{"id":"1fbcbacc-77d3-4652-a109-527e5632caaf","arxiv_id":"2608.04093","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"An optical-fibre-integrated recirculating-loop buffer stores polarisation-encoded qubit payloads for up to 47 μs and reads a classical header that sets the storage time.","lead":"This paper demonstrates an all-fibre, telecom-compatible buffer that stores polarisation-encoded quantum pulses for up to 47 microseconds while reading out a classical packet header that decides when the pulses are released. The advance matters because practical quantum networks will need such fibre-integrated buffers, which previous loop memories could not provide.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The reported QBER values rely on per-cycle manual polarization re-optimization and, in the stability run, on an automatic VOPC that recalibrates whenever QBER exceeds 5%; the central polarization-qubit storage claim is therefore not yet separated from active polarization correction.","rationale":"The reader's weakest_assumption identifies exactly the concern I consider most load-bearing: the polarization transformation per storage cycle is assumed unitary, reciprocal, and stable over each acquisition so that manual PC and VOPC compensation are unbiased. My stress-test agrees and sharpens the issue. The paper's QBER values are obtained after manual per-cycle/basis optimization, and the long-term stability result includes automatic recalibration whenever QBER exceeds 5%, so the quoted numbers characterize the corrected system rather than the raw buffer. This does not make the buffer demonstration invalid, but it means the central claim of high-fidelity polarization storage is conditional on the adequacy of active polarization correction. A fixed-settings run and a Mueller-matrix check would resolve the ambiguity. I do not find a more serious flaw: the loss budget, exponential decay, header readout, and packet-processing demonstrations are internally consistent, and the absence of a single-photon source or of fabrication details for the poled fibre are important limitations but not fatal to the proof-of-principle. The reader's CONDITIONAL verdict is therefore appropriate, and my analysis does not move it.","tokens_in":13592,"tokens_out":8574,"duration_ms":98240,"concrete_test":"Perform a fixed-settings QBER characterization: after one initial polarization optimization at cycle 1, acquire 30 s runs for all four BB84 states at cycles n=1,3,4,6,8 with no per-cycle manual re-optimization beyond the automatic VOPC, and compare the resulting QBER values with Fig. 5b. In parallel, measure the round-trip Mueller matrix of the storage line plus Sagnac switch with classical light at 1546.9 nm for each cycle and check whether the transformation is unitary to within the QBER error budget. If QBER stays below the reported values without per-cycle manual alignment and the Mueller matrix is unitary, the concern is resolved; if QBER degrades or the matrix is non-unitary, the quoted fidelity overstates buffer performance.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that a polarization-encoded qubit payload is stored and retrieved with 1.85% average QBER over up to 8 storage cycles. This requires that the round-trip polarization transformation accumulated per cycle is unitary, reciprocal, and stable over each 30 s acquisition, so that a single manual polarization controller can compensate it without bias and the PBS projection measures the prepared qubit. The paper states that the manual PCs compensate residual polarization transformations for each storage cycle just before Fig. 4, and Fig. 5 data are acquired after choosing the measurement basis with a manual PC for each cycle. The 12-hour stability run in Fig. 6 pauses and actively recalibrates the VOPC whenever the QBER exceeds 5%. No zero-storage baseline QBER is reported, and no direct measurement of the round-trip Mueller matrix or of PDL in the storage path is given. If the residual transformation is non-unitary (e.g., polarization-dependent loss in the FBG, coupler, or poled-fibre modulator) or drifts within a 30 s acquisition, a unitary PC cannot fully compensate it, and the quoted QBER would not represent the intrinsic buffer fidelity. This is load-bearing because the headline contribution is specifically the storage of polarization qubits in a fully fibre-integrated buffer; the fidelity numbers are the quantitative evidence for that claim.