{"id":"6323271f-4df4-4a4f-aa53-c94b19d8bb54","arxiv_id":"2412.03495","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"In a discrete lattice, noninteracting fermions tunnel symmetrically through an asymmetric barrier, while interactions make tunneling strongly direction-dependent and spin-configuration-dependent.","lead":"Two fermions on a tiny wire feel each other's presence so strongly that they pass through an asymmetric energy barrier much more easily from one side than from the other. The paper proves the effect comes from particle interactions and shows how it could be used to design controllable nanoscale switches.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The general claim that interactions break tunneling symmetry and produce U=h/2 resonances is supported only by exact diagonalization for the two-site sawtooth barrier of Eq. (2) at selected parameters; no theorem or parameter scan rules out shape-specific artifacts.","rationale":"The U=0 theorem and the triplet-sector argument are strong, so the concern is not about internal inconsistency but about the scope of the interacting claims. The reader's weakest assumption correctly identifies that the asymmetry and resonance effects are demonstrated for a single two-site sawtooth barrier at selected parameters. I add two sharpenings: the Falicov-Kimball supplementary argument only covers an immobile particle and therefore does not generalize the mobile singlet case, and the resonant-trapping explanation rests on an energy-matching argument rather than on spectral or scattering evidence. These are evidence/scope limitations rather than demonstrated errors, so they do not warrant rejection; they support keeping the verdict conditional. An explicit test with a smooth barrier would settle whether the general wording in the abstract is justified or whether the claims should be narrowed to the studied geometry.","tokens_in":14351,"tokens_out":14511,"duration_ms":154558,"concrete_test":"Recompute the time-averaged asymmetry of Fig. 4 and the U=h/2 trap of Fig. 5 with the same L=4 Hamiltonian but replace Eq. (2) by a smooth barrier V_j = h exp[-(j-2.5)^2/2] on the two central sites, scanning U/h over [0.1, 1.0] and checking T=100/J and T=500/J. If the asymmetry or the peak at U=h/2 disappears or changes sign for the smooth barrier, the abstract's general statements must be restricted to the specific sawtooth potential; if it persists, the conditional verdict can be upgraded.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The noninteracting symmetry theorem (Supp. Note 2) and the exact triplet-sector argument are solid: in the triplet channel the two doublon-producing hopping paths have opposite fermionic signs and cancel, so U never acts and the noninteracting symmetry survives. The load-bearing gap is the interacting singlet case and the U=h/2 effects. The abstract and introduction assert that inter-particle interactions generically break tunneling symmetry and create pronounced asymmetry, but the evidence is exact diagonalization for one two-site sawtooth potential [Eq. (2)] at L=4 and L=6, with h=10J/20J and U=0.5J/10J. No theorem analogous to the U=0 proof is given, and the Falicov-Kimball limit in Supp. Note 1 treats only an immobile particle, reducing the problem to a single-particle Hamiltonian with a modified onsite energy; it cannot certify the mobile two-particle case. In addition, the 'resonant' labels at U=h/2 are inferred from energy matching (initial doublon energy U equals potential energy h/2 at site j*), with no spectral or scattering analysis showing a resonance, and the argument assumes both the trapped particle and its partner have negligible kinetic energy. Finally, the asymmetry is measured as a finite-time population difference with a fixed averaging window T=100/J in a closed finite lattice; the statement that results are 'largely unaffected by increasing final time T' is not documented. If the asymmetry or the U=h/2 peak is special to the sawtooth shape or to the chosen time window, the abstract's sweeping conclusions overstate what the numerics establish.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies two spin-1/2 fermions on a one-dimensional Fermi-Hubbard chain with a two-site asymmetric barrier. It proves that for noninteracting fermions the tunneling probability is symmetric in finite discrete systems, shows that spin-triplet initial states preserve this symmetry while spin-singlet states break it, and identifies a regime U=h/2 where the authors report underbarrier resonant trapping and enhanced many-body resonant tunneling. The evidence is exact diagonalization for L=4,6,20 with the sawtooth potential of Eq. (2), supplemented by a rigorous path-based theorem for U=0 and a Falicov-Kimball limit for an immobile particle.","tokens_in":14714,"tokens_out":5200,"duration_ms":55237,"significance":"The noninteracting symmetry theorem in Supplementary Note 2 and the exact cancellation argument for triplet states are genuine and clean contributions, and the numerical code is publicly available. If the interaction-induced asymmetry and the U=h/2 phenomena are robust beyond the single sawtooth potential studied, the results would be interesting for few-fermion transport and possible device applications. The main weakness is that the broad claims in the abstract and introduction are supported by a narrow set of numerical examples rather than by a theorem or a systematic parameter scan.","major_comments":[{"comment":"The abstract states that inter-particle interactions break