Enantioselective copper catalysed 1,4-conjugate addition reactions using chiral N-heterocyclic carbenes

Article abstract of DOI:10.1016/j.jorganchem.2005.07.024

The preparation of a variety of chiral N-heterocyclic carbene (NHC) precursors is described. The relative merits of imidazolinium salts and silver carbenes as NHC precursors are discussed with respect to their synthesis, stability and performance in the copper catalysed conjugate addition of dialkyl zinc reagents to a variety of Michael acceptors. Enantioselectivities of up to 93% were achieved using as little as 4% of chiral ligand.

Full text of DOI:10.1016/j.jorganchem.2005.07.024

Journal of Organometallic Chemistry 690 (2005) 5672–5695
www.elsevier.com/locate/jorganchem
Enantioselective copper catalysed 1,4-conjugate addition
reactions using chiral N-heterocyclic carbenes
a,1
b
c,2
,c,2
*
´ ´
Caroline L. Winn , Frederic Guillen , Julien Pytkowicz , Sylvain Roland
,
Pierre Mangeney c,2, Alexandre Alexakis
b,
*
a
School of Chemistry, University of Bristol, Cantock’s Close, Bristol BS8 1TS, UK
b
` `
Universit e´ de Geneve, Departement de Chimie Organique, 30, quai Ernest-Ansermet, CH-1211 Geneve 4, Switzerland
´
Laboratoire de chimie des Organoe´le´ments, UMR 7611, Universite´ P. et M. Curie, 4, place Jussieu, 75252 Paris cedex 05, France
c
Received 29 April 2005; accepted 4 July 2005
Available online 24 August 2005
Abstract
The preparation of a variety of chiral N-heterocyclic carbene (NHC) precursors is described. The relative merits of imidazolinium
salts and silver carbenes as NHC precursors are discussed with respect to their synthesis, stability and performance in the copper
catalysed conjugate addition of dialkyl zinc reagents to a variety of Michael acceptors. Enantioselectivities of up to 93% were
achieved using as little as 4% of chiral ligand.
ꢀ 2005 Elsevier B.V. All rights reserved.
Keywords: Imidazolinium salts; Silver(I); Chiral N-heterocyclic carbenes; Enantioselective; Copper catalysed; Conjugate addition
1. Introduction
chiral and achiral carbenes have been prepared and tested
in a variety of reactions [8]. In some cases, excellent enan-
tiomeric excesses have been achieved [9,10].
During the early 1970s, Lappert et al. [1] reported that
there was a striking similarity between the metal coordi-
nation chemistry of N-heterocyclic carbenes and electron
rich organophosphanes. Initially, their use as ligands was
limited, due to their relatively difficult preparation. The
first applications involved the use of free carbenes,
obtained by deprotonation of the corresponding imidaz-
olinium salt, but these compounds are extremely air and
moisture sensitive. More recent investigations have
shown that it is possible to deprotonate and trap the carb-
enes in situ [2–7]. Since this time, much work has been
carried out using N-heterocyclic carbenes as ligands; both
One such reaction in which N-heterocyclic carbenes
have been applied is the copper catalysed 1,4-addition
of a dialkyl zinc reagent to an a, b-unsaturated ketone
[11–16]. More recently, the copper catalysed c-selective
allylic substitution using Grignard or dialkyl zinc re-
agents has been reported using N-heterocyclic carbenes
as ligands [17,18]. Early work by Alexakis et al. [19]
demonstrated that a trivalent phosphorus ligand could
have a strong accelerating effect on the 1,4-conjugate
addition reaction, and in 2001 Woodward showed [11]
that the addition of diethyl zinc to cyclohexenone could
also be accelerated using the achiral Arduengo diamino-
carbene [20]. Chiral phosphorous-based ligands have
been shown to induce a high degree of enantiomeric ex-
cess in the 1,4-adduct [21,22] and, bearing in mind the
similarity between phosphorous- and heterocyclic car-
bene-based ligands, we hoped to induce enantioselectiv-
ity using chiral heterocyclic carbenes (Scheme 1).
*
Corresponding author. Tel.: +41 22 379 65 22; fax: +41 22 379 32
15/328 73 96.
E-mail addresses: caroline.winn@bristol.ac.uk (C.L. Winn),
sroland@moka.ccr.jussieu.fr (S. Roland), alexandre.alexakis@chiorg.
unige.ch (A. Alexakis).
1
Fax: +44 117 9251295.
Fax: +33 1 44277567.
2
0022-328X/$ - see front matter ꢀ 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2005.07.024

C.L. Winn et al. / Journal of Organometallic Chemistry 690 (2005) 5672–5695
5673
Table 1
O
O
Conjugate addition of Et2Zn on cyclohexenone using Cu(OTf)2 and a
L* (4%), CuX (4%), nBuLi, toluene
Et2Zn (1.5 eq), -20˚C, 2h
variety of novel ligands in toluene for 2 h at ꢀ20 ꢁC
O
O
*
L* (4%), Cu(OTf)2 (4%), nBuLi, toluene
Et2Zn (1.5 eq), -20˚C, 2h
Scheme 1. General method for 1,4-conjugate addition.
Yielda (%)
eeb (%)
2. Results
Entry
Imidazolinium salt
1
1
1
1
2
3
4
5
6
92
99
15
99
97
99
99
41
22 (S)
14 (S)
20 (S)
4 (S)
2c
3d
4
Using conditions based upon those reported by
Woodward [11], preliminary studies showed that it
was possible to achieve an enantiomeric excess of
22% for the addition of diethyl zinc to cyclohexenone
in the presence of Cu(OTf)2 using the diaminocarbene
derived from the Simpkinsꢀ diamine [23,24]. Deprotona-
tion of imidazolinium salt 1 was carried out in situ
5
6
7
8
7 (S)
13 (S)
13 (R)
14 (S)
For representative procedure, see experimental details.
n
using BuLi at ꢀ20 ꢁC in toluene. In order to investi-
a
Determined by GC/MS.
Determined by chiral GC (Lipodex E).
Preformed ligand.
Two equivalents of imidazolinium salt were used.
b
gate the influence of several ligand parameters, such
as the steric bulk of the nitrogen substituents, the pres-
ence of a chiral backbone, of chiral substituents on the
nitrogen atoms, and ring size; a range of novel imidaz-
olinium salts (1–6) were synthesised in good yields by a
reported procedure [25] from the readily available chi-
ral diamines (Scheme 2) [26–29]. Although the presence
of a bulky alkyl or aryl substituent on each nitrogen
was thought to be necessary to stabilise the free diami-
nocarbene [30,31], it was proposed that the in situ trap-
ping of the carbene derived from imidazolinium salt 2
should also give access to the desired copper-carbene
complex.
c
d
lents of the ligand, which usually leads to an improve-
ment in selectivity when using phosphorus based
ligands [32], caused a dramatic decrease in reactivity;
the addition product was obtained in only 15% yield
after 14 h (Entry 3). This could be due to the forma-
tion of an inactive Cu complex in which there is a
2:1 ratio of carbene to metal, and suggests that the
active species for the carbenes is different from that
when using the phosphorus based ligands.
It was observed that ligand 1 gave the best results
(22% enantiomeric excess) under these conditions.
Ligands 2 and 3 give particularly low levels of asym-
metric induction (Entries 4 and 5), presumably because
Firstly, each of the imidazolinium salts was tested
under the original conditions (Scheme 1, Table 1).
Investigations were also made into pre-deprotonation
of the diaminocarbene (Entry 2), but this led to a
decrease in enantiomeric excess. The use of 2 equiva-
R2
R2
N
R1
R1
NH
HC(OEt)3, NH4BF4
1: 95%
2: 88%
3: 96%
4: 80%
5: 79%
6: 84%
-
n
BF4
n
120˚C, 15 min, - 2 h
NH
N
R1
R1
R2
R2
n = 0, 1
Ph
Ph
Ph
Ph
Ph
Ph
Ph
N
N
Ph
N
N
N
N
Cy
Cy
-
-
-
BF4
BF4
BF4
2
1
3
Ph
N
N
Ph
N
N
Ph
N
N
Ph
-
-
-
BF4
BF4
BF4
4
5
6
Scheme 2. Synthesis of imidazolinium salts.

5674
C.L. Winn et al. / Journal of Organometallic Chemistry 690 (2005) 5672–5695
the chiral centres are relatively distant from the react-
ing centre. The diastereomeric pair of ligands 4 and 5
gave equal but opposite levels of stereochemical induc-
tion (Entries 6 and 7) as anticipated, despite the
change in ring size which presumably does not effect
the conformation of the chiral centres in the ligand.
Compound 6 however, which had been reported by
Hartwig to give high levels of enantiomeric excess
when used as a ligand for an aryl amination reactions
[28], gave disappointingly low levels of stereochemical
induction (14%, Entry 8). It would appear that this
species is a more effective ligand for Pd-catalysed reac-
tions that for Cu. The low yield observed can be
attributed to the hygroscopic nature of the imidazolin-
ium salt.
(maximum ee = 11%) [12]. A range of solvents was also
screened, and a significant rate increase was observed
when the reaction was carried out in diethyl ether (Table
2). The reaction took only 15 min to reach completion at
ꢀ20 ꢁC, as opposed to 2 h in toluene. This allowed us to
reduce the temperature of the reaction; at ꢀ50 ꢁC, the
reaction took 2 h to reach completion, but there was no
real increase in enantiomeric excess. However, on
decreasing the temperature to ꢀ78 ꢁC, the enantioselec-
tivity improved to 50% using ligand 1.
Under these low temperature conditions, a range of
cyclic and acyclic enones was tested to probe the versa-
tility of the reaction (Table 3). Comparable results were
obtained for the acyclic substrate decenone (11, Entry
5), but in all other cases, the enantiomeric excess of
the 1,4-adduct did not exceed 26% [12]. Indeed, under
these conditions, 5-methyl-3-hexen-2-one (9) and benzy-
lacetone (10) did not react at all.
Under these new, low temperature conditions, fur-
ther copper salts were also screened (Table 4). The
choice of salts was based on related work in the group
(conjugate addition reactions using phosphorus-based
ligands), which had shown that enantioselectivities
could be enhanced by the use of both copper(I) and
copper(II) carboxylates [33]. Although copper(II) salts
are usually cheaper and easier to handle, it has been
proposed that the mechanism involves reduction of
the Cu(II) species by Et2Zn to give the true catalytic
entity; Cu(I). In this work, CuTC (copper thiophene
In order to optimise the conditions, several Cu(I) min-
eral salts were tested; CuI, CuCN, CuBF4 Æ 4CH3CN and
CuPF6 Æ 4CH3CN were used but on the whole, the yields
were moderate and the enantioselectivities were poor
Table 2
Influence of the reaction temperature in diethyl ether using ligand 12
O
O
1 (4%), Cu salt (4%), nBuLi, Et2O
Et2Zn (1.5 eq), -78 C, 16 h
˚
Entry
Temperature (ꢁC)
Time (h)
Yielda (%)
eeb (%)
1
2
3
ꢀ20
ꢀ50
ꢀ78
0.25
2
16
97
97
91
19 (S)
20 (S)
50 (S)
carboxylate;
a
Cu(I) salt) [34] and anhydrous
Cu(OAc)2 were tested. However, for this particular li-
gand, no improvements in enantiomeric excess were
observed.
a
Determined by GC/MS.
Determined by chiral GC (Lipodex E).
b
Table 3
Reaction using variety of substrates with ligand 1 under low temperature conditions
O
O
O
O
Ph
7
8
9
10
O
O
11
12
Entry
Substrate
Yielda (%)
eeb (%)
1
2
3
4
5
6
a
7
8
91
66
0
0
99
99
50 (S)
12 (S)
–
9
10
11
12
–
51 (+)
26 (R)
Determined by GC/MS.
Determined by chiral GC (Lipodex E).
b

C.L. Winn et al. / Journal of Organometallic Chemistry 690 (2005) 5672–5695
5675
Table 4
Ph
Ph
Ph
Ph
Influence of Cu salt using ligand 1 under low temperature conditions
NH4BF4, MeCN
O
O
Ph
N
N
Ph
Ph
N
N
Ph
X
-
X-
BF4
NH4Cl, MeCN
ligand (4%), Cu salt (4%), nBuLi
13 X = Cl or I
15
*
Et2Zn (1.5 eq)
Scheme 3. Anion exchange.
Entry
Cu salt (4 mol%)
Yielda (%)
eeb (%)
1
2
3
Cu(OTf)2
Cu(OAc)2
CuTC
91
78
88
50 (S)
42 (S)
29 (S)
Initially, the iodo-imidazolinium salts were synthes-
ised [35], and the corresponding tetraflouroborate com-
pounds 15 and 16 were prepared by anion exchange
(Scheme 3) using NH4BF4 in acetonitrile overnight
[36,37]. The results of 1,4-addition experiments (see
Table 5) showed that in fact the ‘‘anti-Simpkins’’ ligand,
15, yielded a 1,4-adduct with an enantiomeric excess of
only 8% (although this time the product had an R con-
figuration, which corresponded to all 4 stereogenic cen-
tres in the ligand), and therefore was, itself, the
mismatched ligand (with respect to ligand 1). The meso
ligand performed slightly better, inducing an enantiose-
lectivity of 21% with an R configuration (which in this
case correlates with the configuration of the chiral sub-
stituents on the nitrogen atoms).
Based on the results in hand, a new range of imidaz-
olinium salts were synthesised as carbene precursors in
an attempt to probe the electronic as well as the steric
effects of the ligand on the selectivity of the reaction.
Following the work of Herrmann et al. [38,39], the
unsaturated ligands 17–20 were prepared in a one pot
reaction starting from 2 equivalents of the correspond-
ing chiral amine (Scheme 4). It should be noted that
despite the unsaturation, the ligands show little p aro-
maticity In bonding terms, they can be better described
as diaza-allyl systems [8].
a
Determined by GC/MS.
Determined by chiral GC (Lipodex E).
b
In each case, it was noted that the absolute stereo-
chemistry of the product was related to the chiral centres
of the ligand employed. For example, when ligand 4 was
used, with chiral centre of S configuration adjacent to
the nitrogen atom, the absolute stereochemistry of the
product was also S, while the diastereomeric ligand with
R stereocentres (ligand 5) gave the product with the
opposite but equal degree of enantioselectivity. Likewise
for ligands 2 and 3, an S configuration in the ligand in-
duced an S configuration in the product. It was surpris-
ing to observe, therefore, that for ligand 1, it was not the
stereogenic centres adjacent to the nitrogen ((R) config-
uration), which dictated the absolute stereochemistry of
the product, but the stereocentres in the backbone of the
ligand (which were of (S) configuration). This led us to
believe that, despite achieving promising levels of
enantioselectivity under these conditions, ligand 1 was
actually the mismatched case of a diastereomeric pair
of ligands. With this in mind, the diastereomeric com-
pound, 15; derived from the anti Simpkinsꢀ diamine,
was prepared. A by-product of its formation was isomer
16, which had meso stereochemistry in the backbone,
and this was also tested.
Initial tests with ligand 17, derived from (R)-phenyl-
ethylamine, showed promising results; under the low
temperature conditions it was possible to achieve enan-
tiomeric excesses of up to 54% using anhydrous
17 : 60%
NH2
Ar
N
N
Ar
HBF4, toluene
18 : 48%
19 : 100%
20 : 53%
O
(CH2O)n
-
O
BF4
Ar
R
R
R
N
R or S
Ph
N
N
N
N
Ph
Ph
N
Ph
-
-
18
17
BF4
BF4
Me
Me
Me
Et
R
Et
R
1-naphthyl
1-naphthyl
2-naphthyl
N
N
2-naphthyl
-
-
BF4
BF4
19
20
Me
Scheme 4. Synthesis of unsaturated imidazolinium salts.

