Effect of Oxide Support on Catalytic Performance of FeNi-based Catalysts for CO2-assisted Oxidative Dehydrogenation of Ethane

Article abstract of DOI:10.1002/cctc.201901585

CO₂-assisted oxidative dehydrogenation of ethane over Fe₃Ni₁ catalysts supported on SiO₂, ZrO₂, and CeO₂, as well as CeO₂-supported Fe_3xNi_x with different metal loadings (x=0.25, 0.5, 1, 2, 4), were investigated to illustrate the effects of the oxide support and the metal loading amount in tuning the catalytic activity and selectivity. The oxidation state and the dispersion of the Ni and Fe species on the supported catalysts were determined by using in-situ X-ray diffraction (XRD) and X-ray absorption fine structure (XAFS) measurements. Among these FeNi-based catalysts, Fe₃Ni₁/CeO₂ exhibited the best catalytic activity and maximum ethylene yield for the oxidative dehydrogenation of ethane. Excessive loading amount of active metals changed the reaction selectivity from oxidative dehydrogenation to dry reforming. The high activity and selectivity for CO₂-assisted oxidative dehydrogenation of ethane over Fe₃Ni₁/CeO₂ are likely due to the formation of Ni-FeO_x/CeO_x interfacial sites, while the presence of Ni-ZrO₂ interfacial sites on Fe₃Ni₁/ZrO₂ and NiFe-CeO_x interfacial sites on Fe₁₂Ni₄/CeO₂ likely tune the reaction selectivity toward dry reforming of ethane.

Full text of DOI:10.1002/cctc.201901585

Accepted Article
Title: Effect of oxide support on catalytic performance of FeNi-based
catalysts for CO2-assisted oxidative dehydrogenation of ethane
Authors: Binhang Yan, Siyu Yao, and Jingguang G. Chen
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.201901585
Link to VoR: http://dx.doi.org/10.1002/cctc.201901585
A Journal of
www.chemcatchem.org

10.1002/cctc.201901585
ChemCatChem
Effect of oxide support on catalytic performance of FeNi-based
catalysts for CO2-assisted oxidative dehydrogenation of ethane
Binhang Yan, [a, b] Siyu Yao, [b] and Jingguang G. Chen *[b, c]
[a] Prof. Dr. Binhang Yan
Department of Chemical Engineering
Tsinghua University
30 Shuangqing Rd. Haidian District Beijing, 100084 (China)
[b] Prof. Dr. Binhang Yan, Prof. Dr. Siyu Yao, Prof. Dr. Jingguang G. Chen
Chemistry Division
Brookhaven National Laboratory
55 Lewis Rd. Upton NY, 11973 (USA)
[c] Prof. Dr. Jingguang G. Chen
Department of Chemical Engineering
Columbia University
500 W. 120th St. New York NY, 10027 (USA)
Telephone: (212) 854-6166, Fax: (212) 854-3054
Email: jgchen@columbia.edu
1
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901585
ChemCatChem
ABSTRACT
CO2-assisted oxidative dehydrogenation of ethane over Fe3Ni1 catalysts supported
on SiO2, ZrO2, and CeO2, as well as CeO2-supported Fe3xNix with different metal
loadings (x = 0.25, 0.5, 1, 2, 4), were investigated to illustrate the effects of the oxide
support and the metal loading amount in tuning the catalytic activity and selectivity.
The oxidation state and the dispersion of the Ni and Fe species on the supported
catalysts were determined by using in-situ X-ray diffraction (XRD) and X-ray
absorption fine structure (XAFS) measurements. Among these FeNi-based catalysts,
Fe3Ni1/CeO2 exhibited the best catalytic activity and maximum ethylene yield for the
oxidative dehydrogenation of ethane. Excessive loading amount of active metals
changed the reaction selectivity from oxidative dehydrogenation to dry reforming. The
high activity and selectivity for CO2-assisted oxidative dehydrogenation of ethane over
Fe3Ni1/CeO2 are likely due to the formation of Ni-FeOx/CeOx interfacial sites, while
the presence of Ni-ZrO2 interfacial sites on Fe3Ni1/ZrO2 and NiFe-CeOx interfacial sites
on Fe12Ni4/CeO2 likely tune the reaction selectivity toward dry reforming of ethane.
