J. Non-Newtonian Fluid Mech.

87 (1999) 283305

The rheology of aqueous polyurethane dispersions

G.L. Flickinger, I.S. Dairanieh b, , C.F. Zukoski a

a

Department of Chemical Engineering, University of Illinois, Urbana, IL 61801, USA

b
Amoco Chemicals, Naperville, IL 60566-7011, USA
Received 4 April 1999; received in revised form 7 June 1999

Abstract
The rheological properties of aqueous polyurethane dispersions are reported from the dilute regime up to mass
concentrations in excess of 40 wt.%. Particle size decreases and the particles are more resistant to shear induced
deformation with increasing ionic strength and, at very low ionic strength, the reduced viscosity passes through a
minimum with increasing concentration. As the mass fraction of polyurethane increases, the dispersions shear thin.
The zero shear rate viscosity diverges with the development of an apparent yield stress at the highest concentrations
probed. Stress sweep experiments show that deviation from linear viscoelasticity occurs at strains of less than 0.05%
for the moderately concentrated dispersions. The CoxMerz rule is obeyed by these dispersions only at strains in
the linear viscoelastic region and at concentrations below the gel point. As the ionic strength is increased at fixed
particle concentration, viscosities pass through a minimum. These phenomena are discussed in terms of particle
deformability and an interplay of electrostatic and steric repulsive forces. 1999 Elsevier Science B.V. All rights
reserved.
Keywords: Polyurethane

1. Introduction
Driven by environmental regulations, water-borne coatings and adhesives are of steadily growing
importance. Replacement of the organic solvent by water not only protects the environment but also
leads to reduced product costs. A major class of water-borne polymeric dispersions where flow properties
are poorly studied are those prepared by polycondensation followed by water dispersion [1]. A prime
example of these materials are aqueous rheological polyurethanes.
The chemical difference between the conventional water insoluble polyurethanes and those dispersible
in water lies in the incorporation of hydrophilic segments in the polyurethane backbone. These built-in

Dedicated to Professor David V. Boger on the occasion of his 60th birthday.

Corresponding author.

0377-0257/99/$ see front matter 1999 Elsevier Science B.V. All rights reserved.
PII: S 0 3 7 7 - 0 2 5 7 ( 9 9 ) 0 0 0 6 8 - 3

284

G.L. Flickinger et al. / J. Non-Newtonian Fluid Mech. 87 (1999) 283305

internal emulsifiers are nonionic, cationic or anionic; examples of such moieties can be found in a review
paper by Dieterich [2]. Depending on the chemical structure and the concentration of the monomers,
aqueous polyurethane (PU) dispersions can be tailor made for a variety of applications [2,3]. The influence
of variations in preparation procedures and/or reactants have been correlated with the performance of these
dispersions in coatings and adhesives [47]. The synthesis process results in the polymers being spherical
particles of relatively narrow size distribution suspended in an aqueous electrolyte. Upon drying, these
dispersions coalesce to form barrier coatings. Despite the industrial significance of PU dispersions, little is
understood about their rheological properties [5,6,8,9]. This paper presents the results of a detailed study
of the rheological properties of anionic aqueous PU dispersions and is written in honor of Dave Boger
on the event of his sixtieth birthday. From Dave we have learned much about non-Newtonian mechanics
and the excitement of developing fundamental understanding in pursuit of technological advances.
As a starting point, we anticipate that the particles will have flow properties similar to soft spheres.
This assumption is based on morphological studies indicating that aqueous PU dispersions are well
characterized as consisting of spherical particles [1]. These particles have a narrow but not monodisperse
size distribution and contain internal pockets of water. The base case rheology for these systems is that
of hard spheres where the zero shear rate viscosity (o ) data has been correlated over a wide range of
volume fractions () by


o
2
ro =
= 1
,
(1)
c
m
where c is the continuous phase viscosity and m is the volume fraction at close packing [10]. For hard
spheres in the low shear limit, m = 0.63. This correlation describes the concentration dependence of
particles exhibiting soft interactions such as electrostatic or steric repulsions, but the parameter m , must
be adjusted to account for the effective increase in particle size due to repulsion [11,12].
As increases, the viscosity of suspensions of hard and soft particles displays high shear rate and high
shear rate limiting plateau viscosities, o and , respectively. This shear thinning behavior is often well
described by the Cross equation [13] in terms of critical stress ( c ):
1

=
,
o
1 + (/c )

(2)

or in terms of critical shear rate (c )

1

=
,
o
1 + ( /c )

(3)

where and are order one quantities.

The rheological properties of a variety of suspensions containing deformable particles have been investigated where the thickness of the soft or deformable layer increases relative to the size of the particles
hard core. Hard particles with thin steric layers are well described by the hard sphere scaling with m
being adjustable to account for the adsorbed layer thickness [12,1418]. Hard core particles with thick
soft layers, on the other hand, show deviations from hard sphere scalings. In particular, they display a
more rapid increase in o as increases. The ability of Eq. (1) to describe the volume fraction dependence
of o also breaks down near m where these suspensions exhibit drastically different dynamic properties
due to particle interactions [12,19].

G.L. Flickinger et al. / J. Non-Newtonian Fluid Mech. 87 (1999) 283305

285

The rheology of suspensions containing soft or gel particles is exemplified by crosslinked polymer
particles swollen by good solvent. These suspensions show similar behavior to hard core particles with
thick stabilizing layers up to m [2023]. However above m , as the particles remain deformable, the
suspensions respond as if they are a continuous polymer network. As the particle crosslink density
increases, these suspensions approach the hard sphere limit [21].
In comparing PU dispersion rheology with observations on other systems one of the key issues is the
determination of the sample concentration in terms of volume fraction. This is typically done by assuming
that the reduced viscosity (r ) of the system in the dilute limit is described by the Einstein equation:

= 1 + 2.5eff = 1 + []c,
(4)
lim r =
eff 0
c
where c is the continuos phase viscosity and eff is the effective volume fraction. From Eq. (4), the
measured mass concentration (c) can be linked to volume fraction through
[]
c,
(5)
2.5
where [] is the intrinsic viscosity. Although this link is established in the dilute regime, the connection
between and c is extrapolated into the dense suspension region by assuming that the particles maintain
the same size and shape [18,2429]. The PU dispersions investigated here contain highly charged particles
suggesting that repulsive interactions will be important. In addition, these particles deform under shear. As
the system ionic strength increases, the particles shrink, become less deformable and ultimately become
attractive and aggregate. As a consequence, care must be taken in assuming behavior at low c can be
extrapolated to high c.
Below in Section 2 we describe our experimental techniques which include static and dynamic light
scattering to determine particle molecular weight and second virial coefficient and size. In addition flow
birefringence is used to establish that the particles deform under flow. In Section 3 we first discuss
properties of dilute PU dispersions prior to characterizing dense dispersion properties. In Section 4 we
draw conclusions.
eff =

2. Experimental
2.1. Materials
The polyurethane particles studied in this work are commercially available from Zeneca Resins (Wilmington, MA) under the trade name Neorez R-9637. This system is an aliphatic polyester resin that is
widely used both as a stand-alone dispersion and in blends with other dispersions. The particles are
synthesized by the prepolymer mixing process and contain in-built carboxyl stabilizing groups which are
provided by dimethylolpropionic acid (DMPA). These particles are highly charged. Based on the chemical formula for this system, there should be one carboxylic acid group per 1878 atomic weight units
(AMU). A conductometric titration of the actual experimental system suspended in DI water carried out
in our laboratory indicated that there is one charged group per every 2573 AMU. The number of charges
per particle can be estimated by converting the moles of carboxylic acid groups per gram of PU with the
molecular weight (obtained from static light scattering). The reaction stoichiometry predicts 676 charges
per particle while we calculate 494 charges per particle from conductometric titration data.

