Confining Strings and Glueballs in $\mathbb{Z}_N$ Gauge Theories
Extracted source text · arXiv 2312.03855
% Please make sure you insert your
% data according to the instructions in PoSauthmanual.pdf
\documentclass[a4paper,11pt]{article}
\usepackage{pos}
\usepackage{braket}
\usepackage{amsmath}
\usepackage{amssymb}
\usepackage{multicol}
\usepackage{slashed}
\usepackage{mathtools}
\usepackage{float}
\usepackage{dsfont}
\usepackage{caption}
\usepackage{subcaption}
\usepackage{lipsum}
\newcommand{\Tr}{{\rm Tr}}
\newcommand{\fixme}[1]{{\bf {\color{red}[#1]}}}
\let\lpol\l
\usepackage[utf8]{inputenc}\DeclareUnicodeCharacter{2212}{-}
\usepackage[normalem]{ulem}
%\let\OLDthebibliography\thebibliography
%\renewcommand\thebibliography[1]{
% \OLDthebibliography{#1}
% \setlength{\parskip}{0pt}
% \setlength{\itemsep}{0pt plus 0.3ex}
%}
\title{Confining Strings and Glueballs in $\mathbb{Z}_N$ Gauge Theories}
\newcommand{\be}{\begin{equation}}
\newcommand{\ee}{\end{equation}}
\newcommand{\bea}{\begin{eqnarray}}
\newcommand{\eea}{\end{eqnarray}}
\author[a]{Andreas Athenodorou}
\author[b]{Sergei Dubovsky}
\author*[b]{Conghuan Luo}
\author[c,d]{Michael~Teper}
\affiliation[a]{Computation-based Science and Technology Research Center, The Cyprus Institute, Cyprus}
\affiliation[b]{Center for Cosmology and Particle Physics, Department of Physics, New York University, New York, NY, 10003, USA}
\affiliation[c]{Rudolf Peierls Centre for Theoretical Physics, University of Oxford, Parks Road, Oxford OX1 3PU, UK}
\affiliation[d]{All Souls College, University of Oxford, High Street, Oxford OX1 4AL, UK}
\abstract{ Effective string theory has shown its universal power in the prediction of the spectrum of low-lying excited states of confining strings. Here we study confining flux tubes in $\mathbb{Z}_N$ gauge theories. For the $N=2$ theory, which corresponds to the 3d Ising gauge model, we compute the spectrum of low-lying excitations of confining strings and show that it agrees with the universal Nambu--Goto predictions except for an additional massive scalar resonance. This resonance, however, turns out to be a bulk glueball mixing with the flux tube excitations rather than a genuine string worldsheet state. In general $\mathbb{Z}_N$ gauge theories (dual to clock spin models), we observe a continuous phase transition for $N \geq 4$, while for $N > 5$ it is governed by the $O(2)$ universality class. The critical behavior of the string tension and mass gap is verified to be described by a dangerously irrelevant operator. At large $N$ the glueball spectrum is expected to approach the spectrum of U(1) gauge theory, which is confirmed by our lattice data.}
\FullConference{%
The 40th International Symposium on Lattice Field Theory (Lattice2023),\\
31 July - 4 August, 2023 \\
Fermilab, Illinois, USA
}
%% \tableofcontents
\begin{document}
\maketitle
% *********************************************
% * Introduction
% *********************************************
\section{Introduction}
\label{sec:introduction}
Understanding the color confinement in Quantum Chromodynamics (QCD) has been a long-standing problem ever since early 1970s. Simply put, this puzzle can be phrased as the question of how does a non-Abelian gauge theory, that describes colored quarks and gluons at high energies, produce a spectrum of colorless hadrons at long distances? The concept of confinement becomes precise, for theories like pure glue $SU(N_c)$ Yang--Mills model which enjoy the $\mathbb{Z}_N$ one-form electric ``center" symmetry. The order parameter for the center symmetry is the vacuum expectation value (VEV) of Polyakov loops, which translates into the area law for the Wilson loops $\langle W(C) \rangle \sim e^{-\text{Area(C)}/\ell_s^2}$. This implies the existence of stable long strings of a finite tension \footnote{See, e.g. \cite{Gaiotto:2014kfa,McGreevy:2022oyu} for reviews on higher-form symmetries. } $T=\ell_s^{-2}$. This notion can be generalized to any theory with a one-form symmetry, such as 3D $\mathbb{Z}_N$ gauge theories, which are the main focus of this talk. Studying the behavior of such confining strings may shed light on our understanding of confinement.
Recently, there has been a revival of research on QCD and confinement in terms of effective string theory. This is stimulated in part by the development of high precision Monte-Carlo computations of confining string spectrum for different theories \cite{Athenodorou:2010cs,Athenodorou:2011rx}. %This progress enables the determination of higher order coefficients in the effective string theory.
The spectrum of strings wrapped around a spatial circle of circumference $R$ can be approximated using the derivative expansion based on the effective string theory.
