Systematic study of the asymmetric methoxycarbonylation of styrene catalyzed by palladium systems containing chiral ferrocenyl diphosphine ligands

Article abstract of DOI:10.1002/hlca.200690161

We present the first systematic study of the Pd-catalyzed asymmetric methoxycarbonylation of styrene in the presence of chiral ferrocenyl phosphine ligands. The reaction conditions were optimized, and a screening of different catalyst precursors was performed. A number of 1,1′-bis(phosphino) ferrocenes of the Mandyphos, Josiphos, Walphos, and Taniaphos types were tested in combination with [PdCl₂(NCPh)₂], in the presence of TsOH as the acid source. These systems afforded high enantioselectivities, although the regioselectivity of the reaction was found to be in favor of the (undesired) linear ester. The catalytic system made with the Josiphos ligand 1 gave rise to an enantiomeric excess (ee) of 86%.

Full text of DOI:10.1002/hlca.200690161

1610
Helvetica Chimica Acta – Vol. 89 (2006)
Systematic Study of the Asymmetric Methoxycarbonylation of Styrene
Catalyzed by Palladium Systems Containing Chiral Ferrocenyl Diphosphine
Ligands
by Cyril Godard*, Aurora Ruiz, and Carmen Claver*
Departament de Química Física i Inorgànica, Universitat Rovira i Virgili, Marcelli Domingo w/n,
43007 Tarragona, Spain
(phone: +34-977-559575; fax: +34-977-559563, e-mail: cygodard@yahoo.fr or carmen.claver@urv.net)
Dedicated to Professor Giambattista Consiglio in recognition of his contributions to
enantioselective catalysis
We present the first systematic study of the Pd-catalyzed asymmetric methoxycarbonylation of sty-
rene in the presence of chiral ferrocenyl phosphine ligands. The reaction conditions were optimized, and
a screening of different catalyst precursors was performed. A number of 1,1’-bis(phosphino)ferrocenes of
the Mandyphos, Josiphos, Walphos, and Taniaphos types were tested in combination with
[PdCl2(NCPh)2], in the presence of TsOH as the acid source. These systems afforded high enantioselec-
tivities, although the regioselectivity of the reaction was found to be in favor of the (undesired) linear
ester. The catalytic system made with the Josiphos ligand 1 gave rise to an enantiomeric excess (ee) of
86%.
Introduction. – Recently, chiral ferrocenyl phosphine ligands have been successfully
used in many catalytic processes [1]. These compounds constitute a family of poten-
tially chelating bi- or tridentate ligands, and their success is largely due to their unique
structural features that allow multiple modifications by varying the properties of the
substituents on the cyclopentadienyl (Cp) ligands as well as on the P-atoms. The use
of ligands bearing analogous metallocenes has also been reported [2].
The prototypical ligand 1,1’-bis(diphenylphosphino)ferrocene (dppf) has been
extensively used in Pd-catalyzed CÀC and CÀN bond-forming reactions such as
cross-coupling [3], Heck reaction [4], carbonylation of chloro arenes [5], aryl halide
amination [6], hydroamination of alkynes [7], and methoxycarbonylation of ethene
[8]. High enantioselectivities were reported by Xiao and Zhang [9] for the asymmetric
hydrogenation of acyclic imines with Ir ferrocene-binaphane complexes [9]. In recent
years, 1,2-bis(phosphino)ferrocene has also been used in homogeneous catalysis and
given rise to high enantioselectivities in various reactions. For instance, the Solvias–
Josiphos ligands (Fig. 1), which combine planar chirality with the central chirality of
a side group, constitute versatile ligands that can be used in a number of stereoselective
processes by simply varying the substituents on the P-atoms [10]. Several related Josi-
phos ligands are being fruitfully employed in the Rh-catalyzed asymmetric hydrogena-
tion of alkenes on industrial scale. Typically, the enantiomeric excess (ee) obtained in
the hydrogenation of dimethyl itaconate (DMI) or methyl 2-acetamidoacrylate (MAA)
© 2006 Verlag Helvetica Chimica Acta AG, Zürich

Helvetica Chimica Acta – Vol. 89 (2006)
1611
is in the range 90–99.9%. An industrial process based on Ru catalysts containing Josi-
phos ligands has also been developed for the large-scale production of methyl (+)-cis-
dihydrojasmonate [11].
Fig. 1. Selected ferrocenyl diphosphine ligands used in asymmetric processes
Josiphos-type ligands are also the most-effective ligands for the Ir-catalyzed hydro-
genation of imines, with high enantioselectivities, and turnover-frequency (TOF) values
in the range of 10 to 300000 hÀ1 for aryl imines [12]. The hydrogenation of 2-methyl-6-
ethylphenyl-1’-methyl-2’-methoxyethylimine to the corresponding optically enriched
(80% ee) amine using a Josiphos-containing Ir catalyst is nowadays the largest-scale
enantioselective catalytic process in industry [13], and the Walphos family is produced
on a technical scale by Solvias. The nature of the active Ir species involved in this pro-
cess was recently examined in detail by Togni and co-workers [14]. They showed by X-
ray crystallography that the substrate coordinates in a k2 fashion to an IrI center.
Kagan and co-workers [15] have used Josiphos ligands, differing from the Solvias–
Josiphos type by only displaying planar chirality, in the asymmetric hydrogenation of
(Z)-a-methylacetamidocinnamate, with ee values up to 84%. These ligands were also
screened in various Rh-catalyzed hydrogenation reactions affording enantioselectivi-
ties comparable to those obtained with Solvias–Josiphos ligands. The authors have,
therefore, suggested that the planar chirality may be of major importance in chiral
induction during the hydrogenation of prochiral alkenes [15]. Also, the experimental
data obtained by Hayashi and Kumada [16 ] for Grignard cross-coupling reactions,
when investigating the effect of the central chirality of the stereogenic C-atom of the
Josiphos ligand, led them to the conclusion that, in these reactions, planar chirality
plays the dominant role in terms of product configuration [16]. Furthermore, Pastor
and Togni [17] investigated the effect of ligand chirality on the enantio- and diastereo-
selectivity of the reaction using (R,S)-, (S,S)-, and (S,R)-configured Josiphos ligands;
they concluded that planar and central chirality might act either in a cooperative or
in a non-cooperative way, and that, therefore, both stereogenic elements are important
[17].
The Walphos ligands (Fig. 1) were also used successfully in various homogeneous
asymmetric hydrogenation reactions [18]. In particular, the formation of (R)-2-alkyl-
3-phenylpropanoic acid from the corresponding prochiral alkene was achieved in
95% ee and 100% yield with a Rh catalyst formed in situ in the presence of a Walphos
diphosphine (R=C6H5, R’=3,5-(CF3)2ÀC6H3) [19]. These ligands differ from the Josi-
phos compounds by a C6H4 spacer between a Cp ring and a P-atom. Thus, the coordi-
nation of both P-atoms would lead to the formation of eight-membered metalla rings.

