CCL: Computing AsMe3

[Gerald Knizia <knizia*theochem.uni-stuttgart.de>]

 Sent to CCL by: Gerald Knizia [knizia---theochem.uni-stuttgart.de]
 On Monday 04 October 2010 11:51, osama T Akoubeh dolkeen(_)hotmail.com wrote:
 > I computed the compound AsMe3 " As(CH3)3 " using PM5 and many
 functional
 > methods to get the enthalpy of formation , entropy at various temperature
 > to study its behavior reaction [..]
 >
 > The value from Chemkin data base is 2,9kcal/mol , the difference between
 > both will make the reaction Gibbs favorable at 2.8 and not at 2.9 kcal/mol
 >
 > I computed the compound using PM5 methaod anf got 3.2118 kcal/mol what
 > functional method you suggest to get 2.8 even I tested B88-PW91 with no
 > good results
 This kind of accuracy is far beyond what you could possibly expect from any
 DFT or semiempirical method. In fact, I doubt that there currently are any
 computational methods capable of achieving this kind of accuracy for this
 system, not even for the electronic part alone.
    The least you'd need to even get reliably into the 1kJ/mol accuracy range
 is a compound method (like W4 or HEAT) which includes at least a CCSDT(Q)
 calculation and additionally a F12 or CBS[45] (better CBS[56]) calculation of
 the CCSD(T) basis set limit. My patience would certainly not be sufficient to
 do a CCSDT(Q) calculation on such a system.
 --
 Gerald Knizia