{"id":"ae9aa3d6-030e-482f-bb49-1c67063f5b05","arxiv_id":"2412.05480","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"An interleaved optical pumping scheme boosts storage efficiency of a telecom-band erbium quantum memory by more than an order of magnitude, enabling polarization, frequency, and time-bin qubit storage at moderate field and temperature.","lead":"Researchers stored telecom-wavelength light pulses in an erbium-doped crystal and retrieved them with up to 5% efficiency using only a moderate magnetic field and a 0.9 K cryostat. The work shows a practical route toward quantum memories for fiber networks that avoids ultra-high fields and dilution refrigerators.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Frequency- and time-bin 'qubit storage' is supported only by classical crosstalk and pulse-reordering data; no interferometric or QPT evidence certifies coherence for these two DOFs.","rationale":"The strongest claim in the paper is the headline 'storage of telecom photonic qubits encoded in polarization, frequency, and time-bin bases.' For this to hold, the memory would have to preserve coherences in all three DOFs. The paper provides full QPT only for polarization. For frequency bins, S.I. Sec. C estimates fidelity from a population ratio with one input coefficient set to zero; that is a classical mode-crosstalk measurement and cannot detect dephasing between the two frequency modes. For time bins, no fidelity or interference measurement is given at all; the experiments labelled time-bin storage in Fig. 3 are actually two pulses at different frequencies being delayed and reordered. The phrase 'coherently separated (interfered)' in the main text suggests a capability that is not demonstrated by any measured interference fringe after storage. This is not an internal inconsistency in the AFC physics—the storage medium may well preserve those coherences—but the evidence adduced does not establish the qubit-level claim as stated. I therefore agree with the reader's weakest-assumption analysis. The efficiency enhancement and the polarization QPT are believable and are not affected by this concern; the appropriate disposition is to require the missing qubit-basis characterizations before accepting the multidimensional-qubit claim. Since the reader already reached CONDITIONAL on essentially this basis, no verdict change is needed.","tokens_in":13783,"tokens_out":6815,"duration_ms":75117,"concrete_test":"Perform phase-resolved QPT/interference for both frequency-bin and time-bin qubits: prepare omega1+omega2 superpositions (via EOM sidebands) and early+late superpositions (via an unbalanced interferometer), store them in the double-AFC memory, and measure output in the conjugate bases with a phase-locked readout. A genuine qubit memory should yield visibility/process fidelity well above the classical-mixture bound; if the measured fidelity for either DOF is consistent with a classical mixture (visibility near zero), the corresponding qubit-storage claim fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central assertion that frequency- and time-bin photonic qubits are stored as qubits is not supported by the data. For frequency bins, the only quantitative estimate (S.I. Sec. C) sets beta=0 and computes F_alpha = alpha'/(alpha'+beta'), i.e., the ratio of echo counts in the correct vs wrong AFC window. This is a classical crosstalk/population measurement; it is insensitive to the relative phase between the two frequency modes and therefore cannot certify a coherent superposition. The main-text demonstration (Fig. 3) stores two time bins at two different frequencies and reorders them; this is mode-selective delay, not a demonstration of time-bin qubit storage, which requires a superposition alpha|early>+beta|late> in the same mode and measurement of its interference after retrieval. No such measurement appears anywhere in the paper, including the Supplementary Information. Thus the abstract's 'storage of photonic qubits encoded in frequency and time-bin bases' and the 'multidimensional qubit storage' phrasing overstate what is shown. The polarization QPT is a valid check for one DOF, but it cannot validate the other two.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript reports an atomic-frequency-comb (AFC) optical memory in isotopically purified 167Er3+:YSO operated at 0.9 K and 1.1 T generated by permanent magnets, using an 'interleaved pumping' scheme (periodic ~10 ms pump interruptions) for spectral tailoring. The authors report classical and single-photon-level storage efficiencies of 6.1% and 5.0%, an ~8-fold improvement over their continuous-pumping baseline, and consistent with the standard AFC efficiency formula (predicted ~7.5% from independently measured OD ≈ 3.3, finesse ≈ 3, and background absorption OD ≈ 1.3). They demonstrate polarization-qubit storage with quantum process tomography (single-photon-level process fidelities 92.3% and 95.5% for two time