{"id":"bdb170de-ff62-4d01-9b46-90c3fb61725d","arxiv_id":"2501.06162","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Depth-switching a dipole trap after atom detection raises single-atom filling from the 0.5 collisional-blockade limit to about 0.79.","lead":"This paper shows that switching a single-atom optical trap from shallow to deep after an atom arrives can raise average trap occupancy from 50% to about 79%. The method could reduce the need for atom rearrangement in neutral-atom quantum computers.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 79% filling ratio is a calculated upper bound from Eq. (5), not a measured value; no closed-loop feedback experiment is performed, so detection latency and switching dead time could reduce the realized occupation below the reported value.","rationale":"The reader's weakest assumption correctly identifies the gap between the theoretical filling-fraction formula and the experimental evidence. My independent reading of the manuscript confirms that the experiment is an open-loop, post-selected measurement of survival probabilities, not a closed-loop implementation of the proposed feedback. The 79% value is a calculated estimate, explicitly labeled as 'maximum achievable' and 'estimated' in the text, yet the abstract states 'This method demonstrates an achievable filling ratio of (79±2)%', which overstates the directness of the evidence. The central physical mechanism is sound: the measured shallow-trap dark time (2.07 ± 0.25) s and deep-trap lifetime (7.92 ± 0.15) s are consistent with an alternating renewal model, and even with realistic detection latency the scheme would likely exceed 0.5. The paper also contains independent supporting evidence: exponential lifetime fits, Wilson confidence intervals, and a multi-trap ensemble average. Therefore the result is promising but not fully demonstrated as claimed. The reader's conditional verdict is appropriate, and my stress-test does not identify a reason to move the verdict; I would keep it CONDITIONAL, which maps to UNCHANGED because the reader already chose CONDITIONAL. The concrete test I propose would either validate the 79% figure or quantify the overhead that reduces it, thereby settling whether the abstract's claim is justified.","tokens_in":9458,"tokens_out":12949,"duration_ms":123577,"concrete_test":"Run the actual feedback loop: continuously image fluorescence at about 100 ms per frame; when a single-atom threshold is crossed in the shallow trap, switch the AOM to the deep power; when the fluorescence drops below threshold, switch back to shallow. Record the total time the trap is occupied over at least 100 load-loss cycles and compare the measured filling fraction with eta = tau/(tau_d + tau). Also log the delay from threshold crossing to completed depth change. If the measured occupancy is substantially below 0.79 (for instance, below 0.65), then the neglected latency, false triggers, and switching dead time are material to the central claim.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that depth modulation pushes the time-averaged occupation probability beyond 0.5 is supported only by a theoretical calculation, not by a demonstration of the proposed feedback loop. Equation (5), eta = tau/(tau_d + tau), is the steady-state occupancy of an alternating renewal process in which the trap waits in the shallow state for a detected atom and then instantly switches to the deep state. The experiment instead applies a fixed 30 s open-loop modulation and post-selects intervals that contain an atom at the shallow-to-deep switch (Section V, Figure 6). Under open-loop modulation with 30 s in each state, the unconditional time-averaged occupancy remains near the shallow-regime steady-state value of 0.49-0.5, so the experiment itself does not exceed 0.5. The reported (79 ± 2)% is obtained by inserting separately measured lifetimes into Eq. (5), assuming that detection of an atom is immediate, that the AOM/SLM depth switch adds no dead time, and that false threshold crossings are negligible. These assumptions are not tested: the 250 ms time resolution of the periodic measurement bounds but does not quantify detection latency, and the paper gives no false-trigger rate. If the detection latency is a few hundred milliseconds or the switch takes tens of milliseconds, the effective dark time increases and the realized filling fraction falls; for example, a 0.5 s latency would reduce eta to roughly 0.75, still above 0.5 but below the headline value. Because the claim is explicitly about an achievable filling ratio, the missing closed-loop verification is the load-bearing soft spot. The underlying rate measurements are internally consistent and physically plausible, so the concern is about overreach in the strength of the claim rather than a contradiction in the data.