Chern Number Identities in Geometry & Physics

• Chern number identities are exact relations that express top-degree characteristic classes via diverse methods across algebraic geometry, toric, and combinatorial theories.
• They are established using techniques like ring-level cohomological formulas, index theory with twisted operators, and combinatorial residue computations.
• These identities have practical implications in areas ranging from curvature analysis in complex surfaces to experimental diagnostics in topological band theory and photonic crystals.

Chern number identities are relations in which top-degree characteristic numbers are expressed through other Chern monomials, through indices of elliptic operators, through combinatorial data of fans, flats, or fixed-point weights, or through physical topological invariants and response functions. In current usage, the phrase ranges from exact cohomological equalities such as

$c_k c_{n-k}=\mu_k(n)\,c_n$

on the permutohedral variety, to additivity rules such as

$C_D=C_1+C_2-C_3,$

to measurement formulas in which a Chern number is extracted from opacity, linking number, or magnetic sub-band splitting (Kuwata, 24 Oct 2025, Łącki et al., 2020, Molignini et al., 2022). This breadth is mathematically significant: the same invariant appears in algebraic geometry, toric and matroid theory, index theory, and topological band theory, but the corresponding “identity” may be a ring-level equality, a topological invariance theorem, a residue formula, or a response relation.

1. Ring-level identities on complex varieties

A particularly explicit ring-level identity was established for the permutohedral variety $X_{A_n}$, the smooth projective toric variety associated with the root system of type $A_n$. Using purely combinatorial methods, the product of Chern classes satisfies

$c_k c_{n-k}=\mu_k(n)\,c_n,$

in $H^*(X_{A_n};\mathbb{Q})$, with

$$\mu_k(n)=\sum_{j=0}^{\lfloor k/2 \rfloor}\left(\frac{1}{12}\right)^j \binom{k-j}{j}\binom{n-k-j}{j}.$$

Pairing with the fundamental class yields

$$\langle c_k c_{n-k},[X_{A_n}]\rangle=(n+1)!\,\mu_k(n),$$

because $\langle c_n,[X_{A_n}]\rangle=(n+1)!$ (Kuwata, 24 Oct 2025). The coefficient is obtained by counting block decompositions of exponent vectors arising in the expansion of $c_kc_{n-k}$, together with recursive reductions involving the factors $C_D=C_1+C_2-C_3,$0 and $C_D=C_1+C_2-C_3,$1. In this setting, the identity is literally an equality in the rational cohomology ring, not merely a numerical relation after integration.

Compact complex surfaces furnish another class of exact identities, now linking Chern numbers to curvature and torsion of an Hermitian metric. For a compact complex surface $C_D=C_1+C_2-C_3,$2, one formula is

$C_D=C_1+C_2-C_3,$3

together with the curvature-torsion identity

$C_D=C_1+C_2-C_3,$4

An alternative formula is

$C_D=C_1+C_2-C_3,$5

These identities are used to prove that if $C_D=C_1+C_2-C_3,$6 is a compact Riemannian four-manifold with constant scalar curvature and admits a compatible complex structure $C_D=C_1+C_2-C_3,$7 such that the complexified Ricci curvature is a non-positive $C_D=C_1+C_2-C_3,$8 form, then $C_D=C_1+C_2-C_3,$9 is a Kähler surface (Yang, 15 Aug 2025).

2. Index-theoretic determination and topological constraints

A central topological question asks which linear combinations of Chern numbers are determined by the underlying smooth manifold. For smooth complex projective varieties, a rational linear combination of Chern numbers is an oriented diffeomorphism invariant if and only if it is a linear combination of the Euler and Pontryagin numbers. In complex dimension $X_{A_n}$0, a rational linear combination of Chern numbers is a diffeomorphism invariant if and only if it is a multiple of the Euler number $X_{A_n}$1. In the same framework, the subspace spanned by Euler and Pontryagin numbers intersects the subspace spanned by Hirzebruch–Todd numbers in the span of the Euler number and the signature (Kotschick, 2011). These results separate genuinely topological identities from identities that depend on holomorphic structure.

