Remarks on Non-Shear Thinning Elastic Fluids

R. J. BINNINGTON and D. V. BOGER

Department of Chemical Engineering

The University of Melbourne
Parkville, Victoria, 3052,
Australia

In response to difficulties encountered in measuring and

reproducing the flow properties of Boger fluids (dilute solutions
of polyarcylamide in corn syrup), Choplin, et al. proposed a
particular variety of lubricant (Emkarox) and also solutions of
polyacrylamide (Pusher 700) in mixtures of glycerine and water
as better model fluids for molten polymers than the original
Boger fluid. In response to this work it is argued that the model
fluid systems proposed by Choplin, et al., although interesting
in concept, have insufficient elasticity and hence are of limited
potential for use in investigative work in polymer fluid me-
chanics. Of all the highly elastic constant viscosity fluids avail-
able, the polyisobutylene-polybutenesystem is the most attrac-
tive. Both the properties of this organic Boger fluid and the
original Boger fluid, are entirely consistent with simple fluid
theory.

INTRODUCTION with simple fluid theory in the limit of low shear

n recent years non-shear thinning highly elastic rates. In light of their difficulties, the authors pro-
I fluids (the so called Boger Fluids (1-3)) have been
used by a number of researchers to study the effect
posed that a particular variety of lubricant (Emka-
rox) and also solutions of polyacrylamide (Pusher
of elasticity on flow fields in the absence of any 700) in mixtures of glycerine and water would be
significant shear thinning (4-18). Much of this work better model fluids for molten polymers than the
has been motivated by a need to verify the results original Boger fluid (polyacrylamide in corn syrup).
being generated from numerical solutions for non- In this paper we will demonstrate that the difficul-
shear thinning elastic fluids in polymer processing ties encountered by Choplin, et al. with the poly-
flow fields (19). acrylamide corn syrup fluid system were probably
Other workers (20-24) have been concerned due to experimental technique, that the fluids they
with gaining a better understanding of the flow propose as model materials may not exhibit suffi-
properties of these unique fluid systems. The works cient elasticity to be of interest, and that of all the
of Gupta, et al. (21, 22) and our own (20) have non-shear thinning elastic fluid systems now avail-
clearly shown that non-shear thinning elastic fluids able, the polyisobutylene in polybutene system may
represent a new class of fluid behavior entirely be the most attractive.
consistent with molecular and continuum theory COMMENT ON THE CHOPLIN, ET AL.
and that such ideal elastic fluid systems can be PAPER
constructed by dissolving a number of different
high molecular weight polymers in a number of Choplin, et al. (23) reported the following diffi-
different viscous Newtonian fluid solvents. Possibly culties with the polyacrylamide-corn syrup solu-
the most interesting fluid to emerge from these tion:
works is the organic Boger fluid composed of a high “(i) The rheological properties were not repro-
molecular weight polyisobutylene in a Newtonian ducible from one batch to another. Some
solvent of kerosene and polybutene. The features solutions appeared to be highly elastic, but
of this fluid system are discussed later. no stable readings of the normal force on the
The recent work of Choplin, et al. (23) is of rheogoniometer could be obtained. Stable
particular interest in that a number of difficulties and homogeneous solutions could be pre-
were encountered measuring the rheological prop- pared through considerable stirring, result-
erties of one of the non-shear thinning fluid systems ing in some mechanical degradation of the
(the polyacrylamide-corn syrup system). In addition polymer and lower normal stresses.
these authors perpetuate the idea that the behavior (ii) The corn syrups exposed to air form a thin
of constant viscosity elastic fluids is not consistent coat of crystals at the free surface . . . . . . .