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a scheme to increase the time-averaged single-atom occupation probability of an optical dipole trap beyond the collisional-blockade limit of 0.5 by switching the trap depth after an atom is detected. The authors measure trap lifetimes and dark times as functions of trap depth, introduce the filling-fraction formula eta = tau/(tau_d + tau) (Eq. 5), and estimate a maximum filling ratio of (79 +/- 2)%. To support this, they perform a periodic 30 s open-loop depth-switching experiment and analyze occupation probabilities conditional on an atom being present at the shallow-to-deep switch. They report that this method could avoid atom rearrangement for filling arrays.","tokens_in":9731,"tokens_out":5223,"duration_ms":53756,"significance":"The underlying measurements are careful: lifetime and dark-time data are extracted from exponential fits with Wilson confidence intervals, and the paper checks its main estimate against an independent calculation from Figure 5. The central idea, if validated in closed loop, would be a simple and power-efficient alternative to atom rearrangement for increasing array filling. The formula in Eq. (5) is physically motivated and parameter-free given the independently measured timescales. However, the headline filling ratio is not directly demonstrated by the experiment as written: the reported (79 +/- 2)% is a model-based projection that assumes instantaneous, error-free switching upon detection, while the experiment itself uses open-loop modulation and post-selected conditional occupation probabilities. This gap between claim and evidence is the main correctness risk.","major_comments":[{"comment":"The claimed filling ratio of (79 +/- 2)% is not directly demonstrated. Equation (5) describes a closed-loop renewal process in which the trap waits in the shallow state until an atom is detected and then immediately switches to the deep state. The experiment instead switches the trap depth on a fixed 30 s cycle and analyzes only intervals in which an atom was present at the switch time (Section V, Fig. 6). These conditional occupation probabilities do not by themselves establish that the time-averaged occupation of the proposed feedback protocol would be 0.79. In fact, the unconditional occupation during the shallow phase remains near the steady-state value of about 0.5, and the deep-phase benefit is only realized for intervals that started with an atom present. The abstract and conclusion state that the method 'demonstrates an achievable filling ratio' of 79%, which overstates what the open-loop, post-selected measurement shows. Please reframe the claim as a model-based estimate and either implement the closed-loop protocol or provide a quantitative argument for why the open-loop conditional measurement bounds the closed-loop duty cycle.","section":"Section V, Eq. (5), Fig. 6"},{"comment":"The estimate eta = tau/(tau_d + tau) assumes that the trap depth is switched instantly upon atom detection, with negligible detection latency and no false triggers. The paper reports a 250 ms time resolution for the periodic measurement, which bounds but does not quantify the detection latency, and it provides no false-trigger rate for the atom-detection threshold. If a feedback loop introduces a latency L after an atom enters the trap, the effective dark time becomes tau_d + L, and the filling fraction is reduced to tau/(tau_d + L + tau). For the reported values tau_d = 2.07 s and tau = 7.92 s, a latency of 0.5 s lowers eta from 0.79 to about 0.75. The paper should either report the measured detection latency and false-trigger probability, or present the headline filling fraction as an upper bound under the ideal-switching assumption.","section":"Section V, Eq. (5)"},{"comment":"The ensemble-averaged deep-regime lifetime is (6.67 +/- 0.08) s, which differs from the single-trap value of (7.92 +/- 0.15) s by well over the quoted uncertainties. The reported eta = 0.79 +/- 0.02 is calculated from the single-trap lifetime. Using the ensemble lifetime with the same tau_d gives eta approximately 0.76, still above 0.5 but materially lower. The paper attributes the discrepancy to trap-to-trap variation but does not propagate this variation into the uncertainty of the achievable filling ratio for an array. Please state clearly whether 79% is a single-trap value and, if the method is intended for arrays, include trap-to-trap variability in the reported uncertainty or in the array-level filling estimate.","section":"Section V, Fig. 6(c)"}],"minor_comments":[{"comment":"There is a typo in the apparatus section: 'Semrok' should be 'Semrock' for the bandpass filter, and 'M 2 beam quality factor' should use a superscript (M^2) for clarity.","section":"Section III"},{"comment":"The caption says 'maximum achievable filling fraction' while the text and Section V refer to 'estimated filling fractions.' Please unify the wording so that the model-based nature of these values is clear throughout.","section":"Figure 5 caption"},{"comment":"The phrase 'demonstrates an achievable filling ratio' should be softened to 'indicates an achievable filling ratio' or 'predicts an achievable filling ratio,' since the closed-loop protocol itself was not implemented and the 79% value is derived from Eq. (5) using separately measured timescales.","section":"Abstract and Conclusion"},{"comment":"The Wilson score intervals for survival probabilities are appropriate, but the text should state the number of trials and the number of exponential fits used for each lifetime extraction, as small sample sizes can make the quoted standard deviations optimistic.","section":"Section IV"},{"comment":"The conditioning on 'an atom present upon switching' should be stated directly in the main text and in the figure caption. As written, the occupation probabilities in Fig. 6(b,c) could easily be misread as unconditional filling fractions, which they are not.","section":"Section V, Figure 6"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within the journal's scope and the core lifetime/dark-time data appear solid. The main issue is that the central numerical claim (79% filling ratio) is framed as a demonstration when the experiment only provides a conditional, open-loop measurement plus a model-based estimate. I believe this is fixable with a major revision that either adds a closed-loop demonstration or carefully rephrases the claims and quantifies the idealizing assumptions."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: the paper has a useful set of lifetime-vs-depth measurements and a plausible idea—switch a single-atom trap from shallow to deep after capture—but the headline 79% filling ratio is a calculated upper bound from Eq. (5), not something the experiment actually closed the loop on. Worth refereeing, but the abstract and conclusions need to be pulled back.\n\nWhat's new: I don't know of prior work that systematically measures both dark time and lifetime as a function of trap depth in the collisional-blockade regime, and uses those rates to estimate the benefit of depth switching. The data in Figures 3 and 4 look solid: exponential fits, Wilson intervals, multiple traps. The open-loop 30 s modulation trace in Figure 6 is a reasonable proof-of-concept that the two regimes have the expected time constants. The method is cheap relative to rearrangement and the array-level extension via SLM is plausible.\n\nWhere it's soft: the central claim as stated, 'demonstrates an achievable filling ratio of (79±2)%', is not what the experiment shows. They never run the feedback loop. They switch on a fixed 30 s cycle and post-select intervals with an atom at switch time. The 79% comes from plugging independently measured tau_d = 2.07 s and tau = 7.92 s into eta = tau/(tau_d+tau). That formula assumes instant depth switch and perfect detection with zero latency. The paper gives no false-trigger rate and only bounds detection latency by the 250 ms time resolution. A few hundred ms of dead time would pull the realized filling below 79%, though still above 0.5. Also 'without the need of rearranging atoms' overstates: you still have probabilistic per-trap occupation across an array; you just wait instead of moving atoms. Those are fixable framing issues, not fatal flaws. The model itself is standard renewal-process arithmetic and the measured rates are credible.\n\nThe citation pattern looks fine. The paper engages with the relevant loading and rearrangement literature. I don't see circularity: the filling estimate is not fit to the claimed outcome.\n\nBottom line: for a reader who wants to know whether depth-switching is worth trying, this is a useful data point. It deserves peer review, but a serious referee should ask for a true closed-loop demonstration or a heavily qualified abstract.","headline":"Useful lifetime-vs-depth data and a plausible switching scheme, but the 79% filling ratio is a calculated bound from Eq. (5), not a demonstrated closed-loop result.","tokens_in":10337,"tokens_out":2009,"would_cite":true,"duration_ms":19398,"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":"This paper claims that modulating the depth of a single-atom optical dipole trap can push its average occupation probability beyond the 0.5 limit set by collisional blockade, reaching a filling fraction of (79±2)% without rearranging atoms.","keywords":["single-atom traps","optical dipole traps","collisional blockade","filling fraction","trap depth modulation","atom loading","tweezer arrays","87Rb"],"falsifier":"A closed-loop test that switches the trap to its deep configuration immediately upon atom detection, and then measures the realized fraction of time the trap is occupied over many cycles, would settle the claim. If detection latency, false triggers, or finite switching time add dead time comparable to the 2.07 s dark time, the observed filling fraction will fall below 0.79 toward the 0.5 baseline.","tokens_in":9200,"feed_emoji":"⚛️","tokens_out":5701,"duration_ms":46884,"temperature":0.7,"pith_summary":"This paper demonstrates a method to beat the 0.5 time-averaged occupation probability that limits single-atom optical dipole traps operating in the collisional blockade regime. The idea is to load atoms in a shallow trap, where the average wait for a loading event is short, and then deepen the trap once an atom arrives, so the trapped atom lives longer. From measured lifetimes, the authors derive an achievable filling fraction of (79±2)% using a shallow-trap dark time of (2.07±0.25)s and a deep-trap lifetime of (7.92±0.15)s. This could allow near-deterministic filling of tweezer arrays without the need to rearrange atoms.","feed_headline":"Trap-depth switch lifts atom occupancy past 0.5 limit","feed_subtitle":"Shallow-trap loading plus deep-trap holding yields a 79% filling fraction without atom rearrangement.","key_machinery":"The central object is the depth-modulated optical dipole trap: a 1064 nm tweezer whose depth is switched by an acousto-optic modulator between a shallow loading configuration (about 0.8 mK) and a deeper holding configuration (about 1.7 mK). The identity carrying the argument is η = τ/(τ_d+τ), which expresses the achievable filling fraction in terms of the trap's dark time and lifetime; in the collisional blockade regime these are governed by independent Poisson processes, so the shallow trap loads quickly and the deep trap holds long, and switching at capture converts fast loading into long retention.","core_discovery":"The paper claims that