Applied Catalysis B: Environmental 270 (2020) 118859

Contents lists available at ScienceDirect

Applied Catalysis B: Environmental

journal homepage: www.elsevier.com/locate/apcatb

Dry reforming of methane over the cobalt catalyst: Theoretical insights into T
the reaction kinetics and mechanism for catalyst deactivation
Shuyue Chena,b,c, Jeremie Zaffrana, Bo Yanga,*
a
School of Physical Science and Technology, Shanghai Tech University, Shanghai 201210, China
b
CAS Key Laboratory of Low-Carbon Conversion Science and Engineering, Shanghai Advanced Research Institute, Chinese Academy of Sciences, Shanghai, 201210, China
c
University of Chinese Academy of Sciences, Beijing, 100049, China

A R T I C LE I N FO A B S T R A C T

Keywords: Cobalt shows high catalytic activity for the important dry reforming of methane (DRM) reaction. However, it is
Dry reforming of methane prone to deactivation and the corresponding mechanism remains controversial. In this work, we combined
Cobalt catalyst density functional theory calculations and microkinetic modeling to study the active site and reaction me-
Catalyst deactivation chanism of Co catalyzed DRM reaction, employing face centered cubic Co(111) and Co(211) as models. It was
Reaction kinetics
found that the step site over Co(211) is the active site for the reaction, and on Co(111), the C + O and CH + O
DFT
paths are the preferred reaction pathways, while the C + O path is dominant on Co(211). The dissociation of
CH4 is the rate-controlling step of DRM over both Co(111) and Co(211). We found that Co(111) is mainly
deactivated due to carbon deposition whilst Co(211) undergoes surface oxidization. In addition, Co(111) tends
to follow the surface carbon coupling mechanism, and surface carbon clusters formed will lead to catalyst de-
activation.

1. Introduction
According to the global greenhouse gas emissions data released by
Intergovernmental Panel on Climate Change (IPCC) in 2014, carbon
dioxide accounts for 65 % of total greenhouse gas emissions, followed
by methane. The greenhouse effect caused by greenhouse gas emissions
can have a great impact on the ecological environment. In addition to
emissions reduction, developing a process to utilize greenhouse gases
provides a new direction for solving this problem. Methane reforming
with carbon dioxide, i.e. dry reforming of methane (DRM), converts
these two greenhouse gases into two clean fuel gases, carbon monoxide
and hydrogen, i.e. synthetic gas (syngas) [1,2], which can be further
used in the Fischer-Tropsch process to synthesize desired hydrocarbons
[3,4]. However, due to the lack of highly active and stable catalysts, the
attractive DRM reaction is still not maturely industrialized.
It was found that highly active and deactivation-resistant precious
metal catalysts, such as Rh and Ru, can be used for DRM [5–7], but the
high price and low reserves limit their further development. In addition
to the noble metals, cheap and earth-abundant transition metals are
considered as alternatives [8]. Ni has attracted considerable attentions
due to its high activity and large reserves, but the disadvantage of the
Ni catalyst is also obvious that it is easily deactivated through carbon
deposition [9–15]. Another widely used inexpensive transition metal
catalyst is cobalt [16,17]. Previous studies indicated that Co could show
a nickel-like activity for the DRM reaction, but the stability problems
still exists, and the deactivation mechanism of Co is still under debate
[18–22].
In comparison to the detailed analysis on the deactivation me-
chanism of Co catalysts in Fischer-Tropsch synthesis studies reported in
the literature [23–25], similar discussions from a computational per-
spective under DRM reaction conditions is missing. Experimental evi-
dence showed that, in the DRM process, the Co catalyst can be deac-
tivated mainly through two mechanisms, namely oxidation and carbon
deposition [21,26–29]. It was demonstrated experimentally that the
deactivation mechanism is strongly dependent on the loading of Co
[21,26–29]. While highly metal-loaded catalysts are mainly deactivated
via coke formation, low metal loading mainly leads to Co oxidation
[26–28]. Jose-Alonso et al. found that the 1 wt % Co catalyst was
abruptly deactivated by oxidation, while the catalyst with a Co content
of 9 wt % showed high carbon deposition due to the large size of Co
particles present on the support [21]. Therefore, the deactivation might
be related to the size of metallic nanoparticles. It could be further in-
ferred that differences in size may result in different ratios of surface
active sites, resulting in different deactivation mechanisms.
In general, the support can affect the size and structure of the metal
particle [30,31], form interfacial active sites with the metal [32], and

⁎
Corresponding author.
E-mail address: yangbo1@shanghaitech.edu.cn (B. Yang).

