Cooperative attractive interactions in asymmetric hydrogenations with dihydroxydiphosphine Rh(I) catalysts - A competition study

Article abstract of DOI:10.1016/S0022-328X(00)00071-1

Studies for controlling rate and enantioselectivity of the asymmetric hydrogenation with Rh-diphosphine catalysts by cooperative attractive interactions within the framework of catalyst-substrate complexes are represented. In strong contrast to a Rh(I)[threo-1,4-bis(diphenylphosphino)butane-2,3-diol] catalyst an extremely fast reaction took place applying the complex of the analogue erythro ligand. This is likely due to a strong intramolecular hydrogen bond between the vicinal HO groups impeding the hemilabile coordination of one of the hydroxy groups on the metal center during the hydrogenation. When a substrate with strong hydrogen bond acceptor properties such as (E)-methyl 3-dimethoxyphosphorylbut-2-enoate was hydrogenated even the threo catalyst exhibited a fast reaction. The product was obtained in 83% ee. In contrast, when in the ligand both HO groups were replaced by MeO groups only poor conversion and 24% ee were achieved. While with the complex bearing HO groups in methanolic solution the corresponding diastereomeric catalyst substrate complexes were dominant, analogue MeO groups bearing catalyst substrate complexes could not be detected by NMR. For the latter, only the Rh(I)-solvent complex was found revealing the high importance of the additional functional groups on this equilibrium.

Full text of DOI:10.1016/S0022-328X(00)00071-1

www.elsevier.nl/locate/jorganchem
Journal of Organometallic Chemistry 603 (2000) 61–68
Cooperative attractive interactions in asymmetric hydrogenations
with dihydroxydiphosphine Rh(I) catalysts — a competition study
Jens Holz, Renat Kadyrov *1, Susanne Borns, Detlef Heller, Armin Bo¨rner *2
Institut fu¨r Organische Katalyseforschung an der Uni6ersita¨t Rostock e.V., Buchbinderstraße 5/6, D-18055 Rostock, Germany
Received 18 November 1999; received in revised form 27 January 2000
Abstract
Studies for controlling rate and enantioselectivity of the asymmetric hydrogenation with Rhꢀdiphosphine catalysts by
cooperative attractive interactions within the framework of catalyst–substrate complexes are represented. In strong contrast to a
Rh(I)[threo-1,4-bis(diphenylphosphino)butane-2,3-diol] catalyst an extremely fast reaction took place applying the complex of the
analogue erythro ligand. This is likely due to a strong intramolecular hydrogen bond between the vicinal HO groups impeding the
hemilabile coordination of one of the hydroxy groups on the metal center during the hydrogenation. When a substrate with strong
hydrogen bond acceptor properties such as (E)-methyl 3-dimethoxyphosphorylbut-2-enoate was hydrogenated even the threo
catalyst exhibited a fast reaction. The product was obtained in 83% ee. In contrast, when in the ligand both HO groups were
replaced by MeO groups only poor conversion and 24% ee were achieved. While with the complex bearing HO groups in
methanolic solution the corresponding diastereomeric catalyst substrate complexes were dominant, analogue MeO groups bearing
catalyst substrate complexes could not be detected by NMR. For the latter, only the Rh(I)–solvent complex was found revealing
the high importance of the additional functional groups on this equilibrium. © 2000 Elsevier Science S.A. All rights reserved.
Keywords: Asymmetric synthesis; Cooperative effect; Hydrogenation; Phosphines; Rhodium
1. Introduction
the prochiral substrate. By these interactions a differ-
ence in the thermodynamic stabilities and reactivities of
diastereomeric intermediates yields being the precondi-
tion for the achievement of enantioselectivity in the
product. It is interesting to note that additional attrac-
tive interactions e.g. electrostatic interactions or hydro-
gen bonds within the framework of the catalyst–
substrate assembly have been seldom employed to con-
trol conformational fluxionality [4]. One crucial prob-
lem of this type of recognition widely distributed in
enzymes consists is the large number and variety of
competing binding sites in relevant polyfunctionalized
catalyst–substrate complexes. Therefore, modeling of
individual interactions, studies of equilibria and evalua-
tion of their effect on rate and enantioselectivity repre-
sent a great challenge.
One of the most fascinating features in homoge-
neously metal catalyzed asymmetric synthesis is the
transfer of the chiral information residing in the back-
bone of the chiral catalyst to the prochiral substrate.
The understanding of this mechanism is essential for a
rational design of new ligands and catalysts. Several
rationalizations have been suggested to explain results
with selected catalytic systems accumulated over the
years in terms of product configuration and degree of
stereoselection. Besides electronic effects [1] and kinetic
features [2], rigid or restrained conformations of the
ligand have been identified to be of pivotal importance
[3]. Inspired by these ‘steric’ models for the construc-
tion of new ligands, main attention is usually given to
the incorporation of bulky groups in order to increase
repulsive interactions between parts of the catalyst and
Since some years we elaborate the effect of hydroxy
groups incorporated in different positions of conven-
tional chiral Rh(I) diphosphine catalysts on rate and
enantioselectivity of the asymmetric hydrogenation. We
and others showed that the introduction of one or more
hydroxy groups into a conventional catalyst changes its
1*Corresponding author. Fax: +49-381-4669324.
