Vanadium promoted molybdenum phosphate catalysts for the vapour phase partial oxidation of methanol to formaldehyde

Article abstract of DOI:10.1016/j.apcata.2014.07.029

The catalytic properties of (MoO₂)₂P ₂O₇ promoted with vanadium have been investigated for the partial oxidation of methanol, and structure-activity relationships probed using a range of characterization techniques. All unpromoted and promoted molybdenum phosphate catalysts were active, with higher vanadium content achieving both high activity and high formaldehyde selectivity at reaction temperatures around 400 °C. The association between increasing vanadium content and the enhanced activity towards methanol oxidation was attributed to the formation of mixed phase catalysts, in particular VOHPO₄·0.5H₂O/ VOPO₄·2H₂O with (MoO₂)₂P ₂O₇. The dispersion of vanadium phosphate phases on the surface of (MoO₂)₂P₂O₇ was found to substantially enhance the catalytic properties of the molybdenum phosphate catalyst. The data from this study indicate that molybdenum phosphate based catalysts are promising candidates for selective oxidation, and hence worthy of further investigation.

Full text of DOI:10.1016/j.apcata.2014.07.029

Applied Catalysis A: General 485 (2014) 51–57
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Vanadium promoted molybdenum phosphate catalysts for the vapour
phase partial oxidation of methanol to formaldehyde
Gareth T. Whiting, Jonathan K. Bartley, Nicholas F. Dummer,
Graham J. Hutchings, Stuart H. Taylor∗
Cardiff Catalysis Institute, Cardiff University, School of Chemistry, Main Building, Park Place, Cardiff, CF10 3AT, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 13 June 2014
Received in revised form 22 July 2014
Accepted 23 July 2014
Available online 1 August 2014
The catalytic properties of (MoO2)2P2O7 promoted with vanadium have been investigated for the par-
tial oxidation of methanol, and structure-activity relationships probed using a range of characterization
techniques. All unpromoted and promoted molybdenum phosphate catalysts were active, with higher
vanadium content achieving both high activity and high formaldehyde selectivity at reaction tempera-
tures around 400 ◦C. The association between increasing vanadium content and the enhanced activity
towards methanol oxidation was attributed to the formation of mixed phase catalysts, in particular
VOHPO4·0.5H2O/VOPO4·2H2O with (MoO2)2P2O7. The dispersion of vanadium phosphate phases on the
surface of (MoO2)2P2O7 was found to substantially enhance the catalytic properties of the molybdenum
phosphate catalyst. The data from this study indicate that molybdenum phosphate based catalysts are
promising candidates for selective oxidation, and hence worthy of further investigation.
© 2014 Elsevier B.V. All rights reserved.
Keywords:
Selective oxidation
Methanol
Formaldehyde
Molybdenum phosphate
Vanadium phosphate
1. Introduction
lithium and sodium batteries [10]. For catalysis, they are mainly
reported for use in the partial propane oxidation reaction, where
Formaldehyde is an important intermediate in the chemical
industry, as it is used in the production of thermosetting resins,
antiseptics, and adhesives. It is also essential for the manufacture
of a range of other materials such as plywood, carpeting, paper and
fertilizers amongst others [1]. Presently there are two competing
industrial processes to produce formaldehyde from the oxidation
of methanol, which are based on silver and ferric molybdate cat-
alysts [1–5]. The silver process uses a methanol-rich feed (around
40%) and reaction temperatures around 650 ◦C, whereas the fer-
ric molybdate process uses a methanol-lean feed (around 8%), and
reaction temperatures in the region of 300 ◦C [6,7]. Both catalysts
produce a high yield of formaldehyde, and the choice of process
is determined by the operating and capital costs, as well as prod-
uct end use, plant size and type of operation [5,8,9]. Over the last
decade, the ferric molybdate-catalysed process has dominated the
market as a consequence of the higher formaldehyde selectivity.
This paper reports for the first time, the use of molybde-
num phosphate catalysts for methanol oxidation to formaldehyde.
Molybdenum phosphate materials (MoPO) have received increas-
ing interest in the last decade for use as new cathode materials for
they are usually promoted by metals such as silver or cerium, and
they produce high selectivity to propene at relatively low conver-
sion [11–15]. In general, MoPO phases are usually comprised of
(PO4)3− tetrahedra linked mostly with (MoO6)6− octahedra [16].
One of the main characteristics of these phosphate materials is
their ability to stabilize molybdenum in various oxidation states,
i.e. Mo6+, Mo5+, Mo3+ and even mixed valencies such as Mo5+/Mo6+
.
These redox properties make them ideal catalysts for oxidation
reactions, such as the partial oxidation of propane, and potentially
for the partial oxidation of methanol. Transition metal phosphate
catalysts have been reported in the literature to be active for
numerous partial oxidation reactions, e.g. vanadyl pyrophosphate
for butane oxidation to maleic anhydride [17,18], and iron phos-
phate for the oxidative dehydrogenation of isobutyric acid into
methacrylic acid [19,20]. There is some evidence to suggest that
the role of phosphate tetrahedra is to enhance the redox properties
of the catalyst, possibly due to the increase in mobility of the lattice
oxygen through the bulk to the surface, where it can re-oxidise the
reduced surface during the reaction [21]. Hence, considering the
high selectivity to formaldehyde using molybdenum oxide based
catalysts, and considering the possible improvement as an oxida-
tion catalyst by incorporating phosphate groups, the investigation
of molybdenum phosphates as catalysts for methanol oxidation is
an interesting concept.
