Homogeneous and supported copper complexes of cyclic and open-chain polynitrogenated ligands as catalysts of cyclopropanation reactions

Article abstract of DOI:10.1002/(SICI)1099-0682(199912)1999:12<2347::AID-EJIC2347>3.0.CO;2-7

Cu(I) and Cu(II) complexes of cyclic and open-chain polyaza compounds have been tested as catalysts in the benchmark cyclopropanation reaction of styrene with ethyl diazoacetate. In general, only small amounts of copper are needed to promote the reaction. The catalytic activity depends on the structure of the ligand, e.g. amine-amides are more efficient than polyamines, and on the oxidation state of copper, Cu(II) being more active than Cu(I). Given that Cu(I) is the active species, these changes of behavior must be related to the stabilities of the complexes. The nature of the counterion also has a noticeable influence on the catalytic activity, the role of which is discussed. XAS measurements suggest the formation of oligomeric species. Some of the chiral ligands lead to small enantiomeric excesses, open-chain ligands can easily be supported on organic polymers and their complexes can be used as catalysts. Furthermore, cyclic and acyclic complexes can be supported on clays by cation exchange and the solids obtained tend to promote the reaction with a decrease in the trans/cis ratio.

Full text of DOI:10.1002/(SICI)1099-0682(199912)1999:12<2347::AID-EJIC2347>3.0.CO;2-7

FULL PAPER
Homogeneous and Supported Copper Complexes of Cyclic and Open-Chain
Polynitrogenated Ligands as Catalysts of Cyclopropanation Reactions
[a]
[a]
[b]
[b]
[b]
[c]
´
´
´
´
˜
F. Adrian, M. I. Burguete, J. M. Fraile, J. I. Garcıa, J. Garcıa, E. Garcıa-Espana,
S. V. Luis,*[a] J. A. Mayoral,*[b] A. J. Royo,[a,b] and M. C. Sanchez
[b]
´
Keywords: Cyclopropanation / Copper / N ligands / Macrocycles / Supported catalysts
CuI and CuII complexes of cyclic and open-chain polyaza has a noticeable influence on the catalytic activity, the role
compounds have been tested as catalysts in the benchmark
cyclopropanation reaction of styrene with ethyl diazoacetate.
In general, only small amounts of copper are needed to
promote the reaction. The catalytic activity depends on the
structure of the ligand, e.g. amine–amides are more efficient
than polyamines, and on the oxidation state of copper, CuII
being more active than CuI. Given that CuI is the active
species, these changes of behavior must be related to the
stabilities of the complexes. The nature of the counterion also
of which is discussed. XAS measurements suggest the
formation of oligomeric species. Some of the chiral ligands
lead to small enantiomeric excesses. Open-chain ligands can
easily be supported on organic polymers and their complexes
can be used as catalysts. Furthermore, cyclic and acyclic
complexes can be supported on clays by cation exchange
and the solids obtained tend to promote the reaction with a
decrease in the trans/cis ratio.
Given the presence of nitrogen atoms related by the sym-
metry elements of the molecule, they can be grafted onto a
flexible support without losing their symmetry.[5] It is im-
portant to note that in a significant number of examples of
heterogeneous chiral catalysts, the symmetry of the chiral
ligand is lost when it is grafted onto the support, which
may be one of the reasons for the lower efficiency of these
catalysts in inducing asymmetry.[4]
Porphyrin complexes have been described as catalysts in
several reactions,[6] including cyclopropanations.[7] Com-
plexes of other cyclic polynitrogenated systems have also
been used.[8] Since open-chain polynitrogenated com-
pounds have not been used in this context, we felt that it
would be interesting to compare several of these ligands
(cyclic and open-chain, non-chiral and chiral, amines and
amides, unsupported and supported) in a benchmark reac-
tion. In this paper, we present the results obtained using
copper complexes of these polyaza derivatives (Figure 1) as
catalysts of the cyclopropanation reaction between styrene
and ethyl diazoacetate. The effect of the oxidation state of
the copper and the nature of the counterion are also ana-
lyzed.
Introduction
The search for new ligands suitable for incorporation into
homogeneous and heterogeneous catalysts for organic reac-
tions is a field of considerable interest. As far as possible,
such ligands should be easy to prepare and they should be
amenable to facile chemical modification so as to obtain
chiral derivatives from simple, accessible starting materials.
On the other hand, their incorporation into a hetero-
geneous support has to be easily accomplished with mini-
mum structural modification. In this regard, polynitrogen-
ated compounds fulfil such requirements. These ligands can
selectively bind metal cations and can distinguish between
oxidation states.[1] They can be obtained as open-chain or
macrocyclic compounds, and also in chiral forms by the
incorporation of chiral diamines, amines, or amino acid de-
rivatives. Furthermore, the chiral auxiliaries can be incor-
porated at different sites in the molecules leading to ligands
with different symmetries. In particular, they can easily be
obtained with C2 symmetry.[2] They can also be supported
on organic or inorganic materials, which opens up the pos-
sibility of using their complexes as solid catalysts in the fine
chemicals industry.[3] The chiral compounds would, of co-
urse, open the way to chiral heterogeneous catalysts.[4]
Results and Discussion
Polyamines as Ligands
[a]
´
´
´
Departamento de Quımica Inorganica y Organica,
E.S.T.C.E, Universitat Jaume I,
´
P. O. Box 224, E-12080 Castellon, Spain
Fax: (internat.) ϩ 34-964/728214
E-mail: luiss@qio.uji.es
We first tested a cyclic ligand, 14,15,17,18-tetramethyl-
2,5,8,11-tetraaza[12]paracyclophane (L1, Figure 1) since li-
gands of this type have been show to stabilize CuI with
respect to its disproportionation into CuII and Cu0.[10] We
prepared its CuI complex in situ and used it as a catalyst
in the cyclopropanation reaction of styrene (1) with ethyl
diazoacetate (2) (Scheme 1). The results obtained are pre-
sented in Table 1.
