High strain rate testing of elastomers and

thermoplastic elastomers
by Sean Teller, Veryst Engineering though quasi-static test methods are well-understood, impact
rate testing is a growing field with many test methods. Testing
Elastomers and TPEs are used in many different applications elastomers and TPEs at these rates is difficult; the low sample
and industries, including automotive, aerospace, consumer forces, low density and low wave speed (or acoustic imped-
products, electronics, biomedical devices and heavy industry. ance) make testing more difficult. Careful experimental test
These materials are increasingly being used in high strain rate machine design and selection are necessary to obtain accurate,
and impact applications. Due to their inherent molecular struc- reliable results. This article aims to introduce and explain the
ture, elastomers and TPEs exhibit rate-dependent response: As different test methods available to test elastomers and TPEs, as
the material is deformed faster, the material response changes, well as explain advantages and disadvantages. Brief discussion
becoming stiffer. This effect needs to be taken into account by is also given on how the data can be used to select and calibrate
an engineer when designing a part that experiences high strain a material model for use in a finite element (FE) analysis.
rates. Without correctly understanding this rate-dependent be-
havior, a part can be over- or under-designed, performing Force and strain measurement
poorly during use or increasing production costs due to extrane- Accurately measuring force and sample strain (or displace-
ous material. Figure 1 shows the rate-dependent response of ment) is critical to a high quality test method and test plan.
Santoprene, a TPV material made from an EPDM and PP Modern force and strain measurement systems have made this
blend. As the strain rate is increased to 250 strain/s, the mate- task easier.
rial is ~60% stiffer at high elongation. In addition to the chang- Digital image correlation (DIC) is an increasingly used, ad-
ing stress-strain behavior of the material, many materials have vanced method to measure sample strain and displacement. The
different failure behaviors at high rates. Most materials become DIC method uses samples that have a random speckle pattern
brittle at high strain rates, with some materials undergoing a applied with either paint or ink. Figure 2 shows an exemplar
drastic change. tensile specimen with a speckle pattern. A computer-controlled
Understanding how a material behaves is critical to choos- camera records sample images throughout the test, saving them
ing the best material and optimizing the design of a component for post-processing. The sample image is divided into smaller
to the loading conditions in use. Thus, testing a material at areas, similar to an FE mesh; displacements and strains for each
multiple strain rates that cover the applied rates is critical. Al- area are calculated using
image analysis methods. The
individual areas are com- Figure 2 - DIC
Figure 1 - tensile test data for Santoprene; bined, and full field sample speckle pattern on
at 150% engineering strain, the material strains and displacements are an ASTM D638 type
is 60% stiffer calculated. The strain history IV tensile specimen
3 for all the samples can be
measured and analyzed.
DIC was originally de-
2.5 veloped as a single camera
system, so the systems could
Engineering stress (MPa)

2 not measure out-of-plane

displacements. With a 2D
system, the test engineer
1.5
must use care during test
specimen setup; any motion
1 perpendicular to the camera
introduces significant error
0.5 in the calculation. Stereo
camera systems and 3D DIC
are now common. With these
0 systems, two cameras record
0.5 0 0.25
0.75 1 1.25 1.5
Engineering strain the test from two different
--- Monotonic uniaxial tension, --- Monotonic uniaxial tension, angles, and full 3D displace-
0.001/s (experimental) 0.1/s (experimental) ment fields are measured,
--- Monotonic uniaxial tension, --- Monotonic uniaxial tension, including out-of-plane dis-
0.01/s (experimental) 250/s (experimental) placements.