2*Corresponding author. Fax:
armin.boerner@ifok.uni-rostock.de
+49-381-4669324; e-mail:
0022-328X/00/$ - see front matter © 2000 Elsevier Science S.A. All rights reserved.
PII: S0022-328X(00)00071-1

62
J. Holz et al. / Journal of Organometallic Chemistry 603 (2000) 61–68
enantiodifferentiating ability [5]. In several cases the
intrinsic enantioselectivity of the parent catalyst could
be improved [6]. A similar effect was observed when a
properly situated methoxy group in the catalyst was
replaced by a hydroxy group [7]. Unfortunately, the
additional functionality in the diphosphine Rh(I) cata-
lyst caused in most examples a lowering of the reaction
rate [8,9]. Particularly pronounced effects were ob-
served in the hydrogenation of N-acetylaminocinnamic
acid and itaconic acid derivatives with a conformation-
ally-flexible Rh(I) catalyst based on (R,R)-1,4-bis(-
diphenylphosphino)butane-2,3-diol (threo-1-OH) as
ligand serving therefore as a model [10]. Spectroscopic
studies, X-ray analysis and DFT calculations gave evi-
dence that one of the hydroxy groups can interact with
the metal center forming a coordinatively saturated
h3-coordinated 18 e− Rh complex (Scheme 1, A)
[11,12]. Complexes of this type are catalytically inac-
tive. Only the dissociation of the hydroxy group from
the metal center affording a h2-coordinated species
allows the oxidative addition of hydrogen (C). In other
words, the concentration of complex A and the rate of
decomplexation, respectively, can influence the rate of
the hydrogenation process decisively.
metal center should be weakened. A strong hydrogen
bond acceptor in a suitable position of the prochiral
substrate could result in the same effect.
In order to prove these hypotheses herein we report
on the synthesis of the requisite (achiral) (2R,3S)-1,4-
bis(diphenylphosphino)butane-2,3-diol (erythro-1-OH)
and its application in the Rh-catalyzed hydrogenation
of several standard olefins. For comparison results ob-
served with the chiral (R,R)-configurated counterpart
threo-1-OH and its corresponding dimethylether threo-
1-OMe will be also detailed [10]. As a chelating sub-
strate with strong hydrogen bond acceptor properties a
prochiral bifunctional a,b-unsaturated phosphonate
was chosen for the subsequent asymmetric hydrogena-
tion with chiral Rh(I) complexes of threo-1-OH and
threo-1-OMe, respectively.
In our model system the formation of the h3-coordi-
nated species is forced by the threo arrangement of the
hydroxy groups. In the analogous mono-hydroxy com-
plex no interactions between the HO group and metal
center could be observed both in solution and in the
solid state [7]. Nevertheless, unique catalytic effects
attributed to the HO group were found. The interaction
in the dihydroxy catalyst is dependent on the solvent
[11]. Thus, by changing from methanol to methylene
chloride, the second (non-coordinated) hydroxy group
competed with the metal for the (metal)coordinated
hydroxy group by formation of an intramolecular hy-
drogen bond. By 31P{1H}-NMR and IR spectroscopy
in CD2Cl2 the coordinatively unsaturated h2-complex B
was found to be dominant (Scheme 1). Interestingly,
when the hydrogenation was performed in methylene
chloride the consumption rate of hydrogen increased.
Based on these results we anticipated that by a change
of the geometry of the threo 2,3-diol moiety to erythro
the tendency for the formation of the intramolecular
hydrogen bond should be forced. As a result the rate
limiting interaction between hydroxy group and the
2. Results and discussion
2.1. Cooperati6e effects on the rate of the
hydrogenation
2.1.1. Synthesis of the new ligand
The synthesis of (2R,3S)-1,4-bis(diphenylphos-
phino)butane-2,3-diol (erythro-1-OH) was accom-
plished as detailed in Scheme 2. In general, protected
erythritol 3 being a key compound in the whole synthe-
sis can be prepared from commercially available meso
tartaric acid [13]. But due to the high price of the
starting material this approach is quite expensive. An
alternative pathway starting with maleate, however,
requires several synthetic steps [14]. Therefore, we
based our approach on commercially available (+)-
2,3-O-cyclohexylidene- -erythruronic acid (2) which
L
was reduced with LiAlH4 to give 3. Esterification of the
HO groups with tosyl chloride afforded the ditosylate
4. Replacement of both tosyl groups by lithium phos-
Scheme 1. HOꢀRh interactions of threo-1-OH complexes and their influence on the reactivity towards H2 (L,L%=diolefin, MeOH, bidentate
substrate).

