Revisiting optically active quaternary derivatives made from prolinol as phase transfer catalysts

Article abstract of DOI:10.1016/S0957-4166(98)00213-4

New non-racemic ammonium salts derived from prolinol have been prepared which are stereogenic both at N and at C. Absolute configurations rest on X- ray structures for 3c and 7a. The new compounds have been tested in a number of known enantioselective phase transfer catalytic (PTC) reactions to gain insight into steric control factors in such processes. Only very moderate e.e.s were observed. A SAMP hydrazone related catalyst (7a) reported by other authors was fully characterized. Racemic reaction products were obtained in two alkylations of diphenylmethylene benzylimine in the presence of this pure catalyst 7a. Very high e.e.s reported previously by others could not be reproduced.

Full text of DOI:10.1016/S0957-4166(98)00213-4

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 9 (1998) 2235–2244
Revisiting optically active quaternary derivatives made from
prolinol as phase transfer catalysts1
Eckehard Volker Dehmlow,∗ Robert Klauck, Stephanie Düttmann,† Beate Neumann and
Hans-Georg Stammler ‡
Fakultät für Chemie, Universität Bielefeld, Postfach 100131, D-33501 Bielefeld, Germany
Received 30 January 1998; accepted 25 May 1998
Abstract
New non-racemic ammonium salts derived from prolinol have been prepared which are stereogenic both at N
and at C. Absolute configurations rest on X-ray structures for 3c and 7a. The new compounds have been tested in
a number of known enantioselective phase transfer catalytic (PTC) reactions to gain insight into steric control
factors in such processes. Only very moderate e.e.s were observed. A SAMP hydrazone related catalyst (7a)
reported by other authors was fully characterized. Racemic reaction products were obtained in two alkylations
of diphenylmethylene benzylimine in the presence of this pure catalyst 7a. Very high e.e.s reported previously by
others could not be reproduced. © 1998 Elsevier Science Ltd. All rights reserved.
1. Introduction
Two years ago, J. J. Eddine and M. Cerqaoui reported in this journal enantiomeric excesses of up
to 94% in C-alkylations of benzophenone benzylimine using (2S)-1-methyl-1-[N-(diphenylmethylene)]-
2-hydroxymethylpyrrolidine hydrazonium iodide 7a or the analogous derivative prepared from SAMP
7b as a phase transfer catalysts.2 This information met with strong interest from workers in the PTC
field3 because high e.e.s are relatively rare in PTC. They are realized most often only after very
extensive optimization work (reviews: lit.4a–d). We paid special attention to the mentioned publication
as it was closely related to our ongoing work which we shall cover here first. Generally, derivatives
of cinchona alkaloids have been found to be the most promising enantioselective PT catalysts. The
earlier work of our group with various substituted non-racemic catalysts was disappointing in terms
of the enantioselectivity.5 Using the cinchona catalysts as standards and L-proline as one of several
∗
†
‡
Corresponding author. E-mail: DEHMLOW@HRZ.Uni-Bielefeld.DE
Stereochemical control experiments.
X-Ray structural work.
0957-4166/98/$ - see front matter © 1998 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(98)00213-4

2236
E. V. Dehmlow et al. / Tetrahedron: Asymmetry 9 (1998) 2235–2244
Fig. 1. ORTEP drawing of the molecular structure of 3c in the crystal
chiral motives, we therefore tried to find out which structural elements were responsible for effective
enantioselective catalysis.
2. Results and discussion
When (L)-N-methylprolinol 1 was reacted with benzyl bromide, p-trifluorobenzyl bromide, or 2-
bromomethylnaphthalene, difficult to separate mixtures of epimers 2:3 (1:1 to 1:1.5) were obtained.
Repeated recrystallization resulted in the isolation of a pure diastereoisomer 2c only in the case of the
methylnaphthyl compound. Polymerization sensitive 2-bromomethylquinoline and 1 led to a darkish oil,
from which 2d could nevertheless be crystallized. There was no indication for the epimer in the mother
liquid here.
The direct preparation of the other pure isomer became possible in each case, however, when the
introduction of the two groups on nitrogen was reversed. Thus, conversion of N-benzylprolinol, N-(2-
methylnaphthyl)prolinol and N-(2-methylquinolinyl)prolinol with methyl iodide gave highly diastereo-
selective alkylations allowing the isolation of pure iodides 3a, 3c, and 3d. Ion exchange transformed these
into the chlorides which are more suitable as PT catalysts. Compounds 3c,d proved to be the epimers of
2c,d (for assignment of absolute configurations, see below).
The reaction between 4-bromomethylquinoline (which is even more sensitive than the 2-bromomethyl
compound) and 1 gave a tarry mass from which a 40% yield of 3e (R=4-methylquinolyl) could be
obtained in crystalline form. N-(4-Methylquinolyl)prolinol could not be made accessible. The ether
formation of benzylprolinol with 1-naphthol by a Mitsunobu type reaction and further conversion
furnished 4, and a simple etherification by 2-bromomethyl quinoline plus conventional elaboration gave
5. These two new PTs were designed to provide additional sites for interaction with substrates in catalytic
transition states.