286

G.L. Flickinger et al. / J. Non-Newtonian Fluid Mech. 87 (1999) 283305

Dynamic light scattering data analyzed by Contin [30] and the method of cumulants indicates that
the average particle diameter in a dispersion dialyzed to equilibrium against deionized water is approximately 34 nm with a standard deviation of 15 nm. Solution density measurements indicate that the
particle density is 1.15 0.01 g/ml (gram dry polymer per ml occupied in solution). This value is in close
agreement with the value of 1.14 g/ml reported in the technical information bulletin provided with the
R-9637 samples. The as-received dispersion is at a pH of 8 and has a solids concentration of 35% by
weight. The dispersion contains the following volatile organic compounds on a weight percent basis:
8.4% N-methyl-2-pyrrolidone and 2.0% triethylamine. The N-methyl-2-pyrrolidone is a high boiling coalescent solvent used to enhance film formation [3]. The triethylamine is a neutralizing base and a chain
extender used in the polyurethane synthesis [31].
The manufacturer suggests that the system stabilization arises solely from the ionized carboxylic groups
[1,32]. Polyurethanes stabilized by carboxylic groups display a distinct instability at a pH near the pKa of
the stabilizing acid group. The R-9637 dispersion aggregated at a pH near 5. Increases in ionic strength
to 0.2 M also induced aggregation.
2.2. Sample preparation
Prior to experimental use, the dispersions were thoroughly washed by dialysis to remove the excess
organic solvents and any possible unknown salts. To accomplish this, the dispersions were placed in
individual dialysis membranes and submerged in deionized (DI) water. Due to large osmotic pressure
differences, the dispersions swell. The dialysis sacks were then submersed in polyethylene glycol (PEG)
solution with a higher osmotic pressure and the suspensions were allowed to equilibrate resulting in an
increase in PU concentration. This procedure is repeated in fresh solutions of PEG and DI water until the
conductivity of the dialysate held constant at the value measured initially for DI water. The sample is then
reconcentrated using PEG. This concentrated sample is then diluted with DI water to obtain the desired
concentrations for study. When salt is added to the system, the same procedure as outlined above was
followed with the only exception being that the solution outside the membrane was held at the desired salt
concentration. In this case, the concentrated master batch was diluted to the appropriate concentration
using the matching molarity KCl solution.
The diluted specimens were mixed at high speed for 1 h and then left to sit for 2 days, ensuring a
homogenous concentration. The mass concentration of the samples was determined by the weight loss
of drying at 100 C. A sample volume fraction could be determined from the measurements of the mass
concentration using the measured solution density and the density of the particles as a conversion.
2.3. Particle characterization measurements
Dynamic light scattering experiments were performed using a Spectra-Physics 60 mW HeNe laser at
a wavelength of 632.8 nm, a Brookhaven Instruments BI-200SM goniometer, and a photomultiplier tube
with the output signal processed by a BI-9000AT digital correlator. The data were analyzed by Contin and
the method of cumulents to obtain an average particle size and an approximate particle size distribution.
Static light scattering experiments were performed with a DAWN DSP-F laser photometer manufactured by Wyatt Technology Corporation. The apparatus is equipped with a vertically polarized, 5 mW,
632.8 nm wavelength laser. The scattered intensity of the incident laser light is measured by an array of
18 photodiodes mounted at fixed angles. Toluene was used as the reference for calibration of absolute

G.L. Flickinger et al. / J. Non-Newtonian Fluid Mech. 87 (1999) 283305

287

intensities. The data were plotted on a standard Zimm plot in order to determine the particle molecular
weight (Mw), radius of gyration (Rg ), and second virial coefficient (A2 ).
Static light scattering experiments were performed on sample concentrations ranging from 2.0 104
to 2.0 105 g/ml. Extremely small PU concentrations were necessary, because sample scattering was
intense. Using a Zimm plot, we extrapolate to c = 0 and apply the fundamental equation for static light
scattering [33]:
Kc
1
2B2 NA
=
+
c,
2
R
MW
MW

(6)

where K is the systems optical constant

K=

4 2 [no (dn/dc)]2
.
NA 4

(7)

c is the particle concentration, R is the Rayleigh ratio, MW is the particle weight average molecular
weight, no is the solvent refractive index, dn/dc is the refractive index increment, o is the wavelength of
the light source in vacuum, is the detector angle, and NA is Avogadros number.
Flow birefringence experiments were performed using a system constructed in the laboratory of A.J.
McHugh. A complete description of the optical bench, the experimental procedures and data analysis
procedures used for the birefringence measurements is available elsewhere [34]. A couette flow cell with
shear rates in the range of 0.001600 s1 was used to measure the flow birefringence and the orientation
angle of several dispersions. The requirement of a clear, transparent sample limited the measurements to
the dilute range.
Solution densities were determined using a Mettler/Kem DA-100 density meter. Particle charge measurements were conducted using a standard conductometric titration procedure with 0.024 M HCl as the
titrating acid and 0.02 M NaOH as the titrating base. A YSI model 34 conductance-resistance meter with
a standard dip type cell having a constant of 1.0 mho/cm was used to measure solution conductivities.
2.4. Rheological measurements
Dilute rheological data were collected using a standard CannonFenske 50 capillary viscometer. The
viscometer was immersed in an insulated, constant temperature bath at 25 C. DI water was used as the
reference standard, since it is the continuous phase for the PU dispersion. All samples were temperature
equilibrated to 25 C prior to data collection. Capillary viscometry experiments were repeated 34 times
to ensure measurement accuracy.
A constant stress (Bohlin CS-10) and a constant shear rate (Bohlin VOR) rheometers were used to
obtain the steady state shear flow and dynamic oscillatory measurements. Both rheometers are equipped
with a temperature control unit capable of maintaining the sample temperature at 0.2 C of the set point.
For the constant stress rheometer, two measuring systems were used to cover the stress range required
for obtaining flow curves of the dispersions with various concentrations. A cone and plate (cp 4/20)
with 4 angle was used for the concentrated samples whereas a cup and bob (c14) was used for the less
concentrated specimens.
The samples were loaded in the rheometers using a syringe with no needle (cp 4/20 geometry) or a
pipette (c14 geometry). The samples were covered with a solvent trap to ensure no evaporation took place
and measurements were taken immediately. Even after more than 2 h of measurements, no drying of the

