NiMoO4 preparation from polyampholytic hybrid precursors: Benefiting of the memory effect in the oxidative dehydrogenation of propane

Article abstract of DOI:10.1016/j.cattod.2012.03.070

NiMoO₄ are prepared thanks to the calcination of hybrid precursors. These catalysts are synthesised from nickel and molybdate ions hybridised with a polyampholytic amphiphilic alternated copolymer. The idea is to take advantage of the copolymer supramolecular organisation to improve the physico-chemical properties and the catalytic behaviour of the catalysts. Homogeneous hybrid materials have been successfully synthesised thanks to interactions between inorganic ions and copolymer charges. The polymer matrix exothermic decomposition favours the crystallisation of nickel molybdate at lower temperatures. Depending on the amount of ions added into the copolymer matrix, either β-NiMoO₄ or α-NiMoO₄ is stabilised at RT after the calcination of hybrid precursors. The obtention of β-NiMoO₄ is unusual as it is generally supposed to be unstable at RT. Moreover, the mixed oxides morphology seems to be related to that of the original copolymer matrix. The three latter points-early crystallisation temperature, phase stabilisation and copolymer preserved morphology, argue in favour of the existence of a memory effect of the copolymer matrix on the final mixed oxides. These NiMoO₄ proved to be active in propane oxidative dehydrogenation. As expected, β-NiMoO₄ stabilised at RT is more selective in propene than α-NiMoO₄ pointing the potentialities of our preparation method.

Full text of DOI:10.1016/j.cattod.2012.03.070

Catalysis Today 203 (2013) 24–31
Contents lists available at SciVerse ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
NiMoO4 preparation from polyampholytic hybrid precursors: Benefiting of the
memory effect in the oxidative dehydrogenation of propane
B. Farin, C. Swalus, M. Devillers, E.M. Gaigneaux∗
Institute of Condensed Matter and Nanosciences (IMCN), Division “MOlecules, Solids and reactiviTy” (MOST), Université catholique de Louvain, Croix du Sud 2/17,
1348 Louvain-la-Neuve, Belgium
a r t i c l e i n f o
a b s t r a c t
Article history:
NiMoO4 are prepared thanks to the calcination of hybrid precursors. These catalysts are synthesised
from nickel and molybdate ions hybridised with a polyampholytic amphiphilic alternated copolymer.
The idea is to take advantage of the copolymer supramolecular organisation to improve the physico-
chemical properties and the catalytic behaviour of the catalysts. Homogeneous hybrid materials have
been successfully synthesised thanks to interactions between inorganic ions and copolymer charges.
The polymer matrix exothermic decomposition favours the crystallisation of nickel molybdate at lower
temperatures. Depending on the amount of ions added into the copolymer matrix, either ␤-NiMoO4
or ␣-NiMoO4 is stabilised at RT after the calcination of hybrid precursors. The obtention of ␤-NiMoO4
is unusual as it is generally supposed to be unstable at RT. Moreover, the mixed oxides morphology
seems to be related to that of the original copolymer matrix. The three latter points – early crystallisation
temperature, phase stabilisation and copolymer preserved morphology, argue in favour of the existence
of a memory effect of the copolymer matrix on the final mixed oxides. These NiMoO4 proved to be active in
propane oxidative dehydrogenation. As expected, ␤-NiMoO4 stabilised at RT is more selective in propene
than ␣-NiMoO4 pointing the potentialities of our preparation method.
Received 28 October 2011
Received in revised form 23 February 2012
Accepted 27 March 2012
Available online 5 May 2012
Keywords:
Hybrid precursor
Memory effect
Nickel molybdate
Propane ODH
© 2012 Elsevier B.V. All rights reserved.
1. Introduction
be the most active but the ␤-phase tends to be the most selective
in propene and thus gives better propene yields [6,8]. However, the
The need for propene has been continuously increasing for the
last decades. Its numerous applications in various industrial sec-
tors (i.e. plastic and textile industries), contribute to the dynamic
development of the global demand for propylene. In such an eco-
nomic context, catalytic propane transformation into propene is an
interesting alternative besides cracking processes. This allows the
valorisation of the abundant, natural and low cost propane feed-
stock [1–3]. The catalytic oxidative dehydrogenation of propane
into propene (ODHP) is an interesting process that avoids draw-
backs of the catalytic direct dehydrogenation of propane [1,4].
Nickel molybdate is usually considered as an effective catalyst
in the ODHP [5,6]. Two active crystallographic phases of NiMoO4
are often distinguished: ␣-NiMoO4 and ␤-NiMoO4 which are
respectively stable and metastable at room temperature (RT). A
molybdenum coordination number (CN) of 4 is the main struc-
tural feature which differentiates ␤-NiMoO4 from ␣-NiMoO4 (CN
Mo = 6), while Ni is octahedrally coordinated in both phases [7].
About the catalytic behaviour in the ODHP, the ␣-phase seems to
great stability of light alkanes makes their activation difficult. Fur-
thermore, the catalytic formulations must also be selective enough
so as to avoid propane total oxidation which is the most thermo-
dynamically favourable pathway [1,4]. These contingencies lead to
a never-ending necessity of catalysts improvement. A widespread
way to achieve it is the development and optimisation of cata-
lysts formulations. Incidentally, Cavani et al. have identified and
sorted the greatest recently developed family of catalysts active
in the ODHP [9]. An alternative consists in improving the cata-
lysts physico-chemical properties (structure, texture. . .) that are
responsible for good performance. In continuation with some pre-
vious work, the possibility envisaged here is the formation of mixed
oxides (Ni Mo O) through the thermal degradation of hybrid pre-
cursors made from the association of an organic copolymer matrix
and inorganic ions [10,11].
