{"id":"69fe2b18-9496-4d1e-ad1e-6b6faf781f11","arxiv_id":"2603.02361","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"Improved lithium potentials predict that 6Li–7Li Feshbach resonances are narrow (~0.01–0.1 G), closed-channel-dominated, and mostly triplet in character, unlike homonuclear Li2 resonances.","lead":"This paper refines the interaction potentials for pairs of lithium atoms and uses them to predict Feshbach resonances in the mixed 6Li–7Li system, finding them narrow, triplet-dominated, and closed-channel-dominated. This provides a theoretical foundation for making ultracold lithium molecules and for tuning interactions in quantum-gas experiments.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"6Li–7Li predictions rely on an untested arithmetic mean of homonuclear shift parameters; high-field resonance residuals already exceed experimental errors, so the quantitative catalog is not yet validated.","rationale":"The paper's strongest substantive advance is the homonuclear potential improvement: the reduced χ² drops from ~123 to 1.41, the last singlet level of 6Li2 is reproduced to ~0.01 MHz, and triplet levels to ~1 MHz, with external validation from Ref. [14]. Those results are well supported and I do not object to them. The headline heteronuclear claim, however, is conditional on how the 6Li–7Li potentials are constructed. The arithmetic mean of shift parameters is not derived from heteronuclear data or from a physical reduced-mass scaling; the self-identified limitation in Sec. IV makes this explicit. The high-field resonance residuals in Table IV provide a direct quantitative warning: predicted positions at 544 and 552 G miss the measured values by several gauss, well outside the quoted uncertainties. This does not undermine the qualitative conclusion that the lowest-channel 6Li–7Li resonances are narrow, closed-channel dominated, and triplet in character — that conclusion follows from the robust last-bound triplet level assignment and is supported by the closed-channel fraction analysis. But accepting the precise resonance catalog, and especially the higher-field positions, requires resolving the shift-mean assumption. The reader's verdict was already CONDITIONAL with the same weakest assumption, so my stress test does not move the verdict; it strengthens the conditionality and identifies a concrete refit that would settle the issue.","tokens_in":20802,"tokens_out":4609,"duration_ms":44566,"concrete_test":"Refit the 6Li–7Li shift parameters S_0^(6,7) and S_1^(6,7) as free parameters to the four measured (1,1) resonance positions of Ref. [16] (226.3, 246.0, 539.9, 548.6 G), holding the MLR curves and all homonuclear shift parameters fixed at Table I. If the best-fit values lie outside the intervals S_0^(6,7) ∈ [16.03, 16.80] and S_1^(6,7) ∈ [1.47, 1.70] — or, more directly, if the resulting B_0 residuals at 544 and 552 G remain larger than the combined experimental and theory uncertainties — then the arithmetic-mean prescription is falsified and Tables IV–V require recomputation. If the best-fit shifts fall inside these intervals and the residuals shrink, the concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central 6Li–7Li results are computed from V_shift,S^(6,7)(R) = S_S^(6,7)(R−R_e,S)^2 with S_S^(6,7) = (S_S^(6,6)+S_S^(7,7))/2 and uncertainty |S^(6,6)−S^(7,7)| (Sec. III B, Eq. 16). This arithmetic mean is an interpolation assumption, not a derived consequence of how beyond-Born-Oppenheimer corrections scale with reduced mass. It is load-bearing because every predicted resonance position, width, and channel fraction in Tables IV and V is obtained from these averaged potentials. The paper's own Sec. IV concedes that the mean 'may not faithfully represent the threshold-specific correction required for the 6Li−7Li system.' The consequence is already visible in Table IV: the predicted (1,1) poles at 544.23 G and 552.43 G differ from the measured 539.9 G and 548.6 G by 4.3 G and 3.8 G, far outside the quoted experimental uncertainties (~0.8–0.9 G) and the reported theory uncertainties (~0.57 G). Thus the quantitative resonance catalog for the headline heteronuclear system is not yet validated, even though the qualitative triplet/closed-channel character may be robust because the last-bound-state assignment is not sensitive to the averaging details.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper refines Li–Li interaction potentials by adding a quadratic inner-wall shift to previously published MLR potential curves and fitting four shift parameters to homonuclear 6Li2 and 7Li2 threshold data. The fitted potentials give a reduced chi-squared of 1.41 over 18 observables (compared with ~123 for Ref. [13]) and reproduce the last bound levels of 6Li2 from Ref. [14] to ~0.01 MHz (singlet) and ~1 MHz (triplet). The authors then construct 6Li–7Li potentials by taking the arithmetic mean of the homonuclear shift parameters and use those potentials in coupled-channel calculations to predict s-wave Feshbach resonances. In the lowest hyperfine channel they find narrow, strongly closed-channel-dominated, predominantly