Synthesis and use in asymmetric hydrogenations of solely planar chiral 1,2-disubstituted and 1,2,3-trisubstituted ferrocenyl diphosphines: A comparative study

Article abstract of DOI:10.1021/om700405a

A total of 12 enantiopure 1,2-disubstituted and 1,2,3-trisubstituted ferrocenyl diphosphines [1-Ph₂P-2-R₂PCH ₂-3-R′-Fc, R = Cy, t-Bu, 3,5-(CH₃)₂C ₆H₃), R′ = H, CH₃, Ph

Full text of DOI:10.1021/om700405a

3530
Organometallics 2007, 26, 3530-3540
Synthesis and Use in Asymmetric Hydrogenations of Solely Planar
Chiral 1,2-Disubstituted and 1,2,3-Trisubstituted Ferrocenyl
Diphosphines: A Comparative Study
Yaping Wang and Walter Weissensteiner*
Institute of Organic Chemistry, UniVersity of Vienna, Wa¨hringer Strasse 38, A-1090 Vienna, Austria
Felix Spindler
SolVias AG, Catalysis Research, P.O. Box, CH-4002 Basel, Switzerland
Vladimir B. Arion§
Institute of Inorganic Chemistry, UniVersity of Vienna, Wa¨hringer Strasse 42, A-1090 Vienna, Austria
Kurt Mereiter§
Institute of Chemical Technologies and Analytics, Vienna UniVersity of Technology,
Getreidemarkt 9/164SC, A-1060 Vienna, Austria
ReceiVed April 26, 2007
A total of 12 enantiopure 1,2-disubstituted and 1,2,3-trisubstituted ferrocenyl diphosphines [1-Ph2P-
2-R2PCH2-3-R′-Fc, R ) Cy, t-Bu, 3,5-(CH3)2C6H3), R′ ) H, CH3, Ph, 3,5-(CH3)2C6H3)] have been
synthesized, characterized, and tested in asymmetric rhodium- and ruthenium-catalyzed hydrogenations
of four alkenes and two ketones. The performance of these ferrocene derivatives in catalytic hydrogenations
was compared to those of catalysts based on Josiphos, PPF-t-Bu2, and Xyliphos [1-Ph2P-2-R2P(CH3)-
CH-Fc, R ) Cy, t-Bu, 3,5-(CH3)2C6H3)]. Dichloro palladium(II) complexes of Xyliphos and its analogues
lacking the stereogenic center were synthesized as model compounds, and their molecular structures
were studied in solution and, for four complexes, in the solid state. In hydrogenation reactions the
replacement of Xyliphos as the catalyst ligand with its analogues lacking the stereogenic center leads to
changes in product configurations or significant increases or decreases in ee values.
extended backbones (for example, Taniaphos, 5,7 or Walphos-
type ligands, 6).8 Even small changes in ligand structure have
a marked influence on catalytic performance, and as a result,
tuning ligands with respect to their electronic and steric
properties is a necessary task in catalyst optimization. While
electronic properties are usually changed by attaching electron-
withdrawing or electron-donating substituents to either the ligand
backbone or, more frequently, the phosphine substituents,
changes in steric properties can be achieved by a number of
different methods. These include changes in the size of
phosphine substituents, addition of buttressing substituents in
close proximity to the phosphine units, or limiting or enhancing
the conformational flexibility of the ligand backbone.9-12
Recently, in our search for novel ligand backbones, we
questioned whether restricting the conformational flexibility of
Josiphos-type ligands by transforming the ethyl side chain into
either a homo- or heteroannular bridge (e.g., ligands 7 and 8)
Introduction
Since the pioneering work of Kumada and Hayashi in 1974,1
chiral ferrocenyl-based diphosphines have seen enormous
development and now constitute a class of catalyst ligands that
give outstanding performance in a wide variety of enantiose-
lective reactions.2,3 Among these systems, Josiphos-type
diphoshines are certainly the most prominent examples, mainly
since three representatives [Josiphos, 1; PPF-t-Bu2, 2; and
Xyliphos, 3 (Chart 1)] are used in industrial hydrogenation
processes.4,5 In recent years, these successful applications of
Josiphos-type ligands have boosted the further development of
ferrocenyl-based diphosphines, and this has led to a number of
novel ligands with backbones that are either closely related to
that of Josiphos (for example, diphosphine 4)6 or have more
* Corresponding author. E-mail: Walter.Weissensteiner@univie.ac.at.
§ Crystallography.
(1) Hayashi, T.; Yamamoto, K.; Kumada, M. Tetrahedron Lett. 1974,
4405.
(2) Jacobsen, E. N., Pfaltz, A., Yamamoto, H., Eds. ComprehensiVe
Asymmetric Catalysis, Vol I-III; Springer: Vienna, 1999.
(3) Ojima, I., Ed. Catalytic Asymmetric Synthesis; Wiley-VCH: Wein-
heim, 2000.
(4) Blaser, H. U., Schmidt, E., Eds. Asymmetric Catalysis on Industrial
Scale: Challenges, Approaches and Solutions; Wiley-VCH: Weinheim,
2004.
(5) Blaser, H.-U.; Brieden, W.; Pugin, B.; Spindler, F.; Studer, M.; Togni,
A. Top. Catal. 2002, 19, 3.
(6) Yasuike, S.; Kofink, C. C.; Kloetzing, R. J.; Gommermann, N.; Tappe,
K.; Gavryushin, A.; Knochel, P. Tetrahedron: Asymmetry 2005, 16, 3385.
(7) Ireland, T.; Grossheimann, G.; Wieser-Jeunesse, C.; Knochel, P.
Angew. Chem., Int. Ed. 1999, 38, 3212.
(8) Sturm, T.; Weissensteiner, W.; Spindler, F. AdV. Synth. Catal. 2003,
345, 160.
(9) Gomez Arraya´s, R.; Adrio, J.; Carretero, J. C. Angew. Chem., Int.
Ed. 2006, 45, 7674.
(10) Atkinson, R. C. J.; Gibson, V. C.; Long, N. J. Chem. Soc. ReV.
2004, 33, 313.
(11) Colacot, T. J. Chem. ReV. 2003, 103, 3101.
(12) Togni, A., Hayashi, T., Eds. Ferrocenes: Homogeneous Catalysis,
Organic Synthesis, Materials Science; VCH: Weinheim, 1995.
10.1021/om700405a CCC: $37.00 © 2007 American Chemical Society
Publication on Web 06/12/2007

Planar Chiral Ferrocenyl Diphosphines
Organometallics, Vol. 26, No. 14, 2007 3531
Chart 1
Chart 2
could influence catalytic performance.13 The results obtained
for a number of hydrogenation reactions supported the view
that catalysts expected to perform successfully with different
substrates, or even in a range of different reactions, require a
controlled flexibility rather than very stiff ligand backbones.
A few years ago, Kagan’s group prepared very close
analogues of Josiphos-type ligands with planar chirality as the
sole source of chirality (e.g., ligands 9a and 9b, Chart 2).14 In
hydrogenation reactions, high enantioselectivities were obtained
only for a very limited number of alkenes (e.g., for itaconic
acid or methyl itaconate), and in general the scope of these
diphosphines seems to be much more limited than that of the
parent Josiphos-type ligands.
In view of our previous results on homo- or heteroannular
bridged ligands 7 and 813 we considered how the lack of a
stereogenic center in ligands 9 would actually be reflected in
the catalyst performance. For 9 and Josiphos-type ligands one
might imagine that in the formation of transition metal
complexes the flexibility of the chelate ring or equilibria of
chelate ring conformers are strongly influenced by the presence
or absence of a stereogenic center and substituents thereon,
especially when such a stereogenic center is part of the chelate
ring itself. In addition, we wondered whether one could also
exert an influence on the chelate ring conformations or on
conformer equilibria by attaching additional Cp substituents in
close proximity to the chelate ring and how such structural
changes would be reflected in the catalyst performance.
Therefore, it seemed to be of interest to study the influence of
Cp substituents in position 3 on the conformational flexibility
and enantioselectivity and to ascertain whether this could provide
an additional parameter for ligand tuning.
We report here the synthesis of one 1,2-disubstituted (9c)
and nine 1,2,3-trisubstituted (10-12) ferrocenyl diphosphines
(Chart 2), all of which are analogues of the 1,2-disubstituted
diphosphines 9a and 9b. In addition, in order to study the
coordination behavior of such ligands, dichloro palladium(II)
complexes of Xyliphos (3), i.e., 9c, 10c, 11c, and 12c, were
prepared. Structural features of these new ligands and complexes
are described along with a number of catalysis results obtained
with ligands 9-12 in different asymmetric hydrogenation
reactions of alkenes and ketones. The data obtained are
compared to those of Josiphos-type ligands 1-3.
