{"id":"2725e9f1-1c5c-4520-b3b4-a29759b16c4a","arxiv_id":"2507.16357","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Using fast electronic modulation of the carrier-envelope phase of near-infrared pulses, the authors isolate phase-coherent sub-cycle photocurrents in an STM junction from an incoherent background about 200 times larger.","lead":"This paper demonstrates that near-infrared laser pulses can create tiny currents in a scanning tunneling microscope that depend on the exact phase (carrier-envelope phase) of the light wave. It introduces a fast electronic way to modulate that phase and a lock-in readout that isolates these coherent currents from a much larger background.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Residual amplitude modulation at the lock-in frequency is not quantitatively excluded; the Fig. 3e integrated-spectrum comparison is too weak a bound for a 7th-order nonlinear junction, so the isolation claim in Eq. (1) is not yet established.","rationale":"The paper's central new capability is the claim that lock-in demodulation at the CEP modulation frequency isolates dI/dphi and rejects all power and thermal artifacts. The load-bearing assumption is exactly that the modulation is phase-pure. The authors do provide supporting evidence: a linear pump-current-to-CEP response (Fig. S1), saturation of EDFA2 as a power-compression stage, and a measured lock-in trace that oscillates with CEP and inverts. These are real and nontrivial. However, the quantitative sufficiency of the power-amplitude rejection is not demonstrated. The integrated-spectrum Fourier analysis in Fig. 3e is a visual comparison without an explicit upper bound, and the 0.2 Hz characterization does not address the 100 Hz lock-in condition. Given the extreme nonlinearity of the detection (7-photon photoemission and field-driven tunnelling), the tolerance on residual AM is very tight, below about 0.1%. I therefore do not see an internal inconsistency; I see an unverified quantitative premise that is directly testable. This is exactly the kind of issue that a conditional verdict should require, and it does not change the reader's CONDITIONAL recommendation.","tokens_in":11129,"tokens_out":5722,"duration_ms":66956,"concrete_test":"Apply the same electronic CEP modulation at the actual lock-in frequency (e.g., 127 Hz) and measure the output power with a fast photodiode and lock-in amplifier before and after EDFA2, reporting residual AM in dBc. Then use the measured 7th-order power-law exponent to estimate the spurious photocurrent as n*(delta P/P)*I_mean. If this exceeds about 20% of the measured lock-in signal amplitude, or if a phase-resolved measurement of the CEP swing at >100 Hz is not provided, the isolation claim is not supported. As a cross-check, record the lock-in X and Y quadratures during the CEP sweep: a pure derivative signal should produce a straight line through the origin with fixed phase, whereas AM contamination changes the trajectory and phase.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim rests on the premise that modulating EDFA1's pump current produces a pure CEP modulation at nu_CEP > 100 Hz after saturated EDFA2. Section 3 validates this at 0.2 Hz via f-2f tracking and an integrated-spectrum Fourier comparison (Fig. 3e) that shows 'no appreciable difference' from intrinsic noise. That comparison is not a quantitative bound on residual AM at the lock-in frequency, and it is made in the laser output spectrum, not at the STM junction. The junction amplifies the risk: the isolated-tip photocurrent follows a 7th-order power law (Fig. 2b), while the total-current CEP modulation is only about 0.5% (Fig. 2c). A residual power modulation of about 0.07% at nu_CEP would therefore create a photocurrent modulation of about 0.5%, comparable to the entire coherent signal being extracted. This AM component would appear at the same lock-in frequency, with amplitude and phase set by the pump/gain transfer function, and would contaminate the demodulated signal attributed to dI/dphi in Eq. (1). The zero crossing of the Fig. 4c trace is consistent with a pure derivative signal, but it is not decisive: a constant AM vector merely shifts the zero-crossing and phase of the sinusoid, and no quadrature components or error bars are reported. Additionally, the CEP modulation amplitude is only characterized at 0.2 Hz; at >100 Hz the pump-to-CEP transfer function, including roll-off and phase delay, is unverified.