Holographic Timelike Entanglement and Subregion Complexity in Localized AdS3*S3*T4 Black Holes

Technical Summary: Holographic Timelike Entanglement and Subregion Complexity in Localized AdS$_3 \times S^3 \times T^4$ Black Holes

Problem Statement
The paper investigates how holographic observables sensitive to Lorentzian structure can detect the internal angular geometry of localized black holes in type IIB supergravity. While the standard BTZ black hole uplift is homogeneous over the internal $S^3 \times T^4$, localized solutions (specifically the "black pole") exhibit horizons that are non-uniform and localized on the internal sphere. The authors ask whether timelike entanglement entropy and timelike subregion complexity—observables constructed from spacelike and timelike Lorentzian branches rather than a single spatial extremal surface—can probe the specific angular dependence of the localized geometry, which is absent in the standard BTZ description.

Methodology
The authors employ a "localized timelike prescription" adapted from previous work on localized RT surfaces. The methodology involves several distinct steps:

1. Geometry: The study focuses on the black-pole solution of the D1-D5 system with asymptotic $AdS_3 \times S^3 \times T^4$. The metric depends on the internal Hopf angle $\theta$ through warp factors $K_y(r, \theta)$ and $G(r, \theta)$. This geometry features a "cap-side" (smooth degeneration of the spatial circle) and a "horizon-side" (degeneration of the time circle), separated by a transition angle $\theta_\star$.
2. Lorentzian Branch Construction: For a boundary interval extended in time ($-T/2 \le t \le T/2$), the bulk surface is constructed from two branches in the effective $(t, r)$ plane:
   • A spacelike branch reaching the asymptotic boundary.
   • A timelike branch ending at a radial turning point $r_0$.
   • Crucially, the reduced branch profile is first solved at a fixed angular label $\theta_0$. The physical internal angle $\theta$ is only reintroduced during the "lift" to ten dimensions, where the area or volume is integrated over the full internal sphere.
3. Fixed-Boundary-Interval Selection: Because the exact black-pole geometry renders the time map $T(r_0, \theta_0)$ non-monotonic, a single boundary interval $T$ can correspond to multiple radial branches and angular labels. The authors enforce a fixed-boundary-interval condition: for a target $T_{target}$, they solve for all admissible $(r_0, \theta_0)$ pairs and then select the saddle by minimizing the relevant observable.
4. Observables:
   • Timelike Entanglement Entropy (TEE): Calculated as a lifted area. The result is generally complex, with the real part determined by UV-renormalized spacelike contributions and the imaginary part arising from timelike branches and sign changes in the lifted integrand.
   • Timelike Subregion Complexity: Calculated as a finite, renormalized lifted volume. This observable is real and defined by a signed combination of spacelike and timelike weighted volumes after subtracting the universal boundary divergence.

Key Contributions and Results

• Non-Monotonic Time Map and Branch Multiplicity: In the exact black-pole geometry, the relationship between the boundary interval $T$ and the turning point $r_0$ is non-monotonic. For a fixed angular label $\theta_0$, a single $T$ can be realized by two distinct radial branches (a "smaller-$r_0$" and a "larger-$r_0$" branch relative to the peak of the time map). This necessitates a two-step minimization: first solving the time equation to find admissible branches, then minimizing the observable among them.
• Angular Accessibility and the Transition Region: The maximum accessible boundary interval $T_{max}(\theta_0)$ depends on the angular label. As $T_{target}$ increases, the admissible angular labels are restricted to a narrow region near the cap/horizon transition angle $\theta_\star$. The authors attribute this to a "crossover scale" $r_c(\theta_0) \sim \ell |\theta_0 - \theta_\star|$. Near $\theta_\star$, a transition radial interval opens where the timelike branch kernel has a larger logarithmic coefficient than the spacelike one, generating a positive logarithmic contribution to the time interval. This mechanism allows larger intervals to be supported near $\theta_\star$ from both the cap and horizon sides.
• Timelike Entanglement Results:
  • The lifted area becomes complex. The real part is minimized to select the saddle, while the imaginary part is evaluated on the same surface.
  • The selected saddle moves inward (decreasing $r_0$) and its angular label $\theta_0$ approaches $\theta_\star$ as the boundary interval grows.
  • Unlike the reduced branch level where spacelike is real and timelike is imaginary, the ten-dimensional lift mixes these contributions: both branches can contribute to the real and imaginary parts of the final area depending on the sign of the lifted integrand over the internal sphere.
• Timelike Subregion Complexity Results:
  • The observable is a real, finite renormalized volume.
  • The selected complexity also exhibits an inward motion of the turning point and an angular concentration near $\theta_\star$ as $T_{target}$ increases.
  • The selected saddle remains on the "larger-$r_0$" branch (relative to the time map peak) throughout the studied range, even as the numerical value of $r_0$ decreases.
  • The non-zero finite volume is a localized-geometry effect arising from the exterior patch of the black pole, distinct from the interior volume growth typically associated with eternal BTZ black holes.
• Asymptotic vs. Exact Regimes: In the large-$r$ (asymptotic) regime, the angular dependence drops out, the time map becomes single-valued, and the results recover standard short-interval BTZ-like behavior (logarithmic real part for TEE, vanishing volume for complexity). The distinct features (non-monotonicity, angular restriction, inward motion) only emerge when the exact localized functions $K_y$ and $G$ are restored.

Significance and Claims
The paper claims that timelike Lorentzian observables provide a complementary probe to spatial extremal surfaces for detecting localized geometric structures. Specifically:

1. Detection of Internal Structure: The exact localized geometry leaves a clear imprint on timelike observables that is absent in the BTZ benchmark and the leading large-$r$ description. The observables are sensitive to the cap/horizon transition region, which controls the admissible branch families.
2. Complementary Probes: Timelike entanglement and timelike complexity probe the same Lorentzian branch geometry but respond differently. TEE records the localized effects through the complex structure of a lifted area (real and imaginary parts), while complexity records them through a real, finite, signed volume.
3. Localized Geometry Effects: The results demonstrate that the "branch folding," angular restrictions, and inward motion of saddles are direct consequences of the exact functions $K_y(r, \theta)$ and $G(r, \theta)$. These effects are intrinsic to the localized black-pole solution and are not artifacts of the asymptotic approximation.
4. Limitations: The authors note that their complexity calculation is a finite renormalized volume associated with the exterior patch and a fixed boundary interval, distinct from the global interior complexity of an eternal black hole. The observed non-zero volume is a property of the localized exterior geometry.

The paper concludes that these timelike observables offer a robust method to probe the breaking of internal $S^3$ symmetry and the non-uniform distribution of horizons in localized black holes, suggesting that Lorentzian observables are well-suited for detecting boundary signatures of such localized structures beyond universal asymptotic data.