J. Holz et al. / Journal of Organometallic Chemistry 603 (2000) 61–68
63
Scheme 2. Synthesis of (2R,3S)-1,4-bis(diphenylphosphino)butane-2,3-diol (erythro-1-OH).
phide gave rise to the diphosphine 5, which represents
an achiral analogue of DIOP [15]. In the 31P{1H}-NMR
spectrum in CDCl3 this phosphine was characterized by
a singlet at l −20.6. Cleavage of the cyclohexylidene
acetal was performed with methanesulfonic acid in
aqueous methanol affording erythro-1-OH in 56%
yield. As a result of the liberation of the HO groups a
highfield shift took place in the 31P{1H}-NMR. The
phosphorus resonance was observed at l −23.0.
The functionalized diphosphine erythro-1-OH was
reacted with [Rh(COD)2]BF4 (COD=1,5-cycloocta-
diene) to give the rhodium(I) complex as a yellow
powder (Scheme 3).
field characterizes the second phosphorus, which is part
of the fused six-membered chelate. The small difference
in the chemical shift of Dw˜ 21.9 ppm is in remarkable
contrast to the value observed for [Rh(threo-1-
OH)(COD)]BF4 (Dw˜ =46.7 ppm), but corresponds to
the difference found with [Rh(threo-1-OMe)(COD)]BF4
(Dw˜ =16.1 ppm) [10]. For the latter a higher dominance
of the bidentate coordination mode was found in
deuterated methanol. Obviously, also with [Rh(erythro-
1-OH)(COD)]BF4 besides h3-coordinated species A-
COD (Scheme 4) considerable amounts of
h2-coordinated conformations B-COD are present at
ambient temperature.
Since the COD complex could not be cleanly reduced
to the desired bis-methanol complex (catalytically active
species) it was converted into [Rh(erythro-1-
OH)(NBD)]BF4 (NBD=norbornadiene) by treatment
with an excess of norbornadiene (Scheme 5). Subse-
quently, this complex was exposed for 30 min to 0.1
MPa of hydrogen in order to remove norbornadiene by
hydrogenation and to generate [Rh(erythro-1-
OH)(MeOH)2]BF4. To this complex a five-fold molar
excess of methyl (Z)-N-acetylaminocinnamate (AMe)
was added. Due to the 31P{1H}-NMR spectrum the
bis-methanol complex was fully converted into sub-
strate–catalyst complexes (A-AMe). At 297 K an ABX
pattern with different line width for the A and B part
became visible. The signals broadened as the tempera-
ture was lowered further. Finally, two well separated
2.1.2. Spectroscopic in6estigations
In the IR solid state spectrum the HO valence band
was found at 3443 cm−1. This is only a small difference
compared to the threo complex, where the correspond-
ing band was observed at 3441 cm−1 [10]. This indi-
cates that also in the erythro precatalyst the interaction
between one of the hydroxy groups and the rhodium
center takes place, giving rise to a h3-coordinated
complex.
The 31P{1H}-NMR spectrum of [Rh(erythro-1-
OH)(COD)]BF4 in CD3OD was characterized by a
doublet at l 18.2. When the solution was cooled the
signal became broader. At 225 K two new resonances
appeared. Finally at 185 K two doublets of doublets at
l 29.1 (JRh,P=146.1 Hz, JP,P=36.4 Hz) and 7.2
(JRh,P=143.4 Hz) were observed. Following the analy-
sis of the low temperature spectrum of [Rh(threo-1-
OH)(COD)]BF4 detailed in a recent publication [11] the
doublet at low field can be attributed to a phosphorus
atom which is involved in a five-membered chelate ring
being formed by the coordination of one of the hydroxy
groups on the rhodium center. The resonance at high
Scheme 3. Synthesis of [Rh(erythro-1-OH)(COD)]BF4.

64
J. Holz et al. / Journal of Organometallic Chemistry 603 (2000) 61–68
Scheme 4. Alternate interactions of HO groups in [Rh(erythro-1-OH)(COD)]BF4 in CD3OD.
Scheme 5. Formation of diastereomic catalyst–substrate complexes via NBD- and bis-methanol complex.
ABX spectra could be observed characterizing four
phosphorus nuclei of two diastereomeric substrate com-
plexes in a ratio of approximately 1:1. The close chem-
ical shift of all resonances in the range between l 25.6
and 20.1, as well as two small 103Rhꢀ31P coupling
constants of 136.5 and 136.0 Hz [11], indicate the
tripodal coordination of the hydroxyphosphine ligand.
In principle, due to the stereofacial coordination of
prochiral AMe on rhodium the alternate coordination
of both hemilabile HO ligands should produce a spec-
trum characterizing four diastereomeric substrate com-
plexes (A-AMeRe, A-AMeSi, A%-AMeRe, A%-AMeSi). Only
two of them are observed. In agreement with earlier
investigations [11] we suggest that the phosphine trans
to the carbonyl oxygen is involved in the six-membered
OꢀPꢀRh chelate (A-AMeRe, A-AMeSi). The line-shape
analysis indicated that high-field signals exchange with
high-field and the low-field with low-field signals, re-
spectively, and gives proof for the proposed Re/Si-
interconversion.