The idea of this work is to take advantage of a polyampholytic
amphiphilic alternated copolymer matrix (poly[N,N-diallyl-N-
hexylamine-alt-(maleic acid)]). The latter, synthesised through
cyclopolymerisation of diallyl ammonium salts, displays
supramolecular organisation thanks to hydrophobic interactions
between its alkyl side chains [12]. Inorganic ions ([Ni(H2O)6]2+
MoO42−) are planned to be inserted and dispersed along the matrix
a
,
∗
Corresponding author.
E-mail address: eric.gaigneaux@uclouvain.be (E.M. Gaigneaux).
0920-5861/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.cattod.2012.03.070

B. Farin et al. / Catalysis Today 203 (2013) 24–31
25
through interactions with the polymer charged moieties. In this
context, polyampholytes are a powerful copolymer family. Such
macromolecules are simultaneously able to interact with both
metallic cations and anions thanks to the presence of positive and
negative charges spatially separated on a same repeated unit [13].
Carboxylate moieties have been reported to be particularly effi-
cient to create coordination links with positively charged nickel
ions [14]. The copolymer is also able to make strong electro-
static interactions with anionic species like molybdates through
the positively charged diallylamines. These facts are supposed to
facilitate an efficient alternating dispersion of the inorganic ions
along the copolymer matrix in order to obtain homogeneous hybrid
(organic–inorganic) precursors. Moreover, the diallyamine cycle
makes this copolymer matrix very stable against hydrolysis dur-
ing the synthesis step [12]. The matrix is then removed upon heat
treatment to generate the final mixed oxides active in ODHP.
As various catalysts, nickel molybdates are not highly compet-
itive in ODHP because of the great propane stability and their low
specific area [6]. Our original synthesis way plans obtaining more
effective NiMoO4 catalysts. More precisely, the control and pro-
motion of physico-chemical features responsible of NiMoO4 good
catalytic performance are aimed. Desired structural features of the
mixed oxides are expected. Textural improvements linked to the
volume liberation by the copolymer chain burning, are also con-
sidered [15]. In fact, “memory effects” related to the genesis of the
pre-organised matrix are thought to promote an organisation and
a structuration into the mixed oxide network (i.e. creation of an
organised pore volume into the mixed oxides). Thus, the impact
of the presence and degradation of the copolymer matrix on the
physico-chemical properties (morphology, texture, structure) and
the catalysts behaviour is a key point of this study [15,16].
addition, the pH of the solutions was kept constant with diluted
NH3 (5 mol L−1, Sigma–Aldrich). The final solutions were then
stirred during two hours and lyophilised. The recovered precursors
were firstly decomposed at 300 ◦C under air during 10 h and then
calcined at 550 ◦C under air during 4 h. The amount of added ions
was expressed as molar ratio namely the number of inorganic ions
per copolymer repeat unit (I/R ratio). The following I/R ratios were
chosen so as to evaluate the inorganic ions quantity that the matrix
can fix: 0, 0.25, 0.5, 1, 1.5, 2, 3 and 4. Thereby, for a I/R ratio = 1,
1 mol of C6 repeat units was mixed with 0.5 mol Ni and 0.5 mol
Mo. The hybrid materials were labelled as follows: NiMo-synthesis
pH-I/R ratio (i.e. NiMo-pH6-1). Besides, a reference catalyst was
also synthesised (NiMo-Cop). It was prepared by a conventional
co-precipitation method detailed by Del Rosso et al. [17].
2.2. Hybrid precursors and catalysts characterisation
Fourier transformed infrared spectroscopy was used to study
ions-copolymer interactions. Transparent discs were prepared
after diluting the sample in KBr (>99%, Janssens Chimica) by a
weight factor of 100. The spectra were obtained by recording 600
scans between 400 and 4000 cm−1 with a 4 cm−1 resolution. The
IFS55 Equinox spectrometer (Brucker) used in the transmission
mode was equipped with a DTGS detector.
X-ray diffraction (XRD) measurements were performed with
a Siemens D5000 diffractometer using the K␣ radiation of Cu
(ꢀ = 0.15418 nm). The 2ꢁ diffractograms were recorded between
0.5–10◦ and 10–60◦ at a rate of 0.12 and 0.6◦/min, respectively. The
ICDD-JCPDS database was used to identify the crystalline phases.
Thermogravimetric analyses and simultaneous differential ther-
mal analyses (TGA-SDTA) were performed in a Mettler Toledo
TGA/SDTA 851. The samples (about 4 mg) were heated in an air
stream (40 mL/min) from 30 ◦C to 600 ◦C at a 10 ◦C/min rate.
Thermodiffractometry analyses (XRD-T) were performed under
air on a Siemens D5000 diffractometer apparatus equipped with a
high temperature HTK10 chamber from Anton Paar. The powder
sample was placed on a small platinum strip. The current which
skimmed through the strip was adjusted with a specific controller
and a thermocouple linked to the strip. High temperature stability
and accurate temperature ramp (heating and cooling) could thus
be achieved without over- and down shooting.
2. Experimental
2.1. Hybrid precursors and catalysts synthesis
The polyampholytic amphiphilic copolymer poly[N,N-diallyl-
N-hexylamine-alt-(maleic acid)], hereafter noted C6, was used as
polymer matrix (Fig. 1). It results from the alternation of maleic
acid and diallylhexylamine monomers. The C6 synthesis has been
optimised and detailed by Rullens et al. in a previous paper [12].
(NH4)6Mo7O24·4H2O (>99%, Merck) and Ni(NO3)2·6H2O, (>97%,
Merck) were used as inorganic ions precursors. The polymer
matrix was firstly dispersed overnight in a water solution. Hybrid
precursors were obtained by adding drop-by-drop solutions of
inorganic ions to polymer solutions at pH 6 and 10. During this
During an analysis, the samples were heated by step of 100 ◦C
(2 ◦C/min rising ramp), followed by a one hour waiting time; then
the diffractogram recording took place which lasted 1h30min.
