{"id":"08708467-c838-4c2e-afc3-1550bc3bffba","arxiv_id":"2507.10219","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"PtBi alloys with a few percent Bi show a threefold larger spin Hall efficiency than pure Pt and 32-42% lower threshold current in spin Hall nano-oscillators.","lead":"This paper reports that adding small amounts of bismuth to platinum thin films increases the efficiency of converting charge current into spin current by more than three times, and reduces the threshold current needed to drive nanoscale microwave oscillators by up to 42 percent. The result points to a practical way to make spin-based memory and neuromorphic computing devices more energy-efficient.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The absolute θSH enhancement is not isolated from the Ta seed layer: as PtBi resistivity rises, current redistributes into Ta, which can mimic a Bi-dependent torque increase; a Ta-free control is needed.","rationale":"The reader's weakest assumption is exactly the load-bearing issue. The paper's headline novelty is the >3x increase in θSH (0.07 to 0.24) and the corresponding 42% Ith reduction. The Ith comparison is a direct device-level measurement and is less sensitive to the attribution problem. But the absolute θSH claim depends on the assumption that all detected damping modulation comes from PtBi bulk SHE. The 2.4 nm Ta seed is a plausible additional source of spin current, and the problem is aggravated by the large resistivity change with Bi doping, which changes current sharing. This is a concrete, falsifiable concern rather than a disagreement with consensus. The angular symmetry check and pure-Pt reference are useful but do not resolve it. The four-point side-jump mechanism fit is also weak, but it is secondary to the main quantitative claim. I find no reason to change the reader's conditional verdict: the paper should be accepted only after the Ta contribution is bounded or eliminated experimentally.","tokens_in":16381,"tokens_out":5811,"duration_ms":82073,"concrete_test":"Fabricate the same Pt94Bi6(4 nm)/CoFeB(3 nm) stack on a non-spin-active seed (e.g., 2.4 nm MgO or SiN) and repeat the DC-bias ST-FMR linewidth-modulation measurement and Eq. 4 extraction. If θSH remains approximately 0.24, the Ta seed is not responsible; if it drops substantially, the reported absolute θSH and its Bi-content trend are inflated by Ta. A complementary Ta-thickness series (0–4 nm) under fixed Pt/CoFeB would directly quantify the Ta torque contribution.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claim is θSH(Pt94Bi6) = 0.24 versus 0.07 for Pt. The extraction (Eq. 4) attributes the entire DC-bias ST-FMR linewidth modulation to a spin current generated in the 4 nm PtBi layer, with jdc,HM obtained from a parallel-resistor model. However, every stack contains a 2.4 nm Ta seed (Section 4.1, Figs. 1b–c). Ta is a known spin-Hall metal; even though it is separated from CoFeB by 4 nm PtBi, a fraction of its spin current can reach the ferromagnet, and the paper neither subtracts nor bounds this channel. The bias is not simply a constant additive offset: ρxx rises from 65 μΩ·cm (pure Pt) to 270–301 μΩ·cm with Bi content, so the same total Idc repartitions more current into Ta as Bi increases. A Ta-derived antidamping torque would therefore grow relative to the PtBi torque, partly mimicking the reported θSH versus Bi trend. The sin2φ cosφ angular dependence (Supporting S4) is consistent with bulk SOT symmetry but cannot distinguish PtBi spin current from Ta/PtBi spin transport. The pure-Pt stack also contains Ta, so its 0.07 value does not anchor the absolute scale. This leaves the more-than-threefold absolute enhancement insufficiently established as a property of PtBi, although the device-level Ith reduction is a more robust comparative result because both SHNOs share the same seed layer.