Improved methodology for non-covalent immobilization of tert-butyl-azabis(oxazoline)-copper complex on Al-MCM41

Article abstract of DOI:10.1016/j.apcata.2015.06.005

Abstract Copper sites are supported on Al-MCM41 by thermally treating copper(II) triflate that was deposited using the incipient wetness method. The change of the chemical state of the triflate anions following thermal treatment is monitored using TGA and SEM-EDX. Thermal treatment yields copper sites that then can be modified with chiral tert-butyl-azabis(oxazoline) by adding a solution of this ligand to the support using incipient wetness. The catalyst produced in this way is efficient and stable with good enantioselectivity, and can be used in the cyclopropanation of styrene with ethyl diazoacetate under conditions employing an elevated temperature. The use of thermal treatment in the catalyst preparation method produces catalysts that are superior to the solids prepared by cation exchange or incipient wetness of the pre-formed complex.

Full text of DOI:10.1016/j.apcata.2015.06.005

Accepted Manuscript
Title: Improved methodology for non-covalent immobilization
of tert-butyl-azabis(oxazoline)-copper complex on
Al-MCM41
´
Author: Robert A. Feldman Jose M. Fraile
PII:
S0926-860X(15)30022-3
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.apcata.2015.06.005
APCATA 15417
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
30-4-2015
5-6-2015
7-6-2015
´
Please cite this article as: Robert A.Feldman, Jose M.Fraile, Improved methodology for
non-covalent immobilization of tert-butyl-azabis(oxazoline)-copper complex on Al-
MCM41, Applied Catalysis A, General http://dx.doi.org/10.1016/j.apcata.2015.06.005
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Improved methodology for non-covalent immobilization of tert-butyl-
azabis(oxazoline)-copper complex on Al-MCM41
Robert A. Feldman, José M. Fraile*
Instituto de Síntesis Química y Catálisis Homogénea (ISQCH), Facultad de Ciencias,
C.S.I.C. - Universidad de Zaragoza, E-50009 Zaragoza, Spain. E-mail:
jmfraile@unizar.es
Highlights
- Chiral azabox-copper complex was immobilized on Al-MCM41 in a non-covalent
way.
- Cu(OTf)2 was first impregnated by incipient wetness and then treated at 450ºC.
- Final modification with azabox leads to stable immobilized complexes.
- Better results in enantioselective cyclopropanation are obtained at high temperature.
Abstract
Copper sites are supported on Al-MCM41 by thermally treating copper(II) triflate that
was deposited using the incipient wetness method. The change of the chemical state of
the triflate anions following thermal treatment is monitored using TGA and SEM-EDX.
Thermal treatment yields copper sites that then can be modified with chiral tert-butyl-
azabis(oxazoline) by adding a solution of this ligand to the support using incipient
wetness. The catalyst produced in this way is efficient and stable with good
enantioselectivity, and can be used in the cyclopropanation of styrene with ethyl
diazoacetate under conditions employing an elevated temperature. The use of thermal
treatment in the catalyst preparation method produces catalysts that are superior to the
solids prepared by cation exchange or incipient wetness of the pre-formed complex.
Keywords: asymmetric catalysis; immobilized catalysts; mesoporous materials;
bis(oxazolines); copper;

1. Introduction
Ordered mesoporous materials have been described as supports for single-site catalysts
[1] due to their well-defined textural properties that include their pore size, high surface
area and pore geometry. Chiral catalysts have been immobilized on ordered mesoporous
materials either by covalent bonding or by utilizing non-covalent interactions between
the chiral complex and the support [2-4]. Through these strategies, effects on selectivity
have been observed and explained by the restrictions that their pores impose on the
trajectories of approach of the reacting species to the catalytic center [5,6].
In homogeneous catalysis, bis(oxazolines) and other related compounds are among the
most useful chiral ligands for enantioselective reactions [7] and in efforts to capitalize
on the applications of these ligands as supported catalysis, they have been immobilized
on different supports, using a multitude of methods [8-10]. Examples described in the
literature utilizing ordered mesoporous materials are relatively uncommon but they exist
for the covalent immobilization of complexes on different materials, such as
mesocellular foam silicas [11-15], MCM41 and MCM48 [16], hierarchically-ordered
mesocellular mesoporous silica [17], SBA15 and HMS [18,19], and a periodic
mesoporous phenylene-silica [20]. To date, a few examples are described using non-
covalent methods of immobilization, such as the hydrogen bonding of ion-tagged
bis(oxazoline) ligands on SBA15 [21] and electrostatic interactions with Al-MCM41
and Al-SBA15 [22] or Al-TUD1 [23]. This limited number of examples contrasts with
those dealing with the non-covalent immobilization of other types of complexes, such as
Mn(salen) [24,25] or Rh-phosphines [26,27], on mesoporous ordered materials [4].
In our previous work [28] we demonstrated how the cation-exchange process, which
successfully immobilized bis(oxazoline)- and azabis(oxazoline)-copper complexes on
clays (layered materials) through electrostatic interactions [29-31], was not suitable
when applied using the anionic mesoporous materials Al-MCM41 and Al-HMS. The
best results were obtained when the complex was introduced into the mesoporous
network by incipient wetness, probably due to diffusion problems in the cation
exchange process. A conceptual downside of this approach is that it does not allow the
separation of the triflate anions from the catalytic centres, and so the true nature of the
complex-support interaction is likely to be quite complicated. Indeed, the observations
pointed to there being more than one type of catalytic site, and moreover, the
immobilized catalysts could not be used for more than a few cycles.