288

G.L. Flickinger et al. / J. Non-Newtonian Fluid Mech. 87 (1999) 283305

specimen was observed or detected. The rheological measurements included stress viscometry, stress
sweeps and oscillatory shear flow. All rheometer measurements were conducted at 20 C.
3. Results and discussion
3.1. Dilute dispersions
3.1.1. Static light scattering experiments
Static light scattering measurements of PU dispersions suspended in deionized water as a function
of concentration yield a particle Mw = 1.27 0.08 106 g/mol and an average second virial coefficient
for the system of 3.54 0.06 1015 cm3 . For rigid particles, the second virial coefficient can also be
derived from statistical mechanics [3536],
Z
B2 = 2
x 2 (eV (x)/kT 1)dx,
(8)
0

where x is the distance between particle centers, V(x) is an interparticle pair potential, k is Boltzmans constant, and T is the absolute temperature. To estimate the magnitude of the attractions and repulsions giving
rise to B2 , the measured value is generally compared to the second virial coefficient for an equivalent hard
sphere repulsion B2HS = 2 3 /3 where, is the particle diameter. If B2 /B2HS > 1, the particles have a net
repulsive interaction while if B2 /B2HS < 1, the particles are attractive. Using an average particle diameter
of 34 nm as determined from dynamic light scattering for PU particles suspended in deionized water
B2 /B2HS = 43.0. In 102 M KCl where the particle diameter is 29 nm, B2 = 2.8 0.3 1016 cm3 and
B2 /B2HS = 5.5.The reduction in B2 indicates that with increased ionic strength, interparticle repulsions
are reduced. However, even in 102 M KCl the particles continue to experience a strong enough repulsion
that they remain colloidally stable. Experimental observations indicate that the dispersions are stable to
aggregation until suspended in an electrolyte of ionic strength 0.2 M.
3.1.2. Flow birefringence experiments
Modulated polarimetry was used to generate the flow birefringence data shown in Fig. 1. As is customarily done, the birefringence has been normalized with respect to concentration. The lack of superposition
in Fig. 1 suggests that even at low concentrations nonlinear effects are important. The steady state birefringence, which is a measure of the difference in refractive index between the fast and slow optical axis
of the suspension, increases monotonically with shear rate reflecting the continuous deformation and
orientation of the particles with the flow field. The suspension birefringence is reduced with increasing
ionic strength, suggesting that particle deformability is decreased. These observations are consistent with
the results of Zebrowski and Fuller [37] and Dunlap et al. [38] who studied the rheo-optical behavior of
aqueous solutions of sodium polystyrene sulfonate. The orientation angle of the fast optical axis relative
to the velocity direction for the dispersions in Fig. 1 varied between 2535 and was not a function of
shear rate in the range 50500 s1 . Whereas linear polymers have been known to orient fully with the
flow (i.e., the orientation angle approaches zero), suspensions of soft spheres have been shown to give
orientation angles close to 40 [23]. This is attributable to different elongational and polarization axis.
These results suggest that in dilute dispersions, PU particles deform and partially align in the shear field
at shear rate above 10 s1 .

G.L. Flickinger et al. / J. Non-Newtonian Fluid Mech. 87 (1999) 283305

289

Fig. 1. Change in specific refractive index increment from flow birefringence measurements with shear rate for the R-9637 PU
dispersion at several concentrations and ionic strengths.

3.1.3. Capillary viscometry

Capillary viscometry experiments were performed on PU dispersions ranging in concentration from
0.004 to 0.17 g/ml. The intrinsic viscosity data of PU dispersions at different ionic strengths shows
behavior expected for highly charged polymers [8,3942] or highly charged, rigid, colloidal particles
[41,4346] (Fig. 2). At low suspending medium ionic strength, the upturn in red for c < 0.05 g/ml is
indicative of double layer expansion and the resulting increase in the effective particle size. The minimum
in red near 0.05 g/ml is associated with increased domination of hydrodynamic interactions.
For background ionic strengths capable of swamping contributions to the ionic strength from the
polyurethane particles, we expect:
/c 1
= [] + kH []2 c,
(9)
c
where kH is the Huggins coefficient. For flexible polymer chains in a good solvent, kH is between 0.3 to 0.5
[42,47,48]. Systems with macromolecular associations display higher values of the Huggins coefficient
[42]. For particle systems kH , is a measure of the strength of particle interactions increasing from a value
of approximately unity as attractions or repulsions grow [10]. The measured Huggins coefficients given
in Table 1 are all larger than unity.
As shown in Fig. 3, increases in concentration of PU dispersion (dialyzed to equilibrium against
deionized water) dramatically increases the dispersion conductivity, K. The solid line in Fig. 3 is calculated
from K = Ko Ac with Ko
= 0 and A = 7.56 where K has units of mMho/cm and c has units of g/ml. These
results suggest that when PU particles are suspended in deionized water or 104 M KCl, the effect of the
counter ions on the intrinsic viscosity will only be felt when c  0.01 g/ml. When the suspending medium
has an ionic strength of 103 M KCl, the effect of counter ions can only be neglected for c < 0.03 g/ml.
Not surprisingly, the reduced viscosity curves in Fig. 2 converge for c > 0.05 g/ml for PU dispersions in
deionized water, 104 M KCl and 103 M KCl. The effect of counter ions associated with the PU particles
red =

290

G.L. Flickinger et al. / J. Non-Newtonian Fluid Mech. 87 (1999) 283305

Fig. 2. Reduced viscosity measurements as a function of concentration for PU dispersions suspended in ionic strengths ranging
from DI water to 101 M KCl.
Table 1
Low volume fraction characterization of PU dispersions
[I]

[] (ml/g)

KH

DI water
104 MKCl
103 MKCl
5 103 M KCl
102 M KCl
101 M KCl

9.4a
8.4a
7.73
3.57
2.82
3.04

1.05
6.04
7.19
2.42

Rh (nm)
12.4
11.9
11.6
9.0
8.3
8.5

[]2.5 (ml/g)

C (g/ml)

3.76
3.36
3.09
1.43
1.13
1.22

0.17
0.19
0.21
0.45
0.57
0.52

These values were determined by truncating the capillary viscometry data at low concentration where the viscosity divergence
occurs.

is less important as the ionic strength of the suspending fluid is increased. Thus the reduced viscosity
curves show less convergence with increasing concentration for ionic strengths greater than 5 103 M.
Treating the PU dispersion as consisting of spheres, the particle hydrodynamic radius (RH ) can be
estimated from capillary viscometry data by


3qMw 1/3
,
(10)
RH =
4 NA

G.L. Flickinger et al. / J. Non-Newtonian Fluid Mech. 87 (1999) 283305

291

Fig. 3. Conductivity of PU samples dialyzed against DI water as a function of concentration in g/ml. Horizontal lines indicate
the conductivity of KCl solutions of the indicated ionic strengths.

