Functional-Group-Directed Diastereoselective Hydrogenation of Aromatic Compounds. 21

Article abstract of DOI:10.1021/jo991604s

Diastereoselective liquid-phase hydrogenation of a series of monosubstituted indan substrates was studied on supported rhodium catalysts. Predominantly the cis-cis diastereomer, obtained by hydrogenation from the diastereoface opposite the substituent at the stereogenic center, and the cis-trans diastereomer, obtained by hydrogenation from the diastereoface on the same side as the substituent, were formed. The diastereoselectivity depends on the balance between steric repulsion and electronic attraction of the substituent with the surface of the catalyst. For alkoxy and carboxyl groups (acid, methyl ester, and amide), the steric repulsion dominated and the cis-cis diastereomer was obtained with moderately high selectivity. The diastereoselectivity obtained in the hydrogenation was influenced by the addition of bases to the reaction mixture. Addition of triethylamine caused a small increase in the selectivity to the cis-cis diastereomer in some substrates, whereas the addition of NaOH significantly increased the selectivity toward the cis-trans isomer in all substrates.

Full text of DOI:10.1021/jo991604s

1132
J . Org. Chem. 2000, 65, 1132-1138
F u n ction a l-Gr ou p -Dir ected Dia ster eoselective Hyd r ogen a tion of
Ar om a tic Com p ou n d s. 21
Vidyadhar S. Ranade, Giambattista Consiglio, and Roel Prins*
Laboratory for Technical Chemistry, Swiss Federal Institute of Technology (ETH),
CH 8092 Zurich, Switzerland
Received October 14, 1999
Diastereoselective liquid-phase hydrogenation of a series of monosubstituted indan substrates was
studied on supported rhodium catalysts. Predominantly the cis-cis diastereomer, obtained by
hydrogenation from the diastereoface opposite the substituent at the stereogenic center, and the
cis-trans diastereomer, obtained by hydrogenation from the diastereoface on the same side as the
substituent, were formed. The diastereoselectivity depends on the balance between steric repulsion
and electronic attraction of the substituent with the surface of the catalyst. For alkoxy and carboxyl
groups (acid, methyl ester, and amide), the steric repulsion dominated and the cis-cis diastereomer
was obtained with moderately high selectivity. The diastereoselectivity obtained in the hydrogena-
tion was influenced by the addition of bases to the reaction mixture. Addition of triethylamine
caused a small increase in the selectivity to the cis-cis diastereomer in some substrates, whereas
the addition of NaOH significantly increased the selectivity toward the cis-trans isomer in all
substrates.
In tr od u ction
Stereoselective syntheses of cyclic compounds by dias-
teroselective hydrogenation of the corresponding substi-
tuted aromatic compounds on heterogeneous catalysts
has received growing attention in recent years.2-5 The
mechanism of face selection during hydrogenation of
modified aromatic compounds is, however, not well
understood. It is of interest therefore to study interactions
(or directing effects) of nonreducible functional groups
individually to obtain an insight into the mechanism of
diastereoselection. To this end, we initiated a study into
the hydrogenation of a series of aromatic substrates
differing only in the functional group substituent at the
same stereogenic center. In part 1 of this series1 we
investigated the hydrogenation of racemic indan sub-
strates substituted at the benzylic position with hydroxyl,
amino, and methyl groups (1, Figure 1). In the present
paper we report the results of hydrogenation of indan
substrates having alkoxy and carboxyl groups at the
benzylic position. The choice of the class of substrates
was dictated by two considerations: the selected mol-
ecules had to be flat so that the interaction of substitu-
ents is well expressed during their adsorption on the
catalyst surface, and they had to be conformationally
rigid. The choice of the benzylic-substituted indans was
advantageous in one more respect in that the relative
configuration of the fully hydrogenated products can be
determined relatively easily. Supported rhodium cata-
lysts were used for hydrogenation because of their high
activity under relatively mild reaction conditions.6,7 We
F igu r e 1. Substrates investigated by Thompson and co-
workers (2 and 3)8,9 and substituted indan substrates (1)
investigated in part 1 and the present study.
also continued our investigations into the influence of the
addition of inorganic and organic bases during reaction
on the diastereoselectivity as in part 1.
Although a systematic study of the directing effects of
different functional groups had not been conducted in
aromatic hydrogenation prior to our investigation, a
corresponding study had been conducted in olefin hydro-
genation.8,9 In their study, Thompson and co-workers
hydrogenated a series of substituted tetrahydrofluorenes8
(2, Figure 1) and 2-exo-substituted 7-methylenenorbor-
nanes9 (3, Figure 1), differing only in the functional group
substituent at one stereocenter, over supported palladium
catalysts. Since the two faces of the olefinic substrates
are not identical, hydrogenation from either of the two
faces results in different products. Hydrogenation pro-
ceeds mainly from the side of the functional group
(proximofacial) if the electronic interaction between this
* To whom correspondence should be addressed. Fax: +41 (1)
6321162, Tel: +41 (1) 6325490, E-mail: prins@tech.chem.ethz.ch.
(1) Part 1: Ranade, V. S.; Consiglio, G.; Prins, R. J . Org. Chem.
1999, 64, 8862-8867.
(6) Rylander, P. N. Catalytic Hydrogenation over Platinum Metals;
Academic Press: New York, 1967.
(2) Besson, M.; Pinel, C. Top. Catal. 1998, 5, 25-38.
(3) Besson, M.; Gallezot, P.; Neto, S.; Pinel, C. Chem. Commun.
