Ni-M-O (M = Sn, Ti, W) Catalysts Prepared by a Dry Mixing Method for Oxidative Dehydrogenation of Ethane

Article abstract of DOI:10.1021/acscatal.6b00044

A new generation of Ni-Sn-O, Ni-Ti-O, and Ni-W-O catalysts has been prepared by a solid-state grinding method. In each case the doping metal varied from 2.5% to 20%. These catalysts exhibited higher activity and selectivity for ethane oxidative dehydrogenation (ODH) than conventionally prepared mixed oxides. Detailed characterization was achieved using XRD, N₂ adsorption, H₂-TPR, SEM, TEM, and HAADF-STEM in order to study the detailed atomic structure and textural properties of the synthesized catalysts. Two kinds of typical structures are found in these mixed oxides, which are (major) "Ni_xM_yO" (M = Sn, Ti, W) solid solution phases (NiO crystalline structure with doping atom incorporated in the lattice) and (minor) secondary phases (SnO₂, TiO₂, or WO₃). The secondary phase exists as a thin layer around small "Ni_xM_yO" particles, lowering the aggregation of nanoparticles during the synthesis. DFT calculations on the formation energies of M-doped NiO structures (M = Sn, Ti, W) clearly confirm the thermodynamic feasibility of incorporating these doping metals into the NiO struture. The incorporation of doping metals into the NiO lattice decreases the number of holes (h⁺) localized on lattice oxygen (O^2- + h⁺ → O^?-), which is the main reason for the improved catalytic performance (O^?- is known to favor complete ethane oxidation to CO₂). The high efficiency of ethylene production achieved in these particularly prepared mixed oxide catalysts indicates that the solid grinding method could serve as a general and practical approach for the preparation of doped NiO-based catalysts.

Full text of DOI:10.1021/acscatal.6b00044

Subscriber access provided by UNIV OF CALIFORNIA SAN DIEGO LIBRARIES
Article
Ni-M-O (M=Sn, Ti and W) catalysts prepared from dry
mixing method for oxidative dehydrogenation of ethane
Haibo Zhu, Devon C. Rosenfeld, Moussab Harb, Dalaver H.
Anjum, M. N. Hedhili, Samy Ould-Chikh, and Jean-Marie Basset
ACS Catal., Just Accepted Manuscript • DOI: 10.1021/acscatal.6b00044 • Publication Date (Web): 25 Mar 2016
Downloaded from http://pubs.acs.org on March 29, 2016
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
ACS Catalysis is published by the American Chemical Society. 1155 Sixteenth Street
N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 53
ACS Catalysis
1
2
3
4
5
6
7
8
9
NiꢀMꢀO (M=Sn, Ti and W) catalysts prepared from
dry mixing method for oxidative dehydrogenation of
ethane
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
b
a
c
c
Haibo Zhu , Devon C. Rosenfeld , Moussab Harb , Dalaver H. Anjum , Mohamed Nejib Hedhili ,
a
* a
Samy Ould Chikh , JeanꢀMarie Basset
a
KAUST Catalysis Center, King Abdullah University of Science and Technology, Thuwal 4700,
Kingdom of Saudi Arabia
b
The Dow Chemical Company, 2301 N. Brazosport Blvd., Freeport, TX 77541, USA
c
KAUST Core Lab, King Abdullah University of Science and Technology, Thuwal 4700,
Kingdom of Saudi Arabia
1
ACS Paragon Plus Environment

ACS Catalysis
Page 2 of 53
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ABSTRACT
A new generation of NiꢀSnꢀO, NiꢀTiꢀO, and NiꢀWꢀO catalysts has been prepared by a solid
state grinding method. In each case the doping metal varied from 2.5% to 20%. These catalysts
exhibited higher activity and selectivity for ethane oxidative dehydrogenation (ODH) than
conventionally prepared mixed oxides. Detailed characterisation was achieved using XRD, N2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
adsorption, H ꢀTPR, SEM, TEM, and HAADFꢀSTEM in order to study the detailed atomic
2
structure and textural properties of the synthesized catalysts. Two kinds of typical structures are
found in these mixed oxides, which are (major) “Ni M O” (M = Sn, Ti or W) solid solution
x
y
phases (NiO crystalline structure with doping atom incorporated in the lattice) and (minor)
secondary phases (SnO , TiO or WO ). The secondary phase exists as a thin layer around small
2
2
3
“
Ni M O” particles, lowering the aggregation of nanoparticles during the synthesis. DFT
x y
calculations on the formation energies of Mꢀdoped NiO structures (M = Sn, Ti, W) clearly
confirm the thermodynamic feasibility of incorporating these doping metals into NiO struture.
+
The incorporation of doping metals into the NiO lattice decreases the number of holes (h )
2
ꢀ
+
●
ꢀ
localized on lattice oxygen (O + h
ꢀ
O ), which is the main reason for the improved catalytic
●
ꢀ
performance (O is known to favor complete ethane oxidation to CO ). The high efficiency of
2
ethylene production achieved in these particularly prepared mixed oxide catalysts indicates that
the solid grinding method could serve as a general and practical approach for the preparation of
doped NiO based catalysts.
KEYWORDS: Solid state synthesis, NiO, semiconductor, ethylene production, oxidative
dehydrogenation, single atom
2
ACS Paragon Plus Environment

Page 3 of 53
ACS Catalysis
1
2
3
4
5
6
7
8
9
1
. Introduction
Ethylene is regarded as a very important raw material for the synthesis of a wide variety of
products including polymers, fine chemicals, plastics, and fibers. Its primary use is in the
production of polyethylene, which is the most widely used plastic in the world. In fact, ethylene
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
is the highest volume chemical produced by the petrochemical industry and is primarily
produced by the steam cracking of hydrocarbon feedstock such as ethane and naphtha. The
shortcomings of steam cracking are low olefin selectivity, raw material loss through coke
formation requiring shut downs and maintenance and the large energy demand. In fact, ethylene
production by the steam cracking process (at T>800°C) is considered one of the most energy
2
intensive processes in the petrochemical industry.
With worldwide ethylene demand expected to steadily increase each year for the foreseeable
future, there is an increased focus on the development of new and improved technologies and
processes for ethylene production. The catalytic oxidative dehydrogenation (ODH) of ethane is
3
an attractive alternative technology for the production of ethylene. Firstly, ODH is an
exothermic process and this reaction is carried out at relatively low temperature (typically
4
below 450 °C) in comparison with the steam cracking process. Therefore, the ethylene
production from ethane ODH could save a large amount of energy. Secondly, the large scale
exploitation of shale gas has greatly increased the supply and lowered the cost of ethane. Lastly,
catalyst deactivation from coke formation in the classical steam reforming process is suppressed
5
due to the presence of oxygen and lower temperature operation.
6
A great variety of metal oxide materials are efficient catalysts for the ethane ODH reaction.
Among these oxides, NiO based materials exhibit high activity and selectivity along with good
3
ACS Paragon Plus Environment

