Unraveling mechanistic aspects of the total oxidation of methane over Mn, Ni and Cu spinel cobaltites via in situ electrical conductivity measurements

Article abstract of DOI:10.1016/j.apcata.2020.117901

Mn, Ni and Cu cobaltite spinels were synthesized by coprecipitation, characterized and evaluated in the total oxidation of methane. To gain a deeper insight into their catalytic behavior, their electrical conductivity was studied as a function of temperature and oxygen partial pressure, and was followed with time during sequential exposures to different gases in conditions close to those used for the catalytic tests. The Ni cobaltite is in a metallic conductivity state and cannot be reduced in the reaction conditions, suggesting that it functions via a suprafacial mechanism. The Cu cobaltite behaves as a p-type semiconductor, but it is not reduced under the reaction mixture at the reaction temperature of 350 °C, suggesting that, at least in these conditions, the suprafacial mechanism is operating. The Mn cobaltite has an n-type semiconducting character and is reduced in the reaction conditions, suggesting that a heterogeneous redox mechanism is involved.

Full text of DOI:10.1016/j.apcata.2020.117901

Journal Pre-proof
Unraveling mechanistic aspects of the total oxidation of methane over
Mn, Ni and Cu spinel cobaltites via in situ electrical conductivity
measurements
Marius-Alexandru Mihai, Daniela Cristina Culita, Irina Atkinson,
Florica Papa, Ionel Popescu, Ioan-Cezar Marcu
PII:
S0926-860X(20)30494-4
DOI:
https://doi.org/10.1016/j.apcata.2020.117901
APCATA 117901
Reference:
To appear in:
Applied Catalysis A, General
Received Date:
Revised Date:
Accepted Date:
17 August 2020
22 October 2020
24 October 2020
Please cite this article as: Mihai M-Alexandru, Culita DC, Atkinson I, Papa F, Popescu I, Marcu
I-Cezar, Unraveling mechanistic aspects of the total oxidation of methane over Mn, Ni and Cu
spinel cobaltites via in situ electrical conductivity measurements, Applied Catalysis A, General
(2020), doi: https://doi.org/10.1016/j.apcata.2020.117901
This is a PDF file of an article that has undergone enhancements after acceptance, such as
the addition of a cover page and metadata, and formatting for readability, but it is not yet the
definitive version of record. This version will undergo additional copyediting, typesetting and
review before it is published in its final form, but we are providing this version to give early
visibility of the article. Please note that, during the production process, errors may be
discovered which could affect the content, and all legal disclaimers that apply to the journal
pertain.
©
2020 Published by Elsevier.

Unraveling mechanistic aspects of the total oxidation of methane
over Mn, Ni and Cu spinel cobaltites via in situ electrical
conductivity measurements
1
,2
2
2
2
3
Marius-Alexandru Mihai , Daniela Cristina Culita , Irina Atkinson , Florica Papa , Ionel Popescu , Ioan-
Cezar Marcu1,3,*
1
Laboratory of Chemical Technology and Catalysis, Department of Organic Chemistry, Biochemistry and
Catalysis, Faculty of Chemistry, University of Bucharest, 4-12, Blv. Regina Elisabeta, 030018, Bucharest,
Romania.
2
“Ilie Murgulescu” Institute of Physical Chemistry of the Romanian Academy, 202, Spl. Independentei,
0
3
60021, Bucharest, Romania.
Research Center for Catalysts and Catalytic Processes, Faculty of Chemistry, University of Bucharest,
-12, Blv. Regina Elisabeta, 030018, Bucharest, Romania
4
*
Corresponding author: E-mail address: ioancezar.marcu@chimie.unibuc.ro (I.-C. Marcu).

Graphical abstract
Research highlights





Mn and Cu cobaltites behave as semiconductors of n-type and p-type, respectively.
Ni cobaltite is in a metallic conductivity state.
NiCo
MnCo
CuCo
2
O
4
and CuCo
2
O are not reduced under the reaction mixture: suprafacial mechanism.
4
2
O
4
is reduced in the reaction conditions: heterogeneous redox mechanism.
has the highest density of charge carriers and is the most active catalyst.
2
O
4
ABSTRACT
Mn, Ni and Cu cobaltite spinels were synthesized by coprecipitation, characterized and evaluated in the
total oxidation of methane. To gain a deeper insight into their catalytic behavior, their electrical
conductivity was studied as a function of temperature and oxygen partial pressure, and was followed with
time during sequential exposures to different gases in conditions close to those used for the catalytic tests.
The Ni cobaltite is in a metallic conductivity state and cannot be reduced in the reaction conditions,
suggesting that it functions via a suprafacial mechanism. The Cu cobaltite behaves as a p-type
semiconductor, but it is not reduced under the reaction mixture at the reaction temperature of 350 °C,

suggesting that, at least in these conditions, the suprafacial mechanism is operating. The Mn cobaltite has
an n-type semiconducting character and is reduced in the reaction conditions, suggesting that a
heterogeneous redox mechanism is involved.
KEYWORDS: spinel; cobaltite; electrical conductivity; total oxidation; methane; reaction mechanism
1
. Introduction
Atmospheric pollutants are one of the most serious life threats on our planet. They are mostly
generated in industrial processes and transportation [1]. Some examples are the volatile organic
compounds (VOC) like formaldehyde [2-4] or methane [5-9], inorganic compounds like nitrogen oxides
[
10-12] or even solid microscopic carbon particles [13-15]. The way they harm life differs as ones are
toxic and/or carcinogenic and others affect the global temperature having a greenhouse effect, measured
by the Global Warming Potential (GWP) factor, e.g. methane has a GWP of 21, while N O has a GWP of
10 [7,11,12]. Therefore, it is necessary to convert the harmful compounds into less or even no dangerous
compounds like CO , H O, N etc. In the case of VOC, this can be done either by simply burning the
2
3
2
2
2
exhaust gases which is energy costly, or by catalytic combustion. The latter is currently conducted on
supported noble metal catalysts, but although they show high activity and selectivity toward total oxidation
compounds [3,5-9,11,16,18], they still have some disadvantages like high sintering rates, volatility,
poison-sensitive (especially to sulphur compounds and water), are inhibited by carbon monoxide and are
also very expensive [19]. Nevertheless, they are currently used in the three-way catalysts (TWC) of petrol
or natural gas internal combustion engines [5,7,9,11,17]. An alternative to noble metals is that of variable
oxidation state metal oxides, especially transition-metal oxides, which can have outstanding catalytic
properties in redox reactions [9,11,20]. They are strong competitors to noble metal-based catalysts not
only because of their high abundance and low cost but also due to some other advantages like ease of
preparation, higher resistance to sintering and greater chemical resistance (resistance to poisoning)
[
1,21,22]. Among them, the spinel-structured mixed oxides seem to show quite impressive catalytic
performance in the total oxidation reactions [3,5,7,11,14-18,23,24], a book chapter being already
dedicated to this subject [25]. A key particularity of spinel oxides is the inversion parameter, denoted with
i, expressing the degree of inversion of a specific spinel-structured compound, accurately described by the
푝+ 푞+
푝+ 푞+
푖
general formula [A1 B ] [A B
] O , where usually A is a bivalent cation and B is a trivalent
4
−푖
푖
푇푑
(2−푖) 푂ℎ

