{"id":"e8ea2d33-170b-4e6d-be23-dc03a99703fb","arxiv_id":"2607.05296","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":1,"one_line_summary":"First-principles Monte Carlo simulations show that phonon scattering selection rules and boundary mode conversion compete in semiconductor nanofilms, producing transverse-acoustic-polarized heat currents at temperatures below 100 K, most strongly in InP.","lead":"This paper predicts that nanoscale semiconductor films can polarize heat-carrying phonons by mode, creating transverse-acoustic-dominated heat currents at cryogenic temperatures. A smart generalist might read it to understand how nanostructuring could tune which phonons carry heat, with implications for hot-carrier lifetimes and defect-insensitive thermal management in electronics.","discovery_kind":"unclear","skeptic_critique":{"model":"glm-5.2","headline":"The isotropic projection is the right concern, but the deeper issue is whether it distorts the D_λ·v_λ ratios that drive the mode-conversion asymmetry — the paper never checks this directly.","rationale":"The reader correctly identified the isotropic projection as the most load-bearing assumption. My analysis confirms this: the entire quantitative claim flows through Eq. 2, where the D_λ·v_λ ratios and the Snell's law geometric factor both depend on the isotropic projection preserving the correct relative magnitudes across polarizations. The paper does not validate this preservation directly. However, I recommend UNCHANGED rather than a harsher verdict because: (1) the isotropic projection for cubic crystals has precedent in thin-film transport calculations (Refs. [16-19]), (2) the qualitative mechanism — LA phonons have lower D_λ·v_λ than TA in InP due to velocity separation — is physically robust and unlikely to reverse under anisotropic treatment, (3) the paper correctly frames the result as a prediction without claiming experimental validation, and (4) the diffuse-scattering validation against Fuchs-Sondheimer theory provides a useful internal consistency check on the VRMC framework itself. The concern is real and should be addressed in future work, but it does not rise to the level of invalidating the central claim given the physical plausibility of the mechanism and the established nature of the computational framework. The reader's CONDITIONAL verdict with MODERATE confidence appropriately reflects this state of affairs.","tokens_in":8752,"tokens_out":804,"duration_ms":199312,"concrete_test":"Recompute the mode-conversion probabilities in Eq. 2 using the full anisotropic D_λ(q) and v_λ(q) for InP along at least two high-symmetry directions (Γ→X and Γ→L), without isotropic projection. Compare the resulting D_λ·v_λ ratios (LA vs. TA) and the critical angle θ_c against the isotropic values. If the anisotropic LA/TA ratio changes by more than ~20% relative to the isotropic one, re-run the VRMC simulation for d=50 nm InP and check whether κ/κ_b shifts by more than 5 percentage points from the reported ~50%.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim — that LA phonons are preferentially depopulated by boundary mode conversion in InP nanofilms — depends on the ratio D(ν,TA)·v(ν,TA) / D(ν,LA)·v(ν,LA) being large, as shown in Fig. 3. This ratio enters the mode-conversion probabilities in Eq. 2 and determines the asymmetry between LA→TA and TA→LA conversion. The paper computes D_λ and v_λ from first-principles but then projects them into an isotropic form (paragraph after Eq. 1, citing Ref. [16]) before they enter Eq. 2. The isotropic projection averages over direction, and if the angular averaging changes the LA/TA ratio of D_λ·v_λ differently in InP vs. BP — or differently than the actual directional values along the transport direction — then the predicted ~50% κ suppression in InP could be an artifact. The paper provides no direct comparison between the isotropic-projected D_λ·v_λ ratios and the full anisotropic ones. Additionally, the Snell's law geometric factor uses q(ν,p) magnitudes that are also direction-dependent in real crystals; the isotropic projection assigns a single q per (ν,p), which could distort the critical angle θ_c for TA→LA conversion. Since the entire quantitative claim (κ/κ_b ≈ 0.5 at d=10-100 nm) flows from these ratios, this is the load-bearing assumption.","agreement_with_reader":"agree"},"referee_report":{"model":"glm-5.2","summary":"This manuscript predicts that phonon scattering selection rules and boundary mode conversion compete in nanoscale semiconductor films to produce mode-polarized heat currents at cryogenic temperatures (≤100 K). Using first-principles phonon