Phosphorus-chiral diphosphines as ligands in hydroformylation. An investigation on the influence of electronic effects in catalysis

Article abstract of DOI:10.1021/om0004489

The phosphorus-chiral diphosphine 1,1′-bis(1-naphthylphenylphosphino)ferrocene (1a) and its new electronically modified derivatives 1b-d bearing methoxy and/or trifluoromethyl groups in para positions of the phenyl rings were investigated as ligands in rhodium-catalyzed (asymmetric) hydroformylation. Depending on ligand basicity, high-pressure NMR and IR characterization of the respective (diphosphine) rhodium dicarbonyl hydride precursor complexes revealed subtle differences in the occupation of bis-equatorial (ee) and equatorialapical (ea) coordination geometries. The high ee:ea ratio of the four complexes contrasted with the clear ea preference observed for the related achiral compound dppf (1,1′-bis-(diphenylphosphino)ferrocene). In the hydroformylation of styrene the best result (50% ee) was obtained by employing the best π-acceptor ligand 1c, incorporating two p-trifluoromethyl substituents. Substrate electronic variations using 4-methoxystyrene and 4-chlorostyrene showed a pronounced influence on turnover frequencies, branched/linear aldehyde product ratios, and enantiodiscrimation, whereas in the hydroformylation of 1-octene ligand electronic perturbations did affect only the rate, but not the selectivity of the reaction.

Full text of DOI:10.1021/om0004489

4596
Organometallics 2000, 19, 4596-4607
P h osp h or u s-Ch ir a l Dip h osp h in es a s Liga n d s in
Hyd r ofor m yla tion . An In vestiga tion on th e In flu en ce of
Electr on ic Effects in Ca ta lysis
Ulrike Nettekoven,† Paul C. J . Kamer,† Michael Widhalm,‡ and
Piet W. N. M. van Leeuwen*,†
Institute of Molecular Chemistry, University of Amsterdam, Nieuwe Achtergracht 166,
1018 WV Amsterdam, The Netherlands, and Institute of Organic Chemistry,
University of Vienna, Wa¨hringer Strasse 38, 1090 Vienna, Austria
Received May 30, 2000
The phosphorus-chiral diphosphine 1,1′-bis(1-naphthylphenylphosphino)ferrocene (1a ) and
its new electronically modified derivatives 1b-d bearing methoxy and/or trifluoromethyl
groups in para positions of the phenyl rings were investigated as ligands in rhodium-catalyzed
(asymmetric) hydroformylation. Depending on ligand basicity, high-pressure NMR and IR
characterization of the respective (diphosphine) rhodium dicarbonyl hydride precursor
complexes revealed subtle differences in the occupation of bis-equatorial (ee) and equatorial-
apical (ea) coordination geometries. The high ee:ea ratio of the four complexes contrasted
with the clear ea preference observed for the related achiral compound dppf (1,1′-bis-
(diphenylphosphino)ferrocene). In the hydroformylation of styrene the best result (50% ee)
was obtained by employing the best π-acceptor ligand 1c, incorporating two p-trifluoromethyl
substituents. Substrate electronic variations using 4-methoxystyrene and 4-chlorostyrene
showed a pronounced influence on turnover frequencies, branched/linear aldehyde product
ratios, and enantiodiscrimination, whereas in the hydroformylation of 1-octene ligand
electronic perturbations did affect only the rate, but not the selectivity of the reaction.
In tr od u ction
in asymmetric hydroformylation were obtained using
platinum/stannous chloride catalysts,3 these systems
tend to suffer from low reaction rates, moderate
branched/linear ratios, and product racemization. With
respect to the mentioned criteria, rhodium(I)-based
catalysts are considered superior, but catalyst stability
and especially enantiodiscrimination are features that
deserve improvement. The most successful ligands
reported to date for rhodium-catalyzed asymmetric
hydroformylation are bulky diphosphites based on
enantiopure pentane-2,4-diols4 on one hand and phos-
phine-phosphites bearing atropisomeric binaphthyl
moieties on the other hand (Binaphos, 96% ee for the
hydroformylation of styrene).5
Previously, a new class of rigid diphosphine ligands
was developed in our group, based on 9,10-dimethyl-
xanthene or related backbones and, consequently, bear-
ing a large natural bite angle as a structural charac-
teristic.6 The coordination mode of these Xantphos-type
ligands in rhodium(I) dicarbonyl hydride precursor
complexes was shown to be preferentially bis-equatorial,
The catalytic hydroformylation of alkenes ranks
among the most widely industrially applied homoge-
neous processes. Its asymmetric variant has not yet
reached that level of sophistication; nevertheless, or just
for that reason, many efforts have been directed toward
that goal during the past few years. Attractive features
of the reaction, such as atom economy and potential
catalyst recyclability as well as the need for optically
active aldehyde products as intermediates for fine
chemicals, are counterbalanced by several drawbacks,
among which number unsatisfactory branched/linear
ratios, modest reactivities and enantioselectivities, side
reactions, and insufficient catalyst stability.1
Useful asymmetric systems developed to date include
catalysts based on platinum(II) or rhodium(I), in which
both of the metal centers have been modified by mono-
or bidentate phosphorus ligands.2 Although some of the
highest enantiomeric excess (ee) values ever reported
† University of Amsterdam.
‡ University of Vienna.
(3) (a) Stille, J . K.; Su, H.; Brechot, P.; Parrinello, P.; Hegedus, L.
S. Organometallics 1991, 10, 1183. (b) Csere´pi-Szu¨cs, S.; Bakos, J . J .
Chem. Soc., Chem. Commun. 1997, 635.
(1) For reviews on hydroformylation see for example: (a) Tolman,
C. A.; Faller, J . W. In Homogeneous Catalysis with Metal Phosphine
Complexes; Pignolet, L. H., Ed.; Plenum Press: New York, 1983; p 81.
(b) Beller, M.; Cornils, B.; Frohning, C. D.; Kohlpaintner, C. W. J . Mol.
Catal. A: Chem. 1995, 104, 17. (c) Frohning, C. D.; Kohlpaintner, C.
W. In Applied Homogeneous Catalysis with Organometallic Com-
pounds; Cornils, B., Herrmann, W. A., Eds.; VCH: Weinheim, Ger-
many, 1996; Vol. 1, p 27.
(2) For reviews on asymmetric hydroformylation see for example:
(a) Agbossou, F.; Carpentier, J .-F.; Mortreux, A. Chem. Rev. 1995, 95,
2485. (b) Gladiali, S.; Bayon, J . C.; Claver, C. Tetrahedron: Asymmetry
1995, 6, 1453 and references therein.
(4) (a) Babin, J . E.; Whiteker, G. T. PCT Int. Appl. WO 93 03839,
1993; Chem. Abstr. 1993, 119, 159872h. (b) Buisman, G. J . H.; van
der Veen, L. A.; Klootwijk, A.; de Lange, W. G. J .; Kamer, P. C. J .; van
Leeuwen, P. W. N. M.; Vogt, D. Organometallics 1997, 16, 2929.
(5) (a) Sakai, N.; Mano, S.; Nozaki, K.; Takaya, H. J . Am. Chem.
Soc. 1993, 115, 7033. (b) Nozaki, K.; Sakai, N.; Nanno, T.; Horiuchi,
T.; Takaya, H. J . Am. Chem. Soc. 1997, 119, 4413 and references
therein.
(6) Kranenburg, M.; van der Burgt, Y. E. M.; Kamer, P. C. J .; van
Leeuwen, P. W. N. M. Organometallics 1995, 14, 3081.
10.1021/om0004489 CCC: $19.00 © 2000 American Chemical Society
Publication on Web 10/05/2000

P-Chiral Diphosphines as Ligands in Hydroformylation
Organometallics, Vol. 19, No. 22, 2000 4597
which, at first, was assumed to be an explanation for
the high selectivity toward linear aldehyde observed in
the hydroformylation of 1-octene. When electronically
modified Xantphos ligands were employed, the reactiv-
ity was (expectedly) found to depend on the electron
donor and acceptor capacities of the substituents at the
para position of the phenyl rings of the phosphines,7
whereas the earlier reported correlation8 between Ham-
mett σp constants and the selectivity toward formation
of the linear aldehyde did not apply.
tion is to proceed. Thus, a reduction of competitive
diastereomeric reaction pathways resulting therefrom
is believed to confer high levels of enantiodiscrimina-
tion.12
The ligands we chose for our study were 1,1′-bis(1-
naphthylphenylphosphino)ferrocene (1a ) and electroni-
cally modified derivatives thereof, bearing methoxy and
trifluoromethyl groups in para positions of the phenyl
rings. Consequently, steric alterations imposed by the
In contrast, Casey et al. reported on a distinct
influence of ligand electronic effects on the linear/
branched aldehyde product ratio in the hydroformyla-
tion of 1-hexene.9 They stated that electron-withdrawing
substituents on equatorially coordinated phosphines
slightly favor formation of linear aldehydes, whereas
electron-poor phosphines in the apical position are
supposed to support strongly the opposite regiochem-
istry. This hypothesis was substantiated by the synthe-
sis and application of electronically dissymmetric equa-
torial-apical coordinating ligands, which, due to the
preference of electron-deficient phosphines for the equa-
torial coordination site, showed indeed a higher linear/
branched ratio than their symmetric counterparts.10
Our interest was attracted by a possible extension of
these results to asymmetric hydroformylation and the
effects that might be encountered there. C2-symmetrical
diphosphines do not rank among the ligands inducing
excellent enantioselectivities in that reaction; to the best
of our knowledge, the highest ee values (60%) were
obtained using bdpp (2,4-bis(diphenylphosphino)pen-
tane).11 This ligand was shown to coordinate to trigonal-
bipyramidal rhodium(I) precursor complexes in a pre-
dominantly equatorial-apical fashion, which is ascribed
to the small natural bite angle of around 90°. In
contrast, the application of propeller-shaped diphos-
phine ligands displaying larger natural bite angles
(>100°) in an asymmetric reaction variant has not been
investigated in detail. To render this type of compound
promising, we reasoned that chirality at the coordinat-
ing phosphines might, due to their proximity to the
metal center, induce higher enantioselectivities than if
incorporated in the backbone. Furthermore, provided
that trigonal-bipyramidal intermediates are involved in
the selectivity-determining step of the catalytic cycle,
bis-equatorially coordinated structures of C2-type sym-
metry are more desirable, since they leave only one
equatorial coordination site accessible on which alkene
complexation and subsequent hydride migratory inser-
substituents should be negligible, allowing for an in-
vestigation of mainly electronic effects. Because of their
analogy to the achiral dppf (1,1′-bis(diphenylphosphino)-
ferrocene) ligand, bite angles of around 100° can be
expected for this type of diphosphine,13 which should
favor the formation of bis-equatorially coordinated
complexes. In this paper we report on the stereoselective
synthesis of three new phosphorus-chiral, electronically
varied ferrocenyldiphosphine ligands 1b-d and their
application in the asymmetric hydroformylation of
styrene and derivatives. To evaluate ligand perfor-
mances in comparative context with respect to linear/
branched ratios, hydroformylation experiments employ-
ing 1-octene as the model substrate were conducted as
well. High-pressure IR and NMR measurements are
described, which enabled the characterization of cata-
lytic intermediates. Their implications for the reaction
cycle, the regiochemistry, and enantioselectivities will
be discussed.
(7) van der Veen, L. A.; Boele, M. D. K.; Bregman, F. R.; Kamer, P.
C. J .; van Leeuwen, P. W. N. M.; Goubitz, K.; Fraanje, J .; Schenk, H.;
Bo, C. J . Am. Chem. Soc. 1998, 120, 11616.
(8) (a) Unruh, J . D.; Christenson, J . R. J . Mol. Catal. 1982, 14, 19.
(b) Moser, W. R.; Papile, C. J .; Brannon, D. A.; Duwell, R. A.;
Weininger, S. J . J . Mol. Catal. 1987, 41, 271.
(9) Casey, C. P.; Lin Paulsen, E.; Beuttenmueller, E. W.; Proft, B.
R.; Petrovich, L. M.; Matter, B. R.; Powell, D. R. J . Am. Chem. Soc.
1997, 119, 11817.
(10) Casey, C. P.; Lin Paulsen, E.; Beuttenmueller, E. W.; Proft, B.
R.; Matter, B. R.; Powell, D. R. J . Am. Chem. Soc. 1999, 121, 63.
(11) (a) Masdeu-Bulto´, A. M.: Orejo´n, A.; Castellanos, A.; Castillo´n,
S.; Claver, C. Tetrahedron: Asymmetry 1996, 7, 1829. (b) Dieguez, M.;
Pereira, M. M.; Masdeu-Bulto´, A. M.; Claver, C.; Bayon, J . C. J . Mol.
Catal. A 1999, 143, 111.
Resu lts a n d Discu ssion
Liga n d Syn th esis. Recently we described the prepa-
ration of a series of phosphorus-chiral enantiopure
arylphenylferrocenyldiphosphines.14 On employing the
approach of J uge´ et al.,15 characterized by attachment
of an optically active auxiliary, ephedrine, to a borane-
protected phosphorus center, and its stepwise removal
(14) (a) Nettekoven, U.; Widhalm, M.; Kamer, P. C. J .; van Leeuwen,
P. W. N. M. Tetrahedron: Asymmetry 1997, 8, 3185. (b) Nettekoven,
U.; Kamer, P. C. J .; van Leeuwen, P. W. N. M.; Widhalm, M.; Spek, A.
L.; Lutz, M. J . Org. Chem. 1999, 64, 3996.
(15) J uge´, S.; Stephan, M.; Laffitte, J . A.; Geneˆt, J . P. Tetrahedron
Lett. 1990, 31, 6357.
(12) Buisman, G. J . H.; Vos, E. J .; Kamer, P. C. J .; van Leeuwen, P.
W. N. M. J . Chem. Soc., Dalton Trans. 1995, 409.
(13) For a compilation of crystal structure data of dppf complexes
see for example: Gan, K.-S.; Hor, T. S. A. In Ferrocenes; Togni, A.,
Hayashi, T., Eds.; VCH: Weinheim, Germany, 1995; Chapter 1.