1998, 14, 1432-1433.
(7) Freifelder, M. Practical Catalytic Hydrogenation: Techniques &
Application; J ohn Wiley and Sons: New York, 1971.
(8) Thompson, H. W.; Naipawer, R. E. J . Am. Chem. Soc. 1973, 95,
6379-6386.
(9) Thompson, H. W.; Wong, J . K. J . Org. Chem. 1985, 50, 4270-
4276.
(4) Ranade, V. S.; Consiglio, G.; Prins, R. Catal. Lett. 1999, 58, 71-
74.
(5) Ranade, V. S.; Prins, R. J . Catal. 1999, 185, 479-486.
10.1021/jo991604s CCC: $19.00 © 2000 American Chemical Society
Published on Web 01/28/2000

Hydrogenation of Aromatic Compounds
J . Org. Chem., Vol. 65, No. 4, 2000 1133
Sch em e 1
group and the metal surface dominates, whereas it
proceeds predominantly from the side opposite the func-
tional group (distofacial) if steric repulsion dominates.
The relative magnitude of repulsion and attraction of the
substituents with a particular catalyst was gauged by
hydrogenating under identical conditions a series of
olefinic compounds differing only in the substituent. In
principle, the facial selectivity in the hydrogenation of
substituted aromatic compounds should similarly depend
on the attractive or repulsive properties of the substit-
uent(s) with the catalyst. However, palladium is gener-
ally preferred for olefinic hydrogenation, whereas rhod-
ium is extensively used for aromatic hydrogenation.6,7 It
is therefore of interest to compare the effect of substit-
uents in the hydrogenation of olefinic and aromatic
substrates.
To the best of our knowledge, nothing has been pub-
lished on substituent-directed hydrogenation of aromatic
compounds with homogeneous catalysts, probably be-
cause these catalysts lack activity. The influence of
substituents has been exploited, however, in the hydro-
genation of a variety of olefinic substrates, particularly
with soluble rhodium and iridium complexes. Excellent
selectivities have been achieved for many substrates
utilizing the attractive interaction of polar substitutents
such as the hydroxyl and carboxyl groups with the metal
center in homogeneous catalysts.10,11 Some comparative
data are available for the application of heterogeneous
and homogeneous catalysts in directed hydrogenation
reactions.12-14 It indicates that in general heterogeneous
catalysts lag well behind their homogeneous counterparts
in terms of selectivities, even under comparable reaction
conditions.
Ta ble 1. Hyd r ogen a tion of 1-Alk oxyin d a n s
yield of cis- and yield of cis
trans-perhydro- perhydro
indanb
(%)
productsb
(%)
1, G ) catalysta solvent
drb
OMe Rh/C
OMe Rh/C
OMe Rh/Al2O3 EtOH
OMe Rh/Al2O3 hexane
EtOH
hexane
74
36
4
5
78
7
21
60
91
85
19
88
79:21
75:25
88:12
82:18
81:19
92:8
OPr
OPr
Rh/C
Rh/Al2O3 EtOH
EtOH
a
The substrate-to-metal molar ratio for 1-methoxyindan is 128,
b
and that for 1-propoxyindan is 117. Determined by analyses over
a HP-1 capillary column.
Resu lts
Small to significant amounts of cyclohexene intermedi-
ates were observed (identified by GC-MS analysis)
depending on the substrate hydrogenated and the reac-
tion conditions. For most substrates the hydrogenation
of the cyclohexene intermediate proceeded slower than
that of the substrate, and it caused a change in the
diastereoselectivity, the magnitude of which depended on
the substrate. The cyclohexene intermediate hydroge-
nated slower than the corresponding aromatic substrate
probably because of its higher steric hindrance (at-
tributed to its nonplanar structure) during its adsorption
on the catalyst. The position of the double bond in the
cyclohexene intermediates was identified by NMR analy-
ses to be between the junction carbon atoms, only in the
case of the ester 1d .
Results of the hydrogenation of 1-methoxyindan (1a )
are reported in Table 1. Substantial amounts of the
hydrogenolysis-hydrogenation byproducts cis- and trans-
perhydroindan were obtained in addition to the fully
hydrogenated products on the Rh/C catalyst, especially
in ethanol. The fractional yield of these products is also
reported in Table 1 at 100% conversion of the substrates.
Less hydrogenolysis occurred on the Rh/C catalyst in
hexane and very little on the Rh/Al2O3 catalyst. In
general, a high diastereoselectivity to the cis-cis isomer
was obtained over both catalysts, Rh/Al2O3 being more
stereoselective than Rh/C, and reactions in ethanol being
more stereoselective than those in hexane. Hydrogena-
tion of 1-propoxyindan (1b) in ethanol was studied to
investigate the effect of the size of the alkoxy substituent
on the diastereoselectivity. Like in the case of 1-meth-
Hydrogenation of all substrates yielded predominantly
cis (cis-cis + cis-trans) diastereomers (usually cis to
trans ratio >10) with respect to the two substituents on
the six-membered ring (Scheme 1). Different methods
were employed for the identification of the relative
configuration of the cis-cis and cis-trans diastereomers,
depending on the substrate hydrogenated as described
in the Experimental Section. The diastereoselectivity to
the two cis isomers is expressed as the diastereomeric
ratio (dr) at 100% conversion of the substrate and
intermediates and is defined as dr ) [cis-cis]/ [cis-
trans]. For clarity, the dr is presented as a ratio in which
the sum of the yields of the two diastereomers is
normalized to 100. To demonstrate the variation of
diastereoselectivity with reaction time, selectivity and
incremental selectivity to the cis-cis diastereomer are
presented for a few experiments. Selectivity is defined
as [cis-cis]/[cis], and the incremental selectivity is de-
fined as ([cis-cis]2 - [cis-cis]1)/([cis]2 - [cis]1), where
subscript 2 refers to the point of kinetic analysis at which
the incremental dr is calculated and subscript 1 refers
to the previous point of kinetic analysis, and [cis] ) [cis-
cis] + [cis-trans].
