What does the covariant entropy bound count? The vacuum-entropy saturation conjecture for general boundary structures in quantum gravity

The Bekenstein-Hawking entropy assigns to a horizon an entropy equal to one quarter of its area in Planck units, and Bousso's covariant entropy bound generalizes this to arbitrary spacetimes; but what these formulas physically count at the level of full quantum gravity remains unsettled. Nomura and Weinberg (Phys. Rev. D 90, 104003 (2014)) proposed that this entropy is predominantly an 'entropy of a vacuum': the logarithm of the degeneracy of distinct ways an emergent quantum field theory on a fixed semiclassical background can arise in quantum gravity, with matter-entropy contributions entering only as a perturbation, and conjectured that the vacuum-plus-matter entropy saturates the covariant bound for every boundary-structure sector of the quantum-gravity Hilbert space. This problem asks for a determination of what the covariant entropy counts: a proof, refutation, or delimitation of the saturation conjecture; a microphysical realization of the conjectured vacuum-sector structure within a concrete quantum-gravity model; a decision on whether the vacuum label is operationally accessible to an infalling observer; and a precise notion of boundary-structure equivalence underpinning the counting.

Background

A black hole in general relativity is characterized by the Bekenstein-Hawking entropy $S_{\rm BH}=\mathcal{A}/4l_{\rm P}^{2}$, where $\mathcal{A}$ is the horizon area and $l_{\rm P}$ is the Planck length. Understanding what this entropy microscopically counts is a foundational problem of quantum gravity. Bousso's covariant entropy bound (also called the Bousso bound) generalizes the area rule: the matter entropy flowing across a 'light sheet' — a hypersurface generated by light rays whose cross-sectional area is non-increasing, terminated where the rays caustic or hit a singularity — never exceeds the initial area divided by $4l_{\rm P}^{2}$. Flanagan, Marolf, and Wald proved classical versions of this bound under assumptions on the stress tensor, and Bousso, Freivogel, and Leichenauer showed that matter entropy can saturate the bound in specific semiclassical systems such as flat FRW (Friedmann-Robertson-Walker, i.e., spatially homogeneous and isotropic) cosmologies; a quantum version of the bound in semiclassical gravity follows from the quantum focusing conjecture of Bousso, Fisher, Leichenauer, and Wall. However, all of these results concern the entropy of matter and radiation propagating on a spacetime background; they do not say what fixes the number of states behind the area formula itself at the level of the full, non-semiclassical theory.

The firewall argument of Almheiri, Marolf, Polchinski, and Sully sharpened the stakes: assuming that the Bekenstein-Hawking entropy counts only the states of Hawking radiation in the exterior quantum field theory, unitary evaporation appears to destroy the smoothness of the horizon. Harlow and Hayden further showed that decoding the information in the Hawking radiation is computationally intractable for a young black hole, obstructing the measurement needed to exhibit entanglement monogamy violations. These results motivate asking whether the entropy instead counts degrees of freedom invisible to the exterior quantum field theory.

Nomura and Weinberg answered with the 'entropy of a vacuum' picture. They decompose the quantum-gravity Hilbert space by boundary structure — writing $\mathcal{H}=\bigoplus_{\partial\mathcal{M}}\mathcal{H}_{\partial\mathcal{M}}$, where $\partial\mathcal{M}$ schematically denotes the Planckian boundaries (stretched black-hole horizon, de Sitter horizon, or boundaries at spatial infinity) surrounding a freely falling observer's reference frame. In their construction, the Bekenstein-Hawking entropy is the logarithm of the number of distinct vacuum states: each vacuum $|\psi_{k}\rangle$, labeled by $k=1,\dots,e^{\approx\mathcal{A}/4l_{\rm P}^{2}}$, is a configuration of intrinsically quantum-gravitational degrees of freedom on the stretched horizon (a membrane-like surface slightly outside the horizon where Planckian physics sets in, viewed from a distant reference frame), on which an identical emergent quantum field theory (the low-energy effective theory of matter and radiation) can be built, so that all $e^{S_{\rm vac}}$ vacua look the same to semiclassical physics. A 'stretched-horizon state' $|\tilde{\imath};k\rangle$ denotes the $\tilde{\imath}$-th internal excitation of the $k$-th horizon configuration, and the vacuum states take the thermofield-like form $|\psi_{k}\rangle=Z^{-1/2}\sum_{i}e^{-\beta E_{i}/2}|i\rangle|\tilde{\imath};k\rangle$ (the sum over $i$ implying the corresponding paired sum over $\tilde{\imath}$), entangling each near-exterior mode $|i\rangle$ with the corresponding horizon excitation. Matter entropy $S_{\rm mat}$ — the entropy of particle excitations within a given emergent field theory — enters only as a tiny correction, $S_{\rm mat}\sim O(\mathcal{A}^{n}l_{\rm P}^{-2n})$ with $n<1$, so that $S=S_{\rm vac}+S_{\rm mat}\approx S_{\rm vac}$. Evaporation then transfers vacuum entropy into the entropy of final-state Hawking radiation while the horizon stays smooth, and analogous statements are proposed for cosmological (e.g., de Sitter) horizons.

