Stabilizing Oxygen Vacancies in ZrO2by Ga2O3Boosts the Direct Dehydrogenation of Light Alkanes

Article abstract of DOI:10.1021/acscatal.1c02621

The conversion of light alkanes to olefins (e.g., ethylene, propylene, or butylene) is crucial to the chemical industry. ZrO2 with oxygen vacancies has recently been regarded as a promising catalyst for the direct dehydrogenation of light alkanes. However, the intrinsic mechanism of the effect of oxygen vacancies on catalytic performance has not been completely understood yet, and ZrO2 without promoters generally displays poor activity toward the direct dehydrogenation of light alkanes. In this work, we demonstrate that the oxygen vacancies in ZrO2 can be poisoned by H atoms during the dehydrogenation of light alkanes, and we report a strategy for stabilizing the oxygen vacancies in ZrO2 by Ga2O3. Experimental results and theoretical calculations indicate that ZrO2 with oxygen vacancies is responsible for dehydrogenation, while Ga2O3 prevents the poisoning of oxygen vacancies by dissociated hydrogen atoms which, in the absence of the Ga2O3 component, blocks further dehydrogenation. Consequently, the optimal Zr0.26Ga1 catalyst exhibits superior propane dehydrogenation performance to the industrial Pt-Sn catalyst, the state-of-the-art catalyst for the direct dehydrogenation of light alkanes. We anticipate this work may shed light on both the fundamental research of catalysis and the chemical industry.

Full text of DOI:10.1021/acscatal.1c02621

pubs.acs.org/acscatalysis
Research Article
Stabilizing Oxygen Vacancies in ZrO2 by Ga2O3 Boosts the Direct
Dehydrogenation of Light Alkanes
Yong Xu,* Xuchun Wang, Di Yang, Zeyuan Tang, Muhan Cao, Huicheng Hu, Linzhong Wu, Lijia Liu,
John McLeod, Haiping Lin,* Youyong Li, Yeshayahu Lifshitz,* Tsun-Kong Sham, and Qiao Zhang*
Cite This: ACS Catal. 2021, 11, 10159−10169
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: The conversion of light alkanes to olefins (e.g., ethylene,
propylene, or butylene) is crucial to the chemical industry. ZrO2 with oxygen
vacancies has recently been regarded as a promising catalyst for the direct
dehydrogenation of light alkanes. However, the intrinsic mechanism of the
effect of oxygen vacancies on catalytic performance has not been completely
understood yet, and ZrO2 without promoters generally displays poor activity
toward the direct dehydrogenation of light alkanes. In this work, we
demonstrate that the oxygen vacancies in ZrO2 can be poisoned by H atoms
during the dehydrogenation of light alkanes, and we report a strategy for
stabilizing the oxygen vacancies in ZrO2 by Ga2O3. Experimental results and
theoretical calculations indicate that ZrO2 with oxygen vacancies is responsible
for dehydrogenation, while Ga2O3 prevents the poisoning of oxygen vacancies
by dissociated hydrogen atoms which, in the absence of the Ga2O3
component, blocks further dehydrogenation. Consequently, the optimal
Zr0.26Ga1 catalyst exhibits superior propane dehydrogenation performance to the industrial Pt−Sn catalyst, the state-of-the-art
catalyst for the direct dehydrogenation of light alkanes. We anticipate this work may shed light on both the fundamental research of
catalysis and the chemical industry.
KEYWORDS: light alkanes, light olefins, low-coordinated ZrO2-Ga2O3, oxygen vacancies, H diffusion
ZrO2 for PDH, the intrinsic mechanism of the effects of oxygen
vacancies on PDH performance has not been completely
understood yet, and ZrO2 without promoters generally displays
poor activity toward the direct dehydrogenation of light
alkanes. For instance, amorphous (or highly dispersed) ZrO2
displays poor activity for PDH, even though it has plenty of
oxygen vacancies.24 Composting ZrO2 with other promoters,
such as Cr, La, Y, and Zn,19−23 can significantly promote the
PDH activity, nevertheless, the PDH performance of cheap
oxide catalysts is inferior to those of Pt-based catalysts.
Inspired by those reports, we speculate that the active sites
of defective ZrO2 might be poisoned during dehydrogenation
in the absence of promoters. Herein, we demonstrate an
efficient promoter (Ga2O3) for stabilizing the oxygen vacancies
in ZrO2. Density functional theory (DFT) calculations show
that 4-coordinated Zr atoms at the ZrO2−Ga2O3 interface are
the energetically favorable sites for the adsorption of propane
1. INTRODUCTION
The conversion of light alkanes to olefins (e.g., ethylene,
propylene, or butylene) is crucial to the chemical industry.1−3
Traditionally, light olefins are produced by catalytic cracking of
oil fractions and byproducts (e.g., naphtha). The fast-growing
demand for light olefins and the limited petroleum reserves
have motivated the search for alternative olefin production
methods. Shale gas that mainly consists of light alkanes is
considered an important economical substitute of crude oil.
The catalytic dehydrogenation of light alkanes to generate the
corresponding olefins is therefore of tremendous signifi-
cance.4−7 Pt-based catalysts are widely used for the catalytic
dehydrogenation of light alkanes,8−13 and Pt−Sn catalyst has
been industrialized for propane dehydrogenation (PDH).
However, the Pt−Sn catalyst suffers the disadvantages of
high cost and sintering.14−18 Replacing Pt-based catalysts with
cheap catalysts for the dehydrogenation of light alkanes is
therefore of significance, yet remains a great challenge.
Received: June 11, 2021
Revised: July 22, 2021
Recently, ZrO2 has been deemed as a promising catalyst for
PDH. It has been widely reported that H2 treatment will
produce oxygen vacancies in ZrO2, resulting in the formation
of low-coordinated Zr active sites for significantly enhancing
the dehydrogenation activity.19−23 Despite the experimental
observation and theoretical prediction of oxygen vacancies in
https://doi.org/10.1021/acscatal.1c02621
© XXXX American Chemical Society
ACS Catal. 2021, 11, 10159−10169
10159

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
molecules. The subsequent successive C−H bond activations
are then catalyzed by the low-coordinated (4-coordinated) Zr
atoms. Ga2O3 promotes the diffusion of dissociated hydrogen
atoms from vacancies to regenerate the low-coordinated Zr
atoms. Moreover, Ga2O3 strongly suppresses coke formation
during PDH. Consequently, the optimal Zr0.26Ga1 catalyst
exhibits superior PDH performance to the industrial Pt−Sn
catalyst, the state-of-the-art catalyst for the direct dehydrogen-
ation of light alkanes. Additionally, the suppression of coke
deposition of Ga2O3 has been further validated by other Mo−,
W−, Co−, Ni−, Y−, and Fe− modified Ga oxide composites.
with argon and heated to 150 °C for 60 min. After that, the
sample was heated to 900 °C at a heating rate of 10 °C/min
under a flow of 10% H2/Ar (50 mL/min). Temperature-
programmed oxidation (TPO) was performed on an automatic
chemical adsorption instrument (FINESORB-3010). 100 mg
of the catalyst was put in a U-shape quartz tube and pretreated
by H2/Ar (50 mL/min) at 600 °C for 2 h, followed by flushing
with Ar (50 mL/min) at 600 °C for 0.5 h. Afterward, the
sample was cooled to 30 °C in the same Ar flow. O2/Ar (20
mL/min, 2 vol % in Ar) was introduced into the U-shape
quartz tube, and the sample was heated to 600 °C with a
heating rate of 10 °C/min. The temperature and current of
TCD were 60 °C and 70 mA, respectively. Coke deposition
was determined by using thermogravimetry (Mettler Toledo,
TGA1). Specifically, the preweighted catalyst sample (about
100 mg) was flushed with Ar for 1 h at 150 °C and then cooled
to room temperature. After that, the sample was heated to
about 800 °C in air with the heating rate of 10 °C/min. The
coke deposition was determined by the weight loss in the
temperature range of 300−650 °C. Synchrotron experiments
were conducted at beamline BL01C1 at the National
Synchrotron Radiation Research Centre, Hsinchu, Taiwan.
The storage ring was operated at 1.5 GeV with a current of 300
mA; further measurements were also conducted at beamline
BL14W1 at the Shanghai Synchrotron Radiation Facility. The
storage ring was operated at 3.5 GeV with a current of 250 mA.
