{"id":"0f731d46-5cfd-44eb-994d-b6e3cf1017c4","arxiv_id":"2607.13833","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"In simulations, a Lennard-Jones FPU chain cooled by contact with a gas retains a residual energy E0 at zero temperature that scales approximately as E0 ∼ (ξN)^{2/3}.","lead":"Cooling a Fermi–Pasta–Ulam chain by coupling it to an ideal gas leaves the chain slightly hotter than the gas at low temperature, with a leftover \"zero-point\" energy. The leftover energy grows with system size and cooling rate, roughly as (cooling rate × size)^{2/3}, suggesting a simple model for glassy arrest.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"E0 is read at the end of a 90-step cooling run without a plateau test; if u(T) is still decreasing at the lowest T, the fitted exponents and Eq. (6) are artifacts of the stopping point.","rationale":"The reader's weakest assumption is exactly the plateau issue: E0 in Section III is taken as the lowest-temperature value without demonstrating that u has saturated. I agree this is the most load-bearing concern. If E0 is not the T→0 residual energy, the central claim Eq. (6) collapses, since every fitted exponent is computed from these E0 values. The paper acknowledges the definition is 'by a first approximation' and relies on visual inspection of slopes ('seems'), which is not a quantitative convergence test. The absence of error bars, code, and protocol details (e.g., the cooling-switch threshold) further prevents independent verification, but these are secondary to the plateau issue. The proposed test—holding the thermostat at the final temperature and observing whether u decreases—directly settles the concern. Since the reader already recommends CONDITIONAL and my analysis reinforces that recommendation rather than moving it, the verdict should remain UNCHANGED. I am not claiming the result is false; I am claiming the evidence currently does not establish that the measured E0 is a thermodynamic residual energy, making the scaling law conditional on a nontrivial saturation assumption.","tokens_in":7129,"tokens_out":5912,"duration_ms":68578,"concrete_test":"For N+2=200 and ξ=4×10^{-7}, rerun the full 90-step cooling protocol, then at the lowest reached temperature freeze the thermostat (stop the velocity rescaling, keep the gas at that measured T) and continue the integration for an additional 10×Δt. If the time-averaged specific energy u over the final hold decreases by more than the bin-averaged statistical error, E0 is not converged and Eq. (6) is unsupported. Repeat for the smallest and largest ξ used; this directly tests whether E0 is a plateau value rather than a stopping-point value.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In Section III, E0 is defined 'by a first approximation, as the value of u at the lowest temperature reached' in a protocol with kmax=90 cooling steps. The only evidence that this is the T→0 limit is the qualitative statement that individual orbits 'seem' to have vanishing slope at the lowest temperatures; no quantitative plateau test, no comparison of E0 with a hold at fixed T, and no extrapolation to T=0 are provided. This matters because the paper's own cited results [8–11] imply thermalization times diverge at least as u^{-2}; for the small specific energies reached near the end of cooling, 90 steps with fixed equilibration time Δt may terminate before u stops decreasing. If u would continue to drop with longer equilibration or further cooling, then every E0 used in Figures 5 and 6 is a stopping-time artifact, and the fitted exponents 0.61 and 0.66 (and hence Eq. (6)) are not estimates of a physical residual energy. This is the most load-bearing uncertainty because it precedes the scaling analysis: no plateau, no meaningful E0 to fit. Secondary concerns include the combination of two exponents that differ by more than the claimed 0.01 accuracy and the lack of error bars, but they are moot if the plateau assumption fails.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports numerical simulations of the cooling of a Fermi--Pasta--Ulam (FPU) chain coupled to an ideal-gas thermostat whose temperature is lowered at rate ξ. The authors find that, below the stochastic threshold T≈V0/27, the FPU specific energy u lies above the equilibrium value u=T and appears to approach a nonzero residual value E0 at the lowest temperatures reached. They then study E0 as a function of particle number N and cooling rate ξ, reporting power laws E0 ∼ N^{0.61} and E0 ∼ ξ^{0.66}, which they combine into the scaling law E0 ∼ (ξN)^{2/3} [Eq. (6)]. On this basis they argue that the residual energy survives in the thermodynamic limit unless the cooling rate vanishes faster than 1/N.","tokens_in":7501,"tokens_out":3444,"duration_ms":62750,"significance":"If the claimed scaling were established, it would provide a simple, concrete prediction for