betaFeedback

WorksPhysics and Astronomy

Electrohydrodynamics of a pair of leaky dielectric droplets on the solid substrate: A lattice Boltzmann study

Abstract

In this work, the electrohydrodynamics of a pair of leaky dielectric droplets on a solid substrate is investigated by the phase-field-based lattice Boltzmann method. Different from a pair of suspended droplets that may coalesce or separate, two leaky dielectric droplets on the substrate exhibit more complex modes due to the effects of wettability and electric force. The results show that when a horizontal electric field is applied, five different modes with electrostatic attractive force are observed, including attraction without coalescence, attraction with coalescence, coalescence with bubble entrapment, coalescence followed by suspension, and suspension followed by coalescence. Particularly, if the droplets are in a hydrophilic state, the coalescence mode is usually observed, while for droplets in a neutral or hydrophobic state, the permittivity ratio has an important effect on the droplet modes. Additionally, during the coalescence process, two droplets in a hydrophobic state not only capture bubbles, but may also exhibit suspension at a large permittivity ratio or contact angle. On the other hand, when a vertical electric field is applied, there are three different modes with repulsive electrostatic force, including non-coalescence, coalescence, and suspension followed by repulsion. Specially, a small permittivity ratio or a large contact angle can suppress the horizontal deformation of droplets, preventing their coalescence. Moreover, under superhydrophobic conditions, both horizontal and vertical electric fields suspend the droplets. However, the vertical electric field induces repulsion between the suspended droplets, driving them apart, whereas the horizontal electric field promotes their coalescence.

Introduction

Controlling droplet coalescence is a significant challenge in droplet microfluidics. On one hand, to facilitate chemical reactions through mixing, promoting droplet coalescence is essential for merging specific droplets [1]. On the other hand, to manage a large number of droplets, it is desirable to inhibit coalescence to maintain the intact and distinct partitioning of each droplet [2]. As one of the manipulation technologies, the adoption of an external electric field can be used to control droplet deformation, breakup, coalescence, or repulsion, thereby fulfilling the stringent demands of both scientific research and industrial applications [3].

In recent years, many theoretical, experimental and numerical studies have been performed to investigate the interaction of two droplets under an external electric field [4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]. Sozou [4] first theoretically studied a droplet pair aligned along the electric field direction, where the leaky dielectric droplets are suspended in a leaky dielectric fluid (both droplets and fluid have finite conductivities). It was found that when the distance between the centers of two droplets is reduced to the order of twice droplet diameter, the tangential electric stress at the interfaces of droplets and shear stress induced by the flow field are significantly affected by each other. Then, Atten [5] estimated the coalescence rate and the characteristic coalescence time of two emulsion droplets via considering the interaction of dipoles induced by the electric field. Moreover, Atten et al. [6] theoretically derived an approximate correlation between the critical coalescence distance and electric field strength when investigating electrocoalescence conditions of two droplets suspended in an insulating oil. In addition, by taking into account both hydrodynamic and electrostatic interactions, Zhang et al. [7] utilized the trajectory analysis to track the relative motion of two perfectly conducting droplets, and predicted their collision and coalescence rates. However, due to the perfect dielectric assumption, these theoretical results cannot predict the dynamic behavior of leaky dielectric droplets. To overcome the limitations of theoretical approaches, experimental investigations have been undertaken to explore the complex behaviors of droplet pairs under electric fields. For instance, Dong et al. [8] studied the effect of electric field on the coalescence of two droplets suspended in the organic fluid, and demonstrated that the reduced surface tension promotes droplet deformation, facilitating the contact coalescence of two droplets. They also observed that coalescence time decreases with increasing electric field strength. Bird et al. [9] observed that the leading edges of two close droplets can deform into cones due to the electrical stress acting on the droplet interfaces, and whether they coalesce after contact depends on the cone angle. Recently, Huang et al. [10] studied the dynamic behaviors of two free droplets suspended in low-viscosity oil under a uniform electric field. They found that as the electric field strength increases, the two droplets undergo a transition from coalescence to partial coalescence and then to non-coalescence, and the critical value of electric field strength for droplet coalescence is related to the droplet radius and surface tension. Furthermore, Eow and Ghadiri [11] experimentally investigated the interaction between a pair of suspended droplets, and found that the angle α\alpha between the electric field direction and the line connecting the centers of two droplets has an influence on the attractive or repulsive interaction between the two droplets. Although experimental methods offer a distinct advantage in directly observing and measuring the electrohydrodynamics of a droplet pair under an external electric field, they also suffer from some significant limitations in providing a complete and detailed description of the characteristics of flow and electric fields. Consequently, with advances in computational science, numerical approaches have emerged as powerful alternatives. Early numerical work by Baygents et al. [12] studied the interaction between the leaky dielectric droplets aligned in the direction of electric field. They found that there are some circulation flows driven by the electric stress on the droplet surface, whether the droplets attract or repel each other. Later, Dong et al. [13] investigated the dynamic behavior of the leaky dielectric droplets by using the lattice Boltzmann (LB) method, and the results show that when the permittivity ratio SS between the droplet and the surrounding fluid is much smaller than the conductivity ratio RR, the droplets repel each other. This repulsion is primarily due to the dominant repulsive interaction under a small SS. On the contrary, when SS is large but still smaller than RR, the droplets can attract each other through the attractive electrostatic interaction. For the case where SS is larger than RR, the droplets only attract each other, and then coalesce, resulting from the combined effects of both hydrodynamic and electrostatic interactions. It is worth noting that an estimator to measure the relative strength of electrostatic force to hydrodynamic force has been proposed [14], which indicates that the interaction between two suspended droplets cannot be determined only by the sign of (R−S)(R-S), but also by the relative strength of above two forces and the inertial effect characterized by the Ohnesorge number (OhOh). Here, OhOh also represents the ratio of the viscous capillary to Rayleigh inertial time scale, serving as a metric to assess the influence of viscosity on interfacial dynamics. Furthermore, the inclination angle α\alpha also has a significant effect on the electrohydrodynamics of two leaky dielectric droplets through the dielectrophoretic (DEP) and electrohydrodynamic (EHD) interactions [15]. Specifically, the DEP interaction exerts either an attractive or repulsive influence that is contingent upon the angle α\alpha, whereas the EHD interaction is jointly determined by the angle α\alpha and the sign of (R−S)(R-S).

The aforementioned works have indeed promoted the understanding of the interaction between a pair of droplets under an external electric field. However, when droplets are manipulated under an electric field, they are inevitably in contact with the solid substrate. In this case, the wettability of the solid substrate would affect the equilibrium shapes and dynamic behaviors of the droplets, but this effect has not been comprehensively studied. To fill the gap, in this work we will conduct a numerical study on the electrohydrodynamics of a droplet pair on the solid substrate. For this purpose, the phase-field lattice Boltzmann (LB) method [16] is adopted here owing to its distinct features in the simulation of multiphase EHD flows [17, 18, 19, 20], including a clear physical picture, easy implementation of boundary conditions, and fully parallel algorithm [21, 22, 23]. The rest of this paper is organized as follows. In section II, we first present the mathematical model for the physical problem of a pair of leaky dielectric droplets on the solid substrate, followed by the phase-field LB method. Then in section III, the LB method is tested by two classical multiphase examples. In section IV, the effects of the permittivity ratio SS and wettability on the droplet pair are investigated when the electric field is applied in horizontal or vertical direction. Finally, some conclusions are given in § 0.5.

