I want to know how the analytical solutions of Hartree-Fock can come handy during computation of large systems although the formal scaling order of HF is N^4 compared to N^3 scaling of DFT (where N is an approximate measure of system size). I also want to know how it addresses the convergence problems that might arise with use of DFT for such large systems which can lead to an incorrect description of HOMO-LUMO band gaps and hence convergence issues.
Hi Henk,
I should have added that I am a computational chemist and I posted the question with relevance to the use of Gaussian program. My question was that what advantages one gets on using Hartree-Fock over DFT for large molecular systems, which tend to be infeasible to be calculated with DFT as the system size increases due to the problem of coming closer of HOMO-LUMO band gap description.
Dear Viki,
HF and DFT is a NP-complete problem. Therefore, for large systems(more 1000 atoms) use these methods very difficult.
DFT exchange interaction counts better than HF for many systems, where HF may have error ~ 15% and more.
Best regards, Eugene Kolyvoshko.
To add on to Henk's statement, I came across convergence difficulties for DFT-based methods like BLYP and B3LYP for larger systems which were tackled efficiently by HF.
- Badges
- Science topic
- Similar topics
- Chemistry
- Theoretical Chemistry