CP2K: what is still missing after the first version (#329) #412

The first version (#329, 9.3.0, experimental) runs an energy (optionally with forces) or a geometry optimisation with periodic Kohn-Sham DFT (GPW): PBE/revPBE/BLYP/BP86/LDA, D3, per-element basis and pseudopotential, cutoffs, spin, DFT+U, OT or diagonalisation with smearing. It parses energy, SCF steps, the optimisation trajectory, forces, Mulliken and Hirshfeld charges, spin moments and timing, with progress and failure reasons. This issue collects what is still missing. Develop against real inputs (one is a triclinic clay cell with DFT+U on Fe, read from a PDB, using @set variables).

Run types

• Cell optimisation (CELL_OPT), with the cell trace parsed and shown
• Molecular dynamics (NVE, NVT, NPT; thermostat, time step, steps), with energy/temperature/pressure plots and the trajectory
• Vibrational analysis (frequencies, IR intensities), into the Vibrational Frequencies panel
• NEB / reaction paths

Theory

• Hybrid functionals (PBE0, HSE06) with ADMM
• Non-periodic (molecular) systems: POISSON PERIODIC NONE with a wavelet or MT solver
• Implicit solvent (SCCS) if needed
• k-points

Structure and input

• Triclinic cells from a PDB or CIF (angles), checked against CP2K's cell
• Import an existing CP2K input: @SET/@include preprocessing, COORD_FILE, KIND blocks
• Restart: continue from a .restart file (also for an optimisation that hit its step limit)

Results

• Molecular orbitals and densities in the viewer (MO_CUBES, E_DENSITY_CUBE, or Molden output)
• DOS and PDOS (with Builder: overlap population density of states (OPDOS), with total and partial DOS #171)
• Band structure
• Final orbitals as a guess (Duplicate with Final Orbitals for more codes (9.4): Gaussian 16, NWChem, ECCE-QM, then MOPAC, CP2K, QE #407)

Checks

• Verify: basis/potential coverage of every element in the structure, cutoff sanity, DFT+U elements present

Order to decide; cell optimisation, MD and cubes are the likely first ones.