{"id":"1407851e-e6fd-408c-8e78-1f86a07eaedc","arxiv_id":"1908.08435","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Time-domain switching measurements in FeCoB nanomagnets yield high-precision coercive field, retention time, delta, and nucleation-domain size via a linear Néel-model fit.","lead":"A single-nanomagnet Hall-probe method measures how long magnetization takes to flip in an opposing field, and derives coercive field, retention time, thermal-stability delta, and nucleation-domain size from the waiting-time curve. The method claims higher precision and less systematic error than conventional hysteresis scanning, which matters for MRAM, voltage-control, and spin-orbit-torque research.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Nucleation-domain volume from Eq. (17) requires a strictly field-independent attempt frequency, which the reported linear log-time fits do not establish.","rationale":"The reader's weakest assumption matches the main fragility I find: the slope-to-domain-moment link in Eq. (17) is the load-bearing step that is least supported by the manuscript. I agree with the reader that the Hc and retention-time parts of the method are plausible and likely sound, while the domain-size and model-completeness claims are the fragile parts. The concern lands because linearity alone is not a sufficient test: an exponential field dependence of the prefactor preserves the linear log-time plot but changes the inferred slope and therefore the domain volume. Since the reader already issued CONDITIONAL, my independent read does not move the verdict; it reinforces it. I am not claiming the measurements are wrong; rather, the manuscript needs one concrete check of the prefactor assumption before the domain-size precision claim can be accepted.","tokens_in":18700,"tokens_out":7340,"duration_ms":81832,"concrete_test":"Take one representative nanomagnet with independently measured Hanis, M_s, volume, and damping, and compute the switching rate from Brown's Fokker-Planck/Kramers escape formula with the full field-dependent prefactor over the same field range used in Fig. 2. Evaluate d log10(t_switch)/dH from the numerical solution and compare it with M_domain/(kT ln10) from Eq. (17). If the two slopes differ by more than the claimed 3-nm-equivalent uncertainty in V_domain, the domain-size extraction is not model-independent. A cheaper companion check is to compare Delta from Eq. (19) with Delta obtained by an independent retention-vs-temperature measurement on the same sample.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central quantitative claim is that the slope S = d log10(tswitch)/dH equals M_domain/(kT ln10), so Eq. (17) yields the nucleation-domain volume. This identifies the entire barrier slope with the Zeeman moment of the nucleating region and assumes the pre-exponential factor f0 in Eq. (11) has zero field dependence. That assumption is load-bearing and unverified. Any prefactor of the form f0(H) = A exp(alpha H) would shift the observed slope to (M_domain/(kT ln10)) - alpha, changing the inferred domain size without changing the linearity of the plot. This is not a remote possibility: in the Neel-Brown and nucleation-mediated pictures the attempt rate depends on field through barrier curvature, damping, and the field-dependent critical-nucleus size. The paper's only validation, that all log-time-vs-field data are linear, cannot discriminate the constant-prefactor Arrhenius model from a model with a weak or exponential field-dependent prefactor. The manuscript's Appendix 3 criticizes uncontrolled assumptions about f0, such as the 1 GHz convention, but never rules out field dependence of f0. Consequently the reported 3-nm precision for nucleation-domain size is conditional on an untested premise; the Hc and retention-time extraction, which depend mainly on intercept and zero-crossing, are much less affected.