Rational design of CH/π interaction sites in a basic resolving agent

Article abstract of DOI:10.1021/jo049154d

A novel synthetic basic resolving agent, cis-1-aminobenz[f]indan-2-ol (ABI), was rationally designed by introducing effective CH/π interaction sites to cis-1-aminoindan-2-ol (AI), whose chiral recognition ability has been reported from our laboratory. ABI was applicable to a wide variety of racemic arylalkanoic acids and showed moderate to excellent chiral recognition ability, which was obviously higher than that of AI. The fundamental and important role of CH/π interactions, such as tunable CH(sp²)/π and CH(sp ³)/π interactions, in the chiral recognition by ABI was revealed by X-ray crystallographic study.

Full text of DOI:10.1021/jo049154d

Ra tion a l Design of CH/π In ter a ction Sites in a Ba sic Resolvin g
Agen t
Yuka Kobayashi,† Toshie Kurasawa,‡ Kazushi Kinbara,† and Kazuhiko Saigo*,†,‡
Department of Chemistry and Biotechnology, Graduate School of Engineering, and Department of
Integrated Biosciences, Graduate School of Frontier Sciences, The University of Tokyo, Hongo 7-3-1,
Bunkyo-ku, Tokyo, 113-8656, J apan
saigo@chiral.t.u-tokyo.ac.jp
Received May 19, 2004
A novel synthetic basic resolving agent, cis-1-aminobenz[f]indan-2-ol (ABI), was rationally designed
by introducing effective CH/π interaction sites to cis-1-aminoindan-2-ol (AI), whose chiral recognition
ability has been reported from our laboratory. ABI was applicable to a wide variety of racemic
arylalkanoic acids and showed moderate to excellent chiral recognition ability, which was obviously
higher than that of AI. The fundamental and important role of CH/π interactions, such as tunable
CH(sp2)/π and CH(sp3)/π interactions, in the chiral recognition by ABI was revealed by X-ray
crystallographic study.
In tr od u ction
tion mechanism of enantiopure mandelic acid (MA) in
diastereomeric salt formation, could recognize the chiral-
ity of a variety of racemic arylalkylamines and that its
chiral recognition ability, higher than that of MA, was
supposed to arise from effective CH/π interactions in the
corresponding less-soluble diastereomeric salts.2 Al-
though CH/π interaction is generally considered to be
very weak, compared with other nonbonding interactions
such as hydrogen-bonding interaction and coordination
interaction,3 CH/π interaction has been attracting much
attention in recent years, owing to its remarkable role
for the determination of molecular arrangement in
crystals.4 These facts strongly suggest that the introduc-
tion of CH/π interaction site(s) to a known resolving agent
is very attractive for the development of an efficient
Crystals can offer a suitable environment for chiral
recognition phenomena, since the movement of molecules
in crystals is highly prevented to achieve three-dimen-
sionally dissymmetric fixation. From this point of view,
the sole crystallization of one of a pair of diastereomeric
salts (diastereomeric salt formation) is very fascinating
for the recognition of the chirality of racemates. In the
diastereomeric salt formation, an enantiopure acidic or
basic reagent (resolving agent) is essential, and natural
products, their derivatives, and synthetic mandelic acid
and 1-phenylethylamine are currently used as resolving
agents. However, the number of basic resolving agents,
which are widely applied in diastereomeric salt forma-
tion, is limited, compared with that of acidic resolving
agents; basic resolving agents frequently used are 1-
phenylethylamine, amino acid esters, and alkaloids.
Moreover, these resolving agents are not necessarily
applicable for the chiral recognition of given target
racemates. Therefore, the development of a new synthetic
basic resolving agent, which can be applicable to a wide
variety of racemates, is highly desired.
