betaFeedback

WorksPhysics and Astronomy

A sharp interface approach for wetting dynamics of coated droplets and soft particles

Abstract

The wetting dynamics of liquid particles, from coated droplets to soft capsules, holds significant technological interest. Motivated by the need to simulate liquid metal droplet with an oxidize surface layer, in this work we introduce a computational scheme that allows to simulate droplet dynamics with general surface properties and model different levels of interface stiffness, describing also cases that are intermediate between pure droplets and capsules. Our approach is based on a combination of the immersed boundary (IB) and the lattice Boltzmann (LB) methods. Here, we validate our approach against the theoretical predictions in the context of shear flow and static wetting properties and we show its effectiveness in accessing the wetting dynamics, exploring the ability of the scheme to address a broad phenomenology.

Introduction

The wetting of a solid surface by a liquid coincides with its ability to preserve the contact with the liquid [1, 2, 3, 4]. The wettability of a solid substrate by a pure droplet is quantified by the droplet’s equilibrium contact angle θeq\theta_{eq}, which, in turn, is determined by the balance between adhesive and cohesive forces of the three phases involved (solid, liquid, vapor). At the macroscopic scale, Young’s equation [5] describes this balance as

cos⁡θeq=σsl−σsgσ(1)\cos\theta_{eq} = \frac{\sigma_{sl}-\sigma_{sg}}{\sigma} \tag*{(1)}

where σsl\sigma_{sl}, σsg\sigma_{sg} and σ\sigma are the solid-liquid, solid-gas and liquid-gas surface tensions, respectively. Eq. (1) also estimates the degree of wettability, making the distinction between poor (θeq>90∘\theta_{eq} > 90^\circ) and good (θeq<90∘\theta_{eq} < 90^\circ) wetting regimes. Out of equilibrium, the additional complexities arising from time dependence and viscous dissipation make dynamic wetting critical to a wide range of phenomena including droplet spreading, capillary rise, imbibition, and more complex situations like fluid displacement in porous media or multiphase flow in oil recovery [6, 7, 8, 9]. The recent development of new catalytic devices, for example, requires the usage of liquid metals and metal alloys in the form of catalytic liquid droplets adsorbed on a porous solid support [10, 11]. However, several liquid metals such as gallium and gallium-based alloys oxidize when exposed to air and an inherent oxide layer appears on the top of the surface. This oxide layer acts as a solid-like “skin”, encapsulating a liquid metal core [12, 13] and changes the wetting properties of the droplet [14, 15, 16, 17]. Another example concerns the so called liquid metal marbles, realised by rolling a small liquid droplet in an poorly wetting powder. Because of the layer of powder grains at the liquid-air interface, the wetting of these droplets is inhibited [18, 19], as required in some recent technological and microfluidic applications [20, 21]. The lattice

α=(α1,α2,α3)\alpha= (\alpha_1,\alpha_2,\alpha_3)Modelaabbccdd
Pre-stressed particles
α=αˉ(1,0,0)\alpha= \bar{\alpha}(1,0,0)Pure droplet0.780.420.01.0
α=αˉ(1,0,1)\alpha= \bar{\alpha}(1,0,1)Softly coated droplet0.870.50.01.0
α=αˉ(1,1,1)\alpha= \bar{\alpha}(1,1,1)Rigidly coated droplet0.910.550.01.0
α=(α1,α2,α3)\alpha= (\alpha_1,\alpha_2,\alpha_3)Modelaabbccdd
Non-pre-stressed particles
α=αˉ(0,0,1)\alpha= \bar{\alpha}(0,0,1)Pure elastic capsule0.712.00.20.75
α=αˉ(0,1,1)\alpha= \bar{\alpha}(0,1,1)Non-pre-stressed capsule0.71.60.20.75

Table 1. List of system models that can be explored by tuning the parameter α1\alpha_1, α2\alpha_2, and α3\alpha_3 in the interface model for a generic particle reported in Eq. (2). The left and right tables refer to the pre-stressed and non-pre-stressed particle classes, respectively. For each model we display the corresponding fitting parameters aa, bb, cc and dd appearing in Eq. (27).

Boltzmann (LB) method has been used since decades to address problems in wettability [22, 23, 25, 26, 27, 28, 29, 30, 31, 32, 33] [24]. A typical strategy used to simulate droplets within the framework of LB includes introducing non-ideal interface force models, such as the Shan-Chen [34] and the Free-Energy ones [35]. However, these approaches model a diffuse interface. Simulating droplets with a complex rheology, specifically coated droplets including for example liquid metal ones with an oxide layer or liquid marbles, require the use of a constitutive law for the interface. In this case it is more convenient to use a method that reproduces the sharp-interface limit of hydrodynamics. In addition, pseudopotential or free-energy approaches do not easily allow to model a behaviour that, as it is typical for coated droplets, is intermediate between the case of a pure droplet and that of a capsule, as is the case for coated droplets. For these reasons, here we have opted for combining the LB model with an immersed boundary (IB) method, which naturally preserves the hydrodynamic sharp-interface limit, to simulate the complex droplet’s wetting dynamics (see Figure 1).

Image not extracted

Figure 1. Sketch of the wetting dynamics of a generic particle with initial radius RR initially placed in contact with a flat wall. The interface is resolved with a 3D triangular mesh. On each triangular face jj, some force contributions φ\boldsymbol{\varphi} are computed and distributed to the vertices i∈ji \in j, with the aim to consider (i) the interface elasticity/rigidity, (ii) the volume conservation, and (iii) the wall-particle interaction. We also report the corresponding involved parameters.

Our goal is to introduce a comprehensive numerical approach that allows modelling droplets with complex interfacial properties in a consistent way. Here, we model the interface as a 3D triangular mesh and employ a constitutive law based on the theory of Barthès-Biesel and Rallison [36] to explore the case of coated droplets. This approach allows us to describe in a continuous way the transition from pure droplet to capsule-like models by minimising the number of involved parameters. We provide a validation of our IBLB numerical simulations against the theoretical prediction in case of a simple shear flow experiment. Then, we perform wetting dynamics simulations, which show a good agreement with experimental observations in the case of a pure droplet, and we explore the range of accessible contact angles in terms of the involved parameters and the intensity of the interaction with the wall. With this approach we aim at providing a qualitative approximation of the mechanical behaviour of droplets with a complete and wide range of interfacial properties. Nevertheless, the model can be further refined to include additional interface properties, enhancing its accuracy and applicability, such as an extended model to mimic the oxidized layer thickness when dealing with liquid metal droplets.

The paper is organised as follows: in Sec. II we describe the interface model introduced in the IBLB framework. Then, in Sec. III we summarise the main features of the IBLB model employed. The benchmark of a droplet in a simple shear flow is shown in Sec. IV. In the context of the wetting dynamics, § 0.5 A report a model validation, while all wetting dynamics facets are analysed and discussed in § 0.5 B. Results are summarised in Sec. VI.

Interface model

In this section, we describe the theoretical model employed in this work to simulate a generic soft particle. Following Barthès-Biesel & Rallison [36], we consider a two-dimensional, isotropic and homogeneous elastic interface with no bending resistance. Its mechanical response is characterised by an interface strain energy wS=wS(I1,I2,α1,α2,α3)w_{\mathrm{S}} = w_{\mathrm{S}}(I_1,I_2,\alpha_1,\alpha_2,\alpha_3) which is written in terms of the principal strain invariants I1,2I_{1,2} and three parameters α1,2,3\alpha_{1,2,3} as [36]

wS(I1,I2,α1,α2,α3)=wS,0+12(α1−α3)log⁡(I2+1)+18(α1+α2)log⁡2(I2+1)+α3[12(I1+2)−1],(2)w_{\mathrm{S}}(I_1,I_2,\alpha_1,\alpha_2,\alpha_3) = w_{\mathrm{S},0} + \frac{1}{2}(\alpha_1-\alpha_3)\log(I_2+1) + \frac{1}{8}(\alpha_1+\alpha_2)\log^2(I_2+1) + \alpha_3\left[\frac{1}{2}(I_1+2)-1\right], \tag*{(2)}

