  1.1 (0.0)  2.7 (0.0)  17.0 (0.0)  13.7 (0.1)  3.4 (0.0)  Total Monomeric Sugar 37.9 (0.1)  10  1.3 (0.0)  2.7 (0.1)  15.5 (0.1)  12.8 (0.3)  3.4 (0.1)  35.7 (0.6)  20  1.5 (0.0)  2.7 (0.0)  14.1 (0.1)  11.7 (0.4)  3.2 (0.1)  33.2 (0.6)  30  1.7 (0.0)  2.5 (0.0)  13.0 (0.1)  10.3 (0.4)  2.9 (0.0)  30.4 (0.5)  100  3.3 (0.0)  1.9 (0.0)  4.3 (0.0)  2.1 (0.0)  1.6 (0.0)  13.2 (0.0)  % Bark  73  Figure 3-2: Monomeric sugar concentration (g L" ) in the water-soluble fractions derived from Douglas-fir whitewood/bark chip mixtures (containing 0, 10, 20, 30 or 100% bark by weight) pretreated under medium-severity conditions. Error bars denote the range of values for duplicate analysis.  74  fir, only arabinose increased in concentration, likely reflecting the greater quantity of arabinan in the bark feedstock when compared to the whitewood, as reported previously (Table 3-6).  3.2.2.5  M o n o m e r i c s u g a r recovery  It has been demonstrated that the hydrolysis of oligomeric sugars to their corresponding monomers is a function of pretreatment severity, i.e., reaction time, temperature and the effective acidity during pretreatment.  Prior to pretreatment of the whitewood/bark feedstocks, it was  hypothesized that increased bark content might mitigate pretreatment severity due to its chemical composition. For example, the higher ash content was expected to result in lower pretreatment acidity due to neutralization of acid (Torget et al,  1991).  Furthermore, the ash in  lignocellulosics has been attributed to decreased rates of acid hydrolysis (Saeman et al, 1954). The increased concentration of lignin was also thought to be detrimental to the pretreatment step and the resulting sugar recovery. In light of these phenomena, it was thought possible that bark might result in an effectively lower severity factor than would be expected for a given time and temperature combination.  Although increased bark content may be responsible for such a  mitigation of pretreatment, the effect appears to be relatively small.  A s with the monomeric  sugars, the concentration of oligomeric sugars actually decreased in the water-soluble fraction when greater quantities of bark were included in the feedstock.  However, the proportion of  oligomeric sugars in the water-soluble fraction was not significantly affected by bark over the experimental range (Table 3-10). The oligomeric sugar content was relatively constant over a tested bark content of 0-30%, suggesting that bark had only minimal influence on pretreatment severity at these concentrations. Oligomeric sugar content was notably lower in WS-100 than in all the other water-soluble fractions, including the whitewood-only and mixed whitewood/bark fractions. This was true for all sugars with the exception of arabinose. The good monomer sugar recovery in WS-100 is in agreement with the observation that bark does not mitigate pretreatment 75  VO  CD  o  U  Tt  cd  =  4—> O  c/i  fe £ ^ cd cd  CD  P  oo  P  H CO n  c  "3 cd -a  "O  2 8 S v3 si 4H T3 fl  cd  cd •6 o £  £  7? co  4—»  P oo  b CD  e co  3 * a oo  -rt.  Q O  O  rt  CN  Tt CN  Tt CN  CO  o  d  Tt d  fl  ON  co  in  5  in  Tt CN  rt  p  Tt CN  CO CO  CO  in  o  CO CN  CO CO CN  00 1—1  Tt Tt  o  o  &S  in  O  _ *fl cd ^3  1  vq CO  CN d  co  CN CN  CN CN  in  T—H  oq  rt  TF  d WO Tt  OO  in  CN  CO rt  in  O  vq  —<  d  Tt d  '—^  d  CD  CD 00  O  X  co  m CN Tt  00  © CN  rt  vd CN  vd CN  vd CN  vd CN  Tt CN  VO  CN d  NO «  co  co  O  >n T t  Id  CN CN  rt CN  ON rt  in rt CN  in 00 rt  O  oo  CO  o  CO  CN  —<  d  CN  d  d  r-  in 00  O 00  ON  rt  00  ON  o  o  o  CN  CO  o o  CD oo  O O  rt  e A  rt  ©  ON  c  vd  rt  d  CO  —  00  CN  CD oo  O  •a 00  CN  00 CD  C  oo  vd  O  S _  ao  !>; o CO <N  CN  C O  CD  CN co  i—i  00  oo  d  in  >  S3 00 fl  rt  CO CN  00  o  <u Id  CD  ON CO  oo  -S ~ -3  O  CO  ^ S3 o  rt  d  severity significantly. F o r reasons that were not clear, the oligomeric sugar contents in watersoluble fractions were higher than previously obtained for Douglas-fir under the same pretreatment conditions (Table 3-4). A s mentioned previously, higher oligomeric sugar content is potentially a problem, as these sugars are not fermentable.  3.2.2.6 Inhibitor recovery In addition to sugars, the water-soluble component also contains lignocellulosic degradation products, which may be inhibitory to fermentation. While inhibition of fermentation has been attributed to a wide range of compounds, for simplicity it is easier to consider only two broad classes of compounds:  naturally-occurring (e.g., resin acids, waxes, sterols and other  extractives) and process-derived inhibitors (e.g., sugar and lignin degradation  products).  Optimization o f the pretreatment conditions can minimize the generation o f process-derived inhibitors, as reported earlier.  However, the nature of the lignocellulosic was expected to  influence the recovery of naturally-occurring inhibitors i n the water-soluble component. Consequently, the inclusion o f bark i n the feedstock was anticipated to produce a water-soluble fraction rich in extraneous and potentially inhibitory material.  The strategy of less severe pretreatment resulted in reduced concentrations of process-derived sugar decomposition products, which could influence fermentation. The concentrations of H M F and furfural, which are known inhibitors of fermentation, were relatively l o w in all of the watersoluble fractions (Table 3-11). Both the H M F and furfural concentrations were greatest in the 0% bark sample (1.65 g L " and 0.34 g L " , respectively). The concentration of H M F decreased 1  1  steadily as bark content increased from 0-100%, while the only significant decrease in furfural concentration was observed when bark content increased to 100%. The decrease in H M F and furfural concentration can be explained in part by the decreased carbohydrate content of the bark 77  J  Table 3-11:  a  b  Inhibitor concentration (g L" ) in the water-soluble fractions prepared from steam-exploded Douglas-fir whitewood/bark feedstocks under mediumseverity pretreatment, in terms of sugar-degradation products, lipophilic compounds, and phenolic compounds. Data represent duplicate analysis of the water-soluble fraction. Values in parentheses indicate the range.  % Bark  HMF  Furfural  Total Lipophilic Compounds  Total Phenolic Compounds  0  1.65 (0.03)  0.34 (0.01)  0.22 (0.06)  3.22 (0.04)  10  1.36 (0.05)  0.33 (0.02)  0.27 (0.04)  3.01 (0.05)  20  1.22 (0.08)  0.28 (0.01)  0.39 (0.07)  2.96 (0.01)  30  1.08 (0.05)  0.26 (0.03)  0.37 (0.05)  2.91 (0.05)  100  0.34 (0.01)  0.07 (0.00)  0.43 (0.06)  3.12(0.06)  3  5  Extracted with M T B E , and corrected for the inclusion of H M F Determined by Folin-Ciocalteu method, standardized with vanillin  78  feedstock.  While the production of H M F and furfural is largely dependent on the severity of  pretreatment, the decomposition of sugars under acidic conditions obeys first-order kinetics (Saeman, 1945), and as such, the concentration of available sugar w i l l influence the final concentration of decomposition products. HMF  Similar observations of decreased concentrations of  and furfural have been reported previously for pretreated  pine and spruce bark  (Taherzadeh et al, 1997a).  In contrast to H M F and furfural, lipophilic compounds increased in the water-soluble fraction with the addition of bark (Table 3-11). Lipophilic extractives can also inhibit fermentation (Tran and Chambers, 1986), and were of particular concern due to the high concentration of these compounds in the bark feedstock.  The concentration of total lipophilic compounds (excluding  the furans H M F and furfural) in the water-soluble fractions generally increased with increased bark content, but only to a maximum of 0.43 g L . 1  This was considerably lower than  anticipated, considering that the total extractives content of the bark was approximately 30% of its dry weight (Table 3-6).  It is likely that these compounds simply do not partition to the  aqueous fraction due to limited solubility in water and hydrophobic interactions with the remaining insoluble cellulose and lignin. Lomax et al, (1994) have reported that less than two percent of the extractives reach the water-soluble fraction following steam explosion of Pinus radiata bark.  While some extractives may be volatilized and lost during the steam explosion  process, the high shot yield previous reported (Table 3-7) confirms that most extractives remain in the water-insoluble fraction.  The quantity of extractives recovered from the 100%-bark  feedstock decreased following steam explosion (Figure 3-3). recovered  to each  solvent fraction also changed,  The quantities of extractives  suggesting  chemical changes  during  pretreatment that affected their solubility. These findings are consistent with previous research indicating that the extractive compounds may undergo condensation reactions, either as 79  Total  Hexane  DCM  EtOAc Acetone M e O H  Water  Figure 3-3: Extractive composition of the Douglas-fir bark feedstock (100% bark) before and after steam explosion under medium-severity conditions. Samples were Soxhletextracted sequentially with hexane, dichloromethane ( D C M ) , ethyl acetate (EtOAc), acetone, methanol (MeOH) and water. Error bars indicate the standard deviation (n = 3).  80  consequence of the heat of pretreatment (Lomax et al, 1994), or v i a auto-oxidation processes (Field and Lettinga, 1991), rendering them less soluble. The low concentration of lipophilic extractives recovered in each of the water-soluble components was not expected to cause significant inhibition of fermentation.  Lignin  degradation  products  were anticipated to increase in concentration with  greater  percentage of bark in the feedstock, due to the increased quantities o f lignin. The concentration of total phenolics in each of the water-soluble fractions was estimated using Folin-Ciocalteu reagent, and calibrated using vanillin (4-hydroxy-3-methoxy-benzaldehyde) (Table 3-11). Surprisingly, no significant difference was observed for these water-soluble fractions.  It was  initially thought that the increased bark content in the feedstock would translate to increased phenolic content in the water-soluble fraction.  It is important to consider that the phenolic  compounds present in each of the water-soluble fractions was likely unique. Certain phenolic compounds derived from the bark (e.g., tannins) may simply partition more favourably than others to the water-soluble fraction.  A s the concentration of these compounds would be  expected to increase in the 100% bark feedstock, it should be expected that the concentration of phenolics recovered to the  water-soluble fraction also increase.  These differences  in  concentration could be anticipated to translate into differences in fermentability of the watersoluble fractions.  3.3  Conclusions  It was apparent that both pretreatment severity and feedstock composition (i.e., bark content) had an influence on the composition of the water-soluble component.  Tailoring  pretreatment  conditions towards lower severity was clearly advantageous for the pretreatment o f Douglas-fir wood chips, when the goal is high-yield recovery of the hemicellulose sugars in a fermentable 81  form.  In a d d i t i o n to p e r m i t t i n g the i m p r o v e d r e c o v e r y o f the feedstock (i.e., shot y i e l d ) ,  pretreatment under l o w - s e v e r i t y c o n d i t i o n s also p r o v i d e d for the increased s u r v i v a l and r e c o v e r y o f the h e m i c e l l u l o s e sugars f r o m the starting feedstock.  Consequently, a significantly lower  concentration o f the sugar d e c o m p o s i t i o n products, H M F and furfural, were obtained i n the resulting w a t e r - s o l u b l e c o m p o n e n t .  In contrast, w h e n m o r e severe c o n d i t i o n s were e m p l o y e d ,  more extensive m a t e r i a l losses were observed, resulting i n the p r o d u c t i o n o f increased quantities o f d e c o m p o s i t i o n products and notable fermentation i n h i b i t o r s . H o w e v e r , it was thought l i k e l y that the approach o f m i l d pretreatment m i g h t have consequences for the o v e r a l l b i o c o n v e r s i o n o f Douglas-fir  residues.  Although  high  recovery  of  the  hemicellulose  component  was  demonstrated, there was c o n c e r n that the resulting c e l l u l o s e - r i c h w a t e r - i n s o l u b l e c o m p o n e n t w o u l d not be as r e a d i l y h y d r o l y s a b l e , r e s u l t i n g i n p o o r o v e r a l l carbohydrate r e c o v e r y f r o m the feedstock.  T h e i n f l u e n c e o f b a r k o n h e m i c e l l u l o s e r e c o v e r y was m u c h less p r o n o u n c e d than pretreatment severity.  D u e to its l o w e r sugar content, the i n c l u s i o n o f bark i n the feedstock resulted i n  decreased sugar c o n c e n t r a t i o n i n the water-soluble fraction. H e m i c e l l u l o s e sugar r e c o v e r y f r o m the feedstock was also r e d u c e d s l i g h t l y . unaffected.  D e s p i t e these effects, m o n o m e r sugar recovery was  F u r t h e r m o r e , the concentration o f sugar and l i g n i n degradation products recovered  i n the water-soluble fractions decreased w i t h increased bark content up to 3 0 % , w h i l e the concentration o f l i p o p h i l i c c o m p o u n d s r e m a i n e d l o w i n a l l l i q u i d fractions d e r i v e d f r o m barkc o n t a i n i n g feedstocks.  O n e o f the p r i n c i p l e goals o f this project was to demonstrate h i g h - y i e l d ethanol p r o d u c t i o n f r o m the D o u g l a s - f i r w a t e r - s o l u b l e fractions, and experiments addressing fermentation are reported i n the f o l l o w i n g chapter.  G i v e n the significant v a r i a t i o n i n terms o f sugar and p o s s i b l e i n h i b i t o r 82  concentrations in each of the water-soluble fractions, their fermentability was expected to differ greatly. The higher concentrations of H M F and furfural realized under more severe pretreatment were anticipated to negatively impact the fermentation process. In contrast, the decreased furan and phenolic content and low concentration of lipophilic compounds suggested that bark might not have as detrimental impact on the fermentation process as was originally anticipated. However, in light of the current results it was clear that the fermentation to ethanol would be affected by the dilute sugar concentrations in the water-soluble fractions.  Although it was  possible to define pretreatment conditions that permitted high recovery of the hemicellulose sugars, the final sugar concentration in the water-soluble component was very dilute. In fact, sugar concentration was quite dilute in all of the water-soluble fractions, and only reached a maximum of 49 g L" in the bark-free, high-severity fraction. This dilute sugar concentration 1  would be particularly problematic for an industrial process, and strategies for trying to resolve this issue are described later in the thesis.  83  Chapter 4: Fermentation of the Water-soluble Fraction to Ethanol 4.1 B a c k g r o u n d As mentioned earlier, if the goal of bioconversion is to maximize ethanol production from a given feedstock, then it is essential that the hemicellulose-rich water-soluble fraction recovered following pretreatment be readily fermentable. It was shown previously that high pretreatment severity designed to provide a substrate susceptible to enzymatic hydrolysis results in a significant loss of the available carbohydrate  to decomposition products,  and partial  disintegration of the lignin matrix. These components are largely recovered in the water-soluble fraction where they have been shown to be inhibitory to the fermentation stage. The results presented in the previous chapter show that tailoring pretreatment conditions for hemicellulose recovery can result in improved sugar recovery, in addition to decreased production of these "process-derived", potential inhibitors. However, in addition to the pretreatment severity, the composition of the feedstock can influence the composition of the water-soluble fraction. The inclusion of bark, for example, resulted in significant changes in water-soluble fraction composition following pretreatment.  It appeared that the main consequence of bark was  decreased sugar concentration in the water-soluble fraction, but the concentration of naturallyoccurring, potentially inhibitory material such as extractives also increased.  The work described in this chapter of the thesis reports on the fermentation of the sugars present in the water-soluble fractions derived from steam-exploded Douglas-fir feedstocks, and on the possible influence that pretreatment severity and lignocellulosic composition (i.e., bark content) has on the relative ease of fermentation by an industrial strain of yeast, Saccharomyces cerevisiae.  84  4.2 Results and Discussion 4.2.1  Yeast strain selection  Numerous microorganisms ranging from naturally-occurring to genetically modified yeast and bacteria have been employed for the fermentation of lignocellulose-derived sugars to ethanol. These include, but are not limited to, Saccharomyces shehatae, Escherichia  coli, and Zymomonas mobilis.  cerevisiae,  Pichia  pastoris,  Candida  Non-genetically modified S. cerevisiae is  perhaps the most industrially-relevant ethanologenic organism, capable of high ethanol yields from a variety of hexose sugars, and sufficiently robust to tolerate the presence of potentially inhibitory compounds in lignocellulosic hydrolysates. For the fermentation of the Douglas-fir water-soluble fractions described in this work, two industrial strains of S. cerevisiae were employed. Both strains of yeast were supplied by Tembec Ltd., (Temiscameng, QC, CAN), a sulphite pulp mill that ferments its spent-sulphite liquor (SSL) to ethanol. The first of the strains was a wild-type brewer's yeast, designated "WT". The other strain, designated " T l " was an acclimated strain of S. cerevisiae isolated from spent-sulphite liquor fermentations. This strain has been shown previously by Tembec to ferment its SSL stream with improved productivity compared to the wild-type yeast. The improved fermentation of the SSL stream by strain T l was thought to be the consequence of improved resistance towards inhibitors present in the pulping liquor. It was hoped that the noted improvement during fermentation of SSL would carry over to the fermentation of the steam exploded Douglas-fir water-soluble fractions. However, due to significant differences in the composition of the SSL stream and the steam-exploded Douglas-fir water-soluble fraction (most notably, sulphite concentration), it was possible that there might be no realized advantage.  85  4.2.1.1  H e x o s e utilization by the T1 a n d WT yeast strains  The two strains of yeast were initially compared in parallel fermentation experiments, in order to evaluate the relative abilities of each yeast to ferment sugars to ethanol. This was performed using the water-soluble fraction derived from Douglas-fir whitewood under medium-severity pretreatment.  This fraction was selected in order to evaluate the yeast's performance in the  presence of moderate concentrations of inhibitory compounds. In addition to the Douglas-fir wood hydrolysate, synthetic medium (1% yeast extract, 1% peptone) was formulated to contain the same individual sugar concentration as found in the water-soluble fraction. The results of the fermentations are compared in Figure 4-1.  Both of the yeast strains used in this work are only capable of fermenting hexose sugars and not pentose sugars, thus the data in Figure 4-1 only report the consumption of (A) glucose, (B) mannose, (C) galactose, and (D) combined hexose sugars. The inability to consume pentose sugars by these particular yeast strains does represent a minor concern for the bioconversion of softwood feedstocks. However, as was discussed earlier, it is likely that many of the engineered microorganisms that have been shown able to co-ferment pentoses and hexoses would be able to utilize these pentoses, provided that the hexose sugars in the water-soluble fraction can be utilized.  Regardless of the fermentation medium used, both strains of yeast co-metabolized glucose and mannose, with glucose metabolized at a slightly faster rate than mannose (Figure 4-1, A and B). The metabolism of galactose by most strains of S. cerevisiae is normally stringently controlled (Carlson, 1987; Gancedo, 1992; Ostergaard et al, 2000), and with both of the current strains, galactose utilization was repressed by the presence of both glucose and mannose until the concentrations of these sugars were exhausted at approximately 6 h into the fermentation (Figure 86  0  12  24  36  48  0  12  24  Time in Hours  Time in Hours  Time in Hours  Time in Hours  36  48  Figure 4-1: Comparison of (A) glucose uptake, (B) mannose uptake, (C) galactose uptake and (D) combined hexose uptake by the Tembec SSL-adapted Saccharomyces cerevisiae strain T l and a wild-type S. cerevisiae strain, during fermentation of the water-soluble fraction derived from Douglas-fir whitewood prepared under medium-severity pretreatment, and synthetic medium (1% yeast extract, 1% peptone containing comparable sugar concentrations to the water-soluble fraction). Fermentations were run in duplicate; error bars indicate range.  87  4-1, C). However, even after glucose and mannose were fully utilized and galactose metabolism was finally initiated, the rate of consumption of the galactose was significantly slower than either glucose or mannose. It was also apparent that the fermentation rates for all sugars were reduced in the water-soluble fraction compared to synthetic medium. A part of this decrease may be related to nutrient limitation, as the synthetic medium contained yeast extract and peptone (1% w/v each), whereas only a small ammonium supplement (ammonium phosphate dibasic, 20 m M final concentration) was added to the water-soluble fraction. The decision not to use additional nutrient supplements during the fermentation of the water-soluble component was based on the recognized expense associated with these supplements for a potential commercial process. However, the observed decrease is more likely related to the presence of inhibitory compounds in the water-soluble fraction. It has long been recognized that that compounds that are inhibitory to fermentation are produced or liberated during the pretreatment step (Leonard and Hajny, 1945). These compounds arise due to the decomposition of carbohydrate (Banerjee et al, 1981a; Sanchez and Bautista, 1988; Taherzadeh et al, 2000) and lignin (Clark and Mackie, 1984; Larsson et al, 2000), but also due to the liberation of organic acids and extractive compounds present in the lignocellulosic feedstock (Larsson et al, 1999a; Taherzadeh et al, 1997b). This will be described later in more detail. The initial rates of sugar consumption were assessed (Table 4-1).  Due to the rapid consumption of sugars during fermentation of the synthetic  medium, glucose was fully consumed before the 4-hour sample was taken. Thus, the rate was calculated over the first 2 h of fermentation only. For comparison purposes, the initial rate for mannose and total hexose sugars consumption was also calculated over this same 2-hour period. However, it should be apparent from the graphs that the rate of mannose consumption in synthetic medium improved slightly after the 2-hour sample. This was also true for fermentation of the water-soluble fraction. As previously mentioned, the consumption of galactose sugar did not commence until after approximately 6 h. Thus, the rate was calculated between the 6-hour 88  Table 4-1: Sugar uptake rate by the wild-type (WT) or Tembec SSL-adapted (Tl) strains of Saccharomyces cerevisiae, during fermentation of either the Douglas-fir water-soluble fraction derived from whitewood pretreated under mediumseverity, or synthetic medium containing comparable sugar concentrations. Duplicate samples were fermented; values in parentheses indicate the range.  Initial sugar consumption rate (g L" h" ) 3  Fermented Sugar  Yeast Strain  Combined Hexose  WT Tl  0.3 (0.1) 2.1 (0.3)  Synthetic Medium 8.1 (0.3) 7.4 (0.3)  Glucose  WT Tl  0.2 (0.1) 1.0 (0.1)  6.2 (0.1) 4.6 (0.2)  Mannose  WT Tl  0.1 (0.1) 1.0 (0.1)  2.1 (0.1) 2.9 (0.1)  Water-Soluble Fraction  WT 0.2 (0.0) Tl 0.1 (0.0) Initial rate calculated after 2 h fermentation, except for galactose, which was calculated over the 6- to 21-hour sample. Galactose  89  sample and the next sample point, 21-hours.  This resulted in a rate higher than would be  otherwise calculated if the period during glucose and mannose consumption was also included, but better reflects the yeasts' ability to metabolize galactose once consumption of the sugar had begun.  During fermentation of the synthetic medium, neither yeast showed a distinct advantage over the other. Glucose, for example, was consumed rapidly by both yeasts and was fully utilized within 4 h (Figure 4-1, A), although the rate of consumption was slightly faster for the WT strain. After 4 h, there were still traces of mannose in the synthetic medium, but this was completely consumed within 6 h (Figure 4-1, B). In contrast, no decrease in the measured galactose concentration was observed until hour-6, and metabolism was very slow for both strains (Figure 4-1, C). The wild-type yeast appeared slightly better at consuming galactose, having depleted the galactose within 24 h, whereas approximately half of the galactose remained at this timepoint during fermentation with the adapted strain. In contrast, when the two strains of yeast were used for fermentation of the hemicellulose-rich, water-soluble fraction, it was apparent that sugar consumption by the adapted strain (Tl) was significantly improved compared to the wild-type strain. Both glucose and mannose were depleted from the water-soluble fraction by strain T l at a much faster rate than the wild-type strain (Table 4-1). However, galactose in the water-soluble fraction was poorly consumed by both strains. The presence of residual galactose after 48 h was atypical of the fermentations using the water-soluble fraction, and an explanation for why this occurred is not clear.  The slow utilization of galactose may be due to a weaker induction  response to the presence of galactose. Had samples been taken at 72 h, it is possible that a greater amount of galactose would have been consumed.  