Resolution of β-unsaturated amines with isopropylidene glycerol hydrogen phthalate

Article abstract of DOI:10.1016/S0957-4166(00)00373-6

1-Phenyl-2-propenylamine 2, 1-phenyl-2-propinylamine 3, 1-(1-cyclohexenyl)ethylamine 4, (E)-2-ethylidenecyclohexylamine 5 and α-methylallylamine hydrochloride 6 were selected as candidates for resolution with isopropylidene glycerol hydrogen phthalate 1, previously described as an efficient resolving agent of 1-arylalkylamines. With only the exception of 5, all these substrates were resolved by (S)-1. In particular both the enantiomers of 2 and 3 were obtained in excellent yields and with very high enantiomeric excesses. The absolute configurations of non-racemic forms of 2, 3 and 4, not prepared before except those of 3, were established by correlation with the respective hydrogenation products. The enantiomeric excesses of all the resolved substrates were accurately determined by chiral HPLC analysis. The fact that 1 resolves 4 and 6 but not their saturated analogues and shows higher efficiency in resolving 2 and 3 than 1-phenylpropylamine indicates the positive influence of the presence of β-unsaturation on the resolvability of aminic substrates with such an acid. Copyright (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0957-4166(00)00373-6

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 11 (2000) 4017–4025
Resolution of b-unsaturated amines with isopropylidene
glycerol hydrogen phthalate
Marco Pallavicini,a,* Ermanno Valoti,a Luigi Villaa and Oreste Piccolob
aIstituto di Chimica Farmaceutica e Tossicologica, Universita` di Milano, viale Abruzzi 42, I-20131 Milan, Italy
bStudio di Consulenza Scientifica, via Borno` 5, I-20060 Sirtori (LC), Italy
Received 3 September 2000; accepted 15 September 2000
Abstract
1-Phenyl-2-propenylamine 2, 1-phenyl-2-propinylamine 3, 1-(1-cyclohexenyl)ethylamine 4, (E)-2-ethylid-
enecyclohexylamine 5 and a-methylallylamine hydrochloride 6 were selected as candidates for resolution
with isopropylidene glycerol hydrogen phthalate 1, previously described as an efficient resolving agent of
1-arylalkylamines. With only the exception of 5, all these substrates were resolved by (S)-1. In particular
both the enantiomers of 2 and 3 were obtained in excellent yields and with very high enantiomeric
excesses. The absolute configurations of non-racemic forms of 2, 3 and 4, not prepared before except
those of 3, were established by correlation with the respective hydrogenation products. The enantiomeric
excesses of all the resolved substrates were accurately determined by chiral HPLC analysis. The fact that
1 resolves 4 and 6 but not their saturated analogues and shows higher efficiency in resolving 2 and 3 than
1-phenylpropylamine indicates the positive influence of the presence of b-unsaturation on the resolvability
of aminic substrates with such an acid. © 2000 Elsevier Science Ltd. All rights reserved.
1. Introduction
A new chiral acid, the hydrogen phthalate of isopropylidene glycerol 1, has been shown to
resolve a wide range of 1-arylalkylamines with high efficiencies and with the same stereochemical
outcome.1,2 Some explanation of such a resolving ability has been provided by the study of the
physicochemical and X-ray crystallographic data of the diastereomeric salts formed by (S)-1
with one of the resolved amines, 1-phenylethylamine.3 Furthermore, the effects of structural
modifications of the resolving agent 1 and of the 1-phenylethylamine framework in the tested
substrates on the resolution process have been investigated.2,3 This has allowed some structural
features of 1 and of the amines conditioning the resolvability of the latter by the former to be
identified. In particular, the presence of a phenyl or a vinyl group bonded to the a-carbon of the
amine seems to be critical, as shown by the fact that 1-phenylethylamine and a-methylallylamine
* Corresponding author. E-mail: marco.pallavicini@unimi.it
0957-4166/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(00)00373-6

4018
M. Palla6icini et al. / Tetrahedron: Asymmetry 11 (2000) 4017–4025
are resolved by (S)-1, in contrast to their saturated analogues, 1-cyclohexylethylamine and
2-butylamine, and to 1-phenyl-2-propylamine, which fail to form solid precipitates upon
treatment with (S)-1.
