numerical solution methods
Once we are able to write down the Kohn-Sham equations with a proper expression for the XC-functional, these equations still have to be solved. In this video we’ll discuss some common features of methods that are available for this technical task. You will get a global understanding about basis set methods and the selfconsistency problem, and you’ll get a list of codes that are made to solve the Kohn-Sham equations.
The next video elaborates on the features of two important classes of numerical solutions methods: all-electron methods and pseudopotential methods.
Optional: If you want to learn more about how the procedure of solving the Kohn-Sham equations is cast into matrix algebra (the HC=SCE equation), you can study this video:
On this Wikipedia page, you have an overview of many academic and (semi-)commercial codes that can solve (post-)Hartree-Fock or Kohn-Sham equations. You may sort the table according to the columns HF, post-HF or DFT, and see what’s available. The code that will be used for hands-on exercises in this course is Quantum Espresso (locate it in the table, and have a look at the features it has and hasn’t compared to other codes).
Optional reading resources: review articles that describe some of the more famous general-purpose DFT codes (access to these articles may depend on your institution):
- Quantum Espresso (paper from 2017)
- ABINIT (paper from 2020)
- CRYSTAL (paper from 2023, open access)
- VASP (paper from 2008)
- WIEN2k (paper from 2020, open access)
Another optional resource is the analysis of a community-wide survey (2022) among developers and users of DFT codes, with comments on many different aspects of the codes, their features, their spread and their future.
Finally, fill out this form to tell us which DFT or quantum chemistry codes you used before (if any). This helps us understanding the background of people interested in this course.
You can mark this topic as completed after having watched the video, and after the countdown timer has reached zero.
expected time: 25m
report time spent (page code AW03G)