Black-Hole Thermodynamics and Semiclassical Gravity

Semiclassical gravity asks how a quantum state of matter changes a spacetime that is still described by a classical metric. Its central equation replaces a classical source by a renormalized expectation value, but that replacement is only the beginning: the state must be specified, the stress tensor must be conserved, gravitational counterterms must be fixed, the geometry and state must be solved together, and fluctuations and response must remain controlled. This volume develops that chain from the semiclassical Einstein equation to black-hole thermodynamics, generalized entropy, backreaction, evaporation, cosmological horizons, and chronology. It also marks the point at which a fixed-background calculation, a mean-field metric, or any endpoint extrapolation ceases to answer the physical question being asked.

Helpful background. Quantum Fields in Curved Spacetime supplies states, Hadamard renormalization, and fixed-background Hawking radiation; Einstein Equations supplies Bianchi consistency and classical stress–energy coupling; and Black-Hole Geometry and Mechanics supplies the horizon definitions, generator normalization, charges, and classical first law used throughout.

A black hole gets hotter as it loses energy

Begin with an isolated, neutral, nonrotating black hole whose radius is much larger than the Planck length. In the leading Schwarzschild approximation,

$$r_+=2GM,\qquad T_{\mathrm H}=\frac{1}{8\pi GM},\qquad S_{\mathrm{BH}}=\frac{A}{4G}=4\pi GM^2.$$

Here $M$ is the energy measured at infinity, $A=4\pi r_+^2$ is the horizon area, and $S_{\mathrm{BH}}$ is the black-hole thermodynamic entropy. Doubling the mass halves the temperature and quadruples the entropy. Differentiating shows that $T_{\mathrm H}\,dS_{\mathrm{BH}}=dM$, as required by the first law, but also that

$$C=\frac{dM}{dT_{\mathrm H}}=-8\pi GM^2\lt0.$$

If the hole radiates more energy than it absorbs, it loses mass and becomes hotter. An ordinary body with positive heat capacity cools as it loses energy; this reversed response is why the gravitational boundary conditions and the choice of thermal reservoir matter. A cavity or anti-de Sitter boundary can change the relevant heat capacity. The temperature and entropy relations are reviewed in Wald 2001, §§3–4.

This small calculation raises three different questions. What quantum state produces the outgoing radiation? How does its energy loss change the metric? Does the entropy of the radiation compensate for the decreasing area entropy? The first is developed in Quantum Fields in Curved Spacetime. This volume develops the second and the generalized entropy needed for the third. A decreasing area alone is therefore neither a violation of the generalized second law nor an answer to the fine-grained information problem.

For a first pass, follow surface gravity and temperature, area entropy and the first law, and luminosity and mass loss. Then return to the equations below to see what is required to make that evolving picture self-consistent. The formulas above remain a slowly evolving, large-hole approximation; they do not determine the final endpoint of evaporation.

The semiclassical closure problem

For a collection of quantum fields $\widehat\Phi$ in a state $\omega$, a schematic four-dimensional field equation is

$$G_{ab}+\Lambda g_{ab} +16\pi G\left(\alpha H^{(1)}_{ab}+\beta H^{(2)}_{ab}\right) =8\pi G\,\langle \widehat T_{ab}\rangle_{\omega,\mathrm{ren}}.$$

The local tensors $H^{(i)}_{ab}$ arise by varying curvature-squared terms in the renormalized gravitational action. Their coefficients are measured or matched couplings. Moving a finite local curvature term into the definition of the renormalized stress tensor requires a compensating change in those couplings, so that the complete equation stays the same. The expectation value depends on the state and on the metric whose equation it sources. A solution is therefore not a background plus a convenient stress tensor; it is a compatible data set

$$\mathcal S=\bigl(g_{ab},\omega,\{g_i(\mu)\},\mathcal B, \mathcal G_{\mathrm{ret}},\mathcal O\bigr),$$

where the renormalized couplings ${g_i(\mu)}$, initial and boundary data $\mathcal B$, causal prescription $\mathcal G_{\mathrm{ret}}$, and observables $\mathcal O$ are stated with the geometry and state. Diffeomorphism invariance then gives the indispensable compatibility condition $\nabla^a\langle\widehat T_{ab}\rangle_{\mathrm{ren}}=0$. The effective-action and stochastic formulations of this closure are developed in Hu and Verdaguer 2008, §§2–5, while Parker and Simon 1993, §§2–6 explain when higher-derivative corrections can be reduced consistently to the retained perturbative order.

