Soft Matter

View Article Online

PAPER View Journal | View Issue

Effects of vertical confinement on the flow

of polymer solutions in planar
Published on 16 September 2022. Downloaded by Clemson University on 8/17/2023 8:45:26 PM.

Cite this: Soft Matter, 2022,

18, 7427 constriction microchannels
Mahmud Kamal Raihan, a Sen Wu,ab Heston Dort,a Micah Baghdady,a
Yongxin Song*b and Xiangchun Xuan *a

The flow of polymer solutions under extensional conditions is frequently encountered in numerous
engineering fields. Planar contraction and/or expansion microchannels have been a subject of interest
for many studies in that regard, which, however, have mostly focused on shallow channel structures.
We investigate here the effect of changing the depth of contraction–expansion microchannels on the
flow responses of three types of polymer solutions and water. The flow of viscoelastic polyethylene
oxide (PEO) solution is found to become more stable with suppressed vortex formation and growth in
the contraction part while being less stable in the expansion part with the increase of the channel depth.
These opposing trends in the contraction and expansion flows are noted to have similarities with our
recent findings of constriction length-dependent instabilities in the same PEO solution (M. K. Raihan,
S. Wu, Y. Song and X. Xuan, Soft Matter, 2021, 17, 9198–9209), where the contraction flow gets
Received 30th July 2022, stabilized while the expansion flow becomes destabilized with the increase of the constriction length.
Accepted 16th September 2022 In contrast, the entire flow becomes less stable in deeper channels for the shear-thinning xanthan gum
DOI: 10.1039/d2sm01024k (XG) solution as well as the shear thinning and viscoelastic polyacrylamide (PAA) solution. This
observation aligns with that of water flow, which is attributed to the reduced top/bottom wall stabilizing
rsc.li/soft-matter-journal effects.

1. Introduction published papers have also investigated the effects of constric-

tion configuration (e.g., contraction ratio, corner rounding,
Flow of polymer solutions through extensional microstructures, constriction length, etc.) on the contraction and/or expansion
where the cross-sectional area changes along the length of the polymer flow.18–26 In addition, many other papers have studied
fluid passage, can induce non-linear viscoelastic responses.1–3 the effects of fluid rheological properties such as shear thin-
Such scenarios are frequently encountered in many engineer- ning/thickening and viscoelasticity.28–36
ing fields such as petroleum engineering, groundwater reme- The length and width of the contraction and/or expansion
diation, point-of-care diagnostics, etc.4–7 Understanding the microchannels in previous studies are usually kept much larger
nature of polymer flows in contraction and/or expansion micro- than the depth constructing a pseudo-Hele-Shaw configuration.
channels thus becomes important in making the performance Regardless, earlier studies have shown that the path of a fluid
and design decisions. A series of numerical and experimental packet even for a simple water solution can be complex
studies have been performed over the years to classify the flow and three dimensional mainly due to the inter-planar fluid
responses under various conditions.8–11 Some of these studies exchange in such quasi-two-dimensional structures.37,38
have considered three-dimensional axisymmetric or square Therefore, two-dimensional models are often insufficient to
constriction geometries.12,13 Much more of them have focused correctly predict such flows, sometimes even qualitatively
on two dimensional planar channels due to their relevance in unacceptable.39–41 Moreover, bifurcation or asymmetry may
microfluidic applications and advantages in understanding the occur in the expansion Newtonian flow after a certain Reynolds
fundamentals by ruling out one dimension.14–17 Many of the number is reached in deep channels.42 It is because the
channel walls have been demonstrated to stabilize the fluid
a
flow,37,40,41 whose effects get weakened with the increase of
Department of Mechanical Engineering, Clemson University, Clemson, SC,
29634-0921, USA. E-mail: [email protected]
channel depth.43 These observations particularly raise the
b
College of Marine Engineering, Dalian Maritime University, Dalian, 116026, interest to learn how the viscoelastic responses in contraction
P. R. China. E-mail: [email protected] and/or expansion flows may be impacted by the channel depth.

This journal is © The Royal Society of Chemistry 2022 Soft Matter, 2022, 18, 7427–7440 | 7427
 View Article Online

Paper Soft Matter

Unfortunately, very few studies have so far shed light on this

topic, which are briefly summarized below.
Omowunmi and Yuan44 used a multi-mode Phan–Thien–
Tanner (PTT) model to investigate the three dimensional
viscoelastic flow of a semi-dilute polyethylene oxide (PEO)
solution through abrupt contraction microchannels with vary-
ing depths. They found that with the increase of the channel
depth, which in turn decreases the elasticity number, the
starting location of the vortices shifts from the lips to the Fig. 1 Schematic illustration with dimensions (left) along with the top
Published on 16 September 2022. Downloaded by Clemson University on 8/17/2023 8:45:26 PM.

