Kinetic consequences of methane combustion on Pd, Pt and Pd-Pt catalysts

Article abstract of DOI:10.1039/c6ra21150j

The kinetic consequences of methane combustion on Pd_xPt_1-x oxide surfaces are investigated with kinetic data and density functional theory treatments. The catalytic activity of monometallic palladium drops significantly with time, and this loss is larger at higher temperatures. However, the addition of platinum to palladium catalysts significantly improves their stability. Besides, the Pd-rich bimetallic catalysts show higher activity for methane conversion. Turnover rates are independent of O₂ pressure (>10 kPa) but depend linearly on CH₄ pressure for all the catalysts. The calculated C-H bond activation energies are almost identical for Pd_1.0, Pd_0.75Pt_0.25 and Pd_0.5Pt_0.5. However, the Pt-rich catalysts (Pd_0.25Pt_0.75) show poor activity, and their activation energies and pre-exponential factors increase with increase in the platinum content of the catalyst. In these catalysts, methane dissociation is the only kinetically relevant step, but it occurs on different active sites. H₂O and CO₂ (>3-5 kPa) strongly inhibit methane combustion by titrating surface vacancies in a quasi-equilibrated adsorption-desorption step for Pd-containing catalysts. The DFT-derived C-H bond activation energies for O-saturated Pt surfaces are much larger than the values for PdO(101) surfaces, which is in agreement with the experimental results. Models of Pd-Pt oxide surfaces were built on the basis of XRD results and known structures of oxide Pd and Pt monometallic catalysts. Although the atom arrangement of the PdO(101)/Pt(100) outermost layer is similar to that of PdO(101), the C-H bond activation energy for PdO(101)/Pt(100) is much larger than that of PdO(101) because the surface Pd atoms on PdO(101)/Pt(100) are highly coordinated. However, when the Pd content is high enough (>25 mol%) in the Pd-Pt bimetallic catalysts, the low PdO(101) activation energy (61 kJ mol^-1) is retrieved for two monolayers of PdO(101) on Pt(100) (67 kJ mol^-1).

Full text of DOI:10.1039/c6ra21150j

View Article Online
View Journal
RSC Advances
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: W. Qi, J. Ran, X.
Du, R. Wang, J. Shi, J. Niu, P. Zhang and M. Ran, RSC Adv., 2016, DOI: 10.1039/C6RA21150J.
This is an Accepted Manuscript, which has been through the
Royal Society of Chemistry peer review process and has been
accepted for publication.
Accepted Manuscripts are published online shortly after
acceptance, before technical editing, formatting and proof reading.
Using this free service, authors can make their results available
to the community, in citable form, before we publish the edited
article. We will replace this Accepted Manuscript with the edited
and formatted Advance Article as soon as it is available.
You can find more information about Accepted Manuscripts in the
author guidelines.
Please note that technical editing may introduce minor changes
to the text and/or graphics, which may alter content. The journal’s
standard Terms & Conditions and the ethical guidelines, outlined
in our author and reviewer resource centre, still apply. In no
event shall the Royal Society of Chemistry be held responsible
for any errors or omissions in this Accepted Manuscript or any
consequences arising from the use of any information it contains.
rsc.li/rsc-advances

Page 1 of 34
RSC Advances
View Article Online
DOI: 10.1039/C6RA21150J
Kinetics Consequences of Methane Combustion on Pd, Pt and Pd-Pt
Catalysts
Wenjie Qia, b, Jingyu Rana, b,∗, Ruirui Wanga, b, Xuesen Dua, b, Jun Shia, b
,
Juntian Niua, b, Peng Zhanga, b Mingchu Ranb
a Key Laboratory of Lowꢀgrade Energy Utilization Technologies and Systems,
Ministry of Education of PRC, Chongqing University, 174 Shazheng Street,
Shapingba District, Chongqing, 400044, China
b College of Power Engineering, Chongqing University, Chongqing 400030, China
Abstract: Kinetics consequences of methane combustion on PdxPt1ꢀx oxide surfaces
were investigated with kinetic data and density functional theory treatments. The
activity over monometallic palladium dropped significantly with time, and the loss
was larger at higher temperature. However, the addition of platinum into palladium
catalysts significantly improves its stability. Besides, the Pdꢀrich bimetallic catalysts
showed higher activity for methane conversion. Turnover rates independent of O2
pressure (>10 kPa) but depend linearly on CH4 pressures for all catalysts. Calculated
CꢀH bond activation energies are almost identical for Pd1.0, Pd0.75Pt0.25 and Pd0.5Pt0.5.
On the contrary, the Ptꢀrich catalysts (Pd0.25Pt0.75) showed poor activity, activation
energies and preꢀexponential factors increase with the platinum content in catalyst. In
these catalysts, methane dissociation is the sole kinetically relevant step, but occurs on
different active sites. H2O and CO2 (>3ꢀ5 kPa) strongly inhibits methane combustion
∗ Corresponding Author
Telephone: +86ꢀ 18996085138; Fax: +61ꢀ 23ꢀ65111832. Eꢀmail: Ranjy@cqu.edu.cn (Jingyu Ran).
1

RSC Advances
Page 2 of 34
View Article Online
DOI: 10.1039/C6RA21150J
by titrating surface vacancies in a quasiꢀequilibrated adsorption–desorption step for
the Pd contained catalysts. DFTꢀderived CꢀH bond activation energies for Oꢀsaturated
Pt surfaces are much larger than the values for PdO(101) surfaces, which are in
agreement with the experimental results. Models of PdꢀPt oxide surfaces were built on
the basis of XRD results and known oxide Pd and Pt monometallic structures.
Although the atom arrangement of PdO(101)/Pt(100) outermost layer is similar to that
of PdO(101), CꢀH bond activation energy for PdO(101)/Pt(100) is much larger than
the value for PdO(101) since the surface Pd atoms on PdO(101)/Pt(100) are high
coordinated. However, when the Pd content is high enough (> 25 mol%) in PdꢀPt
bimetal catalysts, the low PdO(101) activation energy (61 kJ molꢀ1) is retrieved in the
case of two monolayers of PdO(101) on (100) (67 kJ molꢀ1).
Key words: methane combustion, bimetallic PdꢀPt catalysts, kinetics, DFT
2

