{"id":"b404cb38-652a-4171-a5c9-91229e9ed0d5","arxiv_id":"2602.20508","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"For a few bosons in a small lattice, an asymmetric barrier plus tuned interactions makes a quenched initial state oscillate through the barrier from one side while remaining an eigenstate (trapped) from the other.","lead":"This paper reports exact-diagonalization results for a few interacting bosons in a small optical lattice with an asymmetric barrier, showing that a carefully prepared initial state tunnels through the barrier from one side but stays trapped from the other. The authors interpret this transient asymmetry as a 'Hilbert-space event horizon' and claim interactions are a new route to directional transport.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Missing U=0 baseline: if noninteracting bosons already yield nonzero Δn for the same quench, the 'purely interaction-driven' directionality claim collapses.","rationale":"The reader's weakest assumption correctly identifies the missing U=0 baseline as the key control: the paper claims interactions are the sole source of directionality, but never isolates the noninteracting contribution. The cited single-particle symmetry does not cover the finite-time quench observable, so a nonzero U=0 Δn would falsify the central claim. However, the reader's secondary concern—that the 'opposite direction' is never simulated—is weaker than stated. Since Hb is the mirror image of Ha, the comparison n_after^a vs n_after^b is mathematically equivalent to comparing left-incident and right-incident initial states on the same barrier, provided the initial states are mirror images. Thus the protocol does address directionality, and the decisive missing control is specifically U=0. The verdict should remain CONDITIONAL: the paper's numerics may be correct, but the headline claim is contingent on a baseline that has not been shown. My recommendation is UNCHANGED relative to the reader because this concern is already embedded in the reader's conditional verdict.","tokens_in":8565,"tokens_out":9761,"duration_ms":96993,"concrete_test":"Run exact diagonalization for L=6, N=4, h=10J with the same cooling/quench protocol at U=0 (and, to check continuity, U=0.1J); plot Δn(t) and compare its maximum to the U=1.42J value. If max|Δn(U=0)| is comparable to the claimed signal, the interaction-driven interpretation fails. If it is zero (or <0.01), the central claim survives this control; as a secondary check, directly evolve a right-localized initial state under Ha and verify n_left equals n_after^b.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim—that directionality emerges 'purely from many-body interactions'—requires that the noninteracting (U=0) limit show no population imbalance Δn for the same quench protocol. The paper never presents this baseline. The cited single-particle left/right transmission symmetry (Refs. [44–46]) applies to asymptotic scattering at fixed energy, not to the finite-time, finite-lattice observable in Eq. (4). Because Ha and Hb are mirror images, comparing n_after^a and n_after^b does encode left- vs right-incidence on the same barrier (n_after^b = P_{Ha,Rψ0→L} after reflection), so the protocol can in principle isolate directionality—but only if the U=0 comparison vanishes. For noninteracting bosons, the initial state is a fragmented condensate in the cooling trap and evolves via single-particle U(t); the two barrier Hamiltonians are related by reflection, and there is no general symmetry forcing P_{ψ0→R}^{Ha}=P_{Rψ0→L}^{Ha} at finite time. Without the U=0 row in Fig. 2(a), the observed Δn(U≈1.42J) could be an interaction-independent scattering asymmetry of the triangular barrier. The eigenstate-overlap mechanism in Fig. 3 would then not be established as the cause. This is a missing control, not an internal inconsistency.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies a Bose-Hubbard model on a one-dimensional lattice with a static asymmetric two-site barrier. An initial state is prepared by loading bosons into a cooling trap on the left, then quenching to one of two mirror-related barrier configurations H_a and H_b. The central observable is the population imbalance Δn = n_after^a - n_after^b (Eq. 4). For L=6, N=4, h=10J, and U≈1.42J, exact diagonalization shows a positive Δn, i.e., more particles cross when the high barrier is encountered first ('vertical side') than when the low barrier is first ('angled side'). The authors attribute this to interaction-induced projection of the initial state onto transport-enabled versus transport-forbidden eigenstates, and present this as a new 'Hilbert space black hole analog' with unidirectional transport in a closed, undriven system. They support the result with a coherent-state initial condition and finite-size scans in supplementary material.","tokens_in":8834,"tokens_out":7812,"duration_ms":73004,"significance":"If