https://doi.org/10.1016/j.apcatb.2020.118859
Received 19 January 2020; Received in revised form 3 March 2020; Accepted 5 March 2020
Available online 06 March 2020
0926-3373/ © 2020 Elsevier B.V. All rights reserved.
S. Chen, et al.
even directly participate in the catalytic reaction, thereby affecting the
activity and stability [33]. However, it was found that the presence of
support influences the activity and stability mainly by varying the ad-
sorption of intermediate and the energy barriers of key elementary
steps in the reaction. Here, we focus on how the activity and stability of
Co in DRM reactions are affected by species adsorption and activation
barriers [34]. These insights should also be applicable to the discussion
of the activity and stability of supported Co catalysts.
We noticed that there are two crystallographic structures of Co, i.e.
face centered cubic (FCC) and hexagon close-packed (HCP), and tran-
sition between these two structures can happen under different condi-
tions [35–38]. Typically, at low temperatures, Co prefers the form of
HCP phase, and when the temperature rises to 700 K and above, HCP Co
will be transformed to FCC Co [35]. Kitakami and co-workers also
studied the effect of Co particle size on the crystalline phase below
673 K. It was found that when the metal particle diameter is smaller
than 200 Å, cobalt is in the form of FCC phase, and when the particle
diameter is larger than 400 Å, cobalt would prefer the HCP crystal
structure [37]. Considering that the typical operating temperature of
the DRM reaction is around 873∼1123 K and the diameter of Co par-
ticles are usually less than 400 Å [39]. It was also found experimentally
that under the conditions of DRM reaction, Co exists as FCC structure on
different oxide supports [40,41]. Therefore, we adopt the FCC Co in the
current study. Since the deactivation process may be structure sensitive,
we also investigate the reactions on different facets of FCC Co. It was
widely reported in the literature that smaller particles would possess
more low-coordination step sites such as fcc-(211), whilst larger par-
ticles show a higher ratio of fcc-(111) terrace sites [42,43]. It was also
found experimentally and theoretically that Co(111) is the dominant
surface of large FCC Co nanoparticles [41], and on small-sized particles
stepped surfaces are abundant which may give rise to a different de-
activation mechanism.
To the best of our knowledge, there is no thorough computational
study in the literature reporting the reaction kinetics of DRM at dif-
ferent sites over Co and a detailed analysis on the catalyst deactivation
process. In this work, we used the (111) and (211) surfaces of FCC Co as
models to represent different surface sites on the catalyst. Using DFT
calculations combined with microkinetic simulations, reaction rates of
possible reaction pathways of DRM on different Co surfaces were cal-
culated, and the corresponding surface dominant species and rate
controlling species were analyzed. Finally, insights into the mechanism
of deactivation of Co catalysts during the DRM reactions on different
surfaces were provided, and possible strategies for future design of
active and stable Co catalysts were suggested.
2. Computational methods
2.1. DFT calculations
All the DFT calculations were performed with the VASP software
[44–46]. The projector-augmented wave (PAW) method was used to
describe the ion core and its interaction with valence electrons [47].
Electron exchange and correlation was treated with the BEEF-vdW
functional, which is widely used in surface properties calculations
[48–55]. The energy cutoff was set to 500 eV. The energy convergence
criterion was set to 10−4 eV, and the force convergence was set to
0.05 eV/Å. We also included the spin polarization in all the calcula-
tions, and transition states were searched using the constrained mini-
mization method [56–58].
We first optimized the lattice of FCC Co. The optimized lattice
parameter was 3.536 Å, which was found close to the experimental
value of 3.550 Å. And the magnetic moment of Co calculated in this
work is 1.72 μB/atom, which is very close to the experimental value
(1.71 μB/atom) and previous theoretical results (1.63 μB/atom)
[59,60]. Then we built four-layer p(3 × 3) Co(111) and twelve-layer p
(1 × 3) Co(211) slabs for the calculations related with the adsorption
2
Applied Catalysis B: Environmental 270 (2020) 118859
and transition states, and the k-point was set to 3 × 3 × 1 and
4 × 3 × 1, respectively [61]. The bottom two and six layers in the slabs
of Co(111) and Co(211), respectively, were fixed at the bulk structure
and the rest of the atoms in the slabs were allowed to relax during the
structural optimizations. The vacuum space introduced for each slab
was with a height of 15 Å to avoid the periodic interactions. The ad-
sorption energy of surface species was calculated using the following
equation.
Ead = Eslab/ads − Eads − Eslab (1)
Where, Ead is the adsorption energy, Eslab/ads, Eads and Eslab is the energy
of slab with adsorbates, the energy of adsorbate in gas phase and the
energy of slab, respectively.
It was found by several studies that the calculations of gaseous CO2
and CO may be inaccurate using DFT [62,63]. One possible approach to
solve this problem is to correct gaseous energies of CO2 or CO to make
sure that the calculated reaction Gibbs free energies with DFT data are
identical to the experimental values. In this work, the standard reaction
Gibbs free energy calculated for the DRM reaction at 298 K is 1.90 eV,
which is very close to the experimental value of 1.77 eV, suggesting that
no correction should be made for the energies of CO2 and CO here.
2.2. Microkinetic modeling
Steady state approximation and transition state theory were used for
microkinetic simulation. The CatMAP software package developed by
Nørskov’s group was used [64], which has been proved to be a powerful
tool for obtaining kinetic insights into surface catalytic reactions by
applying the mean-field assumption [49–54,65,66]. The simulation
temperature was set to 873∼1073 K and the total pressure was 1 bar.
Methane conversion considered was 5% and the ratio between CH4 and
CO2 in the gaseous reactant was 1:1. Further details of the thermo-
dynamic corrections for gaseous species and surface intermediates were
given in the Supporting Information (SI).
We also resorted to the degree of rate control (DRC) analysis method
developed by the Campbell group [67,68], which has been proved as a
useful tool to probe the surface key intermediates and transition states
in the catalytic process [49–54,65,69,70]. The values of DRC can be
determined by calculating the response of reaction rate to changes in
the free energy of the intermediate and transition states using the fol-
lowing equation:
−∂ ln r ⎞
Xi = ⎜⎛ 0 ⎟
⎝ Gi / kB T ) ⎠Gj0≠ i
∂ (
(2)
Where, r is the overall reaction rate and G is the free energy of each
intermediate or transition state, kB is the Boltzmann constant and T is
the temperature. For a transition state, the value of DRC should be
between 0 and 1. The higher the DRC value, the greater the effect of the
transition state on the rate. The value of intermediates should be 0 or
negative and related to the coverage of the intermediate and the
number of site required by the rate controlling transition state (RCTS)
[67,68].
3. Results and discussion
3.1. Reaction network of the DRM reaction on Co surfaces
Scheme 1 shows the possible DRM reaction network, including three
CO2 activation steps and different pathways from methane to the pro-
duct CO, which are similar to those reported in the literature on Ni, Pt
and Pd catalysts [71–73]. We divide the elementary steps into the fol-
lowing types: (1) Direct and H-assisted dissociation of CO2; (2) Hy-
drogenation of O* to OH*; (3) Dehydrogenation of CH4 to CHx
(x = 0–3); (4) CHx* coupling with O*/OH* to form CHxO*/CHxOH*;
(5) Dehydrogenation of CHxO*/CHxOH* and (6) The coupling of
S. Chen, et al. Applied Catalysis B: Environmental 270 (2020) 118859

Table 1
Reaction energy (ΔE) and activation barrier (Ea) of the DRM elementary steps
on Co(111) and Co(211).
No. elementary steps Co(111)　 Co(211)

