Kinetic analysis and process modeling for cellulose valorization in cooperative ionic liquid pairs

Article abstract of DOI:10.1016/j.cattod.2015.07.039

Studies on kinetic analysis and process modeling are important and necessary for the development of novel and efficient technology for cellulose utilization. In this study, a kinetic model for the batch reactor was described for catalytic cellulose degradation in cooperative ionic liquid pairs based on an intensive analysis of the experimental parameters. Our fitting results show that the proposed model agrees well with the experimental data. The kinetic parameters obtained from the above model were subsequently applied to the continuous stirred tank reactor (CSTR) by constructing a mathematical process model for the continuous operation of cellulose degradation combined with product separation. The effects of reaction temperature, retention time, extract flow rate and distribution coefficients were intensively investigated. The numerical results demonstrate that relatively high temperature and larger retention time are favorable in cellulose conversion, whereas the product distribution coefficient and extract flow rate mainly affect the concentration of the hexane-soluble fraction (volatile chemicals). The findings presented in this work will serve as a beneficial reference for further biomass transformation of industrial purpose.

Full text of DOI:10.1016/j.cattod.2015.07.039

G Model
CATTOD-9714; No. of Pages8
ARTICLE IN PRESS
Catalysis Today xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Kinetic analysis and process modeling for cellulose valorization in
cooperative ionic liquid pairs
Sijie Liu, Luo Tang, Jinxing Long, Jianyu Guan∗, Xuehui Li∗
School of Chemistry and Chemical Engineering, Pulp & Paper Engineering State Key Laboratory of China, South China University of Technology, Guangzhou
510640, PR China
a r t i c l e i n f o
a b s t r a c t
Article history:
Studies on kinetic analysis and process modeling are important and necessary for the development of
novel and efficient technology for cellulose utilization. In this study, a kinetic model for the batch reactor
was described for catalytic cellulose degradation in cooperative ionic liquid pairs based on an intensive
analysis of the experimental parameters. Our fitting results show that the proposed model agrees well
with the experimental data. The kinetic parameters obtained from the above model were subsequently
applied to the continuous stirred tank reactor (CSTR) by constructing a mathematical process model
for the continuous operation of cellulose degradation combined with product separation. The effects
of reaction temperature, retention time, extract flow rate and distribution coefficients were intensively
investigated. The numerical results demonstrate that relatively high temperature and larger retention
time are favorable in cellulose conversion, whereas the product distribution coefficient and extract flow
rate mainly affect the concentration of the hexane-soluble fraction (volatile chemicals). The findings
presented in this work will serve as a beneficial reference for further biomass transformation of industrial
purpose.
Received 28 May 2015
Received in revised form 21 July 2015
Accepted 23 July 2015
Available online xxx
Keywords:
Reaction kinetics
Process modeling
Ionic liquids
Catalytic degradation
Cellulose
© 2015 Elsevier B.V. All rights reserved.
1. Introduction
been extensively explored for promoting the liquefaction degree,
product yield and selectivity, and reducing the reaction tempera-
Conversion of biomass substance to bio-fuel or chemicals is
being actively pursued because of its great potential for reliev-
ing the environmental pollution and energy crisis [1]. However,
typical transformation methods for this process are generally cost-
ineffective due to the complex and highly recalcitrant property
of the biomass composition caused by the robust 3D hydrogen
bond network, so the current technologies are unable to meet the
requirement of large scale commercialization [2,3]. Therefore, effi-
cient biomass valorization technology at mild conditions is crucial
and highly desirable. Catalytic liquefaction is regarded as a use-
ful technique for biomass valorization because of its comparatively
mild reaction condition (low temperature and pressure) and con-
centrated production distribution [4]. In general, both chemical
and biochemical catalytic processes are considered to be efficient
strategies for biomass liquefaction [5–8]. However, the liquefaction
efficiencies are still improvable because the lignocellulosic biomass
is normally non-dissolvable in both water and traditional organic
solvents. Various process intensification techniques thereby have
ture and time [9–11].
In 2002, Rogers and coworkers found that cellulose can be
directly dissolved in alkyl-imidazolium ionic liquids (ILs) under a
moderate condition with 10 wt% solubility [12], which opens up
a new perspective for the homogeneous liquefaction of biomass
and its components. ILs as a novel solvent or efficient catalyst
for biomass disposition has thereby received increasing attention
[13–16] and it is regarded as a useful solvent to achieve biomass
valorization. For example, the efficient hydrolysis of cellulose was
reported with more than 77% yield of reducing sugar over the
ILs [bmim]Cl and proton acid catalyst [17]. The conversion of cel-
lulose and real lignocellulosic biomass to fine chemicals, such
as 5-hydroxymethylfurfural (HMF) and levulinic acid, were also
extensively examined [18–20]. However, in most of the above-
mentioned processes, cellulose should be pre-dissolved in the ILs.
