Replica Phase Transition with Quantum Gravity Corrections Jun Nian 1, Yuan Zhong 1,2 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 ) 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 C , K K , and ℰ \mathcal{E} of the 1d effective theory. Introduction.— The study of black hole thermodynamics has revealed that the low-energy dynamics of near-horizon near-extremal Reissner-Nordström (RN) black holes is governed by the AdS 2 JT gravity coupled to a Maxwell theory describing the near-horizon fluctuations Cadoni:2008mw ; Almheiri:2016fws ; Moitra:2018jqs ; Iliesiu:2019lfc ; Iliesiu:2020qvm . This 2d theory, which is dual to a 1d boundary effective theory Sachdev:2019bjn ; Iliesiu:2020qvm ; Davison:2016ngz ; Gaikwad:2018dfc consisting of a Schwarzian mode (encoding gravitational degrees of freedom) and a U(1) phase mode (encoding electromagnetic fluctuations), has become a pivotal model for understanding quantum gravity in low dimensions. Our research is motivated by the 2d bulk replica wormhole mechanism Penington:2019kki , but is executed entirely within this 1d boundary effective theory. We compute the partition function of the coupled Schwarzian and U(1) theory on replica geometries, thereby deriving the entropy from the boundary perspective. This analysis uncovers a rich phase structure, with transitions between connected and disconnected saddles controlled by temperature and effective theory coupling constants C , K , and \mathcal{E} . The novel feature of our work is that, compared to known replica wormhole-sourced phase transitions between connected and disconnected replica geometries, we do not require any additional outer entangled system, such as a thermal bath or end-of-world branes Almheiri:2019qdq ; Penington:2019kki ; Goto:2020wnk ; Marolf:2020xie . Instead, we study a system only consisting of Schwarzian modes and U(1) gauge modes. The phase transition originates from the competition among coupling constants. Effective Theory of Near-Horizon Near-Extremal RN Black Hole.— In this section, we quickly review the effective action describing the near-horizon fluctuations of a near-extremal AdS d+2 RN black hole Sachdev:2019bjn , i.e., the 2d JT gravity coupled to the Maxwell theory. This 2d theory is dual to the boundary 1d effective theory \displaystyle I_{\text{eff}} \displaystyle=-S_{0}-C\int_{0}^{\beta}d\tau\{\tan\pi Tf(\tau),\tau\} \displaystyle\quad+\frac{K}{2}\int_{0}^{\beta}d\tau\left(\partial_{\tau}\phi-i(2\pi T{\mathcal{E}})\partial_{\tau}f\right)^{2}\,, (1) where the effective theory couplings C , K , \mathcal{E} obtained from dimensional reduction are related to the higher-dimensional couplings as C=\frac{s_{d}\Phi_{1}}{\kappa^{2}}=\frac{ds_{d}R_{h}^{d-1}R_{2}^{2}}{\kappa^{2}}\,, (2) K=\frac{(d-1)s_{d}R_{h}^{d-3}[d(d+1)R_{h}^{2}+(d-1)^{2}L^{2}]}{(d+1)g_{F}^{2}}\,, (3) {\mathcal{E}}=\frac{g_{F}R_{h}L\sqrt{d\left[(d+1)R_{h}^{2}+(d-1)L^{2}\right]}}{\kappa\left[d(d+1)R_{h}^{2}+(d-1)^{2}L^{2}\right]}\,, (4) with R_{2}=\frac{LR_{h}}{\sqrt{d(d+1)R_{h}^{2}+(d-1)^{2}L^{2}}} (5) denoting