Reactions of aldehydes with diethylzinc catalysed by polymer-supported ephedrine and camphor derivatives: Comparisons of enantiomeric excesses achieved with various supports: Optimisation of support parameters to enable high enantiomeric excesses to be obtained

Article abstract of DOI:10.1039/a901455a

The reactions of benzaldehyde with diethylzinc catalysed by PS ephedrine or camphor derivatives have been investigated in some depth in order to identify the crucial factors necessary to successfully prepare PS chiral catalysts for such reactions. The most important factor is found to be a favourable interaction of the polymer matrix with the reaction solvent so that the polymer will dissolve or swell to allow the other reactants easy access to the catalytic sites. Accordingly toluene is a better reaction solvent than hexane. The ephedrine-derived catalytic groups reduce the solubility of the linear polymers in toluene and, almost certainly, the swelling properties of the crosslinked polymers. Thus, of the polymers investigated the better linear ones had

Full text of DOI:10.1039/a901455a

View Article Online / Journal Homepage / Table of Contents for this issue
Reactions of aldehydes with diethylzinc catalysed by polymer-
supported ephedrine and camphor derivatives: comparisons of
enantiomeric excesses achieved with various supports: optimisation
of support parameters to enable high enantiomeric excesses to be
obtained†
David W. L. Sung,a Philip Hodge*a and Peter W. Stratfordb
a Chemistry Department, University of Manchester, Oxford Road, Manchester, UK M13 9PL
b Chemistry Department, University of Lancaster, Bailrigg, Lancaster, UK LA1 4YA
Received (in Cambridge) 22nd February 1999, Accepted 16th April 1999
The reactions of benzaldehyde with diethylzinc catalysed by PS ephedrine or camphor derivatives have been
investigated in some depth in order to identify the crucial factors necessary to successfully prepare PS chiral catalysts
for such reactions. The most important factor is found to be a favourable interaction of the polymer matrix with the
reaction solvent so that the polymer will dissolve or swell to allow the other reactants easy access to the catalytic sites.
Accordingly toluene is a better reaction solvent than hexane. The ephedrine-derived catalytic groups reduce the
solubility of the linear polymers in toluene and, almost certainly, the swelling properties of the crosslinked polymers.
Thus, of the polymers investigated the better linear ones had <ca. 1.5 mmol per g of catalyst sites and the better
insoluble ones were 1% crosslinked gels with <ca. 1.0 mmol per g of catalyst sites. Site–site interactions and
microenvironmental effects do not appear to play a major role in these PS reaction systems. For the reaction of
benzaldehyde with diethylzinc, using the best linear PS ephedrine derivatives 10f, 10g or 11e affords 1-phenylpropanol
(1) with 83–88% enantiomeric excesses (ee)s of the (R)-enantiomer; using the best linear PS camphor derivative 21
affords the alcohol 1 with a 98% ee of the (S)-enantiomer; using the best crosslinked PS ephedrine derivative 12a
affords the alcohol 1 with a 78–81% ee and using the best crosslinked PS camphor derivative 22 affords alcohol 1 with
a 97% ee of the (S)-enantiomer. These values are close to those obtained using analogues of non-polymeric catalysts
under similar reaction conditions.
many attempts have been made to prepare PS versions of the
better chiral catalysts, but this is not a trivial exercise and often
Introduction
Merrifield’s method for ‘solid phase’ peptide synthesis was first
the percentage enantiomeric excesses (% ees) achieved have
described in detail in 1963.1 During the following two decades
been lower in the supported systems.11 In most cases this is
polymer-supported (PS) versions of many other synthetic
almost certainly because the reaction conditions have not been
organic reactions were investigated, including innumerable
optimised taking fully into account the three major factors
examples of reactions using PS substrates, PS reagents or PS
noted above.
catalysts.2–8 In this period the fundamental differences between
In recent years a range of chiral catalysts has been discovered
solution reaction systems and PS systems were identified.9 It is
which, in reactions of major synthetic importance, give reaction
clear that three major factors need to be taken into account
products with extremely high % ees.12 Thus, it is now more
when carrying out PS reactions. First, it is necessary to use a
important than before to be able to successfully prepare PS
combination of support and reaction solvent which allows the
soluble reactants to diffuse freely into and out of the support
throughout the reaction, thus allowing good site accessibility.
Second, depending on the choice of support and the loading
versions of selected chiral catalysts. To be able to do this ration-
ally we need to know more about how the conditions used in PS
reactions influence specific reactions, especially as the % ees
obtained in many asymmetric syntheses are extremely sensitive
and distribution of the PS groups, reactions between supported
to the reaction conditions. Moreover, in future, combinatorial
groups (site–site interactions) may be easier or harder than
chemistry techniques will undoubtedly be applied more exten-
reactions with, or between, species in solution. Thirdly, micro-
sively to optimise PS chiral catalyst design,13–15 and it will be
environmental effects can accelerate or slow down PS reactions.
Asymmetric organic syntheses achieved using PS chiral
catalysts are a particularly attractive type of organic reaction.
Thus, the use of the polymer support allows the soluble chiral
important to be able to screen such catalysts knowing that the
observed differences in performance are due to the different
catalytic groups themselves and not to some unidentified poly-
mer effect.
products to be separated easily from the chiral catalyst and the
The present paper is concerned with a detailed investig-
catalyst to be recovered for possible reuse. This latter feature is
ation of the use of PS ephedrine and camphor derivatives as
especially important if the catalyst moiety is expensive. Fur-
catalysts for the reaction of benzaldehyde with diethylzinc to
thermore, PS catalysts may be used in automated systems, for
give 1-phenylpropanol (1), [eqn. (1)], and with closely related
example, in flow systems which allow the chiral products to be
produced continuously.10 Because of these various advantages,
O
Zn(C2H5)2
*
C
CH C2H5
(1)
H
OH
† This project was initiated whilst PH was at the University of
Lancaster.
1
J. Chem. Soc., Perkin Trans. 1, 1999, 1463–1472
1463

View Article Online
Table 1 Reactions of benzaldehyde with diethylzinc under various conditionsa
Mol%
of added
catalyst
Reaction conditions
Chemical
yieldb (%)
Entry
Added catalyst
T/ЊC
t/h
% eec
1
2
3
4
5
6
7
8
9
Nil
Nil
Nil
Nil
—
—
—
—
—
5
—
5
2
0
20
20
50
75
0
0
20
0
48
1
24
1
<1d
2
—
—
—
—
—
2f
—
81
84
76
35e
14
43
22f
14
92
85
88
Nil
1
(R)-1-Phenylpropanol
Blank beadsg
Compound 7
Compound 7
Compound 7
48
24
24
70
20
10
2
23
a Unless indicated otherwise reactions were carried out using 9.4 mmol of benzaldehyde and 10.4 mmol of diethylzinc in 11.5 ml of toluene with
stirring under a dry nitrogen atmosphere. b By gas chromatography using an internal standard. c Estimated by polarimetry of distilled samples of
products. The (R)-enantiomer was the major product. d The same result was obtained when hexane was used as the reaction solvent. e When hexane
was used as the solvent the chemical yield was 10%. f Allowing for recovered catalyst, the enantiomeric excess was in favour of the (R)-alcohol. g 200
mg of 2% crosslinked polystyrene beads added.
R
R
R
R
R
N
O
Zn(C2H5)2
C2H5
R
N
N
OH
N
O
R
Zn
Reaction (i)
Zn
Zn
C2H5
O
2
C2H5
R
6
3
C2H5
C2H5
CH
O
Zn
Higher
aggregates
Zn
O
C2H5
Reaction (ii)
R
R
CH
H5C2
N
O
C2H5
C2H5
Zn
Zn
C2H5
R = alkyl
O
CH
5
4
Scheme 1
reactions. The aims were, first, to investigate a range of linear
and crosslinked PS catalysts, with various catalyst loadings and
various morphologies, so as to determine how, for this particu-
lar catalysed reaction, the choice of support affects the % ee
obtained and, second, to thus determine whether the % ees
achieved using optimal conditions in the PS reactions can
match those achieved using analogous soluble catalysts. PS
versions of eqn. (1) and related reactions have been studied
before on several previous occasions,16–24 but little attention has
been given to the optimisation of the support. Previous work
has been reviewed.25 Silica gel- and alumina-supported catalysts
have also been investigated.26 Previous work will be discussed
below where appropriate. Two recent papers on other types of
PS chiral catalyst highlight how crucial the choice of reaction
conditions can be.27,28
of diethylzinc binds to a lone pair of the oxygen atom in the
catalyst 3, thus enhancing the nucleophilic properties of the
ethyl groups. With both the reactants activated and held in close
proximity, see formula 4, reaction (ii) takes place through a six-
membered ring transition state to form the new C–C bond. The
product prior to “work up” is the alkoxide 5. A major feature
of the system particularly relevant to the present study is that
the catalyst groups 3 tend to reversibly aggregate to give, for
example, dimers 6.29–31 In the context of PS catalysts such
aggregation is important as it would involve site–site inter-
actions and these may be sufficient to serve as crosslinks, so
making diffusion within the matrix more difficult.