where wS,0w_{\mathrm{S},0} is a reference value. I1I_1 and I2I_2 quantify the strain and dilation state of the membrane, respectively. The parameters α1,2,3\alpha_{1,2,3}, instead, characterize the material properties: The pre-stress α1\alpha_1 is an isotropic tension without an applied load; α2\alpha_2 is the resistance against area dilatation and α3\alpha_3 the resistance against shear deformation (i.e., the strain modulus). For the sake of simplicity, hereafter we will refer to these three parameters as α=(α1,α2,α3)\alpha=(\alpha_1,\alpha_2,\alpha_3). Concerning their choice, we distinguish the two main classes of pre-stressed (α1>0\alpha_1>0) and non-pre-stressed (α1=0\alpha_1=0) particles. For the latter class, an appropriate combination of α2\alpha_2 and α3\alpha_3 leads to the well-known Skalak [37] and Neo-Hookean [38] models. By conveniently tuning these parameters, one can switch from a pure droplet (α=(σ,0,0)\alpha=(\sigma,0,0), where σ\sigma is the surface tension) to a pure elastic capsule (α=(0,0,α3>0)\alpha=(0,0,\alpha_3>0)) and describe intermediate and more complex situations. In particular, in this work we consider also the classes of particles with α=(α1>0,0,α3>0)\alpha=(\alpha_1>0,0,\alpha_3>0) and α=(α1>0,α2>0,α3>0)\alpha=(\alpha_1>0,\alpha_2>0,\alpha_3>0). We call the first type of particle a “softly coated droplet” because it has the characteristic surface tension term of a droplet, but also a strain modulus because of α3>0\alpha_3>0. Because of the presence of a dilatational term α2>0\alpha_2>0, we call the second type “rigidly coated droplet”. Here we argue that the latter case could be used to describe the interfacial properties of particles like liquid metal droplets with oxidized surface. In Table I we summarise the type of particles investigated in this work.

In order to check that Eq. (2) leads to the correct particle dynamics, we consider the case of a shear flow experiment, where a particle with initial radius RR and dynamic viscosity μ\mu is placed between two distant moving walls which generate a shear rate γ˙\dot{\gamma} (see top panel of Figure 2). In this setup, time-dependent motion of the particle shape may be distinguished into two contributions: a solid body rotation and a stretching. Notice that the interface may rotate via a tank-treading motion despite of the particle reaching its steady state. This means that at this stage the interface deformation is constant in time at an Eulerian point xx, but it is not constant by looking at a point XX on the interface. Thus, following concepts and notation of Ref. [36], the time-evolution of the dimensionless position xx (i.e., the position divided by the initial radius RR), which is representative of the deformation field, reads

Image not extracted

Figure 2. Simulated experiment of a single particle under shear flow for the pre-stressed (panels (a)-(c)) and non-pre-stressed (panels (d)-(f)) particle models. In all panels, different symbols/colours refer to different models. Top panels: a sketch of the shear experiment and final shape of the particle for Ca=0.3Ca=0.3. Panels (a) and (d): time evolution of the deformation index D(t)D(t) (Eq. (22)) as a function of time for capillary number Ca=0.05Ca=0.05. Time is shown normalised with the shear rate γ˙\dot{\gamma}, and dashed lines refer to the analytical solutions of Eqs. (6). Panels (b) and (e): the steady-state value of the deformation DD as a function of CaCa. Dashed lines draw the theroretical predictions: Eq. (24) (salmon line) for the pure droplet case and Eq. (23) (other colors lines) for all the other models. Panels (c) and (f): the steady-state value of the inclination angle Θ\Theta as a function of CaCa. To validate the model in the case of a pure droplet, we report black crosses from Ref. [67], and we draw dotted lines for Eq. (25).

x=X+β[K⋅X+XX⋅(J−K)⋅X],(3)\mathbf{x}=\mathbf{X}+\beta\left[\mathbf{K}\cdot\mathbf{X}+\mathbf{X}\mathbf{X}\cdot(\mathbf{J}-\mathbf{K})\cdot\mathbf{X}\right], \tag*{(3)}

where β≪1\beta\ll1 is the expansion coefficient around the initial spherical position (we truncate the equation at the leading order in β\beta), while J\mathbf{J} and K\mathbf{K} are two symmetric and traceless second-rank tensors which depend only on time. It follows that the instantaneous external shape of the particle rr can be computed in terms of the norm of Eq. (3), together with its normal n\mathbf{n} [36],

r=∣x∣=1+βX⋅J⋅X=1+βx⋅J⋅xr2,(4)r=\lvert\mathbf{x}\rvert=1+\beta\mathbf{X}\cdot\mathbf{J}\cdot\mathbf{X}=1+\beta\frac{\mathbf{x}\cdot\mathbf{J}\cdot\mathbf{x}}{r^2}, \tag*{(4)}
n=xr+2β[xx⋅J⋅xr3−J⋅xr].(5)\mathbf{n}=\frac{\mathbf{x}}{r}+2\beta\left[\frac{\mathbf{x}\mathbf{x}\cdot\mathbf{J}\cdot\mathbf{x}}{r^3}-\frac{\mathbf{J}\cdot\mathbf{x}}{r}\right]. \tag*{(5)}

Since in Eqs. (4) only tensor J\mathbf{J} appears, it means that J\mathbf{J} describes the overall deformation (i.e., the stretching contribution), while K\mathbf{K} describes the motion on the interface (i.e., the solid body rotation) [36]. We remark that J\mathbf{J} is traceless because of the volume conservation constraint, whereas this property for K\mathbf{K} can be checked a posteriori. In the limit of small deformations, the evolution equations for J\mathbf{J} and K\mathbf{K} are given by [36]

DDtK=5E2λ+3+L2λ+3+M(6λ+4)(2λ+3)(19λ+16),DDt(J−K)=219λ+16M,(6)\begin{aligned} \frac{\mathcal{D}}{\mathcal{D}t}\mathbf{K} &= \frac{5\mathbf{E}}{2\lambda+3}+\frac{\mathbf{L}}{2\lambda+3}+\frac{\mathbf{M}(6\lambda+4)}{(2\lambda+3)(19\lambda+16)},\\ \frac{\mathcal{D}}{\mathcal{D}t}(\mathbf{J}-\mathbf{K}) &= \frac{2}{19\lambda+16}\mathbf{M}, \tag*{(6)} \end{aligned}

where

DADt=dAdt−(Ω⋅A−A⋅Ω)(7)\frac{\mathcal{D}A}{\mathcal{D}t}=\frac{dA}{dt}-(\boldsymbol{\Omega}\cdot A-A\cdot\boldsymbol{\Omega}) \tag*{(7)}

is the Jaumann derivative[36] applied to a generic tensor AA, which takes into account the rotation of the particle with the vorticity of the external fluid. EE and Ω\boldsymbol{\Omega} are the symmetric and asymmetric part of the velocity gradient, respectively, λ\lambda is the viscosity ratio between inside and outside fluids, and

L=4(α2+α3)J−(6α2+10α3)K,(8)L=4(\alpha_{2}+\alpha_{3})J-(6\alpha_{2}+10\alpha_{3})K, \tag*{(8)}
M=−4(α1+2α2+2α3)J+(12α2+16α3)K.(9)M=-4(\alpha_{1}+2\alpha_{2}+2\alpha_{3})J+(12\alpha_{2}+16\alpha_{3})K. \tag*{(9)}

By numerically integrating Eq. (6), it is possible to obtain information on the transient deformation dynamics since the tensor JJ is directly related to the particle-deformation as [39]

D=Ca(J112+J122)1/2,(10)D=\mathrm{Ca}(J_{11}^{2}+J_{12}^{2})^{1/2}, \tag*{(10)}

where Ca=μRγ˙/α\mathrm{Ca}=\mu R\dot{\gamma}/\alpha is the capillary number. In the latter definition, one can consider α=α1=σ\alpha=\alpha_{1}=\sigma for a pure droplet or α=α3\alpha=\alpha_{3} for a pure capsule. For the sake of simplicity, we fix the values of parameters α1\alpha_{1}, α2\alpha_{2}, and α3\alpha_{3} to be equal to the same value αˉ\bar{\alpha} and we will refer to this triad of values simply as α=αˉ(0/1,0/1,0/1)\alpha=\bar{\alpha}(0/1,0/1,0/1), with the vector elements turned on (1) and off (0) with the corresponding model (see Table I).