2.1.3. Hydrogenations
Hydrogenation experiments with achiral [Rh(erythro-
1-OH)(COD)]BF4 were performed with a substrate–
catalyst ratio of 100:1 in methanol with a hydrogen
pressure over the solution of 0.1 MPa. As test sub-
strates (Z)-N-acetylaminocinnamic acid (6a), itaconic
acid (7a) and their methyl esters (6b=AMe, 7b) were
employed. The times measured for the consumption of
50% of hydrogen are summarized in Table 1. Values
measured with the chiral threo catalysts and its
dimethylether are given for comparison. Since the enan-

J. Holz et al. / Journal of Organometallic Chemistry 603 (2000) 61–68
65
Table 1
vidual stable catalyst–substrate complexes does not
allow a reliable conclusion on macroscopically observed
overall rates. Rate constants of preequilibria and rate
determining steps are required in order to enlighten the
special role of the catalytically non-reactive h3-coordi-
nated catalyst–substrate complexes (A-AMe) on the
catalytic cycle. Unfortunately, up to now due to the
great complexity of our model systems chosen the
complete set of data under hydrogenation conditions
were not accessible.
Time (in min) required for 50% consumption of hydrogen with
[Rh(ligand)(COD)]BF4
a
Ligand
Substrate
6a
2.2. Cooperati6e effects on the asymmetric
hydrogenation
6b
7a
7b
erythro-1-OH
threo-1-OH
threo-1-OMe
B2
B2
960
240
B2
B2
–
40
b
c
190
6.0 b
As mentioned in the introduction an accelerating
effect on the hydrogenation of dehydroamino acid
derivatives has been already achieved, when the threo-
1-OH catalyst operated in a nonpolar solvent [11]. In
comparison to the reaction in methanol the asymmetric
hydrogenation proceeded approximately five-fold faster
in methylene chloride without affecting the enantio-
selectivity. The change from methanol to the nonpolar
solvent shifted in the precatalyst the equilibrium from
the h3- to the h2-coordination mode and simulta-
neously an intramolecular hydrogen bond between both
HO groups was formed (Scheme 1, B).
We found that addition of triphenylphosphine oxide
to the solution of the precatalyst [Rh(threo-1-
OH)(COD)]BF4 in CD2Cl2 led to the cleavage of the
intramolecular hydrogen bond in the backbone of the
chiral complex and the concomitant construction of an
intermolecular hydrogen bond between phosphine ox-
ide and one of the HO groups. The other hydroxy
group coordinated on the rhodium. Inspired by this
experiment we anticipated that the PꢁO fragment incor-
porated in an appropriate prochiral substrate could
provide for the same effect by formation of an in-
tramolecular hydrogen bond. For this reason, we have
chosen 3-dimethylphosphono-butenoate (8, Scheme 6)
[16] representing a phosphonate analogue of dimethyl
itaconate.
145 b
3.7 b
a Conditions: 0.01 mmol of precatalyst, 1.0 mmol of substrate in 15
ml of MeOH at 25°C and 0.1 MPa H2 pressure. Times indicated were
measured with an automatically registrating hydrogen consumption
device.
b Times are derived from Ref. [10]. In this reference also the % ee
values are detailed.
c Hydrogenation proceeded only under elevated pressure.
tioselectivities achieved with the latter are not relevant
for the discussion herein, they are omitted from the
table. Browsing the values listed clearly shows that
hydrogenations with the erythro-1-OH complex pro-
ceeded superior. Conversions with this catalyst occured
so fast that the time required for the hydrogenation of
COD has to be taken into consideration [9]. Therefore,
exact specifications of times are not given. It is note-
worthy that hydrogenation times noted with the ery-
thro-1-OH catalyst were even shorter than with the
catalyst bearing only a single hydroxy group [7].
Although in the 31P{1H}-NMR spectra no differences
between catalyst–substrate complexes of erythro-1-OH
and threo-1-OH at low temperature could be found,
which is in contrast to results obtained with precata-
lysts strikingly different consumption rates of hydrogen
were observed. Clearly, only the identification of indi-
Scheme 6. Shift of the equilibrium between solvent complex and catalyst–substrate complex in dependence on the substituent R in CD3OD as
solvent.

66
J. Holz et al. / Journal of Organometallic Chemistry 603 (2000) 61–68
Table 2
the PꢁO group in the major complex could be excluded.
