Electrostatic immobilization of bis(oxazoline)-copper complexes on mesoporous crystalline materials: Cation exchange vs. incipient wetness methods

Article abstract of DOI:10.1016/j.apcata.2014.07.030

Chiral copper complexes with bis(oxazoline) and azabis(oxazoline) ligands have been immobilized by non-covalent interactions on mesoporous materials. The presence of negative charges achieved by the isomorphous substitution of silicon by aluminium is required to retain the complexes on the support. Cation-exchange was found to be not as efficient as observed with layered clays, and produce catalysts which lose their chiral ligand and give low enantioselectivity. A method employing incipient wetness produces more stable catalysts, whose activity varies to some extent according to the structure of the mesoporous material on which it is supported, and the nature of the chiral ligand.

Full text of DOI:10.1016/j.apcata.2014.07.030

Applied Catalysis A: General 485 (2014) 67–73
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Electrostatic immobilization of bis(oxazoline)-copper complexes on
mesoporous crystalline materials: Cation exchange vs. incipient
wetness methods
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
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 28 May 2014
Received in revised form 18 July 2014
Accepted 24 July 2014
Available online 2 August 2014
Chiral copper complexes with bis(oxazoline) and azabis(oxazoline) ligands have been immobilized by
non-covalent interactions on mesoporous materials. The presence of negative charges achieved by the
isomorphous substitution of silicon by aluminium is required to retain the complexes on the support.
Cation-exchange was found to be not as efficient as observed with layered clays, and produce catalysts
which lose their chiral ligand and give low enantioselectivity. A method employing incipient wetness
produces more stable catalysts, whose activity varies to some extent according to the structure of the
mesoporous material on which it is supported, and the nature of the chiral ligand.
Keywords:
Asymmetric catalysis
Immobilized catalysts
Mesoporous materials
Bis(oxazolines)
Copper
© 2014 Elsevier B.V. All rights reserved.
1. Introduction
clays, that provides a flat surface of contact for the complex, has
been proposed to be the origin of this effect.
Enantioselective catalysts immobilized on solid supports [1–3]
have the practical advantages of heterogeneous catalysts, regarding
easy recovery and reuse. However, the additional cost of the
preparation, together with the unpredictable effects of immobiliza-
tion, hinders their application on an industrial scale. Non-covalent
immobilization methods [4] using already existing unmodified lig-
ands and complexes, significantly reduce the synthetic cost and
effort of immobilization. Among the different methods utilizing
non-covalent support–catalyst interactions, electrostatic immo-
bilization is very useful and practical to immobilize charged
complexes, as our group has demonstrated in the case of cop-
per (II) complexes with different types of chiral ligands, such as
bis(oxazolines) [5], azabis(oxazolines) [6], quinolineoxazolines [7],
and pyridine-azaoxazolines [8]. The most successful support has
been laponite, which is a synthetic clay with the intrinsic advan-
tages of having small particles, and reliably consistent composition.
Importantly, the electrostatic immobilization on this support leads
to important effects on diastereo- and enantioselectivities of cyclo-
propanation [5,9] and other reactions [10–13], which have not been
detected with other types of support. The lamellar nature of the
Confinement effects have also been described for mesoporous
materials with different complexes immobilized through both
covalent bonds and non-covalent interactions [14]. In such cases,
the effects have been explained by the support causing restrictions
on the conformational freedom of the complex, and controlling or
limiting the number of attack trajectories of the reaction substrates
[15]. In spite of the usefulness of bis(oxazoline)-copper complexes
in homogeneous catalysis [16], the number of examples of their
immobilization on mesoporous materials is limited. Many works
concern the covalent immobilization of complexes on mesocellular
foam silicas [17–21] or other supports such as MCM41 and MCM48
[22], hierarchically ordered mesocellular mesoporous silica [23],
SBA15 and HMS [24,25]. On the contrary, the examples of non-
covalent immobilization are more scarce, but include hydrogen
bonding of ion-tagged box-ligands on SBA15 [26] and electrostatic
interactions with Al-MCM41 and Al-SBA15 [27] or Al-TUD1 [28]. In
this paper we present our preliminary results of the electrostatic
immobilization of bis(oxazoline)- and azabis(oxazoline)-copper
complexes (Fig. 1), on Al-MCM41 and Al-HMS as supports.
2. Experimental
∗
Al-MCM41 was prepared from sodium aluminate and fumed
silica in a basic medium with cetyl trimethyl ammonium as the
Corresponding author. Tel.: +34 876553514.
E-mail address: jmfraile@unizar.es (J.M. Fraile).
http://dx.doi.org/10.1016/j.apcata.2014.07.030
0926-860X/© 2014 Elsevier B.V. All rights reserved.

68
R.A. Feldman, J.M. Fraile / Applied Catalysis A: General 485 (2014) 67–73
S-4500 I scanning electron microscope (SEM). The order in the sil-
ica network was examined by transmission electron microscopy
(TEM) images obtained using a Philips CM 20 apparatus operating
at 100 keV.
N
N
O
O
O
O
O
O
N
N
N
N
N
N
Ph
Ph
azatBu
azaiPr
2.3. Catalytic tests
boxPh
Fig. 1. Ligands used in this work.