Numerical implementation

The dynamics of the inner and outer fluid is simulated using a single-component lattice Boltzmann (LB) method in terms of the fluid particle populations fi(x^,t)f_i(\hat{\boldsymbol{x}},t). The latter represents the probability distribution function of finding a fluid particle in a discrete lattice (Eulerian) node x^\hat{\boldsymbol{x}} at a discrete time tt. The corresponding macroscopic behaviour is recovered in the long-wavelength limit, which allows the link with the Navier-Stokes equations. Indeed, the solutions of the Navier-Stokes equation for the total density and momentum are easily accessible from the populations as ρ(x^,t)=∑ifi(x^,t)\rho(\hat{\boldsymbol{x}},t)=\sum_i f_i(\hat{\boldsymbol{x}},t) and ρ(x^,t)u(x^,t)=∑icifi(x^,t)\rho(\hat{\boldsymbol{x}},t)\boldsymbol{u}(\hat{\boldsymbol{x}},t)=\sum_i\boldsymbol{c}_i f_i(\hat{\boldsymbol{x}},t), respectively, with ci\boldsymbol{c}_i representing a set of 19 discrete velocities (i=0,…,18i=0,\ldots,18) living on a three-dimensional lattice (i.e., we employ a D3Q19 LB model). The dynamics of fif_i is ruled by a continuous succession of propagation and collision steps, as highlighted by the discretized Boltzmann equation [40] [41]

fi(x^+ciΔt,t+Δt)−fi(x^,t)=−Δtτ[fi(x^,t)−fi(eq)(x^,t)]+wi(1−Δt2τ)((ci−u)⋅Fcs2+(ci⋅F)(ci⋅u)cs4)Δt,(11)f_i(\hat{\boldsymbol{x}}+\boldsymbol{c}_i\Delta t,t+\Delta t)-f_i(\hat{\boldsymbol{x}},t)=-\frac{\Delta t}{\tau}\left[f_i(\hat{\boldsymbol{x}},t)-f_i^{(\mathrm{eq})}(\hat{\boldsymbol{x}},t)\right]+w_i\left(1-\frac{\Delta t}{2\tau}\right)\left(\frac{(\boldsymbol{c}_i-\boldsymbol{u})\cdot\boldsymbol{F}}{c_s^2}+\frac{(\boldsymbol{c}_i\cdot\boldsymbol{F})(\boldsymbol{c}_i\cdot\boldsymbol{u})}{c_s^4}\right)\Delta t, \tag*{(11)}

where Δt\Delta t is the time step. The propagation of fif_i on the lattice is described by the l.h.s. of Eq. (11) with the help of ci\boldsymbol{c}_i, while the single-relaxation-time BGK approximation of the collision operator appears as the first term in the r.h.s. The latter has the aim of modelling the relaxation of fif_i towards the equilibrium distribution fi(eq)(x^,t)f_i^{(\mathrm{eq})}(\hat{\boldsymbol{x}},t), represented as the local Maxwellian distribution

fi(eq)(x,t)=wiρ[1+u⋅cics2+ukuj(ci,kci,j−cs2δkj)2cs4].(12)f_i^{(\mathrm{eq})}(\boldsymbol{x},t)=w_i\rho\left[1+\frac{\boldsymbol{u}\cdot\boldsymbol{c}_i}{c_s^2}+\frac{u_k u_j(c_{i,k}c_{i,j}-c_s^2\delta_{kj})}{2c_s^4}\right]. \tag*{(12)}

The relaxation process lasts for a relaxation time τ\tau. In Eq. (12), fi(eq)f_i^{(\mathrm{eq})} is weighted by the lattice-dependent weights wiw_i3 and depends on the speed of sound cs=Δx^/(3Δt)c_s=\Delta\hat{x}/(\sqrt{3}\Delta t), where Δx^\Delta\hat{x} is the lattice spacing. The last term of Eq. (11) refers to the forcing implementation following the Guo scheme [42], where F\boldsymbol{F} is the force acting on the fluid. Notice that, to guarantee the second-order space-time accuracy, this forcing scheme modifies the fluid velocity as ρ(x^,t)u(x^,t)=∑icifi(x^,t)+FΔt/2\rho(\hat{\boldsymbol{x}},t)\boldsymbol{u}(\hat{\boldsymbol{x}},t)=\sum_i\boldsymbol{c}_i f_i(\hat{\boldsymbol{x}},t)+\boldsymbol{F}\Delta t/2. In our simulations, we keep fixed to unity both Δx^\Delta\hat{x} and Δt\Delta t. Furthermore, the fluid dynamic viscosity μ\mu in LB models is related to the relaxation time τ\tau as μ=cs2ρ(τ−1/2)\mu=c_s^2\rho(\tau-1/2). Here, we keep the viscosity ratio λ\lambda fixed to unity, since the investigation on the role played by λ\lambda goes beyond the purpose of this work. Then, to simulate the interface of a coated droplet or soft particle immersed in the surrounding LB fluid, we model the spherical particle interface using a 3D triangular mesh generated from a recursive refining of an icosahedron. Thus, the mesh resolution is defined in terms of the total number of triangular faces NfN_f (see Figure 6 for a pictorial view of particles with different resolutions). To couple the soft particle dynamics with that of the surrounding fluid, we use the immersed boundary (IB) method, i.e., a fluid-mesh interaction method developed for the first time by Peskin [43] and based on the distinction between interface (Lagrangian) nodes q(t)\mathbf{q}(t) and fluid (Eulerian) nodes x^\hat{\mathbf{x}}. The resulting coupling is distinct in two operations, i.e., interpolation and spreading. The interpolation operation consists of the computation of the ii-th interface-node velocity q˙i(t)\dot{\mathbf{q}}_i(t) from the fluid one (u(x^,t))(\mathbf{u}(\hat{\mathbf{x}},t)) as [41].4

Image not extracted

Figure 6. Resolution test for the wetting dynamics experiment a pure droplet with α=αˉ(1,0,0)\alpha= \bar{\alpha}(1,0,0) (panels (a) and (b)) and for a mixed system with α=αˉ(1,1,1)\alpha= \bar{\alpha}(1,1,1) (panels (c) and (d)). We compare the time evolution of the zz-coordinate of the centre-of-mass ZCMZ_{\mathrm{CM}} for different resolutions, given in terms of the number of mesh triangular faces NfN_f. Time tt is shown in simulation units.

q˙i(t)=∑x^u(x^,t)δD(x^−qi(t))Δx^3.(13)\dot{\mathbf{q}}_i(t)=\sum_{\hat{\mathbf{x}}}\mathbf{u}(\hat{\mathbf{x}},t)\delta_D(\hat{\mathbf{x}}-\mathbf{q}_i(t))\Delta\hat{x}^{3}. \tag*{(13)}

This operation allows to update the node position qi(t)\mathbf{q}_i(t) as:

qi(t+Δt)=qi(t)+q˙i(t)Δt.(14)\mathbf{q}_i(t+\Delta t)=\mathbf{q}_i(t)+\dot{\mathbf{q}}_i(t)\Delta t. \tag*{(14)}

Then, the spreading operation is an interpolation of the interface nodal force to the fluid one which allows to make the latter aware of the presence of the interface: at this step, the total force (volume-)density the particle exerts on the fluid at the Eulerian node x^\hat{\mathbf{x}} is given by

F(x^,t)=∑iφi(t)δD(x^−qi(t)),(15)\mathbf{F}(\hat{\mathbf{x}},t)=\sum_i\boldsymbol{\varphi}_i(t)\delta_D(\hat{\mathbf{x}}-\mathbf{q}_i(t)), \tag*{(15)}

where φi\boldsymbol{\varphi}_i is the total force on the Lagrangian node ii and the sum runs over all Lagrangian nodes. Both operations involve the so-called discrete delta function δD\delta_D, which is used to approximate the Dirac delta function on our lattice and is defined as [43, 44] [41]

