Enantiopure O-substituted phenylphosphonothioic acids: Chiral recognition ability during salt crystallization and chiral recognition mechanism

Article abstract of DOI:10.1021/jo052020v

The chiral recognition ability of enantiopure O-methyl, O-ethyl, O-propyl, and O-phenyl phenylphosphonothioic acids (1a-d) for various kinds of racemic amines during salt crystallization and the chiral recognition mechanism were thoroughly investigated. T

Full text of DOI:10.1021/jo052020v

Enantiopure O-Substituted Phenylphosphonothioic Acids: Chiral
Recognition Ability during Salt Crystallization and Chiral
Recognition Mechanism
Yuka Kobayashi, Fumi Morisawa, and Kazuhiko Saigo*
Department of Chemistry and Biotechnology, Graduate School of Engineering, The UniVersity of Tokyo,
Hongo, Bunkyo-ku, Tokyo 113-8656, Japan
saigo@chiral.t.u-tokyo.ac.jp
ReceiVed September 27, 2005
The chiral recognition ability of enantiopure O-methyl, O-ethyl, O-propyl, and O-phenyl phenylphospho-
nothioic acids (1a-d) for various kinds of racemic amines during salt crystallization and the chiral
recognition mechanism were thoroughly investigated. The chiral recognition abilities of enantiopure 1a-d
for a wide variety of racemic amines varied in a range of 6 to >99% enantiomeric selectivity. Deposited
less-soluble diastereomeric salts were classified into two categories, prism- and needle-type crystals; the
prism-type crystals were composed of a globular molecular cluster, while there existed a 21 column in
the needle-type crystals. In contrast to a general observation of a similar 21 column in the less-soluble
diastereomeric salt crystals of chiral primary amines with chiral carboxylic acids, the globular molecular
cluster is a very unique hydrogen-bonding motif that has never been constructed in diastereomeric salt
crystals. Excellent chiral recognition was always achieved when the less-soluble diastereomeric salts
were prism-type crystals. Significant correlations were found between the degree of the chiral recognition
with 1a-d, the crystal shape of the less-soluble diastereomeric salts, and the hydrogen-bonding motif
(molecular cluster/21 column). The chiral recognition mechanisms via the molecular cluster and the 21
column formations are discussed in detail on the basis of X-ray crystallographic analyses.
Introduction
chiral recognition of racemates by such a method have been
reported so far,1-3 the selection of a resolving agent suitable
for a target racemate is crucial even at present; there is no
The method based on the chiral recognition of a racemate
with a chirality-recognizable enantiopure reagent (resolving
agent) during salt crystallization, so-called the enantioseparation
of a racemate via the diastereomeric salt formation, is one of
the most promising methods for obtaining enantiopure com-
pounds.1 Although a great number of successful results for the
(2) For selected recent papers, see: (a) Sakai, K.; Sakurai, R.; Yuzawa,
A.; Kobayashi, Y.; Saigo, K. Tetrahedron: Asymmetry 2003, 14, 1631-
1636. (b) Nieuwenhuijzen, J. W.; Grimbergen, R. F. P.; Koopman, C.;
Kellogg, R. M.; Vries, T. R.; Pouwer, K.; van Echten, E.; Kaptein, B.;
Hulshof, L. A.; Broxterman, Q. B. Angew. Chem., Int. Ed. 2002, 41, 4281-
4286. (c) Yoshioka, R.; Hiramatsu, H.; Okamura, K.; Tsujioka, I.; Yamada,
S. J. Chem. Soc., Perkin Trans. 2 2000, 10, 2121-2128. (d) Juaristi, E.;
Escalante, J.; Leon-Romo, J. L.; Reyes, A. Tetrahedron: Asymmetry 1998,
9, 715-740. (e) Kozma, D.; Bocskei, Z.; Kassai, C.; Simon, K.; Fogassy,
E. Chem. Commun. 1996, 6, 753-754. (f) Larsen, S.; Lopez, D. H.; Kozma,
D. Acta Crystallogr., Sect. B 1993, 49, 310-316.
(1) (a) Kinbara, K. Synlett 2005, 5, 732-743. (b) Kinbara, K.; Saigo,
K. In Topics in Stereochemistry; Denmark, S. C., Ed.; Wiley & Sons: New
York, 2003; Vol. 23, Chapter 4. (c) Jacques, J.; Collet, A.; Wilen, S. H.
Enantiomers, Racemates, and Resolutions; Krieger Publishing: Malabar,
FL, 1994.
10.1021/jo052020v CCC: $33.50 © 2006 American Chemical Society
Published on Web 12/24/2005
606
J. Org. Chem. 2006, 71, 606-615

O-Substituted Phenylphosphonothioic Acids
SCHEME 1. Synthesis of Enantiopure 1a-d
method for the prediction of the difference in solubility/stability
between a pair of diastereomeric salt crystals of a racemate with
a resolving agent, on which success in chiral recognition strongly
depends. An effective solution to conquering such a situation
is to clarify not only the chiral recognition ability of a resolving
agent for a series of racemates but also the mechanism of the
chiral recognition from a crystallographic point of view. On
the basis of this consideration, we have systematically studied
the chiral recognition abilities of enantiopure arylglycolic acids
for various racemic primary amines and those of enantiopure
2-aminoarylenecyclopentanols for racemic carboxylic acids
during salt crystallization and tried to elucidate the chiral
recognition mechanism on the basis of X-ray crystallography.3
The crystallographic analyses revealed that there were mainly
three kinds of interactions for the stabilization of diastereomeric
salts and that the difference in stability between less- and more-
soluble diastereomeric salts largely depended on the existence
and/or magnitude of the three interactions: (i) the hydrogen-
bonding interaction between carboxylate anions and ammonium
cations to form a helical hydrogen-bonding network (21 column),
(ii) the CH/π interaction(s) between aryl groups in the anions/
cations to allow the T-shape arrangement of the aryl groups,
and (iii) the van der Waals interaction to achieve the close
packing of the hydrogen-bonding columns. Among them, the
hydrogen-bonding interaction is the primary concern because
the other interactions can contribute to the stabilization as a
result of the formation of the hydrogen-bonding column.
Therefore, the understanding of the hydrogen-bonding network,
formed in less-soluble diastereomeric salts, is the most important
for the elucidation of the mechanism for the chiral recognition
of racemates with a resolving agent during salt crystallization.
Compounds containing chiral phosphorus atom(s) (P-chiral)
are well-known for affording a diastereomeric circumstance
upon interacting with chiral substrates, and the P-chiral com-
pounds are used for the determination of the enantiomeric
crystallographic analyses of the less-soluble diastereomeric salts
revealed that the phosphonothioic acid and the amine formed a
globular molecular cluster in which there existed a closed
hydrogen-bonding network at the center. This hydrogen-bonding
motif is very rare and is largely different from an infinitely
hydrogen-bonded 21 column in less-soluble diastereomeric salts,
deposited during the chiral recognition of racemic primary
amines with conventional acidic resolving agents such as
mandelic acid and its derivatives, 2-naphthylglycolic acid, 10-
camphorsulfonic acid, 1-phenylethanesulfonic acid, and so on.
In this paper we report, in detail, the chiral recognition ability
of O-methyl, O-ethyl, O-propyl, and O-phenyl phenylphospho-
nothioic acids for racemic primary amines during salt crystal-
lization as well as the mechanism for chiral recognition.
Results and Discussion
Synthesis of Enantiopure O-Methyl, O-Ethyl, O-Propyl,
and O-Phenyl Phenylphosphonothioic Acids (1a-d). Racemic
O-methyl, O-ethyl, O-propyl, and O-phenyl phenylphosphono-
thioic acids (rac-1a-d) were successfully synthesized from
phenylphosphonothioic dichloride in 96, 92, 68, and 86% total
yields, respectively, according to the procedures described in
the literature,6-12 with some modifications (Scheme 1). The
enantioseparations were carried out by using (R)-PEA for 1a,8
(S)- and (R)-PEA for 1b,10 (1S,2R)-2-amino-1-indanol (AI) for
1c, and (1R,2S)-norephedrine (NE) for 1d as resolving agents,
(4) (a) Besli, S.; Coles, S. J.; Davies, D. B.; Eaton, R. J.; Hursthouse,
M. B.; Kilic, A.; Shaw, R. A.; Ciftci, G. Y.; Yesilot, S. J. Am. Chem. Soc.
2003, 125, 4943-4950. (b) Drescher, M.; Felsinger, S.; Hammerschmidt,
F.; Kahlig, H.; Schmidt, S.; Wuggenig, F. Phosphorus, Sulfur Silicon Relat.