Therefore, the liquefaction efficiency in these processes is still an
issue due to the limited solubility of biomass in ILs. In order to
solve this problem, we previously proposed a dissolution and in-situ
degradation mode for cellulose and lignocellulosic biomass lique-
faction at mild conditions using cooperative IL pairs [21–23]. In
a typical approach [21], the cellulose was first partially dissolved
by the IL solvent to give a homogeneous medium, and then this
∗
Corresponding authors. Tel.: +86 20 87114707.
E-mail addresses: cejyguan@scut.edu.cn (J. Guan), cexhli@scut.edu.cn (X. Li).
http://dx.doi.org/10.1016/j.cattod.2015.07.039
0920-5861/© 2015 Elsevier B.V. All rights reserved.
Please cite this article in press as: S. Liu, et al., Kinetic analysis and process modeling for cellulose valorization in cooperative ionic liquid
pairs, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.07.039

G Model
CATTOD-9714; No. of Pages8
ARTICLE IN PRESS
2
S. Liu et al. / Catalysis Today xxx (2015) xxx–xxx
and in situ degradation method for the transformation of cellulose
Nomenclature
and lignocellulosic biomass was originally focused on the biomass
valorization in batch reactor. Its effectiveness and efficiency in com-
prehensive utilization of energy, convenient separation of products
andmassproductionofbiomassneedto beexaminedandevaluated
in the industrial scope.
In this work, we proposed a reaction kinetic model for the cellu-
lose catalytic liquefaction based on the cooperative IL pairs system.
The kinetic property of the cellulose degradation in a batch reac-
tor was first investigated and analyzed. And then, based on the
parameters obtained from the batch reactor, an intensive process
simulation of the continuous process for cellulose depolymeriza-
tion plus the product separation in an integrated reactor was put
forward to provide an insight into the industrial valorization of
lignocellulosic biomass and its components.
c0
cc
initial cellulose concentration in reactor (g/L)
catalyst concentration (g/L)
cf
fast-reaction part concentration (g/L)
hexane-soluble products concentration (g/L)
methanol-soluble products concentration (g/L)
methanol-soluble products concentration in extract
(g/L)
ch
cm
cmE
chE
hexane-soluble products concentration in extract
(g/L)
crf
crs
cs
k
active concentration of fast-reaction part (g/L)
active concentration of slow-reaction part (g/L)
slow-reaction part concentration (g/L)
rate constant
kꢀ
reaction constant correlated with catalyst concen-
2. Experimental
tration (s−1
)
m0
ms0
R1
initially total cellulose mass (g)
initial slow-reaction part mass (g)
distribution coefficient of methanol-soluble prod-
ucts
distribution coefficient of hexane-soluble products
reaction rate (g/sL)
cellulose degradation rate (g/sL)
solubility of cellulose in ILs (g/100 g ILs)
temperature (K)
Cellulose degradation is carried out in a 150 mL stainless
autoclave as
a batch reactor. Typically, 5 g cellulose, desig-
nated amount of cooperative ILs (20 g bmimCl and 3.5 mmol
C4H8SO3HmimHSO4), 10 mL methanol and 80 mL hexane are added
to the reactor. After displacing the air in the reactor with nitro-
gen, the reactor is heated and kept at designated temperature for
15 min, and then the mixture of IL and products are cooled to room
temperature for separation.
R2
r
rc
S
T
The product from the cellulose degradation forms two phases
in a separating funnel, and is first separated. The heavy component
fraction is then exacted by hexane for three times to remove the
residual volatile chemical in the IL. By collecting the extract and
the upper phase product, the hexane-soluble fraction is obtained.
After a sequence of IL removal operations (extracted by CH2Cl2) and
dilution with methanol, the nonvolatile product from the cellulose
dissolution and degradation process is obtained as the methanol-
soluble fraction. The volatile components are determined with a
gas chromatography mass spectrometry (GC-MS). The molecular
weight distribution of the methanol-soluble fraction is measured
with a gel permeation chromatography (GPC) and the particle size
of this fraction was measured with the Malvern analysis [21].