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports spin-torque ferromagnetic resonance (ST-FMR) and spin Hall nano-oscillator (SHNO) measurements on electron-beam co-evaporated Pt100−xBix/Co40Fe40B20 stacks with x = 0, 3.9, 6.0, and 8.7 at.%. The authors report that alloying Pt with Bi increases the spin Hall efficiency θSH from 0.07 in pure Pt to 0.24 in Pt94.0Bi6.0 and 0.19 in Pt91.3Bi8.7, as extracted from DC-bias ST-FMR linewidth modulation. They attribute this enhancement to bulk extrinsic side-jump scattering based on a ρSH_imp ∝ ρ^2_imp scaling analysis. In 100 nm SHNOs they observe a 42% reduction in threshold current Ith for Pt94.0Bi6.0 (from 1.30 mA to 0.75 mA) and a 32% reduction for Pt91.3Bi8.7 (to 0.89 mA), accompanied by higher output power and narrower linewidth. Structural characterization by GIXRD and TEM shows reduced Pt crystallinity and uniform Bi distribution with slight interfacial enrichment. The paper concludes that PtBi alloys are a promising low-resistivity spin-current source for energy-efficient SOT-MRAM and neuromorphic SHNO applications.","tokens_in":16755,"tokens_out":3981,"duration_ms":43463,"significance":"If the absolute θSH values are correct, the reported threefold enhancement over pure Pt at a modest resistivity increase would be a practically useful result for SOT-MRAM and SHNO applications, and the comparative Ith reduction in identical device geometries is a meaningful demonstration. The manuscript has several strengths: it combines structural, spin-torque, and device-level measurements; the angular dependence of the ST-FMR signal is checked; and the threshold-current comparison is internally consistent and directly relevant to applications. However, the central quantitative claim—that the absolute θSH increases from 0.07 to 0.24 as a property of the PtBi layer—depends on an extraction that does not account for the spin-Hall-active Ta seed layer, and the side-jump mechanism is inferred from a four-point scaling analysis with one forced origin. These issues make the paper's main quantitative claim not yet fully established, although the comparative trends are likely robust.","major_comments":[{"comment":"The absolute θSH values are extracted by attributing the entire DC-bias linewidth modulation to a spin current generated in the 4 nm PtBi layer, with jdc,HM obtained from a parallel-resistor model that includes the Ta seed layer only as a resistive shunt. However, every stack contains a 2.4 nm Ta seed (Section 4.1; Figs. 1b,c) that is itself a spin-Hall metal and is not accounted for as a torque source. Because ρxx of the PtBi layer rises from 65 to 270–301 μΩ·cm with Bi content, the same total current redistributes increasingly into Ta, so a Ta-derived antidamping torque would grow relative to the PtBi torque and could mimic part of the reported θSH-versus-Bi trend. The pure-Pt stack also contains Ta, so the 0.07 reference value does not anchor the absolute scale. The sin2φcosφ angular dependence (Supporting S4) has the same symmetry for interfacial and Ta-generated torques and therefore cannot rule out this channel. A Ta-free control or a quantitative bound on the Ta torque contribution is needed to support the threefold absolute enhancement claim; without it, only the relative trend and the Ith reduction are established.","section":"Section 2.2, Eq. (4); Section 4.1, Figs. 1b–c"},{"comment":"The claim that the enhanced θSH originates from extrinsic side-jump scattering rests on the ρSH_imp ∝ ρ^2_imp scaling shown in Fig. 5f and the extraction described in Supporting S10. This extraction subtracts an assumed composition-independent intrinsic SH contribution equal to that of pure Pt, and it defines ρSH_imp for pure Pt to be zero by construction. With only four data points, one of which is this forced origin, the quadratic fit is not a strong test of the mechanism; in particular, the non-monotonic θSH (0.24 at x=6.0 versus 0.19 at x=8.7) is not reflected in the scaling analysis, and alternative mechanisms (e.g., disorder-enhanced intrinsic SHE) are not quantitatively excluded. A direct test, such as comparing the resistivity scaling of the anomalous Hall effect or measuring a wider composition series, would be needed to establish side-jump dominance.","section":"Section 2.4, Fig. 5f"},{"comment":"The reported θSH=0.19 for Pt91.3Bi8.7 is not reconciled with the previously reported value of 0.10 for Pt92Bi8 cited as Ref. [41] (Hong et al., Adv. Electron. Mater. 