(10) Hoveyda, A. H.; Evans, D. A.; Fu, G. C. Chem. Rev. 1993, 93,
1307-1370.
(11) Brown, J . M. Angew. Chem., Int. Ed. Engl. 1987, 26, 190-203.
(12) Crabtree, R. H.; Davis, M. W. J . Org. Chem. 1986, 51, 2655-
2661.
(13) Stork, G.; Kahne, D. E. J . Am. Chem. Soc. 1983, 105, 1072-
1073.
(14) Schultz, A. G.; McCloskey, P. J . J . Org. Chem. 1985, 50, 5905-
5907.

1134 J . Org. Chem., Vol. 65, No. 4, 2000
Ranade et al.
Sch em e 2
F igu r e 2. Kinetics of the hydrogenation of 1-methoxyindan
in hexane over Rh/Al2O3 under standard reaction conditions:
concentration of 1-methoxyindan (9) and its cyclohexene
intermediate (b) and selectivity (2) and incremental selectivity
(1) to the cis-cis diastereomer.
oxyindan, the Rh/C catalyst showed a lower chemoselec-
tivity as well as diastereoselectivity than the Rh/Al2O3
catalyst (Table 1). The diastereoselectivity in the hydro-
genation of 1-propoxyindan was slightly higher than in
that of 1-methoxyindane. In 1-methoxyindan as well as
1-propoxyindan the diastereoselectivity to the cis-cis
isomer decreased as the reaction proceeded to completion.
The decrease in selectivity and incremental selectivity
to the cis-cis diastereomer with conversion in the
hydrogenation of 1-methoxyindan (and its intermediate)
in hexane over the Rh/Al2O3 catalyst are shown in Figure
2. The dr values corresponding to the initial and final
selectivity and incremental selectivity are also indicated
in the figure. The substrate concentration decreases
continuously, whereas the concentration of the interme-
diate goes through a maximum. Toward the end of the
reaction, the fully hydrogenated cis products are formed
primarily due to hydrogenation of the intermediate and
not of the substrate. The selectivity and incremental
selectivity stay constant during the initial period in which
the cyclohexene intermediate is being formed and de-
crease with the conversion of the intermediate to the
products. This indicates that the decrease in the selectiv-
ity to the cis-cis diastereomer and hence the dr with time
is primarily due to the fact that the cyclohexene inter-
mediate hydrogenates with a much lower selectivity to
the cis-cis diastereomer than 1-methoxyindan.15
Results of hydrogenation of the carboxyl substrates,
viz., indan-1-carboxylic acid (1c), the methyl ester of
indan-1-carboxylic acid (1d ), and indan-1-carboxamide
(1e) (Scheme 2), are presented in Table 2. All substrates
yielded predominantly (92-95%) the cis-cis and cis-
trans diastereomers, with a comparable and relatively
high selectivity to the cis-cis diastereomer. The selectiv-
ity is higher with Rh/Al2O3 than with Rh/C. For the ester,
the selectivity is hardly affected by the change in the
solvent from ethanol to hexane. Reactions of the other
substrates were conducted only in ethanol because of
their limited solubility in hexane. The diastereoselectivity
of all substrates was unaffected by changing the substrate-
to-catalyst ratio. The diastereoselectivity to the cis-cis
Ta ble 2. Hyd r ogen a tion of Ca r boxyl Su bstr a tes
1, G )
catalysta
solvent
drb
COOH
Rh/C
Rh/C
Rh/C
Rh/C
Rh/Al2O3
Rh/Al2O3
Rh/Al2O3
Rh/Al2O3
EtOH
EtOH
EtOH
hexane
EtOH
EtOH
EtOH
hexane
77:23
71:28
80:20
80:20
84:16
81:19
85:15
82:18
CONH2
COOMe
COOMe
COOH
CONH2
COOMe
COOMe
a
The substrate-to-metal molar ratio for acid and amide is 127,
b
and that for ester is 117. Yield of the cis diastereomers (>92%)
and dr determined by analyses over a RTX-200 capillary column.
diastereomer decreased significantly, however, in the
ester hydrogenation especially toward the end of the
reaction. For example, in the hydrogenation of the ester
over the Rh/Al2O3 catalyst in ethanol, the dr decreased
from 89:11 at the beginning to 83:17 at the end of the
reaction. In one experiment, the hydrogenation of the
ester was interrupted at an intermediate conversion, and
a mixture of the unconverted substrate, its cyclohexene
intermediate, and fully hydrogenated products was iso-
lated, by removing the catalyst and the solvent. Column
chromatography over silica gel (hexane:ethyl acetate )
9:1) enabled separation of a mixture of the cyclohexene
intermediate and the fully hydrogenated products from
the unconverted substrate. A normal and a DEPT 13C
NMR spectrum of the mixture enabled us to assign the
position of the double bond between the two junction
carbon atoms (see the Supporting Information). The
cyclohexene intermediate in this mixture was hydroge-
nated further over the Rh/Al2O3 catalyst in ethanol under
standard reaction conditions. A dr value of 71:29 was
obtained in this experiment, and the value changed very
little with increasing conversion of the cyclohexene
intermediate. Thus, hydrogenation of the cyclohexene
intermediate takes place with a constant but different
selectivity than that of the parent aromatic substrate.
