betaFeedback

WorksPhysics and Astronomy

Color Confinement and Random Matrices – A random walk down group manifold toward Casimir scaling –

Abstract

We explain the microscopic origin of linear confinement potential with the Casimir scaling in generic confining gauge theories. In the low-temperature regime of confining gauge theories such as QCD, Polyakov lines are slowly varying Haar random modulo exponentially small corrections with respect to the inverse temperature, as shown by one of the authors (M.~H.) and Watanabe. With exact Haar randomness, computation of the two-point correlator of Polyakov loops reduces to the problem of random walk on group manifold. Linear confinement potential with approximate Casimir scaling except at short distances follows naturally from slowly varying Haar randomness. With exponentially small corrections to Haar randomness, string breaking and loss of Casimir scaling at long distance follow. Hence we obtain the Casimir scaling which is only approximate and holds only at intermediate distance, which is precisely needed to explain the results of lattice simulations. For (1+1)(1+1)-dimensional theories, there is a simplification that admits the Casimir scaling at short distances as well.

Introduction

In the low-temperature regime of confining gauge theory, the Polyakov line must be slowly varying Haar random up to small corrections and gauge transformations [1, 2, 3]. We show how linear confinement potential with Casimir scaling [4, 5, 6, 7, 8] follows from this fact.1 The property of a Haar random distribution is explained below in more detail, repeating some of the arguments in Ref. [1]. It is not based on a strong coupling limit in lattice gauge theory but rather a universal property that emerges for the Polyakov line at a certain range of energy scales.

For concreteness, we consider SU(N)\mathrm{SU}(N) gauge theory, although the same argument applies to any gauge group and any matter content. We quantize this theory via Euclidean path integral at temperature T=β−1T = \beta^{-1}. At each spatial point x⃗\vec{x}, we can define Polyakov line Px⃗P_{\vec{x}} as

Px⃗=Path ordering[ei∫0βdtAt(t,x⃗)].(1)P_{\vec{x}} = \mathrm{Path\ ordering}\left[e^{i\int_{0}^{\beta} dt A_{t}(t,\vec{x})}\right]. \tag*{(1)}

We choose a reference point x⃗0\vec{x}_{0} and define the Wilson line connecting these two points,

U(x⃗0,x⃗)=Path ordering[ei∫Cdx⃗ A⃗],(2)U(\vec{x}_{0},\vec{x}) = \mathrm{Path\ ordering}\left[e^{i\int_{C} d\vec{x}\,\vec{A}}\right], \tag*{(2)}

where CC is a contour connecting x⃗0\vec{x}_0 and x⃗\vec{x} at fixed Euclidean time t=0t = 0. We claim that Px⃗′P'_{\vec{x}} defined by

Px⃗′≡U(x⃗0,x⃗)⋅Px⃗⋅U(x⃗0,x⃗)−1(3)P'_{\vec{x}} \equiv U(\vec{x}_0,\vec{x}) \cdot P_{\vec{x}} \cdot U(\vec{x}_0,\vec{x})^{-1} \tag*{(3)}

(see Figure 1) must be close to slowly varying Haar random, after appropriate smearing of gauge field that will be explained in § 2.3. Namely, statistically, Px⃗′P'_{\vec{x}} is Haar random at each x⃗\vec{x}, and it is slowly varying as a function of x⃗\vec{x} in each typical field configuration that has a large weight in the path integral.

Figure 1. Illustration of Px⃗′≡U(x⃗0,x⃗)⋅Px⃗⋅(U(x⃗0,x⃗))−1P'_{\vec{x}} \equiv U(\vec{x}_0,\vec{x}) \cdot P_{\vec{x}} \cdot(U(\vec{x}_0,\vec{x}))^{-1} defined in (3) in terms of the path on compact space.

A different way to express the slowly varying Haar randomness is to consider a change along a straight line x⃗=x⃗0+Lu^\vec{x} = \vec{x}_0 + L\hat{u} with some unit vector u^\hat{u} and length L>0L > 0. The change of Px⃗P_{\vec{x}} is a smooth analog of random walk on the group manifold.2 For each typical field configuration in path integral, we can write the change of Polyakov line as

Px⃗′−1Px⃗+Lu^′=Path ordering⁡[ei∑α∫0LdL′vα(L′)Tα](4)P_{\vec{x}}^{\prime-1} P'_{\vec{x}+L\hat{u}} = \operatorname{Path\ ordering}\left[e^{i\sum_{\alpha}\int_{0}^{L} dL' v_{\alpha}(L')T_{\alpha}}\right] \tag*{(4)}

where vα(L)v_{\alpha}(L) (α=1,2,⋯ ,N2−1\alpha= 1, 2, \cdots, N^2 - 1) are smooth functions and TαT_{\alpha} are the generators of SU(N)\mathrm{SU}(N). This ‘velocity’ vα(L)v_{\alpha}(L) gradually and randomly changes. We can write a similar expression for any representation r:

R(r)(Px⃗′−1Px⃗+Lu^′)=Path ordering⁡[ei∑α∫0LdL′vα(L′)Tα(r)].(5)R^{(r)}\left(P_{\vec{x}}^{\prime-1} P'_{\vec{x}+L\hat{u}}\right) = \operatorname{Path\ ordering}\left[e^{i\sum_{\alpha}\int_{0}^{L} dL' v_{\alpha}(L')T_{\alpha}^{(r)}}\right]. \tag*{(5)}

The main part of this paper explains how the Casimir scaling of string tension follows from this random-walk nature. In the rest of this introduction section, we will define the Casimir scaling and explain how the computation of string tension is related to the random-walk nature of the Polyakov line. In § 2, we review the previous work and explain why the Polyakov line has such a property. In § 2.4, we estimate the rate of change (i.e., the meaning of ‘slow’) based on a lattice regularization. Note that slowly varying Haar randomness does not apply to the abelianized regime where SU(N)\mathrm{SU}(N) breaks to U(1)N−1\mathrm{U}(1)^{N-1} because the correction to slowly varying Haar randomness is too large. Ref. [16] confirmed that the Casimir scaling does not hold in such a theory.

Casimir scaling

We normalize generators of the Lie algebra TαT_{\alpha} (α=1,2,⋯ ,N2−1\alpha= 1, 2, \cdots, N^2 - 1) in the fundamental representation as

Tr⁡(TαTβ)=2δαβ.(6)\operatorname{Tr}(T_{\alpha}T_{\beta}) = 2\delta_{\alpha\beta}. \tag*{(6)}

For SU(2) and SU(3), we can use Pauli matrices and Gell-Mann matrices, respectively. For irreducible representation r, we define the quadratic Casimir invariant CrC_r by

∑α(Tα(r))2=4Cr⋅1,(7)\sum_{\alpha} \left(T_{\alpha}^{(r)}\right)^2 = 4C_r \cdot1, \tag*{(7)}

where 11 is the identity. For spin-jj representation of SU(2), Cspin-j=j(j+1)C_{\mathrm{spin}\text{-}j}=j(j+1). For SU(3), Cfund=43C_{\mathrm{fund}}=\frac{4}{3}, Cadj=3C_{\mathrm{adj}}=3, C2-sym=103C_{2\text{-}\mathrm{sym}}=\frac{10}{3}, C3-sym=6C_{3\text{-}\mathrm{sym}}=6 for fundamental, adjoint, rank-2 symmetric, and rank-3 symmetric representations, respectively. Casimir scaling means that string tension associated with representation r is proportional to CrC_r. Specifically, the Polyakov loop correlator is

⟨χr(Px⃗)⋅χr(Px⃗+Lu^)⟩∼e−βCrLσ02,(8)\left\langle\chi_r(P_{\vec{x}}) \cdot\chi_r(P_{\vec{x}+L\hat{u}}) \right\rangle\sim e^{-\beta C_r L \sigma_0^2}, \tag*{(8)}

with σ0\sigma_0 independent of r in the regime where Casimir scaling holds. This scaling does not hold at very long distances where strings can break due to pair creation of particles such as quarks or gluons.

