Strictly diabatic states

In this appendix we generalise the expressions of the diabatic quantities first introduced in Sec. 2 for the ideal case of an exact two-level problem to a more realistic description. In a normal situation, the Hamiltonian has an infinite number of eigenstates, and there is no finite number of strictly diabatic states [76] that can describe a given pair of adiabatic states [77-80]. Instead, one can define a unitary transformation of the adiabatic states generating two quasidiabatic states characterised by a residual non-adiabatic coupling, as small as possible, but never zero (see, e.g., [5,24,32-35]). In practice, the electronic Hilbert space is always truncated to a finite number of configurations. In what follows, we consider the case of MCSCF wavefunctions and make use of generalised crude adiabatic states adapted to this. 

In this context, let us briefly return to the two-state case. As discussed above, strictly diabatic states exist in one dimension, e.g. for a diatomic molecule with interatomic distance R. In a typical situation, the adiabatic potentials exhibit an avoided crossing and the diabatic potentials cross each other once a function of R. Assuming that the derivative couplings vanish at infinite internuclear distance, and using Eq. (30) then gives... 

Strictly diabatic states are defined in order to have vanishing derivative cou-phngs [17],... 

It is prerequisite to define localized, diabatic state wave fimctions, representing specific Lewis resonance configurations, in a VB-like method. Although this can in principle be done using an orbital localization technique, the difficulty is that these localization methods not only include orthorgonalization tails, but also include delocalization tails, which make contribution to the electronic delocalization effect and are not appropriate to describe diabatic potential energy surfaces. We have proposed to construct the locahzed diabatic state, or Lewis resonance structure, using a strictly block-localized wave function (BLW) method, which was developed recently for the study of electronic delocalization within a molecule.(28-3 1)... 

The concept of a diabatic state has different definitions. Strictly speaking, a basis of diabatic states (ft, If...) should be such that Equation 10.9 is satisfied for any variation 3Q of the geometrical coordinates (Q). 

The BOVB method is aimed at combining the qualities of interpretability and compactness of valence bond wave functions with a quantitative accuracy of the energetics. The fundamental feature of the method is the freedom of the orbitals to be different for each VB structure during the optimization process. In this manner, the orbitals respond to the instantaneous field of the individual VB structure rather than to an average field of all the structures. As such, the BOVB method accounts for the differential dynamic correlation that is associated with elementary processes like bond forming/breaking, while leaving the wave function compact. The use of strictly localized orbitals ensures a maximum correspondence between the wave function and the concept of Lewis structure, and makes the method suitable for calculation of diabatic states. 

Marcus model and diabatic states make more immediate (even if not strictly necessary) the introduction of the dynamical solvent coordinate, of which in Section 5.2 we have given a definition based on parameters of the continuum model, but other definitions are possible. Actually, the Russian school used this concept without giving formal definitions, at the best of our knowledge (several papers have been published in relatively minor Russian journals, with limited circulation in western countries in those years), and basing it on a description of the solvent as a continuum or a set of os-... 

However, this strict (and useless) definition can be applied only in the case of a complete set of diabatic states. When the study is restricted to a finite number of states of interest, this rule does not apply and the diabatic states must depend on the nuclear coordinates. They can be obtained from the corresponding adiabatic states through a imitary transformation ... 

In general, no strictly diabatic basis exists in a finite set g of interacting electronic states.To see this, let us return to Eqs. (27b) and (32) and... 

Owing to Eq. (35), there is no reason to expect that a strictly diabatic basis exists. Nevertheless, one can construct quasidiabatic states which are extremely useful in solving and understanding many relevant problems abundantly discussed in the literature. With their help it is possible to remove a substantial part of the derivative couplings and make the group-Born-Oppenheimer Eq. (26) more transparent and better amenable to explicit numerical calculations. That part of the derivative couplings which can be removed by an unitary transformation U( (R) is called... 

Apart from special models, Eq. (14) is satisfied only when the whole space of interacting electronic states is considered in the ADT matrix S (in diatomics, with only a single nuclear degree of freedom, there is no curl-condition). This, however, is contradicting the spirit of choosing a small subset of electronic states in the ADT matrix and would lead one back to the crude adiabatic basis discussed in the previous section. Therefore strictly diabatic electronic states, satisfying rigorously Eqs. (12) and (13) do not exist in the multidimensional case. ... 

---