Cis-selective asymmetric cyclopropanation of olefins catalyzed by five-coordinate [RuCl(PNNP)]+ complexes

Article abstract of DOI:10.1021/om010020p

The five-coordinate complex [RuCl(1a)]PF₆ (3a; 1a = (S)-N,N′-bis[2-(diphenylphosphino)-benzylidene]-2,2′- diamino-6,6′ dimethylbiphenylene) catalyzes the cyclopropanation of styrene by decomposition of diazoesters. The cis cyclopropane derivati

Full text of DOI:10.1021/om010020p

2102
Organometallics 2001, 20, 2102-2108
Cis-Selective Asym m etr ic Cyclop r op a n a tion of Olefin s
Ca ta lyzed by F ive-Coor d in a te [Ru Cl(P NNP )]+ Com p lexes
Stephan Bachmann, Maya Furler, and Antonio Mezzetti*
Laboratory of Inorganic Chemistry, Swiss Federal Institute of Technology, ETH Zentrum,
CH-8092 Zu¨rich, Switzerland
Received J anuary 11, 2001
The five-coordinate complex [RuCl(1a )]PF6 (3a ; 1a ) (S)-N,N′-bis[2-(diphenylphosphino)-
benzylidene]-2,2′-diamino-6,6′-dimethylbiphenylene) catalyzes the cyclopropanation of styrene
by decomposition of diazoesters. The cis cyclopropane derivative is formed with moderate
selectivity (cis:trans ratio is 48:52), but high enantioselectivity (90-96% ee). The related
species [RuCl(1b)]PF6 (3b; 1b ) N,N′-bis[2-(diphenylphosphino)benzylidene]-(1S,2S)-diami-
nocyclohexane) gives the cis product with selectivity up to 95% and enantioselectivity up to
>99% ee. High cis selectivity is obtained also with 2,5-dimethyl-2,4-hexadiene as substrate.
Complex 3a yields ethyl chrysanthemate with 94% cis selectivity and enantioselectivity up
to 80% ee. The putative carbene intermediate trans-[RuCl(C(H)COOEt)(1b )]+ (11b) was
prepared and characterized spectroscopically in solution. Its reaction with styrene gives the
cyclopropane derivative with 98:2 cis:trans selectivity. The steric constraints in the transition
states involving 11b and styrene were estimated by means of molecular modeling calcula-
tions. The relative total energies of the four diastereomeric aggregates follow the experimental
enantio- and diastereoselectivity trends. The model proposed also predicts the correct absolute
configuration of both cis and trans cyclopropane products.
Sch em e 1
In tr od u ction
The most striking progress in the asymmetric cyclo-
propanation of olefins is the development of catalysts
affording trans cyclopropane derivatives with high
enantio- and diastereoselectivity.1 With styrene, semi-
corrin and bis(oxazoline) copper catalysts have been
developed that afford both the trans product and the
cis one with high enantioselectivity, but always favor
the trans isomer (Scheme 1).2-7 More recently, ruthe-
nium complexes containing tridentate nitrogen donors
(pybox)8 and cobalt(III) salen systems9 have further
improved the trans diastereoselectivity.
cyclopropane derivatives have been developing slowly.
Doyle has shown that also dirhodium(II) carboxami-
dates can give higher enantioselectivities for the cis
cyclopropane than for the trans isomer. However, the
trans isomer is generally the major product.11 A break-
through has been recently achieved by Katsuki, whose
ruthenium nitrosyl complexes [RuCl(salen)(NO)] give
highly enantio- and cis-selective cyclopropanation of
styrenes upon irradiation with visible light.12
In contrast, after Kodadek’s pioneering work with
rhodium chiral porphyrins,10 catalytic systems for the
efficient diastereo- and enantioselective formation of cis
We recently reported the synthesis of some five-
coordinate species [RuCl(PNNP)]+, where PNNP are
tetradentate ligands with a P2N2 donor set.13 These
complexes catalyze the asymmetric epoxidation of ole-
fins by hydrogen peroxide. We find now that [RuCl-
(PNNP)]+ (PNNP is 1a or the already reported 1b,13b
Chart 1) catalyze the cyclopropanation of styrene with
diazoesters giving the cis isomer 4 with high diastereo-
and enantioselectivity.
* E-mail: mezzetti@inorg.chem.ethz.ch.
(1) For a recent comprehensive review of the literature, see: (a)
Doyle, M. P.; McKervey, M. A.; Ye, T. Modern Catalytic Methods for
Organic Synthesis with Diazo Compounds: From Cyclopropanes to
Ylides; Wiley: New York, 1998; Chapter 4, pp 183-203. (b) Compre-
hensive Asymmetric Catalysis; J acobsen, E. N., Pfaltz, A., Yamamoto,
H., Eds.; Springer: Berlin, 1999; Chapter 16, pp 513-603.
(2) (a) Fritschi, H.; Leutenegger, U.; Pfaltz, A. Helv. Chem. Acta
1988, 71, 1553. (b) Mu¨ller, D.; Umbricht, G.; Weber, B.; Pfaltz, A. Helv.
Chem. Acta 1991, 74, 232.
(3) Evans, D. A.; Woerpel, K. A.; Hinman, M. M.; Faul, M. M. J .
Am. Chem. Soc. 1991, 113, 726.
(4) Lowenthal, R. E.; Abiko, A.; Masamune, S. Tetrahedron Lett.
1990, 31, 6005.
(5) Cho, D. J .; J eon, S. J .; Kim, H. S.; Cho, C. S.; Shim, S. C.; Kim,
T. J . Tetrahedron: Asymmetry 1999, 10, 3833.
(6) Glos, M.; Reiser, O. Org. Lett. 2000, 2, 2045.
(7) Rios, R.; Liang, J .; Lo, M. M. C.; Fu, G. C. Chem. Commun. 2000,
377.
(8) Selected papers: (a) Nishiyama, H. Enantiomer 1999, 4, 569.
(b) Nishiyama, H.; Ito, Y.; Matsumoto, H.; Park, S. B.; Itoh, K. J . Am.
Chem. Soc. 1994, 116, 2223.
