Diastereoselective hydrogenation of (S)-proline-2-methylanilide

Article abstract of DOI:10.1006/jcat.1999.2525

The diastereoselective hydrogenation of o-toluidine covalently linked to the chiral auxiliary (S)-proline has been studied. The hydrogenation of (S)-proline-2-methylanilide on supported noble metal catalysts yielded both the cis and the trans isomers of (

Full text of DOI:10.1006/jcat.1999.2525

Journal of Catalysis 185, 479–486 (1999)
Article ID jcat.1999.2525, available online at http://www.idealibrary.com on
Diastereoselective Hydrogenation of (S)-Proline-2-methylanilide
Vidyadhar S. Ranade and Roel Prins
Laboratory for Technical Chemistry, Swiss Federal Institute of Technology (ETH), CH 8092 Zurich, Switzerland
Received January 8, 1999; revised April 12, 1999; accepted April 12, 1999
alkaloids. It can be formed by hydrogenation of the corre-
sponding substituted aromatic compound. There are two
very important advantages in using a substituted aromatic
compound as a precursor for substituted cyclohexane com-
pounds. One is the ease of introducing various functional
groups into the aromatic ring via electrophilic and nucle-
ophilic substitution, and the other is the availability of a
considerable amount of literature on the hydrogenation of
substituted aromatic compounds using heterogeneous cata-
lysts (5–7). Thus, stereoselective hydrogenation of aromatic
rings should be of potential interest. Although there are
several reports regarding the influence of process condi-
tions and the nature of the catalyst on the cis–trans selec-
tivity in the hydrogenation of substituted aromatics using
heterogeneous catalysts (8), there are very few examples
illustrating diastereoselective hydrogenation of aromatics
(9), using either homogeneous or heterogeneous catalysts
(10, 11). Good results, however, have been obtained by
noncatalytic means (12, 13).
In this paper we present the use of (S)-proline auxiliary-
aided diastereoselective hydrogenation of o-toluidine,
using supported noble metal catalysts. Besson et al. have
already reported the diastereoselective hydrogenation of
o-toluic acid using the same auxiliary (10, 14, 15); moder-
ate diastereoselectivity and a good cis to trans selectivity
were reported. In a recent article, they reported excel-
lent diastereoselectivity with pyroglutamic acid as the chi-
ral auxiliary (11). Our aim was to extend the approach of
diastereoselective hydrogenation from aromatic carboxylic
acids to aromatic amines. Although noble metal catalysts
are frequently employed in hydrogenation reactions in
organic syntheses, very little data are available regarding
the nature of the active site in these kinds of reactions. We
have also attempted to delineate the changes in the nature
of a rhodium catalyst during the hydrogenation reaction.
The diastereoselective hydrogenation of o-toluidine covalently
linked to the chiral auxiliary (S)-proline has been studied. The
hydrogenation of (S)-proline-2-methylanilide on supported noble
metal catalysts yielded both the cis and the trans isomers of (S)-
proline-2-methylcyclohexylamide. Rhodium and ruthenium were
found to be the most active catalysts, rhodium being more selec-
tive than ruthenium. With (S)-proline as the chiral auxiliary, the
1R,2S cis-cyclohexanediyl isomer formed preferentially and with
(R)-proline the 1S,2R cis-cyclohexanediyl isomer formed. No di-
astereoselectivity was obtained for the trans isomers. The activity
of rhodium catalysts was dependent on the rhodium salt used as a
precursor in catalyst preparation and on the support and the process
conditions, while the cis–trans selectivity was independent of these
parameters. The diastereoselectivity was dependent only on the na-
ture of the noble metal. Prereduced and reused rhodium catalysts
were less active than fresh ones. EXAFS and XANES showed that
the catalyst was in an incompletely reduced state during the re-
c
action.
1999 Academic Press
Key Words: stereoselective hydrogenation; auxiliary-aided hy-
drogenation; noble metal catalysts; proline auxiliary; toluidine hy-
drogenation.
INTRODUCTION
The chiral element necessary for a stereoselective reac-
tion has to be incorporated either in the substrate (diastere-
oselective reaction) or in the catalyst (enantioselective re-
action). In the latter case spectacular results have been
achieved with homogeneous catalysts in which the metal
atom is surrounded by chiral ligand(s) (1, 2). In the case
of heterogeneous catalysis, noteworthy results have been
obtained only with a limited number of reactions like hy-
drogenation of -keto compounds on cinchona-modified
platinum catalysts (3) and hydrogenation of -keto com-
pounds on tartaric acid-modified nickel catalysts (4). The
former approach, in which the chiral element is incorpo-
rated in the substrate has received less attention, partly due
to the success of chiral homogeneous catalysts and partly
because the chiral element has to be introduced stoichio-
metrically.