the AdS 2 radius and L standing for the AdS d+2 radius. Here, R_{h} is the horizon radius of the extremal AdS d+2 RN black hole. This theory has four independent parameters (C,K,{\mathcal{E}},\beta) , or equivalently, (\kappa,g_{F},\mu_{0},\beta) , where \kappa , g_{F} , \mu_{0} , \beta denote the gravitational constant, the Yang-Mills coupling constant, the chemical potential, and the inverse temperature, respectively. At finite temperature, the AdS d+2 RN black hole’s entropy is S(T,\mu)=\frac{2\pi s_{d}}{\kappa^{2}}r_{0}(T,\mu)^{d}\,, (6) where r_{0}(T,\mu) is the horizon radius of the AdS d+2 RN black hole at finite temperature. When T=0 , the entropy becomes \displaystyle S_{0}\equiv S(T=0,\mu_{0}) \displaystyle=\frac{2\pi s_{d}}{\kappa^{2}}R_{h}^{d}\,, (7) where the extremal black hole entropy S_{0} depends on the parameters C , K , and {\mathcal{E}} . Using ( 2 ), ( 3 ) and ( 4 ), we can express \kappa , g_{F} , and R_{h} in terms of C , K and \mathcal{E} . Substituting the expression of R_{h} back into ( 7 ), we can express S_{0} in terms of the boundary theory parameters. This is doable, but the result is rather complicated for AdS d+2 with general dimensions. From now on, we focus on the special case AdS 4 with d=2 . In this case, the zero-temperature entropy S_{0} reads S_{0}^{d=2}=\frac{\sqrt{3}\pi\sqrt{C}(\mathcal{E}^{2}K+C)}{L\sqrt{\mathcal{E}^{2}K-C}}\,, (8) which is real and positive, provided {\mathcal{E}}^{2}>C/K. (9) If the lower bound ( 9 ) on \mathcal{E} is violated, the entropy S_{0} will become imaginary. This suggests that there could be a quantum phase transition of the topological term S_{0} in the boundary theory ( Replica Phase Transition with Quantum Gravity Corrections ) at {\mathcal{E}}^{2}=C/K . This phase transition appears to be between AdS 2 JT gravity and dS 2 JT gravity. They both admit the Schwarzian effective theory descriptions, but the S_{0} term is real for AdS while imaginary for dS Cotler:2019nbi ; Maldacena:2019cbz . A nice feature of the effective theory is that the phase mode and the Schwarzian mode decouple in the partition function Mertens:2019tcm : Z=Z_{\text{SL}(2)}[C,\beta]\,\,Z_{U(1)}[K,{\mathcal{E}},\beta]\,, (10) where {\mathcal{E}} only appears only in the phase-mode part Z_{U(1)} through the chemical potential as Z_{U(1)}=\theta_{3}\left(e^{-\beta/K},\,e^{-2\pi{\mathcal{E}}}\right)\,. (11) When \beta/K\gg 1 , or equivalently q=e^{-\beta/K}\to 0 , the function \theta_{3} has the expansion: \theta_{3}(q,e^{2\pi iz})=1+2q\cos(2z)+O(q^{2}) . Therefore, Z_{U(1)}=1+2e^{-\beta/K}\cosh(2{\mathcal{E}})+O(e^{-2\beta/K}). (12) Partition Function of Effective Theory on Replica Geometry.