{"id":"7c155350-e52a-4ccf-bd3d-9ad7768273e3","arxiv_id":"1908.04031","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"For a disordered XY spin chain with power-law interactions, exact diagonalization and finite-size scaling give a critical interaction exponent alpha_c = 1.16 ± 0.17, below which many-body localization is predicted to disappear.","lead":"This paper uses exact computer calculations on chains of up to 18 quantum spins to map where disorder-induced localization disappears when interactions are long-range. It estimates a critical interaction range exponent of about 1.16, below which the localized phase may not exist at all, and uses this to settle a disagreement between two earlier theoretical studies.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The existence of alpha_c≈1.16 rests on an extrapolated power-law fit to Wc(alpha) and nu(alpha) using only data above alpha_f; no measured point approaches alpha_c, and alternative functional forms are not tested, so the 'no MBL below alpha_c' conclusion is not supported.","rationale":"The reader's weakest_assumption identifies the extrapolated power-law divergence as the load-bearing premise. My analysis agrees: the central claim 'no MBL for alpha < alpha_c' depends on Eq. (3) with free alpha_c, fitted only to data above a data-dependent cutoff, and the data points nearest the divergence have such large errors (Wc = 21 +/- 8 at alpha = 1.2) that the power-law form is not distinguished from alternatives. The paper is transparent about this limitation and provides bootstrap errors for the fit parameters, but those errors are conditional on the assumed functional form. The Harris-bound discussion and the super-linear growth of delta_S_E at alpha = 2.5 are supporting observations at large alpha, but they do not pin down alpha_c. For alpha = 0.5, the peak-position analysis suggests no MBL, but there is no scaling fit between alpha = 0.5 and alpha = 1.0, so the location of the boundary remains an extrapolation. The concrete model-comparison test would settle whether the data actually require divergence at 1.16 or are equally consistent with alpha_c = 1 or with a non-divergent critical disorder. Because the reader already marked the paper CONDITIONAL for essentially this reason, no verdict change is needed.","tokens_in":10934,"tokens_out":3599,"duration_ms":33764,"concrete_test":"Perform a model-comparison test on the published Wc(alpha) and nu(alpha) data (Fig. 6): fit (i) Eq. (3) with free alpha_c, (ii) Eq. (3) with alpha_c fixed to 1, (iii) an exponential form Wc = A exp(b/(alpha - 1)), and (iv) a power law with a finite offset, all using the same alpha_f = 1.3 (for Wc) and alpha_f = 1.2 (for nu) and the same bootstrap resampling. Compare chi^2/dof and AIC. If any alternative functional form fits comparably or better (e.g., Delta AIC < 2), then the claim that MBL is absent for alpha < 1.16 is not established, because the data would also be consistent with alpha_c = 1 or with a non-divergent critical disorder near alpha = 1.2. As an additional check, repeat the fits including the alpha = 1.2 point for Wc; if alpha_c shifts by more than the quoted 0.17 uncertainty, the cutoff choice is decisive.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim, Eq. (4), is obtained by fitting Eq. (3), eta = A (alpha - alpha_c)^(-gamma), to data with a data-dependent lower cutoff: alpha_f = 1.3 for Wc and alpha_f = 1.2 for nu (Sec. IV and Appendix B). Thus the fit uses only points with alpha >= 1.3 (or 1.2), while the quoted alpha_c is 1.16; even the excluded alpha = 1.2 point (Wc = 21 +/- 8, nu = 4.2 +/- 1.1) lies far from any evident divergence. The power-law form in Eq. (3) is an assumption, not a derivation: a finite Wc at alpha = 1.2 with a sharp crossover, an exponential divergence, or a divergence at alpha = 1 (as suggested by the dynamics simulation cited as Ref. [31]) would all be consistent with the plotted points within their error bars. The bootstrap errors quoted in Eq. (4) propagate the given Wc and nu uncertainties through the assumed power-law form; they do not test the form itself. The paper itself concedes that 'the divergence ... can not fully manifest itself' and chooses alpha_f by minimizing fitting errors, which risks overfitting. Therefore the thermodynamic statement 'MBL is absent for alpha < alpha_c' is an extrapolation beyond the data, and the single most load-bearing assumption is the power-law divergence with a fitted alpha_c.