(13) Sturm, T.; Weissensteiner, W.; Spindler, F.; Mereiter, K.; Lo´pez-
Agenjo, A. M.; Manzano, B. R.; Jalo´n, F. A. Organometallics 2002, 21,
Results and Discussion
1766.
Synthesis of Diphosphines 9-12 and Dichloro Palladium-
(II) Complexes of Ligands 3 and 9c-12c. In most cases,
(14) Argouarch, G.; Samuel, O.; Kagan, H. B. Eur. J. Org. Chem. 2000,
2885.

3532 Organometallics, Vol. 26, No. 14, 2007
Wang et al.
Scheme 1
Scheme 2
enantiopure 1,2-disubstituted ferrocene derivatives are synthe-
sized enantioselectively from a monosubstituted prochiral
precursor or diastereoselectively from a chiral starting material
through an ortho-directed reaction in which chiral auxiliaries
are frequently used as the ortho-directing group.15,16 Originally,
diphosphines 9a and 9b and related derivatives were synthesized
disubstituted derivatives 19, 20, and 21 [R′ ) CH3, Ph, 3,5-
(CH3)2C6H3]. In another ortho-directed step, a diphenylphos-
phino unit was attached to the second ortho-position next to
the chiral auxiliary O-methylephedrine. Finally, reaction of each
of the trisubstituted ferrocenyl phosphino intermediates 22, 23,
and 24 with dialkyl and diaryl phosphines R2PH [R ) Cy, t-Bu,
3,5-(CH3)2C6H3] in acetic acid led to the replacement of the
chiral auxiliary and resulted in the final diphosphines 10a-
10c, 11a-11c, and 12a-12c [10, 11, 12: R′ ) CH3, Ph, 3,5-
(CH3)2C6H3; a, b, c: R ) Cy, t-Bu, 3,5-(CH3)2C6H3].
The 1,2-disubstituted intermediates 19, 20, and 21 were
synthesized by two different routes. The methyl group was
introduced in a three-step sequence (Scheme 3), while the aryl
groups were attached using only two steps (Scheme 4). Although
derivative 19 (R′ ) CH3) could be prepared easily in one step
by reacting 14 in pentane with t-BuLi followed by methyl iodide,
separation of 19 from unreacted 14 proved to be rather difficult.
For this reason, a more convenient three-step sequence was
chosen (Scheme 3). In the first step, (1R,2S)-14 was lithiated
as described above and treatment with dimethylformamide as
the electrophile gave aldehyde (1R,2S,Rp)-16 in 89% yield.
Reduction with NaBH4 in ethanol gave alcohol (1R,2S,Rp)-17
(98%), which was further reduced with triethylsilane under
acidic conditions18 to give intermediate (1R,2S,Sp)-19 (79%).
Both of the aryl-substituted intermediates 20 and 21 were
synthesized from 14 in two steps (Scheme 4). As reported
by Kagan and co-workers by applying the latter approach.14
A
chiral acetal, 13, was used as the starting material (Scheme 1,
top). Subsequently, in order to shorten the synthesis, we replaced
13 by O-methylephedrine derivative 14, and this approach has
also been used here for the preparation of the novel Xyliphos
analogue 9c.17
We started the synthesis of diphosphine 9c from enantiopure
(1R,2S,Rp)-N-[(2-diphenylphosphino)ferrocenylmethyl]-O-meth-
ylephedrine (15), which we previously obtained from (1R,2S)-
N-ferrocenylmethyl-O-methylephedrine (14)17 and chlorodiphen-
ylphosphine in 81% yield (Scheme 1). In the subsequent step,
reaction of 15 with bis(3,5-dimethylphenyl)phosphine in acetic
acid resulted in the final diphosphine (Rp)-9c (72%).
All of the trisubstituted ligands (10-12) were prepared
according to the same general synthesis route, which is depicted
in Scheme 2. In the first reaction sequence, which started from
ortho-directing O-methylephedrine derivative 14, substituents
R′ were introduced in two or three steps to give the 1,2-
(15) Clayden, J. Top. Organomet. Chem. 2003, 5, 251.
(16) Clayden, J. In Chemistry of Organolithium Compounds; Rappoport,
Z., Marek, I., Eds.; Wiley: Chichester, 2004; pp 495-646.
(17) Xiao, L.; Kitzler, R.; Weissensteiner, W. J. Org. Chem. 2001, 66,
8912.
(18) Olah, G. A.; Narang, S. C.; Gupta, B. G. B.; Malhotra, R. Synthesis
1979, 61.

Planar Chiral Ferrocenyl Diphosphines
Organometallics, Vol. 26, No. 14, 2007 3533
Scheme 3
Scheme 4
previously, iodo derivative (1R,2S,Rp)-18 is accessible from
(1R,2S)-14 by ortho-lithiation and quenching the reaction
mixture with iodine.17 In the second step, the aryl substituents
[R′ ) Ph or 3,5-(CH3)2C6H3] were attached to the ferrocene
unit in a Suzuki-Miyaura reaction between iodide 18 and either
phenyl- or 3,5-dimethylphenylboronic acid.19 These reactions
were carried out in a mixture of toluene and ethanol as the
organic solvents, aqueous Na2CO3 as the base, and Pd(PPh3)4
as the catalyst precursor. The use of these conditions gave
products (1R,2S,Sp)-20 and (1R,2S,Sp)-21 in 50 and 80% yield,
respectively.
Each of the 1,2-disubstituted derivatives (1R,2S,Sp)-19,
(1R,2S,Sp)-20, and (1R,2S,Sp)-21 was further reacted in diethyl
ether with s-BuLi, and the reaction mixture was quenched with
chlorodiphenylphosphine to give the 1,2,3-trisubstituted inter-
mediates (1R,2S,Sp)-22 (74%), (1R,2S,Sp)-23 (69%), and
(1R,2S,Sp)-24 (84%) (Scheme 2). Finally, reaction of each of
these trisubstituted phosphines with dicyclohexylphosphine, di-
tert-butylphosphine, and bis(3,5-dimethyphenyl)phosphine under
acidic conditions resulted in the nine final products 10a-10c,
11a-11c, and 12a-12c, all of which had the Sp absolute
configuration.
shifted to lower field as compared to those of the unsubstituted
reference ligands 9 (CH2PR2, R ) Cy, 12a: 7.3 versus -1.0
ppm for 9a; R ) t-Bu, 12b: 36.2 versus 27.0 ppm for 9b; and
R ) Xyl, 12c: -7.4 versus -14.3 ppm for 9c). It is worth
noting thatsagain compared to the unsubstituted diphosphines
9a-9csa much more marked low-field shift is caused by the
methyl group linked to the stereogenic center of Josiphos-type
diphosphines 1-3 than by the Cp substituents of ligands 10-
12 [CH(CH3)PR2, R ) Cy, 1: 15.7 ppm; R ) t-Bu, 2: 49.9
ppm; and R ) Xyl, 3: 6.5 ppm].
The assignment of the Rp absolute configuration to 9 and 15-
18, as well as the Sp configuration to diphosphines 10-12 and
intermediates 19-24, is based on the observation that depro-
tonation of O-methylephedrine derivative (1R,2S)-14 occurs at
the ortho-Cp-carbon located counterclockwise to the ephedrine
unit (if ferrocene is viewed along an axis defined by the centroid
of the substituted Cp ring and the Fe atom; Schemes 1, 3, and
4). This was confirmed previously by chemical correlation17
and subsequently by X-ray diffraction on complexes [PdCl2-
(L)], L ) (Rp)-9c, (Sp)-10c, and (Sp)-11c (see below). For all
monophosphines (15, 22-24) and diphosphines (1-3, 9-12)
a negative sign for the specific rotation (measured in CHCl3 at
589 nm) correlates with Sp and a positive sign with the Rp
absolute configuration. On complexation of 3, 9c, and 10c-
12c to the respective dichloro palladium(II) complexes, the
specific rotation always changed sign as compared to that of
the corresponding free ligand.
In order to study the coordination behavior of ligands 3 and
9c-12c, their dichloro palladium(II) complexes [PdCl2(L)], L
) (R,Sp)-3, (Rp)-9c, (Sp)-10c-(Sp)-12c, were prepared by
reacting the appropriate diphosphines with dichlorobis(aceto-
nitrile)palladium(II).
