{"id":"3a8462f2-a47b-468b-bc39-3ea8eb445e75","arxiv_id":"2608.09021","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Near-compensated CoGd ferrimagnetic spin Hall nano-oscillators start oscillating at 1.43e7 A/cm2 with a 5 mT field and a 0.61 MHz linewidth, beating Pt/Py controls by more than an order of magnitude.","lead":"This paper builds tiny microwave oscillators from a cobalt-gadolinium alloy tuned so its two magnetic sublattices almost cancel. The devices start oscillating at much lower current and magnetic field and produce a cleaner signal than standard ferromagnetic versions, promising lower-power spintronic microwave sources.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Linewidth comparison is not frequency-matched: the 16x Δf advantage comes with only a 2.3x Q advantage because CoGd runs at 0.51 GHz while Pt/Py runs at 3.44 GHz.","rationale":"The paper's central claim has three quantitative pillars: low threshold current, low operating field, and narrow linewidth. The first two are supported by the data and by the comparison to a standard Pt/Py control; the compensation-point mechanism is plausible and bolstered by MOKE, VSM, ST-FMR, and micromagnetic simulations. The least secure pillar is the linewidth comparison, because the paper reports a Q factor only 2.3x higher while claiming a 16x linewidth improvement; the difference is explained by the 6.8x lower operating frequency of the CoGd device. The reader's stated weakest assumption concerns the compensation composition, which is a confirmation risk rather than an internal inconsistency. My concern is a specific quantitative fairness issue: the 16x linewidth number mixes frequency scaling with coherence. It does not overturn the main demonstration, but it should be tested by a frequency-matched comparison. If the Pt/Py device also achieves ~0.6 MHz at ~0.5 GHz, the 'record narrow linewidth' would be reframed as a low-field/low-frequency consequence; if it remains broad, the material claim is strengthened. The verdict stays CONDITIONAL because the main results and mechanism are credible but the headline linewidth factor needs this check before being treated as definitive.","tokens_in":12771,"tokens_out":22159,"duration_ms":214425,"concrete_test":"Measure the Pt/Py control SHNO at the lowest field at which it auto-oscillates (targeting f ≈ 0.5 GHz, the CoGd frequency at 5 mT) using the same θ, φ, and measurement chain, and extract its minimum linewidth and Q. If Δf(Py, ~0.5 GHz) is close to 0.61 MHz, the 16x linewidth claim is largely a frequency effect rather than a ferrimagnet-specific coherence improvement.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The headline '16-fold narrower linewidth' compares the CoGd device at 5 mT (Δf = 0.61 MHz, Q = 832) with the Pt/Py device at 50 mT (Δf = 9.63 MHz, Q = 357). From these values the CoGd auto-oscillation frequency is f = Q·Δf ≈ 0.51 GHz, whereas the Pt/Py frequency is ≈ 3.44 GHz. Thus the linewidth ratio (≈15.8) is almost entirely accounted for by the frequency ratio (≈6.8) and the Q-factor ratio (≈2.3). Since the nonlinear broadening formula used in the paper, Eq. (5) with N2 from Eq. (4), depends on the operating field and effective magnetization, a device deliberately operated at 5 mT will have a smaller absolute linewidth than one at 50 mT even if the material coherence is comparable. The paper's own nonlinearity model predicts this field dependence. Therefore the 'record-low linewidth' is partly a consequence of the low operating frequency, not purely a material improvement in spectral purity. The claim that low field and low current are achieved simultaneously is supported, but the '16x narrower linewidth' as a material-level advantage is not an apples-to-apples comparison. A frequency-matched comparison is needed before that quantitative headline can be accepted.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports spin Hall nano-oscillators made from Pt/Co1-xGdx ferrimagnetic alloys, with composition tuned near the magnetization compensation point. The authors claim that a near-compensated Pt/Co73.1Gd26.9 device achieves a 16-fold lower threshold current density (1.01×10^7 A/cm^2), a 10-fold smaller operating field (5 mT), and a 16-fold narrower minimum linewidth (0.61 MHz) than a conventional Pt/Py SHNO, with the improvements attributed to reduced effective magnetization, enhanced spin-orbit torque efficiency, low magnetic anisotropy, and suppressed nonlinearity. The evidence includes MOKE and VSM characterization, ST-FMR measurements of damping and torque efficiency, current- and field-dependent auto-oscillation spectra, time-domain phase noise measurements, and micromagnetic simulations. A theoretical model for the nonlinear coefficient N2 is used to connect the measured effective magnetization and external field to the linewidth trend.","tokens_in":12982,"tokens_out":3707,"duration_ms":38350,"significance":"If