Full text of DOI:10.1002/cctc.201701940

DOI: 10.1002/cctc.201701940
Full Papers
Oxidative Dehydrogenation of n-Octane over Niobium-
Doped NiAl2O4: An Example of Beneficial Coking in
Catalysis over Spinel
Majid D. Farahani, Venkata D. B. C. Dasireddy, and Holger B. Friedrich*[a]
NiAl2O4-based materials are known as enhanced catalysts for
several catalytic applications that have coke deposition in
common. In this study, Nb as a dopant was found to occupy
the octahedral sites of the spinel preferably. As a result, the
cation distribution was altered between the two sub-lattices of
the cubic spinel. Therefore, ordered modifications of nickel mi-
gration, electronic profile, redox properties and surface tex-
tures were observed for these catalysts. Also, coke deposition
over these catalysts was controlled using Nb doping and
found to modify the original features of the fresh catalysts and
lower their crystallite sizes. This study suggests reasons why
NiAl2O4 spinel performs so well in the majority of the high-
temperature catalytic processes that suffer from coke deposi-
tion.
Introduction
Nickel aluminate is a member of the spinel family with the
general formula of AB2O4 and is known to have a magneto-
plumbite-like structure.[1–3] Preparation techniques for the syn-
thesis of NiAl2O4 include coprecipitation,[3,4] sol–gel,[5] solid-
state reaction,[6] the alkoxide method,[7,8] Pechini,[9,10] and com-
bustion synthesis.[11,12] NiAl2O4 had been used for the partial
oxidation of methane, methane (steam or dry) reforming,[9] oxi-
dative coupling of methane,[13] diesel steam reforming,[14,15] tar
reforming,[16] and tetradecane reforming.[17] NiAl2O4 is known to
have an inverse spinel structure, and the degree of this inver-
sion can be varied by modifying the preparation technique
and the temperature of the synthesis.[6,9,18–21] Furthermore, mi-
gration of nickel between the two sub-lattices (changing the
degree of inversion) or to the NiAl2O4 surface is well estab-
lished and seems to be highly dependent on catalytic reaction
conditions.[9]
cover the surface and deactivate the catalyst during paraffin
activation.[24]
However, exceptions are known to the above-mentioned
phenomena, and some examples of beneficial coking exist, in
which catalytic reactions were enhanced upon coke forma-
tion.[25–27] The reported reasons for profitable coking are that
deposited coke acts as active sites, the selective deactivation
of nonselective catalytic sites and the participation of the
formed coke in the catalytic reactions.[28] Interestingly, NiAl2O4
is found to be active in catalytic reactions that normally suffer
from coke formation and minimises coke deposition.[13,15]
However, there is no detailed explanation for this behaviour,
and more investigations seem to be required. Furthermore,
NiAl2O4 is highly stable in catalytic applications that involve
coking.
Beside, few reports exist that discuss the effect of foreign-
metal doping in the lattice of these binary metal oxides and
their physicochemical characteristics.[20,21,23] Therefore, this
study focuses on the modification of the physical and chemical
properties of this structure upon coke deposition. Nb was
chosen as a dopant because of its interesting chemistry such
as acting as high-valance dopant (HVD) or low-valance dopant
Substituting the dopant atoms (even in small quantities) in
the lattice of a host oxide can have profound influences on
the properties of the host, such as morphology changes,
changes in electronic and transport properties and lower tran-
sition temperatures. These are important parameters in hetero-
geneous catalysis.[22] Material-based studies on doped NiAl2O4
have shown some significant changes, such as ion migration
upon the occupancy of foreign metals and modifications in
the lattice constant.[20,23]
2À
(LVD) in the form of Nb2O4 and the modification of electro-
conductivity of nickel-based materials by diminishing the
number of positive holes, if doped into the NiO lattice and
used for the oxidative dehydrogenation of ethane.[29–32] Also, it
has a similar atomic radius to those of Ni and Al, making it a
suitable substituent. Several physicochemical characterisation
techniques were used to identify the changes caused by Nb
doping in this system. Then, these catalysts were used for the
ODH of n-octane under “anaerobic conditions”. Afterwards, the
used materials were fully characterised to see how the coking
influenced the composition of these catalysts. Finally, the cru-
cial role of coking in catalysis over NiAl2O4 is discussed. n-
Octane was chosen as a model substrate, because medium-
Coke formation in oxidative dehydrogenation (ODH) (espe-
cially under anaerobic conditions) is expected. The coke can
[a] Dr. M. D. Farahani, Dr. V. D. B. C. Dasireddy, Prof. H. B. Friedrich
School of Chemistry and Physics
University of KwaZulu-Natal
Durban 4000 (South Africa)
E-mail: friedric@ukzn.ac.za
Supporting information and the ORCID identification number(s) for the
author(s) of this article can be found under:
https://doi.org/10.1002/cctc.201701940.
ChemCatChem 2018, 10, 1 – 12
1
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
chain length paraffins are abundant and of low intrinsic
value.[33,34]
XRD patterns of NiAl2O4 (SP) and NiAl2O4 containing 0.02 at.%
Nb (SP-0.02Nb) showed only the peaks of spinel (ICDD
PDF#78-0552) (Figure 1a). The formation of the NiNb2O6 phase
(ICDD PDF#00-032-0694) was detected when higher concentra-
tions of Nb precursors were used for the syntheses of SP-0.06
Nb and SP-0.10 Nb (Figure 1a). This phase belongs to the co-
lumbite family and is known to be inert under oxidative dehy-
drogenation reaction conditions.[35,36] Also, there are a few
small peaks at 2qꢀ388, 438, and 548, which belong to the co-
lumbite phase.[35] In addition, the high intensity of the peaks in
the XRD diffractograms suggest a high degree of crystallinity
for these catalysts, which is driven by the high-temperature
treatment during the synthesis. The calculated average crystal-
lite sizes of the spinels, using the Scherrer equation, for all cat-
alysts showed an increase in crystallite size with the introduc-
tion of niobium (Table 1), except SP-0.06 Nb. The small loading
of Nb in SP-0.02Nb caused a noticeable increase of crystallite
size if compared to SP. However, further increase of Nb doping
(in SP-0.06Nb) and slight formation of NiNb2O6 seems to mini-
mise the crystal growth effect that was observed for SP-0.02
Nb, and result in a slight increase of crystallite size of NiAl2O4
spinel if compared to SP. Also, the catalyst with the highest Nb
doping (SP-0.10Nb) showed the largest crystallite size. The ob-
served pattern is probably the result of induced synergic ef-
fects owing to NiNb2O6 formation and the amount of Nb
doping in different quantities for each catalyst.
Results
X-ray diffraction (XRD) and inductively coupled plasma opti-
cal emission spectroscopy (ICP–OES)
It has been shown that sol–gel auto combustion using oxalyl-
dihdrazide facilitates the synthesis of NiAl2O4.[12] However, a
high calcination temperature (ꢀ10008C) is required for the
synthesis of pure-phase NiAl2O4.[9] The targeted elemental com-
positions were achieved and confirmed by using inductively
coupled plasma optical emission spectroscopy ICP–OES
(Table 1). The powder XRD (PXRD) data of the prepared cata-
lysts are shown in Figure 1a. The absence of any nickel oxide
phase was confirmed by XRD analysis for all four catalysts. The
Table 1. Elemental and phase compositions of the as prepared catalysts.