δD(x^)=1Δx^3ϕ4(x^)ϕ4(y^)ϕ4(z^),(16)\delta_D(\hat{\mathbf{x}})=\frac{1}{\Delta\hat{x}^{3}}\phi_4(\hat{x})\phi_4(\hat{y})\phi_4(\hat{z}), \tag*{(16)}

where ϕ4(r)\phi_4(r) is the “interpolation stencil” involving four Eulerian nodes along each coordinate axis [45] and defined as follows:

ϕ4(x^)={18(3−2∣x^∣+1+4∣x^∣−4x^2)0≤∣x^∣18(5−2∣x^∣−−7+12∣x^∣−4x^2)Δx^≤∣x^∣≤2Δx^02Δx^≤∣x^∣(17)\phi_4(\hat{x})= \begin{cases} \frac{1}{8}\left(3-2|\hat{x}|+\sqrt{1+4|\hat{x}|-4\hat{x}^{2}}\right) & 0\leq|\hat{x}| \\ \frac{1}{8}\left(5-2|\hat{x}|-\sqrt{-7+12|\hat{x}|-4\hat{x}^{2}}\right) & \Delta\hat{x}\leq|\hat{x}|\leq2\Delta\hat{x} \\ 0 & 2\Delta\hat{x}\leq|\hat{x}| \end{cases} \tag*{(17)}

The resulting IBLB method has been largely used to simulate the dynamics of capsules [45, 46, 47, 48, 49] and red blood cells [50, 51, 52, 53, 54]. However, only a few works employed this method also for simulating droplet dynamics [55, 56, 57]. A detailed step-by-step description of the IBLB algorithm implementation can be found in Ref. [41].

In our implementation, the total nodal force φi\boldsymbol{\varphi}_i, appearing in Eq. (15) and acting on the ii-th node at position ri\mathbf{r}_i at time tt, is given by the sum of several contributions, i.e.,

φi=φiS+φiV+φiW.(18)\boldsymbol{\varphi}_i=\boldsymbol{\varphi}_i^{S}+\boldsymbol{\varphi}_i^{V}+\boldsymbol{\varphi}_i^{W}. \tag*{(18)}

Each contribution plays a distinct role. First of all, φiS\boldsymbol{\varphi}_i^{S} incorporates the information on the elastic properties of the interface. Thus, we compute this nodal force term as

φiS=−∂∂{qi}wS({qi}),(19)\boldsymbol{\varphi}_i^{S}=-\frac{\partial}{\partial\{\mathbf{q}_i\}}w_S(\{\mathbf{q}_i\}), \tag*{(19)}

where wSw_S is the generalised strain energy defined in Eq. (2). Eq. (17) is calculated using a first order finite element method as described in Ref. [45]. Then, because we are dealing with incompressible fluids, we need to consider a volume conservation constraint. With this aim, we follow Ref. [45] and we write the nodal volume force contribution φiV\boldsymbol{\varphi}_i^{V} as [58]

φiV=−∂∂{qi}wV({qi}),(20)\boldsymbol{\varphi}_i^{V}=-\frac{\partial}{\partial\{\mathbf{q}_i\}}w_V(\{\mathbf{q}_i\}), \tag*{(20)}

where wV=kV(V−V0)2/2V0w_V=k_V(V-V_0)^2/2V_0 is the volume energy. In this definition of the volume energy, kVk_V refers to the volume-force coefficient and it is kept fixed to 1. V=∑jVjV=\sum_j V_j is the instantaneous total particle volume, with the index jj running over the number of faces NfN_f5, while V0V_0 is the initial total particle volume. Further details on how to compute nodal force contributions in Eqs. (19) and (20) can be found in Ref. [58]. The last contribution in Eq. (18) corresponds to the wall-particle interaction, the key element for wetting dynamics simulations. The IBLB approach used in this work involves only one single fluid component and it is not possible to control the wall-fluid surface tensions σsl\sigma_{sl} and σsg\sigma_{sg} by introducing two different interactions as, for example, Huang and coworkers did in the case of multi-component pseudopotential LB models [24]. However, many implementation of fluid-wall interactions in the case of single-component LB models [59] use a pseudopotential-like fluid-wall interaction which does not control σsl\sigma_{sl} and σsg\sigma_{sg} separately, but only their overall effect. These models work well in describing the wetting dynamics of droplets, despite of this limitation. In this work, we follow the same approach and we introduce a Lennard-Jones interaction on behalf of the wall-particle interaction:

φiw=48ϵ[(ξdi)12−12(ξdi)6]didi2,(21)\varphi_i^{\mathrm{w}} = 48\epsilon\left[\left(\frac{\xi}{d_i}\right)^{12} - \frac{1}{2}\left(\frac{\xi}{d_i}\right)^6\right]\frac{\mathbf{d}_i}{d_i^2}, \tag*{(21)}

where di\mathbf{d}_i is the shortest displacement vector between the centroid of the triangle to which node ii belongs and the wall surface, and di=∣di∣d_i = \lvert\mathbf{d}_i\rvert. It means that the force is computed once for each triangle, and then it is distributed on its vertices. The choice of employing a Lennard-Jones interaction potential results from the necessity to model adhesive and repulsive forces between the droplet or particle and the surface. It follows the large amount of works based on molecular dynamics simulations which successfully studied the behaviour of nanodroplets or ridges on chemically patterned substrates [62]. Notice that in this work, we set ξ=0.5Δx^\xi= 0.5\Delta\hat{x} to have an interface-wall interaction that decays to zero after one lattice spacing, thus respecting as much as possible the microscopic range, and tune the interaction by changing only ϵ\epsilon. The nodal force contributions along with the corresponding parameters are summarised in Figure 1.

By summarizing, the IBLB algorithm implemented in this work matches the following steps [41]:

  1. Compute the nodal force φi\varphi_i on each node ii (Eq. (18));

  2. Spread the nodal force to obtain the force acting on the fluid F(x^,t)\mathbf{F}(\hat{\mathbf{x}},t) via Eq. (15);

  3. Perform the LB integration step: compute equilibrium distributions (Eq. (12)), then apply the collision and perform the propagation. At this stage, F(x^,t)\mathbf{F}(\hat{\mathbf{x}},t) enters in r.h.s. of Eq. (11);

  4. Compute the fluid velocity u(x^,t)\mathbf{u}(\hat{\mathbf{x}},t) from LB populations;

  5. Interpolate the fluid velocity to compute the Lagrangian node velocity (Eq. (13));

  6. Update the position of each node q(t)\mathbf{q}(t) via Eq. (14);

  7. Iterate from step 1.

All simulations have been performed in a periodic domain in the x- and y- directions, while two walls are placed along the (vertical) z-direction. A half-node bounce-back rule implements second-order no-slip boundary conditions at the walls [41]. Dimensional quantities are shown in lattice Boltzmann units (lbu).

Note that, although this method is not able to capture particle breakup and coalescence, it provides an easy way to model different systems by simply tuning the αi\alpha_i parameters, as detailed in Section II. The introduction of the wall-particle interaction (19) induces an accumulation of interface nodes on the wall-particle contact area. Such an aggregation is more prominent for high values of ϵ\epsilon that are required for observing small contact angles, and is responsible for a numerical instability in that regime. Re-meshing technique may help mitigating this problem, but since we are interested in large contact angles, this accumulation does not affect the results presented in this paper.

