Density Functional Theory

Walter Kohn and John Pople shared the 1998 Nobel Prize in Chemistry for their work in Density Functional Theory (DFT). The computer programs that actually solve the DFT equations have a lot in common with Hartree-Fock programs; quite frequently theorists simply adapt their Hartree-Fock program to do DFT computations. Like Hartree-Fock, the DFT equations must be solved self-consistently, making DFT another type of SCF method. In the Q-Chem program [4], both Hartree-Fock and DFT are done using a module called scfman.

Instead of focusing on wave functions and orbitals, DFT focuses on the electron density (although it usually employs orbitals to get the density). It includes an approximate treatment of electron correlation and therefore should be more accurate than Hartree-Fock theory. There are actually very many different DFT methods, depending on the particular treatment of correlation or ``exchange'' (which we will unfortunately leave undefined for this lab). Feel free to experiment with some of the different DFT approaches in the lab.

The drawback of DFT is that nobody knows how to take a given DFT computation and improve it. This contrasts with all other ab initio methods, where an expert can always tell how to keep improving the results until the electronic Schrödinger equation is solved exactly. The reason anyone uses DFT is that it tends to give very accurate results much more cheaply than some competing methods.