{"id":"07f933f9-1934-4a39-a158-d125028709f2","arxiv_id":"2608.12462","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"In simulated massive galaxies, recycled stellar ejecta from inside the host galaxy dominate the metal-rich gas accreted by supermassive black holes, naturally explaining high and weakly evolving quasar metallicities.","lead":"Using 30 computer simulations of massive galaxies, this paper follows every gas particle that ends up inside a supermassive black hole. It finds that most of that fuel is recycled stellar material from the host galaxy, already rich in metals before it arrives.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Subgrid Bondi selection may itself create the recycled-gas dominance; an inflow-based origin census would test whether the 58.7% fraction is robust.","rationale":"The central quantitative claim is the 58.70% recycled fraction of accreted gas. Every downstream statement about the chemical state of the fuel inherits that number. The reader's weakest assumption identifies the soft-Bondi accretion prescription and subgrid AGN feedback as the selection mechanism that defines which particles enter the census. That is exactly the point where the claim is least secure: the accretion model is not a passive filter but a density- and temperature-weighted selector, and the recycled component is plausibly the component it would preferentially select. This is not a criticism of the simulations per se; the authors carefully scope their conclusions in Section 4.4 and acknowledge that torque-limited studies can modify accretion budgets. But the title and summary items make a stronger statement about the fuel of SMBHs than about the fuel selected by one subgrid model. The proposed inflow-based check is a post-processing test that can be run on the existing outputs without new simulations, and it directly determines whether the selection is biasing the origin fractions. If the recycled fraction is similar in the inflow-defined sample, the concern is resolved; if not, the headline claim would need to be weakened to a statement about the subgrid model rather than about SMBH fuel in general. Because this concern is already the basis of the reader's conditional verdict, my assessment does not move the verdict; it reinforces the conditions under which the claim would become robust.","tokens_in":18040,"tokens_out":8722,"duration_ms":89428,"concrete_test":"Use the existing snapshots to compute an accretion-independent origin census: for each galaxy and snapshot, take all gas particles within 0.8 kpc/h (2 times the gas softening length) of the BH that are gravitationally bound to the BH and have inward radial velocity, weight by mass, and assign the same origin categories as Section 2.2. Compare the resulting mass fractions to the Bondi-selected census in Figure 1. If the recycled fraction drops from 58.7% by more than about 10 percentage points, or if the [Fe/H] distribution of the inflow sample differs significantly from the accreted sample, the soft-Bondi selection is driving the headline result. For a stronger test on a few representative galaxies, rerun the tracing with a torque-limited accretion model and recompute the origin fractions.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The headline census is built from gas particles that the stochastic soft-Bondi criterion (Section 2.1) selects for accretion. This is not a neutral sampling of the central gas reservoir: the Bondi rate favors high-density, low-sound-speed gas, and recycled stellar ejecta are preferentially deposited in exactly that dense, enriched, central component. The paper's own Section 4.4 concedes that unresolved small-scale physics, such as torque-limited accretion, can modify the accretion budget. If angular-momentum transport or other sub-resolution processes prevent some of this Bondi-selected recycled gas from actually reaching the black hole, or admit a different component, the 58.70% recycled fraction and the associated 'already metal rich' chemical characterization would not be robust. The logical step from 'recycled gas dominates in this subgrid model' to 'recycled gas dominates the metal-rich fuel of SMBHs' therefore depends on an untested assumption: that the subgrid selection does not preferentially choose recycled gas. The paper presents the census as a property of the cosmological gas supply, but the sample is defined by the subgrid accretion rule rather than by an accretion-independent measure of the gas delivered to the central region.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper analyzes 30 high-resolution cosmological zoom-in simulations of massive galaxies from the Choi et al. (2017) suite, tracing individual gas particles that are accreted by the central supermassive black hole (SMBH) under a soft-Bondi prescription. The accreted gas is classified into four origin categories: Recycled, Early, External, and Smooth. The authors report that, on average, 58.70% of the accreted gas mass is Recycled (originating from stellar mass loss within the main galaxy), that this component is already metal-rich at early epochs, and that the gas reaching the SMBH is more metal-rich than the average central galaxy gas. They further examine the redshift evolution of [Fe/H] and [Mg/Fe], find weak trends, and argue that the simulated abundance ratios are broadly compatible with observed quasar broad-line-region metallicities. The paper concludes that metal-rich SMBH fuel is a natural consequence of