betaFeedback

WorksPhysics and Astronomy

Kinetics of Self-Organization in One Component Active Matter Systems

Abstract

We study an inherently nonequilibrium system consisting of self-propelled particles using molecular dynamics simulation in d=2d=2. The system self-organizes through quorum-sensing interactions characterized by an interaction range λλ and a parameter cthc_{\mathrm{th}}. When the local concentration sensed by a particle, represented by an auxiliary variable ci(t)c_i(t) exceeds cthc_{\mathrm{th}}, its self-propulsion velocity abruptly drops to zero, leading to the emergence of particle rich domains. For small λλ, we observe a droplet-like morphology of clusters, which crosses over to a bicontinuous structure as λλ increases. We confirm this transition by computing the eccentricity of the domains. Next, we characterize the evolving domain morphologies using equal-time correlation function C(r,t)C(r,t) and structure factor S(k,t)S(k,t) of the density field. For a fixed λλ, the scaled C(r,t)C(r,t)'s at different times are indistinguishable from each other, indicating dynamical scaling. However, at a fixed time tt, the scaling functions differ for different values of λλ, confirming a λλ-dependent morphological transition. In the large-wavevector limit (k→∞k\rightarrow\infty), the tail of S(k,t)S(k,t) consistently follows the Porod's law: S(k,t)∼k−3S(k,t)\sim k^{-3}. Furthermore, the average domain size L(t)L(t) exhibits a power-law growth, L(t)∼t1/3L(t)\sim t^{1/3}, independent of λλ.

Introduction

Self-organization in active matter systems has emerged as a rapidly growing field of research over the last two decades [1, 2, 3, 4, 5, 6]. The constituents of an active system are self-propelled agents that can consume energy from their surroundings or convert internal energy into directed motion [5, 7, 8, 9]. Active matter encompasses many examples in nature, ranging from biological tissues, bird flocks [10, 11], fish shoals [12, 13], animal herds [14], and pedestrian crowds [15], to synthetic systems such as microrobots [16], Janus particles [17], and vibrated granular matter [18]. The typical size of these constituents varies from a few millimeters to several centimeters. All of these systems are intrinsically out of equilibrium, which leads to several emergent phenomena, including nonequilibrium pattern formation, giant number fluctuations [19], and spontaneous flows generated by internal stresses [20]. Such phenomena do not exist in passive systems.

In this paper, we focus on the kinetics of self-organization in active Brownian particle (ABP) systems. The existing literature primarily attributes such phenomena to motility-induced phase separation (MIPS) and non-reciprocal interactions [21, 22, 23, 24, 25, 26]. Our study primarily focuses on self-organization driven by MIPS. In the classical MIPS, an active particle transiently reduces its speed due to interactions with its high-density surroundings, leading to an effective accumulation of particles in denser regions. This accumulation enhances the density fluctuations and further reduces the particle speed, thereby establishing a positive feedback loop that drives aggregation of particles [27, 28, 29].

Starting from the seminal particle-based simulations of Vicsek et al. [30], many theoretical [31, 32] and experimental efforts have been made over the past few decades to understand MIPS in ABP systems [33]. Theoretically, various particle-based and coarse-grained models have been proposed [34]. In the context of particle-based simulations, Redner et al. studied MIPS in a system of self-propelled Brownian spheres that interact purely through exclude volume, with no alignment interaction [35]. The system phase separates into dense and dilute phases whose densities depend solely on the Péclet number of the spheres. Agudo-Canalejo and Golestanian studied MIPS in mixtures of chemically interacting particles [36]. Chemical reactions among the species break the action–reaction symmetry of the interactions, leading to separation into phases with distinct densities and stoichiometry. Stenhammar et al. studied the phase-separation kinetics of MIPS in d=3d = 3 and showed that the average domain size, L(t)L(t), follows the classical Lifshitz–Slyozov (LS) growth law, L(t)∼t1/3L(t) \sim t^{1/3} [37]. In addition, they demonstrated that the structure factor obeys Porod’s law in the large wavenumber regime.

The first coarse-grained model for MIPS was proposed by Tailleurr and Cates [34]. They derived an evolution equation for the density field by coarse-graining the single-particle dynamics. They showed that the system undergoes phase separation into dilute and dense phases when the self-propulsion speed vv decreases with the local particle density. Later, Curatolo et al. extended the study for nn-species systems [38]. Recently, Mondal and Das studied the domain kinetics for the model with n=2n = 2. They showed that L(t)L(t) follows the LS law, independent of the relative activities of the species or their compositions [39]. At the phenomenological level, Wittkowski et al. introduced the active model B (AMB) by adding a gradient-squared term of the order parameter, i.e., α∣∇⃗ϕ∣2\alpha|\vec{\nabla}\phi|^2, to the chemical potential of passive model B [40]. This additional term violates the detailed balance condition, a defining signature of nonequilibrium active matter systems. By tuning α\alpha, they controlled the activity of the system. They observed that L(t)L(t) exhibits power-law growth, L(t)∼tϕL(t) \sim t^{\phi}, with an exponent ϕ\phi that depends on α\alpha. Nevertheless, they predicted the L(t)∼t1/3L(t) \sim t^{1/3} growth law irrespective of the value of α\alpha. Recently, Pattanayak et al. [41, 42] provided a detailed study of domain growth kinetics for symmetric and asymmetric active binary mixtures using AMB. They reported that the domain growth exponent exhibits a crossover from 1/31/3 at early times to 1/41/4 at later times. Further extension of this study using AMB+, an extension of AMB, shows that the asymptotic domain growth law is L(t)∼t1/4L(t) \sim t^{1/4} [43].