Random walk and string tension

Let us see how the computation of string tension is related to random walk. Character χr\chi_r associated with irreducible representation r is defined by taking the trace of the representation matrix. Specifically,

χr(Px⃗)=Tr⁡r(R(r)(Px⃗))(9)\chi_r(P_{\vec{x}})=\operatorname{Tr}_r\left(R^{(r)}(P_{\vec{x}})\right) \tag*{(9)}

is the Polyakov loop in representation rr at point x⃗\vec{x}. Because of

χr(Px⃗)=χr(Px⃗′)(10)\chi_r(P_{\vec{x}})=\chi_r(P'_{\vec{x}}) \tag*{(10)}

and

χr′(Px⃗+Lu^′)=χr′(Px⃗′)=χr′(Px⃗′⋅(Px⃗′−1⋅Px⃗+Lu^′)),(11)\chi_{r'}(P'_{\vec{x}+L\hat{u}})=\chi_{r'}(P'_{\vec{x}})=\chi_{r'}\left(P'_{\vec{x}}\cdot\left(P_{\vec{x}}^{\prime-1}\cdot P'_{\vec{x}+L\hat{u}}\right)\right), \tag*{(11)}

the two-point function of Polyakov loops can be written as

⟨(χr(Px⃗))∗⋅χr′(Px⃗+Lu^)⟩=⟨(χr(Px⃗′))∗⋅χr′(Px⃗′⋅(Px⃗′−1⋅Px⃗+Lu^′))⟩.(12)\left\langle\left(\chi_r(P_{\vec{x}})\right)^*\cdot\chi_{r'}(P_{\vec{x}+L\hat{u}})\right\rangle=\left\langle\left(\chi_r(P'_{\vec{x}})\right)^*\cdot\chi_{r'}\left(P'_{\vec{x}}\cdot\left(P_{\vec{x}}^{\prime-1}\cdot P'_{\vec{x}+L\hat{u}}\right)\right)\right\rangle. \tag*{(12)}

So far, ⟨ ⟩\langle\,\rangle meant the expectation value defined by path integral. From here on, we interpret ⟨ ⟩\langle\,\rangle as the average over slowly varying Haar random configurations {Px⃗′}\{P'_{\vec{x}}\}, neglecting corrections to slowly varying Haar randomness. (As we will see, this approximation fails if LL is too large.)

We consider random walk from Px⃗′P'_{\vec{x}} to Px⃗+Lu^′P'_{\vec{x}+L\hat{u}}. The starting point Px⃗′P'_{\vec{x}} is Haar random, and it is not correlated to the displacement Px⃗′−1Px⃗+Lu^′P_{\vec{x}}^{\prime-1}P'_{\vec{x}+L\hat{u}}. Therefore, we can treat Px⃗′P'_{\vec{x}} and Px⃗′−1Px⃗+Lu^′P_{\vec{x}}^{\prime-1}P'_{\vec{x}+L\hat{u}} as independent variables and integrate out Px⃗′P'_{\vec{x}} by using the orthogonality condition of representations under the Haar-random average,

1Vol⁡(SU(N))∫dP(Rij(r)(P))∗⋅Rkl(r′)(P)=dr−1δrr′δilδkj(13)\frac{1}{\operatorname{Vol}(\mathrm{SU}(N))}\int dP\left(R_{ij}^{(r)}(P)\right)^*\cdot R_{kl}^{(r')}(P)=d_r^{-1}\delta_{rr'}\delta_{il}\delta_{kj} \tag*{(13)}

to obtain

⟨(χr(Px⃗))∗⋅χr′(Px⃗+Lu^)⟩=dr−1δrr′⟨χr(Px⃗′−1⋅Px⃗+Lu^′)⟩,(14)\left\langle\left(\chi_r(P_{\vec{x}})\right)^*\cdot\chi_{r'}(P_{\vec{x}+L\hat{u}})\right\rangle=d_r^{-1}\delta_{rr'}\left\langle\chi_r\left(P_{\vec{x}}^{\prime-1}\cdot P'_{\vec{x}+L\hat{u}}\right)\right\rangle, \tag*{(14)}

and hence, by using (8), we get

dr−1δrr′⟨χr(Px⃗′−1⋅Px⃗+Lu^′)⟩∼e−βCrLσ02.(15)d_r^{-1}\delta_{rr'}\left\langle\chi_r\left(P_{\vec{x}}^{\prime-1}\cdot P'_{\vec{x}+L\hat{u}}\right)\right\rangle\sim e^{-\beta C_rL\sigma_0^2}. \tag*{(15)}

Combined with (4) and (5), the computation of the two-point function reduces to a sort of random walk on group manifold. The difference from the vanilla random walk problem is that velocity vv changes slowly.

Organization of this paper

In the rest of the paper, we will show that confinement with Casimir scaling follows naturally from such a random walk on a group manifold. Our argument is based only on the slowly varying Haar randomness, and the choice of gauge group or matter content is irrelevant. However, corrections to Haar randomness depend on these specifications of the theory, and they are relevant for long-distance physics where the Casimir scaling is lost.

This paper is organized as follows. Sec. 2 explains how slowly varying Haar randomness of the Polyakov loop emerges from a generic mechanism of confinement. Sec. 3, the main part of this paper, derives Casimir scaling from a random walk. The simplest, ‘vanilla’ random walk is studied in Sec. 3.1. It is solvable, and Casimir scaling can be analytically derived. A more sophisticated variant of random walk is studied in Sec. 3.2. We find a neat correspondence with a lattice regularization and confirm Casimir scaling and consistency with the microscopic mechanism of confinement based on the similarity to Bose-Einstein condensation [1, 2, 3]. Note, however, that we only model the random walk and details that can affect the short-distance behavior might depend on the specific model. We expect that different confining theories can show different short-distance behaviors because of the dependence on such details. Larger distance behavior, on the other hand, emerges quite universally from the different models of the random walk. In Sec. 4, it is discussed that corrections to exact Haar randomness lead to a breaking of Casimir scaling. Sec. 5 is devoted to discussion.

Microscopic origin of slowly varying Haar randomness

In this section, we review the origin of the slowly varying Haar randomness of the Polyakov line [1]. As a first step, we rewrite the partition function in terms of an extended Hilbert space (Hext\mathcal{H}_{\mathrm{ext}}), including also non-gauge invariant states, and a projection to the gauge-invariant part (Hinv\mathcal{H}_{\mathrm{inv}}) by integration over all gauge variations (gauge orbit). For a gauge theory with gauge group GG, the canonical partition function can be written as

Z(T)=Tr⁡Hinve−H^/T=1vol⁡G∫GdgTr⁡Hext(g^e−H^/T).(16)Z(T) = \operatorname{Tr}_{\mathcal{H}_{\mathrm{inv}}} e^{-\hat{H}/T} = \frac{1}{\operatorname{vol} G} \int_G dg \operatorname{Tr}_{\mathcal{H}_{\mathrm{ext}}}\left(\hat{g}e^{-\hat{H}/T}\right). \tag*{(16)}

where the integral is taken over the Haar measure. A simple but crucial fact is that g^\hat{g} in the expression above corresponds to the Polyakov line [3].

SN invariance and Bose-Einstein condensation

To understand the origin of slowly varying Haar randomness, it is instructive to review the connection between the Bose-Einstein Condensation (BEC) and confinement in gauge theories [1, 3]. In this case the gauge group is the permutations SNS_N of NN particles, the group integral becomes a sum over all possible group elements.

Consider a system of NN indistinguishable bosons with a harmonic potential. The Hamiltonian is given by

H^=12∑i=1N(p^xi2+p^yi2+p^zi2+x^i2+y^i2+z^i2).(17)\hat{H} = \frac{1}{2}\sum_{i=1}^{N}\left(\hat{p}_{xi}^{2}+\hat{p}_{yi}^{2}+\hat{p}_{zi}^{2}+\hat{x}_{i}^{2}+\hat{y}_{i}^{2}+\hat{z}_{i}^{2}\right). \tag*{(17)}

This system has a gauged SNS_N permutation group symmetry. Namely, states connected by SNS_N transformation (exchange of labels 1,2,⋯ ,N1, 2, \cdots, N) must be identified. A generic state spanning the extended Hilbert space Hext\mathcal{H}_{\mathrm{ext}} is given in the Fock basis as

∣n⃗1,n⃗2,…,n⃗N⟩=∏i=x,y,za1i†n1in1i!a2i†n2in2i!⋯aNi†nNinNi!∣0⟩(18)\lvert\vec{n}_1,\vec{n}_2,\ldots,\vec{n}_N\rangle= \prod_{i=x,y,z}\frac{a_{1i}^{\dagger n_{1i}}}{\sqrt{n_{1i}!}}\frac{a_{2i}^{\dagger n_{2i}}}{\sqrt{n_{2i}!}}\cdots\frac{a_{Ni}^{\dagger n_{Ni}}}{\sqrt{n_{Ni}!}}\lvert0\rangle \tag*{(18)}

where n⃗i=(nix,niy,niz)\vec{n}_i=(n_{ix},n_{iy},n_{iz}) denotes the occupation numbers for each boson. The Fock vacuum is described as ∣0⟩=∣0,0,…,0⟩\lvert0\rangle=\lvert0,0,\ldots,0\rangle.