EXPERIMENTAL
Synthesis of Substrate and Reference Compounds
All organic chemicals, except 2-methylcyclohexylamine
A substituted cyclohexane ring is a constituent of several (Aldrich), were supplied by Fluka. (S)-proline-2-methyl-
biologically active molecules like terpenoids, steroids, and anilide was synthesized by coupling the carboxylic acid
479
0021-9517/99 $30.00
c
Copyright
1999 by Academic Press
All rights of reproduction in any form reserved.

480
RANADE AND PRINS
TABLE 1
group of a butyloxycarbonyl (BOC) protected (S)-proline
molecule with the amine group of o-toluidine, according to
the following procedure. BOC-(S)-proline (20 g (93 mmol))
was dissolved in 150 ml THF and 13.0 ml (93 mmol) of
triethylamine (TEA) was added to the solution. The mix-
ture was cooled to 15 C and 13.3 ml (102 mmol) of
isobutylchloroformate was added dropwise with stirring
followed by 20 ml (187 mmol) of o-toluidine after 5 min.
The mixture was allowed to reach room temperature and Rh/Al2O3 III
stirred for a couple of hours. The solution was concentrated
in vacuo to about 50 ml and then poured into 300 ml of a
1 N HCl solution. This solution was extracted twice with
Catalyst Precursors and Solvents Used in Preparation of Rhodium
and Platinum Catalysts by the Incipient Wetness Technique
Catalyst
Precursor
Solvent
H/M
Rh/c
Rh/Al2O3
Rh/Al2O3 II
na.
—
Acetone
Water
0.29
0.27
0.74
0.67–0.70
0.27
0.19
I
Rh(C5H8O2)3
Rh(NO3)3 2H2O
RhCI3 3H2O
Rh(C5H8O2)3
Rh(C5H8O2)3
PtCl4
Water
Rh/TiO2
Rh/ZrO2
Pt/Al2O3
Acetone
Acetone
Water
Not determined
300 ml diethyl ether. The diethyl ether extracts were pooled
together and washed twice successively with a sodium bi-
carbonate solution and water and then dried over sodium
sulfate. The solvent was removed in vacuo to give 22.5 g of
solid BOC-(S)-proline-2-methylanilide (yield, 80% ), which
was purified by recrystallization from hexane.
Deprotection of proline was achieved by adding the
purified BOC-(S)-proline-2-methylanilide to pure trifluo-
roacetic acid at 0 C, with stirring. The stirring was then
continued for a couple of hours at room temperature. The
excess of trifluoroacetic acid was removed in vacuo and
sodium hydroxide solution was added till the pH was 13.
The aqueous solution was extracted twice with diethyl ether
and the extracts were treated as in the previous case. Re-
crystallization from hexane afforded 6 g pure (>98% ) (S)-
proline-2-methylanilide (yield, 40% ). The scheme of the
preparation of the substrate is shown in Fig. 1. The NMR
spectral data for (S)-proline-2-methylanilide are given in
the appendix.
Identification of the absolute configuration was done by
preparing a reference racemic mixture and the 1R,2S opti-
cally pure cis isomer (through synthetic route). The exper-
imental details of preparation of the reference compounds
are given in the appendix.
Catalyst Preparation and Characterization
Rh/C (5 wt% ) (Aldrich, 20,616-4), 5 wt% Ru/C (Fluka,
84031), and 10 wt% Pd/C (Fluka, 75990) were used as
supplied. Several other catalysts were made by the in-
cipient wetness technique using various supports (Al2O3,
TiO2, and ZrO2) and precursors as shown in Table 1. TiO2
(anatase; BET area, 54 m2/g) and ZrO2 (BET area, 54 m2/g)
were used as obtained from Degussa. Al2O3 granules from
Condea were crushed to particles smaller than 60 m (BET
area, 230 m2/g). All rhodium salt precursors were supplied
by Johnson Matthey and the platinum salt was supplied by
Fluka. The catalysts were prepared with a metal loading of
5 wt% and were dried in an oven at 120 C and then calcined
in air at 350 C for 3 h after impregnation. These calcined
catalysts were used for the hydrogenation experiments in
the autoclave. In order to quantify the active surface area of
the catalyst, hydrogen chemisorption was used. The cata-
lysts were first reduced at 300 C for 1 h in hydrogen and
outgassed under vacuum at the same temperature for the
same period of time. The desorption isotherms were mea-
sured at 20 C and the H/M value was determined by ex-
trapolating the part of the isotherm between 10 and 50 kPa
to zero hydrogen pressure. Table 1 lists theH/M values
for various rhodium catalysts. In the case of the Rh/Al2O3
III catalyst, three batches of catalyst were prepared with
dispersions ranging from 0.67 to 0.70.