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a measurement method for thermally activated magnetization switching in FeCoB nanomagnets. For each applied perpendicular field, the switching time is measured repeatedly, and the logarithm of the mean switching time is fitted as a linear function of field. From the fitted slope and intercept the authors extract the coercive field (at a 1 s convention), the retention time, the thermal-stability parameter Δ, and the size of the assumed nucleation domain. The method is demonstrated on Hall-probe devices, and the extracted parameters are studied as functions of gate voltage (VCMA) and bias current (SOT). The paper argues that the observed linearity confirms the classical Néel model with two free parameters.","tokens_in":19009,"tokens_out":7616,"duration_ms":83858,"significance":"If the central extraction is valid, the method is attractively simple: it avoids the systematic timing errors of pulsed or ramped-field methods and gives access to a nucleation-domain volume without imaging. The extraction of Hc and retention time from the linear fit is structurally sound under the stated small-field approximation, and the independent measurement of Hanis for Δ is a good practice. However, the paper's most distinctive quantitative claim — the 3 nm precision of the nucleation-domain size — depends on an unverified identification of the fitted slope with the Zeeman moment of the nucleation region, and the Fig. 5 'confirmation' is partly circular. The precision claims are also not backed by an error budget. The contribution is therefore promising but needs substantial strengthening before the domain-size and model-validation conclusions can be accepted.","major_comments":[{"comment":"The conversion of the fitted slope S = d log10(tswitch)/dH into a nucleation-domain volume assumes that the attempt frequency f0 in Eq. (11) is strictly field-independent and that the field enters the switching rate only through the Zeeman energy of the nucleation region. A prefactor of the form f0(H) = A exp(αH) would change the observed slope to S = M_domain/(kT ln10) − α without breaking the linearity of Fig. 2, so the reported linear fits cannot discriminate the constant-prefactor Arrhenius model from a model with a field-dependent prefactor. Since Eq. (17) is the sole basis for the reported 3 nm precision and for the current-dependence of the domain size in Fig. 7, this is a load-bearing assumption. Please provide an independent check, for example by verifying that S scales as 1/T over a range of temperatures, by estimating the field dependence of f0 from micromagnetic or Néel–Brown calculations, or by directly measuring the prefactor; alternatively, re-state the claim as an effective magnetic moment rather than a geometric volume.","section":"§IV, Eq. (17)"},{"comment":"The apparent linear relation between Δ and the nucleation-domain size is not an independent confirmation of the model. Equation (19) defines Δ ∝ slope, and Eq. (17) defines V_domain ∝ slope, so both plotted axes are functions of the same fitted slope; the positive correlation is largely built into the construction. The statement that 'the data of Fig. 5 confirms such dependence' therefore overstates the evidential value. Please show the relation using independently measured Hanis and M_s values, or present a residual analysis that isolates the scatter not forced by the common slope.","section":"§IV, Fig. 5"},{"comment":"The precision statements (0.6 Oe for Hc, 2% for Δ, and 3 nm for the domain size) are stated without an error budget, confidence intervals, or any error bars in Figs. 2, 6, or 7. Because tswitch depends exponentially on the field, small systematic errors in field calibration, temperature drift, or timing could dominate the statistical spread. Please provide a propagation-of-uncertainties analysis that includes the 200-repetition statistics, field-sensor accuracy, temperature stability, and the covariance of the fitted slope and intercept, and add confidence intervals to the figures.","section":"§II–III"},{"comment":"The derivation of Δ from Eq. (19) treats the zero-field barrier as (1/2) Hanis M_domain, i.e., the barrier of a coherently rotating single-domain particle of the nucleation-domain volume. For a nucleation-mediated reversal the zero-field barrier generally includes domain-wall and magnetostatic contributions that do not scale with Hanis M_domain in this simple way. The paper should state this as an explicit modeling assumption and justify it, or restrict the Δ interpretation to the single-domain case.","section":"§IV, Eqs. (7)–(10) and (19)"}],"minor_comments":[{"comment":"The sentence 'the magnetic field H was applied opposite to the magnetization direction and H reset' is unclear; please rephrase to specify the direction of H relative to the reset field.","section":"§II"},{"comment":"The numerical conversion factor 51717 in Eq. (17) is not derived; please include a worked example with explicit units for slope and M_ferro/V so that readers can reproduce the domain-size