{"id":"16a8e7f9-64be-47fa-984b-505e8c12cae8","arxiv_id":"1908.04039","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"The many-body localization transition in a disordered Ising chain is found to be continuous, with critical entanglement entropy consistent with the thermal volume law and correlation-length exponent nu=0.94.","lead":"This paper uses exact computer simulations of a small disordered quantum spin chain to study the transition between a thermal, ergodic phase and a many-body localized, non-thermal phase. It concludes the transition is smooth, the entanglement at the critical point is close to the thermal maximum, and the two different entropy measures give the same critical point and exponent.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The thermal-critical-point claim rests on an r→0 extrapolation whose smallest-r points have L_A=1,2; this tests L_A→0 at fixed L, not the 1≪L_A≪L thermodynamic limit.","rationale":"The reader's weakest assumption identifies the scaling ansatz Eq. (6). I partially agree: the one-length-scale ansatz is indeed the formal premise. But the more actionable weakness is in how the ansatz is converted into a number. The paper's own text stresses that S_A^E is bounded by L_A ln2 and that the L-dependence is a correction; yet the correction is parametrized only through r, and the extrapolation to r=0 is carried out with data whose smallest-r values are L_A=1 and 2. In the thermodynamic limit one wants L_A large and r small; with L≤16 these two requirements cannot be met simultaneously. The quadratic fit therefore uses exactly the points that are least like a macroscopic subsystem to fix the intercept. The fact that the fitted α exceeds ln2 is a concrete indication of this bias. A clean numerical check—drop the L_A=1,2 points or add a 1/L_A term—would settle the issue. I do not raise the Harris-bound violation or disagreement with other MBL works as a correctness concern, because the paper's exponent is close to Monthus's ν=1 and the consistency of the two probes is independent supporting evidence. The absence of code/data and a quantitative collapse metric is a reproducibility weakness, but not the main logical step. Since the reader already returned CONDITIONAL, my concern does not change the verdict; it sharpens the condition under which the paper would be accepted: the extrapolation must survive the proposed checks.","tokens_in":13143,"tokens_out":10755,"duration_ms":122288,"concrete_test":"Refit the critical amplitude f_c(r) from Fig. 6c (δJ=3.2) in three ways: (i) using all points as in the paper; (ii) excluding L_A=1 and L_A=2; (iii) adding a 1/L_A correction term, f_c(r)=α+βr+γr^2+δ/L_A. Compare the r→0 intercepts to ln2 and to the stated α=0.72±0.03. If the intercept in (ii) or (iii) shifts by more than 0.03 or no longer agrees with ln2, the strictly-thermal conclusion is an artifact of the smallest subsystems and should be weakened to 'consistent within the fit form'.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central conclusion that S_A^E is strictly thermal at criticality rests on Eq. (8): the data in Fig. 6c at δJ=3.2 are fit to f_c(r)=α+βr+γr^2, and the r→0 intercept α=0.72±0.03 is compared to ln2≈0.693. The load-bearing step is not just the scaling form in Eq. (6); it is the use of small-r data to fix the intercept. The sample includes L=10,12,14,16 with L_A from 1 to L/2, so the smallest r values are r=1/16 and 1/8, realized by L_A=1 and L_A=2 at L=16. Thus the extrapolation to r=0 is an extrapolation to L_A→0 at fixed L. The intended limit 1≪L_A≪L requires large L_A and small r simultaneously, which the accessible sizes cannot provide. A further warning sign is that the fitted α exceeds the physical bound ln2; with γ=0.84±0.62 the curvature is poorly constrained, so the intercept is set by the very points that are least representative of a macroscopic subsystem. If the true finite-size correction contains a 1/L_A term, the r-only form Eq. (8) will misestimate the intercept. Consequently, even if Eq. (6) is accepted, the 'thermal' value is not established; a fit restricted to L_A≥3 or with an explicit 1/L_A correction could move α away from ln2.