{"id":"fc751f0f-c1d6-4ad3-818a-7c341510657f","arxiv_id":"1909.02356","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"high","formal_verification":"none","parameter_count":3,"one_line_summary":"A lattice QCD calculation finds strong B B* attraction in the Zb channel and, under an assumed potential, predicts a near-threshold virtual state and a deep bound state.","lead":"Lattice QCD with static bottom quarks shows a strong attractive interaction between B and B* mesons in the Zb tetraquark channel at short separations. Solving a Schrodinger equation with the extracted potential yields a near-threshold virtual bound state that may match Zb(10610), plus a deep bound state that could be searched in Belle data.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Pole predictions rest on an unconstrained extrapolation of V(r) below the lattice cutoff; a refit with alternative short-range forms or with r/a=1 excluded would test whether they survive.","rationale":"The paper's direct lattice result, namely that the red-cross eigenstate lies significantly below m_B + m_B* at short separations, is the strongest and most defensible part of the work. It does not depend on Eq. (5.1), and the exclusivity of the B\\bar B^* component is supported by operator overlaps below 0.02, although the operator basis is not complete. The load-bearing weak point is the step from these four energy shifts to pole positions. The fit has two parameters and four points, so it can represent the data, but p = 3/2 is not derived, and the behavior of the potential between r = 0 and r = a is entirely outside the simulation. The virtual-state result -32^{+29}_{-5} MeV is already consistent with a much deeper or vanishing pole, and the deep state at -403 ± 70 MeV is a striking prediction that has not been experimentally seen. A controlled refit with alternative ansätze, or with the coarsest point removed, would settle whether the pole structure is a consequence of the data or of the chosen functional form. This concern does not undermine the attraction claim, so the appropriate verdict remains CONDITIONAL; the paper should present such a stability test and publish the numerical potential before the pole claims are treated as quantitative.","tokens_in":6561,"tokens_out":11212,"duration_ms":133413,"concrete_test":"Refit the four V(r) values in Fig. 3 (r/a = 1..4, using the actual correlated data and errors where available) under three variants: (i) p = 1, (ii) p = 2, and (iii) a piecewise potential equal to V(a) for r < a with the same exponential tail above a. For each fitted A and d, recompute the S-wave poles of the Schrödinger equation with the physical reduced mass. Also rerun the published p = 3/2 fit with the r/a = 1 point excluded. If the near-threshold pole moves by more than the quoted +29/-5 MeV range, or if the deep pole moves by more than 70 MeV or disappears, the pole predictions are not stable under the uncontrolled short-range extrapolation and should be reported only as illustrative.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Even granting that the red-cross eigenstate is dominated by B\\bar B^* as the normalized overlaps in footnote 2 indicate, the two quantitative pole predictions are not lattice observables: they are outputs of a Schrödinger equation whose potential is V(r) = -A exp(-(r/d)^(3/2)) (Eq. 5.1). The form is chosen by hand with p = 3/2, and A and d are fitted to only four lattice points, r/a = 1..4. Lattice data constrain the potential only for r >= a ≈ 0.124 fm. Below a, the ansatz supplies an unmeasured extrapolation, and it is precisely this short-range part, together with the uncertain point at r/a = 1 which can carry O(a) discretization error, that controls whether and where the near-threshold virtual pole and the deep bound pole appear. A different but equally data-compatible treatment below a, for example capping the potential at V(a), or using p = 1 or p = 2, can move the virtual pole by more than its quoted +29/-5 MeV uncertainty and can shift or remove the -403 MeV state. The paper explicitly defers 'more physically motivated forms' to a forthcoming publication (Sec. 5), so the advertised poles are currently not robust. The direct attraction signal survives, but the Zb-related pole claims do not.