The development of selective organometallic molecular catalysts is one of the most promising areas of molecular chemistry and which presently developes at an unforeseen and still increas-ing pace. Organometallic molecular catalysts have a unprecedented selectivity which is only paralleled by the astounding selectivity of biological catalysts the enzymes. Organometallic catalysts make use of the same type of so called secondary interactions as enzymes to reach this selectivity. They make use of well designed surfaces which embed the catalytic centre and orient the reagent by secondary interactions. The orientation between the catalytically active centre and the substrate is pre(cid:150)organised in such way that only one of many possible products will result. For the rational design of an organometallic catalyst the knowledge of the secondary interactions is a pre(cid:150)requisite therefore. Since this type of interaction can not be reliably calculated by quantum mechanical methods for large molecules chemists have to use other types of models, the most promising approach being the force (cid:2)eld model which describes a molecule in terms of a mechanical machine. The potentials which are necessary to describe the stiffness of the different types of springs and joints of such a model have to be inferred from experimental data. Force (cid:2)elds describing organic molecules have been developed and are highly successful in modelling and predicting the behaviour of organic compounds. On the other hand force (cid:2)eld models for the description of organometallic compounds which would be based on as many experimental data as available are not generally known. The paper describes the development of an appropriate force (cid:2)eld for a class of rhodium compounds which are activein enantioselec- tive catalysis by a novel approach. The parameters of all potentials involving contributions by the rhodium atom are optimised by Genetic Algorithms on the basis of as many experimentally determined structures as available. The ef(cid:2)ciency of of this approach is demonstrated by the comparison of calculated and experimental data with respect to different shapes of the catalyst to their transformations into each other. Predicted and observed energies are found to agree within a few kJ/mol. These results recommend the approach chosen as a tool for the rational design of organometallic catalysts. space analysis on a CRAY T3E at the NIC in J¤ulich using 5000 h of CPU

The stereochemical flexibility of six-membered chelate cycles containing phosphorus donor groups has not yet been analysed in detail. With a number of compounds of the type [(kappa-PR2CH2CH(OH)CH2-K-PR'(2))Rh(eta(4)-COD)]+PF6- at hand, and since the solid state structures and the catalytic properties of these compounds are known, an ideal opportunity to study this topic was available. After assigning all of the important H-1, C-13, and P-31 NMR spectroscopic resonances by a combination of one- and two-dimensional NMR spectroscopic methods quantitative interpretation of the 2D NOESY/EXSY spectra was achieved for four compounds [PR2 = PMes(2); PR'(2) = PPh2 (1), DBP (2), PEt2 (3), P(o-anisyl)(2) (4)]. It is found that the conformations that these compounds adopt in solution are similar to the ones observed for them in the solid state. The conformational ensemble in solution comprises lambda twist and delta twist forms;) and 6 isomers are clearly differentiated by quantitative NOE-based structure analysis (distance geometry). Since the sense of chirality of the ligand is known, an absolute assignment of these conformations is possible. The equilibrium constants at 298 K for the delta reversible arrow lambda, isomerisation of compounds 1-4 are not very different from one another, with reaction enthalpies ranging from -8 to -2 kJ mol(-1) and the reaction entropies in the range of -10 to 8 J K-1 mol(-1). The activation barriers are again similar for 1-4 and lie at DeltaH(not equal) = 64 kJ mol(-1). It is observed that the mesityl entities at the PMes(2) donor groups behave as coupled rotors with two rotational pathways open to them during the 6 reversible arrow delta interconversion while they are rotationally fixed in the lambda and delta conformations themselves. Comparing the above results with the enantioselective discrimination found for the same compounds in hydrogenation experiments it becomes clear that there is no direct correlation between the preference of one twist form over the other in the precatalyst and its chiral performance. (C) Wiley-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002.

Eleven solid state structures of eight compounds 1-8 of the type [{kappa-PAr2CH2CH(OH)CH2-x-PAr'(2))Rh-(eta(4)-COD)]+PF6- are used as a basis for deriving a force field model for this class of compound by Genetic Algorithms. By a complete search in the conformational space of [{kappa-PMes(2)CH(2-) CH(OH)CH2-K-PPh2)Rh-(eta(4)-COD)]+PF6- (1) it is shown that the model is not only capable of reproducing the structures but as well capable to predict the stability of individual conformers and the conformational reaction pathways in full agreement with experimental observations. The model predicts that the k twist conformation is the most stable conformation of 1. The delta twist conformation is calculated 3.1 kJ(.)mol(-1) above this minimum. The experimental value is 3.4 kJ.mol-1. The model predicts that there are two mechanistic pathways for the kJ(.)mol(-1) isomerization process. Both of them are characterised by a strictly coupled rotation of the mesityl groups of the PMes(2) entity of 1. These two pathways differ only in the sense of this rotation. The activation enthalpy for the lambda reversible arrow delta isomerization process as a whole is calculated as DeltaH(#) = 69.1 kJ-mol(-1). The experimental value is DeltaH(#) = 64.4 kJ(.)mol(-1). It is concluded therefore that the novel type of approach as described leads to models of very high predictive power. (C) Wiley-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002.