Keywords: Heterogeneous catalysis; Support effect; Oxidative dehydrogenation; Dry
reforming; Selectivity
2
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901585
ChemCatChem
Introduction
CO2-assisted oxidative dehydrogenation of ethane (C2H6 + CO2  C2H4 + CO +
H2O) selectively cleaves the C-H bonds while preserving the C-C bond of ethane to
produce ethylene. The reaction has attracted significant research interest [1-6] as it
offers a potential way to simultaneously upgrade the underutilized ethane in the shale
gas and mitigate the rising global CO2 emission. However, the competing reaction, i.e.,
CO2 reforming of ethane (C2H6 + 2CO2  4CO + 3H2), breaks all the C-H and C-C
bonds of ethane to produce synthesis gas [7-9]. The dry reforming reaction is
thermodynamically more favorable under the reaction conditions. Therefore, research
efforts should focus on developing new catalysts with high oxidative dehydrogenation
selectivity to produce more desirable ethylene. Many attempts have been made to
promote the oxidative dehydrogenation selectivity by using Ga-based oxide catalysts
[10-11], Cr-based oxide catalysts [4, 6, 12-16], Co-based oxide catalysts [17-18], metal
carbide catalysts [19-20], and other catalysts [21-23]. Although theses catalysts exhibit
high initial dehydrogenation activity, the stability is poor, likely due to the coke
formation. The improvement of the selectivity and stability of CO2-assisted oxidative
dehydrogenation of ethane requires a bifunctional catalyst that can dissociate CO2 into
CO and a reactive surface oxygen species [5, 24] and activate the adsorbed ethane on
metal centers or metal-support interfaces.
The interaction between the active metal and an oxide support presents an
opportunity to design new catalysts with enhanced oxidative dehydrogenation
selectivity. Reducible supports (e.g., CeO2) are commonly used as they can promote
3
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901585
ChemCatChem
the activation of CO2 via direct C=O bond scission [2, 24]. The released reactive
surface/lattice O atoms can readily remove the formed coke deposits and allow the
overall reaction to proceed efficiently. In previous studies Fe3Ni1/CeO2 was identified
as a promising catalyst with high selectivity and excellent stability for CO2-assisted
oxidative dehydrogenation of ethane to produce ethylene [2, 25]. The Ni-FeOx
interfaces and Ni-CeOx interfaces were demonstrated as the most likely active sites for
the oxidative dehydrogenation pathway and the dry reforming pathway, respectively.
The two types of active sites on CeO2-supported FeNi catalysts could be further tuned
to rationally manipulate the catalytic selectivity by changing the Fe/Ni ratio.
Many studies have reported that catalysts using different supports show distinctive
catalytic properties [26-28]. However, the support effects of the FeNi-based catalysts
on the catalytic performances of CO2-assisted oxidative dehydrogenation of ethane,
especially on the properties of the active metal species, have not been studied or fully
understood. SiO2 provides an irreducible and relatively inert substrate [28]. ZrO2 is
debatably considered to be reducible or at least partially reducible, but it is certain that
the support exhibits surface defects or kinks that can strongly influence catalyst activity
and selectivity [29]. CeO2 is classified as a readily reducible oxide with ability to store
and release oxygen, which not only facilitates CO2 activation, but also provides lattice
oxygen for coke suppression [30]. There is a clear need to expand the knowledge
regarding the effect of oxide supports for the reaction of C2H6 with CO2 to develop
highly active and selective catalysts for the oxidative dehydrogenation pathway. Hence
the objective of this work is to study the influences of different supports (SiO2, ZrO2,
4
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901585
ChemCatChem
and CeO2) on the catalytic performances of Fe3Ni1-based catalysts, and to further
understand the optimum loading amount of Fe3xNix/CeO2 (x = 0.25, 0.5, 1, 2, 4)
catalysts for CO2-assisted oxidative dehydrogenation of ethane using a combination of
flow reactors and in-situ XRD and XAFS characterization.
Results and Discussion
Effect of support
To explore the support effect, supported-Fe3Ni1 (same Fe/Ni ratios and weight
loadings) catalysts with different supports (SiO2, ZrO2, and CeO2) were tested at 873 K
for over 13 h (10 mL/min C2H6 + 10 mL/min CO2 + 20 mL/min Ar). The conversion
of C2H6 and CO2, the yield of C2H4 and CO, and the selectivity toward C2H4 and CO
over the three supported-Fe3Ni1 catalysts, are shown in Figure 1 as a function of time
on stream. The steady-state conversion, yield, selectivity, turnover frequency (TOF)
and space-time yield (STY) are listed in Table 1. As compared in Figure 1 and Table 1,
the oxide supports have remarked influence on the catalytic activity of supported-
Fe3Ni1 catalysts.
Fe3Ni1/SiO2 shows poor activity, either for the oxidative dehydrogenation reaction
or for the dry reforming reaction. Severe deactivation on Fe3Ni1/ZrO2 is observed: the
C2H6 (CO2) conversion gradually changes from 10% (27%) at the first 0.5 h to 2% (4%)
after 13 h operation. The C2H4 selectivity is also greatly varied (from 3% to 67%) during
the reaction, indicating that the selectivity evolution on Fe3Ni1/ZrO2 might be due to
the transformation of active sites. Compared to Fe3Ni1/SiO2 and Fe3Ni1/ZrO2,
5
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901585
ChemCatChem
Fe3Ni1/CeO2 shows an enhanced and stable C2H4 yield/selectivity, with an over 80%
C2H4 selectivity. The C2H6 conversion, CO2 conversion, and C2H4 yield over
Fe3Ni1/CeO2 change slightly during the reaction, suggesting that the active sites that
favor oxidative dehydrogenation on the catalyst are stable. The catalytic performance
over Fe3Ni1/SiO2, Fe3Ni1/ZrO2, and Fe3Ni1/CeO2 at different temperatures varying from
873 K to 823 K after 13 h operation is compared in Figure 2. As is seen from Figure 2,
the conversion of C2H6 and CO2, and the yield of C2H4 and CO over the three supported-
Fe3Ni1 catalysts gradually increase with the increase of temperature, while the C2H4
selectivity over Fe3Ni1/CeO2 also changes slightly when the temperature changes.