However, the apparently "stuck"  fermentation may be a consequence of nutrient deficiency, particularly assimilable nitrogen (Bataillon et al,  1996; Salmon, 1989).  Past work by these researchers has shown that 90  insufficient nitrogen can result in arrested protein production and, consequently sugar metabolism can cease. Although the water-soiuble fraction was supplemented with ammonium phosphate, depending on the quantity assimilated and stored during the culturing of either inoculum, this small supplement may not have been sufficient.  Alternatively, galactose  metabolism may be more susceptible to inhibition by compounds present in the water-soluble fraction.  The latter explanation is unlikely, as, in the vast majority of the fermentations  performed using both low- and medium-severity water-soluble fractions, the galactose content was completely exhausted within 48 h, and often within 24 h.  4.2.1.2 Ethanol production Ethanol production after 48 h growth in the water-soluble fraction was comparable for both yeast strains. A calculation of the ethanol yield (based on the available hexose sugars, and expressed as a percentage of theoretical yield) revealed slightly lower yields for the water-soluble fraction (88.5% and 87.0% for T l and WT, respectively) when compared to the synthetic medium (93.5% and 93.0%). The decrease in ethanol production in the water-soluble fraction can be attributed mainly to poor galactose utilization. However, the apparent inhibition observed during fermentation of the water-soluble fraction likely also played a role in the decreased ethanol yield. Various compounds present in the water-soluble fractions have been shown to cause a decrease in the rate of fermentation (e.g., H M F , furfural), while other compounds including organic acids have been attributed to a decrease in ethanol production, and thus ethanol yield (Larsson et al, 1999a, 1999b). Inhibition of fermentation is addressed in more detail in subsequent sections. However, based on the improved rate of consumption of glucose and mannose, it was apparent that the wild-type S. cerevisiae strain was not as robust as the Tembec adapted yeast when fermenting the water-soluble fraction. Thus, the Tembec yeast was used for all subsequent fermentations. 91  4.2.2 Pretreatment severity and fermentation As has been mentioned briefly already, several types of inhibitors arise from the decomposition of carbohydrate and lignin during pretreatment of lignocellulosics, including aldehydes (HMF, furfural, 2-(2-hydroxyacetyl)-furan, 5-methyl-2-furaldehyde, formaldehyde, etc.), monomeric phenolics (Hibbert's ketones, vanillin, coniferyl aldehyde, cinnamic acid, etc.), acids (levulinic, formic, etc.), and various condensation products (Ando et al, 1986; Clark and Mackie, 1984; Lai, 1991; Larsson et al, 1999a, 2000; Nilvebrant et al, 1997; Nishikawa et al, 1987, 1988). The process-derived inhibitors are as numerous as they are diverse, and are recognized for their potential impact on the fermentation of the water-soluble fraction. Despite exhaustive research by other research groups on the topic, a complete understanding of the complex, synergistic interaction between inhibitors and the microbial catalyst remains elusive.  Although there  remains a significant amount of research to be conducted in this area, the objective of this part of the project was not to delve into identification and characterization of the various inhibitors. Rather, it was hoped that by focusing on the pretreatment severity, it would be possible to minimize production of these inhibitors, and effectively ferment the various substrates.  The fermentability of the water-soluble fraction recovered from steam exploded Douglas-fir whitewood under low-, medium- and high-severity pretreatment was assessed, using the Tembec SSL-adapted S. cerevisiae (Figure 4-2, A-C). It was apparent that the pretreatment severity used to treat the wood greatly influenced the utilization of sugars present in the water-soluble fraction. This was most likely related to the production of inhibitors during pretreatment, as it is recognized that acid-catalyzed decomposition of lignocellulosics increases with increased temperature.  After high-severity pretreatment, only minimal amounts of the sugar (~5 g L ) 1  present in the water-soluble fraction were consumed by the yeast over the 48-hour fermentation. 92  Figure 4-2: Hexose sugar consumption and ethanol production during fermentation of the Douglas-fir whitewood water-soluble fractions obtained after pretreatment at various severity conditions. (A) Low-severity; (B) Medium-severity; (C) Highseverity; (D) Ethanol production. Fermentations were preformed in duplicate and error bars denote the range of values.  93  In contrast, when milder pretreatment conditions were employed, fermentation was possible, and hexose sugars were completely consumed by 24 h. As reported previously (Table 3-5), H M F and furfural concentrations in the water-soluble hydrolysate both increased as the pretreatment severity increased. It has been shown that, at a certain threshold concentration, these compounds can completely inhibit yeast metabolism (Sanchez and Bautista, 1988), and can result in a great reduction in the proportion of viable cells in a culture (Chung and Lee, 1985). The specific concentration required for complete inhibition will depend on a number of factors, including the microorganism and the inoculum concentration. However, the assayed concentrations of furans present in the high-severity water-soluble fraction (3.6 g L" and 4.1 g L" for H M F and furfural, 1  1  respectively) were anticipated to result in considerable cell death, based on previous reports. At lower concentrations, cells may remain viable, but due to partial inhibition of enzymes involved in glycolysis (Banerjee et al, 1981b) there may be a prolonged metabolic "lag" at the start of fermentation, or alternatively, a decreased rate of carbohydrate metabolism (Banerjee et al, 1981a, 1981b; Chung and Lee, 1985; Palmqvist, 1998; Taherzadeh et al, 2000). Interestingly, despite notably higher concentrations of furans in the medium-severity water-soluble fraction, sugar consumption was actually slightly faster than in the low-severity fraction (1.2 vs. 1.0 g L"  1  h" , based on consumption of hexose sugars, and calculated over the first 6 h of fermentation). 1  However, due to the low sampling frequency, it is difficult to state with certainty the effect that pretreatment severity had on the rate of sugar consumption.  In addition to sugar consumption, ethanol production was also affected by the selected pretreatment severity (Figure 4-2, D). Ethanol production in all of the Douglas-fir wood-derived water-soluble fractions was considerably slower than in the glucose control ( G M Y P medium). Increased concentrations of inhibitors in the wood hydrolysates, such as H M F and furfural, could account for this decreased productivity. These compounds generally correlate well with reduced 94  ethanol productivity (increased "lag" and slow metabolism) (Banerjee et al, 1981a, 1981b; Sanchez and Bautista, 1988), but do not always cause a reduction in yield. Taherzadeh et al, (1997a) further suggest that the maximum ethanol production rate can be effectively predicted from the sum of H M F and furfural concentration in the water-soluble fraction, although by no means are these compounds the only important inhibitors present.  As the pretreatment  conditions increased from low- to high-severity, the rate of ethanol production decreased from approximately 0.9 to 0.4 to 0 g L" h" (calculated over the first 6 h), while ethanol yields were 1  1  relatively high in both the low and medium-severity hydrolysates (Table 4-2). These results are consistent with inhibition by furans, but based on other researchers' work, it was expected that additional inhibitory compounds were also present in each of the water-soluble fractions. The relatively good fermentation of both the low and medium-severity fraction suggested that the concentrations of these inhibitors were sufficiently low so as to not completely inhibit the yeast. Even with a slower rate of ethanol production in the medium-severity water-soluble fraction, ethanol concentration exceeded that obtained when using the low-severity fraction (Table 4-2).  In contrast to ethanol yields, glycerol yield did not represent a significant proportion of the available substrate. Glycerol, which is usually produced by the yeast to combat the detrimental effects of high osmolarity in the medium, or to counter the redox imbalance caused by the depletion of cellular NAD+ levels (Taherzadeh et al, 2002), attained a maximum yield of only 0.04 g g-hexose" in the low-severity water-soluble fraction. The fact that the yield of glycerol 1  remained low was clearly beneficial, as it meant less carbon was sequestered for cell maintenance and propagation, and consequently, greater ethanol production was obtained.  It was apparent that high-severity pretreatment conditions not only resulted in very poor recovery of sugars from the feedstock (Table 3-3), but also a water-soluble stream that was not 95  Table 4-2: Ethanol and glycerol production during fermentation of the three water-soluble fractions obtained after pretreatment under low-, medium-, and high-severity conditions. Values in parentheses denote the range for duplicate fermentations. Product Concentrations (g L" ) 1  Product Yields (g g-hexose ) 1  Severity Factor  Ethanol  Glycerol  Ethanol  Glycerol  Low  12.6(0.1)  1.1 (0.0)  0.44 (0.01)  0.04 (0.00)  Medium  14.9 (0.0)  0.8(0.1)  0.44 (0.00)  0.02 (0.00)  High  0  0  0  0  96  readily fermentable.  Detoxification of the water-soluble fraction derived under high-severity  pretreatment would likely improve its fermentability, and numerous distinct methods have been reported for the removal of different inhibitors (Jonsson et al., 1998; Larsson et al, 1999b). Economically, it is debatable whether such detoxification procedures could be accommodated in a putative process. However, the US Department of Energy (DOE) Biofuels Program, which is currently investigating the core steps required for the conversion of lignocellulosic biomass (mainly agricultural residues such as corn stover) to ethanol and other chemicals, has incorporated overliming with calcium hydroxide in its current Sugar-Ethanol Platform, to condition the hemicellulose hydrolysate prior to fermentation (Aden, 2002).  In contrast, no  method of detoxification to improve fermentation was explored in the current work. Certainly the economics of a softwood-to-ethanol process differ from those under investigation by the DOE. In addition to this reservation, it would also be difficult to justify the additional cost of detoxification when a considerable fraction of the original hemicellulose (57%) would be lost during pretreatment at high-severity (Table 3-3). In contrast, based on the work presented here, the use of low- and medium-severity pretreatment conditions would obviate the need for a detoxification protocol, while providing improved hemicellulose sugar recovery and allowing better ethanol recovery.  4.2.3 Bark content a n d fermentation Much as pretreatment severity was anticipated to impact the fermentability of the resulting hemicellulose-rich, water-soluble fractions, so too was feedstock composition. It was initially expected that the inclusion of bark would result in significant inhibition of fermentation. Although one of the major physiological roles of bark is as a physical barrier to microbial and insect attack, the extraneous materials in bark (e.g., resin/fatty acids, waxes, terpenes, sterols, etc) are recognized for their "bioactive" nature (Haygreen and Bowyer, 1996; Tran and 97  Chambers, 1986). These extractable materials are especially concentrated in the bark, and have been implicated in reduced fermentation performance.  In addition, debarking wastewaters have  been shown to be inhibitory to methanogenic bacteria, due to their high tannin content and the capacity for these tannins to bind intra- and extracellular enzymes (Field et al, 1989; Field and Lettinga, 1991), thus limiting their effectiveness.  The inclusion of bark in the feedstock was  expected to result in differences in the chemical composition of the water-soluble fraction, and consequently, differences in fermentation. As was reported previously, the inclusion of bark in the feedstock did have an effect on the recovery of both process-derived and naturally-occurring inhibitors (Table 3-11). However, of the compounds analysed, only the lipophilic extractives increased in concentration. Other compounds such as H M F and furfural decreased considerably in the water-soluble fraction as bark content increased, as did total phenolic compounds (up to a bark content of 30%) despite high quantities of available phenolic compounds in the bark feedstock. The concentration of lipophilic compounds doubled for an increase in bark content from 0-100%, but the actual concentration in any of the water-soluble fractions was relatively minor (0.22 to 0.43 g L" ). This suggested that the fermentation of the water-soluble fractions 1  derived from feedstocks containing bark might not be as problematic as initially thought. However, the particular composition of the bark-derived extractives and phenolics recovered in the water-soluble fraction was expected to differ from that in the bark-free, water-soluble fraction.  Thus, the low concentration might still result in inhibition of fermentation.  Furthermore, because only a limited number of compounds were analysed, other compounds that were present in the water-soluble fraction derived from bark-containing feedstocks might result in decreased fermentation performance.  98  4.2.3.1 Influence of bark on hexose sugar consumption The fermentability of the water-soluble fractions derived from wood and bark was evaluated using the Tembec SSL-adapted yeast. It was apparent that increased bark content did not have a significant impact on sugar consumption (Figure 4-3). Consumption of glucose, mannose, and galactose was possible in all of the water-soluble fractions, and with the exception of the whitewood-only water-soluble fraction (WS-0), was virtually complete within 24 h. A l l of the water-soluble fractions were readily fermentable, with all of the available hexose sugars consumed within 48 h, regardless of the bark content. As mentioned before, the Tembec yeast cannot ferment pentose sugars, thus xylose and arabinose were not consumed in any of the fermentations (results not shown). On the basis of total hexose sugars (Figure 4-4, A) the initial rates of consumption appeared comparable for all fermentations. A closer inspection confirmed that the initial rates were nearly identical for all of the water-soluble fractions, with the exception of WS-100 (Table 4-3). It was initially hypothesized that the low rate of sugar consumption for the WS-100 fraction (approximately half that of the other water-soluble fractions) might be a consequence of the increased extractives and phenolic compounds reported previously (Table 311), or possibly a bark-derived compound that was not quantified. In fact, this was shown not to be the case.  The calculated rate was most likely lower because of the low starting sugar  concentration. Both glucose and mannose were already consumed from WS-100 by the 2-hour sample, and the characteristically slow fermentation of galactose had begun (Figure 4-3, E). This effectively reduced the overall rate of hexose consumption compared to the other watersoluble fractions (where glucose and mannose were still present after 2 h), and gave the appearance of decreased fermentation rate. That the 100% bark water-soluble fraction was not inhibitory was confirmed experimentally by supplementing this fraction with galactose, glucose and mannose to achieve the same starting sugar concentration as the 0% bark water-soluble fraction (Figure 4-5, A). The initial rate of sugar consumption in sample WS-100S was slightly 99  16-  16  Glucose -A Mannose - • — Galactose  14  B  Glucose Mannose - Galactose  12  -1—  24  24  Time in Hours  Time in Hours  36  16-  D  - Glucose Mannose - Galactose  S  Glucose Mannose - Galactose  10  o O  24  36  Time in Hours  48  24  36  48  Time in Hours  3.0  E  • its.  2.5  — •  Glucose Mannose - Galactose  12  Time in Hours  Figure 4-3: Sugar consumption during fermentation of the water-soluble fractions derived from Douglas-fir whitewood feedstocks containing 0-100% supplemental bark. (A) WS0; (B) WS-10; (C) WS-20; (D) WS-30; (E) WS-100. Note: due to the very low sugar concentration in WS-100, a different scale is used (Figure E). Error bars denote the range for duplicate fermentations.  100  Figure 4-4: Sugar consumption and ethanol production during fermentation of the watersoluble fractions derived from Douglas-fir whitewood/bark feedstocks pretreated under medium-severity conditions. (A) Combined hexose sugar consumption; (B) Ethanol production. Legend applies to both figures. Error bars indicate the range of values for duplicate fermentations. 101  Table 4-3:  Fermentation of the water-soluble fractions derived from Douglas-fir whitewood/bark feedstocks (containing 0, 10, 20, 30, or 100% bark by weight) pretreated under medium-severity conditions. Values in parentheses denote the range of duplicate fermentations. Hexose concentration (g L" ) t=0h  Volumetric Hexose Uptake  1  Kaie  a  (gL  11.8(0.2)  Total Hexose 28.6 (0.5)  12.4 (0.3)  10.6 (0.2)  25.4 (0.5)  4.0 (0.3)  2.3 (0.0)  12.0 (0.2)  10.0 (0.2)  24.3 (0.4)  4.0 (0.3)  WS-30  2.3 (0.0)  8.9 (0.1)  21.9 (0.2)  4.0 (0.1)  WS-100  1.6 (0.0)  10.7 (0.1) 2.3 (0.1)  5.2 (0.1)  1.8 (0.1)  Reference  0  29.6 (0.2)  1.3 (0.0) 0  29.6 (0.2)  7.0 (0.1)  Feedstock  Galactose  Glucose  Mannose  WS-0  2.4 (0.1)  14.4 (0.2)  WS-10  2.4 (0.0)  WS-20  1  h' ) 1  3.5 (0.4)  Calculated over the first 2 hours of fermentation  102  T i m e in H o u r s  Figure 4-5: Fermentation of the water-soluble fractions WS-0, WS-100 and WS-100S (WS100 supplemented with hexose sugars to achieve a sugar concentration comparable to WS-0). A reference fermentation (medium containing glucose as the sole fermentable sugar) was also used. (A) Hexose sugar consumption; (B) Ethanol production. Error bars indicate the range of values for duplicate fermentations. 103  faster than the glucose control (7.9 ± 0.3 vs. 7.7 ± 0.1 g L" h" , respectively) and noticeably 1  1  faster than WS-0 (3.7 ± 0.4 g L" h" ). The initial rate of ethanol production was also comparable 1  1  in WS-100S and the glucose control (3.7 ± 0.0 and 3.9 ± 0.2, respectively), and approximately double that obtained in WS-0 (1.8 ± 0.1 g L" h" ) (Figure 4-5, B). Lower concentrations of H M F 1  1  and furfural, combined with relatively low concentrations of lipophilic extractives and phenolics may account for the low toxicity and rapid metabolism of the supplemented water-soluble fraction.  4.2.3.2  Influence of bark o n ethanol production  Increased quantities of bark in the feedstock resulted in a significant decrease in ethanol concentration, although ethanol yield was largely unaffected (Figure 4-3, B, and Table 4-4). While the decrease in ethanol concentration can be attributed to the lower sugar concentration in the water-soluble fractions as bark content increased, the high yield is indicative of the relatively low toxicity of the water-soluble fractions.  Ethanol yield was lower than the theoretical  maximum for all fermentation including the reference fermentations ( G M Y P medium), ranging from 0.43 g g" to 0.47 g g" (Table 4-4). However, there was no apparent trend. The relative 1  1  ease of fermentation of these fractions, particularly the WS-100 fraction, is consistent with work performed by Taherzadeh et al, (1997a), who previously reported rapid fermentation of a supplemented spruce bark hydrolysate to a high ethanol yield (0.43 g g" ). However, the current 1  results appear to contradict the findings of previous research using a mixed softwood feedstock (white fir and ponderosa pine) containing bark (Boussaid et al, 2001). A 26% reduction in ethanol yield was reported when bark (-9% w/w) was included in the feedstock compared to a whitewood-only feedstock. The discrepancy between the results for Douglas-fir and this result was thought to be possibly related to the severity of the pretreatment conditions. The White  104  Table 4-4: Ethanol and glycerol production during fermentation of the water-soluble fractions derived from Douglas-fir whitewood/bark feedstocks (containing 0, 10, 20, 30, or 100% bark by weight) pretreated under medium-severity conditions. Values in parentheses denote the range of duplicate fermentations. Product Concentration (g L" ) 3  a  b  1  Product Yield (g g-hexose" ) b  1  Feedstock  Ethanol  Glycerol  Ethanol  Glycerol  WS-0  12.6 (0.1)  2.8 (0.0)  0.44 (0.01)  0.10(0.00)  WS-10  11.7 (0.0)  2.5 (0.0)  0.46 (0.01)  0.10(0.00)  WS-20  10.5 (0.0)  2.6 (0.0)  0.43 (0.01)  0.11 (0.00)  WS-30 WS-100 Reference  9.5 (0.1) 2.4 (0.0) 13.9(0.1)  2.3 (0.0) 1.2 (0.0) 2.7 (0.0)  0.43 (0.01) 0.46 (0.01) 0.47 (0.01)  0.11 (0.00) 0.22 (0.00) 0.09 (0.00)  Maximum concentration attained during fermentation Yield calculation based on maximum product concentration  105  fir/Ponderosa pine sample was pretreated at a lower pretreatment severity (log R = 3.17, 0  corresponding to a temperature of 195°C, time of 2.38 minutes, and S 0 concentration of 3.91%) 2  than was used in the current work (log R = 3.45). When the white-fir/ponderosa pine feedstock 0  was pretreated under higher-severity (log R = 3.76, 215°C, 2.38 minutes, 2.38% S0 ), the 0  2  fermentability did improve significantly. The reported improvement may have been related to increased condensation of possible inhibitors such as tannins and other extractives from the bark, due to the higher processing temperatures, as has been shown to occur during pretreatment of bark from Radiata pine (Lomax et al,  1994).  A clear relationship between extent of  condensation and reactivity of tannins with cellular proteins has been demonstrated with methanogenic bacteria and debarking wastewaters (Field et al, 1989). The relative toxicity of these phenolic compounds is reported to decrease as the degree of polymerization (i.e., molecular weight) increases, due to reduced penetration into the cell, and decreased capacity for hydrogen bonding with the cellular protein/enzymes.  However, Boussaid et al, (2001) still  reported a 10% reduction in ethanol yield (-85% vs. -95% in the whitewood only feedstock, based on theoretical ethanol yield) under the more severe pretreatment conditions. This result suggests that the impact of bark on fermentation may be partly species-dependent.  Bark from  different species is known to vary significantly in composition, in terms of both quantity and type of chemical constituents (Chang and Mitchel, 1955; Jensen et al, 1963). It is likely that the chemical composition of the water-soluble fraction derived from two different bark feedstocks will also be different, which could result in a reduced or augmented degree of inhibition. In addition, bark from different tree species might respond differently to identical pretreatment conditions, much as has been demonstrated previously for whitewood.  106  4.2.4 Pretreatment severity and fermentability of bark-derived WS fractions The possibility that the chosen pretreatment severity might partly explain the good fermentability of the wood/bark-derived water-soluble fractions was pursued further.  If, for example, the  chosen pretreatment severity resulted in improved fermentation due to condensation or degradation of potential bark-derived inhibitors, then pretreatment of the Douglas-fir feedstocks containing bark under less severe conditions might result in the increased survival of these components, and a reduction in fermentability.  This information would be valuable in  establishing whether a mild pretreatment could be employed with bark-containing feedstocks, when maximum hemicellulose sugar recovery is required.  To ascertain whether the chosen pretreatment severity had a beneficial impact on the fermentation of the water-soluble fractions containing bark, a Douglas-fir whitewood feedstock containing bark (30% by weight) was pretreated under low-severity conditions. The resulting water-soluble fraction (WS-30Low) was fermented in parallel to WS-30 obtained using mediumseverity pretreatment (WS-30Med). Prior to fermentation, the composition of WS-30Low was determined (Table 4-5). This water-soluble fraction contained comparable hemicellulose sugar concentrations, but notably lower glucose, owing to the reduced hydrolysis of the cellulose component that is typical of pretreatment under less severe conditions. In terms of inhibitors, the water-soluble fraction derived under low-severity also had a decreased concentration of both H M F and furfural (Table 4-5) compared to the medium-severity water-soluble fraction. The concentration of total phenolics, determined by reaction with Folin-Ciocalteu reagent, was also lower (2.3 g L" vs. 2.9 g L" ), as was the concentration of lipophilic extractives (0.1 g L" vs. 0.4 1  1  1  g L" ). Based on these results, it appeared likely that the water-soluble fraction derived under 1  low-severity pretreatment would be readily fermentable. The two hydrolysates were fermented using the Tembec adapted yeast. To eliminate any difference in starting sugar concentration, 107  Table 4-5: Composition of the water-soluble fraction WS-30Low derived from a whitewood feedstock containing 30% bark by weight, pretreated under low-severity conditions. Samples were analysed in duplicate. Compound  Concentration (SL ) 1  Monomeric sugars: (range < 0.03 for all sugars)  Potential Inhibitors: (range < 0.04 for all components)  Arabinose Galactose Glucose Mannose Xylose  1.9 2.6 5.7 10.3 3.2  HMF Furfural Total Phenolics Lipophilic compounds  0.4 0.1 2.3 0.1  108  monomeric sugars were supplemented to hydrolysate WS-30Low to bring the starting sugar concentration to the same level as found in WS-30Med prior to fermentation.  As suspected, the water-soluble fraction derived under low-severity pretreatment was readily fermentable (Figure 4-6, A). In fact, virtually no difference was observed in terms of sugar consumption during the fermentation of either fraction.  This observation also held true for  ethanol production from these water-soluble fractions (Figure 4-6, B). Ethanol yields for both samples were also comparable, at 0.45 g g" for both the WS-30Low and WS-30 fractions. 