Catalytic Hydrogenation
Experiments were conducted in a 200-ml stainless steel
autoclave at a total pressure of 4.0 MPa and a temperature
of 70 C. In a typical experiment 400 mg of (S)-proline-2-
FIG. 1. Synthesis of (S)-proline-2-methylanilide.

DIASTEREOSELECTIVE HYDROGENATION OF (S)-PROLINE-2-METHYLANILIDE
481
methylanilide was dissolved in 60 ml ethanol. To this so-
lution, 200 mg of supported metal catalyst added in the oxi-
dic form and the slurry was transferred to the autoclave. The
autoclave was closed and flushed three times with nitrogen
and subsequently three times with hydrogen. The hydro-
gen pressure was then increased to 3.7 MPa and stirring
(1000 rpm) was started. The subsequent heating of the
autoclave to 70 C resulted in a total pressure of about
4.0 0.2 MPa. Changing the stirring speed to 800 and
1200 rpm did not change the rate of the reaction, estab-
lishing the absence of diffusion limitations. All subsequent
reactions were conducted with a stirring speed of 1000 rpm.
Since the autoclave wasbeingheated to 70 C duringthe first
few minutes of reaction, it was not possible to obtain reli-
able initial reaction rate values. The activities (ACT) are
therefore reported as reciprocal of the time required for
50% conversion of the substrate (ACT = 1/t50% ).
The autoclave was equipped with a sampling tube, en-
abling periodic sampling of the liquid phase (sample vol-
ume, 0.5–1 ml). The samples were analyzed using a GC
equipped with a -cyclodextrin column and a FID de-
tector. The response factor was assumed to be 1 for
the substrate and the hydrogenated products. In a typi-
cal chromatogram all the diasteromers of (S)-proline-2-
methylcyclohexylamide were separated from each other
as well as from the intermediate cyclohexene derivative
(identified by GC-MS) and the reactant. The position of
the double bond in the cyclohexene intermediate could not
be identified.
FIG. 2. Hydrogenation of (S)-proline-2-methylanilide ( position of
double bond in the cyclohexenic intermediate is not known).
auxiliary. The selectivity between the cis isomers is reported
as the d.e. (diastereomeric excess), where d.e. is defined as
[1R,2S isomer] [1S,2R isomer]
d.e =
.
[1R,2S isomer] + [1S,2R isomer]
Typically, the d.e. was about 45% for the rhodium catalysts
and increased slightly with the conversion of the substrate.
The reproducibility of the diastereoselectivity was about
2% . The concentration of the intermediate went through
a maximum during the reaction. Kinetic data of the reac-
tion indicated that the cyclohexene intermediate washydro-
genated simultaneously with the substrate and, in standard
experiments, it reached a maximum concentration between
4 and 8% of the initial reactant concentration.
X-ray Absorption Experiments
X-ray absorption spectra of the catalysts at the Rh–K
edge were measured in the transmission mode at the syn-
chrotron facility at the Swiss–Norwegian Beamline (SNBL)
at ESRF, Grenoble. The catalyst samples were prepared by
reduction and/or reaction in the autoclave at 4 MPa hydro-
gen pressure and 70 C. After reduction and/or reaction and
after filtering off the solvent, the catalysts were transferred
to the sample cells in a nitrogen environment. The sample
cell was then closed and the measurements were performed
at room temperature without exposure to the atmosphere.
Reference measurements were made on the calcined cata-
lyst and a rhodium foil.