Pages use the site metric and curvature conventions, $c=\hbar=k_{\mathrm B}=1$ with $G$ explicit, and $e^{-i\omega t}$ for Fourier modes. They state the quantum state, regulator or subtraction, finite-renormalization choice, boundary conditions, horizon generator, ensemble, and approximation order whenever one of them changes the conclusion.

Control requires more than small curvature

Small curvature in Planck units is necessary in familiar applications, but it is not sufficient. A useful first parameter is

$$\epsilon_R\sim \ell_{\mathrm P}^{2}\mathcal R,$$

with $\mathcal R$ a representative curvature scale. One must also control loop or large-$N$ counting, derivatives relative to the effective cutoff, secular growth, and fluctuations. For a smeared stress observable $T[f]=\int d^4x\sqrt{-g}\,f^{ab}T_{ab}$, a state can have a modest mean and a large connected variance,

$$\epsilon_F[f] =\frac{\sqrt{\langle T[f]^2\rangle_c}} {\lvert\langle T[f]\rangle\rvert+T_{\mathrm{ref}}}.$$

The reference scale prevents a meaningless division by a vanishing mean; the sampling tensor and scale are part of the test. Linear response adds another question: does the retarded stress-tensor kernel generate a growing, gauge-invariant mode below the cutoff? Stochastic gravity keeps the associated noise kernel when mean-field fluctuations matter. These criteria answer different questions and need not fail together.

The same discipline governs thermodynamics. A Hawking temperature requires a normalized horizon generator; an energy or chemical potential must be defined with the same normalization. A Euclidean action requires boundary data, counterterms or a reference subtraction, and an ensemble, as in Gibbons and Hawking 1977, pp. 2752–2755. Generalized entropy combines renormalized geometric and exterior terms; neither bare term is a separately invariant observable. Evaporation requires flux balance and a slowly evolving geometry before an adiabatic temperature can be used. The standard thermodynamic bridge and its limitations are summarized in Wald 2001, §§2–6.

The volume in twelve chapters

The order moves from the controlled semiclassical expansion to thermodynamic structures and then to dynamical and frontier applications. Each row names the question that must be answered before the next group of claims is safe.

Chapter	Central task	Boundary to keep visible
Semiclassical Regime, Expansion, and Validity	Identify loop, large-$N$, curvature, derivative, fluctuation, and secular control parameters.	A small mean stress tensor does not by itself control its fluctuations or response.
The Semiclassical Einstein Equation	Derive the renormalized equation, its conservation law, initial data, order reduction, and self-consistent solutions.	Finite gravitational couplings and state dependence are distinct inputs.
Linear Response, Fluctuations, and Stochastic Gravity	Build retarded response, noise, Einstein–Langevin dynamics, and stability tests.	A cutoff-scale pole or gauge mode is not a physical low-energy instability.
Quantum Energy Inequalities and Semiclassical Focusing	Replace failed pointwise energy conditions by precisely sampled or information-theoretic bounds.	Worldline QEIs, ANEC, QNEC, and quantum focusing have different hypotheses.
Black-Hole Temperature and Thermodynamic Potentials	Relate surface gravity, temperature, charges, chemical potentials, redshift, and ensemble choice.	Multiple horizons generally do not share a single equilibrium temperature.
Bekenstein–Hawking Entropy and the Generalized Second Law	Develop area entropy, generalized entropy, renormalization, and proof strategies for the generalized second law.	A causal-horizon theorem is not automatically a trapping-horizon theorem.
Euclidean Saddles and Gravitational Thermodynamics	Compute actions, conical entropy, phase competition, and negative modes.	Local saddle stability, heat capacity, and global free-energy preference are separate tests.
Hawking Flux, Backreaction, and Evaporation Dynamics	Convert spectra and greybody factors into conserved mass, spin, charge, and horizon evolution.	Fixed-background flux is an input to backreaction, not already a self-consistent evaporating spacetime.
Generalized Entropy and Quantum-Corrected Horizons	Define quantum expansion, extremality, barriers, and generalized horizon laws.	Semiclassical geometry does not by itself prove microscopic information recovery.
Endpoint Regimes and Remnant Questions	Classify breakdown, near-extremal, remnant, bounce, and singular endpoint possibilities.	A model scenario is not a universal four-dimensional endpoint.
Semiclassical Cosmology and Cosmological Horizons	Apply renormalized stress tensors, anomaly actions, stability, and thermodynamics to cosmological symmetries.	An anomaly fixes a trace, not a unique state or global solution.
Semiclassical Causality, Chronology, and Quantum Horizons	Test causal response, negative-energy geometries, Cauchy horizons, chronology, and horizon fluctuations.	A theorem with compact-generation and Hadamard hypotheses is not universal chronology protection.