4598 Organometallics, Vol. 19, No. 22, 2000
Nettekoven et al.
Sch em e 1
Sch em e 2
therefrom by nucleophilic substitutions, different methyl
arylphenylphosphinite boranes of defined absolute con-
figuration were obtained. Subsequently, we succeeded
in the stereoselective coupling of the latter onto 1,1′-
dilithioferrocene, which gave rise to the first phosphorus-
chiral, enantiopure dppf analogues.
To allow for the introduction of the para substituents
on the 1-naphthylphenylphosphino derivative 1a , a
slightly modified pathway had to be followed. Since the
reaction sequence did not permit easy incorporation of
electron-donating or -releasing groups at a final stage,
the condensation reaction with the auxiliary had to be
performed with the suitably substituted bis(diethylami-
no)phenylphosphines (Scheme 1).16 These were synthe-
sized by reaction of the respective Grignard reagents,
derived from p-(trifluoromethyl)- and p-methoxybro-
mobenzenes with bis(diethylamino)chlorophosphine. Sub-
sequently, condensation with (1S,2R)-ephedrine at el-
evated temperatures, followed by complexation of the
phosphorus centers using BH3-dimethyl sulfide gave
rise to single diastereomers of the respective oxazaphos-
pholidine borane complexes 2b,c.
For the unsubstituted phenyl derivative, the stereo-
chemical course of the several steps during the conden-
sation-substitution sequence has been confirmed by
crystal structure analysis.17 The presence of the p-
methoxy or p-trifluoromethyl groups was assumed not
to interfere severely during diastereo- and enantiodis-
crimination; therefore, descriptors for the asymmetri-
cally substituted phosphorus atoms were assigned in
agreement with the literature findings.
chromatography was achieved easily, thus permitting
the use of stereohomogeneous starting materials for the
next step.
Acidic methanolysis of the phosphorus amides was
performed according to the literature and afforded the
configurationally inverted phosphinite boranes 4b,c. A
considerably higher yield was obtained for the p-
methoxy-substituted derivative.18
In the subsequent reaction, 1,1′-dilithioferrocene served
as the nucleophile, stereoselectively replacing the meth-
oxy groups at phosphorus and giving rise to the respec-
tive diphosphine diborane complexes 5b,c. However,
deprotection of the crude products using nitrogen bases
such as diethylamine and morpholine proved successful
only for the diborane 5c. In the case of 5b, decomplex-
ation was accompanied by considerable epimerization
at phosphorus, as recognized by formation of the meso-
configured diphosphine. Repeated complexation of such
a product mixture and comparison of its NMR data with
those of the original diborane product indicated that
racemization had indeed occurred during deprotection
and did not result from unselective substitution. As an
alternative method, acidic decomplexation was em-
ployed.19 With this procedure no epimerization took
place and, after column chromatographic purification,
the diphosphine ligand 1b was obtained in enantiopure
Nucleophilic attack of 1-naphthyllithium proceeded
in analogy to the described procedure,14b delivering the
ring-opened phosphorus amide boranes 3b,c in good
yields. In comparison to the unsubstituted starting
material, however, the observed diastereoselectivities
were lower, resulting in 88-90% de of the products with
configurationally retained phosphorus stereocenters.
Nevertheless, separation of the isomers by column
(16) In principle, a reversed sequence comprising of condensation
of bis(diethylamino)(1-naphthyl)phosphine and ephedrine, followed by
nucleophilic attack of para-substituted phenyllithium, should also be
possible. Yet, purification of the bis(diethylamino)arylphosphine start-
ing material by distillation was expected to be troublesome; therefore,
this variant was not taken into further consideration.
(17) (a) J uge´, S.; Stephan, M.; Merde`s, R.; Geneˆt, J . P.; Halut-
Desportes, S. J . Chem. Soc., Chem. Commun. 1993, 531. (b) Reference
14b.
(18) Addition of more than 1.1 equiv of sulfuric acid gave no
improvement in the yield of the p-trifluoromethyl-substituted phos-
phinite-borane.
(19) McKinstry, L.; Livinghouse, T. Tetrahedron Lett. 1994, 35, 9319.

P-Chiral Diphosphines as Ligands in Hydroformylation
Sch em e 3
Organometallics, Vol. 19, No. 22, 2000 4599
form, as was diphosphine 1c after removal of borane
using diethylamine.
lation was performed using sec-butyllithium and the
resulting ferrocenyl nucleophile underwent reaction
with the p-methoxyphenyl-substituted methyl phos-
phinite-borane 4b. Although again amounts of mono-
substituted ferrocene derivative were observed, the
desired 1-bromo-1′-phosphinoferrocene-borane complex
6 was obtained as the main product and successfully
purified by column chromatography. Subsequent lithia-
tion of the latter by halogen-metal exchange employing
sec-butyllithium and coupling to phosphinite-borane 4c
gave rise to the ferrocenyldiphosphine-diborane com-
plex. Deprotection was achieved using trifluoromethane-
sulfonic acid and proceeded in a manner similar to that
described for compound 5b. Thus, the enantiomerically
pure C1-symmetrical ligand 1d was isolated as an
orange foam and constitutes one of the first examples
The synthesis of the electronically dissymmetric
ligand 1d was envisaged to proceed by stepwise intro-
duction of the phosphine moieties to the ferrocene
backbone (Scheme 2). In a first attempt, 1,1′-dibromo-
ferrocene was lithiated using 1 equiv of n-butyllithium20
and consecutively made to react with methyl 1-naph-
thyl-(4-(trifluoromethyl)phenyl)phosphinite-borane (4c).
This procedure, however, resulted in a mixture of
several products, including 1-bromo-1′-phosphinofer-
rocene-borane as well as heteroannularly unsubsti-
tuted ferrocenylphosphine-borane, both of which proved
to be largely inseparable by low-pressure column chro-
matographic techniques. In a second approach, meta-
(20) (a) Lai, L.-L.; Dong, T.-Y. J . Chem. Soc., Chem. Commun. 1994,
2347. (b) Butler, I. R.; Davies, R. L. Synthesis 1996, 1350.

