{"id":"dd0cf56f-9227-481e-b5b8-aa45f394dffc","arxiv_id":"2608.07732","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"JWST resolves a rotating circumnuclear disk in NGC 4696 whose kinematic centre is offset 41 pc from the previous AGN position, implying a roughly billion-solar-mass black hole and an inflow of about 18 solar masses per year.","lead":"JWST maps the inner 600 parsecs of NGC 4696 and finds a rotating disk of hot gas feeding the galaxy's central black hole. The new data shift the black hole's likely position by 41 parsecs and give a fresh look at how gas spirals into a giant black hole.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 41 pc offset between the kinematic centre and the VLBA AGN position is only ~2 NIRSpec spaxels, yet the NIRSpec-to-HST registration and the gradient-weighted centroid are both used without a quoted uncertainty; if the combined error is ~0.1-0.2 arcsec, the headline offset is consistent with…","rationale":"The paper does a serious job with a difficult dataset: LOKI fitting, WICKED wiggle correction, and comparison with MUSE and ALMA are real supporting elements, and the companion paper with MHD simulations strengthens the interpretation of a rotating circumnuclear structure. My concern is not with the data reduction or the existence of the CND, but with the central astrometric claim: the offset of the kinematic centre from the VLBA AGN position. This offset (Section 4.2.1, Fig. 18) is the basis for saying the kinematic centre is the true AGN/dynamical centre. It is ~41 pc, i.e. two 0.1\" spaxels, and no error bar is given. Section 2.1.1 states the NIRSpec WCS was set by matching the HST F665N swirl to the Paα morphology; without stellar point sources in the IFU field, such a match can easily be uncertain by 0.1\" or more. The Fabian et al. (2016) AGN position carries its own placement uncertainty. A 1σ total error of ~0.1-0.2\" would make the offset statistically indistinguishable from zero. The kinematic-centre estimator (gradient-weighted centroid) is also not tested against alternative tracers or components, and the Paα map in Fig. 3b is the blueshifted component only. For these reasons the headline 'reassessment of the AGN position' is not yet supported. The M_BH and r_inf values are flagged by the authors as order-of-magnitude, so I do not treat those as the main issue; the offset is presented without any caveat, making it the most load-bearing unsupported element. The proposed test—bootstrap the centroid, propagate the WCS registration residuals, and compare with H2-based centres—would settle whether the offset is real. Since the reader's verdict is already CONDITIONAL, my read does not change it; it sharpens the condition that must be met.","tokens_in":39542,"tokens_out":10722,"duration_ms":105088,"concrete_test":"Compute an end-to-end uncertainty for the 41 pc offset: (1) re-run the NIRSpec-to-HST registration with independent methods, such as cross-correlating the Paα map and the HST F665N image over multiple sub-regions and dither combinations, to obtain a WCS residual map; (2) bootstrap the Garcia-Lorenzo kinematic-centre estimator over spaxels while varying the gradient threshold and the SNR cut, and repeat it using the redshifted Paα component and the H2 1-0 S(1) velocity field; (3) combine these in quadrature with the Fabian et al. (2016) HST/VLBA placement error. If the 1σ total uncertainty on the offset is >=20 pc (0.1\"), the 41 pc offset is not statistically significant and the paper should not claim a reassessed AGN position.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's headline result—that the kinematic centre of the CND is offset by ~41 pc from the VLBA-based AGN position and is therefore the true dynamical centre (Sections 1, 4.2.1, Fig. 18)—rests on an unquantified astrometric and kinematic-centre error budget. The offset is only ~2 NIRSpec spaxels (0.1\" each), yet no uncertainty is quoted for the gradient-weighted centroid or for the registration of the NIRSpec cube to HST described in Section 2.1.1. Because the NIRSpec field contains no stars, the WCS was assigned by matching the extended HST F665N swirl to the Paα morphology; aligning two complex, low-surface-brightness structures without point-source anchors is unlikely to be accurate to better than ~0.1\", which is half the claimed offset. The comparison position (green star in Fig. 18) carries its own HST-registration uncertainty from Fabian et al. (2016). If the combined 1σ error is ~0.2\", the offset is consistent with zero and the 'reassessment of the AGN position' does not stand. The kinematic-centre estimator itself is also sensitive to the choice of Paα component (only the blueshifted component is shown in Fig. 3b), to the SNR threshold, and to the averaging region; none of these are varied. This is the most load-bearing assumption because the derived M_BH and r_inf are explicitly presented as order-of-magnitude, while the offset is presented as a firm, new position.