Highly efficient enantioselective catalysis in supercritical carbon dioxide using the perfluoroalkyl-substituted ligand (R,S)-3-H2F6-BINAPHOS

Article abstract of DOI:10.1016/S0022-328X(00)00778-6

Perfluoroalkyl-substitution of the aryl groups in the BINAPHOS skeleton was achieved using a synthetic strategy based on the Cu(I)-catalyzed cross coupling of arylmagnesium bromide and F(CF₂)₆(CH₂)₂I. The rhodium complexes of the new ligand (R,S)-3-H²F⁶-BINAPHOS (10) exhibited spectroscopic properties and reactivities similar to those of the unsubstituted parent compounds. The substitution provided a high affinity of the ligand and its complexes for scCO₂ allowing the development of ecologically benign protocols for catalytic asymmetric synthesis and even the spectroscopic detection of catalytically active intermediates. Using this new system, a large variety of substrates were hydroformylated in scCO₂ with rates and enantioselectivities comparable to those of the parent system in benzene solution. At the same time, the CO₂-philic substitution pattern resulted in a significantly higher regioselectivity towards the desired chiral aldehydes. Preliminary results indicated for the first time also a remarkable potential of BINAPHOS-derived ligands for asymmetric hydrogenation reactions. The possibility to develop new work-up schemes for product purification and/or catalyst immobilisation based on scCO₂ as the only medium for catalysis and extraction (CESS process) was experimentally verified using a rhodium catalyst containing 10.

Full text of DOI:10.1016/S0022-328X(00)00778-6

www.elsevier.nl/locate/jorganchem
Journal of Organometallic Chemistry 621 (2001) 130–142
Highly efficient enantioselective catalysis in supercritical carbon
dioxide using the perfluoroalkyl-substituted ligand
(R,S)-3-H2F6-BINAPHOS
Giancarlo Francio`, Klaus Wittmann, Walter Leitner *
Max-Planck-Institut fu¨r Kohlenforschung, Kaiser-Wilhelm-Platz 1, D-45470 Mu¨lheim an der Ruhr, Germany
Received 17 August 2000
Dedicated to Professor Henri Brunner on the occasion of his 65th birthday
Abstract
Perfluoroalkyl-substitution of the aryl groups in the BINAPHOS skeleton was achieved using a synthetic strategy based on the
Cu(I)-catalyzed cross coupling of arylmagnesium bromide and F(CF2)6(CH2)2I. The rhodium complexes of the new ligand
(R,S)-3-H2F6–BINAPHOS (10) exhibited spectroscopic properties and reactivities similar to those of the unsubstituted parent
compounds. The substitution provided a high affinity of the ligand and its complexes for scCO2 allowing the development of
ecologically benign protocols for catalytic asymmetric synthesis and even the spectroscopic detection of catalytically active
intermediates. Using this new system, a large variety of substrates were hydroformylated in scCO2 with rates and enantioselectiv-
ities comparable to those of the parent system in benzene solution. At the same time, the CO2-philic substitution pattern resulted
in a significantly higher regioselectivity towards the desired chiral aldehydes. Preliminary results indicated for the first time also
a remarkable potential of BINAPHOS-derived ligands for asymmetric hydrogenation reactions. The possibility to develop new
work-up schemes for product purification and/or catalyst immobilisation based on scCO2 as the only medium for catalysis and
extraction (CESS process) was experimentally verified using a rhodium catalyst containing 10. © 2001 Elsevier Science B.V. All
rights reserved.
Keywords: Homogeneous catalysis; Enantioselective hydroformylation; Enantioselective hydrogenation; Supercritical carbon dioxide; Fluorinated
ligands
wide range of fascinating applications. Of course, the
most obvious advantage of the use of scCO2 is the
replacement of potentially hazardous organic solvents
such as benzene, methylene chloride, diethyl-ether etc.
with environmentally benign and toxicologically harm-
less carbon dioxide. Working beyond the critical tem-
perature of CO2 provides the attractive option to use
the same medium for synthesis, isolation and purifica-
tion of the product by exploiting the extractive proper-
ties of scCO2. Purification by selective extraction with
supercritical fluids (SFE) is well established in various
technical processes, with decaffeination of green coffee
beans being just the most prominent application [3].
Most recently, the successful utilization of scCO2 to
separate and recycle homogeneous transition metal cat-
alysts from reaction products has been demonstrated
for the first time in our group [4–6]. We refer to such
1. Introduction
Supercritical carbon dioxide (scCO2) has recently re-
ceived considerable attention as a reaction medium with
unique properties for chemical reactions and especially
for metal-catalyzed processes [1,2]. The high miscibility
of scCO2 with many reaction gases can help to avoid
mass transport limitations which are frequently encoun-
tered in gas–liquid phase reactions. Further beneficial
effects may arise from direct interaction of CO2 with
catalytically active metal centers. The high compress-
ibility of supercritical solvents offers density as an
additional process parameter for catalysis opening a
* Corresponding author. Tel.: +49-208-3062500; fax: +49-208-
3062993.
E-mail address: leitner@mpi-muelheim.mpg.de (W. Leitner).
0022-328X/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved.
PII: S0022-328X(00)00778-6

G. Francio` et al. / Journal of Organometallic Chemistry 621 (2001) 130–142
131
phosphite ligand (R,S)-BINAPHOS is known to allow
outstanding levels of enantiocontrol in this reaction,
but the established protocols require application of
ecological and toxicological hazardous organic solvents,
in particular benzene [15]. Initial attempts to replace
these solvents and to use the ligand (R,S)-BINAPHOS
directly in compressed CO2 gave preparatively unsatis-
factory results [16]. Detailed control experiments re-
vealed that the poor solubility of the chiral ligand and
of its rhodium complexes in the supercritical phase was
the limiting factor [16a].
As initially shown using the substituent F(CF2)6-
(CH2)2 (H2F6) [17], perfluorinated groups can act as
solubilizers for aryl phosphines making the correspond-
ing metal complexes sufficiently ‘CO2-philic’ for cataly-
sis in many cases [5,10,18,19]. Based on these findings,
we have set out to synthesize the perfluoroalkyl-substi-
tuted ligand (R,S)-3-H2F6-BINAPHOS (10). Although
the fluorine content of 10 is considerable lower than
that of most ligands typically used in this medium (Fig.
1), we were confident that this new ligand would
provide sufficient solubility for enantioselective cataly-
sis in scCO2. Preliminary studies confirmed that the
new ligand 10 formed soluble and efficient catalyst for
the hydroformylation of vinylarenes in compressed (liq-
uid or supercritical) CO2 [20].
integrated processes in homogeneous catalysis as catal-
ysis and extraction using supercritical solutions (CESS)-
processes.
Nowhere is the consideration of ecological and toxi-
cological issues more important than in the synthesis of
biologically active fine chemicals like pharmaceuticals,
food additives, agrochemicals, or cosmetics. Many of
these materials are chiral and the two mirror images
often exhibit different physiological properties. Hence,
there is an increasing demand for efficient and ecologi-
cally benign methodologies to synthesize such com-
pounds in enantiomerically pure or at least in
non-racemic form. Enantioselective homogeneous catal-
ysis using chiral transition metal catalysts has emerged
as one of the most attractive approaches in asymmetric
synthesis but usually involves the use of large quantities
of organic solvents [7]. Clearly, there is a huge potential
for the application of scCO2 as a reaction medium for
such processes, which has been barely explored up to
now.
Several promising examples of asymmetric catalysis
in scCO2 were reported during the last 5 years including
the hydrogenation of CꢀC [8–10] and CꢀN double
bonds [6], the epoxidation [11] and the cyclopropana-
tion of alkenes [12], and the nickel-catalyzed hydroviny-
lation of styrenes [13]. In contrast, asymmetric
hydroformylation in scCO2 has remained elusive until
recently. Rhodium-catalyzed asymmetric hydroformyla-
tion is of great current interest providing for example
viable routes to important anti-inflammatory drugs like
ibuprofen or naproxen [14]. The chiral phosphine–
In the present contribution, we report details of the
synthesis of 10 and its application in asymmetric hydro-
formylation and hydrogenation reactions. The solution
structure of a rhodium hydrido carbonyl complex of 10
relevant to catalysis was elucidated in scCO2 using
Fig. 1. Fluorine content (wt.%) in ‘CO2-philic’ ligands.

