{"id":"389dba79-a700-40ed-95fd-fed0c4371d9b","arxiv_id":"2509.01230","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":7,"one_line_summary":"A real-space dynamical mean-field calculation maps the disordered 2D Bose-Hubbard phase diagram and finds finite-size Bose glass boundaries that deviate from thermodynamic-limit predictions.","lead":"Researchers used a computer model of ultracold bosons in a 2D lattice with random disorder to map out a weird insulating phase, the Bose glass, and where it sits between a superfluid and a Mott insulator. The same calculation also predicts a measurable spectral signature of this phase: damped, localized excitations that experiments could detect with spectroscopy.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"MI/BG boundary rests on hand-set EAOP/compressibility thresholds without threshold-dependence or error analysis, leaving the claimed finite-size deviation from the energy-gap criterion quantitatively unsecured.","rationale":"The reader's weakest-assumption analysis correctly identifies the hand-set EAOP/compressibility thresholds as the most load-bearing weakness. My stress-test agrees: the paper's central claim—that EAOP/compressibility, rather than Eg ≤ ∆, defines the MI/BG boundary in finite systems—depends on threshold values that are not justified by a threshold-dependence study, and the absence of error bars compounds the issue. The paper's qualitative structure is plausible, the SF/BG boundary is benchmarked against QMC, and the spectral comparison to strong-coupling theory is a genuine check, but these do not secure the quantitative MI/BG boundary. Since the reader already returned CONDITIONAL with this concern, my analysis does not move the verdict; it reinforces the need for the proposed threshold-sensitivity and bootstrapping tests before the quantitative boundary can be accepted.","tokens_in":10326,"tokens_out":3390,"duration_ms":45497,"concrete_test":"Along fixed ∆/J cuts (e.g. ∆/J = 20 and 30), compute q(U/J) and κ(U/J) on the 8×8 ensemble and re-extract the MI/BG boundary using q_c ∈ {0.0005, 0.001, 0.002, 0.005, 0.01} and κ_c ∈ {0.001, 0.005, 0.01, 0.02}. Also bootstrap the 1536 disorder realizations to obtain error bars. If the boundary shifts by more than ~10% in U/J, or if q and κ rise gradually rather than steeply near the chosen cutoffs, the claimed deviation from the Eg ≤ ∆ line is not quantitatively reliable; if the boundary is insensitive to these variations, the concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central assertion—that the EAOP (or compressibility), rather than the energy-gap condition Eg ≤ ∆, is the appropriate MI/BG criterion in finite disordered systems—is operationalized through arbitrary cutoffs: q < 0.001 in Fig. 2, q < 0.002 and κ < 0.01 in Fig. 3, with no threshold scan and no statistical error bars. This matters because q in a disordered finite system rises continuously as rare sites develop n=2 occupations; the boundary location is then set by the chosen cutoff, not by a sharp phase transition. The paper defends the percolation cutoff ϕc by showing boundary insensitivity, but no analogous check is provided for q or κ. Furthermore, κ is computed with δµ = 0.01, the same order as the κ threshold, so numerical derivative noise may dominate near the boundary. Without knowing the slope of q(U/J) or κ(U/J) at the cutoffs, the reported boundary—and thus the claimed 'significant deviations' from the thermodynamic-limit Eg ≤ ∆ line—is not quantitatively robust. The finite-size explanation for the deviation is plausible but is supported only by an unshown 4×4-to-16×16 comparison, and the hand-set thresholds are the immediate load-bearing weakness.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the two-dimensional disordered Bose-Hubbard model using real-space bosonic dynamical mean-field theory (RBDMFT) solved by exact diagonalization, with disorder averaging over 1536 realizations on finite square lattices. The superfluid-to-insulator transition is identified through a percolation analysis of the local condensate order parameter, while the Mott-insulator-to-Bose-glass transition is assigned using the Edwards-Anderson order parameter (EAOP) and the compressibility, both compared with the clean-system energy-gap criterion Eg ≤ Δ. The authors report an intermediate Bose-glass phase in all cases, find that the EAOP/compressibility boundary deviates significantly from the Eg ≤ Δ line, attribute this deviation to finite-size suppression of rare regions, and further compute spectral functions that show damped-localized excitations in agreement with a strong-coupling expansion.","tokens_in":10705,"tokens_out":4076,"duration_ms":48463,"significance":"If the claims are quantitatively robust, the paper offers a practical finite-system diagnostic for the Bose-glass phase and a phase diagram for the 2D disordered Bose-Hubbard model that extends earlier mean-field and quantum Monte Carlo studies. The SF/BG boundary is benchmarked against QMC results and appears insensitive to the percolation cutoff, and the spectral-function comparison with strong-coupling theory is a useful cross-check. However, the central MI/BG claim rests on hand-set numerical thresholds for EAOP and compressibility, with no threshold-dependence analysis or error bars; without those tests, the reported 'significant deviations' from the thermodynamic-limit criterion are not quantitatively secured. The work is therefore a valuable contribution but needs additional analysis before its central message can be accepted.","major_comments":[{"comment":"The MI/BG boundary is determined by ad hoc thresholds: q < 0.001 in Fig. 2 and q < 0.002 / κ < 0.01 in Fig. 3. No threshold scan or statistical error bars are provided. In a finite disordered system, q increases continuously as rare sites develop n=2 occupations, so the boundary location is set by the chosen cutoff rather than by a sharp phase transition. The paper's central claim—that the EAOP/compressibility boundary deviates significantly from the Eg ≤ Δ line—is therefore not quantitatively robust. Please provide q(U/J) and κ(U/J) curves at representative disorder strengths, show how the extracted boundary shifts as the thresholds are varied (e.g., q in [0.0005, 0.005], κ in [0.005, 0.02]), and include bootstrap or ensemble error bars based on the 1536 realizations.","section":"Section III, Fig. 2 and Fig. 3; Eq. (3)"},{"comment":"The compressibility is computed via a second RBDMFT run at μ + δμ with δμ = 0.01, while the MI/BG criterion is κ < 0.01. The finite-difference step is the same order as the threshold, so numerical derivative noise can dominate near the boundary. The authors should demonstrate convergence in δμ (e.g., δμ = 0.005 and 0.02) and preferably use a central difference. Without this, the κ-based boundary may be an artifact of the differentiation step.","section":"Section III, compressibility definition"},{"comment":"The text states 'Within the reachable system sizes (4×4 to 16×16) we observe no signs of significant finite-size effects,' but no supporting data are shown. This claim is load-bearing because the paper argues that the deviations from the Eg ≤ Δ criterion arise from the absence of rare regions in finite systems. Please provide the MI/BG boundary (or raw q and κ data) for L = 4, 8, 12, 16 at representative U/J values, or a figure comparing the boundaries. Without these data, the assertion that the effect is not due to ordinary finite-size scaling is unsupported.","section":"Section III, finite-size assertion"},{"comment":"The spectral comparison is made only against the authors' own strong-coupling expansion [20,21]. The agreement is described as 'reasonably accurate' in view of the 'restricted spectral resolution of the exact diagonalization impurity solver,' but no convergence test with respect to the ED bath size or boson occupation cutoff is shown. Since the damped-localized mode interpretation is a central advertised result, the authors should either provide an ED-parameter convergence analysis or quantify the uncertainty in A_k(ω) (e.g., from disorder realizations and frequency resolution).","section":"Section III, spectral function analysis"}],"minor_comments":[{"comment":"Typo: 'RBMDFT' should be 'RBDMFT'.","section":"Abstract"},{"comment":"Duplicate word: 'between between Mott insulator and superfluid phases' in the second paragraph.","section":"Section I"},{"comment":"Typo: 'asymptomatically close' should be 'asymptotically close'.","section":"Section III"},{"comment":"Parenthetical formatting is inconsistent: 'd) (h))' and similar should be 'd) and h)'.","section":"Fig. 4 caption"},{"comment":"'Griffith's type' should be 'Griffiths-type' (the theorem is associated with Griffiths). Also, the paper cites the Ph.D. thesis [21] for part of the analytic spectral comparison; providing a published reference would be preferable.","section":"Section I and references"},{"comment":"No data/code availability statement is given. Given the numerical nature of the work, a statement on availability of the RBDMFT implementation would improve reproducibility.","section":"General"}],"recommendation":"major_revision","confidential_remarks":"The paper has a sound SF/BG boundary benchmarked against QMC and a plausible spectral analysis, but the MI/BG boundary—the main new claim—depends on untested thresholds. The lack of threshold-dependence and error analysis is a solvable but essential issue; I would not accept the paper until those tests are provided. The finite-size comparison mentioned but not shown is also critical. I do not see a fundamental flaw that would require rejection, provided the authors can demonstrate that the reported boundary is not an artifact of the chosen cutoffs."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is a competent RBDMFT study of the 2D disordered Bose–Hubbard model, and the first thing you should know is that the strong parts are the superfluid-to-insulator boundary and the spectral-function comparison. The weak part is the MI/BG boundary, which rests on hand-set thresholds without a stability check.\n\nWhat's actually new: the paper gives finite-system RBDMFT phase diagrams for fixed disorder and for unit filling, uses percolation on the local condensate order parameter to locate the SF/BG transition, and compares three MI/BG criteria: EAOP, compressibility, and the thermodynamic-limit energy-gap condition. The SF/BG boundary agrees well with the QMC results of Söyler et al., which is real evidence that the percolation criterion works. The spectral function in the disordered Mott phase also reproduces the analytic strong-coupling predictions of damped localized modes from Souza et al. That is a genuine numerical check, and the self-citation is not a problem here because the analytic result is independent and externally published.\n\nThe soft spot is exactly where the stress-test note lands. The MI/BG boundary is defined by EAOP thresholds of 0.001 (Fig. 2) and 0.002 (Fig. 3), and a compressibility threshold of 0.01, with no threshold-dependence analysis and no statistical error bars. The compressibility is computed with δµ = 0.01, same order as the threshold, so derivative noise could matter near the boundary. The claimed \"significant deviations\" from the Eg ≤ ∆ line depend on these cutoffs; without knowing the slope of q or κ at the cutoff, that headline result is not quantitatively secured. The finite-size explanation is plausible, but it is supported only by a brief mention of 4×4 to 16×16 with no figure. Also, the ED truncation parameters—local occupation cutoff and bath size—are not stated, which matters for quantitative spectral tails. These are fixable, but currently unquantified.\n\nNone of this sinks the qualitative physics: the intermediate BG between SF and MI is consistent with the theorem of inclusions and with QMC, and the SF/BG boundary is robust. The paper is not overclaiming beyond what the method can deliver.\n\nWho is this for? People working on disordered bosons, quantum gas microscopes, and RF spectral probes. It deserves a serious referee. My recommendation: send it to peer review, and insist that the authors provide a threshold scan for q and κ, error bars, and the ED truncation details. With those, the finite-size deviation claim would be much stronger.","headline":"A useful RBDMFT phase diagram with a solid superfluid boundary and a valuable spectral check, but the MI/BG boundary is set by arbitrary thresholds that need a stability analysis before the finite-size deviation claim is trusted.","tokens_in":11161,"tokens_out":1710,"would_cite":true,"duration_ms":20685,"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":"For finite disordered boson lattices, the Mott-to-Bose-glass boundary is set by the Edwards-Anderson order parameter, not by the thermodynamic-limit energy gap.","keywords":["disordered Bose-Hubbard model","Bose glass","real-space bosonic dynamical mean-field theory","Edwards-Anderson order parameter","compressibility","percolation analysis","spectral function","damped-localized excitations"],"falsifier":"Compute the MI-BG boundary while sweeping the EAOP/compressibility thresholds from 10^-4 to 10^-2 and on larger lattices (32×32, 64×64). If the boundary moves toward the Eg≤Δ line or changes with cutoff or system size, the paper's finite-size-deviation claim fails; if it stays put, the boundary is a robust diagnostic.","tokens_in":10253,"feed_emoji":"🧊","tokens_out":5355,"duration_ms":58407,"temperature":0.7,"pith_summary":"The paper asks which observable actually marks the transition from Mott insulator to Bose glass in a disordered Bose-Hubbard model on a finite two-dimensional lattice. Using real-space bosonic dynamical mean-field theory on lattices up to 16×16 with disorder averages over 1536 realizations, it argues that the Edwards-Anderson order parameter—the realization-to-realization variance of local occupation—or equivalently the compressibility determines the MI-BG boundary, not the thermodynamic-limit energy-gap condition Eg≤Δ. The energy-gap criterion, which relies on exponentially rare disorder regions that are absent in finite systems, overestimates the Bose glass region. The paper also finds that a percolation criterion on the local condensate order parameter gives a superfluid-Bose-glass boundary consistent with Monte Carlo results, and that the RBDMFT spectral function matches strong-coupling predictions of damped-localized excitations. Together these claims make RBDMFT a method that can map the full phase diagram and its excitation spectrum.","feed_headline":"Finite lattices shift the Mott-to-Bose-glass boundary","feed_subtitle":"Real-space DMFT plus percolation maps the 2D disordered Bose-Hubbard phase diagram","key_machinery":"RBDMFT: site-resolved bosonic dynamical mean-field theory maps the lattice model to independent Anderson impurity problems per site, closes self-consistency via the real-space lattice Dyson equation with a local self-energy, and solves by exact diagonalization; it includes second-order 1/z corrections beyond Gutzwiller mean field. The phase boundaries are drawn with three diagnostics: percolation probability of maps of local condensate order parameter above a cutoff φc (superfluid criterion), the Edwards-Anderson order parameter and compressibility thresholds (Mott criterion), and the spectral function A_k(ω) from the disorder-averaged lattice Green's function.","core_discovery":"The central claim is that for finite disordered systems the MI to BG transition is characterized by the Edwards-Anderson order parameter q (variance of local occupation across disorder realizations) and the compressibility κ, both computed directly in RBDMFT, and that this boundary deviates from the Eg≤Δ line derived from the clean system in the thermodynamic limit. In the µ/U-J/U plane at ∆/U=0.5 an intermediate BG phase always separates SF and MI; at ∆/U=1.0 no MI remains. For unit filling, q and κ give the same MI-BG boundary, while the energy-gap criterion gives a substantially larger BG region. The paper interprets this as a finite-size effect: rare Lifshitz regions required by the theo","pith_inferences":["Extension: if the EAOP/κ boundary is the physically relevant one, the energy-gap line should be treated as a thermodynamic-limit benchmark rather than a predictive criterion for finite cold-atom systems; the paper implies this but does not prove convergence for very large lattices.","Extension: a systematic threshold sweep (EAOP and κ cutoffs from 10^-4 to 10^-2) combined with lattice sizes beyond 16×16 would sharpen the central claim; the reported data show no finite-size trend, but do not rule out a slow crossover toward the Eg≤Δ line.","Extension: the spectral-function match suggests that rf-transfer spectroscopy in speckle-disordered gases could observe damped-localized modes as a direct Bose-glass signature, a route the paper mentions experimentally but does not analyze in detail."],"forward_implications":["If the EAOP/compressibility criterion is the correct finite-size MI-BG diagnostic, then the energy-gap line Eg≤Δ should not be used as the MI-BG boundary in real or simulated finite disordered lattices, because it systematically overestimates the Bose glass region.","RBDMFT combined with a percolation analysis reproduces the superfluid-to-insulator boundary of previous Monte Carlo studies, making it a reliable tool for the full 2D disordered Bose-Hubbard phase diagram at low temperature.","The spectral function computed throughout the phase diagram shows damped-localized excitations in the disordered Mott insulator that survive as the Bose glass is approached, providing a spectral signature that can be measured experimentally, e.g., via rf spectroscopy.","At disorder strengths ∆/U ≳ 1 the Mott lobe disappears and the Bose glass dominates the insulating part of the phase diagram, so spectral and compressibility diagnostics become the main ways to identify the insulator."],"supporting_citations":[{"why":"Defines the Bose glass phase and the qualitative SF/MI/BG taxonomy that the paper is testing.","marker":"[3]"},{"why":"Supplies the theorem of inclusions and the Eg≤Δ energy-gap criterion that the paper compares against.","marker":"[7]"},{"why":"Provides the Monte Carlo unit-filling phase diagram used as the quantitative benchmark for both boundaries.","marker":"[9]"},{"why":"Introduces the Edwards-Anderson order parameter as a Bose-glass diagnostic, the basis for the paper's MI-BG criterion.","marker":"[13]"},{"why":"Supplies the percolation-analysis idea that the paper adapts to the condensate order parameter for the SF-insulator transition.","marker":"[19]"},{"why":"Gives the strong-coupling prediction of damped-localized excitations that the spectral function is compared against.","marker":"[20]"},{"why":"Derives the bosonic dynamical mean-field equations and Anderson-impurity mapping that RBDMFT builds on.","marker":"[26]"},{"why":"Provides the real-space BDMFT equations for inhomogeneous systems that the paper uses to treat disorder site-dependently.","marker":"[27]"}],"fun_headline_variants":["Finite-size shifts Mott-glass boundary","Real-space DMFT maps disordered Bose-Hubbard phases","Percolation pins the superfluid-to-insulator line","Damped modes emerge in disordered Bose-Hubbard spectrum"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The MI/BG boundary relies on hand-set numerical thresholds (EAOP below 0.001 or 0.002, compressibility below 0.01) inside an approximate local-self-energy method, with no threshold-sensitivity analysis or statistical error bars; if those cutoffs do not robustly separate finite-size Mott and Bose-glass states, the boundary shifts.","fun_headline_variants_meta":{"raw":{"variants":["Finite-size shifts Mott-glass boundary","Real-space DMFT maps disordered Bose-Hubbard phases","Percolation pins the superfluid-to-insulator line","Damped modes emerge in disordered Bose-Hubbard spectrum"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000291,"raw_usage":{"total_tokens":1544,"prompt_tokens":758,"completion_tokens":786,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":502,"completion_tokens_details":{"reasoning_tokens":732}},"tokens_in":502,"tokens_out":786,"duration_ms":9899,"temperature":1.0,"reasoning_tokens":732,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T12:45:20.817414+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the MI-BG boundary while sweeping the EAOP/compressibility thresholds from 10^-4 to 10^-2 and on larger lattices (32×32, 64×64). If the boundary moves toward the Eg≤Δ line or changes with cutoff or system size, the paper's finite-size-deviation claim fails; if it stays put, the boundary is a robust diagnostic.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the Bose glass phase and the qualitative SF/MI/BG taxonomy that the paper is testing."},{"cited_title":"Pollet, N","cited_arxiv_id":null,"evidence_quote":"Supplies the theorem of inclusions and the Eg≤Δ energy-gap criterion that the paper compares against."},{"cited_title":"Phase Diagram and Spectral Function of the Two-Dimensional Disordered Bose-Hubbard Model: A Real-Space Dynamical Mean-Field Theory Analysis","cited_arxiv_id":"2509.01230","evidence_quote":"Provides the Monte Carlo unit-filling phase diagram used as the quantitative benchmark for both boundaries."},{"cited_title":"Khellil and A","cited_arxiv_id":null,"evidence_quote":"Introduces the Edwards-Anderson order parameter as a Bose-glass diagnostic, the basis for the paper's MI-BG criterion."},{"cited_title":"Buonsante, V","cited_arxiv_id":null,"evidence_quote":"Supplies the percolation-analysis idea that the paper adapts to the condensate order parameter for the SF-insulator transition."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the strong-coupling prediction of damped-localized excitations that the spectral function is compared against."},{"cited_title":"Anders, E","cited_arxiv_id":null,"evidence_quote":"Derives the bosonic dynamical mean-field equations and Anderson-impurity mapping that RBDMFT builds on."},{"cited_title":"Snoek and W","cited_arxiv_id":null,"evidence_quote":"Provides the real-space BDMFT equations for inhomogeneous systems that the paper uses to treat disorder site-dependently."}],"review_version":1}