A sharper finite-ambiguity statement holds for smooth complex projective varieties of dimension at least four. In complex dimension $X_{A_n}$2, the only Chern numbers determined up to finite ambiguity by the underlying smooth manifold are $X_{A_n}$3, $X_{A_n}$4, and $X_{A_n}$5. In complex dimension $X_{A_n}$6, only $X_{A_n}$7 and $X_{A_n}$8 have this property. The dimension of the space of linear combinations determined up to finite ambiguity is at most

$X_{A_n}$9

and the bound is optimal in dimension $A_n$0. In that dimension, the identity

$A_n$1

explains why $A_n$2 is controlled by Pontryagin and Euler data (Schreieder et al., 2015).

Index theory provides constructive identities of a different type. Libgober and Wood showed that $A_n$3 is determined by the Hirzebruch $A_n$4-genus, and a direct proof proceeds by expanding the coefficient of $A_n$5:

$A_n$6

For compact spin almost-complex manifolds, further Chern numbers are recovered from twisted Dirac and signature indices. Among the explicit identities are

$A_n$7

and

$A_n$8

A byproduct is the divisibility statement that

$A_n$9

is divisible by $c_k c_{n-k}=\mu_k(n)\,c_n,$0 for compact spin almost-complex $c_k c_{n-k}=\mu_k(n)\,c_n,$1 (Li, 2010).

3. Combinatorial, toric, and matroidal formulas

For complex flag manifolds $c_k c_{n-k}=\mu_k(n)\,c_n,$2, Bott’s residue formula converts Chern number calculations into explicit sums over isolated fixed points of a circle action. If $c_k c_{n-k}=\mu_k(n)\,c_n,$3 denotes the tangent weights at the fixed point indexed by a decomposition $c_k c_{n-k}=\mu_k(n)\,c_n,$4, then for a symmetric polynomial $c_k c_{n-k}=\mu_k(n)\,c_n,$5,

$c_k c_{n-k}=\mu_k(n)\,c_n,$6

If $c_k c_{n-k}=\mu_k(n)\,c_n,$7, then $c_k c_{n-k}=\mu_k(n)\,c_n,$8. If $c_k c_{n-k}=\mu_k(n)\,c_n,$9, then $H^*(X_{A_n};\mathbb{Q})$0 is constant and equals the corresponding Chern number. In particular,

$H^*(X_{A_n};\mathbb{Q})$1

and the Euler characteristic is

$H^*(X_{A_n};\mathbb{Q})$2

These are combinatorial identities obtained from differential and complex geometry rather than from Schubert calculus alone (Li et al., 2017).

Matroid theory supplies a different combinatorial incarnation. Mannino defines Chern numbers of a matroid $H^*(X_{A_n};\mathbb{Q})$3 via intersections of matroid Chern–Schwartz–MacPherson cycles of López de Medrano, Rincón, and Shaw:

$H^*(X_{A_n};\mathbb{Q})$4

If $H^*(X_{A_n};\mathbb{Q})$5 is realizable as a complex hyperplane arrangement, these coincide with the Chern numbers of the log cotangent bundle of the wonderful compactification. For simple rank $H^*(X_{A_n};\mathbb{Q})$6 matroids,

$H^*(X_{A_n};\mathbb{Q})$7

and the ratio satisfies

$H^*(X_{A_n};\mathbb{Q})$8

If the matroid is orientable, the upper bound improves to $H^*(X_{A_n};\mathbb{Q})$9 (Mannino, 2023).

For toric spaces associated with transpolar pairs of VEX multitopes, standard toric characteristic identities are extended beyond the convex Fano/reflexive setting. The total Chern class is

$$\mu_k(n)=\sum_{j=0}^{\lfloor k/2 \rfloor}\left(\frac{1}{12}\right)^j \binom{k-j}{j}\binom{n-k-j}{j}.$$0

the top Chern number is

$$\mu_k(n)=\sum_{j=0}^{\lfloor k/2 \rfloor}\left(\frac{1}{12}\right)^j \binom{k-j}{j}\binom{n-k-j}{j}.$$1

and the anticanonical self-intersection is

$$\mu_k(n)=\sum_{j=0}^{\lfloor k/2 \rfloor}\left(\frac{1}{12}\right)^j \binom{k-j}{j}\binom{n-k-j}{j}.$$2

Mirror-symmetric transpolar duality exchanges these numbers:

$$\mu_k(n)=\sum_{j=0}^{\lfloor k/2 \rfloor}\left(\frac{1}{12}\right)^j \binom{k-j}{j}\binom{n-k-j}{j}.$$3

The same framework includes generalized Todd–Hirzebruch identities such as the $$\mu_k(n)=\sum_{j=0}^{\lfloor k/2 \rfloor}\left(\frac{1}{12}\right)^j \binom{k-j}{j}\binom{n-k-j}{j}.$$4-, $$\mu_k(n)=\sum_{j=0}^{\lfloor k/2 \rfloor}\left(\frac{1}{12}\right)^j \binom{k-j}{j}\binom{n-k-j}{j}.$$5-, and $$\mu_k(n)=\sum_{j=0}^{\lfloor k/2 \rfloor}\left(\frac{1}{12}\right)^j \binom{k-j}{j}\binom{n-k-j}{j}.$$6-theorems (Berglund et al., 2024).