[b]
[c]
Facultad de Ciencias,
´
Instituto de Ciencia de Materiales de Aragon,
Universidad de Zaragoza - C.S.I.C.,
E-50009 Zaragoza, Spain
E-mail: mayoral@posta.unizar.es
´
´
Departamento de Quımica Inorganica,
Universidad de Valencia,
E-46100 Burjassot (Valencia), Spain
E-mail: Enrique.Garcia-Es@uv.es
Eur. J. Inorg. Chem. 1999, 2347Ϫ2354
WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
1434Ϫ1948/99/1212Ϫ2347 $ 17.50ϩ.50/0
2347

S. V. Luis, J. A. Mayoral et al.
FULL PAPER
Figure 1. Ligands used in this work
of induction is reduced when the ligand/copper molar ratio
is decreased. The addition of an excess of the ligand is of
little consequence. Such an induction period is only ob-
served when the complex is used in a ratio of 0.001:1 with
respect to the styrene and is no longer seen when it is used
in a ratio of 0.01:1 (see entry 14, Table 1). This result indi-
cates the formation of a stable complex of the ligand with
CuI,[10] which reduces the rate of formation of the reactive
copperϪcarbene intermediate. The most rational expla-
nation for this behavior is the formation of aggregates of
the CuIϪL1 complexes, in which the copper cation is much
less accessible for the interaction with the reagents to give
the active metalϪcarbene species. The formation of aggre-
gates has recently been reported to hamper the hydrolytic
catalytic activity of some copperϪpolyamine complexes.[14]
A key feature of tetraaza[n]paracyclophanes is that a maxi-
mum of three of the four nitrogen atoms can be simul-
taneously coordinated to a metal cation. This leaves a non-
coordinated benzylic nitrogen atom, which can interact with
another metal cation (see structure 5, Figure 2), thereby fav-
oring the formation of complex aggregates. This is particu-
larly the case when the solvent is changed from water to a
less polar medium such as CH2Cl2. The formation of spec-
ies such as 5 has been observed in the solid state and has
also been inferred from the results of NMR spectroscopic
studies.[11][12] As a matter of fact, it has been shown that,
in water, tetraaza[12]paracyclophanes such as L1 can form
dinuclear copper(II) complexes (see 6, Figure 2) which, at
least in the solid state, are polymerically associated.[13] The
involvement of complexes of the general structure 6 ac-
counts for the fact that even at low L1/Cu ratios (ഠ0.5)
(entry 9, Table 1) the induction period persists. Only below
this ratio is some free copper available and hence the induc-
tion period drastically decreases (entry 12, Table 1).
Scheme 1
Table 1. Results obtained from the reaction of styrene (1) with ethyl
diazoacetate (2) promoted by CuOTfϪL1 complexes in CH2Cl2 at
25 °C[a]
Entry
Cu/L1
Cu/styrene (%) t (h)[b]
% yield[c] 4:3[c]
1
Ϫ
0.1
0.1
1.5
2.0
2.2
2.5
6.0
1.2
1.3
4.0
1.0
1.1
3.5
0.75
3.0
1.45
26
30
Ϫ
4
30
Ϫ
4
28
Ϫ
4
2.6
2.6
Ϫ
Ϫ
2.3
Ϫ
Ϫ
2.2
Ϫ
2
3
1:1.1
4
5
6
1:0.75
1:0.5
0.1
0.1
7
8
9
10
11
12
13
14
Ϫ
31
11
32
38
2.3
2.3
2.4
2.3
1:0.25
1:1.1
0.1
1
[a]
Reactions carried out under argon atmosphere using a twofold
[b]
excess of 1. Ϫ 2 was slowly added over a period of 45 min.; the
[c]
time is given from the beginning of the addition. Ϫ Determined
by gas chromatography.
Copper(I) compounds are known to be very active cata-
lysts in the cyclopropanation of alkenes with diazo com-
pounds.[9] However, addition of the ligand in an almost
equimolecular ratio results in the appearance of a period of
induction. After this period, the color of the solution
changes to yellow and the reaction ultimately yields a final
result comparable to that obtained with CuOTf. This period
2348
Eur. J. Inorg. Chem. 1999, 2347Ϫ2354

Homogeneous and Supported Copper Complexes of Cyclic and Open-Chain Polynitrogenated Ligands
FULL PAPER
L5) should also be reflected in a decreased tendency for ag-
gregation.
Although the design of efficient chiral auxiliaries was not
the main aim of this work, the enantiomeric excesses
achieved in the reactions using the chiral ligands L4 and L5
were examined. In neither case was any asymmetric induc-
tion observed.
Table 3. Results obtained in the reactions promoted by CuOTfϪli-
near polyamine (L2ϪL5) ϭ 1:1.1 complexes, using 0.1% catalyst
with respect to styrene in CH2Cl2 at 25 °C
Figure 2. Binuclear structural patterns found for tetraaza[n]para-
cyclophanes
Entry Ligand t (h) % yield 4:3
% ee (3)[a] % ee (4)[a]
In order to improve the yields, we added a fresh aliquot
of ethyl diazoacetate (2) (0.5 mol for each mol of remaining
styrene) (Table 2). Surprisingly, the period of induction was
no longer observed after the second addition. This kinetic
behavior is characteristic of product-accelerated reactions,
in which a side product, formed via a non-catalytic path-
way, is the species involved. In this context, we have to con-
sider that decomposition of ethyl diazoacetate gives rise to
the formation of coordinating species, namely ethyl fumar-
ate and maleate. These by-products are produced in large
excess with respect to L1 and, once formed, should favor
disruption of aggregates through coordination to the metal
cation. Displacement of these kinetically labile ligands from
the coordination sphere of the metal by the carbene-for-
ming species thus becomes easier and so the induction per-
iod is much decreased or is not seen at all. To check this
hypothesis, we stirred a solution of CuOTf and L1 for 3 h
in CH2Cl2 with 50 equiv. of ethyl acetate (the same ratio of
ethyl diazoacetate used in the reactions) prior to addition
of the ethyl diazoacetate. Following this treatment, the per-
iod of induction was not seen (entry 5, Table 2).
1
L2
L3
L4
L5
24
Ϫ
23
26
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
0
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
0
2
27
2.1
2.2
Ϫ
Ϫ
2.2
Ϫ
2.5
2.5
Ϫ
2.3
2.3
3
96
4
96[b]
98
5
6
117 29
7
4
8
Ϫ
8
27
33
Ϫ
11
36
9
23
5
0
0
10
11
12
Ϫ
0
0
Ϫ
0
0
8
23
[a]
Determined by gas chromatography using a chiral stationary
[b]
phase. Ϫ
der reflux.