We have recently reported that enantiopure cis-1-
aminoindan-2-ol (AI) showed recognition ability with a
variety of racemic arylalkanoic acids.1 However, the
chiral recognition ability was insufficient; a structural
modification of AI was considered to be one attractive
way to improve the chiral recognition ability of AI. On
the other hand, we have also demonstrated that enan-
tiopure 2-naphthylglycolic acid (NGA), designed on the
basis of the structural characteristics and chiral recogni-
(2) (a) Kinbara, K.; Harada, Y.; Saigo, K. J . Chem. Soc., Perkin
Trans. 2 2000, 1339. (b) Kinbara, K.; Harada, Y.; Saigo, K. Tetrahe-
dron: Asymmetry 1998, 9, 2219.
(3) Nishio, M.; Hirota, M.; Umezawa, Y. The CH/π interaction.
Evidence, Nature and Consequences; Wiley-VCH: New York, 1998.
(4) (a) Matsumoto, A.; Kunisue, T.; Nagahama, S.; Matsumoto, A.;
Sada, K.; Inoue, K.; Tanaka, T.; Miyata, M. Mol. Cryst. Liq. Cryst. 2003,
390, 11. (b) Thuery, P.; J eong, T. G.; Yamato, T. Supramol. Chem. 2003,
15 (5), 359. (c) Suezawa, H.; Yoshida, T.; Umezawa, Y.; Tsuboyama,
S.; Nishio, M. Eur. J . Inorg. Chem. 2002, 12, 3148. (d) Abe, H.;
Miyamura, K. Inorg. Chim. Acta 2000, 298, 90. (e) Umezawa, Y.;
Tsuboyama, S.; Takahashi, H.; Uzawa, J .; Nishio, M. Tetrahedron
1999, 55, 10047.
† Department of Chemistry and Biotechnology, Graduate School of
Engineering.
‡ Department of Integrated Biosciences, Graduate School of Frontier
Sciences.
(1) Kinbara, K.; Kobayashi, Y.; Saigo, K. J . Chem. Soc., Perkin
Trans. 2 2000, 111.
10.1021/jo049154d CCC: $27.50 © 2004 American Chemical Society
Published on Web 09/30/2004
7436
J . Org. Chem. 2004, 69, 7436-7441

Rational Design of CH/ π Interaction Sites
SCHEME 1. Syn th esis of En a n tiop u r e ABI
TABLE 1. Ch ir a l Recogn ition Ability of ABI for
Ar yla lk a n oic Acid s
resolving agent. Thus, we rationally designed a novel
basic resolving agent, cis-1-aminobenz[f]indan-2-ol (ABI),
upon replacing the phenyl group of AI by a naphthyl
group with an expectation that the naphthyl group would
highly contribute to the realization of effective CH/π
interactions in less-soluble diastereomeric salt crystals.
In this paper, we report the synthesis and chiral recogni-
tion ability of enantiopure ABI and the chiral recognition
mechanism for racemic arylalkanoic acids.
Resu lts a n d Discu ssion
P r ep a r a tion of (1R,2S)-1-Am in oben z[f]in d a n -2-ol
(ABI). Racemic cis-1-aminobenz[f]indan-2-ol (rac-ABI)
was prepared by a procedure similar to that for the
synthesis of racemic cis-1-aminoindan-2-ol.5 The starting
material, benz[f]indene, could be obtained in 64% total
yield via five steps from commercially available 2-formyl-
benzoic acid.6 The epoxidation of benz[f]indene, followed
by the Ritter-type reaction, yielded rac-ABI in 44% total
yield as shown in Scheme 1. The enantioseparation of
rac-ABI was achieved in methanol by use of (R)-2-
naphthylglycolic acid2 as a resolving agent. Recrystalli-
zation of the crude less-soluble diastereomeric salt from
methanol three times gave the diastereopure salt. Treat-
ment of the salt with aqueous sodium hydroxide afforded
the crude product, which was purified by recrystallization
from chloroform to give (1R,2S)-1-aminobenz[f]indan-2-
ol (ABI) in 39% overall yield.