— Our goal is to provide the replica partition function for the boundary effective theory ( Replica Phase Transition with Quantum Gravity Corrections ) in the limit n\to 1 . For the Schwarzian part, this procedure has already been done in Penington:2019kki . The n\to 1 on-shell action for the Schwarzian field on the quotient manifold M_{n}/Z_{n} is I^{M_{n}/Z_{n}}_{\text{Sch}}=-\frac{2\pi^{2}C}{\beta}-\frac{8\pi^{2}C}{\beta}\cosh\rho\,(n-1)+O((n-1)^{2})\,, (13) where 2\rho is the geodesic distance between the two Z_{n} -fixed points. The action on the original replica geometry M_{n} before the Z_{n} quotient is \displaystyle I^{n\to 1}_{\text{Sch}} \displaystyle=\left[nI^{M_{n}/Z_{n}}_{\text{Sch}}\right]_{n\to 1} \displaystyle=-\frac{2\pi^{2}C}{\beta}-\frac{2\pi^{2}C}{\beta}(1+4\cosh\rho)\,(n-1)+O((n-1)^{2})\,. (14) The remaining task is to do a similar calculation for the U(1) phase mode. To achieve this goal, we use the following trick. We first calculate the bulk U(1) partition function on the replica geometry and find its dependence on the replica parameter n . With the assumption that the bulk partition function equals the boundary partition function, we can replace the 2d YM parameters by the boundary effective theory parameters to obtain the boundary U(1) partition on the replica geometry. Let us consider an n -replica geometry. For the 2d Maxwell theory, the partition function on this n -hole surface can be obtained from the standard 2d YM method by summing over representations of U(1) , which equals a sum over integers Cordes:1994fc ; Blommaert:2018oue . For the disconnected n -replica geometry, the topology is just n copies of caps (see Fig. 1 ). The partition function on this geometry is the n -th power of the partition function on a single cap with chemical potential \mu , which is related to the gauge field A_{\tau}\big|_{\partial}=i\mu . The corresponding holonomy around the boundary thermal circle is U=e^{\mu\beta} . Hence, \left[Z(a_{1},U)\right]^{n}=\left[\sum_{q\in\mathbb{Z}}e^{\mu\beta q}e^{-\frac{e^{2}}{2}a_{1}q^{2}}\right]^{n}=\theta_{3}\left(e^{-{e^{2}a_{1}}},\,e^{-\mu\beta}\right)^{n}\,, (15) where a_{1}(\beta) is the area of a hyperbolic cap with the thermal circle period \beta . Figure 1 : The connected and disconnected replica geometries For the connected n -replica wormhole, the topology is a hyperbolic surface with n circle boundaries with the asymptotic length \beta for each (see Fig. 1 ). The U(1) partition function for this case is Z[a_{n},U,\cdots,U]=\sum_{q}\left(e^{\mu\beta q}\right)^{n}\,e^{-\frac{e^{2}}{2}a_{n}q^{2}}=\theta_{3}\left(e^{-{e^{2}a_{n}}},e^{-n\mu\beta}\right)\,, (16) where a_{n}(\beta) is the area of the n -connected surface described above. Comparing these expressions with the boundary U(1) partition function ( 11 ), we infer the replica U(1) partition function in terms of boundary effective theory parameters as Z_{U(1)}^{M_{n}}=\theta_{3}\left(e^{-\frac{\beta}{K}\frac{a_{n}}{a_{1}}},e^{-2\pi n{\mathcal{E}}}\right)\,. (17) Assuming that the bulk partition function equals the boundary effective theory partition function, we obtain the boundary phase mode’s partition function from the bulk Maxwell partition function in terms of boundary quantities. In addition, to calculate the entropy, we also employ the replica trick and expand the partition function in the limit n\to 1 . \displaystyle Z^{n\to 