Enantioselective hydrogenation of aromatic ketones over cinchona-modified rhodium: A new opportunity?

Article abstract of DOI:10.1016/j.jcat.2005.01.008

Alumina-supported rhodium modified with cinchonidine has been investigated with regard to its applicability in the enantioselective hydrogenation of various aromatic ketones possessing an α-hydroxy or α-methoxy group. The study revealed that depending on the substrate, rhodium can outperform the catalytic behavior of platinum. With one of the substrates, 2-hydroxy-1-(4- methoxy-phenyl)-ethanone (4), an enantiomeric excess (ee) of 80% at 89% conversion was reached, which is the highest ee reported so far for chirally modified rhodium. However, completely different conditions are required to achieve optimal catalytic performance with rhodium, compared with platinum. Rhodium requires a much higher modifier concentration, and high hydrogen pressure is favorable. The higher modifier concentration required is traced to the much higher activity of rhodium for the hydrogenation of the quinoline ring, which is assumed to be the anchoring moiety of the cinchona modifiers on the platinum group metals. Changing the modifier from cinchonidine to O-phenoxy-cinchonidine resulted in a switch of the major enantiomer of the product, as exemplified for 2-hydroxyacetophenone (1), which showed a switch from 73% ee in favor of the (R)-product to 68% ee for the (S)-product when the modifier was changed from cinchonidine to O-phenoxy-cinchonidine.

Full text of DOI:10.1016/j.jcat.2005.01.008

Journal of Catalysis 230 (2005) 499–506
www.elsevier.com/locate/jcat
Enantioselective hydrogenation of aromatic ketones over
cinchona-modified rhodium: a new opportunity?
1
∗
Otmar J. Sonderegger, Gabriel M.-W. Ho, Thomas Bürgi , Alfons Baiker
Department of Chemistry and Applied Biosciences, Swiss Federal Institute of Technology, ETH-Hönggerberg, CH-8093 Zürich, Switzerland
Received 12 November 2004; revised 23 December 2004; accepted 10 January 2005
Abstract
Alumina-supported rhodium modified with cinchonidine has been investigated with regard to its applicability in the enantioselective
hydrogenation of various aromatic ketones possessing an α-hydroxy or α-methoxy group. The study revealed that depending on the substrate,
rhodium can outperform the catalytic behavior of platinum. With one of the substrates, 2-hydroxy-1-(4-methoxy-phenyl)-ethanone (4), an
enantiomeric excess (ee) of 80% at 89% conversion was reached, which is the highest ee reported so far for chirally modified rhodium.
However, completely different conditions are required to achieve optimal catalytic performance with rhodium, compared with platinum.
Rhodium requires a much higher modifier concentration, and high hydrogen pressure is favorable. The higher modifier concentration required
is traced to the much higher activity of rhodium for the hydrogenation of the quinoline ring, which is assumed to be the anchoring moiety
of the cinchona modifiers on the platinum group metals. Changing the modifier from cinchonidine to O-phenoxy-cinchonidine resulted in a
switch of the major enantiomer of the product, as exemplified for 2-hydroxyacetophenone (1), which showed a switch from 73% ee in favor
of the (R)-product to 68% ee for the (S)-product when the modifier was changed from cinchonidine to O-phenoxy-cinchonidine.

2005 Elsevier Inc. All rights reserved.
1
. Introduction
known that are sufficiently efficient for potential practical
application; among them some heterogenized homogeneous
catalysts [5] and chirally modified metals [6,7] seem most
promising. Chirally modified nickel [8,9], platinum [10–12],
and palladium [13,14] have emerged as effective heteroge-
neous asymmetrical catalysts for the enantioselective hy-
drogenation of specific substrates. Nevertheless, the scope
of reactants that can be hydrogenated with high optical
yield is still rather narrow, and it remains a major chal-
lenge to extend the application range of these catalysts. For
this purpose it seems promising to examine the behavior of
other platinum-group metals and that of new chiral modi-
fiers. With this in mind we have explored the behavior of
alumina-supported rhodium modified by cinchonidine and
O-phenoxy-cinchonidine in the enantioselective hydrogena-
tion of a series of aromatic ketones. Although chirally mod-
ified rhodium has been applied in earlier studies [15–20],
the enantioselectivities reached were low or at best mod-
erate. Here, for the first time, an example is shown where
an enantiomeric excess of 80% at 89% conversion could
Among the various methods for preparing enantiopure
compounds, catalysis is probably the most elegant and
promising [1,2]. Though asymmetrical homogeneous cataly-
sis has become an important tool in asymmetrical synthesis,
relatively few catalytic enantioselective processes are cur-
rently used on a commercial scale [3]. Major problems in
the scale-up are linked with cost effectiveness, process ro-
bustness, and separation and stability of the catalyst. In
principle, heterogeneous catalysis could offer some inter-
esting advantages, such as easier handling and separation
of the catalyst and continuous process operation [4], which
could be favorable for large-scale production. However, to-
day only a few heterogeneous enantioselective catalysts are
*
Corresponding author. Fax: +41 1 632 11 63.
