Solvation computation approaches

Table 7.7 Difference in free energies of solvation (kcal mol 1) for the classical and nonclassical 2-norbornyl cations by three very different computational approaches.

Calculations of time-dependent electromagnetic properties of molecules at the correlated electronic structure level are conveniently carried out by the utilization of modern response theory [43-51], The transition of modern response theory for gas phase molecular systems to solvated molecules has been established [1-6] and these methods include the use of correlated electronic wavefunctions. These methods, reviewed here, have given rise a large number of computational approaches for calculating electric and magnetic molecular properties of solvated molecules. 

The review is largely of rather traditional techniques — fragment methods and correlation between properties. More modem techniques based on wholly a priori computational approaches have not yet yielded methods that are robust in fact, much published material in this area is singularly unconvincing. Why should that be It is because physiochemical properties involve such matters as solvation and intermolecular forces that computational methods frequently fail the energy differences that need to be understood are small and not easily predicted computationally. 

Computational approach. Lee and Houk conducted calculations using a methyl-ammonium ion to mimic the key lysine of the enzyme active site.16 They chose this model because, even though no crystal structures had been solved at the time, a lysine was known to be essential for catalysis.60 The reaction of orotate + CH3NH3 to form a carbene-methylamine complex was thus examined in various dielectrics using the SCI-PCM SCRF method in Gaussian 94.30 31 48 Solvation energies computed at the RHF/6-31 + G level were used to correct gas phase MP2/ 6-31 + G energies and obtain AH values for reaction in solution. 

In considering various computational approaches to solvation, it must first be understood that the ion-water association alone offers a great range of behavior as far as the residence time of water in a hydration shell is concerned. Certain ions form hydrates with lifetimes of months. However, for the ions that are nearly always the goal of computation (ions of groups lA and HA in the Periodic Table and halide ions), the lifetime may be fractions of a nanosecond. 

It is sometimes desirable not only to predict the struaure of a potentially active ligand, but also to perform rigorous quantitative calculations on the proposed ligand-receptor complex. Two major computational approaches have been used for such predictions interaction energy calculations and free energy perturbation techniques. Such calculations should take into account solvation effects because this is often a driving force in the interaction. 

Enthalpic and entropic contributions to diastereofacial selectivity have been explored in the addition of n-butylUthium to the C=N bond of R CH(OR )CH=NSiMe3." ° Using THF or n-hexane as solvent, temperature ranges of up to 150 °C can be covered, over which a change in the anttsyn ratio of the products from 3 1 to 1 3 can be achieved. The results are discussed in terms of stereospecific solvation effects on the reacting r-system, an area in need of an appropriate computational approach. 

The computational approach also makes it possible to gain insight into the effect of solvation on the enthalpy of formation without conducting experiments. A calculation performed in the absence of solvent molecules estimates the properties of the molecule of interest in the gas phase. Computational methods are available that allow for the inclusion of several solvent molecules around a solute molecule, thereby taking into account the effect of molecular interactions with the solvent on the enthalpy of formation of the solute. Again, the numerical results are only estimates, and the primary purpose of the calculation is to predict whether interactions with the solvent increase or decrease the enthalpy of formation. As an example, consider the amino acid glycine, which can exist in a neutral (5) or zwitterionic (6) form, in which the amino group is protonated and the... 

The development of computational approaches to vibrational spectra of systems in the condensed phase requires the analysis of the physics of the solvated system, thus introducing in the model ways for treating the interaction between the system and its surrounding. This is a general requirement for a procedure to be successful in the description (and further prediction) of spectra it is in fact well known that all spectral features (frequencies, intensities, and bandshape) are hugely influenced by the interactions between the system and the environment. 

Computational chemistry has had a powerful impact on the study of proton transfer because many important applications involve small systems and can be studied with high levels of theory. Under these circumstances, very satisfactory results generally have been obtained. The new frontier involves incorporating solvation effects and dynamical effects in a general, straightforward manner. This will extend the applicability of computational approaches and provide a better interface with experimental work. 

---