Comparison of Different Catalysts
The results of the hydrogenation reaction with vari-
ous noble metal catalysts are summarized in Table 2. All
TABLE 2
Activity and Selectivity of Different Noble Metal Catalysts in
the Hydrogenation of (S)-Proline-2-methylanilide at 70 C and
4.0 MPa in Ethanol
RESULTS
Two cis and two trans hydrogenated products as well as a
cyclohexene intermediate were obtained in the hydrogena-
tion of (S)-proline-2-methylanilide (Fig. 2). The cis to trans
ratio was between 4 and 5 for all the experiments. The two
trans isomers were formed without any selectivity. In the
case of the cis isomers, the 1R,2S isomer was always ob-
tained in excess when (S)-proline was used as the chiral
1
Catalyst
ACT (h
)
d.e. (% )
Rh/Al2O3
Rh/C
Ru/C
Pt/Al2O3
Pd/C
I
0.50
0.75
0.47
55
41
11
Negligible
Negligible

482
RANADE AND PRINS
TABLE 3
Influence of Prereduction, Reuse of Catalyst, and Addition
of Diethylamine
Activity and Selectivity of Different Rhodium Catalysts in Hy-
drogenation of (S)-Proline-2-methylanilide at 70 C and 4.0 MPa in
Ethanol
As described in the experimental section, the catalyst
was introduced in the oxidic form in the autoclave along
with the substrate and the solvent. Reduction of the catalyst
took place on introducing hydrogen and subsequent heat-
ing of the autoclave to 70 C. This procedure gave relatively
good activities for the rhodium catalysts. Prereduction of
the Rh/C catalyst under hydrogen (before or after intro-
duction into the autoclave in the absence of ethanol) led to
a drastic reduction in the rate of the reaction. Prereduction
of the catalyst in the solvent under hydrogen pressure led
to an even lower rate and d.e. Similar results were obtained
with the Rh/Al2O3 III catalyst.
1
Catalyst
BET area (m2/g)
H/M
ACT (h
)
d.e. (% )
Rh/TiO2
Rh/C
Rh/ZrO2
Rh/Al2O3
Rh/Al2O3 II
Rh/Al2O3 III
54
944
54
227
227
227
0.27
0.29
0.19
0.27
0.74
0.67
0.75
0.75
0.27
0.50
1.20
1.50
56
41
57
55
49
46
I
reactions were conducted at 70 C and 4.0 MPa, using
ethanol as the solvent. At higher conversion considerable
amounts of side products were obtained over the ruthenium
catalyst.
After a standard experiment with the Rh/Al2O3 III cata-
lyst, the reaction mixture was decanted to remove most of
the solvent and substrate and the catalyst was repeatedly
washed in ethanol, untilthe ethanolwasalmost free ofprod-
uct (as detected by GC analysis). The catalyst was then used
again for another hydrogenation reaction. This reused cata-
lyst exhibited a much lower activity and the reaction almost
stopped at about 50% conversion of the substrate.
In a separate experiment, one equivalent of diethylamine
was added to the reaction mixture before it was pressurized
under hydrogen. Diethylamine was used as the additive be-
cause its basicity is comparable to that of the secondary
amine group present in the substrate. The addition of di-
ethylamine resulted in a marginal increase in the diastere-
oselectivity, but in a large reduction in the reaction rate.
Because of the lower activity of the palladium and plat-
inum catalysts and the low selectivity of the ruthenium cata-
lyst, rhodium metal catalysts were chosen for further stud-
ies. With rhodium acetylacetonate as the salt precursor, a
series of rhodium catalysts were prepared on different sup-
ports (viz. Rh/TiO2, Rh/ZrO2, and Rh/Al2O3 I). The activ-
ities of these catalysts differed widely (Table 3) but they
had an identical diastereoselectivity. With the same Al2O3
as the support, three rhodium catalysts were prepared us-
ing different metal precursors, as indicated in Table 1. The
activity and to some extent the diastereoselectivity of these
catalysts depended on the rhodium salt used in their prepa-
ration (Table 3).
EXAFS and XANES Studies on the Catalyst
EXAFS measurements and analyses were performed on
the Rh/Al2O3 III catalyst to obtain information about the
change in the oxidation state of rhodium during reaction.
The EXAFS data were fitted using the k1 weighted Fourier
transformed (k) function. Details of the analyses and in-
terpretation of the spectra can be found elsewhere (16). The
coordination numbers obtained from the EXAFS analysis
and the percentage of reduced rhodium obtained from the
XANES data are reported for various samples in Table 5.