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the many-body localization (MBL) transition in a disordered transverse-field Ising chain using exact diagonalization with twisted boundary conditions. The authors introduce a scaling form for the subsystem entanglement entropy, S_A^E(L_A,r,δJ) = L_A f(L_A|δJ−δJ_c|^ν, r), where r=L_A/L, and determine the critical disorder and correlation-length exponent from finite-size scaling collapses. They further extract the r→0 limit of the critical entanglement amplitude from a polynomial fit and compare it with ln2, concluding that the critical eigenstates are thermal and that the transition is continuous. They then perform a similar scaling analysis for the participation entropy in a spin/domain-wall basis and report consistent critical point and exponent, suggesting that MBL can be viewed as a localization transition in configuration space.","tokens_in":13538,"tokens_out":3383,"duration_ms":37948,"significance":"If the main claims hold, the paper would provide numerical support for a continuous MBL transition with thermal critical eigenstates, consistent with Grover's argument, and would supply a correlation-length exponent ν≈0.94 that challenges the Harris bound while agreeing with Monthus's prediction. The use of twisted boundary conditions to reduce finite-size fluctuations and the independent consistency between entanglement and participation entropy are genuine strengths. However, the central thermal-volume-law conclusion rests on a small-r extrapolation that may not access the intended thermodynamic limit, and the scaling analysis lacks a quantitative collapse criterion; these issues must be addressed before the claims can be considered established.","major_comments":[{"comment":"The conclusion that the entanglement entropy is strictly thermal at the critical point rests on the fitted intercept α=0.72±0.03 in the polynomial fc(r)=α+βr+γr². The smallest-r data points in this fit correspond to L_A=1 and L_A=2 at L=16, i.e., r=1/16 and 1/8. The extrapolation to r→0 is therefore an extrapolation toward L_A→0 at fixed L, not toward the thermodynamic limit 1≪L_A≪L. Since γ=0.84±0.62 is poorly constrained, and since a plausible finite-size correction of the form 1/L_A is not included in Eq. (8), the intercept is not robustly determined. The authors should demonstrate that the result survives excluding L_A=1,2 or adding an explicit 1/L_A term; otherwise the 'strictly thermal' claim is not established.","section":"Sec. III.B, Eq. (8), Fig. 6(c)"},{"comment":"The critical point δJc=3.19±0.03 and exponent ν=0.94±0.07 are extracted from 'best data collapse' without any quantitative collapse metric or goodness-of-fit criterion. Because the same dataset is used first to fix fc(r) and then to collapse the scaled variable y=S_A^E/(L_A fc(r)), the procedure is susceptible to circularity. A quantitative collapse measure (e.g., minimization of the scatter of the collapsed curves, or a chi-square statistic) and a validation on independent synthetic data would make the error bars meaningful and the exponent determination reproducible.","section":"Sec. III.B, Eqs. (6)-(7), Fig. 7"},{"comment":"The scaling ansatz assumes that S_A^E depends on L only through r and on disorder only through the single combination L_A|δJ−δJc|^ν. However, Fig. 4 shows a clear L dependence of S_A^E near the critical region at fixed L_A. The paper interprets this as an r correction, but since r is changed by varying both L and L_A, the data do not directly test whether an independent L dependence remains. The authors should provide a test of the ansatz, for example by checking whether data with the same r but different (L,L_A) collapse, or by including an explicit correction term and showing it is negligible. Without such a test, the extracted critical point and exponent could be biased by a missing scaling field.","section":"Sec. III.B, Eq. (6)"},{"comment":"The participation entropy scaling gives δJc'=3.16±0.04 and ν'=0.89±0.03, which agree with the entanglement results. This is a valuable cross-check, but it does not by itself resolve the concerns about the entanglement scaling, because the participation entropy analysis uses the same finite-size systems and the same type of visually assessed data collapse. The authors should state explicitly whether a quantitative collapse criterion was used here as well and how the reported errors were obtained.","section":"Sec. IV, Fig. 8"}],"minor_comments":[{"comment":"There are several typographical errors, including 'bondary conditions' in the Section II.B heading, 'wether' in the Introduction, 'entanglement e ntropy' in the title, and 'participation entropy' misspelled in places. A careful proofreading pass is