{"id":"c80539b1-30ae-4d2a-911d-a0477dd8079b","arxiv_id":"2412.01922","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"A semi-analytic model predicts that TDE debris disks undergo repeated thermal-instability accretion cycles for up to about ten years, driving super-Eddington outflows and late-time optical/UV emission.","lead":"This paper models the disk of stellar debris left behind after a black hole shreds a star, and finds the disk can flicker between bright and dim accretion states for years. These cycles could explain the lingering ultraviolet glow, quasi-periodic eruptions, and delayed radio flares seen in some tidal disruption events.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The predicted cycles rest on the contested total-pressure alpha viscosity in Eq. (15): if the gas-pressure or magnetically stabilized prescription is correct, the central claim disappears. The reader's weakest assumption is the deciding issue.","rationale":"I read the paper in good faith. It is a careful semi-analytic study that reproduces prior work (Sec. 3), tests opacity and parameter choices (Sec. 4.3), and explicitly flags its main assumptions in the abstract, Sec. 2.4, and Sec. 6. The internal consistency is good; I found no algebraic or numerical red flag in the one-zone scheme. My stress-test question was whether the central phenomenon would survive the least secure physical input. It would not: the instability that drives the cycles is created by the total-pressure alpha viscosity in Eq. (15). The alternative gas-pressure prescription is not a straw man; it is adopted in published TDE-disk modeling and is motivated by the same radiation-MHD uncertainties the authors cite. Although the paper honestly presents this as a caveat, the caveat is the result. The one-zone coherent-cycling assumption is a secondary concern: a 1D treatment could change front propagation and timing, but the instability criterion is local and the one-zone approximation is standard for this problem, so I do not rank it above the viscosity assumption. I agree with the reader's weakest_assumption. Since the reader already returned CONDITIONAL and I find no reason to strengthen or weaken that, the verdict is unchanged.","tokens_in":20250,"tokens_out":9118,"duration_ms":102302,"concrete_test":"Re-run the fiducial calculation (M_BH = 10^6 M_sun, M* = 1 M_sun, beta = 1.85, alpha = 0.1, p = 0.5, OPAL opacities) with only the viscosity prescription changed from Eq. (15) to nu = alpha P_gas/(Omega rho), the Sakimoto-Coroniti/van Velzen prescription. If the repeated super-Eddington excursions in Fig. 8 disappear and Mdot(t) settles into the low state, the cycle prediction is demonstrated to hinge on the contested total-pressure alpha assumption, upgrading the paper's own caveat from a side note to the deciding factor.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central observable claim — repeated thermal-instability cycles beginning ~100 days and lasting ~10 yr, with super-Eddington outflows and radio flares — requires a local equilibrium curve with dT/dSigma < 0 (Fig. 3). That unstable branch exists only because Eq. (15) adopts nu = alpha P/(Omega rho) with P = P_g + P_rad, making the radiation-pressure-dominated, radiatively cooled disk thermally unstable. The authors are explicit that this is a choice: replacing P by P_g (Sakimoto & Coroniti 1981; van Velzen et al. 2019) removes the instability, and they cite 3D radiation-MHD work and Alush & Stone 2025 as evidence that magnetic fields or convection may stabilize disks in nature. This is not an internal inconsistency; the paper is transparent. It is nonetheless the load-bearing point: every quantitative prediction in Figs. 8-15 (cycle duration, Mflare, UV variability, radio timescales) is downstream of one microphysical assumption. If the real stress tracks gas pressure, or if magnetic pressure or fallback-stream heating stabilizes the disk, the high-state ejections and radio flares do not occur and the late-time UV is instead a quietly growing disk. Since the assumption is unresolved, the paper's central claim cannot yet be treated as robust.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a semi-analytic one-zone model for the long-term (years to decades) evolution of a TDE accretion disk fed by stellar fallback. The model solves for disk mass and angular momentum evolution using a Shakura-Sunyaev alpha-viscosity with total pressure, OPAL or analytic opacities, and an implicit thermal-equilibrium closure, including mass loss via a radius-dependent accretion rate. The central result is that radiation-pressure-dominated disks are thermally unstable and undergo repeated high/low accretion-state cycles beginning around 100 days after the TDE and lasting up to ~10 yr, with super-Eddington high states ejecting ~10^-3 to 10^-1 Msun at ~0.03-0.3c; low states produce late-time optical/UV luminosities within an order of magnitude