Fig. 4. Average PU particle diameters from a second cumulant analysis of dynamic light scattering data () and from intrinsic
viscosity measurements (N).

where q = []/2.5 and NA is Avagadros number. For deionized water and 104 M KCl, [] is determined
by extrapolation of the linear portion of the reduced viscosity curve. Results of these calculations are
summarized in Table 1 where it is seen that RH decreases with increasing ionic strength.
Similar trends are seen for the particle sizes extracted from dynamic light scattering measurements
(Fig. 4). Discrepancies in these two estimates of particle radius are poorly understood, but may be

292

G.L. Flickinger et al. / J. Non-Newtonian Fluid Mech. 87 (1999) 283305

attributable to particle deformability at the high shear rates where [] is measured and the broad particle
size distribution (which influences the value of molecular weight used in Eq. (10)). Both estimates,
however, show a decrease in size as ionic strength increases.
As the particle concentration increases, the effective ionic strength increases. The resulting change
in ionic strength of the suspending medium will result in a shrinking of the particles. The combined
particle size and suspension conductivity data, suggest that for c > 0.15 g/ml all dispersions will exist in
a state similar to that when dilute dispersions are suspended in a continuous phase with an ionic strength
between 101 and 102 M. Table 1 suggests that in this state, the particles have reached a minimum
size with RH = 89 nm (or a dynamic light scattering radius of 14 nm). This observation has a dramatic
impact on the concentration where the dispersion viscosity will diverge. This critical concentration can
be estimated to occur at an effective volume fraction of eff = 0.63. Using dilute suspension properties
this corresponds to a mass concentration C = 0.63/q. As shown in Table 1 for background ionic strengths
of 103 M and below, C drops to values less than 0.17 g/ml, well below the value where the zero shear
rate viscosity diverges. These results add weight to the birefringence and particle size data suggesting
that with increasing ionic strength or concentration, particle repulsions are diminished and the particles
shrink in size.
From the dilute solution characterization, we conclude that when PU dispersions are approximated as
deformable spheres. The sphere diameter and deformability decrease as the background ionic strength
increases. The counter ions balancing the particle charge contribute in a substantial manner to the dispersions effective ionic strength. As a consequence, the effects of the counter ions can only be swamped
when the dispersion concentration is kept below a critical value. For a supporting medium ionic strength
of 103 M, this concentration is approximately 0.02 g/ml while for 102 M this critical value is 0.15 g/ml.
While double layer repulsions will remain a function of PU concentration, the particle size shrinks to a
constant size for c > 0.15 g/ml or ionic strengths greater than 102 M.
3.2. Concentrated dispersions
In characterizing the flow properties of dense dispersions, we focus initial attention on suspensions
dialyzed to equilibrium against deionized water. After discussing these suspensions we turn to the effects
of increasing the background ionic strength.
3.2.1. De-ionized water dispersions
3.2.1.1. Steady state shear viscosity. Viscosity measurements in the concentrated regime were performed at a sample temperature of 20 C with a solvent trap. For T > 20 C viscosity increases with time
were observed. This was attributed to solvent losses due to drying. Shear rate dependant thickening was
not observed for this system at any concentration for the experimentally accessible shear rates.
Even in the most concentrated suspensions studied, steady state viscosity was obtained only seconds
after stress application. At 20 C, suspension viscosities were found to remain constant for time periods
in excess of 2 h. Viscosities were independent of the rheometer tool geometry. The flow curves were
extremely reproducible not only between sample loadings, but also with increasing and decreasing stress
sweeps.
Flow curves for PU dispersions in deionized water are shown in Fig. 5. Three types of behavior are
apparent. The first extends up to a mass concentration c 0.35 g/ml where a Newtonian response is

G.L. Flickinger et al. / J. Non-Newtonian Fluid Mech. 87 (1999) 283305

293

Fig. 5. Relative viscosity of PU dispersions suspended in deionized water as a function of shear rate at several concentrations
expressed in g/ml.

obtained at all shear rates. The second covers the concentration range of 0.36 g/ml < c < 0.42 g/ml where
a Newtonian low shear (zero-shear) plateau viscosity is observed followed by shear thinning. Finally
when c > 0.43 g/ml, the behavior is that of a shear thinning fluid with an apparent yield stress. The zero
shear rate viscosity could not be measured and the viscosity decays as x with x 1 indicating the
onset of a yielding type of response.
The flow properties of PU dispersions will be governed by interparticle thermodynamic and hydrodynamic forces. The thermodynamic forces arise primarily from particle charge, and from the loops and
tails of the polymer comprising the particles. Evidence for the particulate nature of these suspensions can
be developed by comparing the characteristic Peclet number for shear thinning for PU dispersions and for
hard sphere suspensions [10,49]. The Peclet number, Pe characterizes the relative rates of deformation
and diffusion:
Pe =

a 2
6 c a 3
=
,
Do
kT

(11)

where Do is the particle diffusivity, a is the particle radius and the shear rate. A Peclet number characterizing shear thinning, Pec can be defined from the shear rate, c , required for the viscosity to decrease
to halfway between the limiting high and low shear rate viscosities, o and respectively. For dense
suspensions o /  1 and Pec occurs at a shear rate where = 0.5o . Due to difficulties in defining
an effective volume fraction for the PU dispersions, rather than comparing Pec at equal values of , we
compare Pec at equal values of o /c . Rueb and Zukoski [50] report that hard sphere suspensions with zero
shear rate relative viscosities of 104 and 3 105 , have Pec 2 103 and 6 105 respectively. Using
a radius of 14 nm at the same zero shear rate relative viscosities, the PU dispersions have Pec 1 103
and 3 105 respectively. The similarity of values of Pec suggests that shear thinning and the rapid
decrease in Pec as particle concentration is increased have the same physical origins in hard sphere suspensions and in PU dispersions. Thus as c increases, we conclude that the increases in viscosity and rapid

294

G.L. Flickinger et al. / J. Non-Newtonian Fluid Mech. 87 (1999) 283305

Table 2
Cross-equation parameters for concentrated dispersions
[I]

Concentration (g/ml)

o (Pa s)

c (s1 )

DI

0.436
0.415
0.413
0.402
0.385

5024 80
365 2
294 2
12.8 0.1
9.6 0.1

0.096 0.003
1.58 0.04
2.02 0.07
65.1 1.2
91.2 2.5

0.81
0.82
0.82
0.90
0.90

102 M KCl

0.457
0.432
0.425
0.416
0.405
0.401

943 4
219 1
100 1
28.9 0.1
10.4 0.1
4.1 0.1

0.79 0.02
3.1 0.1
6.5 0.1
23.8 0.7
85.5 3.6
248 4

0.81
0.83
0.83
0.86
0.86
0.95

101 M KCl

0.421
0.413
0.403
0.370

5864 175
846 7
455 4
1.62 0.01

0.023 0.002
0.158 0.004
0.20 0.01
106 2

0.80
0.78
0.73
0.79

increases in characteristic stress relaxation times are associated with the approach of the close packing
limit where the particle long time self diffusivity goes to zero [29,49].
The particulate nature of the PU dispersion is further supported by comparing the full shear rate
dependence of the flow curves of PU and hardcore dispersions. Shear thinning in hard core suspensions is
well described by Eq. (3). For our system, o  and the data is well described by the Cross equation
in the following form:
1

=
.
o
1 + ( /c )

(12)

