SHELLQs-JWST perspective on the intrinsic mass relation between supermassive black holes and their host galaxies at z > 6

J. Silverman, Junyao Li, Xuheng Ding, Masafusa Onoue, M. Strauss, Yoshiki Matsuoka, Takuma Izumi, K. Jahnkę, Tommaso Treu, Marta Volonteri, Camryn L. Phillips, I.t Andika, Kentaro Aoki, Junya Arita, Shunsuke Baba, Sarah E. I. Bosman, Anna–Christina Eilers, Xiaohui Fan, Seiji Fujimoto, Mélanie Habouzit, Zoltán Haiman, Masatoshi Imanishi, Kohei Inayoshi, K. Iwasawa, Nobunari Kashikawa, Toshihiro Kawaguchi, Chien‐Hsiu Lee, Alessandro Lupi, Tohru Nagao, Jan–Torge Schindler, Malte Schramm, Kazuhiro Shimasaku, Yoshiki Toba, Benny Trakhtenbrot, Hideki Umehata, M. Vestergaard, Fabian Walter, Feige Wang, Jinyi YangarXiv 2025 · DOI · arXiv

TL;DR

Using the James Webb Space Telescope (JWST), researchers analyzed nine low-luminosity quasars from the early universe ($z > 6$) to see if the relationship between supermassive black holes (SMBHs) and their host galaxies has changed over 13 billion years. They found that while these early black holes appear "over-massive" for their galaxies, this is likely a result of selection bias; the underlying physical relationship is actually consistent with the local universe, albeit with significantly higher statistical "noise" or scatter.

What Problem Does This Paper Address?

In the local universe, there is a tight correlation between the mass of a supermassive black hole ($M_{BH}$) and the stellar mass of its host galaxy ($M_*$). This suggests that black holes and galaxies grow in lockstep—a process called co-evolution.

However, looking back to the first billion years of cosmic history ($z > 6$), the picture gets fuzzy. Early observations often found black holes that seemed far too large for their host galaxies. This led to a "chicken and egg" debate: Did black holes get a head start, or are we just seeing the "monsters" because our telescopes weren't sensitive enough to see the average population?

Before JWST, the intense glare of a quasar (the light from gas falling into a black hole) completely outshone the host galaxy, making it nearly impossible to measure the galaxy's mass. This paper uses JWST’s unprecedented sensitivity to finally "see" the galaxies and determine if the $M_{BH}–M_*$ relation has evolved over cosmic time.

Methods

The authors studied nine quasars from the SHELLQs (Subaru High-z Exploration of Low-luminosity Quasars) survey. They employed two primary JWST instruments:

  • NIRCam (Imaging): Used to perform "2D decomposition." By carefully modeling and subtracting the point-like light of the quasar, they could measure the faint, extended light of the host galaxy to estimate its stellar mass.
  • NIRSpec (Spectroscopy): Used to measure the widths of broad emission lines (like H$\beta$). The faster the gas moves around the black hole, the wider the line, allowing for a "virial" mass estimate of the black hole itself.

The core of their analysis is a forward modeling approach. Instead of just plotting the data points, they simulated an entire underlying population of galaxies and black holes, applied the "selection filters" of their survey (e.g., only picking objects bright enough to see), and then checked which physical models best reproduced their actual observations.

Datasets

  • SHELLQs Survey: A sample of low-to-moderate luminosity quasars at $z \sim 6.2$.
  • JWST Cycle 1 Program (GO 1967): Provided the high-resolution NIRCam and NIRSpec data.
  • Local Baselines: They compared their results against established local relations, specifically Greene et al. (2020) for general galaxies and Kormendy & Ho (2013) for massive ellipticals.
  • Simulations: Results were compared against cosmological simulations like ASTRID, TNG300, Horizon-AGN, and EAGLE.

Key Insights and Ideas

  • The "Lauer Bias" is Dominant: In a flux-limited survey, if there is a lot of scatter in a relationship, you will naturally pick up objects that have exceptionally large black holes for their galaxy size. The paper argues that the "over-massive" black holes seen at high redshift are exactly what you'd expect from a population that follows the local rule but has high variability.
  • Messy Early Growth: The intrinsic scatter in the mass relation at $z > 6$ is roughly 0.8 dex, which is much higher than the 0.3–0.4 dex seen today. This suggests that in the early universe, the "coupling" between black hole growth and galaxy growth was much less regulated.
  • The "Heavy Seed" Alternative: Some theories suggest black holes started from massive "seeds" ($10^4–10^5$ solar masses). While this paper finds the data consistent with the local relation, they note that if the relation were significantly higher (favoring heavy seeds), the scatter would have to be unrealistically low to match the observations.

