1. Introduction
Particles suspended in a shear flow experience diffusion because of irreversible hydrodynamic interactions between particles. This shear-induced diffusion plays a vital role in the rheology of a suspension or emulsion. It appears in two different forms. In a suspension/emulsion, particles constantly interact with each other, and a tagged particle performs a random walk-like motion with its position varying with time as
$\langle \boldsymbol{x}\boldsymbol{\cdot }\boldsymbol{x}\rangle =6D_{s}t , D_{s}$
being the self-diffusivity (Acrivos Reference Acrivos1995), which is present even in a uniformly mixed suspension or emulsion. Additionally, in the presence of a gradient of particle volume concentration
$\phi$
, a smoothing of the gradient with a diffusive flux
$-D_{c}\boldsymbol{\nabla }\phi$
(
$D_{c}$
being gradient/collective diffusivity) occurs (Rallison & Hinch Reference Rallison and Hinch1986; Grandchamp et al. Reference Grandchamp, Coupier, Srivastav, Minetti and Podgorski2013). The shear-induced diffusion is key to understanding enhanced mixing in process industries (Lopez & Graham Reference Lopez and Graham2008), lowering the effective viscosity and pressure drop in particle flow through a channel (Acrivos Reference Acrivos1995), blunting of the velocity profile in a channel flow (Koh et al. Reference Koh, Hookham and Leal1994) and distribution of blood cells in vessels to produce a cell-free layer (Grandchamp et al. Reference Grandchamp, Coupier, Srivastav, Minetti and Podgorski2013). More recently, shear-induced diffusion has also been successful in effectively separating circulating tumour cells from blood using a microfluidic set-up (Zhou et al. Reference Zhou, Tu, Liang, Huang, Fang, Liang, Papautsky and Ye2018, Zhou & Papautsky Reference Zhou and Papautsky2019). Recently, we computed the shear-induced collective diffusivity in an emulsion of viscous drops in a viscous medium (Malipeddi & Sarkar, Reference Malipeddi and Sarkar2019a
,
Reference Malipeddi and Sarkarb
). Here, we compute the same in a viscoelastic medium.
Following the pioneering work of Eckstein et al. (Reference Eckstein, Bailey and Shapiro1977) measuring self-diffusivity from the lateral displacement of individual particles in a Couette flow, many experimental and theoretical works have investigated various aspects of shear-induced diffusion in different flow conditions. A detailed review of previous work relating viscous suspension and emulsion can be found in our previous article (Malipeddi & Sarkar, Reference Malipeddi and Sarkar2019a , Reference Malipeddi and Sarkarb ). There were very few experimental measurements of shear-induced collective diffusion in viscous emulsions (King & Leighton Reference King and Leighton2001; Hudson Reference Hudson2003), which we discussed and successfully compared with in our previous article (Malipeddi & Sarkar, Reference Malipeddi and Sarkar2019b ). Shear-induced diffusion depends on many factors such as particle properties – size, deformability, roughness, concentration, fluid properties – viscosity, density and viscoelasticity, as well as inertia and wall effects, as do single particle motion and particle interactions. Specifically, pair interactions between rigid smooth spheres in Stokes flow are reversible, i.e. two colliding spheres return to their original streamlines, requiring at least three particle interactions to break the symmetry (Marchioro & Acrivos Reference Marchioro and Acrivos2001). The above symmetry is broken by one of many factors, such as deformability (Loewenberg & Hinch Reference Loewenberg and Hinch1997), inertia (Kulkarni & Morris Reference Kulkarni and Morris2008), magnetic effects (Roure & Cunha Reference Roure and Cunha2018) or particle roughness (da Cunha and Hinch Reference Da Cunha and Hinch1996). Note that, in a particle suspension, time reversibility is seen to be violated due to the chaotic nature of the dynamics (Marchioro & Acrivos Reference Marchioro and Acrivos2001). Many of the earlier works have been devoted to understanding shear-induced diffusion of rigid particles in viscous suspensions. Only a few studies have reported rheological studies of viscoelastic emulsions (Zenit & Feng Reference Zenit and Feng2018). Recently, interest has grown in using the viscoelastic properties of a suspending fluid to focus targeted particles and cells in microchannels (Zhou & Papautsky Reference Zhou and Papautsky2020). The first normal stress difference and shear-dependent viscosity play critical roles in the effective focusing of particles in channels of different cross-sections. Fluid viscoelasticity results in the formation of chain-like structures of particles in shear flow that are absent in viscous flows (Won & Kim Reference Won and Kim2004; Scirocco, Vermant et al. Reference Scirocco, Vermant and Mewis2004; Pasquino, Panariello et al. Reference Pasquino, Panariello and Grizzuti2013) and promotes alignment of particles along the flow direction where the line joining three particles aligns along the flow direction (Choi & Hulsen Reference Choi and Hulsen2012; Jaensson et al. Reference Jaensson, Hulsen and Anderson2016).
To our knowledge, there has not been a systematic study of the diffusion of deformable particles in viscoelastic emulsions. Kim et al (Reference Kim, Han and Kim2000) repeated an earlier measurement of axial shear-induced gradient diffusion of rigid spheres in a viscous suspension by Leighton & Acrivos (Reference Leighton and Acrivos1987) in a high molecular weight polymer (polyacrylamide) solution. They found that the diffusion decreased with the polymer concentration and the shear rate. Recently, we have shown that viscoelasticity has significant effects on drop trajectory in pair interactions (Tarafder et al. Reference Tarafder, Malipeddi and Sarkar2022). The usual passing trajectories of a viscous system, where colliding drops slide past each other, change into tumbling ones in a viscoelastic system; collision traps them into a binary pair revolving around each other at low capillary and higher Weissenberg numbers. The tumbling trajectory results from the presence of a large region of spiralling streamlines around a single drop in a viscoelastic medium, helping trap the second approaching drop as well as the strong extensional stresses between drops in the separation quadrant of the collision. This study is motivated by the importance of such pair interactions on shear-induced diffusion. In this work, we numerically investigate the collective diffusion in a layer of concentrated viscous drops in shear at negligible inertia, extending our previous work in a viscous medium (Malipeddi & Sarkar, Reference Malipeddi and Sarkar2019a , Reference Malipeddi and Sarkarb ) to a viscoelastic medium with varying drop deformability and medium viscoelasticity. We note that the diffusivities are anisotropic. Similar to our previous studies, the system chosen here – a layer of drops homogenous in the flow and the vorticity directions in an unbounded shear – is suited to determining the collective diffusivity in the shear direction. In § 2, we describe the numerical procedure and discuss the results in § 3. Finally, we offer concluding remarks in § 4.
2. Mathematical formulation and numerical simulation
The dispersed continuous system is governed by the incompressible momentum conservation equations for velocity
$\boldsymbol{u}$
in the entire domain
$\varOmega$


