{"id":"7059aefb-3e12-47dc-bcc7-c1d966eb6074","arxiv_id":"2412.13718","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":8,"one_line_summary":"THz pump-probe scanning tunneling microscopy captures picosecond charge-state dynamics of single selenium vacancies in WSe2, resolving transient Coulomb blockade and showing Franck-Condon blockade suppresses back tunneling to the tip.","lead":"Using terahertz pump-probe scanning tunneling microscopy, researchers watched single selenium vacancies in a one- and two-layer semiconductor charge and discharge on picosecond time scales. The experiment resolves transient Coulomb blockade at an atomic quantum dot and shows that a phonon-based 'backflow valve' can stop electrons from tunneling back to the tip.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The assumed 1 ps phonon relaxation time is unmeasured and controls the Franck-Condon blockade; no sensitivity analysis is provided, so the central lifetime-vs-bias claim is underconstrained.","rationale":"After reading the paper, the experimental observation of a pump-probe transient that recovers with a timescale matching the dc-measured charge-state lifetime is compelling: the time-domain signal is localized at the defect orbital, and the extracted lifetimes track τ0 from current saturation. The weakest link is not the existence of the transient Coulomb blockade but the attribution of the bias-dependent relaxation to the Franck-Condon blockade. That attribution is mediated entirely by a master-equation model whose most unconstrained input is τ_ph: it is not fitted to any static data, is justified by a room-temperature carrier-cooling reference, and directly sets the effectiveness of the blockade. The reader's weakest assumption listed τ_ph among a bundle of model parameters; this stress-test singles it out as the single most load-bearing because, unlike S and ℏΩ (which are at least fit to the STS lineshape), τ_ph is assumed ad hoc. An independent check is feasible: a sensitivity scan of τ_ph in the existing simulation, without new experimental data, would reveal whether the conclusion is stable. If the simulation is insensitive to τ_ph, the concern can be retired; if not, the Franck-Condon interpretation requires a direct measurement of the defect's vibronic relaxation time. The verdict should remain conditional: the paper's qualitative demonstration is sound, but acceptance should require the sensitivity analysis (or a three-pulse measurement) before the Franck-Condon blockade is established as the mechanism.","tokens_in":22832,"tokens_out":13236,"duration_ms":121863,"concrete_test":"Using the master-equation simulation (Methods, Eqs. 1-6) with all parameters as in the paper, repeat the computation of the pump-probe spectra and the extracted τ_eff(ΔV) for Fig. 5c and 5f with τ_ph = 0.1, 1, 10, 100 ps. Quantify the fit quality (e.g., reduced χ²) against the experimental τ_eff(ΔV) points. If the data constrain τ_ph to 1 ps (i.e., τ_ph ≥ 10 ps is excluded), the concern does not land; if τ_ph can be varied by ≥2 orders of magnitude without degrading the fit, the Franck-Condon interpretation is underdetermined and the claim of lifetimes approaching τ_0 is not supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that Franck-Condon blockade suppresses back tunneling near ΔV ≈ 0 and thereby allows τ_eff to approach the intrinsic τ_0 rests on the charged state having vibrationally relaxed to its ground state before the back-tunneling window opens. The model implements this as a decay to the vibrational ground state with lifetime τ_ph (Methods, Eq. 6), and the paper sets τ_ph = 1 ps as a 'conservative' assumption based on carrier thermalization in few-layer MoS2 (ref. 38). This is an order-of-magnitude extrapolation: the relevant mode is a localized 8 meV defect phonon at 5 K, whose lifetime may be much longer than the free-carrier cooling time. If τ_ph ≥ 10 ps, the electron would occupy excited vibronic levels of the charged state during the ~ps measurement, and the Franck-Condon overlap with low-lying neutral states would be larger, weakening the blockade and reducing τ_eff near ΔV=0. The paper provides no sensitivity analysis of τ_ph, and the single-mode model (ℏΩ=8 meV, S=2.2) is itself only an approximation; the authors use a second mode (ℏΩ2=20 meV, S2=0.7) to fit the static dI/dV (Extended Fig. S6b) and concede the 1-mode model breaks down at ΔV≈10 mV (Methods). Thus the quantitative agreement of τ_eff(ΔV) with the Franck-Condon mechanism is not established independently of this assumption.