1}_{U(1)} \displaystyle=\theta_{3}\left(e^{-\frac{\beta}{K}\frac{a_{n}}{a_{1}}},e^{-2\pi n{\mathcal{E}}}\right) \displaystyle=\theta_{3}\left(e^{-\frac{\beta}{K}},e^{-2\pi{\mathcal{E}}}\right)+\Bigg[\partial_{q}\theta_{3}(q,\eta)\,e^{-\frac{\beta}{K}}\left(-\frac{\beta}{K}\right)\frac{a_{n}-a_{1}}{a_{1}} \displaystyle\qquad+\partial_{\eta}\theta_{3}(q,\eta)\,e^{-2\pi{\mathcal{E}}}\,(-2\pi{\mathcal{E}})\,(n-1)\Bigg]+O((n-1)^{2})\,, (18) where the differentials are evaluated at q=e^{-\beta/K} , \eta=e^{-2\pi{\mathcal{E}}} . To determine the ratio of areas a_{n} over a_{1} . Recall that the quotient geometry M_{n}/Z_{n} can be approximated by the trumpet geometry Penington:2019kki (see Fig. 2 ): ds^{2}=d\sigma^{2}+\cosh^{2}{\sigma}\,\frac{b^{2}}{(2\pi)^{2}}\,d\theta^{2}\,, (19) where the trumpet parameter b is related to the geodesic distance \rho between fixed points as \cosh\left(\frac{b}{4}\right)=\sin\left(\frac{\pi}{n}\right)\,\cosh(\rho)\,, (20) and in the limit n\to 1 b=2\pi i-4\pi i\cosh(\rho)\,(n-1)+O((n-1)^{2})\,. (21) Figure 2 : The quotient geometry M_{n}/Z_{n} can be treated as a trumpet geometry For the ratio of areas, a(M_{n})/a(M_{1}) , near n=1 , note that a(M_{n})=n\,a(M_{n}/Z_{n})=n\int_{0}^{2\pi}d\theta\int_{0}^{\infty}d\sigma\cosh(\sigma)\frac{b_{n}}{2\pi}\propto nb_{n}\,. (22) Hence, the ratio of areas in the limit n\to 1 is \displaystyle\frac{a_{n}-a_{1}}{a_{1}}=\frac{nb_{n}-b_{1}}{b_{1}} \displaystyle=\, \displaystyle\frac{n(2\pi i-4\pi i\cosh(\rho)(n-1)+O((n-1)^{2})-2\pi i}{2\pi i} \displaystyle=\, \displaystyle(1-2\cosh\rho)(n-1)+O((n-1)^{2})\,. (23) Consequently, the U(1) partition function near n=1 is \displaystyle Z^{n\to 1}_{U(1)}=\, \displaystyle\theta_{3}\left(e^{-\frac{\beta}{K}},e^{-2\pi{\mathcal{E}}}\right) \displaystyle+(n-1)\Bigg[\partial_{q}\theta_{3}(q,\eta)\,e^{-\frac{\beta}{K}}\left(-\frac{\beta}{K}\right)\,(1-2\cosh(\rho)) \displaystyle\qquad\qquad+\partial_{\eta}\theta_{3}(q,\eta)\,e^{-2\pi{\mathcal{E}}}\,(-2\pi{\mathcal{E}})\Bigg] \displaystyle+O((n-1)^{2})\,, (24) whose logarithm gives the free energy \displaystyle I^{n\to 1}_{U(1)} \displaystyle=-\log Z^{n\to 1}_{U(1)} \displaystyle=-\log\theta_{3}\left(e^{-\frac{\beta}{K}},e^{-2\pi{\mathcal{E}}}\right)+(n-1)I^{(1)}_{U(1)}+O((n-1)^{2})\,, (25) where the order (n-1) part of the free energy reads \displaystyle I^{(1)}_{U(1)}=\, \displaystyle-\theta_{3}\left(e^{-\frac{\beta}{K}},e^{-2\pi{\mathcal{E}}}\right)^{-1} \displaystyle\cdot\Bigg[\partial_{q}\theta_{3}(q,\eta)e^{-\frac{\beta}{K}}(-\frac{\beta}{K})(1-2\cosh(\rho)) \displaystyle\qquad+\partial_{\eta}\theta_{3}(q,\eta)e^{-2\pi{\mathcal{E}}}(-2\pi{\mathcal{E}})\Bigg]\,. (26) In the low-temperature limit, the total free energy (besides the topological term -\chi S_{0} ) is \displaystyle I^{n\to 1}_{\text{tot}}=I_{\text{Sch}}+I_{U(1)} \displaystyle=\, \displaystyle-\frac{2\pi^{2}C}{\beta}-\frac{2\pi^{2}C}{\beta}(1+4\cosh\rho)\,(n-1)-\log\theta_{3} \displaystyle-(n-1)\left[\theta_{3}^{-1}\partial_{q}\theta_{3}(q,\eta)e^{-\frac{\beta}{K}}\frac{-\beta}{K}(1-2\cosh\rho)-\theta_{3}^{-1}\partial_{\eta}\theta_{3}e^{-2\pi{\mathcal{E}}}2\pi{\mathcal{E}}\right] \displaystyle+O(n-1)^{2}\,, (27) which reaches extremality at \rho=0 . Entropy of Effective Theory.