1. Input file
3. Output files
This INPUT does contain a few new setup keywords. The super_cell keyword expands the unit cell in this case by the unit cell 5 times in the first lattice direction. The super_cell_vac adds 3 unit cells of vacuum to the calculation. The volume is scaled accordingly. The occupation smearing option for the states near the Fermi level is also introduced. The smearing_energy is in eV, and there are 6 options for the smearing method. See the User's Guide for more info. Another different parameter is that the diagonalization method is now Grassmann_metal.
The main feature illustrated here is the Pulay-Thomas-Fermi (PTF) mixing method. The typical mixing method of Pulay-Kerker (PK:the default) performs poorly for elongated inhomogneous systems (such as surfaces). Additional input of the number of initial PK steps before starting the PTF method is input by PTF_num_init_PK. Also the number of conjugate gradient steps for solving the Thomas-Fermi-von Weizsacker equations is input by PTF_max_iter_cg.
Using the PK mixing, convergence is not attained: OUT_PK To switch from Pulay-Kerker method comment out the mix_method pulay_kerker line and uncomment the next four lines. The PTF method does obtain convergence: OUT_PTF. The optimal setting seemed to be using max_iter_diag 5, but the default setting of three was not much worse. The additional lines after the energy analysis give the average correction potential so the Kohn-Sham equations are solved and the following lines give the eigenvalue, gradient, and TFW total energy. Note that your runs may not be identical to this one.