University of Groningen
Kinetic study on the acid-catalyzed hydrolysis of cellulose to levulinic acid
Girisuta, B.; Janssen, L. P. B. M.; Heeres, H. J.
Published in:
Industrial and Engineering Chemistry Research
DOI:
10.1021/ie061186z
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from
it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date:
2007
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Girisuta, B., Janssen, L. P. B. M., & Heeres, H. J. (2007). Kinetic study on the acid-catalyzed hydrolysis of
cellulose to levulinic acid.
Industrial and Engineering Chemistry Research
,
46
(6), 1696-1708.
https://doi.org/10.1021/ie061186z
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the
author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
The publication may also be distributed here under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license.
More information can be found on the University of Groningen website: https://www.rug.nl/library/open-access/self-archiving-pure/taverne-
amendment.
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately
and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the
number of authors shown on this cover page is limited to 10 maximum.
Download date: 09-08-2022
Kinetic Study on the Acid-Catalyzed Hydrolysis of Cellulose to Levulinic Acid
B. Girisuta, L. P. B. M. Janssen, and H. J. Heeres*
Department of Chemical Engineering, UniVersity of Groningen, Nijenborgh 4,
9747 AG Groningen, Netherlands
A variety of interesting bulk chemicals is accessible by the acid-catalyzed hydrolysis of cellulose. An interesting
example is levulinic acid, a versatile precursor for fuel additives, polymers, and resins. A detailed kinetic
study on the acid-catalyzed hydrolysis of cellulose to levulinic acid is reported in this paper. The kinetic
experiments were performed in a temperature window of 150-200 °C, sulfuric acid concentrations between
0.05 and 1 M, and initial cellulose intakes between 1.7 and 14 wt %. The highest yield of levulinic was 60
mol %, obtained at a temperature of 150 °C, an initial cellulose intake of 1.7 wt %, and a sulfuric acid
concentration of 1 M. A full kinetic model covering a broad range of reaction conditions was developed
using the power-law approach. Agreement between the experimental data and the kinetic model is good. The
kinetic expressions were used to gain insights into the optimum process conditions for the conversion of
cellulose to levulinic acid in continuous-reactor configurations. The model predicts that the highest obtainable
levulinic acid yield in continuous-reactor configurations is about 76 mol %, which was obtained when using
reactors with a large extent of backmixing.
1. Introduction
Cellulose is a natural polymer consisting of glucose units. It
is abundantly available on earth, and its annual production is
estimated at 2 × 10
9
tons.
1
Cellulose may be converted to inter-
esting bulk chemicals by an acid-catalyzed hydrolysis reaction.
During hydrolysis, the β-(1f4)-glycosidic bonds of cellulose are
cleaved to give glucose, which can be converted further to vari-
ous organic (bulk) chemicals. One attractive option is the conver-
sion of glucose to levulinic acid (4-oxopentanoic acid) by acid
treatment. Levulinic acid is a versatile building block for fuel
additives, polymer precursors, and resin precursors.
2
Several re-
views have been published describing the properties and poten-
tial industrial applications of levulinic acid and its derivatives.
3-6
Two different approaches are commonly applied for the acid-
catalyzed hydrolysis of cellulose. The first uses high concentra-
tions of mineral acids (e.g., 15-16 N HCl or 31-70 wt %
H
2
SO
4
) as catalysts and low operating temperatures (20-50
°C).
7,8
The major drawbacks are the high operating cost of acid
recovery and the use of expensive construction material for both
the hydrolyzer and the acid recovery system. The second
approach uses highly diluted acids at high operating tempera-
tures (170-240 °C). This method is favored, and research
studies applying this approach are abundant.
9-12
Various kinetic studies on the acid-catalyzed hydrolysis using
a range of cellulosic materials have been reported in the
literature. The first systematic kinetic study on biomass hy-
drolysis to glucose was performed in 1945 by Saeman,
13
who
studied the hydrolysis reaction of Douglas fir in batch reactors.
In this study, the hydrolysis reaction is modeled by the following
two consecutive first-order reactions:
The reaction rate constants are represented by modified Arrhe-
nius equations, including the effects of temperature (T) and acid
concentration (A).
Here, k
i,o
is the frequency factor, m
i
is the reaction order in
acid, R is the ideal gas constant, and E
i
is the activation energy.