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports THz pump–THz probe scanning tunneling microscopy measurements on individual selenium vacancies in monolayer and bilayer WSe2 and claims the first time-domain, atomic-scale observation of transient Coulomb blockade at a single defect. The central observable is the rectified charge per THz transient, measured as a function of dc bias, tip–sample distance, and pump–probe delay; from the delay traces the authors extract an effective charge-state lifetime τeff. They find that τeff grows as the dc gate approaches the defect LUMO and near ΔV ≈ 0 approaches the intrinsic, substrate-limited lifetime τ0. This suppression of back tunneling to the tip is attributed to the Franck–Condon blockade and modeled with a master-equation rate model that includes one vibrational mode (Methods, Eqs. 1–6). The paper also presents orbital-resolved differential images at selected delays, which are interpreted as snapshots of the transient Coulomb blockade.","tokens_in":23125,"tokens_out":5813,"duration_ms":54020,"significance":"If the central claim holds, this is a notable experimental advance: it would be the first real-space, time-domain imaging of Coulomb blockade at an atomic-scale quantum dot, and it would identify Franck–Condon blockade as a practical mechanism for achieving unidirectional charge transfer in lightwave-driven STM. The experimental methodology is strong: multiple defects are studied, several coupling regimes are varied systematically, the THz waveform is calibrated in situ, and the rate-equation model is described in enough detail to be reconstructed. The paper is also commendably transparent about some limitations, such as the acknowledged breakdown of the single-phonon-mode model at ΔV ≈ 10 mV and the presence of trailing THz reflections. However, the attribution of the observed lifetime-versus-bias behavior to Franck–Condon blockade currently rests on several fitted or assumed model parameters, most importantly the unmeasured phonon relaxation time τph, so the quantitative evidence for the mechanism is not yet established independently of those assumptions.","major_comments":[{"comment":"The Franck–Condon blockade mechanism requires the charged defect to relax to its vibrational ground state before the back-tunneling window opens, and the model implements this as a decay to the ground state with lifetime τph = 1 ps. This value is an assumption extrapolated from free-carrier thermalization in few-layer MoS2 (ref. 38) to a localized 8 meV defect mode at 5 K. If τph is an order of magnitude longer, transitions from excited vibronic levels of the charged state to low-lying neutral states would remain available, weakening the blockade and reducing τeff near ΔV = 0. No sensitivity analysis is provided. Because this single parameter controls the central mechanism, the manuscript should report how the predicted τeff(ΔV) and QLW(Vdc) curves change when τph is varied over a plausible range (e.g., 0.1–10 ps), or provide an independent experimental bound on τph.","section":"Methods, Eq. (6) and “Implementation of simulation”"},{"comment":"The master-equation model is not a parameter-free simulation. The Huang–Rhys factors (S = 2.2 for 2 ML, S = 5 for 1 ML), phonon energy ħΩ = 8 meV, 3 meV Gaussian broadening, tip–LUMO+1 coupling factor Λ, effective capacitor-plane distance zVD = 13 Å, and a 0.4 Å z0 offset are all fitted or chosen to match the same sample's static dI/dV and I(z) data. The claimed “quantitative reproduction” of QLW(Vdc), QLW(z), and τeff(ΔV) therefore has limited evidentiary weight for the Franck–Condon mechanism unless the model is validated out-of-sample or the parameter uncertainties are propagated into the predicted observables. The authors' own statement that the 1-mode treatment breaks down at ΔV ≈ 10 mV further narrows the predictive range over which the comparison in Fig. 3c can be taken as support for the mechanism.","section":"Methods, “Implementation of simulation”; Extended Figs. S6b–S6g"},{"comment":"The effective lifetimes are extracted by exponential fits over the restricted window 1.5–15 ps, while the claimed saturation near ΔV = 0 is compared with intrinsic lifetimes τ0 = 55–86 ps that exceed the fit window by a factor of 4–50. With trailing THz reflections at up to 30% of the main pulse amplitude (Extended Fig. S2g) and the authors' own statement that fit errors become large for τeff ≳ 20 ps, the data do not robustly establish that τeff quantitatively reaches τ0. The claim requires either a longer-time measurement with the reflection problem controlled, a quantitative uncertainty analysis of the truncated-window fits, or a demonstration that the same fitting protocol applied to the simulation is insensitive to the window choice.","section":"Fig. 5 and “Time-domain detection of the ultrafast Coulomb blockade”"}],"minor_comments":[{"comment":"The heading “COULOMB BLOCKADE A T A