cation, located in either tetrahedral (T
d
) or octahedral (O ) sites. This parameter can be manipulated by
h
thermal or other treatments [26] and, because spinels preferably expose the octahedral sites, rather than
the tetrahedral ones, as previously demonstrated by Low Energy Ion Scattering (LEIS) analysis [27,28],
the surface population of active sites of a spinel-based catalyst can be controlled [28].
Methane total oxidation is a widely used reaction for testing the performance of different total
oxidation catalysts. Research is intensively conducted on finding efficient catalysts based on transition
metal oxides and mixed-metal oxides, including spinel-structured cobaltites which showed excellent
performance. Indeed, from a series of cobalt-nickel mixed oxides with different Co/Ni ratios, the CoNi
(
50:50) catalyst was shown to be the most active for methane combustion, its performance being close to
that of Pd/γ-Al [6]. This performance was ascribed to the formation of a distorted NiCo spinel
structure which was shown to enhance the adsorption of oxygen [6]. In a series of mesoporous MCo
M = Cu, Zn and Ni) spinels, CuCo was shown to be the most active catalyst in methane combustion,
its superior activity being attributed to a high normalized amount of Co cations on the surface [29].
Among co-precipitated MCo (M = Ni, Co, Mn, Fe, Cu) spinels, NiCo emerged as the best catalyst
for the complete oxidation of a CO-CH mixture, with a total conversion temperature (T100) of only 350
C [7]. This was 70 °C lower than that observed for the complete oxidation of pure methane, obviously
due to a synergistic effect between CO and CH [5]. This effect was also reported in the total oxidation of
a CO-LPG (liquefied petroleum gas) mixture over spinel-structured MCo (M = Ni, Cu, Zn) catalysts
17]. With NiCo , the most active catalyst in this series, total conversion was achieved at 185 °C for
CO-LPG mixture and at 195 °C for LPG alone [17]. In the total oxidation of propane and CO over spinel-
structured Co-Mn mixed oxides with different cation ratios, Co (0 ≤ x ≤ 3), the optimum system
with x = 2.3 was significantly more active than the best Co catalysts reported in the literature [16].
Modifying Co spinel with 2 wt. % ZrO resulted in an enhanced activity in the total oxidation of
methane corresponding to 90 % conversion at a temperature as low as 377 °C due to increased
2
O
3
2
O
4
2
O
4
(
2
O
4
3
+
2
O
4
2
O
4
4
°
4
2
O
4
[
2
O
4
x
Mn3−x
O
4
3
O
4
3
O
4
2
2
+
concentrations of surface Co and adsorbed oxygen species [30]. Spinel oxides, such as Co
3
4
O [3] and
Ni0.8Co2.2
O
4
[18], were shown to be, with T100 ≈ 90 °C, among the most active catalysts for the complete
oxidation of formaldehyde. All these results clearly point out the versatility and high performance of total
oxidation catalysts based on cobaltite spinels and justify a study focused on the deep understanding of
their functioning mechanism.
In the present work, the mechanism involved in the complete oxidation of methane over three
spinel-structured metal cobaltites, i.e. MnCo
2
O
4
, NiCo
2
O
4
and CuCo
2
4
O , is studied by means of in situ
electrical conductivity measurements, a powerful technique that can reveal the phenomena standing

behind an oxidation catalytic process conducted on semiconductor oxide catalysts [31]. This method was
successfully applied to study several total or partial oxidation reactions over different oxide-based
catalysts, such as the total oxidation of methane over perovskite [32,33] and CuFe2−xMn
x
O spinel [34]
4
catalysts, the complete oxidation of different VOC [35] or the CO, ethene and soot combustion [36] over
Cu-Ce oxides, the total oxidation of methane and the preferential oxidation of CO over Ce-Pr oxides [37],
isobutane oxydehydrogenation over phosphated ceria [38], n-butane oxydehydrogenation over metal
pyrophosphates [39], propane ammoxidation over V and Fe antimonates [40] or VAlO mixed oxides [41],
and n-butane oxidation to maleic anhydride over V-P-O catalysts [42].
2
2
. Experimental
.1. Catalysts synthesis
All compounds were used as received from the supplier without any further processing:
Co(NO
Cu(NO
3
3
)
2
2
·6H
2
2
O
(Sigma-Aldrich), Mn(NO
3
)
2
·4H
2
O
(Carl Roth), Ni(NO
3
)
2
·6H
2
O
(Merck),
, NiCo
)
·3H
O (Sigma-Aldrich) and Na CO (Sigma-Aldrich). The three cobaltites, MnCo
2
3
2
O
4
2
O
4
and CuCo
addition of a 3 M metal nitrates (M/Co mol ratio of 1/2, M = Mn, Ni or Cu) solution to the equivalent
volume of a 1.5 M solution of Na CO , at room temperature, under vigorous stirring and careful pH value
monitoring. In the solution of metal nitrates, the Co and metal M contents corresponded to 5 g of MCo
The target pH was set to 9.0, the starting value being of ca. 12, which corresponds to the 1.5 M Na CO
solution. To maintain the pH value to 9.0, excess portions of 1.5 M Na CO
2
4
O , were synthesized by a coprecipitation method. This consisted in the slowly dropwise
2
3
2
O
4
.
2
3
2
3
were dropwise added to the
reaction mixture along with the solution of metal nitrates. The slurry was further mixed for 30 minutes,
and then vacuum filtered and thoroughly washed with double distilled water. Afterwards the washed
precipitate was dried for 22 h at 120 ºC and finally calcined at 450 ºC for 5 h in static air.
It is worth noting that the synthesis method was chosen to obtain improved textural and catalytic
properties of the cobaltites. Indeed, sodium carbonate was shown to generate catalytic materials with
enhanced total oxidation activity compared to other precipitating agents [5-7,10,17]. Moreover, due to its
porogenic effect, increased specific surface areas of the calcined catalysts are expected [3,5]. Although
both the addition of the precipitating agent to the mixed-metal solution [3,6,10] and the addition of the
mixed-metal solution to the precipitating agent [5,7,17] were applied to obtain metal cobaltites, we carried
out the coprecipitation by adding the mixed-metal solution to the precipitating agent in order to avoid the
formation of an inhomogeneous coprecipitate [43].

2
.2. Catalyst characterization
The powder X-Ray diffraction (XRD) analysis was conducted on a Rigaku Ultima IV
diffractometer using a monochromatic Cu K (λ = 1.5418 Å) radiation source operated at 40 kV and 30
α
-1
mA. The diffractograms were registered on a 5 - 85° 2θ range with a 2° min scan speed and a 0.02° step
width. XRD data were analyzed using Rigaku’s PDXL software connected to ICDD PDF-2 database.
Quantitative X-ray powder diffraction analysis was performed using the Reference Intensity Ratio (RIR)
method. The average crystallite size was calculated using the Williamson-Hall method.
Porosity and surface area of the samples were determined by nitrogen adsorption-desorption
analysis at -196 °C using a Micromeritics ASAP 2020 analyzer. The samples were outgassed at 300 °C
for 3 h before analysis. Specific surface areas (SBET) were calculated using the Brunauer-Emmett-Teller
(BET) equation. The total pore volume (Vtotal) was estimated from the amount adsorbed at the relative
pressure of 0.99. The average pore diameter and pore size distribution curves were obtained from the
desorption branch of the isotherm using Barrett–Joyner–Halenda (BJH) method.
Scanning electron micrographs (SEM) and energy-dispersive X-ray (SEM) measurements were
acquired using a field emission gun scanning electron microscope (FEG-SEM) Nova NanoSEM 630
equipped with an EDX module (FEI Company, USA), working at a high voltage of 20 kV. Four different
small areas were analyzed on each sample, as shown in Fig. S1.
Temperature-programmed reduction under hydrogen (H
oxidation under oxygen (O -TPO) measurements were performed on a Chembet 3000-Quantachrome
analyzer connected to a thermal conductivity detector (TCD) through a silica gel trap. The catalyst samples
2
-TPR) and temperature-programmed
2
(ca. 30 mg) were heated up to 150 °C in an inert argon gas flow, cooled down to room temperature and
-1
-1
then subjected to 5 % H in Ar reducing gas mixture, with a 70 mL min flow rate and a 10 °C min
2
heating rate until 820 °C. After cooling down to room temperature, the reduced samples were reoxidized
by a gas mixture of 5 % O in He, using the same temperature program as for the H -TPR.
2
2
2
.3. Catalytic tests
The catalytic tests were conducted in an 18 mm diameter quartz tube fixed-bed down-flow reactor,
3
operating at atmospheric pressure, loaded with 1 cm of finely powdered catalyst supported on quartz