Believing that the presence of b,g-unsaturation positively influences the resolvability of amines
with 1, as suggested by our precedent investigations,2 we decided to explore the applicability of
the same resolution procedure to 1-phenyl-2-propenylamine 2, 1-phenyl-2-propinylamine 3 and
1-(1-cyclohexenyl)ethylamine 4. We speculated that 2 and 3, due to the additional unsaturation,
should be resolved by 1 more efficiently than 1-phenylpropylamine, whose resolution we had
previously described.2 Analogously, 4 should, in principle, be resolvable with 1 unlike 1-cyclo-
hexylethylamine. The interest in the preparation of these b,g-unsaturated amines in non-racemic
form was also aroused by their potential importance as chiral synthons and by the fact that no
report on their synthesis had appeared except that in the case of 3, a very efficient decigram-
scale resolution via lipase catalysed acetylation has been published4 just after we had terminated
our experiments. These included two additional b,g-unsaturated substrates, (E)-2-ethylidenecy-
clohexylamine 5, which was isolated as a side product of the synthesis of 4, and a-methylallyl-
amine hydrochloride 6. The free amine corresponding to the latter had been already resolved
with (S)-1.2 However, its resolution had been stopped at the first crystallisation recovering the
S and R amines with 53 and 39% enantiomeric excess, respectively. Furthermore, those
experiments, made difficult by the volatility of the substrate, had not included an accurate
assessment of the enantiomeric compositions but simply relied on the comparison of the
observed specific rotations with the maximum rotation reported for (R)-a-methylallylamine.5 In
order to overcome such difficulties and limitations, it seemed worthwhile to attempt the
complete resolution of a-methylallylamine using the hydrochloride 6 and to develop an HPLC
analytical method allowing the progress of its enantiomeric excess over the course of the
resolution procedure to be monitored.
2. Results
The racemic amines 2 and 3 were prepared from cinnamyl alcohol and 1-phenyl-2-propyn-1-
ol, respectively, according to literature methods.6–8 For the preparation of amine 4 (Scheme 1),
we decided to start from 1-(1-cyclohexenyl)ethanol, which was in turn obtained by regioselective
1,2-reduction of 1-acetylcyclohexene as reported in the literature.9,10 Intermolecular dehydration

M. Palla6icini et al. / Tetrahedron: Asymmetry 11 (2000) 4017–4025
4019
occurring between the above allylic alcohol and phthalimide on treatment with diethyl azodicar-
boxylate and triphenylphosphine gave N-1-(1-cyclohexenyl)ethylphthalimide 8 and (E)-N-2-
ethylidenecyclohexylphthalimide 9, which were isolated by chromatography in 39.2 and 26.8%
yield, respectively. Subsequent hydrazinolysis of the two N-alkenylphthalimides 8 and 9 led to
the corresponding amines 4 and 5. The E configuration of 5 and of its precursor 9 was
determined by comparison of the NMR spectrum of the former with reported data for the E and
Z isomers of 2-ethylidenecyclohexylamine.11,12 Finally, 6 was obtained by conversion into
hydrochloride of a-methylallylamine, prepared as previously described.13
Scheme 1. (a) Phthalimide, diethyl azodicarboxylate, triphenylphosphine, THF. (b) Hydrazine monohydrate, 95%
ethanol
As for our preceding resolutions of 1-arylalkylamines,1,2 we decided to utilise the S enan-
tiomer of 1 and to combine the amines 2–5 with stoichiometric quantities of such a resolving
acid in methanol or ethanol or 2-propanol.
In the case of 2, treatment of a methanolic solution of the racemic amine with equivalent
amount of (S)-1 produced a white crystalline precipitate, the salt of the levorotatory amine with
(S)-1, in 82% yield (relative to the theoretical amount, i.e. half of the overall starting salt). The
enantiomeric excess of (−)-2 was determined by reversed-phase chiral HPLC analysis of the
precipitated salt, without previous decomposition of the latter, under conditions effective for the
separation of (S)-1, (+)-2 and (S)-1, (−)-2. This describes the observed elution of three
chromatographic peaks corresponding to the resolving agent (S)-1, the dextrorotatory amine
(+)-2 and the levorotatory amine (−)-2. It was assumed higher than 99.6% on the basis of no
chromatographic detection of (+)-2. (−)-2 was recovered from the salt in 80.2% overall yield
(40.1% of the starting 2). The HPLC analysis of the liberated amine under the same conditions
adopted for the precipitated salt confirmed the above enantiomeric purity. (+)-2 was isolated
from the methanolic solution remaining from the crystallisation, precipitated from aqueous
ethanol as a neutral sulphate and finally liberated with >99.6% e.e., determined by HPLC, and
in 42.5% overall yield (21.2% of the starting 2). To assign the absolute configurations to (−)-2
and (+)-2, the levorotatory amine, recovered from the less soluble salt with (S)-1, was
hydrogenated (H2, 10% Pd/C, ethanol) leading to (S)-(−)-1-phenylpropylamine with >99.6% e.e.,
determined by HPLC according to the previously reported method.2 On the basis of this result,
the S configuration was established for (−)-2 and, consequently, the R configuration for (+)-2.