Looking to improve the approach, we hypothesized that incipient wetness of the support
with a solution of a copper salt would ensure that at least most of the copper would be
deposited inside the porous network, and in the most uniform manner possible. The
subsequent thermal treatment would then convert the copper counter anion to a stable
form that would bind the copper sites firmly to the support. A similar idea had in fact
already been described by Hutchings and coworkers [22,32-34] for Cu-exchanged
zeolites and mesoporous materials, using the thermal treatment of copper(II) acetate at
550ºC, followed by treatment with bis(oxazoline) ligands. Nonetheless, in our hands it
was not possible to prepare stable catalysts following the same process with our Al-
MCM41 supports. We decided to continue using copper(II) triflate, which is soluble in
polar organic solvents and additionally, we preferred impregnation by incipient wetness
because it is very quick to implement and gives one precise control over the final cation
loading.
Herein, we describe the preparation of a supported copper catalyst prepared with the
deposition of a copper precursor on Al-MCM41 by incipient wetness and thermal
treatment. The copper sites were then subsequently modified by (S)-tert-butyl-
azabis(oxazoline) ligand (azatBu), and the catalyst used in the enantioselective
cyclopropanation reaction of styrene with ethyl diazoacetate (EDA).
2. Experimental
2.1 Preparation of the catalysts
Al-MCM41 (Si/Al = 15) was prepared from sodium aluminate and fumed silica in a
basic medium with cetyl trimethyl ammonium as the template [35]. To expand the pore
size in the synthesis of Al-MCM41 (Si/Al = 10), 1,3,5-trimethylbenzene was added to
the gel [36].
2.1.1 Preparation of Cu-Al-MCM41.
Al-MCM41 (500 mg) was added to a solution of copper(II) triflate (42 mg, 0.1 mmol)
in methanol (600 µL for Al-MCM41(10), and 300 µL for Al-MCM41(15)) and the
container was shaken to ensure that the solution was evenly distributed. This material
was then calcined (1˚C/min heating until 450˚C, maintained for 4 h) under a 100

mL/min flow of either argon or air to obtain Cu-Al-MCM41(X)(Ar) or Cu-Al-
MCM41(X)(air) respectively (where X is the Si/Al ratio of Al-MCM41).
2.1.2 Modification with chiral ligand.
Cu-Al-MCM41 (500 mg, 0.1 mmol Cu) was added to a solution of azatBu (0.2 mmol) in
dichloromethane (600 µL for Al-MCM41(10), and 300 µL for Al-MCM41(15)), the
container was shaken to ensure that the ligand was evenly distributed, and this solid was
then immediately used as a catalyst.
2.2 Characterization methods
The adsorption-desorption isotherms of nitrogen at 77 K were measured using either a
Micromeritics ASAP 2000 or an ASAP 2020 instrument. Each sample was outgassed at
250˚C after being calcined until a stable static vacuum of 3×10-3 torr was reached.
Particle morphology was observed using a Hitachi S-4500 I scanning electron
microscope (SEM) and the composition was determined using energy-dispersive X-ray
spectroscopy (EDX). Thermogravimetric analysis (TGA) was carried out in a SDT2960
thermobalance (TA Instruments), heating from room temperature to 500˚C at 10˚C/min
in oxygen. Onset temperature (T1), maximum derivative temperature (Tm) and the
temperature of plateau start (T2) were used to characterize TGA profiles.
2.3 Catalytic tests
2.3.1 Homogeneous cyclopropanation.
At room temperature, EDA (2.5 mmol, 285 mg) dissolved in dichloromethane (0.5 mL)
was added over the course of 2 h with a syringe pump to a solution containing the
copper catalyst (0.025 mmol), n-decane (100 mg) and styrene (2 mL for reactions in
excess of alkene or 2.5 mmol, 260 mg, in 2 mL dichloromethane for stoichiometric
reactions). After the addition, the reaction was left to stir for 30 min and then analysed
by CG [28-31].
2.3.2 Heterogeneous cyclopropanation.

The supported catalyst (100 mg, around 0.02 mmol Cu) was suspended in either styrene
or dichloromethane (2 mL) with n-decane (100 mg) and the mixture was heated at the
required temperature. EDA (2.5 mmol, 285 mg) dissolved in dichloromethane (0.5 mL)
was added over the course of 15 min with a syringe pump. After the addition, the
catalyst was separated by centrifugation, the products were analyzed by GC, and the
next reaction was performed by immediately re-suspending the catalyst in the same
solvent.
2.3.3 Continuous flow reaction.
A stainless steel HPLC column (150 mm × 46 mm i.d.) was completely filled using 500
mg of Cu-Al-MCM41(10)(air) and it was sealed with frits (porosity of 2.0 µm). The
column was placed in an oven at 120˚C for 12 h prior to being used. For the
modification with the chiral ligand, a solution of azatBu (20 mL, 0.01 M) in the chosen
solvent was charged into a 25 mL gas tight syringe, fitted with a 51 mm metal hub
needle, and connected to the HPLC column via a Luer Lock Valve Adaptor. This
solution was then fed through the column using a syringe pump. As an example of
operation, a feed of styrene (4.5 g, 43 mmol), decane (0.1 g, 0.5 mmol) and EDA (0.3 g,
2.5 mmol) was passed through a column of unmodified Cu-Al-MCM41(10)(air) at a
flow rate of 1 mL/h. Different fractions were collected along the operation time and
they were then analysed individually by GC.
3. Results and discussion
3.1 Preparation and characterization of Cu-Al-MCM41
The Al-MCM41 supports were prepared by established methods [35,36] and came from
the same batches of the supports used in our previous paper [28]. Copper(II) triflate
dissolved in methanol was introduced into each Al-MCM41 sample (Si/Al = 10 or 15)
using incipient wetness impregnation [28] and then thermally treated. The hypothesis
was that thermal treatment would convert copper(II) triflate into a stable form that
would be strongly bound to the surface of the support.
With the goal of trying to characterize the changes that took place during thermal
treatment, a thermogravimetric study was performed with Cu(OTf)2 in its pure form,

