Table of Contents
Fetching ...

2-Morita Equivalent Condensable Algebras and Domain Walls in 2+1D Topological Orders

Rongge Xu, Holiverse Yang

TL;DR

This work develops a complete framework for classifying 2‑Morita equivalent $E_2$ condensable algebras in a modular tensor category $\mathcal{C}$ by connecting three equivalent viewpoints: (i) left/right/full centers of condensable algebras and their induced $E_2$-modules, (ii) lagrangian algebras in the Drinfeld center $\mathfrak{Z}(\mathcal{C})$, and (iii) $1$-Morita classes of 1d domain walls in $\mathcal{C}$. The authors prove that two condensable algebras $A_1,A_2$ are 2‑Morita equivalent iff there exists a 1d algebra $B$ with $Z_l(B)\simeq A_1$ and $Z_r(B)\simeq A_2$, equivalently that $\mathcal{C}_{A_1}^{loc}\simeq\mathcal{C}_{A_2}^{loc}$ and $\mathfrak{Z}(\mathcal{C}_{A_1})\simeq\mathfrak{Z}(\mathcal{C}_{A_2})$. The framework uses a two-step condensation and folding to map domain-wall data to boundaries of $\mathfrak{Z}(\mathcal{C})$, establishing a universal Trinity of reversible arrows among condensable algebras, walls, and lagrangians. Concrete examples with group doubles (e.g., $\mathfrak{Z}(\mathrm{Vec}_{G})$ for $G=\mathbb{Z}_2,\mathbb{Z}_4,S_3$) demonstrate the method, including lattice realizations of 1d condensables, left/right centers, and lagrangians. The results extend to Witt-equivalent MTCs and open avenues for 0d defects, generalized Morita notions, and broader topological-operator formalisms, offering a robust, model-independent route to understand domain walls and phase transitions in 2+1D topological orders.

Abstract

We classify $E_2$ condensable algebras in a modular tensor category $\mathcal{C}$ up to 2-Morita equivalence. From a physical perspective, this is equivalent to providing a criterion for when different $E_2$ condensable algebras result in the same condensed topological phase in a 2d anyon condensation process. By considering the left and right centers of $E_1$ condensable algebras in $\mathcal{C}$, we exhaust all 2-Morita equivalent $E_2$ condensable algebras in $\mathcal{C}$ and provide a method to recover $E_1$ condensable algebras from 2-Morita equivalent $E_2$ condensable algebras. We also prove that intersecting Lagrangian algebras in $\mathcal{C} \boxtimes \overline{\mathcal{C}}$ with its left and right components generates all 2-Morita equivalent $E_2$ condensable algebras in $\mathcal{C}$. This paper establishes a complete interplay between $E_1$ condensable algebras in $\mathcal{C}$, 2-Morita equivalent $E_2$ condensable algebras in $\mathcal{C}$, and Lagrangian algebras in $\mathcal{C} \boxtimes \overline{\mathcal{C}}$. The relations between different condensable algebras can be translated into their module categories, which correspond to domain walls in topological orders. We introduce a two-step condensation process and study the fusion of domain walls. We also show that an automorphism of an $E_2$ condensable algebra may lead to a nontrivial braided autoequivalence in the condensed phase. As concrete examples, we interpret the categories of quantum doubles of finite groups. We also discuss examples beyond group symmetries. Moreover, our results can be generalized to Witt-equivalent modular tensor categories.

2-Morita Equivalent Condensable Algebras and Domain Walls in 2+1D Topological Orders

TL;DR

This work develops a complete framework for classifying 2‑Morita equivalent condensable algebras in a modular tensor category by connecting three equivalent viewpoints: (i) left/right/full centers of condensable algebras and their induced -modules, (ii) lagrangian algebras in the Drinfeld center , and (iii) -Morita classes of 1d domain walls in . The authors prove that two condensable algebras are 2‑Morita equivalent iff there exists a 1d algebra with and , equivalently that and . The framework uses a two-step condensation and folding to map domain-wall data to boundaries of , establishing a universal Trinity of reversible arrows among condensable algebras, walls, and lagrangians. Concrete examples with group doubles (e.g., for ) demonstrate the method, including lattice realizations of 1d condensables, left/right centers, and lagrangians. The results extend to Witt-equivalent MTCs and open avenues for 0d defects, generalized Morita notions, and broader topological-operator formalisms, offering a robust, model-independent route to understand domain walls and phase transitions in 2+1D topological orders.