1  These results indicated that higher pretreatment severity offered no noticeable improvement in terms of fermentability, unlike for the White-fir/Ponderosa pine mixed feedstock.  4.2.5 Metabolism of HMF and furfural during fermentation It was apparent that, during the fermentation of the water-soluble fractions containing bark, the concentration of H M F decreased steadily from the onset of fermentation (Figure 4-7). This was likely due to the metabolism of H M F to related compounds by the actively growing yeast culture (Sanchez and Bautista, 1988; Taherzadeh et al, 2000).  Taherzadeh et al, (2000) have  demonstrated that the principle conversion product is hydroxymethylfurfuryl alcohol, although the fermentation samples were not analysed to determine concentration of this by-product following fermentation.  Furfural can also be metabolized by yeast, largely via reduction to  furfuryl alcohol, but also oxidation to furoic acid (Palmqvist et al, 1999a; Taherzadeh et al, 1999).  The low initial concentration of furfural present in these water-soluble fractions  (maximum concentration of 0.34 g L" in WS-0) likely accounts for its rapid and complete 1  consumption by the 2-hour sample point (data not shown).  109  Figure 4-6: Fermentation of the water-soluble fractions (WS-30Low and WS-30Med) derived from steam-exploded whitewood/bark feedstock (30% w/w bark) under low- and medium-severity pretreatment conditions. The initial sugar concentration in WS30Low was adjusted with galactose, glucose and mannose to achieve a comparable concentration to fraction WS-30Med. (A) Hexose consumption; (B) Ethanol production. Error bars indicate the range of values for duplicate fermentations. 110  Time (hours)  Figure 4-7:  Metabolism of H M F during fermentation of the water-soluble fractions derived from Douglas-fir whitewood/bark feedstocks (0, 10, 20, 30 or 100% bark by weight) pretreated under medium-severity. Error bars indicate the range of values for duplicate fermentations.  Ill  Although chemically similar to furfural, H M F is generally reported to have a reduced inhibitory effect on microbial fermentations when compared to furfural at the same concentration. In fact, the relative toxicity of H M F is low when compared to most inhibitors present in the watersoluble fraction (Nilvebrant et al., 1997). However, it is apparent that H M F is also found at much higher concentration in these water-soluble fractions than are other potential inhibitors. This fact, combined with its slow breakdown (>24 h in all but WS-100), suggests that H M F could have a prolonged inhibitory effect on fermentation. Indeed, this fact may contribute to the slower rate of galactose fermentation in fraction WS-0 compared to the 100%-bark water-soluble fraction that was supplemented with sugars (Figure 4-5, A). Approximately half of the original concentration of H M F remained at the 6-hour point, when galactose metabolism was just beginning in the WS-0 fraction. Of the other inhibitors analysed, furfural had been exhausted from the fermentation broth by this time, while lipophilic components were found at relatively lower concentration in the 0% bark water-soluble fraction and were not expected to cause substantial inhibition.  Indeed, one would expect lipophilic compounds to have resulted in  stronger inhibition in WS-100. Furfural has been shown to affect several enzymes involved in sugar catabolism, including hexokinase, phosphofructokinase, and alcohol dehydrogenase (Banerjee et al, 1981b), and the mode of action for H M F appears to be similar. Thus, one could anticipate a reduction in the flux of galactose through glycolysis.  Furthermore, given the  multitude of enzymes required for the metabolism of galactose (Carlson, 1987; Gancedo, 1992), it is possible that H M F acts on another enzyme associated with the uptake or interconversion of galactose, thus decreasing the rate of its metabolism.  4.3 Conclusions It was apparent that pretreatment severity played a critical role in determining whether the watersoluble fractions derived from steam exploded Douglas-fir could be fermented.  Only under 112  milder pretreatment conditions (i.e., under low- and medium-severity) was fermentation to high ethanol yield possible.  It was thought likely that toxic by-products derived from the  pretreatment process resulted in decreased rates of fermentation, but ethanol yields of 0.44 g g"  1  were possible for both the low- and medium-severity whitewood hydrolysate. In contrast to pretreatment severity, the feedstock composition (i.e., bark content) had only a limited impact on the fermentability of the resulting process stream. In fact, when corrected for the significant decrease in sugar concentration, the water-soluble fraction WS-100 could actually be fermented more rapidly than the WS-0 fraction derived from a bark-free feedstock. This was likely related to the decreased concentration of process-derived inhibitors, in addition to the low recovery of naturally-occurring inhibitors from the bark. Although bark was evidently quite benign to the fermentation process, it was not clear whether the increased lignin and extractive components remaining in the water-insoluble component would permit the efficient enzymatic hydrolysis of the cellulose. Experiments to assess this issue are explored in the following chapter.  Although the water-soluble fractions were generally readily fermentable, a major concern during the current research was the low sugar concentration in these water-soluble fractions. Low sugar concentration directly translates into low ethanol concentration, which incurs a greater cost for distillation to anhydrous ethanol.  Furthermore, dilute streams can increase the capital and  energy costs associated with the transportation and storage of excess water. From a practical point of view, increasing the fermentable sugar concentration in the water-soluble fraction will likely be necessary. Depending on the particular strategy employed, the fermentability of the resulting water-soluble fraction may be affected. Attempts to increase the water-soluble sugar concentration are addressed in the subsequent sections of the thesis.  113  Chapter 5: Increasing the Sugar Concentration in the Water-soluble Fraction 5.1 B a c k g r o u n d  As detailed in the previous chapter, the water-soluble fractions derived from steam exploded Douglas-fir both whitewood and whitewood/bark feedstocks could be readily fermented to high ethanol yield. Although pretreatment severity did have an impact on the fermentation step, bark had no detrimental impact on fermentation of the water-soluble component. ethanol yields, the final ethanol concentration following fermentation  Despite good  was very low—a  consequence of the low initial sugar concentration in the water-soluble stream.  L o w sugar  concentration is typical of bioconversion processes that are optimized towards hemicellulose sugar recovery.  This arises at least in part because the complete recovery of the solubilized  carbohydrate from the water-insoluble fraction can require copious amounts of water, but also because under conditions that permit high hemicellulose recovery, only limited cellulose is solubilized from the substrate. Additionally, feedstock composition (e.g., high bark content) can further influence the sugar concentration of the water-soluble fraction, as shown already (Table 3-9).  Consequently, the sugar concentration in wood- and wood/bark-derived water-soluble  fractions reported in the literature is typically on the order of 30-60 g L " total sugars. 1  The  fermentable water-soluble fractions described in the current work (low- and medium-severity) ranged from 13-43 g L " total sugars, depending on the bark content, while the concentration of 1  readily fermentable  hexose sugars was only 8-37 g L " . 1  In contrast  with commercial  ethanologenic fermentations, which generally operate at sugar concentrations on the order of 100-200 g L " glucose, the hemicellulose-rich wood water-soluble streams tend to be very dilute. 1  Improving the sugar concentration in the water-soluble fraction would have economic benefits for bioconversion.  Short of major modifications to the steam explosion process, the sugar 114  concentration could be increased by either removing water from the dilute stream (i.e., concentrating the stream), or by supplementing the stream with additional sugars. There are numerous commercially-proven physical concentration techniques available for concentrating process water, but due to generally disappointing economics (Blomgren et al, 1991; Zacchi and Axelsson, 1989), there has been, limited research on their application to bioconversion. Although physical concentration is decidedly energy intensive, it is possible that technological advances could make concentrating the water-soluble fraction economically viable.  The research  presented here thus addresses some practical limitations to physical concentration, rather than economic ones. Furthermore, it was thought that results from these concentration experiments would reveal whether  higher concentration  fractionation would be fermentable.  water-soluble streams recovered  following  The alternative approach of supplementing the water-  soluble fraction with additional sugar may prove more practical for bioconversion than a potentially energy-intensive concentration step. This could be readily accomplished by taking advantage of the carbohydrate remaining in the water-insoluble cellulose component. Provided this component can be effectively converted to glucose, the resulting glucose stream should be fermentable.  The work reported in this section of the thesis addresses the attempts to increase the sugar concentration  of the  water-soluble  fraction  by either  physical concentration,  or by  supplementation with additional carbohydrate derived from the cellulose-rich, water-insoluble fraction.  Initially, two concentration techniques (rotary-evaporation and freeze-drying) were  selected as ways of increasing the sugar concentration of the water-soluble fraction either 2- or 3-fold  prior to fermentation.  Subsequently,  supplementation  with carbohydrate  was  accomplished in two ways: by hydrolysing the cellulose component directly in the water-soluble fraction, or by first hydrolysing the cellulose component and then combining the glucose-rich 115  cellulose hydrolysate (CH) with the water-soluble fraction.  The improvement in sugar  concentration will be described, and the effect of each protocol on the subsequent fermentation to ethanol will be reported on and discussed.  5.2 Results and Discussion 5.2.1 Concentrating the water-soluble fraction 5.2.1.1 Feedstock composition and pretreatment Two water-soluble fractions were prepared from a new Douglas-fir whitewood feedstock (DF-0) and a mixed whitewood/bark feedstock (30% bark by weight, DF-30).  When the chemical  compositions of the original wood and bark feedstocks were determined (Table 5-1), it was apparent that the whitewood was virtually identical to previously utilized feedstocks (Table 3-6). In contrast, the bark component contained notably higher glucan content and a lower lignin and extractive content than the previous bark feedstock. The high glucan content, which was nearly double that determined previously, was thought likely related to the prominent inner bark component of the feedstock, when compared to the outer bark. It is recognized that the inner and outer bark fractions of woods differ significantly in chemical composition (Hafizoglu et al, 1997; Voipio and Laakso, 1992), with the inner bark generally being richer in carbohydrates.  The experimental approach that was followed for the physical concentration experiments is outlined in Figure 5-1. The feedstocks were pretreated using medium-severity conditions, and subsequently processed to recover the highest possible sugar concentration in the water-soluble fraction. This required minimizing the utilization of water to recover the steam exploded slurry from the receiving vessel after pretreatment. The recovered slurry was subsequently fractionated at as high a solids content as possible, to prevent dilution of the recovered water-soluble stream. Despite these efforts,  it was difficult  to achieve a dry weight consistency greater 116  VO  rt  d  00  d  r t  oo  ON O N  h-l  00  c ,  o d  o d  Tt  VO  < d  b0  CN  d i—i  <  d CO  d  ON O N  vd co CN  CL)  CO  in  _ >  O  VO  rt  ° )  ^  CN  c  rt  CN  d d co X oq CO CN 03  a 03 c c  CN  d  03  CL)  03  in  d  ' f t  Tt  d  oo r -  !•§  Tt  CN  , ^  u  c 03  -*—>  o 03  13  O 03  — \  e d  'I  /•—\  CN C N  rt  d CO  \< o o  4-»  oo CL)  ,  <^>  S ^ O 03 ^ m  •a btt.s C  CO  rt  fl  -O  d insol id solu  c  c  << ra co  <<  C3  X)  Douglas-fir wood and bark feedstocks (DF-O, DF-30)  1  Steam Explosion  Water-Solublefraction(WS-0, WS-30)  Water-Insoluble fraction (WI-0, WI-30)  Concentrated water-soluble fraction (RV-0, RV-30 and FD-0, F D 3 0 )  Figure 5-1: Process flow diagram for concentrating the water-soluble fractions derived from Douglas-fir whitewood (DF-O) and mixed whitewood/bark (DF-30), illustrating the key steps of pretreatment, fractionation, and concentration by rotary-evaporation ( R V ) or freeze-drying (FD), as required.  118  than 15% in the slurry prior to fractionation, due primarily to the condensation of steam during pretreatment. The decision to use a relatively high dry matter consistency during fractionation also resulted in the failure to recover all of the soluble sugars from the water-insoluble component, as discussed previously. While subsequent washes could have recovered these sugars, adding them to the water-soluble component would have dramatically reduced the sugar concentration in the resulting stream.  The recovered water-soluble components were essentially as concentrated as the current steamexplosion process would permit. However, analysis of the water-soluble fractions derived from pretreated whitewood containing no bark (WS-0) or whitewood mixed with 30% bark (WS-30) indicated that the sugar concentration was still very low (Table 5-2).  The total sugar  concentration was higher in WS-0 than in WS-30, at 40.5 g L" compared to 33.5 g L" , 1  respectively.  1  The water-soluble fractions were next concentrated 2- or 3-fold by rotary-  evaporation or freeze-drying.  These techniques were selected for several reasons.  procedures were easily implemented in the lab.  Both  Rotary-evaporation is essentially thermal  evaporation under vacuum, and evaporation has been used for the commercial concentration of sugar streams. From a process perspective, thermal evaporation was thought to be suited to the concentration of the water-soluble component following steam explosion, due to the high temperature of the slurry upon exiting the steam gun.  If concentrated while still hot (e.g.,  immediately following recovery/fractionation), the need for additional thermal energy would be reduced.  Furthermore, some of the remaining volatile inhibitors might be lost during  evaporation, affording a mild detoxification of the water-soluble fraction. In contrast, freezedrying was selected because it was expected that the volatile components would be retained during concentration, owing to the low temperature.  Both techniques provided the requisite  increase in sugar concentration. For the whitewood concentrate, this resulted in a combined 119  o I  cd  00  O  rH  d  ]| d  T t  U  oo  O  O  p.l CO  d  I  OH B  cd  00  (U  CD  o d o  PH.  VO  CN  1/3 T 3 -2  VH  O  D  ^£  Q  g>  4H  . 2  oo IH  ^  Cd  =3  CD  loo  °  "cd  ft-S«  42  , VO  d  r  ©  >n i n  d co CO  •4—<  1.2  <D oo  S -n cd CD  C  CD  cd  <D oo O  CD )H  IX  r-H  O  d  d  ON  t--  T t  CO  cd JT^  oo <D  w> >  o co° .y 4-J I *  CN  O  CO  ON  lo, d d  >n  r>  r2  M  IS VH  O  B  CO  CD  c.  t+H  C  O  c O 'OO  o ft £  o  o ° 1 / 1  rC  u  7 3  CD oo O  o  3  CN  >  ON  00  T t  CO  d  o  lo  CD  oo O  -H  d o  •4—1  4 2I °} "cd O  O  d  ^ co co  o  £  —, CD <3 .-a  U cd  o  <  o  o  in  oo  £  oo  ' 2 2,  CN i CO  sugar concentration of approximately 81 and 121 g L" , while the mixed wood/bark feedstock 1  concentrates contained about 67 and 100 g L" total sugars for the 2- and 3-fold concentrates, 1  respectively.  5.2.1.2  Fermentation of the 2-fold water-soluble concentrates  The fermentability of the water-soluble fractions was next compared, using the Tembec SSLadapted yeast for fermentation (Figure 5-2, A and B). As is evident from the results, both of the original-concentration water-soluble fractions (WS-0 and WS-30) were readily fermentable. This was expected, as in earlier results, the Tembec yeast had no difficulty fermenting the equivalent water-soluble fractions. In addition, as previously demonstrated, the water-soluble fraction containing bark was fermented at a slightly faster rate than the whitewood-only fraction. The rate of sugar consumption by the yeast (calculated after the first 3 hours) was 2.3 ± 0.4 g L" h" for the whitewood water-soluble fraction (WS-0), compared to 2.9 ± 0.2 g L" 1  1  1  h" for WS-30. However, virtually all of the hexose sugars (galactose, glucose and mannose) 1  were fully consumed in both water-soluble fractions by 24 h. The slight discrepancy in initial sugar concentration between the two water-soluble fractions (~7 g L" ) was not anticipated to 1  contribute greatly to differences in fermentation. After 48 h fermentation, the calculated ethanol yields from the available hexose sugars were high (Table 5-3). However, given the low starting sugar concentration, ethanol concentration only reached a maximum of 15.2 g L" andl2.4 g L" 1  1  in the water-soluble fractions WS-0 and WS-30, respectively. When the sugar concentrations of the water-soluble fractions were doubled by rotary-evaporation (RV) or freeze-drying (FD) prior to fermentation, there was a dramatic difference in fermentation between whitewood-only concentrates (RV-0, FD-0) and bark-containing concentrates (RV-30, FD-30). The whitewood concentrates were both poorly fermented (Figure 5-2, A ) , and produced negligible amounts of ethanol (Table 5-3). Furthermore, the limited quantity of ethanol produced did not fully account 121  Figure 5-2: Fermentation of the original and 2-fold concentrated water-soluble fractions derived from (A) Douglas-fir whitewood chips and (B) whitewood chips supplemented with 30% bark. Error bars indicate range of values for duplicate fermentations.  122  Table 5-3:  Summary of the fermentations employing the original and 2-fold concentrated water-soluble fractions (obtained by rotaryevaporation and freeze-drying). Values in parentheses indicate the range for duplicate fermentations.  Water Soluble Fraction  Hexose Concentration t = 0h (gL" )  Ethanol Concentration t = 48h (gL" )  Ethanol Yield  WS-0  33.1 (0.1)  15.2 (0.1)  0.46 (0.00)  WS-30  26.3 (0.1)  12.4 (0.0)  0.47 (0.00)  RV-0  66.6 (0.3)  3.1 (1.2)  0.05 (0.02)  RV-30  53.0 (0.3)  24.6(0.1)  0.46 (0.01)  FD-0  67.0 (0.0)  6.3 (0.8)  0.09 (0.01)  FD-30  52.9 (0.1)  25.0 (0.1)  0.47 (0.00)  60.2 (0.0)  28.9 (0.1)  0.48 (0.00)  1  Reference a  b  b  a  (g g ) 1  1  Ethanol yield based on available hexose sugars at t = 0 h Reference fermentation was G M Y P medium supplemented with glucose to achieve a concentration of - 6 0 g L " at the start of fermentation 1  123  for the consumption of sugars from the medium. In contrast, the sugars in both RV-30 and FD30 were rapidly consumed by the yeast, with most sugars consumed after 24 h, and only trace quantities of galactose remaining after 48 h (Figure 5-2, B).  Consequently, ethanol  concentration was much higher in the bark-derived concentrates (Table 5-3), with yields on the order of 0.46 g g" . Owing to the poor fermentation of the 2-fold whitewood concentrates, the 31  fold whitewood concentrates were not used for fermentation experiments.  The decrease in fermentation efficiency following concentration was thought to be related to the accumulation of inhibitory materials in the water-soluble fractions.  The impact that rotary-  evaporation and freeze-drying had on the concentration of potential inhibitors was dependent on the volatility of the inhibitor. In a preliminary test with a synthetic cocktail containing five recognized inhibitors (HMF, furfural, acetic acid, levulinic acid and formic acid at 1 g L" each), 1  the low-volatility inhibitors (HMF and levulinic acid) were observed to increase almost proportionately with concentration factor (Table 5-4).  In contrast, furfural, acetic acid and  formic acid were partially volatilized during concentration, resulting in a lower than anticipated concentration based on a 2-fold concentration factor.  It was apparent that the method of  concentration had a pronounced impact on the final concentration of inhibitors in the cocktail. It was initially suspected that the more volatile compounds would be lost to a greater extent during rotary-evaporation, due to the increased temperature (~50°C), while freeze-drying was anticipated to provide greater retention of these components. However, both techniques resulted in substantial losses. While more furfural was lost by R V than by FD, losses of acetic and formic acid were greater after FD, with approximately a 50% reduction in the concentration of these components.  Overall, freeze-drying resulted in a greater loss of the volatile inhibitors  tested, which may partly explain the improved fermentation of these hydrolysates compared to RV-0 and RV-30. The increased loss of volatiles was likely related to the extended time under 124  Table 5-4: Concentration of inhibitors in a synthetic cocktail following 2-fold concentration by rotary-evaporation or freeze-drying. Inhibitors were present in the original cocktail at a concentration of 1 g L " . Samples were concentrated in duplicate. Values in parentheses indicate the range. 1  Inhibitor Concentration (g L" ) 1  Method  Fold Increase  RV FD  HMF  Furfural  Acetic Acid  Formic Acid  Levulinic Acid  2.09 (0.05)  2.02 (0.05)  0.06 (0.01)  1.57 (0.02)  1.65 (0.02)  2.08 (0.04)  1.99 (0.00)  1.92 (0.01)  0.50 (0.02)  0.56 (0.02)  0.51 (0.02)  1.96 (0.01)  3  'The cocktail was concentrated to a target of 2-fold, but was not adjusted with water to achieve a precise 2-fold increase in concentration.  125  vacuum required for the freeze-drying process (greater than 48 h), compared to rotaryevaporation (~5 min).  Analysis of the inhibitors in the original and concentrated water-soluble fractions used in the current fermentation revealed similar trends for changes to the concentration of inhibitors. The concentration of H M F in WS-0 increased from approximately 1.5 to 3.0 g L" for a 2-fold 1  increase in the concentration (Table 5-5).  Likewise, the concentration of total phenolics  (determined by reaction with Folin-Ciocalteu's reagent) increased almost proportionately with concentration factor. While the furfural concentration decreased following rotary-evaporation and freeze-drying (relative to the concentration in WS-0), this decrease was not as substantial as in the synthetic cocktail. As in the cocktail, an accumulation of acetic acid (-1.5 fold) was observed following a 2-fold concentration by rotary-evaporation, while losses during freeze drying resulted in only the equivalent concentration as found in the original water-soluble fraction.  These differences in retention of inhibitors in the real hydrolysates may have been due  to the presence of other components in the hydrolysates. Poor separation of levulinic and formic acid from other UV-absorbing compounds prevented accurate quantification of these compounds in the hydrolysates. Based on results from the synthetic cocktail, it was anticipated that the levulinic acid concentration would double. Without an alternatively means of quantifying the concentration of formic acid in the hydrolysates, it is difficult to accurately predict what might have happened to its concentration.  However, this preliminary evidence does indicate that  physical concentration can have a pronounced impact on the inhibitor composition of the watersoluble fraction, resulting in the accumulation of non-volatile and moderately-volatile components. In the case of the whitewood concentrates, the accumulation of these inhibitors reached an apparent threshold concentration where performance of the yeast suffered drastically (Figure 5-2).  In contrast, inhibitors likely did not accumulate to a concentration sufficient 126  Table 5-5: Concentration of inhibitors (g L" ) in the original water-soluble fractions (WS-0 and WS-30) and the 2-fold concentrates obtained by rotary-evaporation (RV) and freeze-drying (FD). Data is for analysis of the water-soluble fractions with a single replicate, except for analysis of total phenolics (n = 2) 1  Inhibitor Concentration WS Fraction  HMF  Furfural  Acetic Acid  WS-0  1.55  0.38  2.12  Total Phenolics 3.32 (0.12)  RV-0  3.02  0.14  3.12  6.48 (0.02)  FD-0  3.05  0.18  2.25  6.44 (0.08)  WS-30  1.19  0.33  1.49  2.90 (0.21)  RV-30  2.35  0.16  2.45  5.51 (0.03)  FD-30  2.30  0.18  1.49  5.43 (0.01)  127  inhibit the fermentation when the concentrates RV-30 and FD-30 were used.  This is likely  related to the lower initial concentration of inhibitors in WS-30.  5.2.1.3  Fermentation of the 2- to 3-fold concentrates (WS-30)  Based on positive results for the fermentation of RV-30 and FD-30, the upper limit to concentration was investigated (Figure 5-3).  The water-soluble fraction WS-30 was  concentrated by rotary-evaporation to 3-fold and then diluted with water back to 1-, 2-, 2.25-, 2.5-, 2.75- and 3-fold.  In all cases, to eliminate the influence of sugar concentration on  fermentation, all the concentrates were supplemented with galactose, glucose and mannose to achieve the same level as found in the 3-fold concentrate. The decision to first concentrate the water-soluble fraction and then dilute back to the desired amount meant a potentially greater loss of volatile inhibitors (due to removal of a greater percentage of the starting material), which might translate into better fermentation than if the water-soluble fraction was concentrated incrementally to the desired level. To ascertain whether this would be an issue, a comparison was made between a concentrate obtained by rotary-evaporation to 2-fold only (2.Ox CONC) and one obtained by first concentrating to 3-fold, followed by dilution to 2-fold (2.Ox DIL). No major differences in sugar consumption was observed between these hydrolysates, with only a slight improvement in total sugar consumption in the 2.0x D I L fraction after 72 h. This result suggests that any additional loss of volatile inhibitors had only a relatively minor influence on the fermentation.  