{"id":"ac68efa7-6955-45ba-b5be-73e0d71299e1","arxiv_id":"1908.09696","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Strongly interacting fermionic superfluids carry dissipationless dc Josephson supercurrents across a thin barrier, and the measured critical current yields a phase-sensitive estimate of the pair condensate fraction across the BCS-BEC crossover.","lead":"Ultracold lithium atoms in a superfluid state were split by a thin laser barrier and pushed with a controlled current. Below a critical current the atoms kept flowing without any chemical potential difference, and the measured critical currents were used to infer how many of the atoms were paired into a condensate.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq. (S.5) is the load-bearing link: the factorized Josephson-current model from Ref. 40, not independently derived here, is used both to predict Ic and invert it into ⟨λ0⟩; at unitarity this makes the extracted condensate fraction as trustworthy as that untested factorization.","rationale":"I agree with the reader's conditional verdict. The experimental observation of a dc Josephson current, the zero-resistance plateau, and the sinusoidal current-phase relation for strong barriers are credible and internally consistent, so rejection is not warranted. The soft spot is the quantitative step from Ic to λ0: Eq. (S.5) is a self-cited, non-rederived factorization, and the same theoretical λ0 is used to generate the curves and to check the factorization in Eq. (S.10). The agreement with QMC and Luttinger-Ward λ0 values provides supporting evidence, but it does not isolate the model: an interaction-dependent prefactor error in Eq. (S.5) would be absorbed into the extracted λ0 and still permit agreement to the extent that the prefactor error is small. The concrete mean-field benchmark is the cleanest available test because it uses the same λ0(r) on both sides, isolating the factorization assumption itself. The Eckart-barrier approximation is an additional, smaller error source that can be bounded by recomputing with the exact Gaussian transmission. If the benchmark fails, the paper should be reduced to reporting Ic; if it passes, the central claim is much stronger. Since the reader already identified this assumption and the appropriate conditional posture, I recommend no change to the verdict.","tokens_in":19455,"tokens_out":22639,"duration_ms":250032,"concrete_test":"Run the published mean-field Josephson solvers of Refs. 36 and 37 for the same trap and a Gaussian barrier with V0/μ > 0.6 and w = 0.95 μm at (kFa)^{-1} = 0, and compare the numerically obtained Ic with Eq. (S.5) evaluated using the same mean-field λ0(r), μ(r), and the exact Gaussian transmission (Ref. 69). If the ratio differs from 1 by more than a few percent at unitarity, the factorization underlying Eq. (S.12) fails and the extracted ⟨λ0⟩ is not a model-free measurement. The same check should be repeated at (kFa)^{-1} = ±1 to confirm that the deviation does not grow off resonance.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claim — that the measured dc critical current yields the total condensate fraction via Eq. (S.12), with ⟨λ0⟩ = 0.47(7) at unitarity — rests on Eq. (1)/Eq. (S.5): ℏIc = ∫ d³r λ0(r)n(r)μ(r)|t(μ(r),V0)|/(4k(μ(r))Rx). This formula is neither rederived nor tested within the paper; it is imported from Ref. 40 by two of the same authors. For the extraction to be valid, the current must factor exactly into a bulk condensate-density prefactor and a single-pair transmission amplitude at all couplings, including the strongly correlated unitary regime where 'single-pair' transmission is not an obviously well-defined notion. The paper's numerical check of the factorized form Eq. (S.10) itself uses the theoretical λ0(r) of Ref. 45, so it validates the integration of a slowly-varying envelope but cannot detect an interaction-dependent error in the prefactor of Eq. (S.5). Fig. 3A's agreement is likewise a consistency test: the same model and the same λ0(μ) generate the theory curves and are then inverted to extract λ0. An interaction-dependent prefactor error A(kFa) entering Eq. (S.5) would be absorbed into the reported ⟨λ0⟩ and would not be visible in a fit to Ic. The Eckart-barrier replacement in Sect. S.5.B adds a second, smaller source of model bias: the admitted few-percent deviations of |t|² at ε/V0 > 0.5 are within the energy range probed by the V0/μ > 0.6 data. If the factorization fails at unitarity, the extracted ⟨λ0⟩ is not the condensate fraction but an effective parameter of the untested model.