Abstract

We classify condensable algebras in a modular tensor category up to 2-Morita equivalence. From a physical perspective, this is equivalent to providing a criterion for when different condensable algebras result in the same condensed topological phase in a 2d anyon condensation process. By considering the left and right centers of condensable algebras in , we exhaust all 2-Morita equivalent condensable algebras in and provide a method to recover condensable algebras from 2-Morita equivalent condensable algebras. We also prove that intersecting Lagrangian algebras in with its left and right components generates all 2-Morita equivalent condensable algebras in . This paper establishes a complete interplay between condensable algebras in , 2-Morita equivalent condensable algebras in , and Lagrangian algebras in . The relations between different condensable algebras can be translated into their module categories, which correspond to domain walls in topological orders. We introduce a two-step condensation process and study the fusion of domain walls. We also show that an automorphism of an condensable algebra may lead to a nontrivial braided autoequivalence in the condensed phase. As concrete examples, we interpret the categories of quantum doubles of finite groups. We also discuss examples beyond group symmetries. Moreover, our results can be generalized to Witt-equivalent modular tensor categories.
Paper Structure (34 sections, 55 theorems, 59 equations, 36 figures, 7 tables)

This paper contains 34 sections, 55 theorems, 59 equations, 36 figures, 7 tables.

Key Result

Theorem 1.1

Given a modular tensor category $\EuScript{C}$ with tensor unit $\mathbf{1}$ and consider all indecomposable separable algebras $\{B_i\}$ in $\EuScript{C}$, $L_i\simeq Z(B_i)$, $:A_{l_i} := Z_l(B_i)$ and $A_{r_i} :=Z_r(B_i)$ represent the full, left, right centers respectively,

Figures (36)

  • Figure 1: Gapped domain walls within a topological order $\EuScript{C}$ can be classified by the category of bimodules ${}_{B_i}{\EuScript{C}}_{B_i}$ of 1d condensable algebras $\{B_i\}$ in the trivial domain wall $\EuScript{C}$ (drawn in the dashed line). In particular, ${}_{B_1}{\EuScript{C}}_{B_1} \simeq \EuScript{C}$ for $B_1$ being the tensor unit $\mathbf{1}$ of the fusion category $\EuScript{C}$.
  • Figure 2: The fusion process from (a) to (b) shows that for a gapped domain wall ${}_B{\EuScript{C}}_B \simeq \EuScript{C}_{A_1}\mathop{\mathrm{\boxtimes}}\limits_{\EuScript{C}_{A_1}^{loc}}\Phi \mathop{\mathrm{\boxtimes}}\limits_{\EuScript{C}_{A_2}^{loc}}{}_{A_2} \EuScript{C}$ in $\EuScript{C}$, we have equivalent condensed phases $\EuScript{C}_{A_1}^{loc} \simeq \EuScript{C}_{A_2}^{loc}$ 'hidden inside' this wall ${}_B{\EuScript{C}}_B$. This process gives an intuitive way to understand why $B$ can recover 2-Morita equivalent 2d condensable algebras $A_1$ and $A_2$ in $\EuScript{C}$. Moreover, using folding trick from (b) to (c), the correspondence between 1d condensable algebras in $\EuScript{C}$ and lagrangian algebras in $\mathfrak{Z}(\EuScript{C})$ can be characterized by the correspondence between domain walls in $\EuScript{C}$ and gapped boundaries of $\mathfrak{Z}(\EuScript{C})$.
  • Figure 3: Results of this paper can be summarized by this Trinity, all arrows appear here are reversible. A$\overrightarrow{\text{rrow}}$ 1 was first discussed in DNO12, an alternative proof using 2-step condensation is provided in section \ref{['sec:pf_lag_alg']}; A$\overrightarrow{\text{rrow}}$ 2 was first stated by Davydov Dav10a, and we prove it using results in DNO12; A$\overrightarrow{\text{rrow}}$ 3 is proved in FFRS06; A$\overrightarrow{\text{rrow}}$ 4 is proved by Kong and Runkel in KR09; A$\overrightarrow{\text{rrow}}$ 5 has long been a folklore without enough discussions, we reformulate this 'forget' process in section \ref{['sec:centers']}; A$\overrightarrow{\text{rrow}}$ 6 is first discussed in this work.
  • Figure 4: $\EuScript{X}$ is the category of 0d domain wall conditions, which is a $\EuScript{M}$-$\EuScript{N}$-bimodule category. And a 1d gapped domain wall $\EuScript{M}$ (or $\EuScript{N}$) is described by a closed monoidal $\EuScript{C}_1$-$\EuScript{C}_2$-bimodule.
  • Figure 5: This figure shows the directions of 1d and 2d anyon condensations. For any $B_i$, ${}_{B_i}(\EuScript{C}_A)_{B_i}$ is 1-Morita equivalent as fusion categories to $\EuScript{C}_A$ since the Drinfeld center of ${}_{B_i}(\EuScript{C}_A)_{B_i}$ is equivalent to the Drinfeld center of $\EuScript{C}_A$, i.e. $\mathfrak{Z}(\EuScript{C}_A) \simeq \mathfrak{Z}({}_{B_i}(\EuScript{C}_A)_{B_i}) \simeq \EuScript{C} \mathop{\mathrm{\boxtimes}}\limits \overline{\EuScript{C}^{loc}_A}$Sch01.
  • ...and 31 more figures

Theorems & Definitions (145)

  • Remark 1.1
  • Theorem 1.1
  • Definition 2.1: Morita58
  • Definition 2.2
  • Example 2.3
  • Remark 2.1
  • Remark 2.2
  • Example 2.4
  • Remark 2.3
  • Remark 2.4
  • ...and 135 more