∗
Corresponding author. Tel.: +44 029 2087 4062.
E-mail address: taylorsh@Cardiff.ac.uk (S.H. Taylor).
http://dx.doi.org/10.1016/j.apcata.2014.07.029
0926-860X/© 2014 Elsevier B.V. All rights reserved.

52
G.T. Whiting et al. / Applied Catalysis A: General 485 (2014) 51–57
Against this background we have started to investigate the
efficacy of molybdenum phosphate catalysts for the selective oxi-
dation of methanol to formaldehyde. The effect of adding vanadium
into the molybdenum phosphate structure has also been investi-
gated and the affect on catalyst performance evaluated. Catalysts
have been prepared using a relatively simple co-precipitation tech-
nique, and physico-chemical properties have been studied using a
range of characterisation techniques. Catalytic activity for selective
methanol oxidation is reported for the first time with these cata-
lyst formulations, and performance is related to the structure of the
catalysts.
2. Experimental
2.1. Catalyst Preparation
2.1.1. Unpromoted molybdenum phosphates
The precursor MoO2·HPO4·H2O was obtained by dissolving
MoO3 (15 g, Sigma–Aldrich, >99.5%) in H3PO4 (45 ml, Aldrich, 85%
in H2O, 99.99%) at 180 ◦C. Upon cooling of the viscous solution, con-
centrated HNO3 was added (300 ml, Fisher Chemical, 70% Analytical
grade) and the mixture refluxed for 16 h. After completion of the
reaction, the solid phase was recovered by filtration and washed
with water and acetone, before drying overnight at 110 ◦C in air.
Fig. 1. Powder X-ray diffraction patterns: (a). MoHPO; (b). MoPO; (c). MoPO-V1;
(d). MoPO-V5; (e). MoPO-V10; (f). MoPO-V20.
dual Al/Mg achromatic source. Spectra were acquired over an area
of 700 × 300 ␮m at a pass energy of 40 eV for high resolution scans.
All spectra were calibrated to the C(1s) line of adventitious carbon
at a binding energy of 284.7 eV.
MoO2·HPO4·H2O was calcined (650 ◦C, 6 h, ramp rate 20 ◦C min−1
)
to form (MoO2)2P2O7. The nomenclature for the unpromoted pre-
cursor is Mo-HPO, and for the unpromoted (MoO2)2P2O7 catalyst,
MoPO.
2.3. Methanol oxidation
Vanadium promoted molybdenum phosphate catalysts were
prepared by adding the desired amount of V2O5 (Sigma–Aldrich,
>98%) during the phosphation step of the precursor synthesis,
where MoO3 and V2O5 were dissolved in H3PO4, prior to reflux-
ing with HNO3. The same procedure was then followed as for the
unpromoted precursor and MoPO preparation. The nomenclature
of V promoted (MoO2)2P2O7 is MoPO-Vx, where x denotes either 1,
5, 10 or 20 mol % V, in relation to the molar quantity of Mo.
Catalytic activity for partial gas phase methanol oxidation, was
performed in a fixed bed microreactor. 0.3 g of catalyst was held
between plugs of quartz wool in the centre of a 5 mm i.d. quartz
tube, which was placed vertically into a Carbolite tube furnace, with
the outlet line heated to prevent condensation of products such as
formaldehyde. Mass flow controllers were used to supply the reac-
tant feed mixture of MeOH:O2:He with a molar ratio of 5:10:85,
and a total flow rate of 60 ml min−1 (GHSV = 12000 h−1). To achieve
5 mol. % methanol, helium was passed through a saturator contain-
ing liquid methanol (Aldrich, 99.5%) which was maintained at 8 ◦C
using a thermostatically controlled water bath. The reactor tem-
perature was varied from 25 to 500 ◦C in incremental steps, at each
interval the catalyst was allowed to attain steady state operation
before data were collected. Product analysis was carried out using
a Varian Star 3400 Cx on-line gas chromatograph, which used two
columns in a series/bypass configuration to provide separation of all
reactants and products (calibrated using gas reference standards).
A Carbosieve S-11 (3 m) column was used for the analysis of O2
and CO, accompanied by a Porapak Q (1 m) column to separate
methanol (MeOH), dimethyl ether (DME), methyl formate (MF),
formaldehyde (FA) and CO2. A TCD was used in series with an FID
for product identification and quantification. Methanol conversion
in an empty reactor tube reached around 1% at 500 ◦C.
2.2. Catalyst characterisation
Catalyst surface areas were analysed using a Micromeretics
Gemini 2360 analyser and were determined by multi-point nitro-
gen adsorption at −196 ◦C, prior to data analysis in accordance with
the BET method. All catalysts were degassed under a helium atmo-
sphere (120 ◦C, 2 h) before analysis. Powder X-ray diffraction was
used to identify the crystalline phases present in the catalysts. XRD
patterns were collected using a PANalytical XPert diffractometer,
with a graphite monochromator and a Cu X-ray source operated at
40 kV and 40 mA. Phases were identified by matching the experi-
mental patterns to the ICCD PDF database. Raman spectroscopy was
carried out using a Renishaw inVia Raman microscope equipped
with a 514 nm laser (argon ion) with an average laser power of
25 mW. Before acquisition of catalyst spectra the system was cal-
ibrated using a silicon reference sample. Catalyst samples were
flattened onto an aluminium plate before being analysed.