E-mail address: baiker@chem.ethz.ch (A. Baiker).
1
Present address: Institut de Chimie, Université de Neuchâtel, Av. de
Bellevaux 51, CH-2007 Neuchâtel, Switzerland.
0021-9517/$ – see front matter  2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2005.01.008

500
O.J. Sonderegger et al. / Journal of Catalysis 230 (2005) 499–506
be achieved. The main features of the chirally modified
rhodium are compared with that of the corresponding plat-
inum system.
total conversion was chosen because the amounts of by-
products were generally low. The yields of the alcohols listed
in Table 3 were calculated as yield = chemoselectivity to
alcohol × conversion. The enantiomeric excess [ee (%) =
1
00 × (|R − S|)/(R + S)] was determined with the use
2
. Experimental
of a Merck LaChrom HPLC System and a chiral column
(
Chiralcel OB or Chiralcel OD). The chromatographic sep-
◦
aration was carried out at 15 C and detection at 210 nm.
We calibrated the detector by injecting mixtures of the reac-
tant and the product of known compositions. The absolute
configuration of the major enantiomer of phenylethane-1,2-
diol (product of 1) in the reaction with CD as a modifier
was determined to be a (R)-enantiomer by comparison with
the corresponding commercially available (R)-enantiomer
All reactants (Scheme 1) and solvents were used as re-
ceived: 2-hydroxyacetophenone (1) (Aldrich, 98%), 1-hyd-
roxy-2-methylpropiophenone (2) (Aldrich, 97%), 2-me-
thoxy-acetophenone (3) (Aldrich, 95%), 2-hydroxy-1-(4-
methoxy-phenyl)-ethanone (4) (Bionet Research Ltd.), cin-
chonidine (CD) (Fluka), O-phenoxy-cinchonidine (PhOCD)
(
(
Ubichem, > 95%), tert-butylmethylether (t-BM-ether)
Fluka, > 99.5%), dioxane (Fluka, > 99.5%; over molec-
(
Fluka). The amounts of by-products were determined with
a Thermo Finnigan Trace gas chromatograph (GC) with an
ular sieve (H2O < 0.01%)), ethyl acetate (Merck, > 99.5%),
THF (Riedel-de Haën, Spectranal, > 99.9%), and toluene
(
Optima-5 (30 m×0.25 mm; Macherey–Nagel) column. The
1
H NMR spectra were recorded on a Bruker Avance spec-
Fluka, > 99.7%).
trometer operating at 500 MHz. Hydrogenation of CD was
followed by UV–vis spectroscopy with a Cary 400 spectrom-
eter and a 1-cm-path-length quartz cuvette. The conditions
were as follows: 4 mg CD in 20 ml dioxane; 84 mg cata-
lyst (Rh/Al2O3, Pt/Al2O3); hydrogen pressure, 5 and 90 bar,
respectively. The samples were diluted with dioxane by a
factor of 5 for UV–vis measurements.
Density functional theory (DFT) closed-shell calculations
were performed with the use of Gaussian03 [21]. Complete
optimization of all internal coordinates of the molecule was
performed at the B3PW91 [22] level of theory with the use
of the 6-31G(d,p) basis set [23]. The vibrational frequencies
were calculated for several minimized structures. We simu-
lated absorption and vibrational circular dichroism (VCD)
spectra from calculated normal modes, dipole strengths,
and rotational strengths, assuming Lorentzian band shape
with a linewidth of four wavenumbers (half width at half-
maximum).
The catalysts applied were 5 wt% Rh/Al2O3 (Engelhard
001 ESCAT 34) and 5wt% Pt/Al2O3 (Engelhard 4759). Be-
8
fore use the catalysts were pre-reduced in flowing hydrogen
◦
for 90 min at 400 C. The platinum dispersion was 0.27, as
determined by TEM measurements. Textural and structural
properties of the Rh/Al2O3 catalyst and the effect of the pre-
treatment have been reported in detail elsewhere [20].
The hydrogenation reactions up to 30 bar were carried out
in a multiple reactor (Argonaut Technologies) equipped with
eight 10-ml glass liners, which were mechanically stirred
−1
(
n = 500 min ). Reactions above 30 bar were performed
in a 100-ml stainless-steel autoclave equipped with a 50-ml
glass liner and PTFE cover. The reactor was magnetically
−1
stirred (n = 500 min ). The pressure was held at a constant
value with computerized constant volume–constant pressure
equipment (Büchi BPC 9901). The standard reaction con-
ditions were 21 ± 1 mg pre-reduced catalyst, 0.92 mmol
reactant (0.46 mmol for reactant 1), 6.8 µmol modifier, 5 ml
solvent, a pressure of 20 bar, room temperature, and a reac-
tion time of 6 h.