Further investigations were conducted by Fagan and co-
workers
14
on Kraft paper slurries. A nonisothermal plug-flow
reactor was used to determine the kinetics of the hydrolysis
reaction. Further studies were performed on Solka-Floc
15
and
filter paper
16
in an isothermal plug-flow reactor. Malester and
co-workers
17,18
carried out kinetic studies using municipal solid
waste (MSW) as the cellulose source. The experiments were
carried out ina2Lbatch steel reactor using sulfuric acid in
low concentrations as the catalyst. All these kinetic studies
applied the kinetic model developed earlier by Saeman
13
to
analyze the kinetic data. An overview of kinetic studies
including the range of process conditions and intakes is given
in Table 1. For cellulose decomposition to glucose, the activation
energy is between 172 and 189 kJ mol
-1
. However, large
variations are observed in the order of acid concentration (1.0-
1.78). A similar observation also holds for the decomposition
of glucose to (nonidentified) products, where the order in acid
concentration varies between 0.55 and 1.02.
The acid-catalyzed hydrolysis of cellulose is a heterogeneous
reaction where mass-transfer effects may play an important role
and, under some conditions, may even determine the overall
reaction rate. As such, the dimensions of the cellulosic materials
and their properties (e.g., crystallinity of the cellulose fraction)
may have significant effects on the overall rate of the hydrolysis
reaction. Mass-transfer effects on the overall rate of the
hydrolysis reaction of cellulose were investigated by Saeman
13
by conducting the reaction with various cellulose particle sizes.
The hydrolysis reaction rate was unaffected when using particle
sizes in the range of 20-200 mesh (74-840 µm). Similar results
were obtained by Malester and co-workers.
18
These results imply
that, under these conditions, the hydrolysis reaction of cellulose
can be treated as a homogeneous reaction when the particle size
of cellulose is <20 mesh (840 µm). Sharples
19,20
proposed a
kinetic model including the effects of the degree of crystallinity
of the cellulose on the reaction rate. The cellulose applied in
this study was pretreated with 18 wt % of sodium hydroxide
* To whom correspondence should be addressed. Fax: +31 50 363
4479. Phone: +31 50 363 4174. E-mail: h.j.heeres@rug.nl.
cellulose
9
8
k
1
glucose
9
8
k
2
decomposition products of glucose
(1)
k
i
) k
i,o
A
m
i
exp
-
E
i
RT
i ) 1, 2 (2)
1696 Ind. Eng. Chem. Res. 2007, 46, 1696-1708
10.1021/ie061186z CCC: $37.00 © 2007 American Chemical Society
Published on Web 02/13/2007
solution for 48 h at room temperature A kinetic model with an
inverse relation between the hydrolysis reaction rate constant
and the mean length of the crystalline domains of the cellulose
was proposed. Later investigations
21,22
have shown that the
Sharples model is not valid for virgin, untreated cellulose.
All previous kinetic studies mainly focused on the optimiza-
tion of glucose production. Only a few kinetic reports
23-25
are
available for the acid-catalyzed hydrolysis of cellulose to
levulinic acid. A complete kinetic model describing the acid-
catalyzed hydrolysis of cellulose to levulinic acid, including
byproduct formation and covering a broad range of reaction
conditions and intakes, is lacking. In addition, the acid-catalyzed
decomposition reactions of glucose and 5-hydroxymethylfurfural
(HMF) produce an insoluble-solid product known as humins.
These humins are expected to be formed as well when reacting
cellulose with acids in an aqueous environment. However,
humins formation has never been included in the kinetic models
reported to date. In this paper, we report a systematic kinetic
study on the acid-catalyzed hydrolysis of cellulose to levulinic
acid using sulfuric acid as the catalyst. The effects of temper-
ature, acid concentration, and initial intake of cellulose on the
yield of levulinic acid were assessed, and a kinetic model
including humin formation and covering a wide range of reaction
conditions was developed. The results were applied to optimize
the production of levulinic acid in various reactor configura-
tions.
2. Materials and Methods
2.1. Chemicals. All chemicals used in this study were of
analytical grade and used without purification. Microcrystalline
cellulose [9004-34-6] with an average particle size of 20 µm
was purchased from Sigma-Aldrich. The elemental composition
of the cellulose was determined by elemental analysis (C, 42.2%;
H, 6.1%). The carbon content is less than the theoretical value
for pure cellulose (C, 44.5%; H, 6.2%) because of the presence
of water. On the basis of the elemental composition, the cellulose
applied contains ∼4 wt % water. This was independently
confirmed by thermogravimetric analysis (TGA). Concentrated
sulfuric acid 95-97 wt % [7664-93-9], glucose [14431-43-7],
and formic acid [64-18-6] were purchased from Merck GmbH
(Darmstadt, Germany); 5-hydroxymethylfurfural [67-47-0] and
levulinic acid 98 wt % [123-76-2] were obtained from Acros
Organics (Geel, Belgium). Deionized water was applied to
prepare the various solutions.
