Algebraic conditions for generalized-entropy monotonicity: the open problems of the crossed-product generalized second law

The generalized second law (GSL) states that the generalized entropy — horizon area divided by $4G_N\hbar$ plus the entropy of exterior quantum fields — never decreases. Faulkner and Speranza derived the GSL for cuts of Killing horizons within the crossed-product framework, in which the gravitationally dressed algebra of a horizon exterior is a type II von Neumann algebra whose von Neumann entropy reproduces the generalized entropy, and they computed subleading $O(\epsilon^2)$ corrections to that entropy without being able to determine whether the second law survives them. Kirklin proved a modified all-orders GSL using dynamical cuts; Chandrasekaran and Flanagan proved the quantum focusing conjecture for horizon cuts on Killing-horizon backgrounds; Cao, Faulkner and Wang developed operator-valued weights in two-extremal-surface spacetimes. The construction of semifinite gravitational algebras for rotating and collapse-formed black holes remains untouched, as do the operator-valued weight between nested horizon-cut algebras and monotonicity statements for arbitrary subregions. The problem — the algebraic conditions under which generalized entropy is monotonic beyond leading semiclassical order — remains open.

Background

The generalized second law (GSL) conjectures that the generalized entropy of a black hole horizon, $S_{\rm gen}=A/(4G_N\hbar)+S_{\rm out}$, where $A$ is the horizon area on a spacelike cut and $S_{\rm out}$ the entropy of quantum fields in the exterior, never decreases as the cut is advanced along the horizon. Rigorous proofs in semiclassical gravity, notably Wall's, derive it from monotonicity of quantum relative entropy between the algebras of nested exterior regions, together with the null curvature relation of general relativity.

A modern reformulation comes from operator algebras. The algebra of observables of a quantum field theory (QFT) in a spacetime region is a von Neumann algebra, and for continuum QFT regions it is typically of type III$_1$: it admits no trace and no density matrix, so its von Neumann entropy is ill-defined. The crossed-product construction repairs this: given a type III algebra and a one-parameter group of its automorphisms, one tensors the QFT with an auxiliary quantum system (a clock/observer modeled by $L^2(\mathbb{R})$) and imposes invariance under the automorphisms; the resulting algebra is of type II, which admits a renormalized trace, density matrices, and finite relative-entropy-based entropies. Witten's work on gravity and the crossed product, the de Sitter algebra of Chandrasekaran, Longo, Penington and Witten, and the construction of Jensen, Sorce and Speranza (JSS) for general subregions established that, in the $G_N\to 0$ limit of perturbative gravity, the von Neumann entropy of the gravitationally dressed type II algebra equals the generalized entropy of the region up to a state-independent constant. The JSS construction assumes the existence of states whose modular flow is geometric (a flow generated by spacetime vector fields); for a cut of a Killing horizon — the stationary horizon generated by a Killing vector — such states exist in free and interacting theories because the averaged null energy condition (ANEC), the non-negativity of the null-null stress tensor integrated along a complete null geodesic, makes null translations act as half-sided modular inclusions: inclusions $\mathcal{N}\subset\mathcal{M}$ of von Neumann algebras whose modular automorphism groups move one algebra into the other for positive modular times.

Faulkner and Speranza specialized this program to arbitrary cuts of Killing horizons. They showed that the crossed-product entropy agrees with the generalized entropy of the cut to leading order in the semiclassical expansion parameter $\epsilon$, reproduced Wall's GSL as relative-entropy monotonicity of the underlying QFT algebras, and handled interacting asymptotically flat and de Sitter cases via an algebra at infinity associated with the half-sided translation. Going beyond leading order, they computed the first corrections to the crossed-product entropy, which arise at $O(\epsilon^2)$, but could not determine whether the entropy remains monotone between cuts. They also conjectured a subalgebra structure between nested horizon-cut crossed products, contingent on an operator-valued weight: a positive, unbounded linear map $T:\mathcal{M}\to\mathcal{N}$ generalizing a conditional expectation to the non-normalizable case, whose existence for the relevant half-sided modular inclusions is unproven.

