{"id":"dcdd90fc-31f2-4bb9-a8b6-0e7515cbf8b3","arxiv_id":"2509.06007","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"A finite-difference simulation scheme for coupled magnetic and elastic wave dynamics was implemented, including interface jump conditions for stress and strain in magnetic heterostructures.","lead":"This paper describes a computer simulation method that models how sound waves and magnetic behavior in magnetic films affect each other as they happen. It carefully handles the boundaries where different materials meet, a detail earlier finite-difference simulations struggled with.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The interface jump treatment is incomplete: Eq. 45 enforces displacement/traction continuity but not the continuous-acceleration condition (78), which the paper itself admits in §3.5.1, so the abstract's 'correctly incorporating' overstates the method.","rationale":"The reader's weakest assumption is exactly the load-bearing concern: the jump-corrected first-derivative stencils are reused to build the second-derivative force density in Eq. (45) without enforcing the continuous-acceleration condition, and the paper itself documents the resulting interface error in §3.5.1. This is not a manufactured objection; it is an internal inconsistency between the abstract's strong claim and the method's admitted limitation. The concern is load-bearing because the paper's novelty is the interface-jump treatment, and the heterostructure results rely on it. I see no reason to move the verdict: the method is likely usable and well validated in homogeneous regions, but the interface handling is conditional, requiring either a corrected stencil or an explicit statement of reduced interface accuracy. Hence the reader's CONDITIONAL verdict is appropriate; no adjustment is needed.","tokens_in":20685,"tokens_out":4240,"duration_ms":55869,"concrete_test":"Run a grid-convergence test on the §3.5.1 SH eigenmode at fixed f≈5.26 GHz, λ=600 nm: refine the z-discretization across the GaAs/Ni interface (e.g. 50, 100, 200, 400 cells per wavelength) and compute f_y from Eq. (45) with the iterative B-term procedure. Compare the interface-local L2 error against the analytic f_y satisfying Eq. (78). If the error does not decrease with the scheme's nominal order or remains significant at 400 cells/λ, then the interface treatment is not correct as claimed. As a secondary check, rerun the non-reciprocity loss in Fig. 20 with a corrected interface stencil enforcing Eq. (78); if the reported asymmetry changes materially, the central application is sensitive to this gap.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that the solver correctly incorporates stress and strain discontinuities at material interfaces. That claim rests on the first-derivative jump stencils, Eqs. (40)-(42), and their reuse in the second-derivative / force-density approximation, Eq. (45). The weak point is that Eq. (45) uses only displacement and traction continuity to compute ∇·(σ - σ_m); it does not enforce the additional interface condition that follows for time-harmonic solutions from continuity of u: because u is continuous, ü = -ω²u is also continuous, and since f = ρü, the force density must satisfy the jump condition (1/ρ)(∂σ_yz/∂z + ∂σ_xy/∂x) continuous, i.e. Eq. (78). The paper's own §3.5.1 shows that the proposed Eq. (45) does not reproduce this condition, and the authors call the result 'not a proper solution'; the corrected scheme they plot is explicitly special-case and not adopted. They justify retaining Eq. (45) by asserting the error is small, but no grid-convergence study is provided for the interface-local force density, and the later heterostructure simulations (Rayleigh thin film, SH thick film) depend on this same approximate interface treatment. Thus the abstract's claim of 'correctly incorporating' stress/strain discontinuities is not fully supported by the numerical scheme as presented.