VCD was used to determine the absolute configuration of
the major enantiomer of the hydrogenation product of 4. IR
and VCD spectra were recorded on a Bruker PMA 50 acces-
sory coupled to a Tensor 27 Fourier transform infrared spec-
trometer. A photoelastic modulator (Hinds PEM 90) set at
The conversions given in the tables and figures repre-
sent the total consumption of the ketons corresponding to
all products (alcohol + by-products). Presentation of the
1
/4 retardation was used to modulate the handedness of the
circular polarized light. Demodulation was performed with
a lock-in amplifier (SR830 DSP). An optical low-pass filter
−1
(
< 1800 cm ) put before the photoelastic modulator was
used to enhance the signal/noise ratio. A transmission cell
equipped with KBr windows and a 0.2-mm Teflon spacer
was used. All solutions of product of 4 were prepared at a
concentration of 10 mg in 500 µl CD2Cl2. A VCD spectrum
of racemic product of 4 was subtracted from an enantioen-
riched sample of product of 4 (ee 80%) as obtained from
the enantioselective hydrogenation over CD-modified Rh.
We measured both the racemic and the enantioenriched sam-
−1
ples at 4 cm resolution by accumulating interferograms
over 5 h in time slices of 1 h (total scans 21,500 for each
racemic and enantioenriched sample). The spectra are pre-
Scheme 1. Overview of reactants.

O.J. Sonderegger et al. / Journal of Catalysis 230 (2005) 499–506
501
sented without smoothing or further data processing. More
information about the experimental procedure can be found
elsewhere [24,25].
Table 1
Influence of solvents on enantioselectivity and conversion for the hydro-
genation of 1–4 over CD–Rh and 4 over CD–Pt
N a
T
Reactant
Solvent
E
[26]
ee (%)
Conversion (%)
1
Toluene
t-BM-Ether
Dioxane
0.10
0.15
0.16
0.21
0.23
0.31
0.55
0.65
16
39
51
18
37
3
40
40
36
9
62
48
99
100
3
. Results
THF
To gain some information about the most suitable sol-
Ethyl acetate
Dichloromethane
2-Propanol
Acetic acid
vent, at first hydrogenation of reactants 1–4 (Scheme 1) was
carried out in various solvents on the cinchonidine-modified
Rh/Al2O3 (CD-Rh) catalyst. For comparison, with reactant
23
3
4
the corresponding CD-modified Pt/Al2O3 (CD–Pt) cata-
2
Toluene
t-BM-Ether
Dioxane
0.10
0.15
0.16
0.21
0.23
0.31
0.55
0.65
32
28
44
–
31
7
100
100
21
0
100
58
lyst was applied as well (Table 1). Dioxane proved to be the
most suitable solvent concerning ee. tert-Butylmethylether
THF
(
t-BM-ether) and ethyl acetate showed somewhat lower ee,
Ethyl acetate
Dichloromethane
2-Propanol
Acetic acid
but higher conversion. A striking exception to the solvent be-
havior was observed with THF, which showed very low con-
version for all hydrogenations performed on rhodium. The
28
1
100
87
N
T
results are listed according to increasing E values (empiri-
3
4
Toluene
0.10
0.15
0.16
0.21
0.23
0.55
0.65
21
29
37
24
25
16
2
13
33
25
8
24
67
100
t-BM-Ether
Dioxane
THF
cal parameter of solvent polarity) [26], a parameter that has
been used previously for correlation with catalytic perfor-
mance [27]. However, no significant tendency is discernible,
indicating that in the present enantioselective hydrogena-
tions the catalytic performance cannot be correlated with this
parameter.
Based on the solvent screening results listed in Table 1,
all further investigations were performed in dioxane, if not
specified explicitly. The hydrogenations of substrates 1–4
all showed good chemoselectivity for the corresponding al-
cohols. By-products originated from hydrogenolysis of the
C–O bond and/or saturation of the aromatic ring, as previ-
ously observed by Hess et al. [20]. At standard conditions
the amount of by-products was minor (1.1–6.8%). How-
ever, at low CD concentrations a pronounced increase in by-
products was observed. For instance, with reactant 1, 13.4
and 36.0% by-products were detected at 1/CD ratios of 1350
and 13,500, respectively. Thus the presence of modifier sup-
pressed by-product formation strongly.
Table 2 shows the performance of CD–Rh under stan-
dard conditions. For reactants 1 and 3, CD–Rh showed lower
performance than the CD–Pt catalyst. Interestingly, with re-
actant 2, ee and conversion were the same on both catalysts.
For reactant 4 the ee was the same, whereas the conversion
was higher on the CD–Rh system.
