<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Examples :: Unfolding Documentation</title><link>https://mailhexu.github.io/unfolding/examples/index.html</link><description>Each card links to a walkthrough with input settings and runnable reproduction scripts. Some fixtures are generated locally instead of bundled (large GPAW restarts and VASP’s licensed POTCAR/WAVECAR).&#10;Phonons Phonopy Cu phonons Unfold a 3×3×3 fcc Cu phonon calculation from FORCE_CONSTANTS.&#10;ABINIT DDB phonons Phonon unfolding from ABINIT DDB files (Cu and CaTiO₃).&#10;Magnons TB2J magnons SrMnO3 Downfold TB2J magnon bands from the G-AFM cell onto the pseudo-cubic primitive cell.</description><generator>Hugo</generator><language>en-us</language><atom:link href="https://mailhexu.github.io/unfolding/examples/index.xml" rel="self" type="application/rss+xml"/><item><title>Phonopy Cu phonons</title><link>https://mailhexu.github.io/unfolding/examples/phonopy-cu/index.html</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://mailhexu.github.io/unfolding/examples/phonopy-cu/index.html</guid><description>Unfold the phonon band structure of a 3×3×3 fcc Cu supercell, computed with phonopy, onto the primitive fcc Brillouin zone. For a pristine crystal every supercell phonon folds from exactly one primitive momentum, so the weights are binary: three bold branches trace the primitive Cu dispersion while the other 24 folded copies of the 27-atom cell stay invisible. In a defective or distorted supercell the same plot shows fractional weights.</description></item><item><title>SIESTA Si bands</title><link>https://mailhexu.github.io/unfolding/examples/siesta-si/index.html</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://mailhexu.github.io/unfolding/examples/siesta-si/index.html</guid><description>Unfold a SIESTA supercell calculation onto the primitive-cell band path using the stored Hamiltonian: an 8-atom conventional-cubic Si supercell (SaveHS true) unfolded onto the 2-atom primitive fcc cell. Pristine weights are binary, so the bold branches are the primitive band structure and the independent primitive-cell calculation (red) checks them branch for branch.&#10;The bundle Download siesta-si.tar.gz and unpack it:&#10;tar xf siesta-si.tar.gz &amp;&amp; cd siesta-si Everything needed ships in the bundle: the SIESTA fixtures (data/si_prim.fdf + si_prim.HSX, data/si_sc.fdf + si_sc.HSX — each parser input sits next to its Hamiltonian archive), the primitive cell as a POSCAR (data/si_prim.vasp), the unfold.toml config, the reproduce.py script, and this README-adjacent note on pseudopotentials. Prerequisites: pip install unfolding plus pip install HamiltonIO sisl (the SIESTA parser). No SIESTA run is needed.</description></item><item><title>SIESTA WFSX route</title><link>https://mailhexu.github.io/unfolding/examples/siesta-wfsx/index.html</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://mailhexu.github.io/unfolding/examples/siesta-wfsx/index.html</guid><description>If the supercell run already stored its wavefunctions, unfold from SIESTA’s own eigenvectors and eigenvalues instead of diagonalizing the Hamiltonian: an 8-atom conventional-cubic Si supercell run (SaveWFSX true) unfolded onto the 2-atom primitive fcc cell. Same cell, same path, same weights as the Hamiltonian route — only the eigen-solve is replaced; the weight still uses the .HSX overlap shells.&#10;The bundle Download siesta-wfsx.tar.gz and unpack it:&#10;tar xf siesta-wfsx.tar.gz &amp;&amp; cd siesta-wfsx Shipped: the WFSX path run (data/si_sc_path.selected.WFSX plus its si_sc_path.EIG), the Hamiltonian fixtures for the overlap shells and the primitive reference (data/si_prim.*, data/si_sc.*), the primitive POSCAR, the path-run deck (inputs/si_sc_path.fdf, with SaveWFSX true and a %block WaveFuncKPoints list covering the path), and reproduce.py. Prerequisites: pip install unfolding plus pip install HamiltonIO sisl (needs HamiltonIO ≥ 0.3.6 for SiestaWFSXParser). No SIESTA run is needed.