Two further gaps involve states. A weight on a von Neumann algebra is a nonnegative extended-valued linear functional on its positive cone; a semifinite weight is one on which enough elements take finite values that Tomita–Takesaki theory (the theory of the modular automorphism group canonically attached to a state or weight) still applies. The natural QFT vacuum for rotating black holes is the Frolov–Thorne vacuum, which is regular near the future horizon but singular where the horizon-generating Killing vector becomes spacelike, and the natural QFT vacuum for collapse-formed black holes is the Unruh vacuum, which is regular on the future horizon and empty on past null infinity. The states whose modular flow is geometric on the horizon — the Frolov–Thorne vacuum for rotating black holes and the Hartle–Hawking vacuum for collapse-formed black holes — are not normalizable in the physical representation of the exterior algebra: for rotating black holes the Frolov–Thorne vacuum is singular where the horizon-generating Killing vector becomes spacelike, and for collapse-formed black holes the Hartle–Hawking vacuum is not normalizable in the representation adapted to the Unruh vacuum. Faulkner and Speranza therefore raised the possibility of reinterpreting these singular vacua as semifinite weights — the Frolov–Thorne vacuum for rotating black holes and the Hartle–Hawking vacuum in the representation adapted to the Unruh vacuum for collapse-formed black holes — in order to construct gravitational algebras for rotating and collapse black holes.

Finally, the quantum focusing conjecture (QFC) of Bousso, Fisher, Leichenauer and Wall asserts that the 'quantum expansion' — the derivative of generalized entropy along a light sheet, incorporating the classical null expansion plus a quantum entropy term — is non-increasing, a strengthening of the GSL that reduces to the quantum null energy condition (QNEC) on stationary backgrounds. Whether the QFC can be formulated and proven as a monotonicity statement for crossed-product entropies of general subregions, with UV finiteness supplied by the type II structure, was left open, as was the search for energy conditions guaranteeing states with geometric modular flow for generic subregions.

Problem Statement

The problem is the set of four linked open problems posed by Faulkner and Speranza in the discussion of arXiv:2405.00847 (JHEP 11 (2024) 099), which share a common motivation: determine the algebraic conditions under which generalized entropy — identified as the von Neumann entropy of gravitationally dressed crossed-product algebras of horizon cuts — is monotonic, and thereby place the generalized second law (GSL) and the quantum focusing conjecture (QFC) on a fully algebraic, UV-finite footing. The sub-problems are:

(i) All-orders monotonicity: determine whether the full crossed-product entropy, including the $O(\epsilon^2)$ corrections computed in Section 5 of the source, satisfies a monotonicity result between horizon cuts of a Killing horizon, that is, whether those corrections can be used to violate the leading-order second law. Since the leading-order statement follows from monotonicity of QFT relative entropy, a second law beyond leading order would require a modified bound on the amount by which the quantum-field relative entropy decreases between horizon cuts, a bound that in some cases would improve the standard data-processing inequality.

(ii) Operator-valued weight: prove or disprove the existence of an operator-valued weight $T:\mathcal{M}\to\mathcal{N}$ for the half-sided modular inclusions $\mathcal{N}\subset\mathcal{M}$ of the quantum field algebras of nested horizon cuts; if one exists, construct it explicitly, establish whether it is unique when the relative commutant is trivial (the interacting case) and pinned down by its restriction to the relative commutant (the free case), and determine the resulting second-law inequality for crossed-product entropies and whether it yields a stronger second law.

(iii) Semifinite gravitational algebras for rotating and collapse black holes: determine when the singular vacua of these spacetimes can be reinterpreted as semifinite weights on the physical von Neumann algebra of the black hole exterior — the Frolov–Thorne vacuum for rotating/Kerr black holes and, for collapse-formed black holes, the Hartle–Hawking vacuum in the representation adapted to the Unruh vacuum — including the identification of the correct definition subalgebra and the semicyclic representation analysis; where such a reinterpretation exists, carry out the detailed construction of the corresponding semifinite gravitational algebras, so that the crossed-product GSL extends to these states.