The resolution procedure of 3 paralleled that of 2. The only differences consisted in replacing
methanol with 2-propanol in the diastereoselective crystallisation of the salt of (−)-3 with (S)-1
and in performing two crystallisations of the neutral sulphate of (+)-3 liberated from the
isopropanolic mother liquors. (−)-3·(S)-1 precipitated from the solution of 3 and (S)-1 in 72.1%

4020
M. Palla6icini et al. / Tetrahedron: Asymmetry 11 (2000) 4017–4025
yield allowing to isolate (−)-3 with >99.6% e.e. (determined by reversed-phase HPLC analysis of
both (−)-3·(S)-1 and (−)-3) and in 62.9% yield. (+)-3 was liberated from the recrystallised
sulphate of the amine recovered from mother liquors with 87.5% e.e. and in 49.2% yield. The
respective R and S configurations of (−)-3 and (+)-3 were established on the basis of the fact that
hydrogenation (H2, 10% Pd/C, ethanol) of (−)-3 led to (R)-(+)-1-phenylpropylamine with
>99.6% e.e. This result is in agreement with reported data for the hydrogenation of (+)-3
yielding (S)-(−)-1-phenylpropylamine.4
The resolution of 4 was performed in ethanol by the same procedure used for 2 and 3. The
salt of (+)-4 with (S)-1 precipitated in 60.3% yield and the enantiomeric excess of (+)-4, liberated
from a sample of such a precipitate and transformed into benzamide before chiral HPLC
analysis, proved to be 86%. Recrystallisation from the same solvent raised the enantiomeric
excess to 98.1% with a moderately decreased chemical yield (44.7%). (+)-4 was finally recovered
in 41.8% yield. Unlike the amines (+)-2 and (+)-3, the levorotatory antipode of 4 isolated from
the mother liquors was not converted into sulphate and tentatively submitted to crystallisations
considering its moderate starting enantiomeric excess (37%). The hydrogenation (H2, 10% Pd/C,
methanol) of (+)-4 gave (+)-1-cyclohexylethylamine, to which the S configuration has been
assigned,14 allowing to establish the same configuration for the dextrorotatory enantiomer of 4.
In contrast to 4, the amine 5 turned out to be unresolvable with (S)-1 failing to form solid
precipitates upon treatment with this latter acid.
Finally, the resolution of 6 was accomplished by quantitative displacement of hydrogen
chloride with (S)-1 without isolating the free amine and by two successive crystallisations of the
diastereomeric mixture from 2-propanol. The first precipitate was obtained in 59.2% yield and
contained (S)-a-methylallylamine with 71% e.e. Recrystallisation raised the enantiomeric excess
to 96.5% reducing the chemical yield to 40.5%. (S)-a-Methylallylamine was finally liberated
from the recrystallised salt with (S)-1 and its hydrochloride (S)-6 isolated in 37.7% yield (18.9%
of the starting 6). The progress of the enantiomeric excess over the course of the resolution was
verified by liberating the amine from its salts, i.e. the two salts with (S)-1 and the final
hydrochloride (S)-6, and analysing its benzamide by chiral HPLC. The S configuration was
assigned to the resolved levorotatory hydrochloride on the basis of the fact that the previously
reported partial resolution of a-methylallylamine had demonstrated that (S)-1 generates the less
soluble diastereomeric salt with (S)-a-methylallylamine.2
3. Discussion and conclusion
The definitive data for the preparation of (S)-2, (R)-3, (S)-4 and (S)-6 by selective precipita-
tion of the salts formed with (S)-1 are summarised in Table 1. Interestingly, the stereochemical
outcome of the resolution of 3 is opposite to that of the other resolutions reported herein and
of the previous ones of 1-arylalkylamines,1,2 which had invariably produced the precipitation of
the (S)(S)-salt. Comparison of the total resolution efficiencies (E) observed for 2 and 3, 82 and
72%, respectively, with the lower efficiency (60%) achieved in the case of 1-phenylpropylamine2
addresses the importance of the presence of the additional b,g-unsaturation supporting our
initial hypothesis. This is further substantiated by the resolvability, though with moderate
efficiencies (44 and 39%, respectively), of 4 and 6 which contrasts the unfeasibility of the
resolution of the corresponding saturated amines, 1-cyclohexylethylamine and 2-butylamine. In
this context, the fact that, contrary to our expectation, the b,g-unsaturated amine 5 is not

M. Palla6icini et al. / Tetrahedron: Asymmetry 11 (2000) 4017–4025
4021
Table 1
Preparation of (S)-2, (R)-3, (S)-4 and (S)-6 by selective crystallisation of the salts with (S)-1
Compd
Yield (%)a
Recovered substrate
Ee (%)b
E (%)c
Salt with (S)-1
(S)-2
(R)-3
(S)-4
(S)-6
82.0
72.1
44.7
40.5
80.2
62.9
41.8
37.7
\99.6
\99.6
98.1
82
72
44
39
96.5
a Relative to the theoretical amount, i.e. half of the starting racemate.