— The total ( n\to 1 ) connected geometry boundary action consists of the Schwarzian part, the U(1) part, and the topological part. \displaystyle\log Z_{n}^{\text{conn}}\sim\chi S_{0}-I^{n\to 1}_{\text{Sch}}-I^{n\to 1}_{U(1)} \displaystyle=\, \displaystyle(2-n)S_{0}+\frac{2\pi^{2}C}{\beta}+\log\theta_{3} \displaystyle+(n-1)\left[\frac{10\pi^{2}C}{\beta}+\theta_{3}^{-1}\partial_{q}\theta_{3}e^{-\frac{\beta}{K}}\frac{\beta}{K}+\theta_{3}^{-1}\partial_{\eta}\theta_{3}e^{-2\pi{\mathcal{E}}}(-2\pi{\mathcal{E}})\right] \displaystyle+O((n-1^{2}))\,. (28) We can apply the replica trick to calculate the von Neumann entropy S_{\text{vN}}=\lim_{n\to 1}\frac{\log Z_{n}-n\log Z_{1}}{n-1}=\partial_{n=1}\log Z_{n}-\log Z_{1}\,. (29) The connected entropy can be recomposed as S^{\text{conn}}=\lim_{n\to 1_{+}}\frac{\log Z^{\text{conn}}_{n}-\log Z_{1}^{n}}{n-1}=\lim_{n\to 1_{+}}\frac{\log Z^{\text{conn}}_{n}-\log Z^{\text{disconn}}_{n}}{n-1} . When S^{\text{conn}}>0 , -\log Z^{\text{conn}}_{n}<-\log Z^{\text{disconn}}_{n} for n\approx 1_{+} , thus the disconnected phase has larger free energy and dominates, and vice versa. This agrees with the following discussions comparing entropy. For \log Z_{n}^{\text{conn}} , we get the connected-geometry-sourced entropy \displaystyle S^{\text{conn}} \displaystyle=-S_{0}+\left[\frac{10\pi^{2}C}{\beta}+\theta_{3}^{-1}\partial_{q}\theta_{3}e^{-\frac{\beta}{K}}\frac{\beta}{K}-\theta_{3}^{-1}\partial_{\eta}\theta_{3}e^{-2\pi{\mathcal{E}}}2\pi{\mathcal{E}}\right] \displaystyle\quad-\left(S_{0}+\frac{2\pi^{2}C}{\beta}+\log\theta_{3}\right) \displaystyle=-2S_{0}+\frac{8\pi^{2}C}{\beta}-\log\theta_{3}+\theta_{3}^{-1}\partial_{q}\theta_{3}e^{-\frac{\beta}{K}}\frac{\beta}{K} \displaystyle\quad-\theta_{3}^{-1}\partial_{\eta}\theta_{3}e^{-2\pi{\mathcal{E}}}2\pi{\mathcal{E}}\,. (30) For the disconnected geometry, S^{\text{disconn}}=0\,, (31) which always vanishes since \log Z^{\text{disconn}}_{n}=\log Z_{1}^{n}\propto n . We can distinguish different cases. If -2S_{0}+\frac{8\pi^{2}C}{\beta}-\log\theta_{3}+\theta_{3}^{-1}\partial_{q}\theta_{3}e^{-\frac{\beta}{K}}\frac{\beta}{K}-\theta_{3}^{-1}\partial_{\eta}\theta_{3}e^{-2\pi{\mathcal{E}}}2\pi{\mathcal{E}}>0\,, (32) then S^{\text{conn}}>0 , and the von Neumann entropy of the system is governed by the disconnected one, i.e., S=S^{\text{disconn}}=0 . If -2S_{0}+\frac{8\pi^{2}C}{\beta}-\log\theta_{3}+\theta_{3}^{-1}\partial_{q}\theta_{3}e^{-\frac{\beta}{K}}\frac{\beta}{K}-\theta_{3}^{-1}\partial_{\eta}\theta_{3}e^{-2\pi{\mathcal{E}}}2\pi{\mathcal{E}}<0\,, (33) then S^{\text{conn}}<0 , and the von Neumann entropy of the system is governed by the connected one, i.e., S=S^{\text{conn}}<0 . In the low-temperature limit, \beta/K\gg 1 , we can expand the \theta_{3} function in Z_{U(1)} and obtain \displaystyle\log