wool with the dead volumes filled with quartz chips. The reaction temperature was controlled by an
electric furnace surrounding the reactor and the axial temperature profile (reported catalytic reaction
temperature) was measured using an electronic thermometer placed in a thermowell centered in the
-1
3
catalyst bed. The volume hourly space velocity (VHSV) was set to 20,000 h which for 1 cm of catalyst
-1
corresponds to a 333.3 mL min gas flow rate consisting of 1 vol. % methane in air. The gas flow rates
were controlled by fine needle valves and measured with capillary flow meters. The catalyst masses
3
corresponding to 1 cm bed volume were 0.31, 1.17 and 0.71 g for Mn, Ni and Cu cobaltite, respectively.
Before the catalytic tests, the catalyst was pre-treated for ca. 20 min in a stream of air at 440 °C
-1
(
heating rate of 10 °C min ), in order to clean its surface and to improve the test reproducibility. Activity
measurements were performed by decreasing the reaction temperature from 440 °C at regular intervals of
0 °C until a conversion of ca. 10 % was achieved, and then the interval was reduced to 10 °C in order to
2
obtain accurate data for kinetic analysis. The system was allowed to stabilize to the desired temperature
for at least 15 min before the first product analysis was done. Subsequent analyses were performed until
two consecutive measurements were identical. The analyses were done online by a Clarus 500 gas
chromatograph equipped with a thermal conductivity detector (TCD), using two packed columns in series
(6 ft Hayesep and 10 ft molecular sieve 5 Å). The conversion of methane was calculated based on product
analysis in steady state conditions. Complete selectivity to CO and H O was always observed.
2
2
2
.4. Electrical conductivity measurements
To ensure good electrical contacts between the catalyst grains, the powders were compressed into
pellets prior to the electrical conductivity measurements. The pellets were obtained by mixing the spinel
powder (ca. 1.2 g) with a few drops of 5 % aqueous solution of poly(vinyl alcohol) (PVA) in order to
ensure the necessary lubrication and adhesion between grains, this mixture being then pressed using a
7
-2
Carver 4350.L pellet press at a pressure of ca. 2.76 × 10 N m (4000 psi). Finally, the pellets were
calcined in static air at 450 °C for 5 h, in order to eliminate the PVA residue [44]. The obtained pellet was
placed between two platinum electrodes in a horizontal quartz tube surrounded by an electrical furnace.
The temperature was controlled using Pt-Rh thermocouples soldered to the electrodes. Flow rates of gases
flowing over the pelletized sample were controlled by fine needle valves and were measured by capillary
flow meters. The platinum electrodes were used for the measurement of the electrical resistance of the
pellet with a FLUKE 177 multimeter set to ohmmeter. The electrical resistance value was converted into
electrical conductivity (σ) by using the following equation:

ꢁ
ꢂ
ꢁ ꢃ
∙
ꢀ
=
=
(ꢁ)
푅 푆
where ρ is the electrical resistivity of the material, R is the measured electrical resistance, h is the thickness
of the sample (ca. 0.3 cm) and S is the cross-section area of the platinum electrodes of the cell (whose
diameter is 1.3 cm), i.e. the contact area between the metallic conductor and the semiconductor sample.
The electrical conductivity of a pelletized powder sample is a function of the concentration of the charge
carriers and it is given by the formula [31]:
ꢀ
= 퐴푐
(ꢄ)
where c is the concentration of the main charge carriers and A is a constant that includes the charge of the
electron, the mobility of the charge carriers and the number and quality of the contacts between the grains.
Although the samples do not have similar surface areas (Table 1), since they were compressed at the same
pressure and the electrical conductivity measurements were standardized, A can be considered similar for
all the samples under identical conditions. In all the experiments, in order to eliminate any adsorbed
impurities, the samples were first heated in air to the maximum temperature of 450 °C, cooled down to
-1
2
00 °C and then heated again to the desired temperature with a rate of 5 °C min . In these conditions, the
–
+
concentration of adsorbed ionic species like HO or H
conductivity is negligible.
3
O that are able to induce additional surface
3
3
. Results and discussion
.1. Catalyst characterization
The powder XRD patterns (Fig. 1) confirm the presence of the spinel phase in all cobaltites. Pure
spinel phase was observed only in the case of Mn cobaltite (PDF # 00-023-1237), while Ni cobaltite (PDF
#
0
01-073-1702) and Cu cobaltite (PDF # 00-001-1155) also contain certain amounts of NiO (PDF # 00-
47-1049) and CuO (PDF # 00-048-1548) side phases, respectively (Table 1). The phase separation in the
case of Ni cobaltite is likely due to the harsh calcination conditions used, i.e. 450 °C for 5 h, in line with
the reported instability of the NiCo spinel above 400 °C [6,10,29]. However, it is noteworthy that pure
NiCo spinel phases were reported for calcination temperatures as high as 500 °C [17] or 600 °C [5].
2
O
4
2
O
4
In the case of Cu cobaltite sample, the presence of CuO side phase is primarily caused by the thermal
stability of copper-rich malachite phase that occurs in copper-containing hydroxycarbonate precursors,
inducing phase segregation. This is known as the “impossibility of stoichiometric copper cobaltite being

synthesized from hydroxycarbonate precursors”, as reported by Klissurski and Uzunova [45] and
confirmed by Abu-Zied et al. [9]. Another reason for the presence of CuO side phase is the instability of
CuCo
2
O
4
at temperatures higher than 350 °C [45].
#
# - spinel phase



- NiO phase
+
- CuO phase
CuCo O
2
4
#+
#
#
+
#
#
#
#
+
#
+
+
+
#
+
+
#•
NiCo O
2
4
#
#
#
#
#
#
#
#
#
#
•
#
#
•
#
#
MnCo O
2
4
#
#
#
#
#
10
20
30
40
50
(deg)
60
70
80
2
Figure 1. Powder XRD patterns of the prepared cobaltites.
The specific surface area, pore volume, pore size and crystallite size of the MCo
2
4
O spinels are
listed in Table 1 and their corresponding adsorption-desorption isotherms and pore size distributions are
shown in Figs. 2 and S2, respectively. All the samples reveal type IV isotherms according to the IUPAC
classification, with H3-type hysteresis loop for CuCo
hysteresis loops for MnCo and NiCo spinels, characteristic of mesoporous materials with slit-
shaped pores and more complex pore structures, respectively [46]. The BET surface area (SBET) decreases
2
O
4
spinel and a combination of H3 with H2b types
2
O
4
2
O
4
2
-1
2
-1
from 121 m g for MnCo
2
O
4
to 30 m g for CuCo
2
4
O , in agreement with the observed increase of the
crystallite size of the materials (Table 1). Their pore size distributions (Fig. S2) obtained from the
desorption branch of the isotherms indicate unimodal pore structures with well-defined maxima. The
average pore size increases from 9.6 nm for MnCo
2
O
4
to 35.7 nm for CuCo
2
O , being inversely corelated
4
3
-1
with the evolution of the surface area of these materials. Their pore volume ranges between 0.27 cm g
3
-1
for NiCo
2
O
4
and 0.37 cm g for MnCo
2
4
O (Table 1), irrespective of their surface area and pore size.

MnCo O4
2
NiCo O4
2
CuCo O4
2
0.0
0.2
0.4
0.6
0.8
1.0
Relative presure (P/P0)
Figure 2. Adsorption isotherms for the synthesized cobaltites.
Table 1. Physicochemical characteristics of the prepared spinels.
a
Sample
M/Co mol Phase
S
BET
Pore volume Pore size Crystallite
a
2
-1
3
−1
b
ratio
0.43
0.52
composition (%) (m ·g ) (cm ·g )
(nm)
9.6
size (nm)
7.6
MnCo
NiCo
2
O
4
MnCo
NiCo
NiO (4.8)
CuCo
CuO (23.0)
2
O
4
(100)
(95.2)
121
55
0.37
2
O
4
2
O
4
0.27
15.3
8.7
CuCo O
2 4
0.53
2
O
4
(77.0)
30
0.33
35.7
15.1
a
b
c
From EDX analysis (M = Mn, Ni and Cu, respectively). Average pore size. Calculated for the spinel phase from powder
X-Ray diffraction data using the Scherrer equation.
The SEM images of the spinel cobaltites are shown on Fig. 3. It can be observed that Mn cobaltite
is a fluffy aggregate of small particles, while Ni and Cu cobaltites are agglomerates of intermediate and
large spherical particles, respectively. This is in line with the evolution of their surface area and Scherrer
crystallite size. The compositions of the cobaltite samples were estimated by EDX analysis, the M/Co mol
ratios (M = Mn, Ni and Cu) presented in Table 1 being in good agreement with the theoretical value. The
lower Mn/Co molar ratio suggests a slight excess of Co in the Mn cobaltite whose chemical formula can
be written Mn1-xCo2+x
O . Although the M/Co molar ratios for Ni and Cu cobaltite samples are slightly
4
higher than the stoichiometric 0.5 value, due to NiO and CuO phase separation, respectively, the spinel
phase in these materials also contains an excess of Co and, hence, correspond to the chemical formula M1-
x
Co2+x
O (M = Ni and Co).
4

Mn cobaltite
Ni cobaltite
Cu cobaltite
Mn cobaltite
Ni cobaltite
Cu cobaltite
Figure 3. SEM images at two different magnifications of Mn, Ni and Cu cobaltite samples.

