Ba. Smart et Ch. Schiesser, ON THE EXISTENCE OF SH3, SEH3, AND TEH3 - DISCREPANCIES BETWEEN ALL-ELECTRON AND PSEUDOPOTENTIAL CALCULATIONS, Journal of computational chemistry, 16(9), 1995, pp. 1055-1066
Ab initio calculations using both pseudopotential and double and tripl
e-zeta all-electron basis sets, with and without electron correlation
(MP2, QCISD), have been performed on the lambda(4)-sulfanyl (SH3), lam
bda(4)-selanyl (SeH3), and lambda(4)-tellanyl (TeH3) radicals. All-ele
ctron basis sets of double-zeta quality predict that SH3 and SeH3 corr
espond to transition states on their respective potential energy surfa
ces. In contrast, the pseudopotentials of Hay and Wadt predict that SH
3 and SeH3 correspond to local minima at the QCISD level of theory whi
le the pseudopotentials of Christiansen and Stevens predict transition
states. By comparison, TeH3 proved to be a local minimum at all level
s of theory. Interestingly, when a very large (triple-zeta) all-electr
on basis set was used, SH3 proved to be a transition state; however, i
n this instance the potential energy surface was found to be much flat
ter than in the case for which a double-zeta basis set was used, sugge
sting that further improvements in the basis set may lead to a local m
inimum. Further improvements in the all-electron selenium basis also l
ed to a local minimum for SeH3 at the QCISD level of theory. (C) 1995
by John Wiley & Sons, Inc.