{"id":"48ec2b37-55c5-4d6f-89ef-9e4f044907e5","arxiv_id":"2506.20924","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Mirror feedback from one polariton condensate to another makes two otherwise independent condensates lock phase at a distance, with no in-plane coupling.","lead":"Researchers coupled two separate exciton-polariton condensates by reflecting light from one onto the other through an external mirror, locking their phases together over a distance of about 24 cm. The work is a step toward programmable, long-range connections in polaritonic circuits that could serve as fast analog computers.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Fringes in Fig. 5(a) may be produced by the feedback spot overlapping the condensate image; no quantitative upper bound on this contribution is given.","rationale":"The reader's weakest assumption identifies the same load-bearing point: the fringes in Fig. 5(a) are attributed to phase locking of the condensates, but the exclusion of direct interference from the feedback beam is not quantitative. This is the single most important uncertainty because the entire claim of 'phase locking without planar coupling' rests on it. The experiment has genuine strengths: the no-feedback control in Fig. 5(b) shows that the fringes disappear when the feedback path is blocked, and the path-length calibration in Fig. 4 demonstrates that the interferometer delay is set correctly. The analytical model also provides a plausible mechanism for robust coherence despite the 1.6 ns delay. However, none of these elements strengthens the spatial-exclusion argument, which rests on an unquantified 'tiny fraction' of overlap. The proposed mask test is decisive and straightforward with the existing optics: blocking the feedback spot in the detection path cleanly separates the two interpretations. Because the paper is otherwise internally consistent and the concern is addressable without new physics, the conditional verdict already given is appropriate; no change is needed.","tokens_in":15689,"tokens_out":7452,"duration_ms":87174,"concrete_test":"Place an opaque spatial mask in an intermediate image plane of the collection optics (before the Michelson) that blocks the feedback-spot image while transmitting the image of condensate 2, and record the interferogram under otherwise identical conditions. If fringes persist across the condensate with the feedback spot blocked, direct feedback-beam interference is excluded and phase locking is confirmed; if the fringes vanish, the original contrast was contaminated by the feedback beam. As a quantitative cross-check, compute a 2D fringe-visibility map from the raw Fig. 5(a) interferogram: global phase locking predicts roughly uniform visibility across the condensate, whereas feedback-spot overlap produces a localized visibility peak at the spot position, decaying over the imaging point-spread function.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim requires that the fringes in Fig. 5(a) arise from phase locking of the two condensates as whole entities, not from the external feedback beam being imaged together with the condensate emission. The feedback beam is a coherent copy of condensate 1's field; if any part of it overlaps the overlapped condensate images at the camera, it will interfere with the image of condensate 1 (or with the seeded region of condensate 2) and produce fringes even without global phase locking. Section III, step 5 attempts to exclude this only qualitatively: the feedback spot was moved to one side of condensate 2, and the image overlap included 'only a tiny fraction of this spot.' No value for this fraction, no point-spread function, and no 2D visibility map are given. With the reported fringe visibility of only ~10%, a small residual overlap of the feedback spot with the imaged condensates could easily account for the observed contrast without phase locking of the condensates. The control in Fig. 5(b) shows that the fringes depend on the feedback path, but it does not distinguish injection-induced phase locking from direct coherent illumination of the detector by the feedback beam. Thus the load-bearing condition, that the fringes are a property of the condensate emission and not of the feedback beam, is the least securely established part of the argument.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports a scheme for coupling distant exciton-polariton condensates through an external mirror path rather than in-plane nearest-neighbor coupling. Two geometrically isolated condensates are created in a GaAs/AlGaAs microcavity; vertical emission from each condensate is imaged and retro-injected onto the other. The authors present interferometric evidence: with the feedback path unblocked, fringes appear across the two condensate images (Fig. 5(a)) with about 10% visibility, while blocking the feedback removes the fringes (Fig. 5(b)). They also reduce the delayed coupled GPE to a Kuramoto--Sakaguchi phase model and use a noisy Adler equation to argue that common-mode noise in the mirror path allows robust phase locking despite a round-trip delay (about 1.6 ns) longer than the single-condensate coherence time. The central claim is that this proves external, long-range, pairwise phase locking with no planar coupling, opening a route to programmable polariton networks.","tokens_in":15928,"tokens_out":8851,"duration_ms":91472,"significance":"If the experimental interpretation is correct, the work is a significant proof-of-principle: it would demonstrate a passive, high-bandwidth method for arbitrary long-range coupling between polariton condensates, going beyond the short-range geometric coupling of existing lattice simulators. The paper has several strengths: the control experiment with the feedback blocked is a meaningful check; the theoretical reduction to classical phase-oscillator models is clearly laid out and not used as a fitting target; and the