salient corners of the contraction. They further reported that (upper right) and side (lower right) view pictures of the contraction–
varying the channel depth has similar effects on the contrac- expansion region of the prepared constriction microchannels. The depth
of the microchannel, h, is varied from 30 mm to 50 mm and 90 mm in the
tion flow to varying the fluid properties as the latter also
experiment.
influences the elasticity number. Hidema et al.45 experimen-
tally studied the effects of polymer flexibility and entanglement
on the flow of dilute and semi-dilute sodium hyaluronate 2. Experiment
solutions with and without salt in planar contraction–expan-
2.1. Materials
sion microchannels. They found that for alike elasticity num-
bers, the unstable flow regimes are encountered earlier in the Fig. 1 shows the schematic and pictures of the contraction–
contraction of the deeper channel regardless of the polymer expansion region of the microchannels used in our tests. The
rigidity (which depends on the ionic strength of the solvent) and microchannel molds were prepared with SU-8 25 photoresist
entanglement (which depends on the polymer concentration) (MicroChem, Newton, MA, USA) by standard photolithography.
in the solution. In another study that is not directly related to The microchannels were then made from these molds with
the contraction or expansion flow, Cruz et al.46 conducted polydimethylsiloxane (PDMS) by soft lithography. Further
creeping flow simulations of upper-convected Maxwell and details about the channel fabrication can be found in our
simplified PTT fluids in three dimensional planar cross-slots. previous work.48 The prepared microchannels are all 2 cm long
They found that increasing the Deborah number (equivalent to and 500 (10) mm wide with a 250 mm long and 50 (2) mm
increasing the flow rate) changes the flow from steady sym- wide constriction in the middle. These channels differ only in
metric to time dependent in shallow channels. In contrast, the the vertical direction having the depths of 30 mm, 50 mm and
flow in deep channels first bifurcates from steady symmetric 90 mm, yielding the aspect ratio [depth/width, see the definition
to steady asymmetric and then becomes unsteady with the in eqn (5)] of 0.6, 1.0 and 1.8, respectively, for the constriction.
increase of the Deborah number. Three types of polymer solutions were prepared in deionized
The existing scarce studies are insufficient to elucidate the (DI) water (Fisher Scientific, Waltham, MA, USA): 2000 ppm XG
correlation between fluid rheology and the effect of channel solution (molecular weight, Mw E 2 MDa, and overlap concen-
depth on the contraction/expansion flow of non-Newtonian tration, c* = 254 ppm,49 Tokyo Chemical Industry, Tokyo,
fluids. Hence, we investigate here the flow of three types Japan) is a shear thinning fluid with a negligible elasticity;50
of polymer solutions in three planar constriction micro- 1000 ppm PEO solution (Mw = 2 MDa and c* = 858 ppm,49
channels differing only in their depths. The polymer Sigma-Aldrich, St. Louis, MO, USA) is a viscoelastic fluid with
solutions have distinct rheological properties including the almost no shear thinning effect51 (i.e., Boger fluid52); 200 ppm
shear thinning xanthan gum (XG) solution, viscoelastic PEO PAA solution (poly(acrylamide/sodium acrylate) [70 : 30], Mw =
solution, and the both shear thinning and viscoelastic poly- 18 MDa and c* = 104 ppm,49 Polysciences, Warrington, PA, USA)
acrylamide (PAA) solution. These solutions are noted to is a both shear thinning and viscoelastic fluid.53 The chain
have been frequently used for industrial and microfluidic conformations in all these polymer solutions are in their
applications.3–8,18,19,27 They are each tested along with water respective overlap states, and hence the solutions are in their
from the non-inertial to inertia-dominated flow regime in corresponding semi-dilute regimes.49 DI water was also tested
every channel for a systematic understanding of both the sole as the control experiment for the effect of fluid inertia alone.
and combined effects of fluid shear thinning, elasticity and Fig. 2 shows the shear viscosities of the prepared polymer
inertia on the channel depth-dependent instabilities. Small solutions that were measured using a cone–plate rheometer
fluorescent particles are seeded into each of the prepared (Anton Paar, MCR 302, Graz, Austria) at room temperature. The
solutions for a direct visualization of the flow pattern in the Carreau model54 was used to fit the measured viscosity data of
contraction–expansion region of the microchannels. The flow the shear thinning XG and PAA solutions,
development such as vortices and instabilities in each Z  Z1 h iðn1Þ=2
solution is characterized in dimensionless parameter spaces ¼ 1 þ ðlc g_ Þ2 (1)
Z0  Z1
among the three channels. It is anticipated that our obtained
experimental data will advance the fundamental understanding where Z is the fluid viscosity, ZN is the infinite-shear-
and benefit the numerical modeling of non-Newtonian fluid rate viscosity, Z0 is the zero shear-rate viscosity, lc is a time
flows.47 constant, g_ is the fluid shear rate, and n is the power-law index

7428 | Soft Matter, 2022, 18, 7427–7440 This journal is © The Royal Society of Chemistry 2022
 View Article Online

Soft Matter Paper

lower values in the range to adapt to the faster movements of

the tracer particles. The post processing of the obtained images
was done on the Nikon imaging software (NIS Elements
AR 3.22).
The effect of fluid inertia on the flow is characterized by the
Reynolds number,
rVDh 2rQ
Re ¼   ¼   (2)
Z g_ Z g_ ðw þ hÞ
Published on 16 September 2022. Downloaded by Clemson University on 8/17/2023 8:45:26 PM.

where r is the fluid mass density (assumed equal to that of the

suspending water for all the solutions), V is the average fluid
velocity in the constriction, Dh = 2wh/(w + h) is the hydraulic
diameter of the constriction with w = 50 mm and h being its
 
