{"id":"18d18f91-c6d2-48cd-aa1a-0fcd973fb6e5","arxiv_id":"1909.00976","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"The self-induced inverse spin Hall voltage in permalloy is non-monotonous in temperature and is explained by opposite-sign skew scattering and side-jump contributions to the spin Hall conductivity.","lead":"This paper reports that the self-induced inverse spin Hall voltage in permalloy thin films under ferromagnetic resonance varies non-monotonically with temperature, reaching a minimum near 100 K. The result matters because spin-to-charge conversion inside ferromagnets is often neglected but can rival platinum, so it should be accounted for in spintronic measurements and devices.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Measured IC(T) is compared to bulk σ_SH(T) without removing T-dependent spin-current source terms; the maximum at ~95 K may originate from J_s(T) or l_sf(T) rather than from bulk spin Hall conductivity.","rationale":"The paper presents a well-executed set of experimental controls and a plausible first-principles calculation, and I do not dispute the existence of a non-monotonic T-dependence in the measured transverse signal. The reader's verdict of CONDITIONAL is appropriate because the interpretation depends on assumptions about the spin-current source and on unmeasured T-dependent parameters. My stress-test goes one step further: even if one grants the interfacial-asymmetry source inherited from Tsukahara et al., the measured IC(T) contains T-dependent prefactors from spin pumping (1/α^2, g_r) and transport (l_sf, t*) that are not normalized out before comparison with the bulk σ_SH(T). Since the non-monotonic peak is the paper's key novel observation, the lack of a quantitative conversion model connecting IC to σ_SH is the most load-bearing weakness. This does not invalidate the paper, but it raises the required standard for the central claim. The proposed concrete test (normalizing IC by the measured α^2 and t* before comparison) would settle whether the bulk spin Hall conductivity is actually responsible for the non-monotonic behavior. My recommendation remains CONDITIONAL, matching the reader's verdict; hence verdict_should_be is UNCHANGED. Agreement is partial because the reader identified the spin-current source assumption as weakest, while I emphasize the uncharacterized T-dependence of that source even under the assumed mechanism.","tokens_in":10957,"tokens_out":8447,"duration_ms":94716,"concrete_test":"Recompute the relevant figure by dividing the measured IC(T) by the independently characterized spin-pumping and transport prefactors: define IC^corr(T) = IC(T) × α^2(T) / [t*(T) g_r(T)], using the manuscript's own α(T) from Fig. 2c, l_sf(T) derived from σ_xx(T), and g_r obtained from linewidth enhancement measurements; then plot IC^corr(T) against the first-principles σ_xy^z(T). If the maximum at ~95 K disappears or moves by more than ~20 K, the non-monotonicity is not attributable to the bulk σ_SH(T) alone. Repeat the same normalization for the t=32 nm film; if the normalized T-dependence differs between thicknesses, then l_sf(T) or interface-generated J_s(T) contributes significantly to the effect.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In the discussion of Fig. 4, the experimental charge current IC(T) is plotted alongside the first-principles bulk spin Hall conductivity σ_xy^z(T) and described as being in \"satisfactory qualitative agreement.\" However, IC is not a direct probe of σ_SH. In the self-induced geometry, IC ∝ J_s(T) × [spin-charge conversion efficiency], where J_s(T) is the spin current generated in the NiFe and the conversion efficiency depends on σ_SH(T) and on l_sf(T) through t* = l_sf tanh[t/(2l_sf)]. The paper itself characterizes α(T) (Fig. 2c), l_sf(T) from σ_xx(T) via l_sf = 0.91σ_xx×10^-12, and the spin pumping efficiency contains 1/α^2 and g_r(T). None of these T-dependent prefactors is removed before comparing IC(T) with the SPR-KKR σ_SH(T). The first-principles calculation models only the bulk σ_SH(T); it contains no model for the self-induced spin current generation, whose T-dependence is left unquantified. The observed non-monotonic IC(T) with a maximum at ~95 K could therefore be produced by a monotonic σ_SH(T) combined with a non-monotonic J_s(T) or l_sf(T) (for t=8 nm, t* ≈ l_sf for small l_sf, so an increasing l_sf at low T enhances the conversion efficiency). The controls using different capping layers, field-sweep rate, and AMR/ANE checks rule out several extrinsic artifacts, but they do not calibrate J_s(T). The central microscopic conclusion attributing the non-monotonic T-dependence to opposite-signed skew and side-jump plus intrinsic contributions in bulk NiFe is thus underdetermined by the comparison as presented.