Journal of Non-Newtonian Fluid Mechanics 290 (2021) 104508

Contents lists available at ScienceDirect

Journal of Non-Newtonian Fluid Mechanics

journal homepage: www.elsevier.com/locate/jnnfm

Polymer effects on viscoelastic fluid flows in a planar

constriction microchannel
Sen Wu a, b, 1, Mahmud Kamal Raihan b, 1, Le Song b, c, Xingchen Shao d, Joshua B. Bostwick b,
Liandong Yu c, e, Xinxiang Pan a, f, Xiangchun Xuan b, *
a
College of Marine Engineering, Dalian Maritime University, Dalian 116026, PR China
b
Department of Mechanical Engineering, Clemson University, Clemson, SC 29634-0921 USA
c
School of Instrument Science and Opto-electronic Engineering, Hefei University of Technology, Hefei 230009, PR China
d
Department of Chemical and Biomolecular Engineering, Johns Hopkins University, Baltimore, MD 21218, USA
e
College of Controlling Science and Engineering, China University of Petroleum, Qingdao 257061, PR China
f
Maritime College, Guangdong Ocean University, Zhanjiang 524088, PR China

A R T I C L E I N F O A B S T R A C T

Keywords: A comprehensive understanding of the flow of viscoelastic polymer solutions in a contraction and/or an
Polymer solutions expansion geometry concerns numerous applications. The effects of polymer type, molecular weight, and con­
Elasticity centration are investigated through controlled experiments with polyethylene oxide (PEO), polyvinylpyrrolidone
Shear thinning
(PVP) and hyaluronic acid (HA) solutions in a planar constriction microchannel in a wide range of Reynolds (Re)
Inertia
Microfluidic model
and Weissenberg (Wi) numbers. The polymer structure renders drastic differences amongst the contraction flows
Porous media of 3000 ppm PEO, PVP, and HA solutions despite having similar molecular weights. The expansion flow vortices
of these solutions remain analogous to those in the inertial flow of water. Increasing the molecular weight or
concentration of PEO polymer promotes the elastic instabilities in both the contraction and expansion flows. It,
however, suppresses (and even blocks) the inertial vortices in the expansion flow and makes them start from the
salient corners, in contrast to the lips of the expansion walls in water. Interestingly, a sudden decrease in the size
of inertial vortices is observed in the expansion flow of 500 ppm, 1 megadalton PEO solution when the elastic
disturbances start appearing in the contraction flow. The observed flow regimes and vortex development in the
polymer solutions are summarized in the same dimensionless Re − Wi and Re − χ L parameter spaces, where χ L is
the normalized vortex length. The elasticity number, El = Wi/Re, is found to determine and distinguish the
contraction and expansion flows.

1. Introduction important role in this regard since an abrupt alteration of shape along
the channel is very common in such processes [23–28]. Hence,
Polymer structural dynamics in the flow of non-Newtonian fluids numerous studies have been conducted over the decades for a better
holds significance in many naturally and industrially occurring phe­ grasp of the fluid rheological effects on the flow pattern through
nomena [1–3]. Depending on various inherent factors such as chain axisymmetric, planar or square contractions and/or expansions
length, flexibility, entanglement, ionic strength and relaxation of the [29–48]. There have also been a number of investigations on the influ­
conformation, flow of polymer solutions can behave in different man­ ence of the dimension and topology of the channel concerning the
ners while running through passages having different forms and shapes expansion/contraction ratio, aspect ratio, and length of the constriction
[4–11]. These behaviors often dictate the desired outcome of the pro­ [49–56]. The applications pertain to various analogous fields of flow
cess, may it be fiber spinning [12–14], flow mixing [15–18], or particle through actual porous media such as enhanced oil recovery from un­
separation in lab-on-a-chip devices [19–22]. The change in rheology due conventional reservoirs, inkjet printheads, industrial milling drag
to the shift of the flow regime from shear to extensional also plays an reducing practices, biological and chemical extraction and isolation

* Corresponding author.
E-mail address: xcxuan@clemson.edu (X. Xuan).
1
These authors contributed equally to this work.

https://doi.org/10.1016/j.jnnfm.2021.104508
Received 24 December 2020; Received in revised form 18 February 2021; Accepted 20 February 2021
Available online 23 February 2021
0377-0257/© 2021 Elsevier B.V. All rights reserved.
S. Wu et al. Journal of Non-Newtonian Fluid Mechanics 290 (2021) 104508

processes, and environmental remediation methods [57–59]. interaction with brines during the oil recovery in porous media because
Despite the importance, the influence of polymer structure at the a planar contraction-expansion structure is considered as one of the
molecular scale on the flow behavior at the micro/macroscale is a simplest laboratory models of actual pore-throat networks [28,59].
conundrum yet to be unraveled. The works of Daoud et al. [60], Moreover, our work can be used as the experimental data for numerical
Graessley [61], and Rubenstein and Colby [62] have shed light on the constitutive model improvements as well as microfluidic rheometry
significance of polymer structure on the rheology of neutral polymers. [73–75].
Charged polyelectrolyte solutions were investigated by Dobrynin and
Rubinstein [63], which provides insight on the interaction of charge 2. Experiment
repulsion and counterion condensation. Methods such as changing the
polymer concentration, adding salt to polyelectrolyte solutions to 2.1. Materials
neutralize them, and mixing charged surfactants to neutral polymer
solutions in order to have ions present, have been commonly used in Fig. 1 shows a picture of the planar abrupt contraction-expansion
recent studies to further comprehend the effects of polymer conforma­ microchannel used in the experiment. The length and width of the
tion in the channel flow phenomena. For example, Rodd et al. [64] main channel are 1 cm and 400 μm, respectively, with a uniform height
studied the entry flow of polyethylene oxide (PEO) solutions through a of 55 μm. The constriction is in the middle of the channel with a width of
micro-fabricated planar abrupt contraction-expansion. They demon­ 40 μm and a length of 200 μm. The microchannel was fabricated with
strated the strong influence of PEO concentration on the inertia-elastic polydimethylsiloxane (PDMS) using the standard soft lithography
flow development. Lanzaro et al. [65] observed chaotic-like flow pat­ technique [76]. Three primary types of aqueous polymer solutions
terns in a high-concentration polyacrylamide (PAA) solution while only including PEO (Sigma-Aldrich, St. Louis, MI), PVP (Sigma-Aldrich, St.
symmetric corner vortices appeared in the low-concentration solution. Louis, MI) and HA (Lifecore Biomedical LLC, Chaska, MN) were
Li et al. [66] reported the effect of (Patrick Knappe, 2010)polydispersity employed to investigate various polymer effects on the viscoelastic flow
on the elastic properties of PEO solutions having different molecular pattern in the constriction microchannel. For the effect of polymer type,
weights, and in turn the evolution of vortex formation and stability with we prepared 3000 ppm solutions of PEO [Mw = 0.3 megadalton (MDa)],
respect to the elasticity number. Miller and Cooper-White [67] investi­ PVP (Mw = 0.36 MDa) and HA (Mw = 0.357 MDa) polymers with similar
gated the transition of flow regimes in sodium dodecyl sulfate (SDS) molecular weights. DI water (Thermo Fisher Scientific, Waltham, MA)
surfactant added PEO solution that becomes shear thinning after was also tested as the control experiment. For the effect of polymer
aggregating with the anionic micelles. The effect of adding sodium salt molecular weight, we prepared 1000 ppm solutions of PEO polymer
to the PAA solution was studied by Campo-Deano et al. [68], where a with the molecular weight ranging from 0.3 to 0.6, 1, 2 and 4 MDa. For
change in the viscosity and shear thinning property was observed in the the effect of polymer concentration, we diluted the original 5000 ppm
salt solution with different vortex development ensued. Other groups solution of Mw = 1 MDa PEO polymer to 500, 1000, 1500, 2000 and
such as Kawale et al. [69] and Ekanem et al. [70] studied the salt effects 3000 ppm, respectively.
on the flow of hydrolyzed PAA solution. In addition, Hidema et al. [71] The dynamic viscosity of each of the prepared fluids was measured
found that the polymer entanglement has a stronger effect on the using a cone-plate rheometer (Anton Paar, MCR 302, Graz, Austria) at
constriction flow than the relaxation time and polymer rigidity with room temperature. The measured data for the shear rate ranging from 10
hyaluronic acid (HA) sodium salt dissolved in water and phosphate to 2,000 s− 1 are shown in Fig. 2. Those viscosity data for the fluids
buffer saline solution, respectively. exhibiting shear thinning behaviors are also curve-fitted using the Car­
The existing works have enhanced the comprehension of polymer reau model in order for a quantitative comparison of the power-law
conformation under flow conditions, but further generalized studies index, n [12],
concerning polymer structural properties are required in a single plat­ η − η∞ [ ](n−
(1)
1)/2
form to directly compare their influences and hence unify and bring the = 1 + (λC γ̇)2
η0 − η∞
results under the same roof. In a recent work, we have studied the fluid
rheological effects on the flow of four types of polymer solutions where, η∞ is the infinite-shear-rate viscosity, η0 is the zero-shear-rate
including xanthan gum (XG), polyvinylpyrrolidone (PVP), PEO and PAA
in a planar contraction-expansion microchannel [72]. These
non-Newtonian fluids have dissimilar molecular weights and concen­
trations, leading to significantly different elastic and shear thinning
properties. Their flows were compared among each other and also
against that of Newtonian water in a similar wide range of flow rates in
the same microchannel. It was found that fluid inertia and shear thin­
ning are responsible for the vortex development in the expansion and
contraction flows, respectively. However, we were unable to elucidate a
definitive effect of fluid elasticity from the four polymer solutions that
have significantly different polymer structures without a common
ground.
This work is aimed to provide further experimental data on the fluid
elasticity effects in the flow of polymer solutions through a planar
constriction microchannel. We investigate the respective influences of
polymer type, molecular weight and concentration by conducting
controlled experiments with PEO, PVP and HA solutions. Each test is
performed by varying only the parameter in question while keeping
everything else fixed. The flow development is characterized using
dimensionless numbers for the fluid elasticity, shear thinning, and
inertia effects, respectively. Moreover, the observed flow regimes and
vortex growth are compared in the same dimensionless parameter
spaces for the purpose of elucidating the routine(s). Our findings in this Fig. 1. Perspective view of the planar constriction microchannel with the inset
work can be useful to various analogous systems, for instance, polymer showing the contraction-expansion region with dimensions.

