---
title: K-Theoretic Equivariant 3-Vertex
url: https://www.emergentmind.com/topics/k-theoretic-equivariant-3-vertex
type: topic
---

# K-Theoretic Equivariant 3-Vertex

Searching arXiv for the core papers and closely related recent work on the K-theoretic equivariant 3-vertex.
The **K-theoretic equivariant 3-vertex** is a local enumerative building block in equivariant \(K\)-theory for curve-counting theories on toric and related threefolds. In the Donaldson–Thomas/Pandharipande–Thomas setting, it is a canonical **3-valent tensor** attached to \(X=\mathbb C^3\) or to a corresponding relative geometry, with three legs indexed by partitions or, more canonically, by classes in the equivariant \(K\)-theory of \(\operatorname{Hilb}(\mathbb C^2,\mathrm{points})\) [1905.01523]. In the quasimap and quantum \(K\)-theoretic literature, closely related “vertex functions” appear as **bare** and **capped** generating series, linked by a capping operator that solves a quantum difference equation; this formulation is central for instanton moduli spaces and hypertoric varieties [1612.01048]. Across these settings, the 3-vertex formalism organizes localization, degeneration, gluing, and wall-crossing, and it connects enumerative geometry to quantum toroidal algebras, stable envelopes, and \(q\)-difference systems.

## 1. Geometric frameworks and basic objects

In toric \(3\)-fold Donaldson–Thomas theory, the local model is \(X=\mathbb C^3\), and the vertex is the local contribution assigned to the three coordinate directions. Its three legs encode asymptotic partitions of torus-fixed subschemes, and the resulting tensor serves as the local input for global toric partition functions under degeneration and gluing [1905.01523]. The same trivalent structure persists in the PT formulation, where the moduli spaces parameterize stable pairs
\[
\mathcal O_X \xrightarrow{s} F
\]
with \(F\) pure \(1\)-dimensional and \(\mathrm{coker}(s)\) \(0\)-dimensional [1905.01523].

A different but closely connected formulation arises in equivariant quantum \(K\)-theory via quasimaps. For the instanton moduli space
\[
M(n,r),
\]
the basic geometric object is the moduli of framed rank \(r\) torsion-free sheaves on \(\mathbb P^2\) with second Chern class \(n\), with framing along the line at infinity. The special case \(M(n,1)=\operatorname{Hilb}^n(\mathbb C^2)\) places the Hilbert scheme directly inside the vertex formalism [1612.01048]. Here the torus is
\[
T=A\times (\mathbb C^\times)^2,\qquad A\simeq (\mathbb C^\times)^r,
\]
with framing characters \(a_i\), tangent weights \(t_1,t_2\), and symplectic scaling \(\hbar=t_1t_2\) [1612.01048].

Hypertoric varieties provide another explicit model of the same general phenomenon. Their vertex functions are defined by \(K\)-theoretic counting of quasimaps from \(\mathbb P^1\), and the resulting functions satisfy \(q\)-difference equations in both Kähler and equivariant variables. In this setting the vertex formalism is not a trivalent tensor in the toric DT sense, but a system of quasimap-counting generating series with the same structural roles of localization, capping, and parameter exchange under mirror symmetry [2006.00118].

## 2. Bare vertices, capped vertices, and trivalent tensors

For quasimaps with a descendent insertion \(\tau\), the **bare vertex** and **capped vertex** are generating series of equivariant pushforwards from quasimap moduli spaces. Their defining relation is
\[
\widehat V^{(\tau)}(z)=\Psi(z)\,V^{(\tau)}(z),
\]
where \(\Psi(z)\) is the **capping operator** [1612.01048]. Geometrically, \(V^{(\tau)}\) records the contribution of the open quasimap moduli, \(\Psi(z)\) is the rubber or relative contribution, and \(\widehat V^{(\tau)}\) is the fully capped invariant [1612.01048].

The capping operator is the fundamental solution of the \(K\)-theoretic quantum difference equation
\[
\Psi^{(r)}(zq)\,\mathcal O(1)=\mathcal M^{(r)}(z)\,\Psi^{(r)}(z),\qquad \Psi^{(r)}(0)=1,
\]
with \(\mathcal O(1)\) acting by multiplication by the tautological line bundle and \(\mathcal M^{(r)}(z)\) a difference operator with rational matrix coefficients [1612.01048]. This is the \(K\)-theoretic analogue of the quantum differential equation in quantum cohomology.

