Article

Cite This: Ind. Eng. Chem. Res. 2020, 59, 3702−3716 pubs.acs.org/IECR

Numerical Insights on Controlled Droplet Formation in a

Microfluidic Flow-Focusing Device
Somasekhara Goud Sontti and Arnab Atta*
Multiscale Computational Fluid Dynamics (mCFD) Laboratory, Department of Chemical Engineering, Indian Institute of
Technology Kharagpur, Kharagpur, West Bengal 721302, India

ABSTRACT: In this article, we have developed a computational model

to determine the droplet formation regime and its transition in a square
microfluidic flow-focusing device that eventually dictate the droplet
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
Downloaded via INDIAN INST OF TECH KHARAGPUR on March 27, 2024 at 19:38:59 (UTC).

shape, size, and its formation frequency. We have methodically explored

the influences of various physicochemical parameters on the droplet
dynamics and flow regime transition, which are essential in the
development of new methods for on-demand droplet generation. On
the basis of the droplet formation mechanism, we have formulated flow
maps for different liquid−liquid systems, and have also proposed a
scaling law to predict the droplet length for a wide range of operating
condition resulting from the variation of flow rates, and viscosities of
the continuous phase as well as the interfacial tension. This work can
effectively contribute in providing helpful guidelines on the design and operations of droplet-based flow-focusing microfluidic
systems.

■ INTRODUCTION
Microfluidic devices have emerged as a novel tool for
focusing microchannel and investigated the flow field inside a
droplet. Their findings revealed that droplet topology, internal
conducting chemical and biomedical processes in a mini- velocity magnitude, and recirculation strongly depend on
aturized platform for the past two decades.1−4 Droplet-based viscosity ratio, Ca, and geometric configuration. With
microfluidic systems play a significant role in the field of increasing viscosity ratio, velocity field inside the droplet
chemical, lab-on-a-chip, drug delivery, biochemical reactions, changed in the central zone of the droplet. Their results also
encapsulation, and nanomaterial synthesis.5−8 Typically in revealed that by tuning the viscosity ratio, the strength of
microfluidic devices, reagents volumes can significantly internal circulatory motions can be controlled. Chen et al.19
decrease to micro/nanoliters and reaction time could be developed a mathematical model that described the droplet
shortened to a few seconds.9,10 Furthermore, each microfluidic formation in a flow-focusing device under squeezing regime.
droplet acts as an individual microreactor, isolated from the Kovalchuk et al.20 studied the influence of surfactant
surrounding phase in which mixing or reaction can occur. concentration on droplet formation in a cross junction
Therefore, precise control of droplet size and monodispersity microchannel, where various flow regimes such as squeezing,
are imperatively demanded in such applications. For the dripping, jetting, and threading were identified. They
generation of those bubbles/droplets, the most frequently used constructed and compared the flow regime maps for
microfluidic geometries are coflow, T-junction,11 and flow- surfactant-free and with-surfactant systems, which showed
focusing devices.12,13 Among numerous variants, the flow- transition from dripping to jetting and threading at lower flow
focusing device can allow for high-throughput production of rates of the dispersed phase in the presence of surfactant. Wu
monodisperse droplets because of symmetric rupturing of the et al.21 also presented flow regime maps based on the Ca and
dispersed phase at the cross junction.14,15 Tan et al.16 studied We for five immiscible liquid−liquid systems in flow-focusing
the formulation of monodisperse oil-in-water (O/W) and devices with different hydraulic diameters. They described the
water-in-oil (W/O) emulsions using flow-focusing device, effect of channel hydraulic diameter, dispersed phase viscosity,
where two-phase flow patterns were investigated for various and interfacial tension on flow regime transition. Three distinct
flow rate ratios, continuous phase viscosities, and surfactant flow patterns, viz., tubing/threading, dripping, and jetting, were
concentrations. Fu et al.17 explored droplet breakup dynamics
in viscous liquids and developed flow pattern maps based on Special Issue: Characterization and Applications of Fluidic Devices
dispersed phase Weber number (We) and continuous phase without Moving Parts
Capillary number (Ca). They observed dripping and jetting
flow patterns and proposed a power-law scaling relation for Received: April 19, 2019
droplet size with respect to the variation of dispersed phase Revised: July 2, 2019
thread. With the help of microparticle image velocimetry Accepted: July 4, 2019
(μPIV), Ma et al.18 analyzed droplet formation in a flow- Published: July 4, 2019

© 2019 American Chemical Society 3702 DOI: 10.1021/acs.iecr.9b02137

Ind. Eng. Chem. Res. 2020, 59, 3702−3716
Industrial & Engineering Chemistry Research Article

Figure 1. Schematic of (a) computational domain, (b) typical droplet flow in a flow-focusing microchannel, (c) forces acting on the dispersed
phase during droplet formation at the cross-junction, and (d) plug and spherical droplets with the definition of droplet deformation index (DI).