2
S. Wu et al. Journal of Non-Newtonian Fluid Mechanics 290 (2021) 104508

overlap concentration for dilute PEO solutions. The relaxation time of

3000 ppm PVP solution was estimated from the reported value of 2.2 ms
for 5% PVP solution, which was measured by Liu et al. using a small-
amplitude oscillatory test [78], via the concentration scaling, λ∝c0.8 ,
for a good solvent with solvent quality index of 0.6 [77]. The relaxation
time of 3000 ppm, Mw = 0.357 MDa HA solution was estimated from the
reported value of 0.11 ms for 700 ppm, Mw = 0.9 MDa HA solution,
which was measured by Haward using an extensional flow oscillatory
rheometer [79], via the following molecular-weight and concentration
scaling,

λ∝Mw 1.8 c0.8 (3)

The rheological properties for the prepared polymer solutions are
summarized in Table 1.

2.2. Methods

To visualize the flow pattern in the constriction microchannel,

fluorescent particles of 1 µm diameter (0.05% solids, Bangs Labora­
tories) were seeded into each of the prepared polymer solutions as well
as water. A syringe pump (KD Scientific, Holliston, MA, USA) was used
to drive the particle suspension through the channel in the range of flow
rates from 1 to 70 ml/h. Prior to any test, the channel was primed with
Tween 20 (0.5% vol. Thermo Fisher Scientific, Waltham, MA) for 5 min,
which was found to help minimizing the particles sticking to the channel
walls. The behavior of the tracing particles at the contraction-expansion
region was recorded using an inverted fluorescent microscope (Nikon
Eclipse TE2000U, Nikon Instrument) with a CCD camera (Nikon DS-
Qi1Mc, Nikon Instrument, Lewisville, TX, USA). The exposure time
and intensity of the incident blue light were tuned to obtain good
streakline images at different flow rates. The obtained digital images
were processed using the Nikon imaging software (NIS-Elements, Nikon
Instrument, Lewisville, TX, USA).
The fluid inertial effect on the contraction and expansion flows at the
constriction region of the microchannel is characterized using the local
Reynolds number, Re,
ρVDh 2ρQ
Re = = (4)
η(γ̇) η(γ̇)(w + h)

where ρ is the fluid density assumed equal to the density of the solvent (i.
e., DI water) for dilute polymer solutions, V = Q/hw is the speed cor­
responding to the flow through the constriction with Q being the volu­
Fig. 2. Experimentally measured viscosity data (symbols) for the prepared metric flow rate, h the channel height and w the constriction width,
polymer solutions: (a) 3000 ppm PEO, PVP, and HA solutions with similar Dh = 2hw/(w +h) is the hydraulic diameter of the constriction, and η(γ̇)
molecular weights of around 0.3 MDa; (b) 1000 ppm PEO with varying mo­ is the fluid viscosity estimated at the characteristic shear rate across the
lecular weights; (c) 1 MDa PEO at varying concentrations. The dashed lines on constriction width, i.e., γ̇ = 2V/w. The fluid elasticity effect is charac­
each plot represent the Carreau-model fitting of the viscosity data for the fluids terized by the Weissenberg number, Wi,
exhibiting shear-thinning behaviors.
2λQ
Wi = λγ̇ = (5)
viscosity, λC is a time constant, and γ̇ is the fluid shear rate. Note that the w2 h
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
∑ The elasticity number, El, which is defined as the ratio of the Weis­
standard deviation of the fitted data, σ = (1 − measured/fitted)2 /N,
senberg number to the Reynolds number, measures the relative impor­
was minimized (smaller than 2.5% for all cases) in the curve fitting, tance of the elastic stress over the inertial stress,
where N is the number of data points under consideration. The relaxa­
tion times of the prepared polymer solutions were found in the literature Wi λη(γ̇)(w + h)
El = = (6)
with appropriate scaling for the molecular weight and/or concentration Re ρw2 h
effect if needed. Specifically, the relaxation times, λ, of the PEO solutions
The estimated value of El for each prepared fluid is presented in
were estimated from the reported value of 1.5 ms for c = 1000 ppm, Mw
Table 1 for the flow rate of 10 ml/h, at which the calculated shear rate is
= 2 MDa PEO solution, which was measured by Rodd et al. using
63131.3 s− 1. For the ease of description, we regard a fluid as weakly
capillary breakup extensional rheometry [64], via the following
viscoelastic for 0 < El < 1, mildly elastic for 1 ≤ El < 10, and strongly
molecular-weight and concentration scaling [77],
elastic for El ≥ 10. We admit that the estimated values of Wi and El are
λ∝[η]Mw (c/c∗ )0.65 ∝Mw 2.073 c0.65 (2) probably not very accurate for the tested polymer solutions because of
the lacking of the experimental data for the fluid relaxation time. This
where, [η] = 0.072Mw 0.65 is the intrinsic viscosity and c∗ = 1 /[η] is the approximation is, however, not expected to change the overall trend
that we will report in the next section for at least the various PEO

3
S. Wu et al. Journal of Non-Newtonian Fluid Mechanics 290 (2021) 104508

Table 1
Rheological properties of the prepared polymer solutions. The elasticity number, El, was estimated at the flow rate of 10 ml/h for all the fluids.
Polymer Mw (MDa) c(ppm) λ(ms) η0 (mPa⋅s) η∞ (mPa⋅s) λC (s) n El