After this time, the reaction mixture was heated un-
Amides as Ligands
A further structural change was considered by using the
diamides L6ϪL10. The results obtained with L6 and L7,
when compared to those with L4 and L5 (entries 2 and 6,
Table 4 vs. entries 8 and 11, Table 3), show that the presence
of carbonyl groups leads to a decrease in the induction per-
iod and, in the case of the (S)-phenylalanine derivative (L7),
to a decrease in the trans/cis ratio and a measurable enanti-
oselectivity. These data suggest that, for L6 and L7, the
changes in geometry and donor characteristics associated
with the presence of the amide functionalities lead to an
increased rate of copperϪcarbene intermediate formation.
Table 2. Effect of a second addition of ethyl diazoacetate (2) and
the addition of ethyl acetate
Entry
Cu/L1
1:1.1
Cu/styrene (%) t (h)
% yield
4:3
1
2
3
4
5
6
7
0.1
24.0[a]
30
35
54
57
4
20
28
2.3
2.3
2.3
2.3
Ϫ
2.3
2.4
24.5
26.0
31.0
0.75
2.0
Table 4. Results obtained in the reactions promoted by CuOTfϪli-
near amideϪamine (L6ϪL10) ϭ 1:1.1 complexes, using 0.1% cata-
lyst with respect to styrene in CH2Cl2 at 25 °C
1:1.1
0.1[b]
16.0
Entry Ligand t (h) % yield 4:3
% ee (3)[a] % ee (4)[a]
[a]
After this time, an additional amount of 2 was added over a
1
L6
L7
L8
3.0
4.0
6.5
Ϫ
16
31
Ϫ
Ϫ
0
Ϫ
0
[b]
period of 45 min. Ϫ The CuϪL1 complex was stirred with ethyl
acetate prior to use.
2
2.1
2.2
2.1
Ϫ
1.8
1.8
1.8
Ϫ
1.8
1.8
1.8
1.8
Ϫ
3
0
0
4
24.0 29
0
0
In order to study the influence of ligand topology on the
catalyst activity, we tested several open-chain tri- and tetra-
amines (L2ϪL5). The results obtained (Table 3) show that
the effect of these ligands is to increase the period of induc-
tion compared to that observed using L1 under the same
conditions. Furthermore, this period is longer with triam-
ines, and in particular with L3 (entries 5Ϫ6, Table 3). Again, 13
the formation of aggregates has to be considered. Clearly,
this phenomenon has to be more important for those li-
5
2.5
3.5
6.5
Ϫ
11
29
Ϫ
12
12
12
Ϫ
11
11
0
Ϫ
8
6
7
8
8
8
24.0 33
9
1.0
2.0
3.0
Ϫ
25
29
Ϫ
7
10
11
12
7
L9
0.75 13
1.25 30
0
0
Ϫ
0
0
14
15
L10
24.0
Ϫ
Ϫ
0
48.0 24
2.2
gands (L2 and L3) providing a smaller number of nitrogen
donor atoms. The presence of bulky C-substituents (L4 and rably obtained.
[a] In cases with asymmetric induction, (1R)-cycloadducts are prefe-
Eur. J. Inorg. Chem. 1999, 2347Ϫ2354
2349

S. V. Luis, J. A. Mayoral et al.
FULL PAPER
In comparison to L6 and L7, the period of induction is Structural XAS/EXAFS Studies
clearly reduced with L8, which has a longer bridge, but the
In order to gain further insight into the structural charac-
trans/cis ratio and enantioselectivity remain unchanged.
Two diastereomeric ligands (L9 and L10) were also tested
and showed very different behaviors. Whereas the reaction
is very fast with L9, use of L10 leads to an extremely long
period of induction. Clearly, the cis,trans,cis-arrangement of
the substituents in complexed L9 disfavors aggregation,
while the opposite is true for the all-trans diastereomer
formed from L10, for which the access for the formation of
the carbeneϪcopper species should again be restricted. The
introduction of chirality in the central bridge is always det-
rimental for the enantioselectivity.
teristics of these complexes, we studied the structures of the
species derived from three representative ligands, L1, L5 and
L7, and CuCl2 by X-ray absorption spectroscopy (XAS).
The use of CuCl2 instead of Cu(OTf)2 leads to complexes
with lower catalytic activity (Table 6) and this has been at-
tributed to the presence of the chloride counterion.[15]
However, we decided to use CuCl2 in order to be able to
distinguish between the anion (Cl) and the coordinating
groups in the ligand (N, O).
Table 6. Results obtained in the reactions promoted by CuCl2ϪL ϭ
1:1.1 complexes in CH2Cl2 using 0.1% catalyst with respect to sty-
rene
The Effect of the Oxidation State of Copper: The
Use of CuIL vs. CuIIL Complexes
Entry Ligand t (h) % yield 4:3
% ee (3)[a] % ee (4)[a]
1
2
3
4
5
6
7
8
L1
L3
L5
L7
24[a]
Ϫ
Ϫ
2.5
Ϫ
2.4
Ϫ
2.4
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
0
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
0
It is known that CuI is the active species in these reac-
tions and that CuII is reduced to CuI in the presence of
diazoacetate.[9] Thus, CuII complexes can also be used as
catalysts. In view of this, we tested the cyclic ligand (L1),
two open-chain polyamines (L3 and L5), and an amide-con-
taining ligand (L7) in the same reaction using Cu(OTf)2.
The results obtained (entry 1, Table 5) show that the induc-
tion period with L1 decreases and that it is very short even
on halving the amount of catalyst (entries 3Ϫ4, Table 5).
The induction period is no longer seen with L5 and L7, but
persists with L3. Therefore, other than in the case of L3, the
metal center in the CuII complexes is much more accessible
to the additional exogenous ligands. Once CuII is reduced
to CuI, i.e. in the presence of diazoacetate, the strongly ac-
tive yellow complex is immediately formed.
24[b] 24
24[a]
7[b]
Ϫ
29
Ϫ
28
Ϫ
31
24[a]
7[b]
24[a]
7[b]
Ϫ
0
Ϫ
0
2.2
[a] Reaction at room temperature. Ϫ [b] Reaction under reflux condi-
tions after the initial period at room temperature.