POLYMER ENGINEERING AND SCENCE, JANUARY, 1986, Yo/. 26, No. 2 133
 R. J . Binnington and D.V. Boger

extreme care has to be taken in measuring al. that the first normal stress differences of the
the rheological properties in a cone-and- corn syrup solutions appear to be quite sensitive to
plate instrument.” temperature. A degree of sensitivity of N1 to tem-
The difficulties reported above (i) are most cer- perature is certainly expected. Figure 1 shows the
tainly due to inhomogeneities in the fluid samples. relationship of both N1 and viscosity to temperature
Efficient dispersal of the polymer throughout the for a Boger fluid (at 20°C 77 = 35 Poise, and A0 1=1
solvent is the key to the proper preparation of 0.45 s). As expected both the first normal stress
Boger fluids. The reported observation that “stable difference and the viscosity decrease with increas-
and homogeneous solutions could be prepared ing temperature (measurements made at 100 s-’).
through considerable stirring” is evidence to sug- At higher shear rates it is very difficult if not im-
gest that in some cases the test fluids may not have possible to keep the sample in the gap of the cone
been mixed to a homogeneous state before fluid and plate geometry used for these measurements
property measurements were made. Our normal with the Weissenberg Rheogoniometer. This same
practice is to mix at a low shear rate over a long difficulty is encountered with molten polymers and
period of time which may extend to a number of other polymer solutions at higher shear rates and
days. The first normal stress difference as a function may be the cause of the unusual normal force re-
of shear rate is measured on samples periodically sponse observed by Choplin, et al. At shear rates at
withdrawn from the mixing vessel. When the first and below 274 s-’, the authors report a normal
normal stress difference remains constant with transient response in the normal force measure-
shear rate for successive samples, a uniform sample ment, i.e., the normal force quickly rises to a max-
has been constructed. In a flow visualization exper- imum and constant value. However at higher shear
iment with the polyacrylamide-corn syrup solu- rates (435 s-l and 1727 s-’) they report that the
tions, the shear stress and the first normal stress normal force quickly rises to a large value which
difference are measured before and after the flow then does not remain constant but decreases essen-
field experiment. We have encountered no signifi- tially to zero (at both shear rates) after a few min-
cant difference in the flow properties before and utes of shear. If this response is indeed due to severe
after flow visualization experiments conducted now viscous heating (7”C/minute at .i. = 435 s-l) as the
on a number of different flow fields ( 2 , 5 , 9-1 1). In authors report, then a significant decrease in the
fact the polyacrylamide-corn syrup fluid has been viscosity would have been detected. Such a de-
used recently by two workers to investigate the crease in the viscosity was not reported and in fact
influence of elasticity on the torque required for Choplin, et al. report that the viscosity of the corn
mixing (16, 18). Mechanical degradation of the syrup solutions is not sensitive to temperature, i.e.,
polymer due to mixing is not observed if agitation the viscosity of the 550 ppm solution at 30°C was
is limited to low levels of shear. Inefficient mixing slightly higher than at 25°C. Such unusual viscosity
of these fluids does result in inhomogeneities which temperature behavior is not observed in these fluid
do produce normal stresses which are not stable systems (see Fig. I). Indeed a disadvantage of the
and which may be higher than the equivalent ho- polyacrylamide-corn syrup solutions is the extreme
mogeneous solution. The same problem is also en- sensitivity of the viscosity to temperature.
countered in polyacrylamide in water solutions. If
care and patience are employed, then reproducible
and stable polyacrylamide in corn syrup solutions
can be constructed. The same comments also apply /’

for the preparations of any of the highly elastic /

/’

constant viscosity fluids which are now available.

The second difficulty reported by Choplin, et al.,
that, of “crystallization” of the corn syrup surface
on exposure to air, in the form of a thin skin, does
arise if care is not taken to limit the evaporation at
the free surface. A relevant study on eliminating I
evaporation effects on the Weissenberg Rheogo- /
/