identified. Wu et al.22 investigated ferrofluid droplet formation applications, e.g., bubble rise in viscous liquids,42 bubble
in a flow-focusing device by applying radial and axial magnetic formation on submerged orifices,43 droplet impact on a liquid
fields. Their results provided new insights into the controlling pool,44 Taylor bubble formation,45−47 axisymmetric droplet
of droplet generation under external force. The applied formation,48 droplet coalescence,49 and demulsification.50 In
magnetic field was found to facilitate the formation of longer this work, we present a comprehensive numerical study based
droplets, but the influence of magnetic field decreased with the on the CLSVOF method, which is arguably better in interface
increasing flow rates. Varma et al.23 reported Janus particle tracking for low Capillary number systems than the conven-
formation in the flow-focusing device by applying a magnetic tional VOF technique to delineate the droplet formation
field. Du et al.24 analyzed the breakup dynamics of the mechanism and flow patterns in a microfluidic flow-focusing
viscoelastic dispersed thread in a flow-focusing device, where device. Determining the droplet formation regime and its
four breakup regimes of the viscoelastic thread were observed, transition that eventually dictate the droplet shape, size, and
and the liquid−liquid interface shape evaluation for each types formation frequency is the unique contribution of the present
were presented. Their study provides a foundation for the study. It is noteworthy that few previous studies have identified
formation of viscoelastic droplets in microfluidic devices. the flow regimes and its transition for a limited range of flow
Sonthalia et al.25 demonstrated the generation of extremely rates between 1 and 600 μL/min.51,52 We aim to explain the
fine water droplets in bitumen solutions using a microfluidic influences of physicochemical parameters on the droplet
flow-focusing device. dynamics and flow regime transition for a wider range of
To diminish the necessity of rigorous and exhaustive flow rate (800−2400 μL/min), which are essential in the
experimental investigation, a well-validated computational development of new high throughput methods for on-demand
model can aid in exploring critical aspects of physical droplet generation in the areas of drug delivery, crystallization,
phenomena that are not attainable by experiments. Further- and materials synthesis. Moreover, a deterministic model for
more, a predictive model also assists in interpreting the identifying flow regime transition is presented in our work,
influences of fluid physicochemical properties on the droplet which to the best our knowledge is missing in the literature.
breakup mechanism, and its plausible shape. For the design of From the perspective of achieving desired droplet size and
microfluidic devices operating in droplet flow regime, prior shape by regulating the flow regime resulting from tuning of
knowledge of various parameters such as desired droplet size, operating condition, predictability of the on-demand droplet
shape, velocity, surrounding film thickness, and pressure drop generation is of immense importance. However, such
are of paramount importance. Few numerical studies on operation necessitates the understanding of flow regime
droplet formation in flow-focusing device were carried out transition and its manipulation, which is presented here. We
using the phase field approach,26,27 Lattice Boltzmann method systematically explore the effect of flow rates, viscosity ratio,
(LBM),28,29 volume of fluid (VOF),30,31 and level set (LS)32,33 and interfacial tension on droplet length, velocity, volume, and
techniques. One of the key steps in such methodologies is the shape. We also offer a unified scaling relation to predict the
surface tension modeling, and improper model implementation droplet length for a wider range of operating condition, and
in balance between capillary force and pressure jump across the develop flow maps for different liquid−liquid systems. Such
interface can lead to the evolution of unphysical velocities near details in a microfluidic flow-focusing device are absent in the
the interface, known as spurious currents. 34,35 Several reported literature, which can notably assist in setting
researchers have attempted to reduce spurious currents with guidelines to manipulate droplet size and shape in the flow-
different approaches.36−40 Sussman and Puckett41 developed a focusing devices.
coupled LS and VOF (CLSVOF) technique that smoothly Problem Description. Schematic of a flow-focusing
captured the continuous interface by calculating the radius of microchannel considered in this study is illustrated in Figure
curvature from the LS function. This method has later been 1a, which is of square cross-section having 600 μm width (Wc
successfully utilized to unravel the physics in various = Wd = W) × 600 μm height (h). The lengths of continuous
3703 DOI: 10.1021/acs.iecr.9b02137
Ind. Eng. Chem. Res. 2020, 59, 3702−3716
Industrial & Engineering Chemistry Research Article

Figure 2. (a) Computational grid with near wall mesh refinement, (b) comparison of parasitic currents around droplet interface through VOF and
CLSVOF methods for oil−water system at Qw/Qo = 5 and Qo= 400 μL/min, (c) grid independence analysis with different mesh element sizes, and
(d) comparison of droplet length predictions against the experimental results of Wu et al.28 for Uw = 0.00252, γ = 30 mN/m, and ηw/ηo = 0.416.

and dispersed phase inlets are 3Wc each, and the length of phases are elaborated in Figure 1c. The interplay among these
primary channel is 10Wc. The dispersed phase (oil: octane forces eventually determines the droplet shape, volume and
+2.5% SPAN 80) is introduced through the main channel, and generation frequency. In this study, the droplet shape is
the continuous phase (water/water-glycerol) is entered from quantified by a deformation index (DI), which is based on the
the two side channels. Typically, in a flow-focusing device, length and height of the droplet, as shown in Figure 1d. It
both fluids meet at the cross junction, where the dispersed
indicates an undeformed/spherical shaped droplet at DI = 0,
phase droplet formation occurs due to symmetrical shearing
action of continuous phase fluid. Generated droplets then and deformed droplets with a length greater than its height are
move along the downstream of the microchannel by pushing identified by DI > 0.
the continuous phase, as depicted in Figure 1b. The governing Numerical Method. Here, we present our CLSVOF model
forces such as surface tension, shear stress and pressure build- that combines the LS and VOF method. In such a model, the
up at the cross junction arising from the interaction between mass and momentum equations are dealt using VOF method,
3704 DOI: 10.1021/acs.iecr.9b02137
Ind. Eng. Chem. Res. 2020, 59, 3702−3716
Industrial & Engineering Chemistry Research Article

and the fluid interface between two phases is resolved by LS ∇φ

κ (φ ) = ∇
method. |∇φ| (10)

l
o
Interface Capturing Technique. In LS method, the fluid
o
o
o
o
phase is identified based on the sign of a marker function (φ),
δ(φ) = m
0 if |φ| ≥ a
1 ij πφ y
o
o
o j1 + cosijjj yzzzzzz if |φ| < a
which depends on position vector (χ⃗) and time (t). The LS

o 2a jjk
o
k a {{
z
n
function (φ) represents the signed distance from the interface,
and is calculated as follows:53 (11)
∂φ
+ U⃗ ∇φ = 0 Solution methodology. A finite volume method based
∂t (1)
CFD solver, ANSYS Fluent, is used to solve the aforemen-
where U⃗ and φ are the velocity and LS function, respectively. tioned time-dependent partial differential equations. Pressure
The LS function varies smoothly between positive and negative implicit with splitting operators (PISO) logarithm is used to
magnitudes of d, where d = d(χ⃗) is the shortest distance of a solve the pressure−velocity coupling in momentum equa-
point χ⃗ from interface at time t and zero value is assigned at the tion.57 The spatial derivatives in momentum and level set

l
interface. equations are discretized using the second-order upwind
o
o
o
o
scheme. The volume fraction is solved using piecewise linear

o
d if χ is in the liquid phase
φ (χ ⃗ , t ) = m
interface construction (PLIC) algorithm.58 Subsequently,
o
o
o
variable time step and fixed Courant number (Co = 0.25)
o
o− d if χ is in the gas phase
0 if χ is in the interface

n
are considered for simulating the governing equations. In the
(2) simulation, constant velocity boundary condition is imposed
for both continuous and dispersed phase inlets. At the solid
In this work, the density and viscosity of each fluid are wall, no−slip condition is applied. Pressure boundary is
assumed to be constant, which leads to assigning two different specified at the channel outlet. At the microscale, wall
values depending on the sign of LS function in the solution adhesion property plays significant role, where three phases
domain. To ascertain continuous variation across the interface, (continuous, dispersed, and solid wall) are in contact. Here, it
these properties are consequently calculated using a smoothed is assumed that the continuous phase completely wets the
Heaviside function H(φ).41 channel wall, and the droplet formation occurs at the junction
ρ(φ) = H(φ)ρ2 + (1 − H(φ))ρ1 (3) without spreading on the wall. The fluid and material
properties considered in this work are taken from the
η(φ) = H(φ)η2 + (1 − H(φ))η1 (4) experimental work of Yao et al.,59 in which the static contact
angle of θ = 120° is specified at the solid wall. The

l0
The smoothed Heaviside function (H(φ)) is defined as
o
computational domain is meshed using hexahedral elements,
o
o
o
o Ä É
o 1 ÅÅÅ
o ÑÑ
if φ < −a which is further refined near the wall, as shown in Figure 2a.