The BET specific surface area of the catalysts was measured
by nitrogen adsorption at 77 K with a Micromeritics TriStar 3000
Fig. 1. (a) chemical structure of a repeated unit of the used C6 copolymer ((N,Ndiallyl-N-hexylamine-alt-maleic acide)) and (b) supramolecular organisation model of the
polymer matrix proposed by Rullens et al. [9].

26
B. Farin et al. / Catalysis Today 203 (2013) 24–31
instrument. The samples were outgassed overnight at 150 ◦C under
vacuum before measurement.
SEM analyses were performed with a LEO 983 GEMINI micro-
scope equipped with a field emission gun. The uncoated samples
were exposed to an acceleration voltage of 1 kV.
2.3. Catalytic tests
NiMoO4 catalysts (0.150 g) were tested in the ODHP reaction
in a U-shape quartz reactor at 475 ◦C and 500 ◦C with a total flow
of 40 mL/min (C3H8/O2/N2 = 8/8.75/83.25 (vol.%)). Propane stability
was checked in such conditions and no conversion was observed
without catalyst. Reaction products identification was made with
an on line Varian GC composed by three columns and two detec-
tors. O2, N2, CO, CO2, C3H8 and C3H6 were separated and quantified
with a Hayesep column coupled with a Molecular Sieve column
and a TCD detector. The presence and quantification of oxygenates
(acrolein, acetaldehyde, propionaldehyde and acetone) were man-
aged through a system composed by an EC-Wax column coupled
with an FID detector. All the generated products were identified
and quantified. The absence of by-product (acrylic acid, acetic
acid, ethanol, propylene oxide) related to propene overoxidation
was checked. Moreover, no propane cracking products were sig-
nificantly detected (PR(%) < 0.01%). No carbon loss can be claimed
thanks to calculated carbon balances of the catalytic results. No
coke production can thus be considered. A standard deviation of
10% is accepted due to experimental imprecision linked to the
device. The catalytic performance were calculated in terms of molar
Fig. 2. XRD diffraction patterns of the C6 copolymer matrix, the inorganic ions alone,
and the hybrid precursors. Hybrid precursors are characterised by different I/R ratios
and have been synthesised at pH6. *NH4NO3, ˆ(NH4)4[NiMo6O18(OH)6]·5H2O.
complexation by ammonia ([Ni(H2O)6−x(NH3)x]2+) delays this phe-
nomenon; nevertheless at pH 10 the precipitate formation cannot
be avoided anymore. However, no nickel precipitation occurs at
this high pH value when the C6 matrix is present in the solution
for I/R ratios ≤1. Molybdenum exists in aqueous solution as soluble
molybdate species (MoO42−) between pH 6 and 14 [19].
H
propane conversion (CC 8 (%)) expressed as:
3
xiC
− xfC
H
H
8
3
8
3
H
CC 8 (%) =
× 100
3
xiC
H
The recovered hybrid materials have been characterised in order
to check the success of the hybridisation. Finding evidences of
interactions between inorganic ions and the copolymer matrix
is a key point of this part. The XRD diffractogram of the C6
matrix displays a low-angle peak (2ꢁ = 4.52◦) which reveals its
supramolecular lamellar organisation (Fig. 2). At pH 6 (Fig. 2) and
10 (not shown), the addition of inorganic ions does not affect
the C6 organisation since this low-angle peak is still observable.
However, the low-angle peak position is progressively shifted to
smaller 2ꢁ angles until the added ions amount equals a I/R ratio
of 1. The insertion of inorganic ions into the copolymer matrix
seems to increase the main chains spacing [20]. The peak position
remains then stable for upper I/R ratios (I/R ratios >1) (Fig. 2). For
such high I/R ratios, the stabilisation of the low-angle peak posi-
tion is coupled with the appearance of sharp peaks revealing the
formation of crystalline inorganic phases. According to the ICDD-
JCPDS database, two different phases are identified: NH4NO3 and
(NH4)4[NiMo6O18(OH)6]·5H2O.
3
8
where xi and xf represent respectively the initial and the post-
reaction molar fractions of propane in the total gas flow. Besides,
the product ratio (PRy(%)) of each product “y” generated during
catalytic tests was also evaluated:
ny
y
ꢀ
PR (%) =
× 100
n
where ny is the mole number of the “y” compound in the post-
reaction total flow and ꢂn is the mole number of all the product
generated through the catalytic reaction. Stoichiometric coeffi-
cients were used to calculate PR(%) of products composed of less
than three carbons (CO, CO2). In the present case, selectivities
calculation is not accurate enough because of the low propane
conversion generated by NiMoO4 made from hybrids.
3. Results
FTIR spectroscopy has been used in order to evaluate ions-
copolymer matrix interactions. At pH 6 or 10, the carboxylic acid
moieties of the C6 matrix are deprotonated and exist as carboxy-
late groups (COO−). These chemical functions are characterised by
3.1. Hybrid precursors characterisation
The hybridisation pH is quite important because of the polyam-
pholytic nature of the C6 copolymer matrix. Every repeat unit of
the C6 copolymer is composed by two carboxylic acid functions
(pKa : 2–3 and pKa : 4–5) and one tertiary amine (pKa = 8–9). At
two specific infra-red bands: the symmetric stretching (␯ COO−)
s
at 1410 cm−1 and the antisymmetric stretching (␯ COO−) at
as
1577 cm−1 (Fig. 3) [21,22].−In the presence of inorganic ions, the first
1
2
one is hidden by the NO3 broad and intense band at 1382 cm−1
.