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a finite-difference time-domain scheme for self-consistent magnetoelastic dynamics, implemented as an extension of the magnum.np micromagnetic library. The method couples the Landau-Lifshitz-Gilbert equation to an elastodynamic equation of motion, and derives jump-corrected first-derivative stencils at material interfaces (Eqs. 40-42) together with a second-derivative approximation for the force density (Eq. 45). Neumann boundary conditions are handled through boundary-value formulas (Eqs. 49-51) or an airbox approach. The scheme is validated against analytic solutions for a static FM/NM stack (Eq. 56), bulk shear-wave damping with an energy balance (Eq. 65), and Rayleigh-wave boundary convergence (Eqs. 67-68). The paper then applies the solver to Rayleigh SAW attenuation in a thin Ni film on LiTaO3 and to SH-SAW attenuation in a thick Ni film on GaAs, reproducing the experimentally expected non-reciprocity. The abstract claims that the implementation 'correctly incorporat[es] stress and strain discontinuities at material interfaces.'","tokens_in":21068,"tokens_out":5885,"duration_ms":65934,"significance":"If the scheme performs as stated, this is a useful contribution: it extends finite-difference micromagnetism to self-consistent magnetoelastic wave problems, provides analytic benchmarks for the coupled system, and is shipped in an open-source package (magnum.np ≥ 2.1.0), which supports reproducibility. The bulk-wave energy-balance test and the boundary-convergence study are particularly valuable. The application sections demonstrate qualitative agreement with experimental non-reciprocity, but the central interface-discontinuity claim is weakened by the paper's own admission in Sec. 3.5.1 that the force density does not enforce the continuous-acceleration jump condition, and no convergence study is supplied for this interface-local error. The experimental agreement also depends on fitted and roughly estimated parameters. With the interface treatment clarified and quantified, the paper would be a solid contribution.","major_comments":[{"comment":"The paper states that the force density built from Eq. (45) does not reproduce the continuous-acceleration jump condition (78) at the FM/NM interface; Fig. 17 shows the mismatch in f_y at z=0. The authors call the corrected scheme 'not a proper solution' and retain (45) because 'the error in f_y appears to be small,' but no grid-convergence study is given for this interface-local error. The R-SAW and SH-SAW simulations use the same (45)-based force density, so the abstract's claim that the implementation 'correctly incorporat[es] stress and strain discontinuities at material interfaces' is not supported as stated. Please quantify the error as a function of Δx and either correct the interface treatment or reformulate the claim to the jump conditions that are actually enforced.","section":"§3.5.1, Eq. (45), Eq. (78)"},{"comment":"The in-plane anisotropy field H_k is varied to match the experimental resonance h_ext = 46.5 mT, and the magnetoelastic constants are set to a rough estimate (b1 = b2 = 12 T). Consequently, the angular non-reciprocity shown in Figs. 13-14 is not a parameter-free prediction. The paper should separate fitted from predicted parameters and discuss sensitivity to H_k and b; without this, the experimental agreement is suggestive but not a strong quantitative validation.","section":"§3.4.1, Fig. 11"},{"comment":"The construction of the B-parameters in the jump stencils (40)-(42) requires an iterative update of parallel gradients and magnetic stress. The paper asserts that further iteration steps provide no benefits (§3.3), but gives no convergence criterion or residual measure. Since the scheme is advertised as self-consistent, please provide convergence data for the iteration for at least the validation problems, or state a formal stopping rule.","section":"§2.2, Table 1 and paragraph after Eq. (47)"},{"comment":"The Rayleigh speed is fitted as c ≈ 3104 m/s in §3.3, but Fig. 11 and §3.4.1 use c = 3014 m/s for the dispersion intersection and the H_k fit. This 90 m/s discrepancy is not explained; the resonance condition depends on c. Please clarify which value is correct and ensure consistency throughout.","section":"§3.3 vs §3.4.1"}],"minor_comments":[{"comment":"Typo: 'torch.gardient' should be 'torch.gradient'.","section":"Fig. 3 caption"},{"comment":"Minor language: 'an 200 mT external field' should be 'a 200 mT external field'; 'Therefor' should be 'Therefore'.","section":"§1, §3.4.2"},{"comment":"The magnetoelastic field H_magEl_i is written in index notation but the vector form is not displayed; please define the field vector explicitly for readability.","section":"Eq. (17)"},{"comment":"The 'Boundary Value Error' used in Fig. 9 is not defined in the text or caption. Please define the error measure and specify the exact discretizations corresponding to Nbc = 1, 2, 3.","section":"Fig. 9 and §2.3"}],"recommendation":"major_revision","confidential_remarks":"The self-identified limitation in Sec. 3.5.1 is the decisive point. The paper should either correct the interface treatment or tone down the abstract's 'correctly incorporating' claim. I would not reject because the bulk of the validation is solid, but the load-bearing interface issue needs to be resolved before acceptance."