Cosmic Ultraviolet Baryon Survey (CUBS)

• Cosmic Ultraviolet Baryon Survey (CUBS) is an observational program using UV spectroscopy of QSOs to map diffuse baryons in the CGM and IGM at z ≲ 1.
• The survey employs HST/COS FUV and Magellan/MIKE instruments along with deep galaxy surveys to achieve precise absorber characterization and environmental mapping.
• CUBS reveals a multiphase, chemically diverse, and dynamically complex gaseous medium that informs models of baryon cycling and galaxy evolution.

The Cosmic Ultraviolet Baryon Survey (CUBS) is an observational program designed to map diffuse baryonic structures at redshift $z \lesssim 1$ using absorption-line spectroscopy of 15 UV-bright QSOs with matching deep galaxy survey data (Chen et al., 2020). Its core scientific domain is the diffuse gas in the circumgalactic medium (CGM) and intergalactic medium (IGM), especially the cold-warm and warm-hot phases that ultraviolet spectroscopy probes most directly. In the broader ultraviolet context, UV spectroscopy is uniquely sensitive to diffuse $10^4$–$10^6$ K gas at high spectral resolution and accesses transitions such as H I Ly$\alpha$, O VI, and, with sufficiently short-wavelength coverage, Ne VIII, making it central to baryon censuses, metal-enrichment studies, and environmental diagnostics in galaxies, groups, and the Cosmic Web (Burchett et al., 2019).

1. Survey architecture and observational strategy

CUBS was designed to bridge the gap between prior studies at $z<0.4$ and at $z\sim2$, and to probe how baryons cycling between galaxies and their environments regulate galaxy evolution, star formation, and feedback (Chen et al., 2020). The survey selected 15 UV-bright QSOs at $z_{\rm QSO}>0.8$ based on GALEX NUV brightness, specifically to avoid bias against sightlines containing foreground Lyman limit systems that would obscure QSOs in FUV selection. This NUV-based selection is a defining methodological choice because it preserves sensitivity to optically thick intervening absorbers (Chen et al., 2020).

The absorption-line component of the survey uses HST/COS FUV medium-resolution spectra covering approximately 1100–1800 Å with $S/N>12$ per resolution element and $\sim20$ km/s spectral resolution, supplemented by Magellan/MIKE optical echelle spectra at $\sim8$–10 km/s resolution for transitions such as Mg II, Fe II, and Ca II (Chen et al., 2020). In later CUBS analyses, these data were combined with detailed Voigt-profile decomposition, Bayesian or MCMC-based uncertainty propagation, and photoionization modeling using Cloudy to infer gas density, temperature, metallicity, abundance ratios, and non-thermal broadening on a component-by-component basis (Zahedy et al., 2021).

The galaxy-survey component is comparably deep. VLT/MUSE integral-field observations provide 1'$10^4$01' fields reaching $10^4$1 mag, with sensitivity to galaxies as faint as $10^4$2 at $10^4$3 pkpc in the initial LLS fields, while Magellan imaging and spectroscopy extend environmental mapping to scales of a few Mpc (Chen et al., 2020). This matched architecture is central to the program’s logic: the absorber is measured in the QSO spectrum, and its galactic or group environment is established through a deep, spectroscopically characterized foreground survey.

2. Lyman limit systems, partial Lyman limit systems, and dense absorbers

The first CUBS results focused on Lyman limit systems (LLSs), defined in the survey as absorbers with $10^4$4, the threshold for sufficient FUV opacity at 912 Å (Zahedy et al., 2021). CUBS I reported five new LLSs with $10^4$5 over a total redshift pathlength of $10^4$6, giving $10^4$7, and found $10^4$8 when all absorbers with $10^4$9 were included (Chen et al., 2020). All newly discovered LLSs exhibited multi-component velocity structure and associated metal transitions spanning C II, C III, Mg II, Si II, Si III, and O VI, directly demonstrating a clumpy and multiphase medium (Chen et al., 2020).

The matched galaxy surveys established that no single galaxy archetype explains the CUBS LLS population. The immediate environments of these absorbers include a pair of $10^6$0–$10^6$1 star-forming dwarf galaxies, a $10^6$2 disk galaxy with kinematics consistent with a co-rotating gaseous halo/disk, a $10^6$3 star-forming galaxy with geometry consistent with possible outflows, a massive quiescent galaxy, and a group of 10 galaxies with total group mass $10^6$4 (Chen et al., 2020). The closest galaxies lie at projected distances of 15–72 pkpc and luminosities from $10^6$5 to $10^6$6, supporting the conclusion that LLSs originate in a variety of galaxy environments and trace gaseous structures with a broad range of metallicities (Chen et al., 2020).