The change in diastereoselectivity during the reaction
could not be determined in the case of the acid substrate
1c because of insufficient chromatographic resolution and
in the case of the amide substrate 1e because of the
precipitation of the cis-cis diastereomer.
(15) It could be argued that the change in diastereoselectivity is
related to a change in the catalyst with time. We have eliminated this
possibility, at least in the case of 1-methoxyindan, by conducting an
experiment in which a fresh batch of catalyst was added, after removal
of the original catalyst, to a half-converted reaction mixture. The
diastereoselectivity obtained in this experiment was the same as that
obtained in the normal experiment.

Hydrogenation of Aromatic Compounds
J . Org. Chem., Vol. 65, No. 4, 2000 1135
Ta ble 4. Hyd r ogen a tion in Eth a n ol over Rh /C w ith
Tr ieth yla m in e a n d Na OH a d d itives
Sch em e 3
dr
substrate-
to-rhodium without
TEA
NaOH
substrate
molar ratio additive additive additive
1-OMea
128
64
59
59
64
79:21
79:21
81:19
80:20
71:28
94:6
81:19
49:51
55:45
67:33
61:39
59:41
1-OMeb
1-OPrb
93:7
1-COOMeb
79:21
82:18
95:5
b
1-CONH2
3-propyl-3H-
115
isobenzofuran-1-onea
a
The triethylamine to rhodium molar ratio is 10, and the Na-
b
to-rhodium molar ratio is 6. The triethylamine to rhodium molar
ratio is 5, and Na-to-rhodium molar ratio is 3.
Ta ble 3. Hyd r ogen a tion of
3-P r op yl-3H-isoben zofu r a n -1-on e in Eth a n ol
hydrogen
hydrogen
catalyst pressure (bar) drb
catalyst pressure (bar) drb
Rh/Ca
Rh/Ca,d
Rh/Ca
50
50
20
94:6 Rh/Cc
92:8 Rh/Al2O3
95:5
10
50
95:5
94:6
a
a
b
The substrate to metal molar ratio is 115. Yield of the cis
diastereomers (∼99%) and dr determined by 1H analyses over a
γ-DEX capillary column. c The catalyst-to-substrate molar ratio
d
F igu r e 3. Kinetics of the hydrogenation of 1-methoxyindan
in ethanol over Rh/C with NaOH additive, under standard
reaction conditions: concentration of 1-methoxyindan (9) and
its cyclohexene intermediate (b) and selectivity (2) and
incremental selectivity (1) to the cis-cis diastereomer.
is 60. Reaction conducted at 50 °C.
Results of hydrogenation of 3-propyl-3H-isobenzofuran-
1-one (7, Scheme 3) in ethanol are shown in Table 3. In
all experiments, a very high chemoselectivity (∼99%)
toward the fully hydrogenated cis products (8, Scheme
3) and a very high diastereoselectivity toward the cis-
cis diastereomer was obtained. Rh/C and Rh/Al2O3 were
equally selective. A cyclohexene intermediate was formed
during the hydrogenation, and its maximum concentra-
tion reached exceptionally high values of 20-40% of the
initial substrate concentration, depending on the reaction
conditions. The diastereoselectivity to the cis-cis isomer
increased very slightly as the reaction proceeded to
completion, implying that the intermediate was hydro-
genated with a greater selectivity to the cis-cis isomer
than the parent substrate. Changing the reaction tem-
perature from ambient temperature to 50 °C and the
pressure from 50 to 10 bar and halving the substrate-
to-rhodium ratio had a small effect on the dr with the
Rh/C catalyst.
The influence of the addition of triethylamine (TEA)
during hydrogenation was investigated for all substrates.
In line with previous observations,1 the addition of amine
decreased the activity in all reactions and led to almost
complete suppression of the hydrogenolysis of the 1-al-
koxyindan substrates. With all the substrates but 1-meth-
oxyindan and 3-propyl-3H-isobenzofuran-1-one, the TEA-
to-Rh and substrate-to-catalyst ratios were halved (by
doubling the catalyst amount) because of very low activ-
ity. Results in Table 4 show that the diastereoselectivity
increased in the case of 1-propoxyindan from 81:19 to 93:7
and in the case of indan-1-carboxamide from 71:28 to 82:
18, whereas in all other substrates it remained almost
unaffected. The dr value changed with the conversion for
all substrates; typically the initial dr value was higher
than the corresponding values in the reactions without
additive, but the dr value decreased as the reaction
proceeded to completion.
As in the case of TEA, addition of an aqueous solution
of NaOH (0.3 mL of strength 0.5 N) resulted in a decrease
in the activity accompanied by almost complete suppres-
sion of hydrogenolysis of the 1-alkoxyindan substrates.
Because of the low activity, the substrate-to-catalyst and
Na-to-Rh ratios were halved by doubling the amount of
the catalyst. The dr was significantly reduced in all cases
(Table 4). In the case of 1-methoxyindan, a substantial
reduction is obtained in dr and the value of dr reached
unity in the experiment with higher Na-to-Rh ratio. For
all substrates the value of dr decreased as the reaction
proceeded, probably because of the hydrogenation of the
cyclohexene intermediate. The decrease in dr during
reaction was largest for 1-methoxyindan. Figure 3 shows
the kinetics of hydrogenation of 1-methoxyindan in
ethanol with NaOH additive over the Rh/C catalyst.