b Enantiomeric excess determined by chiral HPLC.
c Total resolution efficiency=chemical yield of the crystallised salt with (S)-1 (%)×enantiomeric excess of the
recovered substrate (%)/100.
resolved by 1 would simply indicate that the only presence of a b,g-unsaturation does not assure
the resolvability, which is obviously conditioned also by other structural features of the aminic
substrates.
Apart from these considerations, we finally wish to point out the very high efficiencies of the
resolutions of 2 and 3, which allow to obtain both the enantiomers of the two amines in excellent
yields using only one antipode of the resolving acid, and the possibility of scaling up such procedures
thanks to their simplicity, as made for the preparation of (R)- and (S)-isopropylidene glycerol
via resolution of 1 with 1-phenylethylamine.15
4. Experimental
Reagentswereobtainedfromcommercialsuppliersandusedwithoutfurtherpurification. Melting
points are uncorrected. 1H NMR spectra were recorded at 200 MHz on a Bruker 220 AC. Optical
rotations were measured in a 1-dm cell of 1 ml capacity using a Perkin Elmer 241 polarimeter.
HPLC analyses were performed on Chiralcel columns (250×4.6 mm i.d.) from Daicel using a Waters
510 pump and a Pye Unicam Pu 4025 UV detector (u=254 nm).
The resolving agent (S)-1 was prepared by resolution of 1 with (S)-1-phenylethylamine.16 Amine
2 was synthesised according to literature procedures.6,7 Base-catalysed addition of trichloroacetoni-
trile to cinnamyl alcohol afforded the corresponding trichloroacetimidic ester, which was submitted
to thermal rearrangement obtaining N-(1-phenyl-2-propenyl)-trichloroacetamide. Final removal
of the trichloroacetyl group by treatment with dilute base yielded 2. Amine 3 was synthesised by
hydrolysis of thecorresponding acetamide, which was in turn obtained from 1-phenyl-2-propyn-1-ol
via Ritter reaction as described by Hacksell.8 1-(1-Cyclohexenyl)ethanol 7 was prepared by
reduction of 1-acetyl-1-cyclohexene with sodium borohydride in the presence of cerium(III) chloride
using the general procedure reported in the literature.9,10 N-a-Methylallylphthalimide 10 was
obtained from 3-chloro-1-butene and potassium phthalimide as described previously.13
4.1. N-1-(1-Cyclohexenyl)ethylphthalimide 8 and (E)-N-2-ethylidenecyclohexylphthalimide 9
Diethyl azodicarboxylate (79.57 g, 456.9 mmol) was added dropwise to a mixture of 7 (57.66
g, 456.9 mmol), phthalimide (67.22 g, 456.9 mmol) and triphenylphosphine (119.84 g, 456.9 mmol)

4022
M. Palla6icini et al. / Tetrahedron: Asymmetry 11 (2000) 4017–4025
in tetrahydrofuran (700 ml) at room temperature. After the resulting solution had been stirred
overnight at room temperature, the solvent was removed in vacuo. Diethyl ether was added to
the residue to precipitate triphenylphosphine oxide and diethyl hydrazinedicarboxylate which
were filtered off. The filtrate was concentrated and the oily residue was applied to a silica gel
column which was eluted with n-heptane:ethyl acetate 95:5 obtaining, in sequence, 45.76 g
1
(39.2%) of 8 and 31.29 g (26.8%) of 9 as white solids. 8: mp 44.5–46.6°C; H NMR (CDCl3): l
1.45–1.65 (m, 4H), 1.60 (d, 3H, J=7), 1.92 (m, 2H), 2.04 (m, 2H), 4.77 (q, 1H, J=7), 5.72 (m,
1
1H), 7.65–7.75 (m, 2H), 7.75–7.88 (m, 2H). 9: mp 121–122°C; H NMR (CDCl3): l 1.25–1.70
(m, 2H), 1.55 (d, 3H, J=6.6), 1.70–2.05 (m, 4H), 2.52 (ddd, 1H, J=25, 12.5, 3.5), 2.80 (m, 1H),
4.68 (m, 1H), 4.89 (q, 1H, J=6.6), 7.65–7.75 (m, 2H), 7.78–7.88 (m, 2H).