Elem. 1998, 140, 79-93. (c) Drabowicz, J.; Dudzinski, B.; Mikolajczyk,
M.; Colonna, S.; Gaggero, N. Tetrahedron: Asymmetry 1997, 8, 2267-
2270. (d) Omelanczuk, J.; Mikolajczyk, M. Tetrahedron: Asymmetry 1996,
7, 2687-2694. (e) Drabowicz, J.; Budzinski, B.; Mikolajczyk, M. Tetra-
hedron: Asymmetry 1992, 3, 1231-1234. (f) Bentrude, W. G.; Moriyama,
M.; Mueller, H.-D.; Sopchik, A. E. J. Am. Chem. Soc. 1983, 105, 6053-
6061. (g) Moriyama, M. W.; Bentrude, G. J. Am. Chem. Soc. 1983, 105,
4727-4733. (h) Mikolajczyk, M.; Omelanczuk, J.; Leitloff, M.; Drabowicz,
J.; Ejchart, A.; Jurczak, J. J. Am. Chem. Soc. 1978, 100, 7003-7008. (i)
Harger, M. J. P. J. Chem. Soc., Perkin Trans. 2 1978, 326-331. (j)
Mikolajczyk, M.; Omelanczuk, J. Tetrahedron Lett. 1972, 16, 1539-1541.
(5) Kobayashi, Y.; Morisawa, F.; Saigo, K. Org. Lett. 2004, 6, 4227-
4230.
1
excesses of the substrates in solutions by H and 31P NMR
spectroscopies.4 Despite the high potential for chiral recognition
in solutions, the chiral recognition ability of P-chiral compounds
during salt crystallization is still unknown. Recently, we have
shown the superior chiral recognition ability of enantiopure
O-ethyl phenylphosphonothioic acid for racemic 1-phenylethyl-
amine (PEA) derivatives during salt crystallization and briefly
reported a plausible chiral recognition mechanism.5 The X-ray
(6) Duddeck, H.; Lecht, R. Phosphorus, Sulfur Relat. Elem. 1987, 29,
169-178.
(7) Hoffmann, F. W.; Kagan, B.; Canfield, J. H. J. Am. Chem. Soc. 1959,
81, 148-151.
(3) (a) Kobayashi, Y.; Kinbara, K.; Sato, M.; Saigo, K. Chirality 2005,
17, 108-112. (b) Kobayashi, Y.; Kurasawa, T.; Kinbara, K.; Saigo, K. J.
Org. Chem. 2004, 69, 7436-7441. (c) Kinbara, K.; Katsumata, Y.; Saigo,
K. Chem. Lett. 2002, 266. (d) Kinbara, K.; Kobayashi, Y.; Saigo, K. J.
Chem. Soc., Perkin Trans. 2 2000, 111-119. (e) Kinbara, K.; Harada, Y.;
Saigo, K. J. Chem. Soc., Perkin Trans. 2 2000, 1339-1348. (f) Kinbara,
K.; Kobayashi, Y.; Saigo, K. J. Chem. Soc., Perkin Trans. 2 1998, 1767-
1775. (g) Kinbara, K.; Sakai, K.; Hashimoto, Y.; Nohira, H.; Saigo, K. J.
Chem. Soc., Perkin Trans. 2 1996, 2615-2622.
(8) Allahyari, R.; Hollingshaus, J. G.; Lapp, R. L.; Timm, E.; Jacobson,
R. A.; Fukuto, T. R. J. Agric. Food Chem. 1980, 28, 594-599.
(9) Yoshikawa, H. Biosci. Biotechnol. Biochem. 1999, 63, 424-426.
(10) Lewis, V.; Donarski, J. W.; Wild, R. J.; Raushel, M. F. Biochemistry
1988, 27, 1591-1597.
(11) DeBruin, K. E.; Tang, C. W.; Johnson, D. M.; Wilde, R. L J. Am.
Chem. Soc. 1989, 111, 5871-5879.
(12) Batra, S. B.; Purnanand. Phosphorus, Sulfur Silicon Relat. Elem.
1993, 85, 169-173.
J. Org. Chem, Vol. 71, No. 2, 2006 607

Kobayashi et al.
TABLE 1. Chiral Recognition of Racemic PEA Derivatives 2a-c with Enantiopure 1a-d during Salt Crystallization and Crystal Shape and
Hydrogen-Bonding Pattern of the Less-Soluble Salts
solventa (mL)
ether/AcOEt/hexane/i-Pr2O
yieldb
(%)
eec
(%)
absolute
crystal
shapef
HB
entry
acid
amine
efficiencyd
configuratione
patterng
1
2
3
4
5
6
7
8
9
1a
2a
2b
2c
2a
2b
2c
2a
2b
2c
2a
2b
2c
1.4:0:0:0
96
61
65
78
52
45
69
50
46
68
63
75
46
97
47
0.44
0.59
0.30
0.76
0.51
0.45
0.66
0.46
0.44
0.46
0.42
0.62
Sp
Sp
Sp
Sp
Sp
Rp
Rp
Rp
Rp
Rp
Rp
Rp
S
R
S
R
R
S
R
S
S
R
R
R
needle
prism
needle
prism
prism
prism
prism
prism
prism
needle
needle
needle
21 columnh
n.d.i
1.3:0:0.9:2.6
2.8:0:0:1.4
2.7:0:0:0
1.5:0.5:0:0
1.4:0:0:0
0:0:5.9:0
0:0:2.2:0
0:0:3.5:1.9
0:1.3:13.4:0
0:1.8:13.4:0
0:5.3:0:0k
n.d.i
1b
1c
1d
98
clusterj
clusterj
n.d.i
>99
>99
95
91
95
68
66
83
n.d.i
clusterj
n.d.i
10
11
12
21 columnh
n.d.i
n.d.i
a Amount of the solvent used in the enantioseparation (mL/1 mmol). b Yield of the crystallized diastereomeric salt based on a half amount of the racemic
amine. c Enantiomeric excess (ee) of the liberated amine, which was determined by a HPLC analysis. d Efficiency is the product of the yield and the ee.
e Absolute configuration: The former is the absolute configuration of enantiopure 1 used in the enantioseparation, and the latter is the absolute configuration
of the major enantiomer of the incorporated amine. The absolute configuration was determined by an X-ray crystallographic analysis or on the basis of the
elution order in a HPLC analysis upon comparing with that of the same amine, of which the absolute configuration had been determined by an X-ray
crystallographic analysis so far. f Crystal shape of the less-soluble salt. g Hydrogen-bonding pattern observed in the less-soluble salt, which was determined
by an X-ray crystallographic analysis. h Columnar hydrogen-bonding network with a two-fold screw axis. i Could not be determined. j Molecular cluster
consisting of four acid molecules and four amine molecules. k Acetone in the amount of 3.6 mL was added.
respectively, to give enantiopure 1a-d in 16-44% yields (based
on a half amount of the racemic phosphonothioic acid used).
The stabilities of 1a-d were largely different from each other.
The phosphonothioic acids 1b and 1c were stable enough to be
reused as resolving agents several times while maintaining their
enantiomeric excesses over 99%, while 1a gradually decom-
posed to contaminate O-methyl phenylphosphonic acid (ca. 2%)
and phenylphosphonothioic acid (ca. 7%) upon standing at rt
for 2 weeks even under an argon atmosphere, and a part of 1d
was transformed to phenylphosphonothioic acid (ca. 15%) after
3 weeks.
Comparison of the Chiral Recognition Abilities of 1a-d.
