New ligands for asymmetric diethylzinc additions to aromatic aldehydes, demonstrating substrate-dependent nonlinear effects

Article abstract of DOI:10.1039/b208810j

The development of aziridine-based ligands for the asymmetric addition of diethylzinc to a selection of aromatic aldehydes was discussed. One zinc atom was chelated by the aminoalcohol ligand and associated with the aldehyde oxygen atom. The x-ray analysi

Full text of DOI:10.1039/b208810j

View Article Online / Journal Homepage / Table of Contents for this issue
New ligands for asymmetric diethylzinc additions to aromatic
aldehydes, demonstrating substrate-dependent nonlinear effects
Philip C. Bulman Page,* Steven M. Allin, Suzanne J. Maddocks and Mark R. J. Elsegood
1
Department of Chemistry, Loughborough University, Loughborough, Leicestershire, UK
LE11 3TU. E-mail: p.c.b.page@lboro.ac.uk
Received (in Cambridge, UK) 9th September 2002, Accepted 30th October 2002
First published as an Advance Article on the web 26th November 2002
The development of several aziridine-based ligands for the asymmetric addition of diethylzinc to a selection of
aromatic aldehydes is described. Positive nonlinear effects were observed, and shown to display substrate dependency.
The success of aziridines containing a β-amino alcohol moiety
as effective chiral ligands for the addition of diethylzinc to
aromatic aldehydes is well documented both in solution and in
the solid phase.1 There have been many reports exploring the
substitution of such aziridines at all the R groups indicated in
Fig. 1; Tanner, for example, has produced a range of these
Fig. 3
acid in methanolic solution. Substitution at the nitrogen atom
to give ligands 3a–d was then accomplished using a range of
alkyl halides with potassium carbonate as base (Scheme 1).
Fig. 1
aziridines.2 We wish to report here our investigation into related
ligands with substitution on the nitrogen and the oxygen atoms;
other aminoalcohol ligands substituted at nitrogen and/or
oxygen are also known,3 as are related aminothiol ligands.4
The accepted mechanism for the reaction between diethylzinc
and benzaldehyde in the presence of 1,2-aminoalcohols as
ligands involves a two-zinc species in the diastereofacially
selective step. One zinc atom is chelated by the aminoalcohol
ligand and is also associated with the aldehyde oxygen atom; a
second zinc atom, of a diethylzinc unit, is associated with both
the ligand and aldehyde oxygen atoms; ethyl group transfer
then takes place (Fig. 2).5–7
Scheme 1
Ligand 4 was prepared by O-methylation of 1 with iodo-
methane in the presence of sodium hydride over five days
and was detritylated under the same conditions as 1, to give
Fig. 2
ligand 5.
This model suggests that if the ligand oxygen atom is part of
a hydroxy group, then the ligand will be more successful than if
it is part of an ether moiety.3 However, we believe that an
ether-containing ligand may be used if the aziridine nitrogen
atom is unsubstituted, potentially resulting in reversal of the
stereoselectivity of the reaction through the alternative transi-
tion state shown in Fig. 3. Indeed, reversal of the sense of
induced stereochemistry upon substitution at nitrogen in the
aminoalcohol ligand has previously been observed in these
reactions.7–9 It should also be noted that recent computational
work has pointed to an alternative transition state assembly.10
The aziridine 1 was prepared from -serine by the procedure
of Zwanenburg,11 and was de-tritylated using trifluoroacetic
The N-alkylation reaction used in the preparation of ligands
3 can require several days of heating under reflux in toluene
solution, to produce only moderate yields. For this reason, we
have developed a new synthetic route for 3c, our most
successful ligand, and this is shown in Scheme 2.