The data indicate that the reduction of rhodium metal is
Influence of Process Conditions: Temperature, Pressure,
and Solvent
The reaction temperature had a considerable influence
on the activity of the catalyst. Thus, after about 500 min
of reaction time, only 40% conversion of the reactant was
obtained when the reaction was carried out at 50 C with
the Rh/Al2O3 III catalyst, while at 70 C complete conver-
sion was obtained. With the Rh/Al2O3 III catalyst, the re-
action rate as well as the selectivity remained almost un-
affected on increasing the reaction pressure from 4.0 to
6.0 MPa.
TABLE 4
Solvents are known to play a significant role in deter-
mining the stereoselectivity in hydrogenation reactions. Be-
sides ethanol, which was used as the standard solvent, the
use of tetrahydrofuran, ethyl acetate, and chloroform was
investigated. In the case of chloroform, the reactant was
slowly converted to an unidentified product without any
selectivity toward the ring hydrogenation product. The ac-
tivity in tetrahydrofuran and ethyl acetate was much lower
than that in ethanol but a slightly higher diastereoselectivity
was obtained (Table 4).
Effect of Solvent on the Selectivity in the Hydrogenation of
(S)-Proline-2-methylanilide with Rh/Al2O3 III Catalyst at 70 C
and 4.0 MPa
1
Solvent
ACT (h
)
d.e. (% )
Ethanol
1.50
0.50
0.11
—
46
56
52
Tetrahydrofuran
Ethyl acetate
Chloroform

DIASTEREOSELECTIVE HYDROGENATION OF (S)-PROLINE-2-METHYLANILIDE
483
TABLE 5
EXAFS and XANES Results for the Rh/Al2O3 III Catalyst
Coordination
numbera
Percentage of
reduced Rhb
Rh–Rh
Rh–O
Catalyst after reduction in EtOH
Catalyst after reaction in EtOH
Catalyst as calcined
6.1
5.3
0.0
1.2
1.4
5.3
69
53
0
a Obtained from EXAFS.
b Obtained from XANES.
incomplete during the reduction or reaction, under the typ-
ical process conditions. The catalyst is reduced to a lesser
extent during reaction (53% ) than during reduction (69% ).
DISCUSSION
The 1R,2S cis diastereomer was formed in excess when
(S)-proline was used as the chiral auxiliary, while on hy-
drogenating a substrate made by coupling of o-toluidine
with (R)-proline instead of (S)-proline, the 1S,2R diastere-
omer was formed in excess. This demonstrates that the pro-
line auxiliary is responsible for the chiral induction in the
hydrogenation reaction. The values of d.e. obtained with
(S)-proline and (R)-proline were approximately equal but
opposite in sign. Thus we can think of the following two
speculative structures of the substrate on the rhodium sur-
face for (R)-proline-2-methylanilide and for (S)-proline-2-
methylanilide (top two structures in Fig. 3) which result
in the formation of the R0,1S,2R and the S0,1R,2S diastere-
omers, respectively, on addition of hydrogen from the metal
face of the rhodium catalyst. In these two structures, the ni-
trogen atom of the proline ring and the aromatic ring lie in
the same plane on the catalyst surface and the amide bond
is approximately parallel to the surface of the catalyst. The
other cisdiastereomer isformed from these substrateswhen
the adsorption through the benzene ring takes place by its
opposite face (bottom two structures in Fig. 3). However,
this configuration leads to more steric crowding because of
the interaction of the methyl group with the carbonyl group
and hence is the less preferred structure.
FIG. 3. Top views of the adsorption-structure of proline-2-meth-
ylanilide. From top left clockwise, formation of R0,1S,2R; S0,1R,2S;
S0,1S,2R, and R0,1R,2S isomers.
Partialhydrogenation ofa substituted aromaticringleads
to the formation of cyclohexene and cyclohexadiene deriva-
tives. Intermediate cyclohexadiene species have rarely been
observed in hydrogenation of aromatic compounds (17),
but cyclohexene species have often been detected. Hydro-
genation of these cyclohexene species is responsible for the
formation of trans isomers. Although many different cyclo-
hexene intermediates can be formed, only the five species
shown in Fig. 4 are relevant in the present reaction for the
formation of trans isomers. In the case of intermediate I,
trans formation can result only if the addition of the two
FIG. 4. Formation and interconversion of cyclohexenic intermediates.

484
RANADE AND PRINS
All the catalysts prepared from the same rhodium pre-
cursor but with different oxide supports give similar results
in terms of diastereoselectivity (viz. Rh/TiO2, Rh/ZrO2,
and Rh/Al2O3 I, Table 3). The difference in diastereoselec-
tivity between the oxide-supported catalysts and carbon-
supported catalyst could be due to the difference in the
nature of the surface of the two supports. However, this
difference in selectivity may also be due to the different pre-
cursor, which has been used in the preparation of the cata-
lysts. (Since the Rh/C catalyst was purchased from Aldrich,
the rhodium salt precursor is not known for this catalyst).
