Identification of True Minima

The structures investigated for the water trimer up to this point have been idealized, with no guarantee that any are true minima on the full potential energy surface (PES). There are 

A sequential dimer of the d-a type in Fig. 5.22 is not a true minimum on the full PES of the trimer. Optimizations beginning with such a structure bring the two terminal molecules closer together until the trimer closes up to the cyclic arrangement. The d-d structure in Fig. 

22 does in fact represent a minimum on the PES. Its total binding energy is less than that of the cyclic trimer by 5-6 kcal/mol. In contrast, the double acceptor geometry a-a in Fig. 

22 is not a minimum at all, but rather a second-order saddle point. On the other hand, the 

The structure of the cyclic water trimer was optimized also using correlation-consistent polarized basis sets, augmented with additional diffuse functions . Whereas the three R(0 0) distances are all different at the SCF level, MP2 correlation brings them within 0.002 A of being equal to one another. The three H-bonds are all nonlinear by about 30°, that is, e(OH--O) 150°. 

---