Catalyst
Ni[a]
Nb[a]
L[b] [nm]
fresh
used
SP
1.00
0.98
0.97
0.95
–
22.9
29.4
25.1
31.8
12.00
12.6
15.6
21.5
SP-0.02 Nb
SP-0.06 Nb
SP-0.10 Nb
0.02
0.06
0.10
No change in the XRD peak positions for the fresh and used
catalysts is observed, which is proof of the thermal stability of
these materials under the reaction conditions (after 48 h at
5008C). The intensity of all the spinel peaks had decreased
after the reactions and small shoulders appeared on the left
side of each peak that belong to different spinel facets (Fig-
ure S9). This implies smaller crystallite sizes for the used cata-
lysts compared to what was observed for the fresh catalysts
and slight structural alteration during the catalytic application.
The crystallite sizes for the used catalysts are tabulated in
Table 1. Generally, the crystallite sizes increased with Nb con-
tent for these used materials. Further studies on the XRD pat-
terns of these catalysts were performed to probe the possibili-
ty of Nb substitution into the spinel cubic lattice.
[a] The atomic ratio between the elements according to ICP–OES.
[b] Average crystallite size of NiAl2O4 estimated using the Scherrer equa-
tion.
According to reports, the planes (220) and (422) are sensi-
tive to cation distribution in the tetrahedral (Td) sub lattice of
the cubic spinel.[20,23,37] Normally, an increase in intensities of
the peaks corresponding to these planes suggests substitution
in the Td sites. Otherwise, the octahedral (Oh) site is the host
for the foreign element.[20,23,37] As shown in Figure 2, the inten-
sities of both peaks decreased if 0.02 atomic% nickel was sub-
stituted with niobium in SP-0.02Nb, indicating the replace-
ment of nickel in the Oh sub-lattice in this molecule. However,
a different behaviour was observed if the molecular formula
was chosen to result in Nb replacing both Al and Ni ions. The
data, in Figure 2, shows that substitution can take place in
both sub-lattices based on the available ionic vacancies in the
spinel structure. Also, the formation of NiNb2O5 shows that the
substitution of Nb in the spinel lattice is partial. The same
trend of substitution in the sub-lattices was observed for the
used catalysts, suggesting the retention of the original struc-
Figure 1. PXRD patterns of a) as-prepared catalysts, b) used catalysts.
&
ChemCatChem 2018, 10, 1 – 12
www.chemcatchem.org
2
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
seems to replace both Al3+ and Ni+2 ions in the Oh and Td sub-
lattices and results in weaker peaks for both sites, than for the
other catalysts (Figure 3). The highest concentration of Nb
doping (and the lowest molar ratio of Ni2+), in SP-0.10Nb,
showed the lowest Ni+2 concentration in Oh sites, whereas Td
coordinated Ni2+ occupancy was similar to that of the undop-
ed spinel (Figure 3). Thus, the UV–DRS and XRD data suggest
Nb prefers to be octahedrally coordinated and motivates the
Ni2+ migration if any ionic vacancies are present in the tetrahe-
dral sub-lattice (based on the molar ratio between Ni and Al).
Nitrogen physisorption analysis
The BET surface areas and pore volumes of the catalysts de-
creased with the addition of niobium (Table 2). However, no
regular trend was observed in the change of pore diameter for
Figure 2. Change in intensities of plane (220) and (422) in the fresh and
used catalysts. (The intensity of the peak with an hkl value of (311) was
used as a reference in each diffractogram).
Table 2. Surface area, pore volume and pore diameter of SP-x Nb cata-
lysts (x: 0.0–0.1).
tures. However, some nickel ions seem to have migrated from
the Oh to the Td sites or left the lattice of spinel, as the intensi-
ties of both planes were consistently higher than what was ob-
served for the fresh catalysts. This change shows that levels of
inversion for the under study spinels decrease during the cata-
lytic reactions.
Catalyst
BET surface area
[m2 gÀ1
Pore volume
[cm3 gÀ1
Pore diameter
[nm]
]
]
fresh
used
fresh
used
fresh
used
SP
17
13
11
8
22
15
12
8
0.04
0.03
0.02
0.01
0.10
0.08
0.05
0.01
10.4
8.2
8.5
17.7
19.9
17.9
6.6
SP-0.02Nb
SP-0.06Nb
SP-0.10Nb
8.3
Ultraviolet diffuse reflectance spectroscopy (UV–DRS) analy-
sis
the different catalysts. Also, the N2 adsorption–desorption iso-
therms showed that the observed porosity for these catalysts
originated from the voids between the particles, as the hyste-
resis loops were at a high relative pressure (Figure S10). The ni-
trogen physisorption measurements of the used catalysts
showed higher surface areas, pore volumes and pore diame-
ters than the fresh catalysts. The difference between the physi-
cal properties of the fresh and used catalysts is significant for
SP-0.02Nb, and the difference becomes smaller as the concen-
tration of the niobium phase increases in the catalyst mixtures.
To investigate the coordination of Ni2+ in these catalysts fur-
ther, ultraviolet diffuse reflectance spectroscopy (UV–DRS)
spectra were obtained. The bands at 716 nm and 640 nm
result from d–d electron transition of the Ni2+ in the Oh and Td
sub lattices of spinel, respectively.[13,38] The low concentration
of introduced Nb in SP-0.02Nb seems to replace the Ni2+ in Oh
positions and motivate the migration of Ni2+ in Oh sites to the
Td sites, as the intensities of the peaks at 716 and 640 in the
UV spectrum of SP-0.02Nb are, respectively, lower and higher
than those of SP (Figure 3). However, niobium in SP-0.06Nb
TEM and SEM analysis
These catalysts showed hexagonal morphology as shown in
Figure 4a. Both NiAl2O4 and NiNb2O6 have this morphology,
and hence they cannot be separately identified.[35] To investi-
gate the distribution of all three elements in these catalysts,
energy dispersive X-ray spectrometry (EDX) mapping analysis
was performed in DFSTEM (dark-field scanning transmission
electron microscopy) mode using high-resolution transmission
electron microscopy (HRTEM). Formation of NiAl2O4 always re-
sults in a good dispersion of nickel. Also, no clusters containing
Nb were detected in SP-0.06Nb Figure 4b, even if some
NiNb2O6 was detected as the additional phase in the PXRD pat-
tern (Figure 1). The HRTEM images show that both the NiAl2O4
and NiNb2O6 phases positively contributed to the observed
high dispersion for these catalysts. SEM images showed
Figure 3. UV-DRS data of the SP-x Nb catalysts (x: 0, 0.02, 0.06 and 0.10).
ChemCatChem 2018, 10, 1 – 12
www.chemcatchem.org
3
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
Figure 5. SEM images of SP-xNb catalysts [x: a) 0.00, b) 0.02, c) 0.06 and
d) 0.10].
vation of the NiO/Al2O3 XPS peak in SP, which was not detect-
ed by XRD analysis, possibly because of the detection limit of
that technique. SP-0.02Nb has mainly octahedrally coordinated
Ni2+ and a very small quantity of Ni3+ on the surface of cata-
lyst. SP-0.06 Nb showed only Ni2+ with octahedral coordina-
tion.