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports an all-fibre optical buffer for packet-switched quantum networks, built from a Sagnac-loop switch controlled by a poled-fibre phase modulator and a 100 m storage line terminated by a fibre Bragg grating. It demonstrates storage and retrieval of 16-pulse weak-coherent polarization-encoded payloads for up to eight recirculations (47 μs), with a per-cycle efficiency of about 56%, cycle-resolved quantum bit error rates between 1.06% and 2.74% (average 1.85%), header-controlled retrieval time, and a 12-hour stability run under active polarization realignment. The central claim is that this establishes the first fully fibre-integrated buffer capable of packet-level processing for quantum networks.","tokens_in":13784,"tokens_out":5232,"duration_ms":58533,"significance":"If the fidelity claims hold, this is a useful step forward: it moves recirculating-buffer demonstrations from bulk-optical implementations to a fully fibre-integrated, telecom-compatible platform and introduces poled-fibre phase modulators as low-loss, polarization-insensitive switches for quantum information processing. The paper has several concrete quantitative strengths: a transparent per-cycle loss budget (2.5 dB from listed components, giving about 56% efficiency), explicit Poissonian error bars on the QBER values, a direct comparison with a hypothetical commercial LiNbO3 modulator in Fig. 4, and a clear proof-of-principle packet-processing demonstration in Fig. 7 with header-triggered variable delays. The 12-hour run, although feedback-assisted, is a useful systems-level data point. I also note that there is no fitted-parameter circularity: the only fit in the paper is the exponential decay in Fig. 4, and it is not used to define a target quantity.","major_comments":[{"comment":"The central qubit-storage claim rests on the QBER values in Fig. 5b, but the text states that 'the manual polarisation controllers are needed to compensate residual polarisation transformations for each storage cycle' and that the measurement basis is selected with a manual PC for each cycle. No zero-storage baseline QBER is reported, and no Mueller-matrix or polarization-dependent-loss characterization of the storage path is given. As written, the quoted QBER is not separated from per-cycle manual polarization alignment, so it cannot be assigned to the buffer alone. Please report a zero-cycle baseline for the same states, specify how many manual adjustments were made per storage cycle and basis, and provide a direct measurement of the round-trip polarization transformation (or at least an upper bound on PDL) showing that a single unitary polarization controller can in principle compensate it without bias.","section":"Experimental results, Fig. 5"},{"comment":"The 12-hour stability claim is made with an active VOPC recalibration routine that pauses the measurement whenever the QBER exceeds 5%. Consequently, Fig. 6 demonstrates the stability of the feedback-stabilized system rather than of the passive buffer itself, and the frequency of recalibration events is not reported. Please report the QBER before each recalibration, the total number of recalibrations over the 12 hours, and, if possible, a comparison segment with the feedback disabled, so that the reader can distinguish intrinsic drift from loop performance.","section":"Experimental results, Fig. 6"},{"comment":"The QBER values are computed as QBER = 1 - P_{|i>} with P_{|i>} = N_{D|i>}/(N_{D|i>}+N_{D|i>⊥}), but the per-state probabilities for H, V, D, and A at each cycle are not tabulated; only normalized curves are shown in Fig. 5a. The stated error bars are Poissonian counting errors only, and they do not include the systematic uncertainty from the manual PC settings, which is the procedure-sensitive part of the measurement. Please provide the raw probability matrix for each storage cycle, or a tomographic state or process fidelity, so that the reported error model matches the actual measurement procedure.","section":"Experimental results, QBER calculation"},{"comment":"The claim of about 56% per-cycle efficiency is supported by a component loss budget, but the exponential fit in Fig. 4 is not quantified. Please report the fitted per-cycle decay constant with its uncertainty and compare it with the loss budget. This would also substantiate the statement that retrieval events are visible after ten storage cycles and would make the efficiency claim independently checkable.","section":"Experimental results, Fig. 4"}],"minor_comments":[{"comment":"The sentence 'the needed voltage to apply a π phase shift decreases from approximately (Vπ ≈ 5 kV) for 950 V to our device' is garbled and should be rephrased to state the poled and unpoled Vπ values clearly.","section":"Poled optical fibre phase modulator"},{"comment":"The red LiNbO3 comparison curve should state the assumed per-cycle insertion loss explicitly, rather than only 'typical 3 dB loss', so that the comparison is reproducible.","section":"Experimental results, Fig. 4"},{"comment":"The notation 'M1' in