properties and a variance-reduced Monte Carlo (VRMC) solution of the steady-state Peierls-Boltzmann equation, the authors show that in InP nanofilms, the AAA # 2 selection rule amplifies LA phonon contributions to bulk κ, while boundary mode conversion at specular surfaces preferentially depopulates LA phonons, yielding a TA-polarized heat current with κ suppressed to ~50% of bulk at 10–100 nm. The effect is weaker in BP, where TA phonons dominate. The relaxation time approximation is justified by comparison to full iterative solutions (within ~12%). The central physical mechanism — that the D_λ·v_λ ratio and Snell's law geometric factor both favor LA→TA conversion — is clearly articulated and supported by Figs. 1–3.","tokens_in":9416,"tokens_out":1406,"duration_ms":151963,"significance":"The paper identifies a previously unexplored mechanism for generating mode-polarized thermal phonon currents in nanoscale films, with potential implications for symmetry-selective phonon engineering. Strengths include: (1) no fitted free parameters — all phonon properties derive from first-principles DFT calculations (Ref. [2]); (2) the RTA is validated against full iterative bulk solutions; (3) the VRMC method is well-established; (4) the prediction is falsifiable — the κ/κ_b ratio and mode polarization are specific quantitative predictions testable in InP films with controlled surface specularity. The claim that the effect is material-dependent (InP vs. BP) and temperature-dependent (Fig. 4) adds further testability. However, the significance of the quantitative predictions rests on the isotropic projection approximation, which is not directly validated for the specific ratios that drive the mode-conversion asymmetry.","major_comments":[{"comment":"Paragraph following Eq. 1: The isotropic projection of wave-vector-dependent group velocities and relaxation times (citing Ref. [16]) is the load-bearing approximation for the central claim. The mode-conversion probabilities in Eq. 2 depend on the ratio D(ν,TA)·v(ν,TA) / D(ν,LA)·v(ν,LA), and the Snell's law critical angle depends on q(ν,TA)/q(ν,LA). If the isotropic projection distorts these ratios differently for InP vs. BP — or differently from the directional values along the transport direction — the predicted ~50% κ suppression in InP could be an artifact. The manuscript provides no direct comparison between isotropic-projected and full anisotropic D_λ·v_λ ratios. A figure or table comparing the projected vs. anisotropic ratios for both materials, at least along representative high-symmetry directions, would substantially strengthen the claim. Without this, the reader cannot assesss","section":null},{"comment":"Eq. 2 and the Snell's law geometric factor: The mode-conversion model assumes elastic scattering (fixed ν) and flux conservation. The manuscript states that evanescent/surface LA fields for θ_in > θ_c are localized to ~nm-thick regions and do not contribute to heat current, but this is stated without calculation or citation. For THz phonons at 100 K, the penetration depth and energy storage in these evanescent modes could be non-negligible relative to a 10 nm film. A brief estimate or reference supporting the 'localized to ~nm' claim would strengthen this argument, particularly since the smallest film thicknesses studied (d=10 nm) are comparable to this length scale.","section":null},{"comment":"The T-independent small-d limit of κ/κ_b (Fig. 4 and accompanying text) is attributed to T-independent relative LA/TA contributions to κ_b (~40% LA, ~60% TA) between 50–100 K. However, this explanation is qualitative. A more direct demonstration — e.g., showing that the mode-converted phonon distribution from Eq. 2 is indeed T-independent when expressed in terms of the normalized spectral contributions — would make this claim more convincing. As stated, the reader must take the T-independence on faith rather than seeing it derived.","section":null}],"minor_comments":[{"comment":"Fig. 2(a): The inset schematic showing incidence/reflection angles is small and difficult to read. Enlarging or separating it into its own panel would improve clarity.","section":null},{"comment":"The phrase 'first-principles calculations' is used throughout, but the phonon relaxation times are taken from Ref. [2] (Ravichandran & Broido, PRX 2020). While this is legitimate, a brief note clarifying that the first-principles inputs are from prior published DFT work, with the present contribution being the VRMC simulation and mode-conversion model, would help the reader.","section":null},{"comment":"The discussion of evanescent modes (end of the paragraph containing