ΔE/eV Ea/eV ΔE/eV Ea/eV

1 CO2 + 2* → CO* + O* −0.90 0.50 −1.25 0.09

2 CO2 + H* → COOH* 0.24 1.36 −0.14 1.40
3 COOH* + * →CO* + OH* −1.10 0.58 −1.24 0.60
4 CO2 + H* → HCOO* −0.40 0.82 −1.11 0.73
5 HCOO*+ * → CHO* + O* 0.58 1.12 0.51 2.45
6 CH4 + 2* → CH3* + H* 0.25 1.17 0.09 0.84
7 CH3* + * → CH2* + H* 0.36 0.89 0.58 0.79
8 CH2* + * → CH* + H* −0.23 0.37 −0.32 0.36
9 CH* + * → C* + H* 0.39 1.23 0.05 0.77
10 CH3* + O* → CH3O* + * 0.08 1.56 0.36 2.00
11 CH3O* + * → CH2O* + H* 0.71 0.34 0.75 0.95
12 CH3* + OH* → CH3OH* + * 0.65 2.12 1.20 2.38
13 CH3OH* + * → CH3O* + H* −0.62 1.02 −0.71 0.90
14 CH3OH* + * → CH2OH* + H* 0.50 1.22 0.29 1.05
15 CH2* + O* → CH2O* + * 0.43 1.24 0.53 1.58
16 CH2O* + * → CHO* + H* −0.08 0.34 −0.02 0.45
17 CH2* + OH* → CH2OH* + * 0.79 1.49 0.91 1.48
18 CH2OH* + * → CH2O* + H* −0.40 0.80 −0.25 0.94
19 CH2OH* + * → CHOH* + H* −0.03 0.63 0.17 0.48
20 CH* + O* → CHO* + * 0.58 1.29 0.83 1.67
21 CHO* + * → CO* + H* −1.09 0.16 −0.65 0.05
22 CH* + OH* → CHOH* + * 0.99 1.62 1.41 2.19
23 CHOH* + * → H* + COH* −0.61 0.12 −0.21 0.03
24 CHOH* + * → H* + CHO* −0.46 0.90 −0.44 1.30
25 C* + O* → CO* + * −0.90 1.57 0.13 1.75
26 C* + OH*→ COH* + * −0.01 1.61 1.15 1.79
27 COH* + * → CO* + H* −0.93 0.32 −0.88 0.63
28 O* + H* → OH* + * 0.04 1.22 −0.13 1.56
29 CO* → CO + * 1.46 – 1.46 –
30 H* + H* → H2 + 2* 0.75 – 0.59 –
Scheme 1. (a) Three pathways of carbon dioxide activation; (b) Reaction
scheme from methane to CO, the blue (red) arrows indicate the reactions be-
tween CHx (x = 1–3) and OH* (O*), whilst the black arrows represent dehy- to break the CeH bond; (4) The barrier of CHx* coupling with O* or
drogenation reactions. The dash and solid lines indicate the breaking of OeH
OH* on Co(211) is usually higher than on Co(111). However, it is im-
and CeH bonds, respectively.
practical to determine the reactivity and main reaction pathways only
from the barriers and reaction energies listed in Table 1. Therefore, we
surface H* to form gaseous H2. The configurations of adsorption and further carried out microkinetic simulations over Co(111) and Co(211).
transition states are shown in Figs. S1–S3 of the SI.
3.2. Microkinetic analysis of the DRM reaction on Co surfaces
3.1.1. CO2 activation steps
Scheme 1(a) shows different CO2 activation pathways for CO pro- 3.2.1. Reaction mechanism and active sites
duction. One can see that CO2 can be directly decomposed into CO* and In the microkinetic simulations, all the elementary steps presented
O*, or combined with the H* produced from methane dehydrogenation, in Table 1 are taken into consideration, and the reaction rates of the
to give COOH* or HCOO*. COOH* is then decomposed into either CO* elementary steps at 973 K are listed in Table 2. After analyzing the
and OH* or COH* and O* followed by COH* dehydrogenation to reaction rates, the reactivity of the two Co surfaces and the main DRM
generate CO. Dissociation of HCOO* would produce O* and CHO* reaction pathway can be determined. The reaction rate of the dominant
followed by CHO* dehydrogenation to generate CO. The reaction bar- reaction pathways of DRM are plotted in Fig. 1 as a function of tem-
riers of different elementary steps in the pathways of CO2 activation are perature. It can be found from this figure that the reaction rate on Co
listed in Table 1. One can see from Table 1 that the barrier of CO2 direct (111) is lower than that on Co(211) at the temperatures considered
dissociation (R1) is much lower than that of H* coupling with CO2 (R2 here. In addition, the reaction rate on Co(111) is almost identical to that
and R3) on both Co(111) and Co(211), indicating that the direct dis- obtained by our group on Ni(111) [71]. The difference in the reaction
sociation of CO2 might be favored in comparison with the H-assisted rate of two cobalt surfaces decreases from about 1.5 to 0.5 orders of
activation pathways. magnitude when the temperature increases from 873 K to 1073 K. This
indicates that the step sites should be the active site at low reaction
3.1.2. From CH4 to CO temperatures, but the reactivity of the terrace and step sites becomes
Scheme 1(b) shows the different reaction pathways from CH4 to H2 identical at high reaction temperatures.
and CO. Methane can be dehydrogenated to produce CHx (x = 0∼3) Regarding the mechanism of CO2 activation, one can see from
species. Then the CHx species can couple with O * or OH * to form CHxO Table 2 that the rate of CO2 direct dissociation is at least five orders of
or CHxOH. These surface intermediates will undergo further dehy- magnitude higher than that of H-assisted dissociation on both Co(111)
drogenation until the final product CO is formed. The energetics listed and Co(211), indicating that the direct dissociation of CO2 is favored.
Table 1 for the two Co surfaces reflect the following trends: (1) For This is the same with the way of carbon dioxide activation on Ni,
CHx*, the dehydrogenation barrier is lower than the corresponding whereas Pd and Pt favor the hydrogen-assisted carbon dioxide dis-
coupling barrier of with O*/OH*; (2) Compared with CHx* coupling sociation [71,73,74]. The dominant reaction path on Co(111) is also the
with OH*, the association with O* is kinetically favored; (3) For same as that on Ni(111) [71]. On Co(111), there are two dominant
CHxOH* species, the barrier for the OeH bond breaking is higher than reaction pathways for CO formation, namely the C + O path and

3
S. Chen, et al. Applied Catalysis B: Environmental 270 (2020) 118859

Table 2 The rates of other reaction paths are all at least 3 orders of magnitude
The net reaction rate of each elementary step calculated at the steady state and lower than the rate of the C + O path. In contrast, as shown in Fig. 1(b),
973 K on Co(111) and Co(211). the C + O path can be considered as the dominant reaction path on Co
No. elementary steps reaction rate/site−1 s−1 (211). The rates of other paths are at least 1.5 orders of magnitude
lower than those of the C + O path. Based on our previous work, we
Co(111) Co(211) found that the main reaction paths on Co are the same as on the Ni
surfaces. [75] However, the reaction rate of Ni catalyst is higher, which
1 CO2 + 2* → CO* + O* 9.58 × 10° 4.16 × 101
2 CO2 + H* → COOH* 3.09 × 10−5 5.02 × 10−5 is consistent with the experimental observation [76,77]. Meanwhile, as
3 COOH* + * →CO* + OH* 3.09 × 10−5 5.02 × 10−5 we mentioned in the introduction, the deactivation performance of Ni
4 CO2 + H* → HCOO* 6.27 × 10−5 7.48 × 10−9 and Co is different, which will be discussed in the following sections.
5 HCOO*+ * → CHO* + O* 6.27 × 10−5 7.48 × 10−9
6 CH4 + 2* → CH3* + H* 9.58 × 10° 4.16 × 101
7 CH3* + * → CH2* + H* 9.58 × 10° 4.16 × 101 3.2.2. Coverage of surface species and DRC analysis
8 CH2* + * → CH* + H* 9.58 × 10° 4.16 × 101
We plotted in Fig. 2 the steady state coverage of dominant surface
9 CH* + * → C* + H* 1.05 × 10° 4.14 × 101
10 CH3* + O* → CH3O* + * 4.92 × 10−6 3.03 × 10−6 intermediates, including CO*, C* and O* on two Co surfaces. The
11 CH3O* + * → CH2O* + H* 4.91 × 10−6 3.04 × 10−6 coverage of other surface intermediates is close to 0 and thus omitted
12 CH3* + OH* → CH3OH* + * 1.13 × 10−8 4.80 × 10−1 from Fig. 2. On Co(111), the coverage of CO* is around 0.2 monolayer
13 CH3OH* + * → CH3O* + H* −1.03 × 10-8 5.69 × 10−10
(ML) at 873 K, and decreases to 0 when the temperature is higher than
14 CH3OH* + * → CH2OH* + H* 1.13 × 10−8 1.02 × 10−8
15 CH2* + O* → CH3O* + * −4.44 × 10-6 −3.00 × 10-9
973 K. The coverage of C* on the surface stays at 0 ML at different
16 CH2O* + * → CHO* + H* 5.03 × 10−205 3.80 × 10−118 temperatures, whilst the coverage of O* slightly increases from 0 to
17 CH2* + OH* → CH2OH* + * 6.40 × 10−7 4.40 × 10−7 0.1 ML when the temperature rises from 873 to 1073 K. On Co(211), the
18 CH2OH* + * → CH2O* + H* −4.65 × 10-7 −1.30 × 10-8 coverage of O* increases from 0.6 ML (873 K) to about 0.8 ML (1073 K)
19 CH2OH* + * → CHOH* + H* 1.12 × 10−6 4.53 × 10−7
with temperature. Therefore, one can find that the high coverage of
20 CH* + O* → CHO* + * 8.53 × 10° 2.00 × 10−1
21 CHO* + * → CO* + H* 8.53 × 10° 2.00 × 10−1 oxygen on Co(211) may facilitate the oxidation of Co. The coverage of
22 CH* + OH* → CHOH* + * 7.37 × 10−4 4.11 × 10−6 both CO* and C* on Co(211) is low at 873 K (about 0.1 ML) and drops
23 CHOH* + * → H* + COH* 7.36 × 10−4 4.37 × 10−6 to about 0 ML when the temperature rises.
24 CHOH* + * → H* + CHO* 2.59 × 10−6 1.91 × 10−4
Since the reverse water gas shift (RWGS) reaction, i.e. CO2 + H2 →
25 C* + O* → CO* + * 1.04 × 10° 4.09 × 101
26 C* + OH*→ COH* + * 6.51 × 10−3 4.80 × 10−1
CO + H2O, was reported as a major side reaction in the DRM reaction
27 COH* + * → CO* + H* 7.25 × 10−3 4.80 × 10−1 [78,79], we investigated the effect of this reaction on the coverage of
28 O* + H* → OH* + * 7.22 × 10−3 4.80 × 10−1 O* over Co(111) and Co(211). We added the elementary step of H2O
29 CO* → CO + * 1.92 × 101 8.32 × 101 formation to the kinetic model, with other simulation conditions un-
30 H* + H* → H2 + 2* 1.92 × 101 8.32 × 101
changed. The obtained coverage of O* at steady state are shown in Fig.
S6. One can find from the figure that, upon considering the RWGS re-
CH + O path, as shown in Fig. 1. In the C + O Path, CH4 is dehy- action, the coverage of O* is almost unchanged on Co(111), while on Co
drogenated to C* sequentially, and then C* couples with O*, formed (211), the coverage is reduced by around 0.1 ML. However, the cov-
from CO2 dissociation, to produce CO*. In the CH + O path, CH* erage of O*on Co(211) is still much higher than that on Co(111), even
couples with O* to form CHO* and finally CO* is obtained through with the RWGS reaction included in the modeling.
dehydrogenation. As one can see from Fig. 1(a), on Co(111), the rate of The DRC analysis results are also shown in Fig. 2. One can find from
CH + O path is around 1.5 orders of magnitude higher than that of the the figure that, on both Co(111) and Co(211), transition state of CH4
C + O path at 873 K, and the difference is narrowed down at 1073 K. dissociation is the RCTS at all the temperatures considered. On Co(111),
the DRC value of CO* is around -0.5 at 873 K, and gets close to zero