4.2. 1-(1-Cyclohexenyl)ethylamine 4
A mixture of 8 (40.26 g, 157.7 mmol), hydrazine monohydrate (16 ml) and 95% ethanol (400
ml) was heated to 60°C for 90 min, cooled to 20°C, treated with concd hydrochloric acid (40 ml)
and filtered. The filtrate was concentrated and the resulting solid residue treated with water (200
ml) and diethyl ether (150 ml). The aqueous phase was made alkaline with KOH and, after
separating the organic layer, extracted with diethyl ether (3×150 ml) again. The ethereal extracts
were combined, dried and concentrated to give 19.1 g (96.7%) of 4 as a pale yellow oil; 1H NMR
(CDCl3): l 1.11 (d, 3H, J=6.5), 1.38 (bs, 2H), 1.40–1.70 (m, 4H), 1.80–2.10 (m, 4H), 3.31 (q,
1H, J=6.5), 5.53 (bs, 1H).
4.3. (E)-2-Ethylidenecycloexylamine 5
Obtained, in 73.3% yield, from 9 as a yellow oil by the same procedure used to prepare 4 from
1
8; H NMR (CDCl3): l 1.1–1.9 (m, 9H), 1.56 (d, 3H, J=6.6), 2.40–2.55 (m, 1H), 3.10–3.20 (m,
1H), 5.27 (q, 1H, J=6.6).
4.4. h-Methylallylamine hydrochloride 6
A mixture of 10 (86.75 g, 431.1 mmol), hydrazine monohydrate (22 ml) and 95% ethanol (350
ml) was heated at 60°C for 3 h, cooled at 5°C, treated with concd hydrochloric acid (40 ml) and
filtered. The filtrate was concentrated and the resulting solid residue treated with water (50 ml)
and diethyl ether (150 ml). The aqueous phase was made alkaline with KOH and, after
separating the organic layer, extracted with diethyl ether (5×150 ml) again. The ethereal extracts
were combined, acidified by addition of 4N HCl/EtOH and concentrated to give a solid (31.4 g),
which was crystallised from ethanol and diethyl ether obtaining 20.58 g (44.4%) of 6 as a white
1
solid: mp 124.8–125.4°C; H NMR (DMSO-d6): l 1.32 (d, 3H, J=6.7), 3.79 (m, 1H), 5.27 (d,
1H, J=10.5), 5.36 (d, 1H, J=17.4), 5.95 (ddd, 1H, J=17.4, 10.5, 6.4), 8.46 (bs, 3H). Anal. calc.
for C4H10ClN (107.58): C, 44.66; H, 9.36; Cl, 32.96; N, 13.02. Found: C, 44.36; H, 9.28; Cl,
32.49; N, 12.96.
4.5. (S)-1-Phenyl-2-propenylamine (S)-2
(S)-1 (17.74 g, 63.29 mmol) and 2 (8.43 g, 63.29 mmol) were combined in methanol (70 ml).