Next we investigated the influence of the modifier/sub-
strate ratio on the catalytic performance. In these ex-
periments the substrate concentration was kept constant,
whereas the modifier (cinchonidine; for reactant 1 also O-
phenoxy-cinchonidine) was varied. Fig. 1 shows the depen-
dence of ee on the modifier/substrate ratio. With the CD–Rh
catalyst the ee decreased steadily with a higher substrate/CD
ratio. The strength of the decrease depended on the sub-
strate. For the hydrogenation of 1 over the same catalyst,
but modified by PhOCD (PhOCD–Rh), a similar tendency
was observed. In contrast, hydrogenation of 4 over CD–Pt
showed no decrease in ee up to a 4/CD ratio of about 540,
Ethyl acetate
2-Propanol
Acetic acid
Dioxane
THF
Ethyl acetate
Acetic acid
0.16
0.21
0.23
0.65
75
–
64
1
31
0
61
47
4
(Pt)
Dioxane
THF
Ethyl acetate
Acetic acid
0.16
0.21
0.23
0.65
75
59
64
3
10
10
31
24
Standard conditions: 21 mg catalyst, 0.92 mmol reactant (0.46 mmol for
reactant 1), 6.8 µmol CD, 5 ml solvent, 20 bar, room temperature (21 C),
and 6 h reaction time.
◦
a
Empirical parameter of solvent polarity.
Table 2
Performance of CD–Rh in the enantioselective hydrogenation of 1–4 under
standard reaction conditions with dioxane as solvent. For comparison the
performance of the corresponding CD–Pt catalyst is also given
Reactant
CD–Rh
CD–Pt
Conversion (%)
ee (%)
Conversion (%)
ee (%)
1
2
3
4
35.5
21.2
20.0
30.5
50.5
44.0
36.9
75.0
41.3
21.3
20.4
10.2
70.2
44.3
68.8
75.2
Standard conditions are specified in Table 1. The data characterizing the
performance of CD–Pt in the hydrogenation of reactants 1–3 are taken
from [35].
but also decreased strongly at higher substrate/modifier ra-
tios.
The effect of the hydrogen pressure on ee and conver-
sion in the hydrogenation of 1 over cinchonidine- and O-
phenoxy-cinchonidine-modified rhodium is shown in Fig. 2.
Enantiomeric excess and conversion increased up to 60 bar

502
O.J. Sonderegger et al. / Journal of Catalysis 230 (2005) 499–506
Fig. 1. Enantiomeric excess of 1, 2, 3, and 4 over CD or PhOCD modi-
fied Rh/Al O and 4 over CD modified Pt/Al O as a function of reac-
2
3
2 3
tant/modifer ratio. Reaction conditions: 21 mg catalyst, 0.92 mmol reactant
0.46 mmol for reactant 1), 5 ml solvent, 20 bar, room temperature, and 6 h
reaction time.
Fig. 3. Dependence of enantiomeric excess and conversion of 4 over CD
modified Pt/Al2O3 or Rh/Al2O3 as a function of hydrogen pressure. Reac-
tion conditions: 3.4 µmol CD (Pt/Al2O3), 54.4 µmol CD (Rh/Al2O3), other
conditions as in Fig. 1.
(
Table 3
Comparison of catalytic performance of rhodium and platinum modified
with CD or PhOCD in the hydrogenation of 1–4 under optimised reaction
conditions
Reactant
Modifier
µmol)
Pressure
(bar)
Conversion
(%)
Yield
alcohol
ee
(R, %)
(
(%)
CD–Rh
1
2
3
4
54.4
54.4
54.4
54.4
60
30
90
90
82.0
32.1
68.2
88.4
78.8
31.5
64.5
86.9
73.1
51.2
63.6
80.4
CD–Pt
1
2
3
3.4
6.8
6.8
3.4
5
5
5
5
52.6
61.6
69.3
4.0
49.1
60.5
64.3
3.9
82.3
56.6
81.1
78.9
Fig. 2. Dependence of enantiomeric excess and conversion of 1 over CD
a
or PhOCD modified Rh/Al O on hydrogen pressure. Reaction conditions:
4
2
3
54.4 µmol modifier, other conditions as in Fig. 1.
PhOCD–Rh
1
54.4
60
99.2
92.3
68.4 (S)
with both modifiers, CD and PhOCD. However, a striking
difference is that PhOCD–Rh favored the opposite enan-
tiomer ((S)- instead of (R)-product) over CD–Rh. The ee
in the hydrogenation of reactants 2 and 3 indicated only
little dependence on pressure, whereas conversion steadily
increased, reaching a plateau at around 30 bar. The pressure
dependence of the hydrogenation of 4 has been investigated
on both CD–Rh and CD–Pt (Fig. 3). The difference in be-
havior is striking. With CD–Rh the ee increased up to 10 bar
and remained virtually constant at this level up to 90 bar. The
corresponding conversion increased monotonically over the
whole pressure range of 2–90 bar. This contrasts with the be-
havior of CD–Pt, for which the ee decreased with increasing
pressure, whereas the conversion reached a plateau at pres-
sures greater than 5 bar.
