An α-aminomethyl carbanion equivalent via a novel Barbier reaction: (1H-naphtho[ 1,8-de]-1,2,3-triazin-2-yl)methyl anion

Article abstract of DOI:10.1016/S0040-4039(00)00689-4

A novel sonication-promoted Barbier reaction putatively generated the titled species from the corresponding naphthotriazinylmethyl chloride and magnesium in THF: its formal addition to a variety of carbonyl compounds in situ occurred in excellent yields. Subsequent catalytic hydrogenolysis of the triazine moiety demasked the amine, thus defining a route to various phenylethylamines (including the alkaloid 'mescaline'), or ethanolamines (in two cases), in excellent overall yields. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0040-4039(00)00689-4

Tetrahedron Letters 41 (2000) 4685±4688
An a-aminomethyl carbanion equivalent via a novel Barbier
reaction: (1H-naphtho[1,8-de]-1,2,3-triazin-2-yl)methyl anion
Sosale Chandrasekhar* and Malayalam Sridhar
Department of Organic Chemistry, Indian Institute of Science, Bangalore 560 012, India
Received 17 April 2000; accepted 27 April 2000
Abstract
A novel sonication-promoted Barbier reaction putatively generated the titled species from the corre-
sponding naphthotriazinylmethyl chloride and magnesium in THF: its formal addition to a variety of
carbonyl compounds in situ occurred in excellent yields. Subsequent catalytic hydrogenolysis of the triazine
moiety demasked the amine, thus de®ning a route to various phenylethylamines (including the alkaloid
`mescaline'), or ethanolamines (in two cases), in excellent overall yields. # 2000 Elsevier Science Ltd. All
rights reserved.
Keywords: alkaloids; a-aminomethyl carbanion; Barbier reaction; mescaline; naphthotriazine; phenylethylamine; ultra-
sound.
We report herein a novel, ultrasound-promoted Barbier reaction1 3 and demonstrate its
synthetic utility. The reaction results in the formal addition of a masked a-(primary-amino)-
methyl carbanion to a variety of carbonyl compounds, the amine function then being released
hydrogenolytically. The two-step sequence is relatively facile and occurs in excellent overall yields.
The synthetic potential of the process derives from the fact that the generation of an a-amino
carbanion is a topic of current interest,4 and that the above sequence de®nes a novel route to the
phenylethylamine group of alkaloids.5
In the course of other studies with 1H-naphtho[1,8-de]-1,2,3-triazine 1, we were drawn to the
possibility of generating the ylidic species 3 via the deprotonation of the 2-methyl derivative 2a
(Scheme 1). Interest in 3 derives from the fact that its alkylationÐand the subsequent reductive
cleavage of the triazine moietyÐwould de®ne a novel synthetic route to a-substituted amines: 3
would therefore serve as an a-aminomethyl carbanion equivalent.
In fact, early work by Perkins6a had not only established that methylation of 1 yields substantial
amounts of 2a (alongside the N1-isomer), but also that 2a condenses with benzaldehyde in the
presence of ethanolic sodium ethoxide: although the resulting N2-styryl derivative was obtained in
* Corresponding author. Fax: (+91-80) 360 0529; e-mail: sosale@orgchem.iisc.ernet.in
0040-4039/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(00)00689-4

4686
Scheme 1.
low yields, the acidity of the N-methyl group in 2a had been proven. (The above results6a on the
methylation of 1 were extended to include a variety of alkylating agents by Tavs and co-workers,6b
but since then no major development in the above chemistry of 1 has apparently been reported.6c)
In fact, the triaza-ylidic moiety in 3 is isoelectronic with the nitromethane anion; thus, the
analogy between the above reaction and the classical Henry `nitroaldol' reaction7 becomes
apparent. (Indeed, the Henry reactionÐfollowed by the reduction of the nitro groupÐde®ned
one of the earliest known aminomethyl carbanion equivalents.4b) Clearly then, further elabora-
tion of the above results of Perkins,6a particularly in aprotic media, was of interest. However,
sodium hydride-mediated deprotonation of 2a in THF led to the ethyl derivative 2b, apparently
via methyl transfer from unreacted 2a to 3 (which is apparently faster than the deprotonation);
similar results were obtained with lithium diisopropylamide in THF.
We then explored the possibility of preparing the 2-chloromethyl derivative 4 in the hope of
e€ecting a chlorine±lithium exchange with n-butyllithium. Treatment of the triazine 1 with
bromochloromethane in dimethylformamide in the presence of potassium carbonate over 2 days
at 25ꢀC a€orded 4 in 63% yield. (Interestingly, none of the N1-isomer was observed, although a
bis-N2-alkylation productÐderived from 4 and 1Ðwas isolated in 7% yield.) However, 4 was