As the concentration of the water-soluble fraction WS-30 was increased incrementally up to 3fold, a significant lag in hexose sugar consumption was observed (Figure 5-3, A). This lag was most likely caused by the accumulation of inhibitory material in the water-soluble fraction during the concentration step, and is in agreement with findings by Lee et al, (1999), working 128  Figure 5-3: Fermentation of the water-soluble fraction concentrated 3-fold by rotaryevaporation, followed by dilution to 2.75-, 2.5-, 2.25-, 2-, and 1-fold. A 2-fold concentrate was also prepared without dilution from the 3-fold stock for comparison (2.Ox CONC), as described in the text. Monomeric sugars were supplemented to each fraction to achieve a comparable starting sugar concentration. A simulated water-soluble fraction (Sim-WS, containing sugars in water) and reference fermentation ( G M Y P medium) were also used. (A) Hexose sugar concentration; (B) Ethanol concentration. Error bars indicate ranse of duplicate fermentations. 129  with concentrated oak wood hydrolysates.  Although increased sugar concentration in the  medium can also contribute to slower metabolic rates, the reference fermentation for this experiment was rapidly metabolized. The reference fermentation medium consisted of G M Y P containing glucose at ~75 g L" as the sole fermentable substrate, but with additional, non1  fermentable arabinose and xylose to ensure a comparable total carbohydrate concentration to the water-soluble fractions. Up to a concentration factor of 2.5-fold, it remained possible for the yeast to consume virtually all glucose and mannose within 24 h. Beyond this concentration, only minimal amounts of sugar were consumed prior to the 24-hour point in the fermentation. However, after this lag period, carbohydrate metabolism resumed and both glucose and mannose were fully consumed. It is well known that microorganisms such as yeast can actively acclimate to and detoxify their environment, metabolizing various inhibitory compounds such as furfural, HMF  and  phenolics  such  as  vanillin  to  less  toxic  compounds  (e.g.,  furfuryl,  hydroxymethylfurfuryl, and vanillyl alcohols) (De Wulf et al, 1987; Larsson et al, 2000; Taherzadeh et al, 1999; Taherzadeh et al, 2000).  Although the increased toxicity of the  concentrates may result in the death of many cells, due to the relatively high inoculum that was used (~6 g O D W L" ), sufficient cells remained viable to re-establish the culture population and 1  permit fermentation of the available sugars.  Previously, it was suggested that a high cell  inoculum could be used as a viable strategy to overcome inhibition during fermentation of wood hydrolysates (Chung and Lee, 1985).  It is worth noting that galactose was not effectively metabolized during fermentation of these concentrates.  This is a potential problem, as residual sugars represent both a loss in terms of  potential ethanol production and a cost in terms of effluent treatment.  It is possible that  galactose metabolism was more severely affected by the increased concentrations of inhibitors or even the higher ethanol concentrations following fermentation of both glucose and mannose. 130  However, a more likely explanation for this observation was the low nitrogen content.  As  discussed briefly previously, a low assimilable nitrogen content can also result in incomplete fermentation.  Although sugar concentration was tripled, due to an oversight, no additional  nitrogen supplement was added. The results obtained for fermentation of the simulated watersoluble fraction support this conclusion. The concentration of sugars in the simulated watersoluble fraction (i.e., water containing arabinose, galactose, glucose, xylose and mannose) was comparable to each of the concentrated Douglas-fir hydrolysates, but was free of inhibitors. Galactose metabolism in this culture was also incomplete, supporting the claim that the deficient metabolism of galactose in the wood hydrolysates was not caused by the presence of inhibitors. The metabolism of galactose via the Leloir pathway requires the expression of numerous proteins upon exhaustion of glucose and mannose from the culture medium (Ostergaard et al, 2000) and the de novo synthesis of proteins under nitrogen-limiting conditions can be a significant obstacle for the yeast to overcome.  Ethanol production followed trends closely related to the consumption of sugars (Figure 5-3, B). It is interesting that despite the initial repression of cell metabolism, ultimately all of the concentrates were fermentable to ethanol, regardless of initial concentration factor. However, it was apparent that productivity was very low in the higher-concentrated water-soluble fractions (e.g., 2.75- and 3-fold). This metabolic inhibition is consistent with inhibition by compounds such as the furans H M F and furfural, which have previously been reported to inhibit C O 2 production (Banerjee et al., 1981a; Sanchez and Bautista, 1988), and consequently, ethanol production under anaerobic conditions. Due to the significant removal of furfural from the water-soluble fractions during rotary-evaporation, it is possible that H M F might account for the majority of this lag. Its concentration increased from approximately 1.2 g L" in WS-30 to 3.4 g 1  L" in the 3-fold concentrate. However, numerous possible inhibitors including total phenolics 1  131  (which increased to 8.2 g L" in the 3-fold concentrate) and bark-derived extractives were also concentrated by rotary-evaporation, and may have further contributed to the observed lag.  Ethanol production and yield during these fermentations was determined, and in all cases, yields for ethanol were lower than would be expected based on the sugar concentration present in the water-soluble fractions (Table 5-6). This was due mainly to the incomplete metabolism of the hexose sugars.  Yields calculated on the available hexose sugars in the various concentrates  ranged from 0.36 to 0.43 g g" , with generally decreased yields for an increase in hydrolysate 1  concentration.  The yield for the simulated water-soluble fraction was comparable to that  established for the water-soluble fractions concentrated between 1- and 2.5-fold, and was actually lower than the 2- and 2.25-fold concentrates owing to a slightly greater residual sugar concentration after 72 h. In contrast, the yield for the reference glucose fermentation (GMYP medium) was notably higher at 0.46 g g" due largely to the complete metabolism of the 1  carbohydrate component. This yield was lower than the theoretical maximum of 0.51 g g" , due 1  in part to the production of new biomass during the fermentation, which reduces the carbon available for ethanol production.  To account for the slight differences in galactose metabolism in each of the concentrates, and the impact that this would have ethanol production, the ethanol yield was also calculated based on consumed hexose sugars (Table 5-6).  It was apparent that ethanol production from the  metabolized sugars was quite high for most fermentations, despite the accumulation of inhibitory components during rotary-evaporation. At lower concentration factors (1- to 2.5-fold), ethanol yields ranged from 0.44 to 0.46 g g^-consumed sugar. Above a concentration factor of 2.5-fold, a decrease in ethanol yield was noted, although the reduction was not large. This lends support to the argument  that other compounds in addition to H M F may have been partly 132  Table 5-6: Ethanol yield from the water-soluble fraction WS-30 concentrated 1-, 2-, 2.25-, 2.5-, 2.75-, and 3-fold by rotary-evaporation prior to fermentation. A simulated water-soluble fraction (Sim-WS, consisting of sugars in water) and reference fermentation ( G M Y P medium) are included. Values in parentheses indicate the range for duplicate fermentations.  a  b c  Concentration Factor  Max. Ethanol Concentration (g L" )  l.Ox 2.0x  29.8 (0.0) 32.1 (0.2)  Ethanol Yield (Avail. Hex.) (g g ) 0.41 (0.00) 0.43 (0.01)  2.25x  31.1 (0.2)  0.42 (0.01)  0.46 (0.01)  2.5x  29.6 (0.1)  0.40 (0.00)  0.44 (0.00)  2.75x  29.1 (0.4)  0.38 (0.01) 0.36 (0.04)  0.42 (0.02)  1  a  b  1  3.0x 26.9 (2.5) Sim-WS 30.5 (0.1) 0.41 (0.00) Reference 33.6 (0.0) 0.46 (0.01) Determined after 72 h; reference fermentation after 24 h Calculated as grams ethanol per gram available hexose sugars Calculated as grams ethanol per gram consumed hexose sugars  Ethanol Yield (Cons. Hex.) (g g"') 0.45 (0.00) 0.45 (0.02)  c  0.41 (0.07) 0.45 (0.01) 0.46 (0.01)  133  responsible for inhibition of the fermentation. The impact of H M F and furfural on fermentation typically results in decreased growth rates, sugar consumption (Taherzadeh et al, 2000) and ethanol productivity (Palmqvist, 1998), while ethanol yield is reported to remain largely unchanged (Larsson et al, 1999a) or even slightly improved (Palmqvist, 1998; Taherzadeh et al, 2000). In contrast, other inhibitors present in the wood hydrolysates such as weak acids (e.g., formic, acetic and levulinic acid) have been implicated in a reduction of ethanol yield (Larsson et al,  1999a), and both acetic and levulinic acids were previously observed to increase in  concentration following rotary-evaporation of the water-soluble component.  A n alternative  explanation for the decrease in ethanol yield during fermentation of the 2.75- and 3-fold concentrated water-soluble fraction may be related to the re-growth of the culture, which will have suffered significant loss of viability during the initial hours of the fermentation due to the toxicity of the water-soluble fraction (Chung and Lee, 1985).  A preliminary investigation  indicated significant cell death when the fermentation of a non-concentrated Douglas-fir woodand bark-derived water-soluble fraction was assessed. (Kwong, 2000).  5.2.1.4  Final observations  Although accumulation of the inhibitory products was anticipated to be a potential problem for fermentation, it was not expected to play such a significant role at the relatively low concentration factors employed in this work (2- to 3-fold). Despite the approach of employing mild to moderate pretreatment severity in an effort to minimize the production of fermentation inhibitors, inhibitors appeared to remain too concentrated to permit even a doubling of the sugar concentration in the whitewood-derived water-soluble fraction. However, the accumulation of inhibitors was of less concern with the water-soluble concentrates derived from the water-soluble fraction WS-30.  The WS-30 fraction was previously shown to ferment more rapidly than  whitewood water-soluble fractions prepared under comparable conditions (WS-0) and the same 134  is true of the concentrates RV-30 and FD-30. Although concentrating the water-soluble fractions generally resulted in increased inhibitors levels, the initial inhibitor concentration in WS-30 was apparently sufficiently low that even after a 2-fold increase, fermentation was not severely inhibited. Exactly why the concentration of inhibitors was so much lower in fraction WS-30 than in fraction WS-0 is unclear. The decrease in available carbohydrate in the bark-containing feedstock may partly account for the decreased H M F and furfural concentration in the watersoluble fraction, but it is thought likely that other mechanisms were also involved which might explain the low concentration of both phenolic compounds and extractives recovered in the barkderived water-soluble fractions. As postulated earlier, it is possible that the inclusion of bark in whitewood feedstocks results in the partial removal of inhibitors, due to hydrophobic interactions and possibly chemical condensations. Whether a similar mechanism accounts for the decreased toxicity of these bark-derived, water-soluble components remains to be fully resolved.  Based on the results presented here, it appears likely that attempts to increase the overall ethanol concentration derived from the water-soluble fraction by concentration of the hemicellulosederived sugar stream, will suffer from both the cost and technical complexity of the concentration method used, and the possibility that potentially inhibitory components may also be increased in concentration to inhibitory levels. The fermentation of the concentrated watersoluble fractions would likely benefit from the use of a detoxification step prior to fermentation (e.g., overliming, solvent extraction, etc.), as has been reported (Lee et ai, 1999). However, in the interest of minimizing additional complexity and costly processing of the water-soluble component, this approach was not investigated. Rather, alternative strategies for increasing the sugar concentration were next investigated.  135  5.2.2  Supplementation of the water-soluble fraction with carbohydrate  Given the current research's emphasis on reducing fermentation inhibitors to improve the fermentation step, neither rotary-evaporation nor freeze-drying seems particularly well suited to bioconversion.  Without the use of a detoxification step, only water-soluble fractions with  significantly reduced toxicity could ever be concentrated and fermented.  Thus, an alternative  strategy of using supplemental carbohydrate derived from the water-insoluble cellulose/lignin component may prove to be a more successful way of increasing sugar concentration in the water-soluble fraction. As mentioned previously, two different approaches were evaluated. In the first, the water-insoluble component was enzymatically hydrolysed directly in the watersoluble fraction. However, it was not known whether the selected steam explosion conditions (medium-severity) were sufficient to permit efficient enzymatic hydrolysis. Thus, experiments were first conducted to evaluate hydrolysis of the water-insoluble components (i.e., WI-0 and WI-30) in buffer.  5.2.2.1  Delignification component  and  enzymatic  hydrolysis  of  the  water-insoluble  Prior to the enzymatic hydrolysis of the water-insoluble components, it was necessary to perform a delignification of the substrate. This was to ensure efficient hydrolysis with relatively low enzyme loading. Pretreatment by steam explosion results in only minimal solubilization and removal of lignin, and high lignin content has been implicated with decreased efficacy of hydrolysis (Cowling and Kirk, 1975; Millett et al, 1975; Sewalt et al, 1997). Previously, it has been shown that the impact that high lignin content has on hydrolysis can be at least partly overcome by the use of greater enzyme loadings. For example, near-complete hydrolysis of Douglas-fir whitewood pretreated under medium-severity conditions could be achieved with a relatively high loading of 60 F P U g-cellulose  1  (Boussaid et al, 2000; Wu et al,  1999).  136  However, it is unlikely that such a high enzyme loading would be commercially possible, due to the recognized cost of the enzymes. The goal of the current work was to achieve hydrolysis of the steam-exploded Douglas-fir pulp using minimal cellulase loading, and experiments were designed to use only 10 F P U g-cellulose" . 1  Delignification was performed following a method described by Yang et al, (2002).  This  method was developed for the delignification of steam exploded Douglas-fir whitewood feedstocks, and it was not clear whether the conditions would be directly applicable to delignification of the bark-containing water-insoluble component, WI-30. Delignification by alkaline peroxide treatment resulted in the significant removal of lignin from the water-insoluble components WI-0 and WI-30. On a percentage basis, the lignin content in the original, steam exploded water-insoluble components was high, at approximately 44% and 56% for WI-0 and WI-30, respectively (Table 5-7). The high proportion of lignin in the water-insoluble fraction was due to the solubilization and removal of the hemicellulose sugars during pretreatment and fractionation.  Delignification by alkaline peroxide treatment significantly reduced the lignin  content of both water-insoluble components, but the residual lignin content of WI-AP-30 (18%) was approximately double that remaining in the whitewood-only fraction (9%). Furthermore, the decrease in lignin content after alkaline peroxide treatment was greater for the whitewood feedstock than for the mixed whitewood/bark feedstock. This is most likely related to the higher initial lignin content of the water-insoluble component containing bark.  5.2.2.2  L o s s of c e l l u l o s e by delignification  The hydroxyl radical, which arises due to the decomposition of hydrogen peroxide, is reported to indiscriminately attack both lignin and carbohydrate (Kadla et al, 1999).  Under alkaline  conditions at high temperatures, cellulose may be degraded by an endwise, "peeling" 137  oo co c x o  s  c  ©  cd  d —  cd  o 2 > CO  S  o  H  OH '« 23 ^  fi  Cd  3 x  c  t>0 -O  £1 .  SL  JO  3co CO oo ccd "§ 8 CO "V 2 cd 5x •a  <  ©  r-;  ON  ON  ON  o  o  o  d  in  VO  Tt  to  d d  d  d  00  CO  d  CM  r-l  cd  --  Tt Tt  O VD >T)  rH  4-»  »^  cd  oo  g  ©  ©  k  ^ O O  CO  a a  .  1  Si <*> c S  a  3 7  •^  <0 _3  cd  x ^ o O  r-H  ^ ^  ^  Cd  T3  r—<  O  K>  (3  C  O  £  co  4—*  cd '— T3  X! O X  b JO _ OO  CO  ft* <  OH  X  CO  Se o  •  d  o  o  o  d  Co  d d CN  d  in  oo  Tt  o  Tf  o  d  d  in d  o  d  vo r-  d d in p  d  in d  in r-  vd  o  i—i i—i  '  d  d  o  o  i>  o  d  d  d  rH CU O , CU  o  .5 ^  3 =3 &  £  r ^ CO  o 4—> OO  CU  es H  d d  o  d 1  r-H  d  d  o  d  o  o  d  o  11  a o U  X  CO  o  o  bo rn  ©  CO  cd  OO  OO  co  d  VH  X  •w  U  cd  CO  w  cd  —'  o a  Q <4-H O  •H  ON  CO  d  '.  K  S.  ">•>  d  ^ £ r=H O  d  ©  d  oo vo  6 8  co  d d  d  3 a, * 9 < v  t>  d d  d  CO  y  e  co cd  > • -H IH  d — V  ON  VO  OH  CO  Co  T—1  00  « "S  8  T P  d  Y  co  Pu  o  co  •a .a 00 X  o  HO  oo  ."2  o  co i  oo  Cu  OH  •i-H  <  If- i-U  M ri!  3 ©  i  c  X!  u  'o  mechanism, involving isomerization, enolization, and ultimately 6-alkoxy elimination of the terminal glucose from the cellulose chain (Sjostrom, 1993).  Thus, delignification by alkaline  peroxide could potentially result in significant losses of cellulose, which would obviously be detrimental to the sugar recovery and bioconversion in general. In the current work, the loss of cellulose following delignification was relatively small (-7-10%) (Appendix 4).  This was  determined by gravimetric analysis of the residual fibre, in combination with Klason analysis of the material, both before and after treatment, and is in agreement with results published previously using steam-exploded Douglas-fir (Yang et al, 2002; Cullis, 2003). Furthermore, it has been suggested that mild interactions with the cellulose fraction during delignification of lignocellulosics may be beneficial, as it can result in the modification of cellulose, resulting in decreased crystallinity of the feedstock (Gould, 1985). This effect may also contribute to the observed improvement in hydrolysis, in terms of both rate and yield.  5.2.2.3  Enzymatic h y d r o l y s i s of the original a n d delignified water-insoluble components  The original water-insoluble solids fractions, WI-0 and WI-30, and the delignified substrate, WIAP-0 and WI-AP-30, were enzymatically hydrolysed to glucose using a complete cellulase system derived from the fungi Trichoderma reesei (Celluclast 1.5L supplemented with additional Novozym-188 to alleviate end-product inhibition). The impact of the delignification process on the yield of glucose following enzymatic hydrolysis was dramatic.  In trial hydrolysis  experiments (2% w/v consistency in sodium acetate buffer) using the non-delignified waterinsoluble components, very poor glucose yields were observed (Figure 5-4). After 48 h, only -32% and -13% of the available glucose was liberated from the feedstocks WI-0 and WI-30, respectively, with minimal improvement after as long as 90 h hydrolysis. These results were much worse than those previously published for non-delignified medium-severity Douglas-fir  139  100  Time in Hours Figure 5-4: Enzymatic hydrolysis of the water-insoluble component (2% consistency in A C buffer) derived from steam exploded Douglas-fir whitewood (WI-0) and whitewood/bark feedstocks (WI-30) pretreated under medium-severity conditions, with or without delignification by alkaline peroxide (AP) treatment. Error bars indicate standard deviation for hydrolyses performed in triplicate.  140  feedstocks (Boussaid et al, 2000; Wu et al, 1999). However, only one-sixth of the enzyme loading was used in the current work (10 vs. 60 F P U g-cellulose" ). The poor hydrolysis, 1  characterized by both a decreased rate and extent of hydrolysis, was likely a consequence of the higher lignin content, but also its nature and location. It was apparent that the pretreatment conditions used in this work did not result in significant removal of lignin from the feedstock. Furthermore, the liquefaction and coalescence of lignin during pretreatment can result in increased quantities of lignin on the surface of the fibre, which blocks enzyme accessibility to the cellulose (Donaldson et al, 1988; Ooshima et al, 1990). In addition, the lignin of softwoods is quite reactive and has been shown to undergo condensation reactions at the a- and 6-carbon positions via the C-4 hydroxide and free C-5 position during pretreatment (Glasser et al, 1983; Shevchenko et al, 1999). This condensed lignin can further restrict access to the substrate by the cellulase enzymes. In contrast, when the water-insoluble fractions were first delignified via alkaline peroxide treatment, the enzymatic hydrolysis was greatly improved, with glucose yields after 48 h of 96% and 83%, for WI-AP-0 and WI-AP-30, respectively. Although the higher quantities of residual lignin in the WI-30 feedstock likely contributed to the observed decrease in hydrolysis, the effect was not severe.  5.2.2.4  Hydrolysis of the water-insoluble component directly in the watersoluble fraction  Based on the good preliminary hydrolysis results at a 2% (w/v) consistency in acetate buffer, hydrolysis of the water-insoluble component directly in the water-soluble fraction was next assessed. In principle, hydrolysis in the water-soluble fraction should be similar to hydrolysis in acetate buffer.  However, in practice, compositional differences between the buffer and the  water-soluble fraction were expected to impact the hydrolysis of the cellulose component. The water-soluble fraction contains numerous chemical compounds {e.g., phenolics, extractives, 141  organic acids and sugar decomposition products) that could potentially inhibit the cellulolytic enzymes employed during hydrolysis. Furthermore, soluble sugars present in the water-soluble fraction could result in increased end-product inhibition of the cellulase and B-glucosidase enzymes. Previously, various researchers reported that removal of the "inhibitor" component in the water-insoluble fraction by fractionation (to remove the water-soluble component), and extensive washing of the solids was necessary for high yield hydrolysis (Mes-Hartree and Saddler, 1983; Schwald et al, 1989; Sinitsyn et al, 1982). By recombining the water-insoluble fraction with the water-soluble fraction, the inhibitors removed from the water-insoluble component during previous washing steps would be effectively replenished. Although this was a significant concern, it was thought that the approach of reduced pretreatment severity would minimize inhibitor generation, and consequently, the relatively low inhibitor content in the water-soluble fraction would not impact the hydrolysis of the cellulose component.  The experimental approach used in these experiments is outlined in Figure 5-5.  Initial  experiments were conducted using only the whitewood substrate (DF-0), which was pretreated using medium-severity conditions. The decision to assess the mixed whitewood/bark feedstock (DF-30) would depend on the success of the preliminary work using the whitewood substrate. The slurry recovered after steam explosion was fractionated and the water-insoluble component was delignified prior to enzymatic hydrolysis. Although it would have been possible to add enzymes directly to the slurry recovered after pretreatment, this approach was not anticipated to yield adequate hydrolysis. As illustrated previously (Figure 5-4), the enzymatic hydrolysis of the non-delignified cellulose component in acetate buffer (2% consistency) resulted in only minimal conversion of the cellulose polymer to glucose. Additionally, the extraneous materials present in the water-soluble component were expected to further reduce the hydrolysis yield. In comparison, the use of the delignified residue (WI-AP-0 and WI-AP-30) resulted in dramatically 142  Douglas-fir wood feedstock (DF-0)  1  Steam Explosion  Water-Soluble fraction (WS-0)  Water-Insoluble fraction (WI-0)  Delignified, Water-Insoluble fraction (WI-AP-0)  Enzymatic Hydrolysis t  Fermentation  Figure 5-5: Process flow diagram for supplementing the water-soluble fractions derived from Douglas-fir whitewood (DF-0) with the delignified cellulose component (WI-AP0) prior to enzymatic hydrolysis and fermentation. The key steps of pretreatment, fractionation, delignification, enzymatic hydrolysis and fermentation are illustrated.  143  improved hydrolysis, and it was further thought that this improvement might partly counter any inhibition effects caused by the presence of inhibitor compounds in the water-soluble fraction.  The preliminary hydrolysis work (Figure 5-4) was conducted at 2% (w/v) consistency, which would produce a theoretical maximum glucose concentration of 19.2 g L" (due to the presence 1  of residual lignin in the WI-AP-0 component). The final hexose sugar concentration following enzymatic hydrolysis of WI-AP-0 in the WS-0 would thus only reach a theoretical maximum of approximately 53 g L ' (neglecting trace amounts of mannose liberated during hydrolysis). To 1  realize a greater increase in sugar concentration, a higher consistency hydrolysis would be required. Hydrolyses were thus conducted at a 5% (w/v) consistency, allowing a theoretical combined hexose sugar concentration of -82 g L" , or slightly more than double the hexose sugar 1  concentration present in the water-soluble component alone.  This calculated value assumes  complete hydrolysis of the cellulose component.  As mentioned earlier, enzymatic hydrolysis in the water-soluble fraction is fundamentally similar to hydrolysis in buffer. However, analysis of the liberated reducing sugars was complicated by the presence of both monomeric and oligomeric sugars derived from the hydrolysis medium itself.  Hydrolysis yield data reported in this chapter are corrected for the "background"  monomeric sugars already present in the water-soluble fraction, by incorporating additional control hydrolyses in the experimental design. Furthermore, because the water-soluble fraction contained oligomeric sugars (upwards of 20% in the Douglas-fir whitewood, water-soluble fraction), it was necessary to compensate for hydrolysis of these sugars by the added enzymes. The hydrolysis of these oligomeric sugars by the addition of the cellulolytic enzymes would otherwise result in higher-than-theoretical glucose concentrations, and the over-estimation of the hydrolysis yield from the cellulose component.  Control hydrolyses containing no water144  insoluble material, but an equivalent enzyme loading (10 F P U Celluclast and 30 IU Novozym188 per gram of cellulose) revealed that monomeric glucose concentration did increase slightly (~1 g L" ) over the course of the hydrolysis due to cleavage of the soluble, oligomeric sugars 1  (Figure 5-6). Glucose was not the only sugar to increase in concentration. Of the other sugars, only mannose concentration increased appreciably. Although the enzymes in Celluclast and Novozym-188 exhibit their greatest hydrolytic activity towards (3-1,4-linked glucosides, they may also exhibit effective hemicellulase activity due to the partial activity towards other glycoside substrates.  Furthermore, the presence of additional hydrolytic enzymes (e.g.,  mannanase, xylanase, etc.) was also likely in the preparations (Appendix 1). It should be noted that the increased concentration of mannose would not influence the calculation of hydrolysis yield, as only the glucose concentration was the used in the determination. Consequently, the results reported in this section for hydrolysis in the water-soluble fraction have been corrected for all monomeric glucose originating from the water-soluble fraction.  Initial attempts to hydrolyse the delignified whitewood cellulose component (WI-AP) at a 5% (w/v) consistency in the fraction WS-0 were marked by a significant reduction in the hydrolysis yield compared to the equivalent control hydrolysis performed in sodium acetate (AC) buffer (Figure 5-7, A). A maximum yield of -65% was realized after 72 h when the enzymatic hydrolysis was preformed in the water-soluble fraction. In contrast, when A C buffer was used for the hydrolysis, greater than 93% conversion of the feedstock was observed after the same time period. The rate of hydrolysis was also significantly slower when performed in the watersoluble fraction (maximum rate of 1.2 g L ' h" compared to 5.9 g L" h" , calculated after 1 h). 1  1  1  1  It should be noted that, although the water-soluble fraction was adjusted to pH 4.8 prior to hydrolysis, it was not buffered with sodium acetate. Consequently, the pH was slightly different 145  45  Figure 5-6: Hydrolysis of oligomeric sugars derived from the water-soluble fraction WS-0 during enzymatic hydrolysis in the water-soluble fraction. Reaction flasks contained no water-insoluble material, but were supplemented with the equivalent enzyme loading. Error bars indicate the range of values for duplicate hydrolyses.  146  - • — A C Buffer - A — S - A C Buffer WS-0  24  36  48  r 72  Time in Hours  -A— A C 3x O  A C 2x  - c ^ A C 1x  W S - 0 3x W S - 0 2x - • — W S - 0 1x -A—  24  36  48  Time in Hours  Figure 5-7: Enzymatic hydrolysis of the delignified, water-insoluble component (WI-AP-0) at 5% (w/v) consistency. (A) Hydrolysis in the water-soluble fraction (WS-0, nonbuffered, pH 4.8), acetate buffer (AC, 50 m M , pH 4.8), or a simulated watersoluble fraction (S-AC, 50 mM, pH 4.8) at 10 F P U g-cellulose. (B) Hydrolysis in the WS-0 or A C buffer at 10, 20 or 30 F P U g-cellulose-1. A l l hydrolyses were supplemented with 6-glucosidase to achieve a 3:1 IU to F P U ratio. Error bars indicate the range of values for duplicate hydrolyses. 147  during hydrolysis in the water-soluble compared to acetate buffer. Whereas the reaction pH in acetate buffer was observed to increase slightly by the end of hydrolysis (from pH 4.8 to 5.1 after 72 h), the hydrolysis pH decreased slightly (to pH 4.6) when the water-soluble fraction was used as the hydrolysis medium. This change in pH was not expected to contribute greatly to the extent of hydrolysis, as both Celluclast and Novozym-188 are stable with high activity over this range (Novozymes Product Specifications Sheet). However, to confirm that buffering the watersoluble fraction would not result in improved hydrolysis, a second hydrolysis was performed in which the fraction WS-0 was supplemented  with crystalline sodium acetate to a final  concentration of 50 m M prior to hydrolysis. Crystalline sodium acetate was used to prevent dilution of the sugars present in the water-soluble fraction. Buffering the sample in this manner had no appreciable impact on the hydrolysis rate or yield (data not shown).  Consequently,  subsequent hydrolyses performed in the water-soluble fraction were not buffered.  In light of the poor hydrolysis of the delignified feedstock WI-AP-0 in the water-soluble fraction, an attempt was made to overcome the observed inhibition and improve the rate and yield of hydrolysis by increasing the enzyme loading. The hydrolysis was repeated, using one of three enzyme loadings: the original loading (lx, 10 FPU g-cellulose" and 30 IU g-cellulose" ), or 1  1  double and triple this loading (Figure 5-7, B). When the enzyme loading was increased, the rate and yield of hydrolysis improved significantly. The maximum rate of glucose production during hydrolysis in the water-soluble fraction (calculated after 1 h) improved from 1.0 g L" h" to 3.7 g 1  1  L" h" as the enzyme loading was increased. When a 3-fold increase in enzyme loading was 1  1  used, the rate of hydrolysis compared closely to the hydrolysis performed in A C buffer at 10 FPU (AC lx), although the glucose yield was lower after 72 h. Although the increased enzyme loading was clearly beneficial, it should be noted that the rate of hydrolysis in A C buffer with these higher loadings was notably faster compared to that observed in the water-soluble fraction. 148  At a 3-fold enzyme loading in A C buffer, the initial rate was found to be 9.8 g L" h' . 1  1  Furthermore, greater than 90% conversion of the substrate was realized within 24 h. In contrast, a yield of only 86% was obtained in WS-0 after 72 h at the same enzyme loading. Given the relatively low productivities during the hydrolyses in the water-soluble fraction even with increased enzyme loading, an investigation of the possible causes of the inhibition was undertaken.  5.2.2.5 Inhibition of enzymatic hydrolysis It was initially suspected that hydrolysis in the water-soluble fraction would be affected by the presence of sugars and potential inhibitors produced or liberated from the feedstock during the pretreatment step.  For example, monomeric glucose is known to reduce the activity of  cellulolytic enzymes by way of end-product inhibition, as discussed in detail previously. This is true for (3-glucosidase, which catalyses the hydrolysis of cellobiose to glucose, but also the "true" cellulase enzymes (e.g., exoglucanase, endoglucanase, and exoglucosidase), which act directly on the cellulose substrate. Both lignin and its chemical constituents (e.g., monoaromatic phenolics such as vanillin, ferulic acid, etc.) have also been implicated in the inhibition of cellulolytic enzymes, as described in more detail in subsequent sections. In contrast, compounds such as H M F and furfural have been shown to be only minimally inhibitory at the concentrations typically found in the water-soluble fraction of steam-exploded wood (Mes-Hartree and Saddler, 1983; Tengborg et al, 2001). The observed reduction in hydrolysis yield by nearly 30% was likely a consequence of the combined influence of sugars and lignin-derived phenolics. However, given the relatively low sugar concentration in the water-soluble fraction (less than 40 g L " total sugars), it was initially suspected that the non-carbohydrate component would be !  found responsible for a sizeable part of the reduction in hydrolysis yield.  149  5.2.2.5 a . Simulated water-soluble fraction To determine the relative influence of the sugars present in the water-soluble fraction on the rate and yield of hydrolysis compared to other compounds present in the water-soluble fraction, a simulated water-soluble fraction consisting of acetate buffer (50 m M , pH 4.8) supplemented with monomeric sugars (arabinose, galactose, glucose, mannose, xylose) to the same concentration as found in the fraction WS-0 was formulated (S-AC buffer). When hydrolysis of the delignified substrate, WI-AP, was performed in this buffer, a greater hydrolysis of the cellulose to glucose was realized, in comparison to the hydrolysis performed in the water-soluble fraction (Figure 57, A). However, the glucose yield after 72 h reached a maximum of only 76%, which was still significantly lower than was achieved in the sugar-free A C buffer (Figure 5-7, A). Thus, despite the low combined sugar concentration in the water-soluble fraction (less than 40 g L" ), and the 1  even lower glucose concentration (less than 15 g L" ), the monomeric carbohydrate component 1  accounted for roughly two thirds of the reduction in hydrolysis yield when the water-soluble fraction was used as hydrolysis medium.  It was not immediately clear whether the presence of supplemental monomeric sugars (glucose, in particular) in the hydrolysis medium resulted in inhibition of the cellulase or 6-glucosidase enzymes. Both groups of enzymes are inhibited by glucose, although 6-glucosidase is generally reported to be more susceptible to inhibition by glucose than are the cellulase enzymes. Complicating the interpretation of the results was the fact that a reduction in 6-glucosidase activity results in an increased cellobiose concentration, which is a potent inhibitor of the cellulase enzymes. In the current work, hydrolyses were supplemented to a relatively high final 6-glucosidase activity (3:1 IU to F P U ratio). However, in the presence of supplemental sugars, the 6-glucosidase activity would likely suffer from increased inhibition. Accordingly, it was  150  thought likely that increasing the (3-glucosidase loading would alleviate much of the observed inhibition of hydrolysis. Furthermore, by establishing a (3-glucosidase loading that was nonlimiting to the hydrolysis, it should be possible to qualitatively evaluate the extent of inhibition of the cellulase enzymes. For example, if after adding more (3-glucosidase, hydrolysis yield was improved to the same level as the control hydrolysis in A C buffer, then (3-glucosidase alone was subject to inhibition by the supplemental sugars in the S-AC buffer.  Alternatively, if the  hydrolysis yield did not improve to the same level as the control hydrolysis in A C buffer with additional (3-glucosidase loading, then inhibition of the cellulase enzymes should account for the difference compared to the control. This should be true, provided protein-protein interactions at the higher (3-glucosidase loadings did not play a significant role in limiting the hydrolysis, and cellulase activity (i.e., filter paper activity) in the (3-glucosidase preparation was insufficient to significantly improve the hydrolysis.  The hydrolysis reactions were performed over a range of supplemented (3-glucosidase enzyme concentrations. To establish the base-case, hydrolysis was performed using no supplemental (3glucosidase. Hydrolyses were also performed with a (3-glucosidase (Novozym-188) loading at a 3:1, 6:1 and 9:1 IU:FPU ratio. Based on the target Celluclast loading of 10 FPU g-cellulose" in 1  all hydrolyses, this required a (3-glucosidase activity of 30, 60 and 90 IU g-cellulose" for the 1  three ratios, respectively.  Results for the hydrolysis are presented in Figure 5-8.  With no  supplemental (3-glucosidase, the hydrolysis of the delignified water-insoluble component (WIAP-0) was severely hindered.  A low level of native (3-glucosidase activity in the Celluclast  preparation (and, possibly, exoglucosidase activity) accounted for the production of glucose. However, even after ~3 days (76 h), hydrolysis in A C buffer yielded only -53% of the theoretical glucose yield. Hydrolysis yield at this enzyme loading was even lower (32%) when 151  100 -r  •  AC — A — AC • AC —o-S-AC -O—S-AC S-AC  —o—  12  24  36  48  60  72  84  96  108  90 IU 60 IU 30 IU 90 IU 60 IU 30 IU 0 IU  S-AC  120  Time in Hours  Figure 5-8: The influence of B-glucosidase loading (0, 30, 60 or 90 I U g-cellulose ) on the enzymatic hydrolysis of the delignified, water-insoluble component (WI-AP) in S - A C buffer. Celluclast was added to an activity of 10 F P U g-cellulose" in all reactions. Control hydrolyses were conducted in A C buffer. Error bars indicate the range of values for duplicate hydrolyses. -1  1  152  S-AC buffer was used. Hydrolysis at a 30 IU g-cellulose" loading (the (3-glucosidase loading 1  normally used in this research) resulted in a hydrolysis comparable to that reported previously in Figure 5-7, with a yield of -75% after 76 h. At higher (3-glucosidase loadings, the hydrolysis yield was improved, but only modestly. When a (3-glucosidase loading of 60 IU g-cellulose"  1  was used, the hydrolysis yield after approximately the same time period (72 h) was 82%. Higher (3-glucosidase loadings (90 IU g-cellulose" ) did not further increase hydrolysis yield (81% after 1  76 h). The hydrolyses in S-AC buffer all had significantly lower glucose yields than those in A C buffer, both at low Novozym-188 supplementation (93%) and high supplementation (95%). As expected, there was not a substantial difference in hydrolysis yield in A C buffer with different (3glucosidase loadings. Hydrolyses differed by only - 3 % for all but the 14-hour and 24-hour samples. This result confirmed high (3-glucosidase loading did not result in improved hydrolysis of the cellulose, owing to its low filter paper activity (Appendix 1). It is also unlikely that protein-protein interactions played a role during hydrolysis at the high (3-glucosidase loading, based on the comparable hydrolyses. Thus, the results suggest that, at or above a loading of 90 IU g-cellulose" , (3-glucosidase activity was no longer limiting the hydrolysis, and the observed 1  decrease in yield compared to the control hydrolyses in A C buffer was caused by inhibition of the cellulase enzymes by the supplemented monomeric sugars. From these results, it appears that inhibition of the cellulase enzymes actually accounts for a greater reduction in the hydrolysis yield than inhibition of (3-glucosidase. While the presence of sugars accounted for a decrease in glucose yield from 93% to 75% (at 30 IU g-cellulose" loading after 76 h), supplementation with 1  (3-glucosidase to 90 IU g-cellulose" provided a maximum conversion of 81% compared to 95% 1  in A C buffer. Higher loadings of Celluclast enzyme would undoubtedly help to overcome this inhibition (as was shown in Figure 5-7, B), but may not be a viable commercial strategy due to the current high cost of the enzymes.  153  It was apparent that the presence of sugars in the hydrolysis medium resulted in only a decreased rate of hydrolysis. When the hydrolysis at 90 IU g-cellulose" (3-glucosidase loading was 1  allowed to proceed for as long as to 120 h, the hydrolysis yield approached that of the sugar-free hydrolyses (-90%). Performance was slower at the lower enzyme loading of 30 IU g-cellulose" , 1  but reached 84% of the theoretical glucose yield. These results are consistent with the accepted models for inhibition by carbohydrates in that carbohydrates do not result in the irreversible inhibition of the enzymes. It is likely that, given sufficient time, all hydrolyses would have eventually reached yields of 100%. However, performing the hydrolysis for the required length of time (>120 h) would be impractical. For a commercial process, high productivity is often more important that high yield where economics are concerned.  5.2.2.5  b. M o n o m e r i c s u g a r s  Based on the poor hydrolysis results in the presence of sugars (both in S-AC buffer and the water-soluble fraction), the role that individual sugars play in the inhibition was evaluated. Of the carbohydrates derived from steam exploded Douglas-fir whitewood, the water-soluble fraction contains 5 principle neutral sugars (arabinose, galactose, glucose, xylose and mannose). As mentioned previously, glucose is well recognized for its inhibition of cellulose hydrolysis. Other sugars such as xylose have also been implicated in the reduced hydrolysis of lignocellulosics to glucose (Dekker, 1986; Marsden et al., 1983). However, the influence of a particular sugar will largely depend on its concentration, and may not have a pronounced effect at the concentrations normally present in the water-soluble fraction (Mes-Hartree and Saddler, 1983). Of the sugars present in the water-soluble fraction, it was thought likely that mannose would have the greatest impact on the hydrolysis (after glucose) due to its similar structure to glucose, and its relatively high concentration in the water-soluble fraction derived from steam exploded Douglas-fir (-10-15 g L " , depending on pretreatment severity). 1  To evaluate the 154  susceptibility of the enzyme system to sugars, a pulp feedstock was hydrolysed at a 2% (w/v) consistency in buffer supplemented with a single monomeric sugar present in the water-soluble fraction (10, 20 and 40 g L" ). 1  The pulp chosen for hydrolysis was an oxygen-delignified hemlock Kraft pulp, supplied by Howe Sound Pulp and Paper. The decision to use this pulp was based on the limited supply of steam exploded Douglas-fir pulp. The composition of the hemlock pulp was determined (Table 5-8), and as expected, its composition was quite different from the delignified Douglas-fir whitewood feedstock used for previous hydrolysis experiments (WI-AP-0, Table 5-7).  The  hemlock pulp had a significantly lower lignin content (2.4% vs. 9.1%) compared to the delignified Douglas-fir whitewood feedstock.  Due to the decreased solubilization of the  hemicellulose sugars during the pulping process, the hemlock pulp also retained greater mannose and xylose content, but had comparably low arabinose and galactose content (less than 0.5%). Cellulose content was slightly lower in the hemlock pulp (81.0% vs. 86.5%). The low lignin content was expected to contribute to greater ease of hydrolysis. This would mean results obtained in these hydrolyses would not be directly comparable to the enzymatic hydrolysis of fraction WI-AP-0. However, it would illustrate the impact that individual sugars could have on the hydrolysis.  Hydrolyses of the kraft pulp were conducted using a cellulase loading of 10 F P U g-cellulose"  1  supplemented with (3-glucosidase at a ratio of 3:1 IU:FPU (Figure 5-9 and Table 5-9). Each of the monomeric sugars tested caused a reduction in hydrolysis yield. The severity of inhibition was dependent on the particular sugar and its concentration. At all tested concentrations (10, 20 and 40 g L" ), glucose had the most significant impact on hydrolysis. At a glucose concentration 1  155  Table 5 - 8 :  Composition (%) of the ( V d e l i g n i f i e d Hemlock kraft pulp used for selected enzymatic hydrolysis experiments. Values in parentheses indicate standard deviation (n = 3). Carbohydrate  a  b  Lignin  Arabinan  Galactan  Glucan  Mannan  Xylan  AIL  0.5 (0.1)  0.2(0.1)  81.0(0.3)  6.8 (0.1)  7.2 (0.1)  2.4(0.1)  a  ASL  b  0.4(0.0)  Total 98.5 (0.3)  A I L : A c i d insoluble lignin A S L : A c i d soluble lignin  156  12  24  36  48  Time in Hours  24  Time in Hours  Figure 5-9: Enzymatic hydrolysis of the Hemlock kraft pulp (2% w/v consistency) in acetate buffer supplemented with individual monomeric sugars (arabinose, galactose, glucose, mannose or xylose) to achieve a concentration of (A) 10 g L" , (B) 20 g L" , or (C) 40 g L" at the start of hydrolysis. Error bars denote range of duplicate values, except for A C , where they indicate standard deviation of duplicates run on 2 separate occasions. 1  1  1  157  Table 5-9: Reduction in the hydrolysis yield after 24 h (%, relative to control hydrolyses performed in A C buffer) during hydrolysis of the 0 -delignified Hemlock kraft pulp in A C buffer supplemented with individual monomeric sugars (arabinose, galactose, glucose, mannose or xylose). Values in parentheses indicate range of values for duplicate hydrolyses. 2  Supplemented sugar concentration (g L" ) 1  Sugar Added Galactose Arabinose Xylose Mannose Glucose  10 3.8 (1.6) 4.9 (1.4) 8.5 (0.8) 10.8 (1.1) 25.1 (0.8)  20 7.0 (0.7) 8.7 (1.5) 15.0 (1.8) 17.5 (0.8) 44.9(1.3)  40 18.0 (1.4) 19.0 (1.5) 26.6 (0.8) 27.5 (0.9) 71.6 (2.4)  158  of 10 g L" , a hydrolysis yield of 64.2% was realized after 24 h, compared to a yield of 85.8% in 1  the A C buffer control. At 40 g L , only a 24.3% conversion of the pulp to glucose was observed 1  (71.6% reduction compared to the control in A C buffer). Mannose was the next most inhibitory sugar, closely followed by xylose, but both were far less inhibitory than glucose at comparable concentrations. Arabinose and galactose caused the least reduction in hydrolysis yield. Even at the high concentration of 40 g L" (more than 10 times higher than the concentration in the 1  water-soluble fraction WS-0), it was still possible to achieve greater than 80% of the yield achieved in sugar-free A C buffer. At the lowest concentration tested (10 g L" ), the effect of 1  arabinose and galactose was detectable, but relatively insignificant (less than 5% reduction in hydrolysis yield, compared to the hydrolysis in A C buffer). It should be noted that the watersoluble fraction (WS-0) used for hydrolysis in Figure 5-7 had an arabinose and galactose concentration closer to 1.5 g L" and 4 g L" , respectively (Table 5-2). Xylose concentration in 1  1  this water-soluble fraction was also closer to 5 g L" . 1  This meant that the contribution by  arabinose, galactose and xylose towards inhibition would also be less than that reported in Figure 5-9, A . Based on these results, it was evident that only glucose and mannose had a significant effect when the water-soluble fraction WS-0 was used as the hydrolysis medium (e.g., Figure 57).  5.2.2.5  c. Non-carbohydrate components  It is clear that monomeric sugars can have a pronounced effect on the hydrolysis of lignocellulosics.  As illustrated previously in Figure 5-7, carbohydrate-mediated inhibition of  cellulase and (3-glucosidase activity resulted in the sizeable reduction in the extent of hydrolysis (18.4%) when S-AC buffer was used in lieu of A C buffer.  However, a further reduction in  hydrolysis yield (11.6%) was observed when the water-soluble fraction was used as hydrolysis medium. It was evident that the inhibition caused by the non-carbohydrate component resulted 159  in a decreased rate of hydrolysis. However, even long term hydrolyses (up to 6 days) resulted in only a minor improvement to the hydrolysis yield compared to 72-hour hydrolyses (Figure 5-10). In contrast, sugar inhibition during hydrolysis in S-AC buffer could be largely overcome by increasing the duration of hydrolysis. This suggested an additional mechanism of inhibition, capable of permanently reducing enzyme activity. As postulated previously, this decrease was thought to be caused by phenolic compounds (Mandels and Reese, 1963; Martin and Blake, 1989) derived from lignin, extractives and possibly carbohydrate  decomposition and  condensation reactions (Lai, 1991). These compounds have been implicated in the inhibition of numerous enzymes, including hydrolytic enzymes, through reversible and irreversible binding, precipitation and deactivation of the enzymes (Jung and Fahey, 1983).  To evaluate this apparent decrease in hydrolytic potential, the enzyme preparation (Celluclast supplemented with Novozym-188) was incubated in the water-soluble fraction (WS-0, pH 4.8), A C buffer, or S-AC buffer without pulp. Enzyme activities in these mixtures were determined by sampling the mixtures periodically, and using this sample to hydrolyse a hemlock kraft pulp prepared in A C buffer (2% w/v consistency).  Hydrolysis yields were greatest after pre-  incubation in A C buffer (Figure 5-11, A). The enzyme activities following pre-incubation in WS-0 and S-AC were consistently lower than the control in A C buffer.  This systemic decrease  was likely caused by the carry-over of both sugars and inhibitors from the pre-incubation medium to the hydrolysis reaction.  Furthermore, the activity of the pre-incubated enzymes  decreased with greater incubation time regardless of the medium used (Figure 5-11, A). This decrease was likely a result of the enzymes' susceptibility to prolonged exposure to temperature and sheer effects during incubation. To differentiate between this "background" loss of activity over time, and any losses caused by components in the water-soluble fraction (e.g., sugars and inhibitors), the data were normalized relative to the A C buffer hydrolysis (Figure 5-11, B). It 160  1009080 H  Time in Hours Figure 5-10:  The effect of extended hydrolysis time on the glucose yield from hemlock pulp enzymatically hydrolysed in the water-soluble fraction (WS-0). Error bars denote range for hydrolyses performed in duplicate.  161  Figure 5-11: Decrease in the hydrolytic power of the enzyme preparation (Celluclast and Novozym-188) after prolonged pre-incubation in the water-soluble fraction (WS-0), acetate buffer (AC), or a simulated water-soluble fraction (S-AC) prior to hydrolysis of the Hemlock kraft pulp in A C buffer. (A) Hydrolysis yield. (B) Hydrolysis yield expressed as a percentage of the hydrolysis in A C buffer. A l l hydrolyses were performed in duplicate; error bars denote range of values.  162  was clear that the decrease in hydrolysis yield over time was more significant when the enzymes were pre-incubated in the water-soluble component. With only limited exposure of the enzymes to the water-soluble fraction (less than 2 minutes), a 2-hour hydrolysis in the water-soluble fraction attained only 43.0% of the yield in the A C buffer.  Although the enzymes appeared  initially quite stable in the water-soluble faction, hydrolysis yields did decrease with longer initial exposure to the water-soluble fraction, and after 50 h pre-incubation, the equivalent hydrolysis attained only 21.9% of the yield in A C buffer. In contrast, pre-incubation in S-AC buffer did not result in a significant change in activity over the 50-hour pre-incubation period.  