Figure 4. a) General TEM image of the as-prepared catalysts. b) Mapping
analysis of Sp-0.06Nb using HRTEM (DFSTEM mode).
The XP spectrum of O1s is shown in Figure 6. The binding
energy of 529.2 eV is typical of the metalÀoxygen bond.[39] The
observed peak at approximately 531.5 eV represents oxygen
ions in low coordination at the surface.[39] SP and SP-0.10Nb
showed very similar types of oxygen on the surface and ionic
oxygen seems to be dominant for SP-0.02Nb. However, only
oxygen atoms from the MÀO bond are seen for SP-0.06 Nb.
The ionic oxygen refers to the available lattice oxygens (O2À)
that are exposed on the surface in ionic form and normally are
present in the vicinity of defect sites.[39,41] The observed order
sponge-shaped morphology, and the introduction of niobium
seems to decrease the size of these spongy particles (Figure 5).
X-ray photoelectron spectroscopy (XPS) analysis
The XP spectra of the spinel-based catalysts are shown in
Figure 6. The Ni2p scan normally shows two main peaks,
Ni2p1/2 and Ni2p3/2, with their related satellite
peaks.[9] A separation of 6.3Æ3.0 eV was found be-
tween the main peaks and their representative satel-
lite, and this is known to be indicative of nickel
bound in NiAl2O4, whereas pure NiO has a separation
of 7.1Æ1.0 eV and was not observed in this study.[3]
The main peaks of Ni2p3/2 were deconvoluted to fur-
ther investigate the available nickel species in these
catalysts. The red graphs in the Ni2p3/2 spectra
(Figure 6) with the binding energy of 853.9–855.8 eV
in all catalysts shows the available Ni2+ in the octa-
hedral sites of the spinel structure (Table 3).[39] The
pink graph in the Ni2p3/2 spectra (Figure 6) with the
binding energy of 856.4–857.1 eV is representative of
strongly bonded Ni3+ on the surface of the NiAl2O4
spinel (Table 3).[39,40] The green graph in the Ni2p3/2
spectrum of SP (Figure 6) with the binding energy of
858.9 eV shows the NiO/Al2O3 (Table 3) and is only
seen for SP as also detected in H2 temperature-pro-
grammed reduction analysis (H2-TPR). SP and SP-
0.10Nb have similar structures, except for the obser-
Figure 6. XP spectra of SP-xNb. [x: a) 0.00, b) 0.02, c) 0.06 and d) 0.10] Deconvolutions of
the core-level spectra Ni2p, O1s are provided.
&
ChemCatChem 2018, 10, 1 – 12
www.chemcatchem.org
4
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Table 3. Surface compositions of the fresh catalysts determined by XPS.
Catalyst
Ni2p3/2
Ni3+/NiAl2O4
O1s
Nb3d
(doublet)
Theo.
(Nb/Ni)
Exp.
(Nb/Ni)
NiO
Ni2+
O(ion)
MÀO
[Oct]
SP
858.9
856.7
856.4
–
855.8
854.3
853.9
855.2
531.7
529.9
–
530.7
528.9
529.2
530.7
–
0/1
0/1.17
SP-0.02 Nb
SP-0.06 Nb
SP-0.10 Nb
–
–
–
210.5, 207.3
210.5, 207.4
210.5, 207.3
0.02/0.98
0.06/0.97
0.10/0.95
0.03/1.14
0.07/1.20
0.12/1.14
857.1
531.7
for O2À on the defect sites is SP-0.02Nb>SPꢀSP-0.10Nb>SP-
0.06Nb. Therefore, this data shows that a higher concentration
of Ni3+ on the surface does not necessarily create more defect
sites, and that there is a limitation to which doping can gener-
ate more defect sites. The Nb3d scan showed a doublet at
210.5 and 207.3 eV for all samples, which indicates the pres-
ence of Nb5+ in all catalysts.[42,43] No change in the binding en-
ergies of the observed doublet for Nb3d in all catalysts implies
an indirect role of this element in the changing of the physico-
chemical properties of these catalysts.
showed very similar temperatures for SP[Oh] and SP[Td], with a
smaller reduction peak seen for SP[Oh] in SP than for SP-0.10 Nb
as the only difference. The reduction peak of SP[Oh] became
smaller and shifted to a higher temperature in SP-0.02Nb than
for SP, which suggests a lower population of nickel ions in Oh
sites for this catalyst. The shift of the SP[Oh] peak in SP-0.06Nb
to higher temperature followed the same trend as seen for SP-
0.02Nb and resulted in the same peak positions of SP[Oh] and
SP[Td] for this sample.
A study of the facets that are sensitive to substitution in Td
sites in the XRD and UV–DRS analyses suggested the following
order for the population of nickel in the Oh environment: SPꢀ
SP-0.10Nb>SP-0.02Nb>SP-0.06Nb. However, the XPS data
suggested the following order of increase of defect surface for
the spinel: SP-0.02Nb>SPꢀSP-0.10Nb>SP-0.06Nb. Zhang
et al. reported that nickel ions in octahedral sites are reduced
more easily than those in tetrahedral sites.[7] Considering the
high chance of Ni2+ exposure in Oh sites and the suggested
population of defect sites from the XPS, it can be stated that
the defect sites belong to the Oh environment on the surface
of the catalysts and the combination of both effects resulted
in complex systems for the observed first TPR analyses.
The catalysts were re-oxidised using oxygen as oxidant. It is
well established that nickel atoms in the metallic state leave
the lattice of spinel.[11] Each type of generated metallic nickel
atom will result in a different type of nickel oxide, and a
second TPR can be viewed as a good technique to identify
these species. As shown in Figure S2, three different types of
NiO were identified after the reduction of the catalysts for the
second time. The total peak areas for two of the three formed
nickel oxides decrease as the concentration of Nb increases in
these catalysts. Furthermore, the peak areas of SP[Oh] decrease
as the Nb content of these catalysts increases, suggesting that
the presence of nickel in the octahedral sites is diminished by
the addition of Nb.
The surface compositions of these catalysts are shown in
Table 3. The concentrations of niobium and nickel were found
to be higher on the surface than in the bulk, which is in con-
trast to the results of the reported Pechini synthesis by Sievers
and co-workers.[9] This can be counted as a major advantage of
using sol–gel auto-combustion synthesis (SGCS) to synthesize
NiAl2O4-based catalysts, as a high nickel (active sites) exposure
on the surface can be achieved. SP-0.06Nb shows the largest
difference between the surface and bulk composition, suggest-
ing that Nb introduction can enhance the nickel exposure on
the surface of catalysts to some extent. The NiNb2O6 phase in
these catalysts can also contribute to this observation.
H2 temperature-programmed reduction (H2-TPR) data
H2-TPR analyses were performed to further investigate the
effect of Nb doping on the redox properties of the catalysts.
