{"id":"d9e61569-9990-452b-be6e-2d8dbfe593c5","arxiv_id":"2412.12379","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Experiments achieve 28.5% storage efficiency and a 630 MHz bandwidth in a Tm:YAG atomic frequency comb memory at 3.5 K, and propose a scheme for wider bandwidth.","lead":"This paper reports a thulium-doped YAG crystal quantum memory that stores light pulses with 28.5% efficiency at 3.5 kelvin, the best reported for this setup without a cavity or dilution refrigerator. It also demonstrates storage across a 630 MHz frequency span and proposes a pumping scheme that could make such memories faster and more efficient for quantum networks.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 28.5% measured result is plausible, but the forward-looking claim of commensurate intrinsic pumping rests on untested linear Zeeman coefficients; if the splitting ratio drifts with field, temperature, or ion site, the central generalization fails.","rationale":"I read the paper as making two distinct contributions: a measured record AFC storage efficiency of 28.5±0.2% at 30 MHz in a free-space Tm3+:YAG crystal at 3.5 K, and a proposed commensurate intrinsic pumping scheme intended to extend bandwidth by orders of magnitude. The measured efficiency appears internally consistent: the stated optical depth, finesse, and background absorption produce efficiencies of the order shown, the noise counts are characterized, and the multi-pass geometry is a plausible route to the reported enhancement. I do not see a concrete internal inconsistency that would invalidate the 28.5% result. The weakest load-bearing assumption is in the forward-looking proposal. Section III.D assumes strict linear Zeeman scaling with fixed coefficients from an external reference and treats the mismatch map as if it provides a robust parameter choice for any storage time. The manuscript's own observation of two ion classes in Section II means that a single ratio may not describe all participating ions, and the coefficients were not verified at the lower fields and elevated temperatures relevant to the proposed operating points. The reader's weakest_assumption identifies this same step, and I agree. I would keep the verdict CONDITIONAL rather than moving to ACCEPT or REJECT: the experimental efficiency can be trusted pending public raw data, but the proposed generalization should be treated as a simulation-based hypothesis until the Zeeman-ratio assumption is tested. I also note that the abstract's phrase 'without compromising the memory bandwidth' is not supported as a simultaneous efficiency-bandwidth claim; the conclusion correctly separates the 28.5% result at 30 MHz from the 5.0% result at 630 MHz.","tokens_in":9581,"tokens_out":12316,"duration_ms":110126,"concrete_test":"Re-measure the ground and excited Zeeman splittings spectroscopically, using hole and anti-hole positions as in Fig. 1(c), on the same Tm3+:YAG crystal at fields of approximately 100, 200, 370, 630, and 4500 G at T=3.5 K and at the elevated operating temperature. Then recompute the Fig. 5(c) mismatch map using the measured field-dependent Δg/Δe instead of the fixed 4.75 ratio from Ref. [26]. If the ratio deviates from 4.75 by more than about 1% anywhere in the proposed operating range, the claim that one can always choose a field to satisfy Eq. (3) fails; if it remains 4.75, the proposal survives this test.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing step is Section III.D, Eq. (3): the commensurate intrinsic pumping proposal assumes that Δe and Δg scale strictly linearly with magnetic field, with fixed μe=0.006 MHz/G and μg=0.0285 MHz/G taken from Ref. [26], so that Δg/Δe=4.75. The mismatch map in Fig. 5(c) and the statement that one can always choose a field to reach a given storage time depend on this ratio being constant over the proposed operating range, roughly 100–700 G, at 3.5 K and above. A few percent variation with field, temperature, or ion site would shift the hole and anti-hole positions relative to the AFC teeth and break the four conditions in Eq. (3). The manuscript itself reports two classes of ions in Section II, and the proposed scheme is only simulated, not experimentally demonstrated. Thus the claimed route to multi-GHz bandwidth with >30% efficiency is an extrapolation from unverified parameters. This concern does not undermine the measured 28.5±0.2% efficiency at 30 MHz, but it is the weakest point on which the paper's central generalization depends. Separately, the abstract's wording can be read as claiming simultaneous high efficiency and high bandwidth, whereas the paper reports 28.5% at 30 MHz and 5.0±0.3% at 630 MHz.