Furthermore, the C2H4 selectivity over Fe3Ni1/CeO2 still remains comparatively steady
when the gas hourly space velocity (GHSV) is changed (by varying the total flow rate)
at low C2H6 conversions, as shown in Figure S1.
6
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901585
ChemCatChem
Figure 1. Conversion of (a) C2H6 and (b) CO2, yield of (c) C2H4 and (d) CO, and
selectivity toward (e) CH4 and (f) CO over Fe3Ni1/SiO2, Fe3Ni1/ZrO2, and Fe3Ni1/CeO2
plotted versus time on stream for reaction of C2H6 and CO2 (10 mL/min C2H6 + 10
mL/min CO2 + 20 mL/min Ar, 100 mg catalyst) at 873 K.
7
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901585
ChemCatChem
Figure 2. Conversion of (a) C2H6 and (b) CO2, yield of (c) C2H4 and (d) CO, and
selectivity toward (e) CH4 and (f) CO over Fe3Ni1/SiO2, Fe3Ni1/ZrO2, and Fe3Ni1/CeO2
for reaction of C2H6 and CO2 (10 mL/min C2H6 + 10 mL/min CO2 + 20 mL/min Ar,
100 mg catalyst) at different temperatures with 10 K temperature increment.
8
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901585
ChemCatChem
Table 1 Summary of flow reactor results for CO2 + C2H6 reaction (10 ml/min CO2 + 10 ml/min C2H6 + 20 ml/min Ar) at 873 K. Values of
conversion, yield, selectivity, TOF and STY calculated by averaging data points between 11-13 h on stream
Conversion, % Yield, % Selectivity, % TOF1, time/site/min TOF2, mol/molNi/min
STY, mol/gcat/min
Catalyst
C2H6
0.4
2.6
3.5
2.2
3.1
3.4
3.5
CO2 C2H4 CO C2H4
CO
-
C2H6
-
CO2
-
C2H6
0.2
1.3
1.7
4.2
3.0
0.8
0.4
CO2
0.5
2.7
3.3
9.0
5.9
2.1
1.6
C2H4
13.9
63.6
116.8
71.8
101.0
92.6
45.3
CO
Fe3Ni1/SiO2
Fe3Ni1/ZrO2
0.9
5.6
6.9
4.7
6.2
8.7
13.1
0.3 0.1
-
57.9
1.6 1.0 61.7
2.9 0.6 82.0
1.7 0.5 76.7
2.5 0.6 79.2
2.3 1.1 67.4
1.1 2.4 31.9
37.9
17.8
23.0
20.5
32.3
67.7
3.5
2.7
1.9
2.6
2.6
2.1
7.5
5.2
4.1
5.1
6.7
8.1
310.3
346.1
246.3
316.2
465.2
762.8
Fe3Ni1/CeO2
Fe0.75Ni0.25/CeO2
Fe1.5Ni0.5/CeO2
Fe6Ni2/CeO2
Fe12Ni4/CeO2
9
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901585
ChemCatChem
Effect of metal loading amount
Time-on-stream catalytic activities of the CeO2-supported Fe3xNix (x = 0.25, 0.5,
1, 2, and 4) catalysts with different metal loadings are shown in Figure 3. The steady-
state conversion, yield, selectivity, turnover frequency (TOF) and space-time yield
(STY) are also listed in Table 1. When the loading amounts of the active metals (i.e.,
Fe and Ni) are low, the C2H6 conversion monotonically increases while the C2H4
selectivity remains stable (80%) with the increase in the loading amount. However, if
the amount of loaded Fe (Ni) exceeds 1.4 (0.5) wt%, the conversion of C2H6 remains
stable but the dominant pathway transforms from favoring the oxidative
dehydrogenation reaction (e.g., Fe3Ni1/CeO2) to favoring the dry reforming reaction
(e.g., Fe12Ni4/CeO2), greatly lowering the C2H4 selectivity (from 82% to 32%). The
maximum C2H4 yield is observed on Fe3Ni1/CeO2 with 1.4 wt% Fe and 0.5 wt% Ni.
The conversion, yield, and selectivity over the CeO2-supported Fe3xNix (x = 0.25, 0.5,
1, 2, and 4) catalysts at different temperatures varying from 873 K to 823 K after 13 h
operation are compared in Figure 4. For Fe3Ni1/CeO2 and Fe12Ni4/CeO2, the conversion
of C2H6 and CO2, and the yield of C2H4 and CO gradually increase with increasing
temperature while the selectivity of C2H4 and CO remains nearly constant.