a glassy residual energy in a Hamiltonian model, and it would make a nontrivial connection between cooling rate, system size, and the breakdown of thermalization. The numerical setup is carefully described, including a Hamiltonian gas thermostat, symplectic integration, and two averaging procedures (moving average and bin average) over 128 orbits. However, the central quantitative claim rests on (i) the identification of E0 with the value of u at the end of a fixed 90-step cooling protocol without a demonstrated plateau, and (ii) power-law fits to very few points, without error bars or a joint fit. These issues must be resolved before the scaling law can be considered supported.","major_comments":[{"comment":"The load-bearing assumption is that the value of u at the lowest temperature of the 90-step cooling run equals the T→0 residual energy. The manuscript only states that individual orbits 'seem' to have vanishing slope at the lowest temperatures, with no quantitative plateau test, no runs with longer equilibration at fixed T, and no extrapolation to T=0. Since the paper's own cited results [8–11] imply thermalization times growing at least as u^{−2}, the final steps of the protocol may terminate before u stops decreasing. If u would continue to drop with longer equilibration, every E0 in Figs. 5–6 is a stopping-time artifact and the exponents in Eq. (6) are not physical. This must be tested directly, e.g. by comparing E0 for different kmax and by holding the temperature fixed at the lowest value for several equilibration times.","section":"Section III, definition of E0"},{"comment":"Equation (6) is obtained by declaring the two fitted exponents 0.61 and 0.66 equal. The authors state that both exponents have an accuracy of 0.01, but the difference 0.05 is five times that claimed accuracy. No statistical test is provided for the equality of the two exponents, and no combined fit of E0 against ξN is shown. To support Eq. (6), the authors should fit E0 as a function of ξN over the full data set (all N and ξ jointly) and show a collapse of the data, e.g. E0/(ξN)^{2/3} versus a relevant variable.","section":"Section III, Eq. (6)"},{"comment":"The power-law exponents are estimated from only three data points in each figure (N+2 = 100, 178, 316 for Fig. 5 and ξ = 4×10^{−7}, 1.265×10^{−6}, 4×10^{−6} for Fig. 6), with no error bars on E0 and no description of the fitting procedure. The claimed accuracy of 0.01 is therefore not substantiated. The authors should provide error estimates from the 128 orbits and the two averaging methods, and ideally additional intermediate system sizes and cooling rates. Without these, the exponents and hence Eq. (6) have unknown statistical uncertainty.","section":"Figures 5 and 6; Section III"},{"comment":"The cooling protocol does not implement a single uniform cooling rate: above a threshold the temperature is decreased by a fixed ΔT = 0.05 per step, while below the threshold the target temperature is 3/4 of the current temperature, so the rate ΔT/Δt is not constant over the run. The manuscript defines ξ = ΔT/Δt but does not specify at what stage or how ξ is computed for the data in Figs. 4 and 6. Since the ξ-dependence is one of the two pillars of Eq. (6), the cooling schedule and the effective ξ must be defined precisely, and the sensitivity of E0 to the threshold and to the two-regime protocol should be examined.","section":"Section II, definition of ξ"}],"minor_comments":[{"comment":"There is a typo 'whit' in the equation line; also the notation for the number of particles is inconsistent: the text says 'N+2 points, the first and last kept fixed' but Eq. (1) uses j=0,...,N+1 with q0 and qN+1 fixed, so there are N+1 moving particles and N+2 total sites. Please clarify the counting.","section":"Section II, Eq. (1)"},{"comment":"The caption lists N+2 = 100, 178, 316, but the text refers to these as 'the size of the FPU system'; it would help to state N explicitly as well.","section":"Section III, Fig. 3 caption"},{"comment":"The caption says 'Cooling rate is rescaled by a factor 10^{-7}', but the x-axis labels do not show the rescaling explicitly. Please make the axis labels self-explanatory.","section":"Section III, Fig. 6"},{"comment":"The term 'zero-point energy' may be confused with quantum zero-point energy; 'residual energy' or 'trapped energy' would be less ambiguous.","section":"Section III"},{"comment":"The interaction potential Vint in Eq. (4) is said to be chosen 'by trial and error' to ensure thermalization; this is plausible, but the values V1=100 and α=50 should be reported with units and a brief justification of their effect on the thermostat properties.","section":"Section II"}],"recommendation":"major_revision","confidential_remarks":"The manuscript's central claim is potentially interesting, but the unverified plateau assumption for E0 and the very sparse, error-free exponent fits are load-bearing. The work needs additional simulations (plateau checks, more N and ξ values) and a proper joint fit before it can support Eq. (6). I recommend major revision rather than rejection, because the required tests are within the scope of the numerical method and the paper's framing."