Stark–Heegner Cycles in Number Theory

• Stark–Heegner cycles are constructions that generalize Heegner points using p-adic integration and topological cycles across CM, real quadratic, ATR, and Bianchi settings.
• They employ modular symbols and optimal embeddings to connect Abel–Jacobi images with Selmer groups and derivatives of L-functions.
• The theory underpins advances in confirming Bloch–Kato and Stark-type conjectures through base-change theorems and reciprocity laws.

Stark–Heegner cycles are a family of constructions that transfer the Heegner-point paradigm from CM settings to real quadratic, ATR, and Bianchi settings, typically by replacing globally visible algebraic points or cycles with objects defined first through $p$-adic integration, modular symbols, or topological cycles, and then relating their Abel–Jacobi images to Selmer groups and derivatives of $L$-functions. In the Bianchi setting they are local cohomology classes attached to Bianchi modular forms; in ATR settings they are null-homologous cycles on Hilbert modular varieties paired with Eisenstein forms; in real quadratic and plectic settings they appear as $p$-adic local points, higher-dimensional cohomology classes, or S-arithmetic cycles. A persistent theme is that Bloch–Kato and Stark-type conjectures predict global rationality, reciprocity, and control by central derivatives or higher derivatives of automorphic $L$-functions (Venkat et al., 2019, Venkat, 2021, Charollois et al., 2014, Fornea et al., 2021).

1. Genealogy and conceptual scope

The modern theory begins from the analogy with classical Heegner points, whose arithmetic significance is controlled by Gross–Zagier and Kolyvagin, and from Darmon’s construction of $p$-adic Stark–Heegner points for real quadratic fields. In higher weight and in non-CM settings, the relevant objects cease to be ordinary points on modular curves and instead become local cohomology classes, higher-codimension cycles, or S-arithmetic cycles on Bianchi or Hilbert modular quotients. In the Bianchi context, Stark–Heegner cycles are described as $p$-adic, Hecke-equivariant algebraic cycles attached to Bianchi modular forms, with Abel–Jacobi images in semistable local cohomology; in the ATR setting, they are null-homologous topological cycles on Hilbert modular varieties; in the plectic setting, they are conjecturally related to wedge powers of Mordell–Weil or Selmer classes (Venkat et al., 2019, Charollois et al., 2014, Fornea et al., 2021).

A common source of confusion is that the expression “Stark–Heegner cycle” does not denote a single geometric object across the literature. The Bianchi papers work with a “local Chow” substitute

$$(\Delta_0 \otimes \mathrm{Div}(\mathcal H_{\mathfrak p}^{ur}) \otimes V_{k,k})_\Gamma,$$

whereas the ATR paper works with cycles $A_T$ in $H_{n-1}(X,\mathbf Z)$, and the equidistribution paper studies S-arithmetic cycles

$$\Delta_\psi \subset \mathrm{PGL}_2^+(\mathbf Z[1/p])\backslash (H \times H_p).$$

This suggests that the unifying feature is not a single ambient geometry but a common package: optimal embeddings, ring class fields, Abel–Jacobi regulators, and conjectural reciprocity laws (Venkat et al., 2019, Pérez-Piña, 2024).

The theory is also explicitly designed to match sign conditions in functional equations. In the Bianchi setting, under the Stark–Heegner hypothesis one has $L$0, so the central value vanishes and Bloch–Kato predicts nontrivial Selmer classes. In the ATR setting, the order of vanishing $L$1 is reflected in the exactness of a projected Eisenstein form and in the torsion of the cycle $L$2. In the plectic setting, the anticyclotomic $L$3-adic $L$4-function vanishes to order at least $L$5, and the theory is built to detect $L$6-fold phenomena rather than rank one alone (Venkat et al., 2019, Charollois et al., 2014, Fornea et al., 2021).

2. Bianchi Stark–Heegner cycles and their local Abel–Jacobi images

Let $L$7 be an imaginary quadratic field, $L$8 an unramified rational prime in $L$9, and $p$0 or $p$1 a cuspidal Bianchi eigenform of parallel weight $p$2 and level $p$3, new at the prime above $p$4. The associated $p$5-dimensional $p$6-adic Galois representation $p$7 or $p$8 is the target of the local construction. The basic source object is the $p$9-coinvariant module

$L$0

where $L$1, $L$2 is the $L$3-adic upper half-plane, and $L$4. Optimal embeddings $L$5 of an order $L$6 in a quadratic extension $L$7 into an Eichler order $L$8 produce fixed points $L$9, a polynomial $p$0, and an element $p$1. The Stark–Heegner cycle is then

$p$2

well defined in the $p$3-coinvariants and independent of the chosen cusp $p$4 after passage to the class $p$5 (Venkat et al., 2019, Venkat, 2021).