The total stress
$\boldsymbol{\tau }$
has pressure, viscoelastic and viscous parts

where p is the pressure,
${\mu} _{s}$
is the solvent viscosity and
$\boldsymbol{D}=(\boldsymbol{\nabla }\boldsymbol{u})+(\boldsymbol{\nabla }\boldsymbol{u})^{T}$
is twice the deformation rate tensor. The superscript T represents the transpose and
$\boldsymbol{T }^{p}$
is the viscoelastic stress due to the presence of polymer. In (2.1),
$\rho$
is the density,
$\Gamma$
is the surface tension (constant) along the drop surface
$\partial B$
consisting of
$\boldsymbol{x}_{B}$
points,
$\kappa$
is the local curvature,
$\boldsymbol{n}$
is the outward normal and
$\delta (\boldsymbol{x}-\boldsymbol{x}_{B})$
is the three-dimensional Dirac delta function. We use a modified Chilcott–Rallison-type (Chilcott & Rallison Reference Chilcott and Rallison1988; Matos et al. Reference Matos, Alves and Oliveira2009) constitutive equation (also called the finite extensible nonlinear elastic modified Chilcott–Rallison or FENE-MCR) to model the viscoelasticity of the continuous phase. In our investigations of the effects of viscoelasticity, we have always chosen simple constitutive equations to explain the underlying physics. Unlike the Oldroyd-B equation used in our earlier studies (Aggarwal & Sarkar, Reference Aggarwal and Sarkar2007, Reference Aggarwal and Sarkar2008a
,
Reference Aggarwal and Sarkarb
; Mukherjee & Sarkar Reference Mukherjee and Sarkar2009, Reference Mukherjee and Sarkar2010), the FENE-MCR model in our recent studies (Mukherjee & Sarkar Reference Mukherjee and Sarkar2011, Reference Mukherjee and Sarkar2014) has a finite extensible viscosity; like Oldroyd-B, it has a constant shear viscosity. It has been extensively used in modelling different viscoelastic flows (Szabo et al. Reference Szabo, Rallison and Hinch1997; Ramaswamy & Leal Reference Ramaswamy and Leal1999; Dou & Phan-Thien Reference Dou and Phan-Thien2003; Kim et al. Reference Kim, Kim, Chung, Ahn and Lee2005). Furthermore, this constant viscosity viscoelastic model is applicable to a Boger fluid, which is also routinely used in experiments to investigate viscoelastic behaviours. e.g. for pair interactions between spheres in Boger fluids (Snijkers et al. Reference Snijkers, Pasquino and Vermant2013). Details of the implementation can be found in our previous work (Mukherjee et al. Reference Mukherjee, Tarafder, Malipeddi and Sarkar2022; Tarafder et al. Reference Tarafder, Malipeddi and Sarkar2022). The viscoelastic stress is determined by a simple rate type of equation,

