Inverse Lieb Lattice (ILL)

• The inverse Lieb lattice (ILL) is a 2D antiperovskite structure with two crystallographically equivalent sublattices and distinct J2 paths that induce anisotropic exchange.
• It serves as a minimal analytical model for d‑wave altermagnetism, where frustration and chiral magnon splitting arise via symmetric exchange without relying on spin–orbit coupling.
• First-principles and spin-wave analyses underscore how orbital filling and exchange anisotropy determine the magnetic phase stability in layered oxides and oxychalcogenides.

Searching arXiv for the cited paper and closely related altermagnetism context. The inverse Lieb lattice (ILL) is a two-dimensional square antiperovskite motif realized by transition-metal cations in layered environments with in-plane oxygen and chalcogen ligands. Originally treated as a theoretical counterpart to the Lieb lattice proposed for cuprate superconductors, it has become a concrete materials platform in which altermagnetism, frustration, and chiral magnons arise already from short-range symmetric exchange. In contemporary usage, the ILL functions as a minimal analytical model for d-wave altermagnetism and as a crystallographic motif now identified in real layered oxides and oxychalcogenides (Chang et al., 6 Aug 2025).

1. Geometry, sublattices, and symmetry

The ILL is the “anti” of the layered perovskite square lattice associated with the canonical Lieb-lattice construction. Its transition-metal sites form a square net with two crystallographically equivalent sublattices, conventionally denoted $A$ and $B$. Nearest neighbors connect opposite sublattices via $J_1$, whereas second neighbors lie on the same sublattice and are connected by two crystallographically inequivalent paths, $J_{2a}$ and $J_{2b}$. Each transition-metal site has four nearest neighbors on the opposite sublattice and four next-nearest neighbors on the same sublattice (Chang et al., 6 Aug 2025).

The inequivalence of the next-nearest-neighbor exchanges is intrinsic to the motif. One path, $J_{2a}$, is a TM–O–TM superexchange route with a nearly $180^\circ$ bond angle; the other, $J_{2b}$, is a TM–Ch–TM route, with $\mathrm{Ch}=\mathrm{Se},\mathrm{Te}$, and bond angles near $90^\circ$. This built-in anisotropy persists even in a perfectly tetragonal lattice and is central to both the magnetic phase structure and the chiral splitting of magnons.

The symmetry criterion relevant to altermagnetism is equally specific. A mirror $B$0 relates the two spin sublattices and flips the spin projection, while inversion $B$1 maps each transition-metal sublattice onto itself. The resulting compensated collinear order breaks time reversal $B$2, has net magnetization $B$3, and nonetheless admits nonrelativistic spin splitting at generic momentum because a unitary crystal symmetry, rather than a translation or inversion, maps spin-up on $B$4 to spin-down on $B$5. In this setting, spin degeneracy is lifted generically, with $B$6 at the same $B$7.

Real-material realizations occur in layered tetragonal compounds. Examples explicitly identified for ILL planes include $B$8 in $B$9, $J_1$0 in $J_1$1, and vanadium oxychalcogenides such as $J_1$2, $J_1$3, and $J_1$4. Nonsymmorphic glide planes may also be present; in $J_1$5, glide $J_1$6 mirrors were identified as compatible with the ILL altermagnetic order.

2. Minimal spin model and phase structure

The basic magnetic description is a Heisenberg model with inequivalent next-nearest-neighbor couplings and, optionally, longer-range exchange:

$J_1$7

with the energy convention

$J_1$8

without double counting bonds and with $J_1$9 unit vectors parallel to the local moments (Chang et al., 6 Aug 2025).

A defining feature of the ILL is frustration-induced decoupling. All nearest-neighbor pairs connect opposite sublattices, whereas all next-nearest-neighbor pairs lie within the same sublattice. Consequently, a dominant antiferromagnetic $J_{2a}$0 can fully frustrate and effectively decouple the two sublattices at the Heisenberg level, making the inter-sublattice energy independent of the relative rotation angle $J_{2a}$1. This produces continuous degeneracy manifolds.

Within this model, several central phases appear. The altermagnetic phase (AM, $J_{2a}$2) is collinear, with all nearest-neighbor pairs antiparallel. It is stabilized when $J_{2a}$3 is the dominant antiferromagnetic exchange, i.e. $J_{2a}$4 if both $J_{2a}$5 terms are antiferromagnetic, or when both $J_{2a}$6 and $J_{2a}$7 are ferromagnetic of any magnitude. In particular, the entire quadrant $J_{2a}$8, $J_{2a}$9 is altermagnetic.