of observed TDE plateau emission. The authors validate their code in the Shen & Matzner (2014) limit, explore variations of MBH, M*, alpha, and p, discuss stabilization by fallback heating and magnetic fields, and compare with late-time UV and radio observations qualitatively.","tokens_in":20547,"tokens_out":8921,"duration_ms":84355,"significance":"If the repeated-cycle picture is correct, it would provide a unified explanation for late-time UV variability, delayed radio flares, and state-dependent TDE disk behavior, making TDE disks a useful laboratory for radiation-pressure instability. Strengths of the paper are its transparency: the equations and time-stepping are clear, the comparison to Shen & Matzner (2014) is an external benchmark, the use of OPAL opacities is a concrete improvement, and the caveats are prominently acknowledged. However, the predictive content is heavily conditioned on an unresolved microphysical assumption—total-pressure alpha-viscosity with radiatively cooled, thermally unstable disks—and on the neglect of fallback-stream heating. The paper is therefore a well-executed parameter study of a hypothesized regime rather than a robust prediction; its significance will be settled by the sensitivity tests and targeted observations it proposes.","major_comments":[{"comment":"The central claim of the paper—that TDE disks repeatedly cycle between high and low accretion states—is produced by the choice nu = alpha P/(Omega rho) with P = P_g + P_rad in Eq. (15); this is what creates the dT/dSigma < 0 unstable branch in Fig. 3. The authors correctly note (Section 2.3 and Section 6) that using P = P_g (Sakimoto & Coroniti 1981; van Velzen et al. 2019) removes the instability and that magnetic stresses or fallback heating may stabilize the disk (Alush & Stone 2025). Because every quantitative prediction in Figures 8-15 is downstream of this choice, the paper needs a quantitative sensitivity study: for representative TDE parameters, recompute the evolution with a gas-pressure-only viscosity and with a prescription in which magnetic pressure stabilizes the disk, and report the fraction of parameter space in which cycles survive. The abstract and conclusions should then be phrased as explicitly conditional on this prescription rather than as a definite finding.","section":"Section 2.3, Eq. (15), Fig. 3"},{"comment":"Fallback-stream heating with eta of order unity removes the thermal instability at early times (Fig. 4), and the authors state in Section 2.4 that they have run time-dependent calculations with different heating levels but do not present them. This is a load-bearing omission because the claimed onset at ~100 days occurs early in the fallback epoch, when Mdot_fb is at its largest. I request that the time evolution be shown for representative values of eta (including the eta << 1 case motivated by Bonnerot et al. 2021) and that the interval of fallback-to-Eddington ratios over which cycles are quenched be quantified. Without this, the paper cannot claim that cycles are typical.","section":"Section 2.4, Eq. (19), Fig. 4"},{"comment":"The one-zone model assumes that all disk mass and angular momentum reside at a single radius R_d and that the disk is in instantaneous thermal equilibrium (Appendix A). The authors themselves note in Section 6 that a real disk would undergo an outside-in transition to the low state on a ~1 yr timescale, with implications for AT2018fyk and AT2021ehb. Because the predicted cycle phase, duration, and high-state mass ejections are all timed by the one-zone clock, the paper should either (i) validate the cycle epoch against a simple 1D diffusion calculation for a representative model, or (ii) explicitly state an uncertainty budget on cycle onset times and flare epochs arising from this approximation. As written, the comparison of cycle behavior to observations in Section 5 is not calibrated against this known systematic.","section":"Section 6 and Appendix A"},{"comment":"The comparison to late-time UV observations combines 12 different TDEs into a single panel and is not a fit; the model tracks are generated with hand-selected values of alpha, p, cos theta, M_BH, M_*, and beta. The text already cautions that Fig. 14 should not be viewed as a light curve, but the paper repeatedly refers to matching observations. I recommend making the qualitative nature of the comparison explicit in the abstract and conclusions, or replacing Fig. 14 with an event-by-event comparison (or a likelihood-style envelope test) so that the reader can judge whether the low-state luminosity is truly matched or merely plausible.","section":"Section 5.1, Fig. 14"}],"minor_comments":[{"comment":"The standard Shakura-Sunyaev effective-temperature profile has T_eff^4 proportional to 1 - (R_i/r)^{1/2}, not [1 - (R_i/r)]^{1/2}; please check the bracket and correct the SED normalization accordingly.","section":"Eq. (22)"},{"comment":"The