and deposited on different supports, to identify where the transition points (T1, Tm, T2)
would occur. The principle is that if there should be even a partial electrostatic
interaction between the copper and the support, this would generate a chemical state
whose TGA would be distinguishible from the pure material by its different temperature
transition points. Furthermore, one would anticipate that the extent of the change would
vary in parallel with the strength of the interaction.
The TGAs of copper(II) triflate in pure form and deposited on pure silica materials
(Figure 1) produce the T1, Tm, T2 values that are summarized in Table 1. A property
seen with pure copper(II) triflate is that there are two separate weight losses at 100˚C
and 120˚C that probably correspond to a two step loss of water, also seen in the
dehydration of copper methanesulfonate [37]. These signals are well below the
temperature at which the triflate decomposes (460˚C) and therefore, they should not
interfere with this region. When lanthanide (III) triflates undergo thermal decomposition
the corresponding lanthanide fluoride is produced (Eq. 1) [38], and the same is
supposed to happen with copper(II) triflate. The weight-loss in the decomposition of
copper(II) triflate is 65%, which is close to the theoretical weight loss of 72% that
would be expected from the conversion to copper(II) fluoride.
(Eq. 1) Ln(CF3SO3)3 → LnF3 + 3 SO2 + CO2 + CF3OCF3
Figure 1 near here.
Table 1 near here.
Copper(II) triflate was deposited on two samples of chromatography-silica (500 m2/g)
with different particle size, and MCM41, using a copper loading of 1.5 wt% Cu. These
materials do not contain aluminium and therefore, one would anticipate a weak mode of
interaction between silica and the triflate, such as hydrogen bonding. Indeed, the TGA
shows that there is a detectable interaction, since the Tm of the supported Cu(OTf)2 in
all the three cases is lower than that of the pure salt (Table 1). It can be seen that Tm for
the chromatography silicas is at around 360˚C, which is 100˚C lower than pure
copper(II) triflate. This finding might not be unexpected considering it was similarly
found that there is a decrease in the temperature transition points when copper(II)
triflate was supported on Kieselgel-K100 [39]. In that case, the explanation for the drop
in Tm is that triflic acid was formed following the cation-exchange and this decomposes

at around 380˚C [40]. The curve produced when MCM41 is the support stands out from
the chromatography-silicas, since its Tm value is at around 400˚C and is so broad that it
eclipses both the decomposition regions of the chromatography silicas and the
decomposition of pure Cu(OTf)2. Based on this it could be inferred that there are at least
two types of interaction going on, which are the interaction of isolated copper(II) triflate
with the silica wall, and secondly, there are probably clusters of non dispersed
copper(II) triflate. Thus, under these circumstances the chromatography silicas provided
a more efficient level of dispersion, which is surprising given the difference in surface
areas (1000 vs 500 m2/g). The culprit could be a confinement effect inside the pores of
MCM-41 that prevents the copper from dispersing.
When the same amount of Cu is deposited on Al-MCM41(15) the decrease in the Tm
value (306˚C) is even more significant (Table 1) and would seem to indicate the
existence of copper/aluminium interaction that is likely to be electrostatic in nature. Due
the use of sodium aluminate in the precursor gel for Al-MCM41, the aluminium sites
are counter-balanced by sodium cations. This means that triflic acid would not be
expected to be liberated from the copper(II) triflate and this is confirmed by the much
higher decomposition point of triflic acid (380˚C) [40]. Pure sodium triflate decomposes
between 253°C and 256°C [41] and therefore, the formed species must not be sodium
triflate either. Therefore, it seems reasonable that what actually happens when
copper(II) triflate is added to our support is that the place of sodium cation is taken up
by an ion par such as [CuNa(OTf)2]+.
To examine this idea, the behaviour of another metal triflate was tested. Thus, zinc (II)
triflate was deposited (1.5 wt% Zn) on Al-MCM41(15) and assessed by TGA. The
finding was that it gave the same Tm value as supported Cu(OTf)2 (around 310˚C),
whereas the Tm value of pure Zn(OTf)2 is significantly higher than Cu(OTf)2, 528˚C vs
460˚C respectively (Table 1). The Tm value obtained when the two salts are deposited
on Al-MCM41 would indicate that both triflates interact with the support in the same
way.
The composition (Si/Al ratio) and textural properties (surface area and porosity) of Al-
MCM41 do not modify the behaviour of supported Cu(OTf)2 assessed by TGA (Table
1). On the other hand, the level of copper loading does have some influence. Different
copper loadings were added to Al-MCM41(15) to see if TGA could reveal anything
about the maximum loading capacity of the support. At loadings equal to 3 wt% or
below, the decomposition temperatures are the same (± 2˚C), whereas at loadings higher

than 3 wt% the Tm and T2 temperatures increase. It can be deduced that 0.5 mmol Cu
per gram of support saturates the Al sites in Al-MCM41(15), which contains around 1
mmol Al per gram. Beyond this level, from the increase witnessed in the value of Tm, it
appears that Cu(OTf)2 begins to interact more with the silica walls of the support. For
the actual preparation of the catalyst (denoted as Cu-Al-MCM41), 450˚C was chosen as
the temperature used for the thermal treatment since it is high enough to finish the
decomposition of the triflate anion.
A semiquantitative analysis was carried out by SEM-EDX, with the goal of seeing
whether the preparation method would achieve a homogeneous distribution of copper
and to further characterise the species on the support. Given the known empirical
composition of the synthesis gel, the analysis of Cu-Al-MCM41(10) and Cu-Al-
MCM41(10) gave Si/Al ratios lower than expected, and the O/(Si+Al) ratios were
higher. These two disparities are explained by an understimation of the silicon content
analyzed by EDX. On the other hand, the Cu/Al ratios were close to the value that
corresponds to 1.5 wt% Cu added in the impregnation process (1.7 wt% Cu detected in
Cu-Al-MCM41(10) and 1.3 wt% Cu in Cu-Al-MCM41(15)). Both Cu-Al-MCM41
samples showed significant amounts of fluorine and sulfur, and the extent of disparity
between the expected and the experimental values shows that they can only be used, at
best for qualitative analysis. The S/Cu ratio was 0.44 in Cu-Al-MCM41(10) and 0.59 in
Cu-Al-MCM41(15), which suggests that part of the sulfur is removed in the thermal
treatment of the catalyst (with a starting S/Cu ratio of 2). The F/Cu ratios were much
higher than the theoretical initial value of 6, which must come from an overstimation of
fluorine in the analysis. It can be especulated that the fluorine from the triflate remains
in the solid by reaction during the thermal treatment.
Using TGA and SEM-EDX data, a hypothesis can be proposed to explain the
transformation that Cu(OTf)2 undergoes during the catalyst preparation (Figure 2). Once
introduced onto the support, copper(II) triflate interacts with the sodium-counterion and
the aluminium site to form an ion-pair. In the course of thermal treatment the fluorine of
the triflate reacts with the surface silanol groups, to produce a partially fluorinated
surface, and the carbon atom is released in the form of CO2. Some sulfur could be
eliminated as either SO2 or SO3, but part of it could also remain on the solid as sulfate
anions, which are highly stable (decomposition at 850˚C on silica [42]).
Figure 2 near here.

The stability of the MCM41 structure following the deposition of copper and thermal
treatment was studied by nitrogen BET isotherms (Figure 3), and the textural properties
are summarized in Table 2. The hysteresis, as well as the surface area, pore volume and
pore size appear to be not significantly affected by the catalyst preparation.
Figure 3 near here.
Table 2 near here.
3.2 Complexation of Cu-Al-MCM41(Ar)
Having chosen 450˚C as the suitable temperature for the thermal treatment, the time
duration and the atmosphere needed to be decided on. Preparations were conducted
using either 4 or 8 h, and the same catalytic results were obtained in both cases, and so 4
h was deemed to be the optimum condition. The options considered for the atmosphere
were either a stream of air or argon.
Cu-Al-MCM41(10) and Cu-Al-MCM41(15) (1.5 wt% Cu) were heated at 450˚C, for 4h,
in a stream of flowing-argon and the solids obtained from this process will henceforth
be referred to as Cu-Al-MCM41(X)(Ar) (with X = Si/Al ratio). (S)-tBu-azabox (azatBu,
Figure 4) was then added by incipient wetness in order to form a chiral immobilized
complex, and this solid was used directly as the catalyst in the cyclopropanation of
styrene with ethyl diazoacetate (Figure 5) in dichloromethane at room temperature,
using a stoichiometric quantity of styrene (Table 3).
Figure 4 near here.
Figure 5 near here.
On the basis of the yield and selectivity, the results are very similar to those obtained by
the parent homogeneous phase catalyst, and so there is no evidence of a surface or
confinement effect. Regarding the performance, when these results are compared to
those obtained with the solid prepared by direct incipient wetness impregnation with the
azatBu-Cu(OTf)2 complex [28], this new method of preparation gives a catalyst much
more active, albeit with slightly lower selectivities. Therefore, one can assert that this
newer process with thermal treatment provides a more efficient way of binding the
complex to the support. Unfortunately, when the catalyst is reused, the yield falls

significantly over the course of three cycles, but nonetheless, two interesting facts were
observed. Firstly, leaching was not detected and secondly, the enantioselectivities were
stable over the course of the recycles. The apparent lack of leaching reinforces the point
that the bond between the copper and the support is indeed very strong, and the stable
performance in the enantioselectivities, but concurrent loss of yield, shows that the
deactivation mechanism does not involve the separation of the ligand from the copper
sites.
Table 3 near here.
In our previous work we described how the supported azatBu-Cu complex performed
better when the reaction was carried out in styrene at 90ºC [28]. Using just supported
copper treated by different methods, without modification by chiral ligand, permits one
to see how the support influences the results of reactions, and we studied the effect of
the reaction temperature on unmodified Cu-Al-MCM41(15)(Ar). Using a strategy that
was originally used to address the problem of improving the reduction of Cu(II) to
Cu(I), (a pre-requisite of the catalytic cycle) [31], different heating regimes on the
cyclopropanation reaction were applied (Table 4): heating at 90˚C only in the beginning
of the reaction (method A), continuos heating at 90˚C (method B) and 100˚C (method
C). The results at 90 and 100˚C are very similar. A remarkable trend is that trans
selectivity smoothly reverses in the course of the cycles, probably due to a change in the
nature of the catalytic site. At the beginning when the catalyst is unused, the copper may
be at its most tightly bound state to the surface, and this imparts the most significant
effect on diastereoselectivity [43]. Gradually, as the catalyst is re-used, the copper can
complex with by-products of the cyclopropanation reaction (mainly diethyl maleate or
analogous oligomers) and this might move the copper further from the surface. For the
homogenous phase reaction, EDA is generally added in 2 h in order to keep the level of
side products formed at the minimum. An important consequence of using excess
styrene, at an elevated temperature, is that EDA can be added at a much faster rate (15
min), without observing a loss of yield due to side products being formed.
Table 4 near here.

