Skip to content

4.3 Noncovalent interaction and dispersion choices

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.

4.3.1 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: Noncovalent interaction and dispersion choices. No numerical results are claimed.
Original schematic: Noncovalent interaction and dispersion choices. No numerical results are claimed.

4.3.2 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.

4.3.2.1 Input block 1

&VDW_POTENTIAL
  POTENTIAL_TYPE PAIR_POTENTIAL
  &PAIR_POTENTIAL
    TYPE DFTD3
    PARAMETER_FILE_NAME dftd3.dat
    REFERENCE_FUNCTIONAL PBE
  &END PAIR_POTENTIAL
&END VDW_POTENTIAL

4.3.3 Worked investigation

Intuition. A hydrogen-bonded dimer combines electrostatics, induction, exchange repulsion and dispersion. Adding a dispersion correction changes the energy model, not just the convergence settings. Use two water molecules with fixed internal geometry, a converged nonperiodic box and a basis study. A rigid scan separates interaction shape from monomer deformation.

Original setup. Put one copy of A near (7,10,10) angstrom in a 24 angstrom cubic box. Orient a second water so that one O–H vector points toward the first oxygen; document the rotation matrix and use O–O distances 2.6, 2.9, 3.2, 3.6 and 4.2 angstrom. These are sampling positions, not predicted equilibrium distances. Evaluate pure PBE and, separately, PBE plus one documented D3 damping variant. A common zero-damping D3 input delta under DFT/XC is:

See input block 1 above.

Verify the installed dftd3.dat and the version-specific TYPE option. D3 and D3(BJ) are distinct parameterizations; do not rename one as the other or add both simultaneously.

Workflow. 1. Generate and visually inspect all dimers for the same orientation. 2. Compute isolated monomer energies with identical geometry and model settings. 3. Form rigid interaction energy Eint(R)=EAB(R)−EA−EB for each model. 4. Print or extract the dispersion contribution if available and check that it is not counted twice in post-processing. 5. Repeat selected distances with a larger basis and counterpoise from case 16. 6. Only then relax the dimer and separately report deformation energy.

Interpretation and checks. A more attractive corrected curve does not prove greater agreement with experiment; validation requires an appropriate reference. The long-range interaction should approach zero under the stated finite-box and numerical tolerance, but a finite endpoint is not infinity. The short-range repulsive wall can expose basis or SCF problems. Gas-phase dimer binding energy is not a liquid-water free energy, and an optimized electronic minimum does not include entropy or zero-point motion.

Exercise and pitfalls. Compare the same correction at two basis sizes. Decide whether a change in apparent well depth comes mainly from dispersion or basis borrowing. Make a signed ledger that separates interaction at frozen geometry, monomer deformation, BSSE correction and optional ZPE. Never fold all of them into an unexplained “binding energy.”

Diagram. Two monomers with fixed internal bonds and an adjustable R arrow; below, separate qualitative components rather than a fabricated total curve. EN: rigid monomers, separation, model correction, deformation excluded.

4.3.5 Sources and further reading