H (φ ) = o ÅÅ1 + φ + 1 sinijj πφ yzzÑÑÑ if |φ| ≤ a

To capture the droplet surrounding liquid film precisely, at
m Å j z
o
o 2 ÅÅÇ π k a {ÑÑÑÖ
first, its thickness is estimated from analytical correlations60 for
o
o
o
o
the corresponding fluid properties and operating conditions.
o
o1
a

n
Thereafter, near wall mesh refinement is performed, which
if φ > a (5) typically varied in the range of 40 μm from the wall in all
corners. A cross-sectional view of the computational grid in the
where a is the interface thickness.
microchannel is also shown in Figure 2a. Although the
Fluid Flow Governing Equations. In CLSVOF approach,
advantage of CLSVOF method over VOF was elaborated in
a single set of conservation equations is solved for immiscible
our earlier works with gas−liquid systems;45−47 in this work,
fluids, as follows.54
the spurious currents around the droplet is further compared
Equation of continuity:
for establishing the efficacy of CLSVOF in liquid−liquid
∂ρ system. With identical element size, the magnitude of velocity
+ ∇(ρU⃗ ) = 0
∂t (6) fluctuations at the interface is found to be significantly lower in
CLSVOF method than the normal VOF method, as shown in
Equation of motion: Figure 2b. Prior to the parametric studies, mesh sensitivity
∂(ρU⃗ ) analysis on the droplet length is conducted with different
⃗ ⃗ ) = −∇P + ∇τ + FSF
+ ∇(ρUU ⃗ element sizes of 25, 30, 35, 40, and 45 μm in the core region,
∂t (7)
as shown in Figure 2c. Considering both computational
τ = ηγ ̇ = η(∇U⃗ + ∇U⃗ )
T accuracy and time, mesh element size of 30 μm in the core and
(8)
6 μm near the wall are utilized in this study.
where U⃗ , ρ, η, P, and F⃗ SF are velocity, density, dynamic
viscosity of fluid, pressure, and surface tension force,
respectively.
■ RESULTS AND DISCUSSION
Model Validation. To examine the efficacy of the
Surface Tension Modeling. In CLSVOF, the volumetric developed CLSVOF model, it is, at first, validated with the
surface tension force (F⃗ SF in eq 7) is evaluated by modifying experimental results of Wu et al.28 for oil−water system in a
the continuum surface force (CSF) model,55 as follows:56 flow-focusing device. Numerical simulations were performed
⃗ = σκ(φ)δ(φ)∇φ
FSF with the same fluid properties and channel dimensions,
(9)
considered by Wu et al.28 Figure 2d demonstrates the
where κ(φ) and δ(φ) are the interface curvature and the quantitative comparison of model prediction for droplet
smoothed Dirac delta function, respectively, defined as lengths with the experimental observations reported by Wu
3705 DOI: 10.1021/acs.iecr.9b02137
Ind. Eng. Chem. Res. 2020, 59, 3702−3716
Industrial & Engineering Chemistry Research Article

et al.28 It shows close agreement with a maximum deviation of the continuous phase viscosity. In the droplet growth stage, the
7% from the experimental results that affirms the model nose of the droplet partially blocks the outlet channel, which
efficacy for further investigation. With this validated model, impedes the flow of the continuous phase, as shown in Figure
systematic numerical investigations are carried out to elaborate 3a at t = 0.035 s. Consequently, droplet head expands slowly in
the influence of various physicochemical parameters on the the downstream and the dispersed phase neck is formed. At the
droplet formation in a flow-focusing device. Five oil−water cross junction, continuous phase squeezes the dispersed phase
systems with different glycerol concentrations are considered symmetrically and pushes the droplet thus formed in the radial
for which the physical properties are obtained from the and axial directions. Figure 3a also shows the droplet pinch-off
experimental results of Yao et al.59 It can be clearly seen from at t = 0.060 s. At the same operating condition for oil−water
Table 1 that on increasing glycerol concentration, the +40% glycerol system, droplet breakup process is similar to the
oil−water system, but an early droplet breakup at t = 0.051 s is
Table 1. Physical Properties of Continuous and Dispersed observed (Figure 3b). This is ascribed to the increased viscous
Phase at 25 °C Used in This Study59 force that resists the growth of dispersed phase in the main
channel. Therefore, droplet formation time decreases with
fluid (glycerol−wt density, ρ viscosity η interfacial tension with
% conc.) (kg/m3) (mPa s) oil γ (mN/m) increase in the continuous phase viscosity. Additionally, it is
apparent that droplet pinch-off position shifts with fluid
water 995.6 0.89 5.37
properties. Because of higher viscous resistance in the
5% glycerol 1000.8 0.94 5.24
continuous phase, dispersed phase thread length increases,
20% glycerol 1042.3 1.42 5.04
30% glycerol 1068.7 2.03 4.95
and droplet pinch-off occurs toward the downstream of the
40% glycerol 1090.4 3.32 5.06
main channel in case of oil−water+40% glycerol system, unlike
octane+2.5% 698 0.53 -
the oil−water system, where the droplet detaches close to the
SPAN80 cross-junction. Typically, the pinch-off position can also be
influenced by the operating condition and geometric
configuration.
continuous phase viscosity increases from 0.89 to 3.32 mPa Figure 4a demonstrates the effect of continuous phase
s; however, interfacial tension alters negligibly. Therefore, viscosity on nondimensional droplet length with respect to
increase in glycerol concentration eventually demonstrates the channel width. It is apparent from Figure 4a that with
effect of continuous phase viscosity. increasing continuous phase viscosity, droplet length decreases
Effect of Continuous Phase Fluid Viscosity. In this but its velocity increases, as shown in Figure 4b. This can be
section, the influence of continuous phase viscosity has been attributed to increased liquid film thickness around the droplet
systematically explored on droplet breakup process at a fixed and its nonuniform shape i.e., different radii of curvatures at
operating condition. Droplet formation in oil−water and oil− the nose and rear, which results in pressure difference around
water+40% glycerol systems are illustrated in Figure 3. The the droplet. The developed pressure gradient leads to increase
evolution of dispersed phase growth is evidently influenced by in droplet velocity with increasing continuous phase viscosity.

Figure 3. Droplet evolution in a flow-focusing microfluidic device containing (a) oil−water, and (b) oil−water+40% glycerol at Qw/Qo = 2 and Qw
= 400 μL/min (blue, oil; red, water).

3706 DOI: 10.1021/acs.iecr.9b02137

Ind. Eng. Chem. Res. 2020, 59, 3702−3716
Industrial & Engineering Chemistry Research Article

Figure 4. Effect of continuous phase fluid viscosity on (a) nondimensional droplet length, (b) droplet velocity, (c) droplet volume, and (d)
pressure drop at Qw/Qo = 2 and Qw = 400 μL/min.

Figure 5. (a) Three-dimenstional iso-surface view of simulated droplet formation in the dripping regime, where the droplet along with four vertical
slices are shown in zoomed view, and (b) visualization of surrounding liquid film thickness around the droplet (left to right: rear to nose) for Qw/
Qo = 2 and Qw = 400 μL/min.