t
time (s)
tR
V
v
˚
ꢀ
average retention time of feedstock (s)
volume (L)
mass flow rate (g/s)
ratio of slow-reaction part
density (g/L)
Subscripts
f
IL
m
h
s
fast-reaction
ionic liquid
methanol-soluble products
hexane-soluble products
slow-reaction
1
methanol-soluble products conversion to hexane-
soluble products
3. Modeling
−1
hexane-soluble products conversion to methanol-
soluble products
3.1. Kinetic model in batch reactor
The dissolution and in situ catalytic degradation of cellulose by
cooperative IL pairs are a complex process [21], in which part of the
feed cellulose is dissolved by the IL solvent to yield a homogeneous
solution at first, and then it can sufficiently contact with another
acidic IL catalyst to form an efficient process. At the same time,
the depolymerization of cellulose in the first step can be divided
into two fractions: fast-reaction and slow-reaction parts to obtain
the methanol-soluble products. Then, the second step about further
degradation of the methanol-soluble products, as a homogeneous
reaction, can be simplified as a reversible first order process. This
two-step reaction model in a batch reactor is described as Eq. (1):
IL/cellulose homogenous medium was catalytically degraded to
useful chemical while a coordinated acidic IL was used as catalyst
simultaneously. With the synergic effect of the cooperative IL pairs
(one IL serving as the solvent for cellulose dissolution and the other
as the catalyst for the degradation), cellulose can be completely
converted to useful industrial chemical in a single batch reaction
at mild conditions (473 K for 15 min). The dragging of the disso-
lution equilibrium, combined with the rapid, in situ acid-catalyzed
degradation of bulk insoluble cellulose, overcame the long intrinsic
problem of cellulose insolubility encountered in the conversion of
biomass to biochemical. However, as the most prominent problems
in current biomass conversion processes, complex combination
and optimization of process factors, such as product separation,
comprehensive energy utilization, solvent recycling and realiza-
tion of continuous operation are far from being resolved [3] and
thus should be comprehensively understood and intensively inves-
tigated. In contrary to catalysis chemistry studies, kinetic modeling
and whole process simulation are able to provide deep under-
standing about the optimal design of the process. Such efforts
are currently still in great need [24]. Our previous dissolution
(1)
where MP and HP represent the methanol-soluble and hexane-
soluble products, respectively. The rate constants kf and ks are for
the fast reaction and slow reaction, whereas k1 and k denote the
−1
Please cite this article in press as: S. Liu, et al., Kinetic analysis and process modeling for cellulose valorization in cooperative ionic liquid
pairs, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.07.039

G Model
CATTOD-9714; No. of Pages8
ARTICLE IN PRESS
S. Liu et al. / Catalysis Today xxx (2015) xxx–xxx
3
Fig. 1. Molecule structure of cellulose and its degradation products.
rate constants for the degradation of the methanol-soluble prod-
ucts and its reversible process.
process of biomass degradation [21–23]. Therefore, the same cor-
relation of cellulose crystallinity and temperature is adopted to
describe the ratio of the slow-reaction ˚, and the interrelation
between the crystallinity and the reaction temperature T can be
expressed in Eq. (6), according to the experimental data obtained
under temperature of 460 K [26]. At high temperature (higher than
460 K), the value of ˚ is set to be equal to zero.
During the degradation, the cellulose is first depolymerized to a
nonvolatile oligomer (the methanol soluble fraction), and then this
oligomer is further degraded to the final products, which may be
extracted by hexane. Fig. 1 describes the molecule structure of cel-
lulose and the representative degradation products. The methanol
soluble fraction has a molecular weight of 788–1025 [21], which is
considered to be oligomer of the products and the part degradation
cellulose. The Malvern analysis demonstrates their size distribution
within 20–300 nm. The lighter fractions that dissolve in the hexane
are small molecule compounds such as levulinic acid, furfural, butyl
acetate and 2-(diethoxymethyl)furan. All of them have a high sol-
ubility in hexane. Cellulose and all of the degradation products are
dissolvable in the IL, but due to extracting by hexane and methanol,
only a trace of products existed in the IL, therefore the products in
the IL may be neglected in this batch reactor model.
ms0
˚ =
= −0.0153T + 7.0437,
(6)
m0
where ms0 is the initial slow-reaction part mass, and m0 is the
initially total cellulose mass.
Because the detailed quantitative effect of acidic IL catalyst on
this process is difficult to obtain, the influence of catalyst is assumed
to be contained in the rate constants. Furthermore, taken into
consideration of the feature of heterogeneous reactions, the con-
centration of cellulose should be adjusted in the first step of the
reactions. In fact, cellulose particles are dissolved in the IL pairs
prior to reacting, so the concept of an active reaction concentration
that relates to the solubility is adopted. The rate equations for the
batch reactor can thereby be modified and listed in Eqs. (7)–(10):
According to this model, the kinetics of cellulose degradation
can be proposed by the following power-law equations including
the reaction constant, reactant concentration and catalyst concen-
tration (Eqs. (2)–(5)).
dcf
dcf
rf
=
= −kꢀ crf
(7)
(8)
f
rf
=
= −kf cc˛cf
(2)
(3)
(4)
(5)
dt
dt
dcs
dt
dcs
dt
rs =
= −ksꢀ crs
rs =
= −kscˇcs
c
dcm
dt
dcm
dt
r1
r
=
= kꢀ crf + ksꢀ crs − k1ꢀ cm + k−ꢀ 1ch
(9)
f
r1
r
=
= k c˛c + kscˇcs − k1cꢁ cm + k−1cϕc
f
f
h
c
c
c
c
dch
dt
= k1ꢀ cm − k−ꢀ 1ch
(10)
dch
dt
=
=
= k1cꢁ cm − k−1cϕc
−1
−1
h
c
c
where kꢀ represents the reaction constant correlated with the cata-
lyst concentration, crf is the active concentration in the fast reaction
and crs is the active concentration in the slow reaction. The value
of crf and crs may be calculated as follows:
where cc, cf, cs, cm and ch stand for the concentration of catalyst, cel-
lulose in fast reaction, cellulose in slow reaction, methanol-soluble
products and hexane-soluble products, respectively, r is the reac-
tion rate, and ˛, ˇ, ꢁ and ϕ are the corresponding exponent of the
catalyst concentrations.