values.","section":"§IV, Eq. (17)"},{"comment":"The arguments of the error functions in Eq. (A2.5) appear to mix dimensionless combinations and field quantities; please check the dimensional consistency of this equation.","section":"Appendix 2"},{"comment":"The phrase 'There are n’t no any facts or experimental evidences' contains a double negative and should be rewritten.","section":"Appendix 3"},{"comment":"The conclusion that the nucleation-domain size is nearly independent of nanomagnet size for effective sizes above 200 nm should be qualified by the absence of error bars and by possible sample-to-sample variations in material parameters.","section":"§IV, Fig. 4"}],"recommendation":"major_revision","confidential_remarks":"The Hc and retention-time extraction is a plausible and useful contribution, but the paper's headline domain-size measurement and the 'perfect confirmation' of the Néel model are currently overclaimed. If the authors can provide an independent check of the slope-to-domain-moment identification (or remove the geometric interpretation), the paper would be much stronger. I would not reject on the model assumptions alone, but the required revision is substantive rather than cosmetic."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague, quick take on Zayets 1908.08435. The useful core is a direct time-resolved switching measurement on a single FeCoB nanomagnet: apply reverse field, wait, record tswitch, repeat 200 times, plot log10(tswitch) vs H. For their samples that plot is linear, and the intercept gives retention time, the zero-crossing gives Hc at 1 s, and the slope feeds Delta and a nucleation-domain volume. That is a real experimental contribution—cleaner than sweep-rate coercivity methods and avoids the pulsed-field time-integration error they dissect in Appendix 3. If the precision claims (0.6 Oe, 2%, 3 nm) survive scrutiny, this is useful for MRAM/VCMA/SOT device work.\n\nSoft spots, in rough order. First, there are no error bars anywhere and no error-propagation budget. The precision claims are stated but not demonstrated; a referee needs the statistics of the 200 repeats and the uncertainty in Hanis and Mferro. Second, the domain-size extraction assumes the slope of log10(tswitch) vs H equals M_domain/(kT ln10) with a strictly field-independent prefactor f0. The paper correctly attacks the 1 GHz f0 convention, but never tests field dependence of f0. Any prefactor f0(H) = A exp(alpha H) changes the inferred domain size without changing linearity. So the 3 nm precision is conditional on an untested premise. The Hc and retention time, which come from intercept and zero-crossing, are much less affected. Third, Fig. 5 is circular: both axes are derived from the same fitted slope, so the linear trend is baked in. It doesn't independently confirm the domain-size model. The claim that the classical Néel model 'perfectly describes' everything is also stronger than linearity can support—linear log-time is necessary but not sufficient to rule out field-dependent prefactors.\n\nNone of this kills the paper. The measurement method itself is likely sound and the VCMA/SOT data are a useful demonstration. It's the domain-size inference and the model-completeness conclusion that need caution and more evidence.\n\nWho this is for: experimentalists working on single-nanomagnet stability and voltage/current control of magnetic anisotropy; not for someone looking for new physics. Worth a serious peer review, but with a request for error bars, a sensitivity analysis on f0(H), and a reworked Fig. 5.","headline":"Useful high-precision switching-time method for single nanomagnets, but the nucleation-domain-size inference leans on an unverified field-independent attempt frequency and a circular confirmation plot.","tokens_in":19474,"tokens_out":2052,"would_cite":false,"duration_ms":21340,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["75.60.Jk","75.75.