The Zr K-edge and Ga K-edge were measured using the same
configuration. The samples were pressed into pellets and sealed
with Kapton tape. The thicknesses of the pellets were adjusted
to reach the optimum absorption thickness. All spectra were
recorded at room temperature in a transmission mode. The
spectral analysis was performed following the standard
procedure using ATHENA and IEFFIT software package.25−27
The extended X-ray absorption fine structure (EXAFS)
function, χ, was obtained by subtracting the postedge
background from the overall absorption and then normalizing
with respect to the edge jump step.
2. METHODS
2.1. Materials. Zr(NO3)4·5H2O, Ga(NO3)3, Fe(NO3)3·
9H2O, Ni(NO3)2·6H2O, Co(NO3)2·6H2O, (NH4)6Mo7O24·
4 H 2 O , S n C l 2 · 2 H 2 O , K N O 3 , Y ( N O 3 ) 3 · 6 H 2 O ,
(NH4)6H2W12O40, Cr(NO3)3·9H2O, and Al2O3 were obtained
from Sinopharm Chemical Reagent Co., Ltd. H2PtCl6·6H2O
was purchased from Aldrich. The reference materials of ZrO2
(99%) and Ga2O3 (99.99%) were purchased from Aladdin
(China). The aluminosilicate supports of SIRAL10 (SiO2/
Al2O3 = 10/90, weight percentage) and SIRAL30 (SiO2/Al2O3
= 30/70, weight percentage) were obtained from Sasol as gifts.
The industrial Al2O3-supported Pt−Sn catalyst (Pt−Sn(I))
was provided by UOP.
2.2. Catalyst Preparation. All catalysts were prepared
using the wet impregnation method, and the loading
percentage of the catalyst was kept at 4% in weight. To
prepare Zr0.26Ga1, for instance, stoichiometric Zr(NO3)4·5H2O
and Ga(NO3)3 were dissolved in deionized water, and then the
solution was dropwise added into the support to form a slurry.
After that, the slurry was aged at 60 °C for 1 h, followed by
drying at 100 °C for 12 h. The obtained powder was calcinated
in air under 800 °C for 2 h. The CrOx/Al2O3 catalyst was
prepared using a wet impregnation process. Specifically,
stoichiometric Cr(NO3)3·9H2O dissolved in deionized water
was added into a commercial Al2O3 support (purchased from
Sinopharm Chemical Reagent Co., Ltd) dropwise, and the
slurry was dried for 12 h at 100 °C. After that, the catalyst was
heated to 600 °C in air for 2 h.
2.3. Catalyst Characterization. The compositions of
different catalysts were determined by inductivity-coupled
plasma optical emission spectrometry (Variance, VISTA-
MPX). X-ray diffraction (XRD) patterns were collected on a
PANalytical Empyrean machine. The surface area of the
catalysts was measured by the Brunauer−Emmett−Teller
method using nitrogen adsorption and desorption isotherms
on a Micrometrics ASAP 2020 system. X-ray photoelectron
spectroscopy (XPS) spectra were collected on a KRATOS
Analytical-KRATOS AXIS Ultra DLD spectrometer (Kratos
Analytical-A Shimadzu group company) using a monochro-
matic Al Kα source (1486.6 eV). The binding energy scale was
calibrated according to the C 1s peak (284.8 eV) of
adventitious carbon on the analyzed sample surface. Trans-
mission electron microscopy images were collected on a FEI-
Tecnai G2 F20 (200 kV) microscope equipped with energy-
dispersive X-ray spectroscopy (EDX) functionality. Electron
paramagnetic resonance (EPR) spectra were obtained by using
the JEOL machine (JES-FA200) with the microwave frequency
range of 8750-9650 MHz (X-band). H2 temperature-
programmed reduction (H2-TPR) was performed using a
TPR analyzer equipped with a thermal conductivity detector
(TCD) detector. The catalyst sample (40 mg) was flushed
EXAFS fitting over the Zr K-edge was performed on ZrO2
on the Zr0.26Ga1 catalysts (before and after H2 treatment) and
on the metallic Zr reference over a k-range of 2−15 Å−1 and an
R-range of 1−5 Å. The fitting was performed in the R-space
with a k-weight of 2. Fitting of ZrO2 was performed using
monoclinic ZrO2 (space group P21/c) as a model. Monoclinic
ZrO2 features Zr coordinated by seven O atoms, all with
unique bond lengths. Since many of these bond lengths differ
by less than 0.05 Å (essentially the resolution of EXAFS), we
simplified the model by grouping the O atoms into three
coordination shells. Reassuringly, the fit was excellent (r-factor
0.044) and the bond lengths were within error of those in the
reported monoclinic crystal structure.
Because the Zr EXAFS of the Zr−Ga catalysts is almost the
same as that of ZrO2 and because the features in the Fourier
transformed EXAFS of the Zr−Ga catalysts are aligned with
those in ZrO2, to fit the EXAFS from the Zr−Ga catalysts, we
constrain the bond lengths, energy shift (E0), and amplitude
2
reduction factor (S0 ) to those obtained from fitting the ZrO2
EXAFS, and instead we allow the coordinate number
(previously fixed for ZrO2) to be fitted. The Debye−Waller
disorder factor is fitted for all cases. Because we do not expect
H2 treatment to remove Zr atoms, we also constrain the Zr−Zr
coordination number to be consistent with the EXAFS of the
Zr−Ga catalysts before and after H2 treatment. The fitting
space, range, and k-weight were kept constant with those for
10160
https://doi.org/10.1021/acscatal.1c02621
ACS Catal. 2021, 11, 10159−10169

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
ZrO2. The fit was good (r-factor 0.075 for simultaneously
fitting the spectra from catalysts before and after H2
treatment), although better r-factors could be obtained by
changing the variables that were fit. This procedure often
resulted in unphysical fitted values. As our present approach
results in physically valid parameters that suggest a structural
model in agreement with the EXAFS and is closely tied to the
ZrO2 reference fit, we believe our approach is the most
appropriate method for studying this system.
(model: TM-Al2O3/S), and N2 and H2 were detected by a
GC with TCD using a packed column (model: TDX-01).
Catalytic performance was evaluated using eqs 1−7. 0.4 g of
the catalyst with a size of 60−80 mesh mixed with 1.2 g of
quartz glass (60−80 mesh) was fixed in the middle of the
reactor. Each ethane dehydrogenation (EDH) test consisted of
(i) a 15 min reduction step at 600 °C (heating rate: 10 °C/
min) using a H2/N2 flow (volume ratio: 10:90%), (ii) a 15 min
EDH step at 650 °C using a C2H6/H2/N2 flow (flow rate:
12:12:100 mL/min) without specific note, and (iii) a 15 min
dry air treatment step at 750 °C; each PDH test consisted of
(i) a 15 min reduction step at 600 °C (heating rate: 10 °C/
min) using a H2/N2 flow (volume ratio: 10:90%), (ii) a 15 min
PDH step at 600 °C using a C3H8/H2/N2 flow (flow rate:
12:12:100 mL/min) without specific note, and (iii) a 15 min
dry air treatment step at 750 °C; each i-butane dehydrogen-
ation (BDH) test consisted of (i) a 15 min reduction step at
600 °C (heating rate: 10 °C/min) using a H2/N2 flow (volume
ratio: 10:90%), (ii) a 15 min i-BDH step at 550 °C using a i-
C4H10/H2/N2 flow (flow rate: 12:12:100 mL/min) without
specific note, and (iii) a 15 min dry air treatment step at 750
°C. The propane conversion and propylene selectivity were
measured three times to get the mean values. A series of blank
experiments were performed by using only quartz glass to
study the possible cracking of light alkanes. It should be
pointed out that no conversion of alkanes was obtained during
one cycle (15 min). The propane conversion and propylene
selectivity were obtained in 15 min without specific note.
2.5. Definition of Conversion, Selectivity, and Yield.
EXAFS fitting over the Ga K-edge was performed on Ga2O3
on both Zr0.26Ga1 catalysts (before and after H2 treatment) and
the metallic Ga reference over a k-range of 2−15 Å−1 and an R-
range of 1−5 Å. The fitting was performed in the R-space with
a k-weight of 2. Fitting of ZrO2 was performed using
monoclinic Ga2O3 (space group C2/m) as a model.
