{"id":"2e8ce0f4-1d4c-494c-834f-99febebd3709","arxiv_id":"2607.21515","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"A fiber microcavity now holds up to about 36 individually trapped and imaged atoms with strong single-atom coupling, enabling high-cooperativity atom arrays.","lead":"This experiment places optical tweezer arrays of individual rubidium atoms inside a high-finesse fiber microcavity and images them with single-atom resolution. It shows that strong cavity coupling and site-resolved detection can now work together in arrays of tens of atoms, opening cavity QED to many-body experiments.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Collective-cooperativity extraction from averaged VRS is not per-shot validated; direct arithmetic from displayed Ω, N̄, and κ,γ does not reproduce quoted C values.","rationale":"The paper's core experimental achievements—single-atom C≈90, background-free site-resolved imaging, and compatibility of cavity readout across a 20×5 footprint—are well supported by the presented data. The weak step is the quantitative leap from a collective vacuum-Rabi spectrum averaged over stochastic arrays to a per-atom cooperativity for those arrays. This leap directly supports the 'up to N̄≃36' high-cooperativity claim, so it is load-bearing. The reader flagged state preparation and dark atoms as the main assumption; I agree but would sharpen it: the relation g=Ω/√N̄ is applied to averaged spectra without per-shot correlation, and the displayed numbers do not reproduce the quoted C values. This is an internal-consistency concern rather than a disagreement with consensus. It can be settled by straightforward re-analysis of the existing data—binning by N and recomputing the relation—so a conditional acceptance is appropriate. If the re-analysis confirms the quoted C values, the concern disappears; if not, the array-level cooperativity claims require revision. The single-atom and imaging claims should not be cast into doubt; the conditional status applies specifically to the collective-array cooperativity numbers.","tokens_in":10831,"tokens_out":16357,"duration_ms":167594,"concrete_test":"First, recompute C=(Ω/√N̄)^2/(κγ) for each configuration using the stated Ω, N̄, κ, and γ; if the quoted C values do not follow, the fitting convention or the stated formula is inconsistent. Then re-analyze the raw spectra: use single-shot fluorescence images to bin the VRS data by measured atom number N (or occupancy pattern), verify Ω(N)=g√N with g equal to the 62.1-MHz single-atom value, and fit a model that includes a dark/imperfect-state fraction. If the binned data deviate from √N scaling at fixed g, report the effective coupled-atom number and corrected mean C.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The quantitative basis for the central 'high-cooperativity ... up to N̄≃36' claim is the conversion g=Ω/√N̄ applied to the vacuum-Rabi spectra in Fig. 4b. This conversion is load-bearing but not secured. (1) The spectra are averages over ~30 cycles with stochastic atom number. Since transmission depends nonlinearly on n, the averaged spectrum is not the spectrum of N̄ atoms; the fitted Ω can be biased. The paper never conditions a spectrum on the measured atom number from the fluorescence image in the same cycle. (2) As written, the numbers do not close: with κ/2π=14.2 MHz and γ/2π=3.0 MHz, the chain values N̄=9.41(6) and Ω/2π=249(2) MHz give g/2π=81.2 MHz and C≈155, not the quoted C=92(3); for the 20×4 array, N̄=36.5(1) and Ω/2π=194(1) MHz give g/2π=32.1 MHz and C≈24, not C=39(1). Either the displayed Ω values are not the half-splitting used in g=Ω/√N̄, or the fit implicitly corrects for dark state-preparation, inhomogeneous coupling, and number fluctuations in a way not described. The text itself notes asymmetry from imperfect stretched-state preparation, so these effects are not negligible. The single-atom C=88(4) is robust; the array-level C values are not independently supported as presented.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports the combination of optical tweezer arrays of single 87Rb atoms with a high-cooperativity fiber Fabry–Perot microcavity. The authors characterize the single-atom interface, measuring a vacuum Rabi splitting that yields g/2π = 62.1(5) MHz and C = 88(4), and demonstrate cavity-based hyperfine-state readout with ~99% fidelity. They then implement a background-free two-photon fluorescence imaging scheme, achieving single-site detection fidelities of 99(1)% and survival probabilities of 96(1)% across a 20×3 array. Finally, they couple 20×1, 20×2, 20×3, and 20×4 arrays to the cavity and, from collective vacuum Rabi spectra, report mean single-atom cooperativities of C = 92(3), 63(2), and 