Unraveling the contribution of structural phases in Ni-Nb-O mixed oxides in ethane oxidative dehydrogenation

Article abstract of DOI:10.1016/j.cattod.2011.12.022

The role of each structural phase present in highly active and selective Ni_0.85Nb_0.15 catalyst in ethane oxidative dehydrogenation is investigated. Based on previous studies, Ni_0.85Nb_0.15 consists of a Ni-Nb sol

Full text of DOI:10.1016/j.cattod.2011.12.022

Catalysis Today 192 (2012) 169–176
Contents lists available at SciVerse ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Unraveling the contribution of structural phases in Ni–Nb–O mixed oxides in
ethane oxidative dehydrogenation
a
b,∗
a,b,∗∗
Z. Skoufa , E. Heracleous , A.A. Lemonidou
a
Department of Chemical Engineering, Aristotle University of Thessaloniki, University Campus, 54124 Thessaloniki, Greece
Chemical Process Engineering Research Institute (CPERI), Centre for Research and Technology Hellas (CERTH), 6th km Charilaou – Thermi Road, P.O. Box 361,
7001 Thessaloniki, Greece
b
5
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 19 September 2011
Received in revised form
The role of each structural phase present in highly active and selective Ni0.85Nb0.15 catalyst in ethane
oxidative dehydrogenation is investigated. Based on previous studies, Ni0.85Nb0.15 consists of a Ni–Nb
solid solution and a Nb-rich amorphous phase. A series of Ni-based mixed oxides with low Nb content
simulating the Ni–Nb solid solution and an amorphous Nb-rich sample were synthesized and studied
separately. The catalytic tests demonstrated that the introduction of even a tiny amount of Nb (atomic
ratio: Nb/(Ni + Nb) = 0.01) leads to a drastic increase in ethylene selectivity (150%) compared to pure NiO.
Selectivity gradually increases with increasing Nb loading up to 15%, in parallel with the decrease in
oxygen desorbed as measured by O2-TPD. Based on the results of combined catalytic testing and physic-
ochemical characterization, Ni–Nb solid solution can be identified as the key component for the high
ethylene selectivity of the Ni0.85Nb0.15 catalyst, as the insertion of even a small amount of Nb drastically
decreases the electrophilic oxygen species responsible for the total oxidation reactions. Co-existence of
the Ni–Nb solid solution with the amorphous Nb-rich phase does not influence significantly the catalytic
properties of the material.
1
3 December 2011
Accepted 14 December 2011
Available online 17 January 2012
Keywords:
Ethane oxidative dehydrogenation
Nickel-based mixed oxides
Solid solution
Temperature programmed oxygen
desorption
CO-temperature programmed reduction
©
2011 Elsevier B.V. All rights reserved.
1
. Introduction
Its major advantages compared to currently applied technology
are high energy efficiency due to comparatively low operation
temperature and exothermicity associated with the presence of
oxygen [2,3]. The successful industrial implementation of such a
process is however linked to the development of a catalytic sys-
tem able to selectively convert ethane to ethylene and at the
same time prevent ethane primary and ethene secondary oxi-
dation to carbon oxides. The key issue therefore is improving
catalyst selectivity and productivity leading to an economically
feasible and competitive with currently used technologies process
[7].
Ethylene constitutes the fundamental building unit of world-
wide petrochemical industry with numerous applications in a
spectrum of end use areas. Although Western European produc-
tion and demand dropped during the last 2 years, world ethylene
demand is expected to cross 160 MMT by 2015 [1]. Ethylene is
currently produced via steam cracking of naphtha and ethane
feedstocks, a process considered as the most energy consum-
ing process in the chemical industry; energy cost accounts for
approximately 70% of production costs in typical ethane- or naph-
tha based olefin plants [2]. In this context, from both economic
and environmental perspectives – the latter associated with COx
and NOx emissions, the production of ethylene through alterna-
tive routes constitutes one the most studied topics in the recent
decades.
Generally, niobium has been reported to have beneficial effect
when used as a dopant in selective oxidation catalysts [8]. In the
case of ethane ODH in particular, Nb-doped NiO oxides constitute
low temperature, highly active and selective catalysts, exhibiting
one of the best catalytic performances [9,10] compared to most
catalysts reported in literature [3]. The Ni–Nb–O system has also
been investigated recently by Ba n˜ ares and co-workers for the
direct ammoxidation of ethane to acetonitrile with extremely
promising results [11]. During the last years, our group has
extensively studied the structure, physicochemical characteristics,
ethane ODH catalytic performance and kinetics and mechanis-
tic aspects of Ni–Nb–O oxides [9,10,12]. The catalyst with the
optimum formulation was found to be Ni0.85Nb0.15, exhibiting an
Catalytic oxidative dehydrogenation of ethane is an attractive
alternative route for the production of ethylene, exhibiting par-
ticular economic, technological and environmental interest [3–6].
∗
Corresponding author. Tel.: +30 2310 498345; fax: +30 2310 498380.
Corresponding author. Tel.: +30 2310 996273; fax: +30 2310 996184.
