LFMM model

Computer programs for empirical force-field calculations that use other concepts have been tested, and some of these will be discussed elsewhere in this book. Among these approaches are one based on a pure central force-field model, used for simple organic compounds [208], an equipotential surface force-field model, used for carbonyl cluster complexes [99,103], and a program that includes ligand field terms, developed for transition metal complexes in the LFMM model (ligand field molecular mechanics) [21, 105, 108]. 

The LFMM FF for the oxidized Cu(II) centers was designed around suitable homoleptic species, viz., [Cu(imidazole)4]2+, [Cu(SCH3)4]2-, [Cu(S(CH3)2)4]2+, and [Cu(0=CH2)4]2+ (37). These complexes represent models for Cu-histidine, Cu-cysteine, Cu-methionine, and Cu-glutamine O/peptide respectively. Only the first of these species is known experimentally. However, it is amply documented that DFT gives excellent structures for metal complexes (64,65) so we can access the remaining species computationally (Fig. 20). 

Fig. 20. Overlay of DFT-optimized (blue) and LFMM-optimized (yellow) model structures, clockwise from top left [Cu(DMS)4]2+, [Cu (SCH3)4]2, [Cu(0=CH2)4]2+, [Cu(imid)2(SCH3)(DMS)]+, and [Cu...

The main driver behind developing LFMM is computational efficiency. Although DFT has undoubtedly revolutionized the theoretical treatment of TM systems, there are many instances where all QM methods, even DFT, are simply not viable. Comprehensive conformational searching, molecular dynamics, and virtual screening represent hundreds of thousands of individual calculations and QM methods are too expensive. We are forced to turn to classical models but, for TM centers, this presents a whole new set of challenges. However, since we cannot easily make QM orders of magnitude faster, our only option is to make MM smarter and thus able to cope with the extra demands of coordination complexes. 

---