2.3.1. Homogeneous cyclopropanation
At room temperature, ethyl diazoacetate (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 analyzed by CG [5–9].
template [29]. In one case (Al-MCM41 with Si/Al = 10), 1,3,5-
trimethylbenzene was also added to the gel to expand the pore size
[30]. Al-HMS was prepared from tetraethoxy silane and aluminium
isopropoxide with hexadecylamine in alcoholic solvent [31].
2.1. Immobilization of the catalysts
2.3.2. Heterogeneous cyclopropanation
The complex was prepared by dissolving the ligand (0.15 mmol)
in dichloromethane (1 mL) and adding it to copper (II) triflate
(40 mg, 0.1 mmol) suspended in dichloromethane (2 mL). After
stirring for 20 min, the solution was microfiltered to ensure the
removal of uncomplexed copper (II).
The supported catalyst (50 mg) was suspended in either styrene
or dichloromethane (2 mL) with n-decane (100 mg) and ethyl
diazoacetate (2.5 mmol, 285 mg) dissolved in dichloromethane
(0.5 mL) was added over the course of 2 h with a syringe pump. After
the addition, the reaction was left to stir for 30 min and then ana-
lyzed by CG. The catalyst was filtered, rinsed with dichloromethane
(5 mL) and then added directly to the next reaction.
2.1.1. Immobilization by cation-exchange of the complex
After the preparation of the complex the solvent was evapo-
rated under a flow of argon to yield a green solid. The complex
was redissolved in the exchange solvent (methanol, acetonitrile or
nitromethane, 5 mL) and the pre-calcined support (1 g) was added
to this solution. The suspension was gently stirred for 24 h at room
temperature, and the solid was filtered, and then dried under vac-
uum for 12 h prior to use.
3. Results and discussion
The mesoporous supports were prepared by well-established
methods [29–31] and the textural properties and composition
(Table 1) were in agreement with the expected values. The absence
of extraframework octahedral Al in Al-MCM41 was confirmed by
27Al MAS NMR, but a small amount was detected in Al-HMS, prob-
ably because the preparation protocol requires that washing is not
done after the filtration of the as-synthesized material. In the case
of Al-MCM41(10) the expansion in the pore size has the effect of
severely disrupting the symmetry of the support, such that hexag-
onal symmetry features of the MCM41 structure, are not clearly
observed by small angle XRD.
2.1.2. Immobilization by cation-exchange of copper and
modification with chiral ligand
In a typical experiment, copper (II) triflate (110 mg, 0.3 mmol)
was dissolved in acetonitrile (5 mL), calcined Al-MCM41(15) (1 g)
was added to the solution, and the suspension was stirred for 24 h
at room temperature. The solid was filtered, washed with methanol
in a Soxhlet extractor for 24 h and then dried under vacuum for 12 h.
This solid can be used as catalyst for non-enantioselective reactions
or modified with chiral ligand. In order to be modified by chiral
ligand, the supported copper(II) (1 g) was added to a solution of the
ligand (0.45 mmol) in dichloromethane (5 mL), and the suspension
was stirred for 24 h at room temperature. The solid was filtered and
then dried under vacuum for 12 h prior to use.
3.1. Immobilization by cation-exchange of the pre-formed
complex
The simplest method of electrostatic immobilization is the
direct cation exchange of the pre-formed complex on the sup-
port (Scheme 1). This method has been commonly used in the
case of immobilization on laponite clay [5–13]. The bis(oxazoline)-
copper(II) triflate complex is stirred with the support in a solvent
with a high dielectric constant, able to dissolve the sodium triflate
generated in the exchange process, which should then subse-
quently eliminated by filtration. The amount of copper complex
used in the exchange process was selected to be 0.3 mmol/g, far
below the theoretical exchange capacity, and was chosen in an
attempt to prevent immobilization by other kinds of interactions,
and to prevent pore blocking with excess copper complex. The
green solids were tested in the benchmark cyclopropanation reac-
tion of styrene with ethyl diazoacetate (Scheme 2).
2.1.3. Immobilization by incipient wetness of the complex
After the preparation of the complex the solvent was evapo-
rated under a flow of argon to yield a green solid. The complex
was redissolved in the required amount of methanol (0.6 mL for
Al-MCM41(10) and 0.3 mL for Al-MCM41(15)), the calcined sup-
port (0.5 g) was added and the flask was shaken to ensure the even
distribution of the solution. The solid was dried under vacuum for
12 h prior to use.
2.2. Characterization methods
Powder X-ray diffraction (XRD) data were obtained on a CGR
Theta-60 diffractometer with an Inel drive, using monochromated
Cu K␣ radiation. The adsorption–desorption isotherms of nitro-
gen 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. The aluminium analysis was deter-
mined using Energy-dispersive X-ray spectroscopy (EDX) on the
calcined solid. Particle morphology was observed using a Hitachi
Table 1
Composition and textural properties of the mesoporous supports.
