Entropy-Enthalpy Competition and Topological Phase Transition in SU(3)₃ Anyon Condensation

1. Introduction: From Entropy-Enthalpy Competition to Anyon Condensation

Physical systems universally exhibit two opposing tendencies: entropy-driven diffusion and disorder, and enthalpy-driven aggregation and order. This “entropy-enthalpy competition” principle manifests at the macroscopic level as the minimization of the free energy $\mathrm{F}=\mathrm{E}-\mathrm{T}\mathrm{S}$, but at the microscopic level it may have a deeper origin --- namely, transitions between different “entropy modes” (distinct topological phases and their associated vacuum topological entropy, see Sec. 3.4 for details), a perspective that continues the entropy-enthalpy competition framework introduced in Ref. [1].

In condensed matter physics, the discovery of topological order provides the cleanest mathematical laboratory for understanding “entropy modes”. The ground state of a topological phase possesses nontrivial long-range entanglement, and its entanglement entropy contains a universal constant term --- the topological entanglement entropy ${\mathrm{S}}_{\mathrm{t}\mathrm{o}\mathrm{p}\mathrm{o}}=\ln \mathcal{D}$, where $\mathcal{D}$ is the total quantum dimension [2,3]. This entropy is not a product of thermal motion, but an intrinsic property of the topological structure of the ground-state wave function, and can be regarded as the “vacuum entropy” of that topological phase.

When a system transits from one topological phase to another, a distinctive phase transition mechanism -- anyon condensation -- can occur [4,5,6]. In this process, a bosonic anyon undergoes macroscopic condensation, causing excitations that have nontrivial mutual statistics with it to become “confined”, while those that are “locally commuting” with it either survive or are identified. Across the condensation, the total quantum dimension $\mathcal{D}$ changes discontinuously, and the topological entanglement entropy changes accordingly.

The central thesis of this paper is: anyon condensation is a precise realization of entropy-enthalpy competition in topological order. In recent years, high-energy theorists have discovered that four-dimensional gauge theories (including QCD) possess non-invertible symmetries whose algebra is described by fusion categories [7,8]; the confinement-deconfinement phase transition can be reformulated precisely as the condensation of a condensation algebra in that category [9]. From this perspective, the calculation we perform in the $\mathrm{S}\mathrm{U}(3{)}_3$category can be viewed as an explicit demonstration of this program in a low-dimensional exactly solvable model. The condensation algebra $\mathrm{A}$represents the “enthalpy” aspect --it requires the system to pay the price of structure formation; while the decrease of topological entropy represents the “entropy” aspect --the system sacrifices a portion of its micro-degrees of freedom in exchange for overall stability. The net result of the condensation process is that the system transits from a high-entropy mode to a low-entropy mode, which is exactly the core picture of the entropy-enthalpy theory.

We choose the $\mathrm{S}\mathrm{U}(3{)}_3$ modular tensor category as the concrete object of study. This choice has a threefold significance:

1)

Algebraic isomorphism with QCD: The objects of the $\mathrm{S}\mathrm{U}(3{)}_3$ category carry a ${\mathbb{Z}}_3$ triality charge, which is isomorphic to the ${\mathbb{Z}}_3$ center symmetry of $\mathrm{S}\mathrm{U}\left(3\right)$ gauge theory. The quark representation $\left(\mathrm{1,0}\right)$ has triality $1$, the anti-quark $\left(\mathrm{0,1}\right)$ has triality $2$, and three quarks can fuse to the vacuum. Hence, this category serves as a precise algebraic projection of three-dimensional $\mathrm{S}\mathrm{U}\left(3\right)$ gauge theory onto a two-dimensional cross section.

2)

Connection with the frontier of non-invertible symmetries: Recent studies have shown that the full symmetry algebra of four-dimensional $\mathrm{S}\mathrm{U}\left(\mathrm{N}\right)$ gauge theories is a fusion category, and the confinement phase transition corresponds to the condensation of a condensation algebra in that category [7,8,9]. The present work can be regarded as an explicit realization of this program in a concrete solvable model.

3)

Providing a bridge to broader theoretical frameworks: The algebraic mechanism uncovered here—condensation of a bosonic algebra leading to confinement of non-zero triality objects—is a structural motif that may recur in other physical contexts, including generalized symmetries in quantum field theory. The $\mathrm{S}\mathrm{U}(3{)}_3$ model serves as an exactly solvable instance of this motif.

The structure of this paper is as follows: Sec.2 reviews the basic structure of the $SU(3{)}_3$ category; Sec.3 executes the anyon condensation procedure, computing the post-condensation object spectrum and entropy change; Sec.4 constructs the extended Levin-Wen model Hamiltonian, discussing the phase transition driven by the parameter $J$; Sec.5 presents a discussion: Secs.5.1--5.2 elucidate the connections of this work with QCD confinement and non-invertible symmetries; Sec.5.3 discusses the thermodynamic significance of the computed entropy change and its conceptual implications beyond the topological limit; Sec.5.4 provides a scaling analysis of the free-energy competition; Sec.5.5 discusses the limitations of the present calculation and its possible extensions to finite temperature and higher categories; Sec.5.6 provides an outlook on non-equilibrium entropy production based on quench dynamics. Sec.6 offers some heuristic (speculative) insights connecting anyon condensation to spacetime emergence, and Sec.7 gives the conclusion.

2. $\mathrm{S}\mathrm{U}(3{)}_3$ Modular Tensor Category: Basic Structure

The objects of the $\mathrm{S}\mathrm{U}(3{)}_{\mathrm{k}}$ category are labeled by two non-negative integers (Dynkin labels) $(\mathrm{n},\bar{\mathrm{n}})$, satisfying $\mathrm{n}+\bar{\mathrm{n}}\le \mathrm{k}$. For $\mathrm{k}=3$, there are ${\displaystyle \frac{(3+1)(3+2)}{2}}=10$ simple objects [10]. Each object carries a ${\mathbb{Z}}_3$ triality charge $\mathsf{\tau }=\mathrm{n}+2\bar{\mathrm{n}}\left(\mathrm{m}\mathrm{o}\mathrm{d}3\right)$.

2.1. Object List and Core Data

The general formula for the quantum dimension ${\mathrm{d}}_{(\mathrm{n},\bar{\mathrm{n}})}$[10] is:

$${\mathrm{d}}_{(\mathrm{n},\bar{\mathrm{n}})}={\displaystyle \frac{\sin \frac{(\mathrm{n}+1)\mathsf{\pi }}{\mathrm{k}+3}\sin \frac{(\bar{\mathrm{n}}+1)\mathsf{\pi }}{\mathrm{k}+3}\sin \frac{(\mathrm{n}+\bar{\mathrm{n}}+2)\mathsf{\pi }}{\mathrm{k}+3}}{{\sin }^2\frac{\mathsf{\pi }}{\mathrm{k}+3}\sin \frac{2\mathsf{\pi }}{\mathrm{k}+3}}}$$

(1)

Here and throughout this paper, the quantum dimensions $d_{(n,\bar{n})}$ are those of the simple objects in the fusion category, not the dimensions of the corresponding $SU\left(3\right)$ Lie algebra representations. For instance, the object $\left(\mathrm{1,1}\right)$ has quantum dimension $d_{\left(\mathrm{1,1}\right)}=3$, whereas the adjoint representation of $SU\left(3\right)$ has dimension $8$. This discrepancy arises because the $SU(3{)}_3$ fusion category is a truncated (level-$k$) deformation of the representation theory of $SU\left(3\right)$; only a finite set of representations with $n+\bar{n}\le k$ survive as simple objects, and their quantum dimensions are determined by the modular data of the associated conformal field theory, not by the classical Weyl dimension formula.

For $\mathrm{k}=3$, $\mathrm{k}+3=6$ the above formula yields the values listed in Table 1.