The 31Pꢀ31P coupling constant of 14.0 Hz gave proof
for the trans arrangement of the vinylphosphonate moi-
ety and the concerned phosphine. The large shift differ-
ences of the observed signals indicate that no
interaction between HO groups and metal center takes
place. We suggest that the RhꢀO interaction is sus-
pended due to a intramolecular hydrogen bond between
HO group and PꢁO functionality (D). Unfortunately,
this feature could not be verified by spectroscopy be-
cause a clear differentiation between hydrogen bonds
between PꢁO and HO groups of the ligand (intramolec-
ular) or between PꢁO and HO groups of the solvent
(intermolecular) was not possible.
Besides the pattern of the major catalyst–substrate
complex discussed in detail a second complex with
similar chemical shifts and coupling constants was ob-
served in very low concentration [l 3.0, (PPh2, ddd), l
57.5 (PPh2, dd)]. Apparently, in two other minor cata-
lyst–substrate complexes also detected coordination of
the PꢁO group on the metal took place which was
indicated by rather large 103Rhꢀ31P and 31Pꢀ31P cou-
plings of 19.0 and 6.6 Hz, respectively.
Asymmetric hydrogenation of (E)-methyl 3-dimethoxyphosphoryl-
but-2-enoate a
b
Run Ligand
Solvent
t/2
ee c (%)
(min)
1
2
3
[Rh(threo-1-OMe)(COD)]BF4 Methanol \1000
[Rh(threo-1-OH)(COD)]BF4 Methanol
[Rh(threo-1-OH)(COD)]BF4 CH2Cl2
24 (S)
83 (S)
–
220
No conversion
a For conditions of the hydrogenation compare with Table 1.
b Time required for 50% consumption of hydrogen.
c For the analysis of the hydrogenation product, see Ref. [16].
2.2.1. Spectroscopic studies
In order to study interactions in relevant catalytic
intermediates 31P{1H}-NMR spectra in deuterated
methanol were recorded. When [Rh(threo-1-OMe)-
(MeOH)2]BF4 generated by prehydrogenation of the
NBD complex was treated with an five-fold excess of
(E)-methyl 3-dimethoxyphosphorylbut-2-enoate the bis-
solvent complex was found exclusively which is charac-
terized by a doublet at l 45.4 (JRh,P=199.8 Hz)
(Scheme 6, B). In this complex no intramolecular inter-
action between MeO groups and rhodium occurred
[11]. The observation of the bis-solvent complex gives
proof for the low stability of the diastereomeric cata-
lyst–substrate complexes belonging to it and fits well
to our results obtained with [Rh((R,R)-DIOP)-
(MeOH)2]BF4 and methyl 3-dimethylphosphono-
butenoate as substrate in hand where the relevant bis-
methanol complex was also dominant.
In contrast, by addition of (E)-methyl 3-
dimethoxyphosphorylbut-2-enoate to [Rh(threo-1-OH)-
(MeOH)2]+ the resonances of catalyst–substrate com-
plexes were observed revealing their high thermody-
namic stability. In deuterated methanol a variety of
complexes were formed. The major complex present in
a concentration of more than 70% was characterized by
three sets of signals at l −6.3 (PPh2, ddd, JRh,P=121.0
Hz, JP,P=31.6 Hz, JP,PꢁO=14.0 Hz), l 58.4 (PPh2, dd,
2.2.2. Asymmetric hydrogenations
Results of the hydrogenation are listed in Table 2. As
can be clearly seen with [Rh(threo-1-OMe)(COD)]BF4
very poor conversion took place (run 1). The hydro-
genation product was obtained in 24% ee. In striking
contrast [Rh(threo-1-OH)(COD)]BF4 reduced the
prochiral substrate much faster (run 2). (S)-Methyl
3-dimethoxyphosphoryl-butyrate was obtained in 83%
ee. It is remarkable that by application of the latter in
methylene chloride as solvent no conversion was ob-
served (run 3). We assume that in both solvents an
intramolecular hydrogen bond between one of the HO
groups and the PꢁO group of the substrate is estab-
lished. While in MeOH the other HO group is engaged
by the solvent allowing the oxidative addition of hydro-
gen to proceed; in methylene chloride this HO group
couples on the rhodium center. The h3-coordinated
complexes formed are catalytically inactive and there-
fore interrupt the hydrogenation. This assumption ex-
plains the dramatic reversal of the hydrogenation
activity by changing from N-acetylaminocinnamate to
3-dimethoxyphosphorylbut-2-enoate as substrate in
MeOH and methylene chloride, respectively. With the
former the catalytically inactive h3-complex is domi-
nant in MeOH, whereas in the latter this h3-complex
dominates in methylene chloride.