the Snell's law analysis) would benefit from a quantitative estimate or a reference. As written, the claim that these modes are 'localized to ~nm-thick regions' is unsupported.","section":null},{"comment":"Supplementary figures are referenced (S1, S2, S3) but not provided in the main text. While standard, ensuring these are included in review materials is essential for full assessment.","section":null},{"comment":"The abstract states 'cryogenic temperatures (≤100 K).' The choice of 100 K as the upper bound is reasonable but could be briefly justified — e.g., noting that above this temperature, intrinsic scattering rates increase and mask the selection-rule effect.","section":null},{"comment":"Ref. [9] appears to be a 2026 Advanced Materials article; if this is an in-press or preprint reference, the authors should verify publication status and update the citation.","section":null}],"recommendation":"major_revision","confidential_remarks":"The reader's concern about the isotropic projection is well-placed and is the primary reason for the major_revision recommendation. The concern is not that the approximation is invalid per se — it has precedent for cubic crystals — but that the specific quantity driving the central claim (the D_λ·v_λ ratio between LA and TA branches) has not been shown to survive the projection faithfully. This is a quantitative concern, not a qualitative one; the physical mechanism is sound and the paper's core idea is interesting and novel. I note that one of the authors (Ravichandran) is a co-author on Ref. [2], from which the phonon scattering rates are taken; this is a legitimate use of published first-principles data and not a circularity concern, but the authors should be transparent about this relationship in the text."},"author_rebuttal":{"model":"glm-5.2","summary":"We thank the referee for a careful and constructive report. The referee correctly identifies that the isotropic projection approximation is load-bearing for our central quantitative claim, and we agree that direct validation of this approximation for the specific ratios driving mode-conversion asymmetry is needed. We will add a comparison between isotropic-projected and full anisotropic D_λ·v_λ ratios for both materials. We will also add a quantitative estimate of evanescent mode penetration depth and a more direct demonstration of the T-independence of the small-d limit. One comment regarding the T-independence derivation we address with a partial revision, as the full analytical derivation is complex but a numerical demonstration is feasible and will be added.","responses":[{"response":"The referee is correct that the isotropic projection approximation is load-bearing for the mode-conversion probabilities in Eq. 2, and that direct validation of this approximation for the specific D_λ·v_λ ratios is absent from the current manuscript. We agree this is a gap. In the revised manuscript, we will add a figure or table comparing the isotropic-projected D_λ·v_λ ratios against full anisotropic calculations along representative high-symmetry directions (Γ→X, Γ→K, Γ→L) for both InP and BP. This will allow readers to directly assess whether the isotropic projection distorts the LA/TA contrast differently between the two materials. We note that the isotropic projection has been previously validated for predicting κ in thin films of cubic crystals (Refs. [17–19]), and the RTA bulk κ values we report agree with full iterative solutions to within ~12%, which provides indirect evidence that the projection does not severely distort the relevant phonon properties. However, we agree that the mode-conversion ratios specifically warrant direct validation, and we will provide it.","revision_made":"yes","referee_comment":"Paragraph following Eq. 1: The isotropic projection of wave-vector-dependent group velocities and relaxation times is the load-bearing approximation. No direct comparison between isotropic-projected and full anisotropic D_λ·v_λ ratios is provided."