{"id":"e2c11720-3740-46ed-a265-5e90014e18af","arxiv_id":"2412.13518","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Long-range hopping, previously thought to suppress reentrant localization, can instead induce it in both staggered and uniform disorder regimes, with four transition points showing distinct critical exponents.","lead":"This paper studies a one-dimensional lattice model with quasi-periodic disorder and long-range hopping, and finds that long-range hopping can bring back a reentrant localization transition where states localize, delocalize, and localize again as disorder grows. It matters because it challenges the earlier view that long-range hopping only destroys reentrant localization, and it adds a uniform-disorder case where such transitions were previously thought absent.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claim of four distinct universality classes rests on correlation-length exponents with no quoted uncertainties, extracted by a non-specified collapse procedure in post-hoc eigenstate windows; the reported differences may be within numerical error.","rationale":"The reader's weakest assumption identifies exactly the same load-bearing concern: the distinct-universality-classes claim depends on ν values obtained from a collapse procedure with no error bars and on post-hoc eigenstate windows. I read the paper in good faith and find the primary reentrant-localization evidence credible. The phase diagrams, the eigenstate-resolved IPR/NPR maps, the direct spatial distributions at selected λ, and the multi-size insets all support the central observation that long-range hopping can induce reentrant localization under both staggered and uniform disorder. That part of the argument does not need to be rejected. However, the abstract and conclusion go further and assert that the four transitions belong to distinct universality classes. That assertion is not secured by the presented analysis. The R-function gives γ/ν with small formal errors, but the determination of ν is descriptive ('adjusting' for optimal overlap) rather than a reproducible optimization, and no uncertainty is quoted for ν. The reported ν values are close enough that the classification into different universality classes is fragile. The additional use of different eigenstate windows for different transitions introduces a selection effect that can change both the crossing point and the extracted exponent. I also note that the hyperscaling check is not independent, since β/ν and γ/ν are derived from the same NPR scaling data. A concrete independent collapse with bootstrap uncertainties, or a transfer-matrix calculation, would settle whether the four ν values are truly distinct. Given the strength of the reentrant-localization evidence and the fragility of the universality-class claim, the reader's CONDITIONAL verdict remains appropriate; I would not change it to ACCEPT or REJECT based on this review. No code or data are provided, so independent verification is especially important here, but that is a reproducibility limitation rather than a separate scientific objection.","tokens_in":16786,"tokens_out":11747,"duration_ms":109781,"concrete_test":"Perform a standardized finite-size scaling analysis at each of the four λc values: compute ν by minimizing a defined collapse residual for Eq. 10 (for example, the normalized mean-square distance of the rescaled data to a master curve), with bootstrap resampling over the available system sizes and over eigenstate windows shifted by ±0.005 in m/L. Alternatively, extract ν independently using transfer-matrix localization-length scaling for the same parameters. If the 95% confidence intervals for the four ν values overlap, or if shifting the eigenstate window changes ν by more than the reported differences, the claim of four distinct universality classes should be withdrawn and the paper should be revised to claim only that long-range hopping modifies the critical behavior.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The reentrant-localization observation itself is reasonably supported: the η phase diagrams, eigenstate-resolved IPR/NPR panels, spatial distributions, and the size dependence in the insets of Figs. 4 and 11 all point to genuine reentrant behavior. The load-bearing weakness is the stronger claim, stated in the abstract and conclusion, that the four transitions belong to distinct universality classes. In Sec. III, γ/ν is obtained from R[L,L′] crossings with small formal errors, but ν is then obtained by 'adjusting' ν in Eq. 10 until curves for L = 8362, 13530, 21892, 35422 overlap optimally, with no collapse metric, no goodness-of-fit, no bootstrap, and no sensitivity analysis. The staggered values ν1 = 0.9, ν2 = 1.1, ν3 = 1.8, ν4 = 0.9 are separated by amounts comparable to typical scaling-collapse uncertainties for this method, and the uniform-disorder values ν1 = 0.814, ν2 = 1.10, ν3 = 0.938, ν4 = 0.680 are even closer. The analysis also selects different eigenstate windows for different transitions (m/L ∈ [0.238, 0.243] for the first, m/L ∈ [0.48, 0.52] for the others), which can bias both λc and ν if the window is not justified independently. Finally, because β/ν and γ/ν are both fit from the same NPR-derived data, the stated verification of the hyperscaling relation (Eq. 12) is not an independent check. If the true uncertainties on ν are ≥0.2, the universality-class conclusion is unsupported even though the reentrant phase diagram survives.