JOURNAL OF APPLIED POLYMER SCIENCE VOL. 14, PP.

2753-2762 (1970)

The Dynamic Viscosity of Polypropylene Melt

A. N. DUNLOP and H. LEVERNE WILLIAR/LS,

Department of Chemical Engineering and Applied Chemistry,
University of Toronto, Toronto 181 , Ontario, Canada

Synopsis
The dynamic viscosity of polypropylene melt was determined as a function of fre-
quency and temperature. The temperature-shift Arrhenius energy was calculated in
several ways. The influence of temperature on the elastic and viscous components was
quantitatively discussed.

INTRODUCTION
The temperature dependence of the melt viscosity of polypropylene
has been studied by several investigator^.'-^ These studies have been
made under various conditions. Mendelson' carried out his experiments
using capillary flow techniques; van der Vegt3 subjected the polymer to
rotational shear as well as capillary flow; and Adamese and co-workers'
sheared the polypropylene melt in an oscillatory manner. The values for
the Arrhenius energy of flow (Eq) as determined by the first two tech-
niques are not in agreement with the third; the present authors considered
that this discrepancy, for a linear polymer such as polypropylene, was
worthy of further work. In this investigation, the temperature depen-
dence under oscillatory shear was studied. A number of different methods
of interpretation of data were considered, and the Arrhenius energy values
were compared with literature.

EXPERIMENTAL
The instrument used was a Weissenberg rheogoniometer Model R. 18,
incorporating an air-bearing head.6 A 2-degree cone of 5.0 cm diameter
and a flat platen of the same diameter were used in oscillation. The tem-
perature range was 195"-26OoC. Nitrogen gas was passed through the
furnace system during each run. The frequency range and amplitude
range were 0.006 to 6 sec-' and 50 to 500 microns, respectively; thus these
experiments were considered to be small-amplitude sinusoidal deformations.
The viscosities plotted in Figure 1 %-ere measured from stress-strain
amplitudes taken as an average over several wavelengths. These amplitudes
were reproducible to *l%. Viscosity was found to be independent of
the amplitude of the applied strain.
2753
@ 1970 by John Wiley & Sons, Inc.
 2754 DUNLOP AND WILLIAMS

4. I

4.6

4.5

4.4
t
'f 4.3
2
-I
4.2

4.1

4.0

3.9

3.8

3.7
-2.2 -2.0 -1.8 -1.6 -1.4 -1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1.0
Log Freq -P

195°C; (@) 210°c; ( 0 )222OC; ( 0 )231OC; ( 0 )246OC; (a)

Fig. 1. Dynamicviscosity: (0)
260°C.

Testing of polypropylene under oscillatory shear can be evaluated in

the following manner (see Fig. 2). The oscillation of the cone can be
described' by
y ( t ) = yoeiw' (1)
where yo is the amplitude of the applied strain, w is the frequency, and
t is the time. Then the rate strain is

?(t) = iwy(t). (2)

The stress can also be described by a sinusoidal function with the same
frequency as the rate of strains:
p*12= Poeiw*+(+ (3)
where Po is the amplitude of the tangential stress and t$ is the phase dif-
ference between the stress and strain. The viscosity is the real part 7' of
the complex viscosity':

Thus the dynamic viscosity is

 DYNAMIC VISCOSITY OF POLYPROPYLENE MELT 2755

Fig. 2. Typical strain and stress curves for a viscoelastic material.

Bogie and Harriss have pointed out that the amplitude is subject to at-
tenuation, and the phase difference, to a shift at higher frequencies due to
the filter system of the rheogoniometer. The present authors have found,
as did van Rijn,'O that this can be compensated for in most cases. Where
there was attenuation, corrections were made.
The material studied was Enjay polypropylene grade 115 supplied by
Esso Chemical Canada. This material is FDA stabilized and has a melt
flow index of 5.0.

RESULTS AND DISCUSSION

Oscillatory measurements were carried out at different temperatures.
Plots of log viscosity versus log shear are shown in Figure 1. As can be
seen from the curves, molten polypropylene has a pseudoplastic rheological
behavior, the viscosity decreasing with increasing shear rate. The effect
of temperature is apparent.