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports the observation of dc Josephson supercurrents in strongly interacting 6Li fermionic superfluids connected through a thin, tunable optical barrier. The authors measure the current-imbalance (I-Δμ) characteristic and the current-phase relation, observe a zero-resistance plateau below a critical current Ic, a sinusoidal current-phase relation for strong barriers, and a sign-reversal symmetry Is(φ) ≈ -Is(-φ). They use the RCSJ model to extract Ic and compare it with an analytic model for the critical current, Eq. (1) and Eq. (S.5), which is then inverted via Eq. (S.12) to obtain the total condensate fraction ⟨λ0⟩ across the BCS-BEC crossover, including ⟨λ0⟩ = 0.47(7) at unitarity. The central claim is that the measured dc Josephson current is controlled by the pair condensate density rather than the superfluid density, thereby providing a phase-sensitive determination of the superfluid order-parameter amplitude in strongly correlated superfluids.","tokens_in":19776,"tokens_out":4716,"duration_ms":47655,"significance":"If the central quantitative claim holds, this experiment provides the first phase-sensitive measurement of the pair condensate density in a strongly correlated fermionic superfluid, and the agreement with non-perturbative Luttinger-Ward and quantum Monte Carlo results would be a notable advance. The experimental work has clear strengths: careful calibration of the barrier height by two independent methods, direct extraction of the current-phase relation from matter-wave interferometry, observation of the zero-resistance plateau, a direct check of the sign symmetry of the supercurrent, and systematic mapping of Ic and the conductance G across the crossover. The RCSJ-based determination of Ic is model-free in the sense that only Ic and G are fit parameters. However, the quantitative condensate-fraction extraction rests on an analytic factorization formula from a companion paper by two of the same authors that is not rederived or independently tested here; this makes the reported ⟨λ0⟩ values less secure than the raw transport observations.","major_comments":[{"comment":"The load-bearing relation ℏIc = ∫ d3r λ0(r)n(r)μ(r)|t(μ(r),V0)|/(4k(μ(r))Rx) is imported from Ref. 40, a companion paper by two of the present authors, and is not rederived in this manuscript. The claim that the critical Josephson current factorizes exactly into a bulk condensate-density prefactor and a single-pair transmission amplitude is especially delicate at unitarity, where the notion of a single-pair transmission amplitude is not obviously well defined. Because Eq. (S.5) is the sole link between the measured Ic and the inferred ⟨λ0⟩, please provide a derivation or an independent numerical validation of this factorization at strong coupling, for example against a microscopic fermionic calculation of the junction current across the crossover.","section":"S.5, Eq. (S.5)"},{"comment":"The numerical check of the factorized form Eq. (S.10) uses the same theoretical λ0(r) from Ref. 45 that is later compared with the extracted ⟨λ0⟩ in Fig. 3B. This verifies only that the trap integral of a slowly varying envelope is well approximated by Eq. (S.10); it cannot detect an interaction-dependent error in the prefactor of Eq. (S.5). Since the inversion Eq. (S.12) divides the experimental current by a theory current computed with the same λ0, an unknown prefactor A((kFa)^-1) multiplying the integrand of Eq. (S.5) would be absorbed into the reported ⟨λ0⟩, and the agreement in Fig. 3A would not reveal it because the same model generates both the theory curves and the extracted quantity.","section":"S.5.A, Eqs. (S.10)-(S.12)"},{"comment":"The Eckart-barrier approximation is stated to deviate from the Gaussian-barrier transmission by a few percent for incident energies ε > 0.5V0. The data used for the ⟨λ0⟩ extraction are restricted to V0/μ > 0.6, so incident kinetic energies up to the local μ can exceed 0.5V0 in the barrier-height range selected. The resulting few-percent systematic in |t| propagates directly into ⟨λ0⟩ and does not appear to be included in the quoted uncertainty of 0.47(7); please quantify and add this model-bias contribution to the error budget.","section":"S.5.B, Fig. S9"}],"minor_comments":[{"comment":"The quantity V0' is used in the figure panels but is not defined in the caption; the main text only says it is 'the calculated barrier height for which Ic = |Iext|'. Please define it explicitly and state how it is computed for each barrier width and interaction strength.","section":"Fig. 2E-F"},{"comment":"The shaded regions in Fig. S7 are described as the standard confidence interval of the critical velocity obtained from imbalance measurements, but the underlying imbalance data and fits are not shown in the figure; either add them or point to the section where they appear.","section":"Fig. S7"},{"comment":"The statement that n(r) coincides with the superfluid pair density is an approximation; please state the expected size of the error from identifying total pair density with superfluid density at T/TF = 0.06, since this approximation enters the condensate-fraction normalization.","section":"S.5, Eq. (S.5)"},{"comment":"Reference [44] is listed as the Supplementary Materials, but the main text also refers to the Supplementary file implicitly; spell out 'see Supplementary Materials' in the main text rather than citing only '[44]'.