Chaos-Bound Violations in Einstein Gravity

• The paper demonstrates that extended gravitational theories can lead to explicit chaos-bound violations, where the classical Lyapunov exponent surpasses the universal bound near the event horizon.
• Detailed near-horizon analyses using effective Lagrangians and metric expansions show how surface gravity and subleading corrections determine the instability parameters in various black hole geometries.
• The findings bridge classical chaos and quantum information limits, offering new insights into the role of curvature corrections and orbital constraints in refining holographic duality models.

Chaos-bound violations in Einstein gravity designate instances where the classical Lyapunov exponent governing the instability of a particle or string probe near a black hole exceeds the "universal bound" $\lambda \leq 2\pi T_H$ conjectured by Maldacena–Shenker–Stanford for quantum many-body systems. In gravitational backgrounds described by Einstein gravity and its generalizations, the bound is generally saturated at the event horizon, but explicit violations occur for extended theories, nontrivial matter couplings, and specific geometrical regimes. The phenomena have substantial implications for understanding the relationship between classical chaos in strong gravity, curvature corrections, and quantum information-theoretic bounds.

1. Chaos Bound: Fundamentals and Near-Horizon Behavior

The chaos bound in gravitational systems formalizes a maximum Lyapunov exponent $\lambda$ for the exponential growth of perturbations, conjectured to satisfy $\lambda \leq 2\pi T_H$ with $T_H$ the Hawking temperature of the black hole horizon. For charged massive particles in static, spherically symmetric backgrounds, the equilibrium condition arises from the balance of gravitational and Lorentz forces. Small radial perturbations near the equilibrium point $r_0$ are governed by an effective Lagrangian,

$$L_{\mathrm{eff}} \supset \frac{1}{2\sqrt{h(r_0)}f(r_0)} \left(\dot{\epsilon}^2 + \lambda^2 \epsilon^2\right),$$

with $\lambda^2$ determined by the local derivatives of metric and gauge potential functions (Zhao et al., 2018).

Expanding the metric near the event horizon $r_+$ yields $f(r) = f_1(r - r_+) + f_2(r - r_+)^2 + \ldots$ and similarly for other functions. The Lyapunov exponent admits a near-horizon expansion

$$\lambda^2 = \kappa^2 + \gamma(r_0 - r_+) + \mathcal{O}\left((r_0 - r_+)^2\right),$$

where $\lambda$0 is the surface gravity. Universally, $\lambda$1 as $\lambda$2, and the bound is saturated. The sign of the subleading coefficient $\lambda$3 determines the possibility and locality of violations near the horizon.

2. Exact Satisfaction and Systematic Violations in Einstein Gravity Extensions

Specific black-hole backgrounds demonstrate either exact satisfaction or violation of the chaos bound:

• In Reissner–Nordström (RN) and RN–AdS geometries, saturation or $\lambda$4 is manifest for all physically admissible unstable equilibria, with explicit closed forms available for $\lambda$5 and $\lambda$6 (Zhao et al., 2018, Targema et al., 28 Dec 2025).
• In RN–de Sitter (RN–dS), Einstein–Maxwell–Dilaton (EMD), Einstein–Born–Infeld (EBI), and Einstein–Gauss–Bonnet–Maxwell backgrounds, explicit analysis reveals domains or parametric regimes where $\lambda$7:
  • For RN–dS and extremal black holes ($\lambda$8), $\lambda$9 remains finite, producing an infinite ratio $\lambda \leq 2\pi T_H$0.
  • EMD and Gauss–Bonnet coupling can cause $\lambda \leq 2\pi T_H$1 to exceed $\lambda \leq 2\pi T_H$2 away from the horizon, with near-extremality facilitating the violation (Zhao et al., 2018, Xie et al., 2023).
  • In EBI, sufficiently large nonlinear parameter or charge triggers domains with $\lambda \leq 2\pi T_H$3, especially near extremality.

These findings indicate that the chaos bound is neither globally nor absolutely universal among all Einstein gravity extensions. A refined scaling bound $\lambda \leq 2\pi T_H$4 (order-one constant) may still hold at large $\lambda \leq 2\pi T_H$5, but is not fundamental in the extremal regime.

3. Angular Momentum Constraints and Apparent vs. Genuine Violations

The computation of the Lyapunov exponent and detection of chaos-bound violation critically depend on how the angular momentum $\lambda \leq 2\pi T_H$6 of the probe is treated:

• If $\lambda \leq 2\pi T_H$7 is regarded as a free parameter, as in many older analyses, $\lambda \leq 2\pi T_H$8 can exceed $\lambda \leq 2\pi T_H$9 for unphysical near-horizon circular orbits of charged particles, producing apparent (spurious) violations (Targema et al., 28 Dec 2025).
• Using a fully constrained approach, $T_H$0 must satisfy the exact circular-orbit condition derived from the background geometry, ensuring only physically realizable orbits are considered. In this framework, genuine chaos-bound violations can be isolated to higher-curvature extensions where the metric itself allows for the instability at admissible $T_H$1 and $T_H$2 (e.g., quadratic-curvature, Gauss–Bonnet) (Targema et al., 28 Dec 2025).

