Skip to content

2.4 Build and audit crystalline silicon

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.

2.4.1 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: Build and audit crystalline silicon. No numerical results are claimed.
Original schematic: Build and audit crystalline silicon. No numerical results are claimed.

2.4.2 Worked investigation

Intuition. A periodic cell is a generator of an infinite structure. Its coordinates, lattice vectors and boundary conditions must be interpreted together. An eight-atom conventional diamond cell and a two-atom primitive cell can describe the same crystal with different Brillouin zones. Begin with baseline B and do not compare their raw cell energies without normalization and equivalent reciprocal-space sampling.

Input and prerequisites. Use B's eight fractional coordinates and cubic lattice 5.43 angstrom. Set both atomic and Poisson periodicity to XYZ. SCALED T makes coordinate entries dimensionless fractions of lattice vectors, not angstrom. The q4 potential yields 32 valence electrons in the eight-atom neutral cell. For an initial static demonstration OT at Gamma is acceptable; quantitative small-cell energetics require case 12's k-point convergence.

Workflow. 1. Reconstruct Cartesian coordinates from the lattice matrix and fractional positions. 2. Replicate the cell visually and verify tetrahedral local coordination rather than trusting the atom count alone. For the ideal geometry, the nearest-neighbor distance follows a√3/4; this is geometry, not a computed bond-length result. 3. Run one single-point calculation and check 8 atoms, 32 explicit electrons and the expected restricted occupations. 4. Build a 2×2×2 supercell by explicit coordinate replication with an external crystallographic tool or script; record the tool/version and preserve atom ordering. 5. Compare energy per atom only after choosing equivalent sampling. A Gamma supercell and primitive Gamma cell generally do not integrate the same k-space set.

Interpretation and tests. A perfect high-symmetry structure may have nearly zero forces even at the wrong volume. Forces test internal coordinates, whereas pressure/stress probes cell mismatch. Therefore zero atomic force is not evidence for an equilibrium lattice constant. If printing stress, add STRESS_TENSOR ANALYTICAL at FORCE_EVAL for supported Quickstep settings and inspect its units/sign convention. Confirm consistency by small isotropic strains rather than memorizing a pressure sign.

Pitfalls. Duplicating atoms on both equivalent cell boundaries creates an unphysical density. Cartesian coordinates mistakenly read as fractional may place atoms far apart or on top of each other. Rotating a cell without rotating positions changes the physical model. Crystalline silicon can have a small semilocal band gap; inadequate sampling or an unsuitable electronic solver can complicate interpretation even though the real material is a semiconductor.

Exercise. Verify the nearest-neighbor geometry analytically, then construct a slightly strained cell with a increased by 1% while keeping fractional coordinates fixed. Explain why forces may remain symmetry-suppressed while the stress changes. Record energy per atom, not only energy per cell.

Diagram. Draw a cubic cell and highlighted tetrahedral neighbors, alongside a lattice-vector/fractional-coordinate conversion. Labels: conventional cell, periodic neighbor, fractional position, energy per atom, cell strain. Avoid displaying a fabricated electronic band structure.

2.4.4 Sources and further reading