cosmological galaxy evolution and stellar recycling, and it explicitly frames the results as providing cosmological boundary conditions rather than a direct model of the sub-parsec accretion flow.","tokens_in":18242,"tokens_out":4407,"duration_ms":41479,"significance":"If the reported census is robust, the paper makes a valuable contribution by linking SMBH fueling to internal stellar recycling and by providing a physically grounded explanation for the high, weakly evolving metallicities inferred in quasar environments. The strengths include the large 30-galaxy zoom-in sample, the explicit particle tracing with an 11-element chemical network, the systematic treatment of four origin channels, and the candid enumeration of limitations in Section 4.4. The paper also shows appropriate restraint in treating [Mg/Fe] as a secondary diagnostic and in labeling the observational comparison as qualitative. The central concern is that the headline quantitative result, the 58.70% recycled fraction, is not yet shown to be robust to numerical resolution or to the choice of subgrid accretion model, which is load-bearing for the paper's main claim.","major_comments":[{"comment":"The central quantitative claim that Recycled gas constitutes 58.70% of the accreted mass rests on a single simulation resolution and a single subgrid accretion prescription; no resolution-convergence test is reported, even though the gas distribution and feedback behavior in such simulations can change with resolution. Please add a convergence test on a subset of runs at higher resolution, or explicitly reframe the headline result as conditional on the adopted numerical setup.","section":"§2.1, Figure 1"},{"comment":"The Early category is defined by first accretion onto the main halo at z≥3, and the text states that recycled gas produced at z>3 is deliberately assigned to Early rather than Recycled. This makes the origin categories non-exclusive and could bias the chemical comparison between Early and Recycled gas reported in Section 3. Please quantify how many z>3 recycled particles are placed in Early and demonstrate that the main conclusions are unchanged if they are reclassified.","section":"§2.2, item 2"},{"comment":"The paper concedes that small-scale physics can modify the accretion budget, yet the title and summary assert that Recycled gas dominates the metal-rich fuel of SMBHs. Because the traced sample is selected by the stochastic soft-Bondi criterion, which preferentially chooses dense, low-sound-speed gas, the census may overrepresent centrally enriched recycled ejecta relative to an inflow-based accounting. A minimal remedy is to report the origin distribution of gas within r10 as an accretion-independent reference, and to adjust the conclusions to the demonstrated 'gas reservoir delivered to the smallest resolved scales' rather than to the final accretion flow.","section":"§4.4"},{"comment":"The paper states that the comparison with quasar BLR abundances should be interpreted qualitatively, and the footnote notes that the observational values are not direct intrinsic abundances. This is a reasonable caution, but the abstract's phrase 'broadly compatible' is stronger than the analysis supports; please make the abstract and the conclusions reflect the qualitative nature of this comparison more explicitly.","section":"§4.2, footnote 7"}],"minor_comments":[{"comment":"The sample-averaged recycled fraction is quoted as 58.70% without a scatter or range; given the visible system-to-system variation, please quote the standard deviation or interquartile range as well.","section":"Figure 1"},{"comment":"The definition of r10 as 0.1×r_vir is used to define the central galaxy, but the text does not state whether the same r10 criterion is applied consistently at all redshifts for the classification; please clarify.","section":"§2.2"},{"comment":"The abstract says the abundance ratios 'are broadly compatible with the high metallicities inferred for quasar broad-line regions,' but the body's limitation statement is stronger; aligning the abstract with the body would improve precision.","section":"Abstract"},{"comment":"The draft header includes the string 'LATEXtwocolumnstyle', which appears to be a formatting artifact; this should be cleaned up before final submission.","section":"General"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe thing to know: this paper reports a clean census of SMBH fuel origins in 30 zoom-in simulations and adds chemistry that Paper I did not have. The headline number—58.7% of accreted gas mass recycled—is real within the model, but it is a product of the soft-Bondi accretion prescription, not a direct measurement of the gas that reaches the black hole. The authors acknowledge the subgrid dependence in Section 4.4, yet the title and abstract outrun the evidence.\n\nWhat is new and good: the [Fe/H] and [Mg/Fe] histories split by origin category, the weak redshift evolution of the accreted gas, and especially Figure 4, which shows that gas reaching the BH is more metal-rich than the average galaxy gas at z~1–3. That offset is a genuine result: the accretion model is selecting centrally enriched material, and the paper is transparent that the BLR comparison is qualitative (footnote 7). I also appreciate the Spearman tests with the explicit warning that large particle counts drive significance.