SRO SiO₂

A 30 Å cubic SiO₂ supercell (2052 atoms) built with the short-range order regime (grain_size = 12 Å, displacement_sigma = 0.010), then refined through the four-stage pipeline.

To re-generate:

python scripts/regen_dftb_examples.py --regime sro --box 30.0

g(r) per stage

Si–O bond length distribution

stage	Si–O peak (Å)	⟨r⟩ (Å)	σ (Å)	# Si–O bonds
Voronoi	1.61	1.61	0.070	2224
after orient	1.61	1.61	0.070	2224
after FIRE	1.67	1.70	0.090	2415
after DFTB+ (60 steps)	1.65	1.72	0.076	2342

DFTB+ pulls the peak back from 1.67 Å to 1.65 Å and narrows σ by ~16 % (0.090 → 0.076). Residual fmax at step 60 is 9.32 eV/Å — the 60-step cap stops the relax short of the 1.0 eV/Å target, but energy descent is monotonic across all 60 steps.

Angle distributions

Energy ladder

stage	E_total (eV)	ΔE vs previous
Voronoi	-141093.01	—
after orient	-141093.01	0
after FIRE	-145637.05	-4544
after DFTB+ relax	-145981.82	-345 (≈ -0.17 eV/atom)

Cell volume change Voronoi → DFTB+: +0.1 %. DFTB+ total at 60 steps is -71.1 eV/atom, matching the nanocrystalline and smaller-cell results.

DFTB+ relax wallclock: 11,670 s = 3.2 hr for the seed SP + 60 FIRE+UCF steps (~2.6 min/step on average).

Reproduction summary

knob	value
cell side	30 Å
atoms	2052
grain_size	12 Å
FIRE steps (tricor stage 3)	200
pre-DFTB cleanup	bond_relax(20) + enforce_hard_core(40)
DFTB+ optimizer	ASE FIRE, maxstep = 0.03 Å
DFTB+ cell DOFs	volume only (hydrostatic_strain=True)
DFTB+ fmax target	1.0 eV/Å, 60-step cap
charge restart	seed SP + ReadInitialCharges=Yes
SK set	matsci-0-3
k-points	(1, 1, 1) Γ-only

Per-stage trajectory (extxyz): docs/_static/dftb/sro_traj.xyz