bins) and demonstrate storage of frequency- and time-bin-encoded pulses at the single-photon level, including FIFO/FILO reordering via frequency-dependent delays, with crosstalk-based fidelity estimates of 99.7% and 97.2% for the frequency windows. The paper claims 'multidimensional qubit storage' in polarization, frequency, and time-bin bases.","tokens_in":14008,"tokens_out":18188,"duration_ms":161797,"significance":"If the results hold, the practical advances are real: an erbium AFC memory operating at 0.9 K and 1.1 T with table-top permanent magnets rather than a dilution refrigerator and superconducting magnet, and an interleaved pumping scheme giving an in-experiment ~8-fold efficiency gain over continuous pumping. A genuine strength is the efficiency consistency check: Eq. (1) is the standard AFC expression, evaluated with separately measured parameters (OD, finesse, background absorption) to predict 7.5% versus a measured 6%, so the agreement is a legitimate test rather than a fit. The polarization quantum process tomography (92.3% and 95.5% at the single-photon level) is a correctly executed certification of one degree of freedom, and the stated claims are falsifiable (e.g., the projected >30% efficiency with extended ground-state lifetimes). The paper's current significance rests mainly on the efficiency and practicality improvements and on the polarization-qubit demonstration; the multidimensional-qubit claim, if substantiated, would raise the impact considerably, but as presented it exceeds the evidence.","major_comments":[{"comment":"The frequency-bin fidelity estimate in S.I. Sec. C, F_alpha = alpha'/(alpha'+beta'), is evaluated with beta = 0, i.e., with a single-frequency input pulse, so it measures only the classical population (crosstalk) ratio between the two AFC windows. This quantity is independent of the relative phase between the two frequency modes and therefore cannot certify that a superposition alpha|0>_f + beta|1>_f remains coherent through storage and retrieval. The Fig. 3 demonstrations are mode-selective delays and reordering of pulses launched into separate windows, which is classical pulse routing. To support the abstract's claim of storing qubits in the frequency basis, the authors should either present a phase-sensitive characterization (for example, quantum process tomography in the frequency basis or an interference-fringe visibility measurement on a retrieved superposition) or revise the claim to 'storage of frequency-encoded pulses.'","section":"S.I. Sec. C; Section III.B; Fig. 3"},{"comment":"The experiments described in Section III.B and Fig. 3 store individual pulses in two successive time bins at two different carrier frequencies and reorder them through frequency-dependent AFC delays (FIFO/FILO). This demonstrates storage and reordering of time-encoded pulses, but it does not test the defining property of a time-bin qubit: the coherence of a superposition alpha|early> + beta|late> in a single spectral mode. No unbalanced-interferometer measurement or equivalent phase-sensitive test appears in the main text or the S.I., so the claim of 'time-bin qubit storage' is unsupported by the presented data. The authors should add such a measurement or state explicitly that only storage of time-encoded pulses (classical delays) was demonstrated.","section":"Section III.B; Fig. 3"},{"comment":"The abstract and the conclusion assert storage of photonic qubits in polarization, frequency, and time-bin bases, with quantum process tomography 'achieving a fidelity exceeding 92%.' As Section III.C makes clear, the QPT and the 92-96% fidelities apply only to the polarization degree of freedom; for the frequency and time-bin encodings the Introduction itself (Section I) states that fidelity was evaluated by 'intensity noise estimation.' The headline claims therefore overstate the demonstrated capabilities. The manuscript should either add the missing coherence measurements for the frequency and time-bin degrees of freedom or rescale the central claims to match what was measured: high-fidelity polarization-qubit storage plus classical-level crosstalk and pulse-routing demonstrations for the other two encodings.","section":"Abstract; Section V"}],"minor_comments":[{"comment":"Section II contains a corrupted passage — 'consist of 8 and 4I15/26 Krathemers doublets, respectively, and then7. These doublets become Applying a strong magnetic field ...' — which must be repaired to restore the intended statement about the Kramers doublets of the 4I15/2 and 4I13/2 multiplets, and the typo 'Krathemers' should be corrected to 'Kramers.'","section":"Section II"},{"comment":"The abstract says the interleaved scheme 'improves storage efficiency by over an order of magnitude,' but Section III.A reports a factor of about 8 relative to continuous pumping in this work; the order-of-magnitude statement refers to comparison with Ref. [43]. The baseline for the abstract claim should be stated explicitly, and