Improve the pseudopotential extrapolation
The simple linear extrapolation that we use for the pseudopotential (to get the point at zero) is not very accurate. I have changed this for quadratic extrapolation. Two tests had to be updated.
Edited by Xavier Andrade