Galaxy mergers in the epoch of reionization – II. Merger-triggered star formation and AGN activities at z = 4.5−8.5

Qiao Duan, Christopher J. Conselice, T. Harvey, Qiong Li, D. Austin, Nathan Adams, Leonardo Ferreira, K. J. Duncan, A. James Trussler, Robert G. Pascalau, Rogier A. Windhorst, Benne W Holwerda, Tom Broadhurst, D. Coe, S. H. Cohen, Xiaojing Du, Simon P Driver, B L Frye, Norman A Grogin, Nimish P Hathi, Rolf A Jansen, Anton M Koekemoer, Madeline A. Marshall, M. Nonino, Rafael Ortiz, Nor Pirzkal, Aaron Robotham, Zoltan G. Levay, Jake Summers, J. C. J. D'Silva, Christopher N A Willmer, Haojing YanMonthly 2026 · DOI

TL;DR

This study uses JWST data to investigate the impact of galaxy mergers on star formation and AGN activity at very high redshifts ($z=4.5-8.5$). They find that close galaxy pairs ($r_p < 20$ kpc) show a modest enhancement in star formation, but no significant increase in sSFR or stellar mass. Interestingly, they also find a notable excess of AGNs in these close pairs, suggesting mergers play a role in fueling active galactic nuclei even in the early universe.

What Problem Does This Paper Address?

Galaxy mergers are thought to be crucial drivers of galaxy evolution, triggering both intense star formation and the growth of supermassive black holes that power active galactic nuclei (AGNs). While this has been observed at lower redshifts, understanding these processes at the very early universe, specifically during the Epoch of Reionization ($z \approx 4.5-8.5$), is challenging due to limited observational data. This paper aims to fill this gap by using the unprecedented capabilities of the James Webb Space Telescope (JWST) to investigate how galaxy mergers influence star formation and AGN activity at these high redshifts. Specifically, they want to quantify the extent of star formation enhancement and the fraction of AGNs in merging or closely interacting galaxies compared to isolated ones.

Methods

The researchers employed a "close-pair" method to identify potential mergers. This involves selecting galaxies that are close to each other in both projected physical separation ($r_p < 100$ kpc) and redshift difference ($\Delta z < 0.3$). For major mergers, an additional criterion of a mass ratio greater than 1:4 was used. They then compared the star formation properties (SFR, sSFR) and stellar masses of these close pairs to a control sample of non-merging galaxies. To identify AGNs, they used two methods: photometric spectral energy distribution (SED) fitting and spectroscopic BPT diagnostics. They then calculated AGN fractions and an "AGN excess" factor to quantify the over-representation of AGNs in close pairs compared to isolated galaxies. They also compared their observational results to predictions from the SC-SAM simulation.

Datasets

The study utilizes data from eight JWST fields: CEERS, JADES GOODS-S, NEP-TDF, NGDEEP, GLASS, El-Gordo, SMACS-0723, and MACS-0416, covering a total unmasked area of 189 arcmin². This dataset includes deep JWST NIRCam imaging and, for some fields, spectroscopic data from NIRSpec. The sample consists of 3,330 galaxies at redshifts $z = [4.5, 8.5]$.

Key Insights and Ideas

  • Mergers trigger star formation, but only up close: The most significant finding is that star formation is enhanced in galaxy pairs, but only when they are extremely close ($r_p < 20$ kpc). This suggests that the physical processes driving starbursts, like gas inflow, are highly localized.
  • No boost for sSFR or stellar mass: Despite the SFR enhancement, the specific star formation rate (sSFR) and stellar mass show no statistically significant increase in close pairs. This implies that while mergers might temporarily boost star formation, they don't necessarily lead to a sustained increase in the galaxy's overall star formation efficiency or its mass accumulation over these timescales.
  • AGNs love company: A key takeaway is the significant "AGN excess" in close-pair galaxies. This means AGNs are much more likely to be found in pairs than would be expected by chance, suggesting that mergers are a significant factor in fueling these active galactic nuclei even at these early cosmic epochs.
  • Simulations lag behind observations: The SC-SAM simulation, used for comparison, fails to reproduce the observed star formation enhancement in close pairs, suggesting that current models might underestimate the impact of mergers on star formation processes.

Key Results

  • SFR Enhancement: A statistically significant SFR enhancement of $0.25 \pm 0.10$ dex (factor of ~1.77) and $0.26 \pm 0.11$ dex (factor of ~1.76) was observed for close pairs ($r_p < 20$ kpc) at $z = [4.5, 6.5]$ and $z = [6.5, 8.5]$, respectively. No enhancement was seen at larger separations ($r_p \ge 20$ kpc).
  • sSFR and Stellar Mass: No statistically significant enhancements in sSFR or stellar mass were found for close pairs at any projected separation or redshift bin.
  • AGN Fraction: The fraction of AGNs found in pairs is significantly higher than the fraction of galaxy pairs in the general population. Nearly all AGNs were found to have a companion within 100 kpc.
  • AGN Excess: An AGN excess factor of $1.26 \pm 0.06$ (BPT method) and $1.34 \pm 0.06$ (SED method) was measured for close pairs relative to isolated galaxies, indicating a notable over-representation of AGNs in interacting systems.
  • Simulation Discrepancy: The SC-SAM simulation showed no significant SFR enhancement in close pairs, highlighting a potential gap in current theoretical models.