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents JWST/NIRSpec IFU observations of the central ~620x620 pc^2 of NGC 4696, the brightest cluster galaxy of the Centaurus cluster. The authors map the kinematics of ionized (Pa-alpha) and warm molecular hydrogen (H2 1-0 S(1), S(3), S(5)) gas, as well as stellar kinematics derived from CO absorption band heads, at a pixel scale of ~20 pc. They report a compact rotating circumnuclear disk (CND) with a radius of ~60-120 pc, a kinematic centre offset by ~41 pc from the previously adopted VLBA-based AGN position, a peak Pa-alpha velocity dispersion of ~449 km/s implying a SMBH mass of ~10^9 Msun and a sphere of influence of ~60 pc, a steep velocity gradient along an inflowing filament (~4.7 km/s/pc, R^2=0.986), and an accretion rate lower bound of ~18 Msun/yr. They compare these data with MUSE H-alpha observations, showing that the CND is unresolved at MUSE's angular resolution, and with ALMA CO(2-1) data, which reveal only compact clumps without an extended counterpart. The paper concludes that the stellar kinematics are decoupled from the multiphase gas and that the observations demonstrate JWST's power to probe SMBH feeding in BCGs.","tokens_in":39878,"tokens_out":11504,"duration_ms":104809,"significance":"If the central claims hold, this paper would provide one of the first spatially resolved views of a circumnuclear disk and inflowing gas within the sphere of influence of a SMBH in a BCG, with direct implications for chaotic cold accretion and AGN feeding models. The data reduction is careful and transparent: the authors detail the use of LOKI, F-tests for multiple components, WICKED wiggle correction, and HST-based astrometric re-registration, and they make use of public software throughout. The MUSE comparison, where the JWST data are degraded to MUSE resolution and reproduce the MUSE velocity field, is an instructive cross-check. The PV diagrams and the velocity-gradient measurement (4.7 km/s/pc) are suggestive of accelerating inflow and are an interesting contribution. However, the headline results—the 41 pc offset of the kinematic centre from the AGN and the SMBH mass / resolved sphere of influence—are currently supported by order-of-magnitude analyses and unquantified systematics, and the accretion rate relies on masses deferred to a companion paper. The significance of the paper is accordingly conditional on these issues being addressed.","major_comments":[{"comment":"The claimed 41 pc offset between the kinematic centre and the AGN position established by Fabian et al. (2016) is presented without any uncertainty budget. The offset corresponds to only ~2 NIRSpec spaxels (0.1'' each, or ~20.6 pc), while the NIRSpec WCS was calibrated by matching the HST F665N 'swirl' to the Pa-alpha morphology (Section 2.1.1) with no quantitative accuracy estimate, and the kinematic centre is a gradient-weighted centroid (García-Lorenzo et al. 2015) with no quoted statistical or systematic error. In addition, the comparison AGN position from Fabian et al. (2016) carries its own HST-registration uncertainty. The paper should propagate these errors (e.g., via bootstrapping the centroid, measuring the registration scatter against independent tracers, or using the HST absolute astrometry) and demonstrate that the offset is significant; otherwise the 'reassessment of the AGN position' is not established.","section":"Section 4.2.1 (Fig. 18) and Section 2.1.1"},{"comment":"The SMBH mass estimate uses the maximum Pa-alpha velocity dispersion sigma=449±12 km/s measured at the kinematic centre, but this dispersion is derived from the line profile in a region where the paper itself finds a double-component structure with a high-dispersion redshifted component (Section 3.2, Fig. 4) and where non-gravitational motions are plausible. The paper correctly labels this an order-of-magnitude estimate in the text, but the Abstract and Conclusions present 'a SMBH mass of ~10^9 Msun' and 'r_inf ~60 pc, resolved by our observations' without this caveat. Since r_inf is proportional to M_BH, an order-of-magnitude uncertainty in M_BH translates directly into r_inf (from ~60 pc down to ~20 pc, i.e., one pixel), so the 'resolved' claim is not robust. Please either provide a proper dynamical model of the CND (as planned in the forthcoming paper) or explicitly qualify the headline claims as order-of-magnitude and avoid stating that the sphere of influence is resolved.","section":"Section 4.2.2 (Eq. 3) and the Abstract"},{"comment":"The accretion rate of ~18 Msun/yr is a headline result presented in the Abstract, but the warm H2 masses M_3, M_4, M_5 used in Eq. (6) are not derived in this paper; the text states 'We derive these quantities in detail in O. Pereira et al. (in preparation)' without presenting the column densities, excitation