ACS Catalysis
Page 4 of 53
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
stability in ethane ODH, especially at temperatures <450 °C. Despite the superior low
temperature ODH performance of NiO based materials further studies are necessary to assess
their potential for use beyond the laboratory scale. The properties of NiO including ODH
performance can be systematically tuned by doping with a secondary metal. The incorporation of
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
7
aꢀe
8
9
10
7f
“dopants’ such as Nb, W, Zr , Sn and Ta into NiO results in formation of a new type of
bimetallic oxide solid solution. The properties and catalytic performance of doped NiO are
affected not only by the identity of the dopant but also by the method used to prepare the mixed
metal oxide. We have shown that the NiꢀNbꢀO and NiꢀTaꢀO synthesized from a solꢀgel method
exhibits higher ethylene selectivities than those obtained by an evaporation or precipitation
7
c, 7f
method.
Furthermore, we have recently developed a more effective and sustainable method
for the synthesis of NiꢀNbꢀO and NiꢀTaꢀO catalyst by a simple dry grinding of Ni and Nb or Ta
11
precursors. This “solid state grinding” synthetic method is green since it does not require any
solvent, even water, in the preparation procedure, which shows great benefit in the practical
application. In this paper, we found that this grinding method could be also extended to the
synthesis of NiꢀTiꢀO, NiꢀSnꢀO and NiꢀWꢀO catalysts, proving that “solid state grinding” could
serve as a general and versatile approach for the preparation of doped NiO. Full characterization
of the mixed oxides and DFT calculations allow us to propose a theory on the origin of improved
ethylene selectivity in ODH with this new generation of dopants.
2. Experimental Section
2.1 Catalysis synthesis
A dry solid state procedure was used to synthesize three series of NiꢀSnꢀO, NiꢀTiꢀO and NiꢀWꢀ
O, catalysts with the doping metal atomic content of 2.5%, 5%, 10%, 15% and 20% in each
4
ACS Paragon Plus Environment

Page 5 of 53
ACS Catalysis
1
2
3
4
5
6
7
8
9
series. The following preparations are typical examples of the synthesis of Ni . M0.05O (M=W,
0
95
Ti and Sn) catalysts. (1) Synthesis of Ni . W0.05O catalyst: 8.20 g of tungsten (VI) ethoxide in
0 95
ethanol solution (W(OCH CH ) , 5% w/v in ethanol) was dried at 70 °C to evaporate ethanol.
2
3 6
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The obtained powder was then mixed with 4.74 g of nickel nitrate hexahydrate (Ni(NO ) .6H O)
3
2
2
and 1.13g of oxalic acid (H C O ) in a mortar bowl. Subsequently, the mixture was mixed and
2
2
4
ground in the mortar bowl by using a pestle at room temperature (nominally 25 °C) for 30
minutes to get a uniform paste, and this paste was submitted to a drying process at 90 °C for 2
hours. Finally, the dried paste was calcined under static air at 300 °C for 4 hours to produce a
black solid. (2) Synthesis of Ni . Ti0.05O catalyst: Replicate the synthesis of Ni0.95W0.05
O
0
95
substituting tungsten(VI) ethoxide with 0.20 g of titanium ethoxide (Ti(OCH CH ) ) in the
2
3 4
synthesis. (3) Synthesis of Ni0.95Sn0.05O catalyst: Replicate the synthesis of Ni0.95W0.05
O
substituting tungsten(VI) ethoxide with 0.30 g of tin(IV) acetate (Sn(CH CO ) ) in the synthesis.
3
2 4
2.2 Catalysts characterization
Xꢀray diffraction (XRD) measurements were recorded using a Bruker D8 Advanced A25
diffractometer with Cu Kα radiation (λ=1.5406 Å) at 40 kV and 40 mA. The data sets were
collected in stepꢀscan mode over the 2θ range 10ꢀ90°, using a step interval of 0.05° and a
counting time of 1s per step.
N adsorption/desorption isotherms were carried out on a Micromeritics@ ASAP2420
2
instrument at 77 K. The samples were degassed in vacuum for 2 h at 150 °C before
measurement. Surface area of the samples were evaluated by the multipoint BET analysis
method in the P/P = 0.05ꢀ0.30 pressure range.
0
5
ACS Paragon Plus Environment

ACS Catalysis
Page 6 of 53
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The temperatureꢀprogrammed reduction (TPR) experiments were performed with an Altamira
Instrument equipped with a thermal conductivity detector (TCD), and the acquired data were
analyzed using the AMI software provided by the instrument company. 50 mg of fresh sample
was loaded in a Uꢀshaped quartz tube, and then was pretreated under high purity (99.9%) argon
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ꢀ1
(30 ml min ) at 350 °C for 4 hours in order to remove adsorbed water and gas. Prior to the test,
the sample was cooled down to the ambient temperature. Standard TPR analysis was done under
ꢀ1
5
% H /Ar with a flow of 30 mL min from room temperature to 1100 °C with a ramp rate of 20
2
ꢀ1
°C min .
XPS studies were carried out in a Kratos Axis Ultra DLD spectrometer equipped with a
monochromatic Al Kα xꢀray source (hν=1486.6 eV) operating at 150 W, a multichannel plate
−
9
and delay line detector under a vacuum of 1~10 mbar . The survey and highꢀresolution spectra
were collected at fixed analyzer pass energies of 160 eV and 20 eV, respectively and quantified
using empirically derived relative sensitivity factors provided by Kratos analytical. Samples were
mounted in floating mode in order to avoid differential charging. Charge neutralization was
required for all samples. Binding energies were referenced to the C 1s peak of (CꢀC, CꢀH) bond
which was set at 284.8 eV. The data were analyzed with commercially available software,
CasaXPS.
The morphologies of the catalysts were analyzed using an FEI Quanta 600 FEG environmental
scanning electron microscope (ESEM). Each sample was sputterꢀcoated with gold under an
argon atmosphere before the SEM image was taken.
Low magnification transmission electron microscopy (TEM) images were obtained on a Tecnai
T12 operated at 120 kV. While, the high resolution TEM and HAADFꢀSTEM were conducted on
6
ACS Paragon Plus Environment

Page 7 of 53
ACS Catalysis
1
2
3
4
5
6
7
8
9
aberration corrected Ttitan G2 80ꢀ300 ST instrument. Samples were dispersed in ethanol with the
help of ultrasound treatment, and then supported onto a 200 mesh carbon coated copper grid
before TEM measurement.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2.3 Catalytic testing
The catalytic ethane ODH reaction was carried out using a fixed bed reactor operating at
atmospheric pressure. 100 mg catalyst was introduced into the reactor with glass wool as
support. The feedstock composed of 10% C H and 5% O in He was passed through the
2
6
2
catalytic bed at a constant flow rate of 10 ml/min (W/F = 0.6 g s/mL). The evaluation of the
catalyst performance was performed from 200 to 400 °C. The gas products were analyzed by an
onꢀline Varian 490 microꢀGC equipped with TCD detectors and two columns: MolSieve 5Å
column (Ar as carrier gas) used to quantify O (CO is not produced under our conditions), and a
2
poraPLOT Q column (He as carrier gas) to analyze CO , C H and C H . The ethane conversion
2
2
4
2
6
and selectivity to ethylene were calculated on a carbon basis, and the carbon mass balance was
above 96ꢁ% in all cases. In order to investigate ethane conversion versus ethylene selectivity at
the constant temperature of 330 °C, the total flow rate was changed from 10 mL/min to 120
mL/min, corresponding to the W/F from 0.05 to 0.6 g s/mL.
3. Computational details
3.1. Supercell models and structure optimization calculations
For simulations of Tiꢀ, Snꢀ, and Wꢀdoped NiO structures, we considered the (2 x 2 x 2) cubic
NiO supercell (space group is FMꢀ3M) containing 32 NiO functional units (Ni O ) or 64
3
2
32
atoms. We modeled several structural configurations by substituting the dopant elements at
various Ni sites accompanied by the removal of additional Ni species. For Tiꢀdoped NiO, one
7
ACS Paragon Plus Environment

