Bekenstein Bound in QFT and Gravity

• Bekenstein Bound is a theoretical limit on a system's entropy based on its energy and size, originally proposed to safeguard the generalized second law in black-hole physics.
• Modern formulations use modular Hamiltonians and relative entropy in quantum field theory to resolve UV divergences and ambiguities, providing a precise statement of the bound.
• Extensions in AQFT, holography, and gravitational thermodynamics reveal its role in constraining hidden bulk information and setting operational limits on information transfer.

The Bekenstein bound is a proposed upper limit on the entropy $S$ of a physical system of total energy $E$ and characteristic size $R$, conventionally written as

$S \le 2\pi E R$

in units with $\hbar=c=k_B=1$. It originated in black-hole thought experiments designed to protect the generalized second law, but in modern quantum field theory it is usually interpreted through vacuum subtraction, modular Hamiltonians, and relative entropy rather than as a naïve inequality for global thermodynamic entropy. In that modern form, the bound becomes a precise statement about localized distinguishability from the vacuum, with rigorous versions in algebraic quantum field theory, explicit realizations in conformal field theory and holography, and a range of extensions and controversies concerning species, negative energy, generalized entropies, and operational information measures (0804.2182).

1. Original statement and physical content

Bekenstein’s original proposal concerns an isolated, weakly self-gravitating system of entropy $S$, energy $E$, and circumscribing radius $R$, with the canonical inequality

$S \le 2\pi E R.$

Its original motivation is the generalized second law: if a system is lowered into a black hole, the increase in black-hole entropy must compensate for the disappearance of ordinary entropy from the exterior region (Page, 2018).

For a Schwarzschild black hole, the Bekenstein–Hawking entropy

$S_{BH}=\frac{k_B c^3 A}{4G\hbar}$

saturates the bound when $E$0 is taken to be the Schwarzschild radius and $E$1. This saturation is one of the main reasons the bound has often been regarded as a universal entropy limit rather than a property of a special class of systems (2207.13652).

From the beginning, however, the flat-space formulation was beset by ambiguities. In continuum QFT, it is nontrivial to define the “system,” the relevant entropy, the energy assigned to a spatial region, and even the appropriate size $E$2. Entanglement across boundaries makes reduced-state von Neumann entropies ultraviolet divergent, and different prescriptions for localization lead to inequivalent notions of $E$3, $E$4, and $E$5 (0804.2182). Page emphasized that one natural “vacuum-outside-$E$6” definition of localized complete systems in Minkowski space still admits explicit counterexamples to the naïve global inequality $E$7, even when the state is vacuum-like outside a ball of radius $E$8 (Page, 2018).

This historical tension is central: the original bound is physically compelling, but its mathematically precise meaning depends on the framework in which entropy, energy, and localization are defined.

2. Relative entropy and the modern QFT formulation

The modern formulation is due to the observation that, for a region $E$9, the relative entropy between a state $R$0 and a reference state $R$1 obeys

$R$2

where $R$3 is the modular Hamiltonian. Taking $R$4 to be the vacuum reduced to $R$5, one obtains the finite, regulator-independent inequality

$R$6

which Casini identified as the precise QFT version of the Bekenstein bound (0804.2182).

This reformulation resolves several classical difficulties at once. First, the subtraction by the vacuum removes the universal UV-divergent boundary contribution to localized entropy. Second, the right-hand side is not an ambiguous product $R$7, but a modular-energy expectation value determined by the geometry of the region. Third, the species problem is softened because the relevant quantity is relative entropy, whose convexity and monotonicity control distinguishability from the vacuum rather than raw state counting (0804.2182).

The same relative-entropy perspective also clarifies why the naïve global bound can fail while the modular version remains valid. Page’s counterexamples target the global statement with total von Neumann entropy and total energy of “vacuum-outside-$R$8” states, whereas the proven inequality is subregional and vacuum-subtracted: $R$9 In this sense, the modern bound is not merely a correction to Bekenstein’s proposal; it is a redefinition of what the bound means in relativistic QFT (Page, 2018).

A further refinement uses monotonicity of relative entropy for nested regions. Blanco and Casini derived

$S \le 2\pi E R$0

where $S \le 2\pi E R$1 is a “free” entropy combination that is insensitive to boundary entanglement. This strengthened Bekenstein-type inequality constrains entropy localized between two boundaries and sharpens bounds on the spatial dispersal of negative energy (Blanco et al., 2013).

3. Explicit modular Hamiltonians, geometric realizations, and holography

The bound becomes especially transparent when the vacuum modular Hamiltonian is local. For the Rindler wedge in Minkowski space,

$S \le 2\pi E R$2

and for a ball of radius $S \le 2\pi E R$3 in a CFT,

$S \le 2\pi E R$4

In these cases, the modular-energy term reduces to an explicitly geometry-weighted energy integral, recovering the traditional $S \le 2\pi E R$5 scaling for suitably localized excitations (0804.2182).

