Endpoint Regimes and Remnant Questions

The late stages of evaporation are not one prediction but a set of regimes separated by explicit control conditions. This chapter asks what semiclassical gravity can still calculate as curvature, temperature, fluctuations, or state dependence approach their limits. It compares cold, remnant, explosive, bounce, and regular-core scenarios without promoting any of them to a microscopic endpoint.

Helpful background. Quantum Fields in Curved Spacetime supplies renormalized states and fluxes; Einstein Equations supplies geometric constraints; Black-Hole Geometry and Mechanics supplies horizon invariants; Semiclassical Gravity supplies backreaction and approximation control; and Black-Hole Information supplies the distinct fine-grained correlation question.

The question before the endpoint

An adiabatic calculation commonly evolves a mass scale $M(u)$ using an outgoing luminosity,

$$\frac{dM}{du}=-L[M,\rho],$$

where $\rho$ denotes the quantum state. Extrapolating this equation to $M=0$ is justified only while the hierarchy used to derive $L$ survives. Useful diagnostics include a curvature ratio $\ell_*^2|R_{abcd}|$, an adiabaticity ratio $|\dot\kappa|/\kappa^2$, stress fluctuations relative to the mean, and the size of omitted higher-curvature or loop terms. The first controlled statement at late time is therefore the location of a breakdown surface, not the nature of the physics beyond it.

This discipline distinguishes three claims. A trajectory claim follows a solution within its controlled domain. A matching claim supplies data on a last reliable hypersurface. An endpoint claim specifies what replaces the failed effective description. Only the first two are generally available from semiclassical gravity. Microscopic singularity resolution, Hilbert-space dimension, topology change, and final unitarity belong to Quantum Gravity; fine-grained recovery and Page-curve interpretation belong to Black-Hole Information.

Six comparisons, one control hierarchy

The pages proceed from the most model-independent diagnostic to increasingly conditional scenarios:

1. Approach to Planckian Curvature builds invariant curvature and derivative-expansion tests and explains why small mass alone is not a covariant breakdown criterion.
2. Near-Extremal and Cold Endpoints separates vanishing surface gravity from dynamical accessibility, stability, and quantum-state regularity.
3. Remnant Thermodynamic Constraints tests long-lived small objects against entropy capacity, production, species, lifetime, and coupling constraints.
4. Explosive, Thunderbolt, and Bounce Scenarios compares three causal continuations and identifies which assumptions, rather than semiclassical evolution, select them.
5. Regular Black-Hole Evaporation Models checks whether bounded curvature, stress conservation, horizon evolution, and a globally specified state coexist in a proposed geometry.
6. Agnostic Endpoint Data for Quantum Gravity assembles invariant geometry, charges, fluxes, state data, correlations, causal boundaries, and error bars on the last controlled surface.

No single scalar orders every scenario. A near-extremal solution may have low Hawking temperature yet large stress in an ill-chosen state. A regular metric may have bounded polynomial invariants yet require an unphysical effective source. A remnant may satisfy an energy balance while failing a state-counting or production test. Each page therefore states its state, renormalization scheme, causal assumptions, and first failed control parameter.

What the models actually establish

Two-dimensional solvable models demonstrate that backreaction can be followed to a sharply defined singular or matching boundary. In the counterterm-modified Russo–Susskind–Thorlacius model, the large-$N$ equations are exactly tractable, but the counterterm, endpoint boundary prescription, dimension, and matter content remain part of the result rather than dispensable details Russo, Susskind, and Thorlacius 1992, §§ 1–4, PDF. Four-dimensional effective geometries provide useful consistency laboratories, yet regularity imposed on a metric does not by itself derive the quantum stress tensor that supports it; Frolov’s explicit nonsingular models make the required metric assumptions visible Frolov 2016, §§ II–V, PDF.

The comparison can be summarized by the question each scenario must answer:

Scenario	Semiclassical question	Information still missing
Cold or extremal	Does $\kappa\to0$ occur dynamically in a regular state?	Stability and microscopic degeneracy
Long-lived remnant	Can energy, entropy, production, and lifetime bounds be met together?	Fundamental state spectrum and couplings
Burst or thunderbolt	Is a large final flux compatible with integrated energy and controlled stress?	UV propagation through the high-curvature region
Bounce or regular core	Do constraints and a conserved renormalized source support the proposed causal geometry?	Singularity resolution and topology-changing dynamics

These are competing hypotheses, not mutually exhaustive outcomes. The unresolved region begins where their required continuation depends on omitted ultraviolet operators, uncontrolled fluctuations, or boundary conditions not fixed by the incoming state.

Review the chapter

Suppose an evaporating solution reaches a surface $\Sigma_*$ where $\ell_*^2\sqrt{R_{abcd}R^{abcd}}=0.08$, $|\dot\kappa|/\kappa^2=0.6$, and the rms fluctuation of a smeared invariant expansion observable divided by its background scale is $0.15$. A calculation declares tolerance $0.1$ for every diagnostic. Is $\Sigma_*$ a last controlled surface, and may the mass-loss law be extrapolated beyond it?

Solution

Only the curvature ratio passes. The adiabaticity ratio exceeds tolerance by a factor of six and the metric-observable fluctuation ratio by a factor of $1.5$. Thus $\Sigma_*$ is already outside the declared controlled region: move the matching surface earlier, where every required test passes, and propagate the uncertainties in the data exported there. Previously integrated fluxes are usable only over their controlled interval. Extrapolating the adiabatic luminosity from the failed surface is unjustified, and these numbers select no remnant, burst, bounce, or regular core.

What completion enables

A reader who completes the chapter can locate the first semiclassical failure without confusing it with a singularity, compare endpoint proposals on shared thermodynamic and causal tests, and prepare a noncircular data record for a more microscopic theory. The durable result is not a preferred ending. It is a disciplined separation between controlled evolution, conditional continuation, and the genuinely unresolved endpoint.

References

• Frolov, V. P. (2016). “Notes on Nonsingular Models of Black Holes.” Physical Review D 94, 104056. DOI. Open PDF.
• Russo, J. G., Susskind, L., and Thorlacius, L. (1992). “End Point of Hawking Radiation.” Physical Review D 46, 3444–3449. DOI. Open PDF.

Dated Research assessment. For open questions and recent results, follow the link below. Each assessment states the date through which it considers results and the limits of its conclusions.

• Evaporation Endpoints and Remnants