A 2026 FIRE-simulation study of gas accretion and angular momentum across halo scales (Sultan et al. 2026) found that the accretion mode tracks the virialization of the inner circumgalactic medium: cold inflows dominate before virialization, while hot inflows dominate afterwards, circularizing and cooling onto galaxies that then form thin stellar disks — even in massive high-redshift halos where cold streams were predicted to prevail. The authors explicitly leave open the robustness and completeness of this picture: whether cold streams actually penetrate to galaxy radii in representative cosmological environments; how unresolved small-scale instabilities, entrainment, and fragmentation alter stream survival and specific-angular-momentum delivery; how AGN feedback, cosmic rays, and magnetic fields reshape CGM-scale cooling flows and the accretion-mode dependence of disk structure; whether the early-hot-accretion/thick-disk result holds in larger high-redshift suites; and which synthetic observables discriminate hot- versus cold-mode feeding in real galaxies.

Background

Galaxies grow by accreting gas from the intergalactic medium, and this gas must pass through the circumgalactic medium (CGM) — the diffuse, multiphase gas enveloping a galaxy out to roughly the virial radius RvirR_{\mathrm{vir}} of its dark-matter halo. Theory distinguishes two feeding channels. In cold-mode accretion, gas remains at T≲104T\lesssim 10^4 K while flowing along cosmic-web filaments; in hot-mode accretion, gas is shock-heated near the virial temperature (T∼106T\sim 10^6 K in a 1012 M⊙10^{12}\,\mathrm{M}_\odot halo) and must cool radiatively before it can fuel star formation. In the influential cold-stream picture, narrow, dense cold filaments were predicted to penetrate the hot CGM of massive (≳1012 M⊙\gtrsim 10^{12}\,\mathrm{M}_\odot) halos at high redshift — the epoch of peak cosmic star formation, z∼1z\sim 1–5, often called cosmic noon — and to directly sustain the high observed star-formation rates. Idealized high-resolution simulations of cylindrical streams show that such streams can survive interface (Kelvin–Helmholtz-type) instabilities and can even grow by entraining hot CGM gas that cools in the turbulent mixing layer and condenses onto the stream (Aung et al. 2024). The specific angular momentum (sAM, angular momentum per unit mass) carried by accreting gas is central to disk formation, since a thin, rotationally supported disk can only build up if gas arrives with coherent, aligned angular momentum. The FIRE project (Feedback In Realistic Environments) performs cosmological zoom-in simulations of individual halos embedded in the full cosmological context, with resolution reaching parsec scales in the galaxy. Sultan et al. (2026) analyzed seven FIRE halos spanning 1010.510^{10.5}–1013 M⊙10^{13}\,\mathrm{M}_\odot and z=0z=0–5, tracing inflows from ∼Rvir\sim R_{\mathrm{vir}} down to ≲0.05 Rvir\lesssim 0.05\,R_{\mathrm{vir}}. They found that the accretion mode tracks the virialization of the inner CGM (ICV) — the transition in which the inner halo fills with a quasi-static, hot, pressure-supported atmosphere: inflows are almost entirely cold before ICV, while hot inflows dominate after ICV, even in massive high-redshift halos where cold streams were predicted to dominate. Although cold inflows carry more sAM than dark matter or hot gas, in bursty low-mass galaxies (whose star formation oscillates under feedback regulation) the cold gas does not circularize — i.e., settle into rotational support — before forming stars, whereas in steady massive galaxies the hot gas circularizes in the inner CGM while cooling, in a rotating cooling flow (a slow, subsonic inflow driven by radiative losses; the idealized counterpart is analyzed in Sultan et al. 2025), and then forms stars with thin-disk kinematics. This generalizes the earlier FIRE result of Hafen et al. (2022) linking thin disks to hot-mode accretion at z∼0z\sim 0. Sultan et al. (2026) also measure that accreted gas resides in steady galaxies for up to ∼25\sim 25 galaxy free-fall times before forming stars, versus ≲5\lesssim 5 in bursty galaxies, and that high-redshift steady disks are systematically thicker and more turbulent than their low-redshift counterparts. The analysis, however, deliberately omits physics expected to matter at these masses — feedback from active galactic nuclei (AGN), cosmic rays (relativistic particles supplying an additional non-thermal pressure that can restructure the CGM: in FIRE-2 runs with cosmic-ray transport, cosmic-ray pressure dominates Milky-Way-mass halos and makes the CGM cool and volume-filling; Ji et al. 2020), and magnetic fields — rests on a small sample of four high-redshift massive halos, and identifies inflows by temperature diagnostics rather than by explicit stream reconstruction. The surrounding literature is actively divided: in NIHAO/HELLO simulations more than 75% of the gas participating in star formation at z∼2z\sim 2–4 is accreted through cold filaments even in halos well above 1012 M⊙10^{12}\,\mathrm{M}_\odot (Waterval et al. 2025), while a dedicated stream-finding analysis of TNG50 finds that cold streams typically disrupt in the outer halo and do not penetrate efficiently to the central galaxy (Medlock et al. 2026).