CUBS III extended this analysis to four optically thick LLSs at $10^6$7–0.6 and found that higher-column-density components with $10^6$8 comprise dynamically cool gas with mean temperature $10^6$9 K and modest non-thermal broadening of $\alpha$0 km/s (Zahedy et al., 2021). The inferred gas densities imply spatially compact clouds with a median line-of-sight thickness of $\alpha$1 pc, while the cool-phase gas has a median metallicity of $\alpha$2 with a 16–84 percentile range of $\alpha$3 (Zahedy et al., 2021). The same study reported wide ranges in $\alpha$4, $\alpha$5, and $\alpha$6, indicating diverse chemical enrichment histories and inefficient mixing within the low-redshift CGM (Zahedy et al., 2021).

CUBS IV showed that even lower-column absorbers, specifically partial Lyman limit systems with $\alpha$7–$\alpha$8 at $\alpha$9, contain kinematically aligned multi-phase gas that masquerades as a single-phase structure (Cooper et al., 2021). In those systems, inferred $z<0.4$0-element abundances range from $z<0.4$1 to near solar, while densities span more than two decades, from $z<0.4$2 to $z<0.4$3 (Cooper et al., 2021). One pLLS is located 55 kpc from a star-forming galaxy of stellar mass $z<0.4$4, whereas another resides in an over-dense environment of 11 galaxies within 750 kpc, illustrating that chemically and thermodynamically complex CGM absorbers arise in both isolated and group settings (Cooper et al., 2021).

A distinct dense-absorber case is the H$z<0.4$5-bearing DLA discovered in CUBS II at $z<0.4$6, with $z<0.4$7 and $z<0.4$8 (Boettcher et al., 2020). A redshift survey identified nine galaxies within 600 pkpc and $z<0.4$9 km/s, and the closest is a massive early-type galaxy at $z\sim2$0 pkpc containing approximately 70% of the total stellar mass identified at $z\sim2$1 pkpc (Boettcher et al., 2020). The absorber’s Fe enhancement, dust depletion, and proximity to the quiescent galaxy were interpreted as evidence for a physical connection, explicitly contrasting with a picture in which DLAs are primarily associated with gas-rich dwarfs (Boettcher et al., 2020).

3. Thermodynamic state, multiphase structure, and turbulence in the cool CGM

CUBS V provided a systematic thermodynamic analysis of the cool CGM around 26 galaxies at $z\sim2$2 (Qu et al., 2022). The derived hydrogen density spans more than three decades, from $z\sim2$3 to $z\sim2$4, while the temperature remains confined to a narrow range of $z\sim2$5 (Qu et al., 2022). The study found a weak anti-correlation between density and temperature,

$z\sim2$6

consistent with the expectation of photoionization equilibrium (Qu et al., 2022).

The same analysis emphasized that non-thermal broadening is common at low redshift. More than 30% of the cool-CGM components at $z\sim2$7 exhibit line widths driven by non-thermal motions, compared to $z\sim2$8 at $z\sim2$9–3 (Qu et al., 2022). Massive quenched galaxies show, on average, higher non-thermal broadening and turbulent energy in their CGM than star-forming galaxies, and in multiphase absorbers the inferred pressure between phases may differ by a factor of $z_{\rm QSO}>0.8$0 (Qu et al., 2022). This directly challenges simple static pressure-equilibrium descriptions of the cool halo gas.

CUBS later extended the kinematic analysis to a velocity–size relation for cool circumgalactic clumps. Using spectrally resolved absorption profiles and component-wise ionization analysis, the survey identified cool clumps with sizes as small as $z_{\rm QSO}>0.8$1 pc and density lower than $z_{\rm QSO}>0.8$2 (Chen et al., 2023). The analysis decomposed line widths using

$z_{\rm QSO}>0.8$3

and estimated clump sizes via

$z_{\rm QSO}>0.8$4

showing empirically that the non-thermal term scales as $z_{\rm QSO}>0.8$5 over three decades in size from $z_{\rm QSO}>0.8$6 pc to $z_{\rm QSO}>0.8$7 kpc (Chen et al., 2023). The best-fit relation is consistent with Kolmogorov at $z_{\rm QSO}>0.8$8 kpc, with a roughly constant energy transfer rate per unit mass of $z_{\rm QSO}>0.8$9 and a dissipation timescale of $S/N>12$0 Myr (Chen et al., 2023). No significant difference was found between massive quiescent and star-forming halos on scales less than 1 kpc (Chen et al., 2023).

