C2-symmetric chiral disulfoxide ligands in rhodium-catalyzed 1,4-addition: From ligand synthesis to the enantioselection pathway

Article abstract of DOI:10.1002/chem.201001749

A family of chiral C₂-symmetric disulfoxide ligands possessing biaryl atropisomeric backbones has been synthesized by using the Andersen methodology. Complete characterization includes X-ray crystallographic studies of all ligands and some of t

Full text of DOI:10.1002/chem.201001749

FULL PAPER
DOI: 10.1002/chem.201001749
C2-Symmetric Chiral Disulfoxide Ligands in Rhodium-Catalyzed 1,4-
Addition: From Ligand Synthesis to the Enantioselection Pathway
Ronaldo Mariz,[a] Albert Poater,[b] Michele Gatti,[a] Emma Drinkel,[a] Justus J. Bꢀrgi,[a]
Xinjun Luan,[a] Sascha Blumentritt,[a] Anthony Linden,[a] Luigi Cavallo,*[c] and
Reto Dorta*[a]
Abstract: A family of chiral C2-sym-
metric disulfoxide ligands possessing
biaryl atropisomeric backbones has
been synthesized by using the Ander-
sen methodology. Complete characteri-
zation includes X-ray crystallographic
studies of all ligands and some of their
rhodium complexes. Their synthesis,
optical purity, electronic properties,
and catalytic behavior in the prototypi-
cal rhodium-catalyzed 1,4-addition of
phenylboronic acid to 2-cyclohexen-1-
one are presented through an in depth
study of this ligand class. Density func-
tional theory calculations on the step
of the catalytic cycle that determines
the enantioselectivity are presented
and reinforce the first hypothetical ex-
planations for the high levels of asym-
metric induction observed.
Keywords: asymmetric catalysis
·
chiral auxiliaries · conjugate addi-
tion · enantioselectivity · sulfoxides
Introduction
the last few years have other systems appeared, for example
chiral dienes, in combination with rhodium and iridium cata-
lysis.[1b–c,4] An alternative class that has been more sporadi-
cally applied in catalysis are ligands that contain sulfur
donors.[5] Among these compounds, sulfoxides are especially
appealing, owing to their inherent chirality at sulfur, their
high optical stability, and their facile synthetic access in
enantiomerically pure form.[6] Although, the sulfinyl group
has played an important role as an efficient chiral auxiliary
in numerous asymmetric transformations, the application of
this moiety as a ligand for transition-metal catalysts has re-
mained neglected.[7–10]
The increasing demand for enantiomerically pure com-
pounds in many industrial sectors, combined with the need
for atom economy and efficiency, has catapulted asymmetric
transition-metal catalysis to the forefront of research ef-
forts.[1,2] Among the plethora of chiral ligands developed so
far, those possessing a C2-symmetric axis are often the most
successful at inducing high degrees of selectivity in cataly-
sis.[3] The overwhelming majority of these ligands have phos-
phorus, nitrogen, and oxygen as donor atoms, and only in
We surmised that the combination of a rigid, C2-symmet-
ric backbone framework and two enantiopure chiral sulfox-
ide donors would create an efficient chelating ligand envi-
ronment. The extremely powerful atropisomeric biaryl-type
backbones that have been so successful for diphosphine li-
gands seemed to be an ideal first choice. In contrast to the
BINAP-type systems, we also anticipated that the diastereo-
meric ligands that we would create, when switching achiral
phosphine moieties with enantiopure chiral sulfoxides,
would allow us to separate the atropisomeric backbone moi-
eties, thus enabling the use of racemic precursor molecules
of these fragments (Scheme 1).
[a] R. Mariz, M. Gatti, E. Drinkel, J. J. Bꢁrgi, X. Luan, S. Blumentritt,
Dr. A. Linden, Prof. Dr. R. Dorta
Organisch-chemisches Institut
Universitꢂt Zꢁrich
Winterthurerstrasse 190, 8057 Zꢁrich (Switzerland)
E-mail: dorta@oci.uzh.ch
[b] Dr. A. Poater
Catalan Institute for Water Research (ICRA)
H2O Building, Scientific and Technological Park
University of Girona
Emili Grahit 101, 17003 Girona (Spain)
[c] Prof. Dr. L. Cavallo
Dipartimento di Chimica
Universitꢃ degli Studi di Salerno
Via Ponte don Melillo, Fisciano (SA), 84084 (Italy)
To gain access to these ligands we chose Andersenꢀs ap-
proach, which involves the nucleophilic addition of an or-
ganometallic reagent (organolithium or organogrignard re-
agent) to a sulfinylating agent containing an electrophilic
The Supporting Information for this article contains all computational
and experimental details and is available on the WWW under http://
dx.doi.org/10.1002/chem.201001749.
Chem. Eur. J. 2010, 16, 14335 – 14347
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
14335

Results and Discussion
General synthetic strategy: The general synthetic strategy,
outlined in (Scheme 2), involves nucleophilic substitution on
a sulfinate ester (Andersen method). Starting with backbone
Scheme 1. Disulfoxides with C2-symmetric backbone.
sulfur atom of known configuration.[11] Based upon previous
reports using this approach,[12,13] we have recently shown
that two chiral disulfoxide ligands that are analogues of
Noyoriꢀs BINAP ligand,[14] and its derivative BIPHEMP,[15]
can be used successfully as chiral ligands in the rhodium-cat-
alyzed addition of arylboronic acids to a,b-unsaturated car-
bonyl compounds (Miyaura–Hayashi reaction).[16]
A common pathway for tuning catalytic properties in C2-
symmetric atropisomeric biaryl ligands is the variation of
their dihedral angle, induced by changes in the substitution
of the backbone units.[17] The modification of geometry and,
as a consequence, the electronic distribution around the
metal center in many cases alters the activity and/or selectiv-
ity of a given catalytic system.[18] Another way to tune cata-
lyst performance of well-established ligand families relies on
changes of the stereoelectronic properties (i.e., s basicity or
p acidity) of the groups directly attached to the donor atom.
For instance, the difference in donor–acceptor abilities has
been investigated for diphosphine ligands and shown to pro-
mote significant changes in both activity and selectivity for a
given reaction.[19]
Scheme 2. General synthetic strategy.
molecules containing two bromides, the first reaction with
either an organolithium compound or magnesium metal
leads to either a dilithiated or di-Grignard derivative. These
nucleophiles are then used in consecutive substitution reac-
tions on the sulfinate ester to give the desired products in
one synthetic step.
Synthesis of C2-symmetric backbone precursors: As men-
tioned above, we wanted to get a clearer picture regarding
the behavior of our disulfoxide ligands with respect to modi-
fications of the atropisomeric backbone and selected three
well-known structures (Scheme 3) for the present study. The
Our initial findings on the successful use of two disulfox-
ide ligands in the Miyaura–Hayashi reaction serve as a basis
for the study we report herein. Owing to the novelty of this
ligand family, and in line with early studies done on other
ligand classes, we wanted to see how the substitution pat-
terns of both the backbone and the sulfoxide groups affect
the synthesis of the ligands and their coordination ability to
rhodium, as well as their catalytic performance in the proto-
typical 1,4-addition reaction of phenylboronic acid to 2-cy-
clohexen-1-one. The results reported below show that the
synthesis of these disulfoxides is not always as straightfor-
ward as expected. Their coordination ability with respect to
electronic ligand modifications is investigated by synthesiz-
ing a series of carbonyl-containing rhodium complexes, and
the catalytic performance of the respective ligands was com-
pared. Within the context of the selectivity pathway in the
Miyaura–Hayashi reaction, preliminary results indicate an
unusual mechanism in the enantioselection when employing
these atropisomeric disulfoxide ligands. To understand the
origin of selectivity, density functional theory (DFT) calcula-
tions were performed with the disulfoxide–rhodium systems
showing the best selectivity in catalysis and the results are
presented here.[20] Further details on the origin of the stereo-
selectivity by these disulfoxide–rhodium and related cata-
lysts can be found in reference [21].
Scheme 3. Well-known structures for the backbone.
racemic dibromo precursors shown are either commercially
available (rac-1),[22] or can be easily synthesized by using es-
tablished synthetic methods. To access rac-2, the synthetic
pathway developed by Schmid and Frejd was slightly modi-
fied.[23] Compound rac-3 was obtained in one step from rac-
1 by means of selective partial hydrogenation by using Ru/
C.[24]
Synthesis of sulfinate esters: To gain access to electronically
modified sulfoxide ligands with similar steric properties that
incorporate the parent atropisomeric backbone rac-1, we
trapped racemic sulfinyl chlorides 4–9,[25] with either com-
mercially available diacetone-d-glucose (DAG, Method A)
or cheap l-menthol (Method B), as stereo-controlling alco-
hols (Scheme 4).
In all cases, the corresponding sulfinates (10–15) were ob-
tained in good to excellent yields (see the Supporting Infor-
mation). Some of these sulfinates are known and were syn-
thesized according to the literature procedures [(S)-12 and
(S)-13].[26,27] Sulfinate (R)-13 was purchased and used as re-
14336
www.chemeurj.org
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 14335 – 14347