Chiral catalysts were prepared using both Cu-Al-MCM41(10)(Ar) and Cu-Al-
MCM41(15)(Ar) and tested in cyclopropanation using the newly optimized reaction
conditions (Table 5). When compared with Table 3, yields are higher, the
enantioselectivities are lower (10-13% ee lower than at room temperature), and almost
50:50 cis/trans ratio is obtained. There is a remarkable difference in the yields obtained
on the different supports, but the selectivities are nearly the same in both cases. The
yields with Al-MCM41(10) are significantly higher than with Al-MCM41(15), and this
may be related to its larger pore size (92 vs 36 Å of pore diameter). The larger pore
could mean that there is a more free flow of reagents to and products around the
catalytic sites. Regarding recovery, the yield drops from the 4th cycle and by the 6th
cycle modified Cu-Al-MCM41(10)(Ar) is almost inactive. Nevertheless, a certain
number of re-uses are possible when using the supported complex whereas, at elevated
temperature with excess styrene, the reaction is impractical in homogenous phase. Thus,
it is not possible to compare the homogeneous and the heterogeneous catalysts under
identical conditions, because when copper(II) triflate is heated in styrene it initiates a
polymerisation reaction. A certain portion of the inactivation will inevitably arise from
pore blocking by the by-products of side reactions, such as oligomerization of styrene.
It is remarkable that, taking into account the recycling, these catalysts are very
productive with stable selectivities. The modified Cu-Al-MCM41(15) produces a TON
of 152 (6 runs), with total cis/trans selectivity 48/52, 70% ee trans and 65% ee cis. Also
in 6 runs, Cu-Al-MCM41(10) leads to a TON of 284, with cis/trans = 49/51, 74% ee
trans and 69% ee cis.
Table 5 near here.
3.3 Complexation of Cu-Al-MCM41(air)
In the literature, it is described how the thermal treatment of supported copper in a flow
of inert gas can produce reductions of Cu(II) to Cu(I) and Cu(0) [44,45]. In case this
could be detrimental for the activity of the catalyst, the use of a stream of air instead of
argon was also tested. This catalyst will be referred as Cu-Al-MCM41(10)(air). This
material was then treated in the same way as before by incipient wetness with a solution
of azatBu and tested in the same benchmark cyclopropanation reaction. The most

relevant results after treatment in air are compared with those obtained after treatment in
Ar in Figure 6.
Depending on the stream used during the catalyst preparation, there is a difference in
the enantioselectivities and a different pattern of behaviour in the yields. After thermal
treatment in air and modification with azatBu ligand, the yield in the first run is lower
than what was obtained in further cycles and also lower than what was obtained in
further cycles and also lower than that obtained with the solid treated in argon. From
these results it can be inferred that at least a proportion of the copper sites in Cu-Al-
MCM41(Ar) are in the form of Cu(I), which would be expected to react immediately
with diazoacetate in the cyclopropanation. For Cu-Al-MCM41(air) one would expect to
find Cu(II), which needs to be reduced to Cu(I) in order to react, and this could explain
why the first use of the catalyst does not give the highest yield. Notably, upon recycling
Cu-Al-MCM41(air) gives consistent yields of more than 60% in runs 2-5. There is a
similar consistency in the diastereoselectivity, whose absolute value is a little bit lower
than seen in the homogeneous phase, however, it is not clear whether this could be
related to either the effect of mesopores or the reaction temperature. At the same time
the enantioselectivity remains stable along the five cycles in values of 60% ee for the
trans and 50-56% ee for the cis isomers, values consistently 12-15% ee lower than those
obtained with Cu-Al-MCM41(Ar). Thus thermal treatment in air gives better stability
than the argon-treatment, but leads to a slightly lower enantioselectivity.
Figure 6 near here.
3.4 Continuous flow reactions
Given the advantages of the continuous flow systems over batch reactions, these solids
were tested in a simple continuous reaction setup to see how they would perform. At
room temperature, the reagents were fed into a HPLC column packed with the Cu-Al-
MCM41(air) (1.5wt% Cu) catalyst. Although the flow rates of solvent through the
column can be as high as 40 mL/h, the application of a feed containing the reagents
produced a significant resistance and flow rate had to be reduced significantly, down to
1 mL/h. In the course of one experiment, fractions were collected every 5 min and
analysed by GC. Diastereoselectivity was found to be constant, at around 50:50
cis/trans ratio. Regarding yield (Figure 7), it remains quite stable for 30 min at values of