The experimental EXAFS signal was extracted from the
raw spectra according to standard methods.[16] The thresh-
old energy E0, defining the zero wavevector, was taken at
the inflection point of the absorption edge. Figure 3 shows
the Fourier transform (FT) of the Cu-K edge k χ(k) spectra
of the three complexes. The FT was calculated using a
Gaussian window in the range 2.2Ϫ12.5 AϪ1. EXAFS spec-
˚
Table 5. Results obtained in the reactions promoted by
Cu(OTf)2ϪL ϭ 1:1.1 complexes in CH2Cl2 at 25 °C
tra of the three samples are very similar and show little
resolution of structure, indicating structural disorder. The
moduli of the FTs of the L1 and L5 complexes show two
main peaks, while that of the L7 complex shows an ad-
Entry Ligand Cu/1 (%) t (h) % yield 4:3 % ee (3)[a] % ee (4)[a]
1
L1
L1
0.1
1.0
21
2.0
2.1
2.4
2.3
2.1
Ϫ
2.1
2.1
2.1
1.8 10
1.8 10
1.7 10
1.7 10
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
6
˚
ditional peak in the region of 3Ϫ4 A.
2
1.75 33
3
0.05
1.3
2.5
6
15
4
5
24.0 25
6
L3
L5
L7
L11
0.1
0.1
0.1
0.1
26.0
Ϫ
7
29.5 23
8
1.0
2.5
16
26
9
10
11
12
13
0.75 18
2.5
0.75 19
1.5 33
29
6
6
6
[a] In cases with asymmetric induction, (1R)-cycloadducts are prefe-
rably obtained.
In the reactions promoted by L7ϪCuI and CuII com-
plexes, the trans/cis ratios and enantioselectivities are the
same, and so it would seem reasonable to assume that the
reactive intermediate is the same L7ϪCuI carbene complex.
It can be concluded that the only difference between CuI
and CuII is the decrease in the induction period, whereas
the selectivities and the relative catalytic activities of the
different complexes are the same.
Figure 3. Fourier transforms of the EXAFS spectra: (a) L1ϪCuCl2
complex; (b) L5ϪCuCl2 complex; (c) L7ϪCuCl2 complex; (d)
L12ϪCuCl2 complex
2350
Eur. J. Inorg. Chem. 1999, 2347Ϫ2354

Homogeneous and Supported Copper Complexes of Cyclic and Open-Chain Polynitrogenated Ligands
FULL PAPER
Table 7. Results obtained from analyses of the EXAFS spectra of L1, L5, L7 and L12ϪCuCl2 complexes
CuϪN (O)
CuϪC
CuϪCl
2
3
2
2
3
2
2
3
2
˚
˚
˚
˚
˚
˚
Ligand n°
R (A)
∆σ 10 A
n°
R (A)
∆σ 10 A
n°
R (A)
∆σ 10 A
L1
5.4±0.3
2.03±0.01
2.01±0.01
2.02±0.02
2.09±0.02
1.98±0.02
1.97±0.02
4.97±0.003 3.4±0.7
6.75±0.003 4.2±0.8
8.0±0.003
7.0±0.003
6.3±0.003
3.2±0.004
2.93±0.03
2.93±0.03
2.87±0.03
2.87±0.03
5.21±0.005
5.17±0.006
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
L5[a]
L7[b]
L7
5.2±0.3
6.2±0.4
3.1±0.4
2.6±0.4
2.0±0.6
3.3±0.4
3.3±0.4
11.7±0.005 2.1±0.5
3.85±0.04
3.85±0.04
12.9±0.006
12.9±0.006
11.7±0.005 2.1±0.5
L12
2.0±0.4
2.21±0.01
2.7±0.003
[a]
[b]
Considering CuϪN contributions only. Ϫ Considering CuϪN and CuϪO contributions.
The first-shell contribution was extracted by Fourier fil-
˚
After background subtraction from the raw spectra, the
tering of the FT spectra in the range 0.8Ϫ2.2 A, and the XANES spectra were normalized taking unity to indicate
corresponding structural parameters were obtained by le- atomic absorption. The XANES spectra of the complexes
ast-squares fitting of the filtered k-weighted spectrum to the of L1 and L5 with CuCl2 were found to be almost superim-
EXAFS formula:[16]
posable, which was not the case with the L7ϪCuCl2 com-
plex. This fact is in good agreement with the differences
observed by EXAFS.
(k) ϭ Σ (ρ/k) (Ni/Ri2) Ai(k) sin[2kRiϩ␾i(k)]
The formation of copperϪL7 oligomeric species might
suggest that each amineϪamide moiety of the ligand par-
ticipates in the complexation of different copper atoms.
This suggestion led us to examine the behavior of the sim-
plest ligand L11. Unfortunately, the corresponding XAS
spectrum could not be obtained. However, the close simi-
larity of their catalytic behavior (Table 5) lends support to
this hypothesis.
The amplitudes Ai(k) and phase shifts ␾i(k) used in the
best fit were generated from the FEFF3.11 code.[16e] The
best-fit parameters obtained for the different complexes are
reported in Table 7. The data for the L1 and L5 complexes
were found to be very similar and showed the copper to be
surrounded by five nitrogen or/and oxygen atoms at an av-
˚
erage distance of 2.02 A. For the second shell, the best fit
˚
was obtained solely with carbon atoms at 2.93 A. We did
The results obtained from the EXAFS measurements
show the absence of chlorine atoms in close proximity to
the copper ions, which would seem to be at variance with
the influence of this counterion on the catalytic activity.
Thus, we investigated whether the same structural behavior
is observed with other ligands, such as bis(oxazolines),
where the same influence of the counterion has been de-
scribed.[15] To this end, we studied the EXAFS spectrum of
the L12ϪCuCl2 complex. The EXAFS and the FT of the
Cu-K edge k χ(k) spectrum is shown in Figure 3 and the
best-fit parameters are given in Table 7. As can be seen, the
spectra show clear differences and chlorine is observed at
not observe CuϪCl contributions in these complexes. On
the other hand, the high DW values obtained for both the
first and second shells, together with the large error margins
associated with the fit of structural parameters of the se-
cond shell indicate a high degree of structural disorder in
these compounds. The absence of a CuϪCl contribution
in these complexes may be accounted for in terms of the
significant structural disorder, the presence of different
CuϪCl and CuϪC distances in the second shell, or a very
long CuϪCl distance with a large vibrational DW factor.
The presence of a large number of donor atoms in close
proximity to the metal supports the formation of aggregate
species with more than one ligand being involved in the
complexation of the copper.
˚
a distance of 2.21 A from the copper, justifying its strong
influence on the catalytic activity.