width and depth, Z g_ is the fluid viscosity at the characteristic
Fig. 2 Experimentally measured (symbols) and theoretically fitted (lines,
shear rate, g_ ¼ 2V=w, across the width of the constriction, and
with the Carreau model) viscosity data of the prepared polymer solutions.
The vertical dashed-dotted line highlights the characteristic shear rate Q is the volumetric flow rate. The effect of fluid elasticity is
inside the constriction for the flow rate of 0.1 mL h1 in the 50 mm deep characterized by the Weissenberg number,
contraction–expansion microchannel.
2lQ
Wi ¼ l_g ¼ (3)
w2 h
Table 1 Summary of the fitting parameters in the Carreau model and the The relative importance of fluid elasticity over inertia is char-
relaxation times for the prepared polymer solutions. The elasticity number, acterized by the elasticity number,
El, was calculated from eqn (4) for the flow rate of 10 mL h1 in the 50 mm
deep constriction microchannel Wi lZðg_ Þðw þ hÞ
El ¼ ¼ (4)
Re rw2 h
Solution Z0 (mPa s) ZN (mPa s) lC (ms) n l (ms) El
Water 1.0 1.0 — 1 0 0
whose value for each of our prepared fluids at one of the
XG 1740 1.8 6.6 0.33 B0a B0 moderate flow rates tested, Q = 10 mL h1, in the 50 mm deep
PEO 2.2 2.2 — B1 1.5b 2.88 microchannel is listed in Table 1. The corresponding charac-
PAA 4900 2.1 151 0.25 95c 162
teristic shear rate, g_ , inside the constriction is around 4.4 
a
Haase et al.55 b
Rodd et al.56 c Poole and Escudier.57 104 s1, for which the three polymer solutions have similar
values of viscosity and hence Re. The effect of fluid shear
characterizing the shear thinning effect. We also attempted to thinning is characterized by the power-law index, n, in the
measure the relaxation times, l, of the polymer solutions, but Carreau model. The effect of vertical confinement on the flow is
were unable to obtain consistent data using our rheometer. compared in terms of the aspect ratio of the constriction,
We therefore extracted the values from the earlier papers55–57 h
AR ¼ (5)
that reported the relaxation times for the same solutions as in w
this work. Table 1 summarizes the fitting parameters in the
Carreau model and the relaxation times for the prepared
solutions.
3. Results and discussion
2.2. Methods 3.1. Newtonian water
To visualize the flow pattern in the contraction–expansion Fig. 3(a) shows the snapshots for the flow of DI water at the
region of the microchannels, we seeded 1 mm diameter fluores- expansion part of the constriction microchannels having three
cent particles (0.05% v/v, Sigma-Aldrich, St. Louis, MO, USA) different depths. In the shallowest channel with AR = 0.6, lip
into each of the prepared solutions. A syringe pump (KD vortices start at around the flow rate of 7 mL h1 with Re = 48.6,
Scientific, Holliston, MA, USA) was used to drive the particle and then extend to the salient corners at around 20 mL h1
suspension through the channels. A plastic tube was inserted with Re = 139. With the further increase of fluid inertia, the
into the inlet reservoir, which carried the fluids from the symmetric corner vortices grow in size along the channel length
syringe pump to the channel, and another one into the outlet with the cores moving downstream. In contrast, the lip vortices
reservoir, which led the fluid out of the channel to a vial. The appear at a smaller Re = 27.8 in the deeper AR = 1.0 channel.
motion of tracer particles was recorded in the constriction The full extension till the salient corners happens at around
regions through an inverted fluorescent microscope (Nikon Re = 55.6, which remain symmetric while growing till the
Eclipse TE2000U, Nikon Instruments) with a CCD camera highest flow rate tested in this channel with Re = 278. In the
(Nikon DS-Qi1Mc). The exposure time of the camera was varied deepest channel with AR = 1.8, the lip and corner vortices each
between 0.5 s and 3 s for the captured snapshots. It was set at set off at an even lower Re than that in the AR = 1.0 channel.

This journal is © The Royal Society of Chemistry 2022 Soft Matter, 2022, 18, 7427–7440 | 7429
 View Article Online

Paper Soft Matter

Published on 16 September 2022. Downloaded by Clemson University on 8/17/2023 8:45:26 PM.

Fig. 3 The expansion flow of Newtonian water in constriction microchannels having three different depths: (a) snapshot images, where the highlighted
dimension, Lv, on the top-row image indicates the measurement of vortex length in the flow direction and the dashed lines on the bottom-row images
highlight the flow asymmetries; (b) plot of the measured vortex length, Lv (normalized by the width of the microchannel) against Re, where the solid lines
are the trendlines best fitted for the experimental data (symbols) and used here to guide the eyes only; (c) summary of the flow events in the
dimensionless AR–Re space: ‘circles’ for no flow disturbances, ‘triangles’ for lip vortices, ‘squares’ for symmetric corner vortices, and ‘crosses’ for
asymmetric corner vortices. The arrow on the chart indicates that the fluid inertia-induced expansion-flow vortices in water get destabilized in deeper
channels.

However, the flow symmetry breaks down at the flow rate of 3.2. Shear thinning XG solution
50 mL h1 with Re = 198 though the flow still remains Fig. 4(a) shows the snapshots for the contraction-flow of XG
temporarily steady. This asymmetric expansion flow becomes solution in the three constriction microchannels. In the shallow-
more pronounced with the increase of Re, consistent with the est channel with AR = 0.6, lip vortices are formed at 0.01 mL h1
previous reports.37,40 Fig. 3(b) compares the measured vortex with Re = 0.00238, which is significantly earlier than the onset
lengths, Lv (normalized by the width of the microchannel), of lip vortices in the expansion water flow because of the
against Re in the three channels. Overall, the vortices start XG solution’s strong shear thinning effect. These lip vortices
earlier and the vortex length is greater throughout the range of extend till the side walls at Re = 0.0906 and grow until Re = 22.7,
Re in the deeper channels. Fig. 3(c) presents the flow events in at which the corner vortices remain confined in the horizontal
the dimensionless AR–Re space. The onset of fluid inertia- plane of the microchannel. The recirculations are observed to
induced vortices is pushed forward to a smaller Re in a deeper overlap at Re = 33.6 exhibiting an apparently 3D pattern. They,
channel, indicating that increasing the channel depth destabi- however, remain approximately symmetric and do not seem to
lizes the inertial flow of water because of the reduced top/ further grow in size with the increase of Re. A similar develop-
bottom wall stabilizing effects.37–43 ment of flow events is also observed in the deeper AR = 1.0 and

7430 | Soft Matter, 2022, 18, 7427–7440 This journal is © The Royal Society of Chemistry 2022
 View Article Online

Soft Matter Paper

Published on 16 September 2022. Downloaded by Clemson University on 8/17/2023 8:45:26 PM.

Fig. 4 The contraction flow of shear-thinning XG solution in constriction microchannels having three different depths: (a) snapshot images; (b) plot of
the measured vortex length (normalized by the width of the microchannel) against Re, where the solid lines are the trendlines best fitted for the
experimental data (symbols) and used here to guide the eyes only; (c) summary of the flow events in the AR–Re space: ‘triangles’ for lip vortices, ‘squares’
for 2D corner vortices, and ‘filled crosses’ for 3D overlapping corner vortices. The arrow on the chart indicates that the fluid shear thinning induced
contraction flow vortices in the XG solution get destabilized in deeper channels.