132
G. Francio` et al. / Journal of Organometallic Chemistry 621 (2001) 130–142
Scheme 1.
Scheme 2.
high-pressure NMR spectroscopy. Furthermore, we de-
scribe the utilization of the tuneable solvent properties
of CO2 for product purification and catalyst recycling
with this system.
selective mono-substitution of the bistriflate of binaph-
thol 6 with a diaryl-phosphineoxide through a palla-
dium-catalyzed C–P coupling reaction. We decided to
use this step for the introduction of the perfluoroalkyl
chains via the diaryl-phosphineoxide (4). The key inter-
mediate 4 is readily available from simple starting
materials in good yield using the synthetic strategy
shown in Scheme 1.
Copper-catalysed reaction of F(CF2)6(CH2)2I with
the Grignard reagent 1 obtained from meta-dibromo-
benzene provided the perfluoroalkyl-substituted aryl-
bromide 2 [17,21]. Lithiation with n-BuLi and reaction
with dichloro-diethylamino-phosphine produced the di-
aryl-diamino-phosphine (3) that was not isolated but
hydrolized directly to obtain the desired phosphine-ox-
2. Results and discussion
2.1. Synthesis of (R,S)-3-H2F6–BINAPHOS (10)
The synthesis of (R,S)-3-H2F6–BINAPHOS (10) fol-
lows a reaction sequence that is largely adapted from
the synthesis of (R,S)-BINAPHOS [15], although some
steps needed significant changes of the reaction condi-
tions (Scheme 2). The key step of this sequence is the

G. Francio` et al. / Journal of Organometallic Chemistry 621 (2001) 130–142
133
ide (4). As with the parent phenyl derivative, the Pd-
catalyzed coupling of 4 with 6 occurred very selectively,
although higher temperatures and larger catalyst and
base loadings were required to obtain the mono-sub-
stitued intermediate 7 in excellent yield [22]. Hydrolysis
with lithium hydroxide in aqueous THF and subse-
quent reduction with trichlorosilane afforded the corre-
sponding phosphine-alcohol (9). The synthesis was
completed by the reaction of 9 with the chlorophos-
phite derived from S-binaphthol. The overall yield for
the desired product (R,S)-3-H2F6–BINAPHOS (10)
based upon 6 was 42%.
troscopy appears not to be sensitive enough to reflect
this electronic change in the case of 11, where only two
chains are introduced compared to the parent com-
pound (Table 1, entry 3 and 4) [25].
Despite the increasing number of successful applica-
tions of phosphine complexes as catalysts in scCO2, the
direct spectroscopic detection of catalytically active spe-
cies has not yet been reported for this class of com-
pounds under supercritial conditions [26]. Therefore the
reaction of 11 with synthesis gas (20 bar) in scCO2 was
investigated by high pressure NMR using a 5 mm tube
made from single crystal sapphire [27]. This study re-
vealed clean formation of the five-coordinated hydrido
carbonyl complex [H(CO)2Rh(10)] (12) which was read-
2.2. Synthesis and high pressure NMR in6estigation of
catalyst precursors
1
ily identified on the basis of its 31P{1H}- and H-NMR
data (Fig. 2, Table 2). As for the parent BINAPHOS
complex, this species can be expected to enter directly
into the catalytic cycle by dissociation of one of the
carbonyl ligands [15].
The complex [{(R,S)-3-H2F6–BINAPHOS}Rh(acac)]
(11) was synthesized by adding equimolar amounts of
ligand 10 to [Rh(acac)(CO)2] in THF solution at −
60°C according to an established procedure [23]. The
spectroscopic data of 11 and the non-fluorinated parent
complex (Table 1) are virtually identical even for the
103Rh nucleus, which is known to provide a very sensi-
tive probe for variations in the steric and electronic
properties of coordinating ligands [24]. For example,
the difference of 23 ppm in the 103Rh spectrum of
complexes derived from the chelating ligands dppe and
(3-H2F6)-dppe (Table 1, entry 1 and 2) indicates a small
but significant electron donating effect of the four
perfluoroalkyl chains. In contrast, 103Rh-NMR spec-
In general, the NMR data of five-coordinated 12 in
scCO2 are very similar to those of the unsubstituted
parent complex in benzene (Table 2), albeit some differ-
ences are observed for example in the coupling con-
stants JPH. It is interesting to note that these data of
complex 12 are in fact more similar to the correspond-
ing (R,S)-BIPHEMPHOS complex, which gives re-
gioselectivities similar to 10 and higher than
(R,S)-BINAPHOS in the hydroformylation of vinylare-
nes (see Section 2.3). The detection of the hydride
signal of 12 is quite remarkable because it indicates that
Table 1
Comparison of NMR data for Rh-complexes containing chelating aryl phosphines substituted with polyfluoroalkylchains and their unsubstituted
parent compounds (acac=acetylacetonate; hfacac=hexafluoroacetylacetonate)
Entry Compounds
l (PB) (ppm)
l (PA) (ppm)
JRhP (Hz) JRhP (Hz) JP (Hz)
l (103Rh) (ppm)
B
A
AB
1
2
3
4
[(dppe)Rh(hfacac)] a
72.1
71.8
49.3
49.6
196
196
175
175
438
461
273
271
[{(3-H2F6)-dppe}Rh(hfacac) b
[{(R,S)-BINAPHOS}Rh(acac)] c
[(10)Rh(acac)], 11
162.2
161.9
332
330
84
84
a Ref. [23].
b Ref. [17].
c Ref. [15].
Fig. 2. The hydride region of the 1H-NMR spectrum of 12 obtained in scCO2.

