Dinuclear ruthenium(I) complexes of the type [Ru2(CO) 4L2] with carboxylate or 2-pyridonate ligands: Evaluation as catalysts for olefin cyclopropanation with diazoacetates

Article abstract of DOI:10.1016/j.jorganchem.2005.06.043

Dinuclear ruthenium(I,I) carboxylate complexes [Ru₂(CO) ₄(μ-OOCR)₂]_n (R = CH₃ (1a), C₃H₇ (1b), H (1c), CF₃ (1d)) and 2-pyridonate complex [Ru₂(CO)₄(μ-2-pyridonate)₂] _n (3) catalyze efficiently the cyclopropanation of alkenes with methyl diazoacetate. High yields are obtained with terminal nucleophilic alkenes (styrene, ethyl vinyl ether, α-methylstyrene), medium yields with 1-hexene, cyclohexene, 4,5-dihydrofuran and 2-methyl-2-butene. The E-selectivity of the cyclopropanes obtained from the monosubstituted alkenes and the cycloalkenes decreases in the order 1b > 1a > 1d > 1c. The cyclopropanation of 2-methyl-2-butene is highly syn-selective. Several complexes of the type [Ru₂(CO)₄(μ-L¹) ₂]₂ (4) and (5), [Ru₂(CO)₄(μ- L¹)₂L²] (L² = CH₃OH, PPh₃) (6)-(9) and [Ru₂(CO)₄(CH ₃CN)₂(μ-L¹)₂] (10) and (11), where L¹ is a 6-chloro- or 6-bromo-2-pyridonate ligand, are also efficient catalysts. Compared with catalyst 3, a halogen substituent at the pyridonate ligand affects the diastereoselectivity of cyclopropanation only slightly.

Full text of DOI:10.1016/j.jorganchem.2005.06.043

Journal of Organometallic Chemistry 690 (2005) 5562–5569
www.elsevier.com/locate/jorganchem
Dinuclear ruthenium(I) complexes of the type [Ru2(CO)4L2]
with carboxylate or 2-pyridonate ligands: Evaluation as catalysts
for olefin cyclopropanation with diazoacetates
*
Thorsten Werle, Lutz Schaffler, Gerhard Maas
¨
Abteilung Organische Chemie I, Universita¨t Ulm, Albert-Einstein-Allee 11, D-89081 Ulm, Germany
Received 13 May 2005; accepted 30 June 2005
Available online 19 August 2005
Dedicated to Manfred Regitz on the occasion of his 70th birthday.
Abstract
Dinuclear ruthenium(I,I) carboxylate complexes [Ru2(CO)4(l-OOCR)2]n (R = CH3 (1a), C3H7 (1b), H (1c), CF3 (1d)) and
2-pyridonate complex [Ru2(CO)4(l-2-pyridonate)2]n (3) catalyze efficiently the cyclopropanation of alkenes with methyl diazoacetate.
High yields are obtained with terminal nucleophilic alkenes (styrene, ethyl vinyl ether, a-methylstyrene), medium yields with 1-hex-
ene, cyclohexene, 4,5-dihydrofuran and 2-methyl-2-butene. The E-selectivity of the cyclopropanes obtained from the monosubsti-
tuted alkenes and the cycloalkenes decreases in the order 1b > 1a > 1d > 1c. The cyclopropanation of 2-methyl-2-butene is highly
syn-selective. Several complexes of the type [Ru2(CO)4(l-L1)2]2 (4) and (5), [Ru2(CO)4(l-L1)2L2] (L2 = CH3OH, PPh3) (6)–(9)
and [Ru2(CO)4(CH3CN)2(l-L1)2] (10) and (11), where L1 is a 6-chloro- or 6-bromo-2-pyridonate ligand, are also efficient catalysts.
Compared with catalyst 3, a halogen substituent at the pyridonate ligand affects the diastereoselectivity of cyclopropanation only
slightly.
ꢀ 2005 Elsevier B.V. All rights reserved.
Keywords: Ruthenium complexes; Catalysis; Cyclopropanation; Diazo compounds; Metal–carbene intermediates
1. Introduction
dinuclear rhodium(II,II) complexes of the type
[Rh2(l-L)4], where L represents bidentate carboxylate,
amidate, or phosphate ligands, and it is well known
that ligand tuning, without changing the coordination
motif of the complex, can be used to influence, e.g.,
the chemoselectivity and enantioselectivity of several
types of carbene transfer reactions [1–5]. For the cata-
lytic cyclopropanation of olefins with diazo com-
pounds, the diastereoselectivity is an important
aspect. Systematic comparisons [1,6] have shown that
this particular selectivity issue can be controlled only
to a limited extent by the catalytically active metal
and its ligands, and those catalysts are rare which pro-
mote a highly trans- or (even more difficult) cis-selectiv-
ity in the cyclopropanation of a simple olefin such as
styrene with (m)ethyl diazoacetate.
A variety of rhodium and copper catalysts are
widely used for several different carbene transfer reac-
tions involving aliphatic diazo compounds, e.g., cyclo-
propanation, cyclopropenation, C–H and X–H
insertion, and ylide formation [1]. Since the catalytic
transformations proceed through short-lived transition
metal–carbene intermediates, the variation of ligands
at the metal is expected to have an impact on the reac-
tivity and selectivity of those intermediates. For exam-
ple, most of the currently used rhodium catalysts are
*
Corresponding author. Tel.: +49 731 5022790; fax: +49 731
5022803.