niometer was reported by Boger and Rama Murthy /

I

in 1969 (2.5). For cone-and-plate characterization

of any evaporatively sensitive system, including the
corn syrup fluids, a thin film of nonvolatile immis-
cible material (e.g. silicone oil) of comparable vis-
cosity should be applied to the free surface of the
test sample to eliminate evaporation effects. If care
is taken in characterizing these fluids then problems Fluid:0.0175 w/w% separan AP30
4% water, 95.98% sweetol C.S.
due to evaporation can be rendered insignificant.
Test shear rate:100 s-’
However, it is true that the polyacrylamide in corn l . O l ’ I * ~ ’ I ’ ” I ’ ~ . I1.0
syrup fluid is not a good choice for conducting a 2.9 3.0 3.1 3.2 3.3 3.4 3.5
free surface experiment, like the observation of die (I/T)x lo3, ( O K - ’ )
swell. Fig. 1 . Temperature dependence of the viscosity and first normal
It was also reported in the paper by Choplin, et stress dgference for a Bogerfluid.

134 POLYMER ENGINEERING AND SCIENCE, JANUARY, 7986, Vol. 26, No. 2
 Remarks on Non-Shear Thinning Elastic Fluids

In complete contrast to the apparent thixotropic shear stress measurements were then performed on
characteristics of the normal force observed by the sample by Professor Bohme and his colleagues
Choplin, et al. (23) at high shear rates, a phenom- at the Hochschule der Brundeswehr in Hamburg.
enon of anti-thixotropy in both the shear stress and The sample was then characterized dynamically in
first normal stress difference has been observed by terms of storage modulus (G') and loss modulus,
Jackson, et al. (24) at the upper shear rate limit for (G"), at the Technical University of Darmstadt by
Boger fluid property measurements in both the Professor Spurk's research team.
cone and plate and parallel plate configurations of The data shows excellent agreement in the shear
the Weissenberg Rheogoniometer. According to stress from the independent laboratories. Compar-
Jackson, et al. (24) the phenomenon is reproduci- ison of the dynamic data of Spurk to the steady
ble. At a particular (high) shear rate both the shear shear property measurements confirms the expec-
stress and the first normal stress difference slowly tation of simple fluid theory in that the following
increase with time from a pseudo-stable level to a relationships hold
secondary stable level which appears to be history
dependent. After examining this phenomenon in Lim n' = Lim rl = v0 (1)
4 7-0
both the cone and plate and parallel plate geome-
tries, Jackson, et al. concluded that the effect can and
be eliminated at a given rim shear rate in the
parallel plate system by both halving the gap and
rotational speeds. There is some evidence then
suggesting that for highly elastic fluids that the flow The validity of E q 1 for this fluid system has been
field may no longer be viscometric at higher shear established by numerous researchers (10, 17, 20,
rates in the cone and plate and parallel plate con- 24) but the correct limit for the storage modulus
figuration, i.e., a secondary flow may form. Al- and the first normal stress difference was estab-
though this explanation is pure speculation, we lished for the first time by this data which was
have also observed this apparent anti-thixotropy obtained in 1982. Jackson, et al. (24) in their recent
behavior at high shear rates for the corn syrup- publication were unable to substantiate E q 2 but in
polyacrylamide solutions and for the polybutene- their own words their results were not inconsistent
polyisobutylene solutions. There are also other ex- with it. We conclude that there is nothing unique
amples of shear-thickening in polymer solutions in about the low shear rate, low frequency behavior
the literature (26, 27). It is therefore very difficult of these fluid systems. They do conform to our
to understand and interpret the high shear rate understanding of simple fluid theory.
normal stress data of Choplin, et al. Further evidence to substantiate the reproduci-
bility of Boger fluid data is shown in Fig. 3 (2). A
LOW SHEAR RATE BEHAVIOR test fluid containing 0.048 percent Separan MG500
The reproducibility of the steady shear property in corn syrup and water was prepared and charac-
measurements for a Boger fluid is shown in Fig. 2 terized in the Chemical Engineering Department
along with dynamic data which clarifies misconcep- at Monash University. The sample was then sent to
tions about the lower limit for the storage modulus the Department of Chemical Engineering at the
(in relation to the first normal stress difference). A University of Delaware where J. C. Chang per-
fluid sample (0.02 percent polyacrylamide, 4 per- formed an independent characterization using a
cent water and 95.98 percent corn syrup) was pre- Weissenbery Rheogoniometer. Note the excellent
pared in our laboratory and characterized with an
R-19 Weissenberg Rheogoniometer. Independent
3.2% water, 96.75% corn syrup MCy41N
I I ' ' """I 1
Temp:22.0 "C

712
. -
G- 102 r
E
t
c
:
6 -
e
c
L
m

0 Bohme
A A Boger
0 Spurk

10'I p/. , , , ....

