Donaldson–Thomas Invariants

Updated 29 June 2026

Donaldson–Thomas invariants are integer or refined counts of stable objects in 3-dimensional Calabi–Yau categories, capturing core enumerative structure.
They leverage a symmetric perfect obstruction theory and the Behrend function to compute weighted Euler characteristics and track wall-crossing behavior.
The Kontsevich–Soibelman formula and Hall algebra frameworks facilitate precise computation and connect DT invariants with modular, physical, and combinatorial applications.

Donaldson–Thomas (DT) invariants are integer (or refined motivic) invariants of moduli spaces of objects in 3-dimensional Calabi–Yau (3CY) categories, encoding their enumerative structure and wall-crossing behavior. They play a central role in enumerative algebraic geometry, representation theory, and mathematical physics, where they provide rigorous counts of BPS states, connect Gromov–Witten theory and sheaf counting, and underlie the wall-crossing phenomena in stability spaces.

1. Foundational Definition and Theoretical Framework

Let $\mathcal{C}$ be a 3CY triangulated category (e.g., $\mathcal{D}^b(\mathrm{Coh}(X))$ for a Calabi–Yau 3-fold $X$ ), and fix a Bridgeland stability condition $\sigma$ on $\mathcal{C}$ . For each class $\gamma\in K_0(\mathcal{C})$ , the DT invariant $\mathrm{DT}(\gamma)$ counts (with weighted signs and multiplicities) $\sigma$ -stable objects of class $\gamma$ .

The DT theory is fundamentally based on the existence of a symmetric perfect obstruction theory on the moduli space $\mathcal{M}(\gamma)$ , yielding a zero-dimensional virtual fundamental class. The integer DT invariant is

$\mathcal{D}^b(\mathrm{Coh}(X))$ 0

where $\mathcal{D}^b(\mathrm{Coh}(X))$ 1 is the Behrend function, encoding the local contributions via Euler characteristics of vanishing cycle complexes (Zhu, 2014, Li et al., 2010).

The generating function of DT invariants is often packaged as a formal product in non-commuting variables, reflecting their wall-crossing behavior: $\mathcal{D}^b(\mathrm{Coh}(X))$ 2 where $\mathcal{D}^b(\mathrm{Coh}(X))$ 3 are the DT invariants (BPS indices), $\mathcal{D}^b(\mathrm{Coh}(X))$ 4 is the quantum dilogarithm, and the ordering is according to the phase of the central charge $\mathcal{D}^b(\mathrm{Coh}(X))$ 5 (Weng, 2016, Zhu, 2014, Goncharov et al., 2016).

2. Wall-Crossing and the Kontsevich–Soibelman (KS) Formula

DT invariants generically jump across walls in the space of stability conditions, i.e., when the central charges of two classes align. The wall-crossing phenomenon is governed universally by the KS formula: $\mathcal{D}^b(\mathrm{Coh}(X))$ 6 This quantized identity encapsulates the transformation of the generating function of DT invariants under wall-crossing, and has become a cornerstone linking enumerative invariants, cluster algebras, and scattering diagrams (Zhu, 2014, Goncharov et al., 2016, Weng, 2016, Cheung et al., 2019).

The motivic (refined) theory further expresses these wall-crossing phenomena via the motivic Hall algebra and transformation identities for the associated "epsilon-motives" and their numerical realization as DT invariants (Meinhardt, 2015, Bu, 26 Mar 2025).

The Hall algebra approach, due to Joyce, Kontsevich–Soibelman, and others, organizes stack functions on moduli spaces into an (associative or Lie) algebra reflecting exact sequences and extensions. DT invariants correspond to certain "logarithmic" generators ("epsilon-motives") in the Hall algebra, and their motivic/integral versions live in the Grothendieck group of varieties with powers of the Lefschetz motive inverted: $\mathcal{D}^b(\mathrm{Coh}(X))$ 7 and satisfy composition and wall-crossing identities in the completed $\mathcal{D}^b(\mathrm{Coh}(X))$ 8-ring (Meinhardt, 2015).

Motivic DT theory allows for refined counts valued in $\mathcal{D}^b(\mathrm{Coh}(X))$ 9 or $X$ 0, and captures not only numerical but also cohomological and motivic information. The integrality conjecture (proved in many cases) asserts that the motivic DT invariants in fact lie in the integral subring without needing denominators.

4. Explicit Geometric and Algebraic Examples

DT invariants were introduced as curve-counting invariants on smooth projective Calabi–Yau threefolds; for instance, the moduli of ideal sheaves parameterize subschemes of dimension $X$ 1. For such $X$ 2,

$X$ 3

where $X$ 4 is the Hilbert scheme of curves of class $X$ 5 and Euler characteristic $X$ 6 (Zhu, 2014, Calabrese, 2011, Jiang, 2015).