The complete reduction of bulk NiAl2O4 requires a high tem-
perature (>9008C) as shown in Figure S1 in the Supporting In-
formation, which is in good agreement with reports.[9,44,45] De-
convolution of the spinel peaks was performed based on the
Gaussian model for qualitative investigation purposes.
A small peak at ꢀ3008C was observed for all catalysts,
which is attributed to the reduction of nickel oxide on the sur-
face of bulk spinel.[9] In addition, the SP catalyst presented a
very small peak at 4238C, which is likely owing to a small
quantity of the NiO phase (without any interaction with the
spinel surface) that could not be detected by XRD and other
characterization techniques used in this study, except XPS.[46]
As seen from the deconvolution in Figure S1, two peaks are as-
signed to the spinel reduction for all catalysts. According to
the literature, most of the cubic spinel lattice contains ions in
the Oh sub lattice and these have a greater chance of exposure
on the surface of these materials.[47] Therefore, it can be postu-
lated that the first reduction peak of spinel (SP[Oh]) is related to
the available nickel in the Oh sites, and the second corresponds
to the nickel ions in Td sites (SP[Td]). The SP and SP-0.10Nb
Based on the literature, the first reduction peak (ꢀ1908C) in
the second TPR cycle data can be assigned to the reduction of
Ni3+ [46]
This peak was only observed in SP and SP-0.02Nb,
.
which suggests that Ni3+ species possibly originated from the
sites with octahedral coordination. Also, the peaks at ꢀ2308C
and 3508C are assigned to the reduction of different crystallite
sizes of NiO with no interaction with the support.[46] It is also
well accepted that the NiO peak at temperatures below 4008C
originated from octahedral coordination, whereas the higher
temperature peak corresponds to tetrahedrally coordinated
NiO supported on alumina.[48] Therefore, the observed peak at
5508C for SP-0.10Nb is attributed to large particles of nickel
ChemCatChem 2018, 10, 1 – 12
www.chemcatchem.org
5
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
oxide supported on alumina with tetrahedral coordination.[46,48]
The oxygen atoms that are in the Oh sub lattice seem to con-
tribute to the redox properties of these materials, as the peak
for Ni in Oh sites was shifted to lower temperatures with
shrunken areas when the first and second TPR cycles were
compared.
adverse effect on the selectivity towards octenes, suggesting
consecutive cracking and combustion of the primary products
to secondary products (aromatics, cracked and COx) at high
temperatures.
Except for SP-0.06Nb, the selectivity towards aromatics de-
creased as the Nb content of these catalysts increased
(Figure 8). Also, there is a direct correlation between pore
volume and selectivity to aromatics, and SP-0.06Nb showed
the lowest selectivity towards aromatics. Furthermore, the in-
crease of the temperature positively affects the aromatics for-
mation over the two least active catalysts (SP-0.06Nb and SP-
0.10Nb), whereas the most active catalysts showed a high pro-
pensity to crack the formed octenes or aromatics if the tem-
perate was increased to 5508C from 5008C.
Olefin cracking is highly dependent on the surface acidity of
the catalyst.[24] The catalytic data suggests the lowest selectivi-
ty to cracked products was obtained over SP-0.02Nb, whereas
SP showed the highest selectivity to cracked products
(Figure 8). This suggests that the substitution of Ni[Oh] sites
with Nb resulted in lower selectivity towards cracked products;
however, the contributions of NiNb2O6 to the surface acidity
and area, as well as doping, created a complex set of results
for interpretation of this catalytic data, which is out of the
scope of the current study.
Catalytic data on n-octane conversion
A carbon-to-oxygen ratio (C/O) of (8:1) was chosen for this
study because conditions that tend more towards anaerobic
seem to be more suitable for the ODH of medium-chain paraf-
fins.[49] The under study spinel based catalysts showed relative-
ly good conversions under the chosen operating conditions
(Figure 7). At 4508C and 5008C, the activities of these catalysts
The selectivity towards COx seems to exactly follow the
order of Nb substitution in the Oh and Td sites. The two cata-
lysts with similar Ni[Oh]/Ni(Td) ratios, SP and SP-0.10Nb, showed
very similar selectivity to COx, whereas the catalyst with low
nickel occupation in both sites (SP-0.06Nb) and low oxygen va-
cancies as shown by XPS data seems to accelerate the COx for-
mation (Figure 8). However, the low concentration of Nb in the
Oh lattice of SP-0.02Nb seems to be beneficial for minimizing
selectivity to COx and clearly shows an enhanced ODH activity
over the catalysts with the highest concentration of defect
sites.
Figure 7. Catalytic conversion of n-octane over SP-xNb (x: 0, 0.02, 0.06 and
0.10) catalysts. (Reaction conditions: C/O=8:1, GHSV=12000 hÀ1, concen-
tration of n-octane in the feed=11%).
The selectivity to aromatic products could provide evidence
for the reaction pathways in this system. Ethylbenzene and o-
xylene are the results of further dehydro-aromatization of dif-
ferent octadienes over the active sites of the catalysts through
1,6 and 2,7 cyclisation.[50] In addition, other isomers of xylene
cannot form under the chosen reaction conditions in this
study because those require branched octadienes, which form
over acidic sites and under a high H2 pressure.[50] Styrene nor-
mally forms from the ODH of ethylbenzene. It seems high tem-
perature (5508C) is required to form o-xylene over all catalysts
(Figure 9). Ethylbenzene seems to be the dominant aromatic
over SP, whereas catalysts containing Nb seem to give styrene
as the dominant aromatic over all temperature ranges
(Figure 9). However, a volcano type behaviour was observed in
selectivity to ethylbenzene over all catalysts at the different
temperatures, suggesting the participation of non-oxidative
dehydrogenation (DH) (to form ethylbenzene from octene)
over these catalysts, which was in agreement with an increase
in the detection of hydrogen (data available in supporting in-
formation). The breakdown of selectivity to different octene
isomers, CO and CO2 also can be found in the supporting infor-
mation. In addition, a comparison of the calculated quantity of
seem to decrease with an increase in the amount of niobium
as a dopant. Indeed, the catalysts containing the highest con-
centration of Nb atoms showed the lowest activity at 5508C.
NiNb2O6 seems to be an inert phase (from the redox aspect) in
ODH under the chosen reaction conditions. Also, the detected
NiO phase in the SP catalyst (based on the first H2-TPR cycle)
in Figure S1, which showed similar catalytic activity as the
other catalysts, thus does not appear to contribute to the cata-
lytic process over SP, possibly owing to its fast deactivation
caused by hard coke deposition, which will be discussed
below with the TGA results.
In addition, the presence of this hard coke likely resulted in
the observation of a plateau for conversion over SP if the reac-
tion temperature was increased from 500 to 5508C.
Selectivity to different products
The selectivity towards octene isomers increased with increas-
ing Nb doping, at all temperatures (Figure 8). However, the se-
lectivity to octene over SP-0.02 Nb was higher than over SP-
0.06Nb. Elevating the reaction temperature seems to have an
&
ChemCatChem 2018, 10, 1 – 12
www.chemcatchem.org
6
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Figure 8. Catalytic selectivity to different products over SP-xNb (x: 0, 0.02, 0.06 and 0.10). (Reaction conditions: C/O=8:1, GHSV=12000 hÀ1, concentration of
n-octane in the feed=11%).
