Kinetics of hydrogenation of acetic acid to ethanol

Article abstract of DOI:10.14233/ajchem.2019.22277

The intrinsic kinetic behaviour of catalytic hydrogenation of acetic acid in vapour phase was studied over a multi-metallic catalyst. The rate expression was derived from the sequence of elementary reaction steps based on a Langmuir-Hinshelwood-model involving two types of active sites. Experiments were carried out in a fixed bed reactor, which is similar to an isothermal integral reactor designed to excluding the negative effects of internal and external diffusion. The reaction conditions investigated were as follow:reaction temperature 275-325 oC, reaction pressure1.5-3.0 MPa, liquid hourly space velocity (sv) 0.3-1.2 h^-1, molar ratio of hydrogen to acetic acid (H/AC) 8:20. The results show that conversion of acetic acid increases with increasing the reaction temperature and pressure, but decreases with increasing the space velocity and H/AC. Furthermore, reducing the reaction pressure and increasing reaction temperature, space velocity and H/AC can improve the reaction selectivity of acetic acid to ethanol. The established kinetic model results agreed with experimental results. The relative difference between the calculated value and the experimental value is less than 6 %. The values of model parameters are consistent with the three thermodynamic constraints. The study provided evidence that the intrinsic kinetic model is suitable both mathematically and thermodynamically, and it could be useful in guiding reactor design and optimization of operating conditions.

Full text of DOI:10.14233/ajchem.2019.22277

Asian Journal of Chemistry; Vol. 31, No. 12 (2019), 2915-2923
A
SIAN
J
OURNAL OF HEMISTRY
C
https://doi.org/10.14233/ajchem.2019.22277
Kinetics of Hydrogenation of Acetic Acid to Ethanol
*,
QIANG CHEN , XUEBING ZHANG, SHUXUN TIAN, JUNYING LONG, XIANGKUN MENG, QI SUN and YONGLONG LI
National Institute of Clean-and-Low-Carbon Energy, Beijing 102211, P.R. China
*Corresponding author: E-mail: qiang.chen.cx@chnenergy.com.cn
Received: 18 June 2019;
Accepted: 20 August 2019;
Published online: 16 November 2019;
AJC-19641
The intrinsic kinetic behaviour of catalytic hydrogenation of acetic acid in vapour phase was studied over a multi-metallic catalyst. The
rate expression was derived from the sequence of elementary reaction steps based on a Langmuir-Hinshelwood-model involving two
types of active sites. Experiments were carried out in a fixed bed reactor, which is similar to an isothermal integral reactor designed to
excluding the negative effects of internal and external diffusion. The reaction conditions investigated were as follow:reaction temperature
275-325 ºC, reaction pressure1.5-3.0 MPa, liquid hourly space velocity (sv) 0.3-1.2 h-1, molar ratio of hydrogen to acetic acid (H/AC)
8:20. The results show that conversion of acetic acid increases with increasing the reaction temperature and pressure, but decreases with
increasing the space velocity and H/AC. Furthermore, reducing the reaction pressure and increasing reaction temperature, space velocity
and H/AC can improve the reaction selectivity of acetic acid to ethanol. The established kinetic model results agreed with experimental
results. The relative difference between the calculated value and the experimental value is less than 6 %. The values of model parameters
are consistent with the three thermodynamic constraints. The study provided evidence that the intrinsic kinetic model is suitable both
mathematically and thermodynamically, and it could be useful in guiding reactor design and optimization of operating conditions.
Keywords: Kinetics, Hydrogenation, Esterification, Integral reactor, Conversion, Selectivity, Model.
as biomass, coal, or natural gas, to syngas, which can then be
converted catalytically to ethanol and higher alcohols [2,10].
During the past few years, urban air pollution has become
a serious problem and attracted great public attention in China.
The main source of emissions is vehicle exhaust [11,12]. Fuel
ethanol is one of the most promising choices that can reduce
emissions [13]. The plan to promote ethanol gasoline for motor
vehicles throughout the country by 2020 was published by
the government of China in September 2017. Due to the
shortage of oil and gas resources, ethanol is mainly produced
by fermentation of corns in China. In this process, feedstock
cost normally accounts for over 80 % of the total cost. It can be
found that corn price has been more than doubled over the past
decade. As the subsidy for corn ethanol producers gradually
phased out over recent years, the profit of producing corn ethanol
became lower or even negative [14]. Coal accounts for over
60 % of China’s total primary energy consumption and will
inevitably remain the overwhelming indigenous energy reso-
urce for the foreseeable future. The method of ethanol production
using syngas from coal gasification is suitable for China’s
INTRODUCTION
Ethanol is currently the largest produced alternative energy
world-widely to petroleum-derived engine fuels due to its comp-
atibility within existing spark-ignition engines and its relatively
well-developed production technologies [1]. Most of the studies
have broadly asserted that vehicle efficiency increases with
ethanol use to justify reducing the greenhouse gas impact of
ethanol [2,3].
Ethanol is produced by hydration of ethylene [4-6] or
fermentation of biomass that has sugar, starch or cellulose [3,7].
Hydration of ethylene is a proven industrial process [4], but it
relies on availability and cost of ethane. Presently, the produc-
tion of ethanol by fermentation of carbohydrates is the primary
route for gasoline additive inAmerica, Brazil and Europe [8,9].
However, it depends on availability of land area, soil, water,
price of feedstocks and even the local policy, which is very limited
for most of the countries [3]. Hence, there is a drive to explore
new fuel production methods to meet the increasing fuel demand.
One method is the conversion of carbon-based feedstocks, such
This is an open access journal, and articles are distributed under the terms of the Attribution 4.0 International (CC BY 4.0) License. This
license lets others distribute, remix, tweak, and build upon your work, even commercially, as long as they credit the author for the original
creation. You must give appropriate credit, provide a link to the license, and indicate if changes were made.