Effect of the Method of Determining 7'o

The viscosity of zero shear rate, 7'0, for polypropylene was determined
in three ways: (1) from extrapolation of the log 7' versus logf plot (Fig. 1) ;
(2) from a method developed by Cross"; and (3) from a method employing
the theory developed by Walters and Kemp. l2
The 7'0 as obtained from extrapolation gave rise to an Arrhenius energy
of flow (E7) of 10.3 kcal/mole (Fig. 3). It has been pointed out by Men-
delson' that, for constant shear stress,
UT = i w ( r 4 / i w ( T )= T ~ ~ ~ (ref)
( T ) / ~ ~ (6)
~ ~
where uT is a temperature shift factor. A capillary flow shear rate,
i,, is considered at an arbitrarily chosen reference temperature, Yw (ref),
and the temperature in question, i w ( T ) . Similarly, 7app is a viscosity,
taken from steady shear studies. Using uT in an Arrhenius-type plot,
one obtains a shift factor energy, E,. Applying this concept to the oscil-
2756 DUNLOP AND WILLIAMS

latory data at zero shear, the E, in the present case is 10.3 kcal/mole.
Mendelson's work yielded an E, = 10.0 kcal/mole for polypropylene.
Another earlier capillary flow study2 resulted in an Eq value of 9.3 kcal/
mole. An Arrhenius energy of 11 kcal/mole was obtained by van der

c 0.6
.-0
Q)

n
Y

-
-f
Oe4

0.2

0
0 0.4
Freq'lz -
0.8 1.2
(sec-'/Z)
1.6 0

Fig. 4. Test of Cross theory, polypropylene at 195OC.

Vegt.3 The above measurements were likely carried out on polymers of

comparable properties; thus it seems probable that the dependence of
% on temperature is the same for oscillatory as for steady shear, at zero
shear.
 DYNAMIC VISCOSITY OF POLYPROPYLENE MELT 2757

0.6

0.4
t
I.

2
-I
0.2

-
0.0
1.8 1.9 2.0 2.1 2.2
10*/T (“K-1)

Fig. 5. Arrheniw plot of relaxation time from Cross calculation.

Another method for handling the data is to use the following equation
correlated by Cross for non-Newtonian bulk flow:

where b is chosen empirically in order to get the data to fit the eduation,
110 and 11- are the limiting viscosity values at frequency f = 0 and f = a ,
respectively, and r can be related to relaxation time. The present data
+
fit b = (Fig. 4) using the least-squares met,hod, and the correlation
coefficients in all cases were greater than 0.99, the standard error in 1/11’
was less than *4%. This value of b compares favorably with other poly-
mer melts under oscillatory shear.” If one recalls Ferry’s definition of
aT, i.e., the ratio of relaxation t,imes at t,wo different temperatures, then

5.0 I .
t 4*8 ‘

M
3 4.6 I

4.4

I 1 1 1

-
4.2

10*/T (“K-1)

Fig. 6. Arrhenius plot of TO from Cross calculation.

2758 DUNLOP AND WILLIAMS

I
4.8 -
4.6 -
t
c
8 4.4
-I
.
4.2 -
4.0
1.8 1.9 2.0 2.1 2.2
+ (OK-*)

Fig. 7. Arrhenius plot for 9’0 from Walters theory.

one should be able to relate an Arrhenius-type plot of 7 to that of aT.

From Figure 5, a value of 9.2 kcal/mole is obtained for E, (shift energy),
which is in fair agreement with the values obtained by Mendelson and
van der Vegt. However, an Arrhenius-type plot of qo‘ produces a value
of 13.0 kcal/mole for the energy of flow (Fig. 6). Adamese and co-workers4
obtained a value of 16.0 kcal/mole from oscillatory shear, although this
value was obtained using only two temperatures.
An attempt was also made t o relate the data to the theory of Walters
and Kemp.12 However, the data only fit at the very limit as the shear
approaches zero; Walters and Kemp suggest “that this interpretation of
the experimental results is best suited to reasonably viscous liquids that

0.6

t
*
0.4
M
-I

0.2

0.0
1.8 1.9 2.0 2.1 2.2
10S/T 4 (OK-‘)
Fig. 8. Arrhenius plot relaxation time from Walters theory.
 DYNAMIC VISCOSITY OF POLYPROPYLENE MELT 2759

1.6

1.5

1.4

C
.-
m
TI
d 1.3
v

t
CI
8
Y

- 1.2
0

a
P)

m
.
I=
P
1.1

1.0

3
0 1.0 2.0 3.0
Freq (f) -+ (sec-1)

Fig. 9. Plot of phase difference vs. frequency: (0)

195°C; ( 8 ) 231°C; ( 0 ) 246°C; ( 0 )
270°C.

are not highly elastic. . .if 7’ and G are known over a limited (low) fre-
quency range.” Predictions of the theory yield

7’ = lm N(T)d7 - 4T2$: (8)

where

N(7)dT = 7’0

J- 7N( T ) d7 = 7’ 0 7 *,
T* being a “characteristic” relaxation time. This can be considered to be
a “measure” of the important relaxation times. The other “moments”
2760 DUNLOP A N D WILLIAMS

r 2 and r 3 were not determined owing to lack of fit of the theory. G is the
dynamic rigidity and N ( T ) is a distribution function of relaxation time T .
Plots of q' versus$ and of G/4r2f2 versus f 2 in the very low frequency range
yielded the above integrals from which Arrhenius energy of flow and shift
were obtained (Figs. 7 and 8). These were found to be 10.0 and 9.9 kcal/
mole, respectively. This close agreement with previous data, it should be
observed, is better than the actual theoretical fit.