Benchmark: shear flow dynamics

To showcase the versatility of the interface model proposed in Sec. II, we perform a double analysis by measuring the deformation of coated droplets and soft particles undergoing a shear flow. Indeed, on the one hand, we benchmark our model with what is known in the literature for the case of a pure droplet, while, on the other hand, we explore the different scenarios associated to each particle case listed in Table I. In this setup, we run simulations for particles with an initial radius R=19R = 19 lbu, placed in a channel with a distance between the two walls H=128 lbu. The system has the same size along the other two directions, x and y. In order to vary the capillary number Ca, we systematically tune the values of αˉ\bar{\alpha}, keeping fixed the shear rate γ˙\dot{\gamma} by the constraint of low Reynolds number (Re⁡=10−2\operatorname{Re}=10^{-2}). Without loss of generality, we set the fluid density ρ=1\rho=1 lbu and the relaxation time τ=1\tau=1 lbu, resulting in a dynamic viscosity of the particle μ=1/6\mu=1/6 lbu. In Figure 2(a) and (d), we report simulation data for the time-evolution of the deformation index defined as

D(t)=r1(t)−r3(t)r1(t)+r3(t),(22)D(t)=\frac{r_1(t)-r_3(t)}{r_1(t)+r_3(t)}, \tag*{(22)}

where r1r_1 and r3r_3 are the main particle semi-axes in the shear plane (see the top of Figure 2 for a sketch). The simulation time is normalised with γ˙\dot{\gamma}. Results show a very good agreement between simulations and the time-evolution of the deformation DD defined in Eq. (10), obtained from the analytical solutions of Eqs. (6) (dashed lines). In addition, the steady-state value of the deformation DD can be analytically estimated as a function of the triad of α\alpha as

D=[5α1(3α2+4α3)4(3α1α2+5α1α3+2α2α3+2α32)]Ca,(23)D=\left[\frac{5\alpha_1(3\alpha_2+4\alpha_3)}{4(3\alpha_1\alpha_2+5\alpha_1\alpha_3+2\alpha_2\alpha_3+2\alpha_3^2)}\right]Ca, \tag*{(23)}

where Ca=μRγ˙/α1\mathrm{Ca}=\mu R\dot{\gamma}/\alpha_1. Since Eq. (23) has been computed with α2,α3≠0\alpha_2,\alpha_3\ne0, it does not hold for a pure droplet, for which α=αˉ(1,0,0)\alpha=\bar{\alpha}(1,0,0). In the latter case we have [36]

D=19λ+1616λ+16Ca,(24)D=\frac{19\lambda+16}{16\lambda+16}\mathrm{Ca}, \tag*{(24)}

with λ=1\lambda=1 in the present work. Figure 2(b) and (e) confirm the agreement between simulation data and Eqs. (23) and (22) in the limit of small deformations (i.e., small Ca\mathrm{Ca}), while it diverges for larger values of DD. Note that in Figure 2(b) the theoretical prediction for cases α=αˉ(1,0,1)\alpha=\bar{\alpha}(1,0,1) and α=αˉ(1,1,1)\alpha=\bar{\alpha}(1,1,1) are so close to not being distinguishable.

In addition, to complete the picture of soft particle dynamics under shear flow, we report in Figure 2(c) and (f) the inclination angle Θ\Theta (see top panel) as a function of the capillary number Ca\mathrm{Ca}. In the limit of small deformations and the case of a pure droplet with λ=1\lambda=1, these results are again in agreement with what is expected from simulations [67] (black crosses) and the theory of Chaffey and Brenner [68] (dotted black line) which reads

Θ=π4−(19λ+16)(2λ+3)80(λ+1)Ca.(25)\Theta=\frac{\pi}{4}-\frac{(19\lambda+16)(2\lambda+3)}{80(\lambda+1)}\mathrm{Ca}. \tag*{(25)}

For non-pre-stressed particle models, we observe a stronger dependency on Ca\mathrm{Ca}, probably due to the higher rigidity. To summarize, we find a good agreement for the time-evolution of the particle deformation D(t)D(t) and its steady-state value DD between the analytical solution of the model equations (5) and our numerical model for both coated droplets and soft particles. This is true in the limit of small deformation, which is the basic assumption behind the theory [36]. Furthermore, in the case of a pure droplet, both DD and the inclination angle Θ\Theta follow the analytical predictions. It is worth noting that this benchmark contribute also to the validation of our generalised interface model against the interface response to an external flow.

Wetting dynamics

Model validation

We now analyse the wetting dynamics of coated droplets and soft particles simulated using the interface model discussed in Sec. II. In this kind of experiment, we consider a single particle with initial radius RR and placed close to a flat wall, i.e., its initial position is such that the zz-coordinate of its centre-of-mass ZCMZ_{\mathrm{CM}} is at a distance RR from the wall to let it feel the action of an attractive wall-interface interaction with intensity ϵ\epsilon (see Eq. (21) and Figure 1 for a pictorial view). In our implementation, since we do not have direct control on σsl,σsg\sigma_{\mathrm{sl}},\sigma_{\mathrm{sg}} appearing in Eq. (1), we consider ϵ\epsilon playing the role of an effective solid surface tension as ϵ∝σsl−σsg\epsilon\propto\sigma_{\mathrm{sl}}-\sigma_{\mathrm{sg}}.

Before entering into the details of the wetting dynamics of pre-stressed and non-pre-stressed particles, we validate our implementation by quantitatively investigating the spreading dynamics of a pure droplet, α=αˉ(1,0,0)\alpha= \bar{\alpha}(1, 0, 0), by comparing the time evolution of the radius rr of the contact area with the literature. Indeed, it has been observed that this observable scales in time as

r=Ctδ,(26)r = Ct^\delta, \tag*{(26)}

where both the prefactor CC and the exponent δ\delta can vary. When capillary forces drive the droplet spreading and inertial effects are negligible, Eq. (26) coincides with the Tanner’s law [69], predicting an exponent δ=1/10\delta= 1/10. Contrariwise, when capillary and inertial forces are balanced, it has been observed that the value of the exponent can vary with some factors, such as viscosity [70], surface tension [70], droplet initial shape [71] and wettability [70, 72, 73, 74]. In particular, a value of δ=1/2\delta= 1/2 has been observed in the case of very small contact angles. In Figure 3(a) we report the time evolution of rr, normalised to the initial radius RR at varying equilibrium contact angles θeq\theta_{\mathrm{eq}}. A scaling law following Eq. (26) is observed, with the exponent δ\delta slightly decreases at increasing θeq\theta_{\mathrm{eq}}, in agreement with Ref. [72] (see Figure 3(b)). This implies that in our simulations of wetting dynamics, the inertia is not negligible and it plays a role in resisting the deformation. Note that, as later highlighted in Figure 4(a), our model can capture only cases of large contact angles (88∘≤θeq≤180∘88^\circ\leq\theta_{\mathrm{eq}} \leq180^\circ), indeed δ\delta approaches but does not reach a value of close to 1/21/2 which is characteristic for small contact angles [70, 72, 73, 74]. This is comparable, for example, to the case of a liquid metal droplet, which has been observed to never assume values of θeq<100∘\theta_{\mathrm{eq}} < 100^\circ also in the case of oxidization, suggesting that our model can be used to study such kind of system. Concerning the prefactor CC (Figure 3(c)), it decreases as θeq\theta_{\mathrm{eq}} increases, once again in agreement with Ref. [72]. Note that the “jumps” in rr that are visible in Figure 3(a) for θeq=157∘\theta_{\mathrm{eq}} = 157^\circ originate from the numerical error in measuring very small variations of the contact area.

Image not extracted

Figure 3. Spreading experiment for a pure liquid droplet, α=(α1,0,0)\alpha=(\alpha_1,0,0), on a flat surface. Panel (a): radius of the contact area rr as a function of time tt, where rr and tt are reported normalised to the initial radius RR and the characteristic time t∗=(ρR3/σ)1/2t^*=(\rho R^3/\sigma)^{1/2}, respectively. Different symbols/colours refer to different values of the equilibrium contact angle θeq\theta_{\mathrm{eq}}. In all cases, we observe a scaling law r/R=C(t/t∗)δr/R=C(t/t^*)^\delta. The solid line indicates the scaling with δ=3/10\delta=3/10, while the dotted line refers to scaling δ=3/20\delta=3/20. Panels (b) and (c) show the value of the dimensionless exponent δ\delta and the dimensionless prefactor CC, respectively, as a function of θeq\theta_{\mathrm{eq}}.

Image not extracted

Figure 4. Experiment of wetting dynamics. Panels (a) and (b): Equilibrium contact angle θeq\theta_{\mathrm{eq}} as a function of the wall-particle interaction intensity ϵ\epsilon, normalised to αˉ\bar{\alpha}. Panels (c) and (d): Corresponding values of cos⁡θeq\cos\theta_{\mathrm{eq}}. Left panels ((a) and (c)) refer to pre-stressed particle models, while right panels ((b) and (d)) refer to non-pre-stressed particle models. In all panels, different symbols/colours refer to different models, while dashed lines indicate fitting curves with Eq. (27) (values of fitting parameters are listed in Table I). Data refer to simulations with a number of triangular faces equal to Nf=16820N_f = 16820.

