Cohesion mediated layering in sheared grains
Khushi Mahajan, Chamkor Singh · 2025
Read Cohesion mediated layering in sheared grains by Khushi Mahajan, Chamkor Singh (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-23573.txt.
Introduction
Khushi Mahajan and Chamkor Singh studied how grains behave when a mix of sticky and loose grains is sheared, in a 2025 paper. Sticky grains tend to clump, but the team found a different pattern under the right conditions.
As the share of sticky grains and their stickiness rose, the sticky grains split into bands or layers that ran through the mix, instead of scattered clumps. The size of the clumps and the stress in the mix followed one curve when plotted against the two factors combined. The authors model the split much like the way oil and water part, with the shearing switching the effect on. Their model matched the steady state layers in tests.
Catalog record
| CREATOR | Khushi Mahajan, Chamkor Singh |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 36 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-23573
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Cohesion mediated layering in sheared grains Khushi Mahajan, Chamkor Singh Abstract We consider pattern formation in a sheared dense mixture of cohesive and non-cohesive grains. Our findings show that cohesive grains, which would typically form distributed agglomerates, instead segregate into percolating stripes or layers when the cohesive grain concentration ( c o c_{o} ) and cohesion strength ( C C ) increase – in a way that the average agglomerate size and the average normal stress collapse onto a single curve when plotted against c o C c_{o}C . Our central proposal is that the development of interfaces between cohesive and non-cohesive grains is akin to phase separation in binary molecular mixtures driven by an effective free energy, although we are dealing with a non-equilibrium system; we setup the segregation flux such that the effect of this free energy is activated only upon application of the external driving. By constructing the segregation flux proportional to the gradient of the variational derivative of the free energy, we closely reproduce the layering in the steady-state limit. We find a robust correspondence between the parameter c o C c_{o}C in the discrete simulations and the parameters in the free energy. Cohesion in granular mixtures is an important aspect for a multitude of natural and industrial settings: agglomeration in powders, mixing of seeds in soil in agriculture, wet soil aggregates in landslides and debris flows, or aggregate formation in snow avalanche deposits. For example, by simply adding moisture as a cohesive agent, one can switch from mixing of grains to segregation of grains, utilizing differential surface wetting properties of particles [ 1 ] , or one can alter the extent of mixing or segregation by changing the volume or viscosity of the added moisture [ 2 ] . Dry binary mixtures usually segregate due to size difference between the two species, due to differential drag or differential substrate friction acting on the two species, or due to difference between the surface or dissipative properties of the two species. This leads to pattern formation, for example, the emergence of streak patterns due to size difference [ 3 ] , band formation due to size difference [ 4 ] or due to differential frictional properties of the particles [ 5 ] , and stripe formation due to differential substrate friction acting on the particles [ 6 ] . Although effect of size or particle property difference is intensively and extensively explored, little is known about pattern formation in a granular system that is binary in the sense that one of its components is cohesive. In this work, we show that cohesion mediation in conjunction with shear, which would normally lead to distributed and irregular shape agglomerates, gives rise to percolating layer or stripe formation at sufficiently increased cohesive grain concentration and cohesion strength. We predict the patterns by constructing an appropriate free energy functional and then deriving the segregation fluxes. This approach accurately describes the steady-state features of the layering patterns. Most importantly, we establish correspondence between the microscopic parameters of the discrete particle simulations and the parameters in the free energy. Figure 1: (a) Mixture of cohesive (green) and non-cohesive (white) grains subjected to shear. The topwall moving with a velocity v_{\mathrm{top}}\mathbf{i} (gray) while the bottomwall remains stationary. The gravity is made to act only on the topwall particles, resulting in a pressure in the bulk. Here, c_{o} is the overall volume fraction or concentration of cohesive grains, C is a cohesive force parameter, and t is the simulation time. (b) Irregular agglomerates at relatively lower c_{o} and higher C , while in (c) we observe ordered layers at higher c_{o} and C . Only cohesive grains are shown in (b, c). (d) Coarse-grained concentration field corresponding to (c). The coarse-graining helps us extract the average interfacial energy for comparison with the continuum model. We subject the mixture of cohesive and non-cohesive grains to plane shear using discrete element method (DEM) simulations and tune: ( i ) the overall volume fraction or concentration of the cohesive grains, c_{o} , and ( ii ) a cohesive force parameter, C . Cohesion originates due to various physical or physicochemical mechanisms: wetting or moisture absorption [ 7 ] ; due to surface treatment of particles with polymeric agents [ 8 ] ; chemical salinization [ 9 ] ; or because of Van der Waals or electrostatic interactions [ 10 , 11 , 12 , 13 ] . It alters the dynamics and rheology in granular flows [ 14 , 15 , 16 , 17 , 18 , 19 , 20 , 21 ] , and often inhibits mixing. Instead of considering specific nature of a cohesive interaction, we use an inter-particle cohesive force that increases linearly as particles approach each other. The parameter C in our simulations represents the net interparticle attractive force in static balance measured in units of particle weight – also known as the granular Bond number. The details of interactions, DEM simulations, and parameters are given in the Supplementary. Figure 1 shows quasi steady-state configurations for (a) relatively low c_{o} , low C ; (b) low c_{o} , high C ; and (c) high c_{o} , high C . By quasi steady-state, we imply that we are away from the initial transients in the system, but dynamic events such as intermittent agglomerate formation and breakup are permitted. The cohesive grains remain homogeneous at low c_{o} and low C [Fig. 1 (a)]. At moderate c_{o} and C , they form irregularly shaped lumps, as expected [Fig. 1 (b)]. However, upon increasing c_{o} and C , we reach a point where the system segregates into percolating stripes or layers – remarkably similar to the classical Cahn-Hilliard phase-separating system under shear [Fig. 1 (c)]. Figure 2: (a) Normalized linear size of clusters, \ell/L , as a function of the overall concentration of cohesive grains c_{o} . (b) The data for \ell/L collapses onto a single curve \ell/L\sim c_{o}C before the agglomerates show percolation ( \ell/L\approx 1 ). (c) The space-time average of non-cohesive pressure is presented as a function of c_{o} , which also collapses onto a single curve when plotted against c_{o}C (d). The error bars represent standard deviations over a quasi-steady-state time window. A characteristic normalized linear size of the agglomerates, defined as {\ell}/{L}\equiv\max\left[\langle{\ell_{x}}/{L_{x}}\rangle_{t},\>\langle{\ell_{y}}/{L_{y}}\rangle_{t},\>\langle{\ell_{z}}/{L_{z}}\rangle_{t}\right] , is shown in Fig. 2 (a) as a function of c_{o} . Here \ell_{k} and L_{k} are the agglomerate span and the flow bed size in the k^{\mathrm{th}} direction, respectively, and \langle\cdot\rangle_{t} implies time average over a quasi steady-state simulation period ( 1500\leq t\leq 2000 ). The data for \ell/L collapses on a single curve when plotted against c_{o}C [Fig. 2 (b)]. We observe a power law scaling close to \ell/L\sim(c_{o}C)^{1} for c_{o}C<2 ; for c_{o}C>2 we start to observe percolating agglomerates ( \ell/L\approx 1 ). To obtain the concentration or the stress field, the microscopic DEM data is coarse-grained onto a mesoscopic grid [Fig 1 (d), Supplementary]. This allows us to obtain the average stress in the system from interparticle force data, and the average interfacial energy between cohesive and non-cohesive phases from the concentration field and its gradients. The data for average normal stress or pressure