Relationship of the Phase Difference with Temperature

The phase difference between the stress (output) and the strain (input)
gives a measure of the elastic to viscous components of the polymer (or
vice versa). The effect of the temperature on a number of representative
curves is seen in Figure 9, which is a plot of phase angle 4 versus frequency
f. I n addition to the kinetic runs, shear experiments were carried out
below T,, that is, on rigid polypropylene, and a t 270°C in order to extend
the range. The tests below T , were carried out on the same instrument
as the melt work, but solid sample holders were used in place of the platens.
All of the curves were found to fit the equation
4 = Afm + C (10)
where C = 1.57 (phase angle in radians = 90"). The above equation
was tested by means of a statistical F-test, which checked the linear, quad-
ratic, and the higher-order regressions, i.e. , the linear regression component
was found to be significant, the quadratic regression components and
residual (high orders) were found to be not significant. The fit was ex-
tremely good in all cases for the melt. Those experiments performed on
the solid samples had poor correlation coefficients, i.e., 0.60-0.70, but
they did satisfy the F-test. A and m were determined and plotted against
temperature in Figures 10 and 11. In Figure 11, A would be asymptotic
to A = 0, i.e., a viscous fluid, neglecting fluid inertia,12gives a stress 90"

0.0 - ~ * 0 -

-0.4 -
t 8
a
-0.8 - l

8
8

8
-1.2 - 8

- 1.6 .
---o d:*
.
 DYNAMIC VISCOSITY OF POLYPROPYLENE MELT 2761

0.6 -

E
t
0.4

0.2
-
- t
t
7
0

0 - - ' -- -o
*
t

-0.2 > . . a . . . .

out of phase with the strain (a line 4 = 1.57). As the compound approaches

-
a completely elastic state, A approaches -1.57. Considering this in
conjunction with Figure 11, one can see that m approaches zero at the
low temperatures, i.e., 4 0, giving the expected result of an elastic
compound having the stress and strain in phase. It can be seen that in
the region of T, there is a definite transition from a primarily viscous to
an elastic response. The overall temperature effect on A and m is well
described in Figures 10 and 11.

SUMMARY OF RESULTS
The shift Arrhenius energy for polypropylene melt under oscillatory
shear is 10 kcal/mole. It appears that, if the Arrhenius energy is con-
sidered in terms of flow, it may be dependent on the method of calculation,
whereas the Arrhenius energy measured from the shift in the curves (re-
laxation) a t different temperatures apparently is not. It also appears
that small-amplitude sinusoidal shear gives the same Arrhenius energy of
flow as capillary and steady shear methods.
The effect of temperature on the elastic and viscous components of
polypropylene can be quantitatively discussed using the equation 4 =
Af" + C.
This study was supported by Imperial Oil Enterprises Limited and the National Re-
search Council of Canada.

References
1. R. A. Mendelson, J . PoZym. Sci. B, 5,295 (1967).
2. G . V. Vinogradov and N. V. Prosorovskaya, Rheol. Acta, 3,156 (1964).
3. A. K. van der Vegt, Trans. Plast. Inst., 32,165 (1964).
4 . J. W. C. Adamese, H. Jareschitz-Kriegl, J. L. Den Otter, and J. L. S. Wales, J .
Polym. Sn'. A-%,6,871(1968).
5. G. Petraglia, A. Coen, and V. Sangermano, Soc. P h t . Eng., Tech. Papers, 14.
168 (1968).
2762 DUNLOP AND WILLIAMS

6. The WeissenbergRheogoniometer Model R.18 Instruction Manual) Sangamo Con-

trols Ltd., Sussex, England.
7. J. D. Ferry, Viswehtic Properties of Polymers, Wiley, New York, 1961, p. 33.
8. K. Weissenberg, The Testing of Materials by Means of the Rheogawmeter, San-
gamo Controls, Sussex, 1966, p. 7.
9. K. Bogie and J. Harris, Rheol. Ada, 5,212 (1966).
10. C. F. H. van Rijn, Rhwl. Ada, 6,295 (1967).
11. M. M. Cross, J . Appl. Polym. Sci., 13,765 (1969).
12. K. Walters and R. A. Kemp, in Polymer Systems, R. E. Wetton and R. W. Whor-
low, Eds., Macmillan, London, 1968, p. 237.
Received May 4, 1970
Revised July 9, 1970