PVP 0.36 3000 0.23 1.18 1.18 - ≈1 0.29

HA 0.357 3000 0.067 16 1.5 0.0018 0.62 0.33
PEO 0.3 3000 0.06 2.09 2.09 - ≈1 0.14
0.3 1000 0.03 1.14 1.14 - ≈1 0.04
0.6 1000 0.12 1.5 1.5 - ≈1 0.19
1 1000 0.36 1.62 1.62 - ≈1 0.63
2 1000 1.50 2.4 1.5 0.01 0.85 3.04
4 1000 6.31 3.26 1.9 0.013 0.73 14.73
1 500 0.30 1.15 1.15 - ≈1 0.37
1 1500 0.46 2.15 1.4 0.004 0.90 0.92
1 2000 0.56 2.7 1.6 0.0045 0.84 1.26
1 3000 0.73 4.2 2.2 0.0052 0.80 2.29

solutions. The fluid shear-thinning effect is characterized by the power- fluid vortices formed in the expansion and/or contraction flow is char­
law index, n, in Table 1, from which we notice that our dilute polymer acterized by the dimensionless vortex length, χ L = Lv /W, where the
solutions exhibit negligible to weak shear thinning effects with 0.75 ≤ n vortex length, Lv , is measured directly from the experimental images and
< 1 except the salt-free HA and 4 MDa PEO solutions that are both W is the width of the microchannel.
mildly shear thinning with 0.5 ≤ n < 0.75. In addition, the growth of the

Fig. 3. Downward flows of Newtonian water

(a) and viscoelastic solutions of 3000 ppm PEO,
PVP and HA polymers with similar molecular
weights of Mw ≈ 0.3 MDa (b) through the
constriction microchannel. The numbers
labeled on the left and right sides of the channel
constriction on each image indicate the corre­
sponding Reynolds and Weissenberg numbers,
respectively. The right half of the middle image
in (a) shows the simulated fluid streamlines of
water flow under the experimental conditions.
The dimension, Lv , highlighted on the image in
(a) measures the fluid vortex length. The scale
bar on the right-most image in the last row
represents 100 µm.

4
S. Wu et al. Journal of Non-Newtonian Fluid Mechanics 290 (2021) 104508

3. Results and discussions 3.2. Effect of polymer molecular weight

3.1. Effect of polymer type For the effect of polymer length, we tested the solutions of 1000 ppm
PEO with the molecular weight ranging from 0.3 MDa to 0.6 MDa, 1
We tested the solutions of 3000 ppm PEO, PVP and HA polymers MDa, 2 MDa and 4 MDa in the constriction microchannel. The first three
with similar molecular weights, Mw ≈ 0.3 MDa, for the effect of polymer solutions are all weakly elastic (El < 1) with negligible shear thinning.
type on the flow in the constriction microchannel. These solutions The 2 MDa PEO solution is a mildly elastic (1 < El = 3.04 < 10 at 10 ml/
exhibit a roughly equally weak elasticity with El ∼ 0.2. However, as h) and a weakly shear thinning (n = 0.85 > 0.75) fluid. The 4 MDa PEO
these polymers have different physical properties (e.g., contour length, solution becomes strongly elastic (El = 14.73 > 10 at 10 ml/h) and
coil size, etc.), the rheological properties of their aqueous solutions (e.g., mildly shear thinning (n = 0.73 < 0.75) (see Table 1). Fig. 4 shows the
viscosity, relaxation time, etc.) should also differ. In practice, the PEO flow patterns at the contraction-expansion regions in these solutions. For
and PVP solutions are both like Boger-like fluids [80] with negligible the expansion flow, the vortex developments in the two lowest
shear thinning while the HA solution is mildly shear thinning (n = 0.62 molecular-weight solutions, i.e., 0.3 MDa and 0.6 MDa, are both anal­
< 0.75). We also tested the flow of DI water (i.e., 0 ppm polymer) in the ogous to the purely inertial flow of water in Fig. 3(a), where the
same microchannel as the control experiment. Fig. 3 shows the tracing expansion flow vortices start from lip vortices and grow into large corner
particle streaklines in these fluids for flow rates of up to 70 ml/h. Like vortices extending downstream with increasing flow rates. Moreover,
the purely inertial flow of water in Fig. 3(a), the expansion flow sepa­ the size of these vortices are also comparable between the two solutions
ration in each of the polymer solutions in Fig. 3(b) initiates with the at similar values of Re. Increasing the molecular weight to 1 MDa makes
symmetric lip vortices under a value of around Re = 20, which grow in the fluid vortices in the expansion flow start from the salient corners of
both the sideway and along the length of the channel with the increase the expansion walls at around Re = 72.2 (20 ml/h). This phenomenon
of Re. However, a little delayed formation of these vortices in the PEO seems to be associated with the comparable fluid elasticity and inertia
flow is perhaps associated with the larger-magnitude of stretching or effects in the 1 MDa PEO solution with El = 0.63 ∼ 1. However, as no
deformation of the flexible coils of the PEO polymer. Thus, despite the such salient-corner vortices have been observed in the strongly more
similar contour lengths, polymer concentrations, and time scales of the elastic 5% PVP (0.36 MDa) solution (El = 17.2) in our recent work [72],
relaxation processes, the PEO molecules can expand and consequently we speculate it is probably mainly the consequence of the greater
suppress the inertial effects more at the expansion part compared to the extensional stretching and re-orientation of the longer PEO polymer.
relatively rigid PVP or HA molecules. Once the lip vortices reach the These vortices grow with the increase of Re both inward from the salient
salient corners of the expansion walls, they continue to grow only in the corners to the channel center and downward following the flow. Further
primary flow direction. Such stable corner vortices in the expansion flow increasing the molecular weight to 2 MDa and 4 MDa causes the bending
of water are found in Fig. 3(a) to match the numerically predicted fluid of fluid streamlines in the expansion flow at an earlier stage, but sup­
streamlines from a three-dimensional model in COMSOL® with the presses the expansion-flow vortices to a point of complete absence.
regular Navier-Stokes and continuity equations. The contraction flow pattern of 0.3 MDa PEO solution maintains a
However, the three types of polymer solutions exhibit different be­ good similarity with that of the purely inertial fluid water. No bending or
haviors in their contraction flows as viewed from the images in Fig. 3(b). vortices are found before the channel constriction in Fig. 4 as far as the
The shear-thinning HA solution shows the development of small vortices flow rates are concerned. With the increase of molecular weight to 0.6
in contrast to the no vortices at all in the non-shear-thinning PVP and MDa and 1 MDa, a similar bending of fluid streamlines like that in Fig. 3
PEO solutions. Small lip vortices first form in the HA solution at around (b) for 3000 ppm, 0.3 MDa PEO solution is observed when Re reaches
Re = 5.4 (5 ml/h) on the re-entrant corners of the constriction and then 155.9 (40 ml/h) and 54.1 (15 ml/h, not shown in Fig. 4), respectively.
grow on to reach the salient corners of the contraction walls at around Re Note that both these values of Re are smaller than that in the 3000 ppm,
= 21.4 (15 ml/h). The mild shear thinning effect present in the HA so­ 0.3 MDa PEO solution, Re = 195.9 (70 ml/h), because of the longer
lution results in the formation of these steady symmetric vortices in the polymer chains in the 0.6 MDa and 1 MDa PEO solutions. When the
contraction flow, which are rather absent in constant-viscosity elastic molecular weight is increased to 2 MDa and 4 MDa, there is a single
flows through such structures, in compliance with our earlier observa­ vortex forming at one side in the stagnant salient corner of the
tions in the contraction-flow of PAA and XG solutions through a similar contraction part shortly after the appearance of streamline bending. This
planar constriction geometry [72]. Their size, however, does not grow phenomenon is speculated to be associated with the increasing shear
much along the channel length like what we have observed in the thinning effect in these two highest molecular-weight PEO solutions.
strongly shear-thinning PAA and XG solutions [72]. This may be because The size of the single corner vortex increases with the increasing value of
the fluid shear thinning effect of the HA solution is not strong enough to Re in both solutions. Its position, however, remains stably at the same
dominate its inertial effect at higher values of Re. The short chain of HA side of the contraction after the formation, in contrast to the bistable
polymers (as compared to those of the PAA and XG solutions) may also phenomenon reported by Rodd et al. [64]. At an even higher value of Re,
contribute to this behavior, which requires further studies. There are no which is 162.9 (50 ml/h) for 2 MDa and 69.5 (25 ml/h) for 4 MDa, the
such events other than the regular symmetric streamlines in the status changes from a single corner vortex to asymmetric corner vortices
contraction flow of the PVP solution for the range of Re tested. with another smaller vortex formed in the opposite corner. However,
Comparing it to the flow of Newtonian water, the viscoelastic PVP so­ this asymmetric state is unstable as the two vortices jump from one side
lution shows similar starting Re for the lip vortices and resembling to another of the contraction part at random intervals, which is different
development of them in size and appearance with the increase in Re. from the quasi-stable symmetric or asymmetric state reported by Rodd
This is consistent with our recent observation of 5% PVP solution in a et al. [64]. It even becomes chaotic in the 4 MDa PEO solution at flow
similar contraction-expansion microchannel [72]. In contrast, the rates higher than 40 ml/h. Such unstable upstream eddies were reported
contraction flow of the roughly equally elastic PEO solution shows weak by Browne et al. [81] to occur in an inertialess (Re≪1) flow of 18 MDa
bending streamlines at the salient corners starting under the highest PAA solution through a single pore-throat. The stronger influence of the
flow rate of 70 ml/h (Re = 195.9) in our test. This may be a consequence enhanced fluid elasticity and as well the reorientation and tumbling of
of the dissimilar structures between the PVP and PEO polymers, which longer polymer chains in the constriction region may have caused the
again requires further studies. observed flow pattern changes with the increasing molecular weight.