3. Results and discussion
Scanning electron microscopy (SEM) was conducted using a
Carl Zeiss EVO 40 microscope, with each sample dispersed on an
adhesive carbon disc. Temperature programmed reduction (TPR)
experiments were performed using a Quantachrome ChemBET
chemisorption analyzer equipped with a TCD detector. Samples
were pre-treated in an argon atmosphere at 120 ◦C for 1 h, prior
to analysis under a reducing atmosphere of 10% H2 in Ar, with
a flow rate of 50 ml min−1. The temperature ranged from room
temperature to 750 ◦C, at a specific ramp rate. XPS analysis was
performed using a Kratos Axis Ultra DLD photoelectron spectrom-
eter, equipped with an aluminium monochromatic source and a
3.1. Catalyst characterisation
The diffraction patterns of both Mo-HPO and MoPO are shown
in Fig. 1. The pattern of the highly crystalline Mo-HPO material was
observed, and corresponds well with that of the monoclinic struc-
ture reported by Kierkegaard [22]. The structure consists of parallel
chains, where each chain of PO4 tetrahedra binding together MoO6,
are linked by hydrogen atoms. The calcination of the Mo-HPO pre-
cursor produced a crystalline orthorhombic (MoO2)2P2O7 phase,
(Fig. 1b). In contrast to the parallel chains of the precursor, MoPO

G.T. Whiting et al. / Applied Catalysis A: General 485 (2014) 51–57
53
To investigate the bulk redox behaviour of both materials, TPR
analysis was performed, and results are presented in Fig. 4. The
reduction process for Mo-HPO is initiated at 380 ◦C, preceded by a
further slow increase up to a maximum at 557 ◦C, with an overall
hydrogen consumption of 1421.5 ␮mol/g. The reduction of Mo6+ to
Mo4+ is a two step process [27], as can be observed by the two peaks
at 557 and 579 ◦C, corresponding to the reduction of Mo6+ to Mo5+
and Mo4+ respectively. The onset temperature of the reduction pro-
cess of MoPO is initiated at a higher temperature of around 450 ◦C,
with a sharp increase in hydrogen consumption (1646.0 ␮mol/g),
to a maximum at 579 ◦C.
The diffraction patterns of MoPO promoted with vanadium
are shown in Fig. 1c–f. Each of the materials analysed are highly
crystalline, with the main phase corresponding to orthorhombic
(MoO2)2P2O7. The XRD pattern of MoPO-V1 presents no indication
that vanadium ions were incorporated into the MoPO structure,
as the diffraction pattern was identical to the (MoO2)2P2O7 phase
without vanadium added. It is possible that vanadium could be
present as an amorphous phase, or as a highly dispersed separate
crystalline phase. The addition of a higher quantity of vana-
dium (MoPO-V5) resulted in additional reflections, visible at 15.5◦,
19.6◦ and 30.4◦ 2Â. These are attributed to the lattice planes
(0 0 1), (1 0 1) and (1 3 0) respectively of a VOHPO4·0.5H2O phase
[28–30]. A further increase in vanadium loading (MoPO-V10) not
only led to the formation of reflections assigned to (MoO2)2P2O7
and VOHPO4·0.5H2O phases, but additional reflections at 11.9◦
and 28.7◦ 2Â are attributed to the respective (0 0 1) and (1 0 1)
planes of VOPO4·2H2O [31]. High concentrations of vanadium
(MoPO-V20) led to the disappearance of reflections associated
with the VOHPO4·0.5H2O phase, and reflections only assigned to
VOPO4·2H2O and the main phase (MoO2)2P2O7 were present. The
clear formation of separate crystalline vanadium phosphate phases
confirm that solid solutions are not formed during catalyst synthe-
sis, instead the formation of biphasic catalysts is favoured.
Fig. 2. Laser Raman spectra: (a). MoHPO; (b). MoPO; (c). MoPO-V1; (d). MoPO-V5;
(e). MoPO-V10; (f). MoPO-V20.
forms zig-zagged chains built up of MoO6 octahedra, where each
octahedron shares two Mo-O vertices with other MoO6 octahe-
dra. The remaining three out of four vertices are shared with PO4
tetrahedra, which link together to form P2O7 groups [23,24].
The Raman spectra for both Mo-HPO and MoPO are shown in
Fig. 2. As presented, the main bands associated with Mo-HPO are
at 962 and 885 cm−1, and these are assigned to (PO4)3-
stretch-
sym
ing mode and (Mo
O Mo) stretching respectively [25]. Bands at
both 1142 and 1079 cm−1 can be related to the (PO4)3-
bonds,
asym
and a weak Raman band displayed at 1001 cm−1 is associated with
Mo O stretching. The main band present at 824 cm−1 in the spec-
trum of MoPO (Fig. 2b), is that of (Mo
O
Mo)asym stretching. A
Psym mode.
The Raman spectra of MoPO and promoted MoPO-Vx sam-
ples are shown in Fig. 2c–f. The analysis of MoPO-V1 provided
no evidence of vanadium species present, with the only bands
present attributed to the main phase, MoPO. MoPO-V5 and MoPO-
V10 both have very similar and complex spectra (Fig. 2d and e),
where some bands associated with MoPO are no longer present.
weak band present at 725 cm−1 is associated with P
O
Further weak bands observed at 860 and 1016 cm−1 are assigned to
symmetric and asymmetric Mo O modes respectively, with those
observed at 974 and 1154 cm−1 assigned to symmetric and asym-
metric (PO4)3− stretching respectively [25]. The Raman spectra
were consistent with the phases identified by XRD.