the claims hold, the paper would provide a compelling demonstration that near-compensated ferrimagnets can relax two long-standing constraints on SHNO operation, namely high threshold currents and large external magnetic fields. The work is strengthened by combining several independent measurement techniques (MOKE, VSM, ST-FMR, PSD, time-domain phase noise) and by using a nonlinearity model whose inputs (Meff and Hext) are independently measured rather than fitted to the linewidth data. The proposed materials-design route, based on balancing PMA against the demagnetizing field while keeping Ms small, is physically plausible and could inspire further work. However, the headline '16-fold narrower linewidth' compares devices at very different operating frequencies and fields, and the absence of the Supporting Information leaves key extraction details unverifiable; these issues currently prevent the quantitative performance claims from being accepted as stated.","major_comments":[{"comment":"The identification of xMC at 25–26% Gd rests on MOKE loop polarity reversal between 24.5% and 26.9%, plus VSM minima in Ms and Ku. The quantitative model for Δf in §3 depends sensitively on the value of |μ0Meff| for the operating composition Co73.1Gd26.9, but no direct measurement of Meff for that exact composition is presented in the main text. If the true magnetization compensation point differs significantly from the assumed range, the proposed mechanism (reduced Meff leading to suppressed N2 and lower Jth) would be weakened. The authors should provide a direct estimate of |μ0Meff| for the operating composition and discuss the sensitivity of their conclusions to the exact compensation composition.","section":"§1, Figure 1c and §3"}],"minor_comments":[{"comment":"Reference [5] includes a DOI with a date that appears inconsistent with the manuscript timeline; please check the bibliographic details.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The manuscript has a potentially interesting materials result, but the central quantitative claim about linewidth is currently framed in a way that overstates the material-level improvement, and the Supporting Information—which contains essential extraction details—was not available for review. The paper should not be accepted until the SI is provided and the linewidth comparison is either frequency-matched or reframed in terms of Q-factor or normalized linewidth."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First: this is the first near-compensated ferrimagnetic SHNO I know of, and the composition sweep is a genuine step forward. The authors show that Pt/CoGd near the magnetization compensation point gives low threshold current, low operating field, and a narrow linewidth, with the properties tracking |Meff| as expected. The evidence is multi-technique: MOKE, VSM, ST-FMR, PSD, time-domain phase noise, and micromagnetic simulations, and the modeling uses independently measured parameters (damping, Ms, spin Hall conductivity) rather than fitting the headline values. That is solid work.\n\nThe soft spots, in proportion. The biggest is the linewidth comparison. The stress-test note is correct: the CoGd device at 5 mT oscillates at roughly 0.51 GHz, while the Pt/Py reference at 50 mT is at ~3.44 GHz. So the 16x linewidth difference is mostly the frequency difference; the quality factor is only 2.3x better. The paper does acknowledge the frequency difference and uses Q to argue coherence, but the '16-fold narrower linewidth' headline is not an apples-to-apples material comparison. A frequency-matched comparison, or quoting phase noise and Q as primary metrics, would fix this. Relatedly, there are no error bars or device statistics; the quantitative claims rest on single devices per composition. The Supporting Information containing key extraction details is unavailable, so I could not check the ST-FMR and threshold-extraction procedures.\n\nA smaller concern is the placement of the compensation point. The composition x=26.9% is inferred to be near xMC from MOKE polarity reversal and VSM minima, and the mechanism depends on that. It is plausible, but the inferred xMC has uncertainty, and if the true compensation is further away, the reduced-Meff story weakens. Also, the threshold-current comparison is between different fields (5 vs 50 mT), so the '16-fold lower Jth' includes the low-field benefit—which is a real advantage, but the comparison should be explicit about field dependence.