Here,
${ \mu} _{p}$
is the polymeric viscosity,
$\lambda$
the relaxation time and
$L$
the finite extensibility, introduced by the FENE-CR model, which limits the maximum length of the end-to-end vector for the polymer molecule. In the limit of
$L\rightarrow \infty$
we obtain the Oldroyd-B equation with
$f\rightarrow 1$
in (2.4). We use
$L=20$
, increasing which was shown to make little difference to the results (Mukherjee & Sarkar Reference Mukherjee and Sarkar2013). We have provided additional details about the modified FENE-MCR model, noting that the nonlinearity does not affect the linear viscoelastic response, and results in a finite extensional viscosity and a constant shear viscosity (Oliveira Reference Oliveira2003; Tarafder et al. Reference Tarafder, Malipeddi and Sarkar2022). In spite of its simplicity, we feel that the model is appropriate for the present purpose of explaining a phenomenon. However, note that shear thinning may qualitatively influence the problem of the particle dynamics in a viscoelastic medium. Studies of shear-induced string-like structure formation in a viscoelastic particulate flow have found no such structure in non-shear-thinning Boger fluids (Scirocco et al. Reference Scirocco, Vermant and Mewis2004, Won & Kim Reference Won and Kim2004). Equations (2.1) and (2.2) along with the viscoelastic constitutive equation (2.4) are solved by a semi-implicit finite difference projection method in a Cartesian domain. An alternating direction implicit scheme is applied to ease the restriction on the time step. A multigrid method is used to solve the pressure position equation. A semi-analytic time integration method is used, which automatically achieves an elastic viscous stress splitting, alleviating some of the numerical stability issues common to viscoelastic numerical schemes (Sarkar & Schowalter Reference Sarkar and Schowalter2000; Izbassarov & Muradoglu Reference Izbassarov and Muradoglu2015). Details of the implementation of the above algorithm can be found in Li & Sarkar (Reference Li and Sarkar2005), Aggarwal & Sarkar (Reference Aggarwal and Sarkar2007) and Mukherjee & Sarkar (Reference Mukherjee and Sarkar2013).
The simulation details for computing the shear-induced collective diffusivity are identical to our previous publications on viscous systems (Malipeddi & Sarkar Reference Malipeddi and Sarkar2019a
,
Reference Malipeddi and Sarkarb
, Reference Malipeddi and Sarkar2021). A total of N = 70 drops with radius a are placed in a random configuration in an initially compact layer at the middle (width ∼ 0.2
$L_{y}$
) of the computational domain (figure 1) of size
$L_{y}=28a$
in the velocity gradient direction (y) and
$L_{x}=L_{z}=14a$
along the flow (x) and the vorticity (z) directions, leading to an initial volume fraction of ∼25 % in the compact layer. The independence of the result on domain dimensions, drop numbers and initial compact layer volume fractions (in the range of ∼25 %–43 %) has been investigated in a previous study of a viscous system (Malipeddi & Sarkar Reference Malipeddi and Sarkar2019b
), indicating that the current configuration is sufficient for the simulation. Specifically, the domain length of 28a along the velocity gradient direction has been shown to be sufficient to neglect any wall effects (results differed negligibly from those obtained with a domain length of 42a). Note that this approach differs from previous studies of hydrodynamic diffusivities (King & Leighton Reference King and Leighton2001; Hudson Reference Hudson2003), where drops uniformly distributed between walls experience competing effects of shear-induced diffusion and wall-induced migration. The computational domain is discretised with
$96\times 192\times 96$
grid points (∼14 grid points per drop diameter), which was shown to be sufficient in our earlier studies (Srivastava et al. Reference Srivastava, Malipeddi and Sarkar2016). The top and the bottom walls move with equal and opposite velocity U to obtain a shear rate
$\dot{\gamma }=2U/L_{y}$
. Initially, the drops are placed in a fully developed shear flow (appropriate for Stokes flow) with no polymeric stresses. Periodic boundary conditions have been applied in the x and z-directions. The drop layer has a homogeneous distribution along the x and z directions. The drop radius a and the inverse shear rate
$\dot{\gamma }^{-1}$
are used as the length and the time scales to define the Reynolds number
${Re}=\rho _{m}\dot{\gamma }a^{2}/{ \mu} _{m}$
, capillary number
$Ca={ \mu} _{m}\dot{\gamma }a/{\unicode[Arial]{x0393}}$
and Weissenberg number
$Wi=\lambda \dot{\gamma }$
. The viscosity ratio
$\lambda _{{ \mu} }={ \mu} _{d}/{ \mu} _{m}$
and the density ratio
$\lambda _{\rho }=\rho _{d}/\rho _{m}$
have been kept at unity. The polymeric to total matrix viscosity ratio
$\beta ={ \mu} _{pm}/{ \mu} _{m}$
is fixed at 0.5 (in Appendix, we briefly investigated the effects of
$\beta$
variation). Subscripts m and d denote fluid and drop phases, respectively. The total viscosity of the surrounding fluid
${ \mu} _{m}$
comprises the polymeric and the solvent viscosities
${ \mu} _{m}={ \mu} _{sm}+{ \mu} _{pm}$
. Because of the explicit nature of the code and thereby the diffusion limitation on time stepping, Re has been kept to 0.1 as a proxy for Stokes flow. We have previously shown it to be sufficient for matching with boundary element simulation of Stokes flow of viscous emulsions (Srivastava et al. Reference Srivastava, Malipeddi and Sarkar2016). Note that the front-tracking implementation is a smoothed-interface method, with the drop’s surface moving with the local velocity interpolated from the Eulerian grid to the Lagrangian drop front, naturally avoiding interpenetration; no physical drop coalescence is considered. The above code was run with the help of George Washington University’s High Performance Computing cluster PEGASUS.