tunneling symmetry and create pronounced asymmetric tunneling behavior as a general property, but the supporting evidence is exact diagonalization for one specific two-site sawtooth potential at selected parameters (h=10J,20J; U=0.5J,10J; L=4,6,20). The noninteracting theorem applies to arbitrary barriers, but no analogous theorem or systematic scan over barrier shapes, heights, interaction strengths, or system sizes is given for the interacting case. Please either restrict the claims to the studied potential or add evidence, such as scans over h, U, and L and tests with other asymmetric potentials, that the asymmetry is generic.","section":"Abstract and Sec. II.C; Eq. (2)"},{"comment":"The label resonant in underbarrier resonant trapping is not demonstrated. The explanation equates the initial doublon energy U with the potential energy h/2 at site j*, but no spectral, scattering, or avoided-crossing analysis is presented, and the energy-conservation argument assumes that both the trapped particle and its partner carry negligible kinetic energy. A peak in trapping probability as a function of U with a resolvable width, or an analysis of the relevant eigenstates, would justify the term resonant; otherwise the phenomenon is better described as energy-matched trapping.","section":"Sec. III.A and Fig. 5"},{"comment":"The asymmetry is quantified by time averages over a fixed interval T=100/J in a closed finite lattice. The statement that the results are largely unaffected by increasing the final time T is not supported by any shown data. Since finite-system recurrences can make such time averages strongly T-dependent, the paper should include a convergence check, such as a T-scan or long-time average, before interpreting these averages as tunneling probabilities.","section":"Sec. III, Figs. 4 and 6"},{"comment":"The triplet symmetry argument is exact for the specific initial state |↑1↓2⟩+|↓1↑2⟩, where the double-occupancy terms cancel, but the text generalizes to spin-triplet states preserve tunneling symmetry without proving the statement for arbitrary triplet superpositions or other triplet components. Please make the claim precise by specifying the class of triplet initial states covered and, if necessary, extend the proof beyond the nearest-neighbor example shown.","section":"Sec. II.C and Eq. (5)"}],"minor_comments":[{"comment":"The Falicov-Kimball limit treats an immobile particle, so it cannot certify the mobile two-particle case; this limitation should be acknowledged where this limit is used as intuition in the main text.","section":"Supplementary Note 1"},{"comment":"The statement The code is available on GitHub provides no repository identifier or persistent version; please add a link or DOI so the exact code version can be retrieved for reproducibility.","section":"Code Availability"},{"comment":"The claim that the two curves overlap perfectly would be more convincing with a difference plot or a numerical estimate of the maximum deviation between the two curves.","section":"Fig. 2"},{"comment":"The time-averaged data would benefit from a discussion of numerical precision of the QuSpin time-evolution method, including any truncation error tolerance used in the Taylor-series expansion.","section":"Figs. 4 and 6"},{"comment":"The phrase treasure chest in the Introduction is informal and could be replaced with a more neutral description of the results.","section":"Sec. II.C"}],"recommendation":"major_revision","confidential_remarks":"The paper contains a solid noninteracting theorem and an elegant triplet argument, but the central interaction-induced claims are broader than the numerical evidence. I would be willing to accept a revised version that either narrows the claims to the studied sawtooth barrier or provides the requested parameter scans and resonance analysis."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper is worth engaging. The genuinely new pieces are the finite-lattice twin-path proof of tunneling symmetry for noninteracting fermions, the spin-triplet versus spin-singlet dichotomy, and the numerical observations of underbarrier trapping and enhanced tunneling at U = h/2. Credit where due: the noninteracting theorem in Supplementary Note 2 is clean and self-contained, and the triplet-sector argument is exact, since no double occupancy means U never acts, so the noninteracting symmetry survives. The exact diagonalization results in Figs. 2-6 are consistent with those claims. The soft spot is scope. The abstract and introduction assert that interactions generically break tunneling symmetry and produce pronounced asymmetric tunneling, but the evidence is exact diagonalization for one two-site sawtooth potential at selected parameters (h = 10J/20J, U = 0.5J/10J, L = 4 and 6). There is no theorem for the interacting singlet case, and no systematic scan over barrier shapes or heights. The Falicov-Kimball limit in Supplementary Note 1 treats only an immobile particle, so it cannot certify the mobile two-particle case. The 'resonant' labels at U = h/2 are inferred from energy matching (initial doublon energy U equals potential h/2 at site j*), not from a spectral or scattering analysis, and the argument assumes both particles carry negligible kinetic energy. The fixed averaging window T = 100/J is another soft spot; the statement that results are 'largely unaffected by increasing final time T' is not documented. Finally, the code is announced but no link or reproducibility artifact is provided. None of this sinks the core results. The noninteracting