pH 6, the C6 number of charges and thus the matrix potential abil-
ity to link inorganic ions, are both maximised. Syntheses at pH 10
have also been realised in order to study the ability of Mo-based
species to interact with the copolymer matrix at pH 6. In fact, the
C6 matrix is only negatively charged at pH 10 because the amine
functions are not protonated anymore. Therefore, molybdate ions
are not supposed to be able to interact with the C6 matrix at pH
10. In agreement with their Pourbaix diagram, nickel ions exist for
pH lower than 7 as water soluble [Ni(H2O)6]2+ complexes. Nickel
hydroxide begins to precipitate between pH 7 and 8 [18]. The nickel
The second one is clearly visible. Moreover, for I/R ≤1 ratios, its posi-
tion is progressively shifted to lower wavenumbers when inorganic
ions are hybridised to the C6 matrix. Then, for I/R ratios >1, the posi-
tion of the band remains stable at 1562 cm−1 (Fig. 3). According to
Strathmann and Myneni, such a band shift represents the nickel
complexation by the carboxylate functions of the C6 matrix [22].
Nickel ions are thus interacting with the C6 matrix. At pH 10, car-
boxylic functions are also deprotonated. Therefore, concerning the
␯
as
COO−, a same shift is logically observed (not shown).

B. Farin et al. / Catalysis Today 203 (2013) 24–31
27
at 320 ◦C) suggests that the C6 copolymer matrix is mainly burnt at
such a temperature. The SDTA plot displays three main exother-
mic peaks at 190, 320 and 460 ◦C respectively (Fig. 4). Through
a correlation with the TGA derivative profile, these exothermic
signals turned to be−linked respectively to the degradation of inor-
ganic ions like NO3 at 190 ◦C, and the C6 matrix burning at 320
and 460 ◦C. The second one (320 ◦C) can be correlated to the main
weight loss observed at 320 ◦C on the TGA derivative profile. Con-
sequently, the copolymer matrix degradation can be considered
to be highly exothermic. Besides, an endothermic SDTA signals is
observed between 30 and 100 ◦C which may correspond to the
water evaporation.
On the one hand, XRD-T measured on the NiMo-pH6-1 sam-
ple shows that NiMoO4 crystals are formed from 300 ◦C (Fig. 5).
After heating, a major ␤-NiMoO4 phase is stabilised at RT even
if some ␣-NiMoO4 crystals are also detected. On the other hand,
␣-NiMoO4 is the main phase obtained after calcination of the
NiMo-pH6-4 hybrid precursor (Fig. 5). Nevertheless, small quan-
tities of ␤-NiMoO4 crystals are also present on its diffractogram.
After calcination at 550 ◦C, the XRD pattern of the NiMo-pH10-1
sample presents a mix of several phases (Fig. 6). The ␤-NiMoO4
phase is clearly observable. Besides, molybdenum oxide phases
(Mo9O26, MoO3) also seem to be stabilised. After these identi-
fications, four major XRD peaks (19.1◦, 25.8◦, 29.1◦ and 32.8◦)
remain unexplained. Nickel oxides and ␣-NiMoO4 could be two
reasonable possibilities. As nickel oxide does not present main
diffraction peaks below 32◦, it cannot be associated to the unknown
peaks. On the other hand, the latters could be explained by ␣-
NiMoO4 even if their position is shifted in comparison with the
theoretical value. This phenomenon could be explained by the exis-
tence of impurities and defects through the ␣-NiMoO4 crystalline
network. Besides, the XRD diffraction pattern of the reference cat-
alyst (NiMo-Cop) reveals the stabilisation of ␣-NiMoO4 even if
traces of ␤-NiMoO4 crystals are also detected after calcination
(Fig. 7, a).
Fig. 3. FTIR spectra of the C6 copolymer matrix, the inorganic ions alone, and the
hybrid precursors. Hybrid precursors are characterised by different I/R ratios and
have been synthesised at pH6.
3.2. Hybrid precursors transformation to mixed oxides
characterisation
As the transformation of the hybrid precursors into their corre-
sponding mixed oxides is crucial, it has been studied by TGA/SDTA
and XRD-T. Theoretically, 22% of the initial NiMo-pH6-1 hybrid
precursor weight are supposed to be transformed into NiMoO4
and should thus be recovered. According to the TGA profile of this
sample, the residue at 550 ◦C amounts to 24% of the initial weight
(Fig. 4). This result is in agreement with the expected value, mean-
ing that the great majority of the C6 matrix is indeed burnt at such
a temperature. The TGA derivative profile (dM/dT) of the NiMo-
pH6-1 sample presents a first weight loss between 40 and 110 ◦C
(Fig. 4). It is probably assigned to the water evaporation of the
hybrid precursors. Then, between 180 and 500 ◦C, other weight
losses are noticeable and must probably correspond to the precur-
sors ions (NH4+, NO3−) and C6 copolymer matrix degradation. The
main weight loss observed between 300–350 ◦C (maximum peak
NiMoO4 crystallinity does not qualitatively change after a
catalytic test. The same combinations of crystalline phases are sys-
tematically observed on the XRD diffraction patterns (Fig. 7). Our
original NiMoO4 synthesis method allows thus the formation of
␤-NiMoO4 stable at RT even after a catalytic test. Some further crys-
tallisation of both ␤-NiMoO4 and ␣-NiMoO4 is only observed. The
diffraction signals of both phases follow indeed a similar evolu-
tion. This fact has likely no impact on the catalysts performance in
ODHP.
3.3. Mixed oxides characterisation
Scanning electron microscopy has been used to study the
morphology and the macroscopic organisation of the studied
materials. The polymer matrix morphology tends to be relatively
heterogeneous. In fact, it presents various patterns like layered
structures which are sometimes irregular, hollow spheres and
tubules. These two latter structures are characteristic shapes that
may result from the polymer matrix folding [23]. The NiMo-pH6-1
hybrid precursors present a similar morphology and the same
patterns observed with the C6 copolymer matrix. After calcination
at 550 ◦C, the morphology of NiMoO4 seems to be related to the
one observed with hybrid precursors and the C6 copolymer matrix.