Mathematical model and numerical method

The consistent and conservative phase-field-based model for two-phase EHD flows

In this work, we only consider the electrohydrodynamics of two droplets that have no net charge but finite electrical conductivity. In the framework of the leaky dielectric theory, a conservative Allen-Cahn equation is used to capture the fluid-fluid interface, the incompressible Navier-Stokes equations are adopted to describe the fluid motion, and the Poisson equation is applied to depict the distribution of electric potential [17, 24]:

∂C∂t+∇⋅(Cu)=∇⋅[M(∇C−λn)],(1)\frac{\partial C}{\partial t}+\nabla\cdot(C\mathbf{u})=\nabla\cdot\left[M(\nabla C-\lambda\mathbf{n})\right], \tag*{(1)}
∇⋅u=0,(2)\nabla\cdot\mathbf{u}=0, \tag*{(2)}
∂(ρu)∂t+∇⋅(ρuu+mCu)=−∇p+∇⋅μ[∇u+(∇u)T]+F,(3)\frac{\partial(\rho\mathbf{u})}{\partial t}+\nabla\cdot(\rho\mathbf{u}\mathbf{u}+\mathbf{m}C\mathbf{u})=-\nabla p+\nabla\cdot\mu\left[\nabla\mathbf{u}+(\nabla\mathbf{u})^{\mathrm{T}}\right]+\mathbf{F}, \tag*{(3)}
∇⋅(σ∇ϕ)=0(4)\nabla\cdot(\sigma\nabla\phi) = 0 \tag*{(4)}

where CC is the order parameter used to label the liquid phase (marked as ll) with Cl=1C_l = 1, and gas or another liquid phase (marked as gg) with Cg=0C_g = 0. u\mathbf{u} is the fluid velocity, MM is the mobility, n=∇C/∣∇C∣\mathbf{n} = \nabla C/\lvert\nabla C \rvert is the unit vector normal to interface, and λ=2β/κ(Cl−C)(C−Cg)\lambda= \sqrt{2\beta/\kappa}(C_l-C)(C-C_g) with β=12γ/[W(Cl−Cg)4]\beta= 12\gamma/[W(C_l-C_g)^4] and κ=3γW/[2(Cl−Cg)2]\kappa= 3\gamma W/[2(C_l-C_g)^2], β\beta and κ\kappa are two physical parameters related to the interface thickness WW and the surface tension coefficient γ\gamma. The mass diffusion mC\mathbf{m}^C between different phases can be expressed as

mC=−ρl−ρgCl−CgM(∇C−λn),(5)\mathbf{m}^C = -\frac{\rho_l-\rho_g}{C_l-C_g}M(\nabla C-\lambda\mathbf{n}), \tag*{(5)}

which can be used to preserve the mass and momentum consistency [22]. ρ\rho is the density, pp is the pressure, μ\mu is the dynamic viscosity, and F\mathbf{F} is the total external force. σ\sigma is the conductivity, ϕ\phi is the electrical potential. Under the assumption of ∇×E=0\nabla\times\mathbf{E} = 0, the electric field strength E\mathbf{E} can be expressed as a gradient of the electrical potential, i.e., E=−∇ϕ\mathbf{E} = -\nabla\phi.

When the gravity is neglected, the total external force F\mathbf{F} is composed of the surface tension force FS\mathbf{F}_S and electric stress FE\mathbf{F}_E, which can be given by

FS=μC∇C,FE=−12E⋅ρe∇ε+ρeE,(6)\mathbf{F}_S = \mu_C\nabla C,\qquad\mathbf{F}_E = -\frac{1}{2}\mathbf{E}\cdot\rho_e\nabla\varepsilon+\rho_e\mathbf{E}, \tag*{(6)}

where μC\mu_C is the chemical potential defined by μC=4β(C−Cl)(C−Cg)(C−Cl+Cg2)−κ∇2C\mu_C = 4\beta(C-C_l)(C-C_g)(C-\frac{C_l+C_g}{2})-\kappa\nabla^2C, ε\varepsilon is the permittivity, and ρe=∇⋅(εE)\rho_e = \nabla\cdot(\varepsilon\mathbf{E}) is the volume charge density. In order to simplify and improve the accuracy in computing electric stress FE\mathbf{F}_E, we can rewrite the volumetric charge density ρe\rho_e as

ρe=∇⋅(εE)=ε∇⋅E+∇ε⋅E=−εσ∇σ⋅E+∇ε⋅E,(7)\rho_e = \nabla\cdot(\varepsilon\mathbf{E}) = \varepsilon\nabla\cdot\mathbf{E}+\nabla\varepsilon\cdot\mathbf{E} = -\frac{\varepsilon}{\sigma}\nabla\sigma\cdot\mathbf{E}+\nabla\varepsilon\cdot\mathbf{E}, \tag*{(7)}

where Eq. (4) has been used.

To ensure that the distribution of each physical quantity or parameter χ\chi in the two-phase system is continuous, the following relation is adopted [25],

χ=χl−χgCl−Cg(C−Cg)+χg,(8)\chi= \frac{\chi_l-\chi_g}{C_l-C_g}(C-C_g)+\chi_g, \tag*{(8)}

where χ\chi can be the fluid density ρ\rho, the dynamic viscosity μ\mu, the conductivity σ\sigma or the permittivity ε\varepsilon.

Boundary conditions

As shown in Figure 1, when the droplets contact with the solid substrate, the wettability must be considered. In this work, the following wetting boundary condition is used [26, 27],

Figure 1. Schematic of a droplet pair placed on the solid substrate under an external electric field.

nw⋅∇C=−2βκ(Cl−C)(C−Cg)cos⁡θ,(9)\mathbf{n}_{w}\cdot\nabla C=-\sqrt{\frac{2\beta}{\kappa}}(C_l-C)(C-C_g)\cos\theta, \tag*{(9)}

where nw\mathbf{n}_{w} denotes the unit vector normal to the solid substrate, and θ\theta represents the contact angle. In addition, for the electric and flow fields, the Dirichlet boundary conditions and the no-slip boundary conditions are imposed on the electrode plates, respectively.

The consistent and conservative phase-field-based LB method

LB models for the Allen-Cahn equation, Navier-Stokes equations and Poisson equation

Compared to the popular and simple single-relaxation-time LB model [28], here we consider the multiple-relaxation-time (MRT) LB model for its better numerical accuracy and stability [29]. The evolution equations of the MRT-LB model for Allen-Cahn equation Eq. (1), Navier-Stokes equations Eq. (2) and Poisson equation Eq. (4) can be given by [22]

fi(x+ciΔt,t+Δt)−fi(x,t)=−Λijf[fj(x,t)−fjeq(x,t)]+Δt(δij−Λijf2)Fj(x,t),(10)f_i(\mathbf{x}+\mathbf{c}_i\Delta t,t+\Delta t)-f_i(\mathbf{x},t)=-\Lambda_{ij}^{f}\left[f_j(\mathbf{x},t)-f_j^{\mathrm{eq}}(\mathbf{x},t)\right]+\Delta t\left(\delta_{ij}-\frac{\Lambda_{ij}^{f}}{2}\right)F_j(\mathbf{x},t), \tag*{(10)}
gi(x+ciΔt,t+Δt)−gi(x,t)=−Λijg[gj(x,t)−gjeq(x,t)]+Δt(δij−Λijg2)Gj(x,t),(11)g_i(\mathbf{x}+\mathbf{c}_i\Delta t,t+\Delta t)-g_i(\mathbf{x},t)=-\Lambda_{ij}^{g}\left[g_j(\mathbf{x},t)-g_j^{\mathrm{eq}}(\mathbf{x},t)\right]+\Delta t\left(\delta_{ij}-\frac{\Lambda_{ij}^{g}}{2}\right)G_j(\mathbf{x},t), \tag*{(11)}
hi(x+ciΔt,t′+Δt)−hi(x,t′)=−Λijh[hj(x,t′)−hjeq(x,t′)],(12)h_i(\mathbf{x}+\mathbf{c}_i\Delta t,t^{\prime}+\Delta t)-h_i(\mathbf{x},t^{\prime})=-\Lambda_{ij}^{h}\left[h_j(\mathbf{x},t^{\prime})-h_j^{\mathrm{eq}}(\mathbf{x},t^{\prime})\right], \tag*{(12)}

where fi(x,t)f_i(\mathbf{x},t) and gi(x,t)g_i(\mathbf{x},t) are the distribution functions for phase and flow fields at position x\mathbf{x} and time tt, hi(x,t′)h_i(\mathbf{x},t') is the distribution function for electric field at position x\mathbf{x} and pseudo time t′t'. ci\mathbf{c}_i is the discrete velocity, Λijf=(Mf−1SfMf)ij\Lambda_{ij}^{f}=(\mathbf{M}_f^{-1}\mathbf{S}_f\mathbf{M}_f)_{ij}, Λijg=(Mg−1SgMg)ij\Lambda_{ij}^{g}=(\mathbf{M}_g^{-1}\mathbf{S}_g\mathbf{M}_g)_{ij} and Λijh=(Mh−1ShMh)ij\Lambda_{ij}^{h}=(\mathbf{M}_h^{-1}\mathbf{S}_h\mathbf{M}_h)_{ij} are the collision matrices. Here Mf\mathbf{M}_f, Mg\mathbf{M}_g and Mh\mathbf{M}_h are the orthogonal transformation matrices, Sf\mathbf{S}_f, Sg\mathbf{S}_g and Sh\mathbf{S}_h are the diagonal relaxation matrices. To recover the Allen-Cahn equation, Navier-Stokes equations and Poisson equation correctly, the equilibrium distribution functions fieq(x,t)f_i^{eq}(\mathbf{x},t), gieq(x,t)g_i^{eq}(\mathbf{x},t) and hieq(x,t′)h_i^{eq}(\mathbf{x},t') can be designed as follows [30, 31, 32, 33]