3707 DOI: 10.1021/acs.iecr.9b02137

Ind. Eng. Chem. Res. 2020, 59, 3702−3716
Industrial & Engineering Chemistry Research Article

Figure 6. Effect of continuous phase fluid viscosity on velocity and pressure field distributions in the microchannel for (a) oil−water, (b) oil−water
+20% glycerol, and (c) oil−water+40% glycerol at Qw/Qo = 2 and Qw = 400 μL/min.

Figure 7. Sequence of images displaying formation of a droplet in oil−water+40% glycerol system: (a) dripping regime at Ca = 0.025, and (b)
jetting regime at Ca = 0.041, for Qw/Qo = 2 and Qw = 400 μL/min.

The volume of individual droplet is also calculated, which Figure 6 demonstrates the influence of viscosity on velocity
significantly decreases with increasing viscosity ratio, as field and pressure distributions in the microchannel. It is found
depicted in Figure 4c. It can be also noted from Figure 4c that with an increase in continuous phase liquid viscosity,
insets that droplet nose curvatures decreases with increasing velocity magnitude inside the droplet increases. This is mainly
continuous phase viscosity. Figure 4d shows the effect of ascribed to the combined effect of enhanced liquid film
viscosity ratio on the total pressure drop, where an increase in thickness, change in droplet shape and viscous stress on the
viscosity ratio results in almost linear enhancement of total dispersed droplet. Interestingly, velocity magnitude at the nose
pressure drop. It is well-known that the thin film between the of a droplet is found to be significantly lower than that of the
channel wall and the droplet is favorable for heat and mass middle in all cases. This observation is in close agreement with
transfer process. This liquid film thickness depends on a the experimental results of Li and Angeli61 identified using μ-
number of parameters, including interfacial tension, viscosity PIV. A similar trend is also observed in the liquid slugs.
ratio of the two flowing phases, and flow rates. Furthermore, it Furthermore, corresponding pressure fields are also analyzed,
is non-uniform, which varies in both lateral and axial directions as shown in Figure 6 for three different systems with increasing
from the droplet nose to its tail. Figure 5a illustrates the three- continuous phase liquid phase viscosity. It is evident from the
dimensional iso-surface visualization in dripping flow regime. results that pressure inside the droplet and liquid slug also
To visualize the liquid film thickness around the droplet, we increases with increasing continuous phase liquid viscosity.
created four slices along the YZ plane between the nose and Effect of Interfacial Tension. To understand the effect of
tail of the droplet, which are shown in the inset of Figure 5a (in interfacial tension on droplet formation and flow regime, a set
zoomed view of a droplet with slices). Figure 5b shows the 2D of simulations is carried out keeping other properties
planar views of those four planes. These results further unaltered. As disused in the Figure 1c, the droplet formation
establish the efficacy of CLSVOF method in precise capture of in microfluidic devices has a strong dependence on interfacial
the liquid film thickness around the droplet. tension and viscous forces. Typically, the interfacial tension of
3708 DOI: 10.1021/acs.iecr.9b02137
Ind. Eng. Chem. Res. 2020, 59, 3702−3716
Industrial & Engineering Chemistry Research Article

Figure 8. Effect of interfacial tension on (a) nondimensional droplet length, (b) droplet velocity, (c) droplet volume, and (d) deformation index
(DI) at Qw/Qo = 2 and Qw = 400 μL/min.

Figure 9. Effect of interfacial tension on droplet shape and flow regime iso-surface visualization for (a) oil−water, and (b) oil−water+40% glycerol
system at Qw/Qo = 2 and Qw = 400 μL/min.

solution decreases with increasing the surfactants concen- device, the formation can be generally classified into the
tration. In this work, the interfacial tension for three different squeezing, dripping and jetting regime. Figure 7a shows the
liquid−liquid systems is varied from their respective reference sequence of three-dimensional iso-surface visualizations in the
values 5 mN/m to 3 mN/m, which are in the practical range59 dripping regime at Ca = 0.025 for relatively low flow rate.
that make our results more relevant for application. However, decrease in the interfacial tension results in shifting
Subsequently, the results are quantified based on the Capillary of the flow regime from dripping to jetting, as depicted in
number (Ca = ηUc/γ). Figure 7 demonstrates the effect of Figure 7b. In the jetting regime, a long thread of the dispersed
interfacial tension on flow patterns for oil−water+40% glycerol phase is formed and propagated into the downstream of the
system. Based on the dominant forces in the flow-focusing microchannel, which reaches an approximately constant value
3709 DOI: 10.1021/acs.iecr.9b02137
Ind. Eng. Chem. Res. 2020, 59, 3702−3716
Industrial & Engineering Chemistry Research Article

Figure 10. Effect of dispersed phase viscosity (a) nondimensional droplet length, (b) droplet volume, (c) droplet volume, and (d) deformation
index (DI) at Qw/Qo = 2 and Qw= 400 μL/min.

(Figure 7b marked by red color box). Subsequently, the jet interfacial tension, droplet length decreases and the shape
propagates fluctuations induced by the Rayleigh-Plateau changes from elongated plugs to smaller plugs. For the studied
instability. Interestingly, in this regime the generated droplet range of Ca, dripping regime is evident from Figure 9a for oil−
diameter is larger than the dispersed thread diameter. Figure 8 water. However, for oil−water+40% glycerol system, two
shows the effect of interfacial tension on nondimensional different flow regimes are observed with increasing Ca as
droplet length, velocity, volume, and deformation index for shown in Figure 9b. In the range of 0.0243 ≤ Ca ≤ 0.0307, the
three different oil−water systems. In all cases, droplet length is dripping regime is noticed, and for 0.0307<Ca ≤ 0.0351, it
found to decrease with increasing Ca, as shown in Figure 8a. changes from dripping to jetting. Monodispersity of droplets is
This is attributed to the weak cohesive forces among liquid one of the important properties in droplet generation.
molecules. Furthermore, to predict the droplet length, simple Interestingly, a significant change in droplet shape is also
power-law scaling relations are proposed as a function of Ca in observed for oil−water+40% glycerol. Droplet shape is
the considered operating range. These relations are similar to changed from plug to near spherical with increasing Ca.
the previously reported results for various gas−liquid and Effect of Dispersed Phase Viscosity. This section
liquid−liquid systems but with different prefactor and systematically explores the effect of dispersed phase viscosity
exponent.21,62,63 Additionally, a unified scaling relation (LD/ on the droplet length and velocity by altering the oil viscosity
Wc = 0.244 Ca−0.430) is formulated to encompass the studied (ηo = 0.53−4 mPa.s) and keeping other properties and
range of Ca resulting from the variation of interfacial tension, operating conditions constant. Figure 10a shows the influence
and it shows sound agreement with a maximum deviation of of oil phase viscosity on droplet length and it indicates that
24%. Droplet velocity is found to increase with increasing Ca with an increase in oil viscosity, the droplet length decreases
(i.e., decreasing interfacial tension) for all cases, as shown in for all three systems. This can mainly be attributed to the
Figure 8b. As mentioned earlier, this phenomenon can be increased viscous resistance of dispersed phase in the main
attributed to increased liquid film thickness and change in channel. Consequently, a greater shear force acts on the
droplet shape from elongated plug to near spherical shape. dispersed phase, which tends to produce smaller droplets at the
Figure 8c shows the change in droplet volume with Ca for cross-junction. A significant difference in droplet length is
three different systems. Similar to the droplet length, droplet observed for a higher viscous system of oil−water+40%
volume also decreases with increasing Ca. Furthermore, glycerol as compared to the oil−water system because of
droplet shape is quantified by droplet deformation index higher viscosity of the continuous phase liquid. It is worth
(DI) based on its width and height (Figure 1d). It can be seen noting that the oil−water system has relatively higher
from Figure 8d that the droplet sphericity increases with interfacial tension than oil−water+40% glycerol. Therefore,
increasing Ca. This is further elaborated in Figure 9, which the droplets would be longer for oil−water system as compared
demonstrates the 3D iso-surface view of droplet shapes in to oil−water+40% glycerol, which is discussed in the previous
microchannel for oil−water and oil−water+40% glycerol section. This difference in droplet length is attributed to the
systems. It can be inferred from Figure 9a that with decreasing complex effect of liquid properties such as dispersed phase
3710 DOI: 10.1021/acs.iecr.9b02137
Ind. Eng. Chem. Res. 2020, 59, 3702−3716
Industrial & Engineering Chemistry Research Article

