K-Theoretical Hall Algebra

• K-theoretical Hall algebra is a convolution algebra defined via pull–push operations in equivariant K-theory on moduli spaces of quiver representations and sheaves.
• It extends classical Hall algebras by replacing cohomology with K-theory, accommodating singularity categories and matrix factorizations.
• Its framework connects geometric representation theory to quantum groups, Hopf structures, and categorification through PBW theorems and shuffle models.

K-theoretical Hall algebra most commonly denotes a Hall-type convolution algebra defined in equivariant $K$-theory, or in the Grothendieck group of categories of singularities or matrix factorizations, for moduli of quiver representations, sheaves, or related correspondence spaces. In geometric representation theory, this means an algebra whose multiplication is induced by pull–push along Hall correspondences, in direct parallel with cohomological Hall algebras but with $K$-theory replacing cohomology. The expression also has broader and older uses: it can refer to Hall algebra structures organized by the Grothendieck group $K_0$ of an abelian category, or to settings in which a Hall algebra and a Waldhausen-style algebraic $K$-theory are developed in parallel from the same exact-like category rather than one being defined on the other (Schiffmann et al., 2010, Pădurariu, 2019, Walker, 2010, Eppolito et al., 2018).

1. Terminology and basic paradigm

In the modern geometric sense, a $K$-theoretic Hall algebra is a $\mathbb Z_{\ge 0}^I$-graded convolution algebra built from stacks of representations or sheaves, with multiplication defined by pull–push along the standard Hall correspondence. For a symmetric quiver $Q$ with vertex set $I$, dimension vector $\gamma$, representation space

$$M_\gamma:=\prod_{i,j\in I}\mathbb A^{\alpha_{ij}\gamma_i\gamma_j},$$

and group

$K$0

one basic form is

$K$1

with Hall multiplication

$K$2

along the correspondence

$K$3

When a potential $K$4 is present, the $K$5-theory side is replaced by $K$6 of singularity categories or matrix factorizations on the derived zero locus of $K$7, giving

$K$8

This is the standard contemporary meaning of the subject (Küng et al., 11 Jul 2025, Pădurariu, 2021).

The same pull–push pattern appears in other geometric settings. For a smooth quasi-projective surface $K$9, the $K_0$0-theoretic Hall algebra studied by Kapranov–Vasserot, Zhao, and later work has underlying graded vector space

$K_0$1

with multiplication defined by flag Quot correspondences, induction from a parabolic, and a refined Gysin pullback. In this sense, the subject is not tied to quivers alone: it is a general Hall-type formalism in equivariant $K_0$2-theory of geometric correspondences (Peng et al., 2020).

A persistent terminological issue is that not every appearance of $K_0$3 in Hall algebra theory means geometric $K_0$4-theory. Several papers explicitly distinguish the modern meaning from Grothendieck-group grading or from side-by-side Hall-theoretic and $K_0$5-theoretic constructions. This suggests that “K-theoretical Hall algebra” is best understood as a family of related but non-equivalent notions, with geometric equivariant $K_0$6-theory as the dominant current usage (Walker, 2010, Eppolito et al., 2018).

2. Geometric realizations on curves, surfaces, and commuting varieties

One of the earliest genuinely geometric $K_0$7-theoretic realizations appears for curves. For a smooth connected projective curve $K_0$8 of genus $K_0$9, the spherical Hall algebra $K$0 is related to a convolution algebra in equivariant $K$1-theory of the genus-$K$2 commuting varieties

$K$3

The paper constructs a graded $K$4-algebra

$K$5

whose spherical subalgebra is generated by degree one, and proves that, modulo torsion, it is anti-isomorphic to the universal twisted spherical Hall algebra of the curve. The comparison is mediated by identical shuffle kernels after parameter identification, and is presented as a form of geometric Langlands duality in the formal neighborhood of the trivial local system (Schiffmann et al., 2010).

For surfaces, the geometry is based on Quot and flag Quot schemes of $K$6-dimensional sheaves. The degree-one classes

$K$7

play the role of current generators. A central structural result is the cubic Serre relation

$K$8

proved for every smooth quasi-projective surface. The argument uses refined Gysin maps, Cohen–Macaulay and normality properties of small flag schemes, and exact sequences on auxiliary moduli spaces $K$9. The result gives direct evidence that the $K$0-theoretic Hall algebra of surfaces obeys quantum-group-like relations even when an explicit global shuffle presentation is unavailable (Peng et al., 2020).

A broader oriented-cohomology framework places these examples in a common formalism. The preprojective CoHA attached to a quiver $K$1 and an algebraic oriented cohomology theory $K$2 specializes, for $K$3-theory, to a $K$4-theoretic Hall algebra of moment-map zero fibers. In the Jordan-quiver case, the moment-map equation

$K$5

identifies the zero fiber with the commuting variety

$K$6

so the construction generalizes the Schiffmann–Vasserot $K$7-theoretic Hall algebra of commuting varieties. In this specialization the shuffle multiplication is governed by the multiplicative formal group law

$K$8

and the Hall product becomes a multiplicative shuffle formula in the equivariant line-bundle variables (Yang et al., 2014).