Monoclinic Ga2O3 has two distinct Ga sites, one of which is
4-coordinated and one of which is 6-coordinated. In principle,
this presents a difficulty for EXAFS fitting as each site may
have a distinct onset energy. However, because both Ga sites
are expected to be Ga3+, and as model full-potential, all-
electron calculations suggest that the difference in the X-ray
absorption onset for the two sites is only around 0.1 eV, we
simplify and use only a single energy shift E0 in our fitting. We
also simplify the structure by grouping atoms with very similar
bond lengths into a single coordination shell and fixing the
coordination number of these shells as the average between the
two Ga sites. The fit was excellent (r-factor = 0.037). The
fitted bond lengths were generally slightly larger than those of
the ideal crystal structure but not by much. This may be due to
the simplifications employed in our model mentioned above
(forming approximate coordination shells, treating both Ga
sites as having identical onset energies) or may simply be due
to local heating by the X-ray beam expanding the Ga2O3
lattice.
n(C3H8)in − n(C3H8)out
C3H8 con. (%) =
× 100%
n(C3H8)in
(1)
(2)
n(i‐C4H10)in − n(i‐C4H10)out
n(i‐C4H10)in
For the Ga EXAFS of the Zr−Ga catalysts, we again
i‐C4H10 con. (%) =
2
constrained the amplitude reduction factor (S0 ) to those
obtained from the Ga2O3 reference fitting. The k- and r-ranges
and the fitting space are the same as used for Ga2O3. We also
constrained the Debye−Waller disorder parameters (we
employ one for scattering from O and one for scattering
from Ga) to those obtained from Ga2O3 fitting. This choice
was motivated by simplifying the fitting model; as the resulting
fit is quite good and provides physically reasonable parameters,
we believe this simplification was justified. We further reduce
the number of scattering paths as the EXAFS transform from
the Zr−Ga catalysts has fewer features than that from Ga2O3.
However, for the Zr−Ga catalysts, we fit both the coordination
number and the bond length as the Fourier transform (FT)
suggests small shifts in bond length compared to Ga2O3. We
also fit the E0 as the EXAFS for the Zr−Ga catalysts features a
noticeably different absorption onset compared to Ga2O3. E0
and the bond lengths are, however, fitted simultaneously for
both Zr−Ga catalysts. The fitted energy shift places the
absorption onset 1.3 0.6 eV lower than in Ga2O3, consistent
with what we expect from the EXAFS. The fit was excellent (r-
factor = 0.039) for simultaneous fitting of spectra from before
and after H2 treatment.
× 100%
n(C2H6)in − n(C2H6)out
C2H6 con. (%) =
C3H6 sel. (%) =
× 100%
n(C2H6)in
(3)
(4)
n(C3H6)out
× 100%
n(C3H8)in − n(C3H8)out
n(i‐C4H8)out
n(i‐C4H10)in − n(i‐C4H10)out
i‐C4H8 sel. (%) =
× 100%
(5)
n(C2H4)out
C2H4 sel. (%) =
× 100%
n(C2H6)in − n(C2H6)out
m(C3H6)
(6)
C3H6 STY=
mCat. × t
(7)
(8)
(9)
m(i‐C4H8)
mCat. × t
2.4. Catalytic Evaluation. All catalysts were evaluated in a
fixed quartz reactor with the inner diameter of 10 mm. The
flow rates of C3H8, N2, and H2 were controlled by mass flow
controllers. The outlet gases were detected by two online gas
chromatographs (GCs). Hydrocarbons were analyzed by a GC
with flame ionization detector using a capillary column
i‐C4H8 STY =
C2H4 STY =
m(C2H4)
mCat. × t
10161
https://doi.org/10.1021/acscatal.1c02621
ACS Catal. 2021, 11, 10159−10169

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Figure 1. (a) EPR spectra of the Zr0.26Ga1 catalyst before (black line) and after (red line) H2 treatment. In situ IR spectra for PDH on ZrO2 (b),
Ga2O3 (c), and Zr0.26Ga1 catalysts (d). In situ IR measurements were performed in the mixture of C3H8, N2, and H2. IR conditions: Vtotal = 3 mL,
Ptotal = 0.12 MPa, T = 600 °C, and m(Zr0.26Ga1) = 50 mg.
where n and m represent the mole and mass, respectively. For
example, n(C3H8)in and n(C3H6)out represent the mole of
propane introduced and the mole of propylene produced,
respectively. The mole of propylene was obtained through the
division of the peak area of propylene in the GC pattern and
that of the standard (propylene with a different volume
fraction). N2 was used as a standard to calibrate the volume
changes for PDH (C3H6 → C3H6 + H2).
the reaction temperature (823.15 K) and ambient pressure
according to the following equation
G = H − TΔS = EDFT + EZPE
+
873 Cv dT − TΔS
∫
0
where EDFT is the DFT-optimized energy, EZPE is the zero-
point energy, ∫ 0873Cv dT is the heat-capacity, T is the
temperature, and ΔS is the entropy. The ideal gas model
was used when calculating the thermodynamic properties of
gas-phase molecules like H2, C3H8, and C3H6, while adsorbates
on surfaces were treated using harmonic approximation.
2.6. DFT Calculations. All DFT calculations were carried
out using the Vienna Ab Initio Simulation code.28,29 The
interaction between the valence electrons and the ions were
described with the project-augmented-wave30 scheme and a
planewave energy cutoff of 400 eV. The exchange−correlation
interactions were treated within the generalized gradient
approximation in the form of Perdew−Burke−Ernzerhof
functional.31 The van der Waals interactions were described
using the empirical correction in Grimme’s scheme (DFT-
D3).32 As Ga2O3 and ZrO2 are both semiconductors, the
Gaussian smearing with a width of 50 meV was used for the
occupation of electronic levels. The climbing-image nudged
elastic band33 calculations and the dimer method34 were
employed to determine the configurations of transition states.
In all calculations, the surfaces were modeled with periodic
slabs consisting of a vacuum region of 15 Å and several atomic
layers. The two bottom layers of atoms were kept fixed, while
the rest of atoms were fully relaxed until the atomic force was
less than 0.02 eV/Å. The convergence threshold of self-
consistent filed iterations was set to 10−4 eV. Taking the
experimental conditions into consideration, the β-Ga2O3(100)
3. RESULTS AND DISCUSSION
Zr−Ga catalysts with different molar ratios of Zr and Ga were
impregnated into an alumina or an aluminosilicate support
(Tables S1 and S2 and Figures S1 and S2). Moreover, we
characterized the Zr−Ga catalyst by using (EDX, Figure S3),
(XPS, Figure S4), X-ray absorption near edge structure
(XANES), and the FT of the EXAFS (Figures S5 and S6)
measurements. XPS and XANES also confirmed the oxidation
nature of the Ga and Zr constituents. EDX elemental mapping
clearly shows that both Zr and Ga were well dispersed on the
support (Figure S3). We thus concluded that the catalysts
consist of oxides with an extremely small grain size. No
obvious H2 consumption peak was observed in the H2-TPR
profile (Figure S7), which was in good agreement with the
results from the XPS measurement. The intensities of the EPR
signals are proportional to the free electron density, indicating
the formation of oxygen vacancies. The EPR signal intensities
for both ZrO2/SIRAL10 and Zr0.26Ga1/SIRAL10 catalysts
strongly increase after H2 treatment, suggesting that the H2-
treated catalyst contains a significant amount of oxygen
vacancies (Figures 1a and S8). In contrast, Ga2O3 seems to
be stable under H2 treatment, and its EPR signal before and
after H2 treatment is weak (Figure S8). This indicated that the
oxygen vacancies in the Zr−Ga catalysts were formed in the
H2-treated ZrO2 component.19,21 XANES further substantiated
̅
and m-ZrO2(111) were selected to represent the surfaces of
Ga2O3 and ZrO2. The parameters are listed in Tables S10 and
S11. The optimized lattice constants for β-Ga2O3 and m-ZrO2
are in good agreement with the previous theoretical
calculations and experimental measurements. The atomic
simulation environment was used to calculate all thermody-
namic properties.35 The Gibbs free energy G was calculated at
10162
https://doi.org/10.1021/acscatal.1c02621
ACS Catal. 2021, 11, 10159−10169

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Figure 2. PDH performance over Zr−Ga/SIRAL10 catalysts. (a) Propane conversion and propylene selectivity and (b) STY using different Zr−
Ga/SIRAL10 catalysts. The PDH conditions were C3H8/H2/N2 = 12:12:100 mL/min, T = 600 °C, Ptotal = 0.12 MPa, and WHSV = 3.5 h−1.