The reducibility of the spinel-structured cobaltites, investigated by H
. The CuCo composite spinel oxide shows a large reduction profile with several maxima
corresponding to the reduction of copper and cobalt species. Thus, the two low-temperature maxima can
2
-TPR, is represented in Fig.
4
2
O
4
2
+
+
+
0
be assigned to the reduction of Cu to Cu and Cu to Cu from surface highly dispersed copper species,
3
+
2+
2+
while the high-temperature maxima can be attributed to the stepwise reduction of Co to Co and Co
0
2+
0
to Co , together with the Cu to Cu in bulk-phase co-reduction [47, 48]. The reduction process ends at
57 °C. In the case of NiCo spinel, two obvious reduction events can be observed in the low and high
temperature regions. The low-temperature hydrogen consumption event with the maximum located at 323
4
2
O
4
3
+
2+
°
C can be assigned to the reduction of Co to Co . The high-temperature reduction signal is broad and
2
+
0
correspond to two superimposed processes, i.e. the reduction of Ni to Ni from both separated NiO and
2
+
0
spinel NiCo
2
O phases, and the reduction of Co to Co [5-7,49,50]. The reduction temperature range for
4
Ni cobaltite extends to higher temperature, i.e. 520 °C, suggesting its lower reducibility compared to Cu
cobaltite. Two reduction events, in the low and high temperature regions, can also be observed for
MnCo
2
4
O . The first reduction peak, with several maxima, extending from ca. 130 to 430 °C, corresponds
3
+
2+
4+
3+
2+
0
to the successive reduction of Co to Co , Mn to Mn , and Co to Co , while the broad peak with the
3
+
2+
maximum at 519 °C corresponds to the reduction of Mn to Mn [51, 52]. Among the three cobaltite
spinels, Mn cobaltite is the last to finish the reduction process, with the high-temperature peak positioned
at a temperature where both Cu and Ni cobaltites are already totally reduced. The data in Table 2 show
that the latter were completely reduced, while the degree of reduction of Mn cobaltite is limited to 88.4
0
%
, likely due to the limited capability of Mn cations to reach Mn oxidation state in the studied temperature
interval. The tail observed in the TPR profile of MnCo
2
4
O after the last reducing peak (Fig. 4) indicates
that the reduction process is not over, most likely due to the ability of the newly formed metallic Co
particles to dissociate the hydrogen molecules and thus favoring the further reduction process of Mn2+
species. Indeed, this hydrogen spillover effect was reported for other Co-containing mixed oxides, such
as Co
3
O
4
-CeO systems [53] and Co-Mn-Al spinel oxides [54].
2

Figure 4. Normalized H -TPR curves of the cobaltites.
2
Table 2. Hydrogen consumptions of MCo
2
O
4
spinels in H -TPR experiments and oxygen consumptions
2
of the reduced spinels in O -TPO experiments.
2
Catalyst Maximum hydrogen Measured hydrogen Degree of Measured
Degree
reoxidation
of
a
consumption
μmol/g)
consumption
(μmol/g)
reduction oxygen
(
(%)
consumption (%)
μmol/g)
(
MnCo
NiCo
CuCo
2
O
4
16,893
16,628
16,299
14,933
16,637
16,643
88.4
100.1
102.1
6,333
7,431
7,839
84.8
89.4
96.2
2
O
4
2
O
4
a
2 4
Calculated for the MCo O stoichiometry and considering the reduction of all cations to their zero-valence state.
To check the capacity of reoxidation of the reduced spinel cobaltites, O
2
-TPO runs were carried
out after the H -TPR measurements, the TPO profiles obtained being shown in Fig. 5. Broad features,
2
with several maxima, extending up to 650 °C for Mn cobaltite and to 800 °C for Cu and Ni cobaltites can
be observed reflecting both the complex redox chemistry of the involved transition metals [55] and the
existence of metallic particles of different sizes in the reduced material [56]. The low-temperature peaks
at 226 and 248 °C for reduced Cu and Ni cobaltites, respectively, could be attributed to the reoxidation of
copper and nickel metallic particles resulted from the reduction of CuO and NiO side phases, respectively,
identified in the initial material. The successive peaks above 300 °C can be attributed to the oxidation of
metallic species from the M-Co (M = Mn, Co and Ni) metallic alloy particles likely formed by the
2
+
3+
reduction of MCo
2
O spinel phase, to M (with M = Mn, Co, Ni and Cu) and, then, to M (with M = Mn
4

4
+
and Co) and M (with M = Mn). The total oxygen consumptions calculated from the TPO profiles and
the degrees of reoxidation calculated considering a degree of reduction of 100 % for Cu and Ni cobaltites
and of 88.4 % for Mn cobaltite, are reported in Table 2. It can be observed that, although the degrees of
reoxidation are very high, the reduced cobaltites were not totally reoxidized. This suggests that the bulk
of the reoxidized particles contains transition metal species in lower valence state, likely including metallic
zerovalent state. The reoxidation ability of the three samples in terms of oxygen consumption increases
following the order MnCo
2
O
4
< NiCo
2
O
4
< CuCo
2
4
O .
3
70
550
500
450
400
350
300
250
200
150
100
430
2
40
685
735
515
CuCo O
2
4
460
660
725
264
390
515
NiCo O
2
4
600
50
MnCo O
2 4
0
100
200
300
400
500
600
700
800
900
o
Temperature ( C)
Figure 5. Normalized O -TPR curves of the reduced cobaltites.
2
The H
2
-TPR together with subsequent O -TPO experiments evidenced high reduction and
2
reoxidation abilities of the studied spinel cobaltites.
3
.2. Catalytic properties
The conversion vs. temperature curves for the total oxidation of methane, as a model molecule,
over the cobaltite catalysts, limited to a temperature lower than their calcination temperature, are depicted
in Fig. 6. For comparison purposes, the curve of a reference Pd/Al
and previously reported in Ref. [57] is also shown in Fig. 6. It can be observed that the best catalyst is the
copper cobaltite, CuCo , which, in the low conversion domain, is even better than the reference
Pd/Al system. For conversion values higher than ca. 10-15 %, the methane conversion increases slowly
2
O
3
catalyst tested in similar conditions
O
2 4
2
O
3