2.2. Fusion Rules and Triality Structure

The fusion rules of the category [10,11,12] are given by the tensor product decomposition: $a\otimes b=\underset{c}{⨁}{N}_{ab}^{c}c$, where $N_{ab}^c\in \mathbb{Z}$are the fusion coefficients. For the subsequent analysis of this paper, the following fusion rules are of crucial importance:

1)

$\left(\mathrm{1,0}\right)\otimes \left(\mathrm{0,1}\right)=\left(\mathrm{0,0}\right)\oplus \left(\mathrm{1,1}\right)$ (quark-antiquark form a meson);

2)

$\left(\mathrm{1,0}\right)\otimes \left(\mathrm{1,0}\right)\otimes \left(\mathrm{1,0}\right)=\left(\mathrm{0,0}\right)\oplus \dots$(three quarks form a baryon);

3)

$\left(\mathrm{3,0}\right)\otimes \left(\mathrm{3,0}\right)=\left(\mathrm{0,3}\right)$(the self-fusion of the boson yields another member of the ${\mathbb{Z}}_3$ orbit);

The triality charge is additive modulo 3 under fusion: $\mathsf{\tau }(\mathrm{a}\otimes \mathrm{b})=\mathsf{\tau }\left(\mathrm{a}\right)+\mathsf{\tau }\left(\mathrm{b}\right)$ (mod 3).This means that objects with triality $0$ form a subcategory, while objects with non-zero triality cannot fuse to the vacuum by themselves -- this is precisely the categorical formulation of “color singlet” in QCD.

2.3. Modular $\mathbf{S}$ Matrix

The modular $S$ matrix [13,14,15] describes the mutual statistics among anyons, with matrix elements:

$$S_{(n,\bar{n}),(m,\bar{m})}={\displaystyle \frac{1}{\mathcal{D}}}{d}_{(n,\bar{n})}{d}_{(m,\bar{m})}{\displaystyle \frac{\sin \frac{(n+1)(m+1)\pi }{k+3}\sin \frac{(\bar{n}+1)(\bar{m}+1)\pi }{k+3}}{{\sin }^2\frac{\pi }{k+3}\sin \frac{2\pi }{k+3}}}$$

(2)

where $\mathcal{D}=\sqrt{\sum _{\mathrm{a}}{\mathrm{d}}_{\mathrm{a}}^2}$ is the total quantum dimension [16].

3. Anyon Condensation: From High-Entropy Mode to Low-Entropy Mode

3.1. Choice of the Condensation Algebra

In a modular tensor category $\mathcal{C}$, anyon condensation is characterized by a connected étale algebra $A=\underset{a}{⨁}{n}_{a}a$, where each $a$ must satisfy the bosonic statistics condition $h_a\in \mathbb{Z}$.

From Table 1, in $SU(3{)}_3$ the objects with integer topological spin are only (0,0), (3,0) and (0,3). Since (3,0) is a simple current of order three--$\left(\mathrm{3,0}\right)\otimes \left(\mathrm{3,0}\right)=(0,3)$ and $\left(\mathrm{3,0}\right)\otimes (0,3)=(0,0)$--the correct condensation algebra must include the full ${\mathbb{Z}}_3$ orbit generated by (3,0):

$$A=\left(\mathrm{0,0}\right)\oplus \left(\mathrm{3,0}\right)\oplus \left(\mathrm{0,3}\right)$$

Its quantum dimension is ${\mathrm{d}}_{\mathrm{A}}=1+{\mathrm{d}}_{\left(\mathrm{3,0}\right)}{+\mathrm{d}}_{(0,3)}=1+1+1=3$. The fusion rule $\left(\mathrm{3,0}\right)\otimes (0,3)=(0,0)$ and $\left(\mathrm{3,0}\right)\otimes (3,0)=(0,3)$ guarantee that AA is a legitimate connected algebra, which has been rigorously proved in Ref. [5].

3.2. Confinement Criterion and the Fate of Quarks

After condensation, whether an object $X$of the original category $\mathcal{C}$ can survive in the condensed phase $\mathcal{D}$depends on its mutual statistics with the condensation algebra $A$[5]. Specifically, $X$ is confined if and only if there exists some $a\in A$ such that the modular $S$ matrix element does not satisfy the “local commutativity” condition:

$$S_{X,a}\ne {\displaystyle \frac{d_X{d}_a}{\mathcal{D}}}$$

(3)

Take the quark $X=\left(\mathrm{1,0}\right)$ and the boson $a=\left(\mathrm{3,0}\right)$ as an example. Substituting into the $S$matrix formula:

$$S_{\left(\mathrm{1,0}\right),\left(\mathrm{3,0}\right)}={\displaystyle \frac{1}{\mathcal{D}}}{d}_{\left(\mathrm{1,0}\right)}{d}_{\left(\mathrm{3,0}\right)}\cdot {\displaystyle \frac{\sin \frac{\left(2\right)\left(4\right)\pi }{6}\sin \frac{\left(1\right)\left(1\right)\pi }{6}}{{\sin }^2\frac{\pi }{6}\sin \frac{2\pi }{6}}}={\displaystyle \frac{1}{\mathcal{D}}}{d}_{\left(\mathrm{1,0}\right)}{d}_{\left(\mathrm{3,0}\right)}\cdot (-2)\ne {\displaystyle \frac{d_X{d}_a}{\mathcal{D}}}$$

(4)

Since the factor $-2\ne 1$, the mutual statistics between the quark $\left(\mathrm{1,0}\right)$ and $\left(\mathrm{3,0}\right)$ is nontrivial; therefore, after condensing $\mathrm{A}$, the quark is confined. By the same reasoning, all objects with non-zero triality ($\left(\mathrm{1,0}\right)$,$\left(\mathrm{0,1}\right)$,$\left(\mathrm{2,0}\right)$,$\left(\mathrm{0,2}\right)$, etc.) are confined. This is the categorical realization of color confinement: excitations carrying non-zero triality (color charge) cannot exist as independent local states.

For the non-Abelian objects $\left(\mathrm{2,1}\right)$ and $\left(\mathrm{1,2}\right)$, which carry non-zero triality ($\tau =1$ and $\tau =2$, respectively), the triality rule already indicates confinement. In the specific fusion rules of $SU(3{)}_3$, an object with non-zero triality cannot split into components with zero triality under condensation, because triality is additive modulo 3 and is preserved under fusion. Consequently, $\left(\mathrm{2,1}\right)$ and $\left(\mathrm{1,2}\right)$ are fully confined and do not contribute to the condensed category; a detailed verification is provided in Appendix A.4. The qualitative conclusion--confinement of all non-zero triality objects and a decrease of topological entropy--is robust and does not depend on any refinement.

3.3. Object Spectrum of the Condensed Category $\mathcal{D}$

Carrying out the full condensation procedure (including the identification step), we obtain the simple objects of the condensed category $\mathcal{D}$, listed in Table 2.

3.4. Transition of Entropy Modes: Computing the Topological Entropy Change

In this context, an “entropy mode” refers to a topological phase and its corresponding vacuum topological entropy $S_{\mathrm{t}\mathrm{o}\mathrm{p}\mathrm{o}}=\ln \mathcal{D}$. Different topological phases possess different total quantum dimensions $\mathcal{D}$, hence different topological entropies. When a system transits from one topological phase to another via anyon condensation, the vacuum topological entropy undergoes a jump --- we call this process a “transition of entropy modes”.