%Consider a straight long string configuration in the spacetime which would not break, it spontaneously breaks the D-dimensional spacetime Poincar\'e symmetry $ISO(1, D-1) \to ISO(1,1) \times O(D-2)$, and the fluctuations along the strings can be described by $D-2$ Goldstone bosons $X^i$. One can write down the most general low energy effective worldsheet action that preserves reparametrization invariance and D-dimensional target spacetime Poincar\'e symmetry:
%\begin{equation}
% S = -\int d^2 \sigma \sqrt{-\det h_{\alpha\beta}} \left[ \ell_s^{-2} + \mathcal{R} + a_2 K^2 + b_2 K_{\alpha\beta}^{\mu} K^{\alpha\beta}_{\mu} + O(\ell_s^2) \right] \,,
%\end{equation}
%where $\ell_s^{-2}$ is the string tension, and $K_{\alpha\beta}^{\mu}$ is the extrinsic curvature tensor and we will in the following assume static gauge $X^{\mu} = (\sigma^{\alpha}, X^i)$. Ricci scalar $\mathcal{R}$ is topological in two dimensional spacetime, and both $K^2$ terms can be eliminated using equation of motions and Gauss-Codazzi equations. As a result, the low energy behavior is governed by the Nambu-Goto action, and the first non-universal term appears at the level of $O(K^4)$.
Assuming that translational Goldstone modes are the only massless degree of freedom on the worldsheet, the Nambu--Goto action fixes the form of the string energy levels up to subleading non-universal corrections which start at the $O(\ell_s^6/R^7)$ order in the derivative expansion~\cite{Aharony:2013ipa,Dubovsky:2012sh}.
On the other hand, for a while it was observed that the following expression
\begin{equation}
E_{\text{GGRT}}(N_L, N_R)=\sqrt{\frac{4 \pi^2(N_L - N_R)^2}{R^2}+\frac{R^2}{\ell_s^4}+\frac{4 \pi}{\ell_s^2}\left(N_L + N_R-\frac{D-2}{12}\right)} \,,
\label{GGRT}
\end{equation}
(dubbed the Goddard–Goldstone–Rebbi–Thorn (GGRT) formula) provides an exceptionally accurate approximation for the energies of many low-lying string states, even when the strings are relatively short and the $\ell_s/R$ expansion fails. Originally this formula was obtained by the light-cone quantization of the Nambu-Goto action~\cite{goddard1973quantum,arvis1983exactqq}. However, this quantization is known to be incompatible with the spacetime Poincar\'e symmetry. This puzzle has been resolved when it was realized that (\ref{GGRT}) arises as the leading order answer in the Thermodynamic Bethe Ansatz (TBA) method of calculating the string spectrum~\cite{Dubovsky:2012wk,Dubovsky:2013gi,Dubovsky:2014fma}, which has better convergence properties than the straightforward $\ell_s/R$ expansion.
The resulting improved theoretical control provides an opportunity to search for novel non-universal string excitations associated with the states which strongly deviate from (\ref{GGRT}). For example, the 4d Yang-Mills confining string is found to support an extra massive pseudo-scalar mode ~\cite{Dubovsky:2013gi,Dubovsky:2014fma}.
It is natural to ask whether massive resonances are also present on the worldsheet of $\mathbb{Z}_N$ strings. Here we present the results for the Ising model, which corresponds to $N=2$. As we will see, a new subtlety arises for $\mathbb{Z}_N$ strings. Namely, bulk glueballs may mix with the genuine string excitations and mimic the presence of a resonance.
In the Yang-Mills theory such a mixing is suppressed at large $N_c$. No such suppression is present in the $\mathbb{Z}_N$ case so we need to find a way to distinguish bulk glueballs from genuine worldsheet states.
Another motivation for studying $\mathbb{Z}_N$ gauge theories is to understand how the $U(1)$ gauge theory is recovered in the $N \to \infty$ limit, where confinement comes out as a result of monopole condensation \cite{polyakov1975compact,polyakov1977quark}. In particular, what is the mechanism for the confinement in $\mathbb{Z}_N$ gauge theories, and why deconfining transitions are only present for $\mathbb{Z}_N$ gauge theories, unlike in the $U(1)$ case?
The rest of this note is organized as follows. We sketch our methodology of lattice simulations in section~\ref{sec:lattice}. Then we review the results of 3d Ising strings in section~\ref{sec:Ising} based on \cite{Athenodorou:2022pmz}. In section~\ref{sec:zn} we summarize some new results for $\mathbb{Z}_N$ gauge theories. The details of the data and analysis will be presented in a forthcoming publication \cite{zngaugetheory}.
\section{Lattice setup}
\label{sec:lattice}
We use the Wilson lattice gauge theory action
\begin{equation}
S = \beta \sum_{\rm plaq} \{1- \operatorname{Re} %\operatorname{Tr}
U_{\rm plaq}\} \,, \quad U_{\rm plaq}(n,\mu,\nu) = U_{\mu}(n) \cdot U_{\nu}(n+\hat{\mu}) \cdot U_{\mu}^{\dagger}(n+\hat{\nu}) \cdot U_{\nu}^{\dagger}(n)\;,
\end{equation}
where the link fields take values $U_{\mu}(n) = e^{\frac{2\pi i k}{N}}, \quad k \in \mathbb{Z}/N\mathbb{Z}$. This is called the (vector) $\mathbb{Z}_N$ gauge theory. The simulation is performed using the Metropolis algorithm. We typically make over $10^5$ measurements for each lattice system and coupling.