J
J
P,Rh=159.5 Hz, JP,P=31.6 Hz) and l 34.6 (PꢁO, dd,
P,P=14.0, JP,Rh=2.8 Hz). In principle, (E)-methyl
In conclusion we have shown that the decelerating
effect frequently observed in the asymmetric hydro-
genation with Rh(I) complexes bearing hydroxyphos-
phines as ligands can be fully suspended by the
cooperative effect of a properly placed second hydroxy
3-dimethoxyphosphorylbut-2-enoate affords the possi-
bility of competitive or cooperative binding by the a-
and b-functional groups. However, due to the small
103Rhꢀ31P(O) coupling of 2.8 Hz the coordination of

J. Holz et al. / Journal of Organometallic Chemistry 603 (2000) 61–68
67
group. Obviously, the latter competes with the metal
3.2. (2R,3S)-2,3-O-Cyclohexylidene-1,4-di-O-p-
toluenesulfonyl-butan-1,2,3,4-tetrol (4)
for the hydroxy functionality by formation of an
intramolecular hydrogen bond. The RhꢀOH dissocia-
tion step seems to be essentially for the oxidative addi-
tion of hydrogen on the rhodium center. Surprisingly, a
strong hydrogen bond accepting group like PꢁO incor-
porated in the substrate may also induce the accelera-
tion of the hydrogenation. Simultaneously, a dramatic
increase of the enantioselectivity was observed. Thus,
with (E)-methyl 3-dimethoxyphosphorylbut-2-enoate as
prochiral olefin 83% ee could be achieved with the
Rh(I) catalysts based on the conformationally flexible
threo-1,4-bis(diphenylphosphino)butane-2,3-diol as lig-
and. This represents an increase of ca. 60% compared
to the result obtained with the analogous dimethoxy
catalysts. It has to be verified if the superior enan-
tioselectivities noted several times in the asymmetric
hydrogenation of difunctionalized olefins like (Z)-N-
acetylaminocinnamic acid, itaconic acid and their
methyl esters is also caused by an intramolecular hy-
drogen bond between hydroxy groups of the ligand and
the carboxylic group of the substrate. In principle, our
results clearly demonstrate the high potential of hy-
droxy groups in chiral metal catalysts for tuning of
reactivity and enantioselectivity of asymmetric transfor-
mations. It opens entirely new perspectives for the
enantioselective discrimination of prochiral substrates.
To a stirred solution of the diol 3 (4.60 g, 22.7
mmol) in pyridine (20 ml) was added p-toluenesul-
fonyl chloride (10.0 g, 52.3 mmol) in small portions
at 0°C. The reaction mixture was stirred overnight at
room temperature before it was poured on ice.
The product was extracted with methylene chloride
(3×100 ml) and the organic layers were washed with
5% aqueous sulfuric acid until the water phase stayed
acidic. After washing with water and drying with
Na2SO4 the solvent was removed and the residue
was recrystallized with ethanol to give 4 as white
needles.
Yield: 8.15 g (70%); m.p. 108–110°C; 1H-NMR
(CDCl3) l 7.76 (m, 4H, arom. H), 7.34 (m, 4H,
arom. H), 4.29 (m, 2H, CHO), 4.02 (m, 4H, CH2O),
2.44 (s, 6H, CH3), 1.55–1.30 (m, 10H, (CH2)5); 13C-
NMR (CDCl3) l 145.2 (arom. C), 132.6 (arom. C),
130.0 (arom. C), 128.0 (arom. C), 110.6 (OCO), 73.6
(CHO), 67.1 (CH2OTs), 37.2 (CH2O), 34.5 (CH2O),
24.9 (CH2), 23.8 (CH2), 23.5 (CH2), 21.6 (CH3); MS
(EI, m/z) 481 (M+−C2H5, 27), 467 (M+−C3H7,
95), 91 (C7H+7 , 100); C24H30O8S2 (510.62) calc.: C,
56.45; H, 5.92; S, 12.56; found: C, 56.44; H, 6.00;
S, 12.69.
3.3. (2R,3S)-2,3-O-Cyclohexylidene-1,4-bis-
(diphenylphosphino)-butan-2,3-diol (5)
3. Experimental
3.1. (2R,3S)-2,3-O-Cyclohexylidene-butan-1,2,3,4-tetrol
(3)
A freshly prepared solution of LiPPh2 [from
ClPPh2 (4.3 ml, 23.5 mmol) and Li (0.50 g, 71 mmol
in THF (15 ml)] was added slowly to a solution of
the ditosylate 4 (3 g, 5.88 mmol) in THF (15 ml).
The addition was stopped when the red color re-
mained. After an additional 2 h the THF was re-
moved under vacuum and the residue was treated
with water (10 ml) and methylene chloride (50 ml).
The organic phase was dried with Na2SO4. After
evaporation of the solvent the crude product was
purified by column chromatography (n-hexane–
AcOEt=19:1, Rf=0.5) to give the phosphine 5 as a
colorless syrup.
(+)-2,3-O-Cyclohexylidene-
L
-erythruronic acid
2
(6.00 g, 28.0 mmol) was dissolved in THF (50 ml). The
solution was slowly dropped to a stirred suspension of
LiAlH4 (2.13 g, 56.0 mmol) in THF (100 ml). Then the
reaction mixture was stirred under reflux for 1 h. The
excess of hydride was carefully destroyed with water
(2.2 ml), 15% aqueous NaOH (2.2 ml) and water (6.6
ml). After filtration the precipitated hydroxide was
extracted with boiling methylene chloride overnight.
