Star-of-David Charge-Density-Wave

• Star-of-David charge-density-wave is a structural motif where atomic displacements and charge modulations form a six-pointed cluster in transition-metal dichalcogenides and kagome metals.
• Studies employ DFT, tr-ARPES, and X-ray diffraction to reveal detailed electron–phonon interactions, precise atomic displacements, and charge instabilities underlying the motif.
• This structural archetype can trigger varied electronic phases—from Mott insulation to correlated metallicity and unconventional magnetism—offering insights into complex quantum materials.

Searching arXiv for recent and foundational papers on star-of-David charge-density-wave states across transition-metal dichalcogenides and kagome metals. The star-of-David charge-density-wave is a commensurate charge–lattice reconstruction in which atomic displacements and charge or bond modulations organize into a six-pointed-cluster motif. In layered $1T$ transition-metal dichalcogenides, the canonical form is a $\sqrt{13}\times\sqrt{13}$ supercell containing one central transition-metal atom, six inner-ring atoms, and six outer-ring atoms; in kagome metals it usually appears as a $2\times2$ or $2\times2\times2$ triple-$Q$ breathing pattern centered on kagome hexagons, with a closely related inverse star-of-David or tri-hexagonal counterpart. Across these materials, the motif is not a single electronic phase but a structural archetype that can accompany Mott localization, correlated metallicity, unusual ferromagnetism, orbital-flux order, or superconductivity (Pasquier et al., 2018, Uykur et al., 2021, Hu et al., 2022).

1. Structural motif and order-parameter geometry

In the triangular $1T$ dichalcogenide lattice, the standard star-of-David reconstruction is generated from primitive vectors

$$a_1=a(1,0),\qquad a_2=a\left(\tfrac12,\tfrac{\sqrt3}{2}\right),$$

with a commensurate supercell such as

$$A_1=3a_1+a_2,\qquad A_2=-a_1+4a_2,\qquad |A_i|=\sqrt{13}\,a.$$

In monolayer $1T$-NbSe$_2$, each supercell contains thirteen Nb atoms reorganized into a central Nb surrounded by two concentric rings of six Nb each; the six nearest Nb are displaced radially inwards by $\sqrt{13}\times\sqrt{13}$0 Å and the outer six by $\sqrt{13}\times\sqrt{13}$1 Å, while the Se sublattice relaxes so that the local symmetry is reduced from $\sqrt{13}\times\sqrt{13}$2 to $\sqrt{13}\times\sqrt{13}$3. Single-layer $\sqrt{13}\times\sqrt{13}$4-NbS$\sqrt{13}\times\sqrt{13}$5 realizes the same $\sqrt{13}\times\sqrt{13}$6 geometry, again as a 13-atom cluster with radial inward displacements of the two shells (Pasquier et al., 2018, Tresca et al., 2019).

The kagome realization is structurally different but order-parameter equivalent in the sense of a triple-$\sqrt{13}\times\sqrt{13}$7 commensurate modulation. In $\sqrt{13}\times\sqrt{13}$8V$\sqrt{13}\times\sqrt{13}$9Sb$2\times2$0, the ordering vectors are the three $2\times2$1-point wavevectors,

$2\times2$2

or symmetry-equivalent conventions, and the charge or bond modulation is written as

$2\times2$3

The star-of-David corresponds to $2\times2$4 up to common phase. In real space, six V atoms around a kagome hexagon move inward, while the complementary inverse star-of-David or tri-hexagonal pattern reverses the breathing sense. In a single kagome layer, some analyses assign the inverse star-of-David to the fully symmetric $2\times2$5 channel of $2\times2$6, whereas three-dimensional $2\times2$7-shifted stacking lowers the crystal symmetry to $2\times2$8 or $2\times2$9 (Uykur et al., 2021, Miao et al., 2021, Deng et al., 10 Mar 2025).