2916 Chen et al.
Asian J. Chem.
actual condition. For this reason, many Chinese researchers
have carried out the study of coal to ethanol technology.
Ethanol can be produced from both syngas directly and
other various sources such as acetic acid, ethyl acetate, methyl
acetate and dimethyl ether, which are conversed from syngas
[15]. Research on the production of alcohols from syngas directly
has been going on for decades, however the reaction exhibits
poor selectivity of ethanol and remains challenging [2]. The
processes of syngas to methanol [16,17] and carbonylation of
methanol to acetic acid [18] are mature technologies. Thus,
the hydrogenation process of acetic acid to ethanol is a promi-
sing choice.
acetaldehyde appear to be kinetically significant steps. The
investigations showed that Pt/Sn-based catalysts are selective
for conversion of acetic acid to form acetaldehyde and ethanol,
whereas platinum catalysts completely decompose acetic acid
to gas productions [25-27]. Additionally, Zhang [27] added
esterification of acetic acid and ethanol to ethyl acetate occurred
on the oxide surface to the elementary steps, therefore, selectivity
of ethanol was led into the kinetics model.A more recent study
of hydrogenation of acetic acid over alumina or silica supported
Cu/In and Ni/In catalysts was carried out, and the results showed
that the activity dependence on the reactant partial pressures
denotes the rate-determining surface reaction in terms of
Langmuir-Hinshelwood kinetics [28,29].
Many patents have reported the catalysts for hydrogena-
tion of acetic acid. In these studies, most of the catalysts were
one or more noble metals in Group VIII, dispersed on Group
III or IV metal oxides. Rachmady andVannice [19-23] carried
out a series of researches on platinum catalysts supported on
TiO2, SiO2, Al2O3 and Fe2O3, and the results were compared
with that obtained without support. Several observations were
made about the kinetic behaviour of acetic acid hydrogenation
to form organic compounds over supported platinum catalysts:
(a) the reaction requires both metal and an appropriate oxide
phase in the catalyst; (b) oxides that are active for ketoniza-
tion are the best supports, implying the reaction can occur on
the oxide surface; (c) platinum acts as a source of activated
hydrogen, presumably hydrogen atoms [19]. Based on these
findings, a Langmuir-Hinshelwood-type mechanism was
proposed, in which acetaldehyde is formed as the first initial
product, and it can react with additional hydrogen atoms to
form ethanol [19,22]. The reaction model similar to that used
to describe acetic acid reduction over Pt/TiO2 was applied to
this reaction over Fe/SiO2, with the only difference being that
the rate-determining steps involved the addition of the first H
atom to an acetate species [19] rather than an acyl species [22].
Non-local density functional theory (DFT) calculations were
used to examine alternative mechanisms for the hydrogenation
of acetic acid to ethanol over different catalysts [24]. Using
the overall reaction energies to deduce a plausible mechanism
for acetic acid hydrogenolysis, Pallassana and Neurock [24]
found that the acetyl formation and acetyl hydrogenation to
Within all reported catalysts, it was difficult to obtain a
high conversion of acetic acid while keep a high selectivity of
ethanol simultaneously. Our research group developed a multi-
metallic based catalyst, which was proven to be more efficiency
in both conversion and selectivity. At this point, investigating
the kinetics behaviour of hydrogenation reaction of acetic acid
to ethanol over this multi-metallic based catalyst becomes
critical to scale up in the future.
EXPERIMENTAL
Acetic acid (99.5 % purity) was supplied by Sinopharm
Chemical Reagent Co. Ltd., China. Hydrogen from Air Liquide
(China) Holding Co. Ltd. was 99.99 vol. %. The kinetic tests
were carried out on a multi-metallic catalyst.
Procedure: Hydrogenation of acetic acid system is shown
in Fig. 1. High pressure hydrogen from the cylinder wasdepre-
ssurized by a pressure reducing valve, and the hydrogen flow
rate was regulated by a mass flow controller. The acetic acid
was pumped into the reaction system and was well mixed with
hydrogen by an on-line mixer. A fixed bed reactor similar to
an isothermal integral reactor was used for the kinetics testing.
Acetic acid was heated to vapour phase in the front of reactor.
Hydrogenation reaction was down in the catalyst bed located
in the mid of reactor. The final products were cooled through
a condenser and entered a gas liquid separator tank. The liquid
products including ethanol, ethyl acetate, acetaldehyde, other
trace components and unreacted acetic acid were collected at
FC
Condenser
Reactor
Mass flow
controller
Hydrogen
Coolant
Coolant
Furnace
Acetic acid
tank
Pressure indicator
PI
Gas products
Pump
Gas liquid
separator
TE
Liquid products
Thermocouple
Fig. 1. Schematic diagram of acetic acid hydrogenation system

