<%BANNER%>

Simplifying the Lignocellulose to Ethanol Process Through Efficient Pretreatment and Improvement of Biocatalyst

Permanent Link: http://ufdc.ufl.edu/UFE0041890/00001

Material Information

Title: Simplifying the Lignocellulose to Ethanol Process Through Efficient Pretreatment and Improvement of Biocatalyst
Physical Description: 1 online resource (121 p.)
Language: english
Creator: Geddes, Claudia
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: cellulase, cellulose, ethanol, lignocellulose, pretreatment, sscf, ssf, viscosity, xylose
Microbiology and Cell Science -- Dissertations, Academic -- UF
Genre: Microbiology and Cell Science thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: A low level of phosphoric acid (1% w/w on dry bagasse basis, 160masculine ordinalC and above, 10 minutes) was shown to effectively hydrolyze the hemicellulose in sugar cane bagasse into monomers with minimal side reactions and to serve as an effective pre-treatment for the enzymatic hydrolysis of cellulose. Up to 45% of the remaining water insoluble solids (WIS) was digested to sugar monomers by a low concentration of Biocellulase W (0.5 filter paper unit/g WIS) supplemented with ?-glucosidase, although much higher levels of cellulase (100-fold) were required for complete hydrolysis. After neutralization and nutrient addition, phosphoric acid syrups of hemicellulose sugars were fermented by ethanologenic E. coli LY160 without further purification. Fermentation of these syrups was preceded by a lag that increased with increased pretreatment temperature. Further improvements in organisms and optimization of steam treatments may allow co-fermentation of sugars derived from hemicellulose and cellulose, eliminating the need for liquid-solid separation, sugar purification, and separate fermentations. Consolidation of bioprocessing steps with lignocellulose are limited in part by hydrolysate toxicity, material handling associated with fibrous suspensions, and the low activity of cellulase enzymes. Combinations of enzyme dose and treatment conditions were shown to improve flow properties and pumping of acid pretreated sugarcane bagasse slurries (10% dry weight). Low levels of cellulase enzyme (0.1 and 0.5 FPU/g dry weight acid pretreated bagasse) were found to reduce the viscosity of these slurries by 77-95% after 6 h of incubation, solubilizing at least 3.5% of the bagasse dry weight. Flow of slurries through small funnels was a useful predictor of success with centrifugal and diaphragm pumps. Equations were derived that describe changes in viscosity and solubilized sugars as a function of time and cellulase dosage. Blending of acid pretreated bagasse (10% dry weight) with suspensions of acid pretreated bagasse (10% dry weight) that had been previously digested with cellulase enzymes (low viscosity) did not increase viscosity in a linear fashion. Viscosity of these mixtures remained relatively constant until a threshold level of new fiber was reached, followed by a rapid increase with further additions. Up to 35% of fresh acid pretreated bagasse could be blended with enzyme-digested fiber (5.0 FPU/g dry weight acid pretreated fiber; 6 h) with only a modest increase in viscosity. A simple model is described to explain this phenomenon. The smooth surfaces of enzyme-treated fiber are proposed to hinder the frequency and extent of interactions between fibrils of fresh fiber particles (acid pretreated) until a threshold concentration is achieved, after which fiber interactions and viscosity increase dramatically. These results were used to model the viscosity in an ideal continuous stirred tank reactor (liquefaction) as a function of residence time and enzyme dosage. A hydrolysate-resistant mutant of E. coli LY180 was selected by sequential transfers in AM1 mineral salts medium containing hemicellulose hydrolysate from a dilute phosphoric acid pretreatment (160?C) of bagasse. The resulting strain was designated MM160. Strain MM160 was also more resistant than the parent to individual inhibitors such as furfural, 5-hydroxymethylfurfural, and acetate. With this mutant, process steps such as the separation of hemicellulose hydrolysate and fiber after dilute acid pretreatment (washing) and toxin removal from the hemicellulose hydrolysate were not required prior to the fermentation. Pretreated bagasse was fermented in a single vessel without separation using strain MM160. A liquefaction step with cellulase was included to improve pumping and mixing (L+SScF), analogous to a corn ethanol process. Furans and other compounds in the hemicellulose hydrolysate remained as a barrier to potential contaminants during fermentation. Bagasse slurries containing 10% and 14% dry weight (fiber plus solubles) were tested in fermentations using pretreatment temperatures of 160-190?C (1% phosphoric acid, 10 min). Saccharification efficiency and inhibitor production increased with treatment temperature. Bagasse samples pretreated at temperatures below 190?C were fermentable. The highest yields and titers were obtained after 122 h of incubation using 14% dry weight slurries of pretreated bagasse (180?C), 0.21 g ethanol/g bagasse (dry weight) and 30 g/L, respectively. These results are similar to those that have been previously reported using pretreatments with ammonia and other bases in SScF processes.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Claudia Geddes.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Ingram, Lonnie O.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041890:00001

Permanent Link: http://ufdc.ufl.edu/UFE0041890/00001

Material Information

Title: Simplifying the Lignocellulose to Ethanol Process Through Efficient Pretreatment and Improvement of Biocatalyst
Physical Description: 1 online resource (121 p.)
Language: english
Creator: Geddes, Claudia
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: cellulase, cellulose, ethanol, lignocellulose, pretreatment, sscf, ssf, viscosity, xylose
Microbiology and Cell Science -- Dissertations, Academic -- UF
Genre: Microbiology and Cell Science thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: A low level of phosphoric acid (1% w/w on dry bagasse basis, 160masculine ordinalC and above, 10 minutes) was shown to effectively hydrolyze the hemicellulose in sugar cane bagasse into monomers with minimal side reactions and to serve as an effective pre-treatment for the enzymatic hydrolysis of cellulose. Up to 45% of the remaining water insoluble solids (WIS) was digested to sugar monomers by a low concentration of Biocellulase W (0.5 filter paper unit/g WIS) supplemented with ?-glucosidase, although much higher levels of cellulase (100-fold) were required for complete hydrolysis. After neutralization and nutrient addition, phosphoric acid syrups of hemicellulose sugars were fermented by ethanologenic E. coli LY160 without further purification. Fermentation of these syrups was preceded by a lag that increased with increased pretreatment temperature. Further improvements in organisms and optimization of steam treatments may allow co-fermentation of sugars derived from hemicellulose and cellulose, eliminating the need for liquid-solid separation, sugar purification, and separate fermentations. Consolidation of bioprocessing steps with lignocellulose are limited in part by hydrolysate toxicity, material handling associated with fibrous suspensions, and the low activity of cellulase enzymes. Combinations of enzyme dose and treatment conditions were shown to improve flow properties and pumping of acid pretreated sugarcane bagasse slurries (10% dry weight). Low levels of cellulase enzyme (0.1 and 0.5 FPU/g dry weight acid pretreated bagasse) were found to reduce the viscosity of these slurries by 77-95% after 6 h of incubation, solubilizing at least 3.5% of the bagasse dry weight. Flow of slurries through small funnels was a useful predictor of success with centrifugal and diaphragm pumps. Equations were derived that describe changes in viscosity and solubilized sugars as a function of time and cellulase dosage. Blending of acid pretreated bagasse (10% dry weight) with suspensions of acid pretreated bagasse (10% dry weight) that had been previously digested with cellulase enzymes (low viscosity) did not increase viscosity in a linear fashion. Viscosity of these mixtures remained relatively constant until a threshold level of new fiber was reached, followed by a rapid increase with further additions. Up to 35% of fresh acid pretreated bagasse could be blended with enzyme-digested fiber (5.0 FPU/g dry weight acid pretreated fiber; 6 h) with only a modest increase in viscosity. A simple model is described to explain this phenomenon. The smooth surfaces of enzyme-treated fiber are proposed to hinder the frequency and extent of interactions between fibrils of fresh fiber particles (acid pretreated) until a threshold concentration is achieved, after which fiber interactions and viscosity increase dramatically. These results were used to model the viscosity in an ideal continuous stirred tank reactor (liquefaction) as a function of residence time and enzyme dosage. A hydrolysate-resistant mutant of E. coli LY180 was selected by sequential transfers in AM1 mineral salts medium containing hemicellulose hydrolysate from a dilute phosphoric acid pretreatment (160?C) of bagasse. The resulting strain was designated MM160. Strain MM160 was also more resistant than the parent to individual inhibitors such as furfural, 5-hydroxymethylfurfural, and acetate. With this mutant, process steps such as the separation of hemicellulose hydrolysate and fiber after dilute acid pretreatment (washing) and toxin removal from the hemicellulose hydrolysate were not required prior to the fermentation. Pretreated bagasse was fermented in a single vessel without separation using strain MM160. A liquefaction step with cellulase was included to improve pumping and mixing (L+SScF), analogous to a corn ethanol process. Furans and other compounds in the hemicellulose hydrolysate remained as a barrier to potential contaminants during fermentation. Bagasse slurries containing 10% and 14% dry weight (fiber plus solubles) were tested in fermentations using pretreatment temperatures of 160-190?C (1% phosphoric acid, 10 min). Saccharification efficiency and inhibitor production increased with treatment temperature. Bagasse samples pretreated at temperatures below 190?C were fermentable. The highest yields and titers were obtained after 122 h of incubation using 14% dry weight slurries of pretreated bagasse (180?C), 0.21 g ethanol/g bagasse (dry weight) and 30 g/L, respectively. These results are similar to those that have been previously reported using pretreatments with ammonia and other bases in SScF processes.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Claudia Geddes.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Ingram, Lonnie O.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041890:00001


This item has the following downloads:


Full Text





SIMPLIFYING THE LIGNOCELLULOSE TO ETHANOL PROCESS THROUGH
EFFICIENT PRETREATMENT AND IMPROVEMENT OF BIOCATALYST




















By

CLAUDIA GEDDES


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2010

































2010 Claudia Geddes



























To my husband who has believed in me every step of the way









ACKNOWLEDGMENTS

I thank my advisor, Dr. Lonnie Ingram, for allowing me to join his lab and for his

mentorship throughout my graduate studies. I thank my committee members for all their

invaluable advice and support. I thank Mike Mullinnix, Dr. James Peterson, Dr. Ismael

Nieves, Ralph Hoffman, Dr. Zhaohui Tong, Sean York, Lorraine Yomano, Dr. Laura

Jarboe and Dr. Elliot Miller for their assistance and advice in experiments. Mike

Mullinnix and Sean York were instrumental in creating the hydrolysate-resistant strain

MM160. Lorraine Yomano carried out the removal of FRT sites in strain MM105. Dr.

Ismael Nieves, Ralph Hoffman, and Dr. Elliot Miller assisted in carrying out

fermentations. Dr. James Peterson was instrumental in carrying out compositional

analyses of sugarcane bagasse. I thank all my family for their love and support that has

given me confidence to achieve my goals. Last but not least, I thank my husband, Ryan

Geddes, for believing in me and for his unwavering support.









TABLE OF CONTENTS

page

ACKNOW LEDG M ENTS..................................... ........................ ... .. 4

LIS T O F T A B LE S ........................................................ 8

L IS T O F F IG U R E S ................................................................................ 9

A B S T R A C T ....................................................... 10

CHAPTER

1 IN T R O D U C T IO N ................................................................................... 13

Lignocellulosic Biomass ................ ........ ........... ........ ................ 13
Comparison of Major Lignocellulosic Substrates ......... ....... .. ............. 14
Lignocellulosic Biomass Pretreatment ........................ .......... ..................17
Inhibitor Formation and Detoxification Methods ....... ...... ... .. .................. 19
Biological Methods .......................... ..................22
Chemical Methods ................ ......... ........ .........23
Adaptive Evolution ..... ............ .... .......................24
Example Process for the Conversion of Lignocellulosic Biomass to Ethanol ..........25
Feedstock Storage, Handling and Size Reduction ................. ................. 25
Hem icellulose Hydrolysis .............. ... ...............27
Solid-Liquid Separations ......... ......... ......... .. ........................... 28
Hydrolysate Conditioning ........................ ....... ........... ..... ..... ...30
Cellulose Hydrolysis and Fermentation ............. ........... ........ 30
D is tilla tio n .................... .. .. ............. .. .............................................. 3 2
W astew ater Treatm ent ...................... ...... ....... .. .. ............................ 34
R research O objectives .................... .................................... ...................... ..... 35
Reduce the Toxicity of Sugarcane Bagasse Hemicellulose Hydrolysate and
Eliminate the Need for Separation of the Solid Lignin-Cellulose Rich
Residue .................................... 35
Establish a Correlation between Solubilized Sugar, Viscosity, and Ability to
Pump for Pretreated and Enzymatically Digested Sugarcane Bagasse
Slurries ........... ..... ...... ......... ............... ........................35
Design a Unified Fermentation Process that Combines Cellulose
Hydrolysis, Hemicellulose Hydrolysate, and Fermentation in One Vessel.....36

2 OPTIMIZING THE SACCHARIFICATION OF SUGARCANE BAGASSE USING
DILUTE PHOSPHORIC ACID FOLLOWED BY FUNGAL CELLULASES ............... 37

In tro d u c tio n ................................................................................................... 3 7
M materials and M methods ........................................... ................................. 39
M materials ................................................................................ ...... ......................39
Standard Analysis of Sugarcane Bagasse Composition ................ ................40









Effect of Phosphoric and Sulfuric Acid Concentrations at 145C .................40
Effect of Temperature on Steam Treatment of Sugarcane Bagasse with 1%
(w /w ) H 3 P O 4 ...................... .. ......... .. ..... .......... ................. ... ........................ 4 0
Enzymatic Hydrolysis of Cellulose after Steam Treatment with Phosphoric
A c id ................. ........ ... .......... .. ....... ..... ...................4 1
Phosphoric Acid Method for Analysis of Bagasse Composition....................... 41
Toxicity and Fermentation of Hemicellulose Hydrolysate (Phosphoric Acid) ....42
Analytical Methods ............... ....... ......... ............ ................ 43
Statistical Analysis ............... ......... ........ ......... 43
R results and D discussion ................................. ............................... ... .. ............ ..... 44
Effects of Phosphoric and Sulfuric Acid on Hemicellulose Hydrolysis
(1450C 1 h) ......... .. ............. .... ............. ........ ............................ 44
Increasing the Temperature to Reduce Treatment Time and Phosphoric
A c id U s a g e ............................................................................. ..........................4 5
Effect of Phosphoric Acid Pretreatment on the Enzymatic Hydrolysis of
Bagasse .............................. ....... ... .......... ............. 47
Phosphoric Acid Method for Analysis of Bagasse Composition....................... 48
Toxicity of Phosphoric Acid Hydrolysate..........................................................49
C conclusions ............................................................... .... ..... ... ...... 50

3 OPTIMIZING CELLULASE USAGE FOR IMPROVED MIXING AND
RHEOLOGICAL PROPERTIES OF ACID-PRETREATED SUGARCANE
B A G A S S E ............... ....................................... ........................... 57

Introd auction ............ ............. ............................. ........................... 57
M materials and M methods ......... .............. .................... ......... .... .... .. 59
M materials ................... .... .... ...... ............................... ......... ................ 59
Dilute Acid Pretreatment of Sugarcane Bagasse ............... .. ............ .. 59
Saccharification with Biocellulase W and 3-glucosidase............. ........... 60
Relative Viscosity Measurement ...... ............. ........... ................ 61
Flow Property Test using Graded Funnels ........ ................. ................ 61
Carbohydrate Composition and Analyses............... ...... .................... .. 61
Statistical Analysis ............... ............. ......................... ... ...... 62
Results and D discussion ....................... ................... ...... .....................62
C om position ......................................................................... ...... .. ..... ................62
Effect of Cellulase on Relative Viscosity........................ ..... ..... ................ 63
Effect of Cellulase Loading on Extent of Saccharification................................ 65
Effect of Enzyme Treatment on Flow through Graded Funnels.........................66
Correlation between the Extent of Saccharification and Relative Viscosity.......67
Effect of Mixing Acid Pretreated Fiber (No Enzyme Digestion) with Enzyme-
digested Acid Pretreated Fiber (pH 5.0, 55 oC, 6 h) on Viscosity.................. 68
Modeling an Ideal Continuous Stirred Tank Reactor (CSTR) to Decrease
V is c o s ity ................................................... ................................ 7 0
C conclusions ............................................................................................. .............. 73









4 A SIMPLIFIED PROCESS FOR ETHANOL PRODUCTION FROM
SUGARCANE BAGASSE USING HYDROLYSATE-RESISTANT
ESCHERICHIA COLI STRAIN MM160 AND PRETREATMENT WITH
PHOSPHORIC ACID ........... ........... ......... .................. ................ 81

Introd uctio n ............ ........... .......................................................... 8 1
M materials and M methods ......... ..... ......................... ................ .. 83
Materials ............................... ..................83
Steam Treatment of Bagasse with 1% (w/w) H3PO4 ..................................... 83
Organisms and Growth Conditions .............. ...... .. .............. ................ 84
Genetic M methods ................ ......... .......... ........... 84
Isolation of Hydrolysate-resistant Biocatalysts ................................................. 84
Removal of Extraneous DNA Segments from Strain MM105 ..........................85
Tolerance to Hydrolysate Toxins ...................................................................... 86
Liquefaction Followed by Simultaneous Saccharification and Co-
fermentation (L+SScF) of Acid-pretreated Sugarcane Bagasse .................... 86
A n a ly s e s ................ ............................................................ ....... 8 7
S ta tistic a l A n a lys is ............... ........................................................ 8 8
R e s u lts .......................... ............. .... ................. ............ ..... .. ............... .. 8 8
Composition of Materials Used as Substrates for Fermentation....................... 88
Development of Hydrolysate-resistant Strain MM160 ...................................... 89
Liquefaction Prior to Simultaneous Saccharification and Co-fermentation
(L+SScF) .............. .............. ............... ........ ................ 90
D is c u s s io n ...................... .. ............. .. ................................................... 9 2

5 GENERAL CONCLUSIONS AND FUTURE DIRECTIONS.............................. 107

G general A ccom p lishm ents ..................... .............. ...................... ................107
Future D directions ............... ....... ...... ............... 1.......... .......... 109

LIST O F R E FER EN C ES ............. ......................... ........ .. ............ ................ 111

B IO G RA P H IC A L S K ET C H ..................................................... ..................................... 121









LIST OF TABLES


Table page

2-1 Sugar composition of sugarcane bagasse (% dry weight) ............................... 52

2-2 Solubilization of sugars from sugarcane bagasse by steam treatments with
phosphoric acid followed by 96 h enzyme (Biocellulase W) digestiona ...............52

2-3 Composition and fermentation of hemicellulose hydrolysates ................................53

3-1 Sugar composition of sugarcane bagasse, washed acid pretreated fiber and
hem icellulose hydrolysate ........... ...................... ................................ 75

3-2 Effects of cellulase enzymes on enzyme-solubilized sugars and theological
properties ................ ..................................................... ... ... .. .... ... 76

3-3 Predicted upper bound viscosity for an ideal continuous stirred tank reactor
(C S T R ) ........................... ........ ..................... 77

4-1 Strains, plasm ids and prim ers. .................................. ........................... .......... 94

4-2 Components of sugarcane bagasse after steam pretreatment with
phosphoric acid. ................................. ............... ......... .... ....... 96

4-3 Effect of enzyme treatment on sugar monomers in hemicellulose hydrolysate... 97

4-4 Effect of enzyme treatment on selected inhibitors in hemicellulose
h y d ro ly s a te ................................................................................... 9 8

4-5 Maximum ethanol concentrations and yields using L+SScF process.................99

4-6 Comparison of ethanol yields from SScF processes. .................................... 100









LIST OF FIGURES


Figure page

2-1 Effect of acid concentration on monomer sugars from dilute acid treatments
(1 h, 145C) w ith H3PO 4 and H2S04 ............... ............................................. 54

2-2 Effect of acid concentration on the production of side products furanss and
organic acids) during steam treatment ........... ............ ...... ........ ........... 55

2-3 Effect of treatment temperature on bagasse hydrolysis (hemicellulose) and
production of side products (1% w/w H3PO4, 10 min)..................... ................ 56

3-1 Effect of cellulase enzyme loading on saccharification and viscosity................. 78

3-2 Scatter plot of viscosity versus saccharification (10% w/w slurries of acid
pretreated fiber in hemicellulose hydrolysate)......................................... 79

3-3 Effect of acid pretreated fiber additions on the viscosity of cellulase-digested
slurries containing 10% dry weight acid pretreated fiber. ................ ................ 80

4-1 Comparison of ethanol production from lignocellulose and corn. .................. 101

4-2 Selected examples of improvements in fermentation during serial transfers of
MM105 in hydrolysate medium ......... .... ......... ...................... ............... 102

4-3 Resistance to selected inhibitors ........... .......... ...... ............. ................ 103

4-4 Composition of fermentation broth at the time of inoculation (after liquefaction
for 6 h w ith enzym es). ....................... ................................ ... .. 104

4-5 Fermentation of bagasse pretreated with phosphoric acid at 16D (L+SScF). 105

4-6 Fermentations of bagasse (14% w/v) pretreated with phosphoric acid at
various tem peratures ............. ................ ....... ....... ................ ........ 106










Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

SIMPLIFYING THE LIGNOCELLULOSE TO ETHANOL PROCESS THROUGH
EFFICIENT PRETREATMENT AND IMPROVEMENT OF BIOCATALYST

By

Claudia Geddes

August 2010

Chair: Lonnie O. Ingram
Major: Microbiology and Cell Science

A low level of phosphoric acid (1% w/w on dry bagasse basis, 1600C and above,

10 minutes) was shown to effectively hydrolyze the hemicellulose in sugar cane

bagasse into monomers with minimal side reactions and to serve as an effective pre-

treatment for the enzymatic hydrolysis of cellulose. Up to 45% of the remaining water

insoluble solids (WIS) was digested to sugar monomers by a low concentration of

Biocellulase W (0.5 filter paper unit/g WIS) supplemented with 3-glucosidase, although

much higher levels of cellulase (100-fold) were required for complete hydrolysis. After

neutralization and nutrient addition, phosphoric acid syrups of hemicellulose sugars

were fermented by ethanologenic E. coli LY160 without further purification.

Fermentation of these syrups was preceded by a lag that increased with increased

pretreatment temperature. Further improvements in organisms and optimization of

steam treatments may allow co-fermentation of sugars derived from hemicellulose and

cellulose, eliminating the need for liquid-solid separation, sugar purification, and

separate fermentations.









Consolidation of bioprocessing steps with lignocellulose are limited in part by

hydrolysate toxicity, material handling associated with fibrous suspensions, and the low

activity of cellulase enzymes. Combinations of enzyme dose and treatment conditions

were shown to improve flow properties and pumping of acid pretreated sugarcane

bagasse slurries (10% dry weight). Low levels of cellulase enzyme (0.1 and 0.5 FPU/g

dry weight acid pretreated bagasse) were found to reduce the viscosity of these slurries

by 77-95% after 6 h of incubation, solubilizing at least 3.5% of the bagasse dry weight.

Flow of slurries through small funnels was a useful predictor of success with centrifugal

and diaphragm pumps. Equations were derived that describe changes in viscosity and

solubilized sugars as a function of time and cellulase dosage. Blending of acid

pretreated bagasse (10% dry weight) with suspensions of acid pretreated bagasse

(10% dry weight) that had been previously digested with cellulase enzymes (low

viscosity) did not increase viscosity in a linear fashion. Viscosity of these mixtures

remained relatively constant until a threshold level of new fiber was reached, followed

by a rapid increase with further additions. Up to 35% of fresh acid pretreated bagasse

could be blended with enzyme-digested fiber (5.0 FPU/g dry weight acid pretreated

fiber; 6 h) with only a modest increase in viscosity. A simple model is described to

explain this phenomenon. The smooth surfaces of enzyme-treated fiber are proposed to

hinder the frequency and extent of interactions between fibrils of fresh fiber particles

(acid pretreated) until a threshold concentration is achieved, after which fiber

interactions and viscosity increase dramatically. These results were used to model the

viscosity in an ideal continuous stirred tank reactor (liquefaction) as a function of

residence time and enzyme dosage.









A hydrolysate-resistant mutant of E. coli LY180 was selected by sequential

transfers in AM1 mineral salts medium containing hemicellulose hydrolysate from a

dilute phosphoric acid pretreatment (160C) of bagasse. The resulting strain was

designated MM160. Strain MM160 was also more resistant than the parent to individual

inhibitors such as furfural, 5-hydroxymethylfurfural, and acetate. With this mutant,

process steps such as the separation of hemicellulose hydrolysate and fiber after dilute

acid pretreatment (washing) and toxin removal from the hemicellulose hydrolysate were

not required prior to the fermentation. Pretreated bagasse was fermented in a single

vessel without separation using strain MM160. A liquefaction step with cellulase was

included to improve pumping and mixing (L+SScF), analogous to a corn ethanol

process. Furans and other compounds in the hemicellulose hydrolysate remained as a

barrier to potential contaminants during fermentation. Bagasse slurries containing 10%

and 14% dry weight (fiber plus solubles) were tested in fermentations using

pretreatment temperatures of 160-190C (1% phosphoric acid, 10 min). Saccharification

efficiency and inhibitor production increased with treatment temperature. Bagasse

samples pretreated at temperatures below 190C were fermentable. The highest yields

and titers were obtained after 122 h of incubation using 14% dry weight slurries of

pretreated bagasse (180C), 0.21 g ethanol/g bagasse (dry weight) and 30 g/L,

respectively. These results are similar to those that have been previously reported using

pretreatments with ammonia and other bases in SScF processes.









CHAPTER 1
INTRODUCTION

Lignocellulosic Biomass

The U.S. consumes 20 million barrels of petroleum daily of which 65% is imported

(U.S. Energy Information Agency, 2005). In 2005, 71% of the total petroleum consumed

was for the transportation and industrial sector. Most of the imported oil comes from

politically unstable countries which can cause considerable fluctuations in the price of

crude oil (U.S. Energy Information Agency, 2005). The price of crude oil reached an all-

time high of $147 per barrel in 2008 due to concern over Iran's missile testing as well as

other factors. In addition to price fluctuations, the supply of oil could be disrupted based

on changes in foreign relations and the finite supply of oil. The burning of fossil fuels

also generates greenhouse gases which are believed to be the cause of global

warming. To gain fuel independence and protect our environment, policy makers are

working on implementing new laws. The Energy Policy Act of 2005 passed by the U.S.

Congress mandates that 7.5 billion gallons of alternative fuels be produced per year by

2012. In 2007, Congress also mandated that 36 billion gallons of ethanol be produced

per year by 2022 of which 44% must come from cellulosic biomass.

Lignocellulosic biomass is the most abundant renewable biomass resource on

Earth. There are three major types of lignocellulosic biomass: agricultural crops and

residues, hardwoods and softwoods. Lignocellulosic biomass is composed of three

major polymers: hemicellulose, cellulose and lignin. The hemicellulose is a branched

heteropolymer made up of C5 and C6 sugars such as xylose, arabinose, glucose,

mannose and galactose. Cellulose is a highly crystalline homopolymer of glucose

connected by 3-1, 4-glycosidic bonds. The U.S. generates about 200 million dry tons of









biomass per year which could potentially be converted to 16 billion gallons of ethanol

based on a yield of 80 gallons of ethanol per dry ton (Perlack et al., 2005). World

production of ethanol from biomass doubled between 2004 and 2007 reaching 13.2

billion gallons (Perlack et al., 2005). Lignocellulosic biomass provides a relatively cheap

alternative to starch-based feedstocks and it does not interfere with food sources. The

conversion of lignocellulosic biomass to fuels and chemicals also provides a way to

reduce greenhouse gas emissions and our dependence on foreign oil sources.

However, the natural structure of lignocellulosic biomass is designed to be less

accessible to chemical and biological degradation than starch based biomass.

Therefore, a pretreatment step is required to effectively release the sugars present in

the lignocellulosic biomass. The production cost of biomass to ethanol processes is

mostly influenced by the overall conversion of sugars to ethanol and the ethanol

concentration in the fermentation broth (Ohgren et al., 2006). The key to making the

lignocellulose to ethanol process economically feasible lies in releasing the sugars for

conversion to ethanol in a cost effective way (Hahn-Hagerdal et al., 2006).

Comparison of Major Lignocellulosic Substrates

Lignocellulosic biomass contains 40-50% cellulose (a glucose polymer), 25-35%

hemicellulose (a sugar heteropolymer), 15-20% lignin (a non-fermentable phenyl-

propene unit), and lesser amounts of minerals, oils, soluble sugars and other

components. The compositions and percentages of cellulose, hemicellulose and lignin

vary from one plant species to another and also depend on the part of the plant, age of

the plant and growth conditions (Demirbas, 2005).

Grasses are generally composed of 25-50% cellulose, 25-50% hemicellulose, 10-

30% lignin, and other lesser constituents. Sugarcane is an example of a grass and the









bagasse is made up of 40-50% cellulose, 23-35% hemicellulose, and the remainder

(~30%) is made up of lignin with smaller amounts of minerals, waxes, and other

compounds. Hardwoods are made up of 45 (2)% cellulose, 30 (5)% hemicellulose, 20

(4)% lignin, 0.6 (0.2)% ash, and 5 (3)% extractives. Hardwood barks are made of

22-40% cellulose, 20-38% hemicellulose, 30-55% lignin, 0.8 (0.2)% ash, and 6 (2)%

extractives. Softwoods are made up of 42 (2)% cellulose, 27 (2)% hemicellulose, 28

( 3)% lignin, 0.5 (0.1)% ash, and 3 (2)% extractives (non-cell wall components such

as fatty acids and terpenoids). Softwood barks are made up of 18-38% cellulose, 15-

33% hemicellulose, 30-60% lignin, 0.8 (0.2)% ash, and 4 (2)% extractives. Some of

these ranges are large; however, it can generally be said that the cellulose content of

grasses, hardwood barks, and softwood barks is within the same range. Also, the

hemicellulose content of grasses is slightly higher than that of hardwood barks and

softwood barks. The lignin content of hardwood and softwood barks is much higher than

in grasses with the lignin content of softwood barks being the highest (Demirbas, 2005).

Lignin content in softwoods is generally higher than in hardwoods. Softwoods have

a more compact fibrous structure than hardwoods. Softwood lignins are more cross-

linked than hardwood lignins because the methoxy group in the 5-carbon position of the

phenylpropanoid unit is missing. Softwood hemicellulose has a higher proportion of

mannose and glucose units than hardwood hemicellulose which usually contains a

higher proportion of xylose units. Hemicellulose is more acetylated in hardwoods than in

softwoods (Jeffries, 1994).

Hemicelluloses of softwoods are primarily made of galactoglucomannans, and the

secondary cell wall can be made of up to 20% of this hemicellulose. The glucose









mannose ratio for gymnosperms is usually 1:3 (glucose:mannose). The extent of

galactose substitution can vary greatly from 1 galactose unit for every glucose unit to

0.1 galactose unit for every glucose unit. The galactose is attached to the glucomannan

backbone via 1>6 linkages. Also the galactoglucomannans have O-acetyl groups at the

C2 and C3 positions of mannose instead of hydroxyls. The degree of substitution is

usually one O-acetyl group per 3-4 mannose units. The second most abundant

hemicellulose in softwoods is arabinoglucuronoxylans. The structure of this

hemicellulose is very similar to glucuronoxylan discussed below for hardwoods except

that arabinose units are a-linked to the C2 and C3 positions of xylose units and occur

once in every ten xylose units (Jeffries, 1994).

Hemicelluloses of hardwoods are primarily made of glucuronoxylan. This

hemicellulose can make up 15-30% of the dry weight of the wood. The hemicellulose

backbone is made up of 3-1,4-linked xylose units with 4-O-methylglucuronic acid a-

linked at the C2 position of xylose. There are also acetyl esters at the C2 and C3

positions of the xylan backbone. The degree of O-acetyl group substitution is high in

glucoronoxylan with on average, 7-O-acetyl groups per 10 xylose units. The 4-0-

methylglucurnoic acid side chains occur about one in every ten xylose units (Jeffries,

1994).

Softwoods and hardwoods commonly contain glucomannans in their secondary

cell walls. However, in softwoods they make up 5-8% of the dry weight of the cell wall

and in hardwoods they make up 2-5% of the dry weight of the cell wall. Also, softwood

glucomannans have a-(1>6) linked galactose unit to the mannose backbone. The









structure of glucomannan hemicellulose is very similar to that of the

galactoglucomannan (Jeffries, 1994).

The main hemicellulose of hardwood and germinaceous plants is glucuronoxylan

and that of softwoods is glucomannan. The glucuronoxylan of hardwoods and

germinaceous plants is made up of a 3-(1>4)-D xylan backbone with substituents in

approximately 10% of the xylose units that could be a-(1>2)-D-glucuronic acid or a-

(1>2)-4-O-methyl-D-glucuronic acid. Unlike hardwood xylans, germinaceous plant and

softwood xylans have a-(1>3) linked arabinose furanose to the xylose backbone. As

mentioned earlier, hardwood xylans are highly acetylated through ester linkages at the

C2 and C3 positions of the xylan backbone. Grass xylans and softwood

galactomannans are acetylated but there is less acetylation in softwoods (Jeffries,

1994).

Lignocellulosic Biomass Pretreatment

Prior to the cellulose hydrolysis step, the lignocellulosic biomass must be

pretreated to render the cellulose more accessible to the cellulolytic enzymes.

Cellulases are a class of enzymes that include endocellulases and exocellulases.

Endocellulases cleave 3-1,4-glycosidic bonds found in the internal structure of cellulose

while exocellulases (i.e., cellobiohydrolases) cleave 3-1,4-glycosidic bonds found at the

ends of cellulose polymer chains. Cellobiase (3-glucosidase) is a third enzyme used in

the conversion of lignocellulosic biomass to soluble sugars and it cleaves cellobiose into

two glucose molecules.

There are a variety of pretreatment methods available that have been developed

each with its own advantages and disadvantages. An effective pretreatment disrupts

cell wall physical barriers as well as cellulose crystallinity and its association with lignin,









and produces monomeric hemicellulose sugars with minimal production of inhibitory

compounds (Mosier et al., 2005a). The pretreatment step is very important because it

affects downstream processes such as fermentation toxicity, enzymatic hydrolysis rates,

enzyme loadings, mixing power, product concentrations, product purification, waste

treatment demands, and power generation (Mosier et al., 2005a). Examples of

pretreatment methods include dilute acid, ammonia fiber explosion (AFEX), ammonia

recycle percolation, N-methylmorpholine-N-oxide, and alkali pretreatments. In acid

pretreatment, it is the hemicellulose that is primarily hydrolyzed, and in alkali catalyzed

pretreatment, it is the lignin that is primarily removed and a separate step is needed to

render the hemicellulose sugar monomers available by the use of hemicellulases

(Mosier et al, 2005a). Dilute acid pretreatment has the advantage of producing a

pentose-rich syrup (i.e., hydrolysate) that can be fermented by microorganisms without

the use of hemicellulases or cellulases. Sulfuric acid is most commonly used in dilute

acid pretreatment however, other acids can also be used such as phosphoric and

hydrochloric acid (Israilides et al., 1978). Phosphoric acid releases hemicellulose sugars

and improves enzymatic hydrolysis of the resultant solid with the production of lower

amounts of inhibitors compared to sulfuric acid (Geddes et al., 2010). In addition,

phosphoric acid can be used by biocatalysts for nutritional purposes at appropriate pH.

Other pretreatment methods are AFEX and water under pressure to penetrate the cell

wall of biomass, hydrate cellulose, and remove the hemicelluloses (Wyman et al.,

2005). Ammonia is able to be recycled which makes it economically attractive as a

pretreatment agent. AFEX treatment decreases the crystallinity of cellulose and

removes acetyl linkages of cellulose. This increases the ability of enzymes to access









cellulose during hydrolysis allowing the use of lower amounts of enzyme. The

association of lignin with cellulose is also disrupted with AFEX pretreatment and

hemicellulose oligomers are formed that need to be further hydrolyzed after

pretreatment (Holtzapple et al., 1991). Treatment of hemicellulose with water under

pressure seeks to reduce the concentration of sugar degradation products by avoiding

the formation of monosaccharides (Mosier et al., 2005b). This pretreatment requires the

addition of enzymes after pretreatment to release the monosaccharides in cellulose as

well as hemicellulose. This approach has the advantage of high yields due to the

reduction in sugar degradation (Mosier et al., 2005a).

The effectiveness of the pretreatment step can be assessed by two parameters:

the digestibility of the solid material by enzymatic hydrolysis and by the fermentability of

the liquid to determine its inhibitory potential on the fermenting organism. The

fermentation and enzymatic hydrolysis can be combined in simultaneous

saccharification and fermentation. The type of pretreatment will also vary depending on

the type of biomass. AFEX, wet oxidation, and liquid hot water pretreatments have been

shown to be effective for agricultural residues (Hahn-Hagerdal et al., 2006). Dilute acid

steam pretreatment is effective for forestry and agricultural residues (Hahn-Hagerdal et

al., 2006).

Inhibitor Formation and Detoxification Methods

Inhibitory compounds are formed as byproducts of certain types of pretreatment

and can inhibit fermentation of the solubilized sugars. Therefore, the solid fraction is

usually separated from the liquid hemicellulose hydrolysate fraction after pretreatment.

The cellulose is separately hydrolyzed by cellulolytic enzymes. The hemicellulose

hydrolysate can then be detoxified by various methods.









The future of the cellulosic biomass to ethanol industry depends on advances that

help to integrate steps in the overall ethanol production process. Integrating the

hydrolysis of cellulose and fermentation while eliminating the need for detoxification of

the hemicellulose hydrolysate and therefore the solid-liquid separation step prior to

detoxification would create a simpler, more economical process. This could be

accomplished by a combination of the following: developing a pretreatment process that

reduces the concentration of inhibitory compounds in the hydrolysate, by adapting a

microbial strain to the presence of the inhibitory compounds, by using a strain that can

detoxify the hydrolysate as well as hydrolyze the cellulose portion through native or

heterologous gene expression, and by using an organism that will ferment all the

pentoses and hexoses available in the lignocellulosic biomass.

Depending on the biomass and pretreatment type you can have varying degrees

of inhibitory compounds in the hydrolysate. Also, microorganisms vary in their

resistance to the different inhibitory compounds found in hemicellulose hydrolysates.

Sulfuric acid hydrolysates of hemicellulose lead to the formation and release of a variety

of inhibitory compounds such as furfural, acetic acid, formic acid, hydroxymethylfurfural,

levulinic acid, and phenolic compounds (Miller et al., 2009a, Zaldivar and Ingram,

1999a).

Acetic acid, typically the most abundant organic acid, can be a structural part of

the hemicellulose (acetylxylan) as well as a sugar degradation product and is released

during acid hydrolysis. At high temperatures and pressures, xylose is degraded to

furfural, and hexoses are degraded to hydroxymethyl furfural. Furfural is the most

abundant aldehyde found in sugarcane bagasse hydrolysates and was more toxic than









hydroxymethyl furfural when using ethanologenic E. coli strains KO11 and LY01 in rich

medium (Zaldivar and Ingram, 1999a). The combination of furfural with other organic

acids (Zaldivar and Ingram, 1999b) and aldehydes (Zaldivar and Ingram, 1999a) had a

synergistic inhibition of growth and fermentation (Zaldivar and Ingram, 1999b).

Further degradation products include formic acid which is formed from the

breakdown of furfural and hydroxymethyl furfural. Levulinic acid is a degradation

product of hydroxymethyl furfural. Phenolic compounds are produced from the

degradation of lignin during acid hydrolysis. Examples of inhibitory lignin degradation

compounds are vanillic acid, 4-hydroxybenzoic acid, and syringic acid. Hydrophobic

parts of proteins, enzymes, or membrane transport systems are possible sites of

inhibitory action of some of the toxic compounds found in sulfuric acid hydrolysates of

hemicelluloses (Palmqvist and Hahn-Hagerdal, 2000b). Studies using Saccharomyces

cerevisiae have shown that the more hydrophobic the inhibitory compound the higher

the inhibition of volumetric ethanol productivity (Palmqvist and Hahn-Hagerdal, 2000b).

Phenols and furans can have aldehyde, ketone, or acid functional groups. Higher pKa

value of the phenol hydroxyl group of aldehydes and ketones means that the phenolic

proton is not dissociated at neutral pH and this makes the phenol more hydrophobic

(Klinke et al., 2004). Researchers believe that the low molecular weight phenolic

compounds are highly inhibitory. It may be possible to specifically remove these

compounds alone and sufficiently reduce fermentation inhibition (Palmqvist and Hahn-

Hagerdal, 2000a).

There are different mechanisms of toxicity that have been proposed for the various

toxic compounds in hydrolysates and some mechanisms are still unknown. Aldehydes









are highly reactive chemical species and can disrupt the structure of nucleic acids,

proteins, and lipids in microorganisms (Boucher, 1972; Miller et al., 2009a). Phenolic

compounds are believed to disrupt the structural integrity of biological membranes

(Palmqvist and Hahn-Hagerdal, 2000b).

The temperature, reaction time, and acid concentration during dilute-acid

hydrolysis influence the concentration of degradation products of cellulose,

hemicellulose, and lignin in hemicellulose hydrolysates. The severity of different

pretreatment methods could be compared using a combined severity factor that takes

these three parameters into account and also the influence of hydrolysis pH (Chum,

1990). Using this as a starting point a researcher could find the optimal combined

severity factor for biomass pretreatment where sugar yield is optimized and inhibitor

formation is minimized.

There are different detoxification methods that can be grouped into two main

categories: biological, and chemical. Each method has its advantages and

disadvantages.

Biological Methods

Treatment of willow hemicellulose hydrolysate with peroxidase and laccase

enzymes from Trametes versicolor was found to increase the maximum ethanol of

Saccharomyces cerevisae (Jonsson et al., 1998). The researchers also noticed that

laccase selectively removed phenolic monomers and acids almost to completion.

Laccase detoxified the hydrolysate by the oxidative polymerization of low molecular

weight phenolic compounds. When Trichoderma reesei is added to hemicellulose

hydrolysates it can remove acetic acid, furfural, and benzoic acid (Palmqvist et al.,

1997). Instead of having a separate organism to detoxify the hydrolysate, such as T.









versicolor, the peroxidase and laccase genes could be cloned into the fermenting

organism. This would eliminate the need for separating the detoxification and

fermentation steps and the maintenance of a separate organism. Palmqvist et al. (1997)

used T. reesei to produce cellulolytic enzymes and to remove inhibitory compounds

found in the hydrolysate prior to ethanologenic fermentation by S. cerevisiae. However,

an improvement on this process would be the elimination of the use of T. reesei

because consumes pentose sugars found in the hemicellulose that could be used to

make ethanol. Furthermore, this organism needs its own set of nutritional requirements

that also increase costs. Instead, Larsson et al. (2001) developed an S. cerevisiae

strain that was tolerant to phenolic inhibitors in lignocellulose hydrolysate by

overexpressing the T. reesei laccase gene from a plasmid inside the yeast along with

the Sso2 protein which is involved in protein excretion. However, this involved the

maintenance of the plasmid and laccase requires oxygen for its activity. As an

alternative genetic approach, the gene(s) that E. coli uses to detoxify hydrolysate can

be identified and then overexpressed to potentially confer more resistance to

concentrated hydrolysate. The detoxification of phenol aldehydes by conversion to

alcohols in anaerobic cultures have also been shown in S. cerevisiae and in Klebsiella

pneumoniae (Larsson et al., 2001). Therefore, these genes could also be identified and

overexpressed to confer resistance to inhibitors in these ethanologenic bacteria.

Chemical Methods

Extraction of a spruce hydrolysate with ether at a pH of 2 increased ethanol yield

to that of a reference fermentation (Palmqvist and Hahn-Hagerdal, 2000a). The

inhibitory compounds extracted by the ether were acetic acid, formic acid, levulinic acid,

furfural, hydroxymethylfurfural and phenolic compounds. Extraction by ethyl acetate had









similar effects on hydrolysate fermentation and removed acetic acid, vanillin, 4-

hydroxybenzoic acid and completely removed furfural (Palmqvist and Hahn-Hagerdal,

2000a).

Detoxification of hydrolysate by alkali treatment is a popular method and is often

referred to as overliming. It involves increasing the pH of the hydrolysate to 9-10 with

Ca(OH)2 The pH of the hydrolysate is subsequently decreased to the desired

fermentation pH with acid. The detoxification is believed to be due to precipitation of

toxic compounds. Overliming decreases the concentration of Hibbert's ketones, furfural

and hydroxymethylfurfural in spruce hemicellulose hydrolysates (Palmqvist and Hahn-

Hagerdal, 2000a). Some researchers found that the detoxification of willow

hemicellulose hydrolysate was most effective when they used a combination of

overliming and heated sulfite treatment (Palmqvist and Hahn-Hagerdal, 2000a). The

combined detoxification removed Hibbert's ketones, aldehydes, and volatile compounds

(Palmqvist and Hahn-Hagerdal, 2000a). Also, treatment with ion-exchange resins and

charcoal can effectively remove inhibitors (Carvalho et al., 2006). Detoxification of

sugarcane bagasse hydrolysate with anion resins effectively removed 84% of the acetic

acid (Chandei, 2007).

Adaptive Evolution

An additional strategy in combating the inhibitory effects of hemicellulose

hydrolysate compounds is to adapt cells to the inhibitors. This can be done by

maintaining continuous cultures to metabolically evolve the cells in the presence of

hydrolysate and select for cells that are more robust having a high cell and ethanol

yield. Escherichia coli can naturally remove the inhibitory effects of furfural by reducing

it to furfuryl alcohol in anaerobic conditions or to furic acid in aerobic conditions. An









evolved E. coli mutated to silence NADPH-dependent oxidoreductase genes, yqhD and

dkgA, which are responsible for detoxifying furfural (Miller et al., 2009b). Yeasts

resistant to benzoic acid showed lower uptake rates of benzoic acid (Piper et al., 2001).

The change in uptake rate could be due to changes in membrane permeability (Piper et

al., 2001). The identification of potential genes involved in the conferred resistance to

inhibitors in hydrolysate could lead to a genetic approach to engineer a more robust

organism for hydrolysate fermentations with minimal or no prior detoxification.

Example Process for the Conversion of Lignocellulosic Biomass to Ethanol

A typical process to convert cellulosic biomass into ethanol involves the following

steps: 1. feedstock storage and handling, 2. size reduction, 3. pretreatment, 4.

detoxification of the hemicellulose fraction, 5. enzyme production, 6. hydrolysis of the

cellulose fraction, 7. hexose and pentose fermentation, and 8. distillation (Meade and

Chen, 1977). However, these steps can vary depending on how integrated the process

is and other factors. The following is a short description of the main steps involved in a

typical process and some important issues associated with them.

Feedstock Storage, Handling and Size Reduction

Assuming that a typical sugar mill produces about 285,000 tons of bagasse per

year and that 15-25% of the bagasse is not burned for energy (bagasse heating value is

18.1 GJ/ton), then a typical sugar mill can provide about 57,000 tons of bagasse per

year. Over a million tons of dry sugarcane bagasse is produced by the Florida sugar

industry. Sugarcane crushing season in Florida (and others) begins in October-

November and ends in April or May. Sugarcane deteriorates rapidly from the time it is

cut and harvested so harvesting, crushing and processing are done concurrently

(Meade and Chen, 1977). Therefore, sugarcane bagasse will be available to use at 50%









moisture during sugarcane crushing season and for the rest of the time (i.e., ~165 days)

bagasse needs to be stored. It has been proposed that bagasse can be dried to

approximately 46-47% moisture using waste heat in the boiler flue gases but this may

have high maintenance and energy costs associated with it. Solar drying may be a

cheaper alternative. The sugarcane bagasse may be kept on concrete slabs to reduce

the amount of standing water in the storage area and will prevent contact with dirt

(Kadam, 2000). The feedstock transport radius is an important parameter in determining

the costs associated with this unit.

Trucks are used to carry the bagasse bales and they can carry about 10 metric

tons of wet bagasse each. The amount to be carried by the trucks can vary from this

number but road weight limits must be followed. The handling areas consist of weigh

stations where the trucks are weighed and unloaded by forklifts, solids conveyor

systems, and others. Some of the bales are sent to storage bunkers while others are

put onto the conveyors where they would be unwrapped automatically. Bulldozers are

also needed to make piles of bagasse for future handling. The actual amount of trucks

and forklifts has to be worked out for the individual plant's storage requirements (Aden

et al., 2002).

The amount of size reduction for the bagasse is dependent on the type of

pretreatment to be used and the heat and mass transfer equations associated with the

pretreatment. The type of size reduction will also determine the amount of energy that

will need to be put into this unit operation. The type of size reduction will determine the

amount of energy that will need to be put into milling the bagasse to the right size as

well as the efficiency of pretreatment (Wyman, 2007). The bagasse is often milled to an









mean size of 15 mm. The bagasse is taken from the storage bunker via a screw feeder

and fed into the pretreatment process (Aden et al., 2002).

Cellulosic biomass is relatively cheap and an estimate average figure could be

given for sugarcane bagasse of about $30/dry ton. Increasing the amount of feedstock

tends to drop the processing costs. However, as the amount of feedstock in the

processing facility is increased, the cost associated with storing it, handling it, reducing

its size, and transporting it also increases. Therefore, based on cost analysis that has

been done for cellulosic biomass to fuels and chemicals, researchers have found that

these two cancel each other and that the overall cost stays relatively the same with the

amount of feedstock used (Wyman, 2003). However, when determining the costs

associated with the amount of feedstock that will be used it is important to know the cost

of feedstock transportation and what effect plant size has on the capital and operating

cost of the plant (Wyman, 2007).

Hemicellulose Hydrolysis

The effectiveness of a particular pretreatment depends on the type of biomass

used. Biomass is naturally resistant to breakdown into component sugar monomers,

because it has a structure that makes it resistant to enzymatic and chemical

degradation. Each biomass type has its own characteristic components and ratios of

components. The pretreatment should be chosen based on laboratory and pilot plant

data specific for each biomass.