10
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901585
ChemCatChem
Figure 3. Conversion of (a) C2H6 and (b) CO2, yield of (c) C2H4 and (d) CO, and
selectivity toward (e) CH4 and (f) CO over CeO2-supported Fe3xNix (x = 0.25, 0.5, 1, 2,
and 4)catalysts plotted versus time on stream for reaction of C2H6 and CO2 (10 mL/min
C2H6 + 10 mL/min CO2 + 20 mL/min Ar, 100 mg catalyst) at 873 K.
11
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901585
ChemCatChem
Figure 4. Conversion of (a) C2H6 and (b) CO2, yield of (c) C2H4 and (d) CO, and
selectivity toward (e) CH4 and (f) CO over CeO2-supported Fe3xNix (x = 0.25, 0.5, 1, 2,
and 4)catalysts for reaction of C2H6 and CO2 (10 mL/min C2H6 + 10 mL/min CO2 + 20
mL/min Ar, 100 mg catalyst) at different temperatures with 10 K temperature increment.
12
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901585
ChemCatChem
Ex-situ and in-situ X-ray diffraction (XRD) characterization
The structures of the fresh and spent Fe12Ni4/CeO2, Fe6Ni2/CeO2, Fe3Ni1/CeO2, and
Fe3Ni1/ZrO2 catalysts were investigated using ex-situ XRD. As shown in Figure 5, the
XRD patterns exhibit the typical diffraction peaks of CeO2 with a cubic fluorite
structure and ZrO2 with a monoclinic crystalline structure. The crystalline sizes of CeO2
and ZrO2 are estimated about 33 nm and 14 nm, respectively, based on the XRD peak
width. XRD peaks of NiO are not observed in the fresh samples, indicating that NiO is
well dispersed on both support with relatively small crystallite sizes. When the loading
amount of Fe exceeds 2.9 wt%, crystalline Fe3O4 or Fe2O3 are observed in the fresh
CeO2-supported Fe3xNix catalysts with weak (311), (400) and (440) reflections at 5.45°,
6.58° and 9.31°, respectively. After over 13 h reaction at 873 K, the diffraction intensity
of crystalline Fe3O4 or Fe2O3 becomes stronger and the corresponding peak width gets
narrower, indicating that the crystal size is increased. Weak peaks at 4.65˚ and 5.70˚,
which are attributed to the (220) and (222) reflections of crystalline Fe3O4 or Fe2O3, are
also observed. New diffraction peaks at 6.75˚ and 7.79˚ appear in the XRD patterns for
the spent CeO2-support FeNi catalysts, implying that metallic Ni (or Ni-rich NiFe alloy)
is formed under reaction conditions.
13
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901585
ChemCatChem
Figure 5. Ex-situ XRD profiles of the fresh and spent Fe12Ni4/CeO2, Fe6Ni2/CeO2,
Fe3Ni1/CeO2, and Fe3Ni1/ZrO2 catalysts in the 2θ range of 3.8˚ to 9.8˚.
14
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901585
ChemCatChem
The structural evolution of Fe12Ni4/CeO2, Fe6Ni2/CeO2, Fe3Ni1/CeO2, and
Fe3Ni1/ZrO2 under reaction conditions was examined using in-situ XRD measurements.
Figure S2 and Figure 6 show the variation of diffraction peaks in different 2θ ranges
with time for the four catalysts. For CeO2-supported FeNi catalysts under reaction
conditions, the diffraction peaks at 5.42°, 5.65°, 6.54°, 8.02°, and 9.25° are assigned to
the (311), (222), (400), (422) and (440) reflections of crystalline Fe3O4 or Fe2O3,
respectively. According to a previous work [2], the (111) and (200) reflections of
metallic Ni with the FCC (face-centered cubic) structure are located at 6.74° and 7.77°,
based on which the calculated lattice constant was 3.558 Å, slightly larger than the
value of 3.524 Å for bulk Ni due to thermal expansion. Compared to metallic Ni, the
diffraction peaks of the reduced metal on CeO2-supported Fe3xNix catalysts under
reaction conditions are slightly shifted to lower 2θ values with a calculated lattice
constant of 3.570 Å. The larger cell parameter suggests that Ni particles are alloyed
with Fe (FCC structure), thus the peaks at 6.71° and 7.75° are assigned to the (111) and
(200) reflections of Ni-rich NiFe alloy. For CeO2-supported Fe3xNix catalysts, the
diffraction peak centered at 7.75° under reaction conditions slightly shifts to 7.77°
under CO2 treatment conditions as shown in Figures S3-S5 with the calculated lattice
constant of 3.561 Å (more close to metallic Ni 3.558 Å), indicating that the Fe species
in Ni-rich NiFe alloy formed under reaction conditions can be re-oxidized in a CO2/He
flow to form an amorphous FeOx thin layer [31]. The peak intensities of crystalline
Fe3O4 or Fe2O3 and metallic NiFe alloy increase when increasing the metal loading
amount. Based on the evolution of the peaks centered at 6.71°, it can be concluded that
15
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901585
ChemCatChem
the growth of metallic NiFe crystal particles starts at a temperature of 833 K in the
presence of reactants (2.5 mL/min C2H6 + 2.5 mL/min CO2 + 5.0 mL/min He). The
intensity and position of this peak do not change after 10-minute reaction, indicating
that the formed metallic NiFe alloy particles are relatively stable under reaction
conditions. The diffraction peaks of crystalline Fe3O4 or Fe2O3 and metallic Ni (or Ni-
rich NiFe alloy) are barely observed over Fe3Ni1/ZrO2, suggesting that both Fe and Ni
species are highly dispersed on ZrO2 under reaction conditions.