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript studies a one-dimensional dimerized (SSH) chain with next-nearest-neighbor hopping and a quasi-periodic potential, under either staggered (λA = −λB) or uniform (λA = λB) disorder. Using exact diagonalization up to L = 35422, it computes IPR/NPR-based diagnostics and η phase diagrams. It reports that with J11 = J3 = 0.1 and J33/J1 = 2.7 (staggered case) the system shows two critical regions, and with J33/J1 = 0.7 (uniform case) the band-edge states show localized–extended–localized reentrant behavior. It also extracts critical exponents (λc, γ/ν, β/ν, ν) at four transitions in each case and concludes that the transitions belong to distinct universality classes.","tokens_in":17169,"tokens_out":5409,"duration_ms":47045,"significance":"The reentrant localization observation is potentially interesting because it extends reentrant behavior to a model with next-nearest-neighbor hopping and to uniform disorder, where previous work (e.g., Refs. [47,65]) suggested it is absent or requires large detuning. The finite-size checks in the insets of Figs. 4(b) and 11 support the persistence of the reentrant feature. However, the universality-class conclusion is not supported by the current analysis: the ν values are obtained by a visually adjusted scaling collapse without quoted uncertainties, in post-hoc eigenstate windows, and the hyperscaling check is not independent. The paper's main numerical evidence for reentrant localization is credible, but the stronger claim about distinct universality classes needs substantial additional analysis.","major_comments":[{"comment":"The correlation-length exponent ν is determined by 'adjusting' ν to achieve 'optimal overlap' of the scaling curves σ²L^{1−γ/ν} versus εL^{1/ν}, with no collapse metric, no goodness-of-fit, and no uncertainty estimate. The reported values (ν1 = 0.9, ν2 = 1.1, ν3 = 1.8, ν4 = 0.9) are separated by amounts comparable to typical scaling-collapse uncertainties for this method, so the claim that the four transitions belong to distinct universality classes is not supported. Please provide error bars, for example via bootstrap or by reporting the sensitivity of ν to the range of system sizes and to the eigenstate window.","section":"Sec. III, Eq. (10) and Fig. 7"},{"comment":"The same issue applies to the uniform-disorder case, where the ν values (ν1 = 0.814, ν2 = 1.10, ν3 = 0.938, ν4 = 0.680) are even closer together. In addition, the eigenstate windows used for the average NPR are selected post hoc (m/L ∈ [0.167, 0.177] for the first transition and m/L ∈ [0, 0.1] for the others), with no independent justification. This can bias both λc and ν. The analysis should demonstrate that the extracted exponents are stable under reasonable variations of these windows.","section":"Sec. IV, Fig. 13"},{"comment":"The hyperscaling relation 2β/ν + γ/ν = 1 is presented as a confirmation, but both β/ν and γ/ν are obtained by least-squares fits to the same NPR-derived data, so the check is not independent. Moreover, the quoted uncertainties (e.g., γ1/ν1 = 0.5735 ± 8e−05) are only fit standard errors and do not include systematic uncertainties from finite-size corrections or eigenstate-window choices; they should not be used to argue for the precision of the exponents. Please either obtain one of the exponents from an independent quantity or explicitly state that Eq. (12) is a consistency condition rather than an independent verification.","section":"Sec. III, Eq. (12) and Fig. 8"},{"comment":"The finite-size scaling analysis assumes a single correlation-length exponent ν for each transition, but the identified 'critical regions' are intervals where localized and extended states coexist (single-particle mobility edges), not simple critical points. The R-function crossing method is designed for a transition at a single λc. The manuscript should justify its applicability to these mobility-edge transitions or test it with a direct scaling collapse that includes the energy dependence explicitly.","section":"Sec. II, Eqs. (8)–(10)"}],"minor_comments":[{"comment":"The captions state that the color represents log10(η), but η is already defined as log10(IPR × NPR) in Eq. (11), so the label appears to be a double logarithm. Please clarify whether the color bar shows η or log10(IPR × NPR).","section":"Figs. 2, 3, 9, 15, 16 captions"},{"comment":"The finite-size check is shown only for the average NPR; the text says this rules out finite-size effects for the reentrant localization feature, but the corresponding IPR behavior for the second