Fits to Eq. (12) are summarized in Table 2. The values compare well with those obtained by Jones et. al.
[18] on hard sphere suspensions 0.50 < < 0.73 and those obtained by Jones et al. [19] 0.547 < < 0.815
and Mewis et al. [12] 0.708 < < 1.15 on sterically stabilized suspensions. Mewis et al. [12] were able
to measure high shear plateau viscosities ( ) for particles with diameters of 1220, 475, and 84 nm. High
shear plateau viscosities were obtained for Pe > 10, suggesting that will not be reached for the PU
dispersions studied here until 105 s1 .
The flow properties of the PU dispersions at different volume fractions and ionic strengths can be
collapsed onto a single master curve (Fig. 6). The master curve can be constructed by normalizing the
viscosity and shear rate with o and c , respectively. This allows one to capture the flow behavior of the
PU system from Eq. (12) with only two parameters (o (c), c ). This curve is well approximated with
= 0.8 independent of c or ionic strength. To achieve a similar collapse for the sterically stabilized system
of Jones et al. [19], was found to be a function of concentration.
The stress characterizing shear thining can be calculated from c = 0.5o c To compare suspensions
composed of particles of different size we work with the hard sphere scaling, where the reduced critical
shear stress is define:
rc =

c a 3
.
kT

(13)

G.L. Flickinger et al. / J. Non-Newtonian Fluid Mech. 87 (1999) 283305

295

Fig. 6. Master curve created by normalizing the sample viscosity by the zero shear viscosity(o ) and the shear rate by the critical
shear rate (c ) Values of o and c were determined by fitting Eq. (12) To the flow curve data of PU dispersions suspended in
deionized water (0.385 c 0.436), 102 M KCl (0.405 c 0.457) and 101 M KCl (0.370 c 0.413). A summary of the
values for o and c and are given in Table 2.

Papir and Krieger [51] originally assumed that rc = 1 for hard sphere suspensions independent of particle
size and concentration. Later, de Kruif et al. [24] showed that rc is concentration dependant with a maximum at approximately = 0.50. This maximum occurs at the concentration where hard spheres undergo
a disorder to order transition. Although the magnitude of the peak for a model soft sphere suspension was
smaller than obtained for hard spheres, Mewis et al. [12] argued that soft sphere suspensions displayed the
same transition and concluded that rc was not very sensitive to particle hardness. Rueb and Zukoski [50]
found that rc is weakly dependant on the strength of particle attraction.For their micogel system, Wolfe
and Scopazzi [21] report that the magnitude of the maximum in rc in agreement with linear polymer
behavior. Wolfe and Scopazzi also showed that the volume fraction at the maximum in rc increased with
increasing particle crosslink density (hardness). This was attributed to two ompeting effects: (1) particle
deswelling at high eff and (2) polymer segment tails at the particle surface which enhance longer range
particle interactions. Particle deswelling will dominate the behavior for particles with high cross-link
densities, acting to increase eff at rc maximum. With low particle cross-link densities, polymer tails are
expected to dominate the behavior eff at rc maximum. The critical reduced stress for the PU system
suspended in deionized water is plotted with the calculated eff = cq in Fig. 7. In this work we use q
determined from intrinsic viscosities measured for particles suspended at an ionic strength of 102 M.
For purposes of comparison in Fig. 7 are also plotted the data of Wolfe and Scopazzi [21] for microgel
particles containing 0.25 and 4.0% crosslinker and data from Jones et al. [19] on 59 nm diameter soft
spheres. These results suggest to PU dispersions have a softness between the microgel particles of Wolfe
and Scopazzi.
The divergence of the zero-shear viscosity is indicative of a transition from a disordered fluid to a
glassy solid. This transition will occur as the system approaches the limit of close packing of equivalent

296

G.L. Flickinger et al. / J. Non-Newtonian Fluid Mech. 87 (1999) 283305

Fig. 7. Reduced critical stress calculated from Eq. (13) as a function of the concentration expressed in effective volume fraction.
Also presented are data on microgel particles containing 0.25% and 4.0% crosslinker from Wolfe and Scopazzi (1989) and data
on sterically stabilized soft spheres from Jones et al. (1992).

Fig. 8. Relative zero shear viscosity as a function of effective volume fraction. Also plotted are data on charged spheres from
Buscall et. al. (1982). The solid line shows the prediction for model hard sphere suspensions. The line through the data of Buscall
et al. is drawn from Eq. (1) with m = 0.144. Also shown is Eq. (1) with m = 0.50 (dash-dotted line) intermediate concentrations.
A better fit is obtained if the exponent in Eq. (1) is set at 3 (dotted line).

hard spheres ( meff ) [49,52,53]. An example of this divergence for highly charged, hard core particles
can be found in the work of Buscall et al. [11] who fit their data exceptionally well with Eq. (1) by letting
m = 0.144 (Fig. 8). Data for model soft sphere systems with small soft layers has also been fit effectively

G.L. Flickinger et al. / J. Non-Newtonian Fluid Mech. 87 (1999) 283305

297

using an effective hard sphere volume fraction [12,14]. As the particle softness increases, Mewis et al.
[14] showed that Eq. (1) becomes less effective at describing the data.
Eq. (1) was used to fit the PU data in Fig. 8 with m = 0.50. Eq. (1) under predicts the viscosity at intermediate concentrations and over predicts the viscosity high concentrations. The smaller value of do /d
at high concentration is attributable to the PU particles ability to deform and compress. The enhancement
of viscosity at intermediate concentrations is suggestive of longer range interactions. Wolfe and Scopazzi
[21] observed this same behavior for low crosslink density polymethylmethacrylate microgels. For this
system, the viscosity enhancement was attributed to increased interparticle interactions from low polymer
segment density tails at the particle surface. This effect was more pronounced for the lower cross-link
density microgels. The behavior of the PU data in Fig. 8 can be captured much more effectively if the
exponent in Eq. (1) is changed from (2) to (3).
Despite the failure of hard sphere based models to capture the divergence of o as c is increased, we
anticipate that at m the particles become close packed. This assumes that in this low shear environment,
electrostatics will be the dominant interaction and thus the deformation seen at elevated shear rates will
not be important. If we choose a hardcore approach, we can relate the effective maximum packing fraction,
meff , to that where hard spheres of diameter d form a glass by
 3
2a
meff = 0.64
,
(14)
d
where a is the particle radius and d is the effective hard sphere particle diameter [10]. For hardcore
particles of radius a experiencing electrostatic repulsions, can be estimated from



1
(15)
d = ln

ln(/(ln . . . ))
where 1 is the Debye length and is defined by
=

4 o s2 a 2 k exp(2ak)
,
kT

(16)

where o is the permittivity of free space, is the relative permittivity of the suspending medium, and s is
the particle surface potential [10]. For the data of Buscall et al. [11], a surface potential of 2535 mV yields
meff = 0.15 and Eq. (1) predicts experimental data well. For the PU system, using meff = 0.50, a = 14 nm,
and an ionic strength of 5 102 M, Eqs. (1416) yield d = 30.4 nm and s = 22 mV. The surface potential
can be related to the number of charges on the particle surface by the linearized PoisonBoltzman Equation
[10,54] yielding 200 charges per PU particle. The calculated charge is lower than the 494 charges per
particle obtained from a conductometric titration. As much of the particle charge will be internal to the
particle and thus screened, it is not surprising that the estimation from meff yields a lower charge than
that determined by titration.