Figure 1. Schematic of the computational set-up.
The shear-induced collective diffusion of the drops in an unbounded shear (ensured by a large enough computational domain; see above) leads to a spreading of the initial layer of drops concentrated in the y-direction (homogeneous in the other two directions). It is governed by a coarse-grained one-dimensional diffusion equation for the drop volume fraction
$\phi (y,t)$
(Hudson Reference Hudson2003)

where
$D_{yy}^{c}=\dot{\gamma }\phi a^{2}f_{2}$
is the collective diffusivity coefficient, which is a function of local volume fraction
$\phi (y,t)$
and
$f_{2}$
is the non-dimensional collective diffusivity. The imposed unbounded shear here eliminates the advection term due to drop migration found in Hudson (Reference Hudson2003). As has been discussed in detail in our previous article (Malipeddi & Sarkar, Reference Malipeddi and Sarkar2019b
), the linear dependence of
$D_{yy}^{c}$
on the volume fraction
$\phi$
is predicated on the dominance of the pairwise interactions validated a posteriori by the simulation results (also see Malipeddi & Sarkar, Reference Malipeddi and Sarkar2019a
, Reference Malipeddi and Sarkar2021). We have shown that
$f_{2}$
does not change with changing initial compact layer volume fraction (in the range of 25 %–43 %). Equation (2.5) for a fixed number of particles spreading by diffusion yields a self-similar parabolic volume fraction (Grandchamp et al. Reference Grandchamp, Coupier, Srivastav, Minetti and Podgorski2013)

with b as a free parameter, showing a 1/3rd scaling of the half-width
$\underline{w}$
of the profile with time

allowing determination of
$f_{2}$
. Here,
$\underline{w}_{o}$
is the initial width and
$N_{0}$
is a conserved quantity representing the number of drops diffusing out from a layer, calculated as
$N_{0}=NV/aL_{x}L_{y}$
. Malipeddi & Sarkar (Reference Malipeddi and Sarkar2019b
) have identified an alternative half-width
$w$
computed from the drop positions
$y_{i}^{\prime}, i=1,\ldots, N$

It can be related to the first moment of the analytical solution (2.6) leading to

Using either expression for finding
$f_{2}$
led to identical results within estimation uncertainties (Malipeddi & Sarkar, Reference Malipeddi and Sarkar2019b
). However, the second method avoids the intermediate step of obtaining the coarse-grained volume fraction from the drop location. We plot
$w^{3}-w_{0}^{3}$
vs.
$t^{\prime}(=t\dot{\gamma })$
and calculate the slope to find the collective diffusivity. The simulations are run up to
$t^{\prime}$
∼ 180 inverse shear unit times and
$t^{\prime}\leq t_{0}^{\prime}=20$
is discarded to make sure the drops are deformed and a linear region of
$w^{3}-w_{0}^{3}$
vs.
$t^{\prime}$
is reached. Here,
$w_{0}$
represents drop layer width at
$t^{\prime}=t_{0}^{\prime}$
. Similar to Malipeddi & Sarkar (Reference Malipeddi and Sarkar2019b
), the time evolution data excluding the initial transient are divided into smaller overlapping time intervals (∼80 non-dimensional times) with
$f_{2}$
computed in each of them; their average is reported as an ensemble average with corresponding standard deviation as the uncertainty in its estimation. Previously, we have shown that
$f_{2}$
does not depend on the initial volume fraction (in the range of ∼25 %–43 %) of the drop layer (Malipeddi & Sarkar, Reference Malipeddi and Sarkar2019b
).
Shear-induced particle diffusivity, both self and gradient (or collective), can also be obtained from the dynamic structure factor (Rallison & Hinch Reference Rallison and Hinch1986; Morris & Brady Reference Morris and Brady1996; Marchioro & Acrivos Reference Marchioro and Acrivos2001; Leshansky & Brady Reference Leshansky and Brady2005, Leshansky et al. Reference Leshansky, Morris and Brady2008). The dynamic structure factor has been useful for particle sizing in dynamic light scattering techniques, where the scattering of a monochromatic laser from the suspension volume is analysed to obtain the fluctuation autocorrelation. The exponential decay time of the autocorrelation is related to the diffusivity, which in turn is related to the particle size. Measurement of gradient diffusivity using the dynamic structure factor has previously been performed in homogeneous suspensions where the decay of spontaneously appearing fluctuation underlies the wavenumber or scale-dependent diffusivities. However, we have convincingly demonstrated that it can be used in a non-homogeneous system such as the centrally packed layer considered here to get useful information. It offered an excellent match with the diffusivity obtained by the continuum diffusion equation method (Malipeddi & Sarkar, Reference Malipeddi and Sarkar2019a
,
Reference Malipeddi and Sarkarb
, Reference Malipeddi and Sarkar2021). A detailed description of the method, including a discussion of the various issues arising from inhomogeneity of the system, has been given in Malipeddi and Sarkar (Malipeddi & Sarkar, Reference Malipeddi and Sarkar2019a
). Briefly, the method is based on the particle autocorrelation function in the Fourier (
$\boldsymbol{k}$
wavenumber) space of
$N$
drops located at
$\boldsymbol{x}_{\alpha }^{\prime}(t'),\alpha =1,2,\ldots, N$