Commercial -serine was esterified, and the nitrogen
protected using benzhydryl chloride. This material was then
cyclized to the aziridine using mesyl anhydride and triethyl-
amine in the presence of catalytic DMAP. If mesyl chloride is
used, as in the previous syntheses,6 the aziridine ring suffers
cleavage through nucleophilic attack by chloride to produce a
chlorine-substituted secondary amine. No such ring cleavage
occurs in the trityl substituted material. The methyl ester was
DOI: 10.1039/b208810j
J. Chem. Soc., Perkin Trans. 1, 2002, 2827–2832
This journal is © The Royal Society of Chemistry 2002
2827

View Article Online
Table 1 Diethylzinc addition to substituted benzaldehydes catalysed by ligands 1–5
Ligand
Aldehyde
Conversiona
eeb
Configuration
15
2
Benzaldehyde
Benzaldehyde
p-Fluorobenzaldehyde
p-Anisaldehyde
Benzaldehyde
Benzaldehyde
Benzaldehyde
p-Fluorobenzaldehyde
p-Anisaldehyde
Benzaldehyde
> 99
85
95
62
44c
73
55
49
96
87c
95
77
12
55
R
R
R
R
R
R
R
R
R
R
R
S
2
2
60
90
90
> 99
> 99
> 99
98
3a
3b
3c
3c
3c
3d
4
Benzaldehyde
Benzaldehyde
24
55
5
a Determined by 1H NMR spectroscopy. b Determined by HPLC, Chiracel OD column. c Determined by GC, Chrompak Chirasil CB column.
Fig. 4 Ligand 3c.
Scheme 2
then subjected to a double addition of Grignard reagent to
produce the tertiary alcohol 3c. Lacking any additional
deprotection/reprotection steps, this new route is significantly
more efficient than the standard method.
Ligands 1–5 were tested in a standard addition reaction of
diethylzinc to a selection of aromatic aldehydes, carried out
using 5 mol% of ligand in toluene solution at room temperature
over 24 h (Table 1).11
For ligand 2, where both the nitrogen and the oxygen atoms
are unsubstituted, the R enantiomer of the alcohol is preferen-
tially formed, fitting the expected transition state. Using ligand
4, where both the oxygen and nitrogen atoms are substituted,
the R enantiomer of the alcohol is again formed preferentially,
but in poor ee, perhaps suggesting that neither atom is able to
bind strongly to the zinc. Use of ligand 5, which is substituted
only at the oxygen atom, results in the opposite sense of
enantioselectivity, inducing the S enantiomer of the product
alcohol to be formed preferentially (coincidentally with an
identical ee to the reaction using ligand 3a), suggesting the
adoption of an alternative transition state such as that shown
in Fig. 3.
Fig. 5 Ligand 3d.
be true in the crystalline state (Figs. 4 and 5), †although of course
the conformation of the ligand–metal complexes in solution
may be very different. It can clearly be seen that in structure
3c the phenyl rings of the benzhydryl group are orthogonal.
However, in 3d the equivalent rings are co-planar. The
orthogonal arrangement found in 3c may impart a greater
steric effect in the transition state, resulting in higher observed
enantioselectivity in the reaction.
Ligand 3c shows the greatest versatility of those tested, with
even the electron-deficient p-fluorobenzaldehyde undergoing
the reaction with a very high ee. Ligand 3d was expected to give
a similar result, due to the similarity in the structure; the fact
that this is not the case suggests that the conformation of the
molecule in the transition state must be significantly different.
Single crystal X-ray analysis of these two ligands shows this to
Single crystal X-ray analysis was also carried out on ligands 2
and 4. Ligand 2 contains an intermolecular hydrogen bond
which holds the nitrogen and oxygen atoms in close proximity,
† CCDC reference numbers 198143–198146. See http://www.rsc.org/
suppdata/p1/b2/b208810j/ for crystallographic files in .cif or other elec-
tronic format.
2828
J. Chem. Soc., Perkin Trans. 1, 2002, 2827–2832

View Article Online
It can be seen from Fig. 8 that ligand 1 has the greatest
catalytic activity. Ligand 3c is only a little less active. Ligand 2
provides a significantly lower rate of reaction, with the reaction
not reaching completion within the 24 h test reaction time. In
order to compare results from different ligands meaningfully,
product ees should be measured at the same point of conver-
sion, rather than after the same length of time. In the case of
ligand 2, the ee of the alcohol product was measured after
24 h even though the reaction had not reached completion. In
order to establish whether or not the product ees vary during
the reaction, we monitored the ee of the product in a reaction
using ligand 2 over 50 h (Fig. 9).
Fig. 6 Ligand 2.
perhaps aiding transition state formation (Fig. 6). However, in
ligand 4, with both the oxygen and nitrogen atoms protected,
the oxygen atom is now at the opposite side of the molecule to
the nitrogen atom, so that the conformation in the crystal must
be different from that in the active catalyst (Fig. 7).
