The large-scale ionization cones in the Galaxy

TL;DR

This paper presents evidence that a Seyfert flare from Sgr A* about 3.5 million years ago created ionization cones affecting the Magellanic Stream, explaining various ionization signatures observed today.

Contribution

It introduces a model of bipolar ionization cones from Sgr A*'s past Seyfert activity, explaining ionization features in the Magellanic Stream and timing consistent with x-ray/gamma-ray bubbles.

Findings

Ionization cones from Sgr A*'s past activity explain Stream ionization.

Timing of the flare (~3.5 Myr ago) matches x-ray/gamma-ray bubble formation.

UV absorption lines indicate high ionization consistent with a Seyfert flare.

Abstract

There is compelling evidence for a highly energetic Seyfert explosion (10^{56-57} erg) that occurred in the Galactic Centre a few million years ago. The clearest indications are the x-ray/gamma-ray "10 kpc bubbles" identified by the Rosat and Fermi satellites. In an earlier paper, we suggested another manifestation of this nuclear activity, i.e. elevated H-alpha emission along a section of the Magellanic Stream due to a burst (or flare) of ionizing radiation from Sgr A*. We now provide further evidence for a powerful flare event: UV absorption line ratios (in particular CIV/CII, SiIV/SiII) observed by the Hubble Space Telescope reveal that some Stream clouds towards both galactic poles are highly ionized by a source capable of producing ionization energies up to at least 50 eV. We show how these are clouds caught in a beam of bipolar, radiative "ionization cones" from a Seyfert nucleus associated with Sgr A*. In our model, the biconic axis is tilted by about 15 deg from the South Galactic Pole with an opening angle of roughly 60 deg. For the Stream at such large Galactic distances (D > 75 kpc), nuclear activity is a plausible explanation for all of the observed signatures: elevated H-alpha emission and H ionization fraction (X_e > 0.5), enhanced CIV/CII and SiIV/SiII ratios, and high CIV and SiIV column densities. Wind-driven "shock cones" are ruled out because the Fermi bubbles lose their momentum and energy to the Galactic corona long before reaching the Stream. The nuclear flare event must have had a radiative UV luminosity close to the Eddington limit (f_E ~ 0.1-1). Our time-dependent Seyfert flare models adequately explain the observations and indicate the Seyfert flare event took place T_o = 3.5 +/- 1 Myr ago. The timing estimates are consistent with the mechanical timescales needed to explain the x-ray/gamma-ray bubbles in leptonic jet/wind models (2-8 Myr).