PhOCD–Pt
b
1
6.8
5
57.4
54.6
34.5 (S)
◦
Rhodium-catalysed hydrogenations were performed in dioxane at 15 C
and 6 h reaction time. Platinum-catalysed hydrogenations were carried out
◦
in t-BM-ether at 0 C and 6 h reaction time.
a
◦
Reaction performed in dioxane at 15 C.
b
◦
Reaction performed in t-BM-ether at 21 C.
eral higher ee’s were achieved with CD–Pt. Interestingly, for
reactant 4 the same ee (80%) was achieved on both CD–Rh
and CD–Pt. However, the corresponding difference in con-
version is striking: 89% over the CD–Rh compared with 4%
over CD–Pt. Note the remarkable differences in the condi-
tions (modifier concentration, hydrogen pressure) required
to reach best performance for the rhodium and platinum sys-
tem.
Table 3 compares the catalytic performance of the CD–
Rh and CD–Pt systems under conditions that led to the best
results. Note that an extensive optimization was not carried
out, leaving the possibility for further improvement. In gen-
Hydrogenations performed with 1 and PhOCD as a chi-
ral modifier afforded the (S)-product as the major enan-
tiomer, independently of the metal catalyst used. This

O.J. Sonderegger et al. / Journal of Catalysis 230 (2005) 499–506
503
Fig. 5. Time dependence of decrease of CD concentration in solution during
hydrogenation over Rh and Pt. Absorption at 280 nm was taken as a measure
of CD concentration. Reaction conditions: 3.4 µmol CD, 21 mg catalyst,
Fig. 4. Hydrogenation of the quinoline ring of CD over Rh/Al O at 90 bar
followed by UV–vis. Reaction conditions: 3.4 µmol CD, other conditions as
in Fig. 1.
2
3
5
ml dioxane, 5 or 90 bar H .
2
change in enantioselectivity when cinchonidine is replaced
by O-phenyl-cinchonidine had been observed earlier in the
platinum-catalyzed hydrogenation of ketopantolactone [28].
Particularly striking is the behavior of the rhodium-catalyzed
reaction, where almost similar ee’s (73, 68%) were achieved
for the (R)- and (S)-enantiomer of the product with CD and
PhOCD, respectively.
existed in solution compared with the corresponding exper-
iment with Pt/Al2O3 as a catalyst. This behavior was less
drastic for hydrogenations performed at 5 bar, where 2.8
times less CD existed in solution for the rhodium-catalyzed
hydrogenation than with the platinum-catalyzed reaction af-
ter 360 min. NMR analysis from a reaction solution where
54.4 µmol CD was hydrogenated at 90 bar for 360 min over
rhodium (optimum condition for reactant 4) showed about
equal amounts of CD and fully hydrogenated CD (quino-
line ring fully hydrogenated). Partially hydrogenated CD
1
With reactant 4, H NMR spectra were recorded for d8-
dioxane to see if the keto or the enol form prevails in solu-
tion. The experiments were carried out with the following
mixtures, mimicking reaction conditions: (i) 7.6 mg 4 in
(
1,2,3,4-tetrahydroquinoline) was found as a minor species.
0
.5 ml dioxane; (ii) 7.6 mg 4, 0.1 mg CD in 0.5 ml dio-
xane; and (iii) 7.6 mg 4, 2 mg CD in 0.5 ml dioxane. In all
reaction mixtures (i)–(iii) only the keto form could be ob-
served.
3.2. Determination of the absolute configuration of alcohol
produced from
2
-hydroxy-1-(4-methoxy-phenyl)-ethanone (4)
3
.1. Hydrogenation of cinchonidine over rhodium and
platinum
Vibrational circular dichroism (VCD) has recently emerg-
ed as a reliable technique for assigning the absolute con-
figuration of chiral samples in solution [24,25,32–34]. This
strategy relies on the comparison between theoretical and
calculated spectra and has become possible since the calcu-
lation of VCD spectra of organic molecules has predictive
character. As a first step the most stable conformers have to
be determined. DFT calculations showed that the alcohol of
4 is relatively flexible, adopting multiple minima. The rel-
evant degrees of freedom are the position of the methoxy
group, torsions around two C–C bonds, and the position of
the hydroxyl hydrogens. It turned out that the two possi-
ble positions of the methoxy group have only a minor effect
on stability. The CH2OH group prefers the position out of
the aromatic plane of the molecule, and the O–H groups
form a hydrogen bond, where each group can act as donor
or acceptor. We have located eight stable conformers. The
most stable four of them are shown in Fig. 6. The other
four conformers are at least 1.4 kcal/mol less stable than the
most stable conformer and are thus expected to be barely
detectable at room temperature.
The hydrogenation of the aromatic quinoline ring of CD,
which is known to be the anchoring group on Pt [29–31],
was investigated over both Rh/Al2O3 and Pt/Al2O3 under
conditions similar to those used in the enantioselective hy-
drogenation of substrates 1–4 (3.4 µmol CD, 5 ml dioxane, 5
and 90 bar H2, respectively). From observation of the inten-
sity of the absorption at 280 nm, indicative of the aromatic
quinoline ring, the hydrogenation of CD could be followed.
Fig. 4 shows UV–vis spectra for CD (0 min) and spectra for
CD and its hydrogenation products after different time inter-
vals. Note that 3.4 µmol CD is hydrogenated quite quickly
over Rh at 90 bar. Fig. 5 shows a decrease in the intensity
of the absorption at 280 nm versus time, which is a mea-
sure for the hydrogenation rate of the quinoline ring of CD.
