---
title: Exact Overlap Formulas with Bethe States
url: https://www.emergentmind.com/topics/exact-overlap-formulas-with-bethe-states
type: topic
---

# Exact Overlap Formulas with Bethe States

Exact overlap formulas with Bethe states quantify scalar products between specially constructed integrable matrix product states (MPS) and the Bethe eigenstates of integrable quantum spin chains. In the context of the ABJM alternating $SU(4)$ spin chain, these formulas probe the interplay between integrable boundary or initial states—such as those described by boundary reflection (“$K$-”) matrices—and the bulk integrability encoded via $R$-matrices and the Yang–Baxter equation. The development and explicit evaluation of such formulas reveal structural constraints on nonvanishing overlaps, their explicit determinant structure (typically of Gaudin or Slavnov type), and allow for systematic construction of integrable subspaces within the Hilbert space.

## 1. ABJM Alternating Spin Chain and Bethe Ansatz Construction

The ABJM integrable spin chain encodes the two-loop planar dilatation operator of $\mathcal{N}=6$ Chern–Simons–matter theory as an alternating $SU(4)$ chain, with odd (even) sites carrying the fundamental $\mathbf{4}$ (anti-fundamental $\bar{\mathbf{4}}$) representation. Four elementary $R$-matrices, 
$$
R_{ab}(u) = u\,I_{ab} + P_{ab}, \quad R_{a\bar{b}}(u) = -(u+2)I_{ab} + K_{ab}, \quad etc.
$$ 
act on adjacent quantum spaces and satisfy eight mixed Yang–Baxter relations, guaranteeing the integrability of the model [2410.05987]. The periodic Hamiltonian is recovered through derivatives of transfer matrices built from products of $R$-matrices along the chain, and the spectrum is accessible via the algebraic Bethe Ansatz. Bethe eigenstates are characterized by sets of rapidities $(\mathbf u, \mathbf w, \mathbf v)$ constrained by parity pairing: $\mathbf{u} = -\mathbf{u}$, and similarly for $\mathbf{w}$ and $\mathbf{v}$ [2602.01697].

## 2. Reflection Equations and Integrable Boundary States

Integrable boundary (or initial) states are constructed using solutions to the boundary reflection (“$K$-”) equations. There are two principal types relevant for the ABJM context:
- **Soliton-preserving (SP)**: $K_a(u)$ acts within the same fundamental/antifundamental representation.
- **Soliton-non-preserving (SNP)**: $\tilde K_a(u)$ maps between fundamental $\mathbf{4}$ and antifundamental $\bar{\mathbf{4}}$.

The SNP reflection equation for $\tilde K_a(u)$ reads
\begin{align*}
R_{12}(u-v)\,\tilde K_1(u)\,\bar R_{21}(u+v)\,\tilde K_2(v)
= \tilde K_2(v)\,\bar R_{12}(u+v)\,\tilde K_1(u)\,R_{21}(u-v),
\end{align*}
and similarly for $\tilde K_{\bar a}(u)$ [2602.01697]. The general $c$-number solution is
$$
\tilde K_a(u) = (1+2u)\,S + A, \quad S^T=S,\;A^T=-A,
$$
where $S$ and $A$ are arbitrary symmetric and antisymmetric $4\times 4$ matrices, respectively.

## 3. Fusion Formalism and $2n$-Site Chiral Matrix Product States

The fusion method systematically constructs higher-bond-dimension integrable MPS as the building blocks of chiral integrable boundary states. For the ABJM chain, the fusion of $n$ fundamental reflection matrices yields the $n$-fold $K$-matrix $K^{(n)}(u)$ acting on a $2n$-site block:
$$
R_{(\bar1\,2\cdots\overline{n-1}\,n),a}(u)
=R_{n\,a}(u)\,\bar R_{n-1,a}(u)\cdots R_{2a}(u)\,\bar R_{1a}(u),
$$
with the fused $K$-matrix satisfying a corresponding fused reflection equation [2602.01697]. 
Concrete realization for the four-site ($n=2$) case:
$$
\tilde K_{(\bar1\,2)}(u) = \tilde K_2(u)\,R_{12}(2u)\,\tilde K_{\bar1}(u),
$$
which determines the four-site chiral MPS building block at rapidity $u$. Translationally invariant $2n$-site MPS are then constructed by concatenating these blocks.