Page 3 of 34
RSC Advances
View Article Online
DOI: 10.1039/C6RA21150J
1. Introduction
Catalytic combustion is widely used for methane elimination and power generation
without NOx emission[1, 2]. Noble metal catalysts were found to be high activity
despite its high cost. The noble metals involve Pt, Pd, Ru, Rh, Ir and Ag, etc, in which
Pt owns the highest catalytic activity [3ꢀ7]. Moreover, the Palladium catalyst is also
widely used for methane oxidation due to its high activity at low temperature.
However, the palladium catalyst has poor stability at high temperature, which limits
its practical application.
Many researchers found that PdOx is the active phase for the methane oxidation
reaction despite the most active phase has been debated over the years between the
metal and the oxide [8ꢀ11]. However, when the temperature reaches to around 973 K
depending on the support material and the atmosphere, the palladium oxide
decomposes into metallic palladium. Besides, the activity of Pd/Al2O3 catalyst for
methane combustion was unstable even at very low operation temperature [9, 12].
Narui et al. [11] reported that the poor stability of monometallic Pd catalyst mainly
attributed to sintering of PdO particles during operation. However, Van Giezen et al.
[13] considered that the formation of Pd(OH)2 during methane combustion causes the
deactivation of Pd/Al2O3 catalyst.
The addition of a second metal may be beneficial to improve the stability of
palladium catalysts [12, 14ꢀ17]. Persson et al. [12] reported the influence of coꢀmetals
on palladium catalysts for methane oxidation. They found that bimetallic PdꢀPt
catalyst is the most promising catalyst that owns both high activity and stability.
3

RSC Advances
Page 4 of 34
View Article Online
DOI: 10.1039/C6RA21150J
Besides, a large number of reports also considered that bimetallic PdꢀPt catalyst
extend the activity time for methane combustion compare with monometallic Pd
catalyst. They found that the effect was attributed to the suppression of the particle
growth [11]. Some experimental results also showed that PdꢀPt bimetal catalyst owns
the higher activity than monometallic Pd catalyst especially at high temperature. The
activity of the bimetal catalysts are highly depend on molar ratio content of Pd:Pt.
Lapisardi et al. [18] found that the high Pd content Pd0.93Pt0.07/Al2O3 owns the highest
activity for methane combustion. Moreover, Persson et al. [17] reported that the most
promising catalysts are Pd0.67Pt0.33 and Pd0.5Pt0.5 with respect to both stability and
activity. However, the experimental results of Kinnunen et al. [19] have shown that
the activity of the fresh Pd is higher than that of the fresh Pd0.85Pt0.15. Besides, almost
all researches verified that platinumꢀrich catalysts own very poor activity for methane
combustion.
In order to predict the effect of the pressure, the composition and the temperature on
methane conversion, it is necessary to develop a kinetic model. Generally, the
elementary steps of the fist CꢀH bond activation in CH4 as the solo kinetically relevant
step, especially when the catalyst surface was saturated covered by oxygen atom
[20ꢀ22]. Many studies were reported in the literature about the methane kinetics over
Pd catalyst. However, the mechanisms of methane combustion over it is still a debate
issue due to its structure is complicated and unstable.
The mechanisms of Langmuir Hinshelwood [23], EleyꢀRideal [24] and Mars van
Krevelen [8, 10, 25] have been proposed for Pd catalyst. The structure of Pt catalyst is
4

Page 5 of 34
RSC Advances
View Article Online
DOI: 10.1039/C6RA21150J
more stable and simpler than that of Pd catalyst due to the bulk oxide of Pt catalyst is
not exist during typical methane combustion[20, 26]. Hence, the surface reactions of
the methane combustion on Pt are always limited to Langmuirian treatment [20, 27].
Even though the kinetic mechanisms of monometallic Pd and Pt catalysts have
substantially studied, the kinetic mechanisms of bimetallic PdꢀPt catalysts are still
deficient. The purpose of this paper is to get insight into the kinetic mechanisms for
methane combustion on bimetallic PdꢀPt catalysts.
2. Experimental and theoretical methods
2.1 Catalyst preparation and characterization
Bimetallic catalysts on γꢀAl2O3 (0.45 cm3 gꢀ1 proe volume, 153 mꢀ2 g BET surface
area) were prepared by incipient wetness impregnation with aqueous solutions which
were manufactured by coꢀimpregnation of PtCl4 and Pd(NO3)2 solution. The
bimetallic catalysts were formulated to contain equal molar amounts of Pd, Pt and
PdꢀPt. The details of these catalysts were summarized in Table 1. The γꢀAl2O3
(0.125ꢀ0.178 mm diameter) was treated by heating to 973 K at 5 K minꢀ1 and holding
for 4 h in flowing dry air before impregnation. Metal solutions were dripped onto the
γꢀAl2O3 powders and carefully mixed. This procedure was repeated thrice, with a
drying step in flowing air at 373 K for 10 h in between. Thereafter these powers were
calcined in flowing dry air by heating to 923 K at 5K minꢀ1 for 3 h. Xꢀray diffraction
(XRD) on the catalyst powers was carried out using the same instrument and the data
were obtained using CuKα radiation.
5