Fig. 1. Reaction rates of the dominant reaction pathways (highlighted in the above panel) and the total rates of the DRM reaction on Co(111) and Co(211) as a
function of temperature.

4
S. Chen, et al.
Fig. 2. Coverage of surface intermediates and the calculated DRC values of transition states and surface intermediates on Co(111) and Co(211), as a function of
temperatures.
when temperature rises due to the decrease of CO coverage on the
surface. In contrast, the DRC value of O* is close to 0 at 873 K, and
becomes more negative to -0.25 at 1073 K. On Co(211) surface, O* is
shown as the only rate controlling intermediates and becomes more
rate-controlling with the increase of temperature. Although oxygen is
not directly participating in the dissociation of CH4, the presence of
oxygen on the surface will influence the surface free sites that can be
used for the dissociation reaction where one more free site is needed.
3.3. Deactivation mechanism of Co
3.3.1. Stability of C* and O* on two Co surfaces
In order to understand the behavior of C* and O* during the de-
activation of Co catalysts, we hereby calculate the adsorption of C* and
O* at different coverages on Co(111) and Co(211). For the extra cal-
culations, we constructed four-layer p(2 × 2) Co(111) and twelve-layer
p(1 × 4) Co(211) slabs and placed different numbers of adsorbates at
these sites. The k-points used for the above two slabs were (3 × 3 × 1)
and (4 × 2×1) for Co(111) and Co(211), respectively. Other para-
meters and convergence criteria are the same with all the previous
calculations.
Typically, carbon deposition can follow the dissolving mechanism
or the surface coupling mechanism. In the dissolving mechanism, the
surface C* species will dissolve into the bulk of metals to form carbide,
and then precipitates at the rear side of the nanoparticles to form a
carbon nanofilament [80]. In the surface coupling mechanism, the
surface carbon species will couple with each other and form stable
surface clustered structures [80]. In this section we will consider both
mechanisms at different Co sites by determining the behavior of carbon
at the surface and subsurface sites with different coverages. Regarding
the oxidation process, as mentioned above, the coverage of O* at the
step sites over Co(211) are much higher than that on Co(111), we
therefore further verify the coverage dependent adsorption of O* at the
surface sites of both surfaces.
The average adsorption Gibbs free energy of C* or O* adsorbed at
the surface or subsurface site at different coverage can be calculated
using the following equation:
5
Applied Catalysis B: Environmental 270 (2020) 118859
E Nads /slab − Eslab − Nμads
ΔG (T ,p,Nads) =
N (3)
Where, ΔG (T , p , Nads) , hereinafter referred to as ΔGC* or ΔGO*, is the
adsorption free energy of C* or O*, N is the number of adsorbates,
ENads/slab and Eslab is the energy of slab with N number of adsorbates and
the energy of clean surface, respectively. μads is the chemical potential
of referenced C or O, which can be calculated from μC = μCH4 - 2μH2 and
μO = μCO2(gas) - μCO(gas), respectively. The chemical potential can be
calculated from the DFT energy with thermodynamic corrections at
973 K, and the pressures of gaseous species is the same with those used
in the microkinetic simulations.
We summarize in Fig. 3 the average adsorption free energy of C*
and O* on Co(111) and Co(211) as a function of coverage. The surface/
subsurface ΔGC* will determine whether the adsorbed carbon can dif-
fuse from the surface sites to subsurface and which carbon deposition
mechanism to follow on different Co surfaces [81]. We find that on Co
(111), the surface/subsurface ΔGC* at 0.25 ML are almost identical,
indicating that the diffusion of C* between the surface and subsurface
sites is thermodynamically neutral. As the coverage increases, the dif-
ference between ΔGC* at the subsurface and surface becomes larger, and
the surface carbon becomes more stable, indicating that the diffusion of
C* from surface to subsurface becomes thermodynamically prohibited.
In addition, we find that when the coverage is higher than 0.25 ML,
surface carbon atoms tend to form clusters on Co(111) (see Fig. 3a), and
are therefore stabilized. However, such cluster structures are not ob-
served on Co(211), where C* adsorption is destabilized with coverages
increased. More importantly, surface C* is more stable than subsurface
C* at any coverage on Co(211), indicating that the diffusion of C* to
subsurface is also thermodynamically prohibited.
For surface O*, it is also shown in Fig. 3 that higher coverage will
result in weaker adsorption on the two cobalt surfaces studied. More
importantly, the adsorption of O* on Co(211) is always more stable
than that on Co(111) at all the coverages considered here. Combining
with the microkinetic simulation results obtained from low coverage
DFT energies shown in Fig. 2, these results suggest that O* deposition
will preferentially take place on Co(211). Furthermore, the dissociation
barrier of CO2 on Co(211) is lower than on Co(111), indicating that the
produce of O* on Co(211) surface is also kinetically favored. The
S. Chen, et al.
Fig. 3. Average adsorption free energy (973 K) of carbon and oxygen atoms at
different coverages on (a) Co(111) and (b) Co(211). Here sur and sub are the
surface and subsurface sites, respectively. The average adsorption free energies
are calculated with Eq. (3). Figures inserted are the corresponding optimized
configurations of carbon adsorbed at the surface sites.
surface oxidation process of a catalyst is rather complex and may in-
clude several steps, such as deposition of oxygen, oxygen permeation
into the bulk and surface reconstruction induced by oxygen. Among all
these steps, deposition of surface oxygen should be the first, and higher
coverage of oxygen deposition would indicate that the surface might be
easier to be oxidized
3.3.2. Kinetics of carbon deposition
After discussing the thermodynamic properties of C* at the surface
or subsurface sites on two Co facets, the kinetics of carbon deposition
on the surface of p(3 × 3) Co(111) and p(1 × 4) Co(211) are further
studied. We calculated the barriers of the first step of surface coupling
mechanism (CeC coupling), the first step of carbon dissolve mechanism