Enantioselective Hydrogenations. Diphosphine ligands (Rp)-
9a-(Rp)-9c and (Sp)-10a-c-(Sp)-12a-c were tested in catalytic
hydrogenations of olefins and ketones (for the catalytic results
see Table 1; for substrates tested see Chart 3). For the sake of
comparison, all reactions were also run with Josiphos, (R,Sp)-
1, PPF-t-Bu2, (R,Sp)-2, and Xyliphos (R,Sp)-3. In performing
the hydrogenation reactions we primarily focused on two
aspects: (i) on the overall performance and enantioselectivity
of the solely planar chiral ligands 9-12 as compared to the
Josiphos-type ligands 1-3 bearing an additional stereogenic
center in the side chain and (ii) on the enantioselectivity of
complexes of trisubstituted ligands 10-12 as compared to the
disubstituted ligands 9. All catalyst precursors were formed in
The structural integrity of all compounds was assessed by
NMR spectroscopy, and in the cases of [PdCl2(L)], L ) (R,Sp)-
3, (Rp)-9c, (Sp)-10c, and (Sp)-11c, single crystals were also
studied by X-ray diffraction (see below). The 31P NMR spectra
of diphosphines 9-12 show some interesting features. The
signals of the PPh2 phosphorus directly attached to the Cp ring
all lie within a very narrow range of chemical shifts (δ ) -22.2
to -23.7 ppm) and are essentially unaffected by the nature of
the Cp substituents in position 3. However, the CH2PR2
phosphorus signals are rather sensitive to 3-substitution and are
(19) Jensen, J. F.; Søtofte, I.; Sørensen, H. O.; Johannsen, M. J. Org.
Chem. 2003, 68, 1258.

3534 Organometallics, Vol. 26, No. 14, 2007
Wang et al.
Table 1. Rhodium- and Ruthenium-Catalyzed Hydrogenation of Alkenes and Ketones with Ligands (R,Sp)-3, (Rp)-9, and
(Sp)-10-12
substrate
ligand
MAC
MCA
DMI
ee
EAC
MPG
EOP
entry
conv ee conf conv ee conf conv
conf conv ee conf conv ee conf conv ee conf
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Josiphos
1
100 94
R
R
R
R
R
R
R
S
S
S
R
S
S
26 26
21 9
R
R
R
R
R
R
R
R
R
R
R
R
100 99.5
100 98
100 99
100 95
100 97
100 10
55 18
49 42
43 25
50 37
96 92
82 21
69 16
54 27
68 24
S
S
S
S
S
R
R
R
R
R
S
S
S
S
S
70
64
20
84
85
48
23
60
20
40
55
42
43
35
34
61
51
36
41
43
43
31
44
26
17
19
12
27
31
26
S
S
S
S
S
S
S
S
S
S
R
S
S
S
S
47
94
32
80
90
40
39
38
32
37
13
18
15
18
18
23
15
15
12
13
57
36
38
30
33
3
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
17
14
10
1
1
25
5
7
5
1
10
10
<5
17
14
53
26
50
nd
nd
29
36
45
25
nd
41
nd
nd
nd
26
R
R
R
RCp-3 ) H 9aa
100
100
100
100
25
9
6
5
4
2
CH3
Ph
Xyl
10a
11a
12a
2
13 20
100 21
100 19
50 10
57 25
46 24
50 25
100 15
PPF-t-Bu2
R
R
R
R
RCp-3 ) H 9ba
51 16
CH3
Ph
Xyl
10b
11b
12b
3
53
51
22
6
5
4
Xyliphos
100 14
100 49
100 67
100 83
100 84
7
6
4
6
6
8
7
nd
nd
nd
R
RCp-3 ) H 9ca
12
9
6
CH3
Ph
Xyl
10c
11c
12c
S
S
6
S
a For the sake of comparison the absolute configuration of the product is given for ligands with the Sp configuration. For details on reaction conditions
and methods of ee determination see the Experimental Section.
Chart 3. Olefins and Ketones Used as Substrates in Catalytic Hydrogenations.
situ using an appropriate Rh or Ru source, and for a given
substrate all reactions were carried out under identical reaction
conditions (see Experimental Section).
respective catalyst-substrate combination. In the majority of
reactions tested, changing the reference ligands 1-3 to their
analogues led to only small to moderate changes in enantiose-
lectivity. For example, replacement of catalyst ligand 1 by 9a-
12a in the hydrogenation of DMI resulted in neither a significant
change in enantioselectivity nor conversion (1: 99.5% ee, 11a:
95% ee; Table 1, entries 1-5). However, in comparison to 1
the reaction of MAC with catalysts modified by the ligands 9a-
12a led to a sharp drop in enantioselectivity [from 94% (1) to
<10% ee; Table 1, entries 1-5], while conversion still remained
quantitative. A similar, though less pronounced, drop in enan-
tioselectivity was observed in the hydrogenation of DMI in
combination with ligands 3 and analogues 9c-12c [from 92%
(3) to <28% ee; Table 1, entries 11-15].
Interestingly, the opposite trend was observed in the reaction
of MAC with Xyliphos-type ligands. The hydrogenation of
MAC with a Xyliphos (3)-modified rhodium catalyst gave the
corresponding amino acid derivative with only 14% ee, but the
use of ligands 9c-12c led to a steady increase in the enanti-
oselectivity to 84% (9c: 49%, 10c: 67%, 11c: 83%, and 12c:
84%; Table 1, entries 11-15). In this case the change from
Xyliphos to its analogous ligand without a stereogenic center,
9c, led not only to higher ee values but also to the product with
the opposite absolute configuration. The use of (R,Sp)-3 gave
the product of R configuration, while use of ligand (Sp)-9c
resulted in the product of S configuration. A similar trend is
seen in the hydrogenation of the substrate ethyl (Z)-3-(aceta-
mido)crotonate (EAC). The replacement of ligand (R,Sp)-3 by
(Sp)-9c again resulted in a change in the product configuration
from R to S (Table 1, entries 11 and 12).
It is well known for the reference ligands 1-3 that, under
given reaction conditions, only certain catalyst-substrate com-
binations result in both high conversion and high enantioselec-
tivity. For the substrates tested in our study this is true only for
methyl R-(acetamido)cinnamate (MAC) and dimethyl itaconate
(DMI) as the substrates and Josiphos (1) as the catalyst ligand
(Table 1, entry 1) as well as for the combination of DMI and
Xyliphos (3, Table 1, entry 11). In all other cases either the
enantioselectivity or both the conversion and enantioselectivity
are low to moderate. In view of these results it was of interest
to see how structural changes in 1-3, for example the lack of
a stereogenic center (ligands 9a, 9b, and 9c) or the lack of
stereogenic center in conjunction with additional Cp substituents
(10-12), would influence the catalytic performance. As one
might expect, the influence of structural changes on conversion
and enantioselectivity is not uniform. For example, the conver-
sions remained quantitative in the hydrogenation of substrates
MAC and DMI on changing from ligand 1 to ligands 9a-12a
(Table 1, entries 1-5), whereas with DMI the change from
either ligand 2 or 3 to ligands 9b-12b and 9c-12c, respectively,
resulted in a slight decrease in conversion (Table 1, entries 6-10
and 11-15). On the other hand, in reactions that were rather
slow with reference ligands 1-3 the change from 1, 2, or 3 to
the analogous ligands without a stereogenic center could, in a
few cases, give increased conversion. An example of the latter
phenomenon is the reaction of MCA with 1 as compared to
11a and 12a (Table 1, entries 1, 4, and 5).
As with the conversion, the influence of structural changes
in the ligands on enantioselectivity strongly depends on the
In summary, comparison of the performance of reference
ligands 1, 2, and 3 with that of their solely planar chiral

Planar Chiral Ferrocenyl Diphosphines
Organometallics, Vol. 26, No. 14, 2007 3535
Table 2. Selected Geometrical Parameters of Complexes
[PdCl2(L)] (L ) (R,Sp)-3, (Rp)-9c, (Sp)-10c, and (Sp)-11c)
analogues 9-12 shows that for each of the reference ligands
the removal of the stereogenic center is reflected differently.
For example, in all hydrogenations tested it was found that
catalysts with the reference ligand Josiphos (1) gave higher enan-
tioselectivity than those with diphosphines 9a-12a, but, apart
from the substrate MAC, changes in ee values remained small
(Table 1, entries 1-5). As is the case with Josiphos, replacement
of the reference ligand PPF-t-Bu2 (2) with its analogues 9b-
12b resulted in only small changes in the overall catalyst
performance, but in this case, except for the substrate methyl
phenylglyoxylate (MPG), slight increases in enantioselectivity
were observed (Table 1, entries 6-10). Only when Xyliphos
(3) was replaced by its analogues 9c-12c were more significant
changes found for several substrates. As already mentioned, in
the case of DMI this led to a sharp drop in ee values, while for
EAC a small decrease and for MAC a very substantial increase
in enantioselectivity was found. In addition, in the hydrogenation
of substrates MAC, EAC, and ethyl 3-oxo-pentanoate (EOP)
the replacement of reference ligand 3 by 9c-12c caused a
change in the absolute configuration of the products.
Comparison of ligands 9a-c with their 3-substituted ana-
logues 10a-c-12a-c shows that the influence of Cp substit-
uents adjacent to the phosphinomethyl unit is rather small. Only
in the case of substrate MAC and ligands 9c-12c is a significant
increase in ee values observed.