Support
Si/Al
Surface area
(m2/g)
Pore volume
(mL/g)
Pore diameter
(Å)
Al-MCM41(10)
Al-MCM41(15)
MCM41
10
15
∞
628
860
967
1.2
0.6
0.7
0.9
92
36
36
33
Al-HMS-(25)
25
1029

R.A. Feldman, J.M. Fraile / Applied Catalysis A: General 485 (2014) 67–73
69
2+
X
O
O
1. Direct cation-
exchange of
the complex
L*Cu(OTf)2
MeOH
N
N
Cu
Na
R
R
Na
Cu(OTf)2
Al-MCM41
MeOH
2+
2. Cation-exchange
X
O
O
L*
CH2Cl2
of copper and
[(MeOH2)4Cu]2+
N
N
modification with
the chiral ligand
Cu
R
R
Scheme 1. Processes of immobilization by cation-exchange.
The triflate counter-ion has been described as acting as a linker
through hydrogen bonding to the support for bis(oxazoline)–metal
complexes on the silica surface [32–35]. To be sure that the alu-
minium sites in Al-MCM41 were playing some role in supporting
the complex, the complex was deposited on a purely siliceous
MCM41 solid with similar textural properties to Al-MCM41(15)
(Table 1). The green solid that was obtained following the imple-
mentation of the cation-exchange conditions lost its colour in an
instant when it was rinsed with a small amount of methanol and the
material left over was not active in the cyclopropanation reaction.
This is in contrast to when the complexes are supported on Al-
MCM41 using the same conditions, but washed with much larger
volumes of methanol. Using Al-MCM41, the solids always retain
their green colour after washing and are active for at least one reac-
tion. Thus, in this way one can infer that the aluminium sites play
a decisive role in the immobilization of our Cu complexes.
Logically, the loss of chiral ligand during the exchange pro-
cess may be a consequence of the steric interactions of the bulky
complex with the material walls, within the mesopores. When
Al-MCM41(10) was used as support the results were significantly
better than those obtained with Al-MCM41(15) (Table 2). The yield
was improved over the result obtained on Al-MCM41(15), the
diastereoselectivity was very similar, albeit with a slightly higher
amount of cis isomer being formed. The enantioselectivity was also
better such that a 50% ee was obtained in the trans cyclopropanes
Scheme 2. Cyclopropanation of styrene with ethyl diazoacetate.
In the beginning, an attempt was made to support the boxPh-Cu
complex on Al-MCM41(15), since this is a complex has been com-
monly supported on laponite. The yield produced by the supported
complex (Table 2) was much lower than that of the homogeneous
complex, and the enantioselectivity was significantly lower. If this
result were caused by a confinement effect, as happened in the case
of laponite [5,9], a significant variation in cis/trans diastereoselec-
tivity would be expected. Since this is not the case, a more obvious
cause for the reduction in enantioselectivity would be there being
an almost complete decomplexation of the chiral ligand from the
copper during the cation-exchange process, which results from a
low stability of the complex. The low yield could potentially result
from a partial pore blocking of the support.
Table 2
Results of the cyclopropanation of styrene with ethyl diazoacetate catalyzed by solids prepared by cation-exchange of the pre-formed box-Cu complexes.
Ligand
boxPh
Support
Reaction cond.a
Run
Yield (%)
TONb
cis/trans
% ee
cis
trans
None
Al-MCM41(15)
Al-MCM41(10)
Homog.
Stoich.
Stoich.
–
1
1
2
3
63
10
24
23
23
63
17
40
38(78)
38(116)
30:70
34:66
39:61
43:57
46:54
51
16
32
32
28
72
18
50
44
37
azaiPr
None
Al-MCM41(10)
Homog.
Stoich.
–
1
2
3
4
1
2
3
49
44
34
43
11
44
7
49
73
57(130)
72(202)
–
73
12(85)
–
30:70
37:63
36:64
36:64
39:61
47:53
48:52
45:55
70
42
44
44
0
46
38
n.d.
83
65
64
63
0
52
33
n.d.
Styrene
1
a
Homog.: reaction in homogeneous phase with the LCu(OTf)2 complex. Stoich.: heterogeneous reaction with stoichiometric amounts of styrene and ethyl diazoacetate.
Styrene: heterogeneous reaction with a large excess of styrene.
b
TON, mmol of cyclopropanes per mmol of Cu. In parenthesis the cumulative TON in all the runs with similar enantioselectivities.

70
R.A. Feldman, J.M. Fraile / Applied Catalysis A: General 485 (2014) 67–73
vs. 18% ee with the same complex supported in the smaller pores
of Al-MCM41(15). This result could imply that there is a very
important role of the pore size in determining the catalytic perfor-
mance of our catalysts. The higher enantioselectivity in the larger
pore would show that the complex is more stable, and one would
also expect there to be lower diffusion limitations on the com-
plex and reaction substrates than there would be in smaller pores.
Nonetheless, in spite of these improved results, upon recycling, the
enantioselectivity steadily decreases following each run. This loss
of enantioselection would be explained by the chiral ligand pro-
gressively being separated from the copper centres due to a lack of
stability, or being displaced by the by-products of the cyclopropa-
nation reaction [36].