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports exact diagonalization results for a disordered one-dimensional XY spin chain with power-law interactions V_ij ∝ |i-j|^{-α}, for chain lengths up to L=18. Using three diagnostics (gap statistics, half-chain entanglement entropy, and its uncertainty), the authors perform finite-size scaling to extract the critical disorder W_c(α) and correlation-length exponent ν(α). They then fit these quantities to a power-law divergence, Eq. (3), obtaining a common critical interaction exponent α_c = 1.16 ± 0.17, and conclude that for α < α_c many-body localization is absent for any disorder strength. This result is presented as resolving a discrepancy between a perturbative prediction α_c = 3/2 and a recent dynamics simulation suggesting α_c ≈ 1.","tokens_in":11254,"tokens_out":4127,"duration_ms":40824,"significance":"If correct, the result would provide the first exact-diagonalization-based determination of α_c for the long-range XY chain and would support the dynamics-based value α_c ≈ 1 over the perturbative 3/2. The paper's strengths are its systematic use of three independent diagnostics, the transparent description of the finite-size scaling and the bootstrap error analysis in Appendix C, and the honest discussion of the difficulties at small α (e.g., the lack of a crossing for α=0.5). The main weakness is that the headline quantity α_c is not measured directly but is an extrapolated zero of a power-law fit over a restricted range, so the load-bearing assumption must be tested more thoroughly before the thermodynamic claim can be considered established.","major_comments":[{"comment":"The central result α_c = 1.16 ± 0.17 is obtained by fitting the assumed power-law form η(α) = A_η (α - α_c)^{-γ_η} to W_c(α) and ν(α). This fit cannot by itself establish that a divergence occurs at a finite α_c, since the functional form already contains a divergence as an input. The authors should present a comparison with alternative models (e.g., an exponential divergence, a divergence at α=1, or a sharp crossover to a finite W_c) and demonstrate that α_c is robust, or alternatively soften the conclusion that MBL is absent for all α < α_c.","section":"Sec. IV, Eq. (3)"},{"comment":"The fit is performed only over the data with α > α_f, with α_f chosen to minimize fitting errors (α_f = 1.3 for W_c and 1.2 for ν). This choice excludes the α = 1.2 point for W_c (W_c = 21 ± 8) and means that no measured point lies within the fitted range of the inferred α_c ≈ 1.16. The divergence is therefore an extrapolation beyond the data, not a direct observation. The authors should show the sensitivity of α_c to the choice of α_f and discuss whether the excluded points are consistent with the fitted form.","section":"Sec. IV and Appendix B"},{"comment":"The bootstrap resampling propagates the statistical errors in W_c(α) and ν(α) through the assumed power-law form, but it does not test the validity of that form or the systematic errors in the finite-size scaling at marginal parameters such as α = 1.2, where W_c = 21 ± 8 and ν = 4.2 ± 1.1. The very large uncertainties near the putative α_c should be interpreted as a warning that the data are nearly insensitive to the divergence; the discussion should explicitly acknowledge that the bootstrap errors do not include model-form uncertainty.","section":"Appendix C"}],"minor_comments":[{"comment":"The citation placeholder '[ ? ]' is unresolved and should be replaced with a proper reference.","section":"Sec. IV, paragraph on universality class"},{"comment":"The word 'Poisson' is misspelled as 'Possion' in the phrase defining the localized limit.","section":"Sec. III, text near Eq. (2)"},{"comment":"The word 