the cross-platform comparison with Ref. [43] (a nanophotonic device) should be discussed with its caveats.","section":"Abstract; Section III.A"},{"comment":"The optical coherence time is reported as T2 = 169 µs in the main text but as 'T2 = 169 ms' in the caption of S.I. Fig. S1(c); these values are inconsistent and must be reconciled with the correct measured value and its uncertainty.","section":"Section III.A; S.I. Fig. S1(c)"},{"comment":"The S.I. contains unfilled citation placeholders — '[ ? ]' in Sec. C and '[ ? ]' and '[ ? ? ]' in Sec. E — that must be replaced with the intended references or removed.","section":"S.I. Secs. C and E"},{"comment":"The main text should state the actual mean photon number per pulse for the measurements labelled 'single-photon' (the S.I. reports detected counts as large as n_in = 2.557 with a Poisson-based saturation correction, but the input mean photon number at the crystal is not given in the main text), and the term 'single photons' should be qualified as weak coherent pulses; the uncertainty on the reported 5.03% single-photon efficiency should also be given.","section":"Section III.C; S.I. Sec. B"},{"comment":"The sentence 'The frequency bins can be coherently separated (interfered) when they enter the memory at the same (different) time bins' is unclear, and since no interference measurement appears in Fig. 3, the coherence claim in that sentence should be removed or supported by quantitative data.","section":"Section III.B"}],"recommendation":"major_revision","confidential_remarks":"The core technical content — interleaved pumping, the efficiency consistency check, and the polarization QPT — is sound and publishable, but in my reading the frequency- and time-bin 'qubit storage' claims rest only on classical crosstalk and pulse-reordering measurements. I would encourage the editor to require either phase-sensitive data for those two degrees of freedom or a substantial rescaling of the title, abstract, and conclusions; otherwise the gap between claims and evidence is too large for a quantum memory paper. The comparison with Ref. [43] spans different platforms and should be scrutinized during revision. The corrupted sentence in Section II also suggests the manuscript needs careful proofreading before resubmission."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Punchline: the interleaved pumping scheme is a real, practical improvement, and the polarization process tomography is solid. But the abstract overstates the evidence: frequency- and time-bin 'qubit storage' is demonstrated as classical mode-selective delay, not as coherent qubit storage. The stress-test note lands, and I agree with the reader's conditional verdict.\n\nWhat is new and good: interleaved pumping—periodically switching the pump off for ~10 ms so excited atoms decay into shelving states—is simple and yields 6.1% classical and 5.0% weak-coherent-pulse efficiency at 0.9 K and ~1.1 T, roughly an order of magnitude better than continuous pumping and previous low-field work in Er:YSO. The consistency check with Eq. (1) (predicted ~7.5% vs observed ~6%) uses independently measured OD, finesse, and background absorption; that is a legitimate test, not a circular fit. The single-photon-level polarization QPT (92.3% and 95.5% fidelity for two time bins) is real evidence for one DOF, and the analysis of the non-identity components is careful.\n\nWhere it is soft: the frequency-qubit fidelity in S.I. Sec. C is a classical crosstalk ratio. Setting beta=0 and computing F_alpha = alpha'/(alpha'+beta') measures how much echo lands in the correct AFC window; it is insensitive to relative phase between the two frequency modes, so it cannot certify a coherent superposition. The time-bin 'storage' shown in Fig. 3 is FIFO/FILO reordering of two pulses that already occupy different frequency windows; there is no measurement of a superposition alpha|early>+beta|late> interfering after retrieval, and no interferometric or two-photon characterization. So the abstract's claim of storing photonic qubits in frequency and time-bin bases goes beyond the data. Also, 'single-photon' pulses are weak coherent states with mean photon number ~0.4; that should be stated honestly, though it is common in memory demonstrations. Minor: T2 is 169 µs in the text but 169 ms in the Fig. S1 caption; fix that.