needed.","section":"General"},{"comment":"The caption states that 26 data points are obtained for L=10,12,14,16 and L_A=1,...,L/2, which gives 5+6+7+8=26 points; however, the text in Section III.B refers to 'all data points in Fig. 5' when determining fc(r). Clarify whether the L=8 data are included in the fit and why the caption omits L=8.","section":"Fig. 6 caption"},{"comment":"The statement that 'fc(r)<ln2 for nonzero r' in Fig. 6(c) is made without showing error bars or statistical significance on the individual points. Since the fitted γ has a large uncertainty, this qualitative claim should be supported by confidence intervals.","section":"Sec. III.B, discussion after Eq. (8)"},{"comment":"The choice of the spin-z basis for the participation entropy is justified by the domain-wall picture, but the statement that 'any two choices of basis connected by transformations that commute with the Hamiltonian will give the same behavior' is too strong; it holds only for exact symmetries of the Hamiltonian, not for arbitrary local basis rotations. Clarify the intended scope of this remark.","section":"Sec. IV"}],"recommendation":"major_revision","confidential_remarks":"The paper addresses a timely and contested problem, and the TBC approach is a useful methodological contribution. My main reservation is that the headline claim (thermal critical entanglement entropy) is not robustly supported by the small-system extrapolation, and the scaling analysis would benefit from a more transparent collapse criterion. These are fixable in a revision, but they are load-bearing for the paper's central conclusion. I would not recommend rejection at this stage."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nShort version: this is a useful paper on the MBL transition in the disordered Ising chain, and the scaling-in-L_A/xi idea is worth taking seriously. But the paper's strongest claim—that the critical entanglement entropy is thermal, volume-law with amplitude ln2—is not established by its own fits. The consistency between entanglement and participation entropy is the most convincing part.\n\nWhat is actually new: they avoid half-chain quantities and write S_A^E = L_A f(L_A |δJ−δJ_c|^nu, r), treating total size only via r=L_A/L. Twisted boundary conditions improve the data quality, and they cross-check with participation entropy in the domain-wall spin basis. The agreement δJ_c≈3.19 vs 3.16 and ν≈0.94 vs 0.89 is a genuine, non-trivial consistency check. If this scaling form is right, it is a useful tool. The paper is also honest about assumptions—it assumes continuity, cites Grover, and discusses the Harris bound.\n\nSoft spots, in order of seriousness. First, the thermal claim rests on an intercept α=0.72±0.03 from Eq. (8), fit to f_c(r) at r values down to 1/16 and 1/8, which are L_A=1 and 2 at L=16. So the r→0 extrapolation is really an L_A→0 extrapolation at fixed L. The intended limit 1≪L_A≪L is not sampled. The fitted α is above ln2≈0.693, and the curvature γ=0.84±0.62 is poorly constrained; a 1/L_A correction could easily shift the intercept. Second, the data-collapse quality is judged visually; no quantitative collapse criterion or error on the collapse is given, so δJc and ν carry unknown systematic uncertainty. Third, disorder-averaging details (number of realizations, angle averaging) are absent, and no code or data are provided, so the numerics are not independently checkable. These concerns do not invalidate the scaling method, but they do undermine the specific 'strictly thermal' headline. The subthermal-versus-thermal debate needs a careful finite-size study with L_A large enough to test the limit, and this paper does not provide that.\n\nWho should read it: people working on MBL scaling, especially those interested in subsystem-size scaling and participation entropy. It deserves a serious referee—the method is worth engaging even if the conclusion is not yet proven. I would not cite it as evidence for a thermal critical point as-is.