16
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901585
ChemCatChem
17
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901585
ChemCatChem
18
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901585
ChemCatChem
Figure 6. Evolution of the in-situ XRD profiles under various conditions: (a) Fe12Ni4/CeO2, (b) Fe6Ni2/CeO2, (c) Fe3Ni1/CeO2, and (d) Fe3Ni1/ZrO2
in the 2θ range of (1) 5.2˚ to 7.0˚, (2) 7.6˚ to 8.2˚, and (3) 9.0˚ to 9.6˚. Magenta line, temperature ramping stage: from 300 K to 873 K, 2.5 mL/min
C2H6 + 2.5 mL/min CO2 + 5.0 mL/min He, 10 mg catalyst; blue line, reaction conditions: 873 K, 2.5 mL/min C2H6 + 2.5 mL/min CO2 + 5.0
mL/min He.
19
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901585
ChemCatChem
In-situ X-ray absorption fine structure (XAFS) measurements
In order to further determine the oxidation state of Ni and Fe on Fe3Ni1/ZrO2,
Fe3Ni1/CeO2, and Fe12Ni4/CeO2 under reaction conditions, in-situ X-ray absorption
near edge structure (XANES) spectra were collected, as shown in Figure 7. Based on
the Ni K-edge XANES spectra, the Ni species in all three fresh catalysts show similar
near edge features as the standard NiO. The Fe-K edge XANES spectra suggest that the
Fe species in all fresh samples exhibit near edge features between Fe3O4 and Fe2O3. For
all three catalysts under reaction conditions, the oxidation state of Ni is metallic and the
Fe species are also reduced and show near edge features between FeO and Fe3O4. The
detailed EXAFS fittings of the Ni and Fe K-edge spectra of Fe3Ni1/ZrO2, Fe3Ni1/CeO2,
and Fe12Ni4/CeO2 under reaction conditions are summarized in Table 2, Table 3,
Figures S6 and S7. The EXAFS fitting results reveal that the coordination numbers
(CNs) of Ni-Ni(Fe) on the three catalysts under reaction conditions are 9.8, 9.8 and 11.3,
respectively, indicating that the Ni-rich NiFe alloy resides in the form of particles with
diameters of 2.2-4.0 nm (based on hemisphere FCC models). According to the EXAFS
fitting results of the Fe K-edge spectra, the total CNs of Fe-O-Fe (2.88 Å and 3.07 Å)
in Fe12Ni4/CeO2 are 3.6. The Fe-Fe(Ni) coordination (2.50 Å) is observed in
Fe12Ni4/CeO2, indicating that there is a part of Fe species in the form of metallic Fe.
This conclusion is consistent with the formation of Ni-rich NiFe alloy with the FCC
structure revealed by XRD results as the XRD peaks of metallic Fe with the BCC (body-
centered cubic) structure are not observed. In contrast, all the Fe species in the
Fe3Ni1/CeO2 catalyst are oxidized and the total CNs of Fe-O-Fe (2.85 Å and 3.04 Å) is
20
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901585
ChemCatChem
also slightly reduced to only 2.8. The CNs of Fe-O-Fe (2.95 Å and 3.34 Å) on
Fe3Ni1/ZrO2 are fairly low (0.9 for each), suggesting that Fe oxides might be highly
dispersed on Fe3Ni1/ZrO2. The bond lengths of both Fe-O-Fe shells expand, indicating
that the ZrO2 support may influence the geometry of supported Fe species.
Figure 7 In-situ XANES spectra: (a) Ni K-edge of fresh Fe3Ni1/ZrO2, Fe3Ni1/CeO2,
and Fe12Ni4/CeO2 with Ni foil and NiO as references; (b) Ni K-edge of Fe3Ni1/ZrO2,
Fe3Ni1/CeO2, and Fe12Ni4/CeO2 under reaction conditions with Ni foil and NiO as
21
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901585
ChemCatChem
references; (c) Fe K-edge of fresh Fe3Ni1/ZrO2, Fe3Ni1/CeO2, and Fe12Ni4/CeO2 with
Fe foil, FeO, Fe3O4 and Fe2O3 as references; (d) Fe K-edge of Fe3Ni1/ZrO2,
Fe3Ni1/CeO2, and Fe12Ni4/CeO2 under reaction conditions with Fe foil, FeO, Fe3O4 and
Fe2O3 as references. Reaction conditions: 873 K, 5 mL/min C2H6 + 5 mL/min CO2 +
10 mL/min He, 100 mg catalyst.