experiment is presented with a step-by-step logic that is easy to follow. However, the load-bearing experimental identification of phase locking is not yet quantitatively secure, and the proposed noise-cancellation mechanism for surviving the long delay appears inconsistent with the mutual-injection geometry. These issues are addressable with additional measurements and analysis, but they currently prevent full acceptance of the central claim.","major_comments":[{"comment":"The central claim that the fringes in Fig. 5(a) arise from phase locking of the two condensates, rather than from direct interference of the feedback beam with the condensate images, is not quantitatively established. The text states that the feedback spot was moved to one side of condensate 2 and that the image overlap included 'only a tiny fraction of this spot,' but no upper bound on this fraction is given, and no point-spread function or two-dimensional visibility map is provided. Because the reported fringe visibility is only about 10%, a small residual overlap of the feedback spot with the condensate images could account for the observed contrast without any global phase locking. A concrete test would be to record the interferometer output with condensate 2 unexcited while the feedback spot is present, or to map the fringe visibility and phase spatially across the full condensate area to show that the entire condensate participates coherently.","section":"Section III, step 5, Fig. 5"},{"comment":"The 'no in-plane coupling' null result is asserted from the absence of interference fringes in Fig. 2(b), but no quantitative upper bound on the coupling is given. Without an estimate of the maximum coherent fraction that could have been hidden in the noise, the later attribution of phase locking exclusively to the external mirror path is not falsifiable. The authors should fit the overlapping-image interferogram with a model that includes a small coherent component and report an upper bound on the fringe visibility (or on the in-plane coupling strength) under the same experimental conditions used for the locking measurement.","section":"Section III, step 2, Fig. 2(b)"},{"comment":"The explanation of how phase coherence survives the 1.6 ns delay relies on a common-mode noise cancellation that appears inconsistent with the experimental geometry. The text states that 'both traps receive the same delayed field,' but in the mutual-injection configuration described by Eq. (2) and Section III, each condensate receives a delayed copy of the other condensate's field, not a common external field. The independent phase noises ξ1 and ξ2 of the two condensates therefore enter the relative-phase equation as (ξ2−ξ1)/2, and the claim that the effective relative-phase diffusion σ_ϕ is far smaller than the single-condensate diffusion σ is not justified. This undermines Eq. (18) and the statement that the locked linewidth is 'orders of magnitude narrower' than σ. The authors should either reformulate the noise model for mutual injection and show that the Adler-equation locking condition is met with realistic parameters, or explicitly restrict the common-mode claim to path-length noise such as mirror vibrations.","section":"Section V, Eqs. (15)-(18)"},{"comment":"The experimental evidence for 'deterministic phase locking' is based on a single line cut with a fringe visibility of about 10%, and no error bars, no number of acquisitions, and no statistical analysis are reported. A time-integrated fringe pattern at such low contrast does not by itself demonstrate a stable, deterministic relative phase; it could also reflect partial or intermittent correlation. The authors should provide multiple independent frames or a histogram of the extracted relative phase over time, together with the uncertainty in the visibility, to support the word 'deterministic' and to allow the reader to assess the significance of the 10% visibility.","section":"Section III, step 6 and Fig. 5(c)"}],"minor_comments":[{"comment":"The phrase 'pure, high-bandwidth analogt element' contains a typo; it should read 'analog element.'","section":"Abstract"},{"comment":"The word 'explicity' should be 'explicitly.'","section":"Section IV, first paragraph"},{"comment":"The sentence defining J and Jeff is grammatically awkward and should be split for clarity: 'J is the coupling strength between condensates; when coherence filtered, it becomes Jeff = J e^{-τ/Tc}.'","section":"Eq. (2)"},{"comment":"The phrase 'after only 200 ps < Tc < 1 ns' could be misread; it should be clarified that the single-trap coherence time Tc lies in the range 200 ps to 1 ns, so the delay τ ≈ 1.6 ns exceeds Tc.","section":"Section V, first paragraph"},{"comment":"The manuscript uses 'an spatial light modulator'; the correct article is 'a spatial light modulator.'","section":"Supplementary Material, Section II"}],"recommendation":"major_revision","confidential_remarks":"The manuscript fits the journal's scope and the idea is timely. The main uncertainty is not the theoretical reduction but the experimental identification of the fringe source; the lack of a quantitative bound on stray feedback overlap is the key risk. The Section V common-mode noise argument should be corrected or qualified regardless of the outcome of additional measurements, because as written it is physically questionable and is used to support the 'robustness' claim. If the authors can supply the requested control measurements and a corrected noise model, the paper could become a solid contribution."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my take on arXiv:2506.20924. The genuinely new thing is experimental: two geometrically isolated polariton condensates, with no detectable in-plane coupling, get phase-locked when their vertical emission is reflected back through an external imaging path and injected near the other condensate. The control experiment—blocking the feedback path kills the fringes—is the right one, and it makes the basic effect credible. The theory is a clean reduction to delay-coupled Kuramoto and noisy Adler equations; nothing surprising in the mathematics, but the common-mode noise argument for why locking can survive a 1.6 ns delay is worth reading.