\n\nRecommendation: send to peer review with a request for quantitative collapse analysis, error methodology, and a critical look at the small-L_A extrapolation. Conditional.","headline":"A plausible scaling scheme with a genuine participation-entropy cross-check, but the 'strictly thermal at criticality' conclusion rests on an r→0 extrapolation that is really L_A→0, so the headline is not established.","tokens_in":14021,"tokens_out":2385,"would_cite":false,"duration_ms":25209,"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":"The many-body localization transition is continuous, and at the critical point the entanglement entropy is strictly thermal, matching the same critical behavior in the participation entropy.","keywords":["many-body localization","entanglement entropy","participation entropy","finite-size scaling","disordered quantum spin chain","exact diagonalization","twisted boundary conditions","critical exponent"],"falsifier":"Measure the von Neumann entanglement entropy for subsystems up to $L_A = 10$ in systems of size $L = 20$ to $24$ at the estimated critical disorder $\\delta J \\approx 3.2$, keeping $r \\le 0.1$ (using, for example, tensor-network or DMRG methods that go beyond exact diagonalization). If $S_A^E/L_A$ extrapolated to $L \\to \\infty$ and $r \\to 0$ does not approach $\\ln 2$ within error bars, or if a two-parameter scaling with an additional relevant field gives a significantly better collapse, the thermal-critical-point and single-length-scale claims would be refuted.","tokens_in":12912,"feed_emoji":"🧲","tokens_out":13934,"duration_ms":112971,"temperature":0.7,"pith_summary":"This paper studies the transition between many-body localized (MBL) and thermal states in a disordered transverse-field Ising chain, using exact diagonalization of eigenstates under twisted boundary conditions. It argues that the transition is continuous and that the entanglement entropy of a small subsystem at the critical point is exactly the thermal volume law, about $\\ln 2$ per spin, once finite-size corrections are removed. The same critical disorder and correlation-length exponent are found for the participation entropy in the domain-wall spin basis, suggesting that MBL is a localization transition in many-body configuration space. If correct, earlier reports of a 'subthermal' critical point would be finite-size artifacts of using half-system partitions.","feed_headline":"At MBL criticality, entropy per spin hits ln 2","feed_subtitle":"Scaling of a disordered spin chain puts critical entanglement entropy at ln 2 per spin.","key_machinery":"The central object is the finite-size scaling ansatz for the entanglement entropy, $S_A^E(L_A, r, \\delta J) = L_A f(L_A |\\delta J - \\delta J_c|^\\nu, r)$, where $r = L_A/L$ is the partition ratio. The ansatz states that the only relevant scaling variable is $L_A/\\xi$ with $\\xi$ the divergent correlation length, that the total size $L$ enters only through $r$ as a correction, and that the critical amplitude $f_c(r) = f(0,r)$ extrapolated to $r \\to 0$ gives the thermal value $\\ln 2$. The extraction proceeds by data-collapsing $S_A^E/L_A$ versus $r$ to determine $\\delta J_c$ and $f_c(r)$, then scaling $y = S_A^E/(f_c(r)L_A)$ versus $x = L_A|\\delta J - \\delta J_c|^\\nu$ to determine $\\nu$.","core_discovery":"In the thermodynamic limit $L \\gg L_A \\gg 1$, the paper claims the entanglement entropy per subsystem size at the critical disorder obeys $S_A^E/L_A \\to \\ln 2$, the fully thermal value, with the extrapolated intercept $\\alpha = 0.72 \\pm 0.03$ consistent with $\\ln 2 \\approx 0.693$. The correlation length diverges as $\\xi \\sim |\\delta J - \\delta J_c|^{-\\nu}$ with $\\nu = 0.94 \\pm 0.07$; this violates the Harris bound but agrees with a theoretical prediction of $\\nu = 1$. The participation entropy in the $z$-spin (domain-wall occupation) basis yields the same critical point $\\delta J_c' = 3.16 \\pm 0.04$ and exponent $\\nu' = 0.89 \\pm 0.03$, leading the authors to conclude that the transition is continuous and describable as localization in many-body configuration space.","pith_inferences":["If the scaling theory is correct, the same $L_A/\\xi$ analysis with partition-ratio corrections could sharpen critical-exponent estimates in other disordered models where exact diagonalization is limited to small sizes.","A testable extension: in a global quench starting from a product state at the critical disorder, the entanglement entropy of a small interval should saturate at $\\ln 2$ per spin, reproducing the thermal value predicted here.","The basis-dependence argument implies that participation-entropy scaling will only match the entanglement-entropy result when the chosen basis aligns with the quasiparticle occupation basis of the model's ordered ground state.","The study is performed at high energy density ($\\epsilon = 59/60$); whether the strictly thermal critical amplitude persists at lower energy densities remains an open question."],"forward_implications":["The MBL transition is a continuous eigenstate phase transition described by a single divergent length scale set by $L_A/\\xi$.","At the critical point a finite subsystem is thermal: $S_A^E/L_A \\to \\ln 2$, so the critical eigenstates carry maximal entropy per site.","The exponent $\\nu \\approx 0.94$ supports the view that the Harris criterion does not apply to this transition, consistent with a predicted $\\nu = 1$.","The matching scaling of the participation entropy suggests the MBL transition is equivalent to Anderson localization in the many-body configuration space.","The subthermal entanglement seen in prior half-chain studies is reinterpreted as a finite-size effect of large $r$, not a thermodynamic property."],"supporting_citations":[{"why":"sets the theoretical premise that a continuous transition makes the critical eigenstates thermal, which the paper's scaling analysis aims to confirm.","marker":"[22]"},{"why":"introduces the disordered transverse-field Ising chain with second-neighbor coupling used for exact diagonalization.","marker":"[9]"},{"why":"provides the error-analysis method for the scaling fits and earlier half-chain entanglement data reinterpreted as finite-size effects.","marker":"[8]"},{"why":"gives the predicted correlation-length exponent ν = 1, the benchmark for the paper's extracted ν = 0.94.","marker":"[28]"},{"why":"argues for an area-law critical regime, the opposing view the paper's thermal-critical-point result is designed to test.","marker":"[23]"},{"why":"demonstrates similarities between participation entropy and entanglement entropy, motivating the basis-space analysis.","marker":"[29]"},{"why":"supplies the twisted-boundary-condition method used to reduce finite-size effects.","marker":"[31]"}],"fun_headline_variants":["MBL criticality: entropy per spin hits ln2","Disordered chain: critical entropy reaches thermal value","Participation entropy locks onto same MBL critical point","Continuous MBL transition from entropy scalings","Entanglement entropy saturates to ln2 per spin at MBL"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The scaling ansatz $S_A^E(L_A, r, \\delta J) = L_A f(L_A |\\delta J - \\delta J_c|^\\nu, r)$ must hold: the subsystem entropy can depend on the total size only through the ratio $r$, and on disorder only through the single combination $L_A |\\delta J - \\delta J_c|^\\nu$.","fun_headline_variants_meta":{"raw":{"variants":["MBL criticality: entropy per spin hits ln2","Disordered chain: critical entropy reaches thermal value","Participation entropy locks onto same MBL critical point","Continuous MBL transition from entropy scalings","Entanglement entropy saturates to ln2 per spin at MBL"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000324,"raw_usage":{"total_tokens":1840,"prompt_tokens":992,"completion_tokens":848,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":608,"completion_tokens_details":{"reasoning_tokens":770}},"tokens_in":608,"tokens_out":848,"duration_ms":9182,"temperature":1.0,"reasoning_tokens":770,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:55:43.278235+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the von Neumann entanglement entropy for subsystems up to $L_A = 10$ in systems of size $L = 20$ to $24$ at the estimated critical disorder $\\delta J \\approx 3.2$, keeping $r \\le 0.1$ (using, for example, tensor-network or DMRG methods that go beyond exact diagonalization). If $S_A^E/L_A$ extrapolated to $L \\to \\infty$ and $r \\to 0$ does not approach $\\ln 2$ within error bars, or if a two-parameter scaling with an additional relevant field gives a significantly better collapse, the thermal-critical-point and single-length-scale claims would be refuted.","supporting_citations":[{"cited_title":"Monthus ,\\ 10.3390/e18040122 journal journal Entropy \\ volume 18 ( year 2016 ),\\ 10.3390/e18040122 NoStop","cited_arxiv_id":null,"evidence_quote":"gives the predicted correlation-length exponent ν = 1, the benchmark for the paper's extracted ν = 0.94."},{"cited_title":"Bera , author H","cited_arxiv_id":null,"evidence_quote":"demonstrates similarities between participation entropy and entanglement entropy, motivating the basis-space analysis."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"supplies the twisted-boundary-condition method used to reduce finite-size effects."}],"review_version":1}