Characteristic patterns like tubules and hollow spheres are still
observable on the SEM pictures of the mixed oxides (Fig. 8). The
textural properties of NiMoO4 synthesised from hybrid precursors
are presented Table 1. The specific area of the calcined hybrids
NiMo-pH6-1 and NiMo-pH6-4 is not so important, namely 12.0
and 18.3 m2/g respectively. On the other hand, the NiMo-Cop
sample is characterised by a higher specific area: 38.4 m2/g.
Fig. 4. TGA (continuous line), derivative of TGA (dotted line) and SDTA (line with
black hollows (•) profiles of the NiMo-pH6-1 sample.

28
B. Farin et al. / Catalysis Today 203 (2013) 24–31
Fig. 5. Thermodiffractometry (XRD-T) patterns of the NiMo-pH6-1 and NiMo-pH6-4 samples. (*) ␤-NiMoO4, (ˆ) ␣-NiMoO4.
3.4. Mixed oxides catalytic behaviour
reaction temperature. Propene PR(%) of 33.3 and 45.4% are respec-
tively measured at 475 ◦C and 500 ◦C. The NiMo-pH6-1 catalyst
is thus significantly more effective to produce propene at higher
temperatures. CO2 is systematically the main product generated
Catalytic performance of the synthesised NiMoO4 has been
measured in ODHP. The results recorded at 475 ◦C and 500 ◦C
are shown Table 1. At 475 ◦C, calcined NiMo-pH6-1 and calcined
NiMo-pH6-4 display 2.2% and 5.7% of propane conversion respec-
tively. NiMoO4 generated through co-precipitation (NiMo-Cop) is
significantly more active in ODHP (C(%) = 12%) at the same reac-
tion temperature. Propane conversion systematically increases
at 500 ◦C even if the same tendency than at 475 ◦C is observed
between the catalysts activity (Table 1). Besides, calcined NiMo-
pH6-1, calcined NiMo-pH6-4 and NiMo-Cop generate at 475 ◦C
propene PR(%) of 33.3, 20.3 and 14.9% respectively (Table 1).
Hybrid precursors burning leads to mixed oxides more effective
to produce propene. Moreover, calcined NiMo-pH6-1 has got the
largest propene PR(%). Note that the latter increases with the
Fig. 7. XRD diffraction patterns of fresh NiMo-Cop (a), tested NiMo-Cop (b), fresh
calcined NiMo-pH6-4 (c), tested calcined NiMo-pH6-4 (d), fresh calcined NiMo-pH6-
1 (e) and tested calcined NiMo-pH6-1 (f). *␤-NiMoO4, ˆ␣-NiMoO4.
Fig. 6. XRD diffraction pattern of the calcined NiMo-pH10-1 sample. (*) ␤-NiMoO4,
(ˆ) ␣-NiMoO4, (ꢀ) Mo9O26 and (+) MoO3.

B. Farin et al. / Catalysis Today 203 (2013) 24–31
29
Table 1
Specific area and catalytic performance (in terms of conversion C(%), and product ratio PR(%)) of the calcined NiMo-pH6-1 and NiMo-pH6-4 hybrid precursors and the
reference catalyst (NiMo-Cop).
Catalyst
Crystalline phase
after calcination
Textural property
Test Condition
Catalytic performance
Specific
Temperature
(◦C)
C(%)
propane (%)
PR(%)
propene (%)
PR(%) CO
(%)
PR(%) CO2 (%)
PR(%)
oxygenates (%)
area (m2/g)
12.0
18.3
38.4
475
500
2.2
3
33.3
45.4
2.3
1.8
58.7
49.4
1.9
2.1
NiMo-pH6-1
NiMo-pH6-4
NiMo-Cop
␤-NiMoO4
␣-NiMoO4
␣-NiMoO4
475
500
5.7
10.5
20.3
22.8
3.8
8.3
75.5
66.7
4.2
5.3
475
500
12.0
17.1
14.9
13.9
17.2
13.9
67.1
67.7
0.8
0.8
by the catalysts (Table 1). For example, CO2 PR(%) of 58.7, 75.5
and 67.1% have been measured at 475 ◦C for the samples NiMo-
pH6-1, NiMo-pH6-4 and NiMo-Cop respectively. CO and traces of
oxygenates have also been detected and quantified (Table 1).
evidenced that nickel ions are interacting with the carboxylate
moieties of the C6 copolymer matrix. Indeed, the shift to lower
wavenumbers of the COO− anti-symmetric band indicates that
nickel ions are coordinated to the carboxylate functions [22]. It
seems to be proportional to the amount of the carboxylate func-
tions that interacts with a nickel ion. This effect can be correlated
to XRD results. One can imagine that the increase of the polymer
chains spacing is induced by the interaction and insertion of inor-
ganic ions into the hydrophilic areas of the matrix organisation.
Fig. 9 presents the evolution of the Bragg distance and the position
of the COO− anti-symmetric band functions of the amount of added
ions (I/R ratio). Both curves are characterised by a slope breaking
at the I/R ratio = 1. The latter seems to correspond to the maximum
amount of inorganic ions that the matrix can fix. In such a situa-
tion, the C6 copolymer matrix seems to be saturated in inorganic
ions. If larger ions amounts are added (I/R ratios > 1), the Mo-
rich ((NH4)4[NiMo6O18(OH)6]·5H2O) crystalline phase containing
nickel and molybdenum is detected on the XRD diffractograms
(Fig. 2). This means that all the added ions cannot be fixed by the
copolymer matrix and thus interact with themselves to generate
separate crystalline phases. Such hybrid materials are considered
to be heterogeneous. It should be underlined that these conclusions
result from two independent characterisation techniques which
reveal the same phenomenon.