3. Quivers with potential, singularity categories, and categorical KHA

For quivers with potential, the decisive categorical input is the replacement of vanishing-cycle cohomology by categories of singularities or matrix factorizations. Given $K$9, Pădurariu defines the categorical Hall algebra

$\mathbb Z_{\ge 0}^I$0

with multiplication

$\mathbb Z_{\ge 0}^I$1

where $\mathbb Z_{\ge 0}^I$2 is Thom–Sebastiani, $\mathbb Z_{\ge 0}^I$3 is induced by diagonal equivariance, and $\mathbb Z_{\ge 0}^I$4 come from the Hall correspondence of short exact sequences. Its Grothendieck group is the $\mathbb Z_{\ge 0}^I$5-theoretic Hall algebra

$\mathbb Z_{\ge 0}^I$6

equivalently

$\mathbb Z_{\ge 0}^I$7

The monoidality theorem gives an associative categorical Hall algebra and, after decategorification, a genuine KHA for any quiver with potential (Pădurariu, 2021).

The same body of work establishes several structural refinements. For symmetric quivers with potential, one can define a filtration whose associated graded is a deformation of a symmetric algebra: $\mathbb Z_{\ge 0}^I$8 A second filtration yields

$\mathbb Z_{\ge 0}^I$9

and its first step is the $Q$0-theoretic BPS Lie algebra. This is the $Q$1-theoretic analogue of the Davison–Meinhardt PBW structure for CoHA, obtained by semi-orthogonal decompositions and singularity categories rather than by the decomposition theorem (Pădurariu, 2019).

A later refinement proves a PBW-type theorem for a natural quotient of $Q$2. Writing

$Q$3

the theorem states that $Q$4 is generated by primitive-type spaces $Q$5, and for $Q$6 these generators are identified with a version of intersection $Q$7-theory: $Q$8 This places $Q$9-theoretic Hall algebras in direct analogy with BPS/intersection-cohomology descriptions on the cohomological side (Pădurariu, 2021).

4. Bialgebras, Hopf structures, and comparisons with CoHA

A central problem in the subject is that the Hall product alone does not automatically provide a satisfactory coproduct or Hopf structure. For symmetric quivers with potential satisfying a Künneth-type assumption,

$I$0

positive and negative extensions of the KHA can be formed by adjoining Cartan currents $I$1. Their defining relations involve the kernel

$I$2

and the resulting extensions are bialgebras. In the tripled-quiver, preprojective case one further obtains a nondegenerate Hopf pairing between the positive and negative halves and hence a Drinfeld double algebra. This gives a precise quantum-group-type framework for KHAs of quivers with potential (Pădurariu, 2021).

There is also a conceptual comparison map from $I$3-theoretic to cohomological Hall algebra. For a symmetric quiver with $I$4, and more generally for quivers with potential under the assumption that $I$5 is the only critical value of $I$6, the natural bridge is not the bare Chern character but the corrected map

$I$7

or in the potential case

$I$8

This is an algebra morphism from the KHA to a suitably twisted completed CoHA, where the twist is dictated by Grothendieck–Riemann–Roch and the factorization of Todd classes along Hall correspondences. The comparison shows that the relationship between KHA and CoHA is intrinsic and functorial rather than an ad hoc modification of the Chern character (Küng et al., 11 Jul 2025).

These constructions clarify the algebraic status of KHA. In the modern theory, the multiplication is not merely a decategorified count of extensions; it participates in a braided, twisted, or doubled structure whose natural targets include quantum loop algebras, Drinfeld doubles, and completed CoHAs. This suggests that the difference between $I$9-theoretic and cohomological Hall algebra is not only one of coefficients, but one of intrinsic multiplicative and Hopf-theoretic geometry (Pădurariu, 2021, Küng et al., 11 Jul 2025).

5. Quantum groups, toroidal structures, Coulomb branches, and abelian categorification

A major line of development identifies KHAs with positive halves of quantum groups. For a preprojective algebra $\gamma$0 of affine type, the nilpotent $\gamma$1-theoretic Hall algebra $\gamma$2 is proved to be isomorphic to the positive Drinfeld half of the corresponding quantum toroidal quantum group. The proof uses a deformed torus-equivariant KHA, a shuffle realization, and a torsion-freeness theorem ensuring an embedding into the shuffle algebra. The same work develops a super analogue for quivers with potential and categories of singularities, giving a homomorphism from a super toroidal quantum group to the corresponding KHA (Varagnolo et al., 2020).

There is also a Dynkin-type realization without potential. For the type $\gamma$3 Dynkin quiver, the ordinary K-theoretic Hall algebra is isomorphic to the positive part of Arkhipov–Mazin’s $\gamma$4-affine quantum group: $\gamma$5 The proof uses a shuffle presentation, generation by degree-one Hall classes, and a restricted-sector dimension argument. The resulting algebra acts categorically on derived categories of partial flag varieties and produces semiorthogonal decompositions of the corresponding weight categories (Hsu, 9 Dec 2025).