RSC Advances
Page 6 of 34
View Article Online
DOI: 10.1039/C6RA21150J
The PdꢀPt dispersion (fraction of Pd and/or Pt atoms residing at cluster surfaces) in
these samples were measured by CO chemisorption. The samples of 1.0 g were
preheated under H2 at a flow rate of 200 cm3 minꢀ1 and then heated to 723 K at a
heating rate of 10 K minꢀ1 for 1.5 h. These samples then cooled to ambient
temperature under He condition. Thereafter a dualꢀisotherm method was then used
for the CO chemisorption analyses. Furthermore, the dispersions of these samples
were determined by using a stoichiometric factor of 1 for Pt and 2 for Pd [14, 28,
29].
Table 1 The metal loading, the CO uptake and the metal cluster dispersion of the different
catalysts
CO uptake (ꢁmol
(gꢀ1 catalysts))
PdꢀPt dispersion
Pd loading Pt loading
Sample
Composition
(%)
(wt%)
(wt%)
from CO uptake
Pd1.0Pt0
Pd0.75Pt0.25
Pd0.50Pt0.50
Pd0.25Pt0.75
Pd0Pt1.0
1:0
2.1
1.6
1.0
0.6
0
0
9.5
11.3
9.8
9.6
9.3
7.3
6.2
8.8
0.75:0.25
0.5:0.5
0.25:0.75
0:1
0.9
1.7
2.6
3.4
10.0
15.0
2.2 Catalytic rate measurement
Methane combustion rates were measured in a quartz tube (8.0 mm inner diameter)
fitted with a Kꢀtype thermocouple. CO2 and CH4, concentrations were measured by
gas chromatography. The catalysts were mixed with γꢀAl2O3 (treated in flowing dry
air by heating to 923 K for 5 h at 5 K minꢀ1) to obtain a 50 intraꢀparticle
diluentꢀtoꢀcatalyst ratio (λ) for Pd/Al2O3 (2.0 wt% Pd). Mass transfer and
Interꢀparticle heat restrictions were ruled out by diluting PdxPt1ꢀx/γꢀAl2O3 with a 27
excess of γꢀAl2O3. In order to obtain the same residence time for gas flow, the mass
6

Page 7 of 34
RSC Advances
View Article Online
DOI: 10.1039/C6RA21150J
of catalysts mixtures for methane combustion were 0.17 g. As a result, Catalyst
mixtures that contain 0.13ꢀ0.17 mg of metal atoms. Besides, activity and stability of
the catalysts in methane combustion were measured under lean burn conditions.
Kinetic measurements and the rate data were carried out at less than 10% methane
conversions.
2.3 Density functional theory methods
DFT calculations were performed with the program package Cambridge Sequential
Total Energy Package 8.0 (CASTEP 8.0) [30, 31]. All DFT calculations were carried
out spinꢀpolarized within the generalized gradient approximation (GGA) using the
PerdewꢀBurkeꢀErnzerhof (PBE) and ultrasoft pseudopotential (USꢀPP) to describe
interaction between core and valence electrons. Plane waves were included for the
electronic wave functions up to a cutoff energy of 380 eV. The MonkhorstꢀPack (MP)
meshes 6×6×1 kꢀpoints for all surfaces of the catalysts. The convergence criteria for
structure optimization was set to the tolerance for the selfꢀconsistent field (SCF),
energy, and maximum force of 2.0×10ꢀ6 eV/atom, 2.0×10ꢀ5 eV/atom, 0.05 eV/Å and
2.0×10ꢀ3 Å, respectively. A Fermi smearing of 0.2 eV was utilized. Unless stated
otherwise, A p(2×2) superꢀcell was employed for the metal substrate, which had been
commonly used in the previous studies. In all simulations, the vacuum gap is larger
than 12 Å between the slab surface models.
Finally, the linear synchronous transit/quadratic synchronous transit (LST/QST)
method was utilized to locate the TS [32].All TSs were converged with a tolerance of
0.15 eV Åꢀ1 in the gradient root mean square.
7

RSC Advances
Page 8 of 34
View Article Online
DOI: 10.1039/C6RA21150J
3.Results and discussion
3.1 XRD and activity tests
Figure 1. XRD patterns of the catalysts samples (a) Pd1.0Pt0, (b) Pd0.75Pt0.25, (c) Pd0.5Pt0.5,(d)
Pd0.25Pt0.75,(e) Pd0Pt1.0.
The effect of Pt addition on the XRD of the Pd /Al2O3 catalyst and the characteristic
reflections of the PdyPt1ꢀyO, Pt, Pd, PdPtꢀalloys and alumina support were presented in
Fig. 1. The monometallic Pd/Al2O3 catalysts is similar to those of other studies [14,
33]. Addition of Pt causes a new reflection at 2θ with a range of 39.7°−40.0°. The
8

Page 9 of 34
RSC Advances
View Article Online
DOI: 10.1039/C6RA21150J
PdyPt1ꢀyO and PdPt alloy phases can be observed in the Pd0.5Pt0.5, which coexist in the
PdꢀPt catalysts [17]. Significantly, 2θ of PdyPt1ꢀyO decreses with the amount of Pt
increasing, which can be attributed to the lager atomic diameter for platinum atoms.
Besides, no any PdyPt1ꢀyO and Pd phases were detected in Pd0.25Pt0.75/Al2O3. Instead,
the alloys (111) and (100) are the most abundant facets in Pd0.25Pt0.75/Al2O3, which is
similar to Pt/Al2O3.
9

RSC Advances
Page 10 of 34
View Article Online
DOI: 10.1039/C6RA21150J
Figure 2. (a)Activity tests and for PdxPt1ꢀx/Al2O3 (Pd1.0 (△), Pd0.75Pt0.25(◆), Pd0.5Pt0.5(▲),
Pd0.25Pt0.75 (◄) and Pt1.0 (►)) during heating (400Kꢀ873 K, 5 K/min, 2kPa CH4 and 20 kPa O2,
3.34 ml sꢀ1), (b)Steadyꢀstate tests for PdxPt1ꢀx/Al2O3 (723 K, 2 kPa CH4 and 20kPa O2). (c)
Steadyꢀstate tests for PdxPt1ꢀx/Al2O3 (873 K, 2 kPa CH4 and 20 kPa O2). All of the catalysts were
not diluted with Al2O3.
The results from the activity tests were displayed in Fig. 2. Activity of the catalyst
decreases with increasing Pt content. In addition, the addition of Pt also can greatly
improves the durability of the catalyst. Pt/Al2O3 owns the lowest activity even when
the temperature reaches up to 873 K, since Pt is a less active catalyst for methane
combustion at low temperature [34]. However, a small amount of Pd addtion to a Pt
catalyst has beneficial effect on the activity, which is different from the result of
Pesson [17]. Fig. 2c shows the methane conversion on PdꢀPt catalysts at 873 K as a
function of time on stream. The methane conversion on Pd/Al2O3 catalyst experienced
a sharp decrease with time on stream while that on bimetal PdꢀPt/Al2O3 catalysts only
underwent a slight decrease.Significantly, the activity of Pd/Al2O3 catalyst was
actually lower than that of Pd0.5Pt0.5/Al2O3 after only 50 min for the samples
10