E-mail address: gerhard.maas@uni-ulm.de (G. Maas).
0022-328X/$ - see front matter ꢀ 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2005.06.043

T. Werle et al. / Journal of Organometallic Chemistry 690 (2005) 5562–5569
5563
Recently, ruthenium complexes have emerged as no-
R
R
O
vel catalysts for olefin cyclopropanation with diazo
compounds [7] and it appears that the rich coordination
chemistry of ruthenium may offer a larger range of
potentially useful catalysts than in the case of rhodium.
However, dimeric ruthenium(II) acetate, [Ru2(OOC-
Me)4], which represents the immediate structural relative
of the highly useful rhodium(II) acetate dimer but is
electronically different (presence of a Ru–Ru double
bond, two unpaired electrons), appears to be a less effec-
tive cyclopropanation catalyst due to the competing
metathetical activity of the ruthenium–carbene interme-
diate [8]. In contrast, cyclopropanation of cyclooctene
with ethyl diazoacetate was quantitative when catalyzed
by the trifluoroacetate complex [Ru2(OOCCF3)4] [9].
We have identified the dinuclear ruthenium(I,I)
acetate complexes [Ru2(CO)4(l-OAc)2]n, which is a
coordination polymer, and the related bis(acetonitrile)
complex [Ru2(CO)4(CH3CN)2(l-OAc)2] as well suited
catalysts for cyclopropanation of olefins with alkyl dia-
zoacetates [10,11], a-silyl-a-diazoacetates [10,12] and
aryl- or silyl-diazomethane derivatives [13]. The struc-
turally similar triazenide complexes [Ru2(CO)6(l-ArN-
NNAr)2] were also found to catalyze efficiently and
effectively the cyclopropanation of nucleophilic terminal
alkenes, but were less suited for the cyclopropanation of
internal alkenes [14].
O
O
O
O
R
R
O
O
O
Ru
Ru
L
Ru
Ru
L
C
C
C
C
O
O
O
O
C
O
C
O
C
O
C
O
n
1
2
1,2
a
b
c
d
L = ligand (see text)
R
CH3 C3H7
H
CF3
Scheme 1.
C3H7COOH (excess), ∆
(1)
Ru3(CO)12
[ Ru2(CO)4(OOC C3H7)2 ]
n
- CO, - H2
1b
CF3COOH (excess), ∆
(2)
2a
(L =CH3CN)
[Ru2(CO)4(CH3CN)2(OAc)2]
- CH3COOH
160 oC/
0.001 mbar
1d
2d (L = CH3CN)
Scheme 2.
Encouraged by the good performance of [Ru2(CO)4-
(l-OAc)2]n in catalytic olefin cyclopropanation with dia-
zoacetate, we decided to investigate the effect of ligand
variation on the effectiveness and diastereoselectivity
of these cyclopropanation reactions. To this end, we
examined several catalysts of the type [Ru2(CO)4-
(l-L)2], where L is a bidentate carboxylate or 2-pyrido-
nate (pyridin-2-olate) ligand.
plexation of the acetonitrile ligands from the initially
formed 2d (L = CH3CN) [17] (Scheme 2, Eq. (2)). It
has been reported [18] that 1d is obtained as a ‘‘white
polymer’’ from the reaction of Ru3(CO)12 with excess
trifluoroacetic acid in toluene at 90 ꢁC. This product
was characterized by IR spectroscopy only and was di-
rectly converted into the bis(triphenylphosphane) ad-
duct 2d (L = PPh3). The color of the ‘‘white polymer’’,
its solubility in diethyl ether, and an IR absorption at
2153 cmꢀ1, which is typical for axial carbonyl ligands
in complexes of type 2 [15], are not in accord with the
properties of 1d prepared by us, and we assume that
these authors had prepared the hexacarbonyl complex
2d (L = CO). In fact, we found that heating of
Ru3(CO)12 with trifluoroacetic acid at reflux tempera-
ture yields 2d (L = CO) which even after prolonged
treatment at 180 ꢁC/0.001 mbar does not completely re-
lease the two axial CO ligands.
2. Results and discussion
2.1. Catalysts
In this study, two series of dinuclear ruthenium(I)
complexes are investigated as cyclopropanation cata-
lysts: the di(l-carboxylato)-tetracarbonyl-diruthenium
complexes 1a–d (Scheme 1) and the di-(l-pyridin-2-ola-
to)-tetracarbonyl complexes 3–11 (Scheme 3).
The carboxylato complexes 1a and 1c are already
known [15]. We have prepared butyrato complex 1b
analogously by heating of Ru3(CO)12 in an excess of bu-
tyric acid (Scheme 2, Eq. (1)). In a related published pro-
cedure, 1b was not isolated but directly converted into
its bis(triphenylphosphane) adduct (2b, L = PPh3) [16].
The orange-red trifluoroacetato complex 1d was ob-
tained from bis(acetonitrile) complex [Ru2(CO)4(CH3-
CN)2(l-OAc)2] (2a, L = CH3CN) by ligand exchange
with trifluoroacetic acid, followed by thermal decom-
Like 1a and 1c, the new complexes 1b and 1d are
coordination polymers which dissolve with depolymer-
ization only in solvents with good donor properties,
e.g., acetonitrile and DMSO, yielding complexes of type
2. Therefore, the NMR data reported in Section 4 are in
fact those of 2b and 2d (L = CD3CN).