For the system of NN bosons under consideration here with a gauged SNS_N symmetry, the partition function takes the form,

Z(T)=1N!∑σ∈SNTr⁡Hext(σ^e−H^/T)=1N!∑σ∈SN∑n⃗1,n⃗2,…,n⃗N⟨n⃗1,n⃗2,…,n⃗N∣σ^e−H^/T∣n⃗1,n⃗2,…,n⃗N⟩=1N!∑n⃗1,n⃗2,…,n⃗Ne−(En⃗1+⋯+En⃗N)/T∑σ∈SN⟨n⃗1,n⃗2,…,n⃗N∣n⃗σ(1),n⃗σ(2),…,n⃗σ(N)⟩=1N!∑n⃗1,n⃗2,…,n⃗Ne−(En⃗1+⋯+En⃗N)/T∑σ∈SNδn⃗1n⃗σ(1)δn⃗2n⃗σ(2)⋯δn⃗Nn⃗σ(N)=1N!∑n⃗1,n⃗2,…,n⃗Ne−(En⃗1+⋯+En⃗N)/TV∣n⃗1,n⃗2,…,n⃗N⟩(19)\begin{aligned} Z(T) &= \frac{1}{N!}\sum_{\sigma\in S_N}\operatorname{Tr}_{\mathcal{H}_{\mathrm{ext}}}\left(\hat{\sigma}e^{-\hat{H}/T}\right) \\ &= \frac{1}{N!}\sum_{\sigma\in S_N}\sum_{\vec{n}_1,\vec{n}_2,\ldots,\vec{n}_N}\langle\vec{n}_1,\vec{n}_2,\ldots,\vec{n}_N\lvert\hat{\sigma}e^{-\hat{H}/T}\lvert\vec{n}_1,\vec{n}_2,\ldots,\vec{n}_N\rangle\\ &= \frac{1}{N!}\sum_{\vec{n}_1,\vec{n}_2,\ldots,\vec{n}_N}e^{-(E_{\vec{n}_1}+\cdots+E_{\vec{n}_N})/T}\sum_{\sigma\in S_N}\langle\vec{n}_1,\vec{n}_2,\ldots,\vec{n}_N\lvert\vec{n}_{\sigma(1)},\vec{n}_{\sigma(2)},\ldots,\vec{n}_{\sigma(N)}\rangle\\ &= \frac{1}{N!}\sum_{\vec{n}_1,\vec{n}_2,\ldots,\vec{n}_N}e^{-(E_{\vec{n}_1}+\cdots+E_{\vec{n}_N})/T}\sum_{\sigma\in S_N}\delta_{\vec{n}_1\vec{n}_{\sigma(1)}}\delta_{\vec{n}_2\vec{n}_{\sigma(2)}}\cdots\delta_{\vec{n}_N\vec{n}_{\sigma(N)}} \\ &= \frac{1}{N!}\sum_{\vec{n}_1,\vec{n}_2,\ldots,\vec{n}_N}e^{-(E_{\vec{n}_1}+\cdots+E_{\vec{n}_N})/T}V_{\lvert\vec{n}_1,\vec{n}_2,\ldots,\vec{n}_N\rangle} \tag*{(19)} \end{aligned}

Here, V∣n⃗1,n⃗2,…,n⃗N⟩≡∑σ∈SNδn⃗1n⃗σ(1)δn⃗2n⃗σ(2)⋯δn⃗Nn⃗σ(N)V_{\lvert\vec{n}_1,\vec{n}_2,\ldots,\vec{n}_N\rangle}\equiv\sum_{\sigma\in S_N}\delta_{\vec{n}_1\vec{n}_{\sigma(1)}}\delta_{\vec{n}_2\vec{n}_{\sigma(2)}}\cdots\delta_{\vec{n}_N\vec{n}_{\sigma(N)}} is the volume of the stabilizer subgroup of the state ∣n⃗1,n⃗2,…,n⃗N⟩\lvert\vec{n}_1,\vec{n}_2,\ldots,\vec{n}_N\rangle i.e., the number of the elements of SNS_N that leave the state invariant. If all the particles have different occupation numbers (n⃗i≠n⃗j\vec{n}_i\ne\vec{n}_j for all i≠ji\ne j), then the stabilizer of such a state is just the identity element σ=1\sigma=1. The contribution of such a state to the partition function thus remains suppressed by a factor of 1/N!1/N!. However, if all the particles have the same occupation number (n⃗1=⋯=n⃗N=n⃗\vec{n}_1=\cdots=\vec{n}_N=\vec{n}), then the stabilizer of such a state is the entire permutation group SNS_N with V∣n⃗,…,n⃗⟩=N!V_{\lvert\vec{n},\ldots,\vec{n}\rangle}=N!. The suppression factor in the partition function thus experiences an enhancement and becomes equal to 1 and such a state is then favoured. This is exactly what happens in the ground state ∣0⟩\lvert0\rangle (n⃗1=⋯=n⃗N=0⃗\vec{n}_1=\cdots=\vec{n}_N=\vec{0}). More generally, if the energy is distributed to a part of particles and the rest remains in the one-particle ground state (say, n⃗1,⋯ ,n⃗M\vec{n}_1,\cdots,\vec{n}_M becomes nonzero while n⃗M+1=⋯=n⃗N=0⃗\vec{n}_{M+1}=\cdots=\vec{n}_N=\vec{0}), a large enhancement factor (N−M)!(N-M)! is obtained. In general, the optimal value of MM between 0 and NN is realized so that the combination of such an enhancement factor and the entropy of the excited sector is maximized. This is the BEC of N−MN-M particles. This mechanism works with finite interactions as well, and unbroken SN−MS_{N-M} symmetry in the extended Hilbert space characterizes BEC. For M=0M=0, all ‘Polyakov line’ σ∈SN\sigma\in S_N contribute equally, which means the Haar-randomness.

SU(NN) invariance and confinement

Coming back from BEC to our original question about gauge theories including SU(NN) pure Yang-Mills theory, a similar enhancement can be observed. In general, an SU(MM) subgroup can be deconfined, while the rest of the degrees of freedom remain confined [17, 18, 19, 20] as illustrated in Figure 2. Roughly speaking, such partial deconfinement is characterized by an unbroken SU(N−MN-M) symmetry. Since this is considered in the extended Hilbert space, there is no conflict with gauge invariance; simply, all embeddings of SU(MM) into SU(NN) are equivalent. The completely confined phase we are discussing in this paper is characterized by unbroken SU(NN) symmetry in the extended Hilbert space. Let us elaborate on this point for the case of Yang-Mills theory and see how the slowly varying Haar randomness of the Polyakov line emerges. The differences from the example of BEC are that the gauge group is bigger (SU(NN) instead of SNS_N) and that we need to deal with spatial coordinates because we deal with QFT rather than quantum mechanics.

Partial deconfinement in SU(N) QCD with fundamental quarks

Figure 2. Partial deconfinement in SU(NN) QCD with NfN_f fundamental quarks. [Left] Gauge field. M×MM \times M sub-matrix is deconfined. [Right] Quarks. MM components are deconfined. Any embeddings of SU(MM) into SU(NN) are equivalent because of gauge symmetry. This figure is taken from Ref. [20].

For concreteness, we consider pure Yang-Mills theory and use a lattice regularization with lattice spacing aa and unitary link variables Un⃗,μU_{\vec{n},\mu} which is related to the Hermitian gauge field AμA_\mu as Un⃗,μ=eiagYMAn⃗,μU_{\vec{n},\mu}=e^{iag_{\rm YM}A_{\vec{n},\mu}}. Here, n⃗\vec{n} is an integer-valued vector that specifies a spatial site on the lattice, and μ\mu is a spatial direction x,y,…x,y,\ldots. In the extended-Hilbert space picture, wave functions are defined on ∏μ,n⃗[SU(N)]μ,n⃗\prod_{\mu,\vec{n}}[\mathrm{SU}(N)]_{\mu,\vec{n}}, where [SU(N)]μ,n⃗[\mathrm{SU}(N)]_{\mu,\vec{n}} corresponds to each link. Specifically, on each link, we consider a linear combination of coordinate eigenstates ∣Un⃗,μ⟩\lvert U_{\vec{n},\mu}\rangle. The confined vacuum in or sufficiently close to the continuum limit is a wave packet localized around Un⃗,μ=1U_{\vec{n},\mu}=1 up to gauge transformation. (As we will see shortly, a large enhancement factor is associated with such a state, which justifies its dominance in the partition function.) Wave packets connected by a (spacial) gauge transformation Ω∈SU(N)\Omega\in\mathrm{SU}(N) are equivalent. Therefore, we should consider wave packets localized around Un⃗,μ=Ωn⃗−1Ωn⃗+μ^U_{\vec{n},\mu}=\Omega_{\vec{n}}^{-1}\Omega_{\vec{n}+\hat{\mu}}.

