Quantum Corner Symmetries

• Quantum corner symmetries are a framework where codimension-2 boundaries encode nonlocal symmetry data beyond bulk observables.
• They manifest in tensor networks, higher-order topological phases, and gravitational theories with distinct spectral, crystalline, and algebraic signatures.
• These symmetries provide practical diagnostics for phase transitions and enable controlled manipulation of quantum states through geometric and symmetry principles.

Searching arXiv for recent and foundational papers on quantum corner symmetries across gravity, higher-order topology, tensor networks, and related algebraic frameworks. Quantum corner symmetries denote a family of research programs in which codimension-2 corners, tensor-network corner objects, or higher-order topological corner states carry symmetry data that is not exhausted by bulk local observables. In current usage, the phrase encompasses at least three technically distinct settings: tensor-network corner transfer matrices and corner tensors whose spectra encode universality and edge conformal structure; corner-localized states in higher-order topological phases whose existence, counting, and manipulability are governed by crystalline, spectral, or refined subsystem symmetries; and the corner-charge algebras of gravity, where gauge transformations acquire non-vanishing Noether charges on codimension-2 surfaces and become physical symmetries (Huang et al., 2017, Imhof et al., 2017, Freidel et al., 2023, Ciambelli et al., 2024). Taken together, these directions suggest a common theme: corners act as symmetry-sensitive interfaces at which global, emergent, or gauge-theoretic structure becomes sharply encoded.

1. Conceptual scope of the term

In tensor-network many-body physics, “corner” refers to corner transfer matrices (CTMs) in two dimensions and corner tensors (CTs) in three dimensions. There the central claim is not that corners support an abstract standalone symmetry algebra, but that universal bulk information is encoded holographically in corner spectra, including degeneracy multiplets, spectral branches, and conformal-tower counting (Huang et al., 2017). In higher-order topological matter, by contrast, the corner is a literal codimension-2 boundary at which localized states appear, and “corner symmetry” refers to the crystalline, chiral, mirror, rotational, or refined subchiral structures that protect, stabilize, or organize those states (Imhof et al., 2017, Wang et al., 2022, Ding et al., 16 Jun 2026).

In gravity and quantum field theory, the term is more algebraic. A corner is a codimension-2 surface carrying Noether charges associated with gauge transformations that cease to be pure redundancies in bounded regions. This leads to finite and infinite-dimensional corner symmetry algebras, central extensions, induced representations, and gluing rules for local subsystems (Freidel et al., 2023, Ciambelli et al., 2023, Ciambelli et al., 2024). A broader QFT reinterpretation places a $d$-dimensional theory itself at the corner of a nested higher-dimensional bulk-edge system of relative symmetry theories, so that symmetry data descends through a hierarchy of boundaries and corners rather than through a single symmetry topological field theory (Cvetič et al., 2024).

A recurrent misconception is that all corner phenomena are of the same kind. The literature instead separates at least three logically distinct notions: corner spectra as holographic diagnostics, corner-localized topological states in higher-order phases, and corner charge algebras in gravity. Their commonality is structural rather than definitional.

2. Corner spectra, holography, and emergent symmetry in tensor networks

The tensor-network formulation begins from the observation that corner objects used for contraction algorithms also define physically meaningful spectra. For a two-dimensional tensor network on a square lattice,

$Z = {\rm tr}(C_1C_2C_3C_4),$

and in the symmetric case $C_1=C_2=C_3=C_4\equiv C$. If $C_d=PCP^{-1}$ has eigenvalues $\nu_\alpha$, then

$$Z={\rm tr}(C_d^4)=\sum_{\alpha=1}^{\chi}\nu_\alpha^4.$$

This motivates a corner Hamiltonian

$Z={\rm tr}(e^{-H_C}),\qquad H_C\equiv-\log(C_d^4),$

with corner energies $\varepsilon_\alpha=-\log\omega_\alpha$, $\omega_\alpha=\nu_\alpha^4$. In one-dimensional quantum systems, the same data reproduce the half-chain entanglement spectrum through

$$\lambda_\alpha=\nu_\alpha^2,\qquad \omega_\alpha=\lambda_\alpha^2=\nu_\alpha^4,$$

so the corner spectrum is directly the reduced-density-matrix spectrum of a half-infinite chain (Huang et al., 2017).