In toric DT/PT theory, the local vertex is defined directly as a trivalent generating function. For prescribed boundary conditions \((\lambda,\mu,\nu)\) and \(M\in\{\mathrm{DT},\mathrm{PT}\}\), the \(K\)-theoretic vertex is
\[
V^{M,K}_{\lambda,\mu,\nu}
=
\sum_{n\in\mathbb Z}
Q^n\,
\chi\!\left(M_{(\lambda,\mu,\nu),n},\,\mathcal O^{vir}\right),
\]
where the moduli spaces carry symmetric perfect obstruction theories \(\operatorname{Ext}_{\mathbf X}^*(I,I(-D))\) [2311.15697]. The theory uses the Nekrasov–Okounkov twist
\[
\#1\mathcal O^{vir}:=\mathcal O^{vir}\otimes K_{vir}^{1/2},
\]
and the twist is emphasized as essential for rigidity and localization in equivariant \(K\)-theory [2311.15697].

These two viewpoints are structurally parallel. The quasimap formalism packages relative geometry into a capping operator, while the toric DT/PT formalism packages local relative conditions into a trivalent tensor. In both cases, the vertex is the local object from which global invariants are assembled.

## 3. Leg structure and explicit formulas

The **leg decomposition** is one of the most concrete aspects of the theory. A **2-leg vertex** means that exactly two of the three asymptotic partitions are nonempty and one is empty. In the PT-based \(K\)-theoretic DT theory of toric \(3\)-folds, this case admits an explicit formula in a gauge adapted to
\[
X=S\times \mathbb C,\qquad S=\mathrm{Blow\text{-}up}_{(0,0)}(\mathbb P^1\times \mathbb C),
\]
with relative divisors \(D_1,D_2\) and torus
\[
T=\mathbb C_x^\times\times \mathbb C_y^\times\times \mathbb C_z^\times
\]
[1905.01523]. The result is expressed plethystically as
\[
\pi_*\mathcal O^{\mathrm{vir}}
=
\mathcal S^\bullet\!\left(
\frac{q}{\{y\}(1-qxyz)}p_1
+
\frac{q}{\{y\}(1-qxyz)}\bar p_1
-
\frac{q}{\{y\}\{xz\}\frac{1-q\kappa}{1-q/\kappa}\,p_1\bar p_1
\right),
\]
together with an equivalent exponential expansion over all \(n\ge 1\) [1905.01523]. The \(1\)-leg specialization is obtained by taking the constant term in the \(\bar p_k\)-variables, and its proof uses stable envelopes [1905.01523].

A more recent contour-integral formulation computes both the equivariant DT and PT \(3\)-vertices from the **same JK-residue integrand**. The Witten index of a \(1\)d \(\mathcal N=2\) SQM is written as
\[
\mathcal Z=\frac{1}{|W_G|}\oint_{\mathrm{JK}}
\prod_{i=1}^k \frac{d\phi_i}{2\pi i}\,\mathcal Z(\phi),
\]
and the distinction between DT and PT is made solely by the **reference vector** of the JK prescription: \(\eta_0=(1,\dots,1)\) extracts the DT poles, while \(\tilde\eta_0=(-1,\dots,-1)\) extracts the PT poles [2508.12125]. In this framework, one-leg and two-leg PT cases involve only simple poles, whereas the genuine three-leg PT vertex exhibits second-order poles [2508.12125].

The same contour formalism analyzes three asymptotic limits of the PT vertex. The limit \(q_4\to 1\) recovers the normalized unrefined topological vertex; an anisotropic scaling limit yields the refined topological vertex; and a distinct scaling of \((q_1,q_2)\) produces the Macdonald refined topological vertex [2508.12125]. This places the \(K\)-theoretic equivariant PT vertex as a common refinement of several established vertex formalisms.