Comparison of this result with those obtained for the po-
lynitrogenated compounds shows that the use of the latest
ligands leads to oligomeric complexes with considerable dis-
order in the environment of the copper, and with chlorine
at greater distances. In spite of these differences, the coun-
terion has the same influence on the catalytic activity. It
may be speculated that, in accordance with the period of
induction, the oligomeric species are fragmented to yield
the reactive copperϪcarbene intermediate, and that the re-
activity of this intermediate depends on the nature of the
counterion. Unfortunately, it has not been feasible to design
an EXAFS experiment to check this hypothesis.
The best-fit results for the L7 complex reveal a slightly
different structure. CuϪO and CuϪN cannot be clearly dis-
tinguished by EXAFS, hence we considered both possibil-
ities in the theoretical fit of the experimental spectra. If we
consider only CuϪN contributions, we obtain 6 nitrogen
˚
atoms at a distance of 2.02 A. A similar fit is obtained con-
sidering CuϪN and CuϪO contributions at two different
distances. An important difference between L7 and L1 and
L5 is the observation of two distant chlorine atoms in the
complex of the former. The presence of these chlorine
atoms suggests less structural disorder and a higher accessi-
bility of the copper in this complex. These structural
changes, which may be due to the planarity imposed by the
amide system, the presence of oxygen atoms, or both these
factors, must also be responsible for the changes in the
trans/cis ratio and the enantioselectivity.
Supported Complexes as Catalysts
As a final avenue of investigation, we tested the behavior
of some of these ligands supported in various ways. The
Eur. J. Inorg. Chem. 1999, 2347Ϫ2354
2351

S. V. Luis, J. A. Mayoral et al.
FULL PAPER
formation of aggregates should be disfavored when using L3, the catalytic activity of the supported catalyst is similar
such immobilized ligands. This, along with the changes in to that seen for the homogeneous complex, while for L1 it
the micro-environments of the ions, could lead to signifi- is even superior. However, a simple comparison cannot be
cant changes in reactivity. Open-chain systems can be made because in the supported catalyst the counterion is
grafted onto polystyreneϪdivinylbenzene through the pri- changed from chloride to the clay surface, and the nature
mary amine functionalities and ligands L13 and L14 were of the counterion has a decisive influence on the catalytic
obtained in this way.[5] In accordance with our expectations, activity. The trans/cis ratio decreases, hence the proportion
the reactions took place without an induction period. The of the less stable product increases. This has previously been
results obtained (Table 8) show that the stereochemical co- observed with clay catalysts and has been explained in
urse of the reaction is influenced by the presence of the terms of the isolation of the catalytic sites and the dimen-
support. A decrease in the ee, relative to that achieved with sionality of the clay, which should also be related to the
the same ligand in solution, was observed. Similar changes higher catalytic activity.[17]
in the stereochemical course of reactions upon immobiliz-
Table 10. Results obtained in the reactions promoted by CuCl2ϪL1
Table 8. Results obtained in the reactions promoted by CuII using
polymeric ligands (L13ϪL14) in CH2Cl2 at 25 °C
and L3 (1:1.1) complexes exchanged in laponite, in CH2Cl2 using a
Cu/1 ratio of 0.001
Entry Ligand Cu/1 (%) t (h) % yield 4:3 % ee (3) % ee (4)
Entry
Ligand
t (h)
% yield
4:3
[a]
24[a]
7[b]
Ϫ
Ϫ
1
2
3
4
5
L13
L14
0.1
0.1
1.5
2.0
12
19
2.0
2.0
1.8
1.6
1.6
Ϫ
Ϫ
0
Ϫ
4
Ϫ
Ϫ
0
Ϫ
3
1
2
3
4
5
6
L1
L3
20
30
Ϫ
27
29
1.1
1.1
Ϫ
1.1
1.1
24[b]
24[a]
7[b]
24.0 28
[a]
1.0
23.0 31
17
24[b]
L12 contains 0.30 mmol ligand gϪ1 and L13 0.46 mmol ligand
[a]
gϪ1
.
[a] Reaction at room temperature. Ϫ [b] Reaction under reflux condi-
tions after initial period at room temperature.
ation of a chiral auxiliary have repeatedly been observed.[3]
Given that the complexes used as catalysts are cationic,
they can also be supported using of a different strategy, i.e.
they can be electrostatically entrapped using anionic solids.
To test this approach, we used a synthetic laponite with a
high swelling ability. Cation exchange was carried out in
methanol using the CuCl2 complexes of a cyclic (L1) and
an open-chain (L3) ligand. Table 9 shows the results of
analyses of the clays. In both cases, 0.1 mmol Cu gϪ1 are
introduced, the Li:Cu stoichiometry is 1:1, and the presence
of the complex produces a decrease in the surface area.
With L1, the basal spacing increases, indicating that the
complex is accommodated into the interlamellar space, at
least in part. With L3, no change in basal spacing is ob-
served, which may either be due to the smaller size of the
complex or to a disposition of this complex on the external
surface of the clay. In view of the result obtained with L1,
the first explanation seems more plausible.
Conclusions
The results presented here open the way for the design of novel
polynitrogenated ligands for use in copper-catalyzed processes and
for understanding the different factors affecting the activity of such
species. The copper(I) and copper(II) complexes of some cyclic and
non-cyclic polyamines, as well as of some diamides, have been as-
sayed in the benchmark cyclopropanation reaction of styrene with
ethyl diazoacetate. First and foremost, all the complexes examined
showed an appreciable catalytic activity, even at very low catalyst
concentrations. All data suggest that formation of the carbeneϪc-
opper complex is the rate-limiting step. For polyamines, aggre-
gation hinders the formation of the active carbeneϪcopper com-
plexes and as a result induction periods are observed in non-coordi-
nating solvents (CH2Cl2). The addition or the in situ generation of
kinetically labile coordinating species such as EtOAc or ethyl fum-
arate (formed by decomposition of the diazoacetate) disrupts ag-
gregation and allows the facile formation of the active complexes.
Aggregation is more significant with polyamines having a smaller
number of nitrogen donor atoms and with less bulky ligands, and
is much reduced with amides.
Table 9. Analysis of the laponites exchanged with the complexes of
CuCl2 with L1 and L3
Entry
Ligand Cu (mmol gϪ1) N/Cu
d001 (•) S (m2 gϪ1
)
The induction periods are more important for CuI than for CuII
complexes, according to the different coordination geometries. This
seems to be associated with the formation of more accessible CuII
sites that readily interact with the carbene-forming species and are
reduced to CuI complexes, which are the active catalytic species.
No other differences (e.g. in terms of selectivity) are associated with
the use of CuI or CuII and, accordingly, the use of the more stable
CuII compounds is clearly favored.
