Dear all,
I am working on testing excitation of very simple molecules such as Ag dimer, and examining both occupied and unoccupied states calculated. However, I find when I change the basis from SVP to TZVPP, the energy of the lowest band increases a littble bit, then decreases if I furtherly change to QZVPP basis (all basis are from BASIS_MOLOPT_UZH collection, I set CUTOFF to 600 Ry, REL_CUTOFF 30 Ry):
basis SVP TZVPP QZVPP
e(1) -96.3005 -95.3219 -96.1675
energies are in eV.
To my knowledge if the completeness of basis set increases, the energy should be lower.
I also convert the pseudopotential to UPF file (by CP2K ATOM module) and perform pw calculation with Quantum ESPRESSO, I set the kinetic energy cutoff to 300 Ry but still find the energy is higher than any of what I obtained from CP2K. But in principle pw is a very huge basis set, so I get confused again...
basis PW
e(1) -94.8071
Very best wishes,
Yike