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The central quantitative claim is hard to evaluate because the governing formula comes from a companion theory paper by two of the same authors. If the factorization in Eq. (S.5) is not valid at unitarity, the condensate-fraction numbers cannot be supported. The editor may wish to ensure that the review process for Ref. 40 specifically assesses the strong-coupling validity of the factorization, since the present manuscript inherits that assumption without rederiving it. The experimental observations themselves are sound and valuable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know two things about this paper before reading it. First, the core experimental observation is the real deal: a dc Josephson supercurrent in a strongly interacting fermionic superfluid, with a zero-resistance plateau, a sinusoidal current-phase relation for strong barriers, and a sign reversal consistent with time-reversal symmetry. That has not been seen before in degenerate atomic gases, and it is a milestone for quantum-gas transport. Second, the quantitative headline — the condensate fraction across the BCS-BEC crossover, with ⟨λ0⟩ = 0.47(7) at unitarity — is less solid than it looks, because it is extracted by inverting the same analytic model that generated the theory curves. The stress-test note is right about that, and it is the main soft spot.\n\nWhat the paper does well: the experimental methodology is careful. The I–Δμ characteristics, the RCSJ fits, the barrier-width collapse in Fig. 2E-F, and the direct measurement of the current-phase relation via matter-wave interference are all convincing. The critical current is essentially model-free in the sense that it comes from the sharp onset of a chemical potential difference, not from a fit to a many-body theory. The authors also give a fair comparison to existing theoretical and experimental numbers for the condensate fraction, and they acknowledge the mean-field discrepancy.\n\nWhere it gets shaky: Eq. (1)/Eq. (S.5), taken from Ref. 40 by two of the same authors, factorizes the critical current into a bulk condensate-density prefactor times a single-pair transmission amplitude. That factorization is physically motivated, but it is neither rederived nor independently tested in the strongly correlated unitary regime, where 'single-pair transmission' is not an obviously well-defined notion. The numerical check of the factorized form (Eq. S.10) uses the same theoretical λ0(r) from Ref. 45, so it validates the LDA integration of a slowly varying envelope but cannot catch an interaction-dependent error in the prefactor of Eq. (S.5). An error there would be absorbed into the reported ⟨λ0⟩. The Eckart-barrier replacement adds a smaller, admitted source of bias. The agreement with Luttinger-Ward and QMC in Fig. 3B is reassuring, but it is partly a consistency test, not an independent confirmation.\n\nThat said, the central experimental claim holds up. The dc Josephson effect is real, and the sinusoidal current-phase relation is a clean result. The condensate fraction number is worth reporting, but it should be framed as a model-dependent estimate rather than a direct measurement. The paper deserves a serious referee; I would send it out with the expectation that the authors soften the order-parameter claim, show a sensitivity analysis of Eq. (S.5) to alternative factorizations, and ideally validate ⟨λ0⟩ against an independent probe. For my own work, I would cite the experimental observation, not the condensate-fraction value.","headline":"A genuinely first dc Josephson supercurrent in a strongly interacting Fermi gas, with a clean sinusoidal current-phase relation; the condensate-fraction extraction is clever but rests on an untested factorization, so take that number as a model-dependent estimate.","tokens_in":20404,"tokens_out":1639,"would_cite":true,"duration_ms":19324,"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 claims that the dc Josephson critical current through a barrier between two strongly interacting Fermi superfluids measures the pair condensate density, and uses this to extract the condensate fraction across the BCS-BEC…","keywords":["dc Josephson effect","BCS-BEC crossover","condensate fraction","unitary Fermi gas","superfluid order parameter","critical current","strongly correlated fermions","weak link"],"falsifier":"Measure the critical current at unitarity in junctions whose barrier transmission is independently calibrated (for instance from the normal-state conductance), and compare the extracted $\\langle\\lambda_0\\rangle$ with a direct independent determination of the condensate fraction at the same temperature, such as rapid-ramp or momentum-resolved photoemission; a significant disagreement would show that the factorization is biased. A second decisive test is to track $I_c$ as the temperature approaches $T_c$: the condensate density vanishes at $T_c$ while the superfluid density does not, so the fate of $I_c$ near $T_c$ separates the two order parameters.","tokens_in":19188,"feed_emoji":"⚛️","tokens_out":7626,"duration_ms":64873,"temperature":0.7,"pith_summary":"This paper reports the observation of dc Josephson supercurrents between two strongly correlated fermionic superfluids and uses them as a phase-sensitive probe of the superfluid order parameter. The authors claim that the measured critical current $I_c$ is set by the density of condensed pairs $n_c$ — the order-parameter amplitude — rather than by the superfluid density, and that this holds through the BCS-BEC crossover all the way to unitarity. Working with a tunable optical-barrier junction in $^6$Li, they observe a zero-resistance plateau and a sinusoidal current-phase relation, extract $I_c$, and compare it with an analytic model with no free parameters. From that comparison they determine the total condensate fraction $\\langle\\lambda_0\\rangle$, finding $0.47(7)$ at unitarity, in agreement with non-perturbative many-body predictions. If correct, this makes dc Josephson transport a direct, essentially model-free window onto strongly correlated order parameters.","feed_headline":"Josephson supercurrents measure the pair condensate fraction","feed_subtitle":"At unitarity the extracted condensate fraction is 0.47(7), matching many-body theory.","key_machinery":"The load-bearing object is the factorization of the Josephson critical current density into a bulk thermodynamic prefactor $\\mu n_c/(2k)$ times the single-pair transmission amplitude $|t(\\mu)|$. This separates single-particle tunnelling from many-body physics and, extended via the local-density approximation to the trapped gas, produces an analytic prediction for $I_c$ with no free parameters once $\\mu(r)$ and $\\lambda_0(r)$ are taken from non-perturbative many-body theory. Experimentally, $I_c$ is extracted from the measured current-imbalance characteristics using the resistively-and-capacitively-shunted junction (RCSJ) model, a lump-element circuit with a capacitive and a resistive channel in parallel with the Josephson element. The inversion to $\\langle\\lambda_0\\rangle$ relies on the approximate factorization $I_c \\approx \\langle\\lambda_0\\rangle I_{c,\\mathrm{sup}}$, which the authors check numerically to within a few percent, and the barrier transmission $|t|$ is evaluated with an Eckart approximation to the Gaussian barrier.","core_discovery":"The paper's central claim is that a dc Josephson junction between superfluid Fermi gases measures the condensate density directly through $\\hbar j_c = \\mu n_c |t(\\mu)| / (2 k(\\mu))$, where $n_c = n\\lambda_0$ is the pair condensate density, $\\mu$ the pair chemical potential, and $|t(\\mu)|$ the single-pair barrier transmission amplitude. Extended to the harmonically trapped gas via the local-density approximation, this gives the total critical current $I_c$; the authors show that the measured $I_c$ across the BCS-BEC crossover is non-monotonic and peaked near unitarity, and that their model reproduces the data without free parameters. This implies that $I_c$ tracks $n_c$ rather than the superfluid density. Inverting the relation yields the total condensate fraction $\\langle\\lambda_0\\rangle$, which at unitarity comes out as $0.47(7)$, consistent with zero-temperature many-body calculations and clearly below the mean-field value around $0.7$.","pith_inferences":["If the factorization survives closer to $T_c$, $I_c$ could serve as a direct thermometer for the condensate fraction where other order-parameter probes become ambiguous; this is a testable extension, not a claim of the paper.","The same junction geometry applied to a two-dimensional Fermi gas could disentangle condensate fraction from superfluid density, since Berezinskii-Kosterlitz-Thouless physics changes their relationship.","Directly measuring the single-pair transmission $|t|$ (for instance from conductance in a regime where its connection to $|t|^2$ is known) would provide an independent test of Eq. (1) without relying on theoretical input for $n_c$.","For imbalanced or topological superfluids, the current-phase relation measured through the same weak link could expose unconventional order-parameter symmetries; the paper points toward such extensions but does not carry them out."],"forward_implications":["If $I_c$ is controlled by $n_c$, then Josephson critical currents give a phase-sensitive measurement of the order-parameter amplitude in any weakly linked superfluid, including systems where other probes are indirect.","The condensate fraction can be mapped across the BCS-BEC crossover from transport data alone, without rapid-ramp or photoemission calibrations.","The observed zero-resistance branch and sinusoidal current-phase relation for strong barriers confirm Josephson's original prediction in a strongly correlated fermionic superfluid.","With the