5
S. Wu et al. Journal of Non-Newtonian Fluid Mechanics 290 (2021) 104508

Fig. 4. Downward flows of viscoelastic solutions of 1000 ppm PEO with molecular weights ranging from 0.3 MDa to 4 MDa through the constriction microchannel in
a range of flow rates. The numbers labeled on the left and right sides of the channel constriction on each image indicate the corresponding Reynolds and Weissenberg
numbers, respectively. The scale bar on the bottom-right image represents 100 µm.

3.3. Effect of polymer concentration the overall trend observed is the elastic suppression of the fluid inertia-
induced vortices with the increase in polymer concentration. In the so­
We tested the solutions of 1 MDa PEO at 500 ppm, 1000 ppm, 1500 lution with the lowest PEO concentration of 500 ppm, the vortex growth
ppm, 2000 ppm and 3000 ppm for the effect of polymer concentration pattern is similar to water in Fig. 3(a). However, the inertial distur­
on the flow in the constriction microchannel. These fluids are negligibly bances to the expansion flow initiate at around Re = 25, which is
to weakly shear thinning (0.75 ≤ n < 1) and exhibit a weak-to-mild delayed than in the flow of water. Moreover, the length of the expansion-
elasticity with the elasticity number, El, on the order of 1. Fig. 5 flow corner vortices exhibits an interesting non-monotonic trend with
shows the flow patterns in these PEO solutions at the contraction- the increase of Re. This should be attributed to the competition of the
expansion region, which exhibits a similar development to the cases of fluid inertia (which promotes the vortex size) and elasticity (which
increasing polymer molecular weights in Fig. 4. For the expansion flow, suppresses the vortex size) effects. With the increase of PEO

6
S. Wu et al. Journal of Non-Newtonian Fluid Mechanics 290 (2021) 104508

Fig. 5. Downward flows of 1 MDa PEO solutions having different concentrations in the constriction microchannel. The numbers labeled on the left and right sides of
the channel constriction on each image indicate the corresponding Reynolds and Weissenberg numbers, respectively. The scale bar on the bottom-right image
represents 100 µm.

concentration to 1000 ppm and 1500 ppm, the initiating location of the 54.1 (15 ml/h) and 31.5 (10 ml/h), respectively. These diverging
expansion-flow vortices is suppressed from being at the re-entrant lips streamlines grow in intensity with the rise in Re but remain approxi­
and shifted to the salient corners of the expansion walls. For the 2000 mately symmetric for the rest of the tested flow rates. However, an
ppm and 3000 ppm flows, there are no vortices but bending of increased concentration of 2000 ppm renders the diverging streamlines,
streamlines for as far as the highest flow rate is concerned because of the which start developing at Re < 28.0 (10 ml/h), to collapse into a single
strong fluid elasticity effect. vortex on one side near the salient corner of the contraction under
For the contraction flow, the onset of elastic disturbances occurs at higher flow rates. The transition from the diverging streamlines to the
Re = 152.6 (30 ml/h) in the lowest PEO (Mw = 1 MDa) concentration of single vortex pattern takes place at around Re = 57.3 (20 ml/h), and the
500 ppm. This starting value of Re, which is comparable to that in the flow stays in this regime till the end of our tested cases (Re = 178.2 at 60
1000 ppm, 0.6 MDa PEO solution (Re = 155.9 at 40 ml/h) in Fig. 4, ml/h) for the flow of 2000 ppm PEO solution. These observed phe­
decreases for the higher concentration solutions of 1000 ppm and 1500 nomena may be attributed to the greater normal stress differences in the
ppm as the diverging streamlines are observed in them at Re as early as flow of higher concentration solution with a larger elasticity. The