theorem and the triplet-singlet distinction are solid, and the U = h/2 trapping is plausibly real. But the strongest claims in the abstract overstate what the numerics establish. A careful referee should ask for either a broader parameter scan or a toned-down generalization, plus the actual code. This paper deserves a serious referee. The exact theorem and the clean triplet argument merit publication even if the broader resonance claims need revision. I would cite it for the symmetry proof and the spin dichotomy. For a reading group, it is a decent example of a mostly solid paper where the headline overreaches the evidence.","headline":"Solid exact theorem plus convincing numerics, but the broad claims about interaction-induced asymmetry rest on a single barrier shape and a narrow parameter window.","tokens_in":743,"tokens_out":889,"would_cite":true,"duration_ms":21587,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["71.10.Fd","73.40.Gk","67.85.-d"],"model":"deepseek-v4-flash","headline":"For noninteracting fermions, tunneling through an asymmetric barrier is exactly symmetric; adding interactions breaks that symmetry, with spin-singlet pairs tunneling asymmetrically while spin-triplet pairs remain symmetric.","keywords":["Fermi-Hubbard model","quantum tunneling","asymmetric barrier","spin singlet and triplet","resonant tunneling","underbarrier trapping","few-fermion dynamics","cold-atom lattices"],"falsifier":"Run exact-diagonalization dynamics for two interacting spin-singlet fermions on a lattice with an asymmetric barrier of a different shape, such as a single-site delta barrier or a smooth ramp, and compare left-to-right and right-to-left tunneling probabilities; if the asymmetry or the $U = h/2$ trapping peak disappears for some barrier shape, the general claim fails. A direct cold-atom experiment with ultracold fermions could test whether singlet pairs tunnel preferentially from one side at $U = h/2$.","tokens_in":1629,"feed_emoji":"⚛️","tokens_out":7256,"duration_ms":106942,"temperature":0.7,"pith_summary":"The paper asks whether tunneling through an asymmetric potential barrier can be directional in a few-fermion lattice system. It proves that for noninteracting particles the tunneling probability is exactly the same from either side of any barrier, even on a finite chain, extending the established continuum result. It then argues that on-site interactions break this left-right symmetry, with the effect depending on the pair's spin state: triplet pairs tunnel symmetrically, singlet pairs do not. At an interaction strength equal to half the barrier height, the same Hamiltonian produces underbarrier resonant trapping and many-body resonant tunneling through a single barrier. If these findings hold, interactions give a new control knob for directional transport in nanoscale and cold-atom devices.","feed_headline":"Interactions give fermion tunneling a direction preference","feed_subtitle":"Singlet pairs pick a direction; at resonance, trapping holds fermions mid-barrier.","key_machinery":"The central object is the one-dimensional Fermi-Hubbard Hamiltonian with nearest-neighbor hopping $J$, on-site interaction $U$, and an asymmetric two-site barrier potential with heights $h$ and $h/2$; this model generates all the reported tunneling dynamics. The noninteracting symmetry theorem is carried by a discrete path-sum (time-slicing) argument in which each contributing sequence of sites has a reversed twin with identical amplitude, using the reality and symmetry of the single-particle Hamiltonian. The spin dichotomy is carried by expanding the time-evolution operator: for the triplet state, terms that place two fermions on one site cancel at every order, so the dynamics never feels $U$, while for the singlet state they do not. The trapping and resonant tunneling regimes are carried by energy conservation at the resonance $U = h/2$, where a particle parked on the half-height site has energy equal to the initial doublon energy.","core_discovery":"The central claim is that the left-right symmetry of single-particle tunneling, known to be exact in one-dimensional continuous systems, survives on finite discrete lattices but is destroyed by interactions. The proof of the noninteracting theorem uses a path-reversal argument: every Feynman path that carries a particle from left to right through an arbitrary barrier has a mirror 'twin' path of equal amplitude from right to left. For two spin-$1/2$ fermions, the triplet initial state retains this symmetry because no doubly occupied site is ever created, so interaction terms cancel term by term in the time-evolution expansion; the singlet state does create doublons and shows pronounced asymmetry. At $U = h/2$, energy conservation confines one fermion at the barrier site with potential energy $h/2$, producing underbarrier resonant trapping, and in a three-particle setup the same resonance enables many-body resonant tunneling through a single barrier.","pith_inferences":["One testable extension is a systematic scan over barrier shapes (single-site delta, smooth Gaussian ramps, wider barriers) to see whether the singlet asymmetry and the $U = h/2$ resonance survive outside the sawtooth geometry; the paper reports only one shape.","If the spin dichotomy extends to larger systems, preparing spin-polarized versus spin-singlet pairs could act as an interaction-tunable direction selector for cold-atom transport experiments.","The near-linear growth of trapping time with lattice size shown in the supplementary material suggests a practical