Relative and absolute stereochemistries of compounds 2 (1R,2S) and 3–5 (1S,2S) rest on the known
configuration of (L)-proline, an X-ray structure determination of 3c (Fig. 1),6 and the NMR spectroscopic
correlation of this compound with the others. Chemical shifts of the two diastereomeric hydrogens of the
CH2–Ar residues are positioned 0.62–0.79 ppm apart in compounds 2 due to differential shielding by the
cis hydroxymethyl group. The ∆δ value of the respective signals in compounds 3 is much smaller, only
0.1–0.39.
It is noteworthy that methyl iodide is taken up with high preference from the same side as the
hydroxymethyl group, possibly due to pre-complexation by hydrogen bonding, whereas the more
voluminous alkyl groups do not exhibit much steric preference.

E. V. Dehmlow et al. / Tetrahedron: Asymmetry 9 (1998) 2235–2244
2237
The 6 new catalysts were evaluated in several enantioselective PTC reactions well documented in the
literature (1–5):
Reaction (1): Ph2C_N–CH2–COOR+Br–CH2–Ph-→Ph2C_N–C*H(CH2Ph)–COOR (R=Et, t-Bu)
Reaction (2): Ph–CO–t-Bu (NaBH4)-→Ph–CH*(OH)–t-Bu
Reaction (3): Phthalimid–K+Br–CH(CH3)–COOEt-→PhthN–CH*(CH3)–COOEt
Reaction (4): Michael addition: 2-methoxycarbonyl-1-indanone+butenone
Reaction (5): Michael addition: 2-ethoxycarbonylcyclohexanone+butenone
Reaction (6): Ph2C_N–CH2–Ph-→Ph2C_N–CH*(R)–Ph-→H2N–CH*(R)–Ph
The highest previous e.e.s mostly rest on polarimetry or NMR shift reagent studies whereas most of our
results were obtained by GC or HPLC on chiral columns. Reaction (1) was scrutinized by O’Donnell and
gave up to 81% e.e. for the t-Bu ester with N,O-dibenzylcinchonidinium bromide.7 Use of a chiral crown
ether catalyst led to only 24% e.e.8 Under the present conditions (powdered NaOH, CH2Cl2, 10 mol%
catalyst, r.t., 24 h) the ethyl ester derivative exhibited 59% e.e. with N-benzylcinchonidinium chloride
(BCDC), and 23 or 19% with 2d and 3e, respectively. Even lower enantiomeric excesses were found with
the other catalysts.
Interestingly, the isomer formed in excess was always (S) irrespective of the stereochemistry at
nitrogen. The two best catalysts contain two structural elements of the cinchona alkaloids, a flat quinoline
ring region and a hydroxyl group positioned opposite to it. Such an arrangement is believed to form part
of a binding pocket for substrates in cinchona catalysts.
Reaction (2) has been executed by many PTC workers.4 Colonna reported an e.e. of 32% us-
ing N-benzylquininium chloride (QUIBEC) in benzene.9 We found 28% in benzene and 27% in
dichloromethane with the same catalyst and 23% with BCDC in toluene/CH2Cl2 (always with a pref-
erence for the (R) alcohol). Testing our new catalysts only 3e reached 15% e.e., the others were even
less suitable. Here the (S) alcohol was preferred. It should be noted that we reported a 39.5% e.e. of the

2238
E. V. Dehmlow et al. / Tetrahedron: Asymmetry 9 (1998) 2235–2244
(S) isomer in 1994 using catalyst 6 as an inseparable diastereomeric mixture at the quaternary nitrogen.5
Using the stereochemical information obtained in the present study we now know that the catalyst sample
was 90% (S), (as drawn in 6) and 10% (R) at the quaternary N.
Reactions (3–5) gave only negligible enantiomeric excesses with our new catalysts. Maximum e.e.
values in the literature are 19.5 [reaction (3)],10 27 [reaction (4)]11,12 and 22 [reaction (5)].13 We found
that PT catalysts accelerate the Michael additions only moderately under most sets of reaction conditions.
Thus, a simple deprotonation/addition without the help of a catalyst competes with the chiral PTC
process, and (low) enantioselectivity can be achieved only in a small ‘reaction window’. All in all, our
results reinforce previous generalizations: enantioselective PTC is very sensitive to detailed reaction
conditions and the structure of substrates. A relatively ‘good’ catalyst in one reaction may be worthless
in another conversion.
In view of these results we turned to a reconsideration of Eddine and Cherqaoui’s work.2 Reaction (6)
was described as proceeding with 91% e.e. (R=benzyl or ethyl) using a catalyst with an unspecified con-
figuration at N (possibly 7a) with a mixture of powdered KOH and K2CO3 as base in dichloromethane.
Similarly, three other alkylations with allyl, butyl, and isopropyl halides were stated as proceeding with
90% or more e.e. The authors reported that no conversion occurred with normal quaternary ammonium
salts. We chose the methylation as a test reaction because both enantiomers of phenethylamine are
commercially available for comparison purposes. This particular reaction was not performed by Eddine
and Cherqaoui. In addition, the ethylation was scrutinized.