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports an ultrafast scanning tunnelling microscopy experiment driven by near-infrared single-cycle pulses, with two intertwined claims. First, CEP-dependent (coherent) photocurrents are observed across the STM junction, with a small sinusoidal modulation as the CEP is slowly swept by wedge insertion. Second, the authors introduce an 'electronic CEP modulation' scheme in which the pump current of the first Er:fiber amplifier is modulated at frequencies above 100 Hz, and the photocurrent is demodulated with a lock-in amplifier. They argue that this isolates the derivative dI/dphi of the photocurrent with respect to CEP, yielding a signal that shows full sign inversion and a claimed 200-fold enhancement in modulation depth relative to the total photocurrent, while remaining compatible with an engaged STM feedback loop. The central experimental evidence is the sign-flipping demodulated signal in Figure 4c, whose period matches the wedge calibration.","tokens_in":11463,"tokens_out":4763,"duration_ms":54841,"significance":"If the isolation claim is correct, this is a valuable methodological step for lightwave-driven STM at near-infrared frequencies: it provides a way to read out the small coherent, sub-cycle component of the tunnelling current while suppressing slow thermal artifacts and allowing the feedback loop to remain active. The derivative-sampling idea is simple and attractive, and the use of an f-2f interferometer to calibrate the CEP shift is a strength. The sign-flipping lock-in trace in Figure 4c is internally consistent with a pure derivative response and is not explained by a simple DC offset. However, the central quantitative claims—'100% modulation with full inversion around zero' and a '200-fold enhancement'—rest on the assumption that the electronic modulation produces a pure CEP modulation with negligible residual amplitude modulation at the lock-in frequency. That assumption is not quantitatively supported by the data presented, and the nonlinearity of the junction amplifies the risk of AM contamination.","major_comments":[{"comment":"The central claim that the demodulated signal is proportional to dI/dphi requires that the electronic CEP modulation does not introduce a significant residual amplitude modulation at the lock-in frequency. The manuscript supports this with an integrated-spectrum Fourier comparison showing 'no appreciable difference' between modulation on and off, but this is not a quantitative bound. Given the 7th-order power-law dependence of the isolated-tip photocurrent reported in Figure 2b, a residual power modulation of only about 0.07% at nu_CEP would produce a photocurrent modulation of about 0.5%, comparable to the entire coherent signal being extracted. The comparison must be quantified as an upper bound on residual AM, ideally measured at the STM junction or at least at the laser output with a specified noise floor, before the isolation claim in Eq. (1) can be considered established.","section":"Section 3, Figure 3e"},{"comment":"The CEP modulation amplitude delta_phi = 0.1 pi is characterized at a modulation frequency of 0.2 Hz using the f-2f interferometer, but the lock-in experiments are performed at nu_CEP > 100 Hz. The pump-to-CEP transfer function, including its amplitude roll-off and phase delay, is not measured at the operating frequency. If the actual delta_phi at >100 Hz differs from 0.1 pi, the calibration of the derivative signal and the claimed '200-fold enhancement' are not quantitatively meaningful. At minimum, the authors should report the lock-in reference settings and either measure delta_phi at the operating frequency or provide an explicit argument that the transfer function is flat up to the lock-in frequency.","section":"Sections 3 and 4, Figure 3b and Eq. (1)"},{"comment":"The key experimental evidence for the isolation method is the demodulated photocurrent trace in Figure 4c, but the manuscript provides no error bars, confidence intervals, or statistical test for the zero crossing and the sine fit. The gray individual scans show substantial scatter, and the fit is described only as a 'guide to the eye.' The claims of '100% modulation with full inversion around zero' and a '200-fold enhancement of modulation depth' are therefore not supported by a quantitative uncertainty analysis. The authors should report the lock-in time constant, the number of independent averages, the noise floor of the detection, and a meaningful uncertainty on the fitted amplitude and phase, ideally including a comparison of the in-phase and quadrature components to exclude a constant AM vector.","section":"Section 4, Figure 4c"},{"comment":"The paper claims that as the tip-sample distance is reduced from the isolated-tip regime to tunnelling proximity, the CEP-modulated current amplitude 'doubles,' and this is used to support the interpretation that field-driven tunnelling sets in at small gaps. However, Figure 2d shows no error bars on the extracted amplitudes, and the exponential fits are only guides to the eye. Given that the underlying data in Figure 2c are noisy at small gaps, the doubling could be within the scatter. This claim should either be quantified with propagated uncertainties or softened to a qualitative observation.","section":"Section 2, Figure 2d"}],"minor_comments":[{"comment":"The sentence 'The results of the measurements are shown in figure 2b' should refer to Figure 2c, since