2.2. Experimental Procedures. 2.2.1. Kinetic Experiments.
The reactions were carried out in two types of glass ampules
with a wall thickness of 1.5 mm and a length of 15 cm, differing
in internal diameter (3 and 5 mm). The ampules were filled
with the predetermined amount of cellulose. Subsequently, the
acid-catalyst solution (0.2-0.5 cm
3
) was added. The ampules
were sealed with a torch. The sealed ampules were placed in a
constant-temperature oven ((1 °C). At various reaction times,
ampules were taken from the oven and quenched in an ice-
water bath (4 °C) to stop the reaction. The ampule was opened,
and the liquid was separated from the solids using a microcen-
trifuge (Omnilab International BV) for ∼15-20 min at 1200
rpm. A certain amount of the clear solution was taken (100-
200 µL) and diluted with water (2 cm
3
). The composition of
the solution was determined using high-performance liquid
chromatography (HPLC).
The composition of the gas phase after the reaction was
determined using GC-MS. Gas samples were obtained by
placing an ampule in an airtight plastic bag. The plastic bag
was flushed with helium and placed under vacuum. Subse-
quently, the glass ampule was broken, and the released gas was
mixed with 10 cm
3
of helium gas.
The solid products were washed with water several times and
dried overnight in the oven at a temperature of 60 °C. The ele-
mental composition of the dried solid products was determined
using elemental analysis. The particle structure of the solid
products was analyzed using a scanning electron microscope
(SEM).
2.2.2. Heat-Transfer Experiments. At the start-up of the
reaction, the ampules were placed in a constant-temperature oven
and the contents were heated up to the predetermined oven
temperature. To determine the temperature profile at the start
of the reaction and to compensate for this nonisothermal
behavior in the kinetic modeling, the temperature inside the
ampule during the heating-up phase was determined experi-
mentally. For this purpose, a special ampule with a thermocouple
was developed. The ampule was filled with a representative
reaction mixture (without catalyst) and closed tightly using a
special bolt-and-screw system to prevent evaporation of the
liquid. The ampule was subsequently placed in the oven, and
the temperature of the reaction mixture was followed in time.
Before and after each experiment, the amount of liquid inside
the ampule was measured to determine the amount of evapora-
tion. In all cases, the loss of water was <1 wt %, indicating
that the results were not biased by water evaporation.
The experimental profiles at different temperatures were
modeled using a heat balance for the contents in an ampule:
Table 1. Literature Overview of Kinetic Parameters for the Acid-Catalyzed Hydrolysis of Cellulose
cellulose hydrolysis glucose decomposition
substrate k
1,o
(min
-1
) m
1
E
1
(kJ mol
-1
) k
2,o
(min
-1
) m
2
E
2
(kJ mol
-1
) reaction conditions
Douglas fir
13
1.73 × 10
19
1.34 179.5 2.38 × 10
14
1.02 137.5 T ) 170-190 °C
C
substrate
) 10 wt %
A ) 0.4-1.0 wt %
Kraft paper slurries
14
28 × 10
19
1.78 188.7 4.9 × 10
14
0.55 137.2 T ) 180-240 °C
C
substrate
) 2.5 wt %
A ) 0.2-1.0 wt %
Solca-Floc
15
1.22 × 10
19
1.16 177.8 3.79 × 10
14
0.69 136.8 T ) 180-240 °C
C
substrate
) 10 wt %
A ) 0.5-2.0 wt %
filter paper
16
1.22 × 10
19
1.16 178.9 3.79 × 10
14
0.69 137.2 T ) 200-240 °C
C
substrate
) 2wt%
A ) 0.4-1.5 wt %
MSW
18
1.16 × 10
19
1.0 171.7 4.13 × 10
15
0.67 142.4 T ) 200-240 °C
C
substrate
) 1wt%
acid ) 1.3-4.4 wt %
d(MC
p
T)
dt
) UA
t
(T
oven
- T) (3)
Ind. Eng. Chem. Res., Vol. 46, No. 6, 2007 1697
When assuming that the heat capacity of the reaction mixture
is constant and not a function of temperature, rearrangement of
eq 3 will give the following:
Solving the ordinary differential eq 4 with the initial value t )
0, T ) T
i
, leads to
The value of h was determined by fitting all experimental data
at different oven temperatures (100-160 °C) using a nonlinear
regression method and was found to be 0.596 min
-1
(for small
ampules) and 0.359 min
-1
(for large ampules). Figure 1 shows
experimental and modeled temperature profiles performed at
an oven temperature of 140 °C using both types of ampules.