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: this is a solid, genuinely useful methods paper with clean analytic validations, but the abstract's claim that the scheme correctly incorporates stress and strain discontinuities outruns the method. The paper's own §3.5.1 shows that the force-density stencil (Eq. 45) does not enforce continuity of acceleration (Eq. 78) at FM/NM interfaces, and the authors keep the approximate stencil anyway because the error 'appears to be small.' No grid-convergence study is given for that interface-local error. That is the main soft spot, and the stress-test note lands on it.\n\nWhat is new: the jump-condition-aware finite-difference stencils with the iterative B-parameter construction, which goes beyond the existing FD magnetoelastic solvers (mumax+, Vanderveken et al.). The validations in §3.1–3.3 are real: the Ni/Al static stack matches the analytic displacement, the bulk-shear energy balance closes exactly against the Gilbert-loss expression, and the Rayleigh-mode boundary study, including the airbox approach, is carefully done. These benchmarks are external and analytic, so the core method is not circular. The thick-film SH-SAW simulation with non-uniform out-of-plane magnetization is a new demonstration, and capturing the non-reciprocal SAW attenuation is a meaningful physical target. The citation pattern is reasonable; relevant prior work is engaged and the magnum.np self-citation is appropriate since the solver ships in that library.\n\nSoft spots, in proportion. First, the interface force-density issue above. Credit where due: it is disclosed honestly, the authors propose a special-case correction for the SH mode and reject it for good reasons, and the deviation between their two loss estimates in Fig. 20 is attributed to this error. But the abstract overstates, and either a proper second-derivative jump stencil or a plain statement of first-order interface accuracy is needed. Second, the §3.4.1 resonance match is partly calibration: Hk is varied to hit the observed field and λ is a rough estimate — the paper says so, and the angular non-reciprocity is a qualitative reproduction, not a fit. Minor. Third, the forward-difference boundary formula is order-zero and they avoid it; known and stated.\n\nThe CONDITIONAL verdict is about right. The central method holds up; the interface treatment is incomplete but disclosed. I would accept this for peer review and expect a revise-and-resubmit: soften the abstract or strengthen the interface scheme. Anyone building FD magnon-phonon simulations should read it, especially §3.5.1.","headline":"A solid, genuinely useful finite-difference magnetoelastic solver with clean analytic validations, but the abstract's 'correctly incorporating' claim outruns the method; the paper's own §3.5.1 shows the FM/NM interface force-density jump is not handled exactly.","tokens_in":21585,"tokens_out":5193,"would_cite":true,"duration_ms":50493,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["75.80.+q"],"model":"deepseek-v4-flash","headline":"A finite-difference micromagnetic solver can now advance magnetization and elastic displacement together, with jump-corrected derivatives at interfaces, reproducing non-reciprocal surface-acoustic-wave attenuation in magnetic films.","keywords":["magnetoelastic dynamics","finite-difference micromagnetism","Landau-Lifshitz-Gilbert equation","surface acoustic waves","magnon-phonon interaction","interface jump conditions","nonreciprocal attenuation","magnetostriction"],"falsifier":"Compute the Section 3.5.1 shear-horizontal mode on a Ni/GaAs stack at successively refined meshes and compare the finite-difference force density f_y at the interface with the analytic eigenmode (Eqs. 73-77) and with the acceleration-jump condition (Eq. 78). If the interface error fails to converge to zero with mesh refinement, or if an experimental measurement of the angular loss contrast in a 300 nm Ni film disagrees with the predicted φ=90° versus φ=-90° non-reciprocity beyond the estimated discretization error, the central claim is falsified.","tokens_in":20596,"feed_emoji":"🧲","tokens_out":9451,"duration_ms":102758,"temperature":0.7,"pith_summary":"The paper aims to show that a finite-difference micromagnetic solver can simulate magnetoelastic dynamics