39(1) for the largest configurations, with a mean atom number up to N̄ ≃ 36. The central claim is that this establishes a high-cooperativity platform for many-body cavity-QED with site-resolved detection and control.","tokens_in":11145,"tokens_out":10581,"duration_ms":99361,"significance":"If the array-level cooperativity claims are correct, the paper represents a substantial advance: it brings high-cooperativity cavity-QED (C ~ 90) together with tens of individually resolved and controlled atoms, enabling studies of collective light-matter interactions, cavity-mediated spin models, and quantum-network nodes with site-resolved readout. The single-atom result (C = 88(4)) is supported by a standard vacuum Rabi measurement, and the fluorescence-imaging and cavity-readout demonstrations are independently valuable, with high fidelities and low loss. The machine-checked internal consistency of the numbers, however, is currently incomplete, and the central array-level quantitative claims need clarification before the significance can be fully accepted.","major_comments":[{"comment":"The displayed numbers are internally inconsistent as written. In Fig. 4b the caption states that the separation between the two polaritonic peaks is 2Ω, and Fig. 4c defines the measured single-atom coupling as g = Ω/√N̄. Using the quoted values and (κ,γ)/2π = (14.2,3.0) MHz, the 20×1 case (Ω/2π = 249(2) MHz, N̄ = 9.41(6)) gives g/2π = 81.2 MHz and C = g²/κγ ≈ 155, not C = 92(3). For the 20×4 case (Ω/2π = 194(1) MHz, N̄ = 36.5(1)) the formula gives g/2π = 32.1 MHz and C ≈ 24, not C = 39(1). Either Ω is not the half-splitting used in the conversion, or the extraction uses additional corrections (state-preparation efficiency, inhomogeneous coupling, atom-number statistics) that are not described. This point is load-bearing for the central claim of high-cooperativity arrays; please clarify the exact definition of Ω, the fitting model, and the relationship between Ω and the quoted C values.","section":"Fig. 4b, 4c"},{"comment":"The vacuum Rabi spectra are averaged over ~30 experimental cycles with stochastic atom number (loading probability ~56%). Because the cavity transmission depends nonlinearly on the atom number, the averaged spectrum is not the spectrum of N̄ atoms, and a simple double-Lorentzian fit to the averaged data can systematically bias the extracted splitting. The paper does not state that the fit accounts for the atom-number distribution, nor does it provide per-shot conditioning of the spectra on the fluorescence-measured atom number. Please either include a model for the number statistics in the fit, show simulated averaged spectra with the claimed parameters, or present a conditional analysis. This is necessary to validate the quantitative C values for the arrays.","section":"Fig. 4b"},{"comment":"The text mentions that 'the observed asymmetry between the two polariton peaks may arise from imperfect preparation in the stretched state and/or from a small impurity of the probe polarization.' Such effects can also shift the fitted peak positions and alter the inferred splitting. Since the authors themselves identify these imperfections as non-negligible, the analysis should quantify how the extracted C values change under plausible amounts of state-preparation error and polarization impurity, or justify that the effect is small compared to the reported uncertainties.","section":"Fig. 4c, final paragraph"}],"minor_comments":[{"comment":"The notation for Ω is used inconsistently: the text sometimes refers to 'collective vacuum Rabi splitting' and the figure caption defines 2Ω as the peak separation, but the formula g=Ω/√N̄ uses Ω as a half-splitting. Please define Ω explicitly in one place and keep the same convention throughout.","section":"Throughout"},{"comment":"The labels 'Fluorescence 1' and 'Fluorescence 2' in Fig. 3a are helpful, but the color scale in Fig. 2b is in photons/pixel and the right panels are not fully described; a scale bar or axis label would improve readability.","section":"Fig. 2 and Fig. 3"},{"comment":"Reference [48] is cited only in a 'Note added'; if the independent work is directly relevant, consider citing it in the introduction alongside [17-26].","section":"Introduction, first paragraph"}],"recommendation":"major_revision","confidential_remarks":"The reader's report recommends acceptance, but the arithmetic inconsistency in Fig. 4c is a genuine load-bearing issue that should be resolved before publication. The single-atom and imaging parts of the paper are strong, so I do not recommend rejection; however, the array-level cooperativity values need a rigorous fit model and clear definition of Ω. Please also check whether the quoted Ω values in Fig. 4b are actually the half-splittings; if they are, the resulting C values do not match the text."