Extractive and Phase Switching Hydrolysis in Chemical Process Development.

Phenols may be separated from neutral substances by liquid/liquid extraction with aq. base if the phenol's molecular weight is not too high.  This is not a guaranteed success because phenols are only weak acids and the alkali phenolate, particularly as its molecular weight increases, may simply be water-insoluble. Because the free phenol in this situation is lipophilic, the phenolate in the presence of both water and an organic phase may substantially hydrolyze back to sodium hydroxide and the free phenol. The portion that remains neutral phenol 'happily jumps' into the organic layer.  For example, when a 10 ml.  of 0.01 moles of 2,4-dimethylphenol in water  is treated with an equivalent of alkali  and is then shaken with 20 ml of ethyl ether for about 10 minutes, the amount of the phenol found in the ether is 43% and the water is strongly basic.  The amount extracted depends upon the ratio of alkali to phenol, the ratio of the phases, and the particular organic solvent used. In the case of 2-isopropyl-5-methyl-phenol (thymol) the amounts extracted by different solvents under the above conditions are: ether 88; benzene 38; carbon tetrachloride 25; and pet. ether 22 percent.

In an extreme case. di-ortho substituted phenols have steric hindrance to the solvation shell that is needed around the oxygen anion, and this makes anion formation energetically disfavoured. Consequently, di-ortho phenols, even when their molecular weight is rather low will not dissolve in aqueous sodium hydroxide.  For that reason, such species are called cryptophenols. The name arose in the days before spectroscopic testing because these phenols did not give the characteristic qualitative test for a phenol. Cryptophenols can be dissolved in methanolic-KOH, a liquid medium called Claisen’s alkali. KiloMentor has written a separate article about Claisen’s Alkali.

Phase Switching Hydrolysis

In some situations, another trick can be employed to separate properly functionalized weak phenols or cryptophenols from a non-phenols.  Suppose for example you are trying to separate two carboxylic acid esters that differ only because one also contains a free phenol while the other, say, contains instead a phenol alkyl ether.  If one places the compound mixture in a two-phase solution of toluene and water, adds sodium hydroxide to the water and stirs the phases gently, after some time the phenolic ester will be hydrolyzed to the carboxylate salt and transferred to the aqueous base phase but the ether-ester will remain untouched in the toluene phase. What has happened is that partial phenolate formation has pulled its intramolecular ester functionality into more contact with the aqueous alkali than the substrate with an ether group.  Once in the aqueous alkaline layer, the phenolic ester substance is quickly hydrolyzed. Subsequently, in the form of the sodium carboxylate, it is stuck quantitatively in the water.  The ether-ester, on the other hand, is comparatively insoluble in the water. It cannot “see” the alkali because the stirring is gentle and there is little interface so it remains unreacted in the toluene. Conditions for the separation can be optimized by adjusting the organic solvent, the stirring rate, and the temperature of the two-phase mixture.

Although I have not tried the method with any combinations other than phenol-esters and ether-esters, other functional groups might be useful to replace the phenol by creating this initial small water solubility. Perhaps thiol, primary and secondary sulfonamide, imide, terminal acetylene, alpha unsubstituted alkyl nitro or dithiane might work. Any compound that can act as a weak acid in aqueous alkali has a good chance to make an intramolecular ester selectively hydrolyzable.

Separation of Diethyl malonate and Diethyl methylmalonate

In Organic Synthesis Coll. Vol. II pg. 279 there is a procedure for monomethylation of ethyl malonate (to diethyl methylmalonate). In the workup the crude, neat diester product is "shaken for exactly one minute with a cold solution of 10 g. of sodium hydroxide in 30 cc. of water."

The corresponding note reports that "[t]he ester is treated in this manner to remove any unchanged ethyl malonate [diethyl malonate]. Michael [the original author] has shown that this treatment will completely remove unchanged diethyl malonate while hardly attacking diethyl methylmalonate. No diethyl dimethylmalonate is formed when methyl bromide is used as the methylating agent.... A separation of the desired product from traces of unchanged starting material and from ethyl dimethylmalonate cannot be accomplished by distillation as the boiling points of the three esters lie within three and one-half degrees of one another." 

The notes contain this additional information. "Michael found that unchanged malonic ester can be removed completely by taking advantage of the greater ease with which it is hydrolyzed by alkali...." The references are Michael, J. Prakt. Chem. (2) 72, 537 (1905) and Fieser and Novello, J. Am. Chem. Soc. 62, 1856 (1940).

This might be an extraordinary example of separation by competitive reaction using heterogeneous media. For this to work the rate of hydrolysis of the starting material, under these conditions, must be at least one hundred times faster than that of the monomethyl product.  Why would this be and how can other separations take advantage of this type of behavior? It might be that the deprotonation and extraction into the aqueous basic phase of diethyl malonate in a vigorously agitated two-phase mixture is much faster because it has an unsubstituted methylene than for diethyl malonate that is substituted, even with a very small group like methyl. 

It is very much less likely to relate to a difference in solubility or even solubility of their respective anions. According to this hypothetical explanation, once diethylmalonate gets deprotonated and carried partially into the aqueous phase it is quickly further hydrolyzed to give carboxylates which keep it there where it is much more vulnerable to being further hydrolyzed than the monomethyl product which has preferentially 'hidden' in the organic layer. Thus this may be an example of 'phase switching hydrolysis'!