Skip to content

7.1 HF and DFT compare questions before numbers

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.

7.1.1 Model, units and provenance

Geometry in Å; electronic energy in hartree; vibrational wavenumbers in cm⁻¹. Check each printed field and keep thermal and standard-state terms distinct.

Shared inputs, conventions and evidence

Original schematic: HF and DFT compare questions before numbers. No numerical results are claimed.
Original schematic: HF and DFT compare questions before numbers. No numerical results are claimed.

7.1.2 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.

7.1.2.1 Input block 1

%chk=g25_water_hf.chk
%mem=2GB
%nprocshared=2
#p HF/cc-pVTZ Opt=Tight Freq SCF=Tight

Water HF model comparison own stationary geometry

0 1
O  0.000000  0.000000  0.000000
H  0.758000  0.000000  0.586000
H -0.758000  0.000000  0.586000

7.1.3 Worked investigation

7.1.3.1 Intuition and prerequisites

Comparing two methods is an experiment about an approximation, not a contest to produce the most negative total energy. Hartree–Fock and approximate Kohn–Sham DFT treat electron interaction differently; their absolute energies are not an interchangeable accuracy ranking. Water provides a controlled first comparison of geometry, dipole, and harmonic frequencies. A meaningful design separates a fixed-geometry comparison from a fully relaxed comparison. Prerequisites: cases 05, 08, and 10; you should already distinguish basis error, geometry error, and finite numerical tolerance.

7.1.3.2 Original worked input

See input block 1 above.

Create a second input replacing HF/cc-pVTZ with PBE1PBE/cc-pVTZ, adding Int=UltraFine, and changing the checkpoint/title to g25_water_pbe0. Gaussian's PBE1PBE keyword denotes the PBE0 hybrid functional. A third branch can use B3LYP/cc-pVTZ with the same explicit integration grid. These are independent optimized surfaces, not repeated evaluations of one stationary geometry. /

7.1.3.3 Workflow

  1. Run all selected geometry/frequency jobs with a shared basis and inspect optimization completion plus the expected three water vibrations.
  2. Record bond lengths, angle, dipole, and mode assignments. Keep raw harmonic frequencies before applying any externally justified scaling.
  3. At the HF optimized geometry, evaluate each DFT method as a separate SP with the correct geometry copied explicitly or read through a branch. Compare that with its own relaxed geometry result.
  4. Repeat one DFT calculation with a larger integration grid using its documented keyword, keeping the geometry fixed, if the target quantity is numerically delicate.
  5. When comparing experiment, specify phase, temperature, isotopologue, and whether the experimental geometry or spectrum includes vibrational averaging.

7.1.3.4 Output interpretation checks and limits

The method table should have separate columns for fixed-geometry electronic results and each method's optimized structural/spectral results. Never add an HF thermal correction to a DFT energy without explicitly defining it as a composite approximation and assessing its suitability. Harmonic frequencies are curvature predictions, not exact band positions. A successful three-method water exercise does not establish reliability for bond dissociation, transition metals, dispersion, or long-range charge transfer. Report disagreement as sensitivity to the model; do not convert a small spread into a statistically calibrated confidence interval.

Exercise Design a method-selection memo for predicting a carbonyl IR shift. Identify what water taught you and which validation molecule is still needed.

7.1.5 Sources and further reading