critical region is not shown for all sizes. Please include IPR or another order parameter in the finite-size analysis.","section":"Sec. III, Fig. 4(b) inset"},{"comment":"The term 'long-range hopping' is used for next-nearest-neighbor hopping. This is consistent with some prior literature, but since it can be confused with power-law or algebraically decaying hopping, a brief clarification in the introduction would help.","section":"Sec. I"},{"comment":"The definition of NPR(m) appears to have a formatting issue with parentheses; please ensure the equation reads NPR(m) = ( L ∑_i |φ_i^{(m)}|^4 )^{-1}.","section":"Eq. (5)"},{"comment":"The text refers to 'Fig 3(c)', but Fig. 3 has only panels (a) and (b); the intended reference is likely Fig. 3(b). Similarly, 'Fig 9(c)' is correct.","section":"Sec. V, first paragraph"},{"comment":"The phrase 'In a word that long-range hopping introduces additional pairs of mobility edges' is grammatically awkward; please rephrase.","section":"Sec. V, paragraph after Fig. 15"},{"comment":"The phrase 'each characterized by unique critical exponents' is ambiguous; 'distinct critical exponents' would be clearer, since 'unique' might be read as 'each having exactly one exponent', which is not the intended meaning.","section":"Sec. VI"}],"recommendation":"major_revision","confidential_remarks":"The central reentrant-localization observation appears to be supported by the numerical evidence, and the paper is a reasonable candidate for publication after revision. However, the universality-class claim is the headline result and currently rests on a non-quantified scaling-collapse procedure. I would ask the authors to provide error bars or a robustness analysis for ν, and to address the post-hoc eigenstate window issue, before this claim can be accepted. The hyperscaling check should be reframed as a consistency condition rather than a confirmation."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's the take: the paper's main observation is real—long-range hopping can induce reentrant localization in this 1D quasi-periodic SSH model, both under staggered disorder (J33/J1=2.7) and under uniform disorder (J33/J1=0.7). That is new relative to Refs. [47,65,69], and the finite-size checks up to L=35422 give it real weight. The phase diagrams, IPR/NPR colormaps, and spatial distributions are consistent and convincing.\n\nWhat it does well: it corrects the simple picture that long-range hopping only suppresses reentrant localization. The ladder reformulation is a plausible way to understand why the phenomenon can reappear. The two pairs of mobility edges in the staggered case are also a concrete, checkable finding. Honest, careful work.\n\nWhere it goes soft: the claim in the abstract and conclusion that the four transitions belong to distinct universality classes. The ν values (0.9, 1.1, 1.8, 0.9) come from “adjusting” ν until curves overlap, with no quoted uncertainty, no collapse metric, no bootstrap. Typical scaling-collapse uncertainties for this method are 0.2 or larger; the differences here are barely bigger than that. The uniform-disorder values are even closer. The eigenstate windows are chosen after inspection (e.g., m/L in [0.238,0.243]), which can bias both λc and ν at the level of the claimed differences. And the hyperscaling check is not independent: β/ν and γ/ν come from the same NPR-derived data, so Eq. 12 is satisfied by construction, not verified.\n\nThat said, none of this undermines the reentrant localization observation itself. The exponent story is an overreach, not a fatal flaw. If the authors either provide a proper collapse analysis with error bars (bootstrap or least-squares) or soften the universality-class language, the paper is in good shape.\n\nI would send this to peer review. The main result is worth publishing, and the exponent claims can be fixed through normal refereeing. Who benefits: anyone working on 1D localization, mobility edges, or the interplay of dimerization and quasiperiodic disorder. I'd cite it for the uniform-disorder result. Bring to reading group? Maybe—it is a useful example of a numerical localization paper where the physics is solid but the FSS claims need scrutiny.","headline":"Reentrant localization finding is solid and new; the four-universality-class claim overreaches the data.","tokens_in":17728,"tokens_out":2262,"would_cite":true,"duration_ms":20719,"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":"Adding long-range hopping to a quasi-periodic chain can make reentrant localization appear, not just suppress it.","keywords":["reentrant localization","quasi-periodic lattice","long-range hopping","Su-Schrieffer-Heeger model","mobility edge","finite-size scaling","universality class","inverse participation ratio"],"falsifier":"A transfer-matrix calculation of the localization length at the four claimed critical points, with bootstrap error bars on the correlation-length exponent, would settle the universality-class claim: if the four exponents overlap within uncertainty, the