fieq(x,t)=ωiC(1+ci⋅ucs2),(13)f_i^{eq}(\mathbf{x},t)=\omega_i C\left(1+\frac{\mathbf{c}_i\cdot\mathbf{u}}{c_s^2}\right), \tag*{(13)}
gieq(x,t)={ρ0+(ω0−1)pcs2+s0(ρ,C,u),i=0,ωipcs2+si(ρ,C,u),i≠0,(14)g_i^{eq}(\mathbf{x},t)= \begin{cases} \rho_0+(\omega_0-1)\frac{p}{c_s^2}+s_0(\rho,C,\mathbf{u}), & i=0,\\ \omega_i\frac{p}{c_s^2}+s_i(\rho,C,\mathbf{u}), & i\ne0, \end{cases} \tag*{(14)}
hieq(x,t′)={(ω0−1)ϕ(x,t′),i=0,ωiϕ(x,t′),i≠0,(15)h_i^{eq}(\mathbf{x},t')= \begin{cases} (\omega_0-1)\phi(\mathbf{x},t'), & i=0,\\ \omega_i\phi(\mathbf{x},t'), & i\ne0, \end{cases} \tag*{(15)}

where ρ0=1\rho_0=1, csc_s is the sound speed, ωi\omega_i is the weight coefficient, and si(ρ,C,u)s_i(\rho,C,\mathbf{u}) can be given by

si(ρ,C,u)=ωi[ci⋅(ρu)cs2+(ρuu+mCu):(cici−cs2I)2cs4].(16)s_i(\rho,C,\mathbf{u})=\omega_i\left[\frac{\mathbf{c}_i\cdot(\rho\mathbf{u})}{c_s^2}+\frac{(\rho\mathbf{u}\mathbf{u}+\mathbf{m}^{C}\mathbf{u}):(\mathbf{c}_i\mathbf{c}_i-c_s^2\mathbf{I})}{2c_s^4}\right]. \tag*{(16)}

The source and force distribution functions Fi(x,t)F_i(\mathbf{x},t) and Gi(x,t)G_i(\mathbf{x},t) can be expressed as [22, 30, 31]

Fi(x,t)=ωici⋅[∂t(Cu)+cs2λn]cs2,(17)F_i(\mathbf{x},t)=\frac{\omega_i\mathbf{c}_i\cdot[\partial_t(C\mathbf{u})+c_s^2\lambda\mathbf{n}]}{c_s^2}, \tag*{(17)}
Gi(x,t)=ωi[u⋅∇ρ+ci⋅(F+Fc)cs2+Q:(cici−cs2I)2cs4],(18)G_i(\mathbf{x},t)=\omega_i\left[\mathbf{u}\cdot\nabla\rho+\frac{\mathbf{c}_i\cdot(\mathbf{F}+\mathbf{F}_c)}{c_s^2}+\frac{\mathbf{Q}:(\mathbf{c}_i\mathbf{c}_i-c_s^2\mathbf{I})}{2c_s^4}\right], \tag*{(18)}

where Fc=−∇⋅(mCu−umC)/2\mathbf{F}_c=-\nabla\cdot(\mathbf{m}^{C}\mathbf{u}-\mathbf{u}\mathbf{m}^{C})/2 and Q=(F+Fc)u+u(F+Fc)+∂t(mCu+umC)/2+cs2(u∇ρ+∇ρu)\mathbf{Q}=(\mathbf{F}+\mathbf{F}_c)\mathbf{u}+\mathbf{u}(\mathbf{F}+\mathbf{F}_c)+\partial_t(\mathbf{m}^{C}\mathbf{u}+\mathbf{u}\mathbf{m}^{C})/2+c_s^2(\mathbf{u}\nabla\rho+\nabla\rho\mathbf{u}).

In addition, the order parameter CC, fluid velocity u\mathbf{u}, the hydrodynamic pressure pp and electrical potential ϕ\phi are calculated by [22, 33]

C=∑ifi,(19)C=\sum_i f_i, \tag*{(19)}
ρu=∑icigi+Δt2(F+Fc),(20)\rho\mathbf{u}=\sum_i\mathbf{c}_i g_i+\frac{\Delta t}{2}(\mathbf{F}+\mathbf{F}_c), \tag*{(20)}
p=cs21−ω0{∑i≠0gi+12Δtu⋅∇ρ+s0(ρ,C,u)+13cs2Δt[2(F+Fc)⋅u+∂t(mC⋅u)]}.(21)p=\frac{c_s^2}{1-\omega_0}\left\{\sum_{i\ne0}g_i+\frac{1}{2}\Delta t\mathbf{u}\cdot\nabla\rho+s_0(\rho,C,\mathbf{u})+\frac{1}{3c_s^2}\Delta t[2(\mathbf{F}+\mathbf{F}_c)\cdot\mathbf{u}+\partial_t(\mathbf{m}^{C}\cdot\mathbf{u})]\right\}. \tag*{(21)}
ϕ=∑i≠011−ω0hi.(22)\phi= \sum_{i\ne0} \frac{1}{1-\omega_0}h_i. \tag*{(22)}

It should be noted that, to calculate the gradient and Laplacian terms in the LB models, the following second-order isotropic central schemes are used,

∇ξ(x)=∑i≠0ωiciξ(x+ciΔt)cs2Δt,(23)\nabla\xi(\mathbf{x}) = \sum_{i\ne0} \frac{\omega_i \mathbf{c}_i \xi(\mathbf{x}+\mathbf{c}_i\Delta t)}{c_s^2\Delta t}, \tag*{(23)}
∇2ξ(x)=∑i≠02ωi[ξ(x+ciΔt)−ξ(x)]cs2Δt2.(24)\nabla^2 \xi(\mathbf{x}) = \sum_{i\ne0} \frac{2\omega_i[\xi(\mathbf{x}+\mathbf{c}_i\Delta t)-\xi(\mathbf{x})]}{c_s^2\Delta t^2}. \tag*{(24)}
Discrete velocity model

In this study, we only consider two-dimensional problems, and the nine-velocity (D2Q9) lattice model [28] is adopted for both phase and flow fields, while for simplicity, the five-velocity (D2Q5) model [17] is applied for the electric field. In the D2Q9 model, the discrete velocity ci\mathbf{c}_i, weight coefficient ωi\omega_i, orthogonal transformation matrix Mf=Mg\mathbf{M}_f=\mathbf{M}_g, and diagonal relaxation matrices Sf\mathbf{S}_f and Sg\mathbf{S}_g are given by

ci=(010−101−1−110010−111−1−1)c,(25)\mathbf{c}_i = \begin{pmatrix} 0 & 1 & 0 & -1 & 0 & 1 & -1 & -1 & 1 \\ 0 & 0 & 1 & 0 & -1 & 1 & 1 & -1 & -1 \end{pmatrix}\mathbf{c}, \tag*{(25)}
ωi=(49,19,19,19,19,136,136,136,136),(26)\omega_i = \left(\frac{4}{9},\frac{1}{9},\frac{1}{9},\frac{1}{9},\frac{1}{9},\frac{1}{36},\frac{1}{36},\frac{1}{36},\frac{1}{36}\right), \tag*{(26)}
Mf=Mg=(111111111−4−1−1−1−122224−2−2−2−21111010−101−1−110−20201−1−110010−111−1−100−20211−1−101−11−10000000001−11−1),(27)\mathbf{M}_f=\mathbf{M}_g= \begin{pmatrix} 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 \\ -4 & -1 & -1 & -1 & -1 & 2 & 2 & 2 & 2 \\ 4 & -2 & -2 & -2 & -2 & 1 & 1 & 1 & 1 \\ 0 & 1 & 0 & -1 & 0 & 1 & -1 & -1 & 1 \\ 0 & -2 & 0 & 2 & 0 & 1 & -1 & -1 & 1 \\ 0 & 0 & 1 & 0 & -1 & 1 & 1 & -1 & -1 \\ 0 & 0 & -2 & 0 & 2 & 1 & 1 & -1 & -1 \\ 0 & 1 & -1 & 1 & -1 & 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & 0 & 0 & 1 & -1 & 1 & -1 \end{pmatrix}, \tag*{(27)}
Sf=diag⁡(s0f,s1f,s2f,s3f,s4f,s5f,s6f,s7f,s8f),(28)\mathbf{S}_f=\operatorname{diag}\left(s_0^f,s_1^f,s_2^f,s_3^f,s_4^f,s_5^f,s_6^f,s_7^f,s_8^f\right), \tag*{(28)}
Sg=diag⁡(s0g,s1g,s2g,s3g,s4g,s5g,s6g,s7g,s8g),(29)\mathbf{S}_g=\operatorname{diag}\left(s_0^g,s_1^g,s_2^g,s_3^g,s_4^g,s_5^g,s_6^g,s_7^g,s_8^g\right), \tag*{(29)}