The type of steel used for this process needs to be determined through

manufacturer's data on the acid and temperature resistance of the steel. The reduced

particle size bagasse is typically first steamed at low-pressure to 1000C. This provides

about 1/3 of the heat requirement for the hydrolysis reaction with a typical reactor









having residence time of 2 minutes, 1.1% sulfuric acid concentration, and a hydrolysis

temperature of 1900C. The bagasse is fed into the pre-steamer by screw conveyors that

have variable frequencies depending on the feed rate desired. The pre-steamer is a

horizontal vessel and has paddles inside to move the bagasse around. After the

bagasse has been pre-steamed it is taken out and added to another screw conveyor

that transfers the bagasse to a blow tank. The blow tank uses steam to inject the

bagasse into the hydrolysis reactor via a live bottom screw and pressurizes the bagasse

along the way. The reactor is designed to be 95% full and operates at 30% solids in the

reactor. The reactor is directly injected with steam to 1900C after the concentrated

sulfuric acid is added and diluted to 1.1% with water. The bagasse is flashed cooled

after the hemicellulose hydrolysis by discharging into a lower pressure flash tank for 15

minutes. Acetic acid, furfural, and hydroxymethylfurfural present in the acid pretreated

bagasse are removed in the form of evaporated condensate by flash cooling (Aden et

al., 2002).

Hemicellulose hydrolysis affects the concentration of sugars and products that can

be obtained. Also, depending on the microorganism used, the sugars released during

hemicellulose hydrolysis may or may not be used during fermentation with glucose

present in the medium. Hemicellulose hydrolysis also affects the quality of the lignin

containing solids that may be used as fuel for the process, as exported heat and

electricity, or used to make a specialty lignin-derived product.

Solid-Liquid Separations

There are different types of liquid solids separations. For slurries such as the ones

encountered in a biomass to ethanol process, filtration is the preferred technique for

separation. The particle size distribution of the slurry is very important for the









effectiveness of the filtration and must be accurately determined and optimized in the

field. There are three basic types of filtration: depth, cake, and membrane. The rate of

filtration is dependent on the force of the selected equipment type and on the resistance

of the cake that is being formed. The flow of filtrate through a cake follows Poiseuille's

equation and can be used to design the liquid solid separations. The contributing

variables to the filtration rate are filter surface area, pressure across the filter medium,

average specific cake resistance, weight of cake, resistance of the filter medium, and

viscosity (Vogel and Todaro, 1997).

There are three important steps involved in the separation of liquids from solids in

the biomass to ethanol process. The steps are the separation of the cellulose and lignin

containing solids after pretreatment from the hydrolysate liquor, the separation of

gypsum (if it is used for conditioning) from the conditioned hydrolysate, and the

dewatering of the lignin residue in the stillage after distillation. It is difficult to separate

the solid cellulose and lignin containing fraction from the hydrolysate liquor because the

solid behaves like a sponge and tends to absorb and retain water. Another complication

is the high temperature of the slurry and stillage that must be filtered. The high

temperature causes an increase in the corrosiveness of the slurry and stillage.

Therefore, the type of steel used for this process needs to be determined through

manufacturer's data on the acid and temperature resistance of the steel (Aden et al.,

2002).

Once the hydrolysate has been flashed cooled in the flash tank, the hydrolysate

slurry is approximately 21% insoluble solids. The slurry is taken to a pressure filter to

separate the solids containing the cellulose and lignin from the hemicellulose sugar and









toxic compound containing liquid. A pressure belt filter press was found to provide the

best solids recovery with the least amount of water (Aden et al., 2002). Compressed air

is passed through the hydrolysate slurry on a filter. This displaces liquid and increases

the solid content of the filter cake. Using less water used to wash the cake reduces the

energy required for downstream processes such as product recovery. The final cake is

conveyed to a slurry tank and mixed with conditioned hydrolysate liquor and recycled

water. Gypsum is separated by hydrocyclone and rotary drum filters in series. This

provides gypsum at around 80% solids (Aden et al., 2002).

Hydrolysate Conditioning

After the hemicellulose hydrolysis the hydrolysate could be flash cooled to

vaporize a large amount of water along with some acetic acid, furfural, and

hydroxymethylfurfural (Aden et al., 2002). Hydrolysate liquor that is to be conditioned is

pumped from the filtrate tanks of the belt filter and sent to a heat exchanger to be

cooled to 500C. The liquor is sent to a tank and lime is then added to raise the pH to 10

and allowed to stir for 1 hour. The conditioned hydrolysate is then adjusted to pH 4.5 in

a separate tank and stays there for 4 hours to allow the gypsum to crystallize (Aden et

al., 2002).

Cellulose Hydrolysis and Fermentation

Prior to carrying out full-scale fermentations in a plant, researchers must first do

laboratory and pilot plant testing. There are several steps involved in this process: 1)

developing of strains for physical and genetic improvements, 2) optimizing medium

conditions and culture conditions such as pH and temperature to the conditions desired,

3) determining the amount of oxygen supply required by the organisms during the

fermentation (although ideally you would want anaerobic conditions because it costs









less), 4) selecting the culture process operation mode, 5) measuring theological

properties of the culture, 6) modeling and formulation of process control strategies, and

7) manufacturing of the various bioreactors, sensors, agitators and other accessory

equipment needed for the process (Vogel and Todaro, 1997).

Fermentation can be batch or continuous. Batch fermentation has a low risk of

contamination but it is labor intensive, and has high investment costs. Batch

fermentation with recirculation of cells reduces water and energy demand but causes

inhibitor accumulation. Continuous fermentation reduces equipment costs and

decreases inhibition but has a higher potential for contamination than batch

fermentation. The ethanol productivity is influenced by product inhibition, cell mass

concentration, and lignocellulosic hydrolysate inhibitor concentrations. The seed culture

for fermentation can be grown in a separate fermentor(s) or it can come from cells that

have been recirculated from a previous fermentation. Cell recirculation can be

advantageous because it may produce cells that are better adapted to the inhibitory

environment of the fermentation (Palmqvist and Hahn-Hagerdal, 2000a).

Current research efforts are focused on the combination of detoxification, cellulose

hydrolysis and hemicellulose fermentation in one vessel, termed consolidated

bioprocessing (Lynd et al., 2005). Lignocellulose derived sugars are a mixture of

pentoses and hexoses with sugarcane bagasse hemicellulose hydrolysate containing

mostly xylose. Saccharomyces cerevisiae is the most commonly used industrial

microorganism in the conversion of lignocellulosic biomass to ethanol. It can ferment

hexoses to ethanol and some strains have a high tolerance to ethanol and to other

inhibitory compounds found in the hydrolysate (Hahn-Hagerdal et al., 2007). However,









S. cerevisiae is not naturally able to ferment pentoses. Therefore, the pentose sugars

are not used. However, recombinant yeast strains have been developed that can

ferment both hexoses and pentoses (Hahn-Hagerdal et al, 2007). The organisms, E.

coli and Klebsiella oxytoca, can naturally ferment both pentoses and hexoses and

strains have been developed through the introduction of ethanologenic genes from

Zymomonas mobilis that can anaerobically ferment the mixed sugars with ethanol as

the primary fermentation product (Jarboe et al., 2007; Wood and Ingram, 1992; Yomano

et al., 2009).

Distillation

Distillation is used to separate components in a liquid solution. The distillation

depends on the distribution of these components in the vapor and liquid phase.

Therefore, the composition of the vapor must be different from the composition of the

liquid with which it is in equilibrium at the boiling point of the liquid. There are different

types of distillation methods. The first type consists of boiling a liquid to produce vapor

in a single stage and collecting the condensing vapors without the return of any liquids

back to the single stage to contact the rising vapors. The second type of distillation

consists of boiling a liquid to produce vapor and collecting the condensate on a series of

trays in the distillation column that allow the flow of condensate counter-currently to the

rising vapors. The second type of distillation is the one used in industry to concentrate

volatile compounds such as ethanol and butanol and is called fractional distillation,

distillation with reflux, or rectification. For separating two compounds, the McCabe-

Thiele Method is used to calculate the theoretical number of trays necessary to achieve

a specific separation of a binary mixture in the second type of distillation (Geankoplis,

1993).









After fermentation is complete the cells are usually removed from the broth, and

the ethanol (or other volatile product such as butanol) in the fermentation supernatant

can be removed by distillation. It is desirable to have high ethanol concentration for

efficient recovery and to minimize the amount of minerals present and energy required

in this step. Some problems associated with distillation of ethanol are problems with an

azeotrope, unfavorable phase equilibrium as pure ethanol is approached, and mineral

fouling. Current research is focused on creating more efficient product recovery, and a

drying process has been developed that has the potential to be more desirable than

distillation because it avoids the azeotrope problem (Bungay, 2004).

In a typical process, ethanol is separated from the fermentation broth by distillation

and is subsequently dehydrated by molecular sieve technology. Distillation involves two

columns. The first column removes the dissolved C02 and most of the water. The

second column is used to concentrate the ethanol to a near azeotropic condition (96%

w/v). The 99.7% (w/v) ethanol is made with molecular sieve technology and sent to fuel

storage. The fermentors have vents, and the gases from these vents as well as the

gases from the first distillation column vent are scrubbed with water to remove any

ethanol in these streams. The scrubber effluent is then fed back into the first distillation

column. The still bottoms containing the lignin residue is de-watered by solid liquid

separation and burned to cogenerate steam and electricity for the process. The excess

electricity can be sold to the grid. The filtrate from the solid-liquid separation process is

used for recycling to other steps in the process, but some may be concentrated in a

multiple effect evaporator. The concentrated filtrate is used for fuel, and the condensate

is used as clean recycle water for the process. The amount of recycled water coming









from the stillage must be accurately controlled because this stream contains organic

salts and inorganic compounds that may inhibit the fermenting organisms (Aden et al.,

2002).

Wastewater Treatment

A variety of process streams are combined into the wastewater treatment. These

streams include flash vapor from various flash cooling steps in the process, condensate

from the belt pressure filter vent, multi effect evaporator condensate not recycled, and

boiler blowdown (Aden et al., 2002). These streams must be treated depending on

where they will go. The United States Environmental Protection Agency (US EPA) has

mandated strict discharge standards that must be met by industry. There are two

general types of wastewater treatment: biological and physical. The biological treatment

involves using microorganisms to reduce the pollutants in the waste streams to limits

acceptable by the EPA. The physical method of treatment involves physical procedures

such as stripping, ion exchange, and membrane separation to remove pollutants from

the streams (Vogel and Todaro, 1997).

The wastewater streams are combined and large particles are removed and sent

to a landfill. The remaining liquid is primarily composed of the evaporator condensate,

so it is first cooled in a heat exchanger and then anaerobically digested. The anaerobic

digestion produces methane that is combusted and then the wastewater is treated by

aerobic digestion. After the aerobic digestion step is complete, the sludge produced is

separated from the water. The water is recycled for use in the process and the biomass

containing sludge is combusted or sent back to the aerobic digestion stage. The total

chemical oxygen demand (COD) of the wastewater before treatment needs to be

determined so that calculations of the nutrient requirement and residence time for the









subsequent digestions can be made. The COD is the amount of oxygen required to

combust all soluble organic components to carbon dioxide, nitrate, and water (Aden et

al., 2002).

The amount of wastewater could be different depending on the final product

chosen. Some of the cells that are filtered from the aerobically digested water could be

used as fertilizer and that would change the amount of cells put back into the aerobic

digester or combusted.

Research Objectives

Reduce the Toxicity of Sugarcane Bagasse Hemicellulose Hydrolysate and
Eliminate the Need for Separation of the Solid Lignin-Cellulose Rich Residue

The dilute acid hydrolysis of hemicellulose can be catalyzed by most mineral

acids. The most commonly used acid is sulfuric acid, and it produces a hemicellulose

hydrolysate containing compounds that inhibit microbial fermentation. Therefore, it is

often necessary to separate and detoxify the hydrolysate prior to fermentation.

Dilute suspensions of corn stover, sugarcane bagasse, and sorghum straw have

been demonstrated to be effectively hydrolyzed by phosphoric acid. This study

investigates the effectiveness of phosphoric acid as a catalyst in the dilute acid

hydrolysis of sugarcane bagasse at elevated temperatures (140-1900C), higher solids

loading, and lower reaction time (10 minutes). The ability of the resulting hydrolysate to

be fermented by ethanologenic E. coli without a prior detoxification step is also

investigated.

Establish a Correlation between Solubilized Sugar, Viscosity, and Ability to Pump
for Pretreated and Enzymatically Digested Sugarcane Bagasse Slurries

The most cost effective process configurations in a lignocellulose to ethanol

process involve consolidation of unit operations. The use of phosphoric acid has been









shown to facilitate the fermentation of hemicellulose hydrolysates avoiding detoxification

and a costly liquid-solid separation of cellulose-enriched fiber. However, the

combination of solid and liquid in bagasse slurries presents mixing and pumping

challenges. Considerable bridging has been observed among fibers of sugarcane

bagasse that severely inhibits mixing and pumping. Attempts at pumping these slurries

typically results in dewatering and blockage.

This study investigates the effects of cellulase treatments on the relative viscosity

and flow properties of acid pretreated sugarcane bagasse fiber. Viscosity results will be

correlated with enzymatic sugar solubilization under a wide range of conditions.

Design a Unified Fermentation Process that Combines Cellulose Hydrolysis,
Hemicellulose Hydrolysate, and Fermentation in One Vessel

Research focused on the development of more robust biocatalysts with increased

resistance to inhibitors (e.g., furfural, hydroxymethylfurfural, acetate) present in

hemicellulose hydrolysates, have facilitated the fermentation of undetoxified

hydrolysate. The investigation of the genetic basis for the increased tolerance has

yielded potential genetic targets that can be used to further improve industrial

biocatalysts.

This study describes the development of the ethanologenic strain of E. coli,

MM160, which is resistant to inhibitors present in hemicellulose hydrolysates. A

fermentation process (L+SScF) will be developed that includes an initial liquefaction

step with subsequent fermentation of both hemicellulose and cellulose derived sugars.

Strain MM160 will be used in the investigation of the fermentation of varying solids

loadings of undetoxified phosphoric acid pretreated sugarcane bagasse slurries

pretreated at various temperatures with phosphoric acid.









CHAPTER 2
OPTIMIZING THE SACCHARIFICATION OF SUGARCANE BAGASSE USING DILUTE
PHOSPHORIC ACID FOLLOWED BY FUNGAL CELLULASES

Introduction

Lignocellulosic biomass (LCB) represents a potential carbohydrate feedstock for

the fermentative production of renewable fuels and chemicals (Demirbas, 2005; Hahn-

Hagerdal et al., 2006; Ohgren et al., 2007). Success of the emerging LCB to ethanol

industry will require process simplification to reduce the investment in construction and

a societal commitment to provide a stable market for renewable fuels. Much of the

process complexity results from fundamental differences in the biological function and

structure of the feedstock. Starch produced in grain serves as a temporary energy

storage polymer with glycosidic linkages that can be readily hydrolyzed to mobilize

glucose for germination and plant growth. In contrast, LCB has been designed by nature

to serve as a more permanent structural composite of lignin and carbohydrates that

resist deconstruction by bacteria and fungi. As a consequence, more aggressive

chemical and physical conditions and more complex processes are required for the

disassembly and fermentation of LCB-derived carbohydrates than are required for grain.

LCB of terrestrial plants is composed primarily of the thermoplastic lignin (15-25%)

and two carbohydrate polymers, cellulose (35-50%) and hemicellulose (20-35%)

(Demirbas, 2005; Jeffries, 1994). Both carbohydrates must be depolymerized to soluble

sugars prior to fermentation. Ethanol production from LCB by current small commercial


1This chapter was reprinted with permission from Elsevier. Geddes, C.C., Peterson, J.J., Roslander, C.,
Zacchi, G., Mullinnix, M.T., Shanmugam, K.T., Ingram, L.O., 2010. Optimizing the saccharification of
sugar cane bagasse using dilute phosphoric acid followed by fungal cellulases. Bioresour. Technol. 101,
1851-1857.









plants involves combinations of the following steps (Galbe and Zacchi, 2007; Ingram et

al., 1997, 1999): 1) size reduction; 2) chemical pretreatment with acid or base; 3)

hydrolysis of pentose-rich hemicellulose (enzymes or acid); 4) liquid-solid separation of

hemicellulose hydrolysate from insoluble fiber; 5) purification of the hemicellulose

sugars limingg or ion exchange); 6) enzymatic or acid hydrolysis of cellulose; 7)

separate or combined fermentation of cellulose-derived hexose and pentose-rich

(hemicellulose) hydrolysates; 8) liquid-solid separation of beer; 9) production (or

purchase) of enzymes for saccharification; 10) distillation; 11) waste-water treatment for

reuse or discharge; and 12) use of residual fiber solids as boiler fuel. Process

improvements are needed that eliminate some of these steps to reduce the complexity

and cost of LCB to ethanol processes.

LCB can be chemically treated to render cellulose more accessible to hydrolytic

enzymes (Chandra et al., 2007; Mosier et al., 2005a; Ohgren et al., 2007). While there

are a variety of methods available, steam treatments with dilute mineral acids have the

added advantage of depolymerizing hemicellulose into a pentose-rich hydrolysate of

sugar monomers. However, dilute acid hydrolysis is not as selective as enzymes and

side products are produced that retard microbial growth and fermentation (Palmqvist

and Hahn-Hagerdal, 2000a, 2000b; Zaldivar and Ingram, 1999a, 1999b). The presence

of these side products is the primary reason for a liquid-solid separation prior to

fermentation, a step that facilitates purification of hemicellulose sugars by over-liming,

charcoal, or ion exchange (Frazer and McCaskey, 1989; Ingram et al., 1999; Larsson et

al., 1999).









Most mineral acids can be used to catalyze hemicellulose hydrolysis (Bhandari et

al., 1984; Mosier et al., 2005a; Ohgren et al., 2005; Sassner et al., 2005; Schell et al.,

2003). Sulfuric acid has been widely investigated and is the least expensive. Phosphoric

acid is approximately 20-times more expensive but may also provide a co-product

opportunity as a valuable component of plant fertilizers. Initial studies concerning animal

feeds have shown that the hemicellulose of rye grass straw can be hydrolyzed with

phosphoric acid to release sugars which supported the growth of yeasts after

neutralization with ammonia (Israilides et al, 1978). More recent studies have

demonstrated that the hemicellulose in dilute suspensions of corn stover (Um et al.,

2003), sugarcane bagasse (Gamez et al., 2004, 2006), and sorghum straw (Vazquez et

al, 2007) can be effectively hydrolyzed in solutions containing 2-6% phosphoric acid at

100-1220C.

The lower level of toxins associated with phosphoric acid and the reduced need for

exotic metallurgy offer potential advantages that justify further investigation. In this part

of the study, investigations of phosphoric acid were extended to include higher

temperatures, high solids conditions, and lower acid usage with sugarcane bagasse as

a model substrate.

Materials and Methods

Materials

Sugarcane bagasse was generously provided by Florida Crystals Corporation

(Okeelanta, FL). Corn steep liquor (approximately 50% solids) was purchased from

Grain Processing Corporation (Muscatine, Iowa). Biocellulase W (164 mg protein/ml; 50

filter paper units/ml) was generously provided by Kerry Biosciences (Cork, Ireland).

Novozyme-188 3-glucosidase (277 cellobiase units/ml) was purchased from Sigma-









Aldrich (St. Louis, MO). Thymol, phosphoric acid, potassium hydroxide, and other salts

were purchased from Thermo-Fisher Scientific (Waltham, MA).

Standard Analysis of Sugarcane Bagasse Composition

Samples of bagasse were solubilized in sulfuric acid and analyzed using

procedures developed by the U. S. National Renewable Energy Laboratory (Sluiter et

al., 2008).

Effect of Phosphoric and Sulfuric Acid Concentrations at 1450C

Sugarcane bagasse (0.5 g) was placed in a 250-ml flask containing 100 ml of

dilute mineral acid (0 to 35 g/L). After soaking for 4 h (220C), excess water was removed

by filtering through a 150 mesh polyester fabric. The resulting acid-soaked bagasse

(88% moisture) was autoclaved in a Hirayama HA-305M Hiclave autoclave (Saitama,

Japan) for 1 h at 1450C (2 h cycle time). After cooling, sufficient deionized water was

added to adjust the contents to 120 g (total). Flask contents were mixed for 2 h and

sampled for analysis of sugars, furans, and organic acids.

Effect of Temperature on Steam Treatment of Sugarcane Bagasse with 1% (w/w)
H3P04

Sugarcane Bagasse (55% moisture) was soaked for 2 hours in a 10-fold excess of

phosphoric acid solution (1% w/w including moisture in bagasse). Acid-impregnated

bagasse was dewatered to 52% moisture using a Tinkturenpressen (Hubert Schwank,

Germany) hydraulic press (approximately 10.8 g acid/kg dry bagasse) prior to loading

into a steam gun. Design and operation of this steam gun have been previously

described (Palmqvist et al., 1996). Large valves allowed samples to be rapidly heated

(140C, 150C, 160C, 170C, 180C, and 190C) and discharged (10 min at

temperature; 12 min total cycle time). A total of 3 kg dry weight of acid-impregnated









bagasse was processed in six runs at each temperature, pooled, and mixed. Processed

samples were stored at 4C. Material balances were estimated using an average of the

bagasse dry weight introduced into the steam gun and the dry weight of the recovered

products as the reference value to express components as g/kg bagasse (dry weight).

Enzymatic Hydrolysis of Cellulose after Steam Treatment with Phosphoric Acid

Water-insoluble solids (WIS) remaining after phosphoric acid treatment were

determined after water washing. Digestibility of washed fiber was measured using

fungal Biocellulase W (164 mg protein/ml; 50 filter paper units/ml) and Novozyme 188 3-

glucosidase (277 cellobiase units/ml). Biocellulase W contains an unrefined

concentrated mixture of enzymes including cellulases and xylanases from fungal culture

broth. Samples of WIS (1.0 g) were added to 250-ml flasks containing a single glass

marble (mixing aid). Thymol (10 mg) was added as a preservative. Sufficient sodium

acetate buffer (50 mM; pH 4.8) containing Biocellulase W (0.5 FPU/g, 5.0 FPU/g, or

50.0 FPU/g) and 0.1 ml of Novozyme 188 was added to adjust the fluid contents to 100

g (total). Flasks were incubated at 500C and 120 RPM for 4 days. Samples (1 ml) were

removed at 24-h intervals after adjustment for any evaporative loss and stored at -20C

until analyzed.

Phosphoric Acid Method for Analysis of Bagasse Composition

Samples of sugarcane bagasse were hydrolyzed with dilute phosphoric acid and

cellulase to estimate structural carbohydrates and sugar yields. Samples (0.5 g for

cellulose analysis and 5.0 g for analysis of soluble sugars and hemicellulose sugars)

were weighed into 250-ml flasks and resuspended in sufficient dilute phosphoric acid

(2% w/w) or deionized water (water soluble sugars) to bring the contents to 100 g









(total). These were soaked for 3 h and autoclaved for 1 h (1450C dilute acid treatment;

1200C for water soluble sugars). After adjustment for evaporative losses, samples were

removed to measure soluble sugars deionizedd water; primarily glucose, fructose, and

sucrose) or hemicellulose sugars (2% w/w phosphoric acid; primarily xylose).

Cellulosic glucose content of the sugarcane bagasse was determined using flasks

containing a single marble to aid in mixing. Bagasse (0.5 g bagasse dry weight) was

autoclaved with 2% (w/w) phosphoric acid (100 g total weight of contents) for 1 h at

1450C. After cooling and adjustment with 45% (w/v) KOH to pH 5.0, thymol (10 mg),

Biocellulase W (1.0 ml; 100 FPU/g), Novozyme 188 3-glucosidase (0.5 ml), and water

were added to adjust the contents to 100 g (total). Cellulose digestion was carried out

for 72 h at 500C (100 RPM). Soluble sugars were analyzed after adjustment for any

evaporative losses. Structural carbohydrates were estimated from the sum of sugars

released during the combination of acid hydrolysis (steam treatment) and cellulase

digestion after subtraction of the water-solubilized sugars (sucrose, glucose and

fructose). Results are expressed as sugar or polymer per kg bagasse (dry weight).

Toxicity and Fermentation of Hemicellulose Hydrolysate (Phosphoric Acid)

The toxicity of hemicellulose hydrolysates prepared using the steam gun (1%

phosphoric acid at 140C, 150C, 170C, 180C, and 190C; 10 min treatments) was

evaluated by measuring ethanol production using E. coli LY160 (Yomano et al., 2008).

Inocula were grown in shake flasks (125 ml, 120 RPM, 16 h, 370C) containing 50 ml of

AM1 medium (Martinez et al., 2007) supplemented with xylose (50 g/L).

Hydrolysates were separated from fibrous residues using a Centra CF basket

centrifuge (International Equipment Company, Needham Heights, MA; 4 minutes at

3000 RPM) and neutralized to pH 6.5 with 45% (w/v) KOH. Fermentation tests were









conducted in small pH-controlled vessels with a 100-ml working volume (150 RPM,

370C) as previously described (Ohta et al., 1991). Fermentation broths contained 70 ml

of hydrolysate, 25 ml sterile corn steep liquor solution (10% dry weight), 0.15 ml AM1

trace elements (Martinez et al., 2007), 0.15 ml 1 M MgSO4*7H20, and 5 ml of inoculum

(grown in shake flasks at 120 RPM for 16 h). Samples were removed for analysis at 24-

h intervals.

Analytical Methods

Moisture content was determined using a Kern model MLB 50-3 moisture analyzer

(Balingen, Germany). Sugars were measured by high-performance liquid

chromatography (HPLC) using an Agilent Technologies 1200 series HPLC system

(Santa Clara, CA) equipped with a model G1314B refractive index detector. Sugars

were separated using a BioRad (Hercules, CA) Aminex HPX-87P ion exclusion column

(300 x 7.8 mm i.d.) fitted with a Phenomenex (Torrance, CA) Carboh-Ca 4 guard

column (4 x 3 mm). Sugars were analyzed at 800C using nano-pure water as the mobile

phase (0.6 ml/min). Organic acids and furans were measured by HPLC using an Agilent

Technologies 1200 series HPLC system equipped with dual detectors (refractive index

and UV210nm) and a BioRad Aminex HPX-87H column (45C; 4 mM H 2SO4 as the mobile

phase, 0.4 ml/min). Total furans furfurall plus hydroxymethyl furfural) were measured by

ultraviolet absorption (Martinez et al., 2000b). Ethanol was measured using an Agilent

Technologies 6890N Network gas chromatography system equipped with a wide bore

HP-PLOT Q column (0.5 mm diameter x 30 meters; J&W Scientific, Folsom, CA).

Statistical Analysis

Results are reported as means with standard deviations (n=3) unless otherwise

indicated. Graphpad Prism (Graphpad Software, San Diego, CA) was used to perform









two-way ANOVA (analysis of variance) with a 95% confidence level. A student t-test

was used to compare the experimental results for the two methods of compositional

analysis (Table 2-1). Differences in means were judged significant when P values for

the null hypothesis were 0.05 or less. Data for soluble sugars released at different

treatment temperatures (Figure 2-3A) was fitted to a linear regression and plotted with

95% confidence intervals (dashed lines).

Results and Discussion

Effects of Phosphoric and Sulfuric Acid on Hemicellulose Hydrolysis (1450C, 1 h)

Initial tests were conducted with excess acid (approximately 88% moisture) for 1 h

at 1450C (Figure 2-1A). Under these conditions, hydrolysis was concentration

dependent from 0-1.0% (w/w) phosphoric acid. Steam treatments with greater than or

equal to 1% (w/w) phosphoric acid and above solubilized approximately 76% of the

xylan sugars based on the NREL analysis of composition (Table 2-1). Yields of

individual sugars (Figure 2-1 B) remained essentially constant for all phosphoric acid

concentrations above 1% (w/v) (p > 0.05), less than half the optimal phosphoric acid

concentration needed for pretreatment at 100-1200C (Gamez et al., 2004, 2006;

Israilides et al., 1978; Vazquez et al., 2007).

Sulfuric acid has been widely investigated for hemicellulose hydrolysis and was

included for comparison (Figures 2-1A and 2-1C). Although sulfuric is a stronger acid

than phosphoric, maximum sugar yields (total) for both were similar at 1450C (1 h),

25710 g sugar/kg bagasse (dry weight) and 24613 g sugar/kg bagasse (dry weight),

respectively (p > 0.05). As previously reported (Um et al., 2003), sulfuric was more

effective than phosphoric at low concentrations. The increase in glucose observed with

3.5% (w/w) sulfuric acid is probably due to cellulose hydrolysis (Figure 2-1C) and did









not occur with phosphoric acid. This increase in glucose was accompanied by a sharp

decline in xylose.

Differences in the concentrations of individual sugars with sulfuric and phosphoric

acid can be explained in large part by the formation of degradation products (Figure 2-

2). Furan dehydration products remained low with all concentrations of phosphoric acid,

approximately 1/3 the levels observed with sulfuric acid (primarily furfural from pentose

dehydration; p < 0.05). Other sugar degradation products (levulinic acid and formic acid)

were abundant in sulfuric acid hydrolysates and essentially absent with phosphoric acid

(p < 0.05). Acetic acid, a natural constituent of hemicellulose (Jeffries, 1994; Mitchell et

al., 1990), was present in hydrolysates with both acids. The higher levels of acetic acid

observed with sulfuric acid (p < 0.05) may also include products from sugar destruction,

consistent with the presence of formic and levulinic acids. Both acetate and the

degradation products of sugars represent potential inhibitors of microbial growth and

fermentation (Palmqvist and Hahn-Hagerdal, 2000a, 2000b; Zaldivar and Ingram,

1999a, 1999b; Zaldivar et al., 2000).

Increasing the Temperature to Reduce Treatment Time and Phosphoric Acid
Usage

Treatment times (1 h), phosphoric acid usage (73 g/kg bagasse dry weight), and

the high moisture content (88% moisture) used in our initial investigations are clearly

unsuitable for most practical applications. The high moisture content in particular would

greatly increase the energy required for processing. Based on previous studies of

others defining the quantitative relationships between sulfuric acid concentration, time,

and temperature (Chum et al, 1990), it seemed reasonable to assume that most of

these could be assigned values judged desirable for industrial application (50%









moisture, 10 g phosphoric acid/kg bagasse dry weight, and 10 min treatment time) and

optimized for treatment temperature. For these tests, acid usage near the desired value

of 1% (w/w) was conveniently obtained by soaking in 1% (w/w) phosphoric acid and

dewatering to approximately 50% solids using a hydraulic press. Tests have confirmed

that a similar dewatering of bagasse can be readily achieved with a Vincent Corporation

model CP-4 screw press (Tampa, FL). The steam gun provided rapid heating and

cooling (Palmqvist et al., 1996) and allowed reproducible use of a short treatment time

(10 min). Condensate accumulated within the steam gun during treatment and

increased the moisture content of the expelled samples to around 80%. Six steam

treatments (1 kg each) at each temperature were accumulated in the cyclone receiver.

These were mixed to provide a large pooled sample for further study.

The yield of sugar monomers was nearly linear with temperature from 1400C to

around 1800C (Figure 2-3A) and can be described using a linear regression (R2 =

0.9298; 95% confidence interval as dashed lines, y = 3.512x 359.8).

Although more scattered (Figure 2-3B), corresponding decreases (17-34%) in the

weight of cellulose-enriched residue (WIS) were also observed. Most of the acid-

solublized sugar was derived from hemicellulose with xylose dominating (Figure 2-3A).

The small increase in glucose at 1800C and 190C may reflect limited hydrolysis of

cellulose. Total sugar yields of 215-299 g/kg bagasse (untreated dry weight) were

obtained between 1600C and 190C.

Lower levels of products from sugar destruction were observed at all treatment

temperatures with phosphoric acid (Figure 2-3C) as compared to sulfuric acid (Figure 2-

2B) (Grohmann et al., 1985, 1986; Martinez et al., 2000a, 2001). Furfural production









was very low at temperatures of 160C and below, but increased at higher

temperatures. Hydroxymethyl furfural remained low with a small rise at 190C. Formic

acid, a minor component, increased from treatments at 140C to 170C and remained

constant while acetate (constituent of hemicellulose) levels increased with temperature

throughout the full range tested. Furfural and other sugar degradation products are

known to retard the growth and fermentation of many biocatalysts (Palmqvist and Hahn-

Hagerdal, 2000a, 2000b; Zaldivar and Ingram, 1999a, 1999b; Zaldivar et al., 2000). A

treatment temperature of 160C provided reasonable sugar yields with minimal side

products from sugar destruction.

Effect of Phosphoric Acid Pretreatment on the Enzymatic Hydrolysis of Bagasse

Table 2-2 contains the results from digestion of WIS with Biocellulase W after

phosphoric acid treatment at different temperatures. Two concentrations of cellulase

were investigated (0.5 FPU/g WIS and 50 FPU/g WIS). Results obtained after 96 h of

saccharification (g glucose/kg untreated bagasse (dry weight) are presented. More than

half of the enzymatic hydrolysis occurred during the initial 24-h period (not shown).

Cellulase activity is known to be nonlinear with respect to enzyme dosage (Bommarius

et al., 2008; Xu and Ding, 2006). The 100-fold difference in enzyme levels altered sugar

yields by 3-fold or less. At the high cellulase loading, near quantitive yield of glucose

was obtained for bagasse treated at 1600C or above (Table 2-2). The mean glucose

yield for steam treatments at 160-1900C (50 FPU/g WIS, 96 h; p > 0.05) was 42028

g/kg bagasse dry weight, similar to that measured using the NREL method and

phosphoric acid methods (Table 2-1). Bagasse samples treated at 1400C and 1500C









were significantly more recalcitrant to enzyme digestion at both enzyme loadings (Table

2-2) than bagasse treated at higher temperatures (p < 0.05).

Incubation of WIS with the high concentration of Biocellulase W (50 filter paper

units/g WIS) released up to 55 g/kg bagasse of additional hemicellulose sugar xylosee

and arabinose) (Table 2-2). Since Biocellulase W is an unrefined fungal broth, it is not

surprising to find that these hemicellulase activities are also present. Hemicellulose

sugars liberated by Biocellulase W declined with increasing treatment temperature,

consistent with lower residual levels of this polymer. Total sugar yields from acid

hydrolysis and enzyme treatment (Table 2-2) were essentially constant at 160C -

180C (p > 0.05), similar to the total sugar content estimated using the NREL

procedures for bagasse analysis (Table 2-1).

It is interesting to note that total sugar yields were 60% of the total in untreated

bagasse using the lowest concentration of Biocellulase W tested (0.5 filter paper units/g

WIS) and a treatment temperature of 160C and above. Assuming an overall efficiency

of 90% (combination of sugar conversion to ethanol and ethanol recovery),

approximately 254 L (67 gal U.S.) of ethanol could be produced/metric ton of dry

bagasse using phosphoric acid treatment and the lowest level of Biocellulase W (+3-

glucosidase). A maximum of 416 L (110 gal U.S.) of ethanol could be produced using

100-fold higher levels of cellulase enzymes.

Phosphoric Acid Method for Analysis of Bagasse Composition

The high yields of sugars from phosphoric acid-hydrolyzed hemicellulose and the

subsequent high yields of glucose from enzymatic digestion of residual fiber (WIS) were

investigated as a method to estimate bagasse composition and potential sugar yields.

For this analysis, a steam-treatment temperature of 1450C (1 h) and an enzyme loading









of 100 FPU/g of bagasse (dry weight) was used. Table 2-1 shows a comparison of

results from analyses using the NREL procedures in which samples are first dissolved

in concentrated sulfuric acid and the phosphoric acid method. Although the results for

most individual sugars (except galactose) were not statistically different (p > 0.05), all

values with the phosphoric acid method were lower than with the NREL method. The

sum of sugars measured with the phosphoric acid method was only 90% of the value

obtained for the NREL method, a difference judged to be significant (p < 0.05). Note

values reported with the NREL method also include assumptions and corrections for

sugar destruction that are absent in the phosphoric acid method. Values reported with

the phosphoric acid method represent sugars that can be recovered and are potentially

available for bioconversion.

Toxicity of Phosphoric Acid Hydrolysate

Phosphoric acid treatments were previously shown to produce lower levels of

inhibitory compounds than sulfuric acid at 100C-120C (Gamez et al., 2004, 2006;

Israilides et al., 1978; Um et al., 2003; Vazquez et al., 2007). This was confirmed for

higher temperatures by our studies (Table 2-3). The toxicity of hemicellulose

hydrolysates was tested in pH-controlled fermentations using an ethanologenic

derivative of E. coli K011 (Ohta et al., 1991), strain LY160 (Yomano et al., 2008).

Hemicellulose hydrolysate was prepared from each steam treatment (140C-190C; 1%

w/w phosphoric acid, 10 min) using a basket centrifuge to separate fiber solids.

Neutralization, nutrients, and inoculum resulted in a 30% (v/v) dilution of hydrolysate.

Concentrations of sugars and inhibitors in the fermentation broth are presented in Table

2-3.









All hydrolysates prepared at temperatures below 180C were fermented with

yields of 77-91% of the theoretical maximum (Table 2-3). Fermentation of hemicellulose

hydrolysate from 140C was near complete after 24 h. Hydrolysates from pretreatment

at higher temperatures exhibited significant lags in fermentation that were related to

treatment temperature and the abundance of potential inhibitors. Similar lags have been

previously observed with other microorganisms in response to the addition of furfural to

fermentation broths (Almeida et al., 2007; Zaldivar and Ingram, 1999a). Steam

treatment at 160C represented an exception, exhibiting a longer lag than hemicellulose

hydrolysate from 170C despite lower concentrations of furfural and acetate. Once

fermentation began, rate of ethanol production was surprisingly similar for all.

Hemicellulose hydrolysates from treatments at 180C and 190C were not fermented

during 1 week of incubation. Although the fermentation of phosphoric acid hydrolysates

without sugar purification was slow, further improvements may allow omission of early

liquid separations leading to process simplification.

Conclusions

Steam treatment of sugarcane bagasse with low level of phosphoric acid (1% w/w

of bagasse dry weight) at elevated temperatures is an effective method to hydrolyze

hemicellulose and increase the accessibility of cellulose for enzymatic digestion.

However, phosphoric acid ($1.75/kg) is currently 20 times more expensive than sulfuric

acid. At a usage rate of 10 kg per metric ton biomass dry weight (80 gal U.S. ethanol

per metric ton), phosphoric acid alone would cost $0.22 per gal ethanol. Thus it is

essential for phosphoric-based processes to recover most of the value of added

chemicals (phosphate, N, Mg, K, and trace metals) as a fertilizer, while minimizing

environmental impact of any residual phosphate. Calcium can be used to precipitate









around 80% of the phosphate. Remaining phosphate and mineral nutrients could be

used beneficially during irrigation of nearby energy crops. The added expense of

phosphoric acid may be further offset by a reduction in the cost of plant construction.

For temperatures and pressures used in this study, a variety of stainless steel alloys

should be acceptable (confirmed by coupon tests; unpublished).

Side reactions of sugars were minimal with phosphoric acid pretreatments

resulting in only low levels of potential inhibitors. Sugar production from hemicellulose

increased with temperature reaching a peak at 180C (300 g total sugar/kg bagasse).

Hemicellulose hydrolysates from treatment temperatures below 180C could be

fermented (slowly) by ethanologenic E. coli without further purification. With further

improvements, this may allow elimination of costly equipment for the separation of

hemicellulose hydrolysate from fiber and the need for separate fermentations. The

hemicellulose sugar mixture xylosee, glucose, arabinose, mannose, and galactose) was

effectively metabolized with fermentation yields of 77-91%. Low doses of cellulase

enzymes (Biocellulase W and Novozyme 188 3-glucosidase) were remarkably effective

in solubilizing glucan and residual xylan after phosphoric acid pretreatments when

tested as very dilute suspensions of fiber. From 23-45% of the glucan (100-200 g

glucose/kg bagasse) was solubilized using an enzyme loading of only 0.5 FPU/g of

steam-treated fiber (WIS).









Table 2-1. Sugar composition of sugarcane bagasse (% dry weight)
Sample Arabinose
(method) Glucose Xylose + Mannose Galactose Total
(method) + Mannose
FL bagasse 43.8 1.4 23.4 2.1 5.6 0.7 3.5 0.4 76.3 2.7
(NREL)
FL bagasse 41.2 1.7 23.0 2.7 3.8 2.7 2.0 0.5 70.1 4.2
(phosphoric)
Reprinted with permission from Elsevier. Geddes, C.C., Peterson, J.J., Roslander, C., Zacchi, G.,
Mullinnix, M.T., Shanmugam, K.T., Ingram, L.O., 2009. Optimizing the saccharification of sugar cane
bagasse using dilute phosphoric acid followed by fungal cellulases. Bioresour. Technol. 101, 1851-1857.
(Page 1853, Table 1). Note: Results shown as mean standard deviation (n=4).

Table 2-2. Solubilization of sugars from sugarcane bagasse by steam treatments with
phosphoric acid followed by 96 h enzyme (Biocellulase W) digestion
Steam Acid Biocellulase W + 3-glucosidase Total
treatment hydrolysateb (50 filter paper units/g WIS) (g sugar/kg
temperature (g sugar/kg Glucose Xylose+Arabinose C bagasse)
(C; 10 min) bagasse) (g sugar/kg (g sugar/kg
bagasse) bagasse)
140 130 +1 338 24 41 18 509 30
150 154 6 341 19 55 4 550 20
160 215 8 452 11 53 3 720 14
170 234 6 419 18 52 4 705 19
180 299 4 383 9 38 3 720 10
190 285 + 2 424 5 39 4 748 6
Biocellulase W + 3-glucosidase
(0.5 filter paper units/ g WIS)
140 130 +1 102 1 43 2 275 2
150 154 6 133 8 38 1 325 10
160 215 +8 185 5 37 2 437 10
170 234 6 176 3 29 3 439 7
180 299 +4 164 10 11 7 474 13
190 285 + 2 198 5 7 7 490 9
Reprinted with permission from Elsevier. Geddes, C.C., Peterson, J.J., Roslander, C., Zacchi, G.,
Mullinnix, M.T., Shanmugam, K.T., Ingram, L.O., 2009. Optimizing the saccharification of sugar cane
bagasse using dilute phosphoric acid followed by fungal cellulases. Bioresour. Technol. 101, 1851-1857.
(Page 1855, Table 2). a Mean standard deviation (n=3). b Total sugars released by steam treatment with
dilute phosphoric acid. Additional hemicellulose sugars released during enzymatic digestion.











Table 2-3. Composition and fermentation of hemicellulose hydrolysates
Steam Ethanol
treatment Max. yield


Fermentation


temperature Furfural (g/L) Acetate (g/L) Sugar (g/L) ethanol (g/L) (% theor.) lag (h)
1400C 0,0 1.73, 1.76 20.54, 21.25 9.52, 9.01 91,83 0
1500C 0.18, 0.23 2.13, 2.14 25.43, 26.14 10.6, 10.33 82, 77 24
1600C 0.50, 0.57 3.45, 3.46 34.95, 36.20 16.3, 16.42 91,89 96
1700C 1.31, 1.38 4.01,4.04 35.63, 36.47 14.4, 15.02 79,81 72
1800C 1.82, 1.88 3.44, 3.48 39.84, 40.25 No Growth No Growth >120
1900C 2.80, 2.94 3.90, 4.00 34.84, 35.12 No Growth No Growth >120
Reprinted with permission from Elsevier. Geddes, C.C., Peterson, J.J., Roslander, C., Zacchi, G.,
Mullinnix, M.T., Shanmugam, K.T., Ingram, L.O., 2009. Optimizing the saccharification of sugar cane
bagasse using dilute phosphoric acid followed by fungal cellulases. Bioresour. Technol. 101, 1851-1857.
(Page 1856, Table 3).Note: Results from two experiments (both shown).













300- 200 200
A B C

S150- 150-
U) /-U-- Xylose U)
S200 -- Arabinose -- Xylose
AGlucose -AGlucose
-0 100 -A- Galactose 100- A Galacose
SGalactose /
-0- Mannose --- Mannose
)100- l
1h 145C Hydrolysis (12% Solids) 1 hr 145yC Hydrolysis (12% Solids)

S-D]- Excess 1% H2S04 hydrolysis (12% Solids) C)

i 0 0

0 1 2 3 0 1 2 3 0 1 2 3
Acid (%, w/w) H3PO4 (%, w/w) H2SO4 (%, w/w)



Figure 2-1. Effect of acid concentration on monomer sugars from dilute acid treatments (1 h, 145C) with H3PO4 and
H2S04. Results shown as mean standard deviation (n=3): A) Total sugars, B) Individual sugars released by
dilute H3PO4, C) Individual sugars released by dilute H2S04. Reprinted with permission from Elsevier. Geddes,
C.C., Peterson, J.J., Roslander, C., Zacchi, G., Mullinnix, M.T., Shanmugam, K.T., Ingram, L.O., 2009.
Optimizing the saccharification of sugar cane bagasse using dilute phosphoric acid followed by fungal
cellulases. Bioresour. Technol. 101, 1851-1857. (Page 1853, Figure 1).













A 1.0% Phosphoric Acid B 1.0% Sulfuric Acid
a" 3.5% Phosphoric Acid 3.5% Sulfuric Acid
S50- 503.5% Sulfuc Acid


_040- _040.
I) I )

30- ~ 30-


20- 20-
0 0





Formic Acetic Levulinic Propionic Furans Formic Acetic Levulinic Propionic Furans


Figure 2-2. Effect of acid concentration on the production of side products furanss and organic acids) during steam
treatment (lh, 145C, n=3): A) Dilute H3PO4, and B) Dilute H2S04. Reprinted with permission from Elsevier.
Geddes, CC., Peterson, J.J., Roslander, C., Zacchi, G., Mullinnix, M.T., Shanmugam, K.T., Ingram, LO., 2009.
Optimizing the saccharification of sugar cane bagasse using dilute phosphoric acid followed by fungal
cellulases. Bioresour. Technol. 101, 1851-1857. (Page 1854, Figure 2).


















Xylose

-A Glucose
250 --Arabinose
Mannose
-- Total Sugars
200


150


100-


50-


50
140 150 160 170 180 190
Treatment Temperature ( C)


800




z600




400



200


140C 150C 160C 170C 180C 190C


s
, 20



a 15



S10
o



0
s 5-



0-


C Acetate
Furfural
Formate
HMF
Levulnic


140 150 160 170 180 190
Treatment Temperature ( C)


Figure 2-3. Effect of treatment temperature on bagasse hydrolysis (hemicellulose) and production of side products (1%

w/w H3PO4, 10 min): A) Hemicellulose sugars (n=3), B) Water insoluble solids (single value from pooled

sample), and C) Furans and organic acids (n=3). Reprinted with permission from Elsevier. Geddes, C.C.,

Peterson, J.J., Roslander, C., Zacchi, G., Mullinnix, M.T., Shanmugam, K.T., Ingram, L.O., 2009. Optimizing the

saccharification of sugar cane bagasse using dilute phosphoric acid followed by fungal cellulases. Bioresour.

Technol. 101, 1851-1857. (Page 1854, Figure 3).









CHAPTER 3
OPTIMIZING CELLULASE USAGE FOR IMPROVED MIXING AND RHEOLOGICAL
PROPERTIES OF ACID-PRETREATED SUGARCANE BAGASSE

Introduction

Lignocellulosic biomass (LCB) could serve as a carbohydrate feedstock to partially

replace petroleum-based fuels and chemicals (Demirbas, 2005; Hahn-Hagerdal et al.,

2006; Ohgren et al., 2007; Overend and Chornet, 1987). The LCB of terrestrial plants is

composed of the thermoplastic lignin (15-25%) and two carbohydrate polymers,

cellulose (35-50%) and hemicellulose (20-35%) (Demirbas, 2005; Jeffries, 1994).

Processes for depolymerization and fermentation of LCB are more complex and more

capital intensive than established technologies for cornstarch or cane hydrolysate.

Unlike starch, LCB has been designed by nature to serve as a structural element that

resists microbial deconstruction. Pretreatments such as dilute mineral acids or base

treatments are needed to render cellulose polymers accessible to enzymes (Chandra et

al., 2007; Mosier et al., 2005a; Ohgren et al., 2007). Steam treatment with dilute mineral

acids hydrolyzes hemicellulose into a pentose-rich hydrolysate (Ingram et al., 1997,

1999; Ohgren et al., 2007). This process is accompanied by side reactions and the

production of inhibitors furanss, organic acids and phenolics) that retard fermentation

(Palmqvist and Hahn-Hagerdal, 2000a, 2000b; Zaldivar and Ingram, 1999a, 1999b).

Progress has been made in developing pretreatment conditions with phosphoric

acid that minimize side products (Geddes et al., 2010) and comparatively more robust

biocatalysts that have increased resistance to furans (Heer et al., 2009; Liu et al., 2009;

Miller et al., 2009a, 2009b). These improvements could facilitate the simultaneous

fermentation of hemicellulose hydrolysate and cellulose-enriched fiber in a single

vessel, avoiding a complex (costly) liquid-solid separation. However, physical handling









of LCB fiber suspensions remains a critical issue. Considerable bridging occurs among

the fibers of sugarcane bagasse (dry solid or in an aqueous slurry; with or without acid

pretreatment) that severely limits mixing and pumping. At 10% or higher solids (dry

weight), fibrous suspensions of acid pretreated bagasse retains most of the water and

pour as a single tangled unit from a laboratory beaker (Dasari et al. 2009). Warwick et

al. (1985) reported that following mild acid treatment of lignocellulose, small acid

molecules penetrate cell wall capillaries (spaces between microfibrils and cellulose

molecules) in the amorphous regions. This external fibrillation greatly increases the

water-holding capacity of LCB slurries by enhancing the abundance of capillary-like

regions (external surface area) and potential bonding areas between fibrils and fibrils

and water. Attempts to pump slurries containing 10% solids content or higher at pilot

scale typically result in dewatering and blockage. Although it remains to be seen if this

problem persists in very large commercial scale plants, mixing and pumping of LCB

slurries represent significant challenges during pilot testing and scale-up.

Previous studies have investigated the effects of particle size on theological

properties of red-oak sawdust (Dasari and Berson, 2007; Rezania et al., 2009) and the

effect of initial solids loading on power consumption, glucose yield, and theological

properties of dilute acid pretreated corn stover slurries (Dasari et al., 2009). Decreasing

the particle size of red-oak sawdust from 590-850 tm to 33-75 tm appeared to improve

the efficiency of enzymatic saccharification by over 50% (i.e., conversion of cellulose to

glucose) as well as reduce viscosities by as much as 98% using an initial solids

concentration of 10% (w/w) (Dasari and Berson, 2007). A subsequent study by Rezania

et al. (2009) reported that reducing the particle size of red-oak sawdust to < 1 tm by









sonication did not improve glucose release by cellulases and increased the viscosity.