2018). Both are nearly the same composition, yet the present value is roughly twice as large. The discrepancy may arise from different measurement techniques (inverse spin Hall effect versus DC-bias ST-FMR) or from the Ta seed or interface differences, but the manuscript does not discuss it. Since the absolute magnitude of θSH is central to the paper's main claim, this omission leaves the reader unable to assess systematic uncertainty in the extraction.","section":"Section 2.2, Fig. 3; Ref. [41]"}],"minor_comments":[{"comment":"Fig. 5 caption labels the composition axis as 'x in Pt100−xBix (wt.%)', but the text and deposition description use atomic percent; the unit should be corrected for consistency.","section":"Fig. 5 caption"},{"comment":"There is a typo in the phrase 'Both stacks exhibit well-defined, uniform layers tructures' — 'tructures' should be 'structures'.","section":"Section 2.1"},{"comment":"The inline notation in the Fig. 3 panels (e.g., '≈ 0.07 ± 0.01q SH' and 'q SH') appears to be a rendering artifact of the θSH symbol; the figure should be regenerated with consistent mathematical notation.","section":"Fig. 3"},{"comment":"The statement that 'the output power is visibly higher at 6% and 8.7% Bi' would be more convincing if quantified in the main text rather than referenced only to Supporting S9.","section":"Section 2.4"}],"recommendation":"major_revision","confidential_remarks":"The manuscript falls within the journal's scope and the device-level comparative results are valuable. The main weakness is the absolute θSH scale: the Ta seed layer is not treated as a torque source, and the discrepancy with the prior Pt92Bi8 value (Ref. [41]) is unaddressed. The side-jump mechanism claim is also based on a four-point analysis with a forced origin. These issues are potentially fixable with additional control experiments or a careful error budget, so major revision rather than rejection seems appropriate."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: the paper gives useful comparative data—Bi alloying cuts SHNO threshold current by 32–42% in the same stack geometry—but the headline absolute θSH values are less secure than the abstract implies. The Ta seed layer, present in every stack, is not subtracted out, and because ρxx rises sharply with Bi content the current redistribution into Ta can mimic part of the reported θSH trend. That said, this is a solid experimental paper with real measurements, not a toy or a fit to a conclusion.\n\nWhat's actually new: new measured θSH for Pt100-xBix at x=3.9, 6.0, and 8.7, plus SHNO threshold data on the same films. The structural work (GIXRD, TEM, EDS) is careful, and the EDS shows uniform Bi with slight interface enrichment. The parallel-resistor model is standard practice. The fact that both ST-FMR linewidth modulation and SHNO Ith move in the same direction is a decent internal consistency check. The ρimp^2 scaling fit for side-jump is suggestive, though it rests on four points after subtracting an assumed intrinsic term.\n\nSoft spots, in order of importance. First, the Ta seed layer: 2.4 nm Ta is a known spin-Hall source, and the analysis attributes all damping modulation to the PtBi layer. Since PtBi resistivity rises from 65 to about 270–300 μΩ·cm, current repartitions into Ta as Bi increases, so a Ta-derived antidamping torque would grow relative to the PtBi torque. The pure-Pt reference also contains Ta, so it doesn't calibrate the absolute scale. This doesn't kill the relative comparison, because both SHNOs share the same seed, but it does mean the abstract's \"0.24 vs 0.07\" is not established as a PtBi-only property. Second, the mechanism: side-jump is inferred from a four-point fit of ρSH_imp versus ρimp^2 after subtracting an assumed composition-independent intrinsic contribution. That is a weak constraint; it's consistent but not compelling. Third, the paper doesn't reconcile with Chi et al.'s θSH≈0.10 for similar composition or discuss why the present values are so much higher. That needs addressing.