1612
Helvetica Chimica Acta – Vol. 89 (2006)
Eight-membered chelates are also formed with the Taniaphos ligands (Fig. 1),
which display a stereogenic center between a Cp ring and a C6H4 spacer, in addition
to the planar chirality. These diphosphines have been shown to be particularly efficient
in Rh-catalyzed hydrogenation reactions of functionalized C=C bonds in alkenes, b-
keto esters, 1,3-diketones, and hydrazones [11b].
Such ligands were also used in carbonylation processes. Rh-catalyzed hydroformy-
lation of styrenes was reported using Solvias–Josiphos diphosphines, with ee values of
up to 76% [20]. In these systems, both activity and selectivity were shown to be affected
by the phosphorus substituents. The highest ee values were, in this case, obtained with
diphenyl-alkyldiaryl diphosphines. Van Leeuwen and co-workers [20] also reported the
use of Rh systems with chiral ferrocenyl diphosphines for this process (Fig. 2). How-
ever, relatively low yields and regioselectivities were obtained, although both TOF
and ee values could be improved with electron-withdrawing groups on the P-atoms
[21].
Fig. 2. Chiral ferrocenyl diphosphine ligands used in the asymmetric hydroformylation of styrene
The asymmetric Pd-catalyzed methoxycarbonylation of vinyl arenes has been
extensively studied as a route to 2-arylpropanoic acids (Scheme), the most-important
class of non-steroidal anti-inflammatory drugs [22]. Catalysts containing bidentate
diphosphine ligands have been frequently used for this transformation, but generally
give rise to relatively low regioselectivities of the branched product, despite good enan-
tioselectivities [23]. With systems containing bidentate diphosphines, the regioselectiv-
ity of the reaction was shown to vary dramatically with the length of the carbon chain
separating the two P-atoms [24]. Nevertheless, the highest ratio of branched to linear
ester are usually obtained with monophosphine-based systems, although these gener-
ally afford low enantioselectivities [25].
Scheme. Palladium-Catalyzed Methoxycarbonylation of Styrene
Recently, a few successful chiral systems have been reported for the asymmetric
methoxycarbonylation of styrene with 1,1’-bis(phosphino)ferrocene ligands (Fig. 3)
[26]. Using an oxazoline-substituted diphosphine (Fig. 3,a), Wang et al. [26b] obtained

Helvetica Chimica Acta – Vol. 89 (2006)
1613
Fig. 3. Ferrocenyl diphosphine ligands used by a) Wang and co-workers [26b] and b) Inoue and co-
workers [26a]
40% of the branched ester and 64% ee, while Inoue and co-workers [26a] reported 44%
regioselectivity for the branched product with 86% ee for R=(Me)2N (Fig. 3,b).
G
In this paper, we report a study on the catalyzed methoxycarbonylation of styrene
(Scheme) with the aid of PdII and Pd0 precursors and Solvias-type ferrocenyl-1,2- and
ferrocenyl-1,1’-diphosphine ligands. These ligands can be classified in four distinct fam-
ilies: Josiphos (1–5), Mandyphos (6), Walphos (7 and 8), and Taniaphos (9 and 10).
Optimization of the reaction conditions and comparison of the performance of various
Pd precursors were performed with ligand 1. The results of the Pd-catalyzed methoxy-
carbonylation of styrene are discussed, and a comparison of the rate of conversion and
selectivity as a function of the different types of ligands is made.
Experimental. – 1. General. All reactions were carried out in dried and distilled solvents under N2
atmosphere using standard Schlenk techniques. The catalyst precursor [PdCl2(NCPh)2] was prepared
according to [27]. Compounds [PdCl2], [Pd2(dba)3], [Pd(acac)2], and [Pd(OAc)2] were purchased from
Johnson Mattey, and used without further purification. The diphosphine ligands were offered by Solvias.
Carbonylation reactions were performed in a 100-ml stainless-steel Bergof autoclave magnetically stirred
and electrically heated. Other chemicals and solvents were purchased from commercial sources and used
without further purification, unless otherwise stated. Carbon monoxide (CO; 99.99%) was supplied by
Air Liquide. All NMR measurements were performed on Varian Mercury VX-400 or Varian Gemini-
300 spectrometers in 5-mm triple or 10-mm high-pressure tubes. High-pressure experiments were carried
out in a 10-mm sapphire tube; chemical shifts d (in ppm) for 1H- and 13C-NMR rel. to residual solvent
signals, coupling constants J in Hz. 31P-NMR Chemical shifts were referenced to H3PO4 as external stan-
dard. Elemental analyses were carried out on a Carlo Erba Microanalyser EA-1108. The chemical and
regioselectivity of reactions were determined by gas-chromatographic (GC) analyses on a Hewlett–Pack-
ard 5890A instrument (split/splitless injector, J&W Scientific; Ultra-2 25-m column, i.d. 0.2 mm, film
thickness 0.33 mm; 150 kPa He as carrier gas; FID detector) equipped with a Hewlett-Packard 3396 ser-
ies-II integrator. The enantiomeric excess (ee) was determined by HPLC analysis (Daicel CHIRACEL
OJ; hexane/i-PrOH 95 :5, 1.5 ml minÀ1).
2. Typical Procedure for Methoxycarbonylation Reactions. A mixture of THF/MeOH 1:1 (5 ml) con-
taining the Pd precursor, the appropriate diphosphine, and the acid was prepared under N2 atmosphere
using Schlenk techniques. Immediately after addition of the substrate, the soln. was introduced into the
evacuated autoclave by aspiration, and then heated and pressurized with CO. After 24 h, the autoclave
was cooled and slowly depressurized. After separation by column chromatography (CC), the obtained
soln. was analyzed by GC, and the ee was determined by HPLC.