Molecular Structures and Conformational Variability of
Complexes [PdCl2(L)] [L ) (R,Sp)-3, (Rp)-9c, and (Sp)-10c-
12c]. It was observed in the hydrogenation reactions described
above that structural changes made to Xyliphos (3) as the chiral
catalyst ligand, and in particular the removal of its stereogenic
center, led to significant changes in ee values. On the basis of
this information we questioned how these structural changes in
the ligands could be reflected in their respective metal com-
plexes. Therefore, palladium dichloride complexes [PdCl2(L)]
of ligands 3 and 9c-12c were prepared as model compounds,
and their structures were determined in the solid state (L ) 3,
9c-11c) as well as in solution.
The molecular structures of complexes [PdCl2(L)] [L )
(R,Sp)-3, (Rp)-9c, (Sp)-10c, and (Sp)-11c] were determined in
the solid state by X-ray diffraction. Details on X-ray crystal-
lography are given in Table 3 and in the Experimental Section.
The absolute configuration of each compound was determined
from the X-ray anomalous dispersion effects and was consistent
with the chemical evidence. Views of the molecular structures
of these compounds are shown in Figures 1 and 2, and selected
geometrical data are given in Table 2.
In all solid-state structures of mononuclear metal complexes
of Xyliphos (3) reported to date Xyliphos acts as a bidentate
ligand.20,21 As in the complex [Ir(cod)((S,Rp)-3)][BF4]21 (Figure
1, right), the chelate ring formed on coordination to the metal
always adopts a conformation that places the metal unit in an
exo-position with the phosphorus atom linked to the ferrocenyl
ethyl side chain and the metal atom positioned above the
substituted Cp ring. In this conformation the carbon atom of
the side chain methyl group (C11A) is located almost in the
least-squares plane of the substituted cylopentadienyl ring (Cp)
and the diphenylphosphino unit is oriented in such a way that
one of the phenyl rings (C12-C17) is placed in close proximity
to the unsubstituted cylopentadienyl ring (Cp′).
3
9c
10c
11c
Bond Lengths (Å)
Pd-Cl1
Pd-Cl2
Pd-P1
2.351(3)
2.336(3)
2.214(3)
2.256(3)
1.813(5)
1.834(9)
2.358(1)
2.362(1)
2.246(1)
2.256(1)
1.776(2)
1.845(2)
2.353(1)
2.343(1)
2.239(1)
2.240(1)
1.787(2)
1.846(3)
1.499(3)
2.363(1)
2.342(1)
2.251(1)
2.239(1)
1.793(5)
1.847(5)
1.507(7)
Pd-P2
P1-C1
P2-C11
C2-C11
1.529(11) 1.490(3)
Bond Angles (deg)
Cl1-Pd-Cl2
P1-Pd-P2
88.6(1)
93.6(1)
91.58(2)
93.50(2)
125.8(2)
125.6(2)
124.5(2)
128.1(2)
117.2(1)
91.51(2)
94.48(2)
125.9(2)
125.1(2)
125.6(2)
126.3(2)
118.1(2)
126.4(2)
126.3(2)
91.64(5)
94.20(5)
123.5(4)
126.9(4)
125.1(4)
127.0(4)
117.3(3)
127.1(4)
125.3(4)
C1-C2-C11
C3-C2-C11
C2-C1-P1
C5-C1-P1
C2-C11-P2
C2-C3-C3A
C4-C3-C3A
126.1(6)
125.0(6)
128.4(4)
123.4(4)
116.1(6)
Torsion Angle (deg)
C18-P1-C1-C5 -88.1(6) 99.2(2)
-88.1(2)
-90.2(5)
Tilt Angle (deg)
Cp/Cp′
4.2(2)
5.2(1)
4.7(1)
4.4(2)
Distance from Least-Squares Plane (Å)
-0.87(2) -1.081(5) -0.803(5) -0.940(11)
0.11(2) 0.068(3) 0.125(4) 0.056(8)
-0.53(2) -0.798(4) -0.651(5) -0.631(10)
Pd/Cp
P1/Cp
P2/Cp
C11/Cp
C3A/Cp
H7-Cp′/Xyl2
0.23(2)
0.183(3)
0.216(4)
0.000(5)
2.616(3)
0.323(9)
-0.013(9)
2.618(6)
2.70(1)
2.553(3)
3)][BF4], in palladium dichloride complexes [PdCl2(3)], [PdCl2-
(9c)], [PdCl2(10c)], and [PdCl2(11c)] (Figures 1 and 2) the
respective chelate rings adopt a totally different conformation,
with the palladium dichloride unit now in an endo-position. This
places both the side chain-linked phosphorus P2 and the
palladium atom below the Cp plane. In this conformation one
of the xylyl rings (Xyl2: C24-C29) is positioned in close
proximity to Cp′ (for details see Table 2), and this interaction
clearly causes significant out-of-plane bending of the methylene
carbon C11, to which the dixylylphosphino unit is linked
(normal distance of C11 to the Cp least-squares plane: [PdCl2-
(3)] 0.23 Å, [PdCl2(9c)] 0.18 Å, [PdCl2(10c)] 0.22 Å, [PdCl2-
(11c)] 0.32 Å) as well as an increase in the bond angles C2-
C11-P2 to a maximum of 118.1° [PdCl2(10c)]. In the case of
complex [PdCl2(3)], as compared to the exo chelate ring
conformation of [Ir(cod)((S,Rp)-3)][BF4], the endo chelate ring
conformation is accompanied by a change in the side chain
conformation that results in the side chain methyl group being
positioned above the substituted Cp ring (normal distance of
C11A to the Cp least-squares plane: 1.79 Å). In addition, the
Cp rings are slightly tilted against each other, and this tilt also
leads to the release of steric strain between Cp′ and the
phosphorus-linked Xyl2 ring. A comparison of the molecular
structures of complexes [PdCl2(L)] (L ) 9c, 10c, and 11c) shows
that attaching substituents to the Cp carbon C3 (ligands 10c
and 11c as compared to 9c) leads to an increase in the out-of-
plane deformation of C11, which in turn causes the palladium
dichloride unit to move closer to the Cp plane (distance between
Pd and the Cp least-squares plane: 3 -0.87 Å, 9c -1.08 Å,
10c -0.80 Å, 11c -0.94 Å). This change is associated with a
counterclockwise rotation of the diphenylphosphino unit about
the P1-C1 bond of about 10° [torsion angle C18-P1-C1-
C5: -88.1° (10c) and -90.2° (11c) versus -99.2° for the
enantiomer of 9c].
Like Xyliphos (3), the analogous diphosphines 9c, 10c, and
11c act as bidentate ligands, but, as compared to [Ir(cod)((S,Rp)-
(20) Dorta, R.; Togni, A. Organometallics 1998, 17, 5441.
(21) Dorta, R.; Broggini, D.; Stoop, R.; Ru¨egger, H.; Spindler, F.; Togni,
A. Chem.-Eur. J. 2004, 10, 267.
In summary, in the solid state the chelate ring in all palladium
dichloride complexes [PdCl2(L)] (L ) 3, 9c, 10c, and 11c)

3536 Organometallics, Vol. 26, No. 14, 2007
Wang et al.
Table 3. Details of the Crystal Structure Determinations of Complexes [PdCl2(L)] (L ) (R,Sp)-3, (Rp)-9c, (Sp)-10c, and (Sp)-11c)
b
[PdCl2(3)]‚2C2H4Cl2
[PdCl2(9c)]‚1/2CH2Cl21/2C2H5OH)
[PdCl2(10c)]‚solvb
[PdCl2(11c)]‚CH2Cl2
formula
fw
cryst size, mm
space group
a, Å
C44H48Cl6FeP2Pd
1013.71
0.41 × 0.14 × 0.02
P212121 (no. 19)
9.348(3)
18.578(6)
26.781(9)
4651(3)
4
1.448
297(2)
1.142
2064
C40.5H42Cl3FeO0.5P2Pd
867.28
C40H40Cl2FeP2Pd
815.81
C46H44Cl4FeP2Pd
962.80
0.38 × 0.36 × 0.14
P212121 (no. 19)
11.5451(5)
17.7582(8)
18.3153(8)
3755.0(3)
4
1.534
100(2)
1.194
1768
0.37 × 0.18 × 0.16
P212121 (no. 19)
9.7526(5)
19.0518(10)
22.4483(12)
4171.0(4)
4
1.299
100(2)
1.008
1664
0.28 × 0.04 × 0.04
P212121 (no. 19)
9.7252(5)
18.5716(10)
23.1105(14)
4174.0(4)
4
1.532
100(2)
1.144
1960
b, Å
c, Å
V, Å3
Z
F
calc, g cm-3
T, K
µ, mm-1 (Mo KR)
F(000)
θ
max, deg
25
25 035
8142
2976
353
0.0778
0.1628
30
52 642
10 927
10 802
445
0.0234
0.0238
30
61 556
12 044
11 563
420
0.0312
0.0331
28.3
93 507
10 379
8279
491
0.0497
0.0738
no. of rflns measd
no. of unique rflns
no. of rflns I > 2σ(I)
no. of params
R1 (I > 2σ(I))a
R1 (all data)
wR2 (all data)
0.1420
0.0556
0.0756
0.1215
Flack abs str param
diff Fourier peaks
min./max., e Å-3
-0.06(5)
-1.01/0.92
-0.003(10)
-0.40/0.77
-0.017(12)
-0.71/1.44
-0.02(2)
-1.19/1.21
2
2
2
a R1 ) ∑||Fo| - |Fc||/∑|Fo|, wR2 ) [∑(w(Fo - Fc )2)/∑(w(Fo )2)]1/2
.
b Disordered solvents (1,2-dichloroethane for [PdCl2(3)] and Et2O/CH2Cl2 for
[PdCl2(10c)]) were squeezed with the program PLATON.25 Chemical formula and derived quantities for [PdCl2(10c)] are given without solvent content.