134
G. Francio` et al. / Journal of Organometallic Chemistry 621 (2001) 130–142
Table 2
NMR data of the complexes [H(CO)2Rh(ligand-PA,PB)]
Ligand-PA,PB
Phosphine
Phosphite
Hydride
l (PB)
J (PBH)
J (RhPB)
l (PA)
J (PAH)
J (RhPA)
J (PAPB)
l (1H)
J (RhH)
(ppm)
(Hz)
(Hz)
(ppm)
(Hz)
(Hz)
(Hz)
(ppm)
(Hz)
(R,S)-BINAPHOS a
26.9
28.0
25.0
23
16
17
119
124
120
184.9
183.6
183.5
160
170
173
182
181
177
40
38
43
−8.85
−9.56
−8.85
9.8
9.2
8.7
3-H2F6-(R,S)-BINAPHOS b
(R,S)-BIPHEMPHOS a
a Ref. [15]; solvent C6D6.
b T=40°C; solvent CO2–THF-d8.
Scheme 3.
there is no measurable interaction of CO2 with the
rhodium hydride bond even under high pressure of
CO2. This lack of reactivity of 12 towards CO2 is in
sharp contrast to the situation observed with the
isoelectronic complex [{Ph2P(CH2)3PPh2}2RhH], which
inserts CO2 to give the corresponding formate complex
even at ambient pressure [28]. The formation of phos-
phine rhodium hydride intermediates that do not inter-
act with CO2 was recently also inferred for
enantioselective hydrogenation in scCO2 using indirect
mechanistic tools [10].
highly enantioselective and regioselective hydroformyla-
tion of vinylarenes in scCO2 [20]. The results summa-
rized in Table 3 demonstrate that this conclusion is
valid for a broad range of substrates as depicted in
Scheme 3.
The catalytic system was conveniently prepared in
situ by mixing THF solutions of the complex
[Rh(acac)(CO)2] and the ligand 10 in a window-
equipped high pressure reactor. A ratio of 10/Rh=2:1
was found to be sufficient to ensure maximum values of
enantioselectivity, whereas 4:1 is typically required with
the parent BINAPHOS ligand. After stirring the com-
bined THF solutions for ca. 5 min, the organic solvent
was evaporated and the substrate was introduced fol-
lowed by synthesis gas and CO2 (see Section 4 for
details). The final reaction temperature was reached
within less than 15 min, ensuring that no significant
conversion occurred before temperature equilibration.
2.3. Rhodium-catalyzed asymmetric hydroformylation
and hydrogenation using (R,S)-3-H2F6-BINAPHOS
(10) in sCO2
Our initial results using 10 as a ligand for catalysis in
scCO2 revealed that this system allows for the first time

G. Francio` et al. / Journal of Organometallic Chemistry 621 (2001) 130–142
135

136
G. Francio` et al. / Journal of Organometallic Chemistry 621 (2001) 130–142
and for rhodium-catalyzed enantioselective hydro-
formylation in scCO2 are significantly higher than
those achieved with unsubstituted BINAPHOS in ben-
zene where comparable data are available [29]. In pre-
vious work, it has been unambiguously shown that the
regioselectivity of hydroformylation reactions using
ligand-free cobalt [26a,30] or rhodium [5] catalysts can
be affected by changing the reaction medium from
liquid organic solvents to scCO2. On the other hand,
the regioselectivity of hydroformylation using ligand-
modified catalysts is also highly sensitive to very subtle
structural variations of the phosphines [31]. As men-
tioned above, the increase in regioselectivity with 10
was found to be independent of the phase behavior
and remained pertinent also in conventional hydro-
formylation solvents such as benzene (Table 3) [20].
The higher regioselectivities must therefore be at-
tributed to the substitution pattern of the CO2-philic
ligand. Although steric effects cannot be ruled out at
this point [32], the increase in selectivity can be fully
rationalized on basis of a small electron-donating ef-
fect of the (CH2)2(CF6)2F substituent (see Section 2.2),
applying the recent model of Casey and coworkers
[31b]. Since the PAr2 units of 10 and BINAPHOS
occupy an equatorial position in the five-coordinate
hydrido carbonyl intermediate, the higher preference
for the formation of the branched aldehydes 10 is in
accord with a slightly more electron-rich nature of this
donor group in 10 as compared to the unmodified
BINAPHOS ligand.
With the notable exception of 2-vinylnaphtalene
(13d, see Section 2.4), fully homogeneous reaction
mixtures were observed during hydroformylation un-
der the standard conditions of Table 3. Typically, car-
bon dioxide densities were adjusted to d(CO2)\0.55 g
ml−1 at reaction temperatures T=40–60°C (critical
density dc(CO2)=0.47 g ml−1 at Tc=31°C), with ini-
tial pressures of synthesis gas at room temperature
ranging from 40 to 60 bar. For the vinyl arenes 13a–d
the reaction rates and selectivties of the homogeneous
catalyst 10–Rh system in scCO2 compare well to those
of the BINAPHOS-derived catalyst in benzene. In the
case of styrene (13a), the phase behaviour was found
to have little influence on the regio- and enantioselec-
tivity [20] and the high selectivties are retained also in
benzene solution (Table 3, entry 1 and 2). Similarly,
the overall selectivities during hydroformylation of 13d
were largely independent of the number of phases,
although the rate was of course lower at the much
lower concentrations required for single phase condi-
tions (Table 3, entry 6 and 7). The homogeneous hy-
droformylation of vinylacetate (13e), however,
represents an example where the use of scCO2 as a
solvent improves the enantioselectivity significantly as
compared to benzene (Table 3, entry 8 and 9). Acetic
acid 1-methyl-2-oxo-ethyl ester (14e) was obtained in
scCO2 with excellent regioselectivity and an ee\95%,
the highest ee ever obtained with this substrate [29].
The homogeneous hydroformylation of acetic acid
allyl ester (13f) proceeded also smoothly in scCO2 at
60°C, but with a disappointingly low ee (Table 3,
entry 11). Lowering the reaction temperature to 42°C
lead to a dramatic increase in enantioselectivity, but
the reaction rate decreased strongly at the same time
(Table 3, entry 10). A similar influence of the reaction
temperature was observed with indene 13h (Table 3,
entry 13 and 14), which confirmed to be a difficult
substrate for hydroformylation in scCO2 just like in
conventional solvents [14,29].
Although the BINAPHOS-ligand has found a num-
ber of highly interesting applications in asymmetric
catalysis, there are currently no reports in literature
about the use of this ligand system in hydrogenation
reactions [33]. This may not be surprising, as it is
often assumed that different design principles are re-
quired for chelating ligands to be highly efficient in
either rhodium-catalyzed asymmetric hydroformylation
or hydrogenation. Our recent experience with the C1-
We also attempted to use the novel catalytic system
for desymmetrization of the norbornadiene skeleton
(13g) by asymmetric hydroformylation. A 30/70 mix-
ture of mono- and bis-hydroformylated products was
formed under standard reaction conditions in a single
homogeneous phase (Table 3, entry 12). The first hy-
droformylation proceeds with virtually no enantio-dis-
crimination (ee of 16gB1%) and the diastereo-
selectivity of the second hydroformylation is also very
small, yielding a mixture of meso-product 15g and
chiral product 14g of 59:41. Nevertheless, the bis-alde-
hyde 14g was obtained with an ee of 20.6%, indicating
that a significant enantio-discrimination occurs during
formation of the chiral bis-hydroformylated product.
The regioselectivities in favor of the branched chiral
aldehydes using perfluoroalkyl-substituted 10 as a lig-
symmetric
phosphine–phosphoramidite
ligand
QUINAPHOS [34] indicated that this paradigm may
be less restricting than anticipated and prompted us to
explore the potential of ligand 10 also for hydrogena-
tion reactions in scCO2. We were very pleased to find
that addition of 1.1 equivalents of ligand 10 to the
cationic rhodium complex [Rh(cod)2]BF4 formed an
active hydrogenation catalyst that gave very high lev-
els of enantiocontrol in the asymmetric hydrogenation
of 2-acetamido methyl acrylate (17a) and dimethyl ita-
conate (17b) in compressed CO2 as a solvent (Table 4).
In contrast to previous systems [6,8,10] no exchange of
the BF4 anion was necessary to obtain a highly effi-
cient asymmetric hydrogenation catalyst based on 10
under these conditions.

