THEORETICAL-STUDY OF TUNGSTEN CARBONYL-COMPLEXES W(CO)(N)(-6) - STRUCTURES, BINDING-ENERGIES, AND IMPLICATIONS FOR GAS-PHASE REACTIVITIES()(N=1)

The electronic structure and geometry of W(CO)(n)(+) (n = 1-6) have be en studied at the B3LYP and ab initio levels. We find that the ground state of W(CO)(+) is linear with a sextet spin state, that a linear se xtet and a bent quartet are nearly degenerate for W(CO)(2)(+), and tha t doublet states are unambiguously the ground states of W(CO)(3)(+) to W(CO)(6)(+). Successive (CO)(n-1)W+-CO binding energies have been com puted to be larger than any of those previously determined for other t ransition metals, We compare our results with available experimental d ata. Electron transfers are very important: (i) sigma donation from th e CO's to the metal is found to be more favorable when involving 5d ra ther than 6p leading to a preference for the bent rather than linear s tructures for W(CO)(n)(+) (n = 2-4); (ii) pi back-donation plays a cru cial role in shaping these molecules. These effects provide the drivin g force for the spin changes as the number of ligands increases. Spin lowering is associated with an increasing number of doubly rather than singly occupied 5d(pi) metal orbitals, which enhances the back-donati on ability while reducing the repulsion between sigma metal electrons and CO lone pairs. On the basis of our results, we propose an interpre tation of the observed differences in gas phase reactivity of W(CO)(n) (+) with small hydrocarbons as a function of n. The rationale for this interpretation is that the initially formed (CO)(n)W+-(hydrocarbon) c omplex should either have a ground or a low-lying excited state bearin g at least two unpaired electrons on the metal to be able to further a ctivate the hydrocarbon efficiently.