Models discrete solvation

It came out immediately clear that the supermolecule approach cannot represent the method to be used in extensive studies of solvent effects. The computational costs increase in the ab initio versions with more than the fourth power of the number of basis set functions, at a given nuclear geometry of the supermolecule. Even more important it has been the recognition that, when the size of the solvation cluster exceeds some very low limits, the number of different nuclear conformations at an equivalent energy increases exponentially computational costs increase in parallel, and the introduction of thermal averages on these conformations becomes necessary. These facts, and some attempts to overcome them, are well summarized in a dementi s monograph (Clementi, 1976). The problem of multiple equivalent minima still plagues some discrete solvation models. 

Traditional continuum, discrete, and mixed discrete-continuum models developed during the past few years have been reviewed in a number of papers (Bonaccorsi et al., 1982 Claverie, 1982 Tapia, 1982). In spile of their obvious shortcomings, there is still hope that the models proposed so far may be suitable for describing, at least partially, some of the features of the solvation phenomena. We shall present an application of theoretical models of solvation to the estimation of the influence of the environment on the shift of the tautomeric equilibrium A B. First, some comments on the models are given. 

These results may partly validate discrete solvation models, where environmental effects are mainly represented by electrostatic effects. In both complexes correlation term as well as three- or four- body effects due to additional solvent molecules are negligible. More approximate results could be obtained using Effective Fragment Potentials AEefp [31] or CAMM estimate Eel,mtp ... 

In the Polarizable Continuum Model for solvation, the molecular solute is hosted in a cavity of a polarizable dielectric medium representing the solvent The cavity is accurately modeled on the shape of the molecular solute (Miertus et al. 1981), and the dielectric medium is characterized by the dielectric permittivity e of the bulk solvent The physics of the model is very simple. The solute charge distribution polarizes the dielectric medium, which in turn acts back on the solute, in a process of mutual polarization that continues until self-consistence is reached. The polarization of the solvent is represented by an apparent charge distribution (ASC) spread on the cavity surface. In computational practice the ASC is discretized to a set of NTS point charges and the solute-solvent interaction is expressed as in terms of the interaction between these and the charge distribution of the molecular solute. 

Basis sets augmented with hydrogen-like orbitals are within 5 ppm of the experimental values (measured within 2 ppm) for the discrete solvated model. This model explicitly includes several deuterated methanol molecules to cater for the specific hydrogen bonding interactions. 

Specific solute-solvent interactions involving the first solvation shell only can be treated in detail by discrete solvent models. The various approaches like point charge models, siipennoleciilar calculations, quantum theories of reactions in solution, and their implementations in Monte Carlo methods and molecular dynamics simulations like the Car-Parrinello method are discussed elsewhere in this encyclopedia. Here only some points will be briefly mentioned that seem of relevance for later sections. 

The mixed solvent models, where the first solvation sphere is accounted for by including a number of solvent molecules, implicitly include the solute-solvent cavity/ dispersion terms, although the corresponding tenns between the solvent molecules and the continuum are usually neglected. Once discrete solvent molecules are included, however, the problem of configuration sampling arises. Nevertheless, in many cases the first solvation shell is by far the most important, and mixed models may yield substantially better results than pure continuum models, at the price of an increase in computational cost. 

For molecules and molecular ions, such as the cations of 8-methyl-N5-deazapterin and 8-methyl-pterin, the charge distribution (which is represented in MD simulations by a set of discrete atomic charges) will be dependent on the chosen quantum chemical model. Differences in the charge distributions of these cations may influence both the relative binding and solvation thermodynamics. Consequently, we studied the relative solvation thermodynamics of similar DHFR-binding molecular ions.30 Atomic charges... 

It can be seen from Table 26.1 that various methods used to model the effect of a solvent can be broadly classified into three types (1) those which treat the solvent as continuous medium, (2) those which describe the individual solvent molecules (discrete/explicit solvation), and (3) combinations of (1) and (2) treatments. The following section provides a brief introduction to continuum models. 

Quantitative models of solute-solvent systems are often divided into two broad classes, depending upon whether the solvent is treated as being composed of discrete molecules or as a continuum. Molecular dynamics and Monte Carlo simulations are examples of the former 8"11 the interaction of a solute molecule with each of hundreds or sometimes even thousands of solvent molecules is explicitly taken into account, over a lengthy series of steps. This clearly puts a considerable demand upon computer resources. The different continuum models,11"16 which have evolved from the work of Bom,17 Bell,18 Kirkwood,19 and Onsager20 in the pre-computer era, view the solvent as a continuous, polarizable isotropic medium in which the solute molecule is contained within a cavity. The division into discrete and continuum models is of course not a rigorous one there are many variants that combine elements of both. For example, the solute molecule might be surrounded by a first solvation shell with the constituents of which it interacts explicitly, while beyond this is the continuum solvent.16... 

The level of accuracy that can be achieved by these different methods may be viewed as somewhat remarkable, given the approximations that are involved. For relatively small organic molecules, for instance, the calculated AGsoivation is now usually within less than 1 kcal/mole of the experimental value, often considerably less. Appropriate parametrization is of key importance. Applications to biological systems pose greater problems, due to the size and complexity of the molecules,66 156 159 161 and require the use of semiempirical rather than ab initio quantum-mechanical methods. In terms of computational expense, continuum models have the advantage over discrete molecular ones, but the latter are better able to describe solvent structure and handle first-solvation-shell effects. 

It is clear from the above that the continuum model can simulate only those aspects of the solvent which are somewhat independent from hydrophobicity, hydrophyUicity, generally the first solvation shell, and specific interactions with the solute. The physical problem is a general one namely, it relates to the validity to use quantities, correctly described and defined at the macroscopic level, in the discrete description of matter at the atomic level. For such study, one needs explicit consideration of the solvent, for example the molecules of water. This can be done either at the quantum-mechanical level, as in cluster computations. Another approach is to simulate the system at the molecular dynamics (or Monte Carlo) level these techniques allow us to consider... 

Discrete and continuum models for the solvent involvement have been employed to steady equilibrium and non-equilibrium solvation effects on bromination of ethylene. Two mechanisms were identified that lead to transition states of different symmetry. One mechanism operates in the gas phase and non-polar solvents. The second one, that leads to the typical C2V transition state, holds in medium-to-very polar solvents. In water, the solvent molecules participate actively and non-equilibration solvations effects proved to be substantial and larger than those previously reported for the >SN2 reaction.23... 

To appreciate the basic reasons why continuous models are so versatile and promising for more applications, however, we have to consider again the simple systems and the simple properties mentioned above. The best way to gain this initial appreciation is to contrast the procedures given by discrete and continuum methods to obtain the solvation energy in a very dilute solution. 

---