A second point relevant to the present work is the extent to
which the reaction shown in eqn. (1) proceeds in the absence of
added catalyst. This is relevant because the product from such a
reaction would be racemic. Since data on this topic are scarce,
in the present study benzaldehyde was treated with diethylzinc
under various conditions. The results are summarised in Table
1. It is evident, entries 1–3, that over a period of 24 h at 0 ЊC,
whether the reaction solvent is toluene or hexane, the chemical
yield of alcohol 1 is less than 1%. At 20 ЊC, however, significant
reaction occurs, the yields of 1-phenylpropanol (1) after 24 h
being 35% when toluene is the reaction solvent and 10% when
hexane is the solvent. Raising the temperature further, entries 4
and 5, results in more rapid reactions. These results are consist-
ent with the previous fragmentary literature reports.16,30,32
Since the catalysts 3 are alkylzinc alkoxides and the initial
reaction products 5 are also compounds of this general type, it
Results and discussion
The mode of action of the catalysts: some relevant observations
The mechanism by which β-amino alcohols (2) catalyse the
reaction shown in eqn. (1) and related reactions has been dis-
cussed by Noyori29,30 and the accepted mechanism is outlined in
Scheme 1. Essentially the β-amino alcohol first reacts with
diethylzinc, reaction (i), to generate the true catalyst 3. The zinc
centre in the catalyst 3 behaves as a Lewis acid and binds to a
lone pair on the oxygen atom of the carbonyl group, thus acti-
vating the latter to nucleophilic attack. A second molecule
1464
J. Chem. Soc., Perkin Trans. 1, 1999, 1463–1472

View Article Online
was of interest to determine whether the initial product from
eqn. (1), i.e. the alkoxide 5, would itself serve as a catalyst. It is
evident from the results summarised in Table 1, entries 1 and 6,
that the chiral R-alkoxide 5, generated in situ from (R)-1-
phenylpropanol (1), is indeed a catalyst for eqn. (1), but that, at
5 mol%, it is not chemically a very effective catalyst. Moreover,
it achieves essentially zero asymmetric synthesis. These results
are consistent with the brief literature reports on this topic.33,34
The poor chemical reactivity no doubt results, at least in part,
because this sterically uncrowded unchelated alkoxide has a
much greater tendency than those alkoxides which are the
more successful chiral catalysts to reversibly aggregate to give
catalytically inactive species.30,31 The very poor stereochemical
performance is no doubt associated with the lack of the stiff
well-organised arrangement that results from the presence of
the five-membered chelate ring in the better chiral catalysts. A
potential problem when the better chiral catalysts are used for
eqn. (1) is, however, that as the alkoxide 5 is the reaction prod-
uct, even though it is a relatively poor catalyst, this is partly
offset by the fact that it is eventually present in stoichiometric
amounts. Furthermore, it may also interfere by forming mixed
aggregates with the chelated alkoxides 3, so reducing their cata-
lytic efficiency.
CH3
C6H5
H
CH3
C6H5
H
H
H
CH2
CH3
CH2
N
OH
N
OH
CH3
7
8
x
1–x
CH3
C6H5
H
H
CH2
CH3
R
N
OH
9: x = 1.00
10: R = H ; x = 0.01–0.66
11: R = CH3; x = 0.08–0.52
12: R = H; x = 0.06–0.51; prepared
from crosslinked polymer beads
It is clear from the various results presented above, therefore,
that in the present work the PS catalysts operate in a competitive
situation and that if the aldehydes and diethylzinc cannot dif-
fuse freely into the polymer supports to reach the catalytic sites,
then they will tend to react together anyway, albeit slowly, to
give a racemate, and this will lead to a reduction in the overall
stereochemical performance of the supported catalysts.
Finally, it is of interest to note, compare Table 1 entries 1 and
7, that blank 2% crosslinked polystyrene beads also catalyse
eqn. (1), presumably by absorbing, and thus concentrating, the
reactants into the beads. Whether this is also true once the
beads are functionalised with the catalyst residues is not clear.
2 that, as expected given the structural and electronic similar-
ities of the polymerisable centres, the compositions of the
final copolymers are closely similar to the feed compositions.
Molecular weights of the polymers were determined by gel
permeation chromatography relative to polystyrene standards.
The optical rotations of the various polymers were measured
for solutions in chloroform: see Table 2. The specific rotations
obtained for the copolymers 10 agree well with those Fréchet
et al. reported for similar copolymers (different proportions of
the monomers) obtained by copolymerising monomers 8 with
styrene.36
Preparation of linear polymers containing (1R,2S)-N-benzyl-
ephedrine residues: the use of these polymers to catalyse eqn. (1)
Catalyst and solubility properties of the polymers. Initially
polymer 9 and the copolymers 10 were used as catalysts for eqn.
(1). The reactions were run for 24 hours in toluene at 20 ЊC
using benzaldehyde, diethylzinc and the catalyst in the mole
ratios 1.00:1.10:0.05. In each case the final reaction mixture
was quenched with methanol, then the polymer was filtered
off. The soluble products were extracted from the filtrate
Initially linear polymers35 containing (1R,2S)-N-benzyl-
ephedrine residues were investigated as catalysts for eqn. (1).
Although linear polymers are less easily separated from the
final reaction mixtures than crosslinked, and therefore in-
soluble, polymer beads, providing the linear polymers are sol-
uble in the reaction medium the possible diffusion problems
that are encountered with crosslinked polymers are avoided.
Thus, information on site–site interactions and on solubilities
may be obtained. The latter are of interest as they are closely
related to the swelling properties of the corresponding cross-
linked polymers.
1
and analysed by H NMR spectroscopy. 1-Phenylpropanol (1)
was the major product in all cases. Benzyl alcohol, evident from
a singlet at δ 4.65 ppm, was a minor product (≤9%) in most
cases: see Table 3. The presence of benzyl alcohol was con-
firmed by gas chromatography in comparison with an authentic
sample. The alcohol probably arises by reduction of the alde-
hyde by diethylzinc in the absence of an effective alkylation
catalyst.16,29,37 Distillation of the crude product gave pure 1-
phenylpropanol (1) and the % ee was then determined by polar-
imetry. In all cases the (R)-enantiomer of alcohol 1 was the
major product. The results of these experiments are summar-
ised in Table 3, entries 1 to 8.
It is evident from the results that as the loading of catalytic
sites decreases from 3.52 mmol gϪ1 down to 1.32 mmol gϪ1, the
chemical yields and the % ees increase. At and below 1.32 mmol
gϪ1 the yields of 1-phenylpropanol (1) are essentially quanti-
tative and the % ees plateau at 83%. Use of (1R,2S)-N-
benzylephedrine (7) as a catalyst in place of the polymers,
under otherwise similar reaction conditions, gave 1-phenyl-
propanol (1) in 92% chemical yield and 81% ee: Table 1, entry 8.
Thus, the polymeric catalysts 10d–10f, all with a loading of
≤1.32 mmol gϪ1, gave, within experimental error ( 1%), the
same % ee as the model catalyst 7.
Polymer synthesis and characterisation. (1R,2S)-Ephedrine
was selected as the starting material for the syntheses of the
initial set of catalysts investigated because it is readily available
and the use of (1R,2S)-N-benzylephedrine (7), or its enantio-
mer, as catalysts for eqn. (1) typically gives, depending on the
reaction conditions, ees of 70–85%.16,25 There is, therefore,
scope for the polymeric systems to give better, as well as poorer,
% ees than the non-polymeric systems.