\n\nBottom line: the device-level Ith reduction is the robust contribution; the absolute θSH number should be treated as an upper bound pending a Ta-free control or a proper subtraction. The paper deserves peer review and would benefit from the control experiment. I'd bring it to a reading group for the methodology question alone.","headline":"The device-level threshold-current win is likely real; the absolute θSH numbers are not yet isolated from the Ta seed layer, so treat 0.24 as an upper bound.","tokens_in":17341,"tokens_out":1694,"would_cite":true,"duration_ms":19619,"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":"Alloying Pt with 6% Bi raises the spin Hall efficiency from 0.07 to 0.24 and cuts the threshold current of 100 nm spin Hall nano-oscillators by 42%.","keywords":["spin-orbit torque","spin Hall effect","spin Hall nano-oscillator","PtBi alloy","threshold current","side-jump scattering","ST-FMR","auto-oscillation"],"falsifier":"Fabricate the same Co40Fe40B20/PtBi bilayer on a seed layer that produces no spin-orbit torque, or with tantalum seed thickness varied from zero to several nanometres, and repeat the DC-bias ST-FMR linewidth-modulation measurement; if the extracted spin Hall efficiency drops substantially when tantalum is removed or thinned, part of the claimed enhancement comes from the seed layer rather than the PtBi alloy.","tokens_in":16210,"feed_emoji":"🧲","tokens_out":8956,"duration_ms":90556,"temperature":0.7,"pith_summary":"Spin Hall nano-oscillators are nanoscale microwave sources driven by a spin current produced in a heavy-metal layer, and their usefulness depends on keeping the threshold current low. This paper aims to establish that adding a few percent of bismuth to platinum, specifically 6.0 at% Bi, raises the spin Hall efficiency (the conversion of charge current into spin current) from 0.07 in pure Pt to 0.24, and that this directly lowers the threshold current of a 100 nm oscillator from 1.30 mA to 0.75 mA, a 42% reduction. The authors attribute the gain to a bulk, extrinsic side-jump scattering mechanism rather than to interface effects, based on structural imaging, torque symmetry, and resistivity scaling. If right, PtBi alloys offer a practical, low-resistivity alternative to ion-implanted or insulating spin Hall layers for energy-efficient spintronic devices such as SOT-MRAM and neuromorphic oscillator networks.","feed_headline":"6% Bi in Pt triples spin Hall efficiency, cuts oscillator current 42%","feed_subtitle":"Spin Hall efficiency reaches 0.24 and SHNO threshold current drops to 0.75 mA, cutting power for spintronic devices.","key_machinery":"The central quantity is the spin Hall efficiency θSH, defined by js = (ℏ/2e) θSH (jc × σ̂), which the paper extracts from the DC-bias-induced linewidth modulation δ(µ0ΔH)/Idc using Eq. (4): θSH = (2e/ℏ)[(H0 + Meff)/2] μ0 Ms t |Δαeff/Δjdc,HM| / sin ϕ. The argument for the mechanism rests on the scaling relation ρSH_imp ∝ ρ_imp², the signature of extrinsic side-jump scattering (the spin-dependent sideways deflection of electron trajectories at impurities), together with the parallel-resistor estimate of the current density in the PtBi layer. Structural tools (GIXRD, cross-sectional TEM, and EDS) do the supporting work of showing that Bi is uniformly incorporated, that Pt crystallinity decreases, and that the largest θSH appears near the crystalline-to-disordered transition.","core_discovery":"The central claim is that Pt100−xBix alloys with small Bi concentrations act as efficient bulk spin