{"id":"be22fc4e-9c9f-442e-a651-775777a62401","arxiv_id":"2508.01416","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"An InAsP/InP nanowire quantum dot was coupled to an erbium-doped fiber atomic frequency comb memory at 980 nm, storing and recalling single photons without spectral tuning.","lead":"This paper reports the first hybrid light-matter interface between a chip-integrated quantum dot and an erbium-doped fiber quantum memory at 980 nm, with storage and recall of single photons without spectral tuning. The result points to a practical way to connect deterministic quantum light sources to fiber-based memories for quantum networks.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Quantum-dot storage claim rests on unverified spectral purity and stability of X1- in the 8 GHz AFC: intrinsic QD linewidth is unmeasured, and the recalled echo's nonclassicality is only estimated.","rationale":"The reader's weakest assumption captures the key gap: the QD storage demonstration does not directly verify the spectral content or stability of the light stored in the AFC. The paper itself admits the intrinsic QD linewidth is unmeasured (Section IV, Fig. 5a) and that g_out(0) is an estimate (Eq. 2). These are explicit limitations flagged in the text. The alignment via 8 GHz hole burning is clever but only proves that some light at the hole frequency is present; it does not prove that light is the X1- single-photon line at the operating power. The low SNR (1.92) makes the recalled echo susceptible to contamination by prompt or background light. A high-resolution spectral measurement of the filtered emission at storage conditions is the single check that would resolve whether the central claim holds. This does not invalidate the solid spectroscopy and weak coherent pulse storage results; it only means the QD storage component is conditional on the spectral attribution. Hence the reader's CONDITIONAL verdict is appropriate and unchanged.","tokens_in":15511,"tokens_out":19414,"duration_ms":240348,"concrete_test":"Measure the high-resolution spectrum of the filtered QD emission under the exact storage conditions (excitation power >1.7 µW, B = 0.06 T) using a scanning Fabry-Perot or heterodyne setup with <100 MHz resolution, taken immediately before and after a 1200 s storage run. Confirm that the X1- line is the only spectral component within the 8 GHz AFC window, that its linewidth is within the comb, and that its center frequency drift over the run is less than the comb period. If any additional line or a drift larger than the AFC tooth spacing is observed, the storage claim is not supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claim—deterministic storage and recall of X1- single photons from a chip-integrated QD in an EDF AFC memory—requires that the light transmitted through the EDF during the 1200 s storage runs is exclusively the X1- line and that this line remains within the 8 GHz AFC for the whole integration. Both conditions are unverified. The alignment procedure (Fig. 5a) burns an 8 GHz spectral hole and observes an APD count increase with a 50 GHz-resolution coarse spectrometer; the paper explicitly states that the resulting 8.05 GHz Lorentzian does not correspond to the intrinsic QD linewidth and that the linewidth was not measured. At the excitation power used (>1.7 µW, above saturation), Fig. 4a shows additional QD emission features near 972 and 977 nm; no filtered spectrum at the storage power is presented to show these are rejected by the 0.5 nm grating. If another spectral line or broadband background falls inside the 0.5 nm passband and the 8 GHz comb, the 'X1-' echo is misattributed. Moreover, the recalled signal has SNR 1.92 and the single-photon character is an estimate from Eq. 2, not a measurement; the estimate assumes all signal is stored single-photon light and all noise is uncorrelated, an assumption untested by any control experiment (e.g., detuning the filter or blocking the QD excitation). The 1200 s integration also assumes no spectral drift of the QD line; no stability monitor is reported.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports a spectroscopic characterization of the 4I15/2 ↔ 4I11/2 transition of Er3+ in an erbium-doped fiber (EDF) at 980 nm and 10 mK, including absorption, spin population lifetime, and optical coherence time. It then implements an 8 GHz atomic frequency comb (AFC) memory in this transition and demonstrates storage and retrieval of 59 weak coherent temporal modes with an efficiency of about 1%, consistent with a forward calculation from the measured optical depth and finesse. The central claim is the deterministic storage and recall of the X1- emission from a chip-integrated InAsP/InP nanowire quantum dot in this AFC memory, without spectral tuning of the quantum dot. The recalled photon statistics are not measured directly but are estimated from the measured signal-to-noise ratio via Eq. (2).","tokens_in":15720,"tokens_out":11564,"duration_ms":124498,"significance":"If