SINGE SELENIUM V ACANCY” contains several typographical errors and should be corrected.","section":"Section heading"},{"comment":"The sign convention for ΔV is inconsistent across the paper: some passages and Fig. 5 axes use ΔV = Vdc − VLUMO, while other captions and the Methods define the offset as VLUMO − Vdc. A single, clearly stated convention should be used throughout.","section":"Definitions of ΔV"},{"comment":"Extended Fig. S6b shows that a two-mode Franck–Condon model with an additional 20 meV mode (S2 = 0.7) is needed to reproduce the static dI/dV spectrum, but the simulations in Figs. 3c, 4c, and 5 use only the 8 meV mode. The main text should explicitly state that the 20 meV mode is omitted from the dynamical simulations and discuss how its inclusion would affect the predicted back-tunneling asymmetry.","section":"Methods, “Implementation of simulation”"},{"comment":"The color scale for the ΔQLW images in Fig. 4b is not defined in the caption; please state the reference used for the differential quantity.","section":"Fig. 4b caption"},{"comment":"The data and code availability statements say materials are available “upon reasonable request.” Given the number of fitted parameters in the model, depositing the master-equation code and representative raw pump–probe traces would substantially strengthen reproducibility.","section":"Data availability"}],"recommendation":"major_revision","confidential_remarks":"The experimental work is impressive and likely of significant interest to the readership. My main concern is not the validity of the observation but the evidentiary weight of the Franck–Condon blockade interpretation: the key mechanism rests on an unmeasured τph and on model parameters fitted to the same sample. I would be comfortable with a major revision that adds a sensitivity analysis, explicitly limits the claims to the range where the model is valid, and quantifies the uncertainty of the extracted lifetimes. The paper fits the journal's scope; I have no concerns about novelty disclosure."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: this is a genuine experimental step forward, and the data deserve a close look from referees. The authors show picosecond time-domain sampling of a transient Coulomb blockade at a single Se vacancy in WSe2, and they demonstrate that the extracted effective lifetime approaches the intrinsic value when the dc bias is tuned to the LUMO. That is the kind of measurement people have been trying to get with lightwave-driven STM.\n\nWhat is genuinely new: the THz pump-probe snapshots of the defect orbital at different delays, the bias-dependent lifetime curve, and the claim that Franck-Condon blockade suppresses back tunneling. The model is also described transparently; the authors state which parameters come from static STS, from approach curves, and which are assumed. The consistency across several defects and coupling regimes gives me confidence the main observation is real.\n\nThe soft spots are in the mechanism, not the measurement. The Franck-Condon blockade argument rests on a single-mode vibronic model with tau_ph=1 ps, which is an order-of-magnitude guess from carrier cooling in MoS2. No sensitivity analysis is given, and the paper itself admits the 1-mode model breaks down at Delta V~10 mV. If the true phonon relaxation time is significantly longer, the excited vibronic population in the charged state would weaken the blockade and bias the extracted tau_eff near Delta V=0. That doesn't invalidate the pump-probe observation, but it means the central physical explanation is underdetermined. I also agree with the reader that the 'first direct observation of ultrafast electron dynamics at atomic scale with LW-STM' is overstated relative to refs 28 and 29, and the 'data available upon reasonable request' policy is not acceptable for a result of this importance—the data and simulation code should be released.\n\nThis is a paper for the lightwave-STM and 2D-defect communities. It deserves peer review and, after revision, likely publication. The revision should add a sensitivity analysis on tau_ph and the phonon spectrum, release the code and data, and temper the novelty claim. I would not desk-reject it.","headline":"A real experimental milestone—time-domain Coulomb blockade at a single defect—with a plausible but under-constrained Franck-Condon mechanism; referee it, but insist on sensitivity analysis and data release.","tokens_in":23767,"tokens_out":2158,"would_cite":true,"duration_ms":19177,"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":"This paper reports direct, real-space, time-domain observation of transient Coulomb blockade at a single selenium vacancy in WSe2 using THz pump–THz probe scanning tunneling