Corresponding to ⟨n⃗1,n⃗2,…,n⃗N∣n⃗σ(1),n⃗σ(2),…,n⃗σ(N)⟩\langle\vec{n}_1,\vec{n}_2,\ldots,\vec{n}_N \mid\vec{n}_{\sigma(1)},\vec{n}_{\sigma(2)},\ldots,\vec{n}_{\sigma(N)}\rangle, we have a factor ⟨Φ∣g^∣Φ⟩\langle\Phi|\hat{g}|\Phi\rangle in the partition function, where we can take ∣Φ⟩|\Phi\rangle to be a wave packet. By definition, this factor measures overlap between ∣Φ⟩|\Phi\rangle and g^∣Φ⟩\hat{g}|\Phi\rangle; see Figure 3. Because the Polyakov lines correspond to g^\hat{g}, such Polyakov lines that do not move the wave packet too much can contribute to the enhancement factor. Let us consider the action of the Polyakov loop corresponding to ‘global’ SU(N)SU(N) transformation, Pn⃗≡Ωn⃗−1VΩn⃗P_{\vec{n}}\equiv\Omega_{\vec{n}}^{-1}V\Omega_{\vec{n}}. It is easy to see that the wave packet under consideration does not move regardless of the choice of VV, because

Wave packet overlap illustration

Figure 3. For a wave packet ∣Φ⟩|\Phi\rangle, such gg that does not move ∣Φ⟩|\Phi\rangle too much can lead to an enhancement factor and give a large contribution to the partition function.

Pn⃗−1(Ωn⃗−1Ωn⃗+μ^)Pn⃗+μ^=Ωn⃗−1V−1Ωn⃗(Ωn⃗−1Ωn⃗+μ^)Ωn⃗+μ^−1VΩn⃗+μ^=Ωn⃗−1Ωn⃗+μ^.(20)P_{\vec{n}}^{-1}(\Omega_{\vec{n}}^{-1}\Omega_{\vec{n}+\hat{\mu}})P_{\vec{n}+\hat{\mu}}=\Omega_{\vec{n}}^{-1}V^{-1}\Omega_{\vec{n}}(\Omega_{\vec{n}}^{-1}\Omega_{\vec{n}+\hat{\mu}})\Omega_{\vec{n}+\hat{\mu}}^{-1}V\Omega_{\vec{n}+\hat{\mu}}=\Omega_{\vec{n}}^{-1}\Omega_{\vec{n}+\hat{\mu}}. \tag*{(20)}

Thus, the vacuum is invariant under this ‘global’ SU(N)SU(N) transformation, leading to an enhancement factor as in the BEC case. The enhancement factor scales with NN as eN2e^{N^2}. For partially-deconfined states shown in Figure 2,

V=(1M00V~)(21)V=\begin{pmatrix} 1_M & 0 \\ 0 & \widetilde{V} \end{pmatrix} \tag*{(21)}

with any V~∈SU(N−M)\widetilde{V}\in SU(N-M) could leave the states invariant. This leads to an enhancement factor ∼e(N−M)2\sim e^{(N-M)^2}.

Let us elaborate on the argument above and show that the enhancement factor can increase exponentially with volume. Roughly speaking, we just promote a ‘global’ transformation to ‘slowly varying’ transformation. Let us start with the case of the confined vacuum (M=0M=0). We consider a ‘local’ SU(N)SU(N) transformation with VV now being dependent on the site n⃗\vec{n} with the Polyakov line taking the form Pn⃗≡Ωn⃗−1Vn⃗Ωn⃗P_{\vec{n}}\equiv\Omega_{\vec{n}}^{-1}V_{\vec{n}}\Omega_{\vec{n}}:

Pn⃗−1Un⃗,μPn⃗+μ^=Ωn⃗−1Vn⃗−1Ωn⃗(Ωn⃗−1Ωn⃗+μ^)Ωn⃗+μ^−1Vn⃗+μ^Ωn⃗+μ^=Ωn⃗−1Vn⃗−1Vn⃗+μ^Ωn⃗+μ^.(22)P_{\vec{n}}^{-1}U_{\vec{n},\mu}P_{\vec{n}+\hat{\mu}}=\Omega_{\vec{n}}^{-1}V_{\vec{n}}^{-1}\Omega_{\vec{n}}(\Omega_{\vec{n}}^{-1}\Omega_{\vec{n}+\hat{\mu}})\Omega_{\vec{n}+\hat{\mu}}^{-1}V_{\vec{n}+\hat{\mu}}\Omega_{\vec{n}+\hat{\mu}}=\Omega_{\vec{n}}^{-1}V_{\vec{n}}^{-1}V_{\vec{n}+\hat{\mu}}\Omega_{\vec{n}+\hat{\mu}}. \tag*{(22)}

If Vn⃗V_{\vec{n}} is slowly varying, i.e., Vn⃗−1Vn⃗+μ^V_{\vec{n}}^{-1}V_{\vec{n}+\hat{\mu}} is close to 11, and we can still have a significant enhancement factor as before. For the confined vacuum, we can choose Vn⃗V_{\vec{n}} arbitrarily as long as it is slowly varying, indeed leading to a large enhancement factor that increases with spatial volume. Intuitively, we can split the volume into multiple pieces that are large enough so that the correlation between Polyakov lines in different pieces is negligible, and then, we can associate ∼eN2\sim e^{N^2} to each piece, which leads to the scaling of the enhancement factor eN2×volumee^{N^2\times\mathrm{volume}}.

Continuum limit and smearing

In the previous subsection, we explained how the bare Polyakov line and symmetry of wave function on each spatial link are related. When we take continuum limit (a→0a\to0), we need to see how renormalization affects the logic.

In the continuum limit with fixed temperature, the expectation value of the bare Polyakov loop vanishes even in the deconfined phase for any representation. A multiplicative renormalization factor is needed to keep the expectation value finite [9, 10, 11]. This means that Polyakov line becomes Haar random even in the deconfined phase. In terms of symmetries, this means that it is hard to see the difference between the ground state and excited states as long as we focus on the wave function on each link. Note also that, associated with ultraviolet divergence, the size of the wave packet can grow as a→0a \to0 for 3d and 4d theories,3 which may destroy the slowly varyingness of Polyakov line. Still, however, infrared properties can be significantly different because the number of links increases and small change at each link can accumulate.

To capture the symmetries and properties relevant for thermodynamics, we should not use bare Polyakov line with too small lattice spacing. To renormalize the Polyakov line, we can combine the facts that the Polyakov line can be written as a sum of characters and each character can be renormalized in a standard manner [9, 10, 11]. We could also use a smeared gauge field and define a smeared version of the Polyakov loop.

As a concrete example, let us consider the smearing via gradient flow [21, 22, 23]. In this smearing scheme, smeared gauge field is defined at each link on the original lattice. By replacing the original gauge field with smeared one, the smeared Polyakov line is obtained. Because we are interested in correlators at not-so-short distance and at finite-temperature (specifically, in the confined phase), we should take smearing length at most order of ΛQCD−1\Lambda_{\mathrm{QCD}}^{-1}. Then, the smearing does not affect the long-distance behavior and hence the string tension. The smearing makes the spatial-coordinate dependence of Px⃗′P'_{\vec{x}} mild and introduces the slowly varying Haar randomness starting at short distance; see also § 2.4. We cannot use such a Polyakov line made of smeared field to study short-distance physics below the smearing length, but that is not a problem for us as long as we discuss Casimir scaling at sufficiently long distance.4 What we will claim is the existence of a smearing scheme consistent with slowly varying Haar randomness is a sufficient condition for the Casimir scaling to be valid.

Slowly varying Haar randomness

So far, we have seen that a large enhancement factor is indeed associated with the confined vacuum. We have also seen that Polyakov lines contributing to the enhancement factor are written as P⃗n⃗≡Ωn⃗−1Vn⃗Ωn⃗\vec{P}_{\vec{n}} \equiv\Omega_{\vec{n}}^{-1}V_{\vec{n}}\Omega_{\vec{n}}, where Vn⃗V_{\vec{n}} can be arbitrary as long as it is slowly varying. Typical Polyakov lines dominating the partition function (equivalently, typical Polyakov lines obtained from dominant configurations in the Euclidean path integral) are the ones that contribute to the enhancement factor. Therefore, typical Polyakov lines are slowly varying SU(NN) Haar random up to gauge transformation [1]. For the same logic, the Polyakov line restricted to the confined sector of the partially-deconfined phase should be slowly varying SU(N−MN-M) Haar random up to gauge transformation.

‘Slowly varying’ is admittedly a vague expression. Below, we give an order estimate for the rate of change. We want to know how slowly a typical Polyakov line changes, assuming the confined vacuum represented by a wave packet localized around a pure-gauge configuration. Because a typical Polyakov line must lead to a large enhancement factor, Vn⃗−1Vn⃗+μ^V_{\vec{n}}^{-1}V_{\vec{n}+\hat{\mu}} has to be sufficiently close to the identity so that the ground-state wave function does not move too much (Figure 3). Specifically, if Vn⃗−1Vn⃗+μ^V_{\vec{n}}^{-1}V_{\vec{n}+\hat{\mu}} is written as Vn⃗−1Vn⃗+μ^∼eiℓX~V_{\vec{n}}^{-1}V_{\vec{n}+\hat{\mu}} \sim e^{i\ell\widetilde{X}} where X~\widetilde{X}’s entries are of order 1, then ℓ\ell must be smaller than the radius of the wave packet.