Taken together, these results define a cool CGM that is multiphase, chemically inhomogeneous, and often dynamically non-thermal. A common misconception is that aligned low- and high-ion absorption implies a single gas phase; the partial-LLS analysis showed precisely the opposite, namely that kinematically aligned absorption can conceal multiple physically distinct phases (Cooper et al., 2021).

4. Warm-hot CGM and environmental regulation at intermediate redshift

The warm-hot component of the CUBS halo census was developed most explicitly in CUBS VII, which assembled a sample of 103 unique galaxies or galaxy groups at $S/N>12$1 for joint O VI and Ne VIII analysis (Qu et al., 2024). In this sample, 30 galaxies or galaxy groups exhibit associated O VI $S/N>12$2 absorption within $S/N>12$3 km/s, while only five show Ne VIII $S/N>12$4 absorption (Qu et al., 2024). The limiting non-detection thresholds are approximately $S/N>12$5 and $S/N>12$6, and the galaxies span stellar masses from $S/N>12$7 to $S/N>12$8 (Qu et al., 2024).

The radial and kinematic behavior of the warm-hot CGM varies systematically with galaxy type. Star-forming galaxies with $S/N>12$9–11 show a significant concentration of metal-enriched warm-hot CGM within the virial radius, whereas massive quiescent galaxies exhibit flatter radial profiles of both column densities and covering fractions (Qu et al., 2024). In addition, O VI absorption around galaxies of $\sim20$0 within the virial radius is broad, with $\sim20$1 km/s, implying a more dynamic warm-hot halo in those systems (Qu et al., 2024). The survey further suggested that the warm-hot CGM probed by O VI and Ne VIII is the dominant phase in sub-$\sim20$2 galaxies with $\sim20$3–10 (Qu et al., 2024).

CUBS VI approached the same redshift regime from the perspective of environment-selected systems at $\sim20$4. Its CUBSz1 sample contains 19 unique galaxies and galaxy groups, selected solely by projected proximity to the QSO sightline, without prior knowledge of whether absorption is present (Qu et al., 2023). Among these systems, nine show absorption features, while ten show no detectable absorption, with $\sim20$5 upper limits of $\sim20$6 and $\sim20$7 (Qu et al., 2023). Environmental properties—galaxy overdensity, total stellar mass, and the summed gravitational potential of neighbors—were found to have a significant impact on the observed CGM absorption properties (Qu et al., 2023).

A key environmental diagnostic introduced in CUBS VI is the projected gravitational potential,

$\sim20$8

where the sum is taken over all nearby galaxies (Qu et al., 2023). This quantity was found to be the best single predictor for the presence and properties of CGM absorption, and the maximum density detected along a sightline scales linearly with it,

$\sim20$9

The observed cool-gas density profile declines from the inner regions to the outskirts and is consistent with a pressure balance scenario between the cool photoionized phase and a hot ambient halo (Qu et al., 2023). In galaxy groups, the absorption properties appear to be driven by the galaxy closest to the QSO sightline, rather than by the most massive galaxy or by mass-weighted properties, and one sightline showed an unusually high $\sim8$0 ratio, suggesting a local enhancement or hardening of the ionizing field (Qu et al., 2023).

5. Dwarf galaxies, luminous quasars, and the range of CUBS environments

CUBS IX extended the survey to the low-mass end of the galaxy population through a dedicated dwarf-galaxy analysis. The CUBS-Dwarfs sample comprises 91 isolated, star-forming field dwarf galaxies at $\sim8$1 with median stellar mass $\sim8$2 (Mishra et al., 2024). The quasar spectroscopy uses 15 background quasars observed with HST/COS G130M and G160M, providing $\sim8$3 per resolution element across approximately 1100–1800 Å and enabling measurement of H I, low and intermediate ions such as C II, Si II, C III, Si III, and high-ionization O VI (Mishra et al., 2024). The survey size is $\sim8$4 larger than past dwarf CGM/IGM studies with similar ionic coverage (Mishra et al., 2024).