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper reports ferromagnetic-resonance spin-pumping measurements on single permalloy (NiFe) films with various capping and buffer layers. The symmetric transverse voltage V_sym is measured as a function of temperature, magnetic-field angle, NiFe thickness, and adjacent material, and converted to a charge current I_C. The data show a non-monotonic I_C(T) with a maximum near 95 K, independence of the adjacent material, and an increase of I_C with NiFe thickness. The authors compare I_C(T) with first-principles SPR-KKR calculations of the bulk spin Hall conductivity of NiFe and decompose the calculated sigma_xy into skew-scattering and side-jump-plus-intrinsic contributions, finding opposite signs and similar magnitudes, which they propose as the origin of the non-monotonic T-dependence. Pt buffer/capping experiments are used to determine the sign of the self-induced spin current and to estimate the Pt spin Hall angle.","tokens_in":11352,"tokens_out":7652,"duration_ms":80542,"significance":"If the interpretation survives the requested analysis, the paper establishes a previously underappreciated contribution: ferromagnet self-induced spin-charge conversion can be as efficient as Pt at low temperature and should be included in spin-pumping analyses. The experimental dataset is unusually complete, comprising temperature dependence, a thickness series, angular dependence, and a set of capping and buffer materials including Pt reference samples; the sweep-rate and heat-sinking checks address the main thermal artifacts. The SPR-KKR first-principles calculations are a genuine strength: they are not fitted to the data, and they provide an independent qualitative microscopic scenario with a decomposition into skew and side-jump plus intrinsic contributions. The main weakness is that the central comparison between I_C(T) and sigma_SH(T) is not yet a direct comparison, because the temperature dependence of the spin-current source term is not calibrated or removed.","major_comments":[{"comment":"The central microscopic claim is supported by a qualitative side-by-side comparison between the measured charge current I_C(T) and the calculated bulk spin Hall conductivity sigma_xy(T), but I_C is not a direct probe of sigma_SH. In the self-induced geometry I_C is proportional to the spin-current source J_s(T) times a conversion factor that itself contains l_sf(T) and the spin-pumping efficiency. The paper reports alpha(T) in Fig. 2(c), infers l_sf(T) from sigma_xx(T), and shows t*/alpha^2(T) in the inset of Fig. 6(b), yet I_C(T) in Fig. 4(a) is not normalized by any of these T-dependent prefactors before comparison with sigma_xy(T). Since the SPR-KKR calculation models only the bulk sigma_xy(T) and contains no model for the generation of J_s,self(T), the observed maximum near 95 K could be produced by a monotonic sigma_xy(T) combined with a non-monotonic J_s(T) or l_sf(T). Please re-analyze the data with the measured T-dependent prefactors removed, for example by dividing I_C by t*/alpha^2, or provide an explicit model of J_s(T), and state the resulting uncertainty.","section":"Fig. 4 and surrounding text"},{"comment":"The Pt buffer/capping experiment is used to assign the sign and direction of the self-induced spin current, but the quantitative analysis assumes a temperature-independent behavior that is not established. The spin mixing conductance g_r^updown is calculated only at 300 K from linewidth broadening, and its T-dependence is not measured; the mechanism of asymmetric spin-dependent scattering at the two NiFe interfaces is inherited from ref. [13] and is not independently tested here. Thus the bulk-origin interpretation of the non-monotonic T-dependence, which is the paper's main claim, relies on an unquantified source term. Please either measure or bound the T-dependence of the spin-current source, for example through a temperature- and thickness-dependent linewidth analysis, or narrow the claim to consistency rather than demonstration.","section":"Fig. 5 and Pt-reference analysis"},{"comment":"The decomposition of sigma_xy into skew-scattering and side-jump-plus-intrinsic contributions is performed by varying the alloy composition at each temperature and using sigma_xy = S sigma_xx + sigma_sj+intr, but no uncertainty is reported for the slope S or the intercept, and the number of compositions used is not given. Because the conclusion that the two contributions are nearly equal and opposite is the microscopic explanation of the non-monotonic T-dependence, the robustness of S and of the intercept to the composition set should be documented, together with an estimate of the error on the ratio sigma_sk / sigma_sj+intr quoted as about -1.2.","section":"Fig. 4(c) and scaling decomposition"}],"minor_comments":[{"comment":"The phrase 'where ΔHpp is the the peak-to-peak line width' contains a duplicated article; please fix the typo.","section":"Experimental details"},{"comment":"The word 'non-monotonous' should be replaced by 'non-monotonic' for consistency with standard usage.","section":"Throughout"},{"comment":"Error bars are absent for I_C(T); given that I_C is an extracted quantity involving two V_sym measurements, a statement of run-to-run reproducibility or an error