transforming lignocelluloses to sugars and liquid fuels by li shuai a dissertation submitted in
TRANSCRIPT
Transforming Lignocelluloses to Sugars and Liquid Fuels
By
LI SHUAI
A dissertation submitted in partial fulfillment of
the requirements for the degree of
Doctor of Philosophy
(Biological Systems Engineering)
at the
UNIVERSITY OF WISCONSIN-MADISON
2012
Date of final oral examination: 06/28/2012
The dissertation is approved by the following members of the Final Oral Committee:
Xuejun Pan, Associate professor, Biological Systems Engineering
John Ralph, Professor, Biological Systems Engineering and Biochemistry
Sundaram Gunasekaran,Professor, Biological Systems Engineering
Troy Runge, Assistant Professor, Biological Systems Engineering
John Grabber, Research Agronomist, US Dairy Forage Research Center
ii
Abstract
Extensive research has been done on the development of biofuel from low-cost and
abundant lignocelluloses. Unfortunately, cost-effectively producing sugars and sugar
derivatives still remains a barrier to developing a biorefining industry. In order to
overcome this barrier, a few innovative processes were developed for converting
lignocelluloses into sugars and liquid fuels, and are presented in this thesis.
First, a sulfite pretreatment (SPORL–Sulfite Pretreatment to Overcome
Recalcitrance of Lignocelluloses) developed by our group was compared with diluted acid
pretreatment (DA) to investigate the efficacy of this new pretreatment method on
enzymatic saccharification of spruce. Results show that addition of sulfite along with
sulfuric acid could remove more lignin, retain more carbohydrates in the substrate and
reduced formation of inhibitors in spent liquor than dilute acid pretreatment due to the
reaction of sulfite with lignin and the buffer effect of sulfite. Cellulose in SPORL-
pretreated spruce was completely digested by enzymes as compared to the cellulose
conversion of 60% in DA-pretreated spruce. Additionally, the formation of hydrophilic
sulfonic groups on lignin surface was believed to decrease non-productive adsorption of
enzymes on lignin, facilitating the enzymatic hydrolysis of cellulose.
Second, considering the acid corrosion and expense of enzymes involved in
pretreatment and enzymatic saccharification, preliminary work was conducted to
synthesize reusable cellulase-mimetic solid acid with both cellulose binding and
hydrolyzing domains for cellulose hydrolysis. The binding domain (-NH2, -OH or -Cl) on
the synthesized solid acids facilitated the association of substrates onto the catalyst surface,
which increased collision chance of substrate with acid sites (-SO3H) and therefore
iii
accelerated the cellulose hydrolysis rate. Cellulose hydrolysis reactions catalyzed by the
synthesized solid acids showed much lower apparent activation energies than the ones
catalyzed by traditional liquid acids and general solid acids without binding domains.
Third, to avoid energy-intensive pretreatment and expensive cellulases involved in
traditional enzymatic saccharification of lignocelluloses, a one-step process to produce
concentrated sugar solution from lignocelluloses was developed. This process directly
converted cellulose and hemicellulose in lignocelluloses into sugars at moderate
temperature (100-160 °C) without pretreatment and enzymatic hydrolysis. Concentrated
LiCl, LiBr and CaBr2 solutions were found to have good cellulose dissolution abilities and
to be able to dissolve/hydrolyze cellulose from lignocelluloses at moderate temperatures.
Addition of small amount of acid into the concentrated salt solution accelerated the
hydrolysis of cellulose and hemicellulose. The batch-feeding of biomass allowed a high
final sugar concentration. After the saccharification, insoluble lignin was separated from
sugars and salt solution by filtration or centrifugation. Sugars and salt were separated
through a combination of organic solvent extraction of the salt and ion-exchange
chromatography. Solvent extraction separated approximately 95% of the salt from the
sugars, and the residual salt was removed by ion-exchange resins. In order to obtain a
purified sugar syrup of high concentration, air instead of water was used to push the sugar
stream through the ion-exchange columns. Ultimately, a sugar solution with a
concentration higher than 50% was recovered.
Fourth, a novel one-pot process of directly converting lignocelluloses into
hydrocarbon precursors without pretreatment and enzymatic saccharification was
developed. The reaction was conducted in a LiBr/acetone reaction system with small
iv
amount of acid and water. Because of the deficiency of water in the LiBr/acetone system,
unsolvated Li+ and Br
- were able to disrupt the hydrogen bonding in cellulose crystals,
facilitating the hydrolysis of cellulose and hemicelluloses to monosugars. The Br- also
catalyzed the dehydration of the sugars into HMF (or furfural), which immediately reacted
with acetone to form furan-based hydrocarbon precursors with 5-21 carbons in high yield
and with high selectivity. Use of acetone as solvent prevented the self-condensation of
HMF (or furfural) into byproduct humins, thereby improving the selectivity of the sugars
to the precursors. Meanwhile, lignin was extensively depolymerized and dissolved in
acetone during the process. Because of very low molecular weight, the lignin could be
hydrodeoxygenated into hydrocarbon fuels (or fuel additives) without further
depolymerization, separately or jointly with the furan-based hydrocarbon precursors
derived from cellulose and hemicellulose.
v
Acknowledgements
The work shown here was only possible under the support, guidance and
collaboration of many important people in the past five years. Firstly, I would like to thank
my advisor, Professor Xuejun Pan for his meticulous guidance on my research and his
cares and support for my life in the States. I feel very fortunate and grateful to work with
such an amiable, thoughtful and inspirational advisor.
I would like to thank our group members for their help, collaboration and
friendships. A special thank to Qiang who I worked with over the last five years. We are
the first two persons working in this group, and experienced and shared a lot together. I
would like to thank Dongsheng Zhang, Daeun Kim, Syrym Abylgaziyev, Sasikumar
Elumalai, Lis Nimani, and Chaoqun Mei for their collaboration and discussions.
In our department, I would like to thank Professor John Ralph for letting me access
NMR and GC-MS facilities and Dr. Fachuang Lu for NMR analysis. A thank-you to
Professor Sundaram Gunasekaran for letting me access his lab for the use of FT-IR and
particle size analyzer. A special thanks to Debby Sumwalt for her help and encouragement
in my life here.
My family deserve a huge thank-you. My mother and brother gave me strong
encouragement and support in my life. There is no way I would be where I am today
without them. Particularly, I want to thank my brother, Ke Shuai, who has taught me a lot
in all aspects of my life and who I always look up to as my role model. A special thank to
my fiancée, Ying Li, for her love and support every day. You are a beautiful girl and I am
blessed to have you in my life.
vi
Table of Contents
Abstract
Acknowledgements
List of Figures
List of Tables
Chapter 1: Introduction to Biofuels .............................................................................................. 1
1.1 Introduction ................................................................................................................ 1
1.2 Chemistry of Biomass ................................................................................................. 4
1.3 Technical issues in biomass conversion ................................................................... 21
1.4 Project description ................................................................................................... 30
Chapter 2: Comparative Study of SPORL and Dilute Acid Pretreatments of Spruce for
Enzymatic Saccharification .................................................................................. 32
2.1 Introduction .............................................................................................................. 32
2.2 Experimental ............................................................................................................ 38
2.2.1 Materials ..................................................................................................... 38
2.2.2 Pretreatments .............................................................................................. 38
2.2.3 Enzymatic Hydrolysis .................................................................................. 39
2.2.4 Analytical Methods ...................................................................................... 40
2.2.5 Whole Cell-Wall NMR of substrates ........................................................... 41
2.2.6 Degree of Polymerization of Cellulose ....................................................... 42
2.2.7 Fermentability of SPORL and DA Pretreatment Liquors ........................... 42
2.3 Results and Discussion ............................................................................................. 43
2.3.1 Changes in Cell-Wall Components after pretreatments .............................. 43
2.3.2 Mass balance of sugars after pretreatments ............................................... 52
2.3.3 Enzymatic digestibility of pretreated spruce ............................................... 54
2.3.4 Fermentability of spent pretreatment liquors .............................................. 58
2.4 Conclusion and recommendations ........................................................................... 61
vii
Chapter 3: Synthesis of Cellulase Mimetic Solid Acid for Cellulose Hydrolysis ..................... 63
3.1 Introduction .............................................................................................................. 63
3.2 Experimental ............................................................................................................ 69
3.2.1 Chemicals and materials ............................................................................. 69
3.2.2 Synthesis of cellulase mimetic solid acid .................................................... 69
3.2.3 Hydrolysis of biomass with solid acid ......................................................... 71
3.2.4 Adsorption of glucose and cellobiose on CP-SO3H .................................... 72
3.2.5 Determination of glucose ............................................................................ 72
3.2.6 FT-IR spectra of prepared resins ................................................................ 72
3.3 Results and discussion ............................................................................................. 73
3.3.1 Screening of binding groups ....................................................................... 73
3.3.2 Mechanism study ......................................................................................... 78
3.3.3 Hydrolysis of cellulose with CP-SO3H ........................................................ 82
3.4 Conclusion and recommendations ........................................................................... 90
Chapter 4: Saccharification of Lignocellulose in Concentrated Salt Solution ........................ 91
4.1 Introduction .............................................................................................................. 91
4.2 Experimental ............................................................................................................ 97
4.2.1 Materials and Chemicals ............................................................................ 97
4.2.2 Liquefaction of lignocellulose in concentrated LiBr solution ..................... 98
4.2.3 Hydrolysis of lignocellulose in acidic concentrated LiBr solution ............. 98
4.2.4 Extraction of LiBr by organic solvents........................................................ 99
4.2.5 Removal of salt by ion-exchange chromatography ................................... 101
4.2.6 Quantification of sugars and sugar derivatives ........................................ 102
4.2.7 Determination of LiBr amount .................................................................. 103
4.3 Result and discussion ............................................................................................. 103
4.3.1 Description of whole process .................................................................... 103
4.3.2 Liquefaction of lignocellulose in concentrated LiBr solution ................... 105
viii
4.3.3 Dissolution mechanism of cellulose in concentrated LiBr solution .......... 109
4.3.4 Hydrolysis of lignocellulose in acidic concentrated LiBr solution ........... 113
4.3.5 Hydrolysis of lignocellulose through batch feeding .................................. 117
4.3.6 Saccharification of lignocellulose in concentrated solution of different salts
................................................................................................................... 119
4.3.7 Hydrolysis of lignocellulose in concentrated LiBr solution with different
acids ......................................................................................................... 124
4.3.8 Separation of LiBr and sugars by different methods ................................. 125
4.3.9 Removal of residual LiBr from sugar stream ............................................ 138
4.4 Conclusion and recommendations ......................................................................... 147
Chapter 5: Conversion of Lignocellulose into Hydrocarbons ................................................. 150
5.1 Introduction ............................................................................................................ 150
5.2 Experimental .......................................................................................................... 156
5.2.1 Chemicals and materials ........................................................................... 156
5.2.2 Production of hydrocarbon precursors from biomass .............................. 156
5.2.3 Determination of residual LiBr ................................................................. 157
5.2.4 Determination of sugars and sugar derivatives ........................................ 157
5.2.5 Qualitative analysis of hydrocarbon precursors using GC-MS ................ 157
5.2.6 Quantitative analysis of hydrocarbon precursors using ESI-MS .............. 158
5.2.7 Estimation of lignin molecular weight ...................................................... 160
5.2.8 Characterization of LiBr/acetone solvent systems .................................... 160
5.3 Results and Discussion ........................................................................................... 160
5.3.1 Description of the HDA process ................................................................ 160
5.3.2 LiBr/water and LiBr/acetone systems ....................................................... 163
5.3.3 One-step conversion of biomass into hydrocarbon precursors ................. 166
5.3.4 Effect of temperature and time on conversion of spruce powder .............. 171
5.3.5 Effect of different salts on conversion of spruce powder .......................... 172
5.3.6 Effect of different acids on conversion of spruce powder ......................... 173
ix
5.3.7 Identification of products .......................................................................... 174
5.3.8 Quantification of conversion by ESI-MS ................................................... 181
5.3.9 Carbon number distributions of hydrocarbon precursors ........................ 185
5.3.10 Decomposition of lignin during HDA process .......................................... 187
5.3.11 Reaction mechanism of glucose to HMF ................................................... 188
5.3.12 Recycling/recovery of solvents and LiBr ................................................... 192
5.3.13 Hydrodeoxygenation of hydrocarbon precursor and lignin into
hydrocarbons ............................................................................................. 193
5.4 Conclusion and recommendations ......................................................................... 195
Reference ................................................................................................................................ 196
x
List of Figures
Figure 1.1 A structure model for plant cell wall 6
Figure 1.2 Chemical structures of (a) cellulose and (b) starch 9
Figure 1.3 Hemicelluloses characterized by a β-(1→4)-linked backbone with an equatorial
configuration at C1 and C4 12
Figure 1.4 Schematic illustration of the types of hemicelluloses found in plant cell walls 14
Figure 1.5 Chemical structure of lignin monomers 15
Figure 1.6 Formation of resonance-stabilized phenoxyl radical catalyzed by enzyme 16
Figure 1.7 Structures of main linkages in lignin 18
Figure 1.8 A proposed model structure of lignin 19
Figure 1.9 Flowchart of biorefining 21
Figure 2.1 Flowchart of experiment and analysis 37
Figure 2.2 HSQC NMR spectra of untreated and pretreated spruce cell walls 51
Figure 2.3 Mass balance of saccharides during the DA and SPORL pretreatments 54
Figure 2.4 Comparison of time-dependent enzymatic hydrolysability of
SPORL and DA pretreated spruce at different levels of enzyme loading 55
Figure 2.5 Inhibitors formation from cellulose and hemicellulose during pretreatment 58
Figure 2.6 Fermentability of SPORL and DA pretreatment spent liquors by in
vitro ruminal fermentation assay 60
Figure 3.1 A proposed model of cellulase-mimetic solid acid and cellulose interactions 68
Figure 3.2 Synthesis of CP-SO3H 70
xi
Figure 3.3 A model for cellulose hydrolysis on solid acid (CP-SO3H) surface 71
Figure 3.4 FT-IR spectra of (a) CP resin and (b) CP-SO3H resin 73
Figure 3.5 FT-IR spectra of solid acids with different binding domains 75
Figure 3.6 Cellobiose hydrolysis catalyzed by four types of solid acids as a function of time
76
Figure 3.7 Time course of adsorption curve of glucose and cellobiose onto resins in
aqueous solution 80
Figure 3.8 Comparison of cellobiose hydrolysis catalyzed by (a) CP-SO3H and (b) PS-
SO3H 81
Figure 3.9 Hydrolysis of cellobiose catalyzed by CP-SO3H and sulfuric acid 83
Figure 3.10 Cellobiose hydrolysis catalyzed by recycled CP-SO3H resin 85
Figure 3.11 Arrhenius plot for cellulose hydrolysis catalyzed by CP-SO3H 87
Figure 4.1 Process flow chart of biomass saccharification in concentrated salt solution 104
Figure 4.2 Hydrolysis way of biomass in concentrated salt solution 105
Figure 4.3 Hydrolysis of spruce powder in LiBr solution (no acid) as a function
of LiBr concentration at different temperatures 107
Figure 4.4 Models for cellulose dissolution in salt solutions of varied concentrations 111
Figure 4.5 The picture of the separated hydrolysate 119
Figure 4.6 The proposed structures of LiCl hydrates 123
Figure 4.7 Mechanism of extraction of glucose with boronic acid 126
xii
Figure 4.8 Schematic diagram for ion-exclusion chromatography for separating
sugar and salt 128
Figure 4.9 Separation of LiBr and sugar solution using ion-exclusion chromatography 129
Figure 4.10 Separation of residual LiBr from sugar solution using ion-exclusion
chromatography 130
Figure 4.11 Flowchart for separating LiBr and sugars by solvent extraction 132
Figure 4.12 The picture of extraction of LiBr from hydrolysate with butanol-hexane 135
Figure 4.13 Formed sugar syrup after butanol extraction of hydrolysate 137
Figure 4.14 Sugars precipitated from solvent 142
Figure 4.15 The pictures of columns packed with cation and anion exchange resins 145
Figure 4.16 The picture of purified concentrated sugar solution 146
Figure 5.1 Reaction pathway of biomass to hydrcarbon precursors in
HDA process 160
Figure 5.2 Process flow chart of converting lignocelluloses into hydrocarbons 162
Figure 5.3 Interaction of LiBr with different solvents 164
Figure 5.4 FT-IR spectra of LiBr/acetone solution 165
Figure 5.5 The Hydrocarbon precursors from HDA process in acetone 171
Figure 5.6 Proposed self-condensation mechanism of acetone 175
Figure 5.7 Proposed condensation mechanism of furfural with acetone 175
Figure 5.8 GC-MS spectra of hydrocarbon precursors in CH2Cl2 178
xiii
Figure 5.9 ESI-MS spectra of hydrocarbon precursors in CH2Cl2 180
Figure 5.10 Quantificaiton of the hydrocarbon precursors from lignocelluloses using 2D-
NMR method 183
Figure 5.11 Mechanism of quantifying hydrcarbon precursors by oxidation method 184
Figure 5.12 Carbon number distributions of hydrocarbon precursors from
different feedstocks 186
Figure 5.13 Gel permeation chromatograph of HDA lignin from spruce 188
Figure 5.14 Dehydration of glucose to HMF derivative at different LiBr concentrations 189
Figure 5.15 Dehydration of fructose to HMF derivative at different LiBr concentrations 189
Figure 5.16 The proposed mechanism of glucose to HMF in acetone/LiBr system 191
Figure 5.17 Images of products before and after hydrodeoxygenation 194
xiv
List of Tables
Table 1.1 Composition of common lignocelluloses on dry basis 7
Table 1.2 Content of main linkages in lignin 18
Table 2.1 Chemical analyses of original spruce, pretreated spruce substrates
and spent pretreatment liquors 46
Table 2.2 Concentrations of major fermentation inhibitors in pretreatment
spent liquors 48
Table 2.3 Viscosity of cellulose solutions from untreated- and pretreated-spruces 49
Table 3.1 The apparent activation energies of cellobiose hydrolysis catalyzed
by four types of solid acids 77
Table 3.2 Hydrolysis of starch and Avicel cellulose catalyzed by CP-SO3H
and sulfuric acid 83
Table 3.3 Apparent activation energies of cellulose hydrolysis catalyzed by different
catalysts 89
Table 4.1 Composition analyses of different feedstocks 100
Table 4.2 LiBr-hydrolysis of various types of feedstock 106
Table 4.3 Monosaccharides in the hydrolysates from LiBr hydrolysis of different
feedstocks without acid 108
Table 4.4 Monosaccharides in hydrolysate from LiBr hydrolysis of different
feedstocks after autoclaving at 120 °C for 1h 109
Table 4.5. Hydrolysis of spruce powder under various conditions 116
Table 4.6 Hydrolysis of spruce powder in batch-feed mode 118
xv
Table 4.7 Effect of different salts on hydrlysis of biomass 121
Table 4.8 Effect of different acids on hydrlysis of biomass 125
Table 4.9 Separation of glucose and LiBr by extraction with a mixture of
butanol and hexane 133
Table 4.10 Separation efficiency of LiBr and glucose by solvent extraction 133
Table 4.11 Effects of LiBr and water on the precipitation of glucose by ethanol 141
Table 5.1 Conversion of lignocelluloses or sugars into hydrocarbon precursors in
acetone/LiBr system 168
Table 5.2 Effect of temperature on conversion of spruce powder 172
Table 5.3 Effect of time on conversion of spruce powder 172
Table 5.4 Effect of different halogen salts on conversion of spruce powder 173
Table 5.5 Effect of different acids on conversion of spruce powder 174
Table 5.6 Proposed structure of products and corresponding molecular weight 176
Table 5.7 Quantitation of hydrocarbon precursors derived from carbohydrates 184
Table 5.8 Proton affinity of different functional groups in hydrocarbon precursors 186
1
Chapter 1: Introduction to Biofuels
1.1 Introduction
A biofuel is a fuel that is derived from biological materials, and can include a
series of different fuels forms, such as solid carbon, liquid fuels, and gases. Biofuels are
typically deemed to be sustainable because the carbon in biofuel is in a closed carbon
cycle where carbon dioxide generated from combustion of biofuel could be fixed and
converted into biomass again through photosynthesis.
There has been a growing interest in liquid biofuels worldwide over the past few
years, mostly due to the rising price of fossil fuels, concerns over the rising CO2 and
other greenhouse gas emissions, climate change, and the expected depletion of world oil
reserves. It is believed that producing transportation fuels from renewable biomass
resources can reduce the dependence on traditional fossil fuel, relieve the energy crisis,
create new job opportunities, improve local economies, and reduce greenhouse gas
emissions. Recent legislation, government investment, and technology advances have
greatly promoted biofuels production and development. For example, the United States
produced 9 billion gallons corn ethanol and 500 million gallons biodiesel in 2009.
2
Technologies for new generation cellulosic ethanol are under development and being
scaled-up in demonstration projects. While current research effort and investment
primarily focus on ethanol (from both starch and cellulose), biobutanol, and biodiesel,
the next generation liquid fuels, biohydrocarbons (gasoline, diesel and jet fuel from
biomass) also show promise because of their advantages over last generation liquid
biofuels (Demirbas, 2005; Demirbas, 2009; Limayem and Ricke, 2012).
Bioethanol including corn ethanol and cellulosic ethanol, is produced by the
fermentation of sugars. Commercialized bioethanol is currently produced from corn
kernels in the United States and sugarcane in Brazil. Corn kernels as grain have limited
yield and the massive consumption of corn kernels to produce bioethanol might
endanger the food security. Therefore, researchers now are focusing on obtaining sugars
from cheap and abundant lignocelluloses, such as wood, agricultural residues, and
dedicated bioenergy crops. A typical process for cellulosic ethanol production consists of
three steps: feedstock pretreatment to enhance the accessibility of cellulose to cellulases,
enzymatic saccharification of the cellulose to glucose, and fermentation of glucose to
ethanol. One of the barriers to the commercialization of cellulosic ethanol is the lack of
economical and effective technologies for the feedstock pretreatment. The significance
3
and importance of the pretreatment step cannot be overemphasized, as the effectiveness
of pretreatment affects the upstream selection of biomass, the yield of fermentable sugars,
and the chemical and morphological characteristics of the pretreated substrate that in
turn govern downstream hydrolysis. An ideal pretreatment method should be both
economical (in terms of capital and operating costs) and effective for a variety of
lignocelluloses. Specifically, it should require minimal feedstock preparation and
preprocessing (such as size reduction) prior to pretreatment, maximally recover all
lignocellulosic components in usable forms with minimal formation of fermentation
inhibitors, and produce a readily digestible cellulosic substrate that can be easily
hydrolyzed with a low loading of enzymes. In the last several decades, research and
development efforts have made significant progress in pretreatment technologies for
lignocellulosic feedstocks (Chandra et al., 2007; Hendriks and Zeeman, 2009; Lynd et al.,
2002; Mosier et al., 2005; Wyman et al., 2005). Many pretreatment technologies, such as
lime, dilute acid, hot water, ammonia, steam explosion, SPORL (Sulfite Pretreatment to
Overcome Recalcitrance of Lignocelluloses), ionic liquid and organosolv pretreatments
(Chandra et al., 2007; Gupta and Sehgal, 1979; Kim et al., 2009; Monavari et al., 2009;
Nguyen et al., 2000; Pan et al., 2005; Sendich et al., 2008; Sierra et al., 2009; Wingren et
4
al., 2008), have achieved varying levels of success.
Compared to bioethanol, next-generation liquid biofuels-hydrocarbons (or bio-
hydrocarbons) have advantages of high energy density and immiscibility with water. In
addition, limited water use for processing, short production cycle, and the elimination of
energy-intensive distillation will possibly lead to a low production cost of
biohydrocarbons. The biohydrocarbons can be produced through different processes
using a variety of technologies, including pyrolysis followed by upgrading of bio-oil,
gasification followed by Fischer-Tropsch synthesis, liquid-phase reforming of sugars or
sugar derivatives, decomposition and hydrodeoxygenation of lignin, and oligomerization
of alkenes derived from biomass. These methods typically involve multiple steps where
biomass is firstly converted into purer and simpler chemical states, such as sugars,
syngas, HMF, or levulinic acid, and then these chemicals are subsequently processed
into fuels catalytically.
1.2 Chemistry of Biomass
Lignocellulose is a scientific term for plant roots, stems and leaves, and is the
most abundant source of organic material on earth. Lignocelluloses include, for example,
5
agricultural residues such as corn stover and wheat stalks, energy crops such as poplar
trees and switchgrass, and municipal solid waste.
The cell walls of lignocelluloses are composed of three major components:
cellulose, hemicelluloses, and lignin. The distributions of three components in cell walls
are schematically depicted in Figure 1.1. A cell wall is generally divided into several
layers, including the primary wall (P), the thin outer layer of the secondary wall (S1), the
substantial middle layer (S2), and the very thin inner layer or tertiary wall (S3).
Cellulose chains and other cell wall constituents are aggregated into bundles
called microfibrils within each layer of the secondary wall (S). The microfibrillar groups
are in helixes alternately crossed in the S1 layer; are oriented in bands nearly parallel to
the cell axis in the S2 layer; and in the S3 layer, is nearly perpendicular to that in the S2
layer. The primary wall (P) has an irregular helical arrangement around the cell axis. The
fibers are surrounded by the heavily lignified middle lamellae (M) which is shared by
adjacent fibers. These fibers are made more waterproof by lignin or waxy compounds,
which offer chemical and disease resistance of resulting plants. Hemicelluloses provide
an intimate interlacing and even bonding between the lignin and cellulose (Fan et al.,
1982).
6
Figure 1.1 A structure model for plant cell wall (Fan et al., 1982).
The composition of lignocellulose depends on its source. Softwood, hardwood,
and herbaceous plants have different cellulose, hemicelluloses and lignin contents.
7
Generally, cellulose and hemicelluloses account for 35-45% and 15-25% of the dry
matter, individually. Softwood has the highest lignin content of 25-35%, followed by
content of 18~25% in hardwood and content of 10-15% in herbaceous plant. Other than
the different lignin contents, another significant difference between woody biomass and
herbaceous biomass is that herbaceous biomass contains 5-10% ash, whereas woody
biomass hardly contains any ash. The composition of some common lignocelluloses is
summarized in Table 1.1.
Table 1.1 Composition of common lignocelluloses on dry basis (Sun and Cheng, 2002)
Lignocelluloses Cellulose (%) Hemicelluloses (%) Lignin (%)
Hardwoods stems 40–55 24–40 18–25
Softwood stems 45–50 25–35 25–35
Nut shells 25–30 25–30 30–40
Corn cobs 45 35 15
Grasses 25–40 35–50 10–30
Paper 85–99 0 0–15
Wheat straw 30 50 15
Newspaper 40–55 25–40 18–30
Waste papers 60–70 10–20 5–10
Swine waste 6.0 28 NA
Solid cattle manure 1.6–4.7 1.4–3.3 2.7–5.7
Coastal Bermuda grass 25 35.7 6.4
Switchgrass 45 31.4 12.0
Lignocellulose has a complicated structure, which makes it hard to utilize cost-
effectively. More processing steps and harsher conditions are required to convert
8
lignocelluloses into usable sugars or sugar derivatives than those required for the
conversion of starch in corn kernels. The structures and chemical properties of each
component will be discussed below in detail.
(1) Cellulose
Cellulose with the formula of (C6H10O5)n is a strong and unbranched polymer of
glucose that is found in plant cell walls, as shown in Figure 1.2. Cellulose is produced by
terrestrial plant, from single-celled algae in the oceans to trees on the land. Cellulose is
the most abundant natural polymer on earth. A study completed by the USDA and the
U.S. Department of Energy indicated that at least 1 billion tons of cellulose in the form
of wheat straw, corn stover, other forages and residues, and wood wastes could be
sustainably collected and processed each year in U.S.. This resource represents an
equivalent of 67 billion gallons of ethanol and could replace 30% of the gasoline
consumption in the U.S. (Mosier, 2007).
9
(a)
(b)
Figure 1.2 Chemical structures of (a) cellulose and (b) starch.
Unlike the α(1→4) and α(1→6) glycosidic linkages in the amorphous starch
polymer, glucose units in cellulose are linked together by β(1→4) glycosidic linkages.
The nature of β(1→4) glycosidic bond allows the cellulose chain to be straight. The
regular arrangement of these straight chains together with the abundant hydroxyl groups
favors the formation of orderly hydrogen bonding among cellulose chains, and allows
cellulose chains to form fibers through several orders of organization. In this way, most
of cellulose in plants exists in the form of a tight crystalline structure. The rigid crystal
structure of cellulose imparts exceptional strength to cellulose and the plant.
10
However, the crystal structure becomes one of the barriers to the utilization of
lignocelluloses. Due to the strong intra- and inter-molecular hydrogen bonding in
cellulose crystals, cellulose is insoluble in water and also in dilute acid solution at low
temperature. The dissolution of cellulose depends on the disruption of hydrogen bonding
in crystalline cellulose. Cellulose can dissolve in concentrated acid at low temperature
where strong protons can penetrate into cellulose crystallites to break hydrogen bonding
and partially hydrolyze cellulose. In alkaline solutions, cellulose can swell and dissolve
when the polymerization degree of cellulose is lower than 200 (Krassig and Schurz,
2002). The previously mentioned cellulose dissolution is generally accompanied by
significant hydrolysis and degradation of the cellulose. In industry and laboratory,
cellulose solvents such as DMF/lithium halides complex, Cadoxen, and
cupriethylenediamine hydroxide, which work under mild condition, have been widely
investigated (Dawsey and Mccormick, 1990; Turbak et al., 1977). These solvents can
dissolve cellulose without substantially depolymerizing cellulose so that dissolved
material can retain its strength and be regenerated for fiber spinning. Salt solution such
as zinc chloride can also dissolve limited amounts of cellulose at certain temperature and
concentration, but may still lead to cellulose degradation because of the strong Lewis
11
acidity of zinc ion. In summary, when extracting sugars from lignocelluloses, severe
conditions such as high temperature and concentrated acid are needed to disrupt
cellulose crystallinity to expose single cellulose chain to chemicals or enzymes for
hydrolysis.
(2) Hemicelluloses
Hemicelluloses are made up of highly branched polymers of glucose, arabinose,
galactose, mannose, xylose and uronic acids. The backbones of hemicelluloses include
xyloglucans, xylans, mannans, and glucomannans. These types of hemicelluloses are
present in the cell walls of all terrestrial plants. The detailed structure and content of the
hemicelluloses varies widely between different species and cell types. Hemicelluloses
are a heterogeneous group of polysaccharides. The term was coined at a time when the
structures were not well understood and biosynthesis was completely unknown.
Currently, most working with biomass uses the term hemicelluloses as a convenient
denotement for a group of cell wall polysaccharides that are characterized by having β-(1
→4)-linked backbones of glucose, mannose, or xylose. These polysaccharides all have
12
an equatorial configuration at C1 and C4 and hence the backbones have significant
structural similarity, as shown in Figures 1.3 and 1.4.
Figure 1.3 Hemicelluloses characterized by a β-(1→4)-linked backbone with an equatorial
configuration at C1 and C4 (Scheller and Ulvskov, 2010).
13
Xyloglucan [β-D-Glcp-(1→4)]n backbone substituted with side chains as seen in pea and Arabidopsis.
The arrow indicates the typical β-glucanase cleavage site.
Mixed linkage β-glucan [β-D-Glcp-(1→4)]n-β-D-Glcp-(1→3)-[β-D-Glcp-(1→4)]m, where n and m are 3 or 4;
typical of Poales.
Glucuronoarabinoxylan, GAX, typical of commelinid monocots.
Glucuronoxylan, typical dicot structure.
Galactomannan, typical of Fabaceae seeds.
14
Galactoglucomannan, typical of conifer wood.
: :D-Glucose Glcp :D-Galactose Galp :D-Mannose Manp
:L-Arabinose Araf :D-Xylose Xylp :L-Fucose Fucp
:L-Rhamnose Rhap :D -Glucuronic acid GlcAp
Figure 1.4 Schematic illustration of the types of hemicelluloses found in plant cell walls (Scheller
and Ulvskov, 2010).
Note: The structure of the hemicelluloses varies greatly in different plant species and tissue types. ―Fer‖
represents acylation with ferulate acid (3-methoxy-4-hydroxycinnamate acid), which is characteristic of
xylans in commelinid monocots.
The most common types of polysaccharides that belong to the hemicelluloses are
xylan and mannan. As some minor sugars such as arabinose and galactose, as well as
acetyl groups, branch from the backbone chain, the resulting hemicelluloses are irregular
and therefore amorphous. Grasses generally contain a large amount of glucourono-
arabinoxylan and no mannan. Mannan is generally found in woody biomass, and
softwood contains higher amount of mannan than hardwood. The amorphous structure of
hemicelluloses allows them to be easily accessed by chemical reagent or enzymes.
15
Hemicelluloses can be extracted by hot water or alkali solutions, and can also be easily
hydrolyzed into monosaccharides at low temperature by acid or enzymes.
(3) Lignin
Lignin occupies the interstitial space of plant cell walls, filling in the left space to
bond cellulose and hemicelluloses together, which results in the strength of the
lignocellulosic matrix, and thus of the entire plant. Lignin has very complicated
structure and is an amorphous polymer mainly composed of three basic monomer units
(monolignols), specifically p-coumaryl alcohol, coniferyl alcohol, and sinapyl alcohol,
as shown in Figure 1.5.
Figure 1.5 Chemical structure of lignin monomers.
16
Figure 1.6 Formation of resonance-stabilized phenoxyl radicals catalyzed by enzyme.
