Practical resolution of 1-phenyl-2-(4-methylphenyl)ethylamine using a single resolving agent controlled by the dielectric constant of the solvent

Article abstract of DOI:10.1016/j.tetasy.2004.08.035

A practical resolution of 1-phenyl-2-(4-methylphenyl)ethylamine 1 for obtaining (R)- and (S)-enantiomers by a dielectrically controlled resolution (DCR) method using a single resolving agent, (S)-mandelic acid 2, has been studied. The configuration of the excess enantiomer 1 in the less-soluble diastereomeric salt was found to vary depending on the solvent controlled by adjusting the solvent dielectric constant ε and the water content of the alcohol solvent. It was found that the presence of water molecules and the solvent molecular structures was key to controlling the configurational change phenomenon in the chiral discrimination between molecules 1 and 2. Moreover, the reaction environment controlled by the solvent dielectric constant plays a very distinct role in utilizing water molecules for crystallizing the less-soluble diastereomeric salt during the resolution. Based on the results, a practical resolution process for (R)-1 and (S)-1 production has been developed involving a simple solvent switch method; the (S)-1?(S)-2?H₂O salt (33% yield, 96% de, E 64%) was obtained by the crystallization from 74% (w/w) aqueous ethanol in the first resolution, while the (R)-1?(S)-2 salt (23% yield, 98% de, E 46%) was crystallized from 100% ethanol of the condensate of the mother liquor after the first resolution.

Full text of DOI:10.1016/j.tetasy.2004.08.035

Tetrahedron:
Asymmetry
Tetrahedron: Asymmetry 15 (2004) 3495–3500
Practical resolution of 1-phenyl-2-(4-methylphenyl)-
ethylamine using a single resolving agent controlled
by the dielectric constant of the solvent
Kenichi Sakai,a,* Rumiko Sakurai,a Hiroyuki Nohira,b
Rumiko Tanakac and Noriaki Hirayamac,*
aR&D Division, Yamakawa Chemical Industry Co., Ltd, Kitaibaraki, Ibaraki 319-1541, Japan
bInnovative Research Organization for New Century, Saitama University, Sakuraku, Saitama 338-8570, Japan
cBasic Medical Science and Molecular Medicine, Tokai University School of Medicine, Isehara, Kanagawa 259-1193, Japan
Received 22 July 2004; accepted 30 August 2004
Available online 28 October 2004
Abstract—A practical resolution of 1-phenyl-2-(4-methylphenyl)ethylamine 1 for obtaining (R)- and (S)-enantiomers by a dielectri-
cally controlled resolution (DCR) method using a single resolving agent, (S)-mandelic acid 2, has been studied. The configuration of
the excess enantiomer 1 in the less-soluble diastereomeric salt was found to vary depending on the solvent controlled by adjusting the
solvent dielectric constant e and the water content of the alcohol solvent. It was found that the presence of water molecules and the
solvent molecular structures was key to controlling the configurational change phenomenon in the chiral discrimination between
molecules 1 and 2. Moreover, the reaction environment controlled by the solvent dielectric constant plays a very distinct role in
utilizing water molecules for crystallizing the less-soluble diastereomeric salt during the resolution. Based on the results, a practical
resolution process for (R)-1 and (S)-1 production has been developed involving a simple solvent switch method; the (S)-1Æ(S)-2ÆH2O
salt (33% yield, 96% de, E 64%) was obtained by the crystallization from 74% (w/w) aqueous ethanol in the first resolution, while the
(R)-1Æ(S)-2 salt (23% yield, 98% de, E 46%) was crystallized from 100% ethanol of the condensate of the mother liquor after the first
resolution.
ꢀ 2004 Elsevier Ltd. All rights reserved.
1. Introduction
tain the opposite stereoisomer of the target substrate. If
both enantiomers can be resolved as the less-soluble
salts by a single enantiomer of the resolving agent, the
resolution process would become much more cost-
effective.