E-mail addresses: eheracle@cperi.certh.gr (E. Heracleous),
∗
∗
◦
alemonidou@cheng.auth.gr (A.A. Lemonidou).
ethylene yield of 46% at 400 C. Recently, the research group of
0
920-5861/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.cattod.2011.12.022

1
70
Z. Skoufa et al. / Catalysis Today 192 (2012) 169–176
2. Experimental part
2.1. Catalyst preparation
A series of bulk mixed Ni–Nb–O oxides, with Nb/(Ni + Nb) ratios
equal to 0.01, 0.03, 0.05 and a Nb-rich phase with Nb/(Ni + Nb)
atomic ratio equal to 0.67 were synthesized by the evaporation
method. The oxides were prepared by mixing aqueous solutions of
nickel nitrate hexahydrate (Merck) and ammonium niobate oxalate
(
Aldrich) used as precursors. In all cases, the solution contain-
◦
ing both metals was heated at 70 C under continuous stirring
for 1 h to ensure complete dissolution and good mixing of the
starting compounds. The solvent was then removed by evapora-
tion under reduced pressure, and the resulting solids were dried
Fig. 1. HRTEM images from the Ni0.85Nb0.15 catalyst, showing the two distinct mor-
phologies [12].
◦
◦
overnight at 120 C and calcined in synthetic air at 450 C for
h.
Pure NiO and Nb O5, obtained from the decomposition of
5
2
Millet also confirmed the enhanced catalytic performance of these
materials [13]. According to detailed structural and isotopic kinetic
measurements conducted by our group, due to the favorable ionic
radii, valence and electron donor properties of niobium, part of Nb
cations fill the cationic vacancies and/or substitute nickel atoms
in the NiO lattice, forming a Ni–Nb solid solution [9,10]. This
substitution process is most likely responsible for the reduction of
the non-stoichiometry and the cationic defects and, consequently
for the decrease of the unselective oxygen species, leading to
enhanced ethane ODH activity. The systematic study of a series of
NiO-based mixed oxides with other than Nb non-reducible doping
metals from each periodic group (valences ranging from +1 to +5)
nickel nitrate hexahydrate (Ni(NO ) ·6H O, Merck) and niobic acid
3
2
2
◦
(
Nb O5·nH O, CBMM) respectively at 450 C for 5 h in synthetic air,
2
2
as well as Ni0 Nb
, Ni
Nb
and Ni
Nb
mixed oxides
.90
0.10
0.85
0.15
0.80
0.20
with Nb/(Ni + Nb) atomic ratios of 0.10, 0.15 and 0.20 reported in
9] are further investigated in this work.
[
2.2. Catalyst characterization
Surface areas of the samples were determined by N2 adsorp-
tion at 77 K, using the multipoint BET analysis method, with
an Autosorb-1 Quantachrome flow apparatus. Prior to the mea-
◦
surements, the samples were dehydrated in vacuum at 250 C
+2
with similar ionic radii to Ni allowed generalization of the above
conclusions [14]. Moreover, interesting results have been reported
for the addition of tungsten [15] and cerium [16] in NiO.
overnight.
X-Ray Diffraction (XRD) patterns were obtained using a Siemens
D500 diffractometer, employing Cu K␣ radiation. Analysis of the
XRD patterns and application of the Scherrer equation allowed the
calculation of the average particle size of the samples. The lat-
tice constant, ␣, of the cubic nickel oxide crystalline phase in the
synthesized mixed oxides was also determined from the obtained
diffractograms with a Lorentzian peak analysis generating an
error always below ± 0.0001. Rietveld analysis [23] was employed
for the determination of phase content using NaCl as internal
standard.
In-depth structural characterization of the optimum formu-
lation Ni0.85Nb0.15 catalyst by HRTEM led to the identification
of the nanostructural features formed upon doping of NiO with
Nb cations at a Nb/(Ni + Nb) atomic ratio of 0.15. The formation
of two distinct phases, clearly shown in Fig. 1, was evidenced:
a NiO-rich (EDS analysis: Nb/Ni ≤ 0.179 [12]), crystallized phase
with Nb cations incorporated in the NiO lattice, which retains
its original crystallographic structure, identified as Ni–Nb solid
solution and a highly distorted Nb-rich phase formed upon inser-
tion of Nb cations in areas of high density cationic deficiency of
the host NiO [12]. According to calculation of the d-spacing of
this phase, it is identified as a precursor phase for the forma-
The oxygen desorption properties of selected catalysts were
studied by O -TPD measurements. The catalyst sample (200 mg)
2
◦
was pretreated in a flow of He at 450 C for 0.5 h and cooled to
room temperature under the He flow. The system was subsequently
flushed with helium for 1 h and the temperature was raised to
tion of NiNb O , while EDS analysis provided a Nb/Ni ratio = 1.5–2
2
6
[
12]. Millet and co-workers confirmed the existence of Nb-doped
◦
◦
3
8
50 C at a heating rate of 15 C/min in He (50 cm /min). The reac-
tor exit was monitored on-line by a quadrupole mass analyzer
Omnistar, Balzers) and the desorbed oxygen was detected by fol-
NiO particles in the Ni0.85Nb0.15 catalyst, although according to
EDX analysis Nb content was limited close to 1% Nb/(Ni + Nb)
(
[
13].
It is of great interest to clarify the role of the different struc-
lowing the 32 (m/z) fragment.
Temperature programmed reduction (TPR) using carbon
monoxide as reducing agent was also employed. Prior to each run,
tural phases present in the Ni–Nb–O catalyst by determining the
contribution of each separately in the ethane oxidative dehydro-
genation reaction. In general, solid solutions are considered to play
a vital role in oxidation catalysis research [17], including oxida-
tive dehydrogenation reactions. This is confirmed by the numerous
publications on the activity of several solid solution systems (see
for example [18–22]).