self-consistently—advancing the Landau-Lifshitz-Gilbert equation with a strain-induced field while the elastic equation of motion is advanced at the same time—provided the discretization accounts for stress and strain jumps at material interfaces. The authors derive jump-corrected first-derivative stencils for the displacement, build the elastic force density from them, and implement open-boundary conditions either with boundary-value formulas or by adding a zero-stiffness 'air' layer. They demonstrate the need for these stencils in a static Ni/Al stack, in bulk shear-wave energy accounting, and in two surface-acoustic-wave settings, where the computed attenuation shows the experimentally expected non-reciprocity under field reversal. The paper's own limitation is acknowledged: at a ferromagnet/nonmagnet interface, the force density from the second-derivative formula does not enforce continuity of acceleration, leaving a small interface error.","feed_headline":"Magnon-phonon waves simulated with jump-aware stencils","feed_subtitle":"Self-consistent magnetization and elastic-wave dynamics in magnetic heterostructures, validated by non-reciprocal SAW attenuation.","key_machinery":"The load-bearing object is the jump-corrected first-derivative stencil family for displacement gradients—one-sided forward and backward differences, Eqs. (40)-(41), averaged into Eq. (42), whose offset parameters (Table 1) are built from the stiffness tensor and the magnetically induced stress. These stencils turn the interface condition that the traction be continuous into usable finite-difference formulas, and they supply the second-derivative approximation (45) that produces the force density f = ∇·(σ−σ_m). Because the jump parameters depend on other gradient components and on the magnetization at the interface, the gradients are assembled iteratively; open boundaries are handled either b","core_discovery":"The central claim is that the coupled system of magnetization direction, displacement, and momentum density can be integrated in time on a regular finite-difference grid if the spatial derivatives of the displacement are replaced by stencils that explicitly impose the interface conditions: displacement continuity and traction continuity. The paper derives one-sided differences, Eqs. (40)-(41), whose jump parameters C and B encode the stiffness tensor and the magnetic stress; since B depends on gradient components parallel to the interface, the strain and force density are computed iteratively. The method is validated against a static analytic solution for a Ni/Al bilayer, against an energy-b","pith_inferences":["Beyond the paper: the residual acceleration-continuity error at ferromagnet/nonmagnet interfaces could be removed by extending the iterative B-parameter scheme to the force density, as the authors sketch for the shear-horizontal mode; a general formulation would make the method exact at the same order as the underlying stencils.","Beyond the paper: the same solver could be applied to composite multiferroic or SAW sensor stacks, where interface traction matching determines how efficiently a piezoelectric strain is transferred to a magnetic layer; the static Ni/Al test is the simplest version of that transfer.","Beyond the paper: because the non-reciprocal loss pattern is computed without any unidirectional anisotropy, matching measured angular loss curves could be used to extract magnetoelastic coupling constants b1 and b2 from transmission data, avoiding the amplitude-calibration ambiguity noted in the paper.","Beyond the paper: the restriction to stiffness tensors of the form (28) excludes materials like trigonal LiTaO3, so the authors fit an effective isotropic Rayleigh profile; generalizing Table 1 to full elastic anisotropy is the evident next step for exploring true substrate-film combinations."],"forward_implications":["A finite-difference micromagnetic code can now treat magnon-phonon coupling without precomputed or measured stress fields, opening rectangular, periodically repeated device geometries to self-consistent study.","Static magnetostriction in ferromagnet/nonmagnet stacks is reproduced only by the jump-corrected stencils; ordinary 3-point gradients give qualitatively wrong interfacial stress.","In bulk shear-wave simulations, the energy lost through Gilbert damping exactly equals the energy injected by the traction boundary condition, confirming that the coupled solver conserves energy apart from the intended damping.","The Rayleigh-mode and shear-horizontal-mode simulations both