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The one thing to know: this paper reports a new numerical scaling law, E0 ~ (ξN)^{2/3}, for the leftover specific energy in an FPU chain after it is cooled by contact with a gas. The observation that the residual energy grows with system size is new and, if it holds up, has a clean consequence: the effect survives the thermodynamic limit only if the cooling rate vanishes faster than 1/N. That is a compact, testable statement, and it is the paper's real contribution.\n\nWhat the paper does well: the model is simple and clearly described, the cooling protocol is explicit, and the authors are appropriately cautious in their language — they say the data \"seem to suggest\" the law, and they define E0 explicitly as a first approximation. The connection to earlier work on stochastic thresholds and divergent thermalization times is sound, and they do not oversell the physics. The numerical experiments are easy to understand, and the figures tell a coherent story.\n\nWhere the soft spots are: the load-bearing issue is the definition of E0 at the lowest temperature of a 90-step cooling run. The paper's own cited results imply thermalization times grow at least as u^{-2}, so it is entirely possible that u is still decreasing when the protocol stops. If that is the case, every E0 in Figures 5 and 6 is an artifact of the stopping point, and the fitted exponents are meaningless. The qualitative statement that the slope \"seems\" to vanish is not a substitute for a plateau test, a hold at fixed temperature, or an extrapolation to T=0. This needs to be fixed before the scaling law can be accepted.\n\nSecondary issues are minor by comparison but still matter: the two fitted exponents, 0.61 and 0.66, are each quoted with 0.01 accuracy, yet they differ by five times that; they are then combined into a 2/3 law. The fits use a narrow range of N and ξ, no error bars are shown, and the averaging methods differ between the two figures. The cooling-switch threshold and kernel width are not reported. These are all fixable in revision.\n\nWho this is for: anyone working on glassy dynamics, nonequilibrium statistical mechanics, or the FPU model. The paper is short, readable, and raises an interesting question even if the answer is not yet nailed down.\n\nMy recommendation: send it to review. The central claim is specific and falsifiable, the numerical setup is transparent enough to reproduce, and the plateau concern is exactly the kind of thing a referee can push on. With a saturation test and error bars, this could become a solid paper. Without them, it is a useful conjecture but not a result.","headline":"A new numerical scaling law for residual energy in cooled FPU chains, E0 ~ (ξN)^{2/3}, is worth taking seriously, but the stopping-time definition of E0 needs a plateau test before the exponents can be trusted.","tokens_in":7963,"tokens_out":1709,"would_cite":false,"duration_ms":33237,"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":"A chain of particles meant to model a solid keeps a nonzero 'zero-point energy' after cooling, scaling approximately as (ξN)^{2/3} with cooling rate ξ and particle number N.","keywords":["Fermi-Pasta-Ulam","cooling rate","zero-point energy","glassy behavior","stochastic threshold","thermodynamic limit","power-law scaling","non-equilibrium dynamics"],"falsifier":"Cool the same system to a lower final temperature (more steps, same ξ) or hold it at the lowest temperature for increasing times; if E0 continues to decrease toward u=T rather than reaching a plateau, the scaling law fails. Alternatively, test the combined law E0∼(ξN)^{2/3} by varying N and ξ while keeping the product ξN fixed — if the residual energy does not stay constant, the factorization into a single variable is wrong.","tokens_in":6971,"feed_emoji":"🧊","tokens_out":5001,"duration_ms":55123,"temperature":0.7,"pith_summary":"The paper asks what happens when a Fermi–Pasta–Ulam (FPU) chain — the simplest model of a crystal — is cooled by contact with a gas whose temperature is lowered at a finite rate ξ. Below a weak-stochasticity threshold, the chain's energy stops tracking the gas temperature, and at zero temperature it retains a residual 'zero-point energy' E0. Numerical simulations show that E0 grows with both the number of particles N and the cooling rate ξ, approximately as (ξN)^{2/3}. If this scaling survives in the thermodynamic limit, even a macroscopic solid could retain a measurable frozen-in energy unless the cooling rate vanishes faster than 1/N. The result is offered