This increase in viscosity was proposed to result from dominant frictional effects at the

tested particle size range.

In this part of the study, the effect of cellulase treatment on the relative viscosity

and flow properties of acid pretreated bagasse fiber (10% dry weight slurries in

hemicellulose hydrolysate) without particle size reduction were investigated. Relative

viscosity (single phase exponential decay) was correlated with the extent of

saccharification under a wide range of conditions. These results were used to model

viscosity changes in an ideal continuous stirred tank reactor with different amounts of

cellulase.

Materials and Methods

Materials

Sugarcane bagasse (from milled sugarcane) was provided by the Florida Crystals

Corporation (Okeelanta, FL) without any further size reduction. The bagasse is fibrous

and composed of a range of particle sizes. Fine dust-like particles make up

approximately 10% of the dry weight and larger (25-50 mm) sized particles make up

approximately 40% of the dry weight of the bagasse. Kerry Biocellulase W (164 mg

protein/ml; 50 filter paper units/ml) was provided by Kerry Biosciences (Cork, Ireland).

Novozyme 188 3-glucosidase (277 cellobiase units/ml) was purchased from Sigma-

Aldrich (St. Louis, MO). Thymol, phosphoric acid and other chemicals were purchased

from Thermo-Fisher Scientific (Waltham, MA).

Dilute Acid Pretreatment of Sugarcane Bagasse

Sugarcane bagasse was received at approximately 50% moisture. Bagasse was

soaked in a 10-fold excess of 1% (w/w) phosphoric acid (2 h) and dewatered to 33% dry









weight using a Centra CF basket centrifuge (International Equipment Company,

Needham Heights, MA; 4 minutes at 3000 RPM). This acid-impregnated bagasse was

autoclaved (500 g batches divided among three 1-L Pyrex beakers) at 145C (1 h), and

cooled to room temperature. Sufficient deionized water was added to adjust the total

weight to 3 kg (6 times the initial dry weight of untreated bagasse). The resulting slurry

contained a total of 167 g/L acid pretreated bagasse fiber (approximately 112 g

insoluble fiber/liter volume). After soaking and manual mixing to allow equilibration (2 h),

most of the hydrolysate was removed by centrifugation. Resulting acid pretreated fiber

(33% dry weight) and hemicellulose hydrolysate were stored at 40C. Samples of acid

pretreated fiber were washed with water prior to carbohydrate analysis.

Saccharification with Biocellulase W and p-glucosidase

Saccharification of acid pretreated fiber was tested using Kerry Biocellulase W and

Novozyme 188 3-glucosidase. Samples of acid pretreated fiber (20 g) were added to

500-ml flasks containing thymol (10 mg) as a preservative. Sufficient pH-adjusted

hydrolysate (pH 3-8) containing Kerry Biocellulase W (0-5.0 FPU/g dry weight acid

pretreated fiber) and Novozyme 188 p-glucosidase (0-3 cellobiase units/g dry weight

acid pretreated fiber) was added to adjust the contents to 200 g (10% dry weight acid

pretreated fiber). A Kerry Biocellulase W loading of 5 FPU/g dry weight acid pretreated

fiber is equivalent to 0.1 ml/g dry weight acid pretreated fiber. Most experiments were

conducted at pH 5.0 (adjusted with 45% KOH) and 550C.

Flasks with enzymes and acid pretreated fiber were incubated at 300 RPM (25-

800C) for 1 h and at 200 RPM thereafter. After adjustment for evaporative loss with

deionized water, samples were removed and stored at -200C until analyzed. The extent









of saccharification was measured as enzyme-solubilized sugar after subtraction of

sugars present before enzyme addition.

Relative Viscosity Measurement

Measurements of relative viscosity were used to compare theological properties of

acid pretreated fiber slurries following enzymatic digestion using a Brookfield DV-II +

Pro Viscometer equipped with a T-bar (T-C spindle, 100 RPM). Although values are

reported in centipoise (cP), these are only useful on a comparative basis. The non-

Newtonian properties of this fibrous material preclude a more rigorous interpretation.

Values > 20,000 cP represent near immobilization of the spindle. The relative viscosity

of water, acetate buffer and hemicellulose hydrolysate were also measured for

comparison (1 cP for each).

Flow Property Test using Graded Funnels

Plastic laboratory funnels with internal stem diameters 7 mm, 12 mm and 17 mm

were used to compare the flow properties of acid pretreated fiber after various

treatments with Kerry Biocellulase W. Flow was assisted by gentle tapping. The slurries

(10% dry weight acid pretreated fiber) were found to either pass through the funnel or to

form a plug that resisted flow.

Carbohydrate Composition and Analyses

Carbohydrate composition of bagasse (as received), acid pretreated fiber, and

hemicellulose hydrolysate were determined as previously described (Geddes et al.,

2010). Moisture content was measured using a Kern model MLB 50-3 moisture analyzer

(Balingen, Germany). Sugars, organic acids, and furans were measured by high-

performance liquid chromatography (HPLC) using an Agilent Technologies 1200 series

HPLC system (Santa Clara, CA).









Statistical Analysis

Graphpad Prism (Graphpad Software, San Diego, CA) was used to derive

equations that simulate various relationships. This program was also used to perform

two-way ANOVA (analysis of variance) of compositional analysis using the two-tailed

student t-test. Differences in means were judged significant when P values for the null

hypothesis were 0.05 or less.

Results and Discussion

Composition

Bagasse samples were analyzed for carbohydrate composition over a 2-year

period before and after acid pretreatment (1% phosphoric acid, 1 h, 145C). Sugar

compositions are expressed as g/kg dry weight (Table 3-1). Steam pretreatment with

dilute phosphoric acid solubilized an mean of 360 g/kg bagasse dry weight. Analysis of

the acid pretreated fiber confirmed that the hemicellulose had been selectively

solubilized. Differences in composition were judged significant for all sugars except

galactose (p < 0.05). As expected, glucan content of the insoluble fiber was increased

by acid pre-treatment. Due to the mild conditions used, approximately 19% of the

xylose, 38% of the galactose and 15% of the arabinose remained associated with acid

pretreated fiber.

Hemicellulose hydrolysates were separated from acid pretreated fiber and also

analyzed (Table 3-1). These contained 35 g/L total sugar and low concentrations of

potential inhibitors of fermentation (4.6 g/L acetate, 0.5 g/L furfural and 0.3 g/L formate.

Soluble sugars recovered in the hydrolysate represented 21% of the initial bagasse dry

weight.









Effect of Cellulase on Relative Viscosity

The effect of incubation time on viscosity was examined using three cellulase

enzyme loadings (0.25 FPU, 0.5 FPU and 5.0 FPU/g dry weight acid pretreated fiber).

For the highest enzyme loading, the reduction in relative viscosity was nearly complete

after only 1 h (Figure 3-1A). Longer times were required for lower enzyme loadings.

Relative viscosities declined to plateau values that were inversely proportional to the

level of added cellulase, with no tendency to converge even after longer incubation

times. This plateau to a constant relative viscosity with continued incubation is similar to

saccharification and is not well-understood (Warwick et al., 1985). The relationship

between viscosity and incubation time for saccharification can be represented by a one

phase exponential decay for each enzyme loading (Figure 3-1A, 0.25 FPU-Equation 3-

1, 0.5 FPU-Equation 3-2, 5.0 FPU-Equation 3-3).

v= 23674e-0 978t +2196 (3-1)

v = 24366e-1592t +1504 (3-2)

v= 25614e 4050t + 255.1 (3-3)

For Equations 3-1 to 3-3, v represents the relative viscosity (cP) and t is

incubation time (hours). R-squared values were calculated as 0.991, 0.989 and 0.975

for 0.25, 0.5 and 5.0 FPU/g dry weight acid pretreated fiber, respectively, indicating

excellent agreement with experimental results.

Acid pretreated fiber was slurried in hemicellulose hydrolysate to simulate process

conditions in which solids and liquids were not separated. Acetate and phosphate

present in the hydrolysate served as buffers for pH adjustment. Although the tangled

mass of fiber is far from the ideal solutions described by viscosity theory, measurements









of relative viscosity can provide useful information regarding changes in fluid properties.

Preliminary experiments were conducted with a variety of cones, paddles and spindles

(data not shown). A small T-bar spindle was found to be the most useful. Slurries of

hydrolysate containing 10% (w/w) acid pretreated fiber were digested with various levels

of Biocellulase W (Figure 3-1B). Temperature and pH optima for fungal cellulases (pH

5.0 and 500C; Ou et al., 2009) were similar for Biocellulase W (pH 5.0 and 60C; Table

3-2). Prior to enzyme addition, the T-bar was unable to rotate and registered values

exceeding 20,000 cP. The extrapolated initial viscosity (t = 0) value from Figure 3-1A

(i.e., 25,870 cP) was used as a maximum value in Figure 3-1B. After 6 h incubation with

Biocellulase W, relative viscosity was reduced by 77% with an enzyme loading of 0.1

FPU/ g dry weight acid pretreated fiber, and by 95% with an enzyme loading of 0.5

FPU/g dry weight acid pretreated fiber. Previous studies have reported that theological

properties of cellulose derivatives are related to molecular structure parameters such as

molar mass and particle size (Clasen et al., 2001; Gautier et al., 1991). Viscosity

reduction can be accomplished by reducing the molar mass through enzymatic

degradation of polysaccharide chains. Enzymatic treatment of biomass disrupts the

interaction of fiber polymers such as cellulose chains creating smaller particles, which

also decrease the viscosity. The relationship between enzyme loading and relative

viscosity (6 h incubation) can be represented by a one phase exponential decay (Figure

3-1B, Equation 3-4):

v = 247986e 15751 +1049 (3-4)

In Equation 3-4 v represents the relative viscosity in centipoise and 1 is enzyme

loading (FPU/g dry weight acid pretreated fiber). The R-squared value was 0.9969,









indicating a good fit with experimental results. Confidence limits have been included for

each curve (p < 0.05). With this equation the viscosity of the acid pretreated fiber slurry

(10% solids) can be estimated for any enzyme loading (6 h incubation).

Effect of Cellulase Loading on Extent of Saccharification

Surprisingly little saccharification was required to reduce viscosity (Figures 3-1 B,

3-1C and 3-1D). After 6 h, a very low enzyme loading of 0.1 FPU/ g dry weight acid

pretreated fiber (3.5% of the fiber dry weight solubilized) reduced relative viscosity by

77% (Figure 3-1B). With 5.0 FPU/ g dry weight acid pretreated fiber (6 h), viscosity was

reduced by 99% accompanied by the saccharification of 17.6% of the dry weight. The

relationship between enzyme loading and saccharification (Figure 3-1B) and the time

course for saccharification (Figure 3-1C) can be represented by fourth (Equation 3-5)

and third (Equation 3-6) order polynomials, respectively. Confidence limits have been

included for each curve (p < 0.05).

s = 0.09071+ 49.80 -161.912 +194.713 32.8414 (3-5)

In Equation 3-5, s represents the amount of enzyme-solubilized sugars (% dry

weight acid pretreated fiber) and 1 is enzyme loading as FPU/g dry weight acid

pretreated fiber. The R-squared value is 0.9980, indicating an excellent agreement with

experimental results. The enzyme loading required for a desired sugar concentration (6

h of incubation) can be estimated using Equation3-5.

s = 0.6496 + 7.520t 0.8492t2 + 0.02458t3 (3-6)

s = 0.1498 + 2.414t 0.1735t2 + 0.00390 t3 (3-7)

s = 0.01837 +1.939t 0.1452t2 + 0.003399t3 (3-8)









In Equations 3-6 to 3-8, s represents the amount of enzyme-solubilized sugars (%

dry weight acid pretreated fiber) and t is time of enzymatic saccharification (hours). The

R- squared values are 0.99 for all three enzyme loadings (5.0, 0.5 and 0.25 FPU/ g dry

weight acid pretreated fiber), indicating excellent agreement with experimental results.

Using these equations, the amount of sugar that will be solubilized by a specified

enzyme loading and incubation time (< 24 h) can be estimated. Similar trends were

observed for individual sugars (Figure 3-1D). Under the mild treatment conditions used,

part of the hemicellulose remained associated with the fiber (Table 3-1). This

hemicellulose was solubilized during incubation with Biocellulase W consistent with the

presence of additional enzymatic activities (Figure 3-1D). Curves defining individual

sugars were not modeled.

Effect of Enzyme Treatment on Flow through Graded Funnels

The handling and transferring of fibrous slurries represent significant challenges

for LCB conversion to fuels and chemicals. Three different funnels with internal stem

diameters of 7 mm, 12 mm and 17 mm were used to compare the flow properties of

acid pretreated fiber slurries (10% w/w). Flow was tested before and after enzyme

treatments (Table 3-2). Acid pretreated fiber slurries failed to flow through all three

funnels prior to enzyme treatment. After 2 h of incubation, all enzyme concentrations

allowed the fiber slurries to flow through the 17 mm stem even though viscosity

measurements were above the measurable range 20,000 cP) in some cases. Only

the two highest enzyme concentrations (0.5 and 5.0 FPU/ g dry weight acid pretreated

fiber) permitted flow through the 12 mm stem. None permitted flow through the 7 mm

stem (2 h). After 6 h of incubation, the highest enzyme concentration permitted flow









through the 7 mm stem. Flow properties followed the same trends (pH, temperature,

enzyme dosage, time) observed for relative viscosity in most cases and were improved

by higher level of saccharification and longer incubation times. Flow through the 17 mm,

12 mm, and 7 mm stems occurred at relative viscosities of > 20,000 cP, < 3,000 cP, and

5 200 cP, respectively. These corresponded to the saccharification of approximately

1%, 5% and 17% of acid pretreated fiber.

More practical tests were conducted with acid and enzyme treated bagasse using

a centrifugal pump (Jabsco, White Plains, NY; Model 18690-0000,115 V; 7.2 AMPS; 1-

1/2 in. inlet diameter, % in outlet) and a pneumatic diaphragm pump (IDEXAodd Inc.,

Mansfield, Ohio; Sandpiper Model SIF Metallic Design Level 1; 1 in. inlet and outlet

diameter). Positive results for flow through the 12 mm funnel stem were found to be an

excellent predictor of successful pumping. Enzyme dose and treatment conditions can

be used in combination to improve flow properties and pumping of acid pretreated

sugarcane bagasse.

Correlation between the Extent of Saccharification and Relative Viscosity

Extent of saccharification and relative viscosity were measured under a variety of

conditions using six independent samples of bagasse. This data has been assembled

into a scatter plot (Figure 3-2). Relative viscosity was dramatically reduced by a small

amount of saccharification. With 5% saccharification, relative viscosity was reduced by

almost 90%. However, saccharification of atleast 13% of the dry weight was required to

achieve the lowest viscosity. An equation was developed to model this data (Figure 3-

2).

The decline in viscosity during saccharification was represented by a one phase

exponential decay (R-squared = 0.9316). Confidence limits were also included (p <









0.05). Equation 3-9 can be used to estimate the viscosity of the slurry based on

enzyme-solubilized sugar, a property that can be correlated with pumping (Table 3-2).

v = 25319e0 4412s +592.2 (3-9)

In Equation 3-9 v represents the relative viscosity in centipoise, and s is

solubilized sugar as a percentage of acid pretreated fiber. A relative viscosity of 3,000

cP or less was needed for flow through a funnel with a 12 mm ID stem (Table 3-2) and

is indicated by a horizontal dotted line on Figure 3-2. Reduction of viscosity to this level

required the solubilization of at least 50 g sugar/kg acid pretreated fiber.

Effect of Mixing Acid Pretreated Fiber (No Enzyme Digestion) with Enzyme-
digested Acid Pretreated Fiber (pH 5.0, 55 C, 6 h) on Viscosity

The physical appearance of acid pretreated fiber slurries was dramatically altered

by enzyme treatments. Initially, slurries with 10% fiber occupied the entire volume,

poured as a single tangled unit and failed to settle indicating extensive bridging and

strong interactions between fibers or between fibers and water. After enzyme digestion,

slurries readily settled and behaved as a suspension of independent particles with lower

viscosities.

Additional experiments were conducted to determine the effect of combining

enzyme-digested slurries of acid pretreated fiber and undigested slurries on the relative

viscosity of the mixture (Figure 3-3). Acid pretreated fiber slurries (10% dry weight fiber

in hydrolysate) were digested for 2 h and 6 h using different amounts of Biocellulase W

(0.25, 0.5 and 5 FPU/g dry weight acid pretreated fiber) and cooled to room temperature

to minimize further enzyme action. Relative viscosity was measured before and

immediately after mixing (within 1 min) with various amounts of acid pretreated fiber (no

enzyme treatment). Relative viscosities of all enzyme treated bagasse were low (200-









2,500 cP) in comparison to undigested materiae(20,000 cP). Addition of small

amounts of acid pretreated fiber (no enzyme digestion) had little effect on relative

viscosity until a threshold value was reached (Figure 3-3). At this point, further addition

of acid pretreated fiber resulted in an abrupt increase in viscosity. The proportion of acid

pretreated fiber (no enzyme digestion) that could be accommodated below each

threshold value varied considerably (5-35%) and was directly related to the initial

viscosity of enzyme-digested bagasse. Using 3,000 cP as a conservative maximum for

pumping based on funnel experiments (12 mm stem), up to 35% fresh acid pretreated

fiber could be blended with enzyme-digested acid pretreated fiber (5.0 FPU/g dry weight

acid pretreated fiber; 6 h). At 0.5 FPU/g dry weight acid pretreated fiber, up to 23%

undigested acid pretreated fiber could be accommodated. At 0.25 FPU/ g dry weight

acid pretreated fiber cellulase loadings, the addition of more than 5% to 6% undigested

acid pretreated fiber resulted in a dramatic rise in viscosity. The relationship between

acid pretreated fiber additions (fresh) to enzyme-digested fiber and relative viscosity can

each be represented by equations sigmoidall curves) for each level of enzyme (Figure

3-3, 0.25 FPU-Equation 3.10, 0.5 FPU-Equation 3-11, 5.0 FPU-Equation 3-12).

874046
v= 1324+ 874046 (3-10)
1 + 10(330 0067f)

313535.7
v = 224.3 3 (3-11)
1 +10(2 70 0028f)

645898.4
v= 811.6 (3-12)
1 + 10(5 670 091f)

In Equations 3-10-3-12, v represents the relative viscosity (cP), and f is the

fraction of undigested fiber as a percentage of total (10% dry weight). The R- squared

values are 0.9795, 0.9461 and 0.9946 for the equations above respectively, indicating









excellent agreement with experimental results. The large confidence limits for sigmoidal

curves obscured the curves and were omitted. Two-hour treatments were judged to

have too few data points to develop a model (Figure 3-3A).

These data indicate that acid pretreated fiber that has been partially digested with

enzymes has the ability to accommodate or passivate additions of fresh acid pretreated

fiber (no cellulase treatment; up to 30% of total fiber) with little increase in viscosity. This

observation together with the rapid decline in viscosity resulting from limited

saccharification suggests a simple mechanism. The viscosity of acid pretreated fiber is

proposed to result from the tangling interactions of surface-exposed micro-fibers as

observed previously (Rezania et al., 2009; Wood et al., 1997) and the bonding between

fibrils and water (water-holding capacity) in the amorphous regions of fibers (Warwick et

al., 1985). Digesting these small fibers with enzymes provides a smooth surface (Wood

et al., 1997), reducing viscosity with limited saccharification. These smooth enzyme-

treated particles would also serve as a diluent that physically hinders associations

between small fibers on the surface of fresh acid pretreated fiber (no enzymes) until the

threshold concentration for random associations is reached leading to increased

viscosity (Figure 3-3).

Modeling an Ideal Continuous Stirred Tank Reactor (CSTR) to Decrease Viscosity

The experimental data from Figure 3-2 and 3-3 were used to estimate the upper

bound viscosity of an ideal CSTR for liquefaction at different mean residence times (r).

The residence time distribution (RTD) of the reactor is a probability function describing

the length of time the fluid elements of the tank spend inside the reactor. The ideal

CSTR assumes that the material at the inlet is instantly and completely mixed into the









bulk material of the reactor and that the contents of the reactor have the same

composition as the outlet at all times. The ideal CSTR has an exponential residence

time distribution (E(t), Equation 3-13; Fogler, 1986):

1
E(t)= -eT
(3-13)

In Equation 3-13, c represents the mean residence time, defined by T = V/Q

where V is the volume of the tank and Q is the inlet volumetric flow rate, and t

represents residence time. The fraction of the reactor contents that has a retention time

between t and t+dt inside the reactor is given by E(t)dt. The fraction of the reactor

contents that has a retention time less than ti and greater than tl are given by Equations

3-14 and 3-15 respectively (Fogler, 1986).

lE(t)dt (3-14)


SE(t)dt = 1-- 1 E(t)dt
^Jo (3-15)

Using Equations 3-13-3-15 above, the fraction wi of the reactor contents with a

residence time t such that iAt < t < (i+1)At can be expressed by Equation 3-16.


w, =- -e T *e T
(3-16)

If the assumption that viscosity is additive is made, then the viscosity ([t) of the

reactor contents can be expressed in terms of the viscosity |i of fraction i using

Equation 3-17.


,o0 (3-17)









However, when Figure 3-3 data is plotted with Equation 3-17 for each enzyme

loading, the experimental data fall below the linear viscosity curve (Equation 3-17) until

a maximum viscosity value, pmax, is reached. Using Equations 3-1-3-3 Equation 3-19 is

obtained, where a, b and c are constants derived from the best fit model of the

experimental data (Figure 3-1A) for each enzyme loading using the constraint that the

initial viscosity for undigested material at t = 0 is the same for all three enzyme loadings.

This constraint allowed the derivation of a relative viscosity value (i.e., 25,870 cP) for

undigested material, which was not measurable with the instrumentation but could be

used to model an ideal CSTR.


=0o (3-18)

S= a +be sA (3-19)

Equation 3-20 is obtained by substituting Equation 3-19 for pi and Equation 3-16

for wi into Equation 3-18. Equation 3-20 can be rearranged and expressed by Equation

3-21.


(a< a+be-A), 1-e e
,=0 I (3-20)

At
1-e
/u t-e 32(3-21)

Taking the limit At 0 of Equation 3-21 gives -, so L'Hopital's Rule can be used

to obtain Equation 3-22.










-A
e T r


-e c+-
(3-22)

b
=> I+c (3-23)

Equation 3-23 can be used to estimate an upper bound viscosity in a CSTR with a

specific residence time and enzyme loading when the viscosity falls below a certain

pmax. The maximum viscosities that an upper bound viscosity could be establish for (i.e.,

pmax) at each enzyme loading were determined graphically by transposing Equation 3-

17 with Equations 3-10-3-12 and were 5,500 cP, 12,000 cP and 11,200 cP for 0.25

FPU, 0.5 FPU and 5 FPU respectively. Predicted upper bound values for viscosity are

given in Table 3-3. From this model, the predicted mean residence times to produce a

slurry that can be pumped (i.e., viscosity< 3,000 cP) are 30 h, 9 h and 2 h for 0.25 FPU,

0.5 FPU and 5 FPU respectively.

This model was experimentally verified to some extent by a manual simulation of

the CSTR for liquefaction (5 FPU cellulase/g fiber; 10% dry weight acid pretreated fiber;

550C; pH 5.0; 60 L working volume). At hourly intervals, 16.7% of the volume was

removed and replaced with new acid pretreated bagasse (and cellulase). After a 3-h

startup incubation, viscosity remained relatively constant (45834 arbitrary relative

units) for the subsequent 9 hours of the experiment. Samples were removed hourly

using a peristaltic pump confirming that the biomass suspension remained pumpable.

Conclusions

The simultaneous saccharification and co-fermentation (SScF) process requires

high concentrations of solids to produce economical levels of ethanol. For fibrous LCB,









these high solids are difficult to stir and transport. This study has demonstrated that low

levels of cellulase enzymes are sufficient to reduce viscosity and improve the flow

properties of acid pretreated sugarcane bagasse slurries. Flow properties and success

in pumping can be predicted by measuring either the extent of enzymatic hydrolysis

(soluble sugar), relative viscosity, or by a simple flow test using laboratory funnels.

Relatively little saccharification was needed to cause a large decline in viscosity. The

decline in viscosity was directly related to cellulase dosage and incubation times and

can be modeled as one phase exponential decays. Third and fourth order polynomials

were used to describe the extent of saccharification of acid pretreated fiber as a function

of enzyme dose and incubation time.

Similar relationships could be established for other types of LCB and treatments

using this approach. The addition of non-enzyme digested fiber to enzyme-digested

fiber slurries was found to have little effect on viscosity until a threshold amount was

reached. This phenomenon can be explained by a simple model in which surface-

exposed micro-fibers are digested by cellulases (limited saccharification) to create

smooth-surfaced particles that serve as barriers to tangling associations between

microfibrils of undigested particles. These results were used to model viscosity changes

(liquefaction) in an ideal continuous stirred tank reactor at different mean residence

times and enzyme dosages. Based on this model, continuous addition of enzyme,

makeup water and acid pretreated fiber to a mixed vessel with 2 h to 6 h mean

residence time (5 FPU/ g dry weight acid pretreated fiber, 600C, pH 5.0) should provide

a continuous supply of low viscosity slurry (10-15% dry weight) for fermentation,

analogous to the liquefaction step in corn dry milling.









Table 3-1. Sugar composition of sugarcane bagasse, washed acid pretreated fiber and
hemicellulose hydrolysate

Material Glucose Xylose Galactose Arabinosea Total
sugars
Bagasse as
received (g/kg) 387 20 212 20 26 11 34 26 659 40
received (g/kg)
Washed fiber after
pretreatmentc 593 +17 64 +19 15 +8 8 +10 680 36
(g/kg)
Hemicellulose
hdrolysated(/ 4+1 27+2 1+1 3+1 35+2
hydrolysate (g/L)
aArabinose may also include mannose and fructose which co-elute. b mean SD
(n=14). mean SD (n=18). dlnhibitors present in hemicellulose hydrolysate included
(g/L): furfural (0.49 0.10), hydroxymethylfurfural (0.02 + 0.04), format (0.27 + 0.04),
and acetate (4.58 0.23). An mean of 36% of the bagasse dry weight was solubilized
by acid pretreatment; mean SD (n=54).












Table 3-2. Effects of cellulase enzymes on enzyme-solubilized sugars and theological
properties


Cellulase level
(FPU/g dry
weight) and
incubation
conditions


) o
-5 Oo

I & I
E Z
N O,
tD -o
S0)0

L U
Q-


0
o




0)
t -
m



o




0
-)


_0


o 0)
-o




0



0)=





U-


Flow testing
(funnel stem
Treatment diameter in
time (h) mm)

7 12 17

2 N N N
2 N N Y
2 N N Y
2 N N Y
2 N Y Y
2 N Y Y
6 N N N
6 N N Y
6 N N Y
6 N Y Y
6 N Y Y
6 Y Y Y
2 N N N
2 N N Y
2 N Y Y
2 N N Y
2 N N Y
2 N N Y
2 N N N
2 N N N
6 N N Y
6 N Y Y
6 N Y Y
6 ND ND ND
6 N Y Y
6 ND ND ND
6 N N N
6 N N N
2 N N Y
2 N N Y
2 N N Y
2 N Y Y
2 N Y Y
2 N N Y
2 N N N
6 N N Y
6 N N Y
6 N Y Y
6 N Y Y
6 ND ND ND
6 ND ND ND
6 N N Y


Enzyme-solubilized
sugars (% dry
weight acid
pretreated fiber)


0.0 +0.0
1.8 0.3
2.0 0.0
4.1 +0.8
4.9 0.3
12.9 +1.2
0.0 +0.0
2.5 +0.0
3.5 0.2
5.4 0.4
6.8 0.1
17.6 0.4
1.4 0.1
3.7 0.2
8.5 3.4
ND
6.5 1.6
ND
1.9 0.5
0.5 +0.6
3.2 0.1
5.9 0.1
8.4 2.1
ND
6.7 0.1
ND
0.7 0.4
0.5 0.4
0.9 0.1
2.3 0.1
3.6 0.0
4.9 0.3
5.8 0.3
4.5 0.1
1.0+1.0
1.4 0.3
4.1 +0.3
7.5 0.1
11.8 0.4
ND
ND
2.0 0.1


Relative viscosity
(cP)


>20000
>20000
>20000
6000
1500
400
>20000
>20000
6000
2300
1300
200
>20000
7500
1500
2000
6000
>20000
>20000
>20000
5000
1400
500
ND
600
ND
>20000
>20000
>20000
15000
5000
1500
2000
>20000
>20000
>20000
8000
3000
1700
ND
ND
>20000


The N, Y, and ND indicate no flow through the funnel, flow through the funnel, and data
that was not determined respectively.










Table 3-3. Predicted upper bound viscosity for an ideal continuous stirred tank reactor
(CSTR)
Enzyme Viscosity at different mean residence times (T, hours)
loading
(FPU/g 0.5 1 2 3 4 5 6 7 8 9 10 20 30
fiber)
0.25 18096 14166 10206 8214 7016 6216 5643 5213 4879 4611 4392 3347 2976

0.50 15069 10903 7326 5721 4810 4222 3812 3510 3277 3093 2943 2245 2003

5.00 8723 5327 3069 2203 1744 1461 1268 1128 1022 939 872 568 464

Note: Italicized values are above the confidence limit (i.e., viscosity > Pmax) and
therefore cannot be used as upper bound viscosity predictions.

















-- 0 25 FPU/g dry WA acid pretreated fiber
-- 0 5 FPU/g dry WA acid pretreated fiber
-- 5 0 FPU/g dry WA acid pretreated fiber


0 -

00

0

00

0

00


3000

2800

2600

S2400

S2200

2000
, 1200

1000
< 800
600

400
200


20
20B -- Sdubilized Sugar
18 B o- scosity (cP)

16



12

10

8

6 -- --


2,

0 -- ---7 ---- 7--- 7-?
00 01 02 03 04 05 064 5 6
Enzyme Loading (FPU/g dry wt acid pretreated fiber)


C -A-0 25 FPU/g dry wt acid pretreated fiber
0 5 FPU/g dry wt acid pretreated fiber
--5 0 FPU/g dry wt acid pretreated fiber


-e- Celloblose
A Glucose
-Xylose
-Arabinose
Mannose


-I-

0 1 2 3 4 5 6 23 24 25 0 1 2 3 4 5 6 23 24 25
Time (hours) Time (hours)



Figure 3-1. Effect of cellulase enzyme loading on saccharification and viscosity: A)

effect of incubation time on viscosity (Equations 3-1,3-2 and 3-3), B) effect of

enzyme loading on viscosity and saccharification (6 h incubation; Equation 3-

4 and Equation 3-5 respectively), C) effect of incubation time on

saccharification (Equations 3-6, 3-7 and 3-8), D) enzyme-solubilized sugars.

Polynomial equations were developed that described saccharification. The

decline in viscosity was modeled as a one phase exponential decay.

Confidence limits (dashed lines) have been included for most curves (p <

0.05). The thick continuous lines were generated using model equations. For

saccharification, sugars present at zero time have been subtracted. Reported

sugars were produced solely by enzymatic action. A Kerry Biocellulase W

loading of 5 FPU/g dry weight acid pretreated fiber is equivalent to 0.1 ml/g

dry weight acid pretreated fiber.


_s




ng
IS

m~
-o


E5
S"R


0
00
0-
00

00
0

0 1 2 3 4 5 6 2324 25
Time (hours)


40




S30




1 Q 20




S10


30000
28000
26000
24000 ;u
22000 S
20000
18000
16000
14000
12000
10000
8000
-u
6000
4000
2000
0











30000


28000

-26000
0_
-24000
O
.0
2-22000
( -
,20000
>12000 -

o 10000
> 8000-
S Flow through
2. 1 17 mm funnel
M 6000-

r 4000
-------------------------------------
2000 Flow through
S -- .--------- m-mTu-nel
0 ----- -------- -

0 5 10 15 20 25 30 35
Enzyme-solubilized Sugar
(% dry wt acid pretreated fiber)


Figure 3-2. Scatter plot of viscosity versus saccharification (10% w/w slurries of acid
pretreated fiber in hemicellulose hydrolysate). Sugars present at zero time
have been subtracted. Reported sugars were produced solely by enzymatic
action. Dashed lines indicate the 95% confidence limits (p < 0.05). The
horizontal dotted line indicates the point at which flow through the 12 mm
funnel occurred. Above the horizontal dotted line, the slurry flowed through the
17 mm funnel but not the 12 mm funnel. Below the horizontal dotted line, the
slurry flowed through the 12 mm funnel. The thick continuous line was
generated using the model equation for a one phase exponential decay
(Equation 3-9). Flow through 12 mm diameter funnel stems was correlated
with a viscosity of 3,000 cP or less.












30000 A 0 25 FPU/g drywt acd retreated fiber 30 --025 FPU/gdry wt acd retreated fiber
-300- 0 5 FPU/g dry wt acid pretreated fiber 30000- 0 5 FPU/g dry wt acid pretreated fiber
28000- -- 5 FPU/g dry wt acid pretreated fiber 28000- -- 5 0 FPU/g dry wt acid pretreated fiber
S26000- D A o 26000- 0 A
24000- 24000-
S22000 22000-
.- 20000 20000
S18000- 18000
> 16000- 16000
0 14000- 14000- ;
12000- 12000-
10000- 10000
S8000- 8000-
6000- 6000
4000 4000
2000- 2000

0 10 20 30 40 50 0 10 20 30 40 50
Fiber Mixture (% undigested) Fiber Mixture (% undigested)



Figure 3-3. Effect of acid pretreated fiber additions on the viscosity of cellulase-
digested slurries containing 10% dry weight acid pretreated fiber. Enzyme-
digested slurries of acid pretreated fiber were prepared by incubating for A) 2
h and B) 6 h (Equations 3-10, 3-11 and 3-12) and cooled to room temperature
to slow enzymatic action. These were mixed with 10% dry weight slurries of
acid pretreated fiber that had not been treated with enzymes. Viscosities were
measured immediately (within 1 min). These data were modeled as equations
for sigmoid curves (Equations 3-10, 3-11 and 3-12), shown as thick black
lines. Insufficient data points were available to model the 2-h treatment (Fig.
3a). Values connected with thin solid lines are within range of instrumentation
(i.e., < 20,000 cP). Curves derived from Figure 3-1A were used to estimate
the value at the immeasurable point for each curve and plotted as open
symbols connected by a dashed arrow.









CHAPTER 4
A SIMPLIFIED PROCESS FOR ETHANOL PRODUCTION FROM SUGARCANE
BAGASSE USING HYDROLYSATE-RESISTANT ESCHERICHIA COLI STRAIN
MM160 AND PRETREATMENT WITH PHOSPHORIC ACID

Introduction

Lignocellulosic biomass represents a potential source of carbohydrates (cellulose

and hemicellulose) for microbial fermentation to ethanol and other chemicals. Steam

pretreatment with dilute mineral acids serves as an efficient approach to depolymerize

hemicellulose into monomeric sugars and to enhance fiber digestion by cellulase

enzymes (Mosier et al., 2005a; Ohgren et al., 2007). However, soluble side products

furfurall, 5-hydroxymethyl furfural, acetate, phenolics, and others) in dilute acid

hydrolysates inhibit microbial growth and retard fermentation (Mills et al., 2009;

Palmqvist and Hahn-Hagerdal, 2000). Potential engineering solutions to this problem

(separation of hemicellulose hydrolysate, washing of fibers, and mitigation of toxins) add

process complexity (Figures 4-1A and 4-1B). Efficient washing requires either

countercurrent systems that minimize dilution or energy intensive steps to re-

concentrate sugars (Sassner and Zacchi, 2008). Although pretreated and washed fiber

can be effectively fermented by yeast and bacterial biocatalysts with added cellulase

(Dien etal., 2008; Hahn-Hagerdal et al., 2007), mitigation of toxins in hemicellulose

hydrolysates generally requires overliming, charcoal, ion exchange, etc. prior to

fermentation (Larsson et al., 1999; Martinez et al., 2001; Mills et al, 2009).

Pretreatment with base is also effective in increasing the digestibility of

lignocellulose by enzymes with minimal formation of toxic side products (Lau et al.,

2009). Previous studies with ammonia have highlighted the effectiveness and potential

savings from the co-fermentation of cellulose and hemicellulose-derived sugars in a









single vessel (Lau and Dale, 2009; Sassner and Zacchi, 2008). Inhibitors formed during

dilute acid pretreatment hindered analogous co-fermentations of cellulose and

hemicellulose hydrolysates without additional steps for toxin mitigation (McMillan et al.,

1999).

The extent of inhibitor production is affected by the severity of steam pretreatment

conditions (Kabel et al., 2007) and by the choice of acid. Although sulfuric has been

most commonly investigated (Mosier et al., 2005a), weaker acids such as phosphoric

acid can also be used and produce lower levels of toxic side products and reduce the

need for exotic metal alloys (Geddes et al., 2010). Since mineral acids are not

consumed by the pretreatment process, the higher cost of phosphoric acid as compared

to sulfuric acid could be offset in part by recovery and reuse as a fertilizer.

Recent progress has been made in the development of more robust biocatalysts

with increased resistance to dilute acid hydrolysates of hemicellulose (Heer and Sauer,

2008; Liu et al., 2009) and to furans present in dilute acid hydrolysates (Heer et al.,

2009; Miller et al., 2009a). Several genes involved in resistance to furfural and 5-

hydroxymethylfurfural have been identified as NADP(H)-dependent oxidoreductases

(Laadan et al., 2008; Miller et al., 2009a).

In this part of the study, an ethanologenic Escherichia coli (strain MM160) with

increased resistance to inhibitors in hemicellulose hydrolysates is described. Using

phosphoric acid pretreatment and strain MM160, sugarcane bagasse hemicellulose

hydrolysate and enzyme-hydrolyzed cellulose could be effectively fermented in a single

vessel in a variation of the simultaneous saccharification and co-fermentation process

(SScF) that included a liquefaction step, termed L+SScF.









Materials and Methods


Materials

Sugarcane bagasse was provided by Florida Crystals Corporation (Okeelanta,

FL). Biocellulase W (164 mg protein/ml; 50 filter paper units/ml) was provided by Kerry

Biosciences (Cork, Ireland). Novozyme-188 3-glucosidase (277 cellobiase units/ml) was

purchased from Sigma-Aldrich (St. Louis, MO). Sulfuric acid hydrolysates of sugarcane

bagasse hemicellulose were provided by Verenium Corporation and contained from 50-

80 g monomer sugar/L (Boston, MA) (Martinez et al., 2001). Phosphoric acid

hydrolysates of sugarcane bagasse hemicellulose were prepared at the University of

Florida. Laboratory supplies and chemicals were purchased from Thermo-Fisher

Scientific (Waltham, MA).

Steam Treatment of Bagasse with 1% (w/w) H3P04

Bagasse (approximately 55% moisture) was soaked for 4 hours in a 14-fold

excess of phosphoric acid solution (1% w/w including moisture in bagasse), dewatered

to approximately 50% moisture using a model CP-4 screw press (Vincent Corporation,

Tampa, FL), and loaded into a steam reactor. This dewatered bagasse contained 10 g

phosphoric acid per kg dry weight. Design and operation of this steam reactor have

been previously described (Palmqvist et al., 1996). Large valves allowed samples to be

rapidly heated (160C, 170C, 180C, and 190C) and discharged (0.5 kg dry weight of

bagasse, 9.5 min at temperature, 10 min total cycle time). After steam treatment, fibrous

suspensions contained approximately 30% dry matter including condensate. A minimum

of 3 kg dry weight of acid-impregnated bagasse was processed at each temperature,

blended into a single sample, and used for fermentation.









Hemicellulose hydrolysate was also prepared from pretreated bagasse

(approximately 70% moisture) using a screw press. Emerging solids contained

approximately 48% moisture and were discarded. Fine particulates were removed from

the extruded liquid using a glass fiber filter (Whatman GF/D, 15 mm diameter, 27 pm

pore size). Resulting clarified hydrolysate was stored at4 and used for the selection

of resistant mutants and as a substrate for seed cultures.

Organisms and Growth Conditions

Strains used in this study are listed in Table 4-1. These were stored as frozen

stocks in 40% glycerol at -800C. Working stocks were maintained by daily serial

transfers (3% inoculum) into AM1 medium (Martinez et al., 2007) containing

hemicellulose hydrolysate (160C pretreatment). Fermentations (37C, 1 50 RPM) were

maintained at pH 6.5 by the automatic addition of 2 N KOH.

Genetic Methods

E. coli strain LY180 was the parent organism for this study (Table 4-1). This strain

was previously constructed from E. coli ATCC 9637 (Miller et al., 2009; Yomano, et al.,

2009). Additional genetic modifications were made to a hydrolysate-resistant derivative

of LY180 (strain MM105) to produce strain LY195 (carried out by Lorraine Yomano)

using standard protocols (Sambrook and Russell, 2001) and those provided by

manufacturers (Invitrogen, Carlsbad, CA; New England Biolabs, Ipswich, MA; Qiagen,

Valencia, CA; and Stratagene, La Jolla, CA). DNA sequencing was performed by the

University of Florida Interdisciplinary Center for Biotechnology Research.

Isolation of Hydrolysate-resistant Biocatalysts

Hydrolysate-resistant derivatives of LY180 were selected by sequentially

transferring broth cultures (3% inoculum) in AM1 mineral salts medium containing









dilutions of hemicellulose hydrolysate (160C pretreatment). Each day, hydrolysate was

neutralized by adding 5N ammonium hydroxide to a final concentration of 47 mM and

adjusting the pH to 6.3 using 45% (w/w) KOH. After filter sterilization (Nalgene PES

filter, 0.45 |tm pore), the hydrolysate was diluted as needed with water. Remaining

components of AM1 medium (NH4H2PO4, (NH4)2HP04, MgSO4*7H20, trace metals)

were then added. Xylose was added as needed to provide a monomer sugar

concentration of approximately 50 g/L. Cultures were incubated at 37(150 RPM) in

small pH-controlled fermentation vessels (250 ml). Cultures were transferred when

ethanol concentrations exceeded 5 g/L (approximately 24 h). The concentration of

hydrolysate was increased in the selection medium when at least 3 successive transfers

produced over 10 g ethanol/L in 24 h.

A resistant clone was isolated after 322 sequential transfers in sulfuric acid

hydrolysate and designated MM105. Extraneous DNA regions (non-coding FRT

recombinase sites) from prior genetic constructions were removed from MM105 to

produce LY195. LY195 was then subjected to further sequential transfers in AM1

medium containing hemicellulose hydrolysate produced at 1500C, and subsequently

from hydrolysate prepared at 1600C. After 139 serial transfers of LY195 in phosphoric

acid hydrolysate, a second resistant clone was isolated and designated strain MM160.

This clone is the product of over 2 years of subculturing (461 transfers) and more than

2,000 generations of selection.

Removal of Extraneous DNA Segments from Strain MM105

During the construction of LY180, non-coding FRT recombinase sites were left

behind in the chromosome at the site of gene deletions and these remained in the









hydrolysate-resistant mutant, MM105. Three of these extraneous segments (one each

in mgsA, ackA, and pflB) were removed from MM105 by double homologous

recombination (carried out by Lorraine Yomano) using a cat-sacB cassette and Red

recombinase as previously described (Datsenko and Wanner, 2000; Jantama et al.,

2008). Chloramphenicol resistance (cat) was used to select for the initial integration of

linear DNA. Resistance to sucrose was used to select for a second integration event in

which sacB was deleted. Primers and plasmids used during these constructions are

listed in Table 4-1.

Tolerance to Hydrolysate Toxins

Tolerance to selected compounds present in hemicellulose hydrolysate was tested

using 13 x 100 mm culture tubes containing 4 ml of AM1 medium (50 g/L xylose) as

described by Miller et al. (2009a). Tubes were inoculated at an initial density of 0.05

OD550nm. Growth was measured after incubation (37C, 60 RPM) for 48 h using a

Spectronic 20D+ spectrophotometer (Thermo-Fisher, Waltham, MA).

Liquefaction Followed by Simultaneous Saccharification and Co-fermentation
(L+SScF) of Acid-pretreated Sugarcane Bagasse

Sugarcane bagasse was pretreated with 1% (w/w) H3P04 at varying temperatures

(160C, 170C, 180C, and 190C). Moisture content was used to calculate the amount

required for a 1-L fermentation volume at 10%, 12%, or 14% steam pretreated bagasse

(total dry weight of fiber plus solubles). After transferring into 3-L fermentation vessels

(BioFlo 110, New Brunswick Scientific, Edison, NJ), sufficient deionized water was

added to adjust the total volume to approximately 900 ml. Ammonium hydroxide (9.4 ml;

5 N) and 45% (w/v) KOH were added to raise the pH to 5.0.









The resulting slurry was too viscous to mix without further treatment. Mixing was

improved by partial saccharification (liquefaction) using Kerry Biocellulase W (5 FPU/g

dry weight acid pretreated bagasse) and Novozyme 188 3-glucosidase (2.77 cellobiase

units/g dry weight acid pretreated bagasse) during a 6 h pre-incubation at 55C (250

RPM) prior to inoculation. After cooling to 370C and adjustment to pH 6.5 with 45% (w/v)

KOH, the remaining salts in AM1 medium (NH4H2PO4, (NH4)2HP04, MgSO4*7H20 and

trace elements) were added. Additional water was added to bring the total volume to

950 ml. The resulting fermentation broth was inoculated with 50 ml of seed culture.

Seed cultures of MM160 were grown in AM1 medium (3% inoculum from working

stock; 500-ml vessels, 37C, 200 RPM, 24 h) containing 40% phosphoric acid

hydrolysate (160C hydrolysate) and added xylose (5% w/v total monomer sugars).

LY180 was grown without hydrolysate. At the time of inoculation (5% inoculum), seed

cultures typically contained 6-10 g/L ethanol. Fermentations were monitored for 240 h

(37C, 200 RPM) and maintained at pH 6.5 by the automatic addition of KOH (45% w/v).

Analyses

Carbohydrate compositions were determined as previously described (Geddes et

al., 2010). Moisture content was determined using a Kern model MLB 50-3 moisture

analyzer (Balingen, Germany). Sugars, furans, and organic acids were analyzed using

an Agilent Technologies 1200 series HPLC system (Geddes et al., 2010). Ethanol was

measured using an Agilent Technologies 6890N Network gas chromatography system

(Geddes et al., 2010).









Statistical Analysis

Prism software (Graphpad, San Diego, CA) was used to perform two-way ANOVA

(analysis of variance) for comparisons. Differences were judged significant when P

values for the null hypothesis were < 0.05.

Results

Composition of Materials Used as Substrates for Fermentation

The composition of bagasse pretreated with steam and dilute phosphoric acid at

different temperatures is summarized in Table 4-2. Pretreated bagasse exiting the

reactor was approximately 30% dry weight (fiber plus solubles). This material was used

as the primary substrate for fermentation without washing, separation, or treatments to

reduce toxicity. Solubles and fiber were separated only for analyses. Solubles,

hemicellulose sugars, sugars from fiber digestion with excess cellulase, and inhibitors

all increased with increasing treatment temperature. The highest yield of total sugars

(pretreatment and cellulase digestion) was obtained for pretreatments at 180C and

190C (difference not significant), although the level of combined inhibitors was 28%

higher at 190C than at 180C. Total sugars recovered by phosphoric acid pretreatment

followed by digestion with excess cellulase (100 FPU/g dry weight acid pretreated fiber)

were equivalent to 70% of the bagasse dry weight (approximately 63% glycan) with a

maximum theoretical yield of 357 kg ethanol/metric ton bagasse (dry weight).

Tables 4-3 and 4-4 contains the results on the effect of enzyme digestion on

concentration of individual sugar and inhibitor in the hemicellulose hydrolysates. These

hydrolysates were used as a source of inhibitors to select resistant mutants. The

concentrations of all monomer sugars and inhibitors increased with treatment

temperature. A broad oligosaccharide peak was observed during HPLC sugar analysis









which decreased in size with increasing treatment temperature (data not shown). This

peak was resolved into monomers xylosee and glucose) and acetate by treatment with

Biocellulase W (Table 4-3), consistent with the presence of xylanase, 3-glucosidase,

and acetyl esterase activities. Enzyme treatment increased the concentration of

monomer sugars by 32% for hydrolysate prepared at 160C indicating an abundance of

oligosaccharides, but had little effect on hydrolysates prepared at 180C or 190C. The

concentration of acetate increased during incubation with Biocellulase W (Table 4-4).

Acetyl esters are natural constituents of hemicellulose, the primary source of this

inhibitor.

Many different sulfuric acid hydrolysates of sugarcane bagasse were provided by

Verenium Corporation and used for mutant selection during the initial year. Although the

details of preparation are proprietary, these hydrolysates contained monomer sugars

(80-110 g/L), furans (1-2 g/L), acetate (4-13 g/L), and 1% to 4% sulfuric acid (Martinez

et al., 2001).

Development of Hydrolysate-resistant Strain MM160

The genetics and construction of strain LY180 have been previously described

(Yomano et al. 2009; Table 4-1). This strain can efficiently ferment all laboratory sugars

that are constituents of hemicellulose. Using LY180 as the parent, resistant mutants

capable of growth in dilutions of bagasse hemicellulose hydrolysate were selected by

repeated subculture. After 322 transfers in dilutions of sulfuric acid hydrolysate (200-600

g/L in AM1 medium), a clone was isolated and designated MM105. Three extraneous

DNA regions (recombinase sites from the construction of LY180) were then removed

from strain MM105 by double homologous recombination, and the resulting strain

designated LY195. Sequential transfers of LY195 were continued using AM1 medium









containing phosphoric acid hydrolysates prepared at 1500C initially and then shifted to

hydrolysates prepared at 1600C after tolerance improved. Examples of improvements in

fermentation during selection are shown in Figure 4-2. After 139 sequential transfers in

phosphoric acid hydrolysates, a clone was isolated and designated MM160. Strain

MM160 grew and fermented well in AM1 medium containing 600 g/L phosphoric acid

hydrolysate (1600C), conditions that were toxic to the parent strain LY180 (Figure 4-2B).