5676
C.L. Winn et al. / Journal of Organometallic Chemistry 690 (2005) 5672–5695
Cu(OAc)2 (Entry 5). In order to observe the effect of
unsaturation in the ring, the reaction was also carried
out using ligand 17 in toluene at ꢀ20 ꢁC with Cu(OTf)2.
An increase in selectivity from 13% to 39% was observed
(Entries 1 and 2, ligands 4 and 17, respectively, Table 6)
and there was also a decrease in yield. This was probably
due to the increased hygroscopicity of the unsaturated
imidazolinium salt 17 with respect to its saturated coun-
terpart 4; nBuLi deprotonation to form the reactive car-
bene species may have been incomplete for ligand 17 (see
Table 6).
From screening the initial set of ligands, it appeared
to be important to have chiral substituents on the N-
atom (which would be closer to the metal centre than
any stereocentres present in the backbone of the ligand),
and efforts were made to vary the steric bulk of the
nitrogen substituents. It was interesting to observe that
on replacing the methyl group with an ethyl (leaving
the phenyl in place), exactly the same degree of enanti-
oselectivity was observed for each of Cu salts tested
(compare Entries 3–5 with Entries 6–8). The difference
in absolute stereochemistry of the product was due to
the configuration of the starting amine, but again, it
was possible to predict the stereochemical outcome of
the reaction; starting with chiral amine of S configura-
tion, it was possible to obtain the 1,4-adduct with an S
configuration, and vice versa.
On changing the size of the aromatic group on the
N-substituent from a phenyl to a naphthyl (both the
1- and 2-naphthyl analogues were synthesised in order
to probe the steric effects/requirements of each), there
was no increase in the enantiomeric excess observed
in the 1,4-adduct. In fact, when the 1-naphthyl substi-
tuted ligand 19 was employed (Entries 9–11) there was
a considerable decrease in enantiomeric induction. The
difference in the results obtained when the two different
ligands (19 and 20) were used can be explained by their
different steric requirements. This effect had previously
been reported by Seebach who had investigated the 1-
and 2-naphthyl derivatives of the TADDOL auxiliary
[40]. In some cases, it was observed that even the ste-
reochemical course of a reaction could be reversed by
the use of the other naphthyl isomer [41]. Evidence
for this was also found in crystal structures of the
two TADDOL compounds which showed that the 2-
naphthyl is orientated in the same direction as a simple
phenyl ring, whereas the 1-naphthyl is more sterically
demanding. In some cases this was responsible for
reducing the catalytic activity of the ligand during a
reaction [42].
Table 5
Influence of ligand counter-anion under low temperature conditions
Ph
Ph
Ph
Ph
Ph
N
N
Ph
Ph
N
N
Ph
X-
X-
13 X = I
14 X = I
15 X = BF4
16 X = BF4
Entry
Ligand
Cu salt
Yielda (%)
eeb (%)
1
2
3
a
1
15
16
Cu(OTf)2
Cu(OTf)2
Cu(OTf)2
91
100
55
50 (S)
8 (R)
21 (R)
Determined by GC/MS.
Determined by chiral GC (Lipodex E).
b
Table 6
1,4-addition using a range of unsaturated ligands with different Cu salts
O
O
ligand (4%), Cu salt (4%), nBuLi
*
Et2Zn (1.5 eq)
Entry
Ligand
Cu salt
Tempertaure (ꢁC)
Solvent
Time (h)
Yielda (%)
eeb (%)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
a
4
17
17
17
17
18
18
18
19
19
19
20
20
20
Cu(OTf)2
Cu(OTf)2
Cu(OTf)2
CuTC
Cu(OAc)2
Cu(OTf)2
CuTC
Cu(OAc)2
Cu(OTf)2
CuTC
Cu(OAc)2
Cu(OTf)2
CuTC
ꢀ20
ꢀ20
ꢀ78
ꢀ78
ꢀ78
ꢀ78
ꢀ78
ꢀ78
ꢀ78
ꢀ78
ꢀ78
ꢀ78
ꢀ78
ꢀ78
Toluene
Toluene
Et2O
Et2O
Et2O
Et2O
Et2O
Et2O
Et2O
Et2O
Et2O
Et2O
Et2O
Et2O
2
2
99
60
80
88
75
38
46
51
44
76
57
87
81
94
13 (S)
39 (R)
42 (R)
52 (R)
54 (R)
42 (S)
52 (S)
54 (S)
27 (S)
38 (S)
2 (S)
16
16
16
16
16
16
16
16
16
16
16
16
37 (S)
51 (S)
46 (S)
Cu(OAc)2
Determined by GC/MS.
Determined by chiral GC (Lipodex E).
b

C.L. Winn et al. / Journal of Organometallic Chemistry 690 (2005) 5672–5695
5677
Table 7
Although no crystal structures were obtained of the
imidazolinium salts (in each case, the salt was isolated
as a fine granular powder), the results obtained in the
conjugate addition reactions suggest that it is the aro-
matic group which is closer to the reacting metal centre
than the alkyl, R, group (changing from Me to Et has no
effect on the reactions selectivity, whereas changing to
the more sterically bulky naphthyl group we see a
change, albeit a reduction in enantiomeric excess).
Roland et al. [14], have shown that it is possible to
carry out the same reaction with low levels of asymmet-
ric induction using a chiral silver(I) diaminocarbene 21
as the carbene precursor (Scheme 5).
The original investigation by Roland focused on the
influence of the reaction temperature and solvent effects.
In each case, the substrate employed was cyclohexenone
and the ligand was compound 21. In contrast to the
work carried out by Alexakis et al. [12], in which 4
mol% of ligand and Cu salt were used, these results were
obtained using only 2 mol% of each.
It was observed that the rate of reaction using a silver
carbene at 0 ꢁC was vastly increased with respect to the
analogous reactions using the imidazolinium salts at
ꢀ20 ꢁC (Table 7, entries 10–13). Using imidazolinium
salt 1 as a ligand, the reaction in toluene took 2 h to reach
completion at ꢀ20 ꢁC (Entry 10), and it is thought that if
the reaction temperature had been decreased to ꢀ78 ꢁC,
the reaction may have taken several days! We observed
another difference for the reaction in dichloromethane.
When the imidazolinium salt was employed, the deproto-
nation step was carried out in a small volume of toluene
(to prevent solvent deprotonation), and then the mixture
was diluted with dichloromethane (to the extent that it
was the major solvent). The reaction rate appeared to
be very similar to that observed in toluene, and the reac-
tions were both complete in 2 h with a comparable degree
of enantioselectivity (Entry 10 and Entry 12). However,
for the silver carbenes, we can see that the rate of reac-
tion is slower in dichloromethane than for toluene, and
there is a complete loss of selectivity (Entry 5). The re-
sults in THF are comparable however; using imidazolin-
ium salt 1, the reaction proceeds with a low yield and
negligible enantioselectivity (Entry 13). For silver car-
bene 21, the same effect is observed; the reaction rate is
much lower than for the other solvents, and again, the
product obtained is racemic (Entry 6).
Influence of solvent and temperature on 1,4-addition reaction using
silver carbene 21 as ligand
O
O
ligand 21 (2 mol%), Cu(OTf)2, (2 mol%)
*
Et2Zn (1.5 equiv.)
Entry
Solvent
Temperature
(ꢁC)
Time
(h)
Yielda
(%)
eeb
(%)
1
2c
3
4
5
6
7
8
9
Toluene
Toluene
Hexane
Et2O
DCM
THF
Toluene
Toluene
Toluene
Toluene
Et2O
0
0
0
0
0
0.25
0.25
0.25
0.25
1
1
1
1
1
98
98
99
95
62
28
99
98
78
92
97
90
24
23 (S)
20 (S)
16 (S)
15 (S)
0
0
0
ꢀ20
ꢀ40
ꢀ78
ꢀ20
ꢀ20
ꢀ20
ꢀ20
16 (S)
10 (S)
4 (S)
22 (S)
19 (S)
27 (S)
3 (S)
10d
11d
12d
13d
a
2
0.25
2
Toluene/DCM
THF
2
Determined by GC/MS.
Determined by chiral GC (Lipodex E).
Reaction carried out using 2 equivalents of copper salt with respect
b
c
to the carbene (i.e., 4 mol% copper species, 2 mol% ligand).
d
Reaction carried out with 4 mol% ligand 1 and 4 mol% copper
species.
also decreased (Entries 7–9), but this effect could not be
tested for imidazolinium salts due to their low rate of
reaction in toluene. Another interesting observation
was that the use of 2 equivalents of the silver carbene
with respect to the copper salt (Entry 2, 4 mol% in total)
led to very similar results as had been observed with the
usual 1:1 ratio of carbene to copper salt; the yield was
almost quantitative after 15 min and there was only a
small decrease in enantioselectivity. This is in stark con-
trast with the results obtained for the chiral imidazolin-
ium salts in which there was a large decrease in reactivity
of the system when two equivalents of ligand were used
(Table 1, Entry 3).
The first silver carbene complexes in the literature
were prepared in 1993 by Arduengo et al. [43], and in-
volved reaction of the free carbene with silver triflate.
In this case however, the silver carbene 21 was prepared
by an alternative method, described by Wang and Lin
[44], in which the imidazolinium salt was treated with
Ag2O [45]. Silver carbene species are known to act as
effective carbene transfer agents for the synthesis of pal-
It was surprising to observe that as the reaction tem-
perature was decreased to ꢀ78 ꢁC, the enantioselectivity
O
O
tBu
tBu
ligand, Cu(OTf)2
21
Me
N
N
*
Me
Et2Zn (1.5 equiv.)
AgI
Scheme 5. 1,4-Conjugte addition using silver carbenes as ligands.

5678
C.L. Winn et al. / Journal of Organometallic Chemistry 690 (2005) 5672–5695
ladium or gold carbene complexes, and this has recently
been applied to the synthesis of palladium–carbene
complexes which have been used successfully as cata-
lysts for C–C coupling reactions [46]. There are several
advantages to using the diaminocarbene in this form,
despite the synthetic route being one step longer than
that for the corresponding imidazolinium salts. The sil-
ver diaminocarbene has been shown to act as an effi-
cient transfer agent to copper(II) triflate, and the
catalytic species could be easily prepared at room tem-
perature in various solvent systems. Unlike the imidaz-
olinium salts, the silver carbenes do not require in situ
deprotonation, therefore, the use of strong bases and
polar solvents is avoided, and perhaps more impor-
tantly, in terms of reaction yield, and in contrast to
the imidazolinium salts, the silver carbenes are not gen-
erally hygroscopic.
methane using MeI to yield imidazolinium salt 24. The
desired product was isolated by precipitation with
Et2O and pentane in 95% yield. Using the Wang and
Lin procedure [44], the silver(I) diaminocarbene was
prepared in excellent yield by treatment with 0.5 equiv-
alents of Ag2O in dichloromethane (Scheme 6).
A crystal structure of the silver carbene 21 was ob-
tained (displayed in Fig. 1 as an ORTEP diagram) which
clearly shows that the compound exists in the solid state
as a dimeric species in which the silver atoms are bridged
by two iodine atoms.
Our attention then turned to the synthesis and appli-
cation of a wider range of silver carbenes, initially this
involved a comparison of the enantiomeric excesses ob-
tained using the unsaturated imidazolinium salts as li-
gands with the corresponding silver carbenes. In order
to prepare the desired, unsaturated, silver carbenes, it
was necessary to prepare the chloro-imidazolinium ana-
logues of the tetra-fluoroborate salts (17–20) described
above. Unfortunately, it was not possible to transform
the BF4 salts directly into silver carbenes and it was only
possible to prepare the chloro-imidazolinium salts in
very low yield by the same one-pot procedure (17% for
The imidazolinium salt precursor to the silver carbene
21 was prepared from the known (R,R)-4,5-di-tert-buty-
limidazolidine (23), which itself was prepared from the
aminal 22 [47], by a one-pot palladium-mediated
hydrogenolysis and oxidation procedure [48]. Com-
pound 23 was readily alkylated at 20 ꢁC in dichloro-
tBu
tBu
tBu
tBu
tBu
tBu
tBu
tBu
10% Pd(OH)2/C
HCO2NH4, EtOH,
reflux, 6h
MeI, K2CO3, DCM
20˚C, 95%
Ag2O, DCM
20˚C, 95%
Ph
N
N
Ph
N
NH
N
N
N
N
Me
Me
Me
Me
I-
95%
22
23
24
21
AgI
Scheme 6. Synthesis of silver carbene 21.
Fig. 1. ORTEP diagram of ligand 21.