The analytic input is a system of Bianchi modular symbols and double integrals. There is a Hecke-equivariant injection from cuspidal Bianchi forms to modular symbols, and Williams’ control theorem lifts the classical symbol to an overconvergent symbol. On the larger Ihara group one constructs a canonical harmonic modular symbol

$p$6

with one-dimensional $p$7-isotypic component. Double integrals define pairings $p$8 and $p$9, and these in turn produce a filtered $p$0-module $p$1 or $p$2. Under semistability, and conditionally in general or unconditionally in the base-change case, this module is identified with Fontaine’s semistable module $p$3 (Venkat et al., 2019, Venkat, 2021).

The $p$4-adic Abel–Jacobi map is defined by

$p$5

lifting the double-integral pairing. Composing with the Bloch–Kato exponential gives classes in

$p$6

and the local Stark–Heegner classes are

$p$7

The original Bianchi paper proves that these values are independent of the chosen Abel–Jacobi lift on Stark–Heegner cycles, using the vanishing $p$8 (Venkat et al., 2019).

3. Selmer conjectures, reciprocity, and the base-change theorem

The global conjectural framework is formulated in terms of semistable Bloch–Kato Selmer groups over ring class fields. If $p$9 is the abelian subextension of the ring class field $$(\Delta_0 \otimes \mathrm{Div}(\mathcal H_{\mathfrak p}^{ur}) \otimes V_{k,k})_\Gamma,$$0 cut out by a character $$(\Delta_0 \otimes \mathrm{Div}(\mathcal H_{\mathfrak p}^{ur}) \otimes V_{k,k})_\Gamma,$$1, the expected global class is

$$(\Delta_0 \otimes \mathrm{Div}(\mathcal H_{\mathfrak p}^{ur}) \otimes V_{k,k})_\Gamma,$$2

and the conjectural local-global compatibility is

$$(\Delta_0 \otimes \mathrm{Div}(\mathcal H_{\mathfrak p}^{ur}) \otimes V_{k,k})_\Gamma,$$3

In the earlier formulation this is accompanied by a Shimura reciprocity law under the action of $$(\Delta_0 \otimes \mathrm{Div}(\mathcal H_{\mathfrak p}^{ur}) \otimes V_{k,k})_\Gamma,$$4, and by the prediction that $$(\Delta_0 \otimes \mathrm{Div}(\mathcal H_{\mathfrak p}^{ur}) \otimes V_{k,k})_\Gamma,$$5 should force

$$(\Delta_0 \otimes \mathrm{Div}(\mathcal H_{\mathfrak p}^{ur}) \otimes V_{k,k})_\Gamma,$$6

The sign condition comes from the Stark–Heegner hypothesis: $$(\Delta_0 \otimes \mathrm{Div}(\mathcal H_{\mathfrak p}^{ur}) \otimes V_{k,k})_\Gamma,$$7 is inert in $$(\Delta_0 \otimes \mathrm{Div}(\mathcal H_{\mathfrak p}^{ur}) \otimes V_{k,k})_\Gamma,$$8, every prime dividing the tame level part $$(\Delta_0 \otimes \mathrm{Div}(\mathcal H_{\mathfrak p}^{ur}) \otimes V_{k,k})_\Gamma,$$9 splits in $A_T$0, and hence the functional equation has sign $A_T$1 at the center (Venkat et al., 2019).

The principal unconditional result currently available is the base-change theorem. Assume $A_T$2, let $A_T$3 be the quadratic base-change of a classical newform $A_T$4, write $A_T$5, require the Heegner hypothesis on the factorization of $A_T$6, assume $A_T$7, and impose

$A_T$8

Then there exists

$A_T$9

such that

$H_{n-1}(X,\mathbf Z)$0

This proves, in the base-change scenario, the rationality conjecture for the trivial character and establishes that the local Stark–Heegner class is the restriction of a global Selmer class defined over $H_{n-1}(X,\mathbf Z)$1 (Venkat, 2021).

The proof combines several analytic and arithmetic ingredients. The paper constructs a two-variable base-change Bianchi $H_{n-1}(X,\mathbf Z)$2-adic $H_{n-1}(X,\mathbf Z)$3-function, proves a $H_{n-1}(X,\mathbf Z)$4-adic Artin factorization [

(D_K)^{\lambda_\kappa/2}L_p(\mathbf F/K,\psi_K,\lambda_\kappa)

\eta\, L_p(\mathbf F,\chi_1,\lambda_\kappa)\,L_p(\mathbf F,\chi_2,\lambda_\kappa),