It can be shown that the number density
$n(\boldsymbol{x}',t')$
satisfies an advection-diffusion equation (Leshansky & Brady Reference Leshansky and Brady2005)

in a shear flow
$\boldsymbol{U}+\dot{\boldsymbol{\varGamma }}\boldsymbol{\cdot }\boldsymbol{x}$
(
$\boldsymbol{U}$
is the average flow and
$\dot{\boldsymbol{\varGamma }}$
is the velocity gradient tensor), with the diffusivity in the gradient direction being related to
$F(\boldsymbol{k},t')$
as

Note that, as noted in our previous publication (Malipeddi & Sarkar Reference Malipeddi and Sarkar2019b
). the advection term in (2.11) does not affect the dynamics in a simple shear due to the orthogonality of the
$\boldsymbol{k}$
(
$=k\hat{\boldsymbol{y}}$
) vector to the velocity field. We showed before that, even in this initially non-homogeneous system, the method can be applied to obtain a meaningful diffusivity
$D_{yy}^{c}$
in the limit of
$k\rightarrow 0$
which can be matched with the results obtained from the continuum method described before. Note that, similar to the two different methods to compute
$f_{2}$
described before, the autocorrelation can also be computed either directly from the particle positions (2.10) or using the coarse-grained volume fraction
$\phi (y,t)$
satisfying the continuum (2.5). Both were shown to yield the same
$D_{yy}^{c}$
(Malipeddi & Sarkar Reference Malipeddi and Sarkar2019b
). To compute
$F(\boldsymbol{k},t')$
, signals for
$t'\lt t_{0}$
are discarded and the rest are used to calculate the autocorrelation in 80 % overlapping intervals. They are averaged with the corresponding standard deviation used as an error of estimation. We will compare the gradient diffusivity represented by
$f_{2}$
and
$D_{yy}^{c}$
.
3. Results
Previously, we offered a detailed comparison with the existing literature on hydrodynamic diffusivities for viscous systems (Malipeddi & Sarkar Reference Malipeddi and Sarkar2019a
,
Reference Malipeddi and Sarkarb
, Reference Malipeddi and Sarkar2021). Specifically, our computed collective diffusivity compared favourably with the experimental measurement (Hudson Reference Hudson2003; Malipeddy & Sarkar 2019b) found by the computed collective diffusivity matching with the zero capillary number limit of Ramachandran et al (Reference Ramachandran, Loewenberg and Leighton2010) and it is 8–9 times higher than the self-diffusivity computed by Boundary Element Method simulations of Loewenberg and Hinch (Reference Loewenberg and Hinch1997), in conformity with the theoretical prediction of da Cunha and Hinch (Reference Da Cunha and Hinch1996). We are not aware of any studies of hydrodynamic shear direction collective diffusivities of drops in a viscoelastic medium. Specifically, figures 2 and 3 show snapshots of a layer of closed-packed drops spreading with time in the velocity gradient direction due to shear-induced diffusion. In figure 2, for Ca = 0.2, both Wi = 0.0 and 2.0 cases show that the drops spread in the shear direction due to diffusion, with the spreading being less for the higher Wi (
$=2.0$
) case. We observe that, at the same time instant, drops are more deformed and inclined in the flow direction for Wi = 2.0 compared with Wi = 0.0, leading to less spreading of the layer for the more viscoelastic case. In our previous study of viscous emulsion, we noted that shear-induced collective diffusivity is a non-monotonic function of Ca due to the competing effects of increased drop deformation and stronger alignment with the flow at higher Ca. At a lower Ca = 0.02, decreased deformation leads to smaller diffusion in a viscous emulsion (Malipeddi & Sarkar Reference Malipeddi and Sarkar2019b
) as can also be seen in figure 3. The lower Ca also further strengthens the viscoelastic effects, leading to near elimination of diffusion at higher Wi = 2.0. The thickness of the compact layer of drops nearly retains its original value, although drops are constantly interacting with each other, changing their positions. As discussed below, the decreased diffusion of drops with increasing viscoelasticity of the surrounding medium (more prominent at lower Ca) can be explained by our previous study of pair interaction between drops in a viscoelastic medium (Tarafder et al. Reference Tarafder, Malipeddi and Sarkar2022). It showed that surrounding viscoelasticity results in high extensional stresses between separating drops. Furthermore, the elastic tension around a drop created a larger region of spiralling streamlines, eventually trapping a second drop into tumbling trajectories at high Wi and low Ca values. These characteristics of pair interaction led to decreased separation between drops at low Ca and high Wi seen here.

Figure 2. Spreading of drops in the velocity gradient direction at Ca = 0.2 for Wi = 0.0 (a) and Wi = 2.0 (b).