TGA analysis of the coked catalysts
Slight coke formation in the unsteady state was expected be-
cause the reactions were run under somewhat anaerobic con-
ditions (C/O=8:1). To determine the catalyst stability and to
estimate the possible coking, all catalysts were subjected to a
time-on-stream analysis at 5008C. After reaching steady state
(no change in the activity and selectivity observed), the cata-
lysts were run for a further 48 h, showing the stability of these
catalysts in time on line applications. The reactor was switched
off after 48 h and the catalysts were cooled down under nitro-
gen and subjected to coke analysis.
The TGA data showed mass losses of 4%, 2%, 2% and 1.5%
for the used SP, SP-0.02Nb, SP-0.06Nb and SP-0.10Nb catalysts,
respectively (Figure S6). The data in Figure 10 shows the differ-
Figure 9. Selectivity towards different aromatic products; a) SP, b) SP-
ent types of coke formed on the surface of these catalysts. SP
0.02Nb, c) SP-0.06Nb and d) SP-0.10Nb. (Reaction conditions: C/O=8:1,
GHSV=12000 hÀ1, concentration of n-octane in the feed=11%).
showed two different types of coke that could be removed at
4628C (soft coke) and 5968C (hard coke), respectively
(Figure 10). However, the small peak attributed to hard coke
for this catalyst seems to be for the coke formed on the avail-
able NiO phase detected by XPS and H2-TPR (Figure 6 and Fig-
ure S1) in this catalyst because this TGA peak was absent for
the other catalysts (Figure 10). This type of coke possibly deac-
tivated the available NiO/Al2O3 detected for the SP catalyst be-
H2 produced in the product formation and detected H2 shows
(Figure S5) that ODH is dominant over all catalysts and Nb in-
troduction further suppresses the already minor dehydrogena-
tion pathway over SP and promotes the ODH pathway.
ChemCatChem 2018, 10, 1 – 12
www.chemcatchem.org
7
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
Figure 11. XP spectra of the used SP-0.02Nb. Deconvolutions of the core-
level spectra Ni2p, O1s are provided. The used catalysts were obtained after
48 h time-on-stream at 5008C.
oxygen in NiO (Table 4). The other two oxygen peaks are the
same as for the fresh catalyst. Also, no change was observed in
the binding energy of the surface niobium in the XPS of the
used SP-0.02Nb catalyst, which proves that this element is not
directly involved in the chemistry over the chosen catalytic
model reaction.
Figure 10. TGA analysis of the used SP-xNb (x: 0, 0.02, 0.06 and 0.10) under
flow of air. The used catalysts were obtained after 48 h time-on-stream at
5008C.
cause its combustion temperature is well above the operating
temperature (5008C) for the coking study. The peak corre-
sponding to the soft coke moved slightly to higher tempera-
ture with a decrease in peak area upon introduction of Nb in
the catalysts. This soft coke is common for all four catalysts,
whereas the quantity is different depending on the amount of
loaded Nb. This suggests that the caused physicochemical
changes in the used catalysts are caused by coking and not
vice versa. The observation of mainly soft coke over the used
catalysts suggests that in situ coke removal took place under
the steady-state conditions, which prevents the formation of
hard coke. In addition, other characterisation data of the used
catalysts showed a decrease in the crystallite sizes, increase in
surface areas and nickel migration between the two sub-latti-
ces of the cubic spinel catalysts during the catalytic testing.
Further characterisation was performed to discover the main
factors causing these changes during the catalytic applications.
Table 4. Surface compositions of the used SP-0.02Nb catalyst determined
by XPS. The used catalysts were obtained after 48 h time-on-stream at
5008C.
Ni2p3/2
O1s
O
Nb3d
(doublet)
Ni3+
Ni2+
NiO
O(ion)
MÀO
[Oct]
856.4
854.8
853.5
530.1 529.6 528.8
210.5–207.4
TEM imaging of the used catalyst
The TEM image of used SP-0.02Nb was compared to that of
the fresh one (Figure 12). This comparison shows the presence
of very small crystals (4–5 nm), which are absent in the fresh
catalyst. The d spacing in the observed small crystals of the
used catalyst (0.2094 nm) is slightly higher than for the fresh
catalyst (0.2013 nm), which confirms the presence of NiO[51] of
very small size in the used catalysts, as was also evidenced
using XPS.
In situ XRD analysis
The in situ XRD data for the fresh SP-0.02 Nb under reduction Discussion
and oxidation conditions from 100–6008C is available in Figur-
Physicochemical and catalytic properties of Nb-doped spinel
es S7 and S8. This data suggests no nickel migration or crystal-
lite size change owing to a temperature change under reduc-
ing or oxidising conditions and, therefore, some other factors
probably cause the observed differences between fresh and
used catalysts.
Nb has a similar atomic radius to Ni and Al, and was shown to
be to partially substituted into the lattice of NiAl2O4, using the
SGCS technique. The Nb increased the crystallite sizes of the
NiAl2O4-based catalysts and decreased their BET surface areas.
These simple physical changes seem to affect the catalytic
ODH performances and coke deposition over these catalysts.
These changes created the opportunity to investigate the
effect of coking over this family of catalysts, systematically.
Nb substitution also modified the structure of the NiAl2O4
spinel. XRD, UV–DRS and XPS data were in excellent agree-
ment to show these structural alterations. Nb prefers to
occupy the Oh sub-lattice of spinel if only nickel was replaced
by it. Replacing both Ni2+ and Al3+ with Nb5+ resulted in the
substitution of this element in both sub-lattices, depending on
the existing cationic vacancies in the spinel. Furthermore, and
XPS of the used catalyst
The XPS data of the used SP-0.02 Nb is shown in Figure 11.
The Ni2p data shows that three different types of nickel are
present on the surface of this catalyst. Ni3+ and nickel in octa-
hedral sites seem to be present as was seen for the fresh cata-
lyst, but there is a new peak at 853.5 eV that can be assigned
to bulk NiO.[46] Also, the detection of this newly formed NiO
was confirmed by the detection of new oxygen species on the
surface of this catalyst at 529.6 eV, which can be related to the
&
ChemCatChem 2018, 10, 1 – 12
www.chemcatchem.org
8
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Nb content of these catalysts. The TPR studies
showed that the reduction of nickel in NiAl2O4 is
only possible at high temperatures (>8008C),
whereas the coke analysis study was done on the
catalyst after reaction at 5008C. The detection of
NiO using HRTEM and XPS in the used catalyst indi-
cates that the coking was found to reduce the NiOx
in the lattice of NiAl2O3 at lower temperatures.
Coking also resulted in the reduction of crystallite
sizes of these catalysts and improve their surface
areas. In addition, coke formation added small
shoulders to each peak in the XRD, which implies
the formation of NiAl2O4 with altered characteristics,
likely small crystals of NiAl2O4 with a different lattice
strain. The used catalysts also have more Ni in Td
sites than in Oh sites, which is an indication of nickel
migration or removal from Oh environments during
the catalytic reaction. The lattice fringes of the small
observed crystals in the used SP-0.02 Nb proved the
formation of small crystals of NiO during the catalyt-
ic reactions.