claim fails. For the reentrant phenomenon itself, a waveguide experiment with engineered next-nearest-neighbor hopping could look for the localized-extended-localized sequence in band-edge states as the disorder amplitude is swept.","tokens_in":16575,"feed_emoji":"🌀","tokens_out":11794,"duration_ms":96076,"temperature":0.7,"pith_summary":"Reentrant localization is the sequence in which stronger disorder first localizes a state, then lets it become extended again, and then localizes it once more. The paper asks whether this counterintuitive sequence survives when the chain also contains long-range (next-nearest-neighbor) hopping, which earlier work suggested only weakens reentrant behavior. It argues that, with the right hopping parameters, long-range hopping actually induces reentrant localization, for both staggered and uniform disorder, and that the four localization transitions in each case belong to distinct universality classes. If correct, this widens the class of one-dimensional quasi-periodic systems in which mobility edges and reentrant phases can be engineered.","feed_headline":"Long-range hopping creates reentrant localization, not just weakens it","feed_subtitle":"In a 1D quasi-periodic chain, disorder can localize, spread, then localize states again under both disorder types.","key_machinery":"The central machinery is the dimerized SSH chain (a one-dimensional tight-binding chain with alternating nearest-neighbor hoppings) extended by next-nearest-neighbor hopping terms $J_{33}$ and $J_3$, viewed as a two-leg ladder. The quasi-periodic potential is applied with opposite signs on the two sublattices (staggered disorder) or equal signs (uniform disorder). Localization is diagnosed with the inverse participation ratio (IPR) and normalized participation ratio (NPR), combined into $\\eta=\\log_{10}(\\mathrm{IPR}\\times\\mathrm{NPR})$ to mark critical regions, and with the $R[L,L']$ function whose size-crossing gives the critical disorder strength and the exponent ratio $\\gamma/\\nu$. The correlation-length exponent $\\nu$ is then fixed by collapsing $\\sigma^2=L^{\\gamma/\\nu-1}G(\\varepsilon L^{1/\\nu})$ for different system sizes. The proposed mechanism is that each ladder leg retains the dimerization-disorder competition that drives reentrant localization, while the inter-leg hopping $J_{11}$ and $J_3$ tunes that competition, which is why long-range hopping can either suppress or create the effect.","core_discovery":"The central discovery is that long-range hopping does not simply compete with dimerization to erase reentrant localization; under specific conditions it is the ingredient that makes reentrant localization appear. For staggered disorder $\\lambda_A=-\\lambda_B$ with $J_{11}=J_3=0.1$ and $J_{33}/J_1=2.7$, the phase diagram in $\\lambda$ shows two intervals, $0.98<\\lambda<2.018$ and $2.599<\\lambda<3.6$, in which localized and extended states coexist, separated by a fully localized band; eigenstates in the middle of the spectrum re-extend as $\\lambda$ grows. For uniform disorder $\\lambda_A=\\lambda_B$ with $J_{33}/J_1=0.7$ and $J_{11}=J_3=0.1$, eigenstates near the band edges undergo localized-extended-localized reentrance in a narrow window near $\\lambda\\simeq 2.3$-$2.4$. The paper interprets the long-range chain as a two-leg ladder: reentrant localization reflects competition between dimerization and disorder within each leg, with inter-leg hopping acting as a tunable perturbation. Finite-size scaling of the $R$-function yields four critical points in each disorder class whose exponents $\\nu$ and ratios $\\gamma/\\nu$, $\\beta/\\nu$ satisfy the hyperscaling relation $2\\beta/\\nu+\\gamma/\\nu=1$, and the paper concludes the four transitions belong to distinct universality classes.","pith_inferences":["Editorially, the two-leg ladder picture predicts that at $J_{11}=J_3=0$ the spectrum is just the union of two decoupled chains; checking whether the two observed pairs of mobility edges coincide with the individual legs' mobility edges would cleanly separate a superposition effect from a genuine long-range-hopping effect.","An implication the paper leaves implicit: the reentrant window location (band center for staggered disorder, band edges for uniform disorder) could serve as a spectroscopic fingerprint of the disorder type in experimental realizations.","A testable extension is to compute the dynamics: a wave packet launched at a band-edge state should show non-monotonic spreading as $\\lambda$ is swept through the reentrant window, which would make the effect visible in time-dependent cold-atom or waveguide experiments.","The distinct-universality-class claim would be strengthened by an independent transfer-matrix computation of localization lengths, since the paper's exponents come from eigenstate windows selected after inspecting the data."],"forward_implications":["In the staggered-disorder case, long-range hopping removes the earlier coincidence of universality classes: the second and third localization transitions