Strains LY180 and MM160 were also compared for resistance to individual

chemicals present in acid hydrolysates (Figure 4-3). MM160 was more resistant to

furfural, 5-hydroxymethyl furfural (HMF), and acetate. Both strains were equally

sensitive to growth inhibition by format. Previous studies are consistent with furans and

acetate as the primary toxins in dilute acid hydrolysates of hemicellulose (Miller et al.,

2009a; Mills et al., 2009).

Liquefaction Prior to Simultaneous Saccharification and Co-fermentation
(L+SScF)

Phosphoric acid pretreated bagasse was fermented using a modified SScF

process without separation of the soluble hemicellulose sugars and inhibitors from fiber

or treatments to remove toxins (Figure 4-1 B). This procedure included an initial 6-h

enzyme treatment at 550C (pH 5) to reduce viscosity prior to inoculation (L+SScF),

analogous to the liquefaction step in corn ethanol processes (Figure 4-1C; Ingledew et

al., 2009). This liquefaction step doubled the initial level of monomer sugars available

for fermentation (Figure 4-4A) and tripled the level of acetate (Figure 4-4B) as

compared to an equivalent dilution (30% solids; 1:3 dilution at 10% solids) of

hemicellulose hydrolysate (Tables 4-3 and 4-4).









Figure 4-5 shows a comparison of L+SScF fermentations with different levels of

solids using bagasse pretreated at 1600C. With these, most of the ethanol was

produced during the initial 72 h and continued to increase slowly during further

incubation. Sugars present during fermentation (predominantly xylose and glucose)

represent a balance between depolymerization and metabolism. Glucose was used

preferentially. Xylose utilization declined with increasing amounts of bagasse and

inhibitors (Table 4-4; Figures 4-5B, 4-5C, and 4-5D). Part of the xylose remained

unfermented even after 240 h. Previous research with LY180 (parent) demonstrated

that furfural tolerance was also higher with glucose than with xylose (Miller et al.,

2009a).

Additional L+SScF experiments were conducted with 14% bagasse (dry weight of

fiber plus solubles) using higher pretreatment temperatures (Figure 4-6). Bagasse

pretreated at 1900C was toxic for MM160 and was not fermented. The highest titer

(Figure 4-6A) and highest yield of ethanol were produced using 14% bagasse (dry

weight) pretreated at 1800C (Table 4-5), although xylose was only partially metabolized

(Figure 4-6D). The slower fermentation of samples from higher treatment temperatures

is consistent with inhibition by toxins, which also increased with pretreatment

temperature (Table 4-2 and 4-4).

Significant amount of ethanol was produced from sugarcane bagasse by the

L+SScF process (Table 4-5), despite the lagging utilization of xylose. Under the

conditions tested, ethanol yields from bagasse ranged from 169 kg/metric ton to 207

kg/metric ton (51-63 gal/US ton), up to 57% of the maximum theoretical yield based on

composition (Table 4-2).









Discussion

Previous studies with base pretreatments have demonstrated the advantages of

low inhibitor production and co-fermentation of sugars derived from cellulose and

hemicellulose in a single vessel (Table 4-6). Base pretreatments are very effective in

opening structure and improving enzyme digestion but produce little fermentable sugar.

Saccharification of both cellulose and hemicellulose are dependent on added enzymes

with base treatments. Dilute acid hydrolysis (typically sulfuric) has the advantage of

substantially hydrolyzing hemicellulose into monomer sugars. This advantage is offset

in part by the production of fermentation inhibitors such as furans from side reactions

(Mills et al., 2009). A weaker acid (phosphoric acid) was used that produces lower level

of furans and other inhibitors (Geddes et al., 2010). With phosphoric acid pretreatments

of bagasse, xylanase and 3-glucosidase activities present in Biocellulase W were

needed to complete the hydrolysis of hemicellulose (Table 4-2). Even with the lower

level of inhibitors (Figure 4- 3B) formed by phosphoric acid pretreatment (as compared

to sulfuric), sufficient levels were present to inhibit fermentation by the parent strain,

LY180 (Figure 4-2B).

Processes based on dilute acid pretreatment and cellulase typically include

additional steps for liquid-solid separations and cleanup steps to remove inhibitors from

hemicellulose hydrolysates as illustrated in Figure 4-1A (McMillan et al., 1999). This

requirement for solid-liquid separation and further steps to remove toxins was

eliminated by developing a more resistant biocatalyst (E. coli MM160) and by

minimizing inhibitor production using phosphoric acid. A simplified process (L+SScF)

was developed in which pretreated bagasse was used for fermentation by strain MM160

without further purification (Figure 4-1B). Phosphoric acid pretreated bagasse was









partially liquefied by adding cellulase enzymes to improve mixing prior to inoculation. An

analogous liquefaction step is also used in corn ethanol processes (Figure 4-1C, similar

to the pre-saccharification proposed by Lau and Dale (2009) for AFEX-treated corn

stover. After liquefaction, the bagasse slurry containing hemicellulose and cellulose-

derived sugars was fermented by strain MM160 in a single vessel using mineral salts as

nutrients. Glucose and xylose were co-fermented effectively when the lowest level of

toxins was present (10% bagasse dry weight; 1600C pretreatment). Xylose utilization

lagged with increasing levels of inhibitors. Even with incomplete xylose fermentation, up

to 207 kg ethanol was produced per metric ton of bagasse (dry weight) representing

57% of the maximum theoretical yield based on carbohydrate analysis (Table 4-2).

Hydrolysate toxins remaining in the broth during fermentation may serve as a natural

barrier to retard the growth of adventitious contaminants.

The maximum yield (0.21 g ethanol/g bagasse dry weight) with phosphoric acid

pretreatment and MM160 was similar to that reported for ammonia fiber expansion

(AFEX) and caustic pretreatments of various biomass materials (Alizadeh et al, 2005;

Lau and Dale, 2009; Table 4-6). High pH pretreatment and phosphoric acid

pretreatment should not require the special alloys needed for sulfuric acid processes.

Ethanol titers with phosphoric acid pretreated bagasse (180C) and strain MM160

averaged 29 g/L, although MM160 can produce 65 g ethanol/L using pure xylose in the

laboratory (data not shown). With high pH pretreatments and paper sludge, ethanol

titers of 40-45 g/L have been reported in other studies (Table 4-6). Further

improvements in titer and volumetric productivity are needed for a phosphoric acid

based process.











Table 4-1. Strains,
Escherichia coli Strain,
Plasmid, or Primer
TOP10F'
LY180



MM105
LY190
LY191
LY192
LY193
LY194
LY195
MM160



Plasmids
pCR2.1-TOPO

pKD46
pL014162
ackA region
pL014810
pL014815

pLOI4823

mgsA region
pLOI3937
pLOI4819

pLOI4821

PflB-ycaK region
pL014813
pLOI4814
pLOI4822

Primers


plasmids and primers.
Relevant characteristics


Source or citation


F' (laclq Tnl0 tet)
AfrdBC::(Zm frg ce/Y, ), AldhA::(Zm frg
casABKo),AadhE::(Zm frg estZpp FRT), AackA::FRT,
AmgsA::FRT, pflB+ (pflA-FRT-ycaK), rrlE::(pdc adhA
adhBzm FRT)
Hydrolysate resistant mutant of LY180
MM105 mgsA-cat-sacB-mgsA', cat
LY190 AmgsA
LY191 ackA'-cat-sacB-ackA', cat
LY192 AackA
LY193 pflA '-cat-sacB-ycaK', cat
LY194 pflA-ycaK, native sequence
Hydrolysate resistant mutant of LY195
AfrdBC::(Zm frg ce/Yc ), AldhA::(Zm frg casABKo),
AadhE::(Zm frg estZpp FRT), AackA, AmgsA, pflB+,
rrlE::(pdc adhA adhBzm FRT)

bla kan lacZ Plac
Para Y P exo (Red recombinase), temperature-conditional
replicon, bla
Pacl flanked cat-sacB cassette, bla, cat

LY180 ackA'-FRT-ackA' in pCR2.1 TOPO, bla, kan
pLOI4810 inside out PCR, self ligated, internal ackA
deletion, bla, kan
pLOI4810 inside out PCR, pLOl4162 cat sacB Paclfrg, cat,
bla, kan

LY180 mgsA'-FRT-mgsA'in pCR2.1-TOPO, Ap, Kan
pLOI3937 inside out PCR, self ligated, internal mgsA
deletion, Ap, Kan
pLOI3937 inside out PCR, pLOl4162 cat sacB Pac/frg, cat,
bla, kan

LY180 pflA-trm-FRT-ycaK' in pCR2.1 TOPO, bla, kan
ATCC9637 pflA-ycaK' in pCR2.1 TOPO, bla, kan
pLOI4813 inside out PCR, pLOl4162 cat sacB Pac/frg,
cat, bla, kan
Sequence


Invitrogen
Yomano et al., 2009



This study
This study
This study
This study
This study
This study
This study
This study




Invitrogen
Datsenko and Wanner,
2000
Jantama et al., 2008

This study
This study

This study


This study
This study

This study


This study
This study
This study


ackA cloning, deletion and sequencing
ackA cloning For 5'- ctggttctgaactgcggtag
Rev 5 cgcgataaccagttcttcgt
ackA deletion Up 5'- gcatgagcgttgacgcaatc
Down 5'- gactcttccggcatagtctg
yfbV-pta sequencing For 5'- ggcgttgacatgcttcacct
Rev 5'- tcgcacgcacgatagtcgta
mgsA cloning, deletion and sequencing
mgsA cloning For 5'- tattgcgctggtggcacacg
Rev 5'- acggtccgcgagataacgct
mgsA deletion Up 5'- cagcaggttggcgcattgat
Down 5'- accggtagtgcctgttgcat
yccT-helD sequencing For 5'- atggcgatgcgacgccgatt
Rev 5'- aacacgctggccgaagttgc
pfl region cloning, deletion and sequencing
pflA-ycaK2 cloning For 5'- cagatgaacgacgagatcca
Rev 5'- gagctgcttgaacatgacac
pflA-ycaK deletion Up 5'- ggtgaacgctctcctgagta
Down 5'- gcagaatgaagcgcggaata


This study

This study

This study


This study

This study

This study


This study

This study











Table 4-1. Continued
Escherichia coli Strain,
Plasmid, or Primer
ycaO-focA sequencing

focA-pflB sequencing

pflB-pflA sequencing

pflA-ycaK sequencing

ycaO int sequencing

focA-pflB2 sequencing


Relevant characteristics


Source or citation


5'- cgctggcttctgcacttggt
5'- ggccattgcagcaggaagta
5'- ggcctataagccaggcgaga
5'- gtggaggtacgaccgaagga
5'- gcgttgcgctgtacggtatc
5'- gccgccggaagcgttcataa
5'- gcaccaacacggcctcagat
5'- gtgcgctccagaacttaacg
5'- tggcgtagcactggaacgta
5'- gctggcgatcttcttcctgt
5'- accactggcacaggcacaat
5'- aacaaaacttcaatctataa


This study

This study

This study

This study

This study

This study











Table 4-2. Components of sugarcane bagasse after steam pretreatment with
phosphoric acid.
Composition after pretreatmenta
Components (g/kg SD bagasse dry weight)
160C 170C 180C 190C
Hemicellulose hydrolysate a

Solubles after pretreatment 215 12 239 18 284 21 273 14

Sugar monomers after pretreatment 115 3 161 9 222 13 232 9

Sugar monomers after enzyme digestionb 133 7 197 12 241 17 241 10

Glycan (hemicellulose+cellulose) 121 6 179 11 219 15 219 9

Inhibitors after pretreatment 8 0 12 1 20 1 30 1

Inhibitors after enzyme digestionb 14 1 18 1 25 2 34 2

Ash 8 1 14 1 133 104

Lignin plus other (by difference)' 72 13 28 21 27 26 10 17

Acid-treated fiber

Insoluble fiber after pretreatmenta 774 24 750 14 713 27 716 8

Solubles after enzyme digestion 461 23 493 20 492 24 497 13

Sugar monomers after enzyme digestion 41519 445 14 457 23 472 12

Glycan (hemicellulose+cellulose) 377 17 405 13 415 21 429 11

Inhibitors after enzyme digestiond 8 3 4 2 4 4 3 2

Ash 80 21 52 5+4

Lignin plus other (by difference)' 381 30 339 19 289 34 279 14

Combined hemicellulose hydrolysate + fiber

Combined sugar monomers 548 20 642 19 698 28 713 14

Combined glycan 498 18 584 17 635 25 648 13

Combined inhibitors C 22 3 22 2 29 4 37 3

Combined ash 16 1 16 1 18 1 15 1

Combined lignin plus other (by difference)' 464 18 378 17 318 25 300 13
Theoretical ethanol (liters/tonne) 354 414 451 460

a After steam pretreatment, soluble (hemicellulose hydrolysate) and insoluble fiber fractions (washed) were separated prior to
analysis. Sugars represent observed monomers and dimers without corrections for degradation (rt 4). b Hemicellulose sugars after
treatment with Biocellulase W (20 ml/L, pH 5, 50C, 6 h) to hydrolyze oligosaccharides. Acetate (inhibitor) also increased during
enzyme digestion. Glycan was estimated by dividing total sugars by 1.1 to correct for water addition. cLignin+other = Solubles -
(sugar after digestion + inhibitors after digestion +ash). dSugar monomers from fiber after digestion with excess Biocellulase W (50
FPU/g fiber, pH 5, 50C, 96 h). Thymol crystals (10 mg) were added to prevent microbial growth. eSum of selected inhibitors:
organic acids (acetate, format) and furans (fufural and 5-hydroxymethylfurfural). 'Lignin+other = Insoluble fiber (sugar after
igestion + inhibitors after digestion +ash)










Table 4-3. Effect of enzyme treatment on sugar monomers in hemicellulose
hydrolysate.


Treatment Enzyme % Dry
temp.a treatedb weightc


160C

170C

180C

190C

160C

170C

180C

190C


Dimerd


Sugars (g/L)

Glucose Xylose Arabinose Galactose Total
sugars


No 30.9 3.0 7.4 0.7 1.4 0.1 36.6 3.2 3.0 0.7

No 31.01.2 6.00.3 2.70.2 56.02.0 3.41.4

No 27.4 2.1 3.6 0.2 4.2 0.6 62.8 0.5 4.6 1.5

No 27.6 0.8 2.1 0.2 6.7 0.6 63.2 1.1 5.6 0.1

Yes 30.9 3.0 1.7 0.7 6.4 0.4 53.8 4.2 2.6 0.8

Yes 31.0 1.2 1.3 0.2 7.0 0.2 64.0 0.9 3.6 1.1


Yes

Yes


27.4 2.1 0

27.6 0.8 0


7.3 0.3 65.9 2.4 4.4 1.1

8.9 0.4 65.0 0.9 4.6 0.8


aReaction times were 10 minutes at indicated temperatures (n3)


1.2 0.2 50 3

1.8 0.1 70 2

2.3 1.3 78 2

2.4 0.1 80 1

1.4 0.3 66 4

2.0 0.4 78 2

2.2 0.2 80 3

2.2 0.1 81 1


The liquid phase


(hemicellulose hydrolysate) was separated using a screw press. bHydrolysate was
adjusted to pH 5 and treated for 6 h (50C) with Biocellulase W (20 ml/L hydrolysate) to
hydrolyze remaining oligomers. CSolids content of bagasse slurry (fiber and non-volatile
solutes) emerging from the reactor after dilute acid steam pretreatment. dDimer was
calculated using cellobiose as a standard and contains xylose and glucose oligomers.










Table 4-4. Effect of enzyme treatment on selected inhibitors in hemicellulose
hydrolysate.


Reaction Enzyme % Dry
conditiona treatedb weightc


30.9 3.0

31.0 1.2

27.4 2.1

27.6 0.8

30.9 3.0

31.0 1.2

27.4 2.1


5-HMF

0.07 0.01

0.10 0.01

0.15 0.02

0.30 0.04

0.09 0.03

0.11 0.03

0.15 0.05


160C

170C

180C

190C

160C

170C

180C

190C


Furfural

0.31 0.05

0.78 0.16

1.57 0.29

3.48 0.27

0.32 0.10

0.74 0.14

1.59 0.31

3.41 0.22


Inhibitors (g/L)

Formic acid

0.40 0.04

0.53 0.06

0.78 0.05

1.05 0.10

0.24 0.07

0.51 0.12

0.66 0.18

1.09 0.19


Acetic acid

2.2 0.1

3.1 0.1

3.8 0.3

5.3 0.3

4.9 0.2

5.8 0.3

5.9 0.3

6.7 0.1


Total

3.0 0.2

4.5 0.2

6.3 0.4

10.2 0.4

5.6 0.2

7.2 0.4

8.3 0.5

11.5 0.3


a All pretreatment reaction times were 10 minutes at the indicated temperature (n > 3).
The liquid phase (hemicellulose hydrolysate) was separated using a screw press. b
Hydrolysate was adjusted to pH 5 and treated for 6 h (50C) with Biocellulase W (20
ml/L hydrolysate; 1000 FPU/L hydrolysate). c Solids content of bagasse slurry (fiber and
non-volatile solutes) emerging from the reactor after dilute acid steam pretreatment.


27.6 0.8 0.28 0.02









Table 4-5. Maximum ethanol concentrations and yields using L+SScF process.
Conditions Ethanol Ethanol yield
Steam Fermentation (g/L) (n) (g/metric (gal/U.S.
pretreatment ton) ton)
160C, 10 min 10% solids 19.1 0.4 (3) 191.0 58.1
160C, 10 min 12% solids 20.9 0.1 (2) 174.2 53.0
160C, 10 min 14% solids 23.6 1.3 (4) 168.6 51.3
170C, 10 min 14% solids 25.7 0.7 (4) 183.6 55.9
180C, 10 min 14% solids 29.0 1.5 (4) 207.1 63.1
Note: n = number of replicates










Table 4-6. Comparison of ethanol yields from SScF processes.
Ethanol Ethanol yield
Feedstock Pretreatment Biocatalyst titer (g/g untreated Reference
(g/L) feedstock)


No additional
Paper sludge treatments
treatments
NaOH,
Sugarcane NaOH,
bagae peracetic acid,
bagasse
wash
NaOH,
Hybrid poplar peracetic acid,
wash
Aqueous
Hybrid poplar ammonia,
wash
Aqueous
Barley hull ammonia,
wash
Aqueous
Corn stover ammonia,
wash
Switcharass AFEX


Forage
sorghum
Sweet
sorghum
bagasse
Rice straw

Corn stover

Poplar

Corn stover

Corn stover


AFEX

AFEX

AFEX

AFEX
Dilute sulfuric
LE-OL0
Dilute sulfuric
S02
impregnation


Sugarcane Dilute
bagasse phosphoric


45 0.26 Calculateda Zhang etal., 2009


24 Not available


Not available


S. cerevisiae
RWB222
Z. mobilis
CP4/pZB5

Z. mobilis
CP4/pZB5

E. coli KO11


E. coli KO11


E. coli KO11


S. cerevisiae
NRRL-D5A
S. cerevisiae
424A(LNH-ST)
S. cerevisiae
424A(LNH-ST)

S. cerevisiae
424A(LNH-ST)
S. cerevisiae
424A(LNH-ST)
Z. mobilis
pZB4L
S. cerevisiae
424A(LNH-ST)
S. cerevisiae
TMB3400
E. coli MM160
(KO11
derivative)


a Results presented were used to calculate yields on an original biomass basis. bData
presented was insufficient to calculate yields on an original biomass basis. After
sulfuric acid pretreatment, hemicellulose sugars and fiber were separated. Toxic side
products in hemicellulose hydrolysate were removed by a combination of liquid
extraction (LE) and overliming (OL). Both fiber and purified hydrolysate were combined
during fermentation (SScF).


100


16 0.24 Calculated


25 0.18 Calculated


19 0.18 Calculated

17 0.20

31 0.17

42 0.16

37 0.19

40 0.20

34 0.23 Calculated

42 0.21 Calculated

37 0.21 Calculated

29 0.21


Teixeira et al., 2000


Teixeira et al., 2000


Gupta and Lee, 2009


Kim et al., 2008


Kim and Lee, 2007

Alizadeh et al., 2005

Li et a., 2010

Li et a., 2010

Zhong et al., 2009

Lau and Dale, 2009

McMillan et al., 1999

Toon et al., 1997

Ohgren etal., 2006

This study











Conversion of Biomass to Fuel Ethanol & Chemicals

A Sulfuric Lignocellulose Process ------------------- ---
Washing Cellulose
Sashng+Lignin
Lignocellulose uDiluteAcid -- Liquid/solid
Hydrolysis Separation
Fermentation
(Zirconium Reactor) Hemicellulose Cellulose+
Syrup Cellulase
Separate hemicellose + 1 IPurification
SSF process for cellulose Hemicellulose Hemicellulose
Cleanup Fermentation


B Modified L+SScF Process with Phosphoric acid
_---------------------
I LFermentation I
DiluteAcid Liquefaction I
Lignocellulose el -- (+ lacellulase& -- Purification
ydrolysis (+cellulase hemicellulase
(Stainless Steel Reactor) ------------------------

C Mature Corn to Ethanol Industry

Sh o Liquefaction Fermentation
Sokm (+ amylase& amylasee& Purification
Sglucoamylase glucoamylase
(Stainless Steel Cooker) -- -


Figure 4-1. Comparison of ethanol production from lignocellulose and corn. A) Sulfuric
acid hydrolysis of hemicellulose (with washing and toxin mitigation) and
enzymatic hydrolysis of cellulose. Separate fermentations of cellulose (SSF)
and hemicellulose (HF). B) Simplified process using phosphoric acid
hydrolysis of hemicellulose and enzymatic hydrolysis of cellulose termed
liquefaction prior to co-fermentation in a single vessel (L+SScF). C)
Enzymatic liquefaction of hydrated corn prior to simultaneous saccharification
and fermentation (L+SSF).











-'in ---------------------


0 1 2
Time (days)


3 4


Time (days)


Figure 4-2. Selected examples of improvements in fermentation during serial transfers
of MM105 in hydrolysate medium. A resistant clone, MM160, was isolated
from transfer number 139 in medium containing phosphoric acid hydrolysate:
A) hemicellulose hydrolysate prepared at 150C (transfers 1, 4, and 9 in
broth containing 90% hydrolysate) and B) hemicellulose hydrolysate prepared
at 160C (transfers 33, 72, and 84 in broth containing 60% hydrolysate). The
parent organism, strain LY180, was incapable of growth and fermentation
under either condition.


102


-B-90% 150C T9
-- 90% 150C T4
-*-90% 1500C T1




















o 0
1 1


o \ I o
1- \ \ 1-



0 0
--------- ---------------------
0 -- ------ ---- 0
0.0 0.5 1.0 1.5 2.0 0 1 2 3 4
Furfural (g/L) 5-HMF (g/L)
4 4
--- LY180 D ----- LY180
-- MM160 .... MM160

3- 3--

E E
1- 2-






0 0, 0 ,
0---------------- --0-------------"
0 4 8 12 16 0 5 10 15 20
Formate (g/L) Acetate (g/L)



Figure 4-3. Resistance to selected inhibitors: A) furfural, B) 5-hydroxymethylfurfural, C)
format, D) acetate. Legend for all (n = 4): LY180, dashed line; MM160, solid
line.


103













60

9 50

S40- 1600C

30-E 170'C |
o 1800C

2 1900C
10


Sugar Dimers Glucose Xylose Arabinose Total Sugars

6

B
5-


.o 4-

S3-
0 M ^ 1600C
o
EM 170C

2 --180C

01--1 --,i .C


HMF Furfural Formate Acetate Total Inhibitors


Figure 4-4. Composition of fermentation broth at the time of inoculation (after
liquefaction for 6 h with enzymes): A) sugars and B) selected inhibitors.












104














25
B
20


0 48 96 144
Time (h)


10% Solids
12% Solids
-14% Solids

192 240


25 C

20

I15
-0- Sugar Dimers
a 10 Glucose
v --Xylose
S-*- Galactose
5
Arabinose


0 48 96 144 192 240
Time (h)


-0-Sugar Dimers
- Glucose
- Xylose
-- Galactose
-*-Arabinose


0 48 96 144 192 240
Time (h)

25
D
20
S--Sugar Dimers
15 -A Glucose
S15
0 -Xylose
a Galactose
10- -+-Arabinose

5


0
0 48 96 144 192 240
Time (h)


Figure 4-5. Fermentation of bagasse pretreated with phosphoric acid at 160C
(L+SScF). Graphs include results for 6 h prior to inoculation (t = -6 h).
Fermentors were inoculated at t = 0 h (after 6 h of liquefaction). A) Ethanol
production using different concentrations of acid-pretreated bagasse.
Concentrations of sugars during fermentations with B) 10% solids, C) 12%
solids and D) 14% solids bagasse.


105















30 B


25

20

15
--14% Solids (160'C)
10
/ / 14% Solids (170C)
5 / /14% Solids (180C)
--- 14% Solids (190C)

0 48 96 144 192 240
Time (h)
35

30 C


0 48 96
Time (h)


144 192 240


Xylose
S20 --Sugar Dimers 20 aXylose
ED Glucose 20 Galactose
15 Xylose 15- Arabinose
S--- Galactose
S10 e Arabinose 10

5 5-

0 -
0 48 96 144 192 240 0 48 96 144 192 240
Time (h) Time (h)



Figure 4-6. Fermentations of bagasse (14% w/v) pretreated with phosphoric acid at
various temperatures: A) ethanol, B) sugars during fermentation of bagasse
pretreated at 160C, C) sugars during fermentation of bagasse pretreated at
170C, and D) sugars during fermentation of bagasse pretreated at 180C.
Note that bagasse pretreated at 190C could not be fermented due to toxicity.


106









CHAPTER 5
GENERAL CONCLUSIONS AND FUTURE DIRECTIONS

General Accomplishments

Lignocellulosic biomass (LCB) is naturally resistant to chemical and biological

degradation. Mineral acids are most commonly used to pretreat LCB in order to release

hemicellulose sugars as well as provide a hydrolysable substrate for cellulolytic

enzymes. During the dilute acid steam pretreatment of LCB side products are generated

that are inhibitory to microbial growth. Therefore, the resulting liquid after pretreatment

is typically separated from the lignin-cellulose rich solid and can be treated by a variety

of methods to remove the soluble inhibitory compounds. Phosphoric was shown to

produce low levels of inhibitors during dilute acid pretreatment of rye grass straw, corn

stover, sorghum straw, and sugarcane bagasse (Gamez et al., 2004, 2006; Um et al.,

2003; Vazquez et al., 2007). Yeasts were able to ferment the hemicellulose hydrolysate

produced after hydrolyzing rye grass straw with phosphoric acid and neutralizing with

ammonium hydroxide (Israilides et al., 1978). The presence of lower level of inhibitors in

such a hydrolysate compared to sulfuric acid treatment, offered the opportunity to

reduce the need for exotic metallurgy and to simplify processing of LCB to ethanol.

Previous studies had shown that phosphoric acid could be used to hydrolyze sugarcane

bagasse and these studies were extended to include higher solids loadings, higher

temperatures, and lower acid usage during dilute acid pretreatment. Steam treatment of

sugarcane bagasse at low levels of phosphoric acid (1% of bagasse dry weight) and

high temperatures (160-1900C) was shown to be an effective method to hydrolyze

hemicellulose sugars. The generation of inhibitory side-products during such a

pretreatment was less than with sulfuric acid and these hemicellulose hydrolysates were









fermented to ethanol by the ethanologen, E. coli LY160, without the removal of these

compounds. However, the hydrolysate produced at 1900C could not be fermented. Low

levels of enzymes were able to hydrolyze residual glucan and xylan present in the solid

portion after pretreatment. Up to 45% of the residual glucan was solubilized at an

enzyme loading of 0.5 FPU/g of steam-treated fiber (1900C pretreatment).

The simultaneous saccharification and co-fermentation (SScF) process has been

shown to be more economical than the separate hydrolysis and fermentation process.

However, SScF requires high solids loadings in order to achieve adequate ethanol

concentrations that will make the process economically feasible. Low levels of cellulase

were shown to be effective at improving flow properties of acid pretreated sugarcane

bagasse slurries by sufficiently reducing their relative viscosity. The ability to pump

slurries of acid pretreated bagasse was correlated to enzyme-solubilized sugar, relative

viscosity, and flow through funnels of varying diameters. The solubilization of a small

amount of sugar (3.5% of the dry weight of bagasse) was able to dramatically reduce

viscosity. Viscosity was modeled as a function of cellulase dosage and incubation time.

Enzyme-solubilized sugar was modeled as a function of enzyme dose and incubation

time.

A simplified fermentation process was developed that combined dilute phosphoric

acid pretreatment and the use of a glucose and xylose co-fermenting hydrolysate-

resistant E. coli strain MM160. This process eliminated the need for solid liquid

separations and was used to ferment undetoxified phosphoric acid treated bagasse

slurries into ethanol. An initial liquefaction step (6 h) was used prior to inoculation to

sufficiently reduce viscosity and improve mixing of the acid pretreated bagasse slurry.


108









Such a fermentation produced a maximum yield of 0.21 g ethanol/g bagasse dry weight

corresponding to 29 g/L in the fermentation broth when using bagasse (14% solids)

pretreated at 1800C.

Future Directions

Process simplification has been made to convert lignocellulosic biomass to

ethanol. However, further improvements are needed to make the process commercially

viable. The economics of the process is largely dependent on the efficient

depolymerization of lignocellulosic biomass and the subsequent yield in converting the

resulting sugars into ethanol. The ethanol produced must be at high yield and titer.

The ethanol titer needs to be increased to above 40 g/L. The titer could potentially

be increased by increasing the solids loading of the fermentation. However, this would

also increase the concentration of inhibitory compounds. Further metabolic evolution

through continuous culturing in the presence of hydrolysate is needed. In addition,

bisulfite, an antioxidant that has been shown to be capable of improving hydrolysate

fermentability (Leonard and Hajny, 1945), can be added to the pH-adjusted (pH 6.5)

slurry at elevated temperatures prior to fermentation. Increasing solids loading would

also make it more difficult to pump and mix the acid pretreated sugarcane bagasse

slurries. Improving enzymatic saccharification by sonication or particle size reduction of

acid pretreated sugarcane bagasse may improve overall sugar and ethanol yields as

well as the flow properties of high solids slurries.

Glucose and xylose were both fermented by strain MM160 when inhibitory

compounds were at the lowest concentration (1600C pretreatment temp., 10% solids

loading). Xylose utilization decreased as pretreatment temperature and solids loading


109









increased most likely due to increasing concentrations of inhibitory compounds in the

hydrolysate. Xylose transport in E. coli under the conditions tested is through the high-

affinity (for xylose) ABC transporter XylFGH (Hasona et al., 2004). This transporter

hydrolyzes ATP to transport xylose. The use of ATP for xylose transportation coupled

with phosphorylation of xylose during metabolism (Hasona et al., 2004) may be

depleting ATP reserves in MM160. Overexpressing the low-affinity (for xylose) proton

symport transporters, XylE or AraE, of E. coli after the deletion of xylFGH could prevent

ATP hydrolysis during xylose transportation. Catabolite repression due to the presence

of glucose could be alleviated by the deletion of ptsG (gene coding for glucose specific

permease). Overexpression of the galactose permease, GalP, capable of transporting

glucose without either ATP hydrolysis or the phosphoenolpyruvate:glucose

phosphotransferase system, would allow the transportation of glucose in a ptsG- strain.


110









LIST OF REFERENCES


Aden, A., Ruth, M., Ibsen, K., Jechura, J., Neeves, K., Sheehan, J., Wallace, B., 2002.
Lignocellulosic biomass to ethanol process design and economics utilizing co-
current dilute acid prehydrolysis and enzymatic hydrolysis for corn stover
(NREL/TP-510-32438). National Renewable Energy Laboratory, Golden, Colorado.

Alizadeh, H., Teymouri, F., Gilbert, T.I., Dale, B.E., 2005. Pretreatment of switchgrass
by ammonia fiber explosion (AFEX). Appl. Biochem. Biotechnol. 121, 1133-1141.

Almeida, J.R.M., Modig, T., Petersson, A., Hahn-Hagerdal, B., Liden, G., Gorwa-
Grauslund, M.F., 2007. Increased tolerance and conversion of inhibitors in
lignocellulosic hydrolysates by Saccharomyces cerevisiae. J. Chem. Technol.
Biotechnol. 82, 340-349.

Bhandari, N., MacDonald, D.G., Bakhshi, N.N., 1984. Kinetic studies of corn stover
saccharification using sulphuric acid. Biotechnol. Bioeng. 26, 320-327.

Bommarius, A.S., Katona, A., Cheben, S.E., Patel, A.S., Ragauskas, A.J., Knudson, K.,
Pu, Y., 2008. Cellulase kinetics as a function of cellulose pretreatment. Metab.
Eng. 10, 370-381.

Boucher, R.M.G., 1975. On biocidal mechanisms in the aldehyde series. Can. J. Pharm.
Sci. 10, 1-7.

Bungay, H.R., 2004. Confessions of a bioenergy advocate. Trends Biotechnol. 22(2),
67-71.

Carvalho, G.B.M., Mussatto, S.I., Candido, E.J., Almeida e Silva, J.B., 2006.
Comparison of different procedures for the detoxification of eucalyptus
hemicellulosic hydrolysate for use in fermentative processes. J. Chem. Technol.
Biotechnol. 81, 152-157.

Chandei, A.K., 2007. Detoxification of sugarcane bagasse hydrolysate improves ethanol
production by Candida shehatae NCIM 3501. Bioresour. Technol. 98, 1947-1950.

Chandra, R.P., Bura, R., Mabee, W.E., Berlin, A., Pan, X., Saddler, J.N., 2007.
Substrate pretreatment: the key to effective enzymatic hydrolysis of
lignocellulosics? Adv. Biochem. Eng./Biotechnol. 108, 67-93.

Chum, H.L., Johnson, D.K., Black, S.K., Overend, R.P., 1990. Pretreatment-catalyst
effects and the combined severity parameter. Appl. Biochem. Biotechnol. 24-25, 1-
14.

Clasen, C., Kulicke, W.M., 2001. Determination of viscoelastic and rheo-optical material
functions of water-soluble cellulose derivatives. Prog. Polym. Sci. 26, 1839-1919.









Dasari, R.K., Berson, R.E., 2007. The effect of particle size on hydrolysis reaction rates
and theological properties in cellulosic slurries. Appl. Biochem. Biotechnol. 136-
140, 289-299.

Dasari, R.K., Dunaway, K., Berson, E.R., 2009. A scraped surface bioreactor for
enzymatic saccharification of pretreated corn stover slurries. Energy Fuels 23,
492-497.

Datsenko, K. A., Wanner, B.L., 2000. One-step inactivation of chromosomal genes in
Escherichia coli K-12 using PCR products. Proc. Natl. Acad. Sci. USA 97, 6640-
6645.

Demirbas, A., 2005. Bioethanol from cellulosic materials: a renewable motor fuel from
biomass. Energy Sources 27, 327-337.

Dien, B.S., Ximenes, E.A., O'Bryan, P.J., Moniruzzaman, M., Li, X., Balan, V., Dale, B.,
Cotta, M.A., 2008. Enzyme characterization for hydrolysis of AFEX and liquid hot-
water pretreated distillers' grains and their conversion to ethanol. Biores. Technol.
99, 5216-5225.

Fogler, H.S., 1986. Distributions of residence times for chemical reactors, in: Elements
of chemical reaction engineering. Prentice Hall, New Jersey, pp.629-679.

Frazer, F.R., McCaskey, T.A., 1989. Wood hydrolyzate treatments for improved
fermentation of wood sugars to 2,3-butanediol. Biomass 18, 31-42.

Galbe, M., Zacchi, G., 2007. Pretreatment of lignocellulosic materials for efficient
bioethanol production. Adv. Biochem. Eng./Biotechnol. 108, 41-65.

Gamez, S., Ramirez, J.A., Garrote, G., Vazquez, M., 2004. Manufacture of fermentable
sugar solutions from sugar cane bagasse hydrolyzed with phosphoric acid at
atmospheric pressure. J. Agric. Food Chem. 52, 4172-4177.

Gamez, S., Gonzales-Cabriales, J.J., Ramirez, J.A., Garrote, G., Vazquez, M., 2006.
Study of the hydrolysis of sugar cane bagasse using phosphoric acid. J. Food Eng.
74, 78-88.

Gautier, S., Lecourtler, J., 1991. Structural characterization by 13C nuclear magnetic
resonance of hydrolysed carboxymethylcellulose. Polym. Bulletin 26, 457-464.

Geankoplis, C.J., 1993. Transport processes and unit operations. Prentice Hall, New
Jersey.


112









Geddes, C.C., Peterson, J.J., Roslander, C., Zacchi, G., Mullinnix, M.T., Shanmugam,
K.T., Ingram, L.O., 2010. Optimizing the saccharification of sugar cane bagasse
using dilute phosphoric acid followed by fungal cellulases. Bioresour. Technol. 101,
1851-1857.

Grohmann, K., Torget, R., Himmel, M., 1985. Optimization of dilute acid pretreatment of
biomass. Biotechnol. Bioeng. 15, 59-80.

Grohmann, K., Torget, R., Himmel, M., 1986. Dilute acid pretreatment of biomass at
high solids concentrations. Biotechnol. Bioeng. Symp. 17, 135-
151.

Gupta, R., Lee, Y.Y., 2009. Pretreatment of hybrid poplar by aqueous ammonia.
Biotechnol. Prog. 25, 357-364.

Hahn-Hagerdal, B., Galbe, M., Gorwa-Grauslund, M.F., Liden, G., Zacchi, G., 2006.
Bio-ethanol-the fuel of tomorrow from the residues of today. Trends Biotechnol. 24,
549-556.

Hahn-Hagerdal, B., Karhumaa, K., Fonseca, C., Spencer-Martins, I., Gorwa-Grauslund,
M.F., 2007. Towards industrial pentose-fermenting yeast strains. Appl. Microbiol.
Biotechnol. 74, 937-953.

Hasona, A., Kim, Y., Healy, F.G., Ingram, L.O., Shanmugam, K.T., 2004. Pyruvate
format lyase and acetate kinase are essential for anaerobic growth of Escherichia
coli on xylose. J. Bacteriol. 186(22), 7593-7600.

Heer, D., Sauer, U., 2008. Identification of furfural as a key toxin in lignocellulosic
hydrolysates and evolution of a tolerant yeast strain. Microb. Biotechnol. 1, 497-
506.

Heer, D., Heine, D., Sauer, U., 2009. Resistance of Saccharomyces cerevisiae to high
furfural concentration is based on NADPH-dependent reduction by at least two
oxidoreductases. Appl. Environ. Microbiol. 75, 7631-7638.

Holtzapple, M.T., Jun, J.H., Ashok, G., Patibandla, S.L., Dale, B.E., 1991. The ammonia
freeze explosion (AFEX) process. Appl. Biochem. Biotechnol. 28-29(1), 59-74.

Ingledew, W.M., Kelsall, D.R., Austin, G.D., Kluhspies, C., 2009.The Alcohol Textbook:
A Reference for the Beverage, Fuel and Industrial Alcohol industries, fifth ed.
Nottingham University Press, Nottingham, UK.

Ingram, L.O., Lai, X., Moniruzzaman, M., Wood, B.E., York, S.W., 1997. Fuel ethanol
production from lignocellulose using engineered bacteria. In: Saha, B.H.,
Woodward, J. (Eds.), Fuel and chemicals from biomass. American Chemical
Society, Washington, D.C., pp. 57-73.


113









Ingram, L.O., Aldrich, H.C., Borges, A.C.C., Causey, T.B., Martinez, A., Morales, F.,
Saleh, A., Underwood, S.A., Yomano, L.P., York, S.W., Zaldivar, J., Zhou, S., 1999.
Enteric bacterial catalysts for fuel ethanol production. Biotechnol. Prog. 15, 855-
866.

Israilides, C.J., Grant, A., Han, Y.W., 1978. Sugar level, fermentability, and acceptability
of straw treated with different acids. Appl. Environ. Microbiol. 36, 43-46.

Jantama, K., Zhang, X., Moore, J.C., Shanmugam, K.T., Svoronos, S.A., Ingram, L.O.,
2008. Eliminating side products and increasing succinate yields in engineered
strains of Escherichia coli C. Biotechnol. Bioeng. 101(5), 881-893.
Jarboe, L.R., Grabar, T.B., Yomano, L.P., Shanmugam, K.T., Ingram, L.O., 2007.
Development o fethanologenic bacteria. Adv. Biochem. Engin./Biotechnol. 108,
237-261.

Jeffries, T.W., 1994. Biodegradation of lignin and hemicelluloses. In: Ratledge, C. (Ed.),
Biochemistry of microbial degradation. Kluwer Academic Publishers, Dordrecht,
Netherlands, pp. 233-277.

Jonsson, L.J., Palmqvist, E., Nilvebrant, N.O., 1998. Detoxification of wood hydrolysates
with laccase and peroxidase from the white-rot fungus Trametes versicolor. Appl.
Microbiol. Biotechnol. 49, 691-697.

Kabel, A.K., Bos, G., Zeevalking, J., Voragen, A.G.J., Schols, H.A., 2007. Effect of
pretreatment severity on xylan solubility and enzymatic breakdown of the remaining
cellulose from wheat straw. Biores. Technol. 98, 2034-2042.

Kadam, K.L., 2000. Environmental life cycle implications of using bagasse-derived
ethanol as a gasoline oxygenate in Mumbai (Bombay) (NREL/TP-580-28705).
National Renewable Energy Laboratory, Golden, Colorado.

Kim, T.H., Lee, Y.Y., 2007. Pretreatment of corn stover by soaking in aqueous ammonia
at moderate temperatures. Appl. Biochem. Biotechnol. 137, 81-92.

Kim, T.H., Taylor, F., Hicks, K.B., 2008. Bioethanol production from barley hull using
SAA (soaking in aqueous ammonia) pretreatment. Biores. Technol. 99, 5694-5702.

Klinke, H.B., Thomsen, A.B., Ahring, B.K., 2004. Inhibition of ethanol-producing yeast
and bacteria by degradation products produced during pretreatment of biomass.
Appl. Microbiol. Biotechnol. 66, 10-26.

Laadan, B., Almeida, J.R.M., Radstrom, P., Hahn-Hagerdal, B., Gorwa-Grauslund, M.,
2008. Identification of an NADH-dependent 5-hydroxymethylfurfural-reducing
alcohol dehydrogenase in Saccharomyces cerevisiae. Yeast 25, 191-198.


114









Larsson, S., Reimann, A., Nilvebrant, N.O., Jonsson, L.J., 1999. Comparison of different
methods for the detoxification of lignocellulose hydrolysates of spruce. Appl.
Biochem. Biotechnol. 77-79, 91-103.

Larsson, S., Cassland, P., Jonsson, L.J., 2001. Development of a Saccharomyces
cerevisiae strain with enhanced resistance to phenolic fermentation inhibitors in
lignocellulose hydrolysates by heterologous expression of laccase. Appl. Environ.
Microbiol. 67(3), 1163-1170.

Lau, M.W., Dale, B.E., 2009. Cellulosic ethanol production from AFEX-treated corn
stover using Saccharomyces cerevisiae 424A (LNH-ST). Proc. Natl. Acad. Sci.
USA 106, 1368-1373.

Lau, M.W., Gunawan, C., Dale, B.E., 2009. The impacts of pretreatment on the
fermentability of pretreated lignocellulosic biomass: A comparative evaluation
between ammonia fiber expansion and dilute acid pretreatment. Biotechnol.
Biofuels 2, 30.

Leonard, R.H., Hajny, G.J., 1945. Fermentation of wood sugars to ethyl alcohol. Ind.
Eng. Chem. 37(4), 390-395.

Li, B.Z., Balan, V., Yuan, Y.J., Dale, B.E., 2010. Process optimization to convert forage
and sweet sorghum bagasse to ethanol based on ammonia fiber expansion (AFEX)
pretreatment. Biores. Technol. 101, 1285-1292.

Liu, Z.L., Ma, M., Song, M., 2009. Evolutionarily engineered ethanologenic yeast
detoxifies lignocellulosic biomass conversion inhibitors by reprogrammed pathways.
Mol. Genet. Gen. 282, 233-244.

Lynd, L.R., van Zyl, W.H., McBride, J.E., Laser, M., 2005. Consolidated bioprocessing
of cellulosic biomass: an update. Curr. Opin. Biotechnol. 16, 577-583.

Martinez, A., Rodriguez, M.E., York, S.W., Preston, J.F., Ingram, L.O., 2000a. Effects of
Ca(OH)2 treatments ("overliming") on the composition and toxicity of bagasse
hemicellulose hydrolysates. Biotechnol. Bioeng. 69, 526-536.

Martinez, A., Rodriguez, M.E., York, S.W., Preston, J.F., Ingram, L.O., 2000b. Use of
UV absorbance to monitor furans in dilute acid hydrolysates of biomass.
Biotechnol. Prog. 16, 637-641.

Martinez, A., Rodriguez, M.E., Wells, M.L., York, S.W., Preston, J.F., Ingram, L.O.,
2001. Detoxification of dilute acid hydrolysates of lignocellulose with lime.
Biotechnol. Prog., 17, 287-293.


115









Martinez, A., Grabar, T.B., Shanmugam, K.T., Yomano, L.P., York, S.W., Ingram, L.O.,
2007. Low salt medium for lactate and ethanol production by recombinant
Escherichia coli B. Biotechnol. Lett. 29, 397-404.

McMillan, J.D., Newman, M.M., Templeton, D.W., Mohagheghi, A., 1999. Simultaneous
saccharification and cofermentation of dilute-acid pretreated yellow poplar
hardwood to ethanol using xylose-fermenting Zymomonas mobilis. Appl. Biochem.
Biotechnol. 77-79, 649-665.

Meade, G.P., Chen, J.C., 1977. Cane sugar handbook. John Wiley and Sons, New
York, New York.

Miller, E.N., Jarboe, L.R., Yomano, L.P., York, S.W., Shanmugam, K.T., Ingram, L.O.,
2009a. Silencing of NADPH-dependent oxidoreductase genes (yqhD and dkgA) in
furfural-resistant ethanologenic Escherichia coli. Appl. Environ. Microbiol. 75(13),
4315-4323.

Miller, E.N., Jarboe, L.R., Turner, P.C., Pharkya, P., Yomano, L.P., York, S.W., Nunn,
D., Shanmugam, K.T., Ingram, L.O., 2009b. Furfural inhibits growth by limiting
sulfur assimilation in ethanologenic Escherichia coli strain LY180. Appl. Environ.
Microbiol. 75(19), 6132-6141.

Mills, T.Y., Sandoval, N.R., Gill, R.T., 2009. Cellulosic hydrolysate toxicity and tolerance
mechanisms in Escherichia coli. Biotechnol. Biofuels 2,26.

Mitchell, D.J., Grohmann, K., Himmel, M.E., Dale, B.E., Shroeder, H.A., 1990. Effect of
the degree of acetylation on the enzymatic digestion of acetylated xylans. J. Wood
Chem. Technol. 10, 111-121.

Mosier, N., Wyman, C., Dale, B., Elander, R., Lee, Y.Y., Holtzapple, M., Ladisch, M.,
2005a. Features of promising technologies for pretreatment of lignocellulosic
biomass. Bioresour. Technol. 96, 673-686.

Mosier, N., Hendrickson, R., Ho, N., Sedlak, M., Ladisch, M.R., 2005b. Optimization of
pH controlled liquid hot water pretreatment of corn stover. Bioresour. Technol. 96,
1986-1993.

Ohgren, K., Galbe, M., Zacchi, G., 2005. Optimization of steam pretreatment of SO2-
impregnated corn stover for fuel ethanol production. Appl. Biochem. Biotechnol.
121-124, 1055-1067.

Ohgren, K., Bengtsson, O., Gorwa-Grauslund, M.F., Galbe, M., Hahn-Hagerdal, B.,
Zacchi, G., 2006. Simultaneous saccharification and co-fermentation of glucose
and xylose in steam-pretreatment corn stover at high fiber content with
Saccharomyces cerevisiae TMB3400. J. Biotechnol. 126, 488-498.


116









Ohgren, K., Bura, R., Saddler, J., Zacchi, G., 2007. Effect of hemicellulose and lignin
removal on enzymatic hydrolysis of steam pretreated corn stover. Bioresour.
Technol. 98, 2503-2510.

Ohta, K., Beall, D.S., Mejia, J.P., Shanmugam, K.T., 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. Appl. Environ. Microbiol. 57, 893-900.

Ou, M.S., Mohammed, N., Ingram, L.O., Shanmugam, K.T., 2009. Thermophilic
Bacillus coagulans requires less cellulases for simultaneous saccharification and
fermentation of cellulose to products than mesophilic microbial biocatalysts. Appl.
Biochem. Biotechnol. 155, 379-385.

Overend, R.P., Chornet, E., 1987. Fractionation of lignocellulosics by steam-aqueous
pretreatments. Philos. Trans. Royal Soc. Lond. A 321, 523-536.

Palmqvist, E., Hahn-Hagerdal, B., Galbe, M., Larsson, M., Stenberg, K., Szengyel, Z.,
Tengborg, C., Zacchi, G., 1996. Design and operation of a bench-scale process
development unit for the production of ethanol from lignocellulosics. Bioresour.
Technol. 58, 171-179.

Palmqvist, E., Hahn-Hagerdal, B., Szengyel, Z., Zacchi, G., Reczey, K., 1997.
Simultaneous detoxification and enzyme production of hemicellulose hydrolysates
obtained after steam pretreatment. Enzyme Microb. Technol. 20, 286-293.

Palmqvist, E., Hahn-Hagerdal, B., 2000a. Fermentation of lignocellulosics hydrolysates.
I: Inhibition and detoxification. Bioresour. Technol. 74, 17-24.

Palmqvist, E., Hahn-Hagerdal, B., 2000b. Fermentation of lignocellulosics hydrolysates.
II: Inhibitors and mechanisms of inhibition. Bioresour. Technol. 74, 25-33.

Perlack, R.D., Wright, L.L., Turhollow, A.F., Graham, R.L., Stokes, B.J., Erbach, D.C.,
2005. Biomass as feedstock for a bioenergy and bioproducts industry: The
technical feasibility of a billion-ton annual supply. ORNL/TM-2005/66. Oak Ridge
National Laboratory, Oak Ridge, TN.