estimate would strengthen the comparison.","section":"Fig. 4"},{"comment":"The sentence 'This finding also seems to infer that the intrinsic contribution to the ISHE is negligible' uses 'infer' where 'imply' is intended; please rephrase.","section":"Page 7, discussion of intrinsic contribution"},{"comment":"The relation l_sf,NiFe = 0.91 sigma_xx x 10^-12 is taken from ref. [40] and applied to 8-32 nm films without discussing its range of validity; please add a sentence justifying its use for these thicknesses.","section":"Page 9, l_sf estimation"},{"comment":"The absence of the effect in CoFeB is reported without comment; a brief explanation that this could reflect a different spin diffusion length or spin Hall angle would avoid overinterpretation.","section":"Supplemental material, Fig. S1(g)"}],"recommendation":"major_revision","confidential_remarks":"This is a solid experimental paper with an interesting and timely claim. The main reservation is that the comparison between I_C(T) and sigma_SH(T) is not yet a demonstration of the bulk origin of the T-dependence, because the T-dependent spin-current source term is not calibrated. The requested normalization appears feasible with data already in the paper, such as alpha(T), l_sf(T), and t*/alpha^2, so I do not see this as requiring new physics; hence major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know before you read it. The dataset is genuinely useful: a non-monotonous temperature dependence of the self-induced transverse voltage in single NiFe films, with careful controls against PHE, ANE, sweep-rate effects, and interface artifacts. The central interpretation, however, is softer than the authors make it look. The stress-test note lands: IC(T) is compared directly to the first-principles bulk sigma_SH(T) without factoring out the temperature dependence of the spin current J_s(T) or the spin-charge conversion efficiency (which contains l_sf(T) through t*). The peak at ~95 K could in principle come from those prefactors.\n\nWhat is new: prior self-induced ISHE work was at room temperature. The 50–300 K sweep, the thickness series, and the SPR-KKR decomposition into skew scattering and side-jump plus intrinsic contributions are all new. The Pt buffer/capping experiments are a nice control; they give a sensible theta_SH l_sf for Pt and strengthen the claim that self-induced conversion in the ferromagnet has to be included when analyzing F/Pt spin pumping.\n\nSoft spots. The experiment measures a voltage, not a spin current. J_s(T) is assumed, inherited from Tsukahara's asymmetric-interface picture, and never independently calibrated. The paper itself has the ingredients to correct for alpha(T) and l_sf(T) — it quotes alpha and estimates l_sf from sigma_xx — but it does not use them in Fig. 4. So the “satisfactory qualitative agreement” is not a quantitative test of the bulk sigma_SH curve, and the skew/side-jump cancellation is supported mainly by the calculation, not by the data. The bulk-origin claim is on firmer ground: the invariance under capping materials and the thickness dependence are good evidence that the conversion step is bulk.\n\nMinor: no data/code and some missing error bars on IC; not disqualifying for an experimental paper of this vintage.\n\nWho this is for: anyone doing spin pumping or spin-charge conversion in F/NM systems. I would cite it for the experimental phenomenology, not for the microscopic mechanism.\n\nRecommendation: yes, send it to a serious referee. Ask the authors to either remove the J_s(T) ambiguity by dividing by the known prefactors, or at least state clearly that the comparison is indicative rather than demonstrative. If they do that, the paper is a solid contribution.","headline":"Solid experimental paper with a real non-monotonic temperature dependence in permalloy self-induced ISHE, but the comparison to bulk spin Hall conductivity is underdetermined because J_s(T) is not calibrated; still worth refereeing.","tokens_in":11927,"tokens_out":5666,"would_cite":true,"duration_ms":63585,"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":"In a single permalloy film, the self-induced inverse spin Hall voltage peaks near 95 K and is governed by the bulk spin Hall conductivity.","keywords":["inverse spin Hall effect","spin pumping","ferromagnetic resonance","permalloy","spin Hall conductivity","skew scattering","side jump","temperature dependence"],"falsifier":"Measure the self-induced transverse voltage versus temperature for permalloy films across a composition span (for instance Ni85Fe15 to Ni70Fe30). In the paper's decomposition the skew-scattering contribution is proportional to $\\sigma_{xx,\\mathrm{NiFe}}$ while the side-jump-plus-intrinsic contribution is not, so the temperature of the extremum should shift with composition; a composition-independent extremum would rule out the proposed cancellation mechanism.","tokens_in":10768,"feed_emoji":"🧲","tokens_out":15950,"duration_ms":154684,"temperature":0.7,"pith_summary":"Plain permalloy films, with no attached