Replica Phase Transition with Quantum Gravity Corrections

Abstract: Motivated by bulk replica wormholes, we study the boundary effective theory that describes the near-horizon fluctuations of a near-extremal Reissner-Nordström black hole. This theory consists of a Schwarzian mode and a $U(1)$ phase mode. We compute the partition function of this boundary theory on replica geometries, from which the entropy is derived. Our analysis reveals a rich phase structure, in which the dominance of connected or disconnected replica configurations leads to a phase transition controlled by the temperature and the coupling constants $C$, $K$, and $\mathcal{E}$ of the 1d effective theory.

• The paper analyzes the replica wormhole mechanism in near-extremal Reissner-Nordström black holes, identifying a rich phase structure controlled by inverse temperature β, and three couplings, C, K, and ε
• Key findings show that the connected vs. disconnected saddle dominance is determined by specifics of Schwarzian-U(1) coupling, not requiring auxiliary entangled systems.
• The paper highlights a possible unified boundary description for AdS2 and dS2 JT gravity, conditioning quantum gravity transitions.

Overview

This paper studies the replica wormhole mechanism for near-extremal Reissner–Nordström (RN) black holes entirely within the one-dimensional boundary effective theory, rather than in the two-dimensional bulk. The effective theory consists of a Schwarzian mode coupled to a $U(1)$ phase mode, dual to AdS$_2$ JT gravity coupled to a Maxwell theory in the near-horizon region (Sachdev, 2019, Iliesiu et al., 2020). The central result is a rich phase structure: the dominance of connected versus disconnected replica saddles — and hence whether the von Neumann entropy is nonzero or vanishes — is controlled by the inverse temperature $\beta$ and by the three couplings $C$, $K$, and $\mathcal{E}$ of the boundary theory.

A distinguishing feature of this construction is that no auxiliary entangled system is required. Prior analyses of replica-wormhole phase transitions typically introduce an outer entangling system such as a thermal bath or end-of-the-world branes (Penington et al., 2019, Almheiri et al., 2019, Goto et al., 2020, Marolf et al., 2020). Here the transition arises purely from competition among the intrinsic couplings of the Schwarzian–$U(1)$ system.

Effective theory and its parameters

The boundary action comprises a Schwarzian term with coupling $C$, a $U(1)$ kinetic term with coupling $K$, and a chemical-potential-like coupling $_2$0 that couples the phase mode to the reparametrization field:

$_2$1

The couplings are fixed by dimensional reduction of the higher-dimensional Einstein–Maxwell action in terms of the gravitational constant $_2$2, Yang–Mills coupling $_2$3, extremal horizon radius $_2$4, and AdS radius $_2$5. The four independent parameters $_2$6 are equivalent to $_2$7.

A key structural result is that the partition function factorizes,

$_2$8

with the $_2$9 piece given exactly by a Jacobi theta function, $\beta$0. This factorization makes the replica computation tractable mode by mode.

For AdS$\beta$1 ($\beta$2), the zero-temperature (extremal) entropy takes the closed form

$\beta$3

which is real and positive only when $\beta$4. Below this bound $\beta$5 becomes imaginary, signaling a quantum phase transition of the topological term at $\beta$6, interpreted as a transition between AdS$\beta$7 and dS$\beta$8 JT gravity descriptions, whose topological terms are real and imaginary respectively (Cotler et al., 2019, Maldacena et al., 2019). The authors note this as suggestive rather than established: it indicates a possible unified boundary description of de Sitter and anti-de Sitter gravity, but the connection is left as an open question.

Replica partition functions

The Schwarzian contribution on the quotient manifold $\beta$9 follows the standard treatment (Penington et al., 2019); expanding to first order in $C$0 gives

$C$1

where $C$2 is the geodesic distance between the $C$3-fixed points.

The novel ingredient is the $C$4 phase mode on replica geometries. The authors compute the bulk 2d Maxwell partition function on the $C$5-replica surface via the standard 2d Yang–Mills sum over representations [hep-th/9411210, (Blommaert et al., 2018)]. For the disconnected geometry (a union of $C$6 caps), the result is $C$7; for the connected $C$8-boundary wormhole it is $C$9. Assuming equality of bulk and boundary partition functions, and using the trumpet approximation to the quotient geometry with parameter $K$0, the area ratio evaluates to

$K$1

This yields the boundary $K$2 free energy on the connected replica geometry, expanded around $K$3 in terms of derivatives of $K$4 evaluated at $K$5 and $K$6. The total on-shell action reaches its extremum at $K$7, consistent with the standard replica wormhole saddle.