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This proceedings paper reports a preliminary N_f=2 lattice QCD study of the \\bar b\\bar b d\\bar u system with static b quarks and heavy-quark spin S_heavy=1. The authors compute eigen-energies E_n(r) as a function of the b-\\bar b separation using six interpolating operators and GEVP; they identify an eigenstate dominated by B\\bar B^*. Its energy lies significantly below m_B+m_{B*} for r between 0.1 and 0.4 fm, which is interpreted as a sizable attractive potential. Assuming this state is exclusively B\\bar B^*, they extract V(r)=E(r)-m_B-m_{B*}, fit it to V(r)=-A exp(-(r/d)^{3/2}) using four lattice separations, and solve the non-relativistic Schrödinger equation with physical B meson masses. They find an s-wave virtual bound state 32^{+29}_{-5} MeV below threshold and a deep bound state 403±70 MeV below threshold. The paper emphasizes that the attraction result is the robust conclusion and that the pole positions rest on additional simplifying assumptions.","tokens_in":6841,"tokens_out":5576,"duration_ms":56234,"significance":"The direct observation of strong attraction in the static limit is the paper's main strength and is genuinely valuable: it is visible in the raw eigen-energies and does not depend on the potential fit. The paper is also explicitly worded about its approximations and labels the study as preliminary. If the short-range potential were constrained reliably, the Schrödinger-equation step would be a promising route to connect lattice potentials to Z_b phenomenology. However, the quantitative pole predictions are not yet robust lattice observables: they depend on an unmeasured extrapolation of V(r) below the lattice cutoff and on the exclusive B\\bar B^* assumption. In its current form the paper therefore supports the qualitative conclusion, while the specific pole masses require additional robustness checks before they can be considered established.","major_comments":[{"comment":"The fit of the potential uses only lattice separations r/a = 1..4 and fixes the exponent p = 3/2 by hand. Since the lattice data constrain V(r) only for r >= a ~ 0.124 fm, the bound-state poles obtained from the Schrödinger equation are controlled by an unmeasured extrapolation below a. The quoted uncertainties on W_B reflect only the statistical fit errors of A and d, not the systematic choice of the functional form. A sensitivity test with alternative forms (e.g., p=1, p=2, or a potential capped at V(a) for r<a) is necessary; without it, the claims of a virtual bound state at -32^{+29}_{-5} MeV and a deep bound state at -403±70 MeV are not robust. Please add such tests or, if they are not available, explicitly demote these numbers to illustrative values.","section":"Section 5, Eq. (5.1)"},{"comment":"The extraction V(r)=E(r)-m_B-m_{B*} assumes that the red-cross eigenstate is exclusively B\\bar B^*. The normalized overlaps quoted in footnote 2 support dominance, but they do not exclude small admixtures of \\Upsilon\\pi or \\Upsilon b1 that could shift the eigen-energy and hence V(r). Since the pole positions in the single-channel Schrödinger equation inherit this assumption, the paper should either quantify the effect of possible mixing or state clearly that the pole prediction assumes exactly zero mixing; the current text mentions the assumption but does not assess its impact.","section":"Section 5, Fig. 2 and footnote 2"},{"comment":"The deep bound state at -403±70 MeV is presented as a 'surprising' finding and a potentially falsifiable prediction. Given that it is driven entirely by the extrapolated short-range part of the ansatz (5.1), and that the authors state more physical potentials will appear in a forthcoming publication, presenting this number with a statistical-only uncertainty of ±70 MeV is misleading. At minimum the systematic uncertainty from the choice of V(r) below a should be included; otherwise the deep state should be described as a model-dependent consequence of the ansatz rather than a lattice prediction.","section":"Section 5, paragraph containing W_B = -403±70 MeV"}],"minor_comments":[{"comment":"The sentence beginning 'The eigenstate dominated by B\\bar B^*...' contains the typo 'but has it has'; it should read 'but it has'.","section":"Section 4, second paragraph"},{"comment":"'cmf momenta' should be written as 'center-of-mass momenta' (c.m.).","section":"Section 5, paragraph on scattering matrix"},{"comment":"The units of A and d are not specified; they should be stated explicitly (e.g., A in GeV and d in fm) to make the fit parameters unambiguous.","section":"Eq. (5.1)"},{"comment":"The relation to the