7
S. Wu et al. Journal of Non-Newtonian Fluid Mechanics 290 (2021) 104508

increase in fluid shear thinning may also have played a part in the for­ ppm PVP solution also resembles that of water despite its El = 0.29. This
mation of vortices in the contraction flow of 2000 ppm, 1 MDa PEO observation is backed up by our recent study of 5% PVP solution with
solution, like those in the 1000 ppm, 2 MDa and 4 MDa PEO solutions El > 10 in a similar constriction microchannel [72], which may be
(Fig. 5). Further increasing the PEO concentration to 3000 ppm shifts the associated with its particular polymer structure and requires further
flow transitions to an even smaller value of Re. The single corner vortex studies. For the polymer solutions with 0.1 < El < 1.0, symmetric
state is maintained in both 2000 ppm and 3000 ppm PEO solutions for diverging streamlines occur in the contraction flow if the Weissenberg
the whole range of tested flow rates (up to 60 ml/h). Its size increases number is large enough. Moreover, the greater El renders the diverging
continuously with the increasing value of Re, similar to that of the single streamlines to start at a smaller Reynolds number. This flow regime
vortex in 2 MDa and 4 MDa PEO solutions (Fig. 5). covers 1500 ppm, 1000 ppm and 500 ppm PEO-1M, 1000 ppm
PEO-0.6M, and 3000 ppm PEO-0.3M solutions. The only exception is
3.4. Summary of flow regimes 3000 ppm HA solution with El = 0.33, where first lip and then corner
vortices are induced in its contraction flow by the fluid shear thinning
A summary of the observed expansion flow regimes for the tested effect. For the polymer solutions with 1.0 < El < 2.5, which include
polymer solutions and water in the constriction microchannel is plotted 3000 ppm and 2000 ppm PEO-1M solutions, diverging streamlines can
in the Re − Wi space in Fig. 6. As the Newtonian water flow (assigned a collapse into a single corner vortex when Wi is above a certain value.
constant small Weissenberg number, Wi = 0.1, for the purpose of Further increasing the value of El to above 2.5 forces the contraction
graphing) is observed to exhibit lip vortices near the re-entrant corners flow to change from a single vortex to unstable asymmetric vortices in
of the expansion walls at around Re = 20 in both the experiment and 1000 ppm PEO-2M and PEO-4M solutions when Wi (or equivalently Re)
simulation, we highlight this Re on the plot as the start of the flow becomes sufficiently high.
regime that fluid inertia plays a significant role. Regarding fluid visco­
elasticity, we find it useful to divide the Re − Wi space into three regimes 3.5. Summary of vortex development
based on the value of the elasticity number, El. The polymer solutions
with El < 0.5 (estimated at 10 ml/h, see Table 1) show a similar flow Vortex development in the tested polymer solutions is quantified by
development to that of water because of the dominant inertial effect the measured length of stable vortex, Lv [highlighted on the image in
over the elasticity effect. Specifically, the disturbances to the expansion Fig. 3(a)], in the contraction and/or expansion part of the constriction
flow of 500 ppm PEO-1M, 1000 ppm PEO-0.6M, 1000 ppm and 3000 microchannel. Fig. 8 shows the normalized vortex length, χ L = Lv /W,
ppm PEO-0.3M, 3000 ppm PVP and HA solutions all start with small lip against Re in the expansion flow. As viewed from the highlighted values
vortices and then develop into stable corner vortices with the increase of of El on the plot, the curves for the polymer solutions with El < 0.35
Re. For the more viscoelastic polymer solutions with El < 1.0, the (including 1000 ppm and 3000 ppm PEO-0.3M, 1000 ppm PEO-0.6M,
expansion flow vortices instead initiate at the salient corners. This 3000 ppm PVP and HA solutions) almost coincide with that of the
transition should be associated with the comparable fluid elasticity and purely inertial vortices in water, illustrating an increasingly larger
inertial effects in 1000 ppm and 1500 ppm PEO-1M solutions. For those vortex with the increase of Re. The slight discrepancy among the curves
polymer solutions with El > 1.0, the fluid inertia-induced expansion is perhaps caused by the experimental error in the measurement of fluid
flow vortices are completely blocked by the strong elasticity effect. Only viscosity or vortex length. For 500 ppm PEO-1M solution with El = 0.37,
diverging streamlines are observed in 1000 ppm PEO-4M, 1000 ppm lip-initiated corner vortices are still available in the expansion flow.
PEO-2M, 3000 ppm and 2000 ppm PEO-1M solutions. Their length, however, reaches a local maxima at around Re = 127.1 (at
Fig. 7 presents a summary of the observed contraction flow regimes 25 ml/h) and starts increasing again after this point without any
for the tested polymer solutions in the Re − Wi plot, where the whole anomaly. This phenomenon may be associated with the elastic distur­
space is divided into four regimes based on the value of El. Similar to the bances to the contraction flow that initiate at Re = 152.6 (at 30 ml/h) in
Newtonian water flow, no disturbances to the contraction flow are the form of bending streamlines. For 1000 ppm and 1500 ppm PEO-1M
observed in 1000 ppm PEO-0.3M solution because of the insignificant solutions with El > 0.5, only corner-initiated vortices are observed (see
fluid elasticity effect at El < 0.1. However, the contraction flow of 3000 Fig. 6). Their length becomes significantly smaller than that of the

Fig. 6. Summary of the flow regimes in the Re − Wi

space for the expansion flow of viscoelastic polymer
solutions (a: 1000 ppm PEO-4M; b: 1000 ppm PEO-2M;
c: 3000 ppm PEO-1M; d: 2000 ppm-1M; e: 1500 ppm
PEO-1M; f: 1000 ppm PEO-1M; g: 500 ppm PEO-1M; h:
3000 ppm PVP; i: 1000 ppm PEO-0.6M; j: 3000 ppm
PEO-0.3M; k: 1000 ppm PEO-0.3M; l: 3000 ppm HA)
and Newtonian water (the Weissenberg number is
assumed to be 0.1 for the purpose of graphing) in the
constriction microchannel: dashes for undisturbed
flows; crosses for lip bending streamlines; circles for lip
vortices; triangles for lip-initiated corner vortices; di­
amonds for corner-initiated vortices; asterisks are for
bending and diverging streamlines. The vertical
dashed-line at Re = 20 marks the transition to the flow
regime where the fluid inertia-induced flow distur­
bances start occurring at the re-entrant corners. The
other two straight dashed-lines highlight the transitions
from the lip-initiated corner vortices to the corner-
initiated vortices at El ≈ 0.5 and to the vortex-free
bending streamlines at El ≈ 1.0.

8
S. Wu et al. Journal of Non-Newtonian Fluid Mechanics 290 (2021) 104508

Fig. 7. Summary of the flow regimes in the Re − Wi

space for the contraction flow of viscoelastic polymer
solutions (a: 1000 ppm PEO-4M; b: 1000 ppm PEO-2M;
c: 3000 ppm PEO-1M; d: 2000 ppm-1M; e: 1500 ppm
PEO-1M; f: 1000 ppm PEO-1M; g: 500 ppm PEO-1M; h:
1000 ppm PEO-0.6M; i: 3000 ppm PEO-0.3M) in the
constriction microchannel: dashes for undisturbed
flows; crosses for lip bending streamlines; circles for lip
vortices; triangles for lip-initiated corner vortices; as­
terisks are for bending and diverging streamlines; di­
amonds for single corner vortices; squares for
asymmetric corner vortices. The three straight dashed-
lines highlight the transitions from the undisturbed
flow to the diverging streamlines at El ≈ 0.1, to the
single corner vortices El ≈ 1.0, and to the asymmetric
corner vortices at El ≈ 2.5.

Fig. 8. Normalized vortex length, χ L = Lv /W, against the Reynolds number, Re, for the expansion flow of viscoelastic polymer solutions and Newtonian water in the
constriction microchannel. The estimated value of the elasticity number, El, at the flow rate of 10 ml/h is highlighted for each fluid.

inertial vortices and decreases with the increasing polymer concentra­ vortices in the four PEO solutions (all in the regime of single vortices) are
tion. It, however, increases with the rise of Re in both solutions because all significantly larger than those in the HA solution. Their sizes, how­
of the enhanced inertial effect though to a much smaller extent than that ever, all grow with the increase of Re, in distinct contrast to the non-
of the inertial vortices. No expansion flow vortices are observed in the monotonic variation in the HA solution. These observations may be
polymer solutions with El > 1.0. attributed to both the stronger fluid elasticity and the longer polymer
Fig. 9 shows the normalized vortex length, χ L , against Re in the chains of the PEO solutions. It is also viewed from Fig. 9 that the onset of
contraction flow of the constriction microchannel. Only five of the tested the contraction flow vortices in the PEO solutions moves ahead and their
polymer solutions exhibit vortices, among which one is the 3000 ppm length grows with the increase of the polymer length or concentration.
HA solution with El = 0.33 and all the others are PEO solutions with El However, neither of these changes is a simple function of the elasticity
> 1.0. The fluid shear thinning-induced symmetric corner vortices in the number. This may be because only single vortices appear in the 2000
contraction flow of the HA solution are much smaller than those of the ppm and 3000 ppm PEO-1M solutions while asymmetric vortices are
fluid inertia-induced expansion flow vortices (see Fig. 8). Moreover, formed in the 1000 ppm PEO-2M and PEO-4M solutions.
they exhibit a first increasing and then decreasing trend with the tran­
sition occurring at around Re = 50 because of the increasingly dominant
inertial effect over the shear thinning effect. The contraction flow