The precipitate, which immediately separated, was redissolved by heating the stirred mixture to

M. Palla6icini et al. / Tetrahedron: Asymmetry 11 (2000) 4017–4025
4023
boiling temperature. Stirring of the resultant solution was continued overnight at room
temperature and then at 5°C for 30 min to yield a white precipitate of (S)-1·(S)-2 (10.73 g),
which was collected by filtration and rinsed with cold methanol: mp 144.8–145.2°C; e.e. of
(S)-2 >99.6% (by HPLC analysis of the salt on a Chiralcel OD-R column; 80/20 1 M
1
NaClO4/CH3CN, 0.4 ml/min); H NMR (DMSO-d6): l 1.31 (s, 3H), 1.36 (s, 3H), 3.83 (dd, 1H),
4.06 (dd, 1H), 4.21 (d, 1H), 4.36 (m, 1H), 4.80 (d, 1H, J=6), 5.24 (d, 1H, J=10), 5.31 (d, 1H,
J=17), 6.13 (ddd, 1H, J=17, 10, 6), 7.30–7.55 (m, 8H), 7.75 (dd, 1H). The salt was decomposed
by treatment with 10% HCl (50 ml) and dichloromethane (50 ml). The organic layer was
separated and extracted with 10% HCl (3×30 ml) again. The aqueous phases were combined,
made alkaline by addition of NaOH and extracted with ethyl acetate (4×50 ml). Removal of the
solvent from the organic extracts, previously washed with brine and dried over Na2SO4, gave
(S)-2 (3.38 g, 80.2% of the theoretical amount) as a pale yellow oil: [h]2D5=−10.45 (c 4,
1
chloroform); e.e.>99.6% (by HPLC under the same conditions as described for (S)-1·(S)-2); H
NMR (CDCl3): l 1.58 (bs, 2H), 4.53 (d, 1H, J=7), 5.12 (d, 1H, J=10), 5.20 (d, 1H, J=17), 6.04
(ddd, 1H, J=17, 10, 7), 7.20–7.40 (m, 5H).
4.6. (R)-1-Phenyl-2-propenylamine (R)-2
The methanolic solution remaining from the isolation of (S)-1·(S)-2 was concentrated and the
residual salt decomposed in the same way as described for (S)-1·(S)-2 obtaining the R-enriched
amine (3.86 g): [h]2D5=+8.37 (c 4, chloroform). The recovered amine was dissolved in 95%
ethanol (25 ml) and combined with a solution of concentrated sulphuric acid (0.81 ml) in 95%
ethanol (25 ml). The solution was concentrated and the resultant solid dissolved in boiling
ethanol (18 ml). The alcoholic solution was slowly cooled to room temperature under stirring to
yield a white precipitate (2.46 g), which was isolated by filtration and treated with 1 M NaOH
(30 ml) and ethyl acetate (20 ml). The organic layer was separated and the aqueous phase
extracted with ethyl acetate (3×20 ml) again. The organic extracts were combined, washed with
brine, dried over Na2SO4 and concentrated to give (R)-2 (1.79 g, 42.5% of the theoretical
amount, i.e. half of 8.43 g of starting racemic amine) as an oil: [h]2D5=+10.23 (c 4, chloroform);
1
e.e.>99.6% (by HPLC under the same conditions as described for (S)-1·(S)-2); H NMR
identical to (S)-2.
4.7. (R)-1-Phenyl-2-propinylamine (R)-3
(S)-1 (18.48 g, 65.94 mmol) and 3 (8.65 g, 65.94 mmol) were combined in 2-propanol (175 ml).
The resultant solution was stirred overnight at room temperature to yield a white precipitate of
(S)-1·(R)-3 (9.78 g), which was collected by filtration and rinsed with cold 2-propanol: mp
105.3–106.5°C dec.; e.e. of (R)-3>99.6% (by HPLC analysis of the salt on a Chiralcel OD-R
1
column; 85/15 1.5 M NaClO4/CH3CN, 0.5 ml/min); H NMR (DMSO-d6): l 1.31 (s, 3H), 1.36
(s, 3H), 3.50 (bs, 1H), 3.82 (dd, 1H), 4.08 (dd, 1H), 4.25 (d, 2H), 4.36 (m, 1H), 4.92 (bs, 1H),
7.30–7.50 (m, 3H), 7.50–7.70 (m, 5H), 7.78 (m, 1H). The salt was decomposed in the same way
as described for (S)-1·(S)-2 yielding (R)-3 (2.72 g, 62.9% of the theoretical amount) as a white
low-melting solid: [h]2D0=−27.9 (c 10, ethanol), [h]D23=−30.9 (c 1, chloroform) [lit.4 [h]D23=−31.3
(c 1, chloroform)]; e.e.>99.6% (by HPLC under the same conditions as described for (S)-1·(R)-
1
3); H NMR (CDCl3): l 1.85 (s, 2H), 2.50 (d, 1H, J=2), 4.79 (d, 1H, J=2), 7.30–7.45 (m, 3H),
7.50–7.60 (m, 2H).