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports an atomic frequency comb (AFC) memory in a Tm3+:YAG crystal at 3.5 K, using tapered/adibatic pumping and six-pass beam routing to achieve a single-photon-level storage efficiency of 28.5±0.2% at a 30 MHz bandwidth. The authors also demonstrate two-frequency-window storage separated by 300 MHz with efficiencies around 3.5–4.3%, and a 630 MHz bandwidth AFC with 5.0±0.3% efficiency. In addition, they propose a 'commensurate intrinsic pumping' scheme intended to make holes and anti-holes coincide with AFC peaks and valleys, with a simulation showing that field strength and AFC spacing can be chosen to approximately match the level splittings. The central claims are the record-high efficiency without a cavity or dilution refrigerator and the proposed route to broader-bandwidth AFC memories.","tokens_in":9898,"tokens_out":9021,"duration_ms":79920,"significance":"If the measured efficiency and the theoretical analysis are correct, the 28.5±0.2% result at 3.5 K in a free-space, cavity-free geometry is a notable advance for rare-earth-ion solid-state quantum memories, especially combined with the systematic optimization of pumping pulses and the demonstration of 630 MHz bandwidth. The commensurate intrinsic pumping proposal, if validated, could point toward AFC memories with gigahertz-scale bandwidth and higher efficiency. The paper is clearly written and the experimental setup is described in enough detail for reproduction of the apparatus. However, the central theory–experiment agreement is not independently checkable from the information given: the parameters d, F, and d0 are stated without estimation procedures or uncertainties, and the numerical implementation of Eq. (1) appears inconsistent with the quoted parameters. The forward-looking proposal also rests on an unverified linear Zeeman assumption for a crystal that exhibits two ion classes. These issues make the paper interesting but currently not fully substantiated.","major_comments":[{"comment":"The manuscript states that for Δ=6 MHz the finesse is F≈4.5, the background OD is d0≈0.4, and the theoretical storage efficiency is 30.4%. However, substituting these values together with the stated effective optical depth d≈12 (Section II) into Eq. (1) gives η=(12/4.5)² exp(-12/4.5) sinc²(π/4.5) exp(-0.4) ≈ 28.1%, not 30.4%. The authors should correct the numerical example or specify the exact values of d, F, and d0 used in the calculation; as written, the claimed agreement with the measured 28.5±0.2% is not reproducible.","section":"Section III.A, Eq. (1)"},{"comment":"The parameters d, F, and d0 are introduced without any description of how they are estimated or what uncertainties they carry. If they are extracted from the same tailored AFC spectrum that is used for the storage efficiency measurement, then the comparison with Eq. (1) is not an independent test of the model. Please state the measurement/estimation method for each quantity, report uncertainties, and propagate them into the theoretical efficiency. This is load-bearing for the central claim that the 28.5% result is consistent with the standard AFC model.","section":"Section III.A, Eq. (1)"},{"comment":"The commensurate intrinsic pumping proposal assumes that the Zeeman splittings scale strictly linearly with magnetic field with the single ratio Δg/Δe=4.75 from Ref. [26]. This ignores the fact that Section II reports two distinct ion classes in the same crystal, with excited-state