Page 11 of 34
RSC Advances
View Article Online
DOI: 10.1039/C6RA21150J
maintained at 873 K. However, the time increases to 600 min when the operating
temperature is 723 K, as shown in Fig.2b.
Many researchers have reported that the surface composition of PdꢀPt catalysts is
generally different from the bulk composition due to the segregation processes
[35ꢀ37]. Pd surface segregation enhanced in oxygen atmosphere by heating the alloy
to 673 K, as was illustrated by Van den Oetelaar et al [35]. In many cases, it leads to
the formation of superficial oxide phases. Besides, the bulk composition of the PdPt
alloys are metallic, which makes the bimetallic PdꢀPt catalysts more stable than
monometallic Pd catalyst. Hence, the stabilities of PdꢀPt catalysts increase with the
increasing Pt content.
3.2 Kinetically relevant of methane combusition on PdxPt1ꢀxOδ surfaces
3.2.1 Effects of O2 and CH4 concentrations on methane combustion
Figure 3. Firstꢀorder rate constant for methane combustion as a single valued function of O2
11

RSC Advances
Page 12 of 34
View Article Online
DOI: 10.1039/C6RA21150J
pressure on PdꢀPt surface with 2.0 kPa (◆), 3.5 kPa (▲), and 4.5 kPa (△) CH4 and the pressure of
O2 is 10 kPa<O2<35 kPa. Color code: green, pink, blue, red and black for Pd1.0Pt0.0
(723K) ,Pd0.75Pt0.25 (723K) ,Pd0.5Pt0.5 (723K) ,Pd0.25Pt0.75 (873K) and Pd0Pt1.0 (873K), repectively.
Turnover rates for CH4 oxidation were measured over a broad range of O2 (10−35
kPa) and CH4 (2.0−4.5 kPa) pressures. Fig. 3 shows the effect of O2 pressure on
firstꢀorder constant with O2 pressure. In order to maintain the oxidation turnover rates
did not vary with time, all catalysts were kept at reaction conditions over 14 h,
indicating that the surfaces of catalyst clusters did not reconstruct during catalytic
measurements [25].
Turnover rates ultimately reached the constant values (>14 kPa) of 3.3, 6.9, 3.6, 4.2
and 1.1 sꢀ1 siteꢀ1 (kPa CH4)ꢀ1 for Pd1.0, Pd0.75Pt0.25, Pd0.5Pt0.5, Pd0.25Pt0.75 and Pd0Pt1.0,
respectively. The surface oxygen atoms gradually dissolved in the bulk of the catalyst
with the increasing O2 pressure (<14 kPa). The Pd→PdO transitions and the the
kinetic consequences of the Pd catalyst already have been reported for the dispersed
Pd clusters [25] and Pd foils [38]. Turnover rates for methane combustion on these
catalysts proportional to CH4 pressure and independent of O2 pressure, reflecting the
firstꢀorder rate constants for methane combustion.
0
rCH = kP 1P
CH4 O2
4
(1)
The reaction rate increases with the increasing Pd content for all catalysts expect for
the Pd catalyst due to its poor stability. The activation barriers and preexponential
factors for methane combustion were calculated by the Arrhenius equation:
12

Page 13 of 34
RSC Advances
View Article Online
DOI: 10.1039/C6RA21150J
ꢀE
RT



(2)
k = A exp −
0



Figure 4. Arrhenius plots of the methane turnover rates versus 1000/T for methane combustion on
catalysts surfaces. Rꢀsquare for these experimental data are larger than 0.97.
Regression fitting of the temperature dependence rate constant from experimental
results were obtained with samples were shown in Fig. 4. The kinetic parameters for
these catalysts listed in Table 2. The apparent activation energies for methane
combustion on Pd1.0, Pd0.75Pt0.25 and Pd0.5Pt0.5 are almost idential (Table 2). This
should be due to these catalysts own a similar surface texture for the formation of
coꢀadsorbates H3C*−OH* from CH4 (g) chemisorption dissociation. Activation energy
and preꢀexponential factor for methane combustion on Pd0.25Pt0.75 are significantly
different with both Pd and Pt catalysts. Hence, the PdꢀPt surface models should be
reconstructed based on Pd and Pt catalysts.
Table 2 Measured apparent activation barrier energies and preꢀexponential factors for CꢀH bond
13

RSC Advances
Page 14 of 34
View Article Online
DOI: 10.1039/C6RA21150J
activation in methane combustion on these catalysts.
Sample
Rate expression
Apparent activation
Preꢀexponential factor (sꢀ1 kPaꢀ1)
energy (kJ molꢀ1)
Pd1.0Pt0
65
62
2.9×105
4.1×105
2.2×105
9.3×106
8.4×109
Pd0.75Pt0.25
Pd0.5Pt0.5
Pd0.25Pt0.75
Pd0Pt1.0
r=kx,1ꢀx(CH4)1.0(O2)0
62
101
160
3.2.2 Effects of CO2 and H2O concentrations on methane combustion
14