</description></item><item><title>SIESTA Si:P dopant</title><link>https://mailhexu.github.io/unfolding/examples/siesta-p-doped/index.html</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://mailhexu.github.io/unfolding/examples/siesta-p-doped/index.html</guid><description>Unfold a substituted SIESTA supercell — one Si atom of the 8-atom conventional-cubic cell replaced by P — onto the primitive fcc cell band path. The substitution breaks the translation symmetry: host bands keep weight 1 while donor-derived states appear with fractional weight.&#10;The bundle Download siesta-p-doped.tar.gz and unpack it:&#10;tar xf siesta-p-doped.tar.gz &amp;&amp; cd siesta-p-doped Shipped: the Si₇P supercell fixtures (data/si_sc_p.fdf + si_sc_p.HSX), the pristine primitive cell (data/si_prim.fdf + si_prim.HSX, POSCAR copy in data/si_prim.vasp), the P pseudopotential (pseudos/P.psml), and reproduce.py. Prerequisites: pip install unfolding plus pip install HamiltonIO sisl. No SIESTA run is needed.</description></item><item><title>SIESTA spinors</title><link>https://mailhexu.github.io/unfolding/examples/siesta-spinor/index.html</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://mailhexu.github.io/unfolding/examples/siesta-spinor/index.html</guid><description>Non-collinear (nspin=4, spinor) SIESTA supercell runs unfold through the same machinery as the collinear example — the parse → relabel → weight chain on the doubled basis (each PAO times two spin components). Here: the 8-atom conventional-cubic Si supercell run non-collinear, unfolded onto the 2-atom primitive fcc cell. The scalar Si pseudopotential has no SOC channels, so the spinor bands equal the scalar bands with Kramers degeneracy — a clean end-to-end seal of the spinor machinery, not spinor physics.</description></item><item><title>ABINIT WFK Si</title><link>https://mailhexu.github.io/unfolding/examples/abinit-wfk/index.html</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://mailhexu.github.io/unfolding/examples/abinit-wfk/index.html</guid><description>Unfold ABINIT supercell calculations directly from their wavefunction files (norm-conserving plane waves: the weight is a pure reciprocal-coset projection — no overlaps or atom maps are involved). This example presents pristine Si in the 8-atom conventional cell, unfolded onto the primitive fcc cell along Γ–X–W–Γ–L–W–X. Its weights are binary — the unfolded bands ARE the primitive band structure.&#10;The bundle Download abinit-wfk-si.tar.gz and unpack it:&#10;tar xf abinit-wfk-si.tar.gz &amp;&amp; cd abinit-wfk-si Shipped: the ABINIT decks (inputs/, two-dataset SCF + non-SCF path runs for Si8 and Si₇P, plus the 2-atom primitive reference), the Troullier–Martins fhi pseudopotentials (pseudos/14-Si.nlcc.fhi, 15-P.LDA.fhi), unfold.toml, the primitive-cell POSCAR (data/si_prim.vasp), and reproduce.py. Not shipped: the WFK files themselves (each dense path WFK is hundreds of MB) — regenerate them with ABINIT as below; this is the bundle’s one user-supplied input. Prerequisites: pip install unfolding; ABINIT ≥ 9 with netCDF support only for the regeneration step.</description></item><item><title>ABINIT DDB phonons</title><link>https://mailhexu.github.io/unfolding/examples/abinit-ddb/index.html</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://mailhexu.github.io/unfolding/examples/abinit-ddb/index.html</guid><description>Unfold phonons from an ABINIT DDB (derivatives database) onto a primitive cell: the route runs ABINIT’s anaddb through abipy to obtain the supercell eigenvectors along your path, then computes the unfolding weights with gauge-robust Bloch-sum projectors and degenerate-group resolution, so eigenvector storage gauges do not change the weights. Two bundled systems: fcc Cu from a conventional-cubic-cell DDB, and CaTiO₃ from its 20-atom Pnma cell onto the 5-atom pseudo-cubic cell.