{"id":"f8f5797c-8b6d-4a90-997c-ddc338a4c5bd","arxiv_id":"2501.10260","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Measurements after subtracting single-layer FM signals find Ta's dampinglike torque is negative and nearly unchanged for Ni, Permalloy, Fe, CoFeB, and FePt, indicating no orbital current torque.","lead":"By re-examining spin-torque ferromagnetic resonance data in Ta/ferromagnet bilayers, the authors find that the positive torque previously attributed to orbital currents in Ta/Ni is actually a self-generated signal from the Ni layer. The paper concludes that Ta's orbital Hall effect produces no detectable torque on the adjacent magnet, and only Ta's spin Hall effect contributes.","discovery_kind":"replication","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Subtractive correction may over-subtract the Ni self-torque because the 1-nm-Ta control has a different FM current density than the 5-nm-Ta bilayer; no normalization is reported.","rationale":"The reader's weakest assumption correctly identifies the subtraction of the FM self-induced signal as the load-bearing step. I sharpen it to a specific, testable mechanism: the control sample has a different current distribution than the bilayer, so even if the FM self-torque is intrinsic and unchanged, the measured S and A amplitudes are not directly subtractable. The paper reports a large sign flip for Ni after subtraction (from +0.041 to -0.028); this is precisely the kind of change that a 10–20% over-subtraction of a strong positive self-torque could produce. The paper includes valuable independent checks, such as multiple FM types, harmonic Hall data, and a thermal-sink control, but those checks do not resolve the current-density normalization issue, and the thermal-sink result is itself based on the same unnormalized subtraction. A reanalysis with current-density rescaling is straightforward and would settle the matter. I therefore recommend keeping the reader's CONDITIONAL verdict, with the condition being the demonstration that the subtraction is performed after proper current-density normalization, or a revised conclusion if it is not.","tokens_in":12046,"tokens_out":5072,"duration_ms":57711,"concrete_test":"Use the measured resistivities and thicknesses to compute the FM current fraction in Ta(5)/Ni(t) and in the Ta(1)/Ni(t) control, and rescale each control S and A by j_FM^bilayer/j_FM^control before subtracting; then rebuild Fig. 3b and the harmonic-Hall analysis. If the corrected ξ for Ta/Ni stays near -0.028, the concern is settled. If it moves toward zero or becomes positive, the reported 'absence' is an artifact of over-subtraction.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The decisive evidence for a negative Ta/Ni torque is the subtraction of the FM self-induced ST-FMR signal (Fig. 3a-b). The control is an FM single layer with only 1 nm Ta, whereas the bilayer has 5 nm Ta. The S and A amplitudes extracted by Eq. (1) are linear in the rf current density in the FM. Because the conductive Ta layer shunts part of the rf current, the FM current density in the bilayer is smaller than in the control. For the reported Ta resistivity (200 μΩ cm) and a Ni thickness of 2–4 nm, the Ni current fraction is roughly 80–90% in the bilayer versus 95%+ in the control. The paper does not state that the control S and A signals were scaled to the bilayer FM current density before subtraction. Over-subtracting a large positive Ni self-torque would shift the residual Ta contribution negative, possibly explaining the observed sign reversal from +0.041 to -0.028. The same issue would affect the harmonic-Hall correction if similarly unnormalized. Thus, unless the subtraction was current-density-normalized, the central claim is not established.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports ST-FMR and harmonic-Hall measurements on Ta(5 nm)/FM bilayers with several ferromagnets (Ni, Ni81Fe19, Fe, Fe60Co20B20, Fe50Pt50) and on control FM single layers grown on a 1-nm Ta adhesion layer. The authors find that, without correction, Ta/Ni yields a positive dampinglike torque efficiency of about +0.041, whereas the other Ta/FM systems yield negative values around -0.03. They attribute the positive Ni value to a thickness-dependent self-induced ST-FMR signal of the Ni layer itself. After subtracting the self-induced S and A signals measured on the single-layer controls, they obtain a negative Ta-contributed efficiency of about -0.028 for Ta/Ni, similar to the other FMs, and conclude that the orbital Hall effect of Ta makes no detectable contribution to the interfacial torque, so that previously reported positive orbital