Across condensation, the total quantum dimension changes [5,16]. Before condensation:

$${\mathcal{D}}_{\mathcal{C}}^2=\sum _{a\in \mathcal{C}}{d}_a^2=36,{\mathcal{D}}_{\mathcal{C}}=6,S_{\mathrm{t}\mathrm{o}\mathrm{p}\mathrm{o}}^{\mathrm{d}\mathrm{e}\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{f}\mathrm{i}\mathrm{n}\mathrm{e}\mathrm{d}}=\ln 6\approx 1.79$$

(5)

The condensation algebra is $A=\left(\mathrm{0,0}\right)\oplus \left(\mathrm{3,0}\right)\oplus \left(\mathrm{0,3}\right)$ with quantum dimension $d_A=3$. The total quantum dimension of the condensed category is given by the standard condensation formula:

$${\mathcal{D}}_{\mathcal{D}}={\displaystyle \frac{{\mathcal{D}}_{\mathcal{C}}}{\sqrt{{\mathrm{d}}_{\mathrm{A}}}}}={\displaystyle \frac{6}{\sqrt{3}}}=2\sqrt{3}$$

Thus:

$${\mathcal{D}}_{\mathcal{D}}^2=12,{\mathrm{S}}_{\mathrm{t}\mathrm{o}\mathrm{p}\mathrm{o}}^{\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{f}\mathrm{i}\mathrm{n}\mathrm{e}\mathrm{d}}=\ln (2\sqrt{3})\approx 1.2425$$

(6)

Topological entropy change:

$$\Delta {\mathrm{S}}_{\mathrm{t}\mathrm{o}\mathrm{p}\mathrm{o}}={\mathrm{S}}_{\mathrm{t}\mathrm{o}\mathrm{p}\mathrm{o}}^{\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{f}\mathrm{i}\mathrm{n}\mathrm{e}\mathrm{d}}-{\mathrm{S}}_{\mathrm{t}\mathrm{o}\mathrm{p}\mathrm{o}}^{\mathrm{d}\mathrm{e}\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{f}\mathrm{i}\mathrm{n}\mathrm{e}\mathrm{d}}=-{\displaystyle \frac{1}{2}}\ln 3\approx -0.549$$

(7)

As a reference, in the toric code the condensation of the $m$ anyon gives $\Delta S_{topo}=-\ln 2\approx -0.69$. The value $-0.549$ obtained here corresponds to a mild confinement --- quarks are confined, while colorless excitations such as glueballs survive. This negative topological entropy change provides a quantitative signal for entropy-enthalpy competition: by forming the $\left(\mathrm{3,0}\right)$ condensate (the “enthalpy” process driven by the parameter $J$), the system sacrifices a portion of topological entropy and enters a more ordered low-entropy mode. The condensation energy $\Delta E_{\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{d}}$ is the energy released or absorbed when the system transits from the deconfined phase to the confined phase due to the formation of the condensate. Although we have not independently computed the numerical value of the condensation energy in the pure topological limit, the correspondence between the condensation algebra $A$ and the topological entropy change $\Delta S_{\mathrm{t}\mathrm{o}\mathrm{p}\mathrm{o}}$ lays the foundation for a complete quantification of the free-energy competition in a finite-temperature Hamiltonian. This is the first explicit realization of the entropy-enthalpy theoretical framework in a topological phase transition.

This work continues the entropy-enthalpy competition framework introduced in Ref.[1]. In that framework, the dynamics of physical systems is understood as a competition between two tendencies, entropy and enthalpy: entropy drives diffusion and disorder, enthalpy drives condensation and structure. The anyon condensation process described above provides an exactly solvable topological instance of this competition: through the formation of the $\left(\mathrm{3,0}\right)$ condensate (represented by the condensation algebra $A$, playing the role of “enthalpy”), the system sacrifices a portion of its topological entropy ($\Delta S_{\mathrm{t}\mathrm{o}\mathrm{p}\mathrm{o}}\approx -0.549$, playing the role of “entropy”), thereby entering a more ordered low-entropy mode. This exactly computable entropy change provides the first example of quantitative investigation of the entropy-enthalpy competition framework in a topological phase transition.

4. Extended Levin-Wen Model: Entropy Mode Transition Driven by Parameter $\mathbf{J}$

4.1. Standard Levin-Wen Hamiltonian

The Levin-Wen model [17,18] is an exactly solvable model defined on a two-dimensional honeycomb lattice, whose low-energy excitations are described by the Drinfeld center $Z\left(\mathcal{C}\right)$. Its Hamiltonian is:

$$H_{\mathrm{L}\mathrm{W}}=-\sum _v{Q}_v-\sum _p{B}_p$$

(8)

where $Q_v$ constrains the strings meeting at a vertex to satisfy the fusion rules, and $B_p$ drives the string-net condensation on the plaquettes. The ground state of this model possesses topological order described by the Drinfeld center $Z\left(\mathcal{C}\right)$.

4.2. Extended Hamiltonian with a Condensation Term

To realize parameter-driven anyon condensation, we adopt the construction of Christian et al. [19] and add a tunable condensation term to the standard Hamiltonian:

$$H(\mathcal{C},A,J)=H_{\mathrm{L}\mathrm{W}}-J\sum _p{B}_p^{\left(A\right)}$$

(9)

where the condensation operator is defined as:

$$B_p^{\left(A\right)}=\sum _{a\in A}{\displaystyle \frac{d_a}{{\mathcal{D}}_A}}{B}_p^{\left(a\right)}$$

(10)

Here $B_p^{\left(a\right)}$ is the operator that inserts a closed $a$-loop on the plaquette $p$, and ${\mathcal{D}}_A=\sqrt{\sum _{a\in A}{d}_a^2}$ is a normalization factor.

4.3. $\mathit{J}$

as the Control Parameter for “Enthalpy”

At the conceptual analogy level, $J$ drives the formation of the condensate and thus plays the role of “enthalpy” in the entropy-enthalpy theory --- it tends to drive the system into a more ordered low-entropy phase.

-

$\mathit{J}=0$ phase:The system is in the ground state of the standard Levin-Wen model, with low-energy excitations described by the Drinfeld center Z($\mathcal{C}$) of the $SU(3{)}_3$ modular tensor category.All anyons exist as independent excitations.This is the high-entropy mode, corresponding to the deconfined quark-gluon plasma phase.

-

$\mathrm{J}\gg 1$ phase: The condensation term dominates; the system tends to create a condensate of the $\left(\mathrm{3,0}\right)$ boson on every plaquette.The low-energy effective theory becomes the Drinfeld center $Z\left({\mathcal{C}}_{cond}\right)$ of the condensed category, and quarks are confined.This is the low-entropy mode, corresponding to the confined hadron phase.

-

Critical point$\mathbf{J}={\mathbf{J}}_{\mathbf{c}}$: There is a topological quantum phase transition between the two phases, accompanied by a gap closing (in the case of a continuous transition) and a jump of the topological entanglement entropy.

4.4. Physical Observables

To establish a quantitative connection with QCD physics, we define the following observables:

1)

String tension $\sigma \left(J\right)$: In the $J\gg 1$ phase, $\sigma >0$; in the $J\ll 1$ phase, $\sigma \to 0$.

2)

Expectation value of the Wilson loop: Follows an area law in the confined phase and a perimeter law in the deconfined phase.

2)

Topological entanglement entropy $S_{topo}\left(J\right)$: Undergoes a jump $\Delta S_{topo}\approx -0.549$ at $J=J_c$.

5. Analysis and Discussion: QCD Confinement, Thermodynamic Interpretation, and Future Directions

5.1. As an Algebraic Projection of Three-Dimensional QCD Confinement on a Two-Dimensional Cross Section

Although anyon condensation is strictly a phenomenon of two-dimensional topological order, the triality structure of the $SU(3{)}_3$category is exactly isomorphic to the ${\mathbb{Z}}_3$center symmetry of three-dimensional $SU\left(3\right)$ gauge theory [20,21,22,23,24,25,26]. Hence, the confinement mechanism discovered in this paper can be viewed as an algebraic projection of three-dimensional QCD confinement:

In three dimensions, the quark flux tube appears on a two-dimensional cross section as a string with an endpoint carrying a non-zero triality anyon.

In three dimensions, the condensation of bosonic bound states (such as glueballs) appears on a two-dimensional cross section as the condensation of the $\left(\mathrm{3,0}\right)$ anyon.