text says high-state flares occur out to approximately 6 yr, while the abstract quotes cycles lasting up to ~10 yr; please reconcile these values explicitly.","section":"Section 4.2 and Fig. 9"},{"comment":"The validation against Shen & Matzner (2014) uses modified prescriptions (Sigma = 2 rho H, nu = 2 alpha c_s H/3, etc.); please state explicitly that this reproduces the previous work only in that comparison limit and does not by itself validate the main-model assumptions.","section":"Section 3"},{"comment":"The claim that the radio flare parameters are similar to those from Cendes et al. (2024) is not supported by a quantitative comparison; please add a table or explicit ranges of radii, densities, and synchrotron luminosities, or soften the statement.","section":"Section 5.2"},{"comment":"There is a typo in the phrase 'especially the the UV with missions like ULTRASAT'; remove the duplicated article.","section":"Section 6"},{"comment":"The convergence criterion epsilon < 10^-3 is stated without demonstrating convergence; adding a short convergence test for one representative model would help.","section":"Appendix A"}],"recommendation":"major_revision","confidential_remarks":"The paper is transparent and well-suited to the journal; my main concern is that the abstract's wording overstates the robustness of the cycle prediction. If the authors provide the requested sensitivity calculations and soften the wording accordingly, I would be satisfied with the revised version."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nRead Piro & Mockler on late-time TDE disk evolution. Quick take: this is a solid, transparent theory paper that maps out the parameter space for thermal-instability cycles in TDE disks, but the headline phenomenon is only as strong as the contested total-pressure alpha-viscosity assumption.\n\nWhat's actually new: the systematic survey across BH mass, stellar mass, alpha, and outflow parameter, plus the use of OPAL opacities. The iron-opacity bump produces a two-tiered low state that shows up in the UV light curves. They give concrete numbers: cycles begin ~100 days, last up to ~10 yr, high-state ejecta of 10^-3 to 10^-1 solar masses at 0.03-0.3c, radio flares from ejecta collisions. Those are testable predictions, and the paper is explicit about where they come from.\n\nThe paper does well: the equations are laid out, the comparison to Shen & Matzner 2014 checks out, and the numerical scheme is simple enough to follow. The authors also flag their own caveats clearly—fallback heating can quench the instability (Section 2.4), and magnetic fields may stabilize disks (Section 6). That is honest and should be credited.\n\nThe soft spot is the load-bearing assumption: Eq. (15) uses nu = alpha P/(Omega rho) with total pressure. That choice creates the dT/dSigma < 0 branch and all the cycles. Swap in gas-pressure-only viscosity, as van Velzen et al. do, and the cycles disappear. The authors know this and say so, but it means every quantitative prediction downstream of the cycles rests on an unsettled microphysical question. The comparisons to UV and radio observations are overlays, not fits, so they don't independently validate the mechanism. That's normal for this kind of paper, but worth keeping in mind.\n\nAlso, no code or data files are released, so reproducing the figures takes reimplementation. Minor issue.\n\nBottom line: if you want a comprehensive, self-aware map of what thermal-instability cycles would look like in TDE disks under standard alpha theory, this is the paper. It deserves a serious referee, and the right referee will push on the viscosity prescription and the one-zone assumption. I'd take it in a reading group and would cite it for the parameter survey and the OPAL opacity treatment.","headline":"A clear, honest semi-analytic survey of TDE disk thermal cycles; the physics is standard, the predictions are concrete, and the central caveat is openly acknowledged.","tokens_in":21064,"tokens_out":1566,"would_cite":true,"duration_ms":15197,"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":"The paper argues that debris disks left by tidal disruption events do not fade steadily: about 100 days after the disruption they become thermally unstable and cycle between super-Eddington flares and slowly re-brightening low states for…","keywords":["tidal disruption events","accretion disks","thermal instability","super-Eddington accretion","alpha-disk viscosity","late-time TDE emission","delayed radio flares","fallback accretion"],"falsifier":"Monitor one TDE from roughly $100$ days to $10$ years after disruption with UV/optical cadence of days. The model predicts repeated abrupt brightenings by more than an order of magnitude lasting one to two days, each followed by a slow rise over months to years. A light curve with no such repeated cycles, or with the first state transition arriving more than about a year after