(10) See, for instance: (a) Maxwell, J . L.; O’Malley, S.; Brown, K.
C.; Kodadek, T. Organometallics 1992, 11, 645. (b) O’Malley, S.;
Kodadek, T. Organometallics 1992, 11, 2299.
(11) (a) Doyle, M. P.; Hu, W. J . Org. Chem. 2000, 65, 8839. (b) Doyle,
M. P.; Zhou, Q.-L.; Simonsen, S. H. Synlett 1996, 697.
(12) (a) Uchida, T.; Irie, R.; Katsuki, T. Synlett 1999, 1793. (b)
Uchida, T.; Irie, R.; Katsuki, T. Tetrahedron 2000, 56, 3501. (c) Niimi,
T.; Uchida, T.; Irie, R.; Katsuki, T. Tetrahedron Lett. 2000, 41, 3647.
(13) (a) Stoop, R. M.; Mezzetti, A. Green Chem. 1999, 39. (b) Stoop,
R. M.; Bachmann, S.; Valentini, M.; Mezzetti, A. Organometallics 2000,
19, 4117.
(9) Fukuda, T.; Katsuki, T. Tetrahedron 1997, 53, 7201.
10.1021/om010020p CCC: $20.00 © 2001 American Chemical Society
Publication on Web 04/19/2001

Cis-Selective Asymmetric Cyclopropanation
Organometallics, Vol. 20, No. 10, 2001 2103
Ch a r t 1
binaphthylene, 1d ) are completely unreactive under the
same conditions.13b In view of the similarity between
ligands 1a and 1d , and 1b and 1c, these reactivity
differences are surprising. Indeed, the bis(imino) com-
plex [RuCl2(1b)] and the bis(amino) derivative [RuCl2-
(1c)] have very similar metrical parameters, despite the
different donor properties of the PNNP ligands.15
We have already reported the preparation of [RuCl-
(1b)]PF6 (3b) from 2b and its high reactivity toward
most oxygen donors and, in particular, water.13b In
contrast, the biphenyl-bridged [RuCl(1a )]PF6 (3a ) is
nearly unreactive toward water or other oxygen donors,
such as THF. This is analogous to what was previously
observed for the five-coordinate derivative with the
binaphthyl-bridged ligand 1d . The reduced oxophilicity
in the biphenyl and binaphthyl derivatives is apparently
due to a combination of steric and electronic factors.
Asym m etr ic Cyclop r op a n a tion . The five-coordi-
nate complexes 3a and 3b were used as catalyst
precursors for the asymmetric cyclopropanation of sty-
rene and 2,5-dimethylhexa-2,4-diene with different
diazo esters (Scheme 1). The precatalysts 3a and 3b
were prepared in situ from the corresponding dichloro
derivatives 2a and 2b by chloride abstraction. The
complexes 2a or 2b were treated with Tl[PF6] (1 equiv)
in CH2Cl2 overnight, after which TlCl was filtered off
over Celite, and the resulting red-brown solution was
added to the olefin (and decane, as internal standard
for GC analysis). Complex 3a , containing the biphenyl-
bridged ligand 1a , gives quantitative precipitation of
TlCl, whereas TlCl tends to precipitate slowly from the
solutions of [RuCl(1b)]PF6 (3b). In general, the chloride
abstraction is necessary for catalytic activity, as the
dichloro complexes 2a and 2b are catalytically unreac-
tive.16
The cyclopropanation reactions were carried out add-
ing slowly a CH2Cl2 solution (1 mL) of N2C(H)COOR
(R ) Et or But) to a solution containing styrene (1 equiv)
and the catalyst in CH2Cl2 (1 mL) at room temperature
over 6 h. The reaction solutions were protected from
light. Precatalyst 3a (5 mol %) gave the cis and trans
cyclopropane derivatives in 40:60 ratio (Table 1, run 1)
with a total yield of 84%. The catalytic system shows a
similar performance when a substrate-to-catalyst ratio
of 100:1 is used (run 2).
The analysis of the stereochemical course of the
reaction catalyzed by 3a is intriguing. Indeed, the
enantiomeric excess of the cis derivative 4 is generally
higher than 90%, whereas the trans product 5 is formed
with enantioselectivities not exceeding 55%. Apart from
some exceptions,10-12 this observation is contrary to
what is observed with most catalytic systems.1-9 This
prompted us to test the already known five-coordinate
complex [RuCl(1b)]PF6 (3b). The reaction catalyzed by
3b (5 mol %) afforded the cis product 4 with good
diastereo- and enantioselectivity (91:9 cis:trans ratio,
87% ee, run 5). As observed with 3a , the residual trans
isomer is formed with low enantioselectivity. Higher ee
Sch em e 2
Resu lts a n d Discu ssion
[Ru Cl(P NNP )]+ Com p lexes. The new ligand (S)-
N,N′-bis[2-(diphenylphosphino)benzylidene]-2,2′-diamino-
6,6′-dimethylbiphenylene (1a ) was prepared by conden-
sation of P(o-C(H)O-C6H4)Ph2 (2 equiv) with enantiomeri-
cally pure (S)-2,2′-diamino-6,6′-dimethylbiphenylene by
the procedure previously reported for N,N′-bis[2-(diphen-
ylphosphino)benzylidene]-(1S,2S)-diaminocyclohexane
(1b).13b As for other PNNP ligands, the yield is nearly
quantitative (95%). The 31P NMR spectrum of 1a shows
a singlet at δ -13.6 for the two equivalent phosphorus
atoms. The two imino protons give a doublet at δ 8.9
(J P,H ) 5.5 Hz) in the 1H NMR spectrum. The IR
spectrum (KBr) shows one band at 1624 cm-1 for the
CdN stretching vibration. The CH2Cl2, CHCl3, or
toluene solutions of 1a and 1b are stable toward
oxidation and hydrolysis when exposed to the atmos-
phere for days.
