Semiempirical theories of bonding in molecules containing transition metals are often merely interpretive. That is, given the experimental fact that a molecule exists and has a particular shape, theory is used to provide a plausible ad hoc justification. Much interpretation of this type uses molecular orbital theory1 and ascribes bonding effects to orbital mixing. Earlier approaches such as ligand-field theory have been more or less abandoned along with the alternative insights they provided. Modern ab initio methods, on the other hand, have the possibility of being quantitatively predictive. They can be used to study the relative energies of the observed structure and nonobserved plausible alternatives. They can also be used to examine * Author to whom correspondence should be addressed. i Deceased, March 28, 1991.

No abstract

No abstract

2129ometry drastically. The proton approaches oxygen in the C2, symmetry, while the optimal geometry of OH3+ is C3". From the study of geometry variation along the R(N-0) bond the conclusion can be drawn that the proton always belongs to the one subunit or the other, the areas of "sharing" proton are very narrow.Orbital pictures for extreme cases (2.707 and 5.0 A) are significantly different, which may affect the structure of potential surfaces. The correlation between orbital pictures of separated systems and the complex is obvious for the long N -0 distance. However an intermediate region is recognized also. The orbital pictures are more complicated for shorter distances (2.7-3.2 A).In the case of R(N-0) from 2.707 to 3.2 A the ground state is purely one-determinantal. The gap between ground and excited states is too large to allow interactions. In the case of long distances (5.0 A) the function for the ground state is contaminated by configurations characteristic for the "avoided" region. The gap between ground and excited states is much smaller than for shorter distances, and the interactions are much bigger. The curve in regions around atoms N or 0 resembles the situation in separate molecules. Energy potential curves calculated by coupled cluster method agree well with those from MRD-CI calculations. However, the multireference character of the wave function a t N-H+ distance of 2.9 A (for N -0 distance 5.0 A) is manifested by lack of convergence for the excited singlet state, while starting from a single determinantal ground-state function.Mulliken gross atomic population indicates that the reaction proceeds as a proton transfer, while at excited states the proton carries more electron density. Similarly to the geometry optimization, in the ground state the proton belongs to OH2 or NH3 and not to a "common" zone. This picture supports the concept of proton transfer as a tuneling between two structures. Electrons in excited states are much more delocalized, and the proton carries an electron acquired partially from ammonia and partially from water. Electrostatic properties of the excited-state systems are completely different from that of the ground state. Acknowledgment.We investigate the molecular orbital self-consistent-field model of bonding in Cr(CO),. The energetics and electron density are examined using a large range of tools. The change in density compared to a promoted 3dt22 'A,, chromium atom and six CO molecules is primarily charge transfer from the tlg orbitals of chromium to the empty 2r*t2, orbital of the (CO), cage. This mixing is counterintuitive, as the largest increase in electron density is in the oxygen p r orbitals. The restricted Hartree-Fock energy is actually repulsive compared to that of ground-state fragments by 11 1 kcal/mol. This energy change consists of +266 kcal/mol of fragment promotion energy, +67 kcal/mol of (CO), cage formation energy, -272 kcal/mol of electrostatic attraction, +359 kcal/mol of overlap repulsion between Cr and (CO),, and finally -329 kcal/mol of orbital r...

No abstract