{"id":"7f36d869-438a-40fb-be33-112d1c814d4f","arxiv_id":"2607.28905","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Thermal-conductivity response to targeted phonon excitation switches sign with normalized excitation frequency and Knudsen number, unifying enhancement in thin films with suppression in bulk.","lead":"This paper predicts that selectively exciting specific phonon frequencies can either enhance or suppress heat conduction in crystals, with the sign controlled by film thickness. The result is a frequency–Knudsen map that could guide active, switchable thermal management in electronics.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Frozen-population model may predetermine the sign reversal: fixing the excess α while self-consistently increasing scattering (Eqs. 10–17) ignores that a fixed pump would yield α∝1/(1+β), potentially erasing the nanofilm enhancement.","rationale":"The reader’s weakest assumption—the frozen-population approximation—is exactly the point I find most load-bearing. I agree that the paper is internally consistent and the first-principles Monte Carlo machinery is credible, but the central prediction’s observability hinges on whether a quasisteady spectrally selective population can be sustained. I sharpen this concern by pointing to the self-consistency inconsistency: the model fixes α and then computes β from the modified occupations, effectively assuming the external source supplies whatever power is needed. A fixed pump power would reduce the excess population of the very modes that supposedly cause enhancement, since their lifetimes shrink. This is not addressed by the paper’s acknowledgments of neglecting electrons, photons, or higher-order scattering; it is a more fundamental issue with the driving protocol itself. The proposed concrete test—a constant-pumping BTE iteration—would settle whether the positive nanofilm response survives under a more physical driving condition. I do not believe this warrants changing the reader’s CONDITIONAL verdict: the concern is real but not yet demonstrated to invalidate the claim, and the paper is transparent about its modeling assumptions. Secondary issues (unavailable SI parameters for the RSTM, lack of error bars) are worth noting but are less central than the driving self-consistency, which affects the physical relevance of the entire phase diagram. I also credit the paper for using first-principles scattering rates and for explicitly stating the frozen-profile limitation; that transparency is a positive feature, not a reason to reject.","tokens_in":13045,"tokens_out":9526,"duration_ms":107640,"concrete_test":"Implement a constant-pumping steady-state calculation for Si at 300 K, H=100 nm: add a generation term G_λ = G0 exp[-(ωλ−ωt)^2/(2σ^2)] to the phonon BTE and iterate n_λ^(k+1) = n0_λ + G_λ τ(n^(k))_λ to convergence, choosing G0 so the integrated excess energy equals ΔE3 = 2.0×10^7 J m^-3. Then recompute (κ−κ0)/κ0 from the converged n and the corresponding linearized scattering for ωt/2π = 1.2 THz and 10 THz. If the low-frequency enhancement at 1.2 THz remains positive, the frozen-profile idealization is not the source of the sign reversal; if it drops below zero or becomes negligible, the predicted phase map overstates the switching achievable under realistic constant-power excitation.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim—low-frequency excitation switches from suppressive in bulk to enhancing in nanofilms—rests on a prescribed frozen Gaussian population excess (Eq. 11) with amplitude A determined by the injected energy ΔE (Eq. 14). The lifetime τ_exc is then recomputed self-consistently from the modified Bose factors (Eqs. 15–16), yielding β>0. But maintaining a fixed α in the face of increased intrinsic scattering requires the external pump power to scale roughly as (1+β); a constant-power source would instead produce a steady-state excess α_ss ∝ G τ_exc = α0/(1+β). Since β is largest for modes in the target window, the very modes that drive the positive response lose population under fixed pumping. The frozen profile therefore assumes the pump compensates exactly for the scattering it induces—an assumption explicitly acknowledged in Methods/Discussion but never tested. If compensation is incomplete, the positive (κ−κ0)/κ0 in 100-nm films (Fig. 2d–f) could vanish or become negative, undermining the Knudsen-frequency phase diagram (Figs. 4a–c). This is the load-bearing linchpin because the paper’s novelty is the sign reversal, and its practical relevance depends on achievability with realistic driving.