the quantum-dot storage claim holds, this would be a valuable step toward a coherent hybrid interface between a deterministic on-chip single-photon source and a solid-state quantum memory at 980 nm. The spectroscopic characterization of the 980 nm EDF transition at millikelvin temperatures is a useful contribution, and the weak-coherent-pulse AFC memory with 59 multimode storage is a solid, internally consistent result. The paper's forward efficiency calculation from independently measured parameters is a positive feature. However, the evidence for the central quantum-dot storage claim is incomplete: the spectral overlap between the X1- line and the 8 GHz AFC is established only with a coarse spectrometer and an 8 GHz spectral hole, the stability of the QD line over the 1200 s integration is assumed rather than measured, and the single-photon character of the recalled light is only estimated, with a numerical inconsistency in the stated formula.","major_comments":[{"comment":"The spectral overlap between the X1- emission and the 8 GHz AFC is not established with sufficient resolution. The coarse wavelength alignment uses a spectrometer with 50 GHz resolution (Methods IV), which is more than six times the AFC bandwidth, and the subsequent transmission measurement through an 8 GHz spectral hole yields a fitted FWHM of 8.05 GHz that the authors explicitly state does not correspond to the intrinsic QD linewidth. At the operating excitation power (>1.7 µW), additional QD emission features appear near 972 and 977 nm (Fig. 4a), and no high-resolution spectrum of the filtered light is shown to demonstrate that only the X1- line falls within the AFC passband. The authors should provide a sub-GHz spectrum of the filtered QD emission, a direct linewidth measurement, or a control experiment (e.g., detuning the AFC or blocking the QD excitation) to rule out contamination of the recalled echo by other spectral components. This is load-bearing for the claim of deterministic storage and recall of the X1- single-photon emission.","section":"Section II, Fig. 5a, and Methods IV"},{"comment":"The assumption that the X1- emission frequency remains fixed for the entire 1200 s integration period is not experimentally verified. The storage sequence is repeated 480 times over 1200 s, and if the QD line drifts out of the 8 GHz AFC window during this time, the recalled echo would be suppressed or diluted. The paper states that 'the frequency of QD emission remains fixed throughout the experiment' but provides no monitoring of the QD wavelength or of the transmission through the spectral hole during the integration. The authors should report a time-resolved measurement of the QD emission position or a stability trace over the full 1200 s, or reduce the integration time to match a measured stability window.","section":"Section II, Fig. 5b"},{"comment":"The estimate of the recalled photon statistics is not directly measured, and the presented numerical result is not reproducible from the stated formula. With S = 1.92 and g_in(0) = 0.207, the formula as printed (g_out = 1 + (S^2 + g_in)/(1+S)^2) gives a value above 1, while the paper reports g_out(0) = 0.5547. Even under common definitions of SNR for signal mixed with Poissonian noise, the result is typically 0.66 rather than 0.55. The authors should clarify the definition of S, correct the typographical rendering of Eq. (2), and either provide a consistent numerical estimate or, preferably, measure the g(2) of the recalled echo directly with a Hanbury-Brown-Twiss detection setup. This is important because the single-photon character of the retrieved light is a central part of the claim that the interface stores non-classical light from the quantum dot.","section":"Eq. (2) and accompanying text"}],"minor_comments":[{"comment":"The identification of the X, XX, and X1- peaks in the main panel of Fig. 4a is not clear; please add explicit labels or arrows in the main spectrum so the reader can associate each feature with the corresponding emission line, especially given the additional peaks at 972 and 977 nm at high excitation power.","section":"Fig. 4a"},{"comment":"Please specify the cutoff wavelengths of the two free-space high-pass filters and the effective bandpass of the 1200 grooves/mm grating (0.5 nm resolution), so the reader can assess which of the QD emission lines observed in Fig. 4a are rejected in the storage experiments.","section":"Methods IV"},{"comment":"The statement that 'all subsequent experiments were conducted with an excitation power of more than 1.7 µW (above saturation power)' should be reconciled with the g(2)(0) = 0.207 value; please clarify whether the g(2) and storage measurements used the same power and discuss the effect of above-saturation excitation on the single-photon purity and on the appearance of the 972/977 nm features.","section":"Section II, paragraph on QD excitation"},{"comment":"The acronym MEMS is used without definition at