timescale metric for designing delay or memory elements based on underbarrier trapping.","The localized-itinerant limit described in the supplement shows that the interaction effectively reshapes the barrier felt by the moving fermion, suggesting the asymmetry mechanism could be reformulated as interaction-induced barrier reshaping in other lattice models."],"forward_implications":["Noninteracting fermions on a finite lattice tunnel through any potential barrier with equal left-to-right and right-to-left probabilities, so the symmetry theorem is not limited to continuous infinite systems.","For two spin-$1/2$ fermions, triplet states preserve tunneling symmetry while singlet states show strong interaction-induced asymmetry.","At $U = h/2$, the system exhibits underbarrier resonant trapping: a fermion dwells on the half-height barrier site, suppressing tunneling, and this persists as the system size grows.","With an additional particle beyond the barrier, the $U = h/2$ resonance produces many-body resonant tunneling through a single barrier, an order-of-magnitude enhancement of tunneling from the steep side.","These effects arise solely from inter-particle interactions and have no single-particle resonant-tunneling analogue, which requires two or more barriers."],"supporting_citations":[{"why":"Establishes the continuous one-dimensional result that tunneling probability is symmetric under barrier reversal, the baseline the paper extends.","marker":"[41]"},{"why":"Provides a modern treatment of the same continuous symmetry, used as supporting context for the discrete theorem.","marker":"[42]"},{"why":"Documents earlier work on left-right tunneling symmetry breaking by interactions, which the paper explains in few-fermion systems.","marker":"[43]"},{"why":"Reports a related trapping effect for bosons, which the paper contrasts with and extends to fermions.","marker":"[44]"},{"why":"Defines single-particle resonant tunneling as requiring two or more barriers, the baseline against which the paper distinguishes its single-barrier many-body resonant tunneling.","marker":"[45]"},{"why":"Observes pair tunneling in a one-dimensional two-particle Hubbard model, the cold-atom experimental setting most relevant to the predicted effects.","marker":"[32]"},{"why":"Supplies the open-source exact-diagonalization and time-evolution code used for the numerical simulations.","marker":"[55]"},{"why":"Documents the time-propagation method in that code, which produces the tunneling dynamics reported in the paper.","marker":"[56]"}],"fun_headline_variants":["Interactions tilt fermion tunneling odds","Singlet fermions tunnel asymmetrically, triplets don't","Underbarrier trapping emerges from interacting fermions","Many-body resonance boosts fermion tunneling"],"cache_read_input_tokens":17280,"weakest_assumption_plain":"The broad claim rests on numerical simulations of one sawtooth barrier shape at selected heights and interaction strengths, plus an energy-conservation argument that assumes the non-trapped fermion carries negligible kinetic energy.","fun_headline_variants_meta":{"raw":{"variants":["Interactions tilt fermion tunneling odds","Singlet fermions tunnel asymmetrically, triplets don't","Underbarrier trapping emerges from interacting fermions","Many-body resonance boosts fermion tunneling"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000293,"raw_usage":{"total_tokens":1674,"prompt_tokens":878,"completion_tokens":796,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":494,"completion_tokens_details":{"reasoning_tokens":738}},"tokens_in":494,"tokens_out":796,"duration_ms":7340,"temperature":1.0,"reasoning_tokens":738,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T22:20:58.680711+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run exact-diagonalization dynamics for two interacting spin-singlet fermions on a lattice with an asymmetric barrier of a different shape, such as a single-site delta barrier or a smooth ramp, and compare left-to-right and right-to-left tunneling probabilities; if the asymmetry or the $U = h/2$ trapping peak disappears for some barrier shape, the general claim fails. A direct cold-atom experiment with ultracold fermions could test whether singlet pairs tunnel preferentially from one side at $U = h/2$.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reports a related trapping effect for bosons, which the paper contrasts with and extends to fermions."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes the continuous one-dimensional result that tunneling probability is symmetric under barrier reversal, the baseline the paper extends."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides a modern treatment of the same continuous symmetry, used as supporting context for the discrete theorem."},{"cited_title":"Lindberg, N","cited_arxiv_id":null,"evidence_quote":"Documents earlier work on left-right tunneling symmetry breaking by interactions, which the paper explains in few-fermion systems."},{"cited_title":"Razavy, Quantum Theory of Tunneling(World Sci- entific, Singapore, 2013)","cited_arxiv_id":null,"evidence_quote":"Defines single-particle resonant tunneling as requiring two or more barriers, the baseline against which the paper distinguishes its single-barrier many-body resonant tunneling."},{"cited_title":"Mukherjee, M","cited_arxiv_id":null,"evidence_quote":"Observes pair tunneling in a one-dimensional two-particle Hubbard model, the cold-atom experimental setting most relevant to the predicted effects."}],"review_version":1}