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes that the sign of the thermal-conductivity response to targeted phonon excitation is controlled by a Knudsen-number-dependent competition between increased transport weight of long-mean-free-path phonons and excitation-enhanced intrinsic three-phonon scattering. It uses DFT plus ShengBTE three-phonon rates and phonon-tracking Monte Carlo simulations for Ge, Si, and 3C–SiC in bulk and nanofilm geometries. The central finding is that bulk systems are predominantly suppressed, while 100-nm films show a positive response under low-frequency excitation and a negative response under high-frequency excitation. The results are organized in a (ω_t/ω_D, Kn) phase diagram, and a reduced spectral transport model (RSTM) is introduced to explain the common topology. The calculation prescribes a frozen Gaussian population excess (Eq. 11) and recomputes the scattering rates with modified Bose factors (Eqs. 15–16).","tokens_in":13388,"tokens_out":8633,"duration_ms":104896,"significance":"If the sign reversal is robust, the paper offers a generic design principle for active thermal switching that combines spectral selectivity with geometric confinement. A clear strength is that the first-principles Monte Carlo maps in Figs. 4(a)–4(c) are independent of the RSTM, so the core sign reversal is not an artifact of fitting the reduced model. The RSTM transparently identifies the physical competition and explains why an intermediate Knudsen regime should maximize enhancement. However, the practical relevance of the central claim depends on whether the prescribed population can actually be sustained under realistic driving, and the paper provides neither a self-consistent pump model nor numerical uncertainty estimates. The significance is therefore conditional on closing this gap.","major_comments":[{"comment":"The frozen-profile approximation fixes the Gaussian excess α while Eqs. (15)–(16) compute an enhanced scattering rate β. In a real experiment with a constant pump, the steady-state excess of the target modes would scale roughly as α_ss ∝ α0/(1+β), so the same modes that provide the positive transport weight are also the ones whose population excess is depleted by the scattering the model computes. This could reduce or even eliminate the positive (κ−κ0)/κ0 values in Figs. 2(d)–2(f) and shrink the enhancement regions in Figs. 4(a)–4(c). The Discussion acknowledges that excitation powers and dynamics are not modeled, but this is load-bearing for the central claim. The authors should either (i) test a constant-power approximation, e.g., α_ss = α0/(1+β), and show whether the sign map survives, or (ii) explicitly restrict the claim to a prescribed maintained population and state what frequency","section":"Methods, Eq. (11); Discussion"},{"comment":"No statistical error bars, convergence checks, or q-grid sensitivity are reported for the Monte Carlo results. Because the reported enhancements are as small as +5–10% for Ge and Si, the crossover boundaries in Figs. 4(a)–4(c) could shift or even disappear within numerical uncertainty. The authors should report the statistical uncertainty of the MC estimates and at least one convergence test (e.g., number of trajectories, q-grid density, or film-boundary sampling) to support the quantitative positioning of the sign-switching map.","section":"Results, Figs. 2–4"},{"comment":"The RSTM map in Fig. 4(d) is central to the claim of a common response topology, but the dimensionless spectral functions and parameter values are only stated to be in SI Appendix Sec. S5, which is not available in the main text. This makes the key model non-reproducible. The authors should either provide the explicit functions and values in the main text or make the SI available, and include a brief sensitivity analysis showing that the topology in Fig. 4(d) is not an artifact of a particular parameter choice.","section":"RSTM, Eq. (9); SI Sec. S5"}],"minor_comments":[{"comment":"The quantity ΔE is called 'injected energy density', but it is actually the excess energy density of the driven state relative to equilibrium. This naming could be confused with absorbed pump energy per unit volume; consider renaming it to 'excess energy density of the prescribed state'.","section":"Methods, Eq. (14)"},{"comment":"The caption states vertical dashed lines mark ⟨ℓ0⟩ and ℓ_max, but the main text says 'vertical dashed lines mark the average intrinsic mean free path ⟨ℓ0⟩ and maximum intrinsic mean free path ℓ_max'. Please ensure the notation is consistent and that the reader knows which line corresponds to which quantity.","section":"Fig. 3"},{"comment":"The discussion of previous predictions is brief. Given that the positive response in nanofilms is the paper's main novelty, a slightly more detailed comparison to Refs. [26]–[28] (especially the assumptions made there about the nonthermal population) would help the reader understand what is new.","section":"Introduction, Refs. [26]–[28]"}],"recommendation":"major_revision","confidential_remarks":"The frozen-profile concern raised in the stress-test is serious and should be addressed quantitatively before publication. An iterative or constant-power check within the MC framework would substantially strengthen the paper. Also, the hidden RSTM parameters in SI S5 are essential for evaluating the general-principle claim; without