Reaction of (1R,2S)-ephedrine with a commercial mixture of
3- and 4-vinylbenzyl chloride (mole ratio, 70:30) gave mono-
mers 8 and this mixture was used to prepare a range of linear
polymers: see Table 2. Homopolymerisation of monomers 8 in
toluene using a free radical initiator gave polymer 9. Under
similar reaction conditions monomers 8 were copolymerised
with styrene in various proportions to give copolymers 10a–
10g, and with 4-methylstyrene in various proportions to give
copolymers 11a–11e. The compositions of the copolymers 10
and 11 were determined by elemental analysis and by 1H NMR
spectroscopy. It is evident from the results summarised in Table
It is of interest to know why the % ees obtained vary with the
loading of catalyst groups. A major reason for using the linear
polymers was an attempt to avoid possible diffusion problems
J. Chem. Soc., Perkin Trans. 1, 1999, 1463–1472
1465

View Article Online
Table 2 Synthesis of linear polymers containing (1R,2S)-N-benzylephedrine residues
Molecular
Fraction of
catalyst monomer
Loading/
weights of
b
polymers (×10Ϫ3
)
[α]D20
—
—
Entry
1
Polymer
In feed
In polymer
mmol gϪ1
/deg cm2 gϪ1 a
Mn
Mw
9
1.00
1.00
3.52
Ϫ24.6
2.9
5.1
(a) With styrene as the diluent monomer
2
3
4
5
6
7
8
10a
10b
10c
10d
10e
10f
10g
0.75
0.50
0.25
0.12
0.06
0.03
0.01
0.66
0.52
0.28
0.18
0.05
0.03
0.01
2.98
2.65
1.81
1.32
0.44
0.27
0.09
Ϫ23.7
Ϫ23.2
Ϫ18.0
Ϫ14.3
Ϫ5.4
Ϫ4.6
Ϫ2.4
7.8
6.6
10.0
15.9
5.2
6.3
2.9
11.5
12.5
19.6
26.5
10.1
15.3
6.0
(b) With 4-methylstyrene as the diluent monomer
9
10
11
12
13
11a
11b
11c
11d
11e
0.50
0.40
0.30
0.20
0.10
0.52
0.46
0.36
0.17
0.08
2.56
2.38
2.03
1.16
0.61
Ϫ21.2
Ϫ18.7
Ϫ15.4
Ϫ10.3
Ϫ6.0
7.1
5.6
5.1
4.9
3.3
13.8
9.8
11.6
9.2
8.6
a In chloroform at 20 ЊC, concentration 5 g per 100 ml. b Determined by gel permeation chromatography. Values relative to polystyrene standards.
Table 3 Use of linear polymers containing (1R,2S)-N-benzylephedrine residues as catalysts for eqn. (1)a
Solubility in toluene
Starting
Toluene as reaction solvent
Yield
Hexane as reaction solvent
Yield
Loading of
catalyst
(%) of
alcohol
1
Yield (%)
of benzyl
alcohol
(%) of
alcohol
1
Yield (%)
of benzyl
alcohol
residues/
Entry
1
Polymer
mmol gϪ1
polymersb
Complexc
% eed
44
% eed
—
9
3.52
I
I
47
9
—
—
(a) With styrene as the diluent monomer
2
3
4
5
6
7
8
10a
10b
10c
10d
10e
10f
10g
2.98
2.65
1.81
1.32
0.44
0.27
0.09
I
I
S
S
S
S
S
I
I
I
I
I
S
S
56
66
80
95
98
100
100
6
5
4
2
1
0
0
56
67
79
81
82
83
83
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
(b) With 4-methylstyrene as the diluent monomer
9
10
11
12
13
11a
11b
11c
11d
11e
2.56
2.38
2.03
1.16
0.61
S
S
S
S
S
I
I
I
I
I
65
67
74
92
92
6
7
7
7
7
69
74
76
81
86
60
63
91
93
94
12
10
5
3
5
20
37
54
63
65
a Reaction carried out for 24 hours at 20 ЊC with toluene or hexane as the reaction solvent and mole ratios Zn (C2H5)2 :C6H5CHO:cata-
lyst = 1.10:1.00:0.05. The catalysts were prepared using (1R,2S)-ephedrine and the 1-phenylpropanol produced had the (R)-configuration. The % ees
were determined by polarimetry. See Experimental section for further details. b Solubility in toluene prior to reaction: I = insoluble; S = soluble. All
polymers were insoluble in hexane. c Complex formed on addition of diethylzinc solution in toluene. d Believed to be correct to 1%.
resulting from insolubility of the catalyst. However, even the
starting polymers with the higher loadings, i.e. polymer 9 and
copolymers 10a and 10b, were, unlike unfunctionalised linear
polystyrene, insoluble in toluene, presumably due to the pres-
ence of the polar β-amino alcohol residues and also the con-
sequent possibility of hydrogen-bonding between the polymer
chains. With the less highly loaded polymers that were soluble
in toluene, i.e. polymers 10c–10e, the addition of the diethylzinc
brought about precipitation. This may result (i) from the simple
change in the functional moieties from amino alcohol to chel-
ated amino alkoxide, (ii) from crosslinking due to dimerisation
of the chelated residues as indicated in formula 6 and/or (iii)
from the zinc alkoxide of one moiety complexing with the
amine group of a different moiety. Possibilities (ii) and (iii) both
involve site–site interactions. It is very difficult to distinguish
between these three effects because they are all expected to
become more important as the loading increases. The addition
of benzaldehyde would be expected to decrease the extent of
crosslinking due to reason (ii) but not necessarily to totally
remove it. In practice all the reaction systems involving polymer
9 and copolymers 10a–10e remained insoluble on the addition
of benzaldehyde, i.e. the alkylation reactions took place under
heterogeneous conditions. The very lightly loaded polymers
10f and 10g are of particular interest because the systems
involving these polymers remained soluble throughout the
entire reaction process. These homogeneous systems gave the
best chemical yields and stereochemical results, though not
significantly better than the heterogeneous systems involving
polymers 10c–10e. The latter, though insoluble, unlike the more
heavily loaded polymers, probably interact sufficiently well with
toluene to swell well under the reaction conditions.
Linear polymers 11a–11e, prepared using the readily avail-
1466
J. Chem. Soc., Perkin Trans. 1, 1999, 1463–1472

View Article Online
Table 4 Reactions of benzaldehyde with diethylzinc catalysed by various crosslinked polymer-supported ephedrinesa
Reaction
conditions
Toluene as
reaction solvent
Hexane as
reaction solvent
Catalyst
loading/
mmol gϪ1
Degree of
substitution
Chem.
yield (%)
Chem.
yield (%)
Entry
Polymer
Starting polymer
T/ЊC
t/h
% ee
% ee
1
2
3
4
5
6
7
8
9
12a
12a
12b
12b
12c
12d
12d
12e
12f
12g
12h
13
1% Crosslinked gel
1% Crosslinked gel
1% Crosslinked gel
1% Crosslinked gel
1% Crosslinked gel
1% Crosslinked gel
2% Crosslinked gel
2% Crosslinked gel
Amberlite XAD-4
Amberlite XAD-4
Polyhipe
0.93
0.93
1.13
1.13
2.30
2.30
1.06
2.62
0.90
0.50
1.51
0.69
0.12
0.12
0.15
0.15
0.40
0.40
0.14
0.51
0.11
0.06
0.21
—
0
22
0
23
0
23
0
0
70
20
70
20
70
20
70
70
70
22
70
24
81
85
77
85
80
86
70
71
59
75
59
67
81
78
77
74
74
73
72
69
28
28
42
36c
81
—
83
81
82
97
—
76
62
70
57
83
64
—
62
65
62
65
—
62
39
39
23
13
0
22
0
10
11
12
Amberlite XAD-4
20
siloxane graftb
a Unless indicated otherwise mole ratios C6H5CHO:Zn(C2H5):Catalyst were 1.00:1.10:0.02. The catalysts were prepared using (1R,2S)-ephedrine
and the 1-phenylpropanol produced had the (R)-configuration. The % ee were determined by polarimetry. The % ee are believed to be correct to 1%.
b Results reported in reference 22. c When 8 mol% of catalyst was used a 65% ee was obtained.
able 4-methylstyrene as the comonomer, were prepared in an
attempt to obtain catalyst systems which were soluble in the
reaction solvent at higher catalyst loadings than the 0.27 mmol
Reaction of three commercial samples of 1% crosslinked gel-
type polystyrene beads, containing 1.01, 1.40 and 3.20 mmol gϪ1
of chloromethyl groups respectively, with (1R,2S)-ephedrine
in toluene in the presence of potassium carbonate gave 1%
crosslinked beads 12a–12c containing different loadings of
(1R,2S)-N-benzylephedrine residues. 2% Crosslinked gel-type
polystyrene beads 12d and 12e were prepared similarly.