measured underdamped regime ($\\beta_c \\sim 10^3$), ac driving should produce Shapiro resonances, enabling frequency-based transport diagnostics.","The critical current shows no detectable decrease at unitarity up to $T\\approx 0.1T_F$, indicating that the extracted condensate fraction is robust to small temperature changes."],"supporting_citations":[{"why":"Supplies the analytic model Eq. (1) expressing the critical current density as the product of the condensate density and the single-pair transmission amplitude.","marker":"[40]"},{"why":"Provides the non-perturbative chemical potential and condensate fraction used to evaluate the local-density model across the crossover.","marker":"[45]"},{"why":"Provides the quantum Monte Carlo condensate fraction used as a comparison for the extracted values.","marker":"[39]"},{"why":"Gives the weak-link theory for $I_c$ and the second-harmonic contribution $I_2$ included in the model.","marker":"[35]"},{"why":"Contains the supplementary derivation and numerical check of the factorization $I_c \\approx \\langle\\lambda_0\\rangle I_{c,\\mathrm{sup}}$.","marker":"[44]"},{"why":"Reports the earlier Josephson plasma-frequency measurement whose non-monotonic trend across the crossover the critical-current data parallel.","marker":"[31]"},{"why":"Provides the mean-field computation of $I_c$ across the crossover that serves as the baseline for the observed non-monotonic behavior.","marker":"[36]"}],"fun_headline_variants":["Supercurrents measure condensate fraction","Josephson supercurrents gauge condensate fraction","Condensate fraction from dc supercurrents","dc supercurrents reveal condensate fraction","Josephson supercurrents pin down condensate fraction"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the critical Josephson current factorizes exactly into the bulk condensate density times a single-pair barrier transmission amplitude, even at unitarity, and that the additional factorization $I_c \\approx \\langle\\lambda_0\\rangle I_{c,\\mathrm{sup}}$ holds; if either step fails, the extracted $\\langle\\lambda_0\\rangle$ is not the true condensate fraction.","fun_headline_variants_meta":{"raw":{"variants":["Supercurrents measure condensate fraction","Josephson supercurrents gauge condensate fraction","Condensate fraction from dc supercurrents","dc supercurrents reveal condensate fraction","Josephson supercurrents pin down condensate fraction"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001347,"raw_usage":{"total_tokens":5448,"prompt_tokens":895,"completion_tokens":4553,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":511,"completion_tokens_details":{"reasoning_tokens":4484}},"tokens_in":511,"tokens_out":4553,"duration_ms":36376,"temperature":1.0,"reasoning_tokens":4484,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:05:20.993657+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the critical current at unitarity in junctions whose barrier transmission is independently calibrated (for instance from the normal-state conductance), and compare the extracted $\\langle\\lambda_0\\rangle$ with a direct independent determination of the condensate fraction at the same temperature, such as rapid-ramp or momentum-resolved photoemission; a significant disagreement would show that the factorization is biased. A second decisive test is to track $I_c$ as the temperature approaches $T_c$: the condensate density vanishes at $T_c$ while the superfluid density does not, so the fate of $I_c$ near $T_c$ separates the two order parameters.","supporting_citations":[{"cited_title":"Zaccanti, W","cited_arxiv_id":null,"evidence_quote":"Supplies the analytic model Eq. (1) expressing the critical current density as the product of the condensate density and the single-pair transmission amplitude."},{"cited_title":"Haussmann, W","cited_arxiv_id":null,"evidence_quote":"Provides the non-perturbative chemical potential and condensate fraction used to evaluate the local-density model across the crossover."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the quantum Monte Carlo condensate fraction used as a comparison for the extracted values."},{"cited_title":"Meier, W","cited_arxiv_id":null,"evidence_quote":"Gives the weak-link theory for $I_c$ and the second-harmonic contribution $I_2$ included in the model."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Contains the supplementary derivation and numerical check of the factorization $I_c \\approx \\langle\\lambda_0\\rangle I_{c,\\mathrm{sup}}$."},{"cited_title":"Valtolina, et al., Science 350, 1505 (2015)","cited_arxiv_id":null,"evidence_quote":"Reports the earlier Josephson plasma-frequency measurement whose non-monotonic trend across the crossover the critical-current data parallel."},{"cited_title":"Spuntarelli, P","cited_arxiv_id":null,"evidence_quote":"Provides the mean-field computation of $I_c$ across the crossover that serves as the baseline for the observed non-monotonic behavior."}],"review_version":1}