{"id":"7d5f860d-28dd-4718-8f22-1f2102cb7124","arxiv_id":"1908.11061","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Single-atom electron spin resonance on Fe/MgO is demonstrated at zero external magnetic field, using the spin-polarized STM tip's stray field for the Zeeman splitting and swept tip fields for constant-frequency ESR.","lead":"This paper shows that electron spin resonance (ESR) of a single iron atom on a surface can be driven and tuned using only the magnetic field of the microscope tip, with no external magnet. The result makes a promising single-atom sensing and quantum control technique accessible to standard scanning tunneling microscopes.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Single-point 50 mT calibration cannot support the linear B_tip(I) assumption across the full current sweep; SI §6 itself admits nonlinear contributions.","rationale":"The reader and I identify the same weakest assumption: the linearity and single-valuedness of B_z^tip as a function of tunneling current. The paper's own SI (§6) admits a nonlinear contribution to B_z^tip(σ), so a single-point calibration at 50 mT cannot establish the linear scaling assumed in §8 over a current range roughly four times wider than the calibration interval. This is a genuine load-bearing weakness for the quantitative claim of a ~300 mT calibrated field sweep, and it would be worth testing directly. However, the central qualitative claim — that single-atom ESR can be performed at zero external field using the magnetic tip field — is supported by the systematic shift of the resonance with frequency (Fig. S6c) and by the extracted moment being consistent with the known Fe value within ~2σ, so this concern does not overturn the main result. The reader's verdict is already CONDITIONAL, and this concern falls within that conditional status; no further verdict change is warranted.","tokens_in":17487,"tokens_out":14623,"duration_ms":149200,"concrete_test":"Re-measure the current-to-field conversion over the full sweep range by recording ESR spectra at several fixed frequencies (or fixed currents) under zero external field and at B_z^ext = 25, 50, 75, and 100 mT, all at V_DC = 50 mV and T = 0.4 K. Plot the resonant current vs B_z^ext for each fixed frequency; a constant slope across the 100–500 pA range would validate the linear scaling, while a slope varying by more than 20% would invalidate the ~300 mT field axis and require reanalysis of Fig. 4b.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The zero-field constant-frequency sweep (Fig. 4b) is converted to a tip-field axis using one calibration point: at B_z^ext = 50 mT and f = 19 GHz the resonance current shifts by (58.7 ± 2.1) pA, yielding (0.85 ± 0.03) mT/pA, and SI §8 assumes \"a full linear scaling across the whole current range.\" But SI §6 (Evaluation of the tip magnetic field as a function of conductance) explicitly states that \"not all measurements extrapolate to zero tip-field for zero conductance. This is likely caused by a non-linear contribution of B_z^tip(σ).\" The single calibration point only samples a narrow current interval (roughly 90–150 pA based on the 50 mT shift at 19 GHz), while the claimed ~300 mT sweep extends to 500 pA. If B_z^tip(I) deviates from linearity at these closer tip-sample separations, or if the Fe moment changes on approach (the zero-field slope gives μ_Fe = 4.29 ± 0.79 μB vs 5.35 ± 0.14 μB from the external-field data), the x-axis in Fig. 4b and the ~300 mT range are not quantitatively reliable. This does not disprove the qualitative zero-field ESR demonstration, but it weakens the specific claim that the constant-frequency sweep is a calibrated tip-field scan.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports a systematic electron spin resonance (ESR-STM) study of single Fe atoms on MgO/Ag(001) using a spin-polarized STM tip and a vector magnet that provides both out-of-plane and in-plane external fields. The authors measure the resonance frequency as a function of the out-of-plane external field at several in-plane fields, extract a constant Fe moment of (5.35 ± 0.14) Bohr magnetons consistent with prior work, and characterize how the ESR peak amplitude depends on both field components. They then show that increasing the tunneling conductance, which strengthens the tip's stray field, improves the ESR driving efficiency and permits ESR at zero in-plane field. The central demonstration is single-atom ESR at zero external magnetic field, in which the tip field alone provides the Zeeman splitting; this is shown both in frequency sweeps and in constant-frequency conductance sweeps that are converted to a tip-field axis claimed to span roughly 300 mT. The authors conclude that these operating modes eliminate the external-magnet requirement and enable fixed-frequency ESR, potentially broadening the accessibility of ESR-STM.","tokens_in":17803,"tokens_out":22875,"duration_ms":207273,"significance":"If the claims hold, this is a significant experimental advance: zero-field single-atom