1
2
3
Ϫ
L1
L3
Ϫ
0.1
0.1
Ϫ
3.9
3.8
16
19
16
289
245
239
Table 10 shows the results obtained with these clay-sup-
ported catalysts. As can be seen in Table 6, the use of chlo-
ride in place of triflate significantly reduces the catalytic
activity in solution and it is necessary to carry out the reac-
tion under reflux. This can only be related to the effect of
Polyamine or amide ligands of this type can easily be supported
by anchoring on functionalized polymers or by cation exchange in
clays. The resulting supported CuII complexes show higher activi-
the more coordinating chloride counterion. In the case of ties than their homogeneous analogues. This can be rationalized in
2352 Eur. J. Inorg. Chem. 1999, 2347Ϫ2354

Homogeneous and Supported Copper Complexes of Cyclic and Open-Chain Polynitrogenated Ligands
FULL PAPER
terms of the pseudodilution effects operative in the solid materials,
which will disfavor the aggregation processes.
Ϫ C32H34N4O2•H2O: calcd. C 73.3, H 6.9, N 10.7; found C 73.4,
H 7.0, N 10.8.
Finally, small enantioselectivities have been observed on using
some of the chiral ligands. However, more work is needed in this
regard if novel chiral ligands able to improve the observed enanti-
oselectivities are to be developed
20
L11: M.p. 67Ϫ71°C. Ϫ [α]D ϭ Ϫ69.1 (c ϭ 1, CHCl3). Ϫ 1H
NMR: δ ϭ 1.56 (br. s, 2 H), 2.78 (dd, J1 ϭ 9.8, J2 ϭ 13.8 Hz, 1
H), 3.27 (dd, J1 ϭ 4.5, J2 ϭ 13.8 Hz, 1 H), 3.62 (dd, J1 ϭ 4.6, J2 ϭ
9.4 Hz, 1 H), 4.43 (d, J ϭ 6.5 Hz, 2 H), 7.25 (m, 10 H), 8.19 (t,
J ϭ 6.3 Hz, 2 H). Ϫ 13C NMR: δ ϭ 41.1, 43.2, 56.5, 126.8, 127.4,
127.7, 128.3, 129.4, 138.0, 173.4. Ϫ C16H18N2O: calcd. C 75.6, H
7.1, N 11.0; found C 75.5, H 7.1, N 11.3.
Experimental Section
Preparation of the Ligands: L2 and L3 are commercial compounds.
Preparation of the Cationic Complexes Supported on Clays: A solu-
tion of the complex was first prepared by stirring CuCl2•2H2O
(1 mmol) and the appropriate ligand (L1 or L3, 1 mmol) in meth-
anol (50 mL) for 45 min. To this green solution, laponite (10 g) was
added and the mixture was shaken for 22 h. After this time, the
blue clay was separated by filtration and washed with methanol.
The residue obtained by evaporation of the solvent from the color-
less solution was found to contain NaCl.
[18]
[5]
L1,[12] L4ϪL7
and L13ϪL14 were prepared as described pre-
viously. L8, L9 and L10 were obtained by reaction of the N-hydroxy-
succinimide esters of the N-Cbz protected amino acids with the
corresponding diamines, followed by deprotection of the N-Cbz
groups.
General Procedure for the Preparation of N-Hydroxysuccinimide Es-
ters of Amino Acids: The appropriate N-Cbz-protected amino acid
(3.34 mmol) and N-hydroxysuccinimide (3.34 mmol) were sus-
pended in anhydrous DME at 0°C. Once a clear solution had been
obtained, DCC (3.67 mmol) was added and the resulting solution
was stirred at 0Ϫ5°C for 20 h. The dicyclohexylurea formed was
filtered off and the filtrate was concentrated to dryness. The crude
product was recrystallized from propan-2-ol to furnish the pure
product. N-Hydroxysuccinimide ester of N-Cbz--phenylalanine:
yield 87%; m.p. 137Ϫ140°C. Ϫ 1H NMR: δ ϭ 2.80 (br. s, 4 H),
3.13Ϫ3.41 (m, 1 H), 5.09 (s, 4 H), 7.29Ϫ7.4 (m, 10 H). Ϫ 13C
NMR: δ ϭ 25.4, 37.8, 52.9, 67.2, 127.3, 128.0, 128.2, 128.5, 128.7,
The copper content was analyzed by plasma emission spectroscopy
using a PerkinϪElmer Plasma 40 emission spectrometer. Carbon
and nitrogen analyses were carried out using a PerkinϪElmer 2400
elemental analyzer. Step-scanned X-ray diffraction patterns of ori-
ented samples were collected at room temperature from 3° in 2&
thetas; up to 60° using a D-max Rigaku system with a rotating
anode. The diffractometer was operated at 40 kV and 80 mA and
Cu-Kα radiation was selected using a graphite monochromator.
Surface areas were calculated from BET nitrogen isotherms deter-
mined at 77 K.
25
129.5, 134.3, 167.5, 168.7. Ϫ [α]D ϭ Ϫ11.1 (c ϭ 2, dioxane). Ϫ
Cyclopropanation Reactions with Homogeneous Catalysts: Under
argon, a mixture of the requisite amount of the ligand (Tables 1Ϫ5)
and CuOTf•1/2C6H6 or Cu(OTf)2 (0.1 mmol) in dry CH2Cl2
(60 mL) was stirred at room temperature for 30 min. Then, a solu-
tion of styrene (100 mmol) and n-decane (2 g) in dry CH2Cl2
(50 mL) was added and the resulting solution was stirred for
10 min. After this time, ethyl diazoacetate (50 mmol) was slowly
added by means of a syringe pump over a period of 45 min. The
reactions were monitored by gas chromatography using n-decane
as an internal standard; FID from Hewlett-Packard 5890 II; cross-
linked methyl silicone column 25 m ϫ 0.2 mm ϫ 0.33 µm; helium
as carrier gas, 20 psi; injector temperature 230°C; detector tem-
perature 250°C; over temperature program: 70°C (3 min)Ϫ15°C/
minϪ200°C (5 min); retention times: ethyl diazoacetate (2)
4.28 min, styrene (1) 5.03 min, n-decane 6.93 min, diethyl fumarate
8.73 min, diethyl maleate 9.04 min, cis-cyclopropanes (3)
11.84 min, trans-cyclopropanes (4) 12.35 min.