the text is confusing; please define whether it is a specific destination user or the set of M users, and use the notation consistently with the figure.","section":"Quantum network buffering, Fig. 1"},{"comment":"The header is described as 'three extra pulses' that allow encoding of up to eight addresses 'assuming direct digital amplitude modulation'; please clarify how three pulses encode eight addresses and how the on/off encoding shown in Fig. 7 maps to the storage-cycle delay.","section":"Experimental results, packet processing"},{"comment":"The caption reports an average QBER of 1.85±0.13% over all cycles, but the main text lists individual QBER values without defining whether the average is weighted or unweighted; please make the definition explicit.","section":"Experimental results, Fig. 5 caption"},{"comment":"The phrase 'is about ≈6 μs' contains a redundant approximation symbol; please write 'about 6 μs' or 'approximately 6 μs'.","section":"Fibre-optical active loop memory"}],"recommendation":"major_revision","confidential_remarks":"The main open question is whether the reported QBER values can be attributed to the buffer rather than to the per-cycle manual polarization alignment and the active VOPC feedback. I do not regard this as a fatal flaw, because polarization compensation is standard in fibre-loop experiments, but the manuscript needs a zero-storage baseline and a characterization of the round-trip polarization transformation before the central fidelity claim is fully supported. I would also ask the editor to ensure the 'first fully fibre-integrated buffer' novelty claim is properly scoped against references [31-33], which are discussed in the paper but could be described more explicitly."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's the short version: this is a real experimental advance - the first fully fibre-integrated buffer that reads a classical header and delays a polarization-qubit payload accordingly - and the authors are straight about its limits. The poled-fibre phase modulator is the key enabler: 0.4 dB loss, roughly 25 ns switching, polarization-insensitive. They show 16-pulse payloads stored up to 8 cycles (47 us) with about 56% per-cycle efficiency and QBER 1.06-2.74%, plus a 12-hour stability run. That is worth a serious referee.\n\nWhat's new: prior recirculating-loop memories exist, including a fibre-loop EO-switch buffer [33], but none do header-controlled packet processing in an all-fibre architecture. The comparison plot (Fig 8) with qubit yield Y = eta*N is a fair way to benchmark against matter memories and loop memories.\n\nWhere it's soft, in proportion: the headline QBERs are taken after manually optimizing the polarization controllers for each storage cycle. The paper says this explicitly. So the numbers tell you the buffer can preserve polarization when someone actively compensates the fibre birefringence; they don't tell you the intrinsic round-trip polarization stability. The 12-hour stability run uses an automatic VOPC that recalibrates whenever QBER exceeds 5%, so 'stable operation' means stable with active feedback, not passively stable. No zero-storage baseline QBER is reported, so preparation/measurement error isn't separated from buffer error; that's a minor omission because the n=1 numbers are low, but it should be added. The custom poled fibre isn't commercially available and no fabrication recipe, data, or code is shipped, which limits reproducibility, though the poling technique is published. Weak coherent states rather than true single photons are fine for a proof of principle, but should be flagged as a limitation in a QKD context.\n\nThe stress-test worry - that non-unitary PDL or drift within a 30 s acquisition could bias the QBER - does not land as a fatal objection. The paper never claims a measured Mueller matrix; it claims that with active compensation the QBER is low, and that is exactly what it demonstrates. A more careful paper would characterize the round-trip PDL and add a no-storage baseline, but the central engineering claim holds.