Figure 3. PDH performance over Pt−Sn(I), 20 wt % CrOx/Al2O3, and Zr0.26Ga1/SIRAL10 catalysts. (a) Propane conversion, (b) propylene
selectivity, and (c) STY of propylene. The PDH conditions were C3H8/H2/N2 = 12:12:100 mL/min, T = 500−625 °C, Ptotal = 0.12 MPa, and
WHSV = 3.5 h−1. (d) Long-term stability in 300 cycles using the Zr0.26Ga1/SIRAL10 catalyst for PDH. The PDH conditions were C3H8/H2/N2 =
12:12:100 mL/min, T = 600 °C, Ptotal = 0.12 MPa, and WHSV = 3.5 h−1.
this observation showing a much stronger intensity change of
Zr K-edge than Ga K-edge in the H2-treated Zr0.26Ga1/
SIRAL10 sample (Figure S5). EXAFS confirmed the lack of a
complete second shell around Zr in the catalyst and the
decrease of Zr−O coordination number after H2 treatment
(Figure S6 and Table S3).
Ga2O3 was active for the PDH reaction. In a striking contrast, a
rapid evolution of the propylene characteristic signals was
observed when the optimized Zr0.26Ga1/SIRAL10 was used.
The intensity of peaks in the range of 2920−2980 cm−1
(Figure 1d) rapidly decreased after 3 min, while the intensity
of the peak at 3015 cm−1 increased accordingly. These results
indicated that the synergy between ZrO2 and Ga2O3 strongly
contributed to the improved PDH performance.
We used in situ infrared (IR) spectrometry to monitor the
PDH performance over ZrO2/SIRAL10, Ga2O3/SIRAL10, and
Zr0.26Ga1/SIRAL10 catalysts The characteristic peaks associ-
ated with C−H vibrations of C3H8 and C3H6 could be easily
distinguished in the IR spectrum (Figure 1b). The character-
istic C−H vibration positions in the IR spectra were
determined by injecting propane (2920−2980 cm−1) and
propylene (3015 cm−1) separately into the in situ IR cell
(Figure S9). When ZrO2/SIRAL10 was used as the catalyst,
only the characteristic peaks corresponding to the C−H
vibrations of propane were observed in the range of 2920−
2980 cm−1, suggesting that ZrO2/SIRAL10 was not active for
the dehydrogenation reaction. In comparison, a small peak at
about 3015 cm−1 corresponding to the C−H vibrations of
propylene was observed after 20 min when pure Ga2O3/
SIRAL10 was used as the catalyst (Figure 1c), indicating that
The catalytic performance of the Zr−Ga catalysts supported
on SIRAL10 with varying Zr/Ga molar ratios was measured in
a lab-scale reactor. The molar ratio of Zr/Ga was found critical
to PDH performance (Figure 2a,b). When ZrO2/SIRAL10 was
used as the catalyst, only 1.8% of propane was converted with a
low selectivity (74.3%). Both the conversion of propane and
the selectivity of propylene increased significantly with the
addition of Ga2O3 (Figure 2a). The propane conversion
reached 50.1% at a propylene selectivity of 95.1% for a Zr/Ga
molar ratio of 0.26/1 (e.g., Zr0.26Ga1/SIRAL10). A further
increase of the amount of Ga2O3 led to a decrease of the
catalytic activity. The propylene yield showed a similar
dependence on the ratio of Zr/Ga. As depicted in Figure 2b,
−1
the space time yield (STY) of propylene was 0.045 kg kgCat.
10163
https://doi.org/10.1021/acscatal.1c02621
ACS Catal. 2021, 11, 10159−10169

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Figure 4. (a) i-BDH performance using the Pt−Sn(I), 20 wt % CrOx/Al2O3, and Zr0.26Ga1/SIRAL10 catalysts. The i-BDH conditions were i-
C4H10/H2/N2 = 12:12:100 mL/min, T = 550 °C, Ptotal = 0.12 MPa, and WHSV = 3.5 h−1. (b) EDH performance using the Pt−Sn(I), 20 wt %
CrOx/Al2O3, and Zr0.26Ga1/SIRAL10 catalysts. The EDH conditions were C2H6/H2/N2 = 12:12:100 mL/min, T = 650 °C, Ptotal = 0.12 MPa, and
WHSV = 3.5 h−1.
−1
h−1 for ZrO2/SIRAL10, which increased to 1.61 kg•kgCat.
h−1 for the optimized Zr0.26Ga1/SIRAL10. It should be noted
that the PHD activity of Zr0.26Ga1/SIRAL10 surpassed that of
the industrial Pt−Sn catalyst [noted as Pt−Sn(I) and obtained
from UOP] and most of the reported catalysts for PDH (Table
S4).
stage was used and no H2 was fed during the PDH stage
(Table S7). The conversion increased to 46.9% when the H2
treatment was performed at 600 °C and further increased to
50.1% after the H2 treatment and the addition of H2 to the
PDH feeding gas (Figure S18a). The observed enhancement in
activity after H2 treatment might be attributed to the formation
of O vacancies at the ZrO2−Ga2O3 interface leading to 4-
coordinated Zr active sites. The effect of H2 introduction to
the PDH feeding gas may be associated with the reported
ability of H2 to retard the coke formation during dehydrogen-
ation.37,38 Our experiments have indeed confirmed that the
coke deposition on various Zr−Ga catalysts significantly
decreased after the addition of H2 to propane (Table S8).
Furthermore, similar results of suppression of coke deposition
due to H2 feeding gas were observed on the 20 wt % CrOx/
Al2O3 and Pt−Sn(I) catalysts (Figure S18b,c and Table S9).
In addition to PDH, the Zr0.26Ga1/SIRAL10 catalyst was a
universal catalyst for the catalytic dehydrogenation of light
alkanes to light olefins. The present work also studied the
properties of Zr0.26Ga1/SIRAL10 for i-BDH and EDH
reactions. As shown in Figure 4, Zr0.26Ga1 exhibited a superior
catalytic performance for the EDH reaction compared to the
Pt−Sn(I) and CrOx/Al2O3 catalysts. For example, Zr0.26Ga1/
SIRAL10 allowed a 95.9% selectivity to ethylene at 37.8%
ethane conversion, while the Pt−Sn(I) catalyst allowed a
96.2% ethylene selectivity at 33.9% ethane conversion. When
CrOx/Al2O3 was used as the catalyst, the ethane conversion
and ethylene selectivity were 30.1 and 92.4%, respectively.
Similarly, Zr0.26Ga1/SIRAL10 presented a higher i-butane
conversion and comparable i-butylene selectivity in i-BDH
(Figure 4b). In addition, the Zr0.26Ga1 catalyst was tested for
20 consecutive dehydrogenation−regeneration cycles for both
EDH and i-BDH. No decay in conversion, selectivity, and yield
was observed, indicating the applicability of Zr0.26Ga1/
SIRAL10 to dehydrogenations of a variety of different alkanes
(Figures S19 and S20).
Moreover, home-made Pt−Sn and CrOx/Al2O3 were used as
another reference (Figures S10−S13 and Table S5). It was
found that Zr0.26Ga1/SIRAL 10 yielded a larger improvement
of the PDH performance with respect to Pt−Sn(I) and 20 wt
% CrOx/Al2O3 under different temperatures (Figure 3a−c).
Similar superior PDH properties of Zr0.26Ga1 with respect to
Pt−Sn(I) and 20 wt % CrOx/Al2O3 were observed when PDH
was performed under different weight hourly space velocities
(WHSVs) (Figures S14). Zr0.26Ga1/SIRAL10 initiated a
propylene STY of 1.61 kg kgCat.−1•h−1, while Pt−Sn(I) and
20 wt % CrOx/Al2O3 catalysts produced lower propylene STYs
(1.33 kg kgCat. h−1 for Pt−Sn(I) and 1.04 kg kgCat. h−1 for
20 wt % CrOx/Al2O3) when the temperature and WHSV were
600 °C and 3.5 h−1, respectively (Figure 3c). In order to
highlight the long-term stability, Zr0.26Ga1/SIRAL10 was tested
for PDH in 300 consecutive dehydrogenation−regeneration
cycles. The propane conversion decreased to about 38.9% after
75 cycles (Figure 3d) and then was almost constant within the
next 225 cycles, indicating the high stability of Zr0.26Ga1/
SIRAL10 for PDH. Note that the Zr0.26Ga1/SIRAL10 catalyst
requires only calcination in air to remove the deposited coke,
while the industrial Pt−Sn catalyst requires additional
oxychlorination.4,13 This is expected to significantly reduce
the cost of PDH when using Zr0.26Ga1/SIRAL10 [in addition
to the reduction due to the much cheaper cost of Zr0.26Ga1/
SIRAL10 with respect to Pt−Sn(I)]. Moreover, additional
EPR measurements were performed on other Zr−Ga catalysts
to further verify the effects of vacancies on PDH. It was found
that the areas of the EPR peak of different catalysts were
strongly correlative with the propane conversion (Figure S15).