ACS Catalysis
Page 8 of 53
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
substitutional Ti at Ni sites in the presence of one or two additional Niꢀvacancies (labeled by
1
Ti @Ni+1V and 1Ti @Ni+2V ) were considered. For Snꢀdoped NiO, one substitutional Sn at
s Ni s Ni
Ni sites (labeled by 1Sn @Ni) as well as one substitutional Sn at Ni sites in the presence of one
s
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
or two additional Niꢀvacancies (labeled by 1Sn @Ni+1V and 1Sn @Ni+2V ) were taken into
s
Ni
s
Ni
account. For Wꢀdoped NiO, one substitutional W at Ni sites in the presence of two or three
additional Niꢀvacancies (labeled by 1W @Ni+2V and 1W @Ni+3V ) were studied. Note that
s
Ni
s
Ni
for NiO, we also explored a defective structure in Ni (labeled by 1V ) which was modeled by
Ni
removing one Ni from the Ni O supercell, leading to the stoichiometry Ni0.96O. For a generic
3
2
32
supercell of doped NiO with the formula Ni(32ꢀk)O M (M = Ti, Sn and W) containing initially 64
32
p
atoms in total, and modified with p M atoms and k Ni vacancies, the atomic concentrations of Mꢀ
p
k
impurities and of Niꢀvacancies are defined as y = 3 and x =
respectively. In the
2
32
2
+
1
Sn @Ni model, Sn species are created, while in all the other structural models explored in our
s
4
+
4+
6+
calculations, the oxidation states of the elements examined are Ti , Sn and W . Note for NiO,
2+
2ꢀ
Ni and O species are formally Ni and O . The various Mꢀdoped NiO materials considered in
our calculations together with their stoichiometries are given in Table 1. Their structural
configurations are displayed in Figure 1.
The various crystal structures were fully optimized using the spinꢀpolarized periodic density
1
2
functional theory (DFT) within the plane wave (PW) approach implemented in VASP program.
The generalized gradient approximation (GGA) within the PerdewꢀBurkeꢀEmzerhof (PBE)
13
14
exchangeꢀcorrelation functional and the projectorꢀaugmented plane wave (PAW) approach
were employed to describe the electronꢀelectron and the electronꢀion interactions, respectively.
8
2
2
4
The valence atomic configurations adopted in PAW potentials are 3d 4s for Ni, 2s 2p for O,
8
ACS Paragon Plus Environment

Page 9 of 53
ACS Catalysis
1
2
3
4
5
6
7
8
9
3
1
2
2
2
4
3
d 4s for Ti,5s 5p for Sn and 6s 5d for W. Cutoff energies of 400 eV and 605.4 eV were used
for wave functions and charge augmentations, respectively. Integration of the band structure
energy over the Brillouin zone was performed with the tetrahedron method with Blöchl
corrections. In all cases, the Brillouin zone was sampled with 5 x 5 x 5 MonkhorstꢀPack kꢀpoint
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
5
grid. The ion coordinates and lattice parameters were fully relaxed until all components of the
residual forces were less than 0.01 eV/Å.
Table 1. Doping models and stoichiometries (including p and k values) of the various Mꢀdoped
NiO materials (M = Ti, Sn and W) with 3 atom % of dopant impurities.
doping model
Structure
Ni(32ꢀk)O M
Ni(1ꢀx)OMy
32
p
stoichiometry
supercell model
1
Ti @Ni+1V
(1a)
(1b)
k=2; p=1
k=3; p=1
k=2; p=1
Ni0.93Ti0.03
O
O
s
Ni
1Ti @Ni+2V
Ni0.90Ti0.03
s
Ni
(
1a)
1Sn @Ni+1V
Ni0.93Sn0.03
Ni0.90Sn0.03
Ni0.96Sn0.03
O
O
O
s
Ni
(1b)
k=3; p=1
k=1; p=1
k=3; p=1
k=4; p=1
1Sn @Ni+2V
s
Ni
1
Sn @Ni
s
(
1c)
1W @Ni+2V
s
Ni
Ni0.90W0.03
Ni0.87W0.03
O
O
1W @Ni+3V
(1d)
s
Ni
3
.2. Formation energy calculations
Identifying the thermodynamic feasibility of assembling the various explored materials from
individual elements was performed by considering the following chemical reaction:
1
2
(
1 − x) Ni (s) + yM (s) + O ( g ) → Ni
OM (s) (M = Ti, Sn and W) (1)
(1− x ) y
2
9
ACS Paragon Plus Environment

ACS Catalysis
Page 10 of 53
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The formation energy (or reaction energy) was computed using the following expression:
0
form
K
Eform = E
− ꢁꢀO (2)
0
form
K
where
E
is the electronic energy at 0 K and expressed by:
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
0
form
K
1
2
E
= E (Ni OM ) − (1− x)E (Ni) − yE (M ) − E (3)
tot
(1−x)
y
tot
tot
O
2
Equation (3) includes the total energies at 0 K for the Ni(1ꢀx)OM , Ni and M (Ti, Sn and W)
y
solids in their groundꢀstate structures as well as for the gas phase O molecule.
ꢁ
ꢀO in equation
2
(2) is the thermal part of the chemical potential of oxygen which depends on temperature (T) and
pressure (p) via the enthalpy (h) and entropy (s) corrections as follows:
p(O2 )
ꢁ
ꢀO = h (T)−Ts (T)+ RTLn(
)
(4)
O
O
2
2
p0
The zero point vibrational energy, the enthalpy correction (h) and the entropy(s) of O as a
2
1
6
function of the temperature (T) and the pressure (p) were calculated using DMolprogram within
17
the PBE exchangeꢀcorrelation functional and the DNP basis set .The zero point vibrational
energy was systematically included in the enthalpy and entropy corrections. The thermal
contributions of the solids were neglected. All electronic energies (for solids and molecule) were
calculated using VASP program. In what follows,
ꢁꢀO was fixed at ꢀ0.22 eV for T = 298 K and
p(O ) = 1 atm (standard thermodynamic conditions). Negative or positive formation energy
2
corresponds to stable or unstable material. Lower or higher formation energy represents more or
less stable material. Note that the formation energy for the defective NiO material (Ni0.96O) was
obtained from eq. (3) for y=0 and x=0.03.
4. Results and discussion
4.1 Catalyst characterization
10
ACS Paragon Plus Environment

Page 11 of 53
ACS Catalysis
1
2
3
4
5
6
7
8
9
We have prepared 3 groups of NiO based catalysts by the simple solid state method. In each
group the doping metal content (atomic percentage) was changed systematically in the ranges of
2.5%, 5%, 10%, 15% and 20%. The XRD patterns of NiꢀTiꢀO, NiꢀSnꢀO and NiꢀWꢀO materials
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
are shown in Figure 2. All the samples show the strong diffraction peaks relevant to the crystal
7
c, 9
structure of the bunsenite NiO (JCPDS 78ꢀ0643).
These characteristic reflection peaks of NiO
are located at 2θ around 37.1°, 43.2°, 62.5°, 74.8°, and 78.7°, which are due to the diffractions
from (111), (200), (220), (311) and (222) planes, respectively. Actually, the width of the NiO
diffraction peaks changes with the variation of doping metals loading. As the content of dopant
increases, these reflection peaks become broad. The peaks widening suggests smaller crystallite
sizes in the samples with higher dopant loading. The pure NiO reference prepared by a
decomposition of Ni(NO ) has a mean crystallite size of 12 nm. In contrast to pure NiO, these
3
2
mixed oxide samples have much smaller particle size. The particle size estimated using the
Scherrer equation shown in Figure 3 clearly indicates an increase of doping metal loading leads
to a decrease of particle size. With the increase of dopant content, the particle size of NiO
reduces to around 4 nm in the samples with 20% dopant metal loading. As proposed by Millet et
al. the decrease of particle size in mixed oxide could be due to a “mutual protective effect”
between NiO and doping metal oxide, which inhibits the growth of particle size during the
7
a
crystallization process.
For the NiꢀTiꢀO samples, only the characteristic peaks related to NiO structure are observable
when the Ti content is 15% and below. An additional peak at 2θ around 25.1° appears in the
sample with Ti loading of 20%. This peak can be indexed to the (101) plane of anatase
1
8
TiO (JCPDS 21ꢀ7212), which suggests the anatase TiO phase is formed in the NiꢀTi mixed
2
2
oxide sample with high Ti content. However, the SnO phase is detectable in all the NiꢀSnꢀO
2
11
ACS Paragon Plus Environment