temperatures, or other inputs needed to evaluate Eq. (5). This makes the inflow rate unverifiable from the present manuscript. Please include the derivation of the H2 masses (or clearly mark the result as preliminary and remove it from the Abstract/Conclusions until the companion paper is available).","section":"Section 4.2.3 (Eq. 5) and the Abstract"},{"comment":"The paper assumes that the kinematic centre of the CND coincides with the SMBH position because the gas is governed by the SMBH gravitational potential. This assumption is not tested. The data show significant non-gravitational features—double-component Pa-alpha, triple-peaked H2 S(1) profiles (Appendix C), and possible outflow/inflow components—that could shift the apparent kinematic centre. Please test the robustness of the kinematic centre, for example by comparing the centre derived from different tracers (Pa-alpha vs. H2 1-0 S(1)) and by excluding spaxels with multiple components, and discuss how non-gravitational motions could affect the claimed offset.","section":"Section 3 (first paragraph) and Section 4.2.1"}],"minor_comments":[{"comment":"Equation (5) contains an undefined symbol 'A' in the term 'A 2 m_H'; please define all variables or remove the stray symbol.","section":"Section 4.2.3, Eq. (5)"},{"comment":"The Abstract states that the redshifted Pa-alpha component reaches sigma~600 km/s, but Section 3.2 reports typical values of ~370 km/s within the central 50 pc; please clarify whether 600 km/s is a maximum from the dispersion map (Fig. 4e) and reconcile the statements.","section":"Section 3.2 / Abstract"},{"comment":"The quoted uncertainties on the stellar and Pa-alpha redshifts (z=0.00984±0.00005 and z=0.00999±0.00002) correspond to velocity uncertainties of ~15 and ~5 km/s, which are smaller than the NIRSpec instrumental resolution (~110 km/s); please state that these are statistical only and that absolute calibration systematics are larger.","section":"Table 1"},{"comment":"The v_c/sigma ratio uses the central Pa-alpha velocity dispersion (sigma=449 km/s), which is dominated by the SMBH's sphere of influence and not representative of the disk's local dynamical support; consider using the disk's velocity dispersion at radii outside the nucleus (or justify why the central value is the appropriate comparison).","section":"Section 4.3.1, Eq. (7)"},{"comment":"The caption says 'only the blueshifted component is shown' for Pa-alpha, but the blueshifted component's velocity field in Fig. 4c extends to positive velocities; please define this component more clearly (e.g., 'the lower-velocity component').","section":"Fig. 3b caption"}],"recommendation":"major_revision","confidential_remarks":"The manuscript relies on two companion papers (Hlavacek-Larrondo et al. 2026 for the CND discovery, and Pereira et al. in preparation for H2 masses and line ratios) for key results. The SMBH mass and the kinematic-centre offset are the two main claims, and both are presented in the Abstract/Conclusions more strongly than the body text warrants. The abstract's phrase 'r_inf~60 pc, resolved by our observations' is particularly problematic given the acknowledged order-of-magnitude nature of the mass estimate and the unquantified astrometric errors. If the authors can add an uncertainty budget for the offset and a more careful qualification of the mass/resolved-SOI claims (or a forward model), the paper would be a valuable contribution. I would recommend that the editor require the authors to address the major comments before considering acceptance."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: the rotating CND is real and well shown; the 41 pc kinematic-centre offset is the interesting new claim but it is not yet supported to the level the paper presents it.\n\nWhat is new: this is the follow-up to the companion letter that announced the CND. The genuinely new bits here are the kinematic-centre estimate for the AGN, the gas-based SMBH mass, the resolved infall gradient, and the systematic comparison with MUSE, ALMA, and stellar kinematics. The data reduction is careful: LOKI, F-tests, wiggle correction, and WCS alignment to HST are described in enough detail to reproduce. The exercise of degrading the JWST data to MUSE resolution and showing that the disk disappears is a nice, honest demonstration. The paper also flags that the BH mass is order-of-magnitude; that is the right call.\n\nSoft spots. The stress-test note lands. The 41 pc offset is only ~2 spaxels, and the NIRSpec WCS is assigned by matching the HST F665N swirl to the Pa-alpha morphology, with no stars in the field. No uncertainty is quoted for that registration, nor for the gradient-weighted centroid. If the combined error is ~0.1-0.2 arcsec, the offset is consistent with zero. The authors present the offset as a firm reassessment of the AGN position in the abstract and conclusions; that overstates what the data currently support. The BH mass from a single-pixel dispersion and a virial estimator is fine as an order-of-magnitude estimate, but the phrase \"resolved SOI\" implies more than a 3-pixel diameter at the resolution limit can deliver. The inflow rate and gradient are clearly rough, and the paper says so. The ALMA non-detection is too weak to be a strong result on its own.