Double Affine Grassmannian Slices
- Double affine Grassmannian slices are finite-dimensional transversal structures that generalize the classical affine Grassmannian associated with reductive groups.
- They are realized via moduli spaces of G-instantons on surface singularities and as Coulomb branches of 3d N=4 affine quiver gauge theories, linking geometry and representation theory.
- These slices encapsulate rich combinatorial and geometric data with practical implications for gauge theory, RG flow analysis, and the construction of Hecke operators.
The double affine Grassmannian and its slices form a conjectural algebro-geometric structure that generalizes the classical affine Grassmannian associated to a reductive group , with far-reaching implications in representation theory, gauge theory, and the geometric formulation of RG flows in supersymmetric field theories. While the "double affine Grassmannian" itself resists direct construction due to infinite dimensionality, its finite-dimensional transversal slices are concretely realized by means of instanton moduli spaces on surface singularities and Coulomb branches of 3d supersymmetric affine quiver gauge theories. These slices encapsulate deep combinatorial, geometric, and representation-theoretic data, linking the geometry of singularities, weight spaces of integrable modules, and the partial order structure of RG flows in physical theories.
1. Affine and Double Affine Grassmannians: Foundational Constructions
The ordinary affine Grassmannian for a complex reductive group is the ind-projective ind-scheme whose Schubert strata and transversal slices are labeled by dominant coweights . Classical transversal slices arise as intersections of Schubert cells inside the Kashiwara flag-type ind-scheme and can be modeled by moduli data for -bundles on with prescribed behavior at points $0$ and , carrying natural Poisson (and typically symplectic) structures, with dimensions given by .
By contrast, the double affine Grassmannian is heuristically defined as for the untwisted affine Kac-Moody group associated to . The object itself is too infinite to be constructed directly, so attention focuses on its finite-dimensional transversal slices , labeled by affine coweight pairs in the affine lattice . These slices are realized as Uhlenbeck partial compactifications of -equivariant -instantons on for level , or equivalently as Coulomb branches of affine quiver gauge theories (Finkelberg, 2017, Fazzi et al., 2023, 0711.2083).
2. Transversal Slices: Moduli of Instantons and ADHM Data
The key realization of double affine Grassmannian slices is via moduli spaces of framed -bundles (or instantons) on quotient surfaces for the cyclic group . These moduli, denoted , are stratified by monodromy data and instanton number , and their closures serve as the analogs of Schubert varieties.
For , -equivariant vector bundles correspond to Nakajima-type cyclic quiver varieties, with ADHM data in graded vector spaces subject to moment map and stability conditions. The geometric invariant theory quotient yields smooth quiver varieties , and the affine quotient realizes the Uhlenbeck closure (0711.2083).
Slices are constructed as intersections , inheriting contracting -actions and forming a Hasse diagram with embeddings corresponding to defect addition and partial order .
3. Coulomb Branch Realization and the BFN Construction
Transversal slices of the double affine Grassmannian are conjecturally isomorphic to Coulomb branches of 3d affine quiver gauge theories. For any symmetric Kac-Moody graph (with no loops), fix dimension vectors and framing vectors indexed by the quiver nodes , constructing the gauge group and matter representation combining bifundamental hypermultiplets and framing multiplets.
The Coulomb branch algebra is defined by equivariant Borel-Moore homology,
where parameterizes bundles with trivialization and sections of the associated bundle (the "raviolo"), and the Coulomb branch is
Conjecturally, for each pair of affine coweights ,
intertwining the residual torus action and identifying hyperbolic stalks with weight spaces of the dual affine Kac-Moody group .
The construction admits extensions: the Coulomb branch for a general complex reductive group and representation is an irreducible normal affine variety of dimension , with a flat integrable system map to Cartan quotient , admitting partial flavor deformations and hyperkähler quotients. The coordinate ring of each slice is generated by monopole operators, graded by dominant cocharacters, with Hilbert series given by the monopole formula (Finkelberg, 2017).
4. Stratification, Partial Order, and Hasse Diagrams
The slices are naturally stratified, indexed by affine coweight pairs , or physical data labeling A-type orbi-instantons in 6d SCFTs (e.g., orbi-instantons). The partial order on coweight labels is determined by the combination (with all ). In the double affine case, the covering relations are classified by Roy’s theorem as either minimal pseudo-Levi steps ( on vanishing subdiagrams) or simple-coroot steps ( with adjacent nodes).
The resulting Hasse diagram encodes all allowed RG flows and degenerations. Minimal edges are labeled by singularity types: minimal ADE, , , or . For symmetric-product degenerations , slices are symmetric products , with further minimal slices stratified by partitions. By composition, one recovers the full inductive RG-flow network (Fazzi et al., 2023).
5. Geometric Properties: Dimensions, Singularities, and Intersection Cohomology
The geometric structure of slices is tightly controlled by root-theoretic data:
- The half-complex dimension is for Weyl vector .
- For minimal pseudo-Levi steps, is isomorphic to the closure of the minimal nilpotent orbit, with Kleinian singularity of corresponding ADE type.
- For simple-coroot steps, .
- For symmetric-product degenerations, , further stratified.
Intersection cohomology stalks at contracting -fixed points recover -analogs of weight multiplicities in integrable modules of , with Poincaré polynomials matching Lusztig-Kostant/Affine combinatorics. For cyclic quiver varieties, Nakajima's formula computes the multiplicity in tensor powers of level-1 fundamentals, with level-rank duality ensuring agreement in ADE settings (0711.2083).
6. Explicit Examples and Computations
Several concrete cases illustrate the general theory:
- For , , the slice coincides with .
- For , , the affine quiver and dimension vectors yield the Uhlenbeck space , with .
- For , minimal slices with data and , yield the hypersurface singularity (type ).
- At level in the case, the dominant coweights and Hasse diagram encode the full RG-flow hierarchy via allowed degenerations and singularity types, with computer verification of a-central charge monotonicity (Fazzi et al., 2023).
In each case, cohomological degrees, fixed points, Hilbert series, and weight multiplicities concur with representation-theoretic expectations and combinatorial formulas.
7. Open Problems and Generalizations
The theory currently provides only the transversal slices and not a global ind-scheme model for the double affine Grassmannian itself. Open problems include:
- Algebraic construction of carrying convolution products, fusion, and a monoidal category of perverse sheaves geometrizing .
- Convolution structures for slices, expected as analogues to classical convolution by merging instanton moduli.
- Extension to twisted or non-simply-laced Kac-Moody groups via alternative surface singularities.
- Global curve models and Hecke operators, connecting instanton moduli to -deformations and higher-dimensional field theories.
- Categorification, -theoretic and elliptic refinements, with connections to double affine Hecke algebras and elliptic quantum groups.
The ultimate research goal is a double affine geometric Satake correspondence, identifying monoidal categories of sheaves on the conjectural double affine Grassmannian with representation categories of . Current constructions via instantons and Coulomb branches establish the atomic building blocks and suggest an inductive geometric and representation-theoretic framework (0711.2083, Finkelberg, 2017, Fazzi et al., 2023).