\n\nWho this is for: people working on spintronic oscillators, SOT devices, and RE-TM ferrimagnets. It deserves a serious referee, with requests for frequency-matched linewidth comparison, error bars, and the SI. I would not desk-reject it.","headline":"First near-compensated ferrimagnetic SHNO with real multi-method support, but the '16x narrower linewidth' headline is largely a low-frequency artifact.","tokens_in":13560,"tokens_out":2318,"would_cite":true,"duration_ms":22557,"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":"Near-compensated CoGd ferrimagnets cut spin Hall nano-oscillator current, field, and linewidth by an order of magnitude.","keywords":["spin Hall nano-oscillator","ferrimagnet","magnetization compensation","CoGd alloy","spin-orbit torque","narrow linewidth","low-power microwave source","auto-oscillation"],"falsifier":"Measure the effective magnetization, threshold current density, and linewidth across a fine Gd-composition series from around 23% to 30% at fixed field. If the minima of threshold current and linewidth do not coincide with the composition where the effective magnetization is smallest, or if the effective magnetization at 26.9% proves not to be near zero, the compensation-based mechanism is refuted.","tokens_in":12544,"feed_emoji":"🧲","tokens_out":12122,"duration_ms":101700,"temperature":0.7,"pith_summary":"The paper's claim is that replacing the ferromagnet in a spin Hall nano-oscillator with a rare-earth–transition-metal ferrimagnet, tuned to the magnetization compensation point, removes the main practical obstacles of these microwave devices. It reports Pt/Co$_{73.1}$Gd$_{26.9}$ SHNOs that self-oscillate at a threshold current density of $1.01\\times10^{7}$ A/cm$^2$ under a 5 mT field, with a minimum linewidth of 0.61 MHz—roughly an order of magnitude better than conventional Pt/Py SHNOs in all three figures of merit. The improvement is attributed to a nearly vanishing effective magnetization, enhanced spin-orbit torque efficiency, weak perpendicular anisotropy that almost cancels the demagnetizing field, and suppressed nonlinear frequency noise near the compensation point. A sympathetic reader would care because these are the parameters that decide whether SHNOs can become practical low-power microwave sources for communications and unconventional computing.","feed_headline":"Near-compensated CoGd cuts oscillator current 16-fold at 5 mT","feed_subtitle":"A 0.61 MHz linewidth at 0.17 mA and 5 mT beats ferromagnetic spin Hall nano-oscillators by an order of magnitude.","key_machinery":"The central object is the Pt (6 nm)/Co$_{73.1}$Gd$_{26.9}$ (7 nm) bilayer patterned into a 150-nm nano-constriction, with the CoGd layer brought close to its magnetization compensation point—the composition at which the antiferromagnetically coupled Co and Gd sublattice moments cancel, so $|\\mu_0 M_{\\mathrm{eff}}| \\to 0$ while the two sublattices remain strongly exchange-coupled. Near this point three mechanisms cooperate: spin-orbit torque efficiency peaks because the small net moment makes the torque relatively more effective; the effective anisotropy collapses because the PMA field (51.8 mT) nearly cancels the out-of-plane demagnetizing field (45.3 mT); and the nonlinear coefficient $N^2$ becomes small because it grows with $|\\mu_0 M_{\\mathrm{eff}}|$ and $|\\mu_0 H_{\\mathrm{ext}}|$. Gilbert damping stays moderate because the operating composition is closer to the magnetization compensation point than to the angular momentum compensation point ($\\sim 23\\%$ Gd), where $\\alpha$ would diverge. These effects together lower $J_{\\mathrm{th}}$, reduce the required field, and narrow the linewidth.","core_discovery":"At the magnetization compensation point ($x \\approx 25$–$26\\%$ Gd), the antiparallel Co and Gd sublattice moments nearly cancel, giving Co$_{73.1}$Gd$_{26.9}$ a saturation magnetization of only 36 emu/cm$^3$ and an effective magnetization close to zero. This paper shows that SHNOs made from this composition inherit three benefits at once. First, the damping-like spin-orbit torque efficiency rises to 0.102 while the Gilbert damping stays moderate (0.042) because the composition sits far from the angular momentum compensation point. Second, the PMA field (51.8 mT) almost exactly cancels the demagnetizing field (45.3 mT), so a small external field of 5 mT suffices to set the precession axis. Third, the nonlinear coefficient $N^2$, which controls phase-noise linewidth, scales with $|\\mu_0 M_{\\mathrm{eff}}|$ and is therefore minimized. The result is a single-constriction SHNO with $J_{\\mathrm{th}} = 1.01\\times10^{7}$ A/cm$^2$, a record-low linewidth of 0.61 MHz, and a quality factor of 832 at 5 mT, versus $1.64\\times10^{8}$ A/cm$^2$, 9.63 MHz, and $Q = 357$ for the Pt/Py reference.","pith_inferences":["The optimal composition is set by the gap between the magnetization and angular momentum compensation points; engineering that gap (via rare-earth choice, alloying, or temperature) could push the trade-off between low effective magnetization and moderate damping further, lowering threshold current and linewidth beyond this demonstration.","A finer composition series than the paper's would reveal whether threshold current and linewidth track the effective magnetization