The previous researchers did not characterize their catalyst salt by C,H,N analysis, spectral data,
or even m.p. Although quaternary methyl iodides of some hydrazones are known, they are relatively
unstable. In the synthesis of optical active ketones via the Corey–Enders SAMP hydrazones, one of the
methods for regenerating the ketone is actually quaternization and subsequent hydrolysis with water, aq.
NaHCO3, or 1.4 N HCl.14–16
The cited paper does not contain a full experimental section, and thus we cannot know exactly
which procedures were used by the authors. In the beginning we had difficulties in quaternizing the
benzophenone-N-aminoprolinol hydrazone. The procedure of Smith and Tan17 apparently used by
Eddine and Cherqaoui (boiling with MeI in ethanol) did not work in our hands. Compound 7a was
obtained, however, after refluxing the respective hydrazone for 3 days in excess methyl iodide (52%
yield). We were not successful in making 7b from benzophenone-SAMP hydrazone in a similar way.
Characterization (sharp m.p., C,H,N analysis, NMR spectra) showed that 7a was a single epimer. To
confirm the place of methylation beyond any doubt, benzophenone N,N-dimethylhydrazone was reacted
with MeI.17 The obtained salt showed the expected 9 proton singlet in its NMR spectrum. As for the
configuration at N in 7a, we expected (1R) in analogy to compound 3c. However, X-ray structure
determination (Fig. 2) showed it to be (1S).
The alkylated imines obtained in reaction (6) are sensitive and difficult to handle. They were hy-
drolyzed to the amines therefore before determination of the stereochemical outcome both by the previous
authors and by us. Our observations are at variance with the ones of the previous workers in several
respects:
(i) Whereas the authors do not mention difficulties we found hydrolysis to be slow. We saw incomplete

E. V. Dehmlow et al. / Tetrahedron: Asymmetry 9 (1998) 2235–2244
2239
Fig. 2. ORTEP drawing of the molecular structure of 7a in the crystal
hydrolysis after 2 days of stirring at room temperature with 10% aqueous HCl/t-BuOMe and still
not a totally complete reaction after 15 h at 70°C. Our liberated amines were then subjected to g.c.
on a chiral column. A separation of enantiomers was possible for the compounds with R=Me and
Et.
(ii) Contrary to the published statement we find that the chemical reaction as such proceeds with many
achiral and chiral PT catalysts including cetyltrimethylammonium bromide, BCDC, and our new
catalysts.
(iii) All catalysts gave only racemic product in methylation (Me2SO4) and ethylation (Et–Br). Specific-
ally, no enantiomeric excesses were observed with 7a as catalyst within analytical error (1–2%),
and the same was true for all other tested catalysts (BCDC, BCNC, QUIBEC, 2c, 2d, 3a, 3c–e, 4
and 5).
Conceivably, the discrepancies between our findings of racemic product and Eddine’s high e.e.s
might have been due to a racemization somewhere in the process. It is known from the literature,
however, that HCl hydrolysis of optically active diphenylmethylene 1-phenethylamine does not result
in racemization.18 To test for it under the present conditions, we prepared the imine from an authentic
(R) compound and hydrolyzed it again. No racemization occurred.
A second possibility consists of a racemization of the imine concurrent with its formation. This
then would make a high enantiomeric excess impossible a priori. Scrutinizing the literature revealed a
classical study on reaction mechanisms by Ingold and coworkers.18,19 In it, rates of parallel equilibration

2240
E. V. Dehmlow et al. / Tetrahedron: Asymmetry 9 (1998) 2235–2244
and racemization were measured for a number of systems, including the one of present concern at
temperatures between 21 and 85°C.
Applying the alkaline conditions of the present study (powdered KOH/K2CO3, CH2Cl2, r.t., 24 h) to
authentic optically active diphenylmethylene 1-phenethylamine we observed a decline in specific rotation
of 45–65%. Partial racemization occured both in the absence and presence of the PT catalyst. As the
alkylation reaction is comparatively slow and needs long reaction times, the generation of substantial
enantiomeric excesses by the present method is impossible. This should apply not only to methylation
and ethylation but similarly to all alkylations.
The earlier authors determined their reported enantioselectivities by polarimetry and reference to
the highest known rotations. Considering that the used catalyst was not well characterized, we must
assume that complex mixtures were submitted to polarimetry (although a chromatographic purification
is mentioned in the text). Possibly compounds were involved which were impurities of the catalyst or
were formed by its decomposition. A β-hydroxyhydrazonium ion can be fragmented with base to give
an epoxyalkylhydrazine, for instance. Errors of this and other kinds were corrected by us before:20–22 if
the impurities have high specific rotations and if the researcher relies on polarimetry alone, an optically
active reaction product may be simulated.
3. Experimental
Enantiomeric mixtures were analyzed either by GC (1 m ‘Methyl-Sil’ retention gap pre-column and 50
m×0.25 mm ‘FS-Hydrodex-β-PM’ column of Macherey & Nagel, Düren, Germany) or HPLC (‘Chirasep
DNBPG’ 250×4 mm, 5 µm particle size of Merck, Darmstadt).