4600 Organometallics, Vol. 19, No. 22, 2000
Ta ble 1. Selected NMR Da ta for (d ip h osp h in e)Rh (CO)2H Com p lexes
Nettekoven et al.
complex
δ(1H)/ppm
1J H-Rh/Hz
2J H-P(av)/Hz
δ(31P)/ppm
1J P(av)-Rh/Hz
ee:ea
(1a )Rh(CO)2H
(1b)Rh(CO)2H
(1c)Rh(CO)2H
(1d )Rh(CO)2H
-8.87
-8.83
-9.04
-8.81
5.2
5.4
5.1
4.0
6.9
7.2
5.0
3.4 (P1)
9.3 (P2)
41.5
23.75
12.16
23.85
17.45
19.33
27.79
135.4
134.6
137.7
136.1
84:16
84:16
88:12
71:20:9a
22:78
(dppf)Rh(CO)2H
-9.09
9.5
121.9
a
Occupation of ee (ee-1 + ee-2), ea-1, and ea-2 coordination modes, respectively (Figure 1).
of a stereogenic diphosphine bearing two differently
substituted phosphorus centers.
complex or its iridium analogue.26 For reasons of
comparison, we were also interested in the behavior of
the achiral ligand dppf, from which diphosphines 1a -d
are derived of. To this aim, experiments with dppf were
included.
After suitable deprotection of diboranes 5b-d , no
signals indicating partial epimerization to the meso
(type) compound were detected by NMR measurements.
Nevertheless, the enantiomeric purity was checked by
applying the NMR shift reagent N-(3,5-dinitrobenzoyl)-
1-phenylethylamine21 to diphosphine dioxides, derived
from the respective ligands 1b-d . Within the NMR
integration errors of (2%, no splitting of signals due to
the presence of other diastereomeric adducts was ob-
served.
High -P r essu r e NMR Mea su r em en ts. A simplified
mechanistic cycle for the generally accepted dissociative
pathway of the diphosphine modified rhodium-catalyzed
hydroformylation is depicted in Scheme 3.22 The trigo-
nal-bipyramidal rhodium diphosphine dicarbonyl hy-
dride may be regarded as a precatalyst, from which CO
decoordination is to occur, followed by alkene complex-
ation to give again a five-coordinated intermediate.
Under conditions of low pressure and temperature,
consecutive hydride migratory insertion is assumed to
be rate- and selectivity-determining,23 as long as sub-
strate isomerization indicating â-hydride elimination
from the rhodium alkyl intermediates remains negli-
gible.24 Since they were the only observable monomeric
species under catalytic conditions, the coordination
mode of bidentate diphosphine ligands in rhodium
dicarbonyl hydride complexes was found to influence the
regiochemical outcome of the hydroformylation reaction
(vide supra).
Concerning the issue of stereoselectivity, the situation
is ambiguous. High enantioselectivities were obtained
with bis-equatorially coordinating diphosphites4 as well
as equatorial-apically coordinating phosphine-phos-
phite5 and, to a lesser degree, diphosphine chelates.11
If the complexation mode of diphosphines in the trigo-
nal-bipyramidal precursor complexes or alkene adducts
was of any relevance to stereodiscrimination, subtle
changes in the equilibrium population of the two
isomers by ligand electronic perturbations should result
in different ee values.
Therefore, the solution structures of rhodium dicar-
bonyl hydrides bearing dppf-analogues 1a -d were
determined by high-pressure NMR techniques. Because
of the fast interconversion between bis-equatorial (ee)
and equatorial-apical (ea) coordination fashions of
chelating diphosphines, presumably proceeding via
turnstile or Berry-type (pseudo)rotations,25 averaged
values for proton-phosphorus and phosphorus-rhodi-
um coupling constants are generally obtained. At low
temperatures the equilibrium may be frozen out, per-
mitting the determination of individual constants for
the coordination isomers and, subsequently, calculation
of the ee:ea ratio for a given (diphosphine)Rh(CO)2H
Reaction of ligands 1a -d with Rh(CO)2acac under 20
bar of syngas (H2/CO ) 1) at 40 °C resulted in exclusive
formation of the respective diphosphine rhodium dicar-
bonyl hydride compounds. This was evidenced by the
appearance of a triplet of doublets for the hydride signal
in 1H NMR and a doublet in the 31P NMR (with
exception of complexes incorporating ligand 1d ; vide
infra). Unfortunately, for all of the complexes investi-
gated, the dynamic equilibrium could not be frozen out
completely at the lowest available temperature of -110
°C. Thus, equilibrium populations of bis-equatorial and
equatorial-apical coordination modes were calculated
2
2
assuming values of J H-P(av) ) (2 Hz and J H-P(av)
)
110 Hz27 for cis and trans hydride phosphorus coupling
constants, respectively (Table 1).
The obvious trend observed for complexes of ligands
1a -d is the dependence of the ee:ea distribution on
ligand basicity. In agreement with previously reported
work,7 an increase in the electron-donating capacity
within the ligand series was accompanied by a decrease
in the ee:ea ratio. Among the complexes ligated by C2-
symmetrical diphosphines, the highest preference for
the bis-equatorial coordination mode was, expectedly,
measured for (1c)Rh(CO)2H, displaying a value of 88:
12. For the complexes incorporating ligands 1a ,b, the
ratios were somewhat lower but, interestingly, almost
equal. The electronic effect of the p-methoxy residues
was found to be rather small, which should be mirrored
in catalysis results (vide infra).
In the case of the complex (1d )Rh(CO)2H, bearing the
C1-symmetrical ligand, the 31P{1H} spectrum showed
two doublets of doublets, originating from direct rhod-
1
ium-phosphorus coupling J P(av)-Rh as well as geminal
2
coupling J P1(av)-P2(av) of the nonequivalent phosphines
(Figure 1). In the 1H spectrum, a very broad doublet
(21) Dunach, E.; Kagan, H. B. Tetrahedron Lett. 1985, 26, 2649.
(22) Evans, D.; Osborn, J . A.; Wilkinson, G. J . Chem. Soc. A 1968,
3133.
(23) Consiglio, G.; Pino, P. Top. Curr. Chem. 1982, 105, 77.
(24) (a) Lazzaroni, R.; Settambolo, R.; Raffaelli, A.; Pucci, S.; Vitulli,
G. J . Organomet. Chem. 1988, 339, 357. (b) Lazzaroni, R.; Raffaelli,
A.; Settambolo, R.; Bertozzi, S.; Vitulli, G. J . Mol. Catal. 1989, 50, 1.
(25) (a) Buisman, G. J . H.; van der Veen, L. A.; Kamer, P. C. J .;
van Leeuwen, P. W. N. M. Organometallics 1997, 16, 5681. (b)
Castellanos-Pa´ez, A.; Castillo´n, S.; Claver, C.; van Leeuwen, P. W. N.
M.; de Lange, W. G. J . Organometallics 1998, 17, 2543.
(26) (a) Yagupsky, G.; Wilkinson, G. J . Chem. Soc. A 1969, 725. (b)
References 9 and 10.
(27) A value of 110 Hz was observed for trans-2J H-P(av) in the
rhodium dicarbonyl hydride complex of ligand 1d at 193 K. This is in
good agreement with data reported for comparably wide bite angle
diphosphine rhodium or iridium compounds; see, for example refs 7
and 9. The common assumption of 2J H-P(av) ) (2 Hz, however, may be
subject to uncertainty.

P-Chiral Diphosphines as Ligands in Hydroformylation
Organometallics, Vol. 19, No. 22, 2000 4601
Ch a r t 1
F igu r e 1. Possible coordination isomers of trigonal-
bipyramidal rhodium dicarbonyl hydrides bearing P-chiral
C1-symmetrical diphosphines.
was observed at δ -8.94 ppm. Broad-band phosphorus
decoupling enabled identification of the rhodium hydride
1
coupling constant J H-Rh and, consequently, determi-
Sch em e 4
nation of 2J H-P1(av) and 2J H-P2(av). Considering the known
preference of electron-poor donor ligands for equatorial
coordination sites, the lower value of the averaged
phosphorus hydride coupling constants of 3.4 Hz was
assigned to the phosphine bearing the trifluoromethyl
residue (P1), whereas the constant of 9.2 Hz implied an
increased occupation of the apical coordination site and
was therefore attributed to the methoxy-substituted
phosphorus donor (P2). Calculation of the equilibrium
concentrations gave a distribution of 71:20:9 between
(not further discerned) ee, ea-1, and ea-2 coordination
modes, respectively. Thus, the electronically dissym-
metric diphosphine 1d induced an ee:ea ratio in the
dicarbonyl hydride complex, which ranks clearly below
the values observed for either of the symmetrical donor
ligands.
phine)-coordinated species (Scheme 4) and the men-
tioned ee:ea equilibration.
Interestingly, the achiral dppf chelate displayed a
different coordination behavior. The NMR spectra in-
dicated that under the conditions applied (20 bar of
syngas, 1/1 CO/H2, [Rh] ) 8 mM) only when equimolar
amounts of dppf and Rh(CO)2acac were employed was
the dicarbonyl hydride species formed as the main
product. At higher ligand concentrations (dppf/Rh-
(CO)2acac ) 1.5-2; [dppf] ) 12-16 mM), appearance
of a broad doublet and, additionally, a broad singlet at
δ 21.19 ppm indicated the prevailing species to be a
dinuclear complex in which three phosphines are bound
to each rhodium center by means of a bridging dppf
molecule (Chart 1).28,29 Such a structure had already
been invoked by Unruh and Christenson in order to
account for the observed increase of the linear/branched
aldehyde ratio and the decrease in reaction rate utilizing
1.5 equiv or more of ligand per mole of rhodium in the
hydroformylation of 1-hexene.8a
A clearly lower ee:ea ratio was calculated for the
chelating dppf ligand than for ligands 1a -d in dicar-
1
bonyl hydride complexes. The smaller J P(av)-Rh coupling
constant likewise suggested enhanced occupation of the
apical coordination site,31 whereas an expected small
1J P(bridge)-Rh coupling might not be resolved in the broad
signal at δ 21.19 ppm. Consequently, tris(phosphine)
carbonyl rhodium hydride species appeared to accom-
modate the bridging or monocoordinating dppf ligand
in the equatorial position. The relatively large ea
contribution to the averaged signals of the chelating
donor seemed surprising, since spatial congestion in
both complexes A and B is better released by tris-
equatorial phosphine coordination. Nevertheless, the
steric accessibility of the phosphines in dppf might, due
to the presence of one five-membered ring, even surpass
that of triphenylphosphine, which in combination with
free rotation around the Cp-Fe-Cp axis could explain
the enhanced formation of bridged ea chelated species.
When these observations are compared with the high-
pressure NMR structures incorporating diphosphines
1a -d , replacement of a phenyl by a 1-naphthyl sub-
stituent seemed to introduce sufficient steric hindrance
to prohibit formation of mono- or dinuclear tris(phos-
phine)-ligated species. Consequently, higher reactivities
The chemical shift values are within the range
previously observed for comparable trigonal-bipyrami-
dal monocarbonyl rhodium hydrides;30 the absence of
additional distinct coupling patterns in 1H as well as
31P{1H} NMR can be ascribed to a combination of rapid
exchange between bridging and mononuclear tris(phos-
(28) At dppf/rhodium ratios of up to 1.5, no signals stemming from
noncoordinating phosphines were observed.
(29) Related observations of bridging phosphines in hydroformyla-
tion have been reported: (a) Reference 25b. (b) Freixa, Z.; Pereira, M.
M.; Pais, A. A. C. C.; Bayo´n, J . C. J . Chem. Soc., Dalton Trans. 1999,
3245.
(30) Hughes, O. R.; Young, D. A. J . Am. Chem. Soc. 1981, 103, 6636.
(31) (a) Sanger, A. R. J . Chem. Soc., Dalton Trans. 1977, 120. (b)
Garrou, P. E.; Curtis, J . L. S.; Hartwell, G. E. Inorg. Chem. 1976, 15,
3094. (c) Meakin, P.; J esson, J . P. J . Am. Chem. Soc. 1973, 95, 7272.