Table 2 Ni K-edge EXAFS fitting results of Fe3Ni1/ZrO2, Fe3Ni1/CeO2 and
Fe12Ni4/CeO2 under reaction conditions (5 mL/min C2H6 + 5 mL/min CO2 + 10 mL/min
He, 100 mg catalyst, 873 K)
Sample
Shell
Bond length (Å)
2.47 ± 0.01
CN
σ2 (Å2)
E0 shift (eV)
Fe3Ni1/ZrO2 Ni-Ni(Fe)
Fe3Ni1/CeO2 Ni-Ni(Fe)
Fe12Ni4/CeO2 Ni-Ni(Fe)
9.8 ± 0.9 0.016 ± 0.001
9.8 ± 1.2 0.018 ± 0.001
11.3 ± 1.4 0.019 ± 0.001
2.9
1.3
2.47 ± 0.03
2.47 ± 0.01
-9.8
Table 3 Fe K-edge EXAFS fitting results of Fe3Ni1/ZrO2, Fe3Ni1/CeO2 and
Fe12Ni4/CeO2 under reaction conditions (5 mL/min C2H6 + 5 mL/min CO2 + 10 mL/min
He, 100 mg catalyst, 873 K)
Sample
Shell
Fe-O
Bond length (Å)
CN
σ2 (Å2)
E0 shift (eV)
1.93 ± 0.02
3.9 ± 0.4 0.010 ± 0.001
0.9 ± 0.3 0.012 ± 0.001
Fe3Ni1/ZrO2
-4.6
Fe-O-Fe(Ni)
2.95 ± 0.03
22
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901585
ChemCatChem
Fe-O-Fe(Ni)
Fe-O
3.34 ± 0.03
1.92 ± 0.03
2.85 ± 0.04
3.06 ± 0.04
1.96 ± 0.03
2.50 ± 0.03
2.88 ± 0.04
3.07 ± 0.04
0.9 ± 0.3 0.012 ± 0.001
3.0 ± 0.3 0.013 ± 0.001
1.0 ± 0.5 0.012 ± 0.001
1.8 ± 0.6 0.012 ± 0.001
3.9 ± 0.3 0.009 ± 0.001
0.9 ± 0.3 0.010 ± 0.001
1.8 ± 0.4 0.010 ± 0.001
1.8 ± 0.3 0.010 ± 0.001
Fe3Ni1/CeO2 Fe-O-Fe(Ni)
Fe-O-Fe(Ni)
-9.1
-1.9
Fe-O
Fe-Fe(Ni)
Fe12Ni4/CeO2
Fe-O-Fe
Fe-O-Fe
Discussion of active sites
As shown in Figure 1 and Table 1, the supported-Fe3Ni1 catalysts with different
oxide supports present distinguishable catalytic activity. As an inert and irreducible
support, SiO2 cannot promote the activation of CO2 via direct C=O bond scission, which
is very different from the reducible supports (e.g. CeO2). Therefore, Fe3Ni1/SiO2 shows
little activity for either the oxidative dehydrogenation or dry reforming of ethane.
Fe3Ni1/ZrO2 shows high initial activity and selectivity for dry reforming of ethane but
accompanied by severe deactivation. Fe3Ni1/CeO2 exhibits excellent activity and
selectivity for oxidative dehydrogenation of ethane.
According to the previous work [2], the Ni-CeOx interfaces are identified as the
most likely active sites for dry reforming of ethane with CO2 (efficient C-C bond
scission to produce synthesis gas), while the Ni-FeOx/CeOx interfaces are identified as
the most likely active sites for CO2-assisted dehydrogenation of ethane (selective C-H
23
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901585
ChemCatChem
bond cleavage to produce ethylene). The oxygen species for dry reforming at the Ni-
CeOx interfacial sites might mainly come from the Ce(III)/Ce(IV) redox couple while
the oxygen species for oxidative dehydrogenation at the Ni-FeOx/CeOx interfacial sites
might be originated from the Fe(II)/Fe(III) redox couple. For Fe3Ni1/ZrO2, Ni is in a
state of metallic nickel (Ni0) while Fe exists in the form of oxides (FeOx, shows near
edge features between FeO and Fe3O4) under reaction conditions (Figure 7). However,
the CNs of Fe-O-Fe on Fe3Ni1/ZrO2 are fairly low (Table 3) and the diffraction peaks
of any Ni0 or FeOx species are not observed on Fe3Ni1/ZrO2 under reaction conditions
(Figure 6), indicating that both the Ni0 and FeOx species distribute relatively sparsely
and evenly over the ZrO2 support. Therefore, the Ni-ZrO2 interfacial sites likely
dominate the conversion of ethane with CO2, exhibiting a high initial CO selectivity
(over 95%). According to the thermogravimetric analysis (TGA) results (Figure S8),
the deactivation of Fe3Ni1/ZrO2 is suggested to be mainly caused by coking on the
surface. As the Ni-ZrO2 interfacial sites are gradually covered by the coke during the
reaction, the C2H4 selectivity continuously increases during the course of reaction, as
shown in Figure 1.
