Prelude - A Critical Assessment from an Industrial Point of View

Hans-Ulrich Blaser

1 Some Introductory Remarks

This preface has been written from the personal viewpoint of the author who has been active in industrial research and the development of catalytic methodologies for the production of fine chemicals, agrochemicals, and pharmaceuticals but who also has a strong interest in mechanistic aspects of catalysis [1].

While catalysis is THE key technology for the production of base and bulk chemicals, this is not (yet) the case for the fine chemical and pharmaceutical industry where classical organic synthesis dominates. Nevertheless, catalysis is able to play a valuable role in the endeavor to minimize waste production in this industry by:

• Replacing reagents which produce wastes with detrimental effects to the environment (e.g. using hydrogen instead of other reducing agents).
• Offering access to shorter, more efficient synthetic routes (e.g. direct synthesis avoiding the use of protective groups).
• Combining several different catalytic and non-catalytic steps to a "one-pot reaction," thereby avoiding workup procedures.
• Taking advantage of unique selectivity properties of a catalytic system (e.g. the possibility to introduce stereogenic centers by enantioselective transformation of prochiral substrates).

Among the different catalytic methods, catalytic hydrogenation is arguably the most valuable catalytic transformation with the highest practical impact. The versatility of this method allows for the selective conversion of an amazing number of functional groups, even in highly functionalized molecules with high yield under often relatively mild conditions. Up to now, heterogeneous catalysts such as Pd, Pt, Rh, Ru, and Ni on various supports have dominated the field of chemoselective hydrogenation [2], while soluble Rh, Ru, and Ir complexes with chiral ligands are applied predominantly for stereoselective hydrogenations [3]. These well-established catalytic systems will be the standard reference when comparing and assessing the various alternative catalysts presented in this monograph [4-12].

The dominance of classical catalysts can in part be explained by the historical development of the mechanistic understanding of the activation and transfer of dihydrogen. For decades, the prevailing opinion was that ensembles of metal atoms, i.e. heterogeneous catalysts, were necessary for the dissociative activation of dihydrogen and its transfer to an unsaturated function (also adsorbed and thereby activated on the same metal surface). It was only in the 1960s (Halpern, Wilkinson, and others) that it was shown that properly ligated transition (mostly noble) metal complexes could actually perform the same reactions. Still, the presence of a transition metal was considered indispensable. This hypothesis was challenged by the finding that certain main group metal complexes display hydrogenation activity and even more by the discovery by Stephan in 2006 that certain frustrated Lewis pairs (FLPs) are able to activate dihydrogen.

Considering the rapid development of "nontraditional" hydrogenation catalysts, the topic of this monograph is timely, allowing a closer look at the development of alternative homogeneous catalytic systems both in respect of the mechanistic understanding as well as how their performance compares to the established competition described [13]. In the following, I undertake this comparison from the point of view of their synthetic application and their potential for eventual use on a technical scale.

1.1 What Is the Motivation for Developing Non-Noble Metal Catalysts?

Besides the interest in mechanistic aspects such as new modes of hydrogen activation, the following practical aspects and expectations are most often mentioned in the various chapters:

• Cost of the metal (e.g. Pd is about 2000 times more expensive than Ni). However, this relation has to be qualified, since, very often, similar ligands are used and especially chiral ligands are often just as (sometimes even more) expensive as the noble metal. Further, much of the higher costs of the noble metal complexes are often compensated for by their much higher catalyst activities and productivities (and that is what counts for the process chemist).
• Abundance of the metal. There is no doubt that noble metals are scarce and their supply is limited. The often high catalytic activity, the typically relatively low production volume of fine chemicals, and a usually very high recovery rate (often >90% for Pd or Rh) make their application more acceptable. Nevertheless, there is a green chemistry aspect in avoiding scarce metals (and maybe even all metals) which deserves consideration.
• Metal toxicity. Due to the relatively high toxicity of (noble) metals, there are strict regulations in place for the removal of metals to the low parts per million level from active pharmaceutical intermediates [14]. Of the several classes of toxicities defined, Au, Pd, Pt, Ir, Os, Rh, Ag, Ru, as well as Co, Mo, and V belong to class 2 (appreciable toxicity); Sb, Ba, Li, Cr, Cu, Sn, and Ni to class 3 (relatively low toxicity); and Al, B, Fe, Zn, K, Ca, Na, Mn, Mg, and W to class 4 (low inherent toxicity). However, it has to be stressed that the toxicity of a metal strongly depends on the nature of the metal complex as well as on how it is administered [15].

  And, just a reminder: While organocatalytic methods obviously do not apply metals, many of the (chiral) catalysts used are actually quite toxic and have to be carefully removed from the product as well. Furthermore, the Lewis acid component in FLPs are mostly heavily fluorinated B compounds which cannot be considered as being particularly green.

• Catalyst separation and trace metal removal. In any case, metallic catalysts have to be removed from the product mixture. In order to separate the bulk of a heterogeneous catalyst, filtration is generally unproblematic. For soluble catalysts, this is usually more of a problem. Classical solutions are distillation for volatile (which also removes trace metals) and crystallization for solid products. In rare cases, the catalyst can be precipitated and removed via filtration. There are several strategies to remove trace metals to get acceptable rest concentrations; recrystallization and adsorption by scavengers which are then either extracted or filtered off are the most useful ones [16].
• Different reactivity of non-noble metals. Obviously, the various metal complexes have quite different chemical properties compared to the established noble metal complexes. Often, this will be a disadvantage for their hydrogenation properties which has to be compensated for, e.g. with suitable ligands or higher reaction temperatures and/or hydrogen pressures. However, there are also contrasting examples as, for example, the higher reactivity of low-valent Ni complexes which, compared to Pd catalysts, often react faster with unreactive organic molecules such as aryl chlorides.
• No need for high-pressure equipment for transfer hydrogenation s (TH s). This argument invoked for organo- and biocatalytic reductions might be valid when no such equipment is available, a case rather seldom in modern fine chemicals development and production. With hydrogen donors such as isopropanol, there might be a problem with incomplete conversion due to equilibria. Hantzsch esters are relatively expensive hydrogen donors and the resulting pyridines have to be separated; and in the case of the cofactors used in biocatalysis, sophisticated regeneration strategies are required. This means that such catalysts will be used only when their catalytic performance is superior to alternative solutions, which for selected biocatalytic reductions is indeed the case.

1.2 Crucial Parameters for Process Development

In process development, there is usually a hierarchy of goals (or criteria) to be met. It is simply not possible to reach all the requirements for a technically useful process in one step. As depicted in Table 1, the catalyst selectivity (combined, of course, with an acceptable activity) is the first criterion - just as in academic research. However, when a reasonable selectivity has been obtained, other criteria will become important: catalyst activity, productivity and stability, catalyst separation (and, maybe, recycling). Then, questions like the cost and availability of the (chiral) catalyst and other materials have to be addressed. The final process is usually a compromise since often not all of these requirements can be fulfilled maximally. It is useful to divide the development of a manufacturing process into different phases; however, it is rarely possible to proceed in a linear manner and very often one has to go back to an earlier phase in order to answer additional questions before it is possible to go on.

Table 1 Catalyst choice: Criteria and requirements during different process development phases.