Rhodium-Catalyzed 1,4-Addition
FULL PAPER
spectroscopy, analysis of both isomers presented a complex
set of signals resulting in no direct assignment of configura-
tion for these sulfinates.
Although (S)-10 and 15 were not obtained as diastereo-
merically pure compounds, they were nevertheless used for
the synthesis of electronically modified disulfoxide ligands
below.
Disulfoxide synthesis
Backbone variations: The syntheses of disulfoxides 16, 17,
and 18 are summarized in Scheme 5. The addition of menth-
yl sulfinate ester (S-13) to the dilithiated intermediate de-
Scheme 4. Synthesis of DAG and l-menthol sulfinates 10–15.
ceived. The 4-methoxyphenyl
counterpart (S)-14 is also
known, but its synthesis was
modified to give highly diaster-
eopure material in 63% overall
yield.[27] The new sulfinate (S)-
11 was formed with high selec-
tivity and the optically pure
compound was obtained in
88% yield. Unequivocal deter-
mination of the absolute config-
Scheme 5. Synthesis of p-Tol-disulfoxides 16, 17, and 18.
uration of both (S)-11 and (S)-
14 was established by an X-ray
diffraction study and their
rived from rac-1 afforded p-Tol-BINASO (16), as a pair of
diastereomers in approximately 90% yield.[28] The P and M
atropisomers of 16 were easily separated by using silica-gel
chromatography. For convenience, we adopt for each disulf-
oxide ligand a lower case letter after the compound number
to indicate the order in which these isomers are collected in
the chromatographic purification of their crude reaction
mixture. Therefore, 16a refers to the first compound isolat-
ed and 16b refers to the diastereoisomer collected in later
fractions. For unknown reasons, only traces of the desired
disulfoxide were formed after the addition of 13 to the dili-
thiated species of rac-2 (independent of whether nBuLi,
nBuLi/TMEDA, sBuLi or tBuLi was used). Sulfinate 13 did
structures are shown in Figure 1. The reaction of phenyl sul-
finyl chloride with DAG gave (S)-10 in 98% yield, as a dia-
stereomeric mixture in a 10:1 ratio (S:R). Unfortunately,
column chromatography or recrystallizations from pentane/
diethyl ether did not affect the relative amount of the two
isomers. Nevertheless, we could establish the major isomer
as (S)-10 through single-crystal X-ray studies (Figure 1). The
4-(trifluoromethyl) analogue 15 was obtained in 96% yield
as an approximate 3:2 mixture of two diastereoisomers that
were separated by column chromatography. In this case, the
products were obtained as oils and did not allow crystallo-
graphic analysis. Furthermore, using H, 13C, and 19F NMR
1
Figure 1. The solid-state molecular structures of S-configured sulfinates a) (S)-10, b) (S)-11, and c) (S)-14.
Chem. Eur. J. 2010, 16, 14335 – 14347
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
14337

R. Dorta, L. Cavallo et al.
undergo nucleophilic substitu-
tion when treated with the di-
Grignard derivative of rac-2.
The di-Grignard reagent, how-
ever, was exceedingly difficult
to generate (see the Supporting
Information). Nevertheless, the
desired disulfoxide ligand p-
Tol-MeBIPHESO (17) could be
obtained in 50–60% overall
yield, after separation of the
pair of diastereoisomers by
column chromatography. Nota-
bly, for both p-Tol-BINASO
and p-Tol-MeBIPHESO, we
also synthesized the corre-
sponding (R)-configured disulf-
oxides by using commercially
available sulfinate (R)-13 and
named the respective products
16’ and 17’.
For the third variation of the
backbone residue included in
our study, as described for H8-
BINAP,[24b] the di-Grignard re-
agent was obtained upon heat-
Figure 2. The solid-state molecular structures of a) (M,R,R)-p-Tol-BINASO (16a’),[16a] b) (M,S,S)-p-Tol-MeBI-
PHESO (17b),[16b] and c) (M,S,S)-p-Tol-H8-BINASO (18b).
ing a mixture of rac-3 and magnesium in THF–toluene (1:3)
at refluxing temperature. The resulting organometallic spe-
cies was then allowed to react with 13 at À408C and slowly
warmed to room temperature. Surprisingly, instead of the
expected pair of diastereomers, only one of the diastereoiso-
mers (18b) and another compound (18a) were formed in
72% overall yield. This byproduct was separated from the
predicted diastereoisomer by chromatography and analyzed.
A set of two sharp signals of equal intensity around d=2.5–
Sulfinyl group variations: To understand the influence of the
substitution pattern at the sulfoxide moiety, the sulfinate
esters 10–12 and 14, 15 were incorporated into the backbone
framework rac-1 to provide five other disulfoxide ligands
complementary to the three mentioned above (Scheme 6).
Under the same reaction conditions employed to obtain
16, ligands 19–22 were synthesized in 65–85% overall yield.
In all of these cases, the two expected atropisomers were
separated and isolated after column chromatography along-
side a third fraction that showed the (DNHS)-conformers at
sulfur. Attempts to form the 4-(trifluoromethyl)phenyl de-
rivative 23 using the lithiation path failed. We suppose that
side reactions through ortho-metalation of the acidic proton
in the 4-CF3PhSO core are the reason for the complex mix-
ture furnished by this experiment.[29] Nevertheless, 23 could
be readily accessed when substituting the lithium nucleo-
phile by its Grignard equivalent (78% yield). As described
1
2.7 ppm in the H NMR spectrum attributed to the methyl
groups of the p-tolyl fragment, together with a more compli-
cated spectrum in the aromatic region, meant that a set of
diastereomers possessing non-homochiral stereogenic sulfur
centers (e.g., M- or P-S,R, from this point onwards abbrevi-
ated DNHS) represented the most likely assignment of its
structure.
Crystals suitable for X-ray crystallography were obtained
for ligands 16a’, 17b and 18b
and allowed an unambiguous
assignment of all sites of the
molecules (Figure 2). Compari-
son
of
(M,S,S)-p-Tol-H8-
BINASO (18b) with its fully
conjugated analogue and the bi-
phenyl derivative shows no sig-
À
nificant variation in S O bond
distances (1.4922(16) ꢅ for
BINASO; 1.4988(16) ꢅ for Me-
BIPHESO; 1.495(2) ꢅ for H8-
BINASO).
Scheme 6. Variation of substituents on the sulfoxide moiety of rac-1 derived ligands.
14338
www.chemeurj.org
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 14335 – 14347

