CuO cell optimization - changes too big

[mdsimula...@gmail.com]

Hello,

I am attempting to cell_opt a 2x3x2 super cell of CuO using PBE-D3.  During the optimization, the cell parameters change drastically.  For instance, one cell angle changes from ~96 deg to ~90 deg.

I am wondering if I am setting up the system wrong as I have seen some literature that uses PBE and does not have massive cell changes.

I have attached my input files.

Thanks,

Frank

[Lucas Lodeiro]

I see three things. First, use higher CUTOFF. With a REL_CUTOFF of 60 and oxygen pseudo (higher gaussian exponent ~12) a CUTOFF of 720 is necessary. Second, you are usign the LBFGS optimizer with a little cell (just 96 atoms), it is better BFGS. And the third and the most important, CuO is antiferromagnetic, so you have to set the magnetization for Cu centers, no just the whole system multiplicity.

Maybe this paper is useful: https://journals.aps.org/prb/abstract/10.1103/PhysRevB.87.115111

Regards - Lucas

--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/949076b9-6c81-4bd3-872d-494efc37a791n%40googlegroups.com.

[mdsimula...@gmail.com]

Hi Lucas,

Thank you for the information!

I'm not too familiar with antiferromagnetic materials but it seems that I need to use broken-symmetry.  I've attempted to set it up in the attached input file.  Do things seem correct?

Thank you for your help,

Frank