splittings of 6 MHz and 27 MHz at 4500 G; the ratio is not 4.75 for the 6 MHz class. The mismatch simulation and the statement that 'we can always choose a field to reach a given storage time' are therefore not established for the actual sample. The authors should either provide experimental verification of the linear scaling and the relevant ratio for the class used in the proposal, or explicitly present the scheme as contingent on these assumptions and discuss how the second class affects the hole/anti-hole structure.","section":"Section III.D, Eq. (3) and Fig. 5(c)"}],"minor_comments":[{"comment":"The caption reads 'The splitting are µe=0.006 MHz/G and µe=0.0285 MHz/G [26]'; the second coefficient should be µg, not µe.","section":"Fig. 5(c) caption"},{"comment":"The sentence 'the ground state splitting is 4.75 times of the excited state spitting' applies only to the ion class with 27 MHz excited-state splitting. Please qualify this statement to avoid ambiguity given the two classes reported in Section II.","section":"Section III.D"},{"comment":"The phrase 'without compromising the memory bandwidth' in the abstract is ambiguous: the 28.5% efficiency is reported for a 30 MHz bandwidth, while the 630 MHz demonstration has an efficiency of 5.0±0.3%. Clarify in the abstract which bandwidth accompanies the 28.5% result.","section":"Abstract"},{"comment":"The relation between the 9 MHz hole–anti-hole separation and the choice of AFC spacing Δ=18 MHz follows from Eq. (2) but is not explained; a brief sentence on the factor of two (Δ/2) would help readers unfamiliar with intrinsic pumping.","section":"Section III.C"}],"recommendation":"major_revision","confidential_remarks":"The arithmetic inconsistency in Section III.A is the most serious issue: the quoted parameters do not produce the stated 30.4% theoretical efficiency. This is likely correctable, but the authors must report how d, F, and d0 are determined and with what uncertainties. The paper's forward-looking proposal in Section III.D is also under-supported by the existence of two ion classes; I would ask the authors to either add a simple spectral-hole-burning measurement at a few field values or to clearly delineate the proposal as conditional. The measured 28.5% result, if the analysis is corrected, appears plausible and worthwhile."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First, the core result. The 28.5±0.2% efficiency at 30 MHz AFC storage in a free-space Tm:YAG crystal at 3.5 K is credible and a genuine step forward over previous Tm:YAG demonstrations. It matches the standard AFC efficiency formula (30.4%) given the reported optical depth, finesse, and background absorption. The noise characterization (SNR ~270 at single-photon level) is careful. The 630 MHz multi-tone comb with 5.0±0.3% efficiency is also a useful demonstration, though efficiency drops as expected.\n\nThe commensurate intrinsic pumping scheme in Section III.D is the most speculative part. The idea of using the ratio Δg/Δe = 4.75 to satisfy the four conditions in Eq. (3) is interesting, but it is supported only by a simulation using μe and μg from Ref. [26], with no experimental demonstration and no discussion of how those coefficients vary with field, temperature, or ion site. The paper itself reports two classes of ions in Section II, so the assumption of a single rigid ratio is not obviously safe. The mismatch map in Fig. 5(c) shows that near-perfect matching can be found for chosen storage times, but that is a parameter search, not evidence the physical system actually behaves that way. The stress-test note is right to flag this; I would add that this does not undermine the measured 28.5% result, because that measurement is independent of the proposal.