A.E. Lewandowska et al. / Journal of Catalysis 255 (2008) 94–103
95
that allowed us to measure IR spectra under dynamic condi-
tions.
and 15- to 35-kHz MAS spectra acquisition. The single pulse
sequence with rf-pulse length of 1 µs (12◦ flip angle) and recy-
cle delay of 1–5 s was used to acquire 27Al and 51V spectra. 1H
spectra were measured with pulse length of 10 µs (90◦) and re-
cycle delay of 1 s. The chemical shift values were referred to
TMS for 1H, 0.1 M Al(NO3)3 for 27Al, and VOCl3 for 51V.
27Al 3QMAS (Triple-quantum MAS) experiments were per-
formed using a basic 3-pulse sequence with z-filter. The se-
quence starts with an excitation pulse p1 (3.5 µs, correspond-
ing to 180◦) that creates 3Q coherence, which is allowed to
evolve during the evolution period, τ. A subsequent conver-
sion pulse, p2 (1.0 µs, corresponding to 50◦) flips magnetization
back along the z-axis, which, after a short (20 µs) delay (to al-
low dephasing of undesired coherences), is read out with a weak
CT-selective 90◦ pulse, p3 (6.5 µs).
We used multinuclear 27Al, 51V, and 1H solid-state NMR for
structural characterization of alumina-supported vanadia, nio-
bia, and vanadia–niobia catalysts. Today, 27Al and 51V solid-
state NMR are indispensable to the study of many alumina- and
vanadia-based catalysts and related materials. The coordination
number of Al sites in AlOx can be readily determined from the
value of the isotropic chemical shift [6,7]. 1H NMR of hydroxyl
groups is very informative regarding the extent of dispersion of
supported metal [8].
In this paper, we report on the preparation and characteriza-
tion of structural and reactive properties of alumina-supported
vanadia and niobia, and the effect of total V + Nb coverage on
alumina.
The samples were dehydrated before the NMR experiments.
The catalyst samples were placed into 7-mm glass ampules
and dehydrated under vacuum (<10−3 Torr) for 4 h at 623 K.
Then they were calcined for 2 h at 723 K in dry O2 and finally
flame-sealed. Just before the measurements, the ampules were
unsealed and the samples loaded into standard 5-mm rotors in
a dry box under an argon flow. Hydration of the samples was
achieved by exposing them to the humidity of the ambient air.
1H MAS NMR measurements were performed with dehydrated
samples; 27Al NMR measurements, with hydrated ones. 51V
MAS NMR was performed with both hydrated and dehydrated
samples.
2. Experimental
2.1. Preparation of samples
The vanadium and niobium oxides supported on alumina
were prepared using different V precursors. The VS series was
prepared using a VOSO4 precursor, and the VM series was
prepared using an ammonium metavanadate precursor. Vana-
dia on alumina catalysts were prepared from an aqueous solu-
tion of VOSO4 (Aldrich, 99.99%), which was maintained under
stirring at 323 K for 50 min, after which γ -Al2O3 (SASOL,
SBET = 193 m2/g) was added. The suspension thus obtained
was evaporated in a rotatory evaporator at 338 K. The result-
ing solid was dried at 388 K for 20 h and then calcined at
673 K for 4 h in air. The rate of heating was 5 K/min. The
impregnation of alumina-supported V–Nb and Nb oxides was
done using NH4VO3 (Sigma, 99.99%) and NH4NbO(C2O4)2
(Aldrich, 99.99%) solutions. Oxalic acid (Panreac, 99.5%) was
added to an aqueous solution of ammonium metavanadate and
of ammonium niobate (V) oxalate to facilitate dissolution of
salts. The same niobium precursor was applied in the prepa-
ration Nb/γ -Al2O3 system. Oxalic acid also was used for the
impregnation of niobium on alumina. As a reference value, the
monolayer value (understood to be the dispersion limit load-
ing of the supported oxides) was estimated as a total number
of 8 atoms (V + Nb)/nm2 of alumina support. The amounts of
V, Nb, and V + Nb were calculated so that the total coverage
of metals ranged from ca. 1/4 of a monolayer to above mono-
layer. The V/Nb atomic ratio was kept at 1. The general xVS/Al
(or xVM/Al), xNb/Al and xVNb/Al nomenclature was applied,
where x indicates the number of atoms per nm2 of V and Nb, S
represents vanadyl sulfate, and M represents ammonium meta-
vanadate.
2.3. IR study
The IR experiments were performed with Vector 22 (Bruker)
FTIR spectrometer. IR spectra were recorded at a resolution of
2 cm−1 using 64 scans. The ∼20 mg catalyst samples were
pressed into thin wafers and placed in the in situ flow cell
equipped with KRS-5 windows (insensitive to humidity). The
activation, adsorption, and oxidation of methanol were car-
ried out in this in situ flow cell. Spectra were registered in
a temperature range of 373–573 K. The spectra without any
sample (“gas phase”) were scanned before each catalyst spec-
trum. Before the measurement, the samples were activated at
623 K in a 50 cm3/min flow of oxygen and helium, which had
been passed through a molecular sieve trap to remove mois-
ture traces. The O2/He molar ratio was 20/80. The pretreated
materials (2 h) were gradually cooled to 373 K. Adsorption of
methanol was performed at 373 K in a ∼40 cm3/min stream
with a CH3OH/He molar ratio 1.2/98.8. The methanol partial
pressure was controlled by a methanol bubbler. IR spectra were
recorded after 15 and 35 min of methanol adsorption. Methanol
oxidation was carried out using gas mixture of CH3OH/O2/He
(molar ratio 1.2/18.8/80) at a flow rate of ca. 40 cm3/min. The
temperature was gradually increased up to 573 K. IR spectra
were obtained at given temperatures after 15 min of dwell time.
2.2. NMR study
Solid-state NMR experiments were performed with Bruker
AVANCE-400 (9.4 T) spectrometer at resonance frequencies
400.13, 104.26, and 105.20 MHz for 1H, 27Al, and 51V, respec-
tively. A Bruker 4.0 mm, 2.5 mm MAS probes and NMR Rotor
Consult ApS (Denmark) 5 mm MAS probe were used for static
2.4. Methanol probe reaction
Methanol oxidation reaction was performed in a glass fixed-
bed reactor equipped with a thermocouple. First, 0.03 g of cat-