Figure 1. Molecular structures of [PdCl2(3)] and [Ir(cod)(3)][BF4] (from ref 21, anion omitted for clarity).
Figure 2. Molecular structures of [PdCl2(L)] (L ) 9c, 10c, and 11c) in the solid state.
adopts a conformation in which the palladium dichloride unit
is in an endo-position. This is particularly striking for the parent
ligand Xyliphos (3) since in this case the chelate ring conforma-
tion depends on the coordinating metal-ligand unit M(L)n. With
Ir(cod) ([Ir(cod)(3)][BF4]) the exo chelate ring conformation is
preferred and with PdCl2 [PdCl2(3)] the endo conformation

Planar Chiral Ferrocenyl Diphosphines
Organometallics, Vol. 26, No. 14, 2007 3537
Chart 4. Numbering Scheme and Main NOE Interactions
for [PdCl2(9c)]
(Figure 1). For complexes of ligands Josiphos (1) and PPF-t-
Bu2 (2) such conformational changes are very unlikely. Regard-
less of the metal-ligand unit, in the solid state in mononuclear
complexes of ligands 1 and 2 the chelate ring has always been
found to have an exo-conformation. It is clear that the dixyl-
ylphosphino unit of 3 is sterically much less demanding than
the Cy2P and t-Bu2P substituents of 1 and 2, and therefore both
exo and endo chelate ring conformations are feasible. This view
is supported by the fact that a chelate ring conformation with
the metal unit in an endo-position has been reported in one
complex of a Xyliphos-related diphosphine ligand, [PdCl(CH3)-
(L)] (L ) 1-Ph2P-2-Ph2PCH(CH3)-Fc), in which the dixyl-
ylphosphino group of 3 is replaced by a Ph2P unit.22 In
complexes of 1 and 2, in which the chelate ring adopts an endo-
conformation, one would expect strong steric interactions
between one cyclohexyl (1) or one tert-butyl substituent (2) and
the respective Cp′ ring, a fact that strongly disfavors the endo-
conformation.
The replacement of ligand 9c by either 10c or 11c causes
comparably small conformational changessmainly in the chelate
ringsand these are attributed to steric interactions between the
Cp substituents in position 3 and the dixylylphosphinomethyl
side chain. However, the presence of the Cp-C3 substituents
CH3 (10c) and Ph (11c) seems to have a very similar influence
since both complexes [PdCl2(10c)] and [PdCl2(11c)] have almost
identical molecular structures.
Overhauser enhancements are indicated in Chart 4, and most
of these interactions are also characteristic of complexes with
ligands 10c-12c. In all of these complexes, the observed NOEs
are consistent only with a chelate ring conformation with the
palladium dichloride unit in an endo-position. This follows
primarily from the fact that strong interactions between the Cp′
protons and two different ringssone xylyl (Xyl2) and one phenyl
ring (Ph2)sare observed. For a conformer with the palladium
dichloride unit in an exo-position, a strong interaction between
Cp′ and xylyl ortho protons would not be expected since such
an arrangement must lead to both xylyl rings being located
above the substituted Cp ring.
The molecular structures of complexes [PdCl2(L)] (L ) 3,
9c, 10c-12c) in solution were determined by NMR spectros-
copy, mainly by analyzing nuclear Overhauser interactions. The
assignment of 1H, 13C, and 31P signals was made using standard
one- and two-dimensional techniques. The 31P spectra appear
as one would expect; on complexation all phosphorus signals
are shifted to lower field by about 40 ppm. However, the main
structural information was obtained from proton and predomi-
nately from NOESY spectra. On the basis of the results
described above one might expect that, as in the solid state, in
solution complexes [PdCl2(L)] (L ) 3, 9c, 10c-12c) could
adopt two chelate ring conformations with the palladium
dichloride unit either in an exo- or in an endo-position.
Furthermore, dynamic equilibria of such conformers must also
be considered. A comparison of the proton NMR spectra of all
complexes shows some common features (for the ring and
proton numbering see Chart 4). As compared to the free ligands,
in each case on complexation the signals of the methyl groups
of xylyl ring 1 (Xyl1) and the Cp-H5 and the Cp′ signals are
shifted to higher field, while the methyl protons of xylyl ring 2
(Xyl2), Cp-H3 (3 and 9c), and Cp-H4 (10c-12c) are shifted to
lower field. In addition, for complexes [PdCl2(L)] (L ) 9c, 10c-
12c) the CH2 proton Hin (Chart 4) always appears at higher field
than the CH2 proton Hout. Significant differences in proton
spectra are seen only in the aromatic region. For complexes
[PdCl2(L)] (L ) 9c, 10c-12c) the appearance of this region of
the spectrum is almost identical, with the signals of the ortho
protons of xylyl ring 1 (Xyl1-ortho) always located at highest
field and the ortho protons of xylyl ring 2 (Xyl2-ortho) at lowest
field. However, in the case of the Xyliphos complex [PdCl2(3)]
the appearance of the spectrum is totally different, with the ortho
protons of phenyl ring 2 (Ph2-ortho) now at lowest field and
the Xyl1-para proton at highest field.
However, for the Xyliphos complex [PdCl2(3)] the interpreta-
tion of the observed set of NOEs is less obvious. In this case
the signal of the Cp′ protons overlaps with the signal of the
side chain CH proton, and therefore interactions between xylyl
ortho protons and Cp′ protons (indicative of an endo chelate
ring conformation) could not be assigned unequivocally. Only
one cross-peak between protons of the side chain methyl group
and the otho protons of one phenyl group allows the differentia-
tion between exo and endo chelate ring conformations. Such
an interaction is expected only when the side chain methyl group
[C(11A)H3] is located above the least-squares plane of the
substituted Cp ring (as in [PdCl2(3)]) and not when this methyl
group is located in close proximity to this plane (as in [Ir(cod)-
(3)][BF4], Figure 1). Therefore, we tentatively assign the endo-
conformation to the chelate ring of complex [PdCl2(3)] in
solution.
In summary, in the solid state and in solution all palladium
dichloride complexes [PdCl2(L), L ) 3, 9c-12c] adopt a
molecular structure in which the chelate ring is in an endo-
arrangement. In the solid state, both the removal of the side
chain methyl group (9c versus 3) and addition of substituents
in position 3 of the Cp ring (10c, 11c versus 9c) lead only to
small structural changes. However, the fact that in Xyliphos
complexessdepending on the coordinated metal-ligand unit
(MLn)sthe chelate ring can adopt both the exo- and the endo-
arrangement seems to indicate a generally higher conformational
flexibility for Xyliphos-type complexes as compared to Josiphos-
or PPF-t-Bu2-type complexes.
Like the 1H spectra, the NOESY spectra of complexes
[PdCl2(L)] (L ) 9c, 10c-12c) also show strong similarities.
For complex [PdCl2(9c)] the main interactions that cause nuclear
(22) Gambs, C.; Consiglio, G.; Togni, A. HelV. Chim. Acta 2001, 84,
3105.

3538 Organometallics, Vol. 26, No. 14, 2007
Wang et al.