Vol. 31, No. 12 (2019)
Kinetics of Hydrogenation of Acetic Acid to Ethanol 2917
the bottom of tank, while the gas products were discharged
into the vent system from the top of tank.
shown in Fig. 3. The results show that as increased F/w, the
deviation of conversion obtained from reactor A and B was
gradually reduced till the conversions obtained in two reactors
became almost identical when the F/w was higher than 0.45
mol g-1 h-1. The deviation was decreased to less 1 % at F/w of
0.33 mol g-1 h-1, corresponding to the linear velocity of 0.19
cm s-1 in reactor B. Therefore, in this kinetics study, the total
feeding mole flow was determined as faster than 0.33 mol g-1 h-1
(or the linear velocity was faster than 0.19 cm s-1).
Analysis: During the experiments, the liquid products were
taken out every 4 h for analysis by off-line Agilent Techno-
logies model 7890A GC-FID with an Innowax19091N-133
column (30 m length × 0.25 mm inner diameter × 0.25 µm
film thickness). The column temperature was ramped from an
initial oven temperature of 40 to 80 ºC with a heating rate of
2 ºC min-1, then heated to 125 ºC at 6 ºC min-1. Hydrogen served
as the carrier gas at a flow rate of 30 mL min-1.
100
Pre-experiments: The study of intrinsic kinetics needs
to eliminate the influences of internal and external diffusion
first. To investigate the effects of internal diffusion of reaction,
we loaded reactor with the same amount of catalyst with particle
size (dp) of 4.1, 1.3 and 0.6 mm. The acetic acid conversions
(x) were obtained and correlated with the particle sizes of catalyst.
All the experiments were carried out at conditions as follow:
reaction temperature (t ) of 275 ºC, pressure (p) of 2.5 MPa, the
space velocity (sv) of acetic acid (the ratio of volumetric flow
rate of acetic acid to volume of catalyst loaded, space velocity)
of 0.6 h-1, the mole ratio of hydrogen to acetic acid (H/AC) of
16. The results as plotted in Fig. 2 show that the conversion of
acetic acid is increased with the decrease of catalyst particle
size, indicating the effects of the internal diffusion are reduced.
The deviation of acetic acid reaction conversions on the catalyst
with particle size of 1.3 mm and 0.6 mm is less than 1 %, which
denotes the effect of internal diffusion has been basically
eliminated. In the kinetics experiments, the catalyst was ground
to particle size of 0.6 mm.
A
B
98
96
94
92
90
0.1
0.2
0.3
(F/w) (mol g–1 h–1)
Fig. 3. Conversion in various F/w of reactor A and B
0.4
0.5
0.6
0.7
Design of experiments: The kinetics experiments were
carried out in a fixed bed reactor system as shown in Fig. 1.
The tests were performed at various combinations of the condi-
tions as shown in Table-1. Totally 45 sets of data were collected
for the kinetics study.According to the conditions at 0.344 MPa
(a) the maximum partial pressure, the boiling point temperature
of acetic acid is 196.7 ºC [30] which is lower than the minimum
reaction temperature 250 ºC, implying the reaction was studied
in the vapour phase.
100
90
80
70
60
50
TABLE-1
RANGE OF PROCESS CONDITIONS
Process conditions
Level
number
t (°C)
250
275
300
325
p (MPa)
sv (h–1)
0.3
0.6
0.9
H/AC 1
L1
L2
L3
L4
1.5
2.0
2.5
3.0
8
12
16
20
1.2
0
1
2
3
4
5
sv = space velocity
Particle size (mm)
Fig. 2. Conversion in various particle sizes
RESULTS AND DISCUSSION
To eliminate the effect of external diffusion, we conducted
reactions in two fixed bed reactors (reactor A and B) with the
same size and configuration to ensure comparability. The
amount of catalyst used in reactor B was 60 % of the catalyst
amount used in reactorA. Two sets of data were obtained under
the following conditions: t = 325 ºC, p = 2.5 MPa, space
velocity = 0.3 ~ 0.9 h-1, H/AC = 16, and the ratio of total feeding
mole flow to the mass of catalyst (F/w) was the same between
reactorA and B, so that the linear velocity of the feed in reactor
B was slower than that in A. The acetic acid conversion data
collected with various F/w ratios in both reactor A and B are
Total experimental error of acetic acid conversion is about
0.38 %, while the total experimental error of ethanol selectivity
is 0.24 %. When the reaction temperature was lower than 300
ºC, the total selectivity of ethanol and ethyl acetate was higher
than 97.5 %, the selectivity of acetaldehyde was lower than
2.5 %, and the total by-products was lower than 0.5 %. The
selectivity of acetaldehyde significantly related to the reaction
temperature, when the temperature was increased to 325 ºC,
acetaldehyde selectivity was raised up to 6 % while the total
selectivity of ethanol and ethyl acetate was around 93 %. In

