H. Benaissa et al. / Journal of Catalysis 253 (2008) 244–252
245
of methacrolein to methacrylic acid is utilised commercially
over these catalysts [7–9]. There is compelling evidence that
this reaction occurs via redox Mars–Van Krevelen mechanism,
with catalyst acidity (controlled by ion exchange, e.g., with Cs+
cations) playing an important role [4,7–9]. To our knowledge,
no attempt to use these HPAs as hydrogenation catalysts has
been made so far.
Here we report that the H3+n[PMo12−nVnO40] acids (n =
0–2) and their Cs+ salts are active catalysts for the vapour-
phase hydrogenation of propanoic acid. It is demonstrated that
the performance of these catalysts in hydrogenation bears close
resemblance to that in selective oxidation.
measured in a hydrogen flow (25 mL/min), ramping the tem-
perature from room temperature to 400 ◦C at a rate of 5 ◦C/min.
Thermal stability of heteropoly compounds was estimated us-
ing a Setaram TG-DSC 111 differential scanning calorimeter
from exothermic peak corresponding to the formation of oxides.
Ammonia adsorption measurements were performed using the
Setaram TG-DSC 111 instrument at 100 ◦C by pulse method.
Catalyst samples were pretreated at 350 ◦C/1 h in dry He flow
(30 mL/min). Then the temperature was lowered to 100 ◦C. Af-
ter stabilisation, successive 2 mL-pulses of gaseous NH3 were
injected in the He flow using a loop fitted in a 6 port valve. Cat-
alyst reduction was performed at 340 ◦C under hydrogen flow
using the Setaram TG-DSC 111 instrument. In a typical experi-
ment, a catalyst sample (30 mg) was packed into the TGA-DSC
crucible. Prior to reduction, the sample was heated under N2
flow (40 mL/min) up to 350 ◦C at a rate of 5 ◦C/min and held
for 1 h at that temperature. Then the temperature was decreased
to 340 ◦C and stabilised for 30 min. This operation was to pre-
vent any additional water loss because of the slight heating of
the sample caused by exothermic reduction. After temperature
stabilisation, reduction was commenced by switching nitrogen
flow to a mixture of H2 and N2.
2. Experimental
2.1. Chemicals and catalysts
H3[PMo12O40] hydrate (PMo) and propanoic acid were
purchased from Aldrich. Heteropoly acids H4[PMo11VO40]
(PMoV) and H5[PMo10V2O40] (PMoV2) were prepared as
crystalline hydrates by refluxing stoichiometric amounts of
MoO3, V2O5 and H3PO4 in water for 48 h to yield orange solu-
tions of HPAs, followed by removing water in a rotary evapora-
tor and drying the solids in an oven at 110 ◦C overnight [10,11].
Cs salts of these HPAs, CsxH3−x[PMo12O40] (CsxPMo),
CsxH4−x[PMo11VO40] (CsxPMoV) and CsxH5−x[PMo10-
V2O40] (CsxPMoV2), were prepared using the Izumi’s proce-
dure for the preparation of CsxH3−x[PW12O40] [12] by adding
dropwise 0.25 M aqueous solution of Cs2CO3 to an aqueous so-
lution of HPA at 70 ◦C, followed by ageing the aqueous slurry
overnight at room temperature. Water was rota-evaporated at
50 ◦C to afford solid Cs salt which was finally dried in an oven
at 110 ◦C overnight.
2.3. Catalyst testing
The hydrogenation of propanoic acid was carried out in a
Pyrex glass fixed-bed reactor (9 mm internal diameter) with
on-line GC analysis (Varian Star 3400 CX gas chromatograph
equipped with a 30 m × 0.25 mm HP INNOWAX capillary
column and flame ionisation detector). The reactor was placed
in a vertical tubular furnace and fed from the top. The reac-
tion temperature was controlled by a Eurotherm controller us-
ing a thermocouple placed in the centre of the catalyst bed.
The hydrogenation of propanoic acid to propanal is slightly
endothermic (ꢀH0 = 28.3 kJ/mol), therefore temperature gra-
dients across the catalyst bed could be neglected. The reactor
was packed with 0.2 g catalyst powder. Prior to reaction, the
catalyst was pretreated in situ in a hydrogen flow (40 mL/min),
ramping the temperature to 400 ◦C at a rate of 10 ◦C/min then
dwelling at 400 ◦C for 2 h, unless otherwise stated. The hy-
drogenation of propanoic acid was carried out at 350 ◦C, 1 bar
pressure, 80 mL/min H2 flow rate and 2 vol% acid concen-
tration in the gas flow. The hydrogen flow was controlled by
Brooks mass flow controllers. Propanoic acid was supplied by
bubbling the H2 flow through a glass saturator containing the
acid, which was kept at a certain temperature to maintain the
acid concentration in the gas flow. All gas lines were made of
stainless steel. The gas lines after the saturator and sampling
valves were heated at 170 ◦C to prevent reactant and product
condensation. At regular time intervals, the downstream gas
flow was analysed by the on-line GC which was calibrated for
each product. The products were collected in an ice trap and
analysed by the off-line GC-MS. Propanoic acid conversion (%)
was calculated as X = 100 × ([acid]0 − [acid])/[acid]0, where
[acid]0 is the concentration of propanoic acid in the feed flow,
which was measured using a bypass line. Product selectivities
were calculated as the molecular percentage of all organic prod-
2.2. Techniques
The BET surface area, SBET, and porosity of catalysts were
measured by nitrogen physisorption at 77 K on a Micromerit-
ics ASAP 2000 instrument. Before the measurement, the sam-
ples were evacuated at 250 ◦C for 4–6 h. Water content in the
catalysts was measured by thermogravimetric analysis (TGA)
using a Perkin–Elmer TGA 7 instrument. Diffuse reflectance
IR spectra were recorded on a Nicolet NEXUS FTIR spectrom-
eter using powder mixtures of catalysts with KBr. 31P MAS
NMR spectra of catalysts were recorded at room temperature
and 4 kHz spinning rate on a Bruker Avance DSX 400 NMR
spectrometer in a 4 mm sample probe using 85% H3PO4 as
a reference. ICP-AES chemical analysis was carried out on
a Spectro Ciros CCD emission spectrometer. Powder X-ray
diffraction (XRD) patterns of catalysts were recorded on a
Siemens D-5000 diffractometer with a monochromatic CuKα
radiation in the angular range 10◦ ꢁ 2θ ꢁ 80◦, with a step width
of 0.036◦ and count-time of 1 s per step. XRD patterns were at-
tributed using JCPDS database. In situ XRD experiments were
carried out using the same diffractometer in a XRK900 re-
actor chamber (Anton Paar GmbH). The XRD spectra were