Problem Statement

Sultan et al. (2026) explicitly identify the following open questions about how gas accretes through the circumgalactic medium and delivers angular momentum to galaxies across the ∼1012 M⊙\sim 10^{12}\,\mathrm{M}_\odot halo-mass transition and from high redshift to z=0z=0: (1) whether narrow, long-lived cold streams from the cosmic web survive penetration through the virialized CGM and effectively feed galaxy radii in representative massive high-redshift halos — requiring quantitative identification and statistical characterization of cold streams that is robust to the stream-finding algorithm, since measured penetration depths range from roughly 0.70.7 to 0.2 Rvir0.2\,R_{\mathrm{vir}} across simulations and vary strongly from halo to halo; (2) how small-scale hydrodynamic instabilities that are unresolved in current cosmological runs, together with the entrainment of cooling hot gas and the fragmentation of cold gas, alter stream survival, morphology, and specific-angular-momentum delivery; (3) how the inclusion of additional physical processes absent from the source analysis — AGN feedback, cosmic rays, and magnetic fields — changes CGM-scale cooling flows, the angular-momentum content and circularization of inflowing gas, and the linked quenching and disk-structure outcomes for massive galaxies; (4) whether the finding that hot-mode inflows circularize and form stars with thin-disk kinematics already at early times, while high-redshift disks are systematically thicker owing to enhanced turbulence, is robust in larger, high-resolution suites of high-redshift halos; and (5) which observable diagnostics — for example absorption, emission, and kinematic tracers across wavelengths, and signatures of the differing gas residence times in bursty versus steady galaxies — can discriminate hot- versus cold-mode feeding in real galaxy samples. These five directions are independent open questions, each resolvable on its own.

