{"id":"feaae797-3aa6-4f08-88fe-6dd50ff92642","arxiv_id":"1909.00003","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Cosmic ray pressure can dominate over thermal pressure in the halos of Milky Way-mass galaxies, producing cool, photo-ionized gas and matching observed O VI absorption.","lead":"This paper uses cosmological galaxy simulations to show that cosmic rays can dominate the pressure in the gas halos of Milky Way-mass galaxies, making the halo gas cooler and more diffuse. The result offers an explanation for observed oxygen and hydrogen absorption around galaxies, and suggests cosmic rays are an essential ingredient in how galaxies recycle gas.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"CR-dominated CGM rests entirely on extrapolating kappa_parallel=3e29 cm2/s from the ISM/inner CGM to 30-300 kpc; if CGM diffusivity rises with radius, the central phase/O VI results collapse.","rationale":"I agree with the reader's weakest_assumption. The paper is a well-executed simulation study, with explicit CR transport, gamma-ray calibration, resolution checks, and a clear analytic model in Section 4.1, and it honestly lists the CGM transport extrapolation as its major caveat. The central prediction of CR-dominated halos and photoionized O VI is not contradicted by any internal inconsistency; it is conditional on an assumed transport coefficient that is not directly observed in the CGM. The self-cited robustness study of Hopkins et al. (2020) does not obviously cover the rapid radial increase in kappa that would let CRs escape, and Section 5.2(i) explicitly states that possibility. A targeted kappa(r) experiment would settle whether this is a genuine failure mode or a benign extrapolation. Since the reader already rendered a conditional verdict and the identified weakness is the same one, no verdict change is needed.","tokens_in":28739,"tokens_out":4584,"duration_ms":45166,"concrete_test":"Run m12i with the same initial conditions and code but with a spatially dependent parallel diffusivity kappa_parallel(r) = 3e29 (r/10 kpc)^n cm^2/s for n = 0.5, 1, and 2, plus constant cases of 1e30 and 1e31 cm^2/s, and compare the radial CR pressure gradient to gravity at 50-200 kpc and the resulting O VI columns. If the ratio |nabla P_cr| / |rho nabla Phi| drops below unity in any of these observationally allowed variants, the headline result is not robust to CGM transport; if it remains above unity, the extrapolation is not the decisive weakness.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 3.1 and Fig. 1 present the central claim: in MW-mass halos the CR pressure gradient exceeds the thermal pressure gradient by more than an order of magnitude and balances gravity, so the CGM becomes cool, volume-filling, and photoionization-dominated with O VI columns ~1e14.5 cm^-2 at ~150 kpc. This result is a direct consequence of the CR transport model in Section 2: constant parallel diffusivity kappa_parallel = 3e29 cm^2/s and streaming at the Alfvén speed. The gamma-ray calibration constrains kappa mostly in the ISM and inner CGM (less than about 10 kpc), exactly as Section 5.2(i) states. If the diffusivity increases rapidly with radius in the diffuse, high-beta CGM, CRs escape more freely, the CR pressure gradient in Fig. 1 falls below gravity, and the phase, pressure, and O VI predictions all revert toward the MHD+ picture. The paper cites Hopkins et al. (2020) for robustness across allowed transport models, but Section 5.2(i) itself admits the rapid-escape scenario, which is precisely the regime that would break the central claim. This is not an internal inconsistency; it is an external-physics sensitivity, but it is load-bearing because every downstream conclusion depends on it.