C21H20N2O6: calcd. C 63.6, H 5.1, N 7.1; found C 63.5, H 5.5,
N 7.4.
General Procedure for the Preparation of N,NЈ-Bis(N-Cbz-
aminoacyl)diamines: The N-hydroxysuccinimide ester of the appro-
priate amino acid (7.44 mmol) was dissolved in anhydrous DME
cooled in an ice bath, and the diamine (3.72 mmol) was added
dropwise. The reaction mixture was stirred at room temp. for 18 h.
The white solid formed was filtered off and washed with water and
cold methanol.
Deprotection of N-Cbz Groups: The N,NЈ-bis(N-Cbz-aminoacyl)-
1,3-diaminopropanes were added to HBr/AcOH (33%) and the re-
spective mixtures were stirred at room temp. until CO2 evolution
ceased. At this point, diethyl ether was added to the clear solution,
which led to the deposition of a white precipitate. This was filtered
off and dried to afford the pure product.
20
L8: M.p. 140Ϫ143°C. Ϫ [α]D ϭ Ϫ89.42 (c ϭ 1, CHCl3). Ϫ 1H
Asymmetric induction in the reactions using chiral ligands was also
assessed by gas chromatography; cyclodex-B column 30
m ϫ 0.25 mm ϫ 0.25 µm; oven temperature program: 125°C iso-
thermal; retention times: (1S,2R)-cyclopropane 28.9 min, (1R,2S)-
cyclopropane 29.8 min, (1R,2R)-cyclopropane 34.3 min, (1S,2S)-
cyclopropane 34.9 min. The peaks were assigned to the correspond-
ing enantiomers by comparison with previous results.
NMR: δ ϭ 1.5 (m, 4 H), 2.7 (dd, J1 ϭ 9.2, J2 ϭ 14 Hz, 2 H), 3.25
(m, 6 H), 3.6 (dd, J1 ϭ 3.8, J2 ϭ 9.2 Hz, 2 H), 7.3 (m, 10 H). Ϫ
13C NMR: δ ϭ 27.0, 38.7, 41.1, 56.5, 126.8, 128.7, 129.4, 138.0,
174.3. Ϫ C22H30N4O2: calcd. C 69.1, H 7.9, N 14.6; found C 69.3,
H 8.3, N 15.0.
20
L9: M.p. 137Ϫ141°C. Ϫ [α]D ϭ Ϫ17.16 (c ϭ 1, CHCl3). Ϫ 1H
NMR: δ ϭ 1.31 (d, J ϭ 6.5 Hz, 4 H), 2.6 (dd, J1 ϭ 9.5, J2 ϭ
13.5 Hz, 2 H), 3.32 (dd, J1 ϭ 3.8, J2 ϭ 13.5 Hz, 2 H), 3.55 (m, 2
H), 5.28 (m, 2 H), 7.2 (m, 20 H), 8.2 (m, 2 H). Ϫ 13C NMR: δ ϭ
40.7, 56.5, 59.0, 126.8, 127.6, 127.7, 128.5, 128.7, 129.3, 137.8,
138.6, 174.7. Ϫ C32H34N4O2•1/2 H2O: calcd. C 74.5, H 6.8, N 10.9;
found C 74.4, H 6.9, N 10.5.
In general, the onset of the reactions is signified by a color change
from green to yellow. In one case (entry 4, Table 3), the reaction
did not take place at room temperature and had to be carried out
under reflux. In another case (entry 2, Table 2), after the consump-
tion of the ethyl diazoacetate an additional amount of this reagent
was added in the same manner. In another reaction (entry 5, Table
2), the L1ϪCuOTf complex was stirred with ethyl acetate (5 mmol)
for 3 h prior to addition of the reagents.
20
1
L10: M.p. 213Ϫ217°C. Ϫ [α]D ϭ Ϫ80.02 (c ϭ 1, CHCl3). Ϫ H
NMR: δ ϭ 1.3 (m, 4 H), 2.6 (dd, J1 ϭ 10, J2 ϭ 14 Hz, 2 H), 3.18
(dd, J1 ϭ 4.5, J2 ϭ 14 Hz, 2 H), 3.58 (m, 2 H), 5.3 (m, 2 H), 7.2
(m, 20 H), 8.26 (m, 2 H). Ϫ 13C NMR: δ ϭ 41.5, 56.8, 56.9, 58.7,
Cyclopropanation Reactions with Polymeric Catalysts: The appro-
126.9, 127.6, 127.7, 128.5, 128.7, 128.8, 129.3, 138.0, 138.4, 174.7. priate amount of polymer (0.37 g of L12 or 0.24 g of L13) was
Eur. J. Inorg. Chem. 1999, 2347Ϫ2354 2353

S. V. Luis, J. A. Mayoral et al.
FULL PAPER
in Organic Synthesis (Ed.: K. Smith ), Ellis Horwood, Chiches-
ter, 1992.
stirred under argon in a solution of Cu(OTf)2 (1 mmol) in CH2Cl2
for 24 h. After this time, the solution was decanted off and the
remaining solid was washed with CH2Cl2. To a suspension of the
catalyst in dry CH2Cl2 (60 mL), the internal standard and the re-
agents were added and the reaction was monitored as described
above.
[4] [4a]
[4b]
H.-U. Blaser, Tetrahedron: Asymmetry 1991, 3, 843. Ϫ
H.-U. Blaser, B. Pugin in Chiral Reactions in Heterogeneous Ca-
talysis (Eds.: G. Jannes, V. Dubois), Plenum Press, New York,
[4c]
1995, p. 33. Ϫ
See ref.[2], p. 346.
[5]
F. M. Adrian, B. Altava, M. I. Burguete, S. V. Luis, R. V. Sal-
´
vador, E. Garcıa-Espan˜a, Tetrahedron 1998, 54, 3581.
[6] [6a]
G. Pozzi, M. Cavazzini, S. Quici, S. Fontana, Tetrahedron
Cyclopropanation Reactions with Clay-Supported Catalysts: To a
suspension of the clay (1 g) in dry CH2Cl2 (60 mL), the reagents
and the internal standard were added and the reaction was moni-
tored as described above.
Lett. 1997, 38, 7605. Ϫ
pf1[b] D. E. de Vos, T. Bein, Chem. Commun. 1996, 917.
[7a] E. Galardon, P. Le Maux, G. Simmoneaux, Chem. Commun.
