{"id":"12065eaf-24da-4e39-87eb-41d1942a62e7","arxiv_id":"2504.21207","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Global 3D simulations with separate electron and proton temperatures show that the inner region of a thin accretion disk becomes a hot two-temperature flow, starting at a radius that grows as the accretion rate drops.","lead":"Simulations of the hot gas falling into a spinning black hole show that the thin, cool disk of gas can turn into a thick, hot flow close to the black hole, depending on how fast the disk feeds the hole. The result matters for interpreting X-ray observations of black holes, including estimates of black hole spin and the origin of the X-ray corona.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 10^10 K cooling target and Ω^-1 cooling time set the density that controls Coulomb coupling, so the quantitative truncation radii in Fig. 11 (e.g.","rationale":"The reader's weakest_assumption identifies the same load-bearing issue: the electron cooling law is imposed rather than derived, with a target temperature ten times the physical Compton temperature and a cooling time set to the orbital time rather than the physical cooling time. This prescription controls the disc density and scale height, which in turn set the Coulomb coupling efficiency that drives the reported truncation. I find no reason to move away from the reader's CONDITIONAL verdict: the simulations are internally consistent, the implicit Coulomb solver is tested in Appendix A, the timescale ordering in Sec. 3.1.1 gives a clear mechanism for two-temperature decoupling inside the ISCO, and the trend across four accretion rates is qualitatively plausible. But the quantitative transition radii, and especially the claim that mdot~1e-2 produces an inner truncated disc at 4–5 r_g, depend on an artificial cooling target. The proposed sensitivity run — reducing the target to 10^9 K and shortening the cooling time — would directly test whether the r>r_ISCO truncation survives under more physical cooling. No separate concern about fraud, reproducibility, or internal inconsistency is warranted; the paper itself flags the target-temperature dependence in Secs. 4.1 and 5, and the current conditional framing is appropriate.","tokens_in":19763,"tokens_out":7116,"duration_ms":76886,"concrete_test":"Repeat the M2 run (mdot≈1e-2) with the same setup but replace Eq. (2) with T_e,target=10^9 K and t_cool set to the local optically-thin Compton cooling time (or, as a controlled sensitivity step, 0.1Ω^-1), using angular resolution adequate for H/r≈0.02 or a target-temperature profile that keeps H/r resolved. Measure r_tr from Eq. (16) over the final 2000 r_g/c and compare with the published 3.6 r_g. If r_tr shifts inward by more than ~1 r_g or falls to r_ISCO, the Fig. 11 intermediate-rate truncation radii are artifacts of the imposed cooling law, and only the two-temperature-plunging claim stands; if r_tr remains above r_ISCO, the concern is mitigated.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The quantitative part of the central claim — that the truncation radius grows from ~2 r_g at mdot~1 to ~4–5 r_g at mdot~1e-2 (Fig. 11) — rests on the imposed electron cooling law of Eq. (2): T_e,target=10^10 K and t_cool=Ω^-1. The target temperature sets H/r (Sec. 2.1 notes that a realistic 10^9 K target would give H/r~0.02, unresolved), hence the midplane density; the Coulomb equilibration rate entering t_Coul (Eq. 11) depends steeply on density and electron temperature. A physical inverse-Compton target near 10^9 K and the much shorter physical cooling time acknowledged in Sec. 4.5 would produce a thinner, denser disc with stronger Coulomb coupling, plausibly moving r_tr inward toward r_ISCO. Sec. 4.1 and Sec. 5 explicitly concede that the exact transition radii depend on the target temperature. Therefore the two-temperature plunging region (t_infall < t_Coul inside the ISCO, Sec. 3.1.1) is robust, but the intermediate-accretion-rate truncation at r>r_ISCO and the specific functional dependence in Fig. 11 are partly set by an imposed, non-physical cooling prescription. The qualitative accretion-rate trend may survive, but the quantitative radii and the claim that mdot~1e-2 truncates at 4–5 r_g are not established.