Sorry, for wasting bandwidth, but in my last post to CCL. I forgot to mention, that if the metal-metal bond distance was fixed to the experimentally observed shorter distance, the energy was found to be ca. 5 kcal/mol higher than the geometry optimized structure with the longer M-M distance. Also, I forgot to mention that I tried polarisation functions as well (also numerical basis sets i.e. Dmol). I have not tried STO's yet. Perhaps I should give ADF a try?
The answer is a definite YES. Furthermore, did you also include relativistic effects ? I guess when you look at Tungsten/Molybdenum, these can be quite significant.Again this is a YES to ADF, since in this program these effects are easily taken into account.
Marcel.