Partial waves matching radii for PAW dataset generation

Boris · Post by **Boris** » Tue Apr 19, 2011 12:52 am

Dear all,

I looked on the abinit website and I noticed that for a number of elements, matching radii are set to a bit lower than the PAW radius. For instance, in the cerium dataset, the PAW radius is 2.9 and all partial waves matching radii are set to 2.85. I did that for all my PAW dataset and it gave fairly reasonable results for a number of compounds. By curiosity, I looked at the PAW datasets from the vasp code, and the partial waves matching radii sometimes differ a lot from the PAW radius. Usually, for s- and p- partial waves, it is significantly lower.

So I would like to know whether the partial waves matching radii have a significant influence on the accuracy of the results or whether it is just a parameter that one has to tune in order to make more efficient dataset, in the sense that it will decrease the number of plane waves used in an abinit calculation.

Thank you for your help

Boris