In three dimensions, the confinement of quarks appears on a two-dimensional cross section as the confinement of anyons with non-zero triality.

In physical scenarios with effective dimensional reduction, such as at finite temperature or under strong magnetic fields, this projection may have a direct physical correspondence, rather than being merely a formal analogy.

5.2. Connection with the Frontier of Non-Invertible Symmetries

Recent studies on generalized symmetries in four-dimensional gauge theories [9,27,28,29] have revealed a profound mathematical structure: $SU\left(N\right)$Yang-Mills theory not only possesses the traditional ${\mathbb{Z}}_N$ 1-form symmetry, but also exhibits non-invertible symmetries arising from the fusion of time-reversal symmetry and the 1-form symmetry. The algebraic structure of these non-invertible symmetries is described by fusion categories (or higher categories), whose objects correspond to different topological operators (including Wilson lines, ‘t Hooft lines, and their non-invertible fusion products).

Within this framework, the confinement-deconfinement phase transition can be reinterpreted as the condensation of a condensation algebra in the symmetry category [27,28,30] --- which is algebraically fully parallel to the condensation procedure we have carried out in the $SU(3{)}_3$category. Our $SU(3{)}_3$ category captures the ${\mathbb{Z}}_3$ grading structure and the fusion rules of the triality charges of the symmetry category of $SU\left(3\right)$ gauge theory; therefore, our condensation calculation can be viewed as an algebraic projection and a proof of concept of the four-dimensional non-invertible symmetry phase transition on a two-dimensional cross section.

5.3. Two Different Kinds of Entropy

It is worth noting that the anyon condensation process described in this paper leads to a decrease of topological entanglement entropy ($\Delta {\mathrm{S}}_{\mathrm{t}\mathrm{o}\mathrm{p}\mathrm{o}}<0$), whereas the classical cosmological picture tells us that the overall evolution of the universe is a process of ever-increasing entropy. This apparent contradiction reveals one of the most profound insights of the entropy-enthalpy theory: the evolution of the universe is a dual process in which the thermodynamic entropy continually increases while the vacuum topological entropy decreases in a stepwise fashion.

First, it is essential to clearly distinguish two essentially different “entropies”:

1)

Thermodynamic entropy${\mathbf{S}}_{\mathbf{t}\mathbf{h}\mathbf{e}\mathbf{r}\mathbf{m}}$: The logarithm of the number of microscopic states, measuring the degree of internal disorder of the system. According to the second law of thermodynamics, the thermodynamic entropy of an isolated system never decreases; the total thermodynamic entropy of the universe increases continuously during expansion and structure formation.

2)

Topological entanglement entropy${\mathrm{S}}_{\mathrm{t}\mathrm{o}\mathrm{p}\mathrm{o}}=\ln \mathcal{D}$[2,3]: A measure of the long-range entanglement structure of the ground-state wave function, reflecting the “complexity” of the vacuum itself as a topological phase. It is not a product of thermal motion, but an intrinsic topological property of the ground state. The condensation process in this paper causes a decrease of the total quantum dimension $\mathcal{D}$, and hence ${\mathrm{S}}_{\mathrm{t}\mathrm{o}\mathrm{p}\mathrm{o}}$ decreases. This signals that the vacuum enters a more “ordered” and “rigid” phase.

In the actual physical process of anyon condensation, the decrease of the ground-state topological entropy ($\Delta S_{\mathrm{t}\mathrm{o}\mathrm{p}\mathrm{o}}<0$) does not violate the second law of thermodynamics, because it is necessarily accompanied by the release of latent heat, the increase of entropy of the excited states, or the increase of entropy of the environment, so that the total thermodynamic entropy does not decrease. This is analogous to water freezing into ice: the microscopic order of the ice increases (“configurational entropy” decreases), but the latent heat released increases the entropy of the environment by an even greater amount, so the total entropy still increases.

Therefore, the “entropy decrease” discussed in this paper refers specifically to the decrease of the complexity of the vacuum topological structure, whereas the “entropy increase” in the overall cosmic evolution refers to the increase of the total thermodynamic entropy. The two are not only not contradictory, but together constitute a complete picture of cosmic evolution: the vacuum is gradually “frozen” into a stable background structure through a series of condensation phase transitions (topological entropy decrease), and each freezing event releases more thermodynamic entropy into the universe (total entropy increase).

5.4. Free-Energy Competition and Scaling Analysis of $\Delta E_{cond}$

Within the entropy-enthalpy competition framework, the complete criterion for the phase transition is given by the free-energy difference

$$\Delta F\left(J\right)=\Delta E_{\mathrm{c}ond}\left(J\right)-T\Delta S_{topo}$$

where $\Delta E_{\mathrm{c}ond}\left(J\right)\equiv E_0\left(J\right)-E_0\left(0\right)$ is the $\mathit{J}$ -dependent condensation energy (the energy cost or gain of deforming the ground state from the $J=0$ Levin-Wen fixed point), and $\Delta S_{topo}=-{\displaystyle \frac{1}{2}}\ln 3\approx -0.549$ is the exactly computed topological entropy change. For $J>J_c$, $\Delta E_{\mathrm{c}ond}\left(J\right)<0$, indicating that the system gains energy by forming the condensate. According to $F=E-TS$, the topological entropy decrease ($\Delta S_{topo}<0$) makes $-T\Delta S_{topo}>0$, which acts as a free-energy barrier. The phase transition can occur when the negative condensation energy overcomes this barrier, driving the system into the low-entropy phase.

A scaling argument for $\mathsf{\Delta }{\mathit{E}}_{\mathbf{c}\mathit{o}\mathit{n}\mathit{d}}\left(\mathit{J}\right)$. Although a full numerical computation of $\Delta E_{\mathrm{c}ond}\left(J\right)$ is beyond the scope of this work, we can establish its qualitative $J$-dependence via a simple variational argument. Consider the extended Hamiltonian $H\left(J\right)=H_{LW}-J\sum _p{B}_p^{\left(A\right)}$. In the limit $J\to \infty$, the condensation term dominates, and the ground-state wave function is forced to satisfy $B_p^{\left(A\right)}|{\mathsf{\Psi }}_A⟩=|{\mathsf{\Psi }}_A⟩$ on every plaquette, up to corrections that vanish as $J\to \infty$. This defines the “fully condensed” variational state $|{\mathsf{\Psi }}_A⟩$.

The variational energy density ${ϵ}_0\left(J\right)\equiv ⟨{\mathsf{\Psi }}_A\left|H\right(J\left)\right|{\mathsf{\Psi }}_A⟩/N$ (where $N$ is the number of plaquettes) then takes the form

$${ϵ}_0\left(J\right)=-J+{ϵ}_{\mathrm{e}ff}+o\left(1\right),J\to \infty$$

where the term $-J$comes from the condensation term acting on $|{\mathsf{\Psi }}_A⟩$ (since $B_p^{\left(A\right)}|{\mathsf{\Psi }}_A⟩=|{\mathsf{\Psi }}_A⟩$), and ${ϵ}_{\mathrm{e}ff}=⟨{\mathsf{\Psi }}_A\left|H_{LW}\right|{\mathsf{\Psi }}_A⟩/N$is a finite constant representing the residual energy cost of the Levin-Wen plaquette and vertex constraints in the fully condensed background. The $o\left(1\right)$ correction represents virtual fluctuations into uncondensed sectors, which are suppressed in the large-$J$ limit.

Therefore, the ground-state energy in the condensed phase scales as $E_0\left(J\right)=N(-J+{ϵ}_{\mathrm{e}ff})+o\left(N\right)$. Comparing this with the deconfined-phase ground-state energy $E_0\left(0\right)=N{ϵ}_0\left(0\right)+o\left(N\right)$, the $J$-dependent part of the condensation energy is

$$\Delta E_{\mathrm{c}ond}\left(J\right)\equiv E_0\left(J\right)-E_0\left(0\right)=-J\cdot N+o\left(N\right)$$

which is negative and extensive (proportional to the system size $N$), growing linearly in magnitude with $J$. (Any $J$-independent constant offset $N\cdot c$ merely shifts the zero of the free energy and does not affect the phase transition criterion.)

