<?xml-model href='http://www.tei-c.org/release/xml/tei/custom/schema/relaxng/tei_all.rng' schematypens='http://relaxng.org/ns/structure/1.0'?><TEI xmlns="http://www.tei-c.org/ns/1.0">
	<teiHeader>
		<fileDesc>
			<titleStmt><title level='a'>Structural criterion for the onset of rigidity in a colloidal gel</title></titleStmt>
			<publicationStmt>
				<publisher>APS</publisher>
				<date>09/01/2025</date>
			</publicationStmt>
			<sourceDesc>
				<bibl> 
					<idno type="par_id">10647219</idno>
					<idno type="doi">10.1103/gw42-dks6</idno>
					<title level='j'>Physical Review E</title>
<idno>2470-0045</idno>
<biblScope unit="volume">112</biblScope>
<biblScope unit="issue">3</biblScope>					

					<author>C S Dias</author><author>J C Neves</author><author>M M Telo da Gama</author><author>E Del Gado</author><author>N A M Araújo</author>
				</bibl>
			</sourceDesc>
		</fileDesc>
		<profileDesc>
			<abstract><ab><![CDATA[Not Available]]></ab></abstract>
		</profileDesc>
	</teiHeader>
	<text><body xmlns="http://www.tei-c.org/ns/1.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:xlink="http://www.w3.org/1999/xlink">
<div xmlns="http://www.tei-c.org/ns/1.0"><p>Colloidal gels are nonequilibrium structures sustained by a percolating network of long-living bonds between the particles in solution <ref type="bibr">[1]</ref><ref type="bibr">[2]</ref><ref type="bibr">[3]</ref><ref type="bibr">[4]</ref><ref type="bibr">[5]</ref>. They exhibit unique structural <ref type="bibr">[6]</ref><ref type="bibr">[7]</ref><ref type="bibr">[8]</ref><ref type="bibr">[9]</ref>, dynamical <ref type="bibr">[10]</ref><ref type="bibr">[11]</ref><ref type="bibr">[12]</ref>, and mechanical <ref type="bibr">[13]</ref><ref type="bibr">[14]</ref><ref type="bibr">[15]</ref> properties, which have motivated an intensive search for strategies to design them. This effort is relevant to food or cosmetics, and to energy materials or biotechnology and construction <ref type="bibr">[16]</ref><ref type="bibr">[17]</ref><ref type="bibr">[18]</ref><ref type="bibr">[19]</ref><ref type="bibr">[20]</ref>. Recently, one focus has been on the design of low density gels and control of their local structures <ref type="bibr">[21]</ref><ref type="bibr">[22]</ref><ref type="bibr">[23]</ref><ref type="bibr">[24]</ref><ref type="bibr">[25]</ref><ref type="bibr">[26]</ref><ref type="bibr">[27]</ref><ref type="bibr">[28]</ref><ref type="bibr">[29]</ref><ref type="bibr">[30]</ref><ref type="bibr">[31]</ref><ref type="bibr">[32]</ref><ref type="bibr">[33]</ref> using, e.g., vitrimetric materials or patchy particles, which limit the valence of the particles and the directionality of the bonds <ref type="bibr">[34]</ref><ref type="bibr">[35]</ref><ref type="bibr">[36]</ref><ref type="bibr">[37]</ref>.</p><p>A key property of colloidal gels is their capacity to bear load without deforming permanently. However, the necessary condition for a gel to be rigid is still under debate <ref type="bibr">[38]</ref>. This question is even more critical in the design of low-density gels, where the reduced number of bonds per particle may compromise the mechanical stability. During gelation, a percolating network of bonds emerges at a critical fraction of bonds per particle <ref type="bibr">[1]</ref><ref type="bibr">[2]</ref><ref type="bibr">[3]</ref><ref type="bibr">[4]</ref><ref type="bibr">[5]</ref> but at this connectivity onset, the gels are floppy and fragile <ref type="bibr">[39]</ref><ref type="bibr">[40]</ref><ref type="bibr">[41]</ref>. As more bonds are formed, a rigid elastic network emerges at a rigidity percolation transition <ref type="bibr">[13,</ref><ref type="bibr">42]</ref>. It is difficult however to identify which subset of the visible structure is responsible for the stress transmission <ref type="bibr">[13]</ref>. In 2D the onset of rigidity can be identified directly from the bond network with the pebble game algorithm <ref type="bibr">[43]</ref><ref type="bibr">[44]</ref><ref type="bibr">[45]</ref>, but in 3D the criteria remain elusive. Several local and global geometrical criteria have been proposed, such as a minimum average coordination number <ref type="bibr">[40]</ref> or the existence of a directed percolating path <ref type="bibr">[39]</ref>, but a recently correlated rigidity percolation framework highlights the * Contact author: csdias@fc.ul.pt &#8224; Contact author: fc51402@alunos.fc.ul.pt &#8225; Contact author: mmgama@fc.ul.pt &#167; Contact author: emanuela.del.gado@georgetown.edu &#2225; Contact author: nmaraujo@fc.ul.pt nonlocal nature of the emergence of rigidity, casting doubts on the previously proposed criteria <ref type="bibr">[13]</ref>.