Results

With the aim of simulating the wetting dynamics of droplets with complex interface properties, we explore both pre-stressed (α1>0\alpha_1 > 0) and non-pre-stressed (α1=0\alpha_1 = 0) particles. For each system, characterized by α=(α1,α2,α3)=αˉ(0/1,0/1,0/1)\boldsymbol{\alpha} = (\alpha_1,\alpha_2,\alpha_3) = \bar{\alpha}(0/1,0/1,0/1), we apply the same strategy used for the benchmark, which we summarize again for the sake of clarity. First, we fix the value of αˉ\bar{\alpha} to 10−4lbu10^{-4}l_{\mathrm{bu}}. This choice fixes both the surface tension for pre-stressed particles and the strain modulus for non-pre-stressed particles. Then we measure θeq\theta_{\mathrm{eq}} as a function of the wall-particle interaction energy ϵ\epsilon. In Appendix § 0.5.1 we discuss also a resolution test for the wetting dynamics.

In Figure 4(a) and (b) we report the measured θeq\theta_{\mathrm{eq}} as a function of the ratio ϵ/αˉ\epsilon/\bar{\alpha} for pre-stressed and non-pre-stressed particle models, respectively. In Figure 4(c) and (d) we report for convenience the corresponding values of cos⁡θeq\cos\theta_{\mathrm{eq}}. After the largest value of ϵ/αˉ\epsilon/\bar{\alpha} reported in each plot, numerical instabilities appear in the contact area region; these set the limit of applicability of our approach in terms of contact angles that can be modeled. Concerning pre-stressed particles, pure droplet (circles) appear to be marginally more stable with respect to the choice of ϵ\epsilon than the other particles (softly coated particles, α=αˉ(1,0,1)\boldsymbol{\alpha} = \bar{\alpha}(1,0,1), upward triangles; rigidly coated particles, α=αˉ(1,1,1)\boldsymbol{\alpha} = \bar{\alpha}(1,1,1), pentagons), but can reach only a slightly higher contact angle (θeq=88\theta_{\mathrm{eq}} = 88 vs. θeq=79∘\theta_{\mathrm{eq}} = 79^\circ). Notice that the cases α=αˉ(1,0,1)\boldsymbol{\alpha} = \bar{\alpha}(1,0,1) and α=αˉ(1,1,1)\boldsymbol{\alpha} = \bar{\alpha}(1,1,1) are very similar, meaning that when the system is very rigid, the dilatational contribution given by α2\alpha_2 is not relevant for the equilibrium contact angle θeq\theta_{\mathrm{eq}}. This result is in contrast with what we observed in the shear flow. Non-pre-stressed particle models (Figure 4(b) and (d)) are stable for a more limited range of values of ϵ/αˉ\epsilon/\bar{\alpha}. After around the value of ϵ/αˉ=1.5\epsilon/\bar{\alpha} = 1.5 the contact angle does not drop significantly anymore. Similarly to the pre-stressed case, α2\alpha_2 does not seem to have any influence on θeq\theta_{\mathrm{eq}}. The behaviour of cos⁡θeq\cos\theta_{\mathrm{eq}} as a function of ϵ/αˉ\epsilon/\bar{\alpha} follows very well the empirical behavior

cos⁡θeq≃atan⁡−1[b(ϵ/αˉ−c)]−d(27)\cos\theta_{\mathrm{eq}} \simeq a \tan^{-1}\left[b(\epsilon/\bar{\alpha} - c)\right] - d \tag*{(27)}

where aa, bb, cc and dd are fitting parameters depending on the type of particle (see Table I and dashed lines in Figure 4(c) and (d)). Eq. (27) differs from Eq. (1) because, as mentioned above, the model we present in this work misses the direct control of the wall surface tensions but rather drives the mechanical interaction between the particle and the solid surface. Obviously, the fitting constants represent (unknown) functions of the parameters α1\alpha_1, α2\alpha_2, α3\alpha_3 and ϵ\epsilon. By increasing separately by a factor of ten each of the components of α\boldsymbol{\alpha}, as reported in Figure 5(a), we can understand that the leading order behavior is dictated by α1\alpha_1, while α2\alpha_2 and α3\alpha_3 provide relatively minor changes in θeq\theta_{\mathrm{eq}}. In addition, from Figure 5(b) one can see that α1\alpha_1 must enter in Eq. (27) in the ratio with ϵ\epsilon, because data for the increased α1\alpha_1 (downward triangles) collapses onto the original case (pentagons) once plotted as a function ϵ/α1\epsilon/\alpha_1. The remaining parameters α2\alpha_2 and α3\alpha_3, instead, do not appear to provide a similar scaling, thus meaning that these two parameters may functionally enter in the other fitting parameters of Eq. (27).

Image not extracted

Figure 5. Equilibrium contact angle θeq\theta_{\mathrm{eq}} as a function of the wall-particle interaction intensity ϵ\epsilon, normalised to αˉ\bar{\alpha} (panel (a)) and max⁡(αi)\max(\alpha_i) with i=1,2,3i = 1, 2, 3 (panel (b)). All data refer to the case with all components of α=(α1,α2,α3)\boldsymbol{\alpha} = (\alpha_1,\alpha_2,\alpha_3) turned on, but different symbols/colours refer to a different “extreme” cases, where one of the three parameters is increased by an order of magnitude. The number of triangular faces NfN_f is the same as for the data in Figure 4.

Conclusions

In this work we introduced a novel numerical framework to accurately characterize coated droplets and soft particles. This approach is based on the thory of Barthès-Biesel and Rallison [36] and enables us to capture the unique behavior that is intermediate between that of a pure droplet and a capsule. With this generalised constitutive law we are able to capture the special properties of a wide spectrum of coated droplets, for example, liquid metal droplets surrounded by an oxide layer. In the present approach, the interface strain energy is written in terms of three parameters that play the role of material properties, i.e., the pre-stress (α1\alpha_1), the resistance against area dilatation (α2\alpha_2), and the resistance against shear deformation (α3\alpha_3). With the choice of these three parameters, we explore different types of coated droplets, from pure liquid droplets to soft particles. We validate our methodology with the theoretical predictions and recent experiments in both shear flow and wetting experiments, and we explore the limits of the model in terms of α1,2,3\alpha_{1,2,3}. We plan to enrich this description by including new contributions to the presented model, for example, to mimic the thickness of the oxide layer in the case of liquid metal droplets. The latter could be useful to mimic the dynamics of other complex droplets, such as liquid marbles.

Author declarations

The authors have no conflicts to disclose.

Acknowledgments

This work has received financial support from the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) – Project-ID 431791331 – SFB 1452 “Catalysis at liquid interfaces” and research unit FOR2688 “Instabilities, Bifurcations and Migration in Pulsatile Flows” (Project-ID 417989464). This work was supported by the Italian Ministry of University and Research (MUR) under the FARE programme, project “Smart-HEART”. The authors gratefully acknowledge the Gauss Centre for Supercomputing e.V. (www.gauss-centre.eu) for funding this project by providing computing time through the John von Neumann Institute for Computing (NIC) on the GCS Supercomputer JUWELS ([75]) at Jülich Supercomputing Centre (JSC).

Data Availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Resolution test for wetting dynamics

Results shown in Figure 4 and Figure 5 required a test to choose the best resolution in terms of accuracy and computational effort. In Figure 6 we show the time evolution of the zz-coordinate of its centre-of-mass ZCMZ_{\mathrm{CM}}, normalised by the initial radius RR, for a pure droplet case, α=αˉ(1,0,0)\alpha= \bar{\alpha}(1,0,0), shown in Figure 6(a) and (b), and a rigidly coated droplet, α=αˉ(1,1,1)\alpha= \bar{\alpha}(1,1,1), shown in Figure 6(c) and (d). We report two values of ϵ/αˉ\epsilon/\bar{\alpha}, i.e., 0.05, (Figure 6(a) and (c)), and 7 (Figure 6(b) and (d)), resulting in a large and a small equilibrium contact angle for both systems. As long as the contact angle is very large then all resolutions are equivalent. However, moving towards θeq∼80∘\theta_{\mathrm{eq}} \sim80^\circ, a large number of NfN_f is required for more precise contact angle measurements in the case of a pure droplet. The latter statement follows from the way we compute θeq\theta_{\mathrm{eq}}, i.e., by fitting the droplet shape with a circumference cut by a chord (i.e., the wall). To perform the fitting procedure, we take a slice of the particle mesh involving a number of nodes which is roughly 43πNf\frac{4}{\sqrt{3}}\sqrt{\pi N_f}. Furthermore, simulations with Nf=42320N_f = 42320 show the same dynamics as Nf=16820N_f = 16820 but they require a higher computational cost, leading to the choice made to produce the data reported in Figure 4 and Figure 5, to run simulations with Nf=16820N_f = 16820.