We compute the low-lying spectrum of confining strings and glueballs from two point functions of operators that represents the physical objects we are interested in, located at different times. We use Polyakov operators that wind around the spatial dimension once to represent confining strings, and contractable Wilson loops to represent glueballs. The excited states are extracted by constructing a large set of operators, and using variational methods to find out linear combinations of operators that have large overlap with the low-lying states. Smearing and blocking methods are applied. Further details of the simulation can be found in \cite{Athenodorou:2010cs,Athenodorou:2022pmz,Teper:1998te}.
Operators are characterized by a set of quantum numbers determined by their symmetry properties. Operators from different sectors have zero two point functions. For winding strings, the quantum numbers include the longitudinal momentum $p = 2\pi q/R$ (the transverse momentum is fixed to zero in our simulations), the transverse parity $P_{\perp}$, that flips the sign of the transverse spatial coordinate, and the longitudinal parity $P_{\parallel}$ that flips the sign of the longitudinal coordinate\footnote{This latter symmetry corresponds to the $CP$ symmetry of the underlying gauge theory, because it reverses the flux direction.}. Note that $P_{\parallel}$ is not a good quantum number for states with $q \neq 0$. Correspondingly, we label the string states either by $(P_{\perp} P_{\parallel})$ or by $q \neq 0$ and $(P_{\perp})$.
On the other hand, glueball states are characterized by their spin $J$, parity $P$ and charge conjugation $C$ (not present for Ising), and we label the corresponding sectors as $|J|^{PC}$.
\section{3d Ising strings}
\label{sec:Ising}
The Kramers-Wannier duality states that the 3d $\mathbb{Z}_2$ lattice gauge theory (LGT) has the same partition function as the Ising model with no external magnetic field. This duality exchanges the low temperature phase of one model to the high temperature phase of another, and vice versa \cite{savit1980duality}. This implies that the critical behavior of the second order deconfining transition in the $\mathbb{Z}_2$ gauge theory is governed by 3D Ising universality class.
We focus on the confining phase and take the continuum limit of the gauge theory by sending $\beta$ to $\beta_c = 0.7614133(22)$ \cite{blote1999cluster}, where the string tension and the mass gap are small in the lattice unit.
%and more suited for precise measurements of excited %string spectrum.
We find that the lowest-lying states of wound flux tubes agree with the GGRT prediction~\eqref{GGRT}, with some notable exceptions. Specifically, we observe a massive resonance in the $(++)$ sector with $m_{res}\ell_s = 3.825(50)$, shown in Figure~\ref{plotparityplus}. A similar state shows up as a boosted massive resonance in the $q=1;(+)$ sector.
\begin{figure}[hbt]
\centering
\scalebox{1}{\includegraphics[width=0.6\textwidth]{parityplus.pdf}\put(-280,160){\scriptsize $\Delta E\ell_s$}\put(0,0){\scriptsize $R/\ell_s$}}
\caption{$q=0$ excited states in the $(++)$ parity sector as a function of string length. Blue curves are the $(1,1)$ and $(2,2)$ GGRT spectrum. The horizontal red line is the fitted mass for the massive resonance state. }
\label{plotparityplus}
\end{figure}
To determine whether this state is a genuine massive worldsheet resonance or a bulk excitation (glueball), we note that a bulk excitation can have transverse momentum relative to the string. So if this state is a glueball, in principle we should also be able to observe a series of scattering states of a glueball and a string, with quantized relative momentum $p_{\perp} = 2\pi q_{\perp}/l_{\perp}$, but the operators we use may have poor overlap onto them. To probe these state we added ``multitrace" operators of the following form
\begin{equation}
\phi_{\text {scattering }}=\sum_{n, m=1}^{l_{\perp} / a} \phi_P(y+n a) \phi_G(y+m a) e^{\frac{2 \pi i q_{\perp}(n-m) a}{l_{\perp}}} \,,
\end{equation}
where $\phi_P$ represents a string state, and $\phi_G$ is a glueball. Our initial simulation, which suggested the presence of a resonance, was performed without these operators and at fixed $l_{\perp}\approx 4.8\ell_s$.
In order to identify scattering states, recall that $q_{\perp}\neq 0$ scattering states are expected to exhibit a strong transverse volume dependence corresponding to the dispersion relation,
\begin{equation}
E=\sqrt{m_{\text {flux }}^2+p_{\perp}^2}+\sqrt{m_{\text {glue }}^2+p_{\perp}^2}, \quad p_{\perp}=2 \pi q_{\perp} / \ell_{\perp} \,.
\end{equation}
We present the transverse volume dependence of the $(++)$ sector after including scattering operators up to $q_{\perp}=4$ in Figure~\ref{plot_ppvolume_5glue}.
\begin{figure}[hbt]
\scalebox{1}{\includegraphics[width=0.6\textwidth]{ppvolume_5glues_x55.pdf}\put(-280,160){\scriptsize $E \ell_s$}\put(0,0){\scriptsize $\ell_s/l_{\perp}$}}
\centering
\caption{Spectrum in the $(++)$ sector at $R=55a=3.80l_s$ as a function of the inverse transverse volume with scattering operators. The GGRT spectrum is represented by horizontal solid lines of different colors
starting from $N=\tilde{N}=0$.
