An improved theoretical approach to the empirical corrections of density functional theory

Jenn Huei Lii, Ching Han Hu

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

An empirical correction to density functional theory (DFT) has been developed in this study. The approach, called correlation corrected atomization-dispersion (CCAZD), involves short- and long-range terms. Short-range correction consists of bond (1,2-) and angle (1,3-) interactions, which remedies the deficiency of DFT in describing the proto-branching stabilization effects. Long-range correction includes a Buckingham potential function aiming to account for the dispersion interactions. The empirical corrections of DFT were parameterized to reproduce reported ΔH f values of the training set containing alkane, alcohol and ether molecules. The ΔH f of the training set molecules predicted by the CCAZD method combined with two different DFT methods, B3LYP and MPWB1K, with a 6-31G * basis set agreed well with the experimental data. For 106 alkane, alcohol and ether compounds, the average absolute deviations (AADs) in ΔH f were 0.45 and 0.51 kcal/mol for B3LYP- and MPWB1K-CCAZD, respectively. Calculations of isomerization energies, rotational barriers and conformational energies further validated the CCAZD approach. The isomerization energies improved significantly with the CCAZD treatment. The AADs for 22 energies of isomerization reactions were decreased from 3.55 and 2.44 to 0.55 and 0.82 kcal/mol for B3LYP and MPWB1K, respectively. This study also provided predictions of MM4, G3, CBS-QB3 and B2PLYP-D for comparison. The final test of the CCAZD approach on the calculation of the cellobiose analog potential surface also showed promising results. This study demonstrated that DFT calculations with CCAZD empirical corrections achieved very good agreement with reported values for various chemical reactions with a small basis set as 6-31G *.

Original languageEnglish
Pages (from-to)199-213
Number of pages15
JournalJournal of Computer-Aided Molecular Design
Volume26
Issue number2
DOIs
Publication statusPublished - 2012 Feb 1

All Science Journal Classification (ASJC) codes

  • Drug Discovery
  • Computer Science Applications
  • Physical and Theoretical Chemistry

Fingerprint Dive into the research topics of 'An improved theoretical approach to the empirical corrections of density functional theory'. Together they form a unique fingerprint.

  • Cite this