the claim holds—that interactions alone can rectify transport in a closed, undriven system—this would be a conceptually novel mechanism with potential atomtronic applications. The exact-diagonalization data for small systems appear numerically reliable, and the use of QuSpin makes the results reproducible. The main weakness is that the central attribution to interactions is not backed by a U=0 control, and the 'robust unidirectional' claim is stronger than the narrow parameter windows shown. The paper is thus suggestive but not yet definitive; the missing control is a fixable omission.","major_comments":[{"comment":"The central claim—that directionality emerges 'purely from many-body interactions'—requires a zero-interaction control. The paper never computes Δn(t) at U=0 for the same cooling-barrier quench and the same H_a/H_b comparison. The single-particle transmission symmetry cited from Refs. [44–46] applies to asymptotic left-vs-right transmission at fixed energy, not to the finite-time observable in Eq. (4). Because the initial state is left-localized, H_a and H_b are mirror images but the protocol is not mirror-symmetric; there is no general symmetry forcing Δn(U=0)=0. If the U=0 baseline is already nonzero, the observed Δn at U≈1.42J is not purely interaction-driven. Please add the U=0 (and a small-U) baseline and discuss its behavior.","section":"§Model and Methods, Eq. (4); §Discussion, Fig. 2(a)"},{"comment":"The term 'unidirectional' is inferred by comparing two barrier orientations with the same left-incident initial state. To claim one-way transport, the equivalence between this comparison and a true left-vs-right incidence test should be stated explicitly: H_b = R H_a R with R the spatial reflection, and the right-incident initial state should be the reflected cooling-trap ground state. The paper does not define this mapping, so the reader cannot verify that the protocol isolates direction rather than merely an orientation-dependent transient. Add a sentence or a short derivation making this equivalence precise.","section":"§Discussion, Fig. 2; §Model and Methods"},{"comment":"The adjective 'robust' is not supported by the presented data. For the main case (L=6, N=4), the strong positive Δn occurs in a narrow window of width ΔU≈0.05J (3.5% in U), and the paper explicitly excludes 'narrower resonance-like features' at higher U. The supplementary scans show that the structure fragments with N and shifts with L. The conclusions state 'for a broad range of system parameters,' but no quantitative robustness measure (e.g., a range of h and U over which Δn exceeds a threshold) is given. Please either soften the claim or provide such an analysis.","section":"§Discussion, Fig. 2(a); Supplementary Note 1"}],"minor_comments":[{"comment":"Reference [47] is incomplete ('L. Amico and et al.') and reference [54] is truncated after 'P. Thekkeppatt'; please complete the bibliography.","section":"References"},{"comment":"The coherent-state initial condition is not defined unambiguously: the same symbol n_j is used for the coherent-state amplitude and for the particle number. Please clarify the notation, e.g., use α_j for the coherent-state amplitude.","section":"§Discussion, coherent-state paragraph"},{"comment":"The heatmap is described as red/blue regions, but no color scale or numerical Δn values are provided in the caption. Adding a colorbar or stating the saturation level would improve interpretability.","section":"Fig. 2(a) caption"},{"comment":"The text switches between 'vertical side' and 'angled side' without consistently connecting to Eq. (2) vs Eq. (3). Define the mapping once and use it throughout.","section":"§Discussion"},{"comment":"The analogy to a black-hole event horizon is heuristic; the paper does not define any causal structure or horizon in Hilbert space. As a metaphor it is evocative, but it should be presented as an analogy rather than a rigorous result.","section":"Abstract and Conclusions"}],"recommendation":"major_revision","confidential_remarks":"The paper is potentially interesting, but the central interaction-driven claim rests on a missing U=0 baseline that is easy to compute and could falsify the interpretation. The 'robust unidirectional' and 'black-hole horizon' language also outruns the data. I would encourage the authors to add the control and temper the claims before publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"What you should know: this paper reports a genuine, reproducible numerical observation—for four bosons on six sites, a quench from a left-localized initial state into one of two mirror-image asymmetric barriers gives a stark asymmetry in how much population ends up on the right. The strong imbalance at U≈1.42J is real, and the authors show it persists for coherent initial states and in larger systems (L=8,10,12) in the supplement. That specific demonstration is not in the cited literature. The eigenstate-overlap explanation (angled side projects almost entirely onto one eigenstate with no post-barrier weight, vertical side spreads over several) is plausible and consistent with the numbers.