There exists several experimental studies on MIPS. One of the earliest experimental studies is due to Dombrowski et al., who showed that bacterial systems can form large-scale coherent structures in confined geometries [44]. Subsequent experimental studies on synthetic active particles, such as Janus particles, demonstrated various tunable self-assembled structures by controlling the environment and confinement. Linden et al. showed that switching on high activity in a relatively dense system of active Janus colloids leads to rapid clustering, followed by cluster aggregation toward full phase separation [45]. Theurkauff et al. studied dynamic clustering in active colloidal suspensions with chemical signaling and showed that, in addition to solid-like and gas-like phases observed at high and low densities, respectively, a novel cluster phase emerges [46]. Experimental observations of the L(t)∼t1/3L(t) \sim t^{1/3} growth law in MIPS are limited. Aforementioned studies exhibit dynamical clustering or arrested coarsening rather than the unrestricted diffusive coarsening observed in model systems, owing to hydrodynamic effects, alignment interactions, boundaries, or finite-size effects.

Our study here is motivated by the work of Bäuerle et al., who experimentally and numerically studied the self-organization of light-activated active particles [47]. In their model, particles interact via a quorum-sensing interaction of range λ\lambda and self-organize for a certain range of λ\lambda values. However, a systematic study of the self-organization kinetics for large system sizes is still lacking. In this work, we address this issue by investigating the coarsening dynamics over a wide range of λ\lambda and other model parameters. Our results show that L(t)L(t) follows the LS law, irrespective of λ\lambda. In contrast, the domain morphology depends strongly on λ\lambda: domains are nearly circular for small λ\lambda, predominantly elliptical at intermediate values of λ\lambda, and elongated for large λ\lambda. We quantify the transition by computing eccentricity of the domains. Further, we have characterized the domain morphologies by calculating equal-time correlation functions and structure factors of the density fields.

Given this background, the paper is organized as follows. We describe the model and provide the details of simulations in § 2. The simulation results are discussed in § 3. Finally, we summarize our findings in § 4.

Details of Modeling and Simulation

We consider a system of active particles in two dimensions. Particle ii, with position vector r⃗i\vec{r}_i, interacts with particle jj, with position vector r⃗j\vec{r}_j, via a purely repulsive Weeks–Chandler–Andersen (WCA) interaction potential UWCAU_{\mathrm{WCA}}, given by

UWCA(rij)=4ϵ[(σrij)12−(σrij)6+14].(1)U_{\mathrm{WCA}}(r_{ij}) = 4\epsilon\left[\left(\frac{\sigma}{r_{ij}}\right)^{12} - \left(\frac{\sigma}{r_{ij}}\right)^6 + \frac{1}{4}\right]. \tag*{(1)}

Here, rij=∣r⃗i−r⃗j∣r_{ij} = \lvert\vec{r}_i-\vec{r}_j \rvert denotes the distance between the particle pair (i,j)(i,j), σ\sigma is the effective particle diameter, and ϵ\epsilon is the interaction strength. The cutoff distance for UWCAU_{\mathrm{WCA}} is rcut=21/6σr_{\mathrm{cut}} = 2^{1/6}\sigma. In our simulation, we set ϵ=100kBT\epsilon= 100 k_{\mathrm{B}}T to minimize overlap between pairs of particles and ensure numerical stability, where kBk_{\mathrm{B}} is the Boltzmann constant and TT denotes the room temperature. For modeling the quorum-sensing interaction, we follow the approach proposed by Bäuerle [47]. We introduce a time-dependent auxiliary variable ci(t)c_i(t) for the iith particle, defined as

ci(t)=c~∑j≠iσrij(t)exp⁡(−rij(t)λ),(2)c_i(t) = \tilde{c}\sum_{j\ne i}\frac{\sigma}{r_{ij}(t)}\exp\left(-\frac{r_{ij}(t)}{\lambda}\right), \tag*{(2)}

where c~\tilde{c} is a constant. The parameter λ\lambda has the dimension of length and determines the range of the quorum-sensing interaction. The propulsion speed viv_i depends on ci(t)c_i(t): it is v0v_0 when ci(t)c_i(t) is less than some threshold value cthc_{\mathrm{th}}, and it becomes zero when ci(t)>cthc_i(t)>c_{\mathrm{th}}. Since ci(t)c_i(t) encodes information about the local neighborhood of the iith particle, the resulting change in particle motility effectively mimics a density-dependent propulsion mechanism, giving rise to MIPS. Furthermore, we assume that the propulsion orientation θi\theta_i (with xx-axis), remains unaffected by the modulation of motility and it shows purely diffusive dynamics.

In the overdamped limit, the equations of motion for the particles are given by

dr⃗idt=vi(ci(t))n^i−D0kBT∇⃗iUWCA+Λ⃗i(t),(3)\frac{d\vec{r}_i}{dt}=v_i\left(c_i(t)\right)\hat{n}_i-\frac{D_0}{k_{\mathrm{B}}T}\vec{\nabla}_iU_{\mathrm{WCA}}+\vec{\Lambda}_i(t), \tag*{(3)}
dθidt=ηi(t).(4)\frac{d\theta_i}{dt}=\eta_i(t). \tag*{(4)}

Here, n^i=(cos⁡θi,sin⁡θi)\hat{n}_i=(\cos\theta_i,\sin\theta_i) denotes the unit vector corresponding to the orientation angle θi\theta_i of the iith particle. The stochastic term Λ⃗i(t)\vec{\Lambda}_i(t) represents Gaussian white noise with zero mean and covariance

⟨Λiα(t1)Λjβ(t2)⟩=2D0δijδαβδ(t1−t2),\langle\Lambda_{i\alpha}(t_1)\Lambda_{j\beta}(t_2)\rangle=2D_0\delta_{ij}\delta_{\alpha\beta}\delta(t_1-t_2),

where D0D_0 represents the translational diffusion coefficients, α\alpha and β\beta label Cartesian components, and ii and jj denote particle indices. Similar to Λ⃗i(t)\vec{\Lambda}_i(t), ηi(t)\eta_i(t) also has zero mean and satisfies

⟨ηi(t1)ηj(t2)⟩=2DRδijδ(t1−t2),\langle\eta_i(t_1)\eta_j(t_2)\rangle=2D_R\delta_{ij}\delta(t_1-t_2),

with DRD_R being the rotational diffusion coefficients [48, 49].