It emerges from Fig. 5 that the hydrogenation rate was much
faster on rhodium than on platinum. The hydrogenation rate
increased in the following order: 5 bar Pt < 90 bar Pt <
5
bar Rh < 90 bar Rh. After hydrogenation of CD at 90
bar (H2) for 360 min, over Rh/Al2O3 about 5 times less CD

504
O.J. Sonderegger et al. / Journal of Catalysis 230 (2005) 499–506
Fig. 6. Structure of the four most stable conformers of the alcohol produced by hydrogenation of 4, calculated by DFT B3PW91 using a 6-31G(d,p) basis set.
The energies are relative to the most stable conformer 4.
Fig. 7. Experimental and calculated infrared absorption and vibrational circular dichroism (VCD) spectra. The calculated spectra represent the Boltzmann
average of the calculated spectra of the four most stable conformers (Fig. 6). The experimental spectra were recorded from solutions of 10 mg/500 µl CD Cl
2
2
with a path length of 0.2 mm. Asterisks in the experimental VCD spectra represent two regions where the infrared absorption was too high to reliably measure
the VCD signal.
Fig. 7 shows a comparison of the experimental and cal-
culated infrared absorption and VCD spectra. The calcu-
lated spectra were obtained from the calculated spectra of
the individual conformers (considering only the four most
stable ones); these were weighed by the appropriate Boltz-
mann factor at 298 K. The agreement between experimental
and calculated infrared spectra is reasonable, considering
the relatively large flexibility of the molecule. Some bands
can be unequivocally assigned, such as the relatively weak
band slightly below 1600 cm associated with a ring de-
formation mode. This latter mode is actually not sensitive to
conformation. This is also true for the calculated VCD spec-
tra for the conformers shown in Fig. 6, for which the latter
band is always positive. In the experiment this band is also
positive. Since the calculations were performed for the (R)-
enantiomer, the results presented above show that the hydro-
genation over CD-modified Pt provides the (R)-enantiomer
−1
in excess. Furthermore, the region below 1250 cm with a
number of negative bands leaves no doubt about the assign-
ment of the absolute configuration. Note that VCD spectra
of enantiomers have an opposite sign but are otherwise iden-
tical. It should be noted that the VCD signals are rather
small, in agreement with the relatively large flexibility of the
molecule, but the spectral features can be reproduced very
well.
−1
4. Discussion
The present study of the enantioselective hydrogenation
of different α-hydroxy or α-methoxy ketones indicates that
cinchona-modified rhodium can in certain cases show a su-

O.J. Sonderegger et al. / Journal of Catalysis 230 (2005) 499–506
505
perior catalytic behavior compared with platinum. This is
exemplified by the hydrogenation of reactant 4, in which an
ee of 80% was achieved at 89% total conversion (Table 3).
Over platinum at slightly lower ee the conversion was only
steadily to the highest pressure (90 bar), whereas with plat-
inum, above ca. 10 bar conversion was hardly affected by the
hydrogen pressure. In addition, the pressure dependence of
ee shows significantly different behavior; with rhodium the
ee increases and reaches a constant level at a pressure lower
than 10 bar, whereas with platinum the ee decreases steadily
with increasing pressure.
4
%. It is very likely that 4 is hydrogenated in the keto form,
even at high CD concentration, because the corresponding
1
enol could not be observed in the H NMR experiments.
The behavior of rhodium when it is modified with PhOCD in
the hydrogenation of 1 is striking (Table 3). Platinum is far
less suitable for this reaction than rhodium, with respect to
both ee and conversion. With reactant 2, ee and conversion
are higher on the Pt system under optimized conditions (Ta-
ble 3), whereas under standard conditions the performances
of the two metals were similar. Under both optimized and
standard conditions in the hydrogenation of 3, the ee was
higher on platinum, whereas the conversion was the same
with the two systems.
The functional group in the α-position to the carbonyl
plays a crucial role in the enantioselection. This has already
been stated by Vargas et al. based on theoretical calcula-
tions [36], and it has been shown for reactant 1 experimen-
tally [35]. A comparison of the ee of reactant 4 (79%, CD–
ꢀ
Pt) with the ee of 4 -methoxyacetophenone (11%) reported
in an earlier study [37] corroborates this finding.
Finally, the present work indicates that rhodium chirally
modified by cinchona alkaloids can provide considerably
better catalytic performance in enantioselective hydrogena-
tion then previously reported. Further work toward a deeper
understanding of this system may lead to new opportunities
in heterogeneous asymmetrical hydrogenation.
The fact that a comparison of the catalytic behaviors
of the two metals provides a changing picture, depend-
ing on whether standard (Table 2) or optimized conditions
(Table 3) are used, indicates that the most suitable condi-
tions are greatly different for rhodium than for platinum.