triplet resonances, and they catalog additional resonances in excited channels. Comparison with experiment shows that two of the four measured (1,1) resonances are reproduced within about 1 G, while the two near 544 G and 552 G are off by about 4 G.","tokens_in":21256,"tokens_out":8436,"duration_ms":80903,"significance":"The homonuclear fit is a genuine, carefully validated improvement: the reduced chi-squared is reduced by two orders of magnitude and the external check against Ref. [14] is about ten times better than Ref. [13]. The qualitative prediction that the lowest-channel 6Li–7Li resonances are narrow, closed-channel dominated, and predominantly triplet is physically interesting and likely robust, since it follows from the least-bound triplet state being near threshold. The paper is also transparent about the limitations of the mixed-isotope potential. However, the quantitative 6Li–7Li resonance catalog is not yet validated: the two high-field pole positions differ from experiment by several gauss, far outside the quoted uncertainties, and the excited-channel inelastic widths in Table V have a sign inconsistency. The work is a solid contribution to the homonuclear potentials and a promising but provisional treatment of the heteronuclear system.","major_comments":[{"comment":"The 6Li–7Li potentials are constructed by setting S_S^(6,7) = (S_S^(6,6)+S_S^(7,7))/2 and taking |S^(6,6)−S^(7,7)| as the uncertainty. This is an interpolation assumption, not a derived property of how beyond-Born-Oppenheimer corrections scale with reduced mass. It is load-bearing because every resonance position, width, and channel fraction in Tables IV and V is computed from these averaged potentials. Table IV shows the consequence: the predicted (1,1) poles at 544.23(57) G and 552.43(57) G differ from the measured 539.9(8) G and 548.6(9) G by 4.3 G and 3.8 G, respectively—several times the quoted experimental (~0.8–0.9 G) and theoretical (~0.57 G) uncertainties. The quoted theoretical uncertainties therefore cannot be complete unless they include the systematic error of the mean prescription. I request a sensitivity analysis varying S_S^(6,7) over its stated uncertainty, or better, a","section":"Section III B, Eq. (16) and Table IV"},{"comment":"The text preceding Table V states that the resonances in the higher hyperfine channels have Γ_B^inel > 0, but every row of Table V reports a negative Γ_B^inel (e.g., −3.322×10^-6 mG for the 252.38-G (2,1) resonance). A negative width is unphysical in Eq. (9) because Γ_B^inel is associated with the finite lifetime of the quasi-bound state. Unless there is an unexplained sign convention in the complex scattering-length fit, the extracted inelastic widths are internally inconsistent. Please correct the sign convention or the extracted values, and verify that the products a_res Γ_B^inel and (a_bg/a0)Δ remain consistent after the correction.","section":"Section III B, Table V"}],"minor_comments":[{"comment":"The last two rows of the 7Li–7Li block both read S_(7,7)_0; the second should very likely be S_(7,7)_1. Please fix the label.","section":"Table I"},{"comment":"There are typographical errors: 'Mangetically' in the Introduction, 'calcualtions' in the Table IV caption, 'discrepencies' in Section IV, and 'close-channel' in Section III D (should be 'closed-channel').","section":"Several places"},{"comment":"The shift term is defined twice, in Eq. (1) and again as Eq. (16). Please define it once and refer back to avoid redundancy.","section":"Equations (1) and (16)"},{"comment":"The definition of reduced chi-squared as χ^2_ν = χ^2/ν with ν = N−M is standard, but the notation 'χ^2_ν = χ^2_ν' in the text appears garbled. Please clarify.","section":"Section II A"},{"comment":"The text says five poles are identified, but Table IV lists only four and says the 214-G resonance is too narrow to be characterized. Please make the status of the 214.5-G pole explicit in the table caption or text.","section":"Table IV and surrounding text"}],"recommendation":"major_revision","confidential_remarks":"The paper has real value on the homonuclear side and the qualitative heteronuclear spin-character prediction is interesting. The main obstacle is that the quantitative 6Li–7Li predictions rest on an untested arithmetic-mean prescription and already show multi-gauss residuals; the negative inelastic widths in Table V are a further internal inconsistency. I think the paper is publishable after the authors either (i) perform a sensitivity/fit analysis of the mixed-isotope shift parameter and revise the uncertainty budget, or (ii) explicitly reframe the 6Li–7Li results as qualitative predictions pending direct heteronuclear data. I recommend major revision rather than rejection because the methodological core and the homonuclear validation are sound."