3.1. (1R,2S)-2-Hydroxymethyl-N-methyl-N-(2-naphthylmethyl)pyrrolidinium bromide 2c
By stirring (S)-N-methylprolinol with 2-bromomethylnaphthalene in MeCN at r.t. for 1 day, evapor-
ation of the solvent and repeated fractional crystallization from acetonitrile gave the title compound.
1
Yield of pure isomer: 19%; m.p. 135°C; [α]D26=+12.7 (c=1.0, MeOH). H NMR (500 MHz, CDCl3):
δ=2.16–2.35 (m, 4H), 3.17 (s, 3H), 3.31–3.43 (m, 1H), 4.00–4.20 (m, 3H), 4.43–4.62 (m, 1H), 4.62 (d,
J=13.0 Hz, 1H), 5.24 (d, J=13.0 Hz, 1H), 7.46–7.87 (m, 6H), 8.04 (s, 1H). Anal. calcd for C17H22NOBr
(336.27): C, 60.72; H, 6.59; N, 4.17; found: C, 59.90; H, 6.70; N 4.01.
3.2. (1S,2S)-2-Hydroxymethyl-N-methyl-N-(2-naphthylmethyl)pyrrolidinium iodide 3c
The compound was obtained from (S)-N-(2-naphthylmethyl)prolinol (prepared by reaction of prolinol
with the bromide and NaHCO3 in ethanol) with excess MeI by heating in acetonitrile for 24 h. 85%
Yield; m.p. 125°C, [α]D26=−8.9 (c=1, EtOH). 1H NMR (250 MHz, CDCl3): δ=2.16–2.35 (m, 4H), 3.17
(s, 3H), 3.31–3.43 (m, 1H), 4.00–4.20 (m, 3H), 4.43–4.62 (m, 1H), 5.17 (d, J=12.9 Hz, 1H), 5.31 (d,
J=12.9 Hz, 1H), 7.46–7.90 (m, 6H), 8.18 (s, 1H). 13C NMR (62 MHz, CD3OD): δ=20.4 (C-4), 24.8
(C-3), 43.8 (CH3), 60.3 (CH2OH), 65.25 (C-5), 69.1 (CH2–Ar), 75.65 (C-2), 127.1, 128.0, 128.7, 128.8,
129.4, 129.5, 130.15, 134.4, 134.5, 135.25 (arom.). Anal. calcd for C17H22NOI (383.24): C, 53.27; H,
5.79; N, 3.65; found: C, 53.26; H, 5.75; N, 3.63. The chloride was made by passing the iodide in methanol
over an ion exchange column (chloride form).

E. V. Dehmlow et al. / Tetrahedron: Asymmetry 9 (1998) 2235–2244
2241
3.3. (1R,2S)-2-Hydroxymethyl-N-methyl-N-(2-quinolinylmethyl)pyrrolidinium bromide 2d
2-Bromomethylquinoline hydrobromide (m.p. 190°C (dec.)) was prepared from quinaldine by NBS
bromination, extraction by dilute aq. HBr and crystallization from 2-propanol. The free base was very
sensitive towards tarring. It was liberated by dilute NaOH and extracted with CH2Cl2 directly before
reacting with N-methylprolinol (2 h boiling in MeCN). The tarry oil was crystallized from 2-propanol/t-
BuOMe to give a 52% yield, m.p. 142°C. [α]D26=−14.7 (c=0.9, MeCN). 1H NMR (500 MHz, CDCl3):
δ=2.24–2.27 (m, 2H), 2.45–2.47 (m, 2H), 3.48 (s, 3H), 3.51–3.54 (m, 1H), 4.27–4.30 (m, 2H), 4.42–4.66
(m, 2H), 4.69 (d, J=13.3 Hz, 1H), 5.47 (d, J=13.3 Hz, 1H), 7.62 (dd, J=7.3, 7.8 Hz, 1H), 7.77 (dd, J=7.2,
8.0 Hz, 1H), 7.86 (d, J=8.2 Hz, 1H), 8.04 (d, J=8.6 Hz, 1H), 8.06 (d, J=8.4 Hz, 1H), 8.27 (d, J=8.3 Hz,
1H). Anal. calcd for C16H21N2OBr (337.26): C, 56.98; H, 6.24; N, 8.31; found: C, 56.98; H, 6.54; N,
8.16.