Di-(l-pyridin-2-olato)-tetracarbonyl-diruthenium (3)
can be prepared from Ru3(CO)12 and 2-hydroxypyridine
in hot toluene; like the carboxylato complexes, it exists
as a coordination polymer which is depolymerized by

5564
T. Werle et al. / Journal of Organometallic Chemistry 690 (2005) 5562–5569
N2CHCO2Me
R1
R3
R2
H
catalyst (1-3 mol%)
- N2
R
N
O
N
O
CO
O
CO
O
N
R2
H
R3
R1
H
Ru
Ru
Ru
Ru
R
H
N
H
C
C
O
O
C
O
C
O
+
C
O
C
O
R2
CO2Me
R1
R3
CO2Me
2
n
(Z)-12
(E)-12
3
4: R = Cl
5: R = Br
H
CO2Me
+
MeO2C
R
N
O
R
CH3C
N
O
O
E/Z-13
CO2Me
CO
L
CO
O
N
CO2Me
Ru
Ru
N
C
Ru
Ru
R
CCH3
R
N
N
C
O
O
C
O
C
O
C
O
C
O
O
10: R = Cl
11: R = Br
14
15
6: R = Cl, L = CH3OH
7: R = Cl, L = PPh3
8: R = Br, L = CH3OH
9: R = Br, L = PPh3
Scheme 4.
so that reliable yields of isolated products could be ob-
tained. In spite of the very slow addition of the diazo es-
ter to an excess of liquid alkene (10 molar equivalents)
that contained the catalyst, formation of the formal car-
bene dimers, dimethyl fumarate and maleate (E- and Z-
13) was a competitive pathway which accounted almost
for the complete remaining material balance in most
cases. In specific cases, small amounts of other products
were also identified, such as the (dihydrofuranyl)acetate
14 in the carbene transfer reaction with 4,5-dihydrofu-
ran and the C/H insertion product 15 in the case of
2,3-dimethyl-2-butene.
The polymeric complexes 1a–d and 3, which are not
soluble in the liquid alkene, are obviously depolymerized
by the diazoester, as indicated by the formation of a
homogeneous solution after addition of small amounts
of the diazoester. Thus, all catalyses reported here occur
under homogeneous conditions.
Cyclopropanation reactions with catalysts 4–11
(3 mol%) were performed on an analytical scale, and
the alkene was diluted with dichloromethane. For com-
parison, catalyst 3 was also applied under these condi-
tions (Table 2). While complexes 7 and 9 were well
soluble in CH2Cl2, all other complexes went into solu-
tion after addition of a small amount of the diazoester.
Since we were mainly interested to see the influence of
the halogen substituent of the pyridonate ligands on
the diastereoselectivity of cyclopropanation, the experi-
ments were focussed on three olefins with different
degree of substitution (styrene, cyclohexene and 2-
methyl-2-butene).
Scheme 3.
good donor solvents and other Lewis bases [19]. For the
current investigation, we have prepared and structurally
characterized a series of new 6-halopyridin-2-olato com-
plexes 4–11 (Scheme 3) [20]. It should be noted that
these complexes can exist with a head–head as well as
a head–tail arrangement of the two bridging pyridonate
ligands. While complex 3, containing the parent pyridin-
2-olate ligand, exists in the head–tail form, the 6-chloro-
and 6-bromopyridin-2-olato complexes 4 and 5 show a
head–head arrangement and exist as centrosymmetric
coordination dimers in the solid state. Treatment of 4
and 5, respectively, with Lewis bases such as methanol
and PPh3 generates the ‘‘monomeric’’ dinuclear com-
plexes 6–9. In all head–head complexes 4–9, only one
ruthenium atom assumes the complete octahedral coor-
dination while a free coordination site remains at the
ruthenium atom that is shielded by the two neighboring
halogen atoms. Interestingly, the head–tail arrangement
of 6-halopyridin-2-olato complexes can be achieved
when sterically little demanding acetonitrile ligands oc-
cupy the axial site at each ruthenium atom (10 and 11).