2. Microscopic mechanisms

No single mechanism explains all star-of-David CDWs. In monolayer $2\times2\times2$0-NbSe$2\times2\times2$1, the transition from the metallic undistorted $2\times2\times2$2 lattice to the $2\times2\times2$3 commensurate phase is driven by the interplay of Fermi-surface nesting, strong Nb–Se covalency, and moderate Hubbard interactions. Density-functional perturbation theory finds a pronounced softening of an in-plane Nb–Se bond-stretching phonon along $2\times2\times2$4–$2\times2\times2$5, while the bare susceptibility $2\times2\times2$6 peaks at $2\times2\times2$7; the incommensurate instability then locks onto the nearest commensurate $2\times2\times2$8 vector through higher-order free-energy terms (Pasquier et al., 2018).

Bulk $2\times2\times2$9-NbS$Q$0 shows a different balance. There the low-temperature ground state is also a $Q$1 star-of-David CCDW, but the calculated bare susceptibility has no sharp peak at the measured ordering vector. Instead, the phonon linewidth and Eliashberg analysis identify a momentum-selective, strongly softened acoustic branch around $Q$2, with $Q$3 in the undistorted lattice. In that case the principal CDW driver is strong, $Q$4-dependent electron–phonon coupling rather than simple nesting (Wang et al., 2020).

In antiferromagnetic FeGe, the experimentally observed $Q$5 charge order is described as an interaction-assisted phonon-instability scenario. DFT+$Q$6 with $Q$7 eV pushes Fe-$Q$8 flat bands away from $Q$9, enhances Ge-$1T$0 weight at $1T$1, and softens a nearly flat optical phonon branch at $1T$2, which becomes imaginary for $1T$3 eV. The generalized susceptibility does not diverge even up to $1T$4 eV, so electron–electron interactions alone do not drive the CDW; instead they renormalize the electronic structure so as to increase the electron–phonon coupling to the soft $1T$5-phonon (Ma et al., 2023).

In kagome metals, several mechanisms are explicitly in play. ARPES on KV$1T$6Sb$1T$7 identifies a CDW gap tied to inter-saddle-point scattering at the kagome $1T$8-point van Hove singularities, while Landau and DFT analyses find unstable phonon modes at $1T$9, $$a_1=a(1,0),\qquad a_2=a\left(\tfrac12,\tfrac{\sqrt3}{2}\right),$$0, and $$a_1=a(1,0),\qquad a_2=a\left(\tfrac12,\tfrac{\sqrt3}{2}\right),$$1 with strong electronic-temperature dependence, consistent with an electronically driven but lattice-coupled instability. Other studies emphasize V $$a_1=a(1,0),\qquad a_2=a\left(\tfrac12,\tfrac{\sqrt3}{2}\right),$$2–Sb $$a_1=a(1,0),\qquad a_2=a\left(\tfrac12,\tfrac{\sqrt3}{2}\right),$$3 orbital hybridization as the direct mediator of the CDW structural transition, or nearest-neighbor and interlayer Coulomb interactions that first generate charge bond order and then induce star-of-David or inverse star-of-David distortions through lattice coupling (Kato et al., 2022, Christensen et al., 2021, Han et al., 2022, Li et al., 2023).