Amberlite XAD-4 is a type of highly crosslinked macro-
porous polystyrene beads.38 Chloromethylation of samples of
this polymer followed by reaction with (1R,2S)-ephedrine,
using the method outlined above, gave beads 12g and 12f with
different loadings of (1R,2S)-N-benzylephedrine residues.
Polyhipe is an open network of interconnected cells of
highly crosslinked macroporous polystyrene.39 Using the simi-
lar procedures to those used with the Amberlite XAD-4 beads,
polymer 12h was prepared.
g
Ϫ1 of polymer 10f. This was partially successful in that all the
polymers prepared using 4-methylstyrene, including the 50:50
copolymer, were soluble in toluene, but as before the addition
of diethylzinc caused precipitation. On the addition of benz-
aldehyde the reaction systems remained heterogeneous. The
trends in chemical yields and % ees as the loadings decreased
were similar to those observed with polymer 9 and copolymers
10: see Table 3. The % ees obtained with the two polymers with
≤1.16 mmol of catalyst sites per gram i.e. polymers 11d and 11e,
gave 1-phenylpropanol (1) in % ees equal to or greater than that
achieved with the model catalyst 7.
Hexane has often been used as the reaction solvent to carry
out the reaction shown in eqn. (1), or related reactions, with
either soluble18,32 or PS catalysts.17,18 A series of reactions was,
therefore, carried out using linear copolymers 11 in hexane, the
other reaction parameters being the same as those described
above. All the copolymers 11 were found to be insoluble in
hexane and the reaction systems were hetereogeneous through-
out. However, chemical yields were comparable to those
obtained before and again the % ees obtained increased as the
loading decreased. The best % ee, was again obtained with
polymer 11e but now it was only 65%, i.e. considerably less than
the 86% obtained using toluene as the solvent.
Amberlite XAD-4 has a significant content of vinyl groups.38
These arise because in its synthesis relatively large amounts of
divinylbenzene are used as the crosslinking agent and in the
case of many molecules only one of the vinyl groups takes part
in the polymerisation reaction. As reported previously,22
poly(methylhydrosiloxane) can be grafted to these residual vinyl
groups using a platinum-containing catalyst. A portion of the
remaining Si–H group can then be reacted similarly with
monomer 8 to give catalyst 13. The results of using polymer 13
as a catalyst are included here simply for comparison with the
present results.
The main conclusion to be drawn from this study of the
linear polymers are that in these systems toluene is a better
choice of reaction solvent than hexane and that for best results
loadings of catalytic groups should be less than ca. 1.5 mmol
gϪ1. These trends almost certainly mainly reflect the solubility of
the catalysts in the reaction solvent or, when the catalyst systems
are insoluble, their ability to swell in the reaction solvent.
CH CH2
CH2
CH3
Si
CH2
Si
CH3
Si
O
O
O
Preparation of crosslinked polymers containing (1R,2S)-N-
benzylephedrine residues: the use of these polymers as catalysts
of eqn. (1)
a
c
b
CH3
(CH2)7CH3
CH2 CH2
13
CH3
Attention was next turned to the preparation of crosslinked
polymer beads containing catalyst groups and their use as cata-
lysts. As noted above, such polymers, being totally insoluble in
all solvents, are much more convenient to use than the linear
polymers.
C6H5
H
H
CH2
CH3
N
OH
Polymer synthesis. A range of crosslinked polystyrene beads
containing (1R,2S)-N-benzylephedrine residues were prepared
as detailed below. The loadings achieved are as given in Table 4.
Catalyst properties of polymer beads 12a–12h and 13. The
various polymers described in the preceding section were used
to catalyse eqn. (1). The reactions were carried out using
J. Chem. Soc., Perkin Trans. 1, 1999, 1463–1472
1467

View Article Online
Table 5 Reactions of aldehydes with diethylzinc in toluene catalysed by (1R,2S)-N-benzylephedrine 7 or polymer-supported ephedrine 12aa
Model catalyst 7
PS-Catalyst 12a
Entry
Aldehyde
T/ЊC
Chem. yield (%)b
% eec
Chem. yield (%)b
% eec
1
2
3
4
5
6
Benzaldehyde
Benzaldehyde
2-Methoxybenzaldehyde
2-Methoxybenzaldehyde
4-Chlorobenzaldehyde
4-Chlorobenzaldehyde
0
23
0
23
0
83
87
82
90
84
83
84
76
89
85
94
89
81
85
88
91
83
84
81
78
90
83
93
86
23
a Reactions were carried out in toluene with aldehyde:Zn(C2H5)2 :catalyst in mole ratios 1.00:1.10:0.02 for 70 hours at 0 ЊC or 20 hours at 20 ЊC.
b Unless indicated otherwise yield of alcohol is that isolated by bulb-to-bulb distillation. c By polarimetry. Believed to be correct to 1%. See
Experimental section for further details.
benzaldehyde, diethylzinc and the catalyst in mole ratios of
1.00:1.10:0.02 in either toluene or hexane as the reaction
solvent at 0 ЊC or 20–23 ЊC. At the end of the reaction period
the polymer beads were simply filtered off and washed, then the
filtrate and washings analysed by gas chromatography using
internal standards to determine the yield of 1-phenylpropanol
(1). These analyses also showed that up to 8% of the benzalde-
hyde was reduced to benzyl alcohol,16,29,37 the larger percentages
being obtained with the less active catalysts. Removal of the
solvent and bulb-to-bulb distillation of the product gave a
under similar conditions. The reasons for these latter differ-
ences is not clear but is almost certainly related to the fact
hexane does not interact as well as toluene with the support
matrix.
To confirm that catalyst 12a in toluene gave stereochemical
results very similar to those obtained with (1R,2S)-N-benzyl-
ephedrine (7) under similar conditions, reactions were also
carried out with 2-methoxybenzaldehyde and 4-chlorobenz-
aldehyde using the aldehyde, diethylzinc and catalyst in the
mole ratios 1.00:1.10:0.02 at 22 ЊC. The results are summarised
in Table 5 together with those for benzaldehyde previously
included in Table 4. It is evident that for all three aldehydes, at
both 0 ЊC and 23 ЊC, the % ee of the PS catalyst 12a is never
more than 4% less than that obtained with model catalyst (7)
under the same conditions.
1
sample of alcohol 1 pure by H NMR spectroscopy. The % ee
was determined by polarimetry. The results are summarised in
Table 4.
The results obtained with toluene and with hexane as the
reaction solvent differed significantly but since the better stereo-
chemical results were obtained using toluene, these results will
be discussed first. The lightly crosslinked gel-type polymers
proved to be the best supports both in terms of the chemical
yields and the % ees. As expected from the results with the
linear polymers, the % ees achieved were greater with the less
heavily loaded polymers, and the 1% crosslinked polymer, being
more easily swollen by the reaction solvent, gave better % ees
than the 2% crosslinked polymer. The % ees were also greater at
0 ЊC than at 23 ЊC. Thus, the best % ee was obtained under the
reaction conditions summarised in Table 4 entry 1 and was
81%. This value is only slightly less than the 84% ee obtained
using (1R,2S)-N-benzylephedrine (7) under otherwise similar
reaction conditions: Table 1 entry 9. The reactions were, how-
ever, very slow at 0 ЊC. Raising the reaction temperature to
23 ЊC allowed the reaction time for a useful yield to be reduced
to 20 h. The ees were then 78% and 76% respectively for the PS
catalyst 12a and the model catalyst 7.
The results obtained with the catalyst supported on Amber-
lite XAD-4 or Polyhipe, especially the % ees, were substan-
tially lower than those obtained using the gel-type supports
despite the fact the former had catalyst loadings of ≤1.51 mmol
gϪ1. It may be that with these polymers a substantial fraction of
the catalyst sites are present in the more highly crosslinked parts
of the support matrix and are, therefore, not readily accessible.
It is also possible that many of the catalyst sites are at the inner
surfaces of the pores so that the local concentration of catalyst
groups is high and this may result in significant site–site inter-
actions. Access to the catalytic groups on the grafted support 13
also appears to be hindered as raising the mol percentage of
catalyst from 2% to 8% raised the ee from 36% to 65%.22
Perhaps the graft substantially blocks the pores.
Preparation of catalysts containing (؉)-3-exo-(N-benzyl,N-
methylamino)isoborneol residues: the use of these to catalyse eqn.
(1) and related reactions
Having identified suitable conditions for the preparation of
successful linear and crosslinked PS ephedrine catalysts, atten-
tion was next given to preparing other PS catalysts similarly
but starting from a β-amino alcohol which is known to afford
higher % ees. Noyori et al. have shown that (Ϫ)-3-exo-
(dimethylamino)isoborneol (DAIB) (14) is an excellent catalyst
for the asymmetric alkylation of aromatic aldehydes using
diethylzinc,40 so this catalyst and various analogues were pre-
pared and studied.