We simulate systems of size Lx×Ly=512σ×512σL_x\times L_y=512\sigma\times512\sigma. The particle number density is set to ρ=0.2\rho=0.2, and we choose σ=1\sigma=1. These parameters fix the total number of particles to N=ρ×512σ×512σ≈52428N=\rho\times512\sigma\times512\sigma\approx52428. The system is initialized by randomly assigning particle positions r⃗i\vec{r}_i within the simulation box, such that the separation between any pair of particles is greater than 2σ2\sigma. The variables θi\theta_i are initialized by random numbers sampling from a uniform distribution over the interval [−π,π][-\pi,\pi]. Equations (3) and (4) are discretized in time using the Euler discretization method to update particle positions and orientations. A time step of size Δt=0.00005\Delta t=0.00005 is used. To improve computational efficiency, a cutoff distance Rcut=3λR_{\mathrm{cut}}=3\lambda is introduced for the calculation of ci(t)c_i(t), and a neighbor list is implemented and updated at regular intervals. At each time step, we compute ci(t)c_i(t) for all particles using Eq. (2) and check whether ci(t)c_i(t) exceeds the threshold cthc_{\mathrm{th}}. If ci(t)>cthc_i(t)>c_{\mathrm{th}}, we set the propulsion speed of that particle to vi=0v_i=0; otherwise, the particle continues to move with vi=v0v_i=v_0 during that time step, while its propulsion orientation θi\theta_i continues to undergo rotational diffusion. Periodic boundary conditions are employed in all spatial directions. We set c~=1\tilde{c}=1, v0=50v_0=50, D0=1D_0=1, and DR=3D_R=3. Unless otherwise stated, we use cth=5.0c_{\mathrm{th}}=5.0. We treat cthc_{\mathrm{th}} and λ\lambda as independent simulation parameters. All the statistical results presented below are averaged over ten independent runs.

Simulation Results

Figure 1(a)–(Figure 1d) show snapshots of the evolution of particle positions at different times for λ=3\lambda= 3, as indicated. Clearly, the system evolves through the formation of droplet-like dense clusters or particle domains. At early times [cf. Figure 1(a)], the number of clusters is large, while their sizes are small. As time advances [cf. Figure 1(b)–(Figure 1c)], droplets above a certain size grow, while smaller ones dissolve, as observed in the Ostwald ripening process [50]. At later times, clusters with large number of particles emerge, as shown in Figure 1(d). Typically, particles that lie well inside a cluster have zero self-propulsion speed, whereas particles near domain boundaries or in dilute regions have a v0v_0 self-propulsion speed. Also, it appears that the center of mass of the larger clusters that grow over time remains essentially stationary [cf. Figure 1(c)–(Figure 1d)].

Evolution snapshots of the system for $\lambda = 3$ at different times, with scaling plots of the equal-time correlation function and structure factor

Figure 1. Evolution snapshots of the system for λ=3\lambda= 3 at different times: (a) t=50t = 50, (b) t=100t = 100, (c) t=300t = 300, and (d) t=600t = 600, as mentioned. Panel (e) shows the corresponding scaling plot of the equal-time correlation function C(r,t)C(r,t) as a function of r/L(t)r/L(t), where L(t)L(t) is the time-dependent average domain size. Panel (f) shows the scaling plot of the structure factor S(k,t)L(t)−2S(k,t)L(t)^{-2} vs. kL(t)kL(t) on a log-log scale. The dashed line labeled k−3k^{-3} represents Porod’s law, S(k,t)∼k−(d+1)S(k,t)\sim k^{-(d+1)}, for d=2d = 2.

Next, we focus on characterizing the morphology of domains by calculating the equal-time correlation function C(r,t)C(r,t) and structure factor S(k,t)S(k,t) [51, 52]. For the order-parameter ϕ(r⃗,t)\phi(\vec{r},t), C(r,t)C(r,t) is defined as

C(r,t)=⟨ϕ(r⃗1,t)ϕ(r⃗2,t)⟩−⟨ϕ(r⃗1,t)⟩⟨ϕ(r⃗2,t)⟩.(5)C(r,t)=\langle\phi(\vec{r}_1,t)\phi(\vec{r}_2,t)\rangle-\langle\phi(\vec{r}_1,t)\rangle\langle\phi(\vec{r}_2,t)\rangle. \tag*{(5)}

Here, rr is the distance between two points r⃗1\vec{r}_1 and r⃗2\vec{r}_2, and the angular brackets represent spherical averaging over different directions and random initial conditions. S(k,t)S(k,t) is the Fourier transform of C(r,t)C(r,t) and is given by

S(k⃗,t)=∫dr⃗ eik⃗⋅r⃗C(r⃗,t).(6)S(\vec{k},t)=\int d\vec{r}\,e^{i\vec{k}\cdot\vec{r}}C(\vec{r},t). \tag*{(6)}

To define ϕ(r⃗,t)\phi(\vec{r},t), we map the continuum fluid configurations onto a square lattice. The system is divided into boxes of size 4σ×4σ4\sigma\times4\sigma. If the number density of particles within a box exceeds the average density (here ρ=0.2\rho=0.2), we assign a value of ϕ(r⃗,t)=+1\phi(\vec{r},t)=+1; otherwise, we assign ϕ(r⃗,t)=−1\phi(\vec{r},t)=-1. This mapping renders the subsequent analysis analogous to studies of phase-separation kinetics in solid binary mixtures using Monte Carlo simulations of the spin-1/2 Ising model [53, 54].