However in both cases, there is an extra band assigned to P
O
Table 1 shows the BET surface areas of both Mo-HPO and MoPO,
both had low surface areas of 1 m2 g−1. To observe the morphol-
ogy of each material SEM analysis was performed, and typical
images are presented in Fig. 3. The morphology of both Mo-HPO
and MoPO is very similar, and can be likened to the transforma-
tion of VOHPO4·0.5H2O to (VO)2P2O7, whereby the morphology of
the parent precursor is unchanged when forming the active cata-
lyst [26]. The precursor Mo-HPO has a rod-like morphology with
jagged edges (Fig. 3a), whilst after calcination to produce MoPO
more slightly rounded and more well-defined rods were formed
(Fig. 3b).
stretching (981 cm−1) in the VOHPO4·0.5H2O phase. This band is
dominant in VOHPO4·0.5H2O and is characteristically higher in fre-
quency than in (VO)2P2O7 [32]. The bonding of V O to structural
water in VOHPO4·0.5H2O increases the polarizability of the bond
[31], which gives rise to V
O .
P bands at 1109 and 1154 cm−1
Although these can not be observed clearly in the spectra, the
band at 1154 cm−1 assigned to (PO4)3-
stretching in the MoPO
asym
phase [25], is no longer present, but a broad low intensity band is
evident between 1120 and 1155 cm−1. The spectra of MoPO-V20
(Fig. 2f) presents bands, which are not only related to the main
(MoO2)2P2O7 phase, but also to those of VOPO4.2H2O, consistent
Table 1
Physico-chemical properties of unpromoted and V promoted MoPO materials.
TPR
XPS binding energy (eV)
XPS metal surface
concentration (At. %)
Material
SBET (m2 g−1
)
Reduction
peak Tem-
peratures
(◦C)
H2 consumption
Mo(3d)
V(2p3/2
)
O(1s)
Mo
V
(␮mol g−1
)
MoHPO
MoPO
MoPO-V1
MoPO-V5
MoPO-V10
MoPO-V20
1
1
1
1
1.5
3
557
579
571
571
490, 565
490, 571
1422
1646
1693
1746
1806
1923
233.6
233.8
233.6
233.6
233.8
233.6
–
–
513.9
531.9
531.7
531.8
531.9
531.6
9.2
7.5
9.2
8.9
8.4
4.2
–
–
0.2
0.3
1.1
4.2
517.5
517.2, 518.8 (70:30)
517.2, 518.8 (17:83)
518.7

54
G.T. Whiting et al. / Applied Catalysis A: General 485 (2014) 51–57
Fig. 3. Scanning electron microscopy images: (a). MoHPO; (b). MoPO; (c). MoPO-V1; (d). MoPO-V5; (e). MoPO-V10; (f). MoPO-V20.
with results observed in XRD (Fig. 1f). The introduction of a band
at 1035 cm−1 (V
O P stretching mode) accompanied by a broad
band at 925 cm−1 (assigned to (PO4)3-sym, and it is shifted from
940 cm−1), is characteristic of the VOPO4·2H2O phase [33]. As in
the spectra of both MoPO-V5 and MoPO-V10, the band observed
at 1154 cm−1 is assigned to V O, which occurs due to the iso-
lated vanadyl octahedra sharing an equatorial oxygen with a PO4
tetrahedra and forms V O···V O chains in a perpendicular direc-
tion [32]. The appearance of coupled V-O and P-O bending modes in
the VOPO4·2H2O can also be observed at 574 and 533 cm−1 respec-
tively.
The BET surface areas of the V promoted MoPO materials
are shown in Table 1. As seen previously for the unpromoted
(MoO2)2P2O7, the surface areas are all very low (1–3 m2 g−1),
although there is a trend where increasing V content slightly
increases the surface area. The morphology of V promoted MoPO
materials were studied by SEM (Fig. 3c–f). Possibly due to the
low quantity of V content in both MoPO-V1 and MoPO-V5, the
rod shaped morphology of (MoO2)2P2O7 was maintained, with
no characteristic VPO morphology present. This finding, accom-
panied by the absence of reflections and bands in both XRD and
Raman spectroscopy, respectively, related to vanadium phosphate
Fig. 4. Temperature programmed hydrogen reduction profiles: (a). MoHPO; (b).
MoPO; (c). MoPO-V1; (d). MoPO-V5; (e). MoPO-V10; (f). MoPO-V20.

G.T. Whiting et al. / Applied Catalysis A: General 485 (2014) 51–57
55
(for MoPO-V1), suggests it is highly dispersed. The increased con-
tent of 10 and 20 mol % V to MoPO visually confirms the presence
of separate VPO phases to that of (MoO2)2P2O7 (Figs. 3e and f),
as the characteristic platelet morphology of VOHPO4·0.5H2O and
VOPO4·2H2O was clearly observed [34].
Due to the significantly low surface areas of these biphasic V pro-
moted MoPO materials, TPR analysis was used to investigate their
bulk reducibility, presented in Fig. 4c-f. The reduction behaviour
of both MoPO-V1 and MoPO-V5 (Figs. 4c and d) are similar to that
of MoPO (Fig. 4b), confirming that vanadium in both materials is
highly dispersed on the surface (as a separate phase), consistent
with our previous characterization. A single but markedly broad
reduction peak is present in each profile, and it is attributed to
Mo6+ → Mo4+ reduction, however, the increasing V content slightly
lowered the peak maximum temperature (Table 1). Taking into
consideration the H2 consumption, it is clear that the increasing
presence of a vanadium phosphate phase enhances the overall
reducibility of the material. The peak temperature of Mo6+ → Mo4+
reduction in the TPR profile of MoPO-V10 (Fig. 4e) is displayed at
565 ◦C, and at 571 ◦C for MoPO-V20 (Fig. 4f). Although the intensity
of the main Mo reduction peak is lower for both MoPO-V10 and
MoPO-V20 than the other materials, there is also a reduction peak
present at a lower temperature of 490 ◦C, which is assigned to the
reduction of V5+ → V4+ [35]. The H2 consumption (Table 1) for both
MoPO-V10 and MoPO-V20 is substantially increased compared to
that of MoPO, and for MoPO containing lower quantities of V.