G. Francio` et al. / Journal of Organometallic Chemistry 621 (2001) 130–142
137
Hydrogenation of the dehydroamino acid 17a (Table
4, entry 1) occurred in a biphasic system containing a
liquid phase and a gas phase. The formation of the
liquid phase at temperatures below the melting point of
17a (m.p. 52–54°C) can be rationalized using the same
arguments as for 13d (Section 2.4) [36]. The alanine
derivative 18a was formed in quantitative yield with an
ee of 97.2% in favor of the R enantiomer. In the
hydrogenation of dimethyl itaconate (17b) in scCO2
(Table 4, entry 2) the reaction mixture was initially
opaque but became transparent and homogenous after
a few minutes. The ee under these conditions was
slightly lower than using the same catalytic system in
CH2Cl2 at lower temperatures (Table 4, entry 3). These
preliminary results underline the broad utility of the
BINAPHOS skeleton in asymmetric catalysis and its
general compatibility with the use of compressed CO2
as a solvent.
though the reaction temperature was held just around
or even below the melting points of the organic compo-
nents, a clear liquid phase was present rather than a
viscous melt or a solid at the bottom of the reactor. The
formation of the liquid phase results from melting point
depression caused by the solubility of CO2 in the or-
ganic compounds. This effect has been utilized recently
for solventless hydroformylation of 13d using subcriti-
cal CO2 and a non-chiral rhodium catalyst [36]. Subcrit-
ical CO2 cannot dissolve substrates or catalysts and its
only role is to provide a liquid phase at low tempera-
tures in this case. Under our reaction conditions how-
ever, the high density scCO2 phase in the upper part of
the reactor can be expected to dissolve at least parts of
the reaction components [37]. Indeed, two different
crops of product could be isolated when the work-up of
the reaction mixture was carried out carefully. Upon
cooling the reactor to 0°C before venting, visual inspec-
tion of the reactor content showed that the liquid phase
on the bottom solidified and at the same time a white
solid precipitated from the supernatant CO2 phase,
completely covering the walls and the lid inside the
autoclave. After venting and opening the reactor, the
two materials could be mechanically separated (bottom
layer/precipitate=ca. 3:1 by weight) and analyzed indi-
vidually. Both layers were found to contain exclusively
the two aldehydes 14d and 15d as organic products and
the enantiomeric excess of 14d was identical within
experimental error in both samples. However, the pre-
cipitate showed a significant and reproducible enrich-
ment in favor of the branched chiral aldehyde
(14d/15d=98.0:2.0) as compared to the bottom layer
(14d/15d=90.9:9.1).
2.4. Combination of catalysis and extraction using
supercritical solutions (CESS process): product
purification and catalyst recycling in rhodium-catalyzed
hydroformylation using (R,S)-3-H2F6-BINAPHOS (10)
Among the various potential benefits often discussed
for the use of scCO2 in homogeneous catalysis, the
development of ‘solventless’ syntheses including novel
work-up and purification procedures is particularly at-
tractive from a practical point of view [4–6,13,35]. This
may be especially useful for high-boiling products,
where distillation is either impracticable or puts too
much thermal stress on sensitive catalysts. The extrac-
tive properties of scCO2 [3] have an interesting potential
for enhanced protocols for product purification and
catalyst immobilization in enantioselective hydroformy-
lation using rhodium catalysts of (R,S)-3-H2F6-BI-
NAPHOS (10).
The enrichment with chiral product in the upper
phase results from the preferential solubility of the
branched aldehyde over the linear product in supercrit-
ical carbon dioxide [38], as demonstrated unambigu-
ously by a control experiment. In this experiment, a
sample (420 mg) with the composition 14d/15d=
90.9:9.1 was placed in the autoclave, pressurized with
CO2 (d=0.83 g ml−1), and heated to 45°C for 5 min.
In contrast to all other substrates, 2-vinyl-naphtha-
lene (13d) did not form a homogeneous reaction mix-
ture under typical hydroformylation conditions, but
this did not prevent efficient catalysis with 10/Rh. Al-
Table 4
Enantioselective hydrogenation of 17a and 17b
Entry
Substrate
Solvent
d (CO2) (g cm−3
)
T (°C)
pH2 (bar)
t (h)
Conversion (%)
ee (%)
1
2
3
17a
17b
17b
CO2
CO2
CH2Cl2
0.74
0.74
40
40
23
30
30
30
16
16
24
\99
\99
\99
97.2 (R)
96.6 (S)
99.4 (S)

138
G. Francio` et al. / Journal of Organometallic Chemistry 621 (2001) 130–142
Table 5
Rhodium-catalysed asymmetric hydroformylation of styrene (13a) applying the CESS process
Run S/Rh T (°C) PH /CO (bar) t (h) CO2 for extraction a
Conversion b (%) Regioselectivity (%) ee (%) Rh leaching (ppm)
2
T (°C) p (bar) V (l)
1
2
3
4
5
6
1000 55
1000 55
2000 50
1500 54
1400 55
3000 40
1500 50
1000
19
34
40
40
38
61
50
17
15
65
15
15
114
16
41
63
60
61
61
59
60
60
94
100
105
103
107
115
104
200
170
270
200
128
213
173
82.6
97.4
99.9
92.4
95.5
98.6
95.1
89.4
90.9
92.0
92.2
91.2
91.5
92.9
91.4
92.4
86.6
90.4
90.2
82.1
77.6
66.4
70.2
88.2
0.97
0.78
0.45
0.36
1.94
0.95
0.41
7
8 c
a Flow-rate about 1 l min−1, all volumes at ambient conditions.
b Quantitative recovery.
c One additional equivalent of ligand was added.
Then the autoclave was cooled down to 0°C and
vented, whereby the same formation of two different
product phases was observed as described above. The
amount and composition of the two materials proved
also largely identical to the products from the catalytic
reaction (bottom layer: 270 mg, 14d/15d=90.0:10.0;
precipitate: 120 mg, 14d/15d=96.0:4.0). This finding
demonstrates nicely the potential for the use of scCO2
not only as solvent during catalysis, but also as medium
for subsequent purification processes of the reaction
products.
The density-tuneable solvent capacity of scCO2 pro-
vides also the possibility to separate and recycle the
rhodium catalyst based on 10 in active and selective
form as demonstrated for the enantioselective hydro-
formylation of styrene 13a. After each run, the reaction
mixture was cooled to room temperature and CO2 was
vented partly to reduce the density of the solvent. The
mixture was then re-heated, resulting in a two-phase
system consisting of a liquid phase and a compressed
gas phase. As judged from the coloration, the metal-
containing species were contained preferentially in the
liquid phase, but the organic products can be expected
to distribute significantly between both phases (vide
supra). The reactor was now purged from the bottom
with CO2 at constant temperature and pressure, strip-
ping away the products and leaving behind the catalyst
for subsequent use.
a Rh content ranging from 0.36 to 1.94 ppm as deter-
mined from atomic absorption measurements. More-
over, the 10–Rh system was still active after a total
catalytic turnover of more than 12 000 TON/Rh, albeit
with a moderate decrease in selectivity. The decrease in
enantiomeric excess of 14a may be attributed at least
partly to racemization of the product under the experi-
mental conditions [16a], as indicated by the exception-
ally low ee resulting from the long contact time chosen
in run six. The initial high level of enantioselectivity
could be restored by adding only one additional equiv-
alent of ligand in the last run.
3. Conclusions
The results of the present paper illustrate the remark-
able potential of supercritical carbon dioxide (scCO2) as
a reaction medium for asymmetric synthesis using chi-
ral organometallic catalysts under environmentally be-
nign conditions. As shown here for the BINAPHOS
framework, well established and highly efficient ligand
systems containing PAr2 donor groups can be readily
adjusted to this non-conventional solvent by modifica-
tion with perfluoroalkyl substituents whereby the syn-
thetic effort to assemble the chiral ligand framework is
largely similar to that of the parent compound. The
new ligand (R,S)-3-H2F6-BINAPHOS forms highly effi-
cient enantioselective rhodium catalyst for hydroformy-
lation and (as shown for the first time for
BINAPHOS-derivatives) also for hydrogenation.
Choosing the F(CF2)6(CH2)2 group as the CO2-philic
solubilizers results only in a small structural modifica-
tion of the donor group, but this may still induce a
measurable change in selectivity during the complex
framework of the catalytic cycle. On the other hand,
there is mechanistic evidence emerging from this paper
and a related study [10] that the solvent CO2 does not
Eight successive runs were performed according to
this procedure, using 13a under standard conditions
and intervened by the extraction step and by reactants
charging (Table 5). No apparent loss in selectivity or
activity occurred up to the third run. The next three
runs saw a slow decrease in enantioselectivity, but
regioselectivity and conversion remained uniformly
high. Even under these not-optimised simple experi-
mental conditions, the CESS-procedure allowed quanti-
tative recovery of the solvent-free reaction product with