4. Additivity, subsystem decomposition, and dynamical topology

In topological band theory, one prominent identity is the dark-band sum rule for a coherently coupled $$\mu_k(n)=\sum_{j=0}^{\lfloor k/2 \rfloor}\left(\frac{1}{12}\right)^j \binom{k-j}{j}\binom{n-k-j}{j}.$$7 system of three Bloch bands. If bands $$\mu_k(n)=\sum_{j=0}^{\lfloor k/2 \rfloor}\left(\frac{1}{12}\right)^j \binom{k-j}{j}\binom{n-k-j}{j}.$$8, $$\mu_k(n)=\sum_{j=0}^{\lfloor k/2 \rfloor}\left(\frac{1}{12}\right)^j \binom{k-j}{j}\binom{n-k-j}{j}.$$9, and $$\langle c_k c_{n-k},[X_{A_n}]\rangle=(n+1)!\,\mu_k(n),$$0 have Chern numbers $$\langle c_k c_{n-k},[X_{A_n}]\rangle=(n+1)!\,\mu_k(n),$$1, $$\langle c_k c_{n-k},[X_{A_n}]\rangle=(n+1)!\,\mu_k(n),$$2, and $$\langle c_k c_{n-k},[X_{A_n}]\rangle=(n+1)!\,\mu_k(n),$$3, then the dark band satisfies

$$\langle c_k c_{n-k},[X_{A_n}]\rangle=(n+1)!\,\mu_k(n),$$4

The derivation uses conservation of the total Chern number under coupling and the fact that both bright bands are homotopic to the auxiliary band, hence each has Chern number $$\langle c_k c_{n-k},[X_{A_n}]\rangle=(n+1)!\,\mu_k(n),$$5 (Łącki et al., 2020). In this context the identity is additive, but it is not a generic statement about arbitrary hybridization; it relies on the dark-state construction and a maintained gap.

An entanglement-theoretic decomposition appears in the Kane–Mele model with ferromagnetism. If one traces out one spin sector and computes the entanglement Chern numbers of the two spin partitions, the numerical results confirm the sum rule

$$\langle c_k c_{n-k},[X_{A_n}]\rangle=(n+1)!\,\mu_k(n),$$6

The paper emphasizes that this identity is empirically observed and holds when the entanglement spectrum is gapped; when that spectrum is gapless, the entanglement Chern number is undefined (Araki et al., 2016).

Higher-Chern-number flat-band engineering provides a further additive pattern. In a bilayer checkerboard lattice, suitable interlayer coupling transforms two $$\langle c_k c_{n-k},[X_{A_n}]\rangle=(n+1)!\,\mu_k(n),$$7 bands into a single flat band with $$\langle c_k c_{n-k},[X_{A_n}]\rangle=(n+1)!\,\mu_k(n),$$8, summarized in the identity

$$\langle c_k c_{n-k},[X_{A_n}]\rangle=(n+1)!\,\mu_k(n),$$9

The same work presents the many-body Chern number formula under twisted boundary conditions and reports fractional Chern insulator states with $\langle c_n,[X_{A_n}]\rangle=(n+1)!$0 and $\langle c_n,[X_{A_n}]\rangle=(n+1)!$1 in the engineered $\langle c_n,[X_{A_n}]\rangle=(n+1)!$2 band (Ding et al., 18 Dec 2025). This is best read as a band-merging identity within a specific construction.

Quench dynamics yields a topological identity of a different kind. In a periodically driven optical lattice, the Chern number of the post-quench Hamiltonian is identified with the linking number of momentum-space vortex trajectories:

$\langle c_n,[X_{A_n}]\rangle=(n+1)!$3

At the same time, the instantaneous Chern number of the evolving state remains zero under unitary dynamics. The identity therefore relates a static invariant of the Hamiltonian to a dynamical invariant of phase singularities, rather than to the topology of the instantaneous state (Tarnowski et al., 2017).