C.L. Winn et al. / Journal of Organometallic Chemistry 690 (2005) 5672–5695
5679
compound 26) [38,39]. The main by-product of the reac-
tion was the imine, 25 formed by addition of para-form-
aldehyde onto the chiral amine (Scheme 7). In addition,
attempts to carry out counter anion exchange from BF4ꢀ
to Clꢀ proved to be fruitless; it appears that this reaction
can only be carried out in the opposite direction.
It was, however, possible to prepare the compounds
using a modified, two step procedure via the intermedi-
ate bis-imines (27 and 28) which were formed in excel-
lent yield from the reaction between 2 equivalents of a
chiral amine with glyoxal under acid mediated condi-
tions (Scheme 8) [49].
The bisimines (27 and 28) required no purification
and were taken straight through to the corresponding
chloro-imidazolinium salts by reaction with para-form-
aldehyde and HCl (32% aqueous solution) in toluene.
In some cases, trituration with diethyl ether was re-
quired to purify the product, which was then trans-
formed into the unsaturated silver carbene (31–33) by
stirring overnight in anhydrous dichloromethane in the
presence of 0.5 equivalents of silver(I) oxide. The silver
carbenes were isolated in good yield and with a high le-
vel of purity after filtration through celite. No further
purification was required.
Using silver carbene 31 as a ligand for the 1,4-addi-
tion of diethyl zinc onto cyclohexenone under the condi-
tions previously described by Roland and co-workers
[14], an enantiomeric excess of 25% was observed (Entry
1, Table 8). However, using the modified, low-tempera-
ture conditions for the same ligand, eeꢀs of up to 62%
could be achieved (Entry 3). The highest excess was ob-
tained using CuTC as the Cu source, although
Cu(OAc)2 also led to an excess of 59% (Entry 4). In con-
trast to the results obtained with the hygroscopic BF4
imidazolinium salts, the conversion was always
quantitative.
Even more pleasing, was the observation that the
reaction appeared to be applicable to other substrates,
and particularly high enantiomeric excesses were
achieved when cycloheptenone was employed. Using
both CuTC and Cu(OAc)2, the product obtained had
an enantiomeric excess greater than 85% (Entries 6
and 7). As for the six-membered cyclic enone however,
Cu(OTf)2 gave lower results (43 and 47% for the 6 and
H
H
25
O
NH2
HCl (32%)
O
O
Ph
N
Toluene, 45˚C, 16h
O
H
H
n
Scheme 7. Formation of imine.
H
H
n
NH2
MgSO4, Formic acid
DCM, rt, 16h
HCl (32%),
Toluene, 45˚C, 16h
Ar
N
N
Ar
O
O
Ar
R
R
R
27: Ar = Ph, R = Et Yield = 100%
28: Ar =1-Naphthyl, R = Me Yield = 96%
O
H
H
X
n
HCl (32%),
Toluene, 40˚C, 8h
Ag2O, DCM, rt, 16h
Ar
N
N
Ar
Ar
N
N
Ar
Cl-
AgCl
R
R
R
R
26: Ar = Ph, R = Me Yield = 17%
29: Ar = Ph, R = Et Yield = 45%
30: Ar =1-Naphthyl, R = Me Yield = 72%
31: Ar = Ph, R = Me Yield = 95%
32 Ar = Ph, R = Et Yield = 87%
33: Ar =1-Naphthyl, R = Me Yield = 82%
Scheme 8. Synthesis of unsaturated silver carbenes.

5680
C.L. Winn et al. / Journal of Organometallic Chemistry 690 (2005) 5672–5695
Table 8
1,4-addition using ligand 31 with a range of substrates and Cu salts
O
O
N
Ligand 31 (4 mol%),
Cu salt (4 mol%), Et2O
N
O
O
-78˚C, 16 h, Et2Zn
34
Entry
Substrate
Cu salt
Yielda (%)
eeb (%)
1c
2
3
4
5
6
7
8
7
7
Cu(OTf)2
Cu(OTf)2
CuTC
Cu(OAc)2
Cu(OTf)2
CuTC
Cu(OAc)2
Cu(OTf)2
CuTC
Cu(OAc)2
CuTC
92
100
99
100
71
100
99
92
100
100
44
60
67
25 (R)
37 (R)
62 (R)
59 (R)
47 (R)
86 (R)
88 (R)
70 (R)
72 (R)
75 (R)
49
7
7
8
8
8
34
34
34
9
9
10
11
12
13
10
12
CuTC
CuTC
42 (S)
42 (S)
7 = cyclohexenone, 8 = cycloheptenone, 9 = nitrostyrene.
a
Determined by GC/MS.
Determined by chiral GC (Lipodex E).
Using original Roland conditions [14]; 2 mol% ligand and Cu salt, toluene, rt, 1 h.
b
c
7 membered rings respectively, Entries 2 and 5). High
selectivities had also been observed using phosphorus
based ligands with nitrostyrene 34 [50], and with ligand
31, good results were observed for addition to this sub-
strate, even with Cu(OTf)2 which appeared to be as effi-
cient as the other Cu salts in this case. All three different
copper salts gave enantiomeric excesses over 70% (En-
tries 8–10). With the acyclic enones (tested only with
CuTC, Entries 11–13), enantiomeric excesses were con-
siderably higher than those achieved with the imidazo-
linium salt, 1, although for each substrate, it was
observed that the reaction did not go to completion at
ꢀ78 ꢁC, even after 16h of reaction. Presumably, this is
due to the lower reactivity of these substrates with re-
spect to their cyclic counterparts. However, it should
be noted that all the product yields were much higher
than those obtained with the saturated ligands (compare
with results in Table 3) and this is assumed to be due to
the electronic rich nature of the ligands, which renders
the ligand–Cu system more reactive.
As with the imidazolinium salts, investigations were
made into changing the size of the substituents at the
stereogenic centre adjacent to the nitrogen (Table 9).
Firstly, the methyl group was substituted for an ethyl
moiety and the resulting ligand (32) was tested in the
reaction with cyclohexenone, cycloheptenone and nitro-
styrene (Entries 1–9). Unlike the corresponding imidaz-
olinium salt (18) which gave exactly the same results for
the addition to cyclohexenone (see Table 6), the results
obtained in this case were slightly improved for addition
in the presence of Cu(OTf)2 (but only for the cyclic en-
ones, Entries 1 and 4) and decreased in all other cases.
The 1-naphthyl derivative (33) was also prepared and
in contrast to the results observed with the correspond-
ing tetraflouroborate-imidazolinium salt for the addi-
tion to cyclohexenone, the enantioselectivities were
Table 9
1,4-addition using unsaturated silver carbene ligands 32 and 33 with
range of substrates and Cu salts
O
O
ligand (2 mol%), Cu salt (2 mol%)
Et2Zn (1.5 equiv.), Et2O,-78 C, 16 h.
˚
Entry
Ligand
Substrate
Cu salt
Yielda (%)
eeb (%)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
a
32
32
32
32
32
32
32
32
32
33
33
33
33
33
33
33
33
33
7
7
Cu(OTf)2
CuTC
Cu(OAc)2
Cu(OTf)2
CuTC
Cu(OAc)2
Cu(OTf)2
CuTC
Cu(OAc)2
Cu(OTf)2
CuTC
Cu(OAc)2
Cu(OTf)2
CuTC
Cu(OAc)2
Cu(OTf)2
CuTC
85
94
92
84
90
97
91
80
98
58
79
87
60
95
95
25
98
96
51 (S)
52 (S)
55 (S)
69 (S)
73 (S)
76 (S)
38 (S)
52 (S)
55 (S)
48 (S)
54 (S)
59 (S)
75 (S)
89 (S)
93 (S)
11 (S)
69 (S)
67 (S)
7
8
8
8
34
34
34
7
7
7
8
8
8
34
34
34
Cu(OAc)2
Determined by GC/MS.
Determined by chiral GC (Lipodex E).
b

C.L. Winn et al. / Journal of Organometallic Chemistry 690 (2005) 5672–5695
5681
approximately equivalent. For the addition to cyclohep-
tenone (Entries 13–15), the results were in excess of
those achieved for the six-membered ring, and also in
excess of those obtained with the phenyl-derived ligand.
Again, Cu(OAc)2 proved to be the copper salt of choice,
yielding a product with an enantiomeric excess of 93%.
For almost every substrate-ligand combination,
Cu(OAc)2 gave the best enantiomeric excess (Entries 3,
6, 9, 12 and 15).
After probing the steric effects of the groups at the
chiral centres adjacent to the nitrogen, investigations
were made into the effect of having a chiral chelating
group adjacent to the nitrogen. To investigate chelation
effects, the O-protected enantiopure amino-alcohol 35
derived from 2-hydroxy-cyclohexylamine was employed
as the starting material for the synthesis of ligand 38
(Scheme 9).
to the air resulted in absorption of H2O, which caused
the product to become sticky and very difficult to han-
dle. Fortunately, the silver carbene was much easier to
manipulate and could be tested as a ligand under the
usual conditions (Table 10).
Surprisingly, the results obtained with cyclohexenone
were very poor, less than 10% enantiomeric excess was
measured in each case, and quite astonishingly, with this
in mind, up to 73% enantiomeric excess was observed
for cycloheptenone using CuTC. These results are cur-
rently inexplicable and the varying yields are also some-
what surprising. It may be that this particular silver
carbene is hygroscopic and the Cu(OTf)2 contained a
higher percentage of moisture that either the Cu(OAc)2
or CuTC, thus quenching the carbene as it formed and
resulting in a lower yield.
In order to probe the effect of having a sterically bulky
chiral group adjacent to the nitrogen, the unsaturated
analogue of Hartwigsꢀ imidiazolinium salt (42) derived
from a-pinene 39 was prepared (Scheme 10) [29].
In this case, attempted synthesis of the bisimine 40 by
the acid catalysed route described above (Conditions A)
[49] was unsuccessful and led to a complex mixture of
Again, a two-step procedure via the bisimine 36 [45]
was employed to prepare the chloro-imidazolinium salt
37, which was converted to the silver carbene 38 in good
yield. In this particular case, some problems were
encountered due to the highly hygroscopic nature of
the chloro-imidazolinium salt. Exposure, even briefly,
O
H2N
O
(CH2O)n, toluene
HCl, 99%
N
N
MgSO4, Formic acid
BnO
DCM, 100%
OBn BnO
36
35
N
N
N
N
Cl-
BnO
Ag2O, DCM
86%
OBn
AgCl
OBn BnO
37
38
Scheme 9. Synthesis of chelating ligand 38.
Table 10
1,4-addition reaction using ligand 38 with a range of Cu salts and substrates
O
O
N
N
38 (4 mol%), Cu salt (4 mol%)
AgCl
BnO
38
Et2Zn (1.5 equiv.), Et2O,-78˚C, 16 h.
OBn
Entry
Substrate
Cu salt
Yielda (%)
eeb (%)
1
2
3
4
5
6
7
8
9
a
7
7
Cu(OTf)2
CuTC
Cu(OAc)2
Cu(OTf)2
CuTC
Cu(OAc)2
Cu(OTf)2
CuTC
55
89
60
15
64
60
22
70
75
5 (S)
5 (S)
9 (S)
54 (S)
64 (S)
73 (S)
6 (S)
53 (S)
31 (S)
7
8
8
8
34
34
34
Cu(OAc)2
Determined by GC/MS.
b
Determined by chiral GC (Lipodex E).

5682
C.L. Winn et al. / Journal of Organometallic Chemistry 690 (2005) 5672–5695
O
O
Conditions A) MgSO4, formic acid,
NH2
N
N
DCM, 16h
(CH2O)n, toluene, HCl
20%
Conditions B) DCM, 6h
40 Yield: Conditions A 34%
39
Conditions B 100%
N
N
N
N
Ag2O, DCM
83%
Cl-
AgCl
41
Scheme 10. Synthesis of sterically bulky ligand 42.
42
products, from which the desired bisimine could not be
separated. Instead, it was possible to form compound 40
in high yield by simply mixing the chiral amine and gly-
oxal in dichloromethane for 5 h (Conditions B) [29].
After this time, the product was isolated by extraction
and used without further purification.
was prepared in a stepwise manner by coupling the
homochiral amine onto dibromobenzene 43 using
Pd2dba3 in the presence of ( ) BINAP and NaOt-Bu.
The amine-coupled product was obtained in moderate
yield to give the desired diamine, which could be cyclized
to form the chloro-imidazolinium salt 47 using triethyl-
orthoformate under acidic conditions. The resulting imi-
dazolinium salt was very hygroscopic but could be dried
over P2O5 before forming the silver carbene 48, which
was isolated in excellent yield (Scheme 11).
Unfortunately, the enantiomeric excesses observed
when using this ligand in 1,4-addition reactions with
cyclohexenone and cycloheptenone, were very low (Ta-
ble 12), but it was interesting to see that for the first
time, the excess achieved for cyclohexenone was supe-
rior to that for the seven membered cyclic substrate,
and, perhaps more surprisingly, the absolute configura-
tion of the cyclohexenone 1,4-adduct (S) was opposite to
the configuration of the starting amine (R). This may be
due to strong intramolecular p-stacking interactions be-
tween the aromatic backbone and the phenyl groups
present in the nitrogen substituents, which twist the li-
gand into an unusual conformation.
Formation of the silver carbene 42 via the chloro-imi-
dazolinium salt 41 proceeded in a satisfactory yield over
the two steps, and the product was then tested in a 1,4-
conjugate addition reaction using CuTC as the catalyst
(Table 11); employing both cyclohexenone (Entry 1)
and cycloheptenone (Entry 2) as the substrate. Unfortu-
nately, in this case, although the reaction yield was high,
the levels of stereochemical induction observed were
very low, only 5% for cyclohexenone and 15% for cyclo-
heptenone. We may conclude from these results that the
ligand is too sterically hindered to induce high levels of
enantiomeric excess. These enantiomeric excesses are the
only ones that are lower than for the corresponding BF4
imidazolinium salt although these results are not ob-
tained under the same conditions. It is thought that un-
der the low temperature conditions, the enantiomeric
excess would be even higher for the BF4 salt, although
silver carbenes were much more efficient (95% yield as
opposed to 22%).
Moving away from the unsaturated ligands, diamino-
carbenes related to ligand 21, containing chiral tertiary
butyl groups in the backbone were synthesised and
tested (Fig. 2), both under the conditions originally de-
scribed by Roland [14] and the low temperature, ethereal
conditions described above [12].
Another ligand, 48, in which the imidazolinium back-
bone was a phenyl ring, was synthesised according to the
method described by Diver et al. [36]. The diamine 45
Compounds 49, 51, 52, and 54, were prepared from
the chiral aminal 22 via compound 23, and the alkylated
imidazolinium salts 55–58 (see Scheme 12).
Table 11
1,4-addition reaction using ligand 42
O
O
42 (4 mol%), CuTC (4 mol%)
In order to preserve the chiral moieties on the nitro-
gen atoms, a different approach was used to prepare li-
gand 53 [51]. Using iodine and NaHCO3 in
dichloromethane, it was possible to oxidise compound
23 to give the imidazolinium salt 55 (Scheme 13) [48].
This reaction proceeded quantitatively (by TLC) but
the yield of the pure product was somewhat reduced
(58%) due to difficulties in isolation. Due to the low
Et2Zn (1.5 equiv.), Et2O,-78 ˚C, 16 h.
Entry
Substrate
Cu salt
Yielda (%)
eeb (%)
1
2
a
7
8
CuTC
CuTC
95
85
4 (S)
15 (S)
Determined by GC/MS.
Determined by chiral GC (Lipodex E).
b