G. Francio` et al. / Journal of Organometallic Chemistry 621 (2001) 130–142
139
interfere directly with the catalytic cycle in hydroformy-
lation and hydrogenation reactions using chiral
rhodium–phosphine catalysts. In view of the high reac-
tivity of many metal hydrides towards CO2 [39], which
is manifested not the least in efficient catalytic systems
for hydrogenation of scCO2 itself [40], this inert behav-
ior is still somewhat unexpected and may not be
universal.
mass spectra were recorded on a Finnigan MAT 95
instrument. GC measurements were carried out on
Carlo Erba 5300 and Hewlett–Packard 5890 instru-
ments. Optical rotations were measured on a JASCO
DIP-360 spectometer. Rhodium content was deter-
mined by atomic absorption spectroscopy (AAS).
4.2. General procedure for the rhodium-catalyzed
hydroformylation of olefins in scCO2
Even in the absence of any chemical interaction, the
physico-chemical properties of scCO2 can provide a
number of potential benefits for enantioselective cataly-
Hydroformylation experiments were carried out in a
window-equipped stainless-steel high-pressure reactor
(V=10 ml). The reactor was charged under an atmo-
sphere of argon with complex [Rh(acac)(CO)2] (1×102
ml, 0.02 M in THF) and the appropriate amount of
ligand 10 to adjust the desired ratio of 10/Rh. After
stirring the reaction mixture for 5 min, the solvent was
removed in vacuo and then the substrate (typically 2–4
mmol) was introduced under an atmosphere of argon.
The reactor was pressurised with a 1:1 mixture of CO
and H2 and then filled with a weighed amount of CO2
by means of a compressor. The reaction mixture was
heated and stirred with a PTFE stirring bar for the
desired reaction time. After cooling to 0°C, the reactor
was carefully vented. The reaction mixture was col-
lected by extraction of the reactor content with ben-
sis. Using asymmetric hydroformylation with
a
rhodium catalysts based on 10 as a model reaction, we
have laid particular emphasis on its simultaneous use as
a medium for reaction and work-up in the present
study. We were able to confirm that the selective and
density-dependent solubility properties of scCO2 can
lead to enrichment of the desired chiral regioisomer in
the product stream and can also be used to separate
and recycle the precious catalyst in active and selective
form. Even with simple equipment and in non-opti-
mised batch-wise operation, remarkable separation effi-
ciency was observed and rhodium contamination of the
product could be limited to below 1 ppm. More sophis-
ticated work-up schemes including continuous pro-
cesses can be readily envisaged and seem highly
promising on basis of the present developments.
1
zene-d6 and analyzed both by H-NMR spectroscopy
and, after filtration through a short path of silica, by
GC analysis in order to determine conversion and
regioselectivity. The enantiomeric excess of the
branched aldehydes 14a–h was determined by GC anal-
ysis using a diethyl-tert-butylsilyl-b-cyclodextrin capil-
lary column (IVADEX 7).
4. Experimental
4.1. General
Caution: working with highly compressed gases must
be carried out only using suitable equipment and under
appropriate safety precautions.
4.3. Hydroformylation of styrene in scCO2 combined
with product extraction (CESS process)
All manipulations involving air- and moisture-sensi-
tive compounds were carried out using standard
Schlenk techniques under argon. Solvents and olefins
were dried, purified, and degassed according to litera-
ture protocols and stored under argon. The gases CO/
H2 (1:1), argon (99.995%), and CO2 (99.9995%) were
purchased from Messer Griesheim. Compounds 1-
bromo - 3 - (3,3,4,4,5,5,6,6,7,7,8,8,8 - tridecafluoro-octyl)-
benzene [21], (R)-2,2%-bis[trifluoromethanesulfonyl)oxi]-
1,1%-binaphthalene (6) [41], and (S)-4-chloro-3,5-dioxa-
After performing the catalysis, the reaction vessel
(V=100 ml) was cooled down to room temperature
(r.t.) and slowly depressurized until a CO2 density of
about 0.3 g ml−1 was reached, at which point the vessel
was warmed again to about 60°C (p$100 bar). Fresh
CO2 was pumped into the vessel by means of a com-
pressor through a needle valve located at the bottom of
the vessel and product-enriched CO2 was released with
similar flow rate from a valve located on the top of the
vessel through a cold trap, keeping the pressure con-
stant during the extraction. The extraction was carried
out until no more liquid phase was visually detectable
and in all cases, quantitative recovery of the product
mixture was achieved. After the extraction was com-
pleted, the autoclave was filled again with fresh sub-
strate, pressurized with synthesis gas and then filled
with CO2 until a total pressure at reaction temperature
of 19891 bar was obtained.
4-phospha-cyclohepta[2,1-a;3,4-a%]dinaphthalene
[42]
were synthesized according to literature procedures. All
other compounds are commercially available and were
used without further purification. NMR spectra were
recorded on Bruker AC 200, AMX 300, ARX 400, and
AMX 600 spectrometers. Chemical shifts are given in
ppm relative to Me4Si (1H, 13C), 85% H3PO4 (31P), or
CFCl3 (19F) as external standard. The 103Rh shifts are
given relative to ](103Rh)=3.16 MHz and were repro-
ducible within 1 ppm [23]. Electron ionization (70 eV)