Ch ir a l Recogn ition Ability of ABI for Ar yla l-
k a n oic Acid s. At first, we tried the enantioseparation
of various arylalkanoic acids by using ABI as a resolving
agent in order to investigate the chiral recognition ability
of ABI. The enantioseparations were performed under
almost the same conditions: the diastereomeric salt was
allowed to crystallize from ethanol/water at a constant
temperature (30 °C), and the ratio and amount of the
mixed solvent were adjusted so as to control the yield of
the precipitated salt to be as close as possible to a range
of 50-80%. No recrystallization was performed. The
results are summarized in Table 1.
a
Product yield of the diastereomeric salt and the ee of the
b
liberated acid. Values in parentheses are the resolution efficien-
cies when AI was used in place of ABI. c Not determined Not
d
crystallized.
As can be seen from Table 1, ABI showed good to
excellent chiral recognition ability for 2-phenylalkanoic
acids. The chiral recognition ability of ABI was generally
higher than that of (1S,2R)-1-aminoindan-2-ol (AI), as
we expected.7 Especially, the resolution efficiencies for
2-phenylalkanoic acids having a substituent on the
phenyl group were remarkably improved by using ABI
(5) (a) Larrow, F.; Roberts, E.; Verhoeven, R. T.; Ryan, K. M.;
Senanayake, C. H.; Reider, P. J .; J acobsen, E. N. Org. Synth. 1999,
76, 46. (b) Senanayake, C. H.; Michele, L. M.; Liu, J .; Fredenburgh, L.
E.; Ryan, K. M.; Roberts, F. E.; Larsen, R. D.; Verhoeven, T. R.; Reider,
P. J . Tetrahedron Lett. 1995, 36, 7615. (c) Hijfte, L. V.; Little, D. R.;
Petersen, J . L.; Moeller, K. D. J . Org. Chem. 1987, 52, 4647.
(6) Sujit, K.; Ghorai, K. S.; Harza, K. N.; Mal, D. Tetrahedron Lett.
1998, 39, 689.
J . Org. Chem, Vol. 69, No. 22, 2004 7437

Kobayashi et al.
columns in the ABI and AI salt crystals are very
similar: the N‚‚‚O distances are 2.69-2.83 and 2.69-
2.87 Å, and the O‚‚‚O distances are 2.66 and 2.62 Å in
the ABI and AI salt crystals, respectively, strongly
indicating that the hydrogen-bonding networks would
contribute in a similar manner for the stabilization of
the salt crystals.
In the ABI and AI salt crystals there are other
interactions, which stabilize the crystals. The packing
motifs of the 21 columns viewed down from the column
axis are displayed in Figure 2. In the crystals, there exist
CH/π interactions, which can be classified into two
types: (1) intracolumnar CH/π interaction in the 21
column and (2) intercolumnar CH/π interaction between
the 21 columns. According to the definition, CH/π interac-
tion should be discussed on the basis of the distance
between a π-plane and a H atom facing the π-plane.3 On
the other hand, an essential factor contributing to
stabilization energy is known to be determinable on the
basis of the distance between a π-plane and a C atom
facing the π-plane.10 In the present X-ray analyses, the
C‚‚‚π-plane distances could be determined, although the
H atoms could not be precisely positioned. Therefore, we
used the C‚‚‚π-plane distances to explain the role of the
CH/π interactions.11
In tr a colu m n a r CH/π In ter a ction in th e ABI a n d
AI Sa lt Cr yst a ls. The helical columnar hydrogen-
bonding network (21 column) makes the arrangement of
the amine and carboxylic acid molecules helical along the
21 column to realize CH(sp3)/π interaction between the
methyl group attached at the chiral center of the car-
boxylic acid 1c and the π-plane of ABI or AI (Figure 2),
indicating that this CH(sp3)/π interaction plays a very
important role in the chiral recognition. The C‚‚‚π-plane
distance for the CH(sp3)/π interaction is 3.96 Å in the
ABI salt crystal, which is the same as that in the AI salt
crystal (Figure 3). These facts indicate that the π-planes
of ABI and AI directly recognize the chirality of 1c and
that the intracolumnar CH(sp3)/π interactions in the ABI
and AI salt crystals similarly contribute to the stabiliza-
tion of the respective salt crystals.
