Highly Active and Stable Pd?GaOx/Al2O3 Catalysts Derived from Intermetallic Pd5Ga3 Nanocrystals for Methane Combustion

Article abstract of DOI:10.1002/cctc.201801684

Catalytic combustion of methane is widely used in industrial and transportation activities. The commonly used Pd-based catalysts, however, are easily deactivated under hydrothermal conditions. In this work, we adopted a novel strategy to prepare a GaO_x-doped Pd/Al₂O₃ catalyst via the oxidative transformation of intermetallic Pd₅Ga₃ nanocrystals supported on γ-Al₂O₃. We observed a synergistic effect between palladium and gallium oxide via formation of a bi-functional active Pd?O?Ga phase, resulting in a highly active and exceptionally stable catalyst that could markedly suppress the sintering of noble metals under harsh conditions (hydrothermal treatment at 750 °C). Moreover, the presence of a large amount of surface oxygen vacancies and the surface Pd?O?Ga phase could promote methane combustion over the catalysts. It is believed that such a novel dual-metal catalyst is promising in practical applications for methane combustion.

Full text of DOI:10.1002/cctc.201801684

Accepted Article
Title: Highly Active and Stable Pd−GaOx/Al2O3 Catalysts Derived from
Intermetallic Pd5Ga3 Nanocrystals for Methane Combustion
Authors: Zhiquan Hou, Yuxi Liu, Jiguang Deng, Yue Lu, Shaohua Xie,
Xiuzhong Fang, and Hongxing Dai
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemCatChem 10.1002/cctc.201801684
Link to VoR: http://dx.doi.org/10.1002/cctc.201801684
A Journal of
www.chemcatchem.org

10.1002/cctc.201801684
ChemCatChem
Highly Active and Stable PdGaOx/Al2O3 Catalysts Derived from Intermetallic
Pd5Ga3 Nanocrystals for Methane Combustion
Zhiquan Houa, Dr. Yuxi Liua,*, Prof. Jiguang Denga, Dr. Yue Lub, Dr. Shaohua Xiea, Dr. Xiuzhong
Fanga, Prof. Hongxing Daia,*
a
Beijing Key Laboratory for Green Catalysis and Separation, Key Laboratory of Beijing on
Regional Air Pollution Control, Key Laboratory of Advanced Functional Materials, Education
Ministry of China, and Laboratory of Catalysis Chemistry and Nanoscience, Department of
Chemistry and Chemical Engineering, College of Environmental and Energy Engineering, Beijing
University of Technology, Beijing 100124, China
b
Institute of Microstructure and Properties of Advanced Materials, Beijing University of
Technology, Beijing, 100124, China
* To whom correspondence should be addressed:
Dr. Yuxi Liu (yxliu@bjut.edu.cn)
Prof. Hongxing Dai (hxdai@bjut.edu.cn)
Tel. No.: +86-10-6739-6118; Fax: +86-10-6739-1983
1
This article is protected by copyright. All rights reserved.

10.1002/cctc.201801684
ChemCatChem
Abstract:
Catalytic combustion of methane is widely used in industrial and transportation activities. The
commonly used Pd-based catalysts, however, are easily deactivated under hydrothermal conditions.
In this work, we adopted a novel strategy to prepare a GaOx-doped Pd/Al2O3 catalyst via the
oxidative transformation of intermetallic Pd5Ga3 nanocrystals supported on -Al2O3. We observed a
synergistic effect between palladium and gallium oxide via formation of a bi-functional active
Pd–O–Ga phase, resulting in a highly active and exceptionally stable catalyst that could markedly
o
suppress the sintering of noble metals under harsh conditions (hydrothermal treatment at 750 C).
Moreover, the presence of a large amount of surface oxygen vacancies and the surface Pd–O–Ga
phase could promote methane combustion over the catalysts. It is believed that such a novel
dual-metal catalyst is promising in practical applications for methane combustion.
2
This article is protected by copyright. All rights reserved.

10.1002/cctc.201801684
ChemCatChem
1. Introduction
Methane can induce a strong greenhouse effect with a global warming potential 20 times higher
than that of CO2.[1] At present, methane emissions due to human activities mainly come from
industrial processes, vehicles, and waste managements.[2,3] To meet the strict emission regulations,
efforts are required to be made in developing efficient and durable catalysts for reducing methane
emissions. Taking the natural gas-heavy-duty vehicles as an example, exhaust gases contain
o
low-concentration unburned methane at low temperatures (< 550 C).[4] Catalytic combustion of
methane, however, is a promising technology that can convert methane into CO2 and H2O at
relatively low temperatures.[5,6] Therefore, it is highly desired to develop heterogeneous catalysts
that can eliminate methane emissions.
Although supported noble metal (especially Pd)-based catalysts exhibit good performance under
methane-lean conditions,[7] high loadings and activity deterioration caused by sintering of noble
metals remain the major drawbacks of the supported Pd catalysts.[8] Therefore, it is of significance
to develop more efficient substitutes for industrial applications. Recently, various strategies have
been tried to enhance catalytic stability of supported noble metal materials by increasing noble
metal dispersion on high-surface-area supports and to improve thermal stability via the
metalsupport interactions (SMSI).[9‒13] Moreover, a practicable approach of reducing catalyst cost
is modification of the noble metal(s) by adding cheap promoters (e.g., transition metal oxides).[14‒16]
When being used as heterogeneous catalysts, intermetallic compounds (IMCs) have shown
excellent activity and long-term stability in a number of chemical reactions due to their
well-ordered structures as compared with the alloy counterparts.[17‒20] To form an intermetallic
phase, a thermal annealing at high temperatures is generally necessary, which often results in
agglomeration of the nanoparticles (NPs). Therefore, it is critical to choose a suitable synthetic
3
This article is protected by copyright. All rights reserved.

10.1002/cctc.201801684
ChemCatChem
route in controlling the ordering and particle size of the intermetallic NPs.
Herein, we develop a solution-based co-reduction strategy to successfully synthesize the Pd5Ga3
intermetallic nanocrystals (NCs). The as-prepared NCs after a thermal treatment were used as
catalyst for methane combustion. It is found that the catalysts exhibited excellent performance and
(hydro)thermal stability, which was much higher than those of the Pd counterpart.
2. Results
The phase identification of the as-obtained samples was performed using the XRD technique.
The XRD pattern of the Pd5Ga3 sample well matched the standard intermetallic Pd5Ga3 sample
(JCPDS PDF# 65-2405) and possessed an orthorhomibc crystal structure (Figure 1a and Figure S1).
As shown in Figure 1b‒e, the XRD patterns of the supported Pd-based catalysts showed several
broad diffraction peaks assignable to the γ-Al2O3 phase, and some diffraction peaks at 2θ = 33.8o
and 54.7o due to the (101) and (112) planes of the tetragonal PdO phase (JCPDS PDF# 41-1107).
Figure S2 in the supporting information shows the XRD patterns of the Pd and PdGaOx NCs.
Obviously, the XRD patterns confirm appearance of the Pd and PdO phases in the pure Pd NCs,
whereas only the PdO phase existed in the PdGaOx NCs. This result indicates that GaOx
modification could promote oxidation of the Pd NCs during the calcination process. It should be
noted that there were diffraction peaks due to the Pd phase in the XRD pattern of Pd NCs, whereas
no diffraction peaks due to the Pd phase in the XRD pattern of Pd/Al2O3. This was due to the fact
that part of the Pd NCs were oxidized after calcination, resulting in presence of only a less Pd phase
in the Pd NCs (Figure S2). Furthermore, the real Pd loading in Pd/Al2O3 was only 1.28 wt%, which
could also lead to no detection of diffraction peaks due to the Pd phase in the Pd/Al2O3 sample.
4
This article is protected by copyright. All rights reserved.