This construction yields a practical holographic encoding of universality. The same paper shows that corner spectra and corner entropies numerically coincide across matched $Z = {\rm tr}(C_1C_2C_3C_4),$0-dimensional quantum and $Z = {\rm tr}(C_1C_2C_3C_4),$1-dimensional classical systems, including Ising, Potts, and XY/Ising correspondences, and that corner spectra diagnose phase transitions “without observables” (Huang et al., 2017). For the $Z = {\rm tr}(C_1C_2C_3C_4),$2 quantum XXZ model at $Z = {\rm tr}(C_1C_2C_3C_4),$3, two transitions appear at $Z = {\rm tr}(C_1C_2C_3C_4),$4 and $Z = {\rm tr}(C_1C_2C_3C_4),$5; for the perturbed $Z = {\rm tr}(C_1C_2C_3C_4),$6 toric-code PEPS, the topological-to-trivial transition is pinpointed near $Z = {\rm tr}(C_1C_2C_3C_4),$7; for the perturbed $Z = {\rm tr}(C_1C_2C_3C_4),$8 topological PEPS, the trivial phase has only one nonzero eigenvalue, while transitions occur at $Z = {\rm tr}(C_1C_2C_3C_4),$9 and $C_1=C_2=C_3=C_4\equiv C$0 (Huang et al., 2017).

The most explicit symmetry result concerns chiral topological PEPS. Using a single-layer CT renormalization scheme, the half-system entanglement spectrum extracted from corner tensors reproduces the vacuum conformal-tower counting of edge $C_1=C_2=C_3=C_4\equiv C$1 Wess–Zumino–Witten theories. For the $C_1=C_2=C_3=C_4\equiv C$2 PEPS, the largest multiplets have degeneracies $C_1=C_2=C_3=C_4\equiv C$3, matching the vacuum Virasoro tower; for the $C_1=C_2=C_3=C_4\equiv C$4 PEPS, the largest multiplets tend to $C_1=C_2=C_3=C_4\equiv C$5 (Huang et al., 2017). In this setting, “quantum corner symmetry” means that the emergent chiral edge symmetry algebra is encoded in the corner-derived entanglement spectrum. The method does not furnish a fully symmetry-resolved decomposition by momentum or topological charge, and the paper explicitly notes that it does not provide a natural momentum labeling, but it does show that spectral counting and degeneracy structure at corners recover bulk universality and boundary conformal symmetry (Huang et al., 2017).

3. Higher-order topological corner modes: crystalline and spectral protection

A canonical higher-order setting is the quadrupole insulator and its classical analogues. In topolectrical circuits realizing the Benalcazar–Bernevig–Hughes quadrupole model, the protecting symmetry group includes two non-commuting reflections $C_1=C_2=C_3=C_4\equiv C$6 and $C_1=C_2=C_3=C_4\equiv C$7, a $C_1=C_2=C_3=C_4\equiv C$8 rotation symmetry, and a chiral symmetry $C_1=C_2=C_3=C_4\equiv C$9 that pins topological boundary modes to zero admittance. At the design frequency $C_d=PCP^{-1}$0, the circuit Laplacian satisfies

$C_d=PCP^{-1}$1

$C_d=PCP^{-1}$2

and

$C_d=PCP^{-1}$3

with $C_d=PCP^{-1}$4, $C_d=PCP^{-1}$5, and $C_d=PCP^{-1}$6 (Imhof et al., 2017). The phase transition occurs at $C_d=PCP^{-1}$7, with $C_d=PCP^{-1}$8 topological and $C_d=PCP^{-1}$9 trivial, and the corner-state profile obeys

$\nu_\alpha$0

Experimentally, the corner mode appears as a topological boundary resonance in the corner impedance profile (Imhof et al., 2017).