It has been estimated that more than half of chiral drugs
in the pharmaceutical market are produced by the dia-
stereomeric salt formation method using enantiopure
resolving agents because of the ease of operation and
wide applicability.1 Currently, chiral intermediates for
important chiral drugs such as indinavir (anti-viral),2
sertraline (anti-depressant),3 orlistat (anti-obsessional),4
and duloxetine (anti-depressant)5 are widely known to
be efficiently produced by this method on an industrial
scale. It is generally understood that one of the chiral
species of the resolving agent only acts on one stereo-
isomer of the target substrate to crystallize the less-solu-
ble diastereomeric salt.6 Therefore, it is common to
apply the other enantiomer of the resolving agent to ob-
Recently we have found a quite unusual but useful phe-
nomenon; a single resolving agent can recognize both
stereoisomers of the target substrate and give the corre-
sponding less-soluble diastereomeric salts by simply
adjusting the dielectric constants (e) of the solvent.
The first example that we discovered was the resolution
system of (RS)-a-amino-e-caprolactam 3 as a target sub-
strate to be resolved and N-tosyl-(S)-phenylalanine 4 as
the resolving agent.7 Based on this finding, we success-
fully developed a practical continuous resolution for
(R)-3 and (S)-3 production by diastereomeric salt for-
mation using one naturally-based resolving agent (S)-
48,9 (Fig. 1). We designated this procedure as a simple
solvent switch method.
*
Corresponding authors. Fax: +81 29330 11 81; e-mail addresses:
sakai@yamakawa-chemical.co.jp; hirayama@is.cc.u-tokai.ac.jp
0957-4166/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetasy.2004.08.035

3496
K. Sakai et al. / Tetrahedron: Asymmetry 15 (2004) 3495–3500
COOH
NHTs
(S)-4
O
O
O
NH2
(S)-4 H2O
NH2
NH2
(S)-4
HN
HN
(S)-4
(S)-4
HN
MeOH
2-PrOH-H2O
(S)-3
(RS)-1
(R)-3
Figure 1. Resolution of a-amino-e-caprolactam 3 with N-tosyl-(S)-phenylalanine 4.
The crystal structures of the less-soluble salts, (R)-3Æ(S)-
4 and (S)-3Æ(S)-4ÆH2O, showed that the molecular recog-
nition between the resolving agent and the target
substrate is governed by a specific hydrogen-bonding
network due to the difference of the solvent dielectric
constant. If this phenomenon is general, namely if the
hydrogen-bonding system can be controlled by adjusting
the solventꢀs e value, this dielectrically controlled resolu-
tion (DCR) method could be quite promising to a vari-
ety of resolution systems.
head-stirring without seeding in order to avoid the
occurrence of preferential crystallization with specific
seed crystals. Components, diastereomeric excesses,
and absolute configurations of the salts obtained were
determined by elemental, chiral HPLC, and X-ray crys-
tal structure analyses. The results are summarized in
Table 1. The (R)-1Æ(S)-2 salt was crystallized with relatively
higher diastereomeric excess (de)13 and resolution effi-
ciency (E)14 from 1-butanol, 1-propanol and ethanol.
The (RS)-1Æ(S)-2 salt was crystallized from 2-butanol,
2-propanol, and methanol. To our surprise, the (S)-
1Æ(S)-2ÆH2O salt crystallized only from water, although
the amount of water required is impractical for indus-
trial-scale production. The results indicate that the
branched alcohols are not favorable for the chiral dis-
crimination between (S)-2 and 1. The bulky groups in
the branched alcohols may disturb the molecular recog-
nition. The exceptional result for methanol may be due
to its specific physicochemical properties. (S)-1 was crys-
tallized only from water and reveals that water mole-
cules play a distinctive role in the chiral recognition
between (S)-2 and (S)-1 as observed in our previous
studies.8,10 The results shown in Table 1 suggest that
the configuration and enantiomeric excess of 1 could
be controlled by using a mixed solvent of non-branched
alcohols and water.
Nohira et al. reported11 that the chirality of the resolu-
tion product in the system of (RS)-1-phenyl-2-(4-meth-
ylphenyl)ethylamine 1 and (S)-mandelic acid 2 changes
and it depends on the solvent used (Fig. 2). It is interest-
ing to examine whether this system is dielectrically con-
trolled or not and therefore we investigated the effect of
the solvent on this resolution. Herein we report that this
resolution system can be controlled by the dielectric con-
stant of the solvent. We also report that the molecular
structure of the solvent can have significant influence
on the resolution.