(C* diffusion to subsurface) and CO* formation on two cobalt surfaces,
6
Applied Catalysis B: Environmental 270 (2020) 118859
and the results are summarized in Fig. 4.
On Co(111), the barrier of CeC coupling (1.07 eV) is lower than that
of the diffusion of C* from surface to subsurface (1.23 eV) and CO*
formation (1.59 eV), indicating the surface coupling mechanism on this
surface is kinetically favored. On Co(211), the barrier of C* and O*
association (1.74 eV) is the lowest among all three elementary steps
considered, indicating that CO formation might be kinetically favored
on this surface. The above results suggest that carbon deposition is
preferred on Co(111), which is consistent with the experimental results
that catalyst with large Co particles is readily deactivated due to carbon
deposition [26,82,83]. Furthermore, we find that the carbon deposition
over cobalt in DRM reaction follows the surface coupling mechanism
rather than the dissolving mechanism.
3.3.3. Competition between the deposition and removal of surface carbon
According to the results presented above, the surface coupling me-
chanism is kinetically favored but thermodynamically unfavored in
comparison with CO formation on Co(111) (Fig. 4a). To further un-
derstand the competition between the deposition and removal of sur-
face carbon on Co(111), we calculated and compared the corresponding
reaction rates at different temperatures. The relevant elementary steps
considered are C* + C* → C2* + * and C* + O* ↔ CO* + *. It should
be noted that the single direction arrow indicates an irreversible reac-
tion while the double direction arrow for a reversible reaction.
We here consider the CeC bonding reaction as an irreversible re-
action for the following reasons. Firstly, under the temperature condi-
tions we considered in this work, the formation of C2* is always an
exergonic process with the reaction Gibbs free energy closed to
−0.64 eV (see Fig. S7). Secondly, C2* is not the end point of surface C*
coupling due to the thermodynamically favored carbon cluster forma-
tion. Finally, the coverage of surface C2* at the beginning of the surface
carbon deposition process is negligible, meaning that the reverse rate of
the CeC bonding reaction should be low enough to be ignored.
Regarding CO* formation, the reverse reaction could not be ignored
since CO* is the main product and the concentration of CO* is much
higher than that of C2*. Therefore, the ratio between the reaction rates
can be calculated from the following equation
+
rC2 kC-CθC2
= + *−
rCO kC-OθC*θO* − kC-O θCO *θ* (4)
+
Where, kC-C represents the forward rate constant of C2* formation re-
+ −
action, kC-O and kC-O are the forward and reverse rate constant of C* and
O* association reaction, respectively. θC*, θO*, θCO* and θ* is the cov-
erage of C*, O*, CO* and free site from the above microkinetic simu-
lations, respectively, implying that we are taking the steady state of
DRM over the clean Co surface as the starting point of deactivation. We
plot the calculation results of Eq. (4) against the reaction temperature
in Fig. 5.
From Fig. 5, one can see that at 873 K, the rate of C2* formation is
five times faster than that of CO formation. With the increase of tem-
perature, the ratio decreases gradually, and is equal to 1 at the tem-
perature around 973 K. This indicates that below 973 K, the rate of CeC
Fig. 4. Energy profiles of carbon oxidation to
CO (black), carbon-carbon coupling (blue) and
diffusion of surface carbon into subsurface
(red) on Co(111) and Co(211). C** in the fig-
ures represents subsurface carbon. (For inter-
pretation of the references to colour in this
figure legend, the reader is referred to the web
version of this article).
S. Chen, et al.
Fig. 5. Ratio between the rates of carbon-carbon coupling and carbon oxidation
to CO on Co (111) surface, which is calculated with Eq. (4), as a function of
temperature. Higher carbon-carbon coupling rate indicates faster carbon de-
position onto the catalyst surface.
coupling is higher than that of CO formation on Co(111). However, at
high temperatures, i.e. above 973 K, CeC coupling rate tends to de-
crease in comparison with CO formation rate. Hence, our results suggest
that carbon deposition is the main mechanism of deactivation on Co
(111) surface at low temperatures, and that high temperatures prevent
catalyst deactivation. Such observations agree well with experiments.
Indeed, the coke formation on 10 wt % Co/γ-Al2O3 catalyst is 0.25 g
(carbon) g(cat)−1 h−1 at 973 K [84], or 35 % of reacted methane turn
into carbon deposition on 7.6 wt% Co catalyst [85]. At high tempera-
tures, e.g. 1023 K, no carbon deposition and deactivation during DRM
reaction at 1073 K with 10 wt% Co/Al2O3 catalyst [86].
3.4. Insights into future cobalt-based catalyst design
Now we have clarified the reactivity and deactivation mechanism of
the Co catalyst towards the DRM reaction. In general, the trend of both
activity and stability of catalysts with different sizes can be derived
from the ratios of the (211) and (111) surfaces. Smaller Co particles
showing more low coordination sites would give higher DRM activity at
lower reaction temperature condition but also higher deactivation
possibility through oxidation. The possibility of carbon deposition over
larger particles of Co, which are dominated by terrace sites, is strongly
dependent on the reaction temperature, and higher temperatures will
inhibit the coke formation. Based on these understandings, we further
propose below some suggestions for the modification of Co-based cat-
alysts.
In order to obtain balanced activity and stability of the DRM cata-
lysts, we recently found that a good dopant should give rise to a proper
adsorption energy of carbon to ensure that the C* formation process,
e.g. CH4 dissociation, is rate controlling to improve the carbon re-
sistance, and relatively low dissociation barriers of CH4 and CO2 can be
achieved to maintain a good activity [75]. By doping or forming an
alloy, the adsorption of O* at the Co step is regulated and the oxidation
of the Co catalyst can be inhibited. For Co with large particles and more
(111) facets, our simulation suggests that high reaction temperature
should be achieved to reduce carbon deposition.