The combined extracts were dried with anhydrous
Na2SO4 and the solution was concentrated. The residue
was purified by column chromatography (n-hexane–
AcOEt=1:1) to give the crystalline tetrol 3.
Yield: 2.10 g (66%); 31P-NMR (CDCl3) l −20.6;
1H-NMR (CDCl3) l 7.50–7.22 (m, 20H, arom. H),
4.12 (m, 2H, CHO), 2.42 (m, 2H, CH2P), 2.32
(m, 2H, CH2P), 1.61–1.31 (m, 10H, (CH2)5); 13C-NMR
(CDCl3) l 138.6 (d, arom. CꢀP), 138.1 (d, arom.
CꢀP), 133.1–128.3 (arom. C), 108.8 (OCO), 75.4
(m, CHO), 38.3 (CH2O), 35.1 (CH2O), 30.1
(d, 1JC,P=13.1 Hz, CH2P), 25.1 (CH2), 24.0 (CH2), 23.8
(CH2); MS (EI, m/z) 538 (M+, 31), 461 (M+−C6H5,
100), 185 (PPh+2 , 29); C34H36O2P2 (538.61) calc.: C,
75.82; H, 6.74; P, 11.50; found: C, 76.04; H, 6.50;
P, 11.29.
Yield: 4.90
g
(86%); m.p. 43–45°C; 1H-NMR
(CDCl3) l 4.20 (m, 2H, CHO), 3.68 (m, 4H, CH2O),
3.60 (s, 2H, OH, exchangeable with D2O), 1.52–1.32
(m, 10H, (CH2)5); 13C-NMR (CDCl3) l 108.8 (OCO),
76.4 (CHO), 60.6 (CH2OH), 37.3 (CH2O), 34.4 (CH2O),
24.9 (CH2), 23.8 (CH2), 23.4 (CH2); IR (KBr) w˜OH 3232;
MS (EI, m/z) 202 (M+, 22), 171 (M+−CH2OH, 13),
159 (M+−C3H7, 100); C10H18O4 (202.25) calc.: C,
59.39; H, 8.97; found: C, 59.47; H, 8.81.

68
J. Holz et al. / Journal of Organometallic Chemistry 603 (2000) 61–68
3.4. (2R,3S)-1,4-Bis(diphenylphosphino)-butan-2,3-diol
(erythro-1-OH)
Pribbenow and K. Kortus for their skilled technical
assistance. It is a pleasure for us to thank Professor Dr
em. R. Selke for support of this work.
Diphosphine 5 (1.80 g, 3.34 mmol) was stirred in a
anaerobic mixture of methanol (50 ml), water (8 ml)
and methanesulfonic acid (100 ml) under reflux. The
reaction was followed by TLC (CH2Cl2–AcOEt=20:1,
Rf=0.3). After completion of the reaction the solvents
were evaporated and the residue was purified by
column chromatography.
References
[1] T.V. RajanBabu, T.A. Ayers, G.A. Halliday, K.K. You, J.C.
Calabrese, J. Org. Chem. 62 (1997) 6012.
[2] (a) Y. Sun, R.N. Landau, J. Wang, C. LeBlond, D.G. Black-
mond, J. Am. Chem. Soc. 118 (1996) 1348. (b) Y. Sun, J. Wang,
C. LeBlond, R.N. Landau, J. Laquidara, J.R. Sowa Jr., D.G.
Blackmond, J. Mol. Catal. A: Chem. 115 (1997) 495. (c) D.
Heller, S. Borns, W. Baumann, R. Selke, Chem. Ber. 129 (1996)
85. (d) D. Heller, R. Thede, D. Haberland, J. Mol. Catal. A:
Chem. 115 (1997) 273.
[3] (a) H. Brunner, A. Winter, J. Breu, J. Organomet. Chem. 553
(1998) 285. (b) J.M. Brown, P.L. Evans, Tetrahedron 44 (1988)
4905. (c) V.A. Pavlow, E. Klabunovskii, Yu.T. Struchkov, A.A.
Voloboev, A.I. Yanovski, J. Mol. Catal. 44 (1988) 217. (d) D.
Heller, H. Buschmann, Top. Catal. 5 (1998) 159. (e) P. Dierkes,
P.W.N.M. van Leeuwen, J. Chem. Soc. Dalton Trans. (1999)
1519. (f) R. Kadyrov, A. Bo¨rner, R. Selke, Eur. J. Inorg. Chem.
(1999) 705.
Yield: 0.86 g (56%); m.p. 114–116°C; 31P-NMR
1
(CDCl3) l −23.0; H-NMR (CDCl3) l 7.50–7.26 (m,
20H, arom. H), 3.76 (m, 2H, CHO), 2.48–2.32 (m, 6H,
CH2P, OH); 13C-NMR (CDCl3) l 138.2 (d, arom.
CꢀP), 137.5 (d, arom. CꢀP), 133.2–128.5 (arom. C),
72.3 (m, CHO), 30.1 (d, 1JC,P=12.1 Hz, CH2P); IR
(KBr) w˜OH 3475; MS (EI, m/z) 458 (M+, 11), 381
(M+−C6H5, 29), 273 (M+−PPh2, 100), 185 (PPh+2 ,
69); C28H28O2P2 (458.48) calc.: C, 73.35; H, 6.16; P,
13.51; found: C, 73.01; H, 6.47; P, 13.19.
3.5. [Rhodium(erythro-1-OH)(cyclooctadiene)]-
tetrafluoroborate complex
[4] M. Sawamura, Y. Ito, Chem. Rev. 92 (1992) 857.
[5] J. Holz, M. Quirmbach, A. Bo¨rner, Synthesis (1997) 983.
[6] (a) L. Dahlenburg, V. Kurth, J. Organomet. Chem. 585 (1999)
315. (b) D. Carmichael, H. Doucet, J.M. Brown, Chem. Com-
mun. (1999) 261. (c) W. Li, Z. Zhang, D. Xiao, X. Zhang,
Tetrahedron Lett. 40 (1999) 6701.
Diphosphine erythro-1-OH (0.47 g, 1.03 mmol) and
[Rh(COD)2]BF4 (418 mg, 1.03 mmol) were stirred in
THF (8 ml) at r.t. for 1 h. Then the solution was
filtrated and the desired complex was precipitated by
addition of diethylether. Filtration and washing of the
orange powder with diethylether (2×4 ml) yielded the
pure complex.
[7] A. Bo¨rner, A. Kless, R. Kempe, D. Heller, J. Holz, W. Bau-
mann, Chem. Ber. 128 (1995) 767.
[8] (a) J.P. Amma, J.K. Stille, J. Org. Chem. 47 (1982) 468. (b) W.S.
Knowles, W.C. Christopfel, K.E. Koenig, C.F. Hobbs, Studies
of Asymmetric Homogeneous Catalysts, in: E.C. Aleya, D.W.
Meek (Eds.), Catalytic Aspects of Metal Phosphane Complexes,
American Chemical Society, Washington, 1982, pp. 325, 330. (c)
I.D. Kostas, C.G. Screttas, J. Organomet. Chem. 585 (1999) 1.
[9] D. Heller, J. Holz, S. Borns, A. Spannenberg, R. Kempe, U.
Schmidt, A. Bo¨rner, Tetrahedron: Asymmetry 8 (1997) 213.
[10] S. Borns, R. Kadyrov, D. Heller, W. Baumann, A. Spannenberg,
R. Kempe, J. Holz, A. Bo¨rner, Eur. J. Inorg. Chem. (1998) 1291.
[11] S. Borns, R. Kadyrov, D. Heller, W. Baumann, J. Holz, A.
Bo¨rner, Tetrahedron: Asymmetry 10 (1999) 1425.
[12] M. Bu¨hl, W. Baumann, R. Kadyrov, A. Bo¨rner, Helv. Chim.
Acta 82 (1999) 811.
[13] (a) M. Pottie, J. Van der Eycken, M. Vandewalle, Tetrahedron:
Asymmetry 2 (1991) 329. (b) H.-J. Gais, H. Hemmerle, S.
Kossek, Synthesis (1992) 169. (c) J. Holz, A. Kless, A. Bo¨rner,
Synlett (1996) 267.
Yield: 0.49 g (63%); 31P-NMR (CDCl3) l 16.5
(d, 1JRh,P=144 Hz); 1H-NMR (CDCl3) l 7.72–7.33
(m, 20H, arom. H), 4.59 (m, 4H, HꢀCOD), 4.07 (m,
2H, CHO), 3.60 (s, 2H, OH), 2.75–2.10 (m, 12H,
CH2P, CH2); 13C-NMR (DMSO-d6) l 135.7–129.2
(arom. C), 99.2 (m, ꢁCH), 70.9 (m, CHO), 31.9 (d,
1JC,P=20.0 Hz, CH2P), 30.9 (CH2); IR (KBr) w˜OH 3443;
C36H40BF4O2P2Rh (756.38) calc.: C, 57.17; H, 5.33; P,
8.19; Rh, 13.61; found: C, 57.59; H, 5.85; P, 8.43; Rh,
12.99.
Acknowledgements
[14] H.J. Bestmann, U.C. Philipp, Angew. Chem. 103 (1991) 78;
Angew. Chem. Int. Ed. Engl. 30 (1991) 86.
[15] H.B. Kagan, T.P. Dang, J. Am. Chem. Soc. 94 (1972) 6429.
[16] R. Kadyrov, R. Selke, R. Giernoth, J. Bargon, Synthesis 6
(1999) 1056.
We are grateful for the financial support provided by
the Max-Planck-Gesellschaft and the Fonds der
Chemischen Industrie. We wish to thank G. Voß, C.
.
.