The TEAF System

The catalysts are best prepared in situ by mixing a half-equivalent of the di-chloro-metal aromatic dimer with an equivalent of the ligand in a suitable solvent such as acetonitrile, dichloromethane or isopropanol. A base is used to remove the hydrochloric acid formed (Fig. 35.3). If 1 equiv. of base is used, the inactive pre-catalyst is prepared, and further addition of base activates the catalyst to the 16-electron species. In the IPA system the base is conveniently aqueous sodium hydroxide or sodium isopropoxide in isopropanol, whereas in the TEAF system, triethylamine activates the catalyst. In practice, since the amount of catalyst is tiny, any residual acid in the solvent can neutralize the added base, so a small excess is often used. To prevent the active 16-electron species sitting around, the catalyst is often activated in the presence of the hydrogen donor. The amount of catalyst required for a transformation depends on the desired reaction rate. Typically, it is desirable to achieve complete conversion of the substrate within several hours, and to this extent the catalyst is often used at 0.1 mol.% (with SCR 1000 1). Some substrate-catalyst combinations are less active, requiring more catalyst (e.g., up to 1 mol.% SCR 100 1), in other reactions catalyst TONs of 10000 (SCR 10000 1) have been realized. 

The TEAF system can be used to reduce ketones, certain alkenes and imines. With regard to the latter substrate, during our studies it was realized that 5 2 TEAF in some solvents was sufficiently acidic to protonate the imine (p K, ca. 6 in water). Iminium salts are much more reactive than imines due to inductive effects (cf. the Stacker reaction), and it was thus considered likely that an iminium salt was being reduced to an ammonium salt [54]. This explains why imines are not reduced in the IPA system which is neutral, and not acidic. When an iminium salt was pre-prepared by mixing equal amounts of an imine and acid, and used in the IPA system, the iminium was reduced, albeit with lower rate and moderate enantioselectivity. Quaternary iminium salts were also reduced to tertiary amines. Nevertheless, as other kinetic studies have indicated a pre-equilibrium with imine, it is possible that the proton formally sits on the catalyst and the iminium is formed during the catalytic cycle. It is, of course, possible that the mechanism of imine transfer hydrogenation is different to that of ketone reduction, and a metal-coordinated imine may be involved [55]. 

The IPA system does not require a co-solvent, but one can be used if this proves advantageous. In the TEAF system a solvent is normally used, though neat TEAF or formic acid can be used if required. The solvent can have a large effect on the reaction rate and optical purity of the product this may in part be because the substrate seems to bind by weak electrostatic interactions with the catalyst, and is also partly due to the pH of the system. Solvents have a dramatic effect on the ionization of formic acid for example, in water the piCa is 3.7, but in DMF it is 11.5. This is because formation of the formate anion becomes less favorable with less polar solvents (see Table 35.2). The piCa of triethy-lamine is far less sensitive. As a consequence, formic acid and triethylamine may remain unreacted and not form a salt. The variation in formic acid piCa can also have a significant impact on the catalyst and substrate, particularly when this is an imine. 

The transfer hydrogenation methods described above are sufficient to carry out laboratory-scale studies, but it is unlikely that a direct scale-up of these processes would result in identical yields and selectivities. This is because the reaction mixtures are biphasic liquid, gas. The gas which is distilled off is acetone from the IPA system, and carbon dioxide from the TEAF system. The rate of gas disengagement is related to the superficial surface area. As the process is scaled-up, or the height of the liquid increases, the ratio of surface area to volume decreases. In order to improve de-gassing, parameters such as stirring rates, reactor design and temperature are important, and these will be discussed along with other factors found important in process scale-up. 

In the TEAF system there is no problem with any back-reaction, and concentrations up to 10 M are possible. Although neat TEAF has been used satisfactorily, it is quite viscous and so it is preferable to use a diluent. As mentioned above, the solvent may have a marked effect on both reaction rate and enantio-selectivity. 

The TEAF system is usually only slightly exothermic, and so again all components can be mixed together (TEAF is prepared separately as this process is exothermic). In this case the triethylamine in the TEAF is sufficient to activate the catalyst. In the laboratory high conversions are seen, but on scale-up some substrates fail to be completely converted. 

The diamine ligands are successful in the TEAF system, but are poor in the IPA system, whilst 1,2-aminoalcohol ligands work well in the IPA, but not the TEAF system. The reason for this is not well understood, but it illustrates the subtle nature of the electronic environment on the metal. 

Rather surprisingly alcohols are poor at reducing imines, yet TEAF works well. During our studies we rationalized that the TEAF system was sufficiently acidic (pH approximately 4) to protonate the imine (pK l approximately 6) and that it was an iminium that was reduced to an ammonium salt [14]. When an iminium was used in the I PA system, it was reduced albeit with a low rate and moderate enan-tioselectivity. Quaternary iminium salts were also reduced to tertiary amines. Hydrogen will not reduce ketones or imines using the CATHy catalysts, but hydrides such as sodium borohydride have been shown to work. 

The CATHy catalysts are best used below 40 °C, above this temperature we have observed signs of decomposition. In the I PA system, preventing the back-re-action depends on how efficiently acetone is distilled. Normally this would be best done at around 80 °C, the boiling point of isopropanol, but an optimal performance of the catalyst requires ambient temperature or less, and reduced pressure. Whilst acetone can be fractionally distilled, it is simpler to distil the mixture with isopropanol and to maintain constant volume by continuously charging with fresh solvent. In the TEAF system the reaction is normally operated at ambient temperature. Operating at lower temperatures can improve the enantiomeric excess slightly but gives lower rates, for example with 4-fluoroacetophenone the results described in Tab. 3 were achieved. 

Fig. 7 Asymmetric transfer hydrogenation of 4-fluoroacetophenone using the TEAF system. The effect of bubbling nitrogen through the reaction mass.

Water has been shown to enhance the activity of ruthenium and rhodium catalysts in both the TEAF and potassium formate systems [34, 36, 52]. The aqueous systems enable much simpler control of pH this is important, as Xiao has found that a low pH markedly slows the reaction [52]. The pH at which this occurs corresponds with the pKa of formic acid (i.e., 3.7), implying that the formate anion is required for complexation with the catalyst. Xiao has proposed two possible catalytic cycles - one that provides poor ee-values at low pH as a result of ligand decomplexation, and another that gives high ee-values at high pH. 

The enzymic formation of aldehydes, ketones, alcohols, and oxoacids (from linoleic and linolenic acids) on disruption of plant tissues is an important biosynthetic pathway by which fruit and vegetable volatiles are formed. Some examples are (E)-2-hexenal ("leaf aldehyde") and ( )-3-hexenol ("leaf alcohol") in tea (E)-2-hexenal in apples (E,Z)-2,6-nonadienal ("violet Teaf aldehyde") and (E)-2-nonenal in cucumber ( Z)-5-nonenal in musk melon (Z,Z) -3,6-nonadienol in water melon, and 1-octen-3-ol ("mushroom alcohol") in certain edible mushrooms and Fungi. The enzyme system is highly substrate specific to a (Z,Z)-1,4-pentadiene system (like lipoxygenase) splitting the >C = C< double bond at the W - 6 and/or W - 9 position. Therefore linoleic-, linolenic-, and arachidonic acids are natural substrates. It seems to be a common principle in leaves, fruits, vegetables, and basidiomycetes. 

---