Dr Harley Katz

Abstract:

We present VLT/X-shooter spectroscopy of the extremely metal-poor starburst galaxy SBS 0335‑052E, a nearby (D∼54 Mpc) analogue of high-redshift systems, probing its outflow up to a distance of ∼2.6 kpc. Using direct-method oxygen abundances, we find a complex metallicity profile that generally declines with distance, decreasing by 0.37 dex from the galaxy centre out to 2.4 kpc into the outflow. This implies a metal-loading factor roughly an order of magnitude lower than predictions based on the mass–metallicity relation for low-mass galaxies. We also detect extended He II emission, including a broad, redshifted component beyond 2 kpc, which is distinct from the narrow emission associated with star clusters. No Hβ emission associated with the broad redshifted component is detected, which implies extremely high He II /Hβ ratios (0.7 - 4.8). Such extreme values, combined with the decreasing metallicity, challenge our current models of stellar feedback. The contribution of an intermediate-mass black hole could simultaneously account for the declining metallicity and the unusually extended He II emission.

Abstract:

We present a JWST and Atacama Large Millimeter/submillimeter Array (ALMA) detailed study of the interstellar medium properties of high-redshift galaxies. Our JWST/NIRSpec integral field unit spectroscopy targeting three galaxies at z = 6-7 detects key rest-frame optical emission lines, allowing us to derive [O ii] λλ3726, 3729-based electron densities of ne,optical ∼ 1000 cm⁻³ on average and [O iii] λ4363-based metallicities of 12 + log ( O / H ) = 8.0 - 8.2 in two galaxies. New ALMA Band 9 and 10 observations detect the [O iii] 52 μm line in one galaxy but do not in the others, resulting in far-infrared (FIR)-based densities of ne,FIR ≲ 500 cm⁻³ from the [O iii] 52 μm/[O iii] 88 μm ratio, systematically lower than the optical [O ii]-based measurements. These low FIR-based densities are comparable to those at both z ∼ 0 and z > 6 in the literature, including JADES-GS-z14-0 at z = 14.18, suggesting little evolution up to z ∼ 14, in contrast to the increasing trend of optical-based densities with redshift. By conducting a JWST and ALMA joint analysis using emission lines detected with both telescopes, we find that the observed FIR [O iii] 52 and 88 μm luminosities are too high to be explained by the optical-based densities at which they would be significantly collisionally de-excited. Instead, a two-zone model with distinct high- and low-density regions is required to reproduce all observed lines, indicating that FIR [O iii] emission arises predominantly from low-density gas, while the optical [O iii] and [O ii] lines trace both regions. We further demonstrate that the direct-Te method can sometimes significantly underestimate metallicities up to 0.8 dex due to the presence of the low-density gas not fully traced by optical lines alone, highlighting the importance of combining optical and FIR lines to accurately determine gas-phase metallicities in the early Universe.

Abstract:

Cosmic ray (CR) feedback has been proposed as a powerful mechanism for driving warm gas outflows in galaxies. We use cosmological magnetohydrodynamic simulations to investigate the impact of CR feedback on neutral hydrogen (H_I) in a 10¹¹M⊙ dark matter halo at 2<z<4. To this end, we post-process the simulations with ionizing radiative transfer and perform Monte Carlo Lyman-α (Lyα) transfer calculations. CR feedback reduces H_I column densities around young stars, thereby allowing more Lyα photons to escape and consequently offering a better match to the Lyα luminosities of observed Lyα emitters. Although galaxies with CR-driven outflows have more extended H_I in the circumgalactic medium, two Lyα line properties sensitive to optical depth and gas kinematics - the location of the red peak in velocity space (v_red) and relative strength of the blue-to-red peaks (B/R) - cannot distinguish between the CR-driven and non-CR simulations. This is because Lyα photons propagate preferentially along low HI density channels created by the ionizing radiation, thereby limiting the scattering with volume-filling H_I. In contrast, the observed low flux ratios between the valley and peak and the surface brightness profiles are better reproduced in the model with CR-driven outflows because the Lyα photons interact more before escaping, rather than being destroyed by dust as is the case in the non-CR simulation. We discuss the potential cause of the paucity of sightlines in simulations that exhibit prominent red peaks and large v_red, which may require the presence of more volume-filling H_I.