The reaction of 1a with [RuCl2(PPh3)3] in boiling
toluene, benzene, or CDCl3 gives the red six-coordinate
complex trans-[RuCl2(1a -κ4P,N,N,P)] (2a ) as the only
isomer, as supported by the singlet at δ 46.5 in the 31P
NMR spectrum. Complex [RuCl2(1b)] (2b) was prepared
by reaction of 1b with [RuCl2(PPh3)3] in CH2Cl2 at room
temperature. Under these conditions 2b is formed as a
mixture of trans and cis-â isomers in a 3:1 ratio, as
reported previously.13b
The six-coordinate complex [RuCl2(1a )] (2a ) reacts
with Tl[PF6] in CH2Cl2 to give the brown five-coordinate
species [RuCl(1a )]PF6 (3a ) (Scheme 2). The latter
complex is isolated by addition of hexane or pentane to
the reaction solution after filtering off the thallium
chloride. The formulation of 3a is supported by MS
(FAB+) and by the appearance of two doublets in the
31P NMR spectrum at δ 80.8 and 42.9 (J ) 27.6 Hz), in
the range expected for five-coordinate [RuXN2P2]+ com-
plexes (X ) halide).13b,14 It should be noted that both
trans-2a and trans-2b react with Tl[PF6] in CH2Cl2 at
room temperature, whereas the related species trans-
[RuCl2(PNNP)] (PNNP ) N,N′-bis[2-(diphenylphosphi-
no)benzyl]-(1S,2S)-diaminocyclohexane, 1c; (S)-N,N′-
bis[2-(diphenylphosphino)benzylidene]-2,2′-diamino-1,1′-
(15) Gao, J . X.; Ikariya, T.; Noyori, R. Organometallics 1996, 15,
1087.
(16) No reaction is observed at room temperature in the presence
of 2b. A slight catalytic activity can be observed with the dichloro
derivative 2b in dichloroethane at reflux (43% conversion, 3% yield,
86:14 cis-to-trans ratio), but 1 and 2 are formed with low enantiose-
lectivity (4 and 2% ee, respectively), suggesting that the largest
contribution to the reaction is due to the uncatalyzed (thermic) reaction.
(14) (a) Shen, J . Y.; Slugovc, C.; Wiede, P.; Mereiter, K.; Schmid,
R.; Kirchner, K. Inorg. Chim. Acta 1998, 268, 69. (b) Costella, L.; Del
Zotto, A.; Mezzetti, A.; Zangrando, E.; Rigo, P. J . Chem. Soc., Dalton
Trans. 1993, 3001.

2104 Organometallics, Vol. 20, No. 10, 2001
Bachmann et al.
Ta ble 1. Cyclop r op a n a tion of Styr en e Ca ta lyzed
by 3a ,ba
Ta ble 2. Cyclop r op a n a tion of
2,5-Dim eth yl-2,4-h exa d ien e (8) Ca ta lyzed by 3a ,ba
ee (%)
run cat. mol % conv. (%) yield (%) cis:trans ee cis (%)
run cat.
R
conv. (%) yield (%) cis:trans
cis
90
93
91
96
87
92
92
82
61
trans
1b
2c
3
3a
3a
3a
3b
1
1
5
5
27
49
55
9
9
18
18
0
92:8
92:8
94:6
80
75
75
1
3a
3a
3a
3a
3b
3b
3b
3b
3b
3b
Et
Et
99
80
80
89
70
60
36
31
42
20
84
67
53
67
41
31
12
8
40:60
48:52
38:62
3:97
91:9
93:7
55
40
54
38
24
3
2
8
5
0
2b
3c
4
4b
Et
But
Et
a
Reaction conditions: see Table 1, footnote a. The ee of the
trans isomer was not determined. No purification. c The diazo-
ester was dried before the reaction, but not distilled.
b
5
6d
7e
8f
9c
10d
Et
Et
Et
95:5
84:16
57:43
83:17
Sch em e 3
Et
5
13
But
>99
a
Reaction conditions: N2C(H)COOEt (51 µL, 0.42 mol) in
CH2Cl2 (1 mL) was added over 6 h to a CH2Cl2 solution (1 mL) of
the olefin (0.42 mol) and 3 (21 µmol, 5 mol %), 20 h total reaction
b
d
time. 1 mol % catalyst was used. c In THF as solvent. The
diazoester was dried before the reaction, but not distilled. e At 0
°C. f Diazoester added in one portion.
conditions but without added olefin. Complex 3a gave
diethyl maleate with 45% yield and 95% selectivity after
20 h, whereas 3b is less reactive (25% yield after 72 h,
90% selectivity). As complexes 3 catalyze the homocou-
pling of diazoacetate with much higher cis selectivity
than in the homocoupling side reaction during cyclo-
propanation, we speculate that the formation of 6 and
7 in the latter reaction is not catalyzed by 3.18 The
stereoselective formation of maleate is noteworthy, and
few systems have been reported to effectively perform
this transformation.19
The optimization of the reaction solvent was limited
by the high oxophilicity of complex 3b, as only chlori-
nated solvents can be used. Indeed, O or N donors block
the active site of the catalyst effectively and form
relatively stable adducts.13b Accordingly, 3b is an inef-
fective catalyst in THF and gives low yield and enan-
tioselectivity (run 9). In contrast, 3a does not form
adducts with THF, which can then be used as solvent.
However, the selectivity with both catalysts shifts
toward the formation of the trans cyclopropanation
product in THF (runs 3, 9). Aromatic solvents (benzene
or toluene) are ineffective owing to the low solubility of
the cationic species 3.
To assess the scope of the [RuCl(PNNP)]+ catalysts,
we investigated the cyclopropanation of 2,5-dimethyl-
2,4-hexadiene (8), a trisubstituted olefin, with N2C(H)-
COOEt to give ethyl chrysanthemate (Scheme 3). Com-
plex 3a is moderately active with the sterically hindered
diolefin 8, whereas 3b does not give any cyclopropana-
tion product (Table 2). Modest yields of 9 and 10 are
obtained only with dried and distilled ethyl diazoacetate
(runs 2, 3). Increasing the catalyst-to-substrate ratio
from 1 to 5% does not improve the yield. In all cases,
excellent cis selectivity is obtained (>90%). The cis
isomer 9 is formed with fair enantioselectivity (75-80%
ee). To the best of our knowledge, this is the best cis
selectivity ever obtained in the cyclopropanation of 8,
as the reported catalysts give prevalently the trans
isomer.20 Thus, 3a gives access to the cis cyclopropa-
values of 4 (up to 92%) are achieved by using diazoesters
that have been dried over molecular sieves, but not
distilled (run 6). The improvement of the enantioselec-
tivity, however, is at cost of the reaction yield. Running
the reaction catalyzed by 3b at 0 °C gives 95% diaste-
reoselectivity and 92% ee for 4 (run 7).
