Birch Reduction - Reaction Regioselectivity

Reaction Regioselectivity

Birch Reduction has several intricate mechanistic features. These features govern the reaction’s regioselectivity and are considered below. Birch’s rule for aromatics with electron donors such as methoxyl or alkyl is that the product will have the residual double bonds bearing the maximum number of substituents. For aromatics with electron withdrawing groups such as carboxyl, the substituent groups avoid the double bonds. In both cases, with electron donating and with withdrawing groups, the residual double bonds are unconjugated (vide infra). It has been a matter of intense interest to understand reaction mechanisms accounting for this regioselectivity. The essential features are:

• In liquid ammonia alkali metals dissolve to give a blue solution thought of simplistically as having “free electrons”. The electrons are taken up by the aromatic ring, one at a time. Once the first electron has been absorbed, a radical-anion has been formed. Next the alcohol molecule donates its hydroxylic hydrogen to form a new C-H bond; at this point a radical has been formed. This is followed by the second electron being picked up to give a carbanion of the cyclohexadienyl type (i.e. with C=C-C-C=C in a six-ring and charged minus ). Then this cyclohexadienyl anion is protonated by the alcohol present. The protonation takes place in the middle of the cyclohexadienyl system. This (regio-)selectivity is unique and characteristic.

• Where the radical-anion is protonated initially determines the structure of the product. With an electron donor as methoxy (MeO) or alkyl protonation has been thought by some investigators as being ortho (i.e. adjacent or 1,2) to the substituent. Other investigators have thought the protonation is meta (1,3) to the substituent. Arthur Birch favored meta protonation. With electron withdrawing substituents protonation has been thought to come at the site (ipso) of the substituent or para (1,4). Again, there has been varied opinion. A. J. Birch’s empirical rules say that for the donor substituents the final product has the maximum number of substituents on the final double bonds. For electron withdrawing groups the double bonds of the product have avoided the substituents. The placement preference of groups in the mechanism and in the final product is termed regioselectivity.

• The reaction mechanism provides the details of molecular change as a reaction proceeds. In the case of donating groups A. J. Birch's preference for meta protonation of the radical anion was based on qualitative reasoning. And it had been noted that no experimental test of this was known.

• In 1961 a simple computation of the electron densities of the radical anion revealed that it was the ortho site which was most negative and thus most likely to protonate. However, A. J. Birch seemed to overlook this result. Additionally, the second proton had been determined by the computations to occur in the center of the cyclohexadienyl anion to give an unconjugated product.

• Of historical interest is the uncertainty in the chemical literature at this point. Indeed, there were some further computational results reported. These varied from suggesting a preference for meta radical-anion protonation to suggesting a mixture of ortho and meta protonation.

• In 1990 and 1993 an esoteric test was devised which showed that ortho protonation of the radical anion was preferred over meta (seven to one). This was accompanied by more modern computation which concurred. Both experiment and computations were in agreement with the early 1961 computations.

• With electron withdrawing groups there are literature examples demonstrating the nature of the carbanion just before final protonation. This revealed that the initial radical-anion protonation occurs para to the withdrawing substituent.

• The remaining item for discussion is the final protonation of the cyclohexadienyl anion. In 1961 it was found that simple Hückel computations were unable to distinguish between the different protonation sites. However, when the computations were modified with somewhat more realistic assumptions, the Hückel computations revealed the center carbon to the preferred. The more modern 1990 and 1993 computations were in agreement.