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper reports 3D GRMHD simulations of weakly magnetized thin accretion discs around a rapidly spinning black hole, with electrons and protons evolved as separate fluids and cooling applied to the electron fluid only. The central claim is that the plunging region of a thin disc is necessarily two-temperature because the infall time inside the ISCO becomes shorter than the Coulomb equilibration time, and that at intermediate accretion rates the two-temperature region extends outside the ISCO, producing an accretion-rate-dependent truncation radius that approaches the ISCO at high rates and moves outward to roughly 4-5 r_g near 10^-2 Eddington. The paper also reports extended surface cooling, a reduction of thermal emission from the plunging region relative to single-temperature models, and implications for X-ray binary state transitions, coronal origins, and spin measurements. The simulations use an electron cooling function with a fixed target temperature of 10^10 K and a cooling time set to the orbital time, with the main quantitative results concentrated in Sec. 3 and summarized in Fig. 11.","tokens_in":20019,"tokens_out":5162,"duration_ms":57252,"significance":"If the qualitative result holds, this is an important step beyond single-temperature thin-disc treatments: it provides a physical, Coulomb-coupling-based reason for a two-temperature plunging region and for an accretion-rate-dependent truncation that does not rely on magnetic arrest. The paper is honest about the limitations of its cooling prescription, and it includes useful implementation tests (Appendix A) and a clear timescale analysis (Secs. 3.1.1 and 3.2.3). The implicit Coulomb solver and the use of previously benchmarked heating prescriptions are strengths. However, the quantitative transition radii in Fig. 11 and the associated radiative efficiencies in Sec. 3.3 are tied to an imposed, non-physical electron cooling law, so the quantitative claims should be viewed as provisional until their sensitivity to that law is established.","major_comments":[{"comment":"The quantitative truncation radii in Fig. 11 are not independent of the imposed electron cooling law. With Q_cool = 2 u_e (T_e/T_e,target)^0.5 / t_cool, holding T_e,target = 10^10 K sets the electron temperature, hence the sound speed, H/r, and midplane density in the cooled equilibrium, and the Coulomb equilibration rate entering Eq. (11) depends steeply on density and electron temperature. The paper itself states that a realistic 10^9 K target would give H/r ~ 0.02, below the resolution of the simulations (Sec. 2.1), and Sec. 4.1 concedes that the exact accretion rates and radii may not remain the same with a lower target temperature. Therefore the specific claim that r_tr grows from about 2 r_g at mdot ~ 1 to about 4-5 r_g at mdot ~ 10^-2 is partly set by numerical convenience rather than by physical Compton cooling. To establish the quantitative claim, the authors should either repeat the M2 and M1 runs with a target temperature near 10^9 K (with correspondingly higher resolution), or provide an analytic scaling argument showing how r_tr depends on T_e,target and verify that the trend in Fig. 11 survives that scaling.","section":"Sec. 2.1, Eq. (2); Sec. 4.1; Fig. 11"},{"comment":"The choice t_cool = Omega^-1 has the same load-bearing role as T_e,target. Equation (2) uses t_cool directly in the cooling rate, and the manuscript acknowledges in Sec. 4.5 that the physical cooling time is likely much shorter. A shorter cooling time would keep electrons closer to the target temperature, changing the density and H/r and therefore the Coulomb coupling efficiency that controls the truncation in Sec. 3.2. Because r_tr is set by the balance between Coulomb cooling and viscous heating, and Q_coul depends on density, the sign and magnitude of the effect of a realistic cooling time on r_tr should be quantified, for example by varying t_cool in a subset of runs or by an analytic estimate of the resulting density change.","section":"Sec. 2.1 and Sec. 4.5 (Eq. 2)"},{"comment":"The cooling function assumes optically thin, local removal of energy (Eq. 1), yet Fig. 10 (left) shows that the M1 and M0 discs reach midplane Thomson optical depths of order 10 inside the ISCO and are optically thick in some of the same regions where the truncation