Taking into account this hypothesis, we decided to attempt the
immobilization of an azabox-Cu complex, since azabox complexes
are more stable than the analogous box-Cu [6]. Using the support
with the largest pore size, Al-MCM41(10), with the hope to having
the least steric problems, as expected, azaiPr-Cu(II) immobilized on
Al-MCM41(10) produced better results (Table 2). Yield and cis/trans
diastereoselectivity were very similar to those obtained in solu-
tion, but the enantioselectivity was lower than in homogeneous
phase; with a 65% ee in the trans isomers with the supported cata-
lyst in comparison with 83% ee in solution. In contrast to the result
obtained with boxPh, this catalytic performance was stable during
3 consecutive cycles, leading to a cumulative TON of 200, but with
an almost complete deactivation by the fourth cycle. Thus, in this
example, a more stable complex gives better initial activity follow-
ing cation-exchange and better performance when re-used, but the
life time of the catalyst is still limited.
In the case of laponite-supported catalysts, the changes in selec-
tivities of the supported complexes were enhanced by the use of
solvents with low dielectric constant [9], which favour the close
proximity of the support-catalyst ion pair. A reaction was there-
fore performed with a large excess of styrene using azaiPr-Cu(II)
immobilized on Al-MCM41(10) (Table 2). Comparing the result
obtained with Al-MCM41(10) and laponite, the parallel exists that
excess styrene produces a slight increase in the amount of cis-
diastereoisomer formed, but with Al-MCM41(10), it also produces
a slightly lower enantioselectivity in the trans products. There is no
doubt that in terms of enantioselection, using excess styrene with
the complexes supported on laponite, gives a more profound effect
than what is seen with MCM41. Furthermore, with Al-MCM41(10),
the use of the larger amount of styrene does not improve the yield
(representative for chemoselectivity of the process with respect to
side reactions of diazoacetate). On the contrary, the use of excess
styrene produced a rapid deactivation, resulting in an almost com-
plete deactivation by the second cycle. Following the reactions in
excess styrene, one can see unmistakably that the support becomes
coated with hydrophobic polymeric products that can also be seen
by thermal gravimetric analysis, and this would point to side reac-
tions of styrene, such as oligomerization, as being the cause for this
faster deactivation.
Fig. 2. Effect of exchange solvent on the cyclopropanation yield (continuous lines)
and enantioselectivity of the trans isomers (dashed lines): methanol (ꢀ), acetonitrile
(ꢁ), nitromethane (ꢂ). Grey lines represent the values in the homogeneous reaction.
that the dielectric constants of the three solvents are quite similar,
any differences that result from solvent effects could be atributed
to the lower coordinating ability of acetonitrile and nitromethane,
that can be represented by their donor number values (DN
[38]: acetonitrile = 14.1, nitromethane = 2.7, methanol = 30). If the
exchange solvent is less prone to substitute itself for the position of
the chiral ligand, then the overall enantioselectivity of the catalyst
should be higher, at least at the outset.
However, once again, the intrinsic lack of stability of the com-
plex leads to a progressive loss of chiral ligand by substitution with
reaction products and/or by-products. The extent of the drop does
not change with respect to the exchange solvent, and this is shown
by the steady decline in the enantioselection that occurs for all of
the catalysts. One would favour the hypothesis of loss of ligand over
deactivation of the copper(II) centres because while the enantios-
election falls, the yield remains relatively stable. Moreover, when
ethyl diazoacetate was added to the solutions that are obtained
after the separation of the catalyst by filtration, all of the solutions
were catalytically active and this would indicate the leaching of
active copper species.
3.2. Complexation of copper-exchanged mesoporous support
with chiral ligand
Given the problems to prepare selective and stable catalysts
by direct exchange of pre-formed complexes, another strategy
that consists of two steps was formulated (Scheme 1), that
involves the exchange of copper(II) on the support, and the sub-
sequent formation of the complex by the addition of ligand using
a non-coordinating solvent of low dielectric constant, such as
dichloromethane. In this way the diffusion of the chiral ligand
through the pores would be improved by the higher conforma-
tional freedom of the non-bound chiral ligand compared to the full
complex. Once the copper centres are installed, complex forma-
tion could potentially be optimized by using an excess of ligand
with respect to copper. For the first step, calcined Al-MCM41 was
added to a solution of the copper(II) triflate in methanol, contain-
ing a quantity of copper corresponding to a final content of 3.8 wt%
Cu. One half of the copper that was added to the cation exchange
suspension was detected on the support, thus giving a loading of
1.9 wt%. Simple washing with small amounts of MeOH did not affect
the copper loading but on the other hand, Soxhlet extraction with
methanol for 12 h reduced the copper content to 1.3 wt%. After
The choice of solvent is important for cation exchange of com-
plexes [37], as there is a competitive equilibrium that takes place
between the box ligand and solvent coordination on the copper(II)
centres, and if the balance of this equilibrium is disfavourable, it
results in there being a larger number of non-chiral copper sites
complexed with solvent molecules. For the reason of its larger pore
size, Al-MCM41(10) was used as support to immobilize boxPh-
Cu(II) by direct exchange, using different solvents and the results
obtained in the cyclopropanation reaction using these catalysts are
presented in Fig. 2.