The energy of the absolute ground state plus the glueball mass is shown by the lower dashed blue line. The absolute ground state plus the resonance mass is shown by the upper dashed blue line.}
\label{plot_ppvolume_5glue}
\end{figure}
We observe all four scattering states with $q_{\perp} \neq 0$ as expected, but no new $q_\perp=0$ state in addition to the ``resonant" state. The conclusion is that the resonance is in fact a $q_\perp=0$ glueball, rather than a genuine massive resonance on the string.
\section{3d $\mathbb{Z}_N$ gauge theories}
\label{sec:zn}
The Kramers-Wannier duality can be extended to the duality between generalized $\mathbb{Z}_N$ clock spin model and the (vector) $\mathbb{Z}_N$ LGT
%\footnote{Note that this duality is not exact in this case, but the difference is expected to be at percent order. }
\cite{Borisenko:2012nf,ukawa1980dual}. It has been suggested that clock spin models with $N \geq 5$ exhibit a second order phase transition governed by the $O(2)$ universality class \cite{hove2003criticality,Borisenko:2013xna}.
Observables sensitive to the $U(1)\to\mathbb{Z}_N$ breaking typically exhibit an intricate scaling behavior associated with a dangerously irrelevant deformation \cite{oshikawa2000ordered}. The duality dictates then that there should be a deconfining transition in the gauge theory, which is also governed by the $O(2)$ universality class. However, so far relatively little is known about the confining mechanism of the $\mathbb{Z}_N$ LGT and how it approaches the
$U(1)$ limit as $N \to \infty$.
We aim to fill this gap here.
% \subsection{(De)confinement mechanism}
% The confining behavior is associated with the area law for the vev of the Wilson loop, $\langle W(C) \rangle$. In the $U(1)$ gauge theory, the area law is known to be given by the monopole condensation \cite{polyakov1975compact,polyakov1977quark}. We argue that for $N \geq 5$ there are two stable configurations in $\mathbb{Z}_N$ gauge theory which play a role in the (de)confinement: $\mathbb{Z}_N$ junctions and magnetic strings. A magnetic string loop of a charge $k$ mod $N$ can be generated from the vacuum by multiplying the value of the link fields by $e^{\frac{2\pi i k}{N}}$ which is intercepted by the surface encircled by the string loop. The action of this configuration is proportional to the length of the string: $\Delta S_s(l) = \beta [1 - \cos(2\pi/N)] l$. This is in analogy with the vortex string in $U(1)$ gauge theory, except that here the magnetic flux can only be defined through the holonomy of a Wilson loop $\oint_C \vec{A} \cdot d\vec{x}$ (mod $2\pi$), where $C$ has linking number one with the magnetic string loop\footnote{In the setup of lattice gauge theory, the integral transforms into the sum over the exponents of the link fields. }. A $\mathbb{Z}_N$ junction originates from the fact that a $-(N-1)$ magnetic string is indistinguishable from a $1$ magnetic string at the action level, so we can think of a configuration where all magnetic strings emitted from one point have their holonomies summed to $2\pi$. Also note that a $|k|$ magnetic string with $k < N/2$ tends to split into strings with smaller charges, as one can check by exploiting the convexity of $\cos$ function. Therefore, a junction configuration where all the magnetic strings emitted from it bifurcate as soon as possible is a local minimum of the action. This is in analogy to the $U(1)$ monopole instanton.
% The deconfining transition (which is missing in $U(1)$ gauge theory) is triggered by the following mechanism. The density of states of a magnetic string loop also grows exponentially with its length $l$ (in the leading large $l$ limit), so the contribution to the partition function becomes:
% \begin{equation}
% Z_{\text{string}} \approx e^{\gamma l - \beta \left[1 - \cos\left( \frac{2\pi}{N} \right) \right] l} \,,
% \end{equation}
% where $\gamma$ is a constant. Its specific value is not important for our argument. We then see that when the coupling grows larger than $\beta_c = \gamma / [1 - \cos(2\pi/N)]$, the magnetic strings are constrained to be small closed loops. The resulting dipole forces provide the perimeter law for the Wilson loops. On the other hand, when $\beta < \beta_c$, we expect the dominating configurations to be the screening gas of $\mathbb{Z}_N$ junctions and anti-junctions, which shall provide the area law of Wilson loops similarly in $U(1)$ gauge theory. This prediction of deconfining point agrees well with the lattice results \cite{bhanot1980phase}. As $N \to \infty$, the deconfining point is pushed to infinity.
% When $N$ is large, one might not be able to see the discretization effect of the holonomies if one is too close the $\mathbb{Z}_N$ junction. More precisely, one will start to see separated charge 1 magnetic string at the distance $l_T/a \simeq \sqrt{N/4\pi}$. As a result, one cannot see the difference between $\mathbb{Z}_N$ and $U(1)$ on a lattice when $l_D \ll l_T$, where $l_D$ is the screening length and represents the bulk energy scale. Also, we know that $l_D$ decreases power-like as one varies $\beta$ away from $\beta_c$, so when $\beta$ is far away from $\beta_c$, we expect to recover the spectrum of $U(1)$ lattice gauge theory. This is confirmed by our lattice results, shown in Table~\ref{table:recover_u1}.