NMR (400 MHz): δ 1.00 (dd, 3H, J ) 14.1 Hz, J ) 7.3 Hz, CH3),
2.19 (s, 6H, Xyl1-CH3), 2.42 (s, 6H, Xyl2-CH3), 3.62 (s, 5H, Cp′),
3.64-3.72 (m, 1H, CH), 3.99-4.03 (m, 1H, Cp-H5), 4.33-4.39
(m, 1H, Cp-H4), 4.40-4.44 (m, 1H, Cp-H3), 6.98 (s, 1H, Xyl1-
para), 7.17 (s, 1H, Xyl2-para), 7.27 (d, 2H, J ) 10.9 Hz, Xyl1-
ortho), 7.27-7.32 (m, 2H, Ph1-meta), 7.37-7.44 (m, 1H, Ph1-para),
7.48-7.56 (m, 3H, Ph2-meta, Ph2-para), 7.56-7.63 (m, 2H, Ph1-
ortho), 7.71 (d, J ) 11.9 Hz, Xyl2-ortho), 8.00-8.11 (m, 2H, Ph2-
ortho). 13C{1H} NMR (100 MHz): δ 19.90 (d, J ) 1.5 Hz, CH3),
21.33 (2C, Xyl1-CH3), 21.53 (2C, Xyl2-CH3), 32.67 (dd, J ) 27.0
Hz, J ) 7.7 Hz, CH), 67.76 (dd, J ) 57.6 Hz, J ) 11.5 Hz, Cp-
Cq), 69.28 (dd, J ) 6.9 Hz, J ) 2.3 Hz, Cp-C4), 70.69 (5C, Cp′),
70.90 (dd, J ) 7.5 Hz, J ) 5.0 Hz, Cp-C3), 73.91 (d, J ) 3.0 Hz,
Cp-C5), 92.82 (dd, J ) 17.6 Hz, J ) 3.8 Hz, Cp-Cq), 127.40 (d, J
) 51.6 Hz, Ph-ipso), 127.77 (d, 2C, J ) 11.5 Hz, Ph2-meta), 128.04
(d, 2C, J ) 11.5 Hz, Ph1-meta), 128.54 (d, J ) 53.0 Hz, Ph-ipso),
130.73 (d, J ) 3.0 Hz, Ph1-para), 131.24 (d, J ) 3.0 Hz, Ph2-
para), 131.69 (d, J ) 56.9 Hz, Xyl-ipso), 132.11 (d, 2C, J ) 10.9
Hz, Xyl1-ortho), 132.24 (d, 2C, J ) 10.9 Hz, Xyl2-ortho), 132.85
(d, J ) 3.0 Hz, Xyl1-para), 133.42 (d, J ) 3.0 Hz, Xyl2-para),
134.02 (d, 2C, J ) 10.7 Hz, Ph1-ortho), 135.30 (d, 2C, J ) 11.5
Hz, Ph2-ortho), 137.65 (d, 2C, J ) 11.7 Hz, Xyl-Cq), 138.15 (d,
2C, J ) 11.7 Hz, Xyl-Cq). Ph-ipso and Xyl-ipso interchangeable.
31P{1H} NMR (162 MHz): δ +17.11 (d, J ) 7.9 Hz), +48.59 (d,
J ) 7.9 Hz). Anal. Calcd for C40H40Cl2FeP2Pd: C 58.89, H 4.94,
P 7.59. Found: C 58.67, H 4.85, P 7.50.
Summary
A number of solely planar chiral analogues of ferrocenyl
diphosphines Josiphos, 1, PPF-t-Bu2, 2, and Xyliphos, 3, have
been tested as catalyst ligands in hydrogenation reactions of
four alkenes and two ketones. For these hydrogenations, as
compared to catalysts based on the reference ligands 1 and 2,
the influence of ligands lacking the stereogenic center on
enantioselectivity was found to be rather small. Only in one
case was a severe drop in ee values observed. More significant
changes were seen for analogues of Xyliphos. In three cases,
the absolute configuration of the product changed when Xyli-
phos was replaced by ligands without a stereogenic center, and
in one case the enantioselectivity dropped significantly. A model
study of the molecular structures of dichloro palladium(II)
complexes of Xyliphos and analogues revealed that for Xyliphos
complexes the conformation of the chelate ring that is formed
on complexation of the diphosphine ligand to the metal depends
strongly on the co-coordinated metal-ligand unit MLn. In
complexes based on Xyliphos the respective chelate ring can
adopt both the exo- and the endo-conformation. Attempts to
use additional Cp substituents for ligand fine-tuning proved
successful only in one case. In the hydrogenation of methyl
R-acetamido-cinnamate the ee value of 49% obtained with the
1,2-disubstituted ligand 9c could be increased to 84% on using
the 1,2,3-trisubstituted ligand 12c.
[PdCl2((Rp)-9c)]. Yield: 63%. Mp: dec >210 °C. [R]λ20 (nm):
-265 (589), -311 (578), -608 (546) (c 0.074, CHCl3). 1H NMR
(400 MHz): δ 2.06 (s, 6H, Xyl1-CH3), 2.47 (s, 6H, Xyl2-CH3),
2.96 (dd, 1H, J ) 16.1 Hz, J ) 7.1 Hz, CH2-in), 3.50 (ddd, 1H, J
) 16.1 Hz, J ) 16.1 Hz, J ) 5.4 Hz, CH2-out), 3.68-3.72 (m,
1H, Cp-H5), 3.71 (s, 5H, Cp′), 4.36 (dd, 1H, J ) 5.0 Hz, J ) 2.5
Hz, Cp-H4), 4.37-4.40 (m, 1H, CP-H3), 6.61 (d, 2H, J ) 12.4
Hz, Xyl1-ortho), 6.88 (s, 1H, Xyl1-para), 7.25 (s, 1H, Xyl2-para),
7.40-7.50 (m, 4H, Ph1-meta + Ph2-meta), 7.50-7.56 (m, 2H, Ph1-
para + Ph2-para), 7.69-7.76 (m, 2H, Ph1-ortho), 7.76-7.83 (m,
2H, Ph2-ortho), 8.01 (d, 2H, J ) 12.4 Hz, Xyl2-ortho). 13C{1H}
NMR (100 MHz): δ 21.21 (2C, Xyl1-CH3), 21.46 (2C, Xyl2-CH3),
28.23 (dd, J ) 30.7 Hz, J ) 8.3 Hz, CH2), 65.60 (dd, J ) 63.6 Hz,
J ) 10.1 Hz, Cp-Cq), 69.35 (dd, J ) 6.9 Hz, J ) 3.5 Hz, Cp-C4),
70.50 (5C, Cp′), 73.14-73.46 (m, 2C, Cp-C3, Cp-C5), 86.40 (d, J
) 17.6 Hz, Cp-Cq), 127.10 (d, J ) 54.9 Hz, Ph-ipso), 127.57 (d,
2C, J ) 12.2 Hz, Ph2-meta), 128.95 (d, 2C, J ) 11.5 Hz, Ph1-
meta), 129.32 (d, J ) 68.1 Hz, Xyl-ipso), 129.41 (d, 2C, J ) 9.9
Hz, Xyl1-ortho), 130.77 (d, J ) 53.3 Hz, Ph-ipso), 131.16, 131.19,
131.23 (2C, Ph-para), 132.10 (d, J ) 57.5 Hz, Xyl-ipso), 132.16
(d, J ) 3.1 Hz, Xyl1-para), 133.67 (d, 2C, J ) 10.7 Hz, Ph1-ortho),
133.89 (d, 2C, J ) 11.5 Hz, Xyl2-ortho), 134.04 (d, J ) 2.3 Hz,
Xyl2-para), 134.67 (d, 2C, J ) 11.5 Hz, Ph2-ortho), 137.79 (d,
2C, J ) 11.5 Hz, Xyl-Cq), 138.63 (d, 2C, J ) 12.2 Hz, Xyl-Cq).
Ph1-meta and Ph2-meta as well as Ph-ipso and Xyl-ipso inter-
changeable. 31P{1H} NMR (162 MHz): δ +20.29 (d, J ) 7.9 Hz),
+28.29 (d, J ) 7.9 Hz). Anal. Calcd for C39H38FeP2‚PdCl2: C
58.42, H 4.78, P 7.73. Found: C 58.18, H 4.88, P 7.59.
Experimental Section
General Comments. All reactions required inert conditions and
were carried out under an argon atmosphere using standard Schlenk
techniques. All solvents were dried by standard procedures and
distilled before use. 1H, 13C{1H}, and 31P{1H} NMR spectra were
recorded on a Bruker DPX-400 spectrometer in CDCl3. Chemical
shifts are given relative to CHCl3 (1H: 7.26 ppm), CDCl3 (13C:
77.0 ppm), and 85% H3PO4 (31P: 0 ppm). The coupling constants
in 13C spectra were due to 13C-31P or 13C-19F coupling. In the 1H
NMR data br s, d, t, and q refer to broad singlet, doublet, triplet,
and quartet, respectively, and Cq in 13C NMR data stands for
quaternary carbon atom. For the signal assignment for complexes
[PdCl2(L)] (L ) 3c, 9c, 10c, 11c, or 12c) the numbering scheme
depicted in Chart 4 was used. For all other diphenylphosphino-
and dixylylphosphino-substituted derivatives the terms Ph1, Ph2 and
Xyl1, Xyl2 are arbitrarily used to denote the phosphorus-linked
phenyl and xylyl rings, respectively. Melting points were determined
on a Kofler melting point apparatus and are uncorrected. Mass
spectra were recorded on a Finnigan MAT 8230 spectrometer (EI).
Optical rotations were measured on a Perkin-Elmer 241 polarimeter.