I , . . ,,,.,I ' 100
100 10' 102 lo-' 100 10' 102
7 , w , (s-1) Shear rate, s-'
Fig. 2. Steady and dynamic shear properties of a Bogerfluid Fig. 3. Typical steady shear characteristics of a Bogerfluid show-
illustrating low shear rate behavior and reproducibility of data. ing reproducibility of data.

POLYMER ENGINEERING AND SCIENCE, JANUARY, 7986, Vol. 26, No. 2 135
 R. I. Binnington and D. V. Boger

agreement between the two laboratories and that fluids are also optically clear, are less sensitive to
this result was first reported in 1978. temperature, are free from biological attack over
It is now clear that non shear-thinning and highly long periods of storage, and are not sensitive to
elastic fluids can be constructed by dissolving a evaporation. However the major advantage of the
high molecular weight polymer in a very viscous polyisobutylene-polybutene fluids is that an ex-
Newtonian solvent and that this behavior is not tended region of quadratic normal stress behavior
unique to the polyacrylamide in corn syrup solu- is observed. Such behavior is important in an ideal
tions where the behavior was first observed (1, 2). elastic fluid and is illustrated by a comparison of
Such behavior has also been observed for polyiso- the first normal stress data in Fig. 3 for the poly-
butylene dissolved in polybutenes (20, 21) and acrylamide-corn syrup fluid to the data in Fig. 4 for
nicely explained within the framework of dilute the polyisobutylene-polybutenefluid.
solution theory in the paper by Gupta, et al. (21).
The steady and dynamic shear properties of both EMKAROX FLUIDS AND PUSHER 700
these fluid systems are consistent with the Oldroyd SOLUTION
B constitutive equation at low shear rates (20, 28). The Emkarox fluids and the Pusher 700 solutions
Typical shear and first normal stress difference mea- also are constant viscosity elastic liquids. However
surements for a polyisobutylene-polybutenefluid in comparison to the polyacrylamide-corn syrup,
are shown in Fig. 4 . This particular fluid was con- polyisobutylene-polybutene fluids and molten pol-
structed by first dissolving the polyisobutylene (Vis- ymers their level of elasticity is low. The elasticity
tanex MML-120, Exxon Chemicals) in kerosene and level of a fluid or molten polymer is sometimes
then by dissolving this mixture in the polybutene measured in terms of a Maxwell relaxation time.
(Hyvis 30, BP Chemicals). The kerosene is used
simply as an intermediate solvent for solution of the
polyisobutylene before the mixture is added and
dissolved in the polybutene. For the fluid whose
while in a flow field the influence of elasticity is
properties are shown in Fig. 4,the viscosity, 7, is
usually measured in terms of the Weissenberg num-
350 poise and constant with shear rate, while the
Maxwell relaxation time, A, (see E q I ) was 0.125 s ber
and constant for shear rates less than 10 s-'. A large
We = -
x
range of viscosities and relaxation times can be D
generated by varying both the molecular weight of -
the polybutene and the polyisobutylene, and by V
varying the concentration of the solute polyisobu- where D and V are a characteristic dimension and
tylene. Shear thinning is observed at higher solute velocity of the flow system. D / V can be regarded
concentrations. To date we have constructed con- as a characteristic time of the flow system. Numer-
stant viscosity elastic polyisobutylene-polybutene ical calculations using finite element techniques for
fluids with 10 < 9 < 1000 poise and with 0.0043 < convected Maxwell fluids (constant viscosity elastic
A < 0.6 s. In comparison to the corn syrup-poly- fluids) in comparison to experimental observation
acrylamide fluids the polyisobutylene-polybutene in die entry flows, creeping flow around a sphere
and for die exit flows have clearly shown that for
We 5 0.1 elasticity has no significant effect on the
kinematics of these flow fields relative to the New-
Fluid:0.098% PI8
(Vistanex MML-120) tonian creeping flow prediction (9, 11, 15, 20, 29-
4.87% Kerosene, 95.03% 36). Clearly for an experimental fluid to provide
useful and new information, experiments must be
conducted at sufficiently high Weissenberg number
so that deviation from the Newtonian kinematics is
observed. Since the flow fields of most interest are
creeping flows where the characteristic velocity is
small, an ideal elastic fluid can be regarded as one
which not only has a relatively high viscosity and is
non-shear thinning but also has a relatively high
relaxation time.
Figure 5 compares the relaxation times as a func-
tion of shear rate for constant viscosity elastic liq-
uids to a number of polymer melts, including the
data of Oda, et al. (37) as quoted by Choplin, et al.
(labelled DSla and D S l b in Fig. 5) and some rep-
resentative data for low density poIyethylene and
10" 100 10' 102 polystyrene (38).Note that the characteristic times
Shear rate, s-' for the molten polymers are of the order of 1 s in
Fig. 4 . Steady shear characteristics of an organic Boger fluid the shear rate range for which measurements are
illustrating the extended region of quadratic normal stress hehav- available. Also shown in Fig. 5 are similar results
ior. for two polymethylsiloxanes (2). The constant vis-