5.2.2.6  Detoxification of the water-soluble fraction  Prolonged exposure of the enzyme preparation to the non-carbohydrate components (e.g., phenolics, extractives, etc.) in the water-soluble fraction resulted in an apparent loss of enzyme activity. It was thought that treatment of the water-soluble fraction to remove these inhibitors prior to hydrolysis would promote improved digestibility of the substrate, although not all inhibition would be removed (i.e., inhibition due to carbohydrates would remain). This approach could help maintain enzyme activity, and would be beneficial when enzyme recycle is considered. Various detoxification steps have been described for the treatment of lignocellulosederived hydrolysates, in an effort to reduce inhibition of microbial metabolism and ensure efficient ethanol production from otherwise toxic lignocellulosic hydrolysates.  For example,  treatment with activated carbon/charcoal (Lee et al., 1999), ion exchange resins (Clark and Mackie, 1984), fungal culture or specific fungal enzymes (e.g., laccase, peroxidase) (Jonsson et al, 1998), solvent extraction (Frazer and McCaskey, 1989; Wilson et ai, 1989), overliming (Leonard and Hajny, 1945; Ranatunga et ai, 2000), and evaporation/steam-stripping (Larsson et al, 1999b) have been investigated.  Depending on the particular mechanism, treatment may  result in the modification of problem compounds to less inhibitory chemicals, or removal of the 163  responsible compounds entirely. Based on reported improvements of hydrolysate fermentation following detoxification, it was thought that detoxification could likewise benefit the enzymatic hydrolysis step.  In the current work, several detoxification strategies were tested: treatment with anion exchange resin (ANEX), overliming with calcium hydroxide (OL), extraction with ethyl acetate (EX), and combined overliming and extraction (OLEX). After treatment, the water-soluble fractions were corrected for any loss of sugar due to the treatment, if required, and hydrolysis of the hemlock pulp (2% w/v consistency) was conducted in the newly detoxified samples.  Anion exchange resins have proved suitable for the detoxification of the water-soluble fraction derived from various wood species prior to fermentation (Clark and Mackie, 1984; Larsson et al, 1999b; Nilvebrant et al, 2001). Anion exchange resins can effectively remove organic acids, furan aldehydes and phenolic compounds from the water-soluble fraction (Larsson et al, 1999b). Batch treatment of the Douglas-fir water-soluble fraction with anion exchange resin (AG1-X8) at pH 10 resulted in the substantial removal of phenolic compounds.  However, the A N E X  treatment also resulted in the significant removal of sugars from the water-soluble fraction (greater than 90%). This was certainly detrimental, as the goal of performing the hydrolysis directly in the water-soluble fraction was to increase sugar concentration. Previously, the use of this resin for detoxification of spruce hydrolysates resulted in only a 5-26% reduction in available sugars (Larsson et al, 1999b). It would be possible to partly counter the loss of glucose through the addition of sulphate ions, which have a greater affinity for the resin (Nilvebrant et al, 2001). Despite the significant sugar losses, the ANEX-treated water-soluble fraction was screened for improvement in hydrolysis, after first supplementing the stream with sugars to achieve the original sugar concentration. Although it was not expected to be practical 164  (due to cost and the potential loss of sugars), it was thought that this method would at least illustrate the best-case scenario, due to the high inhibitor removal.  Despite effective removal of potential inhibitors, hydrolysis was actually hampered by the proposed detoxification process. While the original water-soluble fraction permitted hydrolysis yields of 58% after 48 h, the ANEX-treated hydrolysate permitted a yield of only 42%. The decrease in hydrolysis yield was thought to be caused by the liberation of an inhibitory component from the resin. Following treatment, a strong ammonia-like smell was detected in the treated hydrolysate. As this resin is based on quaternary ammonium ion functional group, possible degradation of the resin might explain the accumulation of ammonia. Repeat attempts at anion exchange detoxification met with similar consequences. Due to the problems associated with the anion exchange resin, its application to detoxification was not explored any further.  Previously, it was shown that solvent extraction of the water-soluble fraction with ethyl acetate can provide effective removal of low molecular weight phenolics, H M F , furfural, and acetic acid. (Clark and Mackie, 1984; Wilson et al, 1989). Due to the relatively low polarity of ethyl acetate, it should also be expected that lipophilic extractives would be at least partly removed. A drawback with using ethyl acetate is that it is partially miscible in water (8.7% at 20°C). Without taking the appropriate steps to remove the residual ethyl acetate, the subsequent hydrolysis was inhibited. Residual ethyl acetate in the water-soluble fraction was removed via an evaporation step prior to hydrolysis. To ensure that the evaporation procedure was sufficient, the S-AC buffer was also put through the extraction/evaporation procedure (S-AC-EX). The overall procedure resulted in only negligible changes in sugar concentration.  165  Although previous work has shown that extraction with ethyl acetate could dramatically improve wood hydrolysate fermentation, in the current work, extraction of the water-soluble fraction with ethyl acetate prior to hydrolysis (WS-O-EX) resulted in only a 7.5% improvement in the glucose yield compared to the untreated fraction, WS-0, after 72 h (from 68.3% to 75.8%) (Figure 5-12). This yield was far lower than observed in the S-AC buffer (91.9%), indicating that not all inhibitory components in the water-soluble fraction were removed by ethyl acetate. This was further shown by supplementing acetate buffer with the extracted compounds (recovered by rotary-evaporation of the ethyl acetate to near-dryness). Essentially no decrease in hydrolysis yield was observed (Figure 5-12). Although it was likely that the volatile inhibitors were lost from the extract during rotary-evaporation, this still suggests that the components removed did not account for the full inhibition effect.  As reported previously, the detoxification of lignocellulosic hydrolysates by alkali treatment is commonly employed prior to fermentation. Overliming is a common form of alkali treatment where the pH is raised to very alkaline conditions (pH 9-11) with calcium hydroxide (lime). It has been shown to partially remove a range of inhibitors, including H M F , furfural and phenolics, although anion-exchange resins in general can provide greater removal of these compounds (Larsson et al, 1999b; Ranatunga et al, 2000). Unlike the anion exchange resin, treatment of the Douglas-fir water-soluble fraction by overliming had a less detrimental impact on the soluble sugar concentration. However, approximately 25% of the soluble sugars were degraded during treatment, likely a result of alkali-catalyzed enolization and 6-elimination (Sjostrom, 1993). The use of lower temperatures should alleviate sugar losses, although with potentially different detoxification effects as well (Nilvebrant et al, 2003). Following treatment, the glucose yield increased to 70.3% after 72 h (Figure 5-13). In combination with ethyl acetate extraction, which had previously provided a modest improvement to the cellulose conversion, the yield improved 166  Figure 5-12: The influence of ethyl acetate extraction on the enzymatic hydrolysis of the hemlock kraft pulp in the water-soluble fraction. Hydrolyses were performed in the untreated water-soluble fraction (WS-0), the extracted water-soluble fraction (WS-O-EX), sugar-acetate buffer (S-AC), extracted sugar-acetate buffer (S-ACEX), acetate buffer (AC) and acetate buffer supplemented with the ethyl acetate extracted components (AC+I). Hydrolyses were performed in duplicate; error bars denote range of values.  167  Figure 5-13:  The influence of overliming and combined overliming/extraction on the enzymatic hydrolysis of the hemlock kraft pulp in the water-soluble fraction. Hydrolyses were performed in the untreated water-soluble fraction (WS-0), the overlimed water-soluble fraction (OL), the overlimed/extracted water-soluble fraction (OLEX), sugar-acetate buffer (S-AC), and acetate buffer (AC). Hydrolyses were performed in duplicate; error bars denote range of values.  168  to 75.5%. However, the combination treatment offered no noticeable advantage compared to extraction only (Figure 5-12).  Furthermore, the yield was still much lower than could be  achieved in S-AC buffer (91.0% hydrolysis yield), suggesting that inhibitors were still present in the water-soluble fraction.  Based on the current results, none of the detoxification strategies resulted in significantly improved enzymatic hydrolysis. One approach reported by Tengborg et al, (2001) was shown to provide near-complete alleviation of inhibition.  These researchers demonstrated that  fermenting a water-soluble hydrolysate derived from spruce prior to performing the enzymatic hydrolysis in the water-insoluble component improved hydrolysis yields significantly, due to the metabolism of carbohydrates, and presumably, presumably some of the non-carbohydrate components. Ethanol produced during the fermentation is far less inhibitory to the enzymes than are carbohydrates (Holtzapple et al, 1990).  However, this particular approach would not  provide any increase in the sugar concentration in the water-soluble fraction, and furthermore, would involve 2 fermentation steps.  Consequently, it was not explored in the current work.  Rather, an alternative approach to improving the sugar concentration of the water-soluble fraction was investigated.  5.2.3  Supplementation hydrolysate  of  the  water-soluble  fraction  with  the  cellulose  Based on the poor enzymatic hydrolysis results when the water-soluble fraction was used as the hydrolysis medium, it was apparent that this approach to increasing the sugars in the watersoluble fraction would not likely be possible. Inhibition caused by both sugars and inhibitors in the water-soluble hydrolysate resulted in markedly reduced rates and yields for hydrolysis. However, it might still be possible to use the sugars derived from the water-insoluble fraction to  169  increase the sugars in the water-soluble fraction prior to fermentation by first hydrolysing the steam exploded wood and then combining the cellulose hydrolysate and the water-soluble fraction.  The experimental strategy is described in Figure 5-14.  Both of the delignified  Douglas-fir whitewood and whitewood/bark water-insoluble components (WI-AP-0 and WI-AP30) were used in the current experiments. These fractions were enzymatically hydrolysed in acetate buffer, and the resulting liquid streams were combined with the water-soluble fraction, with or without prior concentration by rotary-evaporation (CH-0 with WS-0 and RV-0; CH-30 with WS-30 and RV-30). The subsequent mixtures were fermented to ethanol using the Tembec yeast.  5.2.3.1 Hydrolysis of the water-insoluble component at 5% consistency Both of the delignified Douglas-fir water-insoluble components were previously shown to be readily hydrolysable at 2% (w/v) consistency (Figure 5-4). Based on these preliminary results, hydrolysis experiments were repeated at 5% consistency, to obtain a greater final sugar concentration.  At the higher consistency, it remained possible to achieve relatively high  conversion of cellulose to glucose (Figure 5-15). A glucose yield of 84.3% with the whitewoodonly pulp (WI-AP-0) and 79.6% with the whitewood/bark pulp (WI-AP-30) was possible after 72 h hydrolysis. This translated into a final glucose concentration of 40.5 g L" and 34.4 g L" for 1  1  the respective feedstocks. The decrease in glucose yield in the higher consistency hydrolysis compared to the results obtained at 2% (w/v) consistency was likely related to end-product inhibition caused by the greater quantities of liberated glucose during hydrolysis. Glucose is recognized to be a potent inhibitor of cellulolytic enzymes, as was shown in the previous section, and during the current hydrolysis was roughly double the concentration found in the 2% consistency hydrolysis for most time points. In general, end-product inhibition affects only the rate of hydrolysis. Thus, had a longer time been allocated to the hydrolysis, it is possible that a 170  Douglas-fir wood and bark feedstocks (DF-0, DF-30)  1  Steam Explosion  Water-Soluble fraction (WS-0, WS-30)  Concentration  Water-Insoluble fraction (WI-0, WI-30)  Delignification Delignified, Water-Insoluble fraction (WI-AP-0, WI-AP-30)  Concentrated and original water-soluble fractions (RV-0, RV-30; WS-0, WS-30)  Enzymatic Hydrolysis Cellulose Hydrolysate (CH-0,CH-30)  Fermentation  Figure 5-14: Process flow diagram for supplementing the water-soluble fractions derived from Douglas-fir whitewood (DF-0) and mixed whitewood/bark (DF-30) (with and without concentration by rotary-evaporation) with the cellulose hydrolysate (CH) derived from the delignified water-insoluble component (WI-AP).  171  Figure 5-15: Enzymatic hydrolysis (5% consistency in A C buffer) of the alkaline peroxide (AP) treated water-insoluble component derived from Douglas-fir whitewood (WI-AP-0) and whitewood/bark feedstock (WI-AP-30) pretreated under medium-severity conditions. Closed symbols, glucose yield; Open symbols, glucose concentration. Error bars denote standard deviation (n = 3).  172  greater conversion of the cellulose to glucose would have been realized.  However, little  improvement was noted by extending the reaction time from 48 to 72 h, so it is unlikely that the cellulose would be completely hydrolysed without a significant increase in hydrolysis time. In addition to end-product inhibition, enzyme deactivation due to prolonged incubation at temperature and sheer forces due to mixing may partly account for the less than complete hydrolysis. Furthermore, irreversible binding of the cellulase enzymes to the lignin component, of which there was an increased quantity due to the higher consistency, can result in decreased hydrolysis (Ooshima et al, 1990; Sutcliffe and Saddler, 1986).  5.2.3.2 Fermentation of the cellulose hydrolysate Once it was established that both feedstocks could be hydrolysed, albeit to a lower glucose yield than originally anticipated, fermentations of the resulting sugar streams were conducted to ensure that the cellulose hydrolysates were fermentable (Figure 5-16).  To eliminate the  influence of starting sugar concentration on the fermentation, the enzymatic hydrolysate derived from WI-AP-30 (CH-30) was supplemented with glucose to increase the sugar concentration to that found in the whitewood-only hydrolysate. Both cellulose hydrolysates (CH-0 and CH-30) were rapidly fermented, with all of the glucose utilized within 6 h. Neither the presence of barkderived components nor insoluble wood and bark affected the utilization of the carbohydrate derived from cellulose hydrolysates.  Ethanol yields for the fermentation of the cellulosic  hydrolysates were 0.48 g g" and 0.47 g g" for CH-0 and CH-30, respectively, with ethanol 1  1  concentrations of 18.8 g L" and 17.9 g L" . 1  1  5.2.3.3 Combining the two streams Having established that the cellulose hydrolysates were readily fermentable, the impact of supplementing the water-soluble fraction with the cellulose hydrolysate was investigated. Initial  173  A - CH-30S  0  2  4  6  8  10  Time in Hours  Figure 5-16: Fermentation of the cellulose hydrolysates, CH-0 and CH-30, derived from the enzymatically hydrolysed whitewood (WI-AP-0) and whitewood/bark waterinsoluble components (WI-AP-30), respectively. Monomeric sugars were supplemented to CH-30 to achieve a comparable sugar concentration to CH-0. Solid symbols, hexose sugars; Open symbols, ethanol. Error bars indicate range of values for duplicate fermentations.  174  experiments utilized a 1:1 mixture of the water-soluble fraction and the cellulose hydrolysate. Hydrolysis was initially conducted at 5% consistency, because problems were anticipated with hydrolysis performed at higher solids consistency.  However, it was recognized that sugar  concentration in the cellulose hydrolysate would be comparable to that of the water-soluble fraction.  Consequently, only a relatively minor increase in the sugar concentration for the  combined stream would be realized.  The supplementation of the water-soluble fraction (WS-0) with the cellulose hydrolysate (CH-0) resulted in vastly improved rates of fermentation compared to the original water-soluble fraction, both in terms of sugar consumption and ethanol production (Figure 5-17, A). This improvement can be attributed, in large part, to the effective dilution of inhibitory compounds present in the original water-soluble fraction by the addition of the cellulose hydrolysate. In this fermentation, glucose and mannose were completely consumed by hour 6. In contrast, glucose and mannose were still present in the original water-soluble fraction in the 12-hour sample. Although ethanol productivity improved significantly in the mixture (CH-0 plus WS-0) compared to the original water-soluble fraction, the final ethanol concentration did not increase substantially (16.4 g L"  1  vs. 14.9 g L" ), owing to the limited increase in the initial sugar concentration (Table 5-10). 1  Furthermore, ethanol yield was largely unaffected. To improve the starting sugar concentration, the water-soluble fraction was first concentrated 2-fold by rotary-evaporation prior to supplementation with the cellulose hydrolysate. Whereas previously the 2-fold concentrates of WS-0 were only poorly fermented (Figure 5-2), dilution of the inhibitors in the mixture of RV-0 and CH-0 permitted sugar consumption and ethanol production patterns comparable to that in the original water-soluble fraction. Nearly all of the available hexose sugars were consumed within 24 h (residual galactose concentration 3.2 g L" ), and a final ethanol concentration of 23.4 g L" 1  1  was obtained for a 57% increase in ethanol concentration compared to WS-0). 175  -*— WS-0 and CH-0 -•— RV-0 and CH-0 -^-WS-0  -m— WS-30 and CH-30 - • - RV-30 and CH-30 WS-30  24  36  48  Time in Hours  Figure 5-17:  Fermentation of the water-soluble fractions after supplementation with the cellulose hydrolysate. (A) Whitewood-derived water-soluble fraction and cellulose hydrolysate. (B) Whitewood/bark-derived water-soluble fraction and cellulose hydrolysate. A l l bark-derived hydrolysates were supplemented with monomeric sugars to be comparable to the equivalent whitewood hydrolysate. Closed symbols, hexose sugars; Open symbols, ethanol. Fermentations were run in duplicate; error bars denote range of values.  176  Table 5-10: Summary of the fermentations of the original water-soluble fractions (WS-0 and WS-30) and 2-fold concentrates (RV-0 and RV-30) with and without supplementation with the cellulose hydrolysate (CH-0 and CH-30). A l l barkderived hydrolysates were supplemented with monomeric sugars to be comparable to the equivalent whitewood hydrolysate. Values in parentheses denote the range of duplicate fermentations. Water-soluble Fraction  WS-0  Hexose Concentration t = 0h (gL" ) 33.0 (0.3)  Ethanol Concentration t = 48h (gL" ) 14.9 (0.2)  WS-0 and CH-0 RV-0 and CH-0  37.7 (0.3) 54.3 (0.4)  16.4(0.2) 23.4 (0.1)  0.44 (0.01) 0.43 (0.01)  WS-30  33.5 (0.3)  15.0 (0.2)  0.45 (0.01)  WS-30 and CH-30  36.2 (0.3)  16.6(0.1)  0.46 (0.01)  23.8 (0.2) 53.1 (0.6) Ethanol yield calculated on available hexose sugars at t=0 h  0.45 (0.01)  1  RV-30 and CH-30  a  Ethanol Yield (g g" ) a  1  1  0.45 (0.01)  177  Based on the previous results for fermentation of both WS-30 and the 2-fold concentrate RV-30 (Figure 5-2, B), combining either of these fractions with the cellulose hydrolysate CH-30 was not expected to cause problems for the fermentation.  As anticipated, all hydrolysates were  readily fermentable (Figure 5-17, B). However, supplementation of the water-soluble fraction WS-30 with the cellulose hydrolysate CH-30 did not result in the same degree of improvement in fermentation noted for the whitewood stream. This was likely related to the relatively low toxicity of the water-soluble fraction WS-30, as has been discussed previously. Although inhibitors present in the water-soluble fraction were effectively diluted in half, their impact on the yeast metabolism at full strength (WS-30) was relatively insignificant, and dilution offered no clear advantage.  The results of the preliminary investigation into supplementing the water-soluble fraction with the cellulose hydrolysate were very promising. However, the full potential was not realized due to the chosen hydrolysis consistency (5% w/v). This restricted the sugar concentration to less that 50 g L" , due to the presence of lignin in the feedstock and incomplete hydrolysis. As the 1  goal of supplementation was to obtain significantly higher sugar and ethanol concentrations, the experiments were repeated using a cellulose hydrolysate derived from the enzymatic hydrolysis of the delignified, water-insoluble fraction (WI-AP-0) at 10% (w/v) consistency.  At this  consistency, hydrolysis of the cellulose component was incomplete, but yielded a stream with a glucose concentration of 75.0 g L" . This stream was also readily fermentable with virtually no 1  glucose detected after 6 h fermentation (Figure 5-18). Combining this stream 1:1 with the watersoluble fraction WS-0 resulted in a mixture with much improved sugar concentration compared to the original WS-fraction. Likewise, ethanol production was good, and a final concentration of 23.8 ± 0.1 g L" was attained after 48 h, for a yield of 0.46 ± 0.01 g g" . However, hexose sugar 1  1  concentration only reached a maximum of 51.7 ± 0.5 g L" at the onset of fermentation. While 1  178  Figure 5-18: Fermentation of the water-soluble fraction (WS-0), the cellulose hydrolysate (CH-0) obtained from a 10% (w/v) consistency hydrolysis, and the watersoluble fraction after supplementation with the cellulose hydrolysate. Fermentations were run in duplicate; error bars denote range of values.  179  this represented a 56% increase in the hexose sugar concentration compared to the original water-soluble fraction, it still remained far lower than typical commercial ethanologenic processes. To achieve a final hexose concentration of 100 g L" in the mixture (which would 1  provide a final theoretical ethanol concentration of approximately 5%) would require the hydrolysis of WI-AP-0 to be performed at almost 22% (w/v) consistency (assuming a comparable glucose yield of -78% during hydrolysis). This was not a possibility at bench scale in shake flasks. Although commercial hydrolysis at this consistency should be theoretically possible, it can be expected to result in its own range of process issues.  5.3 Conclusions Efforts to increase the starting sugar concentration in the water-soluble fraction by physical concentration so that subsequently higher ethanol yields could be obtained were largely unsuccessful. Although it was possible to increase the sugar concentration in the water-soluble fractions by either rotary-evaporation or freeze-drying, the resulting whitewood concentrate (RV-0 and FD-0) were only poorly fermented to ethanol at as low as a 2-fold increase in concentration. Fermentations of the bark-derived concentrates RV-30 and FD-30 were notably improved, but beyond a 2.5-fold increase in the sugar concentration, the productivity suffered significantly due to a pronounced lag in sugar metabolism. It was apparent that the current approach would not be amenable to bioconversion, due to the accumulation of inhibitory materials present in the water-soluble fractions.  The alternative approach of supplementing the water-soluble fraction with the water-insoluble cellulose component was also not as positive as was hoped. The reduction in glucose yield following hydrolysis of the cellulose component in the water-soluble fraction was largely a consequence of end-product inhibition caused by the monomeric sugars (primarily glucose, but 180  also mannose and xylose) present in the water-soluble fraction. Given the numerous cellulolytic enzymes available for hydrolysis, it is possible that alternative enzymes might exhibit decreased sensitivity to end-product inhibition, and thus could permit improved hydrolysis. However, in addition to the influence of the sugars on hydrolysis, a non-carbohydrate component present in the water-soluble fraction also resulted in reduced glucose yields. Ideally, the selected enzymes would also need to exhibit reduced sensitivity to these as-yet undetermined components, as attempts to overcome this inhibition by treatment (e.g., overliming, extraction) only met with partial success.  Even after treatment, the hydrolysis was incomplete (<75% of theoretical  glucose yield after 72 h).  A possible means of overcoming the problems associated with both inhibition by the carbohydrate and non-carbohydrate fraction might be to perform the hydrolysis and fermentation at the same time, i.e., simultaneous saccharification and fermentation (SSF). Sugars present in the water-soluble fraction, in addition to those liberated from the water-insoluble component during hydrolysis would be removed, limiting the effects of end-product inhibition. Likewise, many of the potential non-carbohydrate inhibitors would likely be metabolized by the growing yeast culture, alleviating inhibition. While SSF may address these concerns, it is important to recognize that the approach will also introduce several compromises and complications, which would require careful scrutiny prior to adoption.  The only truly successful means of improving the sugar concentration in the water-soluble fraction, while maintaining fermentability, was by combining it with the readily-fermentable cellulose hydrolysate. The mixing of the two process streams resulted in the dilution of the inhibitory compounds and markedly improved fermentation. Increasing the sugar concentration in this manner provided a hexose sugar content at the start of fermentation of 51.7 g L"  1  181  compared to 33.1 g L" in the original water-soluble fraction, when a 10% (w/w) consistency 1  cellulose hydrolysate was employed. This was a slightly lower concentration than was obtained when a 5% consistency cellulose hydrolysate was mixed with RV-0 (54.3 g L" ), but in terms of 1  maximum volumetric ethanol productivity, was approximately 4 times as fast.  