1614
Helvetica Chimica Acta – Vol. 89 (2006)
3. Synthesis of Pd Complexes1). 3.1. Synthesis of Neutral Complexes of the type [PdCl2(PÀP’)] ([PÀ
P’=1, 9, 10). The appropriate ligand (1.1 equiv.) in toluene (10 ml) was added to a soln. of
[PdCl2(NCPh)2] at r.t., and the mixture was stirred for 1 h. During this time, the color of the soln. changed
from dark-yellow to bright-orange. The volume of the soln. was then reduced to ca. 5 ml, and the product
was precipitated by addition of hexane. The resulting solid was filtered, washed with several fractions of
hexane, and dried under vacuum.
Data of [PdCl2(1)]. Yield: 91%. 1H-NMR (CDCl3): 3.61 (s, h5-C5H5); 4.00 (br., 1 H of h5-C5H3); 4.35
(br., 1 H of h5-C5H3); 4.42 (br., 1 H of h5-C5H3); 3.66 (m, MeCH); 0.99 (dd, J(H,H)=7, J(P,H)=13,
MeCH); 2.18 (s, 2 Me of Tol); 2.41 (s, 2 Me of Tol); 7.0–8.1 (16 arom. H). 31P{1H}-NMR (CDCl3): 16.4
(d, J(P,P)=18, PPh2); 47.85 (d, J(P,P)=18, P(Tol)2). 13C{1H}-NMR (CDCl3): 125–140 (arom. signals);
20.8 (d, J(P,C)=2, MeCH); 33.5 (dd, J(PC)=28, 7, MeCH); 71.5 (s, h5-C5H5); 68.8 (d, J(P,C)=11, h5-
C5H3); 70.2 (d, J(P,C)=7, h5-C5H3); 71.8 (dd, J(P,C)=6, 7, h5-C5H3); 74.7 (d, J(P,C)=4, h5-C5H3); 93.5
(dd, J(P,C)=18, 4, h5-C5H3); 22.2 (s, Me of Tol); 22.4 (s, Me of Tol). Anal. calc. for C40H40Cl2FeP2Pd:
C 58.89, H 4.94; found: C 58.68, H 5.58.
Data of [PdCl2(9)]. Yield: 87%. 1H-NMR (CDCl3): 2.35 (s, Me2
(br., 1 H of h5-C5H3); 4.64 (s, h5-C5H5); 5.01 (br., 1 H of h5-C5H3); 5.54 (s, Me2
PÀFc); 23.9 (s, Ph2
31P{1H}-NMR (CDCl3): 17.0 (s, Ph2 PÀC6H4). 13C{1H}-NMR (CDCl3): 21.7 (s,
Me2CH); 44.0 (s, Me2
NCH); 62.75 (d, J(P,C)=12, h5-C5H3); 69.75 (d, J(P,C)=6, h5-C5H3); 71.3 (d, J(P,
ACHTREUNG
AHCTREUNG
A
ACHTREUNG
G
C)=8, h5-C5H3); 72.6 (s, h5-C5H5); 76.2 (d, J(P,C)=2, h5-C5H3); 96.95 (d, J(P,C)=14, h5-C5H3),
120–145 (30 arom. C). Anal. calc. for C43H38Cl2FeP2Pd: C 59.71, H 4.55, N 1.62; found: C 59.64, H
4.90, N 1.72.
1
)
Abbreviations: Cy, cyclohexyl; Fc, ferrocenyl; PÀP’, diphosphine ligands (see chemical formulae);
Tol, tolyl; Ts, tosyl (=(4-methylphenyl)sulfonyl).