ACS Catalysis
Page 12 of 53
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
samples. Three weak peaks at 2θ about 26.0°, 33.4° and 51.1° are visible even in the lowest Sn
loading sample of Ni97.5Sn oxide. These three peaks are ascribed to the (110), (101) and (211)
2.5
1
9
planes in tetragonal phase of SnO (JCPDS 41ꢀ1445). In the case of NiꢀWꢀO samples,
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
additional peaks at 2θ of 23.4° and 33.5° being to emerge in the sample of Ni W from the
95
05
2
0
monoclinic WO (JCPDS 43ꢀ1035) structure. With increasing W concentration, these peaks
3
become more evident.
On the basis of XRD results, the NiO crystal is the dominant phase in all the samples. Apart
from this main phase, the structure related to the doping of the metal oxide begins to appear
when the doping metal content increases to a certain level. In the case of Ti, the absence of
observable TiO phase in samples below 15% Ti leads us to hypothesize that most of the Ti
2
atoms are incorporated into the NiO crystal structure. By contrast, the diffraction peaks ascribed
to SnO or WO crystal structures are observed even with low Sn (2.5%) or W (5%) loading.
2
3
From the XRD data, we’re unable to determine whether some Sn or W is inserted into the NiO
crystal lattice. Highly sensitive electron microscopy techniques will be used to probe for
incorporation of dopants into the NiO crystal lattice.
The small particle sizes for the mixed oxides fit well with the high surface areas measured from
the N adsorption test (Figure 4). The surface area of these mixed oxide samples are in the range
2
2
2
of 100 to 250 m /g, which are much higher than that of pure NiO (48 m /g). The unexpectedly
high surface area of doped samples could be due to the presence of an amorphous layer of the
7c
doping metal oxide, as we suggested for NiꢀNbꢀO. It is noted that the surface area decreases in
the samples with high dopant content, especially in sample Ni W and Ni Sn . This could be
80 20
80
20
due to the formation of separate phases of WO and SnO , and they have higher density than
3
2
NiO. These mixed oxide materials have smaller particle size and bigger surface area than those
12
ACS Paragon Plus Environment

Page 13 of 53
ACS Catalysis
1
2
3
4
5
6
7
8
9
8a, 10, 21
made using the typical evaporation method,
which indicates the advantage of this “dry
mixing” synthesis for obtaining nanoꢀstructured materials.
Recent results suggest that the ethane ODH reaction over the NiO based catalyst takes place via
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
2
a Marsꢀvan Krevelen (MVK) mechanism, with the participation of lattice oxygen. In the MVK
mechanism, a redox process occurring on the surface of NiO catalyst, is an important step for
sustaining the catalytic cycle. Thus, the reduction property of NiO is an important parameter for
understanding catalyst performance. H ꢀTPR was performed on the selected NiꢀTiꢀO, NiꢀSnꢀO
2
and NiꢀWꢀO samples. In all the samples (Figure 5, 6 and 7), a small peak located at 210 °C is
clearly visible. A peak at this same position was reported by Nieto et al in NiOꢀCeO mixed
2
2
3
oxides, which they attributed to the reduction of adsorbed oxygen species on the CeO surface.
2
In our case, all the samples demonstrated this peak at the same position, which is unaffected
from changing the doping metal. Thus, we conclude that this peak is derived from the phase
relevant to NiO structure (pure NiO or MꢀNiꢀO solid solution) rather than the doping metal
oxides. This conclusion is consistent with the report of nonꢀstoichiometric oxygen on an NiO
surface, which confirmed that the weak peak at around 200 °C is due to the reduction of
●ꢀ
24
nonstoichiometric electrophilic oxygen O species. Such oxygen species is believed to be
associated directly with the first CꢀH bond activation of ethane but is inversely correlated to
2
2
ethylene selectivity. This peak is hardly observed in Nb and Ta doped as well as undoped NiO
7
c, 9
catalysts prepared using the solꢀgel method.
Therefore, the concentration of electrophilic
ꢀ
●
oxygen O species in NiO based materials made using the dry mixing method is much higher
than that solꢀgel prepared samples, which could explain their different catalytic performance.
Moreover, the intensity of the peak related to nonꢀstoichiometric electrophilic oxygen decreases
13
ACS Paragon Plus Environment

ACS Catalysis
Page 14 of 53
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
with the increase of doping metal loading. Particularly, this peak almost disappears in the
samples of Ni0.80Ti0.20 and Ni0.80W0.20 oxide. These phenomena prove that doping NiO with a
secondary metal could reduce the overꢀstoichiometric oxygen and thus, change the catalytic
performance.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Besides the small peak for nonꢀstoichiometric electrophilic oxygen, additional peaks above 260
°
C are relevant to the reduction of NiO and the dopant metal oxide. As for the pure NiO, the
8
a, 10
reduction peak is located at around 290 °C with a shoulder at about 320 °C.
The peak for
NiO reduction shifts to higher temperatures in all the doped samples. The NiO reduction
temperature is dependent on the nature of the dopant. The reduction temperature of our MꢀNiO
materials follows a sequence of NiꢀTiꢀO > NiꢀWꢀO > NiꢀSnꢀO. The difference in reduction
temperature among these three kinds of catalysts could be ascribed to the different interactions
between NiO and these doping metal oxides. Also, the reduction peak shifts to high temperature
when the content of Ti or W increases in the mixed oxide. However, the NiO reduction
1
0
temperature is not affected by the Sn content, which is similar to previous observations.
In the samples of NiꢀSnꢀO and NiꢀTiꢀO, only the peak related to the reduction of NiO is
4
+
detectable. As suggested by XPS characterization below, the Ti exists as Ti , while the Sn exists
2
+
4+
as Sn and Sn . It should be mentioned that the reduction of bulk oxygen in tin (II, IV) and Ti
1
0, 25
(
IV) has been confirmed to occur at around 600 °C and 650 °C, respectively.
However, the
4
+
4+
2+
peaks resulting from the reduction of Ti , Sn and Sn species are not detectable in these two
types of samples, indicating that the Sn and Ti species becomes less reducible when they are
doped into NiO. By contrast, a peak at 750 °C due to the reduction of W species is observed in
8
a
the Ni0.85W0.15 and Ni0.80W0.20 samples, further confirming the existence of WO phase in the
3
mixed oxide.
14
ACS Paragon Plus Environment