10.1002/cctc.201801684
ChemCatChem
▲ Al2O PDF#29-0063
Pd-GaOx/Al2O3-HT
○ PdO3PDF#41-1107
▲
○
▲
(e)
▲
▲
○
▲
Pd/Al2O3-HT
(d)
(c)
Pd-GaOx/Al2O3
Pd/Al2O3
(b)
(a)
Pd5Ga3 PDF#65-2405
Pd5Ga3 NCs
20
30
40
50
60
70
80
2 Theta (Deg.)
Figure 1. XRD patterns of (a) Pd5Ga3 NCs, (b) Pd/Al2O3, (c) PdGaOx/Al2O3, (d) Pd/Al2O3-HT, and (e)
PdGaOx/Al2O3-HT.
As observed in the typical TEM and HAADFSTEM images, EDX elemental mapping, and
particle-size distributions (Figures 2, S3, and S4 in supporting information), the as-obtained
Pd-based NCs were uniform in size. The Pd particles (40.3 nm in size) of the hydrothermally aged
Pd/Al2O3-HT sample grew significantly as compared with those (5.5 nm in size) of the fresh
counterpart. When the PdGaOx/Al2O3 sample was hydrothermally treated, however, a slight
growth of PdGaOx NCs from 5.3 to 5.4 nm in mean size happened. After the hydrothermally aged
test, the average particle size increased slightly from 5.3 ± 0.7 to 8.4 ± 2.6 nm for PdGaOx/Al2O3,
and more severe sintering happened over Pd/Al2O3 with an increase in average PdO particle size
from 5.5 ± 0.6 to 35.8 ± 6.8 nm (Figure S4 in supporting information). Furthermore, PdO
distribution with an average particle diameter of 29.2 ± 6.1 nm was observed over the aged
Pd/GaOx/Al2O3 sample (Figure S4(GI) in supporting information). Apparently, the Pd/GaOx/Al2O3
sample prepared by a two-step impregnation method showed a negligible effect on PdO particle size
5
This article is protected by copyright. All rights reserved.

10.1002/cctc.201801684
ChemCatChem
control. Therefore, it is concluded that the PdGaOx/Al2O3 sample possessed better stability than
Pd/GaOx/Al2O3 and Pd/Al2O3 samples.
(A)
(B)
(C)
(D)
(E)
30
20
10
0
30
25
20
15
10
5
35
30
25
20
15
10
5
(F)
(G)
(H)
(I)
(J)
0
0
0
10
20
30
40
50
60
70
2
3
4
5
6
7
8
9
1
2
3
4
5
6
7
8
Particle size (nm)
Particle size (nm)
Particle size (nm)
Figure 2. TEM images and particle-size distributions of (A, F) Pd5Ga3 NCs, (B, G) Pd NCs, HAADF‒STEM
images and particle-size distributions of (C, H) PdGaOx/Al2O3, (D, I) PdGaOx/Al2O3-HT, and (E, J)
Pd/Al2O3-HT.
Analysis on the structure of the dual metal catalyst using the EDX elemental mapping technique
reveals the presence of a mixture of Pd and Ga within each particle, and the gallium species were
present as a thin amorphous film coating near the metal particle on the Al2O3 support (Figure 3D).
This is due to the rapid oxidation of Pd5Ga3 in the muffle furnace, so that there are some of gallium
oxide doping into the Al2O3. The corresponding line-scan analysis result (Figure 3C) indicates a
homogeneous distribution of Pd and Ga species in the entire region. Besides, the quantitative
analysis of EDX line scan is also shown in Figure. 4A and B. Inhomogeneous distributions (Figure
4A) of Pd and Ga atoms were observed when the Al and its associated O atoms of the support were
considered. This result shows that the oxygen was mixed with palladium and gallium, maintaining
the particle structure. However, there was a constant molar content of Pd, Ga, and O atoms in the
6
This article is protected by copyright. All rights reserved.

10.1002/cctc.201801684
ChemCatChem
(A)
(B)
(C)
3 nm
20 nm
(D-1)
(D-2)
(D-3)
(D-4)
(D)
Merge
Al K
Pd L
Ga L
Figure 3. (A, B) Aberration-corrected STEM images, (C) EDX line scan, and (D) EDX elemental mapping of the
PdGaOx/Al2O3 sample.
whole PdGaOx particle when the Al and its associated O atoms of the support were ruled out
(Figure 4B). In other words, the rapid oxidation of the intermetallic metal particles could give rise
to a rather homogeneous structure of the PdGaOx NCs without an obvious phase segregation. The
data of EDX line scan were shown at Supporting Information (Tables S1 and S2). The
corresponding SAED patterns in Figure 4D show a bright ring pattern, suggesting formation of a
polycrystalline structure in the obtained samples. It is observed that there were three diffraction
rings assigned to the (440), (400), and (222) crystal planes of -Al2O3 (Figure S6 in supporting
information), whereas the other diffraction rings could not be attributed to the PdO and GaOx. This
result means that the crystal structure of the PdGaOx nanoparticles was different from that of the
PdO or the GaOx. To further confirm such a deduction, we also measured the lattice spacings of the
PdGaOx nanoparticles using the aberration corrected (Cs) STEM technique, and the corresponding
7
This article is protected by copyright. All rights reserved.

10.1002/cctc.201801684
ChemCatChem
HAADF line profile is shown in Figure 4C-1. The lattice spacing of the marked region was 0.246
nm (Figure 4C). There were obvious differences between the PdGaOx and the standard phase of
PdO or GaOx. It was evidenced that the rapid oxidation of the intermetallic metal particles could
facilely form a kind of the Pd–O–Ga compound, but it was not detected in the XRD pattern.
Figure 4. (A, B) EDX line scan crossing a PdGaOx nanoparticle, and (B) is the molar contents of Pd, Ga, and O
elements in the whole region of the nanoparticle, in which the support was ruled out. (C) HAADFSTEM image
of PdGaOx/Al2O3, and (C-1) corresponding HAADF line profile along the direction marked by the blue region in
(C); and (D) selected-area diffraction (SAED) of the PdGaOx/Al2O3 sample.
The exact real Pd and Ga loading of the as-obtained samples were also determined by the
8
This article is protected by copyright. All rights reserved.