Holography provides a geometrization of this structure. For a spherical boundary cap of opening half-angle $S \le 2\pi E R$6 in global AdS, the associated RT surface obeys

$S \le 2\pi E R$7

and reaches a deepest bulk point with effective radius

$S \le 2\pi E R$8

The region deeper than that point cannot be reconstructed from boundary regions smaller than the cap; this is called the sphere of ignorance (Ilgin, 2018).

For spherically symmetric excitations, the change in the full modular Hamiltonian of complementary boundary regions has the bulk interpretation

$S \le 2\pi E R$9

and the boundary entanglement entropies satisfy

$\hbar=c=k_B=1$0

The same work argues that the bulk entropy inside the sphere of ignorance is bounded by the same quantity, giving a holographic realization of a Bekenstein-type bound in AdS/CFT (Ilgin, 2018).

This holographic picture also sharpens the distinction between pure and mixed excitations. For global pure states, $\hbar=c=k_B=1$1, so the entropy difference vanishes even when the modular-energy difference is nonzero. For thermal or mixed excitations, by contrast, the entropic difference can saturate the bulk $\hbar=c=k_B=1$2-type relation at linear order (Ilgin, 2018). This suggests that, holographically, the Bekenstein bound controls unreconstructable or hidden bulk information rather than generic entanglement alone.

4. Rigorous AQFT formulations and one-particle realizations

A major development is the existence of fully rigorous Bekenstein-type bounds in algebraic quantum field theory. Longo proved that if $\hbar=c=k_B=1$3 is a spacetime region of width $\hbar=c=k_B=1$4 in Minkowski space and $\hbar=c=k_B=1$5 is a vector state localized in $\hbar=c=k_B=1$6, then the vacuum relative entropy on the local algebra satisfies

$\hbar=c=k_B=1$7

where $\hbar=c=k_B=1$8 is the minimal total energy among states indistinguishable from $\hbar=c=k_B=1$9 outside $S$0. The result is model-independent and uses only locality, translation covariance with positive energy, wedge duality, and modular theory; no stress tensor or explicit modular Hamiltonian is required (Longo, 2024).

The proof proceeds through a standard-subspace formulation and a geometric operator inequality

$S$1

with $S$2 the Hamiltonian. This produces a genuine upper bound on vacuum relative entropy in terms of energy and width, complementing Casini’s formulation, which interprets the Bekenstein statement as positivity of relative entropy when the modular Hamiltonian is known (Longo, 2024).

A closely related one-particle realization has been given for Klein–Gordon wave packets. If a wave packet $S$3 is localized in a spatial region $S$4 of half-width $S$5, then

$S$6

where $S$7 is defined through the entropy operator of the standard subspace $S$8, and

$S$9

For wedges one has the exact identity

$E$0

from which the localized bound follows by monotonicity (Hollands et al., 3 Feb 2026).

When the wave packet is not strictly localized in $E$1, the inequality acquires a boundary correction: $E$2 where $E$3 is determined by a variational problem depending only on the boundary values of the Cauchy data on $E$4 (Hollands et al., 3 Feb 2026). This is a particularly sharp formulation because it isolates the exact obstruction to the strict local bound.

These AQFT and one-particle results show that the Bekenstein bound is not merely heuristic or CFT-specific. In an operator-algebraic setting it becomes a theorem about relative entropy, modular operators, and localization width.

5. Black holes, generalized bounds, and operational interpretations

In black-hole thermodynamics the bound is tied not only to entropy storage but also to dissipation rates. Bekenstein derived a universal upper bound on information emission rate, and Hod recast it as

$E$5

with $E$6 the inverse relaxation time and $E$7 the system temperature. For Kerr black-hole ringdowns, this becomes a bound involving the Hawking temperature and the imaginary part of the least-damped quasinormal mode (Carullo et al., 2021).

A time-domain analysis of LIGO–Virgo binary-black-hole remnants tested this Bekenstein–Hod bound observationally. The study reported that the observed remnant population obeys the bound with population-level probability

$E$8

providing an astrophysical confirmation of this dynamical version of the Bekenstein idea (Carullo et al., 2021).

Operationally, the question “what exactly does Bekenstein bound?” has also been investigated in communication-theoretic terms. In the Unruh-channel setting, classical and quantum capacities available to an accelerated decoder obey a Bekenstein-type bound scaling like $E$9, with $R$0 the inverse Unruh temperature and $R$1 the decoder’s accessible energy. However, the zero-bit capacity can violate this decoder-only bound; only when both encoder and decoder are spatially constrained does a Blanco–Casini-type inequality force all capacities, including zero-bits, below a $R$2 bound (Hayden et al., 2023).

This suggests that the bound does not universally limit every operational notion of information. It constrains recoverable information under specified localization and energy restrictions, but not necessarily all communication resources unless both encoding and decoding are geometrically controlled (Hayden et al., 2023).

A different line of work proposes a microscopic origin of the bound from fermionic fundamental degrees of freedom. In a toy model of black-hole evaporation with $R$3 fermionic “Xons,” Pauli exclusion yields a maximal entropy

$R$4

and identifying $R$5 with Planck-area cells reproduces the black-hole saturation of the Bekenstein bound (Acquaviva et al., 2020). This does not derive the field-theoretic modular statement, but it offers a distinct microscopic interpretation of why an entropy cap might arise.

6. Extensions, deformations, and persistent controversies

Several recent works study deformations of the bound rather than its standard QFT form. With a generalized uncertainty principle,

$R$6

the thermodynamic derivation of the entropy limit yields a modified Bekenstein bound. For $R$7,

$R$8

while for $R$9,

$S \le 2\pi E R.$0

Thus positive-$S \le 2\pi E R.$1 GUP tightens the bound and negative-$S \le 2\pi E R.$2 GUP loosens it (Buoninfante et al., 2020).

Generalized black-hole entropies can violate the standard bound outright. For Schwarzschild black holes described by Barrow, Tsallis, or Kaniadakis entropies, the standard $S \le 2\pi E R.$3 may fail over the parameter ranges analyzed, with saturation recovered only in the Boltzmann–Gibbs limit $S \le 2\pi E R.$4, $S \le 2\pi E R.$5, or $S \le 2\pi E R.$6 (2207.13652). A related analysis showed that these violations can be compensated by incorporating GUP corrections, yielding explicit relations between the entropy-deformation parameters and the GUP parameter $S \le 2\pi E R.$7 that restore a generalized bound (Shokri, 2024).

A different resolution has been proposed via the equivalence between generalized entropy and varying-$S \le 2\pi E R.$8 gravity. In that framework, if the entropy is replaced by a generalized form $S \le 2\pi E R.$9, then gravity is modified simultaneously through

$S_{BH}=\frac{k_B c^3 A}{4G\hbar}$0

and the relevant thermodynamic energy differs from the ADM mass. The resulting entropy law obeys a relaxed Bekenstein bound

$S_{BH}=\frac{k_B c^3 A}{4G\hbar}$1

where the coefficient $S_{BH}=\frac{k_B c^3 A}{4G\hbar}$2 is no longer fixed at $S_{BH}=\frac{k_B c^3 A}{4G\hbar}$3 but remains finite and reduces to $S_{BH}=\frac{k_B c^3 A}{4G\hbar}$4 in the GR limit (Lu et al., 6 May 2025).

There is also a complementary geometric program deriving energy–size–charge–angular-momentum inequalities suggested by generalized Bekenstein bounds. For axisymmetric bodies satisfying appropriate energy conditions, one finds

$S_{BH}=\frac{k_B c^3 A}{4G\hbar}$5

which supports the most general Bekenstein-type form from classical geometric analysis rather than entropy arguments (Jaracz et al., 2018).

Taken together, these developments indicate that the Bekenstein bound is best understood not as a single immutable inequality, but as a family of structurally related entropy–energy–size constraints whose precise form depends on the notion of entropy, the localization framework, and whether one is working in semiclassical QFT, AQFT, holography, or deformed gravitational thermodynamics.

7. Status and outlook

The contemporary status of the Bekenstein bound is bifurcated. On one side, the naïve global statement $S_{BH}=\frac{k_B c^3 A}{4G\hbar}$6 is not universally valid in nongravitational QFT under arbitrary definitions of localized entropy and energy (Page, 2018). On the other side, relative-entropy and modular-theoretic formulations provide precise, regulator-independent, and in some settings rigorous inequalities that capture the same physical intuition (0804.2182, Longo, 2024).

This suggests that the lasting content of the bound is not a crude cap on thermodynamic entropy, but a constraint on how much distinguishability from the vacuum can be concentrated in a bounded region at fixed energy scale. In holography, that content becomes a statement about hidden bulk information and entanglement wedges (Ilgin, 2018). In AQFT, it becomes an operator inequality for modular Hamiltonians and local energies (Longo, 2024). In black-hole dynamics, it becomes a bound on dissipation and information emission rates (Carullo et al., 2021). In operational settings, it constrains some communication resources but not all unless localization assumptions are strong enough (Hayden et al., 2023).

Open directions follow directly from this landscape. These include extending rigorous results beyond standard-subspace and one-particle settings, characterizing saturation, understanding generalized entropies without ad hoc deformations, and clarifying how Bekenstein-type bounds interface with QNEC, bulk reconstruction, and nonperturbative quantum gravity. The enduring lesson is that the bound survives, but chiefly in modular, relative, and geometric form rather than as a universally valid inequality for raw global entropy.