3. Transition-metal dichalcogenide realizations

Monolayer $$a_1=a(1,0),\qquad a_2=a\left(\tfrac12,\tfrac{\sqrt3}{2}\right),$$4-NbSe$$a_1=a(1,0),\qquad a_2=a\left(\tfrac12,\tfrac{\sqrt3}{2}\right),$$5 is the clearest SoD–Mott case in the set considered here. The undistorted $$a_1=a(1,0),\qquad a_2=a\left(\tfrac12,\tfrac{\sqrt3}{2}\right),$$6 manifold spans $$a_1=a(1,0),\qquad a_2=a\left(\tfrac12,\tfrac{\sqrt3}{2}\right),$$7 eV and is metallic, but the nonmagnetic star-of-David distortion produces a very narrow band at $$a_1=a(1,0),\qquad a_2=a\left(\tfrac12,\tfrac{\sqrt3}{2}\right),$$8 of width $$a_1=a(1,0),\qquad a_2=a\left(\tfrac12,\tfrac{\sqrt3}{2}\right),$$9 meV. Spin-polarized GGA opens a small $$A_1=3a_1+a_2,\qquad A_2=-a_1+4a_2,\qquad |A_i|=\sqrt{13}\,a.$$0 meV gap with $$A_1=3a_1+a_2,\qquad A_2=-a_1+4a_2,\qquad |A_i|=\sqrt{13}\,a.$$1 per star, and GGA+$$A_1=3a_1+a_2,\qquad A_2=-a_1+4a_2,\qquad |A_i|=\sqrt{13}\,a.$$2 with $$A_1=3a_1+a_2,\qquad A_2=-a_1+4a_2,\qquad |A_i|=\sqrt{13}\,a.$$3 eV gives $$A_1=3a_1+a_2,\qquad A_2=-a_1+4a_2,\qquad |A_i|=\sqrt{13}\,a.$$4 eV, in much better agreement with the experimentally observed $$A_1=3a_1+a_2,\qquad A_2=-a_1+4a_2,\qquad |A_i|=\sqrt{13}\,a.$$5 eV insulating behavior. Energetically, the $$A_1=3a_1+a_2,\qquad A_2=-a_1+4a_2,\qquad |A_i|=\sqrt{13}\,a.$$6 phase gains $$A_1=3a_1+a_2,\qquad A_2=-a_1+4a_2,\qquad |A_i|=\sqrt{13}\,a.$$7 meV/NbSe$$A_1=3a_1+a_2,\qquad A_2=-a_1+4a_2,\qquad |A_i|=\sqrt{13}\,a.$$8, larger than the $$A_1=3a_1+a_2,\qquad A_2=-a_1+4a_2,\qquad |A_i|=\sqrt{13}\,a.$$9 CDW gain of $1T$0 meV/NbSe$1T$1 (Pasquier et al., 2018).

Single-layer $1T$2-NbS$1T$3 is closely analogous structurally but somewhat different electronically. Its $1T$4 reconstruction yields an ultraflat band of bandwidth $1T$5 eV, isolated by $1T$6 eV from all other bands. Spin-polarized GGA stabilizes a ferrimagnetic insulating state with $1T$7 eV and a $1T$8 moment localized on the central Nb, while GGA+$1T$9 with $_2$0 eV enhances the central moment to $_2$1 and produces a fundamental gap $_2$2 eV (Tresca et al., 2019).

Not all SoD dichalcogenides are Mott insulators. Monolayer $_2$3-NbTe$_2$4 develops an unusual $_2$5 star-of-David lattice, characterized as a sparsely occupied SoD pattern. ARPES shows a partial CDW gap $_2$6 eV on the nested $_2$7-centered pocket, but ungapped $_2$8-centered pockets remain, so the system stays metallic and shows no signature of a Mott gap. This is explicitly contrasted with monolayer $_2$9-NbSe$\sqrt{13}\times\sqrt{13}$00 and bulk $\sqrt{13}\times\sqrt{13}$01-TaS$\sqrt{13}\times\sqrt{13}$02, where the $\sqrt{13}\times\sqrt{13}$03 SoD pattern supports a Mott phase (Taguchi et al., 2022).

Monolayer $\sqrt{13}\times\sqrt{13}$04-VTe$\sqrt{13}\times\sqrt{13}$05 exhibits multimorphism rather than a unique SoD phase. STM and STS identify a metallic $\sqrt{13}\times\sqrt{13}$06 CDW with 13 V atoms condensed into a typical star-of-David cluster, and a gapped $\sqrt{13}\times\sqrt{13}$07 CDW with truncated-triangle clusters and a hard gap of $\sqrt{13}\times\sqrt{13}$08 meV. DFT+$\sqrt{13}\times\sqrt{13}$09 with $\sqrt{13}\times\sqrt{13}$10 eV reproduces both the metallic and gapped phases. The CDW-driven reorganization weakens ferromagnetic superexchange, strengthens antiferromagnetic exchange, and suppresses long-range magnetic order (1912.01336).

In $\sqrt{13}\times\sqrt{13}$11-TaSe$\sqrt{13}\times\sqrt{13}$12, the equilibrium star-of-David phase is a $\sqrt{13}\times\sqrt{13}$13 commensurate CDW below $\sqrt{13}\times\sqrt{13}$14 K. The six inner-ring Ta atoms move inward by $\sqrt{13}\times\sqrt{13}$15 Å, the six outer-ring Ta atoms by $\sqrt{13}\times\sqrt{13}$16 Å, and AA stacking enhances interlayer coupling and stabilizes the high-temperature commensurate order (Dharmasiri et al., 12 Feb 2026).