When acetonitrile or nitromethane were used instead of
methanol,
a higher initial enantioselectivity in the trans-
diastereoisomer is observed, even higher than the result obtained
by the homogenous phase catalyst in the case of acetonitrile. Being

R.A. Feldman, J.M. Fraile / Applied Catalysis A: General 485 (2014) 67–73
71
Table 3
Results of the cyclopropanation of styrene with ethyl diazoacetate catalyzed by
Cu(II)-exchanged Al-MCM41.
Support
Reaction cond.a
Yield (%)
TONc
cis/trans
Noneb
Al-MCM41(15)
Styrene
Stoich.
Styrene
Stoich.
Styrene
47
32
43
32
49
47
78
105
78
30:70
44:56
47:53
39:61
53:47
Al-MCM41(10)
120
a
Stoich.: reaction with stoichiometric amounts of styrene and ethyl diazoacetate.
Styrene: reaction with a large excess of styrene.
b
Reaction catalyzed by Cu(OTf)2.
TON, mmol of cyclopropanes per mmol of Cu.
c
Soxhlet extraction, it is reasonable to suppose that weakly or only
partially bound copper had been removed.
Fig. 3. Proposed ion-pairing interaction after incipient wetness immobilization.
Copper(II) triflate was supported on both the Al-MCM41(10) and
Al-MCM41(15) solids in this way and then tested in the cyclopropa-
nation of styrene with ethyl diazoacetate, without the use of chiral
ligand, in order to independently examine the effect of the ligand
and support (Table 3). The result was that in each case, a higher
level of cis-diastereoisomers was produced than the homogenous
phase analogues. One could attribute this effect to the elimination
of weakly supported copper(II) following the soxhlet extraction
phase, and the slightly raised level of the cis-diastereomers being
due to a support interaction such as described with laponite [9]. The
copper supported on Al-MCM41(10) produced the highest level of
cis-diastereomers and one could speculate that the origin of this
effect could be the larger diameter of the support. The larger the
pore diameter would geometrically produce an internal surface
that would be relatively flat, and so in this way, the surface more
closely approximates the surface of laponite.
After one reaction, the copper content of the solid drops from
1.3 to 0.8 wt%. The fact that this occurs after Soxhlet extrac-
tion, and that there is no other species present that would be
expected to exchange with the copper(II) sites, would suggest that
at least some of the copper(II) is only partially exchanged onto the
support. Considering that the first step of the cyclopropanation
reaction is a Cu(II) to Cu(I) reduction, the initial presence of [Al-
MCM41]−[CuOTf]+ species would produce either [Al-MCM41]−Cu+,
that would remain on the solid, or CuOTf, that would have a higher
chance of being leached into the solution.
In spite of anticipating leaching, the formation of the chiral
complex was implemented by treating this Cu(II) supported on Al-
MCM41(10) with 1.5 molar equivalents (with respect to the amount
of copper) of boxPh in dichloromethane at room temperature. In the
beginning the complex formation step was done over the course of
12 h to favour the difusion of the ligand inside the pores, and the
partial sucess of this effort was demonstrated by the cyclopropa-
nation results (Method A, Table 4) that show a moderate initial
activity and enantioselectivity (40% ee in the trans isomers). A sig-
nificant drop that occurs when the catalyst is re-used was due in
part to copper leaching, demonstrated by the activity of the filtrate,
but also to the loss of ligand, demonstrated by the drop in enantio-
selectivity. The effect of the difusion of the ligand was studied by
treatment for a much shorter time of 5 min (Method B, Table 4). This
catalyst was less active in terms of its yield. Should the result of the
shorter dispersion time be that the ligand is densely distributed
around outer periphery of the pores, this would inevitably block
some copper sites from reacting, and would thus reduce the yield
of the reaction. On the other hand, this catalyst gave slightly higher
enantioselectivity and was more stable upon recycling.
as straightforward, and a new methodology based on incipient
wetness was designed.
3.3. Immobilization by incipient wetness of the pre-formed
complex
The deposition of the complex by incipient wetness refers to the
process of addition of a solution of the complex to the dry support,
using a volume that is equal to or less than its internal volume, and
in this way, the complex enters the solid by capillary action.
3.3.1. Immobilization of boxPh-Cu
BoxPh-Cu(OTf)2 was dissolved in the required amount of MeOH
and added to pre-calcined Al-MCM41(15). Adsorption by incipient
wetness was completed by removal of the solvent under reduced
pressure. As a copper loading of 0.3 mmol/g (1.9 wt%) after cation-
exchange resulted in leaching, a smaller amount of copper was
added for incipient wetness in an attempt to address this prob-
lem, and this level was 0.2 mmol/g (1.5 wt%). Whereas the intention
of cation-exchange is to separate the triflate counter-ion from the
copper sites, ion-pairing is more likely to occur in incipient wetness,
with triflate anions remaining on the solid (Fig. 3).
This catalyst was tested in the cyclopropanation between ethyl
diazoacetate and a stoichiometric quantity of styrene (Table 5).
When compared with the solid prepared by cation exchange on
the same support (Table 2) it can be seen that incipient wet-
ness produces a much more active and enantioselective catalyst,
although always with values below those obtained with the homo-
geneous catalyst.