\n\nFor whom: experimental quantum networking people, and anyone needing a low-loss in-fibre phase modulator. It deserves peer review. My recommendation is conditional accept, with requests for the baseline QBER, a PDL/birefringence characterization, and a clearer separation of passive versus active stability.","headline":"A genuinely new all-fibre packet buffer, honestly reported; the QBER numbers rely on active polarization correction, so conditional accept is the right call.","tokens_in":14446,"tokens_out":2253,"would_cite":true,"duration_ms":24234,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["03.67.Hk","42.50.Ex","42.79.Sz"],"model":"deepseek-v4-flash","headline":"The paper demonstrates an all-fibre quantum buffer that stores a 16-pulse polarization-encoded qubit payload for up to 47 microseconds, reads the packet header while the payload is stored, and retrieves the payload with 1.85% average QBER.","keywords":["quantum networks","packet switching","optical fibre buffer","recirculating loop","poled fibre phase modulator","polarisation qubits","Sagnac interferometer","telecom-band quantum memory"],"falsifier":"Measure the QBER of a fixed input state over several storage cycles with no polarisation controller adjustment and with integration times much shorter than 30 seconds; if the QBER rises above the reported 1.85% average as the integration window shrinks, then the quoted error rate is an average over a drifting polarisation rather than an intrinsic property of the stored qubit.","tokens_in":13324,"feed_emoji":"⚛️","tokens_out":7926,"duration_ms":75701,"temperature":0.7,"pith_summary":"Packet-switched quantum networks need buffers that delay a payload of quantum states while routing information is processed; the paper shows that such a buffer can be built entirely in optical fibre. The key device is an ultra-low-loss poled-fibre phase modulator, which makes it possible to switch, store, and retrieve a polarization-encoded qubit payload in a recirculating Sagnac loop, and to split a classical header from the payload inside the buffer. The experiment stores a 16-pulse weak-coherent payload for up to 47 microseconds (eight storage cycles) with about 56% efficiency per cycle and an average quantum bit error rate of 1.85%, and it demonstrates stable operation for over twelve hours. If these results are correct, they point to a practical, telecom-compatible buffering component for future quantum network nodes.","feed_headline":"All-fibre quantum buffer reads headers while payload waits","feed_subtitle":"A poled-fibre modulator lets a recirculating loop store qubit packets up to 47 μs at 1.85% average QBER.","key_machinery":"The central object is the poled-fibre phase modulator: a single-mode silica fibre with two bismuth-filled microchannels that, after poling, induces a π phase shift for a 950 V pulse, with only 0.4 dB loss and 22–26 ns response. It sits inside a fibre Sagnac interferometer that works as a polarization-insensitive optical switch: at zero relative phase a packet remains in the loop, at π it is routed into a storage line made of a 100 m fibre spool and a fibre Bragg grating mirror. Two driving pulses on the modulator, separated by a chosen delay, inject and then retrieve the packet; adjusting that delay sets the number of storage cycles, so the storage time is controlled by the header. A manual polarisation controller per cycle cancels accumulated birefringence, and a voltage-operated polarisation controller (VOPC) automatically realigns the state to the measurement PBS when the detected QBER exceeds 5%, enabling the long stability run.","core_discovery":"The paper's central claim is that a fully fibre-integrated recirculating buffer can store and retrieve polarization-encoded qubit packets while performing header readout, a capability not previously shown in fibre. The authors demonstrate storage of a 16-pulse payload at 0.1 photons per pulse, retrieval after one to eight cycles (the longest being 47 μs), and fast in-buffer separation of a three-pulse classical header from the delayed payload. They report per-cycle efficiency of about 56%—compared with the roughly 50% per-pass loss expected from a commercial LiNbO3 modulator—and an average QBER of 1.85% over the four BB84 states and all storage cycles, leading to an effective qubit yield of approximately nine recovered qubits per storage event. The buffer works at 1546.9 nm and maintains stable QBER (below a 5% threshold with automatic VOPC recalibration) over more than twelve hours, supporting the claim that it can operate as a practical network-layer primitive.","pith_inferences":["The same poled-fibre modulator, with its low loss and nanosecond-scale switching, could be applied to high-speed basis selection or measurement-setting randomisation in device-independent quantum communication; the paper mentions this only as a future direction.","If the manual polarisation controllers were replaced by fast, actively tracked compensation inside the loop, the buffer could accept arbitrary incoming polarisation states without reconfiguration, a step the paper does not claim but that would be needed for field deployment.","The qubit-yield comparison in the paper suggests a design trade-off: this loop buffer maximises recovered qubits for short delays, whereas matter memories provide longer and on-demand storage; a hybrid node combining both would likely be needed for full quantum-network functionality, something the paper