10.1002/cctc.201801684
ChemCatChem
ICP−AES (Table 1) techniques. It should be mentioned that by using the sol immobilization
technique, the centrifugation separation and washing of the as-obtained precipitate NCs (dispersed
in the solution) could result in a loss in NCs, especially for the metal cluster (much smaller than the
bimetallic NPs). Therefore, the ratio of the two metal NCs on the support after the loading process
was different from their nominal ratio. Surface areas of the Pd5Ga3/Al2O3, Pd/Al2O3, and
PdGaOx/Al2O3 samples were in the range of 167176 m2 g1, slightly than those (142153 m2 g1)
of their hydrothermally or thermally aged counterparts. All of the samples possessed a pore volume
of 0.480.61 cm3 g1 (Figure S7 in supporting information and Tables 1 and S3).
It can be clearly observed that the supported PdGaOx catalyst performed better than the
supported Pd and Pd/GaOx/Al2O3 catalysts for methane combustion at a SV of 80,000 mL (g h) 1
(Figures 5A and S8 in supporting information). It is worth noting that the performance of
Pd/GaOx/Al2O3 prepared by the impregnation method for methane combustion was poorer than that
o
of PdGaOx/Al2O3 (T50% over the former was 67 C higher than that over the latter), while the
GaOx/Al2O3 sample was inactive for methane combustion. More importantly, to simulate the
practical applications in methane combustion, we conducted two types of hydrothermally ageing
o
experiments: (i) in the presence of 10 vol% water vapor in air at 750 C for 8 h before methane
combustion; and (ii) in the presence of 10 vol% H2O + 0.5 vol% CH4 + 4 vol% O2 + 85.5 vol% N2
o
at 750 C for 8 h before methane combustion. It is found that the catalytic activity (T10% = 286 oC,
T50% = 350oC and T90% = 410oC) over PdGaOx/Al2O3-HT was much better than that (T10% = 372oC,
o
o
T50% = 461 C and T90% = 507 C) over Pd/Al2O3-HT. In other words, the PdGaOx/Al2O3-HT
sample was catalytically stable under the adopted conditions. In addition, Figure 5B and Table S4
in supporting information show the catalytic performance of the samples after the hydrothermal
pretreatment. It is clearly seen that the PdGaOx/Al2O3-HRT sample also performed a better
9
This article is protected by copyright. All rights reserved.

10.1002/cctc.201801684
ChemCatChem
o
hydrothermal stability, with the T10%, T50%, and T90% being 300, 365, and 422 C, respectively,
slightly higher than those (T10% = 294 oC, T50% = 336 oC, and T90% = 372 oC) over PdGaOx/Al2O3.
o
o
However, the Pd/Al2O3-HRT sample showed a T50% of 430 C and a T90% of 478 C, which were
o
o
much higher than those (T50% = 379 C and T90% = 418 C) over the Pd/Al2O3 sample. It should be
mentioned that O2 pretreatment was conducted prior to the in-situ catalytic performance
measurements, which could remove the trace amounts of water and carbon species adsorbed on the
sample surface. We evaluated catalytic activities of the Pd−GaOx/Al2O3 sample with and without O2
pretreatments. As shown in Figure S9, there were no significant differences between the catalytic
activities of this sample with and without O2 pretreatments. Therefore, there was no direct
correlation between the pretreatment and catalytic performance. These results indicate that doping a
small amount of GaOx to the Pd NPs was beneficial for enhancing the hydrothermal stability. We
further investigate the thermal stability of the PdGaOx/Al2O3-T and Pd/Al2O3-T catalysts for
methane combustion (Figure S10 in supporting information). The T90% values over
o
o
PdGaOx/Al2O3-T and Pd/Al2O3-T were 420 and 500 C, which were 48 and 82 C higher than
those over the fresh samples (Table S4 in supporting information), respectively. In addition,
compared to the other catalysts available in the literature (Table S5 in supporting information), the
PdGaOx/Al2O3 catalyst showed better performance than those with similar Pd loadings, but the
methane conversion slightly lower than those with high Pd loadings under the adopted reaction
conditions.
10
This article is protected by copyright. All rights reserved.

10.1002/cctc.201801684
ChemCatChem
Table 1. BET surface areas, pore volumes, average particle sizes, real metal loadings, surface element compositions, and reducibility of the as-obtained
samples.
H2 consumptionf
[μmol g1]
Surface element compositionc [mol mol1]
BET
Pore
Particle
sizea
Real Pd
Real Ga
Sample
surface area volume
loadingb loadingb
[m2 g1]
176
[cm3 g1]
0.61
Pd0/Pd2+ Pd2+/Pdδ+ d Oads/Olatt Ga0/Gaε+ e Ga+/Ga3+
‒5050 oC 300500 oC
[nm]
[wt%]
[wt%]
Pd5Ga3/Al2O3
PdGaOx/Al2O3
Pd/Al2O3
4.2 ± 0.4 1.31
5.3 ± 0.7 1.37
5.5 ± 0.6 1.28
5.4 ± 0.9 1.25
1.29
1.41

8.70
0.06
0.09
0.01
0.03
0.43
0.77
0.55
0.81
0.45
0.40
0.05

2.70
1.12




169
0.54
8.50
14.29
1.12
10.32
174.4
100.2
198.6
203.1

167
0.54

153
0.52
1.09
0
0.58
87.3
56.3
PdGaOx/Al2O3-HT
Pd/Al2O3-HT
40.3 ±
0.98
9.6
149
0.49



a Estimated according to the TEM images;
b Determined by the ICPAES technique;
c Determined by the XPS technique;
d 2 ˂ δ ≤ 4;
e Gaε+ = Ga+ + Ga3+;
f Calculated by quantitatively analyzing the reduction peaks of the H2-TPR profiles of the samples.
11
This article is protected by copyright. All rights reserved.

10.1002/cctc.201801684
ChemCatChem
(A)
(B)
Figure 5. Methane conversion as a function of temperature over the samples at SV = 80,000 mL (g h)1.
In the last decades, many reports on the kinetic behaviors have demonstrated that methane
oxidation obeys a first-order reaction mechanism with respect to methane concentration under the
oxidizing or stoichiometric conditions.[21] The specific reaction rate (μmol (gPd s)1) is used to
evaluate the catalytic activities of the samples (Table 2). Apparently, the PdGaOx/Al2O3-HT
sample exhibited the highest reaction rate of 40.55 μmol (gPd s)1, much higher than those over the
PdGaOx/Al2O3 (23.32 μmol (gPd s)1), Pd/Al2O3 (0.76 μmol (gPd s)1), and Pd/Al2O3-HT (0.75
μmol (gPd s)1) samples. The result of specific reaction rate was in good agreement with the
enhancement in low-temperature catalytic performance of the samples.
XPS is an effective technique to gain the information of surface element compositions, metal
oxidation states, and adsorbed species of a solid material. Figure 6A illustrates the Pd 3d5/2 XPS
spectra of the as-obtained samples, and their Ga 2p and O 1s XPS spectra are shown in Figure 6B
and C), and Table 1 summarizes the quantitative analysis results. It can be observed from Figure 6A
that the asymmetrical Pd 3d5/2 XPS spectrum of each sample could be decomposed to three
components at BE = 335.2, 336.8, and 337.9 eV, assignable to the Pd0, Pd2+, and Pdδ+ (2 ˂ δ ≤ 4)
12
This article is protected by copyright. All rights reserved.