4024
M. Palla6icini et al. / Tetrahedron: Asymmetry 11 (2000) 4017–4025
4.8. (S)-1-Phenyl-2-propinylamine (S)-3
The filtrate from the crystallisation of (S)-1·(R)-3 was concentrated and the residual salt
decomposed in the same way as described for (S)-1·(S)-2 obtaining the S-enriched amine (4.46
g): e.e. 55% (determined, before the decomposition, by HPLC analysis of the salt under the same
conditions as described for (S)-1·(R)-3). The recovered amine was dissolved in 95% ethanol (22
ml), combined with a solution of concentrated sulphuric acid (0.95 ml) in 95% ethanol (23 ml)
and stirred overnight at room temperature and then at 0°C for 30 min to yield a white
precipitate (4.29 g) which was isolated by filtration and treated with 2.5 M NaOH (20 ml) and
ethyl acetate (25 ml). The organic layer was separated and the aqueous phase extracted with
ethyl acetate (3×25 ml) again. The organic extracts were combined, washed with brine, dried
over Na2SO4 and concentrated to give (S)-3 (2.98 g). The latter was analysed by HPLC under
the same conditions as described for (S)-1·(R)-3 and its enantiomeric excess found to be 69%.
Treatment with concentrated sulphuric acid (0.63 ml) in 95% ethanol (31 ml) was repeated
isolating 3.03 g of precipitated sulphate, which yielded, upon decomposition of the salt as above,
(S)-3 (2.13 g, 49.2% of the theoretical amount, i.e. half of 8.65 g of starting racemic amine) as
a white low-melting solid: [h]2D0=+23.75 (c 10, ethanol); e.e. 87.5% (by HPLC under the same
1
conditions as described for (S)-1·(R)-3); H NMR identical to (R)-3.
4.9. (S)-1-(1-Cyclohexenyl)ethylamine (S)-4
(S)-1 (30.13 g, 107.50 mmol) and 4 (13.46 g, 107.50 mmol) were combined in ethanol (90 ml).
The resultant solution was stirred overnight at room temperature and then at 0°C for 1 h to
yield a white precipitate (13.14 g), which was collected by filtration and rinsed with ethanol
cooled to –5°C. (S)-4, liberated from a sample of this precipitate, was 86% enantiomerically pure
(e.e. determined by HPLC of the corresponding benzamide on a Chiralcel-OD column;
hexane/2-propanol 9/1, 1 ml/min). The solid was recrystallised from ethanol (30 ml) to give 9.75
g of (S)-1·(S)-4: mp 132.8–133.8°C; e.e. of (S)-4 98.12% (determined, after liberation of a
sample of amine and conversion into benzamide, by HPLC analysis under the above-mentioned
1
conditions); H NMR (DMSO-d6): l 1.28 (d, 3H), 1.30 (s, 3H), 1.35 (s, 3H), 1.45–1.70 (m, 4H),
1.90–2.10 (m, 4H), 3.62 (q, 1H), 3.84 (dd, 1H), 4.06 (dd, 1H), 4.18 (d, 2H), 4.35 (m, 1H), 5.74
(bs, 1H), 7.30–7.55 (m, 3H), 7.73 (d, 1H). The salt was decomposed by treatment with 10% HCl
(30 ml) and dichloromethane (30 ml). The organic layer was separated and extracted with 10%
HCl (2×30 ml) again. The aqueous phases were combined, made alkaline by addition of NaOH
and extracted with ethyl ether (4×100 ml). Removal of the solvent from the organic extracts,
previously washed with brine and dried over Na2SO4, gave (S)-4 (2.81 g, 41.8% of the theoretical
amount) as a pale yellow oil: [h]2D5=+1.0 (neat); e.e. identical to that previously determined for
1
the amine liberated from a sample of (S)-1·(S)-4; H NMR identical to 4.
4.10. (S)-h-Methylallylamine hydrochloride (S)-6
A solution of sodium methoxide (9.79 g, 181 mmol) in methanol (20 ml) was added to a
stirred solution of 6 (19.5 g, 181 mmol) in methanol (30 ml) at 5°C. After 15 min, precipitated
sodium chloride was quickly removed by suction filtration, maintaining the filtrate at 5°C. The
latter was added with a methanolic solution of (S)-1 (50.73 g, 181 mmol) and concentrated to
give a yellow viscous oil (62 g), which was dissolved in 2-propanol (160 ml). The solution was

M. Palla6icini et al. / Tetrahedron: Asymmetry 11 (2000) 4017–4025
4025
stirred overnight at room temperature and then at 0°C for 30 min to yield a white precipitate
(18.84 g), which was collected by filtration and rinsed with 2-propanol cooled to 5°C.
