Synthesis of Quinolines via Pd/C-Catalyzed Cyclization of 2-Aminobenzyl Alcohol with Ketones

Abstract

It is known that quinoline plays an important role as a basic skeleton for the design of many pharmacologically active compounds such as antiasthmatic, anti-inflammatory and antimalarial. During the course of our studies directed towards C-N bond activation, we have reported the ruthenium-catalyzed synthesis of quinolines via an alkyl or alkanol group transfer from alkylamines or alkanolamines to N-atom of anilines (amine exchange reaction), followed by cascade isomerization and cyclization of 3-(2-aminophenyl)1-arylprop-2-yn-1-ols. In connection with this report, several routes for the coupling of carbonyl compounds and alcohols have recently been reported as exemplified in Scheme 1. The coupling of ketones A with primary alcohols B preferentially afforded the coupled ketones C (Scheme 1, route a) or the coupled secondary alcohols D (Scheme 1, route b) which depend on the molar ratio of B to A. In addition, secondary alcohols E was also found to be coupled with B to give D (Scheme 1, route c). These reactions could be applied to modified Friedlander quinoline synthesis via ruthenium-catalyzed consecutive coupling and cyclization of 2-aminobenzyl alcohol with ketones and secondary alcohols, which is superior to conventional Friedlander method in a sense of price and stability of 2aminobenzyl alcohol. Under these circumstances, this report describes an alternative palladium-catalyzed route for Friedlander quinoline synthesis. Table 1 shows several attempted results for the oxidative coupling and cyclization of 2-aminobenzyl alcohol (1) with acetophenone (2a). Generally, treatment of 1 with 2a in dioxane in the presence of a catalytic amount of 5% Pd/C (0.5 mol%) and KOH at 100 C afforded 2-phenylquinoline (3a) with concomitant formation of direct transfer hydrogenation product, 1-phenylethanol (4). The yield of 3a increased with the reaction time up to 20 h (runs 1-4) and the amount of KOH employed (runs 4-6). As has been observed in our recent ruthenium-catalyzed version, the molar ratio of 2a to 1 also affected the yield of 3a, higher molar ratio up to [2a]/[1] = 2 resulting in the effective formation of 3a (runs 4 and 7). This could be due to the acceleration of the initial oxidation of 1 to 2-aminobenzaldehyde by transfer hydrogenation from 1 to excess 2a. However, performing the reaction in the presence of 1-decene as a hydrogen acceptor instead of excess 2a gave no significant change on the product yield and distribution (run 8). Having established reaction conditions, various ketones 2 were subjected to react with 1 in order to investigate the reaction scope and several representative results are summarized in Table 2. From the reactions between 1 and aryl(methyl) ketones (2a-2h), the corresponding 2-arylquinolines (3a-3h) were produced in the range of 43-77% yields. Here again, the conventional transfer hydrogenated aryl(methyl) carbinols were produced in considerable amounts on GLC analysis. The position and electronic nature of the substituent on the aromatic ring of aryl(metnyl) ketones had no relevance to quinoline yield. The reaction proceeds likewise with heteroaryl(methyl) ketone 2i and 2'-