torques in Ta/Ni are artifacts.","tokens_in":12218,"tokens_out":2722,"duration_ms":29337,"significance":"If the central claim holds, the paper would resolve a prominent controversy by showing that the positive dampinglike torques reported for Ta/Ni bilayers in the orbital-torque literature arise from an overlooked self-induced bulk torque in the Ni layer, not from orbital-to-spin conversion. The paper's strength is its multi-pronged experimental approach: ST-FMR across five FM types and a range of thicknesses, a Si3N4 thermal-sink test for anomalous Nernst contributions, a power-dependence check, and independent harmonic-Hall measurements. The conclusion is falsifiable and quantitatively testable. However, the load-bearing subtraction step that separates the Ta contribution from the self-induced FM contribution rests on assumptions about the equivalence of the control and bilayer samples that are not yet validated in the manuscript.","major_comments":[{"comment":"The central subtraction of the single-layer FM self-induced S and A signals from the bilayer signals requires that the two measurements be normalized to the same radio-frequency current density in the ferromagnet. The control samples have only 1 nm Ta, while the bilayers have 5 nm Ta, so the FM current fraction differs between the two stacks; with the reported Ta resistivity of 200 µΩ cm and Ni thicknesses of 2-4 nm, the Ni current density in the bilayer is roughly 10-20% lower than in the control. The manuscript does not state that the control S and A amplitudes were scaled to the bilayer FM current density before subtraction. Without such normalization, subtracting the large positive Ni self-torque from the bilayer would over-subtract and could artificially shift the residual Ta contribution negative. The same issue applies to the harmonic-Hall correction described in Supplementary Note 1. This is the decisive step for the paper's main claim and must be documented and justified.","section":"Methods and Fig. 3a-b"},{"comment":"The subtraction assumes that the self-induced ST-FMR signal measured in a single FM layer grown on a 1-nm Ta adhesion layer is identical in amplitude and phase to the self-induced signal inside the Ta(5 nm)/FM bilayer. However, the 5-nm Ta underlayer can change the texture, strain, interface electronic structure, and possibly the magnetic damping of the FM compared with the 1-nm-Ta control. The manuscript offers no independent validation of this equivalence, such as a control with 5 nm Ta on the back side of the substrate or a comparison of magnetic properties, damping, or FMR linewidth between the control and bilayer samples. If the self-induced signal differs between the two stacks, the residual S and A attributed to Ta are not established as the true spin Hall torque.","section":"Fig. 3 and Supplementary Note 2"},{"comment":"Equation (2) defines ξ_FMR using the heavy-metal thickness d_HM, but for the single-layer control samples there is no heavy-metal layer, and the text does not specify what thickness is inserted when computing the apparent self-induced ξ_FMR for those controls. Since the correction procedure depends on converting S and A into efficiency units before subtraction, the exact convention used for the control samples must be stated explicitly; otherwise the subtracted quantities are ambiguous.","section":"Eq. (2) and Fig. 3a"}],"minor_comments":[{"comment":"The phrase 'triggers bloomed interest' in the opening sentence is ungrammatical and should be revised.","section":"Introduction"},{"comment":"The figure caption should state how many devices and how many field sweeps contribute to each standard deviation, and whether the error bars include the uncertainty from the linear fits in Fig. 1d and Fig. 3b.","section":"Fig. 2a"},{"comment":"The placeholder 'available at xxx' for the Supplementary Information should be replaced with a working link or reference before publication.","section":"Additional information"},{"comment":"The word 'unambiguous' in the abstract overstates the strength of the evidence given that the key subtraction relies on assumptions that are not yet validated; a more measured claim would better match the presented data.","section":"Abstract and Discussion"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within scope for a condensed-matter physics journal and addresses a topical controversy. The missing current-density normalization for the subtraction is a fixable but essential gap: the authors should either provide the normalization explicitly, or perform a control experiment that removes the shunting ambiguity, before the central claim can be accepted. If the authors can show that the negative Ta/Ni efficiency survives proper normalization, the paper would be a strong contribution."