disruption for a $10^6\\,M_\\odot$ black hole, would contradict the model; so would a decade of radio monitoring with no delayed flares despite repeated predicted high states.","tokens_in":20055,"feed_emoji":"🌠","tokens_out":11410,"duration_ms":92870,"temperature":0.7,"pith_summary":"The paper argues that the debris disk left after a star is torn apart does not settle into a steady fading flow. Roughly $100$ days after the disruption, the disk becomes thermally unstable and then cycles between a short, super-Eddington high-accretion state and a long, dim low-accretion state for up to about ten years. In the high state the disk launches outflows of $\\sim10^{-3}$ to $10^{-1}\\,M_\\odot$ at speeds of $\\sim0.03$ to $0.3c$ over one to two days, and collisions between successive ejections could be the source of delayed radio flares. In the low state, continued fallback slowly rebuilds the disk and its optical/UV luminosity climbs to $\\sim10^{41}$ to $10^{42}\\,\\mathrm{erg\\,s^{-1}}$, matching late-time observations. If this picture is right, several long-standing puzzles of late-time TDE emission would be different manifestations of a single recurring disk instability.","feed_headline":"TDE disks cycle between faint and flaring for up to a decade","feed_subtitle":"Repeated super-Eddington flares hurl 0.001–0.1 solar masses outward and may drive delayed radio flares.","key_machinery":"The load-bearing mechanism is the thermal instability of a radiation-pressure-dominated, radiatively cooled accretion disk. The model is a one-zone disk: it tracks total disk mass and angular momentum at a characteristic radius $R_d$, with the accretion rate set by local energy balance $9\\nu\\Sigma\\Omega^2/8 = acT^4/(3\\kappa\\Sigma) + \\dot M c_s^2 \\xi/(2\\pi R_d^2)$, and uses the $\\alpha$-disk viscosity $\\nu=\\alpha P/(\\Omega\\rho)$ with the total pressure $P=P_g+P_r$. Along a sequence of equilibrium solutions, the branch with $dT/d\\Sigma<0$ is thermally unstable, so when the evolving disk crosses into that region it cycles counterclockwise in the $\\Sigma$–$T$ plane. The one-zone time evolution determines when the disk enters the unstable region and therefore sets the onset, duration, and recurrence of the flares.","core_discovery":"The paper's central claim is that a standard $\\alpha$-disk with viscosity proportional to the total (gas plus radiation) pressure, $\\nu=\\alpha P/(\\Omega\\rho)$, is thermally unstable in the radiation-pressure-dominated regime, and TDE disks necessarily enter this regime because fallback keeps feeding them for months to years. On the unstable branch, heating and cooling no longer balance stably: the disk jumps to a hot, super-Eddington state, drains its mass quickly, then drops to a cool low state where fallback slowly rebuilds it until the next jump. The authors find the cycles typically begin $\\sim100$ days after the TDE and continue for up to $\\sim10$ years, with the high states lasting only about one to two days and ejecting $10^{-3}$ to $10^{-1}\\,M_\\odot$ at $0.03$ to $0.3c$. Using OPAL opacities, the low state acquires a two-tiered structure from the iron-opacity bump, which helps it reach the UV/optical luminosities seen years after some TDEs. The authors also note that magnetic fields or heating from the fallback stream could stabilize the disk; they explore fallback heating and find it can suppress the instability while the fallback rate is high.","pith_inferences":["Beyond the paper: if the cycles are real, single snapshot UV observations are biased toward the brighter moments of the low-state rise, so the observed late-time UV sample may systematically overrepresent disks caught at high mass; high-cadence UV monitoring should catch the predicted dim phases.","Beyond the paper: the predicted ejections repeat with a waiting time set by the disk cycle, so radio light curves should show quasi-periodic flares rather than a single event; sparse radio sampling could miss most of them, making targeted radio follow-up after an optical/UV state transition a sharper test.","Beyond the paper: the same thermal-instability cycling should operate in any fallback-fed radiation-pressure-dominated disk, including disks formed by partial disruptions or by eccentric accretion onto lower-mass black holes; measuring whether such systems show similar episodic outflows would test the mechanism outside the original TDE setting.","Beyond the paper: a direct measurement of the high-state outflow mass and velocity from radio afterglow modeling would calibrate the model's ejection efficiency parameter $p$, which currently sets the mass-loss rate through Equation (10)."],"forward_implications":["Thermal-instability cycles begin roughly $100$ days after the TDE and can persist for up to about ten years, so late-time optical/UV emission from TDE disks should be strongly variable rather than steady.","Each high state is super-Eddington, lasts one to