10.1002/cctc.201801684
ChemCatChem
o
Table 2. Catalytic activities, specific reaction rates of the as-obtained catalysts at 290 C for 0.5 vol % methane
combustion at SV = 80,000 mL (g h)1.
Methane combustion at 290 oC
Methane combustion
Sample
Specific reaction rate
[× 106 mol (gPd s)1]
T10% [oC] T50% [oC]
T90% [oC]
294
336
286
372
350
336
379
350
461
403
372
418
410
507
442
23.32
PdGaOx/Al2O3
Pd/Al2O3
0.76
40.55
0.75

PdGaOx/Al2O3-HT
Pd/Al2O3-HT
Pd/GaOx/Al2O3
species,[22,23]respectively. As summarized in Table 1, the Pd0/Pd2+ molar ratios (0.01‒0.09) of the
calcined samples were much lower than that (8.7) of the intermetallic sample, indicating that the Pd
in the intermetallic sample was present in metallic form to construct the crystal architecture.
Furthermore, the Pd2+/Pdδ+ ratio of the sample decreased with doping of GaOx and hydrothermal
treatment. It can be observed from Figure 6B that the asymmetrical Ga 2p XPS spectrum of each
sample could be decomposed to three components at BE = 1117.0, 1118.3, and 1119.7 eV,
assignable to the surface Ga0, Ga+, and Ga3+ species,[24] respectively. As summarized in Table 1,
the Ga0/Gaε+ (Gaε+ = Ga+ + Ga3+) molar ratio of the intermetallic sample was much higher than that
of the calcined samples. Furthermore, the ratio of Ga+/Ga3+ increased with a rise in calcination
temperature. For each sample, the O 1s spectrum (Figure 6C) could be decomposed to three
components at BE = 530.8, 531.9, and 533.4 eV, attributable to the surface lattice oxygen (Olatt),
adsorbed oxygen (Oads), and adsorbed molecular water,[25,26] respectively. Compared the O 1s XPS
spectra of pure support Pd/Al2O3 and Pd−GaOx/Al2O3 (Figure S11 and Figure 6C), we found that
the binding energy of O in Pd/Al2O3 and pure support is similar while the binding energy of O in
Pd−GaOx/Al2O3 shift to the higher energy. And the surface Oads/Olatt molar ratio increased after the
13
This article is protected by copyright. All rights reserved.

10.1002/cctc.201801684
ChemCatChem
introduction of GaOx (Table 1). In other words, the Pd-GaOx/Al2O3 sample possessed a higher
ability to activate oxygen molecules, as compared to the Ga-free counterpart.
336.8
530.8
Pd 3d5/2
O 1s
Ga 2p
531.9
337.9
(B)
1119.7
1118.3
(C)
(A)
335.2
Pd-GaO /Al O -HT
x
2 3
Pd/Al2O3-HT
Pd/Al2O3-HT
533.4
1117.0
Pd-GaOx/Al2O3-HT
Pd-GaOx/Al2O3-HT
Pd-GaO /Al O
x
2
3
3
Pd/Al2O3
Pd Ga /Al O
Pd/Al2O3
5
3
2
Pd-GaO /Al2O3
x
Pd-GaOx/Al2O3
GaO /Al O
x
Pd5Ga3/Al2O3
2
3
1114
1116
1118
1120
1122
1124 528
530
532
534
536
333
334
335
336
337
338
339
340
Binding energy (eV)
Binding energy (eV)
Binding energy (eV)
Figure 6. (A) Pd 3d, (B) Ga 2p, and (C) O 1s XPS spectra of as-obtained samples.
From the above results, we can realize that there was a strong metalmetal oxide interaction
between Pd(O) and GaOx. Previous HRTEM and DFT studies revealed that Pd was stabilized by
PdOCe bonds on a CeO2 support under an oxidative atmosphere at high temperatures.[12]
Therefore, the oxidized Pd in the PdGaOx particles might form a nanocomposite with Ga under an
oxidative atmosphere. The generation of such a mixed metal oxide is the evidence that there was a
strong interaction between the oxidized Pd (PdO) and GaOx, involving in the formation of a
Pd–O–Ga bond that could stabilize the Pd against sintering. Moreover, the structural defect could
also play a key role in creating oxygen vacancies on the surface of the sample, thus favoring the
activation of the C–H bonds in methane and dissociative adsorption of oxygen molecules.[27]
H2-TPR was used to investigate the reducibility and nature of noble metal sites of the supported
Pd-based samples, and the H2-TPR profiles and H2 consumption are shown in Figure 7A and Table
1. For each sample, there was at least one main reduction peak. The main reduction peak of Pd
o
occurred at the sub-ambient temperature.[28] The reduction peak below 50 C could be ascribed to
the reduction of PdOx (to the Pd0) and the surface adsorbed oxygen species.[28] For the supported Pd
14
This article is protected by copyright. All rights reserved.

10.1002/cctc.201801684
ChemCatChem
samples, there was one asymmetrical reduction signal at ca. 0 oC (which could be decomposed into
two peaks), demonstrating that the reduction of PdOx took place in two steps: the first reduction
Figure 7. (A) H2-TPR profiles, (B) TPO profiles, (C) in-situ DRIFTS spectra of CO adsorption at 25 oC, and (D)
O2-TPD profiles of samples.
peak centered at 0 oC was due to the reduction of Pd2+ to Pd0, whereas the second reduction peak at
o
10 C was due to the reduction of the reducible Pdδ+ species to Pd0.[28] For the supported PdGaOx
sample, its reduction peak slightly shifted to a higher temperature, and the corresponding H2
consumption was 174 μmol g1, which was higher than that (100 μmol g1) of the supported Pd
sample. This result suggests that a large amount of high-valence palladium and higher adsorbed
oxygen species concentration existed in the PdGaOx/Al2O3 sample. If the palladium ions in PdOx
15
This article is protected by copyright. All rights reserved.