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know two things. First, the homonuclear part is genuinely good: refitting the MLR potentials with the inner-wall shift brings the reduced chi-squared from about 123 to 1.41, and the external check against the Semczuk binding energies improves by roughly an order of magnitude over Julienne–Hutson. That is a real contribution. Second, the headline 6Li–7Li predictions are built on an arithmetic mean of the 6Li6Li and 7Li7Li shift parameters, and the paper openly says that mean may not faithfully represent the correction for the heteronuclear system. The stress-test note is accurate: the predicted (1,1) poles at 544.23 and 552.43 G miss the measured values by 4.3 and 3.8 G, far outside the quoted uncertainties. So the quantitative resonance catalog is not yet validated.\n\nWhat is new and useful: this is the first coupled-channel characterization of 6Li–7Li with BBO-corrected potentials, and the spin-character analysis — triplet, closed-channel-dominated, narrow — is a direct output of the wavefunctions and is likely robust, because the last-bound-state assignment (triplet v=10) is not sensitive to the averaging details. The paper also gives a thoughtful discussion of why the shift terms perturb intermediate vibrational levels, and it flags the need for a global fit including heteronuclear data. That is honest and helpful.\n\nSoft spots, in proportion. The main one is the mean-shift assumption; it is load-bearing for every number in Tables IV and V. The paper concedes this, but that does not make the predictions solid. The measured resonances it compares against come from a 2005 conference proceedings, which may have larger systematic errors than the quoted 0.8–0.9 G, but even taking that at face value, several-gauss discrepancies mean the catalog should be treated as indicative. A second, minor issue: the shift parameters are fitted only to homonuclear data, so heteronuclear predictions are extrapolations. The paper says this too.\n\nWho is this for? Anyone working on ultracold Li mixtures or STIRAP to Li2 molecules. The improved homonuclear potentials and the qualitative spin-character picture are worth having; the specific resonance positions and widths should be used with caution until either a global fit or new 6Li–7Li data settles the shift question.\n\nMy recommendation: send it to peer review. The homonuclear improvement is solid, the limiting assumption is clearly stated, and the paper makes testable predictions. A good referee will ask for a more systematic treatment of the BBO scaling or for a fit that includes heteronuclear data, but the work deserves referee time.","headline":"Improved homonuclear Li potentials are a real step forward, but the 6Li–7Li resonance catalog rests on an untested arithmetic mean of shift parameters and should be treated as a qualitative guide, not a validated prediction.","tokens_in":21686,"tokens_out":1459,"would_cite":true,"duration_ms":16582,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["34.20.Cf","34.50.Cx","67.85.-d"],"model":"deepseek-v4-flash","headline":"The paper predicts that all Feshbach resonances in the lowest hyperfine channel of 6Li-7Li are narrow (0.01–0.1 G), strongly closed-channel dominated, and predominantly triplet in electronic spin character, contrasting sharply with homonucl","keywords":["Feshbach resonances","lithium dimers","ultracold collisions","interaction potentials","coupled-channels scattering","scattering length","STIRAP","heteronuclear molecules"],"falsifier":"Measure the predicted 214 G resonance in the (1,1) channel of 6Li-7Li with sub-10-mG resolution; if it is absent or displaced by more than the stated uncertainty, the mixed-isotope shift averaging fails. Alternatively, measure the singlet fraction of the Feshbach molecule near 226 G via radio-frequency spectroscopy; a singlet fraction far from the near-zero predicted value would disprove the predominantly-triplet claim.","tokens_in":20717,"feed_emoji":"🧲","tokens_out":6756,"duration_ms":54967,"temperature":0.7,"pith_summary":"The paper builds improved interaction potentials for lithium dimers by adding small quadratic short-range shifts to spectroscopically accurate Morse/long-range curves for the singlet and triplet electronic states, then fitting the four shift parameters to measured binding energies, scattering lengths, and resonance positions of 6Li2 and 7Li2. The fit reduces the overall reduced chi-squared from about 123 to 1.41 and reproduces the last singlet bound level of 6Li2 to ~0.01 MHz. Using these potentials in coupled-channels calculations, the paper predicts that all Feshbach resonances in the lowest hyperfine channel of mixed 6Li-7Li are narrow (~0.01–0.1 G), strongly closed-channel dominated, and predominantly triplet in electronic spin character—unlike the broad, open-channel, singlet-dominated resonances of the homonuclear systems. This matters because the resonances provide a foundation for designing Raman optical-transfer pathways to produce ultracold Li2 molecules in deeply bound rovibrational levels of both potentials across all three isotopologues.","feed_headline":"Narrow triplet Feshbach resonances predicted for 6Li-7Li","feed_subtitle":"Refined lithium potentials reveal why mixed-isotope resonances are closed-channel and triplet, opening a route to ultracold Li2 molecules.","key_machinery":"The central object is the modified MLR