The catalyst shows a remarkable degree of stability along four
consecutive runs, with a cumulative TON of 300. Considering
that the drop in yield of the fourth may be due to pore block-
ing, the catalyst was washed in a Soxhlet extractor for 12 h with
dichloromethane. This treatment gave an improvement in the yield,
reaching a cumulative TON higher than 400, even the enantio-
selectivity rose, and this shows how by-products can deactivate
the catalyst. The lower enantioselectivity in comparison with the
homogeneus result is likely to be due to a confinment effect within
the pores which modifies the steric environment of the catalytic
centre, however, one cannot discount the possibility of there being
a partial loss or degredation of the ligand during the impregna-
tion process. Both of these scenarios would favour the formation
of a mixture of chiral and non-chiral centres. It would be highly
desirable to determine the proportion of chirally and non-chirally
modified centres however, this would not be achieved as simply as
determinging the elemental composition of the catalyst since the
free ligand is also present on the support.
These results show that direct cation exchange on the meso-
porous Al-MCM41, either of the pre-formed box-Cu complex or
simply Cu(OTf)2, is not as simple as it was with a layered clay
such as laponite. Thus, the immobilization by this method is not
In order to examine how the structure of the support could
affect the activity, Al-HMS(25) was used to immobilize the same

72
R.A. Feldman, J.M. Fraile / Applied Catalysis A: General 485 (2014) 67–73
Table 4
Results of the cyclopropanation of styrene with ethyl diazoacetatea catalyzed by Cu(II)-exchanged Al-MCM41(10) modified with boxPh-Cu.
Methodb
Run
Yield (%)
TONc
cis/trans
% ee
cis
trans
Homog.d
63
46
12
24
23
23
14
63
112
30:70
40:60
37:63
39:61
43:57
46:54
44:56
51
31
17
32
32
28
26
72
40
17
50
44
37
34
A
B
1
2
1
2
3
4
29 (141)
59
56 (115)
56 (171)
34 (205)
a
Reactions carried out in dichloromethane with the stoichiometric amount of styrene.
Method A: treatment with ligand (1/10, w/w with respect to solid) in dichloromethane for 12 h at rt, filtration and drying under vacuum. Method B: treatment with the
b
same amount of ligand for only 5 min, evaporation of solvent and immediate use in the reaction.
c
TON, mmol of cyclopropanes per mmol of Cu. In parenthesis the cumulative TON in all the runs with similar enantioselectivities.
Reaction carried out with boxPh-Cu(OTf)2 in solution.
d
Table 5
Results of the cyclopropanation of styrene (stoichiometric amount) with ethyl diazoacetate catalyzed by boxPh-Cu(OTf)2 immobilized by incipient wetness.
Support
Run
Yield (%)
TONb
cis/trans
% ee
cis
trans
None
Al-MCM41(15)
–
1
2
3
4
5a
1
2
3
4
5
63
37
32
30
22
45
37
11
38
29
17
63
93
30:70
38:62
38:62
37:63
43:57
33:67
36:64
43:57
36:64
50:50
44:56
51
27
24
37
32
45
30
25
32
12
16
72
42
38
52
44
49
55
34
43
5
80 (173)
75(248)
55(303)
112(415)
93
28(121)
95(216)
73
Al-HMS(25)
43
24
a
Before the reaction, the supported catalyst was recovered, dried, and washed in a Soxhlet with dichloromethane for 12 h.
TON, mmol of cyclopropanes per mmol of Cu. In parenthesis the cumulative TON in all the runs with similar enantioselectivities.
b
complex under the same conditions. The performance of this cata-
3.3.2. Immobilization of azatBu-Cu
lyst (Table 5) was quite erratic, with results of yield and selectivities
that were much less stable than those obtained with Al-MCM41. In
addition to differences in the morphology and textural properties
of this support, one must consider that the Al-HMS support was
prepared using Al(OiPr)3 as the aluminium precursor, whereas
the Al-MCM41 materials were made with sodium aluminate. This
means that following the removal of the template by calcination,
the aluminim sites of Al-HMS(25) are present in their protic
form, whereas the Al-MCM41 aluminim sites are stabilized by
sodium. We speculate that the protic form could be potentially be
deleterious to the the stability of the bis(oxazoline) ligand.
Nonetheless, the merits of incipient wetness are that it is quicker
to implement than cation-exchange in a suspension, uses less sol-
vent and the copper complex is more efficiently introduced into
the mesoporous support. Moreover this approach produces more
active and stable catalysts. In view of all those advantages, this
procedure was tested with the most efficient ligand in solution.
Given the steric volume of the tert-butyl substituents, azatBu-
Cu(OTf)2 was deposited inside Al-MCM41(10) by incipient wetness
following the same procedure as with boxPh. On the first occassion
when the catalyst was used in a cyclopropanation reaction, unlike
on all previous occassions, during the addition of ethyl diazoac-
etate, the reaction did not begin immediately. One hour after all
of the ethyl diazoacetate had been added, the reaction suddenly
began and continued for a little while in an uncontrollable manner.