% \begin{table}[htbp]
% \centering
% \begin{tabular}{c|cc|ccc}
% \hline \hline \multicolumn{6}{c}{$Z(N)$ and $U(1)$ lightest masses and string tension } \\
% \hline \hline group & $\beta$ & $L_s^2 L_t$ & $a M_{0^{++}}$ & $a M_{0^{--}}$ & $a \sqrt{\sigma} $ \\
% \hline
% $U(1)$ & 2.20 & $48^3$ & $0.5386(23)$ & $0.2691(14)$ & $0.16646(62)$ \\
% $Z(100)$ & 2.20 & $48^3$ & $0.5320(23)$ & $0.2648(12)$ & $0.16683(50)$ \\
% $Z(10)$ & 2.20 & $48^3$ & $0.5367(23)$ & $0.2673(17)$ & $0.16488(39)$ \\
% $Z(8)$ & 2.20 & $48^3$ & $0.5267(92)$ & $0.2644(17)$ & $0.16469(76)$ \\
% $Z(6)$ & 2.20 & $48^3$ & $0.451(9)$ & $0.2167(18)$ & $0.14252(51)$ \\
% \hline
% \end{tabular}
% \caption{The recovery of $U(1)$ spectrum as $N$ gets larger. }
% \label{table:recover_u1}
% \end{table}
\subsection{(De)confinement mechanism}
The confining behavior is associated with the area law for the vev of the Wilson loop, $\langle W(C) \rangle$. In the $U(1)$ case the partition function can be rewritten as a Coulomb gas of magnetic monopoles \cite{polyakov1975compact,polyakov1977quark}. The area law arises as a result of
the (magnetic) Debye screening of the magnetic field sourced by the external current (the Wilson loop).
This behavior persists at all values of the coupling constant $\beta$, which controls the monopole number density.
One expects the monopoles to be confined in the $\mathbb{Z}_N$ case. One way to argue for this is to realize the $\mathbb{Z}_N$ theory as arising from the spontaneous symmetry breaking $U(1)\to \mathbb{Z}_N$, where the monopole confinement arises as a consequence of the conventional (not the dual one) Meissner effect.
Directly in the $\mathbb{Z}_N$ language one observes that the non-vanishing magnetic flux through a plaquette in the $\mathbb{Z}_N$ model cannot be smaller than $2\pi/N$. Hence, an individual $U(1)$ monopole can be approximate by a $\mathbb{Z}_N$ gauge field configuration only up to a distance of order
$l_T \simeq \sqrt{N/4\pi} a$.
At longer distances a spherically symmetric $U(1)$ monopole turns into a $\mathbb{Z}_N$ junction of $N$ magnetic strings carrying the minimal magnetic flux. To determine the tension of such a string note that a magnetic string loop can be generated from the vacuum configuration by setting all links intersecting a surface bounded by the loop to $e^{\frac{2\pi i }{N}}$. The action of this configuration is proportional to the length $l$ of the loop
\begin{equation}
\Delta S_s(l) = \beta [1 - \cos(2\pi/N)] l\;.
\end{equation}
These considerations suggest the following dynamics of $\mathbb{Z}_N$ models at small values of $\beta$. First, there is a "$U(1)$ region" at the smallest values of $\beta$,
$0<\beta\lesssim\beta_N$, such that the typical monopole separation in the $U(1)$ model is smaller than $l_T$. Here one expects the $U(1)$ and $\mathbb{Z}_N$ spectra to be practically indistinguishable, because in both cases the dynamics is dominated by $U(1)$-like monopole configurations. At larger values of $\beta$, $\beta_N\lesssim\beta<\beta_c$ the model is expected to exhibit $\mathbb{Z}_N$ confinement, where instead of the monopole plasma the Wilson line is screened by a network of $\mathbb{Z}_N$ junctions connected by magnetic strings.
The critical value $\beta_c$, where the theory deconfines can be estimated by noting that the density of states of a large magnetic string loop
grows exponentially with its length $l$, so that the corresponding contribution to the partition function can be estimated as
\begin{equation}
Z_{\text{string}} \approx e^{\gamma l - \beta \left[1 - \cos\left( \frac{2\pi}{N} \right) \right] l} \,,
\end{equation}
where $\gamma$ is a constant. Its specific value is not important for the argument. We then expect that the string network disappears and the theory deconfines at
\begin{equation}
\label{betac}
\beta_c = \frac{\gamma}{1 - \cos(2\pi/N)} \,.
\end{equation}
This functonal form of the deconfining point agrees well with the lattice results of \cite{bhanot1980phase},
which can be fitted by (\ref{betac}) withh $\gamma\approx 1.5$ (which also reproduces the known values of $\beta_c$ at $N\leq 4$).
As $N \to \infty$, the deconfining point is pushed to infinity.
To confirm the existence of the small $\beta$ $U(1)$ region we preformed lattice simulations which determine the mass spectrum and the string tension at the fixed relatively small $\beta$ and different $N$.