Helvetica Chimica Acta – Vol. 89 (2006)
1615
Data of [PdCl2(10)]. Yield: 76%. 1H-NMR (CDCl3): 0–3.5 (PÀCy); 2.74 (s, Me2
ACHTREUNG
h5-C5H3); 4.42 (s, 1 H of h5-C5H3); 4.57 (s, h5-C5H5); 4.61 (s, NH); 5.0 (s, 1 H of h5-C5H3); 7–8 (PÀC6H4).
31P{1H}-NMR (CDCl3): 24.8 (s); 41.1 (s). 13C{1H}-NMR (CDCl3): 0–50 (PÀCy); 63.8 (d, J(P,C)=7.55, h5-
C5H3); 67.9 (d, J(P,C)=7.8, h5-C5H3); 68.45 (d, J(C,P)=39.6, h5-C5H3); 72.2 (s, h5-C5H5); 73.1 (s, h5-
C5H3); 73.65 (s, h5-C5H3); 120–145 (C6H4). Anal. calc. for C43H63Cl2FeNP2Pd: C 58.09, H 7.14, N 1.58;
found: C 57.7, H 9.13, N 1.76.
3.2. Synthesis of [Pd(1)(TsO)(H2O)]+TsOÀ. Prepared according to [2]. To a soln. of [PdCl2(1)] (100
mg, 0.12 mmol) in CH2Cl2 (15 ml) was added TsOAg (72 mg, 0.25 mmol) at r.t. The mixture was stirred
overnight, filtered through a short column of Celite, and concentrated under vacuum. Addition of Et
A
led to the precipitation of the crude title complex as an orange solid, which was washed with Et2O and
AHCTREUNG
dried. Yield: 35%. 1H-NMR (CDCl3): 1.47 (dd, J(P,H)=15, J(H,H)=7, MeCH); 2.03 (s, 6 H, Me2C6H3);
2.23 (s, 6 H, Me2C6H3); 2.31 (s, Me of Ts); d 3.97 (s, h5-C5H5); 4.12 (s, 1 H of h5-C5H3); 4.47 (s, 1 H of h5-
C5H3); 4.85 (s, 1 H of h5-C5H3); 5.03 (very br., MeCH); 6.5–8.5 (16 arom. H). 31P{1H}-NMR (CDCl3): 15.1
(d, J(P,P)=12); 50.3 (d, J(P,P)=12). 13C{1H}-NMR (CDCl3): 120–150 (24 arom. C); 91.6 (dd, J(P,C)=17,
4, h5-C5H3); 73.1 (s, h5-C5H3); 72.25 (m, h5-C5H3); 72.1 (s, h5-C5H5); 70.8 (dd, J(P,C)=8, 3, h5-C5H3); 70.35
(d, J(P,C)=9, h5-C5H3); 30.9 (d, J(P,C)=28, MeCH); 21.55 (s, Me of Ts); 21.4 (s, 2 Me2
Me2Ph).
A
AHCTREUNG
Results. – Synthesis. The PdII precursors [PdCl2(1)], [PdCl2(9)], and [PdCl2(10)]
were obtained by reaction between the bis(benzonitrile) complex [PdCl2(NCPh)2]
and the corresponding diphosphine ligand in toluene at room temperature. Similarly,
reaction of [PdCl2(1)] with TsOAg yielded the cationic complex [Pd(1)(H2O)-
A
N
by multinuclear NMR spectroscopy and elemental analysis (see Experimental). The
‘in situ’ catalytic systems for the methoxycarbonylation reaction were prepared
under N2 atmosphere in THF/MeOH 1:1 containing the catalyst precursor, the appro-
priate diphosphine ligand, and the acid.
Optimization of Reaction Conditions for the System [PdCl2(NCPh)2] and Ligand 1.
The catalytic conditions were first optimized for 0.15 mmol of [PdCl2(NCPh)2] and
TsOH as the source of acid. The reactions were carried out in THF/MeOH 1:1 over
24 h. In Table 1, the conversion, the regioselectivity (percentage of branched vs. linear
esters), the chemoselectivity (formation of acids instead of esters), and the enantio-
meric excess (ee) and configuration of the desired branched esters are collected.
When the reaction was performed with a 10-fold excess of acid to Pd precursor at a
CO pressure of 30 bar and at 508 (Entry 1 in Table 1), no reaction products could be
detected. Increasing the temperature to 908 (Entry 2) resulted in a conversion of
11%, with the expected linear and branched esters as the only reaction products. The
regioselectivity of the reaction was in favor of the linear ester (78%), and a moderate
ee value of 57% was obtained under these conditions. When a 100-fold excess of TsOH
was used at 1008 (Entry 3), the conversion increased to 59%, but the presence of 13% of
byproducts (acids) was detected, the regioselectivity being slightly lowered (16% of
branched product). When the temperature was increased to 1508, a 78% conversion
was achieved, with a chemoselectivity of 75%. The regioselectivity remained basically
constant (15% branched), and 86% ee was observed under these conditions. An
increase in CO pressure to 50 bar (Entry 5) resulted in a conversion of 94%, with the
regio- and chemoselectivities remaining unchanged (15 and 76%, resp). However, in
this case, the ee decreased to 70%. When CF3CO2H was used as the acid (Entry 6),
the conversion and enantioselectivity of the reaction decreased to 48 and 55%, respec-

1616
Helvetica Chimica Acta – Vol. 89 (2006)
Table 1. Methoxycarbonylation of Styrene with [PdCl2(NCPh)2] and Ligand 1 (see Scheme). Conditions:
0.15 mmol [PdCl2(NCPh)2] and 1.1 equiv. of 1 in THF/MeOH 1:1 (5 ml), 1 mol-% of Pd rel. to styrene,
24 h.
Entry T [8] PCO [bar] TsOH/Pd Conversion Regiosel. [%]a)
Chemosel. [%]b) ee [%]
(Config.)
B
L
1
2
3
4
5
6
50
90
100
150
150
150
30
30
30
30
50
30
10
10
100
100
100
100c)
0
11
59
78
94
48
–
–
–
100
87
75
76
–
22
16
15
15
21
78
84
85
85
79
57 (S)
–
86 (S)
70 (S)
55 (S)
100
a) Regioselectivity: branched (B) vs. linear (L) ester. b) Chemoselectivity: percentage of main ester prod-
c
ucts rel. to acid side products. ) CF3CO2H was used instead of TsOH.
tively. However, the regioselectivity was slightly higher (21% branched) than when
TsOH was used, and the chemoselectivity of the reaction was 100% (no side products).
In view of these results, the reaction conditions used in Entry 4, providing high ee
values, were considered to be convenient for this study, and were used for the experi-
ments described in the following sections. In order to compare the activities of different
Pd precursors, the reaction was repeated under the conditions described in Entry 4
using PdII and Pd0 precursors in the presence of the diphosphine 1.
Effect of Palladium Precursor. In Table 2, the conversion, and the regio-, chemo-,
and enantioselectivities for the methoxycarbonlyation of styrene in the presence of
the precursor catalysts [PdCl2(NCPh)2], [Pd(OAc)2], [Pd(acac)2], and [Pd2(dba)3] are
compared. In the case of [Pd2(dba)3], the reaction was also run in the absence of
TsOH to find out whether MeOH could also act as a source of H+ under these condi-
tions. In general, the conversions were between 72 and 83% after 24 h. The highest con-
versions and regioselectivities were obtained with the Pd0 precursor [Pd2(dba)3] (Entry
5 in Table 2), while the lowest was observed for [Pd(acac)2] (Entry 4). However, when
[Pd2(dba)3] was used in the absence of TsOH (Entry 6), no reaction products could be
detected.
Entry 3 in Table 2 shows the product distribution of the reaction after 45 instead of
24 h, using [Pd(OAc)2] as precursor. When compared with the results of Entry 2, very
small variations were observed. Indeed, the conversion was found to only increase by
5% (from 82 to 87%). Neither the regio- nor the enantioselectivity of the reaction was
significantly affected over the extra 21 h, but the chemoselectivity decreased from 46 to
40%.
The regioselectivity of the reaction was found to be in the range 13–20% with the
above Pd precursors. The highest regioselectivity (highest yield of branched ester) was
obtained with [Pd2(dba)3] (Entry 5). However, the chemoselectivity was found to vary
between 46 and 75%, depending on the precursor used. The dichloride complex (Entry
1) afforded the highest ratio of ester, while, in the case of [Pd2(dba)3] and Pd(acac)2
(Entries 4 and 5, resp.), less than 60% of the products were esters. In the case of
[Pd(OAc)2], the major product was the carboxylic acid rather than the ester.

