Bifunctional Hydrogenation Catalysis in Aqueous and Formic Acid/Triethylamine Media

Abstract

Catalytic hydrogenation of organic compounds with molecular hydrogen is fast and efficient, but requires high gas pressures. Recently, there has been a revival in the field of catalytic transfer hydrogenation using formic acid or alcohols as the sacrificial hydrogen donor. These synthetic alternatives are operationally simple and avoid gas handling with specialized equipment. Noyori and co-workers1-3 have developed a highly active and enantioselective catalytic transfer hydrogenation of ketones and imines using (fØ6-arene)Ru(H)TsDPEN (TsDPEN = (1S,2S)- or (1R,2R)-N-tosyl-1,2-diphenylethylenediamine) and a hydrogen donor (HCOOH or 2-propanol). They found the catalysis to follow an unconventional ionic hydrogenation pathway involving the simultaneous transfer of a hydride from ruthenium and a proton from the amino group of the TsDPEN ligand. Studies have shown that this mechanism does not involve the prior coordination of the substrate to the metal center, nor the interaction of the product within the inner coordination sphere of the catalyst. Noyori has coined the term ¡§metal-ligand bifunctional¡¨ to describe the cooperative participation of the metal and ligand in the bond forming process of this reaction. The success of Noyori¡¦s catalyst has triggered extensive studies on this new type of transfer hydrogenation.

A complete catalytic system can be regarded as being composed of three components: the catalyst, the substrate, and the reaction medium. In this dissertation, we examine metal-ligand bifunctional hydrogenation catalysis in terms of these three components. In addition, we report on extensions of bifunctional hydrogenation to other reactions, including reductive amination.

Initially, by studying structure variations on the catalytic activity of the bifunctional hydrogenation we gained insight into the mechanism of this reaction in various media. We believe our discoveries in this area allow the design of more efficient catalytic systems. For example, the crucial role of the ligand amino group on the reaction pathway, the ¡§N-H effect¡¨, has been recognized since the discovery of the reaction by Noyori. The orientation of the ligand N-H bond and its acidity are key to the proton transfer step, and therefore, to overall catalytic performance. We undertook a structure-activity study centered on the ligand N-H group by examining the kinetics of a reaction series containing para-substituted phenyl groups at the N-H nitrogen. Our ligand series included substituents ranging from electron donating to electron withdrawing groups, designed to systematically vary the chemical properties of the N-H groups, especially their pKa values. We observed a correlation of the ligand pKa with catalytic activity, which is consistent with a mechanism involving proton transfer between the N-H group and the substrate. The catalysis was carried out in 2-propanol and the 5:2 formic acid/triethylamine azeotropic mixtures and in both media lower ligand pKa¡¦s suppressed hydrogenation rates, while higher pKa¡¦s promoted catalytic activity. These findings are summarized in Hammett plots and are consistent with proton transfer to the ligand in the rate-determining regeneration step of the catalyst. In addition, we observed linear correlations for catalytic rate constants as a function of ligand N-H pKa in 2-propanol and the azeotropic solvent; however, the sensitivity of the rate constants to the pKa was different in the two media. These findings allow us the design superior catalysts based on modification of the N-H acidity.

In addition, we successfully performed hydrogenation catalysis in aqueous formate solution without any modification of Noyori¡¦s hydrophobic catalyst. We studied the effect of the substrate/catalyst ratio on the reactivity and found that rates were attenuated under conditions of high substrate loading. Further, we determined that the catalyst resides primarily in the organic phase while the hydrogen source is located in the aqueous phase. The use of phase transfer catalysts was explored and in some cases resulted in rate enhancement. For example, dramatic improvements were observed after incorporating long alkyl chain quaternary ammonium salts. Other catalytic conditions, such as stirring speed, temperature, counterions, volume of the solution, and concentration of phase transfer catalyst, as well as sodium formate were screened to optimize this biphasic transfer hydrogenation.

Noyori¡¦s catalyst has been used to efficiently catalyze the hydrogenation of ketones, aldehydes, activated olefins and imines. The imine substrate scope has been limited to cyclic imines probably due to the ease with which linear imines decompose to the corresponding aldehyde and amine. We were excited to find that linear imine reduction can be readily performed by using Noyori¡¦s catalyst in the azeotrope medium. More importantly, the direct reductive amination was successfully achieved in a one-pot catalytic process. However, this method was found to be limited to aromatic aldehydes since the aliphatic aldehydes undergo numerous side reactions in the azeotropic solvent. Further study suggested that 5:3 formic acid/triethylamine mixture was a better reaction medium than the traditional 5:2 azeotropic mixture. Even though the selectivity of direct reductive amination in the 5:3 mixture decreased slightly, the reactivity was dramatically enhanced and the substrate scope was expanded to aliphatic aldehydes. Our results also show the 5:3 mixture to be a more efficient medium for the hydrogenation of ketones and aldehydes. Generally, imine formation or imine reduction reactions require in rigorously dried solvents since water can be detrimental to the reaction; however, our discovery of a direct reductive amination in aqueous sodium formate solution using Noyori¡¦s catalyst challenges this assumption regarding the reactions of organic compounds in aqueous media.