1.1 Four input files, one physical question
These are unexecuted teaching inputs and starting settings to test. Original figures are schematics, not computed results. Use licensed VASP and PAW data, replace every placeholder, record the executable version and validate convergence.
1.1.1 Model, units and provenance
Use eV for energy, Å for length and eV/Å for force; 1 kbar = 0.1 GPa. State normalization per atom, molecule, primitive cell or simulation cell. Record PAW identifiers, release, ZVAL, ENMAX and permitted hashes; never redistribute POTCAR. SCF convergence addresses the chosen electronic problem; convergence of the target property requires separate tests.
Original schematic. Curves explain concepts; blank data areas await verified learner results. No calculation is claimed.
1.1.2 Worked case: procedure, interpretation and checks
Goal and intuition
Before asking “What is the energy?”, ask “Energy of which model?” POSCAR defines the periodically repeated cell and positions. POTCAR specifies how each species is represented by PAW data. KPOINTS determines how the reciprocal-space integral is sampled. INCAR selects the electronic and ionic tasks and their tolerances. Treat these as four linked parts of a contract. A beautiful structure picture cannot reveal that its species order disagrees with POTCAR, and a job that exits normally may still have reached an iteration limit.
A useful first task is a static two-atom silicon calculation. The objective is not a publishable Si energy. It is to show that you can reconstruct the calculation from its inputs and distinguish electronic iterations from ionic steps. Prerequisites: a licensed working VASP installation, an authorized PAW library, access to your site's launcher instructions, and the ability to inspect plain-text files. Do not assume a web tutorial's mpirun command fits your cluster.
Workflow
- Make a new directory containing only the intended inputs. Record the VASP executable version and whether you use the standard, Γ-only, or noncollinear executable. Use the standard executable for the bulk example.
- Inspect the cell visually and independently count atoms and species. Obtain Si PAW data from your licensed installation; confirm that the POTCAR block order matches POSCAR.
- Choose an explicit trial cutoff at least as large as the largest relevant ENMAX. Substitute that number below; do not leave a placeholder.
- Launch using the site's documented scheduler. Save stdout/stderr. Confirm the job did not stop because of wall time, memory, or an electronic-iteration limit.
- Open OUTCAR and verify the settings VASP actually used. Find the number of ions, electron count, k-point sampling, energy records, and final timing information.
Original starter files
Unexecuted POSCAR; 5.43 Å is an illustrative starting conventional lattice parameter, not a predicted equilibrium result.
What to inspect
OSZICAR is a compact progress view; OUTCAR carries detailed diagnostics; CONTCAR stores the last structure; vasprun.xml and, where supported, vaspout.h5 support structured analysis. For a static job, CONTCAR should not be interpreted as an optimized structure. Note the distinction among free energy TOTEN, energy without entropy, and the reported zero-smearing estimate. Record the label beside every energy you extract instead of calling every last number “the total energy.”
Convergence experiment, mistakes and limits
Repeat with EDIFF of 1E-5, 1E-6 and 1E-7 eV, changing nothing else. Tabulate the energy label used, iteration count, and residual force. This tests electronic stopping tolerance only. A common mistake is to tighten EDIFF while leaving a coarse k mesh, then claim overall convergence. Another is to copy a hidden WAVECAR/CHGCAR into a new directory and unknowingly restart from an earlier problem. Keep first-run and restart procedures explicit. Do not replace a missing licensed potential with a downloaded unofficial POTCAR.
Exercise and original figure
Deliver a one-page provenance sheet: physical question, four input roles, species order, PAW identifiers, actual cutoff, executable version, output energy definition, and evidence that the electronic cycle converged. Explain why two calculations with different PAW datasets cannot generally be compared by simply subtracting their absolute energies.
1.1.3 Unexecuted inputs and analysis scaffolds
These are unexecuted teaching inputs and starting settings to test. Original figures are schematics, not computed results. Use licensed VASP and PAW data, replace every placeholder, record the executable version and validate convergence.
1.1.3.1 Input block 1
Si primitive teaching cell
5.43
0.0 0.5 0.5
0.5 0.0 0.5
0.5 0.5 0.0
Si
2
Direct
0.00 0.00 0.00
0.25 0.25 0.25
1.1.3.2 Input block 2
# INCAR: static teaching calculation; replace placeholder
SYSTEM = Si input anatomy
ENCUT = <ECUT_EV>
PREC = Accurate
EDIFF = 1E-6
ALGO = Normal
ISMEAR = 0
SIGMA = 0.05
LREAL = .FALSE.
IBRION = -1
NSW = 0
LWAVE = .FALSE.
LCHARG = .FALSE.