Generally speaking, the content of the slow-reaction cellulose
is mainly dependent on the source of cellulose and the tempera-
ture [25]. Recent studies also shows that the crystallinity, which
usually directly relates to the contents of the slow-reaction and
fast-reaction cellulose, positively depends on the temperature [26],
which is also accords well with our previous studies about the
ꢀ
ꢁ
(1 − ˚)SꢀIL
crf = min cf ,
(11)
(12)
100
ꢀ
ꢁ
˚SꢀIL
crs = min cs,
100
where S is the solubility of cellulose in the IL pairs, and ꢀIL is the
density of IL pairs.
Please cite this article in press as: S. Liu, et al., Kinetic analysis and process modeling for cellulose valorization in cooperative ionic liquid
pairs, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.07.039

G Model
CATTOD-9714; No. of Pages8
ARTICLE IN PRESS
S. Liu et al. / Catalysis Today xxx (2015) xxx–xxx
˚SꢀIL
4
crs
=
(14)
100
In above expressions, only ˚ changes with temperature while
the concentration of active cellulose is constant at a certain temper-
ature. Because the reactant concentration does not change in this
continuous process, the kinetic expression under the steady-state
operation should be modified from Eq. (9) and (10) and is given as
Eqs. (15)–(16).
v4 cm
ꢀ4 VIL
kꢀ crf + ksꢀ crs − k1ꢀ cm + k−ꢀ 1ch − R1
= 0
(15)
(16)
f
v4 ch
k1ꢀ cm − k−ꢀ 1ch − R2
= 0
ꢀ4 VIL
VIL is the volume of IL pairs in the reactor, ꢀi is the density of each
flow i (i = 1–8). R1 is the distribution coefficient of methanol-soluble
products and R2 is the distribution coefficient of hexane-soluble
products, expressed by
Fig. 2. Diagram of the combined process of cellulose degradation and product sep-
aration.
cmE
3.2. Continuous model
R1
R2
=
(17)
(18)
cm
Coupling reaction with a separation unit in an integrated reac-
tor has the potential for enhancing the conversion and selectivity of
an equilibrium reaction, and reducing the cost of equipment [27],
as is always being used in the industry. Hence, a continuous model
of cellulose degradation and product separation in an integrated
reactor is proposed based on the above reaction mechanism and
process analysis. Fig. 2 shows the schematic flowsheet of this pro-
cess. The cellulose particles at a rate of v1 are well mixed with the
reflux cooperative IL pairs to form IL/cellulose slurry. The slurry is
pumped into the integrated continuous stirred tank reactor (CSTR)
at a rate of v2 to conduct the depolymerization reaction. During this
process, the mixture of the lower phase containing solid cellulose,
products, by-products generated by the repolymerization reaction
and IL pairs will be delivered to a settler at a rate of v3 to separate
solid and liquid and to recycle the ILs and other liquid contents.
In the reactor, the extract phase (upper phase), which contains the
value-added volatile products, is transported to a distillation col-
umn to separate the extractant and products at a rate of v4. The
parameter v5 is the flow rate of recycled pure extractant, v6 is the
rate of recycled ILs, v7 is the unreactive cellulose rate, and v8 is
the rate of pure products. Parameters vm and vh are rates of pure
methanol-soluble and hexane-soluble products transferring from
the IL phase to the extract phase.
chE
ch
=
where cmE and chE are the methanol-soluble and hexane-soluble
products concentration in the extract phase.
The mass balance in the whole system, mixer, reactor, settler
and distillation column yields the relations
v1 = v8 + v7
v1 + v6 = v2
v2 + v5 = v4 + v3
v6 + v7 = v3
v4 = v5 + v8
(19)
(20)
(21)
(22)
(23)
Apart from the mass balance equations mentioned above, there
are some equations relating to the concentration and flow rate
by considering the mass conservation of each species. Following
equations show the flow rate of pure methanol-soluble and hexane-
soluble products transferring to the extract phase.
v4
ꢀ4
vm
vh
=
cmE
(24)
(25)
Just as in most biomass-based processes, the lack of correlative
experimental data and precise physical and chemical properties is
the main obstacles to the proper model exploration. Here are some
hypotheses that we propose to simplify the simulation, based on
the experimental data and theoretic analysis.
v4
ꢀ4
=
chE
By comparing the Eqs. (15)–(20) with Eqs. (24)–(25), we can
get Eq. (26) and deduce the following density relation by assum-
ing those three liquids being idea liquid, Eq. (27). The others flow
density can also be calculate using this method.
1. No cellulose particles exist in the extract.
2. When the extract flow rate v4 is ten times higher than cellulose
feeding rate, the product distribution coefficient R1 and R2 equal
0.5 and 1.0, respectively. And the rate of mass transfer is not
taken into account.
