Fast Radio Bursts Reveal Galaxy Evolution ✛ ✦
Context
Ionized gas in galaxies and their halos — from dense HII regions to the diffuse circumgalactic medium — reflects the entire suite of processes governing galaxy growth and transformation. Yet observations in emission lines, quasar absorption spectra, and X-ray data are sensitive only to individual phases, leaving a unified picture elusive. Fast radio bursts (FRBs) — millisecond-duration extragalactic radio pulses — offer a unique way to measure the total electron density along the line of sight via dispersion measure (DM), without assumptions about temperature, ionization, or background sources. Studying the host DM versus redshift allows us, for the first time, to directly trace the co-evolution of density and sizes of ionized structures across the entire galactic evolution. This idea traces back to early work in radio astronomy pioneered by Karl Jansky, and builds on Hubble’s law as the foundation linking redshift to distance.
Methods
To measure the host DM evolution, we employed forward modeling with the zDM package on a sample of 47 FRBs from the Deep Synoptic Array (DSA) and 43 FRBs from the ASKAP/CRAFT ICS survey — 90 events in total, 69 of which have confirmed host redshift. The model accounts for instrumental detection effects for each telescope and models the host DM as a lognormal distribution with a mean evolving as DM_host(z) = DM_host,0 (1+z)^{$n_z$}, where $n_z$ is the evolution index of interest. Using Bayesian MCMC analysis with discrete prior grids for $n_z$, the mean DM_host, the Hubble constant H₀, and the feedback parameter F, we derived posterior distributions for individual and joint samples. To assess the impact of numerical simulations, we also incorporated data from IllustrisTNG and FIRE-2, with statistical significance tracked via likelihood analysis.
Results
The joint analysis of DSA and CRAFT yields $n_z$ = $1.62^{+1.48}_{-1.57}$, ruling out no evolution ($n_z$ = 0) at the 1σ level. Both datasets individually also favor $n_z$ > 0, albeit with larger uncertainties. We find a strong anti-correlation between $n_z$ and the mean host DM: accounting for evolution shifts the log mean DM from 2.18 (no evolution) to ~2.03, corresponding to a ~30% change in the host contribution. This directly impacts the estimated contribution of cosmic plasma and thus the inferred Hubble constant, baryon distribution, and feedback parameters in the interstellar medium and halos. High-z FRBs prove to be the most sensitive indicators: differences between $n_z$ values become noticeable only beyond z~1. Simulations show that reducing the $n_z$ error to ~0.7 will require about 100 localized FRBs from MeerTRAP in coherent mode, or 300–350 from DSA/CRAFT.
Implications
The measured host DM evolution provides, for the first time, direct observational evidence for the growth of integrated ionized gas density in galaxies over cosmic time. Assuming the bulk FRB population traces star formation regions and that galaxy sizes scale as (1+z)^{-1}, the inferred index $n_z$ ≈ 1.6 corresponds to an electron density $n_e$ ∝ (1+z)^{2.6}, closely matching the cosmic star formation history. This points to a dominant role for diffuse ionized gas (DIG) and suggests that star formation sets the overall evolutionary pace. For magnetars as potential FRB progenitors, such a trend aligns with young stellar populations. Moreover, host DM evolution is crucial for correctly subtracting the host contribution in cosmological studies: ignoring $n_z$ overestimates redshifts by Δz ≈ 0.3 for DM_EG ~ 1000–2000 pc cm⁻³. The result also challenges the idea of a large-DM FRB population at low z — the observed excess DM in distant FRBs is now naturally explained by evolution.
Future development
Further accumulation of data with high-precision localizations (CHIME/FRB Outriggers, CHORD, DSA-2000) and massive spectroscopic surveys (DESI, EUCLID) will not only refine $n_z$ but also disentangle contributions from individual gas phases, galaxy sizes, and morphologies. Comparing DM evolution in star-forming and quiescent galaxies will shed light on FRB progenitor channels. Joint analysis with next-generation cosmological simulations (e.g., FIRE-2) will help calibrate subgrid models of feedback and ionization. Potentially, combining FRB data with emission-line observations and X-ray surveys could build a three-dimensional picture of the baryon cycle’s evolution across the galaxy-halo scale.
Impact
The results will directly impact the precision of cosmological estimates using FRBs (Hubble constant, baryon density, helium reionization epoch), as well as the interpretation of data on the intergalactic medium and feedback in galaxies.
Next steps
The immediate next step is to include n_z as a free parameter in standard FRB analysis chains to jointly constrain cosmological and astrophysical parameters. In parallel, cataloging the properties of FRB host galaxies is needed to search for DM correlations with type, mass, and star formation rate.
Key open problems
The evolution of host DM is directly linked to the missing baryon problem: ionized gas in galaxy halos (warm and hot CGM phases) is still poorly accounted for in the overall budget, and FRBs offer a unique integrated way to measure it. Moreover, the DM-redshift dependence is sensitive to feedback processes, which remain a key uncertainty in galaxy formation models.