modulating the depth of a static optical dipole trap can push the time-averaged occupation probability of a single-atom trap in the collisional blockade regime beyond the nominal ceiling of 0.5. The quantitative statement is a filling fraction η = τ/(τ_d+τ), with the trap held shallow while waiting for a loading event (dark time τ_d ≈ 2.07 s) and then switched to a deeper configuration that extends the trapped-atom lifetime (τ ≈ 7.92 s), yielding η = (79±2)%. The demonstration alternates trap power between 10 mW and 21 mW on a fixed 30 s cycle and analyzes the occupation probability in windows centered on the shallow-to-deep switch, conditional on an atom being present at switch time.","pith_inferences":["If detection latency and switching time can be made small relative to the 2.07 s dark time, a true closed-loop version should reach filling fractions closer to the authors' 0.85–0.88 estimates obtained with the cooling light off during holding.","The measured 0.79 is a post-selected upper bound for the fixed-cycle implementation; a fair comparison with rearrangement schemes would require an end-to-end filling-rate measurement including detection and decision time.","The depth-modulation trick may compose with other loading optimizations, such as two-atom collision engineering or gray-molasses loading, because it only requires a contrast between loading and holding timescales."],"forward_implications":["The filling fraction of individual optical tweezers in a collisional-blockade array can exceed 0.5 without atom relocation, so deterministic filling no longer requires rearrangement-equipped setups.","Because the enhancement is local, the scheme can be parallelized: occupation-triggered holograms can deepen only the traps that have captured an atom.","The method remains effective when steady-state filling is below 0.5, as long as the loading-regime dark time is short relative to the holding-regime lifetime.","The approach uses less optical power and involves fewer operations than rearrangement, scaling with array size rather than requiring complex transport sequences."],"supporting_citations":[{"why":"Establishes the collisional blockade regime and the 0.5 loading limit that this work targets.","marker":"[25]"},{"why":"Supplies the standard theory of optical dipole trap loading and loss rates used to motivate the timescale separation.","marker":"[24]"},{"why":"The atom-by-atom assembly reference against which the no-rearrangement filling scheme is contrasted.","marker":"[2]"},{"why":"Demonstrates three-dimensional atom-array assembly via relocation, the conventional approach this method aims to replace.","marker":"[5]"},{"why":"Shows dynamic holographic optical tweezers for in-situ array synthesis, the technology the proposed occupation-triggered hologram scheme extends.","marker":"[23]"},{"why":"Provides the MRAF algorithm used to generate the uniform 2×4 tweezer array in the experiment.","marker":"[30]"},{"why":"The Wilson score interval is used to place confidence bounds on all occupation-probability data.","marker":"[31]"}],"fun_headline_variants":["Trap-depth switch lifts atom filling to 79%","Shallow-to-deep trap ramp yields 79% atom occupancy","Depth-modulated loading beats 0.5 occupation ceiling","Switching trap depth boosts single-atom loading to 79%"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The estimate assumes the trap can be switched instantly and without error at the moment an atom is detected, when the actual experiment only alternates the depth on a fixed 30-second cycle and analyzes data selected for having an atom at the switch.","fun_headline_variants_meta":{"raw":{"variants":["Trap-depth switch lifts atom filling to 79%","Shallow-to-deep trap ramp yields 79% atom occupancy","Depth-modulated loading beats 0.5 occupation ceiling","Switching trap depth boosts single-atom loading to 79%"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000513,"raw_usage":{"total_tokens":2459,"prompt_tokens":874,"completion_tokens":1585,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":490,"completion_tokens_details":{"reasoning_tokens":1514}},"tokens_in":490,"tokens_out":1585,"duration_ms":14494,"temperature":1.0,"reasoning_tokens":1514,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T21:05:59.116919+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A closed-loop test that switches the trap to its deep configuration immediately upon atom detection, and then measures the realized fraction of time the trap is occupied over many cycles, would settle the claim. If detection latency, false triggers, or finite switching time add dead time comparable to the 2.07 s dark time, the observed filling fraction will fall below 0.79 toward the 0.5 baseline.","supporting_citations":[{"cited_title":"Schlosser, G","cited_arxiv_id":null,"evidence_quote":"Establishes the collisional blockade regime and the 0.5 loading limit that this work targets."},{"cited_title":"Grimm, M","cited_arxiv_id":null,"evidence_quote":"Supplies the standard theory of optical dipole trap loading and loss rates used to motivate the timescale separation."},{"cited_title":"McGloin, G","cited_arxiv_id":null,"evidence_quote":"Provides the MRAF algorithm used to generate the uniform 2×4 tweezer array in the experiment."},{"cited_title":"Pasienski and B","cited_arxiv_id":null,"evidence_quote":"The Wilson score interval is used to place confidence bounds on all occupation-probability data."}],"review_version":1}