4602 Organometallics, Vol. 19, No. 22, 2000
Nettekoven et al.
Ta ble 2. Selected IR Da ta (cm -1) for
(d ip h osp h in e)Rh (CO)2H Com p lexes
complex
ν1
ν2
ν3
ν4
(1a )Rh(CO)2H
(1b)Rh(CO)2H
(1c)Rh(CO)2H
(1d )Rh(CO)2H
(dppf)Rh(CO)2H
2041 (m) 1996 (m) 1991 (s)
2037 (m) 1993 (m) 1988 (s)
2043 (m) 2001 (m) 1996 (s)
1955 (m)
1952 (m)
1961 (m)
1957 (s)
2040 (m) 1997 (s)
2040 (w) 1996 (s)
1991 (s)
1984 (w) 1955 (s)
and branched aldehyde selectivities were expected from
catalysts derived from (1a -d ) rhodium dicarbonyl hy-
drides.
High -P r essu r e IR Mea su r em en ts. The presence of
a dynamic equilibrium between ee and ea complexation
modes in (1a -e)Rh(CO)2H compounds was corroborated
by high-pressure IR spectroscopy. Each of the spectra
of the respective complexes showed four absorption
bands in the region between 2100 and 1900 cm-1, which
were attributed to carbonyl vibrations. In analogy to
literature findings, the lowest frequency band ν1 as well
as ν3 were assigned to the ee complex isomer, since upon
H/D exchange only these bands were found to display a
characteristic frequency shift in comparable diphos-
phine rhodium dicarbonyl hydrides.7
For the complexes bearing the structurally related
ligands 1a -d , a characteristic dependence of the vibra-
tion frequencies on ligand basicity was evidenced. With
increasing electron acceptor properties of the phosphine
a shift to higher wavenumbers was observed for all four
absorption bands. 1a -c-chelated complexes showed
comparably strong ν1 and ν3 absorptions, whereas the
ν2 vibrations (partly hidden under ν3) and the ν4 bands
appeared to be relatively weak. This trend was espe-
cially pronounced for the complex incorporating the bis-
(p-(trifluoromethyl)phenyl)-substituted diphosphine 1c,
indicating this compound to exhibit the highest ee
preference (Table 2). In contrast, the complex modified
by ligand 1d displayed a strong ν2 vibration, clearly
distinguishable from the ν3 absorption band. Moreover,
the intensity of the ν4 vibration surpassed the otherwise
strongest ν3 absorption. Both observations conferred a
higher equilibrium concentration of the ea coordination
isomer in comparison to complexes bearing ligands
1a -c (Figure 2).
F igu r e 2. High-pressure IR spectra for diphosphine
rhodium dicarbonyl hydride complexes (carbonyl region):
(1) (1c)Rh(CO)2H; (2) (d p p f)Rh(CO)2H; (3) (1a )Rh(CO)2H;
(4) (1b)Rh(CO)2H; (5) (1d )Rh(CO)2H.
Sch em e 5
Further evidence for this behavior was provided by
the high-pressure IR spectrum of the rhodium dicarbo-
nyl hydride ligated by dppf. NMR measurements sug-
gested a pronounced contribution of the ea chelated
trigonal-bipyramidal complex geometry to the equilib-
rium distribution and this was supported by the ob-
tained IR data. Strong ν2 and ν4 vibrations but weak
absorption bands of ν1 and ν3 are characteristics of a
preferential ea complexation mode, which to a lesser
degree were also recognizable in the spectra of the (1d )-
Rh(CO)2H complex. Because of the low ligand and
rhodium concentrations employed in IR experiments,
the formation of the aforementioned tris(phosphine)
carbonyl rhodium hydride species was less likely. A
single CO resonance at about 1930 cm-1, reported for
mixtures of [HRh(CO)(phosphine)1.5]2 and [HRh(CO)-
(phosphine)3] with CO in a trans relationship to the
hydride, was not observed.30 Thus, the IR spectroscopic
measurements on the respective (1a -d )Rh(CO)2H com-
plexes were found to be in good agreement with the
high-pressure NMR data, substantiating the ee:ea equi-
librium distribution determined by the latter method.
Hyd r ofor m yla tion Resu lts. Hydroformylation of
styrene was carried out in toluene at a pressure of 20
bar of syngas (1/1 CO/H2). In the first instance, the
optimum reaction conditions were determined in experi-
ments employing ligand 1a at variable temperatures,
ligand/rhodium ratios, and substrate concentrations
(Scheme 5, Table 3).
To summarize, rather low reactivities and branched/
linear ratios were obtained, indicating considerable
steric congestion in the active catalyst species. ee values
turned out to be moderate for the catalytic system
investigated. Expectedly, an increase in enantioselec-
tivities was observed on lowering the temperature; the
limit for the reaction to proceed at acceptable rates,
however, was above 30 °C. Equimolar amounts of
rhodium precursor and ligand did not seem to give
quantitative conversion to the diphosphine rhodium
dicarbonyl hydride; the resulting low ee value can be
ascribed to the presence of additional unsaturated

P-Chiral Diphosphines as Ligands in Hydroformylation
Organometallics, Vol. 19, No. 22, 2000 4603
Ta ble 3. Hyd r ofor m yla tion of Styr en e (R ) P h )
Usin g (S,S)-1a /Rh (CO)2a ca ca
decrease in reaction rates was observed for the latter;
nevertheless, the enantioselectivities proved to be equal.
The presence of the p-trifluoromethyl groups of ligand
1c in the catalyst affected turnover frequencies as well
as optical induction. The first were found to be almost
twice as high in comparison to reactions run with ligand
1a and enantiodiscrimination experienced an increase
of 4% to give a maximum of 50% ee. In contrast,
employment of the C1-symmetrical ligand 1d as a
catalyst component had a deleterious impact on stereo-
selectivity. While the presence of only one trifluoro-
methyl group seemed sufficient to promote a comparably
high reaction rate, the ee values dropped to 41%.
To investigate combined influences of substrate and
ligand electronic effects on the hydroformylation out-
come, we also performed reactions employing p-chloro-
and p-methoxystyrene as substrates (Scheme 5). Since
coordination of (substituted) alkenes presumably exerts
a strong influence on charge distribution in catalytic
species as compared to effects imposed by (electronically
modified) phosphine ligation, different mechanistic path-
ways might apply for these substrates. A relation
between, on one hand, reactivity and aldehyde selectivi-
ties and, on the other hand, Hammett σp constants or ø
values is well-established, whereby electron-withdraw-
ing groups in para positions of a styrene substrate were
found to promote higher turnover frequencies and
branched/linear ratios.32 Thus, hydroformylation of
4-methoxystyrene employing ligands 1a -d as catalyst
modifiers necessitated increased reaction temperatures
of 60 °C; results are included in Table 4.
1a /
%
b:l
% eed
Rh(CO)2acac T/°C conversn ratiob TOFc (abs confign)
1.1
2.2
3
2.2
2.2
2.2e
40
40
40
60
30
40
28
25
25
85
12
21
1.7
1.7
1.7
1.5
1.9
1.7
5
7
7
43
1
10
21 (S)
46 (S)
46 (S)
41 (S)
49 (S)
32 (S)
a
Reaction conditions: CO/H2 ) 1; pressure 20 bar, styrene/
b
rhodium ) 500; [Rh] ) 2.30 mM in toluene. Branched/linear
product ratio. c Turnover frequency ) (mol of aldehyde) (mol of
Rh)-1 h-1
.
Determined by chiral GC analysis of the corresponding
d
alcohol. e Styrene/rhodium ) 1000.
Ta ble 4. Hyd r ofor m yla tion of (Su bstitu ted )
Styr en es Usin g Differ en t d ip h osp h in e/
Rh (CO)2a ca c Sp eciesa
substrate
R
b:l
T/°C ratiob TOFc (abs confign)
% eed
ligand
(S,S)-1a
(S,S)-1b
(S,S)-1c
(S,S)-1d
dppf
(S,S)-1a
(S,S)-1b
(S,S)-1c
(S,S)-1d
dppf
(S,S)-1a
(S,S)-1b
(S,S)-1c
(S,S)-1d
dppf
Ph
Ph
Ph
Ph
40
40
40
40
40
60
60
60
60
60
40
40
40
40
40
1.7
1.7
1.5
1.7
2.4
1.2
1.2
1.1
1.1
1.6
2.1
2.4
2.0
2.3
2.6
7e
6e
46 (S)
46 (S)
50 (S)
41 (S)
12e
11e
11e
42f
24f
71f
55f
30f
21g
15g
30g
22g
3h
Ph
4-MeOC6H4
4-MeOC6H4
4-MeOC6H4
4-MeOC6H4
4-MeOC6H4
4-ClC6H4
4-ClC6H4
4-ClC6H4
4-ClC6H4
4-ClC6H4
40 (S)
51 (S)
36 (S)
29 (S)
37 (S)
22 (S)
22 (S)
16 (S)
The trend in turnover frequencies observed for sty-
rene was, expectedly, continued in a pronounced fash-
ion, with the less basic ligands promoting faster reaction
rates. In contrast, optical inductions were affected in a
dissimilar manner. While with respective complexes of
diphosphines 1a ,c,d enantioselectivities for 2-(4-meth-
oxyphenyl)propanal were found to be lower than in the
case of the unsubstituted product, the catalyst compris-
ing ligand 1b effected a noteworthy increase to 51% ee
in the hydroformylation of 4-methoxystyrene. Obviously,
low reaction rates and a favorable combination of
electronic parameters positively influenced enantiodis-
crimination.
A different picture was obtained when utilizing
4-chlorostyrene as the substrate. As anticipated,
branched/linear ratios were found to be higher than
those for the other two substrates investigated, whereby
reactivity and ligand basicity obeyed the correlation
already described.32 The low turnover frequencies ob-
tained on employing dppf-modified complexes were
surprising. Considering a combination of factors such
as reduced σ-donor capacity of the alkene and sterical
overcrowding of a possibly bridged dinuclear rhodium
precursor, this result could point toward alkene coor-
dination as the crucial step in the mechanism. Addition-
ally, catalyst deactivation might occur. The overall
higher reaction rates observed for catalysts of ligands
1a -d were accompanied by a decrease in optical yields,
if compared to the other substrates. The electronically
unmodified 1a complex seemed to be best suited for
asymmetric conversion of the electron-poor substrate,
whereas enantioselectivities effected by the other cata-
a
Reaction conditions: CO/H2 ) 1; pressure 20 bar, 2.2/1 ligand/
rhodium, 500/1 styrene/rhodium; [Rh] ) 2.30 mM in toluene.
Branched/linear product ratio. c Turnover frequency ) (mol of
b
aldehyde) (mol of Rh)-1 h-1
.
Determined by chiral GC analysis
d
of the corresponding alcohol. e Reactions were stopped at ∼30%
conversion. f Reactions were stopped at ∼60% conversion. g Reac-
h
tions were stopped at ∼30% conversion. Reactions were stopped
at ∼10% conversion.
catalytically active species. In contrast, the use of 2.2
equiv of 1a sufficed for complete generation of the actual
catalyst, since when 3 equiv were employed, no further
improvement was obtained. A higher substrate/rhodium
ratio was found to have a deleterious effect on enantio-
discrimination, which might be due to reduced catalyst
stability at higher substrate concentrations.
Utilizing 2.2 equiv of diphosphine and a substrate/
rhodium ratio of 500, hydroformylations with diphos-
phines 1a -d as well as dppf-modified catalysts were
carried out at 40 °C, the results of which are sum-
marized in Table 4. In the presence of predominantly
equatorial-apically coordinated catalysts, the use of
styrene as a substrate is generally accompanied by
relatively high branched/linear aldehyde ratios, a fact
that is postulated to originate from an η3-allyl type
stabilization involving the phenyl ring of the branched
rhodium alkyl species.1a However, a prevailingly bis-
equatorial coordination mode of the diphosphine ligand
employed is associated with an increased preference for
formation of the linear aldehyde,9 which, indeed, is
observed for ligands 1a -d . As anticipated after inspec-
tion of the solution structures, the difference in perfor-
mance of catalysts incorporating diphosphine 1a and its
p-methoxyphenyl derivative 1b was small. A slight
(32) Hayashi, T.; Tanaka, M.; Ogata, I. J . Mol. Catal. 1981, 13, 323.