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe headline is right: this is the first fiber microcavity with single-atom cooperativity around 90 combined with background-free, site-resolved fluorescence imaging of extended multi-row tweezer arrays, reaching mean atom numbers around 36. That is a real step beyond the two-atom high-cooperativity demonstrations and worth taking seriously. The single-atom work is careful: C=88(4), the cavity mode mapping, cavity-assisted hyperfine readout at 99.4(3)%, and the imaging characterization with fidelity ~99(1)% and survival ~96(1)% all look consistent and honest.\n\nThe weak point is the array-level cooperativity extraction. The text and Fig. 4c say g = Ω/√N̄, where Ω is the half-separation of the polariton peaks, and report C=92(3) for the 20x1 chain and C=39(1) for the 20x4 array. But plugging in the displayed values — Ω/2π=249 MHz, N̄=9.41 for the chain, and Ω/2π=194 MHz, N̄=36.5 for the array — with κ/2π=14.2 MHz and γ/2π=3.0 MHz gives C≈155 and C≈24, not the reported numbers. The paper never describes a fit model that could reconcile this, and the asymmetry noted from imperfect state prep suggests the discrepancy is not negligible. This inconsistency is load-bearing because the main claim rests on those array C values.\n\nI don't think the experiment is wrong; the single-atom numbers and imaging are strong, and the collective effect is likely real. But the paper needs to show the actual fit, define Ω precisely, and either correct the formula or correct the quoted C values. The reader's take missed this; the stress-test arithmetic is correct.\n\nThe citation and literature coverage look fair, including the note about the simultaneous preprint. Send it to peer review — it deserves a serious referee — but expect a revision request to make the collective-coupling analysis auditable before the headline number can be trusted.\n\nBest,","headline":"High-cooperativity fiber-cavity tweezers with site-resolved imaging is a genuine advance, but the array-level C values don't close with the paper's own formula.","tokens_in":11689,"tokens_out":7080,"would_cite":true,"duration_ms":68638,"reading_group":"yes","serious_thinker":"unclear","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["42.50.Pq"],"model":"deepseek-v4-flash","headline":"This work demonstrates that tens of single atoms in an optical tweezer array can be trapped inside a high-cooperativity fiber microcavity, imaged with background-free fluorescence, and collectively coupled to a common cavity mode, establish","keywords":["cavity quantum electrodynamics","optical tweezer arrays","fiber Fabry-Perot microcavity","vacuum Rabi splitting","cooperativity","background-free fluorescence imaging","87Rb atoms","collective coupling"],"falsifier":"Measure how the collective vacuum Rabi splitting changes as single atoms are removed from known sites (e.g., by moving a row to a node or emptying sites with a blast beam), and compare the incremental change to the single-atom coupling measured at that site. If the splitting does not follow the expected √N scaling with site-specific couplings, or if the per-atom contribution drops with array size, the coherent-coupling assumption is falsified.","tokens_in":10702,"feed_emoji":"⚛️","tokens_out":6279,"duration_ms":57350,"temperature":0.7,"pith_summary":"This work reports a platform that combines the two ingredients needed for many-body cavity-QED with microscopic control: strong coupling of individual atoms to a common cavity mode, and site-resolved fluorescence imaging of tens of atoms at once. The authors demonstrate single-atom cooperativity of about 90 in a fiber Fabry-Perot microcavity, then show that an array of up to 36 atoms positioned inside the same cavity collectively couples to the mode, with a measured collective vacuum Rabi splitting that matches the expected √N enhancement. They also develop a background-free two-photon fluorescence imaging scheme that works close to the cavity mirrors, achieving about 99% detection fidelity and 96% survival per site. If correct, this establishes a route to cavity-mediated entangling gates, mid-circuit measurements, and quantum simulation with programmable atom arrays.","feed_headline":"36 atoms strongly coupled to one cavity mode","feed_subtitle":"Fiber microcavity plus tweezer imaging opens many-body cavity-QED with site-resolved control.","key_machinery":"The load-bearing mechanism is a hybrid trap: an 808-nm tweezer array positions atoms inside the cavity, and a co-located 1559-nm intracavity lattice, commensurate with the 780-nm probe standing wave, confines each atom at an antinode, reducing