AR = 1.8 channels except that the value of Re for each of these channel depth stabilizes the shear thinning-induced contraction-
events decreases with the increase of channel depth. Fig. 4(b) flow vortices because of the increasing effect of vertical
compares the measured vortex lengths in the three channels, confinement.
where only the data for the visually 2D vortices are included. Fig. 5(a) shows the snapshots for the expansion XG flow in
The vortices appear earlier in terms of Re in the deeper the three channels. For the shallowest AR = 0.6 channel, the
channels. Their sizes also get larger in the deep channels for streamlines are observed to slightly separate from the lips
the same Re. However, the vortex length seems to be restricted and salient corners of the expansion at around Re = 33.6
to around 1.5 times the channel width in all three channels (10 mL h1). These flow separations merge on each side of
before the vortices start exhibiting 3D overlapping patterns. the expansion to form corner vortices at Re = 70.6. The increase
Fig. 4(c) presents the observed flow events in the AR–Re space, in Re from this point forth induces a growth of the vortices
which are clearly seen to be pushed back in Re for the shallower along the channel length only. However, the cores of these
channels. This observation indicates that decreasing the vortices do not apparently shift downstream like in the water

This journal is © The Royal Society of Chemistry 2022 Soft Matter, 2022, 18, 7427–7440 | 7431
 View Article Online

Paper Soft Matter

Published on 16 September 2022. Downloaded by Clemson University on 8/17/2023 8:45:26 PM.

Fig. 5 The expansion flow of shear-thinning XG solution in constriction microchannels having three different depths: (a) snapshot images; (b) plot of the
measured vortex length (normalized by the width of the microchannel) against Re, where the solid lines are the trendlines best fitted for the experimental
data (symbols) and used here to guide the eyes only; (c) summary of the flow events in the AR–Re space: ‘rectangles’ for bending streamlines, ‘triangles’
for lip spikes, and ‘squares’ for corner vortices. The arrow on the chart indicates that the expansion flow of shear-thinning XG solution gets destabilized in
deeper channels.

flow with the increase of Re because of the influence of fluid the critical values.58 This discrepancy may be a consequence of
shear thinning. A similar trend of such flow development is the absent top and bottom walls from a 2D model, which have been
also observed in the deeper channels, where the flow bending demonstrated to destabilize the flow of Newtonian fluids.37–43
and vortices both grow stronger with the increase of channel Fig. 5(b) shows the measured vortex length that increases with Re
depth for comparable values of Re. Moreover, the flow still in each channel and becomes larger in the deeper channels for
maintains the symmetry in the deepest channel for the range of comparable values of Re like in the water flow. A summary of the
flow rates tested. This observation is different from that in the flow events is presented in Fig. 5(c), where a similar flow develop-
inertial water flow because of probably the reported stabilizing ment is viewed in each channel though increasing AR promotes the
effect of fluid shear thinning.58 It seems also inconsistent with occurrences of these events. Therefore, the expansion and contrac-
a recent numerical prediction on the expansion flow of inelastic tion flows of XG solution are both destabilized in deeper channels
shear-thinning fluids in a 2D planar channel, where a steady with enhanced vortex formation and growth, where the effect of
asymmetric flow and even a third vortex may occur if Re reaches vertical confinement becomes weaker.

7432 | Soft Matter, 2022, 18, 7427–7440 This journal is © The Royal Society of Chemistry 2022
 View Article Online

Soft Matter Paper

3.3. Viscoelastic PEO solution corners at Re = 14.5. It grows bigger with the increase of the
Fig. 6(a) shows the images for the contraction flow of PEO flow rate and turns into pseudo-symmetric corner vortices at
solution in the three constriction microchannels. As this Re = 28.9. These double vortices continue growing in size, and
solution is approximately a Boger fluid52 with a constant become asymmetric and bi-stable at Re = 57.9. A visually
viscosity, the elasticity number becomes independent of flow chaotic flow is observed at Re = 116. The AR = 1.0 channel sees
rate and decreases from El = 3.8 to 2.9 and 2.2 with the a similar development of flow events while the corresponding
increasing channel depth. In the shallowest AR = 0.6 channel, Re values are all greater than in the shallower one. Moreover,
diverging streamlines appear at 3 mL h1 with Re = 8.68 and the double vortices therein remain roughly symmetric till the
Wi = 33.3. A single vortex is then formed from one of the re-entrant highest Re = 185 tested. In contrast, the single vortex in the
Published on 16 September 2022. Downloaded by Clemson University on 8/17/2023 8:45:26 PM.

Fig. 6 The contraction flow of viscoelastic PEO solution in constriction microchannels having three different depths: (a) snapshot images; (b) plot of the
measured vortex length (normalized by the width of the microchannel) against Re, where the solid lines are the trendlines best fitted for the experimental
data (symbols) and used here to guide the eyes only; (c) summary of the flow events in the AR–Re space: ‘circles’ for no flow disturbances, ‘rectangles’ for
bending/diverging streamlines, ‘pluses’ for single vortices, ‘crosses’ for symmetric or asymmetric corner vortices, and ‘asterisks’ for unsteady flows.
The arrow on the chart indicates that the contraction flow of viscoelastic PEO solution gets destabilized in shallower channels.