ESR and fixed-frequency tip-field sweeps are new operating modes for ESR-STM, and the demonstration in a commercial microscope is an enabling step for the wider community. The paper's strengths are that the core observations are direct and internally consistent: the externally calibrated Fe moment agrees with previous determinations, the ESR shift is linear in both external and tip fields, and a 50 mT control measurement confirms the spin origin of the zero-field resonance. The core claim is also falsifiable: single-atom ESR should be observable at zero external field in any STM with a spin-polarized tip whose stray field exceeds the resonance linewidth. The authors are candid about the main ambiguities, explicitly stating that the peak-amplitude trend is not yet attributable to a single mechanism and that a crystal-field driving mechanism cannot be excluded. These strengths make the qualitative zero-field demonstration credible; the quantitative claims concerning the calibrated sweep axis and the order-of-magnitude Rabi-rate enhancement require the revision described below.","major_comments":[{"comment":"The quantitative x-axis of the constant-frequency tip-field sweeps rests on a single calibration point: at B_z^ext = 50 mT and f = 19 GHz the resonance shifts by (58.7 ± 2.1) pA, giving (0.85 ± 0.03) mT/pA, and the current-to-field conversion then assumes 'a full linear scaling across the whole current range' (SI §8). This assumption is in tension with SI §6, which states that 'not all measurements extrapolate to zero tip-field for zero conductance,' 'likely caused by a non-linear contribution of B_z^tip(σ).' The supporting linearity evidence, the 'almost perfect linear evolution' of the resonant current with frequency in Fig. S6c, is given without a goodness-of-fit statistic or residual analysis, and the moment extracted from that curve, (4.29 ± 0.79) μB, is consistent with the external-field value only within a large (≈18%) uncertainty. Because the claimed ~300 mT sweep range and its equivalent ~40 GHz frequency window are quantitative selling points of the new method, the authors should either calibrate the sweep at several external fields spanning the full current range, report separate slope and intercept for B_z^tip(σ) with uncertainties, or attach the calibration uncertainty to the stated range and reword the 'full linear scaling' assumption.","section":"SI §8 and Fig. 4b"},{"comment":"The Rabi rates in Fig. 3c and the claim that Ω is 'approximately one order of magnitude higher than previous experiments' rest on a chain of T1 and T2 estimates whose systematic uncertainties are not propagated into the displayed error bars. T1 is measured by pump-probe only at B_z^ext ≥ 0.4 T and is extrapolated to ESR conditions using a slope of ~46 μs/T derived, by the authors' own description, from measurements at only two field values (SI §3); it is then rescaled for the tunnel-current dependence using a factor of 3.5 per 50 pA measured at a single field and tip condition, and for the bias voltage by factors of 3, 10, and 35 whose derivation is not shown (SI §5, Table 1). T2 is computed with the decoherence probability set to P_T2 = 1 'for the sake of simplicity,' while Ref. [10] used 0.7. These choices can shift Ω/V_RF by factors well beyond the plotted error bars; a sensitivity analysis (e.g., Ω for the plausible range of the T1 scaling factors and for P_T2 = 0.7) is needed before the order-of-magnitude enhancement claim is accepted.","section":"SI §5, Eq. (S5), and Fig. 3c"},{"comment":"The statement that 'the two proportionalities Ω ∝ σ and B_z^tip ∝ σ found here imply Ω ∝ B_z^tip' is stronger than the data support. Several of the B_z^tip(σ) datasets in SI Fig. S5 have nonzero intercepts, as SI §6 acknowledges, so the data establish at most an affine relation B_z^tip ≈ aσ + b, and an offset in that relation changes the inferred dependence of Ω on B_z^tip. To justify the inference that underpins the exchange-mechanism discussion in SI §7, the intercepts of the fits should be reported with uncertainties and shown to be consistent with zero, or the argument should be restated in terms of the measured affine relation. The authors' own caveat that a crystal-field mechanism cannot be excluded is appropriate, but as written the Ω ∝ B_z^tip claim in the main text goes beyond the presented fits.","section":"Main text, Fig. 3c"}],"minor_comments":[{"comment":"The slope of the T1-versus-field relation is given as '~ 46 μμs/T,' which appears to be a typesetting error for ~46 μs/T; please correct the unit.","section":"SI §3"},{"comment":"The bias-voltage T1 reduction factors (3, 10, and 35 for 8, 20, and 60 mV) are asserted without derivation or a specific pointer to the measurements in Refs. [7, 10, 28] from which they are estimated; an equation or a short description would make the Rabi-rate estimation reproducible.","section":"SI §5, Table 1"},{"comment":"The bias voltage for the tip-field sweeps is given as V_DC = 50 mV in the Fig. 4b caption but as V_DC = 30 mV in SI §8 (where it is also used to convert current to conductance); the values should be reconciled.","section":"Fig. 4b caption vs SI §8"},{"comment":"The background subtraction of the current-sweep spectra uses a 'rescaled' 1 GHz spectrum, but the rescaling parameters (scale factor and any offset) are not specified; please provide them for reproducibility.","section":"SI §8"},{"comment":"Tip #1 and tip #2 were measured at different RF voltages (15 mV vs 22 mV), so the absolute ESR amplitudes between the two tips are not directly comparable; the caption or text should note that the field dependence should be read within each tip series.","section":"Fig. 2d"},{"comment":"Eq. (S8) is identical to Eq. (S7), so the displayed derivation of the Rabi force is incomplete; the intended intermediate step, in which the exponential form J(z) = J0 exp(-z/l) is inserted and differentiated to obtain F_J = -g μB B_z^tip(z)/l, is missing.","section":"SI §7"}],"recommendation":"major_revision","confidential_remarks":"To the editor: this is an experimental Letter from a group with a leading track record in ESR-STM, and the zero-field single-atom ESR demonstration, if robust in revision, will be of broad interest to the atomic-scale magnetism and quantum-device communities. The qualitative results are credible; the revision should concentrate on the quantitative calibration of the tip-field sweep axis and on quantifying how sensitive the Rabi-rate enhancement is to the assumed T1 and T2 inputs. The paper cites the simultaneously posted independent work (arXiv:1908.03379) fairly, and its reliance on the authors' own earlier T1/T2 measurements is transparent; the revision should nevertheless state explicitly which inputs are taken from prior work and which are re-measured here."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: the central claim—single-atom ESR at zero external field using only the tip field—is well supported, and the paper is honest about what it can't yet nail down. That combination makes this a genuine capability advance, not a reordering of the field but a real step toward making ESR-STM available in more labs.\n\nWhat's new: the vector-field dependence of the ESR amplitude, the constant-frequency tip-field sweep, and zero-external-field ESR. The latter two matter because they remove the need for an external magnet and for wideband RF cabling, which are the main practical barriers to the technique. The experiments are direct: you see the resonance peak shift with conductance, and the fixed-frequency current sweep produces a clean resonance. The Fe moment from the zero-field slope, 4.29 ± 0.79 μB, overlaps with the external-field value 5.35 ± 0.14 μB, which is a good internal consistency check.\n\nSoft spots: the tip-field calibration rests on a single offset measurement at 50 mT. The SI explicitly admits that not all measurements extrapolate to zero tip field at zero conductance, so the linear conversion used to build the top axis of Fig. 4b is an assumption. The ~300 mT sweep range is therefore an estimate, not a calibrated field measurement. That does not break the qualitative result, but the axis should be labeled as inferred.\n\nThe Rabi-rate extraction is the second soft spot. It depends on T1 values rescaled by factors of 3, 10, and 35 from prior work, and it assumes a decoherence probability of 1 for T2. The resulting order-of-magnitude improvement over earlier Fe ESR is model-dependent, not a direct measurement. The authors explicitly leave open the electric-field driving mechanism, so the Ω ∝ B_z^tip inference is reasonable but not settled.\n\nThe citation pattern looks fine: they build on their own prior formulas and measurements, which are inputs rather than circular targets.\n\nOverall, this paper is for the ESR-STM community and for groups considering adopting the technique. It deserves a serious referee. I'd recommend accepting it for peer review and asking the authors to add a sensitivity analysis for the Rabi rates and to relabel the tip-field axis as inferred. But the zero-field demonstration itself is solid.","headline":"Zero-field single-atom ESR is convincingly demonstrated; quantitative tip-field calibration and Rabi-rate extraction are softer than the main text suggests.","tokens_in":18335,"tokens_out":3805,"would_cite":true,"duration_ms":35375,"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":"Single-atom electron spin resonance can be driven entirely by the magnetic stray field of the scanning tunneling microscope tip, with zero external magnetic field applied.","keywords":["scanning tunneling microscopy","electron spin resonance","single atom","vector magnetic field","Fe on MgO","tip-field sweep","Rabi rate","zero external magnetic field"],"falsifier":"At zero external field, record the tunneling current at which the ESR resonance appears for several fixed radio frequencies, then repeat with a small known external field such as 50 mT and compare the implied $B_z^{\\mathrm{tip}}(I)$ relation across the full 100-500 pA range; if the