2.2. Cyclopropanation studies
The cyclopropanation of a series of representative
olefins with methyl diazoacetate catalyzed by ruthenium
carboxylate complexes 1b–d or pyridonate complex 3
was investigated and compared with the published [10]
results obtained with ruthenium acetate 1a as catalyst
(Scheme 4 and Table 1). These experiments were con-
ducted on a preparative scale (20 mmol of diazoacetate)
In terms of cyclopropane yields, Tables 1 and 2 show
that [Ru2(CO)4(l-OAc)2]n (1a) is the most effective of all
investigated catalysts. For the carboxylato-ruthenium

T. Werle et al. / Journal of Organometallic Chemistry 690 (2005) 5562–5569
5565
Table 1
Cyclopropanation of alkenes with methyl diazoacetate catalyzed by 1a–d or 3 (see Scheme 4)a
Entry
Alkene and cyclopropane 12
Yield of cyclopropanes 12, % (E/Z ratiob)c
R1
R2
R3
R4
Catalyst 1ad
Catalyst 1b
Catalyst 1c
Catalyst 1d
Catalyst 3
1
2
3
4
5
6
7
8
a
Ph
Bu
OEt
–(CH2)4–
–(CH2)2O–
Ph
Me
Me
H
H
H
H
H
H
H
H
H
H
H
H
H
H
Me
95 (1.6)
67 (2.0)
89 (4.5)
68 (3.7)
56 (>97:3)
91 (0.67)
61 (0.16)
47
78 (1.9)
55 (2.8)
80 (2.3)
45 (3.6)
50 (6.9)
82 (1.0)
38 (0.12)
28
70 (1.1)
32 (1.4)
62 (0.83)
36 (1.8)
37 (>97:3)
82 (0.59)
31 (0.19)
23
72 (1.5)
54 (1.8)
71 (2.0)
43 (2.2)
41 (>97:3)
81 (0.77)
36 (0.28)
27
78 (1.8)
52 (1.4)
72 (1.4)
23 (3.1)
52 (5.3)
89 (0.91)
33 (0.14)
18
H
H
Me
Me
Me
Me
Me
In neat alkene, T = 20 ꢁC; catalyst:diazoacetate:alkene = 0.01:1:10 (Section 4.3.1).
Anti/syn for cycloalkenes.
Additional products: Dimers Z- and E-13 were formed in all cases; in the following cases, E/Z-13 constituted the major reaction product (catalyst,
b
c
yield of Z-13 (%), yield of E-13 (%)): entry 2: 1c, 41, 16; entry 4: 1b, 35, 14; 1c, 28, 17; 1d, 35, 13; 3, 49, 21; entry 5: 1c, 22, 15; 1d, 29, 12; entry 7: 1b, 41,
16; 1c, 45, 13; 1d, 44, 18; 3, 49, 17; entry 8: 1a, 36, 15; 1b, 42, 16; 1c, 44, 16; 1d, 48, 19; 3, 50, 28. Entry 5: Methyl (2,3-dihydro-4-furyl)acetate (14) was
also formed in 6–10% yield. Entry 8: Methyl 4,5-dimethylhex-4-enoate (15) was also formed in 4 (1b, 1d), 8 (3), and 10% (1c) yield.
d
Taken from lit. [10].
Table 2
Cyclopropanation of alkenes with methyl diazoacetate catalyzed by 3–11a
Catalyst
Yield of cyclopropanes,b % (E/Z or anti/syn ratiob)
From styrene
From cyclohexene
From 2-methyl-2-butene
[Ru2(CO)4(pyO)2]n (3)
73 (1.74)
63 (1.52)
70 (1.41)
63 (1.66)
66 (1.94)
75 (1.44)
38 (1.96)
69 (1.66)
70 (1.39)
24 (2.44)
54 (1.99)
49 (2.12)
52 (2.01)
18 (2.91)
56 (1.96)
15 (3.04)
44 (2.02)
35 (2.00)
33 (0.14)
51 (0.15)
73 (0.16)
64 (0.15)
15 (0.18)
69 (0.15)
11 (0.19)
34 (0.14)
64 (0.16)
[Ru2(CO)4(2-Cl-pyO)2]2 (4)
[Ru2(CO)4(2-Br-pyO)2]2 (5)
[Ru2(CO)4(2-Cl-pyO)2(MeOH)] (6)
[Ru2(CO)4(2-Cl-pyO)2(PPh3)] (7)
[Ru2(CO)4(2-Br-pyO)2(MeOH)] (8)
[Ru2(CO)4(2-Br-pyO)2(PPh3)] (9)
[Ru2(CO)4(2-Cl-pyO)2(CH3CN)2] (10)
[Ru2(CO)4(2-Br-pyO)2(CH3CN)2] (11)
a
In CH2Cl2, catalyst:diazoacetate:alkene = 0.03:1:10, T = 22 ꢁC (Section 4.3.2).
Determined by GC.
b
catalysts, the yields decrease in general in the order ace-
methyl-2-butene and 2,3-dimethyl-2-butene), yields are
lower than expected with a view to the increased elec-
tron density of the olefinic bond as compared to, e.g.,
1-hexene. It is likely that for the more highly substituted
C@C bonds, increased steric hindrance overrides the
favorable aspect of higher nucleophilic character of
these olefins. A comparison with the literature values
[6] suggests that the metal carbene intermediates derived
from Rh2(OOCR)4 catalysts are less sensitive and those
derived from common copper catalysts are more sensi-
tive to these steric effects than the ruthenium carbenes
derived from 1a–d and 3.
tate (1a) > butyrate (1b) ꢁ trifluoroacetate (1d) > for-
mate (1c). With the (2-pyridonato)-ruthenium complex
3, yields are similar to 1b, except for the distinctly lower
yields in the case of cyclohexene and 2,3-dimethyl-2-bu-
tene. Among the complexes with 6-halopyridonate li-
gands (4–11), the 6-bromopyridonate complexes 5, 8
and 11 in general provide somewhat higher yields than
their 6-chloropyridonate counterparts; in particular,
the trisubstituted double bond of 2-methyl-2-butene is
cyclopropanated in higher yields (65–73%) with these
catalysts than even with acetato complex 1a.
