{"id":"9146a164-903c-4082-9770-403822dd6a6e","arxiv_id":"1908.02483","paper_version":2,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":2,"one_line_summary":"In an exactly solvable one-dimensional Fermi gas, two bosonic impurities bind more tightly due to fermion-mediated attraction, and an effective model predicts binding even for repelling bosons.","lead":"Two bosonic impurities in a one-dimensional Fermi gas attract each other more strongly when the gas is present, forming in-medium bound states. The paper computes this exactly for a solvable model and builds an effective model to predict when repelling bosons will bind.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Exact BA benchmark would be strengthened by an independent ground-state check: the Newton branch from Eq. (6) is the only support for E < -c^2/2, and competing trimer states could in principle cross before c = -2.","rationale":"The reader's weakest assumption identifies the uncontrolled effective model in Sec. IV.B, which is indeed soft but does not threaten the exact Bethe-ansatz result. My concern targets the exact-result section itself: the c<0 branch followed by Newton continuation is not independently checked, and the thermodynamic extrapolation is the only bridge from finite-N BA data to the claimed E < -c^2/2. This is the most load-bearing assumption for the paper's strongest claim. However, I do not regard the concern as a demonstrated error: the BA equations are exact, the one-impurity test in Appendix D validates the numerical pipeline, and the trimer branch is expected to become competitive only near c ~ -2.2 because removing a fermion from the Fermi sea costs a chemical potential of order pi^2/2. The concrete DMRG check would settle the question cleanly. Since the concern is a verification gap rather than an identified inconsistency, the reader's ACCEPT verdict need not change.","tokens_in":18027,"tokens_out":25620,"duration_ms":313858,"concrete_test":"Run an independent ground-state calculation (DMRG or exact diagonalization in a momentum-space basis) for two bosons plus Nf = 5, 7, ..., 15 spinless fermions on a ring with the same zero-range interactions at c = -0.5, -1.0, and -1.5. Compare the resulting ground-state energies, after applying the same thermodynamic extrapolation, with the BA branch energies. If the independent energies lie below the BA branch for any c > -2, the central benchmark is invalid; if they match, the branch concern is retired.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central statement that the attractive Bethe-ansatz system has an in-medium pair energy E below the vacuum binding energy -c^2/2 (Fig. 2, Sec. III.A) rests on two numerical steps that are not independently verified. First, for c<0 the BA equations (6) are solved by Newton continuation from the c=0 Fermi-sea state (Appendix B). This assumes the followed branch is the true ground state for all c in (-2,0). For attractive 1D delta gases, competing bound branches exist; the authors themselves quote the two-boson plus one-fermion trimer energy -2c^2 and state that c -> -infinity is dominated by this 'fundamental' trimer. Nothing in Sec. III.A establishes that this trimer branch does not cross the followed branch already for c > -2, in which case the energies in Fig. 2 would be an excited-state branch, not the ground-state in-medium binding energy. Second, the thermodynamic limit is obtained by fitting eps(c)-eps(0) for N up to about 25 with the two forms in Appendix C; no independent method (DMRG, exact diagonalization, or larger-N BA) confirms the extrapolated eps_infinity. Since the gap below -c^2/2 is the central benchmark, this unverified branch/extrapolation is the most load-bearing assumption of the exact-result section.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies two bosonic impurities immersed in a one-dimensional spin-polarized Fermi gas with delta-function interactions. In the exactly solvable case m=M and cII=c, the authors solve the Bethe ansatz equations (6) for finite particle numbers, extrapolate to the thermodynamic limit, and compare the resulting in-medium binding energy E = eps_infinity - 2E with the vacuum two-boson binding energy -c^2/2. Their central result is that for attractive interactions c in (-2,0), E lies below -c^2/2, which they interpret as an attractive boson-boson interaction induced by the Fermi gas; for repulsive interactions c>0 they find no in-medium bound state. They then construct two effective two-body models, a zero-range potential matched to single-phonon exchange and a Born-Oppenheimer potential obtained from static impurities, benchmark them against the exact Bethe ansatz result, and use them to predict in-medium bound states in non-integrable systems with cII != c and/or M != m, including a critical line ccr_II ~ c^2/pi^2 for bosons that repel each other in vacuum.","tokens_in":18376,"tokens_out":18950,"duration_ms":221354,"significance":"If the central exact result holds, this paper provides a rare nonperturbative benchmark for fermion-mediated interactions between two impurities in one dimension and makes concrete, experimentally testable predictions for ultracold Bose-Fermi mixtures. The strengths of the paper are real: the Bethe-ansatz calculation is cross-checked against McGuire's single-impurity result in Appendix D, the thermodynamic limit is examined with two different extrapolation forms, and the effective-potential parameters come from independent single-phonon-exchange and Born-Oppenheimer calculations rather than being fitted to the target quantity. The main caveats are that the followed Bethe-ansatz branch is not proven to be the ground state over the whole attractive interval, and the non-integrable predictions rest on an uncontrolled effective Hamiltonian.","major_comments":[{"comment":"The identification of the followed Bethe-ansatz branch as the ground state is not established. The Newton continuation starts from the c=0 solution (Eq. (7)) and follows one branch, but the authors themselves note that for c -> -infinity a trimer branch with vacuum energy -2c^2 dominates. They do not compute the in-medium trimer branch or show that it lies above the followed branch for all c in (-2,0). If the trimer branch crossed before c=-2, the curves in Fig. 2 would be an