3.2.1.2. Dynamic oscillation. Above meff the dispersion behaves as a viscoelastic solid. Dynamic
stress sweeps were performed at a constant frequency of 1 Hz to obtain values of the storage modulus
(G0 ), the loss modulus (G0 ), and the complex viscosity ( ). Fig. 9 shows the variation of G0 with the
strain for various PU concentrations. For c < 0.41 g/ml ( eff < 0.49 calculated from Eq. (5)) the storage

298

G.L. Flickinger et al. / J. Non-Newtonian Fluid Mech. 87 (1999) 283305

Fig. 9. The elastic modulus of PU dispersions suspended in deionized water as a function of strain at a frequency of 1 Hz. The
PU dispersion concentration is given in g/ml.

modulus is small as would be expected for a disordered fluid. For c > 0.41 g/ml the particles begin to
interact strongly as indicated by the increase in G0 . As the strain increases, this structure is destroyed,
resulting in a decrease in G0 . The extent of the linear viscoelastic response, characterized by a constant G0 ,
is concentration dependent. Higher concentrations lead to a reduction in the critical strain ( crit ) separating
the linear and the nonlinear viscoelastic response. If the cut-off between the two responses (linear and
non-linear) is chosen as the point where G0 is 90% of the plateau value, increasing the mass concentration
from 0.41 to 0.43 g/ml ( eff of 0.49 and 0.51) reduces crit from 0.056 to 0.036. These critical strains are
typical of gels composed of charge stabilized particles as opposed to aggregated suspensions [55,56].
The concentration and strain dependencies of G00 are shown in Fig. 10. Several features may be noted.
First, G00 increases with increasing concentration. Again moderate increments in c lead to orders of
magnitude increase in G00 . Secondly, G00 > G0 for c < 0.42 g/ml ( eff < 0.50). However, for c > 0.43 g/ml
( eff > 0.50) and with the formation of an apparent gel, G0 becomes higher than G00 . Thirdly, a different
behavior is observed for the highest concentration measured. G00 first increases, reaches a maximum and
then decreases; such a behavior has been associated with materials having yield stresses [20].
Dynamic oscillatory flow experiments were conducted at a strain of 2% (ensuring linear viscoelastic
response) and over the frequency range of 0.0130 Hz. Figs. 11 and 12 show, the variation of the G0 and
G00 with , the angular frequency. At low concentrations, terminal behavior is observed with G0 increasing
with 2 and G00 as . As concentration is increased, near c = 0.42 g/ml the relaxation rate (as characterized
where G0 = G00 ) decreases rapidly. This behavior occurs over a very narrow concentration range (i.e., the
characteristic relaxation time moves from 10 s to <103 s as eff increases from 0.50 to 0.51). At high
concentration, the elastic modulus is independent of frequency. At a frequency of 1 s1 , increasing c
from 0.35 to 0.44 g/ml ( eff of 0.42 to 0.52) decreases phase angle () from 82 to 9 These measures
of the onset of glassy behavior agree well with the rapid increases in viscosity seen near c 0.42 g/ml
( eff = 0.50) (Fig. 8).

G.L. Flickinger et al. / J. Non-Newtonian Fluid Mech. 87 (1999) 283305

299

Fig. 10. The loss modulus of PU dispersions suspended in deionized water as a function of strain at a frequency of 1 Hz. PU
concentrations are expressed in g/ml.

Fig. 11. The elastic modulus of PU dispersions suspended in deionized water at a strain of 0.02 as a function of frequency. PU
concentrations are expressed in g/ml.

The CoxMerz correlation is obeyed by this system only in regions where the system can obtain a
Newtonian shear plateau. As can be seen from Fig. 13, the CoxMerz rule where () = ( ) is only
found to be applicable in the region of small and and for c less than the gel concentration or the
onset of an apparent yield point for the system. In the regions of shear thinning, the dynamic viscosity
(0 ) is always greater than the shear viscosity (). This result is observed in several structured systems
[20]. This result lends further support to our essentially particulate description of the PU dispersions.

300

G.L. Flickinger et al. / J. Non-Newtonian Fluid Mech. 87 (1999) 283305

Fig. 12. The loss modulus of PU dispersions suspended in deionized water at a strain of 0.02 as a function of frequency. PU
concentrations are expressed in g/ml.

Fig. 13. Viscosity as a function of shear rate for constant deformation rate experiments (closed symbols) and complex viscosity
(open symbols) as a function of frequency. PU dispersions are suspended in deionized water and concentrations are reported in
g/ml. The CoxMerz rule is only obeyed in the low shear viscosity limit.

3.2.2. Effect of the ionic strength

The results discussed in the previous section are associated with samples diluted with deionized water.
To investigate the effect of screening the double layer repulsion, the flow properties of dense PU samples
were measured in 102 and 101 M KCl. Due to difficulties in dialyzing these PU samples to specific
concentrations, experimentally tested samples were made from a concentrated master batch dialyzed to

G.L. Flickinger et al. / J. Non-Newtonian Fluid Mech. 87 (1999) 283305

301

Fig. 14. Zero shear relative viscosity as a function of the effective volume fraction determined from Eq. (5). Shown are data for
PU dispersions suspended in deionized water, 102 M KCl, and 101 M KCl.

equilibrium with 101 or 102 M KCl by dilution with the appropriate ionic strength solvent. For both
the 101 M and the 102 M KCl dispersions. the trends in the viscosity-shear rate experiments, the stress
sweep experiments, and the frequency sweep experiments were very similar to the deionized dispersion.
The best fit parameters extracted from the Cross equation for each ionic strength are shown in Table
2. The master curve in Fig. 6 clearly shows that shear thinning occurs in a similar manner for all ionic
strengths investigated.
The relative zero shear viscosity (or ) at each ionic strength as a function of effective volume fraction
( eff ) calculated from Eq. (5) is presented in Fig. 14. As expected, at low to medium concentrations, both
the 102 and 101 M KCl samples display lower viscosities than an equivalent concentration dispersion
dialyzed to equilibrium against deionized water. This behavior is consistent with the findings of numerous
studies on dense charge stabilized suspensions [11,55,57]. The decreases in viscosity with increasing
ionic strength are associated with compression of the electrostatic double layer by the inert electrolyte.
A detailed analysis of this result is complicated by a lack of knowledge of the specific ionic condition in
the sample. At higher PU concentrations, the 102 M dispersion continues to follow this trend.
For c > 0.45 g/ml, the zero shear rate viscosities of the dispersion in 101 M KCl are higher than the
respective deionized samples. At an ionic strength of 0.2 M KCl, the PU dispersions aggregate. Second
virial coefficient measurements suggest the particles are less repulsive with increasing ionic strength.
These results suggest that attractions become important as the ionic strength increases. The data in Fig.
14 indicates that these attractions are weak but grow for c > 0.45 g/ml when the background ionic strength
is 101 M.
Reduced critical stresses for each ionic strength and concentration are shown in Fig. 15. The 101 M
samples display slightly different values than the lower ionic strength samples. Using octadecyl silica in
decalin, Rueb and Zukoski [50] found that attractions made very little difference in the reduced critical
stress behavior of hard core systems. The relative magnitude and flattening behavior of the 101 M KCl
data are similar to that obtained on soft 42 nm PMMA particles [16,17]. Frith associated flattening of the