Figure 1(e) shows plots of C(r,t)C(r,t) at four different times, corresponding to the evolution snapshots shown in Figure 1(a)–(Figure 1d). By definition, C(r=0,t)=1C(r=0,t)=1, reflecting perfect self-correlation. As rr increases, C(r,t)C(r,t) initially decays with distance, as the probability of finding opposite order parameters increases. At larger rr, C(r,t)C(r,t) becomes negative, reaching a minimum as regions with opposite order parameters dominate. Upon a further increase in rr, C(r,t)C(r,t) rises again owing to the mixed presence of both order-parameter states. Clearly, C(r,t)C(r,t) at different times collapse onto a single master curve when the distance rr is scaled by the average domain size L(t)L(t), indicating dynamical scaling. Here, the characteristic domain size, L(t)L(t), is defined as the distance at which the C(r,t)C(r,t) decays to a certain fraction of its maximum value, C(0,t)=1C(0,t)=1 [52, 55]. Figure 1(f) shows the scaled structure factor, S(k,t)L(t)−2S(k,t)L(t)^{-2}, plotted as a function of kL(t)kL(t) at different times, as indicated. In the large-kk limit, all master curves approximately exhibit a k−3k^{-3} decay, consistent with Porod’s law: S(k,t)∼k−(d+1)S(k,t)\sim k^{-(d+1)} for d=2d=2 [56, 57]. Furthermore, the tails of the scaled structure factor decrease with time (i.e., S(k,t)L(t)−2S(k,t)L(t)^{-2} is larger for the black curve than for the blue curve), reflecting the growth of the average domain size.

Figures 2(a)–(d) shows snapshots of the system at different times for λ=5\lambda=5, as indicated. Similar to the λ=3\lambda=3 case, the system evolves by forming dense clusters of particles that grow with time. However, in sharp contrast to the λ=3\lambda=3 case, we observe interconnected domain structures. We next characterize the domain morphology by computing the correlation function C(r,t)C(r,t) and the structure factor S(k,t)S(k,t). Similar to the λ=3\lambda=3 case, the numerical data for both C(r,t)C(r,t) and S(k,t)S(k,t) exhibit dynamical scaling, as shown in Figure 2(e) and Figure 2(f), respectively. Further, in the large-kk tail of S(k,t)S(k,t) follows Porod’s law.

Time evolution of the domain morphology and scaling plots

Figure 2. Time evolution of the domain morphology at fixed λ=5\lambda= 5 is illustrated by snapshots taken at different times: (a) t=50t = 50, (b) t=100t = 100, (c) t=300t = 300, and (d) t=600t = 600. Panel (e) presents the corresponding plots of the scaled correlation function C(r,t)C(r,t) versus r/L(t)r/L(t), while panel (f) shows the scaling plot of S(k,t)L(t)−2S(k,t)L(t)^{-2} as a function of kL(t)kL(t) on a log–log scale. The dashed line labeled k−3k^{-3} represents Porod’s law.

Next, we examine how the domain morphologies depend on λ\lambda. We compare snapshots of the system at a fixed time t=300t=300 for different values of λ\lambda, as shown in Figure 3. For smaller λ=3\lambda=3, the system exhibits a circular droplet-like morphology, as shown in Figure 3(a). At intermediate values, for example λ=4\lambda=4, the morphology consists of both circular droplets and elongated domains, as shown in Figure 3(b). As λ\lambda is increased further to λ=5\lambda=5, the system develops an elongated and bicontinuous morphologies, as shown in Figure 3(c). These results indicate that, for a fixed particle density and other model parameters, the domain morphology depends strongly on λ\lambda. This can be explained as follows. For small λ\lambda, the number of particles satisfying ci(t)>cthc_i(t) > c_{\mathrm{th}} during the early stages of evolution is relatively small. Consequently, only a few particles have their self-propulsion velocity reduced to zero, and they are well separated from one another. These particles act as nucleation sites for domain formation, consequently the resulting domains are predominantly circular. As λ\lambda increases, the number of nucleation sites also increases, leading to a reduction in the average separation between them. This enhances the likelihood of coalescence between neighboring circular domains, giving rise to aspherical domain structures. Upon further increasing λ\lambda, the continued increase in the number of nucleation sites and the corresponding decrease in their separation promote the formation of highly elongated domains that can span a significant fraction of the system.

Evolution snapshots of the system at fixed $t = 300$ with different values of $\lambda$: (a) $\lambda = 3$, (b) $\lambda = 4$, and (c) $\lambda = 5$.

Figure 3. Evolution snapshots of the system at fixed t=300t = 300 with different values of λ\lambda: (a) λ=3\lambda= 3, (b) λ=4\lambda= 4, and (c) λ=5\lambda= 5, as mentioned.