\n\nSoft spots, in order of importance:\n\n1. The Bondi-selection concern from the stress test is real. Bondi favors high-density, low-sound-speed gas, and recycled ejecta are deposited in exactly that component. So the origin census is partly constructed by the subgrid rule. The paper cites torque-limited studies but does not run one on the same halos. It does not have to, but then the claim should be 'recycled gas dominates in this subgrid model,' not 'dominates the fuel.' I would ask for a test using an inflow-based or torque-limited accretion census, or for a revised title.\n\n2. No resolution-convergence test. For a simulation paper at this level, that is a gap, though probably not a load-bearing one.\n\n3. The Early category is partially confounded with Recycled by construction. Minor, since Early is ~10%.\n\n4. No data or code release, so the tracing definitions are not independently checkable. Minor.\n\nThe observational comparison is explicitly qualitative, so the paper does not resolve the quasar metallicity puzzle; it shows that the simulated fuel is consistent with the observed BLR metallicities. That is a useful but weaker statement.\n\nBottom line: a solid, honestly written simulation study that deserves peer review. I would send it out, with the request that the authors test the accretion-prescription dependence or temper the title. The central result is internally consistent; the limitation section is candid; the chemistry is a real step forward.","headline":"The chemistry is new and the census is clean, but the 58.7% recycled-gas fraction is a subgrid-Bondi-selected number, so the title should not outrun the model.","tokens_in":18823,"tokens_out":4719,"would_cite":true,"duration_ms":41705,"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":"Across 30 simulated massive galaxies, gas recycled within the host galaxy supplies roughly 59% of the mass accreted by supermassive black holes, and this gas is metal-rich from early cosmic times.","keywords":["active galactic nuclei","supermassive black holes","chemical abundances","hydrodynamical simulations","gas recycling","black hole fueling","quasar metallicity","cosmological zoom-in simulations"],"falsifier":"A zoom-in simulation of the same halo sample that resolves the accretion flow down to sub-parsec scales, rather than using the soft-Bondi subgrid recipe, would settle the claim: if the gas reaching the black hole in such runs is dominated by cold, metal-poor material from the cosmic web or from mergers instead of by recycled stellar ejecta, the 58.7% recycled census would be overturned.","tokens_in":17835,"feed_emoji":"♻️","tokens_out":15991,"duration_ms":112643,"temperature":0.7,"pith_summary":"This paper claims that the gas feeding supermassive black holes in massive galaxies is predominantly recycled material — gas that was once inside stars in the host galaxy and returned to the interstellar medium through stellar winds and supernova explosions. Tracing individual gas particles in 30 cosmological zoom-in simulations, the authors find that recycled gas makes up on average 58.7% of all gas accreted by the black holes, with merger-delivered (external), smooth intergalactic, and early-assembly gas contributing smaller fractions. Because recycled gas is born in a metal-rich central environment, it arrives at the black hole already enriched, which naturally explains why quasar broad-line regions show high, roughly redshift-independent metallicities. If this is right, the metal-rich fuel of black holes is a byproduct of ordinary stellar aging in galaxies rather than of mergers or exotic nuclear-scale enrichment.","feed_headline":"Recycled gas dominates the metal-rich fuel of black holes","feed_subtitle":"Stellar ejecta supply ~59% of black hole fuel, naturally explaining metal-rich quasars.","key_machinery":"The central object is the traced gas particle: each gas element that ever joins the black hole is followed from the moment of black-hole seeding to accretion, with its formation site, trajectory, and chemical history recorded at every simulation snapshot. The argument's load-bearing device is the four-way origin census these trajectories make possible — Recycled, Early, External, and Smooth — combined with the mass-weighted abundance histories that assign each class its characteristic $[{\\rm Fe/H}]$ and $[{\\rm Mg/Fe}]$. This particle-level classification is what converts the raw simulation into the claim that recycled stellar ejecta dominate the fuel budget.","core_discovery":"The paper's central discovery is that the cosmological gas supply to supermassive black holes in massive galaxies is dominated by material recycled through stellar evolution within the main galaxy. By classifying the accreted gas into four origin categories — Early (first entering the main halo at $z\\ge 3$), External (from other halos or subhalos), Recycled (from asymptotic giant branch winds and supernova ejecta inside the main galaxy), and Smooth (from the intergalactic medium) — the authors find sample-averaged accreted-mass fractions of 58.7% Recycled, 19.8% External, 11.8% Smooth, and 9.8% Early. Recycled gas is intrinsically metal-rich already at early epochs, while the other components become enriched gradually while residing in the galactic environment before accretion. The resulting $[{\\rm Fe/H}]$ of the accreted population shows only weak redshift evolution and $[{\\rm Mg/Fe}]$ declines mildly, both broadly consistent with observed quasar broad-line region abundances. The authors conclude that the metal-rich nature of supermassive black hole fuel arises naturally from cosmological galaxy evolution