4604 Organometallics, Vol. 19, No. 22, 2000
Nettekoven et al.
Ta ble 5. Hyd r ofor m yla tion of 1-Octen e (R )
(CH3(CH2)5) a t 80 °C Usin g Differ en t d ip h osp h in es/
Rh (CO)2a ca c Sp eciesa
An interesting feature of these catalytic results was
found within the achieved turnover frequencies. While
reactivities using the first three ligands as catalyst
components expectedly correlated with phosphine donor
capacities, the high activity of catalysts incorporating
ligand 1d represented an unanticipated performance.
A possible explanation may be that the relatively large
bite angles and steric demands of ligands 1a -d promote
faster CO dissociation from trigonal-bipyramidal dicar-
bonyl rhodium hydrides than small bite angle ligands
exhibiting reduced spatial congestion.34 Due to the
rotational freedom around the Cp-Fe-Cp axis, a nearly
trans-chelating stabilization of four-coordinate hydrido
rhodium carbonyl species seems possible,7 whereby a
dissymmetric charge distribution imposed by ligand 1d
could prove advantageous in promoting hydride inser-
tion.
ligand
l:b ratiob
n-ald sel/%
TOFd
1a
1b
1c
1d
7.3
7.4
7.3
7.2
5.4
86.9
86.5
86.7
86.7
83.3
370
224
606
650
87
dppf
a
Reaction conditions: CO/H2 ) 1; pressure ) 20 bar, 4/1 ligand/
rhodium, 637/1 1-octene/rhodium; [Rh] ) 1.18 mM in toluene.
Linear/branched product ratio. c Selectivity toward 1-nonanal.
b
d
Substrate isomerization was found to be <2%. Turnover fre-
quency ) (mol of aldehyde) (mol of Rh)-1 h-1. Reactions were
stopped at ∼30% conversion.
lysts were rather low. Consequently, no clear trend
could be deduced from the latter results.
However, the generally valid relation between reac-
tivity, product selectivities, and Hammett σp constants
in the hydroformylation of styrene derivatives might be
rationalized by invoking an early transition state for the
(presumably selectivity-determining) hydride migra-
tion.33 Negative polarization of the hydride and higher
positive partial charge distribution on the terminal
alkene carbon in comparison to the internal one could
promote faster conversion of electron-poor substrates to
the branched alkyl rhodium species, herewith effectively
counterbalancing a possibly slower alkene coordination.
Con clu sion s
Our investigation on rhodium-catalyzed hydroformy-
lation using electronically perturbed phosphorus-chiral
diphosphines revealed some interesting aspects. When
4-methoxystyrene, styrene, and 4-chlorostyrene were
employed as substrates, reactivity and branched alde-
hyde selectivities were dominated by their correlation
with Hammett σp constants, whereby ligand electronic
variations affected the reaction outcome to a smaller
extent. However, for a given styrene derivative, a clear
trend in turnover frequencies was evidenced with
electron-withdrawing ligands as catalyst modifiers pro-
moting faster reaction rates. High-pressure NMR and
IR measurements showed an inverse relation between
ligand basicity and preference for the bis-equatorial
coordination mode in diphosphine rhodium dicarbonyl
hydride precursor complexes. Worthy of note is the fact
that the observed ee:ea ratios correlated well with the
enantiodiscriminating performance of the respective
diphosphines in styrene hydroformylation, where the
best π-acceptor ligand effected the highest ee values.
As theoretically anticipated, these results indicate the
relevance of diphosphine coordination modes in trigonal-
bipyramidal alkene complexes for the enantioselectivity-
determining step; however, the effects are small and
could easily be overshadowed by more pronounced steric
and electronic variations. Thus, for 4-methoxy- and
4-chlorostyrene substrates no such correlation was
found; in the former case ee values rather (inversely)
depended on reaction rates.
In the hydroformylation of 1-octene, ligand electronic
perturbations influenced only the reactivity, whereas
linear product selectivities and linear/branched ratios
were left unaffected. Consequently, optimization of
turnover frequencies is possible by means of electronic
fine-tuning, but more drastic ligand variations by
introduction of additional electronically perturbing
groups,7,8a possibly in sterically relevant positions,9,10
will inevitably alter catalytic pathways and change
product distributions. To summarize, our results il-
lustrate once more that regio- and enantiodiscriminat-
ing features in hydroformylation are governed by a
subtle interplay of steric and electronic features linked
To place this investigation and the obtained results
in the context of regioselective aliphatic alkene hydro-
formylation, we also performed catalysis experiments
using 1-octene as the model substrate (Scheme 5). The
results of reactions conducted at 80 °C are displayed in
Table 5. Interestingly, 1a -d -modified rhodium com-
plexes effected similar linear/branched ratios (∼7),
whereby the amounts of isomerized internal alkenes
were found to be <2% for all four catalysts. These
findings contrast with previous contributions, detailing
a higher tendency toward isomerization with increasing
ligand basicity.7,8a It might be reasoned, however, that
in the present investigation electronic perturbations
originating from phenyl ring substitution only were too
weak to induce such effects. Virtually the same selec-
tivities toward formation of the linear aldehyde product
were observed, once again indicating that under the
applied conditions electronic parameters might not
affect the regioselectivity-determining step(s). The lower
selectivity obtained using dppf-ligated catalysts could
originate from a smaller bite angle displaying its
influence during alkene coordination to a diphosphine
monocarbonyl rhodium hydride intermediate.7 Purely
steric reasons might also be invoked, featuring the well-
known argument of increased spatial congestion im-
posed by ligands 1a -d favoring the formation of the
linear alkyl rhodium species in comparison to the less
bulky dppf chelate. In contrast, assumption of a RhL2-
LH species (with L2 ) chelating dppf and L ) bridging
dppf) as selectivity relevant intermediate8a seems un-
likely, since for that case the argument of steric encum-
brance would suggest a higher linear/branched ratio in
comparison to results conferred by the CO-ligated 1a -d
rhodium hydrides.
(34) van der Veen, L. A.; Keeven, P. H.; Schoemaker, G. C.; Reek,
J . N. H.; Kamer, P. C. J .; van Leeuwen, P. W. N. M.; Lutz, M.; Spek,
A. L. Organometallics 2000, 19, 872.
(33) Matsubara, T.; Koga, N.; Ding, Y.; Musaev, D. G.; Morokuma,
K. Organometallics 1997, 16, 1065.