Figure 3. Spreading of drop emulsion in the velocity gradient direction at Ca = 0.02 for Wi = 0.0 (a) and Wi = 2.0 (b).
In figures 4(a) and 4(b), we show the Taylor deformation
$D=(L-B)/(L+B)$
(assuming approximately an ellipsoidal shape of the deformed drop; L and B are the major and minor axes, respectively) and inclination angle averaging over all drops for Ca = 0.2 and Ca = 0.02. They show a larger deformation and stronger alignment with the flow for the larger Ca, as expected. We also show that drop deformation and inclination angle decrease with Wi (also see Aggarwal & Sarkar Reference Aggarwal and Sarkar2007, Reference Aggarwal and Sarkar2008a
). The noise in drop deformation and inclination angle indicates frequent interactions with neighbours. At higher Ca, we note a slightly increased drop deformation and alignment towards the flow at higher viscoelasticity, which may facilitate drops passing each other, thereby reducing the post-collision separation in the velocity gradient direction and the shear-induced diffusivity. On the other hand, at lower Ca, the drop deformation is small and does not show significant dependence on Wi. Note that this is consistent with our earlier finding that, rather than drop shapes, the flow around them impacts the pair interaction in a viscoelastic medium; effects of viscoelasticity are maximal at lower Ca, where deformation is minimal. Rather than the deformability, the viscoelastic hindering of drop separation is the dominant effect in determining the diffusivity, indicating similar physics for rigid particles as well. Similar spiralling streamline patterns were seen around a rigid sphere in a viscoelastic fluid (D’Avino et al. Reference D’Avino, Hulsen, Snijkers, Vermant, Greco and Maffettone2008). Our previous pair-interaction study (Tarafder et al. Reference Tarafder, Malipeddi and Sarkar2022) also found hindered drop separation at an increased viscosity ratio, indicating reduced shear-induced diffusion in the limit of rigid spheres at an infinite viscosity ratio.

Figure 4. Average drop deformation and inclination angle for (a) Ca = 0.2, and (b) Ca = 0.02.
In figure 5, we plot drop positions vs. time, showing a random walk-like individual drop motion. At Ca = 0.2, for both Wi = 0.0 and Wi = 2.0, drop positions spread in the velocity gradient direction with time. However, for Ca = 0.02 and Wi = 2.0, the spreading is minimal, indicating no diffusion along the shear direction (as we have also seen in figure 3 bottom panel).

Figure 5. Drop position vs. time at Ca = 0.2 (a,b) and Ca = 0.02 (c,d) for viscous and viscoelastic cases.
Figure 6 shows the concentration profile of the drop layer at three different times. Figures 6(a) and 6(b) show the cases when drops are more deformable, Ca = 0.2. For both Wi = 0.0 and 2.0, the parabolic concentration profile spreads with time, indicating collective diffusion, the spreading being lower for Wi = 2.0 compared with Wi = 0.0 due to increased drop deformation and resulting stronger alignment with the flow, in conformity with figure 5.

Figure 6. Concentration profile of drops at different non-dimensional times for Ca = 0.2 (a, b) and Ca = 0.02 (c, d) for viscous (a, c) and viscoelastic (b, d) cases. Lines are best-fitted parabolas. Insets show concentrations at the same time instants (same colours) scaled with
$t^{1/3}$
(as per (2.6)) collapsing to a single curve due to self-similar evolution.
Figures 6(c) and 6(d) show the case for a low Ca = 0.02 when the drops are less deformable and nearly retain their spherical shape (figure 3). The concentration profiles of the drop layer at Ca = 0.02, Wi = 0.0 (viscous case) at different times are similar to the viscous case at Ca = 0.2 but with a slower spreading of the layer thickness (see also figure 5), i.e. lower diffusivity, also seen before (Malipeddi & Sarkar, Reference Malipeddi and Sarkar2019b ). However, for the viscoelastic case (Wi = 2.0) (figure 6 d), the concentration profile of the drops shows negligible change, indicating minimal diffusion at lower Ca and higher Wi as we saw in figures 3 and 5. In the insets of figure 6, we plot the concentration profile for each case using scaled variables showing approximate collapse of profiles from different time instants to a single curve, in conformity with (2.6), except for the low deformable (Ca = 0.02) high viscoelastic (Wi = 2.0) case (figure 6 d). For this case, profiles from different time instants in the original (non-scaled) variables show little variation over time, indicating minimal diffusion. Therefore, scaling the variables with time rather than collapsing them to a single curve separates them.
As noted before, in our recent work (Tarafder et al. Reference Tarafder, Malipeddi and Sarkar2022) we have shown that increasing viscoelasticity of the surrounding fluid reduces post-collision separation of streamlines of passing drops, eventually trapping them in tumbling trajectories at low Ca and high Wi, where the drops do not separate but rotate around the centre of the line joining them. The transition from a passing to a tumbling trajectory was ascribed to the presence of a large region of spiralling streamlines around a single drop in a viscoelastic shear flow, which traps the approaching second drop. The hoop stress due to the first normal stress difference around the spherical drop (at lower Ca) generates the region of spiralling streamlines. We provided a perturbative theory for the effects of the first normal stress difference in determining the region of the spiralling streamlines that approximately explains the scaling of the critical Ca vs. Wi curve delineating the transition from passing to tumbling trajectories at low Ca and high Wi. We also showed that, during the separation of interacting drops, high polymeric stresses develop in the separation quadrant, hindering their separation. The decreased separation between drops is the primary driving force hindering the shear-induced diffusion of drops seen here retaining their original compact configuration (figures 3 b, 5 d, 6 d) and leading to almost zero spreading at this high Wi and low Ca case.
In figure 7, following (2.9), we plot the cube of the width of the concentration profile with time, Ca = 0.2 and Ca = 0.02 for different Wivalues, showing the
$t'^{1/3}$
scaling noted before (best fitted lines are shown). The rate of spreading indicated by the slope of the curves decreases with increasing Wi for both Cavalues. For Ca = 0.02, the slope eventually goes to zero at Wi = 2.0, indicating zero diffusion. We calculate the collective diffusivity
$f_{2}$
according to (2.9) from the slope and plot it in figure 8(a) for various Ca as a function of Wi. The fitted curves show an approximate linear scaling of
$f_{2}$
with Wi for all the Ca values we have studied. The corresponding plot of
$f_{2}$
with Ca for different Wi values is shown in figure 8(b). As noted in Tarafder et al. (Reference Tarafder, Malipeddi and Sarkar2022), the effects of surrounding fluid viscoelasticity on pair interaction are complex, resulting from the flow field around the drops determined by the viscoelastic stresses competing with the imposed shear flow (Aggarwal & Sarkar, Reference Aggarwal and Sarkar2008a
). The dynamics resulted in the emergent linear decay seen in figure 8(a) (Appendix briefly considers the effects of
$\beta$
showing that
$f_{2}$
decays approximately linearly with
$Wi\beta$
). The non-monotonic variation of
$f_{2}$
seen here with Ca for the viscous case (Wi = 0) was also noted in Malipeddi & Sarkar (Reference Malipeddi and Sarkar2019b
). The diffusion increases with increasing Ca initially due to increased deformation, but it decreases at higher Ca because of increased alignment with the flow direction, facilitating the sliding of drops. At higher Wi, this non-monotonicity is subdued by viscoelasticity.