CH3
Table 1. Resolution of (RS)-1 with (S)-2 in alcohols and water
OH
Entry Solvent
e
Solvent/
(RS)-1a w/w
De, R/S Yieldc Ed
b
%
COOH
NH2
1
2
2-BuOH 16
1-BuOH 17
47
18
1
1
92
RS
R
55
25
4
46
1
32-PrOH 18
49
1-PrOH 22
RS
R
54
(RS)-1
(S)-2
4
5
6
7
21
12
4
23
96
0
2311
16
EtOH
MeOH
Water
24
33
R
31
0
Figure 2.
RS
S
27
24
78 228
98
47
(S)-2/(RS)-1 = 1.0 molar ratio.
a Solvent weight was determined by solubility of the solid substances at
2. Results and discussion
50ꢁC.
The effects of six alcohols (C1–C4; e = 16–33)12 and
water (e = 78)12 on the resolution have been examined
as the resolution solvent while an equimolar amount
of resolving agent (S)-2 to the racemate (RS)-1 was used.
Isobutyl alcohol and tert-butyl alcohol could not be ap-
plied as resolution solvents because of the low water sol-
ubility (16%) and the high melting point (25.6ꢁC),
respectively. Resolutions were performed under over-
b Diastereomeric excess (% de) = [A ꢀ B] · 100/(A + B), where A and B
are both diastereomers.
c Calculated based on (RS)-1.
d Resolution efficiency (E, %) = yield (%) · diastereomeric excess (%
de) · 2/100.
The effects of various mixed solvents of alcohol and
water (e = 16–78)15 on the resolution were examined as

K. Sakai et al. / Tetrahedron: Asymmetry 15 (2004) 3495–3500
Table 2. Resolution of (RS)-1 with (S)-2 in various alcoholic solvents
3497
Entry
Solvent
Alcohol concentration, %
e
Solvent/(RS)-1a w/w
De, %b
R/S
Yieldc
Ed
1
2-BuOH
100
92
16
22
47
19
1
96
RS
S
55
3
2
2
7
4
7
6
71
66
71
69
385
4
25
16
98
S
3
75
65
3
3
2
8
11
9
S
3
3
96
97
5
S
6
7
1-BuOH
2-PrOH
100
99
17
18
18
16
92
19
R
R
25
42
46
16
8
100
95
82
18
21
29
36
42
49
20
1
42
95
97
RS
S
54
45
3
1
38
9
10
11
12
1345
14
12
12
S
8
9
7
72
76
71
70
60
S
3
1397
S
3
51
16
65
100
S
48
51
62
48
15
69
48
S
15
16
1-PrOH
6
100
92
22
26
21
7
23
48
R
S
S
S
S
S
2311
5351
3
17
18
19
20
2
833
96
4
7
6
67
71
70
75
50
25
36
50
64
6
9
97
98
97
3
3
40
28
76
21
22
EtOH
100
95
24
27
12
10
3
96
28
R
S
S
S
S
S
S
S
S
16
3
30
8
4
21
67
2390
24
29
57
47
41
3
33
56
85
74
60
50
30
16
32
38
46
51
62
69
8
68
99
66
25
26
10
11
44
59
65
3
58
57
47
27
28
1348
39
36
98
29
120
4
67
65
30
31
32
33
34
MeOH
Water
100
86
70
44
20
33
39
47
58
69
4
3
0
18
39
68
95
RS
S
27
56
56
44
3
0
20
43
60
6
S
34
100
S
S
4
35
0
78
228
98
S
24
47
(S)-2/(RS)-1 = 1.0 molar ratio.
a Solvent weight was determined by solubility of the solid substances at 50ꢁC.
b Diastereomeric excess (% de) = [A ꢀ B] · 100/(A + B), where A and B are both diastereomers.
c Calculated based on (RS)-1.
d Resolution efficiency (E, %) = yield (%) · diastereomeric excess (% de) · 2/100.