Analogous to Figure 2, the selectivity and incremental
selectivity to the cis-cis diastereomer are plotted as a
function of reaction time along with the concentrations
of 1-methoxyindan and its intermediate, and the initial
and final dr values are also indicated. As in the case of
reactions without additives, toward the end of the reac-
tion, the fully hydrogenated cis products are formed
predominantly by the hydrogenation of the intermediate
and not of the substrate. It is evident from Figures 2 and
3 that the addition of NaOH leads to a lower activity and
a substantial reduction in the initial as well as the final
dr value. The drop in incremental dr is substantial,
especially during the hydrogenation of the intermediate,

1136 J . Org. Chem., Vol. 65, No. 4, 2000
Ranade et al.
indicating that the cyclohexene intermediate is much less
hydrogenated in a distofacial fashion than the parent
substrate.
selectivity because the directing information is carried
individually by each molecule.
Cyclohexene intermediates were observed in the hy-
drogenation of all substrates over the rhodium catalysts.
These intermediates underwent hydrogenation when the
concentration of the parent aromatic substrate was
sufficiently low. The position of the double bond in the
cyclohexene intermediate was identified as between the
junction carbon atoms only for the ester. It is likely,
however, that the double bond is located between the two
junction carbon atoms for other substrates too, this
position being favored from a steric point of view since
the junction carbon atoms of the aromatic substrate stay
farther away from the surface during adsorption than
the other carbon atoms in the six-membered ring. This
position of the double bond is also likely from a thermo-
dynamic point of view (Zaitsev’s rule). In addition,
chromatographic analyses during the reaction showed
only one peak for the intermediate on achiral as well as
chiral columns. The intermediates were hydrogenated
with a lower selectivity to the cis-cis diastereomer than
the parent aromatic substrate in all substrates (except
for 3-propyl-3H-isobenzofuran-1-one), where the cyclo-
hexene intermediate concentration was followed as a
function of time, as can be seen in the kinetics of the
hydrogenation of 1-methoxyindan in Figure 2. The dia-
stereoselectivity values reported for different substrates
are thus contaminated by the results of hydrogenation
of the corresponding cyclohexene intermediates. Separat-
ing the two effects would be possible if an independent
study of the hydrogenation of the cyclohexene intermedi-
ates is conducted. This is of little practical interest,
however, considering that the formation of the interme-
diates seems unavoidable during aromatic hydrogena-
tion. Nevertheless, the hydrogenation of the cyclohexene
intermediate of the ester substrate 1d was studied to
clarify the role of the cyclohexene intermediates in the
hydrogenation of the aromatic substrates. This study
enabled us to calculate the dr values of 89:11 and 71:29
in the direct hydrogenation of the aromatic ring and the
hydrogenation of the cyclohexene intermediate in the
case of the ester. Also, from the first and last points in
the kinetic analysis in Figure 2, the dr values of 88:12
and 59:41 were obtained in the direct hydrogenation of
the aromatic ring and the hydrogenation of the cyclo-
hexene intermediate, respectively, in the case of 1-meth-
oxyindan. Unfortunately, the dr values for the cyclohex-
ene intermediates in the hydrogenation of other aromatic
substrates could not be similarly calculated because the
intermediates underwent hydrogenation together with
the substrate when the concentration of the substrate
was sufficiently low. The lower dr obtained in the
hydrogenation of the cyclohexene intermediate could be
due to the spatial arrangement of the adsorbed cyclo-
hexene intermediate, resulting in a weaker expression
of the steric influence of the functional group. However,
it might also be due to the weaker adsorption of the
olefinic bond as compared to an aromatic ring, resulting
in a stronger expression of the electronic influence of the
group.
Discu ssion
The substrates are hydrogenated distofacially when the
steric repulsion of the substituent dominates, yielding
predominantly the cis-cis diastereomer, whereas they
are hydrogenated proximofacially when the electronic
attraction dominates, yielding predominantly the cis-
trans diastereomer (Scheme 1). The ratio of the cis-cis
to cis-trans diastereomers of substrates with a different
substituent gives an estimate of the relative repulsive
or attractive interaction between the substituent and
rhodium surface. Results in part 1 indicated that only
the amino group in 1-aminoindan interacted very strongly
with the catalyst surface, giving almost exclusively the
cis-trans diastereomeric product. The yield of the cis-
trans diastereomer is thus a measure of the attraction
of the functional group, in the aromatic substrates
hydrogenated, to the catalyst, also termed the haptophi-
licity of the functional group. The hydroxyl group had a
very small haptophilic effect, if at all, because the cis-
cis diastereomer was obtained as the major product. The
methyl group by virtue of its small size produced the cis-
cis diastereomer in moderate excess. We expected the
alkoxy group to have little or no interaction with the
metal surface because of its low polarity. However, the
carboxyl group can in principle interact with the metal
surface on account of its high polarity. This interaction
has often been successfully exploited in homogeneous
catalysis to obtain high selectivities to the proximofacial
products in olefin hydrogenations.10,11
The results for the hydrogenation of 1-methoxyindan
(Table 1) indicate that the cis-cis diastereomer is
obtained with a high selectivity. As observed in the case
of indanol,1 substantial hydrogenolysis takes place over
the Rh/C catalyst with spillover hydrogen.16 The extent
of hydrogenolysis is smaller with hexane as solvent for
both catalysts. 1-Propoxyindan shows the same trends
as 1-methoxyindan. The Rh/Al2O3 catalyst gives a slightly
higher selectivity to the cis-cis isomer with 1-propoxy-
indan than with 1-methoxyindan, in line with the larger
steric hindrance of the propyl group than the methyl
group.