In ter colu m n a r CH/π In ter a ction in th e ABI a n d
AI Sa lt Cr ysta ls. In the ABI and AI salt crystals, there
are two kinds of CH/π interactions between the 21
columns other than van der Waals interaction as attrac-
tive nonbonding interactions; CH(sp3)/π interaction be-
tween the methyl group at the 3-position of the phenyl
group of 1c and the phenyl group of 1c located in the
neighboring column, and T-shaped CH(sp2)/π interaction
between the aromatic rings of two molecules of ABI or
AI in the neighboring columns (Figure 2). The CH(sp3)/π
interaction would be exceptional for 1c, since such CH-
(sp3)/π interaction cannot be found in the other less-
soluble ABI and AI salt crystals we examined. This
CH(sp3)/π interaction is fundamentally similar in the ABI
and AI salt crystals, although the C‚‚‚π-plane distances
and the orientations of the two interacting 1c molecules
are a little different (Figure 4).
F IGURE 1. Hydrogen-bonding networks constructed in (a)
ABI‚1c and (b) AI‚1c salt crystals.
in the place of AI (entries 2 and 4-6). Moreover, ABI
could recognize the chirality of some arylalkanoic acids,
for which AI showed no chiral recognition ability (entries
11-13).
Str u ctu r a l Ch a r a cter istics of th e Less-Solu ble
ABI Sa lt Cr ysta ls. ABI showed obviously higher chiral
recognition ability than did AI for some of arylalkanoic
acids we examined. This result strongly suggests that
ABI would be able to stabilize the less-soluble salts more
efficiently than does AI by additional CH/π interaction-
(s) between ABI molecules and/or between the ABI
molecule and the acid molecule, which would arise from
the enlarged aromatic part of ABI (the naphthyl group
in place of the phenyl group of AI), as was expected. To
clarify the role of the enlarged aromatic part for the
stabilization of the less-soluble salts, we next carried out
X-ray crystallographic analyses of the less-soluble ABI
salts and compared their crystal structures with those
of the corresponding AI salts.
As a typical example for the comparison, we selected
a pair of the less-soluble salts of 2-(3-methylphenyl)-
propionic acid (1c) with ABI and AI.8 In both less-soluble
ABI‚1c and AI‚1c salt9 crystals, a one-dimensional,
helical columnar hydrogen-bonding network (21 column)
is commonly constructed from the ammonium cations,
the carboxylate anions, and the hydroxy groups, as shown
in Figure 1. Moreover, the molecular arrangements in
the 21 columns, which should be determined not only by
the pattern of the 21 columns but also by the molecular
shapes of the amine and carboxylic acid components, are
almost identical so that the packing modes of the 21
columns in the crystals are also very similar to each
other. As a result, the hydrogen-bond distances in the 21
(7) We used (1R,2S)-1-aminobenz[f]indan-2-ol (ABI) and compared
its chiral recognition ability with that of (1S,2R)-1-aminoindan-2-ol
(AI); their absolute configurations are inverse to each other. Conse-
quently, the less-soluble diastereomeric salts with ABI and AI consist
of the amino alcohol and the antipodes of the arylalkanoic acids,
respectively.
(8) Cambridge Crystallographic Data Centre, ref CCDC-140008,
contains the supplementary crystallographic data for (1S,2R)-1-amino-
indan-2-ol‚(R)-2-(3-methylphenyl)propionic acid (AI‚1c) salt.1 These
data can be obtained free of charge via www.ccdc.cam.ac.uk/conts/
retrieving.html.
(10) Tsuzuki, S.; Honda, K.; Uchimaru, T.; Mikami, M.; Tanabe, K.
J . Am. Chem. Soc. 2000, 122, 15, 3746.
(11) We defined CH/π interactions as those with
a C‚‚‚π-plane
distance less than 4.1 Å, according to the intermolecular interaction
potentials of benzene-methane complexes estimated with ab initio
molecular orbital calculations.10
(9) The less-soluble salts are ABI‚(S)-1c and AI‚(R)-1c.
