interesting differences between the different reaction systems,
see Table 3. Almost complete conversion of cyclohexene into
polar and the end product, adipic acid, is water soluble. The
first two steps, the epoxidation (I) and the hydrolytic opening
of the oxirane ring (II) are therefore likely to be true interfacial
reactions, i.e., occurring at the pore openings where the polar and
the apolar regions meet. The subsequent reaction steps (III–VI)
may occur at the interface but since the intermediates 2–5 have
a certain solubility in water these reactions may also take place
in the aqueous phase, i.e., inside the pores of the mesoporous
material. If this is the case, the incompatability problem should
only be an issue for the steps I and II of Scheme 1.
Table 2 lists a number of characteristic properties of the
mesoporous materials. It has not been possible to correlate
parameters such as pore size or surface area with yield of adipic
acid, as given in Table 3. The chemical composition of the
material seems to be much more important than the internal
structure.
One should be aware that the number of parameters that may
influence the reaction is large when comparing such a wide range
of different catalysts and this work only takes into account the
parameters that we have regarded to be the most important.
We have, for instance, not taken particle size into account. It
is conceivable that a smaller particle size will result in more
active sites per unit weight of catalyst being exposed, which may
well lead to a higher reaction rate. This could be a topic for
a future study. Likewise, we have made no attempt to assess
the effect of ions in the reaction medium. Some of the solid
catalysts contain sodium and chloride ions, and such electrolytes
may effect the acid–base properties of the materials. A typical
example where this type of consideration is relevant is in the
adipic acid was obtained when a slurry of mesoporous WO
used as reaction medium. Nearly quantitative yield was also
obtained with the mixed oxide WO /SiO provided the WO to
SiO ratio was high enough.
Reaction with SiO –A and B with added tungstate catalyst
gave a moderately good yield of adipic acid, 44 and 55%,
respectively. Al /SiO , also with added tungstate catalyst,
gave 17% yield but only Al with added catalyst did not give
any adipic acid at all. The reason why Al gives such a poor
yield is not clear. It is the material with the largest pore size
but the material with the second largest pore size is WO , which
3
was
3
2
3
2
2
2
O
3
2
2
O
3
2
O
3
3
was the oxide that gave the highest conversion of cyclohexene to
adipic acid of all the materials. A possible reason for the poor
performance of Al
of zero charge of all the investigated oxides, in the range 8.0–
.5. Thus, at the conditions used for the reaction the pore walls
2
3
O could be that it has the highest point
8
carry a net positive charge. This could be detrimental for the
reaction because such positively charged surface sites may tie up
2
-
the tungstate anion, WO , which is added (as its sodium salt)
4
as catalyst. The catalyst may become inactivated by adsorption
to the walls of the mesoporous structure. It can be seen from
Table 3 that the main product when Al
2
O
3
is used, alone or as
a mixed oxide with SiO , together with the tungstate catalyst is
2
the intermediate diol 2. Evidently, it is the oxidation of this diol
to the hydroxyketone 3 that fails in the alumina system.
According to the literature, titania is an efficient catalyst for
11
oxidation of cyclohexene to adipic acid. In this work TiO
was much inferior to WO , however, and gave a very low yield
of adipic acid. As can be seen from Table 3, there seems to
be a threshold value of WO in the WO /SiO material for
the reaction to proceed well. WO /SiO -3, which gives almost
quantitative yield of adipic acid, contains 5 mol% WO while
WO /SiO -4, which gives a poor yield, contains 2.5 mol% WO
2
3
comparison between the WO
The most important finding of this work is that the meso-
porous material with chemically incorporated WO catalyst
3
and the Na
2
WO
4
materials.
3
3
2
3
3
2
constitutes a very reactive system, even more reactive than
mesoporous silica with added soluble catalyst. This observation
opens for facile reuse of the material, which from a process point
of view is very important. The next section addresses this issue.
3
3
2
3
.
Table 3 also shows the turnover frequency (TOF), calculated as
mole of adipic acid generated per mole of catalytically active
metal (W or Ti). As can be seen, the highest activity is obtained
with the materials that have high surface area and low loading
of catalytically active sites.
Reuse of mesoporous catalyst
A very attractive feature of having the catalytically active site
covalently built into the mesoporous material, as in mesoporous
There is a large difference between WO
–5 materials on the one side and WO /SiO
Both give full conversion of cyclohexene but only WO
WO /SiO -1–5 are capable of catalyzing oxidation of cyclohex-
anediol (intermediate 2). This difference in catalytic activity
may be related to the observation that whereas WO /SiO
–5 contains crystalline WO groups, the oxotungsten silica,
WO /SiO , seems to lack crystallinity. Futhermore, it could be
that WO and the WO3/SiO materials have the majority of the
active sites on the framework while in WO /SiO the active sites
are largely hidden into the bulk of the material. It is interesting
to note the high conversion into adipic acid obtained with WO
3
and the WO
on the other side.
and
3
/SiO -
2
1
4
2
WO
as compared to the situation when a soluble catalyst is added,
such as when mesoporous SiO –A or SiO –B is used together
with added Na WO . In order to test the reusability of the
catalyst, mesoporous WO and the WO /SiO -1–5 materials
3
and WO
3
/SiO -1–5, is that reuse of the catalyst is simplified
2
3
3
2
2
2
2
4
3
2
-
3
3
2
1
3
with varying ratio of the two oxides were subjected to recycling.
After a completed 24 h reaction the solid material was
removed from the reaction mixture by filtration, washed and
reused. Fig. 6 shows analysis results from the two consecutive
4
2
3
2
4
2
runs. As can be seen, WO
with the highest content of WO
2) work at least as well in the second run as in the first. However,
WO /SiO -3, which was an active catalyst in the first run,
performed poorly in the second run. Thus, it seems that the
threshold value of WO content for making the mixed oxide an
3
, as well as the two mixed oxides
3
3
(WO /SiO -1 and WO /SiO
3
2
3
2
-
only despite the fact that this material has a large pore size and
a low surface area. In this case there is evidently no correlation
between surface area and catalytic efficiency.
As illustrated in Scheme 1, oxidation of cyclohexene to adipic
acid involves several consecutive steps. Cyclohexene and the first
intermediate, epoxide 1, are hydrophobic species with very low
solubility in water. The other intermediates are somewhat more
3
2
3
active oxidation catalyst has become higher.
A probable explanation for the decreased efficiency of
the WO
3
/SiO
2
-3 catalyst is that the pore system has been
This journal is © The Royal Society of Chemistry 2010
Green Chem., 2010, 12, 1861–1869 | 1867