Rhodium-Catalyzed 1,4-Addition
FULL PAPER
for 18, the synthesis of 23 afforded the (DNHS)-conformer
and only one of the two expected diastereoisomers.
To confirm the configurations at sulfur, at least one
isomer of each of the disulfoxides 19–23 was characterized
by X-ray diffraction studies (see the Supporting Informa-
tion).[30] Analysis of their solid-state structures shows a grad-
ide ligands by HPLC. Because of the conformational stabili-
ty of the respective atropisomers (M and P), we realized
that another straightforward method consisted of reducing
both sulfoxides to the less polar sulfides and analyzing the
products by HPLC. This path allows an indirect yet elegant
determination of the purity of each homochiral diastereoiso-
mers, as well as the (DNHS)-disulfoxides. By adapting liter-
ature procedures,[31] and by making sure that racemization
of the atropisomeric backbone does not occur,[32] quantita-
tive reduction of all disulfoxide ligand isomers gave their
corresponding disulfides 24–31 (Scheme 7). Table 1 summa-
rizes the HPLC results obtained for the respective disulfides
and the isomeric purity of our disulfoxide ligands. The data
reported validate all of the assumptions made above on
À
ual increase in the S O bond lengths when going from the
most electron-poor substituent at sulfur [23b, (1.483(2) ꢅ)]
to the most donating substituent [21a (1.4986(13) ꢅ)]. In
addition, the crystallographic studies carried out for 19b
[(DNHS)-Ph-BINASO] and 20b [(DNHS)-4-FPh-BINASO]
were the ultimate proof for the formation of non-homochi-
ral diastereoisomers (Figure 3).
Scheme 7. Reduction of disulfoxides into disulfides using Lawessonꢀs re-
agent.
Table 1. Enantiomeric excess of disulfides and corresponding disulfox-
ides.
Disulfide
ee [%][a]
Corresponding disulfoxide[b]
(M,S,S)-p-Tol-BINASO 16b
(P,S,S)-p-Tol-BINASO 16a
98
>99
ACHTUNGTRENNUNG
24
ACHTUNGTRENNUNG
>99
98
ACHTUNGTRENNUNG
25
26
ACHTUNGTRENNUNG
87
>99
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
Figure 3. The solid-state molecular structures of a) (DNHS)-Ph-BINASO
(19b) and b) (DNHS)-4-FPh-BINASO (20b).
95
39
81
ACHTUNGTRENNUNG
27
28
29
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
Overall, the synthesis of the disulfoxide ligands by using
consecutive nucleophilic substitutions at the two electrophil-
ic sulfur moieties is certainly not as straightforward as we
had expected. Although the generation of (DNHS)-con-
formers in some of these reactions is not entirely surprising,
it does pose a serious problem for the unambiguous attribu-
tion of the stereochemistry of all of these compounds.
96
0
99
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
ACHTUNGTRENNUNG
nd[c]
nd[c]
nd[c]
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
96
60
99
ACHTUNGTRENNUNG
Reduction to disulfides and enantiomeric purity of ligands:
The evidence for the presence of a (DNHS)-conformer in
the synthesis of some of the ligands meant that we needed
to find a means to ascertain the optical purity of our ligands.
For instance, contamination of a given (M,S,S)-ligand with
its (P,R,R)-enantiomer would spectroscopically go unno-
ticed. Initial, unsuccessful attempts were made by using
30
31
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
55
40
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
[a] Enantiomeric excess determined by HPLC using ChiralPak-IB and
Chiralcel OD-H columns. [b] Absolute configuration of the major isomer
as assigned by crystallographic and catalytic results [(M,S,S)-(disulfox-
ide)-Rh gives (R)-52aA in the catalytic runs and (P,R,R)-(disulfoxide)-
Rh gives (S)-52aA, reference 16]. [c] Not determined.
1
chiral shift reagents and analyzing the mixture by H NMR
spectroscopy or by trying to analyze the very polar disulfox-
Chem. Eur. J. 2010, 16, 14335 – 14347
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
14339

R. Dorta, L. Cavallo et al.
both the purities of the ligands and the starting sulfinates.
The optical purity of the (M,S,S)- and (P,S,S)-pair of disulf-
oxides [or its equivalent (M,R,R)- and (P,R,R)-pair] corre-
sponded at least to the initial optical purity of the sulfinates.
In the case of the (DNHS)-isomers, we note that the distri-
bution between the possible atropisomers varies from a per-
fectly racemic backbone (DNHS-20b) to one showing
highly enriched atropisomeric distributions (DNHS-18a),
adding another layer of complexity to the disulfoxide syn-
thesis.
The molecular structures of 32, 33, 35, and 38 are dis-
played in Figure 4. The molecular structures of 32, 33, and
38 present the expected {Rh2ACHTNUTRGENNUG(-Cl)2} butterfly shaped core,
whereas complex 35 with the phenyl substituted ligand pos-
sesses an almost perfectly planar arrangement. Data of the
most important bond lengths and angles for the disulfoxides
32, 33, 35, and 38 and their analogous diphosphine com-
plexes
[({(R)-BINAP}RhCl)2],[34]
and
[({(S)-BI-
PHEMP}RhCl)2],[16b] are summarized in Tables S1 and S2
(see the Supporting Information).
Direct comparison of the disulfoxide and diphosphine
complexes reveals ligand-donor–metal bond lengths that are
very similar. Bite angles in the range of 97–988 were found
for our disulfoxide complexes, whereas the phosphorus com-
pounds present a slightly narrower bite angle at the metal
(91–938). In contrast, the dihedral angles of the atropisomer-
ic backbones are very similar for the two ligand classes. Fi-
nally, coordination of the sulfoxide moiety to the metal
leads to a shortening of the S=O bond, a phenomenon that
gives a qualitative indication of the donor abilities of sulfox-
ides.
Synthesis of rhodium complexes: Two equivalents of ligands
16b, 16b’, 17b, 17b’, 18b, 19c, 20c, 21c, 22c, and 23b react-
ed cleanly with the rhodium ethylene dimer [{Rh-
ACHTUNGTRENNUNG(C2H4)2Cl}2] in dichloromethane to afford complexes 32–39
[{(disulfoxide)RhCl}2] in very high yields after appropriate
workup (Scheme 8, Table 2).[33]
Rhodium carbonyl complexes and analysis of electron–
donor properties: Correlation of the electron density on a
metal with the n(CO) frequency of coordinated CO has been
routinely used to evaluate the relative donor strength of var-
ious ligands in metal carbonyl complexes.[19b,35] To quantify
to what extent our sulfoxide ligands are able to donate elec-
tron density to rhodium, we synthesized cationic carbonyl
complexes of general formula [(L–L)Rh(CO)2]+ (where L–
L is one of the investigated bidentate sulfoxide or analogous
phosphine ligands) by treating [{Rh(CO)2Cl}2] with the che-
lating ligand in the presence of AgBF4 (see the Supporting
Information). IR spectroscopic data of these carbonyl com-
plexes were collected and average carbonyl stretching fre-
quencies are given in the right column of Table 3.
Scheme 8. Synthesis of dinuclear chloro-bridged rhodium complexes 32–
39.
Table 2. Ligands, corresponding rhodium complexes and isolated yields
obtained.
Ligand
(P,R,R)-p-Tol-BINASO (16b’)[a]
(M,S,S)-p-Tol-BINASO (16b)[a]
(P,R,R)-p-Tol-MeBIPHESO (17b’)[a]
(M,S,S)-p-Tol-MeBIPHESO (17b)[a]
(M,S,S)-p-Tol-H8-BINASO (18b)
(M,S,S)-Ph-BINASO (19c)
(M,S,S)-4-FPh-BINASO (20c)
(M,S,S)-Cy-BINASO (21c)
(M,S,S)-4-MeOPh-BINASO (22c)
(P,R,R)-4-CF3Ph-BINASO (23b)
[a] Values extracted from reference [16].
Complex
Yield [%]
G
ACHTUNGTRENUN[NG {(16b’)RhCl}2] (32’) 95
ACHTUNGTRENN[NUG {(16b)RhCl}2] (32)
ACHTUNGTRENUN[NG {(17b’)RhCl}2] (33’) 93
T
E
95
E
R
95
97
99
95
96
93
93
R
ACHTUNGTRENNUNG
N
ACHTUNGTRENNUNG
G
ACHTUNGTRENNUNG
Analysis of Table 3, entries 1–3, reveals that substituting
the fused aromatic rings of BINASO with sp3-hybridized
carbon atoms leads to an increase in electron-donation to
T
ACHTUNGTRENNUNG
T
ACHTUNGTRENNUNG
A
ACHTUNGTRENNUNG
Table 3. Summary of IR spectroscopic data of cationic rhodium carbonyl
complexes.
Binding of the sulfoxide ligands is accompanied by signifi-
cant changes in the H NMR spectra. An in situ reaction of
[a]
[a]
[a,b]
Entry Complex
n1(CO)
n2(CO)
nav(CO)
1
1
2
3
N
2100.10 2016.21 2058.16
2096.24 2023.93 2060.09
2097.21 2025.85 2061.53
16b with the rhodium precursor in CD2Cl2 showed fast dis-
placement of the ethylene moieties by the disulfoxide.
Crude reaction mixtures of 32, 33, and 38 were concentrat-
ed, layered with THF, and directly crystallized at low tem-
perature to give analytically pure burgundy-colored material
in high yield, as well as crystals suitable for X-ray diffraction
studies. Compound 35 precipitated cleanly in dichlorome-
thane upon completion of the reaction and crystals of com-
plex 35 were, therefore, directly obtained by allowing a mix-
ture of the rhodium precursor and ligand 19c to react with-
out stirring. Complexes 34, 36, 37, and 39 were obtained as
analytically pure products after precipitation with pentane
and subsequent washings with diethyl ether and pentane.
4
5
6
7
8
9
N
2091.42 2023.93 2057.68
2096.24 2023.93 2060.09
2097.21 2025.85 2061.53
2098.17 2026.82 2062.50
2099.14 2026.82 2062.98
2102.03 2028.75 2065.39
10
11
[{(rac)-BINAP}Rh(CO)2]BF4 (48)[c] 2094.32 2048.89 2071.61
ACHTUNGTRENNUNG
[{(S)-BIPHEMP}Rh(CO)2]BF4
2094.32 2048.03 2071.18
(49)[c]
[a] Values in cmÀ1
[c] Values extracted from reference [16b].
.
[b] navarage(CO) =(n1(CO) [cmÀ1]+n2(CO) [cmÀ1])/2.
14340
www.chemeurj.org
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 14335 – 14347