forms. In an optically thick flow, cooling cannot be treated as local and immediate; photon trapping and reabsorption alter the temperature and density structure and therefore the Coulomb coupling that drives the transition. This limitation is acknowledged in Sec. 4.5, but it is connected to the quantitative transition radii because the density that sets Q_coul is established by exactly this cooling prescription. The authors should state explicitly which results (the qualitative two-temperature plunging region, the quantitative r_tr, or the radiative efficiencies in Table 2) are robust to this assumption, or test the sensitivity with a simplified radiation treatment.","section":"Sec. 2.1 and Sec. 3.3.2; Fig. 10 left"},{"comment":"The reported simulations are not checked for numerical convergence. The MRI quality factors in Table 1 are Q_theta = 7-8 and Q_phi = 12-14 for the thin-disc runs M2, M1, and M0, which are at or below commonly adopted thresholds for resolving the MRI; the paper itself notes in Sec. 4.5 that the low resolution potentially suppressed some turbulence. Since the disc density and scale height in the cooled equilibrium, and hence t_Coul and r_tr, depend on the turbulent dissipation realized by the MRI, a resolution study, at least for M2, is needed before the specific values of r_tr in Fig. 11 can be taken as physical. The statement that accretion and dissipation rates were roughly constant over time does not substitute for a convergence test.","section":"Sec. 2.3 and Table 1"}],"minor_comments":[{"comment":"The sentence describing the simulation domain says it extends 2π radians in π, which is presumably a typo for the azimuthal direction φ; please correct.","section":"Sec. 2.3"},{"comment":"The phrase 'leading the a more ADAF-like disc structure' in the M2 paragraph is grammatically garbled and should be rewritten.","section":"Sec. 3.2.3"},{"comment":"The sentence 'The optically-thin nature could have implications for produces the hard spectrum of XRBs' has a subject-verb error ('produces' should probably be 'producing'); please revise.","section":"Sec. 4.2"},{"comment":"Fig. 11 shows only four simulation points and a lower limit, with no time-averaging window or temporal scatter indicated; adding error bars or a statement of the time variability of r_tr would make the claimed monotonic trend easier to assess.","section":"Fig. 11"},{"comment":"The in-text citation to Naethe Motta et al. (2025) appears as 'NaetheMottaP.' in the reference list, which is a formatting artifact that should be corrected for readability.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The qualitative central claim, that Coulomb decoupling makes the plunging region two-temperature and can produce accretion-rate-dependent truncation, is credible and well supported by the timescale analysis. The main weakness is that the quantitative radii in Fig. 11 and the radiative efficiencies in Table 2 are determined by the imposed cooling law, and the manuscript itself admits the target temperature is an order of magnitude too high and the cooling time may be too long. This is fixable within the scope of the paper if the authors either run the suggested tests or clearly reframe the quantitative claims as sensitivity-dependent. I would not recommend rejection, because the core mechanism is physically motivated and the limitations are stated explicitly."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First, what you should know: this is the first global GRMHD suite I know of that evolves electrons separately and lets Coulomb coupling decide where a thin disc truncates. The central qualitative result—that the plunging region is two-temperature—is solid. The quantitative truncation radii in Fig. 11 are not.\n\nThe paper does several things well. The timescale analysis in Sec. 3.1.1 is clear: inside the ISCO the infall time becomes shorter than the Coulomb equilibration time, so proton-electron decoupling is generic. That argument doesn't depend on the details of the cooling law, and it matches the analytic prediction from Hankla et al. (2022). The implicit Coulomb solver is tested in Appendix A, and the four simulations span six orders of magnitude in accretion rate, which gives the qualitative trend some weight. The authors also flag their own limitations honestly, which I appreciate.