AdS Black Holes and the Holographic Dictionary

AdS black holes translate gravitational saddles into thermal states, conserved charges, heavy sectors, and phase structure in a boundary theory. This chapter builds only the black-hole-specific part of that translation. It begins with the minimal source/response dictionary, passes through Euclidean thermodynamics and ensemble choice, and ends by separating controlled classical results from finite-N, stringy, and nonperturbative corrections. Frontier evidence is assessed through 2026-08-20.

Helpful background. Gauge Theory and String-Theory Preparation supplies large-N and string basics; Quantum Gravity supplies the semiclassical expansion; Quantum Information supplies entropy distinctions; and Black-Hole Geometry supplies horizons, surface gravity, and conserved charges.

Enter this chapter

The organizing equality is a saddle expansion, not an identity between one geometry and one exact state:

$$Z_{\rm CFT}[J] =Z_{\rm string}[\phi\to J] \simeq\sum_s e^{-I^{\rm ren}_E[g_s,\phi_s]} .$$

Boundary sources $J$ fix asymptotic bulk data; normalizable coefficients encode expectation values; and a smooth Euclidean black-hole saddle supplies a thermal contribution with $F=T I_E^{\rm ren}$. Comparing all saddles compatible with the same boundary conditions selects the leading phase. Charges and angular potentials require a declared ensemble and boundary frame. Corrections then ask which part of this chain survives beyond classical supergravity.

Read the six pages in order:

1. The Minimal AdS/CFT Dictionary for Black Holes fixes sources, responses, temperature, entropy, and Ward identities without importing all of CFT.
2. Thermal States and Euclidean Black-Hole Saddles derives periodicity, the renormalized action, and thermodynamic derivatives.
3. Hawking–Page Transition and Deconfinement works through a free-energy crossing in the canonical ensemble on a compact spherical boundary.
4. Charged and Rotating AdS Black Holes compares canonical and grand-canonical variables, boundary frames, and superradiant stability.
5. Black-Hole States and Heavy Operators tests how far energy-dimension matching identifies individual heavy states.
6. Finite-N, Stringy, and Nonperturbative Corrections separates genus, derivative, and exponential sectors rather than calling them one correction.

From boundary conditions to phases

The minimal dictionary begins with the variational problem. A bulk scalar behaving as $\phi=z^{d-\Delta}J+z^\Delta A+\cdots$ identifies $J$ as a source and the renormalized canonical momentum as the response. For the metric, the boundary conformal class sources the stress tensor. Gauge fields similarly pair chemical potentials with charge densities. The source/response rule follows from the generating-functional relation introduced in Gubser, Klebanov, and Polyakov 1998, pp. 105–108.

Euclidean regularity fixes the inverse temperature by removing the conical defect. It does not select the dominant saddle. Thermal AdS, small and large black holes, and charged or rotating competitors must be renormalized with identical boundary data. The resulting first law,

$$dE=T\,dS+\Omega\,dJ+\Phi\,dQ,$$

is simultaneously a normalization check and an ensemble warning: holding $J,Q$ fixed is not the same variational problem as holding $\Omega,\Phi$ fixed. The Hawking–Page transition then becomes a free-energy crossing, with its deconfinement interpretation controlled at large N for a CFT on a compact spatial sphere Witten 1998, §§ 2–3, pp. 511–520.

State of the evidence through 2026-08-20

The comparison table distinguishes results fixed by the duality and saddle calculation from interpretations that require extra spectral or asymptotic assumptions.

assumptions and setting	controlled evidence or claim	adverse limitation or discriminator	primary source
Large-N, strong-coupling boundary theory with a weakly curved asymptotically AdS saddle and fixed sources	Renormalized on-shell bulk variations compute boundary generating-function derivatives and Ward identities	The map does not identify an arbitrary string background with a known microscopic CFT, and loop or derivative corrections can spoil classical accuracy	Gubser, Klebanov, and Polyakov 1998, pp. 105–108
Canonical ensemble with fixed temperature and compact spherical boundary geometry	Thermal AdS and AdS black-hole actions cross, producing a first-order large-N phase transition	Planar horizons have no analogous finite-temperature Hawking–Page crossing, and finite N rounds strict saddle nonanalyticity	Hawking and Page 1983, §§ 3–4, pp. 582–587
Sparse large-central-charge two-dimensional CFT spectrum in a specified heavy-energy regime	Universal thermodynamic behavior constrains the asymptotic density of heavy states	Energy-dimension matching and Cardy growth do not prove that each heavy operator has one smooth interior or that typicality holds for all observables	Hartman, Keller, and Stoica 2014, §§ 2–4
Type-IIB AdS5 black brane with large ’t Hooft coupling and controlled leading higher-derivative term	The first alpha-prime correction changes the strong-coupling thermal free energy by a definite power	This one coefficient neither controls string loops nor determines exponentially small D-brane or topology-changing sectors	Gubser, Klebanov, and Tseytlin 1998, §§ 2–3