In contrast, the topological entropy change $\Delta S_{topo}=-{\displaystyle \frac{1}{2}}\ln 3$ is an intensive constant of order $\mathcal{O}\left(1\right)$, independent of $N$. Substituting into the free-energy difference, we obtain, in the thermodynamic limit $N\to \infty$,

$$\Delta F\left(J\right)\sim -J\cdot N+{\displaystyle \frac{T}{2}}\ln 3+\mathcal{O}\left(N^0\right)$$

Crucially, the $+T\ln 3/2$ term is positive (i.e., it acts as a free-energy penalty), precisely because the system sacrifices topological entropy ($\Delta S_{topo}<0$) to enter the condensed phase. This penalty is, however, intensive and hence subdominant to the extensive condensation energy $-JN$ for any $J>0$ in the thermodynamic limit. This scaling argument rigorously demonstrates that the condensation term acts as the “enthalpy” that drives the system into the low-entropy phase, overcoming the entropic barrier $\Delta S_{topo}<0$. The critical point $J_c$is determined by the balance between the $J$-dependent volume term and subleading corrections; beyond it, the confined phase is thermodynamically stable.

Physical origin of  $\mathsf{\Delta }{\mathit{E}}_{\mathbf{c}\mathit{o}\mathit{n}\mathit{d}}\left(\mathit{J}\right)$. The scaling analysis above establishes that $\Delta E_{\mathrm{c}ond}\left(J\right)$ is negative and extensive for $J\gg J_c$, but it does not specify the mechanism by which the condensation energy emerges. Here we clarify this mechanism. In the extended Hamiltonian $H\left(J\right)=H_{LW}-J\sum _p{B}_p^{\left(A\right)}$, the two terms are in direct competition. At $J=0$, the ground state $|{\mathsf{\Psi }}_0⟩$ of the standard Levin-Wen model is an equal-weight superposition of all string-net configurations satisfying the fusion rules. In this state, $⟨{\mathsf{\Psi }}_0\left|B_p^{\left(A\right)}\right|{\mathsf{\Psi }}_0⟩=0$, because $B_p^{\left(A\right)}$ creates a closed $\left(\mathrm{3,0}\right)$-loop on plaquette $p$, which is orthogonal to the symmetric superposition. Thus there is no energy gain from the condensation term at $J=0$.

As $J$ increases from zero, the system can gain energy $-J⟨B_p^{\left(A\right)}⟩$ by polarizing the ground state toward configurations with non-zero $\left(\mathrm{3,0}\right)$-loop density. This polarization, however, frustrates the plaquette stabilizers $B_p$ in $H_{LW}$, which prefer the flat superposition. The resulting condensation energy takes the form

$$\Delta E_{\mathrm{c}ond}\left(J\right)=\Delta E_{LW}\left(J\right)-J⟨B_p^{\left(A\right)}{⟩}_J$$

where $\Delta E_{LW}\left(J\right)\ge 0$ is the energy cost of deforming the ground state away from the Levin-Wen fixed point, and $⟨B_p^{\left(A\right)}{⟩}_J$ is the expectation value of the condensation operator in the deformed state. As $J$ increases, $⟨B_p^{\left(A\right)}{⟩}_J$grows monotonically, while $\Delta E_{LW}\left(J\right)$ grows quadratically in the condensate order parameter (as in standard Ginzburg–Landau theory for a first-order or continuous transition). The competition between the linear gain from condensation and the quadratic cost from frustration yields a critical value $J_c$ determined by the condition

$$\Delta E_{\mathrm{c}ond}\left(J_c\right)=T_c\Delta S_{topo}$$

where $T_c$ is the critical temperature of the topological phase transition. For $J<J_c$, the entropic barrier $-T\Delta S_{topo}>0$ dominates and the deconfined phase is stable; for $J>J_c$, the condensation energy (which is negative) overcomes this barrier and the confined phase becomes thermodynamically preferred. This establishes $\Delta E_{\mathrm{c}ond}\left(J\right)$ as a monotonically decreasing function of $J$, crossing the critical threshold at a finite $J_c$.

By computing $E_0\left(J\right)$ on a finite lattice using exact diagonalization or tensor network methods, and combining it with the $\Delta S_{topo}\approx -0.549$ obtained in this paper, one can determine $\Delta F\left(T\right)$ and thus fully verify the entropy-enthalpy competition quantitatively. This is left for future work.

5.5. Limitations: From Finite Temperature to Higher Categories

The scaling analysis in Sec. 5.4 establishes the thermodynamic structure of the phase transition in the large-J limit and provides a clear roadmap for future numerical verification. However, the condensation computation performed in the $SU(3{)}_3$ category in this paper provides an exactly solvable microscopic template for the above cosmological picture, but its scope of applicability has clear boundaries: the current work is carried out in the pure topological limit ($T=0$) and within the framework of two-dimensional categories.

Extending to a complete physical theory requires progress in two dimensions.

Finite temperature and free-energy competition. Within the entropy-enthalpy competition framework established in Sec. 5.4, the complete criterion for the phase transition is given by the free-energy difference $\Delta \mathrm{F}=\Delta {\mathrm{E}}_{\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{d}}-\mathrm{T}\Delta {\mathrm{S}}_{\mathrm{t}\mathrm{o}\mathrm{p}\mathrm{o}}$, where $\Delta {\mathrm{S}}_{\mathrm{t}\mathrm{o}\mathrm{p}\mathrm{o}}=-{\displaystyle \frac{1}{2}}\ln 3$is the exactly computable topological entropy change obtained in Sec.3.4, and $\Delta {\mathrm{E}}_{\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{d}}$ is the condensation energy. Although the scaling argument in Sec. 5.4 establishes that $\Delta {\mathrm{E}}_{\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{d}}$ is extensive and negative for $\mathrm{J}\gg {\mathrm{J}}_{\mathrm{c}}$, its precise numerical value requires a full computation of the ground-state energy ${\mathrm{E}}_0\left(\mathrm{J}\right)$ on a finite lattice using exact diagonalization or tensor network methods. Combining this with the $\Delta {\mathrm{S}}_{\mathrm{t}\mathrm{o}\mathrm{p}\mathrm{o}}\approx -0.549$ obtained in this paper, one can determine $\Delta \mathrm{F}\left(\mathrm{T}\right)$ and thus quantitatively verify the entropy-enthalpy competition at finite temperature.

Extension to higher categories. The algebraic mechanism revealed in this paper — that the condensation of a bosonic algebra leads to topological entropy decrease and excitation confinement — has been rigorously verified in two-dimensional fusion categories. However, a fully four-dimensional gauge theory, including QCD, is naturally described by higher-categorical structures. The extension of the present condensation framework to three dimensions requires the language of fusion 2-categories, where the simple objects are 2-morphisms and the condensation algebra becomes a composite of higher-morphisms [27]. A natural conjecture is that the ${\mathbb{Z}}_3$ triality structure of $SU(3{)}_3$ captured here is the dimensional reduction of a more intricate 2-categorical condensation in the full four-dimensional symmetry category. If a similar condensation procedure can be realized in higher categories, the picture of a “ladder of entropy mode transitions” sketched in Sec. 6 -- from spacetime emergence to the QCD phase transition -- will acquire a more solid mathematical foundation.

These two directions--finite-temperature free-energy analysis and higher-categorical extension --together constitute the key steps for the entropy-enthalpy competition framework to progress from a conceptual scheme to a quantitative theory that can be confronted with physical reality.