The table below summarizes the distinction:

Treatment of $T_H$3	Einstein Gravity Bound Status	Violation Type
$T_H$4 arbitrary	$T_H$5 possible	Apparent (spurious)
$T_H$6 circular-orbit	$T_H$7 (RN, Kiselev, ...)	No physical violation
Higher-curvature ext.	$T_H$8 at admissible $T_H$9	Genuine

4. Chaos Bound Violations in Extended Theories and Parametric Dependencies

Explicit analytical and numerical work has revealed how various physical parameters determine the region and magnitude of chaos-bound violations:

• In Einstein–Gauss–Bonnet–AdS backgrounds, increasing the Gauss–Bonnet coupling $r_0$0 drives the system closer to extremality and decreases $r_0$1, facilitating violations at smaller particle charge $r_0$2 and wider ranges of $r_0$3 (Xie et al., 2023). The Lyapunov exponent for circular orbits is

$r_0$4

where the violation criterion is algebraic, depending explicitly on $r_0$5 and $r_0$6.

• Similar parametric studies in Einstein–Euler–Heisenberg, Kiselev, and p-brane gravitational backgrounds expose regions in particle parameters (charge $r_0$7, angular momentum $r_0$8), black hole parameters (mass $r_0$9, charge $$L_{\mathrm{eff}} \supset \frac{1}{2\sqrt{h(r_0)}f(r_0)} \left(\dot{\epsilon}^2 + \lambda^2 \epsilon^2\right),$$0, coupling constants), and spacetime attributes, which delimit whether and where the chaos bound is violated (Chen et al., 2022, Gao et al., 2022, Dutta et al., 2024).

5. Rotating Black Holes: Multimode OTOCs and Apparent Contradictions

In rotating BTZ black holes (3D Einstein gravity), the out-of-time-order correlator (OTOC) displays two Lyapunov exponents, $$L_{\mathrm{eff}} \supset \frac{1}{2\sqrt{h(r_0)}f(r_0)} \left(\dot{\epsilon}^2 + \lambda^2 \epsilon^2\right),$$1 with $$L_{\mathrm{eff}} \supset \frac{1}{2\sqrt{h(r_0)}f(r_0)} \left(\dot{\epsilon}^2 + \lambda^2 \epsilon^2\right),$$2 the angular velocity and $$L_{\mathrm{eff}} \supset \frac{1}{2\sqrt{h(r_0)}f(r_0)} \left(\dot{\epsilon}^2 + \lambda^2 \epsilon^2\right),$$3 the AdS radius (Jahnke et al., 2019). One exponent may exceed the conventional bound $$L_{\mathrm{eff}} \supset \frac{1}{2\sqrt{h(r_0)}f(r_0)} \left(\dot{\epsilon}^2 + \lambda^2 \epsilon^2\right),$$4, but this reflects the splitting into effective thermal subsystems (left and right moving), each saturating its own bound. Thus, when effective inverse temperatures $$L_{\mathrm{eff}} \supset \frac{1}{2\sqrt{h(r_0)}f(r_0)} \left(\dot{\epsilon}^2 + \lambda^2 \epsilon^2\right),$$5 are used, no physical violation occurs. This decoupling is peculiar to 3D gravity and does not naturally occur in higher-dimensional Einstein gravity.

6. Physical Interpretation, Universality, and Implications

Chaos-bound violations in Einstein gravity and its extensions highlight the sensitivity of classical chaos—and its relationship to quantum many-body conjectures—to the detailed structure of spacetime, matter content, and probe parameters. In pure Einstein gravity, once orbital constraints are imposed, the MSS bound is robust for all allowed equilibria. Genuine violations serve as probes of higher-derivative corrections and non-minimal couplings—parametrizing a regime where curvature-induced modifications to instability transcend redshift-suppressed limits imposed by the horizon.

In string and brane settings, both particle and extended-string probes manifest violations in near-extremal and extremal backgrounds (Dutta et al., 2024). These findings delineate the boundary between classical instability and quantum information bounds, crucial for a nuanced understanding of chaos in holographic duality and quantum gravity.

7. Summary Table: Major Results Across Geometries

Geometry	Chaos Bound Status	Mechanism	Key Parameters/Regime
RN, RN–AdS	Saturation or always $$L_{\mathrm{eff}} \supset \frac{1}{2\sqrt{h(r_0)}f(r_0)} \left(\dot{\epsilon}^2 + \lambda^2 \epsilon^2\right),$$6	No violation	All non-extremal
RN–dS, EMD, EBI, GB extensions	Violations at/away from horizon	Geometry or coupling	Large $$L_{\mathrm{eff}} \supset \frac{1}{2\sqrt{h(r_0)}f(r_0)} \left(\dot{\epsilon}^2 + \lambda^2 \epsilon^2\right),$$7, near extremal, strong coupling
Kiselev (Einstein)	No genuine violation with correct $$L_{\mathrm{eff}} \supset \frac{1}{2\sqrt{h(r_0)}f(r_0)} \left(\dot{\epsilon}^2 + \lambda^2 \epsilon^2\right),$$8	Orbital constraint	$$L_{\mathrm{eff}} \supset \frac{1}{2\sqrt{h(r_0)}f(r_0)} \left(\dot{\epsilon}^2 + \lambda^2 \epsilon^2\right),$$9 fixed by $\lambda^2$0
Extended quadratic curvature	Genuine violation possible	Curvature corrections	Large charge-to-mass, higher derivative
Rotating BTZ	Apparent violation resolved	Multiple temperatures	$\lambda^2$1
Black $\lambda^2$2-brane (D=10)	Violation for extremal $\lambda^2$3	Near horizon, extended probe	Extemality, string winding

This synthesis underscores that the chaos bound in Einstein gravity is a robust but not universal constraint, and systematic violation diagnoses encode both the geometric origin of chaos and the demarcation between classical probes and quantum informational limits (Zhao et al., 2018, Targema et al., 28 Dec 2025, Xie et al., 2023, Chen et al., 2022, Gao et al., 2022, Dutta et al., 2024, Jahnke et al., 2019).