At first we examined the chiral recognition abilities of the
phosphonothioic acids 1a-d for racemic PEA (2a) and its
p-substituted derivatives 2b and 2c during salt crystallization
under the same conditions, other than the solvent, to compare
the abilities of 1a-d with each other. The precipitates (the less-
soluble diastereomeric salt), deposited upon stirring a solution
of equimolar amounts of 1 and 2 at rt, were collected by
filtration, washed with hexane, and dried under reduced pressure.
The composition and amount of the solvent were adjusted so
as to control the yield of the precipitates to be as close as
possible to a range of 50-90% (based on a half amount of the
racemate used). To determine the chiral recognition ability of
1, the precipitates were treated with an aqueous alkaline solution,
and the enantiomeric excess of the amine liberated was analyzed
with chiral HPLC.
tiomeric excesses of the amines 2a-c reached almost the stage
of perfection. Thus, the chiral recognition ability of 1b was
found to be superior compared with those of 1a, 1c, and 1d.
Chiral Recognition Ability of 1b for Racemic Amines.
Because the phosphonothioic acid 1b showed an excellent
recognition ability for PEA and two p-substituted derivatives
during salt crystallization, we next examined the chiral recogni-
tion ability of 1b for a wide variety of p-substituted PEA
derivatives as well as for m- and o-substituted PEA derivatives.
To compare the results with each other, the salt crystallization
and analysis were performed under the same conditions as those
mentioned above, other than the solvent. The results are
summarized in Table 2.
The diastereomeric pairs (the less-soluble and more-soluble
diastereomeric salts) of the racemic m- and o-substituted PEA
derivatives 2l-n with 1b had very low crystallinity or gave no
crystals (Table 2, entries 9-11), although the less-soluble
diastereomeric salt of 1-(o-methoxyphenyl)ethylamine (2o) with
1b exceptionally deposited with a diastereomeric purity of 67%
(Table 2, entry 12). These results indicate that 1b is rather
unsuitable for the chiral recognition of m- and o-substituted PEA
derivatives during salt crystallization. In contrast, 1b showed a
moderate to excellent chiral recognition ability for a variety of
racemic p-substituted PEA derivatives during salt crystallization;
1b could recognize the chirality of the amines 2d-i (Table 2,
entries 1-6) as well as that of the amines 2a-c (Table 1, entries
4-6).
As shown in Table 1, the chiral recognition efficiencies (the
yield of the precipitates based on a half amount of the racemic
amine 2 used times the enantiomeric excess of the amine
incorporated in the precipitates) with 1a-d were moderate to
excellent, not depending on the O-substituent. However, 1b gave
better results than did 1a, 1c, and 1d from the viewpoint of the
enantiomeric excesses of the amines 2a-c incorporated in the
deposited salts; in all cases where 1b was applied, the enan-
There was some substituent dependence observed for the
chiral recognition with 1b; namely, the enantiomeric excesses
of the amines, 2, in the less-soluble diastereomeric salts had
the tendency to decrease gradually as the size of the substituent
became large: >99-95% for 2a-f (Table 1, entries 4-6, and
Table 2, entries 1-3), 90% for 2g (Table 2, entry 4), and 73-
31% for 2h and 2i (Table 2, entries 5 and 6). Similar dependence
was also observed when the PEA derivatives had an electron-
608 J. Org. Chem., Vol. 71, No. 2, 2006

O-Substituted Phenylphosphonothioic Acids
TABLE 2. Chiral Recognition of Racemic PEA Derivatives 2d-o with Enantiopure 1b during Salt Crystallization and Crystal Shape and
Hydrogen-Bonding Pattern of the Less-Soluble Salts
solventa (mL)
ether/AcOEt/hexane
yieldb
(%)
eec
(%)
absolute
entry
amine R
efficiencyd
configuratione
crystal shapef
HB patterng
1
2
3
4
5
6
7
8
9
p-F
p-Cl
p-Br
2d
2e
2f
2g
2h
2i
2j
2k
2l
1.4:0:0
84
84
83
60
97
87
79
48
l
96
>99
95
90
31
73
16
12
0.81
0.83
0.79
0.54
0.30
0.64
0.13
0.06
Sp
R
R
R
S
R
S
prism
prism
prism
clusterh
1.5:0.5:0
1.5:0.5:0
0.3:0:2.7
0.5:0:7.2
0.3:0:2.7
2.7:0:0.3
1.4:0:0
0.2:0:0
0.2:0:0
0.2:0:0
0.7:0:16.2
Sp
Sp
Rp
Rp
Rp
Sp
Sp
Rp
Rp
Rp
Sp
clusterh
clusterh
p-Et
needle/prismi
needle/prismi
needle
21 columnj/n.d.k
21 columnj/n.d.k
21 column
n.d.k
p-Pri
p-Pr
p-NO2
p-CN
m-Me
m-OMe
o-Me
o-OMe
n.d.k
n.d.k
n.d.k
n.d.k
n.d.k
10
11
12
2m
2n
2o
l
l
68
67
0.46
R
needle
21 column
a Amount of the solvent used in the enantioseparation (mL/1 mmol). b Yield of the crystallized diastereomeric salt based on a half amount of the racemic
amine used. c Enantiomeric excess (ee) of the liberated amine, which was determined by a HPLC analysis. d Efficiency is the product of the yield and the
ee. e Absolute configuration: The former is the absolute configuration of enantiopure 1b used in the enantioseparation, and the latter is the absolute configuration
of the major enantiomer of the incorporated amine. The absolute configuration was determined by an X-ray crystallographic analysis or on the basis of the
elution order in a HPLC analysis upon comparing with that of the same amine, of which the absolute configuration had been determined by an X-ray
crystallographic analysis so far. f Crystal shape of the less-soluble salt. g Hydrogen-bonding pattern observed in the less-soluble salt, which was determined
by an X-ray crystallographic analysis. h Molecular cluster consisting of four acid molecules and four amine molecules. i Two kinds of crystals were formed
under the same conditions as a result of polymorphism. j Columnar hydrogen-bonding network with a two-fold screw axis. k Could not be determined.
l Could not crystallize.
withdrawing substituent at the p-position. The enantiomeric
excesses of p-halo derivatives 2d-f were highly satisfactory
(Table 2, entries 1-3), while those of the p-nitro and p-cyano
derivatives 2j and 2k were very low (Table 2, entries 7 and 8).
These results clearly indicate that the degree of the chiral
recognition with 1b depends on the size/length of the substituent
at the p-position of 2 rather than on its electronic nature. The
dependence would arise from the general tendency that mol-
ecules pack as closely as possible in crystals, for which the
relative molecular size/length would be the primary concern in
the cases of two-component (acid and amine) crystals.
These successful results of the chiral recognition of p-
substituted PEA derivatives 2a-f with the phosphonothioic acid
1b during salt crystallization prompted us to determine the chiral
recognition ability of 1b for other racemic primary amines such
as 1-arylalkylamines, aliphatic amines, and 1,2-amino alcohols.
The results are shown in Table 3.
The phosphonothioic acid 1b did not show any chiral
recognition ability for 1-(1-naphthyl)ethylamine (3) and 1-(2-
naphthyl)ethylamine (4) (entries 1 and 2) during salt formation,
probably a result of the mismatch of the molecular size/length
of 3 and 4 with that of 1b; the molecular size/length of the
amines is obviously larger/longer than that of 1b. In contrast,
1b could recognize the chirality of 1-phenyl-2-methylpropyl-
amine (5) with an excellent chiral recognition efficiency of 0.85
(entry 3), even though the isopropyl group in 5 is larger than
the methyl group of 2a. Moreover, moderate to excellent chiral
recognition was also achieved for more than half of the amines
we examined (entries 4-13). It is noteworthy that 1b showed
an excellent chiral recognition ability for the aliphatic amines
and the 1,2-amino alcohols 6, 7, 12, and 13 (entries 4, 5, 10,
and 11), of which the chirality is generally difficult to recognize
by conventional resolving agents.
acids are new promising candidates of widely applicable
resolving agents with high performance.
Mechanism of Chiral Recognition Elucidated on the Basis
of X-ray Crystallographic Study. To elucidate the mechanism
for the chiral recognition with the phosphonothioic acid 1b
during salt crystallization, the X-ray crystallographic analyses
of the deposited less-soluble salts would be indispensable.