yield the expected non-reciprocity of surface-acoustic-wave attenuation with respect to field direction, the signature effect used in experimental magnon-phonon devices.","The air-box implementation of open boundaries is the most accurate of the tested variants, reaching relative errors below 1e-2 at 200-300 nodes per wavelength."],"supporting_citations":[{"why":"Supplies the fully coupled mechanical-micromagnetic model and the strain-induced effective field expression that the present scheme extends.","marker":"[18]"},{"why":"Provides the elastic equation of motion and the energy-balance derivation used to build and validate the coupled integrator.","marker":"[25]"},{"why":"Earlier finite-difference magnetoelastic simulator whose damping term and interface-challenge framing are carried over.","marker":"[22]"},{"why":"Supplies the exchange-jump stencil pattern that the authors generalize to elastic displacement gradients.","marker":"[27]"},{"why":"Experimental Ni/LiTaO3 SAW attenuation study that fixes the magnetic parameters and the non-reciprocity baseline.","marker":"[10]"},{"why":"Theory-and-experiment SAW-driven ferromagnetic-resonance reference for the exponential attenuation law and angular loss symmetry.","marker":"[8]"},{"why":"Explains the helicity-mismatch mechanism and strain-ratio dependence behind the non-reciprocal attenuation.","marker":"[9]"},{"why":"Provides the analytic Rayleigh and layered shear-horizontal eigenmode formulas used for validation and injection.","marker":"[33]"},{"why":"Sinc-pulse standard method used to compute spin-wave dispersions and locate resonant fields.","marker":"[38]"}],"fun_headline_variants":["Jump-aware stencils couple magnons and phonons in films","Interface-corrected simulation of magnetoelastic waves","Finite-difference model captures interface jumps in films","Simulating magnon-phonon coupling with traction continuity","Handling stress jumps in magnetic heterostructures"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The load-bearing premise is that enforcing continuity of displacement and traction at an interface, through the modified first-derivative stencils, is enough to compute the force density ∇·(σ−σ_m); the paper's own Fig. 17 shows that this leaves a residual at a ferromagnet/nonmagnet interface because continuity of acceleration is not enforced, so the claim of correct interface handling stands only if that residual is small at the resolutions and materials used.","fun_headline_variants_meta":{"raw":{"variants":["Jump-aware stencils couple magnons and phonons in films","Interface-corrected simulation of magnetoelastic waves","Finite-difference model captures interface jumps in films","Simulating magnon-phonon coupling with traction continuity","Handling stress jumps in magnetic heterostructures"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001231,"raw_usage":{"total_tokens":4854,"prompt_tokens":664,"completion_tokens":4190,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":408,"completion_tokens_details":{"reasoning_tokens":4111}},"tokens_in":408,"tokens_out":4190,"duration_ms":32109,"temperature":1.0,"reasoning_tokens":4111,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T04:38:15.510374+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the Section 3.5.1 shear-horizontal mode on a Ni/GaAs stack at successively refined meshes and compare the finite-difference force density f_y at the interface with the analytic eigenmode (Eqs. 73-77) and with the acceleration-jump condition (Eq. 78). If the interface error fails to converge to zero with mesh refinement, or if an experimental measurement of the angular loss contrast in a 300 nm Ni film disagrees with the predicted φ=90° versus φ=-90° non-reciprocity beyond the estimated discretization error, the central claim is falsified.","supporting_citations":[{"cited_title":"Vanderveken, J","cited_arxiv_id":null,"evidence_quote":"Earlier finite-difference magnetoelastic simulator whose damping term and interface-challenge framing are carried over."},{"cited_title":"Heistracher, C","cited_arxiv_id":null,"evidence_quote":"Supplies the exchange-jump stencil pattern that the authors generalize to elastic displacement gradients."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the analytic Rayleigh and layered shear-horizontal eigenmode formulas used for validation and injection."},{"cited_title":"Venkat, D","cited_arxiv_id":null,"evidence_quote":"Sinc-pulse standard method used to compute spin-wave dispersions and locate resonant fields."}],"review_version":1}