3. The unreacted cellulose settles in the settler entirely.
4. The operation is under steady-state condition.
5. The reactive cellulose concentration in the CSTR remains con-
stant and it equals the saturated concentration.
v8 = vm + vh
(26)
(27)
1
ꢀ8
vm
vh
=
+
v8ꢀm v8ꢀh
There is a characteristic equation to describe the relation
between the volume of IL pairs in the reactor and the rate of feed-
ing, Eq. (28). And the conversion of cellulose can be described as
Eqs. (29)–(30).
v2
ꢀ2
In this system, steady-state operation makes sure that the cellu-
lose concentration does not change with time. Then, Eqs. (13)–(14)
give the two parts of cellulose concentration in the CSTR under
steady-state condition.
VIL
VIL
=
=
tR
(28)
(29)
(30)
c0v2
(−rc)ꢀ2
v8
v1
X
(1 − ˚)SꢀIL
X =
crf
=
(13)
100
Please cite this article in press as: S. Liu, et al., Kinetic analysis and process modeling for cellulose valorization in cooperative ionic liquid
pairs, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.07.039

G Model
CATTOD-9714; No. of Pages8
ARTICLE IN PRESS
S. Liu et al. / Catalysis Today xxx (2015) xxx–xxx
5
Table 1
Determined rate constants of cellulose degradation in cooperative ILs at various
temperatures.
Temperature (K)
Rate constant (s−1
)
kfꢀ
ksꢀ
k1ꢀ
k−ꢀ 1
433
453
473
2.88 × 10−3
3.22 × 10−3
1.15 × 10−2
2.96 × 10−6
1.09 × 10−5
1.30 × 10−5
2.93 × 10−2
1.72 × 10−1
1.73 × 10−1
8.84 × 10−3
4.06 × 10−2
3.93 × 10−2
where tR is the average retention time of feedstock, and X and rc
are the conversion and degradation rate of cellulose. Eqs. (15)–(30)
have a general description of the continuous operation in the
integrated CSTR, and those relative equations were solved by
programming with the software package Matlab to get a clear
recognition of the continuous process.
Fig. 3. Concentration–time profiles of cellulose to methanol-soluble and hexane-
soluble products at different temperatures in the batch reactor.
4. Results and discussion
4.1. Kinetic study in batch reactor
cellulose depolymerization. As shown in Fig. 4, the fast-reaction
cellulose fragment suffers a sharp depolymerisation at each case,
whereas, the concentration of the slow-reaction fraction has no
significant change. Therefore, the high content of the refractory
slow-reaction cellulose at 433 K (measured according to Eq. (8)) is
considered to be responsible for its low conversion. And the high
concentration of slow-reaction fragment significantly hinders its
efficient utilization, which is also a main reason for the fact that
many investigations focus on the cellulose pretreatment [2].
4.1.1. Kinetic parameters estimation in batch reactor
The kinetic constants were estimated by the experimental data
exclusively considering the final reaction chemical equilibrium
using the optimizing program in Matlab. The obtained constants
at different temperatures are summarized in Table 1 when the acid
IL C4H8SO3HmimHSO4 is used as the cellulose degradation cata-
lyst. As shown in Table 1, the value of kf is far higher than that of
ks (more than 103 times) in each case, which is similar to the ratio
that recently reported [25], indicating that the content of the fast-
reaction cellulose is crucial in this cooperative IL pairs system. This
result also explains well with our previous studies, where amor-
phous cellulose and natural cellulose in lignocellulosic biomass
are more flexible than the microcrystal cellulose for degradation
[21,22], because of lower crystallinity and more fast-reaction con-
tent there. Table 1 demonstrates that the reaction constants of the
cellulose degradation to intermediate (methanol-soluble fraction)
are significantly influenced by the reaction temperature. And an
obvious increase is shown when the reaction temperature is ele-
vated from 453 to 473 K, which suggests the sharp disappearance
of the residual solid (unreacted feed) in the cooperative IL pairs
at 473 K [21] where the degradation of cellulose in the presence of
cooperative IL pairs is a continuous process of cellulose depolymer-
ization to methanol soluble fraction and then the degradation of
the intermediate (methanol soluble fraction) to the hexane soluble
chemical (the final products). Table 1 also shows that the reac-
tion rate constant of cellulose depolymerization (kf) is far less than
that of hexane-soluble products (k1) under various reaction tem-
peratures during this two-step continuous process. It also implies
that the heterogenetic reaction of cellulose particles to methanol-
soluble compounds by cooperative IL pairs is the rate-controlling
step as analyzed before [21–23].
4.1.2. Catalyst effect
Fig. 5 shows the variation of cellulose and product concentra-
tion in the batch reactor with time under three different catalysis
conditions (C4H8SO3HmimHSO4, H2SO4 and no catalyst) base on
the optimized kinetic parameters obtained from Matlab. As shown
in Fig. 5, the conversion rate of cellulose without any catalyst is far
away from the catalytic system and with less cellulose conversion
and product formation. The acid IL C4H8SO3HmimHSO4 is the most
efficient catalyst in this system for cellulose degradation, and the
cellulose reach a complete degradation at 500 s. However, when
H2SO4 is used as catalyst, just 72.4% of the cellulose degradation
is attained at the time. Considering that the acid IL has higher cat-
alytic effect than mineral acid and less corrosion of the equipment,
C4H8SO3HmimHSO4 is adopt for the modeling of cellulose catalytic
degradation in the integrated CSTR.
