Near-Extremal and Cold Endpoints

No general semiclassical calculation shows that an evaporating black hole reaches an exactly extremal state in finite physical time. In classical continuous models, vanishing temperature and greybody suppression often make extremality asymptotic; charged-particle emission can instead drive the hole away from it. Before either conclusion becomes exact, the near-extremal level spacing, throat dynamics, and stochastic emission invalidate the naive thermal continuum.

Helpful background. Quantum Fields in Curved Spacetime supplies Hawking emission; Einstein Equations supplies backreaction; Black-Hole Geometry and Mechanics supplies extremal horizons; Semiclassical Gravity supplies loss laws and validity criteria; Black-Hole Information treats fine-grained late radiation; and Quantum Gravity studies the microscopic throat spectrum and endpoint.

Near-extremal scales

For this section only, use geometrized units $G=1$. Thus the symbols $M$ and $Q$ in the next three equations denote mass and charge length parameters, with the Reissner–Nordström charge normalized so

$$f(r)=1-\frac{2M}{r}+\frac{Q^2}{r^2}, \qquad r_\pm=M\pm\sqrt{M^2-Q^2}.$$

The Hawking temperature is

$$T_{\rm H}=\frac{r_+-r_-}{4\pi r_+^2}.$$

Writing the geometrized excess as $\Delta_{\rm geo}=M-|Q|\ll |Q|$, expand $\sqrt{M^2-Q^2}=\sqrt{2|Q|\Delta_{\rm geo}}[1+O(\Delta_{\rm geo}/|Q|)]$. Since $r_+\simeq|Q|$, the leading temperature is

$$T_{\rm H}\simeq\frac{1}{2\pi} \sqrt{\frac{2\Delta_{\rm geo}}{|Q|^3}}.$$

Thus the hole can be very cold even while its radius is macroscopic. Restoring physical energy units, the relevant comparison is not simply $T_{\rm H}/M_{\rm P}$: the long throat introduces a much smaller mass-gap scale of order $M_{\rm P}^4/M_{\rm ext}^3$. Bai and Korwar describe the modified low-temperature energy relation and greybody scaling in Bai and Korwar 2023, §§ 2–5, PDF.

Here $M_{\rm P}=G^{-1/2}$ is the unreduced Planck mass. The crossover scale is not necessarily a literal spectral gap between discrete levels. It marks the scale at which the near-horizon gravitational mode changes the thermodynamics; the existence of an actual gap depends on the theory and charge/spin sector. For example, using $E=M_{\rm phys}-M_{\rm ext}$, the classical expansion gives

$$E\simeq2\pi^2G^2M_{\rm ext}^3T_{\rm H}^2.$$

Hence the ratio of a typical quantum energy $\omega\sim T_{\rm H}$ to $E$ grows like $(G^2M_{\rm ext}^3T_{\rm H})^{-1}$. Individual emissions cease to be small perturbations at this low scale, well before large four-dimensional curvature is required.

In the same geometrized units, Kerr’s distance from extremality is controlled by $M^4-J^2$; if $M_{\rm phys}$ denotes physical mass, the corresponding combination is $M_{\rm phys}^4-J^2/G^2$. Superradiance generally removes spin efficiently, while an electrically charged hole competes between neutral Hawking cooling and Schwinger or Hawking discharge. “Cold” therefore names a regime, not a unique endpoint.

The finite-time criterion

Let $E(t)$ denote energy above an extremal bound while holding the appropriate conserved charge fixed. A local model

$$\frac{dE}{dt}=-C E^p, \qquad C\gt 0,$$

reaches $E=0$ in finite time only for $p\lt 1$. For $p=1$ the approach is exponential, and for $p\gt 1$ it is a power-law tail requiring infinite time. Greybody factors and the shrinking thermal phase space typically make $p$ positive and can push it above unity. But this one-dimensional test is meaningful only if charge really remains fixed and the continuum emission law remains valid.

Hiscock and Weems evolve mass and charge with separate thermal and charged-particle channels in Hiscock and Weems 1990, §§ II–IV. Their trajectories show why neutral cooling toward extremality and electric discharge away from extremality must be solved together. A third-law-like unattainability statement consequently needs hypotheses about regular matter, bounded stress, continuous evolution, and the absence of a channel that changes the extremal bound.