However, the deposited salts were not crystals but were powders
because the salt crystallization in the present study had to be
carried out upon stirring a solution. We then tried to recrystallize
the deposited powders and also to obtain single crystals suitable
for X-ray crystallography. The X-ray diffraction patterns of the
crystals obtained by recrystallization were identical with those
of the deposited powders (see Supporting Information), indicat-
ing that the hydrogen-bonding networks and crystal packing
modes in the crystals are identical with those in the powders,
respectively. Therefore, we carried out the microscopic analyses
of the crystals for the comparison of their crystal shapes as well
as the X-ray crystallographic analyses of the single crystals.
The microscopic analyses of the crystals of the less-soluble
diastereomeric salts with 1b gave valuable information for the
chiral recognition mechanism during salt crystallization. The
crystals of the less-soluble diastereomeric salts were classified
into two categories on the basis of the crystal shapes, prism-
and needle-type crystals, as shown in Figure 1; prism-type
crystals have never been observed in the less-soluble diaster-
eomeric salt crystals of chiral primary amines with chiral acids
as far as we know, although needle-type crystals are the same
as those commonly obtained from the less-soluble diastereomeric
salts of chiral primary amines with enantiopure carboxylic acids.
This observation strongly suggests that the chiral recognition
mechanisms through which the prism- and needle-type crystals
formed are different from each other. The crystal shapes of the
less-soluble salts with 1b are listed in Tables 1-3.
The present results concerning the chiral recognition with 1b
strongly suggest that enantiopure O-substituted phosphonothioic
J. Org. Chem, Vol. 71, No. 2, 2006 609

Kobayashi et al.
TABLE 3. Chiral Recognition of Other Racemic Amines 3-15 with 1b during Salt Crystallization and the Crystal Shape and
Hydrogen-Bonding Pattern of the Less-Soluble Salts
a Amount of the solvent used in the enantioseparation (mL/1 mmol). b Yield of the crystallized diastereomeric salt based on a half amount of the racemic
amine used. c Enantiomeric excess (ee) of the liberated amine, which was determined by a HPLC analysis. d Efficiency is the product of the yield and the
ee. e Absolute configuration: The former is the absolute configuration of enantiopure 1b used in the enantioseparation, and the latter is the absolute configuration
of the major enantiomer of the incorporated amine. The absolute configuration was determined by an X-ray crystallographic analysis or on the basis of the
elution order in a HPLC analysis upon comparing with that of the same amine, of which the absolute configuration had been determined by an X-ray
crystallographic analysis so far. f Crystal shape of the less-soluble salt. g Hydrogen-bonding pattern observed in the less-soluble salt, which was determined
by an X-ray crystallographic analysis. h Columnar hydrogen-bonding network with a twofold screw axis. i Could not be determined.
is generally low to moderate in the cases of needle-type crystals.
It is noteworthy that there was a large difference in the physical
properties of the less- and more-soluble diastereomeric salts,
which largely contributed to the achievement of excellent chiral
recognition with 1b when prism-type crystals were obtained;
the less-soluble salts deposited as crystalline powders, while
the more-soluble salts were always oils.
To make the correlation more clear, we next tried the X-ray
crystallographic analyses of the less-soluble diastereomeric salts
with 1b, belonging to the prism and needle forms, respectively.
Among the less-soluble salts with 1b, prepared in the present
study, we could obtain the single crystals of the less-soluble
diastereomeric salts of 2a, 2b, 2e, and 2f in a prism form and
FIGURE 1. Crystal shapes of the less-soluble salts with enantiopure
1; (a) prism- and (b) needle-type crystals.
2g, 2h, 2i, 2o, 3, 5, 6, and 11 in a needle form, which were
Tables 1-3 indicate that there is high correlation between
suitable for X-ray crystallographic analyses. As a result, in the
needle-type less-soluble salt crystals there commonly exists a
columnar hydrogen-bonding network that consists of hydrogen
bonds between the oxygen/sulfur atoms of the phosphonothioate
anions and the ammonium hydrogen atoms of the ammonium
cations with a two-fold screw axis in the center, a so-called 21
column. The pattern of the columnar hydrogen-bonding network
is almost the same as that generally observed in the less-soluble
diastereomeric salts of chiral primary amines with chiral
carboxylic acids, which afford needle crystals.3 In contrast, all
of the prism-type less-soluble salts have a hydrogen-bonding
pattern quite different from that in the needle-type crystals; a
globular molecular cluster is constructed from four phos-
phonothioate molecules and four ammonium molecules. In the
molecular cluster, hydrophilic hydrogen-bonding sites are
located at the center, and hydrophobic groups are positioned at
the surface of the molecular cluster. Three kinds of T-shaped
CH/π interactions13 exist to stabilize the molecular cluster.
Moreover, other CH/π interactions as well as van der Waals
the crystal shape and the degree of the chiral recognition with
1b. In the case of prism-type crystals, the enantiomeric excesses
of the amines, recovered from the less-soluble diastereomeric
salts with 1b, are over 95% (Table 1, entries 4-6, and Table 2,
entries 1-3), while the enantiomeric excesses are widely spread
from 6 to >99% for the case of needle-type crystals (Table 2,
entries 6 and 12, and Table 3). Among the less-soluble
diastereomeric salts of p-substituted PEA derivatives 2 with 1b,
the salt of the amine with a relatively small substituent tends to
form a prism-type crystal, while that with a relatively large
substituent has a tendency to give a needle-type crystal.
Moreover, in the cases of the salts of 2g and 2h with 1b (Table
2, entries 4 and 5), prism- or needle-type crystals were deposited
in a similar probability under the same conditions for the
recrystallization, most likely a result of polymorphism. This
phenomenon would be reasonably explained in terms of the
medium size of the p-substituents.
These results strongly indicate that excellent chiral recognition
is achieved in the cases of prism-type crystals, while the degree
610 J. Org. Chem., Vol. 71, No. 2, 2006

O-Substituted Phenylphosphonothioic Acids
FIGURE 2. Typical examples of hydrogen-bonding networks observed in the less-soluble salts with enantiopure 1b; (a) globular molecular cluster
in the (Sp)-1b‚(R)-2b salt, (b) its one-dimensional arrangement along with the two-fold screw axis, (c) 21 columns in the (Rp)-1b‚(S)-2f salt, and (d)
the side view of the column. Hydrogen atoms are omitted for clarity. Hydrogen bonds and CH/π interactions are shown by dotted lines and arrows,
respectively. The bond distances are in angstroms.
interaction contributes to the close packing of the molecular
clusters to form a three-dimensional crystal, although the
interactions are rather weak. The hydrogen-bonding patterns,
determined by X-ray crystallography, are listed in Tables 1-3,
and Figure 2 shows the typical hydrogen-bonding patterns of
the globular molecular cluster and the 21 column.
From Tables 1-3, high correlation is observed between the
crystal shape of the less-soluble diastereomeric salts with 1b
and their hydrogen-bonding pattern determined by X-ray
crystallography; as far as we examined, without any exception,
the prism-type crystals consist of a molecular cluster, whereas
there exists a 21 column in the needle-type crystals. This result
suggests that the hydrogen-bonding patterns, which could not
be determined by X-ray crystallography in the present study,
can be deduced on the basis of the crystal shapes.
Thus, in the chiral recognition of racemic primary amines
with 1b, high correlations were found between the degree of
chiral recognition with 1b, the crystal shape of the corresponding
less-soluble diastereomeric salts, and the hydrogen-bonding
pattern in the salt crystals.
Figure 3 shows the crystal structure of the less-soluble (Rp)-
1b‚(1S,2R)-11‚H2O salt. In the crystal there exists a 21 column
that is also composed of hydrogen bonds between the phos-
FIGURE 3. Crystal structure of the less-soluble (Rp)-1b‚(1S,2R)-11‚
H2O salt; (a) top view of the 21 columns and (b) side view of the 21
column. Hydrogen atoms are omitted for clarity. Hydrogen bonds and
CH/π interactions are shown by dotted lines and arrows, respectively.
The bond distances are in angstroms. Gray circles indicate hydrogen-
bonding networks.
phonothioate anions and the ammonium cations. Moreover,
water molecules do play a very important role in the salt crystal;
the water molecule acts as a proton donor and also as a proton
acceptor to form hydrogen bonds with the hydroxy group of
(13) Nishio, M.; Hirota, M.; Umezawa, Y. The CH/π interaction.