Chromatographic separations were performed under gravity either
on silica gel (Merck, 40-62 µm) or on alumina (Merck, activity
II-III, 0.063-0.200 mm). Petroleum ether with a boiling range of
55-65 °C was used for chromatography.
Compounds (Rp)-9a, (Rp)-9b, (1R,2S)-14, (1R,2S,Rp)-15, and
(1R,2S,Rp)-18 were prepared according to ref 17. For the synthesis
of ligands (Rp)-9c, (Sp)-10a-c, (Sp)-11a-c, and (Sp)-12a-c and
their precursors (1R,2S,Rp)-16, (1R,2S,Rp)-17, and (1R,2S,Sp)-19-
24 see the Supporting Information.
20
[PdCl2((Rp)-10c)]. Yield: 56%. Mp: dec >175 °C. [R]λ
1
(nm): +50 (589), +50 (578), +150 (546) (c 0.006, CHCl3). H
NMR (400 MHz): δ 1.93 (s, 3H, Cp-CH3), 2.07 (s, 6H, Ph1-CH3),
2.47 (s, 6H, Ph2-CH3), 2.77 (dd, 1H, J ) 15.8 Hz, J ) 7.4 Hz,
CH2-in), 3.54 (ddd, 1H, J ) 15.8 Hz, J ) 15.8 Hz, J ) 5.6 Hz,
CH2-out), 3.60 (s, 5H, Cp′), 3.62 (t, 1H, J ) 2.0 Hz, Cp-H5), 4.33
(t, 1H, J ) 2.7 Hz, Cp-H4), 6.64 (d, 2H, J ) 12.4 Hz, Xyl1-ortho),
6.90 (s, 1H, Xyl1-para), 7.26 (s, 1H, Xyl2-para), 7.40-7.48 (m,
4H, Ph1-meta + Ph2-meta), 7.48-7.55 (m, 2H, Ph1-para + Ph2-
para), 7.69-7.76 (m, 2H, Ph1-ortho), 7.76-7.82 (m, 2H, Ph2-
ortho), 8.04 (d, 2H, J ) 12.4 Hz, Xyl-ortho). 13C{1H} NMR (100
MHz): δ 13.06 (CH3), 21.23 (2C, Xyl1-CH3), 21.47 (2C, Xyl2-
CH3), 25.91 (dd, J ) 31.5 Hz, J ) 8.8 Hz, CH2), 71.20 (Cp′),
General Procedure for the Synthesis of PdCl2 Complexes.
[PdCl2(L)] L ) 3c, 9c, 10c, 11c, or 12c. A degassed solution of
3c, 9c, 10c, 11c, or 12c (0.1 mmol) in dry benzene (2 mL) was
added to a suspension of dichlorobis(acetonitrile)palladium(II) (26
mg, 0.1 mmol) in dry benzene (2 mL) through a Teflon tube. The
mixture was stirred for 18 h at rt, and the resulting precipitate was
filtered off and washed with benzene (2 × 2 mL) and diethyl ether
(3 × 2 mL) to give the palladium complexes as red solids.
20
[PdCl2((R,Sp)-3)]. Yield: 86%. Mp: dec >270 °C. [R]λ
(nm): +170 (589), +206 (578), +418 (546) (c 0.099, CHCl3). 1H

Planar Chiral Ferrocenyl Diphosphines
Organometallics, Vol. 26, No. 14, 2007 3539
71.41-71.69 (m, 2C, Cp-C4, Cp-C5), 85.35 (d, J ) 16.8 Hz, Cp-
Cq), 127.00 (d, J ) 54.3 Hz, Ph-ipso), 127.52 (d, 2C, J ) 11.5
Hz, Ph1-meta), 128.91 (d, 2C, J ) 11.5 Hz, Ph2-meta), 129.34 (d,
J ) 68.0 Hz, P-Xyl-ipso), 129.48 (d, 2C, J ) 9.9 Hz, Xyl1-ortho),
130.23 (d, J ) 71.6 Hz, P-Xyl-ipso), 131.11 (d, J ) 3.0 Hz, Ph1-
para), 131.18 (d, J ) 3.0 Hz, Ph2-para), 132.07 (d, J ) 57.4 Hz,
Ph-ipso), 132.26 (d, J ) 3.0 Hz, Xyl1-para), 133.69 (d, 4C, J )
11.5 Hz, Xyl2-ortho, Ph1-ortho), 133.98 (d, J ) 3.0 Hz, Xyl2-para),
134.82 (d, 2C, J ) 11.5 Hz, Ph2-ortho), 137.88 (d, 2C, J ) 12.3
Hz, Xyl-Cq), 138.74 (d, 2C, J ) 11.5 Hz, Xyl-Cq). Ph1-meta and
Ph2-meta, Ph1-para and Ph2-para, as well as Ph-ipso and Xyl-ipso,
interchangeable; two Cp-Cq not observed. 31P{1H} NMR (162
MHz): δ +22.3 (d, J ) 7.9 Hz), +28.7 (d, J ) 7.9 Hz).
Standard Procedure for Hydrogenation Reactions. The sub-
strate (2.53 mmol) and the catalyst (formed in situ, see below) were
dissolved separately in the appropriate solvent (5 mL) under argon
(total volume: 10 mL). The catalyst solution was stirred for 15 min.
Both the catalyst and the substrate solution were transferred through
a steel capillary into a 180 mL thermostated glass reactor or a 50
mL stainless steel autoclave. The inert gas was then replaced by
hydrogen (three cycles), and the pressure was set. After completion
of the reaction (total reaction times 1-20 h), the conversion was
determined by gas chromatography and the product was recovered
quantitatively after filtering the reaction solution through a plug of
silica to remove the catalyst. The enantiomeric purity of the product
was determined either by gas chromatography or by HPLC.
20
The following reaction conditions and methods for ee determi-
nation were applied:
[PdCl2((Rp)-11c)]. Yield: 42%. Mp: dec >185 °C. [R]λ
(nm): +209 (589), +257 (578), +504 (546) (c 0.067, CHCl3). 1H
NMR (400 MHz): δ 2.09 (s, 6H, Xyl1-CH3), 2.57 (s, 6H, Xyl2-
CH3), 2.62 (dd, 1H, J ) 16.1 Hz, J ) 6.6 Hz, CH2), 3.64 (s, 5H,
Cp′), 3.68-3.81 (m, 1H, CH2), 3.81-3.88 (m, 1H, Cp-H5), 4.64
(t, 1H, J ) 2.5 Hz, Cp-H4), 6.49 (d, 2H, J ) 12.5 Hz, Xyl1-ortho),
6.84 (s, 1H, Xyl1-para), 7.26-7.35 (m, 5H, Cp-Ph), 7.38 (s, 1H,
Xyl2-para), 7.43-7.61 (m, 6H, P-Ph-meta, P-Ph-para), 7.74-7.89
(m, 4H, P-Ph-ortho), 8.18 (d, 2H, J ) 12.5 Hz, Xyl2-ortho). 13C{1H}
NMR (100 MHz): δ 21.16 (2C, Xyl1-CH3), 21.50 (2C, Xyl2-CH3),
24.80-25.29 (m, CH2), 66.06 (dd, J ) 63.4 Hz, J ) 9.3 Hz, Cp-
C1), 69.75-69.93 (m, Cp-C4), 71.95 (br s, 5C, Cp′), 72.98 (d, J )
3.8 Hz, Cp-C5), 85.18 (d, J ) 16.3 Hz, Cp-C3), 92.50 (t, J ) 6.9
Hz, Cp-C2), 126.03 (d, J ) 55.8 Hz, P-Ph-ipso), 127.55 (d, 2C, J
) 3.8 Hz, P-Ph1-meta), 127.65 (Cp-Ph-para), 128.22 (d, J ) 54.1
Hz, P-Ph-ipso), 128.28 (2C, Cp-Ph-meta), 129.10 (d, 2C, J ) 9.8
Hz, P-Ph2-meta), 129.19 (d, 2C, J ) 10.7 Hz, Xyl1-ortho), 129.83
(2C, Cp-Ph-ortho), 130.29 (d, J ) 51.8 Hz, P-Xyl-ipso), 131.33
(d, 2C, J ) 3.0 Hz, P-Ph1-para + P-Ph2-para), 132.14 (d, J ) 3.0
Hz, Xyl1-para), 132.26 (d, J ) 58.8 Hz, P-Xyl-ipso), 133.73 (d,
2C, J ) 11.5 Hz, P-Ph1-ortho), 134.25 (d, J ) 3.0 Hz, Xyl2-para),
134.54 (d, 2C, J ) 11.5 Hz, P-Ph2-ortho), 134.80 (d, 2C, J ) 11.5
Hz, Xyl2-ortho), 135.48 (Cp-Ph-ipso), 137.77 (d, J ) 12.2 Hz,
P-Xyl-Cq), 138.88 (d, J ) 11.6 Hz, P-Xyl-Cq). P-Ph1 and P-Ph2,
Cp-Ph-meta and Cp-Ph-ortho, as well as P-Ph-ipso and P-Xyl-ipso,
interchangeable. 31P{1H} NMR (162 MHz): δ +23.14 (d, J ) 7.9
Hz), +29.36 (d, J ) 7.9 Hz).