where c=Δx/Δtc = \Delta x/\Delta t is the lattice speed with Δx\Delta x and Δt\Delta t being the lattice spacing and time step, the relaxation parameters s3fs^{f}_{3} and s5fs^{f}_{5} are related to the mobility, and are given by s3f=s5f=1/(M/cs2Δt+0.5)s^{f}_{3} = s^{f}_{5} = 1/(M/c_{s}^{2}\Delta t + 0.5). Similarly, s7gs^{g}_{7} and s8gs^{g}_{8} are related to the kinematic viscosity ν=μ/ρ\nu= \mu/\rho, and can be determined by s7g=s8g=1/(ν/cs2Δt+0.5)s^{g}_{7} = s^{g}_{8} = 1/(\nu/c_{s}^{2}\Delta t + 0.5), cs=c/3c_{s} = c/\sqrt{3} is the sound speed. Additionally, the other relaxation parameters are set as 1 if not specified. In D2Q5 model, the discrete velocity ci\mathbf{c}_{i} and weight coefficient ωi\omega_{i}, orthogonal transformation matrix Mh\mathbf{M}_{h}, and diagonal relaxation matrix Sh\mathbf{S}_{h} can be given as

ci=(010−100010−1)c,(30)\mathbf{c}_{i} = \begin{pmatrix} 0 & 1 & 0 & -1 & 0 \\ 0 & 0 & 1 & 0 & -1 \end{pmatrix} c, \tag*{(30)}
ωi=(13 16 16 16 16),(31)\omega_{i} = \left(\frac{1}{3}\ \frac{1}{6}\ \frac{1}{6}\ \frac{1}{6}\ \frac{1}{6}\right), \tag*{(31)}
Mh=(11111010−100010−101−11−1−41111),(32)\mathbf{M}_{h} = \begin{pmatrix} 1 & 1 & 1 & 1 & 1 \\ 0 & 1 & 0 & -1 & 0 \\ 0 & 0 & 1 & 0 & -1 \\ 0 & 1 & -1 & 1 & -1 \\ -4 & 1 & 1 & 1 & 1 \end{pmatrix}, \tag*{(32)}
Sh=diag⁡(s0h,s1h,s2h,s3h,s4h),(33)\mathbf{S}_{h} = \operatorname{diag}\left(s^{h}_{0},s^{h}_{1},s^{h}_{2},s^{h}_{3},s^{h}_{4}\right), \tag*{(33)}

where the relaxation parameters s1hs^{h}_{1} and s2hs^{h}_{2} are related to the conductivity through the relation s1h=s2h=1/(σ/cs2Δt+0.5)s^{h}_{1} = s^{h}_{2} = 1/(\sigma/c_{s}^{2}\Delta t + 0.5), and other relaxation parameters are fixed to be 1.

Numerical validations

In this section, we validate the LB method by two benchmark problems, and the half-way bounce-back scheme [34, 35] is adopted to treat the no-slip and no-flux boundary conditions, while the general bounce-back scheme [36] is adopted to treat the Dirichlet boundary condition.

The droplet spreading on an ideal wall

We first consider the spreading of a droplet on an ideal wall to test the capability of the LB method in predicting the contact angle. For this problem, the density ratio (ρl/ρg)(\rho_{l}/\rho_{g}) and dynamic viscosity ratio (μl/μg)(\mu_{l}/\mu_{g}) are fixed to be 1000 and 100, the computational domain is [0,200]×[0,100][0,200]\times[0,100], in which a semicircular droplet with the radius R0=35R_0 = 35 is deposited on the bottom wall. The other parameters in numerical simulations are set as Δx=Δt=1\Delta x = \Delta t = 1, W=5W = 5, ρg=1\rho_g = 1, vg=0.1v_g = 0.1, γ=0.2\gamma= 0.2 and M=0.1M = 0.1, the initial distribution of the order parameter is given by

C(x,y)=Cl+Cg2+Cl−Cg2tanh⁡2(R0−(x−100)2+y2)W.(34)C(x,y) = \frac{C_l + C_g}{2} + \frac{C_l - C_g}{2}\tanh\frac{2\left(R_0 - \sqrt{(x-100)^2+y^2}\right)}{W}. \tag*{(34)}

In addition, for the phase field, the periodic boundary condition is applied in the horizontal direction, while the wetting and no-flux boundary conditions are imposed at the bottom and top boundaries. For the flow field, the periodic boundary condition is used in the horizontal direction, while the no-slip boundary condition is adopted at the bottom and top boundaries.

Theoretical30∘30^\circ45∘45^\circ60∘60^\circ90∘90^\circ120∘120^\circ150∘150^\circ
Numerical29.3∘29.3^\circ44.2∘44.2^\circ59.3∘59.3^\circ89.5∘89.5^\circ119.9∘119.9^\circ150.0∘150.0^\circ

Table 1. A comparison of the theoretical and numerical results of the contact angle.

We perform some simulations, and present the results in Figure 2 which shows the equilibrium shapes of the droplets under different contact angles. From this figure, one can see that the droplet on the solid wall can form different steady-state patterns, which depends on the specified contact angle. Based on the equilibrium state of the droplet, we can also numerically calculate the contact angle by using the formula θ=2arctan⁡(2H/L)\theta= 2\arctan(2H/L), where LL and HH are the spreading length and height of the droplet on the solid surface, respectively. As seen from Table I, the predicted values of contact angles are in good agreement with theoretical ones.

Figure 2. The predicted equilibrium shapes of the droplet under different contact angles [(aa) θ=30∘\theta= 30^\circ, (bb) θ=60∘\theta= 60^\circ, (cc) θ=90∘\theta= 90^\circ, (dd) θ=150∘\theta= 150^\circ].

The deformation of a leaky dielectric droplet in an external electric field

The deformation of a leaky dielectric droplet subjected to an external electric field is also used to test the capability of the LB method for the problems with the multi-physical fields. As shown in Figure 3 where a single droplet (marked as ll) with the radius aa is suspended in another immiscible fluid (marked as gg), when an electric field is applied, the droplet deforms into either a prolate or oblate shape with the internal and external circulation flows, which are induced by the electrical stress (FE\mathbf{F}_{\mathrm{E}}) acting on the interface. To quantify the droplet deformation, we adopt the following deformation factor,

Figure 3. Schematic of a droplet suspending in another immiscible liquid under an external electric field.

D=L^−H^L^+H^,(35)D = \frac{\hat{L}-\hat{H}}{\hat{L}+\hat{H}}, \tag*{(35)}

where L^\hat{L} and H^\hat{H} are the end-to-end lenghts of the stable droplet measured along and perpendicularly to the direction of the electric field, respectively. Based on the classic leaky dielectric theory [37], an approximate expression of deformation factor can be given by

D=9CaE16fT(R,S,B)(2+R)2,(36)D = \frac{9Ca_{\mathrm{E}}}{16}\frac{f_{\mathrm{T}}(R,S,B)}{(2+R)^2}, \tag*{(36)}

where fT(R,S,B)=R2+1−2S+35(R−S)2+3B1+Bf_{\mathrm{T}}(R,S,B)=R^2+1-2S+\frac{3}{5}(R-S)^{\frac{2+3B}{1+B}}, which is a function of the conductivity ratio (R=σl/σgR=\sigma_l/\sigma_g), the permittivity ratio (S=εl/εgS=\varepsilon_l/\varepsilon_g) and the dynamic viscosity ratio (B=μl/μgB=\mu_l/\mu_g). The electric capillary number (CaE=εgE02a/γCa_{\mathrm{E}}=\varepsilon_g E_0^2a/\gamma) is used to reflect the influence of the electric field. Similarly, Feng [38] also proposed a first-order approximate formula for 2D droplet deformation under a uniform electric field,

D=CaEfF(R,S)3(1+R)2,(37)D = \frac{Ca_{\mathrm{E}}f_{\mathrm{F}}(R,S)}{3(1+R)^2}, \tag*{(37)}

where fF(R,S)=R2+R+1−3Sf_F(R,S)=R^2+R+1-3S. Moreover, when D>0D>0, the droplet is prolate, while if D<0D<0, the droplet is oblate.