relation is not linear or disagrees with the paper's 0.85 mT/pA conversion factor, the constant-frequency tip-field sweep is not a true field sweep.","tokens_in":17322,"feed_emoji":"🧲","tokens_out":9041,"duration_ms":83559,"temperature":0.7,"pith_summary":"The paper establishes that electron spin resonance of a single atom on a surface can be driven by the magnetic stray field of the spin-polarized scanning tunneling microscope tip alone, with no external magnetic field. Using a two-dimensional vector magnet, the authors map how the resonance frequency and signal amplitude respond to out-of-plane and in-plane external fields, and they show that moving the tip closer to the atom strengthens the tip field, increases the Rabi rate by roughly an order of magnitude over earlier Fe ESR experiments, and shifts the resonance. They convert this shift into a measurement mode: sweeping the tip-to-sample conductance at a fixed radio frequency produces a tip-field sweep. A sympathetic reader would care because, if the claim holds, single-atom ESR becomes practical in commercial STM systems that lack external magnets and at a single fixed frequency.","feed_headline":"Single-atom spin resonance now works with no external magnet","feed_subtitle":"Tip-field sweeps replace frequency sweeps, so lab magnets and broadband RF cables are no longer required.","key_machinery":"The load-bearing identity is the paper's Eq. (1), $f_0 = (2\\mu_{\\mathrm{Fe}}/h)(B_z^{\\mathrm{ext}} + B_z^{\\mathrm{tip}})$, which makes the resonance frequency a direct readout of the sum of the external out-of-plane field and the $z$-component of the spin-polarized tip's stray field. The experimental trick is that $B_z^{\\mathrm{tip}}$ is proportional to the tunneling setpoint current, so a constant-frequency sweep of the conductance is a sweep of the magnetic field. On the driving side, the exchange-coupling model gives a Rabi rate proportional to the tip field gradient, hence $\\Omega/V_{\\mathrm{RF}} \\propto B_z^{\\mathrm{tip}}$, which explains both the enhanced driving at close approach and the use of the tip field as the only source of Zeeman splitting.","core_discovery":"For individual Fe atoms on two atomic layers of MgO on Ag(001), the resonance condition is $f_0 = (2\\mu_{\\mathrm{Fe}}/h)(B_z^{\\mathrm{ext}} + B_z^{\\mathrm{tip}})$, with the iron magnetic moment extracted as $\\mu_{\\mathrm{Fe}} = (5.35 \\pm 0.14)\\,\\mu_B$, independent of the in-plane field. The stray field from the spin-polarized tip, $B_z^{\\mathrm{tip}}$, grows linearly with tunneling conductance, and the normalized Rabi rate $\\Omega/V_{\\mathrm{RF}}$ grows with it, reaching values about an order of magnitude larger than in previous Fe ESR-STM work. The paper demonstrates ESR at zero external magnetic field by using the tip field for the Zeeman splitting, and it shows a constant-frequency mode in which the setpoint current is swept to sweep the tip field across a range equivalent to roughly 40 GHz of frequency. It also finds that a large in-plane external field improves the ESR peak amplitude, with a maximum near $B_{\\parallel}^{\\mathrm{ext}} \\approx \\pm 1.5$ T.","pith_inferences":["If the linear tip-field calibration survives closer approach, the constant-frequency tip-field sweep could be used as an atomic-scale field scanner to map the stray fields of other magnetic nanostructures without any external magnet.","The same tip-field driving mechanism should apply to other high-anisotropy atomic spins on polar insulating films, not just Fe, whenever a magnetic exchange interaction couples tip and atom strongly enough.","Because the sweep axis is the tip-sample distance, this mode couples ESR spectroscopy to topographic feedback, suggesting an automated route to spin-resonance mapping across a surface at fixed RF frequency."],"forward_implications":["ESR-STM no longer requires an external magnet: the tip's own field can provide the Zeeman splitting, so the technique can run in STM systems that have no vector magnet.","Measurements can be performed at one fixed radio frequency, relaxing the need for broadband RF cabling and allowing the field to be swept faster by changing the tip current rather than the generator frequency.","The tip-field sweep covers roughly 300 mT of field, equivalent to a frequency window of about 40 GHz, so a single sweep captures a much wider resonance range than a frequency sweep at fixed field.","Bringing the tip closer increases the Rabi rate by about an order of magnitude for Fe atoms, because the tip field gradient and exchange coupling grow with conductance.","An in-plane external field around $\\pm 1.5$ T maximizes the ESR amplitude, but zero in-plane field is workable when the tip field is strong enough."],"supporting_citations":[{"why":"Supplies the original single-atom ESR-STM method and the tunneling-magnetoresistance detection scheme that this work builds on.","marker":"7"},{"why":"Provides the saturation formalism and decoherence relations used to extract Rabi rates and linewidths from