earlier studies [4,5] could be stated more clearly; the text says the only preliminary study was reported in [4,5], but it is not obvious to the reader which elements are new in the present work beyond the inclusion of \\Upsilon\\pi(\\vec p \\neq 0) operators.","section":"Introduction and Section 5"},{"comment":"The definition of the normalized overlap is terse; one explanatory sentence defining \\tilde Z^i_n and its role in identifying the dominant Fock component would improve readability.","section":"Footnote 2"}],"recommendation":"major_revision","confidential_remarks":"This is a clearly worded preliminary proceedings contribution. The main reason for major revision rather than rejection is that the qualitative attraction signal is robust and the quantitative pole claims can in principle be made defensible with a systematic sensitivity study. I would not recommend acceptance in current form because the abstract and Section 5 present the pole positions as results despite the unconstrained short-distance extrapolation."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The direct result in this proceedings paper is solid: the B B* dominated eigenstate sits well below m_B + m_B* for r between 0.1 and 0.4 fm, which is clear evidence for sizable attraction. That is genuinely new, as far as I know, because earlier static-quark studies did not include nonzero-pion-momentum Upsilon-pi Fock states. The inclusion of those operators is a real step forward, and the extracted potential is a reasonable first pass.\n\nI also want to say what is not a problem: the analysis is not circular. The potential is extracted from lattice eigen-energies, the fit parameters A and d are determined from that potential alone, and the Schrodinger equation is solved afterwards. The Belle data enter only as an interpretive comparison, not as input. The authors are appropriately caveated throughout, and they explicitly flag the small-r potential as provisional.\n\nThe soft spot is exactly the one the stress-test identifies. Everything from Section 5 onward rests on V(r) = -A exp(-(r/d)^(3/2)), with the exponent p=3/2 chosen by hand and A,d fitted to only four lattice separations, r/a = 1..4. The lattice constrains the potential only for r >= a ~ 0.124 fm, and the r/a=1 point can carry O(a) discretization error, but the exponential form supplied below the cutoff is what controls both the shallow virtual pole and especially the deep -403 MeV state. A different but data-compatible treatment below a — capping V at V(a), using p=1 or p=2, or dropping the r/a=1 point — could move or remove those poles. The paper itself says more physically motivated forms are coming in a full publication, which is the right tone but also an admission that the advertised poles are not lattice observables.\n\nThe single-channel assumption, that the red-cross eigenstate is exclusively B B*, is supported by the normalized overlaps in footnote 2, but it remains an approximation; a coupled B B* / Upsilon-pi analysis would be needed to make the poles credible. The Belle comparison in Fig. 4 is suggestive rather than quantitative.\n\nWho is this for: lattice and phenomenological people working on heavy-quark exotics and the Born-Oppenheimer approach to tetraquarks. It is a useful method note with a citable robust attraction result, and it deserves a serious referee despite the fragility of the pole claims. I would send it to peer review, expecting the pole section to become much more hedged or replaced by the improved analysis in the full paper.","headline":"The lattice attraction between B and B* is the robust, citable result; the Zb virtual-state and deep-bound-state poles are fragile outputs of a hand-picked potential extrapolation and should not be quoted as lattice predictions.","tokens_in":7384,"tokens_out":2920,"would_cite":true,"duration_ms":34775,"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":"A lattice QCD computation shows that B and B* attract each other in the Zb tetraquark channel, and with a fitted single-channel potential the two mesons form a near-threshold virtual state about 32 MeV below the BB* threshold.","keywords":["lattice QCD","tetraquark","Zb(10610)","Zb(10650)","Born-Oppenheimer approximation","static heavy quarks","B B* interaction","exotic hadrons"],"falsifier":"Compute the same correlation matrix on a finer lattice ($a\\approx0.05$ fm) so that $r=0.1$ fm is resolved by several lattice points, and check whether the $B\\bar