To characterize the aspherical shape of the domains during evolution, we compute the eccentricity ee of the domains, given by

e=1−b2a2.(7)e = \sqrt{1 - \frac{b^2}{a^2}}. \tag*{(7)}

Here, aa and bb denote the major and minor axes, respectively, of an aspherical shaped domain, as shown in Figure 4. These quantities are obtained from the gyration tensor analysis. For a domain containing, NdN_d particles, we first determine the domain centre of mass and construct the covariance (gyration) matrix from the particle position fluctuations relative to the centre of mass, as described in Ref. [58]. The eigenvalues of the matrix, Γ1\Gamma_1 and Γ2\Gamma_2 (Γ1≥Γ2\Gamma_1 \ge\Gamma_2), correspond to the principal moments of the domain. The lengths of the major and minor axes are then given by a=Γ1a = \sqrt{\Gamma_1} and b=Γ2b = \sqrt{\Gamma_2}, respectively. Figure 4 shows the dependence of ee on λ\lambda, keeping other parameters fixed. Clearly, ee increases with λ\lambda, indicating the increasing of aspherical shape of the domains.

Plot of average eccentricity of the clusters as a function of $\lambda$

Figure 4. Plot of average eccentricity ee of the clusters as a function of λ\lambda at time t=300t = 300, using Eq. (7). The data are obtained by averaging the eccentricities of all clusters in the system and subsequently averaging over 10 independent realizations with different initial conditions.

To further characterize the morphological transition observed in Figure 3, we compute C(r,t)C(r,t) for different values of λ\lambda. Figure 5a shows the scaled C(r,t)C(r,t)s at a fixed time t=300t = 300 for the values of λ\lambda, as indicated. It is evident that the scaled master curves do not collapse onto a single universal function at large distances. This lack of data collapse indicates clear morphological differences in the domain structures as λ\lambda is varied. The data mismatch at large rr translates to the small-kk regime of S(k,t)S(k,t), while the tail of S(k,t)S(k,t) follows Porod’s law, regardless of λ\lambda (not shown here).

Next, we focus on how the average domain size L(t)L(t) varies with time. Figure 5b shows L(t)L(t) as a function of tt on a log–log scale for different values of λ\lambda, as indicated. After an initial transient regime, L(t)L(t) grows as L(t)∼t1/3L(t) \sim t^{1/3}, independent of λ\lambda. In the context of phase separation in passive systems, this growth law is well known and has been observed during domain growth in off-critical binary mixtures, as well as during spinodal decomposition in critical mixtures [50]. Recently, Mondal and Das reported the L(t)∼t1/3L(t) \sim t^{1/3} growth law in the context of MIPS in active binary mixtures, irrespective of the relative activities of the species or their compositions [39].

Further, we compute the density difference Δρ=ρ1−ρ2\Delta\rho= \rho_1 - \rho_2 as a function of λ\lambda at a sufficiently late time. Here, ρ1\rho_1 and ρ2\rho_2 denote the particle densities in the dense and dilute regimes, respectively. We calculate ρi\rho_i as ρi=nπsc2\rho_i = \frac{n}{\pi s_c^2}, where nn is the number of particles within a circle of radius scs_c chosen in the respective dense and dilute regimes. In simulation, we set sc=10σs_c = 10\sigma. The variation of Δρ\Delta\rho with λ\lambda enables us to determine the range of λ\lambda values for which the system exhibits self-organization at a given cthc_{\mathrm{th}}. A value of Δρ≈0\Delta\rho\approx0 indicates that the system remains

Scaling plots of the equal-time correlation functions

(a)

Plot of average domain size

(b)

Figure 5. (a) Scaling plots of the equal-time correlation functions C(r,t)C(r,t) at t=300t = 300 for different values of λ\lambda, as mentioned. (b) Plot of average domain size L(t)L(t) as a function of tt on a log-log scale for three different values of λ\lambda, as mentioned. The black solid line labeled 1/31/3 represents the asymptotic L(t)∼t1/3L(t) \sim t^{1/3} growth law.

homogeneous, i.e., no phase separation occurs. Figure 6 shows the variation of Δρ\Delta\rho with λ\lambda for different cthc_{\mathrm{th}}. Starting from λ>rcut\lambda> r_{\mathrm{cut}}, the system begins to exhibit phase separation. Clearly, Δρ\Delta\rho decreases with increasing λ\lambda for fixed cthc_{\mathrm{th}}. For sufficiently large λ\lambda (≫rcut\gg r_{\mathrm{cut}}), the quorum-sensing interaction sets the speed of all particles in the system to zero, leading to a frozen state for cthc_{\mathrm{th}}. Consequently, phase separation is no longer observed. The value of λ\lambda at which the system freezes depends on cthc_{\mathrm{th}}. For very small-λ\lambda (<rcut< r_{\mathrm{cut}}), the local concentrations ci(t)c_i(t) are unable to reach the quorum-sensing threshold cthc_{\mathrm{th}}, resulting in a state where all particles move with the propulsion speed v0v_0 in random directions, again preventing phase separation. Thus, the system exhibits phase separation only within a finite range of λ\lambda.

Dependence of the average density difference, $\Delta \rho$, between the dense and dilute regimes as a function of $\lambda$

Figure 6. Dependence of the average density difference, Δρ\Delta\rho, between the dense and dilute regimes of the system at a sufficiently late time (t=1000t = 1000) as a function of λ\lambda. The density of particles in the system is ρ=0.2\rho= 0.2.