5.6. Outlook: Theoretical Framework for Quench Dynamics and Non-Equilibrium Entropy Production

The extended Levin-Wen Hamiltonian $H\left(J\right)=H_{LW}-J\sum _p{B}_p^{\left(A\right)}$ constructed in this paper not only describes static topological phase transitions, but also provides a natural platform for studying non-equilibrium dynamics. In this section, we outline a theoretical framework for such studies and identify key phenomena that are expected to emerge. The discussion is intended as a roadmap for future numerical investigations rather than a report of completed simulations.

One direction worth exploring is quench dynamics: the system is initially prepared in the deconfined phase (high-entropy mode) with $J_i\ll 1$, and at $t=0$ the parameter is suddenly changed to $J_f\gg 1$; the system will then undergo a non-equilibrium transition from the high-entropy mode to the low-entropy mode [31,32]. This process can be described within the theoretical framework of non-equilibrium topological order [33].

In this process, the occupation number densities $n_a\left(t\right)$ of the anyons follow a master equation derived from the Hamiltonian [33]:

$${\displaystyle \frac{d{n}_a}{dt}}={\displaystyle \frac{1}{2}}\sum _{b,c}{N}_{bc}^a{W}_{bc\to a}{n}_b{n}_c-\sum _{b,c}{N}_{ab}^c{W}_{a\to bc}{n}_a,$$

where the transition rates $W_{ab\to c}$ are determined by the fusion coefficients and quantum dimensions. The categorical entropy of the system

$$S(t)=\ln \mathcal{D}-\sum _a{n}_a(t)\ln d_a-\sum _a{n}_a(t)\ln n_a(t)$$

and its production rate $\sigma \left(t\right)=dS/dt\ge 0$ provide quantitative measures to track this non-equilibrium process.

Based on this framework, we identify the following phenomena as promising targets for future numerical studies:

Thermalization time and scaling. The characteristic time ${\tau }_{th}$ for the system to thermalize from the high-entropy mode to the low-entropy mode is expected to scale with system size as ${\tau }_{th}\sim L^z$, where the dynamical exponent $z$ reflects the dispersion of the anyonic excitations. In the $SU(3{)}_3$ category, the lightest non-confined excitation is the $\left(\mathrm{1,1}\right)$ glueball with quantum dimension $d=3$; its gap $\Delta \sim |J-J_c|$ near criticality provides a natural scale for ${\tau }_{th}$. This scaling law would provide a direct numerical test of the universality class of the transition.

Entropy production peak. At the moment of the phase transition, the entropy production rate $\sigma \left(t\right)$ is expected to display a peak. The peak position marks the time at which the system crosses the critical region, and its height reflects the amount of entropy generated during the non-equilibrium transition. The peak structure is a generic signature of non-equilibrium phase transitions and can be quantitatively compared with the static entropy change $\Delta S_{topo}\approx -0.549$.

Delayed response of topological entanglement entropy: because the topological entanglement entropy $S_{topo}$ undergoes a jump across the phase transition (computed in this paper as $\Delta S_{topo}\approx -0.549$), its evolution during the quench will be non-trivial. In particular, the settling of the non-local topological entanglement entropy $S_{topo}\left(t\right)$ to its new equilibrium value may be delayed relative to the thermalization of local observables, because topological order is a global property that requires correlations to propagate across the entire system, with a time delay ${\mathsf{\tau }}_{\mathrm{t}\mathrm{o}\mathrm{p}\mathrm{o}}>{\mathsf{\tau }}_{\mathrm{l}\mathrm{o}\mathrm{c}}$.

The numerical verification of these expectations--for example, by performing Runge-Kutta integration on a three- or four-state truncation of the $SU(3{)}_3$ category -- is left for future work. This non-equilibrium dynamics scheme promises to provide the first topological phase transition example in which entropy flow and thermalization can be tracked quantitatively within the entropy-enthalpy competition framework.

The connection between non-equilibrium topological dynamics and quantum computation has been systematically developed in the context of topological quantum computation [34,35].

6. Some Heuristic Insights from Anyon Condensation (Speculative)

We emphasize that the following remarks are speculative and presented only as conceptual motivation for possible future directions; they are not derived from the computation in this paper.

The anyon condensation process studied in this work suggests a broader perspective: the entropy-enthalpy competition framework may apply across multiple physical scales. In the present context, the condensation algebra $A=\left(\mathrm{0,0}\right)\oplus \left(\mathrm{3,0}\right)\oplus \left(\mathrm{0,3}\right)$ plays the role of “enthalpy”--it drives the system toward a more ordered, low-entropy phase by confining non-zero triality objects--while the decrease in topological entropy $\Delta S_{topo}\approx -0.549$ quantifies the “entropic cost” of this reorganization. This same logical structure--condensation → phase transition → emergence of a new, more ordered phase--may recur in other settings where fusion categories provide the underlying mathematical language.

One such setting is the string-net condensation model of Ref. [36], where continuous spacetime itself emerges from a quantum information network via a condensation transition. There, the competition between a cosmological enthalpy term and the topological entropy of the string-net condensate drives the emergence of spacetime geometry and gravitational dynamics. The algebraic isomorphism between string-net condensation and anyon condensation--both are described by fusion categories--suggests that the entropy-enthalpy competition mechanism identified here may be a generic feature of phase transitions in systems with topological order.

More generally, one may envision a hierarchy of such transitions: each condensation event freezes a subset of the available topological degrees of freedom, reducing the vacuum topological entropy and releasing thermodynamic entropy into the surrounding degrees of freedom. The topological entropy change $\Delta S_{topo}\approx -0.549$ computed in this work provides a concrete, exactly computable benchmark for such transitions. Whether similar condensation mechanisms operate in higher-dimensional settings—such as the confinement-deconfinement transition in QCD—remains an open question, but the algebraic correspondence established in Secs. 5.1 and 5.2 suggests that the categorical framework developed here may offer a useful starting point [27,37].

However, these extensions remain conjectural at the conceptual level; the only result rigorously established in this work is the topological entropy change $\Delta S_{topo}\approx -0.549$for the specific $SU(3{)}_3$ condensation. The broader picture is offered as a heuristic perspective--not as a conclusion of the present calculation.

7. Conclusion

Within the framework of the entropy-enthalpy theory, we have reinterpreted anyon condensation as a “transition of entropy modes.” Taking the $SU(3{)}_3$ modular tensor category as an instance, we have accomplished the following:

1)

Identified the bosonic anyon $\left(\mathrm{3,0}\right)$, constructed the condensation algebra $\mathrm{A}=\left(\mathrm{0,0}\right)\oplus \left(\mathrm{3,0}\right)\oplus \left(\mathrm{0,3}\right)$, and proved that its condensation confines all objects with non-zero triality (including quarks);

2)

Computed the topological entropy change across condensation, $\Delta \mathrm{S}\approx -0.549$, providing the first quantitative topological-phase-transition measure for entropy-enthalpy competition;

3)

Constructed an extended Levin-Wen model Hamiltonian $\mathrm{H}\left(\mathrm{J}\right)$and established the parameter $\mathrm{J}$ as the control parameter for “enthalpy”, and discussed the research direction of non-equilibrium entropy production based on quench dynamics;

4)

This work provides an exactly solvable topological phase transition example for the entropy-enthalpy theory, and supplies a categorical computational framework for understanding the structural aspects of QCD confinement;

5)

Stated the limitation that the present work is carried out in the 2+1-dimensional pure topological limit; the extension to higher dimensions and finite temperature has been discussed as future research directions.

This work provides the first exactly solvable topological phase transition example for the entropy-enthalpy theory, supplies a categorical computational framework for understanding QCD confinement.

Appendix A.1. Review of $\mathit{S}\mathit{U}(3{)}_3$ Data

The $SU(3{)}_3$ modular tensor category contains 10 simple objects labeled by Dynkin labels $(n,\bar{n})$ with $n+\bar{n}\le 3$. The relevant data, see Table A1.