Let us start with the two-dimensional Yang-Mills theory, which does not require smearing. Although it is straightforward to solve it in the path-integral formalism, let us use the Hamiltonian formulation to derive the same result. (The same result is derived in Appendix. § 7 by using path integral.) The Hamiltonian contains only the electric term, H^=a22Tr⁡∑x⃗E^x⃗2\hat{H} = \frac{a^{2}}{2}\operatorname{Tr}\sum_{\vec{x}}\hat{E}_{\vec{x}}^{2}, and there is no interaction between links. This is essentially (momentum)2^{2} on the group manifold. As temperature goes up, higher momentum contributes more, or equivalently, typical wave packet shrinks. The energy is proportional to TT, and hence, typical momentum is Ta=(βa)−1/2\sqrt{\frac{T}{a}} = (\beta a)^{-1/2}, and the radius of typical wave packet scales as βa\sqrt{\beta a}. Therefore, we expect Vn−1Vn+1∼eiβaX~V_{n}^{-1}V_{n+1} \sim e^{i\sqrt{\beta a}\widetilde{X}}. Due to the absence of the interaction in the Hamiltonian, there is no correlation between Vn−1Vn+1V_{n}^{-1}V_{n+1} for different nn. In other words, ‘velocity’ is totally random at each link. Therefore, Vm−1Vn∼eiβa∣m−n∣X~V_{m}^{-1}V_{n} \sim e^{i\sqrt{\beta a}|m-n|\widetilde{X}}.

Next, let us consider 3d and 4d Yang-Mills theory, which is more complicated because the bare Polyakov line can become Haar random in the continuum limit even in the deconfined phase. To take the continuum limit, we consider the loops made of smeared gauge field. Because of the randomness of the bare loop, smeared loop should also exhibit the random-walk nature. At a smearing radius of order ΛQCD−1\Lambda_{\mathrm{QCD}}^{-1}, the natural value of vv is of order a0a^{0}.

Let us also comment on the ‘acceleration’, i.e., the rate of change of vv. In 3d and 4d theories, vv in (4) has to be smooth because Px⃗−1Py⃗P_{\vec{x}}^{-1}P_{\vec{y}} depends only on the endpoints x⃗\vec{x} and y⃗\vec{y} and does not change depending on the path connecting these two points. For this reason, the Polyakov lines exhibit random walks with slowly changing velocity, which is different from the most common random walk in which the velocity at each step changes abruptly. In 2d theories, vv does not have to be continuous and the Polyakov line can literally random walk. This has an important consequence as we will discuss in § 3.3.

Casimir scaling from random walk

In this section, we assume the exact Haar randomness and derive the linear confinement potential with Casimir scaling. Hence, our task is to estimate ⟨χr(Px⃗′−1Px⃗+Lu^′)⟩\langle\chi_r(P_{\vec{x}}^{\prime-1}P_{\vec{x}+L\hat{u}}^{\prime})\rangle assuming the random walk (4). To evaluate (4), let us introduce step size aa, where Ka=LKa=L, and take a limit of a→0a\to0. (We use aa because it is analogous to the lattice spacing in lattice regularized theories.) We define WjW_j (j=1,2,⋯ ,Kj=1,2,\cdots,K) by

Wj≡(Px⃗+(j−1)au^′)−1Px⃗+jau^′.(23)W_j \equiv\left(P_{\vec{x}+(j-1)a\hat{u}}^{\prime}\right)^{-1}P_{\vec{x}+ja\hat{u}}^{\prime}. \tag*{(23)}

These can be combined to

Px⃗′−1Px⃗+Lu^′=W1W2⋯WK,(24)P_{\vec{x}}^{\prime-1}P_{\vec{x}+L\hat{u}}^{\prime}=W_1W_2\cdots W_K, \tag*{(24)}

as illustrated in Figure 4.

Figure 4. Contour corresponding to Px⃗′−1Px⃗′P_{\vec{x}}^{\prime-1}P_{\vec{x}}^{\prime} to illustrate the definition of WjW_j in (23).

Below, let us first consider the vanilla random walk as a solvable warm-up exercise, and then consider a more realistic setup that takes into account the correlation between WiW_i’s.

Vanilla random walk

As a crude approximation, let us treat W1,⋯ ,WKW_1,\cdots,W_K totally random, even neglecting the correlation of WiW_i and Wi+1W_{i+1}. This is the vanilla random walk. Apparently, we cannot always take the limit of a→0a\to0 in this case. To keep this in mind, let us use ss instead of aa. Physically, ss is a length scale needed for vv in (4) to change significantly and the correlation between v(L)v(L) and v(L+s)v(L+s) becomes small. With this approximation, the average of a product becomes the product of the average:

⟨W1(r)W2(r)⋯WK(r)⟩=⟨W1(r)⟩⟨W2(r)⟩⋯⟨WK(r)⟩=(⟨W(r)⟩)K.(25)\left\langle W_1^{(r)}W_2^{(r)}\cdots W_K^{(r)}\right\rangle=\left\langle W_1^{(r)}\right\rangle\left\langle W_2^{(r)}\right\rangle\cdots\left\langle W_K^{(r)}\right\rangle=\left(\left\langle W^{(r)}\right\rangle\right)^K . \tag*{(25)}

This expression already tells us an exponential decay. To see the Casimir scaling, we need a little bit more math. We assume W=eiΔXW=e^{i\Delta X}, where XX is random Gaussian with variance 1 and Δ\Delta is a small number. Then,

⟨W(r)⟩=⟨1+iΔxαTα(r)−Δ2xαxβ2Tα(r)Tβ(r)+⋯ ⟩=1−Δ22(Tα(r))2+⋯=1−2Δ2Cr1+⋯ .(26)\begin{aligned} \left\langle W^{(r)}\right\rangle&=\left\langle1+i\Delta x^\alpha T_\alpha^{(r)}-\frac{\Delta^2x^\alpha x^\beta}{2}T_\alpha^{(r)}T_\beta^{(r)}+\cdots\right\rangle\\ &=1-\frac{\Delta^2}{2}\left(T_\alpha^{(r)}\right)^2+\cdots\\ &=1-2\Delta^2C_r1+\cdots. \tag*{(26)} \end{aligned}

Hence, for small Δ\Delta,

⟨W(r)⟩≃e−2Δ2Cr1(27)\left\langle W^{(r)}\right\rangle\simeq e^{-2\Delta^2C_r1} \tag*{(27)}

and hence

⟨W1(r)⟩⟨W2(r)⟩⋯⟨WK(r)⟩≃e−2Δ2CrK1.(28)\left\langle W_1^{(r)}\right\rangle\left\langle W_2^{(r)}\right\rangle\cdots\left\langle W_K^{(r)}\right\rangle\simeq e^{-2\Delta^2C_rK}1 . \tag*{(28)}

Therefore,

⟨(χr(Px⃗))∗χr′(Px⃗+Lu^)⟩≃δrr′e−2Δ2CrK(29)\left\langle\left(\chi_r(P_{\vec{x}})\right)^*\chi_{r'}(P_{\vec{x}+L\hat{u}})\right\rangle\simeq\delta_{rr'}e^{-2\Delta^2C_rK} \tag*{(29)}

up to O(Δ4)O(\Delta^4) terms in the exponent. Therefore, by identifying Δ\Delta with physical parameters as

Δ2=βsσ02,(30)\Delta^2=\beta s\sigma_0^2 , \tag*{(30)}

string tension for representation r is Crσ02C_r\sigma_0^2 if O(Δ4)O(\Delta^4) terms are negligible. The dependence on ss is natural because after KK steps the distance traveled on the group manifold is ∼ΔK∼L\sim\Delta\sqrt{K}\sim\sqrt{L}, which is a typical case of random walk.

That terms of order Δ4\Delta^4 or higher break Casimir scaling would be good news because lattice simulations suggest Casimir scaling is not exact.

Several references [12, 24, 25] assumed the uncorrelated nature of color flux and explained the Casimir scaling in a way similar to the vanilla random walk in the limit of Δ→0\Delta\to0. Our discussion above is based on the microscopic mechanism applicable to generic confining gauge theories, and hence, we can give a more precise picture that takes into account a finite amount of correlations, as we will show next. It would be interesting to see if the hypothesis regarding the QCD vacuum proposed in the past can be justified based on our approach.

As a minor remark, we note that this vanilla random walk works for two-dimensional pure Yang-Mills.