4. Conclusions
In this work, the reactivity and reaction pathway of DRM on Co
(111) and Co(211) were studied with a combined DFT calculation and
microkinetic modeling approach. We found that the step site over Co
(211) is the active site for the DRM reaction. On Co(111), the C + O
and CH + O paths are the preferred reaction paths, while the C + O
7
Applied Catalysis B: Environmental 270 (2020) 118859
path is the only dominant path on Co(211). It was also found that the
rate-controlling step of the DRM reaction is the same over Co(111) and
Co(211), i.e. the dissociation of CH4. We further studied the deactiva-
tion of two Co surfaces and showed the deactivation mechanism is facet
dependent. Compared with Co(111), the stronger O* adsorption and
lower CO2 dissociation barrier on Co(211) leads to much higher O*
coverage on the surface, indicating that O* deposition will pre-
ferentially take place on Co(211). In contrast, coupling of surface
carbon atoms to carbon clusters is thermodynamically and kinetically
favorable on Co(111). We also highlighted that Co deactivation through
carbon deposition is temperature dependent, and the resistance to
carbon deposition would be improved at high temperatures. Our results
provide insights into understanding experimental observations of the
different deactivation behavior of catalysts with different Co particle
size and at different temperatures, based on which strategies for the
design of Co-based catalysts were proposed.
CRediT authorship contribution statement
Shuyue Chen: Investigation, Data curation, Formal analysis,
Writing - original draft. Jeremie Zaffran: Writing - review & editing.
Bo Yang: Conceptualization, Supervision, Funding acquisition, Project
administration, Resources, Writing - review & editing.
Declaration of Competing Interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influ-
ence the work reported in this paper.
Acknowledgements
This work is financially supported by the National Natural Science
Foundation of China (91745102, 21950410522). We thank the HPC
Platform of ShanghaiTech University and Shanghai Supercomputer
Center for computing time.
Appendix A. Supplementary data
Supplementary material related to this article can be found, in the
online version, at doi:https://doi.org/10.1016/j.apcatb.2020.118859.
References
[1] T. Yabe, Y. Sekine, Methane conversion using carbon dioxide as an oxidizing agent:
a review, Fuel Process. Technol. 181 (2018) 187–198.
[2] S.R. Foit, I.C. Vinke, L.G.J. de Haart, R.-A. Eichel, Power-to-Syngas – eine
Schlüsseltechnologie für die Umstellung des Energiesystems? Angew. Chem. 129
(2017) 5488–5498.
[3] S. Golestan, A.A. Mirzaei, H. Atashi, Fischer-Tropsch synthesis over an iron-cobalt-
manganese (ternary) nanocatalyst prepared by hydrothermal procedure: effects of
nanocatalyst composition and operational conditions, Int. J. Hydrog. Energy 42
(2017) 9816–9830.
[4] P.F. Cao, S. Adegbite, H.T. Zhao, E. Lester, T. Wu, Tuning dry reforming of methane
for F-T syntheses: a thermodynamic approach, Appl. Energy 227 (2018) 190–197.
[5] D.H. Mei, V.A. Glezakou, V. Lebarbier, L. Kovarik, H.Y. Wan, K.O. Albrecht,
M. Gerber, R. Rousseau, R.A. Dagle, Highly active and stable MgAl2O4-supported
Rh and Ir catalysts for methane steam reforming: a combined experimental and
theoretical study, J. Catal. 316 (2014) 11–23.
[6] Y. Khani, Z. Shariatinia, F. Bahadoran, High catalytic activity and stability of
ZnLaAlO4 supported Ni, Pt and Ru nanocatalysts applied in the dry, steam and
combined dry-steam reforming of methane, Chem. Eng. J. 299 (2016) 353–366.
[7] N. Kopfle, T. Gotsch, M. Grunbacher, E.A. Carbonio, M. Havecker, A. Knop-Gericke,
L. Schlicker, A. Doran, D. Kober, A. Gurlo, S. Penner, B. Klotzer, Zirconium-assisted
activation of palladium to boost syngas production by methane dry reforming,
Angew. Chem.-Int. Edit. 57 (2018) 14613–14618.
[8] B. Abdullah, N.A.A. Ghani, D.V.N. Vo, Recent advances in dry reforming of methane
over Ni-based catalysts, J. Clean. Prod. 162 (2017) 170–185.
[9] A.E. Castro Luna, M.E. Iriarte, Carbon dioxide reforming of methane over a metal
modified Ni-Al2O3 catalyst, Appl. Catal. A Gen. 343 (2008) 10–15.
[10] S. Damyanova, B. Pawelec, K. Arishtirova, J.L.G. Fierro, Ni-based catalysts for re-
forming of methane with CO2, Int. J. Hydrog. Energy 37 (2012) 15966–15975.
S. Chen, et al.
[11] Y. Lou, M. Steib, Q. Zhang, K. Tiefenbacher, A. Horvath, A. Jentys, Y. Liu,
J.A. Lercher, Design of stable Ni/ZrO2 catalysts for dry reforming of methane, J.
Catal. 356 (2017) 147–156.
[12] Z. Zuo, S. Liu, Z. Wang, C. Liu, W. Huang, J. Huang, P. Liu, Dry reforming of me-
thane on single-site Ni/MgO catalysts: importance of site confinement, ACS Catal.
(2018) 9821–9835.
[13] S. Wen, M. Liang, J. Zou, S. Wang, X. Zhu, L. Liu, Z.-j. Wang, Synthesis of a SiO2
nanofibre confined Ni catalyst by electrospinning for the CO2 reforming of methane,
J. Mater. Chem. A 3 (2015) 13299–13307.
[14] S. Kawi, Y. Kathiraser, J. Ni, U. Oemar, Z.W. Li, E.T. Saw, Progress in synthesis of
highly active and stable nickel-based catalysts for carbon dioxide reforming of
methane, ChemSusChem 8 (2015) 3556–3575.
[15] C.J. Liu, J.Y. Ye, J.J. Jiang, Y.X. Pan, Progresses in the Preparation of Coke Resistant
Ni-based Catalyst for Steam and CO2 Reforming of Methane, ChemCatChem 3
(2011) 529–541.
[16] J.Y. Lu, Y. Guo, Q.R. Liu, G.Z. Han, Z.J. Wang, Co-based catalysts for carbon dioxide
reforming of methane to synthesis gas, Prog. Chem. 29 (2017) 1471–1479.
[17] J.X. Liu, H.Y. Su, D.P. Sun, B.Y. Zhang, W.X. Li, Crystallographic dependence of CO
activation on cobalt catalysts: HCP versus FCC, J. Am. Chem. Soc. 135 (2013)
16284–16287.
[18] A.W. Budiman, S.H. Song, T.S. Chang, C.H. Shin, M.J. Choi, Dry reforming of me-
thane over cobalt catalysts: a literature review of catalyst development, Catal. Surv.
Asia 16 (2012) 183–197.
[19] K. Nagaoka, K. Takanabe, K. Aika, Modification of Co/ TiO2 for dry reforming of
methane at 2 MPa by Pt, Ru or Ni, Appl. Catal. A-Gen. 268 (2004) 151–158.
[20] K. Omata, N. Nukui, T. Hottai, M. Yamada, Cobalt-magnesia catalyst by oxalate co-
precipitation method for dry reforming of methane under pressure, Catal. Commun.
5 (2004) 771–775.
[21] D.S. Jose-Alonso, M.J. Illan-Gomez, M.C. Roman-Martinez, Low metal content Co
