Power-law kinetics of methanol synthesis from carbon dioxide and hydrogen on copper-zinc oxide catalysts with alumina or zirconia supports

Article abstract of DOI:10.1016/j.cattod.2015.11.020

Kinetics of methanol synthesis from carbon dioxide and hydrogen were studied on two catalysts, a copper-zinc oxide-alumina catalyst (CuZA) and a copper-zinc oxide-zirconia (CuZZ) catalyst. Although both catalysts show similar turnover frequencies for the methanol synthesis reaction, CuZZ is more selective for methanol synthesis because the reverse water gas shift reaction occurs more slowly on this catalyst. The results of the catalytic tests were modeled with power-law equations which highlight the strong positive impact of hydrogen partial pressure on methanol synthesis activity and selectivity. The comparison of the experimental results with thermodynamic equilibrium allows separating thermodynamic and kinetic driving forces.

Full text of DOI:10.1016/j.cattod.2015.11.020

G Model
CATTOD-9911; No. of Pages12
ARTICLE IN PRESS
Catalysis Today xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Power-law kinetics of methanol synthesis from carbon dioxide and
hydrogen on copper–zinc oxide catalysts with alumina or zirconia
supports
Kilian Kobl, Sébastien Thomas, Yvan Zimmermann, Ksenia Parkhomenko,
∗
Anne-Cécile Roger
Institut de Chimie et Procédés pour l’Énergie, l’Environnement et la Santé, UMR 7515CNRS-Université de Strasbourg, Groupe “Énergies et Carburants pour
un Environnement durable”, 25 rue Becquerel, 67087 Strasbourg Cedex 2, France
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 1 July 2015
Received in revised form 30 October 2015
Accepted 10 November 2015
Available online xxx
Kinetics of methanol synthesis from carbon dioxide and hydrogen were studied on two catalysts, a
copper–zinc oxide–alumina catalyst (CuZA) and a copper–zinc oxide–zirconia (CuZZ) catalyst. Although
both catalysts show similar turnover frequencies for the methanol synthesis reaction, CuZZ is more
selective for methanol synthesis because the reverse water gas shift reaction occurs more slowly on
this catalyst. The results of the catalytic tests were modeled with power-law equations which highlight
the strong positive impact of hydrogen partial pressure on methanol synthesis activity and selectiv-
ity. The comparison of the experimental results with thermodynamic equilibrium allows separating
thermodynamic and kinetic driving forces.
Keywords:
Methanol synthesis
Carbon dioxide hydrogenation
Kinetics
©
2015 Elsevier B.V. All rights reserved.
Copper catalyst
Zirconia
Alumina
1
. Introduction
trially from synthesis gas mixtures (CO/CO /H ) over Cu/ZnO/Al O
2 2 2 3
catalysts at 50 to 100 bar total pressure and at a temperature
◦
In order to limit global warming, worldwide efforts to reduce
between 200 and 300 C [8]. Pilot and commercial plants for
anthropogenic CO2 emissions are underway [1]. This is particu-
larly relevant to the energy sector, which accounts for over 60% of
global greenhouse gas emissions [2]. While the share of renewable
energies for electricity generation is getting more and more impor-
tant in the global energy mix, the actual electricity production from
renewable sources like wind or solar energy is intermittent, caus-
ing a negative impact on electric grid stability if production and
demand do not match [3,4]. To address these interrelated problems,
the French research project VItESSE2 was initiated [5]. Its goal is to
methanol synthesis using only CO2 as a carbon source and operat-
ing in steady-state have already been launched [9,10]. Among the
requirements for a flexible design of the novel process envisaged
here, sufficiently active and selective catalysts adapted to the inter-
mittency of the process are needed [11]. Furthermore, for reactor
design, knowledge of the kinetics of the catalytic chemical trans-
formation is necessary [12].
Methanol synthesis from carbon oxides can be described using
the following reaction equations: methanol synthesis from CO2
(Eq. (1)), the reverse water gas shift reaction (RWGS, Eq. (2)) and
methanol synthesis from CO (Eq. (3)).
couple two objectives, valorization of CO emissions and electricity
2
grid stabilization, via a flexible methanol synthesis process. In delo-
calized units, CO captured from major industrial emission sources
2
CO +3H2 ꢀ CH OH+H O
(1)
(2)
(3)
would react with hydrogen produced by water electrolysis from
available decarbonized excess electricity from the grid [6].
With a worldwide production of around 60 Mt in 2013,
methanol is an important basic chemical [7]. It is produced indus-
2
3
2
CO + H2 ꢀ CO + H O
2
2
CO + 2H2 ꢀ CH OH
3
As a consequence of the industrial importance of methanol,
numerous kinetic models already exist to describe the reaction.
Early models exclusively accounted for methanol synthesis from
CO in CO/CO2 synthesis gas mixtures [13–16]. For these models, it
∗ Corresponding author.
E-mail address: annececile.roger@unistra.fr (A.-C. Roger).
http://dx.doi.org/10.1016/j.cattod.2015.11.020
0
920-5861/© 2015 Elsevier B.V. All rights reserved.
Please cite this article in press as: K. Kobl, et al., Power-law kinetics of methanol synthesis from carbon dioxide and hydrogen on
copper–zinc oxide catalysts with alumina or zirconia supports, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.11.020