40-45% and then it drops, with an increase in the content of dimerization products
(diethyl fumarate and maleate).
Figure 7 near here.
Next, the modification of the copper sites with the chiral ligand in a fresh column was
attempted by feeding through a dichloromethane solution of azatBu into the column
through the same syringe feed-system. Upon doing so, the column became completely
blocked almost instantly, in spite of using the support with the largest pore size (92 Å
pore diameter). The blocking of the column seems to suggest that dichloromethane is
not capable of distributing ligand to all the copper sites. A different solvent might affect
the distribution of the ligand throughout the column, and less a tendency to blocking.
The requirements of the solvent are to dissolve the ligand, carry it down the pores to the
copper sites, and not compete with the ligand at the copper sites. In batch systems
dichloromethane has been used as typical solvent for impregnation of meso- [46] and
microporous materials [47]. On the contrary, Hutchings et al modified CuHY zeolite
with bis(oxazoline) ligands in acetonitrile [48], a considerably more polar solvent.
In view that the retention factor (Rf) of azatBu on thin layer chromatography plates is
zero when dichloromethane is used as an eluent, pore blocking could have been
expected. As ethyl acetate is used in the column-chromatography purification of azatBu,
this solvent and others with eluotropic strength [49] higher than dichloromethane (ε0 =
0.32), were considered as alternatives. 20 mL of 0.01 M solutions of azatBu in ethyl
acetate, acetone, acetonitrile, and methanol (ε0 values in the range of 0.38-0.73), were
fed into the column and then washed with the same volume of solvent. At the end of the
column all the ligand was collected after just one passing. However, all the solvents
with higher eluotropic strength than dichloromethane also show a higher coordinating
ability index [50], and they can compete more favourably with azatBu in the
coordination to the copper sites. On the other hand, chloroform is a solvent with a
slightly lower coordinating ability and eluotropic strength than dichloromethane. In the
same test described before 90% of the ligand was retained on the column, which could
be slowly eluted by passing larger amounts of solvent. In spite of this, in principle,
better distribution of the ligand inside the pores, the resistance to flow did not
completely disappear, but it allowed the test of the column in an enantioselective
cyclopropanation.

Once the reactants were fed, the column did become blocked very quickly, but it was
possible to obtain a small amount of product. Although the yield was very low (11%)
and the cis/trans ratio was 45/55, the enantiomeric excess were 66% and 77% ee for cis
and trans cyclopropanes, very similar to the results in batch reactions at 90˚C with the
solid calcined in Ar (Table 5) and significant better than those obtained with the solid
calcined in air (Figure 6). Probably in this case the chiral ligand is better distributed
inside the pores, leading to a more efficient enantioselective reaction but with a pore
blocking that hampers the long-term use of the column.
4. Conclusions
In this work, copper was supported on Al-MCM41 using the incipient wetness of a
methanol solution of copper(II) triflate, following by thermal treatment in a flow of
either air or argon. Using the available evidence from TGA and SEM-EDX, we are able
to propose a hypothesis for what the structure of the catalytic site could look like. We
found that the catalysts prepared in argon gave higher enantioselectivities when
modified by the (S)-tBu-azabox ligand, also by incipient wetness impregnation, but the
catalysts prepared in air gave higher and more consistent yields. According to data
described in the literature, it is probable that when the support was thermally treated in
argon, part of the copper is reduced to Cu(I). Although the energy requirement for the
thermal treatment makes this procedure less environmentally benign than the direct
incipient wetness impregnation, it is worth taking into account the improvement in
performance and stability, as well as the need for a calcination in the preparation of the
mesoporous support [35,36]. The Al-MCM41 support that had the larger pore-size
produced yields that were higher than the smaller pored support, which could be
explained by there being better diffusion. Finally, we found that in their present form,
due to the tendency of column blocking, these catalysts are not suitable for use in a flow
reactor.
Acknowledgement
This research was made possible by the financial support from the Spanish Ministerio
de Economía y Competitividad (Project CTQ2011-28124) and the European
Commission (NANO-HOST program no. PITN-GA-2008-215193). Dr. Anne

Galarneau (Institute Charles Gerhardt, CNRS Montpellier) is gratefully acknowledged
for providing the samples of MCM41.
References
[1] D. E. De Vos, M. Dams, B. F. Sels, P. A. Jacobs, Chem. Rev. 102 (2002) 3615–
3640.
[2] C. E. Song, S. Lee, Chem. Rev. 102 (2002) 3495–3524.
[3] C. Li, H. Zhang, D. Jiang, Q. Yang, Chem. Commun. (2007) 547–558.
[4] J. M. Fraile, J. I. García, J. A. Mayoral, Chem. Rev. 109 (2009) 360–417.
[5] J. M. Thomas, R. Raja, Acc. Chem. Res. 41 (2008) 708–720.
[6] J. M. Fraile, J. I. García, C. I. Herrerías, J. A. Mayoral, E. Pires, Chem. Soc. Rev. 38
(2009) 695–706.
[7] G. Desimoni, G. Faita, K. A. Jorgensen, Chem. Rev. 111 (2011) PR284–PR437.
[8] D. Rechavi, M. Lemaire, Chem. Rev. 102 (2002) 3467–3494.
[9] J. M. Fraile, J. I. García, J. A. Mayoral, Coord. Chem. Rev. 252 (2008) 624–646.
[10] J. M. Fraile, J. I. García, C. I. Herrerías, J. A. Mayoral, E. Pires, L. Salvatella,
Catal. Today 140 (2009) 44–50.
[11] J. K. Park, S.-W. Kim, T. Hyeon, B. M. Kim, Tetrahedron: Asymmetry 12 (2001)
2931–2935.
[12] T. M. Lancaster, S. S. Lee, J. Y. Ying, Chem. Commun. (2005) 3577–3579.
[13] S. S. Lee, S. Hadinoto, J. Y. Ying, Adv. Synth. Catal. 348 (2006) 1248–1254.
[14] S. S. Lee, J. Y. Ying, J. Mol. Catal. A 256 (2006) 219–224.
[15] J. Lim, S. N. Riduan, S. S. Lee, J. Y. Ying, Adv. Synth. Catal. 350 (2008) 1295–
1308.
[16] R. J. Clarke, I. J. Shannon, Chem. Commun. (2001) 1936–1937.
[17] J.-M. Lee, J. Kim, Y. Shin, C.-E. Yeom, J. E. Lee, T. Hyeon, B. M. Kim,
Tetrahedron: Asymmetry 21 (2010) 285–291.
[18] A. R. Silva, H, Albuquerque, S. Borges, R. Siegel, L. Mafra, A. P. Carvalho, J.
Pires, Microporous Mesoporous Mater. 158 (2012) 26–38.
[19] A. R. Silva, L. Carneiro, A. P. Carvalho, J. Pires, Catal. Sci. Technol. 3 (2013)
2415–2424.
[20] M. A. O. Lourenço, L. Carneiro, A. Mayoral, I. Díaz, A. R. Silva, P. Ferrerira, J.
Catal. 320 (2014) 63–69.

