{"id":"0bd631c3-24b4-414e-b930-549daa44c63c","arxiv_id":"2507.22795","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Quenched genuine multipartite entanglement (GGM) tracks the ergodic-MBL transition in disordered Heisenberg chains, with three-body DM interactions delaying localization to higher disorder strengths.","lead":"This numerical study uses a multipartite entanglement measure, the generalized geometric measure (GGM), to detect the ergodic-to-many-body-localized transition in a disordered Heisenberg spin chain with Dzyaloshinskii-Moriya interactions. It finds that genuine multipartite entanglement is high in the ergodic phase, vanishes in the localized phase, and that DM interactions, especially three-body ones, push the transition to higher disorder strengths.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central quantitative claim that GGM-derived transition points align with standard indicators is weakened by the absence of a systematic finite-size analysis and error bars, and the GGM and gap-ratio critical disorder values in Table II are not actually in close agreement.","rationale":"The reader's weakest_assumption is the finite-size scaling assumption (power-law correlation length divergence and system-size-independent h*), which I agree is the key load-bearing premise. However, I found an even more direct and internal problem: the quantitative agreement claim fails already within Table II, independently of whether the finite-size scaling ansatz is correct. The discrepancies (GGM vs gap ratio) are large and drift with the interaction setting, and the text gives a different transition value (3.3) in Sec. III.A than Table II (4.0) for the same D = 0 case. This internal inconsistency is a concrete correctness risk. The finite-size scaling assumption cited by the reader is indeed serious, but the paper does not merely weakly support the alignment claim; the data in the paper contradicts it. Hence the qualitative message is credible, but the quantitative 'aligns well' claim is not established. CONDITIONAL is appropriate: the paper should add a direct comparison with bootstrap error bars, a crossing analysis, and a resolution of the internal inconsistency. I agree with the reader's identification of the weakest assumption, but my independent check strengthens the case. I do not see a basis for REJECT because the qualitative use of GGM as a phase indicator is well-supported within this study, and the discrepancy is in the precision of h* estimates, not in the existence of a crossover or the direction of the DM effect.","tokens_in":23397,"tokens_out":2060,"duration_ms":21609,"concrete_test":"Recompute the finite-size scaling for the D = 0 case using the same procedure as in Sec. IV, but for at least twice as many disorder values near the crossing and system sizes N = 8 through 18, with bootstrap error bars on ⟨G⟩ and ⟨r⟩. Then extract h* by (i) the same cost-function data collapse, and (ii) an independent crossing analysis of the N-1 vs N curves. If the two methods disagree, or if adding a 20% larger N changes h* by more than the bootstrap error, the claim that GGM gives a system-size-independent h* that 'aligns well' with the gap ratio is not supported.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The paper's headline claim (Sec. III, Abstract) is that the GGM-identified transition point aligns well with standard indicators such as the gap ratio and long-range correlator. Table II directly contradicts this: for D = D' = 0, h* = 4.0 (GGM) vs 2.9 (gap ratio), a 38% discrepancy; for D = 0, D' = 0.5, h* = 7.9 vs 6.3, a 25% discrepancy; even for D = 0.5, D' = 0, h* = 3.9 vs 3.4, a discrepancy of 15%. Similarly, dynamical h* values in Tables III and IV are inconsistent with each other (e.g., D = 0, D' = 0.5: transient h* = 5.9 vs steady-state h* = 5.7, vs equilibrium h* = 7.9), which is not discussed. The claim that the transition point 'aligns well' is therefore unsupported. Furthermore, the central quantitative estimates rely on the finite-size scaling in Eqs. (8) and (9), which assumes a power-law divergence of the correlation length and a system-size-independent critical point; this is exactly the assumption contested in Ref. [119] (Suntajs et al.), where apparent MBL transitions are shown to drift with system size. The paper cites [119] but does not confront this drift. Given that the model here is studied for N ≤ 18 (with the GGM data in Fig. 6 using only N = 12, 16, 18 for the scaling) and that disorder-averaged quantities have no reported error bars, the extracted h* and ν values are not robust. The qualitative diagnostic claim — that GGM distinguishes ergodic from MBL regimes — is credible, but it is not the same as the 'aligns well' claim. An internal inconsistency in Sec. III.A (text says the transition is near 3.3, while Fig. 2(a) caption and Table II say h* = 4.0) further undermines precision. The numerical protocols (POLFED for N = 16, 18 and Chebyshev dynamics) are useful detail, but they do not address the missing systematic scaling and uncertainty quantification.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies quenched genuine multipartite entanglement, quantified by the generalized geometric measure (GGM) and its single-site approximation G1, in a disordered spin-1/2 Heisenberg chain with two- and three-body Dzyaloshinskii-Moriya (DM) interactions. Using exact diagonalization, POLFED, and Chebyshev time evolution, the authors compute GGM for mid-spectrum eigenstates and for dynamics from a Néel state, and extract critical disorder strengths h* through finite-size scaling. The central claims are that GGM is high in the ergodic phase and vanishes in the many-body localized (MBL) phase; that the GGM-derived transition point aligns well with standard indicators such as the gap ratio and long-range correlator; and that DM interactions, especially the three-body term, delay localization to larger disorder. The paper also proposes the distribution of GGM as a phase marker.","tokens_in":23925,"tokens_out":6655,"duration_ms":73165,"significance":"If substantiated, the paper would establish genuine multipartite entanglement as an experimentally relevant diagnostic of the ergodic-to-MBL transition, going beyond the usual bipartite quantities. The work has clear strengths: the single-site approximation G1 is explicitly validated in Fig. 1; the numerical machinery (ED, POLFED, Chebyshev expansion) is appropriate and state of the art; and the qualitative prediction that DM interactions shift h* to larger values is concrete and falsifiable. However, the quantitative claim that GGM and standard indicators yield consistent transition points is not supported by the numbers reported in the manuscript itself, and the absence of uncertainty quantification makes the specific h* and ν values difficult to assess. The qualitative diagnostic value of GGM is credible, but the degree of quantitative agreement is currently overstated.","major_comments":[{"comment":"The central claim that GGM-derived h* 'aligns well' with the gap-ratio h* is contradicted by the paper's own Table II. For D=D'=0 the values are h*(GGM)=4.0 versus h*(gap ratio)=2.9, a discrepancy of about 38%; for D=0.5,D'=0 they are 3.9 versus 3.4; and for D=0,D'=0.5 they are 7.9 versus 6.3. In addition, Sec. III A states that 'the transition is observed near ~3.3 through ⟨G⟩' for the same D=0 case, while Table II reports h*=4.0. The authors should report uncertainties from disorder averaging and from the fitting procedure, and provide a quantitative measure of agreement (e.g., collapse residuals or bootstrap intervals), or they should revise the claim to reflect the observed spread.","section":"Sec. III, Table II"},{"comment":"The finite-size scaling uses only N=12, 16, 18 and fits both h* and ν freely; no error bars, no collapse-quality landscape, and no test of stability with respect to including N=14 are reported. The manuscript cites Ref. [119] for the known drift of apparent MBL transitions with system size but does not discuss how that drift affects the present h* values. Since both h*(GGM) and h*(gap ratio) are obtained from the same power-law scaling ansatz with free parameters, the agreement between them is not an independent validation of the ansatz. Please provide the cost function C_X as a function of h* and ν, bootstrap or jackknife errors, and a demonstration that h* is stable under the inclusion of additional system sizes or alternative scaling forms.","section":"Sec. IV, Eqs. (8)-(9), Fig. 6"},{"comment":"The dynamical transition points are inconsistent with the equilibrium values and with each other. For D=0,D'=0.5, the transient h*=5.9 (Table III) and steady-state h*=5.7 (Table IV) are far from the equilibrium h*=7.9 (Table II), yet the text says these values 'agree' and 'closely align'. The transient criterion α=0.0025 is set by hand, and the steady-state criterion β≃0 is not defined quantitatively. Please report the sensitivity of h* to these thresholds and reconcile the equilibrium and dynamical estimates before asserting consistency between static and dynamical probes.","section":"Sec. V, Tables III-IV"},{"comment":"The h* values quoted in Sec. III B (h*=3.9 for two-body DM and h*=7.9 for three-body DM) are said to be obtained from the finite-size scaling in Sec. IV, but they are already presented and used in Fig. 3 before the scaling analysis. Moreover, Fig. 6 shows collapses only for the DM cases, while Table II also reports h* for D=D'=0. This makes it difficult to verify how the quoted critical points are extracted. Please show the scaling collapse for all cases, including D=D'=0, and state clearly whether the reported h* values come from the collapse minimization in Eq. (9) or from visual estimates of the raw data.","section":"Sec. III B and Sec. IV"}],"minor_comments":[{"comment":"The caption of Fig. 2 reports h*=4.0 from GGM and h*=2.9 from the gap ratio, while the main text states that the transition is observed near h*~3.3 through ⟨G⟩; these numbers should be reconciled.","section":"Fig. 2 and Sec. III A"},{"comment":"The inequality in Eq. (5) appears to have the wrong sign: a confidence bound of the form P(|λmax_est - λmax| < ε) ≥ 1 - κ would be expected, not '< 1 - κ'.","section":"Sec. II A 1, Eq. (5)"},{"comment":"The notation '|A|max=1' in Eq. (4) is confusing; please define the bipartition constraint explicitly, for example by writing |A|=1 with A a single site.","section":"Sec. II A 1, Eq. (4)"},{"comment":"The paragraph on randomized measurements and classical shadows is not used in any numerical result in the paper; either connect this toolbox to the analysis or remove it to avoid a dangling methodological claim.","section":"Sec. II A 1"},{"comment":"The finite-size scaling figure uses only N=12, 16, 18, although Table I also lists N=14 for static quantities; adding N=14 to the collapse or explaining its omission would strengthen the analysis.","section":"Fig. 6"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within the scope of a quantum many-body physics journal. My main concern is calibration: the qualitative GGM diagnostic is credible and worth publishing, but the quantitative comparison with standard probes is not supported by the reported numbers. A revised version that adds uncertainty quantification, shows all scaling collapses, and reconciles the equilibrium and dynamical h* values would address the core issue."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read this if you care about multipartite entanglement as a probe of MBL: the qualitative story is credible, and the three-body DM delay of localization is a genuinely new model-specific result. What is not credible yet is the quantitative \"aligns well\" claim and the specific h* values quoted in the abstract and tables.\n\nWhat the paper does well: it computes GGM for both eigenstates and time-evolved states, shows that the single-site approximation G1 is exponentially close to the full GGM, and connects GGM estimation to randomized measurements, which makes the probe experimentally plausible. The appendices give enough detail on POLFED and Chebyshev dynamics for a competent group to reproduce the numerics. The references are comprehensive, and the relevant self-citations are legitimate.\n\nThe soft spots are real and centered on overclaiming. Table II directly contradicts the \"good agreement\" language: for D=D'=0, GGM gives h*=4.0 while the gap ratio gives 2.9; for D=0, D'=0.5 they are 7.9 and 6.3. Those are 25-38% discrepancies, not agreement. The text in Sec. III.A even says the transition is near 3.3 while Fig. 2's caption and Table II say 4.0. There are no error bars on any fitted h* or nu, and the dynamical transition thresholds (alpha=0.0025, beta=0) are hand-set. The resulting dynamical h* values also disagree with each other and with the equilibrium values without any discussion.\n\nMore fundamentally, the finite-size scaling assumes a power-law diverging correlation length and a system-size-independent critical point, which is exactly the assumption contested by Suntajs et al. The paper cites [119] but does not confront the drift question. With N <= 18 and only N=12, 16, 18 in the scaling collapse, the extracted critical values are not robust.\n\nNone of this sinks the central qualitative diagnostic: GGM clearly separates the ergodic and MBL regimes, and the G1 simplification is well justified. But the paper needs a serious revision before those h* numbers are usable: error bars, a finite-size drift analysis, and either softened language or a proper defense of the agreement claim.\n\nSend it to a competent referee; it deserves referee time, but expect heavy revision. I would bring it to a reading group, mostly to argue about finite-size scaling in MBL.","headline":"A useful numerical study showing GGM tracks the ergodic-to-MBL transition in disordered Heisenberg chains with DM interactions, but the quantitative 'good agreement' claim and the reported h* values need significantly more support.","tokens_in":24448,"tokens_out":2412,"would_cite":true,"duration_ms":29664,"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":"Genuine multipartite entanglement, measured by GGM, tracks the ergodic-to-MBL transition in a disordered Heisenberg chain: high in the thermal phase, vanishing in the localized phase, with DM interactions delaying localization.","keywords":["many-body localization","genuine multipartite entanglement","generalized geometric measure","Dzyaloshinskii-Moriya interaction","disordered Heisenberg spin chain","gap ratio","finite-size scaling","quantum entanglement dynamics"],"falsifier":"A decisive test would be to push the same GGM finite-size scaling to $N=20$ to $24$ in the three-body DM model and check whether the extracted $h^*$ stays near 7.9 or keeps moving with system size; continued drift would falsify the claimed scale-invariant transition. A complementary check is to compare the equilibrium $h^*$ with the dynamical $h^*$ at those sizes to see whether the 7.9 versus $\\approx 5.9$ discrepancy persists.","tokens_in":2121,"feed_emoji":"🧲","tokens_out":4372,"duration_ms":206454,"temperature":0.7,"pith_summary":"This paper establishes that a computable measure of genuine multipartite entanglement, the generalized geometric measure (GGM), can serve as a reliable indicator of the ergodic-to-many-body-localization transition in disordered spin chains. In the ergodic phase the quenched-averaged GGM saturates near its maximum value, while it falls to zero in the MBL phase, both for mid-spectrum eigenstates and for the long-time state reached from an alternating up/down initial product state. The transition points obtained from GGM agree with the gap ratio and with long-range correlator indicators. The paper also reports that Dzyaloshinskii–Moriya interactions extend the ergodic phase, with the three-body DM term delaying the onset of localization much more than the two-body term. If correct, this gives a practical way to locate the transition in systems where eigenstate probes are hard, and it identifies the localized phase as poor in multipartite-entanglement resources.","feed_headline":"Entanglement vanishes as quantum spin chains localize","feed_subtitle":"A computable multipartite-entanglement measure tracks the transition and shows three-body spin-orbit couplings delay it.","key_machinery":"The load-bearing object is the generalized geometric measure (GGM), defined for a pure state $|\\Psi\\rangle$ as the minimum distance to the set of non-genuinely multipartite entangled states, $G(|\\Psi\\rangle) = 1 - \\max_{|\\chi\\rangle}|\\langle\\chi|\\Psi\\rangle|^2$, equivalently $G = 1 - \\max_{A:B}\\lambda_{A:B}^2$, where $\\lambda_{A:B}$ is the largest Schmidt coefficient across any bipartition. In this model, the maximization can be restricted to single-site bipartitions because the difference between the full measure and its single-site approximation satisfies $|\\langle G-G_1\\rangle| \\approx e^{-N}$. The argument runs on two pairings: in equilibrium, middle-spectrum eigenstates in the ergodic phase have statistics similar to random pure states and therefore carry high GGM, while localized eigenstates are close to product states; in dynamics, the disorder-averaged GGM from an alternating product state grows as $G(t) = a\\ln(t^\\alpha) + b$ and saturates at $G_\\infty = cN^\\beta + d$, with exponents $\\alpha$ and $\\beta$ crossing zero near the transition. The DM terms are two- and three-body antisymmetric exchange interactions that conserve total $S^z$ but break time-reversal symmetry, shifting level statistics from GOE to GUE and lengthening the thermal phase.","core_discovery":"For the spin-1/2 Heisenberg chain with random z-directed fields, the quenched average of GGM over middle-spectrum eigenstates behaves like a two-phase indicator: it rises toward a system-size-independent plateau around 0.5 in the ergodic phase and decays to zero beyond the critical disorder. The authors show that a single-site approximation of GGM, denoted $G_1$, differs from the full GGM by $|\\langle G-G_1\\rangle| \\approx e^{-N}$ in this model, which makes the computation feasible for system sizes up to $N=18$. Finite-size scaling of $\\langle G\\rangle$ gives $h^* = 4.0$ without DM, $h^* = 3.9$ for two-body DM ($D=0.5$), and $h^* = 7.9$ for three-body DM ($D'=0.5$), while the gap-ratio scaling gives 2.9, 3.4, and 6.3. In dynamics starting from an alternating product state, the transient GGM follows $G(t) = a\\ln(t^\\alpha) + b$ and the steady-state value scales as $G_\\infty = cN^\\beta + d$; the disorder strengths where $\\alpha$ or $\\beta$ vanish give dynamical transition estimates 4.0, 4.6, 5.9 and 3.21, 3.4, 5.7, respectively. The paper reads the shift under DM interactions as a consequence of long-range correlations that the DM terms inject, which require stronger disorder to suppress before localization sets in.","pith_inferences":["The equilibrium $h^* \\approx 7.9$ for the three-body DM model is larger than the dynamical values ($\\approx 5.9$ and $\\approx 5.7$); if this gap persists at larger sizes it would suggest that eigenstate and dynamical probes see different crossover scales, a question the paper leaves open.","The same single-site GGM recipe could be applied directly to quasiperiodic disorder models, where finite-size drift in the MBL transition is less contested, offering a cleaner test of whether GGM's transition marker is universal.","Because the GGM distribution over eigenstates narrows near its peak as DM strength grows, the peak position or width of $P(G)$ could serve as a standalone experimental order parameter measurable with randomized measurements.","If the ergodic-phase plateau in GGM reflects typical random-state entanglement, the same probe should also respond to other ergodicity-breaking mechanisms such as quantum many-body scars or Stark localization."],"forward_implications":["Mid-spectrum eigenstates in the MBL phase carry essentially no genuine multipartite entanglement, so localized systems will be poor hosts for protocols that need multipartite-entangled resources.","The single-site GGM approximation $G_1$ is enough to locate the transition, making the diagnostic accessible to randomized-measurement experiments that avoid full tomography.","Including DM interactions, especially the three-body term, shifts the ergodic window to higher disorder, so tuning spin-orbit-like couplings in a quantum simulator can control how easily localization sets in.","Dynamical signatures—the transient growth rate $\\alpha$ and the steady-state exponent $\\beta$—cross near the same disorder strengths as equilibrium probes, so the transition can be detected from short-time and steady-state measurements without preparing high-energy eigenstates.","Deep in the MBL phase the steady-state entanglement shrinks with system size ($\\beta<0$), a sub-area-law behavior that should appear as entanglement freezing in larger simulators."],"supporting_citations":[{"why":"Defines the generalized geometric measure of genuine multipartite entanglement that the paper uses throughout as its central probe.","marker":"[92]"},{"why":"Establishes the ergodic-to-MBL transition in the disordered Heisenberg chain, the baseline model and the conventional behavior the paper's results are compared against.","marker":"[20]"},{"why":"Provides the gap-ratio distribution whose GOE and Poisson limits benchmark the spectral-statistics indicator $\\langle r\\rangle$.","marker":"[7]"},{"why":"Supplies the exact-diagonalization implementation used for eigenstate data up to $N=14$.","marker":"[109]"},{"why":"Introduces the polynomially filtered exact diagonalization routine used to reach middle-spectrum eigenstates at $N=16$ and $N=18$, which the finite-size scaling depends on.","marker":"[110]"},{"why":"Supplies the finite-size scaling methodology and cost-function collapse used to extract $h^*$ and $\\nu$.","marker":"[72]"},{"why":"Cited for the same scaling methodology while also being the source of the known finite-size drift that challenges the assumption of a sharp transition.","marker":"[119]"},{"why":"Gives the imbalance and slow-dynamics diagnostics that the dynamical GGM results are compared against.","marker":"[62]"}],"fun_headline_variants":["Genuine entanglement vanishes at many-body localization","Spin-orbit coupling delays entanglement death in spin chains","Entanglement probe exposes localization transition in spin chains","Three-body DM interactions shift many-body localization boundary","GME tracks ergodic-to-MBL transition in disordered chains"],"cache_read_input_tokens":26240,"weakest_assumption_plain":"The reported transition values assume that the apparent crossover seen in chains of up to 18 spins is a genuine sharp phase transition, not a finite-size artifact that would keep drifting at larger sizes.","fun_headline_variants_meta":{"raw":{"variants":["Genuine entanglement vanishes at many-body localization","Spin-orbit coupling delays entanglement death in spin chains","Entanglement probe exposes localization transition in spin chains","Three-body DM interactions shift many-body localization boundary","GME tracks ergodic-to-MBL transition in disordered chains"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000609,"raw_usage":{"total_tokens":2914,"prompt_tokens":1099,"completion_tokens":1815,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":715,"completion_tokens_details":{"reasoning_tokens":1750}},"tokens_in":715,"tokens_out":1815,"duration_ms":16797,"temperature":1.0,"reasoning_tokens":1750,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T11:16:48.998736+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A decisive test would be to push the same GGM finite-size scaling to $N=20$ to $24$ in the three-body DM model and check whether the extracted $h^*$ stays near 7.9 or keeps moving with system size; continued drift would falsify the claimed scale-invariant transition. A complementary check is to compare the equilibrium $h^*$ with the dynamical $h^*$ at those sizes to see whether the 7.9 versus $\\approx 5.9$ discrepancy persists.","supporting_citations":[{"cited_title":"Fulde, B","cited_arxiv_id":null,"evidence_quote":"Supplies the exact-diagonalization implementation used for eigenstate data up to $N=14$."},{"cited_title":"Magnetic ordering and dynamics in monolayers and bilayers of chromium trihalides: atomistic simulations approach","cited_arxiv_id":"2404.15543","evidence_quote":"Introduces the polynomially filtered exact diagonalization routine used to reach middle-spectrum eigenstates at $N=16$ and $N=18$, which the finite-size scaling depends on."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Cited for the same scaling methodology while also being the source of the known finite-size drift that challenges the assumption of a sharp transition."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the imbalance and slow-dynamics diagnostics that the dynamical GGM results are compared against."}],"review_version":1}