C.L. Winn et al. / Journal of Organometallic Chemistry 690 (2005) 5672–5695
5683
NH2
NH2
Pd2dba3.CHCl3
rac-BINAP, tol
NaOtBu, 135˚C, 2h
Pd2dba3.CHCl3
rac-BINAP, tol
NaOtBu, 95˚C, 48h
Br
Br
Br
HN
Ph
Ph
NH
HN
Ph
H
HN
Ph
43
55%
43%
10%
44
45
46
Ag2O, DCM
95%
HCl, HC(OEt)3
89%
Ph
N
N
Ph
Ph
N
N
Ph
Cl-
47
Scheme 11. Synthesis of ligand 48.
48
AgCl
Table 12
ligand 21, or Entries 3 and 4 using ligand 49 where the
greatest difference is observed, or Entries 6 and 7 using
ligand 51).
1,4-addition reaction using ligand 48
O
O
Ag carbene (x mol%), Cu salt (x mol%)
Contrary to expectation (since there were no chiral
substituents on the nitrogen atom), the silver carbene
49 with N-benzyl substituents gave 58% ee for cyclohex-
enone (Entry 3 using CuTC as the copper source), and
76% ee for the cycloheptenone adduct (Entry 14). The
high degree of stereochemical induction in this case
may be rationalised with the aid of the crystal structures
shown in Figs. 1 and 3 [45]. The t-Bu groups in the back-
bone are anti to one another and although in Fig. 1, we
see that the arrangement of substituents around each
nitrogen atom is planar, the methyl groups themselves
are not large enough to be affected by the back-bone
substituents, so also lie in a planar arrangement. How-
ever, when a proton is replaced by a phenyl group (see
Fig. 3), which is much larger than the proton, the phenyl
group is displaced by the t-Bu groups in an anti fashion
(one up, one down with respect to the plane), and it is
this conformation which leads to increased enantio-
differentiation.
Et2Zn (1.5 equiv.), various conditions
Entry
Substrate
Cu salt
Yielda (%)
eeb (%)
1
2
a
7
8
CuTC
CuTC
100
95
21 (S)
7 (S)
Determined by GC/MS.
Determined by chiral GC (Lipodex E).
b
solubility of this compound and its very high melting
1
point (m.p. > 290 ꢁC), although H NMR data was ob-
tained, complementary analysis could not be performed
[52]. It was however, possible to prepare the correspond-
ing silver carbene in reasonable yield.
From the results in Table 13, many conclusions can
be drawn. Firstly, for a specific ligand, low temperature
conditions using diethyl ether and CuTC yield products
with higher enantiomeric excesses than the toluene/
Cu(OTf)2 conditions (compare Entries 1 and 2 using
tBu
tBu
tBu
N
tBu
tBu
N
tBu
N
N
N
N
49
50
53
21
-
Ph
Ph
AgCl
BF4
AgCl
Ph
Ph
tBu
N
tBu
tBu
N
tBu
tBu
N
tBu
N
N
Ph
N
Ph
51
52
AgCl
AgCl
AgCl
Naphthyl
Naphthyl
m-OMePh
m-OMePh
tBu
tBu
N
N
54
p-MePh
p-MePh
AgCl
Fig. 2. Silver carbenes tested in 1,4-conjugate addition.

5684
C.L. Winn et al. / Journal of Organometallic Chemistry 690 (2005) 5672–5695
tBu
tBu
tBu
tBu
tBu
tBu
10% Pd(OH)2/C
HCO2NH4, EtOH,
reflux, 6h
RX, K2CO3, DCM
20˚C Yield
Ph
N
N
Ph
N
NH
N
N
R
R
95%
X-
23
22
55 R = Bn, X = Br Yield = 92%
56 R = naphthyl, X = Cl Yield = 71%
57 R = m-OMeC6H4CH2, X = Cl Yield = 94%
58 R = p-tolCH2, X = Cl Yield = 58%
tBu
tBu
49 R = Bn, X = Br
Ag2O, DCM
51 R = naphthyl, X = Cl
53 R = m-OMeC6H4CH2, X = Cl
54 R = p-tolCH2, X = Cl
N
N
Yield 88-100%
R
R
Ag
X
Scheme 12. Synthesis of silver carbenes 49, 51, 53, and 54.
In contrast to compound 21, ligand 49 crystallised as
a monomer, but shows the same carbene-Ag-X unit. The
anti arrangement of the benzyl groups with respect to
the t-Bu groups is shown clearly, and it is assumed (as
described above) that in this case, the N atoms are no
longer planar.
For this particular ligand, it was possible to make a
direct comparison between the silver carbene 49 and
the corresponding imidazolinium salt 50 which had also
been tested (Table 13, entry 5). Under exactly the same
experimental conditions, a slight decrease in selectivity
was observed for the imidazolinium salt; 52% enantio-
tBu
tBu
tBu
tBu
tBu
tBu
NaHCO3, I2, DCM
20˚C, 58%
Ag2O, DCM
Yield 80 %
Ph
N
N
Ph
Ph
N
N
Ph
Ph
N
N
Ph
I-
22
55
AgI
53
Scheme 13. Synthesis of silver carbene 53.
Table 13
1,4-additions with a range of t-Bu substituted ligands under various conditions
O
O
Ag carbene (x mol%), Cu salt (x mol%)
Et2Zn (1.5 equiv.), various conditions
Entry
Ligand (mol%)
Cu salt (mol%)
Substrate
Solvent
Tempertaure (ꢁC)
Time (h)
Yielda (%)
eeb (%)
1
2
3
4
5
6
7
8
9
10
11
12
a
21 (4)
21 (2)
49 (4)
49 (2)
50 (4)
51 (4)
51 (2)
52 (4)
53 (4)
54 (2)
49 (4)
52 (4)
CuTC (4)
Cu(OTf)2 (2)
CuTC (4)
Cu(OTf)2 (2)
CuTC (4)
CuTC (4)
Cu(OTf)2 (2)
CuTC (4)
CuTC (4)
Cu(OTf)2 (2)
CuTC (4)
CuTC (4)
7
7
7
7
7
7
7
7
7
7
8
8
Et2O
Toluene
Et2O
Toluene
Et2O
Et2O
Toluene
Et2O
Et2O
Toluene
Et2O
Et2O
ꢀ78
16
1
16
1
16
16
1
16
16
1
99
98
100
100
95
99
98
99
98
30 (S)
23 (S)
58 (S)
31 (S)
52 (S)
14 (R)
0
69 (S)
29 (S)
23 (S)
76 (S)
88 (S)
0
ꢀ78
rt
ꢀ78
ꢀ78
rt
ꢀ78
ꢀ78
rt
95
99
100
ꢀ78
16
16
ꢀ78
Determined by GC/MS.
Determined by chiral GC (Lipodex E).
b

C.L. Winn et al. / Journal of Organometallic Chemistry 690 (2005) 5672–5695
5685
Fig. 3. ORTEP diagram of ligand 49.
meric excess was measured (compared to 58% for the sil-
ver carbene). A difference in yield can be explained by
the stability of the silver carbene with respect to the imi-
dazolinium salt, but to explain the difference in enantio-
meric excess between what should be exactly the same
reactive species (the free carbene), perhaps we need to
examine more closely the effects of the counteranions
and/or the presence of other salts in the reaction mixture
(Ag+ for the silver carbene compared to Li+ from the
BuLi which is used to deprotonated the imidazolinium
salt).
tiomeric excess of 69% (Entry 8), while for addition
to cycloheptenone, the product was found to have
an enantiomeric excess of 85% (Entry 15). At present,
it is unclear whether the increased enantioselectivity is
due to the chelating nature of the OMe groups, or
whether it is due to their electron donating capacity.
Conversely, when the p-Me analogue 54 was tested,
there was a slight decrease in enantioselectivity with
respect to the parent, Bn substituted, compound (49)
under comparable conditions (compare Entries 4 and
10).
The analogous 1-naphthyl substituted silver carbene
51 gave a much lower enantiomeric excess for the addi-
tion to cyclohexenone under both sets of conditions,
with a racemic product being observed in toluene with
Cu(OTf)2 at room temperature (Entry 7), and only
14% under the low temperature conditions (Entry 6).
Presumably in this case, the N-substituents were too
sterically demanding to induce high selectivity. The
difference in sterical requirements of a simple phenyl
and 1-naphthyl group may also account for the change
in absolute configuration of the product which was
observed in this case.
Perhaps more surprisingly, the ligand containing chi-
ral substituents a to the nitrogen (53), also gave a rela-
tively poor enantiomeric excess for the addition to
cyclohexenone; only 29% using CuTC (Entry 9), lower
even that for the simple N-Me analogue, 21. It would
seem that the chiral N-substituents are, in fact, detri-
mental to the level of enantioselective induction, espe-
cially when comparing this result with the N-benzyl
ligand 49. Interestingly for ligand 53, as for the imidaz-
olinium salt derived from the Simpkinsꢀ diamine (1), it
was observed that the absolute stereochemistry of the
adduct was of an S configuration which again, appears
to be controlled from the stereocentres in the backbone
of the carbene rather than the chiral groups adjacent to
the nitrogen (which would be expected to give rise to a
product with R stereochemistry).
Employment of the m-OMe substituted analogue 52
led to a further increase in the enantiomeric excess.
For the addition of Et2Zn to cyclohexenone, the ad-
duct could be obtained in a 95% yield with an enan-

5686
C.L. Winn et al. / Journal of Organometallic Chemistry 690 (2005) 5672–5695
Although the experimental conditions were not al-
Microanalytical department. Optical rotations were
measured with a Perkin–Elmer 241 polarimeter, using
a cell of 1 dm path length. The concentration c is ex-
pressed in g/100 mL. Melting points (m.p.) were deter-
mined using a Reichert hot plate melting point
apparatus and are uncorrected. Solvents were purified
by standard techniques [53], and reactions in non-aque-
ous media were carried out under dry nitrogen or dry ar-
gon unless otherwise stated. All other reagents were used
as received.
ways comparable, it was observed that the enantioselec-
tivities obtained using ligands with t-Bu groups in the
backbone were greater than for their Ph analogues.
For example, if we compare ligand 1 with ligand 50, un-
der comparable experimental conditions i.e., using
CuTC in Et2O at ꢀ78 ꢁC for 16 h, the 1,4-adduct of
diethyl zinc to cyclohexenone has an enantiomeric ex-
cess of 29% when ligand 1 is used (ligand 1 has 4 chiral
centres including chiral substituents on the nitrogen
atoms), but ligand 50 (containing only 2 stereogenic cen-
tres in the backbone) yields a product with 52% ee. This
is thought to be due to the steric bulk of the t-Bu groups,
which displace the N-substituents to a greater, and more
favourable extent (in terms of increased levels of enanti-
oselectivity), than the Ph groups.
4.2. Typical procedure for conjugate addition using
imidazolinium salt in toluene at ꢀ20 ꢁC
A dry Schlenk tube was charged with Cu(OTf)2
(0.04 mmol) and imidazolinium salt (0.04 mmol) and
toluene (3 mL) was added. The resulting solution was
cooled to ꢀ78 ꢁC, upon which n-BuLi (0.05 mmol as
a 2M solution in hexanes) was added. After stirring
for 30 min at low temperature, the reaction mixture
was warmed up to room temperature and stirred for
a further 15 min to ensure complete deprotonation.
After this time, the solution was cooled back to ꢀ20
ꢁC and diethyl zinc was added (1.5 mmol as a 1M solu-
tion in hexane). The mixture was stirred for 15 min
before addition of the substrate (1 mmol, either neat,
or as a solution in toluene). After stirring at ꢀ20 ꢁC
for 2 h, the reaction was quenched at low temperature
by the addition of HCl (2 mL of a 1M aqueous solu-
tion). The resulting mixture was stirred until clear
and the enantiomeric excesses could be measured
directly using chiral GC (capillary column – Lipodex
E, 0.2 km, 50 m, 0.25 mm). The organic phases could
be isolated, dried over MgSO4 and the solvent concen-
trated in vacuo to yield the crude product, which could
be purified by flash column chromatography to provide
the pure product.
3. Conclusions
Optimisation of conditions and ligand modification
has led to good to excellent levels of enantiomeric
excesses in the enantioselective copper catalysed 1,4-
conjugate addition reaction using chiral diaminocarb-
enes as ligands. They allowed us to achieve some of
the highest enantioselectivities (up to 93%) reported to
date for this type of ligand. It should be pointed out that
the synthesis of these chiral NHC is very easy and that
they are less sensitive and much cheaper than phospho-
rus-based ligands. Further work will concentrate on
studying the effects of different aryl substitution patterns
with various electron-donating/withdrawing, chelating
or sterically bulky nitrogen substituents in an attempt
to develop an understanding of the mechanism and
therefore be able to rationally design ligands which will
lead to improved selectivities for a broad range of Mi-
chael acceptors.
4.3. Typical procedure for conjugate addition using silver
carbene in Et2O at ꢀ78 ꢁC
4. Experimental
4.1. General remarks
A dry Schlenk tube was charged with copper salt
(0.04 mmol) and imidazolinium salt (0.04 mmol) and
Et2O (3 mL) was added. After stirring for 15 min at
room temperature, the solution was cooled to ꢀ20
ꢁC and diethyl zinc was added (1.5 mmol as a 1M
solution in hexane). The mixture was stirred for 15
min the reaction mixture was cooled to ꢀ78 ꢁC for
addition of the substrate (1 mmol, either neat, or as
a solution in toluene). After stirring at ꢀ78 ꢁC for
16 h, the reaction was quenched at low temperature
by the addition of HCl (2 mL of a 1M aqueous solu-
tion). The resulting mixture was stirred until clear and
the enantiomeric excesses could be measured directly
using chiral GC (capillary column – Lipodex E, 0.2
1H and proton-decoupled 13C NMR spectra were re-
corded with Brucker DRX-400, Brucker AV-300 and
Varian Gemini 200 spectrometers at ambient tempera-
ture. Chemical shifts are quoted in ppm relative to tetra-
methylsilane (d = 0) and were referenced to the residual
protons in the solvent indicated. Chemical shifts are
quoted relative to tetramethyl silane (d = 0). Mass spec-
tra were recorded at the University of Geneva, using
electrospray techniques with DCM as the carrier sol-
´
vent, or at the Universite Pierre et Marie Curie, Paris
VI, using chemical ionisation (CI) techniques with
NH3 as the carrier gas. Microanalyses were carried out
´
by the staff of the Universite Pierre et Marie Curie