The results are presented in Table~\ref{table:recover_u1}, which indeed indicate that at $\beta=2.20$ the spectrum of $\mathbb{Z}_N$ models with $N\geq 8$ is indeed very close to the $U(1)$ spectrum. On the other hand, the $\mathbb{Z}_6$ theory at $\beta=2.20$ appears to exhibit the $\mathbb{Z}_N$ confinement.
\begin{table}[htbp]
\centering
\begin{tabular}{c|cc|ccc}
\hline \hline \multicolumn{6}{c}{$Z(N)$ and $U(1)$ lightest masses and string tension } \\
\hline \hline group & $\beta$ & $L_s^2 L_t$ & $a M_{0^{++}}$ & $a M_{0^{--}}$ & $a \sqrt{\sigma} $ \\
\hline
$U(1)$ & 2.20 & $48^3$ & $0.5386(23)$ & $0.2691(14)$ & $0.16646(62)$ \\
$Z(100)$ & 2.20 & $48^3$ & $0.5320(23)$ & $0.2648(12)$ & $0.16683(50)$ \\
$Z(10)$ & 2.20 & $48^3$ & $0.5367(23)$ & $0.2673(17)$ & $0.16488(39)$ \\
$Z(8)$ & 2.20 & $48^3$ & $0.5267(92)$ & $0.2644(17)$ & $0.16469(76)$ \\
$Z(6)$ & 2.20 & $48^3$ & $0.451(9)$ & $0.2167(18)$ & $0.14252(51)$ \\
\hline
\end{tabular}
\caption{The recovery of $U(1)$ spectrum as $N$ gets larger. }
\label{table:recover_u1}
\end{table}
\subsection{Critical behavior}
Now we turn to the critical behavior of observables in $\mathbb{Z}_N$ gauge theories. In particular, we focus on the bulk mass gap and the string tension. The question is how to relate the physical quantities we measure to the well-known CFT data in the $O(2)$ CFT. One might expect the critical exponents to be determined by the dimensions of the relevant operators, but this model provides an example to this behavior, associated with dangerously irrelevant operators.
Let us present an renormalization group (RG) flow argument that prescribes the critical behavior. Consider the UV Gaussian fixed point ($u=g=\lambda_N=0$) in the complex scalar field theory,
\begin{equation}
S = \int d^3 x \left[ |\partial_{\mu} \Phi|^2 + u |\Phi|^2 + g |\Phi|^4 + \lambda_N (\Phi^N + \bar{\Phi}^N) \right] \,.
\end{equation}
If one tunes $u=u_c=0$ and inspects a family of theories with different $g>0$, one finds the critical value $g_c$ such that the theory runs into the $O(2)$ IR fixed point. Perturbing this RG trajectory fixed point by introducing $u>0$, one triggers an RG flow which starts at the Gaussian UV fixed point, approaches the vicinity of the $O(2)$ fixed point and then deviates into the Nambu--Goldstone theory,
\begin{equation}
S = \int d^3 x K (\partial_{\mu} \theta)^2 + c\lambda_N K^3 \cos N\theta \,.
\label{NG}
\end{equation}
The latter describes the (pseudo)-Goldstone mode of the $O(2)$ spontaneous symmetry breaking.
Here we introduced also the leading $O(2)$ breaking operator, which is generated if one perturbs the Gaussian theory by turning on $\lambda_N\neq 0$, with $c$ being a dimensionless constant.
Interestingly, this symmetry breaking term is irrelevant both at the Gaussian and the $O(2)$ fixed points, but becomes relevant at the NG fixed point, providing an example of a dangerously irrelevant behavior.
It was proposed in \cite{oshikawa2000ordered} that the confining $\mathbb{Z}_N$ theory near criticality is described exactly by this RG flow.
% Imagine a scenario: one start with a finite $\mathbb{Z}_N$ anisotropic $\lambda_n$ term, and tune a bit away from criticality. Along the RG flow, it goes past the $O(2)$ fixed point and towards the NG fixed point, where the $\lambda_n$ explicitly breaks the $O(2)$ symmetry and generates mass and interactions for the Goldstone mode.
The critical behavior is then governed by the charge $N$ operator as well as the typical length scale $K$ in the $O(2)$ CFT,
\begin{equation}
c\lambda_N \sim (T-T_c)^{\nu |y_N|}, \quad K \sim 1/\xi \sim (T-T_c)^{\nu} \,,
\end{equation}
where $\nu = 0.6718(1)$~\cite{Chester:2019ifh}, $|y_5| = 1.27(1)$, $|y_6|=2.55(6)$~\cite{shao2020monte}. As a consequence of the dangerously irrelevant behavior, we obtain a very weakly coupled massive theory in the deep IR. This is supported by the lattice simulations, where $M_{0^{++}} \approx 2 M_{0^{--}}$ for $\mathbb{Z}_N$ gauge theories.
The scaling of the mass gap and the string tension is then determined by~\eqref{NG},
\begin{equation}
m^2 = \lambda_N K^2 \sim (T - T_c)^{\nu (|y_N| + 2)}, \quad \sigma \propto Km \sim (T - T_c)^{\nu (\frac{|y_N|}{2} + 2)} \,.
\end{equation}
For $N=5$, we verified these scalings quite precisely,
\begin{equation}
a M_{0^{--}}(\beta) = 0.94(13) (2.17961 - \beta)^{1.09(5)}, \quad a^2 \sigma(\beta) = 0.183(8) (2.17961 - \beta)^{1.725(20)} \,.
\end{equation}
For $N=6$, our results are less accurate and we are not able to approach sufficiently close to the critical point, because the mass gap is much smaller in lattice unit, and lattice simulations suffer from severe finite volume effects. Nevertheless, we still get a qualitative agreement (within $3\sigma$),
\begin{equation}
a M_{0^{--}}(\beta) = 0.26(3) (3.0068 - \beta)^{1.90(12)}, \quad a^2 \sigma (\beta) = 0.0212(21) (3.0068 - \beta)^{2.00(11)}
\end{equation}
\section{Conclusion}
In this proceeding, we summarize our simulation results of the 3d Ising string spectrum, and introduce a method to probe scattering states of a glueball and a string, with which we have identified the only "resonance" state we have observed is a glueball mixed with the string, instead of a genuine massive resonance. We also explain the (De)confinement mechanism and the critical behaviors of 3d $\mathbb{Z}_N$ gauge theories and present simulational evidence supporting them.