Additionally, TPO measurements were employed to quantifica-
tionally measure the vacancies, which further confirmed the
significance of vacancies on PDH (Figure S16 and Table S6).
Moreover, kinetic studies show that the reaction order and
apparent activation energy (Ea) of PDH over Zr0.26Ga1/
SIRAL10 are 1.06 and 65.1 kJ/mol, respectively (Figure S17).
H2 was used in Oleflex PDH process to suppress coke
deposition on Pt−Sn catalysts.36 In the present work, we found
that H2 was quite important for activating the Zr0.26Ga1/
SIRAL10 catalyst in the H2 reduction stage (preceding PDH)
and reducing coke formation in the PDH stage. The initial
propane conversion was about 29.9% when no H2 treatment
−1
−1
Considering the existence of H2 in both reactants and
products, the surface of Zr0.26Ga1/SIRAL10 might be
hydroxylated during PDH. We thus performed additional
experiments to check the surface state of Zr0.26Ga1/SIRAL10.
As shown in Figure S21, typical peaks corresponding to OH
appeared at 3180−3400 cm−1 in the IR spectrum of Zr0.26Ga1
at 100 °C, and the peaks completely disappeared when the
treatment temperature was further increased to 300 °C, which
could be attributed to the dehydration under high temperature.
39
Moreover, we synthesized ZrO(OH)2 and GaOOH,40 two
typical compounds with abundant surface OH, to further
confirm this assumption. Thermogravimetric analysis indicated
10164
https://doi.org/10.1021/acscatal.1c02621
ACS Catal. 2021, 11, 10159−10169

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Figure 5. Energy profile for PDH over the Zr−Ga catalyst. The calculated energy profile of the C−C bond activation (a), first C−H bond
activation (b), second C−H bond activation at the carbon radical (c), and the second C−H bond activation at the methylene group (d). The active
site is a 4-coordinated Zr atom. Starting from an adsorbed propane molecule, the C−H bond activation (b) has a much lower reaction barrier (1.11
eV) than the C−C bond activation (2.78 eV) shown in (a), indicating that the first preferential reaction path is to form CH3CH3CH3* (in red).
Once the first C−H bond activation is occurring, the production of propylene (d) has a very low activation barrier (0.10 eV), suggesting that the
second preferential reaction path is to form propylene rather than over dehydrogenation. The energetically preferred reaction pathways are marked
in red. Note that the dissociated H in the FS of (b,d) is attached to the O lattice atom adjacent to the four-coordinated Zr atom, so that the further
C−H activation is blocked. Removal of the dissociated H atom is necessary for further activation. Indeed, the dissociated H atom was removed
from the O lattice atom in the IS of (b,d) to allow C−H activation.
that both ZrO(OH)2 and GaOOH suffered dehydration at
300−400 °C (Figure S22). No weight loss was observed when
the temperature is higher than 450 °C, suggesting the complete
dehydration. XRD patterns show that both ZrO(OH)2 and
GaOOH are converted into ZrO2 and Ga2O3 after thermal
treatment at 500 °C (Figure S23). Consequently, the Zr and
Ga species presented as the oxidation state on the surface of
catalysts during dehydrogenation.41
plus an O lattice atom plus a Ga lattice atom). It is similar to
the active Zr site for PDH (coordinatively unsaturated Zr
cations) reported by Otroshchenko et al.,21 which is well
consistent with our EXAFS fitting results (Table S3). The
PDH process on a 4-coordinated Zr site as derived from DFT
calculations is presented in Figure 5. The initial states (ISs) of
Figure 5a,b show the propane molecule adsorbed on the 4-
coordinated Zr atom. The first C−H activation (1.11 eV,
Figure 5b) is energetically much more favorable than the C−C
activation (2.78 eV, Figure 3a). The scission of C−H bonds is
thus much faster than that of C−C bonds, resulting in the
propyl radical as the major product. Note that the dissociated
H atom is adsorbed on the nearby O lattice atom [final state
(FS) in Figure 5b] and has to be removed to allow further C−
H activation (note that it is absent in the IS of Figure 5c,d).
The second C−H activation may occur either at the first
carbon atom or at the second (central) carbon atom of the
propyl radical. The calculated energy profiles show that the
dissociation of H atom from the second carbon atom of the
propyl group to produce the propylene molecule is a highly
exothermic reaction, with a very small activation barrier of 0.10
eV and a large energy drop of 1.87 eV. By contrast, the C−H
activation on the first carbon atom is an endothermic reaction
with a much larger energy barrier of 0.91 eV. Again, the
dissociated H atom generated via the second C−H activation
is trapped by the nearby O lattice atom and must be removed
to facilitate further C−H activation. The removal of the
dissociated H atom eliminating poisoning of the active sites
and facilitating further dehydrogenation is feasible by diffusion
of the dissociated H atom to the Ga2O3 surface. The diffusion
barrier (calculated by DFT) is 0.33 eV (Figure S25).
Based on the above results from experiments and character-
izations, DFT calculations were performed to elucidate the
reaction pathway of the selective dehydrogenation of propane
molecules at the ZrO2−Ga2O3 interface and at the ZrO2 and
Ga2O3 surfaces (Tables S10 and S11). The calculations refer to
the different physical−chemical aspects relevant to the
dehydrogenation process which include (i) the nature of the
surface and the different activation sites involved, (ii) the
adsorption energies of the different species participating
(propane, propylene, dissociated H, and intermediate radicals),
and (iii) the C−H and C−C activation barriers at the different
stages of the dehydrogenation process. We first studied the
PDH process at the ZrO2−Ga2O3 interface which was
modeled with an adsorbed Zr4O8 cluster on the Ga2O3(100)
surface [a simple configuration which we assume to be
sufficient to describe the PDH properties since they are
determined (as a first order approximation) by the atoms in
the neighborhood of the active site]. This model contains
(Figure S24) two different types of Zr atoms: (i) a 4-
coordinated Zr atom (bonded to O atoms coming from the
ZrO2) and (ii) a 5-coordinated Zr atom bonded to an O atom
coming from the Ga2O3(100) surface. The number of 4-
coordinated Zr atoms is 3 times larger than that of the 5-
coordinated Zr atoms. The PDH active site is marked by a
circle in Figure S24a (containing a 4-coordinated Zr cation
The full energy and Gibbs profiles of PDH (Figure S26,
Tables S12 and S13) show a similar trend. A DFT analysis of
10165
https://doi.org/10.1021/acscatal.1c02621
ACS Catal. 2021, 11, 10159−10169

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Figure 6. Catalytic performance for PDH over different M-Ga/SIRAL10 catalysts. (a) Propane conversion, (b) propylene selectivity, (c) propylene
STY, and (d) catalytic performance in 20 cycles. The PDH conditions were C3H8/H2/N2 = 12:12:100 mL/min, T = 600 °C, Ptotal = 0.12 MPa, and
WHSV = 3.5 h−1.
the 5-coordinated Zr atom reveals that in accordance with the
4-coordinated Zr atom, the first C−H activation is energeti-
cally much more likely than the C−C activation and the
second C−H activation occurs at the second C atom of the
propyl radical. The C−H energy barriers are about the same as
for the 4-coordinated Zr (1.18 and 0.12 eV) (Figure S27).
Similar to the 4-coordinated Zr system, the dissociated H atom
should be removed to allow further PDH. The H diffusion
barrier to the Ga2O3 in the 5-coordinated system is larger than
for the 4-coordinated system (0.87 vs 0.33 eV, Figure S25).