\n\nThe soft spots are real but not disqualifying. First, the fringe visibility is only about 10% and there is no statistical characterization: no error bars, no multiple runs, no 2D visibility map. Second, the paper's own exclusion of the most dangerous artifact—the feedback spot directly reaching the camera and interfering with the image—is qualitative. They say the spot was moved to one side and only a tiny fraction of it overlapped the images, but there is no number, no PSF, no upper bound. With 10% contrast, a small stray overlap could matter. That said, the authors' spatial argument is not silly: fringes are reported across the whole condensate area, and a localized feedback spot would not obviously produce that. Still, a 2D visibility map and a quantitative overlap bound would settle it. Third, the null check for in-plane coupling is shown but not bounded; a modest upper limit on J would make 'no planar coupling' more than a statement about one interferogram. Fourth, the theory has several free parameters (K, Δω, σ_φ, β) and is not quantitatively anchored to the measured device parameters; it is illustrative rather than predictive.\n\nThe citation pattern is fine, and the scalability claims are clearly positioned as outlook. The paper is honest about the low visibility and the proof-of-principle status.\n\nBottom line: this is a credible experimental demonstration of a capability the polariton/photonic-computing community has wanted, and the weak points are addressable. I would send it to a serious referee; with a revision that strengthens the artifact exclusion and adds statistics, it would be a solid specialty-journal paper.","headline":"A credible first demonstration of mirror-mediated long-range phase locking in polariton condensates, with a good control, but the fringe evidence needs a more quantitative artifact check before I'd call it airtight.","tokens_in":16494,"tokens_out":6831,"would_cite":true,"duration_ms":73290,"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":"Two spatially separated exciton-polariton condensates with no planar coupling can be phase-locked by mirror feedback of their vertical emission.","keywords":["exciton-polariton condensates","long-range coupling","phase locking","mirror feedback","time-delayed coupling","Kuramoto model","polaritonic lattice","neuromorphic photonics"],"falsifier":"Measure the fringe visibility as the feedback spot is scanned across the target condensate: if the fringes come from phase locking, the visibility should stay roughly constant wherever the spot hits the condensate, whereas if they come from direct feedback–image interference, the visibility should track the overlap integral of the spot and the condensate image. A second control is to block the feedback and superimpose a copy of the feedback beam on the condensate image; the appearance of fringes in that configuration would show the exclusion step was insufficient.","tokens_in":15469,"feed_emoji":"🪞","tokens_out":5855,"duration_ms":59201,"temperature":0.7,"pith_summary":"This paper reports a new way to couple two exciton-polariton condensates that are spatially separated in a lattice: instead of relying on nearest-neighbor coupling through the plane of the sample, the vertical leakage light from each condensate is imaged by an external lens and reflected back onto the other condensate. The authors show that with this mirror feedback the two condensates become phase-locked, as evidenced by interference fringes that appear only when the feedback path is open and disappear when it is blocked. The result matters because it demonstrates a mechanism to write arbitrary long-range couplings between distant sites in a polariton lattice, a capability that existing short-range coupled lattices lack and that is a prerequisite for scalable polariton-based analog computers and neural networks.","feed_headline":"Mirror feedback phase-locks isolated polariton condensates","feed_subtitle":"External imaging links distant condensates, opening arbitrary couplings for neuromorphic hardware.","key_machinery":"The central object is the mirror-mediated feedback loop: a planar mirror that retro-injects leakage light from each condensate onto the other after a round-trip path of roughly 24 cm (delay τ ≈ 1.6 ns), adding a term J_eff $e^{{iβ}}$ ψ_j(t−τ) to each condensate's driven–dissipative Gross–Pitaevskii equation. Projecting onto the ground trap mode and separating amplitude and phase reduces the system to two delay-coupled Kuramoto–Sakaguchi oscillators whose locking condition is an Adler equation; the key step that makes locking hold is that the delayed field is common to both condensates, so its stochastic component drops out of the phase-difference dynamics.","core_discovery":"The paper's central claim is that two geometrically isolated condensates, verified to have no in-plane coupling by overlapping their images in an interferometer, can nevertheless be locked together in phase by external mirror-mediated feedback. The authors verify this with phase-resolved interferometry: with the feedback blocked, the two condensate images show no interference; with the feedback spots directed from one condensate onto the other, fringes appear across the condensate images, indicating a deterministic phase relationship. To explain how coherence survives a round-trip delay of about 1.6 ns that exceeds the single-condensate coherence time of 200 ps to 1 ns, the paper reduces the delay-coupled Gross–Pitaevskii equations to a Kuramoto–Sakaguchi model and shows that the shared mirror path injects the same delayed field into both traps, converting a large part of the phase