This journal is © The Royal Society of Chemistry 2022 Soft Matter, 2022, 18, 7427–7440 | 7433
 View Article Online

Paper Soft Matter

deepest AR = 1.8 channel never turns into double vortices in our the flow events in the AR–Re space. The flow development is seen
experiment. Fig. 6(b) compares the measured average lengths to take place at lower Re as the channel depth gets smaller,
of stable vortices in the three channels. The vortices start indicating the promoted instability of PEO solution in shallower
appearing at lower Re in shallower channels and their sizes channels.
increase as the channel depth is decreased for the same Re. Fig. 7(a) shows the snapshot images for the expansion PEO
This trend is opposite to that for the fluid shear thinning- flow in the three channels. Symmetric vortices start from
induced contraction-flow vortices in the XG solution (see Fig. 4), the salient corners at around Re = 43.4 (15 mL h1) in the
and as well to those for the inertial expansion-flow vortices in shallowest AR = 0.6 channel. They grow on to extend till the lips
water (see Fig. 3) and XG solution (see Fig. 5). Fig. 6(c) summarizes of the expansion becoming full symmetric corner vortices at
Published on 16 September 2022. Downloaded by Clemson University on 8/17/2023 8:45:26 PM.

Fig. 7 The expansion flow of viscoelastic PEO solution in constriction microchannels having three different depths: (a) snapshot images; (b) plot of the
measured vortex length (normalized by the width of the microchannel) against Re, where the solid lines are the trendlines best fitted for the experimental
data (symbols) and used here to guide the eyes only; (c) summary of the flow events in the AR–Re space: ‘circles’ for no flow disturbances, and ‘diamonds’
for vortices starting from the salient corners. The arrow on the chart indicates that the expansion flow of viscoelastic PEO solution gets destabilized in
shallower channels.

7434 | Soft Matter, 2022, 18, 7427–7440 This journal is © The Royal Society of Chemistry 2022
 View Article Online

Soft Matter Paper

Re = 116. Their size along the channel length mildly increases They remain symmetric over the range of flow rates tested
after this point with the increase of Re. The deeper AR = 1.0 and because of the reported fluid elasticity-induced suppression
AR = 1.8 channels see a similar flow development except that effect on the expansion flow.33 Their cores shift downstream as
the value of Re for each event decreases in deeper channels. It is Re increases though with a much weaker dependence on Re
noted that the corner vortices in each of these channels are than those in the water flow. Fig. 7(b) compares the measured
visually much smaller than those in the water and XG flows. vortex lengths in the three channels. Similar to the water and
Published on 16 September 2022. Downloaded by Clemson University on 8/17/2023 8:45:26 PM.

Fig. 8 The contraction flow of the both shear thinning and viscoelastic PAA solution in constriction microchannels having three different depths:
(a) snapshot images, on which the dashed lines highlight the flow asymmetries; (b) plot of the measured vortex length (normalized by the width of the
microchannel) against Re, where the solid lines are the trendlines best fitted for the experimental data (symbols) and used here to guide the eyes only;
(c) summary of the flow events in the AR–Re space: ‘squares’ for symmetric corner vortices, ‘crosses’ for asymmetric fluctuating vortices, and ‘asterisks’
for unsteady flow. The arrow on the chart indicates that the contraction flow of the both shear thinning and viscoelastic PAA solution gets destabilized in
deeper channels.

This journal is © The Royal Society of Chemistry 2022 Soft Matter, 2022, 18, 7427–7440 | 7435
 View Article Online

Paper Soft Matter

XG flow cases, the vortex size in the PEO solution also increases our recent observations of constriction length-dependent
with Re in each channel and becomes greater in deeper instabilities in the PEO solution,49 where the expansion flow
channels. Fig. 7(c) presents the flow events in the AR–Re space, gets destabilized while the contraction flow becomes more
which exhibit a similar development in each channel except stable in channels with longer constrictions.
that the salient corner vortices occur at smaller Re in deeper
channels. Therefore, the expansion flow of PEO solution 3.4. Shear thinning and viscoelastic PAA solution
becomes less stable in deeper channels with enhanced vortex Fig. 8(a) shows the contraction-flow images for the PAA solution
formation and growth while the contraction flow gets stabi- in the three constriction microchannels. In the shallowest AR =
lized. These opposing trends are noted to be consistent with 0.6 channel, large symmetric corner vortices already appear in
Published on 16 September 2022. Downloaded by Clemson University on 8/17/2023 8:45:26 PM.

Fig. 9 The expansion flow of the both shear thinning and viscoelastic PAA solution in constriction microchannels having three different depths:
(a) snapshot images; (b) plot of the measured vortex length (normalized by the width of the microchannel) against Re, where the solid lines are the
trendlines best fitted for the experimental data (symbols) and used here to guide the eyes only; (c) summary of the flow events in the AR–Re space:
‘circles’ for no flow disturbances, ‘rectangles’ for bending streamlines, ‘triangles’ for lip spikes, and ‘squares’ for corner vortices. The arrow on the chart
indicates that the expansion flow of the both shear thinning and viscoelastic PAA solution gets destabilized in deeper channels.

7436 | Soft Matter, 2022, 18, 7427–7440 This journal is © The Royal Society of Chemistry 2022
 View Article Online

Soft Matter Paper

an inertialess flow at Re = 0.0105 (0.01 mL h1), which should 3.5. Comparison among the fluids
arise from the stronger shear thinning effect of the PAA To obtain an insight of the general role of channel depth in the
solution than the XG solution (see Fig. 4). The size of the contraction–expansion flow, we compare the experimental data
corner vortices keeps growing in the channel length direction across the tested fluids with respect to the channel aspect ratio,
till Re = 3.10 (1 mL h1, Wi = 704, El = 227) when the flow AR. Fig. 10(a) shows the threshold Reynolds number, Reth, for
asymmetry starts. After this point the vortices fluctuate asym- the vortex formation in the contraction XG (corner vortices),
metrically without any period and the flow becomes unsteady at PEO (single vortices), and PAA (asymmetric fluctuating vortices)
Re = 16.2. The flow events are similar for the deeper AR = 1.0 flows. We cannot pick Reth for the onset of symmetric corner
and AR = 1.8 channels with larger vortices at the same flow rates vortices in the PAA solution for this comparison because
Published on 16 September 2022. Downloaded by Clemson University on 8/17/2023 8:45:26 PM.