2918 Chen et al.
Asian J. Chem.
order to simplify the kinetic model, it was assumed that only
products are ethanol and ethyl acetate. Therefore, over present
catalyst, there are two dominant reactions investigated:
tion of acetic acid to ethanol and the esterification of acetic
acid with ethanol to ethyl acetate are two chain reactions. The
first one is a molecular number reduced reaction while the
molecular number of the second reaction is constant. The incre-
ase of pressure allows the hydrogenation equilibrium to proceed
to the right side of the equation, which is beneficial to the prod-
uction of ethanol.
CH3COOH + 2H2 → C2H5OH + H2O
(1)
CH3COOH + C2H5OH → CH3COOC2H5 + H2O (2)
Effect of reaction conditions
Acetic acid space velocity: The effect of space velocity
of acetic acid was studied at 0.3, 0.6, 0.9 and 1.2 h-1 as displayed
in Fig. 6, the conversion dropped significantly with the rise of
space velocity, while the selectivity of ethanol increased grad-
ually. That is because the residence time reduced with the incre-
ase of space velocity which results in lower selectivity of ethyl
acetate via secondary esterification reaction. The selectivity
deviation of ethanol and ethyl acetate indicates that the acetic
acid hydrogenation rate and esterification rate are different at
different space velocity.
Temperature: The effect of reaction temperature on the
activity and selectivity was studied at 275, 300 and 325 ºC
(Fig. 4). The results show that the conversion increases slowly
with increasing the reaction temperature, while the selectivity
of ethanol drops dramatically. The trend of selectivity with
the reaction temperature is against that of conversion, which
is similar with the published results [26,31]. It indicates that
rising temperature leads to the promotion of the reaction rates
of hydrogenation and esterification, but esterification rate is
more sensitive to temperature than hydrogenation.
100
100
90
90
80
80
x-exp
70
s-exp
x-exp
70
x-cal
s-cal
60
s-exp
x-cal
s-cal
325
60
50
0
0.3
0.6
0.9
1.2
1.5
Space velocity of acetic acid (h–1)
50
265
Fig. 6. Effect of acetic acid space velocity (Condition: t = 300 °C, p = 2.0
MPa, H/AC = 12)
275
285
295
305
315
335
Temperature (°C)
Fig. 4. Effect of temperature (condition: p = 2.5 MPa, sv = 0.6 h-1, H/AC =
16)
Mole ratio of hydrogen to acetic acid:As shown in Fig.
7, acetic acid conversion decreased slightly but ethanol select-
ivity increased when the H/AC ratio was increased from 8 to
12, while the selectivity of ethanol increased slowly. When
the total reaction pressure was the same and the H/AC ratio
was increased, the partial pressure of hydrogen increased accom-
Reaction pressure: The impact of reaction pressure was
studied at 2.0, 2.5 and3.0 MPa as displayed in Fig. 5. Both the
conversion and the selectivity increased gradually with the
increase of pressure. For gas phase reactions, partial pressures
of reactants have a positive effect on reaction rate. The reduc-
100
100
90
90
80
80
x-exp
x-exp
70
s-exp
70
s-exp
x-cal
s-cal
60
x-cal
60
s-cal
50
6.0
8.0
10.0
12.0
14.0
50
1.5
2.0
2.5
3.0
3.5
H/AC
Pressure (MPa)
Fig. 7. Effect of mole ratio of hydrogen to acetic acid (Condition: t = 300
°C, p = 2.0 MPa, sv = 0.9 h-1)
Fig. 5. Effect of pressure (condition: t = 275 °C, sv = 1.2 h-1, H/AC = 16)