them the reduced model is not testable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nYou should know about this paper because it gives the clearest statement I've seen of when targeted phonon excitation helps or hurts thermal transport. The claim is that the sign of (κ−κ0)/κ0 is set by a Knudsen-controlled competition: in bulk, excitation-enhanced intrinsic scattering dominates and you get suppression; in nanofilms, low-frequency excitation boosts the quasi-ballistic contribution of long-MFP modes enough to turn the response positive. They demonstrate this with first-principles MC for Ge, Si, and 3C–SiC, and organize it in a (ω_t/ω_D, Kn) map. The MC maps are independent of the reduced model, so the core sign reversal isn't circular.\n\nWhat's genuinely new: prior work in this group showed low-frequency enhancement in graphene and hBN; this paper extends the competition to 3D semiconductors from bulk to films and puts it into a common phase-map topology, plus a minimal spectral model that explains why the sign flips. That's useful.\n\nSoft spots, in order. No error bars or convergence checks on the MC results—standard for this kind of paper but still worth asking. The RSTM parameters are in SI S5, which I can't see; the 'general principle' claim leans on that model, so the parameters need to be public. And the biggest one: the whole calculation prescribes a frozen Gaussian population excess (Eq. 11) and holds it fixed while the scattering rates self-consistently increase. The stress-test note is right that a constant-power pump would give an excess that shrinks as scattering grows, so the positive response in thin films could be weaker or vanish under realistic driving. The authors acknowledge the approximation and say they don't predict excitation powers or switching times. That's honest, but it does mean the practical reach of the map is untested. It doesn't sink the internal logic—as a statement about a maintained nonequilibrium population, the competition is real—but it should be flagged as the main caveat.\n\nI'd send this to review. The simulations are carefully set up, the map is plausible, and the limitations are stated rather than hidden. A serious referee would ask for the SI S5 parameters, error/convergence checks, and ideally a self-consistent pump model or an experimental benchmark. I'd probably cite this if I were working on active thermal switching, mainly for the phase-map idea.\n\nBest regards.","headline":"The ω_t–Knudsen sign-switching map is a real contribution, but the frozen-population assumption needs testing before taking it to the lab.","tokens_in":13885,"tokens_out":3383,"would_cite":true,"duration_ms":39205,"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":"This paper establishes that the sign of the thermal-conductivity response to targeted phonon excitation is set by a Knudsen-controlled competition between added transport weight and enhanced intrinsic scattering, and that this sign reverses","keywords":["thermal conductivity","phonon excitation","Knudsen number","nanofilm","phonon transport","three-phonon scattering","Monte Carlo simulation","active thermal control"],"falsifier":"Measure the thermal conductivity of a 100-nm Si film at 300 K under pump frequencies near 1.2 THz and near 10 THz at equal injected energy density; the theory predicts a positive response for the low-frequency pump and a negative one for the high-frequency pump. Observing suppression in both cases—or no sign reversal when sweeping frequency at fixed Kn—would refute the central claim.","tokens_in":12923,"feed_emoji":"🔥","tokens_out":8676,"duration_ms":76451,"temperature":0.7,"pith_summary":"Targeted phonon excitation—pumping a narrow frequency band of phonons—can either raise or lower a material's thermal conductivity. This paper establishes a general principle that decides which: a Knudsen-controlled competition. In bulk material (Kn≪1), the pumped phonons mostly amplify intrinsic scattering, so conductivity drops. In thin films (Kn≥1), low-frequency pumping adds transport weight to quasi-ballistic long-mean-free-path modes and conductivity rises, while high-frequency pumping stays suppressive. The authors derive this sign rule from first-principles three-phonon rates plus Monte Carlo transport for Ge, Si, and 3C-SiC, and show all three collapse onto one frequency–Knudsen map.","feed_headline":"Flip phonon pumping from heat suppressor to booster with thin films","feed_subtitle":"Low-frequency pumping can raise conduction in nanofilms; high-frequency pumping suppresses it.","key_machinery":"The Reduced Spectral Transport Model (RSTM) condenses the physics into the per-channel response R(ω;Kn,ωt) = [1+α]·(1+Kn·ℓ0/⟨ℓ0⟩)/(1+β+Kn·ℓ0/⟨ℓ0⟩) − 1, where α is the excitation-induced relative increase in modal transport weight and β the relative increase in three-phonon scattering rate; intrinsic and boundary scattering combine via Matthiessen's rule. The Knudsen number appears in the (1+Kn·ℓ0/⟨ℓ0⟩) factors, making