---
title: Monopole-Antimonopole Correlators in nc-QED₃
url: https://www.emergentmind.com/topics/monopole-antimonopole-two-point-function
type: topic
---

# Monopole-Antimonopole Correlators in nc-QED₃

The monopole–antimonopole two-point function encodes the correlation between topological disorder operators that introduce quantized units of $U(1)$ magnetic flux in three-dimensional parity-invariant non-compact quantum electrodynamics (nc-QED$_3$). This two-point function provides direct access to the scaling dimension $\Delta$ of monopole operators at the infrared fixed point, a key nonperturbative observable of the theory's conformal phase structure and critical behavior. Its precise determination offers a stringent test of analytical predictions in the large-$N$ expansion, as well as insight into the universality class of nc-QED$_3$ with massless two-component fermions.

## 1. Definition of Monopole Operator and the Background Gauge Construction

In continuum nc-QED$_3$, a monopole operator $M_Q$ of charge $Q$ is defined to insert $Q$ units of $U(1)$ magnetic flux at a spacetime point. It acts as a source of the topological current $j^{{\rm top}}_\mu = \epsilon_{\mu\nu\rho} F^{\nu\rho}/(4\pi)$, directly coupling to the topological sector of the gauge theory.

To compute the monopole–antimonopole correlator on a Euclidean three-torus $L^3$, a classical background gauge field configuration $\mathcal{A}_\mu^{Q\bar{Q}}(x)$ is constructed so that it contains the appropriate flux from a monopole at position $y$ and an antimonopole at position $y'$, separated by $T = |y - y'|$. This configuration is determined by minimizing the Villain action,
$$
S_v^{Q\bar{Q}}[\mathcal{A}] = \sum_{x, \mu<\nu} \left[\nabla_\mu \mathcal{A}_\nu - \nabla_\nu \mathcal{A}_\mu - 2\pi N^{Q\bar{Q}}_{\mu\nu}(x)\right]^2,
$$
where the integer-valued plaquette fields $N^{Q\bar{Q}}_{\mu\nu}(x)$ are constrained by
$$
\frac{1}{2}\epsilon_{\mu\nu\rho}\nabla_\mu N^{Q\bar{Q}}_{\nu\rho}(x) = Q \delta_{x,y} - Q \delta_{x,y'}.
$$
In the continuum limit, $\mathcal{A}_\mu^{Q\bar{Q}}(x; \tau)$ reduces to the difference of two Dirac monopole gauge potentials, ensuring that the net flux through a surrounding sphere is $Q$ at $x_0$ and $-Q$ at $x_0+\tau\hat{t}$.

## 2. Partition Functions, Correlators, and Free Energy Formalism

With a fixed monopole–antimonopole background, the deformed partition function is
$$
Z_Q = \prod_{x, \mu} \int_{-\infty}^\infty d\theta_\mu(x)\, \det^{N/2}[\slashed{C}^\dagger(U)\slashed{C}(U)]\, \exp\left\{-\frac{L}{\ell} \sum_{x, \mu<\nu}[F_{\mu\nu}(x) - B^{Q\bar{Q}}_{\mu\nu}(x)]^2 \right\},
$$
where $N$ is the number of two-component massless fermion flavors.

The bare monopole–antimonopole two-point function at separation $\tau = T a$ (with $a = \ell/L$) is given by
$$
G_B^{(Q)}(\tau, \ell, a) = \langle M_Q(0) M_Q^\dagger(\tau) \rangle_{\rm bare} = \frac{Z_Q}{Z_0}.
$$
The corresponding bare free energy is defined as $F_B(\tau, \ell, a) = -\ln G_B(\tau,\ell,a)$. To determine $F_B$ numerically, the “flux ramp” method integrates the observable $W(\zeta)=\partial_\zeta\ln Z_\zeta$ over a smooth interpolation parameter $\zeta \in [0,1]$,
$$
F_B = \int_0^1 d\zeta\, W(\zeta), \quad W(\zeta) = \frac{2L}{\ell}\left\langle \sum_x [F_{\mu\nu}(x) - \zeta B^{(1)}_{\mu\nu}(x)] B^{(1)}_{\mu\nu}(x) \right\rangle_\zeta.
$$

The renormalized correlator $G_R(\tau, \ell)$ is defined so as to remove the power-law divergence associated with the "naive" Gaussian fixed-point scaling,
$$
G_R = (\ell/L)^{-2d(L)}\, G_B,
$$
where $d(L)$ is determined by the free-theory fit at $\ell \to 0$. The renormalized free energy is $F_R(\ell) = -\ln G_R(\ell)$.