-c"],"model":"deepseek-v4-flash","headline":"The paper claims that the logarithm of the magnetization switching time is a straight-line function of the applied perpendicular field, and that fitting this line yields coercive field, retention time, thermal-stability factor $\\Delta$…","keywords":["thermally activated switching","N\\'eel model","coercive field","retention time","nucleation domain","anomalous Hall effect","voltage-controlled magnetic anisotropy","spin-orbit torque"],"falsifier":"Directly image the nucleation domain in the same FeCoB nanomagnet by time-resolved XMCD while measuring $t_{\\mathrm{switch}}(H)$, and compare the imaged domain volume with the value from Eq. (17); agreement to within the claimed 3 nm would support the slope-to-moment conversion, while a mismatch would show that the slope is not set by the Zeeman energy of that domain alone. Similarly, measuring $t_{\\mathrm{switch}}(H)$ over a much wider field range would expose any field-dependent prefactor as curvature in $\\log t_{\\mathrm{switch}}$.","tokens_in":18495,"feed_emoji":"🧲","tokens_out":7250,"duration_ms":70425,"temperature":0.7,"pith_summary":"The paper proposes a direct way to measure the parameters that control thermally activated magnetization reversal in a single nanomagnet. Instead of sweeping the field and reading off a switching field, it measures how long a FeCoB nanomagnet takes to flip at each fixed perpendicular field, then fits the logarithm of that waiting time against the field. The central claim is that this curve is a straight line, as the classical N\\'eel model predicts, and that the line's slope and intercept directly give the magnetic moment of the nucleation region, the retention time, and the coercive field at a defined waiting time. If this is right, a two-parameter Arrhenius description is sufficient for non-resonance switching, and small voltage- or current-induced changes in switching parameters can be resolved accurately. The same measurement also yields the size of the nucleation domain where reversal starts.","feed_headline":"One straight-line fit yields a nanomagnet's switching parameters","feed_subtitle":"Measuring flip-wait times gives coercive field, retention time, thermal barrier, and nucleation size to ~3 nm.","key_machinery":"The central object is the Arrhenius relation written as a straight line: $\\log_{10}(t_{\\mathrm{switch}})=\\log_{10}(\\tau_{\\mathrm{retention}})-(M/kT)H$. The slope $M/kT$ converts a measured waiting-time curve into the magnetic moment of the nucleating region; the intercept is the zero-field retention time; the point where the line crosses the $t_{\\mathrm{switch}}=1$ s level is the coercive field; and the thermal-stability factor $\\Delta$ is obtained by combining the slope with the anisotropy field $H_{\\mathrm{anis}}$ measured in a separate magnetostatic experiment. This one identity carries all four extracted parameters, so the measurement reduces to fitting two numbers.","core_discovery":"The core discovery is that for the measured FeCoB/FeB nanomagnets, the average switching time obeys $\\log_{10}(t_{\\mathrm{switch}})=\\log_{10}(\\tau_{\\mathrm{retention}})-(M/kT)H$ over the studied range, with $M$ the moment of the region that nucleates reversal. From the linear fit, the field-axis intercept at $t_{\\mathrm{switch}}=1$ s defines the coercive field $H_c$; the time-axis intercept defines the retention time $\\tau_{\\mathrm{retention}}$; the slope gives $M$, from which the nucleation-domain volume is obtained using the independently measured saturation magnetization; and the thermal-stability factor $\\Delta$ follows from the same slope together with the anisotropy field $H_{\\mathrm{anis}}$. The paper reports fitted precisions of about 0.6 Oe for $H_c$, 2% for $\\Delta$, and 3 nm for the effective nucleation-domain size, and shows that the extracted parameters change linearly with gate voltage (the VCMA effect) and with bias-current polarity and magnitude (the SOT effect). It concludes that the classical N\\'eel model, with only two free parameters, fully describes non-resonance thermally activated switching.","pith_inferences":["If the slope-to-moment identification survives an independent check, the method turns any setup that can time single switching events into a domain-size probe without direct magnetic imaging.","The same straight-line analysis could be applied to current- or voltage-induced switching by replacing the perpendicular field with the effective field from the current or voltage, provided that field enters only through the Zeeman term.","The observed sample-to-sample spread in nucleation-domain size (about 20 to 90 nm) suggests that defects set the nucleation site; if that is causal, controlled defect patterning could stabilize or shrink domain sizes, a prediction this measurement could test directly."],"forward_implications":["A single waiting-time-versus-field scan per nanomagnet determines $H_c$, $\\tau_{\\mathrm{retention}}$, $\\Delta$, and nucleation-domain size at