Helvetica Chimica Acta – Vol. 89 (2006)
1617
Table 2. Variation of Catalyst Precursor in the Methoxycarbonylation of Styrene in the Presence of 1 and
TsOH. Conditions: 0.15 mmol [PdCl2(NCPh)2] and 1.1 equiv. of 1 in THF/MeOH 1:1 (5 ml), 1 mol-% of
Pd rel. to styrene, T=1508, PCO =30 bar, TsOH/Pd 100 :1, 24 h.
Entry
Precursor
Conversion [%]
Regiosel. [%]a)
Chemosel. [%]b)
ee [%]
(Config.)
B
L
1
2
3c)
4
5
6d)
[PdCl2(NCPh)2]
[Pd(OAc)2]
[Pd(OAc)2]
[Pd(acac)2]
[Pd2(dba)3]
[Pd2(dba)3]
78
82
87
72
83
0
15
14
16
13
20
–
85
86
84
87
80
–
75
46
40
58
56
–
86 (S)
52 (S)
55 (S)
55 (S)
59 (S)
–
a) Regioselectivity: branched (B) vs. linear (L) ester. b) Chemoselectivity: percentage of main ester prod-
c
ucts rel. to acid side products. ) Reaction time 45 h. d) Without TsOH.
In terms of enantioselectivity, the highest ee value (86%) was obtained with
[PdCl2(NCPh)2] (Entry 1). When the other precursors were used, the ee reached
50–60%. In view of these results, [PdCl2(NCPh)2] seems to be the catalyst precursor
of choice for the methoxycarbonylation of styrene under these conditions.
Effect of Phosphine Ligand. Next, we compared the effect of the different ligands
1–10 on the methoxycarbonylation of styrene with the catalytic system [PdCl2(NCPh)2]
and ligands 1–10 (Table 3). The conversion was 58–79%, depending on the ligand. The
highest value was achieved with 4 (79% conversion), although very similar values were
obtained for the diphosphines 1, 8, and 9. The percentage of branched ester formed dur-
ing the reaction was found to vary between 15 and 31%, the most-selective system
favoring the ester comprising ligand 6. The formation of 2-phenylpropanoic acid as a
side product varied between 17 and 41%, the lowest amount being produced with 1
and 9 (75%), when 1.1 equiv. of the ligand was used relative to Pd. With ligands 5
and 6, less than 60% of the products were esters. However, when the ratio 6/Pd was
changed from 1:1.1 to 1:2.2, a better chemoselectivity was achieved (83 instead of
59%; Entries 6 and 7). The ee values obtained ranged from 44% (for 6) to 86% (for 1).
Within the series of the Josiphos ligands 1–5, significant variations in the catalytic
results were observed (Entries 1 to 5 in Table 3). The system containing the diphosphine
1 (Entry 1) gave the highest conversion (78%), chemoselectivity (75%), and ee (86%).
The generally modest regioselectivity (15%) obtained during the reaction was very sim-
ilar to that obtained with the other ligands of this type (15–19% of branched product).
The Mandyphos-type ligand 6 yielded a relatively low conversion (58%; Entry 6),
but the use of a higher ligand-to-metal ratio increased the conversion to 73% (Entry
7). In terms of regioselectivity, the reactions carried out in the presence of 1.1 equiv.
of 6 (Entry 6) yielded 31% of branched product. The use of 2.2 equiv. of 6 lowered
the regioselectivity of the reaction (23% branched). The ee values obtained with this
ligand were moderate (44 and 47%).
When the ligands of the Walphos type were used (7 and 8), conversions of 78 and
75% were obtained, with 62 and 75% chemoselectivity, respectively, and with 16%
of branched product. A high ee value (73%) was also obtained with ligand 8. In con-

