Flux-Averaged Force Multipliers

1. Problem Statement

Active Galactic Nuclei (AGN) drive powerful outflows (winds) via radiation pressure, specifically through line driving (momentum transfer from photons to gas via absorption in spectral lines). A critical challenge in modeling these winds is the inconsistency between the Spectral Energy Distribution (SED) used in photoionization calculations and the local radiation field experienced by the gas in hydrodynamic simulations.

• The "Grey" Limitation: Previous studies (e.g., DDD25) often used a "grey" or global SED approach. They assumed a single, fixed SED (e.g., an obscured AGN spectrum) to calculate the force multiplier ($M$)—a measure of how efficiently radiation couples to gas via lines—and applied this globally.
• The Physical Reality: In AGN disc winds, the local SED varies significantly with position. Near the disc, the radiation is dominated by optical/UV blackbody emission from the disc itself, while X-rays dominate from the corona. A global SED fails to capture these local variations, leading to inaccurate predictions of wind launching, mass flux, and velocity.
• The "Line Mismatch" Problem: The strongest spectral lines may not align with the peak of the irradiating SED at a specific location. A global average can erroneously suppress or enhance line driving depending on the mismatch between the assumed SED and the local flux distribution.

2. Methodology

The authors propose a Flux-Averaged Force Multiplier approach to bridge the gap between detailed photoionization physics and computationally feasible hydrodynamic simulations.

• Photoionization Modeling:
  • Used the code XSTAR to generate a grid of photoionization models for two template AGN SEDs (AGN1: unobscured; AGN2: obscured).
  • Calculated force multipliers for four distinct frequency bands: Infrared (IR), Optical (O), Ultraviolet (UV), and X-rays (X).
  • The grid covers a wide range of optical depth parameters ($\log t$) and ionization parameters ($\log \xi$).
• The Flux-Averaged Approximation:
  • Instead of assuming a single global SED, the method calculates the total radiation force as a sum of contributions from each band:
    $$a_{tot} = \frac{\sigma_e}{c} \sum_{\nu} F_{\nu} \frac{1}{f_{\nu}} M_{\nu}(t, \eta)$$
    Where $F_{\nu}$ is the local flux in band $\nu$, $f_{\nu}$ is the fraction of flux in that band for the reference SED, and $M_{\nu}$ is the force multiplier for that band.
  • Key Assumption: The local SED changes the available momentum (via $F_{\nu}$) but does not significantly alter the efficiency of momentum transfer (the force multiplier $M_{\nu}$), which is pre-calculated using a fixed reference SED. This avoids the prohibitive computational cost of recalculating photoionization tables at every time step and spatial cell.
• Hydrodynamic Simulations:
  • Implemented in the Athena++ radiation hydrodynamics code.
  • Simulated AGN disc winds around black holes with masses ranging from $3 \times 10^6$ to $10^9 M_{\odot}$.
  • Compared several models:
    1. Grey: Standard approach using a global force multiplier.
    2. Spectral Models: Models including specific band combinations (UV only, O+UV, UV+X, and a full 4-Band model including IR, O, UV, X).

3. Key Contributions

• Novel Approximation Scheme: Introduced a computationally tractable method to incorporate frequency-dependent radiation forces into hydrodynamic simulations without requiring full time-dependent photoionization coupling at every step.
• Multi-Band Resolution: Explicitly demonstrated that the Optical (O) and X-ray bands, often neglected or treated as secondary, play non-negligible roles in wind dynamics, particularly in specific regions of the disc wind.
• Position-Dependent Physics: Showed how the local radiation field (e.g., blackbody-like near the disc vs. power-law-like further out) fundamentally alters the wind launching mechanism compared to global SED assumptions.

4. Key Results

• Enhanced Wind Strength:
  • The Grey model produced weak, episodic winds with velocities $\sim 5,500$ km/s and low mass fluxes. The force multiplier was only sufficient to launch winds at the very base of the disc.
  • Spectral models (especially the 4-Band and O-UV models) produced robust, sustained winds with velocities reaching 28,000 – 32,000 km/s and mass fluxes ~100 times higher than the Grey model.
• Role of the Optical (O) Band:
  • Contrary to the assumption that UV dominates entirely, the Optical band becomes the dominant driver in regions near the disc (radii $r \sim 1000 r_g$) where the local SED peaks in the optical.
  • Including the O band increased the total force multiplier ($\log M$) from $\sim 2$ (UV-only) to $\sim 3$ in these regions, significantly boosting mass loading.
• Role of X-rays:
  • X-rays contribute modestly ($\sim 20\%$) to the force multiplier in the outer wind regions. While they do not drive mass loading, they are crucial for accelerating the wind to higher terminal velocities.
• Wind Morphology:
  • Grey models produced narrow, radial winds ($\omega \approx 68^\circ$).
  • Spectral models produced wider, more vertical winds ($\omega \approx 52^\circ - 56^\circ$) due to the enhanced radiation pressure from the disc surface (vertical component).
• Mass Dependence:
  • For lower mass black holes ($M < 10^7 M_{\odot}$), X-rays dominate the inner disc, while for higher masses ($M > 10^8 M_{\odot}$), UV dominates. The flux-averaged method correctly adapts to these shifts.

5. Significance and Future Directions

• Validation of Frequency Dependence: The study confirms that "grey" radiation modeling is insufficient for accurate AGN wind predictions. Frequency-dependent effects can change wind properties by orders of magnitude.
• Bridging Scales: The flux-averaged approach offers a practical middle ground between overly simplified analytic fits and prohibitively expensive full Monte Carlo radiation transfer (rad-MC) coupled with photoionization.
• Future Improvements: The authors acknowledge the limitation that the local SED is not used to recalculate the ionization balance (opacity) in real-time. Future work aims to:
  • Develop Machine Learning (ML) surrogates trained on XSTAR data to predict force multipliers based on local SED parameters (e.g., band fractions).
  • Relax the optically thin assumption to include scattering and re-emission effects.
  • Apply these methods to time-dependent scenarios (e.g., reverberation mapping).

In conclusion, this paper establishes that accounting for the local spectral distribution of radiation is critical for modeling AGN disc winds, leading to significantly stronger, faster, and more massive outflows than previously predicted by global SED models.