Paper
NJC
After calcination, the materials retained significantly high
surface areas since their values were about 3 times higher than
that of commercial titania (Degussa E 50 m2 À1).13 Nb
g
addition limited the decrease in the specific surface area
(Table 1). The decrease in the micropore population is respon-
sible for the diminution of specific surface areas, since at low
relative pressure in the N2 adsorption isotherms a decrease in
the nitrogen uptake was observed for all samples (Fig. S3, ESI†).
However the meso and macro porosities were retained after
calcination (Fig. 1c and d).
Powder XRD analyses were performed on the calcined
x%Nb–TiO2 materials. The X-ray diffraction patterns showed
the presence of the anatase TiO2 phase at 2y = 25.3, 38.0, 47.9,
54.0, 55.0, 62.7, 68.9, 70.2, 75.2 and 82.61 (JCPDF 84-1286)
without any Nb-based crystalline phase (Nb2O5 JCPDF 30-0873, Fig. 2 Relationship between the Nb/Ti ratio obtained from XPS analysis
and the Nb/Ti ratio obtained by ICP analysis.
2y = 28.4, 22.6, 36.6, 46.2, 50.9 and 55.11) (Fig. S4, ESI†). The low
Nb loading and/or a high dispersion of Nb species within a TiO2
framework could well explain these results. The structural
ToF-SIMS analyses for all hierarchically porous materials
parameters of x%Nb–TiO2 solids, obtained from XRD analyses,
resulted first in the detection of several TiO2 characteristic
are summarized in Table 1. TiO2 average crystallite size
fragments (Fig. S6, ESI†). No polymeric niobium fragments
decreased with Nb addition. This trend could be an indication
([NbxOy]+) were detected on the solid surface as can be observed
of Nb incorporation which could affect the anatase structure
over Nb–TiO2 samples prepared by an impregnation method.19
somehow. Rietveld refinement was done to estimate the cell
This result can be explained by the high Nb dispersion in the
parameters of the anatase lattice (Fig. S5, ESI†). The unit cell
hierarchically porous TiO2 materials, and thus the statistical
volume obtained from the lattice parameters a and c are close
improbability of forming Nb–O–Nb bonding at such low load-
in value and within the margin of error (Table 1). The fact that
ings. The high mass resolution of ToF-SIMS analyses also
the ionic radii are close in value (rTi4+ = 64 pm compared to
showed the presence of TixNbOy fragments. The detected
rNb5+ = 70 pm) coupled with the fact that doping levels are low
binary ions for 5%Nb–TiO2 in
a positive polarity were
suggest a possible incorporation of Nb species into the anatase
framework by a substitution mechanism of Ti4+ ions by Nb5+
ions.14,15
[TiNbO3]+, [Ti2NbO5]+ and [Ti3NbO7]+ fragments at 188.825
(188.838),‡ 268.876 (268.777)‡ and 348.680 (348.715)‡ u, respec-
tively (Fig. 3). The peak intensities of these fragments were
much more intense for the sample prepared via the self-
assembly method rather than for a mechanical mixture of
TiO2 and Nb2O5 oxides (in a ratio to obtain a 5% At Nb loading
w.r.t. Ti) (Fig. 3). This result indicated that the possible recom-
bination of single ions to give binary ions during the ToF-SIMS
analysis is very low. In the hierarchical Nb–TiO2 materials, the
relative ToF-SIMS intensity ratios of these binary ions were
clearly proportional to the Nb loading (Fig. 4). Since Ti–Nb
mixed fragments show a direct interaction between Ti and Nb
species, the formation of a Nb–TiO2 solid solution on the solid
surface can thus be confirmed by this ToF-SIMS analysis.
3.2. Surface characterisation
The results of the surface characterisation are shown in Table 2
and Fig. 2. Analysis of the O 1s core level showed the presence
of two components located at 530 eV and 532 eV which
correspond respectively to O2À species of the oxide network
(main component) and to oxygen ions with low coordination
À
states (O2 and OÀ), oxygen vacancies and oxygen atoms in
contaminant species (OHÀ and CO32À). The Ti 2p3/2 and Nb 3d5/2
binding energies were characteristic of the presence of Ti4+ and
Nb5+ ions respectively.16,17 The Nb/Ti ratio obtained by XPS was
very close to that obtained by ICP analysis and the corresponding
straight line was very close to the line obtained from the
Kerkhof–Moulijn model (Fig. 2).18
3.3. Acid–base characterisation
The NH3 desorption curves from the analysis of the peak
intensity of m/z = 17 are shown in Fig. 5. A very broad peak of
desorption was obtained, indicating that all the samples pos-
sess a wide distribution of acid site strengths. The temperature
at which the ammonia began to desorb is the same (125 1C),
independent of sample composition. However the desorption
peaks presented a maximum which increased with Nb content
Table 2 Binding energy values obtained from XPS analysis of calcined
x%Nb–TiO2 hierarchically porous materials
N 1sa (eV)
O 1s
(eV)
Nb 3d5/2 Ti 2p3/2
Sample
TiO2
1%Nb–TiO2 531.2 207.9
3%Nb–TiO2 530.1 207.7
5%Nb–TiO2 530.9 207.7
(eV)
(eV)
Lewis
Brønsted
530.3
—
458.9
458.8
458.9
458.9
399.7 (85%) 401.5 (15%)
399.7 (90%) 401.3 (10%) (Fig. 5). It is generally accepted that desorption temperature
399.8 (87%) 401.6 (13%)
399.8 (83%) 401.5 (17%)
can be related to the strength of the acid sites. An increase in
a
Binding energy from N 1s of the pre-adsorbed PEA molecule.
‡ Theoretical ion mass value.
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014
New J. Chem., 2014, 38, 1988--1995 | 1991