302

G.L. Flickinger et al. / J. Non-Newtonian Fluid Mech. 87 (1999) 283305

Fig. 15. Reduced critical stress calculated from Eq. (13) as a function of the effective volume fraction determined from Eq. (5)
for PU suspended in deionized water, 102 M KCl, and 101 M KCl. Also shown are data obtained by Frith (1986).

reduced critical stress with increasing particle softness [16,17]. This conclusion is not consistent with the
PU dispersions where increases in ionic strength appear to increase particle rigidity.
A simple Maxwell model with a single relaxation time could not be used to fit the frequency dependencies of G0 and G00 . The G0 data of Fig. 11 along with the data for the other ionic strengths was fit to a
system of Maxwell elements in parallel such that a high frequency limiting storage modulus, G0 , could
be estimated. As the concentration of the dispersions increases, additional relaxation times were required
to fit the modulus data. As a result, a single characteristic relaxation time for each respective sample was
unattainable. Fig. 16 shows the G0 values obtained for each PU sample. The data of Fig. 16 indicate
that there is virtually no mechanical differences between the glassy structures formed by the PU particles
at elevated concentrations. This result is not surprising as the ion concentrations in the three dispersions
will be very similar at these elevated PU concentrations.
4. Conclusions
Here we present a detailed study of the rheological properties of a commercial aqueous PU dispersion.
These systems are composed of soft particles whose size is a function of both ionic strength and dispersion
concentration. Flow birefringence experiments demonstrate that the PU particles deform and orient in a
shear field. These findings support the soft nature of the particles suggested by Satguru et al. [1,32].
Due to the small particle size, shear thinning does not occur for c < 0.35 g/ml. For higher concentrations,
shear thinning is observed and the zero shear rate viscosity diverges near c 0.42 g/ml ( eff 0.50).
Strong interactions and glassy responses are observed at concentrations below that for randomly packed
hard spheres and many sterically stabilized systems [12,17,19]. However, these PU particles are highly
charged. In addition, their polyelectrolyte nature suggests polymer chains or tails may extend from
the particle core. As a consequence, either through electrostatic repulsions or a combination of charge

G.L. Flickinger et al. / J. Non-Newtonian Fluid Mech. 87 (1999) 283305

303

Fig. 16. The asymptotic high frequency elastic modulus of PU dispersions suspended in deionized water, 102 M KCl and 101 M
KCl as a function of the effective volume fraction determined from Eq. (5).

and steric effects, these systems gel or form glasses at volume fractions near 0.5. The approach to glass
formation is characterized by the rapid change in frequency where G0 = G00 and the rapid decrease in Pec
as the particle concentration is increased.
While we anticipated that background ionic strength would play a significant role in dispersion rheology,
as the particles themselves carry a substantial charge, in a dense suspension, the effects of the addition of
inert electrolyte are muted up to a point where a combination of ions balancing particle charge and those
from the background electrolyte approach an effective ionic strength of 0.1 M. At this point interparticle
attractions can become important.
Acknowledgements
This work was supported by the Materials Research Laboratory at the University of Illinois at Champaign-Urbana by the U.S. Department of Energy through Grant DEFG02-96-ER-45439. Additional support was provided by the 3M Company St. Paul, Minnesota. The authors would like to thank A.J. McHugh
for the use of his flow birefringence apparatus.
References
[1] R. Satgurau, J. McMahon, J.C. Padget, R.C. Coogan, Aqueous polyurethanes for surface coatings, Presented at the Congress
of the Federation of Scandinavian Paint & Varnish Technologists, Copenhagen, 1993.
[2] D. Dieterich, Aqueous emulsions dispersions and solutions of polyurethanes: synthesis and properties, Prog. Organic
Coatings 9 (1981) 281340.
[3] J.W. Rosthauser, K. Nachtkamp, Waterborne polyurethanes, J. Coated Fabrics 16 (1986) 3979 .
[4] P.H. Markusch, R.E. Tirpak, Water-borne polyurethane ionomers, Presented at the Water-borne and Higher-solids coatings
Symposium, New Orleans, LA, 1990.

304

G.L. Flickinger et al. / J. Non-Newtonian Fluid Mech. 87 (1999) 283305

[5] B.Y. Kim, T.K. Kim, Aqueous dispersion of polyurethanes from H12 MDI, PTAd/PPG, and DMPA: particle size of dispersion
and physical properties of emulsion cast films, J. Appl. Polym. Sci. 43 (1991) 393398.
[6] C.K. Kim, B.K. Kim, IPDI-based polyurethane ionomer dispersions: effects of ionic, nonionic hydrophilic segments and
extender on particle size and physical properties of emulsion cast film, J. Appl. Polym. Sci. 43 (1991) 22952301.
[7] H.X. Xiao, S. Yang, J.A. Mclean, J. Kresta, K.C. Frisch, D.P. Higley, R.A. Smiley, Aqueous anionic polyurethane dispersions,
Presented at the Water-borne and Higher-solids coatings Symposium, New Orleans, LA, 1990.
[8] W.C. Chan, S.A. Chen, Conformations of polyurethane cationmers in organic solvents, and in organic/water mixtures,
Polymer 34 (1993) 12651270.
[9] M. Dreja, K. Heine, B. Tieke, G. Junkers, Effects of functionalized latex particles and anionic surfactants on the flow
behavior of aqueous gelatin dispersions, J. Colloid Interface Sci. 181 (1997) 131140.
[10] W.B. Russel, D.A. Saville, W.R. Schowalter, Colloidal Dispersions, Cambridge University, Cambridge, 1989.
[11] R. Buscall, J.W. Goodwin, M.W. Hawkins, R.H. Ottewill, Viscoelastic properties of concentrated latices: I Methods of
examination; II Theoretical analysis, J. Chem. Soc., Faraday Trans. 78, (1982) 28732887, 28892899.
[12] J. Mewis, W.J. Frith, T.A. Strivens, W.B. Russel, The rheology of suspensions containing polymerically stabilized particles,
AIChE J. 35 (1989) 415422.
[13] M.M. Cross, Rheology of non-Newtonian fluids: a new flow equation for pseudoplastic systems, J. Colloid Sci. 20 (1965)
417437.
[14] G.N. Choi, I.M. Krieger, Rheological studies on sterically stabilized model dispersions of uniform colloidal spheres, J.
Colloid Interface Sci. 113 (1986) 101113.
[15] C. Prestidge, T.F. Tadros, Viscoelastic properties of aqueous concentrated polystyrene latex dispersions containing grafted
poly(ethylene oxide) chains, J. Colloid Interface Sci. 124 (1988) 660665.
[16] M.A. Ansarifar, P.F. Luckham, Viscoelastic properties of polyacrylonitrile particles stabilised by a poly-2-vinylpridine/poly
tert butylstrene block copolymer, Colloid Polym. Sci. 267 (1989) 736740.
[17] W.J. Frith, T.A. Strivens, J. Mewis, Dynamic mechanical properties of polymerically stabilized dispersions, J. Colloid
Interface Sci. 139 (1990) 5562.
[18] A.R. Jones, B. Leary, D.V. Boger, The rheology of a concentrated colloidal suspension of hard spheres, J. Colloid Interface
Sci. 147 (1991) 479495.
[19] A.R. Jones, B. Leary, D.V. Boger, The rheology of a sterically stabilized suspension at high concentration, J. Colloid
Interface Sci. 150 (1992) 8496.
[20] R.J. Ketz, R.K. Prudhomme, W.W. Graessley, Rheology of concentrated microgel solutions, Rheol. Acta 27 (1988) 531
539.
[21] M.S. Wolfe, C. Scopazzi, Rheology of swellable microgel dispersions: influence of crosslink density, J. Colloid Interface
Sci. 133 (1989) 265277.
[22] H.N. Nae, W.W. Reichert, Rheological properties of lightly crosslinked carboxy copolymers in aqueous solution, Rheol.
Acta. 31 (1992) 351360.
[23] S. Raquois, C. Fridrikh, J.F. Tassin, S. Rezaiguia, Rheological behavior of concentrated suspensions of soft spheres, J.
Chem. Phys. 93 (1996) 941959.
[24] C.G. De Kruif, E.M.F. van Iersel, Hard sphere colloidal dispersions: viscosity as a function of shear rate and volume fraction,
J. Chem. Phys. 83 (1985) 47174725.
[25] van der Werff, G. deKruif, Hard-sphere colloidal dispersions: the scaling of rheological properties with particle size, volume
fraction, and shear rate, J. Rheol. 33 (1989) 421454.
[26] J.C. Van der Werff, C.G. de Kruif, C. Blom, J. Mellema, Linear viscoelastic behavior of dense hard sphere suspensions, J.
Phys. Rev. A 39 (1989) 795807.
[27] A.T.J.M. Wouterson, C.G. de Kruif, The rheology of adhesive hard sphere dispersions, J. Chem. Phys. 94 (1991) 57395749.
[28] M.C. Grant, W.B. Russel, Volume fraction dependence of elastic moduli and transition temperatures for colloidal silica
gels, Phys. Rev. E. 47 (1993) 26062614.
[29] L. Marshall, C.F. Zukoski, Experimental studies on the rheology of hard sphere suspensions near the glass transition, J.
Phys. Chem. 94 (1990) 11641171.
[30] S. Provencher, A constrained regularization method for inverting data represented by linear algebraic or integral equations,
Comput. Phys. Commun. 27 (1982) 213227.
[31] R.F. Tirpak, P.H. Markusch, Aqueous dispersions of crosslinked polyurethanes, J. Coatings Technol. 58 (1986) 4954.