Results reported for the substrates with carboxylic acid
and ester substituents (Table 2) indicate that these
groups show little attraction to the catalyst surface since
the cis-cis diastereomer is produced with a relatively
high selectivity. Indan-1-carboxamide is also hydroge-
nated with a moderately high selectivity to the cis-cis
diastereomer. The low electronic interaction exhibited by
the relatively polar carboxyl groups is surprising. For all
substrates, higher selectivity is obtained over the Rh/
Al2O3 catalyst than over Rh/C, in line with the results
for the other substrates. Hydrogenation of 3-propyl-3H-
isobenzofuran-1-one proceeds with a very high selectivity
as expected due to the substantial steric hindrance
offered by the propyl group (Table 3). The selectivity is
comparable to the one obtained in the hydrogenation of
1-propoxyindan. As expected in the hydrogenation of all
substrates, a change in the process conditions or the
metal-to-substrate ratio has a very small influence on the
Addition of triethylamine increases the yield of the cis-
cis isomer in some substrates, whereas in other sub-
strates it has no effect on the selectivity (Table 4). The
aromatic molecules undergoing hydrogenation occupy
more space on the metal surface when they are adsorbed
in the proximofacial fashion than in the distofacial
(16) Ranade, V. S.; Prins, R. Chem. Eur. J . 2000, 6, 313-320.

Hydrogenation of Aromatic Compounds
J . Org. Chem., Vol. 65, No. 4, 2000 1137
Ta ble 5. Com p a r ison of Ha p top h ilicities (Exp r essed a s
th e Selectivity to th e P r oxim ofa cia l Dia ster eom er in th e
Hyd r ogen a tion P r od u ct) of Va r iou s F u n ction a l Gr ou p s
in Su bstr a tes 1, 2, a n d 3
a comparable influence in the case of heterogeneous
catalysts. In fact, in complete contrast to the results with
homogeneous catalysts, the distofacial product is ob-
tained predominantly in the hydrogenation of the aro-
matic compounds directed by carboxyl substituents. This
can be explained in terms of the higher Lewis acidity and
the lower steric hindrance around the metal atom in the
case of homogeneous catalysts as compared to their
heterogeneous counterparts. The coordination sphere of
a metal atom embedded in a surface in heterogeneous
catalysts is less flexible than that of a metal center in
homogeneous catalysts. Therefore, unlike in the case of
homogeneous catalysts, an increase in electronic attrac-
tion due to a substituent is accompanied by an increase
in steric hindrance in the case of heterogeneous catalysts.
G )
NH2
1
2a
3b
98c
CH2NH2
CH2OH
OH
63
19
55c
41c
37c
28d
23d
21d
20d
19d
95
Me
CONH2
COOH
OMe
COOMe
OPr
10
18
0
0
15
0
After Thompson and Naipawer,8 Pd/C catalyst, solvent 2-meth-
a
oxyethanol. After Thompson and Wong,9 Pd/C catalyst, solvent
b
d
ethanol. c Part 1, Rh/C catalyst, solvent ethanol. Present study,
Rh/C catalyst, solvent ethanol.
Con clu sion s
The alkoxy groups in benzylic ethers do not interact
significantly with the surface of the catalyst. The interac-
tion is low irrespective of the catalyst support or the
solvent. The same result holds true for all carboxyl (acid,
ester, and amide) substrates, contrary to the intuitive
expectation that these polar groups would interact with
the rhodium surface. The diastereoselectivity is slightly
affected by changing either the substrate-to-catalyst ratio
or the reaction conditions.
The benzylic ethers undergo severe hydrogenolysis over
Rh/C catalysts as observed for 1-indanol, especially in
ethanol. Hydrogenolysis is retarded on addition of organic
or inorganic bases. The addition of triethylamine has a
relatively small influence on the diastereoselectivity
obtained in substrates with either the alkoxy or the
carboxyl substituent. Addition of NaOH reduces the
diastereoselectivity to the cis-cis isomer drastically for
the alkoxy substrates due to interaction of the etheric
oxygen with Na cations on the rhodium surface. The
diastereoselectivity to the cis-cis isomer also reduces
significantly in the case of carboxyl substrates presum-
ably due to the interaction of the carbonyl oxygen with
the Na cations.
fashion. Addition of amine increases the crowding even
more, thus favoring the formation of the cis-cis isomer.
This is not observed in the case of all substrates, probably
because of an opposing selectivity obtained in the hydro-
genation of the corresponding cyclohexene intermediates.
On addition of NaOH the dr reduces drastically for the
1-alkoxyindan substrates and approaches unity especially
for 1-methoxyindan (Table 4), suggesting the interaction
of the etheric oxygen with the Na cations adsorbed on
the rhodium surface. The reduction in dr is accentuated
by the fact that the cyclohexene intermediates produce
the cis-trans diastereomer with a much higher selectiv-
ity than the parent benzylic ether as shown in Figure 3.
The dr values obtained in the direct hydrogenation of the
aromatic ring and the hydrogenation of the cyclohexene
intermediate were 67:33 and 10:90, respectively. The
selectivity is also reduced in the carboxyl substrates,
probably because of similar interactions of the carbonyl
oxygen in carboxyl substrates and their intermediates
with the Na cations. Thus, results of the addition of bases
suggest that the amine influences the facial selectivity
by a steric mechanism while NaOH acts primarily by an
electronic mechanism.