Piper, P., Ortiz Calderon, C., Hatzixanthis, K., Mollapour, M., 2001. Weak acid
adaptation: the stress response that confers yeasts with resistance to organic acid
food preservatives. Microbiol. 147, 2635-2642.

Rezania, S., Zhuoliang, Y., Berson, R.E., 2009. Enzymatic saccharification and
viscosity of sawdust slurries following ultrasonic particle size reduction. Appl.
Biochem. Biotechnol. 153, 103-115.









Sambrook, J, Russel, D.W., 2001. Molecular cloning: A laboratory manual, third ed.
Cold Spring Harbor Laboratory Press, New York.

Sassner, P., Galbe, M., Zacchi, G., 2005. Steam pretreatment of Salix with and without
SO2 impregnation for production of bioethanol. Appl. Biochem. Biotechnol. 121-124,
1101-1117.

Sassner, P., Zacchi, G., 2008. Integration options for high energy efficiency and
improved economics in a wood-to-ethanol process. Biotechnol. Biofuels 1,4.

Schell, D.J., Farmer, J., Newman, M., Mcmillan, J.D., 2003. Dilute sulfuric acid
pretreatment of corn stover in pilot-scale reactor: investigation of yields, kinetics,
and enzymatic digestibilities of solids. Appl. Biochem. Biotechnol 105-108, 69-86.

Sluiter, A., Hames, B., Ruiz, R., Scarlata, C., Sluiter, J., Templeton, D., Crocker, D.,
2008. Determination of structural carbohydrates and lignin in biomass (NREL/TP-
510-42618). National Renewable Energy Laboratory, Golden, Colorado.

Teixeira, L.C., Linden, J.C., Schroder, H.A., 2000. Simultaneous saccharification and
cofermentation of peracetic acid-pretreated biomass. Appl. Biochem. Biotechnol.
84, 111-127.

Toon, S.T., Philippidis, G.P., Ho, N.W.Y., Chen, Z., Brainard, A., Lumpkin, R.E., Riley,
C.J., 1997. Enhanced cofermentation of glucose and xylose by recombinant
Saccharomyces yeast strains in batch and continuous operating modes. Appl.
Biochem. Biotechnol. 63-65, 243-255.

Um, B.H., Karim, M.N., Henk, L.L., 2003. Effect of sulfuric and phosphoric acid
pretreatments on enzymatic hydrolysis of corn stover. Appl. Biochem. Biotechnol.
105-108, 115-125.

U.S. Energy Information Administration, 2005,
http://tonto.eia.doe.gov/FTPROOT/multifuel/038405.pdf U.S. Department of
Energy, Washington, D.C.

Vazquez, M., Oliva, M., Tellez-Luis, S.J., Ramirez, J.A., 2007. Hydrolysis of sorghum
straw using phosphoric acid: Evaluation of furfural production. Bioresour. Technol.
98, 3053-3060.

Vogel, H.C., Todaro, C.L., 1997. Fermentation and biochemical engineering handbook:
principles, process design, and equipment. Noyes Publications, New Jersey.

Warwick, L.M., Gray, P.P., Mandels, M., 1985. Enzymatic hydrolysis of cellulose in
lignocellulosic materials. Crit. Rev. Biotechnol. 3(3), 235-276.


118









Wood, B.E., Ingram, L.O., 1992. Ethanol production from cellobiose, amorphous
cellulose, and crystalline cellulose by recombinant Klebsiella oxytoca containing
chromosomally integrated Zymomonas mobilis genes for ethanol production and
plasmids expressing thermostable cellulase genes from Clostridium thermocellum.
Appl. Environ. Microbiol. 58(7), 2103-2110.

Wood, B.E., Aldrich, H.C., Ingram, L.O., 1997. Ultrasound stimulates ethanol
production during the simultaneous saccharification and fermentation of mixed
waste office paper. Biotechnol. Prog. 13(3), 232-237.

Wyman, C.E., 2003. Potential synergies and challenges in refining cellulosic biomass to
fuels, chemicals, and power. Biotechnol. Prog. 19(2), 254-262.

Wyman, C.E., Dale, B.E., Elander, R.T., Holtzapple, M., Ladisch, M.R., Lee, Y.Y., 2005.
Coordinated development of leading biomass pretreatment technologies. Bioresour.
Technol. 96, 1959-1966.

Wyman, C.E., 2007. What is (and is not) vital to advancing cellulosic ethanol. Trends
Biotechnol. 25(4), 153-157.
Xu, F., Ding, H., 2006. A new kinetic model for the heterogenous (or spatially confined)
enzymatic catalysis: Contributions from the fractal and jamming (overcrowding)
effects. Appl. Catal. A 317, 70-81.

Yomano, L.P., York, S.W., Zhou, S., Shanmugam, K.T., Ingram, L.O., 2008. Re-
engineering Escherichia coli for ethanol production. Biotechnol. Lett. 30,
2097-2103.

Yomano, L.P., York, S.W., Shanmugam, K.T., Ingram, L.O., 2009. Deletion of
methylglyoxal synthase gene (mgsA) increased sugar co-metabolism in ethanol
producing Escherichia coli. Biotechnol. Lett. 31, 1389-1398.

Zaldivar, J., Ingram, L.O., 1999a. Effect of selected aldehydes on the growth and
fermentation of ethanologenic Escherichia coli. Biotechnol. Bioeng. 65, 24-33.

Zaldivar, J., Ingram, L.O., 1999b. Effect of organic acids on the growth and fermentation
of ethanologenic Escherichia coli LY01. Biotechnol. Bioeng. 66, 203-210.

Zaldivar, J., Martinez, A., Ingram, L.O., 2000. Effect of alcohol compounds found in
hemicellulose hydrolysate on the growth and fermentation of ethanologenic
Escherichia coli. Biotechnol. Bioeng. 68, 524-530.

Zhang, J.Y., Shao, X.J., Lynd, L.R., 2009. Simultaneous saccharification and co-
fermentation of paper sludge to ethanol by Saccharomyces cerevisiae RWB222.
Part II: Investigation of discrepancies between predicted and observed performance
at high solids concentration. Biotechnol. Bioeng. 104, 932-938.


119










Zhong, C., Lau, M.W., Balan, V., Dale, B.E., Yuan, Y.J., 2009. Optimization of
enzymatic hydrolysis and ethanol fermentation from AFEX-treated rice straw. Appl.
Microbiol. Biotechnol. 84, 667-676.


120









BIOGRAPHICAL SKETCH

Claudia Geddes was born in Leon, Nicaragua in 1980. Along with her family, she

moved to Miami, Florida in 1987 where she attended schools in Little Havana, Miami

and later Kendall, Miami where she attended G. Holmes Braddock Senior High School.

She graduated from high school in 1998. Claudia joined the Chemical Engineering

Department at the University of Florida in August 1998 and graduated with a Bachelor's

degree (cum laude) and a minor in Chemistry in May 2004. After graduating, she

decided to pursue a graduate degree and joined Dr. Lonnie Ingram's lab at the

University of Florida's Department of Microbiology and Cell Science in August 2005.

She is married to Ryan Geddes whom she met through the Chemical Engineering

program at UF and they now live in Gainesville, Florida with their son, Alessandro.





PAGE 1

1 SIMPLIFYING THE LIGNOCELLULOSE TO ETHANOL PROCESS THROUGH EFFICIENT PRETREATMENT AND IMPROVEMENT OF BIOCATALYST By CLAUDIA GEDDES A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010

PAGE 2

2 2010 Claudia Geddes

PAGE 3

3 To my husband who has belie ved in me every step of the way

PAGE 4

4 ACKNOWLEDGMENTS I thank my advisor, Dr. Lonnie Ingram, for allowing me to join his lab and for his mentorship throughout my graduate studies. I thank my committee members for all their invaluable advice and support. I thank Mike Mullinnix, Dr. James Peterson, Dr. Ismael N ieves, Ralph Hoffman, Dr. Zhaohui Tong, Sean York, Lorraine Yomano, Dr. Laura Jarboe and Dr. Elliot Miller for their assistance and advice in experiments. Mike Mullinnix and Sean York were instrumental in creating the hydrolysate-resistant strain MM160. Lorraine Yomano carried out the removal of FRT sites in strain MM105. Dr. Ismael Nieves, Ralph Hoffman, and Dr. Elliot Miller assisted in carrying out fermen tations. Dr. James Peterson was instrumental in carrying out compositional analyses of sugarcane bagasse. I thank all my family for their love and support that has given me confidence to achieve my goals. Last but not least I thank my husband, Ryan Geddes for b elieving in me and for his unwavering support

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ...................................................................................................... 4 LIST OF TABLES ................................................................................................................ 8 LIST OF FIGURES .............................................................................................................. 9 ABSTRACT ........................................................................................................................ 10 C H APT ER 1 INTRODUCTION ........................................................................................................ 13 Lignocellulos ic Biomass ............................................................................................. 13 Comparison of Major Lignocellulosic Substrates ...................................................... 14 Lignocellulosic Biomass Pretreatment ....................................................................... 17 Inhibitor Formation and Detoxification Methods ........................................................ 19 Biologi cal Methods ............................................................................................... 22 Chemical Methods ............................................................................................... 23 Adaptive Evolution ............................................................................................... 24 Example Process for the Conversion of Lignocellulosic Biomass to Ethanol .......... 25 Feedstock Storage, Handling and Size Reduction ............................................. 25 Hemicellulose Hydrolysis ..................................................................................... 27 Solid Liquid Separations ...................................................................................... 28 Hydrolysate Conditioning ..................................................................................... 30 Cellulose Hydrolysis and Ferm entation ............................................................... 30 Distillation ............................................................................................................. 32 Wastewater Treatment ......................................................................................... 34 Research Objectives ................................................................................................... 35 Reduce the Toxicity of Sugarcane Bagasse Hemicellulose Hydrolysate and Eliminate the Need for Separation of the Solid Lignin -Cellulose Rich Residue ............................................................................................................. 35 Establish a Correlation between Solubilized Sugar, Viscosity, and Ability to Pump for Pretreated and Enzymatically Digested Sugarcane Bagasse Slurries .............................................................................................................. 35 Design a Unified Fermentation Process that Combines Cellulose Hydrolysis, Hemicellulose Hydrolysate, and Fermentation in One Vessel ..... 36 2 OPTIMIZING THE SACCHARIFICATION OF SUGARCANE BAGASSE USING DILUTE PHOSPHORIC ACID FOLLOWED BY FUNGAL CELLULASES ............... 37 Introduction ................................................................................................................. 37 Materials and Methods ............................................................................................... 39 Materials ............................................................................................................... 39 Standard Analysis of Sugarcane Bagasse Composition .................................... 40

PAGE 6

6 Eff ect of Phosphoric and Sulfuric Acid Concentrations at 145C ....................... 40 Effect of Temperature on Steam Treatment of Sugarcane Bagasse w ith 1% (w/w) H3PO4 ...................................................................................................... 40 Enzymatic Hydrolysis of Cellulose after Steam Treatment with Phosphoric Acid ................................................................................................................... 41 Phosphoric Acid Method for Analysis of Bagasse Composition ......................... 41 Toxicity and Fermentation of Hemicellulose Hydrolysate (Phosphoric Acid) .... 42 Analytical Methods ............................................................................................... 43 Statistical Analysis ............................................................................................... 43 Results and Discussion .............................................................................................. 44 Effects of Phosphoric and Sulfuric Acid on Hemicellulose Hydrolysis (145C, 1 h) ....................................................................................................... 44 Increasing the Temperature to Reduce Treatment Time and Phosphoric Acid Usage ........................................................................................................ 45 Effect of Phosphoric Acid Pretreatment on the Enzymatic Hydrolysis of Bagasse ............................................................................................................ 47 Phosphoric Acid Method for Analysis of Bagasse Composition ......................... 48 Toxicity of Phosphoric Acid Hydrolysate ............................................................. 49 Conclusions ................................................................................................................ 50 3 OPTIMIZING CELLULASE USAGE FOR IMPROVED MIXING AND RHEOLOGICAL PROPERTIES OF ACID PRETREATED SUGARCANE BAGASSE ................................................................................................................... 57 Introduction ................................................................................................................. 57 Materials and Methods ............................................................................................... 59 Materials ............................................................................................................... 59 Dilute Acid Pretreatment of Sugarcane Bagasse ............................................... 59 Saccharification with Biocellulase W and glucosidase .................................... 60 Relative Viscosity Measurement ......................................................................... 61 Flow Property Test using Graded Funnels ......................................................... 61 Carbohydrate Composition and Analyses ........................................................... 61 Statistical Analysis ............................................................................................... 62 Resul ts and Discussion .............................................................................................. 62 Composition ......................................................................................................... 62 Effect of Cellulase on Relative Viscosity ............................................................. 63 Effect of Cellulase Loading on Extent of Saccharification .................................. 65 Effect of Enzyme Treatment on Flow through Graded Funnels ......................... 66 Correlation between the Extent of Saccharification and Relative Viscosity ....... 67 Effect of Mixing Acid Pretreated Fiber (No Enzyme Digestion) with Enzyme digested Acid Pretreated F iber (pH 5.0, 55 C, 6 h) on Viscosity ................... 68 Modeling an Ideal Continuous Stirred Tank Reactor (CSTR) to Decrease Viscosity ............................................................................................................ 70 Conclusions ................................................................................................................ 73

PAGE 7

7 4 A SIMPLIFIED PROCESS FOR ETHANOL PRODUCTION FROM SUGARCANE BAGASSE USING HYDROLYSATE-RESISTANT ESCHERICHIA COLI STRAIN MM160 AND PRETREATMENT WITH PHOSPHORIC ACID .................................................................................................. 81 Introduction ................................................................................................................. 81 Materials and Methods ............................................................................................... 83 Materials ............................................................................................................... 83 Steam Treatment of Bagasse with 1% (w/w) H3PO4 .......................................... 83 Organisms and Growth Conditions ..................................................................... 84 Genetic Methods .................................................................................................. 84 Isolation of Hydrolysate-resistant Bi ocatalysts ................................................... 84 Removal of Extraneous DNA Segments from Strain MM105 ............................ 85 Tolerance to Hydrolysate Toxins ......................................................................... 86 Liquefaction Followed by Simultaneous Saccharification and Co fermentation (L+SScF) of Acid pretreated Sugarcane Bagasse .................... 86 Analyses ............................................................................................................... 87 Statistical Analysis ............................................................................................... 88 Results ........................................................................................................................ 88 Composition of Materials Used as Substrates for Fermentation ........................ 88 Development of Hydrolysate-resistant Strain MM160 ........................................ 89 Liquefaction Prior to Simultaneous Saccharification and Co-fermentati on (L+SScF) ........................................................................................................... 90 Discussion ................................................................................................................... 92 5 GENERAL CONCLUSIONS AND FUTURE DIRECTIONS .................................... 107 General Acc omplishments ....................................................................................... 107 Future Directions ...................................................................................................... 109 LIST OF REFERENCES ................................................................................................. 111 BIOGRAPHICAL SKETCH .............................................................................................. 121

PAGE 8

8 LIST OF TABLES Table page 2 -1 Sugar composition of sugarcane bagasse (% dry weight) ................................... 52 2 -2 Solubilization of sugars from sugarcane bagasse by steam treatments with p hosphoric acid followed by 96 h enzyme (Biocellulase W) digestiona ............... 52 2 -3 Composition and fermentation of hemicellulose h ydrolysates ............................. 53 3 -1 Sugar composition of sugarcane bagasse, washed acid pretreated fiber and hemicellulose hydrolysate ...................................................................................... 75 3 -2 Effects of cellulase enzymes on enzyme-solubilized sugars and rheological properties ................................................................................................................ 76 3 -3 Predicted upper bound viscosity for an ideal continuous stirred tank reactor (CSTR) .................................................................................................................... 77 4 -1 Strains, plasmids and primers. .............................................................................. 94 4 -2 Components of sugarcane bagasse after steam pretreatment with phosphoric acid. ..................................................................................................... 96 4 -3 Effect of enzyme treatment on sugar monomers in hemicellulose hydrolysate. .. 97 4 -4 Effect of enzyme treatment on selected inhibitors in hemicellulose hydrolys ate. ............................................................................................................ 98 4 -5 Maximum ethanol concentrations and yields using L+SScF process. ................. 99 4 -6 Comparison of ethanol yields from SScF processes. ......................................... 100

PAGE 9

9 LIST OF FIGURES Figure page 2 -1 Effect of acid concentration on monomer sugars from dilute acid treatments (1 h, 145C) with H3PO4 and H2SO4. ..................................................................... 54 2 -2 Effect of acid concentration on the production of side products (furans and organic acids) during steam treatment .................................................................. 55 2 -3 Effect of treatment temperature on bagasse h ydrolysis (hemicellulose) and production of side products (1% w/w H3PO4, 10 min) .......................................... 56 3 -1 Effect of cellulase enzyme loading on saccharification and viscosity .................. 78 3 -2 Scatter plot of viscosity versus saccharificat ion (10% w/w slurries of acid pretreated fiber in hemicellulose hydrolysate). ...................................................... 79 3 -3 Effect of acid pretreated fiber additions on the viscosity of cellulase-digested slurries containing 10% dry weight acid pretreated fiber. ..................................... 80 4 -1 Comparison of ethanol production from lignocellulose and corn. ...................... 101 4 -2 Selected examples of improvements in fermentation during serial transfers of MM105 in hydrolysate medium. ........................................................................... 102 4 -3 Resistance to selected inhibitors ......................................................................... 103 4 -4 Composition of fermentation broth at the time of inoculation (after liquefaction for 6 h with enzymes) .......................................................................................... 104 4 -5 Fermentation of bagasse pretreated with phosphoric acid at 160 105 4 -6 Fermentations of bagasse (14% w/v) pretreated with phosphoric acid at various temperatures ........................................................................................... 106

PAGE 10

10 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy SIMPLIFYING THE LIGNOCELLULOSE TO ETHANOL PROCESS THROUGH EFFICIENT PRETREATMENT AND IMPROVEMENT OF BIOCATALYST By Claudia Geddes August 2010 Chair: Lonnie O. Ingram Major: Microbiology and Cell Scienc e A low level of phosphoric acid ( 1% w/w on dry bagasse basis, 160C and above, 10 min utes ) was shown to effectively hydrolyze the hemicellulose in sugar cane bagasse into monomers with minimal side reactions and to serve as an effective pretreatment for the enzymatic hydrolysis of cellulose. Up to 45% of the remaining water insoluble solids (WIS) was digested to sugar monomers by a low concentration of -glucosidase, although much higher leve ls of cellulase (100-fold) were required for complete hydrolysis. After neutralization and nutrient addition, phosphoric acid syrups of hemicellulose sugars were fermented by ethanologenic E. coli LY160 without further purification. Fermentation of these syrups was preceded by a lag that increased with increased pretreatment temperature. Further improvements in organisms and optimization of steam treatments may allow co -fermentation of sugars deriv ed from hemicellulose and cellulose, eliminating the need for liquid -solid separation, sugar purification, and separate fermentations.

PAGE 11

11 C onsolidat ion of bioprocessing steps with lignocellulose are limited in part by hydrolysate toxicity, material handling associated with fibrous suspensions, and the low activity of cellulase enzymes Combinations of enzyme dose and treatment conditions were shown to improve flow properties and pumping of acid pretreated sugarcane bagasse slurries (10% dry weight) Low level s of cellulase enzyme (0.1 and 0.5 FPU/g dry weight acid pretreated bagasse) were found to reduce the viscosity of these slurries by 77 -95% after 6 h of incubation, solubiliz ing at least 3.5% of the bagasse dry weight. Flow of slurries through small funnel s was a useful predictor of success with centrifugal and diaphragm pumps Equations were derived that describe changes in viscosity and solubilized sugars as a function of time and cellulase dosage. Blending of acid pretreated bagasse (10% dry weight) with suspensions of acid pretreated bagasse (10% dry weight) that had been previously digested with cellulase enzymes (low viscosity) did not increase viscosity in a linear fashion. Viscosity of these mixtures remained relatively constant until a threshold lev el of new fiber was reached, followed by a rapid increase with further additions Up to 35% of fresh acid pretreated bagasse could be blended with enzymedigested fiber (5.0 FPU/g dry weight acid pretreated fiber; 6 h) with only a modest increase in viscos ity A simple model is described to explain this phenomenon The smooth surfaces of enzymetreated fiber are proposed to hinder the frequency and extent of interactions between fibrils of fresh fiber particles (acid pretreated) until a threshold concentrat ion is achieved, after which fiber interactions and viscosity increase dramatically These results were used to model the viscosity in an ideal continuous stirred tank reactor (liquefaction) as a function of residence time and enzyme dosage

PAGE 12

12 A hydrolysate-resistant mutant of E. coli LY180 was selected by sequential transfers in AM1 mineral salts medium c ontaining hemicellulose hydrolysate from a dilute phosphoric acid pretreatment (160 C) of bagasse. The resulting strain was designated MM160. Strain MM160 was also more resistant than the parent to individual inhibitors such as furfural 5-hydroxymethylfurfural, and acetate. With this mutant, process steps such as the separation of hemic ellulose hydrolysate and fiber after dilute acid pretreatment (washing) and toxin removal from the hemic ellulose hydrolysate were not required prior to the fe rmentation. Pretreated bagasse was fermented in a single vessel without separation using strain MM 160. A liquefaction step with cellulase was included to improve pumpi ng and mixing (L+SScF), analogous to a corn ethanol process. Furans and other compounds in the hemicellulose hydrolysate remained as a barrier to potential contaminants during ferm entation. Bagasse slurries containing 10% and 14% dry weight (fiber plus solubles ) were tested in fermentations using pretreatment tem peratures of 160-190 C (1% phosphoric acid, 10 min). Saccharification efficiency and inhibitor production increas ed with treatment temperature. Bagasse samples pretreated at te mperatures below 190 C were fermentable. The highest yields and titers were obtained after 122 h of inc ubation using 14% dry weight slurries of pretreated bagasse (180 C), 0.21 g ethanol/g bagasse (dry weight) and 30 g/L, respectively. These results are similar to th ose that have been previously reported using pretreatments with amm onia and other bases in SScF processes.

PAGE 13

13 CHAPTER 1 INTRODUCTION Lignocellulosic Biomass The U S consumes 20 million barrels of petroleum daily of which 65% is imported (U.S. Energy Information Agency, 2005) In 2005, 71% of the total petroleum consumed was for the transpor t ation and industrial sector Most of the imported oil comes from political ly unstable countries which can cause considerable fluctuations in the price of crude oil (U S. Energy Information Agency 2005) T he price of crude oil reached an all time high of $14 7 per barrel in 2008 due to concern over Irans missile testing as well as other factors In addition to price fluctuations the supply of oil could be disrupted based on change s in foreign relations and the finite supply of oil The burning of fossil fuels also generates greenhouse gases which are believed to be the cause of global warming. To gain fuel independence and protect our environment policy makers are working on implementing new laws. The Energy Policy Act of 2005 passed by the U S. Congress mandates that 7.5 billion gallons of alternative fuels be produced per year by 2012. In 2007, Congress also mandated that 36 billion gallons of ethanol be produced per year by 2022 of which 44% must come from cellulosic biomass. Lignocellulosic biomass is the most abundant renewable biomass resource on Earth There are three major types of lignocellulosic biomass: agricultural crops and residues, hardwoods and softwoods. Lignocellulosic biomass is composed of three major polymers: hemicellulose, cellulose and lignin. The hemicellulose is a branched heteropolymer made up of C5 and C6 sugars such as xylose, arabinose, glucose, mannose and galactose. Cellulose is a highly crystalline homopolymer of glucose -1 4 glycosidic bonds. The U S. generates about 200 million dry tons of

PAGE 14

14 biomass per year which could potentially be converted to 16 billion gallons of ethanol based on a yield of 80 gallons of ethanol per dry ton (Perlack et al. 2005) World production of ethanol from biomass doubled b etween 2004 and 2007 reaching 13.2 billion gallons (Perlack et al. 2005) Lignocellulosic biomass provides a relatively cheap alternative to starch-based feedstocks and it does not interfere with food sources The conversion of lignocellulosic biomass to fuels and chemicals also provides a way to reduce greenhouse gas emissions and our dependence on foreign oil sources However, t he natural structure of lignocellulosic biomass is designed to be less accessible to chemical and biological degradation than st arch based biomass Therefore, a pretreatment step is required to effectively release the sugars present in the lignocellulosic biomass. The production cost of biomass to ethanol processes is mostly influenced by the overall conversion of sugars to ethanol and the ethanol concentration in the fermentation broth (Ohgren et al. 2006) The key to making the lignocellulose to ethanol process economically feasible lies in releasing the sugars for conversion to ethanol in a cost effective way (Hahn-Hagerdal et a l. 2006) Comparison of Major Lignocellulosic Substrates Lignocellulosic biomass contains 40-50% cellulose (a glucose polymer), 2535% hemicellulose (a sugar heteropolymer), 15 -20% lignin (a non -fermentable phenyl propene unit), and lesser amounts of minerals, oils, soluble sugars and other components The compositions and percentages of cellulose, hemicellulose and lignin vary from one plant species to another and also depend on the part of the plant, age of the plant and growth conditions (Demirbas, 2005) Grasses are generally composed of 25 -5 0% cellulose, 25-50% hemicellulose, 10 30% lignin, and other lesser constituents Sugarcane is an example of a grass and the

PAGE 15

15 bagasse is made up of 40-50% cellulose, 23 35% hemicellulose, and the remainder (~30% ) is made up of lignin with smaller amounts of minerals, waxes, and other compounds Hardwoods are made up of 45 ( 2)% cellulose, 30 ( 5)% hemicellulose, 20 ( 4)% lignin, 0.6 ( 0.2)% ash, and 5 ( 3)% extractives Hardwood barks are made of 2240% cellulose, 20 -38% hemicellulose, 30 55% lignin, 0.8 ( 0.2)% ash, and 6 ( 2)% extractives Softwoods are made up of 42 ( 2 )% cellulose, 27 ( 2)% hemicellulose, 28 ( 3)% lignin, 0.5 ( 0.1)% ash, and 3 ( 2)% extractives (non-cell wall components such as fatty acids a nd terpenoids ). Softwood barks are made up of 1838% cellulose, 15 33% hemicellulose, 30 -60% lignin, 0.8 ( 0.2)% ash, and 4 ( 2)% extractives Some of these ranges are large; however it can generally be said that the cellulose content of grasses, hardwood barks, and softwood barks is within the same range Also, the hemicellulose content of grasses is slightly higher than that of hardwood barks and softwood barks The lignin content of hardwood and softwood barks is much higher than in grasses with the lignin content of softwood barks being the highest ( Demirbas, 2005 ). Lignin content in softwoods is generally higher than in hardwoods Softwoods have a more compact fibrous structure than hardwoods Softwood lignins a re more cross linked than hardwood lignins because the methoxy group in the 5-carbon position of the phenylpropanoid unit is missing Softwood hemicellulose has a higher proportion of mannose and glucose units than hardwood hemicellulose which usually cont ains a higher proportion of xylose units Hemicellulose is more acetylated in hardwoods than in softwoods (Jeffries, 1994) Hemicelluloses of softwoods are primarily made of galactoglucomannans and the secondary cell wall can be made of up to 20% of this hemicellulose The glucose

PAGE 16

16 mannose ratio for gymnosperms is usually 1:3 (glucose:mannose) The extent of galactose substitution can vary greatly from 1 galactose unit for every glucose unit to 0.1 galactose unit for every glucose unit The galactose is at tached to the glucomannan backbone via 1> 6 linkages Also the galactoglucomannans have O acetyl groups at the C2 and C3 positions of mannose instead of hydroxyls The degree of substitution is usually one O acetyl group per 3-4 mannose units The second most abundant hemicellulose in softwoods is arabinoglucuronoxylans The structure of this hemicellulose is very similar to glucuronoxylan discussed below for hardwoods except -linked to the C2 and C3 positions of xylose units and oc cur once in every ten xylose units ( Jeffries, 1994 ). Hemicelluloses of hardwoods are primarily made of glucuronoxylan. This hemicellulose can make up 1530% of the dry weight of the wood The hemicellulose 1,4-linked xylose uni ts with 4 -O linked at the C2 position of xylose. There are also acetyl esters at the C2 and C3 positions of the xylan backbone The degree of O acetyl group substitution is high in glucoronoxylan with on average, 7-O acetyl groups p er 10 xylose units The 4 -O methylglucurnoic acid side chains occur about one in every ten xylose units ( Jeffries, 1994) Softwoods and hardwoods commonly contain glucomannans in their secondary cell walls However, in softwoods they make up 58% of the d ry weight of the cell wall and in hardwoods they make up 25% of the dry weight of the cell wall Also, softwood -(1 > 6) linked galactose unit to the mannose backbone. The

PAGE 17

17 structure of glucomannan hemicellulose is very similar to that of the galactoglucomannan ( Jeffries, 1994 ). The main hemicellulose of hardwood and germinaceous plants is glucuronoxylan and that of softwoods is glucomannan The glucuronoxylan of hardwoods and -(1 > 4) -D xylan backbone wi th substituents in -(1 > 2) -D (1 > 2) -4 -O methyl -D -glucuronic acid. Unlike hardwood xylans, germinaceous plant and -(1 > 3) linked arabinose furanose to the xylose backbone As mentioned earlier hardwood xylans are highly acetylated through ester linkages at the C2 and C3 positions of the xylan backbone Grass xylans and softwood galactomannans are acetylated but there is less acetylation in softwoods ( Jeffries, 1994). Lignocellulosic Biomass Pretreatment Prior to the cellulose hydrolysis step, th e lignocellulosic biomass must be pretreated to render the cellulose more accessible to the cellulolytic enzymes Cellulases are a class of enzymes that include endocellulase s and exocellulases Endocellulases cleave 1 4 -glycosidic bonds found in the i nternal structure of cellulose while exocellul ases ( i.e. cellobiohydrolases) cleave 1,4glycosidic bonds found at the ends of cellulose polymer chains Cellobiase glucosi dase) is a third enzyme used in the conversion of lignocellulosic biomass to soluble sugars and it cleaves cellobiose into two glucose molecules There are a variety of pretreatment methods available that have been developed each with its own advantages and disadvantages An effective pretreatment disrupts cell wall physical barriers as well as cellulose crystallinity and its association with lignin,

PAGE 18

18 and produces monomeric hemicellulose sugars with minimal production of inhibitory compounds (Mosier et al. 2005a) The pretreatment step is very important because it affects downstream processes such as fermentation toxicity, enzymatic hydrolysis rates, enzyme loadings, mixing power, product concentrations, product purification, waste treatment demands, and po wer generation (Mosier et al. 2005a ) Examples of pretreatment methods include dilute acid, ammonia fiber explosion (AFEX) ammonia recycle percolation, N -methylmorpholine-N oxide, and alkali pretreatments In acid pretreatment it is the hemicellulose th at is primarily hydrolyzed and in alkali catal y zed pretreatment it is the lignin that is primarily removed and a separate step is needed to render the hemicellulose sugar monomers available by the use of hemicellulases (Mosier et al. 2005 a ). Dilute acid pretreatment has the advantage of producing a pentose-rich syrup ( i.e. hydrolysate) that can be fermented by microorganisms without the use of hemicellulases or cellulases Sulfuric acid is most commonly used in dilute acid pretreatment however, other acids can also be used such as phosphoric and hydrochloric acid (Israilides et al. 1978) Phosphoric acid releases h emicellulose sugars and improves enzymatic hydrolysis of the resultant solid with the production of lower amounts of inhibitors compared to sulfuric acid (Geddes et al. 2010) In addition, p hosphoric acid can be used by biocatalysts for nutritional purposes at appropriate pH Other pretreatment methods are AFEX and water under pressure to penetrate the cell wall of biomass, hydrate cellulose and remove the hemicelluloses (Wyman et al. 2005) Ammonia is able to be recycled which makes it economically attractive as a pretreatment agent AFEX treatment decreases the crystallinity of cellulose and removes acetyl linkages of cellulose. This incr eases the ability of enzymes to access

PAGE 19

19 cellulose during hydrolysis allowing the use of l ower amounts of enzyme. The association of lignin with cellulose is also disrupted with AFEX pretreatment and hemicellulose oligomers are formed that need to be further hydrolyzed after pretreatment (Holtzapple et al. 1991 ). Treatment of hemicellulose with water under pressure seeks to reduce the concentration of sugar degradation products by avoiding the formation of monosaccharides (Mosier et al. 2005 b ). T his pretreatment requires the addition of enzymes after pretreatment to release the monosaccharides in cellulose as well as hemicellulose This approach has the advantage of high yields due to the reduction in sugar degradation (Mosier et al. 2005a ) The eff ectiveness of the pretreatment step can be assessed by two parameters : the digestibility of the solid material by enzymatic hydrolysis and by the fermentability of the liquid to determine its inhibitory potential on the fermenting organism The fe rmentation and enzymatic hydrolysis can be combined in simultaneous saccharification and fermentation. The type of pretreatment will also vary depending on the type of biomass AFEX, wet oxidation, and liquid hot water pretreatments have been shown to be effective for agricultural residues (Hahn-Hagerdal et al. 2006) Dilute acid s team pretreatment is effective for forestry and agricultural residues ( Hahn -Hagerdal et al. 2006 ). Inhibitor Formation and Detoxification Methods I nhibitory compounds are formed as bypr oducts of certain types of pretreatment and can inhibit fermentation of the solubilized sugars Therefore, the solid fraction is usually separated from the liquid hemicellulose hydrolysate fraction after pretreatment The cellulose is separately hydrolyzed by cellulolytic enzymes The hemicellulose hydrolysate can then be detoxified by various methods

PAGE 20

20 The future of the cellulosic biomass to ethanol industry depends on advances that help to integrate steps in the overall ethanol production process Integra ting the hydrolysis of cellulose and fermentation while eliminating the need for detoxification of the hemicellulose hydrolysate and therefore the solid liquid separation step prior to detoxification would create a simpler, more economical process This could be accomplished by a combination of the following: developing a pretreatment process that reduces the concentration of inhibitory compounds in the hydrolysate, by adapting a microbial strain to the presence of the inhibitory compounds, by using a strai n that can detoxify the hydrolysate as well as hydrolyze the cellulose portion through native or heterologous gene expression and by using an organism that will ferment all the pentoses and hexoses available in the lignocellulosic biomass Depending on the biomass and pretreatment type you can have varying degrees of inhibitory compounds in the hydrolysate. Also, microorganisms vary in their resistance to the different inhibitory compounds found in hemicellulose hydrolysates Sulfuric acid h ydrolysates of hemicellulose lead to the formation and release of a variety of inhibitory compounds such as furfural, acetic acid, formic acid, hydroxymethylfurfural, levulinic acid, and phenolic compounds (Miller et al. 2009a, Zaldivar and Ingram, 1999a) Acetic acid, typically the most abundant organic acid, can be a structural part of the hemicellulose (acetylxylan) as well as a sugar degradation product and is released during acid hydrolysis At high temperatures and pressures, xylose is degraded to f urfural and hexoses are degraded to hydroxymethyl furfural Furfural is the most abundant aldehyde found in sugarcane bagasse hydrolysates and was more toxic than

PAGE 21

21 hydroxymethyl furfural when using ethanologenic E. coli strains KO11 and LY01 in rich medium (Zaldivar and Ingram, 1999a) The combination of furfural with other organic acids (Zaldivar and Ingram, 1999b) and aldehydes (Zaldivar and Ingram, 1999a) had a synergistic inhibition of growth and fermentation (Zaldivar and Ingram, 1999b) Further degradation products include formic acid which is formed from the breakdown of furfural and hydroxymethyl furfural Levulinic acid is a degradation product of hydroxymethyl furfural Phenolic compounds are produced from the degradation of lignin d uring acid hydrolysis Examples of inhibitory lignin degradation compounds are vanillic acid, 4-hydroxybenzoic acid and syringic acid. Hydrophobic parts of proteins, enzymes, or membrane transport systems are possible sites of inhibitory action of some of the toxic compounds found in sulfuric acid hydrolysates of hemicelluloses (Palmqvist and Hahn-Hagerdal, 2000b) Studies using Saccharomyces cerevisiae have shown that the more hydrophobic the inhibitory compound the higher the inhibition of volumetric eth anol productivity (Palmqvist and HahnHagerdal, 2000b) Phenols and furans can have aldehyde, ketone, or acid functional groups Hi gher pKa value of the phenol hydroxyl group of aldehydes and ketones means that the phenolic proton is not dissociated at neutral pH and this makes the phenol more hydrophobic (Klinke et al. 2004 ). Researchers believe that the low molecular weight phenolic compounds are highly inhibitory It may be possible to specifically remove these compounds alone and sufficiently reduce fe rmentation inhibition ( Palmqvist and HahnHagerdal, 2000a ). There are different mechanisms of toxicity that have been proposed for the various toxic compounds in hydrolysates and some mechanisms are still unknown. Aldehydes

PAGE 22

22 are highly reactive chemical species and can disrupt the structure of nucleic acids, proteins, and lipids in microorganisms (Boucher, 1972; Miller et al. 2009a) Phenolic compounds are believed to disrupt the structural integrity of biological membranes (Palmqvist and HahnHagerdal, 2000b) The temperature, reaction time, and acid concentration during dilute acid hydrolysis influence the concentration of degradation products of cellulose, hemicellulose, and lignin in hemicellulose hydrolysates T he severity of different pretreatment m ethods could be compared using a combined severity factor that takes these three parameters into account and also the influence of hydrolysis pH (Chum, 1990) Using this as a starting point a researcher could find the optimal combined severity factor for b iomass pretreatment where sugar yield is optimized and inhibitor formation is minimized There are different detoxification methods that can be grouped into two main categories : biological, and chemical Each method has its advantages and disadvantages B iological M ethods Treatment of willow hemicellulose hydrolysate with peroxidase and laccase enzymes from Trametes versicolor was found to increase the maximum ethanol of Saccharomyces cerevisae (Jonsson et al. 1998) The researchers also noticed that laccase selectively removed phenolic monomers and acids almost to complet ion. L accase detoxified the hydrolysate by the oxidative polymerization of low molecular weight phenolic compounds When Trichoderma reesei is added to hemicellulose hydrolysates it c an remove acetic acid, furfural and benzoic acid (Palmqvist et al. 1997) Instead of having a separate organism to detoxify the hydrolysate, such as T.

PAGE 23

23 versicolor the peroxidase and laccase genes could be cloned into the fermenting organism This would eliminate the need for separating the detoxification and fermentation steps and the maintenance of a separate organism P a l mqvist et al. (1997) used T. reesei to produce cellulolytic enzymes and to remove inhibitory compounds found in the hydrolysate prior to ethanologenic fermentation by S. cerevisiae However, an improvement on this process would be the elimination of the use of T. reesei because consumes pentose sugars found in the hemicellulose that could be used to make ethanol Furthermore, this organ ism need s its own set of nutritional requirements that also increase costs Instead, Larsson et al. (2001) developed an S. cerevisiae strain that was tolerant to phenolic inhibitors in lignocellulose hydrolysate by overexpressing the T. reesei laccase gene from a plasmid inside the yeast along with the Sso2 protein which is involved in protein excretion. However, this involved the maintenance of the plasmid and laccase r equires oxygen for its activity As an alternative genetic approach, the gene(s) that E. coli uses to detoxify hydrolysate can be identified and then overexpressed to potentially confer more resistance to concentrated hydrolysate. The detoxification of phenol aldehydes by conversion to alcohols in anaerobic cultures have also been shown in S. cerevisiae and in Klebsiella pneumoniae (Larsson et a l. 2001 ). Therefore, these genes could also be identified and overexpressed to confer resistance to inhibitors in these ethanologenic bacteria. Chemical M ethods Extraction of a spruce hydrolysate with ether at a pH of 2 increased ethanol yield to that of a reference fermentation (Palmqvist and Hahn-Hagerdal, 2000a) The inhibitory compounds extracted by the ether were acetic acid, formic acid, levulinic acid, furfural, hydroxymethylfurfural and phenoli c compounds Extraction by ethyl acetate had

PAGE 24

24 similar effects on hydrolysate fermentation and removed acetic acid, vanillin, 4hydroxybenzoic acid and completely removed furfural ( Palmqvist and Hahn -Hagerdal, 2000a). Detoxification of hydrolysate by alkali treatment is a popular method and is often referred to as overliming It involves increasing the pH of the hydrolysate to 910 with Ca(OH)2 The pH of the hydrolysate is subsequently decreased to the desired fermentation pH with acid. The detoxification is believed to be due to precipitation of toxic compounds Overliming decreases the concentration of Hibberts ketones, furfural and hydroxymethylfurfural in spruce hemicellulose hydrolysates (Palmqvist and HahnHagerdal, 2000a) Some researchers found that the detoxification of willow hemicellulose hydrolysate was most effective when they used a combination of overliming and heated sul fi te treatment (Palmqvist and Hahn-Hagerdal, 2000a) The combined detoxification removed Hibberts ketones, aldehydes, and volatile compounds (Palmqvist and HahnHagerdal, 2000a). Also treatment with ionexchange resins and charcoal can effectively remove inhibitors (Carvalho et al. 2006) Detoxification of sugarcane bagasse hydrolysate with anion resins effectively remov ed 84% of the acetic acid ( Chandei, 2007) Adaptive Evolution An additional strategy in combating the inhibitory effects of hemicellulose hydrolysate compounds is to adapt cells to the inhibitors This can be done by maintaining continuous cultures to met abolically evolve the cells in the presence of hydrolysate and select for cells that are more robust having a high cell and ethanol yield Escherichia coli can naturally remove the inhibitory effects of furfural by reducing it to furfuryl alchohol in anaer obic conditions or to furic acid in aerobic conditions An

PAGE 25

25 evolved E. coli mutated to silence NADPH -dependent oxidoreductase genes, yqhD and dkgA which are responsible for detoxifying furfural (Miller et al. 2009b) Yeasts resistant to benzoic acid showed lower uptake rates of benzoic acid (Piper et al. 2001) The change in uptake rate could be due to changes in membrane permeability (Piper et al. 2001) The identification of potential genes involved in the conferred resistance to inhibitors in hydrolys ate could lead to a genetic approach to engineer a more robust organism for hydrolysate fermentations with minimal or no prior detoxification. Example Process for the Conversion of Lignocellulosic Biomass to Ethanol A typical process to convert cellulosic biomass into ethanol involves the following steps : 1. feedstock storage and handling, 2. size reduction, 3. pretreatment, 4 d etoxification of the hemicellulose fraction, 5 enzyme production, 6 hydrolysis of the cellulose fraction, 7 hexose and pentose fermentation, and 8 d istillation (Meade and Chen, 1977) However, these steps can vary depending on how integrated the process is and other factors The f ollowing is a short description of the main steps involved in a typical process and some important i ssues associated with them Feedstock Storage, Handling and Size Reduction Assuming that a typical sugar mill produces about 285,000 tons of bagasse per year and that 1525% of the bagasse is not burned for energy (bagasse heating value is 18.1 GJ/ton), th en a typical sugar mill can provide about 57,000 tons of bagasse per year Over a million tons of dry sugarcane bagasse is produced by the Florida sugar industry Sugarcane crushing season in Florida (and others) begins in October November and ends in Apri l or May Sugarcane deteriorates rapidly from the time it is cut and harvested so harvesting, crushing and processing are done concurrently (Meade and Chen, 1977 ). Therefore, sugarcane bagasse will be available to use at 50%

PAGE 26

26 moisture during sugarcane crushing season and for the rest of the time ( i.e. ~165 days) bagasse needs to be stored It has been proposed that bagasse can be dried to approximately 46-47% moisture using waste heat in the boiler flue gases but this may have high maintenance and ener gy costs associated with it So lar drying may be a cheaper alternative. The sugarcane bagasse may be kept on concrete slabs to reduce the amount of standing water in the storage area and will prevent contact with dirt (Kadam, 2000) The feedstock transport radius is an important parameter in determining the costs associated with this unit. Trucks are used to carry the bagasse bales and they can carry about 10 metric tons of wet bagasse each. The amount to be carried by the trucks can vary from this number but road weight limits must be followed The handling areas consist of weigh stations where the trucks are weighed and unloaded by forklifts, solids conveyor systems, and others Some of the bales are sent to storage bunkers while others are put onto the conveyors where they would be unwrapped automatically Bulldozers are also needed to make piles of bagasse for future handling The actual amount of trucks and forklifts has to be worked out for the individual plants storage requirements ( Aden et al. 2002 ). T he amount of size reduction for the bagasse is dependent on the type of pretreatment to be used and the heat and mass transfer equations associated with the pretreatment The type of size reduction will also determine the amount of energy that will nee d to be put into this unit operation. The type of size reduction will determine the amount of energy that will need to be put into milling the bagasse to the right size as well as the efficiency of pretreatment ( Wyman, 2007). The bagasse is often milled to an

PAGE 27

27 mean size of 15 mm The bagasse is taken from the storage bunker via a screw feeder and fed into the pretreatment process ( Aden et al. 2002). Cellulosic biomass is relatively cheap and an estimate average figure could be given for sugarcane bagasse o f about $30/dry ton. Increasing the amount of feedstock tends to drop the processing costs However, as the amount of feedstock in the processing facility is increased, the cost associated with storing it, handling it, reducing its size, and transporting it also increases Therefore, based on cost analys is that ha s been done for cellulosic biomass to fuels and chemicals researchers have found that these two cancel each other and that the overall cost stay s relatively the same with t he amount of feedstock used ( Wyman, 2003). However, when determining the costs associated with the amount of feedstock that will be used it is important to know the cost of feedstock transportation and what effect plant size has on the capital and operating cost of the plant ( Wyman, 2007) Hemicellulose Hydrolysis The effectiveness of a particular pretreatment depends on the type of biomass used Biomass is naturally resistant to breakdown into component sugar monomers because it has a structure that makes it resistant to enzymatic and chemical degradation. Each biomass type has its own characteristic components and ratios of components The pretreatment should be chosen based on laboratory and pi lot plant data specific for each biomass The type of steel used for this process needs to be determined through m anufacturers data on the acid and temperature resistance of the steel The reduced particle size bagasse is typically first steamed at low -p ressure to 100 C This provides about 1/3 of the heat requirement for the hydrolysis reaction with a typical reactor

PAGE 28

28 having residence time of 2 minutes, 1.1% sulfuric acid concentration, and a hydrolysis temperature of 190 C The bagasse is fed into the pr e -steamer by screw conveyors that have variable frequencies depending on the feed rate desired. The pre -steamer is a horizontal vessel and has paddles inside to move the bagasse around. After the bagasse has been pre -steamed it is taken out and added to another screw conveyor that transfers the bagasse to a blow tank The blow tank uses steam to inject the bagasse into the hydrolysis reactor via a live bottom screw and pressurizes the bagasse along the way The reactor is designed to be 95% full and opera tes at 30% solids in the reactor The reactor is directly injected with steam to 190 C after the concentrated sulfuric acid is added and diluted to 1.1% with water The bagasse is flashed cooled after the hemicellulose hydrolysis by discharging into a lowe r pressure flash tank for 15 minutes Acetic acid, furfural, and hydroxymethylfurfural present in the acid pretreated bagasse are removed in the form of evaporated condensate by flash cooling (Aden et al. 2002 ). Hemicellulose hydrolysis affects the concentration of sugars and products that can be obtained. Also, depending on the microorganism used, the sugars released during hemicellulose hydrolysis may or may not be used during fermentation with glucose present in the medium Hemicellulose h ydrolysi s also affects the quality of the lignin containing solids that may be used as fuel for the process, as exported heat and electricity, or used to make a specialty ligninderived product. Solid -Liquid Separations There are different types of liquid solids s eparations For slurries such as the ones encountered in a biomass to ethanol process, filtration is the preferred technique for separation. The particle size distribution of the slurry is very important for the

PAGE 29

29 effectiveness of the filtration and must be accurately determined and optimized in the field There are three basic types of filtration : depth, cake, and membrane The rate of filtration is dependent on the force of the selected equipment type and on the resistance of the cake that is being formed. The flow of filtrate through a cake follows Poiseuilles equation and can be used to design the liquid solid separations The contributing variables to the filtration rate are filter surface area, pressure across the filter medium, average specific cake re sistance, weight of cake, resistance of the filter medium, and visco s ity ( Vogel and Todaro, 1997). There are three important steps involved in the separation of liquids from solids in the biomass to ethanol process The steps are the separation of the cellulose and lignin containing solids after pretreatment from the hydrolysate liquor, the separation of gypsum (if it is used for conditioning) from the conditioned hydrolysate, and the dewatering of the lignin residue in the stillage after distillation. It is difficult to separate the solid cellulose and lignin containing fraction from the hydrolysate liquor because the s olid behaves like a sponge and tends to absorb and retain water Another complication is the high temperature of the slurry and stillage that must be filtered. The high temperature cause s an increase in the corrosiveness of the slurry and stillage. Therefore, the type of steel used for this process needs to be determined through manufacturers data on the acid and temperature resistance of the steel ( Aden et al. 2002). Once the hydrolysate has been flashed cooled in the flash tank the hydrolysate slurry is approximately 21% insoluble solids The slurry is taken to a pressure filter to separate the solids containing the cellulose and lignin from the hemicellulose sugar and

PAGE 30

30 toxic compound containing liquid A pressure belt filter press was found to provide the best solids recovery with the least amount of water (Aden et al. 2002). Compressed air is passed through the hydrolysate slurry on a filter This displaces liquid and increases the solid content of the filter cake. Using less water used to wash the cake reduces the energy required for downstream processes such as product recovery The final cake is conveyed to a slurry tank and mixed with conditioned hydrolysate liquor and recycled water Gypsum is separated by hydrocyclone and rotary drum filters in series This provides gypsum at around 80% solids ( Aden et al. 2002 ). Hydrolysate Conditioning After the hemicellulose hydrolysis the h ydrolysate could be flash cooled to vaporize a large amount of water along with some acetic acid, furfural, and hydroxymethylfurfural (Aden et al. 2002). Hydrolysate liquor that is to be conditioned is pumped from the filtrate tanks of the belt filter and sent to a heat exchanger to be cooled to 50 C The liquor is sent to a tank and lime is then added to raise the pH to 10 and allowed to stir for 1 hour The conditioned hydrolysate is then adjusted to pH 4.5 in a separate tank and stays there for 4 hours to allow the gypsum to crystallize ( Aden et al. 2002 ). Cellulose Hydrolys is and Fermentation Prior to carrying out full -scale fermentations in a plant, researchers must first do laboratory and pilot plant testing There are several steps involved in this process : 1) developing of strains for physical and genetic improvements, 2) optimizing medium conditions and culture conditions such as pH and temperature to the conditions desired, 3 ) determining the amount of oxygen supply required by the organisms duri ng the fermentation (although ideally you would want anaerobic conditions because it costs