References

  1. Sultan, Imran; Faucher-Giguère, Claude-André; Stern, Jonathan; Sun, Guochao (2026). 'Cold versus hot gas accretion and angular momentum in FIRE simulations: from halo to galaxy scales', Monthly Notices of the Royal Astronomical Society, 550, DOI 10.1093/mnras/stag1117, arXiv:2604.14273, https://arxiv.org/abs/2604.14273
  2. Hafen, Zachary; Stern, Jonathan; Bullock, James; Gurvich, Alex B.; Yu, Sijie; Faucher-Giguère, Claude-André; Fielding, Drummond B.; Anglés-Alcázar, Daniel; Quataert, Eliot; Wetzel, Andrew; Starkenburg, Tjitske; Boylan-Kolchin, Michael; Moreno, Jorge; Feldmann, Robert; El-Badry, Kareem; Chan, T. K.; Trapp, Cameron; Kereš, Dušan; Hopkins, Philip F. (2022). 'Hot-mode accretion and the physics of thin-disc galaxy formation', Monthly Notices of the Royal Astronomical Society, 514, 5056–5073, DOI 10.1093/mnras/stac1603, arXiv:2201.07235
  3. Sultan, Imran; Faucher-Giguère, Claude-André; Stern, Jonathan; Rotshtein, Shaked; Byrne, Lindsey; Wijers, Nastasha (2025). 'Cooling flows as a reference solution for the hot circumgalactic medium', Monthly Notices of the Royal Astronomical Society, 540, 1017–1041, DOI 10.1093/mnras/staf786, arXiv:2410.16359, https://arxiv.org/abs/2410.16359
  4. Aung, Han; Mandelker, Nir; Dekel, Avishai; Nagai, Daisuke; Semenov, Vadim; van den Bosch, Frank C. (2024). 'Entrainment of hot gas into cold streams: the origin of excessive star formation rates at cosmic noon', Monthly Notices of the Royal Astronomical Society, 532, 2965–2987, DOI 10.1093/mnras/stae1673, arXiv:2403.00912, https://arxiv.org/abs/2403.00912
  5. Yao, Zhiyuan; Mandelker, Nir; Oh, S. Peng (2026). 'Cold Stream Penetration of Virial Shocks: Fragmentation, Coagulation, and Disruption in the Hot Circumgalactic Medium', arXiv:2607.14090, https://arxiv.org/abs/2607.14090
  6. Medlock, Isabel; Nagai, Daisuke; Mandelker, Nir; Springel, Volker; van den Bosch, Frank C.; Zinger, Elad; Chiang, Barry T. (2026). 'Statistical Properties of Cold Streams in Massive Star-forming Halos in TNG50', The Astrophysical Journal, 1000, 222, DOI 10.3847/1538-4357/ae4c42, arXiv:2511.21814, https://arxiv.org/abs/2511.21814
  7. Waterval, Stefan; Cannarozzo, Carlo; Macciò, Andrea V. (2025). 'Gas accretion at high redshift: cold flows all the way', Monthly Notices of the Royal Astronomical Society, 537, 2726–2751, DOI 10.1093/mnras/staf198, arXiv:2501.19009, https://arxiv.org/abs/2501.19009
  8. Ji, Suoqing; Chan, T. K.; Hummels, Cameron B.; Hopkins, Philip F.; Stern, Jonathan; Kereš, Dušan; Quataert, Eliot; Faucher-Giguère, Claude-André; Murray, Norman (2020). 'Properties of the circumgalactic medium in cosmic ray-dominated galaxy haloes', Monthly Notices of the Royal Astronomical Society, 496, 4221–4238, DOI 10.1093/mnras/staa1849, arXiv:1909.00003, https://arxiv.org/abs/1909.00003
  9. Myrtaj, Olti; Bullock, James S.; Boylan-Kolchin, Michael; Chandra, Vedant; Faucher-Giguère, Claude-André; Feldmann, Robert; Mercado, Francisco J.; Moreno, Jorge; Stern, Jonathan; Wetzel, Andrew; Gandhi, Pratik J. (2026). 'From protogalaxy through thick and thin: Why did the Milky Way evolve in three kinematic phases?', arXiv:2605.22806, https://arxiv.org/abs/2605.22806
  10. Chen, Jianhang; Tacconi, Linda J.; Genzel, Reinhard; Neri, Roberto; Schuster, Karl; Schreiber, Natascha M. Förster; Jolly, Jean-Baptiste; Pastras, Stavros; Scaloni, Letizia; Tozzi, Giulia; Barféty, Capucine; Bolatto, Alberto; Burkert, Andreas; Combes, Françoise; Cox, Pierre; Davies, Ric; Eisenhauer, Frank; Salcedo, Juan Manuel Espejo; Herrera-Camu, Rodrigo; García-Burillo, Santi; Kodama, Tadayuki; Lee, Lilian; Lee, Minju M.; Liu, Daizhong; Lutz, Dieter; Mazzolari, Giovanni; Naab, Thorsten; Shachar, Amit Nestor; Pulsoni, Claudia; Renzini, Alvio; Rubio, Monica; Shimizu, Taro T.; Sternberg, Amiel; Sturm, Eckhard; Übler, Hannah; Usero, Antonio; Wuyts, Stijn (2026). 'NOEMA3D: extended CO, [C I] and dust in massive star-forming main sequence galaxies at cosmic noon', arXiv:2604.18504, DOI 10.1051/0004-6361/202660515, https://arxiv.org/abs/2604.18504