9
S. Wu et al. Journal of Non-Newtonian Fluid Mechanics 290 (2021) 104508

properties amongst them in the larger length scales. Increasing the

molecular weight of PEO polymer causes stronger disturbances to the
streamlines and ultimately the formation of single or even asymmetric
vortices in the contraction flow because of the enhanced fluid elasticity
(0.04 ≤ El ≤ 14.73) and shear thinning (1 ≥ n ≥ 0.73) effects. The
expansion flow encounters a greater resistance on flow separation and
vortex formation as the polymer molecular weight increases. Increasing
the polymer concentration of PEO solution also enhances the fluid
elasticity and shear thinning effects, yielding similar influences on the
flow regime to increasing the polymer molecular weight. Overall, we
find that the fluid elasticity suppresses the fluid inertia-induced vortices
in the expansion flow, and meanwhile draws instabilities into both the
expansion and contraction flows in the form of bending streamlines.
Interestingly in 500 ppm PEO solution, we have seen the inertial
expansion-flow vortices decreasing in size when the streamline bending
starts in the contraction flow. This phenomenon may indicate the
Fig. 9. Normalized vortex length, χ L = Lv /W, against the Reynolds number, Re, dependence of the expansion and contraction flows as the effect of one
for the contraction flow of viscoelastic polymer solutions in the constriction may be propagated to the other through the polymer extensional,
microchannel. The estimated value of the elasticity number, El, at the flow rate reorientational or relaxation process. We have also presented the
of 10 ml/h is highlighted for each fluid. observed flow patterns in each of the tested polymer solutions in the
same dimensionless Re − Wi and Re − χ L parameter spaces. The elas­
3.6. Comparison with previous studies in similar geometries ticity number is found to play a determining role in both the flow re­
gimes and the vortex development for at least the PEO solutions. It is
There have been a good body of works on the flow of viscoelastic hoped that our experimental data can stimulate more numerical studies
fluids, such as PAA [41,54,65,70,72,81], PAA-based Boger [68], PEO (see, e.g., [11]) for a more accurate understanding of fluid rheological
[38,42,52,64,66,72], PVP [72], Na-HA [71], XG [41,72], methylcellu­ effects in polymer solution flows through porous media [28,59].
lose [82], polystyrene [83,84] and worm-like micelle [85] solutions, in a
planar contraction-expansion microchannel. However, the majority of
Declaration of Competing Interest
these studies paid attention to the contraction flow and provided little
information on the expansion flow. Our study advances the under­
None.
standing of the elasto-inertial effects on such flows in at least the
following aspects: (a) We observe for the first time three different re­
gimes in the expansion flow, which are the Newtonian water-like inertial Acknowledgements
vortices initiating from the lips in the polymer solutions with El < 0.5,
the elasto-inertial vortices initiating at the salient corners in the solu­ This work was supported in part by China Scholarship Council (CSC)
tions with 0.5 < El < 1.0, and the vortex-free expansion flow in the so­ - Chinese Government Graduate Student Overseas Study Program (S.W.
lutions with El > 1.0; (b) Our value of El is varied from 0.04 (for 1000 and L.S.), University 111 Project of China under grant number B12019
ppm, 0.3 MDa PEO) to 14.73 (for 1000 ppm, 4 MDa PEO), which (L.Y.), Clemson University through a SEED grant (X.X.), and NSF under
complements the previous studies [38,42,52,64,66,72] that exclusively grant number CBET-1750208 (J.B.B.).
used PEO solutions with El > 1. We observe in the contraction flow of
PEO solutions with 0.1 < El < 1 the onset of symmetric diverging References
streamlines if Re is large enough. No single or asymmetric vortices occur
[1] K.S. Sorbie, Polymer-Improved Oil Recovery, 1st edn., Springer, Netherlands, XII,
in these weakly elastic solutions even when the value of Wi reaches 1991, p. 359.
136.4 for 1000 ppm 1 MDa PEO solution at 60 ml/h (Re = 216.6). In [2] J.F. Steffe, Rheological Methods in Food Process Engineering, Freeman Press, East
contrast, Rodd et al. [64] observed steady contraction flow vortices for Lansing, MI, USA, 1996.
[3] D.S. Roote, Technology Status Report In Situ Flushing, Ground-Water Remediation
100 < Wi < 150 in PEO solutions with El > 1; (c) We propose the use of
Technologies Analysis Center, 1998, pp. 1–164.
El to divide the Re − Wi space for both the contraction (Fig. 6) and [4] D.V. Boger, Viscoelastic flows through contractions, Annual Rev. Fluid Mech. 19
expansion (Fig. 7) flows, in contrast to the critical Wi (or the Deborah (1987) 157–182.
[5] A. Lindner, Flow of complex suspensions, Phys. Fluid. 26 (2014), 101307.
number, De) that Rodd et al. [38,64] used for the contraction flow. As it
[6] F.J. Galindo-Rosales, L. Campo-Deaño, P.C. Sousa, V.M. Ribeiro, M.S.N. Oliveira,
is (nearly) independent of fluid kinematics, El may be viewed as a fluid M.A. Alves, F.T. Pinho, Viscoelastic instabilities in micro-scale flows, Exp. Therm.
property that is more convenient for uses. However, the value of El alone Fluid Sci. 59 (2014) 128–139.
does not indicate the flow regime, which must be determined in [7] X. Shi, G.F. Christopher, Growth of viscoelastic instabilities around linear cylinder
arrays, Phys. Fluids 28 (2016), 124102.
conjunction with the value of Wi (or alternatively Re). Therefore, it will [8] A. Souliès, J. Aubril, C. Castelain, T. Burghelea, Characterisation of elastic
be very useful to identify a single dimensionless number, like the turbulence in a serpentine micro-channel, Phys. Fluids 29 (2017), 083102.
√̅̅̅̅̅̅̅̅̅̅̅̅̅ [9] G. Yao, J. Zhao, H. Yang, M.A. Haruna, D. Wen, Effects of salinity on the onset of
parameter, M ≡ De⋅Wi, developed by Pakdel and McKinley [86], for
elastic turbulence in swirling flow, curvilinear microchannels, Phys. Fluids 31
this purpose if available. (2019), 123106.
[10] P.G. Correa, J.R. Mac Intyre, J.M. Gomba, M.A. Cachile, J.P. Hulin, H. Auradou,
Three-dimensional flow structures in X-shaped junctions: Effect of the Reynolds
4. Conclusion number, crossing angle, Phys. Fluids 31 (2019), 043606.
[11] S. Varchanis, C.C. Hopkins, A.Q. Shen, J. Tsamopoulos, S.J. Haward, Asymmetric
We have experimentally investigated the effects of polymer type, flows of complex fluids past confined cylinders: a comprehensive numerical study
with experimental validation, Phys. Fluids 32 (2020), 053103.
molecular weight, and concentration on viscoelastic fluid flows through
[12] R.B. Bird, R.C. Armstrong, O. Hassager, Dynamics of Polymeric Liquids, Wiley-
a planar constriction microchannel. Changing the polymer type amongst Interscience, 1987, p. 1.
PEO, PVP, and HA with similar molecular weights and concentrations [13] R.G. Larson, Instabilities in viscoelastic flows, Rheolog. Acta 31 (1992) 213–263.
does not affect the fluid inertia-induced vortices in the expansion flow, [14] S.J. Haward, G.H. McKinley, Instabilities in stagnation point flows of polymer
solutions, Phys. Fluids 25 (2013), 083104.
but causes completely different contraction flows because of their [15] R.G. Larson, Flow-induced mixing, demixing and phase-transitions in polymeric
different molecular structures resulting in a diversion in the rheological fluids, Rheolog. Acta 31 (1992) 497–520.