References

  1. [1]P.-G. De Gennes, “Wetting: statics and dynamics,” Reviews of modern physics, vol. 57, p. 827, 1985.DOI
  2. [2]F. Brochard-Wyart and P. De Gennes, “Dynamics of partial wetting,” Advances in colloid and interface science, vol. 39, pp. 1–11, 1992.DOI
  3. [3]D. Bonn, “Wetting transitions,” Current opinion in colloid & interface science, vol. 6, pp. 22–27, 2001.DOI
  4. [4]J. De Coninck, M. J. de Ruijter, and M. Voué, “Dynamics of wetting,” Current opinion in colloid & interface science, vol. 6, pp. 49–53, 2001.DOI
  5. [5]T. Young, “III. an essay on the cohesion of fluids,” Philosophical transactions of the Royal Society of London, vol. 95, pp. 65–87, 1805.DOI
  6. [6]B. Andreotti and J. H. Snoeijer, “Statics and dynamics of soft wetting,” Annual review of fluid mechanics, vol. 52, pp. 285–308, 2020.DOI
  7. [7]J. De Coninck and T. D. Blake, “Wetting and molecular dynamics simulations of simple liquids,” Annu. Rev. Mater. Res., vol. 38, pp. 1–22, 2008.DOI
  8. [8]D. Bonn, J. Eggers, J. Indekeu, J. Meunier, and E. Rolley, “Wetting and spreading,” Reviews of modern physics, vol. 81, no. 2, p. 739, 2009.DOI
  9. [9]J. H. Snoeijer and B. Andreotti, “Moving contact lines: scales, regimes, and dynamical transitions,” Annual review of fluid mechanics, vol. 45, pp. 269–292, 2013.DOI
  10. [10]Taccardi, N., Grabau, M., Debuschewitz, J., Distazo, M., Brandl, M., Hock, R., Maier, F., Papp, C., Erhard, C., & Neiss, C. et . al . (2017). Gallium-rich Pd–Ga phases as supported liquid metal catalysts. Nature Chemistry, 9, 862–867.
  11. [11]A. Hofer, N. Taccardi, M. Moritz, C. Wichmann, S. Hübner, D. Drobek, M. Engelhardt, G. Papastavrou, E. Spiecker, C. Papp, et al., “Preparation of geometrically highly controlled ga particle arrays on quasi-planar nanostructured surfaces as a scalms model system,” RSC advances, vol. 13, pp. 4011–4018, 2023.DOI
  12. [12]Daeneke, T., Khoshmanesh, K., Mahmood, N., De Castro, I. A., Esrafzadeh, D., Barrow, S., Dickey, M., & Kalantar-Zadeh, K. (2018). Liquid metals: fundamentals and applications in chemistry. Chemical Society Reviews, 47, 4073–4111.
  13. [13]F.-M. Allioux, M. B. Ghasemian, W. Xie, A. P. O’Mullane, T. Daeneke, M. D. Dickey, and K. Kalantar-Zadeh, “Applications of liquid metals in nanotechnology,” Nanoscale Horizons, vol. 7, pp. 141–167, 2022.DOI
  14. [14]M. Jeyakumar, M. Hamed, and S. Shankar, “Rheology of liquid metal pastes and alloys,” Journal of non-Newtonian fluid mechanics, vol. 166, pp. 831–838, 2011.DOI
  15. [15]K. Doudrick, S. Liu, E. M. Mutunga, K. L. Klein, V. Damle, K. K. Varanasi, and K. Rykaczewski, “Different shades of oxide: From nanoscale wetting mechanisms to contact printing of gallium-based liquid metals,” Langmuir, vol. 30, pp. 6867–6877, 2014.DOI
  16. [16]I. D. Joshipura, K. A. Persson, V. K. Truong, J.-H. Oh, M. Kong, M. H. Vong, C. Ni, M. Alsafatwi, D. P. Parekh, H. Zhao, et al., “Are contact angle measurements useful for oxide-coated liquid metals?,” Langmuir, vol. 37, pp. 10914–10923, 2021.DOI
  17. [17]L. Zheng, S. Handschuh-Wang, Z. Ye, and B. Wang, “Liquid metal droplets enabled soft robots,” Applied Materials Today, vol. 27, p. 101423, 2022.DOI
  18. [18]E. Bormashenko, “Liquid marbles: properties and applications,” Current Opinion in Colloid & Interface Science, vol. 16, pp. 266–271, 2011.DOI
  19. [19]E. Ni, T. Li, Y. Ruan, Y. Ma, Y. Wang, Y. Jiang, and H. Li, “Modeling of wetting transition of liquid metals on organic liquid surfaces,” Langmuir, vol. 37, pp. 9429–9438, 2021.DOI
  20. [20]Y. Zhao, J. Fang, X. Wang, H. Wang, and T. Lin, “Magnetic liquid marbles: manipulation of liquid droplets using highly hydrophobic fe3o4 nanoparticles,” Advanced materials, vol. 22, pp. 707–710, 2010.DOI
  21. [21]R.-E. Avrămescu, M.-V. Ghica, C. Dinu-Pîrvu, D. I. Udeanu, and L. Popa, “Liquid marbles: From industrial to medical applications,” Molecules, vol. 23, p. 1120, 2018.DOI
  22. [22]Biferale, L., Benzi, R., Sbragaglia, M., Succi, S., & Toschi, F. (2007). Wetting/dewetting transition of two-phase flows in nanocorrugated channels. Journal of Computer-aided Materials Design, 14, 447–456.
  23. [23]Y. Yan and Y. Zu, “A lattice Boltzmann method for incompressible two-phase flows on partial wetting surface with large density ratio,” Journal of Computational Physics, vol. 227, pp. 763–775, 2007.DOI
  24. [25]E. Attar and C. Körner, “Lattice Boltzmann method for dynamic wetting problems,” Journal of colloid and interface science, vol. 335, pp. 84–93, 2009.DOI
  25. [26]Y. Tanaka, Y. Washio, M. Yoshino, and T. Hirata, “Numerical simulation of dynamic behavior of droplet on solid surface by the two-phase lattice Boltzmann method,” Computers & fluids, vol. 40, pp. 68–78, 2011.DOI
  26. [27]J. Hyväluoma, C. Kunert, and J. Harting, “Simulations of slip flow on nanobubble-laden surfaces,” Journal of Physics: Condensed Matter, vol. 23, p. 184106, 2011.DOI
  27. [28]H. P. Jansen, K. Sotthewes, J. van Swigchem, H. J. Zandvliet, and E. S. Kooij, “Lattice Boltzmann modeling of directional wetting: Comparing simulations to experiments,” Physical Review E, vol. 88, p. 013008, 2013.pubmed.ncbi.nlm.nih.gov/23944550
  28. [29]L. Wang and J. Sun, “The application of axisymmetric lattice Boltzmann two-phase model on simulations of liquid film dewetting,” Journal of Applied Physics, vol. 122, p. 085305, 2017.DOI
  29. [30]T. Akai, B. Bijeljic, and M. J. Blunt, “Wetting boundary condition for the color-gradient lattice boltzmann method: Validation with analytical and experimental data,” Advances in Water Resources, vol. 116, pp. 56–66, 2018.DOI
  30. [31]S. Zitz, A. Scagliarini, S. Maddu, A. A. Darhuber, and J. Harting, “Lattice Boltzmann method for thin-liquid-film hydrodynamics,” Phys. Rev. E, vol. 100, p. 033313, Sep 2019.DOI
  31. [32]X. Wang, B. Xu, Y. Wang, and Z. Chen, “Directional migration of single droplet on multi-wetting gradient surface by 3d lattice Boltzmann method,” Computers & Fluids, vol. 198, p. 104392, 2020.DOI
  32. [33]F. Pelusi, M. Sega, and J. Harting, “Liquid film rupture beyond the thin-film equation: A multi-component lattice Boltzmann study,” Physics of Fluids, vol. 34, p. 062109, 2022.DOI