Rhodium-Catalyzed 1,4-Addition
FULL PAPER
also influence the properties of
the ligand. Likewise, varying
the substitution on the sulfox-
ide moiety, while keeping the
backbone unchanged (Table 3,
entries 4–9), follows trends ex-
pected based on electronic ar-
guments. The introduction of
electron-withdrawing
groups,
such as fluorine and trifluoro-
methyl, in the para position of
the phenyl ring of the sulfoxide
leads to diminished electron
density on the metal center
(Table 3, entries 8, 9), whereas
electron-donating
groups
(Table 3, entries 5, 6) or substi-
tution of the aromatic moiety
with
a
cyclohexyl
group
(Table 3, entry 4) increase the
electron-density at rhodium.
Because the electronic char-
acteristics of sulfoxides are
largely unknown in the context
of their coordination to metals
and to understand how this
ligand family compares to the
parent diphosphines, we also
synthesized and recorded the
data for the corresponding
[{(rac)-BINAP}Rh(CO)2]BF4
(48) and [{(S)-BIPHEMP}Rh-
(CO)2]BF4
(49)
complexes
(Table 3, entries 10, 11). Con-
trary to our initial expectations,
the results demonstrate that our
disulfoxides are more electron-
donating than their diphosphine
counterparts.[36,37]
Catalytic studies: We had previ-
ously established that precata-
lysts 32 and 33 performed very
well in the Miyaura–Hayashi
addition reaction of arylboronic
acids to cyclic a,b-unsaturated
substrates, surpassing BINAP
and BIPHEMP ligands in both
reactivity and selectivity. To un-
derstand the effects of electron-
ic and steric variation of this
first generation of chiral disulf-
Figure 4. Full and half view of the solid-state molecular structures of complexes a) [{(P,R,R)-p-Tol-BINA-
SO}RhCl]2 (32’),[16a] b) [({(M,S,S)-p-Tol-MeBIPHESO}RhCl)2] (33),[16b] c) [({(M,S,S)-Ph-BINASO}RhCl)2] (35)
and d) [({(M,S,S)-4-MeOPh-BINASO}RhCl)2] (38).
the rhodium, as reflected by the decreasing stretching fre-
quency observed in complexes 41 and 42, as compared to
40. That p-Tol-MeBIPHESO (complex 41) shows stronger
electron-donation than H8-p-Tol-BINASO (complex 42)
likely indicates that the geometric factors of the backbone
oxide ligands in catalysis, we proceeded with the screening
of the catalytic performances of complexes 32–39 in the
standard 1,4-addition reaction of 2-cyclohexen-1-one (50a)
and phenylboronic acid (51A). The results obtained are en-
closed in Table 4.
Chem. Eur. J. 2010, 16, 14335 – 14347
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
14341

R. Dorta, L. Cavallo et al.
Table 4. Catalytic results with complexes 32–39 in the coupling of 50a
with arylboronic acid 51 A.
sion of the steric effect (i.e., different dihedral angles of the
backbones) on the selectivity of the reaction.[38]
Effects of sulfoxide modification on reactivity and selectivity:
The modification of the p-Tolyl moiety of the parent p-Tol-
BINASO ligand leads to marked differences in reactivity.
Most strikingly, complex 37, which contains the most s-basic
ligand of the binaphthyl series (cyclohexyl groups), is com-
pletely inactive towards the 1,4-addition of substrate 51A to
50a even at catalyst loadings of 5 mol% (Table 4, entry 16).
Comparable effects have been reported with diphosphine li-
gands when going from aromatic to aliphatic substituents on
phosphorus.[39] On the other side of our electronic spectrum,
the p-CF3Ph-derived ligand also leads to poor catalytic con-
version (catalyst 39). The other modifications result in pre-
catalysts that show increased reactivity at 0.75 mol%catalyst
loading when compared to the parent p-tolyl-substituted
ligand 32.
Entry
Complex
Conc.
[mol%]
t [h][a]
Yield of
ee [%][c,d]
N
52aA [%][b]
1
2
3
32
32
32
0.75
0.50
0.25
1
8
24
99
86
55
98 (R)
98 (R)
97 (R)
4
5
6
33
33
33
0.75
0.50
0.25
<0.5
<0.5
0.5
98
99
98
>99 (R)
>99 (R)
>99 (R)
7
8
9
34
34
34
0.75
0.50
0.25
<0.5
<0.5
0.75
99
99
99
>99 (R)
98 (R)
98 (R)
Selectivity differences may also be compared, albeit direct
comparisons with 32 are only possible for the enantiomeri-
cally pure p-MeOPh-substituted (38, Table 4, entries 17–19)
and the p-FPh-derived ligands (36, Table 4, entries 13–15).
Precatalyst 38 shows a clear erosion of selectivity with a
maximum of 79%ee. In contrast, complex 36 not only
showed the best reactivity, but also produced product 52aA
with complete selectivity at 0.75 mol%catalyst loading
(Table 4, entry 13).
Overall, the selectivities decrease for all of the precata-
lysts if catalyst loadings are too low to warrant a short reac-
tion time to product 52aA. If the results here are cumula-
tive, it would also mean that combining the p-FPh-moiety of
sulfinate (S)-11 with the superior backbones of rac-2 and
rac-3 could lead to optimized ligand structures for the pres-
ent transformation.
10
11
12
35
35
35
0.75
0.50
0.25
0.5
1
8
98
96
70
89 (R)
78 (R)
76 (R)
13
14
15
36
36
36
0.75
0.50
0.25
<0.5
1
4
97
99
84
>99 (R)
92 (R)
90 (R)
16
37
5.00
–
0
0
17
18
19
38
38
38
0.75
0.50
0.25
0.5
1.5
5
99
99
72
79 (R)
76 (R)
75 (R)
20
21
22
39
39
39
0.75
0.50
0.25
2
5
8
45
42
29
32 (S)
32 (S)
29 (S)
[a] Reaction is stopped after full conversion or when no further conver-
sion is observed as determined by GC-MS. [b] Yield of isolated product
after column chromatography. [c] Determined by HPLC analysis with
Chiralcel OD-H: flow 0.5 mLminÀ1
, solvent Hexane/iPrOH 98:2.
Catalyst reactivity and selectivity path: Detailed research by
Hayashi and co-workers on the asymmetric 1,4-addition re-
action with BINAP-Rh has established the catalytic cycle
shown in Scheme 9.[40] We have previously compared the ac-
tivities of structurally related disulfoxide and diphosphine
rhodium dimers with chloro and hydroxo bridges to study
the reaction pathway (A in Scheme 9).[16b]
The study has shown that the diphosphine compounds are
distinctly less active than the systems incorporating disulfox-
ides under the reaction conditions outlined in Table 4. It was
also found that for diphosphine ligands, the transformation
from the chloro-bridged dimer to the active species is clearly
more difficult than formation of the monomeric [Rh]-OH
species through dimer dissociation. The inverse trend was
observed with our disulfoxide ligands for which the catalytic
run performed by using the chloro-bridged dimers is faster
and more efficient than starting with the corresponding
[Rh]-OH dimer.[41] In addition to easier accessibility to the
[Rh]-OH active species, the participation of the polarized
oxygen atom of the sulfoxide in the transmetalation step
might facilitate transfer of the aryl group to the rhodium (B
of Scheme 9).[42]
[d] Configuration of major isomer, determined by comparison with re-
ported data.
Effects of backbone modification on reactivity and selectivity:
There is a clear trend in both the reactivity and the selectivi-
ty of the reaction when changing the atropisomeric moiety
of the disulfoxide ligands (Table 4, entries 1–9). The overall
catalytic performance increases significantly when going
from p-Tol-BINASO to its partially hydrogenated derivative
p-Tol-H8-BINASO, both in terms of reactivity and selectivi-
ty. Even better results (albeit only slightly) are observed
with the p-Tol-MeBIPHESO ligand architecture. It is tempt-
ing to rationalize these findings on the basis of the IR
stretching frequencies observed above for the rhodium–car-
bonyl complexes incorporating these ligands. It would,
therefore, appear that decreasing the p acidity of the disulf-
oxide ligand backbone translates into increased reactivity in
the 1,4-addition reaction. Nonetheless, interpreting the se-
lectivity increase we observe for the same set of ligands is
hampered by the fact that we were not able to crystallize
complex 34. In our view, this precludes a meaningful discus-
14342
www.chemeurj.org
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 14335 – 14347