Vol. 31, No. 12 (2019)
Kinetics of Hydrogenation of Acetic Acid to Ethanol 2919
panied with the decrease of the partial pressure of acetic acid.
Because hydrogen was always in excess in the reaction system,
the effect of increasing hydrogen partial pressure was not able
to overtaken the negative effect of reducing acetic acid partial
pressure. Moreover, the shorter residence time with the increase
of H/AC ratio should also be taken into account.
Kinetics model and regression: Previous kinetic studies
of acetic acid hydrogenation over bimetallic catalysts showed
that the reaction can be described by a Langmuir-Hinshelwood-
type mechanism involving two types of active sites, ones are
on the metal to activate hydrogen and another ones are on the
oxide to adsorb and activate acetic acid [19,27].
r2 = k′5θHOAcθEt OH
(16)
where θi is the fractional surface coverage of species i, θHOAc is
the fractional surface coverage of acetic acid on the oxide surface,
EtOH isthe fractional surface coverage of ethanol on the oxide
surface, CH is the concentration of hydrogen atoms on the sites
of oxide surface.
θ
The equilibrium expressions representing adsorption on
the metal surface can be obtained from eqns. 3 and 5:
θH2
KH
=
PH θ*2
(17)
2
2
θAcθH
KAc
=
PHOAcθ*2
(18)
Based on the reported results, a Langmuir-Hinshelwood-
type catalytic elementary reaction sequence was proposed as
in the following eqns. 3-13:
where Pi is the partial pressureof species i, subscripts H, Ac
and * represent hydrogen atoms, acetate and vacant sites on
the metal surface.
The equilibrium expressions representing adsorption on
the oxide surface can be obtained from eqns. 6 and 11:
KH
H2(g) + 2*
2H*
2
(3)
Ksp
KAc
H*
Hsp
(4)
(5)
θHOAc
CH3COOH(g) + 2∗
CH3COO∗ + H∗
KHOAc
=
(19)
PHOAcθS
KHOAc
θEtOH
CH3COOH(g) + S
CH3COOH −S
(6)
(7)
(8)
(9)
KEtOH
=
(20)
PEtOHθS
k1'
CH3COOH − S+ Hsp →CH3COOH2 −S
where subscripts HOAc, EtOH and S representacetic acid,
ethanol and vacant sites on the oxide surface.
k2'
It is assumed that adsorbed hydrogen atoms (H*) and
adsorbed acetate species (CH3COO*) are the predominant surface
species on metal sites; while the molecular acetic acid and
ethanol are the dominated surface intermediates on the oxide
surface sites. Two balances of the fractional surface coverage
on metal and oxide are given in eqns. 21 and 22:
CH3COOH2 − S+ Hsp →CH3CHO − S+ H2O(g)
k3'
CH3CHO − S + Hsp →CH3CHOH − S
k'4
(10)
(11)
CH3CHOH − S+ Hsp →C2H5OH −S
KEtOH
C2H5OH(g) + S
C2H5OH −S
θH + θAc + θ* = 1
(21)
(22)
k'5
CH3COOH −S+ C2H5OH −S →
CH3COOC2H5 −S+ H2O(g) + S
θ
HOAc + θEtOH + θS = 1
(12)
The expression of θ* is given in eqn. 23 evaluated from
k'6
eqns. 17, 18 and 21:
CH3COOC2H5–S
CH3COOC2H5 (g) + S (13)
1
k'-6
θ* =
KAcPHOAc
where * and S represent catalytic sites on the metal and oxide
surface, respectively k' is the rate coefficients of reaction, and
K is the equilibrium adsorption constants.
1+ KH PH
+
(23)
2
2
KH PH
2
2
The expression of θS is given in eqn. 24 evaluated from
The investigations showed that the addition of first H atom
to an acetate species is the rate-determining steps of the sequence
of elementary steps used to describe acetic acid reduction over
Pt/TiO2 [19,22]. The kinetics and mechanism investigations
of acetic acid esterification with ethanol on four zeolites by
Bedard and co-workers [32] showed that experiments data were
consistent with the formation of an energetically favourable
acetic acid and ethanol complex adsorbed on the zeolite active
site and both are involved in the rate-determining step.Applying
the steady-state approximation to the surface intermediates,
and suggesting that both eqns. 7 and 12 are the rate-determining
steps.
eqns. 19, 20 and 22:
1
θs =
(24)
1+ KHOAcPHOAc + KEtOHPEtOH
The equilibrium expressions for H atom concentration
on the oxide surface can be obtained from eqn. 4:
C
H θH
Ksp
=
(25)
The final rate expression of acetic acid disappearance is
obtained by substituting eqns. 17-20 and eqns. 23-25 into eqns.
15 and 16 to give
k1' KHOAcKsp KH PHOAc PH
The activity of hydrogenation and esterification together
determined the rate of acetic acid conversion, i.e., the overall
rate of formation of ethanol and ethyl acetate as shown in eqns.
14-16:
2
2
r =
1

 (26)
KAc
2
PHOAc
(1 + KHOAcPHOAc + KEtOHPEtOH )1 + KH PH
2

KH
PH
2
2


k'5KHOAcKEtOHPHOAcPEtOH
rHOAc = r1 + r2
(14)
(15)
r2 =
(27)
2
(1 + K HOAcPHOAc + K EtOHPEtOH
)
r1 = k′1θHOAcCH

2920 Chen et al.
Asian J. Chem.
The k′ and K found in eqns. 26 and 27 can be written as:
−27432
RT


KHOAc = exp −8.452 −
(36)
−E
RT





a
k' = k0 exp

(28)




−43879
RT


∆S ∆H


Ksp = exp −7.757 −
(37)
(38)
(39)
(40)