10
S. Wu et al. Journal of Non-Newtonian Fluid Mechanics 290 (2021) 104508

[16] J.A. Pathak, D. Ross, K.B. Migler, Elastic flow instability, curved streamlines, [50] M.A. Alves, P.J. Oliveira, F.T. Pinho, On the effect of contraction ratio in
mixing in microfluidic flows, Phys. Fluids 16 (2004) 4028–4034. viscoelastic flow through abrupt contractions, J. Non-newton. Fluid Mech. 122
[17] H.Y. Gan, Y.C. Lam, N.T. Nguyen, K.C. Tam, C. Yang, Efficient mixing of (2004) 117–130.
viscoelastic fluids in a microchannel at low Reynolds number, Microfluid. [51] M.S.N. Oliveira, P.J. Oliveira, F.T. Pinho, M.A. Alves, Effect of contraction ratio
Nanofluid. 3 (2007) 101–108. upon viscoelastic flow in contractions: the axisymmetric case, J. Non-Newton.
[18] S.O. Hong, J.J. Cooper-White, J.M. Kim, Inertio-elastic mixing in a straight Fluid Mech. 147 (2007) 92–108.
microchannel with side wells, Appl. Phys. Lett. 108 (2016) 13–17. [52] L.E. Rodd, D. Lee, K.H. Ahn, J.J. Cooper-White, The importance of downstream
[19] G. D’Avino, F. Greco, P.L. Maffettone, Particle migration due to viscoelasticity of events in microfluidic viscoelastic entry flows: consequences of increasing the
the suspending liquid and its relevance in microfluidic devices, Annu. Rev. Fluid constriction length, J. Non-Newton. Fluid Mech. 165 (2010) 1189–1203.
Mech. 49 (2017) 341–360. [53] P.C. Sousa, P.M. Coelho, M.S.N. Oliveira, M.A. Alves, Effect of the contraction ratio
[20] X. Lu, C. Liu, G. Hu, X. Xuan, Particle manipulations in non-Newtonian upon viscoelastic fluid flow in three-dimensional square–square contractions,
microfluidics: a review, J. Colloid Interf. Sci. 500 (2017) 182–201. Chem. Eng. Sci. 66 (2011) 998–1009.
[21] F. Tian, Q. Feng, Q. Chen, C. Liu, T. Li, J. Sun, Manipulation of bio‑micro/ [54] A. Lanzaro, X.F. Yuan, Effects of contraction ratio on non-linear dynamics of semi-
nanoparticles in non-newtonian microflows, Microfluid. Nanofluid. 23 (2019) 68. dilute, highly polydisperse PAAm solutions in microfluidics, J. Non-Newton. Fluid
[22] M.K. Raihan, D. Li, A.J. Kummetz, L. Song, L. Yu, Vortex trapping and separation of Mech. 166 (2011) 1064–1075.
particles in shear thinning fluids, Appl. Phys. Lett. 116 (2020), 183701. [55] J.E. López-Aguilar, M.F. Webster, H.R. Tamaddon-Jahromi, M. Pérez-Camacho,
[23] M.S.N. Oliveira, L.E. Rodd, G.H. McKinley, M.A. Alves, Simulations of extensional O. Manero, Contraction-ratio variation and prediction of large experimental
flow in microrheometric devices, Microfluid. Nanofluid. 5 (2008) 809–826. pressure-drops in sharp-corner circular contraction-expansions–Boger fluids,
[24] F.J. Galindo-Rosales, L. Campo-Deano, F.T. Pinho, E. van Bokhorst, P.J. Hamersma, J. Non-Newton. Fluid Mech. 237 (2016) 39–53.
M.S.N. Oliveira, M.A. Alves, Microfluidic systems for the analysis of viscoelastic [56] R.M. Matos, M.A. Alves, F.T. Pinho, Instabilities in micro‑contraction flows of
fluid flow phenomena in porous media, Microfluid. Nanofluid. 12 (2012) 485–498. semi‑dilute CTAB and CPyCl solutions: rheology and flow instabilities, Exp. Fluid.
[25] F.J. Galindo-Rosales, M.A. Alves, M.S.N. Oliveira, Microdevices for extensional 60 (2019) 145.
rheometry of low viscosity elastic liquids: a review, Microfluid. Nanofluid. 14 [57] C.D. Meinhart, H. Zhang, The flow structure inside a microfabricated inkjet
(2013) 1–19. printhead, J. Microelectromech. Syst. 9 (2000) 67–75.
[26] E.A. Gryparis, S.D. Gkormpatsis, K.D. Housiadas, R.I. Tanner, Viscoelastic planar [58] A.C. Barbati, J. Desroches, A. Robisson, G.H. McKinley, Complex fluids and
elongational flow past an infinitely long cylinder, Phys. Fluids 31 (2019), 033104. hydraulic fracturing, Annu. Rev. Chem. Biomol. Eng. 7 (2016) 415–453.
[27] A. Phan, D. Fan, A. Striolo, Fluid transport through heterogeneous pore matrices: [59] A. Anbari, H.T. Chien, S.S. Datta, W. Deng, D.A. Weitz, J. Fan, Microfluidic model
Multiscale simulation approaches, Phys. Fluids 32 (2020), 101301. porous media: fabrication and applications, Small 14 (2018), 1703575.
[28] C.A. Browne, A. Shih, S.S. Datta, Pore-scale flow characterization of polymer [60] M. Daoud, J.P. Cotton, B. Farnoux, G. Jannink, G. Sarma, H. Benoit, R. Duplessix,
solutions in microfluidic porous media, Small 16 (2020), 1903944. C. Picot, P.G. de Gennes, Solutions of flexible polymers-neutron experiments and
[29] K. Walters, D.M. Rawlinson, On some contraction flows for Boger fluids, Rheol. interpretation, Macromolecules 8 (1975) 804–818.
Acta 21 (1982) 547–552. [61] W.W. Graessley, Polymeric Liquids & Networks: Structure and Properties, CRC
[30] R.E. Evans, K. Walters, Flow characteristics associated with abrupt changes in Press, 2003.
geometry in the case of highly elastic liquids, J Non-Newton. Fluid Mech. 20 [62] M. Rubinstein, R.H. Colby, Polymer Physics, Oxford University Press Inc., Oxford,
(1986) 11–29. U.K., 2003.
[31] K. Chiba, T. Sakatani, K. Nakamura, Anomalous flow patterns in viscoelastic entry [63] A.V. Dobrynin, M. Rubinstein, Theory of polyelectrolytes in solutions and at
flow through a planar contraction, J. Non-Newton. Fluid Mech. 36 (1990) surfaces, Prog. Polym. Sci. 30 (2005) 1049–1118.
193–203. [64] L.E. Rodd, T.P. Scott, D.V. Boger, J.J. Cooper-White, G.H. McKinley, The inertio-
[32] J.P. Rothstein, G.H. McKinley, Extensional flow of a polystyrene Boger fluid elastic planar entry flow of low-viscosity elastic fluids in micro-fabricated
through a 4:1:4 axisymmetric contraction/expansion, J. Non-Newton. Fluid Mech. geometries, J. Non-Newton. Fluid Mech. 129 (2005) 1–22.
86 (1999) 61–88. [65] A. Lanzaro, Z. Li, X.F. Yuan, Quantitative characterization of high molecular
[33] J.P. Rothstein, G.H. McKinley, The axisymmetric contraction-expansion: the role of weight polymer solutions in microfluidic hyperbolic contraction flow, Microfluid.
extensional rheology on vortex growth dynamics and the enhanced pressure drop, Nanofluid. 18 (2015) 819–828.
J. Non-Newton. Fluid Mech. 98 (2001) 33–63. [66] Z. Li, X.F. Yuan, S.J. Haward, J.A. Odell, S. Yeates, Non-linear dynamics of semi-
[34] S. Nigen, K. Walters, Viscoelastic contraction flows: Comparison of axisymmetric dilute polydisperse polymer solutions in microfluidics: A study of a benchmark