Page 15 of 34
RSC Advances
View Article Online
DOI: 10.1039/C6RA21150J
Figure 5. Effects of (a) H2O and (b) CO2 on turnover rates for Pd1.0 (△), Pd0.75Pt0.25(◆) and
Pd0.5Pt0.5(▲) at 723 K, as well as Pd0.25Pt0.75 (◄) and Pt1.0 (►) at 873K with 2 kPa CH4. The
orders with respect to H2O or CO2 pressures were shown in Table 3.
The effects of CO2 and H2O concentrations on methane combustion rates were
shown in Fig. 5. The presence of H2O in the feed strongly inhibits the methen
combustion on Pd1.0, Pd0.75Pt0.25, Pd0.5Pt0.5 and Pd0.25Pt0.75, as shown in Fig. 5a.
However, H2O inhibition becomes negligible even at high H2O concentration (15 kPa)
for Pt/Al2O3. The orders of H2O for methane combuation on Pd1.0, Pd0.75Pt0.25
,
Pd0.5Pt0.5 Pd0.25Pt0.75 and Pt1.0 were ꢀ1.03, ꢀ1.05, ꢀ0.98, ꢀ0.79. and ꢀ0.04 (Table 3),
respectively. Besides, it was found that H2O removal led to the most recovery of
initial turnover rates for these catalysts. Hence, these effects of H2O and the alomot
complete recovery upon H2O removal indicate that the decreases reflect kinetic
inhibition processes caused by competitive adsorption.
Table 3. Reaction orders of H2O and CO2 on methane combustion.
Samples
Orders(α) of
H2O
Rate expression
Pd1.0Pt0 Pd0.75Pt0.25 Pd0.5Pt0.5 Pd0.25Pt0.75 Pd0Pt1.0
ꢀ1.03
ꢀ1.92
ꢀ1.05
ꢀ1.83
ꢀ0.98
ꢀ1.89
ꢀ0.79
ꢀ0.04
r=kx,1ꢀx(CH4)1.0(H2O)α
Orders (β) of
CO2
ꢀ1.39
ꢀ0.58
r=kx,1ꢀx(CH4)1.0(CO2)β
CO2 has no inhibitory effects on CH4 combution for Pd1.0, Pd0.75Pt0.25 and Pd0.5Pt0.5
15

RSC Advances
Page 16 of 34
View Article Online
DOI: 10.1039/C6RA21150J
except when CO2 concentrations were enough high (>3~5 kPa). An orders
dependence of ꢀ1.92, ꢀ1.83,ꢀ1.89 and ꢀ1.39 are determined for Pd1.0, Pd0.75Pt0.25
,
Pd0.5Pt0.5 and Pd0.25Pt0.75, respectively. Hence, CO2 at high concentration owns
overwhelming competition edge for the active sites compare to CH4, H2O and O2,
namely the adsorbed CO2* or CO3* becomes the masi species.
Two different sequence of elementary steps and describe how kinetically relevant
steps and surface intermediates evolve as CH4, H2O, CO2 and O2 pressures change,
which were proposed in Scheme 1 and Scheme 2. Where (*) stands for oxygen
vacancy and vacancy of Pdꢀlike vs Ptꢀlike surfaces, respectively. The combustion of
methane over Pdꢀlike catalysts follows a Marsꢀvan Krevelen (MvK) mechansim,
while the kinetic models used in this study for the combustion of methane over Ptꢀlike
catalsyts correspond to two types: Langmuir Hinshelwood model (LH) and
EleyꢀRideal model (EꢀR).
Scheme 1. Methane dissociation on (a) Metal−O sites pairs, (b) surfaces saturated with
chemisorbed oxygen O*−O* pairs.
16

Page 17 of 34
RSC Advances
View Article Online
DOI: 10.1039/C6RA21150J
Scheme 2 . Elementary steps for CH4ꢀO2 reaction on supported on Pdꢀlike catalyst and Ptꢀlike catalyst.
Kinetic model
(a and b)
Step 1
Pdꢀlike
Rate and equilibrium
Ptꢀlike
Rate and equilibrium
(MvK)
constant
K1a
(LH and EꢀR)
constant
K1b
*
*
O2+*⇌O2
O2+*⇌O2
*
*
Step 2
Step 3
Step 4
O2 +*→2O*
k2a
O2 +*⇌2O*
k2bf,k2br
k3b
CH4+*⇌CH4
K3a
CH4+2O*→CH3O*+OH*
*
CH4 +O*→CH3 +OH*
k4.1a
*
*
*
*
CH3 +O*→CH2 +OH*
CH3O*+O*→CH2O*+OH*
CH2O*+O*→CHO*+OH*
CHO*+O*→CO*+OH*
CH2 +O*→CH*+OH*
*
CH*+O*→C+OH*
C*+O*⇌CO*+*
K4.2a
K5a
*
*
Step 5
CO*+O*⇌CO2 +*
CO*+O*⇌CO2 +*
K5b
K6.1b
K6.2b
K7b
*
*
Step 6.1
Step 6.2
Step 7
CO2 ⇌CO2+*
K6.1a
K6.2a
K7a
CO2 ⇌CO2+*
CO2+O* ⇌ CO3*
OH*+OH*⇌H2O*+O*
H2O*⇌H2O+*
CO2+O* ⇌ CO3*
OH*+OH*⇌H2O*+O*
H2O*⇌H2O+*
Step 8
K8a
K8b
In the case of Pdꢀlike catalysts, the elementary reaction steps for methane
combustion include the dissociation of O2 onto two oxygen vacancies. Besides, the
17

RSC Advances
Page 18 of 34
View Article Online
DOI: 10.1039/C6RA21150J
rateꢀdetermining step (RDS) involves the dissociation of CH4 on a matched pair sites
which consists of an unsaturated metal site and an adjacent O* (Step 3a and 4a). The
dissociative chemisorption of O2 for Pdꢀlike Catalysts (Step 2a) was assumed to be
irreversible based on the negligible rate of 16O2ꢀ18O2 equilibration during methane
combustion below 800 K [39]. Besides, the steps of CH4 dissociation are irreversible
when OH* species are the most abundant surface intermediate (MASI). Since these
steps are the only several irreversible steps involving the MASI. Subsequently,
surface vacancies (*) are regenerated by the quasiꢀequilibrated desorption of CO2 and
H2O. Besides, when the concentration of CO2 is high enough, CO2 and CH4 compete
for O* that has also been taken into account. Then, a psedudoꢀsteadyꢀstate treatment
of methane combustion on Pdꢀlike catalysts leading to a complex kinetic rate
expression (derived in Supplementary material):
r =
a
k2a K1a P
O2

2


P
P P
P
k6.2ak2a K1a P
O2
k2a K1a
k2a K1a
1
O2
O2 H2O
H2O
21+ K1a P +
+ K3a P
+
+
+
+
P
CO2







O2
CH4

2k4.1ak3a
P
2K7a K8ak4.1ak3a
P
K8a
K6a 2k4.1ak3a P
CH4
CH4
CH4



(3)
This rate equation can be simplified to the experimental rate expression when OH*
are the MASI:
P
CH4
(4)
r
= K7a K8ak4.1ak3a
a−OH*−MASI
P
H2O
Hence, the methane combustion rate becomes proportional to PH2Oꢀ1.When adsorbed
CO2* or CO3* becomes the most abundant surface intermediates, equation (3) can
also be reduced to
18