100
shown in Table 2. The mixture, which gives a homoge-
neous clear solution at ambient temperature, was taken
into account. Concentrations of 1-butanol and 2-buta-
nol were limited with the maximum concentrations
Water
80
60
40
2-BuOH
MeOH
20
being 99% and 65%, respectively. The trends of the con-
figuration changes are illustrated in Figure 3. All alcohol
0
2-PrOH
- 20
1-PrOH
solvents, except methanol, behaved similarly in a range
of e < ca. 40. Their enantiomeric excesses of 1 rapidly
shifted to the (S)-enriched state from the (R) or (RS)-
form state with increase in the solvent dielectric con-
stant. It is noteworthy that the diastereomeric excess
of the salt obtained from the aqueous ethanol solution
dropped to 36% de at e = 62 and recovered at e = 69
(98% de). This phenomenon demonstrates that the sol-
vent dielectric constant seems to be one of the main fac-
tors, which govern the chiral discrimination but some
other factors may also play influential roles in the dis-
crimination process.
- 40
- 60
1-BuOH
EtOH
- 80
- 100
10
20
30
40
50
60
70
80
Dielectric constant ε
Figure 3. Relation between the diastereomeric excess of the less-
soluble diastereomeric salt and the solvent dielectric constant (e).
From a practical point of view, it is extremely useful that
the configuration of the salt was drastically changed

3498
K. Sakai et al. / Tetrahedron: Asymmetry 15 (2004) 3495–3500
from the enantiopure (R)-1 to the enantiopure (S)-1 by
adjusting the dielectric constant of the medium by
changing the water content. A remarkable resolution
efficiency can be achieved when ethanol is used as the
solvent. (R)-1 (96% de) was crystallized from 100% eth-
anol (e = 24), whereas (S)-1 (99% de) from 74% ethanol
(e = 38). On the other hand, in the case of branched
alcohols such as 2-butanol and 2-propanol, the diaste-
reomeric excesses of the salts crystallized were improved
to give the enantiopure (S)-form from the racemic form
at 22 < e < 38 and 29 < e < 42, respectively. It is note-
worthy that the behavior of methanol was completely
different. The diastereomeric excess shifted from the
(RS)-form to the (S)-form linearly according to the in-
crease of the solvent dielectric constant, with the maxi-
mum (95% de) being obtained at a concentration of
20%.
the presence of water but also the solventꢀs molecular
structure plays a very distinctive role in the chiral mole-
cular recognition between molecules 1 and 2. Based on
these findings, we could develop a practical resolution
process for obtaining (R)-1 and (S)-1 using only one
enantiomer of the resolving agent, together with a sim-
ple solvent switch method. In addition to the practical
significance, the solvent dielectric constant e was found
to be a good measure for comparing the resolution
results obtained from different sorts of solvents in the
resolution process development.
4. Experimental
4.1. General
(RS)-1 (CP 98.7%) and (S)-mandelic acid 2 (>99.5%
ee)16 were made of Yamakawa Chemical (Tokyo). Other
reagents were purchased from Junsei Chemical, unless
otherwise indicated. 1H NMR spectra were recorded
on a JEOL JNM-ECP400 spectrometer in DMSO-d6
with Me4Si as an internal reference. IR spectra were
measured on a JASCO IR-700 spectrometer using KBr
pellets. Optical rotations were measured on a JASCO
P-1020 polarimeter with a circular temperature control
unit. High-performance liquid chromatography was per-
formed by a JASCO Intelligent HPLC system equipped
with a UV-970 detector. Melting points were determined
In order to develop the optimum resolution conditions
of (RS)-1 with (S)-2 suitable for an industrial-scale
production, 100% ethanol (e = 24) and 74% ethanol
(e = 38) was selected as resolution solvents. The resolu-
tion was performed in two stages. Since the resolution
efficiency of (R)-1 from ethanol was relatively low, the
first resolution was designed to produce (S)-1 from
74% ethanol following our strategy of the continuous
resolution method described in our previous study.7
The second resolution was designed to produce (R)-1
by using ethanol from mother liquor of the first resolu-
tion containing the (R)-enriched 1. Resolution experi-
ments were performed under over-head-stirring with
seeding in order to effectively enhance the resolution
efficiency. The (S)-1Æ(S)-2ÆH2O was obtained in identical
efficiency (33% yield, 96% de, E 64%) to the single
resolution result shown in Table 2. The result indicates
that seeding did not enhance the resolution efficiency.
The mother liquor of the first resolution was evaporated
to dryness to recover the mixture of the (R)-enriched 1
and the equimolar (S)-2 (49% de), followed by switching
the solvent to 100% ethanol for the second resolution.