## 3. Asymptotic Behavior and Extraction of the Scaling Dimension

At large separation $\tau$, conformal invariance predicts a logarithmic dependence for the renormalized correlator's free energy:
$$
F(\tau) = 2\Delta \ln \tau + {\rm const},
$$
or, equivalently,
$$
\langle M(0) M^\dagger(\tau) \rangle_R \sim \tau^{-2\Delta}.
$$
Thus, the scaling dimension $\Delta$ of the monopole operator is related to the slope of $F(\tau)$ versus $\ln \tau$ as
$$
\Delta = \frac{1}{2} \frac{dF}{d\ln\tau}.
$$
A fit to this form in the asymptotic region enables a direct numerical determination of $\Delta$.

## 4. Numerical Methodology and Lattice Implementation

The computation employs three-dimensional cubic lattices of sizes $L = 16, 20, 24, 28$, with the physical box size $\ell$ varied continuously from $\ell = 1$ up to $\ell \approx 250$, corresponding to lattice spacings as fine as $a = \ell / L = 1/28$. Fermionic degrees of freedom consist of $N = 2, 4, 12$ two-component massless flavors, realized by a single-level HYP-smeared Wilson–SW (Sheikholeslami–Wohlert) fermion operator ensuring parity invariance.

Background flux insertion is achieved by interpolating the monopole–antimonopole flux through 24 intermediate values of the ramp parameter $\zeta \in [0,1]$. For each $\zeta$, the partition function $Z_\zeta$ is sampled using Hybrid Monte Carlo (HMC) for $50\,000$ trajectories. The observable $W(\zeta)$ is evaluated in each sample and integrated numerically (using jackknife blocks to estimate autocorrelation and statistical error) to extract $F_B$.

Renormalization is performed by fitting the bare free energy to $2d(L)\ln L+{\rm const}$ at the Gaussian fixed point and subtracting this leading divergence to yield the physically meaningful scaling dimension, as encapsulated in the renormalized correlator $G_R$. For $N=12$, a direct fit of $F_R(\ell)$ versus $\ln \ell$ in the range $\ell \geq 64$ up to $\ell \approx 250$ is employed. For $N=2,4$, fitting the difference $\Delta F(\ell; N) - \Delta F(\ell; N'=12)$ versus $\ln \ell$ is preferred due to improved numerical stability.

## 5. Scaling Dimension Results and Comparison to Large-$N$ Predictions

Analytical large-$N$ expansion predicts a leading “free-fermion” scaling
$$
\Delta_{\rm free}(N) = 0.265\,N.
$$

The numerical study yields the following results:
- For $N=12$: $\Delta(12)/12 = 0.261(19)$, leading to $\Delta(12) = 3.13(23)$, in agreement with the prediction $3.18$ from the large-$N$ line.
- For $N=4$: Using difference fits, $\Delta(4)/4 - \Delta(12)/12 = 0.053(6)$, so that $\Delta(4)=1.256(48)$, contrasting the free-fermion value of $1.06$.
- For $N=2$: $\Delta(2)/2 - \Delta(12)/12 = 0.153(9)$, yielding $\Delta(2)=0.828(54)$, just above the large-$N$ prediction of $0.53$.

These results indicate that for $N \lesssim 4$, higher order corrections in $1/N$ become mildly important, and the scaling dimensions are systematically above the large-$N$ line. This suggests the presence of positive $1/N^2$ or higher-order terms not captured by the leading free-fermion analysis.

| $N$ | $\Delta(N)$ (numerical) | $\Delta_{\rm free}(N)$ | Difference |
|----|--------------------|----------------|------------|
| 2  | $0.83$             | $0.53$         | $+0.30$    |
| 4  | $1.26$             | $1.06$         | $+0.20$    |
| 12 | $3.13$             | $3.18$         | $-0.05$    |

## 6. Significance and Implications

The precision determination of the monopole–antimonopole two-point function in nc-QED$_3$ establishes the viability of nonperturbative Monte Carlo methods for extracting topologically nontrivial operator data in strongly interacting $2+1$ dimensional gauge theories. Agreement with the large-$N$ prediction at $N=12$ provides nontrivial evidence for the reliability of $1/N$ expansion in capturing the spectrum of topological defect operators. The observed mild positive deviations at small $N$ highlight the importance of subleading corrections and potential limitations of large-$N$ extrapolation for finite $N$.

A plausible implication is that higher-order operator dimensions in gauge theories relevant to quantum critical points and dualities in condensed matter systems can be systematically accessed via lattice implementations of background defect insertions and careful renormalization. This approach opens avenues for detailed studies of nonlocal operators, critical exponents, and universality classes beyond perturbation theory in three-dimensional quantum field theories [1908.05500].

Source: https://www.emergentmind.com/topics/monopole-antimonopole-two-point-function