spin-sink layer, convert their own precessing magnetization into a transverse charge current under ferromagnetic resonance. The paper shows that this self-induced inverse spin Hall voltage is non-monotonous in temperature: it grows as the film cools, reaches a maximum near 95 K, and then falls. The same non-monotonous profile appears whatever material is in contact with the permalloy, and the current grows with film thickness, so the spin-to-charge conversion is argued to take place throughout the bulk of the ferromagnet. First-principles calculations reproduce the profile and trace it to two bulk contributions to the spin Hall conductivity—skew scattering and side-jump-plus-intrinsic—that are comparable in size and opposite in sign. The authors conclude that ferromagnets contribute to their own spin-charge conversion, so measurements of spin Hall effects in adjacent non-magnetic layers must subtract that contribution.","feed_headline":"95 K: permalloy spin-to-charge conversion peaks","feed_subtitle":"The effect is strongest at 95 K, grows with thickness, and rivals platinum's conversion efficiency.","key_machinery":"The load-bearing object is the temperature-dependent bulk spin Hall conductivity of permalloy, $\\sigma^z_{xy,\\mathrm{NiFe}}(T)$, computed with a fully relativistic multiple-scattering Kubo Green-function method and decomposed through the scaling relation $\\sigma^z_{xy,\\mathrm{NiFe}} = \\sigma_{xx,\\mathrm{NiFe}} S + \\sigma^{sj+intr}_{xy,\\mathrm{NiFe}}$, where $S$ is the skewness factor extracted from alloy-composition variations. This decomposition carries the argument: the measured non-monotonous voltage is explained by the two terms having opposite signs and similar amplitudes, so their sum changes non-monotonically with temperature. On the experimental side, the central machinery is the standard symmetric/antisymmetric decomposition of the resonance voltage together with the angular-dependent ISHE expression that relates $V_\\mathrm{sym}$ to the magnetization tilt angle $\\theta_M$, which selects the ISHE contribution from the anisotropic magnetoresistance and the anomalous Nernst effect.","core_discovery":"Under ferromagnetic resonance at 9.6 GHz, a single Ni81Fe19 film produces a transverse voltage whose symmetric part $V_\\mathrm{sym}$ changes sign when the applied field is reversed and follows the angular dependence of the inverse spin Hall effect. Measured between 50 and 300 K, $V_\\mathrm{sym}$ is non-monotonous: its magnitude peaks near 95 K, while the resonance linewidth varies monotonically over the same range, and separate AMR and angular checks rule out magnetoresistance and anomalous Nernst artifacts. The extracted charge current $I_C$ is compared with first-principles calculations of the bulk spin Hall conductivity $\\sigma^z_{xy,\\mathrm{NiFe}}(T)$, which show a non-monotonous profile with an extremum near 100 K. By separating $\\sigma^z_{xy,\\mathrm{NiFe}}$ into a skew-scattering term $\\sigma^{sk}_{xy,\\mathrm{NiFe}} = \\sigma_{xx,\\mathrm{NiFe}} S$ and a combined side-jump-plus-intrinsic term $\\sigma^{sj+intr}_{xy,\\mathrm{NiFe}}$, the authors show that the non-monotonicity comes from two terms of opposite sign and nearly equal magnitude, with the intrinsic contribution appearing negligible in permalloy. Control stacks with different oxides, metals, and growth orders give the same temperature profile, and the signal grows with NiFe thickness, supporting a bulk origin; comparison with Pt reference layers shows that near 95 K the self-induced conversion in NiFe can be as efficient as platinum's.","pith_inferences":["The decomposition suggests that shifting the alloy composition should move the temperature of the extremum, since the skew-scattering term is tied to $\\sigma_{xx,\\mathrm{NiFe}}$; a composition series would test this prediction directly.","The experiments go down to 50 K; because the two opposing terms have different temperature slopes, extending measurements to lower temperatures would show whether the signal continues to fall, flattens, or changes sign.","The conclusion that the intrinsic spin Hall contribution is negligible in permalloy rests on identifying the combined side-jump-plus-intrinsic term; isolating the intrinsic part computationally is a direct way to check that inference."],"forward_implications":["Near 95 K, permalloy's self-induced spin-to-charge conversion is comparable in strength to platinum's, so spin Hall angles extracted from ferromagnet/Pt bilayers without a correction for the ferromagnet's own contribution carry a systematic error.","Because the temperature profile is independent of the material in contact with the permalloy, the conversion cannot be blamed on interface effects such as the anomalous Nernst effect; it is a property of the ferromagnet's bulk.","The non-monotonous temperature dependence means a single-temperature measurement can miss the effect almost entirely; temperature series are needed to characterize a ferromagnet's spin-charge conversion.","The skew-scattering contribution