Page 19 of 34
RSC Advances
View Article Online
DOI: 10.1039/C6RA21150J
k2a K1a P
1
O2
r
=
(5)
a−CO2*+CO3*−MASI
2
P

2
CO2
k6.2ak2a K1a P
1
O2
2
+




K6a 2k4.1ak3a P
CH4


Then, the turnover rate becomes proportional to PCO2ꢀ2, but only at CO2 pressure
obove which CO2 inhibition term become larger than that of H2O. As a result,
kinetics mechanism of methane combustion on catalysts Pd1.0Pt0.0 , Pd0.75Pt0.25 and
Pd0.5Pt0.5 are similar to each other. Significantly, the experimental results show that the
orders for Pd0.25Pt0.75 with respect to CO2 and H2O pressures were ꢀ1.39 and ꢀ0.79,
respectively. This is because the competitive adsorption relationship between OH*
and CO3* (or CO2*). Therefore, kinetically relevant steps and the MASI for
Pd0.25Pt0.75 depend on OH* and CO3* (or/and CO2*) coverages, which are set, in turn,
by both the H2O pressure and the CO2 pressure.
The CH4 turnover rate independent of O2 pressure suggests that the surfaces of Pt1.0
are saturated with chemisorbed O atoms. Moreover, because the vacancies will be
immediately occupied by O atoms after releasing CO2 and H2O, so CH4 activation
steps are assisted by two adjacent O*, to form CHxO (x, 4−0) and OH* (Step 3b).
Besides, the initial Hꢀabstraction step is treated as RDS due to the high stability of
methane and its weak interaction with surfaces [27, 40]. The O2(g) cannot directly
dissociate to form 2O* on an isolated vacancy (*). Instead, it adsorbs as O2* in
equilibrated steps (Step 1b), and then dissociates onto a vicinal O* site resluting in the
formation of a new O2*. This process of O2* activation by vicinal O* allows for the
rapid migration of O* along saturated O* surface until O2* finds a vacancy site (*)
(Step 2b) [41]. Hence, a sequence of elementary steps for Ptꢀlike catalyst was
19

RSC Advances
Page 20 of 34
View Article Online
DOI: 10.1039/C6RA21150J
proposed in Scheme 2 (b), including elementary reactions to form CO2 and H2O. The
releases of H2 and CO were not taken into consideration in this paper due to the high
oxygen pressure. Besides, any H2 and CO were not detected even at higher methane
concentration (PCH4=12 kPa). Then, the CH4 combustion rate including CH4, O2, H2O
and CO2 can be described by a complex rate equation (derived in Supplementary
material)
k3bk2bf K1b
r =
P P
×
b
O2 CH4
2k3b P + k2br
CH4
1

2



k2bf K1b P
P
k2bf K1b P
k2bf K1b P
P
H2O
1
O2
H2O
O2
O2

1+ K1b P +
+
+ 

+ K6.2b
 P
+
O2
CO2





2k3b P + k2br
K7b K8b 2k3b P + k2br
K6.1b
2k3b P + k2br
K8b
CH4
CH4
CH4


(6)
When the O* is the MASI, simplifies eq 5 to the observed experimental rate
equation (eq 1)
(7)
r
= k3b P
b−O*−MASI
CH4
In this study, methane combustion rates on oxide Pt catalyst depend weakly on H2O,
therefore, the product H2O can easily flee away from the saturated O* surface even at
high H2O concentration. When CO2 pressure reaches up to around 8 kPa, an apparent
order of ꢀ0.58 with respect to CO2 was obtained, which was much less than the order
of ꢀ2 if the CO2* (or/and CO3*) were assumed as the MASIs. Hence, the Pt catalyst
surfaces are mainly covered by O*, CO2* or/and CO3* only at CO2 pressures above 8
kPa, in which case of reaction rate is given by
k3bk2bf K1b
r
=
P P
O2 CH4
b−O*+CO2 *+CO3*−MASI


2



k2bf K1b P
k2bf K1bP
O2
1
O2
2k P + k2br
+ 

+ K6.2b
 P
(
)
3b CH4
CO2





2k3b P + k2br
K6.1b
2k3b P + k2br
CH4
CH4


20

Page 21 of 34
RSC Advances
View Article Online
DOI: 10.1039/C6RA21150J
(8)
3.3 Theoretical treatments of CꢀH bond activation on PdxPt1ꢀx surfaces at high
oxygen pressure
3.3.1 CꢀH bond activation on oxide Pd and Pt monometallic catalysts
Figure 6. Sideꢀ and topꢀviews of the optimized structure of different oxide layers. (a) Pt(111)
metal surface covered with 1ML oxygen, (b) Pt(100) metal surface covered with 1ML oxygen,(c)
PdO(101), (d) PdO(100) (O, read spheres; Pd, green spheres; Pt, blue spheres).
The calculations in this study firstly concerning the activation barriers of CH4
chemisorption dissociation on pure Pd and Pt oxide catalysts. The calculations were
performed for the case of subsurface oxygen covergae of 1ML for Pt catalyst due to
the methane combustion process was carried out in high oxygen pressure. Pt surfaces
are covered with 1ML of oxygen, thus requiring methane chemisorption dissociation
on O*ꢀO*. Kinetically relevant CꢀH activation steps are assisted by O* atoms to form
CH3O* and OH*, as shown in Fig. 7a and 7b. Activation energy for CꢀH activation on
21

RSC Advances
Page 22 of 34
View Article Online
DOI: 10.1039/C6RA21150J
Pt(111)ꢀ1ML O is175 kJ molꢀ1, a value much larger than on Pt(100)ꢀ1ML O (132 kJ
molꢀ1) because of the higher coordination Pt atoms in Pt(111) surface. As a result, CꢀH
bond activation energies on O*ꢀO* lie between the activation energies on
Pt(111)ꢀ1ML O and on Pt(100)ꢀ1ML O, which is in agreement with the apparent
activation energy (160 kJ molꢀ1) in this study.
PdO bulk oxide phases are wellꢀknown, and the crystal faces (101) and (100) present
low surface energies, both of which are stable facets [42, 43]. PdO (101) is a
corrugated surface with two types of Pd sites: one coordinated to four oxygen atoms
and another one coordinated to only three oxygen atoms. Hꢀabstraction to form CH3*
and OH* proceeds with a barrier of only 61 kJ molꢀ1 at the underꢀcoordinated
Pdꢀsite.(nonstoichiometric PdꢀO site pairs) on PdO(101). PdO(100) surface is less
active for Hꢀabstraction despite they are more stable than PdO(101) surface. The
DTFꢀderived CꢀH bond activation barriers on PdO (100) are about twoꢀfold higher
than on PdO(101) (138 vs 61 kJ molꢀ1) since its stable stochiometric facet with surface
Pd:O stoichiometries of 1:2.
22

Page 23 of 34
RSC Advances
View Article Online
DOI: 10.1039/C6RA21150J
Figure 7. Structures of reactants, transition states and products (first layer to third layer) for the
initial CꢀH bond activation steps on (a) Pt(111) metal surface covered with 1ML oxygen, (b)
Pt(100) metal surface covered with 1ML oxygen,(c) PdO(101), (d) PdO(100).
3.3.2 CꢀH bond activation on oxide PdꢀPt bimetallic catalysts
Figure 8. Sideꢀ and topꢀviews of the optimized structure of different oxided layers. (a) PdO(101)
layer on Pt(100), (b) PdO(100) layer on Pt(100),(c) PdO2(0001) layer on Pt(111), (d) Pd3O4(100)
layer on Pt(100).
Bulk oxide phases for Pd0.25Pt0.75 did not been detected by XRD. However, the
temperature at 10% conversion (T10) of the Pd0.25Pt0.75 catalyst is found more than 50
K higher than on the Pt1.0 catalyst in this study. Both of the kinetic parameters on
oxide PdꢀPt catalysts are more closer to Pd oxide catalyst rather than Pt catalyst (see
in Tabel 1). On Pd metallic surfaces, chemisorbed oxygen is a p(2×2) periodicity at
low oxygen pressure. However, at higher oxygen temperature, several different
surface oxide phases formed and coexisted, as researched in methane oxidation [25]
and STEM investigations [44, 45]. Hence, the PdꢀPt catalysts own the bulk structure
of Ptꢀlike and the surface structure of PdOꢀlike based on the experimental results.
23

RSC Advances
Page 24 of 34
View Article Online
DOI: 10.1039/C6RA21150J
Because of the lack of relevant experimental information for the surface structures of
PdxPt1ꢀxOδ, the surface oxide models for PdꢀPt were based on the know surface oxide
structures of pure Pd and pure Pt. In particular, the experimentally observed and the
theoretically predicted (√5×√5) PdO(101)ꢀlike/(100) [46, 47], (√5×√5)
PdO(100)ꢀlike/(100)
[46,
48],
PtO2(0001)ꢀlike/(111)
[49ꢀ51]
and
Pt3O4(100)ꢀlike/(100)[52ꢀ54] were considered in this study, since the small lattice
misfits between metallic substrate and oxide surface for these layers [55]. The bulk
phases of Pt and PdꢀPt alloy are more stable than that of Pd with increasing oxygen
supply. Besides, Pd atoms of the outermost layer are more easier to be oxidized than
Pt atoms under the same conditions. This difference reflect stroger bonds with O at Pd
than Pt surfaces. Hence, PdO monolayer attached to Pt(100) and Pt(111) were firstly
studied. The optimized atomic structures of PdO(101)/Pt(100), PdO(100)/Pt(100),
PdO2(0001)/Pt(111), Pd3O4(100)/Pt(100) were shown in Fig.8.
The four surfaces are predicted to have activation energies in the range of 110ꢀ299 kJ
molꢀ1. Although the surface of PdO(101)/Pt(100) is nearly identical to PdO(101),
activation energy on PdO(101)/Pt(100) is much higher (110 kJ molꢀ1 vs 61 kJ molꢀ1).
This is because the underꢀcoordinated Pd atoms on PdO(101) coordinated to only
three oxygen atoms, while the identical Pd atoms on PdO(101)/Pt(100) are higher
coordinated (by two oxygen atoms and four Pt atoms). Significantly, activation energy
of the first CꢀH bond on the low PdO(101)ꢀlike surface is retrieved in the case of two
PdO(101) monolayers on Pt(100) with a value of 67 kJ molꢀ1. Besides, the calculated
reaction energy for CꢀH bond activation on PdO(101)/Pt(100) is weakly
24

Page 25 of 34
RSC Advances
View Article Online
DOI: 10.1039/C6RA21150J
endothermic(+1.8 kJ molꢀ1). Activation energy of 145 kJ molꢀ1 on PdO(100)/Pt(100) is
even higher comparing to PdO(101)/Pt(100), as well as the stronger endothermic (+54
kJ molꢀ1).
For the PdO2(0001)/Pt(111), all metal Pd atoms of the outermost layer are fully
coordinated by six oxygens, consequently, CꢀH activation occurs over two O sites
(Fig 9c). Activation energy calculated for the stochiometric oxide surface
PdO2(0001)/Pt(111) is extremely high with a value of 299 kJ molꢀ1, because every
adsorbed oxygen atoms is coordinated by three Pd atoms. In the case of Pd3O4(100)
layer on Pt(100), CH3 and H fragments absorbed to the high coordinated (by four O
atoms) Pd atom and the under coordinated (by two Pd atoms) O atom, respectively.
The surface Pd atoms in Pd3O4(100)ꢀlike catalysts are highly coordinated, but under
coordinated O atoms are more active compare to PdO(100). Therefore, the activation
energy calculated for Hꢀabstraction on Pd3O4(100)/Pt(100) is 124 kJ molꢀ1, which is
lower than the value calculated for PdO(100) (138 kJ molꢀ1).
In addition to the oxide surfaces as mentioned above, methane activation on
PdO(101)/Pd/Pt(100), PdO(101)/PdPt(100), two monolayers PdO(101)/Pt(100),
Pd0.75Pt0.25 on αO(101)/Pt(100) and Pd0.75Pt0.25 on βO(101)/Pt(100) were also studied (see
in Supplementary material Fig. S1 ). The structure obtained for PdO(101)/Pd/Pt(100)
is qualitatively the same as for PdO(101)/Pt(100) in which the sublayer surface Pt
atoms are replaced by Pd atoms. The CꢀH bond activation energy for
PdO(101)/Pd/Pt(100) (133kJ molꢀ1) is larger than the value for PdO(101)/Pt(100). The
larger activation energy on PdO(101)/Pd/Pt(100) can be understood simply from the
25