microscopy.","keywords":["Coulomb blockade","charge-state lifetime","ultrafast STM","THz pump–probe","Franck–Condon blockade","selenium vacancy","WSe2","lightwave-driven nanoelectronics"],"falsifier":"Measure the THz pump–probe trace on a selenium vacancy with a substantially weaker electron–phonon coupling (small Huang–Rhys factor): the Franck–Condon blockade predicts that τeff should stay short at ΔV ≈ 0 because back tunneling remains strong, so a persistent rise in τeff would rule out the proposed mechanism.","tokens_in":22568,"feed_emoji":"⚡","tokens_out":5623,"duration_ms":47857,"temperature":0.7,"pith_summary":"The paper claims to capture the transient Coulomb blockade of an atomic-scale quantum dot — a selenium vacancy in monolayer and bilayer WSe2 — in real time and real space, using pairs of picosecond THz pulses delivered through the junction of a scanning tunneling microscope. The central experimental result is that the rectified charge measured by a delayed probe pulse relaxes exponentially with an effective charge-state lifetime that approaches the defect's intrinsic lifetime when the dc bias is tuned to the LUMO resonance. The paper attributes this bias-dependent suppression of back tunneling to the tip to the Franck–Condon blockade, and supports the assignment with a master-equation model that reproduces the bias, distance, and time-delay dependence of the lightwave-driven current. A sympathetic reader would care because this is the first time-domain, atomic-scale view of Coulomb blockade at a single defect, and it opens a route to controlling unidirectional electron transfer in lightwave-driven nanoelectronics.","feed_headline":"Ultrafast Coulomb blockade filmed at a single atomic defect","feed_subtitle":"Time-delayed THz pulses clock electron charging of one selenium vacancy in WSe2, key for lightwave nanoelectronics.","key_machinery":"The mechanism that carries the argument is the Franck–Condon blockade, implemented in a time-dependent master equation over three electronic states (neutral ground state, LUMO, LUMO+1) dressed by a single vibrational mode with Franck–Condon factors. The paper defines the parameters — ℏΩ = 8 meV, Huang–Rhys factor S = 2.2 for 2 ML and S = 5 for 1 ML, 3 meV Gaussian broadening, and phonon relaxation time τph = 1 ps — and uses them to compute the rectified charge from the actual THz waveform including reflections. A modified plate-capacitor voltage drop is introduced to model the distance dependence.","core_discovery":"Using THz pump–THz probe time-domain sampling, the authors record atomic-scale snapshots of the transient Coulomb blockade of individual selenium vacancies in WSe2 and extract effective charge-state lifetimes. For a vacancy in 1 ML WSe2 the intrinsic lifetime is about 2.2 ps, while in 2 ML WSe2 it ranges up to 86 ps; the measured effective lifetime rises steeply as the dc bias approaches the LUMO from below and peaks near the intrinsic value at ΔV ≈ 0. The paper's key claim is that this behavior is governed by the Franck–Condon blockade: because vibrational relaxation is fast, transitions occur from the vibrational ground state of the charged defect, and at high bias the overlap with accessible low-lying vibrational states of the neutral defect is strongly reduced, so back tunneling to the tip is suppressed. The same master-equation model, with a single 8 meV vibrational mode and Huang–Rhys factor 2.2, reproduces the rectified-charge curves as functions of bias, tip–sample distance, and pump–probe delay.","pith_inferences":["The sharpness of τeff(ΔV) near resonance is a sensitive measure of electron–phonon coupling, so the technique could be used as a local, time-domain probe of Huang–Rhys factors.","Because the paper's model uses a single phonon mode and the authors note its breakdown near ΔV ≈ 10 mV, quantitative lifetime extraction may carry systematic errors; a two-mode or temperature-resolved treatment could test whether low-energy interlayer phonons dominate the blockade.","The same back-tunneling-suppression strategy might be generalizable to other atomic defects in two-dimensional materials and to other selection rules, such as spin or orbital angular momentum, as the authors suggest."],"forward_implications":["Lightwave-driven STM can now probe localized charge dynamics even when charge-state lifetimes exceed the THz pulse duration, provided back tunneling is suppressed by the Franck–Condon blockade.","Measuring τeff as a function of dc bias near the defect resonance gives access to the intrinsic charge-state lifetime τ0 from a time-domain experiment.","The Franck–Condon blockade acts as a charge backflow valve, so the same junction can be operated in weak-injection, strong-injection, and blockade-limited