C.L. Winn et al. / Journal of Organometallic Chemistry 690 (2005) 5672–5695
5687
km, 50 m, 0.25 mm). The organic phases could be iso-
lated, dried over MgSO4 and the solvent concentrated
in vacuo to yield the crude product, which could be
purified by flash column chromatography to provide
the pure product.
(m, 8H, NCHPh), 3.41 (tt, J = 11.7 Hz, J = 3.5 Hz,
2H, NCH), 2.10–2.00 (m, 2H, CyH), 1.95–1.75 (m, 8H,
CyH), 1.65–1.55 (m, 2H, CyH), 1.35–1.10 (m, 8H, CyH).
13C NMR (100.6 MHz, CDCl3): 155.0 (NCNH), 136.7,
130.0, 126.3, 72.2, 57.7, 32.0, 30.9, 25.0, 24.9, 24.5.
20
½aꢁD ꢀ 319 (c 0.50, CHCl3).
4.4. Synthesis of imidazolines 1–6 from diamines
4.4.4. 1,3-Bis-(1(S)-phenyl-ethyl)-4,5-dihydro-3H-
imidazol-1-ium tetraflouroborate (4)
Into a flask equipped with a magnetic stirrer, was
introduced the diamine (1 equiv.), and ammonium tetra-
fluoroborate (1 equiv.). A distillation cooling head was
fitted with a collection head for the second stage of the
reaction and the system purged with nitrogen. Triethyl-
orthoformate (2 equiv.) was then added via syringe and
the resulting mixture was heated at 120 ꢁC for up to 2 h.
After this time, the system was placed under high vac-
uum and the volatile by-products and excess starting
materials were collected. The resulting products were
purified as described below.
The resulting oil was triturated with ethyl acetate and
the resulting white solid collected by filtration. After
recrystallisation with an ethanol-water mixture, the de-
sired product was obtained in 80% yield as a white crys-
talline solid with melting point = 118 ꢁC.
1H NMR (200 MHz, CDCl3): 8.13 (s, 1H, NCHN),
7.50–7.30 (m, 10H, ArH), 4.80 (q, J = 7.0 Hz, 2H,
NCHMe), 3.70–3.55 (m, 4H, NCH2CH2N), 1.67 (d,
J = 7.0 Hz, 6H, CH3).
13C NMR (50 MHz, CDCl3): 155.6 (NCNH), 139.2,
130.1, 129.8, 128.0, 59.0, 47.7, 19.9.
20
½aꢁD þ 23 (c 0.51, CHCl3).
4.4.1. (4S,5S)-Diphenyl-1,3-bis-(1(R)-phenyl-ethyl)-
4,5-dihydro-3H-imidazolinium tetraflouroborate (1)
The crude product was purified by recrystallisation
with an ethanol-ethyl acetate mixture to give the desired
compound in 95% yield as a white crystalline product.
1H NMR (400 MHz, CDCl3): 9.18 (s, 1H, NCHN),
7.40–7.11 (m, 16H, ArH), 7.00–6.90 (m, 4H, ArH),
4.45 (s, 2H, PhCHN), 4.39 (q, J = 6.9 Hz, 2H,
PhCHMeN), 2.01 (d, J = 6.9 Hz, 6H, CH3).
4.4.5. 1,3-Bis-(1(R)-phenyl-ethyl)-3,4,5,6-tetrahydro-
pyrimidin-1-ium tetraflouroborate (5)
The resulting orange oil was triturated with dichloro-
methane and the resulting solid collected by filtration.
After recrystallisation with ethyl acetate, the desired
product was obtained in 79% yield as a white crystalline
solid.
1H NMR (400 MHz, CDCl3): 8.50 (s, 1H, NCHN),
7.43–7.34 (m, 10H, ArH), 5.09 (q, J = 6.8 Hz, 2H,
NCHMe), 3.16 (m, 4H, NCH2), 1.87 (t, J = 5.8 Hz,
2H, CH2), 1.75 (d, J = 6.9 Hz, 6H, CH3).
13C NMR (100.6 MHz, CDCl3): 153.0 (NCNH),
137.4, 135.4, 130.3, 129.8, 129.4, 129.0, 127.0, 126.9,
72.6, 57.9, 20.1.
20
13C NMR (100 MHz, CDCl3): 151.0 (NCNH), 136.9,
129.2, 128.9, 127.0, 63.4, 40.0, 18.9, 17.5.
½aꢁD ꢀ 174 (c 0.54, CHCl3).
20
½aꢁD þ 40 (c 0.48, CHCl3).
4.4.2. 1,3-Dimethyl-(4S,5S)-diphenyl-4,5-dihydro-3H-
imidazol-1-ium tetrafluoroborate (2)
4.4.6. N,N0-Diisopinocamphenylimidazolinium
tetrafluoroborate (6) [12]
The crude product was recrystallised from ethanol to
give the desired product as a white crystalline solid in
84% yield. All spectroscopic data was in agreement with
published data.
The pale yellow crude product was recrystallised with
an ethyl acetate-chloroform mixture to give the desired
product in 88% yield as a white crystalline product with
melting point = 223–224 ꢁC.
1H NMR (200 MHz, CDCl3): 8.16 (s, 1H, NCHN),
7.55–7.45 (m, 6H, ArH), 7.40–7.30 (m, 4H, ArH), 5.00
(s, 2H, NCHPh), 2.97 (s, 6H, CH3).
1H NMR (400 MHz, CDCl3): 8.40 (s, 1H, NCHN),
4.40 (m, 2H), 4.03 (m, 4H), 2.50 (m, 4H), 2.05 (m,
4H), 1.90 (m, 4H), 1.28 (s, 6H, C(CH3)2), 1.17 (d,
J = 7.1 Hz, 6H, CH3), 1.10 (s, 6H, C(CH3)2), 0.8 (d,
J = 9.6 Hz, 2H).
13C NMR (50 MHz, CDCl3): 159.6 (NCNH), 135.8,
131.0, 130.6, 128.8, 75.8, 33.8.
20
½aꢁD þ 253 (c 0.59, CHCl3).
4.4.3. 1,3-Dicyclohexyl-(4S,5S)-diphenyl-4,5-dihydro-
3H-imidazol-1-ium tetrafluoroborate (3)
4.4.7. (4R,5R)-Diphenyl-1,3-bis-(1(R)-phenyl-ethyl)-
4,5-dihydro-3H-imidazolinium iodide (13)
1H NMR (400 MHz, CDCl3): 10.64 (s, 1H, NCHN),
7.44–7.28 (m, 16H, ArH), 6.92–6.90 (m, 4H, ArH), 5.55
(q, J = 7.2 Hz, 2H, NCH(CH3)Ph), 4.47 (s, 2H,
NCHPh), 1.45 (d, J = 7.2 Hz, 6H, CH3).
The pale yellow crude product was recrystallised with
ethyl acetate to give the desired compound as a white
crystalline solid in 96% yield.
1H NMR (400 MHz, CDCl3): 8.78 (s, 1H, NCHN),
7.55–7.45 (m, 6H, ArH), 7.30–7.20 (m, 4H, ArH), 4.89

5688
C.L. Winn et al. / Journal of Organometallic Chemistry 690 (2005) 5672–5695
13C NMR (100 MHz, CDCl3): 157.6, 137.3, 136.9,
4.6. General method for preparation of unsaturated
tetrafluoroborate-imidazolinium salts (ligands 17–20) [19]
130.0, 129.7, 129.3, 129.2, 127.7, 126.8, 72.1, 58.5, 19.0.
20
½aꢁD þ 207.7 (c 0.32, CDCl3).
A 50 mL flask was charged with (R)- or (S)-chiral
amine (10 mmol) in toluene. Under vigorous stirring
and slight cooling, was added paraformaldehyde (10
mmol). After 30 min, a second equivalent of the amine
was added at 0 ꢁC. Under constant cooling, HBF4 (10
mmol as a 32% aqueous solution) was added dropwise.
After stirring for 15 min, the ice bath was removed and
glyoxal was slowly added (10 mmol as a 40% aqueous
solution). The resulting mixture was stirred for 30 min
at room temperature, then for 12 h at 35–40 ꢁC. After
this time, the mixture was cooled to room temperature
diluted with diethyl ether (10 mL) and a saturated solu-
tion of Na2CO3 (5 mL). The phases were then separated
and the aqueous layer washed with dichloromethane
(2 · 10 mL). The combined organic layers were washed
with brine (25 mL), dried over MgSO4 then concentrated
under reduced pressure. The crude product in each case
was isolated as a viscous brown/orange oil which could
be purified by trituration using diethyl ether.
4.4.8. (4S,5R)-Diphenyl-1,3-bis-(1(R)-phenyl-ethyl)-
4,5-dihydro-3H-imidazolinium iodide (14)
1H NMR (400 MHz, CDCl3): 10.21 (s, 1H, NCHN),
7.55–6.79 (m, 20H, ArH), 5.93 (q, J = 6.8 Hz, 1H,
NCH(CH3)Ph), 5.30 (d, J = 11.6 Hz, 1H, NCHPh),
5.08 (d, J = 12.0 Hz, 1H, NCHPh), 4.47 (q, J = 6.8
Hz, 1H, NCH(CH3)Ph), 2.07 (d, J = 6.8 Hz, 3H,
CH3), 1.53 (d, J = 7.2 Hz, 3H, CH3).
13C NMR (100 MHz, CDCl3): 129.5, 129.4, 129.2,
129.1, 128.7, 128.5, 128.0, 127.0, 77.2, 69.6, 58.5. Car-
bene carbon not observed.
20
½aꢁD þ 72.4 (c 0.21, CDCl3).
4.5. General method for anion exchange (ligands 15 and
16) [18]
In a round-bottomed flask equipped with a magnetic
stirrer bar and rubber septum, were placed iodo-imidaz-
olinium salt (1 equiv., 0.15 mmol) and MeCN (5 mL).
To the resulting suspension was added ammonium tetra-
fluoroborate (1 equiv., 0.15 mmol). The resulting mix-
ture was stirred at room temperature for 16 h before
concentrating under reduced pressure. The crude prod-
uct was triturated with chloroform, from which the yel-
low mother liquor was recovered. Further concentration
under reduced pressure gave the desired compounds
without need for further purification.
4.6.1. 1,3-Bis-(1(R)-phenyl-ethyl)-imidazolium
tetraflouroborate (17)
The desired product was obtained as a hygroscopic
yellow powder in 60% yield.
1H NMR (400 MHz, CDCl3): 9.26 (s, 1H, NCHN),
7.39–7.37 (m, 10H, ArH), 7.13 (s, 2H, NCH@C), 5.75
(q, J = 7.0 Hz, 2H, NCH(CH3)Ph), 1.96 (d, J = 7.0
Hz, 6H, CH3).
13C NMR (100 MHz, CDCl3): 137.8, 133.7, 129.0,
126.9, 121.1, 60.1, 20.6.
4.5.1. (4R,5R)-Diphenyl-1,3-bis-(1(R)-phenyl-ethyl)-
4,5-dihydro-3H-imidazolinium tetraflouroborate (15)
Isolated as a pale yellow solid in 87% yield.
20
½aꢁD þ 18.5 (c 0.96, CHCl3).
MS (Electrospray/DCM) calculated for C19H21N2
277.1704, found 277.3.
1H NMR (400 MHz, CDCl3): 10.37 (s, 1H, NCHN),
7.39–7.25 (m, 16H, ArH), 6.91–6.89 (m, 4H, ArH), 5.48
(q, J = 7.0 Hz, 2H, NCH(CH3)Ph), 4.48 (s, 2H,
NCHPh), 1.43 (d, J = 7.1 Hz, 6H, CH3).
4.6.2. 1,3-Bis-(1(S)-phenyl-propyl)-imidazolium
tetraflouroborate (18)
The desired product was obtained as a hygroscopic
pale yellow powder in 48% yield.
13C NMR (100 MHz, CDCl3): 157.0, 137.2, 136.6,
129.9, 129.6, 129.3, 129.2, 127.7, 126.8, 58.4, 19.0.
20
½aꢁD þ 210.9 (c 1.11, CDCl3).
1H NMR (400 MHz, CDCl3): 9.54 (s, 1H, NCHN),
7.45–7.38 (m, 10H, ArH), 7.17 (s, 2H, NCH@C), 5.51
(t, J = 7.9 Hz, 2H, NCH(CH2CH3)Ph), 2.35 (qd,
J = 13.6 Hz, J = 7.4 Hz, 4H, CH2CH3), 0.94 (t,
J = 7.3 Hz, 6H, CH3).
4.5.2. (4S,5R)-Diphenyl-1,3-bis-(1(R)-phenyl-ethyl)-
4,5-dihydro-3H-imidazolinium tetraflouroborate (16)
Isolated as a white solid in 100% yield.
1H NMR (400 MHz, CDCl3): 9.83 (s, 1H, NCHN),
8.00–6.80 (m, 20H, ArH), 5.69 (q, J = 7.0 Hz, 1H,
NCH(CH3)Ph), 5.43 (d, J = 12.1 Hz, 1H, NCHPh),
5.15 (d, J = 12.1 Hz, 1H, NCHPh), 4.49 (q, J = 7.0
Hz, 1H, NCH(CH3)Ph), 2.04 (d, J = 7.0 Hz, 3H, CH3)
1.58 (d, J = 7.1 Hz, 3H, CH3).
13C NMR (100 MHz, CDCl3): 136.4, 134.9, 129.5,
127.5, 120.6, 66.3, 27.5, 10.6.
20
½aꢁD ꢀ 34.2 (c 1.13, CHCl3).
MS (Electrospray/DCM) calculated for C21H25N2
305.2018, found 305.3.
13C NMR (100 MHz, CDCl3): 156.9, 137.3, 130.2,
129.5, 129.4, 129.3, 129.2, 128.7, 128.5, 128.0, 126.9,
69.5, 67.9, 58.4, 58.2, 21.1, 19.2.
4.6.3. 1,3-Bis-(1(S)-1-naphthyl-ethyl)-imidazolium
tetraflouroborate (19)
The desired product was obtained as a hygroscopic
pale orange powder in 100% yield.
20
½aꢁD þ 49.4 (c 2.63, CDCl3).