3.4. (1S,2S)-2-Hydroxymethyl-N-methyl-N-(2-quinolinylmethyl)pyrrolidinium iodide 3d
Prepared similarly from (S)-N-(quinolinylmethyl)proline, 81% yield, m.p. 135°C. [α]D25=−16.2
1
(c=1.0, MeOH). H NMR (500 MHz, CDCl3): δ=2.01–2.04 (m, 1H), 2.15–2.17 (m, 1H), 2.22–2.26
(m, 1H), 2.37–2.38 (m, 1H), 3.37 (s, 3H), 3.73–3.77 (m, 1H), 4.17–4.19 (m, 1H), 4.23–4.29 (m, 2H),
4.52–4.57 (m, 1H), 5.16 (d, J=13.3 Hz, 1H), 5.47 (d, J=13.3 Hz), 7.61–7.64 (m, 1H), 7.75–7.79 (m, 1H),
7.87 (d, J=8.1Hz, 1H), 7.98 (d, J=8.4 Hz, 1H), 8.05 (d, J=8.4 Hz, 1H), 8.28 (d, J=8.3 Hz, 1H). 13C-NMR
(62 MHz, D3COD): δ=19.45 (C-4), 23.3 (C-3), 43.1 (CH3), 58.8 (CH2OH), 65.0 (C-5), 68.4 (CH2–Ar),
75.7 (C-2), 123.5 (arom. C-3), 127.60 (arom. C-5,6), 127.63 (arom. C-5,6), 127.8 (arom. C-4a), 128.9
(arom. C-7), 130.2 (arom. C-8), 137.7 (arom. C-4), 147.5 (arom. C-8a), 150.0 (arom. C-2). Anal. calcd
for C16H21IN2O (384.26): C, 50.01; H, 5.51; N, 7.29; found: C, 49.76; H, 5.39; N, 7.20.
3.5. (1S,2S)-N-Benzyl-2-hydroxymethyl-N-methylpyrrolidinium iodide 3a
Prepared from benzylprolinol with methyl iodide as before. 85% Yield; m.p. 112°C. [α]D26=−84.1
1
(c=1, EtOH). H NMR (250 MHz, CDCl3): δ=1.26–2.36 (m, 4H), 3.03 (s, 3H), 3.23–3.29 (m, 1H),
3.96–4.15 (m, 3H), 4.28–4.40 (m, 1H), 4.91 (d, J=12.9 Hz, 1H), 5.03 (d, J=12.9 Hz, 1H), 7.43–7.65 (m,
5H). Anal. calcd for C13H20INO (333.21): C, 46.85; H, 6.05; N, 4.20; found: C, 46.89; H, 5.98; N, 4.59.
3.6. (1S,2S)-2-Hydroxymethyl-N-methyl-N-(4-quinolinylmethyl)pyrrolidinium bromide 3e
Prepared similarly from (S)-N-methylprolinol and the extremely sensitive 4-bromomethylquinoline.
Yield: 38%, m.p. 141°C. [α]D26=−22.9 (c=0.8, MeCN). H NMR (500 MHz, CDCl3): δ=1.74–1.78
1
(m, 1H), 1.97–1.99 (m, 1H), 2.25–2.27 (m, 1H), 2.43–2.44 (m, 1H), 3.96 (s, 3H), 3.30–3.39 (m, 1H),
3.99–4.04 (m, 1H), 4.16–4.23 (m, 2H, CH2OH), 4.69–4.72 (m, 1H), 5.47 (d, J=13.5 Hz, 1H), 5.85 (d,
J=13.5 Hz, 1H), 7.76 (dd, J=7.6, 7.7 Hz, 1H), 7.81 (dd, J=7.2, 8.2 Hz, 1H), 7.91 (d, J=4.4 Hz, 1H), 8.17
(d, J=8.4 Hz, 1H), 8.43 (d, J=8.3 Hz, 1H), 8.98 (d, J=4.4 Hz, 1H). 13C NMR (CD3OD): δ=20.7 (C-4),
24.5 (C-3), 43.4 (CH3), 60.7 (CH2OH), 63.4 (C-5), 68.4 (CH2–Ar), 76.4 (C-2), 125.0 (arom. C-3), 128.1
(arom. C-6), 129.4 (arom. C-4a), 129.7 (arom. C-5), 130.8 (arom. C-7), 131.6 (arom. C-8), 136.2 (arom.
C-4), 149.7 (arom. C-8a), 151.1 (arom. C-2). Anal. calcd for C16H21BrN2O (337.26): C, 56.98; H, 6.28;
N, 8.31; found: C, 56.31; H, 6.33; N, 8.86.

2242
E. V. Dehmlow et al. / Tetrahedron: Asymmetry 9 (1998) 2235–2244
3.7. (1S,2S)-N-Benzyl-N-methyl-2-(1-naphthyloxymethyl)pyrrolidinium chloride 4
(S)-N-Benzylprolinol was dissolved together with molar amounts of 1-naphthol and PPh3 in THF and
treated at 0°C with a small excess of diethyl azodicarboxylate under nitrogen. POPh3 was filtered off,
and the product was chromatographed over silica gel with petroleum ether (b.p. 45–60°C):ethyl acetate
(4:1). The eluate was heated with excess MeI in MeCN for 24 h at 50°C, then passed in methanol over
an anion exchange column (chloride form). The resultant product of a honey-like consistance could not
1
be crystallized. H NMR (250 MHz, CDCl3): δ=2.21–2.65 (m, 4H), 3.16 (s, 3H), 3.42–3.60 (m, 1H),
3.72–3.96 (m, 1H), 4.12–4.31 (m, 1H), 4.70–4.85 (m, 2H), 5.20 (d, J=12.8 Hz, 1H), 5.37 (d, J=12.8 Hz,
1H), 7.40–7.99 (m, 12H). IR: µ=3058, 2973, 2937, 1461, 1240, 767 cm−1
.