RRSSCaECa_EEq. (19)Eq. (20)Liu [17]Present
550.20.036700.029600.035240.03535
5600.2-0.40520-0.27590-0.25708-0.25296
120.2-0.04380-0.05000-0.04751-0.04997
5020.20.106900.065200.107560.10689
1.753.50.1-0.02230-0.02070-0.02232-0.02119
3.253.50.10.008500.008000.008330.00874
4.753.50.10.022800.018000.019530.02092

Table 2. A comparison of the numerical and theoretical results on the deformation of the leaky dielectric droplet.

In our simulations, the computational domain is Lx×Ly=400×400L_x \times L_y=400\times400 with Δx=Δt=1\Delta x=\Delta t=1, other parameters are set to be a=50a=50, W=5ΔxW=5\Delta x, ρl=ρg=1\rho_l=\rho_g=1, vl=vg=0.1v_l=v_g=0.1, γ=0.001\gamma=0.001, M=0.1M=0.1, σg=0.1\sigma_g=0.1, εg=0.0004\varepsilon_g=0.0004, E0=CaEγ/(aεg)E_0=\sqrt{Ca_E\gamma/(a\varepsilon_g)}, ϕ0=0\phi_0=0 and ϕmax⁡=ϕ0+E0Lx\phi_{\max}=\phi_0+E_0L_x. The initial distribution of the order parameter is given by

C(x,y)=Cl+Cg2+Cl−Cg2tanh⁡2[a−(x−Lx/2)2+(y−Ly/2)2]W.(38)C(x,y)=\frac{C_l+C_g}{2}+\frac{C_l-C_g}{2}\tanh\frac{2\left[a-\sqrt{(x-L_x/2)^2+(y-L_y/2)^2}\right]}{W}. \tag*{(38)}

The boundary conditions are considered as follows. For the phase and flow fields, the no-flux and no-slip boundary conditions are imposed on the left and right walls, while for the electric potential, the Dirichlet boundary conditions with ϕmax⁡\phi_{\max} and ϕ0\phi_0 are applied on the left and right walls, respectively. In addition, the periodic boundary conditions are employed on the top and bottom boundaries.

As seen from Figure 4 where CaE=0.2Ca_E=0.2, the droplet can form prolate [see Figure 4(a)] and oblate [see Figure 4(b)] patterns under the conditions of (R,S)=(5,0.5)(R,S)=(5,0.5) and (R,S)=(5,60)(R,S)=(5,60). In addition, it is also found that there are four symmetric vortices inside and outside the droplet, and when R>SR>S, the flow direction is from the equator to the poles along the droplet surface, while if R<SR<S, the flow direction is from the poles to the equator along the droplet surface. We also present the distribution of the volume charge density ρe\rho_e in Figure 4, from which one can see that when R>SR>S, the left half of the droplet has a negative charge and the right half has a positive charge, while the distribution is opposite if R<SR<S. Furthermore, we present a comparison of deformation factor among the present results, theoretical solutions [Eqs. (36) and (37)], and numerical data [17] in Table II. This table demonstrates that our numerical results are not only in good agreement with the theoretical solutions for the droplet with small deformation, but also close to the previous numerical data [17]. Here it should be noted that for the droplet with large deformation, the theoretical solutions are not accurate, and there are some differences between the numerical and theoretical data, as reported in the previous work [17].

The deformation of the leaky dielectric droplet and vortex direction

Figure 4. The deformation of the leaky dielectric droplet and vortex direction [(a) (R,S)=(5,0.5)(R,S)=(5,0.5), (b) (R,S)=(5,60)(R,S)=(5,60)].

Numerical results and discussion

In this section, we investigate the electrohydrodynamics of a droplet pair on the solid substrate by the LB method. The schematic of the problem is shown in Figure 1 where a pair of semicircular leaky dielectric droplets (marked as ll) with the same radius aa is initially placed on the solid substrate, and is also surrounded by another immiscible fluid (marked as gg). The distance between the centers of the two droplets is 4a4a, and an external electric field is applied in the horizontal direction,

In the following simulations, the computational domain is Lx×Ly=640×160L_x \times L_y=640\times160, the parameters are given by a=40a=40, W=5W=5, ρl=ρg=1\rho_l=\rho_g=1, νl=νg=0.05\nu_l=\nu_g=0.05, γ=0.004\gamma=0.004, M=0.1M=0.1, R=σl/σg=5R=\sigma_l/\sigma_g=5, σg=\sigma_g= 1/61/6, CaE=0.5Ca_E = 0.5, εg=0.005\varepsilon_g = 0.005, E0=CaEγ/(aεg)=0.1E_0 = \sqrt{Ca_E\gamma/(a\varepsilon_g)} = 0.1, ϕ0=0\phi_0 = 0 and ϕmax⁡=ϕ0+E0Lx\phi_{\max} = \phi_0 + E_0L_x. The initial order parameter is set as

C(x,y)={Cl+Cg2+Cl−Cg2tanh⁡2[a−[x−(Lx/2−2a)]2+y2]w,0≤x≤Lx/2, 0≤y≤Ly,Cl+Cg2+Cl−Cg2tanh⁡2[a−[x−(Lx/2+2a)]2+y2]w,Lx/2≤x≤Lx, 0≤y≤Ly.(39)C(x,y)= \begin{cases} \dfrac{C_l+C_g}{2}+\dfrac{C_l-C_g}{2}\tanh\dfrac{2\left[a-\sqrt{[x-(L_x/2-2a)]^2+y^2}\right]}{w}, & 0\le x\le L_x/2,\ 0\le y\le L_y,\\ \dfrac{C_l+C_g}{2}+\dfrac{C_l-C_g}{2}\tanh\dfrac{2\left[a-\sqrt{[x-(L_x/2+2a)]^2+y^2}\right]}{w}, & L_x/2\le x\le L_x,\ 0\le y\le L_y. \end{cases} \tag*{(39)}

In addition, for the phase field, the no-flux boundary is applied to the left, right and top boundaries, while the wetting boundary condition is imposed on the bottom boundary. For the flow field, the no-slip boundary condition is adopted for all four boundaries. For the electric field, the Dirichlet boundary conditions with ϕmax⁡\phi_{\max} and ϕ0\phi_0 are used at the left and right boundaries, while the no-flux boundary conditions are imposed on the bottom and top boundaries.

A grid-independence test

Before conducting any discussion, a grid-independence test is first performed. As illustrated in Figure 1, the inner-end distance of the two droplets is L1=2aL_1=2a, and its dimensionless form is given as L1∗=L1/2aL_1^*=L_1/2a. Additionally, we also define a dimensionless time t∗=nΔt/tct^*=n\Delta t/t_c, where nn denotes the number of time steps, tc=μg/(εgE02)t_c=\mu_g/(\varepsilon_gE_0^2) represents the characteristic time, with tc=1000t_c=1000 taken as a reference value. We consider the case with θ=90∘\theta=90^\circ and S=εl/εg=1.5S=\varepsilon_l/\varepsilon_g=1.5, and find that the two droplets initially attract each other, and subsequently remain stable. Table III presents the stable values of L1∗L_1^* under different grid resolutions, and the relative errors are calculated based on the reference value with a grid of Nx×Ny=1600×400N_x\times N_y=1600\times400. As shown in the table, the grid resolution 1280×3201280\times320 can give grid-independent results, and the relative error is less than 1%1\%. This grid resolution is adopted in the following simulations considering computational accuracy and efficiency.

Grid size640×160640\times1601280×3201280\times3201600×4001600\times400
L1∗L_1^*0.1570.1640.165
Relative error (%)4.850.61–

Table III. The relative errors of L1∗L_1^* under different grid resolutions.