Rhodium-Catalyzed 1,4-Addition
FULL PAPER
Scheme 9. Catalytic cycle for the rhodium catalyzed 1,4-additon of
PhB(OH)2 to 2-cyclohexen-1-one.
More intriguing than the higher reactivity of these disulf-
oxide ligands is their mode of action during the enantiodis-
criminating step (C of Scheme 9). The stereochemical path-
way in the Miyaura–Hayashi reaction catalyzed by BINAP,
as well as in the overwhelming majority of metal-mediated
asymmetric reactions, is based on the assumption that the
substrates approach the metal so as to minimize steric inter-
actions with the protruding R groups of the chiral ligand
structure.[20,43] However, the half view of our rhodium disulf-
oxide complexes shown above (see partial views in Figure 4)
clearly indicates that our system is devoid of any significant
steric crowding around the metal center. Indeed, the aryl
groups on the sulfoxide units are oriented away from the
metal center and parallel to the atropisomeric backbone,
leaving the oxygen atoms of the sulfoxide moieties as the
sole entities approaching the metal center.
Figure 5. Structure and energy of the most stable square planar and dis-
torted tetrahedral coordination intermediates. Bond lengths from the Rh
atom to the centroid of the Ph group and to the coordinated C atoms of
the substrate are also reported.
These coordination intermediates, however, can also
assume a distorted tetrahedral (dt) geometry with the Ph
group in the apical position, see structures 53-dt-R and 53-
dt-S in Figure 5. In these structures the C=C double bond is
rotated into a face of the tetrahedron, allowing for a some-
what higher back donation from the metal to the C=C
double bond of the substrate, as evidenced by the elongation
of 0.02 ꢅ of the C=C double bond on going from the
square-planar geometry to the distorted tetrahedral (1.43 ꢅ
in the sp geometries versus 1.45 ꢅ in the dt geometries).
À
In-depth DFT computational studies were completed to
allow us to gain insight into step C of the catalytic cycle of
this reaction.[44] Notably, although the Miyaura–Hayashi re-
action represents one of the most straightforward entries
into useful chiral organic building blocks, and has emerged
as an important methodology in organic synthesis, there
have been very few computational studies to elucidate the
pathway.[45]
The enantioselection step (step C of Scheme 9) begins
with coordination of both enantiofaces of substrate 50a to
the [Rh]-Ph complex and proceeds without an energy barri-
er. The corresponding coordination intermediates display a
distorted square planar (sp) geometry around the metal, see
structures 53-sp-R and 53-sp-S in Figure 5 (R or S indicates
that this complex will lead to the R or S enantiomer of
52aA, respectively). In both structures the elongated C=C
double bond of the substrate, 1.43 ꢅ versus 1.34 ꢅ in the un-
coordinated 50a, is almost perpendicular to the mean coor-
dination plane around the metal center. In 53-sp-R, the
The Rh Ph bond length in the tetrahedral geometries is
substantially unchanged relative to the square planar geo-
metries, whereas the substrate is slightly closer to the metal.
These geometries are energetically competitive with the
square planar analogues, since 53-dt-R is only 0.9 kcalmolÀ1
higher in energy than 53-sp-R, whereas 53-dt-S is
0.7 kcalmolÀ1 lower in energy than 53-sp-S. The conversion
of the square planar geometries into the corresponding dis-
torted tetrahedral is rather facile, with barriers of 5.4 and
6.8 kcalmolÀ1 for 53-sp-R and 53-sp-S, respectively (the geo-
metries of these transition states are reported in the Sup-
porting Information). The substantially similar stability of
the complexes, together with the low energy barriers for
their interconversion, underline the remarkable manifold of
structures available after substrate coordination, and that all
these structures are probably in fast equilibrium. Although
these results clearly pertain to a new ligand class, it is cer-
tainly interesting to observe that the initial binding of the
olefin cannot represent the enantio-discriminating step in
the reaction (this is commonly assumed for other ligand
classes, such as chiral diphosphines or dienes).
À
cyclic part of the substrate is oriented away from the Rh Ph
group and towards one of the p-Tol groups, whereas in 53-
À
sp-S it is oriented right above the aromatic ring of the Rh
Indeed, the stereoselective behavior of the present cata-
lyst systems originates in the next catalytic step, namely the
insertion of the C=C double bond of the substrate into the
Ph group. Structure 53-sp-R is only 0.2 kcalmolÀ1 more
stable than 53-sp-S.
Chem. Eur. J. 2010, 16, 14335 – 14347
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
14343

R. Dorta, L. Cavallo et al.
À
Rh Ph bond. The geometries of the transition states with
the correct regiochemistry (labelled as 54-R and 54-S) are
Focusing on the pathways corresponding to the correct re-
giochemistry, transition states 54-R and 54-S collapse into
intermediates 55-R and 55-S shown in Figure 7. Intermedi-
shown in Figure 6.
Figure 6. Structures of the transition states leading to formation of the R
and S enantiomers of product 52aA (distances in ꢅ).
Figure 7. Structures of intermediates 55-R and 55-S, and of intermediates
56-R and 56-S, corresponding to the kinetic and to the thermodynamic
À
products of insertion of an R and S coordinated substrate into the Rh
Ph bond.
Transition state 54-R deviates from planarity, with the
C(Ph) atom lying out of the S-Rh-S plane by 1.23 ꢅ, where-
as in transition state 54-S the C(Ph) atom is (or lies) only
0.03 ꢅ out of the S-Rh-S plane. However, of greater impor-
tance is the fact that transition state 54-R is favored over
transition state 54-S by 4.4 kcalmolÀ1, which is in agreement
with the experimental preferential formation of the R prod-
uct with a (P,S,S) ligand. Analysis of the geometries of
Figure 6 indicates clearly that the most favored 54-R transi-
tion state presents both the Ph and the C=O group of the
substrate in rather open parts of space, which is on the side
of the p-tolyl rings that are bent away from the Rh atom. In-
stead, the competitive 54-S transition state is disfavored by
repulsive steric/electrostatic interactions between both the
Ph and C=O groups of the substrate and the upward-point-
ing S=O groups of the ligand (see the short distances be-
tween these groups in Figure 6).
ate 55-R is more stable than the coordination intermediate
53-R by 7.1 kcalmolÀ1, whereas intermediate 55-S is compa-
rable in energy with the coordination intermediate 53-S
(0.2 kcalmolÀ1 higher in energy). In both intermediates the
substrate wraps around the metal with the Ph group at least
partially coordinated to the metal (average Rh···Ph distances
in 55-R and 55-S are 2.30 ꢅ). The instability of 55-S can
once again be explained by the repulsive interactions be-
tween one of the S=O groups and atoms of the Ph group
(see the short distances in Figure 7). In contrast, in the most
stable 55-R intermediate the substrate is placed nicely away
from the upwardly-pointing S=O groups. However, both in-
termediates 55-R and 55-S evolve toward the more stable
56-R and 56-S intermediates in which the C=O group displa-
ces the Ph group from the metal, and an enolate-type struc-
ture h3-coordinated to the metal through the C···C···O
moiety is formed, see Figure 7. Intermediates 56-R and 56-S
are more stable than 55-R and 55-S by 6.9 and 11.5 kcal
molÀ1, respectively.
Concerning the transition states with the wrong regio-
chemistry (1,3-addition instead of 1,4-addition), in which the
C2 atom of 50a attacks the C(Ph) atom (labelled as 54’-R
and 54’-S), they are 9.7 and 5.2 kcalmolÀ1 higher in energy
than 54-R, respectively. Both these transition states are
À
higher in energy because formation of the C2 Ph bond de-
creases conjugation between the C2 atom and the C=O
At this point, the missing step to complete step C of the
catalytic cycle of Scheme 9 is an H transfer to break the
À
À
group more than formation of the C3 Ph bond, see the
Rh C bond and to release the product. Considering that the
À
longer C2 CO distances in 54’-R and 54’-S relative to 54-R.
reaction is performed in a 10:1 toluene/water mixture, we
investigated if a water molecule can coordinate to the Rh
atom of 56-R and can transfer one of its protons to the sub-
strate. Water coordination to 56-R leads to 57-R with an
In addition, 54’-R is also destabilized by severe steric/elec-
trostatic repulsion between the reacting groups and the
ligand, see the short distances in Figure 6.
14344
www.chemeurj.org
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 14335 – 14347