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper makes a genuinely new claim: the positive dampinglike torque efficiency in Ta/Ni, previously read as evidence for orbital current torque, is actually a self-induced ST-FMR signal from the Ni layer. That is a specific, testable idea, and the authors go after it with a systematic dataset—five ferromagnets, a thickness series, harmonic Hall measurements, a Si3N4 thermal sink, and a power-dependence check. If the central claim holds, it removes one of the most cited experimental pillars of the orbital-torque literature. Credit where it is due: the experiment is well designed, the controls are thoughtful, and the writing is clear about what the raw ST-FMR analysis can and cannot show.\n\nThe soft spot is the subtraction procedure. The paper subtracts the ST-FMR spectrum of a single FM layer (grown on only 1 nm Ta) from the bilayer spectrum (5 nm Ta) to isolate the Ta contribution. That subtraction assumes the self-induced signal in the control matches the self-induced component in the bilayer. The paper does not report normalizing to the FM current density. Because 5 nm Ta shunts current away from the FM, the FM current fraction in the bilayer is visibly lower than in the control. If the control signal is larger than the self-induced component in the bilayer, subtracting it unnormalized over-subtracts, and a large positive self-torque could artificially shift the residual Ta torque negative. The sign reversal from +0.041 to −0.028 is exactly the direction this would produce. The harmonic Hall correction appears to have the same issue. This is not a minor detail; it is the load-bearing joint of the paper's core claim. The authors may well have normalized in the Supplementary Note, but the main text does not say so, and the data are not public.\n\nOther concerns are minor: the control uses 1 nm Ta rather than the bilayer's 5 nm Ta, which changes the interface environment, and the raw data are not deposited. But the paper is honest in scope, and the framework—separating bulk self-torque from interfacial torque—is a useful contribution even if the quantitative conclusion is not yet nailed down.\n\nWho this is for: spintronics experimenters who work on ST-FMR and orbital torque. It deserves a serious referee, because the claim is important and the experimental effort is substantial. But I would not cite it as evidence for the absence of orbital torque until the current-density normalization is either confirmed in the supplementary material or demonstrated in a follow-up. My own verdict is conditional, and I would push the editor to demand the normalization detail before acceptance.","headline":"A potentially important negative result on orbital torque in Ta/FM, but the central subtraction step lacks a reported current-density normalization that could flip the sign.","tokens_in":12783,"tokens_out":3248,"would_cite":false,"duration_ms":34143,"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":"The spin-orbit torque from tantalum in Ta/ferromagnet bilayers is uniformly negative and matches the spin Hall effect alone; the positive torque reported for Ta/Ni is a self-induced ST-FMR artifact of the Ni layer.","keywords":["spin-orbit torque","orbital Hall effect","orbital current torque","tantalum","spin-torque ferromagnetic resonance","self-induced ST-FMR","spin Hall effect","ferromagnet bilayer"],"falsifier":"A decisive check is to measure the Ta/Ni dampinglike torque using harmonic Hall voltage on the same films with the Ni self-torque separately characterized, or to grow Ta/Ni with a series of Ta thicknesses at fixed Ni thickness and perform the same single-layer subtraction; if the corrected efficiency depends on Ni thickness or Ta thickness in a way not matching the Ta spin Hall conductivity, the subtraction premise is falsified.","tokens_in":11809,"feed_emoji":"🧲","tokens_out":9419,"duration_ms":87697,"temperature":0.7,"pith_summary":"This paper aims to settle whether the orbital Hall effect of tantalum produces a detectable orbital-current torque on an adjacent ferromagnetic layer. By measuring spin-torque ferromagnetic resonance across five different ferromagnets and a wide thickness range, the authors find that the torque generated by Ta has essentially the same negative efficiency in every case, matching the sign and magnitude expected from Ta's spin Hall effect alone. They argue that the positive torque previously reported for Ta/Ni, often cited as evidence for orbital-current torque, is an artifact of a strong