\n\nNet assessment: the efficiency result and the polarization tomography deserve referee time, and the interleaved pumping scheme is worth knowing. But the multidimensional qubit-storage claim needs either new coherence measurements or a rewrite to 'mode-selective storage and reordering.' I would send it to peer review with that as the required revision.","headline":"Useful efficiency scheme and solid polarization tomography, but the frequency- and time-bin 'qubit storage' claims outrun the data.","tokens_in":14522,"tokens_out":3066,"would_cite":true,"duration_ms":30778,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Telecom photons stored in three qubit bases at 92% fidelity.","keywords":["quantum memory","atomic frequency comb","erbium-doped crystal","telecom photons","multidimensional qubits","interleaved pumping","quantum process tomography","time-bin qubit"],"falsifier":"A decisive test would be storing a time-bin superposition state (for example one photon in (|early⟩+|late⟩)/√2) and measuring the interference visibility of the retrieved echo after a variable relative phase; a visibility below the classical limit, or a nonzero complex phase in time-bin quantum process tomography, would show the memory is not storing time-bin qubits coherently.","tokens_in":13565,"feed_emoji":"⚛️","tokens_out":7263,"duration_ms":67067,"temperature":0.7,"pith_summary":"This paper reports a quantum memory for telecom-band light that stores single-photon-level pulses in three photonic encodings—polarization, frequency, and time-bin—simultaneously or separately. The memory is an atomic frequency comb in an erbium-167 doped YSO crystal operated at 0.9 K and 1.1 T, conditions reached with a table-top cryostat and permanent magnets rather than a dilution refrigerator or superconducting magnet. The main new mechanism is an interleaved spectral-hole pumping scheme that periodically interrupts the pump to let excited atoms decay into shelving states, boosting storage efficiency by over an order of magnitude compared to previous low-field erbium memories. Polarization qubits are verified with quantum process tomography showing fidelity above 92%; frequency and time-bin storage are shown at the single-photon level, with frequency fidelity estimated from classical intensity ratios. If these results hold, practical telecom quantum memories and multiplexed quantum repeaters become more accessible.","feed_headline":"Telecom photons stored in three qubit bases at 92%","feed_subtitle":"Interleaved pumping lifts efficiency tenfold in a 0.9 K erbium crystal memory.","key_machinery":"The central object is the interleaved optical pumping sequence used to prepare atomic frequency combs in erbium: a batch of complex hyperbolic secant pulses that sweep over the target frequency window to shelve atoms into long-lived hyperfine states, followed by a 10 ms in-loop delay during which the pump is off so that excited atoms can decay into the shelving states rather than being stimulated back. Repeated cycles create deep, narrow spectral holes. This scheme is what raises the comb finesse and lowers background absorption, making the storage efficiency of the AFC memory practical at 0.9 K and 1.1 T. The AFC itself—a periodic series of absorption peaks spaced by 1–2 MHz—carries the storage and re-emission: an absorbed photon is collectively re-emitted as an echo after a time set by the comb spacing, giving a memory with a controllable delay in the 0.5–1 µs range demonstrated here.","core_discovery":"On its own terms, the paper establishes that an AFC memory in 167Er3+:YSO can be initialized efficiently under mild conditions and can store telecom photonic qubits in multiple degrees of freedom. Using interleaved pumping with complex hyperbolic secant pulses and a 10 ms in-loop delay, the authors create AFC combs with finesse F≈3 and background absorption d0≈1.3, achieving up to 6.1% storage efficiency for classical pulses and 5.0% for single-photon pulses—an 8-fold improvement over continuous pumping and more than an order of magnitude over prior low-field work. Quantum process tomography on polarization qubits yields process fidelities of 92.3% and 95.5% for two time-bin echoes at the single-photon level. Two spectrally separated AFC windows allow frequency-dependent delay, and the paper reports FIFO and FILO reordering of time-bin pulses. The central claim is therefore that multidimensional telecom qubit storage and efficient initialization are possible without extreme magnetic fields or millikelvin temperatures.","pith_inferences":["The paper's \"multidimensional qubit storage\" claim currently rests mainly on polarization tomography; until time-bin and frequency-bin coherence are verified with quantum interference or full state tomography, the storage of those encodings should be read as classical-level demonstrations.","If the interleaved pumping principle transfers to other long-lived excited-state ion platforms (for example europium or praseodymium), it could become a standard initialization tool for AFC memories beyond erbium.","A direct extension of this work would be a time-bin two-photon interference experiment: store a photon in a superposition of two time bins and check that the interference visibility after retrieval exceeds the classical bound, which would certify genuine time-bin qubit storage.","The efficiency formula used in the paper suggests that further reducing background absorption d0 while keeping finesse F high could push efficiency to tens of percent; a systematic d0–F optimization using the interleaved pump would test