[21] Z.-H. Li, Z.-M. Zhou, X.-Y. Hao, J. Zhang, X. Dong, Y.-Q. Liu, Chirality 24
(2012) 1092–1095.
[22] Y. Wan, P. McMorn, F. E. Hancock, G. J. Hutchings, Catal. Lett. 91 (2003) 145–
148.
[23] S. Telalović, U. Hanefeld, Appl. Catal. A 372 (2010) 217–223.
[24] As an example of immobilization on inorganic mesoporous ordered materials: P.
Piaggio, P. McMorn, C. Langham, D. Bethell, P. C. Bulman-Page, F. E. Hancock, G. J.
Hutchings, New J. Chem. 22 (1998) 1167–1169.
[25] As an example of immobilization on organically modified mesoporous ordered
materials: S. Xiang, Y. Zhang, Q. Xin, C. Li, Chem. Commun. (2002) 2696–2697.
[26] As an example of electrostatic immobilization on mesoporous ordered materials: H.
H. Wagner, H. Hausmann, W. F. Hölderich, J. Catal. 203 (2001) 150–156.
[27] As an example of immobilization through hydrogen bonding on mesoporous
ordered materials: F. M. de Rege, D. K. Morita, K. C. Ott, W. Tumas, R. D.Broene,
Chem. Commun. (2000) 1797–1798.
[28] R. A. Feldman, J. M. Fraile, Appl. Catal. A 485 (2014) 67–73.
[29] J. M. Fraile, J. I. García, J. A. Mayoral, T. Tarnai, M. A. Harmer, J. Catal. 186
(1999) 214–221.
[30] J. M. Fraile, J. I. García, C. I. Herrerías, J. A. Mayoral, M. A. Harmer, J. Catal. 221
(2004) 532–540.
[31] J. M. Fraile, J. I. García, C. I. Herrerías, J. A. Mayoral, O. Reiser, A. Socuéllamos,
H. Werner, Chem. Eur. J. 10 (2004) 2997–3005.
[32] P. Piaggio, P. McMorn, D. Murphy, D. Bethell, P. C. Bulman Page, F. E. Hancock,
C. Sly, O. J. Kerton, G. J. Hutchings, J. Chem. Soc., Perkin Trans. 2 (2000) 2008–2015.
[33] S. Taylor, J. Gullick, P. McMorn, D. Bethell, P. C. Bulman Page, F. E. Hancock, F.
King, G. J. Hutchings, J. Chem. Soc., Perkin Trans. 2 (2001) 1724–1728.
[34] N. A. Caplan, F. E. Hancock, P. C. Bulman Page, G. J. Hutchings, Angew. Chem.,
Int. Ed. Engl. 43 (2004) 1685–1688.
[35] J. M. Kim, J. H. Kwak, S. Jun, R. Ryoo, J. Phys. Chem. 99 (1995) 16742–16747.
[36] A. Galarneau, M. Cangiotti, F. di Renzo, F. Sartori, M. F. Ottaviani, J. Phys. Chem.
B 110 (2006) 20202–20210.
[37] M. Wang, H. Jiang, Z. C. Wang, J. Therm. Anal. Calorim. 85 (2006) 751–754.
[38] J. Tian, H. Jiang, H. Gong, Z. Sun, J. Therm. Anal. Calorim. 77 (2004) 825–831.
[39] V. Sage, J. H. Clark, D. J. Macquarrie, J. Catal. 227 (2004) 502–511.

[40] K. Wilson, A. Rénson, J. H. Clark, Catal. Lett. 61 (1999) 51–55.
[41] E. Loeser, S. Babiak, M. DelaCruz, P. Karpinski, J. Chromatogr. Sci. 49 (2011)
57–62.
[42] J. C. Halle, K. H. Stern, Corros. Sci. 20 (1980) 1139–1142.
[43] J. I. García, B. López-Sánchez, J. A. Mayoral, E. Pires, I. Villalba, J. Catal. 258
(2008) 378–385.
[44] G. Zhang, J. Long, X. Wang, W. Dai, Z. Li, L. Wu, X. Fu, New J. Chem. 33 (2009)
2044–2050.
[45] A. Tschöpe, M. L. Trudeau, J. Y. Ying, J. Phys. Chem. B 103 (1999) 8858–8863.
[46] H. H. Wagner, H. Hausmann, W. F. Hölderich, J. Catal. 203 (2001) 150–156.
[47] A. R. Silva, H. Albuquerque, A. Fontes, S. Borges, Â. Martins, A. P. Carvalho, J.
Pires, Ind. Eng. Chem. Res. 50 (2011) 11495–11501.
[48] S. Taylor, J. Gullick, P. McMorn, D. Bethell, P. C. Bulman-Page, F. E. Hancock,
F. King, G. J. Hutchings, J. Chem. Soc., Perkin Trans. 2 (2001) 1714–1723.
[49] L. R. Snyder, Principles of Adsorption Chromatography, Marcel Dekker, New
York, 1968.
[50] R. Díaz-Torres, S. Álvarez, Dalton Trans. 40 (2011) 10742–10750.
Table 1. Transition points from TGA of Cu(OTf)2 and Zn(OTf)2 deposited on different
materials by incipient wetness.a
Triflate
Cu
Support
T1 (˚C)
400
323
326
341
280
276
488
284
Tm (˚C)
460
T2 (˚C)
475
367
373
441
329
327
533
340
none
Silica (70-200 µm )
Silica (35-70 µm)
MCM41
356
363
400
Al-MCM41(15)
Al-MCM41(10)
none
306
306
Zn
527
Al-MCM41(15)
310
a 1.5 wt% Cu (or Zn). The solids were dried under vacuum for 12 h prior TGA (from r.t.
to 500˚C at 10˚C/min).