The chemical yields of the catalytic reactions are, at
best, only moderate. This is possibly because of catalyst
deactivation. Indeed, longer reaction times do not
improve the cyclopropane yield. However, it should be
noted that no excess reagent is used, whereas most
published systems use a 4-fold excess of the olefin. In
fact, yield and selectivity did not improve when an
excess either of olefin (2 equiv) or of diazoacetate (10
equiv) was used. The critical factor was instead the slow
and continuous addition of the diazoacetate over several
hours. Under these conditions, the formation of maleate
and fumarate is kept to a minimum. Addition of the
diazoester in one portion gives lower diastereo- and
enantioselectivity (run 8).
We also investigated the effect of changing the ester
substituent and reaction solvent. With both catalysts
3a and 3b, increasing the steric bulk of the ester chain
favors the formation of the trans isomer. In the case of
3a , this leads to the formation of nearly pure trans-4,
but with low enantioselectivity (38% ee, run 4). In the
case of catalyst 3b, N2C(H)COOBut gives enantiopure
cis isomer 4, but with slightly decreased diastereose-
lectivity (83%, run 10). The formation of the trans
isomer 5 is also favored with THF as solvent (see below).
Further, we assessed the formation of maleate (6) and
fumarate (7) by monitoring selected reaction solutions
1
(under standard reaction conditions) by means of H
NMR spectroscopy.17 Maleate and fumarate are first
detected after 4 h reaction time. After 20 h reaction
time, maleate and fumarate are present in a 3:1 ratio
and their total yield does not exceed 10% (based on
starting diazoester). At this time, unreacted ethyl
diazoacetate is present in the reaction solution. To check
whether the formation of 6 and 7 is catalyzed by the
five-coordinate complexes 3, a control reaction was run
with ethyl diazoacetate and 3a or 3b under standard
(18) Decomposition products of 3, or the uncatalyzed thermal
decomposition of the diazoester, can be responsible for the formation
of 6 and 7 during cyclopropanation.
(19) (a) Baratta, W.; Del Zotto, A.; Rigo, P. Chem. Commun. 2997,
2163. (b) Del Zotto, A.; Baratta, W.; Verardo, G.; Rigo, P. Eur. J . Org.
Chem. 2000, 2795.
(17) GC analysis of the coupling products is inadequate, as ethyl
diazoacetate dimerizes to fumarate and maleate during the chromato-
graphic analysis.
(20) (a) Aratani, T. Pure Appl. Chem. 1985, 57, 1839. (b) Kanemasa,
S.; Hamura, S.; Harada, E.; Yamamoto, H. Tetrahedron Lett. 1994,
35, 7985. See also ref 1.

Cis-Selective Asymmetric Cyclopropanation
Organometallics, Vol. 20, No. 10, 2001 2105
Sch em e 4
nation product of a conjugated diolefin even in the
absence of halogen substituents in the vinylic positions.
F igu r e 1. Minimized conformations of the carbene com-
plex [RuCl(C(H)COOEt)((S,S)-1b)] (11b) (with relative
energies).
Some general comments are appropriate at this stage.
The first observation is that the cyclopropanation of
styrene gives the cis derivative 4 with as good a
selectivity and enantioselectivity as the best ever ob-
tained and are comparable with those of the [RuCl-
(salen)(NO)] system.12a,b The [RuCl(PNNP)]+ and [RuCl-
(salen)(NO)] systems are similar, but the five-coordinate
system is intrinsically reactive in view of its 16-electron
configuration and does not need to be photoactivated.
An intriguing feature of the cyclopropanation of styrene
is that catalysts 3 give the cis and trans isomers 4 and
5 with reversed configuration at the cyclopropane C(1)
atom, whereas the configuration at C(2) is the same in
both. Katsuki observed the same stereochemical course
in the reaction catalyzed by [RuCl(salen)(NO)].12a,b
However, this is contrary to what is typically observed
in the cyclopropanation of styrene catalyzed by Cu-
based,2-7 rhodium(II) carboxamidato,21 and ruthenium
pybox8 complexes. With these systems, the configuration
at C(1) is the same in the major cis and trans isomers.
This peculiarity prompted us to investigate the mech-
anism of stereocontrol. With this aim in mind, we
investigated the nature of the carbene intermediate
first.
several experiments carried out at room temperature.23
We conclude that the carbene complex 11b is the
thermodynamic product featuring a C-Ru-Cl arrange-
ment as depicted in Scheme 4. In fact, the mutual trans
arrangement of the carbene ligand (a π-acceptor) and
chloride (a π-donor) is favored in view of the resulting
push-pull interaction.24
Complex 11b reacts instantaneously with styrene (1
equiv) at room temperature to give the cyclopropane
derivatives 4 and 5 (Scheme 4). The reaction is es-
sentially quantitative and yields 4 and 5 in 98:2 ratio,
as determined by GC analysis.
Molecu la r Mod elin g. On the basis of the reaction
of the intermediate carbene complex 11b with styrene,
we modeled the putative “transition states” of the
carbene-transfer step using the Cerius2 program.25 In
the first step, the two conformations of the ester chain
of the carbene ligand were optimized using a Ru-
C(carbene) bond order of 2. This gave optimized bond
lengths of 2.015 Å (A) and 2.092 Å (B) for the two
possible ester chain configurations when the ester group
is ethyl (Figure 1).26 The conformation A leads to the
(1R) cyclopropane derivatives (Figure 2) and is 2.3 kcal
mol-1 more stable than B, which gives the (1S) isomers.