Generally, the modifier concentration that is necessary for
rhodium (Table 3) is much higher than that needed for plat-
inum. This also emerges when the dependence of the ee
on the reactant/modifer ratio is compared, as shown for re-
actant 4 in Fig. 1. For rhodium, the ee decreased steadily
with increasing reactant/modifier ratio. This tendency was
observed for all reactants 1–4, although the loss in ee was
different, depending on the reactant. In contrast with plat-
inum, a broad maximum for the reactant/modifier ratio was
observed and the decrease in ee set in at a much higher
reactant/modifier ratio (Fig. 1). This behavior of the plat-
inum system had already been observed for reactants 1–4 in
a previous study [35].
Remarkable differences between the rhodium- and plati-
num-catalyzed reactions are also evident from the investi-
gated hydrogen pressure dependences. Fig. 2, which com-
pares the hydrogen pressure dependence of ee and conver-
sion on rhodium modified with CD and PhOCD for the hy-
drogenation of 1, shows that both ee and conversion increase
steadily with higher pressure, independently of the modi-
fier used. For reactant 4 a somewhat different behavior was
observed (Fig. 3); the ee reached a constant value at about
5
. Conclusions
Rhodium has been shown to be an interesting candidate
for the extension of the range of cinchona-modified metal
catalysts for enantioselective hydrogenation. This emerges
from a comparison of the hydrogenations of reactants 1–4
on rhodium and platinum. In some cases, rhodium showed
superior catalytic behavior compared with platinum, indicat-
ing that extending the range of metals suitable for hetero-
geneous enantioselective hydrogenation may be rewarding.
A problem is that considerably different conditions have to
be applied for these metals to provide optimal performance.
It seems that with rhodium, the modifier concentration and
hydrogen pressure needed to reach optimal catalytic behav-
ior are higher than those required to reach optimal behavior
with platinum. Specifically, for the α-hydroxy ketones stud-
ied in this work, low CD concentration and low hydrogen
pressure are most suitable for the platinum-catalyzed hydro-
genations, whereas with rhodium high CD concentration and
high pressure are optimal. Moreover, the choice of the op-
timal solvent depends on the metal applied. A disturbing
property of rhodium is its much higher activity in the hy-
drogenation of the aromatic quinoline ring in cinchona-type
modifiers. Saturation of this anchoring group and the accom-
panying loss of adsorption strength may explain the much
higher concentration of modifier required for the rhodium-
catalyzed hydrogenations. The findings of this study corrob-
orate the importance of adjusting and tuning of all parame-
ters of the chirally modified metal system for achieving best
catalytic performance. It is hoped that a better understanding
of the important interactions occurring in these systems may
pave the way to more efficient and versatile chirally modi-
fied metal catalysts.
1
0 bar, whereas conversion increased steadily with hydro-
gen pressure. Thus the pressure dependence of ee seems to
vary depending on the reactant, whereas the increase in con-
version with higher pressure appears to be a more common
feature. The pressure dependence of the rhodium-catalyzed
hydrogenations shows significant differences compared with
that of corresponding platinum-catalyzed reactions. This
emerges from a comparison of the present results with cor-
responding earlier investigations of the platinum-catalyzed
reactions [35]. Fig. 3 illustrates this behavior with the ex-
ample of the hydrogenation of reactant 4. The most strik-
ing difference is that with rhodium, conversion increased

506
O.J. Sonderegger et al. / Journal of Catalysis 230 (2005) 499–506
Acknowledgments
[20] R. Hess, F. Krumeich, T. Mallat, A. Baiker, J. Mol. Catal. A Chem. 212
2004) 205.
21] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb,
(
[
Financial support by the Swiss National Science Foun-
dation is gratefully acknowledged. The Swiss National Su-
percomputer Centre (CSCS) in Manno is acknowledged for
computing time.
J.R. Cheeseman, J.A. Montgomery, T. Vreven, K.N. Kudin, J.C. Bu-
rant, J.M. Millam, S.S. Iyengar, J. Tomasi, V. Barone, B. Mennucci,
M. Cossi, G. Scalmani, N. Rega, G.A. Petersson, H. Nakatsuji, M.
Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T.
Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J.E. Knox,
H.P. Hratchian, J.B. Cross, C. Adamo, J. Jaramillo, R. Gomperts,
R.E. Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C. Pomelli, J.W.
Ochterski, P.Y. Ayala, K. Morokuma, G.A. Voth, P. Salvador, J.J. Dan-
nenberg, V.G. Zakrzewski, S. Dapprich, A.D. Daniels, M.C. Strain,
O. Farkas, D.K. Malick, A.D. Rabuck, K. Raghavachari, J.B. Fores-
man, J.V. Ortiz, Q. Cui, A.G. Baboul, S. Clifford, J. Cioslowski, B.B.
Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R.L. Mar-
tin, D.J. Fox, T. Keith, M.A. Al-Laham, C.Y. Peng, A. Nanayakkara,
M. Challacombe, P.M.W. Gill, B. Johnson, W. Chen, M.W. Wong, C.
Gonzalez, J.A. Pople, Gaussian03, Rev. B05 ed., Gaussian, 2003.