Synthesis of catalysts. The β-amino alcohol 15 was prepared
in four steps (see Experimental section) from (1R)-camphor
(16). Reaction of this amino alcohol with aqueous formalde-
hyde and formic acid gave a sample of DAIB (14), and reaction
with benzaldehyde to give the imine and alkylation of the imine
with methyl iodide gave the secondary amino alcohol 17
(Scheme 2). Reaction of the latter with benzyl chloride gave the
CH3
N
NH2
OH
CH3
O
OH
16
15
14
The results obtained with hexane as the solvent showed a
different pattern, though the gel-type supports were again the
best. The better % ees obtained with these supports were never-
theless significantly lower than those obtained with toluene as
the solvent. Surprisingly the % ees obtained at 0 ЊC were lower
than those obtained at 23 ЊC, and whilst some of the % ees
obtained with the macroporous-type supports were poor, they
were nevertheless better than those obtained using toluene
H
N
CH2
N
CH3
CH3
OH
OH
17
18
Scheme 2
1468
J. Chem. Soc., Perkin Trans. 1, 1999, 1463–1472

View Article Online
Table 6 Reactions of aldehydes with diethylzinc catalysed by catalysts 14, 16 and 20–22a
Polymer
typeb
Loading/
Yield of chiral
alcohol (%)c
Entry
Aldehyde
Catalyst
mmol gϪ1
% eed
1
2
3
4
5
6
7
8
Benzaldehyde
4-Chlorobenzaldehyde
Benzaldehyde
4-Chlorobenzaldehyde
Benzaldehyde
Benzaldehyde
14
14
18
18
20
21
22
22
—
—
—
—
L
L
X
—
—
—
—
3.34
2.86
0.64
0.64
97
88
90
91
89
91
94
94
97e
92f
95e
90f
96e
98e
97f
92e
Benzaldehyde
4-Chlorobenzaldehyde
X
a Reactions carried out for 24 hours at 20 ЊC in toluene using aldehyde, diethylzinc and catalyst in the mole ratios 1.00:1.10:0.05. b L = linear
polymer; X = crosslinked polymer. c Determined by 1H NMR spectroscopic analysis of extracted organic materials. d The predominant enantiomer
was the (S)-alcohol. e Determined by 19F NMR of the Mosher ester derivative:41 see Experimental section. f By polarimetry: see Experimental section.
tertiary amino alcohol 18 which is a low molecular weight
analogue of the PS catalysts. A similar reaction with the com-
mercial mixture of 3- and 4-vinylbenzyl chlorides gave mono-
mer 19. The latter was homopolymerised and copolymerised
with styrene in the same manner as with the ephedrine-derived
analogues, to give homopolymer 20 and the 2:1 copolymer 21.
Reaction with 1% crosslinked polystyrene beads containing
0.70 mmol gϪ1 of chloromethyl groups gave crosslinked beads
22 containing 0.64 mmol gϪ1 of catalyst residues.
well, i.e. dissolves or swells extensively, with the reaction sol-
vent. It assists in achieving this if the catalytic moieties have a
substantial lipophilic character and so interact well with
toluene.
The % ees obtained with the crosslinked polymer beads in
reactions with benzaldehyde and 4-chlorobenzaldehyde were
also high, see entries 7 and 8, and were slightly higher than
those obtained with the model catalyst 18, compare with entries
3 and 4. The % ee of the product from the reaction summarised
in entry 7 was estimated both by polarimetry and by the use of
Mosher esters.41 The % ee obtained with benzaldehyde (97%)
using 5% of catalyst at 20 ЊC is somewhat higher than that
(92%) reported previously using 5 mol% of a similar polymer at
0 ЊC.16 In the present work a 1% crosslinked polymer with a
catalyst loading of 0.64 mmol gϪ1 was used. It is not clear from
the brief details given in the literature report16 just what poly-
mer was used, but it may well have been a 1–2% crosslinked
polymer with a catalyst loading of ca. 1 mmol gϪ1.
Catalyst properties of compounds 14 and 18 and polymers 20–22
The various catalysts were used to catalyse eqn. (1) and, in
some cases, the analogous reactions using 4-chlorobenzalde-
hyde. The reactions were carried out in toluene at 20 ЊC for 24
hours using the aldehyde, diethylzinc and the catalyst in the
mole ratios 1.00:1.10:0.05. The results are summarised in
Table 6. Note that with these catalysts the main product is the
(S)-enantiomer.
The % ees obtained with DAIB (14), entries 1 and 2, were
slightly lower than those reported by Noyori et al.,40 but this is
to be expected because they carried out the reactions at 0 ЊC.
The corresponding results obtained with catalyst 18, an excel-
lent model for the polymers, were in the case of both aldehydes
5% less than that obtained with DAIB (14). Thus, replacement
of an N-methyl group by a more bulky N-benzyl group results
in a small fall in the % ee.
The stereochemical results obtained with the two linear
polymeric catalysts were excellent, see Table 6 entries 5 and 6,
the % ees being very similar to those obtained with the non-
polymeric catalysts 14 and 18. Given the relatively high load-
ings of the catalyst moieties, ≥2.48 mmol gϪ1, this is at first
surprising. However, unlike the ephedrine-derived polymers
of similar loading which showed a relatively poor stereo-
chemical performance, polymers 20 and 21 were soluble in
Conclusions
Many attempts have been made to produce PS versions of the
better chiral catalysts for asymmetric organic synthesis, but in
many cases the % ees achieved have been less than those
achieved using the analogous non-polymeric catalysts.11 In this
paper the reactions of benzaldehyde with diethylzinc catalysed
by PS ephedrine or camphor derivatives have been investigated
in some depth in an attempt to identify the crucial factors
necessary to successfully prepare PS chiral catalysts for such
reactions. The most important factor is found to be a favour-
able interaction of the polymer matrix with the reaction solvent
so that the polymer will dissolve or swell to allow the other
reactants easy access to the catalytic sites. Accordingly toluene
is a better reaction solvent than hexane. The ephedrine-derived
catalytic groups themselves reduce the solubility of the linear
polymers in toluene and, almost certainly, the swelling proper-
ties of the crosslinked polymers. Thus, of the polymers investi-
gated the better linear ones are those with <ca. 1.5 mmol per g
of catalyst sites and the better insoluble ones are 1% crosslinked
gels with <ca. 1.0 mmol per g of catalyst sites. At greater load-
ings and higher percentages of crosslinking a significant frac-
tion of the catalyst sites become inaccessible. The reaction in
eqn. (1) which occurs without being under the influence of the
intended catalyst 3, and so achieves low or zero stereoselect-
ivity, makes an important contribution, and the reduction to
give benzyl alcohol becomes more significant. Site–site inter-
actions and microenvironmental effects may well be present at
the loadings used9 but they do not appear to play a key role in
these PS reaction systems. The reaction of benzaldehyde with
diethylzinc using the best linear PS ephedrine derivatives 10f,
10g or 11e affords 1-phenylpropanol (1) with 83–86% ee of the
(R)-enantiomer; using the best linear PS camphor derivative 21
affords the alcohol (1) with 98% ee of the (S)-enantiomer; using
the best crosslinked PS ephedrine derivative 12a affords the
x
1–x
CH2
N
CH3
CH2
N
OH
CH3
OH
19
20: x = 1.00
21: x = 0.66
22: x = 0.11; prepared from
crosslinked polymer beads
toluene both before and after the addition of diethylzinc.
This strongly suggests that in the present reaction system the
key factor for success is a polymer support which interacts
J. Chem. Soc., Perkin Trans. 1, 1999, 1463–1472
1469

View Article Online
alcohol (1) with 78–81% ee of the (R)-enantiomer and using the
best crosslinked PS camphor derivative 22 gives alcohol 1 with
97% ee of the (S)-enantiomer. These values are close to those
obtained using analogous non-polymeric catalysts under
similar reaction conditions. With other types of reactions and/
or using other types of catalyst the most crucial factor(s) for
success may well be different.
ene, 10.9 mmol) was stirred at 20 ЊC under an atmosphere of
dry nitrogen for 1 h. A solution of benzaldehyde (902 mg, 9.4
mmol) in toluene (1.6 ml) was added and the mixture stirred at
0 ЊC for 48 h. The yield of product was estimated as in (a).