10.1002/cctc.201801684
ChemCatChem
were Pd4+ and completely reduced to Pd0, the theoretical H2 consumption of 1.28 wt% Pd/Al2O3
and 1.37 wt% PdGaOx/Al2O3 would be 240.6 and 257.5 μmol g1 (Table S6 in supporting
information), respectively. Clearly, the actual H2 consumption calculated from the H2-TPR profiles
of the Pd/Al2O3 (100.2 μmol g1) and PdGaOx/Al2O3 (174.4 μmol g1) catalysts were lower than
the corresponding theoretical H2 consumption. This was mainly due to the fact that the Pd ions were
mostly Pd2+ in the Pd/Al2O3 and PdGaOx/Al2O3 samples. However, the maximal H2 consumption
would be 237.9 μmol g1 if the Pd4+ in the 1.25 wt% PdGaOx/Al2O3-HT sample was reduced to
Pd0, while the actual H2 consumption was 285.9 μmol g1, which was higher than the theoretical H2
consumption. We believe that there might be two reasons accounting for such a phenomenon: (i)
The discrepancy in theoretical and actual H2 consumption of PdGaOx/Al2O3-HT might be due to
the presence of a large amount of adsorbed oxygen species that could consume a certain amount of
H2; and (ii), GaOx could be partially reduced in the presence of a small amount of Pd via a hydrogen
spillover process.[29]
After hydrothermal treatment, the PdGaOx/Al2O3-HT or Pd/Al2O3-HT sample shows one broad
o
reduction peak in the range of 310–470 C, which suggests that a small amount of PdOx located at
the interface between the nanoparticle and support would have a strong interaction with Al2O3.[30]
o
For the Pd/Al2O3 or Pd/Al2O3-HT samples, a negative peak below 90 C was due to decomposition
of the formed PdHx. According to the previous literature, the decomposition temperature of metal
hydride and the intensity of these peaks increased linearly with the relatively large noble metal
particle size.[31] In the sharp contrast, the supported PdGaOx sample only displayed a very weak
negative peak at a lower temperature, suggesting the good distribution of Pd in PdGaOx.
The above results demonstrate that the reducibility of the supported Pd-based sample was
dependent on the morphology and composition of the metal crystals. Compared to the pure
16
This article is protected by copyright. All rights reserved.

10.1002/cctc.201801684
ChemCatChem
Pd/Al2O3 sample, the different profiles and H2 consumption of the PdGaOx/Al2O3 sample might be
due to a stronger interaction between palladium and gallium (possible formation of PdGaOx) and/or
a larger amount of coordination-unsaturated surface PdOx.[32]
The transformation of PdOx to Pd is an important feature of the Pd-based catalysts used for
methane combustion since it directly affects the catalytic activity. Figure 7B shows TPO profiles of
the samples. There were one to three positive peaks due to the desorption of oxygen in the range of
o
600–800 C when the temperature was increased, and one or two negative peaks due to the
o
consumption of oxygen when the temperature was decreased from 800 to 100 C. The positive
peaks were attributed to the decomposition of the PdOx species, whereas the negative peaks were
assigned to the re-oxidation of the metallic Pd0 species.[33] It can be seen that addition of gallium
had a significant effect on the PdOx–Pd transition. In the presence of gallium oxide, the
temperatures of PdOx decomposition and metallic Pd0 re-oxidation both shifted to higher
temperatures than those for the Pd/Al2O3 sample, indicating that the formation of PdGaO bonds
could obviously improve the thermal stability of the PdOx. The main difference between the
supported Pd and PdGaOx samples was observed when the temperature was decreased. With the
GaOx as dopant, re-oxidation of the Pd took place in two steps (see Figure 7B): the first smaller
o
oxygen uptake peak centered at ca. 650 C was due to the re-oxidation of the Pd crystallites in
contact with gallium oxide and/or in PdGaOx, and the second oxygen uptake peak centered at ca.
500 oC was due to the re-oxidation of the Pd crystallites in contact with Al2O3 only, but the profiles
o
of the GaOx-free samples showed a single re-oxidation peak at ca. 475 C.[33] After hydrothermal
treatment, the re-oxidation temperature of Pd/Al2O3 shifted significantly towards a lower
temperature, whereas the re-oxidation of the PdGaOx samples occurred in the region of higher
temperatures. The results of TPO experiments demonstrate the presence of the Pd–GaOx interaction
17
This article is protected by copyright. All rights reserved.

10.1002/cctc.201801684
ChemCatChem
after doping GaOx to the Pd sample.
Adsorption of CO on the as-obtained samples was studied by recording their in-situ DFIFTS
spectra (Figure 7C). The adsorption bands at 2090 and 2055 cm‒1 could be assigned to the linearly
adsorbed CO on the metallic Pd0 species, whereas the adsorption bands at 1945 and 1905 cm‒1 were
attributed to the bridged adsorbed CO on Pd0.[34] There were two adsorption bands at 2090 and 1945
cm‒1 on the Pd/Al2O3 and Pd/Al2O3-HT samples, whereas no adsorption bands were observed on
the PdGaOx/Al2O3 and PdGaOx/Al2O3-HT samples. This might be due to formation of the
PdGaO mixed metal oxide after calcination, resulting in isolation of the Pd atoms within the
sample. The more stable Pd sites on the sample surface were surrounded by the gallium and oxygen,
and they were difficult to be reduced at low temperatures. Therefore, there were no obvious CO
adsorption bands on the PdGaOx/Al2O3 and PdGaOx/Al2O3-HT samples. In contrast, significant
CO adsorption was observed on the intermetallic Pd5Ga3 sample. There were three CO vibrational
adsorption bands at 2055, 1945, and 1905 cm−1, assignable to the linearly bound CO, bridged CO
(Pd2–CO), and CO molecules bound to three Pd atoms (Pd3–CO),[35] respectively. There was a
redshift in the in-situ DFIFTS spectrum of the Pd5Ga3 sample. This result suggests that there was
negatively charged Pd species in this sample, which might be a result of the strong electronic
modification.[36]
O2-TPD was used to investigate the bonding strength of oxygen species on the surface of
catalysts that determines the catalytic activity of oxygen for the redox reaction. As shown in Figure
7D, it can be clearly seen that there are two apparent oxygen desorption peaks (i.e., the α and β
o
peaks) for all of the samples. The α peak at low temperature (< 250 C) was attributed to the
desorption of physical adsorption oxygen species and chemisorbed oxygen species, whereas the β
o
peak in the range of 300700 C was assigned to the desorption of lattice oxygen from PdOx and
18
This article is protected by copyright. All rights reserved.

10.1002/cctc.201801684
ChemCatChem
GaOx [37] The relative amount of O2 desorbed was calculated from the peak area by deconvolution of
the O2 desorption curves. As shown in the Table S7, there is no obvious α-O2 desorption peak for
GaOx/Al2O3. After the loading of Pd, an obvious α peak could be observed, and peak area is similar
with each samples. The β peak area of the decomposition of lattice oxygen species at high
temperature can be directly related to the oxidation state of palladium and oxygen coordination
number. The large amount of β-O2 desorbed for the supported Pd−GaOx sample (as compared with
the supported Pd sample) suggests that palladium species in the dual metal may have higher
oxidation states and GaOx may also be partly decomposed at higher temperature, which confirmed
by XPS and TPR results. The β peak of the supported PdGaOx sample shifted to a higher
temperature as compared to the supported Pd sample, also demonstrating that GaOx introduction
could increase the strength of the Pd‒O bond. After hydrothermal treatment, the amount of β-O2
released from PdGaOx/Al2O3-HT, is found to be much lower as compared to the Pd/Al2O3-HT
sample. This phenomenon is also accorded with the TPO result, and it can be explained that the
crystal structure of Pd−O−Ga compound and bond strength between palladium and oxygen could be
further enhanced under the hydrothermal treatment, which results in much higher stability at higher
temperature avoiding structure decomposition. In addition, although the pure Al2O3 support did not
undergo the same measurements (TPO, H2-TPR, O2-TPD, and CO-DRIFT), there were no signals
assignable to Al2O3 in the profiles or spectra of the TPO [38], H2-TPR [39], O2-TPD [40], and
CO-DRIFT [41]
.
Laser Raman technique was employed to obtain global structure information of the samples, and
their spectra are shown in Figure S12 in supporting information. For the Al2O3 support, no Raman
bands were detected in the scanning range. The intense Raman band at 640 cm‒1 and the weak
Raman bands at 280 and 440 cm‒1 were the characteristic Raman bands of crystalline PdO in the
19
This article is protected by copyright. All rights reserved.