1618
Helvetica Chimica Acta – Vol. 89 (2006)
Table 3. Variation of Ligand in the Methoxycarbonylation of Styrene in the Presence of [PdCl2(NCPh)2]
and TsOH. Conditions: 0.15 mmol [PdCl2(NCPh)2] and 1.1 equiv. of ligand in THF/MeOH 1:1 (5 ml),
1 mol-% of Pd rel. to styrene, T=1508, PCO =30 bar, TsOH/Pd 100 :1, 24 h.
Entry
Ligand
Conversion [%]
Regiosel. [%]a)
Chemosel. [%]b)
ee [%]
(Config.)
B
L
1
2
3
4
5
6
7c)
8
9
10
11
1
2
3
4
5
6
6
7
8
78
61
70
79
72
58
73
78
75
65
64
15
17
15
17
19
31
23
16
16.5
28
25
85
83
85
83
81
69
77
84
83.5
72
75
75
74
67
72
57
59
83
62
75
72
70
86 (S)
62 (S)
51 (S)
77 (S)
45 (S)
47 (S)
44 (S)
52 (S)
73 (S)
63 (S)
62 (S)
9
10
a) Regioselectivity: branched (B) vs. linear (L) ester. b) Chemoselectivity: percentage of main ester prod-
c
ucts rel. to acid side products. ) With 2.2 (instead of 1.1) mol-equiv. of 6 rel. to Pd.
trast, conversions obtained in the presence of the Taniaphos compounds 9 and 10 were
moderate (65 and 62%, resp.).
Comparison with Isolated Complexes. In Table 4, a comparison is made for the
methoxycarbonylation of styrene in the presence of isolated neutral and cationic Pd
precursors. Here, the conversions were 70–93%. The highest value (93% conversion)
was obtained with [PdCl2(9)]. The highest regioselectivity was achieved with the Tania-
phos-type ligand 10 (23% branched, Entry 3 in Table 4). The chemoselectivity of the
reaction was found to vary between 53 (Entry 1) and 91% (Entry 2). In terms of enan-
tioselectivity, the highest ee value (70%) was obtained with [PdCl2(1)] (Entry 1). The
lowest ee (52%) was found for the system containing ligand 8.
Table 4. Methoxycarbonylation of Styrene Using Isolated (neutral and cationic) Complexes. Conditions:
0.15 mmol [PdCl2(NCPh)2] and 1.1 equiv. of ligand (1, 9, or 10) in THF/MeOH 1:1 (5 ml), 1 mol-% of Pd
rel. to styrene, T=1508, PCO =30 bar, TsOH/Pd 100 :1, 24 h.
Entry Complex
Conversion [%] Regiosel. [%]a) Chemosel. [%]b) ee (%)
(Config.)
B
L
1
2
3
4
[PdCl2(1)]
[PdCl2(9)]
[PdCl2(10)]
72
93
86
15
20
23
13
85
80
77
87
53
91
71
67
70 (S)
64 (S)
59 (S)
62 (S)
[Pd(1)(TsO)(H2O)]+(TsO)À 86
a) Regioselectivity: branched (B) vs. linear (L) ester. b) Chemoselectivity: percentage of main ester prod-
ucts rel. to acid side products.

Helvetica Chimica Acta – Vol. 89 (2006)
1619
Discussion. – Reaction Conditions. Some 15 mmol of [PdCl2(NCPh)2] and 1.1 equiv.
of ligand 1 were reacted in THF/MeOH 1:1 (for solubility reasons) over 24 h to opti-
mize the reaction conditions (Table 1). With TsOH as the acid source, a 100-fold excess
of acid was necessary to achieve conversions >50%, together with a temperature of
1008 and a CO pressure of 30 atm. Increasing the temperature to 1508 led to an increase
in conversion to 78%, with an ee of 86%. The necessity of elevated temperature and a
large excess of acid could be attributed to the high stability of the [PdCl2(PÀP’)] com-
plexes formed in situ, which are thought to lead to the production of an active species in
the presence of acid [28]. However, other reasons cannot be discarded.
The regioselectivity of the reaction was constant (ca. 15% of branched ester) when
the temperature was above 908. At lower temperatures, a higher proportion of desired
branched product was obtained, but the conversion was poor. Hence, the formation of
the branched ester is favored at lower temperature, reaching a constant value above
908. Under otherwise identical conditions, the use of CF3COOH as the acid source
gave rise to a conversion of 48%, with an ee of 55%, and the regioselectivity increased
to 21% in favor of the branched product. The lower conversion observed could be
rationalized by the weaker acidity of the medium. However, the variations in regio-
and enantioselectivity indicate that the counter-ion may also influence the selectivity
of the reaction. Such an effect had previously been observed [20].
Finally, when the CO pressure was increased to 50 atm, the conversion was found to
be >90%; however, the ee was only 70%. This decrease in chiral induction at higher
CO pressure (Entry 4 vs. 5 in Table 1) had been reported before in an alkoxycarbony-
lation [29].
Effect of Catalyst Precursor. The results obtained with [Pd(OAc)2], [Pd(acac)2], and
[Pd2(dba)3] as catalyst precursors were similar to those previously obtained with
[PdCl2(NCPh)2] in terms of conversion (ca. 80%) and regioselectivity (ca. 15%;
Table 2). However, only [PdCl2(NCPh)2] afforded a high enantioselectivity (86%);
other precursors yielded ee values <60%. The production of side products (acids)
was also minimized with the PdII bis(benzonitrile) complex.
To probe the potential formation of active species from a Pd0 precursor and MeOH,
the reaction was also run with [Pd2(dba)3] in the absence of TsOH (Entry 6 in Table 2).
However, no products were detected after 24 h. Hence, we conclude that a stronger
acid is required for the formation of active species from [Pd2(dba)3] under these condi-
tions.
The catalytic conditions used by the groups of Inoue and Wang, using ferrocenyl
1,1’-diphosphine ligands, were fairly different from those found to be best in our system.
Wang and co-workers [26b] used [PdCl2]/TsOH at 508 and 170 atm of CO. Inoue and co-
workers [26a] achieved 86% ee under formation of 44% of branched product using the
system [Pd(OAc)2]/TsOH at room temperature and 20 atm CO. This latter result was,
however, obtained with a lower substrate/Pd ratio, and the conversion to the esters was
only 17% after 20 h.
Effect of Ligand. The highest conversions after 24 h were obtained for the systems
containing the Josiphos ligands 1 and 4, as well as the Walphos diphosphines 7 and 8
(Table 3), with values >75%. Ligands 1 and 4 both contain a Ph2P unit reducing
A
their electron-donating properties. When compared to other ligands from the Josiphos
family, compounds 3 and 5 gave rise to a conversion of ca. 70%, while the more-basic

