{"id":"e39eecab-d009-4d47-b213-2b31f0dea284","arxiv_id":"2507.20235","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Delayed optical feedback from a polariton condensate's own emission resonantly seeds the next pulse and enhances output intensity by up to 110%.","lead":"A polariton condensate, a coherent half-light half-matter state, becomes up to 110% brighter when part of its own emission is fed back into the cavity with a time delay. This shows that optical feedback, long used in lasers, also works in the strong-coupling regime and could enable recurrent photonic neural networks.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central mechanism is not discriminated: the uncalibrated coherent-seed term in Eq. (2) is not tested against an incoherent-pump alternative, and the model's simulated delay asymmetry is opposite to the measured one.","rationale":"I have read the paper in good faith. The experimental observation of enhanced integrated emission (up to ~110% near threshold, peaking at τ≈−18.6 ps) is plausible and non-trivial. The comparison with and without feedback is a legitimate differential measurement, and the Appendix does show laser stability. However, the central interpretive claim—that the enhancement arises from coherent resonant seeding of the polariton state, analogous to pump-probe parametric amplification—depends entirely on the feedback term in Eq. (2) being a coherent field injection. Two facts make this assumption load-bearing and unsecured: (1) no measurement of the fraction of re-injected light that actually couples into the cavity mode, nor of its spectral overlap with the blueshifted mode; (2) the authors' own rate-equation model, which is the only quantitative support for the mechanism, produces the opposite temporal asymmetry (slow build-up, fast decay) from the measured enhancement (fast build-up, slow decay), and the authors explicitly defer 'a more precise rate-equation model' to future work. Without knowing whether the feedback acts coherently or merely as an additional (resonant or quasi-resonant) excitation, the experiment does not discriminate between stimulated polariton amplification and a two-pulse incoherent pumping scheme. The proposed spectral-filter control would settle this directly. My assessment therefore keeps the reader's CONDITIONAL verdict: the observation is likely sound, but the mechanism is not yet established.","tokens_in":12337,"tokens_out":6633,"duration_ms":83475,"concrete_test":"Insert a narrowband spectral filter in the feedback arm that can block or pass the lower-polariton condensate line while keeping the total injected power constant (using a variable attenuator). Measure the enhancement η at fixed P≈Pth and τ≈−18.6ps under three conditions: unfiltered PL feedback, feedback with the condensate line blocked, and feedback with an equal-power but spectrally flat (incoherent) source. If η collapses when the condensate line is removed, the coherent-seed mechanism is supported; if η persists, the effect is incoherent pumping. Additionally, measure η versus feedback attenuation over ~20 dB; coherent seeding should show a threshold-like nonlinear dependence while incoherent pumping should be near-linear.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim requires the re-injected emission to act as a coherent seed that resonantly stimulates scattering, as implemented in Eq. (2) by the term αψ(t−tFB) with α=0.05. This is the load-bearing assumption of the interpretation. The paper never calibrates the actually coupled feedback fraction nor the spectral overlap between the returning PL and the (blueshift-shifted) condensate mode, and the Appendix only characterizes the laser, not the condensate feedback. Without this calibration, the observed delay-resonant enhancement could equally arise from an incoherent resonant excitation: the feedback light, spectrally overlapping the quantum-well exciton, adds carriers to the reservoir and lowers the threshold, and the τ≈−18.6 ps peak would then simply reflect temporal coincidence between the delayed PL pulse and the next pump pulse. The supporting model also fails to reproduce the temporal asymmetry of the enhancement: the authors state that 'opposite to our measurements, the simulation depicts a slow build-up and fast decay' and conclude 'a more precise rate-equation model is needed.' Thus the model cannot currently validate the coherent-seed mechanism. The experimental observation of up to ~110% integrated-intensity enhancement is likely robust, but its interpretation as coherent polariton amplification is not yet evidenced.