The same integral criterion applies to angular momentum only after the spin-down channel is retained. For a mode of frequency $\omega$, azimuthal number $m$, and charge $q$, the horizon-weighted energy is

$$\widetilde\omega=\omega-m\Omega_{\rm H}-q\Phi_{\rm H}.$$

Superradiant modes can have $\widetilde\omega\lt0$ while removing positive energy and angular momentum at infinity. Dropping that channel can falsely predict approach to extremal Kerr. Near the bound, evolve mass, angular momentum, and charge together.

To restore $G$ in one explicit convention, let $M$ now denote physical mass and normalize $Q$ so the Reissner–Nordström metric contains $+GQ^2/r^2$. Then a useful subextremality diagnostic is

$$\chi(t)=M(t)^4-\frac{J(t)^2}{G^2}-\frac{M(t)^2Q(t)^2}{G},$$

with the declared charge normalization. A positive but shrinking $\chi$ records approach; a numerical step to $\chi\lt 0$ is more likely a breakdown of the emission approximation or time integrator than evidence for a superextremal black hole. Flux conservation and the extremality constraint are therefore independent checks on any claimed finite-time result.

Results and open questions through 2026-08-19

Evidence through 2026-08-19 supports a controlled near-extremal regime and then a change of description, not a universal finite-time extremal endpoint.

Evidence or process	Result	Scope and limitation
Hiscock and Weems 1990, § III, p. 1146	Their finite-mass charged-evaporation trajectories turn away before reaching $T=0$	Model-specific emission channels and particle content; not a general third-law theorem
Hiscock and Weems 1990, §§ II–IV	Coupled evaporation and discharge produce competing trajectories	Historical emission approximations; not the deep quantum throat regime
Bai and Korwar 2023, §§ 2–5, PDF	Near-extremal thermodynamics and greybody factors change below a mass gap	Model and effective-theory assumptions replace the naive thermal continuum
Brown et al. 2026, §§ 2–7, PDF	Quantum throat modes, microcanonical restrictions, and charge channels alter the final approach while maintaining subextremality	Crosses into quantum-gravitational thermodynamics and does not supply a model-independent endpoint theorem

The conservative position says only that ordinary semiclassical evolution loses control before exact extremality, so finite-time attainment is unproven. The competing permissive view uses a specified near-horizon quantum effective theory to continue the evolution and infer subextremal discrete emission. These positions depend on charge spectrum, field content, boundary conditions, ensemble, and whether the throat modes are within the claimed effective theory.

Checks, counterexamples, and limits

Charge-loss counterexample. A model that freezes $Q$ can asymptote to $M=|Q|$. Adding a light charged field can discharge the hole, move the extremal bound, and reverse the trajectory. The fixed-charge result is therefore not universal.

Discrete-emission check. When the typical quantum energy is comparable to $E$, a canonical spectrum can propose an emission that would make $E\lt 0$. A microcanonical or discrete treatment must enforce the remaining energy and charge exactly. Continuing the smooth ODE through that point is not an independent prediction.

Cold-environment check. An ambient bath hotter than $T_{\rm H}$ causes absorption rather than evaporation. The physical time to extremality then depends on cosmological environment, not only on an isolated-hole luminosity.

What remains unresolved is a universal statement across charge spectra, spins, quantum states, and ultraviolet completions. The conclusions could change after an independent calculation reproduces the deep near-extremal spectrum, a theorem establishes or refutes unattainability for renormalized quantum stress, or new greybody and discharge data change the exponent in the finite-time test; a correction to the 2026 quantum-throat calculation could also change the conclusion.

Exercise

Solve $\dot E=-CE^p$ from $E_0$ to zero and determine the attainment time for $p=1/2$, $1$, and $2$. Why does this not settle a charged black hole’s fate?

Solution

For $p\ne1$, $t=(E_0^{1-p}-E^{1-p})/[C(1-p)]$. Thus $p=1/2$ reaches zero at $2\sqrt{E_0}/C$, while $p=2$ takes infinite time. For $p=1$, $E=E_0e^{-Ct}$ and attainment is also asymptotic. A black hole can lose charge, change its greybody spectrum, enter a discrete regime, or leave semiclassical control, so one fixed exponent cannot decide the endpoint.

References

• Bai, Y., and Korwar, M. (2023). “Near-Extremal Charged Black Holes: Greybody Factors and Evolution.” JHEP 03, 151. Open PDF.
• Brown, A. R., Iliesiu, L. V., Penington, G., and Usatyuk, M. (2026). “The Evaporation of Charged Black Holes.” JHEP 01, 109. DOI. Open PDF.
• Hiscock, W. A., and Weems, L. D. (1990). “Evolution of Charged Evaporating Black Holes.” Physical Review D 41, 1142–1151. DOI.