4. Kagome and related realizations

In KV$\sqrt{13}\times\sqrt{13}$17Sb$\sqrt{13}\times\sqrt{13}$18, STM and X-ray diffraction show a $\sqrt{13}\times\sqrt{13}$19 in-plane superstructure below $\sqrt{13}\times\sqrt{13}$20 K. Within each $\sqrt{13}\times\sqrt{13}$21 cell, six V atoms around one kagome-hexagon center move inward by $\sqrt{13}\times\sqrt{13}$22–$\sqrt{13}\times\sqrt{13}$23 Å and the remaining six move outward by roughly the same amount. Optical spectroscopy finds suppression of low-energy conductivity up to $\sqrt{13}\times\sqrt{13}$24 eV and a new peak at $\sqrt{13}\times\sqrt{13}$25 eV, giving $\sqrt{13}\times\sqrt{13}$26 meV. Uykur et al. also report strong phonon anomalies, including softening of the $\sqrt{13}\times\sqrt{13}$27 mode at $\sqrt{13}\times\sqrt{13}$28 cm$\sqrt{13}\times\sqrt{13}$29 and a linewidth described by an electron–phonon form with $\sqrt{13}\times\sqrt{13}$30 cm$\sqrt{13}\times\sqrt{13}$31. DFT-relaxed $\sqrt{13}\times\sqrt{13}$32 supercells find nearly degenerate star-of-David and tri-hexagon solutions, both remaining metallic (Uykur et al., 2021).

The wider $\sqrt{13}\times\sqrt{13}$33V$\sqrt{13}\times\sqrt{13}$34Sb$\sqrt{13}\times\sqrt{13}$35 family does not yet have a single universally accepted three-dimensional CDW structure. Several works argue for inverse star-of-David or tri-hexagonal distortions. ARPES on KV$\sqrt{13}\times\sqrt{13}$36Sb$\sqrt{13}\times\sqrt{13}$37 finds that the low-temperature band reconstruction is better captured by the inverse star-of-David pattern, with a strongly anisotropic gap reaching $\sqrt{13}\times\sqrt{13}$38 meV on the SP1 saddle-point band at $\sqrt{13}\times\sqrt{13}$39 and small three-dimensional pockets near $\sqrt{13}\times\sqrt{13}$40 only at $\sqrt{13}\times\sqrt{13}$41. Temperature-dependent X-ray absorption and first-principles energetics in CsV$\sqrt{13}\times\sqrt{13}$42Sb$\sqrt{13}\times\sqrt{13}$43 identify the inverse-star-of-David as the preferred reconstruction, with $\sqrt{13}\times\sqrt{13}$44 meV/f.u. and $\sqrt{13}\times\sqrt{13}$45 meV/f.u., and interpret V $\sqrt{13}\times\sqrt{13}$46–Sb $\sqrt{13}\times\sqrt{13}$47 orbital hybridization as the microscopic driving force (Kato et al., 2022, Han et al., 2022).

Other measurements support more complicated coexistence or stacking scenarios. ARPES combined with DFT has been interpreted as evidence that AV$\sqrt{13}\times\sqrt{13}$48Sb$\sqrt{13}\times\sqrt{13}$49 hosts intrinsic coexistence of star-of-David and tri-hexagonal distortions, naturally leading to $\sqrt{13}\times\sqrt{13}$50 or $\sqrt{13}\times\sqrt{13}$51 order and two distinct splitting scales at the same momentum. Polarization-resolved Raman, X-ray, and DFT studies on CsV$\sqrt{13}\times\sqrt{13}$52Sb$\sqrt{13}\times\sqrt{13}$53 describe a $\sqrt{13}\times\sqrt{13}$54 structure containing one inverse-star-of-David layer and three consecutive star-of-David layers, with a $\sqrt{13}\times\sqrt{13}$55 distortion as the primary order parameter and $\sqrt{13}\times\sqrt{13}$56 and $\sqrt{13}\times\sqrt{13}$57 distortions as secondary. Time-resolved reflectivity on CsV$\sqrt{13}\times\sqrt{13}$58Sb$\sqrt{13}\times\sqrt{13}$59 further finds that close phonon pairs near $\sqrt{13}\times\sqrt{13}$60 THz and $\sqrt{13}\times\sqrt{13}$61 THz arise from coexistence of star-of-David and inverse star-of-David distortions combined with six-fold rotational symmetry breaking (Hu et al., 2022, Wu et al., 2022, Deng et al., 10 Mar 2025).