the effect of position jitter. The collective signal is interpreted through the identity Ω = √N̄ g, which converts the measured vacuum Rabi splitting into a mean single-atom cooperativity C = g²/(κγ). The background-free imaging relies on a two-photon 'diamond' excitation (780 + 1529 nm) with decay through the 5P1/2 state and photon collection at 795 nm, which filters out the excitation light and eliminates the fluctuating background scattered from the","core_discovery":"The paper reports the combination of two capabilities in one apparatus: single-atom strong coupling to a fiber Fabry-Perot microcavity (cooperativity C≈88, measured via vacuum Rabi splitting) and high-fidelity site-resolved fluorescence imaging of up to 20×4-site tweezer arrays (mean atom number up to N̄≈36). Using an intracavity 1559-nm lattice to pin atoms at antinodes of the probe field, the authors observe collective vacuum Rabi splittings for 1D chains and 2D arrays, and extract a mean single-atom cooperativity from Ω/√N̄ that matches the expected Gaussian mode profile. They also implement a background-free two-photon imaging scheme on the D1 line that suppresses scattered light from th","pith_inferences":["If combined with deterministic rearrangement, the same platform would allow error-corrected cavity-mediated entangling gates and mid-circuit measurements on registers of tens of qubits, not just two.","The two-photon background-free imaging scheme is likely transferable to other photonic interfaces, such as nanophotonic cavities or waveguides, where surface scattering limits imaging.","Site-resolved imaging during collective coupling could directly probe how cavity-mediated interactions modify local observables, e.g., by imaging correlations after a quench.","A direct test of the coherent-coupling assumption would be to measure the collective splitting while blocking individual rows or moving one row to a node; a discrepancy from the predicted change would reveal the fraction of dark atoms."],"forward_implications":["Cavity-based hyperfine-state detection remains above 98.8% fidelity even at the edges of a 20×5-site geometry, so cavity readout can be extended to hundred-site arrays.","The background-free fluorescence scheme works in immediate proximity to macroscopic surfaces, making site-resolved imaging possible inside microcavities where it previously failed.","Collective coupling of about 36 atoms with per-site cooperativity above 5 everywhere is demonstrated, enabling cavity-mediated interactions across a programmable array.","The scaling of the extracted cooperativity with transverse position matches the Gaussian mode profile, confirming that detected atoms are pinned near antinodes.","Based on the measured mode profile, about 60 atoms could be coupled nearly homogeneously with cooperativity above 60, and up to about 150 atoms with controlled inhomogeneity."],"fun_headline_variants":["36 atoms in one high-cooperativity cavity","Fiber cavity couples 36 atoms in an array","Site-resolved strong coupling for 36 atoms","High-C cavity-QED scales to 36-atom array","Many-body cavity-QED with 36 atoms coupled"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The extraction of per-atom cooperativity assumes that every atom seen in the fluorescence image is coherently coupled to the cavity mode, remains near an antinode, and is prepared in the stretched state; if a notable fraction of imaged atoms are dark or out of phase, the inferred cooperativity for the largest arrays is too high.","fun_headline_variants_meta":{"raw":{"variants":["36 atoms in one high-cooperativity cavity","Fiber cavity couples 36 atoms in an array","Site-resolved strong coupling for 36 atoms","High-C cavity-QED scales to 36-atom array","Many-body cavity-QED with 36 atoms coupled"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000344,"raw_usage":{"total_tokens":1679,"prompt_tokens":652,"completion_tokens":1027,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":396,"completion_tokens_details":{"reasoning_tokens":951}},"tokens_in":396,"tokens_out":1027,"duration_ms":9593,"temperature":1.0,"reasoning_tokens":951,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T07:11:56.504280+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure how the collective vacuum Rabi splitting changes as single atoms are removed from known sites (e.g., by moving a row to a node or emptying sites with a blast beam), and compare the incremental change to the single-atom coupling measured at that site. If the splitting does not follow the expected √N scaling with site-specific couplings, or if the per-atom contribution drops with array size, the coherent-coupling assumption is falsified.","supporting_citations":[],"review_version":1}