C.L. Winn et al. / Journal of Organometallic Chemistry 690 (2005) 5672–5695
5689
1H NMR (400 MHz, CDCl3): 9.42 (s, 1H, NCHN),
8.15–7.19 (m, 14H, ArH), 6.89 (s, 2H, NCH@C), 6.56
(q, J = 6.9 Hz, 2H, NCH(CH3)Naphth), 2.12 (d,
J = 6.9 Hz, 6H, CH3).
1H NMR (400 MHz, CDCl3): 7.05 (s, 1H, NCHN),
4.80 (s, 1H, NH), 3.27 (s, 2H, CHt-Bu), 0.86 (s, 18H,
C(CH3)3).
13C NMR (100 MHz, CDCl3): 150.7, 69.1, 33.2,
24.5.
13C NMR (100 MHz, CDCl3): 139.4, 133.9, 132.6,
130.4, 129.2, 128.2, 127.8, 126.5, 126.1, 125.5, 124.5,
123.3.
4.6.7. (4R,5R)-1,3-Dimethyl-4,5-di-tert-butyl-
imidazolinium iodide (24)
20
½aꢁD þ 91.7 (c 0.98, CHCl3).
MS (Electrospray/DCM) calculated for C27H25N2
377.2018, found 377.3.
To a solution of imidazolidine 23 (1 mmol) in DCM (5
mL) were added K2CO3 (0.5 g) and MeI (2.1 mol). The
resulting solution was stirred for 12 h at 20 ꢁC then filtered
through celite and concentrated under reduced pressure.
The residue was dissolved in Et2O (15 mL) and the precip-
itate, which formed on standing, was filtered, washed sev-
eral times with pentane and dried under vacuum to afford
the desired product as a white solid in 95% yield.
1H NMR (400 MHz, CDCl3): 10.07 (s, 1H, NCHN),
3.45 (s, 6H, NCH3), 3.38 (s, 2H, CHt-Bu), 1.02 (s, 18H,
C(CH3)3).
4.6.4. 1,3-Bis-(1(S)-2-naphthyl-ethyl)-imidazolium
tetraflouroborate (20)
The desired product was obtained as a hygroscopic
yellow powder in 53% yield.
1H NMR (400 MHz, CDCl3): 9.47 (s, 1H, NCHN),
7.92–42 (m, 14H, ArH), 7.10 (s, 2H, NCH@C), 5.94
(q, J = 6.9 Hz, 2H, NCH(CH3)Naphth), 2.07 (d,
J = 7.0 Hz, 6H, CH3).
13C NMR (100 MHz, CDCl3): 160.6, 75.5, 38.7, 37.2,
27.5.
13C NMR (100 MHz, CDCl3): 135.0, 134.7, 133.4,
133.1, 129.7, 128.3, 127.7, 127.2, 127.0, 126.7, 123.9,
120.6, 60.5, 20.7.
20
M.p. 270–271 ꢁC. ½aꢁD ꢀ 54.4 (c 2.02, CHCl3). Anal.
20
Calc. for C13H27 IN2 (MW = 338.27): C, 46.16; H,
8.05; N, 8.28. Found C, 45.47; H, 8.36; N, 7.98%.
½aꢁD þ 6.9 (c 1.02, CHCl3).
MS (Electrospray/DCM) calculated for C27H25N2
377.2018, found 377.3.
4.6.8. 1,3-Bis-(1(R)-phenyl-ethyl)-imidazolium chloride
(26)
4.6.5. (4R,5R)-1,3-Dimethyl-4,5-di-tert-butyl-
imidazolin-2-ylidene silver(I) iodide (21)
To a stirred solution of amine (6.05 g, 50 mmol) in
toluene (50 mL), was added paraformaldehyde (1.5 g,
50 mmol). After stirring for 30 min at room tempera-
ture, the mixture was cooled to 0 ꢁC and a further equiv-
alent of amine (6.05 g, 50 mmol) was added. After
stirring for 15 min, HCl (15 mL of a 3.3 M aqueous
solution, 50 mmol) was added. The mixture was warmed
up to room temperature before adding glyoxal (7.0 g of
a 40% aqueous solution, 50 mmol). The heterogeneous
mixture was stirred at 40 ꢁC for 8 h, after which it was
cooled to room temperature, diluted with Et2O (50
mL) and quenched by addition of Na2CO3 (25 mL of
a saturated solution). The organic phase was separated
and the remaining aqueous layer was washed with
DCM (6 · 50 mL). The combined organic layers were
concentrated under reduced pressure to afford the de-
sired product in 17% yield.
To a solution of imidazolinium salt 24 (1 mmol) in
DCM (15 mL) was added Ag2O (0.5 mmol). The mix-
ture was stirred at 20 ꢁC until complete consumption
of the precipitate, filtered through celite and concen-
trated under reduced pressure to yield the desired com-
pound as a crystalline solid in quantitative yield. The
product could be recrystallised from a mixture of pen-
tane, DCM and Et2O to obtain a crystal suitable for
X-ray diffraction analysis.
1H NMR (400 MHz, DMSO-D6): 3.34 (s, 2H,
CHt-Bu), 3.21 (s, 6H, NCH3), 0.87 (s, 18H, C(CH3)3).
13C NMR (100 MHz, CDCl3): 213.7, 80.5, 45.8, 41.7,
32.9.
20
½aꢁD ꢀ 93.0 (c 1.0, CHCl3). Anal. Calc. for C13H2
IAgN2 (MW = 445.1): C, 35.08; H, 5.89; N, 6.29. Found
C, 36.89; H, 6.60; N, 6.37%.
1H NMR (400 MHz, CDCl3): 11.51 (s, 1H, NCHN),
7.50–7.40 (m, 4H, ArH), 7.40–7.30 (m, 6H, ArH), 7.09
(d, J = 1.5 Hz, 2H, NCH@C), 6.05 (q, J = 7.1 Hz, 2H,
NCH(CH3)Ph), 2.03 (d, J = 7.1 Hz, 6H, CH3).
4.6.6. (R,R)-4,5-di-tert-butylimidazolidine (23)
To a solution of imidazolidine 22 [28] (1 mmol) in
EtOH (20 mL), were added Pd(OH)2/C (0.1 mmol)
and ammonium formate (10 mmol). The resulting mix-
ture was refluxed for 6 h, filtered and then concentrated
under reduced pressure. To the residue were added Et2O
(15 mL) and K2CO3 (0.5 g). The resulting suspension
was stirred for 1 h, filtered and then concentrated under
reduced pressure to obtain the desired product as a
white solid in 95% yield.
4.7. General method for formation of diimines
(compounds 27 and 28)
To a stirred solution of amine (15 mmol) in dichloro-
methane (30 mL), was added MgSO4 (30 mmol) and for-
mic acid (50 mL). After stirring for 15 min at room
temperature, glyoxal was added (7.5 mmol of a 40%

5690
C.L. Winn et al. / Journal of Organometallic Chemistry 690 (2005) 5672–5695
aqueous solution) and the resulting mixture stirred at
room temperature for 16 h before filtering through celite
and concentrating under reduced pressure.
13C NMR (100 MHz, CDCl3): 159.0, 137.8, 136.6,
129.5, 127.6, 119.7, 65.9, 27.6, 10.8.
4.8.2. 1,3-Bis-(1(S)-1-naphthyl-ethyl)-imidazolium
chloride (30)
The crude product was washed with Et2O to purify.
The precipitate was collected by filtration and dried to
afford the desired product as a pale brown solid in
72% yield.
4.7.1. N,N-bis[(S)-1-phenyl-propyl]ethanediimine (27)
The product was isolated in quantitative yield as a
yellow oil and required no further purification.
1H NMR (400 MHz, CDCl3): 8.01 (s, 2H, HC@N),
7.31–7.20 (m, 10H, ArH), 4.10 (t, J = 9.3 Hz, 2H,
NCH(CH2CH3)Ph), 1.95 (m, 4H, CH2CH3), 0.85 (t,
J = 6.1 Hz, 6H, CH3).
1H NMR (400 MHz, CDCl3): 11.67 (s, 1H, NCHN),
8.16 (d, J = 8.4 Hz, 2H, ArH), 7.87 (d, J = 7.9 Hz, 4H,
ArH), 7.61–7.45 (m, 8H, ArH), 6.86 (q, J = 6.6 Hz, 2H,
NCH(CH3)Ar), 6.79 (s, 2H, NCH@C), 2.19 (d, J = 6.6
Hz, 6H, CH3).
13C NMR (100 MHz, CDCl3): 161.0, 142.9, 128.5,
127.2, 127.1, 53.4, 30.9, 11.0.
20
½aꢁD ꢀ 5.8 (c 1.17, CHCl3).
13C NMR (100 MHz, CDCl3): 130.4, 129.2, 127.9,
20
4.7.2. N,N-bis[(S)-1-naphthyl-ethyl]ethanediimine (28)
The crude product was a viscous orange, which could
be purified by washing with Et2O. The desired product
was isolated as a pale orange powder in 96% yield.
1H NMR (400 MHz, CDCl3): 8.17 (s, 2H, HC@N),
8.13 (d, J = 9.0 Hz, 2H, ArH), 7.88 (d, J = 8.5 Hz, 2H,
ArH), 7.78 (d, J = 8.2 Hz, 2H, ArH), 7.68 (d, J = 7.1
Hz, 2H, ArH), 7.56–7.46 (m, 6H, ArH), 5.41 (q, J = 6.6
Hz, 2H, NCH(CH3)Ar), 1.75 (d, J = 6.7 Hz, 6H, CH3).
13C NMR (100 MHz, CDCl3): 161.2, 139.4, 134.0,
130.7, 129.0, 127.8, 126.1, 125.6, 125.5, 124.0, 123.3,
64.8, 23.7.
126.6, 125.2, 124.4, 122.4, 120.1, 56.3, 21.6. ½aꢁD þ 32.2
(c 1.08, CHCl3).
4.9. General procedure for synthesis of silver carbene from
chloro-imidazolinium salt (compounds 31–33)
To a solution of chloro-imidazolinium salt (1 mmol) in
DCM (15 mL), was added Ag2O (0.5 mmol). The mixture
was stirred at room temperature for 16 h then filtered
through celite and concentrated under reduced pressure
to yield the desired compound as a crystalline solid.
20
½aꢁD ¼ ꢀ54.4 (c2.02, CHCl3). M/S(Electrospray/DCM)
calculated for C26H24N2 364.1939, found 365.2 (MH+).
4.9.1. 1,3-Bis-(1(S)-1-phenyl-ethyl)-imidazolin-2-ylidene
silver(I) chloride (31)
The desired product was isolated in quantitative yield
and required no further purification.
4.8. General method for preparation of chloro-
imidazolinium salts from diimines (compounds 29 and 30)
1H NMR (400 MHz, CDCl3): 7.40–7.30 (m, 6H,
ArH), 7.30–7.20 (m, 4H, ArH), 6.95 (s, 2H, NCH@C),
5.76 (q, J = 7.1 Hz, 2H, NCH(CH3)Ph), 1.84 (d,
J = 7.1 Hz, 6H, CH3).
To a solution of the starting diimine (5 mmol) in tol-
uene (25 mL) were added, paraformldehyde (5 mmol)
and HCl (5 mmol of a 32% aqueous solution). The
resulting mixture was heated to 45 ꢁC for 16 h, after
which it was cooled to room temperature, diluted with
Et2O (10 mL) and quenched by addition of Na2CO3 (5
mL of a saturated solution). The organic phase was sep-
arated and the remaining aqueous layer was washed
with DCM (3 · 10 mL). The combined organic layers
were dried over MgSO4 and concentrated under reduced
pressure to afford the crude product.
13C NMR (100 MHz, CDCl3): 139.7, 129.1, 128.6,
126.6, 61.0, 21.4.
20
½aꢁD þ 32.7 (c 1.09, CHCl3).
4.9.2. 1,3-Bis-(1(S)-phenyl-propyl)-imidazolin-2-ylidene
silver(I) chloride (32)
The desired product was isolated in quantitative yield
and required no further purification.
1H NMR (400 MHz, CDCl3): 7.40–7.33 (m, 10H,
ArH), 7.01 (s, 2H, NCH@C), 5.45 (m, 2H,
NCH(CH2CH3)Ph), 2.31–2.14 (m, 4H, CH2CH3), 0.89
(t, J = 7.3 Hz, 6H, CH3).
4.8.1. 1,3-Bis-(1(S)-phenyl-propyl)-imidazolium chloride
(29)
The crude product was purified by washing with
Et2O. The precipitate was collected by filtration and
dried to afford the desired product as a pale yellow solid
in 45% yield.
13C NMR (100 MHz, CDCl3): 139.1, 129.1, 128.5,
126.8, 118.9, 67.4, 27.8, 11.1.
20
½aꢁD þ 48.1 (c 0.75, CHCl3).
1H NMR (400 MHz, CDCl3): 11.79 (s, 1H, NCHN),
7.53 (d, J = 1.5 Hz, 2H, ArH), 7.51–7.35 (m, 6H, ArH),
7.11 (d, J = 7.9 Hz, 2H, NCH@C), 5.76 (t, J = 7.9 Hz,
2H, NCH(CH2CH3)Ph), 2.49–2.39 (m, 4H, CH2CH3),
0.96 (t, J = 7.2 Hz, 6H, CH3).
4.9.3. 1,3-Bis-(1(S)-1-naphthyl-ethyl)-imidazolin-2-
ylidene silver(I) chloride (33)
The desired product was isolated in 82% yield and re-
quired no further purification.