Limitations and Future Work

The authors acknowledge several limitations:

  • Photometric Redshifts: The reliance on photometric redshifts means that the $\Delta z < 0.3$ criterion might include some galaxies that are not physically close. Future work with spectroscopic redshifts would provide more precise measurements.
  • Limited Sample Size: While this is one of the largest samples for this redshift range, the sample size is still limited, particularly for statistically robust comparisons across different merger stages or mass ratios. This also hinders the ability to observe trends in AGN fraction with decreasing separation.
  • Photometric SED Fitting: While generally reliable, photometric SED fitting for stellar mass and SFR has inherent uncertainties. The correlation with spectroscopic methods is moderate, and more spectroscopic data would allow for finer calibration.
  • Morphological Selection Bias: The comparison with morphologically selected mergers from another study revealed a discrepancy, which could be due to uncertainties in morphological classification at high redshifts or the late-stage nature of those selected mergers.

Future work could involve:

  • Spectroscopic Follow-up: Obtaining spectroscopic redshifts and detailed line measurements would significantly improve the accuracy of redshift and star formation rate estimates.
  • Larger Surveys: Future, larger JWST surveys or missions like Euclid and Roman will provide larger samples, enabling more detailed studies of massive galaxies and finer separation trends.
  • Broad-Line AGN Identification: Identifying broad-line AGNs at these redshifts would provide a more definitive sample for studying AGN triggering.
  • Improved Simulations: Refining semi-analytic models to better capture the impact of mergers on gas inflow and star formation is crucial for theoretical understanding.

Summary updated

Abstract

ABSTRACT Galaxy mergers are key drivers of galaxy formation and evolution, including triggering active galactic nucleus (AGN) and star formation to a still unknown degree. We thus investigate the impact of galaxy mergers on star formation and AGN activity using 3330 galaxies at $z = 4.5 {\text{-}} 8.5$ from eight James Webb Space Telescope fields (CEERS, JADES GOODS-S, NEP-TDF, NGDEEP, GLASS, El-Gordo, SMACS-0723, and MACS-0416), covering an unmasked area of 189 arcmin$^2$. We focus on star formation rate (SFR) enhancement, AGN fraction, and AGN excess in close pairs defined by $\Delta z &amp;lt; 0.3$ and projected separations $r_p &amp;lt; 100$ kpc, relative to non-merger samples. Close pairs with mass ratios greater than 1:4 are used for the SFR–enhancement analysis, whereas no mass–ratio constraint is applied for the AGN fraction and AGN excess measurements. We find SFR enhancement occurs only at $r_p &amp;lt; 20$ kpc, with values $0.25 \pm 0.10$ dex and $0.26 \pm 0.11$ dex above non-merger medians for $z = [4.5, 6.5]$ and $z = [6.5, 8.5]$, respectively. No other statistically significant enhancements in galaxy sSFR or stellar mass are observed at any projected separation or redshift. We also compare observational results with predictions from the SC-SAM simulation, finding no evidence of star formation enhancement in simulations at any separation. Lastly, we examine the fraction and excess of AGNs identified through photometric spectral energy distributions (SED) fitting (Type-I) and BPT diagnostics (Type-II). We find $48^{+33}_{-7}\%$ of Type-I and $44^{+26}_{-20}\%$ of Type-II AGNs have a close companion within $r_p &amp;lt; 50$ kpc and $\Delta z &amp;lt; 0.3$. Furthermore, $73^{+27}_{-32}\%$ of AGNs have companions within 100 kpc. Relative to isolated galaxies, we measure an AGN excess factor of $1.26^{+0.15}_{-0.04}$ for Type-I and $1.34^{+0.23}_{-0.11}$ for Type-II AGNs in close pairs, suggesting notable AGN enhancement in galaxy pairs at these higher redshifts.
BibTeX
@inproceedings{duan2026galaxy,
  author = {Qiao Duan and Christopher J. Conselice and T. Harvey and Qiong Li and D. Austin and Nathan Adams and Leonardo Ferreira and K. J. Duncan and A. James Trussler and Robert G. Pascalau and Rogier A. Windhorst and Benne W Holwerda and Tom Broadhurst and D. Coe and S. H. Cohen and Xiaojing Du and Simon P Driver and B L Frye and Norman A Grogin and Nimish P Hathi and Rolf A Jansen and Anton M Koekemoer and Madeline A. Marshall and M. Nonino and Rafael Ortiz and Nor Pirzkal and Aaron Robotham and Zoltan G. Levay and Jake Summers and J. C. J. D'Silva and Christopher N A Willmer and Haojing Yan},
  title = {{Galaxy mergers in the epoch of reionization – II. Merger-triggered star formation and AGN activities at <i>z</i> = 4.5−8.5}},
  booktitle = {Monthly},
  year = {2026},
  doi = {10.1093/mnras/stag008},
}
Figure