J. Org. Chem. 1993,58, 5101-5106 5101 Direct Substitution of Aromatic Ethers by Lithium Amides. A New Aromatic Amination Reaction Wolter ten Hoeve,+Chris G. Kruse) Jan M. Luteyn) Janet R. G. Thiecke: and Hans Wynberg'J Syncom B. V., Nyenborgh 4,9747 AG Groningen, The Netherlands, and Solvay Duphar Research Laboratories, PO Box 900,1380 DA Weesp, The Netherlands Received May 25,1993 Reaction of lithiated dialkylamines with methoxy aromatics in refluxing THF leads to products resulting from a direct ipso-substitution. Especially with lithiated secondary amineshigh conversions and selectivities are achieved. Sulfonyl-substituted aromatics react equally well, but halogenated aromatics give rise to side-products arising from a competing pathway via aryne intermediates. The scope and mechanistic implications of this novel nucleophilic amination reaction are described. The introduction of an amino function into aromatic methoxy aromaticss in a completely regioselective fashion compounds by way of nucleophilic substitution is usually and in a good yield. Hence, the substitution of methoxy carried out with halogenated ar0matics.l Such substitu- groups by amines is neither limited to methoxy aromatics tions are regioselective only in the special case that an having activating groups' or an o-oxazolinyl group: nor electron-withdrawinggroup is present in the ortho or para does it have to be performed in the carcinogenic HMPA.' position. The regioselectivity is lost when a benzyne As an example, heating a mixture of veratrole and 1.25 intermediate is involved. Such is the case in many of the equiv of N-lithio-N'-methylpiperazine in THF for 5 h substitutions of halogenated aromatics.' furnishes an 85 7% yield of the direct substitution product In a few scattered reports the nucleophilic substitution 2 as the only positional isomer. Some starting material of unactivated methoxy aromatics has been reported. is also present as well as a small amount of dealkylation Benkeser and DeBoer reported in 1956 that treatment of product (Scheme I). anisole with lithium dimethylamide resulted in the for- Since we expect this direct substitution to become of mation of NJV-dimethylanilinein very low yield.2 The major importance in organic synthesis, we have made an use of sodium amide in boiling piperidine for the dealky- investigation of the scope of this transformation.9 Because lation of aromatic ethers was advocated by Brotherton of the pharmaceutical importance of arylpiperazines, we and Bunnett in 1957. N-Phenylpiperidinewas isolated in have looked especially at this class of compounds which very low yield (along with a high yield of phenol) from the hitherto has usually been prepared from an aniline and reaction with n-butoxybenzene.3 In 1973 Cuvigny and the highly toxic N,N-bis(2-chloroethyl)amines.9 Normant described the reaction of lithium amides with Aromatic Ethers. A number of aromatic ethers has aromatic ethers in HMPA. In addition to dealkylation, been subjected to our amination conditions. The results resulting in phenols, dealkoxylation, providing aromatic of the reaction with 1.1 equiv of N-lithio-N'-methylpip- amines, also occurred.' Acceptable yields were obtained erazine under reflux in THF (generally overnight) are onlywith lithium dimethylamide. Finally, in 1977 Meyers shown in Table I. From this table it can be concluded and Gabel showed that methoxy aromatics having an that our amination reaction is applicable to a wide range 0-oxazolinyl moiety, when treated with lithium amides, of aromatic ethers and that the amination reaction is underwent substitution of the methoxy group, thus completely regioselective. In the case of dialkoxy or affording aromatic amineshaving an o-oxazolinyl m~iety.~ trialkoxy compounds small amounts of disubstitution This latter reaction may, however, be considered an product can usually be detected by GC. Other common example of an activated methoxy aromatic, due to the side-products in the amination reaction are the corre- presence of the electron-withdrawing oxazolinyl moiety sponding phenol, resulting from dealkylation under the (when a second methoxy group is present in the meta reaction conditions, and varying amounts of starting position only the o-methoxy group is substituted). Re- materials. This dealkylation is more prominent in the cently, in two papers the regioselective substitution of a case of ethoxybenzene (entry 7; although with lithium methoxy group for an alkyl group in very specific aromatic piperidide a 35% yield of substitution product was ethers by reaction of the methoxy compound with alkyl- obtained), a feature already cited by Cuvigny and Nor- lithiums has been describede6e7 We have now discovered that lithium amides in THF With 1,2,3-trimethoxybenzene (entry 4; the same holds can efficiently substitute methoxy groups in unactivated for 1,2,3-tris(benzyloxy)benzene)it appears that only the outermostmethoxy group is substituted. This observation f Syncom B.V. t