3.8. (1S,2S)-N-Benzyl-2-N-methyl-2-(2-quinolyloxymethyl)pyrrolidinium chloride 5
(S)-N-Benzylprolinol was dissolved in THF and reacted with NaH. A molar amount of 2-
bromomethylquinoline dissolved in THF was dropped into the refluxing solution which was heated for 6
h. Workup gave a dark-brown oil which was purified by a fast passage through a silica gel column (ethyl
acetate as eluent). The product was treated with excess MeI in MeCN at 40°C for 24 h. Passage through
an anion exchange column in methanol furnished the chloride as an oil which could not be crystallized.
1H NMR (500 MHz, CDCl3): δ=1.91–2.01 (m, 1H), 2.02–2.13 (m, 1H), 2.22–2.33 (m, 1H), 2.39–2.52
(m, 1H), 3.18 (s, 3H), 3.49–3.54 (m, 1H), 4.10 (s, 2H), 4.26–4.30 (m, 1H), 4.70–4.73 (m, 1H), 4.80–5.00
(m, 2H), 5.15 (d, J=12.8 Hz, 1H), 5.23 (d, J=12.8 Hz, 1H), 7.40–7.59 (m, 5H), 7.67–8.24 (m, 6H). IR:
µ=3029, 3004, 2966, 2890, 1616, 1504, 1471, 1367, 1126, 829, 707 cm−1
.
3.9. (S)-1-(Diphenylmethyleneamino)-2-hydroxymethylpyrrolidine
Refluxing molar equivalents of (S)-1-amino-2-hydroxymethylpyrrolidine [b.p. 90°C/40 mbar (Kugel-
rohr), [α]D26=−51.1 (c=1.0, C6H6)] and benzophenone in toluene with a few drops of BF3 ether for 12
h under azeotropic removal of water gave the title compound. Normal workup was followed by chrom-
atography (silica gel, petroleum ether (b.p. 50–70°):ether 1:3) to give an oil; 61% yield. [α]D25=+434.5
(c=1.0, MeCN). High resolution MS for C18H20N2O calcd: 280.15755; found 280.15756. EI, 70 eV: (m/z
(%)): 280 (8), 249 (62), 180 (100), 165 (13), 77 (59). 1H NMR (CDCl3): δ 1.62–1.88 (m, 4H), 2.41–2.48
(m, 1H), 2.68–2.78 (m, 1H), 3.43–3.88 (m, 3H), 7.24–7.81 (m, 10H). 13C NMR: δ 23.2 (C-4), 25.3 (C-3),
26.9 (C-5), 54.5 (C-2), 66.5 (CH2OH), 67.0, 127.4, 128.0, 128.1, 128.19, 128.22, 129.6, 130.1, 137.7,
140.0, 148.7.
3.10. (1R,2S)-1-(Diphenylmethyleneamino)-2-hydroxymethyl-1-methylpyrrolidinium iodide, 7a
The respective hydrazone was refluxed for 3 days in excess methyl iodide. After removal of the reagent,
the residue was crystallized from acetonitrile. M.p. 161°C; 52% yield. [α]D25=+33.7 (c=1.0, MeCN).
1H NMR (500 MHz, CDCl3): δ 2.00–2.09 (m, 1H), 2.10–2.21 (m, 1H), 2.28–2.38 (m, 1H), 2.39–2.49
(m, 1H), 3.58 (s, 3H), 3.97–4.03 (m, 1H), 4.03–4.08 (m, 1H), 4.18–4.20 (m, 1H), 4.37–4.39 (m, 1H),
4.76–4.78 (m, 1H), 7.37–7.63 (m, 10H). 13C NMR: δ 21.5 (C-4), 25.9 (C-3), 54.8 (CH3), 60.9 (CH2OH),
70.6 (C-5), 86.2 (C-2), 129.2, 129.5, 129.8, 129.9, 130.4, 130.5, 132.0, 134.1, 134.4, 138.2, 175.2. Anal.
calcd for C19H23IN2O (422.3): C, 54.03; H, 5.49; N, 6.63; found: C, 54.00; H, 5.62; N, 6.55.
Reaction (1): 267 mg (1 mmol) of ethyl N-(diphenylmethylene)glycinate, 171 mg (1 mmol) of benzyl
bromide and 0.1 mmol of catalyst were dissolved in 10 ml of CH2Cl2 and treated with 200 mg (5 mmol)

E. V. Dehmlow et al. / Tetrahedron: Asymmetry 9 (1998) 2235–2244
2243
of powdered NaOH. The mixture was stirred at r.t. for 24 h. It was filtered and concentrated. The residue
was taken up in petroleum ether:t-BuOMe (5:1) and filtered again. The solution was analyzed by HPLC
for enantiomeric excess (hexane:2-propanol=500:1).
Reaction (2): 1.62 mg (1 mmol) of pivalophenone and 23 mg (0.6 mmol) of NaBH4 were stirred with
0.1 mmol of catalyst in a mixture of 3 ml of CH2Cl2 and 0.5 ml of H2O at 0°C for 3 h. Phases were
separated and the aqueous layer was extracted with more CH2Cl2. The combined organic phases were
dried (Na2SO4). The solvent was removed and the mixture was analyzed by HPLC on the chiral column
(hexane:2-propanol=200:1).