\n\nWhere it gets soft: the electron cooling law is imposed, not derived. T_e,target is 10^10 K, about ten times the realistic Compton temperature, and t_cool is set to the orbital time rather than the physical cooling time, which is likely much shorter. The cooling also assumes optically thin emission even where the disc is optically thick. These choices set the density and scale height, which in turn control the Coulomb coupling rate—so the specific values of r_tr in Fig. 11 (2 r_g at Eddington, 4–5 r_g at 1e-2 Eddington) are not robust predictions. The authors admit this in Secs. 2.1 and 4.1; they say explicitly that a more realistic 10^9 K target would give H/r ~ 0.02, which they can't resolve. That's a real limitation, not a minor one.\n\nThere's also no resolution convergence study, and the MRI quality factors are marginal (Q_theta as low as 7), which is a known concern for thin disc simulations. The data are only available on request, so independent checks are harder.\n\nIn balance: the two-temperature plunging region is probably real. The claim that the truncation radius moves outward at intermediate accretion rates is plausible, but the quantitative mapping between accretion rate and r_tr should be treated as illustrative until someone runs with a more physical cooling law or a sensitivity study. This paper is worth reading for anyone working on XRB state transitions, coronae, or spin measurements—it changes how you think about the inner edge—but it is a qualitative demonstration, not a quantitative model.\n\nI would send it to peer review. It deserves referee time. The main revision I'd ask for is to soften the quantitative claims in the abstract and conclusions, or to add a test showing how r_tr depends on the cooling parameters. A companion paper doing the 10^9 K run at higher resolution would be the natural follow-up.","headline":"First global GRMHD suite to let Coulomb coupling set thin disc truncation: the two-temperature plunging region is solid, but the quantitative transition radii are set by an imposed cooling law.","tokens_in":20583,"tokens_out":5380,"would_cite":true,"duration_ms":46798,"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":"Thin accretion discs decouple into a two-temperature flow inside the plunging region, and at intermediate accretion rates the decoupling starts outside the ISCO, truncating the disc at an accretion-rate-dependent radius.","keywords":["accretion discs","black hole physics","two-temperature plasma","Coulomb decoupling","plunging region","GRMHD simulations","X-ray binaries","disc truncation"],"falsifier":"Run the same simulation suite with $T_{e,\\mathrm{target}}=10^9\\,\\mathrm{K}$ and sufficient resolution to capture $H/r\\sim 0.02$: if the disc remains single-temperature down to the ISCO, or if the truncation radius stops depending on accretion rate, the central claim fails. An observational check would compare the predicted inner edge, drifting from about $2\\,r_g$ at near-Eddington rates to $4$-$5\\,r_g$ near $10^{-2}$ Eddington, against continuum-fitting or reverberation measurements of the inner disc radius across X-ray binary states.","tokens_in":19461,"feed_emoji":"🕳️","tokens_out":9450,"duration_ms":90578,"temperature":0.7,"pith_summary":"This paper argues that a geometrically thin accretion disc cannot remain a single-temperature, thermally radiating flow all the way down to the black hole. In three-dimensional general-relativistic magnetohydrodynamic simulations that evolve electrons and protons separately, with an electron-only cooling function, the protons stop equilibrating with the electrons once the infall time becomes shorter than the Coulomb coupling time, which happens inside the innermost stable circular orbit (ISCO) and, at intermediate accretion rates, already outside it. The authors find a two-temperature plunging region at all rates they model: near Eddington accretion the decoupling begins at $r\\approx 2.04\\,r_g$, close to the ISCO for spin $a=0.9375$, while at $\\dot M \\sim 10^{-2}\\dot M_{\\rm Edd}$ it begins near $3.6\\,r_g$, rising to about $5\\,r_g$ at lower rates. The standard thin-disc model assumes the disc stops radiating at the ISCO, so the result changes predictions for emission from the plunging region, the likely seat of the X-ray