K = exp
−
(29)




R
RT


where Ea is the activation energy of the element reaction, ∆H
is the adsorption enthalpy, ∆S is the adsorption entropy, R is
the gas constant with the value of 8.314 J mol-1 K-1, T is the
thermodynamic temperature with the unit of K.
Considering that the fixed bed reactor as an ideal isothermal
plug-flow reactor, the rate expression of acetic acid around
one differential catalyst mass element (dw) is:
−25865
RT

K
= exp −10.158 −
H2





−44095
RT

KEtOH = exp −5.427 −





−71283
RT

KAC = exp −15.401−


o




d NHOAc (1− x)
dNHOAc
dw
dx
= NoHOAc


(30)
r + r2 = −
= −
Model evaluation: The x and s given by eqns. 32 and 33
were fitted to the experimental data. A histogram bar chart of
the residuals and a normal probability plot were created using
MATLAB function "histfit" and "normplot". They are shown
in Fig. 8a-b. The data appears along the reference line indicates
that the residual data have a normal distribution. The comp-
arison of the experimental and calculated x and s are showed
in Figs. 9 and 10, and the distribution of relative error of x and
s are also described in Figs. 11 and 12. The results show that the
calculated values are in good fitting trend with the experimental
values (blue and red scatters), and all the relative errors except
one are less than 6 %. In addition, both the coefficients and the
adjusted coefficients of determination are calculated: for x, R2
= 0.9256, R2 = 0.8944; for s, R2 = 0.9059, R2 = 0.8665. The
rate expression derived from this model gives the root-mean-
square error (RMSE) of its predicted x and s to be 2.17 and 2.13
%. It suggested that the proposed model fits the experimental
data with an excellent goodness-of-fit.
1
dw
dw
o
NoHOAc
2


d NHOAcx(1−s)
dNEtOAc
=
(1−s)dx − xds




r2 =
=
(31)


dw
2dw
dw


which can be integrated over the entire catalyst mass to give
w
r + r2
1
x =
dw
(32)
o
∫
0 NHOAc
w (1 − s)r − (1 + s)r
1
s =
2 dw
(33)
xNoHOAc
∫
0
where NoHOAc is the molar flow rate of acetic acid at the inlet of
reactor.
Using a numerical integration method based on an explicit
Runge-Kutta formula (function "ode45") and a least-square
non-linear regression method (function "lsqcurvefit") which
are built-in MATLAB, we obtained every k′ and K founded in
eqns. 26 and 27. These optimum parameters are given as follow:
The model should be mathematically reasonable and
thermodynamically limited. The activation energy Ea of the
element reaction must be positive, and the adsorption enthalpy
∆H should be negative. Both absolute values of |Ea| and |-∆H|
should be in the range of 24 to 240 kJ mol-1. Some criteria
comprised of three strong rules and two guidelines to evaluate
whether rate parameters, such as the adsorption equilibrium
47839
RT


k1' = 2.985×104 exp −
(34)
(35)




52411
RT


k'5 = 4.365×103 exp −




18
16
14
12
10
8
0.997
(a) Histogram bar chart
(b) Normal probability plot
0.99
0.98
0.95
0.90
0.75
0.50
0.25
6
0.10
0.05
4
0.02
0.01
2
0.003
-0.05 -0.04 -0.03 -0.02 -0.01
0
Data
0.01 0.02 0.03 0.04
0
-0.08 -0.06 -0.04 -0.02
0
0.02
0.04
0.06
0.08
Fig. 8. Normality test of residuals

Vol. 31, No. 12 (2019)
Kinetics of Hydrogenation of Acetic Acid to Ethanol 2921
10
8
100
6
90
80
70
60
50
4
2
0
-2
-4
-6
-8
-10
50
60
70
80
90
100
50
60
70
80
90
100
sexp (%)
Fig. 12. Distribution of relative error of selectivity
xexp (%)
Fig. 9. Comparison between calculated and experimental conversion
constants appearing in the denominator of Langmuir isotherm
involving the enthalpy and entropy of adsorption were proposed
[33,34]. These criteria are listed in Table-2, where Sg is the
standard total entropy in the gas phase. The calculated para-
meters are shown in Table-3.
100
90
80
70
60
50
TABLE-2
CRITERIA TO EVALUATE PARAMETERS
No.
C1
C2
C3
Criteria unit (J mol-1)
24 × 103 ≤ Ea and –∆H ≤ 240 × 103
0 < – ∆S < Sg
41.8 ≤ |∆S| ≤ 51.0 – 0.0014∆H
These results give the evidences of that model is suitable
thermodynamically, also imply that dissociative acetic acid
adsorption on the oxide surface is much weaker than that on
metallic iron in accordance with previous report [22]. Future
works such as the causes of a low selectivity of acetaldehyde,
which is the intermediate of acetic acid to ethanol, and the mech-
anism of esterification over the multi-metallic catalysts still
require more researches.
The profile of conversion of acetic acid, and the selectivity
of ethanol and ethyl acetate along the axial direction of fixed
bed reactor were calculated using present model. The result is
shown in Fig. 13. The x-axis is the total mass fraction of catalyst
along the flow direction. This pre-calculation can provide guid-
ance for reactor design.
50
60
70
80
90
100
xexp (%)
Fig. 10. Comparison between calculated and experimental selectivity
10
8
6
4
2
Conclusion
0
The hydrogenation of acetic acid to ethanol was con-
ducted in a fixed bed reactor. The pre-experiments results show
that the influence of internal diffusion and external diffusion
could be mostly eliminated by using a catalyst with the particle
size less than 1.3 mm and keeping the linear velocity of greater
than 0.19 cm s-1. The kinetics experiments were investigated
at 275-325 ºC, 1.5-3.0 MPa, 0.3-1.2 h-1 of acetic acid space
velocity (sv) and 8-20 of mole ratio of hydrogen to acetic acid
(H/AC). The results showed that increasing pressure and tem-
perature, and reducing space velocity and H/AC can improve
the conversion. On the other hand, reducing pressure and
-2
-4
-6
-8
-10
50
60
70
80
90
100
xexp (%)
Fig. 11. Distribution of relative error of conversion