As intended, the resolution efficiency for the (R)-1Æ(S)-
2 salt increased dramatically (23% yield, 98% de, E
46%).
with
uncorrected.
a
YAMATO MP-21 instrument and are
4.2. Determination of diastereomeric purity
The diastereomeric purity of the salts, (R)-1Æ(S)-2 and
(S)-1Æ(S)-2ÆH2O, were directly determined by HPLC
using
a
Shiseido Chiral CD-Ph column (ID
4.6mm · 250mm). Analytical conditions for the HPLC
were as follows: 0.5M NaClO4/CH3CN (50/50 v/v),
0.5mL/min, 35ꢁC, detected at 254nm; injection sample
10lL (10mg/10mL); retention times: (S)-1 13.6min,
(R)-1 20.7min, (S)-2 5.5min.
4.3. Strategic continuous resolution: resolution starting
from (S)-1 separation
3. Conclusion
In order to obtain (R)-1 more efficiently, the resolution
was started from (S)-1 resolution. A typical experimen-
tal procedure is as follows: to a 50mL flask were added
(RS)-1 (3.00g, 14.2mmol), (S)-2 (2.16g, 14.2mmol) and
74% ethanol (28.7g), and the mixture was stirred and
heated up to about 50ꢁC to give a clear solution. The
solution was then gradually cooled, seeded (2mg) at
26ꢁC, kept for 1h at around the crystallization temper-
ature, and then cooled again to 20ꢁC. After aging the
suspension at the temperature for 1h, the crystals were
filtered off and washed twice with 74% ethanol to
afford the crude (S)-1Æ(S)-2ÆH2O salt (1.74g, 4.6mmol,
33% yield, 96% de, E 64%). Analytical data for the
In the resolution process development of (RS)-1-phenyl-
2-(4-methylphenyl)ethylamine 1 with (S)-mandelic acid
2 using various alcohol solvents, it was found that the
configuration of the excess enantiomer 1 in the less-sol-
uble diastereomeric salt varies depending on the solvent
used; the configuration and enantiomeric excess of 1 in
the salt were controlled by adjusting the solvent dielec-
tric constant e while controlling water content in the
alcohol solvent. Moreover, it was also found that the
solventꢀs molecular structure affects the chiral recogni-
tion ability of the resolving agent (S)-2 during the reso-
lution; the (S)-1Æ(S)-2ÆH2O salt was obtained by the
crystallization from water or water-contained alcohols,
while the (R)-1Æ(S)-2 salt was crystallized from non-
branched alcohols such as ethanol, 1-propanol, and 1-
butanol. These results clearly indicate that not only
recrystallized salt are as follows: (S)-1Æ(S)-2 saltÆH2O:
20
D
99.6% de; mp 149.5–52.0ꢁC; ½aŠ ¼ þ126:5 (c
1.04, EtOH); IR (KBr) cmꢀ1: 3634, 3324, 2944, 2734,
2590, 2178, 1640, 1574, 1545, 1514, 1450, 1401, 1375,

K. Sakai et al. / Tetrahedron: Asymmetry 15 (2004) 3495–3500
3499
˚
1348, 1328, 1194, 1084, 1060, 933, 811, 778, 740, 702,
618; H NMR (400MHz, DMSO-d6): d 7.16–7.40 (m,
a = 6.135(1), b = 11.901(2), c = 27.907(5)A, V =
2037.5(6)A , Z = 4, Dcalc = 1.243g/cm3, R = 0.042;
1
3
˚
10H), 7.00 (d, J = 8.0Hz, 2H), 6.91 (d, J = 8.0Hz,
2H), 4.65 (s, 1H), 4.28 (dd, J = 8.8, 5.6Hz, 1H), 3.11
(dd, J = 13.2, 5.6Hz, 1H), 2.95 (dd, J = 13.2, 8.8Hz,
1H), 2.21 (s, 3H). Anal. Calcd for C23H25NO3ÆH2O: C,
72.42; H, 7.13; N, 3.67. Found C, 73.32; H, 7.08; N,
3.70.
Rw = 0.042. Number of reflections measured = total
2052, unique 1550.
References
1. Rouhi, A. M. Chemical and Engineering News 2003 (May
5), 46.