  33. [34]X. Shan and H. Chen, “Lattice Boltzmann model for simulating flows with multiple phases and components,” Physical review E, vol. 47, p. 1815, 1993.pubmed.ncbi.nlm.nih.gov/9960203
  34. [35]M. R. Swift, E. Orlandini, W. Osborn, and J. Yeomans, “Lattice Boltzmann simulations of liquid-gas and binary fluid systems,” Physical Review E, vol. 54, no. 5, p. 5041, 1996.pubmed.ncbi.nlm.nih.gov/9965683
  35. [36]D. Barthès-Biesel and J. Rallison, “The time-dependent deformation of a capsule freely suspended in a linear shear flow,” Journal of Fluid Mechanics, vol. 113, pp. 251–267, 1981.DOI
  36. [37]R. Skalak, A. Tozeren, R. Zarda, and S. Chien, “Strain Energy Function of Red Blood Cell Membranes,” Biophysical Journal, vol. 13, pp. 245–264, 1973.pubmed.ncbi.nlm.nih.gov/4697236
  37. [38]D. Matsunaga, Y. Imai, T. Yamaguchi, and T. Ishikawa, “Rheology of a dense suspension of spherical capsules under simple shear flow,” Journal of Fluid Mechanics, vol. 786, pp. 110–127, 2016.DOI
  38. [39]D. Barthes-Biesel and H. Sgaier, “Role of membrane viscosity in the orientation and deformation of a spherical capsule suspended in shear flow,” Journal of Fluid Mechanics, vol. 160, pp. 119–135, 1985.DOI
  39. [40]Benzi, R., Succi, S., & Vergassola, M. (1992). The lattice Boltzmann equation: theory and applications. Physics Reports, 222, 145–197.
  40. [42]Z. Guo, C. Zheng, and B. Shi, “Discrete lattice effects on the forcing term in the lattice Boltzmann method,” Physical Review E, vol. 65, p. 046308, 2002.pubmed.ncbi.nlm.nih.gov/12006014
  41. [43]C. S. Peskin, “The immersed boundary method,” Acta numerica, vol. 11, pp. 479–517, 2002.DOI
  42. [44]B. Kaoui, J. Harting, and C. Misbah, “Two-dimensional vesicle dynamics under shear flow: effect of confinement,” Physical Review E, vol. 83, p. 066319, 2011.pubmed.ncbi.nlm.nih.gov/21797489
  43. [45]T. Krüger, F. Varnik, and D. Raabe, “Efficient and accurate simulations of deformable particles immersed in a fluid using a combined immersed boundary lattice Boltzmann finite element method,” Computers & Mathematics with Applications, vol. 61, pp. 3485–3505, 2011.DOI
  44. [46]O. Aouane, A. Scagliarini, and J. Harting, “Structure and rheology of suspensions of spherical strain-hardening capsules,” Journal of Fluid Mechanics, vol. 911, p. A11, Mar. 2021.DOI
  45. [47]C. Bielinski, O. Aouane, J. Harting, and B. Kaoui, “Squeezing multiple soft particles into a constriction: transition to clogging,” Physical Review E, vol. 104, p. 065101, 2021.DOI
  46. [48]T. Krüger, B. Kaoui, and J. Harting, “Interplay of inertia and deformability on rheological properties of a suspension of capsules,” The Journal of Fluid Mechanics, vol. 751, pp. 725–745, 2014.DOI
  47. [49]F. Guglietta, F. Pelusi, M. Sega, O. Aouane, and J. Harting, “Suspensions of viscoelastic capsules: effect of membrane viscosity on transient dynamics,” arXiv preprint arXiv:2302.03546, 2023.arxiv.org/abs/2302.03546
  48. [50]T. Krüger, D. Holmes, and P. V. Coveney, “Deformability-based red blood cell separation in deterministic lateral displacement devices—A simulation study,” Biomicrofluidics, vol. 8, p. 054114, Oct. 2014.DOI
  49. [51]Kaoui, B., Krüger, T., & Harting, J. (2012). How does confinement affect the dynamics of viscous vesicles and red blood cells?. Soft Matter, 8, 9246.
  50. [52]P. Li and J. Z. Zhang, “Similar but Distinct Roles of Membrane and Interior Fluid Viscosities in Capsule Dynamics in Shear Flows,” Cardiovascular Engineering and Technology, vol. 12, pp. 232–249, Apr. 2021.DOI
  51. [53]Guglietta, F., Behr, M., Biferale, L., Falucci, G., & Sbragaglia, M. (2021). Lattice Boltzmann simulations on the tumbling to tank-treading transition: effects of membrane viscosity. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, 379, 20200395.
  52. [54]F. Guglietta, M. Behr, G. Falcucci, and M. Sbragaglia, “Loading and relaxation dynamics of a red blood cell,” Soft Matter, vol. 17, pp. 5978–5990, 2021.DOI
  53. [55]P. Li and J. Zhang, “A finite difference method with subsampling for immersed boundary simulations of the capsule dynamics with viscoelastic membranes,” International Journal for Numerical Methods in Biomedical Engineering, p. e3200, Apr. 2019.DOI
  54. [56]F. Guglietta, M. Behr, L. Biferale, G. Falcucci, and M. Sbragaglia, “On the effects of membrane viscosity on transient red blood cell dynamics,” Soft Matter, vol. 16, pp. 6191–6205, 2020.DOI
  55. [57]Taglienti, D., Guglietta, F., & Sbragaglia, M. (2023). Reduced model for droplet dynamics in shear flows at finite capillary numbers. Physical Review Fluids, 8, 013603.
  56. [59][59–61]
  57. [62][62–66]
  58. [67]J. Gounley, G. Boedec, M. Jaeger, and M. Leonetti, “Influence of surface viscosity on droplets in shear flow,” Journal of Fluid Mechanics, vol. 791, p. 464–494, 2016.DOI
  59. [68]C. E. Chaffey and H. Brenner, “A second-order theory for shear deformation of drops,” Journal of Colloid and Interface Science, vol. 24, pp. 258–269, 1967.DOI
  60. [69]L. Tanner, “The spreading of silicone oil drops on horizontal surfaces,” Journal of Physics D: Applied Physics, vol. 12, p. 1473, 1979.DOI
  61. [70]D. Legendre and M. Maglio, “Numerical simulation of spreading drops,” Colloids and Surfaces A: Physicochemical and Engineering Aspects, vol. 432, pp. 29–37, 2013.DOI
  62. [71]L. Courbin, J. C. Bird, M. Reyssat, and H. A. Stone, “Dynamics of wetting: from inertial spreading to viscous imbibition,” Journal of Physics: Condensed Matter, vol. 21, p. 464127, 2009.DOI
  63. [72]J. C. Bird, S. Mandre, and H. A. Stone, “Short-time dynamics of partial wetting,” Physical review letters, vol. 100, p. 234501, 2008.pubmed.ncbi.nlm.nih.gov/18643505
  64. [73]A. Carlson, M. Do-Quang, and G. Amberg, “Dissipation in rapid dynamic wetting,” Journal of Fluid Mechanics, vol. 682, pp. 213–240, 2011.DOI
  65. [74]K. G. Winkels, J. H. Weijs, A. Eddi, and J. H. Snoeijer, “Initial spreading of low-viscosity drops on partially wetting surfaces,” Physical Review E, vol. 85, p. 055301, 2012.pubmed.ncbi.nlm.nih.gov/23004813
  66. [75]Centre, J. S. (2019). JUWELS: Modular Tier-0/1 Supercomputer at the Jülich Supercomputing Centre. Journal of Large-scale Research Facilities, 5, A135.

Paper details

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

views · 9

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