Rhodium-Catalyzed 1,4-Addition
FULL PAPER
energy gain of 10.1 kcalmolÀ1 (see Figure 8). After water co-
ordination, the h3-coordinated C···C···O moiety converts into
an almost perfect enolate moiety with a well formed C=C
dinated water molecule, is achieved. Product release,
through H-transfer from the coordinated water molecule of
57-R to the C=C bond of the enolate bond, is an easy pro-
cess that leads to the final complex 59-R, presenting the
product coordinated to the Rh center through the re-estab-
lished carbonyl functionality. The plot also shows why the
product with the wrong stereochemistry is not observed, as
the transition state 54-S lies 4.4 kcalmolÀ1 higher in energy
than the favored transition state 54-R. Finally, the starting
complex can exist as an equilibrium between different iso-
mers corresponding to square planar, 53-sp-R, and distorted
tetrahedral, 53-dt-R, geometries, connected by a low energy
isomerization barrier.
Figure 8. Relevant structures for water coordination and product release.
Conclusion
We have introduced a new family of bidentate sulfoxide li-
gands with C2-symmetric atropisomeric biaryl backbones.
Combining optically pure or enriched sulfinate esters and
racemic biaryl skeletons by using Andersenꢀs methodology
gave access to a series of sterically and electronically modi-
fied C2-symmetric disulfoxides. Although the methodology
used is well documented, a drawback, relating to the optical
integrity of the ligands, was found by the detection, in some
cases, of diastereoisomers possessing non-homochiral sulfur
stereogenic centers. The reduction of the chiral sulfoxide
moiety to achiral sulfides and HPLC analysis of the atropi-
someric backbones has provided the necessary information
on the purity of each isomer from the constituents of the
family.
IR analysis of the electronic properties of the ligands
through analysis of the carbonyl stretching frequencies of
the corresponding [(disulfoxide)Rh(CO)2]+ complexes gave
a clear picture on how different skeletons and sulfoxide sub-
stituents influence the donor/acceptor properties of the
ligand family. The study indicates that a slightly more acidic
character for the sulfoxide substituents and more electron-
rich backbones provide catalytic systems with superior reac-
tivity and selectivity in the 1,4-addition of phenylboronic
acid to 2-cyclohexen-1-one. In addition, smaller R groups in
the para position of the aromatic rings of the sulfoxide sub-
stituents and bulkier atropisomeric backbones are also in-
creasing the catalytic performance.
À
bond not coordinated to the metal, and a short Rh O-
A
coordinated water molecule to the substrate proceeds
through transition state 58-R with the rather low barrier of
5.6 kcalmolÀ1. In the transition state, the C=C enolate is
almost completely transformed into a single C C bond,
whereas the enolate C O bond is almost completely con-
verted into a standard double C=O bond. In the final inter-
mediate 59-R, which is more stable by 3.5 kcalmolÀ1 than
57-R, the product is coordinated to the metal through its C=
O bond (Figure 8). Finally, reaction product 52aA is re-
leased by simple dissociation of 52aA from the metal with
an energy release of 11.6 kcalmolÀ1 coordinating a water
molecule, and finally regenerating the catalytically active
monomeric [Rh]-OH species of Scheme 9.[46]
The energy profile corresponding to the favored reaction
pathway leading to formation of the R product is shown in
Figure 9. The plot clearly indicates that the first step, corre-
sponding to insertion of the C=C bond of the 2-cyclohexen-
À
À
À
1-one into the Rh Ph bond of the square-planar 53-sp-R
complex is rate limiting. After insertion has occurred, rapid
transformation of the resulting intermediate into the Rh
complex 57-R, presenting the Rh-enolate bond and a coor-
Advanced DFT calculations have shown that the initial
binding of the olefin is not the enantio-discriminating step
(at least for the disulfoxide-based catalysts used here). Fur-
thermore, the final protonation step involving the enolate-
metal moiety follows a stepwise mechanism whereby initial
binding of a water molecule to the rhodium center precedes
an intramolecular protonation and release mechanism of the
enolate. Calculations on the reaction path have also uncov-
ered the mechanism by which these disulfoxide–rhodium
catalysts discriminate between the two possible enantiomer-
ic products. Contrary to more traditional chiral ligand
frameworks, electronic factors arising from the sulfoxide
moiety seem to be, at least, partially responsible for the high
Figure 9. Energy profile corresponding to formation of the favored prod-
uct (step C in Scheme 9). The energy of transition state 54-S, leading to
the minor enantiomer, is also indicated.
Chem. Eur. J. 2010, 16, 14335 – 14347
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
14345