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports an experimental and theoretical study of a nonresonantly pumped polariton microcavity coupled to a delayed optical feedback loop. The main observation is that re-injecting part of the condensate emission into the cavity increases the time-integrated output intensity by up to approximately 110% near the condensation threshold, with a sharp resonant enhancement as a function of the feedback delay, peaking near tau = -18.6 ps. The authors propose a classical rate-equation model for the condensate and two excitonic reservoirs, in which the feedback enters as a coherent amplitude term alpha*psi(t-tFB) in Eq. (2). They argue that the model reproduces the pump-power dependence and the resonant delay dependence, and they discuss the role of the blueshift in reducing enhancement. The paper concludes that the method can be extended to coupled polariton condensates for neuromorphic computing.","tokens_in":12693,"tokens_out":3621,"duration_ms":47469,"significance":"If the coherent-seeding mechanism were firmly established, this would be a useful new experimental capability for polariton systems, connecting optical-feedback control to the strong-coupling regime and potentially enabling recurrent signaling between condensates. The manuscript includes a plausible rate-equation framework and the experimental data show a clear, reproducible-looking enhancement effect. The main strength is the experimental observation itself, which is internally consistent across power scans and delay scans. However, the current evidence does not uniquely establish the proposed mechanism over an incoherent reservoir-pumping alternative, and the model fails to reproduce a key temporal asymmetry of the measured enhancement. The significance of the work is therefore conditional on additional calibration and model discrimination, but the experimental result is likely of interest to the polariton and nanophotonics communities.","major_comments":[{"comment":"The feedback term alpha*psi(t-tFB) with alpha = 0.05 is the load-bearing assumption of the paper, but the feedback fraction is never independently calibrated. Appendix B characterizes the laser light reflected from the sample, not the condensate photoluminescence that is actually fed back during the experiment. Without a measurement of the coupled feedback fraction and the spectral overlap between the returned light and the blueshifted condensate mode, the model's agreement is at least partly a fitted outcome, and the observed enhancement could equally arise from an incoherent resonant excitation that adds carriers to the reservoir. Please provide an independent calibration of the feedback power at the sample and the spectrum of the returning light, and test an incoherent-feedback alternative against the data.","section":"Theory, Eq. (2)"},{"comment":"The simulation produces a slow build-up and fast decay of the enhancement as a function of delay, which the authors explicitly state is opposite to the measured behavior (fast build-up, slow decay). The simulated peak is also at tau approximately -6 ps, compared with the experimental value of -18.6 ps. This discrepancy concerns a central experimental feature, and it directly contradicts the Conclusions statement that the model \"explains nearly all features of the study.\" The model needs to be modified to reproduce the correct temporal asymmetry, or the claims about the model's explanatory power must be substantially scaled back.","section":"Results, Fig. 5(d)"},{"comment":"No error bars or statistical uncertainties are shown for the measured enhancement factors eta. The enhancement definition in Eq. (1) involves integrated intensities, but the number of experimental realizations and the integration procedure are not specified for the polariton data. Without uncertainty estimates, the 110% maximum and the sharp resonant peak in the delay scan cannot be quantitatively assessed. Please add error bars from repeated measurements or clearly state the reproducibility of the data.","section":"Results, Figs. 2-4"},{"comment":"The interpretation of the enhancement peak as coherent seeding relies on the assumption that the feedback light is a phase-coherent copy of the condensate field. However, the experimental feedback path contains a long-pass filter and a reflection from a translation-stage mirror, and no measurement of the coherence or polarization of the returned light relative to the condensate is reported. The manuscript should either provide such a characterization or explicitly acknowledge that the observed enhancement could be caused by an incoherent threshold-lowering effect, and explain how the data discriminate between these possibilities.","section":"Results, Fig. 3 and Eq. (1)"}],"minor_comments":[{"comment":"There is a typo in the Introduction: \"excitaiton\" should be \"excitation\" in the sentence about nonresonant excitation schemes.","section":"Introduction"},{"comment":"The text states \"A negative delay tau < corresponds to a feedback seed arriving before the next pump pulse,\" but the inequality is incomplete; it should read \"tau < 0\".","section":"Results, Experimental setup"},{"comment":"The experiment uses a repetition period of t1 = 13 ns (76 MHz), while the Theory section states \"the full t1 = 12 ns gap between pump pulses.\" This numerical inconsistency should be corrected.","section":"Theory, Eq. (4)"},{"comment":"Several figure captions in the arXiv version contain uninterpretable Unicode glyph sequences (e.g., \"uni00000014/uni00000013...\"), which seriously impair readability and should be fixed in the final manuscript.","section":"Figure captions"},{"comment":"The term \"quasiresonantly seeds\" is used in the Introduction, but no spectral measurement of the feedback light relative to the condensate energy is presented. Either add such a measurement or rephrase to avoid implying a resonance that is not directly evidenced.","section":"Introduction"}],"recommendation":"major_revision","confidential_remarks":"The experimental observation of enhanced polariton emission under delayed feedback appears credible and will likely be of interest to the community. However, the manuscript currently overstates the support for the coherent-seeding mechanism: the key feedback parameter is uncalibrated, the model fails on the temporal asymmetry of the enhancement, and the data lack error bars. These issues are fixable within the scope of the paper, in my view, so I recommend major revision rather than rejection. I would also encourage the editor to ask the authors to address the incoherent-pumping alternative explicitly, as this is the main conceptual risk to the interpretation."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Rafał, quick take on arXiv:2507.20235. The genuinely new thing: a delayed feedback loop that feeds a polariton condensate's own emission back into the cavity, in the strong-coupling regime, and a measured ~110% increase in integrated intensity just above threshold. That combination really isn't in the prior laser-feedback literature. The experimental part looks honestly done: clear setup, zero-delay calibrated via interference, and the delay scan showing a resonant peak around -18.6 ps is a believable observation.\n\nWhat it does well: the rate-equation model with two reservoirs plus a feedback term reproduces the power dependence of the threshold and the qualitative shape of the enhancement-vs-delay peak. The authors also test the effect of blueshift, and the idea that mismatch kills enhancement makes physical sense.\n\nThe soft spots are real and not tiny. The coherent-seed term αψ(t−tFB) is the load-bearing assumption, but α=0.05 is hand-set and never calibrated against the actual reflected fraction or the spectral overlap of the returning PL with the blueshifted condensate. Nothing in the paper rules out the simpler story: the feedback light is largely incoherent and just adds carriers to the reservoir, lowering threshold; the delay peak then just means temporal coincidence with the next pump. The model's mismatch in temporal asymmetry is the clearest sign: Fig. 5(d) gives slow build-up and fast decay, while the data show the opposite, and the authors admit that a more precise rate-equation model is needed. That makes the conclusion 'explains nearly all features' overstated. Also, no error bars anywhere on enhancement; the claim might be robust but we can't see the scatter.\n\nProportionately: none of this kills the experimental result. The observation of feedback-enhanced emission is likely correct. What is not yet evidenced is the coherent-stimulation mechanism. A good referee should ask for the diagnostic control: measure the spectrum/resonance of the feedback seed, vary the seed polarization or energy, and calibrate the coupled fraction. Releasing data and code would help too.\n\nFor whom: polariton experimentalists and people thinking about recurrent photonic circuits. I'd cite it as a demonstration of optical feedback in polariton systems, but not as proof of coherent seeding. Yes, send to a serious referee; conditional acceptance seems right.","headline":"Delayed optical feedback clearly enhances a polariton condensate, but the paper doesn't yet distinguish coherent seeding from incoherent re-pumping.","tokens_in":13161,"tokens_out":2086,"would_cite":true,"duration_ms":25649,"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":"Delayed optical feedback raises polariton condensate emission by up to 110% near threshold.","keywords":["exciton-polariton condensate","optical feedback","delayed feedback loop","nonresonant pumping","stimulated scattering","rate-equation model","microcavity","neuromorphic photonics"],"falsifier":"A straightforward test is to inject spectrally filtered feedback: if the ≈110% enhancement persists when the seed is detuned far outside the condensate emission line, the coherent-seed mechanism is wrong.","tokens_in":12188,"feed_emoji":"🔁","tokens_out":5462,"duration_ms":58274,"temperature":0.7,"pith_summary":"The paper reports that re-injecting a polariton condensate's own emission after a fixed delay—an optical feedback loop—raises the condensate's integrated output intensity by up to about 110% when the nonresonant pump is near threshold. The enhancement appears as a sharp resonance versus the feedback delay, and its size depends on pump power. The authors explain the effect with a classical rate-equation model in which the delayed feedback field coherently seeds the condensate and triggers stimulated scattering from an excitonic reservoir, analogous to resonant pump-probe amplification. If the interpretation holds, it shows that strong-coupling condensates respond sensitively to external optical feedback, opening a path to recurrent optical networks with weighted delayed signals.","feed_headline":"Delayed feedback boosts polariton condensates by 110%","feed_subtitle":"Re-injecting a condensate's own emission seeds the next pulse and lowers the condensation threshold.","key_machinery":"The central object is the delayed feedback seed αψ(t − t_FB) added to the condensate amplitude equation: the cavity's own coherent emission, attenuated by strength α and delayed by t_FB, is re-injected into the condensate mode. It sits inside a classical rate-equation model that couples the condensate population to two excitonic reservoirs—an inactive high-momentum reservoir n_I that converts into an active bottleneck reservoir n_R, which scatters into the condensate at rate R. The model also includes a blueshift potential U(t) = g_C |ψ|² + 2g_R(n_R + n_I) that shifts the condensate energy; the paper shows that larger blueshifts reduce the enhancement, indicating that the feedback must remain resonant with the blueshifted state to work.","core_discovery":"The central claim is that delayed optical feedback of the condensate's own emission acts as a resonant seed for a nonresonantly pumped polariton microcavity, lowering the condensation threshold and enhancing the time-integrated cavity emission by up to ≈110% near threshold. The paper obtains this by connecting a fraction of the cavity emission through a feedback arm whose round-trip time matches the laser repetition period, then measuring the enhancement factor η = (⟨I_F⟩ − ⟨I⟩)/⟨I⟩ as a function of pump power and delay τ. The enhancement peaks at τ ≈ −18.6 ps, a negative delay interpreted as the time the condensate takes to build up after the pump pulse, and diminishes at higher pump powers where the nonresonant pump alone can trigger condensation. A rate-equation model for the condensate amplitude ψ(t) coupled to active and inactive excitonic reservoirs, with a feedback term αψ(t − t_FB), reproduces the main features of the enhancement versus power and delay, although it inverts the build-up/decay asymmetry seen in experiment, which the authors attribute to over-simplified reservoir depletion.","pith_inferences":["If the re-injected light is truly a coherent seed, the enhancement should be sensitive to the relative phase and frequency of the feedback field; varying the feedback arm length by half a wavelength (or spectrally filtering the seed) would test this prediction directly.","The model's inverted asymmetry (fast simulated decay vs slow experimental decay) suggests a slower, more gradual reservoir depletion than the single gain-clamping term R|ψ|² allows; time-resolved reservoir photoluminescence could discriminate between the model and the real dynamics.","A natural extension is to measure the enhancement versus feedback strength α and compare it quantitatively with the model after independently calibrating the feedback fraction; this would place the mechanism on firmer footing beyond the single α = 0.05 value used here.","Because the enhancement depends on the condensate's interaction-induced blueshift, the feedback method could be used as a sensitive detector of the local exciton density, mapping the reservoir dynamics in space and time."],"forward_implications":["If the seed mechanism is correct, the effective condensation threshold is lowered, so weaker nonresonant pumps can produce coherent polariton emission.","The sharp delay resonance implies that matching feedback delay to the condensate build-up time gives a tunable, picosecond-scale amplification response, useful for timing-sensitive photonic circuits.","The reduction of enhancement at high pump power and at large blueshift shows that the feedback gain is self-limited by interactions, acting as a classical optical blockade.","The same feedback loop can be extended to multiple spatially separated condensates, where weighted delayed signals from previous pulses drive each condensate's response, the basis for recurrent neuromorphic optical networks.","The measured build-up and decay times of the enhancement give direct access to the relative lifetimes of the exciton reservoir and the condensate, offering a simple probe of carrier relaxation dynamics."],"supporting_citations":[{"why":"Supplies the standard delayed-feedback method and the feedback-term form αψ(t − t_FB) for semiconductor lasers, which the polariton system adopts.","marker":"[1]"},{"why":"Provides the resonant pump-probe polariton amplifier mechanism that the feedback loop emulates to trigger stimulated scattering.","marker":"[9]"},{"why":"Establishes the nonresonant condensation physics and stimulated scattering into the condensate that the rate-equation model relies on.","marker":"[24]"},{"why":"Supplies the interaction constants, Hopfield coefficient, and quantum-well parameters used in the blueshift term of the model.","marker":"[30]"},{"why":"Provides the decay rates of polaritons and excitonic reservoirs and the prior time-delayed polariton response framework on which the parameters are based.","marker":"[31]"}],"fun_headline_variants":["Polariton condensate feedback boosts emission by 110%","Feedback loop seeds polariton condensates, boosting output","Delayed optical feedback enhances polariton condensates 110%","Self-seeding polariton condensates via delayed feedback","Optical feedback boosts polariton condensate emission 110%"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing assumption is that the re-injected light enters as a coherent seed that resonantly stimulates scattering into the condensate, rather than acting as additional incoherent pumping or heating; the paper does not independently calibrate the feedback fraction or verify the seed's spectral overlap with the blueshifted condensate mode.","fun_headline_variants_meta":{"raw":{"variants":["Polariton condensate feedback boosts emission by 110%","Feedback loop seeds polariton condensates, boosting output","Delayed optical feedback enhances polariton condensates 110%","Self-seeding polariton condensates via delayed feedback","Optical feedback boosts polariton condensate emission 110%"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00086,"raw_usage":{"total_tokens":3725,"prompt_tokens":932,"completion_tokens":2793,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":548,"completion_tokens_details":{"reasoning_tokens":2711}},"tokens_in":548,"tokens_out":2793,"duration_ms":22077,"temperature":1.0,"reasoning_tokens":2711,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T13:41:40.209864+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A straightforward test is to inject spectrally filtered feedback: if the ≈110% enhancement persists when the seed is detuned far outside the condensate emission line, the coherent-seed mechanism is wrong.","supporting_citations":[{"cited_title":"Lang and K","cited_arxiv_id":null,"evidence_quote":"Supplies the standard delayed-feedback method and the feedback-term form αψ(t − t_FB) for semiconductor lasers, which the polariton system adopts."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the resonant pump-probe polariton amplifier mechanism that the feedback loop emulates to trigger stimulated scattering."},{"cited_title":"Byrnes, N","cited_arxiv_id":null,"evidence_quote":"Establishes the nonresonant condensation physics and stimulated scattering into the condensate that the rate-equation model relies on."},{"cited_title":"Mirek, M","cited_arxiv_id":null,"evidence_quote":"Supplies the interaction constants, Hopfield coefficient, and quantum-well parameters used in the blueshift term of the model."},{"cited_title":"Mirek, A","cited_arxiv_id":null,"evidence_quote":"Provides the decay rates of polaritons and excitonic reservoirs and the prior time-delayed polariton response framework on which the parameters are based."}],"review_version":1}