Fig. 9 Variation of ee with time, 5 mol% ligand 2.
Fig. 9 shows the change in ee with time for the addition of
diethylzinc to benzaldehyde, in the presence of 5 mol% ligand 2
of 100% ee and 50% ee. For ligand 2 of 100% ee, the ee of the
alcohol product is indeed highest in the initial stages of the
reaction (85% ee after 15 min). After 24 h the product ee has
fallen (to 50% ee), and does not decrease further, showing that
the test reaction time of 24 h is sufficient to indicate the ee of
product at completion under the conditions used.
Many ligand-accelerated reactions display nonlinear effects
in asymmetric catalysis.13 For the addition of diethylzinc to
aromatic aldehydes, a positive nonlinear effect is common. As
highlighted by Blackmond,5 the outcome of nonlinear effects at
the end of the reaction provides information important for
practical applications, although for simplicity of analysis of
reaction kinetics, the initial reaction conditions are usually con-
sidered. An investigation into nonlinear effects displayed by our
ligands for addition to benzaldehyde (Fig. 10) uncovered strong
Fig. 7 Ligand 4.
For these ligands to be useful in asymmetric catalysis, they
must be effective in catalytic quantities. We have compared the
catalytic activities of ligands 2 and 3c at 5 mol% in the addition
of diethylzinc to benzaldehyde with that of ligand 1, which has
previously been reported to be an effective catalyst for this
reaction (Fig. 8).12
Fig. 8 Comparison of ligands 1, 2 and 3c.
Fig. 10 Apparent non-linear effects.
J. Chem. Soc., Perkin Trans. 1, 2002, 2827–2832
2829

View Article Online
Table 2 Variation of ee with reaction time
Ligand 2
Ligand 3c
Time/h
eea
eea
0.5
24
85
50
95
96
a Determined by HPLC, Chiracel OD column.
positive nonlinear effects (ligands 3c and 5) and in one case
(ligand 2) an unexpected apparent negative nonlinear effect.
Table 2 shows that in the presence of ligand 3c the ee is the same
initially as at completion of the reaction, demonstrating a true
positive nonlinear effect in this case. In the case of ligand 2
there may be no true nonlinear effect as a fall in ee of the
product is observed during the reaction, presumably a result of
some form of evolution of the active catalyst species. Indeed,
the 0.5 h samples demonstrate a linear effect (Fig. 9).
Fig. 12 Apparent non-linear effects.
the addition of racemic 1-phenylpropanol (10 mol%) to the
reaction mixture prior to the addition of diethylzinc. Although
the conversion of the reaction was little changed after 24 h,
the change in enantioselectivity was more dramatic. In the
absence of 1-phenylpropanol, the addition of diethylzinc to
p-anisaldehyde proceeds to 60% conversion after 24 h, giving
the product in 73% ee. In the presence of 1-phenylpropanol,
however, after 24 h the reaction proceeds to 54% conversion,
but gives product of only 32% ee. Further investigation showed
that the ee varies throughout the reaction, for example being
at 77% ee after 0.25 h. Product inhibition by newly-formed
1-phenylpropanol may indeed thus be a cause of the variation
in product ee observed during the addition of diethylzinc to
benzaldehyde in the presence of ligand 2.
Fig. 11 shows the remarkable variation in ee of product with
reaction time during the addition of diethylzinc to benzalde-
hyde in the presence of stoicheiometric quantities of ligand 2.
Experimental
Aziridin-2-yldiphenylmethanol 2
N-Tritylaziridin-2-yldiphenylmethanol 1 (16 g, 34 mmol)
was dissolved in a mixture of dichloromethane (80 mL) and
methanol (80 mL). The solution was cooled to 0 ЊC and TFA
(60 mL) was added dropwise. The resulting mixture was con-
centrated, and the residue suspended in saturated aqueous
sodium carbonate (100 mL). Extraction into dichloromethane
(100 mL) was followed by drying over sodium sulfate. Finally
the reaction mixture was evaporated to dryness to give a pale
yellow solid. Trituration in diethyl ether afforded colourless
crystals (6.3 g, 82%), mp 145–147 ЊC. [α]D = Ϫ16.34 (c 0.21,
CH2Cl2); νmax/cmϪ1: 3288, 1671, 1187; δH (250 MHz, CDCl3)
1.72 (d, J = 3.7, 1H); 1.79 (d, J = 6.0, 1H); 2.83–2.87 (m, 1H);
7.19–7.45 (m, 10H); m/z 225.11557 (225.11536 required), 77
(50%); 105 (100%); 183 (60%); 225 (2.5%). Anal Calcd for
C15H15NO: C, 79.9, H, 6.7, N, 6.2. Found: C, 79.8, H, 6.5, N,
6.1%.