Hall materials: the spin Hall efficiency θSH rises from 0.07 ± 0.01 for pure Pt to 0.24 ± 0.02 for Pt94.0Bi6.0 and 0.19 ± 0.01 for Pt91.3Bi8.7, as extracted from DC-bias spin-torque ferromagnetic resonance on Co40Fe40B20/PtBi bilayers. The same alloying reduces the threshold current Ith of 100 nm spin Hall nano-oscillators by 42% (from 1.30 mA to 0.75 mA) at 6.0% Bi and by 32% at 8.7% Bi, despite an increase in the ferromagnet's Gilbert damping. The paper identifies the mechanism as bulk-dominated extrinsic side-jump scattering: the impurity spin Hall resistivity scales quadratically with impurity resistivity, the torque shows the bulk-SOT sin 2ϕ cos ϕ angular symmetry, and TEM-EDS shows Bi distributed uniformly through the Pt layer without interfacial clustering. Structural characterization ties the efficiency peak to the loss of Pt crystallinity as Bi content increases, placing the optimum near 6% Bi.","pith_inferences":["A Ta-free control stack would quantify how much of the 0.24 θSH comes from the PtBi alloy rather than the seed layer, a test the paper does not report.","The side-jump mechanism predicts that θSH should track impurity resistivity as ρxx is tuned by temperature or composition; measuring θSH at low temperatures could confirm the extrinsic picture independently of structural data.","If the threshold-current reduction scales to smaller constrictions, PtBi could be combined with ultra-low-damping ferrimagnets to push SHNO drive currents below the values reported here.","The same co-evaporation recipe could be tested in harmonic Hall or spin-pumping geometries to cross-check θSH without relying on ST-FMR linewidth assumptions."],"forward_implications":["At 6.0% Bi the threshold current of a 100 nm SHNO falls to 0.75 mA, so PtBi-based devices can sustain auto-oscillation at substantially lower drive current than pure-Pt devices.","Because the spin Hall efficiency rises without the very high resistivities seen in ion-implanted or nitrided metals, PtBi alloys improve the power figure of merit ρxx/θSH² used for SOT-MRAM comparisons.","The bulk, side-jump origin means the torque is not tied to a specific PtBi/ferromagnet interface, so the efficiency gain should transfer to other ferromagnets and capping stacks.","The efficiency peaks near 6% Bi and declines at 8.7% Bi, indicating an optimal composition window rather than a monotonic gain with more bismuth.","Even though Bi doping raises Gilbert damping, the enhanced damping-like torque overcompensates, so the threshold current still decreases; pairing PtBi with a lower-damping ferromagnet would make the reduction larger."],"supporting_citations":[{"why":"Supplies the ST-FMR method and Eq. (4) used to extract the spin Hall efficiency from DC-bias linewidth modulation.","marker":"[16]"},{"why":"Prior demonstration that Pt1−xBix alloys improve charge-spin conversion and reduce switching current, the direct precursor for this study.","marker":"[43]"},{"why":"Reports a giant inverse spin Hall effect in Bi-doped PtBi, supporting strong spin-current detection in these alloys.","marker":"[41]"},{"why":"Shows giant spin Hall effect in 5d transition-metal-aluminum alloys from extrinsic scattering, the comparison for the crystalline-to-disorder enhancement mechanism.","marker":"[38]"},{"why":"Recent Pt(P) alloy study showing giant spin-orbit torque from the spin Hall effect, used as a reference for amorphous Pt alloys and scattering analysis.","marker":"[34]"},{"why":"Establishes θSH versus conductivity scaling regimes in Pt, used to argue against a purely intrinsic mechanism.","marker":"[61]"},{"why":"Review of spin Hall effects that provides the side-jump and skew-scattering scaling relations used to identify the mechanism.","marker":"[63]"},{"why":"Device fabrication process for nano-oscillators used to make the 100 nm constriction devices.","marker":"[65]"}],"fun_headline_variants":["PtBi alloy triples spin Hall efficiency, cuts oscillator current 42%","Pt94Bi6 