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents FIRE-2 cosmological zoom-in simulations of the circumgalactic medium (CGM) with explicit cosmic-ray (CR) transport, including supernova injection, anisotropic diffusion and streaming, and collisional and streaming losses. For Milky Way-mass halos at z≲1--2, it finds that the CR pressure gradient exceeds the thermal pressure gradient by more than an order of magnitude and nearly balances gravity, driving the CGM into a cool, volume-filling, largely photoionized phase. The paper predicts H I, O VI, and N V column densities that are consistent with low-redshift absorption-line observations, with O VI columns about 10^14.5 cm^-2 persisting to ~150 kpc, whereas the same halos in MHD-only runs are warm/hot and collisionally ionized. The interpretation is supported by an analytic equilibrium model, and the authors explicitly discuss caveats, most importantly the uncertainty in the CR diffusion coefficient.","tokens_in":29043,"tokens_out":10914,"duration_ms":96668,"significance":"If the CR-dominated state is realized in nature, the paper offers a single mechanism that simultaneously explains the observed under-pressured cool CGM, the large O VI column densities around star-forming galaxies, and the low thermal pressure of photoionized absorbers. The work uses high-resolution cosmological simulations with a comprehensive physics package (MHD, anisotropic conduction/viscosity, FIRE-2 feedback), and the ion columns are computed with the standard Trident post-processing tool. The authors show resolution tests and several Milky Way-mass halos to support generality, and the simulation code and data are publicly available. The central results make concrete falsifiable predictions: flat low/mid-ion column profiles, reduced sightline-to-sightline scatter, and anti-correlation between CR and thermal pressure at fixed radius. The main weakness is that the entire predicted state rests on a constant parallel CR diffusivity calibrated at radii ≲10 kpc and extrapolated to 30--300 kpc; the authors themselves note that a plausible rapid increase of diffusivity would allow CRs to escape and erase the CR-dominated halo.","major_comments":[{"comment":"The central claim that CR pressure balances gravity in the CGM and produces the cool, photoionized phase with O VI columns ~10^14.5 cm^-2 depends entirely on the constant parallel diffusivity κ‖=3×10^29 cm^2/s and Alfvén-speed streaming remaining valid from the ISM out to r~30--300 kpc. Section 5.2(i) states explicitly that the calibrating gamma-ray observations constrain κ only within ~10 kpc and that a plausible rapid increase of diffusivity in the CGM would let CRs escape and destroy the CR-dominated halo. Because every downstream prediction (phase structure, pressure balance, ion columns) follows from the CR-dominated state, this is a load-bearing sensitivity rather than a peripheral caveat. The citation of Hopkins et al. (2020) for robustness across transport models does not fully resolve the issue, since §5.2(i) itself identifies a plausible regime in which the effect collapses. I request a quantitative treatment: either simulations with a radially or plasma-dependent κ (or at least an analytic estimate of the radius at which CR escape becomes important), or a clear reframing in the abstract and conclusions stating that all results are predictions of a specific, observationally calibrated transport model rather than a claim that CRs necessarily dominate the CGM.","section":"§5.2(i); central argument (Figs. 1, 6--8)"}],"minor_comments":[{"comment":"As typeset, Eq. (1) places κ~29 and (1+z)^3/2 in the numerator. This appears inconsistent with the point-source steady-state solution in §4.1, where P_cr ∝ E_dot/(κ~ r) in the diffusion-dominated limit, and with the sentence immediately following the equation, which invokes the (1+z)^3/2 factor to argue that CRs decrease in importance at high redshift. Please verify the equation and correct the typesetting if needed.","section":"Eq. (1), §4.1"},{"comment":"The text refers to dwarf halos 'm10b and m11b', but Table 1 lists the halo as m10q; please reconcile the naming.","section":"§3.4"},{"comment":"The sentence 'so this produces overall less Ne viii column compared to MHD+ case, as seen in Fig. 7' should likely reference Fig. 9, which shows the ion-weighted density-temperature diagrams.","section":"§3.6"},{"comment":"The appendix says it compares 'HM12 and FG19', but the model is FG09 (Faucher-Giguère et al. 2009); please correct the label.","section":"Appendix