Some settings in your file are not reasonable or redundant. For example, ALPHA 0.01, is too aggressive. The settings to tune GAPW calculations are not necessary. Importantly, the cell is very small, and Pt slab it a conductor, considering k-point sampling is indispensable.

I recently tried to calculate the XPS binding energy of some O atoms on a Pt surface using CP2K. I used a big all electron basis for the interested O atoms, a smaller all electron basis for other O atoms, and GTH-MOLOPT basis for Pt with pseudopotential. However, the SCF calculation could not converge. I then tried the ground state calculation with a similar GAPW method and GPW method. For GAPW method, the SCF convergence also failed, and gave me an absurd energy. The GPW method converged after about 100 steps. Strangely, both calculations showed warnings about Fermi-Dirac smearing, although I used METHFESSEL_PAXTON smearing and 1000 ADDED_MO. Any idea on this issue? Any suggestions would be appreciated. Thank you.

Best regards,
Hengyuan