and Ni alumina supported catalysts for the CO2 reforming of methane, Int. J.
Hydrog. Energy 38 (2013) 2230–2239.
[22] F. Mirzaei, M. Rezaei, F. Meshkani, Z. Fattah, Carbon dioxide reforming of methane
for syngas production over Co-MgO mixed oxide nanocatalysts, J. Ind. Eng. Chem.
21 (2015) 662–667.
[23] J. Cheng, P. Hu, P. Ellis, S. French, G. Kelly, C.M. Lok, Some understanding of
Fischer–Tropsch synthesis from density functional theory calculations, Top. Catal.
53 (2010) 326–337.
[24] R.A. van Santen, M.M. Ghouri, S. Shetty, E.M.H. Hensen, Structure sensitivity of the
Fischer–Tropsch reaction; molecular kinetics simulations, Catal. Sci. Technol. 1
(2011) 891–911.
[25] C.Q. Chen, Y.Y. Zhan, J.K. Zhou, D.L. Li, Y.J. Zhang, X.Y. Lin, L.L. Jiang, Q. Zheng,
Cu/CeO2 catalyst for water-gas shift reaction: effect of CeO2 pretreatment,
ChemPhysChem 19 (2018) 1448–1455.
[26] H.Y. Wang, E. Ruckenstein, CO2 reforming of CH4 over Co/MgO solid solution
catalysts - effect of calcination temperature and Co loading, Appl. Catal. A-Gen. 209
(2001) 207–215.
[27] E. Ruckenstein, H.Y. Wang, Carbon deposition and catalytic deactivation during
CO2 reforming of CH4 over Co/gamma-Al2O3 catalysts, J. Catal. 205 (2002)
289–293.
[28] D. San-Jose-Alonso, J. Juan-Juan, M.J. Illan-Gomez, M.C. Roman-Martinez, Ni, Co
and bimetallic Ni-Co catalysts for the dry reforming of methane, Appl. Catal. A-Gen.
371 (2009) 54–59.
[29] M.F.F. Sukri, M. Khavarian, A.R. Mohamed, Effect of cobalt loading on suppression
of carbon formation in carbon dioxide reforming of methane over Co/MgO catalyst,
Res. Chem. Intermed. 44 (2018) 2585–2605.
[30] E. Ruckenstein, H.Y. Wang, Carbon dioxide reforming of methane to synthesis gas
over supported cobalt catalysts, Appl. Catal. A-Gen. 204 (2000) 257–263.
[31] Y. Guo, J.Y. Lu, Q.R. Liu, X.L. Bai, L.J. Gao, W.X. Tu, Z.J. Wang, Carbon dioxide
reforming of methane over cobalt catalysts supported on hydrotalcite and metal
oxides, Catal. Commun. 116 (2018) 81–84.
[32] C.A. Singh, L. Bansal, P. Tiwari, V.V. Krishnan, Strong metal support interactions of
infiltrated Ni with TiO2 in a porous YSZ anode matrix - a possible method for Ni-
stabilization, in: S.C. Singhal, H. Yokokawa (Eds.), Solid Oxide Fuel Cells 11,
Electrochemical Soc Inc, Pennington, 2009, pp. 1897–1904.
[33] L. Yu, Y. Liu, F. Yang, J. Evans, J.A. Rodriguez, P. Liu, CO oxidation on gold-sup-
ported Iron oxides: new insights into strong oxide–metal interactions, J. Phys.
Chem. C 119 (2015) 16614–16622.
[34] J.-H. Park, S. Yeo, T.-S. Chang, Effect of supports on the performance of Co-based
catalysts in methane dry reforming, J. CO2 UTIL. 26 (2018) 465–475.
[35] T. Nishizawa, K. Ishida, The Co (Cobalt) System, Bull. Alloy. Phase Diagr. 4 (1983)
387.
[36] J. Gong, L.L. Wang, Y. Liu, J.H. Yang, Z.G. Zong, Structural and magnetic properties
of hcp and fcc Ni nanoparticles, J. Alloys. Compd. 457 (2008) 6–9.
[37] O. Kitakami, H. Sato, Y. Shimada, F. Sato, M. Tanaka, Size effect on the crystal
phase of cobalt fine particles, Phys. Rev. B 56 (1997) 13849–13854.
[38] R. Lizarraga, F. Pan, L. Bergqvist, E. Holmstrom, Z. Gercsi, L. Vitos, First principles
theory of the hcp-fcc phase transition in cobalt, Sci. Rep. 7 (2017) 8.
[39] K. Takanabe, K. Nagaoka, K. Nariai, K. Aika, Influence of reduction temperature on
the catalytic behavior of Co/TiO2 catalysts for CH4/ CO2 reforming and its relation
with titania bulk crystal structure, J. Catal. 230 (2005) 75–85.
[40] H. Ay, D. Uner, Dry reforming of methane over CeO2 supported Ni, Co and Ni-Co
catalysts, Appl. Catal. B-Environ. 179 (2015) 128–138.
[41] Y.J. Wong, M.K. Koh, M. Khavarian, A.R. Mohamed, Investigation on cobalt alu-
minate as an oxygen carrier catalyst for dry reforming of methane, Int. J. Hydrog.
Energy 42 (2017) 28363–28376.
[42] D. Gao, H. Zhou, J. Wang, S. Miao, F. Yang, G. Wang, J. Wang, X. Bao, Size-
8
Applied Catalysis B: Environmental 270 (2020) 118859
dependent electrocatalytic reduction of CO2 over Pd nanoparticles, J. Am. Chem.
Soc. 137 (2015) 4288–4291.
[43] P. Zhao, Z. Cao, X. Liu, P. Ren, D.-B. Cao, H. Xiang, H. Jiao, Y. Yang, Y.-W. Li, X.-
D. Wen, Morphology and reactivity evolution of HCP and FCC Ru nanoparticles
under CO atmosphere, ACS Catal. 9 (2019) 2768–2776.
[44] G. Kresse, J. Hafner, Ab-initio molecular-dynamics simulation of the liquid-metal
amorphous-semiconductor transition in germanium, Phys. Rev. B 49 (1994)
14251–14269.
[45] G. Kresse, J. Furthmuller, Efficient iterative schemes for ab initio total-energy cal-
culations using a plane-wave basis set, Phys. Rev. B 54 (1996) 11169–11186.
[46] G. Kresse, J. Furthmuller, Efficiency of ab-initio total energy calculations for metals
and semiconductors using a plane-wave basis set, Comput. Mater. Sci. 6 (1996)
15–50.
[47] P.E. Blochl, Projector augmented-wave method, Phys. Rev. B 50 (1994)
17953–17979.
[48] J. Wellendorff, K.T. Lundgaard, A. Mogelhoj, V. Petzold, D.D. Landis, J.K. Norskov,
T. Bligaard, K.W. Jacobsen, Density functionals for surface science: exchange-cor-
relation model development with Bayesian error estimation, Phys. Rev. B 85
(2012) 23.
[49] P.P. Wu, B. Yang, Theoretical insights into the promotion effect of subsurface boron
for the selective hydrogenation of CO to methanol over Pd catalysts, Phys. Chem.
Chem. Phys. 18 (2016) 21720–21729.
[50] K.L. Wang, B. Yang, Theoretical understanding on the selectivity of acrolein hy-
drogenation over silver surfaces: the non-Horiuti-Polanyi mechanism is the key,
Catal. Sci. Technol. 7 (2017) 4024–4033.
[51] P.P. Wu, B. Yang, Significance of Surface Formate Coverage on the Reaction
Kinetics of Methanol Synthesis from CO2 Hydrogenation over Cu, ACS Catal. 7
(2017) 7187–7195.
[52] K.R. Yang, B. Yang, Identification of the active and selective sites over a single Pt
atom-alloyed Cu catalyst for the hydrogenation of 1,3-Butadiene: a combined DFT
and microkinetic modeling study, J. Phys. Chem. C 122 (2018) 10883–10891.
[53] P.P. Wu, J. Zaffran, B. Yang, Role of surface species interactions in identifying the
reaction mechanism of methanol synthesis from CO2 hydrogenation over inter-
metallic Pdln(310) Steps, J. Phys. Chem. C 123 (2019) 13615–13623.
[54] D.Y. Xu, P.P. Wu, B. Yang, Essential role of water in the autocatalysis behavior of