Page 15 of 53
ACS Catalysis
1
2
3
4
5
6
7
8
9
The XPS characterization was further used to identify the chemical state of the selected NiO
samples with 5% and 10% doping metal loading. Figure 8 shows the high resolution XPS spectra
of the Ni 2p core level from the different doped nickel samples of NiꢀWꢀO, NiꢀSnꢀO, and NiꢀTiꢀ
O. A full understanding of the Ni XPS spectrum still remains a matter of debate due to the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
6
controversy in the interpretation of spectrum. The analysis of our spectra was based on the
fitting method from a reference, and a Shirley background was applied across the Ni2p portion
3/2
2
7
of the spectra. The spectrum Ni 2p nickel portion is well fitted with a combination of spectra
3
/2
obtained from standard NiO, Ni(OH) , and NiOOH samples, and the concentration of each
2
species is listed in Table S1. It should be noted that the NiO is the dominant state in all the
samples, which is consistent with the observation from XRD characterization. The presence of
3
+
Ni(OH) could arise from the hydration of surface NiO. The presence of Ni species indicates
2
+
2+
there is nonꢀstoichiometric structure in the NiO structure. The h could be localized on Ni site
3
+
2+
+
3+
to form Ni (Ni + h ꢀ Ni ), and this nonꢀstoichiometric structure results in a low reduction
peak at around 200 °C as observed in the TPR spectra described above. The concentration of
3+
Ni species arising from NiOOH decreases with an increase of doping metal content, suggesting
that doping metal into NiO could reduce nonꢀstoichiometric structure in NiO.
Figure 9 shows the high resolution XPS spectra of the W 4f core level from Ni0.95W0.05 and
Ni0.90W0.10 samples. The W 4f core level was deconvoluted with one doublet W 4f7/2 – W 4f5/2
with a fixed area ratio equal to 4:3 and doublet separation of 2.15 eV. The binding energy of the
6
+
28
W4f7/2 component at 35.7 eV corresponds to W in WO . This XPS result verifies the presence
3
of WO in the NiꢀWꢀO samples, which we observed via XRD analysis.
3
Figure 10 shows the high resolution XPS spectra of the Sn 3d core level from the nickel oxide
doped with 5% and 10% Sn. The Sn 3d5/2 peak was fitted using three components located at
15
ACS Paragon Plus Environment

ACS Catalysis
Page 16 of 53
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
2+
4+
4
84.9, 486.3, and 487.0 eV, which are attributed to Sn metallic tin, Sn in SnO and Sn in
2
9
SnO , respectively. The tin(IV) acetate is the Sn precursor for the preparation of NiꢀSnꢀO
2
4
+
samples. However, most of Sn in tin (IV) acetate gets reduced by the oxalic acid during the
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2+
0
synthesis process, resulting in the formation of Sn and Sn in the mixed oxide. The existence of
4
+
Sn as SnO was observed in our XRD analyses described previously. However, the presence of
2
2
+
2+
Sn as an independent bulk structure had not yet been observed, so the Sn could exist as small
0
clusters or in the NiO crystalline structure. Note the amount of Sn observed through XPS is
negligible, and therefore it is not detected by other techniques.
Figure 11 shows high resolution XPS spectra of the Ti 2p core level from Ni0.95Ti0.05 and
Ni0.90Ti0.10. The Ti 2p3/2 peak was fitted using one component located at 458.4 eV corresponding
3
0
to TiO . This observation is in agreement with our findings from XRD characterization.
2
The SEM images of the Ti, Sn and W doped NiO materials are shown in the supplementary
information (Figure S1). The images show the NiO based materials possess similar morphology.
The primary particles are not uniform; and the sizes range from 1 to 4 ꢀm.
More detailed morphology information of these NiO based materials was obtained by BFꢀTEM
and HAADFꢀSTEM techniques. BFꢀTEM revealed that the big particles (> 1 ꢀm) seen in SEM
micrographs are not single crystallites, but are aggregates of small particles measuring 1ꢀ5
nanometers (Figure 12). The selected area electron diffraction (SAED) characterization was
applied to these areas for the further probe of the structure of the mixed oxides. These SAED
patterns are shown as inserts in the each electron micrograph of Figure 12. It can be noticed from
those SAEDs that all the samples have similar patterns of diffraction rings, indicating they
possess identical structure. Furthermore, the observed diffraction rings match quite well with the
16
ACS Paragon Plus Environment

Page 17 of 53
ACS Catalysis
1
2
3
4
5
6
7
8
9
characteristic structure of cubic NiO phase. These diffraction patterns can be attributed to the
(111), (200), (220), (311), (222), (400), (331) and (422) planes of cubic phase NiO, confirming
the formation of wellꢀcrystalline feature of the NiO structure. SAED did not reveal the
diffraction patterns diagnostic of crystalline SnO , TiO , and WO which were clearly observed
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
2
3
via XRD. However, TEMꢀEDS analysis (Figure S2) on these three samples clearly reveals the
presence of Ti, Sn and Win the mixed oxides. The Ti, Sn and W exist as small oxide clusters or
are incorporated into the NiO crystalline structure, which makes them difficult detect by electron
diffraction.
In order to further investigate the dispersion of doping metals into NiO at the atomic level, high
angle annular dark field (HAADF) STEM imaging analysis was performed. It is equipped with
an aberration corrector to reduce the spherical aberration so that STEM micrographs contain a
subꢀÅ spatial resolution. The aberration corrector improves single atom sensitivity and visibility
in the acquired micrograph, which has been successfully used in identification of single atom
3
1
31d
catalyst. As revealed in a previous report by Wachs et al, individual W atoms and small WO3
clusters are clearly resolved by HAADFꢀSTEM in the WO /ZrO catalyst. Regarding the
3
2
Ni0.90W0.10 catalysts, the W atoms should be in principle observable in HAADFꢀSTEM, because
of a large Z contrast difference arising from different atomic number of W and Ni. In the
HAADFꢀSTEM images small particles approximately 2ꢀ4 nm in diameter are clearly observable,
which possess the characteristic NiO like strucuture (Figure 13). EDS analysis (Figure S3) on the
area in Figure 13a shows the presence of both Ni and W elements, suggesting a intimate
structure comprising Ni and W components was formed. The measured interplanar distance of
2.38 Å is consistent with the (111) plane of NiO. More importantly, there is an array of bright
dots and spots highly dispersed in these small particles, which were never observable by
17
ACS Paragon Plus Environment

ACS Catalysis
Page 18 of 53
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
31d
conventional TEM analysis. From the report by Wachs et al we conclude that these bright dots
are attributed to individual tungsten atoms. Indeed, two different kinds of tungsten arrangements
are distinguishable in the mixed oxide, which are highlighted by yellow and red circles in Figure
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
3c. The bright dots in the yellow circles correspond to isolated tungsten atoms. These atoms are
present inside the NiO crystal lattice forming the solid solution. Those spots in red circles should
be WO clusters (or small particles). Further EDS analysis (Figure S4) on a circle indicates high
3
concentration of W in this area, thus confirming it is polyꢀtungstate species at the surface of NiO.
This observation is also consistent with the XRD and H ꢀTPR characterizations results, which
2
already suggested that there is a separate WO phase in the mixed oxide. On the basis of HAADF
3
in STEM, we can conclude that there are two kinds of tungsten structures in the mixed oxide.
The major portion of the tungsten atoms are incorporated into the NiO structure for the formation
of the NiꢀWꢀO solid solution and a small amount of WO clusters are located at the surface of
3
the NiO particles. It should be noted that this is a typical example of using the HAADF STEM
technique to study the doped NiO based catalysts, and the result here provide a direct and strong
proof that transition metals are incorporated into the NiO crystal lattice.
In the case of our Ni0.90Ti0.10 and Ni0.90Sn0.90 samples, the Ti and Sn atoms are not as clearly
resolved as the W atoms in HAADF micrographs. In STEM, the resolution is highly dependent
on the ability of a target atom to scatter the electron beam. Heavier atoms are more easily
visualized, because the higher atomic mass results in stronger electron scattering. The atomic
number between Ni and Ti or Sn is close, and therefore the Ti and Sn atoms located at NiO
matrix cannot be well identified. (Figure S5 and S6). We can however still find some bright dots
and spots which appear to be the doping metal atoms at the surface of NiO.
4
.2. Structural stabiltiy of doped NiO from DFT calculations
18
ACS Paragon Plus Environment