For CeO2-supported Fe3xNix (x = 0.25, 0.5, 1, 2, 4) catalysts, when the loading
amount of Fe (Ni) is lower than 1.4 (0.5) wt%, the Ni-rich NiFe alloy exists as small
particles (based on the in-situ XRD and XAFS measurements); the FeOx species prefer
to exist underneath the predominantly Ni particles, likely forming the Ni-FeOx/CeOx
interfacial sites that are identified as the most likely active sites for CO2-assisted
dehydrogenation of ethane [2]. More Ni-FeOx/CeOx interfacial sites are formed with
24
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901585
ChemCatChem
the increase of metal loading, leading to higher activity for CO2-assisted oxidative
dehydrogenation of ethane with over 80% ethylene selectivity. When the loading
amount of Fe (Ni) exceeds 5.7 (2.0) wt%, the Ni-rich NiFe alloy resides as lager
particles with diameters of 4.0 nm. The metallic Fe-Fe(Ni) bonding at 2.50Å is only
found on Fe12Ni4/CeO2 while barely seen on Fe3Ni1/CeO2 (Table 3), suggesting that the
ratio of metallic Fe in overall Fe species on Fe12Ni4/CeO2 is much higher than that on
Fe3Ni1/CeO2. This means that FeOx on the perimeter of Ni particles on Fe12Ni4/CeO2 is
likely reduced. Therefore, the NiFe-CeOx interfacial sites that favor the dry reforming
pathway via efficient C-C bond scission to produce synthesis gas are likely formed on
Fe12Ni4/CeO2, leading to a corresponding decrease in ethylene selectivity (from 82%
over Fe3Ni1/CeO2 to 32% over Fe12Ni4/CeO2).
Conclusions
The in-situ XRD and XAFS characterization was combined with flow reactor
evolution to identify the oxidation state and the dispersion of Ni and Fe species on FeNi-
based catalysts for CO2-assisted oxidative dehydrogenation of ethane, aiming to gain a
better understanding of the role of the oxide support and the metal loading amount in
tuning the catalytic activity and selectivity. The formation of Ni-rich NiFe alloy with
the FCC structure and crystalline Fe3O4 or Fe2O3 on CeO2-supported Fe3xNix catalysts
under reaction conditions was revealed by the in-situ XRD and XAFS measurements.
Among these FeNi-based catalysts, Fe3Ni1/CeO2 exhibits the best catalytic activity and
ethylene selectivity. The increased CO selectivity through the dry reforming pathway
25
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901585
ChemCatChem
observed on Fe3Ni1/ZrO2 and Fe12Ni4/CeO2 is likely due to the formation of Ni-ZrO2
and NiFe-CeOx interfacial sites, respectively. The ideal oxidative dehydrogenation
catalyst requires a reducible oxide (CeO2) to facilitate the activation of CO2 via direct
C=O bond scission and optimal loading amounts of active metals (Fe and Ni) to form
the active sites (Ni-FeOx/CeOx interfacial sites) for CO2-assisted dehydrogenation of
ethane via the selective C-H bond cleavage to produce ethylene.
Experimental Section
Catalyst preparation
The supported FeNi catalysts were synthesized by wetness co-impregnation over
commercial supports with an aqueous solution of the respective metal precursors.
Ni(NO3)2∙6H2O (Sigma-Aldrich, 99.999% trace metals basis), Fe(NO3)3∙9H2O (Sigma-
Aldrich, 99.999% trace metals basis), SiO2 (Alfa Aesar, pellets, 250 m2/g), ZrO2 (Alfa
Aesar, pellets, 90 m2/g) and CeO2 (Sigma-Aldrich, nanopowder, 35-45 m2/g) were used.
After impregnation, the catalysts were dried at 363 K for 12 h, then ramped to 723 K
with a rate of 0.8 K/min and calcined for 4 h in air. The metal loading amounts (and
atomic ratio as well) for these catalysts are listed in Table S1.
Catalyst characterization
The pulse CO chemisorption experiments were carried out in an Altamira AMI-
300 ip system to obtain the CO uptake value for each catalyst. 100 mg of fresh catalyst
was firstly reduced at 873 K for 30 min using a mixture of 10% H2 in Ar (30 ml/min),
26
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901585
ChemCatChem
then purged with He (50 ml/min) for degassing and cooled down to 313 K. Pulses of
10% CO in He (590 L loop) were injected into the He stream until the peak area
became constant. The amount of CO flowing out of the reactor was monitored by a
thermal conductivity detector (TCD). The CO uptake offers an approach to estimate the
turnover frequency (TOF) of each catalyst. The CO uptake values for the fresh catalysts
are also listed in Table S1.