136 POLYMER ENGINEERING AND SCIENCE, JANUARY, 1986, Vol. 26, No. 2
 Remarks on Non-Sh,ear Thinning Elastic Fluids

I I I I I I
1 shear-thinning behavior and also a region of quad-
ratic normal stress behavior. The use of these ma-
terials as model fluids is becoming increasingly im-
portant in providing information about the behavior
of viscoelastic fluids in various flow regimes.
The major difficulties encountered by Choplin,
et ul. with the use of Boger fluids appear to stem
from inhomogeneities in their prepared samples
due to inefficient mixing. If prepared carefully, the

i‘lo-2t
corn syrup based Boger fluid provides a very useful,
optically clear highly elastic model fluid for use in
a wide range of flow visualization applications. Care
should be taken to minimize the likelihood of crys-
10-3
tallization of the corn syrup on exposed surfaces.
/...- El Complementing these fluids is an attractive variety
‘.. of model fluids composed of polyisobutylene in
I I I \E2
10-4 1 I
polybutene, which exhibit extended regions of
10-2 10-1 100 101 102 103
quadratic normal stress behavior. The elasticity of
Shear rate, -y (s-’)
both fluid varieties, in terms of the Maxwell relax-
Fig. 5. A comparison of Maxwell relaxation times of constant ation time, A, is comparable in magnitude to real
tiiscosity elastic liquids to typical molten polymer data: DSla-
Dow Styron 666 at 180°C (37);DSlb-Dow Styron 666 at 200°C polymer melt systems and can be varied, virtually
(37);B l , B2 and B3-polyacylamide corn syrup solutions; OB1 independently of viscosity.
and OB2-polyisobutylene polyhutene solutions; P I , P2 and P3- The existence of other model fluids certainly is
Pusher fluids (23); S1 and S2-silicone oils; and E l and E2- not precluded, but when considering the use of
Emkarox fluids. such fluids to investigate elastic effects related to
polymer melt systems, then a significant level of
cosity and extended region of second order behav- elasticity and viscosity in the model fluid is re-
ior in the first normal stress difference of polymeth- quired. Model fluids with high viscosities and low
ylsiloxanes has been known since 1967 (39). Al- relaxation times (low molecular weight silicone oils,
though the low molecular weight silicone oils (pol- and Emkarox solutions), or with low viscosity and
ymethylsiloxanes) are indeed ideal in their behavior moderate to low relaxation times (Pusher 700 so-
and do exhibit a relatively large first normal stress lutions), though interesting in concept, are limited
difference, their characteristic times are of the or- in their potential for use in investigative work into
der of s. Experience has shown that fluids with polymer melt fluid mechanics.
such low relaxation times fail to exhibit any signifi-
