F. Cavani et al.
H3PO4 (Sigma–Aldrich, purity ꢁ98%, either 10.94 g [P/V=1.00] or
13.13 g [P/V=1.20]) in isobutanol (120 mL, Sigma–Aldrich purity
ꢁ99.5%) for 6 h (T=1108C). After filtration, the light blue precipitate
was first washed with isobutanol (20 mL) and acetone (20 mL), to
remove the excess phosphoric acid. The washed precipitate was then
dried at 1208C for 12 h, in static air, and lastly thermally treated accord-
ing to the following procedure: a) pre-calcination step in flowing air, with
temperature gradient from room temperature up to 3008C; then isother-
mal step at 3008C in air for 6 h; b) thermal treatment in flowing N2, with
temperature gradient up to 5508C, and final isothermal step at the latter
temperature for 6 h; and c) equilibration under reaction conditions for
100 h. Equilibration leads to various transformations in calcined V/P/O
catalysts, including: 1) the crystallization of amorphous components, if
any were present in the fresh catalysts, to VPP; 2) the reduction of bulk
VOPO4 to VPP; 3) the oxidation of VIII phosphates to VPP; and 4) the
increase in the crystallinity of VPP.
Conclusions
The effect of the P/V atomic ratio on the nature of the com-
pounds formed on the surface of VPP, both in the presence
of air and dependent on temperature, was investigated by
means of in situ Raman spectroscopy. It was found that a
dynamic equilibrium between various compounds is reached
on the VPP surface. In the intermediate T range (340–
4008C) the main components are aI-VOPO4 and d-VOPO4.
The former develops with the catalyst having the stoichio-
metric P/V bulk ratio; the latter with the catalyst having a
slight excess of P, but only in the presence of steam. For
both catalysts, however, the prevailing compound at high
temperature (400–4408C) is d-VOPO4. Moreover, in P/V1.0
the transformation of aI-VOPO4 into d-VOPO4 is reversible:
the two compounds interconvert.
These results were used to interpret the catalytic behavior
observed experimentally: under conditions giving the pre-
vailing formation of d-VOPO4, both catalysts offered the op-
timal catalytic performance, with high selectivity towards
MA and moderate activity in n-butane conversion. With the
stoichiometric catalyst and at intermediate reaction temper-
ature, that is, under conditions leading to the formation of
aI-VOPO4, the selectivity towards MA was very low, where-
as the catalytic activity was very high (Figure 9).
Raman spectroscopy: In situ Raman analyses were performed using a
Renishaw 1000 instrument equipped with a Leica DMLM microscope,
argon-ion laser source (514 nm), and a commercial Raman cell (Linkam
Instruments TS1500). A small amount of catalyst (1–5 mg) was loaded
into the ceramic crucible of the cell. First, the spectrum was recorded at
room temperature, then the temperature was increased up to the desired
value (heating rate 50 minÀ1), while recording spectra at intermediate
temperatures, and finally maintained in isothermal conditions for a few
hours at the final temperature. A continuous flow of GC-grade air or N2
(flow rate 50 Ncm3 minÀ1) was fed to the cell from the very beginning of
each experiment. A saturator allowed the feeding of steam-saturated air
or nitrogen; the change in the saturator temperature made it possible to
vary the partial pressure of water, from 0.03 to 0.10 bar. In situ spectra
were recorded for “equilibrated” catalysts (samples that had been pre-
liminarily treated in the reactive phase at 4008C for 100 h reaction time).
X-ray photoelectron spectra (XPS): XPS were recorded with a VG ES-
CALAB 220 XL spectrometer equipped with a monochromatic AlKa
(E=1486.6 eV) X-ray source, using the C 1s peak (285.0 eV) as a refer-
ence. Spectra were collected with a pass energy of 40 eV, using the elec-
tromagnetic lens mode low-energy electron flood gun (6 eV) for charge
compensation effect; O 1s/V 2p, C 1s, and P 2p core levels were recorded.
X-ray diffraction (XRD): XRD patterns were recorded using the Philips
PW 1710 apparatus, with CuKa (l=1.5406 ꢂ) as a radiation source.
Chemical analysis: The chemical analysis of samples was performed as
follows. The equilibrated sample was dissolved in concentrated fuming
sulfuric acid. Then its V content was determined by titration of VV with a
Mohr salt solution (FeII), and VIV was determined by titration with a
KMnO4 0.1 N solution. Phosphorus was determined gravimetrically as
quinoline molybdophosphate.
Figure 9. A summary of the effect of the P/V atomic ratio in VPP on
both the catalytic performance and the nature of the active layer.
Catalytic tests were carried out in a quartz continuous-flow reactor, load-
ing 0.8 g of catalyst and feeding 1.7% n-butane and 17% oxygen (re-
mainder N2). Overall GHSV was 2160 hÀ1. Products were analyzed on-
line by sampling a volume of the outlet gas stream and injecting into a
gas chromatograph equipped with an HP-1 column for the separation of
C4 hydrocarbons, formaldehyde, acetic acid, acrylic acid, and MA. A Car-
bosieve SII column was used for the analysis of oxygen, CO, and CO2.
The effect observed experimentally explains why the in-
dustrial catalyst for n-butane oxidation holds a slight excess
of P; it also explains the discrepancies found in the litera-
ture about the nature of the active layer in VPP. These dis-
crepancies can be attributed to small differences in the P/V
ratio of samples. Here, it is also proposed that the small
excess of P is adsorbed on the vanadyl orthophosphate
during the synthetic step, and, lastly, located in the defective
areas of mosaic crystals that form VPP, being responsible in
the end for differences in morphology observed experimen-
tally among samples with different P/V atomic ratios.
Acknowledgements
Prof. Elisabeth Bordes-Richard, Universitꢃ des Sciences et Technologies
de Lille, is acknowledged for XPS measurements.
[1] N. Ballarini, F. Cavani, C. Cortelli, S. Ligi, F. Pierelli, F. Trifirꢄ, C.
[2] P. Arpentinier, F. Cavani, F. Trifirꢄ, The Technology of Catalytic Ox-
idations, 2001, Editions Technip, Paris.
Experimental Section
Synthesis: For the synthesis of the VPP precursor, VOHPO4·0.5H2O, the
“organic procedure” was adopted. The precursor was obtained by heating
under reflux V2O5 (10.04 g, Sigma–Aldrich, purity 99.6%) and 100%
[4] J. C. Volta, C.R. Acad. Sci. Paris 2000, 3, 717.
1654
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 1646 – 1655