C.L. Winn et al. / Journal of Organometallic Chemistry 690 (2005) 5672–5695
5691
20
1H NMR (400 MHz, CDCl3): 8.15–7.42 (m, 14H,
ArH), 6.55 (s, 2H, NCH@C), 6.42 (q, J = 6.2 Hz, 2H,
NCH(CH3)Ar), 2.0 (d, J = 6.9 Hz, 6H, CH3).
½aꢁD ¼ ꢀ0.89 (c 1.01, CHCl3). M/S (Electrospray/
DCM) calculated for C29H36N2O2AgCl 587.1165, found
445 (M ꢀ AgCl) and 997 (C58H72N4O4AgCl2: dimeric
species).
13C NMR (100 MHz, CDCl3): 134.0, 131.0, 130.0,
129.1, 127.3, 126.2, 125.1, 124.4, 122.8, 119.0, 57.4, 22.4.
20
½aꢁD þ 49.8 (c 0.72, CHCl3).
4.9.7. N,N0-bis[isopinocampheyl]ethanediimine (40)
(method B)
4.9.4. N,N0-bis[(1R,2R)-2-benzyloxy-
cyclohexyl]ethanediimine (36)
To a solution of isopinene derived amine 39 (2 mmol)
in DCM (6 mL), was added glyoxal (1 mmol as a 40%
aqueous solution). The resulting mixture was stirred
for 6 h at room temperature before diluting with Et2O
(10 mL) and washing with H2O (15 mL). The organic
layer was dried over MgSO4 and concentrated under
reduced pressure. The crude product was isolated as
an orange solid in quantitative yield and was spectro-
scopically identical to the literature product.
1H NMR (400 MHz, CDCl3): 7.81 (s, 2H, HC@N),
3.40 (m, 2H, CHN), 2.34 (q, J = 5.7 Hz, 2H, CH),
2.22 (t, J = 1.6 Hz, 2H, CH), 2.02 (t, J = 7.3 Hz, 2H,
CH), 1.93 (q, J = 2.5 Hz, 2H, CH), 1.84 (q, J = 4.7
Hz, 4H, CH), 1.19 (s, 6H, CH3), 1.14 (d, J = 9.7 Hz,
2H, CH), 0.99 (s, 6H, CH3), 0.95 (d, J = 7.4 Hz, 6H,
CH3).
The diimine was prepared according to the general
procedure described above for compounds 27–28, and
was isolated in quantitative yield as a brown crystalline
product, which required no further purification.
1H NMR (400 MHz, CDCl3): 8.06 (s, 2H, HC@N),
7.34–7.16 (m, 10H, ArH), 4.47 (dd, J = 22.9 Hz, 10.9
Hz, 4H, CH2 Ph), 3.49 (dt, J = 3.5, 8.9 Hz, 2H,
CHN), 3.21 (dt, J = 4.6 Hz, 6.3 Hz, 2H, CHO), 2.16–
2.13 (m, 2H, CyH), 1.81–1.65 (m, 8H, CyH), 1.38–1.32
(m, 6H, CyH).
13C NMR (100 MHz, CDCl3): 161.9, 138.9, 128.2,
127.5, 127.3, 80.4, 74.9, 71.4, 32.8, 30.6, 24.4, 23.9.
20
½aꢁD ¼ þ8.4 (c 0.89, CHCl3). M/S (Electrospray/
DCM) calculated for C28H36N2O2 432.2777, found
433.3 (MH+) and 218.5 (M ꢀ 2Bn).
13C NMR (100 MHz, CDCl3): 159.3, 69.9, 47.4, 43.3,
41.5, 38.8, 35.6, 33.8, 27.9, 23.5, 19.8.
4.9.5. N,N0-bis (1R,2R)-2-benzyloxy-cyclohexyl
imidazolium chloride (37)
4.9.8. N,N0-diisopinocampheylimidazolium chloride (41)
The chloro-imidazolinium salt was prepared accord-
ing to the general procedure described above and was
isolated in 20% yield as a pale yellow crystalline solid
after trituration with Et2O.
The chloro-imidazolinium salt was prepared accord-
ing to the general procedure described above for
compounds 29–30, and was isolated in 99% yield as a
very hygroscopic brown crystalline product, which
was dried twice over MgSO4 but required no further
purification.
1H NMR (400 MHz, CDCl3): 11.13 (s, 1H, NCHN),
7.55 (s, 2H, NCH@C), 5.19 (dt, J = 7.4 Hz, J = 10 Hz,
2H, CHN), 2.55 (t, J = 4.2 Hz, 2H, CH), 2.25 (m, 2H,
CH), 2.08 (t, J = 2.5 Hz, CH), 1.96–1.92 (m, 6H, CH),
1.26–1.10 (m, 18H, CH3).
1H NMR (400 MHz, CDCl3): 10.56 (s, 1H, NCHN),
7.97–6.98 (m, 10H, ArH), 6.96 (d, J = 1.3 Hz, 2H,
NCH@C), 4.40 (d, J = 16 Hz, 4H, CH2 Ph), 4.13 (d,
J = 3.5, 8.9 Hz, 2H, CHN), 3.21 (m, 2H, CHO), 1.80–
1.20 (m, 16H, CyH).
13C NMR (100 MHz, CDCl3): 138.1, 120.5, 59.8,
47.3, 45.2, 41.3, 38.9, 36.5, 35.2, 27.9, 23.5, 20.2.
13C NMR (100 MHz, CDCl3): 153.0, 137.9, 128.9,
128.1, 127.5, 125.2, 79.4, 71.3, 64.8, 31.7, 30.8, 24.3,
4.9.9. N,N0-diisopinocampheyl-imidazolin-2-ylidene
silver(I) chloride (42)
20
23.9. ½aꢁD þ 351.9 (c 0.67, CHCl3).
The silver carbene was prepared according to the gen-
eral procedure described above and was isolated in 83%
yield, after drying over MgSO4, as a pale yellow crystal-
line solid.
4.9.6. N,N0-bis (1R,2R)-2-benzyloxy-cyclohexyl-
imidazolin-2-ylidene silver(I) chloride (38)
The silver carbene was prepared according to the gen-
eral procedure described above and was isolated in 44%
yield a brown crystalline product after purification by
trituration with Et2O.
1H NMR (400 MHz, CDCl3): 7.21 (s, 2H, NCH@C),
4.85–4.78 (m, 2H, CHN), 2.67–2.58 (m, 4H, CH), 2.20–
1.93 (m, 10H, CH), 1.29–1.11 (m, 18H, CH3). 13C NMR
(100 MHz, CDCl3): 119.2, 62.4, 47.5, 45.5, 41.5, 39.1,
37.1, 35.2, 27.9, 23.7, 20.4.
1H NMR (400 MHz, CDCl3): 7.26–6.98 (m, 10H,
ArH), 6.85 (s, 2H, NCH@C), 4.42 (d, J = 12 Hz, 2H,
CH2 Ph), 4.26–4.19 (m, 2H, CHO), 4.17 (d, J = 12 Hz,
2H, CH2 Ph), 3.67–3.62 (m, 2H, CHN), 2.31–1.33 (m,
16H, CyH).
20
½aꢁD þ 431.9 (c 0.44, CHCl3).
4.9.10. (2-Bromophenyl)-(R)-a-methylbenzylamine (44)
In a resealable Schlenk tube, which had been dried
under vacuum then backfilled with nitrogen, were
13C NMR (100 MHz, CDCl3): 137.99, 128.2, 127.5,
127.4, 118.8, 79.5, 70.3, 66.3, 33.8, 30.9, 24.9, 23.9.

5692
C.L. Winn et al. / Journal of Organometallic Chemistry 690 (2005) 5672–5695
placed Pd2dba3.CHCl3 (26 mg, 0.025 mmol), racemic
BINAP (31 mg, 0.05 mmol) and NaOt-Bu (0.31 g,
3.2 mmol) and toluene (3.5 mL). This was followed
by (R)-a-methylbenzyl-amine (0.38 mL, 2.95 mmol)
and dibromobenzene (0.3 mL, 2.5 mmol). The resulting
mixture was degassed, backfilled with nitrogen, sealed,
and then heated to 95 ꢁC for 45 h. After this time, the
starting material had been completely consumed (by
tlc) and the mixture was allowed to cool to room tem-
perature then diluted with Et2O, filtered through celite
and concentrated under reduced pressure. The crude
product was obtained as a dark brown oil which could
be purified by flash column chromatography (SiO2,
cyclohexane: EtOAc, 98:2). The desired product was
isolated as an orange/yellow oil in 55% yield and
was spectroscopically identical to the literature
compound.
13C NMR (100 MHz, CDCl3): 147.1, 145.1, 129.1,
128.7, 126.9, 125.9, 117.4, 113.5, 53.6, 25.0.
4.9.12. N,N0-Bis (R)-a-methylbenzyl-benzimidazolium
chloride (47)
To a solution of diamine 45 (0.12 g, 0.36 mmol) in
(EtO)3CH (7 mL) under nitrogen, was added HCl (35
mL of a 32% aqueous solution, 0.36 mmol). After stir-
ring for 1 h at room temperature, the mixture was
heated to 80 ꢁC. When condensation was observed on
the neck of the flask, the septum was removed and heat-
ing was continued for 2 h with the flask contents open to
the air. After this time, the mixture was cooled to room
temperature and diluted with Et2O (15 mL). After
allowing the solid to settle out, the supernatant was re-
moved by decantation. The remaining white solid was
washed 5 times with diethyl ether and the 3 times with
hot toluene. The residual solvents were removed under
high vacuum and then the product was dried under vac-
uum at 40 ꢁC over P2O5 to provide the title compound
as a white solid in 89% yield, which was spectroscopi-
cally identical to the literature compound.
1H NMR (400 MHz, CDCl3): 7.47 (d, J = 1.4 Hz,
1H, ArH), 7.46–7.28 (m, 5H, ArH), 7.03 (t, J = 7.5
Hz, 1H, ArH), 6.55 (dd, J = 8.8 Hz, J = 1.5 Hz, 1H,
ArH), 6.44 (dd, J = 8.2 Hz, J = 1.2 Hz, 1H, ArH),
4.77 (br s, 1H, NH), 4.56 (q, J = 6.3 Hz, 1H,
NCH(CH3)Ph), 1.63 (d, J = 6.7 Hz, 3H, CH3).
1H NMR (400 MHz, CDCl3): 12.51 (s, 1H, NCHN),
7.55–7.33 (m, 14H, ArH), 6.29 (q, J = 6.8 Hz, 2H,
NCH(CH3)Ph), 2.35 (d, J = 6.7 Hz, 6H, CH3).
4.9.11. N,N0-Bis (R)-a-methylbenzyl-1,2-diaminobenzene
(45)
In a resealable Schlenk tube, which had been dried
under vacuum then backfilled with nitrogen, were placed
Pd2dba3.CHCl3 (49 mg, 0.047 mmol), racemic BINAP
(58 mg, 0.093 mmol) and toluene (4 mL). The mixture
was degassed and backfilled with nitrogen before heat-
ing to 130 ꢁC for 20 min. After cooling to room temper-
ature, NaOt-Bu (0.24 g, 2.5 mmol) was added, followed
by (R)-a-methylbenzyl-amine (0.38 mL, 2.95 mmol),
amine 44 (0.32 g, 1.2 mmol) and toluene (5 mL). The
resulting mixture was degassed, backfilled with nitrogen,
sealed, and then heated to 130 ꢁC for 2 h. After cooling
to room temperature, the mixture was diluted with
Et2O, filtered through celite and concentrated under re-
duced pressure. The crude product was obtained as a
brown oil which could be purified by flash column chro-
matography (SiO2, cyclohexane:DCM, 80:20). The de-
sired product was isolated as a dark orange oil in 43%
yield and was spectroscopically identical to the literature
compound.
4.9.13. N,N0-Bis-(R)-a-methylbenzyl-benzimidazolin-2-
ylidene silver(I) chloride (48)
was prepared according to the general procedure
described above and was isolated in 95% yield as a fluffy
off-white solid which was purified by triturating with
Et2O.
1H NMR (400 MHz, CDCl3): 7.41–7.13 (m, 14H,
ArH), 6.20 (q, J = 7.1 Hz, 2H, NCH(CH3)Ph), 2.10
(d, J = 7.2 Hz, 6H, CH3).
13C NMR (100 MHz, CDCl3): 138.3, 129.4, 129.1,
128.5, 126.7, 126.6, 123.9, 113.2, 59.9, 19.5.
20
½aꢁD þ 48.6 (c 2.71, CHCl3).
4.9.14. (4R,5R)-1,3-Dibenzyl-4,5-di-tert-
butylimidazolinium tetrafluoroborate (50)
Imidazolinium bromide 58 (482 mg, 1.09 mmol) and
ammonium tetrafluoroborate (114 mg, 1.09 mmol) in a
mixture of acetone and Et2O (1/1, 20 mL) were stirred
overnight at ambient temperature. After filtration and
evaporation of the solvents, the pure imidazolinium tet-
rafluoroborate 50 (490 mg, 100% yield) was obtained as
an off-white solid with melting point = 125 ꢁC.
1H NMR (400 MHz, CDCl3): 8.85 (s, 1H, NCHN),
7.44–7.34 (m, 10H, ArH), 5.09 (d, J = 14.4 Hz, 2H,
NC(H0)HPh), 4.52 (d, J = 14.4 Hz, 2H, NC(H0)HPh),
3.47 (s, 2H, NCHt-Bu), 0.73 (s, 18H, C(CH3)3).
13C NMR (100 MHz, CDCl3): 158.9, 132.0, 130.2,
129.7, 70.6, 54.2, 36.1, 26.7.
1H NMR (400 MHz, CDCl3): 7.47–7.27 (m, 10H,
ArH), 6.65–6.50 (m, 4H, ArH), 4.54 (q, J = 6.6 Hz,
2H, NCH(CH3)Ph), 3.80 (br s, 2H, NH), 1.64 (d,
J = 6.7 Hz, 6H, CH3).
Compound 46 was also isolated from this reaction as
a pale yellow solid in 12% yield.
1H NMR (400 MHz, CDCl3): 7.41–7.15 (m, 5H,
ArH), 7.12 (t, J = 1.04 Hz, 2H, ArH), 6.69 (t J = 7.3
Hz, 1H, ArH), 6.55 (d, J = 7.7 Hz, 2H, ArH), 4.52 (q,
J = 6.7 Hz, 1H, NCH(CH3)Ph), 4.20 (br s, 1H, NH),
1.55 (d, J = 6.7 Hz, CH3).
20
½aꢁD ꢀ 139.3 (c 0.38, CH2Cl2).

C.L. Winn et al. / Journal of Organometallic Chemistry 690 (2005) 5672–5695
5693
4.9.15. (4R,5R)-1,3-Di-[4-methyl-benzyl]-4,5-di-tert-
4.10.3. (4R,5R)-1,3-Di-(3-methoxybenzyl)-4,5-di-tert-
butylimidazolinium chloride (60)
butylimidazolin-2-ylidene silver(I) chloride (54)
Obtained from imidazolinium bromide 61 in quanti-
tative yield after 16 h, using the general procedure de-
scribed above. Product obtained as off-white solid that
decomposed at 75 ꢁC.
To a solution of imidazoline 24 (0.88 mmol) in DCM
(5 mL) were added K2CO3 (0.5 g) and 3-methoxybenzyl
chloride (2.2 mmol, 320 mL). The solution was refluxed
for 12 h, filtered through celite and concentrated under
reduced pressure. The residue was taken up in Et2O
then the white precipitate that formed was filtered,
washed several times with ether and dried under vac-
uum to afford 381 mg of the expected salt (94%) as a
white solid.
1H NMR (400 MHz, CDCl3): 7.37 (d, 4H, J = 7.9
Hz, ArH), 6.96 (d, 2H, J = 7.9 Hz, ArH), 5.15 (d, 2H,
J = 14.3 Hz, NC(H0)HPh), 4.06 (d, 2H, J = 14.3 Hz,
NC(H0)HPh), 2.93 (s, 2H, NCHt-Bu), 2.08 (s, 6H,
ArCH3), 0.41 (s, 18H, C(CH3)3).
13C NMR (100 MHz, CDCl3): 137.8, 134.4, 130.1,
129.6, 70.7, 55.6, 35.2, 27.1, 21.1. Carbene carbon was
not observed.
1H NMR (400 MHz, CDCl3): 11.42 (s, 1H, NCHN),
7.34–6.88 (m, 8H, ArH), 5.47 (d, J 14.0 Hz, 2H,
NC(H0)HPh), 4.43 (d, J = 14 Hz, 2H, NC(H)H0Ph),
3.89 (s, 6H, OCH3), 3.40 (s, 2H, NCHt-Bu), 0.72 (s,
18H, C(CH3)3).
20
½aꢁD ꢀ 89 (c 0.22, CH2Cl2).
4.10. General procedure for alkyl substituted imidazolines
4.10.4. (4R,5R)-1,3-[(S)-1-phenylethyl]-4,5-di-tert-
butylimidazolinium iodide (61)
To a solution of imidazoline 24 (1 mmol) in DCM (5
mL) were added K2CO3 (0.5 g) and the appropriate alkyl
halide (2.1 mmol). The solution was stirred for 12 h at 20
ꢁC, filtered through celite and concentrated. The residue
was taken up in Et2O or pentane (15 mL) and the white
precipitate that formed was filtered, washed several times
with pentane and dried under vacuum to afford the
expected salts (85–95%) as white solids.
To a mixture of aminal 23 (1 mmol) and NaHCO3 (1
mmol) in DCM (5 mL) was added dropwise a solution
of iodine (1 mmol) in DCM (5 mL). The mixture was
stirred for 24 h at 20 ꢁC. Et2O (50 mL) and sodium
bisulfite (aqueous solution) were added. The suspension
was stirred until decolorisation occurred and a white
precipitate appeared in the organic phase. The solution
was filtered and the precipitate was washed with Et2O
then dried to afford 290 mg (58%) of the expected com-
pound as a white solid.
4.10.1. (4R,5R)-1,3-Dibenzyl-4,5-di-tert-
butylimidazolinium bromide (58)
1H NMR (200 MHz, CDCl3): 11.12 (s, 1H, NCHN),
7.56–7.54 (m, 4H, ArH), 7.43–7.38 (m, 6H, ArH), 5.53
(d, J = 7.1 Hz, 2H, NC(H)H0Ph), 4.47 (d, J = 7.1 Hz,
2H, NC(H0)HPh), 3.38 (s, 2H, NCHt-Bu), 0.68 (s, 18H,
C(CH3)3).
1H NMR (CDCl3): 10.21 (s, 1H, NCHN), 7.77–7.25
(m, 10H, ArH), 4.64 (q, J = 7.4 Hz, 2H, NCH(CH3)Ph),
3.35 (s, 2H, NCHt-Bu), 2.28 (d, J = 7.4 Hz, 6H,
NCH(CH3)Ph), 0.60 (s, 18H, C(CH3)3).
13C NMR (50 MHz, DMSO-D6): 159.6, 132.9, 130.0,
129.4, 69.7, 53.5, 35.8, 26.5.
4.11. Typical procedure for the synthesis of 49–54
20
M.p. 249–250 ꢁC. ½aꢁD ꢀ 114.3 (c 1.6, CHCl3). Anal.
To a solution of imidazolinium salt 58–61 (1 mmol)
in DCM (15 mL) was added Ag2O (0.5 mmol). The mix-
ture was stirred at 20 ꢁC until the precipitate was com-
pletely consumed (4–20 h), it was then filtered through
celite and concentrated to give the expected compounds
as crystalline solids in quantitative yields.
Calc. for C25H35BrN2 (MW = 443.46): C, 67.71; H,
7.96; N, 6.32. Found: C, 66.55; H, 7.79; N, 6.15%.
4.10.2. (4R,5R)-1,3-Di-(1-naphthylmethyl)-4,5-di-tert-
butylimidazolinium chloride (59)
To a solution of imidazoline 24 (1 mmol) in DCM (10
mL) were added K2CO3 (0.5 g) and 1-chloromethylnaph-
thalene (2.5 mmol). The solution was refluxed for 12 h,
filtered through celite and concentrated. The residue
was taken up in Et2O. The white precipitate that formed
was filtered, washed several times with ether and dried
under vacuum to afford 357 mg of the expected salt
(71%) as a pale beige solid.
4.11.1. (4R,5R)-1,3-Dibenzyl-4,5-di-tert-
butylimidazolin-2-ylidene silver(I) bromide (49)
The single crystals of 53 suitable for X-ray diffraction
analysis were obtained by recrystallisation from DCM
and hexane mixture.
1H NMR (200 MHz, CDCl3): 7.42–7.3 (m, 10H,
ArH), 5.12 (d, J = 14.5 Hz, 2H, NC(H0)HPh), 4.50 (d,
J = 4.5 Hz, 2H, NC(H)H0Ph), 3.24 (s, 2H, NCHt-Bu),
0.68 (s, 18H, C(CH3)3).
1H NMR (200 MHz, CDCl3): 11.76 (s, 1H, NCHN),
8.65 (d, J = 8.4 Hz, 2H, ArH), 7.89–7.55 (m, 12H,
ArH), 6.08 (d, J = 14.7 Hz, 2H, NC(H0)HPh), 4.48 (d,
J = 14.7 Hz, 2H, NC(H)H0Ph), 3.32 (s, 2H, NCHt-Bu),
0.40 (s, 18H, C(CH3)3).
13C NMR (CDCl3): 135.4, 129.7, 129.2, 128.8, 71.2,
56.5, 35.7, 27.3.