1620
Helvetica Chimica Acta – Vol. 89 (2006)
ligand 2, containing two (Cy)2
containing either two unsubstituted Ph2
P unit (7), both effected high conversions, suggesting a small influence
ACHTREUNG
A
ACHTREUNG
(CF3)2ÀC6H3]2
ACHTREUNG
of the electron-withdrawing CF3 groups. Like the Walphos ligands, the diphosphines
from the Taniaphos family (9 and 10) also contain an aromatic spacer between a Cp
ring and one P-atom, further separated by an alkyl group substituted with a Me
unit. These two ligands both gave rise to conversions of ca. 65%, without a significant
effect of the R2P substituents (R=Ph or Cy). When 1.1 equiv. of 6 was used, only 58%
A
AHCTREUNG
of styrene was consumed after 24 h.
Under the conditions used in this study, the 1,2-bis(phosphino)ferrocene ligands
1–5 and 7–10 gave rise to higher conversions than the 1,1’-bis(phosphino)ferrocene
6. However, as only one ligand of the latter type was probed in this study, no clear con-
clusion can be drawn.
The use of ligand 6 yielded the highest regioselectivity in terms of branched ester
(31%). Interestingly, 2.2 equiv. of this ligand resulted in an increase of the conversion
and chemoselectivity, but also to a decrease in regioselectivity (23%). The formation of
complexes containing more than one diphosphine ligand being sterically more hindered
and, thus, favoring the formation of the linear ester, might rationalize the observed var-
iations in the final product distribution. The systems containing ligands from the Tania-
phos family (Entries 10 and 11 in Table 3) also afforded regioselectivities of >25%
(branched ester), independently of the nature of the substituents R of the R2P units.
A
The systems containing the Josiphos and Walphos ligands, 1–5 and 7–8, respectively,
afforded the lowest regioselectivities, although some of them yielded the highest ee val-
ues (Entries 1, 4, and 9). In an opposite manner, the most-regioselective system, gener-
ated with 6, afforded the lowest ee value (47%) across the series.
Within each family of ligands, the variation in the nature of the R
substantial variations in the final ee values. In the Josiphos family, the ee changed
from 45 to 86%. The system containing ligand 1, which contains two (Ar)2P units,
A
AHCTREUNG
yielded the highest ee value. However, the replacement of these groups by more-elec-
tron-donating substituents such as Cy, t-Bu, or Et led to a decrease in ee. The effect of
the steric hindrance induced by these substituents on the chiral input involved in the
catalytic process was not clearly indicated here. The Walphos ligands 7 and 8, which dif-
fer by the presence or absence of two meta-positioned CF3 substituents on the two
phenyl rings attached to one of the two P-atoms, respectively, gave rise to ee values
of 52 and 73%, respectively. The presence of these electron-withdrawing groups,
thus, resulted in a decrease of enantioselectivity by ca. 20%.
When the isolated neutral complexes [PdCl2(1)], [PdCl2(9)], and [PdCl2(10)] were
used as catalyst precursors in the asymmetric methoxycarbonylation of styrene, the
results differed considerably from those obtained by the in situ method (Table 4). Espe-
cially, higher conversions were obtained with the isolated compounds. The regioselec-
tivities, however, were hardly affected. The ee values obtained were found to be equal
or slightly inferior to those obtained with the in situ method. This could be due to the
presence of a small excess of ligand in case of the in situ method.
Finally, the use of the cationic complex [Pd(1)(H2O)(TsO)]+TsOÀ, when compared
with the neutral complex [PdCl2(1)], yielded a higher conversion and chemoselectivity,

Helvetica Chimica Acta – Vol. 89 (2006)
1621
but a slightly reduced ee value, the regioselectivity of the reaction remaining almost
unchanged (ca. 15% of branched product).
Conclusions. – We have reported a systematic study on the Pd-catalyzed asymmet-
ric methoxycarbonylation of styrene using chiral 1,2-bis(phosphino)ferrocene ligands.
The precursors used for the catalytic reactions, either isolated [PdIICl2]/(PÀP’)] com-
plexes or ‘in situ’ [PdCl2(NCPh)2]/PÀP’ systems, were found to be efficient catalysts.
Although the regioselectivity of the reaction was largely in favor of the (undesired) lin-
ear ester, high conversion and high enantiomeric excess (up to 86% with the Josiphos
ligand 1) were achieved. Investigations into the mechanism involved in this reaction by
means of high-pressure NMR techniques are currently in progress and will be reported
in due course.
We are grateful to the European Community (PALLADIUM network, contract no. HPRN-CT-2002-
00196) for financial support, and to Solvias for providing the ferrocenyl diphosphine ligands used in this
study. We also thank Mickael Lopes de Graça for his contribution to this work, and Prof. Pedro Perez and
Prof. Piet Van Leeuwen for useful discussions.
REFERENCES
[1] P. Barbaro, C. Bianchini, G. Giambastiani, S. L. Parisel, Coord. Chem. Rev. 2004, 248, 2131; T. J.
Colacot, Platinum Met. Rev. 2001, 45, 22; K. S. Gan, T. S. A. Hor, in ‘Ferrocenes: Homogeneous
Catalysis, Organic Synthesis, Materials Science’, Eds. A. Togni, T. Hayashi, Wiley-VCH, Weinheim,
1995, p. 13.
[2] C. Bianchini, A. Meli, W. Oberhauser, S. Parisel, O. V. Gusev, A. Kal’sin, N. V. Vologdin, F. M. Dol-
gushin, J. Mol. Catal., A 2004, 224, 35; C. Bianchini, W. Oberhauser, A. Orlandini, C. Giannelli, P.
Frediani, Organometallics 2005, 15, 3692.
[3] M. Ogasawara, K. Yoshida, T. Hayashi, Organometallics 2000, 19, 1567; B. E. Bosch, I. Brümmer, K.
Kunz, G. Erker, R. Frölich, S. Kotila, Organometallics 2000, 19, 1255; Y. Xie, G. K. Tan, Y. K. Yan, J.
J. Vittal, S. C. Ng, T. S. A. Hor, J. Chem. Soc., Dalton Trans. 1999, 773; S. W. A. Fong, T. S. A. Hor, J.
Cluster Sci. 1998, 9, 351; T. Hayashi, M. Konishi, Y. Kobori, M. Kumada, T. Higushi, K. Hirotsu, J.
Am. Chem. Soc. 1984, 106, 158.
[4] A. Jutand, K.-K. Mimi Hii, M. Thornton-Pett, J. M. Brown, Organometallics 1999, 18, 5367; A. L.
Boyes, I. R. Butler, S. C. Quayle, Tetrahedron Lett. 1998, 39, 7763; J. M. Brown, K.-K. Mimi Hii,
Angew. Chem., Int. Ed. 1996, 35, 657.
[5] W. Mägerlein, A. F. Indolese, M. Beller, Angew. Chem., Int. Ed. 2001, 40, 2856.
[6] J. F. Hartwig, Angew. Chem., Int. Ed. 1998, 37, 2090; J. F. Hartwig, Acc. Chem. Res. 1998, 31, 852; M.
S. Driver, J. F. Hartwig, J. Am. Chem. Soc. 1996, 118, 7217.
[7] K. Li, P. N. Horton, M. B. Hurthouse, K.-K. Mimi Hii, J. Organomet. Chem. 2003, 665, 250.
[8] C. Bianchini, A. Meli, W. Oberhauser, P. W. N. M. van Leeuwen, M. A. Zuidevelt, Z. Freixa, P. C. J.
Kamer, A. L. Spek, O. V. Gusev, A. M. Kal’sin, Organometallics 2003, 22, 2049; O. V. Gusev, A. M.
Kal’sin, M. G. Peterleitner, P. V. Petroskii, K. A. Lyssenko, N. G. Akhmedov, C. Bianchini, A. Meli,
W. Oberhauser, Organometallics 2002, 21, 3637.
[9] D. Xiao, X. Zhang, Angew. Chem., Int. Ed. 2001, 40, 3425.
[10] T. Hayashi, in ‘Ferrocenes: Homogeneous Catalysis, Organic Synthesis, Materials Science’, Eds. A.
Togni, T. Hayashi, Wiley-VCH, Weinheim, 1995, p. 105; H.-U. Blaser, W. Brieden, B. Pugin, F. Spin-
dler, M. Studer, A. Togni, Top. Catal. 2002, 19, 3.
[11] a) M. Lotz, K. Polborn, P. Knochel, Angew. Chem., Int. Ed. 2002, 41, 4708; b) T. Ireland, K. Tappe, G.
Grossheimann, P. Knochel, Chem.–Eur. J. 2002, 8, 843; c) R. Kuwano, M. Sawamura, S. Okuda, T.
Asai, Y. Ito, M. Redon, A. Krief, Bull. Chem. Soc. Jpn. 1997, 70, 2807; d) H. U. Blaser, E. Schmidt,
‘Asymmetric Catalysis on Industrial Scale’, Wiley-VCH, Weinheim, 2004.