Equation 5 was incorporated in the kinetic model to describe
the nonisothermal behavior of the system at the start-up of the
reaction.
2.3. Method of Analyses. The composition of the liquid phase
was determined using an HPLC system consisting of a Hewlett-
Packard 1050 pump, a Bio-Rad Organic Acid column Aminex
HPX-87H, and a Waters 410 refractive index detector. The
mobile phase consisted of an aqueous solution of sulfuric acid
(5 mM) at a flow rate of 0.55 cm
3
min
-1
. The column was
operated at 60 °C. The analysis for a sample was complete
within 55 min. The concentration of each compound in the liquid
phase was determined using calibration curves obtained by
analyzing standard solutions with known concentrations.
The gas composition was analyzed with GC-MS, which
consists of an HP 5890 Series II gas chromatograph and an HP
6890 detector. The composition of the gas phase was determined
using a CP-Porabond-Q column (length ) 25 m and i.d. ) 0.25
mm). The oven temperature was set at 40 °C for 2 min and
increased to 240 °C with an increment of 30 °C min
-1
. Helium
was used as the carrier gas with a flow rate of 1.5 mL min
-1
.
Elemental analyses were performed at the Analytical Department
of the University of Groningen using an automated Euro
EA3000 CHNS analyzer. Solid-products particles were analyzed
using a field emission scanning electron microscope (FESEM)
on a JEOL 6320F.
2.4. Determination of Kinetic Parameters. The kinetic
parameters were estimated using a maximum-likelihood ap-
proach, which is based on minimization of errors between the
experimental data and the kinetic model. Minimization of errors
was initiated by providing initial guesses for each kinetic
parameter. The best estimates were obtained using the MAT-
LAB toolbox fminsearch, which is based on the Nelder-Mead
optimization method.
The calculation of errors was based on the concentrations of
glucose (C
GLC
) and levulinic acid (C
LA
). To compensate for the
large spread in concentrations, the concentrations were scaled
and transformed to the yields of glucose (Y
GLC
) and levulinic
acid (Y
LA
), respectively. By definition, these vary between 0
and 1 and are expressed as
Here, the C
CEL,0
is defined as the initial concentration of
cellulose, expressed as the amount of glucose units present in
cellulose and determined using the following relation:
3. Results and Discussions
3.1. Reaction Products. The generally accepted reaction
pathway for the acid-catalyzed hydrolysis of cellulose to
levulinic acid is schematically given in Scheme 1.
In the first step, the polymer chains of cellulose (1) are broken
down into low molecular weight fragments and ultimately to
glucose (2) by the action of an acid catalyst. The glucose is
decomposed to 5-hydroxymethylfurfural (HMF, 3), which is
further converted in a serial mode to levulinic acid (4) and
formic acid (5). All anticipated products (2-5) were detected
in this study and identified and quantified using HPLC analysis.
A typical example of an HPLC chromatogram is given in Figure
2.
Besides the anticipated products, small amounts of glucose-
reversion products (e.g., levoglucosan, isomaltose, or gentio-
biose) and furfural were detected in the liquid phase. The
formation of the reversion products was also observed in our
previous study
26
on the acid-catalyzed decomposition of glucose.
The maximum amount of glucose-reversion products was very
low (<0.1 wt %). The presence of furfural in the reaction
mixture is, at first sight, surprising. It is a known product of
the acid-catalyzed decomposition of C5-sugars and particularly
of xylose (6), as shown in Scheme 2.
27-29
It is likely that the
cellulose applied in this study is contaminated with C5-sugars,
producing furfural (7). On the basis of the intake of cellulose
and the maximum experimentally observed concentration of
furfural, the amount of C5-sugars in the cellulose applied in
this study is ∼1wt%.
During all experiments, dark-brown insoluble substances
known as humins were formed. These are well-known products
of side-reactions of the acid-catalyzed decompositions of glucose
and HMF. The presence of these humins was confirmed by
elemental analysis on the solid products present after the
reaction. The elemental composition (in wt %) for a typical
product (C, 55.2; H, 4.9) suggests that the solids are a mixture
consisting mainly of humins (typical composition: C, 63.1; H,
4.2)
30
and some unreacted cellulose (C, 42.2; H, 6.1). Further
Figure 1. Heating profile of the reaction mixture at T
oven
) 140 °C for
both types of ampules.