2922 Chen et al.
Asian J. Chem.
TABLE-3
VALIDATION OF MODEL PARAMETERS
k or K
Sg (J mol-1)
Ea or ∆H (kJ mol–1)
47.839
∆S (J mol–1)
–
51.0 – 0.0014 ∆H (J mol–1)
Hydrogenation rate parameter (k1')
Esterification rate parameter (k5')
–
–
–
–
52.411
–
Acetic acid on oxide (KHOAc
H atoms migration to oxide (Ksp)
)
-27.432
-43.879
-25.865
-44.095
-70.3
-64.5
-84.5
-45.1
-128.0
293.4
114.7
130.7
282.7
293.4
86.4
98.0
89.5
85.8
145.3
Hydrogen on metal (KH )
2
Ethanol on oxide (KEtOH
)
Acetic acid on metal (KAC)
-71.283
4. C.R. Nelson, M.A.D. Taylor, D.D. Davidson and L.M. Peters, Olefin
Hydration Process (1951).
100
5. H. Momose, K. Kusumoto, Y. Izumi and Y. Mizutani, J. Catal., 77, 23
(1982);
https://doi.org/10.1016/0021-9517(82)90142-7.
6. N. Katada, Y. Iseki, A. Shichi, N. Fujita, I. Ishino, K. Osaki, T. Torikai
and M. Niwa, Appl. Catal. A Gen., 349, 55 (2008);
https://doi.org/10.1016/j.apcata.2008.07.005.
7. S. Kumar, N. Singh and R. Prasad, Renew. Sustain. Energy Rev., 14,
1830 (2010);
https://doi.org/10.1016/j.rser.2010.03.015.
8. G.W. Huber, S. Iborra and A. Corma, Chem. Rev., 106, 4044 (2006);
https://doi.org/10.1021/cr068360d.
9. Renewable Fuels Association, 2018 Ethanol Industry Outlook, 2018.
http://www.ethanolresponse.com/wp-content/uploads/2018/02/2018-
RFA-Ethanol-Industry-Outlook.pdf.
10. Y. Zhang, X. San, N. Tsubaki, Y. Tan and J. Chen, Ind. Eng. Chem.
Res., 49, 5485 (2010);
https://doi.org/10.1021/ie901882s.
11. X. Wu, Y. Wu, S. Zhang, H. Liu, L. Fu and J. Hao, Environ. Pollut., 214,
556 (2016);
80
60
40
20
0
x
s
0
20
40
60
80
100
Mass fraction of catalyst (%)
Fig. 13. Profile of x and s along the axial direction (Condition: t = 285 °C,
p = 3.0 MPa, sv = 0.8 h-1, H/AC = 16)
https://doi.org/10.1016/j.envpol.2016.04.042.
increasing temperature, liquid hourly space velocity (sv) and
molar ratio of hydrogen to acetic acid (H/AC) can improve
the selectivity of ethanol. The optimized model parameters
were obtained by using a numerical integration method based
on an explicit Runge-Kutta formula and a least-square non-
linear optimization method. Calculated conversion and selec-
tivity were in good agreement with experimental results.The
parameters are consistent with three thermodynamic cons-
traints. Studies in this paper provided evidence that the estab-
lished intrinsic kinetic model was both mathematically and
thermodynamically reasonable, and it could be used to guide
reactor design.
12. Y. Wu, S. Zhang, J. Hao, H. Liu, X. Wu, J. Hu, T.J. Wallington, M.P.
Walsh, K.M. Zhang and S. Stevanovic, Sci. Total Environ., 574, 332 (2017);
https://doi.org/10.1016/j.scitotenv.2016.09.040.
13. J. Jiao, J. Li and Y. Bai, J. Clean. Prod., 180, 832 (2018);
https://doi.org/10.1016/j.jclepro.2018.01.141.
14. H. Hao, Z. Liu, F. Zhao, J. Ren, S. Chang, K. Rong and J. Du, Renew.
Sustain. Energy Rev., 82, 645 (2018);
https://doi.org/10.1016/j.rser.2017.09.045.
15. B. Qian, Chem. Ind., 32, 26 (2014);
https://doi.org/10.3969/j.issn.1673-9647.2014.02.006.
16. A.B. Stiles, F. Chen, J.B. Harrison, X. Hu, D.A. Storm and H.X. Yang,
Ind. Eng. Chem. Res., 30, 811 (1991);
https://doi.org/10.1021/ie00053a002.
17. R.G. Herman, Catal. Today, 55, 233 (2000);
https://doi.org/10.1016/S0920-5861(99)00246-1.
18. N. Yoneda, S. Kusano, M.Yasui, P. Pujado and S. Wilcher, Appl. Catal.
A Gen., 221, 253 (2001);
ACKNOWLEDGEMENTS
This work was supported by National Key Research and
Development Program of China (2018YFB0604703). The
authors thank Ping Miao, Jihong Cheng,Yunjian Hu and Ruili
Tong for discussions about of experimental methods.
https://doi.org/10.1016/S0926-860X(01)00800-6.
19. W. Rachmady and M. Vannice, J. Catal., 192, 322 (2000);
https://doi.org/10.1006/jcat.2000.2863.
20. W. Rachmady and M.A. Vannice, J. Catal., 209, 87 (2002);
https://doi.org/10.1006/jcat.2002.3623.
21. W. Rachmady and M.A. Vannice, J. Catal., 207, 317 (2002);
https://doi.org/10.1006/jcat.2002.3556.
CONFLICT OF INTEREST
22. W. Rachmady and M.A. Vannice, J. Catal., 208, 158 (2002);
https://doi.org/10.1006/jcat.2002.3560.
23. W. Rachmady and M.A. Vannice, J. Catal., 208, 170 (2002);
https://doi.org/10.1006/jcat.2002.3561.
The authors declare that there is no conflict of interests
regarding the publication of this article.
24. V. Pallassana and M. Neurock, J. Catal., 209, 289 (2002);
https://doi.org/10.1006/jcat.2002.3585.
REFERENCES
25. R. Alcala, J.W. Shabaker, G.W. Huber, M.A. Sanchez-Castillo and J.A.
Dumesic, J. Phys. Chem. B, 109, 2074 (2005);
https://doi.org/10.1021/jp049354t.
26. K. Zhang, H. Zhang, H. Ma, W. Ying and D. Fang, Catal. Lett., 144, 691
(2014);
1. X.Yan, O.R. Inderwildi, D.A. King andA.M. Boies, Environ. Sci. Technol.,
47, 5535 (2013);
https://doi.org/10.1021/es305209a.
2. M. Gupta, M.L. Smith and J.J. Spivey, ACS Catal., 1, 641 (2011);
https://doi.org/10.1021/cs2001048.
https://doi.org/10.1007/s10562-014-1210-z.
3. E. Gnansounou and A. Dauriat, J. Sci. Ind. Res. (India), 64, 809 (2005).