In this work, the role of Ni–Nb solid solution is investigated via
the preparation of a series of NiO-based mixed oxides with low Nb
content. A Nb-rich mixed Ni–Nb–O oxide (Nb/(Ni + Nb) ratio = 0.67),
synthesized separately, consists the “model” for the study of the
amorphous Nb-rich phase present in the Ni0.85Nb0.15 catalyst. All
materials were characterized and tested in the ethane ODH reac-
tion, yielding important conclusions on the contribution of each
structural phase in the apparent catalytic performance of Ni–Nb–O
mixed oxide catalysts.
◦
the sample (50 mg) was pretreated under He flow at 450 C for 0.5 h
and cooled to room temperature. Next, the catalyst was heated
◦
◦
up to 830 C at a constant rate of 5 C/min in a 6% CO/He stream
(
3
50 cm /min). After each TPR-CO run, the sample was cooled to
room temperature under pure He flow and subsequently subjected
to temperature programmed oxidation (TPO) from room temper-
◦
ature to 830 C in a 6% O /He flow at a constant heating rate of
5
2
◦
C/min. The reactor exit was monitored on-line by a quadrupole
mass analyzer (Omnistar, Balzers) following (m/z) fragments 2 (H ),
2
1
8 (H O), 28 (CO), 32 (O ) and 44 (CO ).
2 2 2
An SDT Q600 (TA Instrument) thermal gravimetric analysis
TGA) instrument was employed for the characterization of the Nb-
(
rich sample. SDT Q600 works in conjunction with a controller and
associated software to make up a thermal analysis system. A small
quantity of the sample was placed in an aluminium sample cap and

Z. Skoufa et al. / Catalysis Today 192 (2012) 169–176
171
◦
the temperature was raised from room temperature to 800 C at a
rate of 10 C/min in synthetic air flow (100 cm /min).
◦
3
2.3. Reactivity studies
The catalytic performance of the samples was measured in a
fixed-bed quartz reactor. The catalyst particles were diluted with
equal amount of quartz particles of the same size to achieve isother-
mal operation. The temperature in the middle of the catalytic bed
was measured with a coaxial thermocouple. The samples were acti-
◦
vated in flowing oxygen at 450 C for 30 min. The composition of
the reaction mixture used was 10%C H /5%O /85%He.
2
6
2
The oxidative dehydrogenation of ethane was investigated in
◦
◦
the temperature range from 300 C to 425 C. For the determination
of the activity of the catalysts as a function of temperature, the W/F
3
ratio was kept constant at 0.24 g s/cm . In order to obtain different
◦
ethane conversion levels at constant reaction temperature (350 C),
the W/F ratio was varied from 0.03 to 1.35 g s/cm .
3
The reaction products were analyzed on line by a PerkinElmer
gas chromatograph equipped with a thermal conductivity detector
(
TCD). Two columns in a series-bypass configuration were used in
the analysis: a Porapak Q and a MS 5A. The main reaction products
were C H , CO and H O. The concentration of CO as well as that
2
4
2
2
of oxygenates was under detection limits. The ethane conversion
and the selectivity to the reaction products were calculated on a
carbon basis. Closure of the carbon mass-balance was better than
Fig. 2. X-ray diffraction patterns of the Ni–Nb–O catalysts and pure NiO (ꢀ: NiO
crystal structure, ◦: crystalline NiNb2O6, ꢀ: amorphous NiNb2O6).
±
1%. The contribution of gas-phase initiated reactions was tested
by conducting experiments using an empty-volume reactor. The
conversion of ethane at these experiments was lower than 1%, con-
firming that gas-phase reactions are negligible at the experimental
conditions used for the activity tests.
diffraction patterns of pure nickel oxide, Ni0.90Nb0.10, Ni0.85Nb0.15
and Ni0.80Nb0.20 samples are also included for comparison reasons.
Samples Ni0.99Nb0.01, Ni0.97Nb0.03 and Ni0.95Nb0.05, with Nb pro-
motion up to 0.05 relative to total metal content, exhibit only the
diffraction lines corresponding to a “NiO-like” crystal phase. The
◦
3
. Results
broad band centered at 2ꢁ 27 , attributed to the amorphous Nb-rich
phase [9,12], is not detected in any of the synthesized samples and
starts appearing in the materials with Nb loading equal and higher
than 0.10. Further increase of Nb content causes the crystallization
of this phase to crystalline NiNb O6 as observed in the Ni0.80Nb
3
.1. Catalyst characterization
The main physicochemical characteristics of the synthesized
2
0.20
◦
and reference Ni–Nb–O mixed oxides are tabulated in Table 1. BET
measurements point out to a considerable increase of NiO surface
area upon introduction of Nb, as was previously observed [9,13],
with the surface area gradually increasing with increasing Nb con-
tent up to 15 at% Nb metal. This phenomenon might be attributed
to the use of niobium oxalate as starting compound for Nb, since
the decomposition (during calcination) of precursors with organic
nature possibly yields a more porous structure. Another expla-
nation recently provided by Millet and co-workers [13] is that
the surface area enhancement can be attributed to the so-called
catalyst [9], which presents additional diffraction peaks at 2ꢁ 26 ,
◦
◦
◦
30 , 35 and 53 that correspond to NiNb O6 crystal structure.