10.1002/cctc.201801684
ChemCatChem
supported Pd-based samples.[42] The Raman band at 640 cm−1 was attributed to the B1g mode of PdO,
whereas the Raman bands at 280 and 440 cm−1 were ascribed to the Eg mode of PdO. As shown in
Figure S12(B), the addition of GaOx to the supported Pd sample shifted the Raman bands to lower
frequencies by ca. 5 cm−1. The variation in Raman band could be associated with the
non-stoichiometry induced by oxygen deficiency or disorder induced by some minor phases in the
sample.[28,43] The Raman bands of the hydrothermally aged samples shifted to higher frequencies,
indicating the increase in Pd–O bond strength after the hydrothermal treatment. The results were in
good accordance with those of the TPO and in situ DFIFTS investigations.
Water is a product of methane combustion. To probe the water vapor effect on the catalytic
performance, methane oxidation over PdGaOx/Al2O3 in the presence of 3.0 or 5.0 vol% water
o
vapor at 360 C and SV = 80,000 mL (g h)1 was carried out, and the results are shown in Figure
S13(A) in supporting information. Introduction of 3.0 or 5.0 vol% water vapor to the feedstock led
to a decrease by ca. 20 % in methane conversion after a certain time of on-stream reaction. When
water vapor was cut off, CH4 conversion was increased to ca. 78 %, which was ca. 5 % lower than
that in the water-free case. Deactivation by H2O can be attributed to formation of Pd(OH)2 via the
interaction of H2O and PdO exposed and consequence of a competitive adsorption between CH4 and
H2O on the PdO sites.[44]
To probe the effect of CO2 addition on catalytic activity, we performed methane oxidation at 360
oC and SV = 80,000 mL (g h)1 over PdGaOx/Al2O3 in the case of CO2 addition to the feed stream,
and the results are shown in Figure S13(B) in supporting information. It can be seen that methane
conversion decreased by ca. 30 % after 20 h of on-stream reaction. This negative effect of CO2
addition might be a result due to the accumulated adsorption of carbonate species during the
on-stream reaction, which would decrease the activation adsorption of oxygen and/or methane.
20
This article is protected by copyright. All rights reserved.

10.1002/cctc.201801684
ChemCatChem
o
After the deactivated sample was treated in oxygen at 450 C for 3 h, its conversion was almost
restored to its original value in the CO2-free case. This result indicates that the catalyst deactivation
induced by CO2 introduction was reversible, and it can be proved by the result of FT-IR spectra
(Figure S14 in supporting information). According to the literature,[45,46] the absorption bands of the
2‒
‒
CO3 , HCO3 , and C−O species were at 1480, 1370, and 1120 cm1, respectively. Compared to the
FT-IR spectrum of the CO2-accumulated catalyst, there were no carbonate species on the surface of
the catalyst after oxidative activation.
It has been demonstrated in previous studies that sulfur compounds play a key role in the
deactivation of the catalysts used for methane combustion.[47] To examine the effect of SO2 on
catalytic activity, we conducted methane combustion over the PdGaOx/Al2O3 sample at 360 oC and
SV = 80,000 mL (g h)1. As shown in Figure S13C in supporting information, methane conversion
decreased from ca. 80 to 40 % after 20 h of 50 ppm SO2 introduction. After the SO2-treated catalyst
o
was activated in O2 at 450 C for 3 h, methane conversion slightly increased to 48 %. This result
indicates that the catalyst deactivation caused by SO2 introduction was irreversible. To elucidate the
reason for such a deactivation of the catalyst, we recorded the FT-IR spectra, as shown in Figure
S15 in supporting information. There were absorption bands at 1130, 1640, 2450, 2290, and 3400
cm1. According to the literature,[48,49] the broad absorption band at 3330 and 1640 cm1 could be
assigned to the stretching vibration and bending oscillation of −OH species. The appearance of the
absorption band at 2450 cm1 confirms the existence of CO2. The double bands in the range of
2860–2920 cm1 were due to the existence of C−H absorption, and the absorption band at 1130
2
2
cm1 suggests the formation of SO3 or SO4 species. Similar infrared spectra before and after
oxidative activation demonstrate the irreversible deactivation of the PdGaOx/Al2O3 sample
induced by SO2. The negative effect of SO2 addition on methane combustion might be due to the
21
This article is protected by copyright. All rights reserved.

10.1002/cctc.201801684
ChemCatChem
attack of SO2 on the Pd‒O‒Ga compound, which made highly active palladium oxide transform to
less active PdSO3 or PdSO4 via the following reactions of PdO + SO2 → PdSO3 or PdSO4. In the
meanwhile, SO2 could also be strongly adsorbed on the Al2O3 support and formed Al2(SO3)3 or
Al2(SO4)3.[49]
3. Discussion
According to the above results, we propose a structural evolution scheme of the Pd5Ga3/Al2O3
and Pd/Al2O3 catalysts calcined and hydrothermally or thermally aged under different conditions, as
illustrated in Figure 8. The deactivation of the Pd-based catalyst during methane combustion could
be reasonably explained by PdO sintering, PdO decomposition, and Pd(OH)2 formation. The
HAADFSTEM results of the aged catalyst indicate that PdO sintering was one important reason
for stability of the present catalysts, and PdGaOx/Al2O3 was more resistant to PdO sintering (Table
1). On the other hand, the TPO, O2-TPD, and in situ DFIFTS results reveal that the GaOx species
modified the PdO bond strength, avoiding the decomposition of the active PdO phase. According
to the previous literature,[11,12] sintering of the precious metals could be greatly inhibited by
formation of the platinum group metal (PGM)‒O‒ceria complex. Seshadri et al. pointed out that
stability of the PGM ions in an oxide environment could be dramatically enhanced by presence of
the other cations, which could be attributed to the promoted covalent interactions between PGM
ions and oxygen.[50] TPR is an effective technique to investigate the interaction between metal sites
of the supported samples, the different reduction profiles of PdOx in the dual metal sample as
compared to the pure palladium sample could be ascribed to the fact that there was a stronger
interaction between palladium and gallium oxides. Furthermore, the variation (red shift) in the
Raman band of PdO in the PdGaOx/Al2O3 sample as compared to that in the Pd/Al2O3 sample
could be associated with the structure disorder induced by the minor phases in the sample or the
22
This article is protected by copyright. All rights reserved.

