2023-09-10 16:31:49 +02:00
|
|
|
# Pseudopotentials
|
|
|
|
|
|
2026-04-26 12:02:00 +02:00
|
|
|
Most GPW calculations in CP2K use norm-conserving Goedecker-Teter-Hutter ({term}`GTH`)
|
|
|
|
|
pseudopotentials. A pseudopotential removes chemically inactive core electrons from the explicit
|
|
|
|
|
electronic problem and represents their effect on the valence electrons through an effective
|
|
|
|
|
potential. This reduces the number of electrons and avoids the very hard core density that would
|
|
|
|
|
otherwise require extremely fine grids.
|
2023-09-10 16:31:49 +02:00
|
|
|
|
2026-04-26 12:02:00 +02:00
|
|
|
Pseudopotential files are selected in
|
|
|
|
|
[POTENTIAL_FILE_NAME](#CP2K_INPUT.FORCE_EVAL.DFT.POTENTIAL_FILE_NAME), and the actual potential is
|
|
|
|
|
selected for each atomic [KIND](#CP2K_INPUT.FORCE_EVAL.SUBSYS.KIND) with
|
|
|
|
|
[POTENTIAL](#CP2K_INPUT.FORCE_EVAL.SUBSYS.KIND.POTENTIAL):
|
|
|
|
|
|
|
|
|
|
```text
|
|
|
|
|
&FORCE_EVAL
|
|
|
|
|
&DFT
|
|
|
|
|
POTENTIAL_FILE_NAME GTH_POTENTIALS
|
|
|
|
|
&END DFT
|
|
|
|
|
&SUBSYS
|
|
|
|
|
&KIND O
|
|
|
|
|
POTENTIAL GTH-PBE-q6
|
|
|
|
|
&END KIND
|
|
|
|
|
&KIND H
|
|
|
|
|
POTENTIAL GTH-PBE-q1
|
|
|
|
|
&END KIND
|
|
|
|
|
&END SUBSYS
|
|
|
|
|
&END FORCE_EVAL
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
The suffix `q6` in `GTH-PBE-q6`, for example, means that six valence electrons are treated
|
|
|
|
|
explicitly. The chosen basis set should match this valence configuration; for oxygen, a common
|
|
|
|
|
matching basis is `DZVP-MOLOPT-GTH`.
|
|
|
|
|
|
|
|
|
|
## Choosing a Pseudopotential
|
|
|
|
|
|
|
|
|
|
Use a pseudopotential generated for the exchange-correlation functional family used in the
|
|
|
|
|
calculation. For example, `GTH-PBE-q6` is a natural choice for PBE calculations with oxygen. Mixing
|
|
|
|
|
functional families can be acceptable for exploratory work in some cases, but it is not a systematic
|
|
|
|
|
route to high accuracy.
|
|
|
|
|
|
|
|
|
|
The CP2K data directory contains several pseudopotential libraries:
|
|
|
|
|
|
|
|
|
|
- `GTH_POTENTIALS` contains widely used GTH potentials for common GPW calculations.
|
|
|
|
|
- `POTENTIAL_UZH` contains the UZH protocol GTH potentials designed to be used with matching UZH
|
|
|
|
|
basis sets.
|
|
|
|
|
- `NLCC_POTENTIALS` and `GTH_SOC_POTENTIALS` contain more specialized potentials.
|
|
|
|
|
- `ECP_POTENTIALS` contains effective core potentials for Gaussian integral based calculations.
|
|
|
|
|
|
|
|
|
|
For new GPW production inputs, prefer a matching UZH protocol pair from `POTENTIAL_UZH` and
|
|
|
|
|
`BASIS_MOLOPT_UZH` when it is available for the element and functional family. The older
|
|
|
|
|
`GTH_POTENTIALS` library remains important for reproducing established calculations and for cases
|
|
|
|
|
where a matching UZH setup is not available.
|
|
|
|
|
|
|
|
|
|
For all-electron calculations, use `POTENTIAL ALL` together with an all-electron basis set and the
|
|
|
|
|
GAPW method:
|
|
|
|
|
|
|
|
|
|
```text
|
|
|
|
|
&KIND O
|
|
|
|
|
BASIS_SET SVP-MOLOPT-GGA-ae
|
|
|
|
|
POTENTIAL ALL
|
|
|
|
|
&END KIND
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
## Consistency Checks
|
|
|
|
|
|
|
|
|
|
Useful checks when setting up a calculation are:
|
|
|
|
|
|
|
|
|
|
- The basis set and pseudopotential should be available in the files named in the `DFT` section.
|
|
|
|
|
- The pseudopotential valence charge should match the basis set suffix where such a suffix is used.
|
|
|
|
|
- The exchange-correlation functional should be consistent with the pseudopotential family.
|
|
|
|
|
- For heavy elements, decide whether a large-core, medium-core, small-core, or all-electron
|
|
|
|
|
description is appropriate for the property of interest.
|
|
|
|
|
|
|
|
|
|
For a tested minimal GPW input using GTH pseudopotentials, see
|
|
|
|
|
[](../../getting-started/first-calculation).
|
|
|
|
|
|
|
|
|
|
## See Also
|
2023-09-10 16:31:49 +02:00
|
|
|
|
|
|
|
|
- <https://en.wikipedia.org/wiki/Pseudopotential>
|
|
|
|
|
- <https://cp2k.org/static/potentials/>
|
|
|
|
|
- <https://www.cp2k.org/tools:cp2k-basis>
|
|
|
|
|
- [](#Goedecker1996)
|
|
|
|
|
- [](#Hartwigsen1998)
|
|
|
|
|
- [](#Krack2005)
|
2026-04-26 12:02:00 +02:00
|
|
|
- [](#Iannuzzi2026)
|