with reaction temperature for all the cobaltite catalysts indicating that the diffusional influences become
significant. Within all the temperature range studied, the activity of the cobaltites follows the order:
CuCo
2
O
4
> NiCo
2
O
4
> MnCo
2
4
O . It is worth noting that the relative activity of similar catalytic materials
in the total oxidation of different molecules reported in the open literature is quite different from that
reported here for methane total oxidation. Thus, the order of activity of coprecipitated cobaltites calcined
at 400 °C under a 4.5 vol. % CO in air gas mixture in the total oxidation of a 1.5 vol. % CO – 1.5 vol. %
CH
4
in air mixture was: NiCo
2
O
4
> MnCo
2
O
4
> CuCo
2
O
4
[7]. Also, for similarly prepared cobaltites,
in the total oxidation of liquefied
calcined at 300 °C, NiCo
2
O
4
was shown to be more active than CuCo
2
O
4
petroleum gas (2.5 vol. % LPG in air) [17]. On the other hand, for mesoporous spinel cobaltites prepared
by the nanocasting method using mesoporous SBA-15 silica as a hard template and calcined in air at 380
°
C, the activity in the oxidation of CO (1 vol. % CO in air) in the low conversion range followed the order:
MnCo > CuCo > NiCo , the latter also showing a continuous deactivation on stream [23]. These
2
O
4
2
O
4
2
O
4
data show that the catalytic behavior of the cobaltite spinels is a complex function of different factors,
such as the method of preparation, including the temperature and atmosphere of calcination, the nature of
the molecule to be oxidized etc.
Figure 6. Conversion vs. temperature sigmoids for the synthesized materials and the Pd/Al
2
3
O reference
in methane total oxidation.
The temperature corresponding to 10 % methane conversion (T10), both the intrinsic and specific
activities at 360 °C and the apparent activation energy, E , obtained from the Arrhenius plots shown in
Fig. S3, are summarized in Table 3. It can be observed that not only in terms of conversion but also in
a

terms of reaction rate the copper cobaltite is the most active catalyst in this series. Notably, its intrinsic
activity is more than four times higher than that of the nickel cobaltite, while the corresponding activation
energy is only slightly lower. This suggests that the nature of the active site is not too different in the two
catalytic materials, but the copper cobaltite contains a significantly higher density of active sites compared
to the nickel cobaltite. On the other hand, the activation energy for the manganese cobaltite is significantly
higher compared to the two other cobaltites, accounting for its lowest intrinsic activity and suggesting a
quite different nature of the active site in this material. The higher specific activity observed for the
manganese cobaltite compared to the nickel cobaltite is obviously due to the significantly higher surface
area of the former.
Table 3. Catalytic properties of the cobaltite materials in methane combustion.
Catalyst
T
10 (°C) Reaction rate at 360 °C
E
a
-
1
Specific
Intrinsic
(10 mol·s ·m )
(kJ·mol )
-
7
-1 -1
-9
-1
-2
(
10 mol·s ·g )
MnCo
NiCo
CuCo
2
O
4
381
358
341
3.38
2.03
4.84
2.79
3.70
16.18
134.8
78.3
74.3
2
O
4
2
O
4
3
.3. Electrical conductivity measurements
To gain a deep understanding of the catalytic behavior of the cobaltite materials in the complete
oxidation of methane, their redox and semiconductive properties were studied by in situ electrical
conductivity measurements as a function of temperature in the reaction temperature range, of oxygen
partial pressure and of the nature of the gas phase in contact with the solid, including the reaction mixture,
at constant temperature of 350 °C which is in the kinetic domain for all the catalysts.
3
.3.1. Conductivity variations of the cobaltites as a function of temperature
The conductivities of the three cobaltite-based materials were studied under air at atmospheric
pressure in the temperature interval between 200 and 450 °C. The variations of electrical conductivity as
a function of temperature are displayed in Fig. 7a, while the variations of log(σ) as a function of the
reciprocal absolute temperature (Arrhenius plots) are plotted in Fig. 7b. For both Cu and Mn cobaltites
the electrical conductivity increases exponentially with temperature accounting for a typical
semiconducting behavior, while for Ni cobaltite the electrical conductivity is high and does not vary with
temperature accounting for a quasi-metallic conducting behavior. To explain this behavior, it is worth
remembering that in spinels, including cobaltites, a σ*(e ) band is formed by strong covalent B-O-B type
g

interaction between the octahedral B-site cations, as shown for Co
When this band is empty, as in the case of Co spinel (octahedral B-site Co cations have the electronic
configuration 푡 푒 ), or when the charge carriers are present as in the case of Co3-xNi system for small
values of x (octahedral B-site Ni cations have the electronic configuration 푡 푒 ), but they are Anderson-
3
O
4
and Co3-xNi
x
O systems [58,59].
4
3
+
3
O
4
6
0
O
x 4
2
푔 푔
3
+
6
1
2
푔 푔
localized, their density being lower than a threshold percolation concentration, the spinels have a
3
+
semiconducting behavior [58,59]. An increased Ni concentration in the octahedral B-sites of the spinel
structure, which means an increased charge carriers concentration in the σ*(e ) band, leads to the
percolation of the conducting B-O-B chains throughout the lattice and, hence, to a metallic behavior, as
observed for the NiCo spinel [58].
g
2
O
4
Mn cobaltite
Ni cobaltite
Cu cobaltite
0
.014
.012
1.2
1
0
0.01
0.8
0.6
0.4
0.2
0
0.008
0.006
0.004
0.002
0
150
200
250
300
350
400
450
500
Temperature (C)
(a)

1
0
1
2
3
4
Mn cobaltite
Ni cobaltite
Cu cobaltite
-
-
-
-
R² = 0.9982
R² = 0.9973
1.2
1.4
1.6
1
1.8
000/T (K )
2
2.2
-1
(b)
Figure 7. The electrical conductivity of the cobaltites as a function of the temperature (a) and
variation of log() as a function of the reciprocal absolute temperature (Arrhenius plots) (b).
The variation of the electrical conductivity (σ) of semiconductors with temperature is governed by
the Arrhenius law:
−퐸ꢆ⁄ꢇ푇
ꢅ
ꢈ
ꢀ
= ꢀ0푒
(3)
where σ
is the activation energy of conduction. The logarithmation of Eq. (3) yields the following linear
equation:
0
is the pre-exponential factor, R is the universal gas constant, T is the absolute temperature and
E
c
ꢊ ꢁ
ꢋ
ln ꢀ = ln ꢀ ꢉ
∙
(ꢍ)
0
푅 ꢌ
which describes the variations depicted in Fig. 5b and allows determining the σ
cobaltites (Table 4). The σ
concentration, is three orders of magnitude greater for CuCo
values are lower than 0.5 eV and fall within the range observed for similar materials at temperatures lower
than ca. 400-450 °C [60-62]. The E value for NiCo is zero, in line with its quasi-metallic conductivity
59].
0
and E
c
values for the
0
value, which is a material constant depending on the charge carrier
than for the two other cobaltites. The E
2
O
4
c
c
2
O
4
[

Table 4. Pellet densities and Arrhenius parameters for the electrical conductivity of the cobaltite spinels.
3
-1
-1
-1
Sample
MnCo
NiCo
CuCo
Pellet density (cm g )
E
c
(eV)
σ (Ω cm )
0
2
O
4
2.65
3.39
3.68
0.38
0.0
3.03
2
O
4
0.80
1.93 × 103
2
O
4
0.45
3
.3.2. Conductivity variations of the cobaltites as a function of oxygen partial pressure
The electrical conductivity of the extrinsic oxide semiconductors is a function of the oxygen partial
pressure, its variation at constant temperature being dependent on the type of the semiconductor, i.e. the
electrical conductivity either increases with the oxygen partial pressure for p-type semiconductors or
decreases for n-type semiconductors. This behavior is determined by the type of lattice defects in the
material [31]. Therefore, electrical conductivity measurements as a function of oxygen partial pressure
gives information about the type of semiconductor and the nature of point defects in the solid.
Fig. 8 displays the variation of the electrical conductivity with the oxygen partial pressure, P 2, at
O
constant temperature of 350 °C which is in the kinetic domain for all the cobaltite-based catalysts. It can
be observed that for both Ni and Cu cobaltites the electrical conductivity does not vary with the oxygen
partial pressure. While for the Ni cobaltite, which is in a quasi-metallic conductivity state, this is rather
expected [33], for the Cu cobaltite, this behavior suggesting an intrinsic conduction mechanism [31] is
quite unexpected. Indeed, the latter is known to be a p-type semiconductor involving a small polaron
hopping mechanism between the mixed valence cations located in the octahedral sites, as generally
observed in spinel oxides, including cobaltites [63,64].