Finally, to characterize the dependence of Δρ\Delta\rho on λ\lambda in the self-organized phase, we compute the radial distribution function, g(r)g(r), as a function of rr, where rr is the radial distance. It describes the distribution of particles around a reference particle, and is defined as: g(r)=ρ(r)/ρ~g(r) = \rho(r)/\tilde{\rho}. Here, ρ(r)\rho(r) represents the local density profile of particles at a distance rr around the reference particle, and ρ~\tilde{\rho} is the average density of particles [59, 60]. Figure 7(a) shows the plot of g(r)g(r) for the evolution snapshots shown in Figure 2. At r≈σr \approx\sigma, all curves exhibit nearly the same peak height, suggesting that the domains at different times have a similar degree of packing. Furthermore, the decay of the curves slow down as the system evolves from the early time t=50t = 50 (black curve) to the later time t=600t = 600 (blue curve). This shift indicates an increase in the characteristic domain size as the domains grow over time. In Figure 7(b), we plot g(r)g(r) versus rr corresponding to the snapshots shown in Figure 3. At r≈σr \approx\sigma, the black curve for λ=3\lambda= 3 exhibits the highest peak in g(r)g(r), whereas the green curve for λ=5\lambda= 5 shows the lowest peak. The red curve corresponding to λ=4\lambda= 4 exhibits an intermediate peak height. The difference in the peak heights of g(r)g(r) suggests that the degree of domain packing decreases with increasing

Radial distribution function plots

Figure 7. Radial distribution function, g(r)g(r), corresponding to (a) the evolution snapshots shown in Figure 2 at fixed λ=5\lambda= 5 for different times, and (b) the snapshots shown in Figure 3 at fixed t=300t = 300 for λ=3\lambda= 3, 4, and 5.

λ\lambda from 3 to 5, indicating that Δρ\Delta\rho decreases in the self-organized phase as λ\lambda increases. In all cases, g(r→∞)=1g(r \to\infty) = 1, indicating a homogeneous density.

Conclusions

Let us conclude this paper with a summary and discussion of our results. We have investigated the kinetics of self-organization in a one-component active matter system interacting via quorum-sensing interactions with λ\lambda as the characteristic interaction length. In addition to quorum sensing, particles interact through excluded-volume interactions modeled by the WCA potential, and the direction of self-propulsion undergoes rotational diffusion. In this study, all model parameters are held fixed while only λ\lambda and cthc_{\mathrm{th}} are varied. We find that the system undergoes self-organization only within a finite range of λ\lambda for a fixed cthc_{\mathrm{th}}. The system does not self-organize via quorum sensing when λ<rcut\lambda< r_{\mathrm{cut}} or λ≫rcut\lambda\gg r_{\mathrm{cut}}. Starting from random initial conditions, the system phase-separates into dense and dilute regions for appropriate values of λ\lambda.

At fixed cthc_{\mathrm{th}}, we observe a transition from circular to elongated domain structures as λ\lambda increases. We quantify the asphericity of the domains using the eccentricity parameter, ee. Small values of ee are associated with circular domains at low λ\lambda, while larger values of ee correspond to elongated or aspherical domain structures at higher λ\lambda. Intermediate values of ee characterize domain structures that lie between these two extremes, reflecting a gradual transition from droplet-like to elongated structures. Additionally, the compactness of the domains decreases with increasing λ\lambda, as confirmed by the radial distribution function g(r)g(r). Next, we characterize the domain morphology by computing the equal-time density correlation function C(r,t)C(r,t) and the corresponding structure factor S(k,t)S(k,t), both of which exhibit dynamical scaling. However, the form of scaling functions depend on λ\lambda, indicating the occurrence of a morphological transition. sition as λ\lambda is varied. Furthermore, we calculate the average domain size L(t)L(t) from the decay of C(r,t)C(r,t). We find that L(t)L(t) follows a power-law growth, L(t)∼t1/3L(t) \sim t^{1/3}, independent of λ\lambda. This observation suggests that the t1/3t^{1/3} domain-growth law is robust with respect to the choice of physical model describing MIPS.

Acknowledgments: A. Chauhan is grateful to IISER-Mohali, India for the financial support through an institute postdoctoral fellowship.

Conflict of interest: The authors have no conflicts to disclose.