RSC Advances
Page 26 of 34
View Article Online
DOI: 10.1039/C6RA21150J
higher electronegativity of palladium[56]. In the case of PdO(101)/PdPt(100), lattice
constant of the PdPt(100) substrate is estimated by Vegard’s Law [57]. Due to the
surface undergoes compressive caused by the incorporation of the smaller Pd atom
into the Pt lattice shifts the dꢀband downwards relative to the Fermi level. As a
consequence, it will weaken the interaction between the adsorbates and the surface,
resulting in a larger activation energy of 119kJ molꢀ1 than that on PdO(100)/Pt(100).
In the case of two monolayers PdO(101) on Pt(100), Hꢀabstraction barrier is retrieved
with a value of 67 kJ molꢀ1. For Pd0.75Pt0.25 on αO(101)/Pt(100) and Pd0.75Pt0.25 on
βO(101)/Pt(100), a quarter of Pd atoms in oxide surfaces are replaced by Pt atoms.
However, the CꢀH bond activation energy on these two oxide surfaces are slightly
higher than that on PdO(101)/Pt(100).
Figure 9. Structures of reactants, transition states and products (first layer to third layer) for the
initial CꢀH bond activation steps on (a) PdO(101) layer on Pt(100), (b) PdO(100) layer on
Pt(100),(c) PdO2(0001) layer on Pt(111), (d) Pd3O4(100) layer on Pt(100).
26

Page 27 of 34
RSC Advances
View Article Online
DOI: 10.1039/C6RA21150J
Figure 10. Calculated activation energies for the C−H bond activation on the oxide catalysts.
To conclude, the potential active phases of metal supported surface oxides and bulk
metal oxides were investigated by DFT calculations. It is found that underꢀcoodinated
Pd sites and O sites in oxide surfaces are required to obtain low activation barriers for
CꢀH bond activation. Energy barrier for CꢀH bond activation on PdO(101)/Pt(100) is
much larger than the value on Pd(101) due to the higher coordinated Pd atoms on
PdO(101)/Pt(100) surface. Besides, the subꢀlayers oxide on PdPt alloys cannot be
ruled out, especially when the Pd content is enough high in PdꢀPt bimetal catalysts. In
this case, the activation energy is approximate to the value for PdO(101). At last, the
apparent activation energy for methane combustion increases with the increaing Pt
content due to the decrease in PdxPt1ꢀxO content.
27

RSC Advances
Page 28 of 34
View Article Online
DOI: 10.1039/C6RA21150J
4. Conclusions
Pt addition increases the stability of Pd/Al2O3 catalyst for methane combustion. A
sequence of elementary reaction steps and their kinetic relevance are established on
the basis of kinetic dependencies and density functional theory results to describe the
reaction kinetics observed during methane reactions with O2 and H2O (or CO2) on
these catalysts. The conclusions are as follows:
1). Monometallic Pd/Al2O3 catalyst loses a large portion of its initial activity with
continuous methane combustion at 723K and/or 873K, while the activities of
PdꢀPt/Al2O3 and Pt/Al2O3 decrease slightly with time.
2). Turnover rates of methane combustion are independent to O2 pressure (>10 kPa)
and proportional to CH4 pressure for all of the catalysts in this study. Besides,
CꢀH bond activation step on O*−* pair sites of oxide Pdꢀlike catalysts exhibits
low energy barriers and low preꢀexponential factors that are much lower than the
values for oxide Ptꢀlike catalysts.
3). In the case of oxide Pdꢀlike catalysts, H2O strongly inhibits methane combustion
by occupying vacancies in an adsorptionꢀdesorption equilibrium step, while CO2
has no influence on methane combustion only when CO2 concentration is high
(>3~5 kPa)). For oxidized Ptꢀlike catalysts, H2O or CO2 almost has no influence
on methane combustion even when the H2O or CO2 concentration is high (<8
kPa).
4). H2O removal lead to the most recovery of initial turnover rates on these catalysts,
indicating the inhibition processes are caused by competitive adsorption on
28

Page 29 of 34
RSC Advances
View Article Online
DOI: 10.1039/C6RA21150J
kinetics, rather than by deactivation.
5). Pd atoms prefer to segregate to the surfaces for the bimetallic PdꢀPt catalysts. In
the case of Pd0.25Pt0.75 catalyst, hydrogen abstraction occurs on PdO(101)/Pt(100)
with an activation energy of 110 kJ molꢀ1, which is significantly lower than the
value calculated for Ptꢀ1ML O.
6). Although the surface of Pd0.25Pt0.75/Al2O3 is similar to that of Pd/Al2O3, CꢀH bond
activation energy on Pd/Al2O3 is much smaller than the value for
Pd0.25Pt0.75/Al2O3. The oxide subꢀlayers of the PdPt alloys cannot be ruled out in
the case of PdꢀPt bimetal catalysts with high Pd content. For this reason, CꢀH
bond activation energy for PdꢀPt catalysts with high Pd content are very close to
the value for Pd catalyst.
7). Apparent activation energy of methane combustion for Pd0.25Pt0.75 is 101 kJ molꢀ1,
which lies between the activation energies on Pd and Pt catalysts. Bulk PdO
and/or two monolayer PdO phases are hard to exist for PdꢀPt catalysts with low
Pd content.
Acknowledgment
The authors would like to thank the Projected Supported by Graduate Scientific
Research and Innovation Foundation of Chongqing, China (Grant No. CYB16022),
Natural Science Foundation of China with Project No.51276207 and No.50876118.
References
29