\n\nThe soft spot is the one the reader flagged: there is no U=0 baseline. The abstract claims directionality \"emerges purely from many-body interactions,\" but the paper never shows that the noninteracting limit gives zero population imbalance for the same quench. The cited single-particle transmission symmetry is about asymptotic scattering at fixed energy, not the finite-time, open-boundary observable in Eq. (4). It is entirely possible that a noninteracting boson wave packet launched from the same initial trap already shows an asymmetry between the two barrier orientations—the initial state is a fragmented condensate, not a plane wave, and the two Hamiltonians are related by reflection, not by any symmetry of the protocol. Without the U=0 row, the \"interactions as a new route\" claim collapses to \"some interaction strength gives a large imbalance,\" which is weaker and less interesting.\n\nA few smaller issues. The \"opposite direction\" is never simulated directly; the authors infer it by comparing the two barrier orientations with the same left-side initial state. That comparison does encode directionality, but only after reflection of the barrier and initial state—not a direct test of right-incidence. The headline U is a selected peak, and narrower resonance-like features are excluded; the claim of robustness rests on a 3.5% window, which is fine but should be stated as such. The black-hole language is a stretch: the dynamics are transient oscillations in a finite closed system, not an irreversible horizon, and the \"Hilbert-space event horizon\" adds no predictive content. The eigenstate mechanism is interpretive—read off the same exact-diagonalization data—but it is at least a concrete, checkable statement about overlaps.\n\nThe paper is worth a serious referee. The missing baseline is addressable and should be requested; the rest of the numerics look solid and reproducible. If the U=0 control comes out zero, this becomes a nice, modest result about interaction-tuned transport asymmetry in small lattices. If it does not, the central claim needs substantial reworking. My recommendation: send it to review, with a firm request for the U=0 baseline and a direct right-incidence check.","headline":"A concrete exact-diagonalization observation of direction-dependent transport for interacting bosons behind a static asymmetric barrier, but the headline claim that interactions alone cause the effect lacks the U=0 control.","tokens_in":9395,"tokens_out":2301,"would_cite":false,"duration_ms":23150,"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":"Interacting bosons in a static optical-lattice barrier can tunnel in one direction only, with no driving or dissipation—a quantum analog of a black-hole event horizon in Hilbert space.","keywords":["Bose-Hubbard model","unidirectional transport","optical lattice","many-body interactions","eigenstate projection","quantum quench","atomtronics","black hole analog"],"falsifier":"Run the identical quench protocol for Ha and Hb with U=0 (noninteracting bosons) for the same lattice, barrier height, and initial state, and measure Δn(t); any significant nonzero imbalance would falsify the claim that interactions are the source of directionality. A complementary check is to compute single-particle wave-packet transmission through the two finite barriers and see whether the instantaneous left-vs-right difference vanishes.","tokens_in":8395,"feed_emoji":"⚛️","tokens_out":3650,"duration_ms":33931,"temperature":0.7,"pith_summary":"This paper claims that many-body interactions alone can rectify quantum transport in a closed, undriven system. In a one-dimensional optical lattice with a static asymmetric two-site barrier, bosons prepared on one side tunnel across readily, while bosons prepared on the mirror-image side stay trapped—provided the on-site repulsion sits in the right window (around U≈1.42J for the studied parameters). The directionality is traced to how the barrier shape projects the initial state onto eigenstates that either contain or exclude particles on the far side. If correct, this gives a new mechanism for building directional elements in atomtronic circuits without reservoirs or time-dependent potentials.","feed_headline":"Interactions alone push bosons one way through a static barrier","feed_subtitle":"A closed, undriven quantum system may rectify transport, a first step toward interaction-only atomtronic diodes.","key_machinery":"The mechanism is the combination of (i) an asymmetric two-site barrier—two neighboring lattice sites with heights h and h/2, in two mirror configurations Ha and Hb—and (ii) on-site repulsion U in the Bose-Hubbard