as a possible dynamical mechanism for glassy behavior in a Hamiltonian model.","feed_headline":"Residual energy grows as (ξN)^{2/3} in a cooled model solid","feed_subtitle":"A Fermi-Pasta-Ulam chain keeps a frozen-in energy that rises with both cooling rate and size.","key_machinery":"The FPU chain with Lennard-Jones nearest-neighbor interactions is coupled to a one-dimensional ideal gas that acts as a thermostat. Cooling is implemented in discrete steps: gas velocities are rescaled by η=√(T_new/T_old), followed by an equilibration interval Δt; the cooling rate is ξ=ΔT/Δt. Averages are taken over 128 orbits, using both a moving-average and a bin-average procedure. The quantity carrying the argument is the difference between the mechanically measured FPU specific energy and the equilibrium expectation u=T, whose low-temperature plateau defines E0.","core_discovery":"The central claim is that a finite cooling rate prevents a Fermi–Pasta–Ulam chain from fully thermalizing with its thermostat: below a stochasticity threshold, the specific energy u remains larger than the temperature T, and the difference persists as T→0. Defining the residual value E0 as the specific energy at the lowest temperature reached, the paper finds numerical power laws E0∼N^0.61 at fixed cooling rate and E0∼ξ^0.66 at fixed size, and suggests the combined law E0∼(ξN)^{2/3}. The authors state this as a possible law suggested by computations, not a proven result.","pith_inferences":["The product ξN suggests a per-particle cooling budget; a direct test would be to check whether data for different N and ξ collapse onto a single curve when plotted against ξN.","The exponent 2/3, if confirmed, may reflect the u^{-2} divergence of the FPU thermalization time; a simple balance between cooling lag and thermalization rate might predict an exponent and could be checked analytically.","One could test whether the same scaling appears for other anharmonic potentials or for other thermostat models; if it is universal, it would strengthen the analogy with supercooled liquids.","The protocol's dependence on the fixed temperature decrement ΔT=0.05 (above threshold) deserves scrutiny; varying ΔT separately from Δt could reveal whether E0 depends on the full path of the cooling schedule and not just on ξ."],"forward_implications":["If E0∼(ξN)^{2/3} holds, the residual energy per particle grows with system size, so the effect does not vanish in the thermodynamic limit unless ξ decays faster than 1/N.","A real solid coupled to a cooling gas could retain a measurable zero-point energy even at very low temperatures, affecting its thermodynamic response.","The 2/3 exponent links the size dependence and rate dependence, suggesting a single control parameter ξN for the glassy freezing.","The phenomenon provides a purely Hamiltonian, deterministic mechanism for glass-like behavior, without invoking disorder or activation barriers."],"fun_headline_variants":["FPU chain freezes with residual energy scaling as (ξN)^{2/3}","Frozen-in energy in FPU chain obeys (ξN)^{2/3} law","Cooled FPU system keeps energy that scales as (ξN)^{2/3}","Out-of-equilibrium FPU chain stores energy as (ξN)^{2/3}","Residual energy in cooled FPU chain grows like (ξN)^{2/3}"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The entire law rests on treating the specific energy at the lowest temperature reached in the 90-step cooling protocol as the true zero-temperature residual energy; if that energy would continue to drop with longer equilibration or further cooling, the measured exponents are artifacts of the stopping point.","fun_headline_variants_meta":{"raw":{"variants":["FPU chain freezes with residual energy scaling as (ξN)^{2/3}","Frozen-in energy in FPU chain obeys (ξN)^{2/3} law","Cooled FPU system keeps energy that scales as (ξN)^{2/3}","Out-of-equilibrium FPU chain stores energy as (ξN)^{2/3}","Residual energy in cooled FPU chain grows like (ξN)^{2/3}"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000867,"raw_usage":{"total_tokens":3561,"prompt_tokens":675,"completion_tokens":2886,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":419,"completion_tokens_details":{"reasoning_tokens":2783}},"tokens_in":419,"tokens_out":2886,"duration_ms":20836,"temperature":1.0,"reasoning_tokens":2783,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T03:33:24.155986+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Cool the same system to a lower final temperature (more steps, same ξ) or hold it at the lowest temperature for increasing times; if E0 continues to decrease toward u=T rather than reaching a plateau, the scaling law fails. Alternatively, test the combined law E0∼(ξN)^{2/3} by varying N and ξ while keeping the product ξN fixed — if the residual energy does not stay constant, the factorization into a single variable is wrong.","supporting_citations":[],"review_version":1}