A photonic counterpart sharpens the distinction between approximate spectral symmetry and actual protection. In a $\nu_\alpha$1-symmetric breathing honeycomb photonic crystal, topological quantum chemistry and Wilson-loop analysis show that the expanded phase is an obstructed atomic limit with Wannier centers at the $\nu_\alpha$2 Wyckoff position, whereas the contracted phase is trivial with Wannier centers at $\nu_\alpha$3 (Proctor et al., 2020). The finite hexagonal particle exhibits six corner states nested inside the edge gap with $\nu_\alpha$4-$\nu_\alpha$5-$\nu_\alpha$6-$\nu_\alpha$7 multiplet structure. Crucially, long-range photonic interactions break chiral symmetry by generating same-sublattice couplings, shifting corner modes away from exact mid-gap and splitting the sixfold degeneracy. The paper’s conclusion is therefore that the corner modes are protected by lattice symmetries—specifically $\nu_\alpha$8+mirror/$\nu_\alpha$9—rather than by exact chiral symmetry (Proctor et al., 2020).

These works establish two durable principles. First, higher-order corner modes are not simply “corners plus a gap,” but require symmetry-compatible boundary terminations. Second, exact zero-energy pinning and robust corner localization are distinct properties: the former typically needs a spectral symmetry such as chirality, whereas the latter can survive when the true protection mechanism is crystalline.

4. Refined corner-state symmetry: remote control, non-topological corner binding, quasicrystals, and subchiral sectors

More recent work complicates the canonical HOTI picture by showing that corner-state structure can be nonlocal, orientation dependent, or symmetry resolved beyond conventional chiral protection. In a rhombus-shaped Kekulé flake with $$Z={\rm tr}(C_d^4)=\sum_{\alpha=1}^{\chi}\nu_\alpha^4.$$0, the unperturbed HOTI phase has parity invariant $$Z={\rm tr}(C_d^4)=\sum_{\alpha=1}^{\chi}\nu_\alpha^4.$$1 and supports in-gap corner modes only at the two $$Z={\rm tr}(C_d^4)=\sum_{\alpha=1}^{\chi}\nu_\alpha^4.$$2 corners, while the $$Z={\rm tr}(C_d^4)=\sum_{\alpha=1}^{\chi}\nu_\alpha^4.$$3 corners are excluded by “chiral charge cancellation” in the bipartite lattice (Zhou et al., 2022). Because the low-energy states are bonding and antibonding combinations delocalized across the two active corners, a local perturbation at one corner reorganizes a nonlocal two-corner Hilbert space. With local magnetization and local electric potential applied to one corner, the remote corner becomes strongly spin polarized, and the sign of the spin polarization can be reversed at both corners (Zhou et al., 2022). The symmetry protecting the existence of the corner sector is not time reversal; the paper’s explicit bulk diagnosis is inversion based, and the corner selectivity depends additionally on bipartite/chiral structure and corner geometry.

A complementary correction comes from magnetic-field-induced corner states in quantum spin Hall insulators. For realistic zinc-blende quantum wells, the analytic edge theory has a two-component mass vector,

$$Z={\rm tr}(C_d^4)=\sum_{\alpha=1}^{\chi}\nu_\alpha^4.$$4

and the existence of a corner state is controlled not by a generic higher-order invariant but by the domain-wall criterion

$$Z={\rm tr}(C_d^4)=\sum_{\alpha=1}^{\chi}\nu_\alpha^4.$$5

Mirror-graded winding numbers exist only for special high-symmetry edge orientations $$Z={\rm tr}(C_d^4)=\sum_{\alpha=1}^{\chi}\nu_\alpha^4.$$6 and in-plane magnetic fields perpendicular to the relevant mirror line, $$Z={\rm tr}(C_d^4)=\sum_{\alpha=1}^{\chi}\nu_\alpha^4.$$7 mod $$Z={\rm tr}(C_d^4)=\sum_{\alpha=1}^{\chi}\nu_\alpha^4.$$8 (Krishtopenko et al., 2023). Outside those configurations, corner states still often occur, but as generic domain-wall bound states rather than symmetry-protected higher-order topological modes. The paper therefore argues that QSHIs in a magnetic field should not generically be viewed as HOTIs (Krishtopenko et al., 2023).