Four reading routes

Controlled backreaction. Read chapters 1–3 in order, then the evaporation and causality chapters. This route follows the same state and renormalized couplings from the mean equation through response, noise, and evolving horizons.

Black-hole thermodynamics. Begin with temperature and potentials, continue through generalized entropy and Euclidean saddles, and then study quantum-corrected horizons. Record the generator normalization and ensemble at every transition; otherwise identical symbols can denote inequivalent thermodynamic systems.

Negative energy and focusing. Combine the energy-inequality chapter with the generalized-second-law, quantum-horizon, and chronology chapters. Before using an inequality, list its field theory, state class, sampling or null deformation, dimension, curvature assumptions, and whether gravity is dynamical.

Evaporation and endpoints. Read the validity chapters before the flux and endpoint chapters. Use the computational examples to learn convergence, balance, and branch diagnostics, but return to the physical model and its assumptions before drawing a black-hole conclusion.

Connections to quantum fields, information, and quantum gravity

Quantum Fields in Curved Spacetime develops quantization and the Hawking derivation on a prescribed geometry. This volume starts when the renormalized expectation value, response, or fluctuations are allowed to influence the metric. Black-Hole Information develops fine-grained Page curves, islands as information-recovery constructions, and reconstruction; here quantum extremal surfaces are treated only as semiclassical geometric and entropy objects. Quantum Gravity studies microscopic degrees of freedom, singularity resolution, and final endpoint proposals. A breakdown diagnosis is a result of semiclassical analysis, but it is not a microscopic replacement theory.

These boundaries matter most on frontier pages. Evidence is reported with its state, dimension, model, and approximation, together with credible contrary results and a dated reason to revisit the assessment. In particular, a two-dimensional solvable model can reveal exact conservation and state dependence without deciding a four-dimensional endpoint; a recent analytic ansatz can sharpen a question without becoming an independently established solution; and a conjectured focusing principle must not be called a theorem when restricted versions or counterexamples are under active study.

Check your preparation

Suppose a calculation supplies a conserved tensor $\langle\widehat T_{ab}\rangle_{\omega,\mathrm{ren}}$ on a Schwarzschild background and finds an outgoing flux at infinity. Explain why these facts do not yet define a self-consistent evaporating black hole. Give at least five additional pieces of data or checks and identify the neighboring subject that derives the original fixed-background radiation calculation.

Solution

Conservation on the prescribed background is necessary, but the source must be paired with a metric that solves the renormalized semiclassical equation. One must specify the quantum state—normally including its behavior on the past horizon and at infinity—the subtraction prescription and finite gravitational couplings, and the boundary or initial data that select a geometric branch. The negative flux entering the horizon and the positive flux at infinity must satisfy the same energy balance. The evolving mass and any spin or charge must change slowly compared with the local oscillation and light-crossing scales if an adiabatic sequence of stationary geometries is used.

One must also check that curvature and frequencies remain below the effective cutoff, that stress-tensor fluctuations and causal linear response do not invalidate the mean geometry, that apparent-, trapping-, and event-horizon language is not interchanged, and that accumulated secular error remains small over the claimed interval. Greybody data and particle content must be held fixed when integrating the loss law. Near extremality or the endpoint, the hierarchy may fail even while local curvature looks modest. The fixed-background state construction and Hawking spectrum belong to Quantum Fields in Curved Spacetime; this volume develops the subsequent self-consistency and backreaction problem.

References

• Gibbons, G. W., and S. W. Hawking (1977), “Action Integrals and Partition Functions in Quantum Gravity,” Physical Review D 15, 2752–2756. DOI.
• Hu, B. L., and E. Verdaguer (2008), “Stochastic Gravity: Theory and Applications,” Living Reviews in Relativity 11, 3. DOI; Open PDF.
• Parker, L., and J. Z. Simon (1993), “Einstein Equation with Quantum Corrections Reduced to Second Order,” Physical Review D 47, 1339–1355. DOI; Open PDF.
• Wald, R. M. (2001), “The Thermodynamics of Black Holes,” Living Reviews in Relativity 4, 6. DOI; Open PDF.