</p><p>Experiments combining confocal microscopy with microrheology in colloidal gels resulting from depletion forces suggest that the onset of rigidity coincides with the percolation of particles with at least six neighbors <ref type="bibr">[46]</ref>, in line with a local Maxwell rigidity criterion for isostaticity <ref type="bibr">[47]</ref>. However recent experiments and simulations have clarified that this local criterion does not provide the necessary conditions for gel rigidity <ref type="bibr">[38]</ref>. While attractive depletion forces are isotropic and pairs of bonded particles are free to rotate relative to each other <ref type="bibr">[47]</ref>, rotational degrees of freedom can be hindered by molecules grafted on the particle surfaces <ref type="bibr">[48]</ref><ref type="bibr">[49]</ref><ref type="bibr">[50]</ref> and, quite generally, by more complex surface interactions <ref type="bibr">[12,</ref><ref type="bibr">51,</ref><ref type="bibr">52]</ref>.</p><p>Experimentally, the neighbor counting is based on the number of particles within a certain bonding distance. This definition is problematic for very short interaction ranges or when different types of interactions are present, as two neighboring particles are not necessarily bonded. We propose a different strategy, based on a patchy particle model where one can control attractive bonds from purely repulsive neighbors. Using this feature of the patchy particle model, and varying the number of attractive patches, we can proceed to analyze how the onset of rigidity depends on the number of neighbors, disentangling the contributions of bonded particles from those that only provide steric hindrance. We show that this difference is not a mere detail. For gels of limited valence, rigidity does emerge from the interplay of bonds and repulsive interactions (steric constraints). The onset of rigidity coincides with the emergence of a spanning aggregate of particles with at least three bonds for any value of the valence, including the isotropic continuum limit.</p><p>Colloidal particles with limited valence spontaneously form gels of very low density <ref type="bibr">[53]</ref><ref type="bibr">[54]</ref><ref type="bibr">[55]</ref><ref type="bibr">[56]</ref><ref type="bibr">[57]</ref>, with the focus of theoretical and numerical works being the connectivity percolation <ref type="bibr">[21,</ref><ref type="bibr">[58]</ref><ref type="bibr">[59]</ref><ref type="bibr">[60]</ref><ref type="bibr">[61]</ref>. However, for most practical applications, the gel must be rigid. From the findings reported here, the relevant structural criterion is not the onset of connectivity percolation but develops through the slow aging of the gel <ref type="bibr">[13,</ref><ref type="bibr">38,</ref><ref type="bibr">62,</ref><ref type="bibr">63]</ref>. Specific local geometric motifs, which may depend on the particle interactions, contribute to attaining correlated rigidity percolation, however, it is their spatial organization and the way they are embedded in the structure that drives the onset of rigidity. Our results are consistent with recent findings for nanoparticle gels, providing further insight into the role of topological constraints <ref type="bibr">[38,</ref><ref type="bibr">64]</ref>. The identification of the structural criterion for gel rigidity provides critical insight for designing gels tailored to specific applications.</p><p>Model. We consider a monodisperse suspension of spherical colloidal particles of radius R. The valence of each particle is limited by adding pointlike attractive sites equally spaced on the surface of the particles and consider that bonds are only established through the attractive sites. The number of attractive sites sets the valence. In most of this work the valence is set to three, however, we also study the effect of varying the valence. In all cases, the attractive sites are uniformly distributed on the surface of the colloid. The pairwise interaction is then a superposition of an isotropic repulsion and a site-site attraction (see the Supplemental Material <ref type="bibr">[65]</ref>).</p><p>We first consider particles with three attractive sites on a cubic box of linear size L = 32 in units of the particle diameter and periodic boundary conditions in all directions. To form an elastic gel, number density is set to &#961; = 0.2 and the thermostat temperature is such that k B T /&#1013; = 0.025, where &#1013; is the attractive interaction strength. The dynamics evolve over a total time of 10 3 &#964; B , where &#964; B = 4R 2 /D t is the Brownian time, and D t is the diffusion coefficient. These parameters are chosen such that all particles are part of the gel. However, neither the values for the interaction parameters nor the softness of the potential affect the rheological measures in the regime of low frequency considered here (see the Supplemental Material <ref type="bibr">[65]</ref>).