'enegenstates' should be 'eigenstates'.","section":"Ref. [30]"},{"comment":"The notation α_{c,W} and α_{c,ν} is inconsistent with the main text's α_{c,Wc} and α_{c,ν}; please use a uniform notation.","section":"Appendix C"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is honest about its limitations, but the central extrapolation may not survive more stringent tests. Given the current evidence, a major revision requiring alternative fits and robustness checks seems appropriate. The missing reference placeholder in Sec. IV should be fixed before publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Good to have a look at this one. The genuinely new thing is the first systematic ED finite-size scaling of the long-range XY chain across alpha, with sizes up to L=18, using three standard MBL diagnostics that give consistent crossings and collapses. The error handling is transparent, and the paper is honest about the Harris bound and about the divergence not fully manifesting at accessible sizes. That part is solid work.\n\nThe soft spot is exactly where the stress-test note points. The headline alpha_c = 1.16 ± 0.17 is not a measured divergence. It comes from fitting Eq. (3), a power-law divergence, to Wc(alpha) and nu(alpha) using only points above alpha_f = 1.3 (for Wc) or 1.2 (for nu), so the closest measured point (alpha = 1.2) has Wc = 21 ± 8 and nu = 4.2 ± 1.1—far from any evident divergence. The bootstrap errors propagate the scatter in Wc and nu but not the choice of functional form. A finite Wc with a sharp crossover, an exponential divergence, or a divergence at alpha = 1 would all be consistent with the plotted points. The paper itself concedes the divergence cannot fully manifest itself, and alpha_f is chosen to minimize fitting errors. So the statement \"no MBL for alpha < alpha_c\" is an extrapolation beyond the data.\n\nThat said, the qualitative conclusion that MBL disappears at sufficiently small alpha is on firmer ground. At alpha = 0.5, the data show no crossing in <r> or SE/ST, and the peak position of delta_SE grows at least linearly with L, which directly suggests an infinite Wc in the thermodynamic limit. So the paper's qualitative claim is supported; the quantitative alpha_c is a reasonable estimate but not a demonstrated threshold.\n\nThe reference list has one unresolved citation placeholder, a minor issue. No code or data tables are provided, which would help future checks of the fitting procedure.\n\nWho benefits: people working on MBL with power-law interactions and trapped-ion simulators will likely cite this as the ED reference for the XY case. It deserves peer review; a good referee will ask for tests of alternative functional forms, more alpha values near the claimed alpha_c, and ideally release of data and fitting code.","headline":"A solid, honest ED study whose qualitative conclusion—no MBL at sufficiently small alpha—holds up, but the headline alpha_c is an extrapolated power-law fit, not a measured divergence.","tokens_in":11779,"tokens_out":3453,"would_cite":true,"duration_ms":33632,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A disordered XY spin chain with interactions decaying as $1/r^\\alpha$ stops being many-body localized below $\\alpha_c \\simeq 1.16$.","keywords":["many-body localization","XY spin chain","power-law interactions","exact diagonalization","finite-size scaling","random transverse field","entanglement entropy","critical exponent"],"falsifier":"Compute $W_c(\\alpha)$ and $\\nu(\\alpha)$ with exact diagonalization or tensor-network methods for larger chains and for $\\alpha=1.0,1.1,1.15,1.2$: if the scaling-collapse estimates saturate or fail to keep increasing, the inferred $\\alpha_c$ is a fitting artifact. A direct time-evolution probe at $\\alpha=1.1$ with strong disorder on a chain comparable to trapped-ion experiments would also settle the matter, since persistent density imbalance or logarithmic entanglement growth would indicate MBL below the claimed threshold.","tokens_in":1856,"feed_emoji":"🧲","tokens_out":1765,"duration_ms":94167,"temperature":0.7,"pith_summary":"This paper asks whether a one-dimensional XY spin chain with interactions decaying as $V_{ij}\\propto 1/|i-j|^\\alpha$ and a random transverse field can still exhibit many-body localization (MBL), the absence of thermalization in a disordered quantum system. Using exact diagonalization for chains of up to 18 spins, the authors extract the critical disorder strength $W_c(\\alpha)$ and the correlation-length exponent $\\nu(\\alpha)$ from finite-size scaling of gap statistics, half-chain entanglement entropy, and entropy uncertainty. Both $W_c$ and $\\nu$ diverge as $\\alpha$ decreases toward $\\alpha_c\\simeq 1.16\\pm 0.17$, which they interpret as the disappearance of MBL for $\\alpha<\\alpha_c$ in the thermodynamic limit. The result matters because it sits between two prior predictions, $\\alpha_c=3/2$ from a perturbative argument and $\\alpha_c\\approx 1$ from quantum dynamics, and because trapped-ion experiments realize the same model.","feed_headline":"Below alpha ≈ 1.16, long-range spin chains never localize","feed_subtitle":"Exact diagonalization finds the critical disorder and exponent diverge near 1.16, settling a dispute between 3/2 and 1.","key_machinery":"The load-bearing object is the finite-size scaling collapse of the normalized half-chain entanglement entropy $S_E/L$ through the ansatz $S_E(L,W)=L f[(W-W_c)L^{1/\\nu}]$, where the correlation length behaves as $\\xi(W)\\propto |W-W_c|^{-\\nu}$. For each exponent $\\alpha$, the paper locates the transition by collapsing data at chain lengths $L=12,14,16,18$; the same $W_c$ and $\\nu$ also collapse the averaged gap ratio $\\langle r\\rangle$. The extrapolated $W_c(\\alpha)$ and $\\nu(\\alpha)$ are then fitted to a power-law divergence $\\eta=A_\\eta(\\alpha-\\alpha_{c,\\eta})^{-\\gamma_\\eta}$, with the lower cutoff $\\alpha_f$ chosen by minimizing fitting error and uncertainties handled by bootstrap resampling. This two-stage machinery turns finite-size crossing data into a critical interaction exponent $\\alpha_c$.","core_discovery":"Using the finite-size scaling ansatz $S_E(L,W)=L f[(W-W_c)L^{1/\\nu}]$ for the normalized half-chain entanglement entropy, together with matching collapses for the spectral gap ratio, the authors obtain $W_c$ and $\\nu$ for interaction exponents from $\\alpha=1.0$ to $2.5$. Fitting those results to $\\eta(\\alpha)=A_\\eta(\\alpha-\\alpha_{c,\\eta})^{-\\gamma_\\eta}$, with a lower cutoff on the fitted range chosen to minimize errors, gives $\\alpha_{c,W_c}=1.16\\pm 0.17$ and $\\alpha_{c,\\nu}=1.17\\pm 0.14$; the paper reports these conservatively as $\\alpha_c=1.16\\pm 0.17$. It finds no singular behavior at $\\alpha=3/2$ and concludes that below $\\alpha_c$ the system cannot be many-body localized at any disorder strength in the limit $L\\to\\infty$.","pith_inferences":["The divergence of $W_c$ and $\\nu$ is inferred from fitted data with $\\alpha\\ge 1.2$; if larger-system calculations find that $W_c(\\alpha)$ bends over instead of diverging below $\\alpha\\approx 1.2$, the true threshold could be lower or absent altogether.","A direct scaling collapse in $\\alpha$ at fixed strong disorder, using the form $S_E/L = h[L^{1/\\nu}(\\alpha-\\alpha_c)]$, would provide a second, independent route to $\\alpha_c$ that does not rely on the power-law fit to $W_c$.","The same finite-size scaling pipeline could be applied to Heisenberg chains with power-law interactions, where a separate prediction sets $\\alpha_c=2$, testing whether the mechanism behind $\\alpha_c\\simeq 1.16$ is specific to the XY symmetry.","Trapped-ion experiments at $\\alpha\\sim 1$ could test the thermodynamic-limit claim directly: observing persistent density imbalance or slow entanglement growth at strong disorder would contradict the predicted absence of MBL below $\\alpha_c$."],"forward_implications":["If the central claim is right, a disordered one-dimensional XY chain with $1/r^\\alpha$ interactions has no MBL phase in the thermodynamic limit for any disorder strength when $\\alpha<\\alpha_c\\simeq 1.16$.","The predicted $\\alpha_c=3/2$ from resonant spin-pair arguments is ruled out, since the paper sees $W_c(\\alpha)$ and $\\nu(\\alpha)$ vary smoothly across $\\alpha=1.5$.","The earlier quantum-dynamics estimate $\\alpha_c\\approx 1$ is supported by an independent equilibrium, spectrum-based exact-diagonalization method.","At large $\\alpha$ the extracted exponent $\\nu\\approx 1$ matches finite-size studies of short-range MBL, suggesting the long-range transition lies in the same universality class.","At small $\\alpha$ (for example $\\alpha=0.5$) the peak of the entropy uncertainty grows at least linearly with system size, pointing to $W_c\\to\\infty$ and hence no MBL transition."],"supporting_citations":[{"why":"Introduces the averaged ratio of successive gaps that distinguishes GOE from Poisson statistics, the paper's primary level-spacing diagnostic.","marker":"[10]"},{"why":"Establishes the many-body localization phase transition and the entanglement-entropy diagnostics used to locate $W_c$ in finite chains.","marker":"[11]"},{"why":"Supplies the finite-size scaling ansatz for the half-chain entanglement entropy and the short-range reference values for comparison.","marker":"[12]"},{"why":"Gives the perturbative prediction $\\alpha_c=3/2$ from resonant spin-pair excitations that the paper's result is set against.","marker":"[26]"},{"why":"Provides the quantum-dynamics estimate $\\alpha_c\\approx 1$ for the same model that the paper's exact-diagonalization result supports.","marker":"[31]"}],"fun_headline_variants":["Long-range spin chains never localize for α below ≈1.16","α_c ≈ 1.16: critical threshold for MBL breakdown","α_c = 1.16 settles dispute over long-range MBL cutoff","Long-range interactions: no localization below α ≈ 1.16"],"cache_read_input_tokens":13824,"weakest_assumption_plain":"The load-bearing premise is that $W_c(\\alpha)$ and $\\nu(\\alpha)$ truly diverge according to the fitted power law down to $\\alpha_c\\simeq 1.16$, even though every measured point has $\\alpha\\ge 1.0$ and the fits use only data above $\\alpha_f=1.3$ (for $W_c$) or $1.2$ (for $\\nu$).","fun_headline_variants_meta":{"raw":{"variants":["Long-range spin chains never localize for α below ≈1.16","α_c ≈ 1.16: critical threshold for MBL breakdown","α_c = 1.16 settles dispute over long-range MBL cutoff","Long-range interactions: no localization below α ≈ 1.16"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001096,"raw_usage":{"total_tokens":4622,"prompt_tokens":1039,"completion_tokens":3583,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":655,"completion_tokens_details":{"reasoning_tokens":3504}},"tokens_in":655,"tokens_out":3583,"duration_ms":26045,"temperature":1.0,"reasoning_tokens":3504,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:54:14.376629+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute $W_c(\\alpha)$ and $\\nu(\\alpha)$ with exact diagonalization or tensor-network methods for larger chains and for $\\alpha=1.0,1.1,1.15,1.2$: if the scaling-collapse estimates saturate or fail to keep increasing, the inferred $\\alpha_c$ is a fitting artifact. A direct time-evolution probe at $\\alpha=1.1$ with strong disorder on a chain comparable to trapped-ion experiments would also settle the matter, since persistent density imbalance or logarithmic entanglement growth would indicate MBL below the claimed threshold.","supporting_citations":[{"cited_title":"Oganesyan and D","cited_arxiv_id":null,"evidence_quote":"Introduces the averaged ratio of successive gaps that distinguishes GOE from Poisson statistics, the paper's primary level-spacing diagnostic."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the finite-size scaling ansatz for the half-chain entanglement entropy and the short-range reference values for comparison."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the perturbative prediction $\\alpha_c=3/2$ from resonant spin-pair excitations that the paper's result is set against."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the quantum-dynamics estimate $\\alpha_c\\approx 1$ for the same model that the paper's exact-diagonalization result supports."}],"review_version":1}