Page 19 of 53
ACS Catalysis
1
2
3
4
5
6
7
8
9
The characterization results described above indicated that it is possible to incorporate the
doping metal atoms into the NiO crystalline structure. In this section, we will discuss the
thermodynamic feasibility of assembling the mixed metal oxide materials from individual
elements using DFT calculations based on their computed formation energy following the
formalisms described in section 3.2. Figure 1 illustrates the DFTꢀoptimized crystal structures of
the Mꢀdoped NiO materials (M = Ti, Sn and W) explored in our calculations with 3 atom % of
dopant (atomic concentration in the range of the obtained experimental data).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Considering first the pure NiO, the calculated formation energy for the perfect structure is
found to be ꢀ74.5 KJ/mol, while that for the defective structure Ni0.96O associated to two extra
holes is found to be ꢀ87.4 KJ/mol (Figure S7). As the formation energy of Ni0.96O is lower than
that of the perfect one, it can be concluded that the defective structure in NiO is
thermodynamically more favorable than the perfect one. Our computational results are supported
by the fact that NiO is found as a nonꢀstoichiometric material with defects in its structure, and
3
2
this crystallographic defect is the origin of the pꢀtype conductive property of this material.
Since the nonꢀstoichiometric nickle oxide structure with a composion of Ni0.96O is more stable
than the pure NiO, it will be used as our initial structure for establishing the structure of doped
NiO.
The computed formation energies of the structures of Ni0.93Ti0.03 (neutral system) and Ni0.90Ti0.03
4
+
(with two extra holes) associated to created Ti ions are found to be ꢀ104.2 KJ/mol and ꢀ106.8
KJ/mol, respectively (Figure S8). Comparing with Ni0.96O, we can clearly see that both
configurations have lower formation energy, and so, these two compounds are likely to be
thermodynamically more stable than Ni0.96O.
19
ACS Paragon Plus Environment

ACS Catalysis
Page 20 of 53
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
We also calculated the formation energies of Ni0.93Sn0.03O (neutral system) and Ni0.90Sn0.03
O
4
+
(
two extra holes) structures with created Sn ions and values of ꢀ91.1 KJ/mol and ꢀ94.6 KJ/mol
were obtained, respectively (Figure S9). Both values are slightly lower than that for Ni0.96O,
which means that these two compounds are likely to be more favorable than Ni0.96O. In contrast,
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
+
the Ni0.96Sn0.03O structure (neutral system) associated to created Sn ions has given a formation
energy of ꢀ82.8 KJ/mol, which is higher than that obtained for Ni0.96O (Figure S10). This results
means that the Ni0.96Sn0.03O material is likely to be less favourable than Ni0.96O.
For Wꢀdoped NiO compounds, our computed formation energies of Ni0.90W0.03O (neutral
6+
system) and Ni0.87W0.03O (with two extra holes) structures associated to W ions are found to be
ꢀ
99.0 KJ/mol and ꢀ108.7 KJ/mol, respectively (Figure S11). As these two values are clearly
lower than that obtained for Ni0.96O, it can be understood that both structures are likely to be
thermodynamically more stable than Ni0.96O.
As discussed above, the DFT calculation results proved that the incorporation of metal dopant
into NiO is a thermodynamically favorable process. NiO doped with Ti, Sn and W are found to
be more stable than the pure NiO and the nonꢀstoichiometric Ni0.96O. The doping metals can
occupy the Ni crystallographic site, and are incorporated into the crystalline structure of NiO,
resulting in the formation of a new type of solid solution. Moreover, the incorporation of doping
+
metal into NiO leads to a decrease of the amount of holes (h ) in the structure, which will result
in significantly decreasing nonꢀstoichiometric oxygen in the NiO structure. The low
concentration of nonꢀstoichiometric oxygen in the doped NiO is the main reason for improving
the cataltyic performance in ethane ODH.
4
.3 Catalytic reaction
20
ACS Paragon Plus Environment

Page 21 of 53
ACS Catalysis
1
2
3
4
5
6
7
8
9
These three series of NiO based catalysts were tested in the oxidative dehydrogenation of ethane
from 200 to 400 °C. As shown in our previous studies, only ethylene and carbon dioxide were
7
c, 9
detected as the reaction products in the ethane ODH reaction over the NiO based catalysts.
The absence of carbon monoxide as a byꢀproduct may be due to the oxidation of CO into CO by
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
3
3
NiO.
These NiO catalysts were prepared by calcination at 300 ºC, and the catalytic reaction however
was performed up to 400 ºC. A sintering of catalyst may happen when the reaction temperature
exceeds 300 ºC, which could lead to a decrease of the activity. In order to clarify this doubt, a
lightꢀoff behavior of one typical fresh catalyst Ni0.90Ti0.10 in ethane ODH was studied in a
reaction cycle. The reaction cycle consists in one heating step (ramps at 1 ºC/minute from 250 to
3
50 ºC) followed by one cooling step (ramps at 1 ºC/minute from 350 to 250 ºC). Reaction result
in Figure S12 shows that ethane conversion in cooling curve is almost similar to that in heating
curve, which proves that no activity was lost when the catalyst was used above 300 ºC. The spent
catalyst was further studied by TEM and nitrogen adsorption techniques. As shown in Figure
S13, no evident sintering is observed and the particle size remains unchanged after the reaction.
2
2
Moreover, the BET surface areas of the fresh (217 cm /g) and spent (210 cm /g) sample are quite
similar. On a basis of these results, it can be concluded that catalysts prepared by calcination at
3
00 ºC reach a thermodynamically stable state, and their structure does not change when they are
used in the reaction above 300 ºC.
Figure 14, 15 and 16 show the ethane conversion and ethylene selectivity over our NiO based
catalysts, and the oxygen conversions are present in Figure S14, S15 and S16. It should be noted
that ethane conversions reach a plateau accompanied with a slight increase of ethylene selectivity
in most of catalysts, which is due to a large consumption of O in the reaction. A general
2
21
ACS Paragon Plus Environment

ACS Catalysis
Page 22 of 53
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
tendency is observed that the addition of secondary metal by grinding method changes the
activity of NiO in ethane ODH, which could be a result of the alteration of properties of NiO by
the doping metals. Compared with pure NiO catalyst, almost all of the catalysts in the NiꢀTiꢀO
and NiꢀSnꢀO families delivered higher conversion and all of them were more selective to
ethylene. The enhanced activity may be due to the higher surface area and smaller particle size of
these mixed oxides. However, in the NiꢀWꢀO catalyst family only Ni0.975W0.025 exhibits a higher
conversion than pure NiO with all of the samples exhibiting higher ethylene selectivity. This
could be mainly due to the presense of nonꢀstoichiometric oxygen observed by TPR at 200 °C
decreasing in NiꢀWꢀO oxides with increasing tungsten loading. As shown in Figure 7, the
reduction peak for nonꢀstoichiometric oxygen is a slight rise in the baseline for Ni0.85W0.15 and
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
totally disappears in Ni0.80W0.20
.
The selectivity is significantly improved by the doping of NiO with Sn, Ti, or W. In the
reference sample of pure NiO, the selectivity is lower than 50% in all the temperatures
investigated. The addition of a small amount of metal dopant (2.5%) into NiO results in
enhancing the selectivity dramatically. Tungsten is observed to be the most efficient dopant for
improving the selectivity. The incorporation of 2.5% W into NiO increases the selectivity to
65%. This dramatic selectivity increase from a small amount of added W is due to its strong
ability to reduce the number of nonꢀstoichiometric oxygen sites.
The overꢀoxidation of ethane into CO is ascribed to excess nonꢀstoichiometric oxygen at the
2
surface of NiO catalyst. The nonꢀstoichiometric oxygen is evidenced from the H ꢀTPR peak at
2
200 °C, and it systematically decreases with increasing doping metal loading. Therefore, higher
ethylene selectivity is achieved in the sample with higher doping.
22
ACS Paragon Plus Environment