boundary scattering dilute the impact of β for long-MFP modes, which is the mechanism that turns the sign. The paper couples this model with phonon-tracking Monte Carlo simulations driven by first-principles three-phonon rates with modified Bose occupation factors.","core_discovery":"The paper's central claim is that geometric confinement, quantified by the Knudsen number Kn=⟨ℓ0⟩/H, reverses the sign of the relative thermal-conductivity change (κ−κ0)/κ0 under spectrally selective phonon excitation. Exciting low-frequency phonons in a nanofilm (Kn≥1) increases the transport weight of long-mean-free-path modes that remain quasi-ballistic, producing positive response up to +20% for 3C-SiC at 100 nm; exciting high-frequency phonons or working in bulk (Kn≪1) gives predominantly negative response because enhanced three-phonon scattering shortens lifetimes. The sign-switching is organized in a universal (ωt/ωD, Kn) phase map, and a reduced spectral transport model with Eq. (8)","pith_inferences":["The frozen-Gaussian population ansatz (Eq. 11) suggests the mechanism is not tied to a specific excitation pathway; any source that maintains a narrow spectral hole—optical, terahertz, or polariton-mediated—should produce the same map, so the prediction may extend to pump sources beyond those cited.","A sharper test than bulk-vs-film alone: the crossover frequency ωt/ωD at which response changes sign should increase with Kn (thinner films push the positive window to higher normalized frequencies); mapping this shift would confirm the Knudsen mechanism against simple MFP shortening.","Because the model omits temperature redistribution and higher-order scattering, the predicted enhancement is most likely to appear in time-resolved or transient measurements where local heating has not yet built up; steady-state experiments may see a muted version.","The common topology implies a design rule for active phononic switches: choose a material with a high Debye frequency (like SiC) to widen the enhancement window in normalized units, and tune thickness to sit at the intermediate-Kn peak."],"forward_implications":["In films with thickness near the average intrinsic mean free path, low-frequency excitation produces the largest enhancement—up to about +20% in 3C-SiC at 100 nm and 300 K.","High-frequency excitation is predominantly suppressive across confinement regimes, approaching roughly −30% in the bulk limit for Ge and Si.","The response map in the (ωt/ωD, Kn) plane is common to Ge, Si, and 3C-SiC, meaning the sign rule transfers across materials with very different phonon spectra.","Sweeping either the pump frequency or the film thickness across the predicted crossover should reverse the sign of the thermal-conductivity response—a direct, testable prediction.","The RSTM shows that the total response is the conductivity-weighted integral over all spectral channels, so the sign is set by the full spectrum, not by the excited mode alone."],"fun_headline_variants":["Thin films flip phonon excitation from suppressor to booster","Knudsen number decides if phonon pumping heats or cools","Low-frequency phonon pump boosts heat flow in nanofilms","Frequency-Knudsen map predicts sign of heat response to phonon pumping","Phonon excitation: suppress in bulk, boost in thin films"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The prediction rests on the assumption that a narrow, steady Gaussian phonon population excess (Eq. 11) can be maintained while heat current is probed, with only three-phonon Bose occupation factors modified; if real excitation sources cause strong temperature redistribution, higher-order scattering, or electron/phonon coupling, the sign-switching map may not be observable.","fun_headline_variants_meta":{"raw":{"variants":["Thin films flip phonon excitation from suppressor to booster","Knudsen number decides if phonon pumping heats or cools","Low-frequency phonon pump boosts heat flow in nanofilms","Frequency-Knudsen map predicts sign of heat response to phonon pumping","Phonon excitation: suppress in bulk, boost in thin films"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000187,"raw_usage":{"total_tokens":1178,"prompt_tokens":766,"completion_tokens":412,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":510,"completion_tokens_details":{"reasoning_tokens":322}},"tokens_in":510,"tokens_out":412,"duration_ms":4779,"temperature":1.0,"reasoning_tokens":322,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T01:19:25.163559+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the thermal conductivity of a 100-nm Si film at 300 K under pump frequencies near 1.2 THz and near 10 THz at equal injected energy density; the theory predicts a positive response for the low-frequency pump and a negative one for the high-frequency pump. Observing suppression in both cases—or no sign reversal when sweeping frequency at fixed Kn—would refute the central claim.","supporting_citations":[],"review_version":1}