For the three catalysts prepared on the same Al2O3 support
but from different precursors (viz. Rh/Al2O3 I, Rh/Al2O3 II,
and Rh/Al2O3 III), no clear correlation is observed between
the activity of the catalysts and their dispersion (Table 3).
In general a higher activity is obtained at higher dispersion.
The EXAFS data on the calcined Rh/Al2O3 III catalyst
indicated that the chlorine originating from the rhodium
chloride precursor was probably not completely removed
during the calcination step. Thus the possibility of residual
FIG. 5. Example of the mechanisms for interconversion of the cyclo-
hexenic intermediates (R, -pyrrolidine-2-carboxyl).
hydrogen atoms takes place from the opposite faces on the chlorine influencing the activity of this catalyst cannot be
aromatic ring. This is possible if the catalyst has a stepped excluded. The Rh/Al2O3 II catalyst was, however, free of
surface or if the hydrogen atom comes from the solvent (in nitrate as confirmed by FT-IR measurements on the im-
the case of protic solvents or acidic conditions) (18). Since pregnated and the calcined catalyst samples.
the cis to trans ratio was independent of the solvent and the
Rhodium is a very good catalyst for hydrogenation of
metal dispersion, these possibilities do not seem likely. The aromatic compounds since it can catalyze these reactions
maximum concentration of the intermediate detected in at relatively mild process conditions as compared to other
the solution phase ( 6% ) cannot be correlated with the catalysts. However, the present substrate hydrogenated
amount of trans products (20% ) because the intermediate very slowly at room temperature. Hence, the reaction was
and the reactant are hydrogenated simultaneously. Thus the conducted at 70 C in order to get a reasonable activity. The
formation of trans isomers exclusively from the intermedi- relatively strong dependence of the rate of reaction on tem-
ate I via its conversion, on the surface of the catalyst by a
-
perature is indicative of a strong adsorption of the substrate
allyl mechanism or intramolecular hydrogen transfer, to the on the catalyst. The substrate molecule is most probably
intermediates II and III is a possibility (Fig. 5). In addition, preferentiallyadsorbed through the secondaryamine group
the trans products could be formed directly through the hy- on the proline ring, as is evident from the experiments in
drogenation of intermediates II and III or by intramolec- which diethylamine is added to the reaction mixture. The
ular hydrogen transfer followed by hydrogenation (8, 19). product of the hydrogenation reaction is also an amine and
The slow hydrogenation of the substrate probably allows hence the total concentration of amine groups remains con-
equilibrium to be reached between various hydrogenation stant in the reaction mixture. A negligible effect of the hy-
and isomerization processes of the intermediate(s). This ac- drogen pressure on the reaction rate suggests that the re-
counts for the independence of the cis to trans ratio from action is product-desorption controlled. In the diastereose-
either the process parameters or the catalyst preparation. lective hydrogenation of (S)-proline modified o-toluic acid,
The d.e. is strongly dependent only on the nature of the no- Besson et al. (15) have reported enhancement of the di-
ble metal and fairly independent of the process conditions. astereoselectivity and a decrease in the reaction rate on the
A relatively small dependence of the diastereoselectivity addition of an amine. In the present case no big enhance-
on the process parameters but a strong dependence on the ment in the diastereoselectivity was observed because the
molecular structure has been observed in other (nonaro- reaction mixture already contained a significantly high con-
matic) diastereoselective hydrogenations (20).
centration of amine groups.
As has been reported earlier for different kinds of aro-
Use of aprotic solvents like tetrahydrofuran and ethyl
matic substrates (9), ruthenium and rhodium were the most acetate resulted in a marginally higher d.e. (Table 4). As
active catalysts for the present hydrogenation reaction. observed before in the hydrogenation of other aromatic
The platinum and palladium catalysts were almost inactive substrates, the rate was fastest in the alcoholic solvent (6).
probably because of strong adsorption of the amine group The reason for a 14-fold higher activity in ethanol than in
on those catalysts (5).
ethyl acetate is unclear. The hydrogenation failed when

DIASTEREOSELECTIVE HYDROGENATION OF (S)-PROLINE-2-METHYLANILIDE
485
chloroform was used as the solvent, probably because of Preparation of Racemic cis- and Racemic trans-(S)-
poisoning of the active sites by chloride resulting from chlo-
roform (5).