The mother liquor obtained after the first resolution de-
scribed above was evaporated to dryness to give a con-
densate [3.17g; (R)-enantiomer rich 49% de; water
contents 0.14% measured by KF method]. To the con-
densate was added ethanol (21.3g), the slurry was trans-
ferred to a 50mL flask. The mixture was stirred and
heated up to about 60ꢁC to give a clear solution. The
solution was then gradually cooled, seeded (2mg) at
25ꢁC, kept for 1h at around the crystallization temper-
ature and then cooled again to 20ꢁC. After aging the
suspension at the temperature for 1h, the crystals were
filtered off and washed twice with ethanol to afford the
crude (R)-1Æ(S)-2 salt (1.20g, 3.3mmol, 23% yield, 98%
de, E 46%). Analytical data for the recrystallized salt
are as follows.
2. (a) Indinavir: Merck Index, 13ed., No. 4970, 2001; (b)
Murakami, H.; Tobiyama, T.; Sakai, K. JP Kokai 1997-
48762; (c) Murakami, H.; Satoh, S.; Tobiyama, T.; Sakai,
K.; Nohira, H. EP 710652, 1996; (d) Murakami, H.;
Satoh, S.; Tobiyama, T.; Sakai, K.; Nohira, H. USP
5792869, 1996; Chem. Abstr. 1996, 125, 87209, resolution
of (S)-3-tert-butylcarbamoylpiperazine-1-carboxylic acid
tert-butyl ester with N-tosyl-(S)-phenylalanine.
3. (a) Sertraline: Merck Index, 13ed., No. 8541, 2001; (b)
Quallich, G. J. USP 6593496, 2000; Chem. Abstr. 2000,
134, 41980; (c) Mendelovici, M.; Nidam, T.; Pilarsky, G.;
Gershon, N. USP 6552227, 2001; Chem. Abstr. 2001, 135,
242019; (d) Taber, P. G.; Pfisterer, D. M.; Colberg, J. C.
Org. Process Res. Dev. 2004, 8, 385–388, Resolution of
[4-(3,4-dichlorophenyl)-1,2,3,4-tetrahydronaphthalen-1-yl]
methylamine with (R)-mandelic acid.
4. (a) Orlistat: Merck Index, 13ed., No. 6953, 2001; (b)
Karpf, M.; Zutter, U. EP 443449, 1991; Chem. Abstr.
1991, 115, 255980. Resolution of benzoic acid 1-(1-
carboxyheptyl)-3-phenoxymethyltetradecyl ester with (S)-
1-phenylethylamine.
5. (a) Duloxetine: Merck Index, 13ed., No. 3498, 2001; (b)
Berglund, R. A. EP 650965, U.S. 5,362,886, JP 1995-
188065, Chem. Abstr. 1994, 122, 132965. Resolution of 3-
(dimethylamino)-1-(2-thienyl)propan-1-ol with (S)-man-
delic acid; (c) Sakai, K.; Sakurai, R.; Yuzawa, A.;
Kobayashi, Y.; Saigo, K. Tetrahedron: Asymmetry 2003,
14, 1631–1636, Resolution of 3-(methylamino)-1-(2-thi-
enyl)propan-1-ol with (S)-mandelic acid.
20
D
(R)-1Æ(S)-2 salt: >99.9% de; mp 176.5–177.0ꢁC; ½aŠ
¼
ꢀ45:7 (c 1.03, EtOH); IR (KBr) cmꢀ1: 3454, 2890,
2660, 2120, 1894, 1611, 1566, 1529, 1417, 1366, 1332,
1196, 1080, 1060, 937, 811, 787, 756, 703; 1H NMR
(400MHz, DMSO-d6): d 7.15–7.40 (m, 10H), 7.00 (d,
J = 8.0Hz, 2H), 6.92 (d, J = 8.0Hz, 2H), 4.65 (s, 1H),
4.28 (dd, J = 8.8, 6.0Hz, 1H), 3.07 (dd, J = 13.2,
6.0Hz, 1H), 2.94 (dd, J = 13.2, 8.8Hz, 1H), 2.21 (s,
3H). Anal. Calcd for C23H25NO3: C, 76.01; H, 6.93;
N, 3.85. Found C, 76.09; H, 6.98; N, 3.85.
