{"id":"e27b859a-1d50-472f-ac60-756af513cb8c","arxiv_id":"1908.04144","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Numerical statDMFT simulations of the 2D disordered Hubbard model show finite-size hysteresis, coexisting metallic and insulating puddles, and a trend toward smearing of the first-order Mott transition in the thermodynamic limit.","lead":"This paper uses a computer method called statistical dynamical mean-field theory to study how disorder and electron interactions together turn a 2D metal into an insulator. It finds coexisting metallic and insulating patches that multiply as the system grows, supporting the idea that the sharp transition disappears in large samples.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The thermodynamic-limit smearing claim rests on an untested RFIM analogy plus qualitative L≤20 statDMFT snapshots; the site-local self-energy may underestimate the interface energy that Imry–Ma balances, so the extrapolation is a conjecture.","rationale":"The finite-size statDMFT calculations are a legitimate and internally consistent numerical study, and the disorder–puddle correlation in Table I is a genuine supporting observation. The method is exact in the U=0 limit, and the authors do not overclaim the transport calculation beyond its incoherent regime. However, the paper's headline claim—that the first-order transition is smeared in the thermodynamic limit—depends on a step that the data do not directly test. With L≤20 and no quantitative finite-size scaling, the only bridge from the simulations to the thermodynamic limit is the Imry–Ma/RFIM analogy. That analogy is reasonable in broad terms, but the statDMFT approximation is not guaranteed to represent the interfacial free energy that controls the Imry–Ma balance; the site-diagonal self-energy ansatz is engineered for local physics and may systematically soften domain walls. This is a concrete correctness risk, not merely a disagreement with consensus. The reader's conditional verdict correctly captures this: the finite-size results are credible, but the central extrapolation goes beyond the evidence. I therefore see no reason to move the verdict; the concern is the same one the reader identified, and the proposed surface-tension test would settle whether the concern actually lands.","tokens_in":15377,"tokens_out":8274,"duration_ms":103406,"concrete_test":"Compute, within the same statDMFT framework and parameters (W=0.52D, T=0.024D, U=2.27D), the excess free energy of a constrained planar metal–insulator interface in L×L cells (L=8, 12, 16, 20): initialize one half in the converged metallic solution and the other half in the insulating solution, relax the self-consistency loop, and extract the surface tension σ(L) from the excess grand potential per unit interface length relative to the homogeneous coexisting states. Then compare the random-field energy gain ∼W√N to σ(L)L at each L. If the interface energy is never subordinate to the random-field energy for L≤20, the observed bubble proliferation is a finite-size artifact and the thermodynamic smearing claim is not supported; if the random-field term dominates already at L≤10, the Imry–Ma extrapolation is internally consistent.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central assertion—that the 2D disordered Mott transition is smeared in the thermodynamic limit—is supported only by qualitative finite-size data (Figs. 10–11) for L≤20, with no disorder averaging, no bubble-size scaling, and the paper's own admission that L>20 is inaccessible and that no full scaling analysis of puddle sizes is possible. The extrapolation is carried entirely by the Imry–Ma/RFIM analogy of Sec. IV, introduced 'by analogous reasoning' in Eq. (14). That analogy would be valid for a short-range-coupled first-order transition in a random field, but the computed object is a statDMFT solution with a site-diagonal self-energy (Eq. 5). In this approximation the metal–insulator interface free energy—the quantity Imry–Ma balances against the random-field energy gain ∼W√N—is not an input; it is an emergent property that may be severely underestimated precisely because non-local correlations are dropped. If the method's local approximation artificially lowers the domain-wall stiffness, the observed bubble proliferation at L≤20 is an artifact of the solver, not a thermodynamic symptom. The paper provides no diagnostic separating these two possibilities. Thus the 'signaling the smearing' step is the least secure link in the argument; the transport section's ImΣ(iω1)≈ImΣ(ω→0) identification (Eqs. 17 and 22) is also heuristic, but it affects a secondary application and is already acknowledged to fail in the insulator.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the site-disordered two-dimensional Hubbard model at half filling using statistical dynamical mean-field theory (statDMFT) with a Hirsch-Fye quantum Monte Carlo impurity solver, for lattice sizes up to L=20, disorder width W=0.52D, and three temperatures near the clean Mott endpoint. The authors report clean and disordered hysteresis loops, spinodal lines, and a disordered-induced shift and shrinking of the coexistence region. They show spatial maps of the local Green's function and self-energy that display metallic and insulating puddles, and they correlate the puddle character with local disorder fluctuations. They interpret the finite-size increase in puddle proliferation as evidence for Imry-Ma smearing of the first-order transition in the thermodynamic limit. Finally, they map each lattice site to a classical resistor value set by ImΣ_i(iω1) and compute average currents through the resulting resistor network.","tokens_in":15778,"tokens_out":3717,"duration_ms":45789,"significance":"If the central claim holds, the paper would establish a disorder-induced smearing of the two-dimensional Mott transition and connect the resulting puddle landscape to the random-field Ising universality class, with direct relevance to nano-imaging experiments on VO2 and other inhomogeneous strongly correlated systems. The statDMFT implementation with site-resolved QMC impurity solutions, the systematic finite-size comparison up to L=20, and the first statDMFT-based transport calculation via a classical resistor network are useful technical contributions. However, the thermodynamic-limit conclusion currently rests on an extrapolation from small systems and an imported Imry-Ma/RFIM analogy rather than on a derived scaling analysis, so the significance is contingent on strengthening that step.","major_comments":[{"comment":"The transport calculation rests on two unverified identifications: ImΣ_i(iω1) is used as a proxy for ImΣ(k∼k_F, ω→0), and the Fermi velocity is estimated as v_F∼aD with k_F∼1/a. The first identification is particularly fragile because the first Matsubara frequency is not the zero-frequency limit, and the second estimate is order-of-magnitude at best. The paper acknowledges that the resulting description fails in the insulator because ImΣ(iω1) cannot produce the activated temperature dependence of the Mott gap, yet Fig. 14 is then used to draw conclusions about the conductance in the insulating regime. Since the transport section is a secondary application, this issue does not by itself invalidate the main puddle picture, but the claims should be restricted to the metallic and coexistence regions, with the insulator behavior presented as a known limitation rather than as a computed result.","section":"Section V, Eqs. (17)-(22) and Section VI C"}],"minor_comments":[{"comment":"The caption says 'As the temperature increases, the hysteresis loops become smaller,' but Fig. 10 shows hysteresis loops for different lattice sizes at fixed temperature; the caption should instead describe the dependence on system size.","section":"Fig. 10 caption"},{"comment":"Reference 66 is incomplete ('See and e, g, World Scientific p. 277 (1997)'); it needs a full citation with authors, title, and publisher information.","section":"Reference 66"},{"comment":"The reference to Fetter and Walecka is listed as 'Philos. Mag. 21, 863 (1970)'; this appears to be a misattribution, as the standard citation is the textbook 'Quantum Theory of Many-Particle Systems'.","section":"Reference 72"},{"comment":"The notation in Eq. (14) is unclear: ∆ϵ is written with an awkward square-root expression and the definition of N (the number of sites in a region) is not given before the equation; please rewrite the definition cleanly.","section":"Eq. (14)"},{"comment":"The second-neighbor hopping is taken to be purely imaginary (t* = 0.5 i t); the authors should justify this unusual choice more explicitly, since an imaginary hopping is not a standard single-band lattice parameter and its physical interpretation is not explained.","section":"Section II, Eq. (2)"}],"recommendation":"major_revision","confidential_remarks":"The central Imry-Ma smearing claim is presented with more confidence than the data support, and the RFIM analogy is imported from earlier work rather than derived for the disordered Hubbard model. The transport section is heuristic but clearly marked as a first attempt. I would like the authors to either add a scaling analysis that makes the thermodynamic-limit claim quantitative or explicitly demote that claim to a conjecture; either course would make the paper acceptable in principle. The paper is otherwise within scope and contains useful methodological contributions."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my read of Suárez-Villagrán et al. The genuinely new thing is the first statDMFT transport calculation via a classical resistor network, plus the finite-size puddle statistics for the 2D disordered Hubbard model. The spinodals, hysteresis loops, and bubble patterns look like a careful application of a known method. The puddle–disorder correlation in Table I is a nice consistency check; it is not fitted to the conclusion. The paper is also refreshingly explicit about what it cannot do: it says L>20 is out of reach, that no full scaling analysis of puddle sizes is possible, and that the transport description fails in the insulator.\n\nThe soft spot is exactly where the reader and the stress-test note point. The smearing claim in the thermodynamic limit is extrapolated from L≤20 snapshots and hysteresis rounding, with no disorder averaging, no scaling collapse, no error bars. The extrapolation is carried by an analogy to the random-field Ising model, introduced in Eq. (14) by “analogous reasoning.” That analogy is plausible for a short-range first-order transition in a random field, but here the interface energy—the thing Imry-Ma balances—is emergent from a site-diagonal self-energy that drops non-local correlations. You can worry that the local approximation artificially lowers domain-wall stiffness, so the proliferation of bubbles at L≤20 could be a solver artifact rather than a thermodynamic symptom. The paper gives no diagnostic that separates those two readings. So the phrase “signaling the smearing” is doing more work than the data support; I would call it a conjecture, not a demonstration.\n\nThe transport section is a legitimate first step but heuristic: replacing the zero-frequency self-energy by ImΣ(iω1) and taking vF~aD and kF~1/a is a heavy-handed estimate, and the paper admits the mapping fails in the insulator. That is a minor-to-moderate issue because it affects the secondary application, not the main finite-size story.\n\nOverall: this is a useful, incremental contribution worth a serious referee. I'd send it to review, but the referee should push for a more honest statement of what the finite-size data do and do not show, and ideally for a scaling analysis or at least error quantification. The paper is for people working on disordered Mott systems and interpreting nanoscale puddle images; they will get real value from the method and the puddle-disorder correlation, and they should take the Imry-Ma smearing claim as a hypothesis.","headline":"A solid, honest statDMFT study whose finite-size puddle results are credible, but the thermodynamic-limit smearing claim rides on an Imry-Ma analogy the paper does not actually test.","tokens_in":16257,"tokens_out":2380,"would_cite":true,"duration_ms":22661,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["82B44","82B26","82B27","81V70"],"pacs":["71.30.+h","72.15.Rn","71.10.Fd"],"model":"deepseek-v4-flash","headline":"In two dimensions, disorder converts the Mott metal-insulator transition into a proliferating landscape of metallic and insulating bubbles, smearing the transition in the thermodynamic limit.","keywords":["disordered Hubbard model","Mott metal-insulator transition","statistical dynamical mean-field theory","Imry-Ma theorem","random-field Ising universality class","metal-insulator puddles","classical resistor network","finite-size effects"],"falsifier":"A decisive check is to measure the width of the hysteresis and coexistence region as a function of lattice size at fixed temperature and disorder within the same statDMFT calculation. The paper's claim predicts that this width shrinks toward zero as L grows, with both bubble types proliferating; if instead the width saturates to a nonzero value for L larger than about 20, or if only one bubble type appears, the Imry-Ma smearing claim is not correct. The same criterion could be tested in controlled-disorder experiments by looking for the disappearance of a sharp resistance jump as sample size increases.","tokens_in":15176,"feed_emoji":"⚛️","tokens_out":6774,"duration_ms":68783,"temperature":0.7,"pith_summary":"This paper studies the disordered Hubbard model in two dimensions and argues that disorder fundamentally changes the Mott metal-insulator transition. In a finite sample the transition keeps its first-order character, complete with hysteresis and coexisting metallic and insulating solutions, but as the lattice grows, metallic and insulating bubbles multiply until the sharp transition is smeared away in the thermodynamic limit. The mechanism is the same Imry-Ma argument that destroys long-range order in the two-dimensional random-field Ising model, so the paper presents the result as a generalization of the Imry-Ma theorem to the Mott case. It also shows that in this incoherent regime transport can be described by a classical resistor network in which each site's resistance is set by the imaginary part of its local self-energy at the first Matsubara frequency. If correct, the picture explains nanoscale metallic and insulating puddles seen in materials like VO2 as the intrinsic fate of a disordered two-dimensional Mott system.","feed_headline":"Disorder smears 2D Mott transition into metal-insulator puddles","feed_subtitle":"Simulations show metallic and insulating bubbles multiply with sample size, rounding the first-order jump away.","key_machinery":"The central object is the Statistical Dynamical Mean-Field Theory (statDMFT), in which each lattice site carries its own local self-energy Sigma_i(i omega_n), so disorder is kept site-resolved while interactions are treated locally as in DMFT; the impurity problems are solved with a Quantum Monte Carlo algorithm. The order parameter used throughout is the imaginary part of the local self-energy at the first Matsubara frequency, ImSigma_i(i omega_1), which acts simultaneously as a metal/insulator classifier and, through the inelastic mean-free-path estimate l_in/a = D/ImSigma, as the local resistivity in a classical resistor network. The Imry-Ma energy balance of Eq. (14), comparing disorder fluctuations in a region with a critical local metallization scale, is what converts the random-field Ising analogy into a prediction that bubbles proliferate with system size.","core_discovery":"The paper's central claim is that site disorder in two dimensions destroys the first-order Mott transition in the thermodynamic limit. Within statDMFT, finite-size lattices still show hysteresis and coexistence, with per-site loops shifted to higher interactions, but increasing the system size at a fixed interaction produces more metallic bubbles inside the insulator and more insulating bubbles inside the metal, so no sharp jump survives as L grows. This is asserted to be the generalization of the Imry-Ma theorem to the disordered Hubbard model, relying on the mapping of the Mott transition to the random-field Ising universality class. A secondary claim is that the local imaginary self-energy ImSigma_i(i omega_1) is small in the metal and large in the insulator, and that when the inelastic mean free path is about one lattice constant, the same quantity can be used as a local resistivity to build a classical random resistor network; this yields current maps and average conductances across the transition.","pith_inferences":["If the random-field Ising analogy is right, the smearing should be a lower-critical-dimension effect: in three dimensions the same statDMFT calculation should show a first-order transition surviving weak disorder, with bubbles only near the transition. This is a testable extension the paper does not run.","The resistor-network mapping implies a direct connection between spectroscopy and transport: local ImSigma maps from scanning tunneling experiments on a disordered Mott material could be fed into the same network to predict the spatial current distribution, a check that nano-imaging and transport measurements could carry out jointly.","The persistence of the bubble landscape for a fixed disorder realization at different temperatures suggests that quenched disorder, not thermal nucleation, sets the pattern; if so, sample-specific predictions could be made for repeated imaging of the same device."],"forward_implications":["In a macroscopic two-dimensional disordered sample, there is no true first-order Mott transition: the jump in the local order parameter is replaced by a smooth crossover controlled by the statistics of bubbles.","Bulk transport near the transition is a percolation problem through the coexisting puddle landscape, not a single phase transition; current maps should show spatial texture correlated with local disorder fluctuations.","Disorder shifts the coexistence region to larger interactions and shrinks the hysteresis width, so experimentally the transition appears at higher interaction strength and with reduced metastability.","Because the inelastic mean free path is at most one lattice constant at the studied temperatures, classical resistor-network transport is the appropriate description, and the metallic conductance should track the local-resistivity distribution while the insulating conductance will not show the activated temperature dependence."],"supporting_citations":[{"why":"Introduces the statDMFT method used for all disorder calculations.","marker":"[52]"},{"why":"Provides the Quantum Monte Carlo impurity solver that yields the local self-energies.","marker":"[53]"},{"why":"Supplies the Imry-Ma theorem that the paper generalizes to the Mott transition.","marker":"[68]"},{"why":"Provides the random-field Ising scaling analysis of puddles in VO2 that motivates the universality-class analogy.","marker":"[67]"},{"why":"Establishes that the clean Mott transition belongs to the liquid-gas and Ising universality class, the starting point of the analogy.","marker":"[65]"},{"why":"Reports nanoscale imaging of metallic and insulating puddles in VO2, the experimental phenomenon the bubble picture explains.","marker":"[10]"}],"fun_headline_variants":["2D disordered Mott transition smears into puddles","Disorder kills sharp 2D Mott transition, leaves puddles","Mott transition in 2D: disorder rounds the jump, forms bubbles","Spinodal lines fade with size: 2D Mott transition smeared by disorder","Metallic and insulating puddles replace sharp 2D Mott transition"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the finite-temperature Mott transition in the disordered Hubbard model obeys the same physics as the random-field Ising model, so the Imry-Ma energy balance of Eq. (14) applies; if disorder does not act as a random field on a scalar order parameter, the growing bubble proliferation could be a finite-size effect rather than true thermodynamic-limit smearing. The transport calculation additionally assumes that ImSigma_i(i omega_1) is a valid local resistivity.","fun_headline_variants_meta":{"raw":{"variants":["2D disordered Mott transition smears into puddles","Disorder kills sharp 2D Mott transition, leaves puddles","Mott transition in 2D: disorder rounds the jump, forms bubbles","Spinodal lines fade with size: 2D Mott transition smeared by disorder","Metallic and insulating puddles replace sharp 2D Mott transition"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000392,"raw_usage":{"total_tokens":2087,"prompt_tokens":998,"completion_tokens":1089,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":614,"completion_tokens_details":{"reasoning_tokens":993}},"tokens_in":614,"tokens_out":1089,"duration_ms":9136,"temperature":1.0,"reasoning_tokens":993,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:50:37.372678+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A decisive check is to measure the width of the hysteresis and coexistence region as a function of lattice size at fixed temperature and disorder within the same statDMFT calculation. The paper's claim predicts that this width shrinks toward zero as L grows, with both bubble types proliferating; if instead the width saturates to a nonzero value for L larger than about 20, or if only one bubble type appears, the Imry-Ma smearing claim is not correct. The same criterion could be tested in controlled-disorder experiments by looking for the disappearance of a sharp resistance jump as sample size increases.","supporting_citations":[{"cited_title":"Dobrosavljevic and author G","cited_arxiv_id":null,"evidence_quote":"Introduces the statDMFT method used for all disorder calculations."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the Quantum Monte Carlo impurity solver that yields the local self-energies."},{"cited_title":"Imry and author S","cited_arxiv_id":null,"evidence_quote":"Supplies the Imry-Ma theorem that the paper generalizes to the Mott transition."},{"cited_title":"Liu , author B","cited_arxiv_id":null,"evidence_quote":"Provides the random-field Ising scaling analysis of puddles in VO2 that motivates the universality-class analogy."},{"cited_title":"Kotliar , author E","cited_arxiv_id":null,"evidence_quote":"Establishes that the clean Mott transition belongs to the liquid-gas and Ising universality class, the starting point of the analogy."},{"cited_title":"Qazilbash , author M","cited_arxiv_id":null,"evidence_quote":"Reports nanoscale imaging of metallic and insulating puddles in VO2, the experimental phenomenon the bubble picture explains."}],"review_version":1}