When both $J_{2b}$0 and $J_{2b}$1 are antiferromagnetic and dominant, the single-stripe (SS) phase appears. This collinear phase is continuously degenerate with noncollinear $J_{2b}$2–$J_{2b}$3–$J_{2b}$4 states, including the $J_{2b}$5–$J_{2b}$6–$J_{2b}$7 configuration. If one $J_{2b}$8 path is antiferromagnetic and dominant while the other is ferromagnetic, block-checkerboard phases BC1 or BC2 occur; BC1 is degenerate with the noncollinear $J_{2b}$9–$J_{2a}$0–$J_{2a}$1 family, and exchanging the signs of $J_{2a}$2 and $J_{2a}$3 yields BC2 and $J_{2a}$4–$J_{2a}$5–$J_{2a}$6. By contrast, although double-stripe order is degenerate with block-checkerboard phases in FeTe when $J_{2a}$7, it never has the lowest energy in the ILL because the lattice enforces $J_{2a}$8 anisotropy.

The principal ordering vectors are also fixed at the model level: AM has $J_{2a}$9; the $180^\circ$0–$180^\circ$1–$180^\circ$2 state and related SS order have propagation vector $180^\circ$3; and block-checkerboard phases possess $180^\circ$4 real-space periodicity, with precise $180^\circ$5-selection deferred to higher-order terms such as biquadratic interactions and single-ion anisotropy. A recurring misconception is that these states are fully determined by the Heisenberg couplings alone. In fact, the $180^\circ$6-degenerate manifolds are selected by weak anisotropies and biquadratic terms, in direct analogy to iron-based superconductors.

3. First-principles methodology and exchange extraction

The electronic-structure workflow uses numerical-orbital density functional theory in OpenMX with norm-conserving pseudopotentials and the PBE-GGA exchange-correlation functional. For gapped systems, a Hubbard-$180^\circ$7 correction is applied to transition-metal $180^\circ$8 states in an LDA+$180^\circ$9-like manner (Chang et al., 6 Aug 2025).

Exchange constants are primarily extracted through the Green’s-function, or Liechtenstein, method as implemented in OpenMX 3.9. In this linear-response approach, the $J_{2b}$0 parameters are obtained from a single self-consistent magnetic configuration, and the resulting couplings are described as robust for linear spin-wave analysis. For $J_{2b}$1, however, the exchange constants were sufficiently sensitive to the reference magnetic state that total-energy mapping onto the Heisenberg model was used instead.

Spin–orbit coupling does not play a central role in stabilizing altermagnetism in the ILL. The primary symmetry mechanism and the decisive exchange physics are nonrelativistic, and SOC was not essential for the exchange extraction. This directly separates the ILL mechanism from magnetically compensated states whose spin splitting depends on relativistic anisotropies.

Calculated magnetic ground states are reported to agree well with experimental determinations across the surveyed family. The agreement includes the noncollinear phases in $J_{2b}$2 and $J_{2b}$3, altermagnetism in $J_{2b}$4 and $J_{2b}$5, and high-$J_{2b}$6 altermagnetism in $J_{2b}$7.

4. Representative compounds and $J_{2b}$8-shell trends

A central empirical result is that altermagnetism in ILL materials is favored for $J_{2b}$9 and $\mathrm{Ch}=\mathrm{Se},\mathrm{Te}$0 configurations, whereas $\mathrm{Ch}=\mathrm{Se},\mathrm{Te}$1 fillings tend to drive strong antiferromagnetic $\mathrm{Ch}=\mathrm{Se},\mathrm{Te}$2 interactions that destabilize the AM phase in favor of stripe or block-checkerboard manifolds (Chang et al., 6 Aug 2025). Fractional $\mathrm{Ch}=\mathrm{Se},\mathrm{Te}$3-occupancy associated with metallicity promotes a ferromagnetic component in $\mathrm{Ch}=\mathrm{Se},\mathrm{Te}$4, especially $\mathrm{Ch}=\mathrm{Se},\mathrm{Te}$5, and enhances longer-range $\mathrm{Ch}=\mathrm{Se},\mathrm{Te}$6, both of which stabilize AM and elevate $\mathrm{Ch}=\mathrm{Se},\mathrm{Te}$7.

The microscopic rationale given in the source combines Hund’s coupling, crystal-field splitting, and superexchange geometry. Integer $\mathrm{Ch}=\mathrm{Se},\mathrm{Te}$8 fillings in layered tetragonal fields often support robust antiferromagnetic $\mathrm{Ch}=\mathrm{Se},\mathrm{Te}$9 through $90^\circ$0 TM–O–TM paths while allowing ferromagnetic next-nearest-neighbor exchange in itinerant settings with partial $90^\circ$1 filling. The $90^\circ$2 and $90^\circ$3 paths then generate the characteristic $90^\circ$4 anisotropy.

Compound	Exchange parameters (meV)	Ground state
$90^\circ$5 ($90^\circ$6)	$90^\circ$7	AM
$90^\circ$8 ($90^\circ$9)	$B$00	$B$01–$B$02–$B$03 noncollinear
$B$04 ($B$05)	$B$06	$B$07–$B$08–$B$09 noncollinear
$B$10 ($B$11)	$B$12	BC
$B$13 ($B$14)	$B$15	AM
$B$16 ($B$17)	$B$18	AM
$B$19 ($B$20)	$B$21	AM
$B$22 ILL plane ($B$23)	$B$24	AM

From these data, structural and chemical identification guidelines follow directly. Candidate ILL altermagnets should combine layered tetragonal antiperovskite planes with TM–O and TM–Ch links, a sizable antiferromagnetic $B$25, and appreciable anisotropy $B$26. They are especially likely among early transition metals such as V and Cr with partial $B$27 filling, or among $B$28 systems such as $B$29 in insulating oxyselenides. Metallicity correlates with ferromagnetic $B$30, larger $B$31, and higher ordering temperatures.