Figure 11. Effect of dispersed phase viscosity on volume fractions for (a) oil−water, and (b) oil−water+40% glycerol system at Qw/Qo = 2 and Qw=
400 μL/min.

Figure 12. Effect of continuous phase flow rate (a) nondimensional droplet length, (b) droplet velocity, (c) droplet volume, and (d) pressure
profiles along the channel centerline at Qo= 400 μL/min.

viscosity and interfacial tension. In line with the previous 4.50) for oil−water system. In the considered viscosity range,
discussion, droplet velocity is found to increase as a result of dripping regime is experienced and plug shape droplets are
different droplet shape and enhancement in associated near observed. Similarly, for oil−water+40% glycerol a significant
wall liquid film thickness with increasing dispersed phase change in droplet length is also observed, as shown in Figure
viscosity, as shown in Figure 10b. Droplet volume is also 11b. For the oil−water+40% glycerol system, dripping regime
quantified and is found to decrease with increasing dispersed is observed for lower viscosity range (ηo/ηw = 0.30−0.60).
phase viscosity for all the cases, as shown in Figure 10c. The DI With a further increase in viscosity ratio, flow regime shifts
also decreases (Figure 10d) as the viscosity ratio increases. It is from dripping to jetting. Droplet shape also changes from
also found that for lower viscous cases, longer droplets are smaller plugs to near spherical with increasing viscosity ratio
formed, where the length is greater than channel hydraulic (Figure 11b).
diameter. Therefore, the droplet shape is also found to be Effect of Flow Rate. To demonstrate the influence of
influenced by the dispersed phase viscosity. It can be seen from continuous phase velocity on droplet breakup, we performed
the volume fraction contours (Figure 11a) that the droplet systematic investigations by altering the continuous phase flow
length decreases with increasing viscosity ratio (ηo/ηw = 1.12− rate for three different systems. It can be seen from Figure 12a
3711 DOI: 10.1021/acs.iecr.9b02137
Ind. Eng. Chem. Res. 2020, 59, 3702−3716
Industrial & Engineering Chemistry Research Article

Figure 13. Comparison of velocity fields during droplet formation in oil−water+40% glycerol system for (a) Qw/Qo = 2, and (b) Qw/Qo = 4 at Qw=
400 μL/min.

that the droplet length decreases with increasing continuous

phase flow rate at fixed Qo. This is due to the fact that shear
force increases on the dispersed phase at cross junction with an
increase in continuous phase flow rate. Consequently, it leads
to the formation of smaller droplets with higher formation
frequency. It is apparent from Figure 12b that the droplet
velocity increases with increasing continuous phase flow rate
due to increase in liquid film thickness and change in droplet
shape. Similar to droplet length, droplet volume also decreases
with increasing continuous phase flow rate for all cases, as
shown in Figure 12c. The pressure variation along the center
line of the channel is presented in Figure 12d, which shows
that pressure drop increases with increasing continuous phase Figure 14. Scaling relation for nondimensional droplet length with
viscosity (i.e., glycerol concentration) at a fixed flow rate. The the Capillary number for different continuous phase flow rates at Qo=
pressure field in dispersed phase is found to be nearly constant 400 μL/min.
up to the droplet pinch-off point. Thereafter, significant change
in total pressure is observed due to the interaction of two- maximum deviation of 27% for the complete range of our
phases. In any pressure profile, each peak and its width indicate studied operating condition encompassing the influence of
flow rates and viscosities of the continuous phase. From Figure
a droplet position and shape, respectively. Moreover, because
15a, it is observed that with increasing continuous phase flow
of the difference in radius of curvature from rear to nose of a
rate, dripping regime is not influenced for oil−water system.
droplet, there exists a pressure gradient across the droplet,
However, droplet shape significantly changes from slug to near
which is linked with the pressure profile in the continuous spherical at a higher flow rate. However, in oil−water+40%
phase. These results corroborate with the theoretical analysis glycerol system, dripping regime is observed at a lower flow
by Abiev,64 which ensures the link between pressure profile rate, as shown in Figure 15b. With increasing flow rate, the
inside and outside of a droplet through the Laplace equation. regime shifts from dripping to jetting and considerable change
Figure 13 demonstrates the shearing effect by continuous in droplet shape is observed, as depicted in Figure 15b.
phase for two different flow rate ratios at fixed Qw. It can be Furthermore, the flow regime maps are developed for oil−
seen from Figure 13 that the continuous phase flow rate plays water and oil−water+40% glycerol systems as a function of
an important role in symmetrical shearing of dispersed phase continuous phase Capillary number (Cac = ηcUc/γ) and
and also influences the droplet size, formation time, and flow dispersed phase Weber number (Wed = ρU2dW/γ) by altering
regime. For lower flow rate, first droplet breakup is observed at the corresponding phase flow rates. As discussed earlier, three
t = 51 ms. However, in case of higher flow rate, it is noticed at t different flow regimes namely squeezing, dripping, and jetting
= 35 ms. Velocity field around the dispersed phase is also patterns are identified. Figure 16a shows that the oil−water
compared for between those two cases in Figure 13, which system has larger squeezing and dripping regimes than the oil−
shows increased velocity magnitude in higher continuous water+40% glycerol system, where squeezing regime is
phase flow rate as compared to the lower flow rate. completely absent in the range of our studied operating
Additionally, the developed scaling relation (LD/Wc = 0.244 condition. In addition, pressure evolution in the continuous
Ca−0.430 in Figure 8a) is further verified for variation in and dispersed phases, as well as at the pinch-off position is
continuous phase flow rate, as shown in Figure 14. The analyzed during the droplet formation process by creating a XY
proposed unified scaling relation conforms well with a plane in the middle of the microchannel at Z = 300 μm (Figure
3712 DOI: 10.1021/acs.iecr.9b02137
Ind. Eng. Chem. Res. 2020, 59, 3702−3716
Industrial & Engineering Chemistry Research Article

Figure 15. Three dimensional iso-surface evolution of droplet shapes for various flow rate ratio in (a) oil−water and (b) oil−water+40% glycerol
system at Qo = 400 μL/min.