5. Response, geometry, and non-Hermitian reformulations

For non-Hermitian Chern insulators, a magnetic-field diagnostic ties the Chern number to the splitting of parent bands. When the flux per unit cell is $\langle c_n,[X_{A_n}]\rangle=(n+1)!$4, semiclassical quantization gives

$\langle c_n,[X_{A_n}]\rangle=(n+1)!$5

where $\langle c_n,[X_{A_n}]\rangle=(n+1)!$6 is the number of sub-bands into which a parent band splits. The relevant topological invariant is the non-Bloch Chern number,

$\langle c_n,[X_{A_n}]\rangle=(n+1)!$7

and the paper argues that field reversal diagnoses the sign of $\langle c_n,[X_{A_n}]\rangle=(n+1)!$8 (Liu et al., 2024).

Optical response generates a spectral sum rule. For circularly polarized absorption in a two-dimensional insulator, the frequency-resolved Chern spectral function satisfies

$\langle c_n,[X_{A_n}]\rangle=(n+1)!$9

and the integrated identity is

$c_kc_{n-k}$0

The same formalism yields local and nonlocal Chern markers, including

$c_kc_{n-k}$1

and a nonlocal Chern marker whose decay length diverges at topological phase transitions (Molignini et al., 2022).

Geometric identities relate Chern number to quantum metric in even-dimensional Dirac Chern insulators. The quantum metric is

$c_kc_{n-k}$2

so it is the pullback of the metric on the unit $c_kc_{n-k}$3-sphere. In two dimensions,

$c_kc_{n-k}$4

and more generally

$c_kc_{n-k}$5

This expresses the Chern number through the quantum metric and the surface area of the Brillouin zone mapped to the hypersphere (Zhang, 2021).

Odd-dimensional lattice Hamiltonians admit a response-theoretic identity connecting the Chern character of Berry curvature to the Chern–Simons level of the low-energy effective action:

$c_kc_{n-k}$6

The derivation passes through a Green-function winding number and uses a series of Ward–Takahashi identities (Fukaya et al., 2020).

6. Computational equivalences and experimental diagnostics

A major computational theme is the equivalence of twisted-boundary, non-commutative, and Bott-index formulations. For a gapped occupied projector $c_kc_{n-k}$7, the twisted-boundary-condition formula is

$c_kc_{n-k}$8

Perturbatively expanding in $c_kc_{n-k}$9 and $C_D=C_1+C_2-C_3,$00 yields the non-commutative real-space formula

$C_D=C_1+C_2-C_3,$01

while the Bott index is

$C_D=C_1+C_2-C_3,$02

These formulations are derived from one another and are numerically confirmed for the Chern insulator and the quantum spin Hall insulator (Lin et al., 2023).

In photonic crystals, first-principle computation starts from Maxwell’s equations, reformulates the problem as a Hermitian generalized eigenvalue problem, and evaluates Berry curvature on a discretized Brillouin zone. The gauge-invariant lattice formula is

$C_D=C_1+C_2-C_3,$03

and for degenerate groups of bands the composite Chern number is additive,

$C_D=C_1+C_2-C_3,$04

For matched gyroelectric and gyromagnetic parameters, the TE and TM Chern numbers are equal in magnitude and opposite in sign (Zhao et al., 2020).

Recent moiré experiments use Hall quantization and the Středa formula,

$C_D=C_1+C_2-C_3,$05

to identify integer and fractional high-Chern-number states. In twisted rhombohedral trilayer-bilayer graphene, the observed values include $C_D=C_1+C_2-C_3,$06, $C_D=C_1+C_2-C_3,$07, and $C_D=C_1+C_2-C_3,$08, and the reported normalized quantities

$C_D=C_1+C_2-C_3,$09

are presented as scaling patterns relative to a parent flat band with $C_D=C_1+C_2-C_3,$10 (Dong et al., 14 Jul 2025). This suggests a useful distinction between exact topological equalities and experimentally inferred hierarchy relations.

Taken together, these developments show that “Chern number identity” is not a single theorem but a family of structurally different statements: cohomological product formulas, index-determination identities, combinatorial residue expressions, additivity and decomposition rules, and response or measurement formulas. The unifying feature is that each such identity makes the same topological quantity calculable in a different language.