G Model
CATTOD-9911; No. of Pages12
ARTICLE IN PRESS
K. Kobl et al. / Catalysis Today xxx (2015) xxx–xxx
Table 1
2
was assumed that methanol is produced from CO only. This implies
that they cannot be used with pure CO feed streams. Subsequently,
Catalyst and catalyst bed properties.
2
complex kinetic models have been developed, which take into
account the experimental fact that methanol is produced mainly
CuZA
CuZZ
Mass fraction Cu (%)
Mass fraction CuO (%)
Mass fraction ZnO (%)
Mass fraction support (%)
Support type
30
37.5
41
21.5
Al2O3
0.44
135
50–125
82
30
37.5
41
21.5
ZrO2
1.51
135
50–125
71
from CO even using CO-containing feeds [16–20]. Power-law mod-
2
els, on the other hand, require no hypothesis about the reaction
mechanism, which is still under debate in the literature [21,22].
Such models have been proposed by Peter et al. and Ledakowicz
et al. [23,24]. However, these empirical models can only be used in
a defined range of partial pressures, conversions and temperatures.
This work aims to provide recent input data for reactor and pro-
Apparent density (g cm−3)
Catalyst mass (mg)
Particle size (␮m)
2
−1
BET surface area (m
g
)
)
0
2
−1
g
Cu surface area (m
9.6
8.3
cess design for the intermittent CO valorization process described
2
above. Kinetic data of methanol synthesis from CO /H feeds over
SiC mass (mg)
Bed volume (cm )
Bed height (cm)
Internal reactor diameter (cm)
External reactor diameter (in)
–
387
2
2
3
0.307
0.382
1.01
0.5
0.307
0.382
1.01
0.5
a Cu/ZnO/Al O and a Cu/ZnO/ZrO2 catalyst were collected and
2
3
power-law models were developed.
2
. Experimental
catalyst mass was fixed to 135 mg and CuZZ exhibits a higher appar-
ent density than CuZA, CuZZ was diluted with inert SiC in order
to obtain the same GHSV for both catalysts. The reaction condi-
tions are summarized in Table 2. The following procedure was
2
.1. Catalyst characterization
Catalysts 30 wt% Cu/ZnO/Al O (CuZA) and 30 wt% Cu/ZnO/ZrO
2
2
3
(
CuZZ) were prepared by coprecipitation of nitrate solutions of the
used for the experiments: the catalyst was reduced under H flow
respective metals with Na CO3 solution [11].
2
2
−
◦
1
(
6.4 mLSATP min ) at atmospheric pressure (ramp from room tem-
perature to 280 C, rate 1 C min , isotherm for 12 h). The reduced
catalyst was then cooled down to 100 C under the same flow.
The copper surface area of the catalysts was measured by the
◦
−1
N O reactive frontal chromatography method using a Micromerit-
ics AutoChem II analyzer with TCD detector at a total flow rate of
5
2
2
◦
−
min [25]. After reduction of 400 mg of the sample at
SATP
1
The desired reaction gas flow rates were adjusted to a total vol-
0 mL
−
1
◦
◦
−1
umetric flow rate of 40 mLSATP min and verified using an Agilent
ADM 1000 flow meter. Then the reactor was pressurized to the
desired total pressure (usually 50 bar) under reaction flow and the
gas phase composition was allowed to stabilize. The temperature
80 C, ramp 1 C min , isotherm for 12 h under 10% H /Ar flow,
2
◦
N O chemisorption was carried out at 50 C under a 2% N O/Ar flow.
2
2
Specific surface areas measurements were performed
◦
by nitrogen adsorption–desorption at 196 C using the
Brunauer–Emmet–Teller (BET) method on a Micromeritics ASAP
2
3
◦
−1
was ramped at 1 C min to the desired reaction temperature. The
moment when this temperature was reached was defined as the
starting point of the reaction.
◦
420 apparatus. Samples were previously outgassed at 250 C for
h to remove adsorbed moisture.
2.4. Data treatment
2.2. Reaction setup
For the calculation of conversions and selectivities, the amount
of substance integrated over the entire duration of the experiment
was used. For liquid products, the weighed mass of product i was
A gas mixture bottle of molar composition 31.5% CO , 63.5% H2
2
and 5.0% N2 was purchased from Air Liquide. Gas bottles of H , N2
2
and Ar were purchased from Air Liquide and SOL France.
Catalytic tests were carried out on a stainless steel reaction
setup. Gas flows of the mixture bottle, of pure H , N and Ar were
converted to amount of substance using n = m /M . For gases, the
i
i
i
amount of substance n was calculated by integrating the molar
i
2
2
flow rate F over the duration tR of the experiment.
regulated by Brooks SLA 5850S mass flow controllers connected to
Brooks 0254 secondary electronics. Pressure was controlled by a
Brooks 5866 back pressure regulator. The catalyst bed was placed
between two pieces of quartz wool retained by a quartz tube and
a metallic grid with gas flows pointing downwards. Temperature
was regulated by a PID controller with the regulation thermocou-
ple touching the catalyst bed at the effluent side. Liquid products
i
t
R
ꢀ
ni =
Fidt
(4)
0
Conversions (X) were calculated as average values for the total test
duration.
(
methanol and water) were trapped during the reaction in a first
n
+ nCO
CH3OH
trap at room temperature and afterwards in a second, water-cooled
XCO2 = 100% ×
XH2 = 100% ×
(5)
(6)
n
◦
CO2, in
trap at 15 C and analyzed offline using an Agilent 6890 N gas chro-
matograph equipped with a Solgelwax column and FID detector.
Gases CO , CO, H , N and Ar were analyzed online once per hour
2nCH3OH + nH2
O
2
2
2
n
H2, in
by an Inficon 3000 microchromatograph equipped with Molsieve
Selectivities (S) were calculated with respect to carbon-containing
products.
5
Å and Poraplot Q columns and TCD detectors.
No traces of methane or other hydrocarbons, dimethylether or
other alcohols than methanol were detected.
n
CH3OH
SCH
= 100% × n
(7)
(8)
3OH
CH3OH
+ nCO
2.3. Catalytic tests
SCO = 100% − SCH3OH
The catalytic tests were carried out using the catalyst beds
Methanol was assumed to be produced directly from CO2 (Eq.
(1)). This assumption is based on literature data [26–28] and on
the fact that the experiments were conducted at low conversions.
CO production was attributed to the reverse water gas shift reac-
tion (Eq. (2)). In addition, on a catalyst of similar composition to our
described in Table 1. A Cu/ZnO/Al O catalyst (CuZA) and a
2
3
Cu/ZnO/ZrO (CuZZ) catalyst were used. The catalyst powders were
2
sieved to a 50–125 ␮m particle fraction, yielding apparent den-
−3
−3
for CuZA and 1.51 g cm for CuZZ. As the
sities of 0.44 g cm
Please cite this article in press as: K. Kobl, et al., Power-law kinetics of methanol synthesis from carbon dioxide and hydrogen on
copper–zinc oxide catalysts with alumina or zirconia supports, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.11.020

G Model
CATTOD-9911; No. of Pages12
ARTICLE IN PRESS
K. Kobl et al. / Catalysis Today xxx (2015) xxx–xxx
3
Table 2
Overview of reaction conditions.
Inlet pressure (bar)
H2/CO2 ratio
GHSV (h 1
−
)
p (bar)
T ( C)
◦
H2
CO2
N2
Ar
Conditions
A
B
C
D
E
7800
7800
7800
7800
7800
50.0
50.0
50.0
50.0
50.0
200
210
220
230
240
34.9
34.9
34.9
34.9
34.9
8.9
8.9
8.9
8.9
8.9
6.2
6.2
6.2
6.2
6.2
0.0
0.0
0.0
0.0
0.0
3.9
3.9
3.9
3.9
3.9
1
2
3
4
5
7800
7800
7800
7800
7800
50.0
50.0
50.0
50.0
50.0
210/220
210/220
210/220
210/220
210/220
12.7
18.8
24.4
31.3
37.5
6.2
6.2
6.2
6.2
6.2
6.2
6.2
6.2
6.2
6.2
24.8
18.7
13.1
6.3
2.0
3.0
3.9
5.0
6.0
0.0
6
7
3
8
9
7800
7800
7800
7800
7800
50.0
50.0
50.0
50.0
50.0
210/220
210/220
210/220
210/220
210/220
24.4
24.4
24.4
24.4
24.4
12.0
8.1
6.2
4.9
4.1
6.2
6.2
6.2
6.2
6.2
7.4
2.0
3.0
3.9
5.0
6.0
11.3
13.1
14.5
15.3
3
3
a
b
7800
7800
65.0
80.0
210
210
31.7
39.0
8.1
10.0
8.1
10.0
17.1
21.0
3.9
3.9
3
3
.1
.2
11700
23400
50.0
50.0
210
210
24.4
24.4
6.3
6.3
6.3
6.3
13.1
13.1
3.9
3.9
◦
◦
Note: Isothermal experiments 1–9 were carried out at 210 C with CuZA and at 220 C with CuZZ.
samples, CO was shown not to be produced as part of the main reac-
tion pathway of methanol steam reforming, the reverse reaction of
methanol synthesis but through the reverse water gas shift reaction
The catalytic tests were carried out with the goal to obtain con-
versions of less than 15% for both reactants in order to collect data in
the kinetic regime of the reactions. Three experimental data points
do not match this criterion but are still close. As the reactions are
limited by the thermodynamic equilibrium, the approach to equi-
[
28]. In order to calculate TOF values, the total copper surface area
of the catalyst was assumed to be equally active and both reactions
were assumed to occur on the copper surface [29–31]. The assump-
tion of equally active sites can be justified by the experimental fact
that the adsorption sites appear energetically uniform as measured
by adsorption studies on a Cu/ZnO/Al O catalyst [32].
librium ␤ , defined as the ratio of the reaction quotient (calculated
j
with the outlet partial pressures) and the thermodynamic equilib-
rium constant, was added to both kinetic equations in order to take
into account the respective reverse reactions.
2
3
TOFMeOH was calculated using the number of moles of methanol
nMeOH produced during the reaction time t per catalyst mass mcat,
The approach to equilibrium ␤j for the two reactions with
respect to the outlet partial pressures is defined as
Avogadro’s number N , the specific copper surface area of the cat-
alyst SCu and the number of surface copper atoms per unit surface
N .
S
A
◦
(p )
2O,out
2
pMeOH,outpH
ˇMeOH
=
(11)
3
KMeOH
p
p
CO2,out
2,out
H
nCH3OHNA
TOFMeOH
=
(9)
mcatSCuNSt
and
The commonly used literature value for NS is 1.46 × 1019
m
−2
.
pCO,outpH2O,out
ˇCO
=
KCO
(12)
It is obtained by assuming equal proportions of (1 0 0), (1 1 0) and
H2,out
p pCO2,out
(
1 1 1) planes exposed at the copper particle surface [33].
◦
TOFCO was calculated similarly to TOFMeOH
.
with p the standard pressure (1.0 bar). Furthermore, at low conver-
sion, it can be assumed for the reactants that the partial pressure
difference along the catalyst bed is negligible and that therefore
partial pressures of the reactants are homogeneous over the entire
catalyst bed. The reaction orders n and m are then defined with
nCONA
TOFCO
=
(10)
mcatSCuNSt
Expected compositions at thermodynamic equilibrium for the
different experimental conditions from Table 2 were calculated
with ProSim Plus software using the Soave–Redlich–Kwong model
assuming an equilibrated flow reactor by minimizing Gibbs free
j
j
respect to reactant inlet pressures of H2 and CO , respectively.
2
The temperature dependency of the reaction rate is modeled
by an Arrhenius law with an apparent activation energy Ea, j and a
pre-exponential factor k0, j.
energy for CO , CO, H , CH OH and H O, with N and Ar as inerts.
2
2
3
2
2
Experimental results were checked for heat and mass trans-
fer limitations and no such limitations were found. The detailed
calculations are presented in Supplementary material.
Partial pressures were not corrected for non-ideality. Although
slightly more precise results might be obtained, this effect was
judged negligible based on literature results [34]. Severe non-
ideality caused by the apparition of a dew point would not be
expected until exceeding 100 bar total pressure [35].
3
. Theory/Calculation
Based on the aforementioned assumptions, the kinetic equa-
tions were formulated as power-law expressions (Eqs. (13) and
3.1. Power law
(14)).
The power-law expression of the kinetic model is based on sev-
eral assumptions.
Two reactions (index j) are considered: CO2 hydrogenation to
methanol (j = MeOH) and reverse water gas shift reaction (j = CO).
ꢁ
ꢂ
ꢃ
ꢄ
−Ea,MeOH
n
m
MeOH
MeOH
2
TOFMeOH = k0, MeOH exp
p
p
1 − ˇMeOH
H
CO
2
RT
(13)
Please cite this article in press as: K. Kobl, et al., Power-law kinetics of methanol synthesis from carbon dioxide and hydrogen on
copper–zinc oxide catalysts with alumina or zirconia supports, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.11.020

G Model
CATTOD-9911; No. of Pages12
ARTICLE IN PRESS
K. Kobl et al. / Catalysis Today xxx (2015) xxx–xxx
4
4
.3. Thermodynamic calculations
ꢁ
ꢂ
ꢃ
ꢄ
−
Ea,CO
n
m
CO
CO
2
TOFCO = k0,CO exp
p
p
1 − ˇCO
(14)
For different experimental conditions listed in Table 2, Fig. 1
shows the simulated molar composition of the reactor effluent
at thermodynamic equilibrium with the compounds CO , CO, H ,
H
CO
2
RT
For each TOF equation, four parameters (E , n , m , k0,j) have to be
a,j
j
j
2
2
calculated. In these equations, ␤ is not a parameter, as it is entirely
methanol and water (inert gases N2 and Ar are not shown). Table 3
contains the corresponding H2 and CO2 conversions as well as
methanol and CO selectivities.
j
defined by the experimental and thermodynamic data.
The values for the equilibrium constants were calculated by the
equations given in the literature [34].
◦
By rising the temperature from 200 to 240 C (Table 2, rows
A–E), the proportion of reactants H2 and CO2 gets more impor-
tant (Fig. 1a), as CO2 conversion decreases from 40 to 31% and H2
conversion from 30 to 20% (Table 3, rows A–E). At the same time,
methanol selectivity decreases from 96 to 77% because of the high
exothermicity of the methanol synthesis reaction compared to the
endothermic reverse water gas shift reaction.
3
.2. Model parameter calculation
For a same set of inlet partial pressures and for different temper-
atures, the natural logarithm of TOF/(1 − ␤ ) was plotted versus the
j
inverse temperature to determine the apparent activation energy
from the slope of the linearized equation.
For the thermodynamic calculations presented hereafter, the
◦
ꢅ
ꢆ
reaction temperature was fixed to 210 C, temperature at which
TOFj
ꢃ
ꢀ
0
ꢄ
Ea,j
1
T
experiments with CuZA were carried out. As the trends do not
ln
= ln k
−
(15)
,j
◦
1
− ˇj
R
change if the temperature is increased to 220 C (temperature for
experiments with CuZZ), these results are not shown.
At a constant temperature and by varying one of the inlet partial
Increase of the total pressure from 50.0 to 80.0 bar (Table 2,
rows 3, 3a, 3b) increases the proportion of methanol in the efflu-
ent, whereas that of CO is not affected (Fig. 1b). This translates to
an increase in methanol selectivity with pressure from 87 to 94%
(Table 3, rows 3, 3a, 3b), and can be explained by the Le Chate-
lier principle: methanol synthesis is favored by an increase of total
pressure, whereas RWGS is not affected. With the pressure increase,
CO2 conversion rises from 30 to 40% and H2 conversion from 21 to
pressures pCO2 or pH2, partial reaction orders for H2 and CO2 were
obtained from the slope of the linearized equation by linear regres-
sion of natural logarithmic plots of TOF/(1 − ␤ ) versus the natural
j
logarithm of the respective inlet pressures of H2 and CO .
2
ꢅ
ꢆ
TOFj
ꢃ
ꢄ
ꢄ
ꢃ
ꢄ
ꢀꢀ
ln
ln
= ln k
+ n ln pH
(16)
0
,i
j
2
1
− ˇj
ꢅ
ꢆ
3
0%.
When the hydrogen inlet partial pressure is increased from
2.7 to 37.5 bar (Table 2, rows 1–5), the proportion of methanol
TOFj
ꢃ
ꢃ
ꢄ
ꢀꢀꢀ
= ln k
+ m ln p
(17)
0
,j
j
CO2
1
− ˇj
1
ꢀ
ꢀꢀꢀ
increases, too (Fig. 1c). CO2 conversion increases from 15 to 50%
and H2 conversion from 17 to 23%. Methanol selectivity drastically
increases from 65 to 94% (Table 3, rows 1–5).
Changes in CO2 pressure have a less notable effect. When the
CO2 partial pressure is increased from 4.1 to 12.0 bar (Table 2,
rows 3, 6–9), the proportion of the other compounds changes only
slightly (Fig. 1d). CO2 conversion decreases from 37 to 21%, but H2
conversion increases from 17 to 28%. Methanol selectivity is almost
not affected and decreases slightly from 87.5 to 85.6% (Table 3, rows
The pseudo-constants k to k
were discarded and the pre-
0
,j
0,i
exponential factor was obtained by linear regression of all data
points through Eqs. (13) and (14).
4
. Results and discussion
4.1. Catalyst and bed characteristics
The catalyst and bed characteristics are shown in Table 1. The
3
, 6–9).
To summarize, CO2 hydrogenation into methanol is favored by
catalysts used for this study were prepared by coprecipitation from
nitrate solutions of the respective metals [11]. Both catalysts con-
tain 37.5 wt% CuO, 41.0 wt% ZnO and 21.5 wt% of a supporting oxide.
In the case of CuZA, the support is Al O , whereas for CuZZ, ZrO is
used. Similar BET surface areas (82 m g for CuZA and 71 m g
low reaction temperatures, high total pressure and high hydrogen
inlet partial pressure.
2
3
2
−1
2
−1
2
2
−1
for CuZZ) as well as Cu metal surface areas (9.6 m g for CuZA and
8
4.4. Catalytic tests
2
−1
.3 m g for CuZZ) were obtained.
Experimental results for the CuZA catalyst are shown in Table 4.
The experiments and data treatment for the CuZZ catalyst follow
the same logic than for the CuZA catalyst. Experimental results for
CuZZ are shown in Table 5.
4
.2. Choice of experimental conditions
The reaction conditions that were used for the thermodynamic
calculations and the catalyst tests are shown in Table 2. They can
be subdivided into four series. For series A–E, the reaction tem-
perature was varied between 200 and 240 C at a constant total
4
.4.1. Temperature effect
Reaction temperature was varied in order to determine the
◦
pressure of 50 bar and a constant H /CO ratio of 3.9. For series
2
2
temperature at which subsequent isothermal experiments were
to be conducted. As the thermodynamic limitation is far from
being reached, the increase of reaction temperature has a strong
positive effect on CO2 and H2 conversions over CuZA between
2
1
3
–5, the hydrogen partial pressure was varied between 12.7 and
7.5 bar at constant total pressure, CO2 pressure and temperature.
◦
◦
The temperature was set to 210 C for CuZA catalyst and to 220 C
for CuZZ catalyst. For series 6–9, the carbon dioxide partial pres-
sure was varied between 4.1 and 12.0 bar at constant total pressure,
◦
00 and 230 C, with a less pronounced effect between 230 and
◦
2
40 C (Table 4, A–E). This is opposite to the thermodynamic trend
H pressure and temperature. The respective “missing” component
◦
2
where conversions decrease with temperature. At 240 C, experi-
mental CO conversion exceeds 50% of the equilibrium conversion.
Methanol selectivity decreases from 56 to 36% between 200 and
240 C. The activating effect of temperature is also observed over
was compensated by argon to keep the total pressure constant. For
series 3–3b, the total pressure was varied at a constant tempera-
2
ture and an H /CO ratio of 3.9 with CuZA. Finally, two points 3.1
◦
2
2
and 3.2 at higher GHSV were acquired with CuZA.
CuZZ (Table 5, A–E). Along with the increase of conversions with
Please cite this article in press as: K. Kobl, et al., Power-law kinetics of methanol synthesis from carbon dioxide and hydrogen on
copper–zinc oxide catalysts with alumina or zirconia supports, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.11.020

