{"id":"b19cb913-b940-4ebe-95a2-60d4e6111bd8","arxiv_id":"2607.07714","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"A single boundary non-Hermitian bond drives the Hatano-Nelson localization-delocalization transition via total imaginary gauge flux alone; isospectral models share static criticality but exhibit inequivalent dynamics.","lead":"A single non-Hermitian bond on a disordered ring drives the same Anderson delocalization transition as bulk nonreciprocity, controlled only by total imaginary flux. Static criticality is universal across isospectral models while dynamics (scrambling, wavepackets, entanglement) are not, with a multi-terminal experiment proposed.","discovery_kind":"unification","skeptic_critique":{"model":"grok-4.5","headline":"Exponentially large single-bond hopping t_L=e^{Lγ} challenges the thermodynamic-limit interpretation of a 'strictly local' non-Hermitian perturbation while preserving total flux Γ=Lγ.","rationale":"The algebraic isospectrality and the invariance proofs for all static diagnostics (spectrum, W, IPR/D_2, fidelity) are rigorous and leave no internal inconsistency. The numerical collapses for three independent observables are consistent and match the known HN phase boundary. The only load-bearing interpretive gap is precisely the one identified by the reader: whether an exponentially growing single-bond amplitude can still be regarded as a 'strictly local' perturbation once L→∞. That gap justifies retaining CONDITIONAL rather than upgrading to ACCEPT; it does not warrant REJECT because every finite-L statement in the paper is correct. No stronger technical flaw (e.g., failure of the similarity map, incorrect OTOC definition, or breakdown of the multi-terminal construction) appears. Hence the reader's verdict and weakest-assumption diagnosis stand.","tokens_in":22132,"tokens_out":664,"duration_ms":24466,"concrete_test":"Recompute the f_c and ⟨D_2⟩ scaling collapses for the SBN model at fixed γ=0.1 on systems L=300–500 (using a scaled or multiprecision eigensolver to handle the e^{Lγ} entry). If the extracted w_c and ν remain identical to the HN values within statistical error, the finite-size static universality is confirmed; if they drift, the size-dependent coupling alters the apparent critical exponents. Separately, evaluate the L-scaling of the bond current and the weight of the nH link in the steady-state density at criticality; growth faster than O(1) would quantify the non-local character of the 'local' bond.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that a single boundary bond suffices (and that extensive bulk nonreciprocity is unnecessary) rests on concentrating the entire extensive flux Γ=Lγ onto one link, producing t_L=e^{Lγ} (Model Hamiltonian, Eq. (1) and the α_j=(j-L)γ choice). For any fixed γ>0 this amplitude diverges exponentially with L, so the perturbation is not size-independent. Although the similarity map S=exp(∑α_j n_j) rigorously guarantees identical spectra, winding numbers, biorthogonal IPRs and D_2 for every finite L (SM sections on isospectrality, W, IPR), the thermodynamic-limit meaning of 'local' and of the finite-size scaling collapses (ν=2, w_c≈3.2 at γ=0.1) becomes ambiguous: the Lieb-Robinson velocity itself scales as cosh(Lγ) (SM), and the Hermitian sector of H_SBN is spatially anisotropic. Thus the static universality class is identical by construction, yet the physical assertion that a strictly local link drives bulk delocalization in the L→∞ limit is not secured by the existing finite-L data.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The paper constructs an exactly isospectral family of non-Hermitian disordered ring Hamiltonians H({α_j}) related by the similarity S=exp(∑ α_j n_j) that redistributes imaginary gauge flux while keeping the total flux Γ=Lγ fixed. Static diagnostics of the localization-delocalization transition (fraction of complex eigenvalues f_c, winding number W, biorthogonal IPR and fractal dimension ⟨D_2⟩) are shown to be identical across the family and controlled solely by Γ, with clean finite-size collapses yielding ν=2 and a common critical line in the (w,γ) plane. The single-bond non-Hermitian (SBN) endpoint (all flux on one link) therefore exhibits the same spectral, eigenstate and topological criticality as the uniform Hatano-Nelson model. Dynamics, however, are inequivalent: SBN displays rapid OTOC scrambling, oscillatory center-of-mass acceleration, and a double re-entrant steady-state entanglement transition linked to bimodal localization. An experimental multi-terminal QWZ-BHZ conductance realization is proposed.","tokens_in":22494,"tokens_out":1197,"duration_ms":25455,"significance":"If the thermodynamic-limit interpretation holds, the work cleanly separates static universality (gauge-invariant and fixed by total flux alone) from dynamical inequivalence in non-Hermitian Anderson systems, and shows that bulk delocalization need not require extensive bulk nonreciprocity. The similarity map is rigorous, the SM proofs of invariance of W, IPR and fidelity are explicit, and the finite-size collapses for f_c, W and ⟨D_2⟩ are of high quality. The dynamical distinctions (especially re-entrant entanglement and size-dependent Lieb-Robinson velocity) and the concrete multi-terminal proposal are genuine additions. These results organize a broader class of non-Hermitian disordered models and supply falsifiable dynamical signatures.","major_comments":[{"comment":"Model Hamiltonian (Eq. 1) and the SBN limit