To investigate the influence of wettability on the electrohydrodynamics of a droplet pair under an applied electric field, two different cases are considered, the first one is a droplet pair suspended in another fluid (see Figure 5), and the second one is a droplet pair placed on a solid substrate (see Figure 1). For the first case, the periodic boundary conditions are applied to the top and bottom boundaries, and the other boundary conditions and parameters are the same as those for the second case. From Figure 6(a), it can be observed that when two droplets are suspended, they initially approach each other, and subsequently move apart. This is because during this process, the droplets are subjected to both electrostatic force and electrically driven hydrodynamic force [12, 13, 39]. To explain this phenomenon more clearly, we first introduce two time scales for tangential electric stress τμ∼d2ρ/μ\tau_\mu\sim d^2\rho/\mu and the Rayleigh inertial capillary τργ=ρa3/γ\tau_\rho^\gamma= \sqrt{\rho a^3/\gamma}, then the timescale for the electric-stress-driven flow near one droplet to influence the other can be defined as τμ/τργ∼d2/(a2Oh)\tau_\mu/\tau_\rho^\gamma\sim d^2/(a^2 Oh), where the Ohnesorge number OhOh is defined by Oh=μ/ρaγOh = \mu/\sqrt{\rho a\gamma}, dd is the centre-to-centre distance between the droplets [14]. For the case with a small Oh=0.125Oh = 0.125, the droplet pair initially responds to the electrostatic interaction, which can be modeled as the force between two dipoles located at the droplet centers. The horizontal (FhF_h) and vertical (FvF_v) components of this force take the following forms [40],

Figure 5. Schematic of a droplet pair suspending in another immiscible liquid under an external electric field. The subscripts ll and gg represent the inner and outer fluids of droplets, respectively.

Figure 6. The electrohydrodynamics of the droplet pair in two different situations [(a) Suspension of the droplet pair, (b) the droplet pair on the solid substrate].

Fh=12πεgE02a6B^2d4[3(cos⁡α)2−1],Fv=−12πεgE02a6B^2d4sin⁡2α,(40)F_h = \frac{12\pi\varepsilon_g E_0^2 a^6 \hat{B}^2}{d^4}[3(\cos\alpha)^2-1], \qquad F_v = -\frac{12\pi\varepsilon_g E_0^2 a^6 \hat{B}^2}{d^4}\sin2\alpha, \tag*{(40)}

where α\alpha is the angle between the direction of the electric field and the line along the centers of two droplets, and B^=1−3/(S+2)\hat{B}=1-3/(S+2). When the electric field is applied in the horizontal direction (α=0∘\alpha= 0^\circ), the vertical component of electric force is 0, and the horizontal component has a positive sign. This explains why the droplets initially approach each other. Additionally, due to the fact that the tangential component of the electrical stress is non-zero at the interface, some vortices are generated inside and outside the droplet, and the tangential velocity at the interface can be expressed as [12]

uθ^=−9(R−S)5(1+B)(R+2)2cos⁡θ^sin⁡θ^,(41)u_{\hat{\theta}}=-\frac{9(R-S)}{5(1+B)(R+2)^2}\cos\hat{\theta}\sin\hat{\theta}, \tag*{(41)}

where θ^\hat{\theta} is the angle between the direction of applied electric field and the line connecting the droplet center to a point on the interface, and from Eq. (41), one can see that there are two types circulation, with their directions are determined by the value of θ^\hat{\theta} and sign of the term R−SR-S. As pointed out in Ref. [12], when α=0∘\alpha=0^{\circ}, the vortical circulations around one droplet push the other droplet away if R>SR>S, and the circulations around one droplet pull the other one closer if R<SR<S. With the evolution of the flow field, the dominant electrically driven hydrodynamic interaction causes the droplets to gradually separate from each other.

However, for the case of a droplet pair placed on the substrate, it initially attracts to each other, and subsequently coalesces [see Figure 6(b)]. This indicates that the wettability has an important influence on the electrohydrodynamics and final state of the droplet pair, which will be considered in the following part.

The droplet pair in a horizontal external electric field

Based on the above analysis, the permittivity ratio (S) significantly affects the droplet deformation, as well as the intensity of electrostatic interaction and hydrodynamic interaction. Besides, the wettability also exerts an apparent impact on the interaction between two droplets. In the following, we will investigate the effects of permittivity ratio and wettability on the interaction of droplet pair placed on a solid substrate under a horizontal electric field (see Figure 1). We first present the phase diagram for the modes of the droplet pair in Figure 7(a). As seen from this figure, there are five different modes under different values of permittivity ratio and contact angle, including attraction without coalescence (Mode I), attraction with coalescence (Mode II), coalescence with bubble entrapment (Mode III), coalescence followed by suspension (Mode IV) and suspension followed by coalescence (Mode V).

Phase diagram and modes of a droplet pair at different contact angles and permittivity ratios

Figure 7. The modes of the droplet pair at different contact angles (θ\theta) and permittivity ratios (SS) [◦: attraction without coalescence, □: attraction with coalescence, △: coalescence with bubble entrapment, ♢: coalescence followed by suspension, ♦: suspension followed by coalescence].

Effect of permittivity ratio

In this part, we mainly focus on the effect of permittivity ratios (SS) on the interaction of a droplet pair with θ=90∘\theta= 90^\circ. As shown in Figure 8, with the increase of SS, the final state of the droplet pair changes from non-coalescence to coalescence, which indicates that the electrostatic interaction that promotes droplet coalescence becomes more dominant. Moreover, a larger SS brings a more significant deformation during coalescence, causing the upper portions of the droplets to contact first, which further leads to air bubble entrapped [see Figure 8(c)]. To understand this phenomenon more clearly, we present the distribution of the electrical stress (FE\mathbf{F}_E, directed from the droplet to the surrounding fluid) in Figure 9 where t∗=675t^* = 675. From this figure, one can see that for all cases, FE\mathbf{F}_E only plays an important role in the regions near the interfaces of the droplets, resulting in horizontal deformation. However, increasing SS weakens the lateral electrical stress, thereby reducing the horizontal deformation of droplets. Unlike the symmetric configurations in Figure 9(b) and Figure 9(c), the droplets in Figure 9(a) experience asymmetric deformation. This asymmetry arises from the asymmetric distribution of FE\mathbf{F}_E across the droplets. In addition to the distribution of electrical stress, the flow field around the droplets is also shown in Figure 9, and four vortices form inside and outside the droplets. When R>SR > S, the fluid flows from the equator toward the polar regions along the droplet interfaces, while if R<SR < S, the flow direction reverses, moving from the poles to the equator, which is consistent with the vorticity orientation observed for the case of a single droplet [17].

Snapshots of the droplet pair at different dimensionless times

Figure 8. The snapshots of the droplet pair at different dimensionless times (t∗t^{*}) [(a) S=2S = 2, (b) S=4S = 4, (c) $S = 8].

Distribution of electrical stress at different permittivity ratios

Figure 9. The distribution of electrical stress FEF_{\mathrm{E}} at different permittivity ratios [(a) S=2S = 2, (b) S=4S = 4, (c) $S = 8].

Next, we conduct a quantitative study on the effect of SS on droplet coalescence, and focus on Model II. In particular, the normalized spreading length L∗=L/(22a)L^* = L/(2\sqrt{2}a) of post-coalescence hydrophilic droplets is measured under different values of permittivity ratio SS. As seen from Figure 10, the spreading length L∗L^* exhibits a monotonic decrease as SS increases, indicating that the axial elongation of the stable droplet along the electric field direction diminishes, which also agrees well with the theoretical predictions [Eqs. (36) and (37)]. Furthermore, due to the electrowetting of droplet, L∗L^* increases with increasing contact angle θ\theta, as seen from in Figure 10.

Figure 10. The spreading length L∗L^* after the coalescence of two droplets.

Effect of wettability

We further focus on the role of wettability in the coalescence process of droplets. As depicted in Figure 7(a), for the cases with small values of SS (S≤3.5S \le3.5), an increase in contact angle suppresses droplet coalescence, and the system changes from a coalescence to a non-coalescence state. This is because the increase in contact angle reduces the lateral spreading of the droplets, thereby diminishing the likelihood of interfacial contact. To further investigate the influence of contact angle on the final state of the two droplets, we fix the permittivity ratio S=2S = 2 and plot the evolution of dimensionless inner-end distance (L1∗L_1^*) in time (t∗t^*) in Figure 11. As shown in this figure, the distance L1∗L_1^* experiences an apparent increase during the initial stage, which is attributed to the shrinkage behavior of the droplet on a hydrophobic surface. Subsequently, due to the electrostatic interaction between the droplets, they gradually approach each other, resulting in a decrease of L1∗L_1^*. It should be noted that at S=2S = 2, the relatively weak electrostatic interaction and a hydrophobic nature of the solid substrate prevent the coalescence of two droplets, eventually causing L1∗L_1^* to remain constant. Furthermore, the results in Figure 11 also demonstrate that a larger contact angle yields a higher initial peak and slower decay rate, which indicates that the increase of the contact angle would suppress the droplet coalescence.