R. Dorta, L. Cavallo et al.
[7] a) A. J. Walker, Tetrahedron: Asymmetry 1992, 3, 961; b) G. Solladie,
M. C. Carrefio in Organosulphur Chemistry, Synthetic Aspects, Aca-
demic Press, New York, 1995; c) G. H. Posner in The Chemistry of
Sulfones and Sulfoxides, Wiley, New York, 1988; d) G. H. Posner,
Acc. Chem. Res. 1987, 20, 72; e) M. C. CarreÇo, Chem. Rev. 1995,
95, 1717.
enantioselectivities observed in catalysis. These findings
should prove useful for future research in the use of sulfox-
ide based ligands, allowing a rational design of structures
that take advantage of the phenomenon described here.
[8] For achiral catalysis with sulfoxide ligands, see: a) Y. M. Y. Haddad,
H. B. Henbest, J. Husbands, T. R. B. Mitchell, J. Trocha-Grimshaw, J.
Chem. Soc. Perkin Trans. 1 1974, 596; b) B. R. James, R. H. Morris,
J. Chem. Soc. Chem. Commun. 1978, 929; c) A. Greenberg, A.
Gogoll, J.-E. Backvall, J. Org. Chem. 1991, 56, 5808; d) R. C.
Larock, R. Hightower, L. A. Hasvold, K. P. Peterson, J. Org. Chem.
1996, 61, 3584; e) M. S. Chen, M. C. White, J. Am. Chem. Soc. 2004,
126, 1346; f) K. J. Fraunhoffer, N. Prabagaran, L. E. Sirois, M. C.
White, J. Am. Chem. Soc. 2006, 128, 9032.
[9] For attempts showing low selectivities/reactivities: a) B. R. James,
R. S. McMillan, Can. J. Chem. 1977, 55, 3927; b) N. Khiar, I. Fernan-
dez, F. Alcudia, Tetrahedron Lett. 1993, 34, 123; c) R. Tokunoh, M.
Sodeoka, K. Aoe, M. Shibasaki, Tetrahedron Lett. 1995, 36, 8035.
For two very recent examples of high enantioselectivity using disulf-
oxide ligands, see: d) Q.-A. Chen, X. Dong, M.-W. Chen, D.-S.
Wang, Y.-G. Zhou, Y.-X. Li, Org. Lett. 2010, 12, 1928; e) J. Chen, J.
Chen, F. Lang, X. Zhang, L. Cun, J. Zhu, J. Deng, J. Liao, J. Am.
Chem. Soc. 2010, 132, 4552.
[10] For better reactivities and selectivities using sulfoxide/nitrogen or
sulfoxide/phosphine chelates, see: a) J. V. Allen, J. F. Bower, J. M. J.
Williams, Tetrahedron: Asymmetry 1994, 5, 1895; b) K. Hiroi, Y.
Suzuki, Tetrahedron Lett. 1998, 39, 6499; c) K. Hiroi, Y. Suzuki, I.
Abe, Y. Hasegawa, K. Suzuki, Tetrahedron: Asymmetry 1998, 9,
3797; d) D. G. I. Petra, P. C. J. Kamer, A. L. Spek, H. E. Schoemak-
er, P. W. N. M. van Leeuwen, J. Org. Chem. 2000, 65, 3010; e) K.
Hiroi, Y. Suzuki, I. Abe, R. Kawagishi, Tetrahedron 2000, 56, 4701;
f) K. Hiroi, I. Izawa, T. Takizawa, K. Kawai, Tetrahedron 2004, 60,
2155; g) J. Chen, D. Li, H. Ma, L. Cun, J. Zhu, J. Deng, J. Liao, Tet-
rahedron Lett. 2008, 49, 6921; h) J. Chen, F. Lang, D. Li, L. Cun, J.
Zhu, J. Deng, J. Liao, Tetrahedron: Asymmetry 2009, 20, 1953; i) F.
Lang, D. Li, J. Chen, J. Chen, L. Li, L. Cu, J. Zhu, J. Deng, J. Liao,
Adv. Synth. Catal. 2010, 352, 843.
Experimental Section
All computational and experimental details can be found in the Support-
ing Information.
CCDC-781539–781553 contain the supplementary crystallographic data
for this paper (see Supporting Information for further details. These data
can be obtained free of charge from The Cambridge Crystallographic
Data Centre via www.ccdc.cam.ac.uk/data_request/cif. The Crystallo-
graphic data for compounds 16a’, 17b, 32 and 33 have been published
elsewhere.[16]
Acknowledgements
R.D. holds an Alfred Werner Assistant Professorship and thanks the
foundation for generous financial support. R.M. thanks the Swiss Nation-
al Science Foundation (SNF) and E.D. thanks the Roche Research Foun-
dation (RRF) for support. A.P. is grateful for the allocation of a Ramꢆn
y Cajal contract by the Spanish MEC. L.C. thanks the HPC team of
Enea (www.enea.it) for using the ENEA-GRID and the HPC facilities
CRESCO (www.cresco.enea.it) in Portici, Italy.
[1] a) Asymmetric Catalysis in Organic Synthesis (Ed.: R. Noyori),
Wiley, New York, 1994; b) Catalytic Asymmetric Synthesis II (Ed.:
Ojima), Wiley, New York, 2000; c) Comprehensive Asymmetric Cat-
alysis (Eds.: E. N. Jacobsen, A. Pfaltz, H. Yamamoto), Springer,
Berlin, 1999; d) H. U. Blaser, F. Spindler, M. Studer, Appl. Catal. A
2001, 221, 119.
[11] a) K. K. Andersen, W. Gaffield, N. E. Papanikolaou, J. W. Foley,
R. I. Perkins, J. Am. Chem. Soc. 1964, 86, 5637; b) K. K. Andersen,
Int. J. Sulfur Chem., Part B 1971, 6, 69; c) C. Mioskowski, G. Solla-
diꢈ, Tetrahedron 1980, 36, 227; d) C. Mioskowski, G. Solladiꢈ, Tetra-
hedron Lett. 1975, 16, 3341; e) G. Solladiꢈ, J. Hutt, A. Girardin, Syn-
thesis 1987, 173; f) S. C. Benson, J. K. Snider, Tetrahedron Lett. 1991,
32, 5885; g) F. Rebiere, O. Samuel, L. Ricard, H. B. Kagan, J. Org.
Chem. 1991, 56, 5991; h) J. K. Whitesell, M.-S. Wong, J. Org. Chem.
1991, 56, 4552; i) D. A. Evans, M. M. Faul, L. Colombo, J. J. Bisaha,
J. Clardy, D. Cherry, J. Am. Chem. Soc. 1992, 114, 5977; j) W. Op-
polzer, Pure Appl. Chem. 1990, 62, 1241; k) I. Fernꢇndez, N. Khiar,
J. M. Llera, F. Alcudia, J. Org. Chem. 1992, 57, 6789.
[2] a) S. C. Stinson, Chem. Eng. News 1992, 70, 46; b) S. C. Stinson,
Chem. Eng. News 2001, 79, 45; c) S. Borman, Chem. Eng. News
1990, 68, 9.
[3] a) J. K. Whitesell, Chem. Rev. 1989, 89, 1581; b) A. Pfenninger, Syn-
thesis 1986, 89; c) R. Noyori, I. Tomino, M. Yamada, M. Nishixawa,
J. Am. Chem. Soc. 1984, 106, 6709; d) T. R. Kelly, A. Whiting, N. S.
Chandrakumar, J. Am. Chem. Soc. 1986, 108, 3510; e) K. Narasaka,
M. Inoue, N. Okada, Chem. Lett. 1986, 1109; f) E. J. Corey, N. Imai,
H.-Y. Zhang, J. Am. Chem. Soc. 1991, 113, 728; g) D. A. Evans,
K. A. Woerpel, M. M. Hinman, M. M. Faul, J. Am. Chem. Soc. 1991,
113, 726; h) R. E. Lowenthal, A. Abiko, S. Masamune, Tetrahedron
Lett. 1990, 31, 6005.
[12] a) J. Clayden, P. M. Kubinski, F. Sammiceli, M. Helliwell, L. Diora-
zio, Tetrahedron 2004, 60, 4387; b) F. Garcꢉa-Flores, L. S. Flores-
Michel, E. Juaristi, Tetrahedron Lett. 2006, 47, 8235.
[4] For an excellent recent review on the use of chiral olefin containing
ligands, see: a) C. Defieber, H. Grꢁtzmacher, E. M. Carreira,
Angew. Chem. 2008, 120, 4558; Angew. Chem. Int. Ed. 2008, 47,
4482. For selected references, see: b) T. Hayashi, K. Ueyama, N. To-
kunaga, K. Yoshida, J. Am. Chem. Soc. 2003, 125, 11508; c) N. Toku-
naga, Y. Otomaru, K. Okamoto, K. Ueyama, S. Shintani, T. Hayashi,
J. Am. Chem. Soc. 2004, 126, 13584; d) R. Shintani, K. Okamoto, Y.
Otomaru, K. Ueyama, T. Hayashi, J. Am. Chem. Soc. 2005, 127, 54;
e) C. Fischer, C. Defieber, T. Suzuki, E. M. Carreira, J. Am. Chem.
Soc. 2004, 126, 1628; f) C. Defieber, J.-F. Paquin, S. Serna, E. M.
Carreira, Org. Lett. 2004, 6, 3873; g) J.-F. Paquin, C. R. J. Stephen-
son, C. Defieber, E. M. Carreira, Org. Lett. 2005, 7, 3821; h) J.-F.
Paquin, C. Defieber, C. R. J. Stephenson, E. M. Carreira, J. Am.
Chem. Soc. 2005, 127, 10850.
[13] T. Thaler, F. Geittner, P. Knochel, Synlett 2007, 2655.
[14] a) A. Miyashita, A. Yasuda, H. Takaya, K. Toriumi, T. Ito, T.
Souchi, R. Noyori, J. Am. Chem. Soc. 1980, 102, 7932; b) H. Takaya,
K. Mashima, K. Koyano, M. Yagi, H. Kumobayashi, T. Taketomi, S.
Akutagawa, R. Noyori, J. Org. Chem. 1986, 51, 629.
[15] R. Schmid, M. Cereghetti, B. Heiser, P. Schçnholzer, H.-J. Hansen,
Helv. Chim. Acta 1988, 71, 897.
[16] a) R. Mariz, X. Luan, M. Gatti, A. Linden, R. Dorta, J. Am. Chem.
Soc. 2008, 130, 2172; b) J. J. Bꢁrgi, R. Mariz, M. Gatti, E. Drinkel,
X. Luan, S. Blumentritt, A. Linden, R. Dorta, Angew. Chem. 2009,
121, 2806; Angew. Chem. Int. Ed. 2009, 48, 2768.
[17] H. Shimizu, I. Nagasaki, T. Saito, Tetrahedron 2005, 61, 5405.
[18] Z. Freixa, P. W. N. M. van Leeuwen, Dalton Trans. 2003, 1890.
[19] a) K. Mashima, K.-H. Kusano, N. Sato, Y.-I. Matsumura, K. Nozaki,
H. Kumobayashi, N. Sayo, Y. Hori, T. Ishizaki, S. Akutagawa, H.
Takaya, J. Chem. Soc. Chem. Commun. 1991, 609; b) K. Mashima,
[5] a) M. Calligaris, O. Carugo, Coord. Chem. Rev. 1996, 153, 83;
b) J. C. Bayꢆn, C. Claver, A. M. Masdeu-Bultꢆ, Coord. Chem. Rev.
1999, 193, 73; c) I. Fernꢇndez, N. Khiar, Chem. Rev. 2003, 103, 3651.
[6] a) D. R. Rayner, A. J. Gordon, K. Mislow, J. Am. Chem. Soc. 1968,
90, 4854; b) K. Mislow, J. Siegel, J. Am. Chem. Soc. 1984, 106, 3319.
14346
www.chemeurj.org
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 14335 – 14347