noise into common-mode noise that cancels in the relative phase, so the remaining effective diffusion is small enough for the locking condition to hold.","pith_inferences":["A quantitative test of the coupling phase β would be to translate the mirror by ΔL while the condensates are locked, which should shift the interferogram phase by exactly kΔL; observing that shift would confirm that the coupling phase is the controlled parameter.","The common-mode noise argument implies that a single shared mirror will partially suppress relative phase noise for every pair in a many-condensate network, something pairwise-independent active couplings would not provide.","The exponential decay of coupling with e^{-τ/T_c} sets a practical distance limit; extending to much longer mirror paths would likely require an amplifier in the feedback arm, a change that would sacrifice the passive-element advantage.","Because the same mirror serves as a delay line, the scheme naturally supports hierarchical delays (short on-chip coupling, long off-chip coupling) for reservoir computing, though the paper only hints at this."],"forward_implications":["Two condensates with no measurable in-plane coupling can be phase-locked purely through external mirror feedback, verified by interferometry.","The coupling strength and sign can be tuned in principle by changing mirror reflectivity and round-trip optical path length, since the coupling phase β = kL shifts with mirror position.","Because the mechanism uses only a passive mirror, it operates as a high-bandwidth analog element without cameras, modulators, or electronic feedback loops.","Replacing a single mirror with a segmented micro-mirror array would allow arbitrary pairwise couplings J_ij between many sites, limited only by the field of view and numerical aperture of the imaging optics.","The delay-coupled phase dynamics reduce to a Kuramoto–Sakaguchi model, so established results on time-delayed oscillator networks (multiple synchronized states, bistability, oscillatory order parameter) apply to this platform."],"supporting_citations":[{"why":"Establishes Bose–Einstein condensation of exciton polaritons, the platform on which the experiment is built.","marker":"[22]"},{"why":"Shows that polariton simulators realize the classical XY Hamiltonian, providing the motivation for controllable couplings in such lattices.","marker":"[20]"},{"why":"Theoretically proposed controllable pairwise interactions in polariton systems, which the present work claims to realize experimentally for the first time.","marker":"[35]"},{"why":"Provides the time-delayed Kuramoto model results that the paper uses to interpret the delay-induced phase locking phenomena.","marker":"[46]"},{"why":"Supplies the Adler equation, the injection-locking description used to derive the locking condition and narrowed linewidth.","marker":"[48]"},{"why":"Gives the Ornstein–Uhlenbeck process formalism used to compute the effective diffusion of the relative phase.","marker":"[49]"}],"fun_headline_variants":["Mirror feedback phase-locks distant polariton condensates","External mirror couples isolated polariton condensates","Mirror-mediated locking of spatially separated condensates","Mirror feedback synchronizes isolated polariton sites","Long-range polariton phase locking via mirror feedback"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The fringes observed in Figure 5(a) are interpreted as interference of the two condensates' emission, requiring that the feedback spot's contribution to the imaged area is negligible; the paper rules this out only by moving the spot to one side, without giving a quantitative bound on the residual overlap.","fun_headline_variants_meta":{"raw":{"variants":["Mirror feedback phase-locks distant polariton condensates","External mirror couples isolated polariton condensates","Mirror-mediated locking of spatially separated condensates","Mirror feedback synchronizes isolated polariton sites","Long-range polariton phase locking via mirror feedback"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000167,"raw_usage":{"total_tokens":1213,"prompt_tokens":857,"completion_tokens":356,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":473,"completion_tokens_details":{"reasoning_tokens":284}},"tokens_in":473,"tokens_out":356,"duration_ms":3811,"temperature":1.0,"reasoning_tokens":284,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T22:38:23.116620+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the fringe visibility as the feedback spot is scanned across the target condensate: if the fringes come from phase locking, the visibility should stay roughly constant wherever the spot hits the condensate, whereas if they come from direct feedback–image interference, the visibility should track the overlap integral of the spot and the condensate image. A second control is to block the feedback and superimpose a copy of the feedback beam on the condensate image; the appearance of fringes in that configuration would show the exclusion step was insufficient.","supporting_citations":[{"cited_title":"Töpfer, P.Cilibrizzi, W.Langbein,andP.G.Lagoudakis, Realizing the classical XY Hamiltonian in polariton sim- ulators, Nat","cited_arxiv_id":null,"evidence_quote":"Shows that polariton simulators realize the classical XY Hamiltonian, providing the motivation for controllable couplings in such lattices."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Theoretically proposed controllable pairwise interactions in polariton systems, which the present work claims to realize experimentally for the first time."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the time-delayed Kuramoto model results that the paper uses to interpret the delay-induced phase locking phenomena."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the Ornstein–Uhlenbeck process formalism used to compute the effective diffusion of the relative phase."}],"review_version":1}