The paper itself flags what remains unresolved. (i) The microscopic Hamiltonian coupling stretched-horizon and exterior degrees of freedom — the Planckian dynamics implementing the information transfer — is left unspecified; the authors are deliberately agnostic about it. (ii) The central saturation conjecture (Eq. (75) of the paper), $S=S_{\rm vac}+S_{\rm mat}\approx\frac{1}{4l_{\rm P}^{2}}\int d\mathcal{A}(\lambda_{\rm H})$ with $\lambda_{\rm H}(\theta,\phi)=\min\{\lambda_{\rm P},\lambda_{\rm app}\}$ the nearer of the Planckian boundary and the apparent horizon (a locally defined surface where outgoing light rays stop expanding) along each direction, is explicitly presented as a conjecture verified only by consistency in the specific systems considered (black-hole, Rindler, de Sitter, and FRW spacetimes; 'Rindler' denotes the horizon of a uniformly accelerated observer in flat spacetime). (iii) Whether an infalling observer could in principle measure the vacuum label $k$ — projecting onto $|\psi_{k}\rangle$ and thereby seeing a firewall — is left open, with only a speculated dynamical or computational (Harlow-Hayden-type) obstruction. (iv) The equivalence relation defining when two boundary structures count as 'the same' — a prerequisite for the Hilbert-space decomposition and the counting — is acknowledged to be non-obvious. (v) The construction is local and schematic, with no embedding in a concrete quantum-gravity model (such as a string-theoretic or AdS/CFT realization — AdS/CFT being the conjectured duality between quantum gravity in anti-de Sitter spacetime and a conformal field theory on its boundary — reproducing the $k$-sector structure).

Problem Statement

Determine what the covariant entropy bound counts at the level of full quantum gravity, by settling the 'entropy of a vacuum' picture of Nomura and Weinberg in its stated generality. Concretely, for the quantum-gravity Hilbert space decomposed by boundary structure as $\mathcal{H}=\bigoplus_{\partial\mathcal{M}}\mathcal{H}_{\partial\mathcal{M}}$, with Planckian boundaries and apparent horizons defined relative to a freely falling reference frame:

1. Prove, refute, or delimit the saturation conjecture: is the fine-grained (von Neumann) entropy of every boundary-structure sector given by $S=S_{\rm vac}+S_{\rm mat}\approx\frac{1}{4l_{\rm P}^{2}}\int d\mathcal{A}\bigl(\lambda_{\rm H}(\theta,\phi)\bigr),\qquad\lambda_{\rm H}=\min\{\lambda_{\rm P},\lambda_{\rm app}\},$

where $d\mathcal{A}(\lambda_{\rm H})$ is the area element on the surface whose radial position in each direction $(\theta,\phi)$ is $\lambda_{\rm H}$, $S_{\rm vac}$ is the logarithm of the degeneracy of emergent vacuum states and $S_{\rm mat}$ the entropy of matter/radiation excitations, with '$\approx$' holding at leading order since $S_{\rm mat}\sim O(\mathcal{A}^{n}l_{\rm P}^{-2n})$, $n<1$, is subleading — i.e., does the covariant entropy bound count vacuum degeneracy plus matter entropy, and is it always saturated, including dynamical spacetimes, de Sitter horizons, and mixed black-hole/cosmological boundary structures? To count as a delimitation, an answer must state a regime of validity — the class of boundary structures and spacetimes on which the displayed equality is asserted to hold — and establish saturation or its failure at the boundary of that regime; exhibiting saturation only in individual examples, or a counterexample that does not locate the boundary of the regime, is not a delimitation. A single counterexample to the universally quantified equality, however, already refutes the conjecture and discharges this item.

2. Provide (or rule out) a microphysical realization: exhibit, within a concrete quantum-gravity model, the exponentially degenerate $k$-labeled vacuum sectors $\bigoplus_{k}\mathcal{H}^{(k)}$ and the branch-wise Hamiltonian coupling stretched-horizon and exterior degrees of freedom that implements unitary information transfer to Hawking radiation with a smooth horizon.

3. Decide whether the vacuum label $k$ is operationally accessible: can any physical measurement project an infalling observer onto a definite vacuum $|\psi_{k}\rangle$ (thereby revealing a firewall), or is there a dynamical or computational obstruction, and can such an obstruction be proven rather than conjectured?