The changes of cellulose and product concentration under
various reaction temperatures are also simulated using Matlab.
The initial cellulose concentration (c0) is designated to be 200 g/L
according to the feed cellulose ratio in the cooperative IL pairs
reported in [21]. The solubility of cellulose in ILs is set to 12 g/100 g
[3]. As shown in Fig. 3, the reaction temperature has a significant
effect on the cellulose degradation. For example, nearly 50% of
the cellulose is converted when it is treated at 433 K for 1500 s.
However, when the reaction temperature is elevated to 473 K,
100% cellulose can be converted with more than 80% of the
hexane-soluble fraction selectivity during 500 s. These results are
fitted well with our previous experimental results [21], where
cellulose can be converted 100% in 15 min. Fig. 4 further demon-
strates that the species of the cellulose are also a key factor for
Fig. 4. Concentration–time profiles of the fast-reaction and slow-reaction cellulose
in the batch reactor at various temperatures.
Please cite this article in press as: S. Liu, et al., Kinetic analysis and process modeling for cellulose valorization in cooperative ionic liquid
pairs, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.07.039

G Model
CATTOD-9714; No. of Pages8
ARTICLE IN PRESS
6
S. Liu et al. / Catalysis Today xxx (2015) xxx–xxx
Fig. 5. Concentration–time profiles of cellulose to methanol-soluble and hexane-
soluble products by different catalyst in the batch reactor.
Fig. 6. Product concentration and flow rate profile of the methanol-soluble and
hexane-soluble products at various retention time.
Table 2
Simulation parameters for integrated CSTR model.
No.
Parameter
Value
Ref.
1
2
3
4
5
6
7
8
ꢀIL
ꢀ1
ꢀ5
ꢀ7
ꢀm
ꢀh
V1
VIL
1000
1500
690
1500
1050
850
[28]
[29]
[30]
[29]
–
–
–
–
condition. The continuous simulation result shows that cellulose
conversion and the product distribution are significantly affected
by the thermal effect (Table 3). For example, cellulose conver-
sion increases from 14.1% to 24.0% when the reaction temperature
increases from 433 to 453 K, further increasing the temperature to
473 K will enhance the liquefaction of cellulose sharply to 96.4%
conversion. It implies that the relationship between the conver-
sion of cellulose and the reaction temperature is exponential to
some extent. In this process, reaction temperature plays an impor-
tant role in the conversion of cellulose, and there are some key
factors that will push the transformation process forward inten-
sively at relatively higher temperature, for example, the dissolution
cellulose in ILs, the nonvolatile products in methanol and volatile
products in hexane, and the endotherm of depolymerization reac-
tion will also take effect on this conversion. Except for the higher
reaction rate constant at high temperature, this phenomenon is also
greatly caused by the higher content of the fast-reaction cellulose
at elevated temperature, because this fragment is more ready to
be converted (Table 1 and Fig. 4). Moreover, with the increase of
the temperature, the ratio of hexane-soluble to methanol-soluble
products is also increased, which means that the more hexane-
soluble products will be yielded at a high temperature. This well
fits with the simulated results of the kinetic parameters in Table 1
that the value of k1 has a more obvious increasing than that of
100
70
4.2. Integrated CSTR studies
Based on the results from the batch reactor model (Table 1,
Figs. 3 and 4), the simulation of the integrated CSTR was carried out.
It should be noted that this simulation is more complex than that
of the batch reaction. Therefore, in this study, only some important
parameters such as reaction temperature, retention time, extract
flow rate and distribution coefficient are considered and taken
as variables to verify their effect on the cellulose conversion and
product distribution in the integrated CSTR, while other param-
eters such as solubility, density of ILs, extractant, cellulose and
the product are assumed as constant. Table 2 gives the simulation
parameters used in the model.
4.2.1. Effect of temperature
Table 3 shows the effect of the reaction temperature on the cel-
lulose degradation in the integrated CSTR. In this section, the feed
rate of cellulose (v1, g/s) and the volume of ILs (VIL, L) are desig-
nated as constants (Table 2). However, the value of v6 should be
less than 100 v1/S as the cellulose in ILs should maintain its satura-
tion concentration in the CSTR, where S (g) represents the solubility
of cellulose in 100 g IL, therefore, v6 with the value of 100 v1/1.2S
is designed with the assumption that 20% of excessive cellulose
is added. According to our hypotheses 2, the flow rate of reaction
mixture sent to the distillation column (v4) is designated ten times
that of the feed cellulose (v1) in order to maintain a steady-state
k
−1. Then, the following analysis is based on the simulated result
achieved at 473 K because the small molecular weight product in
the hexane-soluble part is preferable.