The main implication of the results presented in the
two parts of this series is that the facial selection in
diastereoselective hydrogenation of aromatics on hetero-
geneous catalysts is primarily determined by the sterical
requirements of the molecule. The electronic interaction
plays an important role only when the molecule bears
an amino group or when the hydrogenation is conducted
in the presence of inorganic basic additives.
The results presented in part 1 and in the present
study encompass directing effects of several nonreducible
substituents encountered in aromatic hydrogenation.
Table 5 summarizes the haptophilicity of substituents
expressed in terms of the selectivity to the proximofacial
product. For comparison we have included the selectivi-
ties to proximofacial products reported by Thompson and
co-workers for substrates 2 and 3 obtained over a Pd/C
catalyst. It is seen that despite differences in substrates,
reaction conditions, and catalysts, the trends in the
haptophilicity are the same. The amino group, the
hydroxymethylene group, and the carboxyl groups exhibit
a high, moderate, and low haptophilicity, respectively.
In addition, in our investigations, the haptophilicities of
the hydroxyl and methyl groups lie between those of the
hydroxymethylene and carboxyl groups, whereas the
haptophilicities of the methoxy and propoxy groups are
comparable to those of the carboxyl groups.
Exp er im en ta l Section
Hyd r ogen a tion Exp er im en ts. Hydrogenation reactions
were conducted in a 60 mL stainless steel autoclave equipped
with a gas-inducing impeller at a stirring speed of 1100 rpm
under efficient mass transport conditions. In a typical experi-
ment, a solution of 0.5 g of substrate in 15 mL of ethanol or
hexane (and basic additive, if any) was added to 50 mg of
catalyst in the autoclave. Rh/Al2O3 (Fluka) and Rh/C (Aldrich)
catalysts (metal loading 5 wt %) were used as supplied. The
autoclave was closed, flushed three times successively with
nitrogen and hydrogen, and then pressurized to 50 bar with
hydrogen. All reactions were conducted at room temperature.
Samples could be taken with a sample tube during the reaction
to detect the completion of the reaction (typically less than 48
h in experiments without base addition). Analyses of samples
for determination of conversion and selectivity were done using
a gas chromatograph equipped with a FID detector and various
It is noteworthy that the directing effects of the
hydroxyl and carboxyl substituents in many substrates
are quite strong in the case of hydrogenation with
homogeneous catalysts. From the results presented in
part 1 and here, as well as in the work of Thompson and
co-workers, it is concluded that these effects are much
too weak (especially for the carboxyl substituents) to have

1138 J . Org. Chem., Vol. 65, No. 4, 2000
Ranade et al.
capillary columns depending on the substrate hydrogenated.
Some hydrogenation experiments were continued overnight,
even after complete conversion of the substrate and the
intermediate. The diastereoselectivity remained unchanged
over this additional period.
d r o-1-p r op oxyin d a n (5b). A reference mixture of the cis-
cis and cis-trans diastereomers was prepared starting from
perhydro-1-indanol using the same method as used in the
preparation of 1-methoxyindan.19 Comparison of the 1H and
13C NMR spectra of the reference mixture with the product of
hydrogenation was used to identify the relative configuration.
For 1-propoxyindan it was assumed that the major product
had the cis-cis configuration, in analogy to the results of
1-methoxyindan.
P r ep a r a tion of th e Su bstr a tes. Racemic 1-methoxyindan
(1a ) was prepared from racemic 1-indanol. To a solution of
1-indanol in dry tetrahydrofuran was added NaH, and the
mixture was heated to 60 °C and maintained under argon for
30 min at the same temperature. Methyl iodide was added to
the mixture, and it was maintained at 60 °C for 4 h. Excess
NaH was hydrolyzed with water, and the mixture was
extracted with ether twice. The ether extracts were pooled
together and washed with water. The ether was removed in
vacuo, and the yellow liquid product was purified by Kugelrohr
distillation. The resulting colorless distillate of 1-methoxyindan
was stored under argon, since it was air-sensitive and was
directly used in hydrogenation experiments. Racemic 1-pro-
poxyindan (1b) was prepared in exactly the same way with
propyl iodide instead of methyl iodide. The yellow liquid
product contained the product as well as the unconverted
adduct in a ratio of about 1:3. Separation was affected by
column chromatography (hexane:ethyl acetate ) 19:1) to yield
pure (>99%) 1-propoxyindan as a colorless liquid. The identity
of 1-methoxyindan was established by comparing its NMR
spectrum to that reported in the literature.17 The identity of
1-propoxyindan was established from its NMR and MS analy-
ses.
Racemic indan-1-carboxylic acid (1c) was prepared by
hydrogenolysis of racemic 3-oxoindan-1-carboxylic acid over
10 wt % Pd/C catalyst (Fluka) in ethanol under 3 bar of
hydrogen pressure (Scheme 2). The resulting product was
isolated by removal of the solvent in vacuo after the catalyst
was filtered off. A quantitative yield of indan-1-carboxylic acid
(white solid) was obtained, and the acid was stored under
argon because it was air-sensitive and used directly without
any purification in further reactions/hydrogenations. The
racemic methyl ester of indan-1-carboxylic acid (1d ) was
prepared by refluxing a solution of the acid in methanol after
addition of excess of thionyl chloride for 2 h (Scheme 2). The
solvent was removed in vacuo, and the yellow liquid product
was distilled under vacuum in a Kugelrohr distillation ap-
paratus. The resulting colorless liquid was used in the
hydrogenation experiments.