B, first paragraph"},{"comment":"The caption reports '8002 sightlines'; this should presumably read '800^2 sightlines' as in the text. Please fix the typography.","section":"Fig. 8 caption"}],"recommendation":"major_revision","confidential_remarks":"The paper is unusually honest about its main uncertainty, and I see no internal inconsistency or circularity in the analysis. The requested revision is essentially a quantitative sensitivity study of the CR transport assumption or a more visible conditional framing of the conclusions; I do not see grounds for rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"I think this paper is worth taking seriously. It adds a new regime to the CGM story: in FIRE-2 MW-mass halos with the fiducial CR transport model, the CR pressure gradient exceeds the thermal one by more than an order of magnitude and roughly balances gravity; the CGM becomes cool, volume-filling, and photoionization-dominated, with O VI columns staying near 1e14.5 out to 150 kpc. That is a concrete mechanism for two observed puzzles: under-pressured cool gas and distant O VI. The paper is not just a claim. It shows pressure-gradient profiles, phase diagrams, pressure anti-correlation, and mock ion columns compared to survey data. It also checks resolution and shows similar CR pressure balance across several MW-mass halos in an appendix. That is real evidence.\n\nThe novelty is moderate but genuine: the CR implementation is from prior group papers, but applying it to CGM phase structure and ion columns is new, and the result is qualitatively stronger than Salem et al. 2016. The analytic model in Section 4.1 is simple, and it is a consistency check, not a fit to the ion-column predictions.\n\nThe soft spots, in proportion. The load-bearing assumption is the constant parallel diffusivity kappa = 3e29, calibrated to gamma-ray observations that mainly constrain radii under about 10 kpc. Section 5.2(i) says this openly: if diffusivity increases rapidly in the CGM, CRs escape and the CR-dominated halo disappears. This is not an internal inconsistency, since the calibration is independent of the predicted CGM columns, but it is a structural sensitivity. Every downstream result, including O VI, sits on it. A referee should press on whether any independent constraint exists at 30-300 kpc, and on whether the Hopkins et al. 2020 transport-variation study actually covers the rapid-escape regime. Second, the O VI columns from photoionized gas are sensitive to the UV background by factors of roughly 2-3 (Appendix B), so the quantitative agreement with observed O VI should be read loosely. Third, CR+ over-predicts Si IV at about 50 kpc and Ne VIII is mixed, so the observational comparison is not a clean win.\n\nThe citation pattern is heavy on the group's own prior work, but that is appropriate because the methods come from those papers, and the underlying simulations and data are public. The authors also state plainly that reality may lie between the MHD+ and CR+ runs; that is the right framing.\n\nWho this is for: CGM observers and simulators. It deserves a serious referee, and with revision focused on the transport assumptions, it could be a valuable paper. I would cite it.","headline":"A serious simulation study proposing CR pressure as the dominant support of MW-mass CGM gas, with testable O VI predictions, but the central claim rests on one extrapolated transport coefficient that the authors themselves flag.","tokens_in":29620,"tokens_out":2106,"would_cite":true,"duration_ms":20075,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"In cosmological simulations, cosmic-ray pressure, not heat, supports the gas halos of Milky Way-mass galaxies.","keywords":["cosmic rays","circumgalactic medium","galaxy halos","cosmic-ray pressure support","O VI absorption","photo-ionization","cosmological zoom-in simulations","galaxy formation feedback"],"falsifier":"Measure the cosmic-ray diffusion coefficient in the outer halo ($\\gtrsim$30 kpc from the disk) of a Milky Way-mass galaxy—for example from the radial gradient of radio synchrotron or gamma-ray emission, or from secondary-to-primary