The principal empirical result is the contrast between rare low ions and common O VI. Low- and intermediate-ionization metal absorption is rare around dwarf galaxies, while O VI is commonly observed in sightlines passing within the virial radius, and its detection rate remains non-negligible at $\sim8$5–$\sim8$6 (Mishra et al., 2024). Within $\sim8$7, all sightlines show strong O VI absorption with $\sim8$8 and covering fraction $\sim8$9; between $10^4$00, the covering fraction is $10^4$01; and at $10^4$02, the O VI detection rate remains non-negligible (Mishra et al., 2024). The O VI-bearing phase is estimated to account for a dominant share of the metal budget of dwarf galaxies, and only $10^4$03 of the O VI-traced gas is formally unbound, implying that low-mass systems at $10^4$04 retain a substantial fraction of their metals within the nearby CGM and IGM (Mishra et al., 2024).

At the opposite end of the luminosity scale, CUBS VIII investigated the group environments of 15 luminous quasars at $10^4$05 with $10^4$06\,erg\,s$10^4$07 (Li et al., 2024). Deep galaxy redshift surveys with VLT/MUSE showed that the CUBS quasars reside in diverse environments: 11 out of 15 occur in $10^4$08 overdense regions, typically with halo masses $10^4$09, extending to $10^4$10, while the remainder inhabit moderate-size galaxy groups (Li et al., 2024). No correlation was observed between overdensity and redshift, black-hole mass, or luminosity (Li et al., 2024). Radio-loud quasars are more likely than radio-quiet quasars to reside in overdense environments, but the paper also reported radio-loud quasars in moderate groups and radio-quiet quasars in overdense environments, indicating large scatter in the environment–radio connection (Li et al., 2024).

These two studies illustrate the breadth of the CUBS program. The survey is not restricted to a single halo mass scale or a single absorber class; rather, it connects quasar absorption to field dwarfs, $10^4$11 galaxies, massive quiescent systems, galaxy groups, and the environments of luminous AGN.

6. Legacy uses of CUBS data and future observational context

CUBS has also functioned as an enabling galaxy-survey infrastructure for work outside the original absorption-line analyses. By combining CUBS spectroscopic galaxy catalogs with MeerKAT Absorption Line Survey (MALS) radio data, a later study assembled a sample of 5986 spectroscopically identified galaxies in 11 fields at $10^4$12 and detected stacked H I 21-cm emission at $10^4$13 significance in all investigated subsamples (DePalma et al., 3 Oct 2025). The mean signal corresponds to $10^4$14, with $10^4$15 for low-mass galaxies at $10^4$16 and $10^4$17 for higher-mass galaxies at $10^4$18 (DePalma et al., 3 Oct 2025). Significant 21-cm signals at distances greater than 75 kpc indicate a substantial reservoir of neutral gas in the extended surroundings of intermediate-redshift galaxies (DePalma et al., 3 Oct 2025). This suggests that the CUBS galaxy-survey component is useful not only for absorber association but also for statistical neutral-gas studies across multiple wavebands.

The broader ultraviolet framework in which CUBS operates was outlined in the white paper on diffuse gas in the largest cosmic structures. There, UV spectroscopy was described as uniquely poised to witness environmental galaxy quenching processes such as strangulation and tidal- and ram-pressure stripping, to directly account for the baryon content of galaxy clusters in cold-warm ($10^4$19 K) gas, to determine the phase structure and kinematics of gas in cluster cores, and to map cold streams and filaments of the Cosmic Web (Burchett et al., 2019). CUBS implements a galaxy-scale realization of that agenda through matched ultraviolet absorption spectroscopy and deep foreground redshift surveys.

A common limitation of present-day CGM constraints is that they come almost exclusively from pencil-beam absorption spectroscopy, leaving the spatial structure and morphology of the gas fundamentally unmapped. A recent mission concept, Ardua, explicitly positions itself as a complement to surveys such as CUBS by proposing wide-field far-ultraviolet spectroscopy together with an X-ray microcalorimeter to obtain comprehensive emission maps across the full CGM temperature range (Vargas et al., 13 Jul 2026). In that context, CUBS represents a mature absorption-line benchmark: it has established the incidence, thermodynamic state, chemical diversity, and environmental dependence of low-redshift diffuse baryons, while also defining the observational need for future spatially resolved UV/X-ray mapping (Vargas et al., 13 Jul 2026).