-
-
-
-
0.05
-2.52
Ni cobaltite
Cu cobaltite
Mn cobaltite
-
0.1
-
-
-
-
2.53
2.54
2.55
2.56
0.15
R² = 0.9975
Slope: -1/27.3
-
0.2
0.25
-
0.3
0.35
-2.57
-
1.6
-1.4
-1.2
-1
-0.8
-0.6
log(PO2
)
-1
-1
Figure 8. The log-log plot of the electrical conductivity (in ohm cm ) of the cobaltite spinels as a
function of the oxygen partial pressure (in atm) at 350 °C.
For the Mn cobaltite the electrical conductivity decreases with increasing the oxygen partial
pressure, i.e. ∂σ/∂P
O
< 0, accounting for its extrinsic n-type semiconductivity [31]. Most likely, this is
2
associated to the existence of anionic vacancies point defects in Mn cobaltite, which form according to
the following equilibria written using the Kröger-Vink notation [65]:
퐾ꢏ
1
×
•
′
ꢎ ↔ ꢎ
ꢐ 푉 ꢐ 푒
(5)
푂
2(푔)
푂
2
퐾ꢑ
1
×
••
′
ꢎ ↔ ꢎ
ꢐ 푉 ꢐ ꢄ푒
(6)
푂
2
2(푔)
푂
×
•
••
where ꢎ represents an oxygen anion in regular lattice points, 푉 and 푉 represent singly and doubly
푂
푂
푂
′
ionized oxygen vacancies, respectively, and 푒 represents a quasi-free electron responsible for the n-type
conductivity of the solid. Depending on which of equilibria (5) or (6) is dominant, the electrical
−
1ꢒ4
ꢑ
−1ꢒ6
conductivity will be a function of 푃푂
or 푃푂ꢑ , respectively, and, hence, the slope of the log  logPO2
plot will be -1/4 or -1/6, respectively. The slope of the log  logPO2 plot for MnCo
2
4
O shown in Fig. 8
was found to be equal to -1/27, which is quite different from -1/4 or -1/6, suggesting that this spinel exhibit
both electronic (ꢀ ) and ionic (ꢀ ) conduction, the total conductivity being:
ꢓ
푖

ꢀ
= ꢀ ꢐ ꢀ푖
(7)
ꢓ
−
1ꢒ4
−1ꢒ6
The electronic conductivity is expected to be a function of 푃푂ꢑ or 푃푂ꢑ depending on which singly or
doubly ionized oxygen vacancies are dominant in the solid, while the ionic conductivity is independent of
the oxygen partial pressure. To identify the most probable dependency of ꢀ on the oxygen partial pressure
ꢓ
and to confirm the conduction model proposed in Eq. (7) for the Mn cobaltite, the total conductivity σ was
−
1ꢒ4
ꢑ
−1ꢒ6
plotted as a function of both 푃푂
and 푃푂ꢑ (Fig. 9a), linear relationships being observed in both cases.
−
1ꢒ6
−1ꢒ4
However, the correlation coefficient being greater for the 푃푂ꢑ than for the 푃푂ꢑ dependency, the doubly
ionized oxygen vacancies are considered to be dominant point defects in MnCo . Therefore, the ꢀ푖
conductivity value for this sample at the temperature concerned, i.e. 350 °C, was calculated from the y-
2
O
4
-3
-1
-1
intercept of the plot in Fig. 9a and is equal to 2.23 × 10 Ω cm . Then, the ꢀ values corresponding to
ꢓ
the different oxygen partial pressures were calculated by subtracting ꢀ from ꢀ according to Eq. (7), and
푖
were represented as a function of the oxygen pressure in a log–log plot (Fig. 9b). The obtained linear plot
has a good correlation coefficient and exhibits the expected -1/6 slope confirming that the doubly ionized
oxygen vacancies are the dominant point defects in the n-type Mn cobaltite. Notably, the lattice oxygen-
anion conductivity (ꢀ ) is higher than the electronic conductivity (ꢀ ) within all the oxygen partial pressure
푖
ꢓ
range studied (Table S1).
(
PO2)-1/4 (atm-1/4
)
0.5
1
1.5
2
2.5
3.0E-03
2.8E-03
2.6E-03
2.4E-03
2.2E-03
2.0E-03
3.4E-03
3.2E-03
3.0E-03
2.8E-03
2.6E-03
2.4E-03
R² = 0.9930
R² = 0.9889
σ = 2.23  10-
3
i
0
0.5
1
1.5
2
(
PO2)-1/6 (atm-1/6
)
(a)

-
3.1
-
3.15
R² = 0.9929
-
3.2
Slope: -1/5.93 ≈ -1/6
-
3.25
-
3.3
-
1.6
-1.4
-1.2
-1
log(PO2
-0.8
-0.6
-0.4
)
(
b)
Figure 9. Total conductivity σ as a function of (P 2
conductivity σ as a function of the oxygen partial pressure in a log-log plot (b) for the Mn cobaltite spinel.
)-1/n with n = 4 and 6, respectively (a), and electronic
O
e
3
.3.3. Study of the redox behavior of the cobaltites by in situ electrical conductivity
measurements
The redox behavior of the spinel cobaltites was studied by electrical conductivity measurement
under different gases flowing over their surface, including the reaction mixture, i.e. 1 vol. % CH in air,
at constant temperature of 350 °C. The oxides were heated from room temperature to 350 °C, at a heating
4
-1
rate of 5 °C min in an air flow at atmospheric pressure, and then they were maintained under air until
reaching the steady state in these conditions before introducing successive sequences of different gas
mixtures. The variations of the electrical conductivity observed are presented in Fig. 10.

-
-
-
-
-
1.8
2.0
2.2
2.4
2.6
1
% CH4
1% CH4
in N
5% CH4
in air
5% CH4
in N
Air
N2
Air
Air
Air
Air
Air
0
.0
in air
2
2
-
-
-
0.5
1.0
1.5
NiCo O
2
4
CuCo O
2
4
MnCo O
2
4
-2.8
0
70
140
210
280
350
420
Time (min)
-1
-1
Figure 10. Variation of the electrical conductivity  (in ohm cm ) of the cobaltite-based materials during
sequential exposures to different gaseous atmospheres at 350 °C.
No change of the electrical conductivity of the Ni cobaltite at constant temperature of 350 °C can
be observed during sequences under different more or less reductive and oxidative gas mixtures, including
1
and 5 vol. % CH in both air and nitrogen, suggesting that this mixed oxide does not have a redox
4
behavior in these conditions, which include the catalytic reaction conditions. This clearly suggests that the
oxidation of methane over the Ni cobaltite-based catalyst takes place via a reaction mechanism only
involving chemisorbed oxygen species, i.e. a suprafacial reaction mechanism, a heterogeneous redox
mechanism involving lattice oxygen species being firmly excluded. This behavior, already observed for a
LaCoO
electrons from the delocalized σ*(e
methane and oxygen, and intermediates, which is correlated with a suprafacial reaction mechanism
66,67]. It is worth noting that the electrical behavior of the Ni cobaltite, which was shown to contain tiny
amounts of NiO, is obviously governed by the NiCo phase. Indeed, NiO is known to be a p-type
3
perovskite combustion catalyst [33], is obviously associated to its metallic conductivity state, the
g
) band being likely involved in the surface bonding of reagents, i.e.
[
2
O
4
semiconducting oxide which is reversibly reduced with a hydrocarbon-air mixture [68]. Or this has not
been observed with the Ni cobaltite sample, which remains in a metallic-like conductivity state whatever
the nature of the gas phase in contact with the solid (Fig. 10), clearly suggesting that the NiCo
is fully percolating and, thus, governs the electrical conductivity of this material.
2
4
O phase