Competing positions and useful discriminators

One position treats the classical dictionary as the universal essence of holography: once energy and charges select a black-hole saddle, thermodynamics and simple correlators follow. This is robust inside a controlled code of states and observables. A stronger position identifies individual heavy eigenstates with the same smooth geometry. Evidence for that extension depends on spectral sparseness, observable class, typicality, and the order of large-N and late-time limits; it is not established by matching the mean density of states.

Correction language creates a second competition. A pragmatic account orders every effect by its size in one example. A structural account keeps independent parameters separate:

$$g_s\ \text{controls genus}, \qquad \alpha'/L^2\ \text{controls derivatives}, \qquad e^{-1/g_s}\ \text{labels nonperturbative sectors}.$$

The structural account is safer because a small derivative correction does not bound a loop term, and no finite power series determines an exponential sector. The decisive test is to state the parameter, observable, and time or charge regime for every truncation.

What remains unresolved is the size of the code subspace admitting one semiclassical geometry, the nonperturbative completion of many AdS gravitational path integrals, and the relation between typical heavy microstates and smooth interiors. The evidence does not establish a universal answer independent of boundary theory or observable.

What could change these conclusions

The conclusions could change when an independent finite-N calculation separates a genuine phase transition from a crossover, when a counterexample heavy-state family violates a proposed universality class, or when a theorem enlarges the controlled spectral domain. A new string-loop or nonperturbative calculation that changes the hierarchy of corrections, or a change of boundary geometry or ensemble that alters saddle dominance, could also change the interpretation. A higher-order coefficient within the same unchanged expansion does not alone revise the conceptual boundaries.

Review the chapter

Determine which conclusion survives an ensemble change. Suppose two saddles have grand potentials $\Omega_1=-4$ and $\Omega_2=-3$ at fixed $(T,\Phi)$, while their charges are $Q_1=5$ and $Q_2=1$ with the same $\Phi=1$. Which dominates grand canonically, and can the result be reused at fixed charge without another calculation?

Solution

At fixed temperature and potential, the smaller grand potential dominates, so saddle one is selected. The corresponding canonical free energies are related by $F_i=\Omega_i+\Phi Q_i$, giving $F_1=1$ and $F_2=-2$ for these numerical data. Saddle two would therefore be favored if both were admissible at the same fixed charge, but their displayed charges differ, so even that comparison is not yet a valid canonical calculation. One must construct competitors with common canonical boundary data. Dominance cannot be carried across ensembles by relabeling the potential.

References

• Gubser, Steven S., Igor R. Klebanov, and Alexander M. Polyakov. “Gauge Theory Correlators from Non-Critical String Theory.” Physics Letters B 428 (1998): 105–114. DOI. Open PDF.
• Gubser, Steven S., Igor R. Klebanov, and Arkady A. Tseytlin. “Coupling Constant Dependence in the Thermodynamics of N=4 Supersymmetric Yang–Mills Theory.” Nuclear Physics B 534 (1998): 202–222. DOI. Open PDF.
• Hartman, Thomas, Christoph A. Keller, and Bogdan Stoica. “Universal Spectrum of 2d Conformal Field Theory in the Large c Limit.” Journal of High Energy Physics 09 (2014): 118. DOI. Open PDF.
• Hawking, S. W., and Don N. Page. “Thermodynamics of Black Holes in Anti-de Sitter Space.” Communications in Mathematical Physics 87 (1983): 577–588. DOI.
• Witten, Edward. “Anti-de Sitter Space, Thermal Phase Transition, and Confinement in Gauge Theories.” Advances in Theoretical and Mathematical Physics 2 (1998): 505–532. DOI. Open PDF.