Figure 11. The evolutions of dimensionless inner-end distance L1∗L_1^* at different contact angles.

When SS is large (S≥4S \ge4), although the two droplets eventually coalesce, there are still some differences under specific contact angles. For instance, when 90∘≤θ≤130∘90^\circ\le\theta\le130^\circ, two droplets would coalesce, and two different modes [Mode III: coalescence with bubble entrapment in Figure 12a and Mode IV: coalescence followed by suspension in Figure 12b] are observed. Unlike the cases in Figure 8(c), the droplets in Figure 12a are in a hydrophobic state, leading to non-bottom contact during coalescence and air bubble entrapment. When SS increase to 20, the pattern shifts from Mode III to Mode IV because the electric stress (FE\mathrm{F}_{\mathrm{E}}, directed from the droplet to the surrounding fluid) concentrates at the droplet apex [see Figure 14(a)]. The deformation in the vertical direction results in significant bubble entrapment, and provides an upward lift, enabling the droplet to suspend after the coalescence. In addition, when θ≥140∘\theta\ge140^\circ, there are also two different patterns, the Mode IV [coalescence followed by suspension in Figure 13(a)] and Mode V [suspension followed by coalescence in Figure 13(b)]. Different from the case in Figure 12a, the droplets in Figure 13(a) are in a superhydrophobic state. Although FE\mathbf{F}_{E} is mainly distributed on both sides of the droplet [see Figure 14(b)], the conversion between surface energy and kinetic energy during coalescence can detach the merged droplet from the substrate (coalescence-induced droplet jumping [41]). Further increasing SS to 20 leads to the droplets subject to an upward electric stress FE\mathbf{F}_{E} [see Figure 14(c)], directly detaching them from the substrate. Since R<SR < S, both electrostatic and electrohydrodynamic interactions make the droplets approach continuously after detachment, leading to the coalescence [Figure 13(b)].

Interaction of the droplet pair at different permittivity ratios, with panels (a) $S=4$ and (b) $S=20$

Figure 13. The interaction of the droplet pair at different permittivity ratios (θ=150∘\theta= 150^\circ).

Distribution diagram of the electrical stress for three contact-angle and permittivity-ratio cases

Figure 14. The distribution diagram of the electrical stress FEF_{\mathrm{E}} at different contact angles and permittivity ratios [ (a) (θ,S\theta, S) = (120∘,20120^\circ, 20), (b) (θ,S\theta, S) = (150∘,4150^\circ, 4), (c) (θ,S\theta, S) = (150∘,20150^\circ, 20)].

Figure 12. The interaction of the droplet pair at different permittivity ratios (θ=120∘\theta= 120^\circ).

The droplet pair in a vertical external electric field

We now investigate the effects of the permittivity ratio and wettability on the dynamic behavior of the droplet pair under a vertically upward electric field. Due to the change in the direction of the electric field, the Dirichlet boundary conditions with ϕ0=0\phi_0 = 0 and ϕmax⁡=ϕ0+E0Ly\phi_{\max} = \phi_0 + E_0L_y are imposed at the top and bottom boundaries. We perform some simulations on the electrohydrodynamics of the droplet pair, and present the phase diagram in Figure 15(a) where three distinct modes are observed, including non-coalescence (Mode I), coalescence (Mode II) and suspension followed by repulsion (Mode III) [see Figure 15(b)]. When the electric field is applied in vertical direction (α=90∘\alpha= 90^\circ), only the horizontal component FhF_h of the electric force with a negative value exists [Eq. (40)], indicating a repulsive electrostatic interaction between the droplets.

Figure 15. Electrohydrodynamics of the droplet pair at different contact angles (θ\theta) and permittivity ratios (SS) [∘\circ: non-coalescence, □\square: coalescence, ⋄\diamond: suspension followed by repulsion].

Effect of the permittivity ratio

Similar to above discussion, we first consider the effect of permittivity ratio on the electrohydrodynamics of the droplet pair in Figure 16 where θ=30∘\theta= 30^\circ. In the absence of an external electric field [see Figure 16(a)], the two droplets contact and eventually coalesce. On the contrary, when an electric field is applied [see Figure 16(b) where S=0.5S = 0.5], the droplets do not coalesce because the horizontal spreading length is reduced due to the repulsion between the droplets. When SS is increased to 6 [see Figure 16(c)], the droplets contact and coalesce due to the significant deformation in the horizontal direction. To give a more detailed analysis of these phenomena, we present the distribution of electrical stress in Figure 17 where t∗=1392.5t^{*} = 1392.5. From this figure, one can see that for both cases of S=0.5S = 0.5 and S=6S = 6, FE\mathbf{F}_{\mathrm{E}} is mainly distributed at the tops of the droplets, reducing the spreading length, compared to the case without electric field. We also note that the magnitude of FE\mathbf{F}_{\mathrm{E}} for the case S=6S = 6 is much weaker than that of the case S=0.5S = 0.5, resulting in an increase in spreading length, and consequently, the coalescence phenomenon can be observed. In addition, the vortex directions near the droplet interfaces in Figure 17 are also consistent with those of a single

Evolutions of the droplet pair at different dimensionless times for the cases without electric field, $S=0.5$, and $S=6$

Figure 16. The evolutions of the droplet pair at different dimensionless times (t∗t^{*}) [(a) the case without electric field, (b) S=0.5S = 0.5, (c) S=6S = 6].

Distributions of the electrical stress at different permittivity ratios

Figure 17. The distributions of the electrical stress FE\mathbf{F}_{\mathrm{E}} at different permittivity ratios [(a) S=0.5S = 0.5, (b) S=6S = 6].

droplet.

To further investigate the influence of SS on the electrohydrodynamics of the droplet pair, we fix the contact angle θ=45∘\theta= 45^\circ and plot the dimensionless inner-end distance (L1∗L_1^*) in Figure 18. Generally, the electrostatic interaction is enhanced as SS increases, leading to an increase in L1∗L_1^*. However, when the permittivity ratio is larger than a critical value, the horizontal deformation of the droplets is increased due to the dominant hydrodynamic effect, resulting in a decrease in L1∗L_1^*.

Figure 18. The dimensionless inner-end distance L1∗L_1^* under different permittivity ratios (SS).

Effect of wettability

We now consider the effect of wettability on the electrohydrodynamics of the droplet pair. As shown in Figure 15(a), two droplets on a hydrophilic substrate are more likely to coalesce. This is because the droplets in a hydrophilic state would be prone to spread under the capillary interaction [42], making them easier to contact and coalesce. It is worth noting that when θ≥140∘\theta\ge140^\circ and SS is small, the droplets exhibit an interesting phenomenon: they first suspend and then move apart, as illustrated in Figure 19. Similar to the results in Figure 13(b), the droplets in a superhydrophobic state can also detach directly from the solid substrate due to the upward electrical stress acting on their tops [see Figure 19], and then gradually move apart under the repulsive electrostatic interaction.

Figure 19. The interaction of the droplet pair under the condition of θ=150∘\theta= 150^\circ and S=0.5S = 0.5.

We also plot the evolutions of dimensionless inner-end distance L1∗L_1^* under different electric field directions in Figure 20 where S=3S = 3, θ=60∘\theta= 60^\circ and 120∘120^\circ. As seen from this figure, regardless of whether the droplets are in a hydrophilic or hydrophobic state, the droplets under a horizontal electric field eventually attract each other, while under a vertical electric field, they repel each other in the final stage. Furthermore, from Figure 7(a) and Figure 15(a), one can observe that apart from the contact coalescence phenomenon caused by droplet spreading and the electrical stress, one can use the direction of the electric field to change the mode of the droplet pair. Additionally, through a comparison between the cases of θ=60∘\theta= 60^\circ and θ=120∘\theta= 120^\circ, we can find that the droplets in a hydrophobic state are more difficult to coalesce when a vertical electric field is applied.