continuously or flatten near compensation, pinning down whether SOT-efficiency enhancement or moment reduction dominates.","The nearly angle-independent auto-oscillation frequency at fixed field implies that at exact compensation the internal field is set almost entirely by the external field, so a deliberately tailored weak anisotropy could enable field-free SHNOs without cryogenic easy-plane layers.","The same compensation strategy should transfer to other rare-earth–transition-metal pairs, such as CoTb, though each pair brings its own separation between the two compensation points and its own damping profile."],"forward_implications":["Near-compensated Pt/Co$_{73.1}$Gd$_{26.9}$ SHNOs need a 16-fold lower threshold current density ($1.01\\times10^{7}$ A/cm$^2$) and a 10-fold smaller operating field (5 mT) than Pt/Py controls while producing the same kind of microwave auto-oscillation.","The minimum linewidth of 0.61 MHz at 5 mT (quality factor 832) is the narrowest reported for a single-constriction SHNO and is more than an order of magnitude below the 9.63 MHz Pt/Py control.","Because the squared nonlinear coefficient grows with effective magnetization and external field, materials with near-zero effective magnetization should show proportionally lower nonlinear phase noise and narrower linewidths.","Materials combining low saturation magnetization with weak perpendicular anisotropy that almost cancels the demagnetizing field can sustain auto-oscillation at small currents and near-zero fields.","Near-compensated ferrimagnets offer a route to bridge ferromagnetic and antiferromagnetic oscillators, potentially extending to other compensated magnetic systems."],"supporting_citations":[{"why":"Demonstrates that spin-orbit torque efficiency is enhanced near the magnetization compensation point in CoGd, the key mechanism invoked for the low threshold current.","marker":"[25]"},{"why":"Shows compositional control of CoGd magnetic properties and places the magnetization compensation composition used to select x = 26.9%.","marker":"[26]"},{"why":"Introduces the nonlinear coefficient N that governs the frequency shift and linewidth of spin-torque oscillators.","marker":"[37]"},{"why":"Supplies the auto-oscillator theory that relates linewidth to N and effective damping, and the model used to estimate threshold current density.","marker":"[38]"},{"why":"Predicts the divergence of Gilbert damping at the angular momentum compensation point, used to explain why damping remains moderate at x = 26.9%.","marker":"[41]"},{"why":"Demonstrates easy-plane SHNO operation at low fields through cancellation of anisotropy and demagnetizing fields, the analogy for the near-compensated design.","marker":"[22]"},{"why":"Shows that reducing nonlinearity narrows SHNO linewidth, the benchmark the 0.61 MHz result improves on.","marker":"[24]"},{"why":"Provides the spin-torque ferromagnetic resonance method used to extract the damping-like torque efficiency and Gilbert damping.","marker":"[39]"}],"fun_headline_variants":["CoGd SHNO: 16x lower current, 0.61 MHz linewidth","Near-compensated CoGd hits 0.61 MHz at 5 mT","CoGd ferrimagnet SHNO: order-of-magnitude gains","5 mT field, 14 MA/cm²: CoGd SHNO sets mark","CoGd compensation yields 0.61 MHz oscillator linewidth"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The improvements depend on the operating composition, 26.9% Gd, lying close enough to the magnetization compensation point (about 25–26% based on MOKE polarity reversal and VSM minima) that the effective magnetization is strongly reduced, while staying far enough from the angular momentum compensation point (around 23%) that the Gilbert damping remains moderate.","fun_headline_variants_meta":{"raw":{"variants":["CoGd SHNO: 16x lower current, 0.61 MHz linewidth","Near-compensated CoGd hits 0.61 MHz at 5 mT","CoGd ferrimagnet SHNO: order-of-magnitude gains","5 mT field, 14 MA/cm²: CoGd SHNO sets mark","CoGd compensation yields 0.61 MHz oscillator linewidth"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000327,"raw_usage":{"total_tokens":1893,"prompt_tokens":1071,"completion_tokens":822,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":687,"completion_tokens_details":{"reasoning_tokens":713}},"tokens_in":687,"tokens_out":822,"duration_ms":8366,"temperature":1.0,"reasoning_tokens":713,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T04:17:22.845238+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the effective magnetization, threshold current density, and linewidth across a fine Gd-composition series from around 23% to 30% at fixed field. If the minima of threshold current and linewidth do not coincide with the composition where the effective magnetization is smallest, or if the effective magnetization at 26.9% proves not to be near zero, the compensation-based mechanism is refuted.","supporting_citations":[],"review_version":1}