once, with no ramp-rate correction.","Because only two parameters define the line, the N\\'eel model's two-parameter description can be tested directly: data that stay linear support it, while curvature would signal resonance-type or non-Arrhenius behavior.","For VCMA studies, a voltage scan changes the intercept but not the slope, indicating that the gate voltage leaves the nucleation-domain size unchanged while shifting the coercive field and retention time.","For SOT studies, reversing the bias-current polarity changes the slope, so the extracted nucleation-domain size itself becomes current-dependent.","The method can distinguish single-domain from nucleation-domain reversal when the extracted domain size approaches the nanomagnet size, identifying the critical size where the reversal mechanism changes."],"supporting_citations":[{"why":"Supplies the Arrhenius/N\\'eel model on which the linear switching-time relation is based.","marker":"[6]"},{"why":"Provides the Landau-Lifshitz-based Fokker-Planck model that the paper sets aside for non-resonance switching.","marker":"[7]"},{"why":"Supports the claim that the simpler N\\'eel model is sufficient when switching is not resonant.","marker":"[8]"},{"why":"Supplies the perpendicular-anisotropy energy expression used to derive the energy barrier and the linear relation.","marker":"[10]"},{"why":"Provides direct time-resolved XMCD observation of nucleation-domain reversal and domain-wall expansion, motivating the domain-size extraction.","marker":"[11]"},{"why":"Demonstrate time-dependence-based evaluation of coercive field and retention in multi-particle systems that the present single-particle method extends.","marker":"[4,5]"}],"fun_headline_variants":["One fit yields coercive field, retention time, thermal barrier, and nucleation size","Four magnetic parameters from one switching-time line in FeCoB","Nanomagnet flip: one straight fit gives Hc, τ, Δ, and volume","Precise thermal-switching parameters from one linear fit in FeCoB","From flip-wait times: coercivity, retention, Δ, and nucleation size in one fit"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The extracted nucleation-domain size assumes that the slope of the switching-time plot is exactly the magnetic moment of one unchanging nucleation region divided by $kT$, with the attempt frequency and prefactor unaffected by the field; the paper gives no independent check of that separation.","fun_headline_variants_meta":{"raw":{"variants":["One fit yields coercive field, retention time, thermal barrier, and nucleation size","Four magnetic parameters from one switching-time line in FeCoB","Nanomagnet flip: one straight fit gives Hc, τ, Δ, and volume","Precise thermal-switching parameters from one linear fit in FeCoB","From flip-wait times: coercivity, retention, Δ, and nucleation size in one fit"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001957,"raw_usage":{"total_tokens":7629,"prompt_tokens":900,"completion_tokens":6729,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":516,"completion_tokens_details":{"reasoning_tokens":6637}},"tokens_in":516,"tokens_out":6729,"duration_ms":47287,"temperature":1.0,"reasoning_tokens":6637,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:39:56.407591+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Directly image the nucleation domain in the same FeCoB nanomagnet by time-resolved XMCD while measuring $t_{\\mathrm{switch}}(H)$, and compare the imaged domain volume with the value from Eq. (17); agreement to within the claimed 3 nm would support the slope-to-moment conversion, while a mismatch would show that the slope is not set by the Zeeman energy of that domain alone. Similarly, measuring $t_{\\mathrm{switch}}(H)$ over a much wider field range would expose any field-dependent prefactor as curvature in $\\log t_{\\mathrm{switch}}$.","supporting_citations":[{"cited_title":"Néel, Some theoretical aspects of rock-magnetism, Adv","cited_arxiv_id":null,"evidence_quote":"Supplies the Arrhenius/N\\'eel model on which the linear switching-time relation is based."},{"cited_title":"Coffey, Y.P","cited_arxiv_id":null,"evidence_quote":"Supports the claim that the simpler N\\'eel model is sufficient when switching is not resonant."},{"cited_title":"Baumgartner, K","cited_arxiv_id":null,"evidence_quote":"Provides direct time-resolved XMCD observation of nucleation-domain reversal and domain-wall expansion, motivating the domain-size extraction."}],"review_version":1}