corona, X-ray binary state transitions, and black-hole spin measurements.","feed_headline":"Coulomb decoupling truncates thin discs at 2–5 r_g","feed_subtitle":"Simulations show the cold disc gives way to a hot inner flow whose radius depends on accretion rate.","key_machinery":"The load-bearing device is an electron-only cooling function coupled to a separate electron fluid: energy is removed from electrons at the rate $Q_\\mathrm{cool}=2u_e(T_e/T_{e,\\mathrm{target}})^{0.5}/t_\\mathrm{cool}$, with $t_\\mathrm{cool}$ set to the local orbital time and the target temperature fixed at $10^{10}\\,\\mathrm{K}$, while protons lose energy only through Coulomb collisions with electrons. The competition that sets the structure is a comparison of timescales: where the infall time is longer than the Coulomb equilibration time and the viscous heating time, the disc collapses to the standard thin, single-temperature state; where the infall time becomes the shortest timescale, protons decouple and the disc transitions to a hot, thick flow. The paper locates this transition with the radius $r_\\mathrm{tr}$ at which more than half of the gas has proton temperature above the target electron temperature, and shows it moves outward as accretion rate drops.","core_discovery":"The paper's central discovery is that the plunging region of a thin accretion disc is necessarily two-temperature, and the decoupling can propagate outside the ISCO to form a truncated disc whose inner edge depends on accretion rate. In the simulations, a disc accreting near the Eddington rate stays single-temperature until $r_\\mathrm{tr}\\approx 2.04\\,r_g$, then decouples; one accreting near $10^{-2}$ Eddington decouples at $r_\\mathrm{tr}\\approx 3.6\\,r_g$, with the transition radius increasing monotonically to $\\sim 5\\,r_g$. The mechanism is the ordering of timescales: inside the region where protons decouple, the infall time is shorter than the proton viscous-heating and Coulomb-cooling times, so the protons heat above the electron target temperature, the disc thickens ($H/r$ rises toward $\\sim 0.2$), and the inner flow becomes hot, thick, and radiatively inefficient. A further result is that cooling is not confined to the midplane: a sandwich-like layer above the disc radiates out to $\\sim 10\\,r_g$, and at intermediate accretion rates about 40% of the cooling occurs above the disc body. Relative to single-temperature models, thermal emission from inside the ISCO is suppressed, reaching roughly 75% of the single-temperature value at the highest rate and less than 50% at intermediate rates, with a large fraction emitted from optically thin, two-temperature regions.","pith_inferences":["If the same timescale ordering holds with a realistic cooling law, the hot two-temperature inner region could itself provide the Comptonizing electrons for the soft-state hard X-ray tail, without requiring a separately heated corona; the present paper identifies the two-temperature region but does not construct the spectral model.","Because the decoupling criterion is a ratio of infall to Coulomb times, the truncation radius should scale with the ISCO radius as black-hole spin changes; a fixed-accretion-rate spin sequence would test whether Coulomb truncation tracks $r_\\mathrm{ISCO}$ rather than magnetic flux.","The predicted $2$-$5\\,r_g$ drift with accretion rate is narrow enough to search for with existing X-ray timing or continuum-fitting samples: soft-state binaries should show an inner edge that moves outward by roughly a factor of two as luminosity drops toward $10^{-2}$ Eddington.","A phenomenological two-zone disc model with the boundary $r_\\mathrm{tr}(\\dot M, a)$ taken from this mechanism would make the prediction directly usable in broad-band spectral fits of X-ray binaries; the paper stops at reporting the simulation result rather than providing such a fitting prescription."],"forward_implications":["At near-Eddington accretion rates, even the canonical thin-disc case has a two-temperature plunging region beginning at $r \\approx 2\\,r_g$, so single-temperature treatments of the inner disc are incomplete.","The inner edge of the thin disc moves outward with decreasing accretion rate, from about $2\\,r_g$ at Eddington to $3.6$-$5\\,r_g$ at $10^{-2}$ Eddington, so disc truncation by Coulomb