EVidence, Nature and Consequences; Wiley-VCH: New York, 1998.
J. Org. Chem, Vol. 71, No. 2, 2006 611

Kobayashi et al.
FIGURE 4. Crystal structure of the less-soluble (Rp)-1b‚(R)-6 salt;
(a) top view of the 21 columns and (b) side view of the 21 column.
Hydrogen atoms are omitted for clarity. Hydrogen bonds and CH/π
interactions are shown by dotted lines and arrows, respectively. The
bond distances are in angstroms.
FIGURE 5. Crystal structure of the less-soluble (Rp)-1c‚(S)-2b salt;
(a) globular molecular cluster and (b) one-dimensional arrangement
along with the two-fold screw axis. Hydrogen atoms are omitted for
clarity. Hydrogen bonds, CH/π interactions, and weak CH/π interactions
are shown in dotted lines, arrows, and red dotted lines, respectively.
The bond distances are in angstroms.
11 and the sulfur atom of 1b, which strongly reinforces the 21
column. The gauche conformation of the two phenyl groups of
11 and the packing mode of the columns are very similar to
those in the less-soluble diastereomeric salts of chiral carboxylic
acids with enantiopure 11.3f On the other hand, the chiral
recognition efficiency was highly improved when 10 was
applied in the place of 11 as a target racemic amine. This
improvement would arise from the more effective hydrogen-
bonding interactions between the phosphonothioate anions of
1b, the ammonium cations of 10, and the water molecules in
the 21 column and/or the more sufficient close packing of the
columns, as was observed in the less-soluble diastereomeric salts
of the chiral carboxylic acids with enantiopure 10.3d
Because all of the less-soluble diastereomeric salts of the
aliphatic amines 6-8 and the aliphatic amino alcohols 12-14
with enantiopure 1b gave needle-type crystals, we anticipated
that a 21 column was commonly formed in the crystals.
However, it was very hard to obtain their single crystals with
good quality for X-ray crystallography, other than the less-
soluble (Rp)-1b‚(R)-6 salt. Thus, we could carry out only the
X-ray crystallographic analysis of the (Rp)-1b‚(R)-6 salt (Figure
4). As we predicted, a 21 column was formed by hydrogen bonds
between the phosphonothioate anions and the ammonium
cations. The molecular arrangement shows that CH(sp3)/π
interactions between the phenyl groups of the 1b molecules and
the methyl groups of 6 contribute to the reinforcement of the
21 column and to the close packing of the columns, respectively.
It is noteworthy that the distance between the carbon at the
1-position of 6 and the phenyl group of 1b in the neighboring
column is very short (3.63 Å) to play a very significant role in
the close packing of the columns.
entries 1, 3, and 10-12). To clarify the mechanism for the chiral
recognition of 2a-c with 1a, 1c, and 1d, we carried out X-ray
crystallographic analyses for several kinds of the less-soluble
diastereomeric salts. As a result, high correlation between the
crystal shape and the hydrogen-bonding pattern was also
realized, as was observed in the cases with 1b; a globular
molecular cluster is formed in the prism-type less-soluble (Rp)-
1c‚(S)-2b salt, whereas there exists a 21 column in the needle-
type less-soluble (Sp)-1a‚(S)-2a and (Rp)-1d‚(R)-2a salts.
The crystal structure of the prism-type (Rp)-1c‚(S)-2b salt is
shown in Figure 5. The hydrogen-bonding pattern and the
packing mode are quite similar to those of the prism-type salts
with 1b. The hydrogen-bonding sites are located at the center
of a closed globular molecular cluster, and there are four kinds
of CH/π interactions, three of which stabilize the molecular
cluster with the other contributing to the close packing of the
molecular clusters. Although the propyl group of 1c does not
have any interaction in the molecular cluster, the propyl group
contributes to make the molecular cluster globular and to
stabilize the packing of the molecular clusters along a two-fold
screw axis; although the C‚‚‚C distance is rather long (4.38 Å),
the propyl group of 1c weakly interacts with the phenyl group
of 2b in the neighboring molecular cluster, as was observed in
the prism-type less-soluble salt with 1b (Figures 2b and 5b).
These analyses indicate that the packing mode itself is one of
the most important factors to realize a globular molecular cluster,
even when the key interaction is weak.
Figure 6 shows the crystal structures of the less-soluble (Rp)-
1d‚(R)-2a and more-soluble (Rp)-1d‚(S)-2a salts. In both salt
crystals, a 21 column is formed by hydrogen bonds between
the phosphonothioate anions and the ammonium cations, and
their columnar hydrogen-bonding networks are very similar to
each other. However, the packing modes of the 21 columns are
obviously different from each other. In the less-soluble (Rp)-
1d‚(R)-2a salt there are four kinds of CH/π interactions between
the phenyl groups of 1d and 2a to stabilize the column (one
kind) and to closely pack the columns (three kinds), while only
one kind of CH/π interaction is observed for the stabilization
of the column in the more-soluble (Rp)-1d‚(S)-2a salt. These
crystal structures suggest that the stability/solubility difference
General relationships were observed between the degree of
chiral recognition, the crystal shape of the less-soluble diaster-
eomeric crystal, and the hydrogen-bonding pattern in the chiral
recognition of 2a-c with enantiopure 1a, 1c, and 1d as well as
with 1b. From the prism-type less-soluble salts, the amines with
a high enantiomeric excess were recovered, 2b with an
enantiomeric excess of 97% from the salt with 1a (Table 1,
entry 2) and 2a-c with an enantiomeric excess of 91-95%
from the salts with 1c (Table 1, entries 7-9). In contrast, the
enantiomeric excesses of the amines 2a-c recovered from the
needle-type less-soluble salts varied from 46-83% (Table 1,
612 J. Org. Chem., Vol. 71, No. 2, 2006

O-Substituted Phenylphosphonothioic Acids
FIGURE 6. Top view of the 21 columns; (a) the less-soluble (Rp)-1d‚(R)-2a salt and (b) the more-soluble (Rp)-1d‚(S)-2a salt. Hydrogen atoms are
omitted for clarity. Gray circles indicate the hydrogen-bonding networks. Hydrogen bonds and CH/π interactions are shown in dotted lines and
arrows, respectively. The bond distances are in angstroms.
between the less- and the more-soluble diastereomeric crystals
is mainly caused by the difference in the number of CH/π
interactions contributing to the stabilization of the packing of
the columns.
in the chiral recognition during salt crystallization. Moreover,
high correlations were elucidated between the degree of the
chiral recognition with 1a-d, the crystal shape of the less-
soluble diastereomeric salts, and the hydrogen-bonding pattern
in the salts, which depended on the relative molecular size/
length of 1a-d and the target racemic amines; prism-type
crystals consisted of a globular molecular cluster without any
exception and gave always very excellent results in chiral
recognition, whereas needle-type crystals were composed of a
21 column, and the chiral recognitions varied from poor to
excellent.
As shown in Figures 2 and 5, the O-substituents of 1b and
1c play an important role for the formation of a globular
molecular cluster. Moreover, when the ethyl/propyl group in
1b and 1c is replaced by a methyl group (corresponds to 1a),
the distance between the substituent and the phenyl group in
the neighboring molecular cluster would be too long to achieve
similar interaction between the molecular clusters, even if the
formation of the molecular cluster is supposed. As a result, some
of the less-soluble salts with 1a would have to deposit in a more
stable infinite form, like a needle-type crystal. On the other hand,
when the O-substituent is a phenyl group (corresponds to 1d),
the globular molecular cluster might no longer be formed as a
result of the steric hindrance between the phenyl groups in the
phosphonothioate anions and/or the ammonium cations which
would result in the deposit of a crystal with a 21 column. These
mean that the formation and packing of globular molecular
clusters is sensitive not only to the shape of the amines 2 but
also to the O-substituent of the phosphonothioic acids 1a-d.