as well as earlier onset of fluctuating and chaotic flows than in they take place at a much lower flow rate than the smallest
the lower AR channels. Fig. 8(b) compares the measured vortex (0.01 mL h1) we have tested. It is apparent that increasing the
lengths among the three channels. The vortices start at smaller channel depth or AR leads to a decrease in Reth for the shear
values of Re in the deeper channels and have larger sizes for the thinning XG and PAA solutions, opposite to the increasing
same Re. The length of these vortices appears to be limited to trend for the viscoelastic Boger PEO solution. Such an observa-
approximately twice the channel width in all three channels tion indicates the opposite influences of fluid shear thinning
before the fluctuating flow sets off. A similar vortex length and elasticity on the contraction flow vortices, between which
threshold is also observed in the XG flow though at a smaller the shear thinning effect should be stronger33 because the PAA
value of 1.5 times the channel width because it has a weaker solution is more viscoelastic than the PEO solution. This
shear thinning effect than the PAA solution and hence smaller phenomenon also explains why the decreasing trend for the
contraction-flow vortices. Fig. 8(c) presents the observed flow XG solution is a little steeper than that for the more shear
events in the AR–Re space, where those in the shallower thinning PAA solution because the latter is both shear thinning
channels are pushed back to greater Re. Such a trend aligns and viscoelastic. Fig. 10(b) shows Reth for the expansion vortex
with the observation in the contraction XG flow while opposite formation in the water, XG, PEO, and PAA flows. Increasing the
to that in the contraction PEO flow, which may validate our channel depth leads to a decrease in Reth regardless of the fluid
earlier hypothesis that fluid shear thinning plays a stronger role rheology because of the reduced wall stabilizing effect on the
than elasticity in the contraction flow.49
Fig. 9(a) shows the snapshots for the expansion PAA flow in
the three channels. The elasticity number does not change
significantly for Re 4 15 in each channel and decreases from
approximately El = 213 to 160 and 125 with the increase of the
channel depth. In the shallowest AR = 0.6 channel, bending
streamlines start at around Re = 32.7 (10 mL h1) and turn into
lip spikes at Re = 65.7. Symmetric vortices then appear in the
salient corners at Re = 82.2 and merge with the lip spikes at
Re = 98.7. Thus formed corner vortices grow with the further
increase of flow rate while maintaining the symmetry through-
out the test. These flow events also take place in the deeper
AR = 1.0 and AR = 1.8 channels though at smaller values of Re
with the increase of the channel depth. Such a trend agrees
with those in the expansion flow of water, XG and PEO
solutions, indicating the dominant role of fluid inertia over
shear thinning and elasticity. Interestingly the expansion-flow
vortices in the PAA solution are viewed to resemble those in the
XG solution. Their cores, however, seem to slowly shift down-
stream at increasing Re like in the PEO solution. Fig. 9(b)
compares the measured vortex lengths in the three channels.
Similar to the observations in the expansion flow of all other
tested fluids, the vortex length in the PAA solution also
increases with Re in each channel and grows larger in deeper
channels for the same Re. Fig. 9(c) presents the flow events in
the AR–Re space, where a similar flow development is once
again viewed in each channel except that all events are delayed
to higher Re in shallower channels. Therefore, the expansion
Fig. 10 Comparison of the threshold Reynolds number, Reth, for the
and contraction flows of PAA solution are both destabilized in vortex formation in the contraction (a) and expansion (b) flows of the
deeper channels because of the weakened vertical confinement tested fluids in constriction microchannels having different depths. The lines
effect, in agreement with the XG solution. are used to guide the eyes only.

This journal is © The Royal Society of Chemistry 2022 Soft Matter, 2022, 18, 7427–7440 | 7437
 View Article Online

Paper Soft Matter

inertia-dominated expansion flow.37–43 Moreover, it is obvious only slightly smaller than those in water, consistent with our
that fluid shear thinning (XG) and elasticity (PEO) both sup- recent observations.26,34 In contrast, the vortices in the PEO
press the formation of expansion vortices in water, which add solution are significantly smaller, which may again indicate the
together to further delay the vortex formation in the PAA stronger impact of fluid elasticity than shear thinning on the
solution. expansion flow. It is, however, unclear to the authors why
Fig. 11(a) shows the vortex length in the contraction flow of the vortices in the PAA solution, which is more shear thinning
XG and PAA solutions at Re = 0.1 as well as the contraction than the XG solution and more viscoelastic than the PEO
vortex length in the PEO solution at Re = 40. Different values of solution, are not the smallest among the tested fluids.
Re are used here for the comparison because the fluid shear
Published on 16 September 2022. Downloaded by Clemson University on 8/17/2023 8:45:26 PM.

thinning induced contraction flow vortices occur at a much

lower flow rate than the fluid elasticity induced ones. The data
are extracted from the fitted trendlines for the experimentally 4. Conclusions
measured vortex length. Increasing the channel depth or AR
enhances the vortex formation in the XG and PAA solutions We have conducted a set of experiments to study the effect of
while suppressing that in the PEO solution. Moreover, the channel depth on the flow of non-Newtonian XG, PEO, and PAA
vortices in the PAA solution are larger than those in the less solutions as well as Newtonian water in planar constriction
shear thinning XG solutions in all depths of channels. microchannels. The tested Reynolds number (so is the
Fig. 11(b) compares the expansion vortex lengths in the four Weissenberg number if applicable) spans multiple orders of
fluids at Re = 100, which all grow with the increase of channel magnitude covering the creeping flow to moderately high
depth because of the reduced wall confinement effect. More- inertial flow. We find that increasing the channel depth desta-
over, the vortices in water are the largest among all the fluids, bilizes the expansion flow of all four fluids with enhanced
indicating the suppressing effect of fluid shear thinning and vortex formation and growth because of probably the weakened
elasticity as noted above. The vortices in the XG solution are stabilizing effect from the vertical confinement. This observa-
tion may also indicate the dominant role of fluid inertia over
both shear thinning and elasticity in the expansion flow.
Increasing the channel depth also destabilizes the contraction
flow of the shear thinning XG and PAA solutions while stabili-
zing that of the viscoelastic PEO solution. These trends may
validate our earlier hypothesis that fluid shear thinning plays a
stronger role than elasticity in the contraction flow.33 Interest-
ingly, we note that the opposing influences of the channel
depth on the contraction and expansion flows of PEO solution
seem to align with those of the constriction length for the same
solution in our recent study.49 We hope that our experimental
results presented here will stimulate more studies on similar
topics, particularly on the theoretical and numerical under-
standing of pertinent polymer flows. For future work we will
vary the molecular weight and concentration of PEO polymer
for an improved understanding of the correlation between fluid
rheology and the effect of vertical confinement on polymer
flows in contraction and expansion microchannels.