Random walk with gradually changing velocity

In reality, we want the ‘velocity’ WiW_i to change gradually (except for 2d theory, as we will see in § 3.3). To incorporate this feature, let us consider a model of random walk with WiW_i defined by

Wi=Zi⋯Zi+p−1,(31)W_i = Z_i \cdots Z_{i+p-1}, \tag*{(31)}

where Zj=eiϵXjZ_j = e^{i\epsilon X_j} and XjX_j is Gaussian random with a unit variance, and take the average of the product of WW’s as before:

⟨W1W2⋯WK⟩.(32)\langle W_1 W_2 \cdots W_K \rangle. \tag*{(32)}

Then, WiW_i changes gradually, and the correlation disappears completely after pp steps. Given the decay of the two-point function of Polyakov loops, ap∼1βσ02ap \sim\frac{1}{\beta\sigma_0^2} is a natural identification. (This means v∼βσ02⋅a0v \sim\beta\sigma_0^2 \cdot a^0, which is consistent with the rough estimate v∼a0v \sim a^0 in § 2.4.) We are interested in the behavior of the correlator as a function of physical distance L=KaL = Ka.

Unfortunately, we were not able to solve this problem analytically. We performed numerical experiments for SU(2) and SU(3). The results are very clean, and we believe our findings apply to other groups as well.

As shown in Figure 5, by taking the horizontal axis to be K⋅(ϵp)2×CrK \cdot(\epsilon p)^2 \times C_{\mathrm{r}}, we can see that exponential decays with different pp, ϵ\epsilon and different representations line up. Therefore, by identifying (ϵp)2(\epsilon p)^2 with βaσ02\beta a\sigma_0^2, we can reproduce e−CrβLσ02e^{-C_{\mathrm{r}}\beta L\sigma_0^2}. This ϵp\epsilon p is a counterpart of Δ\Delta in the vanilla random walk, but now the finite amount of correlation is taken into account.

Plots of normalized Wilson loop traces for SU(2) and SU(3)

Figure 5. Tr⁡⟨W1…WK⟩\operatorname{Tr}\langle W_1\ldots W_K\rangle vs K⋅(ϵp)2×CrK \cdot(\epsilon p)^2 \times C_r for p=10,20,40p = 10, 20, 40 and ϵ2p3=110\epsilon^2p^3 = \frac{1}{10}. [Top] SU(2), spin j=12,1,32,2,52j = \frac{1}{2}, 1, \frac{3}{2}, 2, \frac{5}{2}. [Bottom] SU(3), fundamental, rank-2 symmetric, adjoint, rank-3 symmetric. We normalized Tr⁡⟨W1…WK⟩\operatorname{Tr}\langle W_1\ldots W_K\rangle by dividing dr=Tr⁡1d_r = \operatorname{Tr}1. The number of samples for the averaging is 1 million for both SU(2) and SU(3).

Actually, we observe this clean pattern when ϵ2p3≲1\epsilon^2p^3 \lesssim1. With the indentification (ϵp)2∼βaσ02(\epsilon p)^2 \sim\beta a\sigma_0^2, this means that ϵ2p3∼βσ02×ap≲1\epsilon^2p^3 \sim\beta\sigma_0^2 \times ap \lesssim1, which is consistent with the identification provided above, i.e., ap∼1βσ02ap \sim\frac{1}{\beta\sigma_0^2}.

We emphasize that we only provided a model of the random walk of Polyakov lines satisfying essential features. Specifically, XjX_j may not be exactly Gaussian, or WiW_i may change in a way different from (31).

Simplification for (1 + 1)-dimensional theories

In § 3.2, we have seen that the deviation from vanilla random walk is responsible for the absence of the Casimir scaling at short distance. In fact, however, the situation is different for (1 + 1)-dimensional theories. As commented in § 2.4, (1 + 1)-dimensional theories are special in that vv in (4) does not have to be continuous. On a lattice, the vanilla random walk can be justified for any lattice spacing, as long as the corrections to Haar randomness are neglected. Therefore, quite generally, we expect the Casimir scaling also at short distances.

It is known that the Casimir scaling is exact for (1 + 1)-dimensional pure Yang-Mills theory [26, 27, 28]. With matter fields, infrared behaviors may change drastically. However, we do not expect the short-distance behavior to change, as long as the system remains confined. This provides us with a nontrivial consistency check of the relationship between the Casimir scaling and a random walk on the group manifold.

String breaking and loss of Casimir scaling

String breaking is almost trivial when the corrections to Haar-random distribution are taken into account. If there is no particular global symmetry that forces the Polyakov loop with representation r to become zero, we expect to have a nonzero expectation value of the one-point function of the form

⟨χr(Px⃗)⟩=⟨χr(Px⃗+Lu^)⟩∼e−βmr(33)\langle\chi_r(P_{\vec{x}}) \rangle= \langle\chi_r(P_{\vec{x}+L\hat{u}}) \rangle\sim e^{-\beta m_r} \tag*{(33)}

where mrm_r is the mass of the lowest excitation in representation r. Therefore, we need to take the disconnected part of the two-point function into account:

⟨(χr(Px⃗))∗χr(Px⃗+Lu^)⟩∼e−βCrLσ02+e−2βmr(34)\langle(\chi_r(P_{\vec{x}}))^* \chi_r(P_{\vec{x}+L\hat{u}}) \rangle\sim e^{-\beta C_r L \sigma_0^2} + e^{-2\beta m_r} \tag*{(34)}

where the second term is the disconnected part which is dominant at long distances (CrLσ02≫mrC_r L \sigma_0^2 \gg m_r).

There must be corrections to the connected part as well. For QCD, the details of the corrections to the connected part may not be important because the disconnected part is nonzero for any representations and dominates long-distance physics. The situation is somewhat different for theories with center symmetry such as SU(NN) pure Yang-Mills. As long as the center symmetry is not broken, the disconnected part vanishes. Furthermore, loops with different center-symmetry charges cannot mix with each other, because the connected part must vanish as well. In the random-walk approach, (14) receives some corrections because Px⃗P_{\vec{x}} is not exactly Haar random, and we need to consider a random walk on group manifold with some other weight than the Haar measure. We do not find an immediate reason that forbids the mixing in the same center-symmetry-charge sector, and hence, we expect that string tension takes the same value in the same center-symmetry-charge sector regardless of the representation. Precise dependence on charge requires a better understanding and is out of the scope of this paper. See Refs. [13, 14, 15] for past attempts to build a random-walk model in the long-distance regime.

Discussion

In the low-temperature regime of confining gauge theories such as QCD, Polyakov lines are slowly varying Haar random modulo exponentially small corrections with respect to the inverse temperature [1, 2, 3]. In this paper, we suggested that linear confinement potential with Casimir scaling follows naturally from such a random walk. Strictly speaking, we introduced a model of random walk using a random matrix product. We expect the exponential decay with approximate Casimir scaling to be universal and does not depend much on the details of the random process, while short-distance behaviors depend on details. With exponentially small corrections to Haar randomness, string breaking and loss of Casimir scaling at long distance follow. So, what we have seen is an approximate Casimir scaling in the intermediate distance scale, which is precisely what we expect in confining gauge theories. Our picture can be tested by calculating the Polyakov loops via lattice gauge theory simulation.

The mechanism described for the Polyakov loop may generalize to the case of a rectangular Wilson loop of the size LtL_t for the time direction and LL for the spatial direction. Specifically, we could take a product of Wilson lines like in Figure 6. In this case, (ϵp)2∼Lta(\epsilon_p)^2 \sim L_t a would be expected. For this scenario to work, we need to take LtL_t sufficiently large so that the temporal edge inherits the Haar randomness of the Polyakov loop. Note also that, if LtL_t is too small, the conformal behavior can set in and slowly varying nature can be lost.5 It would be interesting to use lattice simulation to confirm this mechanism.

Figure 6. A Wilson line which can be used to build a rectangular Wilson loop of the size LtL_t for the time direction and LL for the spatial direction.

Note that we essentially assumed the mass gap to claim that the correction to Haar randomness is exponentially small. Therefore, we are assuming one of the important features of confinement, and hence, we do not claim we proved confinement. Still, what we observed is a highly nontrivial consequence of the mass gap, and we can find nontrivial connections between good old results and deepen our understanding of confinement.

More than two decades ago, it was realized that large-NN theory in the weak-coupling limit exhibits confinement/deconfinement transition that resembles some aspects of the counterpart at strong coupling such as the NN-dependence of entropy [29, 30]. Most crucially, deconfinement can be understood as the condensation of long strings both at weak coupling and at strong coupling. However, it was not clear how linear confinement potential can follow from this picture. In the meantime, a series of work on partial deconfinement [17, 18, 19, 20] led to the discovery of a microscopic mechanism of confinement in the large-NN gauge theories [1, 2, 3] which is essentially the same as Bose-Einstein condensation.

(See § 2.) This mechanism is based only on symmetry and is applicable to both weak and strong coupling. Let us call it Confinement = BEC mechanism.6 The slowly varying Haar-randomness is a consequence of the Confinement = BEC mechanism. Therefore, we understood how the Confinement = BEC mechanism, which is based on a very basic property of gauge theory, leads to linear confinement potential with Casimir scaling.

