Charged and Rotating AdS Black Holes

Stability of a charged or rotating AdS black hole is not an intrinsic yes-or-no property of the metric. It depends on which extensive charges or intensive potentials may fluctuate, which angular frame defines the boundary Hamiltonian, and whether dynamical modes can exchange charge or angular momentum with the reflecting AdS boundary. The thermodynamic ensemble selects the Hessian; the field content and boundary conditions select possible superradiant instabilities.

Required background. Gauge Theory and String-Theory Preparation supplies large-N counting, conserved currents, and string-scale control used below.

Helpful background. Quantum Gravity frames the semiclassical expansion, Quantum Information distinguishes ensembles from individual states, Black-Hole Geometry and Mechanics supplies the first law and horizon generators, and Quantum Fields in Curved Spacetime clarifies mode and state boundary conditions.

First-law variables and boundary frames

For a stationary charged rotating solution, choose the horizon generator

$$\chi=\partial_t+\sum_i\Omega_i\partial_{\phi_i}.$$

The potentials entering the first law are measured relative to the declared frame at infinity. Fix the charge sign by $D_\mu=\nabla_\mu-iqA_\mu$. The electric potential is

$$\Phi=(\chi\cdot A)_\infty-(\chi\cdot A)_H.$$

For example, $A_t=-q_0/r$ gives $\Phi=q_0/r_h$; shifting to the Euclidean-regular gauge sets $(\chi\cdot A)_H=0$ and puts the same potential at infinity. This fixes the sign independently Caldarelli, Cognola, and Klemm 2000, §II, Eq. (20). The first law is

$$dE=T\,dS+\sum_i\Omega_i\,dJ_i+\sum_a\Phi_a\,dQ_a.$$

In Kerr–AdS, coordinates that rotate at infinity shift both $E$ and $\Omega$ unless the charge definition is translated consistently. In Reissner–Nordström–AdS, changing the Maxwell normalization rescales $Q$ and $\Phi$ inversely. A formula for a critical potential is incomplete without these choices.

Canonical and grand-canonical tests

At fixed angular momenta and charges, the relevant thermodynamic potential is the Helmholtz free energy

$$F(T,J,Q)=E-TS.$$

At fixed boundary angular velocities and electric potentials, it is the Gibbs potential

$$G(T,\Omega,\Phi) =E-TS-\Omega_iJ_i-\Phi_aQ_a =\frac{I_E}{\beta}.$$

Local stability requires convexity in the variables allowed to fluctuate. In the grand-canonical ensemble, for example, one may test the susceptibility matrix

$$\mathcal H_{AB} =-\frac{\partial^2G}{\partial y^A\partial y^B}, \qquad y^A=(T,\Omega_i,\Phi_a),$$

with signs interpreted through heat capacities and charge or rotational susceptibilities. A branch can have positive heat capacity at fixed $Q$ yet be unstable at fixed $\Phi$. Complete phase diagrams for charged and rotating AdS families make this ensemble dependence explicit in Caldarelli, Cognola, and Klemm 2000, §§IV–V and Hawking, Hunter, and Taylor-Robinson 1999, §§II–IV, pp. 4–10.

Global stability is another comparison: among all saddles with the same boundary data, the one with the lowest potential dominates. A locally stable black hole may remain globally subdominant to thermal AdS or another charged branch.

Superradiance is a dynamical condition

A bosonic mode $e^{-i\omega t+im_i\phi_i}$ carrying charge $q_a$ extracts horizon energy when its real scattering frequency obeys

$$0\lt \operatorname{Re}\omega \lt \sum_i m_i\Omega_i+\sum_a q_a\Phi_a.$$

Here the frequency is measured in the nonrotating boundary frame and the gauge with vanishing boundary electric potential. In the Euclidean-regular gauge its numerical value shifts: the gauge-invariant horizon frequency is $\omega-m_i\Omega_i+q_a(\chi\cdot A^a)_H$. In asymptotically flat space, amplified radiation can escape. Reflecting AdS boundary conditions can return it and produce a growing mode when amplification overcomes losses; the actual complex-frequency boundary problem must be solved. Thermodynamic susceptibility and superradiance often correlate in examples, but neither implies the other without hypotheses about the spectrum, field content, and allowed perturbations. Near a BPS bound, additional light charged fields and supersymmetric boundary conditions can change the relevant comparison.

Checks across limiting families

Three limits expose convention mistakes. Setting $J_i=0$ must recover the charged AdS first law and its fixed-$Q$ versus fixed-$\Phi$ distinction. Setting $Q_a=0$ must recover Kerr–AdS with angular velocity measured in the nonrotating boundary frame. Taking both to zero must reduce to the spherical Hawking–Page branches rather than a planar equation of state.

The Smarr relation provides a dimensional check when the cosmological constant is treated either as fixed or as pressure. These conventions must not be mixed: an extended first law includes a $VdP$ term, while the boundary CFT comparison normally fixes the AdS radius and therefore its central-charge scale. Finally, approaching extremality should send $T$ to zero while leaving the regular chemical potentials finite; a divergent potential usually signals a bad gauge or frame.

Choosing the ensemble

The physical setup answers the page’s question. A boundary path integral with fixed gauge potential and angular shift is grand canonical. A projection onto charge and angular-momentum sectors is canonical or microcanonical. Reflecting boundary conditions make the total system closed, whereas coupling the boundary to reservoirs changes which fluctuations are allowed. The correct statement of stability therefore lists ensemble, frame, gauge, perturbing fields, and AdS boundary condition together.

Exercises

Explain why positive heat capacity at fixed charge does not prove grand-canonical stability for a charged AdS black hole.

Solution

Positive $C_Q$ tests only the energy response to temperature variations while $Q$ is held fixed. In the grand-canonical ensemble the boundary potential $\Phi$ is fixed, so charge can fluctuate and the relevant test also contains the electric susceptibility and mixed thermal-electric derivatives. A negative eigenvalue of the susceptibility matrix $\mathcal H=-\partial^2G/\partial y^A\partial y^B$ defined above signals an instability even when $C_Q$ is positive. The minus sign matters: the grand potential is concave in stable intensive variables. Legendre transforming between $F(T,Q)$ and $G(T,\Phi)$ changes which derivatives are constrained.

A simple matrix illustrates why checking individual entries is insufficient. In dimensionless variables, a symmetric susceptibility matrix with diagonal entries one and off-diagonal entries two has eigenvalues three and minus one. The unstable eigenvector combines thermal and electric fluctuations even though both diagonal entries are positive. A physical black-hole calculation must test the complete matrix with its actual normalization and ensemble.

References

• Caldarelli, Marco M., Guido Cognola, and Dietmar Klemm. “Thermodynamics of Kerr–Newman–AdS Black Holes and Conformal Field Theories.” Classical and Quantum Gravity 17 (2000): 399–420. DOI. Open manuscript.
• Hawking, Stephen W., Christopher J. Hunter, and Marika M. Taylor-Robinson. “Rotation and the AdS/CFT Correspondence.” Physical Review D 59 (1999): 064005. DOI.