MAC: 2.53 mmol (0.25 mol/L) of MAC; [Rh(NBD)2][BF4] +
1.05 equiv of ligand; s/c ) 200; solvent: MeOH (10 mL); p(H2):
1 bar; 25 °C; reaction time: 1 h. ee: GC; Chirasil-L-Val, 170 °C,
isotherm.
DMI: 2.53 mmol (0.25 mol/L) of DMI; [Rh(NBD)2][BF4] +
1.05 equiv of ligand; s/c ) 200; solvent: MeOH (10 mL); p(H2):
1 bar; 25 °C; reaction time: 1 h. ee: GC; Lipodex E, 80 °C,
isotherm.
MCA: 2.53 mmol (0.25 mol/L) of MCA; [Rh(NBD)2][BF4] +
1.05 equiv of Ligand; s/c ) 200; solvent: MeOH (10 mL); p(H2):
5 bar; 25 °C; reaction time: 16 h. ee: as methyl ester; HPLC;
Chiralcel OB, Hex/i-PrOH: 98:2, 0.1 mL/min.
EAC: 2.53 mmol (0.25 mol/L) of EAC: [Rh(NBD)2][BF4] +
1.05 equiv of ligand; s/c ) 200; solvent: EtOH (9.5 mL);
CF3CH2OH (0.5 mL); p(H2): 1 bar; 25 °C; reaction time: 16 h.
ee: GC; Lipodex E, 130 °C, isotherm.
MPG: 2.53 mmol (0.25 mol/L) of MPG; [Rh(NBD)Cl]2 + 2.1
equiv of ligand; s/c ) 200; solvent: toluene (10 mL); p(H2): 80
bar; 25 °C; reaction time: 20.5 h. ee: HPLC; Chrialcel OJ, Hex/
i-PrOH: 90:10, 1.0 mL/min.
EOP: 2.53 mmol (0.25 mol/L) of EOP; [RuI2(p-cymene)]2 +
2.1 equiv of ligand (no pretreatment); s/c ) 200; solvent: ethanol
(10 mL); additive: 1 N HCl(aq), 60 µL; p(H2): 80 bar; 80 °C;
reaction time: 16 h. ee: as TFA derivative; GC, Lipodex E, 80
°C, isotherm.
20
[PdCl2((Rp)-12c)]. Yield: 40%. Mp: dec >210 °C. [R]λ
X-ray Structure Determination. Crystals of [PdCl2(9c)], [PdCl2-
(10c)], and [PdCl2(11c)] in the form of [PdCl2(9c)]‚(1/2CH2Cl2,1/
2C2H5OH), [PdCl2(10c)]‚solv, and [PdCl2(11c)]‚CH2Cl2 were ob-
tained by slow diffusion of Et2O into a solution of the corresponding
complex in CH2Cl2 or CH2Cl2/C2H5OH ([PdCl2(9c)]). In the case
of [PdCl2(3)] this procedure yielded only unsuitable thin needles.
However, when this complex was crystallized from a highly
concentrated solution in 1,2-dichlorethane, a small plate-like crystal
of [PdCl2(3)]‚2C2H4Cl2 was recovered from a polycrystalline mass.
Crystal data and experimental details are given in Table 3. X-ray
data were collected on Bruker Smart CCD area detector diffrac-
tometers using graphite-monochromated Mo KR radiation (λ )
0.71073 Å) and 0.3° ω-scan frames covering complete spheres of
the reciprocal space. Corrections for absorption, λ/2 effects, and
crystal decay were applied.23 The structures were solved by direct
methods using the program SHELXS97.24 Structure refinement on
F2 was carried out with the program SHELXL97.24 All non-
hydrogen atoms were refined anisotropically. Hydrogen atoms were
inserted in idealized positions and were refined riding with the atoms
to which they were bonded. Solvent disorder was taken into account.
Thus, for [PdCl2(3)]‚2CH2Cl2 and [PdCl2(10c)]‚solv (solv ) mainly
(nm): +27 (589), +36 (578), +91 (546) (c 0.135, CHCl3). 1H NMR
(400 MHz): δ 2.01 (s, 6H, Xyl1-CH3), 2.34 (s, 6H, Cp-Xyl-CH3),
2.55 (s, 6H, Xyl2-CH3), 2.70 (dd, 1H, J ) 16.1 Hz, J ) 7.0 Hz,
CH2), 3.68 (s, 5H, Cp′), 3.72-3.84 (m, 1H, CH2), 3.83 (t, 1H, J )
2.0 Hz, Cp-H5), 4.59-4.61 (m, 1H, Cp-H4), 6.50 (d, 2H, J ) 12.4
Hz, Xyl1-ortho), 6.85 (s, 1H, Xyl1-para), 6.87 (s, 2H, Cp-Xyl-
ortho), 6.94 (s, 1H, Cp-Xyl-para), 7.36 (s, 1H, Xyl2-para), 7.43-
7.59 (m, 6H, Ph-meta, Ph-para), 7.74-7.84 (m, 4H, Ph-ortho), 8.12
(d, 2H, J ) 12.4 Hz, Xyl2-ortho). 13C{1H} NMR (100 MHz): δ
21.19 (2C, Xyl1-CH3), 21.50 (2C, Cp-Xyl-CH3), 21.67 (2C, Xyl2-
CH3), 70.01-70.21 (m, Cp-C4), 71.84 (br s, 5C, Cp′), 72.84 (d, J
) 3.8 Hz, Cp-C5), 84.59 (dd, J ) 16.8 Hz, J ) 2.3 Hz, Cp-Cq),
126.96 (d, J ) 54.75 Hz, P-Ph-ipso), 127.62 (d, 2C, J ) 10.7 Hz,
P-Ph1-meta), 127.67 (2C, Cp-Xyl-ortho), 129.12 (d, 2C, J ) 10.7
Hz, P-Ph2-meta), 129.14 (d, J ) 67.3 Hz, P-Xyl-ipso), 129.27 (d,
2C, J ) 11.5 Hz, Xyl1-ortho), 129.32 (Cp-Xyl-para), 130.43 (d, J
) 52.51 Hz, P-Ph-ipso), 131.29 (d, 2C, J ) 3.0 Hz, P-Ph1-para +
P-Ph2-para), 132.03 (d, J ) 58.1 Hz, P-Xyl-ipso), 132.14 (d, J )
3.0 Hz, Xyl1-para), 133.76 (d, 2C, J ) 10.7 Hz, P-Ph1-ortho),
134.36 (d, J ) 3.0 Hz, Xyl2-para), 134.40 (d, 2C, J ) 11.5 Hz,
P-Ph2-ortho), 134.71 (d, 2C, J ) 10.7 Hz, Xyl2-ortho), 135.29 (Cp-
Xyl-ipso), 137.73 (d, 2C, J ) 12.0 Hz, P-Xyl-Cq), 137.83 (2C,
Cp-Xyl-Cq), 138.66 (d, 2C, J ) 12.0 Hz, P-Xyl-Cq). P-Ph1 and
P-Ph2 as well as P-Ph-ipso and P-Xyl-ipso interchangeable. CH2
and two Cp-Cq signals not observed. 31P{1H} NMR (162 MHz):
δ +22.79 (d, J ) 7.9 Hz), +30.34 (d, J ) 7.9 Hz).
(23) Bruker programs: SMART, version 5.629; SAINT, version 6.45;
SADABS, version 2.10; XPREP, version 6.1; SHELXTL, version 6.14; Bruker
AXS Inc.: Madison, WI, 2003.
(24) Sheldrick, G. M. SHELX97: Program System for Crystal Structure
Determination; University of Go¨ttingen: Go¨ttingen, Germany, 1997.

3540 Organometallics, Vol. 26, No. 14, 2007
Wang et al.
Et2O) the solvents were squeezed with the program PLATON.25
For [PdCl2(3)]‚2C2H4Cl2 the cyclopentadienyl and benzene rings
were idealized.
details in CIF format for compounds [PdCl2(3)]‚2C2H4Cl2,
[PdCl2(9c)]‚(CH2Cl2, C2H5OH), [PdCl2(10c)]‚solv, and [PdCl2(11c)]‚
CH2Cl2. This material is available free of charge via the Internet at
http://pubs.acs.org.
Acknowledgment. This work was kindly supported by
Solvias AG, Basel.
OM700405A
Supporting Information Available: Details on the synthesis
(25) Spek, A. L. PLATON: A Multipurpose Crystallographic Tool;
of all ligands as well as complete crystallographic data and technical
University of Utrecht: Utrecht, The Netherlands, 2005.