α_j=(j-L)γ: the single-bond amplitude is t_L=e^{Lγ}. For any fixed γ>0 this diverges exponentially with system size. The central claim that 'a single non-Hermitian boundary bond … suffices' and that 'extensive bulk nonreciprocity is not a necessary ingredient' therefore rests on a size-dependent coupling. While the similarity transformation guarantees identical spectra, W, IPR and D_2 for every finite L (SM), the thermodynamic-limit meaning of a 'strictly local' perturbation, and whether the reported scaling collapses (ν=2, w_c≈3.2 at γ=0.1) still define the same universality class when the non-Hermitian strength itself scales with L, is not secured. The SM Lieb-Robinson bound v_LR=4J cosh(Lγ) likewise grows with L. A clear statement of the L\toà limit (fixed bond strength vs fixed Γ, or an appropriate scaling of γ) is required for the physical","section":null},{"comment":"Dynamics of SBN and End Matter comparison: the Hermitian sector of H_SBN is spatially anisotropic (Eq. 8 and SM acceleration derivation). The oscillatory COM acceleration and the double re-entrant entanglement are traced to this anisotropy plus the boundary current. Because the static criticality is identical by construction, the claim of 'dynamical inequivalence within the same static universality class' is well supported numerically, yet the manuscript should quantify how much of the dynamical distinction survives once the Hermitian anisotropy is removed (the hybrid-SBN model already introduced in the SM). Without that control, it remains possible that part of the reported dynamical separation is an artifact of the Hermitian rather than the non-Hermitian sector.","section":null}],"minor_comments":[{"comment":"Abstract and p. 5: 'double re-entrant steady state entanglement transition' is clear from Fig. 3(d), but the four-phase sequence (extended/critical/bimodal/localized) should be labeled consistently in the main text and figure captions.","section":null},{"comment":"p. 4: 'V on-Neumann' → 'von Neumann'; several other minor typos ('End Matter', 'pbonds', 'la-layer').","section":null},{"comment":"Fig. 2 and SM Fig. 11: the spectral window (central 20 %) and Im(E) threshold (10^{-13}) used for f_c and D_2 should be stated once in the main text for reproducibility.","section":null},{"comment":"Experimental Realization: the claim that the multi-terminal conductance matrix realizes the disordered isospectral family is plausible for the clean case, but a brief remark on how quenched onsite disorder would be introduced (or post-selected) would strengthen the proposal.","section":null}],"recommendation":"major_revision","confidential_remarks":"The thermo-limit issue with t_L=e^{Lγ} is the only load-bearing concern; once clarified the paper is a solid contribution. The distinction from prior impurity papers (Refs. 15-17,68,69) is adequately drawn. Suitable for the journal after revision."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"The punchline is clean: a single non-Hermitian bond of strength e^{Lγ} on a disordered ring produces the same spectral, eigenstate and topological localization-delocalization transition as the uniform Hatano-Nelson model, because the whole family is related by a similarity transformation that preserves total imaginary flux Γ = Lγ. Static diagnostics (f_c, winding number W, biorthogonal IPR and D_2) collapse with the same \nu = 2 and the same phase boundary; dynamics (OTOC scrambling, oscillatory COM acceleration, re-entrant steady-state entanglement) do not. That separation is the real contribution.\n\nWhat they do well is the algebra and the numerics. The map S = exp(∑ α_j n_j) is written out carefully, the SM proves invariance of spectrum, W, IPR and fidelity, and the finite-size collapses for three independent static probes are consistent. The dynamical figures for the single-bond case are concrete and show clear qualitative differences from the uniform case (rapid operator exploration even when COM is suppressed, double re-entrance in entanglement tied to bimodal localization). The multi-terminal QWZ-BHZ proposal with KWANT conductance matrices is a practical route that keeps the idea experimentally grounded.\n\nThe soft spot is exactly the one the stress-test flags: t_L = e^{Lγ} is not a size-independent local perturbation. For fixed γ > 0 it diverges, the Lieb-Robinson velocity scales as cosh(Lγ), and the Hermitian sector is anisotropic. So the thermodynamic-limit claim that “a strictly local bond suffices” needs more care than the paper gives it; the finite-L data and the isospectral construction are solid, but the L \to ∞ interpretation of “local” is not fully secured. That is a genuine caveat, not a fatal flaw. No code or raw data are released, which is a minor practical annoyance.