two days, and ejects $\\sim10^{-3}$ to $10^{-1}\\,M_\\odot$ at $\\sim0.03$ to $0.3c$; collisions among successive ejections are a plausible origin for the delayed radio flares observed in many TDEs.","In the low state the disk mass and accretion rate grow slowly over months to years, giving UV luminosities of $\\sim10^{41}$ to $10^{42}\\,\\mathrm{erg\\,s^{-1}}$; the same model predicts order-of-magnitude UV variations but only factor-of-a-few optical variations, roughly matching existing $g$- and $r$-band light curves.","Lower black hole masses cycle faster and eject less mass per flare, so the timing and energetics of late-time flares could provide a new way to estimate the black hole mass in a TDE.","Fallback-stream heating can stabilize the disk while the fallback rate is high, so the early evolution may be smooth in some events even though the later cycles still occur; whether cycles appear at all depends on unresolved disk physics such as magnetic-field stabilization."],"supporting_citations":[{"why":"Supplies the $\\alpha$-disk viscosity prescription $\\nu=\\alpha P/(\\Omega\\rho)$ used in Equation (15).","marker":"Shakura & Sunyaev 1973"},{"why":"Establishes the thermal-instability criterion for radiation-pressure-dominated, radiatively cooled disks.","marker":"Lightman & Eardley 1974"},{"why":"Companion analysis of the same thermal and viscous instabilities in thin accretion disks.","marker":"Shakura & Sunyaev 1976"},{"why":"Provides the OPAL Rosseland opacities, including the iron-opacity bump that shapes the low-state structure.","marker":"Iglesias & Rogers 1996"},{"why":"Supplies the numerical fallback accretion-rate curves that feed the disk in the time-evolution calculations.","marker":"Guillochon & Ramirez-Ruiz 2013"},{"why":"Previous one-zone TDE disk model used to validate the numerical scheme and to contrast with the higher-viscosity, higher-opacity models that produce cycles.","marker":"Shen & Matzner 2014"},{"why":"Updated TDE disk evolution work that the paper credits for recognizing that higher viscosity and opacity lead to accretion cycles.","marker":"Lu 2022"},{"why":"Compilation of late-time UV observations that the low-state disk models are matched against.","marker":"van Velzen et al. 2019"},{"why":"Delayed radio flare sample whose inferred ejecta parameters the high-state outflow model is intended to explain.","marker":"Cendes et al. 2024"},{"why":"Cited for the possibility that magnetic fields stabilize the disk, the main caveat to the instability cycles.","marker":"Alush & Stone 2025"},{"why":"Supports the assumption that stream self-crossing shocks dissipate much of the orbital energy away from the disk, so fallback heating may be inefficient.","marker":"Bonnerot et al. 2021"}],"fun_headline_variants":["TDE disks flip-flop for a decade in repeated flare cycles","Thermal instability makes TDE disks erupt repeatedly for years","TDE disks cycle between faint and flaring states for a decade","Disk instability drives decade-long cycle of faint and flaring TDEs","Repeated super-Eddington flares from TDE disks for up to a decade"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the disk viscosity follows the classical $\\alpha$-prescription with the total pressure, $\\nu=\\alpha P/(\\Omega\\rho)$; if magnetic fields, convection, or shock heating from the fallback stream stabilize the radiation-pressure-dominated disk, the high/low cycles, the ejections, and the associated radio flares would not occur.","fun_headline_variants_meta":{"raw":{"variants":["TDE disks flip-flop for a decade in repeated flare cycles","Thermal instability makes TDE disks erupt repeatedly for years","TDE disks cycle between faint and flaring states for a decade","Disk instability drives decade-long cycle of faint and flaring TDEs","Repeated super-Eddington flares from TDE disks for up to a decade"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000538,"raw_usage":{"total_tokens":2667,"prompt_tokens":1117,"completion_tokens":1550,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":733,"completion_tokens_details":{"reasoning_tokens":1457}},"tokens_in":733,"tokens_out":1550,"duration_ms":8973,"temperature":1.0,"reasoning_tokens":1457,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T00:04:29.424134+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Monitor one TDE from roughly $100$ days to $10$ years after disruption with UV/optical cadence of days. The model predicts repeated abrupt brightenings by more than an order of magnitude lasting one to two days, each followed by a slow rise over months to years. A light curve with no such repeated cycles, or with the first state transition arriving more than about a year after disruption for a $10^6\\,M_\\odot$ black hole, would contradict the model; so would a decade of radio monitoring with no delayed flares despite repeated predicted high states.","supporting_citations":[],"review_version":1}