Figure 12. Comparison of the a) used and b) fresh SP-0.02Nb. c) The lattice fringe (d
spacing) of the small crystal (which appeared after the catalytic testing) of the used SP-
0.02 Nb was measured using HRTEM and d) compared to the d spacing value obtained
from the PXRD pattern for used SP-0.02 Nb. The used catalysts were obtained after 48 h
time-on-stream at 5008C.
The in situ XRD data showed that thermal
changes under reducing or oxidising conditions do
not cause these changes. In addition, the XPS data
and HRTEM imaging indicated the presence of the
based on UV–DRS data, Nb doping in Oh sub-lattices facilitates
the migration of nickel from Oh to Td environments. It has
been shown that Nb5+ can be a high-valance dopant (HVD) or
NiO phase in the used SP-0.02Nb catalyst, which could not be
detected using XRD analysis. Therefore, the structural changes
of NiAl2O4 during the ODH of n-octane under relatively anaero-
bic conditions are as follows:
2À
low-valance dopant (LVD) in the form of Nb2O4 for the outer-
most lattice of NiO and enhance its ODH catalytic performance
in this way.[31] Interestingly, this study showed the high proba-
bility of the latter type of dopant existing if a low quantity of
Nb was used, because the XPS data showed more ionic
oxygen and better catalytic data for this catalyst. The possible
low concentration of cationic vacancies, as well as a high
degree of NiNb2O6 formation, seems to result in similar struc-
tures for SP and SP-0.10Nb, because the XRD, UV–DRS and
XPS data were similar for them (Scheme S1 in the Supporting
Information).
1. Coke formation over fresh NiAl2O4 catalysts with large crys-
tallite size (XRD, HRTEM).
2. Formation of small Nb-doped NiAl2O4 crystals owing to
coke deposition with different lattice strain.
3. Reduction of available vicinal nickel of the defect sites of
the Oh sub lattice on the surfaces of large, as well as newly
formed small crystals of Nb-doped spinel at low tempera-
ture (5008C) owing to coking.
4. Slow transformation of inverse to normal NiAl2O4 spinel and
formation of NiO under the ODH reaction conditions, upon
the consumption of nickel in Oh sites to form NiO and de-
fects.
The selectivity to COx (undesired product) was found to
strongly correlate to the structural profiles of these catalysts.
XPS and TPR data showed that the defect sites on the Oh, and
these defect sites are believed to be on the outermost surface
of spinel, were highest for SP-0.02Nb. SP-0.02Nb showed the
lowest selectivity to COx, whereas SP and SP-0.10Nb showed
similar selectivity to COx. Interestingly, SP-0.06Nb showed the
highest selectivity to COx and no defect sites were found for
this catalyst using XPS. In addition, the H2-TPR data showed
easier reduction of nickel in Oh than Td sites of spinel, because
nickel in Oh sites was mainly reduced if the second TPR was
performed for these catalysts.
Conclusions
Niobium seems to preferentially replace the nickel in the octa-
hedral sub-lattice of the spinel. This dopant also was able to
replace both Al3+ and Ni2+, possibly because of their similar
ionic radii. This occupancy resulted in modification of nickel
migration, electronic profile, redox properties and surface tex-
tures. The crystallite sizes and degree of inversion of the spinel
were reduced if the used catalysts were compared to the as-
prepared catalysts, indicating that Ni2+ in Oh sites is responsi-
ble for the observed oxidative dehydrogenation (ODH) per-
formance. The defect sites in the octahedral environment of
these catalysts were found to have a positive influence in re-
Effect of coking on Nb-doped NiAl2O4 spinel
Coking was found to influence both physical and chemical
properties of these materials. According to the TGA data,
coking of NiAl2O4 was found to decrease with the increase of
ChemCatChem 2018, 10, 1 – 12
www.chemcatchem.org
9
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
ducing COx formation during ODH of n-octane. The slight coke
deposition resulted in the appearance of in situ formed NiAl2O4
with different lattice strain than the fresh catalysts and possi-
ble migration of nickel from octahedral to tetrahedral sites of
the spinel structure. Coke deposition was also found to facili-
tate the reduction of nickel in octahedral sites and its slow re-
lease into the catalytic system as NiO under the ODH condi-
tions, which may result in the formation of highly defective Oh
sites for these spinel-based catalysts.
Oxidative dehydrogenation of n-octane
A continuous-flow fixed-bed reactor (in vertical flow mode) was
used for catalytic testing. The catalyst particles were loaded in the
middle of the isothermal zone in a stainless-steel tube (10 mm ID
and 200 mm length). The voids were filled using carborundum (40
gritt, Polychem) to minimize the contribution of homogeneous gas
phase reactions.[33] Fresh catalyst, 0.5 mL (ꢀ0.45 g) with a particle
size of 300–600 mm diluted with the same volume of carborundum
(to eliminate mass-transfer limitations), was used for each set of
data. All reactions were performed in the range of 450–5508C at
508C intervals. The molar ratio of carbon to oxygen (C/O) was set
This study shows the high potential of the binary spinel-
based materials with tuneable physicochemical properties and
dynamic nature (depending on the reaction conditions), which
can be used as a strong tool in heterogeneous catalysis and
other applications.
at 8:1 with a gas hourly space velocity (GHSV) of 12000 hÀ1 Air
.
was used as a source of oxygen, nitrogen as the inert “make-up”
gas. The flow rates of n-octane, air and nitrogen were adjusted to
set the GHSV. The concentration of n-octane in the reactant steam
was 11%. The flow rate of nitrogen and air were controlled using
two separate mass flow controllers (Bronkhorst) and n-octane was
pumped (series II HPLC pump) into a heated reactor line (1408C)
to ensure that n-octane is in the gas phase.
Experimental Section
Materials
All products were analysed offline by gas chromatography (GC). H2,
CO and CO2 were analysed using a PerkinElmer Clarus 400 GC
equipped with a TCD detector and PLOT 1010 column with argon
as a carrier gas. Unreacted oxygen was quantified using a Perki-
nElmer Clarus 500 GC equipped with a TCD and PLOT 5A with
helium as a carrier gas. The light products in the gas phase were
analysed using a Shimadzu 2121 GC with FID detector and PONA
capillary column. H2 and N2 were used as carrier gas in this GC. The
collected liquid products in the catch-pot were analysed by the
aforementioned Shimadzu GC. All reported reactions have 100Æ
1% carbon balances and all reported results are the average of at
least two runs under steady state conditions.
The metal precursors used in the synthesis, Ni(NO3)2·6H2O (ACS re-
agent grade), Al(NO3)3·9H2O (ACS reagent grade) and
C4H4NNbO9·xH2O (99.99% trace metals basis), were used as re-
ceived from Sigma–Aldrich. Oxalyldihydrazine was freshly synthes-
ised using a reported method.[52] De-ionized water was used
throughout the synthesis. n-Octane (>98%) was purchased from
Alfa Aesar. Synthetic air (UHP) and nitrogen (UHP) were supplied
from zero air and N2 generators (Peak Scientific), respectively, for
catalytic testing. Hydrogen (base line), Argon (base line) and
Helium (base line) were purchased from Afrox for GC analyses in
this study. 10% H2/Ar and 10% O2/Ar (Afrox) were used in the
TPR–TPO–TPR analysis.