cosmic-ray ratios in halo gas. If the effective diffusivity there is even a few times the assumed constant value of $3\\times10^{29}$ cm$^2$ s$^{-1}$, cosmic rays would escape before building the pressure gradient that balances gravity, and the predicted cool, volume-filling halo would not form.","tokens_in":28583,"feed_emoji":"🌌","tokens_out":29429,"duration_ms":221952,"temperature":0.7,"pith_summary":"This paper argues that cosmic rays—not hot gas—supply the pressure that holds up the gaseous halo (the circumgalactic medium, or CGM) around Milky Way-mass galaxies, and that this reshapes what observations see in that halo. In fully cosmological galaxy-formation simulations that track the injection, streaming, diffusion, and energy losses of cosmic rays from supernovae, the authors find that at radii of roughly 30–300 kpc the cosmic-ray pressure gradient exceeds the thermal pressure gradient by more than an order of magnitude and nearly balances gravity. The halo gas settles into a cool (a few $\\times 10^4$ K), volume-filling, photo-ionized (ionized by ultraviolet background light rather than by collisions) state whose density is set by the balance between cosmic-ray pressure and gravity, independent of temperature. This matters because that state yields columns of the ion O VI (five-times-ionized oxygen) around $10^{14.5}$ cm$^{-2}$ out to $\\sim150$ kpc from star-forming galaxies—columns that models built on hot, collisionally-ionized halos have struggled to reproduce. If the claim is right, a non-thermal component governs the phase structure, ionization balance, and absorption-line signatures of galactic halos.","feed_headline":"Cosmic-ray pressure, not heat, supports galaxy halos in simulations","feed_subtitle":"Milky Way-mass halos match observed gas columns only when cosmic-ray pressure outweighs hot-gas pressure.","key_machinery":"The load-bearing element is the claim that the cosmic-ray pressure gradient balances gravity in the CGM, with a predicted equilibrium gas density $\\rho_{\\rm eq}(r) \\propto \\dot{E}_{\\rm cr}/(V_c^2 \\tilde{v}_{\\rm st} r^2)$ that is independent of gas temperature. The CR treatment injects a fixed fraction $\\epsilon_{\\rm cr} = 0.1$ of supernova kinetic energy into a GeV cosmic-ray fluid that streams along magnetic field lines at the local Alfvén speed and diffuses anisotropically with constant parallel diffusivity $\\kappa_\\parallel = 3\\times10^{29}$ cm$^2$ s$^{-1}$, chosen to match gamma-ray observations, with hadronic, Coulomb, and streaming losses included. The temperature-independence of $\\rho_{\\rm eq}$ is what lets cool and hot gas coexist at the same density at a given radius, making a volume-filling, thermally under-pressured cool phase possible. An analytic scaling, $$\\frac{|\\nabla P_{\\rm cr}|}{|\\rho\\nabla\\Phi|} \\sim 0.5\\,\\$\\alpha$\\,\\epsilon_{\\rm cr}\\,\\frac{(1+z)^{3/2}}{f_{\\rm gas,0.1}\\,\\tilde{\\kappa}_{29}}\\,\\frac{M_*}{f_b\\,M_{\\rm halo}}$$, explains why CR dominance appears only near Milky Way mass and at low redshift, matching the simulated pressure-gradient profiles.","core_discovery":"Comparing otherwise identical cosmological zoom-in simulations with and without explicit cosmic-ray (CR) transport, the paper finds a qualitative transition at halo masses of a few $\\times 10^{11}$ to $10^{12}\\,M_\\odot$ and redshifts $z \\lesssim 1$–2. In the Milky Way-mass runs, the CR pressure gradient exceeds the thermal pressure gradient by more than an order of magnitude and approximately balances gravity across the CGM ($\\approx$30–300 kpc): in the paper's words, 'the CR pressure becomes dominant over thermal (and magnetic) pressure in the CGM, and balances gravity.' In these CR-dominated halos the gas is mostly cool (a few $\\times 10^4$ K), the cool phase is volume-filling rather than confined to dense filaments, and its thermal pressure lies below the level needed for local or virial pressure balance; the cool gas is thermally under-pressured but the total (thermal plus CR) pressure at fixed radius is nearly uniform because the two pressures are locally anti-correlated. The density profile follows a temperature-independent equilibrium density $\\rho_{\\rm eq}$ at which the CR pressure gradient balances gravity, so gas at very different temperatures coexists at the same density. As a result the low and mid ions (H I, Mg II, Si IV, N V, O VI) are predominantly photo-ionized, with O VI columns $\\gtrsim 10^{14.5}$ cm$^{-2}$ out to $\\gtrsim150$ kpc, while Ne VIII columns drop; runs without CRs retain the hot, collisionally-ionized, thermally-supported halo assumed by earlier models and underproduce the observed O VI.","pith_inferences":["The temperature-independent equilibrium density at the heart of the argument is not specific to cosmic rays: any pressure component with a steep radial gradient and weak coupling to gas temperature—magnetic fields, Alfvén-wave pressure, or turbulent pressure—would, if sustained at comparable amplitude, produce the same volume-filling cool phase; cosmic rays are the one candidate the simulations ca","A diagnostic the paper leaves implicit: in the CR-supported halo, the O VI-bearing gas at 100–200 kpc should have thermal pressures below roughly $10^{-3}$ eV cm$^{-3}$ and line widths dominated by bulk motions, so density- or ionization-ratio-sensitive observations of the same absorbers could test the model without any direct cosmic-ray measurement.","Because the CR-to-gravity force ratio in the analytic scaling is inversely proportional to the diffusion coefficient, the mass–redshift window where cosmic rays dominate is the most fragile prediction: if the true diffusivity grows with decreasing density, as microphysical transport models suggest, the window shrinks or vanishes, making scale-dependent transport models the natural next test.","The model also implies that the hot $\\sim10^6$ K phase is strongly suppressed at 50–200 kpc in Milky Way-mass halos; X-ray absorption measurements such as O VII around these galaxies could check this prediction independently of the O VI comparison."],"forward_implications":["O VI columns of $\\gtrsim 10^{14.5}$ cm$^{-2}$ extending to $\\gtrsim150$ kpc around low-redshift star-forming galaxies are produced by photo-ionized cool gas supported by cosmic-ray pressure, not by collisionally-ionized warm gas, resolving the energy-budget problem that collisional models of the observed O VI have faced.","The cool CGM in Milky Way-mass halos is volume-filling and thermally under-pressured, matching the state inferred from observations of low-ion absorbers, and the sightline-to-sightline scatter in low- and mid-ion columns is much smaller than in runs without cosmic rays.","Cosmic-ray dominance is confined to halos near $10^{12}\\,M_\\odot$ at $z \\lesssim 1$–2: dwarf halos and high-redshift Milky Way-mass progenitors show essentially no cosmic-ray effects, so the model predicts a sharp dependence of CGM phase structure and ion columns on halo mass and redshift.","Galaxies whose star formation is quenched would lack the cosmic-ray injection that sustains the cool, volume-filling halo, giving a natural explanation for the observed deficit of O VI around passive galaxies relative to star-forming ones.","The CGM density profile is set by the cosmic-ray pressure gradient rather than by virial temperature, so halo gas can be supported at a few $\\times 10^4$ K while remaining in global force balance."],"supporting_citations":[{"why":"Supplies the cosmic-ray transport implementation and the gamma-ray observations that calibrate the parallel diffusivity of 3 x 10^29 cm^2/s in and around the galaxy.","marker":"Chan et al. 2019"},{"why":"Presents the simulation suite analyzed here and the analytic scaling predicting when cosmic-ray pressure balances gravity as a function of halo mass and redshift.","marker":"Hopkins et al. 2019"},{"why":"Earlier cosmological simulations finding cosmic rays support cool CGM gas; the comparison case this paper extends to full cosmic-ray transport and far stronger CR dominance.","marker":"Salem et al. 2016"},{"why":"Observations showing cool CGM gas is thermally under-pressured; the observational puzzle the cosmic-ray-supported halo is invoked to explain.","marker":"Werk et al. 2014"},{"why":"The