raises spin Hall efficiency to 0.24, cuts threshold 42%","Bulk spin Hall from PtBi alloys triples efficiency, cuts threshold 42%","PtBi alloys: spin Hall efficiency 0.24, 42% lower threshold current"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The analysis assumes that the measured damping modulation comes almost entirely from spin-orbit torque generated in the PtBi layer itself, but the stacks contain a 2.4 nm tantalum seed layer that can also produce spin-orbit torque, and the paper does not subtract or bound its contribution; if that tantalum torque is substantial, the reported spin Hall efficiencies are overestimated.","fun_headline_variants_meta":{"raw":{"variants":["PtBi alloy triples spin Hall efficiency, cuts oscillator current 42%","Pt94Bi6 raises spin Hall efficiency to 0.24, cuts threshold 42%","Bulk spin Hall from PtBi alloys triples efficiency, cuts threshold 42%","PtBi alloys: spin Hall efficiency 0.24, 42% lower threshold current"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001165,"raw_usage":{"total_tokens":4957,"prompt_tokens":1219,"completion_tokens":3738,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":835,"completion_tokens_details":{"reasoning_tokens":3646}},"tokens_in":835,"tokens_out":3738,"duration_ms":26370,"temperature":1.0,"reasoning_tokens":3646,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T17:37:02.168694+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Fabricate the same Co40Fe40B20/PtBi bilayer on a seed layer that produces no spin-orbit torque, or with tantalum seed thickness varied from zero to several nanometres, and repeat the DC-bias ST-FMR linewidth-modulation measurement; if the extracted spin Hall efficiency drops substantially when tantalum is removed or thinned, part of the claimed enhancement comes from the seed layer rather than the PtBi alloy.","supporting_citations":[{"cited_title":"Spin-torque ferro- magneticresonanceinducedbythespinhalleffect","cited_arxiv_id":null,"evidence_quote":"Supplies the ST-FMR method and Eq. (4) used to extract the spin Hall efficiency from DC-bias linewidth modulation."},{"cited_title":"Charge– spin conversion in pt1- xbix alloys for spin–orbit torque switching.APL Materials, 9(6), 2021","cited_arxiv_id":null,"evidence_quote":"Prior demonstration that Pt1−xBix alloys improve charge-spin conversion and reduce switching current, the direct precursor for this study."},{"cited_title":"Giant inverse spin hall effect in bi doped ptbi alloy","cited_arxiv_id":null,"evidence_quote":"Reports a giant inverse spin Hall effect in Bi-doped PtBi, supporting strong spin-current detection in these alloys."},{"cited_title":"Giant spin hall effect and spin–orbit torques in 5d transition metal–aluminum alloys from extrinsic scattering","cited_arxiv_id":null,"evidence_quote":"Shows giant spin Hall effect in 5d transition-metal-aluminum alloys from extrinsic scattering, the comparison for the crystalline-to-disorder enhancement mechanism."},{"cited_title":"Giant spin-orbit torque induced by spin hall effect in amorphous pt (p) alloys","cited_arxiv_id":null,"evidence_quote":"Recent Pt(P) alloy study showing giant spin-orbit torque from the spin Hall effect, used as a reference for amorphous Pt alloys and scattering analysis."},{"cited_title":"Tuning the spin hall effect of pt from the moderately dirty to the superclean regime.Physical Review B , 94(6):060412, 2016","cited_arxiv_id":null,"evidence_quote":"Establishes θSH versus conductivity scaling regimes in Pt, used to argue against a purely intrinsic mechanism."},{"cited_title":"Valenzuela, J","cited_arxiv_id":null,"evidence_quote":"Review of spin Hall effects that provides the side-jump and skew-scattering scaling relations used to identify the mechanism."},{"cited_title":"Fabrication of voltage-gated spin Hall nano-oscillators","cited_arxiv_id":null,"evidence_quote":"Device fabrication process for nano-oscillators used to make the 100 nm constriction devices."}],"review_version":1}