G Model
CATTOD-9911; No. of Pages12
ARTICLE IN PRESS
K. Kobl et al. / Catalysis Today xxx (2015) xxx–xxx
5
a
b
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
2
00 210 220 230 240
10
20
30
40
p(H ) (bar)
T (°C)
2
H2
CO2
H O
2
MeOH
CO
c 0
0
.7
d 0.7
0.6
.6
.5
.4
.3
.2
.1
.0
0
0.5
0
0.4
0
0.3
0
0.2
0
0
.1
.0
0
0
4
6
8
10 12
50
60
70
80
p(CO ) (bar)
2
p (bar)
Fig. 1. Simulated gas phase composition at the reactor exit at thermodynamic equilibrium as a function of (a) temperature (conditions A–E), (b) H2 inlet pressure (conditions
–5), (c) CO2 inlet pressure (conditions 6–9), (d) total pressure (conditions 3, 3a, 3b).
1
Table 3
Conversions (X) and selectivities (S) at thermodynamic equilibrium.
X (%)
S (%)
Conditions
CO2
H2
MeOH
CO
A
B
C
D
E
40.9
37.7
34.9
32.6
30.7
30.5
27.7
25.0
22.4
20.1
95.6
93.0
89.3
84.2
77.4
4.4
7.0
10.7
15.8
22.6
1
2
3
4
5
15.0
22.5
30.4
40.2
48.8
17.0
19.6
21.2
22.5
23.2
64.6
79.7
86.9
91.8
94.3
35.4
20.3
13.1
8.2
5.7
6
7
3
8
9
21.2
26.3
30.2
34.2
37.2
28.4
24.2
21.5
19.1
17.3
85.6
86.4
86.9
87.3
87.5
14.4
13.6
13.1
12.7
12.5
3
3
3
30.1
35.5
40.3
21.2
25.8
29.9
86.9
91.5
94.1
13.1
8.5
5.9
a
b
Please cite this article in press as: K. Kobl, et al., Power-law kinetics of methanol synthesis from carbon dioxide and hydrogen on
copper–zinc oxide catalysts with alumina or zirconia supports, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.11.020

G Model
CATTOD-9911; No. of Pages12
ARTICLE IN PRESS
K. Kobl et al. / Catalysis Today xxx (2015) xxx–xxx
6
Table 4
Results of catalytic tests with CuZA.
Conditions
Conversion (%)
Selectivity (%)
MeOH
103 TOF (s−1
MeOH
)
␤
CO2
H2
CO
CO
MeOH
CO
Reaction time (h)
Mass of liquid (g)
A
B
C
D
E
4.6
7.7
11.1
15.8
17.2
2.4
3.4
4.8
6.3
6.6
56
45
43
38
36
44
55
57
62
64
4.1
5.4
7.5
9.3
9.6
3.3
6.7
9.9
15.3
17.3
<0.01
0.01
0.03
0.09
0.13
0.06
0.17
0.32
0.61
0.64
168
120
113
50
4.8
4.9
6.4
3.7
3.9
51
1
2
3
4
5
5.1
5.9
7.2
9.3
9.9
2.9
3.1
3.1
3.2
3.4
20
33
39
42
48
80
67
61
58
52
1.1
2.2
3.1
4.3
5.1
4.5
4.4
4.8
5.9
5.7
0.03
0.02
0.02
0.02
0.01
0.20
0.16
0.16
0.21
0.18
358
215
188
217
167
4.7
4.3
4.9
7.3
7.3
6
7
8
9
4.8
6.1
8.8
9.8
3.9
3.3
2.9
2.7
37
38
38
40
63
62
62
60
3.7
3.4
2.9
2.8
6.4
5.4
4.7
4.2
0.01
0.02
0.02
0.02
0.14
0.15
0.20
0.20
190
189
195
237
6.1
5.2
4.7
5.4
3
3
a
b
8.6
9.9
4.4
5.3
48
51
52
49
4.5
5.6
5.0
5.4
0.02
0.02
0.21
0.27
137
140
5.0
6.1
3
3
.1
.2
5.6
2.8
2.6
1.4
39
46
61
54
3.6
4.2
5.6
5.0
0.01
<0.01
0.10
0.02
693
501
15.0
5.8
Table 5
Results of catalytic tests with CuZZ.
Conditions
Conversion (%)
Selectivity (%)
MeOH
103 TOF (s−1
MeOH
)
␤
CO2
H2
CO
CO
MeOH
CO
Reaction time (h)
Mass of liquid (g)
A
B
C
D
E
3.9
5.7
8.7
11.8
15.3
2.5
3.5
4.7
6.0
6.6
70
63
56
47
38
30
37
44
53
62
4.9
6.5
8.9
10.1
10.4
2.1
3.9
6.9
11.3
17.2
<0.01
0.01
0.03
0.06
0.11
0.03
0.06
0.14
0.28
0.48
145
135
145
111
110
4.3
5.5
8.0
7.9
8.7
2
4
5
6.1
8.4
10.0
3.5
3.4
3.7
36
49
55
64
51
45
2.8
5.3
7.0
5.0
5.4
5.7
0.04
0.02
0.02
0.13
0.12
0.13
241
121
96
5.5
4.4
4.5
6
7
8
4.9
6.5
9.4
4.0
3.9
3.5
37
39
40
63
61
60
4.4
4.2
3.7
7.6
6.6
5.6
0.02
0.03
0.03
0.12
0.14
0.18
119
119
167
4.0
3.9
4.7
temperature, methanol selectivity decreases from as high as 70%
down to 38%.
Comparing CuZA and CuZZ at the same temperatures, CO2 con-
versions for CuZZ are lower than for CuZA, whereas H conversions
2
are the same for both catalysts. Methanol selectivities obtained
with CuZZ are better than with CuZA over the entire temperature
range. However, this difference is most pronounced at low tem-
◦
perature (210 C) where the selectivity is 63% for CuZZ versus 45%
for CuZA. The selectivity difference is reflected by the turnover fre-
quencies (Fig. 2). TOFMeOH of the CuZZ catalyst is slightly higher
than in the case of CuZA but TOFCO of the CuZZ catalyst is only
two thirds than TOFCO of CuZA. Thus, although CuZZ has compara-
ble methanol synthesis activity to CuZA, it is more selective, as the
RWGS reaction proceeds more slowly on CuZZ compared to CuZA.
4
.4.2. Effect of hydrogen pressure
The increase of H2 partial pressure from 12.7 to 37.5 bar at con-
Fig. 2. Turnover frequencies for methanol synthesis (MeOH, ordinate) and RWGS
reaction (CO, abscissa) of CuZA (red squares) and CuZZ (green circles). The dotted
line indicates 50% selectivity. (For interpretation of the references to color in this
figure legend, the reader is referred to the web version of this article.)
stant CO2 pressure doubles CO2 conversion over CuZA from 5 to
0%, whereas H2 conversion remains almost unaffected around 3%
Table 4, 1–5). This trend is expected from thermodynamic calcu-
1
(
lations although the effect is less drastic than at equilibrium. For
CuZZ, methanol selectivity is more than doubled by the increase of
H2 partial pressure (Table 5, 2–5). Again, the increase of H2 partial
pressure favors CO2 conversion, whereas H2 conversion remains
stable. In parallel, methanol selectivity increases from 36 to 55%.
4
.4.3. Effect of carbon dioxide pressure
The effect of CO2 partial pressure is less pronounced than that
of H . As CO2 partial pressure is increased from 4.1 to 12.0 bar, CO2
conversion over CuZA is halved, whereas H2 conversion slightly
increases (Table 4, 6–9). This trend is also expected at thermody-
2
Please cite this article in press as: K. Kobl, et al., Power-law kinetics of methanol synthesis from carbon dioxide and hydrogen on
copper–zinc oxide catalysts with alumina or zirconia supports, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.11.020