## 4. Factorization, \(q\)-difference equations, and rationality

A central structural result in the capped-vertex literature is factorization under degeneration of framing weights. If the framing torus splits \(r=r_1+r_2\) and the second block of weights is scaled by a parameter \(a\to 0\), then the capping operator satisfies
\[
\lim_{a\to0}\Psi^{(r)}(z)
=
Y^{(r_1),(r_2)}(z)\,
\Psi^{(r_1)}\!\left(z\hbar^{r_2/2}\right)
\otimes
\Psi^{(r_2)}\!\left(z\hbar^{-r_1/2}\right),
\]
where \(Y^{(r_1),(r_2)}(z)\) is a universal lower-triangular operator coming from the quantum toroidal algebra representation [1612.01048]. The bare vertex factorizes compatibly in the same limit. Combined with a large-rank classicality statement, this yields the rationality theorem that the \(K\)-theoretic \(1\)-leg capped descendent vertex is the Taylor expansion of a rational function in
\[
\mathbb Q(u_1,\dots,u_r,t_1,t_2,q,z)
\]
[1612.01048].

For hypertoric varieties, the modified bare vertex satisfies two commuting families of \(q\)-difference equations. Circuits of the hyperplane arrangement control the Kähler-shift equations, and cocircuits control the equivariant-shift equations [2006.00118]. Under \(3\)d mirror symmetry, these two systems are exchanged after swapping Kähler and equivariant variables and reversing the polarization; the mirror transformation matrix is built from elliptic stable envelopes [2006.00118]. This makes the vertex function simultaneously a solution to dual \(q\)-difference systems and a bridge between mirror pairs.

The same rationality pattern reappears in the study of the index vertex for \(T^*\mathrm{Flag}_n\). There the index vertex is identified, for big enough generic slopes, with the equivariant Euler characteristic of a twisted \(K\)-theoretic stable envelope after applying the \(3\)d mirror map, and it follows that the index vertex is the power series expansion of a rational function [2108.07202]. A plausible implication is that rationality is not an isolated feature of special toric calculations, but part of a wider stable-envelope mechanism in equivariant \(K\)-theoretic vertex theory.

## 5. Gluing, DT/PT correspondence, and dimensional reduction

In toric \(3\)-fold geometry, global partition functions factorize into edge and vertex contributions. For a toric \(3\)-fold \(X\),
\[
Z_\beta^{DT,K}(X)
=
\sum_{\lambda}
\prod_e E^K_{\lambda(e)}(t_e)
\prod_v
V^{DT,K}_{\lambda(e_1),\lambda(e_2),\lambda(e_3)}(t_v),
\]
and likewise for PT, with the same edge terms \(E^K\) in both theories [2311.15697]. This concentrates the distinction between DT and PT entirely in the vertices.

The \(3\)-fold \(K\)-theoretic DT/PT vertex correspondence states
\[
V^{DT,K}_{\lambda,\mu,\nu}
=
V^{PT,K}_{\lambda,\mu,\nu}\,
V^{DT,K}_{\emptyset,\emptyset,\emptyset},
\]
and it implies the global toric identity
\[
Z^{DT,K}_\beta(X)=Z^{PT,K}_\beta(X)\,Z^{DT,K}_0(X)
\]
for smooth toric \(3\)-folds [2311.15697]. The proof uses wall-crossing in a family of weak stability conditions and is given in two versions, one following Mochizuki and one following Joyce. A major technical ingredient is the construction of symmetric almost-perfect obstruction theories on auxiliary master spaces via symmetrized pullback [2311.15697].

The \(2\)-leg vertex admits a concrete gluing application to the resolved conifold: gluing two \(2\)-leg vertices produces a reduced operator whose vacuum factor is
\[
\mathsf Z_{\mathrm{resolved\ conifold}}
=
\left(
-\frac{qQ}{(1-q/\kappa)(1-q\kappa)}
\right),
\]
and any matrix element divided by the vacuum matrix element is polynomial in the Kähler parameter \(Q\) [1905.01523].

A \(4\)-fold extension clarifies how the familiar \(3\)-vertex sits inside a larger formalism. For toric Calabi–Yau \(4\)-folds, the local factors \([-v_\alpha]\) and \([-e_{\alpha\beta}]\) specialize to the standard toric \(3\)-fold vertex and edge after setting
\[
y=t_4
\]
and restricting to fixed loci scheme-theoretically supported on the hyperplane \(\{x_4=0\}\) [1906.07856]. More precisely,
\[
V^{\mathrm{DT}}_{\lambda\mu\nu\varnothing}(t,y,q)\big|_{y=t_4}
=
V^{\mathrm{3D,DT}}_{\lambda\mu\nu}(t,-q),
\]
with the analogous PT statement [1906.07856]. This gives an explicit dimensional reduction from the \(4\)-fold vertex formalism to the \(K\)-theoretic equivariant \(3\)-vertex.