</p><p>Rheological measurements. We minimize the energy by running zero-temperature overdamped simulations until the kinetic energy is, at least ten orders of magnitude lower than its initial value <ref type="bibr">[63]</ref>. Then, we apply Lees-Edwards boundary conditions and an oscillatory affine shear deformation, parametrized by &#947; (t ) = &#947; 0 sin(&#969;t ), with frequency &#969; and amplitude &#947; 0 (see the Supplemental Material <ref type="bibr">[65]</ref>), and identify the linear regime from the load curve. To focus on that regime, we set &#947; 0 = 0.07. The first-order storage G &#8242; 1 and loss G &#8242;&#8242; 1 moduli are obtained from the time evolution of the shear stress &#963; xy (t ) as in Ref. <ref type="bibr">[63]</ref>.</p><p>Figures <ref type="figure">1(a</ref>)-1(c) are snapshots at different times, where the color of a particle corresponds to the number of bonded neighbors. Figures <ref type="figure">1(d</ref></p><p>1 and G &#8242;&#8242; 1 on &#969;, for the three different regimes of gelation, where each structure is relaxed at zero temperature and then subject to the oscillatory measurements. Initially, in (a) and (d), most particles have established either one or no bonds. G &#8242; 1 is lower than G &#8242;&#8242; 1 and it scales linearly with &#969;, as in a Maxwell-like fluid <ref type="bibr">[66]</ref>. As time evolves, more bonds are established and aggregates are formed with a characteristic size of several particles, affecting the mechanical response. For (b) and (e), G &#8242; 1 is still lower than G &#8242;&#8242; 1 but both scale as &#969; &#945; , with &#945; &lt; 1; We estimate &#945; = 0.60 &#177; 0.05 <ref type="bibr">[64]</ref>. Finally, in (c) and (f), a percolating network of particles with three bonds is already formed. Operationally, the gel is rigid when the storage (elastic) modulus dominates over the loss (viscous) modulus at low enough frequencies <ref type="bibr">[67]</ref>.</p><p>The time dependence of G &#8242; 1 and G &#8242;&#8242; 1 for a frequency &#8776;2.5 &#964; -1 B is shown in Fig. <ref type="figure">2</ref>(a) [vertical line in Fig. <ref type="figure">1(f)</ref>]. Both moduli increase with time, and the transition to an elastic gel (G &#8242; 1 G &#8242;&#8242; 1 ) occurs around 10&#964; B (vertical red dashed line). The fraction of particles in the largest connected component is plotted in Fig. <ref type="figure">2(b)</ref> (black curve). The onset of rigidity occurs much later than that for the onset of connectivity percolation (vertical black dashed line, from Ref. <ref type="bibr">[68]</ref>). However, if we compute the size of the largest cluster of fully bonded particles (red circles curve), the data indicate that the onset of percolation of this subset coincides with the onset of rigidity.</p><p>To further test if the onset of rigidity coincides with the percolation threshold of particles with three bonds, we performed the following percolation analysis. We started from gel configurations at the final time of the simulations, when the gel is predominantly elastic and the particles with three bonds percolate. The kinetic energy is minimized as described in the protocol above. Then, at each iteration, one particle is selected and one of its attractive sites is removed. Both the particle and the attractive site are selected at random. The gel is not allowed to age, to keep the same structure. We evaluate the mechanical and percolation properties as a function of p 3 , defined as the fraction of particles that still have three attractive sites. Figure <ref type="figure">3</ref> depicts the dependence on p 3 of the size of the largest connected component, the largest cluster of particles with three bonds and the ratio G &#8242; 1 /G &#8242;&#8242; 1 . The data reveals that the onset of rigidity and the percolation of particles with three bonds coincide. Note that particles and attractive sites are selected uniformly at random, discarding possible effects from spatial correlations developed during the dynamics.</p><p>Effect of the valence. To evaluate the robustness of the three-bonded particle percolation criterium, we performed the same percolation analysis for gels of particles with number of bonds per particle (valence) ns from three to twelve, which is the number of neighbors in compact 3D packings (see Fig. <ref type="figure">S8</ref> in the Supplemental Material <ref type="bibr">[65]</ref>). We select particles at random and remove all of their attractive sites but two. Thus, at each iteration, the particles have either ns or two attractive sites. The onsets of connectivity and three-bonded particle percolation transitions depend strongly on the valence ns. While that for connectivity increases with ns, that for three-bonded particles shifts to lower values of p ns . Thus, for particles with higher valence, the two transitions would be very difficult to distinguish experimentally. As in the system with ns = 3 (Fig. <ref type="figure">3</ref>), the G &#8242; 1 is only larger than G &#8242;&#8242; 1 above the percolation transition for particles with at least three bonds. Below this transition, G &#8242;&#8242; 1 is the largest, over the entire range of frequencies, and thus the gel is rather a viscous fluid. Even at high valence, the percolation of three-bonded particles is still a valid criterion. One may therefore expect it to remain valid in the continuum limit for unconstrained valence.