4. Discussion
To generate NiMoO4 catalysts, the synthesis of homogeneous
hybrid precursors is the first step of the original experimental
method proposed in this work. Being sure of ions-C6 interac-
tions is thus a key point. Therefore, hybrid materials have been
characterised by XRD and FTIR spectroscopy. At pH 6 and 10,
the supramolecular organisation of the C6 copolymer matrix is
evidenced by a low-angle peak on its diffractogram (Fig. 2). It
is induced by the alkyl side chains of the diallyamines [12,24].
Hydrophobic interactions tend to promote a lamellar organisation
of the matrix. This leads to the apparition of alternated hydropho-
bic and hydrophilic microdomains along the copolymer matrix. The
position of the low-angle peak represents a characteristic inter-
reticular distance (dBragg) in the C6 matrix organisation (Fig. 1(a)). In
previous works, Rullens et al. proposed the model shown (Fig. 1(b))
to represent this organisation [12]. As the inorganic ions are quite
polar and charged, they are thought to be spontaneously stabilised
in the hydrophilic areas of the copolymer matrix. FTIR results have
Fig. 8. SEM pictures of (A) and (B) C6 copolymer matrix, (C) and (D) NiMo-pH6-1 hybrid precursor, (E) and (F) NiMo-pH6-1 calcined hybrid precursor.

30
B. Farin et al. / Catalysis Today 203 (2013) 24–31
such a phase has already been achieved in the literature [26,27].
Two different explanations can be found to understand this result.
Firstly, an excess of nickel could partially avoid the ␤ → ␣ relaxation
at lower temperature. Mazzocchia et al. reported that calcination
of Ni-rich materials (Ni:Mo = 1.0–1.3) leads to the stabilisation of
the ␤-phase at room temperature due to the excess of nickel in
the crystalline lattice of the molybdate [26]. Secondly, the addi-
tion of organic molecules in mixed oxide precursors also helps the
␤-NiMoO4 formation at RT after calcination. For example, the sta-
bilisation at RT of ␤-NiMoO4 crystals have already been achieved
by Mazzocchia et al. through the quick and exothermic decomposi-
tion of oxalate [27]. This second explanation seems to be consistent
with the present system. However, the interactions and dispersion
of the ions through the copolymer matrix seem to be also cru-
cial in order to stabilise ␤-NiMoO4 at RT. When the inorganic ions
are added in large excess (I/R ratio = 4), only a fraction of them is
able to interact with the C6 matrix. This leads to the formation
of a few of ␤-NiMoO4 crystals. However, as the majority of the
engaged ions interact together, the stabilisation of a main ␣-phase
at RT is observed after calcination. In fact, through interactions, the
C6 copolymer matrix seems to avoid the ions self-reorganisation
which leads to the formation of ␣-NiMoO4 crystals. The diffrac-
tograms of calcined NiMo-pH10-1 and NiMo-Cop also support this
reasoning. The repulsive electrostatic interaction between molyb-
date ions and the C6 matrix at pH 10 prevents their good dispersion.
In consequence, ␤-NiMoO4 mixed with significant amounts of ␣-
NiMoO4 crystals, and molybdenum oxide impurities, is detected
after calcination. About the NiMo-Cop XRD diffraction pattern, a ␣-
phase is logically stabilised after calcination as inorganic ions are
not interacting with a carbon additive. Finally, according to some
authors, the presence of carbon residues after the heat treatment
could also help ␤-NiMoO4 stabilisation at RT [6].
According to SEM pictures, the copolymer matrix morphology
tends to be relatively heterogeneous (Fig. 8). Too short alkyl side
chains might explain such a phenomenon. On the one hand, they
are long enough to confer a supramolecular organisation to the C6
copolymer matrix. On the other hand, they seem to be a little bit too
short to really promote a homogeneous organisation of the copoly-
mer matrix. The NiMo-pH6-1 hybrid precursor presents a similar
morphology and the same characteristic patterns as observed with
the copolymer matrix. Thus, the hybridisation of inorganic ions
does not seem to modify the macroscopic organisation of the C6
copolymer. Besides, the morphology of the calcined hybrid precur-
sors seems to be related to that of the C6 matrix. Characteristic
patterns of the C6 matrix like tubules and hollow spheres are still
observable after calcination. Firstly, such a result is another clue
that argues in favour of a good dispersion of the inorganic ions into
the C6 matrix network. Secondly, as the great majority of the C6
matrix is degraded during the calcination step, these observations
mean that the inorganic ions have crystallised through preserving
the host copolymer matrix morphology. The mixed oxides mor-
phology thus keeps in memory that of the C6 matrix.
Fig. 9. Correlation between XRD and FTIR analyses. Evolution of the Bragg distance
(ꢀ) and the position of the COO− anti-symmetric band (᭹) as a function of the
amount of added ions (I/R ratio).
Molybdate ions are thought to interact with the positively
charged amine moieties through electrostatic interactions. Hybrid
precursors characterised by a I/R ratio = 1 have been synthesised
at pH 6 and 10. As the pKa of the amine moieties is in the range
8–9, the C6 copolymer matrix becomes only negatively charged at
pH 10. Consequently, molybdenum ions cannot interact with the
negatively charged C6 copolymers and are thus not supposed to
be well dispersed along them. A mix of ␤-NiMoO4, ␣-NiMoO4 and
molybdenum oxides is stabilised at RT after calcination of such a
hybrid material. The existence of molybdenum oxide impurities
could be caused by the absence of positive charges on the C6 matrix
at pH 10. On the other hand, at pH 6, molybdate ions are thought
to be well dispersed along the copolymer chains thanks to elec-
trostatic interactions with the positive charges of the C6 matrix. A
␤-NiMoO4 phase is mainly stabilised at RT after burning of such
a hybrid precursor. The absence of molybdenum oxide crystals at
pH 6 argues in favour of the existence of electrostatic interactions
between molybdate ions and the C6 matrix.
