Optical and Ultraviolet Emission from the Gas Photoionized by Gaia BH3: Archival Constraints and Predictions for the 2029 Periastron

Abstract

Gaia BH3, a $32.7\,M_\odot$ black hole (BH) on an 11.6-yr, $e=0.73$ orbit around a metal-poor giant at $\approx590$ pc, accretes from its wind. Neither its mass-loss rate $\dot M_{\rm w}$ nor the fraction of captured gas reaching the BH is known. We ask whether optical and ultraviolet lines from gas photoionized by the hot flow are observable and what they reveal. Unlike X-ray and radio data, they count the ionizing photons the gas absorbs and trace its density and velocity. With Cloudy we compute the H$α$ and He II 4686 emission of the wind and inflow along the orbit for $\dot M_{\rm w}=10^{-11}$-$10^{-8}\,M_\odot\,{\rm yr^{-1}}$, Eddington ratios $f_{\rm Edd}=10^{-8}$-$10^{-5}$ and wind speeds of 30 (slow) and 100 km s$^{-1}$ (fast), and test each model against archival UVES and ESPRESSO spectra. $f_{\rm Edd}$ mainly sets the lines; in dense models they come from gas falling freely onto the BH, spread over thousands of km s$^{-1}$. The archival spectra exclude $\dot M_{\rm w}>2\times10^{-9}\,M_\odot\,{\rm yr^{-1}}$ at $f_{\rm Edd}=2\times10^{-6}$ and $>2.6\times10^{-10}\,M_\odot\,{\rm yr^{-1}}$ at $10^{-5}$ (95%, slow wind), Eddington ratios already excluded by the apastron Chandra limit, $f_{\rm Edd}<7.6\times10^{-7}$. Of two simulated campaigns, only 25 epochs around the 2029 periastron go beyond existing limits. For a slow wind, whose capture rate is constant, a $\ge5σ$ H$α$ detection would imply an accretion state 2-3 times above the 2024 limit. For a fast wind, whose captured fraction rises 14-fold toward periastron, the campaign detects $\dot M_{\rm w}\simeq10^{-8}\,M_\odot\,{\rm yr^{-1}}$ for $f_{\rm Edd}\gtrsim1.2\times10^{-6}$, which the apastron limit allows. Below the Chandra limit, a slow wind produces no detectable line.