Figure 16. Flow pattern maps for different liquid−liquid systems in a flow-focusing microchannel containing (a) oil−water, and (b) oil−water
+40% glycerol.

Figure 17. (a) XY plane in the microchannel at Z = 300 μm, (b) locations of four selected pressure recording points on the XY plane, (c) evolution
of pressures in the microchannel for oil−water, and (d) oil−water+40% glycerol system at fixed operating conditions of θ = 120°, Qw/Qo = 2 and
Qw = 400 μL/min.

17a). Figure 17b shows four points (PC1, PC2, PD1, and PP1), side channels for the continuous phase. PD1 (1500 μm, 300
which are chosen on that XY plane. PC1 (2100 μm, 700 μm) μm) and PP1 (2500 μm, 300 μm) are considered in the
and PC2 (2100 μm, −100 μm) are located at the upstream of dispersed phase and at the downstream edge of the cross
3713 DOI: 10.1021/acs.iecr.9b02137
Ind. Eng. Chem. Res. 2020, 59, 3702−3716
Industrial & Engineering Chemistry Research Article

junction, respectively. As the dispersed phase enters the cross ORCID

junction and its tip partially blocks the main channel, the Somasekhara Goud Sontti: 0000-0002-9007-9124
upstream and downstream pressure of the continuous phase Arnab Atta: 0000-0001-9052-6097
(PC1 and PC2) increases and reaches a maximum value when
Notes
the dispersed phase is completely occupied at the cross
The authors declare no competing financial interest.

■
junction. Subsequently, PC1 and PC2 decrease during the
blocking time until the neck breakup occurs, as shown in
Figure 17b. Afterward, PC1 and PC2 drop suddenly and a sharp ACKNOWLEDGMENTS
peak can be seen for PD1. This is attributed to the rebound This work is supported by IIT Kharagpur under the Centre of
after separation of the droplet. Once the droplet pinch-off Excellence for Training and Research in Microfluidics.
occurs, PC and PD regain their pressure for the next cycle. It is
worth noting that PP1 indicates dispersed phase pressure at the
droplet pinch-off position. Likewise, pressure fluctuations are
■ NOMENCLATURE
Ca, Capillary number (= ηU/γ)
evidenced for oil−water+40% glycerol, as depicted in Figure We, Weber number (= ρU2dW/γ)
17c. The pressure fluctuation in oil−water+40% glycerol is W, width (m)
higher in magnitude as compared to the oil−water system due Q, flow rate (m3/s)
to higher viscous resistance.

■
U, velocity (m/s)
V, volume (m3)
CONCLUSIONS N̂ , unit normal vector
A three-dimensional computational study on droplet formation P, pressure (Pa)
and breakup dynamics in viscous liquids is carried out in a t, flow time (s)
flow-focusing device using CLSVOF method. Oil is considered H, height of the droplet (m)
as dispersed phase and water/glycerol solutions are taken as h, depth of the microchannel (m)
continuous phase. The effects of viscosity ratio, interfacial H(φ), Heaviside function
tension, and flow rate ratio are investigated in detail. For all χ⃗, position vector
cases, the liquid film thickness around the droplet is precisely a, interface thickness (m)
captured to understand its effect on the droplet velocity. d, shortest distance
Results show that droplet length and volume decrease with δ(φ), direct delta function
increasing viscosity and flow rate ratios; however, droplet DI, droplet defromation index

■
velocity and liquid film thickness increase. Scaling laws are
proposed to determine the droplet length based on Ca. GREEK SYMBOLS
Dripping regime is observed at lower Ca and flow rate ratios.
α, volume fraction
However, jetting regime is found at the higher flow rate ratios.
γ̇, shear rate (1/s)
Flow regime transition from dripping to jetting is observed
δ, liquid film thickness (m)
with changing interfacial tension, flow rate ratio, and dispersed
θ, contact angle (◦)
phase viscosity. The impact of continuous phase velocity on
κn, radius of curvature
dispersed phase shearing and droplet breakup is demonstrated.
η, dynamic viscosity (kg/m·s)
Velocity field variations inside the droplet and liquid slug are
ρ, density (kg/m3)
also analyzed for various systems. A parameter defined as
γ, interfacial tension (N/m)
droplet deformation index is used to identify the shape of
τ̅, shear stress (Pa)
droplets, which shows that near-spherical droplets are typically
ϕ, level set function

■
formed in all cases of the jetting regime. Plug-shaped droplets
are obtained in the dripping regime. The evolution of pressure
over time in the continuous and dispersed phases is analyzed. SUBSCRIPTS
From the model predictions, a unified scaling law is proposed D, droplet
to estimate the droplet length for a wide range (0.006−0.072) c, continuous phase
of Ca. Flow regime maps for two different liquid−liquid d, dispersed phase
systems are also developed as a function of continuous phase o, oil
Capillary number and dispersed phase Weber number. w, water
Nonetheless, it is of considerable importance to identify the
liquid film thickness around the droplet, which essentially
dictates the droplet dynamics. None of the previous studies
■ REFERENCES
(1) Yu, X.; Cheng, G.; Zhou, M.-D.; Zheng, S.-Y. On-demand one-
had clearly delineated this influence, whereas our results herein step synthesis of monodisperse functional polymeric microspheres
are demonstrated based on this fundamental aspect. Therefore, with droplet microfluidics. Langmuir 2015, 31, 3982.
this work contributes toward accurate predictions of droplet (2) Dang, T.-D.; Cheney, M. A.; Qian, S.; Joo, S. W.; Min, B.-K.
formation mechanism in various flow regimes. These outcomes Reactivity of poly (dimethylsiloxane) toward acidic permanganate. J.
are likely to guide experiments, where monosized droplet Ind. Eng. Chem. 2013, 19, 1770.
generation and its manipulation with tunable volume are (3) Bennett, J. A.; Kristof, A. J.; Vasudevan, V.; Genzer, J.; Srogl, J.;
Abolhasani, M. Microfluidic synthesis of elastomeric microparticles: A
essential.

■
case study in catalysis of palladium-mediated cross-coupling. AIChE J.
2018, 64, 3188.
AUTHOR INFORMATION (4) Nieves-Remacha, M. J.; Kulkarni, A. A.; Jensen, K. F.
Corresponding Author Hydrodynamics of liquid−liquid dispersion in an advanced-flow
*Email: [email protected]. Phone: +91 (3222) 283910. reactor. Ind. Eng. Chem. Res. 2012, 51, 16251.