\n\nSoft spots besides that: (1) No public data, and the key parameters d, F, d0 are given without estimation procedures or error propagation. The agreement with Eq. (1) therefore cannot be independently checked. This is a common issue in this field but worth noting. (2) The abstract's phrase \"without compromising the memory bandwidth\" is misleading: the 28.5% is at 30 MHz, while the 630 MHz demonstration gives 5%. The paper itself is clear about this in the conclusions, but the abstract overstates. (3) The efficiency prediction uses parameters inferred from the prepared comb, so the agreement with theory is partly self-referential. That is a minor concern; the measured efficiency is the real claim.\n\nWho is this for: experimentalists working on rare-earth-ion quantum memories, particularly Tm:YAG and AFC protocols. They will find the 28.5% result useful and the commensurate pumping proposal worth testing. The paper deserves serious peer review: the measured result is solid enough to justify publication after revision, provided the data availability and abstract wording are addressed. I would recommend sending it to review.","headline":"Solid measured AFC efficiency in Tm:YAG at 3.5 K, but the broadband pumping proposal is a simulation-based extrapolation that needs experimental validation.","tokens_in":10435,"tokens_out":2442,"would_cite":true,"duration_ms":22994,"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":"The paper reports a Tm3+:YAG atomic-frequency-comb memory that stores single-photon-level pulses with 28.5±0.2% efficiency at 30 MHz bandwidth, in free space at 3.5 K, without a cavity or dilution refrigerator, and proposes a pumping…","keywords":["quantum memory","atomic frequency comb","spectral hole burning","Tm3+:YAG","broadband optical storage","intrinsic pumping","rare-earth ions","quantum repeaters"],"falsifier":"Measure the spectral hole and anti-hole positions of Tm3+:YAG as a function of magnetic field between 100 and 700 G and temperature between 1 and 4 K; if the ratio $\\Delta_g/\\Delta_e$ deviates from 4.75 by more than about 1%, or if the mismatch sum in Fig. 5(c) does not approach zero at the predicted field values (e.g. 630 G for 250 ns storage), the commensurate intrinsic pumping scheme fails. A more direct test is to attempt AFC storage at the predicted field and comb spacing: near-zero efficiency at the predicted operating point would disprove the claim that a field can always be chosen for a given storage time.","tokens_in":9395,"feed_emoji":"💎","tokens_out":6493,"duration_ms":52885,"temperature":0.7,"pith_summary":"This paper tries to show that Tm3+:YAG crystals can serve as practical broadband quantum memories without the need for cavities or dilution refrigerators. The authors demonstrate single-photon-level storage with efficiency $28.5 \\pm 0.2\\%$ at a memory bandwidth of 30 MHz in a free-space crystal at 3.5 K, close to the $30.4\\%$ predicted by the atomic-frequency-comb efficiency model. They also show storage in two frequency windows separated by 300 MHz, and a 630 MHz bandwidth comb with $5.0 \\pm 0.3\\%$ efficiency. The central forward-looking claim is a proposed 'commensurate intrinsic pumping' scheme that aligns spectral holes and anti-holes with comb peaks and valleys, which could extend memory bandwidth to the full inhomogeneous broadening, potentially exceeding 2 GHz with efficiency above 30%.","feed_headline":"28.5% storage efficiency in a no-cavity Tm:YAG quantum memory","feed_subtitle":"Free-space AFC at 3.5 K hits record efficiency, with a pumping scheme targeting gigahertz bandwidth.","key_machinery":"The load-bearing object is the atomic frequency comb (AFC): a periodic absorption profile created by spectral hole burning, in which absorbed photons rephase after a storage time $1/\\Delta$. Efficient combs are produced by adiabatic pumping with secant-hyperbolic amplitude and tangent-hyperbolic frequency chirp, and the efficiency is governed by Eq. (1), whose exponential and sinc factors make the finesse $F\\approx 4.5$ near-optimal. For broadband operation, the 'intrinsic pumping' scheme transfers atoms between hyperfine levels so that holes and anti-holes themselves form the comb; the proposed commensurate version requires the Zeeman splittings $\\Delta_e$ and $\\Delta_g$ and the AFC spacing $\\Delta$ to satisfy Eq. (3), i.e. all hole and anti-hole positions fall on the comb grid. The