Key Results

  • Mass Ranges: The black holes weighed in at $10^8–10^9 M_\odot$, while their host galaxies ranged from $10^{9.5}–10^{11} M_\odot$.
  • No Evolution in Normalization: The best-fit model shows that the intrinsic normalization of the $M_{BH}–M_*$ relation at $z > 6$ is consistent with the local relation for all galaxy types.
  • High Dispersion: The intrinsic scatter was measured at $0.80^{+0.23}_{-0.28}$ dex.
  • AGN Fraction: The model predicts that about 2.3% of galaxies at this epoch host an actively growing, unobscured supermassive black hole.

Limitations and Future Work

  • Sample Size: While nine objects is a leap forward for this redshift, it is still a small statistical sample.
  • Fixed Slope: Because the data points cover a relatively narrow range of masses, the authors had to fix the "slope" of the relationship to the local value ($\beta = 1.61$). They couldn't independently determine if the relationship tilts differently at high redshift.
  • Parameter Degeneracy: There is some mathematical overlap (degeneracy) between the normalization of the relation, the amount of scatter, and the total number of active black holes.
  • Future Directions: The authors point toward larger JWST samples and deeper spectroscopy to find the "missing" population of lower-mass black holes that their model predicts should exist in massive galaxies.

Summary updated

Abstract

The relation between the masses of supermassive black holes (SMBHs) and their host galaxies encodes information on their mode of growth, especially at the earliest epochs. The James Webb Space Telescope (JWST) has opened such investigations by detecting the host galaxies of AGN and more luminous quasars within the first billion years of the universe (z > 6). Here, we evaluate the relation between the mass of SMBHs and the total stellar mass of their host galaxies using a sample of nine quasars at 6.18 < z < 6.4 from the Subaru High-z Exploration of Low-luminosity Quasars (SHELLQs) survey with NIRCam and NIRSpec observations. We find that the observed location of these quasars in the SMBH-galaxy mass plane (log M_BH/Msun ~ 8-9; log M_*/Msun ~ 9.5-11) is consistent with a non-evolving intrinsic mass relation with dispersion (0.80_{-0.28}^{+0.23} dex) higher than the local value (~0.3-0.4 dex) of their more massive descendants. Our analysis is based on a forward model of systematics and includes a consideration of the impact of selection effects and measurement uncertainties with an assumption on the slope of the mass relation. While degeneracies between parameters persist, the best-fit solution has a reasonable AGN fraction (2.3%) of galaxies at z ~ 6 with an actively growing UV-unobscured black hole. In particular, models with a substantially higher normalisation in M_BH would require an unrealistically low intrinsic dispersion (~0.22 dex). Consequently, our results predict a large population of AGNs at lower black hole masses, as are now just starting to be discovered in focused efforts with JWST.
BibTeX
@misc{silverman2025shellqs,
  author = {J. Silverman and Junyao Li and Xuheng Ding and Masafusa Onoue and M. Strauss and Yoshiki Matsuoka and Takuma Izumi and K. Jahnkę and Tommaso Treu and Marta Volonteri and Camryn L. Phillips and I.T Andika and Kentaro Aoki and Junya Arita and Shunsuke Baba and Sarah E. I. Bosman and Anna–Christina Eilers and Xiaohui Fan and Seiji Fujimoto and Mélanie Habouzit and Zoltán Haiman and Masatoshi Imanishi and Kohei Inayoshi and K. Iwasawa and Nobunari Kashikawa and Toshihiro Kawaguchi and Chien‐Hsiu Lee and Alessandro Lupi and Tohru Nagao and Jan–Torge Schindler and Malte Schramm and Kazuhiro Shimasaku and Yoshiki Toba and Benny Trakhtenbrot and Hideki Umehata and M. Vestergaard and Fabian Walter and Feige Wang and Jinyi Yang},
  title = {{SHELLQs-JWST perspective on the intrinsic mass relation between supermassive black holes and their host galaxies at z \> 6}},
  year = {2025},
  eprint = {2507.23066},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2507.23066},
  howpublished = {arXiv},
}
Figure