scales with the longitudinal conductivity $\\sigma_{xx,\\mathrm{NiFe}}$, so alloy composition and disorder offer a practical lever to tune the sign and magnitude of the self-induced voltage."],"supporting_citations":[{"why":"Provides the theoretical expression for the angular dependence of the ISHE voltage and the spin-pumping formulas used to model $V_\\mathrm{sym}$ and the spin mixing conductance.","marker":"[10]"},{"why":"The earlier observation and proposed mechanism of self-induced ISHE in NiFe that this study extends; supplies the interfacial asymmetric spin-scattering source picture.","marker":"[13]"},{"why":"The first-principles relativistic Kubo Green-function method used to compute the temperature-dependent spin Hall conductivity of bulk permalloy.","marker":"[27–29]"},{"why":"The resonant-scattering model with split impurity levels that explains why the skew and side-jump contributions enter with opposite signs.","marker":"[31]"},{"why":"Provides reference values for platinum's spin Hall parameters and spin diffusion length, used to compare NiFe conversion and to estimate the error from neglecting the self-induced current.","marker":"[38]"},{"why":"Controls for the anomalous Nernst effect, including sweep-rate independence and heat-sinking variations, used to assign the signal to the ISHE.","marker":"[35]"},{"why":"Earlier work arguing that the ferromagnet's self-induced spin current must be included when extracting spin Hall parameters; the paper's Pt-stack comparison supports it.","marker":"[12]"}],"fun_headline_variants":["Spin-to-charge conversion in permalloy peaks at 95 K","Permalloy's self-induced inverse spin Hall effect peaks near 95 K","Non-monotonous spin Hall conductivity in permalloy: peak at 95 K","Bulk spin-to-charge conversion in permalloy rivals platinum at 95 K","Temperature-dependent spin Hall effect in permalloy peaks at 95 K"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The interpretation assumes that the spin current feeding the effect is generated by asymmetric spin-dependent scattering at the two surfaces of the permalloy film; if the spin current actually comes from another source, the bulk spin Hall explanation of the temperature dependence does not follow.","fun_headline_variants_meta":{"raw":{"variants":["Spin-to-charge conversion in permalloy peaks at 95 K","Permalloy's self-induced inverse spin Hall effect peaks near 95 K","Non-monotonous spin Hall conductivity in permalloy: peak at 95 K","Bulk spin-to-charge conversion in permalloy rivals platinum at 95 K","Temperature-dependent spin Hall effect in permalloy peaks at 95 K"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000997,"raw_usage":{"total_tokens":4238,"prompt_tokens":976,"completion_tokens":3262,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":592,"completion_tokens_details":{"reasoning_tokens":3159}},"tokens_in":592,"tokens_out":3262,"duration_ms":23968,"temperature":1.0,"reasoning_tokens":3159,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T05:31:04.786169+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the self-induced transverse voltage versus temperature for permalloy films across a composition span (for instance Ni85Fe15 to Ni70Fe30). In the paper's decomposition the skew-scattering contribution is proportional to $\\sigma_{xx,\\mathrm{NiFe}}$ while the side-jump-plus-intrinsic contribution is not, so the temperature of the extremum should shift with composition; a composition-independent extremum would rule out the proposed cancellation mechanism.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the theoretical expression for the angular dependence of the ISHE voltage and the spin-pumping formulas used to model $V_\\mathrm{sym}$ and the spin mixing conductance."},{"cited_title":"Tsukahara, Y","cited_arxiv_id":null,"evidence_quote":"The earlier observation and proposed mechanism of self-induced ISHE in NiFe that this study extends; supplies the interfacial asymmetric spin-scattering source picture."},{"cited_title":"Fert and P","cited_arxiv_id":null,"evidence_quote":"The resonant-scattering model with split impurity levels that explains why the skew and side-jump contributions enter with opposite signs."},{"cited_title":"Rojas-Sánchez, N","cited_arxiv_id":null,"evidence_quote":"Provides reference values for platinum's spin Hall parameters and spin diffusion length, used to compare NiFe conversion and to estimate the error from neglecting the self-induced current."},{"cited_title":"Negligible thermal contributions to the spin pumping signal in ferromagnetic metal-Platinum bilayers","cited_arxiv_id":"1905.00771","evidence_quote":"Controls for the anomalous Nernst effect, including sweep-rate independence and heat-sinking variations, used to assign the signal to the ISHE."},{"cited_title":"Azevedo, O","cited_arxiv_id":null,"evidence_quote":"Earlier work arguing that the ferromagnet's self-induced spin current must be included when extracting spin Hall parameters; the paper's Pt-stack comparison supports it."}],"review_version":1}