Figure 2b shows the power-law fit for the isolated tip.","section":"Section 2, text near Figure 2c"},{"comment":"The phrase 'Analogously to the measurements discussed in figure 1c' appears to reference the wrong figure; the CEP-sweep measurements are shown in Figure 2c, not Figure 1c.","section":"Section 4, first paragraph"},{"comment":"The text refers to 'figure 3d' for the integrated spectral intensity, but the figure caption and panels show no panel d; the integrated-spectrum Fourier analysis is in panel e. The citation should be corrected.","section":"Section 3, text near Figure 3e"},{"comment":"There is a typo in the sentence 'We will address this aspect this in the remainder of this work'; the duplicated 'this' should be removed.","section":"Section 3, first paragraph"},{"comment":"The sentence 'The photocurrent is acquired directly from the transimpedance amplifier (no lock-in detection) and averaged over 30 successive CEP sweeps' is clear, but later in Section 4 the text says the total photocurrent modulation was 'around 0.5% in the retracted case (figure 2b, top panel)'; this should reference Figure 2c or 2d, as Figure 2b has no top panel.","section":"Section 2, text near Figure 2b"},{"comment":"Reference [40] (C. Lin et al., ACS Photonics) is a duplicate of reference [22]; the duplicate citation should be removed or replaced with the appropriate source.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within scope for physics.optics and the core idea is promising. The main concern is not circularity or internal inconsistency, but the lack of a quantitative exclusion of residual amplitude modulation at the lock-in frequency, which is load-bearing for the central isolation claim. I believe this can be addressed with additional measurements and analysis, so I recommend major revision rather than rejection. I would also encourage the editor to ensure the figure cross-reference errors are corrected in revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Worth a serious look. The paper demonstrates CEP-dependent photocurrents in a near-IR STM junction and introduces a fast electronic CEP modulation scheme (pump-current modulation of the preamplifier) combined with lock-in derivative sampling. That combination is genuinely new, and the sign-flipping lock-in signal in Fig. 4c is a strong internal consistency check: it behaves like a derivative, inverts around zero, and matches the wedge-calibrated CEP period. The authors also do a careful job of characterizing the slow CEP response and the f-2f calibration. For the field, this is a practical step toward sub-10-fs lightwave STM with feedback engaged, and it deserves referee time.\n\nThe main soft spot is exactly what the stress-test note flags: residual amplitude modulation at the lock-in frequency is not quantitatively excluded. The 7th-order power-law dependence of the multiphoton current means a tiny power ripple (around 0.07%) would produce a photocurrent modulation comparable to the 0.5% coherent signal. The Fig. 3e integrated-spectrum Fourier comparison is weak for that purpose—it is not a calibrated bound at nu_CEP, and it is taken at the laser output, not at the junction. Also, the CEP modulation amplitude is only characterized at 0.2 Hz; the transfer function at >100 Hz (including roll-off and phase) is assumed, not measured. These are addressable, but they are load-bearing for the claim that the lock-in signal is purely dI/dphi.\n\nA second, lesser issue: key quantitative claims—the gap-dependent amplitude doubling and the 200-fold enhancement—come without error bars, and the raw data are not deposited. The paper's own statements are appropriately hedged, but the reader cannot assess the significance of those numbers.\n\nOverall, the central physics is plausible and the method is clever. The concerns do not sink the paper; they call for a revision with added control measurements. I would send this to peer review, asking for a direct measurement of residual AM at the junction (or at least a photodiode-based lock-in bound), a high-frequency CEP characterization, and error bars on the main figures. A competent referee can handle it.\n\nFor my own work, I'd cite it once the AM purity is demonstrated; right now I'd treat the derivative-sampling result as promising but unproven.","headline":"A solid near-IR lightwave-STM method paper with a novel fast CEP modulation readout, but the residual-AM purity claim needs a quantitative check before the central isolation result is fully convincing.","tokens_in":12008,"tokens_out":1789,"would_cite":false,"duration_ms":22308,"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":"Fast electronic CEP modulation plus lock-in detection isolates the coherent sub-cycle part of the photocurrent in a near-infrared STM, turning a ~0.5% ripple into a 100%-modulated, sign-inverting signal.","keywords":["ultrafast scanning tunnelling microscopy","carrier-envelope phase","lightwave-driven STM","single-cycle