The polymer forms through radical polymerization of the monolignols. An
example is shown in Figure 1.6 that a phenoxyl radical formed from coniferyl alcohol
and its resonance forms. It can be seen that stabilization of the radical occurs by
coupling to another radical at the position of the unpaired electrons. Coupling of two
radicals forms a dimer with linkages of β-O-4, 5-5, β-5, β-1, 4-O-5 or dibenzodioxocin,
as shown in Figure 1.7. Repeated radical couplings lead to formation of polymeric lignin,
and a model structure of lignin is shown in Figure 1.8. The typical contents of
dominating linkages in lignin are listed in Table 1.2. It can be seen that although the
content of these linkages varies with the types of wood, more than 2/3 of the linkages in
lignin are ether linkages. Hardwood lignin contains about 1.5 times more β-O-4 linkages
than softwood lignin because it contains more syringyl units which cannot couple at the
17
5-positions. In lignin chemistry, the functional groups that have significant affects on
the reactivity of lignin are the methoxyl, phenolic and aliphatic hydroxyls, and β-O-4
linkages. It has been found that lignin from softwood is derived from more than 90%
coniferyl alcohol with the remaining being mainly p-coumaryl alcohol units. Contrary to
softwoods, lignin from hardwoods is derived from varying ratios of sinapyl and
coniferyl alcohol monomers. Grasses generally contain more p-coumaryl units than
hardwoods and softwoods. In chemical pulping, lignin needs to be removed to separate
fibers. When extracting sugars from lignocelluloses using enzymes, lignin should be
partially removed by cleaving ether linkages to increase the accessibility of the cellulose
to cellulases. After the extraction of sugars, lignin can be generally burned to produce
heat for the conversion process or develop value-added products. Alternatively, through
thermochemical methods, lignin along with cellulose and hemicelluloses can be
simultaneously converted into intermediates (such as syngas and biooil), which could be
catalytically processed into hydrocarbons for fuels.
18
Table 1.2 Content of the main linkages in lignin (Achyuthan et al., 2010; Pandey and Kim, 2011;
Ralph, 2005; Zakzeski et al., 2010)
Linkage type Softwood (spruce) Hardwood (birch)
β-O-4-Aryl ether 46 60
Dibenzodioxocin 25-30 5-10
β-5-Phenylcoumaran 9-12 6
β-β-(Resinol) 2-6 3-12
4-O-5-Diaryl ether <4 <6.5
β-1-(1,2-Diarylpropane) 1-2 1-2
α-O-4-Aryl ether A few A few
Figure 1.7 Structures of main linkages in lignin (Ralph and Landucci, 2010).
19
Figure 1.8 A proposed model structure of lignin.
(4) Ash and other components
Other than the three major components listed above, other organic materials
founded in plants include terpenes, resins, and phenols, which can be extracted by water
and organic solvents such as ethanol, benzene, and acetone, thereby named extractives.
Related to terpenes are terpene alcohols and ketones. The resins include a wide variety of
non-volatile compounds, including fats, fatty acids, alcohols, resin acids, phytosterols,
and less known neutral compounds in small amounts. The phenols consist of a large
20
number of compounds, such as tannins, heartwood phenol, and related substances.
Additionally, low-molecular-weight carbohydrates, alkaloids, and soluble lignin are
extractable as well. The non-cell-wall substances, such as starch, pectin and protein are
not extractable. Generally, extractives do not have significant effects on the
bioconversion process. In gasification or pyrolysis, nitrogen and sulfur from amino acids
will form NH3 and H2S, which can deactivate catalysts in downstream catalytic
processes.
Inorganic components in biomass are usually termed ash, the name given to the
non-aqueous residual components of biomass that remain after it is burned. The
dominating components are alkali and alkali earth carbonates, and oxalates. Silica
deposited as crystals is especially abundant in straws. The ash component of biomass can
forms precipitants such as CaSO4, Ca(OH)2, and Mg(OH)2 in acid or alkali pretreatments.
Additionally, in alkaline pretreatment, silicon dioxide found in high concentration in
grassy plants, can react with alkali to form sodium silicate, which could affect
downstream operations. In gasification, alkali metals also have significant effects on the
syngas composition and might cause slagging, deposition, corrosion and fluidized bed
agglomeration (Hald, 1995; Hedjazi et al., 2009; Lv et al., 2010; Muangrat et al., 2010).
21
1.3 Technical issues in biomass conversion
Biomass from plants offers a potentially abundant source of sugars for ethanol
fermentation , but its complex laminate structure, consisting of cellulose, hemicelluloses
and lignin, often collectively termed lignocellulose, is difficult to disrupt and break down
into fermentable sugars. Pretreatment is required to separate the components, detach
lignin, and discompose the cellulose fibers for efficient conversion into fermentable
sugars. Therefore, the processing required to break down biomass into fermentable
sugars is more energy-intensive and expensive than obtaining sugar from grain starch or
pressing juice from sugar cane.
Figure 1.9 Flowchart of biorefining.
Sugarcane
Starch
Cellulose
Oil
Pretreatment and
Enzymatic hydrolysis
Gasification
Microbial fermentation
Acid hydrolysis
Pyrolysis
Catalytic fuel synthesis
-FT process
-Aldol condensation
-Oligomerization of butene
Catalytic cracking
Transesterification
Methanol
Ethanol
Butanol
……
Gasoline
Jet fuel
Diesel
Biodiesel
l
H2
CH4
DME
……
22
Primary approaches of converting biomass into biofuels are summarized in
Figure 1.9. Fermentation of sugars into ethanol is a very mature technique in the brewing
industry, therefore attracting more attention than other types of liquid biofuel production
processes. Starch, which can be easily hydrolyzed chemically or enzymatically into
sugars at a relatively low cost, was the first feedstock used to produce commercial
bioethanol, but the relatively high price of starch and its potential threat to food security
limit the use of starch as the feedstock to meet the increasing demand for renewable fuels.
Therefore, studies on bioethanol production from cheap and abundant lignocelluloses
have been increasing readily in recent years (Smith, 2008). However, cellulose in
lignocelluloses is enclosed with hemicelluloses and lignin, which makes it more
challenging to be hydrolyzed into glucose than starch; severe conditions, such as high
temperature, and high pressure, are needed to hydrolyze it or activate it (Demirbas,
2005). It is already widely studied that sugar can be produced through swelling of
crystalline cellulose in ground lignocelluloses with concentrated acid, including sulfuric
acid, hydrochloric acid, and phosphoric acid, followed by the hydrolysis of the swelled
cellulose at dilute acid concentration (Miller and Hester, 2007; Zhu et al., 2009).
23
Unfortunately, the acid hydrolysis process has issue with equipment corrosion and the
difficulty of sugar-acid separation, thereby limiting their applicability.
Currently, the two-step process, pretreatment followed by enzymatic hydrolysis,
is the most studied one for extracting sugars from lignocellulose for cellulose ethanol
production. Pretreatment aims at removing recalcitrance of lignocelluloses and
simultaneously decrystallizing and/or prehydrolyzing cellulose. A good pretreatment
method should retain as much cellulose as possible for the subsequent enzymatic
hydrolysis step while improving the enzymatic hydrolysis ability of the pretreated
materials (Piccolo and Bezzo, 2009; Tomas-Pejo et al., 2008; Xu et al., 2009). Generally,
pretreatment conditions are conducted under temperatures of 150-190 °C indicating that
pretreatments are energy-intensive processes and need high-quality facilities. High
temperatures also cause degradation of sugars into byproducts, which not only decreases
sugar yield, but also inhibits organisms in the following fermentation steps. Enzymatic
hydrolysis is to saccharify cellulose with cellulases, unfortunately, which needs costly
enzymes, large amounts of buffer solution, large hydrolysis tank, and long hydrolysis
time (Banerjee et al., 2010). Besides, substrate consistency in enzymatic hydrolysis is
typically lower than 15-20% (w/v) because of stirring and mixing issues, which limits
24
the end sugar concentration to 6-15% (w/v) after hydrolysis, and thereby the ethanol
concentration to 2.5-6% (w/v) after fermentation. Ethanol concentration of 4% (w/v) is
deemed as the minimum for the distillation process to be economical, as the energy
required for distillation is significantly reduced for ethanol concentrations above 4%
(w/v). Therefore, low ethanol concentration will significantly increase the ethanol
distillation cost (Modenbach and Nokes, 2012). In a typical cellulosic bioethanol
production scenario, feedstock, pretreatment, and enzymatic hydrolysis each accounts
for around 1/4 of the total cost and at least 1/8 of total cost is attributed to ethanol
distillation (Aden et al., 2002). Furthermore, pretreatment and enzymatic hydrolysis
require the post-treatment of large volume of waste water from pretreatment and
fermentation (Lingaraju et al., 2012). In addition, a pretreatment method may be
unsuitable to all types of lignocelluloses. For example, AFEX (Ammonia Fiber
Expansion) pretreatment does not work well with woody biomass. These issues together
result in the high production cost of current cellulose ethanol, which is the key factor
retarding the commercialization of cellulose ethanol. In summary, the challenges of
cellulose ethanol include (1) developing effective pretreatment technology to produce
readily digestible lignocellulose substrate, (2) lowering the cost and energy consumption
25
for pretreatment, (3) developing high-activity and low-cost cellulases, (4) reducing
enzyme loading, and (5) developing process and equipment for high-consistency
enzymatic hydrolysis of lignocellulosic substrate to raise the end concentration of sugar
stream for ethanol fermentation.
In order to overcome the corrosion problem of acid and avoid high cost of
cellulose hydrolytic enzymes, solid acids such as zeolite (Zhang and Zhao, 2009),
silica/carbon nanocomposites (Van de Vyver et al., 2011), sulfonated carbon materials
such as graphene (Hara et al., 2009) and CMK-3 (a type of mesoporous carbon)
(Kobayashi et al., 2010; Pang et al., 2010), and layered niobium molybdate (HNbMoO6)
(Takagaki et al., 2008; Takagaki et al., 2010), have been investigated to hydrolyze
cellulose. However, the reported yields of cellulose hydrolysis to glucose were not
satisfactory as the solid acids have limited contact with cellulose, which significantly
slows down the hydrolysis rate.
Due to the limitations and disadvantages of bioethanol, as discussed above,
interests in converting biomass into hydrocarbons is increasing as the hydrocarbons have
the same physicochemical properties as traditional transportation fuels from petroleum
(Elliott and Schiefelbein, 1989; West et al., 2009). Hydrocarbons can be produced from
26
biomass through either biological or chemical processes. In biological processes, sugars
can be fermented into hydrocarbons, but this process still requires a sugar source and
efficient organisms. Bypassing the sugar platform, biomass could be chemically
converted into a series of precursors, such as syngas by gasification of biomass, bio-oil
by pyrolysis of biomass, and HMF or levulinic acid by dehydration of sugars. Although
thermochemical processes seem simple, the purification and upgrading of the precursors
are quite complicated and challenging. In gasification, the yield and composition of
syngas are greatly dependent on feedstock composition such as carbohydrate, lignin, ash,
and extractives. Carbohydrate and lignin content will directly affect how much oxygen is
needed to supply into reactor. Metal ions in ash can catalyze different decomposition
pathways of biomass. Protein in extractives can lead to the formation of H2S and NH3,
which should be removed to avoid deactivating catalysts. For pyrolysis of biomass, the
intermediate bio-oil is a very complicated and instable mixture with more than 300
compounds, of which most are carboxylic acids, aldehydes, and aromatic compounds.
Because limited oxygen is supplied in the pyrolysis process, carbonized byproducts such
as charcoal and tar form. Because of the strong acidity and high oxygen content of bio-
oil, bio-oil needs to be purified, upgraded, and deoxygenated for advanced applications.
27
The mostly investigated upgrading method is cracking bio-oil on zeolite catalysts where
oxygen of bio-oil is removed with formation of aromatic hydrocarbons. However,
because of the strong acidity of zeolite and absence of extra hydrogen supply, a lot of
charcoal caused by strong dehydration will form in the upgrading process, which not
only decreases the recovery yield of product, but also deactivates catalyst. In summary,
thermochemical methods for converting biomass into hydrocarbons still face a lot of
challenges, such as low product selectivity, product purification and upgrading, and
catalyst deactivation (Balfanz et al., 1993; Kirubakaran et al., 2009; Zhang and Wyman,
2011; Zhao et al., 2007).
Extensive work has been done to produce hydrocarbons from lignocelluloses
based on hydroxymethylfurfural (HMF) or levulinic acid platforms. Direct conversion of
glucose into hexane on bifunctional catalyst was reported through hydrodeoxygenation
(Huber et al., 2004). However, hexane was too volatile to be used as transportation fuel.
Currently, one of the potential ways to produce extended hydrocarbon chain from sugars
was through aldol-condensation between furfural (or HMF) and acetone (Chheda et al.,
2007). This process produced HMF from glucose in DMSO (dimethyl sulfoxide); the
resulting HMF was extracted from the reaction mixture with MIBK; and the third step
28
was to conduct aldol condensation of HMF and acetone. The process involved multiple
steps and was very complex. In addition, this process could only use soluble sugars as
feedstock, and efforts were still needed to overcome the barrier of producing sugars from
lignocelluloses.
Furthermore, the intermediates (HMF and furfural) could not be cost-effectively
produced in high yield as HMF tends to polymerize and form insoluble humin in acidic
aqueous solution (Vandam et al., 1986). Many studies were conducted to improve the
selectivity of sugars to furfural/HMF. For example, organic solvent of dimethyl
sulfoxide (DMSO) was used to replace water, and the water-free environment promoted
the dehydration of glucose into HMF (Amarasekara et al., 2008). In another study,
methyl isobutyl ketone (MIBK) was added as extraction solvent to collect HMF
(Roman-Leshkov et al., 2006). The HMF formed was immediately extracted into upper
MIBK layer, which largely reduced the opportunity of the polymerization/condensation
of HMF into insoluble humin. However, it is difficult to separate HMF from the solvents
(DMSO and MIBK) because of their similar boiling points.
It was reported that HMF could be produced in high yield and with high
selectivity using ionic liquids as solvent and chromium halide as catalyst (Zhao et al.,
29
2007). The issues with this process were that ionic liquid is very expensive and difficult
to recycle, and chromium halide is potentially toxic to environment. Partial replacement
of ionic liquid with traditional cellulose solvent was also investigated (Binder and Raines,
2009). High conversions up to 90% were achieved from monosaccharide and pure
cellulose, whereas the yield was only 50% from corn stover powder. This complicated
system involved two expensive organic solvents (dimethylacetamide (DMAC) and ionic
liquid) as well as inorganic catalysts, which made them extremely difficult to recycle. In
addition, the ionic liquid has high viscosity and works well only in water-free
environment, which requires the feedstock to be finely ground and completely dry.
Unfortunately, both grinding and drying of biomass are energy-intensive.
In summary, currently there is still no cost-effective way to produce sugars and
sugar derivatives such as HMF and levulinic acid from lignocelluloses. There are still
strong demands for efficient ways to convert lignocelluloses into sugars, HMF and
levulinic acid selectively.
30
1.4 Project description
After reviewing existing techniques of producing biofuels from lignocelluloses,
sugar is an important platform compound for producing both bioethanol and
hydrocarbons because of its simple and stable chemical state, which makes subsequent
conversion processes more controllable. According to present technologies, sugars could
be fermented or catalytically converted into not only fuels but also important chemicals,
such as lactic acid, diols, and H2 (Davda and Dumesic, 2004; Onda et al., 2008;
Palkovits et al., 2010).
Therefore, a cost-effective way of producing sugars from lignocelluloses is still
crucial to the development of a biorefining industry. Direct transformation of
lignocelluloses to advanced biofuels like hydrocarbons is badly needed. Therefore, in
response to these issues, my dissertation work will focus on four projects:
(1) Investigating the effectiveness of SPORL and dilute acid pretreatments for
enzymatic saccharification of spruce for bioethanol production. (Chapter 2)
(2) Synthesizing cellulase-mimetic solid acids for cellulose hydrolysis. (Chapter 3)
(3) Developing a new process to directly saccharify lignocelluloses into sugars
without pretreatment and enzymatic hydrolysis. (Chapter 4)
31
(4) Developing a new process to directly convert lignocelluloses into hydrocarbon
precursors without prior pretreatment or saccharification. (Chapter 5)
It is the author’s hope that the present work will provide cost-effective ways of
utilizing lignocelluloses, accelerating the development of biorefinery industry.
32
Chapter 2: Comparative Study of Dilute Acid and
SPORL Pretreatments of Spruce for
Enzymatic Saccharification
2.1 Introduction
Bioethanol, produced from lignocelluloses, is a promising alternative to fossil
fuel for vehicular transportation (Carroll and Somerville, 2009; Ragauskas et al., 2006).
The benefits of cellulosic ethanol include, but are not limited to, reduced greenhouse
gases emission, value-added utilization of agricultural and forest residues, enhancement
of the rural economy, and improved national energy independence and security (Farrell
et al., 2006; Scharlemann and Laurance, 2008). A typical process for cellulosic ethanol
production consists of three steps: feedstock pretreatment to enhance cellulose
accessibility to the cellulases, enzymatic saccharification of the cellulose to glucose,
followed by fermentation of the glucose to ethanol. One of the barriers to the
commercialization of cellulosic ethanol is the lack of economical and effective
technologies for the feedstock pretreatment. Pretreatment is a necessary operation
required to achieve optimal bioconversion for all forms and types of lignocellulosic
33
feedstocks to ethanol, but is particularly important for the more recalcitrant softwoods.
The significance and importance of the pretreatment step cannot be overemphasized, as
the effectiveness of the pretreatment affects the upstream selection of biomass, the yield
of fermentable sugars, and the chemical and morphological characteristics of the
pretreated substrate which in turn govern downstream hydrolysis/saccharification.
An effective pretreatment method should be economical (in terms of both capital
and operating costs) and effective for a variety of lignocelluloses. Specifically, it should
require minimal feedstock preparation and preprocessing prior to pretreatment,
maximally recover all lignocellulosic components in usable forms with minimal
formation of fermentation inhibitors, and produce a readily digestible cellulosic substrate
that can be easily hydrolyzed with a low loading of enzymes. In the last several decades,
research and development efforts have made significant progress in pretreatment
technologies for lignocellulosic feedstocks (Chandra et al., 2007; Hendriks and Zeeman,
2009; Lloyd and Wyman, 2005; Lynd et al., 2002; Mosier et al., 2005). Many
pretreatment technologies, such as lime, dilute acid, hot water, ammonia, steam
explosion, SPORL and organosolv pretreatments have achieved varying levels of success
34
(Chandra et al., 2007; Jeong et al., 2009; Marbe et al., 2006; Monavari et al., 2009;
Nguyen et al., 2000; Sendich et al., 2008; Sierra et al., 2009; Wingren et al., 2008).
Dilute-acid pretreatment (DA) is one of the most investigated pretreatment
methods, typically using sulfuric acid at high temperature (160-200 ºC) (Lloyd and
Wyman, 2005; Schell et al., 2003). DA enhances the digestibility of lignocellulose
mainly by hydrolyzing hemicelluloses and partially hydrolyzing cellulose. DA can
achieve satisfactory levels of cellulose saccharification for agricultural residues and
some hardwood species, but is not effective for softwoods. The low pH value of the
dilute acid process causes serious equipment corrosion problem. In addition, high
temperature and low pH lead to formation of significant amounts of fermentation
inhibitors, notably furfurals from hemicellulosic sugars. Furthermore, dilute acid
pretreatment causes extensive condensation of lignin, which diminishes the commercial
value of lignin for co-products development.
Sulfur dioxide (SO2)-catalyzed steam explosion is another acidic pretreatment
using milder (higher pH) condition, which reduces equipment corrosion and minimizes
the generation of fermentation inhibitors. The process has been extensively studied by
Zacchi’s and Saddler’s groups (Chandra et al., 2007; Galbe and Zacchi, 2002).
35
Feedstock is first treated using gaseous SO2 and then subjected to steam explosion. The
process works well with agricultural residues and some hardwoods, but performs less
satisfactorily with softwoods. Two-stage steam explosion can improve the enzymatic
digestibility and overall sugar recovery from softwoods, but the advantages are partially
outweighed by increased energy consumption and operation cost (Galbe and Zacchi,
2002). The toxicity of gaseous SO2 is another concern in application. Furthermore, the
mild SO2 treatment is unable to sulfonate or dissolve lignin significantly. The lignin
extensively condensed during the subsequent steam explosion (Shevchenko, 1999),
leading to recalcitrance of steam exploded softwood substrate (Marbe et al., 2006).
Recently, we developed and reported a novel pretreatment method, Sulfite
Pretreatment to Overcome Recalcitrance of Lignocelluloses (SPORL), for robust
conversion of woody biomass including softwood to sugars (Lu et al., 2009; Melillo et
al., 2009). The pretreatment consists of a short chemical treatment of feedstock followed
by mechanical size reduction (fiberization). Wood chips or other feedstocks first react
with a solution of a sulfite salt (e.g., Na, Mg, or Ca) at 160-180 ºC and pH 2-4 for about
30 min, and are then fiberized (size-reduced) using a disk mill to generate fibrous
substrate for subsequent saccharification and fermentation. SPORL treatments produced
36
readily digestible substrates due to partial removal of lignin and lignin and partial
hydrolysis of cellulose. Because of the decreased acidity of SPORL liquor, most of
hydrolyzed cellulose and hemicellulose sugars were recovered with low amount of
fermentation inhibitors. Energy consumption for post-SPORL size-reduction of wood
chips was about 30 Wh/kg, equivalent to those consumed for size-reduction of
agricultural biomass. In addition, direct pretreatment of commercial wood chips afforded
a low liquid-to-wood ratio (<3) (Lu et al., 2009), which can lead to considerable thermal
energy savings and produce a concentrated hemicelluloses sugar stream. Because the
lignin dissolved in SPORL pretreatment hydrolysate is sulfonated (lignosulfonate), it has
a variety of commercial applications within the established market. It should be noted
that SPORL pretreatment is compatible with SO2-catalyzed steam explosion. One can
conduct sulfite-catalyzed steam explosion by using sulfite in addition to SO2 as catalysts
to significantly improve cellulose enzymatic saccharification, especially for softwoods.
To further understand the fundamentals of the SPORL pretreatment, in this study, we
systematically compared SPORL with the DA for the pretreatment of a softwood, spruce.
In order to investigate the effects of sulfite addition on the pretreatment process, the
same reaction conditions were used for DA and SPORL pretreatments, excepting that
37
sulfite was added in SPORL pretreatment. Softwoods are notoriously more recalcitrant
than other types of biomass and was chosen as feedstock in this experiment. The
experimental flowchart is summarized in Figure 2.1.
Figure 2.1 Flowchart of experiment and analysis.
Air-dried Spruce
Powder (40-mesh)
Extractive-free
Spruce
Pretreated
Spruce
NMR
Analysis
Chemical
Analysis
Saccharide analysis
Acid-insoluble lignin
Acid-soluble lignin
Inhibitor and
Sugar Analysis
DA/SPORL
Pretreatment
Enzymatic
Hydrolysis
Spent Pretreatment
Liquor
Water/Ethanol
Extraction
In-vitro Ruminal
Fermentation
38
2.2 Experimental
2.2.1 Materials
Fresh spruce chips were generously provided by the Wisconsin Rapids mill of
Stora Enso North America (now New Page Corporation, Miamisburg, Ohio). After being
air-dried, the chips were ground passing a 40-mesh screen using a Wiley mill. Chemical
composition of the spruce wood is presented in Table 2.1. Commercial enzymes,
Celluclast and β-glucosidase produced by Novozymes, were purchased from Sigma-
Aldrich (St. Louis, MO) and used as received. All the chemical reagents used in this
study were purchased from Fisher Scientific (Pittsburgh, PA) and used as received.
2.2.2 Pretreatments
Chemical pretreatments were conducted in batch mode using a lab-scale rotatable
reactor, as described previously (Lu et al., 2009; Melillo et al., 2009). Both SPORL and
dilute acid pretreatment were carried out in triplicate; the average results of the three
runs were reported. In general, ground spruce (100 g oven-dry material) was loaded into
a 1 L stainless steel vessel. Prepared pretreatment solution (500 mL 1% H2SO4 for DA
pretreatment or 500 mL 1% H2SO4 + 9 g Na2SO3 for SPORL pretreatment) was then
poured into the vessel. Three sealed 1 L vessels were mounted inside a 23 L stainless
39
steel. The system was heated via the external steam jacket and rotated at a speed of 2
rpm to provide mixing during pretreatments. The temperature was raised to 180 ºC in
about 7 min and maintained for an additional 30 min. At the end of the pretreatment, the
pretreatment spent liquor was separated from the solid (pretreated substrate) by filtration
and stored for fermentation study and chemicals analysis. The solid substrate was
collected in a Buchner funnel on filter paper and washed thoroughly with water.
Substrate yield was determined from the measured wet weight and moisture content of
the washed solid substrate.
2.2.3 Enzymatic Hydrolysis
Enzymatic hydrolysis of the pretreated substrates and original ground spruce was
conducted as described previously (Ahring and Langvad, 2008; Weimer et al., 2005).
Briefly, the hydrolysis was carried out at 50 ºC on a shaking incubator (Thermo Fisher
Scientific, Model 4450, Waltham, MA) at 150 rev/min. Substrate equivalent to 2 g
glucan was loaded into a 250 mL Erlenmeyer flask with 100 mL of 0.05 M sodium
acetate buffer (pH 4.8). Approximately 4 mg of tetracycline chloride was used to control
the growth of microorganisms and prevent consumption of liberated sugars. Cellulase (5
or 15 FPU, Filter Paper Units, per gram glucan) and β-glucosidase (10 or 30 IU,
40
International Units, per gram glucan) were loaded into the flask. Hydrolysates were
sampled periodically and subjected to glucose analysis. The hydrolysis was conducted in
duplicates for each substrate; the average is reported here.
2.2.4 Analytical Methods
Acid-insoluble lignin of spruce and pretreated SPORL and DA substrates was
determined according to National Renewable Energy Laboratory (NREL) Analytical
Procedure with modifications. Acid-soluble lignin was determined by UV at 205 nm
using an extinction coefficient of 110 L·g-1
·cm-1
(Dence, 1992).
Saccharides analysis was conducted using a Dionex High Performance Ion
Chromatography (HPIC) system (ICS-3000) equipped with integrated amperometric
detector and CarbopacTM
PA1 guard and analytical columns at 20 ºC. Eluent was
provided at a rate of 0.7 mL/min, according to the following gradient: 0→25 min, 100%
water; 25.1→35 min, 30% water and 70% 0.1 M NaOH; 35.1→40 min, 100% water. To
provide a stable baseline and detector sensitivity, 0.5 M NaOH at a rate of 0.3 mL/min
was used as post-column eluent.
41
Fermentation inhibitors generated in pretreatment including acetic acid, formic
acid, furfural, levulinic acid and 5-hydroxylmethylfurural (HMF) were analyzed using
the Dionex ICS-3000 equipped with a Supelcogel C-610H column at temperature 30 ºC
and UV detector at 210 nm. Eluent was 0.1% phosphoric acid at a rate of 0.7 mL/min.
Cellulase activity was determined using the filter paper assay recommended by
the International Union of Pure and Applied Chemists (Ghose, 1987) and is expressed in
terms of filter paper units (FPUs). β-Glucosidase activity was determined using p-
nitrophenyl-β-D-glucoside as the substrate (Cyr et al., 1988) and is expressed in terms of
International Units (IUs).
2.2.5 Whole Cell-Wall NMR of substrates
Whole cell-wall NMR of pretreated substrate and original spruce was conducted
according to (Boraston et al., 2003). In brief, 1.5 g of extractive-free spruce powder or
pretreated substrate was loaded into a 50 mL ZrO2 jar and ball-milled on a Retsch PM-
100 ball mill for 10 h (20 min on and 10 min off). The ball-milled sample (600 mg) was
dissolved in 10 mL dimethyl sulfoxide (DMSO) and 5 mL N-methylimidazole (NMI). A
clear solution was formed in approximately 3 h, depending on the sample. Excess acetic
anhydride (3 mL) was added to the solution, and the mixture was stirred for 1.5 h. The
42
resulting clear brown solution was dropped into 2000 mL water by a glass pipette, and
the mixture was allowed to stand overnight. The precipitate was recovered by filtration
through a nylon membrane (0.2 μm). The product was washed with water (250 mL) and
freeze-dried. About 60 mg dried powder was dissolved in CDCl3 for NMR spectrometry.
NMR spectra were acquired on a Bruker DRX-360 instrument fitted with a 5 mm
1H/broadband gradient probe with inverse geometry.
2.2.6 Degree of Polymerization of Cellulose
The degree of polymerization of cellulose was indirectly determined by a
viscometry method. The viscosity of cellulose in cupriethylenediamine solution was
measured using a KV 3000 kinematic viscosity bath (Koehler Instrument Company) with
a Cannon Ubbelohde capillary viscometer, according to TAPPI (Technical Association
of Pulp and Paper Industry) Standard Method T230-om-99. Cellulose samples for
viscosity measurement were prepared from original spruce and SPORL and DA
substrates by delignification using sodium chlorite according to the PPTAC (Pulp and
Paper Technical Association of Canada) Useful Method G.10U.
2.2.7 Fermentability of SPORL and DA Pretreatment Liquors
Fermentability of the pretreatment spent liquors was evaluated according to an in
43
vitro ruminal fermentation assay (Weimer et al., 2005). In vitro ruminal experiments
were conducted at 39 ºC using a single replicate of each sample, and replication was
achieved through a second in vitro run. Triplicates of a standard ryegrass were included
in each run. Incubations were conducted in nominal 60 mL serum bottles (volume was
calibrated to 0.01 mL) and that contained the equivalent of 100 mg (weighed to 0.1 mg)
of biomass material, 6.7 mL of Goering and Van Soest buffer, 0.3 mL of cysteine-sulfide
reducing agent (6.25 g/L each of cysteine HCl and Na2S·9H2O) and a CO2 gas phase.
Gas pressure readings were made at 24 h and at 96 h, using a SenSym digital pressure
gauge modified to accept a 20 gauge hypodermic needle.
2.3 Results and Discussion
2.3.1 Changes in Cell-Wall Components after pretreatments
As described in the Experimental section, ground spruce wood (40 mesh) rather
than wood chips was used as the feedstock, and mechanical size reduction was not
included in either SPORL or DA pretreatment. The consideration of doing so is that a
refining step adds difficulties to mass balance evaluations. In addition, only physical size
changes and no (significant) chemical reactions are expected during the mechanical size
44
reduction. Both SPORL and DA chemical pretreatments were carried out at the same
temperature (180 ºC), pretreatment time (7 min to attain the desired temperature and 30
min at that temperature), ratio of liquor to wood (5:1), and acid loading (5% on a dry
wood basis). The only difference is that 9% (on dry wood) sodium sulfite was added in
SPORL pretreatment.
Chemical composition of the spruce wood is presented in Table 2.1. It had a
composition of 29.0% acid-insoluble lignin and 46.7% glucose (or 42.1% glucan). The
majority of the hemicellulosic sugars were hexoses (10.8% mannose and 2.6% galactose),
accompanied by 5.5% xylose and 1.2% arabinose. The chemical compositions of
SPORL and DA pretreated substrates are compared in Table 2.1. The DA pretreatment
apparently hydrolyzed all hemicellulose from the feedstock spruce, leaving only
cellulose and lignin in the DA substrate. Essentially no delignification (but lignin
condensation) occurred during the DA pretreatment, so the lignin was enriched in the
substrate (48.5% acid-insoluble lignin) after the hemicelluloses and part of the cellulose
were dissolved. On the other hand, SPORL pretreatment retained more cellulose
(SPORL 86.3% vs. DA 71.3%, as shown in Figure 2.3) and a low level of mannan, but
less lignin (32.9% acid-insoluble lignin), in the substrate. The extent of delignification
45
(percentage of original wood lignin dissolved) during the SPORL pretreatment was
approximately 32%, calculated from the total lignin contents of spruce wood and the
SPORL substrate (Table 2.1) and the substrate yield (Figure 2.3). The differences
between DA and SPORL substrates were caused by the addition of the sulfite. First,
sodium sulfite is alkaline, buffering the pH value from 1.2 in the DA pretreatment liquor
to 2.7 in the SPORL liquor, which protected the cellulose and hemicelluloses from
extensive hydrolysis and further degradation at elevated temperature to inhibitors
(Tables 2.1 and 2.2) and prevented lignin from extensive condensation. Second, the
sulfite introduced sulfonic groups at the lignin benzylic carbons, which may (1) partially
depolymerize and dissolve the lignin and (2) increase the hydrophilicity of the residual
lignin that is retained in the pretreated substrate. These two effects are important to the
enzymatic digestibility of the SPORL substrate, as discussed in section 2.3.3.
Table 2.1 Chemical analyses of original spruce, pretreated spruce substrates and spent pretreatment liquors
Note: (1) % on oven-dry material; (2) concentration in the liquor, g/L; (3) ND-not detected; and (4) NA-not applicable.
Arabinose Galactose Glucose Xylose Mannose Acid-insoluble lignin Acid-soluble lignin
Original spruce1 (%) 1.2±0.0 2.6±0.1 46.7±0.9 5.5±0.2 10.8±0.4 29.0±1.1 1.1±0.1
DA substrate1 (%) ND
3 ND 51.9±2.1 ND ND 48.5±1.2 0.8±0.0
SPORL substrate1 (%) ND ND 66.6±3.1 ND 1.2±0.1 32.9±0.8 1.2±0.0
DA liquor2 (g/L) 0.2±0.0 0.8±0.0 6.0±0.2 0.4±0.0 1.7±0.1 NA
4 4.8±0.2
SPORL liquor2 (g/L) 0.8±0.0 2.6±0.0 5.7±0.4 4.3±0.1 9.0±0.1 NA 16.6±0.4
46
47
The concentrations of sugars and lignin in the SPORL and DA liquors, listed in
Table 2.1, provided insight into the effects of sulfite on the reactions of carbohydrates
and lignin, discussed above. The SPORL liquor contained more sugars, except for
glucose, than the DA liquor. The reason is that the higher pH value of the SPORL liquor
limited sugar degradation at high temperature. This is supported by the level of inhibitors
derived from the sugars (Table 2.2) in the spent liquors. The slightly higher content of
glucose in the DA liquor was due to the enhanced acidic hydrolysis of cellulose in the
DA pretreatment, which has been reflected by the low cellulose content in the DA
substrate. The SPORL liquor contained significantly more soluble lignin detected by UV
method than did the DA liquor. The lignin in the SPORL liquor was in the form of
lignosulfonate with a yield of 8.3% (on dry wood, calculated from the amount of the
dissolved lignin in the SPORL liquor estimated by a UV method). Lignosulfonate has
been the most successful lignin product in the market. It has been widely used as
dispersants for carbon black, pesticides, dyestuffs and pigments; emulsifiers for soils,
asphalt, waxes and oil in water; additives for drilling mud and concrete; and adhesives
and binders for animal-feed pellets, minerals, and laminates (Fengel and Wegener, 1984).
48
The SPORL pretreatment therefore has potential for high-value lignin co-products
development.
Table 2.2 Concentrations of major fermentation inhibitors in SPORL and DA pretreatment spent
liquors
Note: ―+‖ and ―-― represent standard deviation.
To investigate the behavior of cellulose during the SPORL and DA pretreatments,
viscometry was used to provide an indication of cellulose depolymerization (Table 2.3).
The viscosity of the cellulose solution from the substrates was only approximately one
tenth of that from original spruce wood, implying that the cellulose was significantly
depolymerized (hydrolyzed) during both pretreatments. This is one of the reasons why
the pretreated substrate had better enzymatic digestibility than the untreated wood (Figure
2.4). No significant difference was found between SPORL and DA substrates in cellulose
viscosity. DA cellulose solution showed a slightly lower viscosity because the lower pH
value of DA pretreatment enhanced hydrolysis of cellulose.
Spent liquor
Inhibitors (g/L)
Acid-soluble
lignin
Formic
acid Acetic acid Furfural HMF
Levulinic
acid
DA liquor 4.8±0.2 7.4±0.3 5.3±0.4 2.9±0.1 4.7±0.1 11.4±0.6
SPORL liquor 16.6±0.4 1.9±0.2 2.7±0.7 1.3±0.2 2.0±0.5 3.2±0.5
49
Table 2.3 Viscosity of cellulose solutions from untreated- and pretreated-spruces
Note: ―+‖ and ―-― represent standard deviation.
To further understand the changes of cell-wall components, in particular lignin,
during the pretreatments, whole cell-wall solution-state two-dimensional HSQC NMR
methods (Boraston et al., 2003) were used to compare the SPORL and DA pretreated
substrates with the original spruce. As shown in Figure 2.2, the cell wall HSQC NMR
spectrum from untreated spruce is typical for a softwood, showing the dominant C-H
correlations (colored in green) from cellulose and hemicelluloses (colored in black) along
with correlations from major lignin substructures (β-aryl ethers A and phenylcoumarans
B). The C-H correlations in the aromatic region of the spectrum also show typical
softwood lignin correlations from (almost) solely guaiacyl (G) units. After pretreatment
with SPORL and DA, the spruce residues contain mainly cellulose and lignin-derived
material as shown by their HSQC NMR spectra, which are consistent with the chemical
analysis results (Table 2.1). For the residual lignin, although the methoxyl signal is
readily seen, the absence of recognizable lignin subunits A and B suggests that
considerable degradation of the side chain has occurred. Although no detectable
Material Viscosity at 25 ºC (mPa·s)
Untreated spruce 25.6±0.3
SPORL substrate 2.7±0.0
DA substrate 2.4±0.0
50
monomeric sugars are in the DA pretreated sample, HSQC NMR shows some
correlations from non-cellulosic carbohydrates (colored in black). These may be
condensed with the residual lignin preventing their release by acid hydrolysis (Yasuda
and Murase, 1995), and thus would be unavailable for further conversion to ethanol.
Meanwhile lignin itself can undergo condensation reactions under acidic conditions. The
most common condensation reactions occur between benzylic α-positions of lignin
sidechains and 5-positions of other aromatic rings, producing diarylmethane structures
(Gierer, 1985). The C-H correlations (colored in pink) in the aromatic regions of these
spectra from pretreated spruce suggest that such condensation reactions indeed occurred
during both pretreatment processes. During SPORL treatment lignin was partially
sulfonated producing lignosulfonate soluble in water so that wood substrates were
partially delignified. That explains why the SPORL treated substrate has less lignin than
the DA treated material. Sulfonated lignin structures in the SPORL treated spruce
residues were not detected by NMR; model work is needed to confirm the presence. The
reason for such observations is that the number of lignin units on the substrate surface
only accounts for very small part of total lignin units and not every lignin unit is
sulfonated.
51
Figure 2.2 HSQC NMR spectra of untreated and pretreated spruce cell walls.
52
2.3.2 Mass balance of sugars after pretreatments
An ideal pretreatment is not only able to produce a readily digestible substrate,
but also to maximize the recovery of sugars in a fermentable form with limited formation
of inhibitors. As shown in Figure 2.3, the feedstock spruce wood was separated into two
fractions after the pretreatments, solid substrate and a liquid stream (spent pretreatment
liquor) containing dissolved sugars, lignin and sugar degradation products. The SPORL
substrate appeared lighter in color than the DA substrate because of its lower lignin
content and less-condensed lignin. SPORL liquor was darker than DA liquor, probably
because the former contained more soluble lignin (Table 2.1) than did the latter. To
compare the performance of SPORL and DA in sugar recovery during the pretreatment, a
mass balance of sugars during the two pretreatments was performed (Figure 2.3). From
100 g of oven-dry spruce wood (containing 46.7 g glucose, 2.6 g galactose, 10.8 g
mannose, 1.2 g arabinose, and 5.5 g xylose), 64.1 g (containing 33.3 g glucose) and 60.5
g (containing 40.3 g glucose and 7.1 g mannose) substrates were obtained from DA and
SPORL pretreatments, respectively. High substrate yield from DA pretreatment was due
in part to the retention of almost all of the original lignin. Total detected sugars in the
pretreatment liquors were 4.6 g (3.0 g glucose, 0.4 g galactose, 0.9 g mannose, 0.1 g
53
arabinose, and 0.2 g xylose) for DA pretreatment and 11.3 g (2.9 g glucose, 1.3 g
galactose, 4.5 g mannose, 0.4 g arabinose, and 2.2 g xylose) for SPORL, respectively.
The calculations from these data indicated that total sugar recovery was 56.7% (DA) and
87.9% (SPORL), and glucose recovery was 77.7% (DA) and 92.5% (SPORL).
Comparing the recovery of hexoses and pentoses, it was found that the recovery of
hexoses was much higher than that of pentoses, 62.8% vs. 4.5% for DA and 93.7% vs.
38.8% for SPORL, respectively, indicating the pentoses were easier to further degrade at
high temperature and low pH than were the hexoses. The results above clearly indicate
that under the same acid loading, temperature and reaction time, SPORL is superior to
DA for the recovery of total sugars–hexoses and pentoses. As discussed previously, this
is likely because of the higher pH value of the SPORL pretreatment liquor formed by the
addition of sulfite. High sugar recovery in SPORL process implies that fewer sugars were
degraded and limited inhibitors were generated, which will benefit the fermentation of
the liquors.
54
Figure 2.3 Mass balance of saccharides during the DA and SPORL pretreatments.
2.3.3 Enzymatic digestibility of pretreated spruce
The substrate characteristics that affect enzymatic hydrolysis of cellulose include
hemicellulose content; lignin structure, distribution and content; cellulose crystallinity
and degree of polymerization; and surface area, pore size and particle size of the substrate
(Lu et al., 2009; Mansfield et al., 1999). Generally speaking, removing hemicellulose and
lignin; swelling cellulose to destroy crystallinity; hydrolyzing cellulose to shorten chain
length (increasing the number of reducing ends for enzymes to attack); and increasing
DA substrate: 64.1 g
Glucose 33.3 g
DA spent liquor
Glucose 3.0 g
Galactose 0.4 g
Mannose 0.9 g
Arabinose 0.1 g
Xylose 0.2 g
SPORL substrate: 60.5g
Glucose 40.3 g
Mannose 7.1 g
SPORL spent liquor
Glucose 2.9 g
Galactose 1.3 g
Mannose 4.5 g
Arabinose 0.4 g
Xylose 2.2 g
100g oven-dry spruce powder
Glucose 46.7 g
Galactose 2.6 g
Mannose 10.8 g
Arabinose 1.2 g
Xylose 5.5 g
DA SPORL
55
surface area or decreasing particle size, are favorable for enzymatic digestibility of
cellulosic substrates.
(a)
(b) Figure 2.4 Comparison of time-dependent enzymatic hydrolysability of SPORL and DA
pretreated spruce at different levels of enzyme loading. (a) 15 FPU cellulase+30 IU β-glucosidase
per gram of cellulose; (b) 5 FPU cellulase+10 IU β-glucosidase per gram of cellulose), 50 ºC, pH
4.8 and on a 250 rpm shaker. CGCY: Cellulose-to-Glucose Conversion
0
20
40
60
80
100
0 10 20 30 40 50
CG
CY
(%
)
Enzymatic hydrolysis time (h)
SPORL
Untreated
DA
0
20
40
60
80
100
0 10 20 30 40 50
CG
CY
(%
)
Enzymatic hydrolysis time (h)
SPORL
Untreated
DA
56
The enzymatic digestibility of SPORL- and DA-pretreated and original spruce
wood was compared in Figure 2.4. At enzyme loadings of 15 FPU (Filter Paper Units)
cellulase and 30 IU (International Units) -glucosidase per gram cellulose (Figure 2.4 a),
the SPORL pretreated spruce displayed much greater hydrolysis than did the DA
pretreated spruce and untreated spruce. For example, only 25% and 55% of the cellulose
in untreated spruce and the DA substrate, respectively, was hydrolyzed to glucose after a
2-day hydrolysis, whereas 93% of the cellulose in SPORL substrates was saccharified
within the same time. When the enzyme loadings were reduced to 5 FPU cellulase and 10
IU -glucosidase per gram cellulose (Figure 2.4 b), the cellulose-to-glucose conversion
after a 2-day hydrolysis of the SPORL substrate (71%) was still substantially higher than
those of the DA substrate and untreated spruce (49 and 17%, respectively). The results
clearly indicated that the SPORL substrate had substantially better enzymatic digestibility
than did the DA substrate under the same hydrolysis conditions. As discussed above, the
DA substrate contained no hemicellulose and had slightly lower viscosity (lower
cellulose degree of polymerization), compared to the SPORL substrate. Based on these
factors alone, the DA substrate might be assumed to have a better digestibility. However,
lignin (both content and nature), and other factors, resulted in quite the opposite situation.
57
The DA substrate contained approximately 50% lignin, and the lignin was highly
condensed, extremely hydrophobic, and covered the surface of the substrate, all of which
enhanced the effect of lignin as physical barrier and non-productive enzyme adsorbent.
On the other hand, the SPORL substrate contained less lignin, and the possible
sulfonation of lignin in the substrate made the lignin less hydrophobic, which may have
reduced the non-productive hydrophobic adsorption of enzymes onto the lignin. Similar
report has shown that addition of surfactants can significantly decrease the non-
productive adsorption of enzymes on lignin because the hydrophobic end of surfactants
was adsorbed onto hydrophobic lignin and the hydrophilic end of surfactants decreased
the adsorption of enzyme onto lignin surface (Qing et al., 2010). Approximately one
third (~33%, Table 2.1) of the SPORL substrate was lignin, but this lignin retarded the
hydrolysis little. The observation implies that the residual lignin in the SPORL substrate
was enzyme-friendly and behaved differently from the acid-condensed DA lignin. This
suggests that costly delignification is not the only way to remove the recalcitrance
(attributable to lignin) to enzymatic degradation of cellulose; less expensive lignin
modification may be more promising. The results also suggest that the action of lignin as
a purely physical barrier played a less important role in its impacts on enzymes,
58
compared to other interactions such as non-productive adsorption, which agrees with our
previous results (Marbe et al., 2006).
2.3.4 Fermentability of spent pretreatment liquors
Potential fermentation inhibitors formed during the DA and SPORL pretreatments
are listed in Table 3, including the soluble lignin, acetic acid released from acetyl groups
on hemicelluloses, furfural derived from pentoses, HMF from degradation of hexoses,
and levulinic and formic acids from successive decomposition of HMF. The formation of
the furfurals and their degradation products (levulinic and formic acids) is shown in
Figure 2.5.
Figure 2.5 Inhibitors formation from cellulose and hemicellulose during pretreatment.
59
The data in Table 2.2 clearly indicated that fewer inhibitors were formed from
degradation of saccharides during the SPORL pretreatment than the DA pretreatment.
Due to the interferences of unknown impurities in spent liquors, the amounts of formic
acid and acetic acid could not be determined accurately. However, this did not affect the
evaluation on these two methods as acetic acid is not derived from saccharide portion of
the carbohydrates and the molar amount of formic acid corresponds equivalently to the
molar amount of levulinic acid. The total of known inhibitors (furfural, 5-hydroxy-
methylfurfural, and formic, acetic and levulinic acids) formed in SPORL were only 35%
of those formed in DA pretreatment. As discussed above, this was owing to the addition
of sulfite in the SPORL pretreatment, which increased the pH value of the pretreatment
liquor, limiting extensive degradation of saccharides. The high lignin concentration in
SPORL liquor is due to the formation of water-soluble lignosulfonate. Low traditional
inhibitor (Table 2.2) and high sugar (Table 2.1) concentrations suggest that good
fermentability can be expected for SPORL liquor.
60
Figure 2.6 Fermentability of SPORL and DA pretreatment spent liquors by in vitro ruminal
fermentation assay. (DM: dry matter)
The fermentability of SPORL and DA liquors was evaluated by an in vitro
ruminal gas production assay, as shown in Figure 2.6. Surprisingly, the two liquors did
not show difference in net gas yield within the first 24 h, while the gas yield of SPORL
liquor decreased beyond this point. The loss of gas was not clear at this time and will be
further investigated in future work. In tests with various perennial grasses, in vitro ruminal
gas production has been shown to provide a reasonable surrogate estimate of ethanol
production potential by an enzyme/yeast SSF (Simultaneous Saccharification and
Fermentation) system (Weimer et al., 2005). In addition, the in vitro ruminal gas
production assay has displayed less inhibition than SSF for certain forages that contain
0
100
200
300
400
500
0 20 40 60 80 100
Ne
t g
as y
ield
(m
L/g
DM
)
Fermentation time (h)
SPORL
DA
61
natural fermentation inhibitors (e.g., some varieties of switchgrass). It was thus a surprise
that in vitro ruminal gas production from the SPORL-pretreatment liquor was less than
that of DA-pretreatment liquor that apparently had higher concentrations of classical
pretreatment-derived inhibitors (e.g., furfural and HMF). This inhibition may have been
due to lignosulfonate (present in much higher concentrations in the SPORL liquor).
Alternatively, inhibition could have been due to the presence of other metabolites that
may be released from wood upon pretreatment.
2.4 Conclusion and recommendations
SPORL pretreatment significantly decreased the recalcitrance of spruce wood and
allowed nearly complete enzymatic hydrolysis (>90%) within 24 h with a cellulase
loading of 15 FPU/g cellulose. SPORL reduced the recalcitrance not only by dissolving
hemicellulose and depolymerizing cellulose, but also by partially (32%) dissolving lignin
and sulfonating the residual lignin in substrate. The latter presumably made lignin more
hydrophilic and thereby reduced the hydrophobic interactions between lignin and the
enzymes. SPORL pretreatment achieved a significantly higher sugar recovery and
produced much lower levels of traditional fermentation inhibitors than dilute acid
62
pretreatment. The overall saccharides (hexoses and pentoses) recovery of SPORL was
87.9%, compared to 56.7% with dilute acid.
63
Chapter 3: Synthesis of Cellulase Mimetic Solid
Acid for Cellulose Hydrolysis
3.1 Introduction
Energy-efficiently and cost-effectively hydrolyzing cellulose to glucose is the
bottleneck to converting lignocelluloses to value-added chemicals and liquid fuels
through the sugar platform (Ragauskas et al., 2006). Acids and enzymes are the most
common catalysts for hydrolyzing cellulose into glucose. Sulfuric acid is an inexpensive
and effective acid catalyst. For example, concentrated sulfuric acid (>65%) can
effectively swell and hydrolyze crystalline cellulose into glucose at moderate temperature
through disrupting strong hydrogen bonds in crystalline cellulose, but acid recovery is
difficult. Dilute acid (0.5-10%) is also able to hydrolyze cellulose, but needs higher
temperature (150-220 °C). The drawback of high-temperature acid hydrolysis is the
undesirable dehydration of glucose to hydroxymethylfurfural (HMF) and further
degradation to levulinic and formic acids, which are fermentation inhibitors. These side
reactions not only reduce glucose yield but also generate fermentation inhibitors. Other
issues with sulfuric acid hydrolysis include process effluent treatment and gypsum
64
handling from acid neutralization (Bienkowski et al., 1984; Harris and Lang, 1947; Li et
al., 2008; Taherzadeh and Karimi, 2007; Taherzadeh and Karimi, 2008). Cellulase is
highly selective and able to hydrolyze cellulose at lower temperature (~50 °C). However,
enzymatic hydrolysis of cellulose is a slow and incomplete reaction, typically taking days
to achieve satisfactory yields. In addition, prior to enzymatic hydrolysis of
lignocelluloses, an energy-intensive pretreatment is necessary to remove the recalcitrance
to enzymes. Finally, cellulases are still expensive enzymes today, making it unaffordable
to pursue high hydrolysis yield by increasing enzyme loading (Chandra et al., 2007;
Huang, 1975; Medve et al., 1998).
The difference in the temperature required for acidic and enzymatic hydrolysis of
cellulose is related to their reaction activation energies. It was reported that the activation
energies for cellulose hydrolysis and glucose dehydration with dilute acid ranged
between 170-180 and 135-145 kJ·mol-1
, respectively, no matter what temperature and
acid concentration were used (Girisuta et al., 2006; Heeres et al., 2007). On the other
hand, the activation energy of enzymatic hydrolysis was only 3-50 kJ·mol-1
(Bravo et al.,
2000; Hsuanyu and Laidler, 1985), which is the reason why enzymatic hydrolysis of
cellulose can be conducted at ~50 °C (Schall et al., 2006; Shuai et al., 2010). However,
65
enzymatic hydrolysis of cellulose is slow and may take days to complete. Because of the
thermal instability of cellulases, the hydrolysis rate cannot be accelerated through
increasing temperature because cellulase will denature at high temperature. An ideal
catalyst for cellulose hydrolysis should lower the activation energy of cellulose
hydrolysis, which will lower hydrolysis temperature, shorten reaction time, and reduce
the undesirable glucose degradation at elevated temperature.
Solid acid catalysts for cellulose hydrolysis have recently drawn a lot of attention
because, on the one hand, the solid acid is less expensive and easier to recover/reuse than
the cellulases, and on the other hand, the solid acid lowers the activation energy of
cellulose hydrolysis thus allows lower hydrolysis temperature than homogenous acids. In
addition, because of the thermal stability of the solid acid, increasing temperature could
be an option to speed up the reaction rate. Sulfuric acid-carbonized biomass is one of the
mostly studied solid acids for hydrolyzing cellulose (Suganuma et al., 2008; Yamaguchi
et al., 2009). Cellulose hydrolysis catalyzed by this carbonized-biomass solid acid had
lower activation energy (110 kJ·mol-1
vs. 170 kJ·mol-1
) than the one catalyzed by sulfuric
acid. Other solid acids studied include zeolite (Zhang and Zhao, 2009), silica/carbon
nanocomposites (Van de Vyver et al., 2011), sulfonated carbon materials (graphene (Hara
66
et al., 2009) and CMK-3 (Kobayashi et al., 2010; Pang et al., 2010)), and layered niobium
molybdate (HNbMoO6) (Takagaki et al., 2008; Takagaki et al., 2010). Recently,
magnetic solid catalysts bearing –SO3H were synthesized for hydrolyzing cellulose. The
uniqueness of the catalysts is that they can be magnetically recovered from the reaction
system (Bond et al., 2010; Fu et al., 2011; He et al., 2010; Lai et al., 2011). The
performance of the solid acid catalysts above for the hydrolysis of cellulose to glucose
varied, with the overall glucose yield of 30-75% (Pang et al., 2010; Van de Vyver et al.,
2011). To achieve a reasonable glucose yield, high temperature (>150 °C) is necessary.
Another issue of the solid acids is the need for very high catalyst/substrate mass ratio
(Van de Vyver et al., 2011). One of the reasons is that the current solid acids, unlike
cellulases, do not have dedicated substrate-binding sites for associating cellulose onto
catalyst surface.
Structurally, most cellulases have two domains, a catalytic domain and a cellulose
binding domain linked by a peptide (Boraston et al., 2003; Coutinho et al., 1993; Gilkes
et al., 1992; Gilkes et al., 1993). The binding domain is responsible for associating the
cellulases to cellulose, while the catalytic domain is to hydrolyze glycosidic linkages of
cellulose chain. In cellulases, acid groups such as carboxylic and phenolic hydroxyl
67
groups of amino acids in the catalytic domain function as proton sources for cleaving
glycosidic bonds (Berti and Tanaka, 2002; Mccarter and Withers, 1994). It is believed
that enzyme-cellulose binding is via ion attraction, hydrophobic affiliation, or hydrogen
bonding between cellulose and aromatic amino acids, such as tryptophan and tyrosine, on
the enzyme (Creagh et al., 1996; Tomme et al., 1996). Mosier et al. (Mosier et al., 2004)
screened potential cellulose binding compounds for designing mimetic catalysts and
found that aromatic, hydrophobic, and planar structures, such as indole and tryphan blue,
have high affinity to cellulose. They also pointed out that hydrophobic interactions and
hydrogen bonding are the most important two contributors to the enzyme adsorption onto
cellulose.
Chemically, it is possible to synthesize a cellulase-mimetic catalyst containing
both catalytic and cellulose-binding domains through molecular design, as shown in
Figure 3.1. For example, acidic groups, such as sulfonic acid (-SO3H), can be introduced
onto the surface of the catalyst and function as catalytic domains. As hydroxyl groups on
cellulose form hydrogen bonds with electronegative groups such as hydroxyl, amine, and
halide, the electronegative groups anchored on the surface of synthesized catalyst can be
expected to function as cellulose binding domains. In addition, the formation of hydrogen
68
bonding between the electronegative groups of catalyst and hydroxyl group proton helps
to disrupt the intra- and inter-molecular hydrogen bonds of cellulose and thereby
enhances the swelling of cellulose (Dawsey and Mccormick, 1990; Turbak et al., 1977).
Figure 3.1 A proposed model of cellulase-mimetic solid acid and cellulose interaction.
So far, most of the studies on solid acids for cellulose hydrolysis have focused
just on the catalytic activity of the catalysts (Hara et al., 2009; Onda et al., 2008;
Suganuma et al., 2008). Little attention has been paid to the interaction/binding between
the catalyst and cellulose and, to our knowledge no research has been reported on
introducing cellulose-binding sites onto synthesized solid acids. In this paper, we report a
simple method to synthesize a solid acid catalyst containing both cellulose binding and
hydrolyzing sites to mimic cellulase for cellulose hydrolysis.
Cl ClClClClCl ClClClCl SO3H
O
H
HO
H
HO
H
HOHH
O
OH
O
H
H
HO
H
HOHH
O
OH
O
H
H
HO
H
HOHH
O
OH
O
H
H
HO
H
HOHH
OH
OH
Cellulose chain
Mimetic cellulase
Binding domain Catalytic domain
Hydrogen bond
Binding domain
69
3.2 Experimental
3.2.1 Chemicals and materials
Chemical reagents used in this study were purchased from Fisher Scientific
(Pittsburgh, PA) or Sigma-Aldrich (St. Louis, MO) and used as received. Chloromethyl-
polystyrene (CP) (100 mesh) with 3.5-4.5 mmol chlorine per gram resin and Amberlyst-
15 (100 mesh) with 4.5 mmol -SO3H per gram resin were dried at 105 ºC overnight prior
to use.
3.2.2 Synthesis of cellulase mimetic solid acid
(1) Synthesis of solid acid with –Cl binding domain (CP-SO3H)
The synthesis of CP-SO3H is schematically shown in Figure 3.2. The oven-dried
CP resin (3 g) and DMF (50 mL) were added into a 100-mL flask and stirred at 120 ºC
for 30 min in an oil bath. Then 2 g of sulfanilic acid was added into the flask and reacted
with the resin at 120 ºC for 48 h. After the reaction, the resin was recovered by filtration
(10 μm retention) followed by washing with 100 mL ethanol and 200 mL water
sequentially. The resin prepared was in the form of a green powder after being dried at
105 ºC overnight.
70
Figure 3.2 Synthesis of CP-SO3H.
(2) Synthesis of solid acid with –NH2 binding domain (NP-SO3H)
Prepared CP-SO3H (2 g) was suspended in 10 mL water and reacted with 1 mL 30%
NH3·H2O at 100 ºC for 2 h. The resulting resin was filtered and acidified with 200 mL of
5% HCl and then washed with deionized water until neutral pH was achieved.
(3) Synthesis of solid acid with –OH binding domain (OP-SO3H)
Prepared CP-SO3H (2 g) was suspended in 10 mL water and reacted with 1 mL of
50% NaOH at 100 ºC for 2 h. The resulting resin was filtered and acidified with 200 mL
of 5% HCl and then washed with deionized water until a neutral pH was achieved.
(4) Synthesis of solid acid without binding domain (HP-SO3H)
Prepared CP-SO3H (2 g) was suspended in 10 mL toluene and reacted with 1 g
NaH at room temperature for 24 h. The resulting resin was filtered and acidified with 200
mL of 5% HCl and then washed with deionized water until a neutral pH was achieved.
71
3.2.3 Hydrolysis of biomass with solid acids
Figure 3.3 A model for cellulose hydrolysis on solid acid (CP-SO3H) surface.
The proposed mechanism of SO3H-catalyzed hydrolysis of the glycosidic bond is
shown in Figure 3.3. A predetermined amount of solid acid and sulfuric acid were mixed
with 100 mg substrate (cellobiose, starch, or Avicel) into 1 mL water in 20-mL glass
vials and mixed well. Vials were heated in oil bath at different temperatures for varied
durations (reaction conditions list in the notes under Figures or Tables). Supernatant
samples were taken to determine the glucose yield.
72
3.2.4 Adsorption of glucose and cellobiose on CP-SO3H
CP-SO3H (600 mg) was added into 12 mL water and vortexed to mix well. The
mixture was divided into 12 equal portions in microtubes. Then 100 mg glucose or 90 mg
cellobiose was loaded into each microtube and mixed well. All microtubes were kept at
room temperature and vortexed every 5 min. One microtube for each was taken and
centrifuged at a time of 10, 20, 30, 60, 90, and 120 min. Glucose and cellobiose in the
supernatant were determined with High Performance Ion Chromatography, as described
below, to calculate the adsorption capacity of CP-SO3H.
3.2.5 Determination of glucose
Procedure is described in section 2.2.6.
3.2.6 FT-IR spectra of prepared solids
Fourier transform infrared (FT-IR) spectra were recorded on a PerkinElmer
Spectrum 100 FT-IR spectrophotometer with a universal attenuated-total-reflection (ATR)
sampling accessory (Waltham, MA). ATR allows samples to be examined directly in the
solid state without further preparation.
73
3.3 Results and discussion
3.3.1 Screening of binding groups
Four types of solid acids with different binding domains were synthesized and
tested for their catalyst activities. The proposed mechanism of cellulose hydrolysis
catalyzed by CP-SO3H is schematically shown in Figure 3.3. The sulfonation of CP was
verified with FT-IR. The spectra of CP and CP-SO3H resins are shown in Figure 3.4.
Wavenumbers (cm-1)
Figure 3.4 FT-IR spectra of (a) CP resin and (b) CP-SO3H resin.
The peak at 1260 cm-1
results from the adsorption of the -CH2Cl groups (Zhang et
al., 2011), while those at 1150 and 1100 cm-1
are from the vibration and stretching of
a
b
-CH2Cl
-SO3H
3
1
74
sulfonic acid group (-SO3H) (Panicker et al., 2006). It is apparent that the intensity of the
-CH2Cl band decreased after the sulfonation, indicating the substitution of -Cl via
sulfanilic acid. The increased intensities at 1150 and 1100 cm-1
further verified the
introduction of the -SO3H group. The content of the acid group (-SO3H) introduced was
approximately 0.0067 mmol per gram of the resin, as determined by electro-conductivity
titration. Considering the fact that the sulfonation reaction occurred primarily on the resin
surface, the surface concentration of -SO3H would be higher than the value above
calculated based on total resin. Surface concentration of -SO3H was further estimated
from the spectra of ATR-FTIR. As shown in Figure 3.4, it was estimated one third of
surface chloride groups were substituted with -SO3H. Further study is needed to the
distribution of -SO3H groups inside and on the surface of the resin. The FT-IR spectra of
NP-SO3H, OP-SO3H and HP-SO3H in Figure 3.5 show that the chloride group was
successfully replaced. The peak representing the -CH2Cl band significantly decreased or
almost disappeared, indicating that solid acid with different binding groups were
synthesized successfully. It was hard to quantitatively and accurately estimate the degree
of substitution as the styrene units only accounted for very small part of the total styrene
75
units. However, as –Cl was easily substituted by alkalis, the substitutions of –Cl by OH-
and NH3 on resin surface were assumed to be complete.
Figure 3.5 FT-IR spectra of solid acids with different binding domains. (a) CP resin, (b) NH2
resin, (c) OH resin, and (d) H resin.
Then effect of binding groups on the hydrolysis of cellulose was tested at
temperatures of 100 and 120 ºC. It can be seen from Figure 3.6 that at both temperatures,
the hydrolysis rate of cellobiose followed the order of Cl>NH2>OH>H. According to
the Arrhenius equation, the apparent activation energies of cellobiose hydrolysis
catalyzed by the four types of solid acids were estimated, as shown in Table 3.1.
76
(a)
(b)
Figure 3.6 Cellobiose hydrolysis catalyzed by four types of solid acids as a function of time.
0
10
20
30
40
50
60
70
80
0 100 200 300 400 500 600
Co
nve
rsio
n(%
)
Reaction time (min)
Cl
NH2
OH
H
0
10
20
30
40
50
60
0 20 40 60 80 100 120
Co
nve
rsio
n(%
)
Reaction time (min)
Cl
NH2
OH
H
77
(a) 100 ºC, (b) 120 ºC.
It can be seen that -Cl lowered apparent activation energy of cellobiose
hydrolysis and therefore was chosen for further investigation for its ability to hydrolyze
cellulose. The reason why the chloride group (-Cl) showed the lowest activation energy
in cellobiose hydrolysis was presumably that chlorine groups could form stronger
hydrogen bonds with hydroxyl groups of cellulose but also disrupt intra- and inter-
hydrogen bonds of cellulose, which would enhance the swelling and dissolution of
cellulose. This is consistent with the wide involvement of chloride in cellulose solvents,
such as DMAC/LiCl and chloride-containing ionic liquids (Kosan et al., 2008;
Mccormick et al., 1985; Remsing et al., 2006). The evidence that the polymers of vinyl
chloride and vinylidene chloride had higher adhesion to cellulose than those of ethylene,
propylene and vinyl acetate (McLaren, 1948) supports the hypothesis that chloride has
cellulose-binding capacity.
Table 3.1 The apparent activation energies of cellobiose hydrolysis catalyzed by four types of
solid acids
Binding group Cl H OH NH2
Apparent activation energy (kJmol-1
) 77.7 88.5 89.5 104.7
78
3.3.2 Mechanism study
When cellobiose is hydrolyzed with a homogeneous acid, such as H2SO4, the
reaction follows collision theory, i.e. the reaction rate depends on the concentration of
both the acid and cellobiose. Lowering acid or cellulose concentration will slow the
reaction rate (Bobletera et al., 1986; Lee et al., 2003). However, the situation is different
when solid acid is used as catalyst. The reaction rate is not only dependent on the
concentrations of cellobiose and acid sites/groups on the solid acids, but also on the
collision probability between cellobiose and the acid sites. Therefore, if cellobiose can be
selectively adsorbed (concentrated) onto the catalyst surface, the hydrolysis rate would be
enhanced. Meanwhile, the resulting glucose molecules need to be desorbed (released)
from the catalyst surface as soon as they are generated to free the binding sites on the
catalyst surface for new cellobiose molecules. It is hypothesized that polysaccharide and
oligosaccharide molecules would be more preferably adsorbed onto the CP-SO3H surface
compared to monosaccharides because the former has more binding locations (hydroxyl
groups) than the latter.
79
In order to verify the above hypothesis, the adsorption of glucose and cellobiose
onto the CP-SO3H surface was studied in aqueous solution at room temperature. As
shown in Figure 3.7, CP-SO3H adsorbed both glucose and cellobiose, which is different
from the solid acid from carbonized biomass that only adsorbed cellobiose but not
glucose (Creagh et al., 1996; Hara et al., 2009; Suganuma et al., 2008). Similarly,
sulfonated activated carbon was able to adsorb oligosaccharides but not monosaccharides,
and the adsorption was dependent on the size of the oligosaccharides (Hara et al., 2009).
It was proposed that the two solid acids above adsorbed the sugars through the interaction
(hydrogen binding) between phenolic hydroxyl groups on the activated carbon and the
oxygen in β-1,4-glycosidic bond, which explained why carbonized biomass only
adsorbed cellobiose but not glucose, and the activated carbon had greater affinity for
longer oligosaccharides. However, this mechanism is unable to explain the adsorption of
glucose on CP-SO3H since there is no glycosidic oxygen in glucose. We believe that CP-
SO3H adsorbs the sugars through the hydrogen bonding between the chloride groups on
the catalyst and the hydroxyl groups of the sugars, as will discussed further below.
When the same amount of glucose units (glucose/cellobiose, 2:1, mol/mol) was
loaded, cellobiose was adsorbed more onto the catalyst surface than glucose (Figure 3.7),
80
suggesting that CP-SO3H has a higher affinity for cellobiose than glucose. This is
presumably because more hydrogen bonds are formed between the catalyst and
cellobiose due to the higher number of hydroxyl groups available in cellobiose. In order
to further verify the higher affinity of cellobiose than glucose to CP-SO3H, the
competitive adsorption of cellobiose and glucose on CP-SO3H was investigated. It was
observed (data not shown here) that when cellobiose and glucose were mixed with CP-
SO3H at equivalent loadings of glucose units, cellobiose was selectively absorbed by CP-
SO3H. Certainly, the preferable adsorption of cellobiose over glucose is desirable and
critical for cellobiose hydrolysis to proceed, which ensures the selective adsorption of
cellobiose to the catalyst and prompt desorption of glucose from the catalyst.
85
90
95
100
105
0 20 40 60 80 100 120 140
Re
sid
ua
l sa
ccha
ride
s(%
)
Time (min)
a
d
c
b
81
Figure 3.7 Timecourse for adsorption of glucose and cellobiose onto resins in aqueous solution.
(a) glucose on PS-SO3H; (b) cellobiose on PS-SO3H; (c) glucose on CP-SO3H; (d) cellobiose on
CP-SO3H.
To demonstrate the importance of the -Cl groups on CP-SO3H in adsorbing
saccharides through hydrogen bonding, the adsorption of glucose and cellobiose onto
sulfonated polystyrene resin (Amberlyst-15, abbreviated as PS-SO3H) was studied. As
expected, PS-SO3H did not adsorb glucose or cellobiose at all because it did not have the
sites (-Cl) for binding the sugars, as indicated in Figure 3.7. This is presumably the
reason why solid Bronsted acids, such as inorganic oxide solid acids, H-mordenite, and
SO3H-bearing polymers like Nafion NR50 and Amberlyst-15 are unable to hydrolyze
cellulose because they do not have the binding capacity to associate the substrate onto the
catalyst surface (Hara et al., 2009; Suganuma et al., 2008).
82
Figure 3.8 Comparison of cellobiose hydrolysis catalyzed by (a) CP-SO3H and (b) PS-SO3H.
(note: 500 mg resins, 100 mg cellobiose, 100 ºC, 1 mL water, 2 h).
The catalytic capacity of CP-SO3H and PS-SO3H was compared for hydrolyzing
cellobiose at the same catalyst loading of 500 mg resin per 100 mg cellobiose. As shown
in Figure 3.8, the two catalysts had comparable glucose yield. More precisely, PS-SO3H
had a slightly higher glucose yield than CP-SO3H. As every styrene unit contains a –
SO3H group in PS-SO3H, PS-SO3H has much higher acid site density on the resin surface
than CP-SO3H. With the similar glucose yields the hydrolysis of cellobiose, CP-SO3H
should have much greater specific catalytic capacity than PS-SO3H, which is probably a
result of the superior cellobiose-binding ability due to -Cl groups on CP-SO3H.
3.3.3 Hydrolysis of cellulose with CP-SO3H
The catalytic capacity of CP-SO3H was first evaluated and compared with sulfuric
acid using a soluble substrate cellobiose. As shown in Figure 3.9, at an equivalent acid
loading of 0.0017 mmol (0.17 mg H2SO4 or 250 mg CP-SO3H) per 100 mg cellobiose,
approximately 73% cellobiose was hydrolyzed into glucose with CP-SO3H in 2 h at 120 ºC,
while only 4% was hydrolyzed with sulfuric acid under the same conditions. When the
acid loading was doubled to 0.0033 mmol (0.34 mg H2SO4 or 500 mg CP-SO3H),
83
cellobiose was completely hydrolyzed with CP-SO3H within 2 h, whereas only 8% was
hydrolyzed by sulfuric acid.
Figure 3.9 Hydrolysis of cellobiose catalyzed by CP-SO3H and sulfuric acid. (a) 0.0033 mmol
CP-SO3H, 120 °C; (b) 0.0017 mmol CP-SO3H, 120 °C; (c) 0.0033 mmol H2SO4, 120 °C; (d)
0.0017 mmol H2SO4, 120 °C.
Table 3.2 Hydrolysis of starch and Avicel cellulose catalyzed by CP-SO3H and sulfuric acid
Entry Feedstock Temp. (°C) Time (h) Yield (%)
Sulfuric acid CP-SO3H
1 Cellobiose 100 2 NA 62
2 Cellobiose 100 4 NA 99
3 Cellobiose 120 2 7.9 100
4 Starch 120 2 3 100
5 Avicel cellulose 120 2 0.5 32
6 Avicel cellulose 120 10 NA 93
7 Starch 120 20 NA 25
8 Avicel cellulose 120 20 NA 27
9 Cellobiose 140 0.5 7 98
10 Cellobiose 160 0.5 15 NA
Note: ―NA‖refers to ― not applicable‖; entry 1 to 8 (100 mg carbohydrate; 1 mL water; 0.0033 mmol
catalyst); entry 9 to 10 (100 mg carbohydrate; 1 mL water; 0.0017 mmol catalyst)
84
The performance of CP-SO3H at different temperatures was also investigated. It
worked well at 100 ºC, though the yield of cellobiose hydrolysis was lowered to 62.1%
after a 2-h reaction. If the reaction was extended to 4 h, complete hydrolysis of cellobiose
(99.3%) was achieved. Low temperature slowed down the reaction, but had the advantage
of avoiding some further degradation of glucose. When temperature was elevated to 140
ºC, the time for complete hydrolysis was shortened to 30 min (as presented in Table 3.2).
These results suggest that CP-SO3H performed better than sulfuric acid in catalyzing
cellobiose hydrolysis. This is presumably attributed to the capacity of adsorbing
cellobiose and high surface acid concentration, as will discussed below.
The durability of CP-SO3H was examined by recycling the catalyst three times in
cellobiose hydrolysis. The catalyst was recovered by filtration after the hydrolysis,
washed, dried at 105 ºC overnight, and reused in the next batch of hydrolysis reactions.
As shown in Figure 3.10, the catalyst activity did not decline after recycling three times,
indicating that CP-SO3H resin had good stability in catalytic activity. The chemically
bonded -Cl and -SO3H on CP-SO3H did not have the leaching problem that happened to
the carbon based-solid acid (Van de Vyver et al., 2010).
85
Figure 3.10 Cellobiose hydrolysis catalyzed by recycled CP-SO3H resin. (Note: 100 mg
cellobiose, 120 ºC, 1 mL water, 2 h).
The catalytic activity of CP-SO3H in hydrolyzing polysaccharides was also
investigated. As shown in Table 3.2, starch and crystalline cellulose (Avicel) were
hydrolyzed with CP-SO3H and sulfuric acid at an equivalent acid loading. It was found
that starch was completely hydrolyzed to glucose by CP-SO3H in 2 h (entry 4), while
only 32% of Avicel cellulose was hydrolyzed into glucose under the same conditions
(entry 5). The crystalline structure of cellulose is considered as the major reason for the
slow hydrolysis and low yield of cellulose hydrolysis to glucose. When hydrolysis was
extended to 10 h, 93% of Avicel cellulose could be hydrolyzed. By contrast, with sulfuric
86
acid at an equivalent acid loading at the same temperature and reaction time, starch was
only 3% hydrolyzed (entry 4), and almost no Avicel cellulose was hydrolyzed (entry 5).
This is because cellulose hydrolysis catalyzed by homogeneous acid hydrolysis has a
higher activation energy than the one catalyzed by CP-SO3H. The capability of CP-SO3H
to hydrolyze Avicel at moderate temperature can be attributed to the presence of -Cl. The
chloride groups, in addition to binding cellulose, are possibly able to decrystallize
cellulose by breaking intra- and inter-molecular hydrogen bonds between the hydroxyl
groups of cellulose through forming stronger hydrogen bonds with the hydroxyl groups.
The mechanism needs further investigation, and is outside the scope of this thesis.
To further understand why CP-SO3H can hydrolyze cellobiose and cellulose at a
decent rate at moderate temperature, the activation energies of cellobiose and cellulose
hydrolysis catalyzed by CP-SO3H were estimated. The activation energy of cellobiose
hydrolysis on two types of synthesized resins was calculated from the Arrhenius equation:
(k = e−Ea RT ). In order to estimate the apparent activation energy, the reactions were
assumed to be pseudo-first order reaction. The rate constant was obtained from equation:
k=ln(2)/t1/2 (t1/2 is half-life time). Half-life time at different temperatures was recorded
and plotted, as shown in Figure 3.11.
87
Figure 3.11 Arrhenius plot for cellulose hydrolysis catalyzed by CP-SO3H.
y = -9.969x + 19.436R² = 0.9995
-7.5
-7
-6.5
-6
-5.5
-5
-4.5
2.4 2.45 2.5 2.55 2.6 2.65 2.7
lnk
1/T (10-3 K-1)
Ea=9.4142×8.31=78 kJ mol-1
Ea=9.969×8.31=83 kJ mol-1
Cellobiose
Crystalline cellulose
88
As shown in Table 3.3, the apparent activation energies for hydrolysis of
cellobiose and Avicel are 78 and 83 kJmol-1
, respectively, at 373-413 K. The similar
activation energies of the two reactions indicate that they can proceed at the same
temperature, which has been verified by the results above. Although hydrolysis of
cellobiose and Avicel by CP-SO3H had a similar activation energy as mentioned above,
the hydrolysis of Avicel was much slower than that of cellobiose under the same
conditions, as shown in Table 3.2 (entries 3 and 5). This is attributed to the solid state of
Avicel, which has fewer collision chances with the catalyst, compared to soluble
cellobiose. The values of the activation energy are significantly smaller than those of the
hydrolysis reactions catalyzed by sulfuric acid (170 kJmol-1
for crystalline cellulose and
130 kJmol-1
for cellobiose) (Bobletera et al., 1986) and sulfonated active carbon (AC-
SO3H) (110 kJmol-1
) (Suganu ma et al., 2008). This is the reason why CP-SO3H was
able to hydrolyze cellobiose and cellulose at a lower temperature, which will minimize
energy consumption and undesirable sugar degradation. Compared with cellulase (3-50
kJmol-1
, depending on source of the enzyme) (Paljevac et al., 2007; Xiao et al., 2002),
CP-SO3H had a higher activation energy and therefore needed higher reaction
temperatures. On the other hand, the high temperature allowed the hydrolysis to be
89
completed in a short time, as discussed previously. The lower activation energy of
hydrolysis catalyzed by CP-SO3H was attributed to the ability of the solid acid to adsorb
cellobiose or cellulose, which lowers the energy barrier for the hydrolysis to proceed.
Table 3.3 Apparent activation energy of cellulose hydrolysis catalyzed by different catalysts
Low activation energy and cellobiose/cellulose binding ability give CP-SO3H
high specific catalyst activity. It was reported that sulfonated active carbon (AC-SO3H)
(110 kJmol-1
) could hydrolyze cellobiose by 12% in 2 h with an acid loading of 54%
(molar/molar, based on cellobiose) (Hara et al., 2010), whereas CP-SO3H achieved a
hydrolysis yield of 62% in 2 h with an acid loading of only 0.59% (molar/molar, based on
cellobiose). It is noteworthy that although CP-SO3H performed well in catalyzing the
hydrolysis of cellulose, the density of acid sites on CP-SO3H was relatively low (0.0033
mmol/g). If the density of acid sites could be increased, the cellulose hydrolysis rate
would be further enhanced from the current level.
Catalyst Apparent activation energy (kJ·mol
-1)
Cellobiose Crystalline cellulose
Sulfuric acid 133 170
Carbon-SO3H / 110
CP-SO3H 78 83
Cellulase 3-50 3-50
90
3.4 Conclusion and recommendations
A cellulase-mimetic solid acid catalyst was synthesized by sulfonating
commercial chloromethyl polystyrene resin. Chloride and sulfonic groups on the
resulting solid acid served as cellulose-binding domains and catalytic domains,
respectively. The catalyst was able to hydrolyze cellobiose, starch, and crystalline
cellulose. Because of its substrate-adsorbing ability, the synthesized resin showed
significantly higher catalytic activity than the homogeneous sulfuric acid or other solid
acids at equivalent acid loadings. The low activation energy of the CP-SO3H-catalyzed
reaction allowed the hydrolysis to proceed at moderate temperature. Preliminary results
indicated that the solid acid catalyst had good stability and could be recycled/reused
without activity loss. Future study is needed to increase the density of catalytic domain (-
SO3H) to reduce the catalyst loading. It is also desirable to develop a strategy to
separate/recover the catalyst from hydrolysis residue, in particular when applied to real
biomass where the presence of lignin would make the catalyst recovery more difficult.
For example, incorporating magnetic material in the solid acid catalyst could make the
catalyst magnetically recoverable.
91
Chapter 4: Saccharification of Lignocellulose in
Concentrated Salt Solution
4.1 Introduction
Currently bioethanol is produced from cornstarch or sugarcane. These processes
are not sustainable and are unable to meet the increasing demand for renewable fuels.
Sustainable production of biofuel needs to rely on abundant, inexpensive, and non-food
lignocelluloses. A core bottleneck of the conversion processes based on the sugar
platform is the effective release of sugars from lignocelluloses at low cost and low energy
input (Smith, 2008). Cellulose in lignocelluloses is wrapped by hemicelluloses and
especially lignin, making cellulose far more difficult to hydrolyze to glucose than starch.
In addition cellulose has a crystalline structure. As a consequence, relatively harsh
conditions such as high temperature and high chemical loadings are needed for
hydrolyzing cellulose (Demirbas, 2005). The primary methods that have been extensively
investigated for saccharification of lignocelluloses include concentrated acid, dilute acid,
ionic liquids, and enzymatic processes.
92
Concentrated acid saccharification is an extensively studied cellulose hydrolysis
process. This process is conducted at relatively mild temperature and can lead to a nearly
theoretical yield of sugars. In this process, cellulose in lignocelluloses is first swollen at
room temperature with concentrated acid (typically sulfuric acid), and then the swelled
cellulose is hydrolyzed in dilute acid at elevated temperature (50-120 C) (Miller and
Hester, 2007; Zhu et al., 2009). However, acid corrosion of equipment and the difficulty
of recycling concentrated sulfuric acid have restricted the development of this technology.
Although ion-exclusion chromatography can be used to separate sugars and sulfuric acid,
the method is costly and energy-intensive. In addition, the acid is extensively diluted
during the sugar-acid separation, and of the recovered sulfuric acid has to be
reconcentrated to 70-80% prior to reuse (Cuzens, 1998; Russo, 1999).
In order to avoid the use of concentrated acid, a saccharification method using
dilute acid at higher temperatures (160-190 ºC) was developed. Unfortunately, the dilute
acid process gives a sugar yield of only about 50% due to the incomplete hydrolysis of
cellulose and sugar degradation at high temperature. Additionally, the sugar degradation
products, such as furfural, hydroxymethylfurfural (HMF), and levulinic acid, can inhibit
the downstream fermentation of the sugars, for example, to produce ethanol. In order to
93
reduce the degradation of sugars, in particular of pentoses, a two-stage process was
developed. In the first stage, hemicelluloses were first extracted at moderate temperature;
and in the second stage, the temperature was elevated to hydrolyze the cellulose to
glucose. Even so, the total yield of sugars was only 60-70%, depending on feedstock and
processing conditions (Harris, 1985; Zerbe et al., 1988; Zhu et al., 2009). Further, the
need to use a two-stage process adds complexity and cost.
In summary, problems encountered with acid processes include low sugar yield
due to the incomplete hydrolysis of cellulose and undesirable degradation of the sugars,
formation of fermentation inhibitors (furfural, HMF, and levulinic acid, etc.), extensively
condensed lignin (which limits the coproducts production potential of the lignin),
equipment corrosion, acid recovery issues, and wastewater treatment concerns.
The enzymatic saccharification of lignocellulose using cellulose and
hemicellulose hydrolytic enzymes is another popular method used to break down
cellulose and hemicelluloses into monosaccharides. Enzymatic saccharification itself is
inexpensive and less hazardous than acid hydrolysis because of the use of mild process
conditions (~50 ºC and pH 4-5). However, enzymatic saccharification of lignocelluloses
is economically less attractive which limits its commercialization. A major obstacle is the
94
unavailability of high-activity and low-cost enzymes (both cellulases and hemicellulases).
Although significant progress has been made in recent decades in improving enzyme
activity and reducing enzyme production cost, enzyme is still a considerable contributor
to the high cost of the sugars from lignocellulose platform (Sukumaran et al., 2005;
Zhang et al., 2006). Additionally, because of the natural recalcitrance of lignocelluloses
to the enzymes, enzymatic saccharification of untreated raw biomass is very difficult and
very slow. In order to achieve a satisfactory level of cellulose hydrolysis, an energy-
intensive pretreatment operation is required. Such pretreatment functions to remove
lignin and/or hemicelluloses thereby exposing cellulose. Pretreatment can result in
destruction of the physical matrix by mechanically grinding or milling to reduce particle
size (and thereby increasing accessible surface area to enzymes), enhancing cellulose
hydrolysis by decrystallization and depolymerization, or combinations thereof.
Representative pretreatment technologies include, for example, acid treatment (e.g., with
dilute acid, concentrated phosphoric acid, etc.); the organosolv process (e.g., U.S. patent
3,585,104), ammonia fiber expansion (AFEX), treatment with ionic liquids, treatment
with alkali, and sulfite processes (Balan et al., 2009; Dadi et al., 2006; McIntosh and
Vancov, 2010; Mosier et al., 2005; Pan et al., 2007; Sathitsuksanoh et al., 2010; Wyman
95
et al., 2005; Yang and Wyman, 2008; Yang et al., 2009). However, due to technical
and/or economic barriers, none of these technologies has yet commercially succeeded. In
addition, unlike chemical reactions, enzymatic hydrolysis is a time-consuming process
and typically takes days to complete. Finally, as high consistency (substrate solid content)
hydrolysis is an engineering challenge, enzymatic hydrolysis typically generates a dilute
sugar stream (5-10%, w/w).
Recently, direct hydrolysis of lignocelluloses in ionic liquids has been reported
from pure cellulose to real biomass, such as untreated corn stover, wheat straws, and
wood powder (Ahring and Langvad, 2008; Binder and Raines, 2010; Li and Zhao, 2007).
The use of ionic liquids can be problematic due to the generally higher cost of these
materials, the poor cellulose dissolution ability of ionic liquids in the presence of water,
and the complexity that can be encountered in separation of the ionic liquids from sugars
and the recycling of ionic liquids.
Hydrolysis of cellulose in salt solutions were also reported. U.S. patent 4,018,620
(Penque) relates to a method of hydrolyzing cellulose to monosaccharides by treating
cellulose with 55% calcium chloride in the presence of acid to hydrolyze newsprint
(newspaper). An overall saccharification yield of 50% was reported, but cellulose was
96
only 20% hydrolyzed (Penque, 1977). Because of its capability of swelling and
dissolving cellulose, ZnCl2 was widely used in cellulose solvent systems (Cao et al., 1994;
Fischer et al., 2003). A two-step process was reported to swell cellulose at high ZnCl2
concentration followed by hydrolyzing cellulose to glucose in dilute ZnCl2 solution in the
presence of acid (Chen, 1984). It was reported that over 90% of pure cellulose could be
saccharified to glucose with the process. However, the process was less effective when
applied to real lignocelluloses where an overall saccharification yield of polysaccharides
(cellulose and hemicellulose) was 60-70%, but that of cellulose was only 30-50%.
U.S. patents 4,713,118 and 4,787,939 relate to a process for modification,
solubilization and/or hydrolysis of a glycosidically linked carbohydrate having reducing
groups. The process used a mixture of water, an inorganic acid, and a halide of lithium,
magnesium or calcium. However, these processes used significant amounts of acid (1-10
M), which elevated the corrosion issue. The method to separate sugars and salt was not
mentioned in these patents, which makes the process inapplicable due to the high salt
concentration.
Whereas processes are known in the art for hydrolyzing lignocelluloses there is
still a significant need for efficient and low-cost processes that hydrolyze lignocelluloses,
97
particularly wood-based materials that are hard to hydrolyze, predominantly to
monosaccharides, with minimal losses to undesired coproducts and preferably without
the need for pretreatment. In response to this need, a new cost-effective way to hydrolyze
cellulose into sugars using acidic concentrated salt solutions at moderate temperature was
developed. This process eliminated the use of energy-intensive pretreatment and costly
enzyme. The sugar yield of over 95% was achieved with limited sugar degradation into
furan compounds. Additionally, several methods of separating salt and sugars that could
maximize sugar stream concentrations and further decrease downstream processing cost
were developed.
4.2 Experimental
4.2.1 Materials and Chemicals
Chemical reagents used in this study were purchased from Fisher Scientific
(Pittsburgh, PA) or Sigma-Aldrich (St. Louis, MO) and used as received. Air-dried
lignocelluloses, ground using a Wiley mill to pass a 40-mesh screen, were used in the
present study. Chemical composition of different types of lignocelluloses is presented in
98
Table 4.1. All the chemical reagents and solvents used in this study were purchased from
Fisher Scientific or sigma-Aldrich and used as received.
4.2.2 Liquefaction of lignocellulose in concentrated LiBr solution
LiBr (700, 1000, 1300, 1400, 1500, 1700, 2000 or 2500 mg) was dissolved in 1
mL water in a 15-mL vial, as shown in Figure 4.3. To the solution, 100 mg 40 mesh
spruce powder was added and then vortexed to mix well. Vials were heated at 100-160 ºC
in an oil bath and stirred with a magnetic stir bar for 2 h. After the reaction, the
hydrolysate was filtered and the residue was washed with water. Filtrate and washings
were collected for glucose and HMF or furfural analysis by HPLC.
4.2.3 Hydrolysis of lignocellulose in acidic concentrated LiBr
solution
(1) Optimization of reaction condition
The chemical reagents list in Table 4.5 were added into 5 mL water and stirred to
dissolve well. To this solution, 0.5 g spruce powder was added and stirred to react for
certain time period list in each entry.
99
(2) Batch feeding to enhance biomass loading
For batch feeding, 7.5 g LiBr and 50 mg concentrated HCl were dissolved into 5
mL water and stirred until clear solution was obtained. To this solution, 0.5 g biomass
was added each batch at an interval of 5 min with stirring.
After the addition of feedstock, hydrolysis was allowed to proceed at a total time
(from the first feedstock addition to the end of reaction) listed in the Table 4.6. For
entries 4 to 9, after the sixth feeding, additional 20 mg HCl were added and reactions
were allowed to proceed for 10-30 min before the next addition. The composition of
original poplar and corn stover is shown in Table 4.1.
4.2.4 Extraction of LiBr by organic solvents
Firstly, the extraction condition was optimized, as shown in Table 4.9 and 4.10.
Then real lignocellulose hydrolysate was subjected to the extraction with butanol/hexane
for defined times in 50 mL Falcon tube. In each reaction, after adding organic solvent, the
tube was vortexed for 1 min to complete the extraction and then was centrifuged to
separate the organic phase and water phases. The organic phase was removed and fresh
organic solvent was added to repeat this extraction procedure. Details are shown in the
note under Table 4.9.
Table 4.1 Composition analysis of different feedstocks
Note: ND, not determined.
Composition (%) Spruce Corn Stover Poplar Switch grass Newspaper Printpaper
Moisture 10.00 4.50 9.56 9.86 ND ND
Extractive 5.00 15.50 6.32 16.58 ND ND
Saccharide
Arabinose 0.98 2.71 0.29 1.93 0.35 0.19
Galactose 2.34 1.05 0.53 0.71 0.40 0.08
Glucose 42.03 35.28 43.45 29.66 63.20 71.28
Xylose 5.18 18.40 13.41 18.01 12.13 14.74
Mannose 10.18 0 2.34 0.33 4.86 3.41
Total lignin 26.88 16.78 19.42 15.51 13.08 3.22
100
101
4.2.5 Removal of salt by ion-exchange chromatography
Two small columns (1 cm in diameter and 5 cm in length) were packed, one with
anion exchange resin (DOWEX-2, Cl- form, 100-200 mesh), and one with cation
exchange resin (Amberlyst 15, 25-50 mesh), as shown in Figure 4.13. Prior to operation,
the anion exchange column was washed with 10 mL of 5% sodium hydroxide solution,
and the cation exchange column was washed with 10 mL of 5% sulfuric acid. The
columns were then washed with water until neutral pH. Both columns were de-watered
by injecting air to prevent the water from diluting the sugar solution. The diluted syrup of
sugar concentration of about 50% (0.5 mL) was loaded onto the cation exchange column
first to remove Li+. The loaded sample was pushed through the column by injecting air.
The recovered solution was then loaded onto the anion exchange column to remove Br-.
Again, air was used to push the sample through the column. The recovered solution was
analyzed for LiBr. The whole process scheme is shown below.
(1) Removal of LiBr
Resin-H+ + LiBr Resin-Li
+ + HBr
Resin-OH- + HBr Resin-Br
- + H2O
102
(2) Regeneration of resins
Resin-Li+ + 1/2 H2SO4 Resin-H
+ + ½ Li2SO4
Resin-Br- + 1/2 Ca(OH)2 Resin-OH
- + 1/2 CaBr2
(3) Recovery of residual LiBr
1/2 Li2SO4 + 1/2 CaBr2 LiBr + 1/2 CaSO4
4.2.6 Quantification of sugars and sugar derivatives
Sugar analysis was conducted using a Dionex HPIC system (ICS-3000) equipped
with integrated amperometric detector and Carbopac™ PA1 guard and analytical
columns at 20 ºC. Eluent was provided at a rate of 0.7 mL/min, according to the
following gradient: 0-25 min, 100% water; 25.1~30 min, 30% water and 70% 0.1 M
NaOH; 30.1-35 min, 100% water. To provide a stable baseline and detector sensitivity,
0.5 M NaOH at a rate of 0.3 mL/min was used as post-column eluent.
5-hydroxylmethylfurural (HMF) and furfural were analyzed using the Dionex
ICS-3000 equipped with a Supelcogel C-610H column at temperature 30 ºC and UV
detector at 210 nm. Eluent was 0.1% phosphoric acid at a rate of 0.7 mL/min.
103
4.2.7 Determination of LiBr amount
The pH value of the solution to be determined was adjusted by 2 M NaOH to 7-9
before titration. The amount of LiBr was determined by the Mohr method (Doughty,
1924). Content of Br- and Cl
- was determined by titration of AgNO3 with KCrO4 as
indicator. The mechanism of titration is shown below.
Br- + Ag
+ AgBr (yellow precipitant)
Cl- + Ag
+ AgCl (white precipitant)
Ag+ + CrO4
- AgCrO4 (red precipitant)
4.3 Result and discussion
4.3.1 Description of whole process
As schematically shown in Figure 4.1, real biomass (such as corn stover,
switchgrass, hardwood, and softwood) was mixed with LiBr, a small amount of acid, and
water at a desired LiBr concentration (e.g., 40-70%). The mixture was heated at elevated
temperature (100-160 ºC) for typically several minutes to several hours (dependent on
acid concentration, salt concentration, biomass species and particle size) with stirring.
104
Both cellulose and hemicelluloses of the biomass were completely hydrolyzed
(saccharified), whereas lignin (up to 30% of the biomass) remained as an insoluble
residue. By filtration or centrifugation, lignin was separated from the solution of sugars
and LiBr. The sugars from cellulose and hemicelluloses and LiBr could be further
separated by ion exchange, extraction, or crystallization methods based on their
differences in ionization and solubility in water and organic solvents. The recovered LiBr
can be reused in the process, and the sugars can be converted into biofuels and chemicals
by biological or chemical approaches.
Figure 4.1 Process flow chart of biomass saccharification in concentrated salt solution.
105
The fast chemical saccharification method was able to simultaneously hydrolyze
cellulose and hemicelluloses without the need for extensive size reduction or any
chemical pretreatment of the biomass, as illustrated in Figure 4.2,. The process could
produce a concentrated sugar solution (>30%, w/w) in high yield with limited
degradation of sugars to furans (furfural and hydroxymethylfurfural). Sugars produced
were predominantly monosaccharides. In addition, the process could directly handle
small sized wood chips without requiring extensive size reduction or any other
pretreatment, which significantly simplified operation and reduced processing cost and
energy consumption. This process will be discussed in detail below.
Figure 4.2 Hydrolysis way of biomass in concentrated salt solution.
4.3.2 Liquefaction of lignocellulose in concentrated LiBr solution
We initially studied the liquefaction of various types of lignocelluloses in a series
of metal halide salts and found that concentrated LiCl, LiBr and CaBr2 solutions (60%,
106
w/w) had very good capabilities to liquefy lignocellulose through the dissolution of
cellulose and the partial hydrolysis of cellulose into oligosaccharides and
monosaccharides. In this thesis, the focus is on the applicability of concentrated LiBr as a
solvent for hydrolysis of lignocellulose. Experimental results show that lignocellulose
could be liquefied/hydrolyzed in concentrated LiBr solution to release saccharides (both
oligosaccharides and monosaccharides) at moderate temperatures (120-160 C), as shown
in Tables 4.2, 4.3 and 4.4. The effect of temperature and LiBr concentration on the
hydrolysis of spruce powder in LiBr solution is summarized in Figure 4.3.
Table 4.2 LiBr-hydrolysis of various types of feedstock
Note: reaction condition: 1 mL H2O, 1500 mg LiBr, 100 mg lignocelluloses powder, 140 ºC, and 2 h.
Pure cellulose such as microcrystalline cellulose (e.g., Avicel® microcrystalline
cellulose, FMC Biopolymer) and dissolving pulp (bleached wood pulp typically
having >95% cellulose) were tested for solubilization and hydrolysis in concentrated
LiBr. The results showed that the pure cellulose dissolved faster (5-30 min) under the
Feedstock Lignin content in biomass (%) Residue left (%)
Switchgrass 16 21
Spruce 27 25.7
Poplar 20 22
Corn stover 17 22.5
107
same conditions (60% LiBr solution at 140 ºC) than real biomass, presumably because of
the absence of blocking from hemicellulose and lignin. It was found that the molecular
weight of cellulose slightly decreased during the dissolution in concentrated LiBr
solution, indicating that some cellulose hydrolysis occurred during the dissolution
process. Increasing temperature or extending time elevated the degree of cellulose
hydrolysis (or depolymerization) as was expected.
Figure 4.3 Hydrolysis of spruce powder in LiBr solution (no acid) as a function of LiBr
concentration at different conditions.
Note: experimental procedures: Specific amounts of LiBr (700, 1000, 1300, 1400, 1500, 1700, 2000 or
2500 mg) were dissolved in 1 mL water in a 15-mL vial. To the solution, 100 mg 40 mesh spruce powder
was added and then vortexed to mix well. Vials were heated at 100-160 ºC in an oil bath and stirred using a
magnetic stir bar to react for 2 h. After the completion of the reaction, hydrolysate was filtered and the
residue was washed with water. Filtrate and washings were collected for glucose and HMF or furfural
analysis by HPLC.
0
25
50
75
100
40 50 60 70 80
Co
ve
rsio
n(m
ola
r%)
LiBr concentration(w%)
Saccharides, 120℃, 2h
Saccharides, 140℃, 2h
Saccharides, 160℃, 2h
HMF, 120℃, 2h
HMF, 140℃, 2h
HMF, 160℃, 2h
LiBr.4H2O
LiBr.3H2O
108
Table 4.3 Monosaccharides in the hydrolysates from LiBr hydrolysis of different feedstocks
without acid
Note: reaction condition: 1 mL H2O, 1500 mg LiBr, 100 mg lignocelluloses powder, 140 ºC, and 2 h.
In light of the excellent dissolution/hydrolysis ability of 60% LiBr solution,
hydrolysis of different feedstocks in this solution at 140 ºC was investigated. The
comparison of lignin content in original biomass and the amount of residue after
hydrolysis is shown in Table 4.2. The results indicate that almost all of the carbohydrates
were dissolved. The carbohydrate composition in the hydrolysates before and after
autoclaving at 120 ºC with 3% dilute acid for one hour was shown in Tables 4.3 and 4.4.
Autoclaving oligosaccharides at 120 ºC with 3% acid can completely hydrolyze
oligosaccharides into monosaccharides. Comparing the results in Tables 4.3 and 4.4, one
can deduce that autoclaved hydrolysates contained much more monosaccharides than
original hydrolysates, implying that most of the carbohydrates dissolved from
lignocelluloses in concentrated LiBr solution without acid existed in the form of water-
soluble oligosaccharides. The hydrolysis of cellulose into oligosaccharides in
Sugar type Corn stover (%) Switchgrass (%) Poplar (%) Spruce (%)
Arabinose 0.66 0.86 0.25 0.59
Galactose 0.39 0.55 0.39 1.31
Glucose 2.079 5.11 8.08 9.00
Xylose 7.059 9.42 7.37 2.00
Mannose 0 0 2.80 6.04
109
concentrated LiBr solution without acid is predicted to be caused by the Lewis acidity of
Li+, which will be further discussed later.
Table 4.4 Monosaccharides in hydrolysate from LiBr hydrolysis of different feedstocks after
autoclaving at 120 °C for 1 h
Note: autoclave procedure: 1 mL hydrolysate was mixed with 1 mL 6% sulfuric acid solution into 2 mL
hydrolysate with 3% acid concentration. This solution was autoclaved at 120 °C for 1 h to hydrolyze
oligosaccharides into monosaccharides.
4.3.3 Dissolution mechanism of cellulose in concentrated LiBr
solution
The conversion of cellulose in spruce powder starts to increase readily when the
concentration of LiBr was larger than 55% at the reaction temperature of 140 ºC. To
explain this phenomenon, mechanism for cellulose dissolution in concentrated LiBr
solution is proposed in Figure 4.4. It is interestingly found that 55% and 60% LiBr
solutions roughly correspond to LiBr·4H2O and LiBr·3H2O, respectively. It can be seen
from Figure 4.3 that either increasing one coordination water or decreasing the reaction
temperature down to 120 ºC caused significantly decreased conversion of cellulose. It is
Sugar type Corn stover (%) Switchgrass (%) Poplar (%) Spruce (%)
Arabinose 0.73 1.07 0.21 0.67
Galactose 0.55 1.09 0.57 2.70
Glucose 30.84 26.43 39.53 40.32
Xylose 16.26 14.06 9.63 3.34
Mannose 0 0 4.72 7.56
110
understandable that either addition of water or decreasing temperature could cause an
increase of coordinated water around ions. Therefore, the number of coordination water
around Li+ and Br
- was thought to play a key role in affecting the cellulose dissolution
ability of LiBr solutions. A lot of research on the coordination number of Li+ in aqueous
solution has been reported and the coordination number of 4 or 6 is disputable (Wolfram
Rudolph, 1995). The coordination number 6 can probably explain our observations. In
the case of LiBr·3H2O (61% LiBr), as shown in Figure 4.4a, six H2O molecules form a
hexahedral structure with Li+ through Li···O bond, and each H2O molecule is shared by
two Li+ ions. It seems that all Li
+ ions are saturated with water molecules in LiBr·3H2O.
However, as the solvation of Br- likely occupies some of the water molecules, some of
the Li+ ions may not be completely solvated. In other words, Li
+ has potential water-
deficit sites. In particular, when the system is heated, Li···O bond could be broken to
form more water-deficit sites. The Li+ with water-deficit spots can coordinate the O of
hydroxyl groups in cellulose, and meanwhile, the ―naked‖ Br- ions in the solution tend to
form hydrogen bonds with hydrogen of hydroxyl groups. As a result, the Li···O and
Br···H bonds replace the inter- and intra-molecular O···H bonds in crystalline cellulose,
and therefore disrupt cellulose crystal structure and enhance the dissolution of cellulose.
111
The situation is very different in the case of LiBr·4H2O (55% concentration). As shown
in Figure 4.4b, because of the presence of an extra water molecule, Li+ ions would be
more solvated than those in LiBr·3H2O. Similarly, Br- ions are more solvated than those
in LiBr·3H2O. Therefore, the dissolution ability of LiBr·4H2O is significantly decreased.
Besides, empty Li+ coordination sites can act as Lewis acid to hydrolyze glycosidic bonds.
Because of the hydrolysis ability of Li+, dissolved polysaccharides can be further broken
down into small units and dissolved away from the lignin.
(a)
(b)
Figure 4.4 Models for cellulose dissolution in salt solutions of varied concentrations. (a)
LiBr·3H2O (61% LiBr concentration) and (b) LiBr·4H2O (55% LiBr concentration).
112
In the experiment, a temperature of 140 ºC (sufficient to overcome the activation
energy) was thought to be the minimum temperature to break Li···O bonds to create
water-deficit sites. Temperatures lower than 120 ºC, as tested in our experiment, could
not provide enough energy to break Li···O bond and create empty sites for dissolution
and hydrolysis of cellulose, so consistently neither dissolution nor hydrolysis was
observed in our experiment. The decreased conversion of sugars observed at higher
temperatures (160 ºC) and higher LiBr concentration (>60%) may be attributed to the
significantly increased Lewis acid sites, which leads to significant condensation of lignin
and degradation of sugars into humins, as indicated by the recovered black residues,
rather than brown residues (lignin). As shown in Figure 4.3, concentrations of LiBr
between 55% and 60% were ideal to dissolve cellulose, as is consistent with the proposed
mechanism. In this range of concentrations, LiBr·4H2O and LiBr·3H2O structures coexist
in solution. With increasing of concentration, LiBr·3H2O increased and thereby
dissolution was enhanced accordingly. The conversion of ~20% in LiBr concentration
lower than 55% is attributable to the dissolution of amorphous cellulose faction in the
spruce powder. In summary, 60% LiBr solution at 140 ºC is the optimum condition to
dissolve/hydrolyze the polysaccharides in spruce.
113
4.3.4 Hydrolysis of lignocellulose in acidic concentrated LiBr
solution
The experimental results in Figure 4.3 indicates that concentrated LiBr was
capable of dissolving and hydrolyzing the polysaccharides in lignocellulose, but the
hydrolysis was incomplete and slow, and most of the hydrolysis products were in the
form of oligosaccharides (Tables 4.2, 4.3, and 4.4). To accelerate the hydrolysis of
polysaccharides, acid was added to the LiBr solution. The process conditions were further
optimized within the range of LiBr concentrations of 44-60% (w/w), temperatures of
110-160 ºC, acid loading of 1-10% (on spruce powder), and reaction time of 5-60 min, as
summarized in Table 4.5. Entries 1 to 5 indicate that the LiBr concentration is vital to the
hydrolysis of cellulose, which is consistent with the conclusion derived from Figure 4.3.
When LiBr concentration was equal to or above 55% (for example, 60, 56, and 55% in
entries 1, 2, and 3), the conversion/hydrolysis of polysaccharides was complete. Lower
LiBr concentration (for example, 50 and 44% in entries 4 and 5) led to a decreased
conversion of polysaccharides. It should be noted that the degradation of sugars to
furfural and HMF was lower at low LiBr concentrations. Although acid addition and high
temperature can enhance the hydrolysis, LiBr concentration remains the most critical
114
parameter. When LiBr concentration was lower than 55%, increasing temperature or acid
loading was unable to achieve a satisfactory rate of hydrolysis in decent time, as shown in
the results of entries 4, 5, and 8.
Acid addition and high temperature can enhance the hydrolysis of biomass in
LiBr solution. For example, compared to the results in Figure 4.3, entries 1 and 6 in Table
4.5 show that addition of 2% hydrochloric acid (w/w on spruce powder) could completely
hydrolyze spruce in 60% LiBr solution in less than 10 min. As shown in entries 10 to 12,
satisfactory conversion could be achieved at high temperatures even with low acid
loading (1% on biomass) in 56% LiBr solution. From entries 1 to 14, it can be seen that
100% conversion/dissolution of polysaccharides was achieved in some entries, but the
yield of sugars was lower than theoretical, because either the polysaccharides were not
completely hydrolyzed to monosaccharides (partially still in the form of oligosaccharides)
or part of the sugars were further degraded to furfural and HMF. It is noteworthy that
extending the reaction time or elevating the temperature generally generated more HMF
and furfural. It is not wise to apply higher concentration of acid and higher temperature
simultaneously. At high LiBr concentration, for example 60%, two options are available
to achieve both high conversion and high sugar yield with a low yield of furan
115
compounds in a short time: (1) high temperature (150-160 ºC) with low acid loading
(1%), such as in entries 10 and 11; (2) high acid loading (5-10% on biomass) at low
temperature (110 ºC), such as in entries 16 and 17. As shown in Table 4.5, up to 96%
sugar yield was obtained under the conditions used in entries 16 and 17 with very limited
degradation of sugars. As low temperature saves energy, the conditions of entries 16 and
17 will be used in our following research.
Table 4.5. Hydrolysis of spruce powder under various conditions
Note: ―Tem.‖ denotes temperature; ―Ara.‖ denotes arabinose; ―Gal.‖ denotes galactose; ―Glu.‖ denotes glucose; ―Xyl.‖ denotes xylose; ―Man.‖ denotes
mannose; ―HMF.‖ denotes hydroxymethylfurfural; ―Fur.‖ denotes furfural; ―Res‖ denote residues (lignin and unhydrolyzed polysaccharides) after reaction;
―Con.‖ denotes conversion; ―TSY‖ denotes total monosaccharide yield; ―TFY‖ denotes total furan compounds (HMF and furfural) yield; Reaction procedure:
the chemical reagents list in each entry in Table 5 were added into 5 mL water and stirred to dissolve well. To this solution, 0.5 g spruce powder was added
and stirred to react for certain time period list in each entry. ―*‖ the weight of pure hydrogen chloride. ―Ara.‖, ―Gal.‖, ―Glu.‖, ―Xyl.‖, ―Man.‖, ―HMF.‖
―Fur.‖, and ―Res‖ (%) are weight percentage based on original feedstocks. ―con‖, ―TSY‖, and ―TFY‖(%) are molar percentage based on total sugar content
in feedstocks.
Entry LiBr
(g)
HCl*
(mg)
Tem.
(°C)
Time
(min)
Ara.
(%)
Gal.
(%)
Glu.
(%)
Xyl.
(%)
Man.
(%)
HMF
(%)
Fur.
(%)
Res.
(%)
Con.
(%)
TSY
(%)
TFY
(%)
0 / / / / 0.98 2.34 42.03 5.18 10.18 / / 26.88 / / /
1 7.50 10 140 10 0.80 1.79 30.51 2.55 6.29 5.29 2.45 24.85 100 67 12
2 6.25 10 140 10 0.80 1.81 30.22 2.44 6.01 6.00 3.25 25.65 100 66 15
3 6.00 10 140 10 0.83 1.84 30.96 2.84 6.39 4.50 2.53 26.12 100 69 11
4 5.00 10 140 10 1.05 2.17 13.01 3.81 7.49 1.86 2.27 42.31 76 44 7
5 4.00 50 140 20 0.18 0.90 13.98 0 1.95 1.62 1.99 49.63 64 27 6
6 7.50 10 140 5 0.89 1.95 34.43 3.73 7.05 3.2 2.02 24.78 100 77 8
7 6.00 50 120 10 0.66 1.72 27.61 1.64 4.79 4.05 2.78 32.31 91 58 11
8 5.00 50 120 10 0.96 2.30 18.26 2.95 7.39 1.69 3.04 36.54 85 51 8
9 7.50 10 120 25 0.81 1.82 31.69 3.00 6.32 3.28 2.24 30.12 95 70 9
10 6.25 5 160 5 1.01 2.08 36.68 4.41 8.05 2.98 1.58 22.65 100 84 7
11 6.25 5 150 5 1.03 2.05 35.20 4.68 8.21 1.65 0.88 23.21 100 82 4
12 6.25 5 140 5 1.10 1.99 27.76 4.94 7.72 0.66 0.57 23.82 100 70 2
13 7.50 10 100 60 0.98 2.02 36.59 4.17 8.48 1.31 1.01 26.32 100 84 4
14 7.50 50 100 10 0.90 1.96 34.23 3.69 7.71 1.68 1.46 25.65 100 78 5
15 5.00 25 140 15 0.32 1.17 17.52 0.06 2.41 2.68 3.06 41.01 78 34 9
16 7.50 50 110 5 1.16 2.36 43.38 5.06 9.97 1.41 1.15 27.12 100 99 4
17 7.50 25 110 5 1.16 2.38 42.46 5.18 9.71 0.98 0.86 26.93 100 98 3
116
3
117
4.3.5 Hydrolysis of lignocellulose through batch feeding
In order to increase production efficiency and decrease LiBr/sugars separation cost
in downstream processing, a high feedstock loading (a low ratio of LiBr solution to
feedstock solid) is desirable, as it will lead to high concentration final sugar solution.
However, addition of too much feedstock caused mixing and mass transfer problems and
therefore affected the hydrolysis efficiency and yield. In order to overcome these problems,
feedstock was fed in multiple steps to ensure that enough liquid is available to wet
biomass and hydrolyze the cellulose at all times. In our experiment, biomass feedstock
was fed into the reactor as soon as the last batch was closed to completely liquefied and
hydrolyzed. From entries 1 to 5 in Table 4.6, 1, 2, 3, 4 and 5 g spruce powder were added
into a 60% LiBr solution (7.5 g LiBr + 5 mL water) (feeding procedure is described in the
notes under Table 4.6). Results showed that when total biomass feeding was below 4 g
(entries 1 to 4), almost all of the polysaccharides in spruce powder was hydrolyzed into
monosaccharides with high selectivity. Figure 4.5 shows the picture of the separated
hydrolysate after centrifugation.
Table 4.6 Hydrolysis of spruce powder in batch-feed mode
Entry LiBr
(g)
HCl*
(mg)
Tem.
(°C)
Bio.
(g)
Time
(min)
Ara.
(%)
Gal.
(%)
Glu.
(%)
Xyl.
(%)
Man.
(%)
HMF
(%)
Fur.
(%)
Res.
(%)
Con.
(%)
TSY
(%)
TFY
(%)
0 / / / / / 0.98 2.34 42.03 5.18 10.18 / / 26.88
1 7.5 50 110 1 20 1.14 2.31 42.89 5.11 10.02 1.22 0.88 26.92 100 98 3
2 7.5 50 110 2 40 1.07 2.36 42.22 5.13 9.69 1.12 1.03 26.34 100 97 3
3 7.5 50 110 3 60 0.98 2.03 41.30 5.04 9.78 1.06 1.25 25.78 100 95 4
4 7.5 70 110 4 80 1.21 2.11 41.75 5.09 9.82 0.94 1.31 25.19 100 95 4
5 7.5 70 110 5 100 0.94 2.05 41.93 5.11 9.78 0.97 1.25 27.07 92 86 3
6 7.5 50 110 4 80 0.24 0.45 42.81 12.24 2.27 0.87 1.01 19.11 100 99 2
7 7.5 50 110 4 80 2.01 0.86 35.22 17.34 0 0.67 1.55 16.24 100 99 5
8 7.5 50 110 4 80 0.25 0.44 39.32 11.35 2.06 0.77 1.23 25.25 90 90 4
9 7.5 50 110 4 80 2.12 0.99 32.21 16.12 0 0.69 1.61 23.27 91 89 4
Note: Entries 1 to 5: spruce powder; entry 6: poplar powder; entry 7: corn stover powder; entry 8: poplar chips (0.2-0.5 cm); entry 9: corn stover chips (0.2-0.5
cm). ―Tem.‖ denotes temperature; ―Bio‖ denotes biomass; ―Ara.‖ denotes arabinose; ―Gal.‖ denotes galactose; ―Glu.‖ denotes glucose; ―Xyl.‖ denotes xylose;
―Man.‖ denotes mannose; ―HMF.‖ denotes hydroxymethylfurfural; ―Fur.‖ denotes furfural; ―Res‖ denote residues (lignin and unhydrolyzed polysaccharide)
after reaction; ―Con.‖ denotes conversion; ―TSY.‖ denotes total monosaccharides yield; ―TFY.‖ denotes total furan compounds (HMF and furfural) yield. ―*‖
the weight of pure hydrogen chloride. Reaction procedure: 7.5 g LiBr and 50 mg concentrated HCl were dissolved into 5 mL water and stirred until clear
solution was obtained. To this solution, 0.5 g biomass was added each batch at an interval of 5 min with stirring. After the addition of feedstock, hydrolysis
was allowed to proceed at a total time (from the first feedstock addition to the end of reaction) listed in the Table. For entries 4 to 8, after the sixth feeding,
additional 20 mg HCl were added and reactions were allowed to proceed for 10-30 min before next addition. The composition of original poplar and corn
stover is shown in Table 1. ―Ara.‖, ―Gal.‖, ―Glu.‖, ―Xyl.‖, ―Man.‖, ―HMF.‖ ―Fur.‖, and ―Res‖ (%) are weight percentage based on original feedstocks. ―con‖,
―TSY‖, and ―TFY‖(%) are molar percentage based on total sugar content in feedstocks.
118
119
However, when 5 g spruce powder was fed, the conversion decreased compared
to previous lower feeding because the high solid loading and the accumulated lignin
residue in the hydrolysate made the stirring and mass transport difficult. If efficient
mixing is provided, the upper limit of total biomass loading may be increased, and the
hydrolysis yield at high feedstock loading may be improved.
Figure 4.5 The picture of the separated hydrolysate.
In entries 6 and 7, polysaccharides in poplar and corn stover powder were also
completely hydrolyzed into monosaccharides using the same conditions and procedure
for as entry 4. Because the hydrolysis was a heterogeneous reaction process, biomass
120
size had a significant influence on the hydrolysis rate. For example, when poplar and
corn stover chips particle size of 2-5 mm was used, they were still hydrolyzed but
hydrolysis yield decreased to around 90%, as shown in entries 8 and 9; this was lower
than those in entries 6 and 7 where 40 mesh biomass powder was used. Extension of
reaction time or efficient mixing could possibly achieve a complete hydrolysis.
4.3.6 Saccharification of lignocellulose in concentrated solution
of different salts
Hydrolysis of biomass in different metal halide solutions was investgiated.
Results in Table 4.7 show that when the concentration of salt was 60%, LiCl, LiBr and
CaBr2 could give complete conversion/hydrolysis of polysaccharides. CaBr2 is as
effective as LiBr in this experiment. In addition, experiments with separation methods
discussed below indicate that the methods for separating LiBr from sugars, also work
with CaBr2. Furthermore, CaBr2 is currently much cheaper than LiBr and as such is
currently preferred. Although LiCl also gave a complete conversion, the solubility of
LiCl in water is significantly lower than that of LiBr. For example, the solubility of LiBr
in water at 90 ºC is about 254 g/100 mL, whereas that of LiCl at 100 ºC is about 128
g/100 mL. As a consequence, 60% LiCl starts crystalizing out at 120 ºC, which makes
121
downstream separation operations difficult. ZnCl2 and CaCl2, consistant with the reports
by Penque and Chen, only gave a conversions of 50~70% (Chen, 1984; Penque, 1977).
Table 4.7 Effect of different salts on hydrlysis of biomass
Note: reaction condition: 1 mL H2O, 1500 mg salt, 100 mg spruce powder, 2 mg HCl, 140 ºC, and 30 min.
It is also instetrestingly found that LiCl concentration of 60% has significantly
higher molar concentration than LiBr concentration of 60%, indicating that the structure
of LiCl and LiBr in aqueous soltuion are different. As we discussed previously,
LiBr·3H2O (60% LiBr concentration) could dissolve cellulose. This is not the case for
LiCl·3H2O (44% LiCl concentration) which could not dissolve cellulose. The possible
explanation is that they have different struture caused by different properties of anions
(Cl- and Br
-). Because Cl
- has higher charge density than Br
- , it can form stronger ionic
bonds with Li+. Additionally, Cl
- and water has comparable affinity to Li
+ and can
compete for coordination sites, which lead to some coordination sites occupied by Cl-. A
Salt Conversion (%)
LiBr 100
LiCl 100
CaBr2 100
ZnBr2 66
ZnCl2 60
MgBr2 75
CaCl2 70
NaBr 68
122
possible struture is shown in Figure 4.6. As we disussed previously, applying heat could
break Li···O bonds to create water-deficit sites. However, in LiCl·3H2O, the situation is
same to that of LiBr·4H2O where the created empty site will be occupied by water first.
Therefore, LiCl·3H2O has very poor dissolution ability. Considering LiCl·2H2O, the
empty site could be occuped by a cellulose hydroxyl group. However, it is observed that
under a LiCl concentration of 54% (LiCl·2H2O), the dissoluton is also very poor even
though sinificant swelling was observed. This might by attribted to the inactive Cl-
which was restricted by Li+ and has decreased activity. Dissoluton of cellulose could be
completed at a LiCl conentration of 60% which is supposed to contain both LiCl·2H2O
and LiCl·H2O and have much more empty sites for cellulose hydroxyl groups, as shown
below. The structure of LiCl·3H2O ,LiCl·2H2O, and LiCl·H2O are shown below in
Figure 4.6. The proposed structure of LiCl·2H2O is consistent with reported crystal
structure of LiCl·2H2O (Brendler et al., 2002). Therefore, we think that the dissolution of
cellulose in 60% LiCl solution should be mostly attributed to rich empty coordination
sites on Li+.
123
LiBr·3H2O LiCl·3H2O LiCl·2H2O LiCl·H2O
Figure 4.6 The proposed structures of LiCl hydrates.
Why concentrated solutions of LiBr, LiCl and CaBr2 are more efficient at
hydrolyzing cellulose than other metal halides. Research has concluded that the
following characteristics mainly determine the dissolution power towards cellulose: (1)
the acidity, (2) the water content of the melts, and (3) the properties of the coordination
sphere of the cations. However, there is no report to explain how these factors affect
dissolution of cellulose. Based on the hypothesis discussed above, these factors can have
a reasonable explanation. Higher acidity metal ions have a higher ability to coordinate
with cellulose hydroxyl group and hydrolyze cellulose. Lower water content can lead to
more empty coordination sites for coordination cellulose hydroxyl group and for
hydrolysis of polysaccharides. Anion will significantly affect the coordination structure
of cation. High charge density, such as Cl-, could allow anion to form strong cation-
anion association and water-anion interactions, leading to inactive anions. Besides, the
124
relaxation time (desolvation kinetics) of salvation shell might also significantly affect the
binding of other ligands such as hydroxyl group of cellulose. Fast and frequent
desolvation of metal ion can automatically create empty coordination sites for cellulose
hydroxyl group. For example, the lifetime of Ca2+
-water coordination is about 18 ps, 10
times longer than the relaxation time previously reported for K+ or Na
+. Even though
Mg2+
is slightly smaller than Ca2+
, the lifetime of water molecules around Mg2+
is on the
order of a few hundreds of picoseconds (Jiao, 2006). This might explain why calcium
halides are better than Magnesium halides. Different salts have different structures
because of the properties of cations and anions and these factors discussed above work
together to determine the dissolution ability of salts.
4.3.7 Hydrolysis of lignocellulose in concentrated LiBr solution
with different acids
It was already found that acid could significantly enhance the hydrolysis of
lignocelluloses in concentrated LiBr solution, as disscused above. Here different acids
were tested for their hydrolysis efficacy on spruce powder in 60% LiBr solution. It can
be seen from Table 4.8 that strong acids such as sulfuric acid and nitric acid are as
effective as hydrochloric acid in catalyzing the hydrolysis of spruce, while weak acids
125
gave a decreased hydrolysis yield. For example, acetic acid had a yield of 90% at the
same conditions.
Table 4.8 Effect of different acids on hydrlysis of biomass
Note: reaction condition: 1 mL H2O, 1500 mg LiBr, 100 mg spruce powder, 2 mg acid, 140 ºC, 30 min.
4.3.8 Separation of LiBr and sugars by different methods
The chemical saccharification process presented here uses concentrated LiBr
solution as a medium. It is obvious that effective separation of sugars and LiBr after the
saccharification is critical to the industrial application of the process. Fortunately, LiBr is
quite different from sugars in physical and chemical properties, which can be used to
separate LiBr from sugars. Three methods including boronic acid extraction, ion
exclusion chromatography, and solvent extraction, were used to separate LiBr from
sugars in this study.
4.3.8.1 Extraction of sugar from LiBr-sugar solution with boronic acid
Boronic acid can react with cis-diol in sugars in the presence of lipophilic
Acid Conversion (%)
HCl 100
H2SO4 100
HNO3 100
H3PO4 95
HCOOH 96
CH3COOH 90
126
quaternary alkyl amines to form a stable boronic acid-sugar complex under alkaline
conditions. The complex is soluble in organic solvent and can be extracted with hexane,
by which sugars can be separated from LiBr. The mechanism of this method is shown in
Figure 4.7. The boronic acid-sugar complex is stable under alkaline conditions but
unstable under acidic condition. Therefore, the sugar extracted to the organic phase can
be recovered by stripping the organic phase with acidic water. The boronic acid remains
in the organic phase and can be used in next batch extraction.
Figure 4.7 Mechanism for extraction of glucose with boronic acid.
Note: Naphthalene-2-boronic acid (N2B) was used in this study. A glucose-LiBr solution (100 mg glucose
and 500 mg LiBr in 1 mL buffer solution at pH = 11) was tested. The organic phase was prepared by
dissolving 100 mg N2B and 200 mg Aliquat 336 into 5 mL hexane/octanol (85:15, v/v). The sugar-LiBr
solution (1mL) was mixed with the organic solution (5 mL) and vortexed for 30 min to reach equilibrium.
The mixture was then centrifuged to separate the aqueous phase and organic phase. The organic phase was
then stripped with 5 mL 1% HCl solution. Analysis of the stripping solution indicated that approximately
10 mg sugar was recovered through the single pass of extraction.
O
HOHO
OH
B
HO
HOB
HO
HO
OH
B
OO
OH
OH- glucose
glucose
aqueous phase
organic phase
OHO
HO
OH
OHOH
aqueous phase
organic phase
OHO
HO
OH
B
OO OH
organic phase
NH4
Alkyl
H+
OHO
HO
OH
B
OO
organic phase
B
HO
HO
LiBr
NH4
Alkyl
LiBr
H2O
NH4
Alkyl
LiBr
organic phase
not stable stable
127
The preliminary results indicate that the boronic acid method was able to separate
LiBr and sugars, but the selectivity and efficiency were not satisfactory. In addition, the
method was expensive and had less potential in industrial applications.
4.3.8.2 Separation of LiBr and sugars with ion-exclusion chromatography
Ion-exclusion chromatography has been applied to separate salts and sugars. The
process does not consume acid or alkaline as ion exchange chromatography does. Ion-
exclusion chromatography uses the resin containing the same anion or cation as that in a
salt as its stationary phase. Because of the exclusion force between the resin and the salt,
and penetration of sugars into the micropores of the resin, the salt will elute faster than
sugars. The mechanism is shown in Figure 2.8.
The experimental procedure was shown in the note under Figure 4.9. A synthetic
LiBr-sugars solution (10 mg arabinose, 10 mg galactose, 80 mg glucose, 20 mg xylose,
20 mg mannose, 50 mg LiBr, and 0.5 mL water) was loaded to a column and eluted with
de-ionized water at a rate of 1.5 mL/min. The profile in Figure 4.9 shows that LiBr and
sugars separated very well, indicating that this method is an efficient way of separating
sugars and LiBr.
128
Figure 4.8 Schematic diagram for ion-exclusion chromatography for separating sugar and salt.
The hydrolysate of spruce was tested with ion-exclusion method as well. When
ethanol is removed/recovered by evaporation, the hydrolysate from section 4.3.9.2 (~80
mg glucose, ~15 mg xylose, ~60 mg mannose, and ~50 mg LiBr) was diluted with water
into a solution of 0.5 mL. This solution was loaded on the column and eluted with de-
ionized water at a rate of 1.5 mL/min. The result in Figure 4.10 shows that LiBr and
sugars could be separated very well, and the recovered sugar and LiBr concentrations
were ~1% and ~1%, respectively.
129
Figure 4.9 Separation of LiBr from sugar solution using ion-exclusion chromatography.
Note:Separation test was performed on a glass column (diameter: 2 cm; length: 50 cm), packed with
anion exchange resin (DOWEX 18-400, Cl- form) at room temperature. Prior to test, the column was
fully converted into Br- form by eluenting with 400 mL of 0.2 N NaBr at a flow rate of 1.5 mL/min,
followed by a thorough rinse with 2000 mL of deionized water, after which it was ready for sugar
separation test. Sugar solution was injected at the top of the column. The sample was then eluted with
deionized water at a flow rate of 1.5 mL/min. AgNO3 solution was used to monitor the elution out of LiBr,
after which 10 fractions of 2.5 mL were collected from the exit of the column, and their sugar profiles
were determined off-line with a HPIC. LiBr amount in each fraction was determined through titration.
The raw hydrolysate from spruce saccharification with LiBr without any
extraction was tested as well. It was possible to separate concentrated LiBr and sugars
directly using ion-exclusion chromatography. However, because of the high
concentration of LiBr, longer column and more water were needed for a good separation;
thus the significant dilution of LiBr and sugars is the major drawback of the method.
15 16 17 18 19 20 21 22 23 24 25 26
Collected fractions
LiBr glucose xylose mannose arabinose galactose
130
Figure 4.10 Separation of residual LiBr from sugar solution using ion-exclusion chromatography.
(~80 mg glucose, ~15 mg xylose, ~ 60 mg mannose, and ~50 mg LiBr),
4.3.8.3 Extraction of LiBr from LiBr-sugar solution by organic solvents
Direct separation of LiBr and sugars in an aqueous solution is very difficult
because they are both highly soluble in water. However, LiBr and sugars have very
different solubility in organic solvent (such as alcohol, ketone and ether) where LiBr is
still highly soluble but sugars are not, making it possible to separate LiBr from sugars by
extraction with water-immiscible organic solvents. Extraction of LiBr from brines by
organic solvent such as TBP (tributylphoshate) (Gou and Zhu, 1998; Huang, 1991) and
15 16 17 18 19 20 21 22 23 24 25 26
Collected fractions
LiBr
Glucose
Xylose
Mannose
131
butanol (Uhlemanna, 1993) and CaBr2 from brines by amine (Grinbaum et al., 2008) has
been reported.
In the present study, n-butanol was chosen as an extraction solvent for LiBr
because of its relatively low price. LiBr is soluble in n-butanol, but sugars are not. As n-
butanol is slightly soluble in water (63.2 g/L), small amount of water will be picked up
into n-butanol layer in the process of extraction. The presence of water increases the
solubility of both salt and sugars in organic phase. As a result, small amount of sugars
will be extracted into the organic phase. To reduce the solubility of sugars in n-butanol,
hexane was used as a phase modifier. However, the inclusion of hexane inevitably
decreases the solubility of LiBr in n-butanol due to the decreased polarity of the organic
phase. Because of strong interaction between LiBr and sugars in thick sugar syrup, it is
very difficult and costly to completely remove LiBr from sugars by extraction alone.
Small amount of LiBr left over in the sugar stream after the extraction should be
removed by other means, such as ion-exchange resin, crystallization of sugars, anti-
solvent precipitation of sugars, precipitation of salt by sodium carbonate, as shown in
Figure 4.11.
132
The results in Tables 4.9 and 4.10 show that not only LiBr but also glucose were
extracted into butanol. It was also observed that small amount of water was also
extracted into the butanol phase. As glucose is insoluble in pure butanol, the extraction
of glucose into butanol should be attributed to the presence of water and LiBr in butanol.
In order to reduce the extraction of glucose into butanol phase, hexane was added as a
phase modifier.
Figure 4.11 Flowchart for separating LiBr and sugars by solvent extraction.
133
Table 4.9 Separation of glucose and LiBr by extraction with a mixture of butanol and hexane
Note: Glucose and LiBr solution (200 mg glucose and 1500 mg LiBr in 1mL water) was extracted with
butanol/hexane in a 10-mL screw capped vial. In each extraction, 1 mL organic solvent was added, and
the vial was vortexed for 1 min for extraction. Organic phase was separated from aqueous phase by
centrifugation. When organic phase was removed, 1 mL fresh organic solvent was added for next
extraction.
Table 4.10 Separation efficiency of LiBr and glucose by solvent extraction
Note: calculation is based on composition analysis of the organic phase from the 1st time extraction.
As shown in Table 4.9, when the ratio of hexane in butanol increased gradually,
less glucose was extracted and more extraction times were needed to achieve satisfactory
Butanol/hexane
(v/v) Extraction times
Glucose retained after
extraction (mg)
Glucose retained
/extraction times
10:0 6 95 16
9:1 7 115 16.5
8.5:1.5 8 140 17.5
8:2 9 160 17.8
7.5:2.5 9 175 19.4
7:3 10 185 18.5
6.5:3.5 12 187 16
6:4 13 189 14.5
5:5 15 190 12.6
Butanol/hexane Glucose extracted (mg) LiBr extracted (mg) LiBr/sugar
10:0 20 383 19
9:1 12 310 26
8.5:1.5 7 277 40
8:2 5.5 252 46
7.5:2.5 5 231 46
7:3 4.5 216 48
6.5:3.5 4 195 49
6:4 3 152 51
5:5 2 141 71
134
extraction of LiBr because of the decreased solubility of LiBr in the polarity-decreased
solvent. When butanol to hexane ratio increased to 7:3, about 10% of the original
glucose (~30 mg out of 320 mg glucose) was extracted into butanol phase. Further
increasing hexane ratio was unable to significantly reduce the extraction of glucose,
instead, more extraction times were needed for a satisfactory extraction, as shown in
Table 4.10. Meanwhile, because of the significant extraction of both LiBr and water into
the organic phase, the volume of the hydrolysate decreased gradually, resulting in a thick
sugar syrup. For example, with the butanol to hexane ratio of 7:3, extracting 10 times
reduced the amount of LiBr in the LiBr-glucose solution (see note under Table 4.9) from
1500 mg to 250 mg. Extracting 20 times and 30 times could decrease the amount of LiBr
from 1500 mg to 50 mg and from 1500 mg to 30 mg, respectively. Further extraction
could remove more LiBr, but it was very difficult to completely remove LiBr from
glucose. This might be caused by the strong interaction between glucose and LiBr. In the
thick sugar syrup, LiBr molecules were surrounded by glucose and water molecules
through hydrogen bonding. Therefore, the organic solvent in which glucose and water
have much lower solubility was unable to penetrate into the thick sugar syrup to extract
the residual LiBr.
135
Figure 4.12 The picture of extraction of LiBr from hydrolysate with butanol-hexane.
As we discussed previously, CaBr2 can work as effectively as LiBr to hydrolyze
lignocelluloses. Therefore, we tested the extraction of CaBr2 with butanol in the same
procedure. Preliminary result shows that CaBr2 could also be extracted into
butanol/hexane (extraction procedure same to that of LiBr extraction); however, lower
butanol/hexane ratio (5:5) was needed to facilitate the separation of CaBr2 and glucose.