Moreover, the diffusion of H in the 5-coordinated system is
reversible (the transition state and the FS have the same
energy), so that the 5-coordinated Zr is more vulnerable to
poisoning. We also calculated the adsorption energies of
propane and propylene on 4- and 5-coordinated Zr (Figure
S28). The 4-coordinated Zr has a lower absorption energy of
propylene (i.e., less over dehydrogenation or other side
reactions). In conclusion, both the 4-coordinated and 5-
coordinated Zr atoms at the ZrO2−Ga2O3 interface are active
for PDH. The 4-coordinated Zr atom is, however, a better
PDH active site. This theoretical analysis is indeed in
accordance with our experimental findings. H2 involvement
generates O vacancies which increase the number of 4-
coordinated Zr atoms, enhancing the number of active sites
and improving the PDH catalytic performance.
Additional DFT calculations probed the PDH activity of
Ga2O3 (Figure S29). The first C−H activation barrier of
Ga2O3 is slightly higher than that of ZrO2−Ga2O3 (1.35
compared to 1.11 and 1.18 eV for the 4- and 5-coordinated Zr
in the ZrO2−Ga2O3 system). The competing C−C activation
reaction is very unlikely with an activation barrier of 3.67 eV.
The C−H activation barrier of the central C atom leading to
propylene production is low (0.25 eV). As for Zr at the ZrO2−
Ga2O3 interface, the dissociated H atom sticks to the O lattice
atom in both the first and second C−H activation and must be
removed to facilitate further PDH. The diffusion energy barrier
needed to remove the dissociated H atom is 1.19 eV (Figure
S30), significantly larger than that of the ZrO2/Ga2O3 system
(0.33 eV, Figure S25). Note that O vacancies do not improve
the PDH catalytic properties of Ga2O3 [the first C−H
activation barrier of defective Ga2O3 is 1.46 eV (Figure S31)
compared to 1.35 eV (Figure S29) for pristine Ga2O3 (Figure
S32)]. This is in accordance with the experimental data which
shows that H2 involvement does not affect the PDH activity of
Ga2O3.
Additionally, DFT calculations were also used to investigate
̅
the pure ZrO2 system (Figure S33). Pristine m-ZrO2(111)
surface (containing 6- and 7-coordinated Zr atoms) has a very
large C−H activation barrier (1.73 eV) which explains its very
low propane conversion (Figure S34). Defective m-ZrO2(111)
̅
with 5- and 6-coordinated Zr sites (Figure S33c) possesses a
very low C−H activation barrier of 0.01 eV (Figure S35). The
poisoning of the active site by strong adsorption of the
dissociated H however blocks the further PDH process. The
trapping of the dissociated H by the adjacent O lattice atom is
very strong and requires at least 2.64 eV for its removal (Figure
S36). Moreover, DFT calculations were conducted to evaluate
the feasibility of H2 formation by two dissociated H atoms.
The calculated activation barrier is up to 2.15 eV, suggesting
that the H2 recombination near the active site is a very slow
step compared to the H diffusion (Figure S37). Note that our
energy profile of the PDH process of the m-ZrO2(1
̅
11) system
is similar to that of Zhang et al.24 They show that the PDH of
the defective m-ZrO (111) catalyst has a very low energy
̅
2
barrier, but the final desorption of H2 required for completion
of the PDH process is endothermic by 1.40 eV.24
The Zr−Ga system indicates that one can design a binary
component catalyst in which Ga2O3 suppresses poisoning and
coke formation. To generalize this strategy and confirm it to
components different than Zr, we first checked the PDH
performance and coke formation of single-component
materials. Ga2O3 obviously has a very low coke formation
rate, orders of magnitude lower than that of the other catalysts
(Table S14). The PDH performance of the single component
10166
https://doi.org/10.1021/acscatal.1c02621
ACS Catal. 2021, 11, 10159−10169

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
catalysts is very low (high coke rate formation leading to low
yield), except Ga2O3 (low coke rate formation and medium
yield). The addition of Ga2O3 to the other metallic
components (Fe, Y, Ni, Co, W, Mo, and Zr) indeed suppresses
coke formation and enhances PDH performance (Figures
S38−S43 and Tables S15−S20). The exact nature of the
processes occurring in the Zr0.26Ga1 nanocomposite catalyst
was thoroughly investigated in the present work. In contrast,
the other M−Ga oxide catalysts may differ for different
components and should be checked in future studies (both
experimentally and by DFT calculations). However, the
rational catalyst strategy was validated. All M−Ga composite
catalysts surpass the activity of Ga2O3, and Mo0.2Ga1/SIRAL10
even compares well with Pt−Sn(I), though it is still inferior
with respect to Zr0.26Ga1/SIRAL10. These optimized catalysts
exhibited an excellent catalytic performance in terms of
propane conversion (31.8−43.2%), propylene selectivity
(>89%), and stability (Figure 6). It is well known that iron-
based materials are poor catalysts for nonoxidative dehydro-
genation of alkanes because they can activate the C−C bond
and form coke quickly, leading to low selectivity and short life
times.41 We show herein that a Fe-based catalyst can however
also activate the C−H bond. so that it can be used to modify
Ga2O3 and form a good alkane dehydrogenation catalyst.
Indeed, the Fe−Ga catalyst shows a good catalytic perform-
ance with a propane conversion of 31.8%, a propylene
AUTHOR INFORMATION
■
Corresponding Authors
Yong Xu − Guangzhou Key Laboratory of Low-Dimensional
Materials and Energy Storage Devices, Collaborative
Innovation Center of Advanced Energy Materials, School of
Materials and Energy, Guangdong University of Technology,
Guangzhou 510006, China; Institute of Functional Nano &
Soft Materials (FUNSOM), Jiangsu Key Laboratory for
Carbon-Based Functional Materials & Devices, Soochow
University, Suzhou 215123, PR China; orcid.org/0000-
0001-5115-6492; Email: yongxu@gdut.edu.cn
Haiping Lin − Institute of Functional Nano & Soft Materials
(FUNSOM), Jiangsu Key Laboratory for Carbon-Based
Functional Materials & Devices, Soochow University, Suzhou
215123, PR China; Email: hplin@suda.edu.cn
Yeshayahu Lifshitz − Institute of Functional Nano & Soft
Materials (FUNSOM), Jiangsu Key Laboratory for Carbon-
Based Functional Materials & Devices, Soochow University,
Suzhou 215123, PR China; Department of Materials Science
and Engineering, Technion-Israel Institute of Technology,
Haifa 3200003, Israel; Email: shayli@technion.ac.il
Qiao Zhang − Institute of Functional Nano & Soft Materials
(FUNSOM), Jiangsu Key Laboratory for Carbon-Based
Functional Materials & Devices, Soochow University, Suzhou
215123, PR China; orcid.org/0000-0001-9682-3295;
Email: qiaozhang@suda.edu.cn
−1
selectivity of 91.1%, and a propylene STY of 0.98 kg kgCat.
h−1, further confirming our proposed catalytic scheme (Figure
6). The modification of Ga2O3 by metal/metal oxides is thus a
universal and versatile approach for preparing inexpensive and
highly efficient catalysts for conversion of alkanes to olefins.
Authors
Xuchun Wang − Institute of Functional Nano & Soft
Materials (FUNSOM), Jiangsu Key Laboratory for Carbon-
Based Functional Materials & Devices, Soochow University,
Suzhou 215123, PR China
Di Yang − Institute of Functional Nano & Soft Materials
(FUNSOM), Jiangsu Key Laboratory for Carbon-Based
Functional Materials & Devices, Soochow University, Suzhou
215123, PR China
Zeyuan Tang − Institute of Functional Nano & Soft Materials
(FUNSOM), Jiangsu Key Laboratory for Carbon-Based
Functional Materials & Devices, Soochow University, Suzhou
215123, PR China
Muhan Cao − Institute of Functional Nano & Soft Materials
(FUNSOM), Jiangsu Key Laboratory for Carbon-Based
Functional Materials & Devices, Soochow University, Suzhou
215123, PR China; orcid.org/0000-0002-7988-7219
Huicheng Hu − Institute of Functional Nano & Soft Materials
(FUNSOM), Jiangsu Key Laboratory for Carbon-Based
Functional Materials & Devices, Soochow University, Suzhou
215123, PR China
Linzhong Wu − Institute of Functional Nano & Soft Materials
(FUNSOM), Jiangsu Key Laboratory for Carbon-Based
Functional Materials & Devices, Soochow University, Suzhou
215123, PR China
Lijia Liu − Institute of Functional Nano & Soft Materials
(FUNSOM), Jiangsu Key Laboratory for Carbon-Based
Functional Materials & Devices, Soochow University, Suzhou
215123, PR China
John McLeod − Institute of Functional Nano & Soft Materials
(FUNSOM), Jiangsu Key Laboratory for Carbon-Based
Functional Materials & Devices, Soochow University, Suzhou
215123, PR China; orcid.org/0000-0002-2040-7540
Youyong Li − Institute of Functional Nano & Soft Materials
(FUNSOM), Jiangsu Key Laboratory for Carbon-Based
4. CONCLUSIONS
In summary, we have demonstrated the stabilization of oxygen
vacancies in defective ZrO2 with the assistance of Ga2O3.