Figure 7. Width of drop layer as a function of time at Ca = 0.2 (a) and Ca = 0.02 (b) for different Wi’ values.

Figure 8. Collective diffusivity varying with Wi for various Ca (a) and varying with Ca for various Wi (b).

Figure 9. (a) Value of
$-k^{2}\ln F$
as a function of time for different k. (b) Slopes of curves in (a) for different Wi. (c) Value of
$D_{yy}^{c}$
vs. Wi for different Ca. (d) Value of
$D_{yy}^{c}$
vs. Ca for different Wi (the stars represent 0.0753f
2 for comparison).
After computing the collective diffusivity from the layer thickness, (2.8) and (2.9), we compute it using the dynamic structure factor, (2.12). It has been plotted as a function of time in figure 9(a) for different
$k$
at
$Ca=0.2$
, showing linear growth with time, which tends to a common line as
$k\rightarrow 0$
. Figure 9(b) plots the slope of the curves in figure 9(a) for different Wi values, displaying a slow variation with wavenumber
$k$
, but reaching a limiting value
$D_{yy}^{c}$
as
$k\rightarrow 0$
, as we saw also for viscous cases (Malipeddi & Sarkar Reference Malipeddi and Sarkar2019a
,Reference Malipeddi and Sarkar
b
) and with red blood cells (Malipeddi & Sarkar Reference Malipeddi and Sarkar2021) (see the discussion in § 2). Figures 9(c) and 9(d) show that
$D_{yy}^{c}$
vs. Wi for different Cavalues and
$D_{yy}^{c}$
vs. Ca for different Wivalues show identical natures as
$f_{2}$
plotted in figure 8. As noted before, the dynamic structure factor is typically applied to a homogeneous system of a certain (unchanging) volume fraction. There, the wavenumber-dependent gradient diffusivity is a property of the spontaneously appearing fluctuation at the length scale corresponding to that wavelength. Here, we computed the dynamic structure factor in a non-homogeneous system with continually changing volume fraction of the drop layer, preventing direct comparison with
$f_{2}$
. However, noting the relationship
$D_{yy}^{c}=\dot{\gamma }\phi a^{2}f_{2}$
, we numerically find a fit
$D_{yy}^{c}/\dot{\gamma }a^{2}f_{2}=\phi =0.0753$
for an overall average volume fraction scale of ∼0.1 during the entire time evolution. It led to a good match between the two computations in figure 9(d).
Finally, figures 10(a) and 10(b) display surface plots of
$f_{2}$
and
$D_{yy}^{c}$
as functions of Wi and Ca showing their identical nature. We also independently obtain curve fits for the surfaces of
$f_{2}$
and
$D_{yy}^{c}$
in figures 10(a) and 10(b)