PAGE 31

31 less), 4 ) selecting the culture process operation mode, 5) measuring rheological properties of the culture, 6 ) modeling and formulation of process control strategies and 7 ) manufacturing of the various bioreactors, sensors, agitators and other accessory equipment needed for the process ( Vogel and Todaro, 1997). Fermentation can be batch or continuous Batch fermentation has a low risk of contamination but it is labor intensive, and has high investment costs Batch fermentation with recirculation of cells reduces water and energy demand but causes inhibitor accumulation Continuous fermentation reduces equipment costs and decreases inhibition but has a higher potentia l for contamination than batch fermentation The ethanol productivity is influenced by product inhibition, cell mass concentration, and lignocellulosic hydrolysate inhibitor concentrations The seed culture for fermentation can be grown in a separate fermentor(s) or it can come from cells that have been recirculated from a previous fermentation. Cell recirculation can be advantageous because it may produce cells that are better adapted to the inhibitory environment of the fermentation ( Palmqvist and Hahn-Ha gerdal, 2000a ). Current research efforts are focused on the combination of d etoxification, cellulose hydrolysis and hemicellulose fermentation in one vessel, termed consolidated bioprocessing ( Lynd et al. 2005 ). Lignocellulose derived sugars are a mixtur e of pentoses and hexoses with sugarcane bagasse hemicellulose hydrolysate containing mostly xylose Saccharomyces cerevisiae is the most commonly used industrial microorganism in the conversion of lignocellulosic biomass to ethanol It can ferment hexoses to ethanol and some strains have a high tolerance to ethanol and to other inhibitory compounds found in the hydrolysate (Hahn-Hagerdal et al. 2007) However,

PAGE 32

32 S. cerevisiae is not naturally able to ferment pentoses Therefore, the pentose sugars are not used However, recombinant yeast strains have been developed that can ferment both hexoses and pentoses (Hahn-Hagerdal et al. 2007) The organisms, E. coli and Klebsiella oxytoca, can naturally ferment both pentoses and hexoses and strains have been developed through the introduction of ethanologenic genes from Zymomonas mobilis that can anaerobically ferment the mixed sugars with ethanol as the primary fermentation product ( Jarboe et al. 2007; Wood and Ingram, 1992; Yomano et al. 2009 ). Distillation Distillation is used to separate components in a liquid solution. The distillation depends on the distribution of these components in the vapor and liquid phase. Therefore, the composition of the vapor must be different from the composition of the liquid with which it is in equilibrium at the boiling point of the liquid There are different types of distillation methods The first type consists of boiling a liquid to produce vapor in a single stage and collecting the condensing vapors without the return of any liquids back to the single stage to contact the rising vapors The second type of distillation consists of boiling a liquid to produce vapor and collecting the condensate on a series of trays in the distillation column that allow the flow of condensat e counter -currently to the rising vapors The second type of distillation is the one used in industry to concentrate volatile compounds such as ethanol and butanol and is called fractional distillation, distillation with reflux, or rectification For separ ating two compounds, t he McCabe Thiele Method is used to calculate the theoretical number of trays necessary to achieve a specific separation of a binary mixture in the second type of distillation ( Geankoplis, 1993)

PAGE 33

33 After fermentation is complete the cel ls are usually removed from the broth, and the ethanol (or other volatile product such as butanol) in the fermentation supernatant can be removed by distillation It is desirable to have high ethanol concentration for efficient recovery and to minimize the amount of minerals present and energy required in this step Some problems associated with distillation of ethanol are problems with an azeotrope, unfavorable phase equilibrium as pure ethanol is approached, and mineral fouling Current research is focused on creating more efficient product recovery and a drying process has been developed that has the potential to be more desirable than distillation because it avoids the azeotrope problem ( Bungay, 2004 ). In a typical process, ethanol is separated fr o m the fermentation broth by distillation and is subsequently dehydrated by molecular sieve technology Distillation involves two columns The first column removes the dissolved CO2 and most of the water The second column is used to concentrate the ethanol to a near azeotropic condition (96% w/v) The 99.7% (w/v) ethanol is made with molecular sieve technology and sent to fuel storage. The fermentors have vents and the gases from these vents as well as the gases from the first distil lation column vent are scrubbed with water to remove any ethanol in these streams The scrubber effluent is then fed back into the first distillation column The still bottoms containing the lignin residue is de watered by solid liquid separation and burned to cogenerate steam and electricity for the process The excess electricity can be sold to the grid. The filtrate from the solid liquid separation process is used for recycling to other steps in the process but some may be concentrated in a multiple eff ect evaporator The concentrated filtrate is used for fuel and the condensate is used as clean recycle water for the process The amount of recycled water coming

PAGE 34

34 from the stillage must be accurately controlled because this stream contains organic salts and inorganic compounds that may inhibit the fermenting organisms ( Aden et al. 2002). Wastewater Treatment A variety of process streams are combined into the wastewater treatment These streams include flash vapor from various flash cooling steps in the pro cess, condensate from the belt pressure filter vent, multi effect evaporator condensate not recycled, and boiler blowdown ( Aden et al. 2002) These streams must be treated depending on where they will go The United States Environmental Protection Agency (US EPA) has mandated strict discharge standards that must be met by industry There are two general types of wastewater treatment : biological and physical The biological treatment involves using microorganisms to reduce the pollut ants in the waste streams to limits acceptable by the EPA The physical method of treatment involves physical procedures such as stripping, ion exchange, and membrane separation to remove pollutants from the streams ( Vogel and Todaro, 1997). The wastewater str eams are combined and large particles are removed and sent to a landfill The remaining liquid is primarily composed of the evaporator condensate, so it is first cooled in a heat exchanger and then anaerobically digested. The anaerobic digestion produces m ethane that is combusted and then the wastewater is treated by aerobic digestion. After the aerobic digestion step is complete, the sludge produced is separated from the water The water is recycled for use in the process and the biomass containing sludge is combusted or sent back to the aerobic digestion stage. The total chemical oxygen demand (COD) of the wastewater before treatment needs to be determined so that calculations of the nutrient requirement and residence time for the

PAGE 35

35 subsequent digestions can be made. The COD is the amount of oxygen required to combust all soluble organic components to carbon dioxide, nitrate and water (Aden et al. 2002) The amount of wastewater could be different depending on the final product chosen Some of the cells t hat are filtered from the aerobically digested water could be used as fertilizer and that would change the amount of cells put back into the aerobic digester or combusted. Research Objectives Reduce the T oxicity of Sugarcane B agasse H emicellulose H ydrolys ate and Elim in ate the N eed for S eparation of the Solid L ignin-C ellulose R ich R esidue The dilute acid hydrolysis of hemicellulose can be catalyzed by most mineral acids. The most commonly used acid is sulfuric acid, and it produces a hemicellulose hydrolysate containing compounds that inhibit microbial fermentation. Therefore, it is often necessary to separate and detoxify the hydrolysate prior to fermentation. Dilute suspensions of corn stover, sugarcane bagasse, and sorghum straw have been demons trated to be effectively hydrolyzed by phosphoric acid. This study investigates the effectiveness of phosphoric acid as a catalyst in the dilute acid hydrolysis of sugarcane bagasse at elevated temperatures (140190C) higher solids loading, and lower reaction time (10 minutes) The ability of the resulting hydrolysate to be fermented by ethanologenic E. coli without a prior detoxification step is also investigated. Establish a C orrelation b etween Solubilized Sugar Viscosity, and A bility to Pump for P retreated and Enzymatically D igested S ugarcane B agasse Slurries The most cost effective process configurations in a lignocellulose to ethanol process involve consolidation of unit operations. The use of phosphoric acid has been

PAGE 36

36 shown to facilitate the ferm entation of hemicellulose hydrolysates avoiding detoxification and a costly liquid-solid separation of cellulose enriched fiber. However, the combination of solid and liquid in bagasse slurries presents mixing and pumping challenges. Considerable bridging has been observed among fibers of sugarcane bagasse that severely inhibits mixing and pumping. Attempts at pumping these slurries typically results in dewatering and blockage. This study investigates the effects of cellulase treatments on the relative vis cosity and flow properties of acid pretreated sugarcane bagasse fiber. Viscosity results will be correlated with enzymatic sugar solubilization under a wide range of conditions. Design a U nified F ermentation P rocess that Co mbines C ellulose H ydrolysis, H emicellulose H ydrolysate, and F ermentation in O ne Vessel Research focused on the development of more robust biocatalysts with increased resistance to inhibitors (e.g. furfural, hydroxymethylfurfural, acetate) present in hemicellulose hydrolysates, have fa cilitated the fermentation of undetoxified hydrolysate. The investigation of the genetic basis for the increased tolerance has yielded potential genetic targets that can be used to further improve industrial biocatalysts. This study describes the development of the ethanologenic strain of E. coli, MM160, which is resistant to inhibitors present in hemicellulose hydrolysates. A fermentation process (L+SScF) will be developed that includes an initial liquefaction step with subsequent fermentation of both hemi cellulose and cellulose derived sugars. Str ain MM160 will be used in the investigation of the fermentation of varying solids loadings of undetoxified phosphoric acid pretreated sugarcane bagasse slurries pretreated at various temperatures with phosphoric acid

PAGE 37

37 1CHAPTER 2 OPTIMIZING THE SACCH ARIFICATION OF SUGAR CANE BAGASSE USING D ILUTE PHOSPHORIC ACID FOLL OWED BY FUNGAL CELLU LASES Introduction Lignocellulosic biomass (LCB) represents a potential carbohydrate feedstock for the fermentative production of renewable fuels and chemicals (Demirbas, 2005; HahnHagerdal et al. 2006; Ohgren et al. 2007). Success of the emerging LCB to ethanol industry will require process simplification to reduce the investment in construction and a societal commitment to provide a stable market for renewable fuels. Much of the process complexity results from fundamental differences in the biological function and structure of the feedstock. Starch produced in grain serves as a temporary energy storage polymer with glycosidi c linkages that can be readily hydrolyzed to mobilize glucose for germination and plant growth. In contrast, LCB has been designed by nature to serve as a more permanent structural composite of lignin and carbohydrates that resist deconstruction by bacteri a and fungi. As a consequence, more aggressive chemical and physical conditions and more complex processes are required for the disassembly and fermentation of LCB derived carbohydrates than are required for grain. LCB of terrestrial plants is composed pri marily of the thermoplastic lignin (15 25%) and two carbohydrate polymers, cellulose (35-50%) and hemicellulose (20 35%) (Demirbas, 2005; Jeffries, 1994) Both carbohydrates must be depolymerized to soluble sugars prior to fermentation. Ethanol production from LCB by current small commercial 1This chapter was reprinted with permission from Elsevier. Geddes, C.C., Peterson, J.J., Roslander, C., Zacchi, G., Mullinnix, M.T., Shanmugam, K.T., Ingram, L.O., 2010. Optimizing the saccharification of sugar cane bagasse using dilute phosphoric acid foll owed by fungal cellulases. Bioresour. Technol. 101, 18511857.

PAGE 38

38 plants involves combinations of the following steps (Galbe and Zacchi, 2007; Ingram et al. 199 7 199 9 ): 1) size reduction; 2) chemical pretreatment with acid or base; 3) hydrolysis of pentose -rich hemicellulose (enzymes or acid); 4) liquid solid separation of hemicellulose hydrolysate from insoluble fiber; 5) purification of the hemicellulose sugars (liming or ion exchange); 6) enzymatic or acid hydrolysis of cellulose; 7) separate or combined fermentation of cellulo se derived hexose and pentose -rich (hemicellulose) hydrolysate s; 8) liquid-solid separation of beer; 9) production (or purchase) of enzymes for saccharification; 10) distillation; 11) waste water treatment for reuse or discharge; and 12) use of residual fiber solids as boiler fuel. Process improvements are needed that eliminate some of these steps to reduce the complexity and cost of LCB to ethanol processes. LCB can be chemically treated to render cellulose more accessible to hydrolytic enzymes (Chandra et al. 2007; Mosier et al. 2005 a ; Ohgren et al. 2007) While there are a variety of methods available, steam treatments with dilute mineral acids have the added advantage of depolymerizing hemicellulose into a pentose-rich hydrolysate of sugar monomers. However, dilute acid hydrolysis is not as selective as enzymes and side products are produced that retard microbial growth and fermentation (Palmqvist and Hahn-Hagerdal, 2000a, 2000b; Zaldivar and Ingram, 1999a, 1999b) The presence of these side products is the primary reason for a liquid-solid separation prior to fermentation, a step that facilitates purification of hemicellulose sugars by over liming, charcoal, or ion exchange (Frazer and McCaskey, 1989; Ingram et al. 1999; Larsson et al. 1999).

PAGE 39

39 Most mineral acids can be used to catalyze hemicellulose hydrolysis (Bhandari et al. 1984; Mosier et al. 2005a ; Ohgren et al. 2005; Sassner et al. 2005; Schell et al. 2003). Sulfuric acid has been widely investigated and is the least expensive. Phosphoric acid is approximately 20-times more expensive but may also provide a co product opportunity as a valuable component of plant fertilizers. Initial studies concerning animal feeds have shown that the hemicellulose of rye grass straw can be hydrolyzed with ph osphoric acid to release sugars which supported the growth of yeasts after neutralization with ammonia (Israilides et al. 1978). More recent studies have demonstrated that the hemicellulose in dilute suspensions of corn stover (Um et al. 2003), sugarcane bagasse (Gamez et al. 2004 2006), and sorghum straw (Vazquez et al. 2007) can be effectively hydrolyzed in solutions containing 26% phosphoric acid at 100-122C. The lower level of toxins associated with phosphoric acid and the reduced need for exotic metallurgy offer potential advantages that justify further investigation. In this part of the study, investigations of phosphoric acid were extended to include higher temperatures, high solids conditions, and lower acid usage with sugarcane bagasse as a m odel substrate. Materials and Methods Materials Sugarcane bagasse was generously provided by Florida Crystals Corporation (Okeelanta, FL) Corn steep liquor (approximately 50% solids) was purchased from Grain Processing Corporation (Muscatine, Iowa) Biocellulase W (164 mg protein/ml; 50 filter paper units/ml) was generously provided by Kerry Biosciences (Cork, Ireland). Novozyme -glucosidase ( 277 cellobi a se units/ml) was purchased from Sigma -

PAGE 40

40 Aldrich (St. Louis, MO). Thymol, phosphoric acid, potas sium hydroxide, and other salts were purchased from Thermo Fisher Scientific (Waltham, MA). Standard A nalysis of Sugarcane B agasse C omposition Samples of bagasse were solubilized in sulfuric acid and analyzed using procedures developed by the U. S. National Renewable Energy Laboratory (Sluiter et al. 2008). Effect of P hosphoric and Sulfuric A cid C oncentrations at 145C Sugarcane bagasse (0.5 g) was placed in a 250ml flask containing 100 ml of dilute mineral acid (0 to 35 g/L). After soaking for 4 h (22C) excess water was removed by filtering through a 150 mesh polyester fabric. The resulting acid -soaked bagasse (88% moisture) was autoclaved in a Hirayama HA 305M Hiclave autoclave (Saitama, Japan) for 1 h at 145C (2 h cycle time) After cooling, sufficient deionized water was added to adjust the contents to 120 g (total) Flask contents were mixed for 2 h and sampled for analysis of sugars, furans, and organic acids. Effect of Temperature o n Steam Treatment of Sugarcane B agasse w ith 1% (w/w) H3PO4 Sugarcane Bagasse (55% moisture) was soaked for 2 hours in a 10-fold excess of phosphoric acid solution (1 % w/w including moisture in bagasse). Acid-impregnated bagasse was dewatered to 52% moisture using a Tinkturenpressen (Hubert Schwank, Germany) hydraulic pre ss (approximately 10.8 g acid/kg dry bagasse) prior to loading into a steam gun Design and operation of this steam gun have been previously described (Palmqvist et al. 1996) Large valves allowed samples to be rapidly heated (140 temperature; 12 min total cycle time). A total of 3 kg dry weight of acid -impregnated

PAGE 41

41 bagasse was processed in six runs at each tempe rature, pooled, and mixed. Processed samples were stored at 4C. Material balances were estimated using an average of the bagasse dry weight introduced into the steam gun and the dry weight of the recovered products as the reference value to express compon ents as g/kg bagasse (dry weight). Enzymatic H ydrolysis of C ellulose after Steam T reatment with Phosphoric A cid Water -insoluble solids (WIS) remaining after phosphoric acid treatment were determined after water washing. Digestibility of washed fiber was m easured using glucosidase (277 cellobi a se units/ml). Biocellulase W contains an unrefined concentrated mixture of enzymes including cellulases and xylanases from fungal culture broth. Samples of WIS (1.0 g) were added to 250ml flasks containing a single glass marble (mixing aid). Thymol (10 mg) was added as a preservative. Sufficient sodium acetate buffer (50 mM; pH 4.8) containing Biocellulase W (0.5 FPU/g, 5.0 FPU/g, o r 50.0 FPU/g) and 0.1 ml of Novozyme 188 was added to adjust the fluid contents to 100 g (total). Flasks were incubated at 50 C and 120 RPM for 4 days. Samples (1 ml) were removed at 24h intervals after adjustment for any evaporative loss and stored at 2 0 C until analyzed. Phosphoric A cid M ethod for A nalysis of B agasse C omposition Samples of sugarcane bagasse were hydrolyzed with dilute phosphoric acid and cellulase to estimate structural carbohydrates and sugar yields. Samples (0.5 g for cellulose analy sis and 5.0 g for analysis of soluble sugars and hemicellulose sugars) were weighed into 250-ml flasks and resuspended in sufficient dilute phosphoric acid (2% w/w) or deionized water (water soluble sugars) to bring the contents to 100 g

PAGE 42

42 (total). These wer e soak ed for 3 h and autoclaved for 1 h (145C dilute acid treatment; 120C for water soluble sugars). After adjustment for evaporative losses, samples were removed to measure soluble sugars (deionized water; primarily glucose, fructose, and sucrose) or hemicellulose sugars (2% w/w phosphoric acid; primarily xylose). Cellulosic glucose content of the sugarcane bagasse was determined using flasks containing a single marble to aid in mixing. Bagasse (0.5 g bagasse dry weight) was autoclaved with 2% (w/w) phos phoric acid (100 g total weight of contents) for 1 h at 145C. After cooling and adjustment with 45% (w/v) KOH to pH 5.0, thymol (10 mg), -glucosidase (0.5 ml), and water were added to adjust the contents t o 100 g (total). Cellulose digestion was carried out for 72 h at 50C (100 RPM ). Soluble sugars were analyzed after adjustment for any evaporative losses. Structural carbohydrates were estimated from the sum of sugars released during the combination of aci d hydrolysis (steam treatment) and cellulase digestion after subtraction of the water -solubilized sugars (sucrose, glucose and fructose ). Results are expressed as sugar or polymer per kg bagasse (dry weight). Toxicity and F ermentation of H emicellulose H ydrolysate ( Phosphoric A cid) The toxicity of hemicellulose hydrolysates prepared using the steam gun (1% phosphoric acid at 140 evaluated by measuring ethanol production using E. coli LY160 (Yomano et al. 2008) Inocula were grown in shake flasks (125 ml, 120 RPM 16 h, 37C) containing 50 ml of AM1 medium (Martinez et al. 2007) supplemented with xylose (50 g/L). Hydrolysates were separated from fibrous residues using a Centra CF basket centrifuge (International Equipment Company, Needham Heights, MA; 4 minutes at 3000 RPM ) and neutralized to pH 6.5 with 45% (w/v) KOH Fermentation tests were

PAGE 43

43 conducted in small pH -controlled vessels with a 100ml working volume (150 RPM 37C) as previously described (Ohta et al. 1991) Fermentation broths contained 70 ml of hydrolysate, 25 ml sterile corn steep liquor solution (10% dry weight ), 0.15 ml AM1 trace elements (Martinez et al. 2007), 0.15 ml 1 M Mg SO4*7H2O and 5 ml of inoculum (grown in shake flasks at 120 RPM for 16 h) Samples were removed for analysis at 24h intervals. Analytical Methods Moisture content was determined using a Kern model MLB 50 -3 moisture analyzer (Balingen, Germany). Sugars were measured by highperformance liquid chromatography (HPLC) using an Agilent Technologies 1200 series HPLC system (Santa Clara, CA) equipped with a model G1314B refractive index detector. Sugars were separated using a BioRad (Hercules, CA) Aminex HPX 87P ion exclusion column (300 x 7.8 mm i.d.) fitted with a Phenomenex (Torrance, CA) Carboh Ca 4 guard column (4 x 3 mm) Sugars were analyzed at 80 C using nano pure water as the mobile phase (0.6 ml/min). Organic acids and furans were measured by HPLC using an Agilent Technologies 1200 series HPLC system equipp ed with dual detectors (refractive index and UV210nm) and a BioRad Aminex HPX -87H column (45 2SO4 as the mobile phase, 0.4 ml/min) Total furans (furfural plus hydroxymethyl furfural) were measured by ultraviolet absorption (Martinez et al. 2000b). Ethanol was measured using an Agilent Technologies 6890N Network gas chromatography system equipped with a wide bore HP -PLOT Q column (0.5 mm diameter x 30 meters; J&W Scientific, Folsom, CA). Statistical A nalysis Results are reported as means with standard deviations (n=3) unless otherwise indicated. Graphpad Prism (Graphpad Software, San Diego, CA) was used to perform

PAGE 44

44 two way ANOVA (analysis of variance) with a 95 % confidence level A student t test was used to compare the experimental results for the two methods of compositional analysis (Table 2 1) Differences in means were judged significant when P values for the null hypothesis were 0.05 or less. Data for soluble sugars released at different treatment temperatures (Figure 2 3 A) was fitted to a linear regression and plotted with 95 % confidence intervals (dashed lines). Results and D iscussion Effects of Phosphoric and S ulfuric A cid on H emicellulose H ydrolysis (145C, 1 h) Initial tests were conducted with excess acid (approximately 88% moisture) for 1 h at 145C (Figure 2-1 A). Under these conditions, hydrolysis was concentration dependent from 01.0% (w/w) phosphoric acid. Steam treatments with greater than or equal to 1% (w/w) phosphoric acid and above solubilized approximately 76% of the xylan sugars based on the NREL analysis of composition (Table 2 -1). Yields of individual sugars (Figure 2 1B ) remained essentially constant for all phosphoric acid concentrations above 1% (w/v) (p > 0.05), less than half the optimal phosphoric acid concentration needed for pretreatment at 100-120C (Gamez et al. 200 4, 2006 ; Israilides et al. 1978; Vazquez et al. 2007). Sulfuric acid has been widely investigated for hemicellulose hydrolysis and w as included for comparison (Figures 2 1 A and 2 -1 C ). Although sulfuric is a stronger acid than phosphoric, maximum sugar yields (total) for both were similar at 145C (1 h), 25710 g sugar/kg bagasse (dry weight) and 24613 g sugar/kg bagasse (dry weight), re spectively (p > 0.05). As previously reported (Um et al. 2003), sulfuric was more effective than phosphoric at low concentrations. The increase in glucose observed with 3.5% (w/w) sulfuric acid is probably due to cellulose hydrolysis (Figure 2 1 C ) and did

PAGE 45

45 not occur with phosphoric acid. This increase in glucose was accompanied by a sharp decline in xylose. Differences in the concentrations of individual sugars with sulfuric and phosphoric acid can be explained in large part by the formati on of degradation products (Figure 2 2) Furan dehydration products remained low with all concentrations of phosphoric acid, approximately 1/3 the levels observed with sulfuric acid (primarily furfural from pentose dehydration; p < 0.05). Other sugar degradation products (l evulinic acid and formic acid) were abundant in sulfuric acid hydrolysates and essentially absent with phosphoric acid (p < 0.05). Acetic acid, a natural constituent of hemicellulose (Jeffries, 1994; Mitchell et al. 1990), was present in hydrolysates with both acids. The higher levels of acetic acid observed with sulfuric acid (p < 0.05) may also include products from sugar destruction, consistent with the presence of formic and levulinic acids. Both acetate and the degradation products of sugars represent potential inhibitors of microbial growth and fermentation (Palmqvist and Hahn-Hagerdal, 2000a, 2000b; Zaldivar and Ingram, 1999a, 1999b; Zaldivar et al. 2000) Increasing the T emperature to R educe T reatment T ime and P hosphoric A cid U sage Treatment times (1 h), phosphoric acid usage (73 g/kg bagasse dry weight ), and the high moisture content (88% moisture) used in our initial investigations are clearly unsuitable for most practical applications. The high moisture content in particular would greatly increase the energy required for processing. Based on previous studies of others defining the quantitative relationships between sulfuric acid concentration, time, and temperature (Chum et al. 1990), it seemed reasonable to assume that most of these could be assigned values judged desirable for industrial application (50%

PAGE 46

46 moisture, 10 g phosphoric acid/kg bagasse dry weight and 10 min treatment time) and optimized for treatment temperature. F or these tests, acid usage near the desired value of 1% (w/w) was conveniently obtained by soaking in 1% (w/w) phosphoric acid and dewatering to approximately 50% solids using a hydraulic press. Tests have confirmed that a similar dewatering of bagasse can be readily achieved with a Vincent Corporation model CP 4 screw press (Tampa, FL). The steam gun provided rapid heating and cooling (Palmqvist et al. 1996) and allowed reproducible use of a short treatment time (10 min). Condensate accumulated within the steam gun during treatment and increased the moisture content of the expelled samples to around 80%. Six steam treatments (1 kg each) at each temperature were accumulated in the cyclone receiver. These were mixed to provide a large pooled sample for furth er study. The yield of sugar monomers was nearly linear with temperature from 140C to around 180C (Figure 2 -3 A) and can be described using a linear regression (R2 = 0.9298; 95% confidence interval as dashed lines 8 359 512 3 x y ). Although more scattered (Figure 2 -3 B), corresponding decreases (17-34%) in the weight of cellulose enriched residue (WIS) were also observed Most of the acid solublized sugar was derived from hemicellul ose with xylose dominating (Figure 2-3 A). The small increase in glucose at 180C and 190C may reflect limited hydrolysis of cellulose. Total sugar yields of 215299 g/kg bagasse (untreated dry weight) were obtained between 160C and 190C. Lower levels of products from sugar destruction were observed at all treatment tem peratures with phosphoric acid (Figure 2 3 C ) as compared to sulfuric acid (Figure 2 2 B) (Grohmann et al. 1985 1986; Martinez et al. 200 0a, 2001 ). F urfural production

PAGE 47

47 was very low at temperatures of 160C and below, but increased at higher temperatures. Hydroxymethyl furfural remained low with a small rise at 190C Formic acid, a minor component, increased from treatments at 140 C to 170C and remained constant while acetate (constituent of hemicellulose) levels increased with temperature throughout the full range tested. Furfural and other sugar degradation products are known to retard the growth and fermentation of many biocatalysts (Palmqvist and HahnHagerdal, 2000a, 2000b; Zaldivar and Ingram, 1999a, 1999b; Zaldivar et al. 2000). A treatment temperature of 160C provided reasonable sugar yields with minimal side products from sugar destruction. Effect of P hosphoric A cid P retreatment on the Enzymatic H ydrolysis of B agasse Table 2 2 contains the results from digestion of WIS with Biocellulase W after p hosphoric acid treatment at different temperatures. Two concentrations of cellulase were investigated (0.5 FPU/g WIS and 50 FPU/g WIS). Results obtained after 96 h of saccharification (g glucose/kg untreated bagasse (dry weight) are presented. More than ha lf of the enzymatic hydrolysis occurred during the initial 24h period (not shown). Cellulase activity is known to be nonlinear with respect to enzyme dosage (Bommarius et al. 2008; Xu and Ding, 2006). The 100 -fold difference in enzyme levels altered sugar yields by 3-fold or less. At the high cellulase loading, near quantitive yield of glucose w as obtained for bagasse treated at 160 C or above (Table 2 -2). The mean glucose yield for steam treatments at 160190 C (50 FPU/g WIS, 96 h; p > 0.05) was 42028 g /kg bagasse dry weight, similar to that measured using the NREL method and phosphoric acid methods (Table 2 1). Bagasse samples treated at 140 C and 150 C

PAGE 48

48 were significantly more recalcitrant to enzyme digestion at both enzyme loadings (Table 2 -2) than bagasse treated at higher temperatures (p < 0.05). Incubation of WIS with the high concentration of Biocellulase W (50 filter paper units/g WIS) released up to 55 g/kg bagasse of additional hemicellulose sugar (xylose and arabinose) (Table 2 2). Since Bioc ellulase W is an unrefined fungal broth, it is not surprising to find that these hemicellulase activities are also present. Hemicellulose sugars liberated by Biocellulase W declined with increasing treatment temperature, consistent with lower residual levels of this polymer. Total sugar yields from acid hydrolysis and enzyme treatment (Table 2 -2) were essentially constant at 160C 180C (p > 0.05), similar to the total sugar content estimated using the NREL procedures for bagasse analysis (Table 2 1). It is interesting to note that total sugar yields were 60% of the total in untreated bagasse using the lowest concentration of Biocellulase W tested (0.5 filter paper units/g WIS) and a treatment temperature of 160C and above. Assuming an overall efficiency of 90% (combination of sugar conversion to ethanol and ethanol recovery), approximately 254 L (67 gal U S. ) of ethanol could be produced/metric ton of dry glucosidase). A maximum of 416 L (110 gal U S. ) of ethanol could be produced using 100-fold higher levels of cellulase enzymes. Phosphoric A cid M ethod for A nalysis of B agasse C omposition The high yields of sugars from phosphoric acid hydrolyzed hemicellulose and the subsequ ent high yields of glucose from enzymatic digestion of residual fiber (WIS) were investigated as a method to estimate bagasse composition and potential sugar yields. For this analysis, a steam -treatment temperature of 145 C (1 h) and an enzyme loading

PAGE 49

49 of 1 00 FPU/g of bagasse (dry weight) was used. Table 2 -1 shows a comparison of results from analyses using the NREL procedures in which samples are first dissolved in concentrated sulfuric acid and the phosphoric acid method. Although the results for most indi vidual sugars (except galactose) were not statistically different (p > 0.05), all values with the phosphoric acid method were lower than with the NREL method. The sum of sugars measured with the phosphoric acid method was only 90% of the value obtained for the NREL method, a difference judged to be significant (p < 0.05) Note values reported with the NREL method also include assumptions and corrections for sugar destruction that are absent in the phosphoric acid method. Values reported with the phosphoric acid method represent sugars that can be recovered and are potentially available for bioconversion. Toxicity of Phosphoric A cid H ydrolysate Phosphoric acid treatments were previously shown to produce lower levels of inhibitory compounds than sulfuric acid at 100C -120C (Gamez et al. 2004 2006 ; Israilides et al. 1978; Um et al. 2003; Vazquez et al. 2007) This was confirmed for higher tem peratures by our studies (Table 2 -3). The toxicity of hemicellulose hydrolysates was tested in pH -controlled fermentations using an ethanologenic derivative of E. coli KO11 (Ohta et al. 1991), strain LY160 (Yomano et al. 2008) Hemicellulose hydrolysate was prepared from each steam treatment (140C 190C; 1% w/w phosphoric acid, 10 min) using a basket centrifuge to separate fiber solids. Neutralization, nutrients, and inoculum resulted in a 30% (v/v) dilution of hydrolysate. Concentrations of sugars and i nhibitors in the fermentation broth are presented in Table 2 -3.

PAGE 50

50 All hydrolysates prepared at temperatures below 180C were fermented with yields of 77 -91% of the theoretical maximum (Table 2 3) Fermentation of hemicellulose hydrolysate from 140C was near complete after 24 h. Hydrolysate s from pretreatment at higher temperatures exhibited significant lags in fermentation that were related to treatment temperature and the abundance of potential inhibitors. Similar lags have been previously observed with oth er microorganisms in response to the addition of furfural to fermentation broths (Almeida et al. 2007; Zaldivar and Ingram, 1999a). Steam treatment at 160C represented an exception, exhibiting a longer lag than hemicellul ose hydrolysate from 170C despit e lower concentrations of furfural and acetate. Once fermentation began, rate of ethanol production w as surprisingly similar for all Hemicellulose hydrolysate s from treatments at 180C and 190C were not fermented during 1 week of incubation. Although the fermentation of phosphoric acid hydrolysates without sugar purification was slow, further improvements may allow omission of early liquid separations leading to process simplification. Conclusions Steam treatment of sugarcane bagasse with low level of pho sphoric acid (1% w/w of bagasse dry weight) at elevated temperatures is an effective method to hydrolyze hemicellulose and increase the accessibility of cellulose for enzymatic digestion. However, phosphoric acid ($1.75/kg) is currently 20 times more expensive than sulfuric acid. At a usage rate of 10 kg per metric ton biomass dry weight (80 gal U S ethanol per metric ton), phosphoric acid alone would cost $0.22 per gal ethanol. Thus it is essential for phosphoric based processes to recover most of the value of added chemicals (phosphate, N, Mg, K, and trace metals) as a fertilizer, while minimizing environmental impact of any residual phosphate. Calcium can be used to precipitate

PAGE 51

51 around 80 % of the phosphate. Remaining phosphate and mineral nutrients could be used beneficially during irrigation of nearby energy crops The added expense of phosphoric acid may be further offset by a reduction in the cost of plant construction. For temperatures and pressures used in this study, a variety of stainless steel alloys should be acceptable (confirmed by coupon tests; unpublished). Side reactions of sugars were minimal with phosphoric acid pretreatments resulting in only low levels of potential inhibit ors. Sugar production from hemicellulose increased with temperature reaching a peak at 180C (300 g total sugar/kg bagasse). Hemicellulose hydrolysates from treatment temperatures below 180C could be fermented (slowly) by ethanologenic E. coli without fur ther purification. With further improvements, this may allow elimination of costly equipment for the separation of hemicellulose hydrolysate from fiber and the need for separate fermentations. The hemicellulose sugar mixture (xylose, glucose, arabinose, mannose, and galactose) was effectively metabolized with fermentation yields of 7791%. Low doses of cellulase glucosidase) were remarkably effective in solubilizing glucan and residual xylan after phosphoric acid pretreatments when tested as very dilute suspensions of fiber. From 23-45% of the glucan (100 200 g glucose/kg bagasse) was solubilized using an enzyme loading of only 0.5 FPU/g of steam -treated fiber (WIS).

PAGE 52

52 Table 2 1. Sugar composition of sugarcane baga sse (% dry weight) Sample (m ethod) Glucose Xylose Arabinose + Mannose Galactose Total FL b agasse (NREL) 43.8 1.4 23.4 2.1 5.6 0.7 3.5 0.4 76.3 2.7 FL b agasse ( p hosphoric) 41.2 1.7 23.0 2.7 3.8 2.7 2.0 0.5 70.1 4.2 Reprinted with permission from Elsevier. Geddes, C.C., Peterson, J.J., Roslander, C., Zacchi, G., Mullinnix, M.T., Shanmugam, K.T., Ingram, L.O., 2009. Optimizing the saccharification of sugar cane bagasse using dilute phosphoric acid followed by fungal cellulases. Bioresour. Technol. 101, 1851 1857. (Page 1853, Table 1). Note: Results shown as mean standard deviation (n=4). Table 2 2. Solubilization of sugars from sugarcane bagasse by steam treatments with p hosphoric acid followed by 96 h enzyme (Biocellulase W) digestiona Steam treatment t emperature (C; 10 min) Acid hydrolysateb (g sugar/kg bagasse) glucosidase (50 filter paper units/g WIS) Total (g sugar/kg bagasse) Glucose (g sugar/kg bagasse) Xylose+Arabinose c (g sugar/kg bagasse) 140 130 1 338 24 41 18 509 30 150 154 6 341 19 55 4 550 20 160 215 8 452 11 53 3 720 14 170 234 6 419 18 52 4 705 19 180 299 4 383 9 38 3 720 10 190 285 2 424 5 39 4 748 6 glucosidase (0.5 filter paper units/ g WIS) 140 130 1 102 1 43 2 275 2 150 154 6 133 8 38 1 325 10 160 215 8 185 5 37 2 437 10 170 234 6 176 3 29 3 439 7 180 299 4 164 10 11 7 474 13 190 285 2 198 5 7 7 490 9 Reprinted with permission from Elsevier. Geddes, C.C., Peterson, J.J., Roslander, C., Zacchi, G., Mullinnix, M.T., Shanmugam, K.T., Ingram, L.O., 2009. Optimizing the saccharification of sugar cane bagasse using dilute phosphoric acid followed by fungal cellulases. Bioresour. Technol. 101, 1851 1857. (Page 185 5 Table 2 ). a Mean standard deviation (n=3) b Total sugars released by steam treatment with dilute phosphoric acid c Additional hemicellulose sugars released during enzy matic digestion.

PAGE 53

53 Table 2 3. Composition and fermentation of hemicellulose hydrolysates Steam t reatment t emperature Furfural (g/L) Acetate ( g/L) Sugar (g/L) Max. e thanol (g/L) Ethanol y ield (% theor.) Fermentation l ag (h) 140 C 0, 0 1.73, 1.76 20.54, 21.25 9.52, 9.01 91, 83 0 150 C 0.18, 0.23 2.13, 2.14 25.43, 26.14 10.6, 10.33 82, 77 24 160 C 0.50, 0.57 3.45, 3.46 34.95, 36.20 16.3, 16.42 91, 89 96 170 C 1.31, 1.38 4.01, 4.04 35.63, 36.47 14.4, 15.02 79, 81 72 180 C 1.82, 1.88 3.44, 3.48 39.84, 40.25 No Growth No Growth >120 190 C 2.80, 2.94 3.90, 4.00 34.84, 35.12 No Growth No Growth >120 Reprinted with permission from Elsevier. Geddes, C.C., Peterson, J.J., Roslander, C., Zacchi, G., Mullinnix, M.T., Shanmugam, K.T., Ingram, L.O., 2009. Optimizing the saccharification of sugar cane bagasse using dilute phosphoric acid followed by fungal cellulases. Bioresour. Technol. 101, 1851 1857. (Page 185 6 Table 3 ). Note: Results from two experiments (both shown).

PAGE 54

54 Figure 21. Effect of acid concentration on monomer sugars from dilute acid treatments (1 h, 145C) with H3PO4 and H2SO4. Results shown as mean standard deviation (n=3) : A) Total sugars, B) Individual sugars released by dilute H3PO4, C) Indiv idual sugar s released by dilute H2SO4. Reprinted with permission from Elsevier. Geddes, C.C., Peterson, J.J., Roslander, C., Zacchi, G., Mullinnix, M.T., Shanmugam, K.T., Ingram, L.O., 2009. Optimizing the saccharification of sugar cane bagasse using dilute phosphori c acid followed by fungal cellulases. Bioresour. Technol. 101, 18511857. (Page 1853 Figure 1). 0 1 2 3 0 100 200 300Excess 1% H3PO4 hydrolysis (12% Solids) Excess 1% H2SO4 hydrolysis (12% Solids) Acid (%, w/w)Total Sugar (g/kg dry bagasse) 0 1 2 3 0 50 100 150 200Glucose Xylose Mannose Galactose Arabinose 1h 145 C Hydrolysis (12% Solids)H3PO4 (%, w/w)Sugar (g/kg dry bagasse) 0 1 2 3 0 50 100 150 200Glucose Xylose Mannose Galactose Arabinose 1 hr 145 C Hydrolysis (12% Solids)H2SO4 (%, w/w)Sugar (g/kg dry bagasse) C B A

PAGE 55

55 Figure 22. Effect of acid concentration on the production of side products (furans and organic acids) during steam treatment (1h, 145C, n=3) : A) Dilute H3PO4, and B) Dilute H2SO4. Reprinted with permission from Elsevier. Geddes, C.C., Peterson, J.J., Roslander, C., Zacchi, G., Mullinnix, M.T., Shanmugam, K.T., Ingram, L.O., 2009. Optimizing the saccharification of sugar cane bagasse using dilute phosphoric acid followed by fungal cellulases. Bioresour. Technol. 101, 18511857. (Page 1854 Figure 2). Formic Acetic Levulinic Propionic Furans 0 10 20 30 40 50 601.0% Phosphoric Acid 3.5% Phosphoric Acid Furans and Organic Acids (g/kg dry bagasse) Formic Acetic Levulinic Propionic Furans 0 10 20 30 40 50 601.0% Sulfuric Acid 3.5% Sulfuric Acid Furans and Organic Acids (g/kg dry bagasse) A B

PAGE 56

56 Figure 23. Effect of treatment temperature on bagasse hydrolysis (hemicellulose) and production of side products (1% w/w H3PO4, 10 min) : A) Hemicellulose sugars (n=3) B) Water insoluble s olids (single value from pooled sample) and C ) Furans and organic acids (n=3). Reprinted with permission from Elsevier. Geddes, C.C., Peterson, J.J., Roslander, C., Zacchi, G., Mullinnix, M.T., Shanmugam, K. T., Ingram, L.O., 2009. Optimizing the saccharification of sugar cane bagasse using dilute phosphoric acid followed by fungal cellulases. Bioresour. Technol. 101, 18511857. (Page 1854 Figure 3). 140 150 160 170 180 190 0 50 100 150 200 250 300 Xylose Glucose Galactose Arabinose Mannose Total Sugars Treatment Temperature ( C)Sugar (g/kg dry bagasse) 140C 150C 160C 170C 180C 190C 0 200 400 600 800 WIS (g/kg dry bagasse) 140 150 160 170 180 190 0 5 10 15 20 25Formate Acetate Levulinic Furfural HMF Treatment Temperature ( C)Furans and Organic Acids (g/kg dry bagasse) A B C

PAGE 57

57 CHAPTER 3 OPTIMIZING CELLULASE USAGE FOR IMPR OVED MIXING AND RHEOLOGICAL PROPERTIES OF ACID PRETREATED SUGARCANE BAGASSE Introduction Lignocellulosic biomass (LCB) could serve as a carbohydrate feedstock to partially replace petroleum -based fuels and chemicals (Demirbas, 2005; Hahn-Hagerdal et al. 2006; Ohgren et al. 2007; Overend and Chornet, 1987) The LCB of terrestrial plants is composed of the thermoplastic lignin (1525%) and two carbohydrate polymers, cellulose (3550%) and hemicellulose (20 -35%) (Demirbas, 2005; Jeffries, 1994) Processes for depolymerization and fermentation of LCB are more complex and more capital intensive than established technologies for cornstarch or cane hydrolysate. Unlike starch, LCB has been designed by nature to serve as a structural element that resists microbial deconstruction. Pretreatments such as dilute m ineral acids or base treatments are needed to render cellulose polymers accessible to enzymes (Chandra et al. 2007; Mosier et al. 2005a ; Ohgren et al. 2007) Steam treatment with dilute mineral acids hydrolyzes hemicellulose into a pentose -rich hydrolysate (Ingram et al. 1997, 1999; Ohgren et al. 2007) This process is accompanied by side reactions and the production of inhibitors (furans, organic acids and phenolics) that retard fermentation (Palmqvist and HahnHagerdal, 2000a, 2000 b; Zaldivar and Ingram, 1999a, 1999 b) Progress has been made in developing pretreatment conditions with phosphoric acid that minimize side products (Geddes et al. 2010) and comparatively more robust biocatalysts that have increased resistance to furans (Heer et al. 2009; Liu et al. 2009 ; Miller et al. 2009a, 2009b). These improvements could facilitate the simultaneous fermentation of hemicellulose hydrolysate and celluloseenriched fiber in a single vessel, avoiding a complex (costly) liquid-solid separation. However, physical handling

PAGE 58

58 of LCB fiber suspensions rem ains a critical issue. Considerable bridging occurs among the fibers of sugarcane bagasse (dry solid or in an aqueous slurry; with or without acid pretreatment) that severely limits mixing and pumping. At 10% or higher solids (dry weight) fibrous suspensi ons of acid pretreated bagasse retains most of the water and pour as a single tangled unit from a laboratory beaker (Dasari et al. 2009). Warwick et al. (1985) reported that following mild acid treatment of lignocellulose, small acid molecules penetrate ce ll wall capillaries (spaces between microfibrils and cellulose molecules) in the amorphous regions This external fibrillation greatly increases the water -holding capacity of LCB slurries by enhancing the abundance of capillary like regions (external surfa ce area) and potential bonding areas between fibrils and fibrils and water Attempts to pump slurries containing 10% solids content or higher at pilot scale typically result in dewatering and blockage. Although it remains to be seen if this problem persist s in very large commercial scale plants, mixing and pumping of LCB slurries represent significant challenges during pilot testing and scale-up. Previous studies have investigated the effects of particle size on rheological properties of red oak sawdust (D asari and Berson, 2007; Rezania et al. 2009) and the effect of initial solids loading on power consumption, glucose yield, and rheological properties of dilute acid pretreated corn stover slurries (Dasari et al. 2009) Decreasing the particle size of redoak sawdust from 590850 m to 33 -75 m appeared to improve the efficiency of enzymatic saccharification by over 50% ( i.e. conversion of cellulose to glucose) as well as reduce viscosities by as much as 98% using an initial solids concentration of 10% (w /w) (Dasari and Berson, 2007) A subsequent study by Rezania et al. (2009) reported that reducing the particle size of red oak sawdust to < 1 m by

PAGE 59

59 sonication did not improve glucose release by cellulases and increased the viscosity This increase in visco sity was proposed to result from dominant frictional effects at the tested particle size range. In this part of the study, the effect of cellulase treatment on the relative viscosity and flow properties of acid pretreated bagasse fiber (10% dry weight sl urries in hemicellulose hydrolysate) without particle size reduction were investigated. Relative viscosity (single phase exponential decay) was correlated with the extent of saccharification under a wide range of conditions. These results were used to model viscosity changes in an ideal continuous stirred tank reactor with different amounts of cellulase. Materials and Methods Materials Sugarcane bagasse (from milled sugarcane) was provided by the Florida Crystals Corporation (Okeelanta, FL) without any furt her size reduction The bagasse is fibrous and composed of a range of particle sizes. Fine dust -like particles make up approximately 10% of the dry weight and larger (2550 mm) sized particles make up approximately 40% of the dry weight of the bagasse. Kerry Biocellulase W (164 mg protein/ml; 50 filter paper units/ml) was provided by Kerry Biosciences (Cork, Ireland). -glucosidase ( 277 cellobiase units/ml) was purchased from Sigma Aldrich (St. Louis, MO). Thymol, phosphoric acid and other c hemicals were purchased from Thermo -Fisher Scientific (Waltham, MA). Dilute A cid Pretreatment of Sugarcane B agasse Sugarcane bagasse was received at approximately 50% moisture. Bagasse was soaked in a 10-fold excess of 1% (w/w) phosphoric acid (2 h) and dewatered to 33% dry

PAGE 60

60 weight using a Centra CF basket centrifuge (International Equipment Company, Needham Heights, MA ; 4 minutes at 3000 RPM ). This acid impregnated bagasse was autoclaved (500 g batches divided among three 1L Pyrex beakers) at 145 cooled to room temperature. Sufficient deionized water was added to adjust the total weight to 3 kg (6 times th e initial dry weight of untreated bagasse). The resulting slurry contained a total of 167 g/L acid pretreated bagasse fiber (approximately 112 g insoluble fiber/liter volume). After soaking and manual mixing to allow equilibration (2 h), most of the hydrol ysate was removed by centrifugation. Resulting acid pretreated fiber (33% dry weight) and hemicellulose hydrolysate were stored at 4 C. Samples of acid pretreated fiber were washed with water prior to carbohydrate analysis. Saccharification with Biocellula se W and -glucosidase Saccharification of acid pretreated fiber was tested using Kerry Biocellulase W and -glucosidase. Samples of acid pretreated fiber (20 g) were added to 500-ml flasks containing thymol (10 mg) as a preservative. Sufficie nt pH adjusted hydrolysate (pH 38) containing Kerry Biocellulase W (0-5.0 FPU/g dry weight acid pretreated fiber) and Novozyme 188 glucosidase (03 cellobi a se units/g dry weight acid pretreated fiber) was added to adjust the contents to 200 g (10% dry w eight acid pretreated fiber) A Kerry Biocellulase W loading of 5 FPU/g dry weight acid pretreated fiber is equivalent to 0. 1 ml/g dry weight acid pretreated fiber. Most experiments were conducted at pH 5.0 (adjusted with 45% KOH) and 55 C. Flasks with en zymes and acid pretreated fiber were incubated at 300 RPM (25 80 C) for 1 h and at 200 RPM thereafter. After adjustment for evaporative loss with deionized water, samples were removed and stored at 20 C until analyzed. The extent

PAGE 61

61 of saccharification was measured as enzyme-solubilized sugar after subtraction of sugars present before enzyme addition. Relative V iscosity M easurement Measurements of relative viscosity were used to compare rheological propert ies of acid pretreated fiber slurries following enzymatic digestion using a Brookfield DV -II + Pro Viscometer equipped with a T -bar (T -C spindle, 100 RPM ). Although values are reported in centipoise (cP), these are only useful on a comparative basis. The n onNewtonian properties of this fibrous material preclude a more rigorous interpretation. Values 20,000 cP represent near immobilization of the spindle. The relative viscosity of water, acetate buffer and hemicellulose hydrolysate were also measured for comparison (1 cP for each ). Flow P ropert y T est using G raded F unnels Plastic laboratory funnels with internal stem diameters 7 mm, 12 mm and 17 mm were used to compare the flow properties of acid pretreated fiber after various treatments with Kerry Biocell ulase W. Flow was assisted by gentle tapping. The slurries (10% dry weight acid pretreated fiber) were found to either pass through the funnel or to form a plug that resisted flow Carbohydrate C omposition and A nalyses Carbohydrate composition of bagasse (as received), acid pretreated fiber and hemicellulose hydrolysate were determined as previously described (Geddes et al. 2010). Moisture content was measured using a Kern model MLB 503 moisture analyzer (Balingen, Germany). Sugars, organic acids and f urans were measured by highperformance liquid chromatography (HPLC) using an Agilent Technologies 1200 series HPLC system (Santa Clara, CA).