However, it is the energy of the “transition states” (TS)
4* and 5* between 11b and the incoming styrene
molecule that dictates the overall stereochemical out-
come of the reaction, rather than the orientation of the
ester chain alone (see below).27
In both conformers A and B, only one trajectory is
available for the approach of the olefin. This occurs
along a Ru-N vector, as attack along the Ru-P vectors
is blocked by the bulky PPh2 groups, and the ester group
blocks the remaining Ru-N trajectory. Figure 2 shows
how the choice of the enantioface of the olefin deter-
mines the configuration at C(2) of the cyclopropane
[Ru Cl(C(H)COOEt)(1b)]P F 6. Treatment of complex
3a with ethyl diazoacetate (1 equiv, CD2Cl2, room
temperature) gave only the starting five-coordinate
complex and a mixture of fumarate and maleate. In
contrast, complex 3b reacts with ethyl diazoacetate
under the above conditions to give the carbene deriva-
tive trans-[RuCl(C(H)COOR)((S,S)-1b)] (11b) (Scheme
4). As 11b is stable for 1-2 h at room temperature or
for ca. 12 h at -80 °C, it was characterized in solution
1
by H, 31P, and 13C NMR spectroscopy.
1
The H NMR spectrum of 11b features the signal of
the carbene hydrogen atom RudCH as a doublet of
doublets (δ 16.94, J P,H ) 7.3 and 13.1 Hz). The reso-
nance of the carbene 13C atom appears as a double of
doublets at δ 324 in the 13C NMR spectrum. The J P,C
coupling constants (25.3 and 31.7 Hz) indicate a cis
arrangement of the carbene ligand with respect to both
P atoms. The NMR data of the RuP2(dC(H)CO2Et)
fragment in 11b are similar to those of the related
complexes [RuCl2(dC(H)CO2R)P2]22 and rule out a trans
arrangement of the carbene and one phosphine ligand.
Furthermore, 11b was formed as a single isomer in
(23) In contrast, the reaction of 3b with ethyl diazoacetate at -50
°C gives two different carbene complexes 11b′ and 11b′′. In both
complexes the carbene ligand is cis to both phosphines, as indicated
by the P,C(carbene) coupling constants (see Experimental Part). Thus,
we assume that these species are the two cis-â isomers featuring a
trans C-Ru-N arrangement.
(24) Caulton, K. G. New J . Chem. 1994, 18, 25.
(25) Cerius2 uses the Universal Force Field (UFF): Rappe´, A. K.;
Casewit, C. J .; Colwell, K. S.; Goddard, W. A., III; Skiff, W. M. J . Am.
Chem. Soc. 1992, 114, 10024. Rappe´, A. K.; Colwell, K. S.; Casewit, C.
J . Inorg. Chem. 1993, 32, 3438.
(26) These values are in reasonable agreement with the experimen-
tal value of 1.861(3) Å found in [RuCl2(dCH-o-OMeC6H4)(PPh3)]2:
Kingsbury, J . S.; Harrity, J . P. A.; Bonitatebus, P. J .; Hoveyda, A. H.
J . Am. Chem. Soc. 1999, 121, 791.
(21) Doyle, M. P.; Winchester, W. R.; Hoorn, J . A. A.; Lynch, V.;
Simonsen, S. H.; Gosh, R. J . Am. Chem. Soc. 1993, 115, 9968.
(22) Schwab, P.; France, M. B.; Ziller, J . W.; Grubbs, R. H. Angew.
Chem. 1995, 107, 2179.
(27) In a MM study of the rhodium(II) carbamidato system, Doyle
has reported similar considerations.21

2106 Organometallics, Vol. 20, No. 10, 2001
Bachmann et al.
Ta ble 3. Rela tive En er gies of “Tr a n sition Sta tes”
F or m ed by [Ru Cl(C(H)COOR)(1b)]+ (11b) a n d
a
Styr en e (k ca l m ol-1
)
R
(1R,2S)
(1S,2S)
(1R,2R)
(1S,2R)
Et
0
11.2
0
4.4
12.6
1.3
7.4
14.0
2.8
7.7
19.5
8.3
But b
But c
a
b
Configurations refer to the cyclopropane product. Values
relative to (1R,2S)-4*, R ) Et. c Values relative to (1R,2S)-4*, R
) But.
of the ester chain, that is, B. This explains the inversion
of configuration at C(1) on going from (1R,2S)-4 to
(1S,2S)-5. The configuration at C(2) is the same as in
(1R,2S)-4, as the olefin approaches with the same
enantioface. In general, the (S)-configuration at C(1) is
disfavored, as (1S,2S)-5 and (1S,2R)-4 derive from the
attack along the sterically hindered trajectory passing
between a phenyl group of the ligand and the axial H
atom on the tertiary carbon of the cyclohexanediyl ring.
The energy gap of 7.7 kcal mol-1 between the TS’s of
the cis isomers, (1R,2S)-4* and (1S,2R)-4*, is larger than
that of 3.0 kcal mol-1 between the TS’s of the trans ones,
(1S,2S)-5* and (1R,2R)-5* (Table 3). This is in qualita-
tive agreement with the much higher enantioselectivity
observed for the cis isomer 4 than for 5. The calculated
energy ordering also reflects the experimental cis
selectivity. Furthermore, it also yields the experimen-
tally observed absolute configurations both for 4 and for
5. The model is also in agreement with the decreasing
cis selectivity on going from the ethyl to the tert-butyl
ester (Table 1, runs 4, 10), as the energy gap between
(1R,2S)-4* and (1S,2S)-5* decreases from 4.4 to 1.3 kcal
mol-1 in this series. Finally, the calculations account
also for the increasing enantioselectivity of 4 with
increasing steric bulk of the ester chain. Indeed, the
energy difference between the enantiomers of 4 in-
creases from 7.7 to 8.3 kcal mol-1 on going from Et to
But.
F igu r e 2. Minimized “transition states” between the
carbene complex [RuCl(C(H)COOEt)((S,S)-1b)] (11b) and
styrene (with relative energies), and absolute configura-
tions of the cyclopropanation product. Only relevant parts
of the complex are shown.
product. The configuration at C(1) results from the
orientation of the ester chain relative to the direction
of attack.