Table A1. Object list and core data.

Label $(n,\bar{n})$	$SU\left(3\right)$ rep.	Triality $\tau$	Quantum dimension $d_{(n,\bar{n})}$ [10]	spin $h$
$\left(\mathrm{0,0}\right)$	$\mathbf{1}$	$0$	$\mathbf{1}$	$\mathbf{0}$
$\left(\mathrm{1,0}\right)$	$\mathbf{3}$	$1$	$2$	$4/15$
$\left(\mathrm{0,1}\right)$	$\overline{\mathbf{3}}$	$2$	$2$	$4/15$
$\left(\mathrm{2,0}\right)$	$\mathbf{6}$	$2$	$2$	$1/3$
$\left(\mathrm{0,2}\right)$	$\overline{6}$	$1$	$2$	$1/3$
$\left(\mathrm{1,1}\right)$	$\mathbf{8}$	$0$	$3$	$2/5$
$\left(\mathrm{3,0}\right)$	$\mathbf{10}$	$0$	$1$	$1$
$\left(\mathrm{0,3}\right)$	$\overline{\mathbf{10}}$	$0$	1	$1$
$\left(\mathrm{2,1}\right)$	$\mathbf{15}$	$1$	$2$	$2/3$
$\left(\mathrm{1,2}\right)$	$\overline{\mathbf{15}}$	$2$	2	$2/3$

Table A1. Object list and core data.

Label $(n,\bar{n})$	$SU\left(3\right)$ rep.	Triality $\tau$	Quantum dimension $d_{(n,\bar{n})}$ [10]	spin $h$
$\left(\mathrm{0,0}\right)$	$\mathbf{1}$	$0$	$\mathbf{1}$	$\mathbf{0}$
$\left(\mathrm{1,0}\right)$	$\mathbf{3}$	$1$	$2$	$4/15$
$\left(\mathrm{0,1}\right)$	$\overline{\mathbf{3}}$	$2$	$2$	$4/15$
$\left(\mathrm{2,0}\right)$	$\mathbf{6}$	$2$	$2$	$1/3$
$\left(\mathrm{0,2}\right)$	$\overline{6}$	$1$	$2$	$1/3$
$\left(\mathrm{1,1}\right)$	$\mathbf{8}$	$0$	$3$	$2/5$
$\left(\mathrm{3,0}\right)$	$\mathbf{10}$	$0$	$1$	$1$
$\left(\mathrm{0,3}\right)$	$\overline{\mathbf{10}}$	$0$	1	$1$
$\left(\mathrm{2,1}\right)$	$\mathbf{15}$	$1$	$2$	$2/3$
$\left(\mathrm{1,2}\right)$	$\overline{\mathbf{15}}$	$2$	2	$2/3$

The total quantum dimension is:

$${\mathcal{D}}_{\mathcal{C}}^2=\sum _{a\in \mathcal{C}}{d}_a^2=36,{\mathcal{D}}_{\mathcal{C}}=6$$

Key observation: $\left(\mathrm{3,0}\right)$ is a simple current ($d=1$, integer spin). It generates a ${\mathbb{Z}}_3$ subgroup:

$$⟨\left(\mathrm{3,0}\right)⟩=\left\{\right(\mathrm{0,0}),(\mathrm{3,0}),(\mathrm{0,3}\left)\right\}$$

Appendix A.2. Confirmation of the Condensation Algebra

The correct condensation algebra must be closed under fusion. Since $\left(\mathrm{3,0}\right)$ is a simple current of order three, its full orbit under fusion must be included:

$$\left(\mathrm{3,0}\right)\otimes \left(\mathrm{3,0}\right)=\left(\mathrm{0,3}\right),\left(\mathrm{3,0}\right)\otimes \left(\mathrm{0,3}\right)=\left(\mathrm{0,0}\right)$$

Therefore, the condensation algebra is:

$$A=\left(\mathrm{0,0}\right)\oplus \left(\mathrm{3,0}\right)\oplus \left(\mathrm{0,3}\right)$$

It is a legitimate connected étale algebra because:

1)

$\left(\mathrm{3,0}\right)$ and $\left(\mathrm{0,3}\right)$ are bosonic: $h_{\left(\mathrm{3,0}\right)}=h_{\left(\mathrm{0,3}\right)}=1\in \mathbb{Z}$

2)

The algebra is closed under fusion: $A\otimes A\supset A$

3)

The algebra is connected: $d_A=1+d_{\left(\mathrm{3,0}\right)}+d_{\left(\mathrm{0,3}\right)}$

Appendix A.3. Fusion Orbits Under the Simple Current $\left(\mathrm{3,0}\right)$

The action of $\left(\mathrm{3,0}\right)$ on all objects by fusion produces the following orbits, see Table A2:

Table A2. Fusion orbits under the simple current (3,0).

Orbit	Objects	Length	Triality
${\mathcal{O}}_1$	$\left(\mathrm{0,0}\right)$,$\left(\mathrm{3,0}\right)$,$\left(\mathrm{0,3}\right)$	3	0
${\mathcal{O}}_2$	$\left(\mathrm{1,0}\right)$,$\left(\mathrm{0,2}\right)$	2	1
${\mathcal{O}}_3$	$\left(\mathrm{0,1}\right)$,$\left(\mathrm{2,0}\right)$	2	2
${\mathcal{O}}_4$	$(2,1)$, $\left(\mathrm{1,2}\right)$	2	1 (for (2,1)), 2 (for (1,2))
${\mathcal{O}}_5$	$(1,1)$	1	0

Table A2. Fusion orbits under the simple current (3,0).

Orbit	Objects	Length	Triality
${\mathcal{O}}_1$	$\left(\mathrm{0,0}\right)$,$\left(\mathrm{3,0}\right)$,$\left(\mathrm{0,3}\right)$	3	0
${\mathcal{O}}_2$	$\left(\mathrm{1,0}\right)$,$\left(\mathrm{0,2}\right)$	2	1
${\mathcal{O}}_3$	$\left(\mathrm{0,1}\right)$,$\left(\mathrm{2,0}\right)$	2	2
${\mathcal{O}}_4$	$(2,1)$, $\left(\mathrm{1,2}\right)$	2	1 (for (2,1)), 2 (for (1,2))
${\mathcal{O}}_5$	$(1,1)$	1	0

Verification of key fusion rules:

$$\begin{gathered} \left(\mathrm{3,0}\right)\otimes \left(\mathrm{1,0}\right)=\left(\mathrm{0,2}\right) \\ \left(\mathrm{3,0}\right)\otimes \left(\mathrm{0,1}\right)=\left(\mathrm{2,0}\right) \\ \left(\mathrm{3,0}\right)\otimes \left(\mathrm{2,1}\right)=\left(\mathrm{1,2}\right) \\ \left(\mathrm{3,0}\right)\otimes \left(\mathrm{1,1}\right)=\left(\mathrm{1,1}\right) \end{gathered}$$

Appendix A.4. Confinement Criterion

For a general anyon condensation with algebra $A$, the definitive criterion for confinement is the mutual statistics condition. Recall from Sec. 3.2 that an object $X$ is confined if and only if there exists some $a\in A$ such that [5]

$$S_{X,a}\ne {\displaystyle \frac{d_X{d}_a}{\mathcal{D}}}$$

For our simple-current condensation algebra $A=\left(\mathrm{0,0}\right)\oplus \left(\mathrm{3,0}\right)\oplus \left(\mathrm{0,3}\right)$, this criterion can be applied directly. For any object $X$ with non-zero triality $\tau \left(X\right)\ne 0$, a direct computation using the $S$-matrix formula (2) gives

$$S_{X,\left(\mathrm{3,0}\right)}\ne {\displaystyle \frac{d_X{d}_{\left(\mathrm{3,0}\right)}}{\mathcal{D}}},$$

so $X$ satisfies the confinement criterion and is therefore confined. Conversely, for any object with $\tau \left(X\right)=0$, the local commutativity condition $S_{X,\left(\mathrm{3,0}\right)}=d_X{d}_{\left(\mathrm{3,0}\right)}/\mathcal{D}$ holds, so $X$ is not confined by $\left(\mathrm{3,0}\right)$ and may survive in the condensed phase. Thus the triality rule $\tau \left(X\right)\ne 0$ serves as a simple diagnostic for confinement, but its rigorous justification is the $S$-matrix condition (A.1).