Z_{U(1)} \displaystyle=\log\theta_{3}(e^{-\beta/K},e^{-2\pi{\mathcal{E}}}2\pi{\mathcal{E}}) \displaystyle=2e^{-\beta/K}\cosh(2{\mathcal{E}})+O(e^{-\beta/K})^{2}\,. (34) Consequently, the connected entropy reads \displaystyle S^{\text{conn}} \displaystyle=\, \displaystyle-2S_{0}+\frac{8\pi^{2}C}{\beta}-\log\theta_{3}+\partial_{q}\log\theta_{3}e^{-\frac{\beta}{K}}\frac{\beta}{K}-\partial_{\eta}\log\theta_{3}e^{-2\pi{\mathcal{E}}}2\pi{\mathcal{E}} \displaystyle\approx\, \displaystyle-2S_{0}+\frac{8\pi^{2}C/K}{\beta/K}+2\cosh(2{\mathcal{E}})e^{-\frac{\beta}{K}}\frac{\beta}{K}+4e^{-\frac{\beta}{K}}{\mathcal{E}}\sinh(2{\mathcal{E}})\,, (35) where we neglect higher-order terms, and S_{0}=\frac{\sqrt{3}\pi\sqrt{CK}(\mathcal{E}^{2}+C/K)}{L\sqrt{\mathcal{E}^{2}-C/K}}\,. (36) Phase Transitions.— Now, we are ready to discuss the phase structure between the connected and the disconnected phases. i) Transition on \beta . For fixed C , K , and {\mathcal{E}}^{2}>C/K , the terms in S^{\text{conn}} other than -2S_{0} is \displaystyle S^{\text{conn}}_{*} \displaystyle\equiv S^{\text{conn}}+2S_{0} \displaystyle=\frac{8\pi^{2}C/K}{\beta/K}+2\cosh(2{\mathcal{E}})e^{-\frac{\beta}{K}}\frac{\beta}{K}+4e^{-\frac{\beta}{K}}{\mathcal{E}}\sinh(2{\mathcal{E}}) \displaystyle\quad+O((e^{-\beta/K})^{2})\,, (37) which monotonically decays when \beta/K increases. The transition point is when S^{\text{conn}} crosses 0 , or equivalently, when S^{\text{conn}}_{*} crosses 2S_{0} . When \beta grows from 0 to \infty , S^{\text{conn}}_{*} decays from \infty to 0 , which always intersects with 2S_{0} at some finite \beta . Therefore, in the relatively high temperature region (small \beta ), S^{\text{conn}}>0 , the disconnected phase S^{\text{disconn}}=0 dominates; in the relatively low temperature region, S^{\text{conn}}<0 , the connected phase dominates. ii) Transition on \mathcal{E} . When we fix C and K , and increase \mathcal{E} from \sqrt{C/K} to \sqrt{3C/K} , S_{0} first decays from \infty to S_{0}^{\text{min}}=\frac{2\sqrt{6}\pi C}{L},\,\text{ at }\mathcal{E}=\sqrt{3C/K}\,; (38) then as \mathcal{E} goes from from \sqrt{3C/K} to \infty , S_{0} increases from S_{0}^{\text{min}} to \infty , and grows linearly on \mathcal{E} when \mathcal{E}\gg\sqrt{C/K} . While the term S^{\text{conn}}_{*} monotonically grows with \mathcal{E} . That means when \mathcal{E}=\sqrt{C/K} , S^{\text{conn}}_{*}\big|_{{\mathcal{E}=\sqrt{C/K}}}<\infty smaller than 2S_{0} ; when \mathcal{E}\gg\sqrt{C/K} , S^{\text{conn}}_{*} grows exponentially in \mathcal{E} and larger than 2S_{0} , which grows linearly with \mathcal{E} . Hence, when \mathcal{E}>\sqrt{C/K} small, S^{\text{conn}}<0 dominates; when \mathcal{E} large and S^{\text{conn}}>0 , S^{\text{disconn}} dominates. We conclude that there must be at least one connected-disconnected phase transition on \mathcal{E} from \sqrt{C/K} to \infty . iii) Transition on C/K with C/K\in(0,\mathcal{E}^{2}) . When \beta/K is relatively large, S^{\text{conn}}_{*} is suppressed and always smaller than 2S_{0} . In this case, the connected phase always dominates, and there