Figure 10. Variation of
$f_{2}$
(a) and
$D_{yy}^{c}$
(b) as functions of Ca and Wi.
These relations are approximately consistent with the volume fraction scale
$D_{yy}^{c}/\dot{\gamma }a^{2}f_{2}=\phi \sim 0.1$
used before in figure 9(d). Note that such correlations as in the viscous cases (Malipeddi & Sarkar Reference Malipeddi and Sarkar2019a
,
Reference Malipeddi and Sarkarb
) are purely phenomenological and are based on the computed diffusivities in the study and are therefore not applicable outside the range of parameters of that study. Nonetheless, the correlation here is approximately close to the viscous case in the limit of
$Wi=0$
. In Appendix, we briefly consider the effects of
$\beta$
variation away from the value of 0.5 studied here.
4. Conclusion
Following our previous investigation of shear-induced gradient/collective diffusion of viscous drops and red blood cells in viscous media, here, we investigate the effects of viscoelasticity of the surrounding medium. As before, we use a front-tracking finite difference method to simulate a compact layer of viscous drops in a shear flow, which presents a sharp concentration gradient at the layer boundary. The surrounding fluid is modelled using a FENE-type constitutive equation. The drops undergo shear-induced collective diffusion because of the concentration gradient and spread in the velocity gradient direction. The diffusivity is computed using a one-third scaling of the layer width with time as well as an independent means of a dynamic structure factor approach, both methods resulting in matching results.
In a viscoelastic fluid, drops spread less, eventually leading to zero diffusion at low capillary numbers and high Weissenberg numbers. The physics underlying the reduced diffusion stems from the specific nature of pair interaction in a viscoelastic fluid flow recently investigated (Tarafder et al. Reference Tarafder, Malipeddi and Sarkar2022). Fluid viscoelasticity due to polymeric stresses changes the local flow field near a drop in a shear flow, hindering post-collision drop-separation. At higher Wi and lower Ca, tension along streamlines around a spherical drop creates a region of spiralling streamlines, eventually transitioning a passing trajectory of drops into a tumbling one where the drops revolve around each other. As a result of this hindered separation, the drop layer experiences reduced dispersion in the velocity gradient direction. Shear-induced diffusivity shows a linear decrease with Wi. Appendix briefly considers the effects of the ratio of polymeric viscosity to the total viscosity
$\beta$
to show that increasing it further enhances the effects of viscoelasticity.
Shear-induced diffusion is an important factor in industrial flows for enhancing mixing as well as in microfluidic particle technologies such as particle separation and flow focusing. This investigation indicates the possibility of using viscoelasticity as an additional means to control particulate flows in these applications. The range of capillary numbers (0.01–0.2) studied here aligns well with past experimental observations and industrial applications involving non-breaking drops. However, Weissenberg number varies widely in practice depending on the fluid chosen. The current study is limited to
$Wi\leq 2.0$
, which is sufficient for the current aim to explore the effects of viscoelasticity on diffusion. The current study is limited to monodisperse viscosity-matched drops in a non-shear-thinning Boger-type viscoelastic fluid. However, we note that our pair-interaction study (Tarafder et al. Reference Tarafder, Malipeddi and Sarkar2022) has shown that increasing the viscosity ratio enhances viscoelastic hindering of drop separation. It indicates that on increasing the viscosity ratio in an emulsion, as well as in its infinite limit, a rigid suspension of rigid particles would experience similar physics and a reduced shear-induced gradient diffusion. This study adequately demonstrates the first-order effects of the matrix viscoelasticity on the diffusion of suspended drops. Future work may consider bi/polydisperse emulsions common in many industrial flows.
Funding
K.S. acknowledges partial support from the National Science Foundation Award 1239105.
Declaration of interests
The authors declare no competing interests.
Author contributions
A.T. and K.S. conceptualised the problem. A.T. and A.R. performed the computations. A.T., A.R. and K.S. contributed to analysing the data, reaching the conclusion, and writing the manuscript. K.S. and A.R. revised the manuscript. K.S. supervised the overall project.
Data availability
The data sets generated and/or analysed during the current study are available from the corresponding author upon reasonable request.
Appendix Effects of
$\boldsymbol{\beta}$
variation
In the interest of brevity and a reasonable computational time for the entire study, we varied only Ca and Wi to investigate their effects on the shear-induced gradient diffusivity, keeping all other parameters fixed, including
$\beta$
= 0.5. In this appendix, we briefly consider the effects of
$\beta$
variation on
$f_{2}$
. Figure 11(a) plots
$f_{2}$
for three different values of
$\beta$
. We note that, except for the lowest value
$\beta$
= 0.1, where the diffusivity values are very close to the viscous case, the others show an approximately linear decreasing trend with
$\beta$
similar to the variation with Wi. Note that
$\beta$
signifies the amount of polymer viscosity and the first normal stress difference in the viscoelastic medium is
$N_{1}=2\lambda { \mu} _{p}\dot{\gamma }^{2}$
with its non-dimensionalised counterpart
$N_{1}/{ \mu} \dot{\gamma }=2Wi\beta$
. Previously, we found that, for the single drop dynamics in shear, the deformation and inclination for different
$\beta$
values collapsed to one curve when plotted against
$Wi\beta$
(Aggarwal & Sarkar, Reference Aggarwal and Sarkar2007, Reference Aggarwal and Sarkar2008a
). Figure 11(b) also shows that
$f_{2}$
approximately decreases linearly with this quantity.

Figure 11. Variation of
$f_{2}$
vs. Wi (a) and
$f_{2}$
vs.
$Wi\beta$
(b) for Ca = 0.2.