G Model
CATTOD-9911; No. of Pages12
ARTICLE IN PRESS
K. Kobl et al. / Catalysis Today xxx (2015) xxx–xxx
7
namic equilibrium. Methanol selectivity remains unchanged in the
kinetic regime as well as at equilibrium. Similar observations are
a
T (°C)
220
2
40
230
210
CO
200
made with CuZZ, when the CO partial pressure is changed (Table 5,
2
-
-
-
-
-
2.5
3.0
3.5
4.0
4.5
6
–8).
-
1
E = 156 ± 12 ꢂk ꢃmo
R = 0.982
40
30
4.4.4. Effect of total pressure
a
2
Further experiments were carried out with CuZA. The increase
2
0
of total pressure increases CO2 and H2 conversions, but the effect
is less pronounced than the predictions at equilibrium (Table 4, 3a
and b). In contrast, in terms of methanol selectivity, the pressure
rise has a higher effect in the kinetic regime than at thermodynamic
equilibrium.
10
5
-5.0
-5.5
MeOH
-
1
E = 61.1 ± 1.4 ꢂk ꢃmo
a
-
-
6.0
6.5
2.5
2
R = 0.998
4.4.5. Effect of GHSV
As can be expected, CO and H conversions over CuZA decrease
1.95
2.00
-12.05
2.10
2.15
2
2
3
-1
as the GHSV is increased, resulting in increasing distance from ther-
modynamic conversion (Table 4, 3.1 and 3.2). On the other hand,
the higher the GHSV is, the higher is methanol selectivity, which
rises from 38 to 46%.
1
0 T (K )
b
p(H ) (bar)
2
1
2
15
20
25
30 35 40
4
.4.6. Approach to equilibrium
-
4.5
5.0
5.5
6.0
6.5
7.0
The approach to equilibrium ␤ is different for methanol syn-
j
thesis and for RWGS reaction. For CuZA (Table 4), ␤MeOH is small
for most data points, the maximum being at 0.13. In contrast, ␤CO
can reach values up to 0.64. These high values are observed at high
temperatures and pressures indicating that reverse reactions rates
cannot be negligible in the whole range of tested conditions.
For almost all experimental points, TOFCO is higher than
TOFMeOH. This means that under identical experimental conditions,
RWGS is faster than methanol synthesis and at the same time, the
RWGS reaction is closer to equilibrium than the methanol synthesis
reaction.
8
6
-
-
-
-
-
4
2
1
CO
n2= 0.2ꢀ ± 0.1ꢁ
R = 0.461
MeOH
n2= 1.40 ± 0.0ꢀ
R = 0.989
For CuZZ (Table 5), the approach to equilibrium reaches up to
.11 in the case of methanol synthesis and up to 0.48 in the case of
2.4
2.6
2.8
3.0
3.2
3.4
3.6
3.8
0
ln(p(H ) /bar)
2
RWGS. As observed with CuZA, RWGS is closer to equilibrium than
methanol synthesis. However, ␤ values are higher for CuZA than
j
c
p(CO ) (bar)
for CuZZ although the partial pressure variation was carried out at
2
◦
◦
4
6
8
10
12
a higher temperature of 220 C with CuZZ, as compared to 210 C
with CuZA.
-4.5
CO
m = 0.ꢁꢀ ± 0.04
R = 0.96ꢁ
4.4.7. Summary
2
8
6
-
-
-
5.0
5.5
6.0
Altogether, concerning partial pressure variation, CuZZ shows
trends which are in accordance with those at thermodynamic equi-
librium and observed with CuZA. In contrast, there is a difference in
performance as a function of temperature. If the methanol synthe-
4
◦
sis reaction on CuZZ is faster than the RWGS reaction up to 220 C,
MeOH
◦
this is only true at 200 C on CuZA. Inversely, this means that CuZZ is
m = 0.26 ± 0.02
2
better suited for methanol synthesis than CuZA in the temperature
R = 0.981
2
◦
range of 200–220 C.
1.5
2.0
ln(p(CO ) /bar)
2.5
4.5. Kinetic models
2
The impact of mass and heat transfer limitations on the experi-
Fig. 3. Parameter regression of the kinetic model of CuZA for the methanol synthesis
(full circles, red line) and the RWGS reaction (open circles, blue line): (a) apparent
activation energy, (b) reaction order with respect to H2, (c) reaction order with
respect to CO2. (For interpretation of the references to color in this figure legend,
the reader is referred to the web version of this article.)
mental data was checked and deemed to be negligible, as shown in
detail in Supplementary material. The inclusion of the approach
to equilibrium in the kinetic equations turned out to be neces-
sary, as the reverse reactions are not negligible, especially as far
as the RWGS reaction is concerned, so that the reaction rates of the
◦
direct reactions could be calculated using ␤ . Preliminary calcula-
the data points at 240 C are clearly not aligned with the points at
j
tions without ␤ , which are not presented here, showed inferior
model correlation with experimental data.
lower temperature. If heat transfer limitations were at the origin of
the misalignment, the real temperature would be higher than the
abscissa value and the obtained TOF would be higher as expected.
This is not the case. However, at this temperature, the reaction is
getting closer to equilibrium, but in our simplified approach the
composition was assumed to be uniform throughout the reactor.
Moreover, the Arrhenius equation is only valid in a limited tem-
j
4.5.1. Apparent activation energies
For CuZA, the apparent activation energies for the two parallel
reactions methanol synthesis and RWGS were determined from the
Arrhenius plot shown in Fig. 3a. Taking into account the error bar,
Please cite this article in press as: K. Kobl, et al., Power-law kinetics of methanol synthesis from carbon dioxide and hydrogen on
copper–zinc oxide catalysts with alumina or zirconia supports, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.11.020

G Model
CATTOD-9911; No. of Pages12
ARTICLE IN PRESS
K. Kobl et al. / Catalysis Today xxx (2015) xxx–xxx
8
synthesis, a linear correlation in the Arrhenius plot is only obtained
between 200 and 230 C. The point at 240 C is at high conver-
sion. In contrast, the correlation is linear for RWGS over the entire
a
T (°C)
220
◦
◦
2
40
230
210
CO
200
◦
-
-
-
-
-
-
-
-
3.0
3.5
4.0
4.5
5.0
5.5
6.0
6.5
temperature range from 200 to 240 C. In both cases, a good cor-
4
0
2
relation is obtained (R ≥ 0.98). The apparent activation energies
-1
30
20
−
1
−1
E = 1ꢁꢀ ± 4 ꢂk ꢃmo
are E
= 53.6 ± 4.5 kJ mol and E
= 137 ± 4 kJ mol . As on
a, MeOH
a, CO
a
2
CuZA, the apparent activation energy for methanol synthesis is less
than half the value obtained for the RWGS reaction.
R = 0.996
1
5
0
4.5.2. Partial reaction orders with respect to hydrogen
MeOH
For CuZA, partial reaction orders with respect to H2 were
obtained from the slopes in the double logarithmic plot in Fig. 3b.
-1
E = 5ꢁ.6 ± 4.5 ꢂk ꢃmo
a
2.5
2
Concerning methanol synthesis, a good correlation is obtained
R = 0.9ꢀ9
2
(
R > 0.98). H2 partial reaction order for methanol synthesis is
1
.40 ± 0.07. The case of the RWGS reaction is more delicate. As the
1.95
2.00
2.05
2.10
2.15
experimental error for this rate is higher and its slope smaller than
for methanol synthesis, the experimental data points are farer away
from the straight line. Including all 5 data points the correlation
coefficient R2 is as low as 0.46. This value may seem too low, but
as it turns out, the final model gives a satisfactory representation
of the experimental data. The partial reaction order with respect to
H2 for RWGS reaction is 0.27 ± 0.13.
3
-1
-1
1
0 T (K )
b
p(H ) (bar)
2
1
2
15
20
25
30 35 40
-
-
-
-
-
-
-
4.8
5.0
5.2
5.4
5.6
5.8
6.0
8
7
In order to determine the reaction orders with respect to
hydrogen for CuZZ, H2 partial pressure was varied between 18.8
and 37.5 bar at constant CO2 pressure and constant temperature
6
5
(Fig. 4b). By regression through all three data points on the double-
CO
logarithmic plot, partial reaction orders of 1.29 ± 0.08 for methanol
synthesis and 0.18 ± 0.05 for RWGS reaction were obtained. The
correlation coefficient R2 of 0.86 for RWGS reaction is lower than
n2= 0.18 ± 0.05
R = 0.859
4
3
MeOH
2
n = 1.29 ± 0.08
for methanol synthesis (R > 0.99).
2
R = 0.99ꢁ
4.5.3. Partial reaction orders with respect to carbon dioxide
In the same way as for hydrogen, partial reaction orders with
respect to CO2 for CuZA are obtained from the double-logarithmic
plot in Fig. 3c. For methanol synthesis, a good correlation is obtained
R > 0.98) and the partial reaction order with respect of CO2 is
2
.4
2.6
2.8
3.0
3.2
3.4
10
3.6
3.8
ln(p(H ) /bar)
2
2
(
obtained as 0.26 ± 0.02. For RWGS, the higher experimental error
c
4
6
8
12
causes a larger distribution of the data points which were all used
2
for regression. A satisfactory correlation is obtained (R = 0.96) and
CO
-4.5
-5.0
-5.5
-6.0
m = 0.25 ± 0.01
R = 0.995
the partial reaction order with respect of CO2 for the RWGS reaction
is determined as 0.37 ± 0.04.
2
8
6
Similarly, for the determination of reaction orders with respect
to carbon dioxide for CuZZ, CO2 partial pressure was varied
between 4.9 and 12.0 bar at constant H2 pressure and constant
temperature (Fig. 4c). Using all three data points in each case for
regression on the double-logarithmic plot, partial reaction orders of
4
MeOH
m2 = 0.19 ± 0.02
R = 0.968
0
.19 ± 0.02 for methanol synthesis and 0.25 ± 0.01 for RWGS reac-
tion were calculated. This time, the correlation is better for RWGS
2
2
(
R > 0.99) than for methanol synthesis (R = 0.97).
1
.5
2.0
ln(p(CO ) /bar)
2.5
2
4.5.4. Preexponential factors
In order to finalize the model for CuZA, the pre-exponential fac-
tors k0, MeOH and k0, CO have to be determined. They were obtained
from the slopes of the plots in Fig. 5, correlating all experimental
Fig. 4. Parameter regression of the kinetic model of CuZZ for the methanol synthesis
full circles, red line) and the RWGS reaction (open circles, blue line): (a) apparent
activation energy, (b) reaction order with respect to H2, (c) reaction order with
respect to CO2. (For interpretation of the references to color in this figure legend,
the reader is referred to the web version of this article.)
(
ꢃ
ꢄ
ꢃ
ꢄ
n
m
j
data points to the products exp Ea,j/RT
p
j
p
1 − ˇj calcu-
H
CO
2
2
lated from experimental conditions and already determined model
parameters. For both reactions a very good correlation is obtained,
2
with R ≥ 0.99 for methanol synthesis and RWGS reaction.
perature range and a change in temperature can induce changes
in apparent activation energy as a consequence of surface cover-
The pre-exponential factors of the reaction rate equations for
CuZZ were obtained from the plot in Fig. 6. In both cases, a very
◦
age [36]. In any case, the data points at 240 C were excluded from
2
2
good correlation with R > 0.99 is obtained.
linear regression. The correlation then obtained is good (R ≥ 0.98).
From the slopes, an E
of 61.1 ± 1.4 kJ mol− and an E
1
of
a,CO
a, MeOH
−1
1
56 ± 12 kJ mol are obtained.
4.5.5. Parity plots
The global correlation between the power law model and the
experimental data for CuZA is illustrated by the parity plots in Fig. 7.
In the first measurement series with CuZZ, the reaction temper-
ature was varied between 200 and 240 C (Fig. 4a). For methanol
◦
Please cite this article in press as: K. Kobl, et al., Power-law kinetics of methanol synthesis from carbon dioxide and hydrogen on
copper–zinc oxide catalysts with alumina or zirconia supports, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.11.020

