J. Mohd Ekhsan et al. / Applied Catalysis A: General 471 (2014) 142–148
147
should realize that the relatively low yield of 1,2-epoxyoctane using
P/Nb/Ti–Si was most probably due to the presence of strong acidic
active site in the sample, hence leading to effective conversion of
species (see Fig. 1) has the highest oxidative activity. Besides, it
was believed that absence of Nb has caused alteration in Ti species
formed due to the direct interaction between PO43 and TiO2 in
−
1
,2-epoxyoctane to 1,2-octanediol. Besides, the higher surface area
P/Ti–Si, bringing in the poorer oxidative catalytic performance.
A sample without PO43 was prepared in order to study its func-
tion as catalyst modifier. Surprisingly, the sample Nb/Ti–Si was a
weak catalyst in epoxidation reaction as it only produced 59 mol
1,2-epoxyoctane (Fig. 4(h)). The epoxy yield was even lesser than
−
in P/Nb/Ti–Si may have provided more active sites for the reaction,
subsequently contributing to its excellent catalytic performance.
Obviously, different synthesis procedures have produced samples
of varied physical–chemical properties and subsequently resulted
in diverse catalytic activities. The proposed structure of Lewis and
Brønsted acid sites in the samples is illustrated in Fig. 5.
Since 15 mmol of 1-octene was used in the reaction, the theoreti-
cal maximum activity was 15 mmol 1,2-octanediol of yield. This can
be achieved only if conversion of 1-octene was 100% and selectivity
of the catalyst toward 1,2-octanediol was 100%. In the present work,
the conversion of 1-octene by using P/Nb/Ti–Si was 50%. The con-
version value could be even lower as some of the 1-octene may have
adsorbed onto the SiO2 supported catalyst. The selectivity of 1,2-
epoxyoctane and 1,2-octanediol was approximately 25 and 75%,
respectively, with detection of trace of some byproducts. If the pres-
ence of byproducts is ignored, the theoretical maximum activity
was 5.6 mmol 1,2-octanediol of yield. As evidenced from pyridine
adsorption analysis, amount of Brønsted acidity in P/Nb/Ti–Si was
that of produced using P/Nb/Si, the sample without TiO . Appar-
2
ently, the PO43 treatment has promoted the oxidative active site
generation or has strengthened the oxidative active sites. Its role
in oxidative active site formation will be discussed in detail in our
next report. Similar to sample P/Ti–Si, this sample was totally inac-
tive in diol formation. Therefore, we conclude that the co-existence
−
of PO43 and Nb O5 is important for Brønsted acidity generation.
−
2
Fig. 4(i) shows that the sample P/Nb/Ti produced limited amount
of 1,2-epoxyoctane and 1,2-octanediol. Apparently, the absence of
SiO2 support in this sample has caused remarkably decrease in
the catalytic activity (16 times lesser in yield) as compared with
−
P/Nb/Ti–Si. Owing to abundant of OH groups available on its sur-
face, SiO provided space for better dispersion of TiO , Nb O5 and
2
2
2
3−
PO4 , bringing to effective interaction among the elements and
hence creating more active sites. It was reported previously that
−
1
2.4 mol g . Thus, there was only 0.12 mol Brønsted acidity in
0.05 g of the catalyst used in the reaction. Since the formation of
1,2-octanediol strongly depends on amount of Brønsted acidity
usage of SiO2 support for Nb O5 has facilitated Brønsted acidity
2
generation [14]. In another work, Ziolek et al. [33] claimed that
silanol group present in SiO2 type support was important to form
Nb OH Si which gave enhanced acidic character to the material.
It is note-worthy that a relatively economical support of SiO2
was used in this work. Previously, it had been demonstrated that
the oxidative-acidic bifunctional catalyst of vanadium-phosphate
impregnated silica–titania aerogel yielded only 151 mol 1,2-
epoxyoctane and 275 mol 1,2-octanediol under the same reaction
condition [3]. The better catalytic performance of niobium-
phosphate impregnated silica–titania in the current work strongly
suggests the usage of fumed silica, a relatively easy prepared SiO2
matrix as effective catalyst support. Besides, the finding may also
available in the catalyst, only 461 mol 1,2-octanediol was yielded.