Ga2O3 suppressed active site poisoning and coke formation
and enabled efficient dehydrogenation. Our most efficient
(Zr0.26Ga1/SIRAL10) catalyst showed superior catalytic
activity of conversion of alkanes to olefins compared with
the industrial Pt−Sn catalyst. DFT calculations of the Zr−Ga
system provided insight into its improved PDH performance,
emphasizing the role of oxygen vacancies (and low coordinated
Zr atoms) in the ZrO2 part of the ZrO2−Ga2O3 interface. The
optimum dehydrogenation activity is achieved by the interplay
between the number of low-coordinated Zr sites at the ZrO2−
Ga2O3 interface and the availability of Ga2O3 to provide a low
energy channel for removal by diffusion of the dissociated H
atoms which otherwise block the reaction sites. The optimized
Zr0.26Ga1/SIRAL10 displayed a higher propylene STY by 21%
compared to the industrial Pt−Sn catalyst. The promising
catalytic performance for the dehydrogenation of light alkanes
over the cheap ZrO2−Ga2O3 catalyst is expected to have a
significant impact on the chemical industry.
ASSOCIATED CONTENT
■
sı
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acscatal.1c02621.
Experimental details, material characterization data, and
computational details (PDF)
10167
https://doi.org/10.1021/acscatal.1c02621
ACS Catal. 2021, 11, 10159−10169

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
(10) Aly, M.; Fornero, E. L.; Leon-Garzon, A. R.; Galvita, V. V.;
Saeys, M. Effect of boron promotion on coke formation during
propane dehydrogenation over Pt/gamma-Al2O3 catalysts. ACS Catal.
2020, 10, 5208−5216.
(11) Nakaya, Y.; Hirayama, J.; Yamazoe, S.; Shimizu, K.-i.;
Furukawa, S. Single-atom Pt in intermetallics as an ultrastable and
selective catalyst for propane dehydrogenation. Nat. Commun. 2020,
11, 2838.
Functional Materials & Devices, Soochow University, Suzhou
215123, PR China; orcid.org/0000-0002-5248-2756
Tsun-Kong Sham − Department of Chemistry, University of
Western Ontario, London, Ontario N6A 5B7, Canada;
orcid.org/0000-0003-1928-6697
Complete contact information is available at:
https://pubs.acs.org/10.1021/acscatal.1c02621
(12) Searles, K.; Chan, K. W.; Mendes Burak, J. A.; Zemlyanov, D.;
Safonova, O.; Copéret, C. Highly productive propane dehydrogen-
ation catalyst using silica-supported Ga-Pt nanoparticles generated
from single-sites. J. Am. Chem. Soc. 2018, 140, 11674−11679.
(13) Im, J.; Choi, M. Physicochemical stabilization of Pt against
sintering for a dehydrogenation catalyst with high activity, selectivity,
and durability. ACS Catal. 2016, 6, 2819−2826.
Author Contributions
The manuscript was written through contributions of all
authors. Y.X., X.W., D.Y., and Z.T. contributed equally.
Notes
The authors declare no competing financial interest.
(14) Kaylor, N.; Davis, R. J. Propane dehydrogenation over
supported Pt-Sn nanoparticles. J. Catal. 2018, 367, 181−193.
(15) Zhu, H.; Anjum, D. H.; Wang, Q.; Abou-Hamad, E.; Emsley,
L.; Dong, H.; Laveille, P.; Li, L.; Samal, A. K.; Basset, J. M. Sn surface-
enriched Pt-Sn bimetallic nanoparticles as a selective and stable
catalyst for propane dehydrogenation. J. Catal. 2014, 320, 52−62.
(16) Iglesias-Juez, A.; Beale, A. M.; Maaijen, K.; Weng, T. C.;
Glatzel, P.; Weckhuysen, B. M. A combined in situ time-resolved UV-
Vis, Raman and high-energy resolution X-ray absorption spectroscopy
study on the deactivation behavior of Pt and Pt-Sn propane
dehydrogenation catalysts under industrial reaction conditions. J.
Catal. 2010, 276, 268−279.
(17) Shi, L.; Deng, G. M.; Li, W. C.; Miao, S.; Wang, Q. N.; Zhang,
W. P.; Lu, A. H. Al2O3 nanosheets rich in pentacoordinate Al3+ ions
stabilize Pt-Sn clusters for propane dehydrogenation. Angew. Chem.
Int. Ed. 2015, 54, 13997−13998.
(18) Lobera, M. P.; Téllez, C.; Herguido, J.; Menéndez, M.
Transient kinetic modelling of propane dehydrogenation over a Pt-Sn-
K/Al2O3 catalyst. Appl. Catal., A 2008, 349, 156−164.
(19) Otroshchenko, T.; Sokolov, S.; Stoyanova, M.; Kondratenko, V.
A.; Rodemerck, U.; Linke, D.; Doz, P.; Kondratenko, E. V. ZrO2-
based alternatives to conventional propane dehydrogenation catalysts:
active sites, design, and performance. Angew. Chem. Int. Ed. 2015, 54,
15880−15883.
(20) Han, S.; Zhao, Y.; Otroshchenko, T.; Zhang, Y.; Zhao, D.;
Lund, H.; Vuong, T. H.; Rabeah, J.; Bentrup, U.; Kondratenko, V. A.;
Rodemerck, U.; Linke, D.; Gao, M.; Jiao, H.; Jiang, G.; Kondratenko,
E. V. Unraveling the origins of the synergy effect between ZrO2 and
CrOx in supported CrZrOx for propene formation in nonoxidative
propane dehydrogenation. ACS Catal. 2020, 10, 1575−1590.
(21) Otroshchenko, T.; Kondratenko, V. A.; Rodemerck, U.; Linke,
D.; Kondratenko, E. V. ZrO2-based unconventional catalysts for non-
oxidative propane dehydrogenation: factors determining catalytic
activity. J. Catal. 2017, 348, 282−290.
(22) Otroshchenko, T. P.; Rodemerck, U.; Linke, D.; Kondratenko,
E. V. Synergy effect between Zr and Cr active sites in binary CrZrOx
or supported CrOx/LaZrOx: consequences for catalyst activity,
selectivity and durability in non-oxidative propane dehydrogenation.
J. Catal. 2017, 356, 197−205.
(23) Otroshchenko, T.; Radnik, J.; Schneider, M.; Rodemerck, U.;
Linke, D.; Kondratenko, E. V. Bulk binary ZrO2-based oxides as
highly active alternative-type catalysts for non-oxidative isobutane
dehydrogenation. Chem. Commun. 2016, 52, 8164−8167.
(24) Zhang, Y.; Zhao, Y.; Otroshchenko, T.; Lund, H.; Pohl, M.-M.;
Rodemerck, U.; Linke, D.; Jiao, H.; Jiang, G.; Kondratenko, E. V.
Control of coordinatively unsaturated Zr sites in ZrO2 for efficient
C−H bond activation. Nat. Commun. 2018, 9, 3794.
ACKNOWLEDGMENTS
■
This work is supported by the National Natural Science
Foundation of China (21673150, 21703146, 51802206, and
51922073), the Natural Science Foundation of Jiangsu
Province (BK20180097 an d BK20180846), and the
Guangdong Provincial Natural Science Fund for Distinguished
Young Scholars (2021B1515020081). We acknowledge the
financial support from the 111 Project, Collaborative
Innovation Center of Suzhou Nano Science and Technology
(NANO-CIC), and the Priority Academic Program Develop-
ment of Jiangsu Higher Education Institutions (PAPD). We
also acknowledge Sasol Co. for providing the catalyst support
and Dr. Jinghua Guo at the Lawrence Berkeley National
Laboratory for helpful discussion.