thickness-dependent self-induced ST-FMR signal inside the Ni layer itself. If correct, this removes a key experimental pillar of orbital-torque claims and restores the standard spin Hall picture for Ta/ferromagnet bilayers.","feed_headline":"Tantalum's orbital current exerts no torque on ferromagnets","feed_subtitle":"Five ferromagnets give the same negative spin Hall torque once Ni's self-induced signal is subtracted.","key_machinery":"The key mechanism is the self-induced bulk spin-orbit torque within the ferromagnetic layer itself, which produces a spurious symmetric and antisymmetric ST-FMR signal that is thickness-dependent. The measurement machinery is three-terminal spin-torque ferromagnetic resonance: the bilayer and control single-layer devices are driven by a radio-frequency current, the mixed voltage is fit to symmetric and antisymmetric Lorentzians, and the dampinglike efficiency is extracted from the inverse intercept of $1/\\xi_{\\mathrm{FMR}}$ versus $1/t_{\\mathrm{FM}}$. The load-bearing step is subtracting the single-layer FM's symmetric and antisymmetric components from the bilayer's components before computing the Ta-only efficiency; without this subtraction, the Ni self-torque dominates the apparent Ta/Ni torque and flips its sign.","core_discovery":"The central claim is that the dampinglike spin-orbit torque produced by a 5 nm Ta layer has essentially the same negative efficiency, about $-0.028$ after correction, for Ni, Ni$_{81}$Fe$_{19}$, Fe, Fe$_{60}$Co$_{20}$B$_{20}$, and Fe$_{50}$Pt$_{50}$, independent of ferromagnet type and thickness, and therefore originates solely from the negative spin Hall effect of Ta. The apparent positive efficiency of Ta/Ni bilayers in the Ni thickness range above about 2 nm is identified as an artifact: the Ni layer itself generates a strong self-induced ST-FMR signal of the same sign and comparable magnitude, and this signal grows with thickness. After subtracting the symmetric and antisymmetric responses of control single-layer ferromagnets from the bilayer responses, the Ta-contributed efficiency becomes uniformly negative and matches the non-Ni ferromagnets. Control experiments with a Si$_3$N$_4$ thermal sink and rf-power dependence exclude anomalous Nernst and spin-pumping explanations. The paper concludes that the orbital Hall effect of Ta, despite predicted conductivities 20 to 50 times larger than its spin Hall conductivity, makes no detectable contribution to the interfacial torque in Ta/ferromagnet systems.","pith_inferences":["A natural extension is to look for the orbital Hall effect of Ta directly in bulk transport, e.g., by detecting the inverse orbital Hall effect in a Cu or nonmagnetic detector layer, which would separate whether an orbital current exists in Ta from whether it exerts a torque on the adjacent ferromagnet.","The same single-layer subtraction logic could be applied to orbital-torque claims in other light metals such as Cr, Ti, and Zr, whose reported torque signs may also be contaminated by self-induced FM signals.","Because the control samples use a 1 nm Ta adhesion layer, a stricter control would eliminate Ta entirely, e.g., by growing the FM on a nonmetallic seed; if that changes the extracted residual, the subtraction is not exact.","Varying the Ta thickness while holding the Ni thickness fixed would provide a quantitative test: the corrected Ta efficiency should track the Ta spin Hall conductivity and stay independent of Ni thickness if the central claim holds."],"forward_implications":["The positive Ta/Ni dampinglike-torque efficiencies cited as evidence for orbital-current torque disappear once the self-induced ST-FMR signal of the Ni layer is subtracted, so those Ni-based claims need re-examination with single-layer controls.","The dampinglike torque from Ta remains negative and nearly constant across FM types and thicknesses, consistent with Ta's spin Hall effect being the sole source of the interfacial torque.","Extracted spin Hall conductivities of Ta from ST-FMR do not need an orbital-current correction of the kind suggested by the 20 to 50 times larger predicted orbital Hall conductivity.","ST-FMR thickness-series determinations of interfacial torques are unreliable for ferromagnets with strong bulk self-torque unless the ferromagnet's own signal is measured and subtracted, even for FM layers thicker than 10 nm."],"supporting_citations":[{"why":"Establishes tantalum as a negative spin Hall metal whose giant spin Hall effect generates dampinglike torques, the baseline against which the measured efficiency is compared.","marker":"[1]"},{"why":"Provides the inverse-thickness linear-fit