B^*$-dominated eigenstate remains tens of MeV below $m_B+m_{B^*}$ at $r\\approx0.2$ fm; also measure its overlaps with the $\\Upsilon\\pi$ operators and verify they stay below about 2%. If the energy shift shrinks, or the overlap to $\\Upsilon\\pi$ grows, the single-channel potential and the two derived poles are artifacts of the assumed form.","tokens_in":6291,"feed_emoji":"⚛️","tokens_out":9606,"duration_ms":89145,"temperature":0.7,"pith_summary":"This paper asks whether the exotic Zb tetraquarks, quark content $\\bar b\\bar b\\bar u d$, can be understood as $B\\bar B^*$ systems held together by a strong attraction. It reports a preliminary lattice QCD calculation in which the heavy $b$ quarks are held fixed at a separation $r$ and the energies of the light $u/d$ quarks and gluons are computed. The key result is that the $B\\bar B^*$-dominated eigenstate lies substantially below $m_B + m_{B^*}$ for $r$ between 0.1 and 0.4 fm, a signal of sizable attraction. Treating that eigenstate as the pure $B\\bar B^*$ channel yields a potential, and solving the non-relativistic Schrödinger equation with a fitted potential produces a virtual bound state about 32 MeV below threshold, a plausible relative of the experimentally observed $Z_b(10610)$, plus a deep bound state about 403 MeV below threshold. The near-threshold result matters because it offers a first-principles route to the $Z_b$ states, while the deep state is a sharp prediction that can be searched for in the $\\Upsilon(1S)\\pi$ invariant-mass distribution.","feed_headline":"Zb tetraquark channel shows strong B–B* attraction","feed_subtitle":"Static-quark lattice data put a virtual pole 32 MeV below threshold, with a deep state 403 MeV down as a sharper test.","key_machinery":"The load-bearing machinery is the Born-Oppenheimer factorization of the $\\bar b\\bar b\\bar u d$ system: the heavy $b$ and $\\bar b$ are static at separation $r$, and the light $u/d$ quarks and gluons form eigenstates $E_n(r)$. The relevant eigenstates are identified from a $6\\times6$ correlation matrix of operators resembling $B\\bar B^*$, $\\Upsilon\\pi(\\vec p=0)$, $\\Upsilon\\pi(|\\vec p|=2\\pi/L)$, $\\Upsilon\\pi(|\\vec p|=4\\pi/L)$, a derivative-type $B\\bar B^*$, and $\\Upsilon b_1$, extracted with the full-distillation method and a variational (GEVP) analysis. The energy $E_{B\\bar B^*}(r)$ supplies $V(r)=E(r)-m_B-m_{B^*}$; the second step promotes the heavy quarks to finite mass and solves the non-relativistic Schrödinger equation to find scattering-matrix poles. The two assumptions carrying the argument are that the $B\\bar B^*$-dominated eigenstate has no other Fock components, and that the potential below the lattice spacing is described by the exponential form with exponent $p=3/2$.","core_discovery":"The paper's central result is the $r$-dependent energy spectrum of the $\\bar b\\bar b\\bar u d$ system with $I=1$, $S_{\\rm heavy}=1$, and conserved product $C\\cdot P=-1$, computed on a 2 fm lattice with two dynamical quark flavors at $m_\\pi\\simeq 266$ MeV. In this spectrum the state whose overlap is dominated by the $B\\bar B^*$ operator (the red-cross eigenstate in the figures) has energy significantly below the non-interacting $B\\bar B^*$ threshold for $r\\in[0.1,0.4]$ fm, and approaches the threshold at larger $r$. Under the stated assumption that this eigenstate consists exclusively of $B\\bar B^*$, the paper extracts the potential $V(r)=E(r)-m_B-m_{B^*}$ and fits it to $V(r)=-A e^{-(r/d)^{3/2}}$ with $A=0.99(5)$ and $d=1.84(10)$. Solving the non-relativistic Schrödinger equation with the measured $B$ and $B^*$ masses then gives an s-wave virtual bound-state pole at $-32^{+29}_{-5}$ MeV below threshold and a deep s-wave bound state at $-403\\pm70$ MeV below threshold. The paper presents the first as a plausible lattice description of $Z_b(10610)$, noting the similarity between the predicted $B\\bar B^*$ rate peak and the observed rate, and the second as a surprising prediction that has not yet been seen experimentally.","pith_inferences":["A finer lattice and larger volume would allow the potential to be measured at $r\\approx0.1$ fm rather than extrapolated; I would expect the near-threshold virtual pole to shift less than the deep pole, since the deep pole is controlled mainly by the unmeasured $r\\to0$ region.","The present result fixes $S_{\\rm heavy}=1$; the physical $Z_b$ states could mix with $S_{\\rm heavy}=0$, so a complete account would