7438 J . Org. Chem., Vol. 69, No. 22, 2004

Rational Design of CH/ π Interaction Sites
F IGURE 2. Crystal structures viewed down from the b axis for (a) ABI‚1c and (b) AI‚1c salt crystals.
ference in the magnitude of the intercolumnar T-shaped
CH(sp2)/π interactions between the ABI and AI salt
crystals would cause the difference in chiral recognition
ability between ABI and AI.
Wid e Ap p lica bility of ABI to th e Ch ir a l Recogn i-
tion of (4-Su bstitu ted Ar yl)a lk a n oic Acid s. Thus, in
ABI‚1c and AI‚1c salt crystals, one-dimensional helical
hydrogen-bonding columns (21 columns) aggregate by
CH/π interactions and van der Waals interaction to give
the corresponding three-dimensional crystals. The most
outstanding characteristic of the ABI salt crystal, which
cannot be observed in the AI salt crystal, is the T-shaped
double CH(sp2)/π interaction, probably leading to higher
crystallinity of ABI salts and better chiral recognition
ability of ABI. This structural feature of the ABI salt
crystal strongly suggests that the T-shaped CH(sp2)/π
interaction would have flexibility to a considerable extent;
in the other less-soluble ABI salt crystals, the T-contact-
ed naphthyl groups of ABI molecules would be able to
slide, depending on the molecular arrangement of ABI
and the carboxylic acid component in the 21 column, with
maintaining T-shaped double or at least single CH(sp2)/π
interaction. Such a slide would be unexpectable in AI salt
crystals. To verify this consideration, we investigated the
crystal structure of the less-soluble salt of ABI with 2-(4-
isobutylphenyl)propionic acid (ibuprofen, 1f), which has
a long substituent at the 4-position of the phenyl group.
F IGURE 3. Intracolumnar CH(sp3)/π interactions in (a) ABI‚
1c and (b) AI‚1c salt crystals.
F IGURE 4. Intercolumnar CH(sp3)/π interactions in (a) ABI‚
1c and (b) AI‚1c salt crystals.
The unit cell lengths of ABI‚1f salt crystal are signifi-
cantly altered, compared with those of ABI‚1c salt
crystal; the c axis is elongated due to the long isobutyl
group in 1f, while the a axis is shortened due to the slide
of the T-contacted naphthyl groups (Figure 6a). As a
result, a new pattern of CH/π interaction appears be-
tween the naphthyl groups; there exists CH(sp3)/π inter-
action between the methylene of the five-membered ring
of ABI and the π-plane of ABI in the neighboring 21
column as well as T-shaped single CH(sp2)/π interaction
between them (Figure 6b). The crystal structures of ABI‚
1c and ABI‚1f clearly demonstrate that the naphthyl
group of ABI plays an important role for its wide
applicability in chiral recognition.
F IGURE 5. Intercolumnar CH(sp2)/π interactions in (a) ABI‚
1c and (b) AI‚1c salt crystals.
In contrast, the T-shaped CH(sp2)/π interactions in the
ABI and AI salt crystals are quite different from each
other; in the ABI salt crystal there are two short contacts,
while only one short contact exists in the AI salt crystal
(Figure 5). This means that the T-shaped double CH-
(sp2)/π interaction in the ABI salt crystal is obviously
stronger than the T-shaped single CH(sp2)/π interaction
in the AI salt crystal and that the ABI salt crystal is
stabilized more effectively than the AI salt crystal by the
intercolumnar T-shaped CH(sp2)/π interaction. This dif-
J . Org. Chem, Vol. 69, No. 22, 2004 7439

Kobayashi et al.
F IGURE 7. (a) Crystal structure of ABI‚2, viewed down from
the b axis, and (b) intercolumnar CH(sp3)/π interaction in the
crystal.12
F IGURE 6. (a) Crystal structure of ABI‚1f, viewed down from
the b axis, and (b) intercolumnar CH(sp2)/π interaction in the
crystal.
relative geometry of ABI and 2 to contribute to the
recognition of the chirality of 2.