Fig. 11 Variation of ee with time, 100 mol% ligand 2.
The 1-phenylpropanol produced after 0.25 h has an ee of 80%,
but, remarkably, of the opposite enantiomer to that preferen-
tially formed in the presence of 5 mol% ligand, where 85% ee
is seen at the beginning of the reaction. The ee drops rapidly
during the stoicheiometric reaction, the product after 24 h
having an ee of 24% of the opposite enantiomer to that
produced at the start of the reaction, and thus of the same
enantiomer preferentially formed in the truly catalytic reaction.
Substrate dependency of nonlinear effects in the addition of
diethylzinc to aldehydes has been recently reported by the
Walsh group,14 which demonstrated that adding an electron-
donating substituent to the aromatic aldehyde causes an
increase in the magnitude of the positive nonlinear effect.
Ligand 2, of varying ee, was tested with several different
aldehyde substrates (Fig. 12). To our surprise, while the effects
induced by ligand 2 in the reaction with benzaldehyde were
enhanced when using p-fluorobenzaldehyde, this behaviour
was replaced by a small positive nonlinear effect with
p-anisaldehyde, consistent with the observations of Walsh.
The cause of the behaviour induced by ligand 2 is currently
unknown, but is perhaps a result of product inhibition by
newly-formed 1-phenylpropanol in the addition of diethylzinc
to benzaldehyde in the presence of ligand 2.6 Unlike the
addition to benzaldehyde, the addition of diethylzinc to
p-anisaldehyde in the presence of ligand 2 (5 mol%) does not
display any change in ee during the course of the reaction under
the conditions used by us. This reaction was repeated, but with
N-Methylaziridin-2-yldiphenylmethanol 3a
Aziridin-2-yldiphenylmethanol 2 (0.5 g, 2.2 mmol) and
potassium carbonate (0.76 g, 5.5 mmol) were suspended in
THF (20 mL). Iodomethane (2.3 mmol, 0.33 g) was added
dropwise, and the mixture stirred at room temperature for 20 h.
Water was introduced into the reaction, followed by extraction
into dichloromethane. The solution was dried over magnesium
sulfate, and evaporated to dryness. The residue was purified by
column chromatography on silica gel (eluent: 4 : 1 petroleum
ether : ethyl acetate) to give colourless crystals (0.15 g, 30%),
mp 88–90 ЊC. [α]D = Ϫ111.84 (c 0.26, CH2Cl2); νmax/cmϪ1: 3424,
1448, 1175; δH (400 MHz, CDCl3) 1.30 (d, J = 6.4, 1H); 1.94
(d, J = 3.6, 1H); 2.25–2.28 (m, 1H); 2.41 (s, 3H); 7.23–7.43
(m, 10H); δC (100 MHz, CDCl3) 31.5, 46.3, 47.3, 74.3, 126.3,
126.6, 127.0, 127.0, 128.0, 128.2, 145.0, 147.9; m/z 239.13101
(239.13101 required), 77 (70%); 105 (100%); 183 (60%); 239
2830
J. Chem. Soc., Perkin Trans. 1, 2002, 2827–2832

View Article Online
(15%). Anal Calcd for C16H17NO: C, 80.3, H, 7.2, N, 5.9.
Found: C, 80.1, H, 7.2, N, 5.8%.
(60 mL), and washed with water (3 × 200 mL) and brine (100
mL). The solution was dried over magnesium sulfate and evap-
orated to dryness. The residue was purified by chromatography
on silica gel (eluent 9 : 1 petroleum ether : ethyl acetate) to give
the product as colourless crystals (0.90 g, 88%), mp 125–128 ЊC.
