{"id":"8c59f7f9-4f42-465b-a6d2-eea2db5e900f","arxiv_id":"2607.26957","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":4.5,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"MBE-grown as-grown InAsP/InP quantum dots under detuned quasi-resonant excitation show telecom single-photon emission with raw multiphoton probability 0.076 and HOM visibilities of 0.094–0.106.","lead":"As-grown InAsP/InP quantum dots under 32 meV detuned excitation emit telecom single photons with raw g^(2)(0)=0.076 and Hong–Ou–Mandel visibilities of only ~0.09–0.11. The result is a materials baseline for MBE dots without cavities, not a competitive source yet.","discovery_kind":"extension","skeptic_critique":{"model":"grok-4.5","headline":"Full-period HOM visibilities rest on a reconstructed distinguishable baseline that is under-constrained by peak overlap and non-monoexponential decay.","rationale":"The reader correctly isolated the load-bearing soft spot: the full-period V numbers that appear in the abstract and strongest claim depend on a reconstructed distinguishable baseline whose fidelity is questionable given documented peak overlap, non-monoexponential decay, and run-to-run parameter scatter. That does not collapse the paper—raw g_raw^(2)(0)=0.076 is on solid HBT ground, the platform/first-MBE-TPI framing is modest and appropriate, and even a somewhat lower V would still be ‘non-zero’—so the CONDITIONAL verdict and medium correctness_risk stay. No stronger internal inconsistency (e.g., in the biexciton assignment or the 32 meV detuning claim) rises to the same level of impact on the central numbers. A single re-analysis of the existing coincidence histograms with alternative baselines would settle whether the quoted V values and the pulse-separation trend survive.","tokens_in":10843,"tokens_out":660,"duration_ms":24777,"concrete_test":"Re-fit both raw HOM histograms with a multi-exponential decay fixed to the SI TRPL curves, reconstruct the distinguishable central peak by (i) pure side-peak scaling and (ii) a Monte-Carlo blinking+overlap simulation matched to the measured side-peak areas; if either method moves V outside 0.094±0.02 / 0.106±0.02 or erases the 5.3 vs 13.1 ns ordering, the abstract visibilities and diffusion interpretation are not robust.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The abstract and strongest claim quote V=0.094(4) and V=0.106(5) as the measured two-photon interference visibilities. Those numbers are not the zero-delay dip fits (V_fit=0.556 and 0.74); they are area ratios of the entire central peak against a synthetic ‘distinguishable’ peak built from neighboring peaks plus the blinking envelope (Figs. 5c, 6c; §2.3). At 5.3 ns the peaks strongly overlap (τ_dec≈1.5 ns), the SI shows the PL decay is non-monoexponential and fit-range dependent (1.1–1.5 ns), and Table 1 reports inconsistent τ_dec and τ_blink across HBT/HOM runs. Under those conditions the reconstructed baseline area is not uniquely determined: residual overlap, blinking sampling, and monoexponential model mismatch can shift the denominator enough to move V by an amount comparable to the reported 0.09–0.11 values and to the small 13.1→5.3 ns increase used to argue for slow spectral diffusion. If the baseline is biased high, the claim of demonstrated non-zero full-period indistinguishability (and the pulse-separation trend) weakens even while raw g^(2)(0) remains solid.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The manuscript reports quantum-optical characterization of as-grown MBE InAsP/InP quantum dots emitting in the telecom C-band under quasi-resonant excitation detuned by 32 meV. PLE identifies suitable higher-energy resonances; power- and polarization-resolved PL support a biexciton assignment (slope ~1.9, FSS 96 µeV). Under pulsed excitation near saturation, HBT yields a raw multiphoton probability g_raw^(2)(0)=0.076(6) (fit value ~0.009). HOM interferometry gives zero-delay fit visibilities V_fit≈0.56 and 0.74, and full-period area-ratio visibilities V=0.094(4) and 0.106(5) for 13.1 ns and 5.3 ns pulse separations. The authors attribute the modest pulse-separation trend to slow spectral diffusion and argue that Purcell enhancement in cavities can improve indistinguishability, while the large detuning aids laser rejection for fiber-compatible sources.","tokens_in":11122,"tokens_out":1696,"duration_ms":41612,"significance":"If the reported single-photon purity