Catalyst descriptors

The single-event microkinetic concept ensures the feedstock independence of the kinetic parameters [8]. Present challenges in microkinetic modelling are the identification of catalyst descriptors accounting for catalyst properties such as acidity [10,11] and shape selectivity [12,13]. 

A single-event microkinetic description of complex feedstock conversion allows a fundamental understanding of the occurring phenomena. The limited munber of reaction families results in a tractable number of feedstock independent kinetic parameters. The catalyst dependence of these parameters can be filtered out from these parameters using catalyst descriptors such as the total number of acid sites and the alkene standard protonation enthalpy or by accounting for the shape-selective effects. Relumped single-event microkinetics account for the full reaction network on molecular level and allow to adequately describe typical industrial hydrocracking data. 

To do this, we define two additional multi dimensional spaces B and C (Figure 1). Space B contains the values of the catalyst descriptors that pertain to these catalysts e.g. backbone flexibility, partial charge on the metal atom, lipophilicity) as well as the reaction conditions (temperature, pressure, solvent type, and so on). Finally, space C contains the catalyst figures of merit (i.e., the TON, TOF, product selectivity, price, and so forth). Spaces B and C are continuous, and are arranged such that each dimension in each space represents one property. 

Cons Data quality for HT synthesis and testing (reproducibility, homogeneity, scale-up ability), as shown for this case of metals/mixed oxides WGS and Selox catalysts, remains a drawback for efficient data-mining. As such, even if the trends for new catalytic formulas may still be detected, even in the presence of poorly reproducible systems or outliers, the ultimate target of predicting the quantitative performance of a catalyst after teaching an ANN with experimental data was found to require new catalysts descriptors that would contain more information than the simple elemental composition of the materials. 

HT technology for catalysts-automated synthesis and testing appears to be reasonably adapted to date, but further improvements are expected for HT catalysts characterization, which is still restricted to costly and in general ex-situ spectroscopic techniques. These tools would provide the new catalyst descriptors needed to improve the ability to predict catalytic performances without testing. 

The blessing of the latter, i.e. the possibility to suppress specific undesired side reactions, renders any kinetic model strongly catalyst dependent which is a curse to the modeler. Accounting in a clever way for so-called catalyst descriptors deserves a separate issue. 

Fig. 5.4 Comparative display of a catalyst-HipHop hypothesis and the pharmacophore field descriptors selected by the minimalist ComPharm overlay-based model. ComPharm key features are pinpointed by arrows, while HipHop feature spheres stand for hydrophobes (light blue) and hydrogen bond acceptors (green).

Here we present an alternative concept for optimizing homogeneous catalysts. Using a virtual synthesis platform, we assemble large catalyst libraries (lO -lO candidates) in silica, and use statistical models, molecular descriptors, and... 

Figure 1. Simplified three-dimensional representation of the multi dimensional spaces containing the catalysts, the descriptor values, and the figures of merit.

Despite the fact that solvent effects on enzyme enantioselectivity appear to resist our efforts to rationalize their outcome using commonly accepted solvent descriptors, the effects are certainly there. An impressive example is provided in a report on the successful resolution of ds/trans-( 1 R,5 R)-bicyclo[3.2.0]hept-6-ylidene-acetate ethyl esters, intermediates in the synthesis of GABA (y-aminobutyric acid) analogs, by the Pfizer Bio transformations and Global R D groups (Scheme 2.2) [136]. From a screening protocol, CaLB was identified as a reactive catalyst for the hydrolysis of the racemic mixture of / //-os lor enantiomers with approximately equal activity for the ds- and tmns-isomers and a rather modest (E = 2.7) preference for the /Z-(lR,5R)-enantiomers. Application of medium engineering resulted in a phenomenal increase in the enantioselectivity (addition of 40% acetone, E > 200), while the ds- and trans-isomers were still converted at an almost equal rate. 

A separate class of experimental evaluation methods uses biological mechanisms. An artificial neural net (ANN) copies the process in the brain, especially its layered structure and its network of synapses. On a very basic level such a network can learn rules, for example, the relations between activity and component ratio or process parameters. An evolutionary strategy has been proposed by Miro-datos et al. [97] (see also Chapter 10 for related work). They combined a genetic algorithm with a knowledge-based system and added descriptors such as the catalyst pore size, the atomic or crystal ionic radius and electronegativity. This strategy enabled a reduction of the number of materials necessary for a study. 

---