> I was wondering if any one could possibly share any rule of thumb or
> technique for choosing appropriate rc and rm values for each shell.
Hi Thomas,
- I'd recommend using 1 zeta for deep levels and 2 zetas for higher
valence.
- Then let SIESTA to generate the PAO.Basis - for the start choose a
fairly good EnergyShift (usually ~50meV is good)
- play around these vales and check how much total Energy goes down vs
changes in EnergyShift. (you have to check that Etotal becomes lower and
due to variational principle this means a better basis).
- Check the computation time and decide whether those changes worth the
spent calculation time
- checking the resulting accuracy of main parameters (lattice constant,
bulk modulus, cohesive energies) should help to get a better evaluation
- ideally one should optimize the basis with simplex. But in the
simplest case, you just need to adjust the cutoff radius only for the
polarization orbital (take the PAO.Basis from the .out file after
default basis generation).
By default SIESTA takes polarization orbital zeta radius equal to the
radius of the first or second zeta (I don't remember exactly, check it
in the manual or just generate a default basis and see the output) of
the previous orbital.
In fact, what I observe in all optimized bases that the radius of the
polarization orbital should be somewhere in between 1st and 2nd zetas of
the previous orbital, even slightly closer to the second zeta (again
check this with the optimized bases on SIESTA site).
So try to modify this only value - it would have the most dramatic
effect on the results than anything else.
Sincerely,
Alexander Vozny