10.1002/cctc.201801684
ChemCatChem
non-stoichiometry induced by oxygen deficiency.[51] Interestingly, we find that the hydrothermal
treatment of the dual metal catalyst was also important for stabilizing the active component.
Figure 8. Schematic illustration of the PdGaOx/Al2O3 catalysts after calcination and hydrothermal ageing
treatments and their catalytic methane combustion process.
O2-TPD results confirm that the amount of lattice oxygen released from PdGaOx/Al2O3-HT was
much lower as compared with the PdGaOx/Al2O3 and Pd/Al2O3-HT samples, suggesting that the
hydrothermal treatment could further benefit enhancement of the PdOGa bond strength. Taking
these considerations together, we believe that during the heating process in air at higher
temperatures, gallium oxide interacted with palladium oxide to yield a homogeneous solid solution
or form a PdGaO compound, causing the lattice distortion of PdOx.[52] In addition, hydrothermal
treatment could also serve as an alternative strategy for preparing more stable supported catalysts.
The presented data were consistent with the low-temperature multistep pathway in which the
intermetallic nanoparticles were oxidized into nanocrystalline multi-metal oxides (Figure 8), which
would result in a very stable catalyst with high resistance to deactivation.
Methane combustion over the Pd-based catalysts has been interpreted in terms of the Mars–van
Krevelen reduction–oxidation pathway.[9] In this proposed mechanism, the catalytic site is a surface
23
This article is protected by copyright. All rights reserved.

10.1002/cctc.201801684
ChemCatChem
oxygen vacancy or a coordinately unsaturated Pd site on the surface of PdO crystallites. Such a
mechanism also involves in the activation of gas-phase oxygen to replenish the surface oxygen
defects formed by methane combustion. The structural defects of PdGaOx/Al2O3 could play a key
role in enhancing the catalytic performance since they were favorable for adsorption and
dissociation of oxygen molecules, which confirmed by the XPS and Raman results. According to
the literature,[53] activation of the first C–H bond in CH4 was the rate-determining step in methane
combustion. Recently, Weckhuysen's group[54] have pointed out that modification of the Pt surface
by gallium oxide primarily led to an increase in catalytic performance for propane dehydrogenation,
and the authors explained that the coordinately unsaturated Ga3+ species were responsible for the
activation of the C–H bond and the dissociative adsorption of propane resulted in formation of a
surface GaH/GaOH species. In the present work, a clear synergistic effect was observed between
the Pd and GaOx components. The H atoms could be abstracted sequentially from the adsorbed CH4
by the surface Ga–O species to form surface hydroxyl groups (Ga–OH and Pd–OH), the additional
−
active oxygen species promoted the dissociation of CH3 to HCOO− that was subsequently oxidized
to CO2, thus completing the catalytic cycle.
4. Conclusions
In summary, we developed a multistep strategy to synthesize the intermetallic Pd5Ga3 NCs with a
uniform particle size and Al2O3-supported nanocrystalline PdGaOx. The PdGaOx/Al2O3 catalyst
was highly resistant to sintering, which showed a higher catalytic activity for methane combustion
(T90% = 372 oC at SV = 80,000 mL (g h)1 and excellent hydrothermal stability (T90% = 410 oC after
o
hydrothermal aging treatment at 750 C). A combination of structural, morphological, and surface
property characterization reveals that PdGaOx/Al2O3 was a complex material with synergistic and
bi-functional characters originating from formation of a Pd–O–Ga bond that could enhance the
24
This article is protected by copyright. All rights reserved.

10.1002/cctc.201801684
ChemCatChem
Pd–O bond strength against sintering. Formation a large amount of surface oxygen vacancies and
surface Pd–O–Ga sites favored the improvement in the activation ability of reactant molecules. We
believe that the rational approach to catalyst development can provide new and exciting
perspectives in attaining highly efficient and durable catalysts with potential implications in a wide
range of catalytic reactions.
5. Experimental Section
Catalyst preparation: Palladium (II) 2,4-pentanedionate (Pd(acac)2, Pd 34.7 %), gallium (III)
chloride, ultra dry (GaCl3, 99.999 %), tert-butylamine borane (TBAB, 97 %), tri-n-octylphosphine
(TOP, 90 %), and benzyl ether (BE, 98 %) were purchased from Alfa Aesar. Oleylamine (OAm,
80-90%) was purchased from Acros Organics. Aluminium oxide (γ-Al2O3, 99.99 %) was purchased
from Aladdin. Ethanol and cyclohexane were obtained from Sinopharm Chemical Reagent Co. Ltd.
(Shanghai, China). Syntheses of the nanocrystals were carried out under inert atmosphere using the
Schlenk line technique.
Synthesis of Pd5Ga3 NCs. Pd(acac)2 (0.05 mmol) and GaCl3 (0.05 mmol) mixed with OAm (10
mL) were loaded into a three-neck flask, equipped with the reflux condenser, degassed and dried
under vacuum at 110 oC for 1 h under stirring. Then the reaction flask was heated to a 180 oC under
a gentle flow of Ar, followed by the drop wise injection of solution containing TBAB (4 mmol) in
degassed OAm (2.5 mL). A visible color change from yellowish to black was observed immediately.
o
o
After holding at 180 C for another 10 min, the reaction flask temperature was rose to 250 C and
o
held at 250 C for 1 h. Finally, the mixture was cooled down to room temperature with Ar
atmosphere protection. Particles collected by centrifugation after the addition of
ethanol/cyclohexane (3 : 1) mixture, were re-dispersed in an apolar organic solvent.
Synthesis of Pd NCs. Pd(acac)2 (0.15 mmol) were solved in benzyl ether (10 mL) together with
25
This article is protected by copyright. All rights reserved.

10.1002/cctc.201801684
ChemCatChem
OAm (1.084 mL) in a 50 mL three-necked flask. After dissolving at room temperature, the flask
was evacuated with nitrogen, and tri-n-octylphosphine (0.73 mL) was added. And then the solution
was degassed and dried under vacuum for 30 min to remove water and other impurities after heating
to 100 oC. Subsequently, the reaction was heated to 290 oC with the nitrogen protection and kept for
30 minutes before it was cooled down to room temperature.
In a typical preparation of supported catalyst, the as-obtained NCs, were loaded on γ-Al2O3
(palladium theoretical loading = 1.5 wt%) in 10 mL cyclohexane by stirring at room temperature for
12 h. The solids were then precipitated using ethanol, separated via centrifugation. Finally, the
catalysts were dried in a vacuum drying box.
Finally, the dried powders were transferred to a ceramic boat in a muffle furnace, and then
o
calcined at 450 C for 5 h. The obtained supported samples were denoted as PdGaOx/Al2O3 and
Pd/Al2O3, respectively. In order to investigate the hydrothermal stability and thermal stability, the
samples were also calcined in 10 vol% H2O/air (flow rate = 200 mL min1), 10 vol% H2O + 0.5
vol% CH4 + 4 vol% O2 + 85.5 vol% N2 at 750 oC for 8 h, and air (flow rate = 200 mL min1) at 800
oC for 10 h in a tubular furnace, and the obtained samples are denoted as PdGaOx/Al2O3-HT
(PdGaOx/Al2O3-HRT), Pd/Al2O3-HT (Pd/Al2O3-HRT), PdGaOx/Al2O3-T, and Pd/Al2O3-T, in
which the “HT”, “HRT”, and “T” denote the “hydrothermal”, “hydro-reaction-thermal”, and
“thermal” treatments, respectively.
For comparison purposes, we also prepared the GaOx/Al2O3 and Pd/GaOx/Al2O3 samples via a
typical impregnation route. The desired amount of Ga(NO3)3 was dissolved in deionized water, a
certain amount of Al2O3 (Ga2O3 theoretical loading = 1.5 wt%) was impregnated with the
Ga-containing solution and then dried in an oven. The as-received solid was calcined in air at 450
oC for 5 h. For the preparation of Pd/GaOx/Al2O3, the above-prepared Pd nanocrystals were loaded
26
This article is protected by copyright. All rights reserved.