[α]D = Ϫ142.19 (c 0.37, CH2Cl2); νmax/cmϪ1: 1076, 1447, 1490,
3085, 3057; δH (400 MHz, CDCl3) 1.20 (s, 1H); 1.44 (s, 1H);
2.12–2.13 (m, 1H); 2.99 (s, 3H); 7.15–7.50 (m, 25H); δC (100
MHz, CDCl3) 25.2, 38.2, 52.0, 75.2, 84.2, 127.0, 127.4, 127.5,
127.5, 127.7, 127.8, 128.3, 128.5, 128.7, 128.9, 129.2, 129.7,
130.1, 130.4, 143.5, 143.7, 144.8.
N-Benzylaziridin-2-yldiphenylmethanol 3b
Aziridin-2-yldiphenylmethanol 2 (0.5 g, 2.2 mmol) and
potassium carbonate (5.5 mmol, 0.76 g) were suspended in
THF (20 mL). Benzyl bromide (2.2 mmol, 0.38 g) was added
dropwise and the resulting mixture was stirred for 20 h at room
temperature. Water (100 mL) was introduced, followed by
extraction into dichloromethane. The mixture was then washed
successively with 1 M HCl (10 mL) and water (200 mL), and
dried over magnesium sulfate. Evaporation to dryness produced
a colourless powder requiring no further purification (0.39 g,
56%), mp 88–92 ЊC. [α]D = Ϫ35.43 (c 0.46, CH2Cl2); νmax/cmϪ1:
3424, 1448; δH (250 MHz, CDCl3) 1.55 (d, J = 6.4, 1H); 2.05
(d, J = 3.4, 1H); 2.55–2.59 (m, 1H); 3.45 (d, J = 13.4, 1H); 3.77
(d, J = 13.4, 1H); 7.15–7.42 (m, 15H); δC (100 MHz, CDCl3)
30.5, 46.3, 62.8, 74.1, 126.3, 126.3, 126.8, 127.0, 127.2, 127.3,
127.9, 127.9, 128.0, 128.2, 128.2, 128.3, 128.4, 128.4, 128.5,
138.0, 122.9, 147.4; m/z 315.16231 (315.16231 required), 77
(29%); 91 (100%); 105 (49%); 183 (34%); 260 (16%).
O-Methylaziridin-2-yldiphenylmethanol 5
A solution of 4 (0.5 g, 1.0 mmol) in dichloromethane (5 mL)
and methanol (5 mL) was cooled to 0 ЊC, and TFA (2 mL) was
added. The reaction was allowed to reach room temperature
and stirred for a further 30 min. The reaction was concentrated
in vacuo. The residue was suspended in dichloromethane
(10 mL) and poured into saturated aqueous sodium hydrogen
carbonate. The organic layer was dried over magnesium sulfate
and evaporated to dryness. The residue was purified by chroma-
tography on silica gel (eluent: 3 : 1 petroleum ether : ethyl
N-Benzhydrylaziridin-2-yldiphenylmethanol 3c
acetate) to give the product as a yellow oil (0.19 g, 75%). [α]D
=
Ϫ36.8 (c 0.10, CH2Cl2); νmax/cmϪ1: 3292, 1076; δH (400 MHz,
CDCl3) 1.47 (d, J = 3.6, 1H); 1.64 (d, J = 5.2, 1H); 2.65–2.67 (m,
1H); 3.18 (s, 3H); 7.27–7.45 (m, 10H); δC (100 MHz, CDCl3)
22.4, 36.6, 51.5, 82.6, 127.4, 127.5, 127.6, 127.7, 127.7, 127.8,
128.0, 128.0, 128.4, 128.6, 141.5, 141.7; m/z 239.13123
(239.13101 required), 77 (55%), 105 (70%), 197 (100%), 206
(95%), 239 (> 1%).
Aziridin-2-yldiphenylmethanol 2 (0.5 g, 2.2 mmol) and
potassium carbonate (0.59 g, 5.6 mmol) were suspended in
toluene (15 mL). Benzhydryl chloride (0.68 g, 3.3 mmol) was
added and the reaction heated to reflux for 9 days. The reaction
was quenched with water and extracted into dichloromethane.
The resulting solution was dried over sodium sulfate and
evaporated to dryness in vacuo. The colourless residue was
recrystallized from dichloromethane–hexanes to give the
product as colourless crystals (0.38 g, 44%), mp 192–194 ЊC.