NMR and NQR provide a different bulk-sensitive perspective. In CsV$\sqrt{13}\times\sqrt{13}$62Sb$\sqrt{13}\times\sqrt{13}$63, $\sqrt{13}\times\sqrt{13}$64V NMR and $\sqrt{13}\times\sqrt{13}$65Sb NQR detect a first-order commensurate transition below $\sqrt{13}\times\sqrt{13}$66 K and report that the observed charge modulation is of star-of-David pattern, not inverse star-of-David, with an additional weaker charge modulation appearing below $\sqrt{13}\times\sqrt{13}$67 K. This result is one reason structural identification remains controversial (Luo et al., 2021).

Antiferromagnetic FeGe extends the SoD concept beyond the nonmagnetic kagome metals. There the $\sqrt{13}\times\sqrt{13}$68 CDW organizes six Fe sites into a real-space star-of-David bond pattern, and in the CDW phase the ground-state current density exhibits a star-of-David loop with amplitude $\sqrt{13}\times\sqrt{13}$69 A together with counter-circulating second-neighbor loops of $\sqrt{13}\times\sqrt{13}$70 A (Ma et al., 2023).

5. Electronic reconstruction, Mottness, magnetism, and topology

A recurrent misconception is that a star-of-David CDW necessarily implies Mottness. In fact, the consequences range from Mott insulating to metallic. Monolayer $\sqrt{13}\times\sqrt{13}$71-NbSe$\sqrt{13}\times\sqrt{13}$72 is explicitly a SoD Mott insulator: in a three-band Wannier DMFT treatment at $\sqrt{13}\times\sqrt{13}$73 K, the effective interaction on the central “type I” Wannier function is $\sqrt{13}\times\sqrt{13}$74 eV, its occupancy is pushed from $\sqrt{13}\times\sqrt{13}$75, and the gap between the upper Hubbard band and the Se–Nb valence bands is reproduced. Single-layer $\sqrt{13}\times\sqrt{13}$76-NbS$\sqrt{13}\times\sqrt{13}$77 likewise develops a spin-$\sqrt{13}\times\sqrt{13}$78 insulating state. By contrast, monolayer $\sqrt{13}\times\sqrt{13}$79-NbTe$\sqrt{13}\times\sqrt{13}$80 remains a correlated metal with no signature of Mott gap, and the kagome metals remain metallic despite clear SoD- or ISD-like reconstruction (Pasquier et al., 2018, Tresca et al., 2019, Taguchi et al., 2022, Uykur et al., 2021).

The SoD motif can also host unconventional magnetism. In monolayer $\sqrt{13}\times\sqrt{13}$81-NbSe$\sqrt{13}\times\sqrt{13}$82, mapping total-energy differences to

$\sqrt{13}\times\sqrt{13}$83

gives $\sqrt{13}\times\sqrt{13}$84 K in GGA and $\sqrt{13}\times\sqrt{13}$85 K in GGA+$\sqrt{13}\times\sqrt{13}$86, with $\sqrt{13}\times\sqrt{13}$87 K and $\sqrt{13}\times\sqrt{13}$88 K, respectively. The positive nearest-neighbor sign is interpreted as a hallmark of flat-band ferromagnetism rather than conventional superexchange. Single-layer $\sqrt{13}\times\sqrt{13}$89-NbS$\sqrt{13}\times\sqrt{13}$90 similarly yields ferromagnetic inter-star couplings $\sqrt{13}\times\sqrt{13}$91 K and $\sqrt{13}\times\sqrt{13}$92 K. Conversely, in monolayer VTe$\sqrt{13}\times\sqrt{13}$93, cluster formation pushes $\sqrt{13}\times\sqrt{13}$94 toward zero or negative values and suppresses the long-range ferromagnetism predicted for the undistorted lattice (Pasquier et al., 2018, Tresca et al., 2019, 1912.01336).