potential: a spectroscopically accurate Morse/long-range analytic potential-energy curve for each of the X1Σ+ and a3Σ+ states, augmented by a quadratic inner-wall shift term V_shift(R) = S_S (R - R_e,S)^2 for R < R_e,S. The four shift parameters (singlet and triplet for each homonuclear isotopologue) are the only free parameters in a weighted least-squares fit to threshold observables; mixed-isotope 6Li-7Li potentials are constructed by arithmetic averaging of the homonuclear shifts, with the difference taken as uncertainty. Resonance classification relies on the dimensionless resonance-strength parameter s_res and on open/closed-channel fractions extra","core_discovery":"The central claim is that the near-threshold physics of 6Li-7Li is governed by the last bound vibrational level of the triplet a3Σ+ potential (v=10), not the singlet potential as in 6Li2 and 7Li2. Consequently, the Feshbach molecules in the (1,1) entrance channel remain predominantly triplet across the entire field range and are closed-channel dominated, with resonance-strength parameters s_res around 10^-5 to 10^-3 and widths of tens of milligauss. The paper validates the underlying potentials by reproducing the last singlet and triplet bound levels of 6Li2 to ~0.01 MHz and ~1 MHz, and by matching four of five predicted (1,1) resonances to measured positions within experimental uncertainty","pith_inferences":["A global fit that includes heteronuclear 6Li-7Li spectroscopy, rather than the arithmetic-mean shift prescription, would likely resolve the few-gauss residuals at ~550 G and remove the reliance on ad hoc inner-wall shifts.","The switch of the least-bound level from singlet to triplet with reduced mass may be a general feature of heteronuclear alkali mixtures, suggesting similarly narrow, triplet-dominated resonances in other mixed isotope combinations.","Because the 6Li-7Li resonance positions are extremely sensitive to the triplet short-range phase, precision measurements of these resonances could serve as a sensitive probe of the triplet potential, complementing spectroscopy.","One could test the predicted spin character directly via rf spectroscopy of the Feshbach molecule near a resonance, measuring the singlet fraction and checking the predicted near-zero values."],"forward_implications":["The 6Li-7Li (1,1) resonances are too narrow for broad magnetic tuning, but four of them match measured positions within ~1 G at ~240 G, providing a benchmark for the triplet interaction potential.","The predominantly triplet Feshbach molecules naturally favor STIRAP transfer to deeply bound triplet levels of a3Σ+; reaching singlet X1Σ+ levels requires an intermediate state with mixed singlet-triplet character.","The predicted extremely narrow 214 G resonance has not been observed and offers a direct, unambiguous test of the model and the mixed-isotope shift assumption.","The broad resonance in the (6,1) channel near 525 G is too short-lived (lifetime ~2 μs) for coherent molecule formation, despite its large width.","The improved homonuclear potentials reduce the reduced chi-squared from ~123 to 1.41 and reproduce recent high-precision bound-state data, so they supersede earlier potentials for threshold scattering calculations."],"fun_headline_variants":["Triplet Feshbach resonances in 6Li-7Li: narrow and closed-channel","Improved potentials predict narrow triplet Feshbach resonances in 6Li-7Li","Why 6Li-7Li Feshbach resonances are triplet and narrow","Narrow triplet resonances in 6Li-7Li enable ultracold Li2"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The mixed-isotope 6Li-7Li potentials are obtained by averaging the fitted short-range shifts of the two homonuclear systems, and if the true 6Li-7Li short-range correction differs from this arithmetic mean—as the paper itself cautions—the predicted resonance positions, widths, and spin characters would shift.","fun_headline_variants_meta":{"raw":{"variants":["Triplet Feshbach resonances in 6Li-7Li: narrow and closed-channel","Improved potentials predict narrow triplet Feshbach resonances in 6Li-7Li","Why 6Li-7Li Feshbach resonances are triplet and narrow","Narrow triplet resonances in 6Li-7Li enable ultracold Li2"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000644,"raw_usage":{"total_tokens":2866,"prompt_tokens":882,"completion_tokens":1984,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":626,"completion_tokens_details":{"reasoning_tokens":1903}},"tokens_in":626,"tokens_out":1984,"duration_ms":13867,"temperature":1.0,"reasoning_tokens":1903,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T19:22:11.780925+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the predicted 214 G resonance in the (1,1) channel of 6Li-7Li with sub-10-mG resolution; if it is absent or displaced by more than the stated uncertainty, the mixed-isotope shift averaging fails. Alternatively, measure the singlet fraction of the Feshbach molecule near 226 G via radio-frequency spectroscopy; a singlet fraction far from the near-zero predicted value would disprove the predominantly-triplet claim.","supporting_citations":[],"review_version":1}