its first occurrence in the storage protocol description; please spell out 'micro-electro-mechanical system' and define the role of the MEMS switches in the timing sequence.","section":"Methods IV, storage protocol"},{"comment":"The caption does not list the five storage durations shown in the panel; if they correspond to the same values used for the weak coherent pulses (e.g., 5 to 100 ns), please state them explicitly.","section":"Fig. 5b caption"},{"comment":"The claim of 'approximately an order of magnitude improvement in spectral hole burning efficiency' relative to conventional techniques should be quantified or referenced in the Supplementary Information, so the reader can evaluate the comparison.","section":"Section II, AFC preparation"}],"recommendation":"major_revision","confidential_remarks":"The manuscript combines a solid weak-coherent-pulse AFC memory demonstration at a new wavelength with a more ambitious quantum-dot storage claim. The spectroscopic characterization of the 980 nm EDF transition is a useful stand-alone contribution, and the efficiency consistency is a strength. However, the central claim in the abstract and introduction is not yet supported by the experimental evidence: the spectral purity and stability of the X1- line inside the 8 GHz AFC are unverified, and the non-classicality of the recalled light is only estimated with a numerically inconsistent formula. These issues can be addressed with additional measurements rather than reanalysis, so I recommend major revision with a clear request for control experiments and a direct g(2) measurement of the recalled echo. The paper is otherwise well written and within the scope of the journal."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a real experimental step—first AFC memory on the 980 nm Er transition in fiber, and first storage of photons from a chip-integrated InAsP QD in an erbium-doped fiber memory. The weak-coherent-pulse storage (59 modes, 1% efficiency) is internally consistent and the efficiency budget closes. The QD storage is the soft spot: SNR 1.92, g(2) of recalled light estimated not measured, QD linewidth not measured, and no control rules out contamination from other emission lines or spectral drift over the 1200 s integration. The paper is transparent about most of this, but the abstract's \"deterministic storage and recall\" is stronger than the evidence.\n\nWhat's new: the 980 nm transition of EDF has not been used for AFC before, and the direct spectral overlap between InAsP nanowire QDs and this transition without tuning is a nice match. The spectroscopy (T1S biexponential, T2O ~2.6 µs) is useful and careful. The AFC preparation with amplitude+phase modulated pumping claims an order-of-magnitude improvement in burning efficiency; that would need a closer look at the supplement, but the demonstrated 8 GHz comb and 59-mode storage are solid.\n\nSoft spots, in order: (1) The QD-storage claim rests on an unmeasured intrinsic linewidth and an alignment procedure that shows only that some light is transmitted through a burned hole; it does not prove the light is exclusively X1-. At the >1.7 µW excitation power used, Fig. 4a shows extra features near 972 and 977 nm. The 0.5 nm grating may or may not reject them; no filtered spectrum at storage power is shown. (2) The g(2) of recalled photons is an estimate from SNR, not a measurement. The estimate is standard but assumes signal is all single-photon and noise is uncorrelated; a control (e.g., QD excitation off, or filter detuned) would rule out leakage. (3) No spectral stability monitor over the 1200 s integration. These are all fixable in revision, not fatal.\n\nThe citation pattern looks fine; the relevant prior work (Tang et al., Thomas et al., Saglamyurek et al.) is cited. No data/code included, which is normal for this kind of experimental letter but limits reproducibility.\n\nMy take: the weak-pulse and spectroscopy parts are solid and worth publishing. The QD storage is a promising proof-of-principle that needs either more evidence or more guarded claims. If I were the editor, I would send it to peer review; the platform combination is new and the community would benefit from a careful referee report. For a reader, it's a worthwhile paper to know about, but I'd wait for the follow-up with measured g(2) before citing the single-photon interface as established.","headline":"A genuine new platform pairing—EDF at 980 nm meets a chip-integrated InAsP QD—with solid weak-pulse AFC storage and a weaker, but fixable, single-photon storage claim.","tokens_in":16440,"tokens_out":2913,"would_cite":true,"duration_ms":34371,"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":"A chip-integrated quantum dot at 980 nm stores and recalls its single photons in an erbium-doped fiber memory, with no spectral tuning.","keywords":["quantum memory","erbium-doped fiber","atomic frequency comb","quantum dot single photons","nanowire quantum dot","980 nm transition","hybrid light-matter interface","multimode storage"],"falsifier":"Measure the X1− linewidth with a high-resolution spectrometer or scanning Fabry-Perot and monitor its center wavelength continuously during a 1200 s storage run; if the line drifts out of the 8 GHz comb, or if the transmission increase in Fig. 5a disappears when the dot is excited above-band, the single-photon storage claim fails. A cleaner control would alternate storage runs with the AFC hole burned on-resonance and off-resonance and compare recall rates.","tokens_in":15215,"feed_emoji":"📡","tokens_out":7322,"duration_ms":78568,"temperature":0.7,"pith_summary":"This paper reports the first experimental hybrid light–matter interface in which photons from a deterministic, chip-integrated quantum dot are stored and recalled in a solid-state quantum memory. The emitter is an InAsP/InP nanowire quantum dot whose negatively charged exciton line (X1−) sits near 980 nm, and the memory is a 10 m erbium-doped fiber cooled to about 10 mK, operated on the 4I15/2↔4I11/2 transition of Er3+ with an 8 GHz atomic frequency comb. The authors achieve spectral overlap without tuning the quantum dot, characterize the erbium transition's spin population lifetime and optical coherence, and demonstrate storage of 59 weak coherent temporal modes as well as the X1− single photons for five storage times. The result matters because it connects a deterministic on-chip single-photon source directly to a fiber-compatible, broadband quantum memory.","feed_headline":"Chip quantum dot photons stored in fiber memory","feed_subtitle":"A 980 nm nanowire emitter sends single photons into an erbium-doped fiber memory without spectral tuning.","key_machinery":"The central object is an atomic frequency comb (AFC) prepared on the 4I15/2↔4I11/2 transition at 980 nm in a 10 m erbium-doped fiber cooled to about 10 mK. An AFC is a periodic series of narrow absorbing teeth carved into an inhomogeneously broadened ensemble by optical pumping, with storage time t_s=1/Δ for comb spacing Δ and bandwidth Γ_AFC=NΔ. The paper's preparation approach modulates both amplitude and phase of the pump pulses to produce squarish teeth with an order-of-magnitude better hole burning efficiency than conventional methods; the resulting 8 GHz combs have finesse about 2 and background optical depth 0.05. Supporting measurements give a spin population lifetime with two components (6.75 s and 385 s) and an optical coherence time of about 2.6 µs, which together make the transition suitable for photon-echo protocols. Expected memory efficiency is 6.7%, with measured total efficiency 1% because of 83% transmission loss from splicing and connectors.","core_discovery":"The central claim is that Er3+ ions in glass silica fiber are a viable quantum memory for a chip-integrated InAsP quantum dot at 980 nm, and the paper demonstrates this by deterministic storage and recall of X1− photons in an 8 GHz bandwidth atomic frequency comb. No electric, magnetic, or strain tuning of the quantum dot is used; compatibility comes from the roughly 10 nm inhomogeneous broadening of the erbium 980 nm transition, which is wide enough to engulf the dot's emission. The recalled echo preserves the input temporal profile, the estimated g(2)_out(0)=0.5547±0.001 lies below the classical limit of 1, and weak-coherent time-bin storage fidelity is estimated at 87.6%, above the classical bound of 66.7%. The paper states that no qubit is encoded; the temporal profile of the single photon is mapped directly onto the memory.","pith_inferences":["A direct measurement of the quantum dot's intrinsic linewidth and wavelength drift during the 1200 s integration would test whether the recalled echo is single-photon storage rather than background or InP emission.","Because 980 nm is a standard erbium amplifier band, this interface could in principle drop into existing fiber networks as a memory module once efficiency and coherence improve.","The same recipe may transfer to other rare-earth transitions that overlap InAsP quantum dot emission, such as 171Yb3+ doped crystals near 978.54 nm, making the hybrid interface a general strategy rather than a single-wavelength fix.","The amplitude-and-phase-modulated AFC preparation could improve hole burning efficiency in other rare-earth-doped fibers, potentially raising multimode capacity."],"forward_implications":["A deterministic chip-integrated source can be temporally stored and recalled in a fiber-based AFC memory without spectral tuning, shifting the wavelength-matching burden onto the memory's broad inhomogeneous bandwidth.","The 980 nm erbium transition in fiber is coherent enough for photon-echo quantum storage, with multi-second spin lifetimes and microsecond optical coherence, adding a new wavelength channel for memories.","The demonstrated 59-mode storage, with a potential 144 modes within the 8 GHz bandwidth, supports temporal multiplexing of deterministic sources.","Retrieved X1− photons have an estimated g(2)_out(0)=0.5547, below the classical threshold of 1, indicating that the recalled light retains nonclassical photon statistics despite the low SNR."],"supporting_citations":[{"why":"Defines the atomic frequency comb protocol and the storage-time/bandwidth relations (t_s=1/Δ, Γ=NΔ) used throughout.","marker":"[44]"},{"why":"Earlier storage of quantum-dot single photons in a solid-state AFC memory, the result this work extends to a chip-integrated source and fiber at 980 nm.","marker":"[16]"},{"why":"Supplies the InAsP/InP nanowire quantum dots integrated on Si3N4 waveguides that act as the deterministic single-photon source.","marker":"[50]"},{"why":"Shows erbium-doped fiber can serve as an AFC quantum memory at telecom wavelength, establishing the platform this paper moves to 980 nm.","marker":"[36]"},{"why":"Demonstrates broadband multimode AFC storage in a rare-earth crystal, providing the multimode benchmark against which the 59-mode storage is set.","marker":"[33]"},{"why":"Provides the method for mapping multiple time-bin photonic qubits into one atomic ensemble, underlying the random time-bin storage test.","marker":"[55]"},{"why":"Supplies the classical fidelity bound used to certify quantum storage of the weak coherent time-bin modes.","marker":"[56]"},{"why":"Interfaces a quantum dot with an atomic memory and supplies the formula used to estimate g(2)_out(0) for the recalled photons.","marker":"[11]"}],"fun_headline_variants":["Quantum dot photons stored in fiber memory at 980 nm","Chip quantum dot beams into erbium fiber memory","First direct coupling of quantum dot to fiber memory","Hybrid memory stores single photons without spectral tuning","Erbium fiber remembers quantum dot photons at 980 nm"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The recalled echoes count as stored single photons only if the quantum dot's X1− line stayed inside the 8 GHz comb for the whole 1200 s run; the overlap is checked by burning a broad spectral hole and seeing transmission rise, but the dot's intrinsic linewidth is not measured, so drift or background emission could mimic the signal.","fun_headline_variants_meta":{"raw":{"variants":["Quantum dot photons stored in fiber memory at 980 nm","Chip quantum dot beams into erbium fiber memory","First direct coupling of quantum dot to fiber memory","Hybrid memory stores single photons without spectral tuning","Erbium fiber remembers quantum dot photons at 980 nm"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00068,"raw_usage":{"total_tokens":3144,"prompt_tokens":1054,"completion_tokens":2090,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":670,"completion_tokens_details":{"reasoning_tokens":2012}},"tokens_in":670,"tokens_out":2090,"duration_ms":16690,"temperature":1.0,"reasoning_tokens":2012,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T05:37:22.826170+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the X1− linewidth with a high-resolution spectrometer or scanning Fabry-Perot and monitor its center wavelength continuously during a 1200 s storage run; if the line drifts out of the 8 GHz comb, or if the transmission increase in Fig. 5a disappears when the dot is excited above-band, the single-photon storage claim fails. A cleaner control would alternate storage runs with the AFC hole burned on-resonance and off-resonance and compare recall rates.","supporting_citations":[{"cited_title":"Afzelius, C","cited_arxiv_id":null,"evidence_quote":"Defines the atomic frequency comb protocol and the storage-time/bandwidth relations (t_s=1/Δ, Γ=NΔ) used throughout."},{"cited_title":"Tang, Z.-Q","cited_arxiv_id":null,"evidence_quote":"Earlier storage of quantum-dot single photons in a solid-state AFC memory, the result this work extends to a chip-integrated source and fiber at 980 nm."},{"cited_title":"Yeung, D","cited_arxiv_id":null,"evidence_quote":"Supplies the InAsP/InP nanowire quantum dots integrated on Si3N4 waveguides that act as the deterministic single-photon source."},{"cited_title":"Saglamyurek, J","cited_arxiv_id":null,"evidence_quote":"Shows erbium-doped fiber can serve as an AFC quantum memory at telecom wavelength, establishing the platform this paper moves to 980 nm."},{"cited_title":"Businger, L","cited_arxiv_id":null,"evidence_quote":"Demonstrates broadband multimode AFC storage in a rare-earth crystal, providing the multimode benchmark against which the 59-mode storage is set."},{"cited_title":"Usmani, M","cited_arxiv_id":null,"evidence_quote":"Provides the method for mapping multiple time-bin photonic qubits into one atomic ensemble, underlying the random time-bin storage test."},{"cited_title":"Gündoğan, P","cited_arxiv_id":null,"evidence_quote":"Supplies the classical fidelity bound used to certify quantum storage of the weak coherent time-bin modes."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Interfaces a quantum dot with an atomic memory and supplies the formula used to estimate g(2)_out(0) for the recalled photons."}],"review_version":1}