Reproducibility of Conformational Energies

Mimic (1.0), and ChemSD Plus (3.0) software, in addition to the abilities of their force fields [TRIPOS, COSMIC, MMX, MM2(87), and MM2, respectively] to reproduce experimental conformational energies. This comparison comprised a series of 31 molecules and their corresponding conformational energy data. In this test, MM2(87) and the two MM2-based methods (MMX and MM2 ) were clearly superior to the TRIPOS and COSMIC force fields. The rms deviations for the MM2 methods were 0.32-0.37 kcal/mol, whereas the corresponding values for the TRIPOS and COSMIC force fields were 0.90 and 0.93 kcal/mol, respectively. 

In connection with the validation of the CFF93 force field, Hwang et al. compared CFF93 to MM3, AMBER, and CVFF for a very limited set of six simple hydrocarbons (eight rotational barriers and conformational energies). In this test, MM3 and CFF93 appear to be somewhat better than the other force fields. The value of the test is limited, however, by the limited size of the dataset and by the use of a number of these data in the parameterization of all the tested force fields. 

In this section, we present some examples of different force fields to see how well they reproduce certain fundamentally important conformational energy 

3-Dimethylbutane is an interesting test case because the two con-formers have different numbers of methyl-methyl gauche interactions—two and three, respectively. In spite of this, the experimental energy difference is essentially zero. This result is attributable to differences in the relaxation of 

The axial-equatorial energy difference in phenylcyclohexane is less well treated by most force fields. UFF is in error by as much as 5 kcal/mol, and half the other force fields overestimate the observed energy difference by 1 kcal/mol 

---

Fig. 21. Crystal structure of the Valinomycin-K+ complex. Reproduced with permission from Ref.100). This crystal structure confirmed within tenths of an Angstrom the structure derived previously in solution 97 98) and by means of conformational energy calculations...

Figure 2.12 Maps of conformational energy of various syndiotactic polymers as function of backbone torsion angles 0 and 0227 (a) syndiotactic polystyrene, (b) polypropylene, (c) poly (1-butene), and (d) poly(4-methyl-l-pentene). Succession of torsion angles. .. 0i 0i 0202 - - -[s(M/N)2 symmetry] has been assumed. Isoenergetic curves are reported every 5 kJ/mol of monomeric units with respect to absolute minimum of each map assumed as zero. Values of energies corresponding to minima (x) are also indicated. Experimental conformations observed for different polymorphic forms of polymers are indicated by triangles. (Reproduced with permission from Ref. 27. Copyright 1992 by the Socicta Chimica Italiana.)...

Figure 13. Map of conformational energy for rotations about the Si-Si skeletal bonds in polysilane, [SiHz-Jn- The energies (in kilocalories per mole) relative to the minima (designated by the plus signs) are shown as contour lines. (Reproduced from reference 60. Copyright 1986 American Chemical Society.)...

In spite of this common problem, there are surprisingly few publications in the literature in which molecular mechanics force fields are compared to assess their ability to reproduce experimental conformational energy data. Usually, authors who report a new force field use a more or less extensive set of experimental data to validate it. However, independent testings of the performance of such data sets are scarce. 

We have searched the literature for publications comparing the performance of several different force fields in regard to their ability to reproduce experimental conformational energies. The results of this search are summarized in Table 1.14,15,17,19,24-27... 

Fig. 20. A. Conformation of the Valinomycin-cation complex derived for solution using a combination of proton magnetic resonance data and conformational energy calculations. This structure agrees within tenths of an Angstrom with the crystal structure subsequently determined (100) and shown in Fig. 21. Reproduced with permission from Ref.99).

Figure 51. Arrhenius plot of ln 1/(3 [ Q t)ldt2]) from data corresponding to Fig. 54. The conformational energy consumed per mole of polymeric segments in the absence of any external electric field (AH) can be obtained from the slope. (Reprinted from T. F. Otero and H.-J. Grande, Reversible 2D to 3D electrode transition in polypyrrole films. Colloid Surf. A. 134, 85, 1998, Figs. 4-9. Copyright 1998. Reproduced with kind permission of Elsevier Science-NL, Sara Burgerhartstraat 25, 1055 Amsterdam, The Netherlands.)...

---