G.L. Flickinger et al. / J. Non-Newtonian Fluid Mech. 87 (1999) 283305

305

[32] R. Satguru, J. Mcmahon, J.C. Padget, R.G. Coogan, Aqueous polyurethanes-polymer colloids with unusual colloidal,
morphological, and application characteristics, J. Coatings Technol. 66 (1994) 4756.
[33] P. Kratochvil, Clasical Light Scattering from Polymer Solutions, Elsevier Science, New York, 1987.
[34] A. Imaneni, Conformational and orientational dynamics of semi-rigid macromolecules in shear flow, Ph.D. Thesis,
University of Illinois, Urbana, IL, 1997.
[35] D. McQuarrie, Statistical Mechanics, Harper Collins, New York, 1976.
[36] P.C. Heimenz, R. Rajagopalan, Principles of Colloid and Surface Chemistry, 3rd ed., Marcel Dekker, New York, 1997.
[37] B.F. Zebrowski, G.G. Fuller, Rheo-optical studies of polyelectrolyte solutions in simple shear flow, J. Rheol. 29 (1985)
943954.
[38] P.N. Dunlap, C.H. Wang, L.G. Leal, An experimental study of dilute polyelectrolyte solutions in strong flows, J. Appl.
Polym. Sci. 25 (1987) 22112238.
[39] S. Nomula, S.L. Cooper, Structural studies of polyurethane ionomer solutions 1. Dynamic light scattering, Macromolecules
30 (1997) 13551362.
[40] M. Hara (Ed.), Polyelectrolytes, Marcel Dekker, New York, 1993.
[41] M. Hara, J. Wu, A.H. Lee, Solution properties of ionomers 1. Counter ion effect, J. Polym. Sci. B: Polym. Phys. 25 (1987)
14071418.
[42] J. Niezette, J. Vanderschueren, L. Aras, Dilute solution behavior of n-butyl methacrylate ionomers, J. Polym. Sci.: Polym.
Phys. Ed. 22 (1984) 18451848.
[43] J. Yamanaka, H. Matsuoka, H. Kitano, N. Ise, Revisit to the intrinsic viscositymolecular weight relationship of ionic
polymers, J. Colloid Interface Sci. 134 (1990) 92106.
[44] M. Hara, J. Wu, A.H. Lee, Solution properties of ionomers 2. Simple salt effect, Macromolecules 22 (1989) 754757.
[45] J. Cohen, Z. Priel, Y. Rabin, Viscosity of dilute polyelectrolyte solutions, J. Chem. Phys. 88 (1988) 71117116.
[46] R.D. Lundburg, R.R. Phillips, Solution behavior of metal sulfonate ionomers II. Effect of solvents, J. Polym. Sci.: Polym.
Phys. Ed. 20 (1982) 11431154.
[47] A. Ait-Kadi, P.J. Carreau, G. Chauveteau, Rheological properties of partially hydrolyzed polyacrylamide solutions, J. Rheol.
31 (1987) 537561.
[48] M.L. Huggins, The viscosity of dilute solutions of long-chain molecules IV. Dependence on concentration, J. Am. Chem.
Soc. 64 (1942) 27162718.
[49] J.F. Brady, Rheological behavior of concentrated colloidal dispersions, J. Chem. Phys. 99 (1993) 567581.
[50] C.J. Rueb, C.F. Zukoski, Rheology of suspensions of weakly attractive particles: approach to gelation, J. Rheol. 42 (1998)
14511475.
[51] Y.S. Papir, I.M. Krieger, Rheological studies on dispersions of uniform colloidal spheres, J. Colloid Interface Sci. 34 (1970)
126130.
[52] M.H. Cohen, D. Turnbull, Molecular transport in liquids and glasses, J. Chem. Phys. 31 (1959) 11641169.
[53] A. Doolittle, Studies in Newtonian flow II. The dependence of the viscosity of liquids on free space, J. App. Phys. 22 (1951)
14711475.
[54] R.J. Hunter, Foundations of Colloid Science, Vol. I, Oxford University Press, New York, 1986.
[55] L.B. Chen, C.F. Zukoski, Flow of ordered latex suspensions: yielding and catastrophic shear thinning, J. Chem. Soc. Faraday
Trans. 86 (1990) 26292639.
[56] C.J. Rueb, C.F. Zukoski, Viscoelastic properties of colloidal gels, J. Rheol. 41 (1997) 197218.
[57] H.M. Laun, Rheological properties of aqueous polymer dispersions, Die Angew Makromol. Chemie 123/124 (1984) 335
359.