Proline-2-Methylcyclohexylamide
(as Reference Compounds)
The catalyst is not reduced completely during either pre-
reduction or reaction as indicated in Table 5. As the catalyst
becomes reduced, it loses its activity as shown by experi-
ments in which the catalyst is prereduced. Accordingly a
reused catalyst also exhibits lower activity than a fresh one.
The loss in activity occurs irrespective of whether carbon or
alumina is used as the support. The EXAFS data reported
for catalyst Rh/Al2O3 III in Table 5 indicate that the par-
ticle size remains unchanged during the reduction of the
catalyst. Therefore sintering of small metal particles during
reaction can be excluded as an explanation for the loss in
catalyst activity. The loss in activity is therefore probably
due to loss of the active sites, either due to a change in the
morphology of the rhodium particles or due to a change in
the oxidation state of the rhodium.
A small amount of BOC-proline was dissolved in 40 ml
chloroform and the solution cooled to 0 C under ni-
trogen. Triethylamine (3 eq.), 1.2 eq. hydroxybenzotria-
zole, 1 eq. 2-methylcyclohexylamine (cis : trans = 1 : 3), and
1.2 eq. N-(3-dimethylaminopropyl)-N0-ethyl-carbodiimide
hydrochloride were added to the cooled solution in suc-
cession. The solution was allowed to warm to room tem-
perature and was stirred overnight under nitrogen. After
washing the solution three times with 1 N HCl and twice
with saturated sodium bicarbonate solution and finally with
a saturated salt solution, the solvent was removed in vacuo
to yield racemic cis- and racemic trans-BOC-(S)-proline-2-
methylcyclohexylamide with a cis–trans ratio of 1 : 3. De-
protection of the racemic amides was carried out in au-
toclavable sample bottles using a solution of dry HCl in
dioxane. Thus, racemic cis- and racemic trans-(S)-proline-2-
methylcyclohexylamide with a cis to trans ratio of 1 : 3 were
obtained. The dioxane was then evaporated in a stream of
air and about 0.5 ml of dichloromethane and 0.05 ml of per-
fluoropropionic anhydride were added. The mixture was
heated to 110 C for 5 min after sealing the sample bot-
tle. The excess anhydride was evaporated in a stream of air
along with dichloromethane and the contents of the sample
bottle were dissolved in fresh dichloromethane. The perflu-
oropropionic acid derivatives of the racemic cis- and trans-
(S)-proline-2-methylcyclohexylamide were then injected
into a GC equipped with a RTX-200 capillary column.
The isolated product of catalytic hydrogenation was deriva-
tized similarly. Comparison of the chromatogram and mass
spectra of the derivatized racemic mixture to those of the
derivatized product of catalytic hydrogenation confirmed
that the cis products were obtained in excess during cata-
lytic hydrogenation.
CONCLUSIONS
Ruthenium and rhodium are the most active metal cata-
lysts for the hydrogenation of the (S)-proline-2-methylani-
lide among the noble metal catalysts investigated. The se-
lectivity is strongly dependent on the substrate structure
since relatively small differences are observed between
the different rhodium catalysts. This has also been ob-
served in other diastereoselective heterogeneous catalytic
hydrogenations. The hydrogenation activity is dependent
on the temperature, the catalyst precursor and support, and
the solvent. The substrate adsorbs preferentially through
the amine group and the diastereoselectivity arises from
adsorption of the substrate on a rhodium surface prefer-
entially through one of the two diastereotopic faces of the
aromatic ring. ZrO2, TiO2, and Al2O3 yield comparable re-
sults in terms of the selectivity in the case of rhodium cata-
lysts. The catalyst deactivates as it is reduced during the
reaction. EXAFS and XANES experiments indicate that
only a part of the rhodium in the catalyst is reduced under
reaction conditions. A partly reduced catalyst is thus more
active than a fully reduced catalyst.