Experimental Section
Synthesis of Enantiopure 1a. To a solution of commercially
available phenylthiophosphonoic dichloride (25.00 g, 0.12 mol) was
slowly added a mixture of sodium methoxide (16.00 g, 0.30 mol)
in MeOH (24 mL) over a period of 1.5 h, and the mixture was
stirred at rt for 3 h. After removal of the solvent under reduced
pressure, the residue was dissolved in C6H6 (300 mL) and washed
with H2O (3 × 300 mL), and the combined aqueous layers were
extracted with C6H6 (3 × 1000 mL). The combined organic layers
were washed with saturated aq NaCl (300 mL), dried over
anhydrous Na2SO4, filtered, and concentrated under reduced pres-
sure to give crude O,O′-dimethyl phenylphosphonothioate (24.10
g, 0.12 mol, quantitative) as an oily substance. IR (NaCl) 1438,
Conclusion
Racemic O-substituted phenylphosphonothioic acids 1a-d
were synthesized from phenylphosphonothioic dichloride, and
the enantiopure forms were obtained very easily via enantio-
separations of the racemic acids. Among the enantiopure
phosphonothioic acids, O-ethyl phenylphosphonothioic acid (1b)
showed chiral recognition ability better than did O-methyl,
O-propyl, and O-phenyl phenylphosphonothioic acids (1a, 1c,
and 1d) for several kinds of racemic amines in salt crystalliza-
tion. A systematic study on the chiral recognition ability of 1b
during salt crystallization revealed that 1b showed very high
performance in the chiral recognition of various racemic amines.
The microscopic and X-ray crystallographic analyses of the
deposited (less-soluble diastereomeric) salts with 1a-d revealed
that the crystal shapes of the less-soluble diastereomeric salts
could be classified into two categories, prism- and needle-type
crystals, and that there were two independent chiral recognition
mechanisms, namely, chiral recognitions through the formation
of a globular molecular cluster and through the formation of a
21 column, of which the cluster formation is the first example
1
1123, 1024, 796, 737 cm-1; H NMR (300 MHz, CDCl3) δ 3.76
(s, 6H), 7.46-7.55 (m, 3H), 7.86-7.94 (m, 2H); 31P NMR (121
MHz, CDCl3) δ 92.11.
A solution of crude O,O′-dimethyl phenylphosphonothioate
(24.10 g, 0.12 mol) in a mixture of 12 M aq KOH (10 mL) and
MeOH (150 mL) was stirred at 80 °C for 24 h. The solution was
concentrated under reduced pressure to remove MeOH and extracted
with CH2Cl2 (3 × 200 mL). The aqueous layer was acidified with
3 M aq HCl (400 mL), and the acidic aqueous solution was
extracted with CH2Cl2 (3 × 200 mL). The combined extracts were
washed with saturated aq NaCl (300 mL), dried over anhydrous
Na2SO4, and filtered. The concentration of the organic solution
under reduced pressure gave crude racemic O-methyl phenylphospho-
nothioic acid (rac-1a; 21.60 g, 0.12 mol, 96%) with satisfactory
purity for the following enantioseparation as an oily substance. IR
(NaCl) 3150, 2460, 2440, 1380, 1050, 980, 780, 730, 710 cm-1
;
1H NMR (300 MHz, CDCl3) δ 3.81 (s, 3H), 7.42-7.57 (m, 3H),
7.86-7.94 (m, 2H); 31P NMR (121 MHz, CDCl3) δ 83.57.
To a solution of rac-1a (21.64 g, 0.12 mol) in a mixture of Et2O
(230 mL) and AcOEt (70 mL) was added (S)-PEA (9.75 g, 0.08
J. Org. Chem, Vol. 71, No. 2, 2006 613

Kobayashi et al.
mol), and the mixture was stirred at rt for 6 h. The deposited powder
was collected by filtration using a membrane filter (T050A047A,
Advantec) and recrystallized five times from AcOEt/hexane (1:3,
200 mL; 1:1, 80 mL; 2:3, 50 mL; 1:2, 60 mL; and then 5:1, 20
mL) to give the diastereopure (Sp)-1a‚(S)-PEA salt (3.94 g, 13
mmol, 22%) as a white powder. To the diastereopure salt thus
obtained was added 3 M aq HCl (500 mL), and the mixture was
stirred for 1 h and extracted with CH2Cl2 (3 × 500 mL). The
combined extracts were dried over anhydrous Na2SO4, filtered, and
concentrated under reduced pressure to afford enantiopure (Sp)-O-
methyl phenylphosphonothioic acid ((Sp)-1a; 2.40 g, 13 mmol,
quantitative) as an oily substance. The enantiomeric excess of (Sp)-
1a thus obtained was determined by a HPLC analysis (Daicel
CHIRALCEL OJ-RH; eluent, HClO4 (pH 2)/CH3CN ) 90:10; flow
rate, 0.5 mL/min; t1 (the Sp isomer) ) 31 min, t2 (the Rp isomer )
36 min); enantiomeric excess, >99%). (Sp)-1a‚(S)-PEA salt: mp
(300 MHz, CDCl3) δ 1.34 (t, J ) 7.0 Hz, 3H), 4.19 (q, J ) 7.0
Hz, 2H), 7.43-7.52 (m, 3H), 7.88-7.96 (m, 2H); 13C NMR (75
MHz, CDCl3) δ 16.10, 62.77, 128.32, 130.59, 132.24; 31P NMR
(121 MHz, CDCl3) δ 81.64.
The mother liquor of the enantioseparation mentioned above was
basified with 1 M NaOH (50 mL), and the (R)-PEA that was
liberated was extracted with CH2Cl2 (3 × 100 mL). After the
aqueous solution was acidified with 1 M HCl (100 mL), (Rp)-
enriched 1b was extracted with CH2Cl2 (3 × 100 mL). The
combined organic layers were dried over anhydrous Na2SO4,
filtered, and concentrated under reduced pressure to give (Rp)-
enriched 1b (3.36 g, 17 mmol, 96%) as an oily substance. To a
solution of (Rp)-enriched 1b (3.36 g, 17 mmol) in Et2O (50 mL)
was added (S)-PEA (2.02 g, 17 mmol), and the mixture was stirred
at rt for 6 h. The deposited solid was collected by filtration using
a membrane filter (T050A047A, Advantec) and recrystallized six
times from AcOEt (50, 25, 35, 40, 40, and then 40 mL) to give the
diastereopure (Rp)-1b‚(S)-PEA salt (0.85 g, 2.6 mmol, 16%) as a
white powder. To the diastereopure salt thus obtained was added 1
M aq HCl (100 mL), and the mixture was stirred for 1 h and
extracted with CH2Cl2 (3 × 100 mL). The combined extracts were
dried over anhydrous Na2SO4, filtered, and concentrated under
reduced pressure to afford enantiopure (Rp)-1b (0.51 g, 2.5 mmol,
96%) as an oily substance. (Rp)-1b‚(S)-PEA salt: mp 134.0-134.5
148.5-149.0 °C. (Sp)-1a:11 [R]25 ) -23.5° (c 0.36, EtOH; lit.
D
[R]24D ) -21.0° (neat));8 IR (NaCl) 3150, 2460, 2440, 1380, 1050,
1
980, 780, 730, 710 cm-1; H NMR (300 MHz, CDCl3) δ 3.81 (s,
3H), 7.42-7.57 (m, 3H), 7.86-7.94 (m, 2H); 13C NMR (75 MHz,
CDCl3) δ 52.91, 128.50, 130.82, 132.54; 31P NMR (121 MHz,
CDCl3) δ 83.57.
Synthesis of Enantiopure 1b. To a solution of phenylthio-
phosphonoic dichloride (50.00 g, 0.24 mol) in C6H6 (300 mL)
heated at 40 °C was slowly added a mixture of triethylamine (53.40
g, 0.53 mol) and EtOH (24.20 g, 0.53 mol) over a period of 3 h,
and the mixture was stirred at 60 °C for 3.5 h. After removal of
unreacted triethylamine and EtOH under reduced pressure, the
solution was washed with H2O (3 × 300 mL). The organic layer
was dried over anhydrous Na2SO4, filtered, and concentrated under
reduced pressure to give crude O,O′-diethyl phenylphosphonothioate
(53.10 g, 0.23 mol, 97%) as an oily substance. IR (NaCl) 1440,
°C. (Rp)-1b: [R]20 ) +15° (c 1.5, EtOH; lit. [R]D ) +15.1°).9
D
1
The IR, H NMR, 13C NMR, and 31P NMR were identical with
those of (Rp)-1b.