Con clu sion
Ap p lica bility of ABI to th e Ch ir a l Recogn ition of
3-P h en ylbu tyr ic Acid (2). The chiral recognition of
â-chiral carboxylic acids and amines is usually difficult
by diastereomeric salt formation, since their chiral
centers are located far from the chiral center of a
resolving agent across a carboxylic acid/amine hydrogen-
bonding site. Thus, AI showed no chiral recognition
ability at all to 3-phenylbutyric acid (2). In contrast, ABI
showed a moderate chiral recognition ability to 2. This
difference in chiral recognition ability between ABI and
AI could be explained on the basis of the X-ray crystal-
lographic analyses of ABI‚2 and AI‚2 salt crystals.
AI‚2 salt crystal consists of hydrogen-bonding columns
having inversion centers (i-columns); the column includes
both enantiomers of 2 alternatively. On the other hand,
ABI salt crystal still consists of 21 columns and subse-
quently includes single enantiomers of 2 (Figure 7a). As
shown in Figure 7b, both R-methylene and γ-methyl of 2
contact with the π-plane of ABI by CH(sp3)/π interac-
tions. It is noteworthy that the R-methylene group shortly
contacts with the π-plane, where the C‚‚‚π-plane distance
is 3.58 Å. These CH(sp3)/π interactions would fix the
We rationally designed a novel basic resolving agent,
(1R,2S)-1-aminobenz[f]indan-2-ol (ABI) on the basis of
the difference in chiral recognition ability between man-
delic acid (MA) and 2-naphthylglycolic acid (NGA). ABI
showed a moderate to excellent chiral recognition ability
for a wide variety of arylalkanoic acids, such as unsub-
stituted/substituted 2-phenyl- and 2-(2-naphthyl)alkanoic
acids and 3-phenylalkanoic acid. The chiral recognition
ability of ABI was generally higher than that of (1S,2R)-
1-aminoindan-2-ol (AI), which was selected as a model
for the design; only the replacement of the phenyl group
in the indane skeleton by a naphthyl group contributed
to the obvious improvement of the chiral recognition
ability. X-ray crystallographic study revealed that the
higher and wider chiral recognition ability of ABI than
that of AI arose from effective and tunable CH(sp2)/π and
CH(sp3)/π interactions, for which the naphthyl group of
ABI plays a very important role.
Exp er im en ta l Section
1,2-Ep oxyben z[f]in d en e. To a suspension of sodium hy-
drogencarbonate (13.6 g, 0.162 mol) and 3-chloroperbenzoic
acid (14.6 g, 0.087 mol) in dichloromethane (200 mL), cooled
with an ice bath, was added dropwise a solution of benz[f]-
indene6 in dichloromethane (400 mL) at 0 °C under Ar
(12) The salt crystal contains single enantiomers of 2 and ABI. This
means that the actually deposited crystals consist of ABI‚(S)-2 and
ABI‚(R)-2 as the major and minor diastereomeric salts, respectively.
7440 J . Org. Chem., Vol. 69, No. 22, 2004

Rational Design of CH/ π Interaction Sites
atmosphere. The suspension was stirred at room temperature
for 4 h, and saturated aqueous NaCO3 (400 mL) was added
dropwise to the reaction mixture at 0 °C. After being separated
from the organic layer, the aqueous layer was extracted with
CH2Cl2 (5 × 300 mL). The combined organic layer and extracts
were successively washed with saturated aqueous NaCO3 (3
× 300 mL) and saturated aqueous NaCl (300 mL), dried over
anhydrous Na2SO4, filtered, and concentrated under reduced
pressure to give crude 1,2-epoxybenz[f]indene as a pale yellow
solid mass (15.90 g, 87 mmol, quant.): mp 148-150 °C; IR
The enantiopurity of ABI thus obtained was determined by
an HPLC analysis (Chiralcel OD; eluent hexane/2-propanol )
9/1; flow rate 1.0 mL/min) of its derivative, N,O-diacetylated
ABI, which was obtained by the reaction with a large excess
amounts of acetic anhydride and pyridine, followed by the
usual workup: (1S,2R)-ABI 23 min, (1R,2S)-ABI 30 min;
enantiopurity >99%.