near-infrared pulses","lock-in detection","sub-cycle photocurrents","coherent control","photoemission"],"falsifier":"With the electronic CEP modulation running at $\\nu_{CEP}>100$ Hz, place a fast photodiode in the beam after the second amplifier (or, inside the STM, use a large-gap multiphoton photoemission channel that is insensitive to CEP) and measure the amplitude modulation at $\\nu_{CEP}$. If the residual power modulation at that frequency is large enough to produce a lock-in signal comparable to the observed 100% inversion, the claim that the demodulated signal is purely $dI/d\\phi$ fails.","tokens_in":10942,"feed_emoji":"⚡","tokens_out":12556,"duration_ms":116157,"temperature":0.7,"pith_summary":"Scanning tunnelling microscopes illuminated by single-cycle near-infrared pulses ought to combine atomic spatial resolution with sub-femtosecond temporal resolution, but the coherent portion of the tunnelling current is buried under a much larger phase-independent background (multiphoton photoemission, thermal expansion, and other intensity-driven signals). This paper shows that the coherent part can be isolated by modulating the carrier-envelope phase (CEP) electronically at frequencies above 100 Hz and demodulating the photocurrent with a lock-in amplifier. The demodulated signal is proportional to $dI/d\\phi$ and exhibits 100% modulation with full inversion around zero, against only about 0.5% modulation of the total photocurrent: a 200-fold contrast enhancement and an order-of-magnitude faster acquisition. If correct, this provides a practical readout for lightwave-driven near-IR STM with the feedback loop engaged, a step toward attosecond-resolved scanning probe microscopy.","feed_headline":"Isolate sub-cycle STM currents with 200x contrast gain","feed_subtitle":"Fast electronic CEP modulation plus lock-in detection turns a 0.5% ripple into a fully phase-invertible signal.","key_machinery":"The load-bearing mechanism is 'electronic CEP modulation' combined with 'derivative sampling'. The pump current of the first erbium-fiber amplifier is modulated at $\\nu_{CEP}>100$ Hz, imprinting a sinusoidal CEP shift $\\delta\\phi\\simeq 0.1\\pi$; driving the second amplifier in saturation is intended to suppress the accompanying power modulation. A slow wedge sweep sets the CEP working point $\\phi_0$, and a lock-in amplifier demodulates the current at $\\nu_{CEP}$. The Taylor-expansion identity above turns the lock-in output into a direct measurement of $dI/d\\phi$, so the method reads out the slope of the photocurrent-phase curve rather than the current itself; this is what rejects the phase-independent background and isolates sub-cycle dynamics.","core_discovery":"The paper's central claim is that sampling the STM current with a small, fast CEP dither yields a lock-in signal proportional to the CEP derivative of the photocurrent, $I(\\phi_0+\\delta\\phi\\sin 2\\pi\\nu_{CEP}t)\\simeq I(\\phi_0)+(dI/d\\phi)|_{\\phi_0}\\delta\\phi\\sin 2\\pi\\nu_{CEP}t$, and that this signal is purely the coherent sub-cycle contribution. All phase-independent channels—multiphoton photoemission, photo-assisted tunnelling, thermal gap modulation—are rejected because they do not oscillate at $\\nu_{CEP}$. The authors demonstrate the method on an isolated tip and at tunnel-gap distances, finding that the CEP-oscillation amplitude roughly doubles as the tip approaches the sample, and that the demodulated photocurrent shows complete contrast inversion while the total photocurrent varies by only about 0.5%. They therefore conclude that coherent field-driven tunnelling at near-infrared carrier frequencies can be detected directly and rapidly with the STM feedback loop engaged.","pith_inferences":["Beyond the paper, a direct junction-level measurement of residual amplitude modulation at $\\nu_{CEP}$ would turn the saturated-amplifier suppression argument from a plausibility into a tested bound; the integrated-spectrum Fourier comparison in the paper is not a quantitative junction-level test.","The same derivative-sampling readout could be transferred to other scanning-probe modulation dithers (bias, polarization, pump current) wherever the observable is a smooth background-dominated function of the modulated parameter.","A systematic study of the sign-inversion pattern on controlled asymmetric versus symmetric tips could provide a quantitative check of the theoretical prediction that a perfectly symmetric junction shows full CEP inversion.","Applied to synthesized two-colour waveforms, the derivative signal could serve as a field-sampling observable, mapping the sub-cycle electric-field waveform rather than only the CEP."],"forward_implications":["The near-IR lightwave-driven STM can be operated with the feedback loop engaged, because the CEP dither at $\\nu_{CEP}>100$ Hz is faster than the $\\sim$100 Hz feedback bandwidth and therefore not compensated by gap adjustments.","The derivative readout rejects all phase-independent