Page 23 of 53
ACS Catalysis
1
2
3
4
5
6
7
8
9
The selectivity is related not only to the reaction temperature, but also to the ethane conversion.
The study of ethylene selectivity versus ethane conversion was carried out by changing the
contact time to get different levels of conversion and selectivity. The relationship between
ethylene selectivity and ethane conversion at 330 °C grouped by metal dopant is shown in
Figures 17, 18 and 19. The ethylene selectivity at the same level of ethane conversion is
dependent on the doping metal loading. High metal oxide doping is beneficial for achieving
improved selectivity. More importantly, the ethylene selectivity keeps almost stable with the
increase of ethane conversion in all the catalysts, which suggests parallel reaction pathways for
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ethylene and CO formation. That selectivity remains unchanged by contact time variation leads
2
us to conclude that CO is mainly produced by an over oxidation of ethane, and not by the
2
secondary reaction of ethylene. The stability of ethylene at the surface of NiO is due to the
34
difficulty of breaking the C=C of ethylene .
Two representative catalysts Ni0.90Ti0.10 and Ni0.90Sn0.10 were investigated in the stability test for
2
4 hours at 330 °C with a feed of 10% C H and 5% O . The catalytic performances of these two
2 6 2
catalysts over 24 hours of timeꢀonꢀstream are presented in Figure S17 and S18. No evident loss
of activity was observed and ethylene selectivity was well maintained in the reaction. These
7
ꢀ10
results are consistent with the findings in references , which proves that NiO based material is
a stable catalyst for ethane ODH.
4.4 Discussion on structureꢀactivity relationship in NiO based catalyst
The reaction mechanism of ethane ODH over NiO based catalyst has been studied by several
7
b, 22a, 35
●ꢀ
groups.
All of these papers concluded that the electrophilic O species, sometimes called
ꢀ
●ꢀ
O , at the NiO surface plays an important role in the ethane ODH reaction. The O species is
23
ACS Paragon Plus Environment

ACS Catalysis
Page 24 of 53
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2ꢀ
+
believed to be formed via the electron exchange between the lattice O and positive h originated
7
a
from the pꢀdoping of NiO. Then, the abundance of nickel vacancies in the pure NiO leads to the
●
ꢀ
formation of a high concentration of O species at the NiO surface. Recently, using a direct
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
●ꢀ
kinetic isotope study, it was proposed that the O species is responsible for activating the CꢀH
bond, leading to a homolytic scission of CꢀH bond that is the rate determining step in the ethane
ꢀ
2
2a
●
ODH reaction (a kinetic isotope effect was observed between C H and C D ). The O species
2
6
2
6
●
attacking a CꢀH bond in ethane results in the abstraction of H from ethane, which forms an ethyl
radical at the “surface” of NiO. It is not fully understood where this ethyl radical coordinates to
2+
2ꢀ
the surface: it could be bonded to a Ni or an O site which are equally abundant on the surface.
On a semiꢀconductor it will find enough electrons in the conduction band to couple with one of
these electrons without looking for localized bonding. We assume that it will coordinate to the
“
surface” as a coordinated ethyl without specifying the coordination site. This surface ethyl
ꢀ
●
radical would lead to ethylene by betaꢀH elimination or it can react further the O to give CO
2
(Scheme 1).
Scheme 1 Ethyl group transformation at the surface of NiO based catalyst.
24
ACS Paragon Plus Environment

Page 25 of 53
ACS Catalysis
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 2 Summarized reaction pathway of ethane ODH over NiO based catalysts from
[
7a, 23]
references.
We have purposely not precisely identified whether the ethyl radical is linked to
surface oxygen or nickel atoms due to a lack of experimental and theoretical evidence.
●
ꢀ
It is clear that the high concentration of O species, generated by the presence of holes leads to
●
ꢀ
increased formation of CO in undoped NiO. Therefore, this electrophilic oxygen O is likely
2
22a
responsible for the formation of CO from a surface bond ethyl group. On the basis of this
2
understanding, a reaction pathway for ethane ODH over NiO can be generalized in Scheme 2. In
●
ꢀ
this simplified reaction pathway, K and K are linked to the concentration of O species at the
1
3
2
6
NiO surface as confirmed by several groups. K is presumably correlated with the reactivity of
2
2
ꢀ 36
lattice O , which can be evaluated from the H ꢀTPR test. The low reduction temperature in H ꢀ
2
2
TPR indicates a high reactivity of the lattice oxygen toward the βꢀhydrogen elimination, and as a
result the high K value for ethylene formation is obtained. Overall, the ethane conversion and
2
ethylene selectivity can be therefore correlated to K , K and K . The ethane conversion is
1
2
3
dependent only on K , since the CꢀH bond cleavage is the rate determining step, whereas the
1
ethylene selectivity is dependent upon K /K .
2
3
Based on the important conclusions discussed above, the key point for the design of highly
efficient NiO catalysts for ethane ODH is rational control of nonꢀstoichiometric oxygen at the
●ꢀ
surface of NiO. The concentration of O species at the surface of NiO needs to be controlled
●
ꢀ
carefully to a reasonable level. Too much O at the surface of NiO will ultimately lead to a total
25
ACS Paragon Plus Environment

ACS Catalysis
Page 26 of 53
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
●ꢀ
oxidation of ethane into CO (K >>K ) On the contrary, a low concentration of O will result in
2
3
2 .
severely limit its ability to activate ethane at low temperature (K close to 0). Therefore, the
1
concentration of nonꢀstoichiometric oxygen should be precisely balanced to merely activate
ethane for its transformation at the surface of NiO, but does not lead to it completely burning into
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
CO (K1>0, and K >> K ). NiO is intrinsically active for ethane oxidation at relatively low
2 3
2
temperature, but it is not selective for the formation of ethylene due to high concentration of nonꢀ
stoichiometric oxygen in its structure. It is proven that transition metal doped into NiO could
largely reduce the nonꢀstoichiometric oxygen concentration in NiO due to a significant decrease
+
of positive h , resulting in a dramatic increase of ethylene selectivity in ethane ODH. We found
that the catalyst preparation method has also significant effect on the structure and property of
7
c, 9
NiO bringing about different catalytic behavior in the ethane ODH.
Accordingly, we have
11
developed a simple but efficient dry grinding method for the synthesis of NiꢀNbꢀO catalysts. In
this report we attempted to extend this preparation method to other NiO based catalysts, and
fortunately it is applicable for synthesizing NiꢀSnꢀO, NiꢀTiꢀO and NiꢀWꢀO catalysts.
Characterization techniques and DFT calculation cleanly indicate that these doping metals were
inserted into the NiO structure, forming a solid solution between NiO and the secondary metal
oxide.
These three series of NiꢀSnꢀO, NiꢀTiꢀO and NiꢀWꢀO catalysts (NiꢀSnꢀO, NiꢀTiꢀO and NiꢀWꢀO)
exhibit enhanced efficiency in ethane ODH reaction as compared with pure NiO, which is
mainly due to the modification of NiO structure by doping metals. Doping NiO with Sn, Ti or W
has three effects on the nature of the NiO, all of which affect the catalyst performance. Firstly,
the concentration of nonꢀstoichiometric oxygen in the mixed oxides largely reduces owing to the
+
decrease of the amount of holes (h ) in the structure, and the amount of nonꢀstoichiometric
26
ACS Paragon Plus Environment