Next, we minimized the conformational energies of
the four diastereomeric “transition states” 4* and 5*.
The C(carbene)-C(2) and C(carbene)-C(1) distances
were kept fixed at 2.4 and 2.6 Å, respectively. The
double bond of the olefin was kept parallel to the Ru-C
vector and approximately perpendicular to the N-Ru
one (Figure 2). This mode of approach is generally
assumed1 in the interpretation of the stereochemical
outcome of cyclopropanation reactions.28 The relative
energy values summarized in Figure 1 show that the
energetically favored TS leads to (1R,2S)-4. This can be
explained by the observation that the vinyl H atoms of
the olefin and the H atom on the nearest sp3 carbon (on
the C6H10 ring) are oriented anti. A contact between the
phenyl group of styrene and that of the PPh2 group
disfavors the attack of the olefin with the opposite
enantioface, which leads to the minor cis isomer (1R,2R)-
5.
As a final observation, ligand 1b has the same
conformation in all structures and directs the approach
of the olefin by means both of the phenyl substituent
and of the cyclohexanediyl ring.
Con clu d in g Rem a r k s
The steric features of the PNNP ligand 1b account
for the remarkable enantio- and cis selectivity observed,
which are not far from the highest ever observed.1,12a,b
As compared with recent applications of phosphorus-
containing ligands in asymmetric cyclopropanation,29
complex 3b and Katsuki’s catalyst [RuCl(salen)(NO)]
(12) show exceptionally high cis selectivity in the
asymmetric cyclopropanation of terminal olefins. At
difference with the Ru- and Co-salen systems mentioned
above,12 3a ,b do not require chiral or bulky ester groups
or any additive to improve the selectivity. The molecular
modeling studies cast some light on the nature of the
cis selectivity in the case of ligand 1b. As shown in
The preferred TS that leads to a trans isomer is
(1S,2S)-5*, whose energy is 4.4 kcal mol-1 higher than
that of (1R,2S)-4*. It derives from the attack onto the
carbene complex 11b having the opposite conformation
(28) The only exception to this general assumption we are aware of
has been suggested by Kodadek to explain the cis selectivity observed
with rhodium porphyrins. According to his model, the olefin attacks
the carbene with the CdC double bond perpendicular to the M-C
vector and parallel to the plane of the porphyrin. However, the olefin
rotates before reaching the transition state, and the side-on approach
also leads to an end-on transition state in which both double bonds
are parallel.10
(29) (a) Song, J . H.; Cho, D. J .; J eon, S. J .; Kim, Y. H.; Kim, T. J .;
J eong, H. H. Inorg. Chem. 1999, 38, 893. (b) Brunel, J . M.; Legrand,
O.; Reymond, S.; Buono, G. J . Am. Chem. Soc. 1999, 121, 5807. (c)
Lee, H. M.; Bianchini, C.; J ia, G.; Barbaro, P. Organometallics 1999,
18, 1961. (d) Stoop, R. M.; Bauer, C.; Setz, P.; Wo¨rle, M.; Wong, T. Y.
H.; Mezzetti, A. Organometallics 1999, 18, 5691. (e) Bianchini, C.; Lee,
H. M. Organometallics 2000, 19, 1833. (f) Braunstein, P.; Naud, F.;
Pfaltz, A.; Rettig, S. J . Organometallics 2000, 19, 2676.

Cis-Selective Asymmetric Cyclopropanation
Organometallics, Vol. 20, No. 10, 2001 2107
[Ru Cl2(1a )], 2a . [RuCl2(PPh3)3] (635 mg, 0.66 mmol) and
1a (500 mg, 0.66 mmol) were refluxed in toluene (20 mL)
overnight. Partial evaporation of the solvent gave a bordeaux-
red solid that was recrystallized from CH2Cl2/pentane. Yield:
536 mg (90%). 1H NMR (250 MHz, CDCl3): 8.6 (d, 1H, Nd
CH, J ) 6.2), 7.6-6.7 (m, 34H, arom), 6.0 (d, 1H, J ) 6.2 Hz,
NdCH), 1.9 (s, 6H, 2 CH3). 31P NMR (101 MHz, CDCl3): 46.5
(s, 2P). MS (FAB+) (m/z): 930 ([M + 2 H]+, 100), 928 (M+, 50),
893 [(M - Cl]+, 22). Anal. Calcd for C52H42Cl2N2P2Ru‚0.5CH2-
Cl2: C, 64.92; H, 4.46; N: 2.88. Found: C, 65.10; H, 4.94; N,
2.58.
Ch a r t 2
[Ru Cl(1a )]P F 6, 3a . Complex 2a (185 mg, 0.199 mmol) and
Tl[PF6] (10 mg, 0.20 mmol) were stirred in CH2Cl2 (15 mL)
for 14 h. The precipitated TlCl was filtered off over Celite.
Addition of pentane to the resulting clear solution and partial
evaporation of the solvent gave a brown solid that was
recrystallized from CH2Cl2/pentane. The complex tenaciously
retains variable amounts of crystallization solvents, resulting
1
in erratic elemental analyses. Yield: 124 mg (60%). H NMR
(300 MHz, CDCl3): δ 8.90 (d, 1 H, J H,H′ ) 9.6 Hz, NdCH),
7.9-6.3 (m, 34H, arom), 4.40 (d, 1 H, J H,H′ ) 7.8 Hz, NdCH).
31P NMR: 80.8 (d, 2P, J P,P′ ) 27.6 Hz), 42.9 (d, 2P, J P,P ) 27.6
Hz). MS (FAB+) (m/z): 893 (M+, 100), 858, ([M - Cl]+, 6).
Typ ica l Ca ta lytic Ru n . Complex 3a (20 mg, 21 µmol, 5
mol %) and Tl[PF6] (7.3 mg, 21 µmol) were dissolved in CH2-
Cl2 (1 mL) and stirred overnight, and then TlCl was filtered
off and the resulting red-brown solution was added to the olefin
(0.42 mmol) and decane (50 mg, as internal standard). A
solution of the diazoester (0.42 mmol) in CH2Cl2 (1 mL) was
added with a syringe pump over 6 h. The solution, which was
protected from light throughout the reaction, was stirred for
an additional 14 h at room temperature and then analyzed by
GC. The reactions with a substrate-to-catalyst ratio of 100:1
(1 mol % catalyst) were performed using more olefin (2.1 mmol)
and diazoester (2.1 mmol). Yields were obtained by GC. In
selected instances, we isolated the cyclopropanation product
by column chromatography. The isolated yields are within
(10% of those obtained by GC.