Y
GLC
)
C
GLC
C
CEL,0
(6)
Y
LA
)
C
LA
C
CEL,0
(7)
C
CEL,0
)
mass of cellulose × wt % of glucose in cellulose
molecular weight of glucose × volume of reaction mixture
(8)
dT
dt
)
UA
t
MC
p
(T
oven
- T) ) h(T
oven
- T) (4)
T ) T
oven
- (T
oven
- T
i
) exp
-ht
(5)
1698 Ind. Eng. Chem. Res., Vol. 46, No. 6, 2007
evidence for the formation of substantial amounts of humins
was obtained from SEM. Typical SEM images of the cellulose
particles applied in this study and the solid products after the
reaction are given in Figure 3. Its shows round-shaped,
agglomerated particles with particle sizes in the range of 5-10
µm. On the basis of their visual appearance, the particles are
likely composed of insoluble humins.
30
Furthermore, some
unreacted cellulose appears to be present, in line with the
elemental analysis data.
Other possible byproducts of the acid-catalyzed hydrolysis
of cellulose are gas-phase components from thermal degradation
reactions of reactants and/or products. To gain insights into the
extent of these reactions, the gas phase after the reaction was
analyzed using GC and GC-MS. Both CO and CO
2
could be
detected; however, the amounts were <0.1 wt % of the cellulose
intake. This implies that the formation of gas-phase compounds
is only a minor reaction pathway under the reaction conditions
applied in our experiments.
The product levulinic acid (LA) is stable at the conditions
applied in this study and is not a source for byproducts (e.g.,
levoglucosan or R-angelica lactone). This was independently
checked by exposing a solution of levulinic acid (0.1 M) in a 1
M aqueous sulfuric acid solution for2hat200°C. After
reaction, only LA was detected in the solution.
3.2. Effects of Process Variables on the Yield of Levulinic
Acid. A total of 26 experiments were performed covering a wide
range of reaction conditions. Three operating temperatures (150,
175, and 200 °C) were used. In all cases, sulfuric acid was used
as the catalyst with concentrations varying between 0.05 and 1
M. The initial intake of cellulose (x
CEL,0
) was varied between
1.7 and 14 wt %.
The composition of the reaction mixture was followed in time
and a typical concentration profile is given in Figure 4. As
anticipated on the basis of Scheme 1, the concentrations of both
glucose and 5-hydroxymethylfurfural (HMF) displayed an
optimum. The maximum C
GLC
was 0.15 M, which equals a
glucose yield (Y
GLC
) of 30 mol %. The maximum C
HMF
obtained
in all experiments were generally much lower than the maximum
C
GLC
, which indicates that the conversion of HMF to levulinic
acid (LA) and formic acid (FA) is much faster than the
conversion of glucose to HMF. In line with the reaction
stoichiometry given in Scheme 1, LA and FA were always
formed in a 1:1 molar ratio. This finding also implies that these
compounds are stable under the reaction conditions employed
and do not decompose to other products.
The yield of levulinic acid (Y
LA
) is a clear function of the
operating temperature, with high temperatures leading to reduced
yields. This is illustrated in Figure 5, where the yields of
levulinic acid are plotted as a function of reaction time at three
different temperatures.
The yield of levulinic acid was improved when applying
higher acid concentrations; see Figure 6 for details. This effect
was substantial, and the yields increased from 31 to 54 mol %
when increasing the acid concentration from 0.1 to 1 M.
A number of experiments were carried out using various
initial intakes of cellulose (1.7-14 wt %) at T ) 150 °C and a
catalyst concentration of 1 M. The initial intake of cellulose
has a significant effect on the yields of levulinic acid (Figure
7). Lower intakes of cellulose resulted in higher yields of
levulinic acid. These findings are in line with previous studies
on the acid-catalyzed decompositions of HMF and glucose,
Scheme 1. Acid-Catalyzed Hydrolysis Reaction of Cellulose to Levulinic Acid
Figure 2. Typical HPLC chromatogram for the acid-catalyzed hydrolysis
of cellulose (x
CEL,0
) 7.7 wt %, C
H
2
SO
4
) 0.05 M, T ) 200 °C, t ) 16
min).
Scheme 2. Acid-Catalyzed Decomposition of Xylose to
Furfural
Ind. Eng. Chem. Res., Vol. 46, No. 6, 2007 1699