3714 DOI: 10.1021/acs.iecr.9b02137

Ind. Eng. Chem. Res. 2020, 59, 3702−3716
Industrial & Engineering Chemistry Research Article

(5) Song, H.; Chen, D. L.; Ismagilov, R. F. Reactions in droplets in (28) Wu, L.; Tsutahara, M.; Kim, L. S.; Ha, M. Three-dimensional
microfluidic channels. Angew. Chem., Int. Ed. 2006, 45, 7336. lattice Boltzmann simulations of droplet formation in a cross-junction
(6) Haeberle, S.; Zengerle, R. Microfluidic platforms for lab-on-a- microchannel. Int. J. Multiphase Flow 2008, 34, 852.
chip applications. Lab Chip 2007, 7, 1094. (29) Gupta, A.; Matharoo, H. S.; Makkar, D.; Kumar, R. Droplet
(7) Chueluecha, N.; Kaewchada, A.; Jaree, A. Enhancement of formation via squeezing mechanism in a microfluidic flow-focusing
biodiesel synthesis using co-solvent in a packed-microchannel. J. Ind. device. Comput. Fluids 2014, 100, 218.
Eng. Chem. 2017, 51, 162. (30) Ong, W.-L.; Hua, J.; Zhang, B.; Teo, T.-Y.; Zhuo, J.; Nguyen,
(8) Campbell, Z. S.; Parker, M.; Bennett, J. A.; Yusuf, S.; Al-Rashdi, N.-T.; Ranganathan, N.; Yobas, L. Experimental and computational
A. K.; Lustik, J.; Li, F.; Abolhasani, M. Continuous Synthesis of analysis of droplet formation in a high-performance flow-focusing
Monodisperse Yolk-Shell Titania Microspheres. Chem. Mater. 2018, geometry. Sens. Actuators, A 2007, 138, 203.
30, 8948. (31) Hoang, D. A.; van Steijn, V.; Portela, L. M.; Kreutzer, M. T.;
(9) Chand, R.; Jeun, J.-H.; Park, M.-H.; Kim, J.-M.; Shin, I.-S.; Kim, Kleijn, C. R. Benchmark numerical simulations of segmented two-
Y.-S. Electroimmobilization of DNA for ultrafast detection on a phase flows in microchannels using the Volume of Fluid method.
microchannel integrated pentacene TFT. J. Ind. Eng. Chem. 2015, 21, Comput. Fluids 2013, 86, 28.
126. (32) Li, Y.; Jain, M.; Ma, Y.; Nandakumar, K. Control of the breakup
(10) Abolhasani, M.; Jensen, K. F. Oscillatory multiphase flow process of viscous droplets by an external electric field inside a
strategy for chemistry and biology. Lab Chip 2016, 16, 2775. microfluidic device. Soft Matter 2015, 11, 3884.
(11) Ma, R.; Zhang, Q.; Fu, T.; Zhu, C.; Wang, K.; Ma, Y.; Luo, G. (33) Lan, W.; Li, S.; Wang, Y.; Luo, G. CFD simulation of droplet
Manipulation of microdroplets at a T-junction: Coalescence and formation in microchannels by a modified level set method. Ind. Eng.
scaling law. J. Ind. Eng. Chem. 2018, 65, 272. Chem. Res. 2014, 53, 4913.
(12) Teh, S.-Y.; Lin, R.; Hung, L.-H.; Lee, A. P. Droplet (34) Popinet, S.; Zaleski, S. A front-tracking algorithm for accurate
microfluidics. Lab Chip 2008, 8, 198. representation of surface tension. Int. J. Numer. Methods Fluids 1999,
(13) Bordbar, A.; Taassob, A.; Zarnaghsh, A.; Kamali, R. Slug flow in 30, 775.
microchannels: Numerical simulation and applications. J. Ind. Eng. (35) Harvie, D. J.; Davidson, M.; Rudman, M. An analysis of
Chem. 2018, 62, 26. parasitic current generation in volume of fluid simulations. Appl.
(14) Anna, S. L.; Bontoux, N.; Stone, H. A. Formation of dispersions Math. Model. 2006, 30, 1056.
using flow focusing in microchannels. Appl. Phys. Lett. 2003, 82, 364. (36) Popinet, S. Gerris: A tree-based adaptive solver for the
(15) Dang, T.-D.; Kim, Y. H.; Kim, H. G.; Kim, G. M. Preparation of incompressible Euler equations in complex geometries. J. Comput.
monodisperse PEG hydrogel microparticles using a microfluidic flow- Phys. 2003, 190, 572.
focusing device. J. Ind. Eng. Chem. 2012, 18, 1308. (37) Francois, M. M.; Cummins, S. J.; Dendy, E. D.; Kothe, D. B.;
(16) Tan, J.; Xu, J.; Li, S.; Luo, G. Drop dispenser in a cross-junction Sicilian, J. M.; Williams, M. W. A balanced-force algorithm for
microfluidic device: Scaling and mechanism of break-up. Chem. Eng. J. continuous and sharp interfacial surface tension models within a
volume tracking framework. J. Comput. Phys. 2006, 213, 141.
2008, 136, 306.
(38) Wörner, M. Numerical modeling of multiphase flows in
(17) Fu, T.; Wu, Y.; Ma, Y.; Li, H. Z. Droplet formation and breakup
microfluidics and micro process engineering: a review of methods and
dynamics in microfluidic flow-focusing devices: from dripping to
applications. Microfluid. Nanofluid. 2012, 12, 841.
jetting. Chem. Eng. Sci. 2012, 84, 207.
(39) Denner, F.; van Wachem, B. G. Fully-coupled balanced-force
(18) Ma, S.; Sherwood, J. M.; Huck, W. T.; Balabani, S. On the flow
VOF framework for arbitrary meshes with least-squares curvature
topology inside droplets moving in rectangular microchannels. Lab
evaluation from volume fractions. Numer. Heat Transfer, Part B 2014,
Chip 2014, 14, 3611.
65, 218.
(19) Chen, X.; Glawdel, T.; Cui, N.; Ren, C. L. Model of droplet
(40) Guo, Z.; Fletcher, D. F.; Haynes, B. S. Implementation of a
generation in flow focusing generators operating in the squeezing height function method to alleviate spurious currents in CFD
regime. Microfluid. Nanofluid. 2015, 18, 1341. modelling of annular flow in microchannels. Appl. Math. Model.
(20) Kovalchuk, N. M.; Roumpea, E.; Nowak, E.; Chinaud, M.; 2015, 39, 4665.
Angeli, P.; Simmons, M. J. Effect of surfactant on emulsification in (41) Sussman, M.; Puckett, E. G. A coupled level set and volume-of-
microchannels. Chem. Eng. Sci. 2018, 176, 139. fluid method for computing 3D and axisymmetric incompressible
(21) Wu, Z.; Cao, Z.; Sundén, B. Liquid-liquid flow patterns and two-phase flows. J. Comput. Phys. 2000, 162, 301.
slug hydrodynamics in square microchannels of cross-shaped (42) Keshavarzi, G.; Pawell, R. S.; Barber, T. J.; Yeoh, G. H.
junctions. Chem. Eng. Sci. 2017, 174, 56. Transient analysis of a single rising bubble used for numerical
(22) Wu, Y.; Fu, T.; Ma, Y.; Li, H. Z. Ferrofluid droplet formation validation for multiphase flow. Chem. Eng. Sci. 2014, 112, 25.
and breakup dynamics in a microfluidic flow-focusing device. Soft (43) Buwa, V. V.; Gerlach, D.; Durst, F.; Schlücker, E. Numerical
Matter 2013, 9, 9792. simulations of bubble formation on submerged orifices: period-1 and
(23) Varma, V.; Wu, R.; Wang, Z.; Ramanujan, R. Magnetic Janus period-2 bubbling regimes. Chem. Eng. Sci. 2007, 62, 7119.
particles synthesized using droplet micro-magnetofluidic techniques (44) Ray, B.; Biswas, G.; Sharma, A. Regimes during liquid drop
for protein detection. Lab Chip 2017, 17, 3514. impact on a liquid pool. J. Fluid Mech. 2015, 768, 492.
(24) Du, W.; Fu, T.; Zhang, Q.; Zhu, C.; Ma, Y.; Li, H. Z. Breakup (45) Sontti, S. G.; Atta, A. Numerical investigation of viscous effect
dynamics for droplet formation in a flow-focusing device: Rupture on Taylor bubble formation in co-flow microchannel. Comput.-Aided
position of viscoelastic thread from matrix. Chem. Eng. Sci. 2016, 153, Chem. Eng. 2017, 40, 1201−1206.
255. (46) Sontti, S. G.; Pallewar, P. G.; Ghosh, A. B.; Atta, A.
(25) Sonthalia, R.; Ng, S.; Ramachandran, A. Formation of Understanding the inuence of rheological properties of shear-thinning
extremely fine water droplets in sheared, concentrated bitumen liquids on segmented ow in microchannel using CLSVOF based CFD
solutions via surfactant-mediated tip streaming. Fuel 2016, 180, 538. model. Can. J. Chem. Eng. 2019, 97, 1208.
(26) Zhou, C.; Yue, P.; Feng, J. J. Formation of simple and (47) Sontti, S. G.; Atta, A. Formation characteristics of Taylor
compound drops in microfluidic devices. Phys. Fluids 2006, 18, bubbles in power-law liquids flowing through a microfluidic co-flow
092105. device. J. Ind. Eng. Chem. 2018, 65, 82.
(27) Bai, F.; He, X.; Yang, X.; Zhou, R.; Wang, C. Three (48) Chakraborty, I.; Rubio-Rubio, M.; Sevilla, A.; Gordillo, J.
dimensional phase-field investigation of droplet formation in Numerical simulation of axisymmetric drop formation using a coupled
microfluidic flow focusing devices with experimental validation. Int. level set and volume of fluid method. Int. J. Multiphase Flow 2016, 84,
J. Multiphase Flow 2017, 93, 130. 54.