cant deviation from Newtonian behavior in a num-
ber of creeping flows (9, 40, 41). Clearly the prop- REFERENCES
erties of the Emkarox fluids fall into this category
with their relaxation times being on the order of 1- 1 . D . V. Boger, 1. Non-Newtonian Fluid Mech., 3, 87 (19771
4 x lop4s in comparison to the silicone oils with X 1978).
on the order of s. Although the relaxation 2. D. V. Boger and H. Nguyen, Polym. Eng. Sci., 18, 1037
times of the Pusher 700 (polyacrylamide-glycerine- ( 1 978).
water) solutions are higher (2.68 X lo-’ to 1.75 x 3. K. Walters, J. Non-Newtonian Fluid Mech., 5 , 113 (1979).
4. D . V. Boger and R. Binnington, Trans. Soc. Rheol., 21, 515
lo-” s as reported by Choplin, et ul.), they are still (1977).
low compared to the polyacrylamide-corn syrup 5 . H . Nguyen and D. V. Boger, J . Non-Newtonian Fluid Mech.,
and polyisobutylene-polybutene fluids, the normal 5, 3.53 ( 1 979).
stress behavior is non-quadratic at the lowest shear 6 . J. C. Chang and M. M. Denn,]. Non-Newtonian Fluid Mech.,
rates where measurements can be made, and the 5, 369 (1979).
7 . €3. Mena. 0.Manero, and D . M. Binding,. -I. Newtonian Fluid
viscosity is very low. Constant viscosity elastic liq- Mech., 5, 427 (1979).
uids were first developed not only to observe elastic 8. D. V. Boeer and M. M. Denn. 1. Non-Newtonian Fluid Mech..
,I

effects in the absence of any significant shear thin- 6, 163 (r980).

ning but also to make these observations in the 9. R. P. Chhabra, P. H. T. Uhlherr, and D . V. Boger, f. Non-
Newtonian Fluid Mech., 6, 187 (1980).
absence of any fluid inertia. For this purpose an 10. D . V. Boger, “Separation of Shear Thinning and Elastic
ideal elastic fluid should not only have a high relax- Effects in Experimental Rheology”,Rheology, Vol. 1, Edited
ation time, but should also have a high viscosity. by G. Astarita, G. Marrucci, and L. Nicholas, p. 19.5 Plenum
The combination of a high viscosity and high elas- Publishing Carp., New York, (1980).
1 1 . R. Binnington and D . V. Boger, Proceedings of’ the 2nd
ticity has allowed elastic effects to be observed in National Conference on Rheology (Australian Branch of the
the absence of significant shear thinning and fluid British Society of Rheology), p. 117 Sydney, (1981).
inertia in a number of flow fields of importance in 12. T. Cochrane, K. Walters, and M. F. Webster, Phil. Trans. R.
polymer processing. Soc. Lond., A 301, 163 (1981).
13. D . V. Boger, “Entry Flows of Inelastic and Viscoelastic
CONCLUSIONS Fluids,” Advances in Transport Processes, Vol. 2, pp. 43-
104, Wiley International, New Delhi (1981).
It is recognized that Boger fluids are a general 14. K. Walters and D . M. Rawlinson, Rheol. Acta., 21, S47
class of materials which exhibit approximately non (1982).