[7]
[8]
[7b]
1997, 927. Ϫ
J. Robbins Wolf, C. G. Hamaker, J.-P. Djukic,
T. Kodadek, L. K. Woo, J. Am. Chem. Soc. 1995, 117, 9194.
For a recent example in enantioselective Michael reactions, see
K. Yasuda, M. Shindo, K. Koga, Tetrahedron Lett. 1997, 38,
3531.
EXAFS and XANES Experiments: X-ray absorption experiments at
the Cu-K edge were carried out at the synchroton radiation source,
Daresbury, U.K. A double-crystal Si(111) monochromator was
used. The storage ring operated at 2 GeV with an average beam
current of 150 mA. The experiments were performed at room tem-
perature in the transmission mode, and the signal was collected by
means of ion chambers. For the XAS measurements, solid samples
of the L1, L5, L7 and L12ϪCuCl2 complexes were pressed into thin
pellets. The thickness of the samples was optimized in order to
obtain the best signal-to-noise ratio.
[9] [9a]
R. G. Salomon, J. K. Kochi, J. Am. Chem. Soc. 1973, 95,
[9b]
3300. Ϫ
Ϫ
M. Hagen, U. Lüning, Chem. Ber. 1997, 130, 231.
[9c]
M. P. Doyle, M. N. Protopopova, Tetrahedron 1998, 54,
7919.
[10] [10a]
´
A. Domenech, J. V. Folgado, E. Garcıa-Espan˜a, S. V. Luis,
´
J. M. Linares, J. F. Miravet, J. A. Ramırez, J. Chem. Soc., Dalton
[10b]
´
A. Domenech, E. Garcıa-Espan˜a, V.
Trans. 1995, 541. Ϫ
Marcelino, B. Altava, S. V. Luis, J. F. Miravet, A. Bianchi, L.
Ferrini, Inorg. Chim. Acta 1996, 252, 123.
[11] [11a]
´
E. Garcıa-Espan˜a, J. Latorre, S. V. Luis, J. F. Miravet, P. E.
´
Pozuelo, J. A. Ramırez, C. Soriano, Inorg. Chem. 1996, 35, 4591.
Samples were prepared using the ligand and CuCl2 in an equimolar
ratio; the amounts of the two components were adjusted in order
to obtain 100 mg of complex. To a solution of the ligand in CH2Cl2
(5 mL), the requisite amount of CuCl2•2H2O was added. The re-
sulting suspension was stirred at room temperature for 5 h, which
led to the gradual disappearance of the solid CuCl2. After this time,
the solution was filtered through a plug of cotton wool and the
solvent was evaporated from the filtrate under reduced pressure.
The solid complexes were dried in vacuo for 24 h at 50°C. All com-
plexes showed the expected IR bands and gave appropriate elemen-
tal analyses.
Ϫ [11b] M. I. Burguete, B. Escuder, J. C. Frıas, E. Garcıa-Espan˜a,
S. V. Luis, J. F. Miravet, J. Org. Chem. 1998, 63, 1810.
´
´
[12]
´
A. Bencini, M. I. Burguete, E. Garcıa-Espan˜a, S. V. Luis, J. F.
Miravet, C. Soriano, J. Org. Chem. 1993, 58, 4749.
[13] [13a]
´
´
A. Andres, C. Bazzicaluppi, A. Bianchi, E. Garcıa-Espan˜a,
´
S. V. Luis, J. F. Miravet, J. A. Ramırez, J. Chem. Soc., Dalton
Trans. 1994, 2995. Ϫ [13b] E. Garcıa-Espan˜a, J. Latorre, V. Mar-
´
´
celino, J. A. Ramırez, S. V. Luis, J. F. Miravet, M. Querol, Inorg.
Chim. Acta 1997, 265, 179.
[14] [14a]
K. A. Deal, A. C. Hengge, J. N. Burstyn, J. Am. Chem.
[14b]
Soc. 1996, 118, 1713. Ϫ
Chem. 1996, 35, 7474.
E. L. Hegg, J. N. Burstyn, Inorg.
[15]
D. A. Evans, K. A. Woerpel, M. M. Hinman, M. M. Faul, J.
Am. Chem. Soc. 1991, 113, 726.
[16] [16a] D. C. Konigsberger, R. Prins, X-ray Absorption: Principles,
Applications, Techniques of EXAFS, SEXAFS and XANES,
Wiley, New York, 1988. Ϫ [16b] X-ray Absorption Fine Structure
(Ed.: S. S. Hasnain), Ellis Horwood, Chichester, 1988. Ϫ [16c] B.
K. Teo, EXAFS: Basic Principles and Data Analysis, Springer,
Berlin, 1986. Ϫ [16d] Appendix to the report on the International
Workshop on Standards and Criteria in X-ray Absorption
Acknowledgments
This work was made possible by the generous financial support of
the C.I.C.Y.T. (Project MAT96Ϫ1053) and the Generalitat Valenci-
ana (GV-D-CN-09Ϫ140Ϫ96).
Spectroscopy, March 7Ϫ9, 1988, Brookhaven National Labora-
[16e]
tory, Physica B 1989, 158, 701. Ϫ
J. J. Rehr, J. Mustre de
´
Leon, S. I. Zabinsky, R. C. Albers, J. Am. Chem. Soc. 1991,
113, 5135.
[17] [17a]
[1]
E. Kimura, Tetrahedron 1992, 48, 6175.
[2]
´
J. M. Fraile, J. I. Garcıa, J. A. Mayoral, Chem. Commun.
[17b]
R. Noyori, Asymmetric Catalysis in Organic Synthesis, Wiley,
´
J. M. Fraile, J. I. Garcıa, J. A. Mayoral, T.
New York, 1994.
1996, 1319. Ϫ
[3] [3a]
Polymeric Reagents and Catalysts (Ed.: W. T. Ford ), A.C.S.
Tarnai, Tetrahedron: Asymmetry 1997, 8, 2089.
[18]
Symposium Series 308, American Chemical Society, Wash-
F. R. Wagler, Y. Fang, C. J. Burrows, J. Org. Chem. 1989, 54,
[3b]
ington D.C., 1986. Ϫ
Synthesis and Separations using Func-
1584.
tional Polymers (Eds.: D. C. Sherrington, P. Hodge ), John Wiley
and Sons, New York, 1988. Ϫ
Received April 21, 1999
[I99140]
[3c]
Solid Supports and Catalysts
2354
Eur. J. Inorg. Chem. 1999, 2347Ϫ2354