&#10;The bundle Download abinit-ddb.tar.gz and unpack it:</description></item><item><title>ABINIT WFK AFM NiO</title><link>https://mailhexu.github.io/unfolding/examples/abinit-nio/index.html</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://mailhexu.github.io/unfolding/examples/abinit-nio/index.html</guid><description>Unfold type-II antiferromagnetic NiO from the 4-atom magnetic primitive cell onto the 2-atom rocksalt primitive cell. The magnetic cell doubles the rocksalt primitive cell along [111] and carries two Ni moments (+m and −m along z) and two O; the collinear WFK (nsppol 2) stores two spin channels, and the unfolding weight of a state is computed within its own spin channel — no spin mixing.&#10;The bundle Download abinit-nio.tar.gz and unpack it:</description></item><item><title>Wannier90 SrTiO3</title><link>https://mailhexu.github.io/unfolding/examples/wannier-sto/index.html</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://mailhexu.github.io/unfolding/examples/wannier-sto/index.html</guid><description>Unfold a real Wannier90 tight-binding Hamiltonian of cubic SrTiO$_3$ — the pristine $\sqrt{2}\times\sqrt{2}\times 2$ supercell (20 atoms, 56 Wannier functions) and, on the same footing, a Ti-vacancy supercell — onto the 5-atom cubic primitive cell along Γ-X-M-Γ-R. The bundled wannier90_hr.dat files are exactly as written by Wannier90; the package’s own reader (hr + win + Wannier-function centres) is all that is needed, minimulti is not required.&#10;The bundle Download wannier-sto.tar.gz and unpack it:</description></item><item><title>TB2J magnons SrMnO3</title><link>https://mailhexu.github.io/unfolding/examples/tb2j-magnon-srmmo3/index.html</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://mailhexu.github.io/unfolding/examples/tb2j-magnon-srmmo3/index.html</guid><description>Downfold magnon band structures computed from TB2J exchange parameters onto the primitive magnon Brillouin zone: the antiferromagnetic DFT cell (√2×√2×√2 G-AFM SrMnO₃, 10 atoms) is larger than the chemical primitive cell (5-atom pseudo-cubic, one Mn), so the magnon bands fold and unfolding assigns each folded branch its primitive-cell momentum.&#10;SrMnO3 VASP/Wannier90 DFT (G-AFM) -&gt; TB2J pickle: JR tensors, 10-atom sqrt2 x sqrt2 x sqrt2 cell, 2 Mn -&gt; TB2J Magnon: BdG magnon bands (collinear reference, Q = 0) -&gt; unfolding: weights on the 5-atom pseudo-cubic primitive cell The bundle Download tb2j-magnon-srmmo3.tar.gz and unpack it:</description></item><item><title>GPAW LCAO Si and Si:P</title><link>https://mailhexu.github.io/unfolding/examples/gpaw-si/index.html</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://mailhexu.github.io/unfolding/examples/gpaw-si/index.html</guid><description>Unfold a GPAW LCAO supercell calculation onto the primitive-cell band path: pristine Si and P-doped Si in the 8-atom conventional cubic cell, unfolded onto the 2-atom fcc primitive cell through HamiltonIO.gpaw.GpawLcaoModel. GPAW’s LCAO matrices already carry every PAW contribution — the overlap is the projector-augmented one and the Hamiltonian includes the dH projector terms — so the unfolding backend consumes them exactly like any other atomic-orbital table: no PAW correction is applied on top.</description></item><item><title>GPAW plane-wave Si</title><link>https://mailhexu.github.io/unfolding/examples/gpaw-si-pw/index.html</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://mailhexu.github.io/unfolding/examples/gpaw-si-pw/index.html</guid><description>Unfold pristine GPAW pseudo plane-wave coefficients by reciprocal cosets with the backend-free plane-wave engine. The 8-atom conventional cubic Si supercell is sampled on the exact 305-point primitive fcc path recorded in the fixture restart and compared with independently computed primitive-cell plane-wave bands. This does not reconstruct PAW all-electron weights.&#10;The Si:P plane-wave restart is Γ-only: it provides one supercell momentum, not a sampled band path. Its sparse doped output is suppressed. The sparse restart remains in the repository’s test fixtures; the public bundle and this page present the adequately sampled pristine path only. This does not affect the GPAW LCAO Si:P example, whose supercell uses an 8×8×8 k-grid.