X-ray photoelectron spectroscopy results for MoPO and V pro-
moted catalysts are summarised in Table 1. As expected, MoPO
presents a Mo(3d) peak at 233.8 eV, which corresponds to Mo6+
species [36,37], and a P(2p) peak at 134.4 eV, attributed to sur-
face P5+ [38]. Even though the detection of a VPO phase was not
observed using other characterisation techniques for MoPO-V1,
XPS results show that a V(2p3/2) peak at 517.5 eV (0.18 at. %) is
attributed to surface V4+, which could be associated with the phase,
VOHPO4·0.5H2O. The XPS results recorded for MoPO-V5 presents
Fig. 5. Performance of catalysts for methanol oxidation; A: Methanol conversion,
B: Formaldehyde selectivity:
= MoPO-V10: = MoPO-V20.
= MoPO:
= MoPO-V1:
= MoPO-V5:
the presence of a biphasic material. Specifically, VOHPO4·0.5H2O
exists in accordance with the main MoPO phase. Catalytic testing of
VOHPO4·0.5H2O during methanol oxidation (under identical reac-
tion conditions) is displayed in Table 3. In comparison with the
pure MoPO catalyst (5.3% conversion), the VOHPO4·0.5H2O catalyst
shows higher conversion at this temperature (8.6%). Therefore, it is
reasonable to suggest that the increase in activity of the biphasic
MoPO-V5 catalyst at low temperatures is due to the presence of
the vanadium phosphate phase. In correlation with this postula-
tion, the increase in conversion with MoPO-V10 and MoPO-V20
catalysts can be attributed to the presence of VOPO4·2H2O, which
achieves 9.6% conversion at 200 ◦C (Table 3).
V
4+ species, which again corresponds to the VOHPO4·0.5H2O phase
observed by XRD and Raman spectroscopy. However, there was also
the presence of a second peak at 518.8 eV, which can be assigned
to V5+ [39]; the ratio of V4+:V5+ was 70:30 (0.33 at. %). MoPO-V10
also showed both surface V4+ and V5+ species, but in a 17:83 ratio
(1.10 at. %), which again corresponds well with the XRD results
where both VOHPO4·0.5H2O and VOPO4·2H2O diffraction patterns
were observed. MoPO-V20 displayed only surface V5+ (4.18 at. %),
associated with the VOPO4·2H2O bulk phase identified by previous
characterization.
3.2. Catalyst performance for methanol oxidation
Focusing on product selectivity at these low conversions, each
catalyst achieves formaldehyde selectivity above 98%. However,
even at these low conversions, there are subtle selectivity changes
towards each product as the V content is increased in the MoPO
catalysts. For the unpromoted MoPO catalyst, the only product pro-
duced at 200 ◦C was formaldehyde, which indicates the presence of
redox sites on the surface, which are reported as essential for high
formaldehyde selectivity [40]. At 200 ◦C, MoPO-V1 produced 99%
formaldehyde selectivity, but also 1% methyl formate selectivity,
which can be explained via the introduction of stronger basic sites
and reduced influence from redox sites than in the unpromoted
catalyst. Increasing basicity increases the time formaldehyde is
adsorbed on the surface, leading to over oxidation to methyl for-
mate [41]. As the vanadium content increases, there is a further
increase of methyl formate selectivity, with MoPO-V20 producing
2.1%, which suggests that the number of stronger basic/redox sites
is increasing with vanadium content.
Blank reactions carried out in an empty reactor tube only
showed very low methanol conversion. Trace conversion was
observed initially at 200 ◦C and conversion increased marginally
with temperature, but it was <1% at 500 ◦C. For the blank reaction
the only reaction product was formaldehyde. Selective oxidation
of methanol was conducted over the unpromoted and V promoted
MoPO catalysts over a temperature range 200– 480 ◦C (Table 2). As
the reaction temperature was increased, all of the catalysts pro-
duced significantly high selectivity towards formaldehyde, even at
high conversion of methanol. Fig. 5 shows both methanol conver-
sion and formaldehyde selectivity over the reaction temperature
profile, clearly presenting the enhanced activity of V promoted
catalysts compared to the unpromoted MoPO catalyst. Table 2
displays the conversions of each catalyst at 200 ◦C, where a min-
imal increase in methanol conversion with increasing V content is
reported. Although these are subtle enhancements, these can be
correlated with the findings from characterization studies. In the
case of MoPO-V5, XRD and Raman spectroscopy results confirmed
Comparing catalytic performance at 400 ◦C provides a clearer
indication of the enhanced activity and selectivity of vanadium pro-
moted catalysts. Methanol conversion of 50.8% was obtained with

56
G.T. Whiting et al. / Applied Catalysis A: General 485 (2014) 51–57
Table 2
Catalytic performance of unpromoted and V promoted MoPO catalysts in methanol selective oxidation.