The temperature-programmed reduction (TPR) experiments were carried out in the
same Altamira AMI-300 ip instrument. 50 mg of fresh catalyst was pre-treated with a
He flow (30 ml/min) at 393 K for 30 min. After cooled down to 303 K, the TPR
measurement was performed with a heating rate of 10 K/min to 873 K, using a mixture
of 10% H2 in Ar with a constant flow rate of 50 ml/min. The variation of hydrogen
consumption with reduction temperature was continuously recorded by a TCD.
The ex-situ and in-situ X-ray diffraction (XRD) patterns for Fe3Ni1/ZrO2,
Fe3Ni1/CeO2, Fe6Ni2/CeO2, and Fe12Ni4/CeO2 were collected at beamline 17-BM of
advanced photon source (APS) at Argonne National Lab (ANL). The X-ray wavelength
was 0.24128 Å. 10 mg of sieved catalyst (60-80 mesh) was loaded into a quartz
capillary with two open ends (1.0 mm OD, 0.9 mm ID and 75 mm in length), packed
by quartz wools at both sides of the sample. The sample was heated to 873 K at a rate
of 20 K/min and held at 873 K for 40 min in the presence of reactants (2.5 mL/min
C2H6 + 2.5 mL/min CO2 + 5.0 mL/min He). Before ramped down to room temperature,
the sample was treated in a CO2/He flow (2.5 mL/min CO2 + 7.5 mL/min He) for 30
min and then in a H2/He flow (5.0 mL/min H2 + 5.0 mL/min He) for 30 min. Two-
27
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901585
ChemCatChem
dimensional XRD images were continuously collected by a PerkinElmer a-Si flat panel
detector. The images were subsequently processed with GSAS-II to obtain XRD
profiles, and Rietveld refinement was performed to extract structural information.
The in-situ Ni K-edge (8333 eV) and Fe K-edge (7112 eV) X-ray absorption fine
structure (XAFS) spectra for Fe3Ni1/ZrO2, Fe3Ni1/CeO2, and Fe12Ni4/CeO2 were
collected at Beamline 9-BM of Advanced Photon Source (APS) at Argonne National
Laboratory (ANL). For each run, 100 mg sample (60-80 mesh) was loaded into a home-
designed in-situ micro-channel reactor with graphite paper as the window material.
Measurements were performed in a fluorescence mode using a vortex detector. The
sample was firstly reduced in a H2/He flow (10 mL/min H2 + 10 mL/min He) at 723 K
for 60 min. The flow was the switched to a mixture of reactants (5 mL/min C2H6 + 5
mL/min CO2 + 10 mL/min He), and the sample was heated to 873 K at a rate of 10
K/min and held for 30 min. The spectra were collected under reaction conditions and
then analyzed using IFEFFIT package. Ni foil, NiO, Fe foil, FeO, Fe3O4, and Fe2O3
were employed as references of EXAFS fitting.
Catalyst evaluation
Flow reactor studies of C2H6 + CO2 reactions over supported FeNi catalysts were
performed in a quartz tube reactor under atmospheric pressure. For steady-state
experiments, 100 mg of sieved catalyst (40-60 mesh) was used. The catalyst was firstly
reduced under a H2/Ar mixture (20 mL/min H2 + 20 mL/min Ar) at 723 K for one hour
prior to reaction. After reduction, the temperature was increased at a rate of 10 K/min
28
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901585
ChemCatChem
and held at 873 K in the presence of reactants (10 mL/min C2H6 + 10 mL/min CO2 +
20 mL/min Ar) for at least 13 hours. Then, the temperature was changed from 873 K to
823 K with 10 K increments to study the effect of temperature on the reaction rates of
C2H6 and CO2. The concentrations of gas products were analyzed on-line by a gas
chromatography instrument (Agilent 7890B) equipped with a flame ionization detector
(FID) and a TCD. Blank experiments were carried out at 873 K using a blank quartz
tube. Little activity was found, indicating that the quartz tube and the gas-phase reaction
have no significant effect on the reaction.
The steady state conversion (X), C2H6 based yield (Y), C2H6 based selectivity (S),
turnover frequency (TOF) and space-time yield (STY) were defined as
Freinalcettant  Foutlet
reactant
Xreactant
=
(1)
(2)
(3)
Finlet
reactant
outlet
C2H4
F
YC H
=
4
inlet
C2H6
2
F
outlet
F
CO originated from C2H6
YCO
=
inlet
C2H6
2 F
Here, the amount of CO produced for C2H6 can be calculated based on the
oxygen balance,
FCoOutlet  F

outlet 
H2O
outlet
outlet
CO
F
= F

(4)
(5)
(6)




CO originated from C2H6
2


Y
C2H4
SC H
=
4
2
X
C2H6
YCO
SCO
=
X
C2H6
29
This article is protected by copyright. All rights reserved.