A potential  drawback of this approach was the slow utilization of galactose. Although glucose and mannose were exhausted by the 6-hour sample, complete consumption of galactose required in excess of 36 h. This arose at least in part due to the dilution of the sugars present in the water-soluble fraction, by supplementation with CH-0. This effect would likely also have an impact on the potential fermentation of arabinose and xylose by recombinant organisms, but could be addressed through the use of organisms exhibiting reduced catabolite repression.  182  Chapter 6: Conclusions and Future Work 6.1 General Conclusions As mentioned previously, the principle goal of the current research project was to address pretreatment of Douglas-fir whitewood and bark feedstocks, with an objective to demonstrate efficient and effective fermentation of the hemicellulose-derived sugars to high yield ethanol. This goal was realized through an approach of reduced severity steam explosion pretreatment, which resulted in improved recovery of the hemicellulose sugars.  Overall, a high yield of  hemicellulose sugars could be recovered from the feedstock using low-severity pretreatment (87%), of which 85% were in readily fermentable monomeric form. This good yield was likely a consequence of the reduction in acid-catalyzed decomposition reactions, which also contributed to the improved fermentability of both the low- and medium-severity water-soluble fractions. Even when the feedstock contained appreciable quantities of bark (30%), hemicellulose sugar recovery only decreased moderately. Furthermore, fermentation was not adversely impacted by the large quantities of naturally-occurring inhibitors in the bark feedstock.  Although the initial objective was met, it was apparent that the final ethanol concentrations achieved following fermentation were likely too dilute for an economically viable process. Initial attempts to overcome this low sugar content with physical concentration techniques were largely unsuccessful, owing to the accumulation of inhibitors in the concentrates, and the associated decrease in fermentability. Enzymatic hydrolysis of the water-insoluble cellulose component directly in the water-soluble fraction resulted in incomplete hydrolysis. Due to the limited success of this approach as a means to improve sugar concentration, the fermentability of the resulting stream was not evaluated. The alternative approach of combining the cellulose hydrolysate with the water-soluble stream proved a more effective way to increase the initial sugar concentration. This simple method to increase ethanol recovery from the water-soluble 183  fraction offered the additional benefit of improving the fermentation rate without the requirement of a detoxification step.  Over the course of the research, a number of issues pertaining to the chosen bioconversion strategy became apparent. In particular, concerns regarding hemicellulose sugar recovery during fractionation and the need for a separate delignification step subsequent to steam explosion were recognized.  These issues, discussed in greater detail below, reveal the need for continued  research in an effort to advance bioconversion technology in general.  The recovery of hemicellulose sugars from the water-insoluble component during fractionation represents a dilemma for the bioconversion process. Given the initial objective to maximize the recovery of hemicellulose sugars from the feedstock (by way of milder pretreatment conditions) it would clearly be beneficial to also maximize recovery of the soluble sugars during the fractionation step. However, the complete recovery the soluble sugars from the water-insoluble matrix was shown to require copious amounts of water, which would obviously contribute to the dilution of an already dilute water-soluble fraction. This could have clear implications for the fermentation step. To avoid this excessive dilution, only the first wash (accounting for -70% of the soluble sugars) was recovered from the water-insoluble component and fermented in the current work.  However, this approach results in the loss of almost 1/3 of the soluble  hemicellulose sugars, because the sugars remaining in the water-insoluble component following fractionation were lost during the delignification stage.  While it is difficult to ignore this  avoidable loss of sugars, it is at the same time questionable to advocate increased dilution of the water-soluble fraction, when increasing its sugar concentration is arguably integral to the commercial success of the process.  184  The need for a separate delignification step prior to the enzymatic hydrolysis is another significant concern for the conversion of Douglas-fir wood residues, and softwood feedstocks in general. Although the strategy of reduced-severity steam pretreatment was an effective means of improving the recovery of the hemicellulose sugars, the steam explosion process does not result in the removal of appreciable quantities of lignin from the starting feedstock, and furthermore can result in the redistribution of condensed lignin on the fibre surface. Although digestibility of the water-insoluble component improved significantly following the delignification, the protocol is costly, and introduced increased complexity to the bioconversion process. A refinement of the current delignification process or selection of an alternative process may be required. Alternatively, it may be necessary to modify the pretreatment process. Although the use of higher-severity pretreatment has previously been shown to improve digestibility of the waterinsoluble component, this approach would not be a viable option without first recovering the hemicellulose sugars. This might be accomplished by employing a two- or multi-stage steam explosion process, although as mentioned earlier, this approach has its own issues. Another strategy might be to explore the use of pretreatment process  (alone, or in combination with  steam explosion) that directly removes greater quantities of lignin (e.g., acetic acid and ethanol pulping pretreatments), thus allowing improved enzymatic hydrolysis.  6.2 F u t u r e w o r k  The work described in this thesis has attempted to address some issues regarding the bioconversion of Douglas-fir residues, but a number of unanswered questions and issues remain. The following list of experiments describes work that would build on the current results, in an effort to improve the pretreatment and fermentation sub-processes of bioconversion.  185  6.2.1 Improved recovery of soluble sugars during fractionation Despite the potential to recover high yields of hemicellulose sugars from the feedstock, the complete isolation of these sugars from the water-insoluble residue has been shown to require significant quantities of water.  This would contribute to the dilution of sugars, but also the  dilution of inhibitory material in the water-soluble fraction. Commercially, a counter-current extraction process will likely be required, to minimize dilution effects. However, it is possible that fermentation and ethanol production will be hindered by the potential increase in sugar- and lignin-degradation products recovered to the water-soluble fraction. Experiments should address optimization of the fractionation process, and the impact that this may have on fermentability.  6.2.2  Galactose fermentation  During the current work, it became apparent that the utilization of galactose by the Tembec T l yeast was poor at best. Galactose consumption was slow, and also suffered from catabolite repression. Although galactose represents only a small component of the water-soluble fraction, it does represent a potentially fermentable substrate that must be used more effectively. The screening of alternative yeast strains for the improved fermentation of galactose would be beneficial, provided their capacity to ferment other sugars and contend with inhibitors be high as well. Yeasts capable of simultaneous catabolism of galactose in the presence of glucose could offer improved productivity, particularly in the case of the combined cellulose hydrolysate and water-soluble fraction.  6.2.3  Selection of cellulolytic enzymes exhibiting reduced end-product inhibition  Performing the enzymatic hydrolysis directly in the water-soluble fraction offers far greater opportunities to increase the starting sugar concentration in the water-soluble fraction than mixing the two liquid streams. However, it was clear that end-product inhibition of the cellulase 186  enzymes rendered this approach impractical. Although end-product inhibition is common with cellulases, it is possible that other cellulase systems are less sensitive to the effect of soluble glucose and other sugars.  The screening of alternative cellulolytic enzymes that display  decreased end-product inhibition could permit improved cellulose hydrolysis in the watersoluble fraction.  6.2.4 S i m u l t a n e o u s saccharification and fermentation (SSF) Related to end-product inhibition of the cellulase enzymes, an alternative approach for improving the hydrolysis of the cellulose component directly in the water-soluble fraction may be to employ SSF. Due to the continued removal of glucose and mannose from the medium by the growing yeast culture, end-product inhibition of the cellulase enzymes can be effectively eliminated. The application of SSF to the bioconversion of Douglas-fir residues would indirectly address the issue of sugar concentration in the water-soluble fraction.  187  References: Adams G.A. (1965) Complete acid hydrolysis. In: Whistler R.L., and Wolfrom M . L . (eds) Methods in Carbohydrate Chemistry. Academic Press, New York, pp 269-276. Aden A . (2002) Technical and economic feasibility assessment. Presentation at N R E L Enzyme Sugar-Ethanol Platform, Gate 3 Review Meeting, Golden, Colorado, Jan. 30-Feb. 3, 2002. Amartey S., and Jeffries T. (1996) An improvement in Pichia stipitis fermentation of acidhydrolysed hemicellulose achieved by overliming (calcium hydroxide treatment) and strain adaptation. World Journal of Microbiology and Biotechnology 12: 281-283. Ando S., Arai I., Kiyoto K., and Hanai S. (1986) Identification of aromatic monomers in steamexploded poplar and their influences on ethanol fermentation by Saccharomyces cerevisiae. Journal of Fermentation Technology 64: 567-570. Asghari A., Bothast R.J., Doran J.B., and Ingram L . O . (1996) Ethanol production from hemicellulose hydrolysates of agricultural residues using genetically engineered Escherichia coli strain K O l l . Journal of Industrial Microbiology 16: 42-47. Asiegbu F.O., Paterson A . , and Smith J.E. (1996) Inhibition of cellulose saccharification and glycolignin-attacking enzymes of five lignocellulose-degrading fungi by ferulic acid. World Journal of Microbiology & Biotechnology 12: 16-21. Bailey B . K . (1996) Performance of ethanol as a transportation fuel. In: Wyman C E . (ed) Handbook on Bioethanol: Production and Utilization. Taylor & Francis, Washington, pp 3760. Bailey M.J., Biely P., and Poutanen K . (1992) Interlaboratory testing of methods for assay of xylanase activity. Journal of Biotechnology 22: 257-270. Baker A.J. (1973) Effect of lignin on the in-vitro digestibility of wood pulp. Journal of Animal Science 36: 768-771. Banerjee N . , Bhatnagar R., and Viswanathan L . (1981a) Development of resistance in Saccharomyces cerevisiae against inhibitory effects of Browning reaction products. Enzyme and Microbial Technology 3: 24-28. Banerjee N . , Bhatnagar R., and Viswanathan L . (1981b) Inhibition of glycolysis by furfural in Saccharomyces cerevisiae. European Journal of Applied Microbiology and Biotechnology 11: 226-228. Bataillon M . , Rico A . , Sablayrolles J.M., Salmon J.M., and Barre P. (1996) Early thiamin assimilation by yeasts under enological conditions: Impact on alcoholic fermentation kinetics. Journal of Fermentation & Bioengineering 82: 145-150. BeMiller J.N. (1967) Acid-catalyzed hydrolysis of glycosides. Advances in Carbohydrate Chemistry 22: 25-108. 188  Blomgren Y . , Jonsson A.-S., and Wimmerstedt R. (1991) Concentration of acid hydrolyzate by reverse osmosis and mechanical vapor recompression evaporation. Journal of Wood Chemistry and Technology 11: 117-135. Boussaid A., Jarvis J., Gregg D.J., and Saddler J.N. (1997) Optimization of hemicellulose sugar recovery from a steam-exploded softwood (Douglas-fir). In: Overend R.P., and Chornet E. (eds)  Making a business from biomass in energy, environment, chemicals, fibers, and materials. Elsevier Science, New York, pp 873-880. Boussaid A . - L . , Esteghlalian A.R., Gregg D.J., Lee K . H . , and Saddler J.N. (2000) Steam pretreatment of Douglas-fir wood chips. Can conditions for optimum hemicellulose recovery still provide adequate access for efficient enzymatic hydrolysis? Applied Biochemistry and Biotechnology 84-86: 693-705. Boussaid A . , Cai Y . , Robinson J., Gregg D.J., Nguyen Q., and Saddler J.N. (2001) Sugar recovery and fermentability of hemicellulose hydrolysates from steam-exploded softwoods containing bark. Biotechnology Progress 17: 887-892. Boyer L.J., Vega J.L., Klasson K.T., Clausen E.C., and Gaddy J.L. (1992) The effects of furfural on ethanol production by Saccharomyces cerevisiae in batch culture. Biomass and Bioenergy 3: 41-48. Brasch D.J., and Free K . W . (1965) Prehydrolysis-kraft pulping of Pinus radiata grown in New Zealand. TAPPI48: 245-248. Breuil C , Chan M . , Gilbert M . , and Saddler J.N. (1992) Influence of (j-glucosidase on the filter paper activity and hydrolysis of lignocellulosic substrates. Bioresource Technology 39: 139-142. Brownell H.H., and Saddler J.N. (1984) Steam-explosion pretreatment for enzymatic hydrolysis. Sixth Symposium on Biotechnology for Fuels and Chemicals, Gatlinburg, Tennessee, May 1518. Biotechnology & Bioengineering Symposium 14: 55-68. Brownell H . H . (1985) Comparison of pretreatment methods for enhancing the enzymatic hydrolysis of wood. International Symposium on Wood and Pulping Chemistry, Vancouver, B C , Canada, August 26-30. pp. 271. Brownell H . H . , Y u E.K.C., and Saddler J.N. (1986) Steam-explosion pretreatment of wood: effect of chip size, acid, moisture content and pressure drop. Biotechnology and Bioengineering 28:792-801. Brownell H.H., and Saddler J.N. (1987) Steam pretreatment of lignocellulosic material for enhanced enzymatic hydrolysis. Biotechnology and Bioengineering 29: 228-235. Brownell H.H. (1988) Steam pretreatment of wood in relation to enzymatic hydrolysis. Report prepared for: Forintek Canada Corp. Report number 05SR.31926-4-5022. Brownell H.H., Schwald W., Smaridge T., and Saddler J.N. (1988) Steam pretreatment of wood for enzymatic hydrolysis - chemical and physical changes involved. IEA Workshop on the Bioconversion of Lignocellulosics, Ottawa, Canada, June 12-16. pp la. 189  Browning B.L. (1963) The composition and chemical reactions of wood. In: Browning B.L. (ed) The Chemistry of Wood. John Wiley & Sons, New York, pp 57-101. Browning B.L. (1967) Methods of Wood Chemistry. Vol 2. Interscience, New York. Bura R., Mansfield S.D., Saddler J.N., and Bothast R J . (2002) SC»2-catalyzed steam explosion of corn fiber for ethanol production. Applied Biochemistry and Biotechnology 99-100: 59-72. Carlson M . (1987) Regulation of sugar utilization in Saccharomyces Bacteriology 169: 4873-4877.  species. Journal of  Carrasco J.E., Saiz M.C., Navarro A . , Soriano P., Saez F., and Martinez J.M. (1994) Effects of dilute acid and steam explosion pretreatments on the cellulose structure and kinetics of cellulosic fraction hydrolysis by dilute acids in lignocellulosic materials. Applied Biochemistry and Bioengineering 45-46: 23-34. Caulfield D.F., and Moore W.E. (1974) Effect of varying crystallinity of cellulose on enzymatic hydrolysis. Wood Science 6: 375-379. Chang Y.-P., and Mitchel R . L . (1955) Chemical composition of common North American pulpwood barks. TAPPI 38: 315-320. Chum H.L., Johnson D.K., Black S.K., and Overend R.P. (1990) Pretreatment-catalyst effects and the combined severity factor. Applied Biochemistry and Biotechnology 24-25: 1-14. Chung I.S., and Lee Y . Y . (1985) Ethanol fermentation of crude acid hydrolyzate of cellulose using high level yeast inocula. Biotechnology and Bioengineering 27: 308-315. Clark T.A., and Mackie K . L . (1984) Fermentation inhibitors in wood hydrolysates derived from the softwood Pinus radiata. Journal of Chemical Technology and Biotechnology 34B: 101-110. Clark T.A., and Mackie K . L . (1987) Steam explosion of the softwood Pinus radiata with sulphur dioxide addition. I. Process optimisation. Journal of Wood Chemistry and Technology 7: 373403. Clark T.A., Mackie K . L . , Dare P.H., and McDonald A . G . (1989) Steam explosion of the softwood Pinus radiata with sulphur dioxide addition. II. Process characterisation. Journal of Wood Chemistry and Technology 9: 135-166. Clausen E.C., and Gaddy J.L. (1987) The production of chemicals and fuels from municipal solid waste. In: Wise D.L. (ed) Global Bioconversions. V o l 2. C R C Press, Inc., Boca Raton, pp 61-73. COFI (2000) British Columbia Forest Industry Fact Book 2000. Report prepared for: Council of Forest Industries.  190  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. Cowling E.B., and Kirk T.K. (1975) Properties of cellulose and lignocellulose materials as substrates for enzymatic conversion processes. Biotechnology and Bioengineering Symposium, Newton and Natick, Massachusetts, Sept. 8-10. Biotechnology and Bioengineering Symposium 6: 95-123. Cullis I. (2003) The effects of steam explosion parameters on softwood delignification efficiency. M.Sc. Thesis, Department of Wood Science, Faculty of Forestry, University of British Columbia. David C , and Atarhouch T. (1987) Utilization of waste cellulose VIII. Enzymatic hydrolysis of spruce bark by cellulases of Trichoderma viride. Applied Biochemistry and Biotechnology 16: 51-59. Dekker R.F.H. (1986) Kinetic, inhibition, and stability properties of a commercial B - D glucosidase (cellobiase) preparation from Aspergillus niger and its suitability in the hydrolysis of lignocellulose. Biotechnology and Bioengineering 28: 1438-1442. Dekker R.F.H. (1988) Inhibitors of Trichoderma reesei B-glucosidase activity derived from autohydrolysis-exploded Eucalyptus regnans. Applied Microbiology and Biotechnology 29: 593-598. Delgenes J.P., Moletta R., and Navarro J . M . (1996) Effects of lignocellulose degradation products on ethanol fermentations of glucose and xylose by Saccharomyces cerevisiae, Zymomonas mobilis, Pichia stipitis and Candida shehatae. Enzyme and Microbial Technology  19: 220-225. Dellweg H . , Rizzi M . , Methner H . , and Debus D. (1984) Xylose fermentation by yeasts. Biotechnology Letters 6: 395-400. De Wulf O., Thonart P., Gaignage P., Marlier M . , Paris A., and Paquot M . (1987) Bioconversion of Vanillin to Vanillyl Alcohol by Saccharomyces cerevisiae. Biotechnology and Bioengineering Symposium 17: 605-616. Dien B.S., Hespell R.B., Wyckoff H.A., and Bothast R J . (1998) Fermentation of hexose and pentose sugars using a novel ethanologenic Escherichia coli strain. Enzyme and Microbial Technology 23: 366-371. Donaldson L.A., Wong K . K . Y . , and Mackie K . L . (1988) Ultrastructure of steam-exploded wood. Wood Science and Technology 22: 103-114. du Preez J . C , Kock J.L.F., Monteiro A.M.T., and Prior B.A. (1985) The vitamin requirements of Candida shehatae for xylose fermentation. FEMS Microbiology Letters. 28: 271-275. du Preez J . C , Bosch M . , and Prior B.A. (1986) The fermentation of hexose and pentose sugars by Candida shehatae and Pichia stipitis. Applied Microbiology and Biotechnology 23: 228-233. 191  du Preez J.C., Bosch M . , and Prior B.A. (1987) Temperature profiles of growth and ethanol tolerance of the xylose-fermenting yeasts Candida shehatae and Pichia stipitis. Applied Microbiology and Biotechnology 25: 521-525. du Preez J.C., van Driessel B., and Prior B.A. (1989) Ethanol tolerance of Pichia stipitis and Candida shehatae strains in fed-batch cultures at controlled low dissolved oxygen levels. Applied Microbiology and Biotechnology 30: 53-58. Dumsday G.J., Zhou B . , Yaqin W., Stanlery G.A., and Pamment N . B . (1999) Comparative stability of ethanol production by Escherichia coli K O I 1 in batch and chemostat culture. Journal of Industrial Microbiology and Biotechnology 23: 701-708. Eklund R., Galbe M . , and Zacchi G. (1988) Two-stage steam pretreatment of willow for increased pentose yield. Journal of Wood Chemistry and Technology 8: 379-392. Eklund R., Galbe M . , and Zacchi G. (1995) The influence of SO2 and H 2 S O 4 impregnation of willow prior to steam pretreatment. Bioresource Technology 52: 225-229. Environment Canada (2001) "The Kyoto Protocol." (Web Page) http://www.ec.gc.ca/pdb/ghg/kyoto_protocol_e.cfm Last updated on: Jan 25, 2001. Accessed on: Mar 14, 2003. Fan L.T., Lee Y . - H . , and Gharpuray M . M . (1982) The nature of lignocellulosics and their pretreatments for enzymatic hydrolysis. Advances in Biochemical Engineering 23: 157-187. Fan L.T., Gharpuray M . M . , and Lee Y . - H . (1987) Cellulose hydrolysis. Springer-Verlag, Berlin. Fengel D., and Wegener G. (1989) Wood: Chemistry, Ultrastructure, Reactions. Walter de Gruyter, Berlin. Field J.A., Kortekaas S., and Lettinga G. (1989) The tannin theory of methanogenic toxicity. Biological Wastes 29: 241-262. Field J.A., and Lettinga G. (1991) Treatment and detoxification of aqueous spruce bark extracts by Aspergillus niger. Water Science and Technology 24: 127-137. Frazer F.R., and McCaskey T.A. (1989) Wood hydrolyzate treatments for improved fermentation of wood sugars to 2,3-butanediol. Biomass 18: 31-42. Galbe M . , and Zacchi G. (2002) A review of the production of ethanol from softwood. Applied Microbiology and Biotechnology 59: 618-628. Gancedo J.M. (1992) Carbon catabolite repression in yeast. European Journal of Biochemistry 206: 297-313. Ghose T.K. (1987) Measurement of cellulase activities. Pure and Applied Chemistry 59: 257268. 192  Glasser W . G . (1982) Lignin. In: Overend R.P., Milne T.A., and Mudge L . K . (eds) Fundamentals of thermomechanical biomass conversion. Elsevier Applied Science, New York, pp 61-76. Glasser W.G., Barnett C.A., Muller P.C., and Sarkanen K . V . (1983) The chemistry of several novel bioconversion lignins. Journal of Agricultural and Food Chemistry 31: 921-930. Gould J.M. (1985) Studies on the mechanism of alkaline peroxide delignification of agricultural residues. Biotechnology and Bioengineering 27: 225-231. Gregg D.J., and Saddler J.N. (1995) Bioconversion of lignocellulosic residue to ethanol: Process flowsheet development. Biomass and Bioenergy 9: 287-302. 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. Grethlein H.E. (1985) The effect of pore distribution on the rate of enzymatic hydrolysis of cellulosic substrates. Bio/Technology 3: 155-160. Grohmann K., Torget R., and M . H . (1985) Optimization of dilute acid pretreatment of biomass. Biotechnology and Bioengineering Symposium 15: 59-80. Gusakov A . V . , and Sinitsyn A.P. (1992) A theoretical analysis of cellulase product inhibition: effect of cellulase binding constant, enzyme/substrate ratio, and (3-glucosidase activity on the inhibition pattern. Biotechnology and Bioengineering 40: 663-671. Hadj-Taieb N . , Chaabouni-Ellouz S., Kammoun A., and Ellouz R. (1992) Hydrolytic efficiency of Penicillium occitanic cellulase: kinetic aspects. Applied Microbiology and Biotechnology 37: 197-201. Hafizoglu H . , Usta M . , and Bilgin O. (1997) Wood and Bark Composition of Picea (L.) Link. Holzforschung 51: 114-118.  orientalis  Hahn-Hagerdal B., Hallborn J., Jeppsson H . , Olsson L., Skoog K . , and Walfridsson M . (1993) Pentose fermentation to alcohol. In: Saddler J.N. (ed) Bioconversion of Forest and Agricultural Plant Residues. C A B International, Oxon, pp 231-290. Hahn-Hagerdal B., Jeppsson H . , Olsson L., and Mohagheghi A . (1994a) An interlaboratory comparison of the performance of ethanol-producing micro-organisms in a xylose-rich acid hydrolysate. Applied Microbiology and Biotechnology 41: 62-72. Hahn-Hagerdal B., Jeppsson H., Skoog K., and Prior B.A. (1994b) Biochemistry and physiology of xylose fermentation by yeasts. Enzyme and Microbial Technology 16: 933-943. Hathaway J., and Hawkins D. (1999) Hearing on oxygenated fuels before the Heath and Environmental Subcommittee, US House of Representatives, May 6, 1999.  193  Hatton T. (1999) Canada's wood residues: a profile of current surplus and regional concentrations. Report prepared for the National Climate Change Process, Forest Sector Table, by the Canadian Forest Service (Industry, Economics and Programs Branch), Mar. 1999. Haygreen J.G., and Bowyer J.L. (1996) Forest Products and Wood Science: Introduction. 3rd edition. Iowa State University Press, Ames, Iowa.  An  Heredia A., Fernandez-Bolanos J., and Guillen R. (1990) Cellulase inhibition by polyphenols in olive fruits. Food Chemistry 38: 69-73. Hillis W.E. (1982) Wood and biomass ultrastructure. In: Overend R.P., Milne T.A., and Mudge L.K. (eds) Fundamentals of thermomechanical biomass conversion. Elsevier Applied Science, New York, pp 1-33. Ho N.W.Y., Chen Z., and Brainard A.P. (1998) Genetically engineered Saccharomyces yeast capable of effective cofermentation of glucose and xylose. Applied and Environmental Microbiology 64: 1852-1859. Ho N.W.Y., Chen Z., Brainard A.P., and Sedlak M . (1999) Successful design and development of genetically engineered Saccharomyces yeasts for effective cofermentation of glucose and xylose from cellulosic biomass to fuel ethanol. Advances in Biochemical Engineering/Biotechnology 65: 163-192. Holtzapple M . , Cognata M . , Shu Y . , and Hendrickson C. (1990) Inhibition of Trichoderma reesei cellulase by sugar and solvents. Biotechnology and Bioengineering 36: 275-287. Hsu T.-A. (1996) Pretreatment of biomass. In: Wyman C E . (ed) Handbook on Bioethanol: Production and Utilization. Taylor & Francis, Washington, pp 179-212. Ingram L.O. (1986) Microbial tolerance to alcohols: Biotechnology 4: 40-44.  