},{"response":"The referee raises a valid point. The statement about evanescent mode localization is currently made without supporting calculation. We will add a brief quantitative estimate in the revised manuscript. The penetration depth of an evanescent LA field can be estimated as δ = 1/√(q_LA² - q_parallel²), where q_parallel = q_TA sin θ_in is the conserved parallel momentum component and q_LA is the LA wave vector magnitude at the same frequency. For THz phonons in InP (q ~ 0.1–1 nm⁻¹), when θ_in modestly exceeds θ_c, the evanescent decay length is on the order of ~1–5 nm. We will include this estimate explicitly and cite relevant literature on evanescent phonon fields at interfaces. We acknowledge that for the smallest film thicknesses studied (d = 10 nm), this length scale is not entirely negligible, and we will discuss this caveat: the fraction of phonons with θ_in significantly exceeding θ_c is small, and the energy stored in evanescent modes is further reduced by the D_λ·v_λ weighting, but the limitation at d = 10 nm will be noted honestly.","revision_made":"yes","referee_comment":"Eq. 2 and Snell's law: The claim that evanescent/surface LA fields for θ_in > θ_c are localized to ~nm-thick regions is stated without calculation or citation. For THz phonons at 100 K in a 10 nm film, penetration depth could be non-negligible."},{"response":"The referee is right that the current explanation is qualitative and that a more direct demonstration would strengthen the claim. The T-independence of the small-d limit follows from two facts: (1) at small d, the phonon distribution is governed entirely by the mode-conversion distribution from Eq. 2, which is T-independent when expressed in terms of normalized spectral contributions, and (2) the relative LA/TA contributions to κ_b are approximately T-independent between 50–100 K. In the revised manuscript, we will add a figure showing the mode-converted phonon distribution from Eq. 2 expressed in terms of normalized spectral contributions, demonstrating its T-independence directly. We will also show the LA and TA fractional contributions to κ_b as a function of T to make the second point quantitative rather than qualitative. We note that a full analytical derivation of the T-independence from Eq. 2 is non-trivial because the mode-conversion distribution depends on the equilibrium energy distribution e^d_λ, which is T-dependent; the T-independence of the normalized ratios emerges from the cancellation of T-dependent factors in the ratio D_λ v_λ e^d_λ / Σ_p D_λ v_λ e^d_λ. We will include this analytical argument alongside the numerical demonstration.","revision_made":"partial","referee_comment":"The T-independent small-d limit of κ/κ_b is attributed to T-independent relative LA/TA contributions to κ_b, but this explanation is qualitative. A more direct demonstration is needed."}],"tokens_in":8570,"tokens_out":1080,"duration_ms":96455,"standing_objections":[]},"desk_editor":{"model":"glm-5.2","letter":"The main thing to know: this paper identifies a genuinely new physical mechanism — the competition between intrinsic phonon scattering selection rules and extrinsic boundary mode conversion in nanofilms — and predicts it produces TA-polarized heat currents in materials like InP. That is a real insight, not a recombination of existing tools. The VRMC framework is well-established, the relaxation times come from published first-principles DFT calculations (Ravichandran & Broido, PRX 2020), and there are no fitted parameters. The RTA is validated against full iterative solutions (bulk κ within ~12% for both BP and InP at 100 K). The Fuchs-Sondheimer agreement for diffuse boundary scattering is a useful sanity check. The physical story is clean and well-told: in InP, the AAA #2 selection rule amplifies LA phonon contributions to bulk κ, but boundary mode conversion preferentially depopulates LA modes because D_λ·v_λ is much larger for TA than LA, producing a TA-polarized current with ~50% κ suppression at 10-100 nm. The Snell's law critical-angle argument for why TA→LA conversion is geometrically restricted adds a second, independent mechanism favoring LA depopulation. This is solid physical reasoning. The soft spot is real but bounded. The isotropic projection (Hua & Minnich, PRB 2014) collapses direction-dependent group velocities and relaxation times into a single isotropic form before they enter the mode-conversion probabilities in Eq. 2. The reader and stress-test both flag this correctly: the entire quantitative claim flows from the D_λ·v_λ ratio between LA and TA branches, and the angular averaging could distort this ratio differently than the actual directional values along the transport direction. The Snell's law geometric factor also uses direction-dependent q(ν,p) magnitudes that get collapsed to a single value per branch. The paper never directly compares the isotropic-projected ratios against full anisotropic ones for either material. That said, the qualitative mechanism — LA modes preferentially convert to TA modes because TA branches have higher DOS and the velocity contrast is larger in InP than BP — is robust to the projection. The isotropic approximation would have to distort the LA/TA ratio of D_λ·v_λ qualitatively (not just quantitatively) to