1622
Helvetica Chimica Acta – Vol. 89 (2006)
[12] H.-U. Blaser, F. Spindler, in ‘Comprehensive Asymmetric Catalysis’, Eds. E. N. Jacobsen, A. Pfaltz,
H. Yamamoto, Springer, Berlin, 1999, p. 247.
[13] F. Spindler, B. Pugin, H. P. Jalett, H. P. Buser, U. Pittelkow, H.-U. Blaser, in ‘Catalysis of Organic
Reactions’, Ed. R. E. Maltz, Marcel Dekker, New York, 1996, p. 153.
[14] R. Dorta, D. Broggini, R. Stoop, H. Rüegger, F. Spindler, A. Togni, Chem.–Eur. J. 2004, 10, 267.
[15] G. Argouarch, O. Samuel, H. B. Kagan, Eur. J. Org. Chem. 2000, 2885.
[16] T. Hayashi, M. Tajika, K. Tamao, M. Kumada, J. Am. Chem. Soc. 1976, 98, 3718; T. Hayashi, M.
Konishi, Fukushima, T. Mise, M. Kagotani, M. Tajika, M. Kumada, J. Am. Chem. Soc. 1982, 104, 180.
[17] S. Pastor, A. Togni, J. Am. Chem. Soc. 1989, 111, 2333.
[18] T. Sturn, L. Xiao, W. Weissensteiner, Chimia 2001, 55, 688.
[19] P. Herold, S. Stutz, to Speedel Pharma AG, WO 02/02500 A1, 2002.
[20] F. A. Rampf, W. A. Herrmann, J. Organomet. Chem. 2000, 601, 138.
[21] U. Nettekoven, P. C. J. Kamer, M. Whidalm, P. W. N. M. van Leeuwen, Organometallics 2000, 19,
4596.
[22] ‘Ullmans Encyclopedia of Industrial Chemistry’, Ed. W. Gerharzt, VCH, Weinheim, 1985, Vol. 3, p.
41.
[23] ‘Catalytic Asymmetric Synthesis’, 2nd edn., Ed. K. Nozaki, I. Ojima, J. Wiley & Sons, New York,
2000, p. 448.
[24] Y. Sugi, K. Bando, Chem. Lett. 1976, 727.
[25] I. Del Rio, C. Claver, P. W. N. M. van Leeuwen, Eur. J. Inorg. Chem. 2001, 2719.
[26] a) S. Oi, M. Nomura, T. Aiko, Y. Inoue, J. Mol. Catal. A 1997, 115, 289; b) H. Zhou, J. Hou, J. Cheng,
S. Lu, H. Fu, H. Wang, J. Organomet. Chem. 1997, 543, 227.
[27] M. S. Kharasch, R. C. Seyler, F. R. Mayo, J. Am. Chem. Soc. 1998, 60, 882; J. R. Doyle, P. E. Slade,
H. B. Jonassen, Inorg. Synth. 1960, 6, 216.
[28] G. R. Eastham, B. T. Heaton, J. A. Iggo, R. P. Tooze, R. Whyman, S. Zacchini, Chem. Commun.
2000, 609; P. J. Perez, J. C. Calabrese, E. E. Brunel, Organometallics 2001, 20, 337; V. V. Grushin,
Chem. Rev. 1996, 96, 2011.
[29] D. Zim, R. F. de Souza, J. Dupont, A. L. Monteiro, Tetrahedron Lett. 1998, 39, 7071.
Received February 9, 2006