140
G. Francio` et al. / Journal of Organometallic Chemistry 621 (2001) 130–142
4.4. General procedure for hydrogenation experiments
in scCO2
122.2 (m, 2F, CF2), −123.2 (s broad, 2F, CF2), −
123.8 (s, 2F, CF2), −126.6 (m, 2F, CF2). MS (EI): m/z
894.
The hydrogenation experiments were performed in
the same reactor (V=10 ml) using a similar procedure
as that employed for the hydroformylation experi-
ments. In this case the catalyst was formed in situ from
[Rh(cod)2]BF4 (1×102 ml, 0.02 M in CH2Cl2) and the
ligand 10 (1×102 ml, 0.022 M in CH2Cl2). After stirring
the catalyst solution for 5 min, the solvent was removed
in vacuo and then the substrate (2 mmol) was intro-
duced under an atmosphere of argon. The reactor was
pressurized with H2 and then filled with a weighed
amount of CO2 by means of a compressor. The reaction
mixture was heated and stirred with a PTFE stirring
bar for 16 h. After cooling to 0°C, the reactor was
carefully vented. The reaction mixture was collected by
extraction of the reactor content with toluene, filtered
through a short path of silica, and analyzed by GC.
The enantiomeric excess was determined by GC analy-
sis using a b-cyclodextrin and a trifluoroacetyl g-cy-
clodextrin capillary column for 18a and for 18b,
respectively.
4.6. Synthesis of (R)-trifluoro-methanesulfonic acid
2%-{bis-[3-(3,3,4,4,5,5,6,6,7,7,8,8,8-tridecafluoro-octyl)-
phenyl]-phosphinoyl}-[1,1%]binaphthalenyl-2-yl ester (7)
Palladium diacetate (150 mg, 0.67 mmol), 1,3-bis-
(diphenylphosphino)-propane (290 mg, 0.68 mmol),
(R)-6 (2.295 g, 4.17 mmol), and 4 (7.5 g, 8.35 mmol)
were placed in a Schlenk tube together with dimethyl-
sulfoxide (25 ml) and diisopropylethylamine (15 ml).
The reaction mixture was heated to 120°C for 12 h.
After cooling to r.t., the reaction mixture was concen-
trated under vacuo to obtain a dark red residue. The
residue was solubilized in ethylacetate and washed twice
with saturated Na2CO3 solution. The organic phase was
dried over MgSO4 and concentrated under reduced
pressure. The residue was purified by chromatography
on silica (hexane–ethylacetate 3:1). Yield: 91% (white
solid).
31P-{1H}-NMR (CDCl3): l=28.2 (s). 1H-NMR
(CDCl3): l=8.03–7.80 (m, 4H, Ar–H), 7.66–7.52 (m,
2H, Ar–H), 7.46–7.19 (m, 11H, Ar–H), 7.11–7.03 (m,
2H, Ar–H), 6.88 (m, 1H, Ar–H), 2.83–2.71 (m, 4H,
CH2), 2.36–2.07 (m, 4H, CH2). 19F-{1H}-NMR
(CDCl3): l= −75.3 (s, 3F, OS(O2)CF3), −81.1 (m,
6F, CF3), −114.8 (m, 4F, CF2), −122.1 (m, 4F, CF2),
−123.1 (m, 4F, CF2), −123.7 (s, 4F, CF2), −126.4
(m, 4F, CF2). LCMS (ESIpos): 1295 [M+H].
4.5. Synthesis of bis-[3-(3,3,4,4,5,5,6,6,7,7,8,8,8-
tridecafluoro-octyl)-phenyl]-phosphinoxide (4)
A pentane solution of n-BuLi (21 ml, 1.7 M) was
added dropwise at −25°C to a solution of 1-bromo-3-
(3,3,4,4,5,5,6,6,7,7,8,8,8 - tridecafluoro - octyl) - benzene
(18.0 g, 3.6×10−3 mol) in diethyl-ether (200 ml). The
mixture was warmed up to 0°C and stirred for 30 min
at this temperature. Diethylamino–dichlorophosphine
(2.6 ml, 0.017 mol) was added at −40°C via syringe to
the resulting intensive yellow solution and then the
reaction mixture was slowly allowed to warm to r.t.,
whereby large amounts of precipitate (presumably
LiCl) were formed. After 2 h, a 31P-{1H}-NMR (THF-
d8–diethyl-ether) spectrum of the supernatant reaction
mixture showed only one peak (l=62.6 s) correspond-
ing to bis-[3-(3,3,4,4,5,5,6,6,7,7,8,8,8-tridecafluoro-oc-
tyl)-phenyl]-diethyl-amino-phosphine (3). The reaction
mixture was then treated with 5 ml of HClconc under
vigorous stirring. The organic phase was separated,
washed with saturated NaCl solution (2×20 ml) and
saturated NaHCO3 solution (2×20 ml), and then dried
over MgSO4. The solvents were removed under reduced
pressure leading to a pale yellow oil. Spectroscopically
pure product was obtained by crystallization from pen-
tane as a colorless microcrystalline compound (yield:
64%).
4.7. Synthesis of (R)-2%-{bis-[3-(3,3,4,4,5,5,6,6,7,7,-
8,8,8-tridecafluoro-octyl)-phenyl]-phosphinoyl}-[1,1%]-
binaphthalenyl-2-ol (8)
A saturated LiOH solution (30 ml) was added to a
solution of 7 (4.8 g, 3.7 mmol) in THF (30 ml) and the
resulting biphasic mixture was stirred vigorously
overnight. The reaction mixture was acidified (pH 2)
with concentrated HCl, extracted with ethylacetate
(2×40 ml), dried over MgSO4, and concentrated under
reduced pressure. The resulting pale yellow solid was
purified by chromatography on silica (hexane–ethylac-
etate 2:1). Yield: 69% (white solid).
31P-{1H}-NMR (CDCl3): l=30.4 (s). 1H-NMR
(CDCl3): l=7.94–7.80 (m, 3H, Ar–H), 7.74 (m, 1H,
Ar–H), 7.59 (m, 1H, Ar–H), 7.56–7.32 (m, 6H, Ar–
H), 7.22–6.99 (m, 5H, Ar–H), 6.90 (m, 1H, Ar–H),
6.71–6.59 (m, 2H, Ar–H), 6.40 (m, 1H, Ar–H), 3.04–
2.93 (m, 2H, CH2), 2.56–2.26 (m, 4H, CH2), 2.17–1.92
(m, 2H, CH2).
31P-{1H}-NMR (CDCl3): l=21.3 (s). 1H-NMR
(CD2Cl2): l=7.94 (d, 1JPH=482.6 Hz, 1H, P–H),
7.57–7.32 (m, 8H, Ar–H), 2.94–2.83 (m, 4H, CH2),
2.43–2.20 (m, 4H, CH2). 19F-{1H}-NMR (CD2Cl2):
l= −81.5 (m, 3F, CF3), −114.9 (m, 2F, CF2), −
19F-{1H}-NMR (CDCl3): l= −81.1 (m, 3F, CF3),
−114.6 (m, 1F, CF2), −114.9 (m, 1F, CF2), −122.1
(m, 2F, CF2), −123.1 (m, 2F, CF2), −123.6 (m, 2F,
CF2), −126.3 (m, 2F, CF2). LCMS (ESIpos): 1163
[M+H].