22] A.D. Becke, J. Chem. Phys. 98 (1993) 5648.
References
[1] R.A. Sheldon, Chirotechnology, Dekker, 1993.
[2] H.-U. Blaser, F. Spinler, M. Studer, Appl. Catal. A: Gen. 221 (2001)
119.
[
[
[
[
3] A.M. Rouhi, Chem. & Eng. News 82 (24) (2004) 47.
4] N. Künzle, R. Hess, T. Mallat, A. Baiker, J. Catal. 186 (1999) 239.
5] P. McMorn, G.J. Hutchings, Chem. Soc. Rev. 33 (2) (2004) 108.
6] A. Baiker, H.U. Blaser, in: G. Ertl, H. Knözinger, J. Weitkamp (Eds.),
in: Handbook of Heterogeneous Catalysis, vol. 5, VCH, Weinheim,
[
[
[
23] R. Ditchfield, W.J. Hehre, J.A. Pople, J. Chem. Phys. 54 (1971) 724.
24] T. Bürgi, A. Vargas, A. Baiker, J. Chem. Soc. Perkin Trans. 2 (2002)
1596.
1997, p. 2422.
[7] A. Baiker, Curr. Opin. Solid State Mater. Sci. 3 (1) (1998) 86.
[8] T. Osawa, T. Harada, O. Takayasu, Top. Catal. 13 (2000) 155.
[9] A. Tai, T. Sugimura, in: D.E. De Vos, I.F.J. Vankelecom, P.A. Jacobs
[25] T. Bürgi, A. Urakawa, B. Behzadi, K.-H. Ernst, A. Baiker, New J.
Chem. 28 (2004) 332.
(Eds.), Chiral Catalyst Immobilization and Recycling, Wiley–VCH,
Weinheim, 2000, p. 173.
10] M. Studer, H.U. Blaser, C. Exner, Adv. Synth. Catal. 345 (2003) 45.
11] A. Baiker, J. Mol. Catal. A 163 (2000) 205.
12] A. Baiker, J. Mol. Catal. A 115 (1997) 473.
13] E. Sipos, A. Tungler, G. Fogassy, J. Mol. Catal. A 216 (2) (2004) 171.
14] T. Mallat, A. Baiker, in: R.A. Sheldon, H. van Bekkum (Eds.), Fine
Chemicals Through Heterogeneous Catalysis, Wiley–VCH, Wein-
heim, 2001, p. 449.
[26] C. Reichardt, in: Solvents and Solvent Effects in Organic Chemistry,
VCH, Weinheim, 1988, p. 359.
[27] Z. Ma, F. Zaera, J. Phys. Chem. B 109 (2005) 406.
[
[
[
[
[
[28] S. Diezi, A. Szabo, T. Mallat, A. Baiker, Tetrahedron Asymm. 14
(2003) 2573.
[29] D. Ferri, T. Bürgi, A. Baiker, Chem. Commun. (2001) 1172.
[30] D. Ferri, T. Bürgi, J. Am. Chem. Soc. 123 (2001) 12074.
[31] J. Kubota, F. Zaera, J. Am. Chem. Soc. 123 (2001) 11115.
[32] P.J. Stephens, F.J. Devlin, A. Aamouche, ACS Symp. Ser. 810 (2002)
[
[
[
[
[
15] H.U. Blaser, H.P. Jalett, D.M. Monti, J.F. Reber, J.T. Wehrli, Stud.
Surf. Sci. Catal. 41 (1988) 153.
16] Y. Huang, J. Chen, H. Chen, R. Li, Y. Li, L.-E. Min, X. Li, J. Mol.
Catal. A Chem. 170 (2001) 143.
17] Y. Huang, Y. Li, J. Hu, P. Cheng, H. Chen, R. Li, X. Li, C.W. Yip,
A.S.C. Chan, J. Mol. Catal. A Chem. 189 (2002) 219.
18] W. Xiong, H.X. Ma, H. Chen, Y.Y. Huang, Y.L. Huang, X.J. Li, X.S.
Li, J.Y. Qi, A.S.C. Chen, Chem. J. Chin. Univ. Chin. 23 (2002) 1758.
19] H.X. Ma, H. Chen, Q. Zhang, X.Y. Li, J. Mol. Catal. A Chem. 196
18.
[33] P.L. Polavarapu, J. He, Anal. Chem. 76 (2004) 61A.
[34] T.B. Freedman, X. Cao, R.K. Dukor, L.A. Nafie, Chirality 15 (2003)
743.
[35] O.J. Sonderegger, G.M.-W. Ho, T. Bürgi, A. Baiker, J. Mol. Catal. A:
Chem., in press.
[36] A. Vargas, T. Bürgi, A. Baiker, J. Catal. 222 (2003) 439.
(2003) 131.
[37] R. Hess, T. Mallat, A. Baiker, J. Catal. 218 (2003) 453.