paper uses the measured splittings $\\mu_e = 0.006$ MHz/G and $\\mu_g = 0.0285$ MHz/G, with a ground/excited ratio of 4.75, to compute a mismatch map over magnetic field and storage time, showing that a field can be chosen for a given $\\Delta$.","core_discovery":"The central discovery is that careful spectral hole pumping in Tm3+:YAG yields record storage efficiency at elevated temperature without a cavity. Using six passes through a 0.1%-doped crystal to reach an effective optical density of about 12, and adiabatic secant-hyperbolic chirped pumping pulses to carve square atomic frequency combs, the authors store weak coherent pulses with measured efficiency $28.5 \\pm 0.2\\%$ for a comb period $\\Delta = 6$ MHz, corresponding to a 30 MHz bandwidth and a 167 ns storage time at a wait time of 5 ms; the signal-to-noise ratio is about 270. The efficiency matches Eq. (1), $\\eta = (d^2/F^2)\\exp(-d/F)\\mathrm{sinc}^2(\\pi/F)\\exp(-d_0)$, with finesse $F\\approx4.5$ and background absorption $d_0\\approx0.4$, predicting $30.4\\%$. The same approach, with an EOM and etalon, prepares two AFCs separated by 300 MHz, storing single photons at about 3.5% and 4.3% efficiency, and intrinsic pumping at 370 G creates a 630 MHz bandwidth AFC with $5.0\\pm0.3\\%$ efficiency. The paper's proposed commensurate intrinsic pumping uses the Zeeman splittings $\\Delta_g = 4.75\\,\\Delta_e$ and requires hole and anti-hole positions to coincide with AFC spacing under Eq. (3), with simulations showing a field can be chosen for any target storage time.","pith_inferences":["If the linear-Zeeman ratio $\\Delta_g/\\Delta_e = 4.75$ holds beyond the measured fields, the commensurate pumping map implies an engineering rule: pick the magnetic field first to satisfy Eq. (3) for a desired storage time, then set comb spacing; this turns spectral hole positions into a tunable resource rather than a fixed defect.","A direct extension would be to test commensurate intrinsic pumping in a host with longer ground-state lifetime, such as Eu3+:YSO, where the predicted bandwidth-efficiency tradeoff could be verified at gigahertz spacing without the 10 ms fast-decay limitation seen in Tm:YAG.","The efficiency model in Eq. (1) suggests that combining six-pass geometry with impedance-matched cavities would trade bandwidth for efficiency, so the finesse and pass number should be jointly optimized; the paper does not pursue this joint optimization."],"forward_implications":["A no-cavity, free-space Tm3+:YAG AFC memory can reach $28.5\\%$ efficiency at 30 MHz bandwidth at 3.5 K, close to the $30.4\\%$ theoretical ceiling set by Eq. (1).","Multi-frequency-window AFC storage with 300 MHz separation is feasible, enabling spectrally multiplexed memories and frequency-bin encoding of qubits.","Intrinsic pumping can produce a 630 MHz bandwidth AFC with $5.0\\%$ efficiency; the efficiency is currently limited by the short ground-state lifetime, not by the protocol.","If ground-state lifetimes are extended at lower temperatures, the same methods should give storage efficiencies near 40% for standard AFC and over 30% for intrinsic pumping at bandwidths up to 2 GHz.","The commensurate intrinsic pumping condition in Eq. (3) is a general recipe for non-Kramers rare-earth ions, so the bandwidth scaling should transfer to other hosts and dopants."],"supporting_citations":[{"why":"Provides the measured Zeeman splittings $\\mu_e = 0.006$ MHz/G and $\\mu_g = 0.0285$ MHz/G (ratio 4.75) that Eq. (3) and Fig. 5(c) rely on.","marker":"[26]"},{"why":"Demonstrated intrinsic pumping AFC in a Tm-doped crystal and the condition $\\Delta_g - \\Delta_e \\simeq \\Delta_{\\mathrm{AFC}}/2$, which the 630 MHz broadband experiment extends.","marker":"[18]"},{"why":"Defines the atomic frequency comb protocol and its rephasing mechanism used throughout the paper.","marker":"[27]"},{"why":"Gives the adiabatic pumping method with secant-hyperbolic amplitude and tangent-hyperbolic chirp, plus the square-comb efficiency expression Eq. (1).","marker":"[28]"},{"why":"Supplies the efficiency optimization treatment