Figure 20. The evolutions of dimensionless inner-end distance L1∗L_1^* under different electric field directions.

Conclusions

In this paper, the electrohydrodynamics of a pair of leaky dielectric droplets on the solid substrate is studied by a consistent and conservative phase-field-based LBM. When a horizontal electric field is applied, there are five different modes of droplet pair, including attraction without coalescence (Mode I), attraction with coalescence (Mode II), coalescence with bubble entrapment ment (Mode III), coalescence followed by suspension (Mode IV) and suspension followed by coalescence (Mode V). For the first three modes, when the droplets are in a hydrophilic state, the spreading of the droplets promotes their coalescence. However, if the droplets are in a neutral or hydrophobic state, whether they can coalesce depends on the permittivity ratio SS. Actually, the larger SS is, the easier they coalesce, this is because the electrostatic interaction becomes more significant for larger SS. In addition, with increasing SS, the horizontal deformation of the droplets decreases, and the bubble is captured during the coalescence process. For Mode IV, there are two mechanisms for droplet suspension. On the one hand, the two droplets are separated from the solid substrate by an upward electrical stress after coalescence. On the other hand, for the droplet pair on a superhydrophobic substrate, surface energy is converted into kinetic energy during coalescence process, leading to suspension after coalescence. For Mode V, the droplets on a superhydrophobic substrate are subjected to an upward electrical stress (FE\mathbf{F}_E), and they first move away from the solid substrate, and then coalesce due to the electrostatic and electrically driven hydrodynamic interactions.

When an electric field is applied in the vertical direction, the droplet pair exhibits three different modes, i.e., non-coalescence (Mode I), coalescence (Mode II), and suspension followed by repulsion (Mode III) in which the electrostatic force is repulsive. Compared to the coalesced droplets on the hydrophilic wall under no electric field, adjusting SS can prevent the droplets from coalescing. This is because the droplets are subjected to the upward electrical stress (FE\mathbf{F}_E), which reduces the spreading length, and the repulsive electrostatic force drives them apart. However, increasing SS induces horizontal deformation of the droplets, leading to coalescence. In addition, for the two droplets on a superhydrophobic substrate, the vertical electric field causes them move away after suspension due to the electrostatic and electrically driven hydrodynamic interactions, which differs from the phenomenon of coalescence after suspension under a horizontal electric field.

We hope that the research on the electrohydrodynamics of droplet pairs on a solid substrate provides guidance for controlling droplet dynamics through the application of an external electric field.

Acknowledgements

The computation is completed in the HPC Platform of Huazhong University of Science and Technology. This work was financially supported by the National Natural Science Foundation of China (Grant No. 123B2018), the Postdoctoral Fellowship Program of CPSF (Grant No. GZB20250714), China Postdoctoral Science Foundation (Grant No. 2025M773077) and the Interdisciplinary Research Program of Hust (Grant No. 2024JCYJ001).

References

References

  1. [1]K. Ahn, J. Agresti, H. Chong, M. Marquez, and D. A. Weitz, Appl. Phys. Lett. 88, 264105 (2006).
  2. [2]J. Hartmann, M. T. Schür, and S. Hardt, Nat. Commun. 13, 289 (2022).
  3. [3]B. W. An, K. Kim, H. Lee, S. Y. Kim, Y. Shim, D. Y. Lee, J. Y. Song, and J. U. Park, Adv. Mater. 27, 4322 (2015).
  4. [4]C. Sozou, J. Fluid Mech. 67, 339 (1975).
  5. [5]P. Atten, J. Electrost. 30, 259 (1993).
  6. [6]P. Atten, L. Lundgaard, and G. Berg, J. Electrost. 64, 550 (2006).
  7. [7]X. Zhang, O. A. Basaran, and R. M. Wham, AIChE J. 41, 1629 (1995).
  8. [8]J. Dong, V. F. de Almeida, and C. Tsouris, J. Dispers. Sci. Technol. 23, 155 (2002).
  9. [9]J. C. Bird, W. D. Ristenpart, A. Belmonte, and H. A. Stone, Phys. Rev. Lett. 103, 164502 (2009).
  10. [10]X. Huang, L. He, X. Luo, K. Xu, Y. Lü, and D. Yang, J. Phys. Chem. B 124, 7508 (2020).
  11. [11]J. S. Eow and M. Ghadiri, Colloid. Surf. A 219, 253 (2003).
  12. [12]J. C. Baygents, N. J. Rivette, and H. A. Stone, J. Fluid Mech. 368, 359 (1998).
  13. [13]Q. Dong and A. Sau, Phys. Rev. Fluids 3, 073701 (2018).
  14. [14]S. K. Das, A. Dalal, and G. Tomar, J. Fluid Mech. 915, A88 (2021).
  15. [15]C. Sorgentone, J. I. Kach, A. S. Khair, L. M. Walker, and P. M. Vlahovska, J. Fluid Mech. 914, A24 (2021).
  16. [16]T. Krüger, H. Kusumaatmaja, A. Kuzmin, O. Shardt, G. Silva, and E. M. Viggen, The lattice Boltzmann method (Springer, 2017).
  17. [17]X. Liu, Z. Chai, and B. Shi, Phys. Fluids 31, 092103 (2019).
  18. [18]X. Liu, Z. Chai, and B. Shi, Commun. Comput. Phys. 30, 1346 (2021).
  19. [19]X. Liu, Z. Chai, B. Shi, and X. Yuan, Physica D 468, 134294 (2024).
  20. [20]H. Hu, X. Liu, J. Peng, and Z. Chai, Acta Phys. Sin. 74, 140201 (2025).
  21. [21]S. Chen and G. D. Doolen, Annu. Rev. Fluid Mech. 30, 329 (1998).
  22. [22]C. Zhan, Z. Chai, and B. Shi, Phys. Rev. E 106, 025319 (2022).
  23. [23]X. Liu, C. Zhan, Y. Chen, Z. Chai, and B. Shi, SIAM J. Sci. Comput. 47, B308 (2025).
  24. [24]J. Hua, L. K. Lim, and C. H. Wang, Phys. Fluids 20, 113302 (2008).
  25. [25]X. He, S. Chen, and R. Zhang, J. Comput. Phys. 152, 642 (1999).
  26. [26]J. J. Huang, H. Huang, and X. Wang, Int. J. Numer. Meth. Fluids 77, 123 (2015).
  27. [27]H. Liang, H. Liu, Z. Chai, and B. Shi, Phys. Rev. E 99, 063306 (2019).
  28. [28]Y. H. Qian, D. d’Humières, and P. Lallemand, Europhys. Lett. 17, 479 (1992).
  29. [29]Z. Chai and B. Shi, Phys. Rev. E 102, 023306 (2020).
  30. [30]Z. Chai, B. Shi, and Z. Guo, J. Sci. Comput. 69, 355 (2016).
  31. [31]F. Ren, B. Song, M. C. Sukop, and H. Hu, Phys. Rev. E 94, 023311 (2016).
  32. [32]Z. Guo, B. Shi, and N. Wang, J. Comput. Phys. 165, 288 (2000).
  33. [33]Z. Chai and B. Shi, Appl. Math. Modell. 32, 2050 (2008).
  34. [34]A. J. C. Ladd, J. Fluid Mesh. 271, 285 (1994).
  35. [35]A. J. C. Ladd, J. Fluid Mesh. 271, 311 (1994).
  36. [36]T. Zhang, B. Shi, Z. Guo, Z. Chai, and J. Lu, Phys. Rev. E 85, 016701 (2012).
  37. [37]G. I. Taylor, Proc. R. Soc. Lond. A 291, 159 (1966).
  38. [38]J. Q. Feng, J. Colloid Interface Sci. 246, 112 (2002).
  39. [39]H. A. Pohl, Dielectrophoresis (Cambridge University Press, 1978).
  40. [40]M. Chiesa, J. A. Melheim, A. Pedersen, S. Ingebrigtsen, and G. Berg, Euro. J. Mech. B 24, 717 (2005).
  41. [41]C. Liu, M. Zhao, Y. Zheng, L. Cheng, J. Zhang, and C. A. T. H. Tee, Langmuir 37, 983 (2021).
  42. [42]M. Pasandideh-Fard, Y. Qiao, S. Chandra, and J. Mostaghimi, Phys. Fluids 8, 650 (1996).

Paper details

vlnc:W_6nuptgzb
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

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