Synthesis of Enantiopure 1c. To a solution of phenylthio-
phosphonoic dichloride (8.00 g, 38 mmol) in benzene (15 mL) were
successively added triethylamine (4.25 g, 42 mmol) and 1-propanol
(25.24 g, 0.42 mol) drop by drop, and the mixture was stirred at
40 °C overnight. After removal of the unreacted 1-propanol under
reduced pressure, the solution was washed with H2O (2 × 300 mL)
and the combined aqueous layers were extracted with C6H6 (3 ×
300 mL). The combined organic layers were washed with aq NaCl
(100 mL), dried over anhydrous Na2SO4, filtered, and concentrated
under reduced pressure to give a crude mixture of O,O′-dipropyl
phenylphosphonothioate and O-propyl phenylthiophosphonic chlo-
ride (9.09 g, 20:80 on the basis of the 31P NMR integrations,
quantitative) as an oily substance.
O,O′-Dipropyl phenylphosphonothioate and O-propyl phenylth-
iophosphonic chloride could be separated by the preparative TLC
of an aliquot of the mixture.
O,O′-Dipropyl phenylphosphonothioate: IR (NaCl) 1462, 1122,
1058, 1009, 744 cm-1; 1H NMR (300 MHz, CDCl3) δ 0.94 (t, J )
7.4 Hz, 6H), 1.63-1.75 (m, 4H), 4.02 (t, J ) 7.8 Hz, 4H), 7.42-
7.53 (m, 3H), 7.87-7.95 (m, 2H); 31P NMR (121 MHz, CDCl3) δ
87.43.
1
1120, 1020, 750, 730 cm-1; H NMR (300 MHz, CDCl3) δ 1.32
(t, J ) 7.0 Hz, 6H), 4.11 (q, J ) 7.0 Hz, 4H), 7.45-7.51 (m, 3H),
7.87-7.94 (m, 2H); 31P NMR (121 MHz, CDCl3) δ 87.54.
A solution of crude O,O′-diethyl phenylphosphonothioate (53.10
g, 0.23 mol) in a mixture of 5 M aq NaOH (200 mL) and EtOH
(50 mL) was stirred at 80 °C for 18 h. The solution was concentrated
under reduced pressure to remove EtOH and then extracted with
CH2Cl2 (3 × 100 mL). The aqueous layer was acidified with 3 M
aq HCl (500 mL), and then the acidic aqueous solution was
extracted with CH2Cl2 (3 × 100 mL). The combined extracts were
washed with saturated aq NaCl (300 mL) and dried over anhydrous
Na2SO4. The concentration of the organic solution under reduced
pressure gave crude racemic O-ethyl phenylphosphonothioic acid
(rac-1b; 44.30 g, 0.22 mol, 95%) with satisfactory purity for the
following enantioseparation as an oily substance. IR (NaCl) 3000,
2380, 1440, 1120, 1020, 760, 730 cm-1 1H NMR (300 MHz,
;
CDCl3) δ 1.34 (t, J ) 7.0 Hz, 3H), 4.19 (q, J ) 7.0 Hz, 2H), 7.43-
7.52 (m, 3H), 7.88-7.96 (m, 2H); 31P NMR (121 MHz, CDCl3) δ
81.64.
O-Propyl phenylthiophosphonic chloride: IR (NaCl) 1463, 1377,
1140, 1060, 720 cm-1; 1H NMR (300 MHz, CDCl3) δ 0.93 (t, J )
7.1 Hz, 3H), 1.64-1.75 (m, 2H), 4.10 (t, J ) 6.6 Hz, 2H), 7.39-
7.54 (m, 3H), 7.83-7.95 (m, 2H); 31P NMR (121 MHz, CDCl3) δ
76.91.
To a solution of rac-1b (6.07 g, 30 mmol) in Et2O (70 mL) was
added (R)-PEA (3.64 g, 30 mmol), and the mixture was stirred at
rt for 6 h. The deposited solid was collected by filtration using a
membrane filter (T050A047A, Advantec) and recrystallized six
times from AcOEt (50, 30, 15, 25, 30, and then 40 mL) to give the
diastereopure (Sp)-1b‚(R)-PEA salt (0.88 g, 2.7 mmol, 18%) as a
white powder. To the diastereopure salt thus obtained was added 1
M aq HCl (100 mL), and the mixture was stirred for 1 h and
extracted with CH2Cl2 (3 × 100 mL). The combined extracts were
dried over anhydrous Na2SO4, filtered, and concentrated under
reduced pressure to afford enantiopure (Sp)-1b (0.54 g, 2.7 mmol,
quantitative) as an oily substance. The enantiomeric excess of (Sp)-
1b thus obtained was determined by a HPLC analysis (Daicel
CHIRALCEL OJ-RH; eluent, HClO4 (pH 2); flow rate, 0.5 mL/
min; t1 (Sp isomer) ) 24 min, t2 (Rp isomer) ) 29 min;
enantiopurity, >99%). (Sp)-1b‚(R)-PEA salt: mp 134.0-134.5 °C.
A solution of the crude mixture of O,O′-dipropyl phenylphospho-
nothioate and O-propyl phenylthiophosphonic chloride (9.09 g, 38
mmol) thus obtained in a mixture of 7 M KOH (8 mL) and
1-propanol (55 mL) was stirred at 100 °C overnight. The solution
was concentrated under reduced pressure to remove 1-propanol and
extracted with CH2Cl2 (3 × 200 mL). The aqueous layer was
acidified with 3 M aq HCl (400 mL), and the acidic aqueous
solution was extracted with CH2Cl2 (3 × 200 mL). The combined
extracts were washed with saturated aq NaCl (100 mL) and then
dried over anhydrous Na2SO4. The concentration of the organic
solution under reduced pressure gave crude racemic O-propyl
phenylphosphonothioic acid (rac-1c; 5.63 g, 26 mmol, 68%) with
satisfactory purity for the following enantioseparation as an oily
(Sp)-1b:11 [R]20 ) -15° (c 3.0, EtOH; lit. [R]D ) -15.2°);9 IR
substance. IR (NaCl) 2980, 2350, 1440, 1380, 1120, 990, 730 cm-1
1H NMR (300 MHz, CDCl3) δ 0.95 (t, J ) 7.5 Hz, 3H), 1.68-
;
D
1
(NaCl) 3000, 2380, 1440, 1120, 1020, 760, 730 cm-1; H NMR
614 J. Org. Chem., Vol. 71, No. 2, 2006

O-Substituted Phenylphosphonothioic Acids
1.76 (m, 2H), 4.06 (t, J ) 7.5 Hz, 2H), 7.46-7.52 (m, 3H), 7.89-
7.96 (m, 2H); 31P NMR (121 MHz, CDCl3) δ 82.13.