G Model
CATTOD-9911; No. of Pages12
ARTICLE IN PRESS
K. Kobl et al. / Catalysis Today xxx (2015) xxx–xxx
9
a
a
1
0
8
6
4
2
0
12
MeOH
MeOH
-
1.66 -1
s
-
1.48 -1
s
k = 90.4 ± 0.8 bar
0
2
k = 28.1 ± 0.4 bar
1
0
8
6
4
2
0
0
2
R = 0.999
R = 0.998
0
4
.0
0.5
nMeOH
1.0
1.5
0
1
2
3
4
nMeOH+mMeOH
mMeOH
nMeOH+mMeOH
4
nMeOH
mMeOH
1
0 ·exp(-Ea,MeOH/RT)·p(H2)
·p(CO2)
·(1-βMeOH) (bar
)
10 ·exp(-Ea,MeOH/RT)·p(H2)
·p(CO2)
·(1-βMeOH) (bar
)
b
2
1
1
0
b
CO
11
-0.43 -1
CO
k = (7.8 ± 0.2) 10 bar
s
0
2
1
5
0
5
0
13
-0.64 -1
k = (9.8 ± 0.3) 10 bar
s
0
2
R = 0.992
5
0
5
0
R = 0.989
1
0
.0
1
0.5
1.0nCO
1.5
2.0
2.5
nCO+mCO
0
.0
0.5
1.0
1.5
2.0
14
mCO
0 ·exp(-Ea,CO/RT)·p(H2) ·p(CO2) ·(1-βCO) (bar
)
1
5
nCO
mCO
nCO+mCO
1
0 ·exp(-Ea,CO/RT)·p(H2) ·p(CO2) ·(1-βCO) (bar
)
Fig. 6. Determination of preexponential factors of the kinetic model of CuZZ for (a)
the methanol synthesis and (b) the RWGS reaction.
Fig. 5. Determination of preexponential factors of the kinetic model of CuZA for (a)
the methanol synthesis and (b) the RWGS reaction.
The parity plots for methanol synthesis and RWGS on CuZZ are
shown in Fig. 8, respectively. For methanol synthesis, all experi-
mental points are described by the model within a deviation well
It can be seen that most experimental data points are described by
the model with a relative error smaller than 10%, with the exception
of the point E at 240 C. Deviation of the point E can be explained by
◦
◦
below the 10% error range except the point E at 240 C, which can
the CO conversion (17.2%) which may be too important to assume
be explained by the same reasons given for the CuZA catalyst. As a
whole, there is also a good correlation for RWGS.
2
a constant CO partial pressure in the whole bed. As a consequence,
2
for this condition, TOFMeOH and TOFCO predicted by the model are
slightly higher than the ones measured experimentally.
Generally, the correlation for methanol synthesis is better than
for RWGS reaction, as TOFMeOH for all conditions cover a wide
a 1
4
b
30
+
10%
+10%
E
1
2
0
8
6
4
2
0
E
25
20
15
10
5
0
-
10%
-10%
1
0
2
4
6
8
10 12 14
0
5
10
15
20
25
-1
30
3
-1
3
experimental 10 TOF(MeOH) (s )
experimental 10 TOF(CO) (s
)
Fig. 7. Parity plots of the kinetic model of CuZA for (a) the methanol synthesis and (b) the RWGS reaction.
Please cite this article in press as: K. Kobl, et al., Power-law kinetics of methanol synthesis from carbon dioxide and hydrogen on
copper–zinc oxide catalysts with alumina or zirconia supports, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.11.020