Rhodium-Catalyzed 1,4-Addition
FULL PAPER
K. Kusano, N. Sato, Y. Matsumura, K. Nozaki, H. Kumobayashi, N.
Sayo, Y. Hori, T. Ishizaki, S. Akutagawa, H. Takaya, J. Org. Chem.
1994, 59, 3064.
[35] In the case of RhI complexes, donor–acceptor properties of diphos-
phines have been investigated by studying the carbonyl stretching
frequencies of [(diphosphine)RhCl(CO)] compounds, see: a) L. Val-
larino, J. Chem. Soc. 1957, 2287; b) W. Hieber, R. Kummer, Chem.
Ber. 1967, 100, 148; c) A. R. Sanger, J. Chem. Soc. Dalton Trans.
1977, 120; d) S. Jeulin, S. Duprat de Paule, V. Ratovelomanana-
Vidal, J.-P. GenÞt, N. Champion, P. Dellis, Angew. Chem. 2004, 116,
324; Angew. Chem. Int. Ed. 2004, 43, 320; e) D. E. Kim, C. Choi,
I. S. Kim, S. Jeulin, V. Ratovelomanana-Vidal, J.-P. GenÞt, N. Jeong,
Adv. Synth. Catal. 2007, 349, 1999.
[20] For the proposed selectivity pathway in the original BINAP–RhI
system, see: a) Y. Takaya, M. Ogasawara, T. Hayashi, M. Sakai, N.
Miyaura, J. Am. Chem. Soc. 1998, 120, 5579. For a review on the
rhodium-catalyzed 1,4-addition: b) T. Hayashi, K. Yamasaki, Chem.
Rev. 2003, 103, 2829.
[21] A. Poater, F. Ragone, R. Mariz, R. Dorta, L. Cavallo, Chem. Eur. J.
2010, 16, 14348.
[22] Comopund rac-3 is reasonably priced when purchased in >100 g
quantities from Dalchem/ABCR (<10$/g). Alternatively, it can be
prepared in good quantities from the corresponding alcohol, see: A.
Miyashita, H. Takaya, T. Souchi, R. Noyori, Tetrahedron 1984, 40,
1245.
[23] a) G. Svensson, J. Albertsson, T. Frejd, T. Klingsted, Acta Crystal-
logr. 1986, C42, 1324; b) Y. Liang, S. Gao, H. Wan, J. Wang, H.
Chen, Z. Zheng, X. Hu, Tetrahedron: Asymmetry 2003, 14, 1267; see
also reference [15a].
[24] For synthesis and structural features of its diphosphine counterpart,
see: a) N. Sayo, H. Kumobayashi EP 479 542, 1992; [Chem. Abstr.
1992, 117, 111803]; b) X. Zhang, K. Mashima, K. Koyano, N. Sayo,
H. Kumobayashi, S. Akutagawa, H. Takaya, Tetrahedron Lett. 1991,
32, 7283; c) X. Zhang, K. Mashima, K. Koyano, N. Sayo, H. Kumo-
bayashi, S. Akutagawa, H. J. Takaya, Chem. Soc. Perkin Trans. 1
1994, 2309.
[25] a) C. Alayrac, S. Nowaczyk, M. Lemariꢈ, P. Metzner, Synthesis 1999,
669. For X-ray analysis see: b) C. Alayrac, J.-F. Saint-Clair, M. Le-
mariꢈ, P. Metzner, M.-T. Averbuch-Pouchot, Acta Crystallogr. Sect.
A 1999, 55, 262.
[36] The differences in average stretching frequencies are big enough to
exclude misinterpretation, owing to different bite angles (and result-
ing geometry/orbital overlap differences) between the complexes.
For stretching frequencies of other rhodium-dicarbonyl complexes
with chelating diphosphine ligands, see, for example: T. A. Betley,
J. C. Peters, Angew. Chem. 2003, 115, 2487; Angew. Chem. Int. Ed.
2003, 42, 2385.
[37] Such a direct comparison of different ligand families can be mislead-
ing. The data obtained here do not imply that the disulfoxide ligands
are binding more tightly to RhI than diphosphines, but merely re-
flect the relative contributions of donor and accepting properties of
different ligand families. For a recent computational analysis on the
subject, see: D. G. Gusev, Organometallics 2009, 28, 763.
[38] For discussions on the effect of dihedral angles on selectivity and re-
activity, see for example: a) X. Zhang, T. Taketomi, T. Yoshizumi,
H. Kumobayashi, S. Akutagawa, K. Mashima, H. Takaya, J. Am.
Chem. Soc. 1993, 115, 3318; b) T. Uemura, X. Zhang, K. Matsu-
mura, N. Sayo, H. Kumobayashi, T. Ohta, K. Nozaki, H. Takaya, J.
Org. Chem. 1996, 61, 5510; c) T. Shinohara, K. Kondo, H. Ogawa, T.
Mori, K. Nozaki, T. Hiyama, Chirality 2000, 12, 425.
[26] Both (1R,2S,5R)-(À)-menthyl (S)-p-toluenesulfinate and (1S,2R,5S)-
(+)-menthyl (R)-p-toluenesulfinate can be purchased in high de ra-
tio from various chemical suppliers.
[27] J. M. Klunder, K. B. Sharpless, J. Org. Chem. 1987, 52, 2598.
[28] The protocol reported in the Supporting Information of this manu-
script gives superior yields of p-Tol-BINASO compared to the pro-
cedures given in reference [12a] and reference [16a].
[29] Ortho-metallation of sulfoxides is well-documented and sometimes
directs the reaction outcome to undesired products, for example:
R. J. Kloetzing, P. Knochel, Tetrahedron: Asymmetry 2006, 17, 116.
[30] The crystal structures of ligands 19a and 19c show that these com-
pounds are not enantiopure as expected (see purity of starting sulfi-
nate 10). The major isomer was determined indirectly by HPLC and
the catalytic results (see: reduction to disulfides and enantiomeric
purity of ligands) to have the S,S-configuration at the sulfur centers.
Likewise, ligand 23b contains the R,R-configured isomer as the
major component.
[31] a) H. Bartsch, T. Erker, Tetrahedron Lett. 1992, 33, 199; b) N.
Tewari, Y. Kumar, R. K. Thaper, J. M. Khanna, Synth. Commun.
1996, 26, 1169; c) K. Shimada, K. Kodaki, S. Aoyagi, Y. Takikawa,
C. Kabuto, Chem. Lett. 1999, 695; d) A. Ishii, R. Yamashita, M.
Saito, J. Nakayama, J. Org. Chem. 2003, 68, 1555; e) D. K. Bates, X.
Li, P. V. Jog, J. Org. Chem. 2004, 69, 2750.
[32] The disulfides 24–31 were subjected to identical reaction conditions
as for their synthesis using elevated temperatures (up to 1008C) and
extended reaction times (up to 2 days). No change in enantiomeric
ratios was found as determined by HPLC analysis.
[33] The other isomers of ligands 16–23 did not form clean complexes of
the corresponding dimers and (DNHS)-conformers of the ligands
were not used in our study. Our hypothesis is that steric repulsion
from the sulfoxide groups hinders formation of the bridged com-
plexes, a phenomenon that was also observed for atropisomeric di-
phosphine ligands incorporating bulky aromatic groups on phospho-
rus atoms, see: T. Ohshima, H. Tadaoka, K. Hori, N. Sayo, K. Mashi-
ma, Chem. Eur. J. 2008, 14, 2060.
[39] T. Shibata, H. Tsuruta, H. Danjo, T. Imamoto, J. Mol. Catal. A 2003,
196, 117.
[40] a) T. Hayashi, M. Takahashi, Y. Takaya, J. Am. Chem. Soc. 2002,
124, 5052; b) A. Kina, H. Iwamura, T. Hayashi, J. Am. Chem. Soc.
2006, 128, 3904.
[41] A recent report on the use of modified atropisomeric diphosphines
as ligands for rhodium-catalyzed 1,4-addition in toluene/water has
come to similar conclusions, see: T. Korenaga, K. Osaki, R. Mae-
nishi, T. Sakai, Org. Lett. 2009, 11, 2325.
[42] A similar effect has been proposed for conjugated additions with
rhodium quinonoid catalysts, see: a) U. S. Son, S. B. Kim, J. A. Rein-
gold, G. B. Carpenter, D. A. Sweigart, J. Am. Chem. Soc. 2005, 127,
12238; b) W. C. Trenkle, J. L. Barkin, S. U. Son, D. A. Sweigart, Or-
ganometallics 2006, 25, 3548.
[43] K. E. Koenig, M. J. Sabacky, G. L. Bachman, W. C. Christopfel,
H. D. Barnstorff, R. B. Friedman, W. S. Knowles, B. R. Stults, B. D.
Vineyard, D. J. Weinkauff, Ann. N. Y. Acad. Sci. 1980, 333, 16.
[44] Details of DFT calculations are included in the Supporting Informa-
tion.
[45] a) T. Itoh, T. Mase, T. Nishikata, T. Iyama, H. Tachikawa, Y. Ko-
bayashi, Y. Yamamoto, N. Miyaura, Tetrahedron 2006, 62, 9610. For
DFT calculations on a related 1,4-addition–enantioselective proto-
nation reaction, see: b) L. Navarre, R. Martinez, J.-P. Genet, S.
Darses, J. Am. Chem. Soc. 2008, 130, 6159.
[46] We also tested if the H-transfer from the coordinated water mole-
cule of 57-R could follow a two-step pathway, in which the H is first
transferred to the Rh with formation of a Rh-hydride species (oxi-
dative addition of water) followed by reductive elimination of the
substrate. The H-transfer to the metal has a barrier of 25.3 kcal
molÀ1, which is 19.7 kcalmolÀ1 higher than the barrier for the direct
transfer from the water molecule to the substrate, supporting the
low energy mechanism pathway shown in Figure 8.
[34] K. A. Bunten, D. H. Farrar, A. J. Poe, A. Lough, Organometallics
2002, 21, 3344.
Received: June 21, 2010
Published online: November 12, 2010
Chem. Eur. J. 2010, 16, 14335 – 14347
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
14347