G. Francio` et al. / Journal of Organometallic Chemistry 621 (2001) 130–142
141
4.8. Synthesis of (R)-2%-{bis-[3-(3,3,4,4,5,5,6,6,7,7,-
8,8,8-tridecafluoro-octyl)-phenyl]-phosphanyl}-[1,1%]-
binaphthalenyl-2-ol (9)
128.28 (s, arom CH), 128.21 (s, arom CH), 127.41 (d,
P(2)C=2.0 Hz, arom CH), 127.38 (s, arom CH),
J
127.27 (s, arom CH), 127.24 (s, arom CH), 127.16 (s,
arom CH), 126.69 (s, arom CH), 126.59 (s, arom CH),
126.44 (s, arom CH), 126.34 (s, arom CH), 125.29 (s,
arom CH), 125.01 (s, arom CH), 124.88 (s, arom CH),
123.35 (d, JP(1)C=1.7 Hz, arom C), 121.98 (s, arom
CH), 121.85 (s, arom CH), 120.98 (d, JP(1)C=8.6 Hz,
Trichlorosilane (1.92 ml, 0.014 mol) and triethy-
lamine (2.8 ml) was added to a solution of 8 (2.03 g,
0.02 mol) in toluene (40 ml) at 0°C. The reaction
mixture was heated at 100°C for 12 h. After cooling to
r.t., the reaction mixture was quenched with degassed 2
N NaOH (1 ml). The organic phase was filtered
through a path of Celite and MgSO4, and then concen-
trated under reduced pressure to give 9 as a white solid.
Yield: 61%.
arom CH), 120.55–106.75 (m, CFx), 32.51 (t, JFC
=
22.0 Hz, CH2CH2CF2), 32.31 (t, FC=21.9 Hz,
J
CH2CH2CF2), 26.27 (s, CH2CH2CF2), 26.12 (s,
CH2CH2CF2). 19F-{1H}-NMR (282.4 MHz, CDCl3):
l= −81.3 (m, 6F, CF3), −114.6 (m, 4F, CF2), −
122.1 (m, 4F, CF2), −123.1 (m, 4F, CF2), −123.5 (m,
2F, CF2), −123.7 (m, 2F, CF2), −126.4 (m, 4F, CF2).
LCMS (ESIpos): 1461 [M+H]. [h]2D0= +177° (ca. 1.0,
toluene).
31P-{1H}-NMR (C6D6): l= −12.7 (s). 1H-NMR
(C6D6): l=8.00–7.28 (m, 10H, Ar–H), 7.22–6.60 (m,
10H, Ar–H), 2.80–2.63 (m, 2H, CH2), 2.57–2.42 (m,
2H, CH2), 2.29–1.80 (m, 4H, CH2). LCMS (ESIpos):
1147 [M+H].
4.10. Synthesis of the complex [Rh(10)(acac)] (11)
4.9. Synthesis of 4-(2%-{bis-[3-(3,3,4,4,5,5,-
6,6,7,7,8,8,8-tridecafluoro-octyl)-phenyl]-phosphanyl}-
[1,1%]-(R)-binaphthalenyl-2-yloxy)-3,5-
A pre-cooled solution of the ligand 10 (146 mg, 0.1
mmol) in THF (10 ml) was added to a solution of
[Rh(acac)(CO)2] (25.8 mg, 0.1 mmol) in THF (10 ml) at
−78°C. After 30 min at −78°C, the reaction mixture
was slowly allowed to warm to r.t. and then stirred for
further 60 min. The reaction mixture was concentrated
to a volume of about 2 ml and pentane was added (20
ml) to form a bright-yellow precipitate that was col-
lected and washed with pentane (10 ml). Yield 98%.
31P-{1H}-NMR (C6D6): l=161.9 (dd, JPP%=84,
dioxa-4-phospha-cyclohepta[2,1-a;3,4-a%]-(S)-
dinaphthalene, 3-H2F6-(R,S)-BINAPHOS (10)
A solution of 9 (2.0 g, 1.74 mmol) and triethylamine
(1.2 ml, 8.6 mmol) in toluene (20 ml) was added
dropwise at 0°C to a solution of (S)-4-chloro-3,5-dioxa-
4-phospha-cyclohepta[2,1-a;3,4-a%]dinaphthalene (0.611
g, 1.74 mmol) in toluene (10 ml). The reaction mixture
was stirred overnight at r.t., filtered off, and concen-
trated under reduced pressure. The residue was ex-
tracted with pentane to obtain the pure title compound.
Yield: 99% (white solid).
J
RhP=330 Hz), 49.6 (dd, JRhP%=175 Hz). 1H-NMR
(C6D6): l=8.18–6.58 (m, 32H, Ar–H), 5.26 (s, 1H,
CH), 2.81–2.63 (m, 2H, CH2), 2.52–1.68 (m, 6H, CH2),
2.18 (s, 3H, CH3), 1.34 (s, 3H, CH3). 103Rh-NMR
(C6D6): l=271.
31P-{1H}-NMR (243.0 MHz, C6D6): l=146.5 (d,
JPP%=27.8 Hz), −13.3 (d). 1H-NMR (600.2 MHz,
C6D6): l=7.88–6.38 (m, 32H, Ar–H), 2.41–2.34 (m,
4H, CH2), 1.95–1.79 (m, 4H, CH2). 13C-NMR (150.9
MHz, C6D6): l=148.20 (d, JP(1)C=4.2 Hz, arom C),
148.06 (dd, JP(1)C=8.2, JP(2)C=1.8 Hz, arom C),
Acknowledgements
This work was supported by the Max-Planck-Society,
the Deutsche Forschungsgemeinschaft (Gerhard–Hess
Award to W. L.), the Fonds der Chemischen Industrie,
and the European Community (COST-D10). We thank
Celanese GmbH, Werk Ruhrchemie, for the AAS mea-
surements and Dr Wolfgang Baumann (University of
Rostock) for the 103Rh-NMR spectra. Special thanks
are due to Professor Kyoko Nozaki for a generous
donation of BINAPHOS and a very open scientific
discussion and to Professor C.J. Elsevier for his support
in establishing high pressure NMR in our group.
147.69 (d, JP(1)C=2.5 Hz, arom C), 141.71 (d, JP(2)C
=
34.3 Hz, arom C), 139.26 (d, JP(2)C=5.6 Hz, arom C),
138.96 (s, arom C), 138.94 (d, JP(2)C=25.1 Hz, arom
C), 137.82 (d, JP(2)C=14.6 Hz, arom C), 137.01 (d,
JP(2)C=12.1 Hz, arom C), 134.91 (d, JP(2)C=31.9 Hz,
arom CH), 134.85 (d, JP(2)C=2.4 Hz, arom C), 134.28
(s, arom C), 134.06 (d, JP(2)C=7.0 Hz, arom C), 133.36
(s, arom C), 133.11 (d, JP(2)C=19.2 Hz, arom CH),
132.91 (s, arom C), 132.69 (d, JP(2)C=11.5 Hz, arom
CH), 132.04 (s, arom C), 131.92 (d, JP(2)C=17.5 Hz,
arom CH), 131.62 (s, arom C), 131.19 (s, arom C),
130.66 (s, arom CH), 130.47 (s, arom CH), 130.41 (s,
arom CH), 129.64 (s, arom CH), 129.03 (d, JP(2)C=5.5
Hz, arom CH), 128.89 (d, JP(2)C=4.7 Hz, arom CH),
128.81 (s, arom CH), 128.76 (s, arom CH), 128.61 (s,
arom CH), 128.51 (s, arom CH), 130.41 (s, arom CH),
References
[1] (a) P.G. Jessop, W. Leitner (Eds.), Chemical Synthesis Using
Supercritical Fluids, Wiley-VCH, Weinheim, 1999. (b) R. Noy-
ori, Chem. Rev. 99 (1999) 353.