Racemic indan-1-carboxamide (1e) was prepared by two
methods (Scheme 2). In one method, a suspension of the
methyl ester of indan-1-carboxylic acid was stirred vigorously
in an aqueous ammonia solution, giving a slightly green-white
precipitate, which was isolated by filtration and washed
successively with water. In the second method, indan-1-
carboxylic acid was converted to the acid chloride by being
heated with thionyl chloride for 20 min at about 40 °C. The
acid chloride was isolated by removing the excess thionyl
chloride in vacuo. It was dissolved in tetrahydrofuran, and
ammonia gas was bubbled through to get a slightly green-
white precipitate. The solvent was removed in vacuo and the
resulting solid washed with water. The crude indan-1-car-
boxamide was recrystallized from water.
Commercially available propylidene phthalide (6, Lancaster,
cis:trans ) 6.7, total purity 96.6%) was hydrogenated on the
Pd/C catalyst at ambient temperature and 10 bar of hydrogen
pressure in ethanol (Scheme 3). The catalyst hydrogenated the
olefinic bond exclusively, and a quantitative yield of pure
racemic 3-propyl-3H-isobenzofuran-1-one (7, >99.6% pure,
identified by NMR18) was obtained as a colorless liquid after
removal of the solvent in vacuo. This was used further in
hydrogenation experiments.
P er h yd r oin d a n -1-ca r boxylic Acid (5c) a n d P er h y-
d r oin d a n -1-ca r boxylic Acid Meth yl Ester (5d ). The prod-
uct of hydrogenation of indan-1-carboxylic acid consisted
primarily of two diastereomers as observed by NMR analy-
1
sis. The H and 13C NMR spectral data of the product mixture
did not fit the NMR spectral data reported for the trans
isomers by Galteri et al.20 For the identification of the rela-
tive configuration of the cis diastereomers, the hydrogenated
carboxylic acid was converted to its methyl ester by treatment
with diazomethane in ether (prepared using the method of
Black21). The mixture of hydrogenated indan-1-carboxylic acid
methyl esters was then injected into a gas chromatograph
equipped with an HP-1 capillary column. Comparison of the
order of elution of the two diastereomers to that reported by
Granger et al.22 was used to identify the absolute configuration.
The cis-trans diastereomer elutes before the cis-cis diaste-
reomer.
P er h yd r oin d a n -1-ca r boxa m id e (5e). For identification
of the cis diastereomeric products obtained in the hydrogena-
tion of indan-1-carboxamide, a reference mixture was prepared
from perhydroindan-1-carboxylic acid in a two-step procedure
via the acid chloride as reported for indan-1-carboxamide. The
reference products were isolated, and comparison of their 1H
and 13C NMR spectra and chromatogram in an RTX-200
column with the products of hydrogenation enabled the
assignment of the cis-cis and cis-trans configurations to the
major and the minor products, respectively.
3-P r op yl-3H-h exa h yd r oisoben zofu r a n -1-on e (8). The
absolute configuration was identified by a two-dimensional
NOESY analysis of the isolated product mixture (see the
Supporting Information). In the two-dimensional 1H NMR
spectrum of the product mixture, a NOE was observed between
the protons at the 3a and 7a positions on the cyclohexanediyl
ring for both the major and the minor diastereomers, indicat-
ing that the cis diastereomers were formed exclusively. A NOE
was observed between the proton at position 3 and the protons
at positions 3a and 7a on the cyclohexanediyl ring, however
only for the major diastereomer. Thus, the cis-cis and cis-
trans configurations were assigned to the major and minor
products, respectively.
Ack n ow led gm en t. We are grateful to F. Bangerter
for the NMR measurements and A. Dutly for the MS
measurements.
Su p p or tin g In for m a tion Ava ila ble: NMR and MS data
of 1-propoxyindan (1b), 3-propyl-3H-isobenzofuran-1-one (7),
2,3,4,5,6,7-hexahydro-1H-indene-1-carboxylic acid methyl ester
(4d ), and the cis-cis and cis-trans diastereomers of all
perhydro products (5a -e and 8) and two-dimensional 1H NMR
NOE spectrum of a mixture of the cis-cis and cis-trans
diastereomers of 3-propyl-3H-hexahydroisobenzofuran-1-one
(8). This material is available free of charge via the Internet
at http://pubs.acs.org.
J O991604S
(19) Morlender-Vais, N.; Mandelbaum, A. J . Mass. Spectrosc. 1997,
31, 1124-1132.
Id en tifica tion of cis-cis a n d cis-tr a n s P r od u ct Dia -
ster eom er s. P er h yd r o-1-m eth oxyin d a n (5a ) a n d P er h y-
(20) Galteri, M.; Lewis, P. H.; Middleton, S.; Stock, L. E. Aust. J .
Chem. 1980, 33, 101-111.
(17) Pincock, J . A.; Wedge, P. J . J . Org. Chem. 1995, 60, 4067-4076.
(18) Canonne, P.; Plamondon, J .; Akssira, M. Tetrahedron 1988, 44,
2903-2912.
(21) Black, H. T. Aldrichimica Acta 1983, 16, 3-10.
(22) Granger, R.; Boussinesq, J .; Girard, J .-P.; Rossi, J . C. Bull. Soc.
Chim. Fr. 1968, 4, 1445-1450.