10.1002/cctc.201801684
ChemCatChem
on GaOx/Al2O3 and subsequently calcined in air at 450 oC for 5 h.
Catalyst characterization: All of the samples were characterized by means of techniques, such as
X-ray diffraction (XRD), inductively coupled plasma atomic emission spectroscopic (ICP‒AES),
transmission electron microscopy (TEM), selected-area diffraction (SAED), high angle annular dark
field-scanning transmission electron microscopy (HAADF‒STEM) and energy dispersive X-ray
spectroscopic (EDX) elemental mapping, electron energy loss spectroscopy (EELS), Laser Raman
(Raman), N2 adsorption‒desorption (BET), X-ray photoelectron spectroscopy (XPS), hydrogen
temperature-programmed reduction (H2-TPR), oxygen temperature- programmed desorption
(O2-TPD), temperature-programmed oxidation (TPO), and in-situ diffuse reflectance Fourier
transform infrared spectroscopy (in situ DRIFT). The detailed characterization procedures are
described in the Supporting information.
Catalytic activity evaluation: Methane combustion activities were carried out in a fixed-bed
quartz tubular microreactor (i.d. 6 mm). 50 mg of the sample (40–60 mesh) was diluted with 0.50 g
of quartz sands (40–60 mesh) to minimize the effect of hot spots. Before the test, each sample was
o
treated in an oxygen flow of 20 mL min1 at 300 C for 1 h. Reaction conditions were as follows:
0.5 vol% CH4 + 4 vol% O2 + 95.5 vol% N2, with a total flow rate of 66.6 mL min1, corresponding
to a space velocity (SV) of about 80,000 mL (g h)1. In the case of CO2 addition, 5.0 vol% CO2
from a CO2 cylinder (balanced with N2) was introduced via a mass flow controller. In the case of
SO2 introduction, 50 ppm SO2 from a SO2 cylinder (balanced with N2) was introduced to the
reaction system through a mass flow controller. In the case of water vapor introduction, 3.0 or 5.0
vol% H2O was introduced by passing the feed stream through a water saturator at a certain
temperature. Reactants and products were analyzed online by gas chromatography (GC-14C,
Shimadzu) equipped with a flame ionization detector (FID) and a thermal conductivity detector
27
This article is protected by copyright. All rights reserved.

10.1002/cctc.201801684
ChemCatChem
(TCD), using a Stabilwax column (30 m in length) for methane separation and a Carboxen 1000
column (3 min length) for permanent gas detection. The balance of carbon throughout the catalytic
system was estimated to be 99.0 ± 1.5 %. Catalytic activities of the samples were evaluated using
the temperatures (T10%, T50% and T90%) required for achieving methane conversions of 10, 50 and
90 %, respectively. The activities were evaluated in terms of CH4 conversion defined as (cin ‒
cout)/cin × 100 %, with cin and cout being methane concentration corresponding to the inlet and outlet,
respectively. In order to better compare the catalytic performance, we calculated the specific
reaction rate (mol gPd‒1 s‒1) according to the activity data and mass of Pd in the as-obtained catalysts
(Table 2). To exclude the heat transfer effect, we calculated the specific reaction rate using the
methane conversions of lower than 10 %.
Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.
Acknowledgements
This work was supported by the National Natural Science Foundation of China (Grant
21876006, 21607005, 21677004, 21622701, 21477005, and U1507108), Beijing Nova Program
(Z141109001814106), Natural Science Foundation of Beijing Municipal Commission of Education
(KM201710005004, KM201410005008), and Scientific Research Base Construction–Science and
Technology Creation Platform National Materials Research Base Construction.
Conflict of Interest
The authors declare no conflict of interest.
Keywords
Supported bimetallic catalyst; Pd5Ga3 nanocrystals; Pd-GaOx/Al2O3; hydrothermal stability;
methane combustion.
28
This article is protected by copyright. All rights reserved.

10.1002/cctc.201801684
ChemCatChem
References
[1] G. Ercolino, P. Stelmachowski, G. Grzybek, A. Kotarba, S. Specchia, Appl. Catal. B 2017, 206,
712.
[2] G. Ercolino, P. Stelmachowski, S. Specchia, Ind. Eng. Chem. Res. 2017, 56, 6625.
[3] N. N. Clark, D.L. McKain, D. R. Johnson, W. S. Wayne, H. Li, V. Akkerman, C. Sandoval, A. N.
Covington, R. A. Mongold, J. T. Hailer, O. J. Ugarte, Environ. Sci. Technol. 2017, 51, 968.
[4] M. Honkanen, T.W. Hansen, H. Jiang, M. Kärkkäinen, M. Huuhtanen, O. Heikkinen, K.
Kallinen, J. Lahtinen, R.L. Keiski, J. B. Wagner, M. Vippol, J. Catal. 2017, 349, 19.
[5] J. H. Chen, W.B. Shi, X. Y. Zhang, H. Arandiyan, D. F. Li, J. H. Li, Environ. Sci. Technol. 2011,
45, 8491.
[6] G. Ercolino, P. Stelmachowski, G. Grzybek, A. Kotarba, S. Specchia, Appl. Catal. B 2017, 206,
712.
[7] T. V. Choudhary, S. Banerjee, V.R. Choudhary, Appl. Catal. A 2002, 234, 1.
[8] M. Cargnello, J. J. Delgado Jaén, J. C. Hernández Garrido, K. Bakhmutsky, T. Montini, J. J.
Calvino Gámez, R. J. Gorte, P. Fornasiero, Science 2012, 337, 713.
[9] Y. Wang, H. Arandiyan, J. Scott, M. Akia, H. X. Dai, J. G. Deng, K. F. Aguey-Zinsou, R. Amal,
ACS Catal. 2016, 6, 6935.
[10] J. Jones, H. F. Xiong, A. T. DeLaRiva, E. J. Peterson, H. Pham, S. R. Challa, G. S. Qi, S. Oh, M.
H. Wiebenga, X. I. P. Hernández, Y. Wang, A. K. Datye, Science 2016, 353, 150.
[11] M. Hatanaka, N. Takahashi, N. Takahashi, T. Tanabe, Y. Nagai, A. Suda, H. Shinjoh, J. Catal.
2009, 266, 182.
[12] S. Colussi, A. Gayen, M. F. Camellone, M. Boaro, J. Llorca, S. Fabris, A. Trovarelli, Angew.
Chem. Int. Ed. 2009, 48, 8481.
29
This article is protected by copyright. All rights reserved.