Reaction (3): 0.5 g (2.7 mmol) of potassium phthalimide was refluxed with 2.0 g (11.04 mmol) of
ethyl 2-bromopropionate in 20 ml of CH2Cl2 in the presence of 5 mol% catalyst for 48 h to give
95% conversion. Solvent and excess reagent were removed in vacuo, and the residue is purified by
passage over silica gel in CH2Cl2, m.p. 61°C. Enantioselectivity assays were performed by polarimetry14
showing only 2–5% e.e. When the reaction was performed in boiling THF with only a twofold excess of
bromoester and BCDC as a catalyst, a maximum e.e. of 18% at 55% conversion was found.
Reaction (4): 258 mg (1.36 mmol) of 2-methoxycarbonyl-1-indanone and 100 mg (1.36 mmol) of
butenone were stirred in toluene in the presence of 5% aqueous NaHCO3 and 5 mol% of catalyst at r.t. or
at −40°C for 24 h. Phases were separated, the aqueous layer was extracted with CH2Cl2. Solvents were
removed, and the residue was chromatographed (silica gel, tert-butyl methyl ether), m.p. 104°C, yield
90–95%. Enantiomeric composition was assayed by polarimetry15 to give a maximum e.e. of 8% with
catalyst BCDC. Shorter reaction times required stronger bases, but again no useful e.e.s were observed.
Reaction (5): 200 mg (1.18 mmol) of 2-ethoxycarbonyl-1-cyclohexanone and 87 mg (1.18 mmol) of
butenone were stirred in 5 ml of toluene in the presence of 165 mg (1.19 mmol) of solid K2CO3 and 5
mol% catalyst for 24 h at r.t. After the usual work-up, the residue was passed over silica gel in tert-butyl
methyl ether, then distilled, b.p. 50°C/0.2 mbar. Analysis was performed by polarimetry17 and gave a
maximum e.e. of 18% with QUIBEC.
Reaction (6): 271 mg (1 mmol) of N-(diphenylmethylene)benzylamine and 127 mg (1 mmol) of
dimethyl sulfate [or 109 mg (1 mmol) of ethyl bromide] were dissolved in 3 ml of dichloromethane
and stirred with a mixture of 0.5 g of powdered KOH and K2CO3 and 10 mol% of catalyst at r.t. for
24 h. 10 ml of H2O and 20 ml of petroleum ether were added, and the phases were separated. The
aqueous phase was washed twice with 10 ml of petroleum ether. Then the combined organic phases were
concentrated, and the residue was stirred with 10 ml of t-BuOMe and 10% aqueous HCl for 16 h at 70°C.
After cooling, the organic layer was washed twice with dilute HCl. The combined aqueous phases were
made alkaline by the addition of conc. NaOH and extracted three times with 10 ml of t-BuOMe. This
extract was dried and concentrated. It contained 1-phenethylamine [or 1-phenylpropylamine] and a little
unhydrolyzed alkylation product. Enantioselectivity assays were performed by g.c. on the chiral column
mentioned above (111°C, 1 bar of N2 pressure) and indicated no e.e.s in both cases.
Racemization tests: (a) 285 mg (1 mmol) of diphenylmethylene (R)-phenethylimine19 was stirred with
a mixture of 10 ml 10% aq. HCl and 10 ml of t-BuOMe for 2 days at r.t. or for 24 h at 70°C. Workup as
described for reaction (6) furnished the amine which was subjected to g.c. on the chiral column (111°C,
1 bar pressure of N2). Only the (R) enantiomer was present in both cases.
(b) A quantity (285 mg,1 mmol) of diphenylmethylene (R)-phenethylimine {[α]D25=+11 (c=0.18,
MeCN)} was stirred with 1.38 g of solid K2CO3 and 0.56 g of powdered KOH in 3 ml of CH2Cl2 for 24
h at r.t. In a parallel run 7 mg (0.02 mmol) of cetyltrimethylammonium bromide were present. Workup
and reisolation furnished products with [α]D24=+6 (c=0.16, MeCN) and [α]D24=+4 (c=0.12, MeCN)
respectively.

2244
E. V. Dehmlow et al. / Tetrahedron: Asymmetry 9 (1998) 2235–2244
Acknowledgements
This work was supported, in part, by Deutsche Forschungsgemeinschaft (project De 173/20-1) and by
the Fonds der Chemischen Industrie. Special thanks are due to Professor P. Jutzi of this faculty who made
possible the X-ray analyses.
References
1. Applications of Phase Transfer Catalysis, Part 67. Part 66: Schrader, S.; Dehmlow, E. V. Org. Prep. Proc. Int., 1996, 28,
634–637.
2. Eddine, J. J.; Cherqaoui, M. Tetrahedron: Asymmetry, 1995, 6, 1225–1228.
3. PTC Symposium at Pacifichem 95, Honolulu, Hawaii, Dec. 17–22, 1995; International Conference “PTC ‘97” Nagoya,
Japan, Sept. 24–27, 1997.