P-Chiral Diphosphines as Ligands in Hydroformylation
Organometallics, Vol. 19, No. 22, 2000 4605
133.27 (d, CH, J C-P ) 13.8 Hz); 136.31 (d, C, J C-P ) 6.1 Hz);
163.05 (br, C) ppm. 31P NMR (121.50 MHz): δ 132.95 (q, br,
J P-B ) 88 Hz) ppm. [R]20D ) -2.52° (c ) 0.60; CH2Cl2). HRMS
(EI+): m/z calcd for C17H23BNO2P, 315.1559; obsd, 315.1566.
Anal. Calcd for C17H23BNO2P: C, 64.79; H, 7.36; N, 4.44.
Found: C, 64.49; H, 7.37; N, 4.52.
to substrate as well as ligand properties. To date, its
origin and consequences appear, however, much too
complex to be reliably predicted by quantum mechanical
and/or molecular modeling treatments.35
Exp er im en ta l Section
(2RP ,4S,5R)-(+)-3,4-Dim et h yl-2-(4-(t r iflu or om et h yl)-
ph en yl)-5-ph en yl-1,3,2-oxaazaph osph olidin e-Bor an e (2c).
Gen er a l Com m en ts. If not stated otherwise, reactions
were carried out under an atmosphere of argon using standard
Schlenk techniques. THF and diethyl ether were distilled from
sodium benzophenone ketyl, CH2Cl2 and acetonitrile were
distilled from CaH2, and toluene and methanol from sodium
wire under nitrogen. NMR spectra were recorded on 250, 300,
and 400 MHz instruments; CDCl3 was used as the solvent if
not mentioned otherwise. Phosphorus-carbon coupling con-
stants (J C-P) were identified by comparison of J -modulated 13C
spectra measured at different magnetic field strengths. Phos-
phorus-boron coupling constants (J PB) were determined be-
tween the central peaks of the nonbinomial quartets. Elemen-
tal analyses were obtained using an Elementar Vario EL
apparatus. Mass spectra were recorded on a J EOL J MS SX/
SX102A four-sector mass spectrometer; for FAB-MS 3-ni-
trobenzyl alcohol was used as matrix. Melting points are
uncorrected. Optical rotations were measured in a thermo-
stated polarimeter with l ) 1 dm. With the exception of the
compounds given below, all reagents were purchased from
commercial suppliers and used without further purification.
Bis(diethylamino)(4-methoxyphenyl)phosphine and bis(diethyl-
amino)(4-(trifluoromethyl)phenyl)phosphine were prepared via
reaction of the respective para-substituted phenyl Grignard
reagents and phosphorus trichloride36 and subsequent conver-
sion of the aryldichlorophosphines in the presence of excess
diethylamine.37 The following compounds were synthesized
according to published procedures: 1,1′-dilithioferrocene,38 1,1′-
dibromoferrocene,39 and (S,S)-1,1′-bis(1-naphthylphenylphos-
phino)ferrocene (1a ).14
Syn th esis of Oxa a za p h osp h olid in e-Bor a n es 2 (Typ i-
ca l P r oced u r e). (1R,2S)-(-)-Ephedrine (0.1 mol) was dis-
solved in 500 mL of toluene and degassed by three freeze-
pump-thaw cycles. Bis(diethylamino)(4-methoxyphenyl)phos-
phine or bis(diethylamino)(4-(trifluoromethyl)phenyl)phos-
phine (0.1 mol) was added, and the solution was heated at 105
°C for 15 h. During this period, the diethylamine produced was
distilled off to drive the condensation toward completion.
Afterward, the reaction mixture was cooled on ice and BH3-
dimethyl sulfide (0.11 mol) was added dropwise. The solution
was stirred at room temperature overnight. The solvent was
removed in vacuo, and the residue was subjected to column
chromatography (SiO2; 3/1 toluene/hexane for 2b; 1/1 toluene/
hexane for 2c). Evaporation of the eluent left oxaazaphospho-
lidine-boranes 2 as white solids.
1
Yield: 38%. Mp: 89 °C. H NMR (400.13 MHz): δ 0.39-1.27
(m br, 3H); 0.70 (d, 3H, J ) 6.6 Hz); 2.58 (d, 3H, J H-P ) 11.1
Hz); 3.54 (m 1H); 5.44 (dd, 1H, J ) 3.0, 6.1 Hz); 7.19-7.26
(m, 5H); 7.61 (d, br, 2H, J ) 7.3 Hz); 7.80 (tr, br, 2H, J ) 8.6
Hz) ppm. 13C NMR (100.62 MHz): δ 13.52 (d, CH3, J C-P ) 3.1
Hz); 29.38 (d, CH3, J C-P ) 8.4 Hz); 59.01 (d, CH, J C-P ) 1.5
Hz); 84.51 (d, CH, J C-P ) 7.6 Hz); 125.43 (m, CF3); 125.43 (m,
CH); 126.56 (CH); 128.43 (CH); 128.51 (CH); 131.20 (d, CH,
J C-P ) 12.1 Hz); 133.92 (dd, C, J ) 2.3; 32.9 Hz); 135.77 (d, C,
J C-P ) 5.4 Hz); 137.48 (d, br, C, J ) 39.8 Hz) ppm. 31P NMR
(161.98 MHz): δ 133.04 (q, br, J P-B ) 80 Hz) ppm. [R]20
)
D
+9.87° (c ) 0.38; CH2Cl2). HRMS (EI+): m/z calcd for C17H20
-
BF3NOP, 353.1328; obsd, 353.1332. Anal. Calcd for C17H20BF3-
NOP: C, 57.82; H, 5.71; N, 3.91. Found: C, 57.64; H, 5.82; N,
3.89.
Syn t h esis of P h osp h in e Am id e-Bor a n es 3 (Typ ica l
P r oced u r e). 1-Bromonaphthalene (21 mmol) was dissolved
in 60 mL of Et2O, and the solution was degassed and cooled
to -78 °C. n-BuLi (22 mmol of a 2.5 M solution in hexane)
was added slowly, and the suspension was warmed to -20 °C
over a period of 2 h to ensure complete lithiation. Consecu-
tively, it was transferred slowly via Teflon cannula into a
Schlenk flask containing a precooled solution (-78 °C) of
oxaazaphospholidine-borane 2 (20 mmol) in 20 mL of THF.
The reaction mixture was warmed to room temperature
overnight and then quenched with water. Extractive workup
with CH2Cl2, filtration, and evaporation of solvent left the
crude product, which was purified by column chromatography
(SiO2, 96/4 toluene/ethyl acetate for 3b; 97.5/2.5 toluene/ethyl
acetate for 3c). The desired products were eluted first, followed
by minor amounts (∼5%) of phosphorus-epimeric byproducts.
(SP ,1R,2S)-(+)-N-Meth yl-N-((1-h yd r oxy-1-p h en yl)p r op -
2-yl)-P -(4-m e t h oxyp h e n yl)-P -(1-n a p h t h yl)p h osp h in -
a m id e-Bor a n e (3b). Yield: 81%. Mp: 56-58 °C. 1H NMR
(400.13 MHz): δ 0.90-1.72 (m, br, 3H); 1.29 (d, 3H, J ) 6.8
Hz); 1.86 (s, br, 1H); 2.62 (d, 3H, J ) 7.6 Hz); 3.82 (s, 3H);
4.45 (m, 1H); 4.99 (d, 1H, J ) 4.0 Hz); 6.90 (m, 2H); 7.14-
7.19 (m, 1H); 7.23-7.56 (m, 10H); 7.85 (d, 1H; J ) 8.1 Hz);
7.94 (d, 1H; J ) 8.1 Hz); 8.22 (d, 1H, J ) 8.6 Hz) ppm. 13C
NMR (100.62 MHz): δ 11.59 (d, CH3, J C-P ) 3.8 Hz); 31.29
(d, CH3, J C-P ) 3.8 Hz); 55.24 (CH3); 58.00 (d, CH, J C-P ) 10.7
Hz); 79.06 (d, CH, J C-P ) 2.3 Hz); 114.06 (d, CH, J C-P ) 11.5
Hz); 123.27 (d, C, J C-P ) 67.3 Hz); 124.60 (d, CH, J C-P ) 9.9
Hz); 126.05 (CH); 126.12 (CH); 126.22 (CH); 127.34 (d, CH,
J C-P ) 4.6 Hz); 127.40 (CH); 127.49 (d, C, J C-P ) 61.9 Hz);
128.29 (CH); 128.76 (d, CH, J C-P ) 0.9 Hz); 132.21 (d, CH,
J C-P ) 2.3 Hz); 132.55 (d, CH, J C-P ) 7.6 Hz); 133.31 (d, C,
J C-P ) 11.5 Hz); 134.01 (C); 134.04 (d, CH, J C-P ) 11.5 Hz);
142.61 (C); 161.63 (d, C, J C-P ) 1.5 Hz) ppm. 31P NMR (121.50
(2RP ,4S,5R)-(-)-3,4-Dim et h yl-2-(4-m et h oxyp h en yl)-5-
p h en yl-1,3,2-oxa a za p h osp h olid in e-Bor a n e (2b). Yield:
47%. Mp: 70-72 °C. 1H NMR (400.13 MHz): δ 0.52-1.32 (m,
br, 3H); 0.81 (d, 3H, J ) 6.5 Hz); 2.63 (d, 3H, J H-P ) 10.5 Hz);
3.68 (m, 1H); 3.83 (s, 3H); 5.56 (dd, 1H, J ) 2.5; 8.0 Hz); 6.97
(m, 2H); 7.28-7.40 (m, 5H); 7.75-7.79 (m, 2H) ppm. 13C NMR
(100.62 MHz); δ 13.42 (d, CH3, J C-P ) 2.3 Hz); 28.44 (d, CH3,
J C-P ) 8.4 Hz); 55.38 (CH3); 59.34 (d, CH, J C-P ) 2.3 Hz); 83.76
(d, CH, J C-P ) 7.6 Hz); 114.12 (d, CH, J C-P ) 10.7 Hz); 124.15
(d, C, J C-P ) 49.8 Hz); 126.54 (CH); 128.23 (CH); 128.33 (CH);
MHz): δ 70.73 (q, br, J PB ) 56 Hz) ppm. [R]20 ) +98.3° (c )
D
0.12; CH2Cl2). HRMS (EI+): m/z calcd for C27H31BNO2P,
443.2185; obsd, 443.2193. Anal. Calcd for C27H31BNO2P: C,
73.15; H, 7.05; N, 3.16. Found: C, 72.80; H, 6.95; N, 3.17.
(SP ,1R,2S)-(+)-N-Meth yl-N-((1-h yd r oxy-1-p h en yl)p r op -
2-yl)-P -(1-n a p h th yl)-P -(4-(tr iflu or om eth yl)p h en yl)p h os-
(35) Casey, C. P.; Petrovich, L. M. J . Am. Chem. Soc. 1995, 117,
6007.
(36) Gilheany, D. G.; Mitchel, C. M. In The Chemistry of Organo-
phosphorus Compounds; Hartley, F. R., Ed.; Wiley: New York, 1990;
Vol. 1, p 151.
(37) Schumann, H. J . Organomet. Chem. 1986, 299, 169.
(38) Bishop, J . J .; Davison, A.; Katcher, M. L.; Lichtenberg, D. W.;
Merrill, R. E.; Smart, J . C. J . Organomet. Chem. 1971, 27, 241.
(39) (a) Kovar, R. F.; Rausch, M. D.; Rosenberg, H. Organomet.
Chem. Synth. 1970/1971, 1, 173. (b) Dong, T.-Y.; Lai, L.-L. J . Orga-
nomet. Chem. 1996, 509, 131.
1
p h in a m id e-Bor a n e (3c). Yield: 72%. Mp: 74 °C. H NMR
(400.13 MHz): δ 0.70-1.75 (m, br, 3H); 1.33 (d, 3H, J ) 6.8
Hz); 1.86 (s, br, 1H); 2.59 (d, 3H, J H-P ) 7.3 Hz); 4.50 (m, 1H);
4.91 (d, br, 1H, J ) 4.6 Hz); 7.22-7.33 (m, 4H); 7.38-7.52 (m,
6H); 7.58-7.66 (m, 3H); 7.88 (d, br, 1H, J ) 8.1 Hz); 7.96-
8.01 (m, 1H); 8.22 (d, br, 1H, J ) 8.6 Hz) ppm. 13C NMR
(100.62 MHz): δ 12.10 (d, CH3, J C-P ) 3.8 Hz); 31.33 (d, CH3,
J C-P ) 3.8 Hz); 58.22 (d, CH, J C-P ) 10.7 Hz); 78.96 (d, CH,