overturn the result, which seems unlikely given that the velocity contrast is a bulk property visible directly in the dispersions. So the concern is primarily about whether the ~50% suppression number is right, not whether the effect exists. This paper is for researchers working on nanoscale thermal transport, phonon polarization engineering, and hot-carrier physics. It deserves a serious referee who can assess whether an anisotropic spot-check for at least one material is feasible and whether the elastic boundary scattering assumption is justified at 100 K.","headline":"New mechanism for mode-polarized heat currents in nanofilms via competition between selection rules and boundary mode conversion; isotropic projection is the load-bearing approximation that needs checking","tokens_in":9712,"tokens_out":670,"would_cite":true,"duration_ms":68315,"reading_group":"no","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"glm-5.2","headline":"Nanofilms polarize heat into transverse phonons, halving InP conductivity","keywords":[],"falsifier":"Measure the in-plane thermal conductivity of 10–100 nm InP films at 100 K with specular boundaries. If κ does not drop to approximately 50% of bulk, or if the spectral contribution of LA phonons is not preferentially suppressed relative to TA, the mode-conversion polarization mechanism is not operating as predicted. Alternatively, if polarized phonon spectroscopy (e.g., inelastic neutron or Raman-based probes adapted for thin films) shows no TA-polarized non-equilibrium phonon population in the films, the central claim is falsified.","tokens_in":8872,"feed_emoji":"🔥","tokens_out":1385,"duration_ms":95668,"temperature":0.7,"pith_summary":"This paper argues that two mechanisms known to govern phonon heat transport in bulk semiconductors — intrinsic three-phonon scattering selection rules and extrinsic boundary mode conversion — compete directly when a semiconductor is made thin enough (10–500 nm), and that this competition produces a heat current dominated by one phonon polarization. In bulk indium phosphide (InP) at cryogenic temperatures (≤100 K), a selection rule called AAA # 2 (arising from the large velocity gap between longitudinal acoustic (LA) and transverse acoustic (TA) branches) weakens the intrinsic scattering of LA phonons, making them the dominant heat carriers. But when those same LA phonons strike the boundaries of a nanofilm, they convert into TA modes with high probability, while the reverse TA→LA conversion is suppressed by a Snell's-law critical-angle effect. The net result is that the LA contribution is drained away, the heat current becomes TA-polarized, and the thermal conductivity drops to roughly 50% of the bulk value for 10–100 nm InP films at 100 K under specular boundaries with mode conversion. The paper establishes this using first-principles variance-reduced Monte Carlo solutions of the steady-state Peierls-Boltzmann equation, comparing InP (where the LA/TA velocity gap is large) against boron phosphide (BP, where acoustic branches are bunched and the effect is weak). The central object carrying the argument is the mode-conversion probability at the boundary, which factorises into a material term D(ν,p)v(ν,p) (density of states times group velocity, evaluated per polarization at fixed frequency) and a geometric term cos θ_out set by Snell's law. Both terms favour LA→TA over TA→LA, and both are stronger in InP than in BP because the larger velocity separation in InP amplifies the contrast between polarizations. The paper also shows a size effect: as film thickness shrinks, boundary mode conversion precedes intrinsic scattering for an increasing fraction of phonons, causing κ/κ_b to saturate at a thickness-independent, temperature-independent limit set by the relative LA and TA contributions to bulk κ.","feed_headline":"Nanofilms polarize heat into transverse phonons, halving InP conductivity","feed_subtitle":"Boundary mode conversion drains longitudinal phonons in 10–100 nm InP films, producing transverse-polarized heat currents at 100 K.","key_machinery":"The argument rests on three linked components. First, the AAA # 2 selection rule: in materials where LA and TA dispersions are well separated (InP, InAs, InSb), three-phonon all-acoustic scattering of low-frequency LA phonons is suppressed, amplifying their bulk κ contribution. Second, the boundary mode-conversion probability [Eq. 2]: elastic reflection at a film boundary redistributes phonons across polarizations at fixed frequency, with relative probabilities proportional to D(ν,p')v_⊥(ν,p'), where D is the density of states and v_⊥ is the group velocity