References

  1. [1]A. Czirók, E. Ben-Jacob, I. Cohen, and T. Vicsek, Formation of complex bacterial colonies via self-generated vortices, Phys. Rev. E 54, 1791 (1996).pubmed.ncbi.nlm.nih.gov/9965259
  2. [2]Z. Neufeld, M. Vicsek, and T. Vicsek, Complex spatiotemporal patterns in two lattice models with instability, Physica A: Statistical Mechanics and its Applications 233, 754-766 (1996).DOI
  3. [3]J. Toner, Y. Tu, and S. Ramaswamy, Hydrodynamics and phases of flocks, Annals of Physics 318, 170-244 (2005).DOI
  4. [4]F. Jülicher, K. Kruse, J. Prost, and J. F. Joanny, Active behavior of the cytoskeleton, Physics Reports 449, 3-28 (2007).DOI
  5. [5]S. Ramaswamy, The Mechanics and Statistics of Active Matter, Annual Review of Condensed Matter Physics 1, 323-345 (2010).
  6. [6]N. J. Lauersdorf, E. Nazockdast, and D. Klotsa, Binary mixtures of active Brownian particles with distinct nonzero activities, Soft Matter 21, 6132-6143 (2025).DOI
  7. [7]M. C. Marchetti, J. F. Joanny, S. Ramaswamy, T. B. Liverpool, J. Prost, M. Rao, and R. A. Simha, Hydrodynamics of soft active matter, Rev. Mod. Phys. 85, 1143 (2013).DOI
  8. [8]J. O’Byrne, Y. Kafri, J. Tailleur, and F. van Wijland, Time irreversibility in active matter from micro to macro, Nature Reviews Physics 4, 167-183 (2022).DOI
  9. [9]S. Mondal and P. Das, Velocity distribution and diffusion of an athermal inertial run-and-tumble particle in a shear-thinning medium, Phys. Fluids 37, 057114 (2025).DOI
  10. [10]D. Chen, Y. Wang, G. Wu, M. Kang, Y. Sun, and W. Yu, Inferring causal relationship in coordinated flight of pigeon flocks featured, Chaos 29, 113118 (2019).DOI
  11. [11]A. Cavagna and I. Giardina, Bird flocks as condensed matter, Annu. Rev. Condens. Matter Phys. 5, 183-207 (2014).DOI
  12. [12]M. Jiang, A. Zhou, R. Chen, Y. Yang, H. Dong and W. Wang, Collective motions of fish originate from balanced local perceptual interactions and individual stochastics, Phys. Rev. E 107, 02441 (2023).DOI
  13. [13]J. T. Bonner, A way of following individual cells in the migrating slugs of Dictyostelium discoideum, Proc. Natl. Acad. Sci. 95, 9355-9359 (1998).pubmed.ncbi.nlm.nih.gov/9689084
  14. [14]C. L. Hueschen, A. R. Dunn and R. Phillips, Wildebeest herds on rolling hills: Flocking on arbitrary curved surfaces, Phys. Rev. E 108, 024610 (2023).DOI
  15. [15]A. Corbetta and F. Toschi, Physics of human crowds, Annu. Rev. Condens. Matter Phys. 14, 311-333(2023).DOI
  16. [16]G. Volpe, I. Buttinoni, D. Vogt, H. Kümmemer, and C. Bechinger, Microswimmers in patterned environments, Soft Matter 7, 8810-8815 (2011).DOI
  17. [17]H. R. Jiang, N. Yoshinaga, and M. Sano, Active motion of a janus particle by self-thermophoresis in a defocused laser beam, Phys. Rev. Lett. 105, 268302 (2010).pubmed.ncbi.nlm.nih.gov/21231718
  18. [18]A. Kudrolli, G. Lumay, D. Volfson, and L. S. Tsimring, Swarming and swirling in self-propelled polar granular rods, Phys. Rev. Lett. 100, 058001 (2008).pubmed.ncbi.nlm.nih.gov/18352433
  19. [19]J. Toner, Giant number fluctuations in dry active polar fluids: A shocking analogy with lightning rods, J. Chem. Phys. 150, 154120 (2019).DOI
  20. [20]C. A Whitfield, T. C. Adhyapak, A. Tiribocchi, G. P. Alexander, D. Marenduzzo, S. Ramaswamy, Hydrodynamic instabilities in active cholesteric liquid crystals, Eur. Phys. J. E. 40, 50 (2017).DOI
  21. [21]Y. Duan. J. Agudo-Canalejo, R. Golestanian, and B. Mahault, Dynamical pattern formation without self-attraction in quorum-sensing active matter: The interplay between nonreciprocity and motility, Phys. Rev. Lett. 131, 148301 (2023).DOI
  22. [22]A. V. Ivlev, J. Bartnick, M. Heinen, C. -R. Du, V. Nosenko, and H. Löwen, Statistical mechanics where Newton’s third law is broken, Phys. Rev. X 5, 011035 (2015).DOI
  23. [23]Y. F. Lin, A. Ivlev, H. Löwen, L. Hong, and C. R. Du, Structure and dynamics of a glass-forming binary complex plasma with non-reciprocal interaction, Europhys. Lett. 123, 35001 (2018).DOI
  24. [24]R. Mandal, S. S. Jaramillo, and P. Sollich, Robustness of traveling states in generic non-reciprocal mixtures, Phys. Rev. E 109, L062602 (2024).DOI
  25. [25]S. Saha, J. Agudo-Canalejo, and R. Golestanian, Scalar Active Mixtures: The Nonreciprocal Cahn-Hilliard Model, Phys. Rev. X 10, 041009 (2020).DOI
  26. [26]Z. You, A. Baskaran, and M. C. Marchetti, Nonreciprocity as a generic route to traveling states, PNAS 117, 19767-19772 (2020).DOI
  27. [27]M. E. Cates and J. Tailleur, Motility-Induced Phase Separation, Annu. Rev. Condens. Matter Phys. 6, 219-244 (2015).DOI
  28. [28]B. Partridge and C. F. Lee, Critical motility-induced phase separation belongs to the Ising universality class, Phys. Rev. Lett., 123, 068002 (2019).DOI
  29. [29]A. Torres-Carbajal and F. J. Sevilla, Motility-induced phase separation of soft Brownian particles, Phys. Fluids 36, 027129 (2024).