4606 Organometallics, Vol. 19, No. 22, 2000
Nettekoven et al.
J C-P ) 3.1 Hz); 124.63 (d, CH, J C-P ) 10.7 Hz); 125.20 (m,
CF3); 126.16 (d, C, J C-P ) 62.0 Hz); 126.22 (CH); 126.42 (CH);
126.62 (CH); 126.88 (d, CH, J C-P ) 6.1 Hz); 127.77 (CH);
128.45 (CH); 129.01 (d, CH, J C-P ) 1.5 Hz); 132.43 (d, CH,
J C-P ) 9.9 Hz); 132.78 (d, CH, J C-P ) 2.3 Hz); 132.92 (d, CH,
J C-P ) 6.9 Hz); 133.24 (d, C, J C-P ) 11.5 Hz); 134.05 (d, C,
J C-P ) 7.6 Hz); 137.20 (d, C, J C-P ) 59.7 Hz); 142.28 (C) ppm.
31P NMR (121.50 MHz): δ 72.38 (q, br, J PB ) 34 Hz) ppm.
(S,S)-(+)-1,1′-Bis((4-m eth oxyp h en yl)(1-n a p h th yl)p h os-
p h in o)fer r ocen e (1b). Acidic decomplexation was performed
for intermediate 5b; this compound (∼2.5 mmol) was dissolved
in 15 mL of toluene, the solution was cooled on ice, and
trifluoromethanesulfonic acid (12.5 mmol) was added dropwise.
The reaction mixture was stirred at 0 °C for 30 min and then
warmed to ambient temperature. After 2 h, the solvent was
removed in vacuo and the residue treated with a degassed
solution of KOH (25 mmol) in 8 mL of 10/1 EtOH/H2O. The
suspension was agitated for 30 min and then repeatedly
extracted with Et2O. The combined organic layers were dried
over Na2SO4, filtered, and concentrated. Column chromatog-
raphy (alumina, 1/1 CH2Cl2/hexane) yielded enantiopure prod-
[R]20 ) +77.6° (c ) 0.30; CH2Cl2). HRMS (FAB+): m/z calcd
D
for C27H29BF3NOP (MH+), 482.2032; obsd, 482.2038. Anal.
Calcd for C27H28BF3NOP: C, 67.38; H, 5.87; N, 2.91. Found:
C, 67.04; H, 6.15; N, 2.73.
Syn th esis of Meth yl P h osp h in ite-Bor a n es 4 (Typ ica l
P r oced u r e). Phosphine amide-borane 3 (10 mmol) was
dissolved in 80 mL of methanol and cooled on ice. Concentrated
sulfuric acid (10.5 mmol) was added dropwise, and the solution
was stirred at room temperature for 15 h. Then the reaction
mixture was concentrated and subjected to column chroma-
tography (SiO2, 9/1 hexane/ethyl acetate for 4b, 95/5 hexane/
ethyl acetate for 4c) to afford the respective methyl phosphin-
ite-boranes as a colorless oil (4b) and white crystals (4c).
(R)-(-)-Meth yl (1-Na p h th yl)(4-m eth oxyp h en yl)p h os-
p h in ite-Bor a n e (4b). Yield: 84%. Oil. 1H NMR (400.13
MHz): δ 3.72 (s, 3H); 3.77 (d, 3H, J H-P ) 12.1 Hz); 6.90 (m,
2H); 7.40-7.49 (m 2H); 7.53-7.58 (m, 1H); 7.61-7.68 (m, 2H);
7.91 (d, br, 1H, J ) 7.8 Hz); 8.06 (d, br, 1H, J ) 8.1 Hz); 8.23-
8.31 (m, 2H) ppm. 13C NMR (100.62 MHz): δ 53.55 (d, CH3,
J C-P ) 3.1 Hz); 55.05 (CH3); 114.11 (d, CH, J C-P ) 11.5 Hz);
122.57 (d, C, J C-P ) 70.4 Hz); 124.55 (d, CH, J C-P ) 13.0 Hz);
126.11 (CH); 126.22 (d, CH, J C-P ) 4.6 Hz); 126.79 (CH);
126.93 (d, C, J C-P ) 57.4 Hz); 128.87 (CH); 132.45 (d, C, J C-P
) 6.1 Hz); 132.94 (d, CH, J C-P ) 12.3 Hz); 133.24 (d, CH, J C-P
1
uct 1b as a yellow powder. Yield: 51%. Mp: 110 °C. H NMR
(400.13 MHz): δ 3.57 (m, 2H); 3.70 (s, 6H); 4.22 (m, 2H); 4.31
(m, 2H); 4.34 (m, 2H); 6.76 (d, 4H, J ) 8.1 Hz); 7.10 (ddd, 2H,
J ) 1.0, 4.8, 6.8 Hz); 7.29-7.41 (m, 10H); 7.74 (d, 2H, J ) 8.4
Hz); 7.77 (d, 2H, J ) 7.8 Hz); 8.29 (dd, br, 2H, J ) 3.8, 7.8 Hz)
ppm. 13C NMR (100.62 MHz): δ 55.05 (CH3); 72.51 (br, CH);
72.64 (CH); 72.97 (d, CH, J C-P ) 6.9 Hz); 75.57 (d, CH, J C-P
)
29.1 Hz); 77.06 (C); 113.77 (d, CH, J C-P ) 8.4 Hz); 125.20 (CH);
125.68 (CH); 125.87 (CH); 126.00 (d, CH, J C-P ) 24.8 Hz);
127.48 (d, C, J C-P ) 5.4 Hz); 128.47 (CH); 128.84 (CH); 130.80
(CH); 133.34 (d, C, J C-P ) 4.6 Hz); 134.45 (d, C, J C-P ) 21.4
Hz); 135.53 (d, CH, J C-P ) 21.4 Hz); 137.42 (d, C, J C-P ) 13.8
Hz); 160.30 (C) ppm. 31P NMR (121.50 MHz): δ -26.80 (s)
ppm. [R]20 ) +262° (c ) 0.45; CH2Cl2). HRMS (FAB+): m/z
D
calcd for C44H37FeO2P2 (MH+), 715.1618; obsd, 715.1613. Anal.
Calcd for C44H36FeO2P2: C, 73.96; H, 5.08. Found: C, 73.82;
H, 5.38.
(S,S)-(+)-1,1′-Bis((1-n aph th yl)(4-(tr iflu or om eth yl)ph en -
yl)p h osp h in o)fer r ocen e (1c). Diphosphine-diborane 5c
(∼2.5 mmol) was dissolved in 10 mL of degassed diethylamine
and stirred overnight at room temperature. Afterward, the
solvent was removed in vacuo and the residue purified by
column chromatography (alumina, 1/1 CH2Cl2/hexane) to
afford the enantiopure ligand 1c as orange foam. Yield: 36%.
) 2.3 Hz); 133.56 (d, C, J C-P ) 6.9 Hz); 134.10 (d, CH, J C-P
)
14.5 Hz); 162.23 (d, C, J C-P ) 2.3 Hz) ppm. 31P NMR (121.50
MHz): δ 110.91 (q, br, J PB ) 78 Hz) ppm. [R]20D ) -49.5° (c )
0.59; CH2Cl2). HRMS (EI+): m/z calcd for C18H20BO2P, 310.1294;
obsd, 310.1294. Anal. Calcd for C18H20BO2P: C, 69.71; H, 6.50.
Found: C, 69.69; H, 6.60.
1
Mp: 95-97 °C. H NMR (400.13 MHz): δ 3.81 (m, 2H); 4.15
(m, 2H); 4.24 (m, 2H); 4.28 (m, 2H); 7.12 (t, br, 2H, J ) 5.8
Hz); 7.34 (tr, 2H, J ) 7.4 Hz); 7.39-7.50 (m, 12H); 7.81 (d,
4H, J ) 7.9 Hz); 8.36 (m, 2H) ppm. 13C NMR (100.62 MHz):
δ 72.45 (br, CH); 72.90 (d, CH, J C-P ) 5.2 Hz); 73.63 (d, CH,
J C-P ) 9.9 Hz); 74.65 (d, CH, J C-P ) 20.8 Hz); 75.73 (d, C,
J C-P ) 6.1 Hz); 124.85 (m, CF3); 125.31 (CH); 125.70 (CH);
126.01 (CH); 126.03 (CH); 126.25 (d, CH, J C-P ) 2.3 Hz);
128.71 (CH); 129.80 (CH); 130.50 (d, C, J C-P ) 32.1 Hz); 132.39
(d, CH, J C-P ) 2.3 Hz); 133.45 (d, C, J C-P ) 2.3 Hz); 133.50 (d,
CH, J C-P ) 19.9 Hz); 134.99 (d, C, J C-P ) 22.2 Hz); 135.20 (d,
C, J C-P ) 13.8 Hz); 143.27 (d, C, J C-P ) 12.2 Hz) ppm. 31P
(R)-(-)-Meth yl (1-Na p h th yl)(4-(tr iflu or om eth yl)p h en -
yl)p h osp h in ite-Bor a n e (4c). Yield: 75%. Mp: 92 °C. 1H
NMR (400.13 MHz): δ 0.62-1.55 (m, br, 3H); 3.82 (d, 3H, J H-P
) 12.1 Hz); 7.41-7.45 (m, 1H); 7.49-7.53 (m, 1H); 7.61-7.67
(m, 3H); 7.73-7.79 (m, 2H); 7.92 (d, br, 1H, J ) 8.1 Hz); 8.06
(d, br, 1H, J ) 8.3 Hz); 8.11 (d, br, 1H, J ) 8.3 Hz); 8.36 (ddd,
1H, J ) 1.0; 7.3; 15.9 Hz) ppm. 13C NMR (100.62 MHz): δ
54.27 (d, CH3, J C-P ) 1.5 Hz); 124.82 (d, CH, J C-P ) 15.3 Hz);
125.38 (m, CH); 125.51 (d, C, J C-P ) 53.5 Hz); 125.91 (d, CH,
J C-P ) 5.4 Hz); 126.53 (CH); 127.51 (CH); 129.29 (CH); 131.07
(d, CH, J C-P ) 11.5 Hz); 132.44 (d, C, J C-P ) 3.8 Hz); 133.21
(dd, C, J ) 2.3; 32.9 Hz); 133.77 (d, C, J C-P ) 6.1 Hz); 134.35
(d, CH, J C-P ) 2.3 Hz); 136.13 (d, CH, J C-P ) 19.9 Hz); 137.20
(d, C, J C-P ) 66.5 Hz) ppm. 31P NMR (121.50 MHz): δ 112.32
NMR (121.50 MHz): δ -25.95 (s) ppm. [R]20 ) +130° (c )
D
0.29, CH2Cl2). HRMS (FAB+): m/z calcd for C44H31F6FeP2
(MH+), 791.1155; obsd, 791.1169. Anal. Calcd for C44H30F6-
FeP2: C, 66.85; H, 3.83. Found: C, 66.97; H, 4.17.
(q, br, J PB ) 74.2 Hz) ppm. [R]20 ) -2.5° (c ) 0.39; CH2Cl2).
D
HRMS (EI+): m/z calcd for C18H17BF3OP, 348.1062; obsd,
348.1088. Anal. Calcd for C18H17BF3OP: C, 62.10; H, 4.93.
Found: C, 62.25; H, 5.23.
Syn th esis of (S)-(+)-Meth yl (1′-Br om o-1-fer r ocen yl)-
p h en ylp h osp h in ite-Bor a n e (6). 1,1′-Dibromoferrocene (12
mmol) was dissolved in 50 mL of THF and cooled to -50 °C.
sec-BuLi (11 mmol of a 1.3 M solution in cyclohexane) was
added slowly via syringe, and the mixture was warmed to -30
°C over a period of 2 h. After this mixture was cooled again, a
solution of methyl phosphinite-borane (R)-4b (10 mmol) in
10 mL of THF was added via Teflon cannula at -70 °C, and
the mixture was allowed to reach ambient temperature
overnight. Then, the reaction was quenched with water, the
solvent was removed in vacuo, and the residue was extracted
with CH2Cl2. Drying over MgSO4, filtration, and evaporation
afforded the crude product, which was purified by column
chromatography (SiO2, 1/3 CH2Cl2/hexane). The desired bromo-
containing phosphine-borane (S)-6 was eluted first, closely
followed by ∼20% of monosubstituted byproduct. Yield: 41%.
Mp: 149-151 °C. 1H NMR (400.13 MHz): δ 0.56-1.78 (m,
br, 3H); 3.82 (s, 3H); 4.03 (m, 1H); 4.15 (m, 1H); 4.22 (m, 2H);
Syn t h esis of F er r ocen yld ip h osp h in es 1b ,c (Typ ica l
P r oced u r e). Dilithioferrocene (5 mmol) was dissolved in a
mixture of 40 mL of Et2O and 10 mL of THF and cooled to
-50 °C. The resulting suspension was added via a Teflon
cannula to a precooled (-50 °C) solution of methyl phosphinite
4 (10 mmol) in 10 mL of THF. The reaction mixture was
allowed to reach room temperature over a period of 15 h and
was then quenched with water. The solvent was removed in
vacuo, and the residue was extracted twice with CH2Cl2. The
combined organic layers were dried (MgSO4) and filtered, and
the solvent was evaporated. Purification by column chroma-
tography removed monosubstituted byproducts (∼10%) and
small amounts (<10%) of isomerized meso diphosphine dibo-
ranes; the desired C2-symmetrical diborane complexes 5b,c
were eluted last and concentrated.