Quasicrystalline systems replace translational symmetry by rotational symmetry. Using a pseudo-Brillouin-zone $$Z={\rm tr}(C_d^4)=\sum_{\alpha=1}^{\chi}\nu_\alpha^4.$$9 theory for Penrose and Ammann–Beenker tilings, one obtains edge masses whose phase depends on edge orientation and field direction,

$Z={\rm tr}(e^{-H_C}),\qquad H_C\equiv-\log(C_d^4),$0

so adjacent edges differ by a fractional mass kink $Z={\rm tr}(e^{-H_C}),\qquad H_C\equiv-\log(C_d^4),$1. This binds corner charge

$Z={\rm tr}(e^{-H_C}),\qquad H_C\equiv-\log(C_d^4),$2

giving $Z={\rm tr}(e^{-H_C}),\qquad H_C\equiv-\log(C_d^4),$3 for the pentagonal Penrose case and $Z={\rm tr}(e^{-H_C}),\qquad H_C\equiv-\log(C_d^4),$4 for the octagonal Ammann–Beenker case (Wang et al., 2022). In the superconducting extension, tuning the in-plane Zeeman field and chemical potential yields Majorana corner modes. The symmetry principle here is rotational rather than translational or mirror based, and the corner response is phase-texture protection rather than simple mass-sign inversion (Wang et al., 2022).

An even finer symmetry resolution is provided by subchiral symmetry in the BBH model. The conventional chiral operator

$Z={\rm tr}(e^{-H_C}),\qquad H_C\equiv-\log(C_d^4),$5

is decomposed as

$Z={\rm tr}(e^{-H_C}),\qquad H_C\equiv-\log(C_d^4),$6

with projected relations

$Z={\rm tr}(e^{-H_C}),\qquad H_C\equiv-\log(C_d^4),$7

Each of the four zero-energy corner modes is tied to one such sector. By adding orbital-selective intercell hoppings

$Z={\rm tr}(e^{-H_C}),\qquad H_C\equiv-\log(C_d^4),$8

one selectively lifts the corresponding corner mode, with energy shift

$Z={\rm tr}(e^{-H_C}),\qquad H_C\equiv-\log(C_d^4),$9

while leaving the other corner modes pinned up to exponentially small finite-size corrections (Ding et al., 16 Jun 2026). This permits adiabatic transfer of a single corner state or a coherent superposition between corners, and the protocols were implemented numerically and on an IBM quantum processor with high fidelity (Ding et al., 16 Jun 2026). In this sense, “corner symmetry” becomes an operational control principle rather than merely a classificatory invariant.

5. Corner observables, angle functions, and many-body quantum geometry

Corner symmetry also appears in subregion observables rather than localized boundary states. For $\varepsilon_\alpha=-\log\omega_\alpha$0-symmetric many-body systems, the second cumulant of the disorder operator defines bipartite charge fluctuations,

$\varepsilon_\alpha=-\log\omega_\alpha$1

and sharp corners contribute a subleading term $\varepsilon_\alpha=-\log\omega_\alpha$2 beyond the boundary law (Wu et al., 2024). Earlier isotropic results suggested a single universal angle function

$\varepsilon_\alpha=-\log\omega_\alpha$3

but the general analysis for charge insulators shows that the corner term depends on both opening angle $\varepsilon_\alpha=-\log\omega_\alpha$4 and orientation $\varepsilon_\alpha=-\log\omega_\alpha$5, with harmonic decomposition

$\varepsilon_\alpha=-\log\omega_\alpha$6

The $\varepsilon_\alpha=-\log\omega_\alpha$7 are orientation-resolved universal angle functions, whereas the coefficients $\varepsilon_\alpha=-\log\omega_\alpha$8 are generally non-universal (Wu et al., 2024).

The small-angle limit collapses much of this complexity. If the inversion-even structure factor begins as

$\varepsilon_\alpha=-\log\omega_\alpha$9

then

$\omega_\alpha=\nu_\alpha^4$0

where $\omega_\alpha=\nu_\alpha^4$1 is the unit vector perpendicular to the lower edge of the wedge (Wu et al., 2024). Orientation averaging projects out the anisotropic harmonics and yields

$\omega_\alpha=\nu_\alpha^4$2

The coefficient $\omega_\alpha=\nu_\alpha^4$3 is the many-body quantum metric, so sharp-corner fluctuations become a real-space probe of many-body quantum geometry rather than merely a symmetry-protected scalar response (Wu et al., 2024).

The same work derives universal bounds,

$\omega_\alpha=\nu_\alpha^4$4

and shows how saturation occurs in anisotropic Landau levels and for a broad class of fractional quantum Hall wavefunctions, including unprojected parton states and composite-fermion Fermi sea wavefunctions (Wu et al., 2024). This shifts the language of “corner symmetry” toward an interplay between shape, orientation, anisotropy, and quantum geometry.

6. Gravitational corner symmetry, edge modes, and quantum geometry

In gravity, corner symmetry is a boundary-charge algebra. Noether’s second theorem implies that gauge symmetries can acquire non-vanishing Noether charges on codimension-2 surfaces, and these surfaces are the corners (Freidel et al., 2023). In the metric formulation, the kinematical corner group is

$\omega_\alpha=\nu_\alpha^4$5

while the extended version includes normal translations,

$\omega_\alpha=\nu_\alpha^4$6

In first-order gravity the corner group is enlarged further by internal Lorentz rotations to

$\omega_\alpha=\nu_\alpha^4$7

and the paper argues that quantum gravitational states across an entangling cut decompose into matched representations of this corner algebra (Freidel et al., 2023).

A foundational classical result is that bulk-equivalent formulations of gravity differ by corner symplectic potentials and therefore by the sectors of corner symmetry they represent non-trivially. Metric Einstein–Hilbert and canonical GR share the same bulk symplectic structure but differ by a relative corner potential

$\omega_\alpha=\nu_\alpha^4$8

with corner symplectic form

$\omega_\alpha=\nu_\alpha^4$9

As a result, canonical GR realizes only $$\lambda_\alpha=\nu_\alpha^2,\qquad \omega_\alpha=\lambda_\alpha^2=\nu_\alpha^4,$$0 non-trivially, whereas Einstein–Hilbert realizes the larger $$\lambda_\alpha=\nu_\alpha^2,\qquad \omega_\alpha=\lambda_\alpha^2=\nu_\alpha^4,$$1 algebra, and the area density becomes the $$\lambda_\alpha=\nu_\alpha^2,\qquad \omega_\alpha=\lambda_\alpha^2=\nu_\alpha^4,$$2 Casimir (Freidel et al., 2020). In Einstein–Cartan–Holst gravity, the corner symplectic structure acquires additional terms,

$$\lambda_\alpha=\nu_\alpha^2,\qquad \omega_\alpha=\lambda_\alpha^2=\nu_\alpha^4,$$3

which is the starting point for the first-order quantum-geometry program (Freidel et al., 2020).

The wider “corner proposal” states that gravity is described by a set of charges and their algebra at corners. In a local analysis near a codimension-2 surface, this leads to the universal corner symmetry group

$$\lambda_\alpha=\nu_\alpha^2,\qquad \omega_\alpha=\lambda_\alpha^2=\nu_\alpha^4,$$4

Because ordinary covariant phase space generically yields non-integrable charges due to corner flux, an extended phase space is introduced with an embedding/edge-mode variable $$\lambda_\alpha=\nu_\alpha^2,\qquad \omega_\alpha=\lambda_\alpha^2=\nu_\alpha^4,$$5 satisfying $$\lambda_\alpha=\nu_\alpha^2,\qquad \omega_\alpha=\lambda_\alpha^2=\nu_\alpha^4,$$6. The resulting extended symplectic form,

$$\lambda_\alpha=\nu_\alpha^2,\qquad \omega_\alpha=\lambda_\alpha^2=\nu_\alpha^4,$$7

makes all diffeomorphisms integrable generators (Ciambelli et al., 2023). In this view, asymptotic BMS and BMSW symmetries become special examples of corner symmetries, and memory effects and soft theorems follow from the corresponding charge-flux algebra (Ciambelli et al., 2023).