When butanol/hexane (7:3) was used to extract CaBr2-glucose solution (1500 mg CaBr2,
1 mL water, and 200 mg glucose), ~95 mg glucose was left in the aqueous phase after 10
times’ extraction. When butanol/hexane (5:5) was used to extract CaBr2-glucose solution
(1500 mg CaBr2, 1 mL water, and 200 mg glucose), extracting 30 times decreased the
136
amount of CaBr2 from 1500 mg to 70 mg, and glucose from 200 mg to only 190 mg. It
can be seen that more extraction times were needed to further decrease residual CaBr2. In
real application, the extraction process could be continuous and the amount of solvent
will be significantly decreased because of fresh extraction solvent in later stage could be
reused in previous stage. It might be more effective and economic to remove the residual
LiBr in other ways.
Based on the result of extracting LiBr from glucose-LiBr solution above, it was
tested to separate LiBr from real biomass hydrolysate with n-butanol. Spruce powder (4
g) was hydrolyzed using the conditions of entry 4 in Table 4.6. After the hydrolysis, the
hydrolysate was centrifuged to separate supernatant and precipitate (lignin and
unhydrolyzed spruce, if any). Precipitate was washed with 5 mL of n-butanol twice to
recover LiBr in the precipitate. The butanol washings were mixed with hexane
(butanol/hexane, 7:3, v/v) and used as extraction solvent. The collected supernatant
(containing 0.048 g arabinose, 0.084 g galactose, 1.64 g glucose, 0.2 g xylose, and 0.4 g
mannose) was shown in Figure 4.5 and extracted with butanol/hexane (7:3) for 20 times
in a 50-mL screw capped bottle. Specifically, 5 mL of organic solvent was added in the
first extraction, as shown in Figure 4.12. After vortexed for 1 min, the bottle was
137
centrifuged to separate organic phase and aqueous phase. When the organic phase was
removed, 5 mL of fresh organic solvent was added. The extraction was repeated 20 times.
After the extraction, the resulting syrup-like sugar mixture was analyzed, which
contained 1.4 g glucose, 0.075 g xylose, 0.3 g mannose, 0.260 g LiBr, and 1 mL water,
as shown in Figure 4.13. From the sugar content of spruce shown in Table 4.6 and initial
LiBr usage, it was calculated that the recovery yields of glucose, xylose and mannose
were 83, 36, and 74%, respectively, and that 96.5% of initially loaded LiBr was
recovered. The unrecovered sugars could be found in two parts. Small amount of the
sugars was degraded to furfural and HMF during the hydrolysis, and the rest was
extracted to the organic layer, which could be recovered in next cycle.
Figure 4.13 Formed sugar syrup after butanol extraction of hydrolysate.
138
4.3.9 Removal of residual LiBr from sugar stream
4.3.9.1 Precipitation of sugars from sugar-LiBr solution by anti-solvent
Because sugar has lower solubility in organic solvent than in water, it is expected
that the sugars in the sugar syrup obtained after n-butanol extraction would precipitate
out and LiBr would retain in the mother liquor if water-miscible anti-solvent such as
methanol, ethanol or acetone, in which sugars are insoluble and LiBr is soluble, is added.
By doing this, sugars and LiBr might be completely separated.
However, it was found that the direct addition of methanol or ethanol did not
work, and no sugars precipitated, because the alcohols and the syrup are miscible. The
presence of water and LiBr in the syrup was thought to increase the solubility of sugars
in the water-alcohol mixture. On the other hand, addition of less polar acetone into the
syrup did not work either. The sugar syrup was partially dissolved in acetone, and the
undissolved part still contained a significant amount of LiBr, indicating that acetone is
not suitable to extract LiBr and precipitate out sugars. The dissolution of sugars in the
organic solvent is attributed to the presence of water in the syrup.
In order to reduce the dissolution of sugars in organic solvent, we tried to remove
the water by evaporation before the precipitation. However, the presence of LiBr made
139
the complete removal of water in the sugar syrup very difficult through vaporization
because of the extremely low vapor pressure of LiBr-water. It was found that after the
water was evaporated, the syrup was still soluble in methanol, but became only partially
soluble in ethanol, and completely insoluble in acetone.
From the observation above, we proposed that dissolving the syrup into methanol
or ethanol first followed by pouring the solution in acetone to precipitate out glucose and
retain LiBr in the solution. Pure glucose syrup with a very small amount of water was
tested first. For example, 320 mg glucose, 50 mg LiBr, and 50 L water were dissolved
in 1 mL methanol at 80 ºC. The resulting solution was then added into 10 mL acetone
dropwise with constant stirring, which precipitated the glucose that could be separated
by centrifugation. The results showed that 240 mg glucose was recovered as solid and
only 4.5 mg LiBr was carried over into the solid.
Then the real sugar syrup from spruce hydrolysis was tested with this method.
The sugar syrup (containing ~280 mg glucose, ~15 mg xylose, ~60 mg mannose, 50 mg
LiBr, and 0.2 mL water) was first evaporated to reduce the water content from 0.2 mL to
0.05 mL. Then the concentrated syrup was dissolved in 1 mL methanol at 80 °C. This
solution was then added into 10 mL acetone dropwise with severe stirring to allow the
140
formation of sugar precipitate. The sugar precipitate were collected by centrifugation or
filtration, leaving LiBr in the solution. Approximately 220 mg glucose and 4 mg LiBr
were found in the precipitate. In other words, 79% glucose in the syrup was recovered
and only 8% LiBr in the syrup was carried over with the sugar.
In summary, dissolving sugar-LiBr mixture in methanol followed by dropping
the solution into acetone avoided the strong interaction among sugars, water, and LiBr
and thereby separate LiBr and sugars. Addition of methanol weakened the interaction of
LiBr, sugar, and water and increased the solubility of LiBr in acetone. It turned out that
most of the residual LiBr in the syrup was dissolved into acetone and most of the
glucose in the syrup were precipitated out. Small amounts of glucose and xylose were
dissolved into organic solvent with LiBr, which still need further separation. In addition,
a large amount of acetone was needed to precipitate sugars in this method. However, it is
worthy pointing out that the dissolved sugar and LiBr in organic solvent can be directly
recycled together and used in next hydrolysis cycle without further separation.
141
4.3.9.2 Crystallization of sugars in anti-solvent
Vaporizing water from the sugar syrup significantly decreased the solubility of
sugar in ethanol, making the crystallization of the sugar possible. When the syrup was
dissolved in ethanol, ethanol molecules broke the strong interactions between sugars,
water, and LiBr. Because of the high solubility of LiBr in ethanol and the low solubility
of glucose in ethanol, glucose is supposed to crystallize or be precipitated out.
Table 4.11 Effects of LiBr and water on the precipitation of glucose by ethanol
Note: Specific amount of LiBr and water (as shown in the Table), 400 mg glucose, and 2 mL ethanol were
mixed and heated to 120 C under stirring until a clear solution formed. Then solution was cooled down to
room temperature naturally under stirring and stirred for additional 20 min. Sugar precipitated out and was
filtrated and then washed with 5 mL ethanol. The precipitate then was dried at 105 C.
To verify this assumption, specific amounts of glucose, LiBr and water, as shown
in Table 4.11, was dissolved into ethanol by heating to 120 °C to facilitate the
dissolution of glucose in ethanol and thereby form a clear solution and then cooled to
Entry LiBr (mg) Water (µL) Precipitated glucose (mg)
1 50 30 330
2 50 50 330
3 50 100 330
4 50 150 330
5 100 150 320
6 150 150 300
7 200 150 250
142
room temperature naturally under stirring for sugar crystallization (or precipitation).
Precipitated glucose was filtered and washed with ethanol. However, the presence of
water and LiBr increased the solubility of glucose in ethanol, and therefore glucose still
could not be fully recovered in this way. Effects of LiBr and water on the precipitation of
glucose by ethanol were investigated, and the results are shown in Table 4.11.
Figure 4.14 Sugars precipitated from solvent.
Then the sugar syrup from spruce hydrolysis was tested. The sample of the sugar
syrup containing ~280 mg glucose, ~15 mg xylose, ~60 mg mannose, ~50 mg LiBr, and
143
0.2 mL water was vaporized to reduce water content from 0.2 mL to 0.05 mL. The same
operation procedure was carried out as did for the glucose-LiBr mixture above. It was
found that approximately 200 mg glucose was precipitated, as shown in Figure 4.14.
Almost all other sugars and LiBr retained in ethanol. It was also found that hemicellulose
sugars such as xylose were very difficult to crystallize. In addition, xylose may
negatively affect the crystallization of glucose.
In summary, precipitation or crystallization in anti-solvent was not an efficient
way to completely separate sugars and LiBr. Some sugars, in particular hemicellulose
sugars, retained unrecovered.
4.3.9.3 Separation of LiBr and sugars by Ion exchange chromatography
Although sugars and LiBr could be separated by ion exclusion chromatography,
as discussed above, recovered LiBr and sugars were extremely diluted (both at
concentration of ~1%), which will consume a lot of energy for reconcentration. The ion
exclusion method may be not feasible in industrial application.
To keep the sugar concentration as high as possible, ion exchange
chromatography might be an alternative method. Ion exchange resin has been widely
used to remove salt by exchanging the cation and anion in salt with the H+ and OH
- in
144
resins. The disadvantage of ion exchange resin is that when H+ and OH
- are consumed,
the resin needs regeneration by flushing cation exchange resin with acid and anion
exchange resin with alkali. Obviously, directly applying ion exchange resin to raw
hydrolysate from biomass saccharification to separate LiBr from sugar is almost
impossible because of the very high concentration of LiBr, which would need huge
amounts of ion exchange resins and very large columns, and consume tremendous
amount of acid and alkali to regenerate the resins. However, when the majority of LiBr is
extracted with organic solvent, as discussed above, LiBr concentration in the sugar syrup
becomes much lower, for example ~5% of original feedstock loading (50 mg residual
LiBr/0.8 g lignocelluloses), which theoretically only needs ~3.5% sulfuric acid (28 mg)
and ~2.6% calcium hydroxide (21 mg) to regenerate the cation and anion resins. The
mechanism of whole process is shown in experimental section.
The consumption of the acid and alkali for resin regeneration would be
comparable to that in the dilute acid hydrolysis of biomass (including the neutralization
of acid). Typically, 1-5% acid loading is common in dilute acid hydrolysis or
pretreatment, and accordingly calcium hydroxide is needed to neutralize the acid to form
gypsum. In our experiment, after extraction of spruce hydrolysate with butanol/hexane,
145
concentration of sugar syrup reached to 67% (1.4 g glucose, 0.075 g xylose, 0.3 g
mannose, 0.260 g LiBr, and 1 mL water), which is too thick to pass the columns packed
with the ion exchange resins. Therefore, the sugar syrup needed to be diluted to ~50%
before added onto columns.
Figure 4.15 The pictures of columns packed with cation and anion exchange resins.
Two small columns (diameter: 1 cm and length: 5 cm) were packed, one with
anion exchange resin (DOWEX-2, Cl- form, 100-200 mesh), and another with cation
exchange resin anion exchange (Amberlyst 15, 25-50 mesh), as shown is Figure 4.15.
Prior to operation, column was washed with 10 mL 5% NaOH solution followed by fresh
146
water until neutral pH. Both columns were de-watered by injecting air. The syrup after
LiBr was extracted with butanol-hexane (sugar concentration: ~67%) was diluted with
water into sugar concentration of 50%. And then 0.5 mL the dilute sugar solution was
loaded into cation (or anion) exchange column and was pushed through the column by
injecting air. Recovered solution was subsequently loaded onto anion (or cation)
exchange column. Analysis of recovered sugars solution showed that LiBr was
completely removed. The sugar stream maintained the original sugar concentration of
~50% because air was used as eluent and no dilution occurred. If necessary, dry solid
sugar products could also be produced by removing water. The picture of purified
concentrated sugar solution was shown in Figure 4.16.
Figure 4.16 The picture of purified concentrated sugar solution.
147
4.3.9.4 Removal of LiBr from sugar solution through precipitation
Another method to remove LiBr from sugar stream is the precipitation of salt
with sodium carbonate according to the following reaction.
2 LiBr + Na2CO3 Li2CO3 + 2NaBr
This is an easy and effective way to remove small amount of LiBr in the sugar
stream after the majority of LiBr is extracted with butanol and hexane. Since lithium
carbonate is insoluble in water, it is very easy to separate it by filtration or centrifugation.
If necessary, lithium carbonate can be converted to lithium bromide for next batch of
saccharification, according to the reaction below.
Li2CO3 + 2HBr 2LiBr + H2O + CO2
4.4 Conclusion and recommendations
Concentrated solutions of varied halide salts were tested for their cellulose
dissolution ability. It was found that LiCl, LiBr and CaBr2 solutions at a concentration of
over 60% had superior cellulose dissolution ability over other salts. It is proposed that
solvation structure/pattern of the salts significantly affects their cellulose dissolution
ability. In concentrated solutions, the cation empty coordination sites were thought to
148
play an important role in coordinating cellulose hydroxyl group, thereby disrupting
hydrogen bonding. In addition, cation could act as Lewis acids to hydrolyze cellulose
and allow depolymerized oligosaccharides to be dissolved away from lignin.
Addition of acid into the concentrated salt solutions accelerated the hydrolysis of
polysaccharides and oligosaccharides into monosaccharides. It was found that the
lignocelluloses could be completely and selectively hydrolyzed into sugars quickly with
limited formation of fermentation inhibitors. Batch-feeding could reduce the ratio of salt
solution to biomass to 1:1, and thereby produce a concentrated sugar stream.
Several methods were investigated for separation of sugars and salt. Ion-
exclusion chromatography could separate sugars from the salt, but the resultant sugar
and salt streams were extremely diluted. Solvent extraction was able to separate LiBr
from sugars, but it was very difficult and costly to completely separate them. It turned
out that the combination of organic solvent extraction and ion-exchange chromatography
was a successful method to separate sugars from salt, for example, extracting 95% of the
salt from sugar solution with butanol-hexane first and removing the residual 5% salt
using ion-exchange resins.
149
The results suggested that saccharification of biomass in concentrated salt
solution at moderate temperature might be a cost-effective way to produce sugars
without pretreatment and enzymatic hydrolysis. The process needs further, for example,
to create mass balance of the process, elucidate and confirm the hydrolysis mechanism
of lignocellulose in concentrated salt solutions and the role of halide salt, modify and
optimize the sugar-salt separation methods, and investigate the changes of lignin during
the saccharification. Appropriate equipment needs to be selected and evaluated for the
saccharification process to improve efficiency and reduce cost, for example, a twin-
screw extruder as the reactor for saccharification and a continuous liquid-liquid extractor
for separating sugars and salt.
150
Chapter 5: Conversion of Lignocelluloses into
Hydrocarbons
5.1 Introduction
Cellulosic ethanol has been receiving a lot of attentions for years. However, its
production still faces a lot of issues. Costly pretreatment of the feedstock under severe
conditions such as high temperature and high pressure is required to remove the
recalcitrance caused by lignin and hemicelluloses prior to enzymatic saccharification of
cellulose (Demirbas, 2005). In addition, the high cost of cellulases, the low efficiency of
fermentation of pentoses, the high energy consumption for ethanol distillation, as well as
long production cycle make cellulosic ethanol economically incomparable to fossil fuels
at this stage. Furthermore, the low heating value and water-absorbent property of ethanol
make it not an ideal substitute for gasoline (Yoon et al., 2009). These issues of fuel
ethanol have been the driving force of developing the next generation liquid biofuels
from biomass. For example, converting biomass into liquid hydrocarbons which have the
same physiochemical properties as the traditional fossil fuels, attracting more and more
attention (Elliott and Schiefelbein, 1989; West et al., 2009). Hydrocarbons (gasoline,
151
diesel and jet fuel) from biomass have great advantages and promise. For example,
hydrocarbons have an overall energy efficiency of 2.1 (ratio of the heating value of
alkanes to the energy required to produce the alkanes), compared to 1.1-1.3 of bioethanol
(Huber and Dumesic, 2006). In addition, limited water use for processing, short
production cycle, and the elimination of energy-intensive distillation allow low-cost
production of hydrocarbons.
There are five main pathways under investigation to convert biomass into liquid
hydrocarbon fuels: (1) biomass gasification (to syngas) followed by Fischer-Tropsch
synthesis (Kirubakaran et al., 2009); (2) biomass pyrolysis (to bio-oil) followed by
cracking and upgrading (Balfanz et al., 1993); (3) dehydration of oxygenates from
biomass under multifunctional heterogeneous catalysts followed by hydrogenation
(Chheda et al., 2007); (4) depolymerization followed by hydrodeoxgenation of lignin
(Pandey and Kim, 2011); and (5) decarboxylation of alkyl carboxyl acids such as
levulinic acid derived from hexoses followed by chain extension (Bond et al., 2010).
Of the approaches above, dehydration and hydrogenation of oxygenates derived
from biomass saccharides have garnered considerable interest for hydrocarbon
production. First, the resulting hydrocarbon fuels are the same as traditional fossil fuels
152
so that modification of existing distribution infrastructure and vehicle engine is
unnecessary. Second, hydrocarbons from biomass have comparable heating value and
gas mileage as gasoline. Third, hydrocarbons are immiscible with water; therefore, the
expensive distillation step is eliminated. Fourth, as the bio-hydrocarbons are produced in
chemical ways, it allows a much shorter production period and higher feedstock
concentration than biological ethanol. Fifth, because heterogeneous catalysts can work at
higher feedstock concentration, processing water needed in hydrocarbon production can
be significantly reduced, compared to the low ethanol concentration caused by limited
ethanol tolerance of yeast. Sixth, catalysts in chemical conversion can be recycled by
simple filtration and reused for months or years, whereas the recovery of enzyme and
yeast in bioethanol production is expensive and incomplete. Besides, as most of sugar
and sugar derivatives are water soluble, the reactions can be conducted in aqueous
solution, which allows automatic separation of the final hydrophobic products from
water (Huber and Dumesic, 2006).
Extensive work has been done on saccharides-derived hydrocarbons via
hydrogenation/dehydration of sugar derivatives, such as hydroxymethylfurfural (HMF)
and furfural. A method of directly converting six-carbon sugars into hexane under a
153
bifunctional catalyst was reported (Huber et al., 2004). Hydrocarbons with carbon chains
longer than five or six carbons could be prepared through aldol-condensation of
furfural/HMF and acetone, resulting in hydrocarbons with up to 20 carbons, which was
very similar to the composition of gasoline and jet fuel (Chheda et al., 2007). However,
the intermediates (HMF and furfural) could not be easily produced in high yield and at
low cost as HMF tended to polymerize and form insoluble humin in acidic aqueous
solutions (Vandam et al., 1986). Many studies were conducted to improve the selectivity
of sugar conversion into furfural/HMF. For example, dimethyl sulfoxide (DMSO) was
used to replace water, and the water-free environment promoted the dehydration of
glucose into HMF (Amarasekara et al., 2008). In another study, methyl isobutyl ketone
(MIBK) was added as an extraction solvent to collect HMF in situ when it was produced
(Roman-Leshkov et al., 2006). The HMF formed was immediately extracted into the
upper MIBK layer, which largely reduced the opportunity for the
polymerization/condensation of HMF into insoluble humin. However, the low solubility
of sugars in the organic solvents decreased the feedstock concentration and therefore the
final product concentration. In addition, it was difficult to separate HMF from the
solvent because of their similar boiling points.
154
It was reported (Zhao et al., 2007) that HMF could be produced in high yields
and with high selectivity using ionic liquids as solvents with chromium halide as catalyst.
The issues related to this process are that ionic liquid is very expensive and difficult to
recover, and chromium halide is potentially toxic to environment. Partial replacement of
ionic liquid with traditional cellulose solvents was also investigated (Binder and Raines,
2009). High conversions of up to 90% were achieved from monosaccharides and pure
cellulose, whereas the yield was only 50% from corn stover powder. In addition, this
complicated system involved two expensive organic solvents (dimethylacetamide and
ionic liquid) as well as inorganic catalysts, which made them extremely difficult to
recycle. Also, the ionic liquid had high viscosity and worked well only in water-free
environment, which required the feedstock to be finely ground and completely dry.
Unfortunately, both grinding and drying of biomass are energy intensive. Currently, one
of the most potentially useful ways to produce extended hydrocarbon chain from sugars
was through aldol-condensation of furfural compounds and acetone, as reported by
Chheda et al (Chheda et al., 2007). This process used DMSO as solvent and MIBK as
extraction solvent to produce HMF and required another step to conduct the aldol-
condensation of HMF and acetone; thus the whole process was very complex. In
155
addition, this process could only use soluble sugars as feedstock, which did not break the
bottleneck of utilizing lignocelluloses for cheap production of sugars.
In this chapter, a process of directly converting lignocelluloses into hydrocarbon
precursors (furfural-/HMF-acetone adducts) for liquid hydrocarbon production under mild
conditions was demonstrated. Using acetone as a solvent in presence of lithium bromide
and a small amount of hydrochloric acid and water, the process integrated the hydrolysis
of polysaccharides (both cellulose and hemicelluloses) to monosaccharides, dehydration of
the monosaccharides to HMF (from hexoses) and furfural (from pentoses), and aldol
condensation of HMF and furfural with acetone into a single step to produce hydrocarbon
precursors. The process was abbreviated as HDA (Hydrolysis-Dehydration-Aldol
condensation). The hydrocarbon precursors from the process could be easily hydrogenated
to hydrocarbons with chain lengths in the range of C5-C21. Meanwhile, lignin in the
biomass was depolymerized and dissolved in acetone in the process. The dissolved lignin
could be easily separated when the acetone was evaporated for recycling. The lignin was
expected to have considerable potential for co-products development. The acetone and
LiBr could be recycled and reused in next batch.
156
5.2 Experimental
5.2.1 Chemicals and materials
Various types of biomass feedstocks were used in the present research, including
softwood spruce, hardwood poplar, crop residue corn stover, energy crop switchgrass, and
paper (newspaper and print paper). After being air-dried, the biomass samples were
ground to pass a 40-mesh screen using a Wiley mill. Sources and chemical composition of
the biomass samples are presented in Table 4.1. Chemical reagents used in this study were
purchased from Fisher Scientific (Pittsburgh, PA) or Sigma-Aldrich (St. Louis, MO) and
used as received.
5.2.2 Production of hydrocarbon precursors from biomass
Exact amounts (see Table 5.1) of LiBr and hydrochloric acid, 1 mL acetone and
100 μL water were loaded into a 15-mL vial and vortexed to mix well. To this solution
100 mg 40 mesh biomass powder or sugar was added and the mixture was vortexed again.
The vials was heated at 80-120 ºC in an oil bath and stirred with magnetic stir bar at 500
rpm for 2 h. At the end of reaction, acetone in the mixture was carefully removed at low
temperature (50 C) using a rotary evaporator. The residue was extracted with CH2Cl2 to
collect hydrocarbon precursor products. The residue after the extraction, consisting of
157
liquid (water + LiBr) and solid (lignin and unreacted biomass), was filtered and washed
with water. The filtrate was collected for analysis of saccharides, HMF, furfural, and
residual LiBr and HCl. The solid residue was redissolved in acetone and filtered. The
filtrate was dried to produce acetone-soluble lignin, and the residue after filtration was
dried for composition analysis.
5.2.3 Determination of residual LiBr
Procedure is shown in Section 2.2.7.
5.2.4 Determination of sugars and sugar derivatives
Procedure is shown in Section 2.2.6.
5.2.5 Qualitative analysis of hydrocarbon precursors using GC-
MS
GC-MS was used to identify components in the products. Approximately 5 mg
product was dissolved into 1 mL CH2Cl2 for GC-MS analysis. GC-MS was performed on
a GC-MS-QP 2010 instrument (Shimadzu Co., Addison, IL) equipped with a 30 m × 0.25
mm i.d., 0.25 m film, SHR5XLB capillary column. Helium was used as carrier gas at a
flow rate of 1 mL/min. GC conditions were as follows: initial column temperature, 100 °C,
158
held for 1 min, ramped at 2 °C/min to 310 °C, and then held for 4 min; injector
temperature, 250 °C; split ratio, 1/20, EI mode at 20 kV for ionization.
5.2.6 Quantitative analysis of hydrocarbon precursors using ESI-
MS and oxidation methods
(1) ESI-MS analysis
ESI-MS in positive-mode was used to quantify the products on an Applied
Biosystems 3200 QTRAP instrument. Product (10 mg) was dissolved in 10 mL
methanol/CHCl3 (1:1, v/v) to prepare 1 mg/mL product solution. Furfural-acetone (1 mg)
was dissolved in 10 mL methanol/ CHCl3 (1:1, v/v) to prepare 0.1 mg/mL standard
solution, which was further diluted into 0.01 mg/mL. Combining 0.5 mL the product
solution and the 0.5 mL 0.1 mg/mL standard solution led to a sample with product
concentration of 0.5 mg/mL and standard concentration of 0.05 mg/mL.
(2) Oxidation analysis
i. Preparation of dimethyldioxirane in acetone: A 250 mL two-necked round-
bottomed flask was connected by a U tube to a receiving flask. The receiving flask was
connected to a vacuum pump. Receiving flask was cooled at -80 °C by means of liquid
nitrogen/ethanol bath. The reaction flask was loaded with a mixture of water (50 mL),
159
acetone (40 mL), and NaHCO3 (15 g) and cooled to 5-10 °C with the help of an
ice/water bath. With vigorously stirring and cooling, solid oxone (40 g) was added in
five portions at 3-min intervals. Waiting for 3 min after the last addition, the ice/water
bath was removed and a moderate vacuum (80-100 Torr) was applied.
Dimethyldioxirane/acetone solution was distilled and collected in the cooled receiving
flask. The volume of dimethyldioxirane (5% yield) in acetone was approximately 40 mL.
ii. Oxidation of the furan-derived hydrocarbon precursors with dimethyldioxirane:
to avoid the oxidation of hydroxyl in the hydrocarbon precursors into carbonyl group,
the hydrocarbon precursors was acetylated with acetyl bromide in CH2Cl2/pyridine prior
to oxidation. After removing CH2Cl2/pyridine, the acetylated hydrocarbon precursors
(from 100 mg sugars) was dissolved in acetone and reacted with excessive
dimethyldioxirane in acetone for 30 min at room temperature. After the completion of
the reaction, acetone was removed; the oxidized hydrocarbon precursors were dissolved
in 50 mL ethanol and ready for coloration assay.
iii. Determination of carbonyl content in the hydrocarbon precursors: The original
or oxidized hydrocarbon precursors in ethanol (1 mL) was reacted with 1 mL 5% (w/v)
dinitrophenylhydrozine in acidic ethanol at 55 °C for 20 min and then cooled in an ice
160
bath. Then 8 mL of 5% (w/v) KOH in ethanol was added and the mixture was
centrifuged to separate precipitant (KCl) and red wine colored supernatant. The
supernatant was taken to determine the carbonyl content using UV at 432 nm.
5.2.7 Estimation of lignin molecular weight
Gel permeation chromatography (GPC) was used to estimate the molecular weight
of lignin. GPC was performed with a Viscotek GPCmax VE-2001 chromatograph using
two columns (VARIAN 5M-POLY-008-27 and VARIAN 5M-POLY-008-32). THF was
used as an eluent with a flow rate of 1 mL/min at 30 ºC. Monodispersed Polystyrene
standards were used for calibration.
5.2.8 Characterization of LiBr/acetone solvent systems
FT-IR was used to investigate the bonding changes caused by the interactions
between acetone/water and LiBr. Spectra were recorded on a PerkinElmer Spectrum 100
FT-IR spectrophotometer with universal attenuated-total-reflection (ATR) sampling
accessory (Waltham, MA).
5.3 Results and Discussion
5.3.1 Description of the HDA process
161
The proposed reaction pathway for direct conversion of biomass to hydrocarbon
precursors is shown in Figure 5.1.
Figure 5.1 Reaction pathway of biomass to hydrocarbon precursors in HDA process (Structures
of R2 are shown in Table 5.6).
The experimental flowchart of the HDA process of converting biomass into
hydrocarbons is shown in Figure 5.2. Biomass feedstock was fed with acetone, water,
salt and acid, into reactor. Reaction was conducted under 100-140 ºC for 1-4 h. After
reaction, unreacted acetone in final liquor was recycled through vaporization. Left
residues including hydrocarbon precursors, salt, acid and water were suspended into
certain amount of CH2Cl2. Extra fresh water was added to facilitate the separation of
162
LiBr and wash CH2Cl2 and lignin. Suspension was centrifuged to separate CH2Cl2 layer
from insoluble lignin and water. CH2Cl2 was recovered through vaporization, and the
hydrocarbon precursors was obtained as thick liquid or solid. The LiBr/water phase was
concentrated and reused. Hydrocarbon precursor from CH2Cl2 was converted into
hydrocarbon through hydrodeoxygenation on bifunctional catalysts.
Biomass
Fuel
1 23
45
6
7 8
a b
c
de
fg
h
i
jk
1
2
3
4
56
7
8
9 Lignin
11
10
4
l
mn
Figure 5.2 Process flow chart of converting lignocelluloses into hydrocarbons. Main facilities:1.
Hydrocarbon precursor synthesis reactor; 2. Acetone evaporator; 3. Centrifuge; 4. Water
evaporator; 5. CH2Cl2 evaporator; 6. Hydrodeoxygenation reactor; 7.Products tank; 8. Hydrogen
gas tank; 9. Acetone tank; 10. Extraction solvent (CH2Cl2) tank; 11. Water tank. Mass flow: a.
Hydrocarbon precursors, acetone, LiBr, water and acid; b. Hydrocarbon precursors, LiBr, water
and acid; c. CH2Cl2 and hydrocarbon precursors; d. Hydrocarbon precursors; e. Hydrocarbons; f.
Acetone; g. Acetone; h. CH2Cl2; i. Water; j. Diluted acidic LiBr solution; k. Concentrated acidic
LiBr solution; l. Hydrogen gas; m. CH2Cl2; n. Water.
163
5.3.2 LiBr/water and LiBr/acetone systems
The interaction of LiBr with different solvents is illustrated in Figure 5.3. In
aqueous solution, polarized -OH bond of water induces partial positive charge on
hydrogen atom that has affinity to Br- and partial negative charge on oxygen that attacks
Li+. Therefore, both Li
+ and Br
- are solvated by water molecules (Figure 5.3a), which not
only separates Li+ and Br
-, but also limits Br
- from attacking positively charged positions.
From mechanism of traditional cellulose solvents of LiBr/DMSO (Figure 5.3b) and
LiBr/DMF (Figure 5.3c), it can be seen how free halide ion is crucial to dissolving
cellulose by destroying hydrogen bonds in cellulose. Different from water, only Li+ is
able to coordinate closely with electronegative O in DMSO or DMF, while Br- is
relatively free in the solvent. The reason is that the sterically hindered positive position
(C or S atom) by neighboring groups cannot interact with Br-. Free halide ion plays a
significant role in destroying the hydrogen bonds and promotes the dissolution of
cellulose in ionic liquid. However, these solvents are expensive, toxic, and hard to
recover due to high boiling point, which excludes or limits their industrial potential as
solvents in directly hydrolyzing lignocelluloses into monosaccharides.
164
Figure 5.3 Interaction of LiBr with different solvents. (a) LiBr in H2O; (b) LiBr in DMSO; (c)
LiBr in DMF or DMAC; (d) LiBr in acetone; (e) ionic liquid (1-Butyl-3-methylimidazolium
bromide).
In addition to these extensively investigated polar aprotic solvents, acetone has
similar structure, as illustrated in Figure 5.3d, and theoretically can coordinate Li+ and
leave Br- free. Besides, acetone has other advantages over DMSO and DMF. First, the
low boiling point makes acetone very easy to separate and recover from the reaction
HO
H
Li+
Br- Br-
O
H
H
H
H
H
H
Li+
S
O
H
H
H
H
H
H
Li+
R N
O
Li+
a: LiBr/H2O
d: LiBr/acetone
c: LiBr/DMF or LiBr/DMACb: LiBr/DMSO
NN Br-
Br- Br-
Br-
f: ionic liquid(1-Butyl-3-methylimidazolium bromide)e:
165
system. The most important one is that acetone can react in situ with newly generated
furfural and HMF through aldol condensation reactions to form the precursors of liquid
hydrocarbons with extended carbon numbers (Figure 5.3e). In addition, the aldol
condensation reduces or prevents the self-condensation of furfural and HMF to humin.
Furthermore, compared to DMSO and DMF, acetone is inexpensive, less toxic, and of
low viscosity. It is also a good solvent for lignin.
165016701690171017301750
wavenumber (cm-1)
acetone+0
acetone+100
acetone+200
acetone+250
acetone+300
acetone+400
acetone+500
acetone+600
Figure 5.4 FT-IR spectra of LiBr/acetone solution.
In order to verify the proposed interaction between LiBr and acetone and the
formation of loosely attached Br-, solutions of LiBr in acetone were studied by FT-IR. In
166
Figure 5.4, the band at 1715 cm-1
represents the stretching vibration of pure acetone.
Because of the interaction of acetone carbonyl group with Li+, partial electron on C=O
bond was donated to Li+, which weakened the C=O bond. The reflection of this on FT-
IR spectra is that a red shift of the peak to low wavenumber region occurred. Reduced
positive charge on Li+ caused by the electron donation from C=O bond weakened the
attraction force between Li+ and Br
-. Thus, loosely attracted or free Br
- was formed in
this situation. The free Br- could effectively catalyze xylose and glucose to furfural and
HMF, respectively, which will be discussed later. It can be seen from Figure 5.4 that
with the increasing of LiBr, peak red shifted more, indicating formation of more loosely
attached Br-.
5.3.3 One-step conversion of biomass into hydrocarbon
precursors
Different types of feedstock, including monomeric saccharides (glucose, xylose,
arabinose, galactose and mannose), polysaccharides (cellulose and starch), papers (filter
paper, newspaper, and print paper), and lignocelluloses (spruce, poplar, corn stover, and
switchgrass), were treated with the HDA process, and the results are summarized in
Table 5.1.