## 6. Algebraic structures underlying the vertex

The \(K\)-theoretic equivariant 3-vertex is closely controlled by representation theory. For instanton moduli,
\[
K_T(M(r))\cong \mathcal F^{(r)}=\mathcal F(a_1)\otimes\cdots\otimes\mathcal F(a_r),
\]
where each factor is a Fock space and the framing parameters \(a_i\) are the evaluation parameters of the corresponding Fock representations [1612.01048]. In this formulation, the capping operator is tied to the fundamental solution of the quantum difference equation, and its factorization is governed by the universal \(R\)-matrix, Heisenberg subalgebras, stable envelopes, and the \(w\mathrm{KZ}\) equation [1612.01048].

In the \(2\)-leg PT theory, the Hilbert-scheme/symmetric-function correspondence and stable envelopes provide the computational mechanism. The paper uses
\[
K_{S(n)\times T}(\mathbb C^{2n})_{\mathrm{cpt}}
\cong
\Lambda_n\otimes K_T(\mathrm{pt}),
\]
identifies irreducibles with Schur functions, and relates stable envelopes to symmetric functions through the BKR/Haiman correspondence and Cherednik-algebra/Verma-module technology [1905.01523]. This is what makes the explicit plethystic formula and the skew-Schur-function description of the \(2\)-leg vertex possible.

The contour-integral approach extends this algebraic picture by constructing an operator version of the PT vertex, termed the **Pandharipande–Thomas \(qq\)-character**. In this language, the residues are rewritten as free-field correlators involving bosonic operators \(\mathsf A(x)\), \(\mathsf S_a(x)\), and \(\mathsf W_{\bar a}(x)\), and the resulting operator series is linked to the shifted quantum toroidal \(\mathfrak{gl}_1\) / quiver \(W\)-algebra framework [2508.12125]. In one- and two-leg cases the PT \(qq\)-character commutes with modified screening charges, while in the three-leg PT case second-order poles force derivative terms such as
\[
:\mathsf H(v)\,v\partial_v\mathsf A^{-1}(v):
\]
to appear [2508.12125]. This clarifies that the three-leg vertex is not merely a combinatorial generating function but also an operator-valued object with nontrivial algebraic singularity structure.

## 7. Terminological boundaries and related but distinct usages

The phrase **“3-vertex”** can be misleading because several neighboring literatures use superficially similar language for different objects. In toric DT/PT theory it refers to a trivalent local tensor with three asymptotic legs [1905.01523]. In quasimap theory it refers more loosely to bare and capped vertex functions, often with one-leg emphasis, whose structure is controlled by a capping operator and \(q\)-difference equations [1612.01048]. These are part of the same broad enumerative framework, but they are not identical constructions.

By contrast, **torus-equivariant \(3\)-point genus-\(0\) \(K\)-theoretic Gromov–Witten invariants** of flag manifolds are not the \(K\)-theoretic equivariant 3-vertex. They concern invariants
\[
\langle \gamma_1,\gamma_2,\gamma_3\rangle_d
\]
in quantum \(K\)-theory of \(G/B\) or \(G/P\), together with a divisor-axiom replacement based on the equivariant quantum Chevalley formula and the quantum Bruhat graph [2505.16150]. The common numeral “3” here refers to the number of marked points, not to a trivalent vertex tensor.

Likewise, **three-parameter** equivariant \(K\)-theoretic constructions in exotic geometry should not be confused with the 3-vertex formalism. In the construction of affine quantum Schur algebras of type \(C\) and affine iquantum groups of type AIII, the three parameters arise from a \((\mathbb C^\times)^3\)-equivariant action on Kato’s exotic representation
\[
\mathbb V=V^{\oplus 2}\oplus \wedge^2V,
\]
and the paper explicitly notes that this is not a standard “three-vertex” in topological vertex theory [2606.15722]. The shared prefix “three-” there refers to torus factors and parameters, not to a trivalent enumerative tensor.

Taken together, these distinctions show that the modern literature uses “vertex” in a family of related but non-identical senses. The most stable core meaning of the **K-theoretic equivariant 3-vertex** remains the local equivariant \(K\)-theoretic building block for toric \(3\)-fold curve counting, together with its capped, \(q\)-difference, mirror-symmetric, and operator-theoretic avatars.

Source: https://www.emergentmind.com/topics/k-theoretic-equivariant-3-vertex