Entropy from the replica trick

Applying the replica trick $K$8, the connected-saddle entropy is

$K$9

while the disconnected entropy vanishes identically, since $\mathcal{E}$0. The sign of $\mathcal{E}$1 determines which saddle dominates: when $\mathcal{E}$2 the disconnected configuration has lower free energy and governs the entropy (which is then zero); when $\mathcal{E}$3 the connected wormhole dominates and the entropy equals $\mathcal{E}$4. In the low-temperature limit $\mathcal{E}$5, this simplifies to

$\mathcal{E}$6

Phase structure

Three distinct transition mechanisms emerge.

Transition in temperature. At fixed $\mathcal{E}$7, $\mathcal{E}$8, and $\mathcal{E}$9, the quantity $U(1)$0 decays monotonically from $U(1)$1 to $U(1)$2 as $U(1)$3 increases from $U(1)$4 to $U(1)$5, so it necessarily crosses $U(1)$6 at some finite $U(1)$7. High temperatures therefore favor the disconnected phase with vanishing entropy, while low temperatures favor the connected wormhole-dominated phase. This is the direct analogue of the Page-type transition, realized here without any external radiation system.

Transition in chemical potential $U(1)$8. As $U(1)$9 increases from $C$0, the extremal entropy $C$1 first decreases from $C$2 to a minimum $C$3 at $C$4, then grows linearly for large $C$5. Meanwhile $C$6 grows exponentially in $C$7 through the $C$8 and $C$9 terms. Since exponential growth eventually overtakes linear growth, there must be at least one connected–disconnected transition as $U(1)$0 ranges over $U(1)$1: small $U(1)$2 favors the connected phase, large $U(1)$3 favors the disconnected phase.

Transitions in $U(1)$4. For large $U(1)$5, $U(1)$6 is suppressed relative to $U(1)$7 everywhere in the allowed interval $U(1)$8, so the connected phase dominates with no transition. For smaller $U(1)$9, up to two transitions can occur: starting from the connected phase at $K$0, increasing $K$1 drives the system to the disconnected phase, and approaching the upper bound $K$2 returns the system to the connected phase because $K$3 diverges there.

Quantum corrections

Including the known one-loop corrections to the Schwarzian disk and trumpet partition functions (Saad et al., 2019, Mertens et al., 2022, Turiaci, 2024) adds logarithmic terms,

$K$4

These corrections constrain the validity of the temperature-driven transition. In the regime $K$5, the ratio $K$6 cannot be taken arbitrarily large; it truncates at $K$7. The corrected combination $K$8 then only decreases to a minimum of approximately $K$9 at $_2$00. If additionally $_2$01, the would-be critical temperature satisfies $_2$02, placing the transition outside the semiclassical working regime; only when $_2$03 does the logarithmic correction merely shift the tree-level transition point. This is a genuine limitation of the analysis: the existence and location of the thermal transition depend on the hierarchy between $_2$04 and $_2$05 in the strongly quantum regime.

Limitations and open questions

Several assumptions underpin the results. First, the derivation of the boundary $_2$06 replica partition function relies on the assumption that the bulk Maxwell partition function equals the boundary phase-mode partition function on replica geometries; this equality is inferred rather than proven from first principles. Second, the explicit closed-form expression for $_2$07 is available only for AdS$_2$08 ($_2$09); the general-$_2$10 case is stated to be "rather complicated" and is not worked out. Third, the interpretation of the $_2$11 point as an AdS$_2$12/dS$_2$13 transition rests on the reality properties of the topological term and remains conjectural. Finally, the analysis is performed at leading order in the $_2$14 expansion, with higher-order terms in $_2$15 neglected throughout.

Conclusion

By computing the replica partition function of the coupled Schwarzian–$_2$16 boundary theory of near-extremal RN black holes, this work derives the black hole entropy from a purely boundary perspective and establishes a multi-parameter phase structure governing connected versus disconnected replica saddles. The transitions are driven by temperature, by the chemical-potential coupling $_2$17, and by the ratio $_2$18, all without introducing external entangled systems. The main open questions are the status of the assumed bulk–boundary equality on replica geometries, the generalization beyond AdS$_2$19, and whether the $_2$20 singularity indeed provides a unified boundary description of AdS$_2$21 and dS$_2$22 JT gravity.