PAGE 62

62 Statistical A nalysis Graphpad Prism (Graphpad Software, San Diego, CA) was used to derive equations that simulate various relationships This program was also used to perform two way ANOVA (analysis of variance) of compositional analysis using the two tailed student t -test Differences in means were judged significant when P values for the null hypothesis were 0.05 or less. Results and D iscussion Composition Bagasse samples were analyzed for carbohydrate composition over a 2year period before and after acid pretreatment (1% phosphoric acid, 1 h, 145C). Sugar compositions are expressed as g/kg dry weight (Table 3 1) Steam pretreatment with dilute phosphoric acid solubilized an mean of 360 g/kg bagasse dry weight. Analysis of the acid pretreated fiber confirmed that the hemicellulose had been selectively solubilized. Differences in composition were judged significant for all sugars except galactose (p < 0.05). As expected, glucan content of the insoluble fiber was increased by acid pre-treatment. Due to the mild conditions used, approximately 19% of the xylose, 38% of the galactose and 15% of the arabinose remained as sociated with acid pretreated fiber Hemicellulose hydrolysate s were separated from acid pretreated fiber and also analyzed (Table 3 1). These contained 35 g/L total sugar and low concentrations of potential inhibitors of fermentation (4.6 g/L acetate, 0. 5 g/L furfural and 0.3 g/L formate). Soluble sugars recovered in the hydrolysate represented 21% of the initial bagasse dry weight

PAGE 63

63 Effect of C ellulase on R elative V iscosity The effect of incubation time on viscosity was examined using three cellulase enz yme loadings (0.25 FPU, 0.5 FPU and 5.0 FPU/g dry weight acid pretreated fiber). For the highest enzyme loading, the reduction in relative viscosity was nearly complete after only 1 h (Fig ure 3 1 A). Longer times were required for lower enzyme loadings Rel ative viscosities declined to plateau values that were inversely proportional to the level of added cellulase, with no tendency to converge even after longer incubation times This plateau to a constant relative viscosity with continued incubation is simil ar to saccharification and is n ot well understood (Warwick et al. 1985) The relationship between viscosity and incubation time for saccharification can be represented by a one phase exponential dec ay for each enzyme loading (Figure 3 -1 A, 0.25 FPU -Equation 31, 0.5 FPU Equation 3 -2, 5.0 FPU -Equation 3 -3 ) 2196 23674978 0 te v (3 -1 ) 1504 24366592 1 te v (3 -2) 1 255 25614050 4 te v (3 -3) For Eq uations 31 to 3 -3 r epresent s the relative viscosity (cP) and is incubation time (hours). R -squared values were calculated as 0.991, 0.989 and 0.975 for 0.25, 0.5 and 5.0 FPU/g dry weight acid pretreated fiber, respectively, indicating excellent agreement with experimental results. Acid pretreated fiber was slurried in hemicellulose hydrolysate to simulate process conditions in which solids and liquids were not separated. Acetate and phosphate present in the hydrolysate served as buffers for pH adjustment. Although the tangled mass of fiber is far from the ideal solutions described by viscosity theory, measurements

PAGE 64

64 of relative viscosity can provide useful information regarding changes in fluid properties. Preliminary experiments were conducted with a variety of cones, paddles and sp indles (data not shown) A small T -bar spindle was found to be the most useful. Slurries of hydrolysate containing 10% (w/w) acid pretreated fiber were digested with various levels of Biocellulase W (Fig ure 31 B). Temperature and pH optima for fungal cellulases (pH 5.0 and 50C; Ou et al. 2009) were similar for Biocellulase W (pH 5.0 and 60C; Table 3 -2) Prior to enzyme addition, the T -bar was unable to rotate and registered values exceeding 20,000 cP The extrapolated initial vis cosity (t = 0) value from Figure 3 -1 A ( i.e. 25,870 cP) was used as a maximum value in Figure 3 1B. After 6 h incubation with Biocellulase W, relative viscosity was reduced by 77% with an enzyme loading of 0.1 FPU/ g dry weight acid pretreated fiber, and by 95% with an enzyme load ing of 0.5 FPU/g dry weight acid pretreated fiber Previous studies have reported that rheological properties of cellulose derivatives are related to molecular structure parameters such as molar mass and particle size (Clasen et al. 2001; Gautier et al. 1991) Viscosity reduction can be accomplished by reducing the molar mass through enzymatic degradation of polysaccharide chains Enzymatic treatment of biomass disrupts the interaction of fiber polymers such as cellulose chains creating smaller particles, which also decrease the viscosity The relationship between enzyme loading and relative viscosity (6 h incubation) can be represented by a one phase exponential decay (Figure 3 -1 B Equation 34 ): 1049 24798675 15 le v (3 -4) In Equ a tion 3 -4 represents the relative viscosity in centipoise and is enzyme loading (FPU/g dry weight acid pretreated fiber) The R -squared value was 0.9969,

PAGE 65

65 indicating a good fit with experimental results. Confidence limits have been included for each curve (p < 0.05). With this equation the viscosity of the acid pretreated fiber slurry (10% solids) can be estimated for any enzyme loading (6 h incubation). Effect of C ellulase L oading on E xtent of S accharificati on Surprisingly little saccharification was required to reduce viscosity (Figures 31 B, 3 -1 C and 3 1 D ). After 6 h, a very low enzyme loading of 0.1 FPU/ g dry weight acid pretreated fiber (3.5% of the fiber dry weight solubilized) reduced relative viscos ity by 77% (Fig ure 31 B). With 5.0 FPU/ g dry weight acid pretreated fiber (6 h), viscosity was reduced by 99% accompanied by the saccharification of 17.6% of the dry weight. The relationship between enzyme loading and saccharification (Figure 3 1 B) and th e time c ourse for saccharification (Figure 3 -1 C ) can be represented by fourth (Equation 35) and third (Equation 3 -6) order polynomials, respectively. Confidence limits have been included for each curve (p < 0.05) 4 3 284 32 7 194 9 .161 80 49 09071 0 l l l l s ( 3 -5) In E quation 3 -5 represents the amount of enzyme-solubilized sugars (% dry weight acid pretreated fiber) and is enzyme loading as FPU/g dry weight acid pretreated fiber The R -squared value is 0.9980, indicating an excel lent agreement with experimental results The enzyme loading required for a desired sugar concentration (6 h of incubation) can be estimated using E q uation35 3 202458 0 8492 0 520 7 6496 0 t t t s ( 3 -6) 3 2003901 0 1735 0 414 2 1498 0 t t t s ( 3 7) 3 2003399 0 1452 0 939 1 01837 0 t t t s ( 3 8)

PAGE 66

66 In E q uations 3 6 to 3 -8, represents the amount of enzyme-solubilized sugars (% dry weight acid pretreated fiber) and is time of enzymatic saccharification (hours). The R squared values are 0.99 for all three enzyme loadings (5.0, 0.5 and 0.25 FPU/ g dry weight acid pretreated fiber), indicating excellent agreement with experimental results Using these equations, the amount of sugar that will be solubilized by a specified enzyme loading and incub ation time ( 24 h) can be estimated. Similar trends were observed for individual sugars (Fig ure 3 1 D ). Under the mild treatment conditions used, part of the hemicellulose remained associated with the fiber (Table 3 1). This hemicellulose was solubilized during incubation with Biocellulase W consistent with the presence of additional enzymatic activities ( Figure 3 1 D ). Curves defining individual sugars were not modeled. Effect of Enzyme T reatment on F low through G raded F unnels The handling and transferring of fibrous slurries represent significant challenges for LCB conversion to fuels and chemicals. T hree different funnels with internal stem diameters of 7 mm, 12 mm and 17 mm were used to compare the flow properties of acid pretreated fiber slurries (10% w /w). Flow was tested before and after enzyme treatments (Table 3 2). Acid pretreated fiber slurries failed to flow through all three funnels prior to enzyme treatment. After 2 h of incubation, all enzyme concentrations allowed the fiber slurries to flow through the 17 mm stem even though viscosity measurements were above the measurable range ( the two highest enzyme concentrations (0.5 and 5.0 FPU/ g dry weight acid pretreated fiber) permitted flow through the 12 mm stem. Non e permitted flow through the 7 mm stem (2 h). After 6 h of incubation, the highest enzyme concentration permitted flow

PAGE 67

67 through the 7 mm stem. Flow properties followed the same trends (pH, temperature, enzyme dosage, time) observed for r elative viscosity in most cases and were improved by higher level of saccharification and longer incubation times. Flow through the 17 mm, 12 mm and 7 mm stems occurred at relative viscosities of 20,000 cP, and 1%, 5% and 17% of acid pretreated fiber. More practical tests were conducted with acid and enzyme treated bagasse using a centrifugal pump (Jabsco, White Plains, NY; Model 186900000,115 V; 7.2 AMPS; 11/2 in. inlet diameter, in outlet) and a pneumatic diaphragm pump (IDEX Aodd Inc., Mansfield, Ohio; Sandpiper Model SIF Metallic Design Level 1; 1 in. inlet and outlet diameter). Positive results for flow through the 12 mm funnel stem were found to be an excellent predictor of successful pumping. Enzyme dose and treatment conditions can be used in combination to improve flow properties and pumping of acid pretreated sugarcane bagasse. Correlation between the E xtent of Saccharification and R elative Viscosity Extent of saccharification and relative viscosity were measured under a variety of conditions using six independent samples of bagasse. This data has been ass embled into a scatter plot (Figure 3 -2). Relative viscosity was dramatically reduced by a small amount of saccharification. With 5% saccharification, relative viscosity was reduced by almost 90%. However, saccharification of atleast 13% of the dry weight was required to achieve the lowest viscosity. An equation wa s de veloped to model this data (Figure 3 2 ). The decline in viscosity during saccharification was represented by a one phase exponential decay (R -squared = 0.9316). Confidence limits were also included (p <

PAGE 68

68 0.05) Eq uation 3 -9 can be used to estimate the v iscosity of the slurry based on enzyme -solubilized sugar, a property that can be correlated with pumping (Table 3 -2). 2 .592 253194412 0 se v ( 3 9) In E quation 3 -9 represents the relative viscosity in centipoise and is solubilized sugar as a percentage of acid pretreated fiber. A relative viscosity of 3,000 cP or less was needed for flow through a funnel with a 12 mm ID stem (Table 3 2) and is indicated by a horizontal d otted line on Figure 3 -2. Reduction of viscosity to this level required the solubilization of at least 50 g sugar/kg acid pretreated fiber. Effect of M ixing A cid P retreated F iber ( N o Enzyme D igestion) with Enzyme digested A cid P retreated F iber (pH 5.0, 55 C, 6 h) o n V iscosity The physical appearance of acid pretreated fiber slurries was dramatically altered by enzyme treatments. Initially, slurries with 10% fiber occupied the entire volume, poured as a single tangled unit and failed to settle indicating extensive bridging and strong interactions between fibers or between fibers and water. After enzyme digestion, slurries readily settled and behaved as a suspension of independent particles with lower viscosities. Additional experiments were conducted to determine the effect of c ombining enzyme -digested slurries of acid pretreated fiber and undigested slurries on the relativ e viscosity of the mixture (Figure 3 -3). Acid pretreated fiber slurries (10% dry weight fiber in hydrolysate) were digested for 2 h and 6 h using different amounts of Biocellulase W (0.2 5, 0.5 and 5 FPU/g dry weight acid pretreated fiber) and cooled to room temperature to minimize further enzyme action. Relative viscosity was measured before and immediately after mixing (within 1 min) with various amounts of aci d pretreated fiber (no enzyme treatment). Relative viscosities of all enzyme treated bagasse were low (200-

PAGE 69

69 2,500 cP) in comparison to undigested material ( amounts of acid pretreated fiber (no enzyme digestion) had little eff ect on relative viscosity until a threshold value was reached (Fig ure 33). At this point, further addition of acid pretreated fiber resulted in an abrupt increase in viscosity. The proportion of acid pretreated fiber (no enzyme digestion) that could be ac commodated below each threshold value varied considerably (535%) and was directly related to the initial viscosity of enzyme digested bagasse. Using 3,000 cP as a conservative maximum for pumping based on funnel experiments (12 mm stem) up to 35% fresh a cid pretreated fiber could be blended with enzymedigested acid pretreated fiber (5.0 FPU/g dry weight acid pretreated fiber; 6 h). At 0.5 FPU/g dry weight acid pretreated fiber up to 23% undigested acid pretreated fiber could be accommodated. At 0.25 FPU / g dry weight acid pretreated fiber cellulase loadings, the addition of more than 5% to 6% undigested acid pretreated fiber resulted in a dramatic rise in viscosity The relationship between acid pretreated fiber additions (fresh) to enzyme-digested fiber and relative viscosity can each be represented by equations (sigmoidal curves) for each level of enzyme (Fig ure 3 -3 0.25 FPU -Equation 3.10, 0.5 FPU -Equation 3 -11, 5.0 FPU -Equation 3 12 ). ) 067 0 30 3 (10 1 874046 1324fv (3 10) ) 028 0 70 2 (10 1 7 313535 3 224fv (3 -11) ) 091 0 67 5 (10 1 4 645898 6 811fv (3 12) In E q uations 3 103 12, represents the relative viscosity (cP) and is the fraction of undigested fiber as a percentage of total (10% dry weight). The R squared values are 0.9795, 0.9461 and 0.9946 for the equations above respectively, indicating

PAGE 70

70 excellent agreement with experimental results. The large confidence limits for sigmoidal curves obscured the curves and were omitted Tw o hour treatments were judged to have too few dat a p oints to develop a model (Figure 3-3 A). These data indicate that acid pretreated fiber that has been partially digested with enzymes has the ability to accommodate or passivate additions of fresh acid pretreated fiber (no cellulase treatment; up to 30% of total fiber) with little increase in vis cosity. This observation together with the rapid decline in viscosity resulting from limited saccharification suggests a simple mechanism. The viscosity of acid pretreated fiber is proposed to result from the tangling interactions of surfaceexposed micro-fibers as observed previously (Rezania et al. 2009; Wood et al. 1997) and the bonding between fibrils and water (water holding capacity) in the amorphous regions of fibers (Warwick et al. 1985) Digesting these small fibers with enzymes provides a smoot h surface (Wood et al. 1997), reducing viscosity with limited saccharification. These smooth enzymetreated particles would also serve as a diluent that physically hinders associations between small fibers on the surface of fresh acid pretreated fiber (no enzymes) until the threshold concentration for random associations is reached lead ing to increased viscosity (Figure 3 -3) Modeling an I deal C ontinuous S tirred T ank R eactor (CSTR) to D ecrease Viscosity T he experimental data from Figure 3 2 and 3 3 were used to estimate the upper bound viscosity of an ideal CSTR for liquefaction at different mean residence times ( ). The residence time distribution (RTD) of the reactor is a probability function describing the length of time the fluid elements of the tank s pend inside the reactor The ideal CSTR assumes that the material at the inlet is instantly and completely mixed into the

PAGE 71

71 bulk material of the reactor and that the contents of the reactor have the same composition as the outlet at all times The ideal CSTR has an exponential residence time distribution ( ( ) Equation 3 -13; Fogler, 1986): /1 ) (te t E ( 3 13) In E quation 3 -13, represents the mean residence time, defined by = / where V is the volume of the tank and Q is the inlet volumetric flow rate, and t represents residence time The fraction of the reactor contents that has a retention time between t and t+dt inside the reactor is given by ( ) The fraction of the react or contents that has a retention time less than t1 and greater than t1 are given by Equation s 3 -14 and 3 15 respectively (Fogler, 1986). 10) (tdt t E (3 14) 1 10) ( 1 ) (t tdt t E dt t E (3 15) Using Equations 3 -13 -3 15 above, the fraction wi of the reactor contents with a t (i+1) t can be expressed by E quation 3 16. t i t ie e w 1 (3 16) If the assumption that viscosity is additive is made, then the viscosity ( ) of the reactor contents can be expressed in terms of the viscosity i of fraction i using Equation 317. i i iw 0 (3 17)

PAGE 72

72 However, when Fig ure 33 data is plotted with Equation 3-1 7 for each enzyme loading, the experimental data fall below the linear viscosity curve (Equation 3 17) until a maximum viscosity value, max, is reached Using Equations 31 3 3 Equation 319 is obtained, where a b and c are constants derived from the best fit model of the experimental data (Figure 3-1 A ) for each enzyme loading using the constraint that the initial viscosity for undigested material at t = 0 is the same for all three enzyme loadings This constraint allowed the derivation of a rel ative viscosity value ( i.e. 25, 870 cP) for undigested material, which was not measurable with the instrumentation but could be used to model an ideal CSTR. i i iw 0 (3 18) t ic ibe a (3 19) Equation 320 is obtained b y s ubstituting Eq uation 3 19 for i and Eq uation 3 -16 for wi into Eq uation 3 -18 Equation 3-20 can be rearranged and expressed by Equation 3 -21. i i t i t t ice e be a01 (3 20) t c te e b a 11 1 (3 21) Taking the limit 0 of Equation 3 21 gives so LHopitals Rule ca n be used to obtain Equation 3 -22.

PAGE 73

73 1 11 lim 0c e e b at c t t (3 22) c b a 1 (3 -23) Equation 323 can be used to estimate an upper bound viscosity in a CSTR with a specific residence time and enzyme loading when the viscosity falls below a certain max. The maximum viscosities that an upper bound viscosity could be establish for ( i.e. max) at each e nzyme loading were determined graphically by transposing Equation 3 17 with Equations 3103 -12 and were 5,500 cP, 12,000 cP and 11,200 cP for 0.25 FPU, 0.5 FPU and 5 FPU respectively Predicted upper bound values for viscosity are given in Table 3 3 From this model, the predicted mean residence times to produce a slurry that can be pumped ( i.e. viscosity 0.5 FPU and 5 FPU respectively. This model was experimentally verified to some extent by a manual simulation of the CSTR for liquefaction (5 FPU cellula se/g fiber; 10% dry weight acid pretreated fiber ; 55C; pH 5.0; 60 L working volume). At hourly intervals, 16.7% of the volume was removed and replaced with new acid pretreated bagasse (and cellulase). After a 3h startup incubation, viscosity remained relatively constant (45834 arbitrary relative units) for the subsequent 9 hours of the experiment. Samples were removed hourly using a peristaltic pump confirming that the biomass suspension remained pumpable. Conclusions The simultaneous saccharification and co -fermentation (SScF) process requires high concentrations of solids to produce economical levels of ethanol. For fibrous LCB,

PAGE 74

74 these high solids are difficult to stir and transport. This study has demonstrated that low levels of cellulase enzymes are sufficient to reduce viscosity and improve the flow properties of acid pretreated sugarcane bagasse slurries. Flow properties and success in pumping can be predicted by measuring either the extent of enzymatic hydrolysis (soluble sugar), relative viscosity, or by a simple flow test using laboratory funnels. Relatively little saccharification was needed to cause a large decline in viscosity. The decline in viscosity was directly related to cellulase dosage and incubation times and can be modeled as one phase exponential decays Third and fourth order polynomials were used to describe the extent of saccharification of acid pretreated fiber as a function of enzyme dose and incubation time. Similar relationships could be established for other types of LCB and treatments using this approach. The addition of nonenzyme digested fiber to enzyme-digested fiber slurries was found to have little effect on viscosity until a threshold amount was reached This phenomen on can be explained by a simple model in which surfaceexposed micro-fibers are digested by cellulases (limited saccharification) to create smooth-surfaced particles that serve as barriers to tangling associations between microfibrils of undigested particles. These results were used to model viscosity changes (liquefaction) in an ideal continuous stirred tank reactor at different mean residence times and enzyme dosages. Based on this model, continuous addition of enzyme, makeup water and acid pretreated fib er to a mixed vessel with 2 h to 6 h mean residence time ( 5 FPU/ g dry weight acid pretreated fiber, 60C, pH 5.0) should provide a continuous supply of low viscosity slurry (1015% dry weight) for fermentation, analogous to the liquefaction step in corn dry milling

PAGE 75

75 Table 3 1. Sugar composition of sugarcane bagasse, washed acid pretreated fiber and hemicellulose hydrolysate Material Glucose Xylose Galactose Arabinosea Total s ugars Bagasse as r eceived b (g/kg) 387 20 212 20 26 11 34 26 659 40 Washed fiber after pretreatmentc (g/kg) 593 17 64 19 15 8 8 10 680 36 Hemicellulose hydrolysate d (g/L) 4 1 27 2 1 1 3 1 35 2 aArabinose may also include mannose and fructose which co elute. b mean SD (n =14) c mean SD ( n=18 ). dInhibitors present in hemicellulose hydrolysate included (g/L) : furfural (0.49 0.10), hydroxymethylfurfural (0.02 0.04), formate (0.27 0.04), and acetate (4.58 0.23) An mean of 36% of the bagasse dry weight was solubilized by acid pretreatment ; me an SD ( n=54 )

PAGE 76

76 Table 3 2. Effects of cellulase enzymes on enzyme-solubilized sugars and rheological properties Cellulase level (FPU/g dry weight) and incubation condtions Treatment t ime (h) Flow t esting (funnel stem diameter in mm)* Enzyme s olubilized s ugars (% dry weight acid pretreated fiber) Relative v iscosity (cP) 7 12 17 Effect of e nzyme l oading (FPU/g dry weight acid pretreated fiber ,pH 5, 55C) 0.0 2 N N N 0.0 0.0 >20000 0.05 2 N N Y 1.8 0.3 >20000 0.1 2 N N Y 2.0 0.0 >20000 0.25 2 N N Y 4.1 0.8 6000 0.5 2 N Y Y 4.9 0.3 1500 5.0 2 N Y Y 12.9 1.2 400 0.0 6 N N N 0.0 0.0 >20000 0.05 6 N N Y 2.5 0.0 >20000 0.1 6 N N Y 3.5 0.2 6000 0.25 6 N Y Y 5.4 0.4 2300 0.5 6 N Y Y 6.8 0.1 1300 5.0 6 Y Y Y 17.6 0.4 200 Effect of pH (55 C, 0.5 FPU/g dry weight acid pretreated fiber) 3.0 2 N N N 1.4 0.1 >20000 4.0 2 N N Y 3.7 0.2 7500 5.0 2 N Y Y 8.5 3.4 1500 5.5 2 N N Y ND 2000 6.0 2 N N Y 6.5 1.6 6000 6.5 2 N N Y ND > 20000 7.0 2 N N N 1.9 0.5 >20000 8.0 2 N N N 0.5 0.6 >20000 3.0 6 N N Y 3.2 0.1 5000 4.0 6 N Y Y 5.9 0.1 1400 5.0 6 N Y Y 8.4 2.1 500 5.5 6 ND ND ND ND ND 6.0 6 N Y Y 6.7 0.1 600 6.5 6 ND ND ND ND ND 7.0 6 N N N 0.7 0.4 > 20000 8.0 6 N N N 0.5 0.4 >20000 Effect of t emperature ( C, pH 5, 0.5 FPU/g dry weight acid pretreated fiber) 25 2 N N Y 0.9 0.1 >20000 40 2 N N Y 2.3 0.1 15000 50 2 N N Y 3.6 0.0 5000 55 2 N Y Y 4.9 0.3 1500 60 2 N Y Y 5.8 0.3 2000 70 2 N N Y 4.5 0.1 >20000 80 2 N N N 1.0 1.0 >20000 25 6 N N Y 1.4 0.3 >20000 40 6 N N Y 4.1 0.3 8000 50 6 N Y Y 7.5 0.1 3000 55 6 N Y Y 11.8 0.4 1700 60 6 ND ND ND ND ND 70 6 ND ND ND ND ND 80 6 N N Y 2.0 0.1 >20000 *The N, Y and ND indicate no flow through the funnel, flow through the funnel, and data that was not determined respectively.

PAGE 77

77 Table 3 3. Predicted upper bound viscosity for an ideal continuous stirred tank reactor (CSTR) Enzyme loading (FPU/g fiber) Vi scosity at different mean residence times ( hours) 0.5 1 2 3 4 5 6 7 8 9 10 20 30 0.25 18096 14166 10206 8214 7016 6216 5643 5213 4879 4611 4392 3347 2976 0.50 15069 10903 7326 5721 4810 4222 3812 3510 3277 3093 2943 2245 2003 5.00 8723 5327 3069 2203 1744 1461 1268 1128 1022 939 872 568 464 Note : Italicized values are above the confidence limit ( i.e. viscosity > max) and t herefore cannot be used as upper bound viscosity predictions.

PAGE 78

78 Figure 31. Effect of cellulase enzyme loading on saccharification and viscosity : A ) effect of incubation time on viscosity (Eq uations 3 1, 3 2 and 3 -3), B) effect of enzyme loading on viscosity and saccharification (6 h incubation; Eq uation 34 and Equation 35 respectively), C ) effect of incubation time on saccharification (Equations 36, 3-7 and 38 ), D ) enzyme-solubilized sugars Polynomial equations were developed that described saccharification The decline in viscosity was modeled as a one phase exponential decay Confidence limits (dashed lines) hav e been included for most curves (p < 0.05). The thick continuous lines were generated using model equations. For saccharification, sugars present at zero time have been subtracted. Reported sugars were produced solely by enzymatic action. A Kerry Biocellul ase W loading of 5 FPU/g dry weight acid pretreated fiber is equivalent to 0.1 ml/g dry weight acid pretreated fiber. 0 1 2 3 4 5 6 0 2000 4000 6000 8000 10000 12000 20000 22000 24000 26000 28000 30000 23 24 250.25 FPU/g dry wt acid pretreated fiber 0.5 FPU/g dry wt acid pretreated fiber 5.0 FPU/g dry wt acid pretreated fiber Time (hours)Relative Viscosity (centipoise,cP) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0 2 4 6 8 10 12 14 16 18 20 Solubilized Sugar Viscosity (cP) 0 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000 22000 24000 26000 28000 30000 4 5 6Enzyme Loading (FPU/g dry wt acid pretreated fiber)Enzyme-solubilized Sugar (% dry wt acid pretreated fiber)Relative Viscosity (centipoise, cP) 0 1 2 3 4 5 6 0 10 20 30 40 0.25 FPU/g dry wt acid pretreated fiber 0.5 FPU/g dry wt acid pretreated fiber 5.0 FPU/g dry wt acid pretreated fiber 23 24 25Time (hours)Enzyme-solubilized Sugar (% dry wt acid pretreated fiber) 0 1 2 3 4 5 6 0 2 4 6 8 10 12 14 16 18 20 Cellobiose Glucose Xylose Arabinose 23 24 25 Mannose Time (hours)Enzyme-solubilized Sugar (% dry wt acid pretreated fiber) A B C D

PAGE 79

79 Figure 32. Scatter plot of viscosity versus saccharification (10% w/w slurries of acid pretreated fiber in hemicellulose hydrolysate) Sugars present at zero time have been subtracted. Reported sugars were produced solely by enzymatic action Dashed lines indicate the 95% confidence limits (p < 0.05) The horizontal dotted line indicates the point at which flow through the 1 2 mm funnel occur ed. Above the horizontal dotted line, the slurry flow ed through the 17 mm funnel but not the 12 mm funnel. Below the horizontal dotted line, the slurry flow ed through the 12 mm funnel. The thick continuous line was generated using the model equation for a one phase exponential decay (Eq uation 3 -9) Flow through 12 mm diameter funnel stems was correlated with a viscosity of 3,000 cP or less. 0 5 10 15 20 25 30 35 0 2000 4000 6000 8000 10000 12000 Flow through 12 mm funnel Flow through 17 mm funnel 20000 22000 24000 26000 28000 30000Enzyme-solubilized Sugar (% dry wt acid pretreated fiber)Relative Viscosity (centipoise, cP)

PAGE 80

80 Figure 33. Effect of acid pretreated fiber additions on the viscosity of cellulase digested slurries containing 10% dry weight acid pretreated fiber Enzyme digested slurries of acid pretreated fiber were prepared by incubating for A) 2 h and B) 6 h (Equations 310, 311 and 312) and cooled to room temperature to slow enzymatic action. These were mixed with 10% dry weight slurries of acid pretreated fiber that had not been treated with enzymes. Viscosities were measured immediately (within 1 min) These data were modeled as equations for sigmoid curves ( Equations 3-10, 311 and 3-12 ), shown as thic k black lines Insufficient data points were available to model the 2 -h treatment (Fig. 3a) Values connected with thin solid lines are within range of instrumentation ( i.e. < 20,000 cP) Curves derived from Fig ure 31A were used to estimate the value at the immeasurable point for each curve and plotted as open symbols connected by a dashed arrow. 0 10 20 30 40 50 0 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000 22000 24000 26000 28000 30000 32000 0.25 FPU/g dry wt acid pretreated fiber 5 FPU/g dry wt acid pretreated fiber 0.5 FPU/g dry wt acid pretreated fiber Fiber Mixture (% undigested)Relative Viscosity (centipoise, cP) 0 10 20 30 40 50 0 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000 22000 24000 26000 28000 30000 32000 0.25 FPU/g dry wt acid pretreated fiber 0.5 FPU/g dry wt acid pretreated fiber 5.0 FPU/g dry wt acid pretreated fiber Fiber Mixture (% undigested)Relative Viscosity (centipoise, cP) A B

PAGE 81

81 CHAPTER 4 A SIMPLIFIED PROCESS FOR ETHANOL PRODUCTION FROM SUGARCANE BAGASSE USING HYDROLYSATE-RESISTANT ESCHERICHIA COLI STRAIN MM160 AND PRETREATMENT W ITH PHOSPHORIC ACID Introduction Lignocellulosic biomass represents a potential source of carbohydrates (cellulose and hemicellulose) for microbial fermentation to ethanol and other chemicals. Steam pretreatment with dilute mineral acids serves as an effic ient approach to depolymerize hemicellulose into monomeric sugars and to enhance fiber digestion by cellulase enzymes (Mosier et al. 2005a ; Ohgren et al. 2007). However, soluble side products (furfural, 5hydroxymethyl furfural, acetate, phenolics, and others) in dilute acid hydrolysates inhibit microbial growth and retard fermentation (Mills et al. 2009; Palmqvist and Hahn-Hagerdal, 2000). Potential engineering solutions to this problem (separation of hemicellulose hydrolysate, washing of fibers, and mi tigation of toxins) add process complexity (Fig ures 4 -1 A and 4 1 B). Efficient washing requires either countercurrent systems that minimize dilution or energy intensive steps to re concentrate sugars (Sassner and Zacchi, 2008). Although pretreated and washed fiber can be effectively fermented by yeast and bacterial biocatalysts with added cellulase (Dien et al. 2008; Hahn-Hagerdal et al. 2007), mitigation of toxins in hemicellulose hydrolysates generally requires overliming, charcoal, ion exchange, etc. pr ior to fermentation (Larsson et al. 1999; Martinez et al. 2001; Mills et al. 2009). Pretreatment with base is also effective in increasing the digestibility of lignocellulose by enzymes with minimal formation of toxic side products (Lau et al. 2009) Previous studies with ammonia have highlighted the effectiveness and potential savings from the co -fermentation of cellulose and hemicellulose derived sugars in a

PAGE 82

82 single vessel (Lau and Dale, 2009; Sassner and Zacchi, 2008). Inhibitors formed during dilute acid pretreatment hindered analogous co -fermentations of cellulose and hemicellulose hydrolysates without additional steps for toxin mitigation (McMillan et al. 1999). The extent of inhibitor production is affected by the severity of steam pretreatment conditions (Kabel et al. 2007) and by the choice of acid. Although sulfuric has been most commonly investigated (Mosier et al. 2005a ), weaker acids such as phosphoric acid can also be used and produce lower levels of toxic side products and reduce the need for exotic metal alloys (Geddes et al. 2010). Since mineral acids are not consumed by the pretreatment process, the higher cost of phosphoric acid as compared to sulfuric acid could be offset in part by recovery and reuse as a fertilizer. Recent p rogress has been made in the development of more robust biocatalysts with increased resistance to dilute acid hydrolysates of hemicellulose (Heer and Sauer, 2008; Liu et al. 2009) and to furans present in dilute acid hydrolysates (Heer et al. 2009; Mille r et al. 2009 a ). Several genes involved in resistance to furfural and 5 hydroxymethylfurfural have been identified as NADP(H) -dependent oxidoreductases (Laadan et al. 2008; Miller et al. 2009a ). In this part of the study, a n ethanologenic Escherichia c oli (strain MM160) with increased resistance to inhibitors in hemicellulose hydrolysates is described Using phosphoric acid pretreatment and strain MM160 sugarcane bagasse hemicellulose hydrolysate and enzymehydrolyzed cellulose could be effectively fer mented in a single vessel in a variation of the simultaneous saccharification and co -fermentation process (SScF) that included a liquefaction step, termed L+SScF.

PAGE 83

83 Materials and Methods Materials Sugarcane bagasse was provided by Florida Crystals Corporati on (Okeelanta, FL). Biocellulase W (164 mg protein/ml; 50 filter paper units/ml) was provided by Kerry Biosciences (Cork, Ireland). Novozymeglucosidase (277 cellobia se units/ml) was purchased from Sigma Aldrich (St. Louis, MO). Sulfuric acid hydrolysates of sugarcane bagasse hemicellulose were provided by Verenium Corporation and contained from 5080 g monomer sugar/L (Boston, MA) (Martinez et al. 2001). Phosphoric a cid hydrolysates of sugarcane bagasse hemicellulose were prepared at the University of Florida. Laboratory supplies and chemicals were purchased from Thermo Fisher Scientific (Waltham, MA). Steam T reatment of B agasse with 1% (w/w) H3PO4 Bagasse (approximat ely 55% moisture) was soaked for 4 hours in a 14-fold excess of phosphoric acid solution (1% w/w including moisture in bagasse), dewatered to approximately 50% moisture using a model CP 4 screw press (Vincent Corporation, Tampa, FL), and loaded into a steam reactor. This dewatered bagasse contained 10 g phosphoric acid per kg dry weight. Design and operation of this steam reactor have been previously described (Palmqvist et al. 1996). Large valves allowed samples to be rapidly heated (160 and 190 bagasse, 9.5 min at temperature, 10 min total cycle time). After steam treatment, fibrous suspensions contained approximately 30% dry matter including condensate. A minimum of 3 kg dry weight of acid-impregna ted bagasse was processed at each temperature, blended into a single sample, and used for fermentation.

PAGE 84

84 Hemicellulose hydrolysate was also prepared from pretreated bagasse (approximately 70% moisture) using a screw press. Emerging solids contained approxim ately 48% moisture and were discarded. Fine particulates were removed from the extruded liquid using a glass fiber filter (Whatman GF/D, 15 mm diameter, 27 m pore size) Resulting clarified hydrolysate was stored at 4 of resis tant mutants and as a substrate for seed cultures. Organisms and G rowth C onditions Strains used in this study are listed in Table 4 -1. These were stored as frozen stocks in 40% glycerol at 80C. Working stocks were maintained by daily serial transfers (3% inoculum) into AM1 medium (Martinez et al. 2007) containing hemicellulose hydrolysate (160 50 RPM ) were maintained at pH 6.5 by the automatic addition of 2 N KOH. Genetic M ethods E. coli strain LY180 was the parent organism for this study ( Table 4 1). This strain was previously constructed from E. coli ATCC 9637 ( Miller et al. 2009 ; Yomano, et al. 2009). Additional genetic modifications were made to a hydrolysate -resistant derivative of LY180 (strain MM105) to produce strain LY195 (carried out by Lorraine Yomano) using standard protocols (Sambrook and Russell, 2001) and those provided by manuf acturers (Invitrogen, Carlsbad, CA; New England Biolabs, Ipswich, MA; Qiagen, Valencia, CA; and Stratagene, La Jolla, CA). DNA sequencing was performed by the University of Florida Interdisciplinary Center for Biotechnology Research. Isolation of H ydrolysa te -r esistant B iocatalysts Hydrolysate-resistant derivatives of LY180 were selected by sequentially transferring broth cultures (3% inoculum) in AM1 mineral salts medium containing

PAGE 85

85 dilutions of hemicellulose hydrolysate (160 hydro lysate was neutralized by adding 5N ammonium hydroxide to a final concentration of 47 mM and adjusting the pH to 6.3 using 45% (w/w) KOH. After filter sterilization ( Nalgene PES filter, 0.45 m pore), the hydrolysate was diluted as needed with water. Remai ning components of AM1 medium (NH4H2PO4, (NH4)2HPO4, MgSO4*7H2O trace metals) were then added. Xylose was added as needed to provide a monomer sugar concentration of approximately 50 g/L. Cultures were incubated at 37 RPM ) in small pH -controlled fe rmentation vessels (250 ml). Cultures were transferred when ethanol concentrations exceeded 5 g/L (approximately 24 h). The concentration of hydrolysate was increased in the selection medium when at least 3 successive transfers produced over 10 g ethanol/L in 24 h. A resistant clone was isolated after 322 sequential transfers in sulfuric acid hydrolysate and designated MM105. Extraneous DNA regions (non-coding FRT recombinase sites) from prior genetic constructions were removed from MM105 to produce LY195. LY195 was then subjected to further sequential transfers in AM1 medium containing hemicellulose hydrolysate produced at 150C, and subsequently from hydrolysate prepared at 160C. After 139 serial transfers of LY195 in phosphoric acid hydrolysate, a second resistant clone was isolated and designated strain MM160. This clone is the product of over 2 years of subculturing (461 transfers) and more than 2,000 generations of selection. Removal of E xtraneous DNA S egments from Strain MM105 During the constr uction of LY180, non -coding FRT recombinase sites were left behind in the chromosome at the site of gene deletions and these remained in the

PAGE 86

8 6 hydrolysate-resistant mutant, MM105. Three of these extraneous segments (one each in mgsA ackA and pflB ) were rem oved from MM105 by double homologous recombination (carried out by Lorraine Yomano) using a cat -sacB cassette and Red recombinase as previously described (Datsenko and Wanner, 2000; Jantama et al. 2008). Chloramphenicol resistance ( cat ) was used to select for the initial integration of linear DNA. Resistance to sucrose was used to select for a second integration event in which sacB was deleted. Primers and plasmids used during these constructions are listed in Table 4 -1. Tolerance to H ydrolysate T oxins Tol erance to selected compounds present in hemicellulose hydrolysate was tested using 13 x 100 mm culture tubes containing 4 ml of AM1 medium (50 g/L xylose) as described by Miller et al. (2009a ). Tubes were inoculated at an initial density of 0.05 OD550nm. G rowth was measured after incubation (37C, 60 RPM ) for 48 h using a Spectronic 20D+ spectrophotometer (Thermo -Fisher, Waltham, MA). Liquefaction F ollowed by S imultaneous S accharification and C o -fermentation (L+SScF) of A cid -pretreated S ugarcane B agasse Sugarcane bagasse was pretreated with 1% (w/w) H3PO4 at varying temperatures (160 required for a 1 L fermentation volume at 10%, 12%, or 14% steam pretreated bagasse (total dry weight of fiber plus solubles). After transferring into 3L fermentation vessels (BioFlo 110, New Brunswick Scientific, Edison, NJ), sufficient deionized water was added to adjust the total volume to approximately 900 ml. Ammonium hydroxide (9.4 ml; 5 N) and 45% (w/v) KOH were added to raise the pH to 5.0.

PAGE 87

87 The resulting slurry was too viscous to mix without further treatment. Mixing was improved by partial saccharification (liquefaction) using Kerry Biocellulase W (5 FPU/g dry weight acid pretreated bagas -glucosidase (2.77 cellobiase units/g dry weight acid pretreated bagasse) during a 6 h preincubation at 55 RPM ) prior to inoculation. After cooling to 37 C and adjustment to pH 6.5 with 45% (w/v) KOH, the remaining salts in AM 1 medium (NH4H2PO4, (NH4)2HPO4, MgSO4*7H2O and trace elements) were added. Additional water was added to bring the total volume to 950 ml. The resulting fermentation broth was inoculated with 50 ml of seed culture. Seed cultures of MM160 were grown in AM 1 medium (3% inoculum from working stock; 500 ml vessels, 37 RPM 24 h) containing 40% phosphoric acid hydrolysate (160 LY180 was grown without hydrolysate. At the time of inoculation (5 % inoculum), seed cultures typically contained 6 10 g/L ethanol. Fermentations were monitored for 240 h (37RPM ) and maintained at pH 6.5 by the automatic addition of KOH (45% w/v). Analyses Carbohydrate compositions were determined as previously d escribed (Geddes et al. 2010). Moisture content was determined using a Kern model MLB 503 moisture analyzer (Balingen, Germany). Sugars, furans, and organic acids were analyzed using an Agilent Technologies 1200 series HPLC system (Geddes et al. 2010). Ethanol was measured using an Agilent Technologies 6890N Network gas chromatography system (Geddes et al. 2010).

PAGE 88

88 Statistical Analysis Prism software (Graphpad, San Diego, CA) was used to perform two-way ANOVA (analysis of variance) for comparisons. Differences were judged significant when P values for the null hypothesis were 0.05. Results Composition of Materials Used as Substrates for Fermentation The composition of bagasse pretreated wit h steam and dilute phosphoric acid at different temperatures is summarized in Table 4-2. Pretreated bagasse exiting the reactor was approximately 30% dry weight (fiber plus solubles). This material was used as the primary substrate for fermentation without washing, separation, or treatments to reduce toxicity. Solubles and fiber were separated only for analyses. Solubles, hemicellulose sugars, sugars fr om fiber digestion with excess cellulase, and inhibitors all increased with increasing treatment temper ature. The highest yield of total sugars (pretreatment and cellulase digestion) was obtained for pretreatments at 180 C and 190 C (difference not significant), although the level of combined inhibitors was 28% higher at 190 C than at 180 C. Total sugars recovered by ph osphoric acid pretreatment followed by digestion with excess cellulase ( 100 FPU/g dry weight acid pretreated fiber) were equivalent to 70% of the bagasse dry weight (approximatel y 63% glycan) with a maximum theoretical yield of 357 kg ethanol/metric ton bagasse (dry weight). Tables 4-3 and 4-4 contains the result s on the effect of enzyme digestion on concentration of individual sugar and inhibitor in the hemicellulose hydrolysates. These hydrolysates were used as a source of inhi bitors to select re sistant mutants. The concentrations of all monomer sugars and inhibitors increased with treatment temperature. A broad oligos accharide peak was observed during HPLC sugar analysis

PAGE 89

89 which decreased in size with increasing trea tment temperature (dat a not shown). This peak was resolved into monomers (xylose and glucose) and acetate by treatment with Biocellulase W (Table 4-3), consistent with the presence of xylanase, -glucosidase, and acetyl esterase activities. Enzyme tr eatment increased the concentration of monomer sugars by 32% for hydrolysate prepared at 160 C indicating an abundance of oligosaccharides, but had little e ffect on hydrolysates prepared at 180 C or 190 C. The concentration of acetate incr eased during incubation with Biocellulase W (Table 4-4). Acetyl esters are natural constituents of hemicellulose, the prim ary source of this inhibitor. Many different sulfuric acid hydrolysate s of sugarcane bagasse were provided by Verenium Corporation and used for mutant selection during the initial year. Although the details of preparation are proprietary, thes e hydrolysates contai ned monomer sugars (80-110 g/L), furans (1-2 g/L), acetate (4-13 g/L), and 1% to 4% sulfuric acid (Martinez et al., 2001). Development of Hydrolysate-resistant Strain MM160 The genetics and construction of strain LY180 have been previously described (Yomano et al. 2009; Table 4-1). This strain can effi ciently ferment all laboratory sugars that are constituents of hemicellulose. Using LY180 as the parent, resistant mutants capable of growth in dilutions of bagasse he micellulose hydrolysate were selected by repeated subculture. After 322 transfers in diluti ons of sulfuric acid hydrolysate (200-600 g/L in AM1 medium), a clone was isolat ed and designated MM105. Three extraneous DNA regions (recombinase sites from the construction of LY180) were then removed from strain MM105 by double homologous recombination, and the resulting strain designated LY195. Sequential transfers of LY195 were continued using AM1 medium

PAGE 90

90 containing phosphoric acid hydrolysates prepared at 150C initially and then shifted to hydrolysates prepared at 160C after tolerance improved. Examples of improvements in fermentation during selection are shown in Figure 4 -2. Afte r 139 sequential transfers in phosphoric acid hydrolysates, a clone was isolated and designated MM160. Strain MM160 grew and fermented well in AM1 medium containing 600 g/L phosphoric acid hydrolysate (160C), conditions that were toxic to the parent strai n LY180 (Fig ure 4 -2 B). Strains LY180 and MM160 were also compared for resistance to individual chemicals present in acid hydrolysates (Fig ure 43). MM160 was more resistant to furfural, 5-hydroxymethyl furfural (HMF), and acetate. Both strains were equally sensitive to growth inhibition by formate. Previous studies are consistent with furans and acetate as the primary toxins in dilute acid hydrolysates of hemicellulose (Miller et al. 2009 a ; Mills et al. 2009). Liquefaction Prior to S imultaneous S accharifi cation and C o -fermentation (L+SScF) Phosphoric acid pretreated bagasse was fermented using a modified SScF process without separation of the soluble hemicellulose sugars and inhibitors from fiber or treatments to remove toxins (Figure 4 -1 B). This procedure included an initial 6h enzyme treatment at 55C (pH 5) to reduce viscosity prior to inoculation (L+SScF), analogous to the liquefaction step in corn ethanol processes (Figure 4 -1 C ; Ingledew et al. 2009). This liquefaction step doubled the initial level of monomer sugars available for fermentation (Figure 44 A) and tripled the level of acetate (Figure 4 -4 B) as compared to an equivalent dilution (30% solids; 1:3 dilution at 10% solids) of hemicellulose hydrolysate (Table s 4 -3 and 4 4).

PAGE 91

91 Figure 4 5 shows a comparison of L+SScF fermentations with different levels of solids using bagasse pretreated at 160C. With these, most of the ethanol was produced during the initial 72 h and continued to increase slowly during further incubation. Sugars present during fer mentation (predominantly xylose and glucose) represent a balance between depolymerization and metabolism. Glucose was used preferentially. Xylose utilization declined with increasing amounts of bagasse and inhibitors (Table 4 4; Fig ures 4 5 B, 4 5 C and 4 -5D ). Part of the xylose remained unfermented even after 240 h. Previous research with LY180 (parent) demonstrated that furfural tolerance was also higher with glucose than with xylose (Miller et al. 2009 a ). Additional L+SScF experiments were conducted w ith 14% bagasse (dry weight of fiber plus solubles) using higher pretreatment temperatures (Fig ure 4 -6). Bagasse pretreated at 190C was toxic for MM160 and was not fermented. The highest titer (Fig ure 4 6 A) and highest yield of ethanol were produced using 14% bagasse (dry weight) pretreated at 180C (Table 4 5), although xylose was only partially metabolized (Fig ure 46 D ). The slower fermentation of samples from higher treatment temperatures is consistent with inhibition by toxins, which also increased wit h pretreatment temperature (Table 4 2 and 4 -4). Significant amount of ethanol w as produced from sugarcane bagasse by the L+SScF process ( Table 4 5), despite the lagging utilization of xylose. Under the conditions tested, ethanol yields from bagasse ranged from 169 kg/ metric ton to 207 kg/ metric ton (5163 gal/US ton), up to 57% of the maximum theoretical yield based on composition (Table 4 2).

PAGE 92

92 Discussion Previous studies with base pretreatments have demonstrated the advantages of low inhibitor production and co -fermentation of sugars derived from cellulose and hemicellulose in a single vessel (Table 4 -6). Base pretreatments are very effective in opening structure and improving enzyme digestion but produce little fermentable sugar. Saccharification of bot h cellulose and hemicellulose are dependent on added enzymes with base treatments. Dilute acid hydrolysis (typically sulfuric) has the advantage of substantially hydrolyzing hemicellulose into monomer sugars. This advantage is offset in part by the product ion of fermentation inhibitors such as furans from side reactions (Mills et al. 2009). A weaker acid (phosphoric acid) was used that produces lower level of furans and other inhibitors (Geddes et al. 2010). With phosphoric acid pretreatments of bagasse, -glucosidase activities present in Biocellulase W were needed to complete the hydrolysis of hemicellulose (Table 4 -2). Even with the lower level of inhibitors (Figure 4 3 B) formed by phosphoric acid pretreatment (as compared to sulfuric), su fficient levels were present to inhibit fermentation by the parent strain, LY180 (Figure 4 2 B). Processes based on dilute acid pretreatment and cellulase typically include additional steps for liquid-solid separations and cleanup steps to remove inhibitor s from hemicellulose hydrolysates as illustrated in Figure 4 1A (McMillan et al. 1999). This requirement for solid-liquid separation and further steps to remove toxins was eliminated by developing a more resistant biocatalyst ( E. coli MM160) and by minimizing inhibitor production using phosphoric acid. A simplified process (L+SScF) was developed in which pretreated bagasse was used for fermentation by strain MM160 wi thout further purification (Figure 4 1 B). Phosphoric acid pretreated b agasse was

PAGE 93

93 partially liquefied by adding cellulase enzymes to improve mixing prior to inoculation. An analogous liquefaction step is also used in co rn ethanol processes (Figure 4-1C, similar to the pre-saccharification proposed by Lau and Dale (2009) for AFEX-treated corn stover. After liquefaction, t he bagasse slurry containing hemicellulose and cellulosederived sugars was fermented by strain MM160 in a single vessel using mineral salts as nutrients. Glucose and xylose were co-ferm ented effectively when the lowest level of toxins was present (10% bagasse dry weigh t; 160C pretreatment). Xylose utilization lagged with increasing levels of inhibitors. Even with incomplete xylose fermentation, up to 207 kg ethanol was produced per metric to n of bagasse (dry weight) representing 57% of the maximum theoretical yield based on carbohydrate analysis (Table 4-2). Hydrolysate toxins remaining in the broth during fermentation may serve as a natural barrier to retard the growth of adventitious contaminants. The maximum yield (0.21 g ethanol/g bagasse dry weight) with phosphoric acid pretreatment and MM160 was similar to that reported for ammonia fiber expansion (AFEX) and caustic pretreatments of various biomass materials (Alizadeh et al 2005; Lau and Dale, 2009; Table 4-6). High pH pretreatment and p hosphoric acid pretreatment should not require the special alloys needed for sulfuric acid processes. Ethanol titers with phosphoric ac id pretreated bagasse (180 C) and strain MM160 averaged 29 g/L, although MM160 can produce 65 g ethanol/L using pure xylose in the laboratory (data not shown). With high pH pretreatments and paper sludge, ethanol titers of 40-45 g/L have been reported in other studies (Table 4-6). Further improvements in titer and volumetric productivity are needed for a phosphoric acid based process.