2
Analysis of the XRD patterns allowed the calculation of the lattice
parameters of the nickel oxide-like phase in the synthesized mixed
oxides, presented in Table 1. In agreement with previous results
[9], in all Ni–Nb–O mixed oxides NiO lattice parameter decreases,
due to the lower lattice constant of Nb relative to Ni, indicating
that in all samples part of Nb inserts NiO crystal forming a solid
solution and thus leading to lattice contraction. Based on all the
above, it can thus be argued that since NiNb solid solution is the
predominant phase in these samples, catalysts with low Nb load-
ing that were synthesized in this work succeed in simulating the
pure Ni–Nb solid solution phase that is present in the Ni0.85Nb0.15
catalyst.
“
mutual protective effect” in co-precipitated binary oxide systems
[
24], according to which the presence of one oxide may prevent
crystallization of the other. A maximum surface area is recorded
for Ni0.85Nb0.15. The subsequent decrease in surface area for the
catalyst Ni0.80Nb0.20 was previously related to the formation of
In the Nb-rich oxide sample the intensity of the broad band at
◦
crystalline NiNb O6 as evidenced by XRD analysis results [9]. The
2ꢁ 27 increases, however the amorphous phase co-exists together
2
bulk Nb-rich phase prepared in this study to simulate the amor-
phous phase present in the Ni0.85Nb0.15 catalyst exhibits a rather
small surface area (13 m ), most probably due to the incomplete
decomposition of carbonaceous species originating from the nio-
bium precursor (see TGA results).
X-ray diffraction (XRD) was employed for investigating the crys-
talline phases formed in the calcined catalysts. The XRD patterns
were also used to estimate the average particle size of the crystal
phases in the synthesized materials using the Scherrer equation.
The results shown in Table 1 evidence that all Ni–Nb–O oxides
consist of “NiO-like” crystallites in the nanorange. Fig. 2 illus-
trates the diffractograms obtained for all the samples. The X-ray
with NiO-like crystallites as evidenced by the characteristic peaks.
According to Rietveld analysis, employed for the determination of
phase content, the amorphous phase accounts for approximately
90 wt% of the synthesized material, leading to the assumption that
indeed the prepared sample satisfactorily simulates the highly dis-
torted Nb-rich phase present in the Ni0.85Nb0.15 mixed oxide.
The oxygen desorption properties of all samples were inves-
tigated by temperature programmed experiments. Pure NiO,
calcined at low temperatures, is a well-known non-stoichiometric
oxide with cationic deficiency [25–27]. The corresponding oxygen
excess of NiO is illustrated by a strong oxygen release upon heat-
ing the catalyst under helium flow, with its maximum located at
2

1
72
Z. Skoufa et al. / Catalysis Today 192 (2012) 169–176
Table 1
Nomenclature, physicochemical characteristics and apparent activation energy in ethane ODH for Ni–Nb–O mixed oxides and NiO.
Catalyst
Nb/(Ni + Nb) atomic
ratio
Surface
area (m /g)
“NiO-like” phase average
crystal size (nm)
“NiO-like” phase
Desorbed O2
(mg/m )
Apparent activation
energy (kJ/mol)
2
´
˚
2
lattice constant (A)
Nominal
ICP
NiO
–
–
12
17
22
29
56
85
59
13
32
28
31
29
21
17
14
n.d.
4.1762
4.1760
4.1757
4.1757
4.1751
4.1725
4.1722
n.d
0.87
0.34
0.21
0.15
0.07
0.01
0.07
n.d.
90 ± 1
83 ± 1
88 ± 2
88 ± 2
84 ± 2
79 ± 1
80 ± 1
85 ± 2
Ni0.99Nb0.01
Ni0.97Nb0.03
Ni0.95Nb0.05
Ni0.90Nb0.10
Ni0.85Nb0.15
Ni0.80Nb0.20
Nb-rich amorphous
0.01
0.03
0.05
0.10
0.15
0.20
0.67
0.01
0.03
0.05
0.11
0.14
0.19
0.67
◦
∼
700 C, as reported previously [9]. The temperature programmed
During reduction of NiO in the presence of CO, CO2 is formed
through the following reaction:
oxygen desorption (O -TPD) profiles of all catalysts of this work
2
and reference catalysts are presented in Fig. 3. All Nb-doped NiO
materials present a gradual shift of the desorption maxima towards
lower temperatures (up to 90 C temperature shift) as Nb loading
0
CO + NiO → Ni + CO2
◦
CO2 can be also formed through the disproportionation of CO
increases from 0% to 20% atomic content, indicating a modification
of the lattice oxygen species. The overall amount of desorbed oxy-
gen, tabulated in Table 1, also reduces with increasing Nb content,
indicative of the over-stoichiometry reduction caused by Nb dop-
ing up to Nb/(Ni + Nb) molar ratio of 0.15. For higher Nb-loading
(
Boudouard reaction) that leads to carbon deposition on the catalyst
surface. Boudouard reaction is known to be catalyzed by metallic
Ni:
2
CO ꢁ CO + C
2
(
Ni0.80Nb0.20) the amount of desorbed oxygen slightly increases.
Water–gas shift (WGS) reaction due to the interaction of CO with
The effect of Nb in the O2 desorption properties of NiO is exten-
sively discussed in the Discussion section, in combination with the
catalytic performance of the materials.
adsorbed hydroxyl groups on the catalytic surface can contribute
to the formation of CO2 and H2:
CO + 2OH− → CO + H + O
2−
The application of the O -TPD technique on the Nb-rich sam-
2
(s)
2
2
(s)
◦
ple resulted in a large CO2 peak centered at 600 C, with no
CO-TPR profiles of all samples, including Nb O5 for compari-
2
detectable O desorption. The CO formation at such high tempera-
2
2
son, are presented in Fig. 5. The reaction onset for all materials
was recorded at 270 C, except for Nb O5 that presents negligible
tures could be attributed to carbonaceous deposits originating from
the oxalate Nb-precursor that survived calcination. The existence of
such remaining organic species was further confirmed by TGA anal-
ysis of Nb-rich sample under air flow. As shown in Fig. 4, a weight
◦
2
reactivity. The WGS reaction occurs at a very low extent over all
materials as in all cases traces of H2 were detected. The differ-
ence in the CO consumption and CO2 production profiles shown
in Fig. 5 indicates that apart from NiO reduction, Boudouard reac-
tion leading to the formation of coke also takes place. The extent
of coke formation increases with temperature due to the increas-
ing formation of metallic Ni (which catalyzes the reaction). Pure
◦
loss of 7% was recorded at around 600 C, in agreement with the
O -TPD results.
2
Although in temperature programmed reduction studies carbon
monoxide is used less frequently as reducing agent than hydrogen,
CO-TPR is one of the best probes to investigate reducibility. The
mild reductive nature of carbon monoxide compared to hydrogen
can be the key to identify between different oxidation states [28]
or between different oxygen species [29]. CO-TPR has also been
used for studying the redox properties of NiO samples of flower-like
morphology [30].
NiO exhibits a single peak of CO consumption/CO production cen-
2
◦
tered at ∼380 C, in good agreement with previously published
results [30,31]. The peak maximum shifts to lower temperature
◦
with increasing Nb content, reaching 330 C over Ni0.80Nb0.20 sam-
ple. Moreover, as niobium content increases, the formation of a high
Fig. 3. Temperature programmed oxygen desorption profiles of Ni–Nb–O catalysts
Fig. 4. Mass spectrum recorded for CO2 during O -TPD experiment (grey line) and
2
and NiO.
TGA analysis plot of the Nb-rich phase (black line).

Z. Skoufa et al. / Catalysis Today 192 (2012) 169–176
173
Fig. 5. CO-TPR profiles of Ni–Nb–O catalysts, NiO and Nb2O5; full line: CO2 produced,
dash line: CO consumed.
Fig. 7. Ethane conversion as
W/F = 0.24 g s/cm , C2H6/O2 = 2/1).
a
function of temperature (reaction conditions:
3
temperature shoulder is observed, which gradually grows into a
distinct second peak at molar ratio Nb/(Ni + Nb) ≥ 0.10. The maxi-
mum of this second high temperature peak can be detected around
The presence of Nb seems to weaken these bonds, as the peak max-
imum is shifted to lower temperatures with increasing Nb content,
indicating that the oxygen can be removed at lower temperatures
than in pure NiO. Taking a step further, it can be postulated that this
peak corresponds to the nickel rich Ni–Nb solid solution (“NiO-like”
◦
4
50 C.
Quantitative calculation of the extent of reduction of oxides
with TPR-CO usually proves to be a challenging task due to the
concurrent occurrence of the Boudouard reaction. Therefore, it is
first necessary to quantify separately coke formation on the sam-
ples. In order to do so, after the TPR-CO measurement, oxidation of
the samples under temperature programmed conditions was per-
formed. Oxidation under oxygen flow resulted in the consumption
of O2 for re-oxidation of metallic Ni to NiO and oxidation of coke,
phase), in which Ni
in close vicinity.
O Ni bonds prevail with the presence of Nb
In view of previous HR-TEM [12] and current XRD results that
point out to the formation of an amorphous Nb-rich phase at Nb-
loadings above Nb/(Ni + Nb) = 0.10, the high temperature peak can
be attributed to the presence of this phase. Apart from covering part
of NiO particles preventing their reduction and thus increasing the
reduction temperature of the solid as also proposed in [13], this Nb-
with subsequent formation of CO . The CO2 production profiles of
2
all samples are presented in Fig. 6. Quantification of the produced
CO2 allowed cross-checking of the CO-TPR results, in terms of car-
bon balance and reduction stoichiometry. The calculations showed
that the amount of CO used only for reduction of NiO is proportional
rich phase contains Ni
at higher temperature than Ni
O
Nb bonds that were found to be reduced
Ni bonds [9]. The TPR-CO profile
O
of the synthesized Nb-rich sample that consists largely of an amor-
phous Nb-rich phase confirms the above, as the low temperature
peak is significantly smaller.
to wt% NiO content in the samples, in good agreement with H -TPR
2
results [9].
The linear relationship established between NiO content and CO
consumption for reduction implies that oxygen is removed from
3.2. Catalytic performance in ethane ODH
Ni
O Ni and Ni O Nb bonds, since Nb O Nb bonds are irre-
ducible (see Fig. 5, Nb O5 profile). Based on the above, the main CO
consumption peak in the TRP-CO profiles at 330–380 C temper-
2
The activity of the Ni–Nb–O mixed oxide catalysts in the ethane
◦
oxidative dehydrogenation reaction was explored under steady
ature range can be attributed to the reduction of Ni
O
Ni bonds.
◦
state conditions between 300 and 450 C with a constant W/F
3
(
0.24 g s/cm ) and ethane-to-oxygen (2/1) ratio. It should be noted
that in all cases oxygen conversion did not exceed 90%. It is also
important to mention that ethylene and carbon dioxide were the
only products detected over all catalysts. Therefore, main reac-
tions taking place over nickel oxide-based catalysts are the primary
selective oxidative dehydrogenation of ethane to ethylene, the pri-
mary unselective total oxidation of ethane to CO and the secondary
2
over-oxidation of the produced olefin to CO2.
Ethane conversion is plotted as a function of temperature in
Fig. 7. All catalysts are active in ethane conversion at relatively low
◦
temperatures (<300 C). Upon insertion of Nb, the catalytic activity
seems to be favored, leading to overall increased ethane conversion
at a given temperature for all Nb-containing catalysts compared to
pure NiO. As discussed below, the enhanced weight activity is due
to the higher surface area that Nb induces. Samples Ni0.99Nb0.01
,
Ni0.97Nb0.03, Ni0.95Nb0.05 and Ni0.90Nb0.10 present similar conver-
sion values. Ni0.85Nb0.15 remains the most active catalyst, while
further increase in Nb loading leads to reduced conversion. The syn-
thesized Nb-rich amorphous phase presents satisfactory activity,
yet as low as that of NiO, due to its low surface area.
Fig. 6. Temperature programmed oxidation profiles of Ni–Nb–O catalysts and NiO.

1
74
Z. Skoufa et al. / Catalysis Today 192 (2012) 169–176
3
varying W/F from 0.03 to 1.35 g s/cm to attain different conversion
levels.
Ethylene selectivity as a function of ethane conversion is pre-
sented in Fig. 9. For all catalytic samples selectivity remained
relatively constant with increasing conversion for the conver-
sion range studied, indicating low reactivity towards ethylene
secondary reactions. Although pure NiO is an effective total oxi-
dation catalyst, demonstrating only 20% ethylene selectivity, even
a small amount of Nb equal to 1% atomic metal content leads to a
remarkable increase in selectivity up to ∼55%. The most impor-
tant observation is that selectivity increases progressively with
increasing Nb loading up to 15% (sample Ni0.85Nb0.15). However, as
previously reported [9], further increase of Nb-loading up to 40%
results in a small decrease in ethylene selectivity, attributed to the
formation of crystalline NiNb O6 and Nb O5. The synthesized Nb-
2
2
rich amorphous phase with a high Nb content (Nb/(Ni + Nb) = 0.67)
follows the trend of decreasing selectivity with increasing Nb con-
tent for high Nb-loadings [9], with a selectivity equal to 70%.
Fig. 8. Specific surface activity (SSAc) as a function of temperature (reaction condi-
3
tions: W/F = 0.24 g s/cm , C2H6/O2 = 2/1).
4. Discussion
To account for the different surface areas of the samples, the
The aim of the present work is to define the role of each struc-
activity of the catalysts in terms of specific surface activity (SSAc)
tural phase of the Ni0.85Nb0.15 mixed oxide in ethane oxidative
dehydrogenation reaction by a systematic study of the physico-
chemical and catalytic properties of the two constituents; Ni–Nb
solid solution and Nb-rich amorphous phase.
2
–
(
expressed as ␮mol C H6 consumed/m s – was also considered
2
Fig. 8). Based on this definition, pure NiO exhibits the highest sur-
face reactivity for ethane activation compared with the Ni–Nb–O
oxides. However, ethane activation on NiO is unselective, leading
to combustion. In the mixed oxides with low Nb-loading (1–20 at%)
surface activity decreases progressively with increasing Nb content.
Using only the experimental values for ethane conversion <10%
allowed the calculation of the apparent activation energy based
on the Arrhenius equation. As tabulated in Table 1, the apparent
activation energy for ethane consumption presents only small dif-
ferences among all mixed oxides. Therefore, the gradual decrease
in surface activity recorded for increasing Nb loading can be mainly
attributed to reduction in the number of surface sites that activate
ethane.
The characterization results obtained by XRD analysis indicate
that for Nb loading up to 5%, Ni–Nb–O mixed oxides exhibit only
the crystallographic structure of a NiO-like phase. Calculation of
this NiO-like phase lattice parameters points out to a relatively
small – yet systematic and therefore not negligible – reduction
+5
of the lattice constant with increasing Nb content. As Nb has
a smaller ionic radius than NiO (0.69 A and 0.64 A, respectively)
32,33], these findings indicate the formation of a solid solution in
˚
˚
[
samples Ni0.99Nb0.01, Ni0.97Nb0.03, Ni0.95Nb0.05, which is the only
phase detected by XRD for these samples. Based on the above
observations, the synthesized oxides with 1–5% Nb content may be
considered as basis for the investigation of the Ni–Nb solid solution
contribution in the catalytic performance of Ni0.85Nb0.15 in ethane
ODH.
The most important requirement for a good ODH catalyst is the
ability to effectively convert ethane to ethylene at low temperature
with a high selectivity. In order to study the selectivity of the mixed
Ni–Nb–O oxides and since selectivity is generally strongly related
to conversion, we conducted a second series of experiments at con-
According to the principle of controlled valence [26], higher than
5+
nickel valence cations (such as Nb ) act as electron donors when
◦
stant temperature (350 C), constant ethane/oxygen ratio (2/1) and
+
dissolved in cation deficient NiO and reduce the positive p hole
−
concentration and consequently the electrophilic O radicals of the
NiO acceptor. These species have been reported to affect, both pos-
itively and negatively, the oxidative dehydrogenation of ethane to
ethylene [10,13,34]. Quantification of the O -TPD results tabulated
2
in Table 1 indeed confirms that the amount of desorbed oxygen
obtained in O -TPD experiments decreases with increasing Nb con-
2
tent up to 15% Nb. As illustrated in Fig. 10, at Nb loadings up to ∼3%,
the excess oxygen undergoes a drastic reduction. Therefore, and in
accordance with the role of Nb as electron donor, the elimination
of excess oxygen (electrophilic oxygen species) is strongly related
to the formation of the Ni–Nb solid solution.
In previous works [9,14], the excess oxygen of the catalysts
measured by O -TPD and identified as electrophilic oxygen, was
2
inversely related to selectivity in ethane ODH. The same trend was
recorded in the synthesized materials of this study. At Nb load-
ings up to ∼3%, selectivity to ethylene rapidly increases, from 20%
(
NiO) to ∼70% (Ni
Nb
) at iso-conversion 10%. As shown in
0.97
0.03
Fig. 10, selectivity to ethylene inversely correlates with the amount
of desorbed oxygen, rendering the formation of Ni–Nb solid solu-
tion crucial for the selective transformation of the catalytic surface.
At higher Nb loadings, up to 15%, the NiNb solid solu-
tion co-exists with a highly distorted Nb-rich phase [9,12]. The
Fig. 9. Ethene selectivity as a function of ethane conversion (reaction conditions:
◦
T = 350 C, C2H6/O2 = 2/1).

Z. Skoufa et al. / Catalysis Today 192 (2012) 169–176
175
co-exists with the growing NiNb solid solution for Nb-loadings up
to 15%. This phase becomes crystalline at 20% Nb metal content.
CO-TPR studies confirmed the gradual formation of two distinct
phases in Ni–Nb–O mixed oxides with increasing Nb-content as
evidenced by the existence of two reduction peaks for samples with
Nb/(Ni + Nb) ratio ≥ 0.10.
The introduction of even
a
tiny amount of Nb
(
(
Nb/(Ni + Nb) = 0.01) leads to a drastic increase in selectivity
150%) compared to pure NiO, demonstrating the crucial role of the
Ni–Nb solid solution in the enhanced ethane ODH performance.
This major selectivity improvement continues for loadings up to
∼
3%. As Nb loading increases further, selectivity also improves
at however reduced pace, possibly due to the appearance of an
additional Nb-rich phase with lower selectivity.
Based on the above, Ni–Nb solid solution can be estab-
lished as the key component for the high ethylene selectivity
of the Ni0.85Nb0.15 catalyst, as the insertion of Nb decreases the
electrophilic oxygen species responsible for the total oxidation
reactions. Co-existence of the Ni–Nb solid solution with the amor-
phous Nb-rich phase does not seem to influence significantly the
catalytic properties of the material.
Acknowledgments
These results have been achieved within the framework of the
ACENET project “Alkanes to light olefins via novel catalysts and pro-
cess Schemes – (AL2OL)”, with funding from the General Secretariat
for Research and Technology (GSRT), Greece.
Fig. 10. Amount of desorbed oxygen per surface area (empty symbols-dash line)
and ethene selectivity (full symbols-solid line) as a function of Nb-loading.
presence of this second phase results in the appearance of a high-
References
temperature reduction peak in CO-TPR profiles of Ni0.90Nb0.10
,
Ni0.85Nb0.15 and Ni0.80Nb0.20. This amorphous Nb-rich phase is sim-
ulated by the prepared Nb-rich sample in this work (that also
demonstrated this high-temperature reduction peak) and exhib-
ited selectivity to ethylene equal to 70%. At these Nb-loadings
[
1] Ethylene production, consumption and trade balance, Association of Petro-
chemical Producers in Europe, http://www.petrochemistry.net/ (September
2011).
[
[
[
2] T. Ren, M. Patel, K. Blok, Energy 31 (2006) 425–451.
3] F. Cavani, N. Ballarini, A. Cericola, Catal. Today 127 (2007) 113–131.
4] G. Centi, F. Cavani, F. Trifirò, Selective Oxidation by Heterogeneous Catalysis,
Kluwer Academic/Plenum Publishers, New York, 2001.
5] F. Cavani, F. Trifirò, Partial oxidation of C2 to C4 paraffins, in: M.
Baerns (Ed.), Basic Principles in Applied Catalysis, Springer Publisher, 2004,
pp. 19–84.
(
3–15%) selectivity to ethylene still increases at a slower pace,
as the main physicochemical transformation, the formation of
Ni–Nb solid solution, though still under development, has already
been achieved to a sufficient extent. The co-existence of the
amorphous phase does not seem to deteriorate the catalytic
properties.
[
[6] X. Liu, H. Zhu, Q. Ge, W. Li, H. Xu, Prog. Chem. 16 (2004) 900–910.
[
[
[
7] V.C. Corberán, Top. Catal. 52 (2009) 962–969.
8] M.O. Guerrero-Pérez, M.A. Ba n˜ ares, Catal. Today 142 (2009) 245–251.
9] E. Heracleous, A.A. Lemonidou, J. Catal. 237 (2006) 162–174.
Further enrichment in Nb leads to the formation of crystalline
NiNb O as evidenced by XRD in the Ni0.80Nb0.20 catalyst [9,12].