does not discuss."],"forward_implications":["A fully fibre-integrated buffer can be spliced directly into existing telecom links, since it needs no free-space optics, cryogenic cooling, or wavelength conversion.","A network node can read a packet's header and set the payload's delay on the fly, which allows collision avoidance and time-slot assignment in a star-topology quantum switch.","Replacing four FC/PC connectors inside the Sagnac loop with splices and using a lower-excess-loss coupler would raise per-cycle transmissivity to roughly 72%, extending the storage-time range of the same architecture.","Because the buffer is designed for entire packets, the payload density can be increased by using narrower, more closely spaced pulses (for example, 45 ps pulses with 400 ps separation) without altering the storage mechanism."],"supporting_citations":[{"why":"Supplies the recirculating-loop quantum memory concept that the all-fibre buffer implements.","marker":"[22]"},{"why":"Gives the polarisation-independent fibre Sagnac switch used as the store/retrieve routing element.","marker":"[37]"},{"why":"Describes the integrated fibre electro-optic switch fabrication underlying the poled-fibre phase modulator.","marker":"[41]"},{"why":"Defines the BB84 protocol and the QBER quantity used to quantify storage fidelity.","marker":"[5]"},{"why":"Demonstrates the closest prior fibre-loop quantum buffer with an electro-optical switch, the baseline this work extends to packet-level processing.","marker":"[33]"},{"why":"Reports a fibre-integrated quantum memory for telecom light, providing comparative context for the fibre-integration claim.","marker":"[31]"},{"why":"Reports a loop-based broadband quantum memory used in the qubit-yield comparison.","marker":"[34]"}],"fun_headline_variants":["All-fibre buffer reads headers, stores qubits up to 47 μs","Fibre loop stores qubit packets, reads headers live","Quantum buffer in fibre: header readout, 47 μs storage","Fibre-integrated buffer reads routing headers, delays qubits","All-fibre loop reads routing headers while qubits wait"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The result depends on the residual polarisation change per storage cycle being unitary, reciprocal, and stable over each 30-second data run, so that one manual polarisation controller per cycle can fully undo it and the final polarising beam splitter measures the prepared qubit without bias.","fun_headline_variants_meta":{"raw":{"variants":["All-fibre buffer reads headers, stores qubits up to 47 μs","Fibre loop stores qubit packets, reads headers live","Quantum buffer in fibre: header readout, 47 μs storage","Fibre-integrated buffer reads routing headers, delays qubits","All-fibre loop reads routing headers while qubits wait"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00242,"raw_usage":{"total_tokens":9305,"prompt_tokens":948,"completion_tokens":8357,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":564,"completion_tokens_details":{"reasoning_tokens":8267}},"tokens_in":564,"tokens_out":8357,"duration_ms":63417,"temperature":1.0,"reasoning_tokens":8267,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-08T00:34:34.435199+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the QBER of a fixed input state over several storage cycles with no polarisation controller adjustment and with integration times much shorter than 30 seconds; if the QBER rises above the reported 1.85% average as the integration window shrinks, then the quoted error rate is an average over a drifting polarisation rather than an intrinsic property of the stored qubit.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the recirculating-loop quantum memory concept that the all-fibre buffer implements."},{"cited_title":"Alarc´ on, P","cited_arxiv_id":null,"evidence_quote":"Gives the polarisation-independent fibre Sagnac switch used as the store/retrieve routing element."},{"cited_title":"Fokine, L","cited_arxiv_id":null,"evidence_quote":"Describes the integrated fibre electro-optic switch fabrication underlying the poled-fibre phase modulator."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the BB84 protocol and the QBER quantity used to quantify storage fidelity."},{"cited_title":"Fook Lee, G","cited_arxiv_id":null,"evidence_quote":"Demonstrates the closest prior fibre-loop quantum buffer with an electro-optical switch, the baseline this work extends to packet-level processing."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reports a fibre-integrated quantum memory for telecom light, providing comparative context for the fibre-integration claim."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reports a loop-based broadband quantum memory used in the qubit-yield comparison."}],"review_version":1}