PAGE 94

94 Table 4 1. Strains, plasmids and primers. Escherichia coli Strain, Plasmid, or Primer Relevant characteristics Source or citation TOP10F' F' (lacI q Tn10 tet ) Invitrogen LY180 frdBC::(Zm frg celY Ec ldhA ::(Zm frg casABKo,), adhE::(Zm frg estZPp ackA ::FRT, mgsA ::FRT, pflB+ ( pflA FRT ycaK ), rrlE::(pdc adhA adhB Zm FRT) Yomano et al. 2009 MM105 Hydrolysate resistant mutant of LY180 This study LY190 MM105 mgsA cat sacB mgsA' cat This study LY191 mgsA This study LY192 LY191 ackA' cat sacB ackA' cat This study LY193 ackA This study LY194 LY193 pflA' cat sacB ycaK' cat This study LY195 LY194 pflA ycaK native sequence This study MM160 Hydrolysate resistant mutant of LY195 frdBC::(Zm frg celYEc ldhA::(Zm frg casABKo,), adhE::(Zm frg estZPp ackA mgsA pflB+, rrlE::(pdc adhA adhBZm FRT) This study Plasmids pCR2.1 TOPO bla kan lacZ P lac Invitrogen pKD46 P ara exo (Red recombinase), temperature conditional replicon, bla Datsenko and Wanner, 2000 pLOI4162 PacI flanked cat sacB cassette, bla cat Jantama et al. 2008 ackA region pLOI4810 LY180 ackA' FRT ackA' in pCR2.1 TOPO, bla, kan This study pLOI4815 pLOI4810 inside out PCR, self ligated, internal ackA deletion, bla, kan This study pLOI4823 pLOI4810 inside out PCR, pLOI4162 cat sacB PacI frg, cat, bla, kan This study mgsA region pLOI3937 LY180 mgsA' FRT mgsA' in pCR2.1 TOPO, Ap, Kan This study pLOI4819 pLOI3937 inside out PCR, self ligated, internal mgsA deletion, Ap, Kan This study pLOI4821 pLOI3937 inside out PCR, pLOI4162 cat sacB PacI frg, cat, bla, kan This study PflB ycaK region pLOI4813 LY180 pflA trm FRT ycaK' in pCR2.1 TOPO, bla, kan This study pLOI4814 ATCC9637 pflA ycaK' in pCR2.1 TOPO, bla, kan This study pLOI4822 pLOI4813 inside out PCR, pLOI4162 cat sacB PacI frg, cat, bla, kan This study Primers Sequence ackA cloning, deletion and sequencing ackA cloning For 5 ctggttctgaactgcggtag Rev 5 cgcgataaccagttcttcgt This study ackA deletion Up 5 gcatgagcgttgacgcaatc Down 5 gactcttccggcatagtctg This study yfbV pta sequencing For 5 ggcgttgacatgcttcacct Rev 5 tcgcacgcacgat agtcgta This study mgsA cloning, deletion and sequencing mgsA cloning For 5' tattgcgctggtggcacacg Rev 5 acggtccgcgagataacgct This study mgsA deletion Up 5 cagcaggttggcgcattgat Down 5 accggtagtgcctgttgcat This study yccT helD sequencing For 5 atggcgatgcgacgccgatt Rev 5 aacacgctggccgaagttgc This study pfl region cloning, deletion and sequencing pflA ycaK2 cloning For 5 cagatgaacgacgagatcca Rev 5 gagctgcttgaacatgacac This study pflA ycaK deletion Up 5 ggtgaacgctctcctgagta Down 5 gcagaatgaagcgcggaata This study

PAGE 95

95 Table 4 1. Continued Escherichia coli Strain, Plasmid, or Primer Relevant characteristics Source or citation ycaO focA sequencing For 5 cgctggcttctgcacttggt Rev 5 ggccattgcagcaggaagta This study focA pflB sequencing For 5 ggcctataagccaggcgaga Rev 5 gtggaggtacgaccgaagga This study pflB pflA sequencing For 5 gcgttgcgctgtacggtatc Rev 5 gccgccggaagcgttcataa This study pflA ycaK sequencing For 5 gcaccaacacggcctcagat Rev 5 gtgcgctccagaacttaacg This study ycaO int sequencing For 5 tggcgtagcactggaacgta Rev 5 gctggcgatcttcttcctgt This study focA pflB2 sequencing For 5 accactggcacaggcacaat Rev 5 agcggagcttcagtctgtag This study

PAGE 96

96 Table 4 2. Components of sugarcane bagasse after steam pretreatment with phosphoric acid. Components Composition after pretreatment a (g/kg SD bagasse dry weight) 160C 170C 180C 190C Hemicellulose hydrolysate a Solubles after pretreatment 215 12 239 18 284 21 273 14 Sugar monomers after pretreatment 115 3 161 9 222 13 232 9 Sugar monomers after enzyme digestionb 133 7 197 12 241 17 241 10 Glycan (hemicellulose+cellulose) 121 6 179 11 219 15 219 9 Inhibitors after pretreatment 8 0 12 1 20 1 30 1 Inhibitors after enzyme digestionb 14 1 18 1 25 2 34 2 Ash 8 1 14 1 13 3 10 4 Lignin plus other (by difference)f 72 13 28 21 27 26 10 17 Acid treated fiber a Insoluble fiber after pretreatmenta 774 24 750 14 713 27 716 8 Solubles after enzyme digestion 461 23 493 20 492 24 497 13 Sugar monomers after enzyme digestionc 415 19 445 14 457 23 472 12 Glycan (hemicellulose+cellulose) 377 17 405 13 415 21 429 11 Inhibitors after enzyme digestiond 8 3 4 2 4 4 3 2 Ash 8 0 2 1 5 2 5 4 Lignin plus other (by difference)f 381 30 339 19 289 34 279 14 Combined hemicellulose hydrolysate + fibera Combined sugar monomers 548 20 642 19 698 28 713 14 Combined glycan 498 18 584 17 635 25 648 13 Combined inhibitors c 22 3 22 2 29 4 37 3 Combined ash 16 1 16 1 18 1 15 1 Combined lignin plus other (by difference)f 464 18 378 17 318 25 300 13 Theoretical ethanol (liters/tonne) 354 414 451 460 a After steam pretreatment, soluble (hemicellulose hydrolysate) and insoluble fiber fractions (washed) were separated prior to analysis. Sugars re present observed monomers and dimers without corrections for degradation (n b Hemicellulose sugars after treatment with Biocellulase W (20 ml/L, pH 5, 50C, 6 h) to hydrolyze oligosaccharides. Acetate (inhibitor) also increased during enzyme digestion. Glycan was estimated by dividing total sugars by 1.1 to correct for water addition. cLignin+other = Solubles (sugar after digestion + inhibitors after digestion +ash) dSugar monomers from fiber after digestion with excess Biocellulase W (50 FPU/g fiber, pH 5, 50C, 96 h). Thymol crystals (10 mg) were added to prevent microbial growth. eSum of selec ted inhibitors: organic acids (acetate, formate) and furans (fufural and 5hydroxymethylfurfural). fLignin+other = Insoluble fiber (sugar after igestion + inhibitors after digestion +ash)

PAGE 97

97 Table 4 3. Effect of enzyme treatment on sugar monomers in hemi cellulose hydrolysate. Treatment temp.a Enzyme treatedb % Dry weightc Sugars (g/L) Dimerd Glucose Xylose Arabinose Galactose Total sugars 160C No 30.9 3.0 7.4 0.7 1.4 0.1 36.6 3.2 3.0 0.7 1.2 0.2 50 3 170C No 31.0 1.2 6.0 0.3 2.7 0.2 56.0 2.0 3.4 1.4 1.8 0.1 70 2 180C No 27.4 2.1 3.6 0.2 4.2 0.6 62.8 0.5 4.6 1.5 2.3 1.3 78 2 190C No 27.6 0.8 2.1 0.2 6.7 0.6 63.2 1.1 5.6 0.1 2.4 0.1 80 1 160C Yes 30.9 3.0 1.7 0.7 6.4 0.4 53.8 4.2 2.6 0.8 1.4 0.3 66 4 170C Yes 31.0 1.2 1.3 0.2 7.0 0.2 64.0 0.9 3.6 1.1 2.0 0.4 78 2 180C Yes 27.4 2.1 0 7.3 0.3 65.9 2.4 4.4 1.1 2.2 0.2 80 3 190C Yes 27.6 0.8 0 8.9 0.4 65.0 0.9 4.6 0.8 2.2 0.1 81 1 aReaction times were 10 minutes at indicated temperatures (n The liquid phase (hemicellulose hydrolysate) was separated using a screw press bHydrolysate was adjusted to pH 5 and treated for 6 h (50C) with Biocellulase W (20 ml/L hydrolysate) to hydr olyze remaining oligomers cSolids content of bagasse slurry (fiber and nonvolatile solutes) emerging from the reactor after dilute acid steam pretreatment. dDimer was calculated using cellobiose as a standard and contains xylose and glucose oligomers.

PAGE 98

98 Table 4-4. Effect of enzyme treatment on sele cted inhibitors in hemicellulose hydrolysate. Reaction conditiona Enzyme treatedb % Dry weightc Inhibitors (g/L) 5-HMF Furfural Formic ac id Acetic acid Total 160C No 30.9 3.0 0.07 0.01 0.31 0.05 0.40 0.04 2.2 0.1 3.0 0.2 170C No 31.0 1.2 0.10 0.01 0.78 0.16 0.53 0.06 3.1 0.1 4.5 0.2 180C No 27.4 2.1 0.15 0.02 1.57 0.29 0.78 0.05 3.8 0.3 6.3 0.4 190C No 27.6 0.8 0.30 0.04 3.48 0.27 1.05 0.10 5.3 0.3 10.2 0.4 160C Yes 30.9 3.0 0.09 0.03 0.32 0. 10 0.24 0.07 4.9 0.2 5.6 0.2 170C Yes 31.0 1.2 0.11 0.03 0.74 0. 14 0.51 0.12 5.8 0.3 7.2 0.4 180C Yes 27.4 2.1 0.15 0.05 1.59 0. 31 0.66 0.18 5.9 0.3 8.3 0.5 190C Yes 27.6 0.8 0.28 0.02 3.41 0. 22 1.09 0.19 6.7 0.1 11.5 0.3 a All pretreatment reaction times were 10 mi nutes at the indicated temperature (n 3). The liquid phase (hemicellulose hydrolysate ) was separated using a screw press. b Hydrolysate was adjusted to pH 5 and treated for 6 h (50C) with Biocellulase W (20 ml/L hydrolysate; 1000 FPU/L hydrolysate). c Solids content of bagasse slurry (fiber and non-volatile solutes) emerging from the reactor after dilute acid steam pretreatment.

PAGE 99

99 Table 4 5 Maximum ethanol concentrations and yields using L+SScF process. Conditions Ethanol (g/L) (n) Ethanol yield Steam pretreatment Fermentation (g/ metric ton ) (gal/U S ton) 160C, 10 min 10% solids 19.1 0.4 (3) 191.0 58.1 160C, 10 min 12% solids 20.9 0.1 (2) 174.2 53.0 160C, 10 min 14% solids 23.6 1.3 (4) 168.6 51.3 170C, 10 min 14% solids 25.7 0.7 (4) 183.6 55.9 180C, 10 min 14% solids 29.0 1.5 (4) 207.1 63.1 Note: n = number of replicates

PAGE 100

100 Table 4 6 Comparison of ethanol yields from SScF processes. Feedstock Pretreatment Biocatalyst Ethanol titer (g/L) Ethanol yield (g/g untreated feedstock) Reference Paper sludge No additional treatments S. cerevisiae RWB222 45 0.26 Calculateda Zhang et al. 2009 Sugarcane bagasse NaOH peracetic acid, wash Z. mobilis CP4/ pZB5 24 Not availableb Teixeira et al. 2000 Hybrid poplar NaOH, peracetic acid, wash Z. mobilis CP4/pZB5 22 Not available Teixeira et al. 2000 Hybrid poplar Aqueous ammonia, wash E. coli KO11 16 0.24 Calculated Gupta and Lee, 2009 Barley hull Aqueous ammonia, wash E. coli KO11 25 0.18 Calculated Kim et al. 2008 Corn stover Aqueous ammonia, wash E. coli KO11 19 0.18 Calculated Kim and Lee, 2007 Switchgrass AFEX S. cerevisiae NRRL D 5 A 17 0.20 Alizadeh et al. 2005 Forage sorghum AFEX S. cerevisiae 424A(LNH ST) 31 0.17 Li et al. 2010 Sweet sorghum bagasse AFEX S. cerevisiae 424A(LNH ST) 42 0.16 Li et al. 2010 Rice straw AFEX S. cerevisiae 424A(LNH ST) 37 0.19 Zhong et al. 2009 Corn stover AFEX S. cerevisiae 424A(LNH ST) 40 0.20 Lau and Dale, 2009 Poplar Dilute sulfuric LE OLc Z. mobilis pZB4L 34 0.23 Calculated McMillan et al. 1999 Corn stover Dilute sulfuric S. cerevisiae 424A(LNH ST) 42 0.21 Calculated Toon et al. 1997 Corn stover SO 2 impregnation S. cerevisiae TMB3400 37 0.21 Calculated Ohgren et al. 2006 Sugarcane bagasse Dilute phosphoric E. coli MM160 (KO11 derivative) 29 0.21 This study a Results presented were used to calculate yields on an original biomass basis bData presented was insufficient to calculate yields on an original biomass basis c After sulfuric acid pretreatment, hemicellulose sugars and fiber were separated. Toxic side products in hemicellulose hydrolysate were removed by a combination of liquid e xtrac tion (LE) and overliming (OL) Both fiber and purified hydrolysate were combined during fermentation (SScF).

PAGE 101

101 Figure 41. Comparison of ethanol production from lignocellulose and corn. A) Sulfuric acid hydrolysis of hemicellulose (with washing and toxin mitigation) and enzymatic hydrolysis of cellulose Separate fermentations of cellulose (SSF) and hemicellulose (HF) B ) Simplified process using phosphoric acid hydrolysis of hemicellulose and enzymatic hydrolysis of cellulose termed l iquefaction prior to co -fermentat ion in a single vessel (L+SScF). C ) Enzymatic liquefaction of hydrated corn prior to simultaneous saccharification and fermentation (L+SSF). Conversion of Biomass to Fuel Ethanol & Chemicals Corn Steam Cooker Fermentation + amylase & glucoamylase Purification C Mature Corn to Ethanol Industry Lignocellulose Washing Dilute Acid Hydrolysis Liquid/solid Separation Hemicellulose Fermentation Cellulose +Lignin Purification Hemicellulose Syrup Hemicellulose Cleanup ASulfuric Lignocellulose Process Lignocellulose Dilute Acid Hydrolysis Purification Fermentation cellulase & hemicellulase BModified L+SScF Process with Phosphoric acid Fermentation Cellulose+ Cellulase Liquefaction (+ amylase & glucoamylase Liquefaction (+cellulase) (Zirconium Reactor) (Stainless Steel Reactor) (Stainless Steel Cooker)Separate hemicellose +SSF process for cellulose

PAGE 102

102 Figure 4-2. Selected examples of improvements in fermenta tion during serial transfers of MM105 in hydrolysate medium. A resistant clone, MM160, was isolated from transfer number 139 in medium c ontaining phosphoric acid hydrolysate: A) hemicellulose hydrolysate prepared at 150 C (transfers 1, 4, and 9 in broth containing 90% hydrolysate) and B) hemicellulose hydrolysate prepared at 160 C (transfers 33, 72, and 84 in broth containing 60% hydrolysate). The parent organism, strain LY180, was in capable of growth and fermentation under either condition. 0 1 2 3 4 0 5 10 15 20 2590% 150 C T9 90% 150 C T4 90% 150 C T1 Time (days)Ethanol (g/L) 0 1 2 3 4 0 5 10 15 20 2560% 160 C T84 60% 160 C T72 60% 160 C T33 60% 160 C LY180 Time (days)Ethanol (g/L)A B

PAGE 103

103 Figure 43. Resistance to selected inhibitor s : A) furfural, B) 5 hydroxymethylfurfural, C ) formate, D ) acetate. Legend for all (n = 4) : LY180, dashed line; MM160, solid line. 0.0 0.5 1.0 1.5 2.0 0 1 2 3 4LY180 MM160 Furfural (g/L)OD550nm 0 1 2 3 4 0 1 2 3 4LY180 MM160 5-HMF (g/L)OD550nm 0 4 8 12 16 0 1 2 3 4LY180 MM160 Formate (g/L)OD550nm 0 5 10 15 20 0 1 2 3 4LY180 MM160 Acetate (g/L)OD550nm A C B D

PAGE 104

104 Figure 44. Composition of fermentation broth at the time of inoculation (after liquefaction for 6 h with enzymes ): A) sugars and B) selected inhibitors. A B

PAGE 105

105 Figure 4-5. Fermentation of bagasse pr etreated with phosphoric acid at 160 C (L+SScF). Graphs include results for 6 h prior to inoculation (t = -6 h). Fermentors were inoculated at t = 0 h (after 6 h of liquefaction). A) Ethanol production using different concentra tions of acid-pretreated bagasse. Concentrations of sugars during fermentations with B) 10% solids, C) 12% solids and D) 14% solids bagasse. 0 48 96 144 192 240 0 5 10 15 20 25 10% Solids 12% Solids 14% Solids Time (h)Ethanol (g/L) 0 48 96 144 192 240 0 5 10 15 20 25 Sugar Dimers Glucose Xylose Galactose Arabinose Time (h)Concentration (g/L) 0 48 96 144 192 240 0 5 10 15 20 25 Sugar Dimers Glucose Xylose Galactose Arabinose Time (h)Concentration (g/L) 0 48 96 144 192 240 0 5 10 15 20 25 Sugar Dimers Glucose Xylose Galactose Arabinose Time (h)Concentration (g/L)AB C D

PAGE 106

106 Figure 4-6. Fermentations of bagasse (14% w/v) pretreated wit h phosphoric acid at various temperatures: A) ethanol, B) sugars during fermentation of bagasse pretreated at 160 C, C) sugars during fermentation of bagasse pretreated at 170 C, and D) sugars during fermentation of bagasse pretreated at 180 C. Note that bagasse pretreated at 190 C could not be fermented due to toxicity. 0 48 96 144 192 240 0 5 10 15 20 25 30 35 14% Solids (180 C) 14% Solids (170 C) 14% Solids (160 C) 14% Solids (190 C) Time (h)Ethanol (g/L) 0 48 96 144 192 240 0 5 10 15 20 25 30 35 Sugar Dimers Glucose Xylose Galactose Arabinose Time (h)Concentration (g/L) 0 48 96 144 192 240 0 5 10 15 20 25 30 35 Sugar Dimers Glucose Xylose Arabinose Galactose Time (h)Concentration (g/L) 0 48 96 144 192 240 0 5 10 15 20 25 30 35 Sugar Dimers Glucose Xylose Galactose Arabinose Time (h)Concentration (g/L)D C B A

PAGE 107

107 CHAPTER 5 GENERAL CONCLUSIONS AND FUTURE DIRECTIONS General Accomplishments Lignocellulosic biomass (LCB) is naturally resistant to chemical and biological degradation. Mineral acids are most commonly used to pretreat LCB in order to release hemicellulose sugars as well as provide a hydrolysable substrate for cellulolytic enzymes. During the dilute acid steam pretreatment of LCB side products are generated that are inhibitory to microbial growth. Therefore, the resulting liquid after pretreatment is typically separated from the lignin-cellulose rich solid and can be treated by a variety of methods to remove the soluble inhibitory compounds. Phosphoric was shown to produce l ow levels of inhibitors during dilute acid pretreatment of rye grass straw, corn stover, sorghum straw, and sugarcane bagasse (Gamez et al. 2004, 2006; Um et al. 2003; Vazquez et al. 2007) Yeast s were able to ferment t he hemicellulose hydrolysate produced after hydrolyzing rye grass straw with phosphoric acid and neutralizing with ammonium hydroxide (Israilides et al. 1978) The presence of l ower level of inhibitors in such a hydrolysate compared to sulfuric acid treatment, offered the opportunity to reduce the need for exotic metallurgy and to simplify processing of LCB to ethanol Previous studies had shown that phosphoric acid could be used to hydrolyze sugarcane bagasse and these studies were extended to include higher solids loadings, higher temperatures, and lower acid usage during dilute acid pretreatment. Steam treatment of sugarcane bag a sse at low levels of phosphoric acid (1% of bagasse dry weight ) and high temperatures (16 0 -190C) was shown to be an effective method to hydrolyze hemicellulose sugars. The generation of inhibitory sideproducts during such a pretreatment was less than with sulfuric acid and the se hemicellulose hydrolysates were

PAGE 108

108 ferm ented to ethanol by the ethanologen, E. coli LY160, without the removal of these compounds. However, the hydrolysate produced at 190C could not be fermented. Low levels of enzymes were able to hydrolyze residual glucan and xylan present in the solid porti on after pretreatment. Up to 45% of the residual glucan was solubilized at an enzyme loading of 0.5 FPU/g of steam -treated fiber (190C pretreatment) The simultaneous saccharification and co -fermentation (SScF) process has been shown to be more economical than the separate hydrolysis and fermentation process. However, SScF requires high solids loadings in order to achieve adequate ethanol concentrations that will make the process economically feasible. Low levels of cellulase were shown to be effective at improving flow properties of acid pretreated sugarcane bagasse slurries by sufficiently reducing their relative viscosity. The ability to pump slurries of acid pretreated bagasse was correlated to enzyme-solubilized sugar, relative viscosity, and flow through funnels of varying diameters. The solubilization of a small amount of sugar (3.5% of the dry weight of bagasse) was able to dramatically reduce viscosity. Viscosity was modeled as a function of cellulase dosage and incubation time. Enzyme -solubilized s ugar was modeled as a function of enzyme dose and incubation time. A simplified fermentation process was developed that combined dilute phosphoric acid pretreatment and the use of a glucose and xylose co -fermenting hydrolysateresistant E. coli strain MM160. This process eliminated the need for solid liquid separations and was used to ferment undetoxified phosphoric acid treated bagasse slurries into ethanol. An initial liquefaction step (6 h) was used prior to inoculation to sufficiently reduce viscosity and improve mixing of the acid pretreated bagasse slurry.

PAGE 109

109 Such a fermentation produced a maximum yield of 0.21 g ethanol/g bagasse dry weight corresponding to 29 g/L in the fermentation broth when using bagasse (14% solids) pretreated at 180C Future Directions Process simplification ha s been made to convert lignocellulosic biomass to ethanol H owever, further improvements are needed to make the process commercially viable. The economics of the process is largely dependent on the efficient depo lymerization of lignocellulosic biomass and the subsequent yield in converting the resulting sugars into ethanol. The ethanol produced must be at high yield and titer. The ethanol titer needs to be increased to above 40 g/L. The titer could potentially be increased by increasing the solids loading of the fermentation. However, this would also increase the concentration of inhibitory compounds. Further metabolic evolution through continuous culturing in the presence of hydrolysate is needed. In addition, bis ulfite, a n antioxidant that has been shown to be capable of improving hydrolysate fermentability (Leonard and Hajny, 1945) can be added to the pH adjusted (pH 6.5) slurry at elevated temperatures prior to fermentation. Increasing solids loading would also make it more difficult to pump and mix the acid pretreated sugarcane bagasse slurries. Improving enzymatic saccharification by sonication or particle size reduction of acid pretreated sugarcane bagasse may improve overall sugar and ethanol yields as well as the flow properties of high solids slurries. Glucose and xylose were both fermented by strain MM16 0 when inhibitory compounds were at the lowest concentration (160 C pretreatment temp., 10% solids loading). Xylose utilization decreased as pretreatment temperature and solids loading

PAGE 110

110 increased most likely due to increasing concent rations of inhibitory compounds in the hydrolysate. Xylose transport in E. coli under the conditions tested is through the highaffinity (for xylose) ABC transporter XylFGH (Hasona et al., 2004). This transporter hydrolyzes ATP to transport xylose. The us e of ATP for xylose transportation coupled with phosphorylation of xylose during metabolism (Hasona et al., 2004) may be depleting ATP reserves in MM160. Overexpressing the low-affinity (for xylose) proton symport transporters, XylE or AraE, of E. coli after the deletion of xylFGH could prevent ATP hydrolysis during xylose transportation. Catabolite repression due to the presence of glucose could be alleviated by the deletion of ptsG (gene coding for glucose specific permease). Overexpression of the galactose permease, GalP, capable of transporting glucose without either ATP hydrolysi s or the phosphoenolpyruvate:glucose phosphotransferase system, would allow the transportation of glucose in a ptsG strain.

PAGE 111

111 LIST OF REFERENCES Aden, A., Ruth, M., Ibsen, K., Jechura, J., Neeves, K., Sheehan, J., Wallace, B., 2002. Lignocellulosic biomass to ethanol process design and economics utilizing cocurrent dilute acid prehydrolysis and enzymatic hydrolysis for corn stover (NREL/TP -51032438). National Renewable Energy Laboratory, Golden, Colorado. Ali zadeh, H., Teymouri, F., Gilbert, T.I., Dale, B .E., 2005. Pretreat ment of switchgrass by ammonia fiber explosion (AFEX). Appl. Biochem. Biotechnol. 121, 1133 1141. Almeida, J.R.M., Modig, T., Petersson, A., Hahn -Hagerdal, B., Liden, G., Gorwa G rauslund, M.F., 2007. Increased tolerance and conversion of inhibitors in lignocellulosic hydrolysates by Saccharomyces cerevisiae J Chem Technol Biotechnol. 82, 340349. Bhandari, N., MacDonald, D.G., Bakhshi, N.N., 1984. Kin etic studies of corn stover saccharification using sulphuric acid. Biotechnol Bioeng. 26, 320 -327. Bommarius, A.S., Katona, A., Cheben, S.E., Patel, A.S., Ragauskas, A.J., Knudson, K., Pu, Y., 2008. Cellulase kinetics as a function of cellulose pretreatment. Metab Eng. 10, 370-381. Boucher, R.M.G., 1975. On biocidal mechanisms in the aldehyde series. Can. J. Pharm. Sci. 10, 1-7. Bungay, H.R. 2004. Confessions of a bioenergy advocate. Trends Biotechnol. 22(2), 67 71. Carvalho, G.B.M., Mussatto, S.I., Candido, E.J., Almeida e Silva, J.B., 2006. Comparison of different procedures for the detoxification of eucalyptus hemicellulosic hydrolysate for use in fermentative processes. J. Chem. Technol. Biotechnol. 81, 152 157. Chandei, A.K., 2007. Detoxification of sugarcane bagasse hydrolysate improves ethanol production by Candida shehatae NCIM 3501. Bioresour. Technol. 98, 19471950. Chandra, R.P., Bura, R., Mabee, W.E., Berlin, A., Pan, X., Saddler, J.N., 2007. Substrate pretreatment: the key to effective enzymatic hydrolysis of l ignocellulosics? Adv. Biochem. Eng /Biotechnol. 108, 6793. Chum, H.L., Johnson, D.K., Black, S.K., Overend, R.P., 1990. Pretreatment -catalyst effects and the combined severity parameter. Appl Biochem Biotechnol. 2425, 114. Clasen, C., Kulicke, W.M., 2001. Determination of viscoelastic and rheooptical material functions of water -soluble cellulose derivatives. Prog. Polym. Sci. 26, 18391919.

PAGE 112

112 Dasari, R.K., Berson, R.E., 2007. The effect of particle size on hydrolysis reaction rates and rheological properties in cellulosic slurries. Appl. Biochem. Biotechnol. 136 140, 289 -299. Dasari, R.K., Dunaway, K., Berson, E.R., 2009. A scraped surface bioreactor for enzymatic saccharification of pretreated corn stover slurries. Energy Fuels 23, 492 -497. Datsenko, K. A., Wanner, B.L., 2000. One-step inact ivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc. Natl. Acad. Sci. USA 97, 6 640 6645. Demirbas, A., 2005. Bioethanol from cellulosic materials: a renewable motor fuel from biomass. Energy Sources 27, 327337. Dien, B.S., Ximenes, E.A., OBryan, P.J., Moniruzzaman, M., Li, X., Balan, V., Dale, B., Cotta, M.A., 2008. Enzyme characterization for hydrolysis of AFEX and liquid hot wat er pretreated distillers grains and their conversion to ethanol. Biores. Technol. 99, 5216 -5225. Fogler, H.S., 1986. Distributions of residence times for chemical reactors, in: Elements of chemical reaction engineering. Prentice Hall, New Jersey, pp.629-679. Frazer, F.R., McCaskey, T.A., 1989. Wood hydrolyzate treatments for improved fermentation of wood sugars to 2,3butanediol. Biomass 18, 3142. Galbe, M., Zacchi, G., 2007. Pretreatment of lignocellulosic materials for efficient bioethanol production. Adv Biochem Eng /Biotechnol 108, 4165. Gamez, S., Ramirez, J.A., Garrote, G., Vazquez, M., 2004. Manufacture of fermentable sugar solutions from sugar cane bagasse hydrolyzed with phosphoric acid at a tmospheric pressure. J Agric Food Chem 52, 4172-4177. Gamez, S., Gonzales -Cabriales, J.J., Ramirez, J.A., Garrote, G., Vazquez, M., 2006. Study of the hydrolysis of sugar cane bagasse using phosphoric acid. J Food Eng 74, 78 -88. Gautier, S., Lecourtler, J., 1991. Structural characterization by 13C nuclear magnetic resonance of hydrolysed carboxymethylcellulose. Polym. Bulletin 26, 457 -464. Geankoplis, C.J., 1993. Transport processes and unit operations. Prentice Hall, New Jersey.

PAGE 113

113 Geddes, C.C., Peterson, J.J., Roslander, C., Zacchi, G., Mullinnix, M.T., Shanmugam, K.T., Ingram, L.O., 2010. Optimizing the saccharification of sugar cane bagasse using dilute phosphoric acid followed by fungal cellulases. Bioresour. Technol. 101, 1851 1857. Grohmann, K., Torget, R., Himmel, M., 1985. Optimization of dilute acid pretreatment of biomass. Biotechnol Bioeng 15, 59 -80. Grohmann, K., Torget, R., Himmel, M., 1986. Dilute acid pretreatment of biomass at high solid s concentrations. Biotechnol. Bioeng. Symp 17, 135151. Gupta, R., Lee, Y.Y., 2009. Pretreatment of hybrid poplar by aqueous ammonia. Biotechnol. Prog. 25, 357 364. Hahn -Hagerdal, B., Galbe, M., Gorwa-Grauslund, M.F., Liden, G., Zacchi, G., 2006. Bio ethanol -the fuel of tomorrow from the residues of today. Trends Biotechnol. 24, 549 -556. Hahn -Hagerdal, B., Karhumaa, K., Fonseca, C., Spencer Martins, I., Gorwa -Grauslund, M.F., 2007. Towards industrial pentose-ferme nting yeast strains. Appl. Microbiol. Biotechnol. 74, 937 953. Hasona, A., Kim, Y., Healy, F.G., Ingram, L.O., Shanmugam, K.T., 2004. Pyruvate formate lyase and acetate kinase are essential for anaerobic growth of Escherichia coli on xylose. J. Bacteriol. 186(22), 75937600. Heer, D., Sauer, U., 2008. Identification of furfural as a key toxin in lignocellulosic hydrolysates and evolution of a tolerant yeast strain. Microb. Biotechnol. 1 497 506. Heer, D., Heine, D ., Sauer, U., 2009. Resistance of Saccharomyces cerevisiae to high furfural concentration is based on NADPH -dependent reduction by at least two oxidoreductases. Appl. Environ. Microbiol. 75, 76317638. Holtzapple, M.T., Jun, J.H., Ashok, G ., Patibandla, S.L., Dale, B.E., 1991. The ammonia freeze explosion (AFEX) process. Appl. Biochem. Biotechnol. 28 -29(1), 59 -74. Ingledew, W.M., Kelsall, D.R., Austin, G.D., Kluhspies, C., 2009.The Alcohol Textbook: A Reference for the Beverage, Fuel and Industrial Alcohol industries, fifth ed. N ottingham University Press, Nottingham, UK. Ingram, L.O., Lai, X., Moniruzzaman, M., Wood, B.E., York, S.W., 1997. Fuel ethanol production from lignocellulose using engineered bacteria. In: Saha, B.H., Woodward, J. (Eds.), Fuel and chemicals from biomass. American Chemical Society, Washington, D. C., pp. 57 73.

PAGE 114

114 Ingram, L.O., Aldrich, H.C., Borges, A.C.C., Causey, T.B., Martine z, A., Morales, F., Saleh, A., Underwood, S.A., Yomano, L.P., York, S.W., Zaldivar, J., Zhou, S., 1999. Enteric bacterial catalysts for fuel ethanol production. Biotechnol. Prog. 15, 855866. Israilides, C.J., Grant, A., Han, Y.W., 1978. Sugar level, fermentability, and acceptability of straw treated with different acids. Appl. Environ Microbiol 36, 4346. Jantama, K., Zhang, X., Moore, J.C., Shanmugam, K.T., Svoronos, S.A., Ingram, L.O., 2008. Eliminating side pr oducts and increasing succinate yields in engineered strains of Escherichia coli C. Biotechnol. Bioeng. 101(5), 881-893. Jarboe, L.R., Grabar, T.B., Yomano, L.P., Shanmugam, K.T., Ingram, L.O., 2007. Development o fethanologenic bacteria. Adv. Biochem. Engin./Biotechnol. 108, 237 -261. Jeffries, T.W., 1994. Biodegradation of lignin and hemicelluloses. In: Ratledge, C. (Ed.), Biochemistry of microbial degradation. Kluwer Academic Publishers, Dordrecht, Netherlands, pp. 233277. Jonsson, L.J., Palmqvist, E., Nilvebrant, N.O., 1998. Detoxification of wood hydrolysates with laccase and peroxidase from the white -rot fungus Trametes versicolor Appl. Microbiol. Biotechnol. 49, 691697. Kabel, A.K., Bos G., Zeevalking, J., Voragen, A.G.J., Schols, H.A., 2007. Effect of pretreatment severity on xylan solubility and enzymatic breakdown of the remaining cellulose from wheat straw. Biores. Technol. 98, 2034 2042. Kadam, K.L., 2000. Enviro nmental life cycle implications of using bagasse derived ethanol as a gasoline oxygenate in Mumbai (Bombay) (NREL/TP -580-28705). National Renewable Energy Laboratory, Golden, Colorado. Kim, T.H., Lee, Y.Y., 2007. Pretreatment of corn stover by soaking in aqueous ammonia at moderate temperatures. Appl. Biochem. Biotechnol. 137, 81-92. Kim, T.H., Taylor, F., Hicks, K.B., 2008. Bioethanol production from barley hull using SAA (soaking in aqueous ammonia) pretreatment. Biores. Technol. 99, 56945702. Klinke, H.B., Thomsen, A.B., Ahring, B.K., 2004. Inhibition of ethanol producing yeast and bacteria by degradation products produced during pretreatment of biomass. Appl. Microbiol. Biotechnol. 66, 10-26. Laadan, B ., Almeida, J.R.M., Radstrom, P., Hahn -Hagerdal, B., Gorwa-Grauslund, M., 2008. Identification of an NADH -dependent 5hydroxymethylfurfural -reducing alcohol dehydrogenase in Saccharomyces cerevisiae Yeast 25 191 198.

PAGE 115

115 Larsson, S., Reimann, A., Nilvebrant, N.O., Jonsson, L.J., 1999. Comparison of different methods for the detoxification of lignocellulose hydrolysates of spruce. Appl. Biochem Biotechnol 7779, 91 103. Larsson, S., Cassland, P., Jonsson, L.J., 2001. Development of a Saccharomyces cerevisiae strain with enhanced resistance to phenolic fermentation inhibitors in lignocellulose hydrolysates by heterologous expression of laccase. Appl. Environ. Microbiol. 67(3), 11631170. Lau, M.W., Dale, B.E., 2009. Cellulosic ethanol production from AFEX treated corn stover using Saccharomyces cerevisiae 424A (LNH ST). Proc. Natl. Acad. Sci. USA 106, 13681373. Lau, M.W., Gunawan, C., Dale, B.E., 2009. The impacts of pretreatment on the fermentability of pretreated lignocellulosic biomass: A comparative evaluation between ammonia fiber expansion and dilute acid pretreatment. Biotechnol. Biofuels 2, 30. Leonard, R.H., Hajny, G.J. 1945. Fermentation of wood sugars to ethyl alcohol. Ind. Eng. Chem. 37(4), 390395. Li, B.Z., Balan, V., Yuan, Y.J., Dale, B.E., 2010. Process optimization to convert forage and sweet sorghum bagasse to ethanol based on ammonia fiber expansion (AFEX) pretreatment. Biores. Technol. 101, 1285 -1292. Liu, Z.L., Ma, M., Song, M., 2009. Evolutionarily engineered ethanologenic yeast detoxifies lignocellulosic biomass co nversion inhibitors by reprogrammed pathways. Mol. Genet. Gen. 282, 233-244. Lynd, L.R., van Zyl, W.H., McBride, J.E., Laser, M., 2005. Consolidated bioprocessing of cellulosic biomass: an update. Curr. Opin. Biotechnol. 16, 577 -583. Mar tinez, A., Rodriguez, M.E., York, S.W., Preston, J.F., Ingram, L.O., 2000a. Effects of Ca(OH)2 treatments ("overliming") on the composition and toxicity of bagasse hemicellulose hydrolysates Biotechnol Bioeng 69, 526536. Martinez, A., Rodriguez, M.E., York, S.W., Preston, J.F., Ingram, L.O., 2000b. Use of UV absorbance to monitor furans in dilute acid hydrolysates of biomass. Biotechnol. Prog 16, 637641. Martinez, A., Rodriguez, M.E., Wells, M.L., York, S.W., Preston, J.F., Ingram, L.O., 2001. Detoxification of dilute acid hydrolysates of lignocellulose with lime. Biotechnol. Prog 17, 287293.

PAGE 116

116 Martinez, A., Grabar, T.B., Shanmugam, K.T., Yomano, L.P., York, S.W., Ingram, L.O., 2007. Low salt medium for lactate and ethanol production by recombinant Escherichia coli B. Biotechnol Lett 29, 397-404. McMillan, J.D., Newman, M.M., Templeton, D.W., Mohagheghi, A., 1999. Simultaneous saccharification and cofermentation of dilute acid pretreated yellow poplar hardwood to ethanol using xylose-fermenting Zymomonas mobilis Appl. Biochem. Biotechnol. 77 79, 649665. Meade, G.P., Chen, J.C., 1977. Cane sugar handbook. John Wiley and Sons, New York, New York. Miller, E.N., Jarboe, L.R., Yomano, L.P., York, S.W., Shanmugam, K.T., Ingram, L.O., 2009a. Silencing of NADPH dependent oxidoreductase genes ( yqhD and dkgA ) in furfural -resistant ethanologenic Escherichia coli Appl. Environ. Mi crobiol. 75(13), 4315 4323. Miller, E.N., Jarboe, L.R., Turner, P.C., Pharkya, P., Yomano, L.P., York, S.W., Nunn, D., Shanmugam, K.T., Ingram, L.O., 2009b. Furfural inhibits growth by limiting sulfur assimilation in ethanologenic Escherichia coli strain LY180. Appl. Environ. Microbiol. 75(19), 6132-6141. Mills, T.Y., Sandoval, N.R., Gill, R.T., 2009. Cellulosic hydrolysate toxicity and tolerance mechanisms in Escherichia coli Biotechnol. Biofuels 2,26. Mitchell, D.J., Grohmann, K., Himmel, M.E., Dale, B.E., Shroeder, H.A., 1990. Effect of the degree of acetylation on the enzymatic digestion of acetylated xylans. J Wood Chem Technol 10, 111 -121. Mosier, N., Wyman, C., Dale, B., Elander R., Lee, Y.Y., Holtzapple, M., Ladisch, M., 2005a Features of promising technologies for pretreatment of lignocellulosic biomass. Bioresour. Technol. 96, 673686. Mosier, N., Hendrickson, R., Ho, N., Sedlak, M., Ladisch, M.R., 2005 b O ptimization of pH controlled liquid hot water pretreatment of corn stover. Bioresour. Technol. 96, 1986 1993. Ohgren, K., Galbe, M., Zacchi, G., 2005. Optimization of steam pretreatment of SO2impregnated corn stover for fuel ethanol production. Appl Biochem Biotechnol. 121 -124, 1055-1067. Ohgren, K., Bengtsson, O., Gorwa-Grauslund, M.F., Galbe, M., Hahn -Hagerdal, B., Zacchi, G., 2006. Simultaneous saccharification and co -fermentation of glucose and xylose in steam pretreatment corn stover at high fiber content with Saccharomyces cerevisiae TMB3400. J. Biotechnol. 126, 488 -498.

PAGE 117

117 Ohgren, K., Bura, R., Saddler, J., Zacchi, G., 2007. Effect of hemicellulose and lignin removal on enzymatic hydrolysis of steam pretreated corn stover. Bioresour. Technol. 98, 2503 -2510. Ohta, K., Beall, D.S., Mejia, J.P., Shanmugam, K.T., 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. Appl Environ Microbiol 57, 893-900. Ou, M.S., Mohammed, N., Ingram, L.O., Shanmugam, K.T., 2009. Thermophilic Bacillus coagulans requires less cellulases for simultaneous saccharification and fermentation of cellulose to products than mesophilic microbial biocatalysts. Appl. Biochem. Biotechnol. 155, 379385. Overend, R.P., Chornet, E., 1987. Fractionation of lignocellulosics by steam aqueous pretreatments. Philos. Trans. Royal Soc. Lond. A 321, 523536. Palmqvist, E., Hahn -Hagerdal, B., Galbe, M., Larsson, M., Stenberg, K., Szengyel, Z., Tengborg, C., Zacchi, G., 1996. Design and operation of a bench -scale process development unit for the production of ethanol from lignocellulosics. Bioresour Technol 58, 171179. Palmqvist, E., Hahn -Hagerdal, B., Szengyel, Z., Zacchi, G., Reczey, K., 1997. Simultan eous detoxification and enzyme production of hemicellulose hydrolysates obtained after steam pretreatment. Enzyme Microb. Technol. 20, 286293. Palmqvist, E., Hahn -Hagerdal, B., 2000a. Fermentation of lignocellulosics hydrolysates. I: In hibition a nd detoxification. Bioresour. Technol. 74, 1724. Palmqvist, E., Hahn -Hagerdal, B., 2000b. Fermentation of lignocellulosics hydrolysates. II: Inhibitors and mechanisms of inhibition. Bioresour. Technol. 74, 2533. Perlack, R.D., Wright, L.L., Turhollow, A.F., Graham, R.L., Stokes, B.J., Erbach, D.C., 2005. Biomass as feedstock for a bioenergy and bioproducts industry: The technical feasibility of a billion -ton annual supply. ORNL/TM 2005/66. Oak Ridg e National Laboratory, Oak Ridge, TN. Piper, P., Ortiz Calderon, C., Hatzixanthis, K., Mollapour, M., 2001. Weak acid adaptation: the stress response that confers yeasts with resistance to organic acid food preservatives. Microbiol. 147, 2635-2642. Rezania, S., Zhuoliang, Y., Berson, R.E., 2009. Enzymatic saccharification and viscosity of sawdust slurries following ultrasonic particle size reduction. Appl. Biochem. Biotechnol. 153, 103115.

PAGE 118

118 Sambrook, J, Russel, D .W., 2001. Molecular cloning: A laboratory manual, third ed. Cold Spring Harbor Laboratory Press, New York. Sassner, P., Galbe, M., Zacchi, G., 2005. Steam pretreatment of Salix with and without SO2 impregnation for production of bioethanol. Appl Biochem Biotechnol 121124, 1101 1117. Sassner, P., Zacchi, G., 2008. Integration options for high energy efficiency and improved economics in a woodto ethanol process. Biotechnol. Biofu els 1,4. Schell, D.J., Farmer, J., Newman, M., Mcmillan, J.D., 2003. Dilute sulfuric acid pretreatment of corn stover in pilot -scale reactor: investigation of yields, kinetics, and enzymatic digestibilities of solids. Appl Biochem Biotechnol 105108, 6986. Sluiter, A., Hames, B., Ruiz, R., Scarlata, C., Sluiter, J., Templeton, D., Crocker, D., 2008. Determination of structural carbohydrates and lignin in biomass (NREL/TP 510 -42618). National Renewable Energy Laboratory, Golden, C olorado. Teixeira, L.C., Linden, J.C., Schroder, H.A., 2000. Simultaneous saccharification and cofermentation of peracetic acidpretreated biomass. Appl. Biochem. Biotechnol. 84, 111127. Toon, S.T., Philippidis, G.P., Ho, N.W.Y., Chen, Z., Brainard, A., Lumpkin, R.E., Riley, C.J., 1997. Enhanced cofermentation of glucose and xylose by recombinant Saccharomyces yeast strains in batch and continuous operating modes. Appl. Biochem. Biotechnol. 6365, 243 -255. Um, B.H., Karim, M.N., Henk, L.L., 2003. Effect of sulfuric and phosphoric acid pretreatments on enzymatic hydrolysis of corn stover. Appl Biochem Biotechnol 105 -108, 115125. U.S. Energy Information Administration, 20 05, http://tonto.eia.doe.gov/FTPROOT/multifuel/038405.pdf U.S Department of Energy, Washington, D.C. Vazquez, M., Oliva, M., Tellez -Luis, S.J., Ramirez, J.A., 2007. Hydrolysis of sorghum straw using phosphoric acid: Evaluation of furfural production. Bioresour Technol 98, 3053 -3060. Vogel, H.C., Todaro, C.L., 1997. Fermentation and biochemical engineering handbook: principles, process design, and equipment. Noyes Publications, New Jersey. Warwick, L.M., Gray, P.P., Mandels, M., 1985. Enzymatic hydrolysis of cellulose in lignocellulosic materials. Crit. Rev. Biotechnol. 3(3), 235276.

PAGE 119

119 Wood, B.E., Ingram, L.O., 1992. Ethanol production from cellobiose, amorphous cellulose, and crystalline cellulose by recombinant Klebsiella oxytoca containing chromosomally integrated Zymomonas mobilis genes for ethanol production and plasmids expressing thermostable cellulase genes from Clostridium thermocellu m Appl. Environ. Microbiol. 58(7), 21032110. Wood, B.E., Aldrich, H.C., Ingram, L.O., 1997. Ultrasound stimulates ethanol production during the simultaneous saccharification and fermentation of mixed waste office paper. Biotechnol. Prog. 13(3), 232 -237. Wyman, C.E., 2003. Potential synergies and challenges in refining cellulosic biomass to fuels, chemicals, and power. Biotechnol. Prog. 19(2), 254-262. Wyman, C.E., Dale, B.E., Elander, R.T., Holtzapple, M., Ladis ch, M.R., Lee, Y.Y., 2005. Coordinated development of leading biomass pretreatment technologies. Bioresour. Technol. 96, 1959 -1966. Wyman, C.E., 2007. What is (and is not) vital to advancing cellulosic ethanol. Trends Biotechnol. 25(4), 153157. Xu, F., Ding, H., 2006. A new kinetic model for the heterogenous (or spatially confined) enzymatic catalysis: Contributions from the fractal and jamming (overcrowding) effects. Appl Catal A 317, 70-81. Yomano, L.P., Yor k, S.W., Zhou, S., Shanmugam, K.T., Ingram, L.O., 2008. Re engineering Escherichia coli for ethanol production. Biotechnol Lett 30, 2097 2103. Yomano, L.P., York, S.W., Shanmugam, K.T., Ingram, L.O., 2009. Deletion of methylglyoxal synthase gene ( mgsA) increased sugar co metabolism in ethanol producing Escherichia coli Biotechnol. Lett. 31, 1389-1398. Zaldivar, J., Ingram, L.O., 1999a. Effect of selected aldehydes on the growth and fermentation of ethanologeni c Escherichia coli Biotechnol. Bioeng. 65, 24-33. Zaldivar, J., Ingram, L.O., 1999b. Effect of organic acids on the growth and fermentation of ethanologenic Escherichia coli LY01. Biotechnol. Bioeng. 66, 203-210. Zaldivar, J., Martinez, A., Ingram, L.O., 2000. Effect of alcohol compounds found in hemicellulose hydrolysate on the growth and fermentation of ethanologenic Escherichia coli Biotechnol Bioeng. 68, 524 -530. Zhang, J.Y., Shao, X.J., Lynd, L.R., 2009. Simultaneous saccharification and co fermentation of paper sludge to ethanol by Saccharomyces cerevisiae RWB222. Part II: Investigation of discrepancies between predicted and observed performance at high solids concentration. Biotechnol. Bioeng. 104, 932 -938.

PAGE 120

120 Zhong, C., Lau, M.W., Balan, V., Dale, B.E., Yuan, Y.J., 2009. Optimization of enzymatic hydrolysis and ethanol fermentation from AFEX -treated rice straw. Appl. Microbiol. Biotechnol. 84, 667676.

PAGE 121

121 BIOGRAPHICAL SKETCH Claudia Geddes was born in Leon, Nicaragua in 1980. Along with her family, she moved to Miami Florida in 1987 where s he attended schools in Little Havana, Miami and later Kendall, Miami where she attended G. Holmes Braddock Senior High School. She graduated from high school in 1998. Claudia joined the Chemical Engineering Department at the University of Florida in August 1998 and graduated with a Bachelors degree (cum laude) and a minor in Chemistry in May 20 04. After graduating she decided to pursue a graduate degree and joined Dr. Lonnie Ingrams lab at the University of Floridas Department of Microbiology and Cell Science in August 2005. She is married to Ryan Geddes whom she met through the Chemical Engi neering program at UF and they now live in Gainesville, Florida with their son Alessandro.