CCL:G: Unrestricted vs. restricted DFT for closed-shell metal complex

 Sent to CCL by: Susi Lehtola [susi.lehtola,alumni.helsinki.fi]
 On 07/01/2017 09:56 PM, Andrew Scott Rosen rosen^u.northwestern.edu wrote:
 
 Sent to CCL by: "Andrew Scott Rosen" [rosen]^[u.northwestern.edu]
 Hello,
 I have a question regarding the choice of unrestricted versus restricted
 calculations in DFT. Let's say that I am dealing with a Cu(I)
 organometallic complex that I anticipate would have a spin multiplicity of
 1 due to the assumed [Ar]3d10 electron configuration of Cu(I) in this
 complex. Under this assumption, the system is not open-shell. As such, it
 seems that a restricted calculation would be most appropriate.
 Would there be any benefit in running an unrestricted calculation (still as
 a singlet) in this case? Additionally, what would an unrestricted
 calculation on a system that's a singlet even mean, especially if I don't
 use specific keywords (e.g. in Gaussian) to try and obtain a broken-
 symmetry state?
If you run spin-unrestricted DFT with a spin-restricted guess, you'll get the same solution as with spin-restricted DFT, only getting it will be 2-3x more expensive. This is because there's nothing that will break the alpha-beta spin symmetry: a restricted guess gives identical alpha and beta Fock matrices, and consecutively identical new alpha and beta orbitals. 
 --
 -----------------------------------------------------------------------
 Mr. Susi Lehtola, PhD             Chemist Postdoctoral Fellow
 susi.lehtola##alumni.helsinki.fi   Lawrence Berkeley National Laboratory
 http://www.helsinki.fi/~jzlehtol  USA
 -----------------------------------------------------------------------