64
R. Gava et al. / Catalysis Communications 40 (2013) 63–65
Table 1
were greater in the case of THF. The least active catalyst appeared to
Supports and catalysts employed in the present work.
be CuOx/TiO2, which could be explained with the low surface area of
this nonporous support (Degussa P25) and with the consequent lower
degree of dispersion of the copper(II) oxide phase. The most active cat-
alysts, namely CuOx/SiAl, CuOx/Al2O3 and CuOx/SiTi2.3, were instead
obtained when deposition of Cu did not clog the pores of the support
(Table 1). The presence of copper(I) species may however also play
a significant role, particularly with THF as the substrate. Thus, the
employed catalyst preparation method allows formation of well dis-
persed CuO on all the supports but one, namely silica–alumina. In this
case, the presence of strong acidic sites similar to those determined
for a H-Beta zeolite [27] makes a pure exchange mechanism of these
sites with the [Cu(NH3)4]2+ solution possible (competitive with the
chemisorption hydrolysis one). This exchange generates isolated cop-
per species, hardly reducible to an oxidation state lower than monova-
lent copper, as evidenced by several techniques [18]. Copperδ+ species
resistant to reduction have been also detected in solvated CuOx/Al2O3
by ATR-IR and XAS [17], and to a minor extent in CuOx/SiTi 2.3 in the
gas phase [19]. Consequently, the increased catalytic efficiency of
these materials seems to correlate also with the presence of copper(I)
sites. This may be due to the fact that, according to the current under-
standing of the mechanism of this reaction, copper(I) species are the
actual catalyst [28], hence when they are present from the beginning
they are directly available for reaction, whereas copper(II) species
have to be prereduced to copper(I) by the nitrene source before enter-
ing the catalytic cycle [25]. Furthermore, such copper(I) species are
present as isolated species on the support, rather than as copper(I)
oxide nanoparticles, hence they are more available for reaction with
the substrate due to their higher degree of dispersion.
To shed some light into the relevance of the oxidation state, we
also tested some of the catalyst in reduced form [16–19] (Table 2).
In the case of the best catalyst, namely CuOx/SiAl a significant
drop in yield was obtained with both dioxane and THF on moving
to reduced Cu/SiAl. On the contrary, the activity of both Cu/SiO2 and
Cu/TiO2 showed only a minor decrease. As already mentioned, on
the surface of CuOx/SiAl a minor part of the metal is present as CuO,
while most of the Cu is present as copper(I). Since reduction of the
CuO fraction on the surface results in a lower yield, it seems that
the presence of both oxidation states on the surface is advantageous
for the reaction (it should be remarked that pore volume measure-
ments show that there is almost no Cu in the pores). A similar positive
effect of the simultaneous presence of copper in different oxidation states
was already observed in the polymerization of 2,6-dimethyl-phenol with
the same kind of catalysts [29]. On the other hand, Cu/SiO2 and Cu/TiO2,
are promptly reoxidized (the catalysts turn greenish under reaction
conditions), which restores their catalytic activity.
Finally, an evaluation of the true nature of catalysis was performed.
Leaching of copper into solution in the course of the reaction, as well as
after reaction, was invariably found to be very low (b5% of added
copper, most commonly 2–3%). However, upon removal of the solid
catalyst at the reaction temperature after only 1 h of reaction (“hot fil-
tration test”) the reaction was found to continue in the solution phase,
reaching after 8 h a final yield fully analogous to that recorded with
the solid catalyst. Thus, it can be concluded that the catalytically compe-
tent species are highly active copper species released into solution by
the solid catalyst. The exact nature of this species is presently unknown;
it needs to be remarked that control reactions performed with dioxane
under identical reaction conditions using as catalyst simple copper salts
such as copper(II) triflate (the most efficient copper salt as catalyst for
this reaction) [22] invariably showed much lower yields of C\H inser-
tion product, reaching after 8 h 30% yield at the best with 10 mol% cop-
per (average TOF 0.37 h−1) to be compared with the average TOF of
27 h−1 estimated with the best heterogeneous catalyst upon taking
into account the low amount of catalytically competent, leached copper
species. Importantly, the solid catalysts maintain some activity upon
recycling. For example, as it can be appreciated in Table 2 (last two
Catalyst
Surface area (m2/g)
Pore volume (mL/g)
Ref.
[16]
SiO2
CuOx/SiO2
Al2O3
480
363
280
238
400
412
483
430
50
0.75
0.68
1.15
1.24
0.77
0.75
1.43
0.86
n.a.
[17]
[18]
[16]
[17]
[19]
CuOx/Al2O3
SiO2−Al2O3
CuOx/SiAl
SiO2−Al2O3 0.6
CuOx/SiAl 0.6
TiO2
CuOx/TiO2
SiO2−TiO2 2.3
CuOx/SiTi 2.3
45
340
349
n.a.
1.20
1.01
We next set out to optimize the reaction conditions using catalyst
CuOx/SiAl (Fig. 1). Analysis of the conversion vs. time curves made it
clear that the maximum yield at 70 °C (58%) was reached already
after 8 h of reaction, hence this reaction time was taken as standard
for the subsequent tests. The yield was found to slightly decrease at
prolonged reaction times, possibly due to some product decomposition
under the reaction conditions. Indeed, the reaction product is an
N-tosylated cyclic hemiaminal (Scheme 1), which has been reported
to undergo decomposition with relative ease (e.g. isomerisation to an
open chain N-tosylated imine) [24]. Variations of the reaction tempera-
ture in the range 50–90 °C affected the reaction rate but not the maxi-
mum yield of the reaction, apart from the highest temperature, at
which perhaps the reaction product again underwent partial decompo-
sition, thereby resulting in lower reaction yields at all times. An increase
of the amount of catalyst to 20 mol% Cu using CuOx/SiAl resulted in no
appreciable increase in yield, whereas with a lower catalyst amount
(5 mol% CuOx/SiAl) a lower yield (35%) was obtained after 8 h. Identical
observations concerning the dependence of the catalytic efficiency on
temperature and amount of catalyst have been previously reported
with homogeneous copper(I) catalysts in a closely related reaction
(styrene aziridination) and attributed to a competing copper-catalyzed
decomposition process of the nitrene source [25]. Remarkably, and in
contrast to most homogeneous catalytic systems described before, the
reaction could be conveniently performed also with commercial chlora-
mine T trihydrate, without the need for previous dehydration of the
same, which according to some reports involves explosion hazards
[26]. This results into a safer and more economical reaction procedure,
with yields equal to those obtained with anhydrous chloramine T.
Using the optimized reaction conditions, a screening of the various
catalysts was performed with dioxane or THF as substrates (Table 2).
As outlined above, the catalysts differ primarily for the chemical nature
of the oxide support and for the surface area and pore volume. The re-
sults obtained with the various catalysts were rather comparable
when dioxane was employed as substrate, whereas the differences
Scheme 1. Amidation reactions via nitrene insertion investigated in the present work.
Gava, Riccardo