3715 DOI: 10.1021/acs.iecr.9b02137

Ind. Eng. Chem. Res. 2020, 59, 3702−3716
Industrial & Engineering Chemistry Research Article

(49) Mino, Y.; Kagawa, Y.; Ishigami, T.; Matsuyama, H. Numerical

simulation of coalescence phenomena of oil-in-water emulsions
permeating through straight membrane pore. Colloids Surf., A 2016,
491, 70.
(50) Kagawa, Y.; Ishigami, T.; Hayashi, K.; Fuse, H.; Mino, Y.;
Matsuyama, H. Permeation of concentrated oil-in-water emulsions
through a membrane pore: numerical simulation using a coupled level
set and the volume-of-fluid method. Soft Matter 2014, 10, 7985.
(51) Costa, A. L. R.; Gomes, A.; Cunha, R. L. Studies of droplets
formation regime and actual flow rate of liquid-liquid flows in flow-
focusing microfluidic devices. Exp. Therm. Fluid Sci. 2017, 85, 167.
(52) Cao, Z.; Wu, Z.; Sundén, B. Dimensionless analysis on liquid-
liquid flow patterns and scaling law on slug hydrodynamics in cross-
junction microchannels. Chem. Eng. J. 2018, 344, 604.
(53) Sussman, M.; Smereka, P.; Osher, S. A level set approach for
computing solutions to incompressible two-phase flow. J. Comput.
Phys. 1994, 114, 146.
(54) Hirt, C. W.; Nichols, B. D. Volume of fluid (VOF) method for
the dynamics of free boundaries. J. Comput. Phys. 1981, 39, 201.
(55) Brackbill, J.; Kothe, D. B.; Zemach, C. A continuum method for
modeling surface tension. J. Comput. Phys. 1992, 100, 335.
(56) Sussman, M.; Almgren, A. S.; Bell, J. B.; Colella, P.; Howell, L.
H.; Welcome, M. L. An adaptive level set approach for incompressible
two-phase flows. J. Comput. Phys. 1999, 148, 81.
(57) Issa, R. I. Solution of the implicitly discretised fluid flow
equations by operator-splitting. J. Comput. Phys. 1986, 62, 40.
(58) Holt, M. Numerical Methods in Fluid Dynamics; Springer
Science & Business Media, 2012.
(59) Yao, C.; Liu, Y.; Xu, C.; Zhao, S.; Chen, G. Formation of
Liquid-Liquid Slug Flow in a Microfluidic T-junction: Effects of Fluid
Properties and Leakage Flow. AIChE J. 2018, 64, 346.
(60) Sontti, S. G.; Atta, A. CFD Analysis of Taylor Bubble in a Co-
Flow Microchannel with Newtonian and Non-Newtonian Liquid. Ind.
Eng. Chem. Res. 2017, 56, 7401.
(61) Li, Q.; Angeli, P. Experimental and numerical hydrodynamic
studies of ionic liquid-aqueous plug flow in small channels. Chem. Eng.
J. 2017, 328, 717.
(62) Sontti, S. G.; Atta, A. CFD study on Taylor bubble
characteristics in Carreau-Yasuda shear thinning liquids. Can. J.
Chem. Eng. 2019, 97, 616.
(63) Sontti, S. G.; Atta, A. CFD analysis of microfluidic droplet
formation in non−Newtonian liquid. Chem. Eng. J. 2017, 330, 245.
(64) Abiev, R. S. Analysis of local pressure gradient inversion and
form of bubbles in Taylor flow in microchannels. Chem. Eng. Sci.
2017, 174, 403.

3716 DOI: 10.1021/acs.iecr.9b02137

Ind. Eng. Chem. Res. 2020, 59, 3702−3716