Typ ica l P r oced u r e for En a n tiosep a r a tion of Ar yla l-
k a n oic Acid s [Table 1, en tr y 1]. A solution of ABI (99.2 mg,
0.5 mmol) and racemic 2-phenylpropionic acid (1a ) (75.8 mg,
0.5 mmol) in EtOH/H2O (6/4 v/v; 5 mL) was refluxed for 1 h,
and the solution was cooled to and kept at 30 °C in a
thermostat for 12 h. The deposited crystals were collected by
filtration, washed with a small amount of cooled EtOH, and
dried in vacuo to give ABI‚1c salt (53.7 mg, 0.15 mmol, 61%)
as white needle crystals. The salt was treated with 1 M
aqueous HCl (50 mL), and the aqueous solution was extracted
with ether (3 × 50 mL). The combined extracts were dried over
anhydrous Na2SO4, filtered, and concentrated under reduced
pressure to give enantioseparated 1c, of which the enantio-
meric excess (ee) was determined by an HPLC analysis
(Chiralcel OJ -R; eluent aqueous HClO4 (pH 2)/CH3CN ) 7/3;
flow rate 0.5 mL/min): 91% ee.
(KBr) 3450, 3060, 1420, 830, 490, 470 cm-1 1H NMR (300
;
MHz, CDCl3) δ 3.14 (1H, d, J ) 18.0 Hz), 3.37 (1H, d, J )
18.0 Hz), 4.21 (1H, t, J ) 2.7 Hz), 4.41 (1H, d, J ) 2.7 Hz),
7.26-7.95 (6H, m).
1,2-Epoxybenz[f]indene thus obtained was used in the next
reaction without further purification.
Ra cem ic cis-1-Am in oben z[f]in d a n -2-ol (r a c-ABI). To
dry CH3CN (200 mL), vigorously stirred with a mechanical
stirrer, were successively added fuming sulfuric acid (20 mL)
at -10 °C in a period of 1 h and a solution of 1,2-epoxybenz-
[f]indene (9.76 g, 54 mmol) in dry dichloromethane (160 mL)
at -10 °C in a period of 1 h, and the reaction mixture was
stirred for 2 h at room temperature. Then, to the solution was
added dropwise water (1 L), and the mixture was stirred for 1
h at room temperature. After the organic solvents were
evaporated under reduced pressure, ethanol (1 L) was added
to the mixture, and the solution was stirred at 100 °C for 50
h. After removal of ethanol, the solution was cooled to room
temperature and extracted with CH2Cl2 (200 mL). The pH of
the aqueous solution, cooled with an ice bath, was adjusted at
about 13 with 12.5 M aqueous NaOH, and the alkaline solution
was extracted with CH2Cl2 (5 × 400 mL). The combined
extracts were dried over anhydrous Na2SO4, filtered, and
concentrated under reduced pressure to give crude rac-ABI
as a white solid mass. Recrystallization from ethyl acetate gave
pure rac-ABI (4.73 g, 24 mmol, 44%): decomp 169 °C; IR (KBr)
3350, 3270, 3170, 3050, 2970, 2940, 2890, 1580, 1420, 1340,
The enantioseparations of the other carboxylic acids were
performed under almost the same conditions: the diastereo-
meric salts were allowed to crystallize at 30 °C from ethanol/
H2O, of which the ratio and amount were adjusted so as to
control the yield of the salts to be as close as possible to a range
of 50-80%. The carboxylic acids, except for 2-phenylpropionic
acid, 2-(4-isobutylphenyl)propionic acid, and 2-phenylbutyric
acid, were converted into their methyl esters by reaction with
(trimethylsilyl)diazomethane prior to their HPLC analyses.