and stellar recycling, and need not invoke sub-parsec enrichment processes.","pith_inferences":["If recycled gas dominates the fuel supply, quasar metallicity should correlate more strongly with a host galaxy's old stellar mass and age than with recent merger activity or gas-richness, a distinction testable with resolved stellar-population observations of quasar hosts.","The sample is restricted to massive, group-scale, largely quenched galaxies at z = 0; in lower-mass, star-forming, or gas-rich systems the recycled fraction is likely smaller, and smooth or external gas may dominate at high redshift.","The weak redshift evolution of the accreted gas iron abundance implies that, at a fixed epoch, the gas reaching a black hole should be as metal-rich as the most enriched gas in the inner circumgalactic medium, which could be tested by comparing quasar broad-line metallicities with host stellar metallicities at matched stellar mass."],"forward_implications":["Mergers are not the dominant long-term fuel source for supermassive black hole growth in massive group-scale galaxies; recycled stellar ejecta sustain the accretion budget instead.","The high, nearly redshift-independent metallicities inferred for quasar broad-line regions need not be produced by sub-parsec enrichment; they follow naturally from galaxy-scale stellar recycling.","Because recycled gas is preferentially metal-rich, gas accreted by the black hole is systematically more enriched than the average interstellar gas of the host galaxy, especially at z ~ 1-3.","The mild [Mg/Fe] decline of accreted gas toward low redshift is a direct prediction of delayed Type Ia supernova iron enrichment that could be checked against quasar abundance-ratio evolution.","In quenched massive galaxies at low redshift, stellar mass loss alone can continue to feed the black hole, providing a physical basis for observed AGN activity in quiescent galaxies."],"supporting_citations":[{"why":"Supplies the 30-galaxy zoom-in simulation suite, black-hole seeding, soft-Bondi accretion, and AGN feedback model that produce the accreted-gas sample.","marker":"E. Choi et al. (2017)"},{"why":"Paper I; develops the particle-tracing method and first showed recycled gas dominates black-hole fuel, which this work extends to chemistry and a larger sample.","marker":"E. Choi et al. (2024)"},{"why":"Provides the star-formation and chemical-enrichment model with 11 tracked species, stellar yields, and metal diffusion used to compute [Fe/H] and [Mg/Fe].","marker":"M. Aumer et al. (2013)"},{"why":"Gives the expectation that recycled gas from stellar mass loss is a major reservoir in old stellar systems, motivating the recycled category.","marker":"S. N. Leitner & A. V. Kravtsov (2011)"},{"why":"Provides the observationally inferred quasar abundance ratios used for the qualitative comparison of simulated [Fe/H] and [Mg/Fe].","marker":"Y. Yoshii et al. (2022)"},{"why":"Earlier particle-tracing study of black-hole gas origins showing low-angular-momentum selection, providing the dynamical comparison for this origin census.","marker":"N. N. Sanchez et al. (2018)"},{"why":"Earlier tracing of gas accreted onto intermediate-mass black holes at z~4, showing black-hole growth reflects host gas composition.","marker":"J. Bellovary et al. (2013)"},{"why":"Isolated-elliptical simulations showing recycled stellar gas alone can sustain episodic AGN activity, supporting the physical plausibility of the main claim.","marker":"L. Ciotti & J. P. Ostriker (2007)"}],"fun_headline_variants":["Stellar recycling feeds most black hole gas","Recycled gas makes up 59% of black hole fuel","Black hole fuel mostly comes from recycled stars","Star ejecta dominate black hole gas supply","Metal-rich quasar fuel from stellar recycling"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The paper assumes that unresolved small-scale physics near the black hole does not change which large-scale gas reservoirs actually feed it, so the simulation's simplified subgrid accretion recipe reliably selects the gas that a real black hole would consume.","fun_headline_variants_meta":{"raw":{"variants":["Stellar recycling feeds most black hole gas","Recycled gas makes up 59% of black hole fuel","Black hole fuel mostly comes from recycled stars","Star ejecta dominate black hole gas supply","Metal-rich quasar fuel from stellar recycling"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000234,"raw_usage":{"total_tokens":1553,"prompt_tokens":1060,"completion_tokens":493,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":676,"completion_tokens_details":{"reasoning_tokens":422}},"tokens_in":676,"tokens_out":493,"duration_ms":4458,"temperature":1.0,"reasoning_tokens":422,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T00:08:10.442236+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A zoom-in simulation of the same halo sample that resolves the accretion flow down to sub-parsec scales, rather than using the soft-Bondi subgrid recipe, would settle the claim: if the gas reaching the black hole in such runs is dominated by cold, metal-poor material from the cosmic web or from mergers instead of by recycled stellar ejecta, the 58.7% recycled census would be overturned.","supporting_citations":[{"cited_title":"N., Bellovary, J","cited_arxiv_id":null,"evidence_quote":"Earlier particle-tracing study of black-hole gas origins showing low-angular-momentum selection, providing the dynamical comparison for this origin census."}],"review_version":1}