In fact, the bifunctional catalytic performance of P/Nb/Ti–Si is much
better compared to other reported bifunctional catalysts including
sulfated titania-silica aerogel [24], phosphate-vanadium impreg-
nated silica–titania aerogel [3] and sulfated zirconia loaded TS-1
[
23] in terms of selectivity and yield of 1,2-octanediol.
As it was claimed, TiO supported SiO materials have only Lewis
2
2
acidity, but no Brønsted acidity [30]. Therefore, it is believed that
existence of Nb or PO43 or combination of two has facilitated
creation of Brønsted acidity in the samples regardless the loading
method applied. In order to further study the role of each element
in the prepared bifunctional catalysts and to explain the catalytic
−
indicate that, as compared to V O5, Nb O5 was definitely a better
2
2
3
−
behavior of these samples, samples without TiO , Nb O5, PO
4
modifier in improving both oxidative and acidic catalytic activities
2
2
and SiO2 were synthesized via the preparation procedure similar
with that of P/Nb/Ti–Si which was the best bifunctional catalyst in
this work. The activity of the samples was evaluated in the same
catalytic reaction. As observed in Fig. 4(f), with absence of TiO2,
the sample P/Nb/Si has produced only 71 mol 1,2-epoxyoctane
and 30 mol 1,2-octanediol. The dramatic drop of more than 56%
in epoxy yield suggested that Ti species contributed as oxidative
site. The formation of 1,2-octanediol strongly indicated the pres-
ence of Brønsted acid sites in the sample since only Brønsted acidity
provides the active site for epoxy to diol transformation.
of TiO –SiO2 materials.
2
4. Conclusions
An excellent oxidative-acidic bifunctional catalyst of P/Nb/Ti–Si
was successfully synthesized. This material was a promising cata-
lyst in the consecutive transformation of 1-octene to 1,2-octanediol
through the formation of 1,2-epoxyoctane. Both tetrahedrally
coordinated Ti and Nb species acted as oxidative sites for the
epoxidation. The phosphoric acid treatment has contributed to
the enhanced oxidative activity. Meanwhile, the direct interaction
Sample P/Ti–Si was the weakest oxidative catalyst (Fig. 4(g)).
Without Nb O5, the sample produced only 8 mol 1,2-
2
between PO43 and Nb2O5 was important for the Brønsted acid-
ity formation, bringing to the high yield of diol. As evidenced, SiO2
support played a key role for the improved bifunctional catalytic
performance of the material.
−
epoxyoctane, with no detection of any 1,2-octanediol after
2
4 h of reaction. Even though this sample has very low surface
area and porosity as compared to that of TiO –SiO , both materials
2
2
have similar catalytic activity in transformation of 1-octene to
,2-epoxyoctane. It may suggest that the surface area and porosity
1
were not important in the epoxidation. As evidenced, the samples
P/Nb/Ti–Si and P/Nb/Si yielded 164 mol and 71 mol, respec-
tively, 1,2-epoxyoctane. It was documented that the SiO2 support
has aided formation of more isolated Nb which acted as active site
in SiO supported Nb O5 catalyst for the oxidation reaction [26,31].
In another study, the authors claimed that the incorporation of
Nb into MCM-41 type silica matrix resulted in formation of active
and selective catalyst for the oxidation of hydrocarbons [32]. Here,
Acknowledgments
The authors acknowledge funding from the Ministry of Higher
Education (MOHE) and Universiti Teknologi Malaysia for the
Research University Grants (Vote no: Q.J.130000.2526.00H08 and
Q.J.130000.2526.02H81). The authors are grateful to Prof. Dr. Sug-
eng Triwahyono from Ibnu Sina Institute for Fundamental Science
Studies, Universiti Teknologi Malaysia for the pyridine adsorption
analysis. J.M.E. is thankful to Universiti Teknologi Malaysia for the
UTM Institutional Zamalah Scholarship.
2
2
our results strongly implied Nb O5 would be a more important
2
active site for oxidation compared to that of originated from TiO .
2
This finding explains why P/Nb/Ti/Si which possesses more Nb5
+