The control of the hybrid precursors transformation into mixed
oxides is thought to be crucial in order to take as much advan-
tage as possible of the precursors organisation. Therefore, the
precursors degradation profile has been studied by XRD-T, TGA
and DTA. According to the literature, the crystalline network of
nickel molybdate synthesised through co-precipitation is detected
around 400–450 ◦C [6]. XRD-T, measured on the NiMo-pH6-1
hybrid precursor, shows that NiMoO4 crystallises from 300 ◦C,
namely 100–150 ◦C earlier than with the co-precipitation method.
Besides, the TGA derivative profile (dM/dT) of this hybrid precursor
presents its biggest peak at 300–350 ◦C, which corresponds to the
matrix degradation. At such a temperature, this peak can be cor-
related with the main exothermic signal on the hybrid precursor
SDTA profile. The exothermic decomposition of the polymer matrix
induces local and punctual temperature increases of the burning
precursor. Consequently, NiMoO4 crystallises at lower oven tem-
peratures namely 300–350 ◦C.
The catalytic performance of NiMoO4 synthesised thanks to
hybrid precursors have been measured in ODHP. In agreement
with the literature, ␤-NiMoO4 stabilised at RT is more effective to
produce propene than ␣-NiMoO4 [6,8]. This result points out the
importance of stabilising ␤-NiMoO4 at RT as realised by the hybrid
synthesis method. Nevertheless, CO2 is systematically the main
product generated in the catalytic tests. Better propene productions
are related in the literature [4,6,8]. In the present case, two experi-
mental reasons sound contributing to low propene PR(%). Catalytic
tests were managed under oxidative conditions (O2/propane = 1.1)
and a low total flow (40 mL/min). These experimental conditions
likely favour both propane and propene overoxidation and COx pro-
duction, and thus low propene PR(%). Indeed, low total flows imply
high reagent-catalyst contact time.
According to XRD-T analyses, ␣-NiMoO4 is the main phase sta-
bilised at RT after calcination of hybrid precursors characterised by
I/R ratio = 4 (NiMo-pH6-4). On the other hand, calcination of the
NiMo-pH6-1 precursor mainly stabilises the ␤-NiMoO4 phase at
RT even if some ␣-NiMoO4 crystals are also detected (Fig. 5). This
result is quite interesting because the ␤-NiMoO4 crystallographic
phase is known to be metastable at RT [25]. The stabilisation of

B. Farin et al. / Catalysis Today 203 (2013) 24–31
31
Concerning propane conversion, NiMoO4 generated from hybrid
precursors appear to be less active than NiMo-Cop. This lower
activity could be related to the lower specific area of NiMoO4
made from hybrids. The calculation of catalysts intrinsic conversion
(conversion/m2 g of catalyst) argues in favour of such a hypothesis.
Intrinsic conversions of calcined NiMo-pH6-1, calcined NiMo-pH6-
4 and NiMo-Cop are equal at 475 ◦C to 1.41, 2.08 and 2.08% of
propane conversion/m2 respectively. According to these results,
mixed oxides generated from hybrid precursors are intrinsically as
active as NiMoO4 synthesised through co-precipitation. This fact is
in agreement with the literature where ␣-NiMoO4 is defined a bit
more active than ␤-NiMoO4 [6]. Although NiMo-Cop is less effective
to produce propene than NiMoO4 made from hybrids (especially
NiMo-pH6-1). This sounds logic as NiMo-Cop crystallinity is char-
acterised by ␣-NiMoO4 after calcination. ␤-NiMoO4 obtained after
NiMo-pH6-1 displays the best ability to convert propane into
propene.
The improvement of the textural properties of mixed oxides
made from hybrid precursors will be a key point of the future work
in order to overcome their low propane activity. Hybrid precursors
will be synthesised with copolymers characterised by longer alkyl
side chains: (poly(N,Ndiallyl-N-dodecylamine-alt-maleic acid)
(C12) and (poly(N,Ndiallyl-N-octadecylamine-alt-maleic acid)
(C18) [12,28]. The idea is to take advantage of a copolymer
matrix better organised thanks to larger hydrophobic interactions
between the side chains. The polymer combustion is also sup-
posed to create larger and more organised internal volume for
pores formation. As a result, structural and textural improvements
are expected. Such advances are thought to markedly improve the
catalytic behaviour of the NiMoO4 we tried to synthesise.
The propane conversion of NiMoO4 made from hybrids is relatively
low. This could be explained by the low specific area of these mixed
oxides. In order to improve this feature, hybrid precursors are cur-
rently synthesised with copolymer characterised by longer alkyl
side chains. The idea is to take advantage of a better organised poly-
mer matrix thanks to better hydrophobic interactions between the
side chains. Such a polymer is expected to lead to mixed oxides
characterised by improved physico-chemical properties and cat-
alytic behaviour.
Acknowledgements
The authors gratefully thank R. Lasselin for providing the
copolymer. The authors gratefully acknowledge the ‘Fonds pour
la formation à la Recherche dans l’Industrie et dans l’Agriculture
(FRIA)’ of Belgium for financial support and B. Farin’s PhD fel-
lowship. The involvement of the authors in the “Inanomat” IUAP
network sustained by the ‘Service public fédéral de programmation
politique scientifique’ (Belgium) is also acknowledged. The Insti-
tute of Condensed Matter and Nanosciences is alsp involved in the
Cost Action D41 sustained by the European Science Foundation.
Finally, the institute is also indebted to the ‘Communauté franc¸ aise
de Belgique’ for financial support through the ARC programme
(08/13-009).