Str u ctu r e of (1R,2S)-ABI‚(S)-1c Sa lt. Crystals of the salt
belong to the monoclinic space group P21 with a ) 12.112(8)
Å, b ) 6.291(4) Å, c ) 13.038(9) Å, and â ) 98.248(9)°. V )
983.1(12) Å3, Dcalcd ) 1.228 g cm-3, Rf ) 0.0590, and Rw
0.0620.
)
1
1260, 1160 cm-1; H NMR (300 MHz, CDCl3) δ 1.78 (2H, s),
Str u ctu r e of (1R,2S)-ABI‚(S)-1f Sa lt. Crystals of the salt
belong to the monoclinic space group P21 with a ) 10.04(2) Å,
b ) 6.261(10) Å, c ) 17.58(3) Å, and â ) 95.48(2)°. V ) 1100.0-
(34) Å3, Dcalcd ) 1.224 g cm-3, Rf ) 0.0900, and Rw ) 0.0890.
Str u ctu r e of (1R,2S)-ABI‚(S)-2 Sa lt. Crystals of the salt
belong to the monoclinic space group P21 with a ) 12.81(2) Å,
b ) 6.099(7) Å, c ) 12.650(10) Å, and â ) 95.28(2)°. V ) 984.1-
(17) Å3, Dcalcd ) 1.226 g cm-3, Rf ) 0.0900, and Rw ) 0.1390.
Str u ctu r e of (1S,2R)-AI‚(R)-2 Sa lt. Crystals of the salt
belong to the triclinic space group P1 with a ) 11.206(3) Å, b
) 13.618(6) Å, c ) 5.869(3) Å, R ) 97.73(3)°, â ) 99.51(3)°,
and γ ) 75.43(3)°. V ) 850.9(6) Å3, Dcalcd ) 1.223 g cm-3, Rf )
0.0510, and Rw ) 0.0630.
3.09-3.26 (2H, m), 4.20-4.50 (2H, m), 7.28 (1H, s), 7.82 (6H,
m).
(1R,2S)-1-Am in o-ben z[f]in d a n -2-ol (ABI). A solution of
rac-ABI (2.06 g, 10.3 mmol) and (R)-2-naphthylglycolic acid2
(1.23 g, 6.1 mmol) in MeOH (420 mL) was gently refluxed for
1 h, slowly cooled to 30 °C, and kept at 30 °C for 12 h. The
deposited solid was collected by filtration and recrystallized
three times from MeOH to give diastereopure salt (0.81 g, 2
mmol, 39%) as a white powder. To the diastereopure salt thus
obtained was added 3 M aqueous NaOH (250 mL), and the
mixture was stirred for 1 h and extracted with CH2Cl2 (5 ×
150 mL). The combined extracts were dried over anhydrous
Na2SO4, filtered, and concentrated under reduced pressure to
give a solid mass, which was recrystallized from chloroform
to afford enantiopure ABI as a white solid (0.39 g, 2 mmol,
quant): decomp 169 °C; [R]D ) 89° (c 0.2, CHCl3); IR (KBr)
3350, 3270, 3150, 3070, 2970, 2930, 2870, 2800, 1580, 1420,
Su p p or tin g In for m a tion Ava ila ble: 1H NMR and IR
spectra of (1R,2S)-1-aminobenz[f]indan-2-ol (ABI) and CIF
files for the crystals of (1R,2S)-ABI‚(S)-1c, (1R,2S)-ABI‚(S)-
1f, (1R,2S)-ABI‚(S)-2, and (1S,2R)-AI‚(R)-2 salts (PDF, CIF).
This material is available free of charge via the Internet at
http://pubs.acs.org.
1
1340, 1260, 1160; H NMR (300 MHz, CDCl3) δ 1.78 (2H, s),
3.09-3.26 (2H, m), 4.20-4.50 (2H, m), 7.28 (1H, s), 7.82 (6H,
m). Anal. Calcd for C13H13NO‚1/4H2O: C, 76.63; H, 6.68; N,
6.87. Found: C, 76.46; H, 6.47; N, 6.89.
J O049154D
J . Org. Chem, Vol. 69, No. 22, 2004 7441