POLYMER ENGINEERING AND SCIENCE, JANUARY, 1986, Vol. 26, No. 2 137
 R. J. Binnington and D. V. Boger
15. K. Walters and M. F. Webster, Phil. Trans. R. Soc. London, 27. M-N. Layek-Raphalen and C. Wolff, J. Non-Newtonian Fluid
A 308, 199 (1982). Mech., 1, 159 (1976).
16. D. R. Oliver, A. W. Ninow, R. J. Mitson, and K. Terry, 28. R. J. Binnington and D. V. Boger, “Constant Viscosity Elastic
Chem. Eng. Research and Des., 62, 123 (1984). Liquids,” Presented at IX international Rheological Con-
17. D. Sigli and A. Maalouf, J. Non-Newtonian Fluid Mech., 9, gress, Acapulco (1984), In press, J . of Rheol.
191 (1981). 29. M. J. Crochet, Proceedings of the Third International Con-
18. R. K. Prud’homme and E. Shaqfeh, A.Z.Ch.E. 1..30, 485 ference on Finite Elements in Flow Problems, I , p. 332 Banff,
(1984). (1980).
19. M. J. Crochet and K. Walters, Ann. Reo. Fluid Mech., 15, 30. R. E. Nickell, R. I. Tanner, and B. Caswell, J. Fluid Mech.,
241 (1983). 65, 189 (1974).
20. D. V. Boger, “Dilute Polymer Solutions and Their Use to 31. M. J. Crochet and R. Keunings, J. Non-Newtonian Fluid
Model Polymer Processing Flows,” Interrelations Between Mech., 7, 199 (1980).
Processing Structure and Properties of Polymer Materials, 32. M. Viriyayuthakorn and B. Caswell, J . Non-Newtonian Fluid
Eds., J. C. Seferis and P. S. Theocaris, Elsevier Science Mech., 6, 24.5 (1980).
Publisher, Amsterdam, 307 (1984) and in Proceedings of 33. M. J. Crochet and M. Bezy, J. Non-Newtonian Fluid Mech.,
the Third National Conference on Rheology, Australian So- 5, 201 (1979).
ciety of Rheology, 11, Melbourne (1983). 34. M. J. Crochet and M. Bezy, Rheology, Vol. 2, Edited by G.
21. R. K. Gupta, G. Prilutski, T. Sridhar, arid M. E. Ryan, J. Astarita, C. Marrucci, and L. Nicolais, p. 53, Plenum Press,
Non-Newtonian Fluid Mech., 12, 233 (1983). New York (1980).
22. R. K. Gupta, T. Sridhar, and M. E. Ryan, Proceedings ofthe 3.5. R. Keunings and M. J. Crochet, J. Non-Newtonian Fluid
Third National Rheology Conference, Australian Society of Mech., 14, 279 (1984).
Rheology, 33, Melbourne (1983). 36. N. Y. Tuna and B. A. Finlayson, J. Rheol., 28, 79 (1984).
23. L. Choplin, P. J. Carreau, and A. Ait Kadi, Polym. Eng. and 37. R. Ola, J. L. White, andE. S. Clark, Polym. Eng. Sci., 18, 25
Sci., 23, 459 (1983). (1978).
24. K. P. Jackson, K. Walters, and R. W. Williams, J. Non- 38. I. Chen, Ph.D. Thesis, University of Tennessee (1971).
Newtonian Fluid Mech., 14, 173 (1984). 39. V. H. Giesekus, Rheol. Acta, 6 , 3 3 9 (1967).
25. D. V. Boger and A. V. Rama Murthy, Trans. Soc. Rheot., 13, 40. G. J. Donnelly and C. B. Weinberger, Ind. Eng. Chem. Fund.,
405 (1969). 14, 334 (1975).
26. J . F. Hutton, Proc. 5th Int. Cong. on Rheol., Kzj’ota, 4 , p. 551, 41. S. J. Lee, M. M. Denn, M. J. Crochet, A. B. Metzner, and
University Park Press, (1 968). G. J. Riggins,]. Non-Newtonian Fluid Mech., 14,301 (1984).

138 POLYMER ENGINEERING AND SCIENCE, JANUARY, 1986, Vol. 26, No. 2