[α]D = Ϫ39.8 (c 0.32, CH2Cl2); νmax/cmϪ1: 3422; δH (250 MHz,
CDCl3) 1.62 (d, J = 6.3, 1H); 2.16 (d, J = 3.7, 1H); 2.76–2.80
(m, 1H); 3.89 (s, 1H); 3.96 (s, 1H); 6.99–7.42 (m, 20H); δC (62.5
MHz, CDCl3) 31.4, 46.9, 74.3, 77.4, 126.4, 126.6, 126.7, 127.3,
127.4, 127.7, 128.0, 128.4, 128.7, 128.8, 142.9, 143.2, 145.7,
147.2; m/z 392.20138 (392.20143 required), 136 (59%); 154
(100%); 392 (16%).
Methyl 2-(N-benzhydrylamino)-3-hydroxypropanoate 6
A suspension of serine methyl ester hydrochloride (5 g, 32
mmol) in chloroform (50 mL) was cooled to 0 ЊC. Triethylamine
(7.7 g, 77 mmol) was added, followed by benzhydryl chloride
(4.4 g, 22 mmol), which was added dropwise over 10 minutes.
The reaction was heated to reflux for 3 days, allowed to cool to
room temperature, quenched with saturated aqueous ammo-
nium chloride solution and extracted into dichloromethane
(50 mL). The solution was dried over magnesium sulfate and
evaporated to dryness. The resulting oil was purified by chroma-
tography on silica gel (eluent 8 : 1 petroleum ether : ethyl acet-
N-(9H-Fluoren-9-yl)aziridin-2-yldiphenylmethanol 3d
Aziridin-2-yldiphenylmethanol 2 (0.5 g, 2.2 mmol) and
potassium carbonate (8.8 mmol, 1.2 g) were suspended in
toluene (20 mL). 9-Bromofluorene (3.3 mmol, 0.81 g) was
added and the reaction was stirred at room temperature for 2
days, then heated to reflux for a further 4 days. The reaction was
quenched with water (10 mL), extracted into dichloromethane
(50 mL) and dried over magnesium sulfate. The solution was
evaporated to dryness and the residue purified by chroma-
tography on silica gel (eluent 9 : 1 petroleum ether : ethyl
acetate), to give the product as yellow crystals (0.2 g, 23%), mp
150–153 ЊC. [α]D = Ϫ61.78 (c 0.23, CH2Cl2); νmax/cmϪ1: 1185,
1449, 3060, 3397; δH (400 MHz, CDCl3) 2.14–2.17 (m, 2H);
2.95–2.99 (m, 1H); 3.89 (s, 1H); 4.15 (s, 1H); 6.42 (d, J = 12,
1H); 6.83–6.89 (m, 1H); 7.24–7.70 (m, 16H); δC (100 MHz,
CDCl3) 28.9, 45.0, 69.8, 75.3, 119.9, 120.3, 125.2, 125.3, 126.4,
126.9, 127.1, 127.5, 127.6, 127.7, 128.5, 128.6, 128.7, 129.1,
141.0, 141.0, 143.1, 143.8, 146.3, 148.3; m/z 389.17846
(389.17796 required), 77 (25%), 105 (50%), 165 (100%), 183
(30%), 389 (4%).
ate) to give the product as a yellow oil (3.08 g, 49%); [α]D
=
Ϫ31.1 (c 0.28, CH2Cl2); νmax/cmϪ1: 3447, 1741; δH (400 MHz,
CDCl3) 3.40–3.43 (m, 1H); 3.64–3.68 (m, 2H); 3.74 (s, 3H); 4.92
(s, 1H); 7.22–7.40 (m, 10H); δC (100 MHz, CDCl3) 52.2, 60.4,
63.1, 65.3, 127.3, 127.4, 127.6, 128.7, 128.7, 142.3, 143.5, 173.7;
m/z 284.12806 (284.12865 required), 77 (50%), 105 (70%), 167
(100%), 182 (90%), 284 (< 1%).
Methyl N-benzhydrylaziridine-2-carboxylate 7
4-(Dimethylamino)pyridine (0.12 g, 1.1 mmol) was added to a
solution of 6 (3 g, 10.5 mmol) in chloroform (200 mL).