5694
C.L. Winn et al. / Journal of Organometallic Chemistry 690 (2005) 5672–5695
20
½aꢁD ꢀ 57 (c 0.68, DCM). Anal. Calc. for C25H34Ag-
BrN2 (MW = 550.3): C, 54.56; H, 6.23; N, 5.09. Found:
C, 55.24; H, 6.42; N, 4.89%.
the Swiss National Research Foundation No. 20-
068095.02 and COST action D/24/003/01 (OFES Con-
tract No. C02.0027) for financial support.
4.11.2. (4R,5R)-1,3-Di-(1-naphthylmethyl)-4,5-di-tert-
butylimidazolin-2-ylidene silver(I) chloride (51)
To a solution of imidazolinium salt (0.715 mmol) in
DCM (15 mL) was added Ag2O (0.55 equiv.). The mix-
ture was refluxed for 12 h, filtered through celite and
concentrated to give 432 mg (100%) of the expected
compound as a beige crystalline solid.
References
[1] D.J. Cardin, B. Cetinkaya, P. Dixneuf, M.F. Lappert, Chem.
Rev. 72 (1972) 545–574.
[2] W.A. Herrmann, C.-P. Reisinger, M. Spiegler, J. Organomet.
Chem. 557 (1998) 93–96.
´
[3] V. Calo, A. Nacci, L. Lopez, N. Mannarini, Tetrahedron Lett. 41
1H NMR (200 MHz, CDCl3): 8.31–7.45 (m, 14H,
ArH), 5.52 (d, J = 15.2 Hz, 2H, NC(H0)HPh), 5.07 (d,
J = 15.2 Hz, 2H, NC(H)H0Ph), 3.32 (s, 2H, NCHt-
Bu), 0.60 (s, 18H, C(CH3)3).
(2000) 8973–8976.
[4] H.M. Lee, S.P. Nolan, Org. Lett. 2 (2000) 2053–2055.
[5] J.P. Morgan, R.H. Grubbs, Org. Lett. 2 (2000) 3153–3155.
[6] C. Zhang, J. Huang, M.L. Trundell, S.P. Nolan, J. Org. Chem. 64
(1999) 3804–3805.
[7] V. Calo´, R. Del Sole, A. Nacci, E. Schingaro, F. Scordari, Eur. J.
Org. Chem. (2000) 869–871.
20
½aꢁD þ 196.0 (c 0.66, CHCl3).
[8] W.A. Herrmann, Angew. Chem. Int. Ed. 41 (2002) 1291–1309.
[9] T.J. Seiders, D.W. Ward, R.H. Grubbs, Org. Lett. 3 (2001) 3255–
3288.
[10] M.T. Powell, D.-R. Hou, M.C. Perry, X. Cui, K. Burgess, J. Am.
Chem. Soc. 123 (2001) 8878–8879.
[11] P.K. Fraser, S. Woodward, Tetrahedron Lett. 42 (2001) 2747–
2749.
[12] F. Guillen, C.L. Winn, A. Alexakis, Tetrahedron Asymmetry 12
(2001) 2083–2086.
[13] A. Alexakis, C.L. Winn, F. Guillen, J. Pytkowicz, S. Roland, P.
Mangeney, Adv. Synth. Catal. 345 (2003) 345–348.
[14] J. Pytkowicz, S. Roland, P. Mangeney, Tetrahedron Asymmetry
12 (2001) 2086–2088.
[15] H. Clavier, L. Coutable, J.-C. Guillemin, M. Mauduit, Tetrahe-
dron Asymmetry 16 (2005) 921–924.
[16] P.L. Arnold, M. Rodden, K.M. Davis, A.C. Scarisbrick, A.J.
Blake, C. Wilson, Chem. Commun. (2004) 1612–1613.
[17] S. Tominiga, Y. Oi, T. Kato, D.K. An, S. Okamoto, Tetrahedron
Lett. 45 (2004) 5585–5588.
[18] A.O. Larsen, W. Leu, C.N. Oberhuber, J.E. Campbell, A.H.
Hoveyda, J. Am. Chem. Soc. 126 (2004) 11130–11131.
[19] A. Alexakis, J. Vastra, P. Mangeney, Tetrahedron Lett. 38 (1997)
7745–7748.
4.11.3. (4R,5R)-1,3-Di-(3-methoxybenzyl)-4,5-di-tert-
butylimidazolin-2-ylidene silver(I) chloride (52)
To a solution of imidazolinium salt (0.83 mmol) in
DCM (15 mL) was added Ag2O (0.55 equiv.). The mix-
ture was refluxed for 12 h, filtered through celite and
concentrated to give 470 mg (95%) of the expected com-
pound as a white crystalline solid.
1H NMR (400 MHz, CDCl3): 7.30–6.80 (m, 8H,
ArH), 5.06 (d, J = 15.2 Hz, 2H, NC(H)H0Ph), 4.49 (d,
J = 15.2 Hz, 2H, NC(H0)HPh), 3.81 (s, 6H, OCH3),
3.28 (s, 2H, NCHt-Bu), 0.75 (s, 18H, C(CH3)3).
20
½aꢁD þ 192.8 (c 0.66, CHCl3).
4.11.4. (4R,5R)-1,3-Bis-[(S)-1-phenylethyl]-4,5-di-tert-
butylimidazolin-2-ylidene silver(I) iodide (53)
1H NMR (CDCl3): 7.72–7.23 (m, 10H, ArH), 4.66 (q,
J = 7.4 Hz, 2H, NCH(CH3)Ph), 2.20 (s, 2H, NCHt-Bu),
2.15 (d, J 7.4 Hz, 6H, NCH(CH3)Ph), 0.61 (s, 18H,
C(CH3)3).
13C NMR (DMSO-D6): 141.3, 128.4, 127.9, 72.1, 58.2,
34.9, 26.0, 22.1. The carbene carbon was not observed.
[20] A.J. Arduengo III, R.L. Harlow, M.A. Kline, J. Am. Chem. Soc.
113 (1991) 361–363.
[21] N. Krause, A. Hoffmann-Roder, Synthesis (2001) 171–196.
[22] A. Alexakis, C. Benhaim, Eur. J. Org. Chem. (2002) 3221–3236.
[23] K. Bambridge, M.J. Begley, N.S. Simpkins, Tetrahedron Lett. 35
(1994) 3391–3394.
20
½aꢁD þ 609.9 (c 0.21, CHCl3).
5. Supplementary material
[24] G. Martelli, S. Mori, D. Savoia, Tetrahedron 56 (2000) 8367–
8374.
Crystallographic data for the structural analysis of
compounds 21 and 49 have been deposited with the
Cambridge Crystallographic Data Centre, CCDC Nos.
160043 (compound 21) and 160044 (compound 49) [45].
Copies of this information may be obtained free of charge
from The Director, CCDC, 12 Union Road, Cambridge,
CB2 1EZ, UK (fax: +44 1223 336033; e-mail: deposit@
ccdc.cam.ac.uk or www: http://www.ccdc.cam.ac.uk).
[25] S. Saba, A.-M. Brescia, M.K. Kaloustain, Tetrahedron Lett. 32
(1991) 5031–5034.
[26] P. Mangeney, A. Alexakis, F. Grojean, J.F. Normant, Tetrahe-
dron Lett. 29 (1988) 2675–2676.
[27] M. Periasamy, G. Srinivas, G.V. Karunkar, P. Bharathi, Tetra-
hedron Lett. 40 (1999) 7577–7580.
[28] R. Hulst, N. Koen de Vries, B.L. Feringa, Tetrahedron Asym-
metry 5 (1994) 699–708.
[29] S. Lee, J.F. Hartwig, J. Org. Chem. 66 (2001) 3402–3415.
[30] A.J. Arduengo III, H.V.R. Dias, R.L. Harlow, M.A. Kline, J.
Am. Chem. Soc. 114 (1992) 5530–5534.
[31] A.J. Arduengo III, J.R. Goerlich, W.J. Marshall, J. Am. Chem.
Soc. 117 (1995) 11027–11028.
Acknowledgements
[32] A. Alexakis, J. Burton, J. Vastra, C. Benhaim, X. Fournioux, A.
The authors thanks Stephane Rosset for analytical
support, BASF for the generous gift of chiral amines,
´
van den Heuvel, J.M. Leveˆque, F. Maze, Eur. J. Org. Chem.
(2000) 4011–4027.

C.L. Winn et al. / Journal of Organometallic Chemistry 690 (2005) 5672–5695
5695
[33] A. Alexakis, C. Benhaim, S. Rosset, M. Humam, J. Am. Chem.
Soc. 124 (2002) 5262–5263.
[43] A.J. Arduengo III, H.W.R. Dias, J.C. Calabrese, F. Davidson,
Organometallics 12 (1993) 3405–3409.
[34] G.D. Allred, L. Liebeskind, J. Am. Chem. Soc. 118 (1996) 2748–
2749.
[35] We had previously observed that changing the counteranion to a
halide led to a decrease in enantiomeric excess and in this case, the
iodocomponds gave less that 3% ee.
[44] H.M.J. Wang, I.J.B. Lin, Organometallics 17 (1998) 972–975.
[45] J. Pytkowicz, S. Roland, P. Mangeney, J. Organomet. Chem. 631
(2001) 157–163.
[46] D.S. McGuiness, K.J. Cavell, Organometallics 19 (2000) 741–748.
[47] S. Roland, P. Mangeney, A. Alexakis, Synthesis (1999) 228–230.
[48] A. Alexakis, I. Aujard, J. Pytkowicz, S. Roland, P. Mangeney, J.
Chem. Soc., Perkin Trans. 1 (2001) 949–951.
[36] F.M. Rivas, U. Riaz, A. Giessert, J.A. Smulik, S.T. Diver, Org.
Lett. 3 (2001) 2673–2676.
[37] F.M. Rivas, U. Riaz, S.T. Diver, Tetrahedron Asymmetry 11
(2000) 1703–1707.
[38] W.A. Herrmann, L.J. Goossen, C. Koecher, G.R. Artus, J.
Angew. Chem. Int. Ed. 35 (1996) 2805–2807.
[49] H.T. Dieck, J. Dietrich, Chem. Ber. 117 (1984) 649–701.
[50] A. Alexakis, C. Benhaim, Org. Lett. 2 (2000) 2579–2581.
[51] G. Domany, A. Gere, M. Paroczai, C. Szantay, G. Ferenczy,
Pharmazie 48 (1993) 941–942.
[39] W.A. Herrmann, M. Elison, J. Fischer, C. Kocher, G.R. Artus, J.
Chem. Eur. J. 2 (1996) 777–780.
[52] 62 was very insoluble in CDCl3 and its solubility in other solvents
was even lower. For this reason, it was not possible to obtain a 13
C
[40] D. Seebach, A.K. Beck, A. Heckel, Angew. Chem. Int. Ed. 40
(2001) 92–138.
[41] D. Seebach, R. Dahinden, R.E. Martin, A.K. Beck, D.A. Plattner,
NMR spectrum with high enough resolution, nor to measure its
optical rotation. Moreover, elemental analysis of 62 performed on
the same sample gave irreproducible results.
F.N.M. Kuhnle, J. Org. Chem. 60 (1995) 1788–1799.
¨
[42] D. Seebach, D.A. Plattner, A.K. Beck, Y.M. Wang, D. Hunziker,
W. Petter, Helv. Chim. Acta 75 (1992) 2171–2209.
[53] W.L.F. Armarengo, D.D. Perrin, Purification of Laboratory
Chemicals, fourth ed., Butterworth and Heinemann, Bath,
1998.