Troubleshooting Extrusion Instabilities with

a Rubber Process Analyzer (RPA)

Keywords: Rubber, Elasticity, Extrusion, RPA RT005

BACKGROUND
14
The Rubber Process Analyzer (RPA) from TA Instruments can
12
be used to troubleshoot processing issues by measuring
the viscoelastic properties of the raw polymer material and 10
rubber compounds. Differences in viscoelastic properties1-3

S’ [dNm]
can lead to difficulties within the production process. In this 8
note, we present a case study where an RPA was utilized to
troubleshoot a processing problem where extrusion instabilities 6
were experienced. We will compare one batch of material that
processed normally to another batch that demonstrated scorch 4
problems in the extruder, resulting in a product with orange skin or Sample 1, Orange Peel Surface
orange peel texture. Both batches of rubber were tested with the 2
Sample 2, Smooth Surface
moving die rheometer (MDR), with representative curves shown
in Figure 1. The cure curves show that sample 1, which exhibited 0
0 2 4 6 8 10 12 14 16 18 20
scorch behavior during the manufacturing process, cured more
slowly than sample 2, which showed no processing issues. These Time [min]
results contradict the processing behavior on the production floor.
Figure 1. Moving die rheometer (MDR) cure curves of two batches of
In addition, both curves show a similar minimum torque around rubber compound show similar curing behavior. The minimum torque of
1 dNm, indicating that these two samples would process similarly. both curves is similar indicating they should process similarly, however
The results from the MDR tests were unable to distinguish behavior in the manufacturing process differed.
the differences between the materials and did not provide the
Both features are imperative for obtaining accurate data in the
information needed to identify the root cause of the processing
non-linear regime, particularly during large amplitude oscillatory
issues.
strain (LAOS) testing. The viscoelastic properties of the samples
This note will demonstrate the use of the RPA to characterize the measured in this note were obtained on the RPA using standard
material properties of the two rubber compounds, and the data oscillation testing, such as frequency sweeps and strain sweeps.
further used to troubleshoot the extrusion instabilities. Frequency sweeps were performed from 0.0628 rad/s to
314 rad/s or 0.01 Hz to 50 Hz using an appropriate strain amplitude
METHOD of 0.1 degrees, which was within the linear viscoelastic regime of
the testing specimen, while strain sweeps were performed from
The RPA elite (TA Instruments, New Castle, DE) is an advanced, 0.1° to 90° degrees of arc using an appropriate frequency of
rotorless rotational shear rheometer dedicated to the complete 0.2 Hz. Approximately 5.5 grams of material (± 5%) was utilized
viscoelastic characterization of polymers, rubber, and rubber and compressed in the testing cavity using a clamping pressure
compounds at all stages of manufacturing. The RPA elite provides of 4.5 bar and a temperature of 90 °C. Samples were allowed to
invaluable data related to the viscoelastic behavior of materials in relax for 10-15 minutes in the testing cavity before viscoelastic
the pre-cure state, through the curing process, and in situ cured measurements were performed in order to allow ample time for
state. Standard rubber testing on the MDR and Mooney viscometer any internal stresses experienced during the loading process to
does not provide additional viscoelastic characterization that is dissipate.
essential to optimizing processing parameters and troubleshooting
many processing issues. Samples are compressed and sealed in VISCOELASTIC DATA
a biconical die under high pressure, as is standard in the rubber
industry. This design results in a closed testing cavity which Frequency sweep tests on the RPA were performed to characterize
prevents melt fracture at the sample edge, and compressive both samples and correlate differences in viscoelasticity1-3 with
forces to prevent sample slippage during testing. their processing behavior. The complex viscosity is shown as a
function of angular frequency in Figure 2. Both samples show
shear thinning behavior, with the complex viscosity decreasing as
the angular frequency, and consequently shear rate, increase.

1 RT005
 delta results indicate that sample 1 is more elastic compared to
sample 2. The increase in elasticity can be attributed to higher
molecular weight or the presence of long chain branching.

Further analysis included a large amplitude oscillatory shear (LAOS)

Viscosity, η* [Pa.s]

test on both samples where the strain was varied from 0.3% to
over 1000% strain. The storage modulus (G’) and the tangent
104 delta for both samples are shown as a function of strain percent
in Figure 4. Sample 1 and 2 have similar viscoelastic behavior at
low strains but at higher strains in the non-linear regime, where the
modulus decreases, the viscoelastic responses begin to deviate.
The storage modulus (G’) for sample 1 is much higher than sample
Sample 1, Orange Peel Surface
2, indicating a stronger elastic structure within the material. The
Sample 2, Smooth Surface higher storage modulus (G’) equates to a stronger microstructure,
103 usually as a result of higher molecular weight, polymer rigidity, and
1 10 100 additive interactions.4-6 This is also corroborated by the tangent
Angular Frequency [rad/s] delta signal, with sample 1 showing a lower tangent delta overall
which corresponds to a larger degree of elasticity. The data at
Figure 2. Data from a frequency sweep shows the complex viscosity as a 90° of arc, or strain percent of 1255%, can be recast by plotting
function of angular frequency for sample 1 and 2. The frequency spectra
show similar behavior in processability between both samples, especially the transient shear stress against the shear rate in the Lissajous
in the high shear rate range which correlates with flow behavior in an plot (inset, Figure 4). The long chain branching index can be
extrusion process. calculated from the Lissajous plot. Both Lissajous plots show a
lack of secondary loops, indicating the presences of long chain
1.0 branching. This is confirmed with positive long chain branching
0.9 index values of 2.73 and 0.18 for sample 1 and 2, respectively. The
long chain branching index indicates that sample 1 has a higher
0.8 degree of long chain branching compared to sample 2, which was
0.7 confirmed through the tangent delta value in the frequency and
amplitude spectra. Recall, sample 1 demonstrated a lower tangent
0.6
delta which corresponds to a higher degree of elasticity.
Tan Delta