4.4. X-ray crystal structure analysis
6. (a) Jacques, J.; Collet, A.; Wilen, S. H. Enantiomers,
Racemates, and Resolutions; Wiley: New York, 1981;
Crystallographic data for the structures of (R)-1Æ(S)-2
and (S)-1Æ(S)-2ÆH2O have been deposited with the Cam-
bridge Crystallographic Data Centre as reference num-
bers CCDC 241593and 241594, respectively. X-ray
diffraction data were measured using the Rigaku
RAXIS-RAPID system.
p
251; (b) Kozma, D. CRC Handbook of Optical
Resolution via Diastereomeric Salt Formation; CRC Press
LLC, 2002.
7. (a) Sakai, K. Symposium Molecular Chirality, Shizuoka,
Japan, IL-8, Oct 19, 2003; (b) Sakurai, R.; Sakai, K.;
Yuzawa, A.; Hirayama, N. Symposium Molecular Chiral-
ity, Shizuoka, Japan, PA-11, Oct 19, 2003.
8. Sakai, K.; Sakurai, R.; Yuzawa, A.; Hirayama, N.
Tetrahedron: Asymmetry 2003, 14, 3713–3718.
9. Sakai, K.; Sakurai, R.; Yuzawa, A.; Hatahira, K. Jpn.
Patent Appl. 2003-338118, 2003.
10. Sakai, K.; Sakurai, R.; Hirayama, N. Tetrahedron:
Asymmetry 2004, 15, 1073–1076.
11. Nohira, H.; Murata, H.; Asakura, I.; Terunuma, D. Japan
Patent Kokai 59-110656, 1984; Chem. Abstr. 1985, 102,
5896.
12. Asahara, S.; Tokura, N.; Okawara, M.; Kumanotani, J.;
Senoo, M. Solvent Handbook; Kodansha Scientific, 2001,
and references cited therein.
(R)-1Æ(S)-2: a colorless platelet single crystal of (S)-1Æ
(S)-2 salt (0.50 · 0.30 · 0.25mm) was grown from EtOH
using recrystallized salt crystals (>99.9% de). The X-ray
intensities were measured up to 2hmax = 68.17ꢁ with
graphite monochromated CuKa radiation (k =
˚
1.5419A) at 293K. C23H25NO3; formula weight
363.46; orthorhombic, P212121 (# 19), a = 6.607(4),
3
˚
˚
b = 15.504(1), c = 20.243(2)A, V = 2073(1)A , Z = 4,
Dcalc = 1.164g/cm3, R = 0.062; Rw = 0.053. Number of
reflections measured = total 2085, unique 1387.
13. Diastereomeric excess (% de) = [A ꢀ B] · 100/(A + B),
(S)-1Æ(S)-2ÆH2O: a colorless platelet single crystal of
(S)-1Æ(S)-2ÆH2O salt (0.30 · 0.25 · 0.15mm) was grown
from water using recrystallized salt crystals (>99.9%
de). The X-ray intensities were measured up
to 2hmax = 68.19ꢁ with graphite monochromated CuKa
where A and B are both diastereomers.
14. Resolution efficiency (E, %) = yield (%) · diastereomeric
excess (% de) · 2/100, where yield is calculated based on
(RS)-1.
15. (a) The values of dielectric constants of mixed solvents
were calculated as the weighted average of the mixture
components, alcohol and water, based on the following
˚
radiation (k = 1.5419A) at 293K. C23H27NO4; formula
weight 381.47; orthorhombic, P212121 (# 19),

3500
K. Sakai et al. / Tetrahedron: Asymmetry 15 (2004) 3495–3500
equation and literatures; e(mix) = (wt%(alcohol) · e(alcohol)) +
(wt%(water) · e(water)), where e(alcohol) and e(water) are the
dielectric constants at 20ꢁC of pure alcohol and water,
respectively. (a) Jouyban, A.; Soltanpour, S.; Chan, H.-K.
Int. J. Pharm. 2004, 269, 353–360; (b) Prakongpan, S.;
Nagai, T. Chem. Pharm. Bull. 1984, 32, 340–343.
16. Murakami, H.; Sakai, K. J. Synth. Org. Chem. Jpn. 1990,
48, 850.