5. $B$32 as an ILL case study

$B$33, crystallizing in $B$34, contains alternating CrO$B$35 perovskite-like planes and Cr$B$36O ILL planes, making it a natural heterostructure in which the inverse Lieb motif coexists with a conventional checkerboard antiferromagnet (Chang et al., 6 Aug 2025). Earlier work suggested $B$37 ($B$38) in the perovskite layer and $B$39 ($B$40) in the ILL layer, whereas the cited calculations yield partial disproportionation with occupations $B$41 and $B$42, with $B$43–$B$44, consistent with metallicity.

The exchange hierarchy is unusually strong. In the ILL plane, the reported values are $B$45 meV, $B$46 meV, $B$47 meV, and $B$48 meV. In the perovskite plane, $B$49 meV and $B$50 meV stabilize checkerboard antiferromagnetism. The material exhibits very high ordering temperatures: neutron data resolve ordering slightly above $B$51 K for the $B$52 sublattice and $B$53 K for $B$54, consistent with the large antiferromagnetic $B$55 values.

Its electronic structure is metallic and shows what the source describes as ultra-strong spin anisotropy of the Fermi surface. DFT Fermi-surface calculations identify two sets of spin-split pockets: one near $B$56, relatively isotropic near $B$57 but becoming anisotropic away from $B$58, and another near Brillouin-zone corners with propeller-like cross-sections on $B$59, where the momentum separation between spin channels is large. The spin textures of the two channels are related by a $B$60 rotation, which is identified as the hallmark of d-wave altermagnetism in the ILL.

The compound is further characterized as a supercell altermagnet. The magnetic unit cell is doubled by the checkerboard antiferromagnetism of the perovskite layers, so that $B$61 globally, while the ILL plane itself retains $B$62 altermagnetism. The global spin point group of the AM ILL, $B$63, remains intact because compatible glide mirrors and fourfold axes through oxygen sites survive when the perovskite layers are incorporated. This suggests that altermagnetic symmetry can persist in layered composite structures even when auxiliary magnetic subsystems enlarge the magnetic cell.

6. Magnons, experimental signatures, and limitations

For the two-sublattice collinear AM state, the linear spin-wave spectrum is written as

$B$64

where $B$65 and $B$66 are lattice structure factors formed from exchange sums. In the ILL, the inequivalent next-nearest-neighbor couplings enter $B$67 asymmetrically on each sublattice, producing $B$68. The chiral or nonreciprocal splitting therefore emerges entirely from symmetric exchange rather than from Dzyaloshinskii–Moriya interactions. Its magnitude scales with $B$69 anisotropy:

$B$70

with $B$71 a geometry-dependent structure factor that is maximal near $B$72 and along $B$73 (Chang et al., 6 Aug 2025).

The material dependence of this effect follows the exchange anisotropy. $B$74, where $B$75, shows only subtle chiral splitting. $B$76, with $B$77 meV and $B$78 meV, exhibits pronounced splitting along high-symmetry lines. $B$79 has the strongest reported splitting, reaching approximately $B$80 of the total magnon bandwidth near $B$81, exceeding half the bandwidth, with weak and frustrated interlayer couplings such that the essential chiral physics is captured by the ILL plane alone.

The expected experimental signatures are correspondingly diverse. Spin-polarized ARPES should reveal momentum-dependent spin splitting with spin textures related by $B$82 rotation and a d-wave spin-momentum locking, as already demonstrated in $B$83. In transport, the cited work points to giant and tunneling magnetoresistance without net magnetization, anisotropic magnetoresistance in zero applied field, and spin-density-wave modulations in metals. In inelastic neutron scattering, the crucial signature is nonreciprocal magnon dispersion $B$84 in the absence of DM interaction, with splitting that tracks the anisotropy $B$85.

The same study also delineates the present theoretical limits. The Heisenberg-only $B$86–$B$87–$B$88 model captures the broad phase boundaries and extensive degeneracies but cannot by itself select among degenerate noncollinear states; single-ion anisotropy, anisotropic exchange, and biquadratic terms are required and may be material-specific. SOC is not necessary for the basic altermagnetic splitting, though weak SOC and crystalline anisotropies fix spin orientations and can renormalize magnon dispersions and gap pseudo-Goldstone modes. Metallic versus insulating character, stoichiometry, filler layers, and strain modify $B$89 anisotropy and $B$90, potentially moving a compound into or out of the AM regime. A plausible implication is that the ILL should be viewed not as a single magnetic phase but as a tunable symmetry-constrained family in which geometry, exchange anisotropy, and filling jointly determine whether the realized state is altermagnetic, noncollinear, or block ordered.