(iv) Algebraic quantum focusing: generalize the crossed-product approach to arbitrary subregions, and formulate and establish the quantum focusing conjecture as a monotonicity statement for crossed-product entropies of subregions bounded by causal horizons; and identify energy conditions (analogous to the role of the ANEC for horizon cuts) that guarantee the existence of states with geometric modular flow for generic subregions, as needed for the crossed-product construction there.

The four sub-problems are independent: a complete resolution of any one of (i)–(iv) is a valid solution to the task, and the shared motivation stated above is not an additional deliverable to be discharged in addition to the sub-problems.

References

1. T. Faulkner and A. J. Speranza, Gravitational algebras and the generalized second law, J. High Energy Phys. 11 (2024) 099, arXiv:2405.00847, doi:10.1007/JHEP11(2024)099, https://arxiv.org/abs/2405.00847
2. K. Jensen, J. Sorce and A. J. Speranza, Generalized entropy for general subregions in quantum gravity, J. High Energy Phys. 12 (2023) 020, arXiv:2306.01837, doi:10.1007/JHEP12(2023)020, https://arxiv.org/abs/2306.01837
3. E. Witten, Gravity and the crossed product, J. High Energy Phys. 10 (2022) 008, arXiv:2112.12828, doi:10.1007/JHEP10(2022)008, https://arxiv.org/abs/2112.12828
4. A. C. Wall, A proof of the generalized second law for rapidly changing fields and arbitrary horizon slices, Phys. Rev. D 85 (2012) 104049, arXiv:1105.3445, doi:10.1103/PhysRevD.85.104049, https://arxiv.org/abs/1105.3445
5. R. Bousso, Z. Fisher, S. Leichenauer and A. C. Wall, Quantum focusing conjecture, Phys. Rev. D 93 (2016) 064044, arXiv:1506.02669, doi:10.1103/PhysRevD.93.064044, https://arxiv.org/abs/1506.02669
6. J. Kirklin, Generalised second law beyond the semiclassical regime, J. High Energy Phys. 07 (2025) 192, arXiv:2412.01903, doi:10.1007/JHEP07(2025)192, https://arxiv.org/abs/2412.01903
7. X. Cao, T. Faulkner and Z. Wang, Gravitational algebras with two areas, arXiv:2512.04435, https://arxiv.org/abs/2512.04435
8. V. Chandrasekaran and É. É. Flanagan, Subregion algebras in classical and quantum gravity, arXiv:2601.07915, https://arxiv.org/abs/2601.07915
9. M. Ali and V. Suneeta, A local generalized second law in crossed product constructions, Phys. Rev. D 111 (2025) 024015, arXiv:2404.00718, doi:10.1103/PhysRevD.111.024015, https://arxiv.org/abs/2404.00718
10. S. Hollands, R. Longo and G. Morsella, A quantum dominant energy condition, arXiv:2608.01814, https://arxiv.org/abs/2608.01814
11. G. Satishchandran and J. Sorce, Uniqueness of null-local modular flow, arXiv:2607.20678, https://arxiv.org/abs/2607.20678
12. N. Engelhardt and H. Liu, The making of von Neumann algebras from bulk focusing, arXiv:2509.05413, https://arxiv.org/abs/2509.05413
13. M. S. Klinger, J. Kudler-Flam and G. Satishchandran, Generalized black hole entropy is von Neumann entropy. II. The complete symmetry group and edge modes, Phys. Rev. D 114 (2026) 065010, arXiv:2601.07910, doi:10.1103/tb6y-lsfv, https://arxiv.org/abs/2601.07910
14. C.-H. Chen and G. Penington, A clock is just a way to tell the time: gravitational algebras in cosmological spacetimes, arXiv:2406.02116, https://arxiv.org/abs/2406.02116
15. A. Rignon-Bret and M. Vilatte, Black hole thermodynamics at null infinity. Part 1: Dual generalized second law, arXiv:2601.03353, https://arxiv.org/abs/2601.03353