that scaling."],"forward_implications":["If the interleaved pumping scheme proves general, erbium-based telecom memories can operate with table-top permanent magnets and a 0.9 K cryostat rather than a dilution refrigerator, lowering the barrier to practical deployment.","The demonstrated simultaneous storage of two pulses in distinct frequency windows enables frequency-dependent delay and first-in-first-out or first-in-last-out reordering, which can serve as a coherent pulse processor for quantum repeaters.","With an extended ground-state lifetime (for example via spin-polarization initialization), the paper projects storage efficiency exceeding 30% under the same field and temperature.","Storing polarization, frequency, and time-bin degrees of freedom in one crystal provides building blocks for multiplexed quantum memories and entanglement distribution over telecom fibers.","Process fidelity above 92% for polarization qubits at the single-photon level indicates compatibility with quantum-repeater protocols that require high-fidelity storage and retrieval."],"supporting_citations":[{"why":"Introduces the atomic frequency comb protocol that provides the storage and re-emission mechanism used throughout the paper.","marker":"[26]"},{"why":"Supplies the complex hyperbolic secant pulse shaping and the AFC efficiency equation used to design and predict the interleaved-pumped comb.","marker":"[42]"},{"why":"Represents the prior low-magnetic-field erbium memory whose efficiency this work improves by over an order of magnitude.","marker":"[43]"},{"why":"Provides the high-field (7 T) erbium memory benchmark of 22% efficiency that this work seeks to approach under milder experimental conditions.","marker":"[34]"},{"why":"Gives the magnetic g-tensor data used to align the field and freeze electron spins, extending the hole lifetime to over 3 seconds.","marker":"[37]"},{"why":"Supplies the standard quantum process tomography routine used to extract the reported polarization process fidelities.","marker":"[45]"}],"fun_headline_variants":["Telecom quantum memory: three bases, 92% fidelity","Tenfold efficiency boost in erbium telecom memory","Solid-state memory stores telecom photons in three formats","Efficient telecom photon storage in three quantum bases"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The claim that frequency and time-bin storage are quantum qubit storage assumes that those encodings remain coherent through the memory; the paper estimates frequency fidelity from a classical intensity ratio (setting β=0) and does not measure time-bin fidelity, so if either encoding loses phase coherence the demonstration reduces to classical pulse delay.","fun_headline_variants_meta":{"raw":{"variants":["Telecom quantum memory: three bases, 92% fidelity","Tenfold efficiency boost in erbium telecom memory","Solid-state memory stores telecom photons in three formats","Efficient telecom photon storage in three quantum bases"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000559,"raw_usage":{"total_tokens":2620,"prompt_tokens":873,"completion_tokens":1747,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":489,"completion_tokens_details":{"reasoning_tokens":1684}},"tokens_in":489,"tokens_out":1747,"duration_ms":14923,"temperature":1.0,"reasoning_tokens":1684,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T20:40:31.129983+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A decisive test would be storing a time-bin superposition state (for example one photon in (|early⟩+|late⟩)/√2) and measuring the interference visibility of the retrieved echo after a variable relative phase; a visibility below the classical limit, or a nonzero complex phase in time-bin quantum process tomography, would show the memory is not storing time-bin qubits coherently.","supporting_citations":[{"cited_title":"Afzelius, C","cited_arxiv_id":null,"evidence_quote":"Introduces the atomic frequency comb protocol that provides the storage and re-emission mechanism used throughout the paper."},{"cited_title":"Jobez, N","cited_arxiv_id":null,"evidence_quote":"Supplies the complex hyperbolic secant pulse shaping and the AFC efficiency equation used to design and predict the interleaved-pumped comb."},{"cited_title":"Craiciu, M","cited_arxiv_id":null,"evidence_quote":"Represents the prior low-magnetic-field erbium memory whose efficiency this work improves by over an order of magnitude."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the high-field (7 T) erbium memory benchmark of 22% efficiency that this work seeks to approach under milder experimental conditions."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the magnetic g-tensor data used to align the field and freeze electron spins, extending the hole lifetime to over 3 seconds."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the standard quantum process tomography routine used to extract the reported polarization process fidelities."}],"review_version":1}