Revealing the link between halo mass and radio jet activities in quasars

TL;DR

This work investigates how the large-scale halo environment controls quasar radio jet power by classifying LoTSS-eBOSS quasars into four physically motivated populations that separate jet-driven emission from host-galaxy star formation. Using a two-component Bayesian model and clustering analyses via the two-point correlation function, the authors find that the correlation length and implied halo mass increase with jet fraction, from SF-dominated to hi-jet quasars, with typical halo masses rising from ~10^{12.5} to ~10^{14} h^{-1} M_\odot. They show that black hole mass and bolometric luminosity have only minor or no systematic effect on halo mass, arguing against a simple BH-mass or luminosity threshold for jet triggering. The results favor a magnetic-flux/MAD-based jet production mechanism as the primary driver linking halo environment to jet power, with implications for AGN feedback and the coevolution of galaxies and their clusters.

Abstract

There is a fundamental lack of understanding as to why quasars that are otherwise very similar can have such a wide range of radio jet powers, and the large-scale environment is thought to play an important role. We investigate the spatial clustering properties of 225,382 quasars from the extended Baryon Oscillation Spectroscopic Survey (eBOSS) within the LOFAR Two-metre Sky Survey (LoTSS) Data Release 2 footprint, split by the statistically-calculated fraction of their radio flux densities contributed by jets (relative to the contribution from star formation). We find a positive correlation between the clustering strengths of quasars and their jet fraction, where quasars with a higher jet fraction have a higher clustering amplitude measured by their two-point correlation functions. We show that this correlation is unlikely related to differences in BH masses or bolometric luminosities. Quasars dominated by powerful jet activities generally reside in dark matter haloes $10-100$ times more massive than those without strong jets, with typical halo masses of $10^{13-14}\ h^{-1}M_\odot$, establishing a robust link between powerful AGN jets and rich cluster environments. Our results demonstrate that halo mass is important for determining the power of radio jets, but suggest that there is no minimum dark matter halo mass or BH mass required for the triggering of jets. The observed correlation suggests that BH spin is likely to play a minor role in jet production; instead, the key driver could be the presence of a strong magnetic flux.

Figures (10)

• Figure 1: A comparison of the sky coverages from LoTSS DR2 (grey) and eBOSS (green) survey catalogues. The masked LoTSS DR2 field for reduced flux variation systematics is shown in blue, according to the definition in hale_cosmology_2024. The orange line encompasses the sky area studied in this paper, which is the overlap between the masked LoTSS DR2 area and the eBOSS survey area in the north galactic cap. Note that the holes in the sky area are due to the uneven spatial coverage in the eBOSS survey.
• Figure 2: Distribution of eBOSS quasars in the $\mathcal{M}_i-z$ plane. The red solid grids show the $\mathcal{M}_i-z$ grids with at least 3,000 quasars within the redshift range $0.8<z<2.2$. We use these grids to obtain best-fits of the two-component model, as well as defining the radio quasar populations in this work. The dashed orange grid outlines the binning used for clustering analysis, with smaller redshift binning and an upper limit of $z<2.2$. We combine the quasars along the $\mathcal{M}_i$ axis and compute the clustering signals in each redshift slice.
• Figure 3: Radio flux density distribution and the corresponding model best-fit in one of the representative $\mathcal{M}_i-z$ bins explored in this study. The orange dotted line and green dashed line represent the SF component from the host galaxy activity (log Gaussian) and the jet component from the AGN activity (single power-law) respectively. The pink dashed-dotted line shows the combined PDF of the two-component radio flux density distribution model, and the brown solid line shows the probability density function (PDF) after convolving with a Gaussian uncertainty characterising noise in the flux density measurements. This convolution spreads the original PDF at the faint end out to negative flux density values, leading to an offset between the convolved and original PDFs at flux densities lower than $\sim100\mathrm{\mu Jy}$. The distribution of measured radio flux densities is shown as blue crosses, indicating a good match between the actual distribution and the best-fit model result. The coloured areas highlight the different radio quasar populations defined with their dominant sources of radio emission, based on the best-fit of our model. From left to right are populations dominated by: host galaxy SF (green), mixture of SF and AGN activities (blue; radio intermediate region with no dominant source), low-power AGN jets (red), and high-power AGN jets (orange).
• Figure 4: The distribution of radio loudness, defined with $R=\log_{10}(S_\mathrm{5GHz}/f_\mathrm{4400\AA})$, across different quasar populations defined in this study. The vertical dashed-dotted line and dashed line marks two different thresholds used to define radio-loud quasars in previous studies: $R=10$ and $R=30$. The grey histograms marks the distribution of FIRST-detected quasars, another definition of the radio-loud quasar population. The optical $f_\textrm{4400\AA}$ flux density is calculated from the luminosity at 3000Å ($L_\textrm{3000\AA}$) in the SDSS DR16Q catalogue, assuming a spectral slope of $f_\lambda\propto\lambda^{0.5}$. The radio $S_\mathrm{5GHz}$ flux density is calculated from LoTSS 144 MHz or FIRST 1.4 GHz (for sources with FIRST-detection) flux densities assuming a spectral slope of $S_\nu\propto\nu^{-0.3}$.
• Figure 5: The best-fit parameters of the auto-correlation TPCF within each redshift bin, for the entire eBOSS quasar sample in our target field. The blue triangles at the top show the best-fit values for the correlation length $s_0/h^{-1}\ \mathrm{Mpc}$, while the orange crosses at the bottom show the best-fit values for the slope $\gamma$, assuming no priors in the fit. The dashed lines trace the average values for each parameter, with the uncertainties marked by the shaded areas.