142
G. Francio` et al. / Journal of Organometallic Chemistry 621 (2001) 130–142
[2] (a) P.G. Jessop, T. Ikariya, R. Noyori, Science 269 (1995) 1065.
[24] W. Leitner, M. Bu¨hl, R. Fornika, C. Six, W. Baumann, E.
Dinjus, M. Kessler, C. Kru¨ger, A. Rufinska, Organometallics 18
(1999) 1196 and refs. cited therein.
[25] Alternatively, the electronic influence may be offset by very small
geometric changes, which influence the Rh-shift most signifi-
cantly [24].
[26] For NMR spectroscopic investigation of catalytically active
metal carbonyl complexess see: (a) J.W. Rathke, R.J. Klingler,
T.R. Krause, Organometallics 10 (1991) 1350. (b) R.J. Klingler,
J.W. Rathke, J. Am. Chem. Soc. 116 (1994) 4772. (c) Ref. [19e].
[27] J.W. Rathke, R.J. Klinger, R.E. Gerald, II, D.E. Fremgen, K.
Woelk, S. Gaemers, C.J. Elsevier, in: P.G. Jessop, W. Leitner
(Eds.), Chemical Synthesis Using Supercritical Fluids, Wiley-
VCH, Weinheim, 1999, pp. 165–194.
(b) W. Leitner, Top. Curr. Chem. 206 (1999) 107.
[3] (a) K. Zosel, Angew. Chem. Int. Ed. Engl. 17 (1978) 702. (b) M.
McHugh, V.J. Krukonis, Supercritical Fluid Extraction, second
ed., Butterworth–Heinemann, Boston, MA, 1994.
[4] A. Fu¨rstner, D. Koch, K. Langemann, W. Leitner, C. Six,
Angew. Chem. Int. Ed. Engl. 36 (1997) 2466.
[5] D. Koch, W. Leitner, J. Am. Chem. Soc. 120 (1998) 13398.
[6] S. Kainz, A. Brinkmann, W. Leitner, A. Pfaltz, J. Am. Chem.
Soc. 121 (1999) 6421.
[7] E.N. Jacobsen, A. Pfaltz, H. Yamamoto (Eds.), Comprehensive
Asymmetric Catalysis, Springer-Verlag, Berlin, 1999.
[8] M.J. Burk, S. Feng, M.F. Gross, W. Tumas, J. Am. Chem. Soc.
117 (1995) 8277.
[9] J. Xiao, S.C.A. Nefkens, P.G. Jessop, T. Ikariya, R. Noyori,
Tetrahedron Lett. 37 (1996) 2813.
[28] T. Burgemeister, F. Kastner, W. Leitner, Angew. Chem. Int. Ed.
Engl. 32 (1993) 739.
[10] S. Lange, P. Trautner, K. Woelk, J. Bargon, W. Leitner, Chiral-
ity 12 (2000) 450.
[11] D.R. Pesiri, D.K. Morita, W. Glaze, W. Tumas, Chem Commun
(1998) 1015.
[12] D.C. Wynne, M.M. Olmstead, P.G. Jessop, J. Am. Chem. Soc.
122 (2000) 7638.
[13] A. Wegner, W. Leitner, Chem. Commun. (1999) 1583.
[14] (a) F. Agbossou, J.-F. Carpentier, A. Mortreux, Chem. Rev. 95
(1995) 2485. (b) S. Gladiali, J.C. Bayo´n, C. Claver, Tetrahedron
Asymmetry 6 (1995) 1453. (c) C. Botteghi, S. Paganelli, A.
Schionato, M. Marchetti, Chirality 3 (1991) 335.
[15] (a) K. Nozaki, H. Takaya, T. Hiayama, Top. Catal. vol. 4, 1997,
p. 175. (b) K. Nozaki, N. Sakai, S. Mano, T. Higashijima, T.
Horiuchi, H. Takaya, J. Am. Chem. Soc. 119 (1997) 4413.
[16] (a) S. Kainz, W. Leitner, Catal. Lett. 55 (1998) 223. (b) I. Ojima,
M. Tzamarioudaki, C.Y. Chuang, D.M. Iula, Z. Li, Catalytic
carbonylations in supercritical carbon dioxide, in: F.E. Herkes
(Ed.), Catalysis of Organic Reactions, Marcel Dekker, New
York, 1998, p. 333.
[29] K. Nozaki, in: E.N. Jacobsen, A. Pfaltz, H. Yamamoto (Eds.),
Comprehensive Asymmetric Catalysis, vol. 1, Springer-Verlag,
Berlin, 1999, pp. 382–413.
[30] Y. Guo, A. Akgerman, J. Supercrit. Fluids 15 (1999) 63.
[31] (a) A.S.C. Chan, C.-C. Pai, T.-K. Yang, S.-M. Chen, Chem.
Commun. (1995) 2031. (b) C.P. Casey, E.L. Paulsen, E.W.
Beutenmu¨ller, B.R. Profi, B.A. Matter, D.R. Powell, J. Am.
Chem. Soc. 121 (1999) 63.
[32] K. Selvakumar, M. Valentini, P.S. Pregosin, A. Albinati, F.
Eisentra¨ger, Organometallics 19 (2000) 1299.
[33] For the use of BINAPHOS in hydrocyanation and copolymeri-
sation of propene and CO see, respectively: (a) T. Horiuchi, E.
Shirakawa, K. Nozaki, H. Takaya, Tetrahedron Asymmetry 8
(1997) 57. (b) K. Nozaki, N. Sato, H. Takaya, J. Am. Chem.
Soc. 117 (1995) 9911.
[34] G. Francio`, F. Faraone, W. Leitner, Angew. Chem. Int. Ed.
Engl. 39 (2000) 1428.
[35] M.F. Sellin, D.J. Cole-Hamilton, J. Chem. Soc. Dalton Trans.
(2000) 1681.
[17] S. Kainz, D. Koch, W. Baumann, W. Leitner, Angew. Chem.
Int. Ed. Engl. 36 (1997) 1628.
[36] P. Jessop, D.C. Wynne, S. DeHaai, D. Nakawatase, Chem.
Commun. (2000) 693.
[18] H. Hori, C. Six, W. Leitner, Macromolecules 32 (1999) 3178.
[19] (a) M.A. Carroll, A.B. Holmes, Chem. Commun. (1998) 1395.
(b) D.K. Morita, D.R. Pesiri, S.A. David, W.H. Glaze, W.
Tumas, Chem. Commun. (1998) 1397. (c) C.A.G. Carter, R.T.
Baker, S.P. Nolan, W. Tumas, Chem. Commun. (2000) 347. (d)
D.R. Palo, C. Erkey, Ind. Eng. Chem. Res. 37 (1998) 4203. (e)
D.R. Palo, C. Erkey, Organometallics 19 (2000) 81.
[20] G. Francio`, W. Leitner, Chem. Commun. (1999) 1663.
[21] S. Kainz, Z. Luo, D.P. Curran, W. Leitner, Synthesis (1998)
1425.
[22] (a) L. Kurz, G. Lee, D. Morgans, Jr., M.J. Waldyke, T. Ward,
Tetrahedron Lett. 44 (1990) 6321. (b) T. Hayashi, H. Iwamura,
Y. Uozumi, Y. Matsumoto, F. Ozawa, Synthesis (1994) 526. (c)
H. Doucet, J.M. Brown, Tetrahedron Asymmetry 8 (1997) 3775.
[23] K. Angermund, W. Baumarm, E. Dinjus, R. Fornika, H. Go¨rls,
M. Kessler, C. Kru¨ger, W. Leitner, F. Lutz, Chem. Eur. J. 3
(1997) 755.
[37] It is impossible to decide whether the enantioselective catalytic
reaction occurs exclusively in the liquid phase, the CO2 phase or
in both. However, the experience using unmodified BINAPHOS
in multiphase systems shows that the perfluoroalkyl substitution
is required for high selectivities even under two-phase conditions
[16a,20].
[38] (a) F.P. Lucien, N.R. Foster in: P.G. Jessop, W. Leitner (Eds.),
Chemical Synthesis Using Supercritical Fluids, Wiley–VCH,
Weinheim, 1999, pp. 37–53. (b) D.K. Dange, J.P. Heller, K.V.
Wilson, Ind. Eng. Chem. Prod. Res. Dev. 24 (1985) 162.
[39] W. Leitner, Angew. Chem. Int. Ed. Eng. 34 (1995) 2207.
[40] P.G. Jessop, Y. Hsiao, T. Ikariya, R. Noyori, J. Am. Chem. Soc.
118 (1996) 344.
[41] Y. Uozumi, A. Tanahashi, S.-Y. Lee, T. Hayashi, J. Org. Chem.
58 (1993) 1945.
[42] G. Francio`, C.G. Arena, F. Faraone, C. Graiff, M. Lanfranchi,
A. Tiripicchio, Eur. J. Inorg. Chem. (1999) 1219.
.