methanol synthesis from CO2 hydrogenation on Cu: a combined DFT and micro-
kinetic modeling study, J. Phys. Chem. C 123 (2019) 8959–8966.
[55] H. Liu, J. Liu, B. Yang, Modeling the effect of surface CO coverage on the electro-
catalytic reduction of CO2 to CO on Pd surfaces, Phys. Chem. Chem. Phys. 21 (2019)
9876–9882.
[56] A. Alavi, P.J. Hu, T. Deutsch, P.L. Silvestrelli, J. Hutter, CO oxidation on Pt(111): an
ab initio density functional theory study, Phys. Rev. Lett. 80 (1998) 3650–3653.
[57] Z.P. Liu, P. Hu, General rules for predicting where a catalytic reaction should occur
on metal surfaces: a density functional theory study of C-H and C-O bond breaking/
making on flat, stepped, and kinked metal surfaces, J. Am. Chem. Soc. 125 (2003)
1958–1967.
[58] A. Michaelides, Z.P. Liu, C.J. Zhang, A. Alavi, D.A. King, P. Hu, Identification of
general linear relationships between activation energies and enthalpy changes for
dissociation reactions at surfaces, J. Am. Chem. Soc. 125 (2003) 3704–3705.
[59] H. Schäfer, J. Donohue, The Structures of the Elements, John Wiley & Sons, New
York, Sydney, Toronto, 1974 436 S., Preis: £ 12.—, 79 (1975) 110–110.
[60] X. Liu, L. Sun, W.-Q. Deng, Theoretical investigation of CO2 adsorption and dis-
sociation on low index surfaces of transition metals, J. Phys. Chem. C 122 (2018)
8306–8314.
[61] M. Methfessel, A.T. Paxton, High-precision sampling for Brillouin-zone integration
in metals, Phys. Rev. B 40 (1989) 3616–3621.
[62] A.A. Peterson, F. Abild-Pedersen, F. Studt, J. Rossmeisl, J.K. Norskov, How copper
catalyzes the electroreduction of carbon dioxide into hydrocarbon fuels, Energy
Environ. Sci. 3 (2010) 1311–1315.
[63] R. Christensen, H.A. Hansen, T. Vegge, Identifying systematic DFT errors in cata-
lytic reactions, Catal. Sci. Technol. 5 (2015) 4946–4949.
[64] A.J. Medford, C. Shi, M.J. Hoffmann, A.C. Lausche, S.R. Fitzgibbon, T. Bligaard,
J.K. Norskov, CatMAP: a software package for descriptor-based microkinetic map-
ping of catalytic trends, Catal. Lett. 145 (2015) 794–807.
[65] P. Wu, B. Yang, Intermetallic PdIn catalyst for CO2 hydrogenation to methanol:
mechanistic studies with a combined DFT and microkinetic modeling method,
Catal. Sci. Technol. 9 (2019) 6102–6113.
[66] P. Wang, B. Yang, Influence of surface strain on activity and selectivity of Pd-based
catalysts for the hydrogenation of acetylene: a DFT study, Chin. J. Catal. 39 (2018)
1493–1499.
[67] C. Stegelmann, A. Andreasen, C.T. Campbell, Degree of rate control: how much the
energies of intermediates and transition states control rates, J. Am. Chem. Soc. 131
(2009) 8077–8082.
[68] C.T. Campbell, The degree of rate control: a powerful tool for catalysis research,
ACS Catal. 7 (2017) 2770–2779.
[69] C.A. Wolcott, A.J. Medford, F. Studt, C.T. Campbell, Degree of rate control ap-
proach to computational catalyst screening, J. Catal. 330 (2015) 197–207.
[70] S. Chen, B. Yang, Theoretical understandings on the unusual selectivity of 1,3-
Butadiene hydrogenation to butenes over gold catalysts, Catal. Today (2018).
[71] B. Wang, S. Chen, J. Zhang, S. Li, B. Yang, Propagating DFT uncertainty to me-
chanism determination, degree of rate control, and coverage analysis: the kinetics
of dry reforming of methane, J. Phys. Chem. C 123 (2019) 30389–30397.
[72] Y.A. Zhu, D. Chen, X.G. Zhou, W.K. Yuan, DFT studies of dry reforming of methane
on Ni catalyst, Catal. Today 148 (2009) 260–267.
[73] L. Foppa, M.C. Silaghi, K. Larmier, A. Comas-Vives, Intrinsic reactivity of Ni, Pd and
Pt surfaces in dry reforming and competitive reactions: insights from first principles
S. Chen, et al.
calculations and microkinetic modeling simulations, J. Catal. 343 (2016) 196–207.
[74] C. Fan, Y.A. Zhu, M.L. Yang, Z.J. Sui, X.G. Zhou, D. Chen, Density functional theory-
assisted microkinetic analysis of methane dry reforming on Ni catalyst, Ind. Eng.
Chem. Res. 54 (2015) 5901–5913.
[75] S. Chen, J. Zaffran, B. Yang, Descriptor design in the computational screening of Ni-
based catalysts with balanced activity and stability for the dry reforming of me-
thane reaction, ACS Catal. 10 (2020) 3074–3083.
[76] B. AlSabban, L. Falivene, S.M. Kozlov, A. Aguilar-Tapia, S. Ould-Chikh, J.-
L. Hazemann, L. Cavallo, J.-M. Basset, K. Takanabe, In-operando elucidation of
bimetallic CoNi nanoparticles during high-temperature CH4/ CO2 reaction, Appl.
Catal. B 213 (2017) 177–189.
[77] J. Horlyck, C. Lawrey, E.C. Lovell, R. Amal, J. Scott, Elucidating the impact of Ni
and Co loading on the selectivity of bimetallic NiCo catalysts for dry reforming of
methane, Chem. Eng. J. 352 (2018) 572–580.
[78] M.S. Aw, I.G. Osojnik Črnivec, P. Djinović, A. Pintar, Strategies to enhance dry
reforming of methane: synthesis of ceria-zirconia/nickel–cobalt catalysts by freeze-
drying and NO calcination, Int. J. Hydrog. Energy 39 (2014) 12636–12647.
[79] X. Du, L.J. France, V.L. Kuznetsov, T. Xiao, P.P. Edwards, H. AlMegren, A. Bagabas,
Dry reforming of methane over ZrO2-supported Co–Mo carbide catalyst, Appl.
Applied Catalysis B: Environmental 270 (2020) 118859
Petrochem. Res. 4 (2014) 137–144.
[80] J.P. Tessonnier, D.S. Su, Recent progress on the growth mechanism of carbon na-
notubes: a review, ChemSusChem 4 (2011) 824–847.
[81] P. Sautet, F. Cinquini, Surface of metallic catalysts under a pressure of hydrocarbon
molecules: metal or carbide? ChemCatChem 2 (2010) 636–639.
[82] K. Nagaoka, K. Takanabe, K. Aika, Influence of the reduction temperature on cat-
alytic activity of Co/TiO2 (anatase-type) for high pressure dry reforming of me-
thane, Appl. Catal. A-Gen. 255 (2003) 13–21.
[83] J.L. Ewbank, L. Kovarik, C.C. Kenvin, C. Sievers, Effect of preparation methods on
the performance of Co/Al2O3 catalysts for dry reforming of methane, Green Chem.
16 (2014) 885–896.
[84] L. Ji, S. Tang, H.C. Zeng, J. Lin, K.L. Tan, CO2 reforming of methane to synthesis gas
over sol–gel-made Co/γ-Al2O3 catalysts from organometallic precursors, Appl.
Catal. A Gen. 207 (2001) 247–255.
[85] D. San José-Alonso, M.J. Illán-Gómez, M.C. Román-Martínez, K and Sr promoted Co
alumina supported catalysts for the CO2 reforming of methane, Catal. Today 176
(2011) 187–190.
[86] Z. Hou, T. Yashima, Supported Co catalysts for methane reforming with CO2, React.
Kinet. Catal. Lett. 81 (2004) 153–159.