The finite-dimensional quantum corner symmetry program specializes the per-point algebra to

$$\lambda_\alpha=\nu_\alpha^2,\qquad \omega_\alpha=\lambda_\alpha^2=\nu_\alpha^4,$$8

with generators $$\lambda_\alpha=\nu_\alpha^2,\qquad \omega_\alpha=\lambda_\alpha^2=\nu_\alpha^4,$$9, and then identifies a unique non-trivial central extension in the translation sector,

$Z = {\rm tr}(C_1C_2C_3C_4),$00

This yields the quantum corner symmetry algebra

$Z = {\rm tr}(C_1C_2C_3C_4),$01

with central Casimir $Z = {\rm tr}(C_1C_2C_3C_4),$02 and modified cubic Casimir $Z = {\rm tr}(C_1C_2C_3C_4),$03 (Ciambelli et al., 2024). A family of unitary irreducible representations is constructed from the Stone–von Neumann representation of $Z = {\rm tr}(C_1C_2C_3C_4),$04,

$Z = {\rm tr}(C_1C_2C_3C_4),$05

and the metaplectic action

$Z = {\rm tr}(C_1C_2C_3C_4),$06

while gluing of corners is implemented by matching a maximal commuting subalgebra and defining the entangling product

$Z = {\rm tr}(C_1C_2C_3C_4),$07

(Ciambelli et al., 2024). Here “quantum corner symmetry” is literal: it is a symmetry algebra of quantum gravitational boundary data.

7. Representation theory, semiclassical limits, and algebraic generalizations

Subsequent work develops this gravitational and algebraic picture into a full representation-theoretic program. The full representation theory of the two-dimensional extended corner symmetry group, including projective representations corresponding to QCS, shows that quantum corner states are described by one-dimensional conformal fields with an additional harmonic-oscillator Fock-space index. Projective ECS representations become ordinary unitary representations of

$Z = {\rm tr}(C_1C_2C_3C_4),$08

with Hilbert spaces of the form $Z = {\rm tr}(C_1C_2C_3C_4),$09 (Varrin, 2024). An independent orbit-method derivation reaches the same conclusion by showing that the coadjoint orbits of $Z = {\rm tr}(C_1C_2C_3C_4),$10 factorize, in adapted coordinates, into an $Z = {\rm tr}(C_1C_2C_3C_4),$11 orbit and a Heisenberg orbit, so geometric quantization reproduces the previously known irreducible unitary representations (Neri et al., 14 Jul 2025).

The semiclassical limit of this corner-symmetry quantum mechanics is then built using generalized Perelomov coherent states and Berezin quantization. For the positive discrete-series representations $Z = {\rm tr}(C_1C_2C_3C_4),$12, the relevant group is again

$Z = {\rm tr}(C_1C_2C_3C_4),$13

and coherent-state expectation values of Lie-algebra generators become classical observables on QCS coadjoint orbits. The explicit scaling

$Z = {\rm tr}(C_1C_2C_3C_4),$14

keeps the classical Casimir finite as $Z = {\rm tr}(C_1C_2C_3C_4),$15, and the expectation value of the quantum Casimir tends to the classical one, $Z = {\rm tr}(C_1C_2C_3C_4),$16 (Varrin, 29 Oct 2025). This turns corner symmetry into a mathematically controlled route from quantum representation theory to classical corner phase space.

Two further algebraic directions extend the notion of corner symmetry beyond gravity. Relative symmetry theories place a $Z = {\rm tr}(C_1C_2C_3C_4),$17-dimensional QFT at a corner of a $Z = {\rm tr}(C_1C_2C_3C_4),$18-dimensional nested bulk-edge system, replacing a single SymTFT by a filtration of relative theories and coupled edge modes (Cvetič et al., 2024). In a distinct but related operator-algebraic development, quantum corner polynomials are introduced as partially symmetric polynomials corresponding to quantum corner VOAs $Z = {\rm tr}(C_1C_2C_3C_4),$19; they generalize Sergeev–Veselov super Macdonald polynomials and realize corner symmetry as blockwise partial symmetricity across three variable sectors (Cheewaphutthisakun et al., 17 Aug 2025).

Taken together, these works indicate that quantum corner symmetries are no longer a single specialized phrase. They now designate a broad technical landscape in which corners serve as carriers of conformal data, crystalline and higher-order topological structure, many-body geometric response, or genuine boundary symmetry algebras. The unifying lesson is not that all corners obey one universal formalism, but that corners repeatedly provide the smallest locus at which bulk organization becomes spectrally, geometrically, or algebraically explicit.