Selectivities (%)
Catalyst
MoPO
Temperature (◦C)
MeOH Conv. (%)
DME
MF
FA
CO
CO2
FA Yield
(mol %)a
200
400
480
200
400
480
200
400
480
200
400
480
200
400
480
5.3
–
1
–
–
–
–
–
1
–
–
–
–
–
–
–
–
–
–
1
–
–
1.4
–
–
1.6
–
–
2.1
–
–
100
91.4
89.5
99
93.2
88
99.6
92.8
81.9
99.4
93.8
81.7
97.9
95.8
82.5
–
–
–
–
–
–
1.1
–
–
1.7
–
–
2.1
–
–
1.5
5.3
50.8
95.8
5.4
71.4
99.5
7.7
62.5
96.4
8
70
99.5
6.4
77.6
99.4
7.6
10.5
–
6.8
10.9
–
7.2
16.4
–
6.2
16.2
–
46.4
85.7
5.3
66.5
87.6
7.7
58.0
78.9
7.9
65.7
81.3
6.3
MoPO-V1
MoPO-V5
MoPO-V10
MoPO-V20
7.6
16
74.3
82.0
a
Formaldehyde per pass yield (mol%).
MoPO, whereas the incorporation of only 1 mol % V increased the
conversion to 71%, whilst both unpromoted and promoted cata-
lysts achieved 91.4 and 93.2% formaldehyde selectivity respectively
(Table 2). The detection of VPO phases in materials with higher
quantities of vanadium, accompanied by XPS data (Table 1), could
indicate that a VPO phase is present in MoPO-V1 but is highly
dispersed, which could account for its absence using bulk char-
acterization techniques, and could also explain the high activity
of this catalyst. MoPO-V5 also produced higher activity than the
unpromoted catalyst with 63% methanol conversion. The increase
in conversion compared with MoPO, can be attributed to the pres-
ence of VOHPO4·0.5H2O (Fig. 1), which alone gave 84.3% conversion
at 400 ◦C (Table 3). However, the conversion reached with MoPO-
V5 in contrast to MoPO-V1 is lower, which could be due to the
lower dispersion of the VPO phase. MoPO-V10 exhibited similar
activity to MoPO-V1 with 70% conversion at the same reaction
temperature, which indicates that there is not a simple relation-
ship between vanadium content and activity. Relating to bulk
characterization of this sample, we can associate these catalytic
features with the presence and dispersion of mixed phases, specifi-
cally, VOHPO4·0.5H2O, VOPO4·2H2O and (MoO2)2P2O7, identified
using XRD (Fig. 1). Although the use of VOPO4·2H2O as a cata-
lyst for methanol oxidation at 400 ◦C is poor in comparison to
MoPO (Table 3), its dispersion on the surface of MoPO and/or
VOHPO4·0.5H2O could enhance its catalytic properties. This pos-
tulation is exaggerated further with MoPO-V20, which achieves
78% methanol conversion and 96% formaldehyde selectivity. XPS
surface composition analysis of this biphasic catalyst (Table 1), dis-
plays a dramatic increase in the ratio of V species at the surface,
equalling the ratio of Mo species. This indicates a greater disper-
sion of the VPO phase near the surface, which in turn enhances
catalytic activity.
The other by-products produced during the oxidation reaction
were DME and CO. Although in relatively low quantities, both MoPO
and MoPO-V5 produce 1% selectivity to DME, which indicates the
presence of acidic as well as basic and redox sites on the surface
[42]. The CO selectivity obtained for each V promoted catalyst at
high reaction temperatures (480 ◦C) also reinforces the postulation
that the introduction of vanadium phosphate phases heavily influ-
ences the catalytic properties of these mixed phase catalysts. The
CO selectivity at this temperature obtained with VOHPO4·0.5H2O
is 71%, in contrast to 10.5% achieved with MoPO, both at >98%
methanol conversion. XRD and Raman spectroscopy both pro-
vided evidence that MoPO-V5 and MoPO-V10 were mixed with
VOHPO4·0.5H2O. Considering that both these catalysts obtained
the highest CO selectivity (16.4 and 16.2%, respectively) of all un-
promoted and V promoted MoPO catalysts, the influence of the
vanadium phosphate phases is clear. No evidence for catalyst deac-
tivation was observed, however, the time-on-stream was relatively
short, and to assess the full potential of these catalysts long-term
stability studies need to be performed.
Comparing the activity of the most commonly reported cata-
lysts for methanol oxidation to formaldehyde (Table 3), the MoPO
and MoPO-Vx catalysts obtain moderate activity. The commercial
(Fe2(MoO4)3) catalyst achieves substantially higher activity than
that of the MoPO-based catalysts reported here, as total methanol
conversion was reached at ∼300 ◦C for Fe2(MoO4)3, compared
to ∼480 ◦C for MoPO catalysts. However, both types of catalyst
Table 3
Catalytic performance of alternative selective methanol oxidation catalysts.
Selectivities (%)
Catalyst
Temperature (◦C)
MeOH
Conv. (%)
DME
MF
FA
CO
–
2.7
7.8
–
CO2
FA Yield (mol%)
Fe2(MoO4)3
200
250
300
200
400
450
200
400
480
200
400
480
6.2
50.6
99.3
1.5
85
99.8
8.6
84.3
95.7
9.6
40.6
99.5
–
1.8
–
100
–
–
–
1.2
–
–
1.4
–
–
100
95.5
92.2
–
91.3
86
–
–
–
–
1.4
4.2
–
1.2
–
6.2
48.3
91.5
–
77.6
85.8
8.6
63.1
27.8
9.6
40.0
78
MoO3
–
2.6
–
4.7
9.8
VOHPO4·0.5H2O
VOPO4·2H2O
––
100
–
74.8
29.0
100
98.6
78.4
22.8
71.0
–
–
13.3
–
–
–
–
–
8.3
–

G.T. Whiting et al. / Applied Catalysis A: General 485 (2014) 51–57
57
demonstrate initial activity at the same temperature (ca. 200 ◦C).