Author contributions
M. K. R. and S. W. performed the experiment; H. D. and M. B.
helped with the experiment; Y. S. and X. X. designed and
supervised the project; M. K. R. wrote the manuscript;
all authors commented on the manuscript; X. X. edited the
manuscript.

Fig. 11 Comparison of the vortex length (normalized by the width of the

microchannel) in the contraction (a) and expansion (b) flows of the tested
fluids at the same values of Re (highlighted on the plots) in constriction Conflicts of interest
microchannels having different depths. The lines are used to guide the
eyes only. There are no conflicts to declare.

7438 | Soft Matter, 2022, 18, 7427–7440 This journal is © The Royal Society of Chemistry 2022
 View Article Online

Soft Matter Paper

Acknowledgements 19 C. A. Browne, A. Shih and S. S. Datta, Small, 2020, 16,

1903944.
This work was supported in part by Clemson University through 20 M. A. Alves, P. J. Oliveira and F. T. Pinho, J. Non-Newtonian
a SEED grant and Creative Inquiry Program (X. X.), China Fluid Mech., 2004, 122, 117–130.
Scholarship Council (CSC) – Chinese Government Graduate 21 L. E. Rodd, J. J. Cooper-White, D. V. Boger and
Student Overseas Study Program (S. W.), and University 111 G. H. McKinley, J. Non-Newtonian Fluid Mech., 2007, 143,
Project of China under Grant number B08046 (Y. S.). 170–191.
22 L. E. Rodd, D. Lee, K. H. Ahn and J. J. Cooper-White, J. Non-
References Newtonian Fluid Mech., 2010, 165, 1189–1203.
Published on 16 September 2022. Downloaded by Clemson University on 8/17/2023 8:45:26 PM.