Vol. 31, No. 12 (2019)
Kinetics of Hydrogenation of Acetic Acid to Ethanol 2923
27. K. Zhang, Ph.D. Thesis, Study on PtSn/Al2O3 Catalysts for Gas Phase
Acetic Acid Hydrogenation to Ethanol, East China University of
Science and Technology, Shanghai, China (2014).
28. G. Onyestyák, S. Harnos,A. Kaszonyi, M. Štolcová and D. Kalló, Catal.
Commun., 27, 159 (2012);
31. S. Tian, J. Cheng, W. Di, Q. Chen, J. Long, X. Luo, Y. Hu, X. Meng, S.
Sun and Q. Sun, J. Fuel Chem. Technol., 44, 862 (2016);
https://doi.org/10.3969/j.issn.0253-2409.2016.07.012.
32. J. Bedard, H. Chiang and A. Bhan, J. Catal., 290, 210 (2012);
https://doi.org/10.1016/j.jcat.2012.03.020.
https://doi.org/10.1016/j.catcom.2012.07.021.
29. G. Onyestyák, S. Harnos, S. Klébert, M. Stolcová, A. Kaszonyi and D.
Kalló, Appl. Catal. A Gen., 464–465, 313 (2013);
33. M. Boudart, D.E. Mears and M.A. Vannice, Ind. Chim. Belge., 32, 281
(1967).
34. M.A. Vannice, S.H. Hyun, B. Kalpakci and W.C. Liauh, J. Catal., 56,
358 (1979);
https://doi.org/10.1016/j.apcata.2013.05.042.
30. R.D. Lide, CRC Handbook of Chemistry and Physics, CRC Press: Boca
Raton, FL (2005).
https://doi.org/10.1016/0021-9517(79)90128-3.