model. The observable is the population imbalance Δn between the two configurations after the same quench. The barrier asymmetry, together with U, reshapes the many-body eigenstates: from one side the initial state lands almost entirely on an eigenstate with no weight beyond the barrier; from the other side it lands on a superposition with weight on Fock states that have particles past the barrier. That eigenstate-selective projection, not time-reversal breaking by driving or dissip","core_discovery":"The central claim is that a purely static asymmetric potential plus bosonic interactions produces unidirectional transport: for a lattice with two adjacent barrier sites of heights h and h/2, swapping which site carries the taller barrier reverses the role of 'easy' and 'hard' tunneling directions. At U≈1.42J and h=10J, bosons tunnel from the 'vertical' side (h then h/2) with large oscillations, while from the 'angled' side (h/2 then h) post-barrier population stays below 0.1. The authors show the same qualitative behavior for both Fock and coherent initial states and across lattice sizes and particle numbers. Eigenstate analysis attributes the asymmetry to the barrier selecting a nearly sin","pith_inferences":["If the mechanism is eigenstate selection, varying U or the barrier shape should allow continuously tuning from one-way to bidirectional transport or even reversing the preferred direction; the paper shows multiple windows with opposite signs at higher U, which supports this but does not map the full phase diagram.","The paper's central attribution to interactions would be tested directly by running the same quench with U=0; that baseline is not shown, and if a nonzero Δn appears there, the Hilbert-space-horizon interpretation would need revision.","A natural extension is to two or three barriers or to fermionic species: the same eigenstate-selection logic might produce directionality in other many-body settings, but that is beyond what the paper demonstrates.","The coherent-state result suggests the effect may survive the classical limit, which would make it observable in relatively warm condensates rather than requiring deep quantum degeneracy."],"forward_implications":["Atomtronic diodes could be built from static lattices with interaction-tuned barriers, requiring only Feshbach control of U rather than engineered dissipation or periodic driving.","The direction of rectification can be chosen by which of the two mirror barriers is used, so a single device could switch transport direction by swapping the barrier profile.","The effect persists for coherent (condensate-like) initial states and for several lattice sizes and fillings, suggesting it is not a fine-tuned single-Fock-state artifact.","The required interaction precision (δU/U ≈ 3.5%) is within reach of Feshbach resonance control in cold-atom experiments.","Because the mechanism is eigenstate selection, it defines a new class of black-hole analogs located in Hilbert space rather than in curved spacetime."],"fun_headline_variants":["Boson interactions create one-way transport through a static barrier","Interacting bosons rectify transport without any driving","Asymmetric barrier plus interactions gives one-way flow","Bosonic interactions alone create a quantum event horizon","Interaction-only diode: unidirectional boson transport"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The claim that interactions cause the directionality assumes that with U=0 the same two barrier configurations give no population imbalance; the paper never computes that noninteracting baseline.","fun_headline_variants_meta":{"raw":{"variants":["Boson interactions create one-way transport through a static barrier","Interacting bosons rectify transport without any driving","Asymmetric barrier plus interactions gives one-way flow","Bosonic interactions alone create a quantum event horizon","Interaction-only diode: unidirectional boson transport"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000571,"raw_usage":{"total_tokens":2484,"prompt_tokens":636,"completion_tokens":1848,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":380,"completion_tokens_details":{"reasoning_tokens":1775}},"tokens_in":380,"tokens_out":1848,"duration_ms":12871,"temperature":1.0,"reasoning_tokens":1775,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T21:19:50.366414+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the identical quench protocol for Ha and Hb with U=0 (noninteracting bosons) for the same lattice, barrier height, and initial state, and measure Δn(t); any significant nonzero imbalance would falsify the claim that interactions are the source of directionality. A complementary check is to compute single-particle wave-packet transmission through the two finite barriers and see whether the instantaneous left-vs-right difference vanishes.","supporting_citations":[],"review_version":1}