References
Considering the activity of MoO3 in relation to MoPO-Vx provides
a clear indication of the potential of these molybdenum phosphate
catalysts, as at a reaction temperature of 400 ◦C, there is comparable
activity and formaldehyde selectivity.
[1] H.R. Gerberich, A.L. Stautzenberger, W.C. Hopkins, Formaldehyde, in: Kirk-
Othmer Encyclopaedia of Chemical Technology, John Wiley and Sons, New
York, 1980, pp. 231–250.
[2] G. Reuss, W. Disteldorf, O. Grundler, A. Hilt, Formaldehyde, in: Ullmann’s Ency-
clopaedia of Industrial Chemistry, VCH Publishers, 1992, pp. 619–620.
[3] J.R. Fair, R.C. Kmetz, Formaldehyde, in: J.J. McKetta, W.A. Cunningham (Eds.),
In Encyclopaedia of Chemical Processing and Design, Marcel Dekker Inc., 1985,
pp. 350–351.
[4] A.R. Chauvel, P.R. Curty, R. Maux, C. Petitpas, Hydr. Proc. 52 (1973) 179–180.
[5] A.B. Stiles, T.A. Koch, Oxidation Catalysts, in: Catalyst Manufacture, Marcel
Dekker, New York, 1995, Chap20.
[6] A. Andreasen, H. Lynggaard, C. Stegelmann, P. Stoltze, Appl. Catal. A. Gen. 289
(2005) 267–273.
Analysing the physico-chemical properties of both the unpro-
moted and the promoted MoPO catalysts, and relating to known
catalyst systems, some structure-activity relationships can be
determined. It is known that related molybdenum oxide catalysts
operate by a Mars-van Krevelen mechanism. The first step of the
reaction has been found to be the formation of methoxy groups with
a surface cation, by dissociative adsorption of methanol on a dual
acid–base site, with a surface oxygen ion providing the basic site
[42,43]. The consecutive transformation of the adsorbed methoxy
group, not only depends on the acid strength of the site to which
it is adsorbed, but also to the nature of neighbouring active cen-
tres. It is widely accepted that the rate-limiting step is C H bond
dissociation, and it is determined on the basic or nucleophilic char-
acter of oxygen species in close proximity to the adsorbed methoxy
group [44–46]. The use of (MoO2)2P2O7 as a methanol oxidation
catalyst is reported here for the first time, and much remains to
be understood. However, considering some of the structural and
chemical similarities to other known effective methanol selective
oxidation catalysts, such as MoO3 [47] and Fe2(MoO4)3 [48], it is
reasonable to assume it operates in a similar manner. For MoO3
and Fe2(MoO4)3 it is reported that the main active centre consists
of terminal Mo O, where the oxidation state of Mo plays an impor-
tant role in the activity of the catalyst, with Mo6+ being favourable
to Mo5+. As the oxidation state of molybdenum in MoPO catalysts
is Mo6+ and Raman spectra features Mo O bands at 860, 1001 and
1016 cm−1, the active site and mechanism of methanol oxidation
may be suggested to be similar to MoO3 and Fe2(MoO4)3 catalysts.
Interestingly, in the fresh MoPO catalysts only one surface
species of oxygen is present at 531.9 eV (Table 1), which is assigned
to that of lattice oxygen, however, in the ex-reactor sample, two
species of oxygen are present at 531.9 and 533.1 eV. The presence
of the binding energy peak at 533.1 eV, is attributed to OH− groups,
which suggests that un-reacted methoxy groups are present on the
surface [49], or that the surface could be hydroxylated.
[7] M. Bowker, R. Holroyd, A. Elliot, P. Morrall, A. Alouche, C. Entwistle, A. Toern-
crona, Catal. Lett. 83 (3–4) (2002) 165–176.
[8] G. Reuss, W. Disteldorf, A.O. Gamer, A. Hilt, Ullmann’s Encyclopaedia of Indus-
trial Chemistry, Wiley-VCH, Weinheim, 2008, pp. 619–652.
[9] B. Crichton in Informally Speaking (Newsletter from Perstorp Formox,
http://www.perstorpformox.com), spring/summer 2003, p. 12-13.
[10] Y. Uebou, S. Okada, J. Yamaki, J. Power Sources 115 (2003) 119–124.
[11] D. Bhattacharyya, S.K. Bey, M.S. Rao, Appl. Catal. A. Gen. 87 (1992) 29–43.
[12] M.A. Chaar, D. Patel, H.H. Kung, J. Catal. 109 (1988) 463–467.
[13] D.S.H. Sam, V. Soenen, J.C. Volta, J. Catal. 123 (1990) 417–435.
[14] R.H.H. Smits, K. Seshan, J.R.H. Ross, Stud. Surf. Sci. Catal. 72 (1992) 221–229.
[15] A. Cherrak, R. Hubaut, Y. Barbaux, G. Mairesse, Catal. Lett. 15 (1992)
377–383.
[16] G. Costentin, L. Savary, J.C. Lavalley, M.M. Borel, A. Grandin, Chem. Mater. 10
(1998) 59–64.
[17] S. Nianxue, Z. Xuan, J.K. Bartley, S.H. Taylor, D. Chadwick, G.J. Hutchings, Catal.
Lett. 106 (2006) 3–4.