Applying this criterion to the 10 simple objects of $SU(3{)}_3$ yields:

-

Confined: $\left(\mathrm{1,0}\right)$, $\left(\mathrm{0,1}\right)$, $\left(\mathrm{2,0}\right)$, $\left(\mathrm{0,2}\right)$, $\left(\mathrm{2,1}\right)$, $\left(\mathrm{1,2}\right)$--all have $\tau \ne 0$ and satisfy $S_{X,\left(\mathrm{3,0}\right)}\ne d_X{d}_{\left(\mathrm{3,0}\right)}/\mathcal{D}$.

-

Surviving (0,0), $\left(\mathrm{3,0}\right)$, $\left(\mathrm{0,3}\right)$, $\left(\mathrm{1,1}\right)$ -- all have $\tau =0$ and satisfy the local commutativity condition.

Among the surviving objects:

-

$\left(\mathrm{0,0}\right)$ is the original vacuum.

-

$\left(\mathrm{3,0}\right)$ and $\left(\mathrm{0,3}\right)$ are identified with the vacuum and merged into $\left[\right(\mathrm{0,0}\left)\right]$

-

$\left(\mathrm{1,1}\right)$ survives as an independent excitation, interpreted as a “glueball” in the condensed phase.

Remark on  $\left(\mathrm{2,1}\right)$ and  $\left(\mathrm{1,2}\right)$. Both objects carry non-zero triality ($\tau =1$ and $\tau =2$, respectively), and hence the triality rule immediately indicates confinement. To verify this rigorously via the $S$-matrix criterion, we compute using formula (2) with $k=3$:

$$S_{\left(\mathrm{2,1}\right),\left(\mathrm{3,0}\right)}=0,{\displaystyle \frac{d_{\left(\mathrm{2,1}\right)}{d}_{\left(\mathrm{3,0}\right)}}{\mathcal{D}}}={\displaystyle \frac{2\cdot 1}{6}}={\displaystyle \frac{1}{3}}$$

and

$$S_{\left(\mathrm{1,2}\right),\left(\mathrm{3,0}\right)}=-{\displaystyle \frac{4}{3}},{\displaystyle \frac{d_{\left(\mathrm{1,2}\right)}{d}_{\left(\mathrm{3,0}\right)}}{\mathcal{D}}}={\displaystyle \frac{2\cdot 1}{6}}={\displaystyle \frac{1}{3}}$$

In both cases, $S_{X,\left(\mathrm{3,0}\right)}\ne d_X{d}_{\left(\mathrm{3,0}\right)}/\mathcal{D}$, so the confinement criterion (A.1) is satisfied. Thus $\left(\mathrm{2,1}\right)$ and $\left(\mathrm{1,2}\right)$ are fully confined and do not contribute to the condensed category $\mathcal{D}$

Appendix A.5. Topological Entropy Change

The total quantum dimension of the condensed category follows from the standard condensation formula [5,16]:

$${\mathcal{D}}_{\mathcal{D}}={\displaystyle \frac{{\mathcal{D}}_{\mathcal{C}}}{\sqrt{d_A}}}={\displaystyle \frac{6}{\sqrt{3}}}=2\sqrt{3}$$

Thus:

$${\mathcal{D}}_{\mathcal{D}}^2=12$$

The topological entropy change is therefore:

$$\Delta S_{topo}={\ln \mathcal{D}}_{\mathcal{D}}-{\ln \mathcal{D}}_{\mathcal{C}}=\ln \left(2\sqrt{3}\right)-\ln 6=\ln {\displaystyle \frac{2\sqrt{3}}{6}}=\ln {\displaystyle \frac{\sqrt{3}}{3}}=-{\displaystyle \frac{1}{2}}\ln 3$$

$$\Delta S_{topo}=-{\displaystyle \frac{1}{2}}\ln 3\approx -0.549$$

This result is derived from the general theory of anyon condensation and does not depend on the detailed splitting of individual non-Abelian objects.

Appendix A.6. Summary

See Table A3.

Table A3. Summary table.

Quantity	Value
Condensation algebra	$\mathrm{A}=\left(\mathrm{0,0}\right)\oplus \left(\mathrm{3,0}\right)\oplus \left(\mathrm{0,3}\right)$
${\mathrm{d}}_{\mathrm{A}}$	3
Confined objects	$\left(\mathrm{1,0}\right),$$\left(\mathrm{0,1}\right),$$\left(\mathrm{2,0}\right),$$\left(\mathrm{0,2}\right),$$\left(\mathrm{2,1}\right),$$\left(\mathrm{1,2}\right)$
Surviving objects	$\left(\mathrm{0,0}\right),\left(\mathrm{3,0}\right),\left(\mathrm{0,3}\right),\left(\mathrm{1,1}\right)$
Condensed vacuum	$\left[\right(\mathrm{0,0}\left)\right]$from$\left(\mathrm{0,0}\right)\oplus \left(\mathrm{3,0}\right)\oplus \left(\mathrm{0,3}\right)$
Condensed glueball	$\left[\right(\mathrm{1,1}\left)\right]$with $\mathrm{d}=3$
${\mathcal{D}}_{\mathcal{C}}$	$6$
${\mathcal{D}}_{\mathcal{D}}$	$2\sqrt{3}\approx 3.464$
$\Delta {\mathrm{S}}_{\mathrm{t}\mathrm{o}\mathrm{p}\mathrm{o}}$	$-{\displaystyle \frac{1}{2}}\ln 3\approx -0.549$

Table A3. Summary table.

Quantity	Value
Condensation algebra	$\mathrm{A}=\left(\mathrm{0,0}\right)\oplus \left(\mathrm{3,0}\right)\oplus \left(\mathrm{0,3}\right)$
${\mathrm{d}}_{\mathrm{A}}$	3
Confined objects	$\left(\mathrm{1,0}\right),$$\left(\mathrm{0,1}\right),$$\left(\mathrm{2,0}\right),$$\left(\mathrm{0,2}\right),$$\left(\mathrm{2,1}\right),$$\left(\mathrm{1,2}\right)$
Surviving objects	$\left(\mathrm{0,0}\right),\left(\mathrm{3,0}\right),\left(\mathrm{0,3}\right),\left(\mathrm{1,1}\right)$
Condensed vacuum	$\left[\right(\mathrm{0,0}\left)\right]$from$\left(\mathrm{0,0}\right)\oplus \left(\mathrm{3,0}\right)\oplus \left(\mathrm{0,3}\right)$
Condensed glueball	$\left[\right(\mathrm{1,1}\left)\right]$with $\mathrm{d}=3$
${\mathcal{D}}_{\mathcal{C}}$	$6$
${\mathcal{D}}_{\mathcal{D}}$	$2\sqrt{3}\approx 3.464$
$\Delta {\mathrm{S}}_{\mathrm{t}\mathrm{o}\mathrm{p}\mathrm{o}}$	$-{\displaystyle \frac{1}{2}}\ln 3\approx -0.549$

Appendix A.7. Consistency with the Main Text

The value $\Delta S_{topo}=-{\displaystyle \frac{1}{2}}\ln 3\approx -0.549$ is the value used in Eq. (7) of the main text. It follows rigorously from the general condensation formula and does not require a complete vertex-lifting coefficient (VLC) computation. The qualitative conclusion—confinement of non-zero triality objects and a decrease of topological entropy--is robust.