Physicochemical characterization
Fresh catalysts were characterized as they were synthesised with-
out any special pre-treatment. Powder X-Ray diffraction (Bruker D8
Advance) with a copper radiation source (l=1.5406 nm) was used
to analyse the phase composition of the synthesised materials. X-
ray photoelectron spectroscopy measurements were performed on
a Thermo Ka XPS equipped with a monochromatic small-spot X-
ray source using an aluminum anode AlKa (hn=1486.6 eV). The
background pressure was 4.9ꢁ10À8 bar, and 4ꢁ10À7 bar argon was
used during measurement to prevent sample charging. Binding en-
ergies were referenced to the sample stage, which contains built in
calibration standards of copper, silver, and gold. The bulk composi-
tion of the synthesised spinel was confirmed with ICP–OES using a
PerkinElmer Precisely Optima 5300DV, after the sample was digest-
ed in H2SO4 (98%, Merck). BET surface area measurements were
done using a Micromeritics Tristar II. Samples were degassed under
a flow of N2 at 2008C with a Micromeritics flow prep 060 overnight
prior to each BET analysis. UV-DRS spectra were collected using a
PerkinElmer (Lambda 35, UV/VIS Spectrometer) that was equipped
with a Labsphere reflectance spectroscopy accessory. Thermogravi-
metric analysis was done using a TA instrument (SDT Q600) under
a positive flow of air. Temperature-programmed reduction/oxida-
tion/reduction was performed using a Micromeritics 2920 Autoch-
em II analyser using methods reported elsewhere.[53] Catalyst mor-
phology was viewed using a Zeiss Ultra plus Scanning Electron Mi-
croscope (SEM). The bulk structures of the catalysts were viewed
using a Jeol JEM-1010 Transmission Electron Microscope. A Jeol
JEM-2100 High Resolution-Transmission Electron Microscope (HR-
TEM) was used in dark field-scanning transmission electron micro-
Sol–gel auto combustion synthesis (SGCS) of nickel alumi-
nate catalysts
The stoichiometric amounts of each precursor (including fuel) were
weighed and mixed in water. The solution was stirred at 808C on a
heater stirrer until all starting materials were fully dissolved. The
mixture was then dehydrated, which resulted in the formation of a
thick gel. The formed gel was put in a muffle furnace (set at
4008C) under static air for 30 min for ignition, followed by a tem-
perature ramp to 5008C and this temperature was held for one
extra hour. The resultant material was crushed and calcined at
10008C under a flow of air for 4.5 h. To insure good homogemicity
of the catalyst and to prevent the formation of NiO, the sample
was crushed every 1.5 h with a mortal and pestel for 10 min during
the calcination at 10008C (the total of 4.5 h includes the crushing
time). The final colour of the catalyst was sky blue.
As the aim of this study was the synthesis of NiAl2O4 spinel with
Nb as a dopant, the atomic ratio between Al and Ni was kept at
approximately 2:1. Nb substitution (of Al or Ni) was calculated
based on the charge balance and the final molecular formula of
the spinel (AB2O4). Therefore, the four catalysts for this study
are NiAl2O4Æd
,
Ni0.98Al2Nb0.02O4Æd
,
Ni0.97Al1.97Nb0.06O4Æd and
Ni0.95Al1.95Nb0.1O4Æd, labelled as SP, SP-0.02Nb, SP-0.06Nb and SP-
0.10Nb throughout this study.
&
ChemCatChem 2018, 10, 1 – 12
www.chemcatchem.org
10
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
[22] E. W. McFarland, H. Metiu, Chem. Rev. 2013, 113, 4391–4427.
[23] C. O. Augustin, K. Hema, L. J. Berchmans, R. K. Selvan, R. Saraswathi,
Phys. Status Solidi A 2005, 202, 1017–1024.
[24] J. J. H. B. Sattler, J. Ruiz-Martinez, E. Santillan-Jimenez, B. M. Weckhuy-
sen, Chem. Rev. 2014, 114, 10613–10653.
scope (DFSTEM) mode for elemental distribution analysis (map-
ping). The used catalysts, after the ODH of n-octane reaction at
5008C after 48 h of time-on-stream, were cooled under nitrogen
and then characterised.
[25] D. Teschner, E. Vass, M. Hꢅvecker, S. Zafeiratos, P. Schnçrch, H. Sauer, A.
Knop-Gericke, R. Schlçgl, M. Chamam, A. Wootsch, A. S. Canning, J. J.
Gamman, S. D. Jackson, J. McGregor, L. F. Gladden, J. Catal. 2006, 242,
26–37.
Acknowledgements
We thank SASOL and the National Research Foundation, South
Africa (NRF) for their financial supports. We also thank the Elec-
tron Microscopy Unit at the University of KwaZulu-Natal (West-
ville campus). Furthermore, M. D. Farahani thanks UKZN for a
PhD bursary.
[26] J. McGregor, Z. Huang, E. P. J. Parrott, J. A. Zeitler, K. L. Nguyen, J. M.
Rawson, A. Carley, T. W. Hansen, J.-P. Tessonnier, D. S. Su, D. Teschner,
E. M. Vass, A. Knop-Gericke, R. Schlçgl, L. F. Gladden, J. Catal. 2010, 269,
329–339.
[27] S. G. Sanz, L. McMillan, J. McGregor, J. A. Zeitler, N. Al-Yassir, S. Al-Khat-
taf, L. F. Gladden, Catal. Sci. Technol. 2015, 5, 3782–3797.
[28] C. H. Collett, J. McGregor, Catal. Sci. Technol. 2016, 6, 363–378.
[29] E. Heracleous, A. A. Lemonidou, J. Catal. 2006, 237, 162–174.
[30] E. Heracleous, A. A. Lemonidou, J. Catal. 2010, 270, 67–75.
[31] X. Sun, B. Li, H. Metiu, J. Phys. Chem. C 2013, 117, 23597–23608.
[32] I. Popescu, Z. Skoufa, E. Heracleous, A. Lemonidou, I.-C. Marcu, Phys.
Chem. Chem. Phys. 2015, 17, 8138–8147.
Conflict of interest
The authors declare no conflict of interest.
[33] H. B. Friedrich, A. S. Mahomed, Appl. Catal. A 2008, 347, 11–22.
[34] M. I. Fadlalla, H. B. Friedrich, Catal. Sci. Technol. 2014, 4, 4378–4385.
[35] S. Lei, C. Wang, D. Guo, X. Gao, D. Cheng, J. Zhou, B. Cheng, Y. Xiao,
RSC Adv. 2014, 4, 52740–52748.
Keywords: carbon · doping · niobium · oxidation · spinel
phases
[36] B. Savova, S. Loridant, D. Filkova, J. M. M. Millet, Appl. Catal. A 2010,
390, 148–157.
[37] B. P. Ladgaonkar, A. S. Vaingankar, Mater. Chem. Phys. 1998, 56, 280–
[1] R. Collongues, D. Gourier, A. Kahn-Harari, A. M. Lejus, J. Thery, D. Vivien,
Annu. Rev. Mater. Sci. 1990, 20, 51–82.