A further advance is abelian categorification. To a quiver $\gamma$6, one can attach a finite-length monoidal abelian category $\gamma$7 whose Grothendieck group is the relevant preprojective K-theoretic Hall algebra of Varagnolo–Vasserot. In this framework the simple objects provide a dual canonical basis of the Hall algebra; for affine quivers this yields a basis of the positive half of the corresponding quantum toroidal algebra. The category also carries renormalized $\gamma$8-matrices in the sense of Kang–Kashiwara–Kim, strengthening the representation-theoretic structure far beyond the decategorified algebra (Cautis, 6 Aug 2025).

More recently, the shuffle model has been used to connect KHA to Coulomb branches. For a quiver $\gamma$9 with equivariant parameters and canonical tripled-quiver potential, one studies the loop-nilpotent KHA

$$M_\gamma:=\prod_{i,j\in I}\mathbb A^{\alpha_{ij}\gamma_i\gamma_j},$$0

where $$M_\gamma:=\prod_{i,j\in I}\mathbb A^{\alpha_{ij}\gamma_i\gamma_j},$$1 is the locus on which all loops $$M_\gamma:=\prod_{i,j\in I}\mathbb A^{\alpha_{ij}\gamma_i\gamma_j},$$2 act nilpotently. The paper constructs a surjective homomorphism from a suitably interpreted shifted Drinfeld double of this loop-nilpotent KHA to the $$M_\gamma:=\prod_{i,j\in I}\mathbb A^{\alpha_{ij}\gamma_i\gamma_j},$$3-theoretic Coulomb branch algebra $$M_\gamma:=\prod_{i,j\in I}\mathbb A^{\alpha_{ij}\gamma_i\gamma_j},$$4. Because the double has no direct geometric definition, the construction is carried out through an integral shuffle double. This gives a concrete Higgs-branch to Coulomb-branch map in $$M_\gamma:=\prod_{i,j\in I}\mathbb A^{\alpha_{ij}\gamma_i\gamma_j},$$5-theory (Jindal et al., 19 May 2026).

6. Alternative meanings, historical variants, and conceptual boundaries

Not every paper with “$$M_\gamma:=\prod_{i,j\in I}\mathbb A^{\alpha_{ij}\gamma_i\gamma_j},$$6” and “Hall algebra” in the same title belongs to the modern geometric KHA theory. A foundational example is Walker’s reinterpretation of the ordinary Ringel–Hall algebra as a Hopf algebra object in the braided category $$M_\gamma:=\prod_{i,j\in I}\mathbb A^{\alpha_{ij}\gamma_i\gamma_j},$$7, where $$M_\gamma:=\prod_{i,j\in I}\mathbb A^{\alpha_{ij}\gamma_i\gamma_j},$$8 is the Grothendieck group of the underlying abelian category. The braiding

$$M_\gamma:=\prod_{i,j\in I}\mathbb A^{\alpha_{ij}\gamma_i\gamma_j},$$9

absorbs the Green-formula twist, so the Hall algebra becomes a Hopf object after changing the ambient monoidal category. This is “$K$00-theoretical” only in the sense of Grothendieck-group grading and Euler-form data on $K$01, not in the sense of Hall algebras built from equivariant $K$02-theory or categories of singularities (Walker, 2010). The same perspective is expanded categorically through groupoidification and spans, again with $K$03 as grading data rather than geometric $K$04-theory (Walker, 2013).

Another adjacent but distinct use occurs for pointed matroids. The category $K$05 of pointed matroids and strong maps is shown to be a finitary proto-exact category, supporting both a Hall algebra and Waldhausen algebraic $K$06-theory. Its Hall algebra counts restriction–contraction decompositions of matroids, while its $K$07-theory is defined by the $K$08-construction and contains injections

$K$09

This is not a $K$10-theoretic Hall algebra in the geometric representation-theoretic sense; rather, Hall algebra and algebraic $K$11-theory coexist in parallel on the same proto-exact category (Eppolito et al., 2018).

These distinctions delimit the subject. In current representation-theoretic usage, a K-theoretical Hall algebra is typically an algebra in equivariant $K$12-theory or in $K$13 of matrix factorization or singularity categories, defined by Hall correspondences on moduli spaces or stacks. Broader uses tied to $K$14-grading, braided monoidal reinterpretation, or parallel Waldhausen $K$15-theory remain mathematically significant, but they mark different points in the genealogy of the term. This suggests that the unqualified phrase “K-theoretical Hall algebra” is intrinsically polysemous, and that precise context—Grothendieck-group grading, equivariant $K$16-theory, singularity categories, preprojective geometry, or Coulomb-branch comparison—is essential (Walker, 2010, Eppolito et al., 2018).