4.2.2. Effect of retention time
In this model, it is assumed that the feeding rate maintains
constant and reactive cellulose concentration is saturate in the ILs
when this process is under the steady state. The cellulose conver-
sion is linearly related to the average retention time according to
Eqs. (28)–(29). In this section, the feeding rate v1 is set to 200 g/L,
due to the high reaction rate under 473 K. It can be calculated that
the cellulose conversion increases from 42.1% to 84.4% linearly,
when the retention time changes from 40 to 80 s. Fig. 6 shows the
change of product concentration and flow rate with the retention
time. It can be seen from Fig. 6 that the four parameters are
increased linearly with the retention time, and among these,
the flow rate of hexane-soluble product increases fastest. Those
calculated results illustrate that increasing the retention time
has a positive effect on the product yielding. On the other hand,
longer retention time means bigger reactor volume, resulting in
Table 3
Product concentrations and cellulose conversion at various temperaturesa.
Temperature (K)
Productconcentration(g/L)
cm ch
6.5 6.5
X (wt%)
433
453
473
14.1
24.0
96.4
5.0
14.1
57.9
19.9
a
v
1 = 100 g/s, VIL = 70 L.
Please cite this article in press as: S. Liu, et al., Kinetic analysis and process modeling for cellulose valorization in cooperative ionic liquid
pairs, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.07.039

G Model
CATTOD-9714; No. of Pages8
ARTICLE IN PRESS
S. Liu et al. / Catalysis Today xxx (2015) xxx–xxx
7
rises from 0.3 to 0.7, the mass flow rate of hexane-soluble products
vh decreases from 87.4 to 77.7 g/s. Hence, the smaller value of R1 is
desirable since the total mass rate of products is not changed. The
mass flow rate of hexane-soluble products vh increases from 75.0 to
82.2 g/s when the value of R2 increases from 0.5 to 1.0. This increase
is important from the view point of industry as much high concen-
tration of the final products in the reaction phase goes against the
depolymerisation process and repolymerization will always take
place between the products. Therefore, it is important to use an
extractant that has a high value of R2 and a much small R1.
5. Conclusion
In this work, a kinetic model was firstly proposed to describe
the catalytic degradation of cellulose over cooperative ionic liq-
uids in a batch reactor in view of the experimental data and the
final reaction chemical equilibrium. Then, a mathematical model
including simultaneous cellulose catalytic degradation and prod-
uct separation has been built for the integrated CSTR based on the
kinetic model in the batch reactor. The simulation result shows
that reaction temperature, retention time, extract flow rate, and
the distribution coefficient are all crucial for cellulose degradation.
The first two factors mainly affect the cellulose conversion, in which
relatively high temperature and larger retention time are favorable
in the cellulose conversion. The extract flow rate and distribution
coefficient have a more significant effect on the product concentra-
tion, with a higher hexane-soluble product distribution coefficient
and lower extract flow rate resulting in higher concentrations of
the products. These simulation results should shed light on better
understanding the catalytic conversion of cellulose in ionic liquid
pairs and forecasting the product distribution in a continuous oper-
ation for the industrial transformation of biomass.
Fig. 7. Product concentration and flow rate profile of the methanol-soluble and
hexane-soluble products at various extract flow rates.
increasing of equipment cost. Therefore, the appropriate value of
retention time should be determined through optimization.
4.2.3. Effects of extract flow rate
Fig. 7 shows the effect of the extract flow rate v4 on the prod-
uct distribution, when the parameters of VIL, T, and v1 are set to
70 L, 473 K and 100 g/s, respectively. As shown in Fig. 7, with the
increase of the extract flow rate (v4), the flow rate of hexane-
soluble products (vh) has a small decrease and methanol-soluble
products (vm) has a corresponding increase. However, the concen-
trations of products both in methanol and hexane (cm, ch) decrease
sharply. Especially, when the extract flow rate increases from 500
to 1500 g/s, the hexane-soluble product decreases sharply from
121.3 to 27.4 g/L. Therefore, we can safely conclude that the hexane-
soluble products are more sensitive than that of methanol-soluble
fraction. Although the operation purpose is to get more hexane-
soluble product, the energy consumption should be taken into
account. Much lower concentration of products in the extract phase
will undoubtedly result in an increased cost for the separation of
products and extractant. It illustrates that the appropriate value
of extract flow needs more calculation by comprehensive consid-
eration of the product yield and energy consumption before it is
decided.
Acknowledgments
The authors gratefully acknowledge financial support of the Nat-
ural Science Foundation of China (Grants 21336002, 21276094, and
51306191), the Doctoral Fund of the Ministry of Education of China
(Grant 20130172110043), and the Natural Science Foundation of
Guangdong Province, China.
References
[1] M.E. Himmel, S.-Y. Ding, D.K. Johnson, W.S. Adney, M.R. Nimlos, J.W. Brady,
T.D. Foust, Science 315 (2007) 804–807.
[2] Q. Zhang, F. Jerome, ChemSusChem 6 (2013) 2042–2044.
[3] A. Brandt, J. Gräsvik, J.P. Hallett, T. Welton, Green Chem. 15 (2013) 550–583.
[4] J. Long, B. Guo, J. Teng, Y. Yu, L. Wang, X. Li, Bioresour. Technol. 102 (2011)
10114–10123.