[18] Y.H. Taufiq-Yap, C.K. Goh, G.J. Hutchings, N. Dummer, J.K. Bartley, Catal. Lett.
130 (2009) 327–334.
[19] P. Bonnet, J.M.M. Millet, C. Leclercq, J.C. Vedrine, J. Catal. 158 (1996) 128–141.
[20] J.M.M. Millet, J.C. Vedrine, G. Hecquet, Stud. Sci. Catal. 55 (1990) 833–841.
[21] J.C. Vedrine, Top. Catal. 11/12 (2000) 147–152.
[22] P. Kierkegaard, Acta. Chem. Scand. 12 (1958) 1701–1703.
[23] P. Kierkegaard, Ark. Kemi. 19 (1962) 1–14.
[24] S.E. Lister, A. Soleihavoup, R. Withers, P. Hodgkinson, J. Evans, Inorg. Chem. 49
(2010) 2290–2301.
[25] D. Boudlich, L. Bih, M.E.H. Archidi, M. Haddad, A. Yacoubi, A. Nadiri, B. Elouadi,
J. Am. Ceram. Soc. 85 (2002) 623–630.
[26] N. Duvauchelle, E. Kesteman, F. Oudet, E. Bordes, J. Solid State Chem. 137 (1998)
311–324.
[27] X. Zhang, H.-L. Wan, W.-Z. Weng, X.-D. Yi, Appl. Surf. Sci. 220 (2003) 117–124.
[28] L. O’Mahony, J. Henry, D. Sutton, T. Curtin, B.K. Hodnett, Appl. Catal. A: Gen. 253
(2003) 409–416.
[29] L. O’Mahony, D. Sutton, B.K. Hodnett, Catal. Today 9/92 (2004) 185–189.
[30] L. O’Mahony, T. Curtin, J. Henry, D. Zemlyanov, M. Mihov, B.K. Hodnett, Appl.
Catal. A: Gen. 285 (2005) 36–42.
[31] Y.C. Wong, Y.H. T-Yap, Asian J. Chem. 23 (2011) 3853–3858.
[32] V.V. Guliants, J.B. Benziger, S. Sundaresan, I.E. Wachs, J.-M. Jehng, J.E. Roberts,
Catal. Today 28 (1995) 275–295.
4. Conclusions
[33] J.K. Bartley, C.J. Kiely, R.P.K. Wells, G.J. Hutchings, Catal. Lett. 72 (2001)
99–105.
[34] L. Griesel, J.K. Bartley, R.P.K. Wells, G.J. Hutchings, J. Mol. Catal. A: Chem. 220
(2004) 113–119.
[35] P. Concepcion, J.M.L. Nieto, J. P-Pariente, J. Mol. Catal. A: Chem. 99 (1995)
173–182.
[36] L.E. Firment, A. Faretti, Surf. Sci. 129 (1983) 155–176.
[37] T.H. Fleisch, G.J. Mains, J. Chem. Phys. 76 (1982) 780–786.
[38] A. Medina, J.L. Solis, J. Rodriguez, W. Estrada, Sol. Energy Mater. Sol. Cells 80
(2003) 473–481.
The structural and chemical properties of novel V promoted
MoPO catalysts were studied, and their catalytic behaviour inves-
tigated for the partial oxidation of methanol to formaldehyde. All
the catalysts achieved high formaldehyde selectivity at relatively
high methanol conversion, but high reaction temperatures were
needed in comparison to the commercial ferric molybdate cata-
lyst. The increasing V content in MoPO catalysts led to a substantial
enhancement in activity compared to that of the unpromoted
catalyst, whilst obtaining high formaldehyde selectivities. It was
observed that varying the V content, formed two additional VPO
phases. Addition of low quantities of V led to the formation of
VOHPO4·0.5H2O, with higher V content leading to VOPO4·2H2O for-
mation. The dispersion of these VPO phases at the MoPO surface,
were key to the enhanced activity of these mixed phased catalysts.
The encouraging activity of MoPO and MoPO/VPO mixed phase cat-
alysts towards methanol oxidation warrants further investigation.
[39] M.P. Casaletto, S. Kaciulis, L. Lisi, G. Mattogno, A. Mezzi, P. Patrono, G. Ruoppolo,
Appl. Catal. A: Gen. 218 (2001) 129–137.
[40] J.M. Tatibouet, Appl. Catal. A: Gen. 148 (1997) 213–252.
[41] G. Busca, J. Mol. Catal. 50 (1989) 241–249.
[42] M. Ai, J. Catal 54 (1978) 426–435.
[43] N. Pernicone, F. Lazzerin, G. Liberti, G. Lanzavecchia, J. Catal. 14 (1969) 293–302.
[44] T.J. Yang, J.H. Lunsford, J. Catal. 103 (1987) 55–64.
[45] C.J. Machiels, A.W. Sleight, J. Catal. 76 (1982) 238–239.
[46] W.E. Farneth, F. Ohuchi, R.H. Staley, U. Chowdhry, A.W. Sleight, J. Phys. Chem.
89 (1985) 2493–2497.
[47] J.M. Jehng, H. Hu, X. Gao, I.E. Wachs, Catal. Today 28 (1996) 335–350.
[48] M. Bowker, R. Holroyd, M. House, R. Bracey, C. Bamroobgwongdee, M. Shannon,
A. Carley, Top. Catal. 48 (2008) 158–165.
[49] J. Haber, E. Lalik, Catal. Today 33 (1997) 119–137.
Acknowledgements
We would like to thank the EPSRC and Cardiff University for
financial support.