4. Make the counting well-defined: specify the general equivalence relation under which two boundary structures $\partial\mathcal{M}$ are 'the same'.

These four questions jointly constitute one scientific objective — validating or refuting the claim that the Bekenstein-Hawking/covariant entropy is an entropy of the quantum-gravitational vacuum. They are logically independent parts of that objective: settling whether the area formula counts $S_{\rm vac}+S_{\rm mat}$ (question 1) does not supply the microphysical realization (question 2), decide the operational accessibility of the vacuum label (question 3), or fix the boundary-structure equivalence relation (question 4).

References

1. Y. Nomura and S. J. Weinberg, 'Entropy of a vacuum: What does the covariant entropy count?', Phys. Rev. D 90, 104003 (2014). DOI: 10.1103/PhysRevD.90.104003; arXiv:1310.7564; https://arxiv.org/abs/1310.7564
2. R. Bousso, 'A covariant entropy conjecture', JHEP 07 (1999) 004. DOI: 10.1088/1126-6708/1999/07/004; arXiv:hep-th/9905177; https://arxiv.org/abs/hep-th/9905177
3. E. E. Flanagan, D. Marolf, and R. M. Wald, 'Proof of classical versions of the Bousso entropy bound and of the generalized second law', Phys. Rev. D 62, 084035 (2000). DOI: 10.1103/PhysRevD.62.084035; arXiv:hep-th/9908070; https://arxiv.org/abs/hep-th/9908070
4. R. Bousso, B. Freivogel, and S. Leichenauer, 'Saturating the holographic entropy bound', Phys. Rev. D 82, 084024 (2010). DOI: 10.1103/PhysRevD.82.084024; arXiv:1003.3012; https://arxiv.org/abs/1003.3012
5. R. Bousso, Z. Fisher, S. Leichenauer, and A. C. Wall, 'Quantum focusing conjecture', Phys. Rev. D 93, 064044 (2016). DOI: 10.1103/PhysRevD.93.064044; arXiv:1506.02669; https://arxiv.org/abs/1506.02669
6. A. Almheiri, D. Marolf, J. Polchinski, and J. Sully, 'Black Holes: Complementarity or Firewalls?', JHEP 02 (2013) 062. DOI: 10.1007/JHEP02(2013)062; arXiv:1207.3123; https://arxiv.org/abs/1207.3123
7. D. Harlow and P. Hayden, 'Quantum Computation vs. Firewalls', JHEP 06 (2013) 085. DOI: 10.1007/JHEP06(2013)085; arXiv:1301.4504; https://arxiv.org/abs/1301.4504
8. Y. Nomura, J. Varela, and S. J. Weinberg, 'Low Energy Description of Quantum Gravity and Complementarity', Phys. Lett. B 733, 126 (2014). DOI: 10.1016/j.physletb.2014.04.027; arXiv:1304.0448; https://arxiv.org/abs/1304.0448
9. Y. Nomura and S. J. Weinberg, 'Black Holes, Entropies, and Semiclassical Spacetime in Quantum Gravity', JHEP 10 (2014) 185. DOI: 10.1007/JHEP10(2014)185; arXiv:1406.1505; https://arxiv.org/abs/1406.1505
10. Y. Nomura, F. Sanches, and S. J. Weinberg, 'Relativeness in Quantum Gravity: Limitations and Frame Dependence of Semiclassical Descriptions', JHEP 04 (2015) 158. DOI: 10.1007/JHEP04(2015)158; arXiv:1412.7538; https://arxiv.org/abs/1412.7538
11. Y. Nomura and N. Salzetta, 'Why Firewalls Need Not Exist', Phys. Lett. B 761, 62 (2016). DOI: 10.1016/j.physletb.2016.08.003; arXiv:1602.07673; https://arxiv.org/abs/1602.07673
12. Y. Nomura, N. Salzetta, F. Sanches, and S. J. Weinberg, 'Spacetime Equals Entanglement', Phys. Lett. B 763, 370 (2016). DOI: 10.1016/j.physletb.2016.10.045; arXiv:1607.02508; https://arxiv.org/abs/1607.02508
13. T. Jacobson, 'Entanglement Equilibrium and the Einstein Equation', Phys. Rev. Lett. 116, 201101 (2016). DOI: 10.1103/PhysRevLett.116.201101; arXiv:1505.04753; https://arxiv.org/abs/1505.04753
14. Y. Nomura, 'Reanalyzing an Evaporating Black Hole', Phys. Rev. D 99, 086004 (2019). DOI: 10.1103/PhysRevD.99.086004; arXiv:1810.09453; https://arxiv.org/abs/1810.09453