3.3 Crystal relaxation, stress and Pulay error
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.
3.3.1 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.
3.3.2 Worked case: procedure, interpretation and checks
Goal and intuition
Moving atoms and changing a cell are different operations. Atomic forces answer whether atoms want to move inside the current cell. Stress answers how the energy changes under strain. A high-symmetry crystal can have zero atomic forces and a badly chosen volume. Cell optimization adds a numerical complication: changing the cell changes the relationship between the finite plane-wave basis and the material, which can introduce Pulay stress.
Use Si or Al from the earlier bulk cases. The target is a defensible equilibrium cell at the chosen approximation, supported by an independent energy–volume check. Prerequisites: converged electronic and sampling settings and a clear decision about which degrees of freedom should be allowed. A molecular box or slab vacuum dimension should not be freely optimized as though it were a bulk lattice constant.
Workflow
- Decide whether the task is fixed-cell internal relaxation or full cell relaxation. Write the decision in words before selecting ISIF.
Unexecuted template
ISIF=2 allows ionic positions to change in a fixed cell during a relaxation; ISIF=3 also allows cell shape and volume. ISIF=8, available from VASP 6.4.1, allows positions and volume while preserving cell shape; do not recommend it to an older installation without checking. For a beginner, an explicit isotropic volume series is often easier to audit than a complex constrained-cell workflow.
Reading stress correctly
VASP's stress-sign convention must be read deliberately: positive diagonal values correspond to compression in the documented convention. Do not import a sign rule from a different code without checking. Hydrostatic pressure is not the same as every tensor component; shear components can reveal a shape problem even when an average pressure is small. Compare kbar and GPa only after conversion.
A negative EDIFFG expresses a force-based ionic stopping criterion. It is not a user-specified pressure tolerance. Therefore reaching the ionic stopping condition is not enough evidence that the lattice is satisfactory. Inspect the final stress tensor and independently state what residual stress you consider acceptable for the intended property.
Experiment, mistakes, exercise
Compare final volumes from two cutoff levels and perform fresh static stress checks at both. Then compare both to an independently sampled energy–volume minimum. Avoid changing the k-grid count discontinuously across nearby volumes without recording that choice; it can create artificial structure in the curve. Do not “fix” pressure by adding an arbitrary PSTRESS until you understand and have measured the correction being applied. Exercise: create a relaxation audit with initial cell, permitted degrees of freedom, final cell, final force maximum, stress tensor and independent minimum check.
3.3.3 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.
3.3.3.1 Input block 1
# Full bulk-cell relaxation; not for an isolated molecule or vacuum slab
ENCUT = <CELL_TESTED_ECUT_EV>
PREC = Accurate
EDIFF = 1E-7
LREAL = .FALSE.
IBRION = 2
NSW = 100
ISIF = 3
EDIFFG = -0.01
# Use the validated insulating or metallic occupation settings for the material
3.3.3.2 Input block 2
# Fresh validation after copying CONTCAR to POSCAR
IBRION = -1
NSW = 0
ISIF = 2
# Keep or improve cutoff and k sampling; inspect full residual stress
3.3.4 Related learning paths
- 1.1 Four input files, one physical question
- 1.2 A convergence laboratory with an error budget
- 4.1 Competing magnetic states in bcc iron
- 5.1 Equation of state and bulk modulus
- 5.2 Elastic tensor from finite strains