Preparation of (S)-Proline-(1R, 2S)-2-methylcyclo-
hexylamide (as Reference Compound)
The optically active BOC-(S)-proline-(1R,2S)-2-methyl-
cyclohexylamide cis isomer was synthesized as described
above except that the optically active (1R,2S)-2-methyl-
cyclohexylamine hydrochloride (synthesized using the
procedure reported by Knupp and Frahm (21)) was used
APPENDIX
Spectral Data of (S)-Proline-2-methylanilide
The NMR spectra were recorded on a Bruker AMX 500 instead of racemic 2-methylcyclohexylamine. Deprotection
instrument at room temperature, using CDCl3 as the sol- ofBOC-(S)-proline-(1R,2S)-2-methylcyclohexylamide and
vent.
subsequent derivatization of the optically pure (S)-proline-
1H NMR : 9.83(br, 1H), 8.12 (d, 1H, J = 7.8), 7.20 (1R,2S)-2-methylcyclohexylamide isomer was carried out
(t, 1H, J = 7.7), 7.15 (d, 1H, J = 7.4), 7.02 (td, 1H, J = 7.5, as reported above for the racemic product. Comparison of
1.1), 3.90 (dd, 1H, J = 9.2, 4.9), 3.10 (m, 1H), 3.00 (m, 1H), the chromatogram and the mass spectrum of the deriva-
2.29 (s, 3H), 2.19 (m, 1H), 2.07 (m, 1H), 1.78 (m, 2H); 13C tized optically pure product to those of derivatized product
NMR : 173.1, 136.0, 130.2, 127.4, 126.8, 124.1, 120.8, 61.3, of catalytic hydrogenation, as for the racemic mixture, re-
47.4, 30.8, 26.3, 17.6; GC-MS: M+ = 204.
vealed the absolute configuration of the cis products.

486
RANADE AND PRINS
A. G., “Stereoselectivity of Heterogeneous Metal Catalysis.” Wiley,
ACKNOWLEDGMENT
New York, 1985.
The authors are grateful to Prof. G. Consiglio for helpful discussions,
F. Bangerter for NMR measurements, and Dr. T. Shido for the EXAFS
and XANES experiments.
9. Besson, M., and Pinel, C., Topics Catal. 5, 25 (1998).
10. Besson, M., Blanc, B., Campelet, M., Gallezot, P., Nasar, K., and Pinel,
C., J. Catal. 170, 254 (1997).
11. Besson, M., Gallezot, P., Neto, S., and Pinel, C., Chem. Commun. 14,
1432 (1998).
REFERENCES
12. Schultz, A. G., Acc. Chem. Res. 23, 207 (1990).
13. Donohoe, T. J., Garg, R., and Stevenson, C. A., Tetrahedron:
Asymmetry 7, 317 (1996).
14. Besson, M., Blanc, B., Gallezot, P., Nasar, K., and Pinel, C., in
“Catalysis of Organic Reactions” (R. E. Malz, Jr., Ed.), Dekker,
New York, 1996.
15. Besson, M., Gallezot, P., Pinel, C., and Neto, S., Stud. Surf. Sci. Catal.
108, 215 (1997).
16. Gloor, A. P., and Prins, R., J. Phys. Chem. 98, 9865 (1994).
17. Stanislaus, A., and Cooper, B. H., Catal. Rev. Sci. Eng. 36, 75
(1994).
1. Noyori, R., “Asymmetric Catalysis In Organic Synthesis.” Wiley,
New York, 1994.
2. Kagan, H. B., in “Comprehensive Organometallic Chemistry”
(G. Wilkinson, F. G. A. Stone, and E. W. Abel, Eds.), Vol. 8.
Pergamon Press, Oxford, 1982.
3. Blaser, H. U., Tetrahedron: Asymmetry 2, 643 (1991)
4. Izumi, Y., Adv. Catal. 32, 215 (1983).
5. Freifelder, M., “Practical Catalytic Hydrogenation: Techniques and
Application.” Wiley, New York, 1971.
6. Rylander, P. N., “Catalytic Hydrogenation over Platinum Metals.”
Academic Press, New York, 1967.
18. Van Rantwijk, F., van Vliet, A., and van Bekkum, H., J. Mol. Catal.
9, 283 (1980).
7. Augustine, R. L., “Heterogeneous Catalysis for the Synthetic
Chemist.” Dekker, New York, 1996.
19. Smith, G. V., Wang, Y., Song, R., and Tyrrell, J., in “Catalysis of Organic
Reactions” (F. E. Herkes, Ed.), Dekker, New York, 1998.
8. Bartok, M., Czombos, J., Felfoldi, K., Gera, L., Gondos, G., 20. Tungler, A., and Fodor, K., Catal. Today 37, 191 (1997).
Molnar, A., Notheisz, F., Palinko, I., Wittmann, G., and Zsigmond, 21. Knupp, G., and Frahm, A. W., Chem. Ber. 117, 2076 (1984).