  30. [30]T. Vicsek, A. Czirók, E. Ben-Jacob, I. Cohen, and O. Shochet, Novel Type of Phase Transition in a System of Self-Driven Particles, Phys. Rev. Lett. 75, 1226 (1995).pubmed.ncbi.nlm.nih.gov/10060237
  31. [31]A. P. Solon, H. Chaté, and J. Tailleur, From Phase to Microphase Separation in Flocking Models: The Essential Role of Nonequilibrium Fluctuations, Phys. Rev. Lett. 114, 068101 (2015).pubmed.ncbi.nlm.nih.gov/25723246
  32. [32]H. Zhao, A. Košmrlj, and S. S. Datta, Chemotactic Motility-Induced Phase Separation, Phys. Rev. Lett. 131, 118301 (2023).arxiv.org/abs/2301.12345
  33. [33]I. S. Aranson, Bacterial active matter, Rep. Prog. Phys. 85, 076601 (2022).DOI
  34. [34]J. Tailleur and M. E. Cates, Statistical Mechanics of Interacting Run-and-Tumble Bacteria, Phys. Rev. Lett. 100, 218103 (2008).DOI
  35. [35]G. S. Redner, M. F. Hagan, and A. Baskaran, Structure and Dynamics of a Phase-Separating Active Colloidal Fluid, Phys. Rev. Lett. 110, 055701 (2013).
  36. [36]J. Agudo-Canalejo and R. Golestanian, Active Phase Separation in Mixtures of Chemically Interacting Particles, Phys. Rev. Lett. 123, 018101 (2019).
  37. [37]J. Stenhammar, D. Marenduzzo, R. J. Allen, and M. E. Cates, Phase behaviour of active Brownian particles: the role of dimensionality, Soft Matter 10, 1489-1499 (2014).DOI
  38. [38]A. I. Curatolo, N. Zhou, Y. Zhao, C. Liu, A. Daerr, J. Tailleur, J. Huang, Cooperative pattern formation in multi-component bacterial systems through reciprocal motility regulation, Nat. Phys. 16, 1152–1157 (2020).DOI
  39. [39]S. Mondal and P. Das, Domain growth kinetics in active binary mixtures, J. Chem. Phys. 161, 134902 (2024).DOI
  40. [40]R. Wittkowski, A. Tiribocchi, J. Stenhammar, R. J. Allen, D. Marenduzzo, and M. E. Cates, Scalar φ⁴ field theory for active-particle phase separation, Nat. Commun. 5, 4351 (2014).DOI
  41. [41]S. Pattanayak, S. Mishra, and S. Puri, Ordering kinetics in the active model B, Phys. Rev. E 104, 014606 (2021).DOI
  42. [42]S. Pattanayak, S. Mishra, and S. Puri, Domain Growth in the Active Model B: Critical and Off-critical Composition, Soft Matter. 19, 286 (2021).DOI
  43. [43]P. K. Yadav, S. Mishra, S. Puri, Coarsening kinetics in active model : Macroscale and microscale phase separation, Phys. Rev. E. 112, 035412 (2025).arxiv.org/abs/2506.14548
  44. [44]C. Dombrowski, L. Cisne, S. Chatkaew, R. E. Goldstein, and J. O. Kessler, Self-Concentration and Large-Scale Coherence in Bacterial Dynamics, Phys. Rev. Lett. 93, 098103 (2004).pubmed.ncbi.nlm.nih.gov/15447144
  45. [45]M. N. van der Linden, L. C. Alexander, D. G. A. L. Aarts, and O. Dauchot, Interrupted Motility Induced Phase Separation in Aligning Active Colloids, Phys. Rev. Lett. 123, 098001 (2019).DOI
  46. [46]I. Theurkauff, C. Cottin-Bizonne, J. Palacci, C. Ybert, and L. Bocquet, Dynamic Clustering in Active Colloidal Suspensions with Chemical Signaling, Phys. Rev. Lett. 108, 268303 (2012).DOI
  47. [47]T. Bäuerle, A. Fischer, T. Speck, and C. Bechinger, Self-organization of active particles by quorum sensing rules, Nat Commun 9, 3232 (2018).DOI
  48. [48]Y. Fily, M. C. Marchetti, Athermal Phase Separation of Self-Propelled Particles with No Alignment, Phys. Rev. Lett. 108, 235702 (2012).DOI
  49. [49]W. J. M. Ridgway, M. P. Dalwadi, P. Pearce, and S. J. Chapman, Motility-Induced Phase Separation Mediated by Bacterial Quorum Sensing, Phys. Rev. Lett. 131, 228302 (2023).DOI
  50. [50]A. Onuki, Phase Transition Dynamics (Cambridge University Press, Cambridge, 2002).
  51. [51]S. Dattagupta and S. Puri, Dissipative Phenomena in Condensed Matter: Some Applications (Springer-Verlag, Heidelberg, 2004).DOI
  52. [52]Kinetics of Phase Transitions, edited by S. Puri and V. K. Wadhawan (CRC Press, Boca Raton, 2009).
  53. [53]S. Ahmad, S. K. Das, and S. Puri, Kinetics of phase separation in fluids: A molecular dynamics study, Phys. Rev. E 82, 040107(R) (2010).DOI
  54. [54]S. Ahmad, S. K. Das, and S. Puri, Crossover in growth laws for phase-separating binary fluids: Molecular dynamics simulations, Phys. Rev. E 85, 031140 (2012).pubmed.ncbi.nlm.nih.gov/22587071
  55. [55]A. J. Bray, Theory of phase-ordering kinetics, Adv. Phys. 43, 357–459 (1994).
  56. [56]G. Porod, Small-Angle X-Ray Scattering, edited by O. Glatter and O. Kratky (Academic Press, New York 1982), Vol. 42
  57. [57]Y. Oono and S. Puri, Large wave number features of form factors for phase transition kinetics, Mod. Phys. Lett. B 02, 861-867 (1988).DOI
  58. [58]S. N. Floor, A. L. Melott, and P. M. Motl, Simulated versus observed cluster eccentricity evolution, The Astrophysical Journal 611, 153–157 (2004).DOI
  59. [59]D. A. McQuarrie. Statistical Physics. Harper & Row, New York, (1975).
  60. [60]A. Chauhan, D. Gogoi, S. Puri, and A. Singh, Effect of amphiphilic polymers on phase separating binary mixtures: A dpd simulation study, J. Chem. Phys. 159, 204901 (2023).DOI

Paper details

vlnc:W_wmcj8cpj
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 10

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents