Purdue University
Purdue e-Pubs
Open Access Dissertations
Theses and Dissertations
Spring 2015
Effect of maleic acid on the selectivity of glucose
and fructose dehydration and degradation
Ximing Zhang
Purdue University
Follow this and additional works at: http://docs.lib.purdue.edu/open_access_dissertations
Part of the Bioresource and Agricultural Engineering Commons, Chemical Engineering
Commons, and the Chemistry Commons
Recommended Citation
Zhang, Ximing, "Effect of maleic acid on the selectivity of glucose and fructose dehydration and degradation" (2015). Open Access
Dissertations. 609.
http://docs.lib.purdue.edu/open_access_dissertations/609
This document has been made available through Purdue e-Pubs, a service of the Purdue University Libraries. Please contact [email protected] for
additional information.
Graduate School Form 30
Updated 1/15/2015
PURDUE UNIVERSITY
GRADUATE SCHOOL
Thesis/Dissertation Acceptance
This is to certify that the thesis/dissertation prepared
By Ximing Zhang
Entitled
EFFECT OF MALEIC ACID ON THE SELECTIVITY OF GLUCOSE AND FRUCTOSE DEHYDRATION AND
DEGRADATION
For the degree of Doctor of Philosophy
Is approved by the final examining committee:
Nathan S. Mosier
Chair
Michael R. Ladisch
Abigail Engelberth
Mahdi M. Abu-Omar
To the best of my knowledge and as understood by the student in the Thesis/Dissertation
Agreement, Publication Delay, and Certification Disclaimer (Graduate School Form 32),
this thesis/dissertation adheres to the provisions of Purdue University’s “Policy of
Integrity in Research” and the use of copyright material.
Approved by Major Professor(s): Nathan S. Mosier
Approved by: Bernard Engel
Head of the Departmental Graduate Program
4/20/2015
Date
i
EFFECT OF MALEIC ACID ON THE SELECTIVITY OF GLUCOSE AND
FRUCTOSE DEHYDRATION AND DEGRADATION
A Dissertation
Submitted to the Faculty
of
Purdue University
by
Ximing Zhang
In Partial Fulfillment of the
Requirements for the Degree
of
Doctor of Philosophy
May 2015
Purdue University
West Lafayette, Indiana
ii
I dedicate this dissertation to my beloved wife and parents.
iii
ACKNOWLEDGEMENTS
I would like to express my deepest appreciation to my committee chair, Professor Nathan
S. Mosier. This dissertation would not have been possible without his vision,
encouragement and advice necessary for me to proceed through the doctoral program and
complete my dissertation. Dr. Mosier himself is the definition of excellence, and I always
regard him as a role model to me. He has been a strong and supportive advisor to me
throughout my graduate school career, but he has always given me great freedom to pursue
independent work.
Special thanks to my committee, Dr. Michael R. Ladisch, Dr. Abigail Engelberth and
Dr. Mahdi M. Abu-Omar, for their support, guidance and helpful suggestions to make me
a better researcher.
Members of LORRE, Dr. Eduardo Ximenes, Dr. Gozdem Kilaz, Linda Liu, Tommy Kreke,
Leyu Zhang, Jinsha Li, Seockmo Ku also deserve my sincerest thanks, their assistance and
friendship has meant more to me than I could ever express. I owe my gratitude to great
many those people who have made this dissertation possible and because of whom my
graduate experience has been one that I will cherish forever. It has been a great privilege
to spend several years in the Laboratory of Renewable Resources Engineering (LORRE)
at Purdue University, and its members will always remain dear to me.
iv
In addition, I want to extend my gratefulness to Dr. Weihua Xiao, Barron Hewetson,
Priya Murria and Yuan Jiang, without whose contribution I could not complete my work.
Last but not least, I would like to dedicate this work to my wife Ning and my parents whose
love and encouragement accompany along the journey. I owe them everything and wish I
could show them just how much I love and appreciate them.
v
TABLE OF CONTENTS
Page
LIST OF TABLES ............................................................................................................ vii
LIST OF FIGURES ......................................................................................................... viii
ABSTRACT
……………………………………………………………………………x
CHAPTER 1． INTRODUCTION .................................................................................... 1
1.1
Research background .................................................................................. 1
1.2
Problem statement ....................................................................................... 2
1.3
Statement of purpose ................................................................................... 3
1.4
Research questions ...................................................................................... 4
1.5
Organization of the dissertation .................................................................. 4
CHAPTER 2. CATALYTIC CONVERSION OF LIGNOCELLULOSE FOR
PRODUCTION OF HIGH VALUE CHEMICALS: A REVIEW ........... 6
2.1
Lignocellulose is an alternative to petroleum ............................................. 6
2.2
Lignocellulose utilizes as alternative energy source .................................. 8
2.3
A synergistic biorefinary ........................................................................... 12
2.4
Conclusions ............................................................................................... 23
CHAPTER 3. KINETICS OF MALEIC ACID AND ALUMINUM CHLORIDE
CATALYZED DEHYDRATION AND DEGRADATION OF
GLUCOSE ............................................................................................. 25
3.1
Introduction ............................................................................................... 25
3.2
Methodology ............................................................................................. 29
vi
Page
3.3
Results and Discussion .............................................................................. 32
3.4
Conclusions ............................................................................................... 46
CHAPTER 4. REACTION MECHANISM OF MALEIC ACID AND ALUMINUM
CHLORIDE CATALYZED GLUCOSE AND FRUCTOSE TO 5(HYDROXYMETHYL) FURFURAL AND LEVULINIC ACID IN
AQUEOUS MEDIA ............................................................................... 48
4.1
Introduction ............................................................................................... 48
4.2
Methodology ............................................................................................. 50
4.3
Results and Discussion .............................................................................. 57
4.4
Conclusions ............................................................................................... 71
CHAPTER 5. CONCLUSIONS AND RECOMMENDATION ................................... 73
REFERENCES …………………………………………………………………………..75
VITA……………………………………………………………………………………..89
PUBLICATIONS………………………………………………………………………...90
vii
LIST OF TABLES
Table ..............................................................................................................................Page
Table 3-1 Conditions for Kinetic Analysis of Hexose Dehydration, HMF Hydrolysis, and
Humin formation. .............................................................................................................. 32
Table 3-2 Reaction Datasets for Determining Kinetic Constants ..................................... 33
Table 3-3 Kinetic rate constant ......................................................................................... 34
Table 4-1 Conditions for Kinetic Analysis of Hexose Dehydration, HMF Hydrolysis, and
Humin ............................................................................................................................... 52
Table 4-2 Kinetic rate constant ......................................................................................... 56
Table 4-3 Reaction comparison at 24 min, 140 oC ........................................................... 61
Table 4-4 Kinetic constant selectivity............................................................................... 62
Table 4-5 Activation energy and Gibbs free energy ......................................................... 64
Table 4-6 Summery for NMR experiment ........................................................................ 69
Table 4-7 Mass spectrometer: complex formed under different condition....................... 71
viii
LIST OF FIGURES
Figure .............................................................................................................................Page
Figure 2-1 Current cellulosic ethanol production process. ................................................. 9
Figure 2-2 Structures for 5-hydroxymethylfurfural, levulinic acid, furfural .................... 13
Figure 2-3 Reaction scheme from lignocellulose to HMF, furfural and levulinic acid.. .. 14
Figure 2-4 Dehydration of glucose and fructose to HMF. ................................................ 16
Figure 2-5 β-O-4 ether linkage the most prevalent linkage in lignin. .............................. 21
Figure 2-6 Propose the novel roadmap for conversion of biomass to value-added
chemicals........................................................................................................................... 23
Figure 3-1 Reaction pathway for glucose upgrade to levulinic acid ................................ 27
Figure 3-2 Reactor heat up profile at various temperatures .............................................. 30
Figure 3-3 Reactant and product yields observed during the reaction.............................. 36
Figure 3-4 Selectivity of levulinic acid as a function of time. .......................................... 43
Figure 3-5 Mole distribution based on 6 min.................................................................... 44
Figure 4-1 Reaction pathway for glucose upgrade to levulinic acid ................................ 54
Figure 4-2 Reactant and product yields observed during the reaction with (a) AlCl3 and
HCl as catalyst at 140 oC; (b) AlCl3 and maleic acid as catalyst at 140 oC; (c) AlCl3 and
HCl as catalyst at 160 oC; (d) AlCl3 and maleic acid as catalyst at 160 oC. ...................... 60
ix
Figure
Figure 4-3
Page
13
C NMR (a) D-Glucose in D2O (b) D-Glucose and maleic acid in D2O (c)
Maleic acid in D2O (d) AlCl3 and glucose in D2O (e) Maleic acid, AlCl3 and glucose in
D2O (f) HCl, AlCl3 and glucose in D2O, pH=1.2. .......................................................... 68
Figure 4-4 Mass spectrometer for glucose mixing with maleic acid and AlCl3. ............ 70
x
ABSTRACT
Zhang, Ximing. Ph.D., Purdue University, May 2015. Effect of maleic acid on the
selectivity of glucose and fructose dehydration and degradation. Major Professor: Nathan
Mosier.
5-Hydroxymethyfurfural (HMF), a platform chemical can upgrade to a variety of fuels
and polymers, can be manufactured from lignocellulose. This study focuses on the Lewis
and Brønsted acid effect on hexose dehydration for HMF production. We report the
positive effect of maleic acid, a dicarboxylic acid used as Brønsted acid, on the selectivity
of hexose dehydration to 5-hydroxymethyfurfural (HMF), and subsequent hydrolysis to
levulinic and formic acids. We also describe the kinetic analysis of a Lewis acid (AlCl3)
alone and in combination with HCl or maleic acid to catalyze the isomerization of glucose
to fructose, dehydration of fructose to HMF, hydration of HMF to levulinic and formic
acids, and degradation of these compounds to humins. Results show that AlCl3 significantly
enhances the rate of glucose conversion to HMF and levulinic acid in the presence of both
maleic acid and HCl. In addition, the degradation of HMF to humins, rather than levulinic
and formic acids, is reduced by 50% in the presence of maleic acid and AlCl3 compared to
hydrochloric acid combined with AlCl3. The results suggest a different reaction mechanism
for the dehydration of glucose and rehydration of HMF between maleic acid and HCl.
xi
Further elevated temperature (140-180 ℃) experiment demonstrates the maleic acid
alone behaves like Lewis acid to isomerization glucose to fructose. Maleic acid also found
facilitating glucose ring open reaction. Compared to HCl combined with AlCl3, calculated
activation energy justifies maleic acid can lower the isomerization step activation energy
when combined with AlCl3.
1
CHAPTER 1． INTRODUCTION
The overall aim of the work presented in this dissertation is to develop a process on
upgrading glucose to 5-hydroxymethylfurfural (HMF) by using maleic acid combined with
aluminum chloride. The novelty of this study is maleic acid changes the reaction pathway
and selectivity of HMF from glucose, which is significantly improved compared to HCl.
The high selectivity of maleic acid combined with AlCl3 and the effect of inhibition on the
byproduct humins generation, makes maleic acid a promising catalyst for conversion
hexose to value added chemicals.
1.1
Research background
Petrochemical industry was established and commercialized for production of
transportation fuel and chemicals for over a century (Boisen et al., 2009). Using crude oil
as feedstock has issues such as high pollution to environment, unsustainable and carbon
dioxide emission, which will intensify the global greenhouse effect (Werpy et al., 2004).
For the environmental and energy security concerns, renewable energy resources such as
wind energy, solar energy and bioenergy are drawing more and more attention since the
late 20th century. As a replacement or supplement to fossil fuels, lignocellulose is unique
since it is renewable and neutral to carbon emission feedstock, which can also generate
transportation fuels and bulk and fine chemicals (Murzin & Salmi, 2012).
2
Lignocellulose, mainly referring to forest resources and agricultural residues, is
primarily composed of cellulose (~40%wt), hemicellulose (~30%wt) and lignin (~20%wt),
extractives and inorganic components (Murzin & Salmi, 2012). The lignin is poly-phenolic
material and cellulose and hemicellulose mainly consist subunits glucose and xylose,
respectively (Murzin & Salmi, 2012). Current research shows that fractionation of
lignocellulose can generate different chemical and fuel streams. C6 and C5 sugars from
cellulose and hemicellulose can be upgraded to value-added chemicals, which can further
use for polymer building blocks and drop-in fuel (vom Stein et al., 2011). A good example
is production of 5-hydroxymethylfurfural (HMF) from glucose. HMF is regarded as a
platform building-block for producing furanic polyesters, polyamides, and polyurethanes
analogous to those derived from the petroleum polymer industry (Pagan-Torres, Wang,
Gallo, Shanks, & Dumesic, 2012). It can be produced from lignocellulose through a
sequential series of catalytic conversions: first is hydrolysis of cellulose to release glucose;
second is isomerization of glucose to fructose; and finally is dehydration of fructose to
make HMF.
1.2
Problem statement
Even though the production of HMF from lignocellulose is attractive, there are obstacles
to overcome. Because fructose is favored over glucose by the Brønsted acid catalyzed
dehydration reaction to make HMF, Lewis acid catalyzed isomerization of glucose to
fructose can improve reaction rates and yields (Choudhary, Mushrif, et al., 2013). However,
this increases the complexity of the catalytic system. Various Lewis acids such as AlCl3,
SnCl4, and VCl3 are effective in isomerizing glucose to fructose (Pagan-Torres et al., 2012).
Abu-Omar’s research group reported using AlCl3 as Lewis acid combined with HCl can
3
catalyze glucose achieved the yield of HMF of 61% using biphasic system (Y. Yang et al.,
2012). However, the insoluble polymeric byproduct, humin, requires special attention as it
can introduce impurities that must be removed in downstream recovery of HMF, deactivate
the catalysts, and even clog the reactor. Humin is a heterogeneous and carbonaceous
product from interpolymerization of hexose and HMF (Hayes & Hayes, 2009).
1.3
Statement of purpose
To minimize the impact of this side reaction, reaction conditions such as acid
concentration and temperatures should be tested and optimized. On the other hand, the
behavior and effect of a Brønsted acid combined with a Lewis acid (such as AlCl3) needs
further study to understand the mechanism of the reactions and how catalyst structure
affects product selectivity. Strong Brønsted acids, commonly used in industrial processes,
are corrosive and can generate large amounts of inorganic waste (Kootstra, Mosier, Scott,
Beeftink, & Sanders, 2009). There is not much study on using organic acid combined with
AlCl3 on catalyzing hexose isomerization and dehydration. Unlike strong Brønsted acids,
weak organic acids, such as maleic acid, are less corrosive with nontoxic end products
(CO2, formic acid and fumaric acid) (Lu & Mosier, 2007, 2008; N. S. Mosier, Ladisch, &
Ladisch, 2002).
Maleic acid as a biomimetic catalyst can mimic the structure of the active site of cellulase
enzymes (N. S. Mosier, Ladisch, & Ladisch, 2002). Previous studies show that maleic acid
has superior selectivity compared to sulfuric acid for the hydrolysis of cellulose and
hemicellulose to monomeric sugars (Lu & Mosier, 2007). In the study of the catalytic
conversion of xylose to furfural, maleic acid was found not follow the specific acid catalyst
mechanism or general acid mechanism (Lu & Mosier, 2008). Both of the studies show that
4
maleic acid has a protective action which inhibits the degradation reactions. This may be
explained by the dicarboxylic acid structure of maleic acid that forms an internal hydrogen
bond with transition-state intermediates, resulting in increased activation energy barrier for
further degradation (Lu & Mosier, 2008).
With many advantages compared to strong Brønsted acids, it is worth conducting a
systematic study to explore the mechanism for Lewis acid (AlCl3) combined with different
type of Brønsted acids (HCl and maleic acid) to produce HMF.
1.4
Research questions
In this dissertation, we propose to answer the following questions:
(1) How does catalysts selection affect the kinetic parameters?
(2) How would different catalyst systems behave? This can be understood by calculating
the catalyst selectivity and testing the reaction under elevated temperatures for different
catalyst combinations.
(3) What is the reaction mechanisms followed by maleic acid and HCl when they are used
alone and combined with Lewis acid on hexose to HMF conversion? Multiple instruments
will be used to understand the reaction mechanisms.
1.5
Organization of the dissertation
This dissertation will propose a novel catalysis mechanism using maleic acid as Brønsted
acid combined with AlCl3 in transformation of hexose to make HMF and levulinic acid.
The investigation in maleic acid function in aqueous solution is tested in elevated
temperatures and by multiple methods. The chapter two will give a literature review on
using lignocellulose to generate different value chemical streams. After review the current
utilization of lignocellulose by cellulosic ethanol industry, a novel roadmap is proposed for
5
lignocellulose utilization by removal lignin prior to conversion carbohydrates in the
lignocellulose. The chapter three will describe the kinetic study of maleic acid combined
with AlCl3 to catalyze the dehydration and degradation of glucose. Multiple comparisons
were made between Brønsted acids (HCl and maleic acid) or one Lewis acid (AlCl3) used
alone and in combination. The results show different pathways that maleic acid and HCl
may follow. Chapter four presents data and analysis from varying reaction temperatures
where maleic acid combined with AlCl3 and HCl combined with AlCl3 were used to
isomerize and dehydrate glucose. Further reaction mechanism analysis was done by using
NMR and mass spectrometer to determine the glucose ring open reaction and chelated
speciation in aqueous solution. Finally, the novel character of maleic acid is confirmed.
Maleic acid acts similar to Lewis acids in facilitating glucose chain open in aqueous
solution while also stabilizing the acyclic form glucose in solution. Chapter five will
summarize the research presented in the dissertation and give the future recommendation.
6
CHAPTER 2. CATALYTIC CONVERSION OF LIGNOCELLULOSE FOR
PRODUCTION OF HIGH VALUE CHEMICALS: A REVIEW 1
2.1
Lignocellulose is an alternative to petroleum
A shift from crude oil to renewable resources as a basis for the chemical industry is
driven by several impetuses: more sustainable and carbon neutral society, minimizing the
uncertainty of crude oil price, and developing new products with improved performance
and functionality (T. H. Parsell et al., 2013). Lignocellulosic materials have enormous
potential as feedstocks for the emerging green chemical industry. Lignocellulose has an
annually production about 10-50 billion tons of dry lignocelluolosics, accounting half of
the global biomass yield (Claassen et al., 1999; Galbe & Zacchi, 2002; X. Zhao, Zhang, &
Liu, 2012). The sufficiency and non-competition with increasing global food demand give
lignocelluolosics the potential to be a competitive candidate in production of transportation
fuel and value-added chemicals and polymers, as complement to the current crude oil
refineries (Kamm & Kamm, 2007).
2.1.1
Component in lignocellulose
Lignocellulose refers to cell wall tissues of forest and agricultural residuals, and
energy crops. Cellulose, hemicellulose and lignin are the three main components of
1
Chapter 2 is adapted from the manuscript “Catalytic Conversion of Lignocellulose for
Production of High Value Chemicals: a Review”, which is in preparation for submission
to Journal of Green Chemistry.
7
lignocellulosic materials (Rowell, 2012). Carbohydrates in lignocellulose account for
around 60-70% by weight (Gellerstedt & Henriksson, 2008). For the annual lignocellulosic
production on earth, carbohydrates including cellulose and hemicellulose are the most
abundant in amount (Kaplan, 1998). Industrially, most of the carbohydrates in
lignocellulose are used for pulp and paper industry (mainly woody material) as well as for
biofuel especially bioethanol production (mainly corn stover and other agricultural residues)
(Zakzeski, Bruijnincx, Jongerius, & Weckhuysen, 2010). Lignin content in softwood and
hardwood tree can vary between 15% to 40% by weight and accounts for around 40% of
energy content of lignocelluolosics (Gellerstedt & Henriksson, 2008). For the annual
lignocellulosic production on earth, lignin is ranked second following cellulose in amount
(Kaplan, 1998). Currently, over 50 million tons of lignin are produced in pulp and paper
industry and only a small portion of these are used as dispersing or binding agents. Largely,
lignin is burned as fuel (Zakzeski et al., 2010). Most production scale cellulosic ethanol
biorefinaries use lignin as a source for boiler fuel to generate steam for heat and possibly
electricity for the plant (Vorotnikova & Seale Jr, 2014).
Cellulose accounting for 40-50% of dry weight of biomass (N. Mosier et al., 2005;
Rowell, 2012). The cellulose builds up by linear polymer of D-glucose subunit linked by
β-1, 4 glycosidic bond. Cellulose mainly forms in crystalline structure with a small extent
of amorphous structure.
Hemicellulose accounts for 23-35% of dry weight biomass. Hemicellulose is mostly
random branched polymer consisting typically of xylose monomers along with other C5
and C6 sugars. Monomeric subunits (such as D-xylose, D-mannose, D-galactose, D-
8
glucose, L-arabinose, D-galacturonic, D-glucuronic and 4-O-methyl-glucuronic acids)
consists the hemicellulose polysaccharide.
The lignin is physically and/or chemically linked to both cellulose and hemicellulose to
form a physical seal in the plant cell wall which would favor the plant strength and stiffness
(Ritter, 2008). The phenyl propanoid subunits of lignin, are covalently linked to form an
amorphous heteropolymer that is not water soluble (Pérez, Munoz-Dorado, de la Rubia, &
Martinez, 2002). Lignin has complex structures; the lignin molecule is random
polymerized by mainly three subunits: p-coumaryl alcohol (hydroxyphynyl unit), coniferyl
alcohol (guaiacyl unit) and sinapyl alcohol (sinapyl unit). The random cross-linked
subunits build up the amorphous three-dimensional lignin molecule biologically by
enzyme-catalyzed oxidation of phenylpropane unit (Freudenberg & Neish, 1968).
2.2
Lignocellulose utilizes as alternative energy source: cellulosic bioethanol
For the past few decades, numerous efforts are focusing on biochemical and
thermochemical conversion of lignocellulose to liquid transportation fuels and chemicals
(Jönsson, Alriksson, & Nilvebrant, 2013). The most prevalent utilization is conversion of
lignocellulose to bioethanol. Bioethanol is of great importance to complement the US oil
market (Lichts, 2011). Brazil and the United States lead the industrial production of ethanol
fuel, accounting together for 87.8 percent of the world's production in 2010 (Lichts, 2011;
McMichael, 2009), and 87.1 percent in 2011 (Vorotnikova & Seale Jr, 2014). Currently,
nearly all of the cellulosic bioethanol produced is made through a biocatalytic process:
pretreatment disrupts the lignin and plant cell wall structure of lignocellulose; enzymatic
hydrolysis releases fermentable sugars from the cellulose and hemicellulose, and finally
bacterial or yeast fermentation converts sugars to ethanol (Figure 2-1).
9
Figure 2-1 Current cellulosic ethanol production process.
2.2.1
Pretreatment and hydrolysis to release monosaccharides
To make the process profitable，steps before sugar fermentation is challenging as the
lignocellulosic structure is recalcitrant and complex. Pretreatment processes aim to
increase the surface area and reducing the crystallinity of lignocellulose to enhance the
following hydrolysis yields of cellulose and hemicellulose to monosaccharides (da Costa
Sousa, Chundawat, Balan, & Dale, 2009; N. Mosier et al., 2005; Severian, 2008).
Pretreatment technologies mainly fall into four categories: biological pretreatment,
physical
pretreatment,
chemical
pretreatment,
physic-chemical
pretreatment.
Hemicellulose and cellulose can hydrolyze to C5 and C6 sugars by either strong inorganic
acids or hydrolytic enzymes. Enzymatic hydrolysis is more selective; however, using this
method would require narrowed operating temperatures, maintained pH by using buffer,
10
and strict feed purification requirements, which would substantially increase the processing
cost. Meanwhile the chemical hydrolysis is relatively cheap, albeit with lower selectivity
due to the severe side reaction of monosaccharides in the presence of chemical catalysts
(Bhosale, Rao, & Deshpande, 1996; El Khadem, Ennifar, & Isbell, 1989; Rinaldi & Schüth,
2009; B. Y. Yang & Montgomery, 1996). Both the C5 and C6 sugars derived through the
hydrolysis process can then fermented by yeast or bacteria to desired product such as
ethanol or butanol.
2.2.2
Techno-economic drawbacks for cellulosic ethanol: inhibitors
During the pretreatment, especially under acidic environment, which is commonly used
in the current and upcoming commercial demonstration facilities (Poet’s plant in
Emmetsburg, IA; DuPont’s plant in Nevada, IA; Abengoa’s plant in Hugoton, KS) (Brown
& Brown, 2013); by-products generated through the pretreatment step can severely inhibit
on the following enzymatic hydrolysis and fermentation. The inhibitors can be classified
as mainly four categories: phenolic compounds, weak acid, furan derivatives and inorganic
compounds (Casey, Sedlak, Ho, & Mosier, 2010).
The phenolic compounds and related aromatics released from lignin have the inhibition
and deactivation effects on cellulolytic enzymes, which requires increased enzyme doses
to achieve high conversion yields. The LORRE research group from Purdue University
conducted indepth studies on the deactivation of enzyme caused by phenols. Results
showed that phenols are major inhibitors and deactivators of cellulolytic enzymes
(Ximenes, Kim, Mosier, Dien, & Ladisch, 2011). They proposed two ways to minimize
the phenolic compounds’ deactivation and inhibition effect in order to have higher glucose
11
yield meanwhile with lower enzyme loading: use of lignin-free cellulose or prevention of
cellulase adsorption on lignin.
Another group of inhibitors is mainly aliphatic acid, referring to acetic acid, formic acid
and levulinic acid. All of these three aliphatic acids relate to sugar degradation during the
pretreatment. Formic acid is the hydrolysis product from HMF (5-hydroxymethylfurfural)
and furfural. Levulinic acid is a degradation product from HMF. The yeast S. cereviae
commonly used for fermentation can be deactivated by binding with all these acids. Acetic
acid is a weak acid generated by hydrolysis of acetyl groups of hemicellulose (Palmqvist
& Hahn-Hägerdal, 2000). Acetic acid has negative impact on fermentative performance of
microorganisms by inhibition of biomass growth, substrate consumption and ethanol
volumetric productivity (Casey et al., 2010). The furan derivatives mainly refer to HMF
(degraded from glucose) and furfural (degraded from xylose). The inhibition caused by
these compounds is similar to acetic acid, including slowed growth rate of yeast and
decreased ethanol yield and productivity (Chung & Lee, 1985; Larsson et al., 1999; Z. Liu
et al., 2004).
Ten to one hundred times of theoretical enzyme loading have to be used to gain high
sugar yield to compensate the inhibition effect (Wyman, 2013). The cellulase dose in order
to have high sugar yield costs around $1.00/gal ethanol ($1.50/gal equivalent gasoline)
(Klein‐Marcuschamer, Oleskowicz‐Popiel, Simmons, & Blanch, 2012), which further
weaken the profitability of cellulosic ethanol specially when the gasoline price is low. To
deal with this issue, many scientists are focusing on mitigating pretreatment side effect on
enzyme and method to reuse the enzyme, on the other hand, adaptation of enzyme to make
it more resistant to inhibitor and lower production cost are research hotspot. However, there
12
is limited scope to lower the production cost for cellulosic ethanol if the product is with
low value.
In summary, the inhibitors released by the pretreatment step can significantly reduce the
yield of alcohol and make the enzymatic hydrolysis step more expensive. Coupled with the
high capital costs associated with pretreatment and handling the low bulk-density
lignocellulose, the increased costs associated with higher enzyme use and lower
fermentation productivity continue to be the most significant economic hurdles to
expanding cellulosic ethanol production (Wyman, 2007). It will be a shortcut for
biorefinary if the lignocellulose selective upgrades by component similar to fractionation
in petrochemical industry. Lignin, cellulose and hemicellulose can generate similar high
value chemical with high functionality derived from petroleum refinery, though it would
took an extensive effort to develop processes and catalysts.
2.3
A synergistic biorefinary: sequentially catalytic conversion carbohydrates and
lignin
This review focuses on emerging strategy to synergistic conversion of lignocellulose
subunits to valuable chemicals; aim to release the application potential of lignocellulose.
Both the carbohydrate utilization (mainly C6 sugar) and lignin utilization technology is
discussed and summarized.
2.3.1
Platform molecules derived from carbohydrates in lignocellulose
Unlike the technology used in cellulosic bioethanol, there is alternative way to
conversion of lignocellulose to platform chemicals through subsequential conversion of
lignocellulosic component. In 2004, US Department of Energy proposed 12 promising
platform molecules derived from lignocellulose, in which includes furans (furfural, 5-
13
hydroxymethylfurfural) and acids (succinic, levulinic, lactic acids) (Werpy et al., 2004).
HMF, furfural and levulinic acid structures are shown in Figure 2-2. For the HMF, the
functional groups located at 2, 5-position of the furanic molecule, a hydroxide and aldehyde,
make HMF chemical structure attractive. The unique structure either can be oxidized to a
dicarboxylic acid or reduced to a diol, both of which can be used for the synthesis of
polymers (furanic polyesters, polyamides, and polyurethanes) (Wang, Nolte, & Shanks,
2014). Further, the unsaturated aromatic compound not only can upgrade to fuel molecules
via hydrogenation but also has pharmaceutical application as it is in an array of biologically
active molecules (T. Wang et al., 2014). Levulinic acid, rehydrated from HMF, could be
readily transformed to product such as acrylate polymers and fuel additives such as γvalerolactone, 2-methyl-tetrahydrofuran, and ethyl levulinate (Chundawat, Beckham,
Himmel, & Dale, 2011). Furfural can be used as precursor molecule for solvent synthesis
such as THF and MeTHF (Kim, Liu, Abu-Omar, & Mosier, 2012).
Figure 2-2 Structures for 5-hydroxymethylfurfural (HMF), levulinic acid (LA), furfural.
14
Furfural and HMF can be directly converted from xylose and glucose, which are
monomeric subunits of hemicellulose and cellulose. The reaction scheme of cellulose and
hemicellulose separately conversion to HMF and furfural is shown in the Figure 2-3.
Figure 2-3 Reaction scheme from lignocellulose to HMF, furfural and levulinic acid.
Adapted from Reference (Climent, Corma, & Iborra, 2014).
The transformation of cellulose to HMF and LA is mainly through three steps:
depolymerization of polysaccharides through a hydrolysis step to monosaccharides, and
then loss of three molecules water through dehydration of monosaccharides; Rehydration
of HMF would add one mole of water to HMF to form LA. To favor the reaction towards
to form HMF and LA with high selectivity, an isomerization step is needed for transform
glucose to fructose, as fructose is much easier to dehydrate to make HMF compared to
glucose (L. Wang et al., 2014). For conversion of hemicellulose to furfural, the
hemicellulose need to hydrolysis to make xylose and dehydrated under acidic condition to
make furfural.
15
2.3.1.1 Reaction systems for HMF and levulinic acid production
To facilitate the reaction mentioned, different catalytic systems have been proposed. The
catalyst system mainly falls into three categories in reaction media: aqueous, organic and
ionic liquid media; and two systems in catalyst: mainly homogeneous (mineral acid, weak
organic acid, Lewis acidic metal halides, Brønsted acidic ionic liquids) and heterogeneous
catalysts (Dutta, De, & Saha, 2012; F. Liu et al., 2014; Zakrzewska, Bogel-Łukasik, &
Bogel-Łukasik, 2010; Zhang, Hewetson, & Mosier, 2015). The pros and cons for different
catalytic system is analyzed in detail in review by Saha and Abu-omar (Saha & Abu-Omar,
2014). Despite the building block mentioned above has many potential use for replacement
for petroleum feedstock, there are obstacles in lowering production cost (e.g. ionic liquids
are expensive) and solving separation issues (e.g. monophasic solvent system generally use
high boiling point organic solvents). For example, the reported high HMF conversion yield
by MClx in ionic liquids is significant higher than competing catalyst system, however, the
commercialization of this technology must face the expensive ionic liquid cost and stability
and purity requirement of the ironic liquid (H. Zhao, Brown, Holladay, & Zhang, 2012).
Overall, biphasic reaction system for production of HMF is promising as it can be used in
both batch and continuous biphasic reactors with less side reaction (Saha & Abu-Omar,
2014).
2.3.1.2 Reaction mechanism: dehydration from hexose
The reaction mechanism for production of HMF from lignocellulose must trace back to
the dehydration chemistry from hexose. Among hexose sugars, both glucose and fructose
are widely studied for dehydration to make HMF; however, no conclusive reaction pathway
16
can be determined and applied to glucose or fructose. The reasons to that are complicated:
under different temperatures, solvent systems and catalyst combinations, the reaction for
dehydration may differ a lot. The side reaction can generate humins and other organic acids
(Kuster, 1990; Ranoux, Djanashvili, Arends, & Hanefeld, 2013). Humins is polymerized
product of hexose with HMF and other intermediates (van Zandvoort et al., 2013). Some
parallel pathways for humins generation includes dehydration forming non-furan cyclic
ethers and condensation reaction (Zheng, Fang, Cheng, & Jiang, 2010). The organic acids
such as levulinic acid and formic acid can self-catalyze the reaction (Ranoux et al., 2013).
All these factors result in the various reaction mechanism proposed.
Figure 2-4 Dehydration of glucose and fructose to HMF.
Fructose as reactant to make HMF is known as a Brønsted acid catalyzed dehydration
reaction (T. Wang et al., 2014). Fructose first studied using mineral, organic, resins and
metal oxides in monophasic system such as water as media, and then extended to biphasic
17
and ionic liquid solutions. The highest reported yields of HMF from hexose are by using
fructose as reactant (T. Wang et al., 2014).
Regarding to utilize lignocellulose, glucose is drawing more attention for study as it is
cheaper and readily available after chemical or enzymatic hydrolysis. However, glucose as
reactant has lower rate of conversion and selectivity to HMF compared to fructose (Figure
2-4). The reason is fructose has both cyclic furanose tautomer and pyranose form in water,
while glucose mainly exists in pyranose form in water. The furanose tautomer facilitates
HMF formation and thus leads to superior reactivity and selectivity for HMF generation
compared to glucose.
To overcome the conversion threshold for glucose, an isomerization step is needed to
convert glucose to fructose first and subsequent dehydration to HMF. There are many
reports about using base or enzyme for conducting glucose to fructose isomerization
reaction, however, the catalysts works under pH from neutral to alkaline condition, which
is not suitable to combine with acidic condition for subsequent dehydration (T. Wang et
al., 2014).
The Lewis acid shows the potential to be used as catalyst combined with Brønsted acid
under acidic condition to isomerize glucose to fructose and dehydrate to HMF with
relatively high HMF yields (>60 mol% directly from glucose) (Y. Yang et al., 2012). Lewis
acid accepts lone electron pairs and Lewis bases donate the electron pair. Brønsted acid
lose or donate a hydrogen cation or proton and Brønsted base gain or accept the H+ (Bohn,
2014). The acidic pH of Lewis acids facilitate the catalytically active metal open the ring
of glucose and H+ in the solution helps the following dehydration. It is believed that the
metal of the Lewis acid would coordinate with the glucopyranose and stabilize the
18
transition state of the ring opened sugar (X. Zhang et al., 2015). Catalyst systems such as
CrClx/ionic liquid, Lewis acidic zeolites and MClx (M represents the metal) in water are
proven effective. For the MClx-type Lewis acidic metal salts, it works with Brønsted acid,
HCl, to synergistic conversion glucose to HMF. The pH provided dominantly by HCl in
the system plays a key role in controlling the Lewis acidity of the MClx catalyst (Kobayashi
et al., 2004). When no Lewis acids are added, no fructose is formed and detected from
glucose. Traditional moisture-sensitive and water-compatible metal salts (e.g., AlCl3,
SnCl4, GaCl3, YbCl3, LaCl3) and HCl were used for glucose conversion in a biphasic
system (Alonso, Wettstein, & Dumesic, 2013). AlCl3 gives the highest selectivity of 67
mol% for HMF production from glucose. The ionic radius of the cation, hardness/softness
of the Lewis acid, metal speciation may contributes to the different activities of the MClx
catalysts (Alonso, Wettstein, & Dumesic, 2013).
To sum up, the catalyst and reaction system used to catalyze lignocellulose to make
HMF is well studied. Among them, application of Lewis acid combined with Brønsted acid
in a biphasic solvent system to integrate glucose isomerization and fructose dehydration
step in one pot is promising as the selectivity for product is quite high while the catalyst
system is not as expensive as ionic liquid.
Even the reaction system is well developed regarding to aspects such as product yield
and sugar conversion, the systematic mechanism is not well understood, as there is not
many intrinsic kinetic analysis is done for the better understanding and further developing
the commercial process and reaction engineering for HMF production. Also, the synergistic
effect of Lewis acid and Brønsted acid need to further studied; the stability and possible
leaching of the Lewis acid to the system need to take caution.
19
2.3.2
Catalytic conversion of lignin from lignocellulose
To utilize the complex composition of lignin, mainly three processes are developed to
isolate lignin from hemicellulose and cellulose. The main but with none value-added
processing is use the lignin derived from pulp and paper industry or cellulosic ethanol
industry as combustion feedstock. Secondly, process to direct use lignin derived from paper
and pulp industry with minimum further treatment as polymer mixture (e.g. dispersants,
binders, adhesives, emulsifiers, resins). Good examples are the commercialized lignin
products known as lignosulfonates derived from the sulfite pulping industry and kraft
lignin from kraft pulping process (Gosselink, De Jong, Guran, & Abächerli, 2004). Last
but not the least, research hotspots are demonstrating new roadmaps in developing new
applications of lignin, addressing the depolymerization of the lignin molecule into valuable
aromatic monomers or the direct use as polymer. The main technologies can fell into four
dominant thermochemical treatments which are: pyrolysis, hydrolysis, hydrogenolysis and
oxidation. In the technical aspect, the chemical catalytic reactions to upgrade lignin to value
added chemicals normally involve several catalytic reaction simultaneously and not limit
to the above four category.
Lignin pyrolysis is a thermochemical decomposition of lignin at elevated temperatures
in the absence of oxygen. The reaction is irreversible and the products are temperature
dependent (from 200 to 1000 oC) with mixture with hydrocarbons, aromatic monomers and
other liquids (e.g. water, methanol) (Dorrestijn, Laarhoven, Arends, & Mulder, 2000). This
process needs huge energy input and generate a mixture of product which is not easy to
separate.
20
Lignin hydrolysis uses base or acid to dissociate lignin subunit linkage in solvent at mild
temperature, generating a mixture of monomers such as catechol, syringol, guaiacol and
vanillin. The acidic hydrolysis of lignin disadvantages is side reaction char formed heavily
as the repolymerization happens severely when acidic media used (Goldstein, 1981). When
base is used, the side reaction is suppressed but with product yield low, mostly less than
10% (Sarkanen & Ludwig, 1971).
Because the hydrolysis of lignin gives low product yield and poor selectivity, oxidative
and hydrogenolytic conditions are used to increase the reaction selectivity. Oxidation aims
to functionalize the lignin polymer or monomer to change the character such as
compatibility and function in either copolymerize with other materials or used as monomer
building blocks for polymer and fuel production. Catalysts such as oxidative enzyme,
biomimetic catalyst and inorganic catalyst conduct different oxidation pathways to
functionalization and depolymerization of lignin. Hydrogenolysis is to break up C-C or
carbon-heteroatom (C-O, C-N, C-S) single bond by effect of hydrogen. Dr. Abu-Omar’s
group showed lignin in poplar can be selective hydrodeoxygenated to cleave the β-O-4
ether linkage (Figure 2-5) by bimetallic palladium on carbon and zinc dissolved in
methanol and H2 environment with yield only two methoxyphenol products between 8090% (T. H. Parsell et al., 2013).
21
Figure 2-5 β-O-4 ether linkage the most prevalent linkage in lignin. Source: Adapted and
redrawn from reference (T. H. Parsell et al., 2013) .
The aromatics derived from lignin have similar properties to major aromatic compounds
in the existing petroleum chemical industry; their goal is turn the aromatics into high-value
molecules that have applications in fragrance, flavoring and high-octane jet fuels.
Removing the lignin can leave nearly pure cellulose behind with an enzymatic hydrolysis
yield of glucose around 95% (T. Parsell et al., 2015). The latest research outcome from Dr.
Shannon Stahl’s group showed by inducing aqueous formic acid into the oxidized lignin
from aspen, depolymerization of lignin would give a results more than 60wt% yield of low
molecular-mass aromatics in a mild condition (Rahimi, Ulbrich, Coon, & Stahl, 2014).
The breakthrough and application of lignin conversion technology is more than just
selective breaking the bonds within the lignin to make aromatic monomers, Dr. Miller’s
group successfully condensed lignin-derived hydroxyaldehyde with pentaerythritol and ditrimethylopropane to produce a new class of cyclic polyacetal ether thermoplastics which
22
have thermal properties is similar to those found in commodity plastics (Pemba, Rostagno,
Lee, & Miller, 2014). The development of lignin fractionation technology makes it possible
to remove lignin ahead and leave cellulose for fuel, bulk, and fine chemicals production.
In summary, facing the challenge of current biorefinary, lignin should be dealt with
carefully. It should play a role of value promoter instead of problem maker. The emerging
catalytic conversion techniques for lignin opens many options for economic, efficient and
environmental friendly production of lignin-derived aromatic compounds, which is a good
compensation for the current petroleum based chemical production system.
2.3.3
Remove lignin prior to conversion carbohydrates
Abu-Omar and coworkers shows the lignin can removed and upgraded to bioplastic,
flavoring and fragrance or even drop-in fuels and leave nearly pure carbohydrates
(Zakzeski et al., 2010). Dumesic and coworkers proposed the technical feasibility of
sequentially production furfural and HMF from hemicellulose and cellulose; both furfural
and HMF can be upgraded to make GVL(γ-valerolactone), and subsequential to make fuels
and chemical commodities (Wettstein, Alonso, Gürbüz, & Dumesic, 2012).
Based on the positive results of both, strategy for an integrated process for utilizing the
monomeric sugar and lignin stream can be developed to produce multiple value-added
products. The proposed roadmap is shown in Figure 2-6: removing lignin first to make
chemicals or bioplastics and then use the leftover (mainly hemicellulose and cellulose)
selectively upgrade to make GVL.
23
Figure 2-6 Propose the novel roadmap for conversion of biomass to value-added chemicals
lignocellulosic biomass(green), fuels(pink) and chemicals(orange), C5 and C6 sugars,
lignin(blue). Source: Adapted from Figure 1 in reference (Wettstein et al., 2012).
2.4
Conclusions
The biorefinary process gives the possibility to produce different chemicals as well as
transportation fuels from lignocellulose materials. However, the industry is currently
focusing on utilization of sugar monomer conversion to low value alcohols. New catalytic
approaches are opening multiple new possible processing pathways to design an
environmental sustainable route to fuels and chemicals. Lignin, which is usually regarded
as an inhibitor and barrier to enzymatic hydrolysis of cellulose, has great potential for fine
and bulk chemical production enabled by new catalysts and catalytic technologies. Novel
roadmaps for C5 and C6 sugars upgrade to HMF and furfural or even drop-in fuels by
chemical catalysis may be more efficient compared to current cellulosic ethanol industry.
24
ACKNOWLEDGEMENT
This material is based upon work supported as part of the Center for Direct Catalytic
Conversion of Biomass to Biofuels (C3Bio), an Energy Frontier Research Center funded
by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences,
Award Number DE-SC0000997.
25
CHAPTER 3. KINETICS OF MALEIC ACID AND ALUMINUM CHLORIDE
CATALYZED DEHYDRATION AND DEGRADATION OF GLUCOSE 2
3.1
Introduction
Over the past century, petroleum has been the main raw material for the production of
fine and bulk chemicals. However, with the rapidly growing demand for unsustainable
petroleum resources, lignocellulosic materials are drawing increasing attention as a
renewable and lower carbon-footprint feedstock (Boisen et al., 2009; Werpy et al., 2004).
Lignocellulose is mainly composed of cellulose (~40%wt), hemicellulose (~30%wt) and
lignin (~20%wt), with extractives and inorganic components comprising the remainder
(Murzin & Salmi, 2012). Glucose and xylose are the main subunits of the cellulose and
hemicellulose. Fractionation of lignocellulose can generate monosaccharide streams,
which can be catalytically converted with high selectivity to value added chemicals for use
as polymer building blocks and drop-in biofuels (vom Stein et al., 2011).
5-Hydroxymethylfurfural, HMF, has been identified as a platform building block for the
production of furanic polyesters, polyamides, and polyurethanes analogous to those
derived from petroleum (Pagan-Torres, Wang, Gallo, Shanks, & Dumesic, 2012). In
addition, bioderived HMF can be hydrolyzed to produce levulinic acid, which is ranked
2
Chapter 3 has been published in Journal of Energy and Fuels with the title of “Kinetics
of Maleic Acid and Aluminum Chloride Catalyzed Dehydration and Degradation of
Glucose” by Ximing Zhang, Barron B. Hewetson, and Nathan S. Mosier. DOI:
10.1021/ef502461s
26
as one of the 12 top value added chemicals produced from biomass by the US Department
of Energy (Boisen et al., 2009; Werpy et al., 2004). Levulinic acid is the building block for
producing a variety products, such as acrylate polymers and fuel additives such as γvalerolactone, 2-methyl-tetrahydrofuran, and ethyl levulinate (Serrano-Ruiz, West, &
Dumesic, 2010).
The process path between inexpensive and abundant lignocellulose and higher value
HMF is bridged by generating streams enriched in glucose from cellulose which is
subsequently isomerized to fructose and dehydrated to HMF (Choudhary, Mushrif, et al.,
2013). Theoretically, as fructose is twice the cost of glucose, using lignocellulose-derived
glucose as the feedstock instead of fructose results in a further reduction of cost ("Fructose
and glucose prices.,"). However, direct conversion of glucose to HMF has low yields when
catalyzed by Brønsted acids. This is because HMF is formed through the dehydration of a
5 member monosaccharide ring. Fructose is more favored in conversion to HMF as it forms
furanose tautomers in aqueous solution to a significantly greater extent than glucose (21.5%
of the molecules at steady-state compared to 1%) (L. Wang et al., 2014). Therefore, glucose
requires an additional isomerization step to achieve high yields of HMF. The
transformation of glucose to fructose is more readily catalyzed by Lewis acids than
Brønsted acids (L. Wang et al., 2014). If levulinic acid is the desired product, HMF can be
hydrolyzed by Brønsted acids (see Figure 3-1). Ideally, both Lewis and Brønsted acids
could be used together to achieve conversion of glucose to HMF or levulinic acid in a
single reactor.
27
Figure 3-1 Reaction pathway for glucose upgrade to levulinic acid Adapted based on
Scheme 1 in Reference (Choudhary, Mushrif, et al., 2013).
Various Lewis acids, such as AlCl3, SnCl4, VCl3, InCl3, GaCl3, LaCl3, DyCl3, and YbCl3
have been shown to be effective in converting glucose to HMF in aqueous media at
temperatures around 443K (Choudhary, Mushrif, et al., 2013). However undesired
reactions produce humin, a group of carbonaceous, heterogeneous and polydisperse
materials, through the condensation of fructose with HMF. These undesired huminforming reactions are challenges that need to be overcome to fully commercialize this
process (van Zandvoort et al., 2013). For example, the Biofine process developed by
BioMetics, Inc. with funding from the U.S. DOE, aimed to commercialize technology to
make intermediate chemicals such as furfural and levulinic acid from lignocellulose by
using two-step mineral acid hydrolysis (Hayes & Hayes, 2009). The process illustrated that
humin formation inhibits the yield of levulinic acid and increases the costs of product
28
separation and reactor construction (Hayes & Hayes, 2009). Weckhuysen et al. pointed out
that while different mechanisms for these reactions have been proposed, the acid-catalyzed
dehydration of pure fructose to HMF yielded the highest amount of humins (van Zandvoort
et al., 2013). It was also shown that humin formation also involved reactions other than
aldol condensation. Results from varied acid and sugar concentration for many
temperatures showed that humin formation was strongly influenced by reaction
temperature and acid concentration, but not sugar concentration (van Zandvoort et al.,
2013).
For the utilization of Brønsted acids in the dehydration of hexose, strong mineral acids,
especially sulfuric acid and hydrochloric acid, are commonly used as homogenous catalysts
as they are effective and inexpensive. On the other hand, these mineral acids are corrosive
to equipment and can generate a large inorganic waste stream (Kootstra, Mosier, Scott,
Beeftink, & Sanders, 2009).Though more expensive compared to strong acids, organic
acids are less corrosive, more selective, and may be thermally decomposed into nontoxic
molecules (CO2, formic acid and fumaric acid) at the end of their use cycle (Kootstra,
Beeftink, Scott, & Sanders, 2009; Lu & Mosier, 2007, 2008; N. S. Mosier, Ladisch, &
Ladisch, 2002). Thus organic acids are also being considered as candidates when
experimenting with the dehydration of hexose.
In this paper we report the kinetics of a Lewis acid (AlCl3) and two Brønsted acids (HCl
and maleic acid) to catalyze glucose isomerization and produce HMF and levulinic acid in
aqueous phase, without extraction of products to an organic phase. In addition, a more indepth study is presented on the effect of combinations of Lewis and Brønsted acid on
29
glucose. Finally, we report the benefit of maleic acid to increase the selectivity of the
reaction kinetics for levulinic acid production and to decrease humin production.
3.2
Methodology
The reactions were conducted using 3.5 mL, 316 stainless steel tubing (8 mm
diameter×2.1 mm wall thickness×7 cm long) fitted with a pair of 1.2 cm Swagelok tube
end fittings (Swagelok, Solon, OH). Each tube had a volume of 3.5 mL and was filled with
2 mL solution to give about 28% free space for liquid expansion and gas production. The
solution is heated at 180 ℃ for up to 8 min by placing the tube in a Tecam SBL-1 fluidized
sand bath (Cole-Parmer, Vernon Hills, IL). Heat up time was measured to be 2 min using
a thermocouple inserted into the center of the reactor. Stated reaction times begin after the
2 min heat up period. All yields and kinetics are reported compared to reaction solutions
heated to 180 ℃ and then immediately cooled to <100 °C within 30 seconds by submersion
in cool water.
For maleic acid to be used as a catalyst in commercial application, the selectivity and
reusability should be key factors in consideration; both are related closely to maleic acid
thermal stability. Kim et al. reported that maleic acid will hydrolyze to malic acid at
elevated temperatures (Kim et al., 2012). They reported that after 10 minutes of reaction at
180 °C, 6% of the initial maleic acid was hydrolyzed to malic acid. The conditions
examined in this paper were shorter than this and we did not observe detectible amounts of
malic acid using the same HPLC method used by Kim et al.
The heat up time is tested to accurately calculate the reaction time. The Figure 3-2 shows
the heating up time trend. We regard 95% of the desired temperature as the requirement
30
for counting for starting reaction time. Elevated temperatures are tested; it shows 2 minutes
would be the heating up time for the reaction conducted under 180 oC.
Figure 3-2 Reactor heat up profile at various temperatures
For conditions using only HCl and maleic acid as catalyst, the possibility of metal ions
leaching from stainless steel tubing reactor may affect the rate kinetics. To examine this,
control experiments were conducted in glass reactors and compared to results from
stainless steel tubing reactors. The glass reactor were 2 mL Sigma Aldrich glass bottles
with a solution loading of 1 mL to give a 50% free space for liquid expansion and gas
production. Each glass reactor was sealed by rubber septa with silver aluminum cap. A
copper shim (add thickness) was added between the rubber septa and aluminum cap to
enhance the strength of the seal. Glass reactors were used with caution as the pressure may
result in reactor explosion. Stated reaction times begin after a 10 second heat up period.
31
All yields and kinetics are reported compared to reaction solutions heated to 180 ℃ and
then immediately cooled to <100°C within 30 seconds by submersion in cool water.
Glucose, fructose, HMF, levulinic acid, formic acid, AlCl3, and maleic acid were
purchased from Sigma-Aldrich (St. Louis, MO) and were used as received. In a typical
experiment, 2 mL of an aqueous solution of reactant was mixed with an aqueous solution
of catalyst: AlCl3·6H2O (0.1 mol/L), HCl (0.1 mol/L) or maleic acid (0.1 mol/L) and then
placed into the reactor and sealed. The reaction was heated up to 180 ℃ and held for a
fixed period of time (see Table 3-1). After quenching the sample in cool water, the solution
was filtered by a syringe fitted with a 0.2 µm nylon filter. The filtered liquid was analyzed
by high pressure liquid chromatography (HPLC), equipped with a Waters 1525 pump and
Waters 2412 Refractive Index detector (Waters, Milford, MA). An HPX-87H AMINEX
column (BioRAD, Hercules, CA) was used for separation with 5 mM aqueous H2SO4 and
5% (w/w) acetonitrile as the mobile phase at a flow rate of 0.6 mL per minute. The
acetonitrile was used to facilitate the separation of hexose and maleic acid (Lu & Mosier,
2008). The column temperature was maintained at 338 K. All concentrations of sugars and
organic products in the aqueous phase were determined by external calibration standards.
The pH value of the reaction solution prior to the reaction was measured on a WD-3563440 Waterproof Double Junction pH (± 0.01 pH units) calibrated with standard buffer
solutions. MATLAB was used to simulate the reactant and product profile. 4th order
Runge-Kutta method was used to solve ODE.
32
Table 3-1 Conditions for Kinetic Analysis of Hexose Dehydration, HMF Hydrolysis, and
Humin formation.
Parameters
Treatment conditions
Temperature
180 ℃
Reaction time
0, 1,2,3,4 and 6 min
Lewis acid (100 mM)
AlCl3
Brønsted acids (100 mM)
HCl
Maleic acid
Reactants
Glucose or Fructose (250 mM)
HMF (63 mM)
3.3
3.3.1
Results and Discussion
Kinetic Parameter Estimation.
We assumed that the reactor was well mixed and isothermal during the reaction period.
We used a monophasic model based on homogeneous, pseudo first-order reaction kinetics.
The glucose was first isomerized to make fructose with the Lewis acid and then dehydrated
to generate HMF. HMF was a reactant being further rehydrated to make levulinic acid.
Glucose, fructose and HMF are reactants theoretically capable of generating humins. The
simplified overall kinetic model used in this study is shown in Figure 3-1.
The kinetic parameter estimation is based on the following set of differential equations
(3-1)-(3-4):
𝑑[𝑔𝑙𝑢𝑐𝑜𝑠𝑒]
𝑑𝑡
𝑑[𝐹𝑟𝑢𝑡𝑜𝑠𝑒]
𝑑𝑡
𝑑[𝐻𝑀𝐹]
𝑑𝑡
= −(𝑘1 + 𝑘4 )[𝐺𝑙𝑢𝑐𝑜𝑠𝑒] + 𝑘−1 [𝐹𝑟𝑢𝑐𝑡𝑜𝑠𝑒]
(3-1)
= 𝑘1 [𝐺𝑙𝑢𝑐𝑜𝑠𝑒] − (𝑘−1 + 𝑘2 + 𝑘5 )[𝐹𝑟𝑢𝑡𝑜𝑠𝑒]
(3-2)
= 𝑘2 [𝐹𝑟𝑢𝑐𝑡𝑜𝑠𝑒] − （𝑘6 + 𝑘3 ）[𝐻𝑀𝐹]
(3-3)
33
𝑑[𝐿𝑒𝑣𝑢𝑙𝑖𝑛𝑖𝑐 𝐴𝑐𝑖𝑑]
𝑑𝑡
(3-4)
= 𝑘3 [𝐻𝑀𝐹]
Reactions were conducted using pure glucose, fructose or HMF as the reactant to
measure the initial reaction rates as shown in Table 3-2.
Table 3-2 Reaction Datasets for Determining Kinetic Constants
Reactant
Glucose
(250 mM)
Fructose
(250 mM)
HMF(63 mM)
k dis (min-1)
k-1 (min-1)
k2 (min-1)
k3 (min-1)
k5,6(min-1)
√
√
√
√
√
√
Note: the kinetic constants shown in the table is described in Figure 3-1, where kdis=k1+ k4.
The rate constants (kx) in the equations were determined by plotting natural logarithm of
reactant/reactant initial (time = 0) versus time (per minute) using the initial time points (0,
1, 2, 3 min) of the reaction. In the case when HMF was used as the reactant, two parallel,
pseudo first-order reactions occurred. The first converts HMF to equimolar amounts of
levulinic and formic acids. Simultaneously a parallel reaction produces humins from HMF
and HMF derived intermediates (Horvat, Klaić, Metelko, & Šunjić, 1985). Apparent firstorder rate coefficients 𝑘3 plus 𝑘6 for reactions in Figure 3-1 was the slope of plotting the
natural logarithm of [𝐻𝑀𝐹]/[𝐻𝑀𝐹]0 vs time. Measuring the negative concentration of
levulinic acid vs concentration of HMF would have a slope equal to 𝑘3 /(𝑘3 + 𝑘6 ). Rate
34
constants 𝑘3 and 𝑘6 could then be determined algebraically. The estimated rate constants
are shown on Table 3-3.
Table 3-3 Kinetic rate constant (kdisappearance rate constant with confidence interval of 95%,
all other rate constant was calculated based on [product]/ [reactant] without confidence
interval.)
Catalyst
kdis (min-1)
k-1 (min-1)
k2 (min-1)
k3 (min-1)
k5(min-1)
k6(min-1)
HCl
0.11±0.02
0
0.51
0.13
0.13
0.05
Maleic acid
0.04±0.01
0
0.08
0.02
0.1
0.02
AlCl3
0.68±0.2
0.09
0.25
0.06
0.23
0.08
AlCl3+HCl
0.93±0.17
0.17
0.6
0.09
0.23
0.09
AlCl3+MA
0.23±0.04
0.09
1
0.11
0.09
0.06
Note: the kinetic constants shown in the table is described in Figure 3-1, where kdis=k1+ k4.
In the study, we used a constant catalyst concentration (100 mM) in the solution. pH
value plays a key role in controlling the Lewis acidity of metal halides to form catalytically
active metal species in aqueous solution, which could significantly impact the glucose ringopening step and subsequent dehydration and rehydration reactions (Fringuelli, Pizzo, &
Vaccaro, 2001). In our experiment, HCl is a strong mineral acid that produces a solution
of pH 1.02 as measured. Maleic acid is a weak organic acid, which results in a solution
with pH 1.85. Aqueous solution of AlCl3 has a pH of 2.76. The mixed catalysts of maleic
acid with AlCl3 gives a pH of 1.44 and HCl combined with AlCl3 gives a pH value of 1.20.
35
3.3.2
Brønsted Acid Comparison.
To determine if metal ions leaching from the stainless steel reactors were affecting the
kinetics, control experiments are conducted in glass reactors. Experimental data from both
maleic acid and HCl are comparable with data from steel tubing reactor, which suggests
that the effects from metal ions leached from stainless steel are negligible.
Reactant conversion and product yields for glucose, at an initial concentration of 34 g/L
reacted by HCl and 36g/L by maleic acid after 2 min heat up time, are showed in Figure 32 a and b. Formic acid yields are not shown due to significant decomposition to CO 2 and
H2 in the reactor (Fukuzumi, Kobayashi, & Suenobu, 2008). The conversion of the reactant
and yield of the product is calculated based on the following equations (3-5)-(3-6):
% converion of reactant =
% yield of product 𝑖 =
(𝐶𝑟𝑒𝑎𝑐𝑡𝑎𝑛𝑡,𝑡=0 −𝐶𝑟𝑒𝑎𝑐𝑡𝑎𝑛𝑡)
𝐶𝑟𝑒𝑎𝑐𝑡𝑎𝑛𝑡,𝑡=0
𝐶𝑖
𝐶𝑟𝑒𝑎𝑐𝑡𝑎𝑛𝑡,𝑡=0
× 100
× 100
Note: C reactant is based on molar concentration.
(3-5)
(3-6)
36
(a)
80%
70%
60%
Yield (%)
50%
40%
30%
20%
10%
0%
0
1
2
3
Reaction time (min)
4
5
6
80%
(b)
70%
60%
Yield (%)
50%
40%
30%
20%
10%
0%
0
1
2
Reaction 3
time (min)
4
5
Figure 3-3 Reactant and product yields observed during the reaction with (a) maleic acid
as catalyst; (b) HCl as catalyst; Shape represents experimental data, line represents
simulated data. Glucose unreacted = ○, fructose = □, HMF = ◊, levulinic acid = ∆, formic
acid is not shown. Reaction conditions: glucose (250 mM); temperature (453 K); maleic
acid (100 mM); HCl (100 mM); AlCl3 (100 mM). Error bars represent standard deviation
for triplicates.
6
37
It has been reported that glucose can be converted to fructose, HMF and humins in
parallel (Pagan-Torres et al., 2012).Because the combination of reaction rate constants are
too complex to be accurately calculated from only the data shown in Figure 3-2, the rate
constants k1 and k4 are not shown in Table 3-3 and instead a lumped rate constant for
glucose disappearance is presented. The overall glucose disappearance rate constant for
HCl (0.11 min-1) is higher than maleic acid (0.04 min-1). HMF and levulinic acid can be
measured by HPLC in moderate amounts when HCl was used as the catalyst, but amounts
are much smaller for maleic acid. However, under the catalysis by maleic acid, the
accumulation of fructose reached at maximum 9.6g/L after 4 min and then decreased to
around 6 g/L at 6 min, while under HCl catalysis, fructose was only detectable in trace
amounts. When catalyzed by HCl, the HMF formation rate and concentrations were higher
compared to reactions catalyzed by maleic acid. It is worth noting that levulinic acid could
be detected in a moderate amount at the start time, defined by heating the solution to 180 ℃
and immediately cooling, in the presence of HCl as catalyst. This suggests that the reaction
pathway may differ between the two kinds of Brønsted acids. It is obvious that if rate
constants for HCl from Table 3-3 are used to simulate glucose dehydration through fructose
as an intermediate, the predicted fructose is much higher than the experimental data, which
justifies the hypothesis that HCl catalyzes the direct dehydration of glucose to HMF.
Pagán-Torres et al. reported that Brønsted acids such as HCl alone can directly convert
glucose to HMF without involvement of metal chloride to make HMF with low efficiency
(selectivity of 30% with 91% glucose conversion) (Pagan-Torres et al., 2012). They
proposed the open-chain form of glucose was dehydrated at the C-2 position, forming a
carbocation with reacts with the hydroxyl group at C-5 position, forming tetrahydro-3,4-
38
dihydroxy-5-(hydroxymethyl)-2-furaldehyde followed by further dehydration to HMF.
The data presented in this paper are consistent with this proposed mechanism.
On the other hand, the experimental and simulated data from catalysis with maleic acid
seemed to facilitate the glucose isomerization to fructose, which is subsequently
dehydrated to HMF. These data suggest that maleic acid behaves more like Lewis acid
metal halides in the reaction pathway from glucose to HMF.
When fructose was the reactant, HCl showed a much higher rate of dehydration to HMF
than hydrolysis of HMF to make levulinic acid (see Table 3-3). Compared to HCl, rates of
fructose dehydration and HMF hydrolysis to levulinic and formic acids were significantly
slower.
3.3.3
Lewis Acid and Brønsted Acid Performance.
AlCl3 is a Lewis acid and an effective catalyst in facilitating glucose isomerization to
fructose. When AlCl3 was used alone, glucose was quickly converted to fructose. The
glucose disappearance rate constant (0.68 min-1) was much larger than either HCl (0.11
min-1) or maleic acid (0.04 min-1) catalyzed reactions. Neither HCl nor maleic acid is as
effective as AlCl3 in isomerization, which is consistent with prior reported results by
Pagán-Torres et al. (Pagan-Torres et al., 2012). When Lewis and Brønsted acids were
mixed, HCl accelerated the glucose disappearance rate from 0.68 min-1 to 0.93 min-1, while
maleic acid moderated the rate constant.
For the Brønsted acids, the results showed that fructose was converted to HMF and
levulinic acid directly, no glucose formation was observed. This indicates that
isomerization rate constants (k-1) of the Brønsted acid are zero, and the Brønsted acid
directly catalyzed the fructose dehydration. In the case of AlCl3 as a catalyst, fructose was
39
isomerized to glucose with a rate constant of 0.09 min-1, much lower than the isomerization
rate constant from glucose to fructose (0.68 min-1). The apparent isomerization rate
constant (k-1) from fructose to glucose was affected by the dehydration of fructose to HMF.
This is mainly because the equilibrium of isomerization is transient as fructose was quickly
dehydrated to HMF, making the reaction appear as an irreversible first-order reaction. HCl
combined with AlCl3 showed a larger isomerization rate constant (0.17 min-1) compared to
maleic acid combined with AlCl3 (0.09 min-1). One possible interpretation is that the
greater added acidity provided by HCl compared to maleic acid facilitates the reaction.
For the dehydration reaction (fructose to HMF) and rehydration reaction (HMF to
levulinic acid and formic acid), significant humin formation was evidenced by the color
change after the reaction (from colorless and transparent before heating to dark brown,
muddy after quenched) as well as the mass balance of the reactant and products. HCl and
maleic acid behaved quite differently. In the dehydration reaction step, HCl had a much
higher rate constant (0.51 min-1) compared to maleic acid (0.08 min-1). However, when
AlCl3 was introduced, the maleic acid had advantages in the rate constant (1 min-1 versus
0.6 min-1) at the same time had slower humins generation (0.09 min-1 compared to 0.23
min-1).
In the rehydration step, the rate constant of HMF to levulinic acid in the presence of
maleic acid was almost twice high as that of HMF toward humins (0.11 min -1 versus 0.06
min-1). However, HCl combined with AlCl3 had an equal rate constant (0.09 min-1) for
HMF disappearance to either levulinic acid or humins. Alone, AlCl3 has the worst rate
constant ratio of levulinic acid (0.06 min-1) to humins (0.08 min-1). The overall rate constant
40
for the rehydration step showed maleic acid combined with AlCl3 had the best selectivity
of levulinic acid to humins.
80%
(c)
70%
Yield (%)
60%
50%
40%
30%
20%
10%
0%
0
1
2
Reaction 3time (min)
4
5
80%
6
(d)
70%
60%
Yield (%)
50%
40%
30%
20%
10%
0%
0
1
2
Reaction 3time (min)
4
5
6
41
(e)
80%
70%
60%
Yield (%)
50%
40%
30%
20%
10%
0%
0
1
2
3
4
5
6
Reaction time (min)
Figure 3-3 Reactant and product yields observed during the reaction with (c) AlCl3 as
catalyst; (d) AlCl3 and maleic acid as catalysts; (e) AlCl3 and HCl as catalysts. Shape
represents experimental data, line represents simulated data. Glucose unreacted = ○,
fructose = □, HMF = ◊, levulinic acid = ∆, formic acid is not shown. Reaction conditions:
glucose (250 mM); temperature (453 K); maleic acid (100 mM); HCl (100 mM); AlCl3
(100 mM). Error bars represent standard deviation for triplicates.
Subsequently we used the estimated rate constants to simulate the reactant and product
profile in MATLAB. The experimental data and simulated trend are shown in Figure 3-2
for all the catalysts combinations. The simulated trend does not fit the data in Figure 3-2
(b), likely because HCl converts some glucose directly to HMF instead first isomerizing
glucose to fructose. For AlCl3, shown in Figure 3-2 (c), the simulated fructose is higher
than experimental data, which suggests that the rate of humin formation is faster than
42
predicted. It is possible that in a single aqueous phase, fructose reacted with humins to form
additional humins through aldol addition and condensation reactions which violates the
assumption of pseudo first order reaction kinetics at the longer residence times (Y. Yang,
Hu, & Abu-Omar, 2012). The above simulations suggest that the reactions are not strictly
pseudo first order, especially at longer residence times. Therefore, the kinetic model should
be used with caution to predict the product yield beyond the data range that was
experimentally verified.
3.3.4
Effects of Catalysts on Selectivity.
With all the reactant and products profiles, an overall trend can be analyzed to gain
quantitative information on the effects of the process conditions on the selectivity of the
reaction. For this purpose, it is convenient to use selectivity, which is defined as the ratio
of the amount of desired product and the amount of humins formed (Equation 3-7).
selectivity =
moles of levulinic acid formed
moles of humins formed
(3-7)
Using selectivity, it is possible to describe the trend of the effect of different catalysts on
HMF. The results (see Figure 3-4) show that AlCl3 combined with HCl has the lowest
selectivity over reaction time. It also shows the selectivity of AlCl3 combined with HCl is
less than either HCl alone or AlCl3 alone. However, maleic acid combined with AlCl3 has
an additive effect on selectivity compared to either maleic acid alone or AlCl3 alone.
43
Figure 3-4 Selectivity of levulinic acid as a function of time. HCl = ∆, AlCl3 = ◊,
AlCl3+HCl = ○, Maleic acid = □, Maleic acid+AlCl3 = ●. Reaction conditions: HMF (63
mM); temperature (453 K); AlCl3（100 mM）；Maleic acid (100 mM); HCl (100 mM).
Error bars represent standard deviation for triplicates.
Figure 3-5 shows the molar distribution in percent of reactants and products for each
catalyst, alone or paired, at 6 min of reaction time. While measured amounts of formic acid
were less than 1:1 molar ratio with levulinic acid, we assumed that a significant amount of
formic acid was degraded to carbon dioxide and water in our experiments. The longer
residence time results in a larger percent of formic acid decomposed (Ribeiro, Marenich,
Cramer, & Truhlar, 2011). AlCl3 alone can generate more humins than levulinic acid;
maleic acid itself generates a small amount of levulinic acid with similar amounts of
humins, while HCl produces the largest amount of the levulinic acid but less humins.
44
When glucose is the reactant, to facilitate the reaction, the combination of the Lewis and
Brønsted acids is required to achieve high product yields. Results show at 180 ℃, 6 min
and an initial HMF concentration of 63 mM, maleic acid combined with AlCl 3 generated
only 50% of total amount of humins were generated compared to hydrochloric acid
combined with AlCl3. Although the rates of reaction are slower, maleic acid with AlCl3
generates much more levulinic acid compared to HCl paired with AlCl3. The results also
showed that maleic acid combined with Lewis acid redistribute the product ratio
significantly.
AlCl3
Maleic acid
HCl
Maleic acid+AlCl3
43±1
41±2
HCl+AlCl3
100%
Product distribution in percentage
90%
12±1.2
23±0.1
80%
29±4
16
70%
60%
34
50%
25
52
31
40%
72±1
30%
20%
43±1
26±2
10%
34±3
19±6
0%
Unreacted HMF
Humins and unquantified compounds
levulinic acid
Figure 3-5 Mole distribution based on 6 min. White section = humins and unquantified
compounds, gray section = unreacted HMF, black section = levulinic acid. Reaction
conditions: HMF (63 mM); temperature (453 K); AlCl3 （100 mM）; Maleic acid (100
mM); HCl (100 mM). Error bars represent standard deviation for triplicates; white
section=100% - gray section – black section.
45
3.3.5
Interpretation of Maleic acid Effect on Selectivity.
Lewis acids has been extensively studied by many groups for isomerization of glucose
to fructose (Pagan-Torres et al., 2012). Abu-Omar et al. illustrated that a Lewis acid such
as AlCl3·6H2O is not only effective for isomerization but also partially effective in
dehydration of fructose to HMF in a biphasic (water-solvent) system (Y. Yang et al., 2012).
For Brønsted acids, Dumesic et al. reported HCl alone can directly convert glucose to
HMF with low efficiency (Pagan-Torres et al., 2012). As a Brønsted acid, maleic acid has
been shown to exhibit reaction kinetics for sugars that differs from mineral acids. Mosier
et al. reported that maleic acid below a concentration of 200 mM resulted in comparable
or lower rates of sugar degradation than pure water (N. S. Mosier, Sarikaya, Ladisch, &
Ladisch, 2001). Lu et al. found that the rate of dehydration was inversely proportional to
the catalyst concentration when maleic acid is used for degradation of xylose to furfural,
unlike strong mineral acids in which the rate of xylose degradation solely depends on the
proton concentration (pH) and increases with acid concentration (Lu & Mosier, 2007,
2008). The maleate ion may form a strong internal hydrogen bond with the xylose
transition-state intermediate, inhibiting the intermediate from further degrading. Kim et al.
reported that maleic acid was found to be an effective catalyst able to selectively hydrolyze
xylan to xylose and catalyze further dehydrated of the separated xylose to furfural at
elevated temperatures with high yield (Kim et al., 2012).
Sievers et al. claimed that saccharides are the main component of humins as strong NMR
similarities were shown between the starting cellulose and the humins derived from it via
hydrolysis in an ionic liquid solution (Sievers et al., 2009). They also concluded that other
species could be involved in the humin formation and final structure. Hovat et al. proposed
46
that HMF transformed to levulinic acid and humins through a different pathway, where
HMF derived 2,5-dioxo-6-hydroxy-hexanal is an important intermediate (Horvat et al.,
1985). More recently Patil et al. clearly showed the reaction of humins growth involved
HMF and 2, 5-dioxo-6-hydroxy-hexanal aldol addition and condensation in the presence
of acid (Y. Yang et al., 2012). Our experiments showed that maleic acid alone substantially
decreases humin generation compared to HCl alone from both fructose and HMF in
monophasic, aqueous reaction. Results also showed that maleic acid paired with AlCl3
generates substantially fewer humins from fructose compared to HCl combined with AlCl3.
This suggests that maleic acid may interact with fructose and HMF to form transient
intermediates that inhibit humin formation. We hypothesize that the internal hydrogen
bond of maleic acid stabilizes the intermediate between HMF and levulinic. The possible
interactions between maleic acid and fructose or HMF and the influence of pH should be
further tested in subsequent work.
3.4
Conclusions
Compared with HCl in a mixed aqueous catalysts system, maleic acid was shown to
significantly change the selectivity of glucose and fructose degradation and dehydration
when combined with the catalyst AlCl3. At the longest reaction time tested (6 min), the
amount of humins generated by maleic acid combined with AlCl3 were only 50% of the
total amount of humins generated by HCl with AlCl3. This results in an increased
selectivity toward levulinic acid formation. It may be concluded that, instead of rate
acceleration under catalysis by strong acids such as HCl, the product distribution is more
strongly affected by organic acids such as maleic acid.
47
ACKNOWLEDGEMENT
This material is based upon work supported as part of the Center for Direct Catalytic
Conversion of Biomass to Biofuels (C3Bio), an Energy Frontier Research Center funded
by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences,
Award Number DE-SC0000997.
48
CHAPTER 4. REACTION MECHANISM OF MALEIC ACID AND ALUMINUM
CHLORIDE CATALYZED GLUCOSE AND FRUCTOSE TO
5-(HYDROXYMETHYL) FURFURAL AND LEVULINIC ACID IN AQUEOUS
MEDIA 3
4.1
Introduction
Lignocellulose is a potential sustainable and low-carbon resource for producing
chemicals and transportation fuels (Werpy et al., 2004). Cellulose, which represents about
40% of lignocellulosic biomass, is mainly made of D-glucose subunits cross-linked by β1,4-glucoside bonds (N. Mosier et al., 2005). D-glucose can be catalytically isomerized to
fructose and consequently dehydration to 5-(hydroxymethyl)furfural (HMF) with high
selectivity and rehydration to make levulinic acid in almost quantitative yield (Zhang et al.,
2015). Both HMF and levulinic acid have been identified as potential building blocks for
biofuels, biochemicals, and biopolymers (Choudhary, Mushrif, et al., 2013; Zhang et al.,
2015). Commercial production of HMF and levulinic acid requires processes that result in
high yields for all of the reactions such as glucose (aldose) to fructose (ketose)
isomerization, fructose dehydration and levulinic acid rehydration steps need to take place
on one spot and in sequence. Fructose conversion to HMF and levulinic acid is mainly
catalyzed by Brønsted acid and has been studied thoroughly,
3
Chapter 4 is adapted from the manuscript “Reaction Mechanism of Maleic Acid and
Aluminum Chloride Catalyzed Glucose and Fructose to 5-(Hydroxymethyl) furfural and
Levulinic Acid in Aqueous Media”, which is in preparation for submission to Journal of
Green Chemistry.
49
however, the isomerization step was still a hotspot for continuing research on (Antal, Mok,
& Richards, 1990; Assary, Kim, Low, Greeley, & Curtiss, 2012; Assary, Redfern, Greeley,
& Curtiss, 2011; Caratzoulas & Vlachos, 2011; Choudhary, Pinar, Lobo, Vlachos, &
Sandler, 2013; Li et al., 2012; Román-Leshkov, Chheda, & Dumesic, 2006).
For the aldose to ketose isomerization step, there is mainly two reaction mechanisms
proposed: proton transfer from C-2 to C-1 and intramolecular hydride shift from O-2 to O1 (Nagorski & Richard, 2001). It is believed that base-catalyzed isomerization takes place
by a proton transfer through a series of enolate intermediates generated after the
deprotonation of the α-carbonyl carbon in water, while metal ion from Lewis acid in acidic
solution may facilitate the reaction go through intramolecular hydride shift by forming
complex with glucose O-1 and O-2 (Bordwell, Zhang, Eventova, & Rappoport, 1997; Rom
á n‐ Leshkov, Moliner, Labinger, & Davis, 2010). Interestingly, research found when
enzyme xylose isomerase is used for isomerization of glucose into fructose for production
of high-fructose corn syrups (HFCS), both reaction mechanisms exist: enolate
intermediates generated by histidine-directed base catalysis can mediate the reaction; and
metal centers in the enzyme can stabilize of the sugar’s open-chain form and facilitate an
intramolecular hydride shift (Allen et al., 1994; Kovalevsky et al., 2010; Rose, O'Connell,
& Mortlock, 1969; Schray & Rose, 1971).
Maleic acid, an organic dicarboxylic acid, in recent study shows its unique performance
in stabilizing glucose in isomerization step and acts like Lewis acid, catalyzing glucose
isomerize to fructose. It is also found can inhibit side product humins generation
substantially, compare to strong Brønsted HCl, alone or combined with AlCl3 (Zhang et
50
al., 2015). Previous study also showed maleic acid has higher selectivity compared to
sulfuric acid in hydrolysis of sugar polymers as less sugar degraded to humins (Lu &
Mosier, 2007, 2008; N. S. Mosier et al., 2002; N. S. Mosier et al., 2001). Hypothesis is
made that the two carboxylic groups in maleic acid can mimic the structure of the active
site in cellulase enzymes, which also has two carboxylic groups catalyzing the hydrolysis
through general acid catalysis (N. S. Mosier et al., 2002; N. S. Mosier et al., 2001). The
carboxylic groups of maleic acid is 2.5 Å , shorter than 4-10 Å of two carboxylic residues in
cellulase enzymes, may generate strong hydrogen bond (Lu & Mosier, 2008).
In this work we investigate the Lewis acid (AlCl3) catalyzed glucose transformation in
water combined with maleic acid at elevated temperature. To further understand the
mechanism, HCl was used combined with AlCl3 to compare. The kinetics for the reactions
were reported and activation energy and Gibbs free energy are estimated and analyzed.
Further, we report the first results of maleic acid in stabilizing glucose by using CarbonNMR. Additionally, mass spectrometer was used to study the configuration of intermediate
formed in the water.
4.2
4.2.1
Methodology
Sample Preparation.
Glucose, fructose, HMF, levulinic acid, formic acid, AlCl3, and maleic acid were
purchased from Sigma-Aldrich (St. Louis, MO) and were used as received. The reactions
were conducted using 3.5 mL, 316 stainless steel tubing (8 mm diameter×2.1 mm wall
thickness×7 cm long) fitted with a pair of 1.2 cm Swagelok tube end fittings (Swagelok,
Solon, OH). Each tube had a volume of 3.5 mL and was filled with 2 mL solution to give
about 28% free space for liquid expansion and gas production. The solution is heated at
51
140, 160 ℃ for up to 40 min by placing the tube in a Tecam SBL-1 fluidized sand bath
(Cole-Parmer, Vernon Hills, IL). Heat up time was measured to be 2 min using a
thermocouple inserted into the center of the reactor. Stated reaction times begin after the 2
min heat up period. All yields and kinetics are reported compared to reaction solutions
heated to 140 and 160 ℃ and then immediately cooled to <100 °C within 30 seconds by
submersion in cool water. Similar experiment was conducted under 180 oC in previous
work (Zhang et al., 2015). The possible metal ions leaching from stainless steel tubing is
tested using only HCl and maleic acid as catalyst. The possibility of metal ions leaching
from stainless steel tubing reactor may affect the rate kinetics. To examine this, control
experiments were conducted in glass reactors and compared to results from stainless steel
tubing reactors. The glass reactor were 2 mL Sigma Aldrich glass bottles with a solution
loading of 1 mL to give a 50% free space for liquid expansion and gas production. Each
glass reactor was sealed by rubber septa with silver aluminum cap. A copper shim (add
thickness) was added between the rubber septa and aluminum cap to enhance the strength
of the seal. Glass reactors were used with caution as the pressure may result in reactor
explosion. Stated reaction times begin after a 10 second heat up period. All yields and
kinetics are reported compared to reaction solutions heated to 180 ℃ and then immediately
cooled to <100 °C within 30 seconds by submersion in cool water.
4.2.2
HPLC Analysis.
In a typical experiment, 2 mL of an aqueous solution of reactant was mixed with an
aqueous solution of catalyst: AlCl3·6H2O (0.1 mol/L), HCl (0.1 mol/L) or maleic acid (0.1
mol/L) and then placed into the reactor and sealed. The reaction was heated up to 140 or
160 ℃ and held for a fixed period of time (Table 4-1).
52
Table 4-1 Conditions for Kinetic Analysis of Hexose Dehydration, HMF Hydrolysis, and
Humin
Parameters
Treatment conditions
Temperature
140, 160 ℃
Reaction time
0, 1, 2, 3, 4, 6, 14, 22, 30, 38 min
Catalysts
AlCl3(100 mM)+HCl(100 mM)
AlCl3(100mM)+maleic
acid(100 mM)
Reactants
Glucose or Fructose (250 mM)
HMF (63 mM)
After quenching the sample in cool water, the solution was filtered by a syringe fitted
with a 0.2 µm nylon filter. The filtered liquid was analyzed by high pressure liquid
chromatography (HPLC), equipped with a Waters 1525 pump and Waters 2412 Refractive
Index detector (Waters, Milford, MA). An HPX-87H AMINEX column (BioRAD,
Hercules, CA) was used for separation with 5 mM aqueous H2SO4 and 5% (w/w)
acetonitrile as the mobile phase at a flow rate of 0.6 mL per minute. The acetonitrile was
used to facilitate the separation of hexose and maleic acid (Lu & Mosier, 2008). The
column temperature was maintained at 338 K. All concentrations of sugars and organic
products in the aqueous phase were determined by external calibration standards.
The pH value of the reaction solution prior to the reaction was measured on a WD-3563440 Waterproof Double Junction pH (± 0.01 pH units) calibrated with standard buffer
solutions. MATLAB was used to simulate the reactant and product profile. 4th order
Runge-Kutta method was used to solve ODE.
53
4.2.3
Mass Spectrometry Experiments.
All mass spectrometry experiments were performed on a Thermo Scientific LTQ linear
quadrupole ion trap (LQIT) mass spectrometer equipped with an electrospray ionization
(ESI) source. Samples were analyzed by direct injection and were introduced into the mass
spectrometer via syringe drive at a flow rate of 10-20μL min−1. The ESI source conditions
used were as follows: 0.8-1 kV spray voltage, flow of 60 (arbitrary units) sheath gas, flow
of 20 (arbitrary units) auxiliary gas, and a 275 °C transfer capillary temperature. All ion
optic voltages were set using the LTQ Tune Plus interfaces tuning features. The nominal
pressure within the instrument, as read by ion gauges, was maintained below 1 ×10−5 Torre
in the linear quadrupole ion trap.
4.2.4
Carbon nuclear magnetic resonance (C-NMR) Analysis.
13
C liquid NMR data were obtained from a Bruker ARX400-MHz NMR spectrometer
(Bruker, Fremont, CA). It is equipped with a 5mm QNP (H1, F19, P31, C13) probe and a
sample temperature control unit. It has an SGI O2 host computer running XwinNMR 2.6.
Spectra were accumulated using a 5 μs pulse width at 300 K. Samples are prepared in
deuterated water.
4.2.5
Kinetic Parameter Estimation.
We assumed that the reactor was well mixed and isothermal during the reaction period.
We used a monophasic model based on homogeneous, pseudo first-order reaction kinetics.
The glucose was first isomerized to make fructose with the Lewis acid and then dehydrated
to generate HMF. HMF was a reactant being further rehydrated to make levulinic acid.
Glucose, fructose and HMF are reactants theoretically capable of generating humins (van
54
Zandvoort et al., 2013). The simplified overall kinetic model used in this study is shown in
Figure 4-1.
Figure 4-1 Reaction pathway for glucose upgrade to levulinic acid
The kinetic parameter estimation is based on the following set of differential equations
(4-1)-(4-4):
𝑑[𝐺𝑙𝑢𝑐𝑜𝑠𝑒]
𝑑𝑡
𝑑[𝐹𝑟𝑢𝑡𝑜𝑠𝑒]
𝑑𝑡
𝑑[𝐻𝑀𝐹]
𝑑𝑡
= −(𝑘1 + 𝑘4 )[𝐺𝑙𝑢𝑐𝑜𝑠𝑒] + 𝑘−1 [𝐹𝑟𝑢𝑐𝑡𝑜𝑠𝑒]
(4-1)
= 𝑘1 [𝐺𝑙𝑢𝑐𝑜𝑠𝑒] − (𝑘−1 + 𝑘2 + 𝑘5 )[𝐹𝑟𝑢𝑡𝑜𝑠𝑒]
(4-2)
= 𝑘2 [𝐹𝑟𝑢𝑐𝑡𝑜𝑠𝑒] − （𝑘6 + 𝑘3 ）[𝐻𝑀𝐹]
𝑑[𝐿𝑒𝑣𝑢𝑙𝑖𝑛𝑖𝑐 𝐴𝑐𝑖𝑑]
𝑑𝑡
= 𝑘3 [𝐻𝑀𝐹]
(4-3)
(4-4)
Reactions were conducted using pure glucose, fructose or HMF as the reactant to
measure the initial reaction rates as shown in Table 4-2.
55
The rate constants (kx) in the equations were determined by plotting natural logarithm
of reactant/reactant initial (time = 0) versus time (per minute) using the initial time points
(0, 1, 2, 3 min) of the reaction. In the case when HMF was used as the reactant, two parallel,
pseudo first-order reactions occurred. The first converts HMF to equimolar amounts of
levulinic and formic acids. Simultaneously a parallel reaction produces humins from HMF
and HMF derived intermediates (Qi & Horváth, 2012). Apparent first-order rate
coefficients 𝑘3 plus 𝑘6 for reactions in Figure 4-1 was the slope of plotting the natural
logarithm of [𝐻𝑀𝐹]/[𝐻𝑀𝐹]0 vs time. Measuring the negative concentration of levulinic
acid vs concentration of HMF would have a slope equal to 𝑘3 /(𝑘3 + 𝑘6 ). Rate constants
𝑘3 and 𝑘6 could be determined algebraically.
In the study, we used a constant catalyst concentration (100 mM) in the solution. It should
be noted that pH plays a key role in controlling the Lewis acidity of metal halides to form
catalytically active metal species in aqueous solution, which could significantly impact the
glucose ring-opening step and subsequent dehydration and rehydration reactions. In our
experiment, the mixed catalysts of maleic acid with AlCl3 gives a pH of 1.2 and HCl
combined with AlCl3 gives a pH value of 1.1.
Table 4-2 Kinetic rate constant (kdisappearance rate constant with confidence interval of 95%, all other rate constant was
calculated based on [product]/[reactant] without confidence interval. )
Temperature
Catalyst
kdis (min-1)
k-1 (min-1)
k2 (min-1)
k3 (min-1)
k5(min-1)
k6(min-1)
AlCl3+HCl
0.93±0.17
0.17
0.6
0.09
0.23
0.09
AlCl3+MA
0.23±0.04
0.09
1.00
0.11
0.09
0.06
AlCl3+HCl
0.15±0.01
0.03
0.09
0.03
0.06
0.02
AlCl3+MA
0.05±0.03
0.03
0.14
0.1
0.02
0.03
AlCl3+HCl
0.02±0.003
0.007
0.022
0.014
0.015
0.01
AlCl3+MA
0.024±0.002
0.01
0.017
0.03
0.005
0.01
180 oC
160 oC
o
140 C
Note: The 180 oC kinetic constants was remade from Reference (Zhang et al., 2015), the kinetic constants shown in the table is
described in Figure 4-1, where kdis=k1+ k4
56
57
4.3
4.3.1
Results and Discussion
Comparison of glucose conversion using aluminum chloride combined with HCl
and maleic acid
Elevated temperatures (140, 160 oC) are tested as illustrated. The results for experimental
and simulated data are shown in Figure 4-2. At 140 oC, HCl and maleic acid behave
differently. For the HCl combined with AlCl3, 30% of the glucose has disappeared and
converted into products or unidentified intermediate during the heating of the reaction
mixture to 140 °C. Glucose isomerizes to fructose gradually from the beginning until 14
min. With increasing reaction time, fructose to glucose ratios stabilize and fructose
concentration is higher than glucose, meanwhile, the HMF and levulinic acid
concentrations are constant, which means that do dehydration is occurring and glucose and
fructose are at equilibrium. However, maleic acid combined with AlCl3 gives a different
reaction characteristics at 140 oC. Only 5% of glucose is reacted during heat up and the
glucose concentration at equilibrium is significantly higher at 14 minutes. HMF and
levulinic acid are formed in small amounts. It is consistent with previous study in glucose
conversion to make HMF, results showed at 180 oC, maleic acid combined with AlCl3 has
different behavior compared to HCl combined with AlCl3 as more glucose is left after the
reaction, and maleic acid has better selectivity compared HCl with both combined with
AlCl3 (Zhang et al., 2015). Maleic acid is also found as effective as sulfuric acid in
hydrolysis of cellulose and cellubiose to glucose but not as easy as sulfuric acid to degrade
glucose to HMF and levulinic acid (N. S. Mosier et al., 2001). The pH value of the solution
is a key factor under alkaline condition, while there is no deep research on equilibrium of
isomerization step between glucose and fructose under acidic condition (Román‐Leshkov
58
et al., 2010). As two combinations of catalysts gives different pH values (HCl with AlCl3
generates pH 1.1, maleic acid with AlCl3 generates pH 1.2), a separated sample was tested
to exclude the possibility of acidity of the solution gives the different reactant and product
profile. The sample was prepared with AlCl3 (100mM) and using HCl to adjust pH to 1.2.
The sample was tested with 24 min, 140 oC, and the results in Table 4-3 shows even at pH
matches between HCl combined with AlCl3 and maleic acid combined with AlCl3, the ratio
of glucose unreacted over fructose yield still has a significant difference (1.9 to 5.6), which
means the different product profile is not caused by acidity of the solution.
For AlCl3 combined with HCl at 160 oC, glucose at initial point only has 66% left
unreacted, less than 70% of glucose unreacted at 140 oC, the glucose decreased rapidly to
fructose and after 30 minutes only trace amount of glucose can be detected. The fructose
concentration rapid increased before six minutes, and gradually decreased as HMF and
levulinic acid were formed. HMF yield increased to 20% and remained consistent until 22
min before decreasing while levulinic acid increased to 8% at 14 min remained consistent
until the end of the reaction. This indicates HMF degraded to humins mostly rather than
conversion to levulinic acid after 22 min. Similar experiment conducted by Yang using a
biphasic system (H2O/THF), AlCl3 and HCl as catalyst under 160 oC for 5 min, the glucose
conversion is about 30%, which is consistent with this experiment (Y. Yang, Hu, & AbuOmar, 2013).
59
90%
(a)
80%
70%
Yield (%)
60%
50%
40%
30%
20%
10%
0%
0
5
10
15
20
25
Reaction time (min)
30
35
40
(b)
90%
80%
70%
Yield (%)
60%
50%
40%
30%
20%
10%
0%
0
5
10
15
20
25
Reaction time (min)
30
35
40
60
90%
(c)
80%
70%
Yield (%)
60%
50%
40%
30%
20%
10%
0%
0
5
10
15
20
25
Reaction time (min)
30
35
90%
40
(d)
80%
70%
Yield (%)
60%
50%
40%
30%
20%
10%
0%
0
5
10
15
20
25
Reaction time (min)
30
35
40
Figure 4-2 Reactant and product yields observed during the reaction with (a) AlCl3 and
HCl as catalyst at 140 oC; (b) AlCl3 and maleic acid as catalyst at 140 oC; (c) AlCl3 and
HCl as catalyst at 160 oC; (d) AlCl3 and maleic acid as catalyst at 160 oC. Shape represents
experimental data, line represents simulated data. Glucose unreacted = ○, fructose = □,
HMF = ◊, levulinic acid = ∆, formic acid is not shown. Reaction conditions: glucose (250
mM); maleic acid (100 mM); HCl (100 mM); AlCl3 (100 mM). Error bars represent
standard deviation for triplicates.
61
Table 4-3 Reaction comparison at 24 min, 140 oC
a
Catalyst
pH
Glucose
unreacted
(%)
AlCl3+HCl
1.1
27.6(3.1)a
34.4(5.2)
0.8
9.5(0.3)
1.4(0.5)
AlCl3+HCl
1.2
48.9(0.2)
25.1(2.1)
1.9
3.5(0.6)
0.5(0.1)
AlCl3+MA
1.2
82.8(0.3)
14.9(1.5)
5.6
2.1(0.8)
0.7(0.01)
Fructose
yield (%)
𝐺𝑙𝑢𝑐𝑜𝑠𝑒 𝑢𝑛𝑟𝑒𝑎𝑐𝑡𝑒𝑑𝑒𝑑
𝐹𝑟𝑢𝑐𝑡𝑜𝑠𝑒 𝑦𝑖𝑒𝑙𝑑
HMF
yield (%)
Levulinic
acid yield
(%)
Standard deviation for triplicate experimental data.
For AlCl3 combined with maleic acid at 160 oC, glucose at initial point has 89% left
unreacted, much more than 66% unreacted under AlCl3 combined with HCl at 160 oC,
slightly lower compared to same catalysts under 140 oC. The glucose decreased gradually
to fructose and at end of reaction, leaving 20% glucose unreacted, which significant amount
is left not reacting. The HMF and levulinic acid experimental and simulated data fits well,
at the end of reaction, 30% levulinic acid yield and 12% HMF yield are got, when glucose
and fructose has significant amount left unreacted.
The comparison of both 140 and 160 oC under AlCl3 both combined with HCl and maleic
acid, shows the different reaction mechanism between HCl and maleic acid in facilitating
isomerization of glucose and fructose, at 140 oC, maleic acid can stabilize glucose and not
easy to convert to fructose and HCl can facilitate the fructose in larger amount and keep
the ratio constant. At 160 oC, as higher temperature applied, glucose and fructose
isomerization equilibrium cannot last, most of the sugars converted to HMF and levulinic
62
acid, this is similar with 180 oC, however, maleic acid combined with AlCl3 has a higher
selectivity from desire product to humins.
4.3.2
Analysis of Kinetics
The estimated kinetic parameters for elevated temperature are shown in Table 4-2. For
AlCl3 combined with HCl, at 180 oC, the glucose disappearance rate constant is much
higher than AlCl3 combined with maleic acid, the difference is smaller at 160 oC, however,
at 140 oC, maleic acid combined with AlCl3 is slightly higher than HCl combined with
AlCl3. The same trend happens for fructose to glucose rate constant (k-1), this indicate
maleic acid has interaction with glucose to slow down the reaction. This phenomenon
shows maleic acid has a behavior not wholly depends on the proton concentration in the
solution. For the selectivity of the reaction, Table 4-4 lists the selectivity for these two
catalysts combination at elevated temperature.
Table 4-4 Kinetic constant selectivity
Catalyst
𝑘2
Fructose to HMF
=
𝑘5 Fructose to Humins
𝑘3
HMF to LA
=
𝑘6 HMF to Humins
Temperature (oC)
140
160
180
140
160
180
AlCl3+HCl (pH 1.1)
1.5
1.5
2.6
1.4
1.5
1.0
AlCl3+MA (pH 1.2)
3.4
7
11
3
3.3
1.8
63
For fructose dehydration to HMF step, maleic acid increases the selectivity as
temperature increase, while HCl has lower selectivity compared to maleic acid. For HMF
rehydrated to make levulinic acid, selectivity decreases for both catalysts combination and
but maleic acid still has higher selectivity to HCl.
To understand the reaction mechanism, activation energy and Gibbs free energy is
calculated based on kinetic constant data. Table 4-5 shows maleic acid significant lower
activation energy for glucose and fructose isomerization. But for fructose to HMF and
fructose to humins step, both has higher activation energy compared to HCl, this suggests
that maleic acid inhibits these reactions. Similar trends are found in Gibbs free energy.
64
Table 4-5 Activation energy and Gibbs free energy
Ea(kJ/mol)
Reaction
step
Glucose to
fructose
Fructose
to Glucose
Fructose
to HMF
HMF to
LA
Fructose
to Humins
HMF to
Humins
MA+
ΔH(kJ/mol)
ΔS(J/molK)
ΔG(kJ/mol)
ΔG(kJ/mol)
enthalpy
entropy
Maleic acid+AlCl3
HCl+AlCl3
HCl+
MA+
HCl+
MA+
HCl+
296.
413.
433.
453.1
296.
413.
433.
453.
AlCl3
AlCl3
AlCl3
AlCl3
AlCl3
AlCl3
15 K
15 K
15 K
5K
15 K
15 K
15 K
15 K
94.5
149
91
146
-94
40
119
130
132
134
134
129
129
128
85.4
124
82
120
-121
-32
118
132
134
137
129
133
134
135
158
128
155
125
60
-12
137
130
129
128
129
130
130
130
51.1
72.1
47
69
-194
-151
104
127
131
135
114
131
134
137
112
106
109
103
-63
-68
128
135
136
138
123
131
132
134
69.9
89.1
66
85
-159
-115
113
132
135
138
119
133
135
137
65
4.3.3
Speciation of complex in solution
Based on hypothesis that maleic acid may have interaction with glucose, 13C NMR was
used to study the isomerization of glucose. A comparison of 13C NMR spectra of samples
is shown in Figure 4-3 and Table 4-6. Results shows when maleic acid mixed with glucose,
a new peak at 189 ppm was detected. The same peak was detected when AlCl3 was mixed
with glucose. But the Peak 189 was not detected either by glucose mixing with maleic acid
combined with AlCl3 or glucose mixing with HCl combined with AlCl3. D-glucose in the
water solution mainly exists in two cyclic forms: α-D-glucopyranose and β-Dglucopyranose. The acyclic form of D-glucose is formed as intermediate for hydride shift
during isomerization to D-fructose. As the acyclic form of D-glucose is thermodynamically
unstable, there is trace amount of acyclic D-glucose can be measured. Previous study
showed glucose can chelate with tin-containing zeolite to form intermediate, in which Dglucose is in chain form. Our results showed both aluminum ion and maleic acid can
interact with glucose to form intermediate. To further confirm
13
C NMR results, mass
spectrometer was used to check intermediate formation (Figure 4-4 and Table 4-7). Mass
spectrum shows two glucose coordination with the one aluminum ion, and two maleic acid
and one glucose can form with one aluminum ion. 13C NMR and mass spectrometer results
showed maleic acid has a catalyst behavior similar as Lewis acid, also maleic acid can
facilitates D-glucose chain open and can stabilize the open chain form glucose.
66
(a)
(b)
67
(c)
(d)
68
(e)
(f)
Figure 4-3 13C NMR (a) D-Glucose in D2O (b) D-Glucose and maleic acid in D2O (c)
Maleic acid in D2O (d) AlCl3 and glucose in D2O (e) Maleic acid, AlCl3 and glucose in
D2O (f) HCl, AlCl3 and glucose in D2O, pH=1.2. Reaction conditions: glucose (250 mM);
maleic acid (100 mM); AlCl3 (100 mM).
69
Table 4-6 Summery for NMR experiment
Sample
β:α
(glucopyranose)
pH
Glucose
1.5
6
MA+Glucose
0.7
1.7
√
AlCl3+Glucose
1.6
2.7
√
MA+AlCl3+Glucose
0.68
1.2
HCl+AlCl3+Glucose
0.53
1.2
Peak 189
70
(a)
(b)
(c)
Figure 4-4 Mass spectrometer for glucose mixing with maleic acid and AlCl3. Reaction
conditions: glucose (250 mM); maleic acid (100 mM); HCl (100 mM); AlCl3 (100 mM).
71
Table 4-7 Mass spectrometer: complex formed under different condition.
Condition
(a)
(b)
(c)
MA+Glucose
AlCl3+Glucose
AlCl3+MA+Glucose
[Al+2GLU-2H]+
Complex
4.4
[Al+2GLU-2H]+
[Al+GLU+2MA-2H]+
Conclusions
Maleic acid combined with AlCl3 shows a superior selectivity and inhibits humins
generation in transformation of glucose to HMF and levulinic acid at elevated temperature,
compared to HCl combined with AlCl3. Calculated activation energy shows maleic acid
lowers energy barrier for the sugar isomerization step. The NMR data suggests maleic acid
has the unique performance to stabilize glucose in acyclic form. Further mass spectrometry
shows complex formed by maleic acid chelated with AlCl3 and glucose may direct the
reaction pathway different from previous proposed catalysis by Lewis acid. Previous study
found that maleic acid alone can behave like Lewis acid, such as AlCl 3, to isomerize
glucose to fructose, while generate much less humins. Combined with the recent data
support, maleic acid is a promising catalyst candidate in upgrading glucose to HMF
production.
ACKNOWLEDGEMENT
This material is based upon work supported as part of the Center for Direct Catalytic
Conversion of Biomass to Biofuels (C3Bio), an Energy Frontier Research Center funded
72
by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences,
Award Number DE-SC0000997.
73
CHAPTER 5.
CONCLUSIONS AND RECOMMENDATION
This work demonstrated that maleic acid can be used as cocatalyst combined with
AlCl3 in conversion glucose to make HMF. Maleic acid, a carboxylic acid, behaves
different from Brønsted acid and follows different mechanistic pathway from previous
proposed Brønsted acid. Maleic acid demonstrates the chelated intermediates formed with
glucose through the isomerization step can help to reduce the humins generation and
redistribute the product profile.
For the glucose isomerization step, maleic acid alone showed much slower reaction
rate constant compared with HCl alone at 180 oC. NMR data justified maleic acid help
glucose open ring reaction and stabilized the open ring form, while HCl did not. The
novel found is maleic acid itself can act as Lewis acid, which facilitates fructose
formation from glucose in a lower rate constant. Elevated temperature tested at 140, 160,
and 180 oC, maleic acid combined with AlCl3 significantly change the selectivity for
HMF production as well as inhibit the humins generation compared to HCl combined
with AlCl3. Calculated activation energy shows maleic acid lowers energy barrier for the
sugar isomerization step. The NMR data suggests maleic acid has the unique performance
to stabilize glucose in acyclic form. Further mass spectrometry shows complex formed
by maleic acid chelated with AlCl3 and glucose may direct the reaction pathway different
from previous proposed catalysis by Lewis acid.
74
For the reaction start from HMF at 180 oC, at the longest reaction time tested (6 min),
the amount of humins generated by maleic acid combined with AlCl3 were only 50% of
the total amount of humins generated by HCl with AlCl3. This results in an increased
selectivity toward levulinic acid formation. It may be concluded that, instead of rate
acceleration under catalysis by strong acids such as HCl, the product distribution is more
strongly affected by organic acids such as maleic acid.
Maleic acid, as previous study showed, is effective in xylose to furfural generation; it
is also used as biomimetic catalyst on cellulose to glucose hydrolysis as well as
hemicellulose to xylose hydrolysis. The unique molecular structure make it has similar
function as enzyme has. The broad study shows maleic acid has a potential to be used as
an effective catalyst through the steps of lignocellulose hydrolysis, sugar isomerization
and dehydration, with a lower degradation product formed. More study should conduct
on the thermal stability and reusability for maleic acid. Synthetic catalyst is also a future
direction, as the results showed maleic acid can form chelated complex with Al ion and
glucose in water. The chelated complex need to be further studied to figure out the spatial
structure.
REFERENCES
75
REFERENCES
Allen, Karen N, Lavie, Arnon, Farber, Gregory K, Glasfeld, Arthur, Petsko, Gregory A, &
Ringe, Dagmar. (1994). Isotopic exchange plus substrate and inhibition kinetics of
D-xylose isomerase do not support a proton-transfer mechanism. Biochemistry,
33(6), 1481-1487.
Alonso, David Martin, Wettstein, Stephanie G, & Dumesic, James A. (2013). Gammavalerolactone, a sustainable platform molecule derived from lignocellulosic
biomass. Green Chemistry, 15(3), 584-595.
Antal, Michael Jerry, Mok, William SL, & Richards, Geoffrey N. (1990). Mechanism of
formation of 5-(hydroxymethyl)-2-furaldehyde from D-fructose and sucrose.
Carbohydrate research, 199(1), 91-109.
Assary, Rajeev S, Kim, Taejin, Low, John J, Greeley, Jeff, & Curtiss, Larry A. (2012).
Glucose and fructose to platform chemicals: understanding the thermodynamic
landscapes of acid-catalysed reactions using high-level ab initio methods. Physical
Chemistry Chemical Physics, 14(48), 16603-16611.
Assary, Rajeev S, Redfern, Paul C, Greeley, Jeffrey, & Curtiss, Larry A. (2011).
Mechanistic insights into the decomposition of fructose to hydroxy methyl
furfural in neutral and acidic environments using high-level quantum chemical
methods. The Journal of Physical Chemistry B, 115(15), 4341-4349.
75
Bhosale, Snehalata H, Rao, Mala B, & Deshpande, Vasanti V. (1996). Molecular and
industrial aspects of glucose isomerase. Microbiological reviews, 60(2), 280-300.
Boisen, A, Christensen, TB, Fu, Wenjing, Gorbanev, YY, Hansen, TS, Jensen, JS, . . .
Ståhlberg, Tim. (2009). Process integration for the conversion of glucose to 2, 5furandicarboxylic acid. Chemical engineering research and design, 87(9), 13181327.
Bordwell, FG, Zhang, Shizhong, Eventova, Irina, & Rappoport, Zvi. (1997). Acidities of
Stable Simple Enols and Bond Dissociation Enthalpies of Their HO Bonds in
DMSO. The Journal of Organic Chemistry, 62(16), 5371-5373.
Brown, Tristan R, & Brown, Robert C. (2013). A review of cellulosic biofuel
commercial ‐ scale projects in the United States. Biofuels, bioproducts and
biorefining, 7(3), 235-245.
Caratzoulas, Stavros, & Vlachos, Dionisios G. (2011). Converting fructose to 5hydroxymethylfurfural: a quantum mechanics/molecular mechanics study of the
mechanism and energetics. Carbohydrate research, 346(5), 664-672.
Casey, Elizabeth, Sedlak, Miroslav, Ho, Nancy WY, & Mosier, Nathan S. (2010). Effect
of acetic acid and pH on the cofermentation of glucose and xylose to ethanol by a
genetically engineered strain of Saccharomyces cerevisiae. FEMS yeast research,
10(4), 385-393.
76
Choudhary, Vinit, Mushrif, Samir H, Ho, Christopher, Anderko, Andrzej, Nikolakis,
Vladimiros, Marinkovic, Nebojsa S, . . . Vlachos, Dionisios G. (2013). Insights into
the interplay of Lewis and Brønsted acid catalysts in glucose and fructose
conversion to 5-(hydroxymethyl) furfural and levulinic acid in aqueous media.
Journal of the American Chemical Society, 135(10), 3997-4006.
Choudhary, Vinit, Pinar, Ana B, Lobo, Raul F, Vlachos, Dionisios G, & Sandler, Stanley
I. (2013). Comparison of Homogeneous and Heterogeneous Catalysts for
Glucose‐to‐Fructose Isomerization in Aqueous Media. ChemSusChem, 6(12),
2369-2376.
Chundawat, Shishir PS, Beckham, Gregg T, Himmel, Michael E, & Dale, Bruce E. (2011).
Deconstruction of lignocellulosic biomass to fuels and chemicals. Annual review of
chemical and biomolecular engineering, 2, 121-145.
Chung, IS, & Lee, YY. (1985). Ethanol fermentation of crude acid hydrolyzate of cellulose
using high‐level yeast inocula. Biotechnology and bioengineering, 27(3), 308-315.
Claassen, PAM, Van Lier, JB, Contreras, AM Lopez, Van Niel, EWJ, Sijtsma, L, Stams,
AJM, . . . Weusthuis, RA. (1999). Utilisation of biomass for the supply of energy
carriers. Applied microbiology and biotechnology, 52(6), 741-755.
Climent, Maria J, Corma, Avelino, & Iborra, Sara. (2014). Conversion of biomass platform
molecules into fuel additives and liquid hydrocarbon fuels. Green Chemistry, 16(2),
516-547.
da Costa Sousa, Leonardo, Chundawat, Shishir PS, Balan, Venkatesh, & Dale, Bruce E.
(2009). ‘Cradle-to-grave’assessment of existing lignocellulose pretreatment
technologies. Current opinion in biotechnology, 20(3), 339-347.
77
Dorrestijn, Edwin, Laarhoven, Lucas JJ, Arends, Isabel WCE, & Mulder, Peter. (2000).
The occurrence and reactivity of phenoxyl linkages in lignin and low rank coal.
Journal of Analytical and Applied Pyrolysis, 54(1), 153-192.
Dutta, Saikat, De, Sudipta, & Saha, Basudeb. (2012). A Brief Summary of the Synthesis
of
Polyester
Building ‐ Block
Chemicals
and
Biofuels
from
5 ‐
Hydroxymethylfurfural. ChemPlusChem, 77(4), 259-272.
El Khadem, Hassan S, Ennifar, Sofiane, & Isbell, Horace S. (1989). Evidence of stable
hydrogen-bonded ions during isomerization of hexoses in alkali. Carbohydrate
Research, 185(1), 51-59.
Freudenberg, Karl, & Neish, Arthur Charles. (1968). Constitution and biosynthesis of
lignin. Constitution and biosynthesis of lignin.
Fringuelli, Francesco, Pizzo, Ferdinando, & Vaccaro, Luigi. (2001). Lewis-acid catalyzed
organic reactions in water. The case of AlCl3, TiCl4, and SnCl4 believed to be
unusable in aqueous medium. The Journal of organic chemistry, 66(13), 4719-4722.
Fructose and glucose prices.). Retrieved Aug
28th, , 2014, from http://www.alibaba.com/showroom/glucose-fructose-syrup.html
Fukuzumi, Shunichi, Kobayashi, Takeshi, & Suenobu, Tomoyoshi. (2008). Efficient
Catalytic Decomposition of Formic Acid for the Selective Generation of H2 and
H/D Exchange with a Water‐Soluble Rhodium Complex in Aqueous Solution.
ChemSusChem, 1(10), 827-834.
Galbe, Mats, & Zacchi, Guido. (2002). A review of the production of ethanol from
softwood. Applied microbiology and biotechnology, 59(6), 618-628.
78
Gellerstedt, Göran, & Henriksson, Gunnar. (2008). Lignins: major sources, structure and
properties: Amsterdam, the Netherlands: Elsevier.
Goldstein, Irving S. (1981). Organic chemicals from biomass. Organic chemicals from
biomass.
Gosselink, RJA, De Jong, E, Guran, B, & Abächerli, A. (2004). Co-ordination network for
lignin—standardisation,
production
and
applications
adapted
to
market
requirements (EUROLIGNIN). Industrial Crops and Products, 20(2), 121-129.
Hayes, Daniel J, & Hayes, Michael HB. (2009). The role that lignocellulosic feedstocks
and various biorefining technologies can play in meeting Ireland's biofuel targets.
Biofuels, Bioproducts and Biorefining, 3(5), 500-520.
Horvat, Jaroslav, Klaić, Branimir, Metelko, Biserka, & Šunjić, Vitomir. (1985).
Mechanism of levulinic acid formation. Tetrahedron letters, 26(17), 2111-2114.
Jönsson, Leif J, Alriksson, Björn, & Nilvebrant, Nils-Olof. (2013). Bioconversion of
lignocellulose: inhibitors and detoxification. Biotechnol Biofuels, 6(1), 16.
Kamm, B, & Kamm, M. (2007). Biorefineries–multi product processes White
Biotechnology (pp. 175-204): Springer.
Kaplan, David L. (1998). Introduction to biopolymers from renewable resources: Springer.
Kim, Eurick S, Liu, Shuo, Abu-Omar, Mahdi M, & Mosier, Nathan S. (2012). Selective
conversion of biomass hemicellulose to furfural using maleic acid with microwave
heating. Energy & Fuels, 26(2), 1298-1304.
Klein‐Marcuschamer, Daniel, Oleskowicz‐Popiel, Piotr, Simmons, Blake A, & Blanch,
Harvey W. (2012). The challenge of enzyme cost in the production of
lignocellulosic biofuels. Biotechnology and Bioengineering, 109(4), 1083-1087.
79
Kobayashi, Shū, Ueno, Masaharu, Saito, Susumu, Mizuki, Yumiko, Ishitani, Haruro, &
Yamashita, Yasuhiro. (2004). Air-stable, storable, and highly efficient chiral
zirconium catalysts for enantioselective Mannich-type, aza Diels–Alder, aldol, and
hetero Diels–Alder reactions. Proceedings of the National Academy of Sciences of
the United States of America, 101(15), 5476-5481.
Kootstra, A Maarten J, Beeftink, Hendrik H, Scott, Elinor L, & Sanders, Johan PM. (2009).
Optimization of the dilute maleic acid pretreatment of wheat straw. Biotechnology
for Biofuels, 2(31).