P-Chiral Diphosphines as Ligands in Hydroformylation
Organometallics, Vol. 19, No. 22, 2000 4607
4.44 (m, 1H); 4.54 (m, 1H); 4.59 (m, 1H); 4.75 (m, 1H); 6.93
(m, 2H); 7.28 (ddd, 1H, J ) 1.2, 6.8, 8.6 Hz); 7.35-7.44 (m,
3H); 7.63 (ddt, 2H, J ) 2.3, 8.8, 10.6 Hz); 7.80 (d, br, 1H, J )
8.1 Hz); 7.88-7.95 (m, 2H) ppm. 13C NMR (100.62 MHz): δ
30.46 (C); 55.30 (s, CH3); 69.59 (d, 2CH, J C-P ) 6.1 Hz); 71.68
(d, 2CH, J C-P ) 13.8 Hz); 74.44 (d, CH, J C-P ) 5.4 Hz); 75.48
(d, CH, J C-P ) 7.6 Hz); 75.62 (d, CH, J C-P ) 6.1 Hz); 76.06 (d,
CH, J C-P ) 13.8 Hz); 77.74 (C); 114.43 (d, CH, J C-P ) 11.5
Hz); 121.27 (d, C, J C-P ) 63.5 Hz); 124.57 (d, CH, J C-P ) 10.7
Hz); 126.15 (d, CH, J C-P ) 22.9 Hz); 127.35 (d, CH, J C-P ) 6.1
Hz); 128.38 (d, C, J C-P ) 55.1 Hz); 128.80 (d, CH, J C-P ) 0.9
Hz); 132.21 (d, CH, J C-P ) 2.3 Hz); 132.74 (d, C, J C-P ) 9.2
Hz); 133.47 (d, CH, J C-P ) 8.4 Hz); 133.90 (d, C, J C-P ) 6.9
Hz); 134.10 (d, CH, J C-P ) 11.5 Hz); 161.93 (d, C, J C-P ) 2.3
Hz) ppm. 31P NMR (121.50 MHz): δ 17.85 (q, br, J PB ) 45 Hz)
High -P r essu r e NMR Exp er im en ts (Typ ica l P r oce-
d u r e). Rh(CO)2acac (0.012 mmol) and the respective ligand
1a -d or dppf (0.026 mmol) were dissolved in 1.5 mL of toluene-
d8, and the solution was degassed and transferred into a 0.5
mL sapphire NMR tube. The tube was pressurized with 18
bar of syngas (1/1 CO/H2) and heated to 40 °C in the
spectrometer. The formation of diphosphine rhodium dicar-
bonyl hydrides was usually completed within 1.5 h. Low-
temperature measurements were conducted employing 1/1
THF-d4/acetone-d6 as the solvent.
High -P r essu r e IR Exp er im en ts (Typ ica l P r oced u r e).
Rh(CO)2acac (0.012 mmol) and the respective diphosphine
(0.026 mmol) were dissolved in 15 mL of cyclohexane, and the
solution was degassed and introduced into the high-pressure
IR autoclave. The mechanically stirred solution was pressur-
ized to 18 bar and heated to 60 °C in order to hasten formation
of diphosphine rhodium dicarbonyl hydrides at the solubility-
limited low concentrations. The reaction was usually completed
within 2-6 h.
ppm. [R]20 ) +112.9° (c ) 0.42; CH2Cl2). HRMS (FAB+): m/z
D
calcd for C27H22BrFeOP (M+ - BH3), 527.9943; obsd, 527.9968.
Anal. Calcd for C27H25BBrFeOP: C, 59.72; H, 4.64. Found: C,
59.98; H, 4.75.
Syn th esis of (S,S)-(+)-1-[(1-Na p h th yl)(4-m eth oxyp h en -
yl)p h osp h in o]-1′-[(1-n a p h th yl)((4-tr iflu or om eth yl)p h en -
yl)p h osp h in o]fer r ocen e (1d ). Methyl phosphinite-borane
6 (1.5 mmol) was dissolved in 3 mL of Et2O and 3 mL of THF
and cooled to -70 °C. sec-BuLi (1.65 mmol) was added via
syringe after 1 h, followed by cannula transfer of a solution of
methyl phosphinite-borane (R)-4c (1.65 mmol) in 2 mL of
THF. The reaction mixture was warmed to ambient temper-
ature over a period of 15 h, quenched with water, and extracted
with CH2Cl2. The combined organic layers were dried over
MgSO4, filtered, and concentrated. Column chromatography
(SiO2, 1/2 CH2Cl2/hexane) afforded the enantiopure diphos-
phine-diborane complex, which was subjected to deprotection
as described for compound 5b. Final column chromatographic
purification (alumina, 1/1 CH2Cl2/hexane) yielded the desired
diphosphine (S,S)-1d as yellow foam. Yield: 70%. Mp: 99 °C.
1H NMR (400.13 MHz): δ 3.71 (m, 1H); 3.75 (s, 3H); 3.77 (m,
1H); 4.19 (m, 1H); 4.25-4.32 (m, 4H); 4.40 (m, 1H); 6.79 (m,
2H); 7.08 (ddd, 1H; J ) 1.0, 4.8, 6.8 Hz); 7.18 (ddd, 1H, J )
1.0, 5.3, 6.6 Hz); 7.28-7.45 (m, 8H); 7.46 (d, 4H, J ) 4.0 Hz);
7.74 (d, 1H, J ) 8.1 Hz); 7.77 (d, 1H, J ) 7.8 Hz); 7.81 (d, 2H,
J ) 7.8 Hz); 8.30 (m, 1H); 8.35 (m, 1H) ppm. 13C NMR (100.62
MHz): δ 55.08 (CH3); 72.15 (CH); 72.67 (d, CH, J C-P ) 2.3
Hz); 72.75 (d, CH, J C-P ) 7.5 Hz); 72.79 (CH); 73.13 (dd, CH
J ) 1.4; 5.5 Hz); 73.29 (d, CH, J C-P ) 7.6 Hz); 74.89 (d, CH,
J C-P ) 22.9 Hz); 75.28 (d, C, J C-P ) 5.5 Hz); 75.43 (d, CH,
J C-P ) 26.8 Hz); 77.54 (d, C, J C-P ) 5.4 Hz); 113.89 (d, CH,
J C-P ) 8.4 Hz); 124.76 (m, CF3); 125.20 (d, CH, J C-P ) 0.9
Hz); 125.33 (d, CH, J C-P ) 1.5 Hz); 125.72 (br, CH); 125.86 (d,
CH, J C-P ) 26.8 Hz); 125.90 (d, CH, J C-P ) 2.3 Hz); 125.96 (d,
CH, J C-P ) 26.0 Hz); 125.98 (d, CH, J C-P ) 1.5 Hz); 126.23 (d,
CH, J C-P ) 2.3 Hz); 127.44 (d, C, J C-P ) 5.4 Hz); 128.50 (d,
CH, J C-P ) 1.5 Hz); 128.68 (d, CH, J C-P ) 1.4 Hz); 128.94 (CH);
129.69 (CH); 130.42 (d, C, J C-P ) 32.1 Hz); 130.88 (CH); 132.24
(d, CH, J C-P ) 2.3 Hz); 133.35 (d, C, J C-P ) 3.8 Hz); 133.42 (d,
C, J C-P ) 4.6 Hz); 133.59 (d, CH, J C-P ) 19.9 Hz); 134.51 (d,
C, J C-P ) 21.4 Hz); 134.93 (d, C, J C-P ) 21.4 Hz); 135.50 (d,
C, J C-P ) 13.8 Hz); 135.53 (d, CH, J C-P ) 21.4 Hz); 137.13 (d,
C, J C-P ) 13.8 Hz); 143.29 (dd, C, J ) 1.4; 12.9 Hz); 160.39
(C) ppm. 31P NMR (121.50 MHz): δ -25.34 (s), -26.94 (s) ppm.
[R]20D ) +117.9° (c ) 0.28; CH2Cl2). HRMS (FAB+): m/z calcd
for C44H34F3FeOP2 (MH+), 753.1386; obsd, 753.1397. Anal.
Calcd for C44H33F3FeOP2: C, 70.23; H, 4.42. Found: C, 70.07;
H, 4.78.
(Asym m etr ic) Hyd r ofor m yla tion Rea ction s (Typ ica l
P r oced u r e). Rh(CO)2acac (0.015 mmol) and diphosphine
ligand (0.033 mmol) were dissolved in 3.5 mL of toluene, and
the solution was degassed by three freeze-pump-thaw cycles.
The solution was transferred into a 100 mL stainless steel
autoclave, equipped with glass insert, substrate inlet vessel,
electronic heating mantle, and liquid sampling valve. The
argon atmosphere was replaced by 18 bar of syngas (1/1 CO/
H2), and the magnetically stirred solution was heated to the
given temperature. After 2 h of catalyst preparation, the
degassed substrate solution (7.5 mmol of (substituted) styrene
in 3 mL of toluene and 0.5 mL of decane as internal standard)
was purged three times with 1/1 CO/H2 and added under
pressure. The progress of the reaction was monitored by
subjecting samples to GC analysis. After the desired degree
of conversion had been reached, catalysis was stopped by
cooling the autoclave on ice and depressurizing. The crude
product mixture was purified by fractional distillation. Sub-
sequently, linear and branched aldehyde products (1.5 mmol)
were subjected to reduction with NaBH4 (1.5 mmol) by stirring
in 10 mL of EtOH for 30 min. Quenching with water,
extraction with a 1/1 solution of ethyl acetate/hexane, drying
of organic layers, filtration, and removal of solvent gave the
corresponding alcohols, for which the enantioselectivity of the
branched derivative was determined by chiral GC (Cyclosil-
B; isothermal; T ) 140 °C for 2-phenylpropanol, tR(R) ) 18.1
min, tR(S) ) 18.8 min; T ) 150 °C for 2-(4-methoxyphenyl)-
propanol, tR(R) ) 42.5 min, tR(S) ) 43.4 min; T ) 150 °C for
2-(4-chlorophenyl)propanol, tR(R) ) 45.1 min, tR(S) ) 46.3 min).
For 1-octene hydroformylation, catalyst precursor solutions
were prepared from 0.01 mmol of Rh(CO)2acac and 0.04 mmol
of diphosphine in 5 mL of toluene. After 2 h of catalyst
preparation at 80 °C and 18 bar of 1/1 CO/H2, reaction was
started by substrate addition (6.37 mmol of 1-octene in 3.5 mL
of toluene and 0.5 mL of decane as internal standard). Samples
were drawn from the reaction mixture, immediately cooled on
ice, and analyzed by GC.
Ack n ow led gm en t. We appreciate financial support
by the IOP Katalyse program of The Netherlands
Ministry of Economic Affairs (Project IKA 94081) and
are thankful for a special grant to U.N.
OM0004489