Bulb-to-bulb distillation (105 ЊC, 0.5 mm Hg) of the crude
product using a Büchi GKR-50 microdistillation apparatus
gave 1-phenylpropanol (1). The optical rotation, [α]D20 was
measured for a 5% solution in chloroform and compared to a
value of Ϫ45.45 (c 5.15, CHCl3) reported42 for the pure
S-enantiomer.
Experimental
(1R,2S)-(Ϫ)-Ephedrine (99%), (1R)-(ϩ)-camphor (99%) and
diethylzinc in toluene were obtained from Aldrich. Chloro-
methylated 1% and 2% crosslinked polystyrene beads were
obtained from Kodak. Polyhipe was a gift from the National
Starch and Chemical Corporation, New Jersey. Other chem-
icals were obtained from Fluka. Solvents and reagents were
dried and purified according to standard procedures. Organic
extracts were dried with magnesium sulfate. Samples were dried
in a vacuum oven at 1.0 mm Hg.
(c) Entry 7. This was carried out as described in (a) above
except that 2% crosslinked gel-type beads (200 mg) were added
to the reaction mixture.
(d) Entries 8–10. These experiments were carried out as
described in (b) above except that catalyst (7) (amounts as
indicated in Table) was used in place of alcohol 1.
(1R,2S)-N-(3- and 4-Vinylbenzyl)ephedrines (8)
Infrared spectra were recorded using either a Nicolet MX1
instrument or a Perkin-Elmer 1720 instrument. Solid samples
were prepared as potassium bromide discs unless stated other-
wise, liquid samples were prepared as thin films between
These were prepared using the procedure described above for
compound 7 except that a commercial mixture of 3- and 4-
vinylbenzyl chloride (70% the 3-isomer; 30% the 4-isomer) was
used. Flash column chromatography with hexane–ethyl acetate
7:1 as the eluant gave the desired product as an oil (28.21g, 35%
yield). It had [α]D Ϫ36.3 (c = 2.35, CHCl3) (lit.,36 [α]D Ϫ41.8
(c = 2.35, CHCl3) for the 4-isomer).
1
sodium chloride plates. H NMR spectra were recorded for
solutions in deuterated chloroform on a Varian Gemini 200
MHz NMR spectrometer using TMS as an internal standard.
Elemental analyses for chlorine were carried out by Butter-
worth Laboratories Limited; analyses for nitrogen were made in
house on a Carlo Erba model 1106 instrument. Optical
rotations were measured using a Perkin-Elmer 141 digital polar-
imeter, in a cell of path length 10 cm. Gas chromatographic
(GC) analyses were carried out using a Pye 204 Chromatograph
equipped with a 10% SP1000 stationary phase at 220 ЊC and a
flame-ionisation detector. Molecular weights and polydispersi-
ties were determined by gel permeation chromatography (GPC)
using a Waters Associates model 502 system equipped with UV
and RI detectors. PL gel columns 100 Å and 500 Å were used.
The eluant was THF, at a flow rate 1 cm3 minϪ1, and data
acquisition and analysis was achieved with a Trilab 2000
chromatography data system.
Preparation of linear polymers 9, 10 and 11
(a) Typical procedure: polymerisation of monomer (8) to give
homopolymer (9). (1R,2S)-N-(3- and 4-Vinylbenzyl)ephedrines
(8) (17.9 mmol) were dissolved in toluene (10 ml) and azo-
bisisobutyronitrile (0.81 mmol, 4.5 mmol%) was added as the
initiator. The solution was degassed thoroughly using three
freeze–thaw cycles. The tube was sealed under vacuum and
polymerisation achieved by heating at 75 ЊC for 24 h. The tube
was then cooled, the toluene removed by rotary evaporation
and the residue dissolved in dichloromethane. The polymer was
precipitated twice into petroleum ether bp (60–80 ЊC). The
product was collected and dried in a vacuum oven at 40 ЊC for
24 h. This gave a pale brown powder (3.58 g). Microanalysis
indicated a loading of 3.52 mmol N per g polymer. The molecu-
lar weight of the product was determined using GPC: the result
is given in Table 2. It had [α]D Ϫ24.6 (c = 2.4, CHCl3); δ (500
MHz) 0.9 (br, 3H, C-CH3), 1.1–2.3 (br, 6H, backbone, N-CH3),
2.8 (br, 1H, C-2), 3.4 (br, 2H, Ar-CH2), 4.8 (br, 1H, C-1) and
6.0–7.4 ppm (br, 9H, Ar-H).
(1R,2S)-N-Benzylephedrine (7)
(1R,2S)-Ephedrine (43.1 mmol) and benzyl chloride (43.1
mmol) were dissolved in pyridine (20 ml) and stirred at room
temperature for 7 days. After reaction, the mixture was
quenched with water (20 ml), the product extracted three times
with ethyl acetate (60 ml). The extracts were dried and the sol-
vent was then evaporated off. The residue was purified by flash
column chromatography using petroleum ether (bp 40–60 ЊC)–
ethyl acetate = 3:1 as the eluant. This afforded white crystals
(5.21 g, 47% yield), mp 49–51 ЊC; [α]D Ϫ29.5 (c = 2.35, CHCl3);
δ 1.02 (d, 3H, C-CH3), 2.23 (s, 3H, N-CH3), 2.96 (m, 1H, C-2),
3.64 (s, 2H, Ar-CH2), 4.90 (d, 1H, C-1) and 7.22–7.45 ppm (m,
10H, Ar-H). Found C, 80.0; H, 8.4; N, 5.5: C17H21NO requires
C, 80.0; H, 8.3; N, 5.5%.
(b). Other polymers were prepared similarly from monomers
(8) and styrene or 4-methylstyrene. The feedstock compositions
and the compositions of the copolymers and the optical rota-
tions were as summarised in Table 2.
Catalytic properties of linear polymers 9–11, 20 and 21
The following procedure is typical of the experiments summar-
ised in Tables 3 and 6 using linear polymers.
Reactions summarised in Table 1
(a) Entries 1–5. A mixture of benzaldehyde (902 mg, 9.4
mmol) and diethylzinc (9.5 ml of a 1.1 M solution in toluene;
10.4 mmol) in toluene (2.0 ml) was stirred magnetically under
an atmosphere of dry nitrogen at the temperatures and for the
times given in the Table. The reactions were then quenched by
the addition of aqueous hydrochloric acid (6 ml of 1 M) and the
organic product extracted into toluene. The extracts were dried
and then analysed by GC using durene (1,2,4,5-tetramethyl-
benzene) as an internal standard.
Entries 1–8 in Table 1. A mixture of the catalyst (0.24 mmol,
5 mol%) in toluene (5 ml) was cooled down to 0 ЊC and stirred
magnetically. Diethylzinc (5.2 mmol as a 1.1 M solution in tolu-
ene) was added and the mixture stirred for an hour before the
temperature was raised to 20 ЊC. Benzaldehyde (4.7 mmol) in
toluene (2 ml) was then added and the reaction mixture stirred
for 24 hours. At the end of the reaction period the solution
was cooled down to 0 ЊC and quenched with methanol (5 ml)
and 1 M hydrochloric acid (5 ml). The precipitated polymer was
filtered off. Toluene (20 ml) was added to the filtrate and the
organic layer separated. The aqueous layer was extracted with
diethyl ether (3 × 15ml), the combined extracts washed with
(b) Entry 6. A mixture of (R)-1-phenylpropanol (1) (64 mg,
0.47 mmol) and diethylzinc (9.9 ml of a 1.1 M solution in tolu-
1470
J. Chem. Soc., Perkin Trans. 1, 1999, 1463–1472

View Article Online
water (5 ml) and finally dried. The organic solvents were care-
fully evaporated off. The residue was analysed by H NMR
4-chlorobenzaldehyde distilled at 135 ЊC/0.1 mmHg and that
1
from 2-methoxybenzaldehyde at 150 ЊC/0.1 mmHg. The
experimental [α]D20 were compared with the literature values of
Ϫ24.2Њ (c = 5, C6H6) for (S)-1-(4-chlorophenyl)propanol43 and
of ϩ54.0 (c = 1.2, toluene) for (R)-1-(2-methoxyphenyl)-
propanol.44
spectroscopy then purified by bulb-to-bulb distillation (110 ЊC,
1 mm Hg). The purity of the distillate was confirmed by GC
and H NMR spectroscopy, then the [α]D20 optical rotation was
1
measured for a solution in chloroform.