\n\nThis is for people who work on non-Hermitian localization, skin effect, or open quantum systems and who care about what is gauge-invariant versus what is representation-dependent. It is not a broad condensed-matter paper, but inside its niche it is useful. I would send it to referees; the math is clean enough and the distinction between static and dynamical universality is worth having on the record. Engage with it if you are writing about non-Hermitian criticality or impurity-driven transitions.","headline":"Single-bond non-Hermitian flux drives the full Hatano-Nelson LDL transition with identical static criticality but inequivalent dynamics; the exponential bond strength is a real thermodynamic-limit caveat but does not erase the finite-L result.","tokens_in":23102,"tokens_out":638,"would_cite":true,"duration_ms":6780,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.5","headline":"A single non-Hermitian bond on a disordered ring drives the Anderson delocalization transition; static criticality depends only on total imaginary flux, while dynamics do not.","keywords":["non-Hermitian Anderson transition","Hatano–Nelson model","imaginary gauge flux","isospectral family","localization–delocalization","operator scrambling","steady-state entanglement","multi-terminal transport"],"falsifier":"Measure the complex-eigenvalue fraction or winding number versus disorder for both the uniform Hatano–Nelson ring and a single-bond ring engineered with the same total flux; if the critical disorder strengths or the extracted exponent ν fail to coincide, the claim of static gauge invariance is false.","tokens_in":22994,"feed_emoji":"🔄","tokens_out":716,"duration_ms":7018,"temperature":0.7,"pith_summary":"Anderson localization in one dimension is famously robust: any weak disorder traps every wavefunction. The classic Hatano–Nelson picture shows that spreading non-reciprocal hopping throughout the bulk can overcome that trapping by means of an imaginary gauge flux. This paper shows that the bulk spreading is unnecessary. Concentrating the entire flux onto one boundary bond of a disordered ring is already enough to trigger the same localization–delocalization transition. The authors construct a continuous family of Hamiltonians, all related by complex gauge transformations, that interpolate between the uniform bulk model and the single-bond extreme while keeping the total flux fixed. Every static diagnostic—complex-eigenvalue fraction, inverse-participation ratios, fractal dimensions, winding number—collapses onto the same critical surface controlled solely by that total flux. Yet the same family displays sharply different real-time dynamics: the single-bond member shows rapid operator scrambling, oscillatory wave-packet acceleration, and a double re-entrant steady-state entanglement transition that are absent from the uniform case. The result separates static universality from dynamical inequivalence in non-Hermitian disordered systems and supplies a multi-terminal topological-transport route to realize the effect.","feed_headline":"One non-Hermitian bond delocalizes a disordered ring","feed_subtitle":"Static criticality depends only on total imaginary flux; dynamics still feel where that flux sits","key_machinery":"An exactly isospectral family of non-Hermitian Hamiltonians generated by the similarity transformation S = exp(∑ α_j n_j). The transformation redistributes non-reciprocity while preserving the total imaginary flux Γ = Lγ, thereby collapsing all static critical behavior onto a single gauge-invariant manifold.","core_discovery":"Extensive bulk non-reciprocity is not required for the non-Hermitian Anderson transition. A single asymmetric boundary bond that carries the entire imaginary gauge flux Γ = Lγ is sufficient to drive the localization–delocalization transition on a disordered ring. All members of the isospectral family generated by complex gauge transformations share identical spectra and identical critical exponents for every static diagnostic; only the total flux matters. Dynamics, however, remain sensitive to the spatial distribution of that flux.","pith_inferences":[],"forward_implications":[],"fun_headline_variants":["Single non-Hermitian bond drives Anderson delocalization on a ring","Total imaginary flux alone fixes static criticality in isospectral models","Isospectral family shares spectra and exponents but not dynamics","One boundary bond triggers non-Hermitian localization-delocalization","Static universality depends only on total flux; dynamics feel its placement"],"cache_read_input_tokens":16512,"weakest_assumption_plain":"That an exponentially large hopping amplitude on one bond, required to keep the total flux fixed while concentrating it, still counts as a strictly local physical perturbation whose finite-size scaling defines the same universality class as the uniform bulk model.","fun_headline_variants_meta":{"raw":{"variants":["Single non-Hermitian bond drives Anderson delocalization on a ring","Total imaginary flux alone fixes static criticality in isospectral models","Isospectral family shares spectra and exponents but not dynamics","One boundary bond triggers non-Hermitian localization-delocalization","Static universality depends only on total flux; dynamics feel its placement"]},"model":"grok-4.5","effort":"low","cost_usd":0.007918,"raw_usage":{"total_tokens":1839,"prompt_tokens":779,"num_sources_used":0,"completion_tokens":89,"cost_in_usd_ticks":79180000,"prompt_tokens_details":{"text_tokens":779,"audio_tokens":0,"image_tokens":0,"cached_tokens":0},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":971,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":779,"tokens_out":89,"duration_ms":8523,"temperature":1.0,"reasoning_tokens":971,"cache_read_input_tokens":0,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-12T07:18:16.758722+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Measure the complex-eigenvalue fraction or winding number versus disorder for both the uniform Hatano–Nelson ring and a single-bond ring engineered with the same total flux; if the critical disorder strengths or the extracted exponent ν fail to coincide, the claim of static gauge invariance is false.","supporting_citations":[],"review_version":1}