background, so the coherent signal can be measured without subtracting the multiphoton and thermal contributions.","Acquisition times drop by roughly an order of magnitude relative to slow mechanical CEP sweeps, making long-term stabilized measurements practical.","The observed doubling of the CEP-oscillation amplitude as the gap shrinks supports field-driven tunnelling across the junction as the source of the coherent signal.","The 100% modulation with full inversion gives a sensitive, sign-discriminating handle for setting and stabilizing the CEP at the junction."],"supporting_citations":[{"why":"Supplies the theoretical prediction that CEP-dependent current oscillations identify coherent laser-driven tunnelling, the signal this paper isolates.","marker":"[11]"},{"why":"Demonstrates sub-cycle optical phase control of nanotunnelling in the single-electron regime, giving the wedge-based CEP control convention and field-tunnelling framework.","marker":"[19]"},{"why":"Provides the Er:fiber laser technology on which the 40 MHz CEP-stable source is built.","marker":"[33]"},{"why":"Describes the passively phase-locked Er:fiber source of single-cycle near-IR pulses whose amplifier chain is used for electronic CEP modulation.","marker":"[34]"},{"why":"Gives the two-dimensional spectral shearing interferometry used to retrieve the spectral phase and the single-cycle waveform.","marker":"[35]"},{"why":"Presents alternative modulation strategies (polarization, envelope duration) that the CEP-modulation scheme is designed to supersede.","marker":"[43]"},{"why":"Reports an earlier near-IR STM experiment with ambiguous evidence for attosecond coherent manipulation, motivating the new readout.","marker":"[16]"}],"fun_headline_variants":["CEP dither isolates sub-cycle STM currents","Fast CEP modulation yields clean STM signal","Lock-in CEP trick gives 200x contrast in STM","Sub-cycle STM: tiny CEP ripple, full contrast flip","Thermal-free photocurrent readout via CEP dither"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument's load-bearing premise is that the fast modulation of the first amplifier's pump current changes only the carrier-envelope phase, with the saturated second amplifier suppressing any accompanying pulse-energy modulation below the level that would masquerade as a phase-driven lock-in signal at the CEP modulation frequency.","fun_headline_variants_meta":{"raw":{"variants":["CEP dither isolates sub-cycle STM currents","Fast CEP modulation yields clean STM signal","Lock-in CEP trick gives 200x contrast in STM","Sub-cycle STM: tiny CEP ripple, full contrast flip","Thermal-free photocurrent readout via CEP dither"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000624,"raw_usage":{"total_tokens":2862,"prompt_tokens":891,"completion_tokens":1971,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":507,"completion_tokens_details":{"reasoning_tokens":1888}},"tokens_in":507,"tokens_out":1971,"duration_ms":16703,"temperature":1.0,"reasoning_tokens":1888,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T15:11:16.744701+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"With the electronic CEP modulation running at $\\nu_{CEP}>100$ Hz, place a fast photodiode in the beam after the second amplifier (or, inside the STM, use a large-gap multiphoton photoemission channel that is insensitive to CEP) and measure the amplitude modulation at $\\nu_{CEP}$. If the residual power modulation at that frequency is large enough to produce a lock-in signal comparable to the observed 100% inversion, the claim that the demodulated signal is purely $dI/d\\phi$ fails.","supporting_citations":[{"cited_title":"Ultrafast Spin Dynamics in Ferromagnetic Nickel,","cited_arxiv_id":null,"evidence_quote":"Supplies the theoretical prediction that CEP-dependent current oscillations identify coherent laser-driven tunnelling, the signal this paper isolates."},{"cited_title":"Photoassisted scanning tunneling microscopy,","cited_arxiv_id":null,"evidence_quote":"Provides the Er:fiber laser technology on which the 40 MHz CEP-stable source is built."},{"cited_title":"Thermal effects in pulsed laser assisted scanning tunneling microscopy,","cited_arxiv_id":null,"evidence_quote":"Describes the passively phase-locked Er:fiber source of single-cycle near-IR pulses whose amplifier chain is used for electronic CEP modulation."},{"cited_title":"Thermal expansion of scanning tunneling microscopy tips under laser illumination,","cited_arxiv_id":null,"evidence_quote":"Gives the two-dimensional spectral shearing interferometry used to retrieve the spectral phase and the single-cycle waveform."},{"cited_title":"Optical antenna properties of scanning probe tips: Plasmonic light scattering, tip-sample coupling, and near-field enhancement,","cited_arxiv_id":null,"evidence_quote":"Presents alternative modulation strategies (polarization, envelope duration) that the CEP-modulation scheme is designed to supersede."}],"review_version":1}