Skip to content

1.4 GPW and all-electron GAPW

These are unexecuted teaching inputs and starting models. Original diagrams are schematics, not calculated results. Validate version-specific syntax, licensed or authorized data, numerical convergence and the scientific model before using this workflow.

1.4.1 Model, units and provenance

Keep basis/potential files and executable versions traceable. Grid controls use Ry; common energy/force outputs use hartree and hartree/bohr. Read the unit in each output heading.

Shared inputs, conventions and evidence

Original schematic: GPW and all-electron GAPW. No numerical results are claimed.
Original schematic: GPW and all-electron GAPW. No numerical results are claimed.

1.4.2 Unexecuted inputs and explicit deltas

Use the accompanying instructions to identify the parent calculation and placement of every delta; a snippet is not automatically a standalone input. Preserve all blank-line and file-provenance requirements.

1.4.2.1 Input block 1

! DFT/QS replacement
&QS
  METHOD GAPW
  EPS_DEFAULT 1.0E-12
&END QS
! Example KIND replacement only after checking the installed record
&KIND O
  ELEMENT O
  BASIS_SET 6-31G*
  POTENTIAL ALL
&END KIND

1.4.3 Worked investigation

Intuition. Smooth pseudodensities and sharply varying core densities place different demands on numerical representations. GAPW partitions the difficult local contribution rather than requiring the global grid to resolve every core oscillation. It is not “GPW with a better cutoff,” and it is not interchangeable with a PAW input from a plane-wave program. This case teaches a controlled model change using one water molecule, not a core-spectrum prediction. Prerequisites: cases 01–03 and access to a verified all-electron basis record in the release data.

Input delta and applicability. Replace DFT/QS/METHOD GPW by METHOD GAPW. For an all-electron oxygen calculation, change its potential to ALL and its orbital basis to an actual all-electron record. A common teaching choice is BASIS_SET 6-31G* from BASIS_SET, but confirm that exact element/name exists in your installation. Add or replace the data-file declaration with BASIS_SET_FILE_NAME BASIS_SET; set POTENTIAL_FILE_NAME POTENTIAL if that installed file supplies ALL. Apply a compatible all-electron basis and POTENTIAL ALL to H as well for this controlled comparison. These are selection instructions, not permission to assume every release packages identical records.

See input block 1 above.

Workflow. 1. Inventory the all-electron records and preserve their source file. 2. Run the all-electron calculation at A's geometry, with a dedicated project and no incompatible old wavefunction restart. 3. Verify that the explicit electron count has changed from the pseudopotential model: all-electron neutral water has ten electrons. 4. Repeat a bond-stretch energy difference in both models, using internally consistent references. 5. Tighten the grid and the documented GAPW numerical controls one category at a time, especially if investigating core-sensitive quantities. 6. Only after these tests consider larger systems or mixed all-electron/pseudopotential kinds.

Output interpretation. The all-electron total energy includes core contributions absent from the valence-only model; its numerical value cannot be treated as an improvement by subtracting it directly from the GPW energy. Compare energy differences, forces or structures on the same physical question. A small difference between the two models could be fortuitous if the basis families have very different quality. The pedagogical objective is to separate representation, potential and basis, not to crown an automatic winner.

Tests and pitfalls. Verify translation/rotation stability of the molecule and the finite-difference force test. For all-electron calculations, quadrature around atoms adds controls beyond the global grid; do not conclude convergence from CUTOFF alone. Mixed AE/PP calculations exist but complicate interpretation and are not the beginner baseline. GAPW capability does not by itself turn a ground-state run into an X-ray spectrum: transition model, core hole, states and response approximation need separate validated setup.

Exercise. Construct a comparison ledger with electron count, core treatment, orbital basis, numerical representation, observable and reference energy convention. Explain which entries must match for a meaningful comparison and which deliberately differ.

Diagram. Split panel: smooth density on a global mesh versus smooth background plus compact atom-centered sharp density. EN labels: valence model, core region, smooth contribution, local augmentation, same physical observable. Use qualitative curves only.

1.4.5 Sources and further reading