\n\nWho this is for: people working on SMBH feeding in BCGs, and those planning NIRSpec IFU observations of nearby nuclei. The disk detection and the multiwavelength comparison are genuinely useful. The quantitative claims about offset and BH mass need a dynamical model or a real error budget before they should be taken as established.\n\nRecommendation: send it to peer review. The data and core detection are solid, the analysis is honest, and the astrometric concern can be fixed with an explicit error budget or a softer presentation. I would want the offset de-emphasized until it is nailed down.","headline":"The rotating disk is real and well presented; the 41 pc kinematic-centre offset is the new claim that needs an error budget before it is sold as a firm result.","tokens_in":40666,"tokens_out":2542,"would_cite":true,"duration_ms":25848,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"JWST finds a rotating disk around a billion-solar-mass black hole","keywords":["gas accretion","stellar kinematics","circumnuclear disk","supermassive black hole","JWST/NIRSpec","NGC 4696","Centaurus cluster","AGN feedback"],"falsifier":"A decisive check would be high-resolution radio imaging of the compact VLBA source: if the radio core aligns with the previously adopted AGN position rather than the kinematic centre, the 41 pc offset is likely an astrometric or projection artifact.","tokens_in":39338,"feed_emoji":"🌌","tokens_out":11674,"duration_ms":90938,"temperature":0.7,"pith_summary":"This paper uses JWST/NIRSpec integral-field spectroscopy to map gas and stellar motions in the central 620 parsecs of NGC 4696, the brightest galaxy in the Centaurus cluster. The authors claim to resolve a compact, rotation-dominated circumnuclear disk traced by hydrogen emission lines, and they use its kinematic centre to re-locate the supermassive black hole, which sits about 41 parsecs from the AGN position previously adopted from radio and HST data. From the gas velocity dispersion at that centre they derive a black hole mass of roughly $10^9$ solar masses with a sphere of influence of about 60 parsecs, marginally resolved by their 20-parsec pixels. They also trace inflowing filamentary gas accelerating toward the nucleus and estimate a lower bound on the accretion rate of about 18 solar masses per year. The result matters because it demonstrates that JWST can directly resolve the dynamical regime where gas falls into the gravitational grasp of a black hole in a nearby galaxy, and it suggests that the AGN position in NGC 4696 should be revised.","feed_headline":"JWST finds a rotating disk around a billion-solar-mass black hole","feed_subtitle":"In NGC 4696, the disk's kinematic centre shifts the AGN by 41 pc and resolves its sphere of influence.","key_machinery":"The central object is the circumnuclear disk (CND): a compact, rotation-dominated structure of radius roughly 120 pc traced by Pa$\\alpha$ and warm molecular hydrogen lines. The kinematic centre of this disk, located by the gradient-weighted centroid of the velocity field, serves as the anchor that re-defines the AGN position; the black hole mass is then estimated from the virial relation $M_{\\rm BH}\\sim\\sigma_{\\rm gas}^2 R/G$ using the peak gas dispersion and the pixel scale, and the sphere of influence follows from $r_{\\rm inf}=GM_{\\rm BH}/\\sigma_*^2$.","core_discovery":"The paper's central claim is that the kinematic centre of the circumnuclear disk, offset by roughly 41 pc from the previously adopted AGN position, is the true dynamical centre of NGC 4696, and that this disk carries a black hole mass of order $10^9\\,M_\\odot$ with a sphere of influence of about 60 pc that the JWST resolution marginally resolves. The evidence comes from Pa$\\alpha$ and H$_2$ 1$-$0 S(1) velocity fields showing a smooth blue- to redshifted rotation pattern, position–velocity diagrams with a velocity gradient of $4.7\\,\\mathrm{km\\,s^{-1}\\,pc^{-1}}$ over the inner 150 pc, a central Pa$\\alpha$ velocity dispersion of $449\\pm12$ km s$^{-1}$, and a rotation-to-dispersion ratio $v_c/\\sigma\\sim1.7$ indicating a settled rotating structure.","pith_inferences":["If the kinematic centre marks the true black hole position, the one-sided radio jet detected by VLBA should be re-registered to this new centre: a jet knot that tracks the old AGN position would reveal a systematic astrometric shift rather than a physical offset.","The method of using the full velocity-gradient field to locate the kinematic centre could be applied to