now have different correlation-length exponents, so each transition is its own universality class.","In the uniform-disorder case, reentrant localization appears where it was previously thought absent, with the re-extended states sitting at the highest and lowest eigenenergies rather than near the band center.","Tuning the inter-leg hopping strength $J_{11}=J_3$ controls the phenomenon: for staggered disorder reentrant localization disappears above $J_{11}=J_3\\simeq0.207$, while for uniform disorder it appears only for nonzero inter-leg hopping and vanishes above about $J_{11}=J_3\\simeq0.35$.","The four critical points in each disorder class satisfy the hyperscaling relation $2\\beta/\\nu+\\gamma/\\nu=1$, so the extracted exponents are internally consistent and describe genuine transitions."],"supporting_citations":[{"why":"supplies the baseline nearest-neighbor staggered-disorder model where reentrant localization was first identified, which this paper extends to long-range hopping.","marker":"[47]"},{"why":"reports that long-range hopping weakens reentrant localization, the prior claim this paper qualifies by showing parameter regimes where long-range hopping induces it.","marker":"[69]"},{"why":"provides the earlier finding that the second and third transitions share a universality class and introduces the critical-exponent analysis that this paper revisits.","marker":"[72]"},{"why":"gives the R-function and finite-size scaling relation used to extract critical disorder strengths and exponents.","marker":"[73]"},{"why":"is the earlier statement that uniform disorder with small detuning shows no reentrant localization, the claim the uniform-disorder results directly challenge.","marker":"[65]"},{"why":"defines the $\\eta$ diagnostic used to map critical regions in the phase diagrams.","marker":"[74]"}],"fun_headline_variants":["Long-range hopping triggers reentrant localization in 1D quasiperiodic chains","Reentrant localization emerges via long-range hopping in disordered 1D lattices","Long-range hopping induces reentrant localization for both disorder types","Long-range hopping alters universality classes in reentrant localization","Long-range hopping doesn't ruin localization; it induces reentrance"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The conclusion that the four transitions are genuinely different in character rests on matching numerical curves at different system sizes, and the key fitted number is quoted without an error bar.","fun_headline_variants_meta":{"raw":{"variants":["Long-range hopping triggers reentrant localization in 1D quasiperiodic chains","Reentrant localization emerges via long-range hopping in disordered 1D lattices","Long-range hopping induces reentrant localization for both disorder types","Long-range hopping alters universality classes in reentrant localization","Long-range hopping doesn't ruin localization; it induces reentrance"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000754,"raw_usage":{"total_tokens":3382,"prompt_tokens":1001,"completion_tokens":2381,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":617,"completion_tokens_details":{"reasoning_tokens":2286}},"tokens_in":617,"tokens_out":2381,"duration_ms":16473,"temperature":1.0,"reasoning_tokens":2286,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T13:08:07.499907+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A transfer-matrix calculation of the localization length at the four claimed critical points, with bootstrap error bars on the correlation-length exponent, would settle the universality-class claim: if the four exponents overlap within uncertainty, the claim fails. For the reentrant phenomenon itself, a waveguide experiment with engineered next-nearest-neighbor hopping could look for the localized-extended-localized sequence in band-edge states as the disorder amplitude is swept.","supporting_citations":[{"cited_title":"Jiang, Y","cited_arxiv_id":null,"evidence_quote":"reports that long-range hopping weakens reentrant localization, the prior claim this paper qualifies by showing parameter regimes where long-range hopping induces it."},{"cited_title":"Vaidya, C","cited_arxiv_id":null,"evidence_quote":"provides the earlier finding that the second and third transitions share a universality class and introduces the critical-exponent analysis that this paper revisits."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"gives the R-function and finite-size scaling relation used to extract critical disorder strengths and exponents."},{"cited_title":"Signature of localization-delocalization in collisional inhomogeneous spin-orbit coupled condensates","cited_arxiv_id":"2403.02027","evidence_quote":"is the earlier statement that uniform disorder with small detuning shows no reentrant localization, the claim the uniform-disorder results directly challenge."},{"cited_title":"Hashimoto, K","cited_arxiv_id":null,"evidence_quote":"defines the $\\eta$ diagnostic used to map critical regions in the phase diagrams."}],"review_version":1}