power-dependent ESR amplitudes.","marker":"10"},{"why":"Gives the exchange-coupling theory predicting a Rabi rate proportional to the tip field gradient, the mechanism invoked for tip-field driving.","marker":"14"},{"why":"Provides the constant-amplitude radio-frequency sweep technique used to deliver the drive to the tunnel junction.","marker":"23"},{"why":"Shows the tip magnetic field as a tunable local field on single atoms and supports the linear tip-field-versus-conductance calibration.","marker":"25"},{"why":"Reports comparable Rabi rates for Ti atoms driven by tip field gradients and establishes the reference for the tip-field driving mechanism.","marker":"26"},{"why":"Establishes the all-electrical pump-probe method and the millisecond spin lifetime of Fe, from which the T1 estimates under ESR conditions are extrapolated.","marker":"28"},{"why":"Explains the perpendicular magnetic anisotropy and large orbital moment of Fe atoms on MgO, underlying the resonance formula and moment estimate.","marker":"27"}],"fun_headline_variants":["No external magnet needed for single-atom spin resonance","Tip magnetic field enables zero-field ESR on single atoms","In-plane fields boost single-atom spin resonance signal","Single-atom ESR at constant frequency via tip-field sweeps","Zero-field electron spin resonance with a scanning tip"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The tip's magnetic field is assumed to be a single-valued, monotonic, and nearly linear function of the setpoint tunneling current over the entire sweep, with that relation calibrated from one 50 mT offset measurement; if the tip's magnetization or its magnetic interaction with the atom changes nonlinearly as the tip gets close, the field axis and the zero-field resonance assignment become distorted.","fun_headline_variants_meta":{"raw":{"variants":["No external magnet needed for single-atom spin resonance","Tip magnetic field enables zero-field ESR on single atoms","In-plane fields boost single-atom spin resonance signal","Single-atom ESR at constant frequency via tip-field sweeps","Zero-field electron spin resonance with a scanning tip"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000876,"raw_usage":{"total_tokens":3826,"prompt_tokens":1019,"completion_tokens":2807,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":635,"completion_tokens_details":{"reasoning_tokens":2730}},"tokens_in":635,"tokens_out":2807,"duration_ms":21244,"temperature":1.0,"reasoning_tokens":2730,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:25:41.183659+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"At zero external field, record the tunneling current at which the ESR resonance appears for several fixed radio frequencies, then repeat with a small known external field such as 50 mT and compare the implied $B_z^{\\mathrm{tip}}(I)$ relation across the full 100-500 pA range; if the relation is not linear or disagrees with the paper's 0.85 mT/pA conversion factor, the constant-frequency tip-field sweep is not a true field sweep.","supporting_citations":[{"cited_title":"Electron paramagnetic resonance of individual atoms on a surface","cited_arxiv_id":null,"evidence_quote":"Supplies the original single-atom ESR-STM method and the tunneling-magnetoresistance detection scheme that this work builds on."},{"cited_title":"Probing quantum coherence in single-atom electron spin resonance","cited_arxiv_id":null,"evidence_quote":"Provides the saturation formalism and decoherence relations used to extract Rabi rates and linewidths from power-dependent ESR amplitudes."},{"cited_title":"Exchange mechanism for electron paramagnetic resonance of individual adatoms","cited_arxiv_id":null,"evidence_quote":"Gives the exchange-coupling theory predicting a Rabi rate proportional to the tip field gradient, the mechanism invoked for tip-field driving."},{"cited_title":"P.; Heinrich, A","cited_arxiv_id":null,"evidence_quote":"Provides the constant-amplitude radio-frequency sweep technique used to deliver the drive to the tunnel junction."},{"cited_title":"J.; Lutz, C","cited_arxiv_id":null,"evidence_quote":"Shows the tip magnetic field as a tunable local field on single atoms and supports the linear tip-field-versus-conductance calibration."},{"cited_title":"Tuning the Exchange Bias on a Single Atom from 1 mT to 10 T","cited_arxiv_id":null,"evidence_quote":"Reports comparable Rabi rates for Ti atoms driven by tip field gradients and establishes the reference for the tip-field driving mechanism."},{"cited_title":"Control of the millisecond spin lifetime of an electrically probed atom","cited_arxiv_id":null,"evidence_quote":"Establishes the all-electrical pump-probe method and the millisecond spin lifetime of Fe, from which the T1 estimates under ESR conditions are extrapolated."},{"cited_title":"P.; Macfarlane, R","cited_arxiv_id":null,"evidence_quote":"Explains the perpendicular magnetic anisotropy and large orbital moment of Fe atoms on MgO, underlying the resonance formula and moment estimate."}],"review_version":1}