[2] W. Chu, W. Yang, L. Lin, Appl. Catal. A 2002, 235, 39–45.
[3] T. H. Gardner, J. J. Spivey, E. L. Kugler, A. Campos, J. C. Hissam, A. D. Roy,
J. Phys. Chem. C 2010, 114, 7888–7894.
283.
ˇ
ˇ
ˇ
[38] P. Priecel, D. Kubicka, L. Capek, Z. Bastl, P. Rysꢃnek, Appl. Catal. A 2011,
397, 127–137.
[4] B. Dou, B. Jiang, Y. Song, C. Zhang, C. Wang, H. Chen, B. Du, Y. Xu, Fuel
2016, 166, 340–346.
[39] J. F. Marco, J. R. Gancedo, M. Gracia, J. L. Gautier, E. Rꢆos, F. J. Berry, J.
Solid State Chem. 2000, 153, 74–81.
[5] N. Salhi, C. Petit, A. Kiennemann, Stud. Surf. Sci. Catal. 2008, 174, 1335–
1338.
[40] H. Ojagh, D. Creaser, S. Tamm, C. Hu, L. Olsson, Ind. Eng. Chem. Res.
2015, 54, 11511–11524.
[6] Y. S. Han, J. B. Li, X. S. Ning, B. Chi, J. Am. Ceram. Soc. 2004, 87, 1347–
1349.
[41] J. Li, S. Xiong, Y. Liu, Z. Ju, Y. Qian, ACS Appl. Mater. Interfaces 2013, 5,
981–988.
[7] K. Zhang, G. Zhou, J. Li, T. Cheng, Catal. Commun. 2009, 10, 1816–1820.
[8] M. Machida, K. Eguchi, H. Arai, J. Catal. 1989, 120, 377–386.
[9] J. L. Rogers, M. C. Mangarella, A. D. D’Amico, J. R. Gallagher, M. R.
Dutzer, E. Stavitski, J. T. Miller, C. Sievers, ACS Catal. 2016, 6, 5873–5886.
[10] J. Deng, M. Cai, W. Sun, X. Liao, W. Chu, X. S. Zhao, ChemSusChem 2013,
6, 2061–2065.
[42] N. ꢇzer, T. Barreto, T. Bꢈyꢈklimanli, C. M. Lampert, Sol. Energy Mater. Sol.
Cells 1995, 36, 433–443.
ˇ
[43] I. Milosev, T. Kosec, H. H. Strehblow, Electrochim. Acta 2008, 53, 3547–
3558.
[44] C. Jimꢂnez-Gonzꢃlez, Z. Boukha, B. de Rivas, J. R. Gonzꢃlez-Velasco, J. I.
Gutiꢂrrez-Ortiz, R. Lꢄpez-Fonseca, Energy Fuels 2014, 28, 7109–7121.
[45] L. Zhou, Y. Guo, J. M. Basset, H. Kameyama, Chem. Commun. 2015, 51,
12044–12047.
[11] N. F. P. Ribeiro, R. C. R. Neto, S. F. Moya, M. M. V. M. Souza, M. Schmal, Int.
J. Hydrogen Energy 2010, 35, 11725–11732.
[12] M. D. Farahani, J. Valand, A. S. Mahomed, H. B. Friedrich, Catal. Lett.
2016, 146, 2441–2449.
[46] S. R. Kirumakki, B. G. Shpeizer, G. V. Sagar, K. V. R. Chary, A. Clearfield, J.
Catal. 2006, 242, 319–331.
[13] Y. Kathiraser, W. Thitsartarn, K. Sutthiumporn, S. Kawi, J. Phys. Chem. C
2013, 117, 8120–8130.
[47] J. P. Jacobs, A. Maltha, J. G. H. Reintjes, J. Drimal, V. Ponec, H. H. Bron-
gersma, J. Catal. 1994, 147, 294–300.
[14] I. E. Achouri, N. Abatzoglou, C. Fauteux-Lefebvre, N. Braidy, Catal. Today
2013, 207, 13–20.
[15] C. Fauteux-Lefebvre, N. Abatzoglou, N. Braidy, I. E. Achouri, J. Power
Sources 2011, 196, 7673–7680.
[16] D. Li, M. Koike, L. Wang, Y. Nakagawa, Y. Xu, K. Tomishige, ChemSusChem
2014, 7, 510–522.
[48] C. Li, Y.-W. Chen, Thermochim. Acta 1995, 256, 457–465.
[49] J. Zhang, X. Liu, R. Blume, A. Zhang, R. Schlçgl, D. S. Su, Science 2008,
322, 73–77.
[50] E. A. Elkhalifa, H. B. Friedrich, J. Mol. Catal. A 2014, 392, 22–30.
[51] Y. Dong, Y. Chen, P. Jiang, G. Wang, X. Wu, R. Wu, RSC Adv. 2016, 6,
7465–7473.
[17] T. H. Gardner, D. Shekhawat, D. A. Berry, M. W. Smith, M. Salazar, E. L.
Kugler, Appl. Catal. A 2007, 323, 1–8.
[52] A. Koch, A. Phukan, O. B. Chanu, A. Kumar, R. A. Lal, J. Mol. Struct. 2014,
1060, 119–130.
[18] Z. Boukha, C. Jimꢂnez-Gonzꢃlez, B. de Rivas, J. R. Gonzꢃlez-Velasco, J. I.
Gutiꢂrrez-Ortiz, R. Lꢄpez-Fonseca, Appl. Catal. B 2014, 158–159, 190–
201.
[53] V. D. B. C. Dasireddy, H. B. Friedrich, S. Singh, Appl. Catal. A 2013, 467,
142–153.
[19] H. S. C. O’Neill, W. A. Dollase, C. R. Ross, Phys. Chem. Miner. 1991, 18,
302–319.
[20] C. O. Arean, M. L. R. Martinez, A. M. Arjona, Mater. Chem. Phys. 1983, 8,
443–450.
[21] E. C. O’Quinn, J. Shamblin, B. Perlov, R. C. Ewing, J. Neuefeind, M. Fey-
genson, I. Gussev, M. Lang, J. Am. Chem. Soc. 2017, 139, 10395–10402.
Manuscript received: December 5, 2017
Revised manuscript received: January 18, 2018
Version of record online: && &&, 0000
ChemCatChem 2018, 10, 1 – 12
www.chemcatchem.org
11
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

FULL PAPERS
M. D. Farahani, V. D. B. C. Dasireddy,
H. B. Friedrich*
&& – &&
Oxidative Dehydrogenation of n-
Octane over Niobium-Doped NiAl2O4:
An Example of Beneficial Coking in
Catalysis over Spinel
Nickel wanders with coke: Nb-doped
NiAl2O4 catalysts are synthesized and
used in the oxidative dehydrogenation
(ODH) reaction of n-octane. Nickel mi-
gration and varying concentration of
defect sites are established upon the
substitution of Nb and coke deposition.
&
ChemCatChem 2018, 10, 1 – 12
www.chemcatchem.org
12
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!