{"id":"0fc3e523-2a69-4c3f-bd78-e11f25f21ba1","arxiv_id":"2511.20379","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"GdFeCo's spin Hall and spin anomalous Hall torques retain their signs through the magnetization compensation temperature, with opposite signs and distinct proposed sublattice origins.","lead":"This paper measured how spin currents generated by the ferrimagnet GdFeCo change across its magnetization compensation temperature. It found that two different spin-current mechanisms keep their sign and oppose each other, and proposes they come from different magnetic sublattices.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"SAHE sign retention across T_M is extrapolated from only two far endpoints (15 K and 300 K); no direct measurement at intermediate temperatures supports the central claim.","rationale":"The reader's verdict is CONDITIONAL, and I agree that the central claim is not fully established. The reader's weakest_assumption highlights both the out-of-plane exactness and the endpoint-only SAHE data. I find the endpoint interpolation the more load-bearing concern because even if the out-of-plane assumption holds perfectly near T_M, the SAHE sign across T_M is still not directly measured—only inferred from 15 K and 300 K. The out-of-plane assumption is likely good near T_M because the anisotropy field diverges, and its failure would mainly affect the SHE sign extraction, which is already supported by the consistent positive lineshape signal across all temperatures. The SAHE sign, by contrast, has no direct supporting data in the intermediate range. The authors' physical explanation for the absence of reversal is explicitly acknowledged to be in tension with Hund's third rule, so it does not provide independent confirmation. Therefore the strongest interpretation—that both mechanisms retain their sign across T_M and that prior self-torque sign changes are due to absorption—goes beyond what the measurements alone establish. A CONDITIONAL verdict is appropriate: the experimental observations are plausible and clearly presented, but the SAHE sign-retention claim should be either directly measured or explicitly framed as an extrapolation supported by a more robust model.","tokens_in":13617,"tokens_out":11464,"duration_ms":125948,"concrete_test":"Extend the DC-bias ST-FMR measurements into the parallel configuration at intermediate temperatures (e.g., 100 K, 120 K, 180 K, 200 K) using a high-field in-plane setup capable of overcoming the GdFeCo anisotropy field slightly away from T_M. If the combined DL efficiency ξ^{SHE+SAHE}_DL remains negative at these temperatures, with the SHE contribution independently fixed as positive from the lineshape analysis, the SAHE sign-retention claim is supported. If the combined sign changes, the claim is falsified. Alternatively, use harmonic Hall voltage measurements with the GdFeCo layer saturated in-plane across the same temperature range to extract the SAHE sign directly.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim requires that the SAHE-driven DL torque retains its sign across the magnetisation compensation temperature. The DC-bias technique yields a combined SHE+SAHE signal only in the parallel configuration at 15 K and 300 K, the two endpoints where GdFeCo can be saturated in-plane. Near T_M (120–160 K), the GdFeCo magnetisation is out-of-plane, so only the SHE contribution is accessible; the SAHE contribution is not measured there. The SAHE sign across T_M is therefore inferred from two far-separated temperatures on either side of compensation, not from direct observation. This extrapolation is not independently supported by the proposed sublattice model: the authors explicitly acknowledge in Section V A that the required λ_SOC signs conflict with Hund's third rule. Moreover, an even number of sign reversals between 15 K and 300 K would be entirely invisible to the endpoint data. The paper's further claim that previously reported self-torque sign changes arise from spin-current absorption rather than generation also extends beyond the measured torque on NiFe, whose absorption properties may differ from those of GdFeCo itself. The SHE sign retention is well supported by the lineshape analysis across T_M; the SAHE sign retention is the load-bearing, under-supported part of the central claim.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports ST-FMR measurements on a GdFeCo(10)/Cu(4)/NiFe(4) heterostructure from 15 K to 300 K, with a magnetisation compensation temperature T_M ≈ 146 K. Lineshape analysis is used to isolate the SHE contribution, while DC-bias measurements detect the combined SHE+SAHE contributions. The central claim is that both the SHE- and SAHE-driven torque efficiencies retain their sign across T_M, and that the SAHE-driven damping-like torque is opposite in sign to the SHE-driven term and dominates in the parallel configuration. A sublattice model is proposed in which the SHE originates from Gd 5d electrons and the SAHE from FeCo 3d electrons with opposite spin-orbit coupling signs. The paper further suggests that previously reported sign changes of the self-torque in single-layer GdFeCo are due to variations in spin-current absorption rather than to sign changes of the generated spin currents.","tokens_in":13983,"tokens_out":5445,"duration_ms":61957,"significance":"If the sign-retention claim holds, this is a valuable advance: it would distinguish spin-current generation from spin-current absorption in ferrimagnets and identify sublattice-specific SHE/SAHE channels. The paper has real strengths: complementary lineshape and DC-bias techniques, a Pt/NiFe control sample, consistent positive SHE signs from lineshape analysis over the entire temperature range, and an explicit discussion of the limitations of the proposed model, including the conflict with Hund's third rule. The main weakness is that the SAHE sign across T_M is inferred from only two endpoint temperatures in the parallel configuration, so the headline claim is not directly supported by the data.","major_comments":[{"comment":"The claim that the SAHE-driven DL torque retains its sign across T_M rests on only two endpoint measurements in the parallel configuration: ξ_DL^{SAHE+SHE} = −0.20±0.03 at 15 K and −0.45±0.07 at 300 K. No SAHE data are reported at intermediate temperatures, and in the perpendicular configuration the SAHE contribution vanishes by symmetry (Eq. 5). Thus an even number of sign reversals between 15 K and 300 K would be entirely invisible to this dataset. The wording in §V A ('the SAHE sign also remains unchanged across both the magnetisation and the angular compensation points') is too strong; at minimum the claim should be softened to 'consistent with no sign change between the two endpoints', or additional in-plane measurements at intermediate temperatures should be provided.","section":"§IV B, Fig. 5, Table I"},{"comment":"The separation of SHE and SAHE in the perpendicular configuration assumes that the GdFeCo magnetisation is exactly out-of-plane, so that the SAHE spin current component along z vanishes. The paper itself raises the possibility of 'a small canting of the magnetisation away from the out-of-plane direction' to explain the amplitude enhancement of ξ_DL near T_M. If such a canting exists, the SAHE contribution is not strictly zero and would enter the 'SHE-only' lineshape/DC-bias data. Since the SAHE DL torque is opposite in sign and can dominate (as at 15 K and 300 K), even a modest canting could affect the extracted SHE sign near T_M. The authors should quantify the maximum permissible canting angle consistent with their measurements, or provide direct evidence (e.g. angle-dependent Hall/FMR data within the ST-FMR field range) that the magnetisation is out-of-plane to within the required acc","section":"§II, Fig. 2(c), Eq. (5), §V B"},{"comment":"In the parallel configuration the DC-bias linewidth measurement gives only the sum ξ_DL^{SAHE} + ξ_DL^{SHE}; an independent SHE value at those temperatures is not reported (Table I lists '-' under ξ_DL^{SHE} for 15 K and 300 K). The decomposition into an opposite-sign SAHE term therefore relies on an assumed or interpolated SHE contribution, and the propagation of uncertainties from the SHE reference into the inferred SAHE value is not shown. The problem is compounded near T_M, where the SHE-only values have large or unstated uncertainties (+0.05±0.06 at 120 K; 'unclear' at 150 K). The paper should report the individual SHE and SAHE efficiencies with full uncertainty propagation at each temperature, or explicitly state the interpolation and its uncertainties.","section":"§III C, Eq. (7), Table I"}],"minor_comments":[{"comment":"Grammar: 'both effects originates' should be 'both effects originate'; §I has 'producting'; §III B has 'is not consider as a reliable method'.","section":"Abstract and §V"},{"comment":"The LS sign entries are reported without confidence or error bars, and the 150 K DC-bias entry is listed as 'unclear'. Providing a measure of sign confidence (e.g. sign of V_S/V_A with uncertainty, or number of repeated spectra) would strengthen the sign-retention argument.","section":"Table I"},{"comment":"The statement that previously reported self-torque sign changes 'can be driven by variations in spin current absorption' is an extrapolation from torques measured on a Cu/NiFe detector to self-torques in GdFeCo itself. It is a reasonable hypothesis, but it should be explicitly labelled as such, since the measured absorption properties are those of NiFe, not necessarily of GdFeCo.","section":"§VI"},{"comment":"The proposed sublattice model is post hoc and, as the authors acknowledge, the inferred λ_SOC signs conflict with Hund's third rule. This is not a fatal flaw, but it means the model does not provide independent support for the sign-retention claim; a reformulation or a caveat that the model is speculative would help.","section":"§V A, Fig. 6"}],"recommendation":"major_revision","confidential_remarks":"The manuscript fits the journal's scope and the central idea is interesting. The main barrier to acceptance is not the model but the evidential support for the headline SAHE-sign-retention claim: it rests on two endpoint measurements with no direct observation across T_M. If the authors can add intermediate-temperature data or conservatively reframe the claim, I would be satisfied. I see no circularity or misrepresentation of prior work."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know. This is a clean temperature-resolved decomposition of SHE and SAHE spin currents generated by GdFeCo, using a NiFe detector separated by Cu, and the SHE sign retention across T_M is well supported. The second thing: the SAHE sign-retention claim, which is the abstract's headline, is inferred from two in-plane endpoint temperatures, 15 K and 300 K. There is no SAHE data between them, so the central claim is weaker than the abstract suggests.\n\nThe experiment is well designed. The Cu spacer magnetically decouples the layers, so the NiFe resonance probes spin currents emitted by GdFeCo rather than self-torques inside the ferrimagnet. That is a real advance over the earlier GdFeCo self-torque measurements. The lineshape analysis isolates SHE, and the DC-bias analysis sees SHE plus SAHE; the combination is enough to fix the relative sign. The signs are consistent across all measured temperatures: SHE DL and FL' positive, combined SHE+SAHE DL negative at the two in-plane endpoints, so SAHE DL is opposite to SHE and dominant there. That part is solid.\n\nThe authors are also honest about limitations. The sublattice explanation is explicitly called a suggestion; they flag the Hund's third rule discrepancy; they mention canting as an alternative explanation for the amplitude trend. No circularity: the efficiencies come from standard ST-FMR susceptibility equations, and the self-citations supply measurement details, which is appropriate.\n\nNow the soft spots. The SAHE sign across T_M is extrapolated, not observed. An even number of sign reversals between 15 K and 300 K would be invisible. The perpendicular-configuration analysis assumes the GdFeCo magnetization is exactly out-of-plane near T_M; canting would mix SAHE into the \"SHE-only\" data. The authors acknowledge canting but do not quantify its effect on the sign conclusions. And the final claim that earlier self-torque sign changes in GdFeCo come from absorption rather than generation goes beyond what this NiFe-detector measurement can prove; NiFe may absorb spin currents differently than GdFeCo.\n\nBottom line: this is a worthwhile experimental paper for the ferrimagnet SOT community. It deserves a serious referee. The referee should ask the authors to either provide SAHE data at intermediate temperatures or reframe the SAHE sign-retention and absorption claims as provisional. I would cite it for the SHE result and the method.","headline":"Clean SHE/SAHE spin-current decomposition in GdFeCo, but the SAHE sign-retention headline rests on two endpoint temperatures and should be treated as provisional.","tokens_in":14436,"tokens_out":3278,"would_cite":true,"duration_ms":35929,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper argues that GdFeCo's spin-Hall and spin-anomalous-Hall torques keep their signs across the magnetisation compensation temperature, with the SAHE torque opposite in sign to the SHE torque.","keywords":["spin current","spin Hall effect","spin anomalous Hall effect","ferrimagnet","GdFeCo","spin-torque ferromagnetic resonance","magnetisation compensation","spin-orbit torque"],"falsifier":"A decisive check would be to measure the DC-bias damping-like torque at intermediate temperatures (for instance 175–275 K) while applying a magnetic field strong enough to pull the GdFeCo magnetisation fully in plane, or to detect any in-plane canting near T_M with Hall or torque magnetometry; if the extracted SAHE sign flips at some intermediate temperature, or the 'SHE-only' signal shows SAHE contamination, the conclusion would fail.","tokens_in":13536,"feed_emoji":"🧲","tokens_out":6926,"duration_ms":68179,"temperature":0.7,"pith_summary":"This paper tries to establish that a ferrimagnet's own spin currents do not change sign when the material passes through its magnetisation compensation temperature, where the net moment vanishes. Using spin-torque ferromagnetic resonance on a GdFeCo/Cu/NiFe stack, the authors separate two spin-current generation mechanisms: the ordinary spin Hall effect (SHE) and the spin anomalous Hall effect (SAHE), in which the spin polarisation is tied to the magnetisation direction. They find that both torque signs stay fixed through compensation, that the SAHE-driven damping-like torque is opposite in sign to the SHE-driven one and dominates far from compensation, and that this can be understood if the SHE originates from Gd 5d electrons and the SAHE from FeCo 3d electrons. If correct, the result reframes earlier reports of sign-changing self-torques in ferrimagnets as absorption effects rather than emission effects, which matters for designing spin-torque devices whose torque direction is temperature stable.","feed_headline":"Spin torques from GdFeCo keep their sign across compensation point","feed_subtitle":"Opposite, temperature-stable torque signs trace to Gd 5d and FeCo 3d electrons, not to net magnetization.","key_machinery":"The key mechanism is the geometrical separation of the spin Hall effect (SHE, charge-to-spin conversion) and the spin anomalous Hall effect (SAHE, magnetisation-tied spin polarisation) using a GdFeCo/Cu/NiFe heterostructure and spin-torque ferromagnetic resonance. Near the compensation temperature (T_M ≈ 146 K) the anisotropy of GdFeCo pins its magnetisation out of plane, so the SAHE spin current—whose polarisation follows the magnetisation—has no component along the detection axis and only the SHE torque appears (perpendicular configuration); away from compensation the magnetisation is pulled in plane and both effects contribute (parallel configuration). Lineshape analysis reads only the SH","core_discovery":"The paper's central claim is that both spin-current generation channels in the ferrimagnet GdFeCo—the spin Hall effect (SHE) and the spin anomalous Hall effect (SAHE)—produce torques whose signs stay fixed across the magnetisation compensation temperature, even though the net magnetisation reverses. In the GdFeCo/Cu/NiFe devices the SHE damping-like and field-like torques are positive at all temperatures, while the SAHE contribution, seen only in the parallel configuration far from compensation, is opposite in sign to the SHE term and larger in magnitude; because the combined sign is the same at 15 K and 300 K, the SAHE does not invert either. The explanation is that the SHE originates from","pith_inferences":["If the sublattice decomposition holds, measurements of the two torque signs on a series of Gd_x(FeCo)_{1-x} alloys should show the SAHE/SHE amplitude ratio tracking the FeCo fraction; that is a direct, testable extension the paper does not report.","The same two-component picture might explain sign behaviour in other RE-TM ferrimagnets, such as FeTb, where multiple self-torque reversals were reported—if those reversals are set by sublattice composition gradients rather than by compensation points, the present framework gives a way to separate those contributions.","A quantitative implication left implicit is that the SAHE torque efficiency should scale with the magnetisation direction of the FeCo sublattice rather than with the net magnetisation; an experiment measuring torque amplitude as a function of a controlled in-plane field angle near T_M could confirm this."],"forward_implications":["If the sign of generated spin currents is fixed across compensation, the sign changes of damping-like self-torque previously observed in single-layer GdFeCo and GdFeCo/Cu bilayers must be attributed to changes in spin-current absorption, not to the emission process.","Ferrimagnetic spin sources can exert torques on an adjacent detector layer whose sign is robust to temperature drift through compensation, which is useful for spin-torque devices operated over wide temperature ranges.","Because the SHE and SAHE channels have opposite signs, the net damping-like torque can be switched from positive to negative by moving between the out-of-plane (SHE-only) and in-plane (SHE+SAHE) configurations, giving a temperature-controlled torque sign switch.","The sublattice assignment predicts that changing the Gd:FeCo ratio or substituting the rare earth should change the relative weights of the two channels while preserving their individual signs.","The SAHE dominance of the damping-like torque far from compensation means that single-layer ferrimagnet self-torque devices will be governed by the FeCo 3d channel, not by the net magnetisation."],"fun_headline_variants":["GdFeCo torques: both spin Hall channels defy magnetization reversal","Compensation point fails to flip GdFeCo spin torque signs","Spin torques from GdFeCo keep sign across compensation","GdFeCo spin currents: sign stable, source split between sublattices"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"Load-bearing premise: near the compensation temperature, where the GdFeCo magnetisation is pinned out of plane by its anisotropy, the SAHE spin current along the detection axis is exactly zero—any small in-plane canting mixes SAHE torque into the 'SHE-only' data and could change the sign conclusion, and the SAHE sign is inferred from only two endpoint temperatures.","fun_headline_variants_meta":{"raw":{"variants":["GdFeCo torques: both spin Hall channels defy magnetization reversal","Compensation point fails to flip GdFeCo spin torque signs","Spin torques from GdFeCo keep sign across compensation","GdFeCo spin currents: sign stable, source split between sublattices"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000354,"raw_usage":{"total_tokens":1781,"prompt_tokens":785,"completion_tokens":996,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":529,"completion_tokens_details":{"reasoning_tokens":932}},"tokens_in":529,"tokens_out":996,"duration_ms":8617,"temperature":1.0,"reasoning_tokens":932,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T20:15:19.825379+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A decisive check would be to measure the DC-bias damping-like torque at intermediate temperatures (for instance 175–275 K) while applying a magnetic field strong enough to pull the GdFeCo magnetisation fully in plane, or to detect any in-plane canting near T_M with Hall or torque magnetometry; if the extracted SAHE sign flips at some intermediate temperature, or the 'SHE-only' signal shows SAHE contamination, the conclusion would fail.","supporting_citations":[],"review_version":1}