and planar configurations, J. Non-Newton. Fluid Mech. 102 (2002) 343–359. flow problem, J. Non-Newton. Fluid Mech. 166 (2011) 951–963.
[35] M.A. Alves, P.J. Oliveira, F.T. Pinho, Benchmark solutions for the flow of Oldroyd- [67] E. Miller, J.J. Cooper-White, The effects of chain conformation in the microfluidic
B and PTT fluids in planar contractions, J. Non-Newton. Fluid Mech. 110 (2003) entry flow of polymer–surfactant systems, J. Non-Newton. Fluid Mech. 60 (2009)
45–75. 22–30.
[36] R.J. Poole, M.P. Escudier, Turbulent flow of viscoelastic liquids through an [68] L. Campo-Deañoa, F.J. Galindo-Rosales, F.T. Pinho, M.A. Alves, M.S.N. Oliveira,
axisymmetric sudden expansion, J. Non-Newton. Fluid Mech. 117 (2004) 25–46. Flow of low viscosity Boger fluids through a microfluidic hyperbolic contraction,
[37] R.J. Poole, M.P. Escudier, A. Afonso, F.T. Pinho, Laminar flow of a viscoelastic J. Non-Newton. Fluid Mech. 166 (2011) 1286–1296.
shear-thinning liquid over a backward-facing step preceded by a gradual [69] D. Kawale, E. Marques, P.L.J. Zitha, M.T. Kreutzer, W.R. Rossen, P.E. Boukany,
contraction, Phys. Fluid. 19 (2007), 093101. Elastic instabilities during the flow of hydrolyzed polyacrylamide solution in
[38] L.E. Rodd, J.J. Cooper-White, D.V. Boger, G.H. McKinley, Role of the elasticity porous media: effect of pore-shape and salt, Soft Matt. 13 (2017) 765–775.
number in the entry flow of dilute polymer solutions in micro-fabricated [70] E.M. Ekanem, S. Berg, S. De, A. Fadili, T. Bultreys, M. Rücker, J. Southwick,
contraction geometries, J. Non-Newton. Fluid Mech. 143 (2007) 170–191. J. Crawshaw, P.F. Luckham, Signature of elastic turbulence of viscoelastic fluid
[39] S. Gulatia, S.J. Muller, D. Liepmanna, Direct measurements of viscoelastic flows of flow in a single pore throat, Phys. Rev. E 101 (2020), 042605.
DNA in a 2:1 abrupt planar micro-contraction, J. Non-Newton. Fluid Mech. 155 [71] R. Hidema, T. Oka, Y. Komoda, H. Suzuki, Effects of flexibility and entanglement of
(2008) 51–66. sodium hyaluronate in solutions on the entry flow in micro abrupt contraction-
[40] S.C. Omowunmi, X.F. Yuan, Modelling the three-dimensional flow of a semi-dilute expansion channels, Phys. Fluids 31 (2019), 072005.
polymer solution in microfluidics-on the effect of aspect ratio, Rheolog. Acta 49 [72] P.P. Jagdale, D. Li, X. Shao, J.B. Bostwick, X. Xuan, Fluid rheological effects on the
(2010) 585–595. flow of polymer solutions in a contraction-expansion microchannel,
[41] P.C. Sousa, F.T. Pinho, M.S.N. Oliveira, M.A. Alves, Extensional flow of blood Micromachines 11 (2020) 278.
analog solutions in microfluidic devices, Biomicrofluid 5 (2011), 014108. [73] C.J. Pipe, G.H. McKinley, Microfluidic rheometry, Mech. Res. Commun. 36 (2009)
[42] T.J. Ober, S.J. Haward, C.J. Pipe, J. Soulages, G.H. McKinley, Microfluidic 110–120.
extensional rheometry using a hyperbolic contraction geometry, Rheolog. Acta 52 [74] X. Hu, P.E. Boukany, O.L. Hemminger, L.J. Lee, The use of microfluidics in
(2013) 529–546. rheology, Macromol. Mat. Eng. 296 (2011) 308–320.
[43] J.E. López-Aguilar, H.R. Tamaddon-Jahromi, M.F. Webster, K. Walters, Numerical [75] S.J. Haward, Microfluidic extensional rheometry using stagnation point flow,
vs experimental pressure drops for Boger fluids in sharp-corner contraction flow, Biomicrofluid 10 (2016), 043401.
Phys. Fluids 28 (2016), 103104. [76] X. Lu, S. Patel, M. Zhang, S.W. Joo, S. Qian, A. Ogale, X. Xuan, An unexpected
[44] S.J. Haward, J. Page, T.A. Zaki, A.Q. Shen, Phase diagram for viscoelastic particle oscillation for electrophoresis in viscoelastic fluids through a microchannel
Poiseuille flow over a wavy surface, Phys. Fluids 30 (2018), 113101. constriction, Biomicrofluid 8 (2014), 021802.
[45] T. Tomkovic, E. Mitsoulis, S.G. Hatzikiriakos, Contraction flow of ionomers and [77] V. Tirtaatmadja, G.H. McKinley, J.J. Cooper-White, Drop formation and breakup of
their corresponding copolymers: ionic and hydrogen bonding effects, Phys. Fluids low viscosity elastic fluids: effects of molecular weight and concentration, Phys.
31 (2019), 033102. Fluids 18 (2006), 043101.
[46] C. Sasmal, Flow of wormlike micellar solutions through a long micropore with step [78] C. Liu, C. Xue, X. Chen, Shan L, Y. Tian, G. Hu, Size-based separation of particles
expansion and contraction, Phys. Fluids 32 (2020), 013103. and cells utilizing viscoelastic effects in straight microchannels, Anal. Chem. 87
[47] L.L. Ferrás, A.M. Afonso, M.A. Alves, J.M. Nóbrega, F.T. Pinho, Newtonian and (2015) 6041–6048.
viscoelastic fluid flows through an abrupt 1:4 expansion with slip boundary [79] S.J. Haward, Characterization of hyaluronic acid and synovial fluid in stagnation
conditions, Phys. Fluids 32 (2020), 043103. point elongational flow, Biopolymers 101 (2014) 287–305.
[48] P.R. Varges, B.S. Fonseca, P.R. de Souza Mendes, M.F. Naccache, C.R. de Miranda, [80] D.F. James, Boger fluids, Annu. Rev. Fluid Mech. 41 (2009) 129–142.
Flow of yield stress materials through annular abrupt expansion–contractions, [81] C.A. Browne, A. Shih, S.S. Datta, Bistability in the unstable flow of polymer
Phys. Fluids 32 (2020), 083101. solutions through pore constriction arrays, J. Fluid Mech. 890 (2020) A2.
[49] S.A. White, A.D. Gotsis, D.G. Baird, Review of the entry flow problem:
experimental and numerical, J. Non-Newton. Fluid Mech. 24 (1987) 121–160.

11
S. Wu et al. Journal of Non-Newtonian Fluid Mechanics 290 (2021) 104508

[82] M.W. Collis, M.R. Mackley, The melt processing of monodisperse and polydisperse [84] P.F. Salipante, S.E. Meek, S.D. Hudson, Flow fluctuations in wormlike micelle
polystyrene melts within a slit entry and exit flow, J. Non-Newton. Fluid Mech. 128 fluids, Soft Matter 14 (2018) 9020–9035.
(2005) 29–41. [85] B.L. Micklavzina, A.E. Metaxas, C.S. Dutcher, Microfluidic rheology of
[83] S.J. Haward, Z. Li, D. Lighter, B. Thomas, J.A. Odell, X. Yuan, Flow of dilute to methylcellulose solutions in hyperbolic contractions and the effect of salt in shear
semi-dilute polystyrene solutions through a benchmark 8:1 planar abrupt micro- and extensional flows, Soft Matter 16 (2020) 5273–5281.
contraction, J. Non-Newton. Fluid Mech. 165 (2010) 1654–1669. [86] P. Pakdel, G.H. McKinley, Elastic instability and curved streamlines, Phys. Rev.
Lett. 77 (1996) 2459–2462.

12