For moduli of stable sheaves on surfaces and curves, especially in abelian categories of homological dimension 1, DT invariants agree with the (compactly supported) intersection cohomology of the moduli space (Meinhardt et al., 2014, Meinhardt, 2015). For quivers, the DT invariant of a dimension vector $X$ 7 is identified (under generic stability) with the intersection complex of the stable locus; this leads to effective combinatorial and geometric computation schemes.

DT invariants for more general objects, such as self-dual or orthosymplectic objects, are defined via "orthosymplectic" Hall modules and realize wall-crossing and integration statements analogous to the linear case (Bu, 26 Mar 2025).

Example Table: Key Geometric Contexts for DT Invariants

Setting	Moduli Space	DT Invariant Interpreted As
Smooth CY 3-fold $X$ 8	Hilbert scheme of ideal sheaves	Weighted Euler characteristic
Quiver with potential $X$ 9	Semistable module moduli	Intersection cohomology
K3-fibered Calabi–Yau 3-folds	Sheaves supported on fibers	Modular generating functions
Self-dual quivers / Higgs bundles	Self-dual moduli stack	Orthosymplectic DT invariants

5. Donaldson–Thomas Transformations and Cluster Varieties

In the cluster setting, DT invariants are encoded in a formal automorphism—the Donaldson–Thomas transformation—of the associated (quantum) cluster variety. For 3d Calabi–Yau categories arising from cluster varieties, Kontsevich–Soibelman define a canonical automorphism (the cluster DT transformation) whose wall-crossing factorization coincides with the mutation factorization of the birational automorphism, connecting the enumerative DT theory to cluster algebra dynamics (Weng, 2016, Goncharov et al., 2016).

The geometric realization of the cluster DT transformation, such as the "hyperplane map" on configuration spaces of lines or the induced automorphism on Grassmannians, reflects the deep symmetry between the wall-crossing and the cluster mutation pattern.

6. Connections and Applications: Modularity, Integrality, and Physical Contexts

DT invariants admit connections to modular forms (notably in low dimensions and special geometries). For instance, generating series for invariants on local surfaces or K3-fibered threefolds are modular (or vector-valued modular) forms, realizing predictions from S-duality and enumerative string theory (Gholampour et al., 2013, Gholampour et al., 2013, Toda, 2014, Bryan et al., 2016).

In many settings, formulas for generating functions of DT invariants involve symmetric products, quantum dilogarithms, or theta-type structures. The modularity and integrality phenomena extend, via the motivic formalism, to refined and cohomological invariants, confirming strong predictions from physics regarding BPS state counts, wall-crossing, and dualities (Meinhardt, 2015, Meinhardt et al., 2014, Gholampour et al., 2013).

Moreover, DT invariants are invariant under birational transformations such as flops, modulo exceptional factors, a property established via Hall algebra comparisons and derived equivalences (Calabrese, 2011, Jiang, 2015, Davison et al., 2012).

7. Variants, Generalizations, and Outlook

Recent developments involve DT invariants counting objects with additional symmetries or constraints—orthosymplectic, self-dual, or with defects—and their wall-crossing and motivic refinements (Bu, 26 Mar 2025, Cirafici, 2013). There exist motivic DT invariants with nontrivial monodromy, exhibiting phenomena not reducible to numerical invariants alone (Davison et al., 2012). In addition, DT invariants of parabolic sheaves or those arising in the context of surface defects in gauge theory provide further generalizations.

Research continues in understanding the categorified, cohomological, and refined versions of DT invariants, their role in enumerative and homological mirror symmetry, and their foundational place in the deeper structure of moduli spaces and their wall-crossing behavior. The unification provided by the KS wall-crossing framework and the linkage to cluster and scattering diagrams underlines their centrality in modern geometry and representation theory (Cheung et al., 2019).

Key References:

(Zhu, 2014): Introduction and wall-crossing for Donaldson–Thomas invariants
(Li et al., 2010): Explicit calculations on Calabi–Yau 3-folds
(Meinhardt et al., 2014, Meinhardt, 2015): DT invariants and intersection cohomology of quiver moduli
(Bu, 26 Mar 2025): Orthosymplectic DT invariants
(Weng, 2016, Goncharov et al., 2016): DT transformations in cluster and moduli settings
(Calabrese, 2011, Jiang, 2015): DT invariants and (orbifold) flop invariance
(Gholampour et al., 2013, Toda, 2014): Modularity and generating functions
(Cheung et al., 2019): Refined DT invariants via tropical disk counts.