role of the cell membrane. Trends in  Ingram L.O., Conway T., Clark D.P., Sewell G.W., and Preston J.F. (1987) Genetic engineering of ethanol production in Escherichia coli. Applied and Environmental Microbiology 53: 24202425. Jecu L . (1998) Effect of phenolic compounds on the growth of Aspergillus niger and activity of hydrolytic enzymes. Revue Roumaine de Biochimie 34: 99-102. Jeffries T.W. (1983) Utilization of xylose by bacteria, yeasts, and fungi. Advances in Biochemical Engineering/Biotechnology 38: 1-32. Jeffries T.W., and Shi N.-Q. (1999) Genetic engineering for improved xylose fermentation by yeasts. Advances in Biochemical Engineering/Biotechnology 65: 117-161. Jensen W., Fremer K . E . , Sierila P., and Wartiovaara V . (1963) The chemistry of bark. In: Browing B.L. (ed) The Chemistry Of Wood. John Wiley and Sons, Inc., New York, pp 587-666.  194  Jonsson L.J., Palmqvist E., Nilvebrant N.-O., and Hahn-Hagerdal B. (1998) Detoxification of wood hydrolysates with laccase and peroxidase from the white-rot fungus Trametes versicolor. Applied Microbiology and Biotechnology 49: 691-697. Jozsa L.A., and Middleton G.R. (1994) A discussion of wood quality attributes and their practical implications. Report prepared for: Forintek Canada Corp. Special Publication SP-34. Jung H.G., and Fahey G.C. (1983) Nutritional implication of phenolic monomers and lignin: a review. Journal of Animal Science 57: 206-219. Juntheikki M.-R., and Julkunen-Tiitto R. (2000) Inhibition of B-glucosidase and esterase by tannins from Betula, Salix, and Pinus species. Journal of Chemical Ecology 26: 1151-1165. Kadla J.F., Chang H., and Jameel H . (1999) The reaction of lignins with hydrogen peroxide at high temperature. Part II. The oxidation of Kraft lignin. Holzforschung 53: 277-284. Keller F.A., Bates D., Ruiz R., and Nguyen Q. (1998) Yeast adaptation on softwood prehydrolysate. Applied Biochemistry and Biotechnology 70-72: 137-148. Kim K . H . , Tucker M.P., Keller F.A., Aden A . , and Nguyen Q.A. (2001) Continuous countercurrent extraction of hemicellulose from pretreated wood residues. Applied Biochemistry & Biotechnology 91-93: 253-267. Kim K . H . , Tucker M.P., and Nguyen Q.A. (2002) Effects of operating parameters on countercurrent extraction of hemicellulosic sugars from pretreated softwood. Applied Biochemistry & Biotechnology 100: 147-159. King K . W . (1966) Enzymic degradation of crystalline hydrocellulose. Biochemical Biophysical Research Communications. 24: 295-298.  and  Kniel L., Winter O., and Stork K . (1980) Ethylene, keystone to the petrochemical industry. Vol 2. Marcel Dekker, Inc., New York. Kurth E.F. (1949) The chemical analysis of western woods. Part II: Douglas fir bark analysis. TAPPI 32: 175-176. Kwong B.W. (2000) The examination of cell recycle batch fermentation of softwood hexose sugars to ethanol. Bachelor's of Applied Science Thesis, Department Chemical and Bio-resource Engineering, The University of British Columbia. Lai Y . - Z . (1991) Chemical degradation. In: Hon D.N.-S., and Shiraishi N . (eds) Wood and Cellulose Chemistry. Marcel Dekker, Inc., New York, pp 455-523. Larsson S., Palmqvist E., Hahn-Hagerdal B., Tengborg C , Stenberg K . , Zacchi G., and Nilvebrant N.-O. (1999a) The generation of fermentation inhibitors during dilute acid hydrolysis of softwood. Enzyme and Microbial Technology 24: 151-159.  195  Larsson S., Reimann A . , Nilvebrant N.-O., and Jonsson L.J. (1999b) Comparison of different methods for the detoxification of lignocellulose hydrolyzates of spruce. Applied Biochemistry and Biotechnology 77-79: 91-103. Larsson S., Quintana-Sainz A . , Reimann A . , Nilvebrant N.-O., and Jonsson L.J. (2000) The influence of lignocellulose-derived aromatic compounds on oxygen-limited growth and ethanolic fermentation by Saccharomyces cerevisiae. Applied Biochemistry and Biotechnology 84-86: 617-632.  Lee W.G., Lee J.S., Shin C.S., Park S . C , Chang H . N . , and Chang Y . K . (1999) Ethanol production using concentrated oak wood hydrolysates and methods to detoxify. Applied Biochemistry and Biotechnology 77-79: 547-559. Lee Y . - H . , and Fan L.T. (1982) Kinetic studies of enzymatic hydrolysis of insoluble cellulose: analysis of the initial rates. Biotechnology and Bioengineering 24: 2383-2406. Lee Y . - H . , and Fan L.T. (1983) Kinetic studies of enzymatic hydrolysis of insoluble cellulose: (II). Analysis of extended hydrolysis times. Biotechnology and Bioengineering 25: 939-966. Leonard R.H., and Hajny G.J. (1945) Fermentation of wood sugars to ethyl alcohol. Industrial and Engineering Chemistry 37: 390-395. Linden T., and Hahn-Hagerdal B. (1989) Fermentation of lignocellulose hydrolysates with yeasts and xylose isomerase. Enzyme and Microbial Technology 11: 583-589. Lindsay S.E., Bothast R.J., and Ingram L.O. (1995) Improved strains of recombinant Escherichia coli for ethanol production from sugar mixtures. Applied and Environmental Microbiology 43: 70-75. Lomax T.D., Mackie K . L . , Meder R., Croucher M . , and Burton R.J. (1994) Steam explosion of Pinus radiata bark. Journal of Wood Chemistry and Technology 14: 539-561. Lynd L.R. (1996) Overview and evaluation of fuel ethanol from biomass: technology, economics, the environment, and policy. Annual Review of Energy and the Environment 21: 403-465. Mackie K . L . , Brownell H.H., West K.L., and Saddler J.N. (1985) Effect of sulphur dioxide and sulphuric acid on steam explosion of aspen wood. Journal of Wood Chemistry and Technology 5: 405-425. Maiorella B., Blanch H.W., and Wilke C R . (1983) By-product inhibition effects on ethanolic fermentation by Saccharomyces cerevisiae. Biotechnology and Bioengineering 25: 103-121. Mamers H., and Menz D.N.J. (1984) Explosion pretreatment of Pinus radiata woodchips for the production of fermentation substrates. Appita 37: 644-649.  196  Mandels M . , and Reese E.T. (1963) Inhibition of cellulases and B-glucosidases. In: Reese E.T. (ed) Advances in enzymic hydrolysis of cellulose and related materials. Pergamon Press, New York, pp 115-157. Marsden W.L., Dunn N.W., and Gray P.P. (1983) Action of Trichoderma reesei QM9414 and C30 cellulase on substrates with varying crystallinity. Enzyme and Microbial Technology 5: 345-351. Martin C , and Jonsson L . (2003) Comparison of the resistance of industrial and laboratory strains of Saccharomyces and Zygosaccharomyces to lignocellulose-derived fermentation inhibitors. Enzyme and Microbial Technology 32: 386-395. Martin S.A., and Blake G.G. (1989) Effect of phenolic compounds on a commercial hemicellulase and two cellulases. Nutrition Reports International 40: 685-693. McCloy B.W., and O'Connor D . V . (1998) Wood Ethanol: A B C Value Added Opportunity. Report prepared for: Ministry of Environment, Lands and Parks, Environment Canada-Pacific &Yukon Region, Dec. 1998. McCloy B.W., and O'Connor D . V . (1999) Wood-ethanol opportunities and barriers. prepared for: Forest Sector Table, Feb. 1999.  Report  McCloy B.W. and Associates (2003) Estimated Production, Consumption and Surplus M i l l Residues in British Columbia, a 2003 Update. Report prepared for: Natural Resources Canada, Feb. 2003. McMillan J.D. (1994) Pretreatment of lignocellulosic biomass. In: Himmel M.E., Baker J.O., and Overend R P . (eds) Enzymatic conversion of biomass for fuels production. American Chemical Society, Washington, pp 292-324. Mes-Hartree M . , and Saddler J.N. (1983) The nature of inhibitory materials present in pretreated lignocellulosic substrates which inhibit the enzymatic hydrolysis of cellulose. Biotechnology Letters 5: 531-536. 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.A., Baker A.J., and Satter L.D. (1974) Pretreatments to enhance chemical, enzymatic, and microbial attack of cellulosic materials. Cellulose as a Chemical and Energy Resource, Berkeley, California, June 25-27, 1974. Biotechnology & Bioengineering Symposium 5: 193219. Millett M.A., Baker A.J., and Satter L . D . (1975) Physical and chemical pretreatments for enhancing cellulose saccharification. Enzymatic conversion of Cellulosic Materials: Technology and Applications, Newton and Natick, Massachusetts, Sept 8-10, 1975. Biotechnology & Bioengineering Symposium 6: 125-153.  197  Mooney C , Mansfield S.D., Tuohy M . , and Saddler J.N. (1998) The effect of initial pore volume and lignin content on the enzymatic hydrolysis of softwoods. Bioresource Technology 64: 113119. Mooney C , Mansfield S.M., Beatson R.P., and Saddler J.N. (1999) The effect of fiber characteristics on hydrolysis and cellulase accessibility to softwood substrates. Enzyme and Microbial Technology 25: 644-650. Neale A.D., Scopes R.K., and Kelly J.M. (1988) Alcohol production from glucose and xylose using Escherichia coli containing Zymomonas mobilis genes. Applied and Environmental Microbiology 29: 162-167. Nguyen Q.A., and Saddler J.N. (1991) An integrated model for the technical and economic evaluation of an enzymatic biomass conversion process. Bioresource Technology 35: 275-282. 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: 133-142. Nguyen Q.A., Tucker M.P., Keller F.A., and Eddy F.P. (2000) Two-stage dilute-acid pretreatment of softwoods. Applied Biochemistry and Biotechnology 84-86: 561-576. Nilvebrant N.-O., Reimann A., de Sousa F., Kleen M . , and Palmqvist E. (1997) Identification of fermentation inhibiting compounds from acid hydrolysis of wood. 9th International Symposium for Wood and Pulping Chemistry (Montreal) 79: 1-4. Nilvebrant N.-O., Reimann A., Larsson S., and Jonsson L.J. (2001) Detoxification of lignocellulose hydrolysates with ion-exchange resins. Applied Biochemistry and Biotechnology 91-93: 35-49. Nilvebrant N.-O., Persson P., Reimann A . , De Sousa F., Gorton L . , and Jonsson L.J. (2003) Limits for alkaline detoxification of dilute-acid lignocellulose hydrolysates. Applied Biochemistry and Biotechnology 108: 615-628. Nishikawa N.K., Sutcliffe R., and Saddler J.N. (1987) The effect of wood-derived inhibitors on 2,3-butanediol production by Klebsiella pneumoniae. Biotechnology and Bioengineering 31: 624-627. Nishikawa N.K., Sutcliffe R., and Saddler J.N. (1988) The influence of lignin degradation products on xylose fermentation by Klebsiella pneumoniae. Applied Microbiology and Biotechnology 27: 549-552. Ohta K , Beall D.S., Mejia J.P., Shanmugan K.T., and Ingram L.O. (1991) Genetic improvement of Escherichia coli for ethanol production: chromosomal integration of Zymomonas mobilis genes encoding pyruvate decarboxylase and alcohol dehydrogenase II. Applied and Environmental Microbiology 57: 893-900.  198  Olsen K., Collas P., Boileau P., Blain D., Ha C , Henderson L . , Liang C , McICibbon S., and Morel-a-l'Huissier L . (2002) Canada's Greenhouse Gas Inventory 1990-2000. Report prepared for: Environment Canada, Greenhouse Gas Division, June 2002. Report number EPS 5/AP/8. Olsson L., Hahn-Hagerdal B., and Zacchi G. (1995) Kinetics of ethanol production by recombinant Escherichia coli K O I 1. Biotechnology and Bioengineering 45: 356-365. Ooshima H., Burns D.S., and Converse A.O. (1990) Adsorption of cellulase from Trichoderma reesei on cellulose and lignacious residue in wood pretreated by dilute sulfuric acid with explosive decompression. Biotechnology and Bioengineering 36: 446-452. Ostergaard S., Olsson L . , and Nielsen J. (2000) In vivo dynamics of galactose metabolism in Saccharomyces cerevisiae: Metabolic fluxes and metabolite levels. Biotechnology and Bioengineering 73: 412-425. Overend R.P., and Chornet E. (1987) Fractionation of lignocellulosics by steam-aqueous pretreatments. Philosophical Transactions of the Royal Society of London "A" 321: 523-536. Palmqvist E. (1998) Fermentation of lignocellulosic hydrolysates: Inhibition and Detoxification. Doctoral Thesis, Department of Applied Microbiology, Lund University. Palmqvist E., Almeida J.S., and Hahn-Hagerdal B. (1999a) Influence of furfural on anaerobic glycolytic kinetics of Saccharomyces cerevisiae in batch culture. Biotechnology & Bioengineering. 62: 447-454. Palmqvist E., Grage H., Meinander N.Q., and Hahn-Hagerdal B . (1999b) Main and interaction effects of acetic acid, furfural, and 4-hydroxybenzoic acid on growth and ethanol productivity of yeasts. Biotechnology & Bioengineering 63: 46-55. Palmqvist E., and Hahn-Hagerdal B. (2000) Fermentation of lignocellulosic hydrolysates. II: inhibitors and mechanisms of inhibition. Bioresource Technology 74: 25-33. Parisi F. (1989) Advances in lignocellulosics hydrolysis and in the utilization of the hydrolyzates. Advances in Biochemical Engineering/Biotechnology 38: 53-87. Poltak R.F., and Grumet J.S. (2001) Health, environmental, and economic impacts of adding ethanol to gasoline in the northeast states. V o l 1. Summary and recommendations. Report prepared for: New England Interstate Water Pollution Control Commission (NEIWPCC), and Northeast States for Coordinated Air Use Management (NESCAUM), July, 2001. Ranatunga T.D., Jervis J., Helm R.F., McMillan J.D., and Wooley R.J. (2000) The effect of overliming on the toxicity of dilute acid pretreated lignocellulosics: the role of inorganics, uronic acids and ether-soluble organics. Enzyme and Microbial Technology 27: 240-247. Roquemore K . G . (1976) Hybrid designs for quadratic response surfaces. Technometrics 18: 419423.  199  Rowland H . (1997) U.S. crude oil imports: Growing U.S. dependence, in "Petroleum 1996 Issues and Trends". Report prepared for: Energy Information Administration, U.S. Department of Energy. Report number DOE/EIA-0615. Russel J.B. (1992) Another explanation for the toxicity of fermentation acids at low pH: anion accumulation versus uncoupling. Journal of Applied Bacteriology 73: 363-370. Saddler J.N., Ramos L.P., and Breuil C. (1993) Steam pretreatment of lignocellulosic residues. In: Saddler J.N. (ed) Bioconversion of Forest and Agricultural Plant Residues. C A B International, Oxon, pp 73-91. Saeman J.F. (1945) Kinetics of wood saccharification: hydrolysis of cellulose and decomposition of sugars in dilute acid at high temperature. Industrial and Engineering Chemistry 37: 43-52. Saeman J.F., Moore W.E., Mitchell R.L., and Millett M . A . (1954) Techniques for the determination of pulp constituents by quantitative paper chromatography. TAPPI 37: 336-343. Salmon J . M . (1989) Effect of sugar transport inactivation in Saccharomyces cerevisiae on sluggish and stuck enological fermentations. Applied & Environmental Microbiology 55: 953958. San Martin R., Perez C , and Briones R. (1995) Simultaneous production of ethanol and kraft pulp from pine (Pinus radiata) using steam explosion. Bioresource Technology 53: 217-223. Sanchez B., and Bautista J. (1988) Effects of furfural and 5-hydroxymethylfurfural on the fermentation of Saccharomyces cerevisiae and biomass production from Candida guilliermondii. Enzyme and Microbial Technology 10: 315-318. Scherrard E C , and Kressman F.W. (1945) Review of processes in the United States prior to World War II. Industrial and Engineering Chemistry 37: 5-8. Schwald W., Smaridge T., Chan M . , Breuil C , and Saddler J.N. (1989) The influence of S 0  2  impregnation and fractionation on product recovery and enzymic hydrolysis of steam-treated sprucewood. In: Coughlan M . P . (ed) Enzyme Systems for Lignocellulose Degradation. Elsevier Science Publishers Ltd., Essex, pp 231-242. Sene L . , Converti A . , Zilli M . , Felipe M.G.A., and Silva S.S. (2001) Metabolic study of the adaptation of the yeast Candida guilliermondii to sugarcane bagasse hydrolysate. Applied Microbiology and Biotechnology 57: 738-743. Sewalt V.J.H., Glasser W.G., and Beauchemin K . A . (1997) Lignin impact on fiber degradation. 3. Reversal of inhibition of enzymatic hydrolysis by chemical modification of lignin and by additives. Journal of Agricultural and Food Chemistry 45: 1823-1828. Sheehan J. (2001) The road to bioethanol: A strategic perspective of the U.S. Department of Energy's National Ethanol Program. In: Himmel M.E., Baker J.O., and Saddler J.N. (eds) Glycosyl Hydrolases for Biomass Conversion. American Chemical Society, Washington, D C , pp 2-25. 200  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 component after steam explosion of softwood chips. Bioresource Technology 72: 207-211. Shleser R. (1994) Ethanol production in Hawaii. Processes, feedstocks, and current economic feasibility of fuel grade ethanol production in Hawaii. Report prepared for: Prepared for the State of Hawaii, Department of Business, Economic Development and Tourism, Energy Division. Report number HD9502.5.B54.S4.1994. Singleton V . L . , and Rossi J.A. (1965) Colorimetry of total phenolic with phosphomolybdicphosphotungstic acid reagents. American Journal of Enology and Viticulture 16: 144-158. Sinitsyn A.P., Clesceri L.S., and Bungay H.R. (1982) Inhibition of cellulases by impurities in steam-exploded wood. Applied Biochemistry and Biotechnology 7: 455-458. Sinitsyn A.P., Gusakov A . V . , and Vlasenko E.Y. (1991) Effect of structural and physio-chemical features of cellulosic substrates on the efficiency of enzymatic hydrolysis. Applied Biochemistry and Biotechnology 30: 43-59. Sinsabaugh R.L., and Linkins A . E . (1987) Inhibition of the Trichoderma complex by leaf litter extracts. Soil Biology and Biochemistry 19: 719-726. Sjostrom E. (1993) Wood Chemistry: Fundamentals and Applications. 2 Press, Inc., San Diego.  nd  viride cellulase  edition. Academic  Skoog K., and Hahn-Hagerdal B. (1990) The effect of oxygenation on xylose fermentation by Pichia stipitis. Applied and Environmental Microbiology 56: 3389-3394. Soderstrom J., Pilcher L., Galbe M . , and Zacchi G. (2002) Two-step steam pretreatment of softwood with SO2 impregnation for ethanol production. Applied Biochemistry and Biotechnology 98-100: 5-21. Stenberg K., Tengborg C , Galbe M . , and Zacchi G. (1998) Optimisation of steam pretreatment of S02-impregnated mixed softwoods for ethanol production. Journal of Chemical Technology & Biotechnology 71: 299-308. Sutcliffe R., and Saddler J.N. (1986) The role of lignin in the adsorption of cellulases during enzymatic treatment of lignocellulosic material. Biotechnology and Bioengineering Supplement 17: 749-762. Swings J., and De Ley J. (1977) The biology of Zymomonas mobilis. Bacteriology Reviews 41: 1-46.  201  Taherzadeh M.J., Eklund R., Gustafsson L., Niklasson C , and Liden G. (1997a) Characterization and fermentation of dilute-acid hydrolyzates from wood. Industrial and Engineering Chemistry Research 36: 4659-4665. Taherzadeh M.J., Niklasson C , and Liden G. (1997b) Acetic acid - friend or foe in anaerobic batch conversion of glucose to ethanol by Saccharomyces cerevisiae. Chemical Engineering Science 52: 2653-2659. Taherzadeh M.J., Gustafsson L . , Niklasson C , and Liden G. (1999) Conversion of furfural in aerobic and anaerobic batch fermentation of glucose by Saccharomyces cerevisiae. Journal of Bioscience and Bioengineering 87: 169-174. Taherzadeh M.J., Gustafsson L., Niklasson C , and Liden G. (2000) Physiological effects of 5hydroxymethylfurfural (HMF) on Saccharomyces cerevisiae. Applied Microbiology and Biochemistry 53: 701-708. Taherzadeh M . , Adler L., and Liden G. (2002) Strategies for enhancing fermentative production of glycerol - a review. Enzyme and Microbial Technology 31: 53-66. Tanahashi M . , Takada S., Aoki T., Goto T., Higuchi T., and Hanai S. (1983) Characterization of explosion wood. 1. Structure and physical properties. Wood Research 69: 36-51. Tengborg C , Stenberg K., Galbe M . , Zacchi G., Larsson S., Palmqvist E., and Hahn-Hagerdal B. (1998) Comparison of SO2 and H 2 S O 4 impregnation of softwood prior to steam pretreatment on ethanol production. Applied Biochemistry and Biotechnology 70-72: 3-15. Tengborg C , Galbe M . , and Zacchi G. (2001) Reduced inhibition of enzymatic hydrolysis of steam-pretreated softwood. Enzyme and Microbial Technology 28: 835-844. Theander O. (1982) Cellulose, hemicellulose and extractives. In: Overend R.P., Milne T.A., and Mudge L . K . (eds) Fundamentals of thermomechanical biomass conversion. Elsevier Applied Science, New York, pp 35-60. Thompson D.N., Chen H . - C , and Grethlein H.E. (1992) Comparison of pretreatment methods on the basis of available surface area. Bioresource Technology 39: 155-163. Torget R., Himmel M . E . , and Grohmann K . (1991) Dilute sulfuric-acid pretreatment of hardwood bark. Bioresource Technology 35: 239-246. Tran A . V . , and Chambers R.P. (1986) Lignin and extractives derived inhibitors in the 2,3butanediol fermentation of mannose-rich prehydrolysates. Applied Microbiology and Biotechnology 23: 191-197. Vasquez G., Antorrena G., Parajo J . C , and Francisco X . L . (1988) Acid prehydrolysis of Pinus pinaster bark with dilute sulfuric acid. Wood Science and Technology 22: 219-225. Verduyn C , Postma E., Scheffers W.A., and van Dijken J.P. (1990) Energetics of Saccharomyces cerevisiae in anaerobic glucose-limited chemostat cultures. Journal of General Microbiology 136: 405-412. 202  Vinzant T.B., Ehrman C L , Adney W.S., Thomas S.R., and Himmel M . , E. (1997) Simultaneous saccharification and fermentation of pretreated hardwoods. Effect of native lignin content. Applied Biochemistry and Biotechnology 62: 99-104. Voipio R., and Laakso T. (1992) Chemical composition of the above ground biomass of smallsized trees. Folia Forestalia 789: 1-22. von Sivers M . , and Zacchi G. (1995) A techno-economical comparison of three processes for the production of ethanol from pine. Bioresource Technology 51: 43-52. Vroom K.E. (1957) The " H " factor: A means of expressing cooking times and temperatures as a single variable. Pulp and Paper Magazine of Canada 58: 228-231. Walch E., Zemannm A . , Schinner F., Bonn G., and Bobleter O. (1992) Enzymatic saccharification of hemicellulose obtained from hydrothermally pretreated sugar cane bagasse and beech bark. Bioresource Technology 39: 173-177. Wayman M . , Tallevi A . , and Winsborrow B. (1984) Hydrolysis of biomass by sulfur dioxide. Biomass 6: 183-191. Wills C. (1990) Regulation of sugar and ethanol metabolism in Saccharomyces Biochemistry and Molecular Biology 25: 245-272.  cerevisiae.  Wilson JJ, Deschatelets L , Nishikawa N K (1989) Comparative fermentability of enzymatic and acid hydrolysates of steam-pretreated aspenwood hemicellulose by Pichia stipitis CBS 5776. Applied Microbiology and Biotechnology 31: 592-596. Wood T M , Bhat K M (1988) Methods for measuring cellulase activities. In: Wood W A , Kellogg ST (eds) Biomass - Part A : Cellulose and Hemicellulose. Academic Press, San Diego, pp 87112. Wood G., and Egyed M . (1994) Risk assessment for the combustion products of methylcyclopentadienyl manganese tricarbonyl (MMT) in gasoline. Report prepared for: Environmental Health Directorate, Health Canada, Dec. 6, 1994. 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 softwoods. Applied Biochemistry and Biotechnology 77-79: 47-54. Yang B., Boussaid A . , Mansfield S.D., Gregg D.J., and Saddler J.N. (2002) Fast and efficient alkaline peroxide treatment to enhance the enzymatic digestibility of steam-exploded softwood substrates. Biotechnology and Bioengineering 77: 678-684. Zacchi G., and Axelsson A . (1989) Economic evaluation of preconcentration in production of ethanol from dilute sugar solutions. Biotechnology and Bioengineering 34: 223-233. Zhang M . , Eddy C , Deanda K., Finkelstein M . , and Picataggio S. (1995) Metabolic engineering of a pentose metabolism pathway in ethanologenic Zymomonas mobilis. Science 267: 240-243. 203  Zhuang Q., and Vidal P.F. (1997) Hemicellulose solubilization from Populus tremuloides via steam explosion and charaterization of the water soluble fraction. II. Alkali-catalytic process. Cellulose Chemistry and Technology 31: 37-49.  204  Appendix 1 Enzyme activities in the enzyme preparations employed for hydrolysis experiments. (3-glucosidase Activity  Mannanase Activity  Xylanase Activity  (FPU mL" )  (IU mL" )  (IU mL" )  (IU mL" )  49  50  30  22  226  44  5  320  N.D.  Total Protein Content  FPU Activity  (mg mL" ) Celluclast 1.5L Novozym-188  1  a  1  1  1  a  1  94  N . D . : Not Determined  205  Appendix 2  Standard curve correlating Tembec T l inoculum optical density and dry weight cell concentration.  200  0  50  100  150  200  250  300  350  Optical Density (600nm)  206  Appendix 3  Standard curve correlating total phenolic concentration (g L" ) and absorbance (760 nm) using vanillin (4-hydroxy-3-methoxybenzaldehyde). 1  i 0.0  1  1  0.1  1  1  0.2  1  1  0.3  1  1  1  0.4  1  0.5  1  1  0.6  Vanillin Concentration (g L" ) 1  207  Appendix 4 Delignification of feedstock WI-0 (Whitewood water-insoluble fraction) Original Material Material Material  material used in each delignification (g recovered following delignification 1 (g recovered following delignification 2 (g recovered following delignification 2 (g  DW) DW) DW) DW)  9.00 5.27 5.19 5.12 Arabinan Galactan Glucan  Oriqinal Material  Lignin ASL AIL  Mannan  Xylan  0.5 0.0  43.1 3.9  99.8 9.0  0.3 0.0 0.0  9.1 0.5 3.4  97.7 5.1 3.8  Total  Composition of WI-0 (%) Weight (g)  0.1 0.0  0.1 0.0  55.0 5.0  0.7 0.1  0.3 0.0  ComDOSition of WI-AP-0 (%) Weight (g) Loss (g) Loss (%)  0.1 0.0 0.0  0.1 0.0 0.0  87.2 4.6 0.4  0.6 0.0 0.0  0.3 0.0 0.0  Deliqnification 2:  ComDOSition of WI-AP-0 (%) Weight (g) Loss(g) Loss (%)  0.1 0.0 0.0  0.1 0.0 0.0  87.7 4.6 0.4 8.0  0.7 0.0 0.0  0.3 0.0 0.0  0.3 0.0 0.0  9.5 0.5 3.4 87.3  98.7 .5.1 3.9  Deliqnification 3:  Comoosition of WI-AP-0 (%) Weight (g) Loss (g) Loss(%)  0.1 0.0 0.0  0.0 0.0 0.0  87.0 4.5 0.5  0.5 0.0 0.0  0.2 0.0 0.0  0.3 0.0 0.0  9.5 0.5 3.4 87.5  97.6 5.0 4.0  Mannan 0.4 0.0  Xylan 0.3 0.0  0.6 0.1  56.0 5.0  Deliqnification 1:  1 I  I  '.J87T&;;.  7.2  io.6"  Delignification of feedstock WI-30 (Whitewood+bark water-insoluble fraction) Original Material Material Material  material used in delignification (g DW) recovered following delignification 1 (g DW) recovered following delignification 2 (g DW) recovered following delignification 3 (g DW)  9.00 4.51 4.37 4.52 Arabinan Galactan Glucan  Lignin ASL AIL  Total 99.4 8.9  Oriqinal Material  Composition of WI-30 (%) Weight (g)  0.1 0.0  0.0 0.0  42.0 3.8  Deliqnification 1:  Composition of WI-AP-30 (%) Weight (g)  0.1 0.0 0.0  0.1 0.0 0.0  78.0 3.5 0.3 6.9  0.5 0.0 0.0  0.5 0.0 0.0  0.5 0.0 0.0  18.3 0.8 4.2 83.6  98.0 4.4 4.5  0.0 0.0 0.0  0.1 0.0 0.0  78.4 3.4 0.4 9.4  0.6 0.0 0.0  0.4 0.0 0.0  0.5 0.0 0.0  18.0 0.8 4.3 844  98.0 4.3 4.7  0.1 0.0 0.0  0.1 0.0 0.0  77.5  0.5  0.3  0.4  3.5 0.3 7.3  0.0 0.0  0.0 0.0  0.0 0.0  18.5 0.8 4.2 83.4  97.4 4.4  Loss (g) Loss(%) Deliqnification 2:  Composition of WI-AP-30 (%) Weight (g) Loss (g) Loss(%)  Deliqnification 3:  Composition of WI-AP-30 (%) Weight (g) Loss (g) Loss(%)  Note: all feedstock compositions are the average of two determinations  4.5  

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.831.1-0075111/manifest

Comment

Related Items