The effectiveness of cyclopropanation also depends
on the nature of the olefin. In this respect, the same
qualitative results are obtained with all ruthenium cata-
lysts used in this study: The highest yields are obtained
uniformly with nucleophilic mono- and 1,1-disubsti-
tuted alkenes, in the sequence a-methylstyrene > sty-
rene > ethyl vinyl ether > 1-hexene. The cyclic olefins
cyclohexene and 4,5-dihydrofuran give intermediate
yields. With tri- and tetrasubstituted alkenes (2-
For the diastereoselectivity of the cyclopropanation
reactions, the dependence on the nature of the olefin is
the same with all ruthenium carboxylate or pyridonate
complexes investigated here. With the exception of 4,5-
dihydrofuran, where the syn-cyclopropane may not be
stable to the catalyst (syn ! anti isomerization as well
as ring-opening leading to 14 [21]), the thermodynami-
cally favored E (or anti) isomer is formed preferentially
from the monosubstituted and 1,2-cis-substituted

5566
T. Werle et al. / Journal of Organometallic Chemistry 690 (2005) 5562–5569
alkenes. While this is the common result with most
cyclopropanation catalysts and the E/Z ratios are in
typical ranges [1,6], the high preference for the syn-
cyclopropane formed from 2-methyl-2-butene (up to
88:12) as well as other trisubstituted alkenes [11,14] is
a unique feature of the [Ru2(CO)4(l-L)2] catalysts.
As Table 1 shows, the diastereoselectivity of cyclo-
propanation is only modestly dependent on the bridging
ligands of the dinuclear ruthenium complexes. Leaving
aside the cases of ethyl vinyl ether and 4,5-dihydrofuran
(where the formed cyclopropanes may be subject to sub-
sequent reaction with the catalyst, see above), the
E-selectivity observed with the terminal alkenes and
cyclohexene in general decreases in the sequence
1b > 1a > 1d > 1c (i.e., butyrate > acetate > trifluoroace-
tate > formate ligands). For 2-methyl-2-butene, the Z-
selectivity decreases in the series 1b > 1a > 1c > 1d. The
stereoselectivities observed with (2-pyridonate)-ruthe-
nium catalyst 3 are found somewhere between those of
1b and 1c.
An often used working hypothesis for the ligand-
dependent chemoselectivity of carbenoid reactions using
dirhodium tetracarboxylate catalysts emphasizes the
influence of electron-withdrawing ligands on the electro-
philic character of the rhodium–carbene intermediate
[2]. Considering the resonance structures LnM@
CR1R2 M LnMꢀ–+CR1R2, it can be argued that elec-
tron-withdrawing ligands enhance the weight of the po-
lar resonance structure due to delocalization of the
negative charge, thereby rendering the metal–carbene
more reactive and less selective. This hypothesis also
seems to work for the diastereoselectivity of rhodium-
catalyzed cyclopropanation reactions; e.g., the E (anti)
selectivity for cyclopropanation of styrene, ethyl vinyl
ether and cyclohexene decreases in the order Rh2-
(acetamide)4 > Rh2(acetate)4 > Rh2(trifluoroacetate) >
Rh2(perfluorobutyrate)4 [6c].
Within the series of carboxylato-ruthenium catalysts
1a–d, however, the formato complex 1c does not fit into
the simple correlation between decreasing diastereose-
lectivity and increasing electron-withdrawing power of
the carboxylate ligand, i.e. lower pKa value of the corre-
sponding carboxylic acid. Although we do not wish to
speculate on the special behavior of 1c, we consider it
likely that the ligand effects on the electrophilic charac-
ter and the stabilization of a metal-bound carbene inter-
mediate are different when the carbene is bound to a
[Rh2(OOCR)4] or a [Ru2(CO)4(OOCR)2] unit. It has
been shown that in rhodium-catalyzed carbenoid reac-
tions, the chemoselectivity is not only correlated with
the electron-withdrawing properties of the carboxylate
ligands but also with their polarizability, and the rele-
vance of metal-to-ligand backbonding (ligand = axial
carbene or carbonyl) has been discussed [5]. In the
D4h-symmetric rhodium case, effects from all four li-
gands uniformly act on the axially coordinated carbene
moiety. In the ruthenium case, the ligand effects and
backbonding affect not only the axial carbene ligand
but also the carbonyl ligands at the dinuclear metal core.
The diastereoselectivities of cyclopropanation
achieved with the pyridonate complexes 3–11 vary only
in a narrow range (Table 2); e.g., for styrene cycloprop-
anation, the E:Z ratio is between 66.2:33.8 (=1.96) and
58.2:41.8 (=1.39). Thus, the introduction and variation
of the halogen substituent at the pyridonate ligand as
well as the structure of the complex (head–head as in
3, 10, 11 vs. head–tail as in 4–9) have only a minor influ-
ence. X-ray crystal structure determination of most of
the head–head complexes 4–9 have shown [20] that the
ruthenium atom which is close to the halogen substitu-
ents (see Scheme 3) keeps a vacant coordination site
even if a small ligand such as methanol would have been
available during the synthesis. Therefore, it could be ar-
gued that steric shielding by the two neighboring halo-
gen atoms also prevents the coordination of the
diazoacetate or the derived methoxycarbonylcarbene
moiety at this ruthenium atom. Coordination of the car-
bene would take place, after displacement of the axial li-
gand, at the ruthenium atom surrounded by two oxygen
and two carbonyl ligands (compare formulae 6–9 in
Scheme 3, L = carbene), and no steric influence of the
halogen substituents on the diastereoselectivity of cyclo-
propanation would exist. Whether the halogen substitu-
ents modulate the electronic ligand effects of the
pyridonate groups and thus the diastereoselectivity of
cyclopropanation, as discussed above, cannot be con-
cluded firmly from the data of Table 2. For styrene in
particular, there are small but significant differences
depending on whether the chloropyridonate complexes
(4) and (6) or bromopyridonate (5) and (8) complexes
were used.