excited-state branch rather than the ground-state in-medium binding energy, and the interpretation in terms of a two-boson bound state induced by the Fermi gas would need revision. Please provide an independent check for small Nf (e.g., exact diagonalization or a DMRG calculation on a lattice analog), or solve the Bethe-ansatz string equations for the trimer branch and compare energies over the full range c in (-2,0).","section":"Sec. III.A and Appendix B"},{"comment":"The thermodynamic-limit extrapolation is not independently verified. The energies for N up to about 25 are fitted with the two forms in Eqs. (C1) and (C2); for c=-1 these give eps_infinity = -5.176 and -5.125, respectively, a difference of about 0.05 in the total energy. Since E is obtained by subtracting two extrapolated quantities, the systematic error in E may be comparable to the small-c gap E + c^2/2 shown in Fig. 4 (bottom). Please either provide a consistency check with larger-N Bethe-ansatz solutions or an independent numerical method, and quantify the extrapolation uncertainty directly in E rather than only through the dot size in Fig. 2.","section":"Sec. III.A and Appendix C"},{"comment":"The predictions for non-integrable systems use the effective Hamiltonian heff with meff=M and with V taken from either the static-impurity Born-Oppenheimer energy or a zero-range term matched to single-phonon exchange. This is an uncontrolled approximation: the good agreement in the integrable case does not by itself validate the model when M != m or cII != c. In particular, the critical line ccr_II ~ c^2/pi^2 in Eq. (21) comes from the zero-range model, and the Born-Oppenheimer model gives a slightly different threshold. Please benchmark the effective model against an independent many-body calculation for at least one mass-imbalanced or interaction-asymmetric case, or clearly state in the abstract and conclusions that these non-integrable results are uncontrolled model predictions.","section":"Sec. IV.B and Eq. (20)"}],"minor_comments":[{"comment":"The symbol E is used both for the single-impurity energy gain and for the in-medium binding energy E = eps_infinity - 2E; this is confusing. Please rename one of the two quantities.","section":"Sec. III.A"},{"comment":"The solid black curve -2c^2 is the vacuum trimer energy, not the in-medium trimer energy. Since it lies below the Bethe-ansatz curve for c < -0.8, the figure should carry a caveat explaining that this comparison is not the in-medium trimer branch and does not by itself indicate a level crossing.","section":"Fig. 2"},{"comment":"For c<0 the square roots in Eq. (7) are imaginary; the text should state explicitly that k1, k2 and Lambda1, Lambda2 become complex and specify the branch choice used in the continuation.","section":"Eq. (7)"},{"comment":"There is a typo: 'desribed' should be 'described'. There are also minor typos in Section IV.A ('coulpings' should be 'couplings') and in Fig. 3's inset ('present' should be 'presents').","section":"Appendix D"},{"comment":"Some references are arXiv-only without journal information (e.g., Refs. [63] and [72]); please update them if they have appeared in print.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The central exact Bethe-ansatz result is likely correct and the paper is well written, but I would like to see the ground-state-branch and thermodynamic-extrapolation questions addressed before publication. The non-integrable predictions, while interesting, should be framed more cautiously unless benchmarked independently."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know before you read it. First, the genuinely new result is an exact Bethe-ansatz benchmark: two bosonic impurities in a 1D Fermi gas in the thermodynamic limit, in the integrable symmetric case, with the in-medium binding energy E = ε∞ − 2E_single. For attractive interactions they find E below the vacuum two-boson binding energy −c²/2, and they interpret the gap as fermion-mediated attraction. Second, the paper keeps the exact and model parts cleanly separated: the effective-model results are labeled as predictions, and the exact part is cross-checked where it matters.\n\nWhat is done well. The BA equations are the standard nested form for the SU(3) delta gas; the single-impurity limit is checked against McGuire; and the thermodynamic limit is obtained with two different extrapolation forms that agree to the third digit. The two effective potentials—zero-range matched to single-phonon exchange, and a Born-Oppenheimer potential from the static-impurity energy—are derived independently and then benchmarked, not tuned, against the exact curve; both reproduce the trend of the exact data. The mass-renormalization effect is checked and found marginal. For the non-integrable cases (mass imbalance, c_II ≠ c), the predictions, including the window where repelling bosons bind with c_cr ≈ c²/π², are presented as model results, with the caveats stated.\n\nSoft spots, in proportion. The branch worry is real: for c<0 the ground state is found by Newton continuation from the c=0 Fermi sea, and nothing independently rules out a trimer-like branch crossing inside (−2,0). The authors are aware of the competitor—they plot the −2c² trimer energy and restrict all claims to c ≳ −2—but they do not close the gap. A couple of DMRG or exact-diagonalization points in that window would settle it. I would ask for that in revision, not reject over it. The repulsive case is numerically checked only to about c=2, padded by the spin-chain limit for c→∞; adequate, but thin. No code or data is released, which makes Fig. 2 less useful as a benchmark than it could be.\n\nWho benefits: cold-atom experimentalists working on Bose-Fermi mixtures and induced interactions—the Li-Cs estimate is concrete—and people using 1D integrable models as reference points. This is a contained, solid paper, not a reshape-the-field one. I would send it to a serious referee; my expectation is a request for the branch check plus code/data, not a redo of the physics. Cite it if you work on induced interactions in one dimension.","headline":"Solid, honest Bethe-ansatz benchmark for two-boson in-medium binding in a 1D Fermi gas; the central attractive-case energy holds up, though the ground-state branch is not independently confirmed in the −2