G Model
CATTOD-9911; No. of Pages12
ARTICLE IN PRESS
1
0
K. Kobl et al. / Catalysis Today xxx (2015) xxx–xxx
a
b
16
14
12
10
8
6
4
2
0
2
0
+10%
+10%
E
-
10%
15
-
10%
1
0
5
0
0
2
4
6
8
10 12 14 16
0
5
10
15
20
3
-1
3
-1
)
experimental 10 TOF(MeOH) (s )
experimental 10 TOF(CO) (s
Fig. 8. Parity plots of the kinetic model of CuZZ for (a) the methanol synthesis and (b) the RWGS reaction.
range of values. The isothermal TOF
reaction are cumulated around 5 × 10
data points of the RWGS
and 32 to 50 bar with feeds containing both CO and CO [8–41]. The
CO
2
−
3
−1
−3
−2 −1
s measured here are higher than
s
. When the other exper-
TOFCO of 3 × 10 to 1.7 × 10
imental conditions are varied at isothermal conditions, the RWGS
reaction rate changes little in contrast to the methanol reaction
rate, as RWGS total reaction order is lower than in the case of
methanol synthesis. On the other hand, the RWGS reaction is
very temperature-dependent because of its comparatively high
activation energy and TOFCO values as a function of temperature
therefore span a wider range than TOFMeOH, taking into account
the temperature-dependent experiments, too.
in the literature and can be explained by the absence of CO in the
feed.
The apparent activation energy of methanol synthesis on CuZA is
−
1
in the lower range of literature values. With 61 kJ mol , it is higher
−
1
than the value of 43 kJ mol given by Ledakowicz et al. for their
−
1
Cu/ZnO/Al O3 catalyst but only half of the value of 113 kJ mol
2
found by Peter et al. [23,24]. These differences can be explained
by the mathematical expression of the kinetic rate equations. As
Ledakowicz et al. do not account for the approach to equilibrium
in their rate equation, the resulting apparent activation energy is
lower than in this work due to exothermicity of the reaction. On
4.5.6. Summary
The model parameters are summarized in Table 6. For the CuZA
the other hand, Peter et al. not only included ␤ , but also accounted
catalyst, the average deviation of the model from experimental data
is 5% for methanol synthesis and 10% for RWGS reaction. For the
CuZZ catalyst, the average deviation between model and experi-
ment is 3% for methanol synthesis and 8% for RWGS.
j
for product inhibition by water, which results in a higher apparent
activation energy for the direct reaction. Furthermore, differences
in catalyst preparation and composition may have an influence on
the numeric results.
Some general trends are observed. In both cases, the appar-
ent activation energy of RWGS is much higher than for methanol
synthesis. The apparent activation energies are higher for CuZA
compared to CuZZ. The influence of the H2 partial pressure on
methanol production rate is much higher than CO2 partial pres-
sure for both catalysts. The higher the H2 pressure, the faster the
methanol synthesis reaction. The reaction orders for methanol syn-
thesis and reverse water gas shift reaction are lower in the case of
CuZZ. In contrast to the methanol synthesis reaction, however, the
RWGS reaction is little dependent on the partial pressure of the
reactants.
The partial reaction order with respect to H2 for methanol syn-
thesis on CuZA of 1.40 is higher than the literature values of 1.17
from Ledakowicz et al. and 1.25 from Peter et al. On the contrary,
the partial reaction order with respect to CO2 of 0.26 is compa-
rable to the value in Ref. [24], whereas it is considerably smaller
than the value of 0.55 found by Peter et al. Lower values than those
reported in Ref. [24] are in good agreement with the fact that the
approach to equilibrium was not taken into account by the authors.
Moreover, partial reaction orders can be influenced by the range of
experimental conditions. As the model of Peter et al. [40] is based
on a wide range of total pressures going down to as low as 5 bar,
differences in the numerical values of partial reaction orders were
expected.
4.6. Comparison with literature
As Cu/ZnO/Al O3 are predominantly used for industrial
2
Although the power-law models cannot explain the observed
methanol synthesis, few kinetic data for catalysts with different
composition like Cu/ZnO/ZrO2 are available in the open litera-
ture. For a Cu/ZnO/Al O /ZrO catalyst, Ledakowicz et al. found
high dependency of the methanol synthesis reaction on hydrogen
partial pressure, it can be related to the findings from the litera-
ture. The mechanism of the methanol synthesis reaction on copper
catalysts is supposed to include several successive hydrogenation
steps, whereas for the RWGS reaction less hydrogenation steps or a
redox-mechanism are suggested [37–39]. The low reaction order
with respect to CO2 for methanol synthesis and RWGS may be
indicative of strong adsorption of this compound on the catalyst
surface.
2
3
2
−
1
an apparent activation energy of 61 kJ mol
for methanol syn-
thesis, a reaction order of 1.71 with respect to H2 and 0.94 with
respect to CO2 [24]. The apparent activation energy of this catalyst
−
1
is higher than the value of 54 kJ mol for the CuZZ catalyst, whereas
partial reaction orders are much higher, especially for CO . This dif-
2
ference may originate from the catalyst properties (simultaneous
presence of Al O3 and ZrO2 as compared to ZrO -only containing
2
2
The methanol turnover frequencies observed over CuZA and
CuZZ) and from different experimental conditions like the maximal
pressure of 30 bar and the presence of CO in the feed. For methanol
−
3
−2 −1
at 50 bar and
CuZZ are situated in the range of 10 to 10
s
◦
between 200 and 240 C. They were calculated as in the literature
by taking into account the total copper surface area of each catalyst
and supposing all copper sites equally active. The obtained values
are comparable to published values of 4 × 10 to 10
synthesis on a Cu/ZnO/ZrO catalyst at a total pressure of 30 bar
2
−
1
and a GHSV of 10,000 h , Bonura et al. give an apparent activa-
−
1
tion energy of 57 kJ mol which compares favorably to our value
−3
−2 −1 ◦
at 240 C
s
Please cite this article in press as: K. Kobl, et al., Power-law kinetics of methanol synthesis from carbon dioxide and hydrogen on
copper–zinc oxide catalysts with alumina or zirconia supports, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.11.020