Confinement = BEC mechanism leads to various phenomena including partial deconfinement and intriguing pattern of global symmetry breaking [31, 32], coexistence of parton-like and string-like degrees of freedom [33],7 instanton condensation [1, 2], and so on. Hopefully, this mechanism leads us to a unified picture of an even richer list of nontrivial phenomena.

Acknowledgement

The authors would like to thank Pavel Buividovich, Nadav Drukker, Jeff Greensite, Jack Holden, Enrico Rinaldi, Andreas Schäfer, Hidehiko Shimada, Yuya Tanizaki, Masaki Tezuka, and Hiromasa Watanabe. V. G. thanks STFC for the Doctoral Training Programme funding (ST/W507854-2021 Maths DTP). M. H. thanks his STFC consolidated grant ST/X000656/1. G. B. is funded by the Deutsche Forschungsgemeinschaft (DFG) under Grant No. 432299911 and 431842497.

Path integral for Two-dimensional pure Yang-Mills

In this appendix, we explain how the mechanism we discussed in the main text works for the simple yet illuminating case of pure Yang-Mills theory in two dimensions. Unlike the main text, we consider the spacetime lattice and Euclidean path integral [34, 35]. Note that we are merely rewriting known facts by using slightly different language so that the connection to the discussion in the main text becomes manifest.

Wilson’s plaquette action is given by

SWilson=−∑n⃗12g2a2Tr⁡{U□,n⃗+U□,n⃗†},(35)S_{\mathrm{Wilson}}=-\sum_{\vec n}\frac{1}{2g^{2}a^{2}}\operatorname{Tr}\left\{U_{\square,\vec n}+U_{\square,\vec n}^{\dagger}\right\}, \tag*{(35)}

where the plaquette term U□,n⃗U_{\square,\vec n} is

U□,n⃗=Un⃗,tUn⃗+t^,xUn⃗+x^,t†Un⃗,x†.(36)U_{\square,\vec n}=U_{\vec n,t}U_{\vec n+\hat{t},x}U_{\vec n+\hat{x},t}^{\dagger}U_{\vec n,x}^{\dagger}. \tag*{(36)}

We take the spatial direction to be noncompact. (We allow the temporal direction to be compact because we want to consider finite temperature.) Then, we can choose the gauge such that Un⃗,x=1U_{\vec n,x}=1 for all n⃗\vec n, or in other words Ax=0A_x=0. In this gauge, the plaquette term reduces to U□,n⃗=Un⃗,tUn⃗+x^,t†U_{\square,\vec n}=U_{\vec n,t}U_{\vec n+\hat{x},t}^{\dagger}. The partition function is

Z=∫(∏n⃗dUn⃗,tdUn⃗,x)e−SWilson[Ut,Ux]=∫(∏n⃗dUn⃗,t)e−SWilson[Ut,Ux=1].(37)Z=\int\left(\prod_{\vec n}dU_{\vec n,t}dU_{\vec n,x}\right)e^{-S_{\mathrm{Wilson}}[U_t,U_x]}=\int\left(\prod_{\vec n}dU_{\vec n,t}\right)e^{-S_{\mathrm{Wilson}}[U_t,U_x=1]} . \tag*{(37)}

The integral is taken over the Haar measure. To see that no extra factor appears, we can write the spatial link variables as Un⃗,x=Ωn⃗†Ωn⃗+x^U_{\vec n,x}=\Omega_{\vec n}^{\dagger}\Omega_{\vec n+\hat{x}} (it is possible because the spatial dimension is noncompact), and use the gauge invariance of the action and the property of Haar measure as follows. First, because of the gauge invariance of the action,

SWilson[Ut,Ux]=SWilson[Ut′,Ux′=1],(38)S_{\mathrm{Wilson}}[U_t,U_x]=S_{\mathrm{Wilson}}[U_t',U_x'=1], \tag*{(38)}

where

Un⃗,t′≡Ωn⃗Un⃗,tΩn⃗+t^†,Un⃗,x′≡Ωn⃗Un⃗,xΩn⃗+x^†.(39)U_{\vec n,t}'\equiv\Omega_{\vec n}U_{\vec n,t}\Omega_{\vec n+\hat{t}}^{\dagger},\qquad U_{\vec n,x}'\equiv\Omega_{\vec n}U_{\vec n,x}\Omega_{\vec n+\hat{x}}^{\dagger}. \tag*{(39)}

The Haar measure is also gauge invariant, and hence,

Z=∫(∏n⃗dUn⃗,t′dUn⃗,x′)e−SWilson[Ut′,Ux′=1].(40)Z=\int\left(\prod_{\vec n}dU_{\vec n,t}'dU_{\vec n,x}'\right)e^{-S_{\mathrm{Wilson}}[U_t',U_x'=1]} . \tag*{(40)}

The integral over Un⃗,x′U_{\vec n,x}' gives a trivial volume factor. This is the case not just for partition function but also for evaluation of gauge-invariant observables. Therefore, we drop dUn⃗,x′dU_{\vec n,x}', and removing the prime, we arrive at (37).

Redefining the variables such that Wn⃗≡Un⃗,tUn⃗+x^,t†=U□,n⃗W_{\vec n}\equiv U_{\vec n,t}U_{\vec n+\hat{x},t}^{\dagger}=U_{\square,\vec n}, the action reduces simply to

SWilson=−∑n⃗12g2a2Tr⁡{Wn⃗+Wn⃗†}.(41)S_{\mathrm{Wilson}}=-\sum_{\vec n}\frac{1}{2g^{2}a^{2}}\operatorname{Tr}\left\{W_{\vec n}+W_{\vec n}^{\dagger}\right\}. \tag*{(41)}

The partition function is

Z=∫(∏n⃗dWn⃗)exp⁡{∑n⃗12g2a2Tr⁡{Wn⃗+Wn⃗†}}=∏n⃗(∫dWn⃗exp⁡{12g2a2Tr⁡{Wn⃗+Wn⃗†}}).(42)\begin{aligned} Z&=\int\left(\prod_{\vec n}dW_{\vec n}\right)\exp\left\{\sum_{\vec n}\frac{1}{2g^{2}a^{2}}\operatorname{Tr}\left\{W_{\vec n}+W_{\vec n}^{\dagger}\right\}\right\}\\ &=\prod_{\vec n}\left(\int dW_{\vec n}\exp\left\{\frac{1}{2g^{2}a^{2}}\operatorname{Tr}\left\{W_{\vec n}+W_{\vec n}^{\dagger}\right\}\right\}\right). \tag*{(42)} \end{aligned}

Again, there is no extra factor due to the property of the Haar measure. From this expression, we see that plaquettes Wn⃗W_{\vec n} behave as independent variables.

The expectation value of a Wilson loop Tr⁡WC\operatorname{Tr}W_C along a contour CC can then be calculated as

⟨Tr⁡WC⟩=1Z∫dWWCe−SWilson(43)\langle\operatorname{Tr}W_C\rangle=\frac{1}{Z}\int dW W_C e^{-S_{\mathrm{Wilson}}} \tag*{(43)}

Here, we use WC\mathcal{W}_{C} to denote the loop before taking a trace, which is an N×NN \times N matrix for SU(N)\mathrm{SU}(N) theory. Any Wilson loop WC\mathcal{W}_{C} can be broken down in terms of the plaquettes contained in it, as shown in Figure 7. More precisely, Wilson loops can be expressed as a product of plaquettes with ‘tails’. Specifically in two dimensions, each plaquette Wn⃗W_{\vec{n}} is independent of the others, and hence, the plaquette with ‘tail’ is also independent. Therefore, the Wilson loop in two dimensions factorizes to the product of plaquettes contained in it. Hence, for a Wilson loop consisting of nplaq.n_{\mathrm{plaq.}} plaquettes, we obtain

Figure 7. Wilson loop expressed in terms of a product of plaquettes with ‘tails’.

⟨WC⟩=(w⋅1)nplaq.=wnplaq.⋅1,(44)\langle\mathcal{W}_{C}\rangle= (w \cdot1)^{n_{\mathrm{plaq.}}} = w^{n_{\mathrm{plaq.}}} \cdot1 , \tag*{(44)}

where ww is the expectation value of single plaquette, i.e. ⟨Wn⃗⟩=w⋅1\langle W_{\vec{n}}\rangle= w \cdot1.