decoupling alone is accretion-rate dependent.","Thermal emission from inside the ISCO is weaker than single-temperature models predict, down to about 75% of the single-temperature value at high rates and below 50% at intermediate rates, with much of it optically thin.","A substantial part of the cooling occurs above the disc body, about 40% at intermediate rates, which bears directly on where an X-ray corona forms.","Spin measurements that map the disc's peak temperature to the ISCO radius can be biased: at intermediate accretion rates the disc appears truncated near $4\\,r_g$ rather than the true $2\\,r_g$, and the reduced plunging-region thermal emission shifts the inferred ISCO location."],"supporting_citations":[{"why":"Provides the canonical thin-disc model whose single-temperature assumption the paper challenges.","marker":"Shakura & Sunyaev 1973"},{"why":"Supplies the standard prediction of zero dissipation and radiation inside the ISCO, the benchmark the simulations exceed.","marker":"Novikov & Thorne 1973"},{"why":"Gives the radial luminosity profile f(r) used to compare simulated cooling to the analytic efficiency.","marker":"Page & Thorne 1974"},{"why":"Analytic model of magnetic torques that predicts dissipation within the plunging region, which the simulations support in modified form.","marker":"Gammie 1999"},{"why":"Introduces the cooling-function approach and the ISCO orbital frequency used to set t_cool and define H/r.","marker":"Noble et al. 2009"},{"why":"Provides the separate electron-fluid evolution scheme on which the two-temperature thermodynamics build.","marker":"Ressler et al. 2015"},{"why":"Gives the Coulomb energy-exchange rate that sets the proton-electron coupling and hence the decoupling radius.","marker":"Stepney & Guilbert 1983"},{"why":"Supplies the reconnection-based electron heating fraction used to partition dissipated energy between electrons and protons.","marker":"Werner et al. 2018"},{"why":"Previously predicted analytic decoupling in the plunging region and the associated hard X-ray tail, motivating the simulation setup.","marker":"Hankla et al. 2022"},{"why":"Defines the transition-radius criterion where more than half of the gas exceeds the target temperature, adopted as Eq. 16.","marker":"Hogg & Reynolds 2018"}],"fun_headline_variants":["Thin discs truncate at 2-5 r_g, rate-dependent","Two-temperature physics sets thin disc inner edge","Plunging region is two-temperature, simulations show","Accretion rate sets where cold disc becomes hot flow","40% of disc cooling happens above the midplane"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The quantitative truncation radii rely on an imposed electron cooling law, a fixed target temperature of $10^{10}\\,\\mathrm{K}$ and a cooling time tied to the orbital time, that sets the disc's density and scale height; a more realistic target near $10^9\\,\\mathrm{K}$ or a shorter cooling time would change the simulated radii.","fun_headline_variants_meta":{"raw":{"variants":["Thin discs truncate at 2-5 r_g, rate-dependent","Two-temperature physics sets thin disc inner edge","Plunging region is two-temperature, simulations show","Accretion rate sets where cold disc becomes hot flow","40% of disc cooling happens above the midplane"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000949,"raw_usage":{"total_tokens":4123,"prompt_tokens":1088,"completion_tokens":3035,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":704,"completion_tokens_details":{"reasoning_tokens":2955}},"tokens_in":704,"tokens_out":3035,"duration_ms":19875,"temperature":1.0,"reasoning_tokens":2955,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T05:10:28.479151+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the same simulation suite with $T_{e,\\mathrm{target}}=10^9\\,\\mathrm{K}$ and sufficient resolution to capture $H/r\\sim 0.02$: if the disc remains single-temperature down to the ISCO, or if the truncation radius stops depending on accretion rate, the central claim fails. An observational check would compare the predicted inner edge, drifting from about $2\\,r_g$ at near-Eddington rates to $4$-$5\\,r_g$ near $10^{-2}$ Eddington, against continuum-fitting or reverberation measurements of the inner disc radius across X-ray binary states.","supporting_citations":[],"review_version":1}