Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Density Functional Theory

  1. Nov 27, 2011 #1
    Is there someone's research related to DFT?
    I'm an undergraduate trying to get into it.
    I hope I could get some help hear if I have any question about that!
    Thanks a lot,Euphemia
  2. jcsd
  3. Nov 27, 2011 #2


    User Avatar
    Staff Emeritus
    Science Advisor

    What would one like to know about DFT?

    The field has been developing over the last 4 or 5 decades.
  4. Nov 28, 2011 #3
    Ya, you're probably right. But I think the theory still not complete yet!
    There are a lot of fine models, and none of them is really based on the first principle.
    So I guess there is still something I can do!
  5. Nov 28, 2011 #4


    User Avatar
    Science Advisor

    Actually, almost all popular density functionals are either directly derived from first principles (e.g. PBE, TPSS) or based on a mixture of physically meaningful ideas with parameter fitting to accurate reference data (e.g., the LYP correlation functional or the mixing factors in B3LYP).

    I don't want to discourage you, but there has been no major breakthrough in (ground state) density functional theory during the last about 10-20 years. These years have seen many very good ideas (most of them based on first principles) which turned out to not actually work any better than PBE or the B3LYP thing. Recent research has actually mostly focused on patching up holes in DFT, like the dispersion problem, but this is not necessarily done in a terribly elegant way; the main goal is typically to get something which more or less works in practice in 90% of the cases. Progress in fundamental problems like a systematic improvability of functionals or the static correlation problem has been very slow. And there are some reasons to believe that obtaining significant further advances in these areas in the context of pure density functional theory is likely impossible.
    Many major DFT gurus have actually given up on that and are now making DFTs which are getting closer and closer to wave function methods... (e.g., optimized effective potential methods, random phase approximation correlation, range-separated hybrids, etc.)
  6. Nov 29, 2011 #5
    That's a fair point.
    Thank you for your notification.
    I am fascinated by the original idea of DFT, but this theory turns out not really work very well and become very messy. I still want to give it a shot! Since I'm still an undergraduate,
    there is no need to be hurry to settle down.
  7. Nov 29, 2011 #6
  8. Nov 29, 2011 #7


    User Avatar
    Staff Emeritus
    Science Advisor

    Apparently Density-Functional Theory of Atoms and Molecules (International Series of Monographs on Chemistry) by Robert Parr and Yang Weitao is recommended.

    These might be of interest -

    Implementation and Application of Advanced Density Functionals



    I suspect methods depend on applications, e.g., I believe applications would be different for molecules found in biological systems than those for metals, intermetallics, alloys and ceramics.

    I'm becoming more familiar with DFT through some projects at work.
    Last edited: Dec 5, 2011
  9. Nov 30, 2011 #8
    Thanks a lot
    You are really thoughtful!
  10. Dec 5, 2011 #9
    bcbwilla: nice :) i also did research using DFT as an undergrad , but i used DFT methods to generate some trial WF to use on monte carlo methods.
    Euphemia: try bcbwilla's suggestion, i found the book at least comprehensible, Density Functional Theory: A Practical Introduction, by scholl.
  11. Dec 19, 2011 #10
    Yep, DFT seems to be stucked now, but that also means that you have the chance to make a great breakthrough. It's extremely hard, though. Try it if you really like it!
  12. Dec 22, 2011 #11


    User Avatar
    Science Advisor

    I'd like to add that even though density functional theory is more or less stuck at the moment, it still is the most important electronic structure theory out there. Even small progress is important -- this is how things evolve.

    Also note that the closely related field of wave function theory, which attempts to achieve basically the same thing as DFT but with a different theoretic approach, is very alive, and has seen immense progress in the recent years. (Wave function methods is what DFT originally replaced in many applications). Contrary to DFT, in those fields wave functions are explicitly constructed from first principles[*], and there are systematic ways to improve the approxmations. If you like first priciples, that might be the right way for you.

    [*] (which turns out to be far less impossible than people would make you believe. At this moment it is quite possible to calculate an explicitly correlated coupled cluster wave function with a decent augmented triple-zeta basis set for a 100-atomic molecule.)
  13. Mar 21, 2012 #12
    Could someone please send me pages 174 and 175 from YAng Parr density functional theory of atoms and molecules 1989?? I really need them.
  14. Mar 22, 2012 #13
    Old thread I know, but I was just at the APS meeting a few weeks ago and it has me thinking about this. There are in fact some methods where you can do some pretty high level wave function stuff on some big systems, which is GREAT. It's not easy though and at least at this point the resources needed aren't readily available to all scientists. For most calculations of most large and relevant systems, DFT is the only game in town.

    There are ways of looking at it where DFT is a very elegant theory (solving an alternative problem to the SE) particularly when you see how accurate it can be for a large number of large systems. Of course, a lot of it rests upon fortuitous cancellation of errors and fitting to experimental parameters, which is decidedly in-elegant.

    My point is, it seems like everyone is trying to take shots at DFT nowadays, because it's a big and easy target. But, as someone pointed out above, it's the most important electronic structure method of the last thirty years.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook