{"id":"1b755d2a-2dd3-4349-ac92-faf878021e81","arxiv_id":"1909.01375","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Persistent self-propulsion of non-convex hard particles on a lattice produces gel-like arrested states, with density profiles described by an ASEP-based mean-field theory.","lead":"Simulations show that self-propelled, cross-shaped hard particles can freeze into gel-like networks and empty voids without any attractive forces. The authors explain the late-time structure with a known exclusion-process model and predict how the gel layer thickens as activity drops.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Quantitative ASEP claim rests on a fitted D/alpha; the mean-field current with a single alpha and no N3 correlations is untested, so the abstract overstates quantitative agreement.","rationale":"The paper has two intertwined claims: (i) persistent activity of repulsive hard crosses produces gel-like arrested states, and (ii) in the DR -> 0 limit the ASEP-based coarse-grained model quantitatively predicts the density profiles. The simulations give credible support for (i): arrested runs show gamma(t) -> 0 while non-arrested runs show gamma -> 2, the snapshots display percolated immobile networks, and the DR = 0.005 runs reproduce the gel morphology at high activity. The load-bearing weakness is in (ii). The coarse-grained currents are mean-field ASEP currents with an unspecified constant alpha, and the extended N3 exclusion is explicitly neglected. The only quantitative comparison is a fit of the predicted interface width to the same simulation data, fixing D/alpha. Thus the theory is not independently predictive for the density profiles, and the phase boundaries built on it are not quantitative, as the authors concede in Section VIII. This does not invalidate the gelation claim, but it means the manuscript should be accepted only with the quantitative claim softened and with code/data or an independent test of the current. The reader's weakest_assumption points to the same ASEP mean-field current; I agree and sharpen it by noting the fitted prefactor. This supports the existing CONDITIONAL verdict; I see no grounds to reject or to accept outright.","tokens_in":23123,"tokens_out":11068,"duration_ms":113333,"concrete_test":"Run a planar-interface simulation at DR = 0 and directly measure the active and diffusive currents between coarse-grained columns: bin hop rates along and against the density gradient and the local density rho(x). Fit JA = alpha * DeltaV * rho * (rho_solid - rho) to these currents and check whether a single alpha describes the data at several (rho, DeltaV); independently measure D from the passive tracer diffusivity. Then compare the predicted xi = D/(alpha * DeltaV) with the observed interface width without fitting. Also compute adjacent-row density correlations across the interface to quantify the neglected N3 correlations. If alpha varies with rho or DeltaV, or if correlations are large, the quantitative ASEP prediction is not established.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claim—that the ASEP coarse-grained model in Section V quantitatively predicts the density profiles of the gel states—is undercut by the way the comparison is made. The active current is written as JA = alpha * DeltaV * rho * (rho_solid - rho) (Eq. 5) with alpha an undetermined proportionality constant, and the interface width is xi = D/(alpha * DeltaV) (Eq. 7). The only quantitative test, Fig. 8, fits xi = 10/DeltaV to the simulated widths, thereby fixing D/alpha from the same data it is then said to predict. This tests the DeltaV^{-1} scaling but not the prefactor. The phase boundaries in Section VI (Eqs. 10 and 13) inherit this fitted constant and also rely on approximating the sigmoidal solution of Eq. (6) by a linear profile. Section V itself notes that adjacent-row exclusion and lateral diffusion give rise to correlations, which we ignore, and Section VIII concedes the ASEP predictions fail at low activity and high densities because only simple exclusion is included. Appendix C restricts the predicted phase diagram to large enough activities. The simulation-based gelation claim is not undermined, but the abstract's quantitative statement is an overclaim: the theory is a one-parameter fit with an acknowledged, untested mean-field current.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies a lattice model of self-propelled hard-core cross-shaped particles (the N3 model) with tunable activity and rotation rate. In the passive limit the model shows a first-order fluid-solid transition and glassy aging. Using kinetic Monte Carlo simulations, the authors show that persistent activity induces two classes of dynamically arrested states: aging glass-like states at high density and gel-like states at lower density, characterized by void-solid coexistence with a wetting active-fluid layer. They propose a coarse-grained mean-field ASEP description that yields a wetting lengthscale xi = D/(alpha DeltaV) and a predicted phase diagram with active-liquid, solid-active-liquid, and solid-active-liquid-void phases. They also analyze finite rotation rates and finite-size effects.","tokens_in":23417,"tokens_out":5933,"duration_ms":56338,"significance":"The simulation-based classification of arrested states into aging glasses and gels is a valuable contribution, and the demonstration of activity-induced gelation without attractions in a non-convex lattice model is conceptually important. The mapping to ASEP is an interesting theoretical step, and the paper offers testable predictions for the DeltaV^{-1} scaling of the interface width, the finite-size dependence of phase boundaries, and a crossover rotation rate. The dynamical-heterogeneity analysis based on stationary times and gamma(t) is careful and well presented. However, the central quantitative claim is weakened because the prefactor D/alpha is fitted to the same data that the theory is said to predict, and the mean-field current is an untested simplification. The phase-boundary derivation also contains a mass-balance inconsistency that needs correction.","major_comments":[{"comment":"The claim that the ASEP coarse-grained model 'quantitatively predicts' the density profiles is not fully supported: the lengthscale xi = D/(alpha DeltaV) contains an undetermined proportionality constant D/alpha, and Fig. 8 fits this constant (xi = 10/DeltaV) to the same simulated widths that the theory is said to predict. The theory therefore tests the DeltaV^{-1} functional form but not the prefactor, and Eqs. (10) and (13) inherit the fitted value when constructing the phase boundaries. The Abstract and Section V should either derive D/alpha from first principles or explicitly state that the theory has one fitted parameter, and soften 'quantitatively predicts' accordingly.","section":"Section V (Eq. (7)), Fig. 8, and Abstract"},{"comment":"The mass balance equation N = rho_solid (xi + l_solid) is inconsistent with the linear density profile shown in Fig. 7. If the interface is a ramp from zero to rho_solid over width xi, its mass is approximately rho_solid xi / 2, not rho_solid xi; as written, Eq. (8) overcounts the interface mass by a factor of two. Since Eqs. (10) and (13) are derived from this balance, the theory of the phase boundaries needs to be re-examined. Please clarify the definition of the interface profile and correct the mass balance.","section":"Section VI (Eq. (8))"},{"comment":"The mean-field current JA = alpha DeltaV rho (rho_solid - rho) assumes that the extended N3 exclusion reduces to simple exclusion with a single constant alpha and no correlations from adjacent rows. This assumption is acknowledged in Section V and Section VIII, where the authors note that the ASEP predictions fail at low activity and high densities, but the paper does not provide a direct test of the assumed current against simulation measurements. Since the quantitative claim rests on this functional form, the authors should either validate the current-density relation in the simulations or restrict the quantitative prediction to the regime where the assumption is expected to be accurate.","section":"Section V (Eqs. (4)-(6)) and Section VIII"}],"minor_comments":[{"comment":"The word 'heterogenities' appears in the Abstract and in Section III (e.g., 'appearance of density heterogenities'); it should be 'heterogeneities'.","section":"Abstract and Section III"},{"comment":"The phrase 'the physics in of void-solid coexistence' appears to have a missing word or a typo; please revise.","section":"Section VI, last paragraph"},{"comment":"The caption states that the dashed black line is the best fit xi = 10/DeltaV but does not report the number of data points, error bars, or the fitting procedure; please include these details.","section":"Fig. 8 caption"},{"comment":"The arrest threshold gamma(tmax) < 0.5 is an arbitrary choice; please discuss the sensitivity of the phase diagram in Fig. 1(g) to this threshold and to the maximum simulation time tmax.","section":"Section IV (Eq. (2))"},{"comment":"The coarse-graining box size L^2/900 (15x15 lattice sites) is used for the density profiles but its choice is not justified; a brief comment on how the results depend on this coarse-graining scale would be helpful.","section":"Section II and Fig. 6"},{"comment":"The notation D0 and D(ρ) are used in the same equation without clarification; using a consistent notation for the density-dependent diffusion coefficient would improve readability.","section":"Eq. (3)"}],"recommendation":"major_revision","confidential_remarks":"This is a solid simulation study with a useful phenomenological classification of active arrested states. The main reservation is that the quantitative ASEP claim is stronger than the evidence supports: the prefactor D/alpha is fitted, and the phase-boundary derivation appears to have a factor-of-two mass-balance error. These issues are fixable in revision, so I do not recommend rejection. Please ask the authors to correct Eq. (8) and to carefully restate the predictive status of the theory in the Abstract and Section V."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this paper does something genuinely new — persistent activity alone, with no attraction, drives repulsive hard crosses on a lattice into gel-like arrested states — and the ASEP mapping gives a clean way to think about the void-solid-active-liquid morphology. The central simulation claim holds. The soft spot is the quantitative theory: the advertised prediction of the density-profile lengthscale carries a fitted prefactor, and the phase boundaries inherit it. The word \"quantitatively\" in the abstract is stronger than what the data support.\n\nThe simulations are the real contribution. Using stationary-time distributions and MSD log-derivatives, the authors distinguish aging glassy states from arrested gel states with percolated immobile clusters, and they show that activity shifts the onset of arrest to lower densities. The void-solid coexistence with a wetting active-liquid layer is clearly documented, and the finite-size study in Appendix C is a nice check that the void phase is not a small-system artifact. The finite-rotation-rate section is mostly qualitative but consistent. The paper is also honest: Section VIII states plainly that the ASEP-based predictions fail at low activity and high densities because only simple exclusion is included, and Appendix C restricts the predicted phase diagram to large enough activities.\n\nWhere the paper oversells is the comparison between the coarse-grained theory and simulation. Equation (7) predicts xi = D/(alpha DeltaV), and Fig. 8 shows \"near perfect agreement\" with a line xi = 10/DeltaV. But that line fixes the single constant D/alpha from the same data being compared, so the test is of the DeltaV^{-1} scaling, not of the prefactor. The phase boundaries in Eqs. (10) and (13) then reuse D/alpha = 10 and additionally approximate the sigmoidal solution of Eq. (6) by a linear profile. None of this undermines the classification of arrested states, but it does mean the theory is a one-parameter fit with an untested mean-field current, not an independent quantitative prediction. The abstract should say \"captures the scaling\" or \"predicts the form\" rather than \"quantitatively predicts.\" The other thresholds (gamma<0.5, epsilon, tmax) are operational choices and are not a problem on their own.\n\nWho should read it: anyone working on active glasses, MIPS, gelation, or lattice models of active matter. It belongs in the conversation on whether activity can stand in for attraction. I would send it to peer review; the right referee ask is to have the authors reframe the quantitative claim, add error estimates on the phase boundaries, and release code/data. Ideally they would also give an independent determination of D/alpha — from a single-particle response or a separate geometry — rather than fitting it. Verdict from me: conditionally useful, more for the simulation phenomenology and the scaling argument than for the fitted phase diagram.","headline":"A solid simulation study of activity-induced gelation without attraction; the ASEP scaling argument is elegant, but the quantitative prediction rests on a fitted prefactor, so the abstract overstates the theory.","tokens_in":23935,"tokens_out":2649,"would_cite":true,"duration_ms":27443,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Persistent activity makes hard repulsive particles form gel-like arrested states, just as attractions do in passive colloids.","keywords":["active matter","motility-induced phase separation","kinetic arrest","gelation","glassy dynamics","asymmetric simple exclusion process","hard-core lattice gas","nonconvex colloids"],"falsifier":"In the zero-rotation limit, measure the density profile of an arrested state at fixed global density for two activities $\\Delta v_1$ and $\\Delta v_2$; the claim predicts interface widths satisfying $\\xi_2/\\xi_1 = \\Delta v_1/\\Delta v_2$ and a solid-onset boundary moving as $1/\\rho$. Any observed interface width not proportional to $1/\\Delta v$, or arrest without a sigmoidal fluid-solid interface, would rule out the ASEP-based description.","tokens_in":22925,"feed_emoji":"🕸️","tokens_out":12172,"duration_ms":110713,"temperature":0.7,"pith_summary":"Could self-propulsion replace the attractive forces that make colloids gel? The paper studies hard, cross-shaped particles on a square lattice—a shape whose interlocking blocks rotation—and shows that persistent active motion drives them into a percolated network of nearly immobile particles, structurally and dynamically resembling gels formed by attractive colloids. Activity lowers the density at which glassy dynamics set in, and in the infinite-persistence limit the system passes from an active fluid or an aging glass into arrested solid-active-liquid-void states. A coarse-grained model based on the asymmetric exclusion process predicts the density profile of these arrested states, including the width of the mobile layer that wets the interface between solid and void. The consequence, if the picture holds, is a concrete route to gelation without any attractive interaction.","feed_headline":"Persistent motion freezes colloids into gels with no attraction","feed_subtitle":"Hard cross-shaped particles freeze into percolated networks; a simple model predicts the width of the fluid layer between solid and void.","key_machinery":"The central object is an active lattice gas of hard, cross-shaped particles whose shape imposes exclusion up to third nearest neighbours (the N3 model) and locks rotations. The argument is carried by viewing the late-time, infinitely persistent dynamics through a one-dimensional asymmetric simple exclusion process (ASEP) oriented perpendicular to a fluid-solid interface, and balancing a diffusive current $J_T = -D\\,\\partial\\rho/\\partial x$ against an active current $J_A = \\alpha\\,\\Delta v\\,\\rho(\\rho_{\\mathrm{solid}}-\\rho)$. The steady-state balance yields the spatial logistic equation $\\partial\\rho/\\partial x = (\\alpha\\,\\Delta v/D)\\,\\rho(\\rho_{\\mathrm{solid}}-\\rho)$, whose sigmoidal solutions have a single activity-controlled length scale $\\xi = D/(\\alpha\\,\\Delta v)$; particle-number conservation then fixes when solid slabs and void regions can fit in the box, producing the predicted phase boundaries.","core_discovery":"In a lattice gas of hard crosses with strong rotational locking, activity acts as an effective attraction. As the self-propulsion strength $\\Delta v$ grows, the system passes from a passive fluid through aging, glassy states into arrested configurations in which a percolated, nearly immobile solid-like network coexists with voids and with a mobile active fluid wetting the interface; the authors identify this as the first activity-induced transition from a repulsive glass to a gel. The arrest is tracked by the log-derivative $\\gamma(t)$ of the mean-squared displacement, which falls to zero in arrested runs, and by bimodal distributions of stationary times that appear at lower densities than in the passive system. In the zero-rotation limit the late-time density profile follows a spatial logistic equation obtained from asymmetric-exclusion mean-field currents, giving an interface width $\\xi = D/(\\alpha \\Delta v)$ and phase boundaries for the solid-active-liquid and solid-active-liquid-void regions that match the simulated density fields. Small finite rotation rates leave the qualitative picture intact, with empty voids replaced by a dilute gas and a two-phase binodal similar to that of active dumbbells.","pith_inferences":["As an extension not pursued in the paper, the same ASEP balance suggests a design rule: any persistent, non-rotating active species with strong excluded-volume frustration should develop void-solid coexistence with an interface width proportional to $1/\\Delta v$, even if the particle shape is not cross-like.","A direct test the paper leaves open is to extract the ratio $D/\\alpha$ from one measured density profile and use it to predict all other phase boundaries, checking the theory's internal consistency.","The Section VII crossover estimate implies a measurable experimental knob: increasing rotational diffusion should erase voids and restore ordinary fluid-solid coexistence, which could be tested in shaped colloidal swimmers.","The finite-size divergence predicted by the theory implies that confinement could select between homogeneous and void-containing arrested states, a possible route to active-gel patterning in small devices."],"forward_implications":["Activity lowers the onset density for glassy dynamics: bimodal stationary-time distributions and aging appear at densities below the passive value $\\rho \\approx 0.1625$.","The interface width diverges as activity is reduced ($\\xi \\sim 1/\\Delta v$), so at weak persistence the active-liquid layer fills the system and the arrested void-solid morphology disappears.","The void size is set by global density and activity, not by coarsening time; the arrested phase separation does not grow without bound.","With a small finite rotation rate the empty voids are replaced by a low-density gas, yet at high activity a percolated arrested solid still forms because rotational locking suppresses reorientation.","Because the predicted boundaries depend on system size $L$, smaller lattices can appear homogeneous where larger lattices develop voids, a finite-size signature that can be checked directly."],"supporting_citations":[{"why":"Provides the hydrodynamic ASEP coarse-graining procedure from which the steady-state density profile and interface width are derived.","marker":"[33]"},{"why":"Supplies the mean-field active and diffusive currents used in the steady-state balance that yields the logistic density equation.","marker":"[64]"},{"why":"Defines the active hard-cross dynamics and supplies the single-tracer ballistic-to-diffusive crossover and rotation-locking rates used throughout.","marker":"[45]"},{"why":"Establishes the glass transition in the passive hard-cross system that serves as the high-density baseline for activity-induced arrest.","marker":"[38]"},{"why":"Provides the passive fluid-solid transition and sublattice-ordered phases that the active system builds on.","marker":"[35]"},{"why":"Supplies the random sequential adsorption and diffusion protocol used to prepare uniform, disordered initial states.","marker":"[37]"},{"why":"Is the active lattice-gas MIPS model whose dynamics and single-tracer mean-square displacement are adapted to the hard crosses.","marker":"[11]"},{"why":"Provides the MSD log-derivative arrest criterion used to classify states as arrested or non-arrested.","marker":"[44]"},{"why":"Supplies the gel-bubble and arrested-phase-separation phenomenology from attractive colloids that the active states are compared with.","marker":"[32]"},{"why":"Reports long-persistence-limit clustering and void-like fluctuations in active Brownian particles, motivating the zero-rotation study here.","marker":"[28]"}],"fun_headline_variants":["No attraction, no problem: activity freezes colloids into gels","Persistent motion turns hard particles into arrested gels","Activity as effective attraction: gels from self-propelled crosses","Pure activity drives gelation in repulsive colloids"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The quantitative theory assumes that, once averaged over large boxes, the density obeys the same simple current equations as a one-lane traffic model, with only one fitted constant and with the extra blocking caused by cross arms ignored; the authors note this fails at low activity and high density.","fun_headline_variants_meta":{"raw":{"variants":["No attraction, no problem: activity freezes colloids into gels","Persistent motion turns hard particles into arrested gels","Activity as effective attraction: gels from self-propelled crosses","Pure activity drives gelation in repulsive colloids"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000226,"raw_usage":{"total_tokens":1493,"prompt_tokens":996,"completion_tokens":497,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":612,"completion_tokens_details":{"reasoning_tokens":431}},"tokens_in":612,"tokens_out":497,"duration_ms":5957,"temperature":1.0,"reasoning_tokens":431,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T05:19:36.480719+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"In the zero-rotation limit, measure the density profile of an arrested state at fixed global density for two activities $\\Delta v_1$ and $\\Delta v_2$; the claim predicts interface widths satisfying $\\xi_2/\\xi_1 = \\Delta v_1/\\Delta v_2$ and a solid-onset boundary moving as $1/\\rho$. Any observed interface width not proportional to $1/\\Delta v$, or arrest without a sigmoidal fluid-solid interface, would rule out the ASEP-based description.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the hydrodynamic ASEP coarse-graining procedure from which the steady-state density profile and interface width are derived."},{"cited_title":"Bodineau and B","cited_arxiv_id":null,"evidence_quote":"Supplies the mean-field active and diffusive currents used in the steady-state balance that yields the logistic density equation."},{"cited_title":"Chatterjee, N","cited_arxiv_id":null,"evidence_quote":"Defines the active hard-cross dynamics and supplies the single-tracer ballistic-to-diffusive crossover and rotation-locking rates used throughout."},{"cited_title":"Eisenberg and A","cited_arxiv_id":null,"evidence_quote":"Establishes the glass transition in the passive hard-cross system that serves as the high-density baseline for activity-induced arrest."},{"cited_title":"Nath and R","cited_arxiv_id":null,"evidence_quote":"Provides the passive fluid-solid transition and sublattice-ordered phases that the active system builds on."},{"cited_title":"Eisenberg and A","cited_arxiv_id":null,"evidence_quote":"Supplies the random sequential adsorption and diffusion protocol used to prepare uniform, disordered initial states."},{"cited_title":"Whitelam, K","cited_arxiv_id":null,"evidence_quote":"Is the active lattice-gas MIPS model whose dynamics and single-tracer mean-square displacement are adapted to the hard crosses."},{"cited_title":"Khalil, A","cited_arxiv_id":null,"evidence_quote":"Provides the MSD log-derivative arrest criterion used to classify states as arrested or non-arrested."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the gel-bubble and arrested-phase-separation phenomenology from attractive colloids that the active states are compared with."},{"cited_title":"Extreme active matter at high densities","cited_arxiv_id":"1902.05484","evidence_quote":"Reports long-persistence-limit clustering and void-like fluctuations in active Brownian particles, motivating the zero-rotation study here."}],"review_version":1}