Synthesis of camphor derivatives
Preparation of polymer-supported catalysts 12a–12e
(a) Primary amino alcohol 15. (1R)-Camphor 16 was con-
verted into amino alcohol 15 in 4 practical steps essentially as
described by Bonner and Thornton.45 It had mp 188–192 ЊC
(lit.,45 mp 190–194 ЊC); [α]D Ϫ6.72 (c = 1.4, MeOH) (lit.,45 [α]D
Ϫ1.3Њ (c = 1.43, MeOH); νmax 3393 (O–H, N–H), 2948 (–CH2–,
–CH3) and 1107 cmϪ1 (C–O); δ 0.79 and 0.95 (2s, 6H, C-8 and
C-9 CH3), 1.06 (s, 3H, C-10 CH3), 1.40–1.85 (m, 4H, C-5 and
C-6 CH2) and 3.38 ppm (d, 1H, CHO); CI MS: m/z 170
(M ϩ H)ϩ.
These were prepared from commercial samples of chloro-
methylated 1% crosslinked polystyrene beads containing 1.01,
1.40, or 3.20 mmol of chlorine per gram. A typical procedure is
as follows.
Preparation of polymer 12a. (1R,2S)-Ephedrine (60.6 mmol),
1% cross-linked chloromethylated polystyrene (20.16 g, 20.2
mmol) of loading 1.01 mmol Cl per g polymer and potassium
carbonate (20.3 mmol) were suspended in toluene (125 ml), and
the mixture mechanically stirred and heated under reflux for 48
h. At the end of the reaction period, the beads were filtered off
and successively washed with methanol, water, tetrahydrofuran,
tetrahydrofuran–water (1:1), and diethyl ether. Since the beads
were to be used as a catalyst they were Soxhlet-extracted with
diethyl ether for 48 h before being dried in a vacuum oven at
40 ЊC for 48 h. This afforded 1% cross-linked polystyrene-
supported ephedrine-derived catalyst 12a (22.15 g), containing,
by elemental analysis, 0.93 mmol N per g polymer.
(b) Tertiary amino alcohol 14. Reaction of amino alcohol 15
with formaldehyde and formic acid as described by Chittenden
and Cooper46 gave amino alcohol 14 as an oil, bp 70 ЊC/0.1 mm
Hg. It had [α]D Ϫ8.0 (c = 4.3, EtOH); δ 0.79, (s, 3H, C-8 CH3),
0.98 (s, 3H, C-9 CH3), 1.16 (s, 3H, C-10 CH3) 1.20–1.95 (m, 4H,
C-5 and C-6 CH2), 2.25 (s, 6H, N(CH3)2), 3.67 (d, 1H, CHN),
3.81 (d, 1H, C-4) and 4.50 ppm (d, 1H, CHO).
(c) Secondary amino alcohol 17. Primary amino alcohol 15
(66.5 mmol) and benzaldehyde (77.9 mmol) were dissolved in
benzene (50 ml). After standing for 24 h the mixture was heated
under reflux for 12 h whilst water was removed using a Dean–
Stark trap. The benzene was then removed by rotary evapor-
ation and the residue dissolved in dimethylformamide (6 ml).
Methyl iodide (80.3 mmol) was added and the mixture allowed
to react at 20 ЊC for 18 hours. Diethyl ether was then added to
this mixture until a permanent cloudiness appeared. Cooling
caused the precipitation of an orange solid. The solids were
collected and heated under reflux with ethanol–water = 1:1 (75
ml) for 4 hours after which the aqueous solution was extracted
once with chloroform (20 ml) to remove any organic by-
product. The aqueous solution was acidified with 1 M hydro-
chloric acid, then made alkaline with 10% sodium hydroxide
and extracted with chloroform (3 × 75 ml). The combined
organic extracts were washed with water (50 ml) and dried.
Evaporation of the solvents gave amino alcohol 17 as a yellow
oil (5.26g, 43% yield). It had [α]D Ϫ12.3 (c = 1.5, CHCl3); νmax
3334 (O-H), 2950 (CH2 and CH3) and 1096 cmϪ1 (C–O); δ 0.75
and 0.92 (2s, 6H, C-8 and C-9 CH3), 1.01 (s, 3H, C-10 CH3),
1.10–1.80 (m, 4H, C-5 and C-6 CH2), 2.42 (s, 3H, N-CH3), 2.57
(d, 1H, CHN), 2.91 (d, 1H, C-4) and 3.41 ppm (d, 1H, CHO);
CI MS: m/z 184 (M ϩ H)ϩ.
Preparation of polymer-supported catalysts 12f–12h
To remove impurities the commercial Amberlite XAD-4 was
Soxhlet-extracted sequentially with water, methanol, THF, hex-
ane and ether for 24 h with each solvent and then dried at 90 ЊC
in the vacuum oven for 2 days. The following preparation of
catalyst 12f is typical. Catalyst 12g was prepared similarly using
half the amount of dimethoxymethane, thionyl chloride and tin
chloride. Catalyst 12h was prepared similarly to polymer 12f
but starting with Polyhipe.
(a) Chloromethylation of Amberlite XAD-4. In a 1 L,
3-necked, round-bottomed flask equipped with a reflux con-
denser, stirrer and addition funnel, Amberlite XAD-4 (50 g)
was stirred in dichloromethane (500 ml) at 35 ЊC under a nitro-
gen atmosphere for 30 minutes. Dimethoxymethane (116 ml)
was then added to the suspension over 10 minutes and the mix-
ture stirred for a further 1 h. After cooling to room temperature
thionyl chloride (73.6 ml) and stannic chloride (2.8 ml) were
added slowly so as to maintain a gentle reflux (ca. 1 h). When
the addition was complete the temperature was raised to 35 ЊC
and the suspension stirred for a further 18 hours. At the end of
the reaction period the suspension was cooled to 20 ЊC and
distilled water (200 ml) was added slowly. The suspension was
allowed to stand to allow any excess chloromethyl methyl ether
(CAUTION: SEVERE CARCINOGEN) to hydrolyse. The
polymer was then filtered off and washed successively with
water, 2 M hydrochloric acid, water, dichloromethane, THF,
THF–water and methanol. The polymer was dried under
vacuum at 20 ЊC overnight. By elemental analysis it contained
1.72 mmol of chlorine per g.
(d) Tertiary amino alcohol 18. This compound was prepared
from secondary amino alcohol 17 using the same procedure as
that given above for the synthesis of compound 7 from ephe-
drine. Compound 18 was obtained as a pale yellow oil (26%
yield). It had [α]D ϩ4.98 (c = 1.20, CHCl3); νmax 3338 (O–H),
2957 (CH2 and CH3), 1455 (-C(CH3)2), 1370 (O-H), 740 and 699
cmϪ1 (5 adjacent Ar-H); δ 0.80 and 1.01 (2s, 6H, C-8 and C-9
CH3), 1.13 (s, 3H, C-10 CH3), 1.20–1.90 (m, 4H, C-5 and C-6
CH2), 2.13 (s, 3H, N-CH3), 2.55 (d, 1H, CHN), 3.57 (d, 1H,
CHO) and 7.20–7.55 ppm (m, 5H, Ar-H).
(b) Reaction of chloromethylated Amberlite XAD-4 with
(1R,2S)-ephedrine. This was carried out using the procedure
given above for the preparation of polymer 12a. The final
polymer 12f contained 0.90 mmol of nitrogen per g.
(e) Monomers 19. These were prepared from secondary
amino alcohol 17 using the same procedure as that given above
for the synthesis of monomers 8. Monomers 19 were obtained
as a pale yellow oil (26% yield). The oil had [α]D Ϫ11.75
(c = 2.35, CHCl3); δ 0.80 and 1.01 (2s, 6H, C-8 and C-9), 1.14 (s,
3H, C-10), 1.43* (s, 2H, Ar-CH2), 2.15 (s, 3H, N-CH3), 2.55 (d,
1H, CHN), 3.58 (d, 1H, CHO), 5.25* (d, 1H, ABX), 5.75*
(d,1H,ABX), 6.72* (d, 1H, ABX) and 7.15–7.45 ppm (m, 4H,
Catalytic properties of crosslinked polymers (12) and (22)
These experiments were carried out similarly to those described
above using the linear polymeric catalysts, except that (i) at the
end of the reaction period the polymer beads were filtered off
prior to treating the filtrate with hydrochloric acid, and (ii) the
organic product was analysed by GC. The product from
J. Chem. Soc., Perkin Trans. 1, 1999, 1463–1472
1471

View Article Online
6 A. Akelah and D. C. Sherrington, Chem, Rev., 1981, 81, 557.
Ar-H). (*Smaller peaks were present at these positions due to
the presence of the 4-isomer.) Found C, 80.2; H, 9.6; N, 5.1;
C20H29NO requires C, 80.2; H, 9.8; N, 4.7%. CI MS: m/z
300 (M ϩ H)ϩ.
