Vortex screening and the fate of the BKT transition with long-range couplings

Technical Summary: Vortex Screening and the Fate of the BKT Transition with Long-Range Couplings

Problem Statement
The stability of the Berezinskii–Kosterlitz–Thouless (BKT) phase and its associated quasi-long-range order (QLRO) in the presence of long-range (LR) interactions remains an open and debated problem in two-dimensional statistical mechanics. While short-range systems with O(2) symmetry exhibit a BKT transition driven by the unbinding of vortex–antivortex pairs, the fate of this mechanism when couplings decay as $r^{-d-\sigma}$ is unclear. Previous studies have yielded conflicting results: some perturbative renormalization group (RG) analyses suggest an intermediate QLRO phase persists for arbitrarily small $\sigma$, while Monte Carlo simulations on lattices and diluted graphs report direct transitions to ordered or disordered phases, or logarithmic QLRO at specific thresholds. A central difficulty lies in the interplay between topological defects (vortices) and the non-local nature of LR couplings, which prevents a simple factorization of spin-wave and vortex contributions.

Methodology
The author investigates the two-dimensional XY model with ferromagnetic O(2)-symmetric couplings decaying as $J(r) \sim r^{-2-\sigma}$ ($\sigma > 0$). The Hamiltonian is decomposed into a short-range gradient term ($S_{sr}$) and a long-range cosine term ($S_g$). The study employs a real-space renormalization-group (RG) treatment near the Gaussian manifold of QLRO states.

Key methodological features include:

1. Full Many-Body Vortex Interaction: Unlike previous approaches that truncate the vortex Hamiltonian or RG-generated couplings, this work retains the full many-body vortex interaction. The field configuration is decomposed into a single-valued spin-wave part ($\theta_{sw}$) and a topological part involving integer point charges (vortices).
2. Vortex Pair Integration: The author integrates out vortex–antivortex pairs with separations $1 < |\xi| < e^{\delta \ell}$. Crucially, they demonstrate that the screening contribution by these small pairs preserves the Gaussian-average representation of the partition function at leading order in the fugacity ($y$), without requiring an expansion in the long-range coupling strength.
3. Perturbative Spin-Wave Integration: Following the vortex renormalization, spin-wave contributions are integrated perturbatively to first order in the LR coupling $g$. This allows for the derivation of flow equations for the spin stiffness $J$, the fugacity $y$, and the LR coupling $g$.

Key Contributions and Results
The primary technical contribution is the derivation of a local stability criterion for the Gaussian QLRO phase that is non-perturbative in the form of the vortex interaction.

1. Preservation of Gaussian Form: The author shows that screening by small vortex–antivortex pairs renormalizes the spin stiffness $J$ exactly as in the short-range Coulomb gas, regardless of the specific form of the long-range action $S_g$. This allows the RG flow to be closed without introducing uncontrolled truncations of the vortex interaction.
2. Stability Criterion ($\sigma > 7/4$): By analyzing the linear stability of the Gaussian fixed line ($g=0, y=0$), the author derives a condition for the existence of a stable QLRO interval. The analysis reveals that a stable Gaussian QLRO phase exists only for $\sigma > 7/4$. For $\sigma < 7/4$, the Gaussian line becomes unstable, ruling out a stable Gaussian QLRO phase in this regime within the retained approximation.
3. Finite-Coupling Fixed Point: For $\sigma < 7/4$, the RG flow predicts a Wilson–Fisher-like fixed point at finite coupling ($J^*, g^*, y^*$). This fixed point merges with the Gaussian line at $\sigma = 7/4$ and the BKT point $J_{BKT} = 2/\pi$.
4. Critical Exponents: Near the merging point $\sigma \to 7/4$, the correlation length exponent $\nu$ scales as $\nu^{-1} \sim \sqrt{7/4 - \sigma}$. This square-root scaling parallels that of the one-dimensional long-range Ising model near its $1/r^2$ interaction limit.
5. Resolution of Discrepancies: The results suggest that the intermediate QLRO phase, if it exists for $\sigma < 7/4$, is extremely narrow or potentially bypassed by finite-coupling corrections, explaining the difficulty in observing it in numerical simulations. The contrasting results in previous literature are attributed to truncations of the vortex interaction or the generation of couplings that were subsequently set to zero.

Significance and Claims
The paper provides a self-consistent description of vortex screening that does not rely on uncontrolled approximations of the many-body vortex interaction, offering a refined perspective on the debate regarding the stability of the BKT transition in long-range systems. The central claim is that the stability of the Gaussian QLRO manifold is strictly bounded by $\sigma = 7/4$. Below this threshold, the mechanism for Gaussian QLRO is destabilized by the interplay of long-range forces and vortex proliferation, leading to a direct order–disorder transition described by a finite-coupling critical point.

The author emphasizes that their approach differs substantially from previous treatments (e.g., Refs. [62, 64]) which suggested QLRO persistence at smaller $\sigma$. By retaining the full form of the vortex interaction under renormalization, the current work establishes that the screening mechanism fundamentally alters the stability landscape, placing the lower threshold for an open interval of stable Gaussian fixed points at $\sigma = 7/4$. The findings offer a route to reconcile conflicting numerical results by suggesting that the intermediate phase may be too narrow to be resolved or that strong-coupling fixed points limit its basin of attraction.