and non-zero two-photon interference hold, the work supplies the first HOM data on MBE-grown InAs(P)/InP QDs in the as-grown (non-cavity) form under experimentally convenient detuned excitation. That combination is relevant for telecom fiber platforms where strict resonance and polarization filtering are costly. The raw g^(2) result is solid and the platform potential is clearly framed. The absolute indistinguishability remains low compared with recent cavity-enhanced C-band results (tens of percent to >90%), so the advance is primarily a materials/excitation benchmark rather than a performance record. Strengths include transparent reporting of both raw area ratios and fit values, explicit blinking envelopes, and Supporting Information on non-monoexponential decay.","major_comments":[{"comment":"§2.3 and Figs. 5c/6c: The abstract and central claim quote full-period visibilities V=0.094(4) and V=0.106(5) obtained as area ratios of the measured central peak to a synthetic “distinguishable” peak reconstructed from neighboring peaks plus the blinking envelope. At 5.3 ns, τ_dec≈1.5 ns produces strong peak overlap; the SI shows the PL decay is non-monoexponential and fit-range dependent (1.1–1.5 ns); Table 1 further shows inconsistent τ_dec and τ_blink across HBT/HOM runs. Under those conditions the reconstructed denominator is not uniquely determined, and plausible model mismatch can shift V by an amount comparable to the reported values and to the small 13.1→5.3 ns increase used to argue for slow spectral diffusion. Please (i) quantify reconstruction systematics (e.g., alternative envelopes, Monte-Carlo peak decomposition, or bounds under mono- vs multi-exponential models), (ii) pro","section":"§2.3, Figs. 5c and 6c, Table 1, Supporting Information"},{"comment":"Table 1 and §2.3: Extracted τ_dec rises from 1.165(7) ns (HBT) to 1.36–1.57 ns (HOM), and τ_blink varies from 46 to 81 ns across nominally identical excitation conditions. The text attributes this to fitting range, residual overlap, and acquisition-time sampling, which is plausible given the SI, but the monoexponential-plus-blinking model is then used to extract both V_fit and T2. If the decay model is misspecified, the zero-delay dip parameters (especially T2=108(25) ps and 214(68) ps) and the blinking-corrected baselines are biased. A single consistent decay model (or an explicit multi-exponential form tied to the TRPL data) should be applied uniformly to HBT and both HOM datasets, with the impact on V_fit, T2, and full-period V reported.","section":"Table 1, §2.3, Supporting Information Figs. S1–S2"},{"comment":"§2.1 and conclusions: The investigated line is assigned to a biexciton on the basis of a power-law slope of 1.9 and FSS of 96 µeV. HOM is performed on this XX transition near saturation. Biexciton photons are in principle usable, but cascade timing jitter, possible residual XX–X cross-talk under the spectral filtering used, and the large FSS relative to the transform limit affect the expected indistinguishability ceiling. The manuscript should state the spectral filter bandwidth relative to FSS and neighboring lines, and briefly discuss whether the reported V is limited by XX-specific physics versus pure dephasing/spectral diffusion, so that the cavity-Purcell outlook is quantitatively grounded.","section":"§2.1, §2.3, §3"}],"minor_comments":[{"comment":"Abstract and §3 quote g_raw^(2)(0)=0.076(6) and later 0.0758(6); keep a single consistent value and uncertainty rounding.","section":"Abstract, §3"},{"comment":"Section numbering in §2.3 repeats “2.1.1” for both pulse-separation subsections; renumber to 2.3.1 / 2.3.2.","section":"§2.3"},{"comment":"Figure 5 caption gives V_fit=0.565(65) while the body text gives 0.556(65); align caption and text.","section":"Fig. 5"},{"comment":"Data Availability Statement currently reads “The authors have no conflicts to disclose,” which belongs under Conflict of Interest; add an actual data-availability sentence (e.g., histograms available on request or repository).","section":"End matter"},{"comment":"Introduction claims “to the best of our knowledge, this work presents the first measurement of the two-photon interference from MBE-grown InAs/InP quantum dots.” Given rapidly appearing 2024–2026 C-band HOM results, a short explicit comparison table (material, growth, cavity, excitation scheme, g^(2), V) would help the reader place the result.","section":"§1"},{"comment":"Clarify