The results obtained with catalysts 7 and 9 suggest
that additional or different sources of diastereoselec-
tion are present. If the preceding assumption about
the constitution of the metal–carbene intermediate
was correct, the same intermediate should result from
all precursor complexes with a given pyridonate li-
gand, e.g., 4, 6 and 7 (i.e., coordination dimer, meth-
anol and PPh3 complex), and identical diastereomer
ratios of the formed cyclopropanes should result.
However, the PPh3 complexes 7 and 9 give rise to sig-
nificantly higher E/Z (anti/syn) ratios with styrene and
cyclohexene than the related MeOH complexes 6 and
8. Currently, we can only speculate about the reasons.
Perhaps, the carbene ligand does not replace the axial
PPh3 ligand in complexes 7 and 9 but coordinates at
the free axial position of the second ruthenium atom;
the in most cases exceptionally low cyclopropane
yields with the halopyridonate/PPh3 complexes 7 and
9 could then be attributed to the hindered access of
the diazoester to this ruthenium atom which is steri-
cally shielded by two halogen atoms. A contribution

T. Werle et al. / Journal of Organometallic Chemistry 690 (2005) 5562–5569
5567
of an uncatalyzed cyclopropanation pathway can also
not be excluded [5].
control of diastereoselectivity by ligand tuning is a
long-standing challenge of catalytic carbenoid cyclo-
propanation reactions. Nevertheless, several correla-
tions between the ligands in the catalytically active
complex and the diastereoselectivity can be extracted
from the obtained data. Unfortunately, the fast rear-
rangement of the head–tail complexes 10 and 11 into
head–head complexes 12 and 13, respectively, under
the reaction conditions did not allow us to observe the
influence of this structural change on yields and diaste-
reoselectivities of the cyclopropanation reactions. Re-
cently, it has been shown for carbenoid CAH insertion
reactions catalyzed by rhodium complexes of the type
[Rh2(OOCR)2L2], where L is a bridging ortho-metal-
lated diphenylphosphanyl-(het)aryl ligand, that the
head–head and the head–tail arrangement of these
unsymmetrical ligands causes indeed distinctly different
results [22].
For the head–tail complexes 10 and 11, which have a
C2-symmetric molecular topology, no ambiguity con-
cerning the coordination site of the carbene moiety
exists because the axial sites at both ruthenium atoms
are symmetry-equivalent. However, we have NMR-
spectroscopic evidence [20b] that complexes 10 and 11
rearrange irreversibly to the head–head isomers 12 and
13, respectively, in chloroform solution between ꢀ20
and 0 ꢁC (Scheme 5). Thus, we encounter the same situ-
ation as with the related head–head complexes discussed
above, and we would expect about the same cycloprop-
anation diastereoselectivities for the chloropyridonate
complexes 4, 6 and 10 on one hand, and for the bromo-
pyridonate complexes 5, 8 and 11 on the other. This is
indeed the case for cyclopropanation of styrene
(d.r. = 1.52–1.66 in the chloro series, and 1.39–1.44 in
the bromo series). For the reactions with cyclohexene
and 2-methyl-2-butene, the two series of complexes give
rise to diastereoselectivities in the same very narrow
range (d.r. = 2.00–2.12 for cyclohexene, 0.14–0.16 for
2-methyl-2-butene).
4. Experimental
4.1. Materials
Ru3(CO)12 was prepared as published [23] or pur-
chased. The following complexes were prepared by
published methods: [Ru2(CO)4(l-OAc)2]n [15], [Ru2-
(CO)4(CH3CN)2(l-OAc)2] [15], [Ru2(CO)4(l-OOCH)2]n
[15], 4–9 [20a], 10 [20b], 11 [20b] and methyl diazoace-
tate [24].