component normal to the boundary. Because D_λ v_λ is larger for TA than LA, LA→TA conversion is favoured and TA→LA is disfavoured. Third,","core_discovery":"The paper's central discovery is that phonon scattering selection rules, which amplify LA phonon heat transport in bulk InP, are overridden at nanoscale film boundaries by mode conversion that preferentially depopulates LA phonons into TA modes. This produces a TA-polarized non-equilibrium heat current and suppresses thermal conductivity to ~50% of bulk for 10–100 nm InP films at 100 K. The asymmetry in mode conversion arises from two compounding factors: the material factor D_λ v_λ is larger for TA than LA (because the large velocity gap gives TA phonons higher density of states and comparable or higher velocities at fixed frequency), and Snell's law restricts TA→LA conversion to phononsInc","pith_inferences":[],"forward_implications":["Nanofilms of InP and related III-V semiconductors could serve as thermal phonon polarizers, producing heat currents with controlled polarization that enable selective coupling to electrons, defects, or strain fields — useful for engineering hot-carrier lifetimes and defect-insensitive thermal properties.","The predicted κ suppression to ~50% of bulk for 10–100 nm InP films at 100 K is directly measurable with existing thin-film thermal conductivity techniques (time-domain thermoreflectance, suspended microbridge) on films already fabricable at sub-10-μm thickness.","The principle extends to any material where a large LA/TA velocity gap activates the AAA # 2 selection rule, suggesting a materials-design criterion: large acoustic-branch separation produces strong thermal polarization in nanofilms.","The temperature-independent saturation of κ/κ_b at small thickness (driven by the T-independent relative LA/TA contributions to bulk κ between 50–100 K) provides a clean experimental signature: below a critical thickness, further thinning or cooling should not change the normalized conductivity.","The finding that diffuse boundary scattering erases all mode-conversion effects (because direction randomisation dominates) means the polarizing effect requires specular or near-specular boundaries — a constraint on surface preparation for any device application."],"fun_headline_variants":["Nanofilm boundaries flip InP heat to transverse phonons at 100 K","Phonon mode conversion at nanofilm boundaries overrides bulk selection rules in InP","InP nanofilms polarize heat by depopulating longitudinal phonons","Boundary mode conversion halves InP nanofilm conductivity via TA-polarized heat","Snell's law and velocity gap drive transverse phonon polarization in InP nanofilms"],"cache_read_input_tokens":0,"weakest_assumption_plain":"The load-bearing premise is that the anisotropic, wave-vector-dependent group velocities and relaxation times of real crystals can be accurately collapsed into a single isotropic representation per polarization branch. If this isotropic projection distorts the relative magnitudes of the density-of-states–velocity product D_λ v_λ between LA and TA branches — or the wave-vector ratio q(ν,TA)/q(ν,LA) that sets the Snell's-law critical angle for mode conversion — then the asymmy","fun_headline_variants_meta":{"raw":{"variants":["Nanofilm boundaries flip InP heat to transverse phonons at 100 K","Phonon mode conversion at nanofilm boundaries overrides bulk selection rules in InP","InP nanofilms polarize heat by depopulating longitudinal phonons","Boundary mode conversion halves InP nanofilm conductivity via TA-polarized heat","Snell's law and velocity gap drive transverse phonon polarization in InP nanofilms"]},"model":"glm-5.2","effort":"high","cost_usd":0.0,"raw_usage":{"total_tokens":663,"prompt_tokens":572,"completion_tokens":91,"prompt_tokens_details":null},"tokens_in":572,"tokens_out":91,"duration_ms":19979,"temperature":1.0,"reasoning_tokens":null,"cache_read_input_tokens":0,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-07T18:57:51.054672+00:00","model_set":{"reader":"glm-5.2"},"falsifier":"Measure the in-plane thermal conductivity of 10–100 nm InP films at 100 K with specular boundaries. If κ does not drop to approximately 50% of bulk, or if the spectral contribution of LA phonons is not preferentially suppressed relative to TA, the mode-conversion polarization mechanism is not operating as predicted. Alternatively, if polarized phonon spectroscopy (e.g., inelastic neutron or Raman-based probes adapted for thin films) shows no TA-polarized non-equilibrium phonon population in the films, the central claim is falsified.","supporting_citations":[],"review_version":1}