[5] P. Biller, A.B. Ross, Bioresour. Technol. 102 (2011) 215–225.
[6] R. Harun, M.K. Danquah, G.M. Forde, J. Chem. Technol. Biotechnol. 85 (2009)
199–203.
[7] A.F. Lee, K. Wilson, Catal. Today 242 (2015) 3–18.
[8] N. Ji, T. Zhang, M. Zheng, A. Wang, H. Wang, X. Wang, J.G. Chen, Angew. Chem.
Int. Ed. 120 (2008) 8638–8641.
[9] Q. Gan, S.J. Allen, G. Taylor, J. Chem. Technol. Biotechnol. 80 (2005) 688–698.
[10] A.I. Torres, M. Tsapatsis, P. Daoutidis, Comput. Chem. Eng. 42 (2012) 130–137.
[11] M. Goto, J.M. Smith, B.J. McCoy, Ind. Eng. Chem. Res. 29 (1990) 282–289.
[12] R.P. Swatloski, S.K. Spear, J.D. Holbrey, R.D. Rogers, J. Am. Chem. Soc. 124
(2002) 4974–4975.
4.2.4. Effects of distribution coefficients
In this model, the distribution coefficients R1, R2 are assumed,
due to the lack of precise mass transfer parameters at high temper-
ature. But the mass transfer plays an important role on this process,
caused by its effect on the product distribution in the two phases
in fact. Table 4 shows the effect of R1 value on the product dis-
tribution while R2 remains 1.0, and the effect of R2 value on the
product distribution while R1 remains 0.5. When the value of R1
Table 4
Product concentration and flow rate at various distribution coefficientsa.
[13] J.B. Binder, M.J. Gray, J.F. White, Z.C. Zhang, J.E. Holladay, Biomass Bioenerg. 33
(2009) 1122–1130.
Distributioncoefficient
Concentration (g/L) and flow rate (g/s)
[14] H. Xie, S. Li, S. Zhang, Green Chem. 7 (2005) 606–608.
[15] S. Hyvarinen, J.P. Mikkola, D.Y. Murzin, M. Vaher, M. Kaljurand, M. Koel, Catal.
Today 223 (2014) 18–24.
[16] W. Liu, J. Holladay, Catal. Today 200 (2013) 106–116.
[17] C. Li, Z.K. Zhao, Adv. Syn. Catal. 349 (2007) 1847–1850.
[18] H. Wang, G. Gurau, R.D. Rogers, Chem. Soc. Rev. 41 (2012) 1519–1537.
[19] J. Long, X. Li, L. Wang, N. Zhang, Sci. Chin. Chem. 55 (2012) 1500–1508.
[20] H. Zhao, J.E. Holladay, H. Brown, Z.C. Zhang, Science 316 (2007) 1597–1600.
[21] J. Long, B. Guo, X. Li, Y. Jiang, F. Wang, S.C. Tsang, L. Wang, K.M.K. Yu, Green
Chem. 13 (2011) 2334–2338.
R1
R2
cm
ch
61.5
57.9
54.7
105.7
79.5
57.9
vm
9.0
14.1
18.7
21.4
17.4
14.1
vh
0.3
0.5
0.7
0.5
0.5
0.5
1.0
1.0
1.0
0.5
0.7
1.0
21.2
19.9
18.8
30.2
24.5
19.9
87.4
82.2
77.7
75.0
79.0
82.2
a
v
1 = 100 g/s, VIL = 70 L, T = 473 K.
Please cite this article in press as: S. Liu, et al., Kinetic analysis and process modeling for cellulose valorization in cooperative ionic liquid
pairs, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.07.039

G Model
CATTOD-9714; No. of Pages8
ARTICLE IN PRESS
8
S. Liu et al. / Catalysis Today xxx (2015) xxx–xxx
[22] J. Long, B. Guo, X. Li, F. Wang, L. Wang, Acta Phys. Chim. Sin. 27 (2011)
995–999, J.
[23] J. Long, B. Guo, X. Li, F. Wang, L. Wang, Green Chem. 14 (2012) 1935–1941.
[24] W. Liu, F. Zheng, J. Li, A. Cooper, AIChE J. 60 (2013) 300–314.
[25] L. Chen, H. Zhang, J. Li, M. Lu, X. Guo, L. Han, Bioresour. Technol. 177 (2015)
8–16.
[26] L. Yan, A.A. Greenwood, A. Hossain, B. Yang, RSC Adv. 4 (2014) 23492–23504.
[27] K.D. Samant, K.M. Ng, AIChE J. 44 (1998) 2212–2228.
[28] H. Machida, R. Taguchi, Y. Sato, R.L. Smith Jr., J. Chem. Eng. Data 56 (2011)
923–928.
[29] C. Sun, J. Pharm. Sci. 94 (2005) 2132–2134.
[30] K.R. Harris, J. Chem. Soc. 78 (1982) 2265–2274.
Please cite this article in press as: S. Liu, et al., Kinetic analysis and process modeling for cellulose valorization in cooperative ionic liquid
pairs, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.07.039