23 A. Lanzaro and X. F. Yuan, J. Non-Newtonian Fluid Mech.,

1 M. A. Alves and R. J. Poole, J. Non-Newtonian Fluid Mech., 2011, 166, 1064–1075.
2007, 144, 140–148. 24 C. A. Browne, A. Shih and S. S. Datta, J. Fluid Mech., 2020,
2 F. J. Galindo-Rosales, L. Campo-Deaño, P. C. Sousa, V. M. 890, A2.
Ribeiro, M. S. N. Oliveira, M. A. Alves and F. T. Pinho, 25 M. Kumar, S. Aramideh, C. A. Browne, S. S. Datta and
Exp. Therm. Fluid Sci., 2014, 59, 128–139. A. M. Ardekani, Phys. Rev. Fluids, 2021, 6, 033304.
3 S. S. Datta, A. M. Ardekani, P. E. Arratia, A. N. Beris, 26 M. Dacus, M. K. Raihan, M. Baghdady, C. Gabbard, S. Wu,
I. Bischofberger, G. H. McKinley, J. G. Eggers, J. E. López- J. B. Bostwick, Y. Song and X. Xuan, Phys. Fluids, 2022, 34,
Aguilar, S. M. Fielding, A. Frishman, M. D. Graham, J. S. 032006.
Guasto, S. J. Haward, A. Q. Shen, S. Hormozi, A. Morozov, 27 X. Lu, C. Liu, G. Hu and X. Xuan, J. Colloid Interface Sci.,
R. J. Poole, V. Shankar, E. S. G. Shaqfeh, H. Stark, 2017, 500, 182–201.
V. Steinberg, G. Subramanian and H. A. Stone, Phys. Rev. 28 L. Campo-Deañoa, F. J. Galindo-Rosales, F. T. Pinho,
Fluids, 2022, 7, 080701. M. A. Alves and M. S. N. Oliveira, J. Non-Newtonian Fluid
4 A. Mohsenatabar Firozjaii and H. R. Saghafi, Petroleum, Mech., 2011, 166, 1286–1296.
2020, 6, 115–122. 29 A. Lanzaro, Z. Li and X. F. Yuan, Microfluid. Nanofluid., 2015,
5 A. C. Barbati, J. Desroches, A. Robisson and G. H. McKinley, 18, 819–828.
Annu. Rev. Chem. Biomol. Eng., 2016, 7, 415–453. 30 S. J. Haward, J. Page, T. A. Zaki and A. Q. Shen, Phys. Fluids,
6 F. Tian, Q. Feng, Q. Chen, C. Liu, T. Li and J. Sun, Microfluid. 2018, 30, 113101.
Nanofluid., 2019, 23, 68. 31 R. M. Matos, M. A. Alves and F. T. Pinho, Exp. Fluids, 2019,
7 J. Zhou and I. Papautsky, Microsyst. Nanoeng., 2020, 6, 113. 60, 145.
8 A. Anbari, H. T. Chien, S. S. Datta, W. Deng, D. A. Weitz and 32 B. L. Micklavzina, A. E. Metaxas and C. S. Dutcher, Soft
J. Fan, Small, 2018, 14, 1703575. Matter, 2020, 16, 5273–5281.
9 Y. Liu, K. Zografos, J. Fidalgo, C. Duchêne, C. Quintard, 33 S. Wu, M. K. Raihan, L. Song, X. Shao, J. B. Bostwick, L. Yu,
T. Darnige, V. Filipe, S. Huille, O. du Roure, M. S. N. Oliveira X. Pan and X. Xuan, J. Non-Newtonian Fluid Mech., 2021,
and A. Lindner, Soft Matter, 2020, 16, 9844–9856. 290, 104508.
10 C. Yuan, H.-N. Zhang, Y.-K. Li, X.-B. Li, J. Wu and F.-C. Li, 34 M. K. Raihan, P. P. Jagdale, S. Wu, X. Shao, J. B. Bostwick,
Proc. Inst. Mech. Eng., Part C, 2020, 234, 4390–4414, DOI: X. Pan and X. Xuan, Micromachines, 2021, 12, 836.
10.1177/0954406220922863. 35 F. Khalkhal and S. Muller, Phys. Rev. Fluids, 2022, 7, 023303.
11 C. C. Hopkins, S. J. Haward and A. Q. Shen, Soft Matter, 36 C.-D. Xue, Z.-Y. Zheng, G.-S. Zhang, D.-W. Zhao and K.-R.
2022, 18, 4868–4880. Qin, Soft Matter, 2022, 18, 3867–3877.
12 S. Nigen and K. Walters, J. Non-Newtonian Fluid Mech., 2002, 37 T. P. Chiang, T. W. H. Sheu and S. K. Wang, Comput. Fluids,
102, 343–359. 2000, 29, 467–492.
13 F. Pimenta, K. Toda-Peters, A. Q. Shen, M. A. Alves and 38 C. A. Klettner and F. T. Smith, J. Fluid Mech., 2022, 934,
S. J. Haward, Exp. Fluids, 2020, 61, 204, DOI: 10.1007/ 1–29.
s00348-020-03036-z. 39 C.-H. Tsai, H.-T. Chen, Y.-N. Wang, C.-H. Lin and L.-M. Fu,
14 E. Miller and J. Cooper-White, J. Non-Newtonian Fluid Mech., Microfluid. Nanofluid., 2007, 3, 13–18.
2009, 160, 22–30. 40 M. S. N. Oliveira, L. E. Rodd, G. H. McKinley and M. A. Alves,
15 S. J. Haward, Z. Li, D. Lighter, B. Thomas, J. A. Odell and Microfluid. Nanofluid., 2008, 5, 809–826.
X.-F. Yuan, J. Non-Newtonian Fluid Mech., 2010, 165, 41 D. Li, L. Song, C. Zhang, L. Yu and X. Xuan, Phys. Fluids,
1654–1669. 2021, 33, 012002.
16 A. M. Afonso, P. J. Oliveira, F. T. Pinho and M. A. Alves, 42 E. Schreck and M. Schafer, Comput. Fluids, 2000, 29, 583–593.
J. Fluid Mech., 2011, 677, 272–304. 43 W. Cherdron, F. Durst and J. H. Whitelaw, J. Fluid Mech.,
17 K. Zografos, W. Hartt, M. Hamersky, M. S. N. Oliveira, 1978, 84, 13–31.
M. A. Alves and R. J. Poole, J. Non-Newtonian Fluid Mech., 44 S. C. Omowunmi and X. F. Yuan, Rheol. Acta, 2010, 49,
2020, 278, 104222. 585–595.
18 D. Yuan, Q. Zhao, S. Yan, S. Tang, G. Alici, J. Zhang and 45 R. Hidema, T. Oka, Y. Komoda and H. Suzuki, Phys. Fluids,
W. Li, Lab Chip, 2018, 18, 551–567. 2019, 31, 072005.

This journal is © The Royal Society of Chemistry 2022 Soft Matter, 2022, 18, 7427–7440 | 7439
 View Article Online

Paper Soft Matter

46 F. A. Cruz, R. J. Poole, A. M. Afonsoa, F. T. Pinho, P. J. Oliveira 54 R. B. Bird, R. C. Armstrong and O. Hassager, Dynamics of
and M. A. Alves, J. Non-Newtonian Fluid Mech., 2016, 227, 65–79. Polymeric Liquids, Wiley-Interscience: Hoboken, NJ, USA,
47 M. A. Alves, P. J. Oliveira and F. T. Pinho, Annu. Rev. Fluid 1987, vol. 1.
Mech., 2021, 53, 509–541. 55 A. S. Haase, J. A. Wood, L. M. J. Sprakel and
48 P. P. Jagdale, D. Li, X. Shao, J. B. Bostwick and X. Xuan, R. G. H. Lammertink, Inelastic non-Newtonian flow over
Micromachines, 2020, 11, 278. heterogeneously slippery surfaces, Phys. Rev. E, 2017, 95,
49 M. K. Raihan, S. Wu, Y. Song and X. Xuan, Soft Matter, 2021, 023105.
17, 9198–9209. 56 L. E. Rodd, T. P. Scott, D. V. Boger, J. J. Cooper-White and
50 M. A. Masuelli, J. Polym. Biopolym. Phys. Chem., 2014, 2, 37–43. G. H. McKinley, J. Non-Newtonian Fluid Mech., 2005, 129,
Published on 16 September 2022. Downloaded by Clemson University on 8/17/2023 8:45:26 PM.

51 V. Tirtaatmadja, G. H. McKinley and J. J. Cooper-White, 1–22.

Phys. Fluids, 2006, 18, 043101. 57 R. J. Poole and M. P. Escudier, J. Non-Newtonian Fluid Mech.,
52 D. F. James, Annu. Rev. Fluid Mech., 2009, 41, 129–142. 2004, 117, 25–46.
53 A. Lanzaro and X. F. Yuan, J. Non-Newtonian Fluid Mech., 58 S. Dhinakaran, M. S. N. Oliveira, F. T. Pinho and M. A. Alves,
2011, 166, 1064–1075. J. Non-Newtonian Fluid Mech., 2013, 198, 48–58.

7440 | Soft Matter, 2022, 18, 7427–7440 This journal is © The Royal Society of Chemistry 2022