</description></item><item><title>ABACUS LCAO Si bands</title><link>https://mailhexu.github.io/unfolding/examples/abacus-si/index.html</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://mailhexu.github.io/unfolding/examples/abacus-si/index.html</guid><description>Unfold an ABACUS LCAO supercell calculation onto the primitive-cell band path: an 8-atom conventional-cubic Si supercell (plus a Si:P variant) with the real-space Hamiltonian and overlap exported (out_mat_hs2 1 writes data-HR-sparse_SPIN0.csr / data-SR-sparse_SPIN0.csr), unfolded onto the 2-atom fcc primitive cell through HamiltonIO.abacus.abacus_wrapper.AbacusParser, exactly like the SIESTA example.&#10;Run parameters and shared options (mode, spin, method, resolve_degenerate, supercell-matrix row convention, k-path coordinates and frames, energy-reference conventions) are explained together in the method, parameters, and k-path guide. Primitive cell input: required for the LCAO abacus route — it supplies the relabel map and the default path cell.</description></item><item><title>OpenMX Si bands</title><link>https://mailhexu.github.io/unfolding/examples/openmx-si/index.html</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://mailhexu.github.io/unfolding/examples/openmx-si/index.html</guid><description>Unfold an OpenMX supercell calculation onto the primitive-cell band path: an 8-atom conventional-cubic Si supercell (plus a Si:P variant) unfolded onto the 2-atom fcc primitive cell from the binary .scfout file (real-space Hamiltonian and overlap, written with HS.fileout on), parsed by the HamiltonIO OpenMX adapter (pure python, no OpenMX runtime needed).&#10;Run parameters and shared options (mode, spin, method, resolve_degenerate, supercell-matrix row convention, k-path coordinates and frames, energy-reference conventions) are explained together in the method, parameters, and k-path guide. Primitive cell input: required for the openmx route — it supplies the relabel map and the default path cell.</description></item><item><title>ABINIT PAW plane-wave unfolding: Si and Si:P</title><link>https://mailhexu.github.io/unfolding/examples/abinit-paw/index.html</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://mailhexu.github.io/unfolding/examples/abinit-paw/index.html</guid><description>Unfold ABINIT PAW plane-wave calculations with the unfold_abinit_paw adapter: 8-atom pristine Si and 8-atom Si:P supercells onto the 2-atom primitive fcc cell, with a separately computed 2-atom primitive WFK supplying the reference bands. The PAW wavefunction is normalized in the overlap metric, so overlaps are evaluated with S = I + Σ_aij |p_ai⟩ ΔS_aij ⟨p_aj| (only the overlap augmentation ΔS enters the projection).&#10;Run parameters and shared options (mode, spin, method, resolve_degenerate, supercell-matrix row convention, k-path coordinates and frames, energy-reference conventions) are explained together in the method, parameters, and k-path guide. Primitive cell input: required for the abinit-paw route — it provides the primitive-cell wavefunctions the supercell is projected onto.</description></item><item><title>VASP PAW plane-wave unfolding: bcc Fe</title><link>https://mailhexu.github.io/unfolding/examples/vasp-paw/index.html</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://mailhexu.github.io/unfolding/examples/vasp-paw/index.html</guid><description>Unfold a VASP PAW calculation — spin-polarized bcc Fe in a 2×2×2 conventional supercell (16 atoms) — onto the 1-atom bcc primitive cell with the unfold_vasp_paw adapter, from WAVECAR + POSCAR + POTCAR. As in the ABINIT PAW route, overlaps use the PAW metric S = I + Σ_aij |p_ai⟩ ΔS_aij ⟨p_aj| built from the POTCAR projector tables; primitive reference wavefunctions are embedded into the supercell G basis (A = B† S C, G = B† S B, weight = diagonal of A† G⁻¹ A).</description></item></channel></rss>