167
The first 14 experiments were conducted with glucose as feedstock to investigate
the effect of LiBr and acid dosage on the conversion of glucose. With the increase of
LiBr or acid dosage, glucose conversion increased. Addition of 300 mg LiBr appeared to
be a turn point. The conversions of 70-80% might be caused by the shortage of effective
free Br- because the amount of LiBr was below 300 mg. However, the conversion
decreased when LiBr dosage increased beyond 300 mg, which was probably caused by
the high viscosity of reaction solvent and the formation of humin. With the same amount
of LiBr, Entry 13 had a higher conversion than Entry 3 because more acid was added.
Entry 4 and Entry 14 had the same conversion though Entry 14 had a higher acid dosage
than Entry 4, implying that 0.25% acid was sufficient to complete a high conversion
without the need of more acid loading in Entry 14. Conversions of Entry 2 and Entry 3
were far lower than that of Entry 4, which might indicate that LiBr dosage less than 200
mg was unable to produce enough free Br- as catalyst (as discussed above, in the
presence of water, most of Br- ions were surrounded and solvated by water). It is worth
pointing out that ~70% conversion was observed even without any addition of LiBr
(Entry 1) whereas no HMF was detected in the final liquor, which was attributable to the
formation of ketal through reaction of acetone with sugar (Pfaff, 1987).
Table 5.1 Conversion of lignocelluloses or sugars into hydrocarbon precursors in acetone/LiBr system
Note: (1) Other reaction conditions: 1 mL acetone, 100 μl water, 120 ºC, 2 h. (2) Acid loading, w% based on weight of solvent. (3) Glu-glucose; Ara-
arabinose; Man-mannose; Xyl-xylose; Cell-cellulose; FP-filter paper; PP-print paper; Pop-poplar; NP-newspaper; CS-corn stover; and SG-switchgrass. (4)
Soluble solid-CH2Cl2 soluble hydrocarbon precursor; insoluble solid-acetone-insoluble residues; residual sugars-sugars left in solution after reaction. (5) A/F
ratio-The molar ratio of reacted acetone to converted sugars.
Entry LiBr
(mg)
Acid
(%,
w/w)
Feedstock
(mg)
Soluble
solid (mg)
Insoluble
solid (mg)
Residual
sugars (mg)
Reacted
acetone (mg)
Sugar
conversion (%)
A/F
ratio
1 0 1 100 (Glu) 46 0 30 13 70 0.6
2 100 0.25 100 (Glu) 50 0 25 21 75 0.9
3 200 0.25 100 (Glu) 64 0 21 25 79 1.0
4 300 0.25 100 (Glu) 83 0 3 39 97 1.3
5 400 0.25 100 (Glu) 92 0 5 45 95 1.5
6 500 0.25 100 (Glu) 105 0 10 55 90 1.9
7 500 0.20 100 (Glu) 99 0 10 51 90 1.8
8 500 0.15 100 (Glu) 87 0 16 46 84 1.7
9 500 0.10 100 (Glu) 63 0 20 31 80 1.2
10 500 0.05 100 (Glu) 51 0 30 28 70 1.3
11 750 0.00 100 (Glu) 43 0 36 4 64 0.2
12 100 0.50 100 (Glu) 67 0 26 34 74 1.5
13 200 0.50 100 (Glu) 135 0 8 97 92 3.3
14 300 0.50 100 (Glu) 155 0 2 136 98 4.4
15 300 0.25 100 (Ara) 110 0 4.3 81 96 2.3
16 300 0.25 100 (Gal) 96 0 14 63 86 2.3
17 300 0.25 100 (Man) 93 0 11.5 57 88 2.1
18 300 0.25 100 (Xyl) 99 0 4.5 68 96 1.9
19 300 0. 50 100 (Cell) 156 0 0 121 98 3.0
20 300 0.75 100 (Xyl) 104 7 0 113 90 2.7
21 300 0.25 100 (Star) 98 0 0 46 98 1.3
22 300 0.5 100 (FP) 141 0 0 92 99 2.6
23 300 2 100 (PP) 50 2 0 34 95 1.1
24 300 1 100 (NP) 113 8 0 53 96 2.0
25 300 0.5 100 (pop)) 95 1 0 76 99 3.8
26 300 0.5 100 (CS) 62 1 0 55 99 2.8
27 300 0.5 100 (spr) 87 1 0 72 99 3.6
28 300 0.5 100 (SG) 74 2 0 61 99 3.5
168
169
Acid promoted the formation of acetone self-condensation product.
Comparing Entry 2-4 and Entry 12-14, with the same LiBr dosage, the later had much
higher glucose conversion than the former, implying that the acid promoted/catalyzed
the reactions. In addition, higher acid loading enhanced the consumption of acetone
caused by self-condensation. In Entry 11, acetone was hardly consumed without the
addition of the acid. In summary, it appeared that 300 mg LiBr with 0.25% acid was
enough to ensure efficient conversions of sugars.
Other saccharides occurring in lignocellulose, including arabinose, galactose,
xylose, and mannose, were studied at the same conditions (300 mg LiBr and 0.25%
acid). The results indicated that these sugars could be converted as easily as glucose.
In addition to the monomeric saccharides above, the HDA process was applied
to polysaccharides and real lignocelluloses at the similar conditions studied above.
Considering that different feedstocks had varied recalcitrance to the reactions, the acid
loading varied with different feedstock. Conversion of 99% was achieved when 0.5%
acid was used for cellulose micron and filter paper. Starch, because of its amorphous
structure, only needed 0.25% acid for a high conversion. However, xylan needed
more acid to get a high conversion probably due to the dry and tough feature of
extracted xylan. Because the inorganic fillers in print paper and newspaper neutralize
acid, they needed more acid to achieve a high conversion (1 and 2% for newspaper
170
and print paper, respectively). For the lignocelluloses investigated (softwood spruce,
hardwood poplar, agricultural residue corn stover, and energy crop switchgrass), 0.5%
acid was enough to break down lignin and convert cellulose and hemicellulose to
sugars for further reaction. A/F ratio (molar ratio of condensed acetone to condensed
furfural or HMF) was used to estimate the average carbon number of hydrocarbon
precursors. High A/F ratio indicates more acetone connected to furfural or HMF, thus
longer chain in final hydrocarbon precursors; low A/F ratio indicates less acetone
connected to furfural or HMF, thus shorter chain hydrocarbon. Although this ratio
was highly dependent on the conversion and acid concentration, it could be used as an
indicator of the extent of acetone self-condensation. For example, higher acid
concentration (>0.5%) gave an A/F ratio larger than 2, indicating a severe acetone
self-condensation, while lower acid concentration (0.25%) resulted in an A/F ratio
less than 2, indicating that acetone condensation was significantly reduced.
The pictures of the product solutions from different feedstocks are shown in
Figure 5.5. The control experiment (acetone and LiBr only) gave a light brown liquid.
The color was supposed to be from the acetone self-condensation products. On the
other hand, the experiments with xylose, glucose, and spruce powder generated
solutions in dark brown or black. The color was probably from the derivatives or
condensation products of furfural or HMF. In all cases, the feedstocks dissolved and
171
formed homogeneous solutions at the end of reaction. After vaporizing acetone, the
hydrocarbon precursors could be easily extracted by methylene dichloride to recycle
catalyst (in aqueous phase), as show in Figure 5.5e.
Figure 5.5 The hydrocarbon precursors from HDA process in acetone. (a) Acetone self-
condensed products; (b) Hydrocarbon precursors from xylose; (c) Hydrocarbon precursors
from glucose; (d) Hydrocarbon precursors from softwood spruce; (e) Hydrocarbon precursors
extracted by CH2Cl2 (bottom layer was water).
5.3.4 Effect of temperature and time on conversion of spruce
powder
Effect of temperature on conversion of spruce powder was investigated in the
range of 80-120 ºC. The results in Table 5.2 indicated that the conversion of glucose
decreased with the decreasing of temperature; however, high conversion could be
achieved by extending the reaction time at lower temperature, which can be seen from
Table 5.3. For example, the conversion of spruce powder was only 67% at 100 ºC for
a b c d e
172
2 h, compared to the 100% conversion at 120 ºC for 2 h. The conversion at 100 ºC
increased to 88% at 3 h and 100% at 4 h, respectively.
Table 5.2 Effect of temperature on conversion of spruce powder
Note: other reaction conditions: 300 mg LiBr, 100 mL water, 1 mL acetone, 100 μL water, 100 mg
spruce powder, 2 h.
Table 5.3 Effect of time on conversion of spruce powder
Note: other reaction conditions: 300 mg LiBr, 100 mL water, 1 mL acetone, 100 μL water, 100 mg
spruce powder, 100 ºC.
5.3.5 Effect of different salts on conversion of spruce powder
The catalysis abilities of different common and inexpensive halide salts on the
conversion were investigated. The results in Table 5.4 shows that CaBr2, NaBr, LiBr
and LiCl had almost the same catalysis abilities in the same reaction time. On the
other hand, transitional metal halide including CuCl2, FeCl3 and ZnCl2 did not give as
good results as alkali metal halide did. The reason probably is that d orbital of
transition metals ion still formed bonds with halide ion in spite of the presence of the
interaction with C=O, which reduced the releasing of free halide ions. Malfunction of
Temperature (°C) Conversion (%)
80 58
100 67
120 99
Time (h) Conversion (%)
1 58
2 67
3 88
4 99
173
LiI might be attributed to its instability, as LiI tends to decompose into LiOH and
iodine. AlCl3 is not an ionic compound that cannot form free Cl- in solution, thus no
obvious change was observed in reaction process.
Table 5.4 Effect of different halogen salts on conversion of spruce powder
Salt Left solid (mg) Carbohydrates Conversion (%)
CaBr2 15 99
NaBr 26 97
NaBr (200mg water) 20 99
LiCl (200mg water) 25 99
ZnBr2 52 65
FeBr3 52 65
MgCl2 40 81
CuCl2 25 /
FeCl3 29 /
CaCl2 41 80
ZnCl2 59 50
LiI Doesn’t work
AlCl3 Doesn’t work
KBr Doesn’t work
Note: other reaction conditions: 300 mg salt, 100 μL water, 1 mL acetone, 100 mg spruce powder, 0.25%
(w/w) HCl (pure hydrogen chloride, based on weight of solvent), 2 h, 120 ºC.
5.3.6 Effect of different acids on conversion of spruce powder
The catalysis abilities of different common mineral acids were investigated.
The results are shown in Table 5.5. It was observed that the conversion of glucose
decreased with decreasing acidity. The conversion reaction proceeded faster when
strong acids were added. Formic acid could only convert 54% spruce powder, but it is
expected that longer reaction time might be able to get a higher conversion. Acetic
acid as weak acid did not result in significant change in the conversion process.
174
Table 5.5 Effect of different acids on conversion of spruce powder
Salt Left solid (mg) Carbohydrates conversion (%)
H2SO4 17 99
HNO3 25 99
H3PO4 12 99
HCOOH 54 54
CH3COOH / /
Note: other reaction conditions: 300 mg LiBr, 100 μL water, 1 mL acetone, 100 mg spruce powder, 2 h,
0.25 % (w/w) acid (pure acid, based on weight of solvent), 120 ºC.
5.3.7 Identification of products
GC-MS was used to identify the products and verify the validity of the
proposed reaction pathway. The proposed mechanisms of acetone self-condensation
and aldol-condensation between furfural and acetone are shown in Figures 5.6 and 5.7,
respectively. From the pathway shown in Figure 5.6, it is expected that the molecular
weight of acetone self-condensation products should be n1 40 + n2 58 (n1, n2=0, 1,
2…, n1 + n2 is the number of condensed acetone) because of possible intramolecular
dehydration. Similarly, molecular weight of HMF-acetone condensation products
should be 126 + n1 40 + n2 58 (n1, n2 = 0, 1, 2…, n1 + n2 is the number of
condensed acetone), while molecular weight of furfural-acetone condensation product
should be 96 + n1 40 + n2 58 (n1, n2=0, 1, 2…, n1 + n2 is the number of condensed
acetone). The products would be a mixture of the products from acetone self-
condensation and HMF- and furfural-acetone condensation with varied molecular
weight. The proposed structures and their molecular weights from these proposed
175
mechanisms is shown in Table 5.6. The GC-MS and positive ESI-MS of acetone self-
condensation products with peaks in black, HMF-acetone condensation products in
red, furfural-acetone condensation products in blue, are shown in Figures 5.8 and 5.9,
respectively. As expected, molecular weight of most peaks observed from GC-MS
agreed with the number calculated from the formulas proposed above, which verified
the proposed reaction pathways. In positive ESI-MS graphics, as an ion was produced
by the addition of a proton, so the observed m/z was the value of molecular weight
plus one. Based on this, most peaks had the m/z fitting the proposed formulas.
Figure 5.6 Proposed self-condensation mechanism of acetone.
Figure 5.7 Proposed condensation mechanism of furfural with acetone.
176
Table 5.6 Proposed structure of products and corresponding molecular weight
Note: ‖R‖ represents side chain of furan ring (see Figure 5.1).
In Figure 5.8a, MW of acetone self-condensation products ranged from 98 to
298. The former was the MW of the dehydration products of 2 acetone molecules, and
the latter was the MW of the dehydration products of 7 acetone molecules. Between
them, the peaks of 138, 178, 218, 258, and 298 were attributed to dehydration
products of 3, 4, 5, and 6 acetone molecules, while those of 120, 160, 200, 240, and
280 were the MWs of the ring structures formed from the compounds of 138, 178,
218, 258, and 298 after further dehydration (Figure 5.8 and Table 5.6).
Condensed
acetone
R (double bond could be
added by water)
MW of HMF-acetone
adduct
MW of furfural-acetone
adduct
1
166, 184, 136, 154,
2
206, 224, 242, 176, 194, 212
3
246,264 216,234
4
286,304 256,274
O
O
O
O
O
O
177
a) Acetone + acetone
c) Furfural + acetone
b) HMF + acetone
d) Glucose + acetone
178
Figure 5.8 GC-MS spectra of hydrocarbon precursors in CH2Cl2.
For HMF-acetone condensation products, Figure 5.8d shows that the
molecular weights were from 126 for HMF to 246 for HMF + 3 acetone - 3 H2O. The
peaks of 166 and 206 were the MWs for HMF + 1 acetone – 1 H2O and HMF + 2
acetone – 2 H2O, respectively.
For xylose-acetone condensation products that are shown in Figure 5.8e, 136
was the MW for furfural + 1 acetone – 1 H2O, followed by 176, 216, and 256 that
were attributed to the MWs for furfural + 2 acetone – 2 H2O, furfural + 3 acetone – 3
H2O, and furfural + 4 acetone – 4 H2O, respectively. Products with incomplete
f) Spruce + acetone
e) Xylose + acetone
179
dehydration were also observed, which were 234 (furfural + 4 acetone – 3 H2O), and
274 (furfural + 5 acetone – 4 H2O), respectively.
For spruce-acetone condensation products, only one condensation product
derived from pentose (furfural-acetone) was observed at 136, the MW of dehydration
product of 1 acetone and 1 furfural. Peaks for hexose-derived products observed were
the condensation products of 1 HMF with 1, 2, and 3 acetone. The peak at 184 was
the condensation product of 1 HMF and 1 acetone without dehydration.
The difference between GC-MS and ESI-MS spectra was that peaks with
higher m/z in ESI-MS spectra were observed, but they were weak and not clear when
m/z was above 300. With m/z below 300, most peaks were from condensation of 1
furfural with a few acetones. Most of the peaks with m/z above 300 observed were
from the condensation of 2 furfural with acetone. Statistically, this happened with
very low possibility since the concentration of acetone was much higher than that of
furfural, and each furfural molecule was surrounded by a large amount of acetone
molecules, as indicated by the very small amount of products with molecular weight
above 300.
180
Figure 5.9 ESI-MS spectra of hydrocarbon precursors in CH2Cl2.
When molecular weight was below 300, carbon number was estimated to be 1-
20. Since these components had the same response to detector, their intensities in
spectra reflected their relative contents (see next section). From Figure 5.9, it can be
seen that intensity of components showed a normal distribution, statistically, which
181
matched the reaction chances of furfural with acetone. In this reaction, the longer the
chain was, the smaller the chance of production would be. Thus, the average carbon
number would be around 12, which was qualitatively given by the largest peak in
ESI-MS graphics, which was the product of 1 HMF + 2 acetone.
In order to verify the proposed furfural acetone condensation pathway, GC-
MS of pure HMF and furfural as feedstock were investigated. When 100 mg HMF
reacted with 1 mL acetone, we found that large amount of humin formed in 30 min at
120 C. Thus, loading of HMF and furfural was decreased to 10 mg, and the reaction
was conducted at 100 C for 30 min. The GC-MS results are show in Figures 5.8b and
5.8c. Furfural showed the peak of 1 furfural + 1 acetone with MW of 136. HMF also
showed the peak of 1 HMF + 1 acetone product. These peaks verified the formulas for
calculating products MW, proposed above.
5.3.8 Quantification of conversion by ESI-MS
The complexity and diversity of the hydrocarbon precursors formed in HDA
process made the quantification of the yield and selectivity of the process very
difficult. Fortunately, the furan rings in furfural or HMF kept unchanged during the
aldol-condensation; therefore the selectivity of sugars to the hydrocarbon precursors
could be estimated through quantifying the furan rings. Two methods (2D NMR and
182
oxidation, respectively) were developed for estimating reaction selectivity, as shown
in Figures 5.10 and 5.11, respectively.
(a)
(b)
183
(c)
Figure 5.10 Quantification of the hydrocarbon precursors from lignocelluloses using 2D-
NMR method (internal standard: pyrazine; (a) glucose; (b) xylose; and (c) spruce (softwood)).
The results of the sugar conversion and selectivity of sugar to the hydrocarbon
precursors of the HDA process are summarized in Table 5.7. Both glucose and xylose
had a conversion of ~95% and selectivity of 85-90%, respectively, which resulted in
an overall yield of ~80% from the sugars to the furan-derived hydrocarbon precursors.
When real biomass (softwood spruce) was used, ~85% sugar conversion and ~85%
selectivity were obtained, respectively, leading to an overall hydrocarbon precursor
yield of approximately 72%. The two methods gave comparable results, whereas the
2D NMR gave slightly higher values than the oxidation method, which might be
attributed to the over-integration of hydrogen peaks that did not belong to the furan
ring in the rectangle area of integration (Figure 5.10).
184
Figure 5.11 Mechanism of quantifying hydrocarbon precursors by oxidation method
Table 5.7 Quantitation of hydrocarbon precursors derived from carbohydrates
Feedstock Conversion (%)a
Selectivity (%)b
2D NMR Oxidation
Glucose 95 90 86
Xylose 93 91 83
Spruce 85 88 84 Note: (a) Conversion represents the molar percentage of converted (consumed) carbohydrates based on
initial carbohydrate; (b) Selectivity represents the molar percentage of the furans formed from the
converted carbohydrates.
Ooxone, NaHCO3
5~15oC
o o
O O
HO
O O
OAc
o
o
O
OAc
O O
NO2
O2N
NHH2N
OAc
N N
NO2
O2N
NH
NO2
O2N
HN
AcBr
N
NO2
O2N
NH
185
5.3.9 Carbon number distribution of hydrocarbon precursors
The carbon number and distribution of the hydrocarbon precursors were
estimated from positive ESI-MS spectra. Carbonyl group is easier to be protonated
than the hydroxyl or oxygen on furan in the hydrocarbon precursors because of its
larger proton affinity, as shown in Table 5.8. In particular, when a double bond
existed in the hydrocarbon precursors, the ketone would have a substantially higher
proton affinity than regular ketone because the formed positively-charged center
([M+1]+) could be stabilized through delocalization. As every hydrocarbon precursor
molecule was a ketone or double bond conjugated ketone (after dehydration), all the
hydrocarbon precursors were ionized in the same way and had the same response to
detector. Therefore, the ion intensity of each peak on the spectra reflected the relative
content of the responding hydrocarbon precursor compound. According to the ESI-
MS spectra shown in Figure 5.9, the carbon numbers of the identified hydrocarbon
precursors from glucose, xylose and spruce were estimated and summarized in Figure
5.12. It can be seen that the carbon numbers of most hydrocarbon precursors were
between 5 and 21, indicating that 0-3 acetone molecules were condensed to furan
rings. The precursor with more than 3 condensed acetone molecules were not detected.
The carbon numbers of the hydrocarbon precursors fell into the range of
186
transportation fuels. In addition, the carbon numbers seemed to follow a normal
distribution with C11/C12 the highest frequency.
Table 5.8 Proton affinity of different functional groups in hydrocarbon precursors
Functional groups Proton affinity (kJ·mol-1
)
R OH
760~800
(Jolly, 1991) O
~820
(Jolly, 1991)
O
~803
(Pan et al., 2010)
O
~870
(Bouchoux et al., 1988)
Note: Definition of proton affinity: the proton affinity, Epa, of an anion or of a neutral atom or
molecule is a measure of its gas-phase basicity. It is the energy released in the following
reactions: B + H+ → BH
+ (Jolly, 1991).
Figure 5.12 Carbon number distributions of hydrocarbon precursors from different
feedstocks.
C6
126
C9
126
C12
126
C15
126
C18
126
C21
126
C8
126
C11
126
C14
126
C17
126
C20
126
C9
126
C12
126
C15
126
C18
126
C5
126
187
5.3.10 Decomposition of lignin during HDA process
In many small-scale experiments above various types of lignocelluloses were
used as feedstock, no insoluble residue was observed at the end of reaction, implying
that lignin was completely soluble in acetone. In addition, the residue after CH2Cl2
extraction was far less than theoretical lignin content, indicating that part of lignin
was soluble in CH2Cl2. The evidence above suggests that the lignin be significantly
degraded during the HDA process as natural lignin is insoluble in acetone or CH2Cl2.
The acetone-soluble part of lignin was investigated for its molecular weight by
GPC. As seen from Figure 5.13, lignin was severely depolymerized into small
fragments with weight average molecular weight around 1000. In other words, the
lignin consisted of approximately only 4-5 monomeric units. The number molecular
weight of the lignin was around 400. Considering the mild condition (120-140 C) of
HDA process, there must have been a unique delignification mechanism involving
LiBr, mineral acid and acetone. This will be investigated in the future research. In
addition, the fraction soluble in CH2Cl2 are supposed to have even smaller molecular
weight than the acetone-soluble fraction. Further study will be conducted to try to
characterize the CH2Cl2-soluble lignin as well.
188
Elution time (min)
Figure 5.13 Gel permeation chromatograph of HDA lignin from spruce. Note: Mn represents number-average molecular weight; Mw represents weight-average molecular
weight
5.3.11 Reaction mechanism of glucose to HMF
The results above (Table 5.1) shows that both acid and Br- played a very
important role in catalyzing sugars to furfurals, which can also be seen from Figure
5.14. It was reported (Aida et al., 2007; Qian et al., 2005) that six-member ring of
glucose experienced a ring opening followed by a ring closing to form five-member
ring of fructose as intermediate in the dehydration process to form HMF. In order to
investigate whether Br- was involved in the dehydration of fructose, converting
fructose to HMF under different conditions was studied. Several test experiments
showed that fructose was dehydrated very quickly so that the reaction temperature
was lowered down to 75 C in 30 min to get a decent data set, as shown in Figure 5.15.
Retention time 21 min
Mw 1000
Mn 400
189
Note:other reaction condtion: 100 ºC, 1 mL acetone, 100 μL water, 100 mg glucose.
Figure 5.14 Dehydration of glucose to HMF derivative at different LiBr concentrations.
Note:other reaction condtion: 75 ºC, 1 mL acetone, 100 μL water, 100 mg fructose.
Figure 5.15 Dehydration of fructose to HMF derivative at different LiBr concentrations.
According to Arrhenius equation ( k = e−Ea RT ), the dehydration rate of
fructose was 100 times faster than dehydration of glucose. In addition, the dosage of
0
20
40
60
80
100
120
0 10 20 30 40 50 60
Le
ft g
luco
se
(%
)
Time (min)
300mg LiBr+0.25% acid
300mg LiBr+0.5% acid
600mg LiBr+0.25% acid
0
20
40
60
80
100
0 5 10 15 20 25 30
Left
fr
ucto
se
(%
)
Time (min)
300mg LiBr+0.25% acid
600mg LiBr+0.25% acid
190
LiBr did not have any significant effect on the dehydration of fructose. Therefore, in
the all reaction steps from glucose to HMF, the effect of the step from fructose to
HMF could be ignored. In other words, the formation of fructose intermediate from
glucose was the rate-determining step where Br- and H
+ were both involved. In kinetic
study, the rate of glucose dehydration was affected by concentrations of both acid and
LiBr.
According to these experiments, the mechanism of acetone-LiBr promoted
conversion of glucose to HMF was proposed in Figure 5.16. Glucose reacted with
acetone first to form a cyclic acetal. Since the O-C-O angle in the cyclic acetal was
not optimum for tetrahedron geometry, the tension force favored the opening of the
ring and would prevent the reverse reaction. In addition, Br- as a strong neucleophile
attacked the positively charged anomeric carbon, resulting in the ring opening to form
enol, which was the key intermediate in the isomerization of glucose. The enol then
closed ring and formed fructose, which was readily and rapidly dehydrated to HMF.
The cyclic acetal always formed at C1 and C2 positions, because the cyclic acetal
formed at other positions was unable to give a stable ring-opening structure due to the
tension force of formed cyclic acetal. Similar mechanism was proposed for boric acid
catalyzed glucose conversion into HMF where boric acid formed a cyclic intermediate
191
to facilitate the ring opening of glucose and stabilize the resultant ring-opening
structure (Stahlberg et al., 2011).
Figure 5.16 The proposed mechanism of glucose to HMF in acetone/LiBr system.
It is well known that furans (furfural and HMF) are instable and tend to
condense and form insoluble humins in acidic environment. It was reported that using
transition metals (e.g. CrClx) (Zhao et al., 2007) as catalyst and DMF, DMSO or ionic
liquid as solvent (Binder and Raines, 2009; Roman-Leshkov et al., 2006; Zhao et al.,
2007) instead of water could prevent HMF from the condensation (polymerization). In
this process, as excessive acetone was used in HDA process, newly formed
furfural/HMF molecules were isolated and surrounded by acetone molecules, which
prevented them from self-condensation. Therefore, the HDA environment favored the
aldol-condensation between furfural/HMF and acetone to form the hydrocarbon
precursors. In summary, acetone in HDA process had three functions: (1) as polar
192
aprotic reagent to promote the formation of free Br-, (2) as reactive reagent involved
in the aldol-condensation reaction with furans, and (3) as solvent to dissolve the
hydrocarbon precursors and lignin degradation products.
5.3.12 Recycling/recovery of solvents and LiBr
Recycling of solvents and LiBr is crucial to the success of the HDA process.
After the reaction, acetone was in a mixture along with hydrocarbon precursors, LiBr,
water, lignin, and, if any, unreacted/residual biomass/sugars. The unreacted acetone
could be easily recovered by evaporation after the reaction. Then water was added to
the mixture to facilitate the separation of LiBr from lignin, followed by the extraction
of the precursors with CH2Cl2. Fresh water could be further added to wash CH2Cl2
and lignin, if necessary. The addition of water allowed the formation of three phases:
LiBr/water aqueous phase, CH2Cl2/precursors organic phase, and lignin solid phase.
After centrifugation, three phases could be easily separated. The LiBr solution
(including HCl) could be reused in the next batch directly or after concentration, if
necessary. CH2Cl2 was recycled through vaporization, and hydrocarbon precursor was
left as thick liquid or solid. The recovery yields of LiBr and HCl were estimated by
titration with Ag+. When water/acetone (50:50) was used as reaction solvent, titration
indicated that 97.5% LiBr was reserved in final LiBr solution after extraction with
CH2Cl2. Washing CH2Cl2 and lignin fraction with 5 mL water could recover
193
additional 1.5% LiBr. Only approximately 1% of LiBr was non-recoverable, which
was probably dissolved in CH2Cl2 or absorbed on the precursors or lignin.
5.3.13 Hydrodeoxygenation of the precursors and lignin into
hydrocarbons
The hydrodeoxygenation of furan-derived hydrocarbon precursors and the
CH2Cl2-soluble low molecular weight lignin was conducted in ethanol with Pd/C and
SiO2-Al2O3 as catalyst. It might be unnecessary to achieve complete hydrogenation to
deoxygenate and saturate the hydrocarbon precursors, which is costly and needs a lot
of hydrogen gas. In other words, partial hydrodeoxygenation is probably enough to
make the hydrocarbon precursors miscible in hydrocarbon fuels. Preliminary results
indicated that both the hydrocarbon precursors and lignin were converted into hexane-
soluble products in 5 h at 250 C with initial hydrogen pressure of 6-7.5 MPa. The
images of the products in ethanol and suspended above water is shown in Figure 5.17.
For comparison, the precursor in ethanol before hydrodeoxygenation was given as
well. Because of the saturation of furan rings and double bonds in hydrocarbon
precursors, the color of products faded. The water-immiscibility and hexane-
miscibility of products indicated that the products were miscible with gasoline and
can be directly used as drop-in transportation fuel. It is anticipated that the
194
hydrodeoxygenated hydrocarbon precursors could be blended in gasoline even
without the removal of solvent ethanol.
Figure 5.17 Images of products before and after hydrodeoxygenation. (a) Hydrocarbon
precursors dissolved in ethanol before hydrodeoxygenation; (b) Hydrodeoxygenated
hydrocarbon precursors dissolved in ethanol after hydrodeoxygenation; (c) Hydrophobic
products were separated from ethanol by adding water; (d) Separated hydrophobic products
(upper layer in C) were miscible in hexane.
Further investigations are required into the optimization of
hydrodeoxygenation, characterization of the deoxygenated hydrocarbon precursors,
and their engine performance when blended with gasoline. It is worthy pointing out
that no coke formed during the hydrodeoxygenation. This is different from the
aqueous phase hydrodeoxygenation where two steps were needed and some coke
formed in the aqueous phase (Huber et al., 2005).
a b c d
195
5.4 Conclusion and recommendations
One-pot process (HDA) for converting lignocelluloses into hydrocarbon
precursors to fuel-grade hydrocarbons was developed. This process was carried out
under mild temperature (~120 °C) with inexpensive and recyclable catalyst (halide
salt) and could directly use real lignocelluloses as feedstock without pretreatment or
fractionation. Carbohydrates in the biomass can be readily converted into furan-
derived hydrocarbon precursors of C5-C21 to hydrocarbons in high yield and with
high selectivity (e.g., 72% for spruce). Lignin was extensively depolymerized during
the HDA process. Because of the very low molecular weight, the resultant HDA
lignin, in particular that from hardwood and herbage, had good potential to be
converted to chemicals and fuels. Not only the furan-derived hydrocarbon precursors
but also the low-molecular-weight fractions of the HDA lignin could be
hydrodeoxygenated (separately or jointly) into hydrocarbon fuels (or fuel additives).
As acetone is currently produced from petroleum and natural gas, renewable source of
acetone or other ketones should be pursued for HDA process in the future, for
example, from biomass through fermentation or pyrolysis process. The
hydrodeoxygenation of the hydrocarbon precursors and engine performance of the
resulting hydrocarbons need further investigation.
196
Reference
Achyuthan, K. E., Achyuthan, A. M., Adams, P. D., Dirk, S. M., Harper, J. C., Simmons, B.
A., and Singh, A. K. (2010). Supramolecular self-assembled chaos: polyphenolic
lignin's barrier to cost-Effective lignocellulosic biofuels. Molecules, 15(12), 8641-
8688.
Aden, A., Ruth, M., Ibsen, K., Jechura, J., Neeves, K., Sheehan, J., Lukas, J. (2002).
Lignocellulosic biomass to ethanol process design and economic utilizing co-current
dilute acid prehydrolysis and enzymatic hydrolysis for corn Stover, National
Renewable Energy Laboratory, Golden, CO, NREL/TP-510-32438.
Ahring, B. K., and Langvad, N. (2008). Sustainable low cost production of lignocellulosic
bioethanol - "The carbon slaughterhouse". A process concept developed by BioGasol.
International Sugar Journal, 110(1311), 184-190.
Aida, T. M., Sato, Y., Watanabe, M., Tajima, K., Nonaka, T., Hattori, H., and Arai, K. (2007).
Dehydration of D-glucose in high temperature water at pressures up to 80 MPa.
Journal of Supercritical Fluids, 40(3), 381-388.
Amarasekara, A. S., Williams, L. D., and Ebede, C. C. (2008). Mechanism of the dehydration
of D-fructose to 5-hydroxymethylfurfural in dimethyl sulfoxide at 150 degrees C: an
NMR study. Carbohydrate Research, 343(18), 3021-3024.
Balan, V., Bals, B., Chundawat, S. P., Marshall, D., and Dale, B. E. (2009). Lignocellulosic
biomass pretreatment using AFEX. Methods in Molecular Biology, 581, 61-77.
Balfanz, U., Rupp, M., and Baldauf, W. (1993). Upgrading of pyrolysis oil from biomass to
petroleum-products. Abstracts of Papers of the American Chemical Society, 205, 56-
Cell.
Banerjee, S., Mudliar, S., Sen, R., Giri, B., Satpute, D., Chakrabarti, T., and Pandey, R. A.
(2010). Commercializing lignocellulosic bioethanol: technology bottlenecks and
possible remedies. Biofuels Bioproducts and Biorefining, 4(1), 77-93.
Berti, P. J., and Tanaka, K. S. E. (2002). Transition state analysis using multiple kinetic
isotope effects: Mechanisms of enzymatic and non-enzymatic glycoside hydrolysis
and transfer. Advances in Physical Organic Chemistry, 37, 239-314.
Bienkowski, P. R., Ladisch, M. R., Voloch, M., and Tsao, G. T. (1984). Acid-hydrolysis of
pretreated lignocellulose from corn residue. Biotechnology and Bioengineering
Symposium Series, 14, 511-524.