Kootstra, A Maarten J, Mosier, Nathan S, Scott, Elinor L, Beeftink, Hendrik H, & Sanders,
Johan PM. (2009). Differential effects of mineral and organic acids on the kinetics
of
arabinose
degradation
under
lignocellulose
pretreatment
conditions.
Biochemical Engineering Journal, 43(1), 92-97.
Kovalevsky, Andrey Y, Hanson, Leif, Fisher, S Zoe, Mustyakimov, Marat, Mason, Sax A,
Forsyth, V Trevor, . . . Carrell, HL. (2010). Metal ion roles and the movement of
hydrogen during reaction catalyzed by D-xylose isomerase: A joint X-ray and
neutron diffraction study. Structure, 18(6), 688-699.
Kuster, B. F. M. (1990). 5-Hydroxymethylfurfural (HMF). A Review Focussing on its
Manufacture. Starch - Stärke, 42(8), 314-321. doi: 10.1002/star.19900420808
Larsson, Simona, Palmqvist, Eva, Hahn-Hägerdal, Bärbel, Tengborg, Charlotte, Stenberg,
Kerstin, Zacchi, Guido, & Nilvebrant, Nils-Olof. (1999). The generation of
fermentation inhibitors during dilute acid hydrolysis of softwood. Enzyme and
Microbial Technology, 24(3), 151-159.
80
Li, Yu-Nong, Wang, Jin-Quan, He, Liang-Nian, Yang, Zhen-Zhen, Liu, An-Hua, Yu, Bing,
& Luan, Chao-Ran. (2012). Experimental and theoretical studies on imidazolium
ionic liquid-promoted conversion of fructose to 5-hydroxymethylfurfural. Green
Chemistry, 14(10), 2752-2758.
Lichts, FO. (2011). Industry statistics: 2010 world fuel ethanol production. Renewable
Fuels
Association,
accessed
in
HYPERLINK
“http://www.
ethanolrfa.
org/pages/statistics# E.
Liu, F, Boissou, F, Vignault, A, Lemée, L, Marinkovic, S, Estrine, B, . . . Jérôme, F. (2014).
Conversion of wheat straw to furfural and levulinic acid in a concentrated aqueous
solution of betaïne hydrochloride. RSC Advances, 4(55), 28836-28841.
Liu, ZL, Slininger, PJ, Dien, BS, Berhow, MA, Kurtzman, CP, & Gorsich, SW. (2004).
Adaptive response of yeasts to furfural and 5-hydroxymethylfurfural and new
chemical evidence for HMF conversion to 2, 5-bis-hydroxymethylfuran. Journal of
Industrial Microbiology and Biotechnology, 31(8), 345-352.
Lu, Yulin, & Mosier, Nathan S. (2007). Biomimetic catalysis for hemicellulose hydrolysis
in corn stover. Biotechnology progress, 23(1), 116-123.
Lu, Yulin, & Mosier, Nathan S. (2008). Kinetic modeling analysis of maleic acid ‐
catalyzed hemicellulose hydrolysis
in
corn
stover.
Biotechnology and
bioengineering, 101(6), 1170-1181.
McMichael, Philip. (2009). Banking on agriculture: a review of the World Development
Report 2008. Journal of Agrarian Change, 9(2), 235-246.
81
Mosier, Nathan S, Ladisch, Christine M, & Ladisch, Michael R. (2002). Characterization
of acid catalytic domains for cellulose hydrolysis and glucose degradation.
Biotechnology and bioengineering, 79(6), 610-618.
Mosier, Nathan S, Sarikaya, Ayda, Ladisch, Christine M, & Ladisch, Michael R. (2001).
Characterization of dicarboxylic acids for cellulose hydrolysis. Biotechnology
progress, 17(3), 474-480.
Mosier, Nathan, Wyman, Charles, Dale, Bruce, Elander, Richard, Lee, YY, Holtzapple,
Mark, & Ladisch, Michael. (2005). Features of promising technologies for
pretreatment of lignocellulosic biomass. Bioresource technology, 96(6), 673-686.
Murzin, Dmitry Yu, & Salmi, Tapio. (2012). Catalysis for Lignocellulosic Biomass
Processing: Methodological Aspects. Catalysis letters, 142(7), 817-829.
Nagorski, RW, & Richard, John P. (2001). Mechanistic imperatives for aldose-ketose
isomerization in water: Specific, general base-and metal ion-catalyzed
isomerization of glyceraldehyde with proton and hydride transfer. Journal of the
American Chemical Society, 123(5), 794-802.
Pagan-Torres, Yomaira J, Wang, Tianfu, Gallo, Jean Marcel R, Shanks, Brent H, &
Dumesic, James A. (2012). Production of 5-hydroxymethylfurfural from glucose
using a combination of Lewis and Brønsted acid catalysts in water in a biphasic
reactor with an alkylphenol solvent. Acs Catalysis, 2(6), 930-934.
Palmqvist, Eva, & Hahn-Hägerdal, Bärbel. (2000). Fermentation of lignocellulosic
hydrolysates. I: inhibition and detoxification. Bioresource technology, 74(1), 1724.
82
Parsell, Trenton H, Owen, Benjamin C, Klein, Ian, Jarrell, Tiffany M, Marcum,
Christopher L, Haupert, Laura J, . . . Miller, Jeffrey T. (2013). Cleavage and
hydrodeoxygenation (HDO) of C–O bonds relevant to lignin conversion using
Pd/Zn synergistic catalysis. Chemical Science, 4(2), 806-813.
Parsell, Trenton, Yohe, Sara, Degenstein, John, Jarrell, Tiffany, Klein, Ian, Gencer,
Emre, . . . Choudhari, Harshavardhan. (2015). A synergistic biorefinery based on
catalytic conversion of lignin prior to cellulose starting from lignocellulosic
biomass. Green Chemistry.
Pemba, Alexander G, Rostagno, Mayra, Lee, Tanner A, & Miller, Stephen A. (2014).
Cyclic and spirocyclic polyacetal ethers from lignin-based aromatics. Polymer
Chemistry, 5(9), 3214-3221.
Pérez, J, Munoz-Dorado, J, de la Rubia, TDLR, & Martinez, J. (2002). Biodegradation and
biological treatments of cellulose, hemicellulose and lignin: an overview.
International Microbiology, 5(2), 53-63.
Qi, Long, & Horváth, István T. (2012). Catalytic Conversion of Fructose to γValerolactone in γ-Valerolactone. ACS Catalysis, 2(11), 2247-2249.
Rahimi, Alireza, Ulbrich, Arne, Coon, Joshua J, & Stahl, Shannon S. (2014). Formic-acidinduced depolymerization of oxidized lignin to aromatics. Nature, 515(7526), 249252.
Ranoux, Adeline, Djanashvili, Kristina, Arends, Isabel WCE, & Hanefeld, Ulf. (2013). 5Hydroxymethylfurfural synthesis from hexoses is autocatalytic. ACS Catalysis,
3(4), 760-763.
83
Ribeiro, Raphael F, Marenich, Aleksandr V, Cramer, Christopher J, & Truhlar, Donald G.
(2011). Use of solution-phase vibrational frequencies in continuum models for the
free energy of solvation. The Journal of Physical Chemistry B, 115(49), 1455614562.
Rinaldi, Roberto, & Schüth, Ferdi. (2009). Acid hydrolysis of cellulose as the entry point
into biorefinery schemes. ChemSusChem, 2(12), 1096-1107.
Ritter, Stephen K. Lignocellulose: A complex biomaterial.
Ritter, Stephen K. (2008). Lignocellulose: A complex biomaterial. Plant Biochemistry,
86(49), 15.
Román-Leshkov, Yuriy, Chheda, Juben N, & Dumesic, James A. (2006). Phase modifiers
promote efficient production of hydroxymethylfurfural from fructose. Science,
312(5782), 1933-1937.
Román‐Leshkov, Yuriy, Moliner, Manuel, Labinger, Jay A, & Davis, Mark E. (2010).
Mechanism of glucose isomerization using a solid Lewis acid catalyst in water.
Angewandte Chemie International Edition, 49(47), 8954-8957.
Rose, Irwin A, O'Connell, Edward L, & Mortlock, Robert P. (1969). Stereochemical
evidence for a cis-enediol intermediate in Mn-dependent aldose isomerases.
Biochimica et Biophysica Acta (BBA)-Enzymology, 178(2), 376-379.
Rowell, Roger M. (2012). Handbook of wood chemistry and wood composites: CRC press.
Saha, Basudeb, & Abu-Omar, Mahdi M. (2014). Advances in 5-hydroxymethylfurfural
production from biomass in biphasic solvents. Green Chemistry, 16(1), 24-38.
Sarkanen, KV, & Ludwig, CH. (1971). Lignins: occurrence and formation, structure,
chemical and macromolecular properties, and utilization: Interscience, New York.
84
Schray, Keith J, & Rose, Irwin A. (1971). Anomeric specificity and mechanism of two
pentose isomerases. Biochemistry, 10(6), 1058-1062.
Serrano-Ruiz, Juan Carlos, West, Ryan M, & Dumesic, James A. (2010). Catalytic
conversion of renewable biomass resources to fuels and chemicals. Annual review
of chemical and biomolecular engineering, 1, 79-100.
Severian, Dumitriu. (2008). Polysaccharides: structural diversity and functional versatility.
Marcel Dekker, New York.
Sievers, Carsten, Musin, Ildar, Marzialetti, Teresita, Valenzuela Olarte, Mariefel B,
Agrawal, Pradeep K, & Jones, Christopher W. (2009). Acid‐catalyzed conversion
of sugars and furfurals in an ionic‐liquid phase. ChemSusChem, 2(7), 665-671.
van Zandvoort, Ilona, Wang, Yuehu, Rasrendra, Carolus B, van Eck, Ernst RH, Bruijnincx,
Pieter CA, Heeres, Hero J, & Weckhuysen, Bert M. (2013). Formation, molecular
structure, and morphology of humins in biomass conversion: influence of feedstock
and processing conditions. ChemSusChem, 6(9), 1745-1758.
vom Stein, Thorsten, Grande, Philipp M, Kayser, Henning, Sibilla, Fabrizio, Leitner,
Walter, & de Marí
a, Pablo Domínguez. (2011). From biomass to feedstock: onestep fractionation of lignocellulose components by the selective organic acidcatalyzed depolymerization of hemicellulose in a biphasic system. Green
Chemistry, 13(7), 1772-1777.
Vorotnikova, Ekaterina, & Seale Jr, James L. (2014). US Ethanol Mandate Is a Hidden
Subsidy to Corn Producers.
85
Wang, Liang, Zhang, Jian, Wang, Xuefeng, Zhang, Bingsen, Ji, Weijie, Meng, Xiangju, . . .
Xiao, Feng-Shou. (2014). Creation of Brønsted acid sites on Sn-based solid
catalysts for the conversion of biomass. Journal of Materials Chemistry A, 2(11),
3725-3729.
Wang, Tianfu, Nolte, Michael W, & Shanks, Brent H. (2014). Catalytic dehydration of C
6 carbohydrates for the production of hydroxymethylfurfural (HMF) as a versatile
platform chemical. Green Chemistry, 16(2), 548-572.
Werpy, Todd, Petersen, Gene, Aden, A, Bozell, J, Holladay, J, White, J, . . . Jones, S.
(2004). Top value added chemicals from biomass. Volume 1-Results of screening
for potential candidates from sugars and synthesis gas: DTIC Document.
Wettstein, Stephanie G, Alonso, David Martin, Gürbüz, Elif I, & Dumesic, James A. (2012).
A roadmap for conversion of lignocellulosic biomass to chemicals and fuels.
Current Opinion in Chemical Engineering, 1(3), 218-224.
Wyman, Charles E. (2007). What is (and is not) vital to advancing cellulosic ethanol.
TRENDS in Biotechnology, 25(4), 153-157.
Wyman, Charles E. (2013). Aqueous pretreatment of plant biomass for biological and
chemical conversion to fuels and chemicals: John Wiley & Sons.
Ximenes, Eduardo, Kim, Youngmi, Mosier, Nathan, Dien, Bruce, & Ladisch, Michael.
(2011). Deactivation of cellulases by phenols. Enzyme and microbial technology,
48(1), 54-60.
Yang, Byung Yun, & Montgomery, Rex. (1996). Alkaline degradation of glucose: effect
of initial concentration of reactants. Carbohydrate Research, 280(1), 27-45.
86
Yang, Yu, Hu, Chang-wei, & Abu-Omar, Mahdi M. (2012). Conversion of carbohydrates
and lignocellulosic biomass into 5-hydroxymethylfurfural using AlCl 3·6H 2 O
catalyst in a biphasic solvent system. Green Chemistry, 14(2), 509-513.
Yang, Yu, Hu, Changwei, & Abu-Omar, Mahdi M. (2013). The effect of hydrochloric acid
on the conversion of glucose to 5-hydroxymethylfurfural in AlCl 3–H 2 O/THF
biphasic medium. Journal of Molecular Catalysis A: Chemical, 376, 98-102.
Zakrzewska, Małgorzata E, Bogel-Łukasik, Ewa, & Bogel-Łukasik, Rafał. (2010). Ionic
liquid-mediated formation of 5-hydroxymethylfurfural
A promising biomass-
derived building block. Chemical reviews, 111(2), 397-417.
Zakzeski, Joseph, Bruijnincx, Pieter CA, Jongerius, Anna L, & Weckhuysen, Bert M.
(2010). The catalytic valorization of lignin for the production of renewable
chemicals. Chemical reviews, 110(6), 3552-3599.
Zhang, Ximing, Hewetson, Barron B, & Mosier, Nathan S. (2015). Kinetics of Maleic Acid
and Aluminum Chloride Catalyzed Dehydration and Degradation of Glucose.
Energy & Fuels.
Zhao, Haibo, Brown, Heather M, Holladay, Johnathan E, & Zhang, Z Conrad. (2012).
Prominent roles of impurities in ionic liquid for catalytic conversion of
carbohydrates. Topics in Catalysis, 55(1-2), 33-37.
Zhao, Xuebing, Zhang, Lihua, & Liu, Dehua. (2012). Biomass recalcitrance. Part I: the
chemical compositions and physical structures affecting the enzymatic hydrolysis
of lignocellulose. Biofuels, Bioproducts and Biorefining, 6(4), 465-482.
87
Zheng, Baohui, Fang, Zhijie, Cheng, Jie, & Jiang, Yuhua. (2010). Microwave-assisted
conversion of carbohydrates into 5-hydroxymethylfurfural catalyzed by ZnCl2.
ChemInform, 41(27), i.
VITA
89
VITA
Ximing Zhang was born in Benxi, Liaoning, China on Feb 23, 1987. After high school,
he entered Shenyang Agricultural University, China, where he studied in college of
engineering with a focus on biosystem and energy engineering. He received his Bachelor’s
degree in July, 2010 and then entered North Carolina State University in the same year for
Master’s degree. He pursued research in process development and optimization for sugar
production from lignocellulosic biomass with pretreatment and enzymatic hydrolysis. He
received his Master of Science degree in Agricultural Engineering in May, 2012. Later, he
joined Laboratory of Renewable Resources Engineering (LORRE) at Purdue University,
West Lafayette, for his PhD study in catalytic conversion of lignocellulose for high value
chemical production.
PUBLICATIONS
90
PUBLICATIONS
1. Zhang, X., Hewetson, B.B., Mosier, N. S., 2015. Kinetics of Maleic Acid and
Aluminum Chloride Catalyzed Dehydration and Degradation of Glucose. Energy
and Fuels. 2015, 29, 2387-2393.
2. Xu, J., Zhang, X., Cheng, J.J., 2012. Pretreatment of Corn Stover for Sugar
Production with Switchgrass-derived Black Liquor. Bioresource Technology. 2012,
111, 256-260.
3. Xu, J., Zhang, X., Ratna Sharma-Shivappa., Mary Eubanks., 2012. Gamagrass
Varieties as Potential Feedstock for Fermentable Sugar Production. Bioresource
Technology. 2012,116, 540-544.
4. Zhang, X., Xu, J., Cheng, J.J., 2011. Pretreatment of Corn Stover for Ethanol
Production by Using the Combination of Alkaline Reagents. Energy and Fuels.
2011, 25(10), 4796-4802
5. Li,Y., Lv, X., Liu Q., Yi W., Zhang, X., 2009. Influence of Biogas Manures on
Nature of Soil. Agricultural Mechanization Research. 2009, 10, 140-142.