4. (a) Dehmlow, E. V. and S. S. ‘Phase Transfer Catalysis’, 3rd Ed., VCH, Weinheim, 1993; (b) Goldberg, Y. ‘Phase Transfer
Catalysis: Selected Problems and Application’, Gordon & Breach Science Publ., Reading, 1992; (c) Starks, Ch.; Liotta,
Ch.; Halpern, M. ‘Phase Transfer Catalysis: Fundamentals, Applications, and Industrial Perspectives’, Chapman & Hall,
New York, 1994; (d) Shioiri, T. ‘Chiral Phase Transfer Catalysis’, Chapter 14 in: Handbook of Phase Transfer Catalysis,
Sasson, Y., Neumann, R. (eds), Blackie Academic & Professional, London, 1997.
5. Dehmlow, E. V.; Schrader, S. Pol. J. Chem. 1994, 68, 2199–2208.
6. X-Ray crystallographic analyses were performed with Siemens P2(1) difractometer. Scantype: Omega-scan; program:
Siemens SHELXTL plus/SHELX-93; direct structure resolution. 3c: C17H22 INO (383.26); crystal size 0.70×0.50×0.50
mm; temp. 293(2) K; wave length 0.71073 Å; space group orthorhombic P2(1)2(1)2(1); unit cell dimensions: a=9.074(3)
Å, α=90°, b=12.059(3) Å, β=90°, c=15.606(4) Å, γ=90°; volume 1707.7(8) Å3; Z=4; calculated density 1.491 mg
m
−3; absorption coefficient 1.872 mm−1; 5656 reflections collected, 4669 of these independent; goodness of fit on
F2: 1.021. Rf=0.048; 183 parameters. 7a: C19H23IN2O (422.29); crystal size 0.7×0.5×0.4 mm; temp. 173(2) K; wave
length 0.71073 Å; space group orthorhombic P2(1)2(1)2(1); unit cell dimensions: a=10.602(2) Å, α=90°, b=12.962(2)
Å, β=90°, c=13.527(3) Å, γ=90°; volume 1858.9(6) Å3; Z=4, calculated density 1.509 mg m−3; absorption coefficient
1.729 mm−1; 6084 reflections collected, 5370 of these independent; goodness of fit on F2: 1.036. Rf=0.043; 234
parameters. Crystallographic data (without structural factors) will be deposited as supplementary material publications
nos. CCDC xxxx and yyyy with the Cambridge Crystallographic Data Centre. Copies of the data maybe obtained free of
charge at the following address: CCDC, 12 Union Road, Cambridge CB21 EZ, GB, (fax: (+44) 1223-336-033; e-mail:
deposit@ccdc.cam.ac.uk).
7. O’Donnell, M. J.; Esikova, I. A.; Mi, A.; Shullenberger, D. F.; Wu, S.; Chapter 10, p. 125–135: Phase Transfer Catalysis;
Mechanisms and Syntheses, M. E. Halpern, ed., A.C.S. Symposium Series #659, Washington, D.C., 1997.
8. Dehmlow, E. V.; Knufinke, V. Liebigs Ann. Chem. 1992, 283–285.
9. Colonna, S.; Fornasier, R. J. Chem. Soc., Perkin Trans. 1 1978, 371–373.
10. Juliá, S.; Ginebreda, A.; Guixer, J.; Masana, J.; Tomás, A.; Colonna, S. J. Chem. Soc., Perkin Trans. 1 1981, 574–577.
11. Hodge, P.; Khoshdel, E.; Waterhouse, J. J. Chem. Soc., Perkin Trans. 1 1983, 2205–2209.
12. A 99% e.e. (based on max. rotation) was reported for reaction (4) in a homogeneous, non-PTC process with a preformed
complex of a chiral crown ether and KOt-Bu: Cram, D. J.; Sogah, G. D. Y. J. Chem. Soc., Chem. Commun. 1981, 625–628.
13. Hermann, K.; Wynberg, H.; J. Org. Chem. 1979, 44, 2238–2244.
14. Corey, E. J.; Enders, D. Chem. Ber. 1978, 111, 1337–1361.
15. Enders, D.; Eichenauer, H. Angew. Chem. 1979, 91, 425–427; Angew. Chem. int. Ed. Engl. 1979, 18, 397.
16. Avaro, M.; Lewisalles, J.; Rudler, H. J. Chem. Soc., Chem. Commun. 1969, 445–446.
17. Smith, P. A. S.; Tan, H. H. J. Org. Chem. 1967, 32, 2586–2591 and previous papers.
18. Hsü, S. K.; Ingold, C. K.; Wilson, C. L.; J. Chem. Soc. 1935, 1778–1785.
19. Ingold, C. K.; Wilson, C. L.; J. Chem. Soc. 1933, 1493–1505.
20. Dehmlow, E. V.; Lissel, M.; Heider, J. Tetrahedron, 1977, 33, 363–366.
21. Dehmlow, E. V.; Singh, P.; Heider, J. J. Chem. Res. (S), 1981, 292–293.
22. Dehmlow, E. V.; Sleegers, A. J. Org. Chem. 1988, 53, 3875–3877.