The knowledge machine's view
This section is addressed to the author of the paper and their colleagues. Neighbouring works were found by meaning-based search across our archive, not by word overlap. We do not review: we show the field around the work.
What the work does
This study examines the evolution of the dispersion measure (DM) of fast radio bursts (FRBs) with redshift as a global probe of ionized gas in galaxies and their halos. Using a sample of 90 FRBs (DSA and ASKAP/CRAFT) and direct modeling that accounts for instrumental effects, the authors derived an evolution index of n_z = 1.62 (+1.48/-1.57), ruling out no evolution at the 1σ level, and showed the impact of this evolution on estimates of the host's DM contribution.
Where its strength is
The strength of this work lies in its novel use of FRBs as a direct probe of the total electron column, independent of gas phase, and its demonstration of statistically significant evolution in host DM, which is critical for cosmological applications.
Where it could go next
- It might be interesting to compare the derived evolution index with predictions from IllustrisTNG numerical simulations, as done in a recent review on hidden baryons, to separate contributions from different gas phases.based on: Cosmic Lightning: How Fast Radio Bursts Illuminate Hidden Baryons arXiv:2607.04106v1
- A related work uses FRBs to estimate the Hubble constant; comparing systematics associated with host DM evolution could help refine cosmological conclusions.based on: In the Rhythm of the Universe: Radio Bursts vs. Hubble Tension arXiv:2606.22390
- There is a review in the archive on the impact of quasar feedback on gas in clusters; it may be worth considering how feedback from active galactic nuclei modulates host DM evolution at high redshifts.based on: Breath of a Quasar: How a Black Hole Controls a Cluster arXiv:2607.24911v1
Worth trying
- Test how including host DM evolution changes constraints on the feedback parameter F when combined with cosmic baryon content data.
- Use future FRB samples with a larger number of confirmed redshifts to separately measure host DM evolution for spiral and elliptical galaxies.
- Compare the derived index n_z with the evolution of the star formation rate density to identify the dominant source of ionized gas.
Nearby works in our archive
- Cosmic Lightning: How Fast Radio Bursts Illuminate Hidden Baryonssimilarity 0.66
- How Young Galaxies Ignite Gassimilarity 0.64
- In the Rhythm of the Universe: Radio Bursts vs. Hubble Tensionsimilarity 0.64
- Radio bursts help measure the expansion of the Universesimilarity 0.63
- Breath of a Quasar: How a Black Hole Controls a Clustersimilarity 0.63
- Radio Bursts to Map the Invisible Universesimilarity 0.63
The closest work in the archive sits at 0.66 by meaning — the usual level for our corpus.
Original abstract
as written by the authors arXiv:2606.25214
The redshift evolution of ionized gas in the full galaxy-halo system is a central open question in galaxy formation, because no existing observable is simultaneously sensitive to all ionized phases. Here we explore fast radio bursts (FRBs) as a probe of the density evolution of this gas through the redshift dependence of the FRB host dispersion measure, $${\rm DM_{host}}(z) \propto (1+z)^{n_z}$$. The host DM denotes the electron column density of all ionized gas in the host along the FRB sightline, providing a unified tracer that complements existing phase-specific diagnostics. We apply a forward-modeling framework that accounts for instrumental effects to 90 localized FRBs (69 with confirmed host redshifts) from the DSA and ASKAP/CRAFT ICS surveys. Our inference yields $$n_z = 1.62^{+1.48}_{-1.57}$$, ruling out the non-evolving scenario ($$n_z = 0$$) at $$1\,\sigma$$, with both datasets independently favoring $$n_z > 0$$. The main $$n_z$$ degeneracy is with the mean host DM and parameters such as $$H_0$$, highlighting the need to account for a host evolution in inference analyses and DM-based host redshift estimates; overestimating redshifts by up to $$\Delta z \approx 0.3$$ for DM$$_{\rm EG} \sim 1000 - 2000\,{\rm pc\,cm^{-3}}$$ otherwise. About 100 ($$300-350$$) localized hosts from MeerTRAP in coherent mode (DSA/CRAFT) will yield $$n_z$$ uncertainties of $$\sim0.7$$. Precise $$n_z$$ measurements compared with the evolution of individual phases and galaxy scaling relations will shed light on ionized gas evolution in galaxies and halos, informing the dominant phase, the driver of the overall evolution, and FRB progenitor channels.
The dispersion measure of a fast radio burst is so precise that it can 'weigh' every free electron along a line of sight stretching billions of light-years. If you could gather those electrons in one place, their total charge would make your hair stand on end even from several kilometers away!
Key numbers
- Evolution index n_z: 1.62 (+1.48/–1.57)
- Number of FRBs with confirmed redshift: 69 out of 90
- Required number of FRBs for ~0.7 error: ~100 (MeerTRAP) or ~300–350 (DSA/CRAFT)
- Shift in mean log host DM when accounting for evolution: from 2.18 to 2.03 (~30% in DM)
- Maximum FRB redshift in the sample (MeerTRAP): 2.15