7 A. Akelah and D. C. Sherrington, Polymer, 1983, 24, 1369.
8 P. Hodge, Annu. Rep. Prog. Chem., Sect. B, Org. Chem., 1986, 283.
9 P. Hodge, Chem. Soc. Rev., 1997, 26, 417.
10 C. Dreisbach, G. Wischnewski, U. Kragl and C. Wandrey, J. Chem.
Soc., Perkin Trans. 1, 1995, 875.
11 P. Hodge, in Innovation and Perspectives in Solid Phase Synthesis,
ed. R. Epton, SPCC (UK) Ltd., Birmingham, 1990, pp. 273–292.
12 R. Noyori, Asymmetric Catalysis in Organic Synthesis, John Wiley,
New York, 1994.
Synthesis of polymer 20
Polymerisation of monomers 19 using the procedure given
above for the synthesis of polymer 9 gave polymer 20. It had
[α]D Ϫ9.8 (c = 2.3, CHCl3); νmax 3341 (O–H), 1606 (C᎐C), 1041
13 G. Liu and J. A. Ellman, J. Org. Chem., 1995, 60, 7712.
14 S. R. Gilbertson and X. Wang, Tetrahedron Lett., 1996, 37, 6475.
15 B. M. Cole, K. D. Shimizu, C. A. Krueger, J. P. A. Harrity, M. L.
Snapper and A. H. Horveyda, Angew. Chem., Int. Ed. Engl., 1996,
35, 1668.
16 S. Itsuno and J. M. J. Frechet, J. Org. Chem., 1987, 52, 4140.
17 K. Soai, S. Niwa and M. Watanabe, J. Org. Chem., 1988, 53, 927.
18 K. Soai, S. Niwa and M. Watanabe, J. Chem. Soc., Perkin Trans. 1,
1989, 109.
19 K. Soai and M. Watanabe, J. Chem. Soc., Chem. Commun., 1990,
43.
20 S. Itsuno, Y. Sakurai, K. Ito, T. Maruyama, S. Nakahama and J. M.
J. Frechet, J. Org. Chem., 1990, 55, 304.
21 M. Watanabe, S. Araki, Y. Butsugan and M. Uemura, Chem.
Express, 1990, 5,761.
22 Z. Zhang, P. Hodge and P. W. Stratford, React. Polym., 1991, 15, 71.
23 M. Watanabe and K. Soai, J. Chem. Soc., Perkin Trans. 1, 1994, 837.
24 D. Seebach, R. E. Marti and T. Hintermann, Helv. Chim. Acta,
1996, 79, 1710.
25 K. Soai and S. Niwa, Chem. Rev., 1992, 92, 833.
26 K. Soai, M. Watanabe and A. Yamamoto, J. Org. Chem., 1990, 55,
4832.
27 K. Burgess and D. Lim, Chem. Commun., 1997, 785.
28 L. Canali, E. Cowan, H. Deleuze, C. L. Gibson and D. C.
Sherrington, Chem. Commun., 1998, 2561.
29 M. Kitamura, S. Okada, S. Suga and R. Noyori, J. Am. Chem. Soc.,
1989, 111, 4028.
30 R. Noyori and M. Kitamura, Angew. Chem., Int. Ed. Engl., 1991, 30,
49.
31 G. E. Coates and D. Ridley, J. Chem. Soc. (A), 1966, 1064.
32 P. A. Chaloner and E. Langadianou, Tetrahedron Lett., 1990, 31,
5185.
33 N. Oguni and T. Omi, Tetrahedron Lett., 1984, 25, 2823.
34 A. H. Alberts and H. Wynberg, J. Am. Chem. Soc., 1989, 111, 7264.
35 D. J. Gravert and D. K. Janda, Chem. Rev., 1997, 97, 489.
36 F. Villedon-Denaide, P. Lecavalier and J. M. J. Frechet, Polym. Bull.,
1986, 15, 491.
᎐
(C–O), 820 (2 adjacent Ar-H) and 796 cmϪ1 (3 adjacent Ar-H);
δ (500 MHz) 0.7–0.9 (br, 3H,C-8, CH3) 0.9–1.2 (br, 6H, C-9,
C-10 CH3), 1.2–2.2 (br, 11H, backbone, N-CH3, C-4, CH, C-5
and C-6 CH2), 2.3–2.6 (br, 1H, CHN), 3.3–3.7 (br, 2H,
Ar- CH2), 4.2–4.7 (br, 1H, CHO) and 5.9–7.1 ppm (br, 4H,
Ar-H). Found C, 80.7; H, 8.7; N, 5.1: C20H29NO3 requires C,
80.2; H, 9.8; N, 4.7%. By GPC it had Mn = 3300, Mw = 5700.
Synthesis of polymer 21
Copolymerisation of monomers 19 with an equimolar amount
of styrene using the procedure given above for the synthesis of
polymers 10 gave copolymer 21. It had [α]D Ϫ8.4 (c = 2.3,
CHCl3); νmax 3379 (O–H), 1604 (C᎐C), 1041 (C–O), 820 (2
᎐
adjacent Ar-H) and 796 cmϪ1 (3 adjacent Ar-H), 759 and 701
ppm (5 adjacent Ar-H); δ (500 MHz) 0.7–0.9 (br, 3H, C-8,
CH3), 0.9–1.2 (br, 6H, C-9 and C-10 CH3), 1.2–2.2 (br, 11H,
backbone, N-CH3, C-4, CH, C-5 and C-6, CH2), 2.3–2.6 (br,
1H, CHN), 3.4–3.7 (br, 2H, Ar-CH2), 4.2–4.8 (br, 1H, CH0) and
6.0–7.2 ppm (br, 9H, Ar-H). By elemental analysis it contained
2.86 mmol N per g corresponding to a composition of 67%
monomers 19 and 33% styrene. By GPC it had Mn = 3800,
Mw = 8200.
Synthesis of polymer 22
This polymer was prepared from 1% crosslinked polystyrene
beads (0.70 mmol chlorine per g) and amino alcohol 17 using
the procedure described above for the preparation of polymer
12a. By elemental analysis it contained 0.64 mmol of nitrogen
per g.
37 B. Marx, E. Henry-Basch and P. Freon, C. R. Seances Acad. Sci.,
Ser. C., 1967, 264, 1967.
38 D. C. Sherrington, Chem. Commun., 1998, 2275.
39 N. R. Cameron and D. C. Sherrington, Adv. Pol. Sci., 1996, 126,
163.
40 M. Kitamura, S. Suga, K. Kawai and R. Noyori, J. Am. Chem.
Soc., 1986, 108, 6071.
Acknowledgements
We thank the SERC/EPSRC for financial support and CASE
Awards (for D. W. L. S. and P. W. S.) in collaboration with
Zeneca (Blackley).
41 R. Noyori, S. Suga, K. Kawai, S. Okada, M. Kitamura, N. Oguni,
M. Hayashi, T. Kaneko and Y. Matsuda, J. Organomet. Chem, 1990,
382, 19.
42 R. H. Pickard and J. Kenyon, J. Chem. Soc., 1914, 1115.
43 J. Capillon and J. Guétte, Tetrahedron, 1979, 35, 1817.
44 A. A. Smaardijk and H. Wynberg, J. Org. Chem., 1987, 52, 135.
45 M. P. Bonner and E. R. Thornton, J. Am. Chem. Soc., 1991, 113,
1299.
References
1 R. B. Merrifield, J. Am. Chem. Soc., 1963, 85, 2149.
2 Polymer-supported Reactions in Organic Synthesis, eds. P. Hodge and
D. C. Sherrington, John Wiley, Chichester, 1980.
3 Polymeric Reagents and Catalysts, ed. W. T. Ford, ACS Symposium
Series 308, Washington, 1986.
4 Syntheses and Separations Using Functional Polymers, eds. D. C.
Sherrington and P. Hodge, John Wiley, Chichester, 1988.
5 A. Akelah and A. Moet, Functionalised Polymers and Their
Applications, Chapman and Hall, London, 1990.
46 R. M. Chittenden and G. H. Cooper, J. Chem. Soc (C), 1970, 49.
Paper 9/01455A
1472
J. Chem. Soc., Perkin Trans. 1, 1999, 1463–1472