## 4. Exact Overlap Formulas: Determinant Structure and Selection Rules

Let $|\Psi\rangle = |\Phi(-1)\rangle^{\otimes L/2}$ denote the four-site chiral integrable MPS, and $|\mathbf{u},\mathbf{w},\mathbf{v}\rangle$ a Bethe eigenstate as constructed above. Nonzero overlaps occur only for parity-paired rapidity sets $\mathbf{u}=-\mathbf{u}$, $\mathbf{w}=-\mathbf{w}$, $\mathbf{v}=-\mathbf{v}$. The overlap assumes the explicit form:
\[
\frac{\langle\Psi^{S}|\mathbf{u},\mathbf{w},\mathbf{v}\rangle}
{\sqrt{\langle\mathbf{u},\mathbf{w},\mathbf{v}|\mathbf{u},\mathbf{w},\mathbf{v}\rangle}}
= 
(-1)^{\tfrac{N_u+N_v}{2}+N_w}(-2)^{L/2}
\prod_{i=1}^{N_u/2}\sqrt{\left|\frac{u_i+i/2}{u_i-i/2}\right|}\cdots
\sqrt{\frac{\det G_+}{\det G_-}\,[P_u\,P_w\,P_v]^{*}},
\]
where $G$ is the Gaudin matrix with elements $G_{ab} = \partial_{x_a}\phi_{x_b}$, $x_a$ running over all rapidities of type $u,w,v$, and $P_{u,w,v}$ are combinations of principal minors of the boundary $K$-matrix [2602.01697, eq.(4.16)]. For the antisymmetric case, the scalar factors involve principal Pfaffians instead of minors [2602.01697, eq.(4.18)]. 

The proof employs the transfer-matrix representation of the MPS, standard integrable techniques for scalar products (such as Slavnov's formula), and the restriction to on-shell, parity-symmetric Bethe states, allowing factorization 
$\det G = \det G_+\,\det G_-$.

## 5. Variants, Dressing Procedures, and Numerical Classification

Dressing the $K$-matrix solutions by insertion of auxiliary spaces and additional $R$-matrices yields operator-valued $K(u)$ and MPS of higher bond dimension, systematically enlarging the set of chiral integrable states [2602.01697, Sec.~3.4]. Exact dimension counts for small system sizes, e.g., $L=2$ and $L=3$, reveal that the explicit product/MPS constructions span a substantial but not exhaustive sector of the space of integrable boundary states. There exist further solutions, including two-site rank-one solutions of a degenerate mixed reflection equation at $u=-1$, not captured by invertible $K$-matrix constructions. For example, at $L=2$ out of $256$ states, imposing integrability leaves $196$ in the chiral subspace, while only $142$ are captured by four-site and two-site constructions. 

## 6. Fused Boundary Formalism and Boundary Effects

The fusion approach generalizes to constructions involving open boundary conditions. Here, double-row transfer matrices and associated $K^-$, $K^+$ matrices satisfy reflection equations guaranteeing integrability, and generate the open-chain Hamiltonian with bulk and explicit boundary contributions [2410.05987]:
$$
t_{open}(u) = \operatorname{Tr}_{00} \left\{ K^+_{00}(u) T_{00}(u) K^-_{00}(u) T_{00}(-u)\right\},
$$
with local Hamiltonian terms given by derivatives of $\ln t_{open}(u)$ at $u=0$ and boundary terms explicitly determined by the structure of $K$-matrices and associated fusion rules. Universal Kronecker-delta factors in boundary densities encode the protein of integrable boundary deformations classified via fusion.

## 7. Applications and Open Problems

Explicit overlap formulas are essential for computing quench dynamics, boundary correlation functions, and for characterizing integrable initial state manifolds in quantum integrable systems. The framework developed in [2602.01697] delivers a systematically extendable toolkit, comprising reflection equation analysis, fusion, and determinant-based overlap evaluation. Yet a complete classification of all chiral integrable states remains unresolved, particularly in higher bond-dimension sectors or for rank-deficient $K$-matrices. Numerical analyses confirm the existence of integrable sectors beyond the reach of current constructions, suggesting directions for further development of exact and numeric classification methods for integrable boundary/interacting states in complex quantum spin chains. 

---
Key references: [2410.05987], [2602.01697].

Source: https://www.emergentmind.com/topics/exact-overlap-formulas-with-bethe-states