REFERENCES
■
(1) Torres Galvis, H. M.; Bitter, J. H.; Khare, C. B.; Ruitenbeek, M.;
Dugulan, A. I.; de Jong, K. P. Supported iron nanoparticles as catalysts
for sustainable production of lower olefins. Science 2012, 335, 835−
838.
(2) Jiao, F.; Li, J.; Pan, X.; Xiao, J.; Li, H.; Ma, H.; Wei, M.; Pan, Y.;
Zhou, Z.; Li, M.; Miao, S.; Li, J.; Zhu, Y.; Xiao, D.; He, T.; Yang, J.;
Qi, F.; Fu, Q.; Bao, X. Selective conversion of syngas to light olefins.
Science 2016, 351, 1065−1068.
(3) Zhong, L.; Yu, F.; An, Y.; Zhao, Y.; Sun, Y.; Li, Z.; Lin, T.; Lin,
Y.; Qi, X.; Dai, Y.; Gu, L.; Hu, J.; Jin, S.; Shen, Q.; Wang, H. Cobalt
carbide nanoprisms for direct production of lower olefins from syngas.
Nature 2016, 538, 84−87.
(4) Sattler, J. J. H. B.; Ruiz-Martinez, J.; Santillan-Jimenez, E.;
Weckhuysen, B. M. Catalytic dehydrogenation of light alkanes on
metals and metal oxides. Chem. Rev. 2014, 114, 10613−10653.
(5) McFarland, E. Unconventional chemistry for unconventional
natural gas. Science 2012, 338, 340−342.
(6) Guo, X.; Fang, G.; Li, G.; Ma, H.; Fan, H.; Yu, L.; Ma, C.; Wu,
X.; Deng, D.; Wei, M.; Tan, D.; Si, R.; Zhang, S.; Li, J.; Sun, L.; Tang,
Z.; Pan, X.; Bao, X. Direct, nonoxidative conversion of methane to
ethylene, aromatics, and hydrogen. Science 2014, 344, 616−619.
(7) Chowdhury, A. D.; Weding, N.; Julis, J.; Franke, R.; Jackstell, R.;
Beller, M. Towards a practical development of light-driven acceptor
less alkane dehydrogenation. Angew. Chem. Int. Ed. 2014, 53, 6477−
6481.
(8) Sattler, J. J. H. B.; Gonzalez-Jimenez, I. D.; Luo, L.; Stears, B. A.;
Malek, A.; Barton, D. G.; Kilos, B. A.; Kaminsky, M. P.; Verhoeven, T.
W. G. M.; Koers, E. J.; Baldus, M.; Weckhuysen, B. M. Platinum-
promoted Ga/Al2O3 as highly active, selective, and stable catalyst for
the dehydrogenation of propane. Angew. Chem. Int. Ed. 2014, 53,
9251−9256.
(9) Sun, P.; Siddiqi, G.; Vining, W. C.; Chi, M.; Bell, A. T. Novel Pt/
Mg(In)(Al)O catalysts for ethane and propane dehydrogenation. J.
Catal. 2011, 282, 165−174.
(25) Ankudinov, A. L.; Ravel, B.; Rehr, J.; Conradson, S. Real-Space
Multiple-Scattering Calculation and Interpretation of X-Ray-Absorp-
tion Near-Edge Structure. Phys. Rev. B: Condens. Matter Mater. Phys.
1998, 58, 7565−7576.
(26) Newville, M. Interactive XAFS Analysis and FEFF Fitting. J.
Synchrotron Radiat. 2001, 8, 322−324.
10168
https://doi.org/10.1021/acscatal.1c02621
ACS Catal. 2021, 11, 10159−10169

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
(27) Ravel, B.; Newville, M. ATHENA,ARTEMIS,HEPHAESTUS:
Data Analysis for X-Ray Absorption Spectroscopy Using IFEFFIT. J.
Synchrotron Radiat. 2005, 12, 537−541.
(28) Kresse, G.; Furthmuller, J. Efficient Iterative Schemes for Ab
̈
Initio Total-Energy Calculations Using a Plane-Wave Basis Set. Phys.
Rev. B: Condens. Matter Mater. Phys. 1996, 54, 11169−11186.
(29) Kresse, G.; Furthmuller, J. Efficiency of Ab Initio Total Energy
̈
Calculations for Metalsand Semiconductors Using a Plane-Wave Basis
Set. Comput. Mater. Sci. 1996, 6, 15−50.
(30) Blöchl, P. E. Projector Augmented-Wave Method. Phys. Rev. B:
Condens. Matter Mater. Phys. 1994, 50, 17953−17979.
(31) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized Gradient
Approximation Made Simple. Phys. Rev. Lett. 1996, 77, 3865−3868.
(32) Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H. A Consistent and
Accurate Ab Initio Parametrization of Density Functional Dispersion
Correction (DFT-D) for the 94 Elements H-Pu. J. Chem. Phys. 2010,
132, 154104.
(33) Henkelman, G.; Uberuaga, B. P.; Jónsson, H. A Climbing
Image Nudged Elastic Band Method for Finding Saddle Points And
Minimum Energy Paths. J. Chem. Phys. 2000, 113, 9901−9904.
(34) Henkelman, G.; Jónsson, H. A Dimer Method for Finding
Saddle Points on High Dimensional Potential Surfaces Using only
First Derivatives. J. Chem. Phys. 1999, 111, 7010−7022.
(35) Larsen, A. H.; Jørgen Mortensen, J.; Blomqvist, J.; Castelli, I. E.;
Christensen, R.; Dułak, M.; Friis, J.; Groves, M. N.; Hammer, B.;
Hargus, C.; Hermes, E. D.; Jennings, P. C.; Bjerre Jensen, P.;
Kermode, J.; Kitchin, J. R.; Leonhard Kolsbjerg, E.; Kubal, J.;
Kaasbjerg, K.; Lysgaard, S.; Bergmann Maronsson, J.; Maxson, T.;
Olsen, T.; Pastewka, L.; Peterson, A.; Rostgaard, C.; Schiøtz, J.;
Schutt, O.; Strange, M.; Thygesen, K. S.; Vegge, T.; Vilhelmsen, L.;
̈
Walter, M.; Zeng, Z.; Jacobsen, K. W. The Atomic Simulation
Environment−A Python Library for Working with Atom. J. Phys.:
Condens. Matter 2017, 29, 273002.
(36) Schäfer, R.; Noack, M.; Kölsch, P.; Stöhr, M.; Caro, J.
Comparison of different catalysts in the membrane-supported
dehydrogenation of propane. Catal. Today 2003, 82, 15−23.
(37) Aguilar-Ríos, G.; Salas, P.; Valenzuela, M. A.; Armendáriz, H.;
Wang, J. A.; Salmones, J. Propane dehydrogenation activity of Pt and
Pt-Sn catalysts supported on magnesium aluminate: Influence of
steam and hydrogen. Catal. Lett. 1999, 60, 21−25.
(38) Larsson, M.; Hultén, M.; Blekkan, E. A.; Andersson, B. The
effect of reaction conditions and time on stream on the coke Formed
during propane dehydrogenation. J. Catal. 1996, 164, 44−53.
(39) Zhang, R.; He, D. Effect of alcohol solvents treated ZrO(OH)2
hydrogel on properties of ZrO2 and its catalytic performance in
isosynthesis. J. Nat. Gas Chem. 2012, 21, 1−6.
(40) Xu, Y.; Chen, J.; Yuan, X.; Zhang, Y.; Yu, J.; Liu, H.; Cao, M.;
Fan, X.; Lin, H.; Zhang, Q. Sintering-resistant Pt on Ga2O3 rods for
propane dehydrogenation: The morphology matters. Ind. Eng. Chem.
Res. 2018, 57, 13087−13093.
(41) Tan, S.; Hu, B.; Kim, W.-G.; Pang, S. H.; Moore, J. S.; Liu, Y.;
Dixit, R. S.; Pendergast, J. G.; Sholl, D. S.; Nair, S.; Jones, C. W.
Propane dehydrogenation over alumina-supported iron/phosphorus
catalysts: structural evolution of iron species leading to high activity
and propylene selectivity. ACS Catal. 2016, 6, 5673−5683.
10169
https://doi.org/10.1021/acscatal.1c02621
ACS Catal. 2021, 11, 10159−10169