0.5
The presence of long chain branching contributes to the elastic
0.4
character of the material. Usually the presence of long chain
0.3 branching improves processability, however, process engineers
saw challenges and scorching problems with sample 1 despite the
0.2
Sample 1, Orange Peel Surface higher degree of long chain branching. The work dissipated during
0.1 Sample 2, Smooth Surface oscillation and shear can be calculated from the Lissajous plot by
0
measuring the area within the ellipse when plotting transient stress
1 10 100 against strain. This work is associated with viscous heating; this
heat must be removed from the material or it will result in a local
Angular Frequency [rad/s]
temperature rise. Analysis at 90° of arc, or strain percent of 1255%,
Figure 3. Data from a frequency sweep shows the tangent delta as a
shows the work dissipated to be 177 J and 169 J for sample 1 and
function of angular frequency for sample 1 and 2. Significant differences in 2, respectively. Compared to sample 2, sample 1 shows a larger
the viscoelastic response between the samples is observed at low angular amount of energy dissipated at high shear which is akin to the
frequency which correlates with polymer architecture. Sample 1 shows a shear rates experienced within the extruder. The larger amount of
lower tangent delta value, indicating a higher degree of elasticity.
energy dissipation could cause local temperature increase within
However, sample 1 is shown to have a slightly lower viscosity the sample as it is sheared within the extruder.
at elevated frequencies which corresponds to high shear rates.
Despite the MDR cure curve showing sample 1 would cure
Materials experience high shear rates during processing steps
more slowly, local temperature elevations within the sample
such as mixing, milling, and extrusion. Conversely, sample 1 also
as a result of energy dissipation from the induced shear in the
shows a higher viscosity at the low angular frequencies. Material
extruder could result in scorching of sample 1 during the extrusion
properties at low angular frequencies representative of material
process. The RPA was able to quantify differences between the
behavior at low shear rates, such as exiting an extruder, sitting on a
two rubber compounds and help process engineers optimize their
shelf, or flowing in a mold. A higher viscosity typically corresponds
processing conditions by decreasing the extruder speed and then
to a stronger resistance to flow, usually due to a higher degree
re-evaluating the compound formulation.
of elasticity. The difference in elasticity between the samples is
highlighted in Figure 3, by replotting the results of the frequency
sweep as tangent delta vs. frequency. The tangent delta between
both samples is significantly different, with sample 1 showing a value
of 0.725 at 1 rad/s compared to 0.89 for sample 2. These tangent

2 RT005
 Sample 1, Orange Peel Surface 6 REFERENCES
Sample 2, Smooth Surface

Sample 1 Sample 2 5 1. Barnes, H. A. “An Introduction to Rheology”, Elsevier, 1989.

Shear Modulus, G’ [kPa]

LCB Index 2.73 0.18

Wdiss 177 J 169 J 2. Macosko, C. W. “Rheology: Principles, Measurements and
4
Strain 90.018 [Grad] Applications” VCH Publishers, New York, 1994.

Tan Delta
150000

10 100000

50000
3 3. Tschoegl, N. W. “The Phenomenological Theory of Linear
Stress σ [Pa]

0
Viscoelastic Behavior”, Chapter 9, pages 443-483. Springer-
Verlag, 1989.
-50000

-100000
2
-20 -10 0 10 20

4. Ewoldt, R.H. & McKinley, G.H. Rheol Acta (2010) 49: 213.
Stress, γ [1/s]

1
https://doi.org/10.1007/s00397-009-0408-2
1 0
0.1 1 10 100 1000 5. Stadler FJ, Leygue A, Burhin H, Bailly C (2008) The potential
Strain [%] of large amplitude oscillatory shear to gain an insight into
the long-chain branching structure of polymers. In: The 235th
Figure 4. Data from a strain sweep is plotted as the elastic modulus (G’) and ACS national meeting, polymer preprints ACS, vol 49. New
tangent delta as a function of strain. Significant differences between the Orleans, LA, USA, pp 121–122
rubber compound samples are observed at high strains, in the non-linear
regime. The Lissajous plot shown for both samples at 90° of arc (inset) 6. Burhin HG, Bailly C, Keunings R, Rossion N, Leygue A,
indicates differences in long chain branching and energy dissipated during
the applied oscillation. Pawlowski H (2008) A study of polymer architecture with FT-
rheology and large amplitude oscillatory shear (LAOS). In: XV
CONCLUSION international congress on Rheology: The Society of Rheology
80th annual meeting, Monterey (California), The Society of
The RPA can be used to measure the viscoelastic properties Rheology, abstract booklet, EM18
of rubber compounds and raw polymers as a function of
shear rate and deflection, which are directly linked to flow and For more information or to request a product quote, please visit
processing behavior. Data obtained from frequency sweeps and www.tainstruments.com/ to locate your local sales office
amplitude sweeps can be used to identify differences in polymer information.
architecture, such as the presence of long chain branching and
variations in molecular weight and molecular weight distribution.
Testing on the RPA was able to distinguish differences between
two batches where the standard Mooney and MDR tests failed.
Testing data from an RPA can be used to identify variations in
polymer architecture, providing insight to the flow behavior so
production parameters can be optimized in advance and waste
can be reduced.

© TA Instruments 3 RT005