197
Binder, J. B., and Raines, R. T. (2009). Simple chemical transformation of lignocellulosic
biomass into furans for fuels and chemicals. Journal of the American Chemical
Society, 131(5), 1979-1985.
Binder, J. B., and Raines, R. T. (2010). Fermentable sugars by chemical hydrolysis of
biomass. Proceedings of National Academy of Sciences of the United States of
America, 107(10), 4516-4521.
Bobletera, O., Schwalda, W., Concina, R., and Bindera, H. (1986). Hydrolysis of Cellobiose
in dilute sulpuric acid and under hydrothermal conditions. Journal of Carbohydrate
Chemistry, 5(3), 387-399.
Bond, J. Q., Alonso, D. M., Wang, D., West, R. M., and Dumesic, J. A. (2010). Integrated
catalytic conversion of gamma-valerolactone to liquid alkenes for transportation fuels.
Science, 327(5969), 1110-1114.
Boraston, A. B., Kwan, E., Chiu, P., Warren, R. A. J., and Kilburn, D. G. (2003). Recognition
and hydrolysis of noncrystalline cellulose. Journal of Biological Chemistry, 278(8),
6120-6127.
Bouchoux, G., Djazi, F., Houriet, R., and Rolli, E. (1988). Gas-phase basicity of olefinic C-5
and C-6 carbonyl-compounds. Journal of Organic Chemistry, 53(15), 3498-3501.
Bravo, V., Paez, M. P., Aoulad, M., and Reyes, A. (2000). The influence of temperature upon
the hydrolysis of cellobiose by b-1,4-glucosidases from Aspergillus niger. Enzyme
and Microbial Technology, 26, 614-620.
Brendler, E., Fisher, S., and Leipner, H. (2002). 7Li NMR as probe for solvent-cellulose
interactions in cellulose dissolution. Cellulose, 8, 283-288.
Cao, N. J., Xu, Q., Chen, C. S., Gong, C. S., and Chen, L. F. (1994). Cellulose hydrolysis
using zinc chloride as a solvent and catalyst. Applied Biochemistry and Biotechnology,
45(6), 521-530.
Carroll, A., and Somerville, C. (2009). Cellulosic Biofuels. Annual Review of Plant Biology,
60, 165-182.
Chandra, R. P., Bura, R., Mabee, W. E., Berlin, A., Pan, X., and Saddler, J. N. (2007).
Substrate pretreatment: The key to effective enzymatic hydrolysis of lignocellulosics.
Biofuels, 108, 67-93.
Chen, L. Y., and Yang, C. (1984). U.S patent Patent No. 4452640.
Chheda, J. N., Huber, G. W., and Dumesic, J. A. (2007). Liquid-phase catalytic processing of
biomass-derived oxygenated hydrocarbons to fuels and chemicals. Angewandte
Chemie-International Edition, 46(38), 7164-7183.
Coutinho, J. B., Gilkes, N. R., Kilburn, D. G., Warren, R. A. J., and Miller, R. C. (1993). The
nature of the cellulose-binding domain affects the activities of a bacterial
198
endoglucanase on different forms of cellulose. Fems Microbiology Letters, 113(2),
211-218.
Creagh, A. L., Ong, E., Jervis, E., Kilburn, D. G., and Haynes, C. A. (1996). Binding of the
cellulose-binding domain of exoglucanase Cex from Cellulomonas fimi to insoluble
microcrystalline cellulose is entropically driven. Proceedings of the National
Academy of Sciences of the United States of America, 93(22), 12229-12234.
Cuzens, W. A. F. a. J. (1998). U.S. patent Patent No. 5820687.
Cyr, N., Elofson, R. M., Ripmeester, J. A., and Mathison, G. W. (1988). Study of lignin in
forages and wood by C-13 CP/MAS NMR. 1. Some evidence of polymerization and
depolymerization. Journal of Agricultural and Food Chemistry, 36(6), 1197-1201.
Dadi, A. P., Varanasi, S., and Schall, C. A. (2006). Enhancement of cellulose saccharification
kinetics using an ionic liquid pretreatment step. Biotechnology and Bioengineering,
95(5), 904-910.
Davda, R. R., and Dumesic, J. A. (2004). Renewable hydrogen by aqueous-phase reforming
of glucose. Chemical Communications, (1), 36-37.
Dawsey, T. R., and Mccormick, C. L. (1990). The lithium lhloride/dimethylacetamide solvent
for cellulose-a literature review. Journal of Macromolecular Science-Reviews in
Macromolecular Chemistry and Physics, 30(3-4), 405-440.
Demirbas, A. (2005). Bioethanol from cellulosic materials: A renewable motor fuel from
biomass. Energy Sources, 27(4), 327-337.
Demirbas, M. F. (2009). Biorefineries for biofuel upgrading: A critical review. Applied
Energy, 86, 151-161.
Dence, C. W. (1992). The determination of lignin. Methods in lignin chemistry, 33-61.
Jiao, D., King, C., Grossfield, A., Darden, T. A., and Ren, P. (2006). Simulation of Ca2+
and
Mg2+
solvation using polarizable atomic multipole potential. Journal of Physical
Chemistry B, 110, 18553-18559.
Doughty, H. W. (1924). Mohr's method for the determination of silver and halogens in other
than neutral solutions. Journal of American Chemical Society, 46(12), 2707-2709.
Elliott, D. C., and Schiefelbein, G. F. (1989). Liquid hydrocarbon fuels from biomass.
Abstracts of Papers of the American Chemical Society, 34(4), 1160.
Fan, L., Lee, Y. H., and Gharpuray, M. (1982). The nature of lignocellulosics and their
pretreatments for enzymatic hydrolysis. Microbial Reactions, 23, 157-187.
Farrell, A. E., Plevin, R. J., Turner, B. T., Jones, A. D., O'Hare, M., and Kammen, D. M.
(2006). Ethanol can contribute to energy and environmental goals. Science,
311(5760), 506-508.
Fengel, D., and Wegener, G. (1984). Wood : chemistry, ultrastructure, reactions. Berlin ;
New York: W. de Gruyter.
199
Fischer, S., Leipner, H., Thummler, K., Brendler, E., and Peters, J. (2003). Inorganic molten
salts as solvents for cellulose. Cellulose, 10(3), 227-236.
Fu, Y., Lai, D. M., Deng, L., Li, J. A., Liao, B., and Guo, Q. X. (2011). Hydrolysis of
cellulose into glucose by magnetic solid acid. Chemsuschem, 4(1), 55-58.
Galbe, M., and Zacchi, G. (2002). A review of the production of ethanol from softwood.
Applied Microbiology and Biotechnology, 59, 618-628.
Ghose, T. K. (1987). Measurement of cellulase activities. Pure and Applied Chemistry, 59,
257-268.
Gierer, J. (1985). Chemistry of delignification. 1. General concept and reactions during
pulping. Wood Science and Technology, 19(4), 289-312.
Gilkes, N. R., Jervis, E., Henrissat, B., Tekant, B., Miller, R. C., Warren, R. A. J., and Kilburn,
D. G. (1992). The adsorption of a bacterial cellulase and its two isolated domains to
crystalline cellulose. Journal of Biological Chemistry, 267(10), 6743-6749.
Gilkes, N. R., Kilburn, D. G., Miller, R. C., Warren, R. A. J., Sugiyama, J., Chanzy, H., and
Henrissat, B. (1993). Visualization of the adsorption of a bacterial endo-beta-1,4-
glucanase and its isolated cellulose-binding domain to crystalline cellulose.
International Journal of Biological Macromolecules, 15(6), 347-351.
Girisuta, B., Janssen, L. P. B. M., and Heeres, H. J. (2006). A kinetic study on the
decomposition of 5-hydroxymethylfurfural into levulinic acid. Green Chemistry, 8(8),
701-709.
Grinbaum, B., Kogan, L., Barnea, E., Harel, G., Semiat, R., and Wahrmann, S. (2008). USA
Patent No. 7452520.
Gupta, R. C., and Sehgal, V. K. (1979). Effect of viscosity and molecular weight of lignin
phenol formaldehyde resin on the glue adhesion Strength of plywood. Holzforschung
Und Holzverwertung, 31(1), 7-9.
Hald, P. (1995). The behavior of alkali-metals in biomass conversion systems. Abstracts of
Papers of the American Chemical Society, 40(3), 753-757.
Hara, M., Kitano, M., Yamaguchi, D., Suganuma, S., Nakajima, K., Kato, H., and Hayashi, S.
(2009). Adsorption-enhanced hydrolysis of beta-1,4-glucan on graphene-based
amorphous carbon bearing SO3H, COOH, and OH groups. Langmuir, 25(9), 5068-
5075.
Hara, M., Suganuma, S., Nakajima, K., Kitano, M., Yamaguchi, D., Kato, H., and Hayashi, S.
(2010). Synthesis and acid catalysis of cellulose-derived carbon-based solid acid.
Solid State Sciences, 12(6), 1029-1034.
Harris, E. E., and Lang, B. G. (1947). Hydrolysis of wood cellulose and decomposition of
sugar in dilute phosphoric acid. Journal of Physical and Colloid Chemistry, 51(6),
1430-1441.
200
Harris, J. F. (1985). Two-stage, dilute sulfuric acid hydrolysis of wood: an investigation of
fundamentals. [Madison, Wis.]: U.S. Dept. of Agriculture, Forest Service, Forest
Products Laboratory.
He, H., Wang, H. Y., Zhang, C. B., and Wang, L. A. (2010). Preparation of magnetic
sulfonated carbon-based solid acid catalysts for the hydrolysis of cellulose. Acta
Physico-Chimica Sinica, 26(7), 1873-1878.
Hedjazi, S., Kordsachia, O., Patt, R., Latibari, A. J., and Tschirner, U. (2009). Alkaline
sulfite-anthraquinone (AS/AQ) pulping of wheat straw and totally chlorine free (TCF)
bleaching of pulps. Industrial Crops and Products, 29(1), 27-36.
Heeres, H. J., Girisuta, B., and Janssen, L. P. B. M. (2007). Kinetic study on the acid-
catalyzed hydrolysis of cellulose to levulinic acid. Industrial and Engineering
Chemistry Research, 46(6), 1696-1708.
Hendriks, A. T., and Zeeman, G. (2009). Pretreatments to enhance the digestibility of
lignocellulosic biomass. Bioresource Technology, 100(1), 10-18.
Hsuanyu, Y., and Laidler, K. J. (1985). Kinetics of the hydrolysis of cellobiose catalysed by
P-glucosidase. Canadian Journal of Biochemistry and Cell Biology, 63, 167-175.
Huang, A. A. (1975). Enzymatic hydrolysis of cellulose to sugar. Biotechnology and
Bioengineering(5), 245-252.
Huber, G. W., Chheda, J. N., Barrett, C. J., and Dumesic, J. A. (2005). Production of liquid
alkanes by aqueous-phase processing of biomass-derived carbohydrates. Science,
308(5727), 1446-1450.
Huber, G. W., Cortright, R. D., and Dumesic, J. A. (2004). Renewable alkanes by aqueous-
phase reforming of biomass-derived oxygenates. Angewandte Chemie-International
Edition, 43(12), 1549-1551.
Huber, G. W., and Dumesic, J. A. (2006). An overview of aqueous-phase catalytic processes
for production of hydrogen and alkanes in a biorefinery. Catalysis Today, 111(1-2),
119-132.
Jeong, S., Joo, S., Kim, H., Kim, S., and Jeong, H. (2009). Effect on two-step polishing
process of electrochemical mechanical planarization and chemical-mechanical
planarization on planarization. Japanese Journal of Applied Physics, 48(6).
Jolly, W. L. (1991). Modern Inorganic Chemistry (2nd ed.). New York: McGraw-Hill.
Kim, Y., Mosier, N. S., and Ladisch, M. R. (2009). Enzymatic digestion of liquid hot water
pretreated hybrid poplar. Biotechnology Progress, 25(2), 340-348.
Kirubakaran, V., Sivaramakrishnan, V., Nalini, R., Sekar, T., Premalatha, M., and
Subramanian, P. (2009). A review on gasification of biomass. Renewable and
Sustainable Energy Reviews, 13(1), 168-175.
201
Kobayashi, H., Komanoya, T., hara, K., and Fukuoka, A. (2010). Water-tolerant mesoporous-
carbon-supported ruthenium catalysts for the hydrolysis of cellulose to glucose.
Chemsuschem, 3, 440-443.
Kosan, B., Michels, C., and Meister, F. (2008). Dissolution and forming of cellulose with
ionic liquids. Cellulose, 15(1), 59-66.
Krassig, H., and Schurz, J. (2002). Ullmann's Encyclopedia of Industrial Chemistry (6 ed.).
Weinheim, Germany, Wiley-VCH.
Lai, D. M., Deng, L., Guo, Q., and Fu, Y. (2011). Hydrolysis of biomass by magnetic solid
acid. Energy and Environmental Science, 4, 55-58.
Lawrence J. Russo, J. (1999). U.S. patent Patent No. 5968362.
Lee, Y. Y., Xiang, Q., Pettersson, P. O., and Torget, R. (2003). Heterogeneous aspects of acid
hydrolysis of alpha-cellulose. Applied Biochemistry and Biotechnology, 105, 505-514.
Li, C. Z., Wang, Q., and Zhao, Z. K. (2008). Acid in ionic liquid: an efficient system for
hydrolysis of lignocellulose. Green Chemistry, 10(2), 177-182.
Li, C. Z., and Zhao, Z. K. B. (2007). Efficient acid-catalyzed hydrolysis of cellulose in ionic
liquid. Advanced Synthesis and Catalysis, 349(11-12), 1847-1850.
Limayem, A., and Ricke, S. C. (2012). Lignocellulosic biomass for bioethanol production:
Current perspectives, potential issues and future prospects. Progress in Energy and
Combustion Science, doi:10.1016/j.pecs.2012.03.002.
Lingaraju, B. P., Lee, J. Y., and Yang, J. Y. (2012). Process and utility water requirements for
cellulosic ethanol production processes via fermentation pathway. Environmental
Progress and Sustainable Energy. doi: DOI 10.1002/ep.11604
Lloyd, T. A., and Wyman, C. E. (2005). Combined sugar yields for dilute sulfuric acid
pretreatment of corn stover followed by enzymatic hydrolysis of the remaining solids.
Bioresource Technology, 96(18), 1967-1977.
Lu, Q., Zhu, X. F., Li, W. Z., Zhang, Y., and Chen, D. Y. (2009). On-line catalytic upgrading
of biomass fast pyrolysis products. Chinese Science Bulletin, 54(11), 1941-1948.
Lv, D. Z., Xu, M. H., Liu, X. W., Zhan, Z. H., Li, Z. Y., and Yao, H. (2010). Effect of
cellulose, lignin, alkali and alkaline earth metallic species on biomass pyrolysis and
gasification. Fuel Processing Technology, 91(8), 903-909.
Lynd, L. R., Weimer, P. J., van Zyl, W. H., and Pretorius, I. S. (2002). Microbial cellulose
utilization: Fundamentals and biotechnology. Microbiology and Molecular Biology
Reviews, 66(4), 739-739.
Mansfield, S. D., Mooney, C., and Saddler, J. N. (1999). Substrate and enzyme characteristics
that limit cellulose hydrolysis. Biotechnology Progress, 15, 804-816.
202
Marbe, A., Harvey, S., and Berntsson, T. (2006). Technical, environmental and economic
analysis of co-firing of gasified biofuel in a natural gas combined cycle (NGCC)
combined heat and power (CHP) plant. Energy, 31(10-11), 1614-1631.
Mccarter, J. D., and Withers, S. G. (1994). Mechanisms of enzymatic glycoside hydrolysis.
Current Opinion in Structural Biology, 4(6), 885-892.
Mccormick, C. L., Callais, P. A., and Hutchinson, B. H. (1985). Solution studies of cellulose
in lithium chloride and N,N-dimethylacetamide. Macromolecules, 18(12), 2394-2401.
McIntosh, S., and Vancov, T. (2010). Enhanced enzyme saccharification of Sorghum bicolor
straw using dilute alkali pretreatment. Bioresource Technology, 101(17), 6718-6727.
McLaren, A. D. (1948). Adhesion of high polymers to cellulose. Influence of structure,
polarity, and tack temperature. Journal of Polymer Science, 3, 652-662.
Medve, J., Karlsson, J., Lee, D., and Tjerneld, F. (1998). Hydrolysis of microcrystalline
cellulose by cellobiohydrolase I and endoglucanase II from Trichoderma reesei:
Adsorption, sugar production pattern, and synergism of the enzymes. Biotechnology
and Bioengineering, 59(5), 621-634.
Melillo, J. M., Reilly, J. M., Kicklighter, D. W., Gurgel, A. C., Cronin, T. W., Paltsev, S.,
Schlosser, C. A. (2009). Indirect Emissions from Biofuels: How Important?. Science,
326(5958), 1397-1399.
Miller, S., and Hester, R. (2007). Concentrated acid conversion of pine softwood to sugars.
Part I: Use of a twin-screw reactor for hydrolysis pretreatment. Chemical Engineering
Communications, 194(1), 85-102.
Modenbach, A. A., and Nokes, S. E. (2012). The use of high-solids loadings in biomass
pretreatment—A review. Biotechnology and Bioengineering, 109(6), 1430-1442.
Monavari, S., Galbe, M., and Zacchi, G. (2009). Impact of impregnation time and chip size on
sugar yield in pretreatment of softwood for ethanol production. Bioresource
Technology, 100(24), 6312-6316.
Mosier, N., Wyman, C., Dale, B., Elander, R., Lee, Y. Y., Holtzapple, M., and Ladisch, M.
(2005). Features of promising technologies for pretreatment of lignocellulosic
biomass. Bioresource Technology, 96(6), 673-686.
Mosier, N. S. (2007). Cellulosic ethanol—biofuel beyond corn. Purdue extension, ID 335.
Mosier, N. S., Wilker, J. J., and Ladisch, M. R. (2004). Rapid chromatography for evaluating
adsorption characteristics of cellulase binding domain mimetics. Biotechnology and
Bioengineering, 86(7), 756-764.
Muangrat, R., Onwudili, J. A., and Williams, P. T. (2010). Alkali-promoted hydrothermal
gasification of biomass food processing waste: a parametric study. International
Journal of Hydrogen Energy, 35(14), 7405-7415.
203
Nguyen, Q. A., Tucker, M. P., Keller, F. A., and Eddy, F. P. (2000). Two-stage dilute-acid
pretreatment of softwoods. Applied Biochemistry and Biotechnology, 84(6), 561-576.
Onda, A., Ochi, T., Kajiyoshi, K., and Yanagisawa, K. (2008). A new chemical process for
catalytic conversion Of D-glucose into lactic acid and gluconic acid. Applied
Catalysis A-General, 343(1-2), 49-54.
Onda, A., Ochi, T., and Yanagisawa, K. (2008). Selective hydrolysis of cellulose into glucose
over solid acid catalysts. Green Chemistry, 10(10), 1033-1037.
Paljevac, M., Primozic, M., Habulin, M., Novak, Z., and Knez, Z. (2007). Hydrolysis of
carboxymethyl cellulose catalyzed by cellulase immobilized on silica gels at low and
high pressures. Journal of Supercritical Fluids, 43(1), 74-80.
Palkovits, R., Tajvidi, K., Procelewska, J., Rinaldi, R., and Ruppert, A. (2010).
Hydrogenolysis of cellulose combining mineral acids and hydrogenation catalysts.
Green Chemistry, 12(6), 972-978.
Pan, X., Xie, D., Kang, K. Y., Yoon, S. L., and Saddler, J. N. (2007). Effect of organosolv
ethanol pretreatment variables on physical characteristics of hybrid poplar substrates.
Applied Biochemistry Biotechnology, 137(12), 367-377.
Pan, X. J., Arato, C., Gilkes, N., Gregg, D., Mabee, W., Pye, K., Saddler, J. (2005).
Biorefining of softwoods using ethanol organosolv pulping: Preliminary evaluation of
process streams for manufacture of fuel-grade ethanol and co-products.
Biotechnology and Bioengineering, 90(4), 473-481.
Pan, Y. J., Liu, P. Y., Hu, N., and Tu, Y. P. (2010). Ion-neutral complexes resulting from
dissociative protonation: fragmentation of alpha-furanylmethyl benzyl ethers and 4-
N,N-dimethylbenzyl benzyl ethers. Journal of the American Society for Mass
Spectrometry, 21(4), 626-634.
Pandey, M. P., and Kim, C. S. (2011). Lignin depolymerization and conversion: a review of
thermochemical methods. Chemical Engineering and Technology, 34(1), 29-41.
Pang, J., Wang, A., Zheng, M., and Zhang, T. (2010). Hydrolysis of cellulose into glucose
over carbons sulfonated at elevated temperatures Chemical Communications, 46,
6935-6937.
Pang, J. F., Wang, A. Q., Zheng, M. Y., and Zhang, T. (2010). Hydrolysis of cellulose into
glucose over carbons sulfonated at elevated temperatures. Chemical Communications,
46(37), 6935-6937.
Panicker, C. Y., Yarghese, H. T., Anto, P. L., and Philip, D. (2006). Potential dependent
SERS profile of sulfanilic acid on silver electrode. Journal of Raman Spectroscopy.,
37, 853-857.
Penque, R. A. (1977). U.S. patent Patent No. 4018620.
Pfaff, K. P., Paust, J., Hartmann, H. (1987). United States Patent No. 4681936.
204
Piccolo, C., and Bezzo, F. (2009). A techno-economic comparison between two technologies
for bioethanol production from lignocellulose. Biomass and Bioenergy, 33(3), 478-
491.
Qian, X. H., Nimlos, M. R., Johnson, D. K., and Himmel, M. E. (2005). Acidic sugar
degradation pathways. Applied Biochemistry and Biotechnology, 121, 989-997.
Qing, Q., Yang, B., & Wyman, C. E. (2010). Impact of surfactants on pretreatment of corn.
Bioresource Technology, 101(15), 5941-5951.
Ragauskas, A. J., Williams, C. K., Davison, B. H., Britovsek, G., Cairney, J., Eckert, C. A.,
Tschaplinski, T. (2006). The path forward for biofuels and biomaterials. Science,
311(5760), 484-489.
Ralph, J., and Landucci, L. L. (2010). Lignin and Lignans Advances in Chemistry, 137-243.
Boca Raton, FL: CRC press (Taylor and Francis group).
Ralph, S. (2005). Conundrums regarding 5-O-4-linkages in softwood MWLs. paper presented
at the 59th appita annual conference and exhibition : incorporating the 13th ISWFPC
(International Symposium on Wood, Fibre and Pulping Chemistry) Auckland, New
Zealand.
Remsing, R. C., Swatloski, R. P., Rogers, R. D., and Moyna, G. (2006). Mechanism of
cellulose dissolution in the ionic liquid 1-n-butyl-3-methylimidazolium chloride: a C-
13 and Cl-35/37 NMR relaxation study on model systems. Chemical Communications,
(12), 1271-1273.
Roman-Leshkov, Y., Chheda, J. N., and Dumesic, J. A. (2006). Phase modifiers promote
efficient production of hydroxymethylfurfural from fructose. Science, 312(5782),
1933-1937.
Sathitsuksanoh, N., Zhu, Z., Ho, T. J., Bai, M. D., and Zhang, Y. H. (2010). Bamboo
saccharification through cellulose solvent-based biomass pretreatment followed by
enzymatic hydrolysis at ultra-low cellulase loadings. Bioresource Technology,
101(13), 4926-4929.
Schall, C. A., Dadi, A. P., and Varanasi, S. (2006). Enhancement of cellulose saccharification
kinetics using an ionic liquid pretreatment step. Biotechnology and Bioengineering,
95(5), 904-910.
Scharlemann, J. P. W., and Laurance, W. F. (2008). Environmental science - How green are
biofuels?. Science, 319(5859), 43-44.
Schell, D. J., Farmer, J., Newman, M., and McMillan, J. D. (2003). Dilute sulfuric acid
pretreatment of corn stover in pilot-scale reactor - Investigation of yields, kinetics,
and enzymatic digestibilities of solids. Applied Biochemistry and Biotechnology, 105,
69-85.
205
Scheller, H. V., and Ulvskov, P. (2010). Hemicelluloses. Annual Review of Plant Biology, 61,
263-289.
Sendich, E., Laser, M., Kim, S., Alizadeh, H., Laureano-Perez, L., Dale, B., and Lynd, L.
(2008). Recent process improvements for the ammonia fiber expansion (AFEX)
process and resulting reductions in minimum ethanol selling price. Bioresource
Technology, 99(17), 8429-8435.
Gou, Z. and Zhu, S. (1998). Extraction of lithium from brine containing high concentration of
magnesium by tri-n-butyl phosphate dissolved in kerosene. Chinese Journal of
Chemical Engineering, 6, 124-129.
Shevchenko, S. M. (2000). Depolymerization of lignin in wood with molecular hydrogen
Iodide. Croatica Chemica Acta, 73(3), 831-841.
Shuai, L., Yang, Q., Zhu, J. Y., Lu, F. C., Weimer, P. J., Ralph, J., and Pan, X. J. (2010).
Comparative study of SPORL and dilute-acid pretreatments of spruce for cellulosic
ethanol production. Bioresource Technology, 101(9), 3106-3114.
Sierra, R., Granda, C., and Holtzapple, M. T. (2009). Short-term lime pretreatment of poplar
wood. Biotechnology Progress, 25(2), 323-332.
Smith, A. M. (2008). Prospects for increasing starch and sucrose yields for bioethanol
production. Plant Journal, 54(4), 546-558.
Stahlberg, T., Rodriguez-Rodriguez, S., Fristrup, P., and Riisager, A. (2011). Metal-free
dehydration of glucose to 5-(hydroxymethyl)furfural in ionic liquids with boric acid
as a promoter. Chemistry-a European Journal, 17(5), 1456-1464.
Suganuma, S., Nakajima, K., Kitano, M., Yamaguchi, D., Kato, H., Hayashi, S., and Hara, M.
(2008). Hydrolysis of cellulose by amorphous carbon bearing SO3H, COOH, and OH
groups. Journal of the American Chemical Society, 130(38), 12787-12793.
Sukumaran, R. K., Singhania, R. R., and Pandey, A. (2005). Microbial cellulases - Production,
applications and challenges. Journal of Scientific and Industrial Research, 64(11),
832-844.
Sun, Y., and Cheng, J. Y. (2002). Hydrolysis of lignocellulosic materials for ethanol
production: a review. Bioresource Technology, 83(1), 1-11.
Taherzadeh, M., and Karimi, K. (2007). Acid-based hydrolysis processes for ethanol from
lignocellulosic materials: a review. Bioresources, 2, 472-499.
Taherzadeh, M. J., and Karimi, K. (2008). Pretreatment of lignocellulosic wastes to improve
ethanol and biogas production: a review. International Journal of Molecular Sciences,
9, 1621-1651.
Takagaki, A., Tagusagawa, C., and Domen, K. (2008). Glucose production from saccharides
using layered transition metal oxide and exfoliated nanosheets as a water-tolerant
solid acid catalyst. Chemical Communications, (42), 5363-5365.
206
Takagaki, A., Tagusagawa, C., Hayashi, S., Hara, M., and Domen, K. (2010). Nanosheets as
highly active solid acid catalysts for green chemical syntheses. Energy and
Enviromental Sciences, 3, 82-93.
Ting Chia Huang, T. H. T. (1991). Study of extraction behavior of lithium chloride by tri-n-
butyl phosphate dissolved in kerosene. Journal of Chemical Engineering Data, 36,
231-235.
Tomas-Pejo, E., Oliva, J. M., & Ballesteros, M. (2008). Realistic approach for full-scale
bioethanol production from lignocellulose: a review. Journal of Scientific and
Industrial Research, 67(11), 874-884.
Tomme, P., Creagh, A. L., Kilburn, D. G., and Haynes, C. A. (1996). Interaction of
polysaccharides with the N-terminal cellulose-binding domain of Cellulomonas fimi
CenC. 1. Binding specificity and calorimetric analysis. Biochemistry, 35(44), 13885-
13894.
Turbak, A. F., Hammer, R. B., Portnoy, N. A., and Davies, R. E. (1977). Critical review of
cellulose solvent systems. Abstracts of Papers of the American Chemical Society, 173,
6.
Bukowskya H. and Uhlemanna E. (1993). Selective extraction of lithium chloride from brines.
Separation Science and Technology, 28, 1357-1360.
Van de Vyver, S., Geboers, J., Jacobs, P. A., and Sels, B. F. (2011). Recent advances in the
catalytic conversion of cellulose. Chemcatchem, 3(1), 82-94.
Van de Vyver, S., Peng, L., Geboers, J., Schepers, H., de Clippel, F., Gommes, C. J., Sels, B.
F. (2010). Sulfonated silica/carbon nanocomposites as novel catalysts for hydrolysis
of cellulose to glucose. Green Chemistry, 12(9), 1560-1563.
Vandam, H. E., Kieboom, A. P. G., and Vanbekkum, H. (1986). The conversion of fructose
and glucose in acidic media - formation of hydroxymethylfurfural. Starch-Starke,
38(3), 95-101.
Weimer, P. J., Dien, B. S., Springer, T. L., and Vogel, K. P. (2005). In vitro gas production as
a surrogate measure of the fermentability of cellulosic biomass to ethanol. Applied
Microbiology and Biotechnology, 67(1), 52-58.
West, R. M., Kunkes, E. L., Simonetti, D. A., and Dumesic, J. A. (2009). Catalytic conversion
of biomass-derived carbohydrates to fuels and chemicals by formation and upgrading
of mono-functional hydrocarbon intermediates. Catalysis Today, 147(2), 115-125.
Wingren, A., Galbe, M., and Zacchi, G. (2008). Energy considerations for a SSF-based
softwood ethanol plant. Bioresource Technology, 99(7), 2121-2131.
Wolfram Rudolph, M. H. B., and Cory C. Pye. (1995). Hydration of lithium ion in aqueous
solution. Journal of Physical Chemistry, 99, 3793-3797.
207
Wyman, C. E., Dale, B. E., Elander, R. T., Holtzapple, M., Ladisch, M. R., and Lee, Y. Y.
(2005). Coordinated development of leading biomass pretreatment technologies.
Bioresource Technology, 96(18), 1959-1966.
Xiao, Z. Z., Wang, P., Qu, Y. B., Gao, P. J., and Wang, T. H. (2002). Cold adaptation of a
mesophilic cellulase, EG III from Trichoderma reesei, by directed evolution. Science
in China Series C-Life Sciences, 45(4), 337-343.
Xu, Q., Singh, A., and Himmel, M. E. (2009). Perspectives and new directions for the
production of bioethanol using consolidated bioprocessing of lignocellulose. Current
Opinion in Biotechnology, 20(3), 364-371.
Yamaguchi, D., Kitano, M., Suganuma, S., Nakajima, K., Kato, H., and Hara, M. (2009).
Hydrolysis of cellulose by a solid acid catalyst under optimal reaction conditions.
Journal of Physical Chemistry C, 113(8), 3181-3188.
Yang, B., and Wyman, C. E. (2008). Pretreatment: the key to unlocking low-cost cellulosic
ethanol. Biofuels Bioproducts and Biorefining, 2(1), 26-40.
Yang, Y., Sharma-Shivappa, R., Burns, J. C., and Cheng, J. J. (2009). Dilute acid
pretreatment of oven-dried switchgrass germplasms for bioethanol Production.
Energy and Fuels, 23(7), 3759-3766.
Yasuda, S., and Murase, N. (1995). Chemical structures of sulfuric-acid Lignin. 12. Reaction
of lignin models with carbohydrates in 72-percent H2SO4. Holzforschung, 49(5), 418-
422.
Yoon, S. H., Ha, S. Y., Roh, H. G., and Lee, C. S. (2009). Effect of bioethanol as an
alternative fuel on the emissions reduction characteristics and combustion stability in
a spark ignition engine. Proceedings of the Institution of Mechanical Engineers Part
D-Journal of Automobile Engineering, 223(7), 941-951.
Zakzeski, J., Bruijnincx, P. C. A., Jongerius, A. L., and Weckhuysen, B. M. (2010). The
catalytic valorization of lignin for the production of renewable chemicals. Chemical
Reviews, 110(6), 3552-3599.
Zerbe, J. I., and Baker, A. J. (1988). Investigation of fundamentals of two-stage, dilute
sulfuric acid hydrolysis of wood. Madison, Wis.: Forest Products Laboratory.
Zhang, T. Y., and Wyman, C. E. (2011). Integrated production of levulinic acid and furfural
from cellulosic biomass. Abstracts of Papers of the American Chemical Society, 241.
Zhang, X., Jin, Q., Dai, L., and Yuan, S. (2011). Facile synthesis of hypercrosslinked resins
via chloromethylation and continuous condensation of simple aryl molecules. Bulletin
of Material Science, 34, 735-738.
Zhang, Y. H. P., Himmel, M. E., and Mielenz, J. R. (2006). Outlook for cellulase
improvement: Screening and selection strategies. Biotechnology Advances, 24(5),
452-481.
208
Zhang, Z. H., and Zhao, Z. B. K. (2009). Solid acid and microwave-assisted hydrolysis of
cellulose in ionic liquid. Carbohydrate Research, 344(15), 2069-2072.
Zhao, H. B., Holladay, J. E., Brown, H., and Zhang, Z. C. (2007). Metal chlorides in ionic
liquid solvents convert sugars to 5-hydroxymethylfurfural. Science, 316(5831), 1597-
1600.
Zhu, Z. G., Sathitsuksanoh, N., Vinzant, T., Schell, D. J., McMillan, J. D., and Zhang, Y. H. P.
(2009). Comparative study of corn stover pretreated by dilute acid and cellulose
solvent-based lignocellulose fractionation: enzymatic hydrolysis, supramolecular
structure, and substrate accessibility. Biotechnology and Bioengineering, 103(4), 715-
724.