Solvay Duphar Reeearch Laboratories. is in contrast to the reactions with theoxazolines of Meyers, (1) March, J. Advanced Organic Chemistry,3rd ed.; Wiley: New York, where only the inner methoxy group is substituted in 2,3- 1985; Chapter 13 and refs cited therein. dimethoxy-l-oxazolinylbenzene.5 This supports our (2) Benkeeer, R. A.; DeBoer, C. E. J. Org. Chem. 1956,21, 366. as- (3) Brotherton, T. K.; Bunnett, J. F. Chem. Znd. (London) 1957,80. (4) Cuvigny, Th.; Normant, H. J. Orgonomet. Chem. 1973,55,41. (8) In an attempt to cause lithiated veratrole to react with meth- (6)Meyers, A. I.; Gabel, R. J. Org. Chem. 1977,42,2663. oxyamine, the methoxyamineWBB generated from ita hydrochloride with (6) Bublaer, P. H. M.; van Doom, J. A. Rec. Trau. Chim. Pays-Bas KOH in DMF. The gaseous producta (apparently containing some 1990,109,443. dimethylamiie) were led through the solutionof the lithioveratrole;from (7) Mateumoto, T.; Kakigi,H.; Suzuki, K. TetrahedronLett. 1991,32, the reaction mixture we isolated 2-methoxy-NjV-dimethylaniie. 4337. (9) Eur. Pat. 92200468 (to Solvay-Duphar,Weesp, The Netherlanda). 0022-3263/93/195&5101$04.00/0 0 1993 American Chemical Society 5102 J. Org. Chem., Vol. 58, No. 19, 1993 ten Hoeve et al. Scheme I dimethoxybenzoicacid, respectively, into the known 2,3- dimethoxyanilineand 2,6-dimethoxyaniline,respectively, n THF aOCH3+ Li-N-N-CH3 through a Curtius reaction.’OJl These anilines were then OCH3 transformed in a standard fashion to the corresponding piperazines using N,N-bis-(2-chloroethyl)methylamine. The piperazine obtained from 2,3-dimethoxybenzoicacid proved to be identical to the piperazine obtained from the reaction of 1,2,3-trimethoxybenzene with N-lithio-N‘- methylpiperazine. Table I. Reaction of Aromatic Ethers with 1.1 Equiv of It can also be seen from the Table I that our amination N-Lithio-N’-methylpiprazine reaction is not limited to methoxy aromatics but that some other aromatic ethers, namely phenyl ethers and benzyl entry aromatic ether product yield ( 5% ) ethers, also undergo the substitution reaction (entries 8 1 44 and 9). The fact that (benzy1oxy)benzene is a very good substrate is particularly surprising because this is in 1 contrast to the reaction of lithium dimethylamide with 2 pCH3 75 (benzy1oxy)benzenein HMPA where a Wittig rearrange- ment takes place resulting in diphenylmethan~l.~ Amines. Veratrole has been subjected to the action of a variety of lithium amides (1.1 equiv of amide in THF 3 73 under reflux). The results are shown in Table 11. In general, secondary amines give a good yield of the substitution product, the exception being the less-nu- cleophilic lithium amide from N-methylaniline (entry 16). 4 65 Primary amines are less suitable candidates for the amination reaction; for instance butylamine gives a 28% yield of the N-butylaniline 22 (entry 18). In this case there is also present a small amount of N-methylated 22 as judged 5 69 from the NMR spectrum of the product. Apparently, N-butyl-N-lithioamine gives some demethylation, the resulting N-butylmethylamine then undergoes lithiation and substitutes the methoxy group (some N-methylated material was also observed in the piperazine case (entry 6 68 4) and the homopiperazine case (entry 10)). Hindered secondary amines such as lithium diisopro- pylamide and lithium tetramethylpiperidide (entries 7 and 6 14) are also unreactive under the standard conditions. 7 20 Actually, the common use of lithium diisopropylamide in deprotonation reactions may have prevented a prior 1 discovery of the nucleophilic amination of alkoxy aro- 8 63 matics. This hindrance is used to advantage in the preparation of the piperazines 14 and 15 (entries 5 and 6), which are otherwise difficultly accessible. 9 70 In an intramolecularvariant of our amination,N-benzyl- N-lithio-2-(2-methoxyphenyl)ethylamine23 is smoothly cyclized to the dihydroindole24 (Scheme11). The primary analog 2-(2-methoxyphenyl)ethylaminedid not undergo 10 86 the intramolecular substitution under the standard con- ditions. 7 Another intramolecular reaction takes place when the 11 22 amine 25 (or substituted with a methoxy group, compound 26) shown in Scheme I11 is treated with n-butyllithium. The product 27 obtained in 20 76 yield (compound 28 where 12 50 the methoxy group is present is obtained in 50 % yield) is similar to the result of a Smiles rearrangement, yet the precursor lacks electron-withdrawing groups.12 Access to 1,2-diaminobenzenederivatives is also possible 13 0 as shown in Scheme IV. Treatment of the reaction product 2 of veratrole and N-lithio-N’-methylpiperazinewith 1.1 equiv of the same lithium amide results in substitution of the second methoxy group although the yield of 29 is only sumption that a mechanism different from the one in our amination reaction is operative in the oxazoline case. Proof (IO) Smith, P. A. S. Org. React. 1946, 3, 337. of the fact that only the outermost methoxy group is (11) Haeaner, A.; Munger, P.; Belinka, B.