3. Conclusion
This study has shown that several ruthenium com-
plexes with the [Ru2(CO)4]2+ core and bridging carbox-
ylate or 2-pyridonate ligands are efficient and effective
catalysts for cyclopropanation of nucleophilic alkenes
with methyl diazoacetate. In terms of yields, the acetate
complex [Ru2(CO)4(l-OAc)2]n is generally superior to
the related propionate, trifluoroacetate and formate
complexes (in this sequence). Among the pyridonate
complexes, those with 6-bromopyridonate ligands
([Ru2(CO)4(2-Br-pyO)2]2 (5), [Ru2(CO)4(2-Br-pyO)2-
(MeOH)] (8) and [Ru2(CO)4(2-Br-pyO)2(CH3CN)2]
(11)) are in general a little more effective than their
relatives with 6-chloropyridonate ligands or with the
parent pyridonate ligand and represent good alterna-
tives to the acetate complex mentioned before. The dia-
stereoselectivities of cyclopropanation of a given alkene
are found in a rather limited range for all investigated
catalysts. This is not unexpected, because effective
4.2. Ruthenium complexes
4.2.1. Di-(l-butyrato)-tetracarbonyl-diruthenium(I,I)
(1b)
A suspension of Ru3(CO)12 (1.28 g, 2 mmol) in buty-
ric acid (100 ml) was heated at reflux (130 ꢁC) until gas
evolution had ceased (7 h). The initially red solution
gradually turned to yellow. The excess of butyric acid
was removed by distillation at ambient pressure, remain-
ing traces were evaporated at 100 ꢁC/0.001 mbar. The
solid residue was washed with methanol (2 · 30 ml)
and ether (2 · 30 ml). After drying at 100 ꢁC/
0.001 mbar, a yellow solid was obtained which decom-
posed at 228 ꢁC; yield: 1.27 g (87%). 1H NMR
(400.1 MHz, CD3CN): d = 0.86 (t, J = 7.3 Hz, 3H,
CH3), 1.53 (mc, 2H, CH2CH3), 2.17 (t, J = 7.3 Hz,
2H, CH3CH2CH2). 13C NMR (100.6 MHz, CD3CN):
d = 14.0 (CH3), 20.2 (CH2), 39.4 (CH2), 187.8 (COO),
203.3 (C„O). IR (KBr): m = 2050 (vs, CO), 1990 (vs,
CO), 1955 (vs, CO), 1945 (vs, CO), 1545 (s, br, OCO),
1390 cmꢀ1 (vs, OCO). MS (EI, 70 eV): m/z
(%) = 488.9/489.9 (15), 460.9/461.9 (17), 432.9/433.9
(33), 404.9/405.9 (35), 376.9/377.9 (100). Anal. Calc.
for C12H14O8Ru2 (488.38): C 29.51, H 2.89. Found C
29.8, H 2.9.
R
O
N
O
R
N
O
CHCl3,
CO
L
-20 0 o
C
CO
O
O
N
+
L
L
C
Ru
Ru
R
Ru
Ru
R
L
N
C
O
C
O
C
O
C
O
C
O
10: R = Cl
11: R = Br
12: R = Cl
13: R = Br
L = CH3CN
Scheme 5.

5568
T. Werle et al. / Journal of Organometallic Chemistry 690 (2005) 5562–5569
4.2.2. Di-(l-trifluoroacetato)-tetracarbonyl-
diruthenium(I,I) (1d)
determined using an authentic sample prepared accord-
ing to Section 4.3.1.
A solution of [Ru2(CO)4(CH3CN)2(l-OAc)2] (1.03 g,
2 mmol) in trifluoroacetic acid (10 ml) and acetic anhy-
dride (1 ml) was heated at 50 ꢁC for 2 h. The carboxylic
acids are evaporated at 40 ꢁC/0.003 mbar, last traces at
100 ꢁC/0.001 mbar. The residue was then kept at
165 ꢁC/0.001 mbar until its weight was constant (ca.
6 h). An orange-red solid was obtained which decom-
posed at 205 ꢁC; yield: 0.96 g (89%). 13C{1H} NMR
Acknowledgments
Financial support by the Deutsche Forschungsgeme-
inschaft and the Fonds der Chemischen Industrie is
gratefully acknowledged. The support and sponsorship
conceded by COST Action D24 ÔSustainable Chemical
Processes: Stereoselective Transition Metal-Catalyzed
ReactionsÕ is kindly acknowledged.
1
(100.6 MHz, CD3CN): d = 115.9 (q, J(C,F) = 286 Hz,
CF3), 170.2 (q, 2J(C,F) = 39 Hz, COO), 201.4 (s,
C„O). IR (KBr): m = 2095 (sh, CO), 2045 (vs, CO),
1990 (vs, CO), 1960 (vs, br, CO), 1640 (vs, br, OCO),
1460 cmꢀ1 (vs, OCO). MS (EI, 70 eV): m/z (%) =
540.8/541.8 (34), 512.8/513.8 (12), 484.8/485.8 (23),
456.8/457.8 (15), 428.6 (100). Anal. Calc. for
C8F6O8Ru2 (540.21): C 17.79. Found C 17.8.
References
[1] M.P. Doyle, M.A. McKervey, T. Ye, Modern Catalytic Methods
for Organic Synthesis with Diazo Compounds: From Cyclopro-
panes to Ylides, Wiley, New York, 1998.
[2] A. Padwa, D.J. Austin, Angew. Chem. 106 (1994) 1881–1899;
A. Padwa, D.J. Austin, Angew. Chem. Int. Ed. Engl. 33 (1994)
1797–1815.
4.3. Catalytic cyclopropanations
[3] (a) M.P. Doyle, D.C. Forbes, Chem. Rev. 98 (1998) 911–936;
(b) H.M.L. Davies, R.E.J. Beckwith, Chem. Rev. 103 (2003)
2861–2903.
[4] C.A. Merlic, A.L. Zechman, Synthesis (2003) 1137–1156.
[5] M.C. Pirrung, A.T. Morehead Jr., J. Am. Chem. Soc. 116 (1994)
8991–9000.
4.3.1. Method A
A solution of methyl diazoacetate (2.00 g, 20 mmol)
in liquid alkene (20 mmol) was added during 12 h, by
means of a syringe pump, to a magnetically stirred solu-
tion of the same alkene (180 mmol) in dichloromethane
(25 ml) containing 1 mol% of catalyst (1b–d). Stirring of
the reaction mixture was continued until the evolution
of N2 had ceased (3–8 h). The solvent and low-boiling
alkenes were removed by distillation at 60 ꢁC/800 mbar.
The residue was separated by column chromatography
(silica gel, Macherey & Nagel, 0.063–0.2 mm; water-
cooled column). The alkene was eluted first with pen-
tane, the cyclopropanes (E/Z mixture) with ether/
pentane mixtures and diethyl fumarate and diethyl
maleate with ether. All products reported in Table 1
are known; the stereochemistry of cyclopropanes 12
´
[6] (a) A.J. Anciaux, A.J. Hubert, A.F. Noels, N. Petiniot, P. Teyssie,
J. Org. Chem. 45 (1980) 695–702;
(b) M.P. Doyle, R.L. Dorow, W.E. Buhro, J.H. Griffin, W.H.
Tamblyn, M.L. Trudell, Organometallics 3 (1984) 44–52;
(c) M.P. Doyle, V. Bagheri, T.J. Wandless, N.K. Harn, D.A.
Brinker, C.T. Eagle, K.-L. Loh, J. Am. Chem. Soc. 112 (1990)
1906–1912.
[7] G. Maas, Chem. Soc. Rev. 33 (2004) 183–190.
[8] (a) A.F. Noels, A. Demonceau, E. Carlier, A.J. Hubert, R.-L.
Ma´rquez-Silva, R.A. Sa´nchez-Delgado, J. Chem. Soc., Chem.
Commun. (1988) 783–784;
(b) A.F. Noels, A. Demonceau, J. Phys. Org. Chem. 11 (1998)
602–609.
[9] A. Demonceau, A.F. Noels, E. Saive, A.J. Hubert, J. Mol. Catal.
76 (1992) 123–132.
was assigned based on H and 13C NMR data.
1
[10] G. Maas, T. Werle, M. Alt, D. Mayer, Tetrahedron 49 (1993)
881–888.
4.3.2. Method B
The catalyst (4–11; 3.0 mol% based on diazoacetate)
was dissolved in a mixture of alkene (10 mmol) and
dichloromethane (4 ml). By means of a syringe pump,
a solution of methyl diazoacetate (0.100 g, 1 mmol) in
dichloromethane (0.9 ml) was added at a rate that was
adjusted to the reactivity of the alkene (ca. 4 h for sty-
rene, ca. 10 h for cyclohexene and 2-methyl-2-butene).
The complete consumption of the diazo compound
was monitored by IR spectroscopy. A defined amount
of naphthalene (for experiments with styrene and cyclo-
hexene) or mesitylene (for 2-methyl-2-butene) was added
as an internal standard, and the yields and diastereomer
ratios of cyclopropanes 12 were determined by gas chro-
matography, using a Varian CP-WAX 52 column (30
m · 0.32 mm, film thickness 0.25 lm) fitted with a reten-
tion gap. The response factor of each cyclopropane was
[11] T. Werle, G. Maas, Adv. Synth. Catal. 343 (2001) 37–40.
[12] G. Maas, M. Alt, D. Mayer, U. Bergstra¨sser, S. Sklenak, P.
Xavier, Y. Apeloig, Organometallics 20 (2001) 4607–4615.
[13] G. Maas, J. Seitz, Tetrahedron Lett. 42 (2001) 6137–6140.
[14] C.-D. Leger, G. Maas, Z. Naturforsch. 59b (2004) 573–578.
[15] G.R. Crooks, B.F.G. Johnson, J. Lewis, I.G. Williams, J. Chem.
Soc. (A) (1969) 2761–2766.
[16] H. Schumann, J. Opitz, J. Pickart, J. Organomet. Chem. 128
(1977) 253–264.
[17] F. Neumann, H. Stoeckli-Evans, G. Suss-Fink, J. Organomet.
¨
Chem. 379 (1989) 139–150.
[18] P. Kalck, M. Siani, J. Jenck, B. Peyrille, Y. Peres, J. Mol. Catal.
67 (1991) 19–27.
[19] P.L. Andreu, J.A. Cabeza, V. Riera, Y. Jeannin, D. Miguel, J.
Chem. Soc., Dalton Trans. (1990) 2201–2206.
[20] (a) L. Scha¨ffler, B. Muller, G. Maas, Inorg. Chim. Acta (2005), in
¨
press;
(b) L. Scha¨ffler, G. Maas (manuscript in preparation).
[21] M.P. Doyle, D. van Leussen, J. Org. Chem. 47 (1982) 5326–5339.

T. Werle et al. / Journal of Organometallic Chemistry 690 (2005) 5562–5569
5569
´
[22] (a) F. Estevan, P. Lahuerta, J. Perez-Prieto, I. Pereira, S.-E.
[23] A. Mantovani, S. Cenini, Inorg. Synth. 16 (1976) 47–48.
Stiriba, Organometallics 17 (1998) 3442–3447;
[24] (a) N.E. Searle, Org. Synth. Coll. Vol. IV (1963) 424–
´
(b) K. Bieger, F. Estevan, P. Lahuerta, J. Lloret, J. Perez-Prieto,
426;
M. Sanau´, N. Siguero, S.-E. Stiriba, Organometallics 22 (2003)
1799–1801.
(b) H.-U. Reißig, J. Reichelt, T. Kunz, Org. Synth. 71
(1992) 189–199.