At this point, GC analysis showed that a significant amount of dia-
zoacetate remained unreacted, a very low amount of cyclopropane
products had been formed, but most importantly, the enantiose-
lection of this catalyst was very similar to the homogenous phase
catalyst result (Table 6).
The Cu(I) is the active species [39] in the cyclopropanation reac-
tion and since we are using a precursor of Cu(II), this Cu(II) needs to
be reduced to Cu(I) [40,41] before the catalytic cycle can proceed. In
general, the ethyl diazoacetate carries out the reduction. In another
Table 6
Results of the cyclopropanation of styrene with ethyl diazoacetate catalyzed by azatBu-Cu(OTf)2 immobilized on Al-MCM41(10) by incipient wetness.
Reaction conditionsa
Temp. (◦C)
Run
Yield (%)
TONb
cis/trans
% ee
cis
trans
Homog.
Stoich.
Styrene
25
25
90
–
1
1
2
3
44
3
56
39
21
44
8
140
98
53
24:76
24:76
24:76
31:69
35:65
87
82
69
15
10
93
91
85
15
2
a
Stoich.: reaction with stoichiometric amounts of styrene and ethyl diazoacetate. Styrene: reaction with a large excess of styrene.
TON, mmol of cyclopropanes per mmol of Cu.
b

R.A. Feldman, J.M. Fraile / Applied Catalysis A: General 485 (2014) 67–73
73
example of electrostatic immobilization of azatBu-Cu(II), for exam-
References
ple on nafion-silica, it was necessary to heat the catalyst at 50 ◦C
in the presence of ethyl diazoacetate for a few minutes in order to
promote the reduction to Cu(I) [6]. Here, the same method was not
adequate and so a series of experiments was performed in which
ethyl diazoacetate was added to a heated suspension of azatBu-
Cu(OTf)2 immobilized on Al-MCM41(10) in styrene. Below 50 ◦C
the reaction does not take place, in the range of 50–80 ◦C reaction
occurs slowly but is not completely under control. The optimal tem-
perature was found to be 90 ◦C, as higher temperatures produced
a syrupy mixture due to styrene polymerization. Using a tempera-
ture of 90 ◦C produces an immediate reaction when the addition of
ethyl diazoacetate is initiated, and this reaction stops the moment
that the addition ethyl diazoacetaet is halted. The need for this high
temperature demonstrates the high stability of this aza-Cu(II) com-
plex against reduction, as already observed in previous work [6].
The activity and selectivities were very good in the first reaction
(Table 6), however, the enantioselectivity is virtually lost following
the first recycle. The rapid loss of activity would imply that although
the complex is very stable, it is still prone to substitution with by-
products and the gradual decrease in yield is probably indicative of
poisoning with styrene oligomers due to the high temperature.
[1] D.E. De Vos, I.F.J. Vankelecom, P.A. Jacobs (Eds.), Chiral Catalysts Immobilization
and Recycling, Wiley-VCH, Weinheim, 2000.
[2] P. Barbaro, F. Liguori (Eds.), Heterogenized Homogeneous Catalysts for Fine
Chemicals Production, Springer, Dordrecht, 2010.
[3] M. Heitbaum, F. Glorius, I. Escher, Angew. Chem. Int. Ed. 45 (2006) 4732–4762.
[4] J.M. Fraile, J.I. García, J.A. Mayoral, Chem. Rev. 109 (2009) 360–417.
[5] J.I. García, B. López-Sánchez, J.A. Mayoral, E. Pires, I. Villalba, J. Catal. 258 (2008)
378–385.
[6] 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.
[7] J.M. Fraile, J.I. García, G. Jiménez-Osés, J.A. Mayoral, M. Roldán, Organometallics
27 (2008) 2246–2251.
[8] J.I. García, G. Jiménez-Osés, B. López-Sánchez, J.A. Mayoral, A. Vélez, Dalton
Trans. 39 (2010) 2098–2107.
[9] A.I. Fernández, J.M. Fraile, J.I. García, C.I. Herrerías, J.A. Mayoral, L. Salvatella,
Catal. Commun. 2 (2001) 165–170.
[10] J.M. Fraile, J.I. García, J.A. Mayoral, M. Roldán, Org. Lett. 9 (2007) 731–733.
[11] J.M. Fraile, P. López-Ram-de-Viu, J.A. Mayoral, M. Roldán, J. Santafé-Valero, Org.
Biomol. Chem. 9 (2011) 6075–6081.
[12] M.J. Fabra, J.M. Fraile, C.I. Herrerías, F.J. Lahoz, J.A. Mayoral, I. Pérez, Chem.
Commun. (2008) 5402–5404.
[13] J.M. Fraile, N. García, C.I. Herrerías, ACS Catal. 3 (2013) 2710–2718.
[14] J.M. Fraile, J.I. García, C.I. Herrerías, J.A. Mayoral, E. Pires, Chem. Soc. Rev. 38
(2009) 695–706.
[15] J.M. Thomas, R. Raja, Acc. Chem. Res. 41 (2008) 708–720.
[16] G. Desimoni, G. Faita, K.A. Jorgensen, Chem. Rev. 111 (2011) PR284–PR437.