G Model
CATTOD-9911; No. of Pages12
ARTICLE IN PRESS
K. Kobl et al. / Catalysis Today xxx (2015) xxx–xxx
11
Table 6
Calculated model parameters.
Catalyst
CuZA
Parameter
Unit
MeOH
Value
CO
sa
Value
sa
kJ mol 1
−
61.1
1.40
0.26
90.4
5
1.5
0.07
0.02
0.8
–
156
12
Ea
n
–
–
bar
%
0.27
0.37
9.8 × 10
10
0.13
0.04
0.3 × 10
–
m
k0
ꢀ
−
(n + m) −1
13
13
s
b
b
rel
kJ mol 1
−
53.6
1.29
0.19
28.1
3
4.5
0.08
0.02
0.4
–
137
4
CuZZ
Ea
n
–
–
bar
%
0.18
0.25
7.8 × 10
8
0.05
0.01
0.2 × 10
–
m
k0
ꢀ
−
(n + m) −1
11
11
s
rel
a
Standard deviation.
Relative mean deviation between model and experiment.
b
[
3
1
42,43]. TOF
values on their catalyst lie in the order of 2 to
at 200 C and are lower than on CuZZ. With a value of
[9] Mitsui Chemicals, Annual report 2010. http://www.mitsuichem.com/ir/
library/ar/pdf/en ar10 all.pdf?1427812726561 (accessed on 29.10.15).
10] Carbon Recycling International Website, http://www.carbonrecycling.is
(accessed on 29.10.15).
MeOH
−
3
−1
◦
× 10
s
[
.5 for H and 0.5 for CO , partial reaction orders only differ slightly
2
2
from those of CuZZ.
. Conclusion
Two power-law models for methanol synthesis from CO /H on
[11] L. Angelo, K. Kobl, L.M.M. Tejada, Y. Zimmermann, K. Parkhomenko, A.-C.
Roger, Study of CuZnMOx oxides (M Al, Zr, Ce, CeZr) for the catalytic
hydrogenation of CO2 into methanol, Comptes Rendus Chim. 18 (2015)
250–260.
5
[
12] S. Arab, J.-M. Commenge, J.-F. Portha, L. Falk, Methanol synthesis from CO2
and H2 in multi-tubular fixed-bed reactor and multi-tubular reactor filled
with monoliths, Chem. Eng. Res. Des. 92 (2014) 2598–2608.
[13] P. Villa, P. Forzatti, G. Buzzi-Ferraris, G. Garone, I. Pasquon, Synthesis of
alcohols from carbon oxides and hydrogen. 1. Kinetics of the low-pressure
methanol synthesis, Ind. Eng. Chem. Process Des. Dev. 24 (1985) 12–19.
2
2
a Cu/ZnO/Al O and a Cu/ZnO/ZrO catalyst have been created. The
2
3
2
turnover frequencies of the methanol synthesis reaction are of the
same order of magnitude for both catalysts. The main difference
between the catalysts is that the turnover frequency of the reverse
water gas shift reaction is lower on the zirconia supported catalyst,
leading to better methanol selectivity. As a perspective, better cat-
alysts for CO2 hydrogenation of methanol, based on Cu/ZnO/ZrO2,
may be obtained by increasing their intrinsic activity for methanol
synthesis through the optimization of the catalyst preparation con-
ditions.
[
14] J.C.J. Bart, R.P.A. Sneeden, Copper–zinc oxide–alumina methanol catalysts
revisited, Catal. Today 2 (1987) 1–124.
[15] J. Skrzypek, M. Lachowska, M. Grzesik, J. Słoczy n´ ski, P. Nowak,
Thermodynamics and kinetics of low pressure methanol synthesis, Chem.
Eng. J. 58 (1995) 101–108.
[16] K.M. Vanden Bussche, G.F. Froment, A steady-state kinetic model for
methanol synthesis and the water gas shift reaction on a commercial
Cu/ZnO/Al2O3 catalyst, J. Catal. 10 (1996) 1–10.
[
17] G.C. Chinchen, P.J. Denny, D.G. Parker, M.S. Spencer, D.A. Whan, Mechanism of
methanol synthesis from CO2/CO/H2 mixtures over copper/zinc
1
4
oxide/alumina catalysts: use of C-labelled reactants, Appl. Catal. 30 (1987)
33–338.
3
Acknowledgement
[
[
18] G.H. Graaf, E.J. Stamhuis, A.A.C.M. Beenackers, Kinetics of low-pressure
methanol synthesis, Chem. Eng. Sci. 43 (1988) 3185–3195.
19] J. Skrzypek, M. Lachowska, H. Moroz, Kinetics of methanol synthesis over
commercial copper/zinc oxide/alumina catalysts, Chem. Eng. Sci. 46 (1991)
The authors thank ANR VItESSE2 for funding.
2809–2813.
[
20] T. Kubota, I. Hayakawa, H. Mabuse, K. Mori, K. Ushikoshi, T. Watanabe, et al.,
Kinetic study of methanol synthesis from carbon dioxide and hydrogen, Appl.
Organomet. Chem. 15 (2001) 121–126.
Appendix A. Supplementary data
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.cattod.2015.11.
[21] J.B. Butt, Reaction Kinetics and Reactor Design, 2nd ed., Marcel Dekker, New
York, 2000.
[
22] M. Behrens, Heterogeneous catalysis of CO2 conversion to methanol on
copper surfaces, Angew. Chem. Int. Ed. 53 (2014) 12022–12024.
23] M. Peter, M.B. Fichtl, H. Ruland, S. Kaluza, M. Muhler, O. Hinrichsen, Detailed
kinetic modeling of methanol synthesis over a ternary copper catalyst, Chem.
Eng. J. 203 (2012) 480–491.
020.
[
References
[
24] S. Ledakowicz, L. Nowicki, J. Petera, J. Nizioł, P. Kowalik, A. Goł e˛ biowski,
Kinetic characterisation of catalysts for methanol synthesis, Chem. Process
Eng. 34 (2013) 497–506.
[25] G.C. Chinchen, C.M. Hay, H.D. Vandervell, K.C. Waugh, The measurement of
copper surface areas by reactive frontal chromatography, J. Catal. 86 (1987)
79–86.
[26] E.L. Kunkes, F. Studt, F. Abild-Pedersen, R. Schlögl, M. Behrens, Hydrogenation
of CO2 to methanol and CO on Cu/ZnO/Al2O3: is there a common intermediate
or not? J. Catal. 328 (2015) 43–48.
[27] F. Studt, M. Behrens, E.L. Kunkes, N. Thomas, S. Zander, A. Tarasov, et al., The
mechanism of CO and CO2 hydrogenation to methanol over Cu-based
catalysts, ChemCatChem 7 (2015) 1105–1111.
[
1] IEA Energy Technology Perspectives, Pathways to a Clean Energy System,
OECD, Publishing, Paris, 2012, http://dx.doi.org/10.1787/energy tech-2012-
en.
[
2] IEA Statistics, CO2 Emissions From Fuel Combustion Highlights 2014 edition.
https://www.iea.org/publications/freepublications/publication/
CO2EmissionsFromFuelCombustionHighlights2014.pdf (accessed 29.10.15).
3] BP Energy Outlook 2035 (2015) http://www.bp.com/energyoutlook (accessed
on 29.10.15).
[
[
4] D. Mignard, M. Sahibzada, J.M. Duthie, H.W. Whittington, Methanol synthesis
from flue-gas CO2 and renewable electricity: a feasibility study, Int. J.
Hydrogen Energy 28 (2003) 455–464.
[
5] ANR Project 10-EESI-0006 description, http://www.agence-nationale-
recherche.fr/?Project=ANR-10-EESI-0006 (accessed 29.10.15).
6] N.A.M. Razali, K.T. Lee, S. Bhatia, A.R. Mohamed, Heterogeneous catalysts for
production of chemicals using carbon dioxide as raw material: a review,
Renew. Sustain. Energy Rev. 16 (2012) 4951–4964.
[28] J.P. Breen, F.C. Meunier, J.R.H. Ross, Mechanistic aspects of the steam
reforming of methanol over a CuO/ZnO/ZrO2/Al2O3 catalyst, Chem. Commun.
(1999) 2247–2248.
[29] J. Szanyi, D.W. Goodman, Methanol synthesis on a Cu (1 0 0) catalyst, Catal.
Lett. 10 (1991) 383–390.
[
[
7] J. Chesko, Methanol Industry Outlook, PROMSUS Workshop 2014,
Gothenburg, Sweden. http://www.marinemethanol.com/publications/
category/5-promsus-workshop (accessed 29.10.15).
[30] J. Yoshihara, C.T. Campbell, Methanol synthesis and reverse water–gas shift
kinetics over Cu (1 1 0) model catalysts: structural sensitivity, J. Catal. 782
(1996) 776–782.
[
8] M. Behrens, F. Studt, I. Kasatkin, S. Kühl, M. Hävecker, F. Abild-Pedersen, et al.,
The active site of methanol synthesis over Cu/ZnO/Al2O3 industrial catalysts,
Science 336 (2012) 893–897.
[31] P.B. Rasmussen, M. Kazuta, I. Chorkendorff, Synthesis of methanol from a
mixture of H2 and CO2 on Cu (1 0 0), Surf. Sci. 318 (1994) 267–280.
Please cite this article in press as: K. Kobl, et al., Power-law kinetics of methanol synthesis from carbon dioxide and hydrogen on
copper–zinc oxide catalysts with alumina or zirconia supports, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.11.020

G Model
CATTOD-9911; No. of Pages12
ARTICLE IN PRESS
K. Kobl et al. / Catalysis Today xxx (2015) xxx–xxx
1
2
[
[
32] J. Sloczynski, R. Grabowski, J. Janas, J. Skrzypek, Adsorption model of
methanol synthesis reactants on Cu–ZnO–Al2O3 catalyst I. Adsorption on the
catalyst, Chem. Eng. Sci. 46 (1991) 2599–2610.
33] J.W. Evans, M.S. Wainwright, A.J. Bridgewater, D.J. Young, On the
determination of copper surface area by reaction with nitrous oxide, Appl.
Catal. 7 (1983) 75–83.
[39] C.V. Ovesen, B.S. Clausen, B.S. Hammershøi, G. Steffensen, T. Askgaard, I.
Chorkendorff, et al., A microkinetic analysis of the water–gas shift reaction
under industrial conditions, J. Catal. 158 (1996) 170–180.
[40] J. Yoshihara, S.C. Parker, A. Schafer, C.T. Campbell, Methanol synthesis and
reverse water–gas shift kinetics over clean polycrystalline copper, Catal. Lett.
31 (1995) 313–324.
[
[
34] G.H. Graaf, P.J.J. Sijtsema, E.J. Stamhuis, G.E.H. Joosten, Chemical equilibria in
methanol synthesis, Chem. Eng. Sci. 41 (1986) 2883–2890.
35] J.G. Van Bennekom, J.G.M.M. Winkelman, R.H. Venderbosch, S.D.G.B. Nieland,
H.J. Heeres, D.G.B. Sebastiaan, Modeling and experimental studies on phase
and chemical equilibria in high-pressure methanol synthesis, Ind. Eng. Chem.
Res. 51 (2012) 12233–12243.
36] H. Lynggaard, A. Andreasen, C. Stegelmann, P. Stoltze, Analysis of simple
kinetic models in heterogeneous catalysis, Prog. Surf. Sci. 77 (2004) 71–137.
37] Z.-M. Hu, K. Takahashi, H. Nakatsuji, Mechanism of the hydrogenation of CO2
to methanol on a Cu (1 0 0) surface: dipped adcluster model study, Surf. Sci.
[41] H. Ahouari, A. Soualah, A. Le Valant, L. Pinard, P. Magnoux, Y. Pouilloux,
Methanol synthesis from CO2 hydrogenation over copper based catalysts,
React. Kinet. Mech. Catal. 110 (2013) 131–145.
[42] F. Arena, G. Mezzatesta, G. Zafarana, G. Trunfio, F. Frusteri, L. Spadaro, Effects
of oxide carriers on surface functionality and process performance of the
Cu–ZnO system in the synthesis of methanol via CO2 hydrogenation, J. Catal.
300 (2013) 141–151.
[43] G. Bonura, M. Cordaro, C. Cannilla, F. Arena, F. Frusteri, The changing nature of
the active site of Cu–Zn–Zr catalysts for the CO2 hydrogenation reaction to
methanol, Appl. Catal. B Environ. 152–153 (2014) 152–161.
[
[
4
42 (1999) 90–106.
[38] L.C. Grabow, M. Mavrikakis, Mechanism of methanol synthesis on Cu through
CO2 and CO hydrogenation, ACS Catal. 1 (2011) 365–384.
Please cite this article in press as: K. Kobl, et al., Power-law kinetics of methanol synthesis from carbon dioxide and hydrogen on
copper–zinc oxide catalysts with alumina or zirconia supports, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.11.020