References
[1] M.M. Barsan, F.C. Thyrion, Catalysis Today 81 (2003) 159–170.
[2] M.A. Botavina, G. Martra, Y.A. Agafonov, N.A. Gaidai, N.V. Nekrasov, D.V.
Trushin, S. Coluccia, A.L. Lapidus, Applied Catalysis A-General 347 (2008)
126–132.
[3] E. Heracleous, A.A. Lemonidou, Journal of Catalysis 237 (2006) 162–174.
[4] V.C. Corberan, Catalysis Today 99 (2005) 33–41.
[5] F. Dury, M.A. Centeno, E.M. Gaigneaux, P. Ruiz, Applied Catalysis A-General 247
(2003) 231–246.
5. Conclusions
[6] L.M. Madeira, M.F. Portela, C. Mazzocchia, Catal. Rev. Sci. Eng. 46 (2004) 53–
110.
[7] J.A. Rodriguez, J.C. Hanson, S. Chaturvedi, A. Maiti, J.L. Brito, Journal of Chemical
Physics 112 (2000) 935–945.
[8] C. Mazzocchia, C. Aboumrad, C. Diagne, E. Tempesti, J.M. Herrmann, G. Thomas,
Catalysis Letters 10 (1991) 181–191.
[9] F. Cavani, N. Ballarini, A. Cericola, Catalysis Today 127 (2007) 113–131.
[10] F. Rullens, M. Devillers, A. Laschewsky, Journal of Materials Chemistry 14 (2004)
3421–3426.
[11] F. Rullens, N. Deligne, A. Laschewsky, M. Devillers, Journal of Materials Chem-
istry 15 (2005) 1668–1676.
[12] F. Rullens, M. Devillers, A. Laschewsky, Macromolecular Chemistry and Physics
205 (2004) 1155–1166.
[13] P. Favresse, A. Laschewsky, Polymer 42 (2001) 2755–2766.
[14] B.L. Rivas, E. Pereira, Macromolecular Symposium 193 (2003) 237–250.
[15] Y. Peng, Z.Y. Liu, F. Ye, X.W. Wei, Chinese Journal of Chemistry 27 (2009)
295–298.
[16] F. Rullens, A. Laschewsky, M. Devillers, Chemistry of Materials 18 (2006)
771–777.
[17] R. Del Rosso, A. Kaddouri, C. Mazzocchia, P. Gronchi, P. Centola, Catalysis Letters
69 (2000) 71–78.
[18] B. Beverskog, I. Puigdomenech, Corrosion Science 39 (1997) 969–980.
[19] N. Mahadevaiah, B. Venkataramani, B.S.J. Prakash, Chemistry of Materials 19
(2007) 4606–4612.
[20] Z.H. Li, P.H. Chang, W.T. Jiang, J.S. Jean, H.L. Hong, L.B. Liao, Journal of Colloid
and Interface Science 360 (2011) 227–232.
[21] G.B. Deacon, R.J. Phillips, Coordination Chemistry Reviews 33 (1980) 227–250.
[22] T.J. Strathmann, S.C.B. Myneni, Geochimica et Cosmochimica Acta 68 (2004)
3441–3458.
[23] H.P. Lin, C.Y. Mou, Accounts of Chemical Research 35 (2002) 927–935.
[24] K. Inomata, Y. Sakamaki, T. Nose, S. Sasaki, Polymer Journal 28 (1996) 992–999.
[25] A. Kaddouri, R. Del Rosso, C. Mazzocchia, D. Fumagalli, Journal of Thermal Anal-
ysis and Calorimetry 63 (2000) 267–277.
[26] C. Mazzocchia, R. Delrosso, P. Centola, Anales De Quimica Serie a-Quimica Fisica
Y Quimica Tecnica 79 (1983) 108–113.
[27] C. Mazzocchia, A. Kaddouri, R. Anouchinsky, M. Sautel, G. Thomas, Solid State
Ionics 63-5 (1993) 731–735.
[28] K. Inomata, Y. Sakamaki, T. Nose, S. Sasaki, Polymer Journal 28 (1996) 986–991.
Hybrid precursors have been synthesised using a polyam-
pholytic amphiphilic alternated copolymer (C6), nickel ions and
molybdate ions. From XRD, FTIR and SEM analyses, hybrid mate-
rials seem to be homogeneous when I/R ratios do not exceed 1. In
such a case, inorganic ions are well-dispersed through the copoly-
mer matrix thanks to electrostatic and complexation interactions
with the charged moieties of the copolymer chains.
According to XRD-T, SDTA, and TGA results, it can be concluded
that the exothermic decomposition of the polymer matrix induce
NiMoO4 crystallisation at lower oven temperatures. ␤-NiMoO4 is
the main phase observed after calcination of a hybrid precursor
characterised by a I/R ratio = 1. Interactions and dispersion of the
inorganic ions through the copolymer matrix coupled with the
presence of carbon during the combustion step may explain the
stabilisation of such a metastable NiMoO4 phase. Besides, when a
large amount of inorganic ions is added (I/R ratio = 4), only a frac-
tion of ions are able to interact with the matrix. Consequently,
␣-NiMoO4 is the main phase formed. SEM pictures reveal that the
characteristic polymer matrix organisation is maintained into the
mixed oxides after the calcination step. The morphology of the
mixed oxides seems to be related to the one of the C6 matrix. The
three latter points – early crystallisation temperature, phase sta-
bilisation and copolymer preserved morphology, argue in favour of
the existence of a memory effect of the copolymer matrix on the
final mixed oxides.
These NiMoO4 synthesised through hybrid precursors are active
in the ODHP. ␤-NiMoO4 stabilised at RT is more effective to produce
propene than ␣-NiMoO4. This points out the importance of stabil-
ising ␤-NiMoO4 at RT as realised by the hybrid synthesis method.