4687
unreactive towards n-butyllithium under a variety of conditions (^78±0ꢀC/THF). Although 4 was
also reluctant to form the corresponding Grignard reagent, it was found possible to e€ect the
Barbier reaction1 3 with it under the in¯uence of ultrasound.8 Thus, when a mixture of 4, magnesium
and benzaldehyde 6a in THF was sonicated for 1 h, the naphthotriazinylbenzyl alcohol 7a was
isolated in 87% yield. The wide extent of the reaction was demonstrated with a number of aromatic
aldehydes 6a±6d bearing electron-donating substituents, with valeraldehyde 6e and with aceto-
phenone 6f, to a€ord the corresponding alcohols 7 in excellent yields (Table 1). The reactionÐ
which occurs via the putative Grignard reagent 5Ðfailed, however, with 3-methoxy- and 3,4-
dimethoxybenzaldehydeÐas it also generally did in the absence of the sonication.
The catalytic hydrogenolyses of the alcohols 7 were then investigated, and found to be facile at
normal temperature and pressure, generally in the presence of 10% Pd±C. Both the triazine
and the benzylic alcohol moieties (in the aromatic cases) were found to be cleaved to a€ord the
corresponding arylethylamine products 8 in excellent yields (Table 1), together with the expected
1,8-diaminonaphthalene by-product.6a However, the triazine moiety could be selectively cleaved
in two cases, 7c (with Raney nickel) and 7e, to furnish the corresponding amino alcohols 8c and
8e, also in high yields (Table 1). The amines 8 possess the skeleton of the phenylethylamine group
of alkaloids, 8d being well known5 as `mescaline'.
Table 1
Percent yields for the formation of the benzyl alcohols 7 and the phenylethylamines 8
The reasons for the above lack of reactivity of the 3-methoxy- and 3,4-dimethoxy-
benzaldehydes are not clear, particularly in view of the normal reactivity of 4-methoxy- and 3,4,5-
trimethoxybenzaldehydes. The Barbier reaction is believed3 to occur via a complex mechanism
involving the radical anions of the reacting partners which are formed at the metal surface, and it is
possible that the redox potential of the carbonyl compound exerts a subtle in¯uence on its reactivity.
It would appear that a 3-methoxy group destabilisesÐand hence suppresses the formation ofÐ
the ketyl radical anion derived from the aromatic aldehyde; however, in the case of 3,4,5-tri-
methoxybenzaldehyde, the relatively low concentration of the derived ketyl species may possibly be
countered by a relatively large increase in its further reactivityÐboth being presumed consequences
of the above destabilising e€ect.
In summary, a novel Barbier reaction has been e€ected, which de®nes a new a-(primary-amino)-
methyl carbanion equivalent, and leads eciently to the naturally occurring phenylethylamine

4688
skeleton: either the 2-arylethylamines or the 2-arylethanolamines may thus be selectively accessed.
Further studies are planned.
Typical procedures: Alcohols 7: the chloromethyltriazine 4, magnesium and the carbonyl com-
pound 6 (0.5 mmol each) in dry THF (2 ml) were treated with a trace of iodine, and the mixture
sonicated for 1 h (on a `Julabo USR 3' instrument at 35 kHz). Water was added to the mixture
and this was extracted with ether. The extracts were washed with water, dried (Na2SO4) and the
solvent evaporated to obtain the crude product, which upon chromatography a€orded the pure 7.
Phenylethylamines 8: the alcohols 7 (0.3 mmol) in ethanol (3 ml) were hydrogenated with 10%
Pd±C (freshly prepared W-2 grade Raney-Ni in the case of 7c) at NTP for 24 h. The catalyst was
®ltered o€, the solvent evaporated, and the resulting crude material chromatographed to furnish
the pure amines 8. All products were characterised by IR, NMR (1H and 13C) and mass spectra,
and physical constants or elemental analyses as appropriate.
Acknowledgements
CSIR and UGC (New Delhi) are thanked for their generous ®nancial support.
References
1. March, J. Advanced Organic Chemistry; John Wiley & Sons: New York, 1992; p. 921.
2. Li, C.-J. Tetrahedron 1996, 52, 5643±5668, and references cited therein.
3. Molle, G.; Bauer, P. J. Am. Chem. Soc. 1982, 104, 3481±3487.
4. (a) Katritzky, A. R.; Qi, M. Tetrahedron 1998, 54, 2647±2668, and references cited therein. (b) Saavedra, J. E. In
Umpoled Synthons; Hase, T. A., Ed.; John Wiley & Sons: New York, 1987; pp. 101±121, and references cited
therein. (c) Saavedra, J. E. J. Org. Chem. 1983, 48, 2388±2392.
5. Finar, I. L. Organic Chemistry; Addison Wesley Longman: Harlow, UK, 1997; Vol. 2, pp. 702±708.
6. (a) Perkins, M. J. J. Chem. Soc. 1964, 3005±3008. (b) Tavs, P.; Sieper, H.; Beecken, H. Liebigs Ann. Chem. 1967,
704, 150±160. (c) Neuenhoe€er, H. In Comprehensive Heterocyclic Chemistry; Katritzky, A. R.; Rees, C. W.;
Boulton, A. J.; McKillop, A., Eds.; Pergamon: Oxford, 1984; Vol. 3, pp. 369±384.
7. Rosini, G. In Comprehensive Organic Synthesis; Trost, B. M.; Fleming, I.; Heathcock, C. H., Eds.; Pergamon:
Oxford, 1991; Vol. 2, pp. 321±340.
8. Ref. 1, pp. 364±365, and references cited therein.