Page 27 of 53
ACS Catalysis
1
2
3
4
5
6
7
8
9
oxygen is inversely associated with the doping metal loading as indicated from H ꢀTPR
2
●
ꢀ
characterization. The decrease of O species in NiO brings about two consequences: (1) the
selectivity to ethylene is largely improved due to the decrease of K ; (2) the ethane conversion
3
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
decreases is some cases especially the NiꢀWꢀO series, because K becomes smaller. Secondly,
1
the particle size of NiO significantly decreases accompanied with an increase of surface area.
The reduced particle size is due to the formation of a protective layer (such as SnO , TiO or
2
2
WO clusters) around NiO surface stabilizing small NiO particles. Smaller NiO particles results
3
in a higher number and enhanced access to active sites increasing ethane conversion. It is found
●ꢀ
that the decrease of active oxygen species (O ) does not lead to a significant loss of activity in
the NiꢀTiꢀO and NiꢀSnꢀO catalysts as compared with pure NiO, because the loss of active oxygen
●
ꢀ
(O ) is compensated by the increased accessibility to the active sites. Lastly, the mixed oxides
exhibit less reducibility than the pure NiO due to the strong interaction between the doping meal
oxide and NiO. This phenomenon should in principle reduce the ethylene selectivity because of
the decrease of K . However, such hypothesis was never confirmed in the catalysts studied. In
2
●ꢀ
fact, the decrease of K is much faster than that of K because of significant loss of O ; therefore
3
2
the ethylene selectivity still increases to a high level. Taking these three aspects into
consideration together, increased ethylene production is indeed achieved by these doped NiO
catalysts synthesized with our solid grinding method.
5. Conclusions
The solid state grinding synthesis method previously disclosed for NiꢀNbꢀO and NiꢀTaꢀO mixed
oxides was successfully extended to three series of NiꢀSnꢀO, NiꢀTiꢀO and NiꢀWꢀO catalysts with
27
ACS Paragon Plus Environment

ACS Catalysis
Page 28 of 53
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
different levels of dopants. Characterization techniques including XRD, N adsorption, H ꢀTPR,
2
2
XSP, bright filed SEM and high angle annular dark field (HAADF) STEM were used to study
the detailed structure and properties of these catalysts. The structure of the mixed oxides has a
certain level of heterogeneity, which is mainly composed of two phases: Ni M O (M = Sn, Ti or
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
x
y
W) solid solution (NiO crystalline structure with doping atom in the lattice) and secondary metal
oxide (SnO , TiO or WO ). DFT calculations performed on the formation energies of solid Niꢀ
2
2
3
MꢀO solution structures clearly confirmed the thermodynamic feasiblility of incoporating these
doping metals into the NiO structure. NiO doped with Sn, Ti and W were found to be more
stable than the pure NiO and the nonꢀstoichiometric Ni0.96O. The incorporation of doping metals
into NiO significantly decreases the particle size to around 4 nm and increases the surface area
2
up to 250 m /g. The presence of nonꢀstoichiometric oxygen was confirmed in the H ꢀTPR test,
2
+
and it dramatically decreases with dopant content because of less holes (h ) in the doped NiO.
The mixed oxides become less reducible due to the strong interaction between NiO and doping
metal oxide.
These catalysts exhibit high performance for the production of ethylene from ethane oxidative
dehydrogenation reaction. The enhanced performance is mainly due to a significant elimination
of nonꢀstoichiometric oxygen and decrease of particle size opening up more active sites. The low
concentration of nonꢀstoichiometric oxygen in the mixed oxide suppresses the formation of CO2
byproduct, resulting in dramatilly increasing the ethylene selectivity. The reduced particle size
●
ꢀ
and increased surface area compensate the loss of actvity arising from the low O content and
therefore, the activity of mixed oxide is still kept at a relatively high level. As a consequence of
these two facts, the high ethylene production efficiency was achieved in the mixed oxide
catalysts.
28
ACS Paragon Plus Environment

Page 29 of 53
ACS Catalysis
1
2
3
4
5
6
7
8
9
Supporting Information.
These include SEM images of NiꢀSnꢀO, NiꢀTiꢀO and NiꢀWꢀO samples, EDS in TEM analysis on
NiꢀSnꢀO, NiꢀTiꢀO and NiꢀWꢀO catalysts, HAADF STEM images from Ni0.90Ti0.10 and
Ni0.90Sn0.10 oxides and formation energies of Mꢀdoped NiO structures (M = Sn, Ti and W). This
material is available free of charge via the Internet at http://pubs.acs.org.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
AUTHOR INFORMATION
Corresponding Author
*
Eꢀmail: jeanmarie.basset@kaust.edu.sa
Acknowledgment
The authors acknowledge financial support from The Dow Chemical Company.
REFERENCES
(1) True, W. R. Oil Gas J. 2010, 108, 34ꢀ38.
(2) Ren, T.; Patel, M.; Blok, K. Energy 2006, 31, 425ꢀ451.
(3) (a) Gärtner, C. A.; van Veen, A. C.; Lercher, J. A. ChemCatChem 2013, 5, 3196ꢀ3217;(b)
Valente, J. S.; ArmendárizꢀHerrera, H.; QuintanaꢀSolórzano, R.; del Ángel, P.; Nava, N.; Massó,
A.; López Nieto, J. M. ACS Catalysis 2014, 4, 1292ꢀ1301; (c) Zhu, H.; Laveille, P.; Rosenfeld,
D. C.; Hedhili, M. N.; Basset, J.ꢀM. Catal. Sci. Technol. 2015, 5, 4164ꢀ4173 (d) Koirala, R.;
Buechel, R.; Krumeich, F.; Pratsinis, S. E.; Baiker, A. ACS Catalysis 2015, 5, 690ꢀ702.
(4) Cavani, F.; Ballarini, N.; Cericola, A. Catal. Today 2007, 127, 113ꢀ131.
(5) Bhasin, M. M. Top. Catal. 2003, 23, 145ꢀ149.
(6) (a) Seshan, K. In Catalysis: Volume 22, The Royal Society of Chemistry: 2010; Vol. 22, pp
1
1
19ꢀ143;(b) Nieto, J. M. L.; Solsona, B.; Grasselli, R. K.; Concepcion, P. Top. Catal. 2014, 57,
248ꢀ1255;(c) Popescu, I.; Heracleous, E.; Skoufa, Z.; Lemonidou, A.; Marcu, I.ꢀC. Phys. Chem.
Chem. Phys. 2014, 16, 4962ꢀ4970;(d) Qin, Z.; Zhou, D. S.; Wu, Y.; Wu, T. H. J. Rare Earths
013, 31, 669ꢀ673.
7) (a) Savova, B.; Loridant, S.; Filkova, D.; Millet, J. M. M. Appl. Catal., A 2010, 390, 148ꢀ
57;(b) Skoufa, Z.; Heracleous, E.; Lemonidou, A. A. Catal. Today 2012, 192, 169ꢀ176;(c) Zhu,
2
(
1
H.; OuldꢀChikh, S.; Anjum, D. H.; Sun, M.; Biausque, G.; Basset, J.ꢀM.; Caps, V. J. Catal. 2012,
29
ACS Paragon Plus Environment