method, $1/\\xi_{\\mathrm{FMR}}$ versus $1/t_{\\mathrm{FM}}$, used to extract the interfacial dampinglike efficiency.","marker":"[5]"},{"why":"Reports positive orbital torque in Ta/Ni magnetic bilayers, the main Ni-based claim being challenged.","marker":"[20]"},{"why":"Reports sign reversal of dampinglike torque in Ta/Ni and Nb/Ni bilayers, another Ni-based orbital-torque observation re-examined here.","marker":"[22]"},{"why":"Theory predicting that Ta's orbital Hall conductivity is 20 to 50 times larger and opposite in sign to its spin Hall conductivity, defining the expected orbital-torque signal.","marker":"[42]"},{"why":"Demonstrates strong bulk dampinglike spin-orbit torque in ferromagnetic single layers, the physical basis for the self-induced ST-FMR correction.","marker":"[45]"},{"why":"Introduces spin-torque ferromagnetic resonance as a way to detect spin Hall torques, the core measurement technique.","marker":"[46]"},{"why":"First-principles calculation of orbital relaxation length used to show that the short-relaxation assumption required by the positive Ni result conflicts with other orbital-current claims.","marker":"[17]"}],"fun_headline_variants":["Ta orbital current torque ruled out in FM bilayers","Orbital Hall torque from Ta absent in ferromagnets","Spin Hall effect only: no orbital torque in Ta/FM","Ta's orbital current gives no torque on ferromagnets","No orbital current torque in Ta/ferromagnet stacks"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The subtraction step assumes that the spin-torque signal produced by a bare ferromagnet layer grown on just 1 nm of tantalum is identical, in size and phase, to the ferromagnet's own signal when it sits on the 5 nm tantalum layer being studied.","fun_headline_variants_meta":{"raw":{"variants":["Ta orbital current torque ruled out in FM bilayers","Orbital Hall torque from Ta absent in ferromagnets","Spin Hall effect only: no orbital torque in Ta/FM","Ta's orbital current gives no torque on ferromagnets","No orbital current torque in Ta/ferromagnet stacks"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001022,"raw_usage":{"total_tokens":4359,"prompt_tokens":1041,"completion_tokens":3318,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":657,"completion_tokens_details":{"reasoning_tokens":3234}},"tokens_in":657,"tokens_out":3318,"duration_ms":19608,"temperature":1.0,"reasoning_tokens":3234,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T19:16:43.993395+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A decisive check is to measure the Ta/Ni dampinglike torque using harmonic Hall voltage on the same films with the Ni self-torque separately characterized, or to grow Ta/Ni with a series of Ta thicknesses at fixed Ni thickness and perform the same single-layer subtraction; if the corrected efficiency depends on Ni thickness or Ta thickness in a way not matching the Ta spin Hall conductivity, the subtraction premise is falsified.","supporting_citations":[{"cited_title":"Liu, C.-F","cited_arxiv_id":null,"evidence_quote":"Establishes tantalum as a negative spin Hall metal whose giant spin Hall effect generates dampinglike torques, the baseline against which the measured efficiency is compared."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the inverse-thickness linear-fit method, $1/\\xi_{\\mathrm{FMR}}$ versus $1/t_{\\mathrm{FM}}$, used to extract the interfacial dampinglike efficiency."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reports positive orbital torque in Ta/Ni magnetic bilayers, the main Ni-based claim being challenged."},{"cited_title":"Dutta and A","cited_arxiv_id":null,"evidence_quote":"Reports sign reversal of dampinglike torque in Ta/Ni and Nb/Ni bilayers, another Ni-based orbital-torque observation re-examined here."},{"cited_title":"Kontani, T","cited_arxiv_id":null,"evidence_quote":"Theory predicting that Ta's orbital Hall conductivity is 20 to 50 times larger and opposite in sign to its spin Hall conductivity, defining the expected orbital-torque signal."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Demonstrates strong bulk dampinglike spin-orbit torque in ferromagnetic single layers, the physical basis for the self-induced ST-FMR correction."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces spin-torque ferromagnetic resonance as a way to detect spin Hall torques, the core measurement technique."},{"cited_title":"Rang and P","cited_arxiv_id":null,"evidence_quote":"First-principles calculation of orbital relaxation length used to show that the short-relaxation assumption required by the positive Ni result conflicts with other orbital-current claims."}],"review_version":1}