regimes by choosing bias and tip height.","Tuning bias within one phonon energy of the resonance switches between back-tunneling-dominated and forward-transfer-dominated regimes, enabling unidirectional charge transport.","The approach can track charge transfer into the substrate with picosecond temporal resolution and atomic spatial resolution."],"supporting_citations":[{"why":"Supplies the layer-dependent average charge-state lifetimes of single selenium vacancies in WSe2 that set the intrinsic τ0 values used throughout the paper.","marker":"[16]"},{"why":"Provides the theory of the Franck–Condon blockade regime that the paper invokes to explain suppression of back tunneling.","marker":"[37]"},{"why":"Establishes the vibronic sidebands and spin-orbit structure of chalcogen vacancies in TMDs and informs the Huang–Rhys parameter choices.","marker":"[18]"},{"why":"Describes the phase- and amplitude-controlled THz-STM method used for state-selective tunneling and voltage calibration.","marker":"[32]"},{"why":"Provides the current-saturation method used to extract average charge-state lifetimes for calibration of τ0.","marker":"[15]"},{"why":"Demonstrates the Franck–Condon blockade in quantum dots experimentally, serving as precedent for assigning the blockade signature.","marker":"[39]"},{"why":"Reports earlier lightwave-STM work on selenium vacancies in WSe2, providing the coherent-dynamics context and comparison for the incoherent charging studied here.","marker":"[28]"}],"fun_headline_variants":["THz pump-probe catches Coulomb blockade in a single defect","Atomic-scale quantum dot's Coulomb blockade clocked by THz","Ultrafast Coulomb blockade in one selenium vacancy","Franck-Condon blockade readies ultrafast single-electron control","THz pulses time electron charging of a single WSe2 vacancy"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The interpretation rests on a single-mode vibronic model with fitted parameters — Huang–Rhys factor 2.2, phonon energy 8 meV, 3 meV broadening, and 1 ps phonon relaxation time — and the paper itself states that the one-phonon-mode treatment breaks down at bias offsets near 10 mV.","fun_headline_variants_meta":{"raw":{"variants":["THz pump-probe catches Coulomb blockade in a single defect","Atomic-scale quantum dot's Coulomb blockade clocked by THz","Ultrafast Coulomb blockade in one selenium vacancy","Franck-Condon blockade readies ultrafast single-electron control","THz pulses time electron charging of a single WSe2 vacancy"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000729,"raw_usage":{"total_tokens":3281,"prompt_tokens":982,"completion_tokens":2299,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":598,"completion_tokens_details":{"reasoning_tokens":2215}},"tokens_in":598,"tokens_out":2299,"duration_ms":18484,"temperature":1.0,"reasoning_tokens":2215,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T12:51:34.942027+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the THz pump–probe trace on a selenium vacancy with a substantially weaker electron–phonon coupling (small Huang–Rhys factor): the Franck–Condon blockade predicts that τeff should stay short at ΔV ≈ 0 because back tunneling remains strong, so a persistent rise in τeff would rule out the proposed mechanism.","supporting_citations":[{"cited_title":"Layer-Dependent Charge State Lifetime of Single Se Vacancies in WSe$_2$","cited_arxiv_id":"2407.04508","evidence_quote":"Supplies the layer-dependent average charge-state lifetimes of single selenium vacancies in WSe2 that set the intrinsic τ0 values used throughout the paper."},{"cited_title":"& Andreev, A","cited_arxiv_id":null,"evidence_quote":"Provides the theory of the Franck–Condon blockade regime that the paper invokes to explain suppression of back tunneling."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes the vibronic sidebands and spin-orbit structure of chalcogen vacancies in TMDs and informs the Huang–Rhys parameter choices."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Describes the phase- and amplitude-controlled THz-STM method used for state-selective tunneling and voltage calibration."},{"cited_title":"& Gross, L","cited_arxiv_id":null,"evidence_quote":"Provides the current-saturation method used to extract average charge-state lifetimes for calibration of τ0."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Demonstrates the Franck–Condon blockade in quantum dots experimentally, serving as precedent for assigning the blockade signature."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reports earlier lightwave-STM work on selenium vacancies in WSe2, providing the coherent-dynamics context and comparison for the incoherent charging studied here."}],"review_version":1}