For the Cu cobaltite, it can be observed that electrical conductivity decreased after the introduction
of the nitrogen flow over the sample, suggesting its p-type character according to the Heckelsberg criterion
[
69], in line with the literature data for this spinel oxide [63]. When air was introduced again over the
sample, the electrical conductivity reached quickly the initial value. Then a gaseous mixture containing 1
vol. % methane in air (the reaction mixture) was passed over the sample but, surprisingly, no change of
the electrical conductivity could be observed. However, when the methane concentration in air was
increased to 5 vol. %, the electrical conductivity decreased compared to that under air, confirming the p-
type semiconducting character of Cu cobaltite. Moreover, when gas mixtures consisting of 1 and 5 vol. %
methane in nitrogen flowed over the sample, the electrical conductivity strongly decreased, accounting
for the reduction of the p-type oxide, which is greater when the concentration of the reducing gas
(methane) is higher. For both methane concentrations in nitrogen the steady state cannot be reached after
ca. 50 min, suggesting a slow diffusion of the oxygen vacancies formed on the surface of the solid into
the bulk. Notably, when an air flow was passed over the reduced sample the electrical conductivity was
completely restored suggesting that the reoxidation of the reduced solid was totally reversible. The
reoxidation was immediate for the solid reduced under 5 vol. % methane in air, suggesting only a
superficial reduction, while it took longer times, i.e. 20 and 35 min for the reduced solid under 1 and 5
vol. % methane in nitrogen, respectively, suggesting deeper bulk degrees of reduction. To conclude, the
Cu cobaltite is a reducible p-type semiconducting oxide, which is, however, not reduced by the reaction
mixture, i.e. 1 vol. % methane in air, at 350 °C, suggesting that, at least in these conditions, a suprafacial
reaction mechanism is involved in the total oxidation of methane over this catalyst. This does not exclude
the possibility of a heterogeneous redox mechanism operating at higher reaction temperatures over Cu
cobaltite. Indeed, it has recently been shown that both suprafacial and intrafacial mechanisms can be
involved separately or simultaneously in the total oxidation of methane over cube-shaped Co
3
O
4
nanoparticles depending on the reaction temperature and the actual dioxygen pressure [70]. It is
noteworthy that the electrical behavior of the Cu cobaltite, which was shown to contain a certain amount
of CuO, is governed by the CuCo
2
4
O phase. Indeed, considering the well-known high reducibility of CuO,
the reduction of Cu cobaltite sample under the reaction conditions would have been expected. Moreover,
based on our previous results [36], CuO switches from p-type to n-type when reduced with hydrocarbon-
air or hydrocarbon-nitrogen mixtures and the reoxidation of the reduced solid is not fully reversible. Or
these phenomena have not been observed with the Cu cobaltite sample (Fig. 10), unambiguously
suggesting that the CuCo
this material.
2
4
O phase is fully percolating and, thus, governs the electrical conductivity of

In the case of Mn cobaltite, the electrical conductivity under nitrogen was higher than in air
confirming its n-type character according to the Heckelsberg criterion [69]. Under air again, the electrical
conductivity reached the initial value. Then, under the reaction mixture flow, i.e. 1 vol. % methane in air,
the electrical conductivity slightly increased clearly suggesting the superficial reduction of the n-type
oxide and, hence, the involvement of a heterogeneous redox mechanism in the total oxidation of methane
over this catalyst. The reduced solid was completely reoxidized when air replaced the reaction mixture,
accounting for the reversibility of the phenomenon. To get further information about the redox behavior
of the MnCo
2
4
O spinel, a gaseous mixture containing 1 vol. % methane in nitrogen was then passed over
the sample. It can be observed that the electrical conductivity first increased during ca. 15 min up to a
maximum value not higher than that under pure nitrogen (dotted line on Fig. 10), in line with the n-type
behavior of the solid, then, surprisingly, it strongly decreased during the next 35 min suggesting a switch
to the p-type behavior. Finally, under a new air sequence the electrical conductivity first quickly jumped
to a value close to the previously observed maximum and then decreased, as for an n-type oxide, reaching
within ca. 15 min a plateau corresponding to the steady state in air which is identical with the initial plateau
under air confirming the complete reoxidation of the MnCo
observed phenomena. To explain this behavior, it is worth remembering that, on one hand, in cobaltite
spinels a σ*(e ) band is formed by strong covalent B-O-B type interaction between the octahedral B-site
cations [58], which are the dominant surface cations [71] and, on the other hand, in MnCo spinel, the
tetrahedral A-sites are preferentially occupied by Co cations, while the octahedral B-sites are occupied
2
O
4
spinel and, hence, the reversibility of the
g
2
O
4
2
+
2
+
3+
3+
4+
by mixed valence Co , Co (in low-spin state), Mn and Mn cations [72]. Among the octahedral B-
2
+
6
1
site cations, only Co has one electron on the 푒 level, as it has the electronic configuration 푡 푒 , being
푔
2푔 푔
thus able to place charge carriers, i.e. electrons, in the σ*(e
g
) band which are likely responsible for the n-
type conductivity observed. However, these charge carriers are Anderson-localized [73] as demonstrated
by the linear behavior of the ln(ρ) vs. T1/4 plot for the MnCo
sample within all the temperature range
from 200 to 450 °C (Fig. S4), accounting for a semiconducting behavior associated to the hopping of
2
O
4
2
+
3+
electrons between Co and Co species on octahedral sites rather than a metallic behavior with itinerant
electrons in the σ*(e
g
) band as observed for NiCo
2
O
4
sample. When methane is in contact with the
3
+
2+
MnCo O sample, the octahedral Co species are reduced to Co leading to an increased charge carriers
2
4
concentration and, hence, to an increased n-type electrical conductivity. However, this increase is limited,
3
+
2+
as observed in Fig. 10, due to a limited number of octahedral Co species to be reduced to Co . Under
the reaction mixture, i.e. 1 vol. % methane in air, the solid remains of n-type, its reduction being obviously
3
+
2+
limited to the Co → Co reduction process, according to the following equation:

×
×
′
••
8
퐶표 ꢐ ꢍꢎ ꢐ 퐶퐻 → 8퐶표 ꢐ ꢍ푉 ꢐ 퐶ꢎ ꢐ ꢄ퐻 ꢎ
(8)
ꢔꢕ
푂
4
ꢔꢕ
푂
2
2
where 퐶표× and 퐶표′ stand for Co and Co species, respectively, sitting on Co octahedral lattice
3+
2+
3+
ꢔꢕ
ꢔꢕ
3+
points. This leads to the conclusion that octahedral Co species represent the catalytic site for the total
oxidation of methane over MnCo spinel in the given reaction conditions which functions via a
heterogeneous redox mechanism. On the other hand, under the methane-nitrogen mixture the reduction of
2
O
4
3
+
2+
all octahedral Co to Co species takes place first, the n-type electrical conductivity passing first through
3
+
2+
a maximum corresponding to an optimum Co /Co ratio in the octahedral sites, and then, when all the
3
+
2+
Co species are reduced to Co , it disappears. Continuing to expose the solid to the methane-nitrogen
mixture, the solid continues losing oxygen with formation of oxygen vacancies but, now, with the
4
+
3+
simultaneous reduction of Mn species on octahedral sites to Mn , according to the following equation:
푀푛• ꢐ ꢍꢎ ꢐ 퐶퐻 → 8푀푛 ꢐ ꢍ푉 ꢐ 퐶ꢎ ꢐ ꢄ퐻 ꢎ
×
×
••
(9)
8
ꢔꢕ
푂
4
ꢔꢕ
푂
2
2
where 푀푛• and 푀푛× stand for Mn and Mn species, respectively, sitting on Co octahedral lattice
4+
3+
3+
ꢔꢕ
ꢔꢕ
4+
3+
points. As the conduction mechanism associated to the Mn /Mn polaron hopping is likely of p-type, i.e.
electron hole hopping [74], this explains the observed switch from n-type to p-type of the MnCo spinel
2
O
4
under the methane-nitrogen mixture (Fig. 10), while the reduction of Mn4+ species accounts for the
observed decrease of the p-type conductivity. These phenomena are successively reversible when the solid
is exposed to an air flow, the oxygen vacancies being filled in according to the following equations:
×
••
•
8
푀푛 ꢐ ꢍ푉 ꢐ ꢄꢎ → 8푀푛 ꢐ ꢍꢎ×
(10)
(11)
ꢔꢕ
푂
2
ꢔꢕ
푂
′
••
×
×
8
퐶표 ꢐ ꢍ푉 ꢐ ꢄꢎ → 8퐶표 ꢐ ꢍꢎ
ꢔꢕ
푂
2
ꢔꢕ
푂
These results explain why MnCo
2
O
4
spinel, which participates to the catalytic process with nucleophilic
lattice O species via a heterogeneous redox mechanism, is less active in the total oxidation reaction than
NiCo and CuCo spinels, which are functioning via a suprafacial mechanism involving more
reactive electrophilic adsorbed O and O species [65]. This is in line with the different nature of the active
2
-
2
O
4
2
O
4
−
–
2
sites in MnCo
activation energies (Table 3) and discussed in Section 3.2. On the other hand, as the latter spinels contain
similar active sites, the remarkably higher intrinsic activity of CuCo compared to NiCo can be
2
O
4
spinel compared to NiCo
2
O
4
and CuCo
2
O
4
spinels, as suggested by the apparent
2
O
4
2
O
4
attributed to its significantly higher density of active sites (Section 3.2). Indeed, if the charge carriers are
involved in the surface bonding of reagents and intermediates, as stated above for a suprafacial
mechanism, then this is consistent with the appreciably greater charge carrier concentration estimated for
CuCo
2
4
O spinel (Table 4). Finally, it should be pointed out that if an oxide is reduced under hydrogen and