Table 2. Composition and textural properties of the mesoporous supports and the Cu-
Al-MCM41 (1.5 wt% Cu) catalysts.
% Cu
Surface
Pore volume
Pore diameter
Support
loading area (m2/g) (ml/g)
(Å)
92
99
36
36
36
Al-MCM41(10)
-
628
532
860
930
967
1.2
1.4
0.6
0.7
0.7
1.5
-
Al-MCM41(15)
MCM41
1.5
-
Table 3. Results of the cyclopropanation of styrene with EDA catalyzed by Cu-Al-
MCM41(10)(Ar) modified with azatBu.
Conditionsa
Run Yield TONb cis/
% ee
(%)
44
3
trans
cis
trans
93
Homogeneous
Incipient wetnessc
Heterogeneous
-
55
4
24:76 87
1
1
2
3
24:76 82
36:64 79
38:62 81
36:64 68
91
46
17
8
58
21
10
87
87
70
a
At room temperature, EDA (2.5 mmol) in CH2Cl2 (0.5 mL) slowly added (2 h) to a
solution of styrene (2.5 mmol) in CH2Cl2 (2 mL) containing the catalyst (0.025 mmol
b
c
Cu). Mol of cyclopropanes produced per mol of copper. Catalyst prepared by
incipient wetness of the preformed azatBu-Cu(OTf)2 complex (ref. [28]).
Table 4. Results of the cyclopropanation of an excess of styrene with EDA catalyzed by
Cu-Al-MCM41(15)(Ar).
Conditionsa
A (initial heating at
90ºC)
Run
1
Yield (%) TONb cis/trans
17
40
43
43
21
50
54
54
55:45
52:48
50:50
47:53
2
3
4

5
1
2
3
4
5
1
2
3
4
5
38
34
60
59
56
61
47
61
59
65
59
48
43
75
74
70
76
59
76
74
81
74
45:55
56:44
54:46
51:49
48:52
43:57
55:45
53:47
49:51
43:57
38:62
B (continuous
heating at 90ºC)
C (continuous
heating at 100ºC)
a
EDA (2.5 mmol) in CH2Cl2 (0.5 mL) slowly added (15 min) to a suspension of
catalyst (0.025 mmol Cu) in styrene (2 mL) heated at the required temperature. b Mol of
cyclopropanes produced per mol of copper.
Table 5. Results of the cyclopropanation of an excess styrene with EDA at 90ºC
catalyzed by Cu-Al-MCM41(Ar) modified with azatBu.a
Support
Run Yield
(%)
TONb cis/
% ee
cis
68
70
71
68
69
69
66
62
65
66
66
65
65
trans
trans
74
76
74
74
72
73
62
68
68
70
72
70
72
Al-MCM41(10)
1
2
3
4
5
6
7
1
2
3
4
5
6
69
53
54
38
33
37
8
86
66
68
48
41
46
10
36
39
43
26
25
21
49:51
50:50
50:50
47:53
47:53
48:52
50:50
44:56
46:54
48:52
51:49
49:51
51:49
Al-MCM41(15)
29
31
34
21
20
17

a Reaction conditions as in Table 4. b Mol of cyclopropanes produced per mol of copper.
Captions for figures
Figure 1. TGA curves of Cu(OTf)2 in pure form and deposited on different solids by
incipient wetness.
Figure 2. Hypothesis for copper(II) triflate transformation after impregnation on Al-
MCM41 and thermal treatment.
Figure 3. Nitrogen BET isotherms of Al-MCM41(10) (blue) and Cu-Al-MCM41(10)
with 1.5 wt% Cu (red).
Figure 4. (S)-tBu-azabox ligand (azatBu).
Figure 5. Cyclopropanation of styrene with ethyl diazoacetate (EDA).
Figure 6. Results (yields in bars, %ee trans in lines) obtained in the cyclopropanation of
styrene with ethyl diazoacetate catalyzed by Cu-Al-MCM41(10) modified with azatBu:
treated in air (black) and Ar (dotted).
Figure 7. Yields obtained in the cyclopropanation of styrene with ethyl diazoacetate
catalyzed by Cu-Al-MCM41(10)(air) (1.5wt% Cu) in the continuous flow setup.
HIDDEN TEXT FOUND PLEASE VERIFY IT FROM INPUT DOC FILE THEN
DELETE IT IF NOT REQUIRE::::::********************
ꢀ
</errorimage>

Figure1 .

Figure2 .

Figure3 .

Figure4 .

Figure5 .

Figure6 .

Figure7 .

Graph-Abst-Fraile .