ABINIT Discussion Forums

Hi,

I'm performing some calculations on the cohesive energies for a few metals like copper and zinc, and I seem to be getting results that converge close to the experimental lattice constant values when I use toldfe and not when I use toldff. The calculated acell values when using toldff differ by more than 30% from the experimental values, and in some cases, I can't even get convergence at all.

However, the help file for the toldfe variable also mentioned 'When the geometry is optimized (relaxation of atomic positions or primitive vectors), the use of toldfe is to be avoided. The use of toldff or tolrff is by far preferable, in order to have a handle on the geometry characteristics.' Can I know how I should set the values of toldff so that I can get convergence at close to experimental values? I'm currently setting toldff at 5.0d-6.

Also, for convergence studies of metals, is it be preferable to optimize the volume only (optcell=1, changing acell only and not rprim) or is it necessary to do full optimization of cell geometry (optcell=2, changes both acell and rprim).

Thanks a lot in advance, I appreciate it.

"It is STRONGLY suggested first to optimize the ionic positions without cell shape and size optimization (optcell=0), then start the cell shape and size optimization from the cell with relaxed ionic positions." - maybe for metals too. Maybe for metals full optimization is not required (but only if it is cubic), but optcell=0 is necessary (maybe )

to check required convergence criteria perform convergence of your (toldfe or toldff - as for me I usually use toldfe) on forces (use static multidataset runs and look at *.out) - if it is not converged at even toldfe 10^-12, then use tolvrs 10^-22. The same procedure for convergence of kpoints grid on forces is also required.

Best wishes.