Chart 2, 3b and 12 possess similar steric features
concerning the possible approach trajectories of the
incoming olefin. In both complexes, bulky groups (either
PPh2 or 1-phenyl-naphthyl) block the approach along
the Ru-P or Ru-O vector, and the cyclohexyl moiety
directs the approach along one Ru-N direction. The
presence of the latter stereogenic element is apparently
pivotal for high cis selectivity. Indeed, the apparently
related complex 13, in which the cyclohexenediyl moiety
is missing, prevalently forms the trans cyclopropane
derivative with selectivity in the range 66-98%.29a
Exp er im en ta l P a r t
Gen er a l Com m en ts. Reactions with air- or moisture-
sensitive materials were carried out under an argon atmos-
phere using Schlenk techniques. Styrene, (1S,2S)-(+)-1,2-
diaminocyclohexane, 2,5-dimethyl-2,4-hexadiene, and ethyl
chrysanthemate (30:70 mixture of cis and trans) were obtained
from Fluka AG. 2-(Diphenylphosphino)benzaldehyde was pur-
chased from Aldrich, Tl[PF6] from Strem Chemicals, and ethyl
trans-2-phenylcyclopropane from Lancaster. Enantiomerically
pure (S)-2,2′-diamino-6,6′-dimethylbiphenylene was obtained
from Solvias AG (Basel). Ligand 1b and complexes 2b and 3b
were prepared as previously reported.13b 1H and 31P NMR
A control reaction without the catalyst indicated that the
cyclopropane derivatives 4 and 5 were formed only in traces
under the conditions used. The yields of the uncatalyzed
reaction were 0.03% and 0.13% for the cis and trans products,
respectively.
An a lyt ic Det a ils (GC, ee, Ab solu t e Con figu r a t ion ).
Eth yl 2-P h en yl-cylopr opan ecar boxylate. Achiral GC analy-
sis: Macherey-Nagel SE 54, 30 m, He carrier (92 kPa).
Temperature program: 50 °C isotherm for 5 min, then to 200
°C at 5 °C min-1. Rt (min): styrene, 10.2; decane, 14.25; ethyl
cis-2-phenylcyclopropanecarboxylate, 28.2; ethyl trans-2-phe-
nylcyclopropanecarboxylate, 29.7. Chiral GC analysis: Supelco
Beta Dex 120, 1.4 mL He min-1; temperature program: 110
°C for 10 min, 5 °C min-1 to 150 °C, isotherm for 20 min. Rt
(min): cis-(1R,2S), 26.0; cis-(1S,2R), 26.38; trans-(1R,2R),
28.09; trans-(1S,2S), 28.34. The GC peaks were attributed by
comparison with an authentic sample with a known E/Z ratio
(99:1, Lancaster). The absolute configurations are (1R,2S) for
4 and (1S,2S) for 5 (by the sign of the optical rotation of the
isolated products).30
1
spectra were recorded on Bruker DPX spectrometers. H and
13C positive chemical shifts in ppm are downfield from tet-
ramethylsilane. 31P NMR spectra were referenced to external
85% H3PO4. Mass spectra were measured by the MS service
of the Laboratorium fu¨r Organische Chemie (ETH Zu¨rich). A
3-NOBA (3-nitrobenzyl alcohol) matrix and a Xe atom beam
with a translational energy of 8 keV were used for FAB+ MS.
Optical rotations were measured using a Perkin-Elmer 341
polarimeter with a 1 dm cell. Elemental analyses were carried
out by the Laboratory of Microelemental Analysis (ETH
Zu¨rich). Molecular modeling calculations were performed on
a Silicon Graphics O2 platform with the Cerius2 program (MSI)
using standard UFF settings.
ter t-Bu tyl 2-P h en ylcyclop r op a n eca r boxyla te. Achiral
GC analysis: Macherey-Nagel SE 54, 30 m, He carrier (92
kPa). Temperature program: 50 °C isotherm for 5 min, then
to 200 °C at 5 °C min-1. Rt (min): styrene, 10.0; decane, 14.10;
tert-butyl cis-2-phenylcyclopropanecarboxylate, 29.95; tert-
butyl trans-2-phenylcyclopropanecarboxylate, 31.29. Chiral GC
analysis: Supelco Beta Dex 120, 1.4 mL He min-1; tempera-
ture program: 110 °C for 10 min, 5 °C min-1 to 150 °C,
isotherm for 20 min. Rt (min): cis-(1R,2S), 27.68; cis-(1S,2R),
28.06; trans-(1R,2R), 31.08; trans-(1S,2S), 31.26. Product
isolated from a catalysis run was used for GC calibration. The
(S)-N,N′-Bis[2-(d ip h en ylp h osp h in o)ben zylid en e]-2,2′-
d ia m in o-6,6′-d im eth ylbip h en ylen e, (S)-1a . (S)(-)-2,2′-Di-
amino-(S)-6,6′-dimethylbiphenylene (292 mg, 1.38 mmol) and
2-diphenylphosphinobenzaldehyde (800 mg, 2.76 mmol) were
refluxed in toluene (15 mL) in a Dean-Stark apparatus for 14
h. Evaporation of the solvent in a vacuum and recrystallization
of the resulting yellow oil from toluene/MeOH (4:1) gave a
yellow solid. Yield: 998 mg (96%). Mp: 69 °C. [R]D20: -224.9
1
( 2 (c ) 1, CHCl3). H NMR: δ 8.9 (d, 2H, NdCH, J P,H ) 5.5
Hz), 7.7-6.75 (m, 32H, arom.), 6.3 (d, 2H, arom.), 1.9 (s, 6H,
2CH3). 31P NMR: δ -13.6 (s, 2P). IR (KBr, cm-1): 1624 (s,
ν
CN). MS (FAB+) (m/z): 757 ([M + H]+, 61), 756 (M+, 24), 468
([M + H - NdCPPh2]+, 100).