observed O VI columns of about 10^14.5 cm^-2 extending to about 150 kpc around star-forming galaxies that the CR+ runs reproduce.","marker":"Tumlinson et al. 2011"},{"why":"Observed O VI and N V/O VI ratios whose path-length argument is resolved by the diffuse, volume-filling cosmic-ray-supported gas.","marker":"Werk et al. 2016"},{"why":"Provides the photo-ionization equilibrium scaling and O VI ionization fractions used to identify the photo-ionized origin of the predicted columns.","marker":"Stern et al. 2018"},{"why":"The low-redshift UV background model used in the photo-ionization calculations, whose high-energy shape shifts the quantitative O VI and Ne VIII columns.","marker":"Faucher-Giguere et al. 2009"},{"why":"The ionization post-processing tool used to compute the ion number densities and column densities from the simulated gas.","marker":"Hummels et al. 2017"}],"fun_headline_variants":["Cosmic rays, not heat, support Milky Way-mass halos","Cosmic-ray pressure dominates in Milky Way-mass halos","Cosmic rays make cool, volume-filling halos","Cool CGM in Milky Way-mass halos from cosmic rays","Cosmic-ray pressure balances gravity in galaxy halos"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entire CR-dominated halo rests on assuming that the cosmic-ray diffusion coefficient inferred within roughly 10 kpc of the disk stays constant all the way out to 300 kpc; if cosmic rays instead escape far more freely through the dilute outer halo, the predicted pressure support—and with it the cool, volume-filling, photo-ionized CGM—would not form.","fun_headline_variants_meta":{"raw":{"variants":["Cosmic rays, not heat, support Milky Way-mass halos","Cosmic-ray pressure dominates in Milky Way-mass halos","Cosmic rays make cool, volume-filling halos","Cool CGM in Milky Way-mass halos from cosmic rays","Cosmic-ray pressure balances gravity in galaxy halos"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00075,"raw_usage":{"total_tokens":3482,"prompt_tokens":1234,"completion_tokens":2248,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":850,"completion_tokens_details":{"reasoning_tokens":2165}},"tokens_in":850,"tokens_out":2248,"duration_ms":15239,"temperature":1.0,"reasoning_tokens":2165,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:05:10.811734+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the cosmic-ray diffusion coefficient in the outer halo ($\\gtrsim$30 kpc from the disk) of a Milky Way-mass galaxy—for example from the radial gradient of radio synchrotron or gamma-ray emission, or from secondary-to-primary cosmic-ray ratios in halo gas. If the effective diffusivity there is even a few times the assumed constant value of $3\\times10^{29}$ cm$^2$ s$^{-1}$, cosmic rays would escape before building the pressure gradient that balances gravity, and the predicted cool, volume-filling halo would not form.","supporting_citations":[{"cited_title":"L., Corlies L., 2016, Monthly Notices of the Royal Astronomical Society, 456, 582","cited_arxiv_id":null,"evidence_quote":"Earlier cosmological simulations finding cosmic rays support cool CGM gas; the comparison case this paper extends to full cosmic-ray transport and far stronger CR dominance."},{"cited_title":"K., et al., 2014, The Astrophysical Journal, 792, 8","cited_arxiv_id":null,"evidence_quote":"Observations showing cool CGM gas is thermally under-pressured; the observational puzzle the cosmic-ray-supported halo is invoked to explain."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The observed O VI columns of about 10^14.5 cm^-2 extending to about 150 kpc around star-forming galaxies that the CR+ runs reproduce."},{"cited_title":"K., et al., 2016, The Astrophysical Journal, 833, 54","cited_arxiv_id":null,"evidence_quote":"Observed O VI and N V/O VI ratios whose path-length argument is resolved by the diffuse, volume-filling cosmic-ray-supported gas."},{"cited_title":"F., Hafen Z., Johnson S","cited_arxiv_id":null,"evidence_quote":"Provides the photo-ionization equilibrium scaling and O VI ionization fractions used to identify the photo-ionized origin of the predicted columns."}],"review_version":1}