whether the HOM setup was polarization-maintained and how the parallel/orthogonal configurations were verified; only the co-polarized (or equivalent) histogram is shown.","section":"§4.3"}],"recommendation":"major_revision","confidential_remarks":"The experimental core (raw g^(2), clear HBT antibunching, first as-grown MBE InAsP/InP HOM attempt under large detuning) is publishable and fits a solid-state quantum-optics / telecom-QD audience. The load-bearing weakness is the full-period V extraction under strong overlap and a demonstrably imperfect monoexponential model; without tighter systematics those abstract numbers and the spectral-diffusion narrative are soft. I would not reject on novelty or performance alone—the “as-grown + detuned” framing is legitimate—but I would require the visibility analysis to be hardened before acceptance. Scope is appropriate for a specialized quantum-photonics or applied-physics journal; for a very high-impact venue the absolute V is too low without cavity data."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"This is a clean materials baseline, not a device paper. They report the first two-photon interference on as-grown MBE InAsP/InP dots (no cavity, no deterministic placement) under a large, filter-friendly 32 meV quasi-resonant detuning. Raw g^(2)(0)=0.076(6) is respectable for a bare mesa at saturation, and they get non-zero full-period HOM visibilities of 0.094 and 0.106 at 13.1 ns and 5.3 ns. That is new relative to the MOVPE InAs/InP and cavity InAs/InAlGaAs numbers they cite, and the framing is honest: potential plus a Purcell roadmap.\n\nWhat they do well is the basic characterization chain. PLE picks the detuning, power law ~1.9 plus 96 µeV FSS supports the biexciton assignment, quasi-resonant drive narrows the line and cleans the background, and both raw area ratios and fit values are shown with uncertainties. Methods are adequate for this subfield. Circularity is low; the claims are measurement claims.\n\nThe soft spot is exactly the one the stress-test flags, and it is real but not fatal. The headline V numbers are not the zero-delay dips (those are 0.56 and 0.74); they are full-central-peak area ratios against a reconstructed “distinguishable” baseline built from neighbors plus the blinking envelope. At 5.3 ns the peaks overlap hard (τ_dec ~1.5 ns), the SI shows non-monoexponential decay that is fit-range dependent, and Table 1 has inconsistent τ_dec and τ_blink across runs. That reconstruction is under-constrained, so the absolute V≈0.09–0.11 and the small pulse-separation trend used for slow spectral diffusion sit on thinner ice than the raw g^(2). Still, the data clearly show some residual interference and good single-photon purity; the central experimental fact survives.\n\nThis is for people already working telecom QD sources who need an MBE InAsP/InP reference point under practical detuned drive. It will not move the brightness–purity–indistinguishability frontier. I would send it to referees; they should demand tighter discussion (or sharing) of the baseline reconstruction and the decay-model mismatch, not a rewrite of the science. Worth a look if that materials niche is yours; otherwise file it.","headline":"First HOM numbers on bare MBE InAsP/InP dots under easy 32 meV detuning; purity is solid, full-period V≈10% is real but soft and not competitive.","tokens_in":11877,"tokens_out":624,"would_cite":false,"duration_ms":18386,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.5","headline":"As-grown InAsP/InP quantum dots under 32 meV detuned excitation produce telecom single photons with measurable two-photon interference.","keywords":["III-V semiconductor","InAsP/InP quantum dots","two-photon interference","indistinguishable single photons","telecom spectral range","detuned quasi-resonant excitation","Hong-Ou-Mandel"],"falsifier":"Repeat the Hong–Ou–Mandel measurement on the same dots with a cavity that shortens the radiative lifetime well below the spectral-diffusion timescale and check whether the full-period visibility rises substantially above 0.1.","tokens_in":11681,"feed_emoji":"📡","tokens_out":871,"duration_ms":19802,"temperature":0.7,"pith_summary":"The paper shows that ordinary, cavity-free InAsP/InP quantum dots grown by molecular beam epitaxy can emit single photons in the third telecom window when driven by a laser detuned 32 meV above the emission line. That large detuning lets the laser light be filtered away easily while still giving the benefits of near-resonant pumping. Under these conditions the dots give a raw multiphoton probability of about 0.076 and Hong–Ou–Mandel visibilities of roughly 0.09–0.11, which rise slightly when the time between successive photons is shortened. The result matters because it proves the material itself already supports indistinguishable telecom photons without demanding cavity integration or strictly resonant excitation, opening a practical route for fiber-compatible quantum light sources.","feed_headline":"Cavity-free telecom dots show two-photon interference","feed_subtitle":"32 meV detuned excitation yields raw g(2)(0)=0.076 and V≈0.1 from as-grown InAsP/InP quantum dots","key_machinery":"Detuned quasi-resonant excitation into a higher QD state (ΔE = 32 meV) combined with full-period Hong–Ou–Mandel visibility extracted by reconstructing the distinguishable central-peak baseline from neighboring peaks and the blinking envelope.","core_discovery":"As-grown MBE InAsP/InP quantum dots, excited quasi-resonantly with a 32 meV detuning and without Purcell enhancement, emit single photons in the telecom C-band with raw g^(2)(0) = 0.076(6) and as-measured two-photon interference visibilities V = 0.094(4) and 0.106(5) for 13.1 ns and 5.3 ns pulse separations.","pith_inferences":["If the same dots are placed in a modest-Q cavity giving a Purcell factor of only a few, the visibility could enter the range already reported for other C-band platforms that required deterministic fabrication.","The blinking times of tens of nanoseconds suggest charge-noise spectroscopy on these as-grown mesas could identify the dominant trap species before cavity integration.","Because the detuning is large enough for simple spectral filtering, the scheme is immediately compatible with existing plug-and-play fiber packages that cannot maintain polarization."],"forward_implications":["As-grown InAsP/InP dots already meet the single-photon purity needed for basic fiber-based quantum optics without cavity fabrication.","Large excitation–emission detuning can replace strict resonance for laser rejection in all-fiber setups where polarization filtering is unreliable.","Shortening the radiative lifetime via Purcell enhancement is predicted to raise indistinguishability by outrunning spectral diffusion.","The modest rise in visibility at shorter pulse separation indicates a nanosecond-scale decoherence process that cavity engineering can target."],"fun_headline_variants":["As-grown InAsP/InP dots show two-photon interference at telecom","Detuned excitation yields V≈0.1 from cavity-free InAsP/InP QDs","Raw g(2)(0)=0.076 and HOM visibility from as-grown telecom dots","32 meV detuned quasi-resonant drive enables telecom two-photon interference","MBE InAsP/InP quantum dots emit indistinguishable C-band photons"],"cache_read_input_tokens":128,"weakest_assumption_plain":"The reconstructed area of the central peak expected for fully distinguishable photons is assumed to be an unbiased baseline even when neighboring peaks overlap strongly and the decay is non-monoexponential.","fun_headline_variants_meta":{"raw":{"variants":["As-grown InAsP/InP dots show two-photon interference at telecom","Detuned excitation yields V≈0.1 from cavity-free InAsP/InP QDs","Raw g(2)(0)=0.076 and HOM visibility from as-grown telecom dots","32 meV detuned quasi-resonant drive enables telecom two-photon interference","MBE InAsP/InP quantum dots emit indistinguishable C-band photons"]},"model":"grok-4.5","effort":"low","cost_usd":0.002874,"raw_usage":{"total_tokens":1046,"prompt_tokens":796,"num_sources_used":0,"completion_tokens":104,"cost_in_usd_ticks":28744000,"prompt_tokens_details":{"text_tokens":796,"audio_tokens":0,"image_tokens":0,"cached_tokens":128},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":146,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":796,"tokens_out":104,"duration_ms":5652,"temperature":1.0,"reasoning_tokens":146,"cache_read_input_tokens":128,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-30T15:49:15.640108+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Repeat the Hong–Ou–Mandel measurement on the same dots with a cavity that shortens the radiative lifetime well below the spectral-diffusion timescale and check whether the full-period visibility rises substantially above 0.1.","supporting_citations":[],"review_version":1}