[17] J.K. Park, S.-W. Kim, T. Hyeonb, B.M. Kim, Tetrahedron: Asymmetry 12 (2001)
2931–2935.
4. Conclusions
[18] T.M. Lancaster, S.S. Lee, J.Y. Ying, Chem. Commun. (2005) 3577–3579.
[19] S.S. Lee, S. Hadinoto, J.Y. Ying, Adv. Synth. Catal. 348 (2006) 1248–1254.
[20] S.S. Lee, J.Y. Ying, J. Mol. Catal. A 256 (2006) 219–224.
[21] J. Lim, S.N. Riduan, S.S. Lee, J.Y. Ying, Adv. Synth. Catal. 350 (2008) 1295–1308.
[22] R.J. Clarke, I.J. Shannon, Chem. Commun. (2001) 1936–1937.
[23] J.-M. Lee, J. Kim, Y. Shin, C.-E. Yeom, J.E. Lee, T. Hyeon, B.M. Kim, Tetrahedron:
Asymmetry 21 (2010) 285–291.
[24] A.R. Silva, H. Albuquerque, S. Borges, R. Siegel, L. Mafra, A.P. Carvalho, J. Pires,
Microporous Mesoporous Mater. 158 (2012) 26–38.
[25] A.R. Silva, L. Carneiro, A.P. Carvalho, J. Pires, Catal. Sci. Technol. 3 (2013)
2415–2424.
Cation exchange was found to be unsuitable as a method for the
electrostatic immobilization of bis(oxazoline)-copper complexes
on mesoporous materials due to a combination of factors, such
as the rather poor stability of the complex, leading to non-chiral
active sites, and the steric constraints for diffusion of the rather
bulky complex inside the pores. The use of azabis(oxazoline) lig-
ands, whose copper complexes have higher formation constants,
solids with larger pores, and less coordinating exchange solvents,
minimizes these problems, but the recyclability of the catalysts in
the cyclopropanation reactions is furthermore limited by the ligand
replacement with reaction by-products.
Deposition of the complex by incipient wetness overcomes
some of the problems of the restricted diffusion in the pores and
the resulting catalyst is more active and more stable against deac-
tivation. The azatBu copper complex deposited on Al-MCM41 in
this way is of such a stability that in order to achieve the necessary
reduction of Cu(II) to Cu(I), a high reaction temperature is needed
and this leads to a fast deactivation due to styrene polymerization.
[26] Z.-H. Li, Z.-M. Zhou, X.-Y. Hao, J. Zhang, X. Dong, Y.-Q. Liu, Chirality 24 (2012)
1092–1095.
[27] Y. Wan, P. McMorn, F.E. Hancock, G.J. Hutchings, Catal. Lett. 91 (2003) 145–148.
[28] S. Telalovic´, U. Hanefeld, Appl. Catal. A 372 (2010) 217–223.
[29] J.M. Kim, J.H. Kwak, S. Jun, R. Ryoo, J. Phys. Chem. 99 (1995) 16742–16747.
[30] A. Galarneau, M. Cangiotti, F. di Renzo, F. Sartori, M.F. Ottaviani, J. Phys. Chem.
B 110 (2006) 20202–20210.
[31] T.R. Pauly, Y. Liu, T.J. Pinnavaia, S.J.L. Billinge, T.P. Rieker, J. Am. Chem. Soc. 121
(1999) 8835–8842.
[32] P. O’Leary, N.P. Krosveld, K.P. De Jong, G. van Koten, R.J.M. Klein Gebbink, Tetra-
hedron Lett. 45 (2004) 3177–3180.
[33] H. Wang, X. Liu, H. Xia, P. Liu, J. Gao, P. Ying, J. Xiao, C. Li, Tetrahedron 62 (2006)
1025–1032.
[34] C. McDonagh, P. O’Leary, Tetrahedron Lett. 50 (2009) 979–982.
[35] E.P. Carreiro, N.M.M. Moura, A.J. Burke, Eur. J. Org. Chem. (2012) 518–528.
[36] J.M. Fraile, J.I. García, J.A. Mayoral, T. Tarnai, M.A. Harmer, J. Catal. 186 (1999)
214–221.
Acknowledgements
[37] J.M. Fraile, J.I. García, C.I. Herrerías, J.A. Mayoral, M.A. Harmer, J. Catal. 221
(2004) 532–540.
[38] Y. Marcus, J. Solut. Chem. 13 (1984) 599–624.
[39] J.M. Fraile, J.I. García, V. Martínez-Merino, J.A. Mayoral, L. Salvatella, J. Am. Chem.
Soc. 123 (2001) 7616–7625.
[40] R.G. Salomon, J.K. Kochi, J. Am. Chem. Soc. 95 (1973) 3300–3310.
[41] R.L. Safiullin, V.A. Dokichev, L.R. Yakupova, R.M. Sultanova, S.L. Khursan, R.N.
Zaripov, Yu.V. Tomilov, Kinet. Catal. 49 (2008) 43–51.
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 pro-
gram no. PITN-GA-2008-215193). Dr. Anne Galarneau (Institute
Charles Gerhardt, CNRS Montpellier) is gratefully acknowledged
for providing the samples of MCM41.