other nearby brightest cluster galaxies with JWST/NIRSpec, potentially revising several AGN positions that currently rely on radio or narrow-band imaging.","A direct comparison of the measured $\\sim18\\,M_\\odot\\,{\\rm yr}^{-1}$ inflow with the Bondi-rate prediction of the hot phase would test whether the CND is fed by chaotic cold accretion or by the hot ICM, a distinction the paper leaves open.","Sub-parsec observations of the same disk, for instance with ALMA at higher resolution or future JWST high-contrast modes, could confirm whether the central dispersion peak is purely gravitational or broadened by an outflow component."],"forward_implications":["If the kinematic centre is the true AGN position, the AGN in NGC 4696 is offset by about 41 pc from the radio-core position, which will affect alignment between multi-wavelength observations and any dynamical modelling of the central engine.","A black hole mass near $10^9\\,M_\\odot$ with $r_{\\rm inf}\\sim60$ pc in a galaxy with stellar dispersion $\\sigma_*\\sim262$ km s$^{-1}$ places NGC 4696 on the $M$–$\\sigma$ relation, supporting the gas-dynamical method as a complement to scaling relations.","The measured inflow lower bound of $\\sim18\\,M_\\odot\\,{\\rm yr}^{-1}$ feeding the CND, compared with the much smaller accretion needed to power the radio source, indicates that most of the inflowing gas never reaches the black hole, consistent with the hidden-cooling-flow scenario.","The demonstration that co-spatial MUSE data fail to detect the disk due to seeing, while the JWST data resolve it, implies that ground-based IFU surveys of BCG cores are systematically missing the innermost accretion structures."],"supporting_citations":[{"why":"Supplies the HST H$\\alpha$ spiral/swirl morphology and the AGN position based on VLBA alignment that the paper re-assesses.","marker":"A. C. Fabian et al. (2016)"},{"why":"Provides the VLBA 5.0 GHz position of the radio core and the one-sided jet orientation used for the AGN position and disk inclination.","marker":"G. B. Taylor et al. (2006)"},{"why":"Companion letter that first reported the rotating circumnuclear disk in the same NIRSpec dataset and gives the morphological/kinematic framework this paper extends.","marker":"J. Hlavacek-Larrondo et al. (2026)"},{"why":"Supplies the MUSE stellar-velocity reference frame, the systemic redshift, and the hot-gas kinematics for comparison.","marker":"XRISM collaboration et al. (2025)"},{"why":"Provides the gradient-weighted centroid method used to define the kinematic centre.","marker":"B. García-Lorenzo et al. (2015)"},{"why":"Source of the stellar velocity dispersion $\\sigma_*\\sim262$ km s$^{-1}$ used in the sphere-of-influence estimate.","marker":"M. Bernardi et al. (2002)"},{"why":"Provides the larger-scale ALMA CO(1$-$0) and MUSE context that the paper compares with its CO(2$-$1) and NIRSpec data.","marker":"V. Olivares et al. (2019)"},{"why":"Analogous ALMA study of NGC 1275 that gives the comparison accretion rate and the CND velocity-twist signature.","marker":"T. Oosterloo et al. (2024)"}],"fun_headline_variants":["JWST resolves black hole's sphere of influence in NGC 4696","Billion-solar-mass black hole disk mapped by JWST in Centaurus","Kinematic centre of NGC 4696's disk shifts AGN by 41 pc","JWST pinpoints black hole feeding via 120-pc rotating disk","Sub-kiloparsec gas dynamics reveal SMBH feeding in NGC 4696"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The kinematic centre of the disk is assumed to coincide exactly with the black hole position, meaning the gas rotation is driven purely by the black hole's gravity with no significant outflow, inflow-stream, or projection contamination, and the astrometric alignment between JWST and HST is accurate to well below 41 parsecs.","fun_headline_variants_meta":{"raw":{"variants":["JWST resolves black hole's sphere of influence in NGC 4696","Billion-solar-mass black hole disk mapped by JWST in Centaurus","Kinematic centre of NGC 4696's disk shifts AGN by 41 pc","JWST pinpoints black hole feeding via 120-pc rotating disk","Sub-kiloparsec gas dynamics reveal SMBH feeding in NGC 4696"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000324,"raw_usage":{"total_tokens":1956,"prompt_tokens":1218,"completion_tokens":738,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":834,"completion_tokens_details":{"reasoning_tokens":635}},"tokens_in":834,"tokens_out":738,"duration_ms":6726,"temperature":1.0,"reasoning_tokens":635,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T00:21:15.471680+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A decisive check would be high-resolution radio imaging of the compact VLBA source: if the radio core aligns with the previously adopted AGN position rather than the kinematic centre, the 41 pc offset is likely an astrometric or projection artifact.","supporting_citations":[],"review_version":1}