likely require a potential matrix spanning both heavy-spin sectors and the $\\Upsilon\\pi$ channel.","A high-statistics search in $\\Upsilon(1S)\\pi^+$ invariant mass is a clean discriminator: a broad structure roughly 400 MeV below the $B\\bar B^*$ threshold would be evidence for the deep state, while a smooth background would disfavor it."],"forward_implications":["A near-threshold enhancement in the $B\\bar B^*$ scattering rate is predicted, with a shape resembling the observed $Z_b(10610)$ peak, making the virtual pole directly testable in measured invariant-mass distributions.","The deep bound state at roughly 400 MeV below the $B\\bar B^*$ threshold is a specific prediction that can be searched for in $\\Upsilon(1S)\\pi^+$ invariant-mass spectra; current data are not flat but do not resolve it.","Because the $B\\bar B^*$ eigenstate approaches the non-interacting energy for $r\\gtrsim0.5$ fm, the attraction is short-range; accurate large-$r$ data would be needed to determine whether one-pion exchange plays any role.","The pole positions depend on the unknown short-distance behavior of $V(r)$; the fitted form $V(r)=-A e^{-(r/d)^{3/2}}$ with $p=3/2$ is an assumption, and a better-motivated short-range potential is the stated next step.","The lower $\\Upsilon\\pi$ and $\\Upsilon b_1$ eigenstates show no statistically significant energy shifts, so within current precision the $\\Upsilon\\pi$ interaction is unresolved and does not visibly mix into the $B\\bar B^*$ state."],"supporting_citations":[{"why":"Discovery of the Zb(10610) and Zb(10650) states and their quantum numbers; the target of the lattice calculation.","marker":"[1]"},{"why":"Measured B B* production rate from Zb decay, used to compare with the predicted near-threshold peak from the virtual pole.","marker":"[3]"},{"why":"Earlier static-quark lattice study of the same system that set up the Fock-component decomposition and operator basis.","marker":"[4]"},{"why":"Full-distillation method used to compute the light-quark correlation matrices from which eigen-energies are extracted.","marker":"[8]"},{"why":"Experimental-data analysis locating a virtual bound-state pole in B B* when other channels are switched off, supporting the interpretation of the near-threshold pole.","marker":"[9]"}],"fun_headline_variants":["Lattice QCD: B–B* attraction in Zb tetraquark channel","Zb tetraquark: lattice predicts bound state near threshold","B–B* potential from lattice: two bound states predicted","Static-quark lattice study finds B–B* bound states"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"Everything after the spectrum—the potential, the virtual pole, and the deep bound state—depends on the belief that the red-cross eigenstate is purely $B\\bar B^*$, so its energy can be read directly as the $B\\bar B^*$ potential, and on the guessed form of that potential for distances smaller than the lattice spacing (about 0.12 fm).","fun_headline_variants_meta":{"raw":{"variants":["Lattice QCD: B–B* attraction in Zb tetraquark channel","Zb tetraquark: lattice predicts bound state near threshold","B–B* potential from lattice: two bound states predicted","Static-quark lattice study finds B–B* bound states"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000547,"raw_usage":{"total_tokens":2704,"prompt_tokens":1127,"completion_tokens":1577,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":743,"completion_tokens_details":{"reasoning_tokens":1501}},"tokens_in":743,"tokens_out":1577,"duration_ms":10340,"temperature":1.0,"reasoning_tokens":1501,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T04:52:34.088402+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the same correlation matrix on a finer lattice ($a\\approx0.05$ fm) so that $r=0.1$ fm is resolved by several lattice points, and check whether the $B\\bar B^*$-dominated eigenstate remains tens of MeV below $m_B+m_{B^*}$ at $r\\approx0.2$ fm; also measure its overlaps with the $\\Upsilon\\pi$ operators and verify they stay below about 2%. If the energy shift shrinks, or the overlap to $\\Upsilon\\pi$ grows, the single-channel potential and the two derived poles are artifacts of the assumed form.","supporting_citations":[{"cited_title":"Investigation of $B\\bar B$ four-quark systems using lattice QCD","cited_arxiv_id":"1602.07621","evidence_quote":"Earlier static-quark lattice study of the same system that set up the Fock-component decomposition and operator basis."}],"review_version":1}