Bulk $\sqrt{13}\times\sqrt{13}$95-NbS$\sqrt{13}\times\sqrt{13}$96 illustrates a different electronic consequence: the CCDW phase opens an in-plane direct gap of $\sqrt{13}\times\sqrt{13}$97 eV, but the nearly flat band just below $\sqrt{13}\times\sqrt{13}$98 retains substantial dispersion along $\sqrt{13}\times\sqrt{13}$99–$2\times2$00, producing one-dimensional metallic behavior along the stacking direction. The top valence state is dominated by the central Nb $2\times2$01 orbital, with real-space charge strongly localized on the central Nb of each star. Under pressure, the CCDW is suppressed by $2\times2$02 GPa, $2\times2$03 falls to $2\times2$04, and the Allen–Dynes estimate yields a peak $2\times2$05 K (Wang et al., 2020).

In kagome and antiferromagnetic kagome systems, the reconstructed state can be topological or orbital-current bearing. In FeGe, the orbital flux through the central hexagon is $2\times2$06, while the small triangles carry $2\times2$07. The CDW phase then hosts Weyl points protected by $2\times2$08 symmetry and a nearly flat Chern band near $2\times2$09, with robust edge modes in the partial bulk gap. In KV$2\times2$10Sb$2\times2$11, the CDW gap has periodicity of the undistorted Brillouin zone along $2\times2$12, again showing that the reconstructed low-energy electronic structure, rather than the high-symmetry lattice, is the correct basis for discussing superconductivity and anomalous transport (Ma et al., 2023, Kato et al., 2022).

6. Ultrafast control, metastability, and unresolved questions

The star-of-David lattice is not only an equilibrium order. In $2\times2$13-TaSe$2\times2$14, femtosecond pumping with $2\times2$15 eV and $2\times2$16 fs can coherently over-drive the equilibrium $2\times2$17 mode into an inverted CDW with the same periodicity but opposite sign of the displacement field. The critical fluence is $2\times2$18 mJ/cm$2\times2$19; the amplitude mode passes through the transient normal state at $2\times2$20–$2\times2$21 fs and overshoots into the inverted regime at $2\times2$22–$2\times2$23 fs. tr-ARPES and TDDFT show that the inverted state has a higher density of states at $2\times2$24 than even the normal metallic state, enhanced metallicity, and altered electron–phonon couplings (Zhang et al., 2020).

Photoinduced metastability is also seen in kagome metals. In vanadium kagome compounds, femtosecond time-resolved X-ray scattering identifies a coherent phonon at $2\times2$25 THz tied to the CDW reflection and shows that an out-of-plane Cs mode is frustrated in the CDW phase. Photoexcitation relieves that frustration, producing a metastable CDW with $2\times2$26 fs, $2\times2$27 fs, and a $2\times2$28 ns lifetime before full thermal recovery. This suggests that phononic frustration, not only static lattice energetics, is part of the star-of-David problem in kagome metals (Heo et al., 2024).

The principal unresolved issue is structural identification in AV$2\times2$29Sb$2\times2$30. The literature summarized here contains mutually inconsistent but individually well-supported assignments: star-of-David from bulk NMR/NQR; inverse star-of-David from XAS, DFT, and some ARPES analyses; coexistence of star-of-David and tri-hexagonal distortions from ARPES plus DFT; and several $2\times2$31-shifted $2\times2$32 or $2\times2$33 stacking patterns from Landau, Raman, X-ray, and effective-Hamiltonian approaches (Luo et al., 2021, Han et al., 2022, Hu et al., 2022, Li et al., 2023). A plausible implication is that the phrase “star-of-David charge-density-wave” now denotes a family of closely competing triple-$2\times2$34 states whose selection depends sensitively on interlayer coupling, orbital hybridization, electron–phonon coupling, and electronic correlations.

Across both dichalcogenides and kagome materials, the star-of-David motif therefore functions less as a unique phase label than as a geometrically recognizable endpoint of several different instability channels. Its importance lies precisely in this variability: the same cluster geometry can support a Hubbard-driven gap, a correlated metal, flat-band ferromagnetism, orbital-flux order, or a laser-created hidden state, depending on how lattice distortion, band topology, and many-body interactions are combined (Pasquier et al., 2018, Taguchi et al., 2022, Ma et al., 2023, Zhang et al., 2020).