</p><p>Bonds versus steric constraints. A snapshot of the largest connected component at the onset of rigidity is shown in Fig. <ref type="figure">4(a)</ref>, where the set of fully bonded (three bonds) particles is represented by the network of bonds (red). The cluster of three-bonded particles forms a spanning network with very low average connectivity and cycles of at least five particles. We study the effect of the two types of interactions: steric constraints (repulsive contacts) and bonds. In Fig. <ref type="figure">4</ref>(b), we plot G &#8242; 1 and G &#8242;&#8242; 1 for the cluster of three-bonded particles, neglecting the interaction with particles outside that cluster during the rheological measurements. We observe that this structure is not rigid. Even considering all particles, switching off the repulsive steric constraints while keeping the attractive bonds, results in a nonrigid structure [see Fig. <ref type="figure">4(c)</ref>]. Only when including the repulsive steric constraints of particles with three bonds the structure becomes rigid [see Fig. <ref type="figure">4(d)</ref>]. Thus, although the onset of rigidity coincides with the percolation of particles with three bonds, rigidity requires a combination of bonds and repulsive interactions. This combination does not depend on local interactions (see Figs. <ref type="figure">S2</ref> and <ref type="figure">S3</ref> in the Supplemental Material <ref type="bibr">[65]</ref>), but rather on a large-scale steric effect that stabilizes the gel.</p><p>Effect of bond sliding. We now consider the effect of bond sliding, assuming a bending energy described by a quadratic potential V &#952; = k A (&#952;&#952; 0 ) 2 , where &#952; is the angle between the bonds at the center of the particle. The equilibrium angle &#952; 0 is 2&#960;/3, and k A is the strength of the bending energy (results in Figs. S9 and S10 in the Supplemental Material <ref type="bibr">[65]</ref>). In the absence of bending energy (k A = 0), the bonds are free to move around the particle, and the gel is never rigid. Otherwise, the gel is rigid for all k A &#824; = 0 and the transition frequency vanishes for k A = 0, that is, in the absence of bending energy. This is consistent with what was found for fiber networks <ref type="bibr">[69]</ref>. A nonmonotonic dependence of the frequency transition as a function of k A was found, indicating a competition between the rigidity promoted by the bond resistance to sliding and the local rearrangement to more stable (rigid) positions when sliding is allowed.</p><p>Conclusion. It is the combination of network bonds and steric constraints that leads to the rigidity of the gel. The development of rigid structures depends strongly on the gelation dynamics, not just on the number of bonds of a particle, as proposed by correlated rigidity percolation <ref type="bibr">[13]</ref>. For all the gels considered here, where the particle valence ranged from three to twelve, we found that the onset of gel rigidity corresponds to the percolation of particles with at least three bonds, suggesting that for the type of interactions considered, this identifies the necessary condition for rigidity percolation. For nanoparticles with ad hoc limited valence, our results show that a sufficient fraction of particles with valence three is required to design rigid structures. More generally, when the valence cannot be controlled precisely, our findings shed light on how the mechanics of colloidal gels result from the interplay of different substructures intertwined with each other, not easily recognizable from local geometrical motifs <ref type="bibr">[70]</ref>. The implications of our discoveries extend far beyond the study of colloidal gels. From biomaterials to advanced manufacturing and beyond, the ability to predict and control gel rigidity opens avenues for the design of customizable materials with targeted mechanical properties. Consider biomedical applications where gels must provide structural support within the body <ref type="bibr">[20,</ref><ref type="bibr">71]</ref> or act as drug delivery systems <ref type="bibr">[72]</ref>, which, by using low-density gels of limited-valence particles, one could understand and manipulate rigidity to revolutionize treatment strategies.</p></div></body>
		</text>
</TEI>