(S)-a-Methylallylamine, liberated from a sample of this precipitate by treatment with ether and
2N HCl, successive alkalisation of the separated aqueous phase with 6N NaOH and final
extraction with ether, was 71% enantiomerically pure (e.e. determined by HPLC, after treatment
of the ethereal solution of the amine with benzoyl chloride in the presence of triethylamine, on
a Chiralcel-OD column; hexane/2-propanol 95/5, 0.7 ml/min). The solid was recrystallised from
2-propanol (40 ml) to give 12.89 g of the salt of (S)-a-methylallylamine with (S)-1: mp
108–110°C; e.e. of (S)-a-methylallylamine 96.50% (determined, after liberation of the amine
from a sample of the salt and conversion into benzamide, by HPLC analysis under the
1
above-mentioned conditions); H NMR (DMSO-d6): l 1.28 (d, 3H, J=6), 1.28 (s, 3H), 1.33 (s,
3H), 3.72–3.84 (m, 2H), 4.04 (dd, 1H, J=8.4, 6.2), 4.16 (d, 2H, J=5.9), 4.33 (m, 1H), 5.20 (dd,
1H, J=10.3, 1.1), 5.30 (dd, 1H, J=17.2, 1.1), 5.94 (ddd, 1H, J=17.2, 10.3, 6.2), 7.34–7.48 (m
3H), 7.69–7.74 (m, 1H), 8.50 (bs, 3H). The salt was decomposed by treatment with 1N HCl (50
ml) and dichloromethane (140 ml). The aqueous phase was separated, washed with
dichloromethane (2×40 ml) and concentrated to give a solid residue (4.03 g), which was
crystallised from ethanol and ether yielding (S)-6 (3.68 g, 37.7% of the theoretical amount) as
a white solid: mp 130–132°C; [h]2D0=−3.5 (c 10, ethanol); e.e. identical to that previously
1
determined for the amine liberated from a sample of the salt with (S)-1; H NMR identical to
6.
Acknowledgements
This work was supported by the Ministero dell’Universita` e della Ricerca Scientifica e
Tecnologica of Italy.
References
1. Pallavicini, M.; Valoti, E.; Villa, L.; Piccolo, O. Tetrahedron: Asymmetry 1996, 7, 1117.
2. Pallavicini, M.; Valoti, E.; Villa, L.; Piccolo, O. Tetrahedron: Asymmetry 1997, 8, 1069.
3. Pallavicini, M.; Valoti, E.; Villa, L.; Piccolo, O.; Marchetti, F. Tetrahedron: Asymmetry 2000, 11, 1957.
4. Messina, F.; Botta, M.; Corelli, F.; Schneider, M. P.; Fazio, F. J. Org. Chem. 1999, 64, 3767.
5. Ringdahl, B.; Dahlbom, R. Chem. Scr. 1977, 12, 47.
6. Overman, L. E. J. Am. Chem. Soc. 1976, 98, 2901.
7. Wolber, E. K. A.; Schmittel, M.; Ruchardt, C. Chem. Ber. 1992, 125, 525.
8. Nilsson, B. M.; Hacksell, U. J. Heterocyclic Chem. 1989, 26, 269.
9. Gemal, A. L.; Luche, J. L. J. Am. Chem. Soc. 1981, 103, 5454.
10. Tanis, P. S.; Head, D. B. Tetrahedron Lett. 1982, 23, 5509.
11. Chaabouni, R.; Laurent, A.; Marquet, B. Tetrahedron 1980, 36, 877.
12. Grigg, R.; Santhakumar, V.; Sridharan, V.; Thornton-Pett, M.; Bridge, A. W. Tetrahedron 1993, 49, 5177.
13. Roberts, J. D.; Mazur, R. H. J. Am. Chem. Soc. 1951, 73, 2509.
14. Lightner, D. A.; An, J. Y. Tetrahedron 1987, 43, 4287.
15. Pallavicini, M.; Valoti, E.; Villa, L.; Piccolo, O. Tetrahedron: Asymmetry 1994, 5, 5.
16. As described in Ref. 1, (S)-1 and (R)-1 can be indifferently prepared either by resolution of 1 with 1-phenylethyl-
amine or by esterification of (R)- and (S)-isopropylidene glycerol with phthalic anhydride, respectively.