When the lattice spacing aa is small, Wn⃗W_{\vec{n}} localizes around 11. Let us introduce a traceless Hermitian matrix Xn⃗X_{\vec{n}} by

Wn⃗=eiaXn⃗.(45)W_{\vec{n}} = e^{iaX_{\vec{n}}}. \tag*{(45)}

By construction, this XX is the same as field strength. Then,

SWilson=12g2∑n⃗Tr⁡Xn⃗2+O(a2).(46)S_{\mathrm{Wilson}} = \frac{1}{2g^{2}}\sum_{\vec{n}}\operatorname{Tr}X_{\vec{n}}^{2}+O(a^{2}). \tag*{(46)}

The Haar measure dWdW is equivalent to the flat measure dXdX. Therefore, Xn⃗X_{\vec{n}} is Gaussian random and the variance is g2g^{2}. The thin loop of the form shown in Figure 4, with one-lattice-unit extension along the spatial direction, is obtained by taking a product of ntn_{t} plaquettes with tails, where ntn_{t} is the number of lattice sites along the temporal direction which is related to temperature by β=T−1=ant\beta= T^{-1} = an_{t}. The sum of ntn_{t} independent Gaussian random numbers with variance g2g^{2} is a Gaussian random number with variance g2ntg^{2}n_{t}. Therefore, the thin loop takes the form of eigantX~=eigβaX~e^{iga\sqrt{n_{t}}\widetilde{X}} = e^{ig\sqrt{\beta a}\widetilde{X}}, where X~\widetilde{X} is Gaussian random with variance 11. This corresponds to Vn−1Vn+1V_{n}^{-1}V_{n+1} in Sec. 2.4. Because all plaquettes are independent, Vn−1Vn+1V_{n}^{-1}V_{n+1}’s are also independent.

References

  1. [1]M. Hanada and H. Watanabe, “On thermal transition in QCD,” arXiv:2310.07533 [hep-th].DOI
  2. [2]M. Hanada, H. Ohata, H. Shimada, and H. Watanabe, “A new perspective on thermal transition in QCD,” arXiv:2310.01940 [hep-th].arxiv.org/abs/2310.01940
  3. [3]M. Hanada, H. Shimada, and N. Wintergerst, “Color confinement and Bose-Einstein condensation,” JHEP 08 (2021) 039, arXiv:2001.10459 [hep-th].DOI
  4. [4]J. Ambjørn, P. Olesen, and C. Peterson, “Stochastic Confinement and Dimensional Reduction. 1. Four-Dimensional SU(2) Lattice Gauge Theory,” Nucl. Phys. B 240 (1984) 189–212.
  5. [5]J. Ambjørn, P. Olesen, and C. Peterson, “Stochastic Confinement and Dimensional Reduction. 2. Three-dimensional SU(2) Lattice Gauge Theory,” Nucl. Phys. B 240 (1984) 533–542.
  6. [6]L. Del Debbio, M. Faber, J. Greensite, and S. Olejnik, “Casimir scaling versus Abelian dominance in QCD string formation,” Phys. Rev. D 53 (1996) 5891–5897, arXiv:hep-lat/9510028.DOI
  7. [7]G. S. Bali, “Casimir scaling of SU(3) static potentials,” Phys. Rev. D 62 (2000) 114503, arXiv:hep-lat/0006022.DOI
  8. [8]S. Deldar, “Static SU(3) potentials for sources in various representations,” Phys. Rev. D 62 (2000) 034509, arXiv:hep-lat/9911008.DOI
  9. [9]S. Gupta, K. Huebner, and O. Kaczmarek, “Renormalized Polyakov loops in many representations,” Phys. Rev. D 77 (2008) 034503, arXiv:0711.2251 [hep-lat].DOI
  10. [10]S. Gupta, K. Huebner, and O. Kaczmarek, “Polyakov loop in different representations of SU(3) at finite temperature,” Nucl. Phys. A 785 (2007) 278–281, arXiv:hep-lat/0608014.DOI
  11. [11]A. Mykkanen, M. Panero, and K. Rummukainen, “Casimir scaling and renormalization of Polyakov loops in large-N gauge theories,” JHEP 05 (2012) 069, arXiv:1202.2762 [hep-lat].DOI
  12. [12]A. M. Brzoska, F. Lenz, J. W. Negele, and M. Thies, “Diffusion of Wilson loops,” Phys. Rev. D 71 (2005) 034008, arXiv:hep-th/0412003.DOI
  13. [13]G. Arcioni, S. de Haro, and P. Gao, “A Diffusion model for SU(N) QCD screening,” Phys. Rev. D 73 (2006) 074508, arXiv:hep-th/0511213.DOI
  14. [14]P. V. Buividovich and V. I. Kuvshinov, “Kramers-Moyall cumulant expansion for the probability distribution of parallel transporters in quantum gauge fields,” Phys. Rev. D 73 (2006) 094015, arXiv:hep-th/0605207.DOI
  15. [15]P. V. Buividovich and M. I. Polikarpov, “Random walks of Wilson loops in the screening regime,” Nucl. Phys. B 790 (2008) 28–41, arXiv:0704.3367 [hep-ph].DOI
  16. [16]E. Poppitz and M. E. Shalchian T., “String tensions in deformed Yang-Mills theory,” JHEP 01 (2018) 029, arXiv:1708.08821 [hep-th].DOI
  17. [17]M. Hanada and J. Maltz, “A proposal of the gauge theory description of the small Schwarzschild black hole in AdS₅×S⁵,” JHEP 02 (2017) 012, arXiv:1608.03276 [hep-th].DOI
  18. [18]D. Berenstein, “Submatrix deconfinement and small black holes in AdS,” JHEP 09 (2018) 054, arXiv:1806.05729 [hep-th].DOI
  19. [19]M. Hanada, G. Ishiki, and H. Watanabe, “Partial Deconfinement,” JHEP 03 (2019) 145, arXiv:1812.05494 [hep-th]. [Erratum: JHEP 10, 029 (2019)].DOI
  20. [20]M. Hanada, A. Jevicki, C. Peng, and N. Wintergerst, “Anatomy of Deconfinement,” JHEP 12 (2019) 167, arXiv:1909.09118 [hep-th].DOI
  21. [21]M. Lüscher, “Properties and uses of the Wilson flow in lattice QCD,” JHEP 08 (2010) 071, arXiv:1006.4518 [hep-lat]. [Erratum: JHEP 03, 092 (2014)].DOI
  22. [22]S. Datta, S. Gupta, and A. Lytle, “Using Wilson flow to study the SU(3) deconfinement transition,” Phys. Rev. D 94 no. 9, (2016) 094502, arXiv:1512.04892 [hep-lat].DOI
  23. [23]P. Petreczky and H. P. Schadler, “Renormalization of the Polyakov loop with gradient flow,” Phys. Rev. D 92 no. 9, (2015) 094517, arXiv:1509.07874 [hep-lat].DOI
  24. [24]A. Di Giacomo, H. G. Dosch, V. I. Shevchenko, and Y. A. Simonov, “Field correlators in QCD: Theory and applications,” Phys. Rept. 372 (2002) 319–368, arXiv:hep-ph/0007223.DOI
  25. [25]J. Greensite, K. Langfeld, S. Olejnik, H. Reinhardt, and T. Tok, “Color Screening, Casimir Scaling, and Domain Structure in G(2) and SU(N) Gauge Theories,” Phys. Rev. D 75 (2007) 034501, arXiv:hep-lat/0609050.DOI
  26. [26]A. A. Migdal, “Recursion Equations in Gauge Theories,” Sov. Phys. JETP 42 (1975) 413.
  27. [27]E. Witten, “On quantum gauge theories in two-dimensions,” Commun. Math. Phys. 141 (1991) 153–209.DOI
  28. [28]M. Nguyen, Y. Tanizaki, and M. Ünsal, “Noninvertible 1-form symmetry and Casimir scaling in 2D Yang-Mills theory,” Phys. Rev. D 104 no. 6, (2021) 065003, arXiv:2104.01824 [hep-th].DOI
  29. [29]B. Sundborg, “The Hagedorn transition, deconfinement and N=4 SYM theory,” Nucl. Phys. B 573 (2000) 349–363, arXiv:hep-th/9908001.
  30. [30]O. Aharony, J. Marsano, S. Minwalla, K. Papadodimas, and M. Van Raamsdonk, “The Hagedorn - deconfinement phase transition in weakly coupled large N gauge theories,” Adv. Theor. Math. Phys. 8 (2004) 603–696, arXiv:hep-th/0310285.DOI
  31. [31]M. Hanada and B. Robinson, “Partial-Symmetry-Breaking Phase Transitions,” Phys. Rev. D 102 no. 9, (2020) 096013, arXiv:1911.06223 [hep-th].DOI
  32. [32]M. Hanada, J. Holden, M. Knaggs, and A. O’Bannon, “Global symmetries and partial confinement,” JHEP 03 (2022) 118, arXiv:2112.11398 [hep-th].DOI
  33. [33]V. Gautam, M. Hanada, J. Holden, and E. Rinaldi, “Linear confinement in the partially-deconfined phase,” JHEP 03 (2023) 195, arXiv:2208.14402 [hep-th].DOI
  34. [34]D. J. Gross and E. Witten, “Possible Third Order Phase Transition in the Large N Lattice Gauge Theory,” Phys. Rev. D 21 (1980) 446–453.
  35. [35]S. R. Wadia, “A Study of U(N) Lattice Gauge Theory in 2-dimensions,” arXiv:1212.2906 [hep-th].

Paper details

vlnc:W_rgu8755d
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents