{"id":"9d844d37-d79a-4b55-9117-2d475d4f9e2c","arxiv_id":"1908.03861","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Strong cross-flow suppresses spanwise vortex structures in 3D electro-convection, forcing a hysteretic transition to 2D streamwise rolls that boosts charge transport.","lead":"This paper uses 3D simulations to show how an electric-field-driven fluid instability between two plates changes when a sideways flow is added: strong sideways flow turns three-dimensional vortex cells into two-dimensional rolling vortices. The result matters for engineering systems that use electric fields to stir or pump fluids, such as desalination and heat transfer devices.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Grid-resolution evidence is missing for the sharp Yc/Yf thresholds; the 3D/2D bracket is only 1% wide, so the reported Y values may shift on finer grids.","rationale":"The reader identified the same weakest assumption: the numerical results rely on a single resolution with no grid-convergence study in the main text, and error analysis is deferred to supplementary materials. The manuscript itself flags this missing support by stating 'Error analysis is provided in supplementary materials' (Section V), and the only accuracy claim cites a companion paper [87] rather than a convergence test for the present 3D cross-flow simulations. I considered whether other concerns are more fundamental: the DMD analysis in the nonlinear transition region is used qualitatively and is not the source of the quantitative thresholds, and the Poiseuille averaging in the Y definition is an explicit convention rather than a hidden assumption. The sharp bifurcation brackets in Section V.4 (3.84 vs 3.88 for Couette, 3.92 vs 3.96 for Poiseuille) make the resolution question the most load-bearing, because a 1% change in u* determines whether a case is classified as 3D or 2D, and the reported Yf values inherit that sensitivity. The paper does have independent support: the linear growth rate (~0.896) agrees with prior linear stability analysis and LBM results, and the qualitative suppression of spanwise structures is physically plausible. Therefore the appropriate verdict remains CONDITIONAL, not REJECT or UNVERDICTED: the qualitative claim is credible, but the specific threshold values should be treated as provisional until grid-convergence evidence is supplied.","tokens_in":19588,"tokens_out":9262,"duration_ms":95765,"concrete_test":"Perform a grid-refinement study for the two bracketing Couette cases (u*_wall=3.84 and 3.88) and the two bracketing Poiseuille cases (u*_center=3.92 and 3.96) from Section V.4, using the same square-pattern initialization, cross-flow application protocol, and steady-state criterion. Run each case at the original grid size and at least two finer grids, ideally doubling the wall-normal resolution and proportional x/y resolution. Record whether the final state is oblique-3D or 2D rolling on each grid. If any of the four cases changes classification, or if the inferred Yf shifts by more than approximately 2%, the reported thresholds are not converged. The paper should also report the actual number of grid points used in Figures 17 and 18.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is quantitative: specific thresholds Yc=772.73 and Yf=438.14 for Couette flow, and Yc=300.75 and Yf=213.90 for Poiseuille flow, as reported in Section V.4 and the Conclusion. These values are extracted from TRT-LBM simulations at a single grid resolution, but the main text never states the grid resolution; the domain is given only in physical units (Lx=Ly=1.22 m, H=1 m in Section V.1). The only resolution statements are 'The numerical method was shown to be 2nd order accurate in space [87]' and 'Error analysis is provided in supplementary materials' (Section V), deferring the key evidence to an external companion paper and unreviewed supplementary material. The threshold between the oblique-3D branch and the 2D rolling branch is extremely sharp: in Section V.4 the Couette bifurcation is bracketed by u*_wall=3.84 (3D) and 3.88 (2D), a 0.04 change in u* (about 1%); the Poiseuille bracket is similarly narrow (3.92 vs 3.96). The reported Yf values depend on resolving this boundary. At such a sharp transition, numerical diffusion or insufficient grid resolution can shift the threshold by more than the bracket width, changing the 3D/2D classification of a given case. Because the strongest claim includes specific Y values, the missing grid-convergence evidence is load-bearing.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents three-dimensional lattice Boltzmann (TRT-LBM) simulations of electro-convection between parallel plates with unipolar charge injection, with and without Couette or Poiseuille cross-flow. Using rolling, square, hexagon, and mixed initial perturbations, the authors characterize equilibrium patterns and use dynamic mode decomposition (DMD) of transient velocity fields to extract growth rates and unstable modes. They report that sufficiently strong cross-flow suppresses spanwise (transverse) vortex structures while leaving streamwise (longitudinal) rolls essentially unaffected, leading to a hysteretic 3D-to-2D transition. The hysteresis is parameterized by the dimensionless quantity Y, a ratio of electrical to viscous forces, with reported thresholds Yc=772.73 and Yf=438.14 for Couette flow and Yc=300.75 and Yf=213.90 for Poiseuille flow. The linear growth rate of about 0.896 is compared with prior linear stability analysis [42] and SRT-LBM simulations [49], and the DMD modes are used to interpret the transition.","tokens_in":19878,"tokens_out":9125,"duration_ms":92853,"significance":"If the quantitative thresholds are reliable, this is a useful extension of shear-affected convection studies to three-dimensional electro-convection, with potential connections to Rayleigh-Benard convection and cloud-street formation. The paper's main strengths are its use of multiple initial perturbation patterns, the independent validation of the linear growth rate against established results (Section V.2, Figs. 4-6), and the fact that DMD is applied to simulation data rather than used to fit parameters. The hysteresis characterization through Y is physically appealing and could be practically useful for predicting when transverse rolls are suppressed. However, the central quantitative claims—the Yc and Yf values—rest on a single-resolution numerical campaign, and the manuscript itself defers error analysis to supplementary materials, so the thresholds are not yet fully supported.","major_comments":[{"comment":"The manuscript never states the grid resolution (grid points, lattice spacing, or dx/H) used for the simulations. It only says the method is second-order accurate in space [87] and that 'Error analysis is provided in supplementary materials.' This is load-bearing because the reported thresholds Yc=772.73 and Yf=438.14 (Couette) and Yc=300.75 and Yf=213.90 (Poiseuille) are extracted from bifurcation brackets only 0.04 in u* wide (3.84 vs. 3.88 for Couette; 3.92 vs. 3.96 for Poiseuille). At such a sharp transition, numerical diffusion at a single resolution could shift the 3D/2D classification by more than the bracket width. Please report the resolution used, provide a grid-convergence study (ideally with at least two refined grids), and give an uncertainty estimate for each of the four Y values.","section":"Section V, first paragraph; Section V.4, Figs. 17-18"},{"comment":"The conclusion states that the transition to 2D rolling vortices is 'observed for all initial perturbation schemes and independent of the domain configurations considered in this work,' but the quantitative threshold analysis is performed for a single configuration: square initial perturbation, Lx=Ly=1.22 m, H=1 m, T=170, C=10, M=10, Fe=3500. The oval, hexagonal, harmonic, and mixed cases shown in Fig. 2 are not used to determine Yc or Yf. The generalization should either be explicitly restricted to the tested configuration or supported by threshold determinations for at least one other pattern and domain size.","section":"Section V.4 and Section VI (Conclusion)"},{"comment":"The definition of Y is not developed with enough algebraic detail for the reader to reproduce the reported values. Equation (12) introduces Y as the product Re*X and relates it to shear stress, but the displayed equations do not explicitly show how the dimensionless ratio follows from the dimensional parameters, and the Poiseuille case uses half-height H/2 while the Couette case uses H. Please state the explicit formula for Y in terms of rho0, phi0, mu, u_ext, and the relevant length scale, and report how u*_wall and u*_center were converted into the Y values quoted in Figs. 17-18.","section":"Section II, Eq. (12)"}],"minor_comments":[{"comment":"There are several typos and OCR-style artifacts that should be corrected, including 'spanwis' after Eq. (23), 'cro ss-flow' in the abstract and introduction, and 'it’s' in the conclusion.","section":"Throughout"},{"comment":"The paper should label the two scenarios consistently: cross-flow applied before the perturbation (thresholds u*wall=2.20 and u*center=2.80) versus cross-flow applied after the vortices are established (thresholds u*wall=3.88 and u*center=3.96). As written, the juxtaposition of u*wall=4 and u*wall=3.84 in nearby paragraphs can be misread as contradictory.","section":"Section V.3 vs. Section V.4"},{"comment":"The conclusion uses 'u*max=2.20' for the Couette threshold, while the body text uses 'u*wall=2.20'. Please unify the notation for the characteristic cross-flow velocity.","section":"Section VI"},{"comment":"The captions contain notation errors such as 'z= /2H' and use 't=7.5' with units; since the time is non-dimensionalized, it should be written as t*=7.5.","section":"Fig. 8 and Fig. 15 captions"},{"comment":"The thresholds are quoted to five significant figures (e.g., Yc=772.73). Given that the underlying brackets are only 0.04 in u* wide and no uncertainty analysis is provided, this precision is unsupported; please report fewer digits or provide error bars.","section":"Section V.4, Figs. 17-18"},{"comment":"The DMD window is described as covering the linear growth region (t*=0-7 for Couette, t*=0-6.25 for Poiseuille), but the text earlier states that the linear growth region ends near t*=5. Please justify why the interval extends into the nonlinear regime, or restrict the DMD analysis to the genuinely linear portion.","section":"Section V.3, DMD analysis"}],"recommendation":"major_revision","confidential_remarks":"The main concern is reproducibility: the paper does not state the grid resolution, does not provide the data or code, and gives over-precise thresholds from narrow brackets. I do not see circularity: the DMD analysis is applied to the simulation output, and the growth-rate check against [42,49] is a genuine independent validation. The reliance on [88] for the Y parameter is acceptable but should be summarized in the main text so the present paper is self-contained regarding its central dimensionless group."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Worth a look if you work on electroconvection or shear-modulated pattern formation. The genuinely new pieces are the 3D simulations with Couette and Poiseuille cross-flow, the use of DMD to characterize the linear and transition regimes, and the quantitative hysteresis map in the Y parameter. The paper also does something right that many numerical studies skip: it validates the linear growth rate against published linear stability analysis and against the SRT LBM code, and the agreement (~0.896) is reassuring. The physical story—cross-flow suppresses spanwise rolls while leaving streamwise rolls intact, with a 3D-to-2D transition and an associated jump in Nusselt number—is coherent and consistent with earlier 2D results.\n\nThe soft spots are real but not disqualifying. The main one is grid convergence, or rather the absence of it in the main text. The domain is given in physical units and the text says the method was shown to be second-order accurate elsewhere, but the actual resolution used here is never stated, and the error analysis is deferred to supplementary material. The stress-test note is right that this is load-bearing: the Couette bifurcation bracket is u* = 3.84 vs 3.88, about a 1% window, and the Poiseuille bracket is similarly narrow. At such a sharp transition, numerical diffusion can shift the threshold by more than the bracket width. So the exact values Yc = 772.73, Yf = 438.14, and the Poiseuille equivalents should be treated as provisional until resolution sensitivity is shown. This is not a reason to reject the paper; it is a reason to ask for the convergence study before the numbers become canon.\n\nA smaller issue: DMD is applied in the nonlinear transition region and the extra 'unstable modes' are used as evidence for the 3D structures. That is heuristic—DMD eigenvalues from nonlinear data are not true linear stability eigenvalues—but the authors are careful enough to support the conclusions with direct flow visualization and Nusselt number curves, so it does not undermine the core claim. Code and data are not released; making them available would help, especially because the supplementary convergence material is not part of the preprint.\n\nWho is this for? People working in EHD convection, electrokinetic instabilities, or shear effects on convection roll patterns. It is a solid within-subfield advance, not a paradigm shift. I would send it to a serious referee, with the explicit request that the referee check the grid sensitivity and the Y thresholds. My own verdict would be conditional acceptance pending that convergence evidence.\n\nIf I were in this subfield I would cite it; as is, I would read the revised version before relying on the specific threshold numbers.","headline":"A credible 3D extension of electroconvection with cross-flow, but the sharp reported hysteresis thresholds rest on a single-resolution simulation with no convergence data in the main text.","tokens_in":20436,"tokens_out":1335,"would_cite":false,"duration_ms":18092,"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":"Strong cross-flow turns 3D electro-convection into 2D rolls via a hysteretic transition controlled by the force ratio Y.","keywords":["electro-convection","electrohydrodynamic instability","unipolar charge injection","cross-flow","hysteresis","dynamic mode decomposition","lattice Boltzmann method","electric Nusselt number"],"falsifier":"Repeat the two bifurcation cases (applying shear to an existing square pattern) on a grid with doubled resolution in all directions and compare the $Y$ values at which the oblique 3D state gives way to 2D rolls; if either $Y_c$ or $Y_f$ moves by more than a few percent, the reported thresholds are numerical artifacts rather than fixed physical transition points.","tokens_in":19364,"feed_emoji":"🌀","tokens_out":12541,"duration_ms":117844,"temperature":0.7,"pith_summary":"The paper aims to establish that a cross-flow imposed on three-dimensional electro-convection—electrically driven motion in a dielectric liquid with ions of one sign injected from an electrode—acts as a pattern filter: strong enough Couette or Poiseuille shear destroys the vortex components lying across the flow, while vortex rolls aligned with the flow survive unchanged. The surviving state is a two-dimensional rolling pattern, and the 3D-to-2D transition is hysteretic, controlled by a single non-dimensional parameter $Y$, the ratio of the electrical (Coulomb) force to the viscous force. The result matters because it identifies a simple control knob, shear strength, for switching a convection state and sharply increasing charge transport. The paper reports thresholds of $Y_c=772.73$ and $Y_f=438.14$ for Couette flow and $Y_c=300.75$ and $Y_f=213.90$ for Poiseuille flow.","feed_headline":"Cross-flow snaps 3D electro-vortices into 2D rolls","feed_subtitle":"One ratio of electric to viscous force predicts whether 3D vortices survive or collapse into 2D rolls under shear.","key_machinery":"The load-bearing object is the dimensionless group $Y = \\mathrm{Re}\\,X$, the product of the Reynolds number of the cross-flow and the ratio of electric to inertial force, i.e., the ratio of the Coulomb force to the viscous force in the streamwise momentum equation. It collapses the effect of cross-flow geometry and strength into one number, and the paper locates the bifurcations by sweeping $Y$ while holding other parameters fixed. The second piece is dynamic mode decomposition (DMD), a data-driven method that extracts growth rates and spatial modes from transient velocity snapshots; the presence of extra unstable modes in weak-cross-flow cases marks the oblique 3D state, and their absence in strong-cross-flow cases marks the pure rolling state. Together, $Y$ provides the control parameter and DMD provides the marker for which side of the bifurcation the system is on.","core_discovery":"Electro-convection between parallel plates with unipolar injection can settle into three-dimensional square, oval, hexagonal, or mixed vortex patterns. The central discovery is that a sufficiently strong streamwise cross-flow selectively suppresses the transverse vortex component while leaving streamwise rolls untouched, so the system reorganizes into purely two-dimensional longitudinal rolls. The mechanism is the interaction of the vortex's streamwise velocity component with the bulk shear, strongest near the walls; once only $y$-$z$ rolls remain, the cross-flow no longer couples to them, which is why the streamwise pattern survives at any cross-flow strength. The transition is not reversible: destroying an existing 3D state requires a stronger shear than preventing one from forming, and this hysteresis is visible in the electric Nusselt number, which jumps when the moving pattern becomes two-dimensional.","pith_inferences":["If $Y$ is truly the controlling group, the same hysteresis should be measurable experimentally by sweeping the applied voltage or the cross-flow velocity while recording current: the $Ne$ versus $Y$ curve should trace different branches depending on sweep direction, with the gap equal to $Y_c - Y_f$.","By analogy with shear-affected buoyancy-driven convection, the threshold ratio $Y_c$ might be expressible through a buoyancy-to-inertia ratio, which would let the thresholds be predicted for other body-force convection systems.","DMD's ability to flag the extra unstable modes before the nonlinear transition suggests a data-driven early-warning method: monitoring a few velocity snapshots while ramping the cross-flow could indicate approach to $Y_f$ before the pattern visibly changes.","The independence of the saturated rolling state from cross-flow strength suggests a superposition principle: the streamwise vortex solution plus the base shear flow may be an exact or near-exact nonlinear solution, which a reduced two-amplitude model could test."],"forward_implications":["A shear strong enough that $Y < Y_f$ converts every three-dimensional electro-convective pattern tested—square, oval, hexagonal, and mixed—into the same two-dimensional streamwise rolling state.","Because streamwise rolls simply superimpose on the cross-flow, their saturated amplitude and charge transport are independent of the cross-flow strength and profile once the transition is complete ($Ne = 1.41$).","The 3D-to-2D transition is hysteretic: the shear needed to destroy an existing 3D state is larger than the shear needed to prevent one from forming, so the final state depends on the system's history.","DMD can identify the bifurcation threshold from simulation snapshots, because the oblique 3D state has extra unstable modes that the rolling state lacks.","The electric Nusselt number rises sharply when transverse structures are suppressed, meaning the 2D rolling state transports charge more efficiently than the 3D oblique state."],"supporting_citations":[{"why":"It supplies the linear-stability analysis of electrohydrodynamic flow with and without cross-flow, including the growth rate used for validation.","marker":"[42]"},{"why":"It supplies a unified lattice-Boltzmann model whose hydrostatic profiles and stability results are used as a comparison.","marker":"[48]"},{"why":"It provides the three-dimensional finite-amplitude electroconvection results that the paper compares with for square and rolling pattern growth rates.","marker":"[49]"},{"why":"It gives the linear result that cross-flow leaves longitudinal rolls unaffected and reduces the transverse roll growth rate.","marker":"[80]"},{"why":"It provides the two-relaxation-time lattice-Boltzmann solver coupled to a fast Poisson solver that the three-dimensional code extends.","marker":"[87]"},{"why":"It supplies the two-dimensional cross-flow study that introduced the parameter $Y$ and the hysteresis behavior extended here.","marker":"[88]"}],"fun_headline_variants":["Shear threshold forces 3D electro-vortices into 2D rolls","Cross-flow flattens 3D vortex cells into 2D streamwise rolls","Shear kills transverse rolls but leaves streamwise rolls","Hysteretic shear collapse of 3D electro-vortices to 2D rolls","Shear flips 3D vortex patterns to 2D longitudinal rolls"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the single-resolution numerical grid is fine enough to pin down the reported hysteresis thresholds, since the main text reports a second-order method but defers grid-convergence and error analysis to supplementary material.","fun_headline_variants_meta":{"raw":{"variants":["Shear threshold forces 3D electro-vortices into 2D rolls","Cross-flow flattens 3D vortex cells into 2D streamwise rolls","Shear kills transverse rolls but leaves streamwise rolls","Hysteretic shear collapse of 3D electro-vortices to 2D rolls","Shear flips 3D vortex patterns to 2D longitudinal rolls"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000729,"raw_usage":{"total_tokens":3281,"prompt_tokens":976,"completion_tokens":2305,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":592,"completion_tokens_details":{"reasoning_tokens":2203}},"tokens_in":592,"tokens_out":2305,"duration_ms":17074,"temperature":1.0,"reasoning_tokens":2203,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:59:21.588234+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Repeat the two bifurcation cases (applying shear to an existing square pattern) on a grid with doubled resolution in all directions and compare the $Y$ values at which the oblique 3D state gives way to 2D rolls; if either $Y_c$ or $Y_f$ moves by more than a few percent, the reported thresholds are numerical artifacts rather than fixed physical transition points.","supporting_citations":[{"cited_title":"Traoré and A","cited_arxiv_id":null,"evidence_quote":"It supplies the linear-stability analysis of electrohydrodynamic flow with and without cross-flow, including the growth rate used for validation."},{"cited_title":"Wu and P","cited_arxiv_id":null,"evidence_quote":"It supplies a unified lattice-Boltzmann model whose hydrostatic profiles and stability results are used as a comparison."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It provides the three-dimensional finite-amplitude electroconvection results that the paper compares with for square and rolling pattern growth rates."},{"cited_title":"Atten and R","cited_arxiv_id":null,"evidence_quote":"It gives the linear result that cross-flow leaves longitudinal rolls unaffected and reduces the transverse roll growth rate."},{"cited_title":"Kato and K","cited_arxiv_id":null,"evidence_quote":"It provides the two-relaxation-time lattice-Boltzmann solver coupled to a fast Poisson solver that the three-dimensional code extends."},{"cited_title":"Zhang, X","cited_arxiv_id":null,"evidence_quote":"It supplies the two-dimensional cross-flow study that introduced the parameter $Y$ and the hysteresis behavior extended here."}],"review_version":1}