reoxidized under oxygen during H
2
-TPR and subsequent O -TPO experiments, respectively, within the
2
reaction temperature range, this not necessarily demonstrates that it is reducible and functions via a
heterogeneous redox mechanism under the reaction conditions.
4
. Conclusions
Spinel-structured MnCo
2
O
4
, NiCo
2
O
4
and CuCo
2
O were synthesized via coprecipitation followed
4
by calcination at 450 °C. While Mn cobaltite consisted of a pure spinel phase, Ni and Cu cobaltites contain
besides the spinel phase certain amounts of NiO and CuO side phases, respectively. Their surface area
2
-1
2
-1
varies within a relatively wide range, from 30 m g for CuCo
2
O
4
to 121 m g for MnCo
2
4
O , following
the inverse variation of the spinel phase crystallite size. All the cobaltites are reducible under hydrogen,
the Ni and Cu cobaltites being completely reduced, while the degree of reduction of Mn cobaltite is limited
to 88.4 %. All three samples also show high reoxidation abilities under oxygen. Their activity in the total
oxidation of methane in terms of both T10 and intrinsic reaction rates, follows the order: Cu cobaltite > Ni
cobaltite > Mn cobaltite. Mn and Cu cobaltites behave as semiconductors of n-type and p-type,
respectively, while the Ni cobaltite is in a metallic conductivity state within all the temperature range
studied. At constant temperature of 350 °C, which is in the kinetic domain for all the catalysts, NiCo
and CuCo are not reduced under the reaction mixture, i.e. 1 vol. % methane in air, in spite of their high
reduction and reoxidation abilities evidenced by H -TPR and O -TPO experiments, respectively. This
clearly suggests that a suprafacial mechanism is operating during the oxidation of methane over these
catalysts. At the same time, MnCo is reduced in the reaction conditions indicating that, in this case, a
2
O
4
2
O
4
2
2
2
O
4
heterogeneous redox mechanism is involved.
Authors contribution
Marius-Alexandru Mihai: Investigation (catalysts preparation, catalytic tests, electrical
conductivity measurements), Writing – original draft. Daniela Cristina Culita: Investigation (textural
analysis). Irina Atkinson: Investigation (XRD analysis). Florica Papa: Investigation (H
2
-TPR and O -
2
TPO measurements). Ionel Popescu: Investigation (catalytic tests, electrical conductivity measurements),
Methodology, Validation. Ioan-Cezar Marcu: Conceptualization, Methodology, Resources, Writing -
review & editing, Supervision.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal
relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
The authors are grateful to Mr. Gabriel Crăciun from National Institute for Research and
Development in Microtechnologies, Bucharest, Romania, for performing the SEM-EDX measurements.
References
1
. Kuśtrowski, P., Rokicińska, A., Kondratowicz, T. Abatement of volatile organic compounds
emission as a target for various human activities including energy production, Adv. Inorg. Chem., 72
2018), 385:419.
(
2
. Salthammer, T., Mentese, S., Marutzky, R., Formaldehyde in the indoor environment, Chem.
Rev., 110 (2010), 2536:2572.
3
. Fan, Z., Zhang, Z., Fang, W., Yao, X., Zou, G., Shangguan, W., Low-temperature catalytic
oxidation of formaldehyde over Co
Catalysis, 37 (2016), 947:954.
3
4
O catalysts prepared using various precipitants, Chinese Journal of
4
. Huang, Y., Ye, K., Li, H., Fan, W., Zhao, F., Zhang, Y., Ji, H., A highly durable catalyst based
nanosheets for low-temperature formaldehyde oxidation, Nano Research, 9 (2016),
on Co
x
Mn3–x
O
4
3
881:3892.
5
. Trivedi, S., Prasad, R., Choice of precipitant and calcination temperature of precursor for
synthesis of NiCo for control of CO–CH emissions from CNG vehicles, Journal of Environmental
Sciences, 65 (2018), 62:71.
2
O
4
4
6
. Lim, T.H., Cho, S.J., Yang, H.S., Engelhard, M.H., Kim, D.H., Effect of Co/Ni ratios in cobalt
nickel mixed oxide catalysts on methane combustion, Applied Catalysis A: General, 505 (2015), 62:69

7
. Trivedi, S., and Prasad, R., Selection of cobaltite and effect of preparation method of NiCo
2
O
4
for catalytic oxidation of CO–CH mixture, Asia‐ Pac. J. Chem. Eng., 12 (2017), 440:453.
4
8
. Vishnyakov, A.V., Gridasova, T.P., Chashchin, V.A., Rodina, K.V., Catalytic Properties of
SnO
2
–TiO Compositions in Total Methane Oxidation, Kinet. Catal. 52 (2011), 733:738.
2
9
. Keav, S., Matam, S. K., Ferri, D., Weidenkaff, A. Structured, Perovskite-Based Catalysts and
Their Application as Three-Way Catalytic Converters—A Review, Catalysts, 4 (2014), 226:255.
1
0. Abu-Zied, B.M., Soliman, S.A., Abdellah, S.E., Enhanced direct N
(0.0≤x≤1.0) spinel-oxide catalysts, Journal of Industrial and Engineering Chemistry, 21
2015), 814:821.
2
O decomposition over
Cu
x
Co1−xCo
2
O
4
(
1
1. Bion, N., Can, F., Courtois, X., Duprez, D., Vedrine, J.C., Metal Oxides in Heterogeneous
Catalysis, Elsevier, 2018, Chapter 6, Transition metal oxides for combustion and depollution processes,
pp. 287-335.
1
2. Li, Z., He-jian, L., Xiu-feng, X., Catalytic decomposition of N
2
O over Mg-Co composite
oxides hydrothermally prepared by using carbon sphere as template, J Fuel Chem Technol, 46 (2018),
5
69:577.
1
3. Neeft, J.P.A., Makkee, M., Moulijn, J.A., Catalytic oxidation of carbon black — I. Activity of
catalysts and classification of oxidation profiles, Fuel, 77 (1998) 111:119.
1
4. Wang, J., Yang, G., Cheng, L., Shin, E.W., Men, Y., Three-dimensionally ordered
macroporous spinel type MCr (M = Co, Ni, Zn, Mn) catalysts with highly enhanced catalytic
performance for soot combustion, Catal. Sci. Technol., 5 (2015), 4594:4601.
2
O
4
1
5. Soloviev, S.O., Kapran, A.Y., Kurylets, Y.P., Oxidation of diesel soot on binary oxide
CuCr(Co)-based monoliths, Journal of Environmental Sciences, 28 (2015), 171:177.
1
6. Faure, B., Alphonse, P., Co–Mn-oxide spinel catalysts for CO and propane oxidation at mild
temperature, Applied Catalysis B: Environmental, 180 (2016) 715:725.
1
7. Prasad, R., Sony, Singh, P., Low temperature complete combustion of a lean mixture of LPG
emissions over cobaltite catalysts, Catal. Sci. Technol., 3 (2013), 3223:3233.