Triethylamine (2.54 g, 25.2 mmol) was added and the solution
cooled to 0 ЊC. Methanesulfonic anhydride (2.4 g, 15 mmol) in
chloroform (20 mL) was added over 30 min. The reaction was
maintained at 0 ЊC for 1 h, then allowed to reach room tempera-
ture and heated to reflux for 24 h. The reaction was allowed to
cool to room temperature, quenched with water and extracted
into dichloromethane. The resulting solution was dried over
magnesium sulfate and evaporated to dryness in vacuo. The
resulting oil was purified by chromatography on silica gel (3 : 1
petroleum ether : ethyl acetate with 1% triethylamine) to give
the product as pale yellow crystals (2.04 g, 73%), mp 84–87 ЊC;
[α]D = Ϫ93.3 (c 0.45, CH2Cl2); νmax/cmϪ1: 1744, 1200; δH
(400 MHz, CDCl3) 1.82 (d, J = 6.4, 1H); 2.25–2.27 (m, 1H); 2.30
(d, J = 3.2, 1H); 3.60 (s, 1H); 3.69 (s, 1H); 7.18–7.42 (m, 10H);
δC (100 MHz, CDCl3) 34.9, 38.1, 52.1, 77.9, 126.6, 127.1, 127.3,
O-Methyl-N-tritylaziridin-2-yldiphenylmethanol 4
N-Tritylaziridin-2-yldiphenylmethanol 1 (1 g, 2.2 mmol) and
sodium hydride (60% dispersion in mineral oil, 5.4 mmol, 0.13
g) were dissolved in DMF (10 mL), and stirred at room
temperature for 5 min. Iodomethane (3.4 mmol, 0.46 g) was
added and the reaction stirred for a further 4 days. The reaction
was quenched with water (20 mL), extracted into ethyl acetate
J. Chem. Soc., Perkin Trans. 1, 2002, 2827–2832
2831

View Article Online
6 M. Kitamura, S. Suga, M. Niwa and R. Noyori, J. Am. Chem. Soc.,
1995, 117, 4832.
7 S. Itsuno and J. M. J. Fréchet, J. Org. Chem., 1987, 52, 4142.
8 C.-T. Qian, F.-F. Gao and J. Sun, Tetrahedron: Asymmetry, 2000, 11,
1733.
127.4, 127.5, 127.6, 128.2, 128.4, 128.5, 128.6, 142.4, 171.0;
m/z 266.11832 (266.11810 required), 77 (18%), 105 (22%), 167
(100%), 266 (12%).
9 J. Beliczey, G. Giffels, U. Krägl and C. Wandrey, Tetrahedron:
Asymmetry, 1997, 8, 1529.
Acknowledgements
10 T. Rasmussen and P. M. O. Norrby, J. Am. Chem. Soc., 2001, 123,
2464.
11 J. G. H. Willems, M. C. Hersmis, R. Gelder, J. M. M. Smits,
J. B. Hammink, F. J. Dommerholt, L. Thijis and B. Zwanenburg,
J. Chem. Soc., Perkin Trans. 1, 1997, 963.
12 C. F. Lawrence, S. K. Nayak, L. Thijis and B. Zwanenburg, Synlett,
1999, 1571.
This investigation has enjoyed the support of the EPSRC and
Charnwood Catalysis Ltd.
References
1 L. Pu, Chem. Rev., 2001, 101, 757.
13 (a) H. B. Kagan, J. Am. Chem. Soc., 1986, 108, 2353;
(b) D. Guillaneux, S. Zhao, O. Samuel, D. Rainford and
H. B. Kagan, J. Am. Chem. Soc., 1994, 116, 9430; (c) C. Girard
and H. B. Kagan, Angew. Chem., Int. Ed., 1998, 39, 2922;
(d ) H. B. Kagan, Synlett, 2001, 888.
14 Y. K. Chen, A. M. Costa and P. J. Walsh, J. Am. Chem. Soc., 2001,
123, 5378.
2 (a) D. Tanner, H. T. Korno, D. Guijarro and P. G. Andersson,
Tetrahedron, 1998, 14213; (b) D. Tanner, Angew. Chem., Int. Ed.,
1994, 33, 599.
3 J. Näslund and C. J. Welch, Tetrahedron: Asymmetry, 1991, 2, 1123.
4 J. C. Anderson and M. Harding, Chem. Commun., 1998, 393.
5 T. Rosner, P. J. Sears, W. A. Nugent and D. G. Blackmond, Org.
Lett., 2000, 2511.
2832
J. Chem. Soc., Perkin Trans. 1, 2002, 2827–2832