8.1 Born–Oppenheimer MD: integrate accurate forces

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.

8.1.1 Model, units and provenance

PW cutoffs and energies use Ry, common force output uses Ry/bohr, and pressure uses kbar. Geometry cards state their coordinate units. Different executables have distinct grammars and time-unit conventions.

Shared inputs, conventions and evidence

Original schematic: Born–Oppenheimer MD: integrate accurate forces. No numerical results are claimed.

8.1.2 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.

8.1.2.1 Input block 1

! Copy final relaxed geometry; new prefix/outdir.
&CONTROL
 calculation='md', prefix='water_md', pseudo_dir='./pseudo',
 outdir='./scratch/water_md', nstep=200, dt=5.0,
 tprnfor=.true., tstress=.true.
/
&ELECTRONS
 conv_thr=1.0d-10, mixing_beta=0.3
/
&IONS
 ion_dynamics='verlet', ion_temperature='not_controlled'
/

8.1.3 Worked investigation

8.1.3.1 Intuition and prerequisites

In Born–Oppenheimer MD, electrons are converged for each ionic configuration before forces advance the nuclei. An apparently smooth trajectory can still drift because of incomplete SCF, a large time step or an inconsistent restart. Use a small insulating system first: the relaxed water molecule is a numerical integrator exercise, not a liquid-water simulation. A liquid requires a separately prepared many-molecule cell and sampling plan.

8.1.3.2 Original delta to the relaxed-water input

See input block 1 above.

8.1.3.3 Checks, pitfalls and exercise

Plot total conserved-energy drift per atom or per molecule versus physical time. Halve dt and tighten SCF separately to identify the controlling error. Monitor bond lengths for unphysical excursions and account for removed translational/rotational degrees of freedom when discussing temperature. Thermostats can conceal integration drift, so establish a stable microcanonical test before using a thermostat for equilibration. A thermostat name alone does not prove the desired ensemble is sampled.

Exercise: run two initial conditions and compare vibrational behavior without interpreting a 48 fs trace as a diffusion coefficient. For a later liquid extension, distinguish equilibration, production, time correlation and independent samples; block-average uncertainties and test finite-size effects. Never use unwrapped diffusion coordinates without checking periodic-boundary conventions.