Please use this identifier to cite or link to this item:
|Title:||Very accurate ab initio binding energies: a comparison between empirical corrections and extrapolation methods||Authors:||MARTIN, Jan||Issue Date:||1997||Publisher:||Elsevier Science B.V.||Source:||JOURNAL OF MOLECULAR STRUCTURE-THEOCHEM, 398-399. p. 135-144||Abstract:||The experimentally very accurately known total atomization energies (TAEs; Σ De values) of 14 small polyatomic molecules have been computed at the CCSD(T) level using basis sets of up to [7s6p5d4f 3g2h/5s4p3d2f1g] quality. Core correlation was corrected for through separate calculations using a special core-correlation basis set. A comparative study of extrapolations to the infinite-basis set limit and empirical correction treatments for basis set incompleteness was made. Up to basis sets of [spdfg] quality, the Martin 3-parameter empirical correction formula significantly outperforms extrapolation techniques based on the asymptotic l-dependence of the correlation energy: for still larger basis sets, the two approaches yield comparable results. The effect of core correlation can be fairly well absorbed in the parameters of the correction, but for very large valence basis sets, it becomes worthwhile to treat it explicitly. 3-parameter-corrected CCSD(T)/aug'-cc-pVQZ calculations represent a particularly good compromise between computational cost and accuracy (mean absolute error 0.23 kcal mol−1; maximum absolute error 0.63 kcal mol−1).||Keywords:||Basis set convergence; Empirical corrections; Coupled cluster; Very accurate thermochemistry||Document URI:||http://hdl.handle.net/1942/7544||DOI:||10.1016/S0166-1280(96)04932-9||Type:||Journal Contribution|
|Appears in Collections:||Research publications|
Show full item record
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.