To a solution of rac-1d (9.50 g, 38 mmol) in AcOEt/hexane (1:
1, 40 mL) was added (1R,2S)-NE (3.48 g, 23 mmol), and the
mixture was refluxed for 10 h. Then the solution was slowly cooled
to rt to give a white solid mass. The solid mass was collected by
filtration using a membrane filter (T050A047A, Advantec) and
recrystallized three times from THF/hexane (3:1, 20 mL; 3:1, 10
mL; 3:1, 5 mL) to give the diastereopure (Rp)-1d‚(1R,2S)-NE salt
(1.82 g, 4.5 mmol, 24%) as a white powder. To the diastereopure
salt thus obtained was added 1 M aq HCl (200 mL), and the mixture
was stirred for 1 h and extracted with CH2Cl2 (3 × 100 mL). The
combined extracts were dried over anhydrous Na2SO4, filtered, and
concentrated under reduced pressure to afford enantiopure (Rp)-O-
phenyl phenylphosphonothioic acid ((Rp)-1d; 1.12 g, 4.5 mmol,
quantitative) as an oily substance. The enantiomeric excess of (Rp)-
1d thus obtained was determined by a HPLC analysis (Daicel
CHIRALCEL OJ-RH; eluent, HClO4 (pH 2)/CH3CN ) 9:1; flow
rate, 0.5 mL/min; t1 (Sp isomer) ) 33 min, t2 (Rp isomer) ) 39
min; enantiomeric excess, >99%). (Rp)-1d‚(1R,2S)-NE salt: mp
155.5-156.0 °C. (Rp)-1d: [R]20D ) +34° (c 0.5, EtOH); IR (NaCl)
To a solution of rac-1c (4.10 g, 19 mmol) in Et2O (60 mL) was
added (1S,2R)-AI (2.80 g, 19 mmol), and the mixture was stirred
at rt for 6 h. The deposited solid was collected by filtration using
a membrane filter (T050A047A, Advantec) and recrystallized once
from AcOEt/hexane (2:1, 160 mL) to give the diastereopure (Rp)-
1c‚(1S,2R)-AI salt (1.51 g, 4 mmol, 44%) as a white powder. To
the diastereopure salt thus obtained was added 1 M aq HCl (300
mL), and the mixture was stirred for 1 h at 0 °C and extracted
with CH2Cl2 (3 × 100 mL). The combined extracts were dried over
anhydrous Na2SO4, filtered, and concentrated under reduced pres-
sure to afford enantiopure (Rp)-O-propyl phenylphosphonothioic
acid ((Rp)-1c; 0.86 g, 4 mmol, quantitative) as an oily substance.
The enantiomeric excess of (Rp)-1c thus obtained was determined
by a HPLC analysis (Daicel CHIRALCEL OJ-RH; eluent, HClO4
(pH 2)/CH3CN ) 8:2; flow rate, 0.5 mL/min; t1 (Sp isomer) ) 28
min, t2 (Rp isomer) ) 34 min; enantiomeric excess, >99%). (Rp)-
1c‚(1S,2R)-AI salt: mp 163.5-164.0 °C. (Rp)-1c: [R]20 ) +8°
D
1
(c 0.6, EtOH); IR (NaCl) 2980, 2350, 1440, 1380, 1120, 990, 730
3180, 2380, 1590, 1490, 1230, 730 cm-1; H NMR (300 MHz,
1
cm-1; H NMR (300 MHz, CDCl3) δ 0.95 (t, J ) 7.5 Hz, 3H),
CDCl3) δ 7.13-7.19 (m, 3H), 7.26-7.33 (m, 2H), 7.48-7.60 (m,
3H), 7.98-8.05 (m, 2H); 13C NMR (75 MHz, CDCl3) δ 121.83,
125.39, 128.49, 129.57, 130.50, 132.72; 31P NMR (121 MHz,
CDCl3) δ 80.53.
An elemental analysis was performed on the salt of 1d with
(1R,2S)-NE as a result of the instability of 1d. Anal. Calcd for
C21H24NO3PS: C, 62.83; H, 6.03; N, 3.49. Found: C, 62.61; H,
6.06; N, 3.30.
Typical Procedure for the Enantioseparation of Racemic
Amines with Enantiopure 1a-d. To a solution of 1b (168 mg,
0.83 mmol) in Et2O (1 mL) was added a solution of racemic PEA
(2a; 100 mg, 0.83 mmol) in Et2O (1 mL), and the mixture was
stirred at rt for 6 h. The deposited powder was collected by filtration,
washed with hexane, and dried under reduced pressure to give the
corresponding diastereomeric salt (104 mg, 0.32 mmol, 78% based
on a half amount of 2a used). The salt was dissolved in 1 M aq
KOH (100 mL), and the aqueous solution was extracted with
chloroform (3 × 100 mL). The combined extracts were dried over
anhydrous Na2SO4 and concentrated under reduced pressure to give
enantio-enriched PEA (39 mg, 0.32 mmol).
The enantiomeric excesses of the amines were determined by
HPLC analyses on a Daicel Chiralcel CrownPak CR(+) for 2a-
g,j,k,o, 10, 11, 12, and 13, on a Daicel Chiralcel OD-H for 2h,i, 5,
6, 7, and 8, on a Daicel Chiralcel OJ-RH for 3, and on a Daicel
Chiralcel AD for 14.
1.68-1.76 (m, 2H), 4.06 (t, J ) 7.5 Hz, 2H), 7.46-7.52 (m, 3H),
7.89-7.96 (m, 2H); 13C NMR (75 MHz, CDCl3) δ 10.31, 23.74,
68.33, 128.45, 130.76, 132.40; 31P NMR (121 MHz, CDCl3) δ
82.13. Anal. Calcd for C9H13O2PS: C, 49.99; H, 6.06. Found: C,
49.82; H, 6.00.
Synthesis of Enantiopure 1d. To a solution of phenylthio-
phosphonoic dichloride (20.00 g, 95 mmol) in CH2Cl2 (40 mL)
was slowly added a mixture of triethylamine (19.20 g, 190 mmol)
and phenol (17.90 g, 190 mmol) in CH2Cl2 (60 mL) over a period
of 1 h, and the mixture was stirred at 40 °C for 4 h. After removal
of the solvent under reduced pressure, the residue was dissolved
in CH2Cl2 (300 mL). The CH2Cl2 solution was washed with H2O
(3 × 300 mL), and the combined aqueous layers were extracted
with CH2Cl2 (3 × 200 mL). The combined organic layers were
washed with aq NaCl (100 mL), dried over anhydrous Na2SO4,
filtered, and concentrated under reduced pressure to give crude
O,O′-diphenyl phenylphosphonothioate (30.96 g, 95 mmol, quan-
titative) as an oily substance. IR (NaCl) 1459, 1376, 1119, 721
1
cm-1; H NMR (300 MHz, CDCl3) δ 7.09-7.20 (m, 2H), 7.26-
7.31 (m, 8H), 7.51-7.63 (m, 3H), 8.09-8.16 (m, 2H); 31P NMR
(121 MHz, CDCl3) δ 83.64.
A solution of crude O,O′-diphenyl phenylphosphonothioate
(30.96 g, 95 mmol) in a mixture of 5 M aq KOH (50 mL) and
dioxane (300 mL) was stirred at 45 °C for 8 h. The solution was
concentrated under reduced pressure to remove dioxane, and the
aqueous solution was extracted with CH2Cl2 (3 × 200 mL). The
aqueous layer was acidified with 3 M aq HCl (400 mL), and the
acidic aqueous solution was extracted with CH2Cl2 (3 × 200 mL).
The combined extracts were washed with saturated aq NaCl (100
mL) and dried over anhydrous Na2SO4. The concentration of the
organic solution under reduced pressure gave crude racemic
O-phenyl phenylphosphonothioic acid (rac-1d; 20.54 g, 82 mmol,
86%) with satisfactory purity for the following enantioseparation
as an oily substance. IR (NaCl) 3180, 2380, 1590, 1490, 1230, 730
Supporting Information Available: General methods. X-ray
diffraction charts of the less-soluble 1a‚2a, 1b‚2a, 1b‚2f, and 1b‚
2o salts deposited in the enantioseparation and obtained under
1
diffusion conditions. H and 31P NMR spectra of (Sp)-1a, (Sp)-1b,
(Rp)-1c, and (Rp)-1d. CIF files for the crystals of (Sp)-1b‚(R)-2g,
(Rp)-1b‚(R)-2h, (Rp)-1b‚(S)-2i, (Sp)-1b‚(R)-2o, (Sp)-1b‚(R)-3, (Rp)-
1b‚(S)-5, (Rp)-1b‚(R)-6, (Rp)-1b‚(1S,2R)-11‚H2O, (Rp)-1c‚(S)-2b,
(Rp)-1d‚(R)-2a, and (Rp)-1d‚(S)-2a salts. This material is available
free of charge via the Internet at http://pubs.acs.org.
1
cm-1; H NMR (300 MHz, CDCl3) δ 7.13-7.19 (m, 3H), 7.26-
7.33 (m, 2H), 7.48-7.60 (m, 3H), 7.98-8.05 (m, 2H); 31P NMR
(121 MHz, CDCl3) δ 80.53.
JO052020V
J. Org. Chem, Vol. 71, No. 2, 2006 615