is no phase transition on C/K . When \beta/K is relatively small, there can exist one or two phase transitions on C/K . When C/K=0 , depending on \beta/K and \mathcal{E} , we start with either the connected phase (relatively medium \beta/K ) or the disconnected phase (relatively small \beta/K ). As we increase C/K , the entropy approaches the disconnected phase. Hence, there is a phase transition on C/K from the connected phase to the disconnected phase, with C/K=0 belonging to the connected phase. When further increasing C/K to the upper bound \mathcal{E}^{2} , we will go back to the connected phase, since S_{0} diverges and dominates when C/K\to\mathcal{E}^{2} (see Fig. 3 ). Figure 3 : The phase structure on \beta/K and C/K with constant \mathcal{E} . Quantum corrections.— Adding quantum correction to the Schwarzian partition function would introduce additional logarithmic terms in S^{\text{conn}} Saad:2019lba ; Mertens:2022irh ; Turiaci:2024cad : S^{\text{conn}}_{\text{quantum}}=S^{\text{conn}}+\frac{3}{2}\log\frac{2\pi C}{\beta}+\frac{1}{2}\log\frac{2\pi C}{\beta}\,, (39) where \frac{3}{2}\log\frac{2\pi C}{\beta} comes from the disk quantum correction, and \frac{1}{2}\log\frac{2\pi C}{\beta} comes from the trumpet correction. In a special limit K\ll C , the ratio \beta/K cannot increase all the way to \infty . Instead, it should be truncated at \beta_{q}/K=\frac{2\pi C}{K} . In this case, for the discussion in case i) on the transition of \beta/K , the sum S^{\text{conn}}+2S_{0} can only decrease to the minimal value at \beta=\beta_{q} about (S^{\text{conn}}+2S_{0})_{\textrm{min}}\approx\left(\frac{8\pi^{2}C}{\beta}+2\log\frac{\beta_{q}}{\beta}\right)_{\beta=\beta_{q}}=4\pi\,. (40) If additionally 4\pi>2S_{0} , the predicted transition point \beta_{c} is larger than \beta_{q} (or equivalently, T_{c}T_{q} ). In contrast, when 4\pi<2S_{0} , the quantum-corrected term 2\log\frac{\beta_{q}}{\beta} only shifts the transition point discussed in Case i), which considers phase transitions without quantum corrections. Discussion.— We considered the boundary effective theory of near-extremal AdS RN black holes in the near-horizon region. The theory consists of a Schwarzian mode and a U(1) phase mode. We computed the partition function on replica geometries and derived the entropy from the boundary perspective. Our analysis uncovered a rich phase structure, with transitions between connected and disconnected replica saddles controlled by the temperature \beta and the couplings ( C , K , \mathcal{E} ) of the effective theory. For future directions, the topological term in the effective theory ( Replica Phase Transition with Quantum Gravity Corrections ) exhibits a transition at \mathcal{E}^{2}=C/K , where the topological entropy S_{0} becomes imaginary. This suggests a possible unified boundary description for both dS 2 and AdS 2 JT gravity theories. Exploring this connection could shed light on the relation between quantum gravity in de Sitter and anti-de Sitter spacetimes from a boundary perspective.