for AFC storage used to compute the theoretical $30.4\\%$ value.","marker":"[29]"},{"why":"Previous algorithmic optimization and multi-pass configuration for Tm:YAG that this work builds on to reach above 28% efficiency.","marker":"[23]"},{"why":"Provides the branching ratio of 0.25 and level-structure parameters that determine pumping dynamics in Tm3+:YAG.","marker":"[22]"}],"fun_headline_variants":["Tm:YAG quantum memory hits 28.5% efficiency without a cavity","Free-space Tm:YAG memory stores quantum light at 28.5% efficiency","28.5% storage efficiency in Tm:YAG via spectral hole pumping","No-cavity Tm:YAG quantum memory: 28.5% efficiency at 3.5 K","Tm:YAG quantum storage: 28.5% efficiency from pumped spectral holes"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proposed broadband pumping scheme assumes that the Zeeman splittings of Tm3+:YAG scale strictly linearly with magnetic field with the measured rates $\\mu_e = 0.006$ MHz/G and $\\mu_g = 0.0285$ MHz/G, so that $\\Delta_g/\\Delta_e = 4.75$ and Eq. (3) can be met by choosing field and comb spacing; if the ratio drifts with field, site, or temperature, the matching condition and the accompanying efficiency projections fail.","fun_headline_variants_meta":{"raw":{"variants":["Tm:YAG quantum memory hits 28.5% efficiency without a cavity","Free-space Tm:YAG memory stores quantum light at 28.5% efficiency","28.5% storage efficiency in Tm:YAG via spectral hole pumping","No-cavity Tm:YAG quantum memory: 28.5% efficiency at 3.5 K","Tm:YAG quantum storage: 28.5% efficiency from pumped spectral holes"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000412,"raw_usage":{"total_tokens":2155,"prompt_tokens":990,"completion_tokens":1165,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":606,"completion_tokens_details":{"reasoning_tokens":1051}},"tokens_in":606,"tokens_out":1165,"duration_ms":9571,"temperature":1.0,"reasoning_tokens":1051,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T14:08:47.077068+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the spectral hole and anti-hole positions of Tm3+:YAG as a function of magnetic field between 100 and 700 G and temperature between 1 and 4 K; if the ratio $\\Delta_g/\\Delta_e$ deviates from 4.75 by more than about 1%, or if the mismatch sum in Fig. 5(c) does not approach zero at the predicted field values (e.g. 630 G for 250 ns storage), the commensurate intrinsic pumping scheme fails. A more direct test is to attempt AFC storage at the predicted field and comb spacing: near-zero efficiency at the predicted operating point would disprove the claim that a field can always be chosen for a given storage time.","supporting_citations":[{"cited_title":"De Seze, A","cited_arxiv_id":null,"evidence_quote":"Provides the measured Zeeman splittings $\\mu_e = 0.006$ MHz/G and $\\mu_g = 0.0285$ MHz/G (ratio 4.75) that Eq. (3) and Fig. 5(c) rely on."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Demonstrated intrinsic pumping AFC in a Tm-doped crystal and the condition $\\Delta_g - \\Delta_e \\simeq \\Delta_{\\mathrm{AFC}}/2$, which the 630 MHz broadband experiment extends."},{"cited_title":"Afzelius, C","cited_arxiv_id":null,"evidence_quote":"Defines the atomic frequency comb protocol and its rephasing mechanism used throughout the paper."},{"cited_title":"Bonarota, J","cited_arxiv_id":null,"evidence_quote":"Gives the adiabatic pumping method with secant-hyperbolic amplitude and tangent-hyperbolic chirp, plus the square-comb efficiency expression Eq. (1)."},{"cited_title":"Jobez, N","cited_arxiv_id":null,"evidence_quote":"Supplies the efficiency optimization treatment for AFC storage used to compute the theoretical $30.4\\%$ value."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Previous algorithmic optimization and multi-pass configuration for Tm:YAG that this work builds on to reach above 28% efficiency."},{"cited_title":"Louchet, J","cited_arxiv_id":null,"evidence_quote":"Provides the branching ratio of 0.25 and level-structure parameters that determine pumping dynamics in Tm3+:YAG."}],"review_version":1}