Task dft optimize failed

[yilin Zhao]

Hello there!

I tried to optimize an MgO slab for 3 layers with bottom layer fixed,trying to optimize and get single point energy with an selected basis and ecp.

The problem came with optimize not converge, I noticed that in 100 iters, the energy didn't climb down a hill steadily, but hopping after several downhill. I tried to use the convergence critria in "LOOSE" but with no noticeably help, I'll also attach LOOSE input and output below.

Is there something to improve in this circumstance? Or maybe this set is hard to converge?

I'll share my inputs and outputs with google drive:

https://drive.google.com/file/d/1-b0HXyAHiIOvftLk0Pz_bmtTlO182ivP/view?usp=sharing, https://drive.google.com/file/d/12kr1zkQAkDXGcZd_UdRLnGoAsifxTaLZ/view?usp=sharing, https://drive.google.com/file/d/1IOE__NdmWKkRwjiQ2pJQG-G5_RH5xCKC/view?usp=sharing, https://drive.google.com/file/d/1h4GAwwXQk5aOq0yKvBsoxJhDPFTaUShG/view?usp=sharing

[yilin Zhao]

I relink these files below:

mgo_clean_default.inp

mgo_clean_default.out

mgo_clean_loose.inp

mgo_clean_loose.out

[Edoardo Aprà]

The NWChem gaussian basis DFT code does not support periodic boundary conditions.

Unfortunately, NWChem version 7.02 (and earlier) does not stop when these conditions are met. A bug fix is present in the master branch. https://github.com/nwchemgit/nwchem/issues/322

The only option in NWChem for periodic boundary condition is to use the plane-wave code https://nwchemgit.github.io/Plane-Wave-Density-Functional-Theory.html

Same topic discussed in the following threads

https://groups.google.com/g/nwchem-forum/c/CYQeUO-zZv0/m/umdMU0dRBQAJ

https://groups.google.com/g/nwchem-forum/c/7mkuxBfbDzM/m/IOBkSwivAQAJ