% *********************************************
% * Simulation details
% *********************************************
% \section{Conclusions}
% \label{sec:conclusions}
% In this proceedings we presented a summary of our study of the string spectrum in the Ising gauge model \cite{Athenodorou:2022pmz} as well the main results of our forthcoming work, which studies the confinement and critical behavior of vector $\mathbb{Z}_N$ gauge theories \cite{zngaugetheory}.
\section*{Acknowledgements}
We thank John Donahue, Yifan Wang for fruitful discussions. AA has been financially
supported by the EuroCC2 project funded by the Deputy Ministry of Research, Innovation and
Digital Policy and the Cyprus Research and Innovation Foundation and the European High-
Performance Computing Joint Undertaking (JU) under grant agreement No 101101903. SD and MT
acknowledge support by the Simons Collaboration on Confinement and QCD Strings. SD is also supported in part by the NSF grant PHY-2210349, and by the BSF grant 2018068.
\bibliographystyle{JHEP}
\bibliography{biblio_NEW}
\end{document}
\providecommand{\href}[2]{#2}\begingroup\raggedright\begin{thebibliography}{10}
\bibitem{Gaiotto:2014kfa}
D.~Gaiotto, A.~Kapustin, N.~Seiberg and B.~Willett, \emph{{Generalized Global
Symmetries}}, \href{https://doi.org/10.1007/JHEP02(2015)172}{\emph{JHEP}
{\bfseries 02} (2015) 172} [\href{https://arxiv.org/abs/1412.5148}{{\ttfamily
1412.5148}}].
\bibitem{McGreevy:2022oyu}
J.~McGreevy, \emph{{Generalized Symmetries in Condensed Matter}},
\href{https://arxiv.org/abs/2204.03045}{{\ttfamily 2204.03045}}.
\bibitem{Athenodorou:2010cs}
A.~Athenodorou, B.~Bringoltz and M.~Teper, \emph{{Closed flux tubes and their
string description in D=3+1 SU(N) gauge theories}},
\href{https://doi.org/10.1007/JHEP02(2011)030}{\emph{JHEP} {\bfseries 02}
(2011) 030} [\href{https://arxiv.org/abs/1007.4720}{{\ttfamily 1007.4720}}].
\bibitem{Athenodorou:2011rx}
A.~Athenodorou, B.~Bringoltz and M.~Teper, \emph{{Closed flux tubes and their
string description in D=2+1 SU(N) gauge theories}},
\href{https://doi.org/10.1007/JHEP05(2011)042}{\emph{JHEP} {\bfseries 05}
(2011) 042} [\href{https://arxiv.org/abs/1103.5854}{{\ttfamily 1103.5854}}].
\bibitem{Aharony:2013ipa}
O.~Aharony and Z.~Komargodski, \emph{{The Effective Theory of Long Strings}},
\href{https://doi.org/10.1007/JHEP05(2013)118}{\emph{JHEP} {\bfseries 05}
(2013) 118} [\href{https://arxiv.org/abs/1302.6257}{{\ttfamily 1302.6257}}].
\bibitem{Dubovsky:2012sh}
S.~Dubovsky, R.~Flauger and V.~Gorbenko, \emph{{Effective String Theory
Revisited}}, \href{https://doi.org/10.1007/JHEP09(2012)044}{\emph{JHEP}
{\bfseries 09} (2012) 044} [\href{https://arxiv.org/abs/1203.1054}{{\ttfamily
1203.1054}}].
\bibitem{goddard1973quantum}
P.~Goddard, J.~Goldstone, C.~Rebbi and C.B.~Thorn, \emph{Quantum dynamics of a
massless relativistic string}, {\emph{Nuclear Physics B} {\bfseries 56}
(1973) 109}.
\bibitem{arvis1983exactqq}
J.~Arvis, \emph{The exactqq{\={}} potential in nambu string theory},
{\emph{Physics Letters B} {\bfseries 127} (1983) 106}.
\bibitem{Dubovsky:2012wk}
S.~Dubovsky, R.~Flauger and V.~Gorbenko, \emph{{Solving the Simplest Theory of
Quantum Gravity}}, \href{https://doi.org/10.1007/JHEP09(2012)133}{\emph{JHEP}
{\bfseries 09} (2012) 133} [\href{https://arxiv.org/abs/1205.6805}{{\ttfamily
1205.6805}}].
\bibitem{Dubovsky:2013gi}
S.~Dubovsky, R.~Flauger and V.~Gorbenko, \emph{{Evidence from Lattice Data for
a New Particle on the Worldsheet of the QCD Flux Tube}},
\href{https://doi.org/10.1103/PhysRevLett.111.062006}{\emph{Phys. Rev. Lett.}
{\bfseries 111} (2013) 062006}
[\href{https://arxiv.org/abs/1301.2325}{{\ttfamily 1301.2325}}].
\bibitem{Dubovsky:2014fma}
S.~Dubovsky, R.~Flauger and V.~Gorbenko, \emph{{Flux Tube Spectra from
Approximate Integrability at Low Energies}},
\href{https://doi.org/10.1134/S1063776115030188}{\emph{J. Exp. Theor. Phys.}
{\bfseries 120} (2015) 399}
[\href{https://arxiv.org/abs/1404.0037}{{\ttfamily 1404.0037}}].
\bibitem{polyakov1975compact}
A.M.~Polyakov, \emph{Compact gauge fields and the infrared catastrophe},
{\emph{Physics Letters B} {\bfseries 59} (1975) 82}.
\bibitem{polyakov1977quark}
A.M.~Polyakov, \emph{Quark confinement and topology of gauge theories},
{\emph{Nuclear Physics B} {\bfseries 120} (1977) 429}.
\bibitem{Athenodorou:2022pmz}
A.~Athenodorou, S.~Dubovsky, C.~Luo and M.~Teper, \emph{{Excitations of Ising
strings on a lattice}},
\href{https://doi.org/10.1007/JHEP05(2023)082}{\emph{JHEP} {\bfseries 05}
(2023) 082} [\href{https://arxiv.org/abs/2301.00034}{{\ttfamily
2301.00034}}].
\bibitem{zngaugetheory}
A.~Athenodorou, S.~Dubovsky, C.~Luo and M.~Teper, \emph{{Confinement and the
Critical Behavior of 3D $\mathbb{Z}_N$ Gauge Theories}}, {\emph{to appear}
(2023) }.
\bibitem{Teper:1998te}
M.J.~Teper, \emph{{SU(N) gauge theories in (2+1)-dimensions}},
\href{https://doi.org/10.1103/PhysRevD.59.014512}{\emph{Phys. Rev. D}
{\bfseries 59} (1999) 014512}
[\href{https://arxiv.org/abs/hep-lat/9804008}{{\ttfamily hep-lat/9804008}}].
\bibitem{savit1980duality}
R.~Savit, \emph{Duality in field theory and statistical systems},
{\emph{Reviews of Modern Physics} {\bfseries 52} (1980) 453}.
\bibitem{blote1999cluster}
H.W.~Bl{\"o}te, L.N.~Shchur and A.L.~Talapov, \emph{The cluster processor: New
results}, {\emph{International Journal of Modern Physics C} {\bfseries 10}
(1999) 1137}.
\bibitem{Borisenko:2012nf}
O.~Borisenko, V.~Chelnokov, G.~Cortese, M.~Gravina, A.~Papa and I.~Surzhikov,
\emph{{Phase structure of 3D Z(N) lattice gauge theories at finite
temperature}},
\href{https://doi.org/10.1016/j.nuclphysb.2013.01.009}{\emph{Nucl. Phys. B}
{\bfseries 870} (2013) 159}
[\href{https://arxiv.org/abs/1212.3198}{{\ttfamily 1212.3198}}].
\bibitem{ukawa1980dual}
A.~Ukawa, P.~Windey and A.H.~Guth, \emph{Dual variables for lattice gauge
theories and the phase structure of z (n) systems}, {\emph{Physical Review D}
{\bfseries 21} (1980) 1013}.
\bibitem{hove2003criticality}
J.~Hove and A.~Sudb{\o}, \emph{Criticality versus q in the (2+ 1)-dimensional z
q clock model}, {\emph{Physical Review E} {\bfseries 68} (2003) 046107}.
\bibitem{Borisenko:2013xna}
O.~Borisenko, V.~Chelnokov, G.~Cortese, M.~Gravina, A.~Papa and I.~Surzhikov,
\emph{{Critical behavior of 3D Z(N) lattice gauge theories at zero
temperature}},
\href{https://doi.org/10.1016/j.nuclphysb.2013.12.003}{\emph{Nucl. Phys. B}
{\bfseries 879} (2014) 80} [\href{https://arxiv.org/abs/1310.5997}{{\ttfamily
1310.5997}}].
\bibitem{oshikawa2000ordered}
M.~Oshikawa, \emph{Ordered phase and scaling in z n models and the three-state
antiferromagnetic potts model in three dimensions}, {\emph{Physical Review B}
{\bfseries 61} (2000) 3430}.
\bibitem{bhanot1980phase}
G.~Bhanot and M.~Creutz, \emph{Phase diagram of z (n) and u (1) gauge theories
in three dimensions}, {\emph{Physical Review D} {\bfseries 21} (1980) 2892}.
\bibitem{Chester:2019ifh}
S.M.~Chester, W.~Landry, J.~Liu, D.~Poland, D.~Simmons-Duffin, N.~Su et~al.,
\emph{{Carving out OPE space and precise $O(2)$ model critical exponents}},
\href{https://doi.org/10.1007/JHEP06(2020)142}{\emph{JHEP} {\bfseries 06}
(2020) 142} [\href{https://arxiv.org/abs/1912.03324}{{\ttfamily
1912.03324}}].
\bibitem{shao2020monte}
H.~Shao, W.~Guo and A.W.~Sandvik, \emph{Monte carlo renormalization flows in
the space of relevant and irrelevant operators: Application to
three-dimensional clock models}, {\emph{Physical Review Letters} {\bfseries
124} (2020) 080602}.
\end{thebibliography}\endgroup