2
6
[10] E. Heracleous, A.A. Lemonidou, J. Catal. 237 (2006) 175–189.
11] E. Rojas, M.O. Guerrero-Pérez, M.A. Ba n˜ ares, Catal. Commun. 10 (2009)
[
[
[
[
The presence of this phase leads to an increase in oxygen excess,
with a simultaneous slight drop in ethylene selectivity (Fig. 10).
In any case, the increase of the Nb-rich phases, either in amor-
phous or crystal state, is responsible for the slight deterioration of
ethylene selectivity as it is accompanied by an increase in oxygen
excess.
1555–1557.
12] E. Heracleous, A. Delimitis, L. Nalbandian, A.A. Lemonidou, Appl. Catal. A: Gen.
325 (2007) 220–226.
13] B. Savova, S. Loridant, D. Filkova, J.M.M. Millet, Appl. Catal. A: Gen. 390 (2010)
148–157.
14] E. Heracleous, A.A. Lemonidou, J. Catal. 270 (2010) 67–75.
[15] B. Solsona, F. Ivars, A. Dejoz, P. Concepción, M.I. Vázquez, J.M. López Nieto, Top.
Catal. 52 (2009) 751–757.
[
16] B. Solsona, P. Concepción, S. Hernández, B. Demicol, J.M. Lopez Nieto, Catal.
Today (2011), doi:10.1016/j.cattod.2011.03.056.
5
. Conclusions
[
[
[
[
17] A. Cimino, F.S. Stone, Adv. Catal. 47 (2002) 141.
18] A. Aaddane, M. Kacimi, M. Ziyad, Catal. Lett. 73 (2001) 47–53.
19] A. Maione, M. Devillers, J. Solid State Chem. 177 (2004) 2339–2349.
20] B. Pillay, M.R. Mathebula, H.B. Friedrich, Appl. Catal. A: Gen. 361 (2009)
In this work, an effort is conducted to elucidate the contribu-
tion of the different structural phases of the Ni0.85Nb0.15 catalyst
in its catalytic performance in ethane ODH. A series of Ni-based
mixed oxides with very low Nb content (Nb/(Ni + Nb) atomic
ratio = 0.01–0.05) simulating the Ni–Nb solid solution and an amor-
phous Nb-rich sample were synthesized and studied separately, in
addition to previously reported Ni–Nb–O mixed oxides with higher
Nb-content. The main conclusions drawn were the following:
Pure nickel oxide is a non-stoichiometric oxide, with cationic
deficiency (NiO1 + x). Upon promotion with Nb up to 3–5%, a Ni–Nb
solid solution, formed upon incorporation of part of Nb ions into
NiO crystal, is the predominant phase. Further increase of Nb-
loading leads to the formation of a second, Nb-rich phase that
5
7–64.
[21] B.M. Reddy, P. Lakshmanan, S. Loridant, Y. Yamada, T. Kobayashi, C.
López-Cartes, T.C. Rojas, A. Fernández, J. Phys. Chem. 110 (2006)
140–9147.
B
9
[
22] B. Savova, D. Filkova, D. Cri s¸ an, M. Cri s¸ an, M. R aˇ ileanu, N. Dr aˇ gan, L. Petrov, J.C.
Védrine, Appl. Catal. A: Gen. 359 (2009) 55–61.
[23] L.B. McCusker, R.B. Von Dreele, D.E. Cox, D. Louër, P.J. Scardi, J. Appl. Crystallogr.
2 (1999) 36.
3
[
24] A.M. Rubinstein, A.A. Dulov, A.A. Slinkin, L.A. Abramova, I.S. Gershenzon,
L.A. Gorskaya, V.J. Danyushevskii, M.I. Dashevskii, A.L. Klyachko-Gurvich, T.K.
Lavrovskaya, L.I. Lafer, V.I. Yakerson, J. Catal. 35 (1974) 80–91.
[
25] B.S. Girgis, W.E. Mourad, T.M. Ghazi, Surf. Technol.
67–373.
7
(1978)
3
[
26] P.A. Cox, Transition Metal Oxides: An Introduction to their Electronic Structure
and Properties, Oxford University Inc., New York, 1992.

1
76
Z. Skoufa et al. / Catalysis Today 192 (2012) 169–176
[
[
27] M.E. Dry, F.S. Stone, Discuss. Faraday Soc. 28 (1959) 192–200.
28] A. Hornés, P. Bera, A.L. Cámara, D. Gamarra, G. Munuera, A. Martínez-Arias, J.
Catal. 268 (2009) 367–375.
29] S.T. Shen, H.S. Weng, Ind. Eng. Chem. Res. 37 (1998) 2654–2661.
30] B. Zhao, X.-. Ke, J.-. Bao, C.-. Wang, L. Dong, Y.-. Chen, H.-. Chen, J. Phys. Chem.
C 113 (2009) 14440–14447.
[31] V.L. Budarin, V.E. Diyuk, N.V. Zakharenko, B.A. Eichis, V.K. Yatsimirskii, Theor.
Exp. Chem. 34 (1998) 283–286.
[32] R.D. Shannon, C.T. Prewitt, J. Inorg. Nucl. Chem. 32 (1970) 1427–1441.
[33] R.D. Shannon, Acta Crystallogr. A32 (1976) 751.
[34] Y. Schuurman, V. Ducarme, T. Chen, W. Li, C. Mirodatos, G.A. Martin, Appl. Catal.
A: Gen. 163 (1997) 227–235.
[
[