(30) Krieger, P. E.; Landgrebe, J . A. J . Org. Chem. 1978, 43, 4447.

2108 Organometallics, Vol. 20, No. 10, 2001
Bachmann et al.
absolute configuration of 5 is (1S,2S), as indicated by the
positive sign of the optical rotation of the isolated product.31
The absolute configuration of 4 is assumed to be (1R,2S) on
the basis of GC retention times and by analogy with the ethyl
derivative.
Complex 2b (40 mg, 48 µmol) and Tl[PF6] (17 mg, 48 µmol)
were stirred at room temperature overnight in 3 mL of CH2-
Cl2. TlCl was filtered over Celite, CH2Cl2 was removed under
vacuum, and CD2Cl2 and ethyl diazoacetate (5 µL, 48 µmol)
were added. 1H NMR (250 MHz, CD2Cl2): 16.9 (d×d, 1 H, J P,H
) 7.3 Hz, J P′,H ) 13.1 Hz). 31P NMR (101 MHz, CD2Cl2): 37.6
(d, 1P, J P,P′ ) 29.8 Hz), 28.1 (d, 1P, J P,P′ ) 29.8 Hz), -144.4
(septet, 1 P, J P,F ) 714 Hz, PF6). 13C NMR (75 MHz, CD2Cl2,
-80 °C): δ 324.1 (d×d, 1 C, J P,C ) 25.3, J P′,C ) 31.7 Hz).
Eth yl Ch r ysa n th em a te. Achiral GC analysis: Macherey
Nagel SE 54, 30 m, He carrier (92 kPa). Temperature
program: 50 °C isotherm for 5 min, then 3 °C min-1 to 150
°C, 10 °C min-1 to 200 °C. Rt (min): 8, 9.95; decane, 16.75;
ethyl cis-chrysanthemate, 30.7; ethyl trans-chrysanthemate,
31.1. The GC peaks were attributed by comparison with a
commercially available authentic sample with a known E/Z
ratio (70:30, as determined by 1H NMR spectroscopy). The
enantiomeric excess of 9 was determined by trans-esterifica-
tion with (-)-menthol according to the following procedure.
The crude reaction solution was eluted over alumina with
hexane/ethyl acetate (9:1). Evaporation of the solvents gave a
colorless oil. A portion (20 mg) of this oil was dissolved in a
toluene solution (0.7 mL, 0.25 M) of pyridine (0.175 mmol).
SOCl2 (0.7 mL of a 0.7 M toluene solution, 0.5 mmol) and (-)-
menthol (0.7 mL of a 1.4 M toluene solution, 0.98 mmol) were
added thereto. The mixture was refluxed for 1 h and then
diluted with Et2O and extracted three times with 0.1 M
phosphate buffer (Na3PO4, pH ) 3), followed by saturated
NaHCO3. The crude product was concentrated and analyzed
by GC. GC analysis: Macherey-Nagel SE 54, 30 m, He carrier
Performing the same reaction at -50 °C gave a mixture of
two different carbene complexes, 11b′ (43%) and 11b′′ (37%),
along with 11b (20%). 11b′: 1H NMR (250 MHz, CD2Cl2): 15.5
(d×d, 1 H, J P,H ) 12.5 Hz, J P′,H ) 5.2 Hz). 31P NMR (202 MHz,
CD2Cl2, -50 °C): 44.1 (br, 1P) (data from a P,H correlation
experiment, no cross-peak for the second P atom due to small
J
P,H). 11b′′: 1H NMR (250 MHz, CD2Cl2): 15.0 (d×d, 1 H, J P,H
) 10.0 Hz, J P′,H ) 4.8 Hz). 31P NMR (202 MHz, CD2Cl2, -50
°C): 45.2 (br, 1P) (see 11b′ for signal of second P atom).
Molecu la r Mod elin g. Molecular modeling calculations
were performed on a Silicon Graphics O2 platform with the
Cerius2 program (MSI) (standard UFF settings). The two
conformations of the ester chain in the carbene intermediate
[RuCl(C(H)COOEt)(3b)] were minimized first. The conforma-
tion A is more stable by 2.3 kcal mol-1 than B. With tert-butyl
as the ester group, the energy difference between A and B
increases to 4.0 kcal mol-1. Then, the transition states were
simulated by positioning the styrene molecule with fixed
C(carbene)-C(2) and C(carbene)-C(1) distances (2.4 and 2.6
Å, respectively). During minimization, the geometry of the
RudC‚‚‚C(1)dC(2) moiety was kept fixed with C(carbene)-
C(2) and carbene-C(1) distances of 2.4 and 2.6 Å, respectively.
The positions of ruthenium and of the P atom anti to the
incoming olefin were also kept constant. The final energy
values were calculated with the contribution of all atoms.
Relative energy values are given in Table 3.
(92 kPa). Temperature program: 50 °C to 200 °C at 1 °C min-1
.
Rt (min): 142.8 (1S,3R), 144.6 (1R,3S). For run 1 of Table 2,
the enantiomeric excess of 9 was independently determined
by integration of the 1H NMR spectrum of the mixture of 9
and 10 in the presence of the chiral shift reagent tris[3-
(propylhydroxymethylene)-d-camphorato]europium(III). Inte-
gration of the cyclopropane methyl signal at δ 1.68 gave the
same value (80% ee). As isolated 9 has negative R, its absolute
configuration is (1S,3R) by correlation with the optical rotation
of the acid.32
Obser va tion of [Ru Cl(C(H)COOEt)((S,S)-1b)]P F 6, 11b.
Ackn owledgm en t. Financial Support from the Swiss
(31) Fukada, T.; Katsuki, T. Tetrahedron 1997, 21, 7201.
(32) Aratani, T.; Yoneyoshi, Y.; Nagase, T. Tetrahedron Lett. 1977,
30, 2599.
National Science Foundation is gratefully acknowledged.
OM010020P