40 RUBBERWORLD.COM • JANUARY 2019

1RW - 40,41,42,43.indd 40 1/21/2019 9:50:29 AM

 Universal test machines
Figure 3 - DIC axial true strain contours of a tensile test on Universal test machines or univer-
Santoprene at low strain rate sal test systems (UTS) are used for
True axial low strain rate testing of polymers,
strain typically up to 10 strain/s. UTS
0.75 0.70 0.66 0.61 0.56 0.52 0.47 0.42 0.38 0.33 0.28 0.23 0.19 0.14 0.09 0.05 0.00 come in two different configura-
tions: servohydraulic or electro-
mechanical machines. Typically,
the two systems are interchange-
able, although servohydraulic ma-
chines are better suited to higher
strain rates and fatigue applica-
tions. Advantages of UTS are that
they are well-characterized and
validated, and include complex
fixturing and precise control sys-
tems. For example, a test engineer
can program complex strain histo-
ries to test the response of the ma-
terial with the control system, in-
cluding loading, unloading, creep
or relaxation, and cyclic loading.
Unfortunately, only moderate
strain rates are achievable in these
systems. A typical electromechani-
cal UTS is presented in figure 4.
The figure shows compression
platens and a temperature chamber
in the background.

Split Hopkinson pressure bar

The split Hopkinson pressure bar
(SHPB), or Kolsky bar, is a dy-
‹‰—”‡͵Ǥ ƒš‹ƒŽ–”—‡•–”ƒ‹ ‘–‘—”•‘ˆƒ–‡•‹Ž‡–‡•–‘ƒ–‘’”‡‡ƒ–Ž‘™•–”ƒ‹Ǧ”ƒ–‡Ǥ namic, stress wave based experi-
‹‰—”‡͵Ǥ
‹‰—”‡͵Ǥ ƒš‹ƒŽ–”—‡•–”ƒ‹
ƒš‹ƒŽ–”—‡•–”ƒ‹ ‘–‘—”•‘ˆƒ–‡•‹Ž‡–‡•–‘ƒ–‘’”‡‡ƒ–Ž‘™•–”ƒ‹Ǧ”ƒ–‡Ǥ
‘–‘—”•‘ˆƒ–‡•‹Ž‡–‡•–‘ƒ–‘’”‡‡ƒ–Ž‘™•–”ƒ‹Ǧ”ƒ–‡Ǥ
mental method to measure the
DIC strain measurement has many advantages versus tradi- high strain rate response of materials up to ~10,000 strain/s.
tional extensometer or laser extensometer systems. DIC sys- SHPBs can be configured to test in compression, tension, tor-
tems can measure sample strains up to and in excess of 1,000% sion or shear, and have a long history of use in research and
strain, as well as accuracy down to 10 microstrain. Addition- development applications.
ally, any strain localization or failure effects can be measured A schematic of an SHPB is shown in figure 5. The system
throughout the test; if a sample experiences necking or tearing consists of a pneumatic cylinder, striker, incident bar and trans-
during testing, these values can be measured and analyzed. mitted bar. The pneumatic cylinder is charged to a set pressure,
Further, all strain measurements are made without sample con- and accelerates the striker. The striker transmits a stress wave
tact, so no effects are introduced due to the measurement sys- to the incident bar, and the velocity of the striker is measured.
tem. Last, with one strain measurement system, all strain The stress wave travels through the bar, into the sample, and
components are measured. Figure 3 shows DIC strain contours into the transmitted bar. The strain in the incident and transmit-
from a tensile test on a TPE material. Multiple commercial DIC ted bars is measured with strain gauges, and the bars are made
software packages exist, as well as open source packages. of a well-known material. Sample stress and strain are calcu-
Load cells for force measurement can be broken down into lated from the strain gauge measurements, and the applied load
two main categories: static load cells and dynamic load cells. and strain history for the sample is calculated.
Static load cells are used for low strain rate testing, and are usu- Careful design of an SHPB is necessary when used for elas-
ally strain gauge based systems. Typical configurations are the tomers and TPEs. If the acoustic impedance of the incident and
s-style or pancake load cells. These can be either single axis transmitted bars is too high, then very little energy (and defor-
measurements or multi-axis measurements. Dynamic load cells mation) is transferred to the sample. SHPB systems are de-
are designed to have a fast response time, and cannot measure signed for the materials they are to test: SHPBs for plastics use
static loads. These load cells are typically piezoelectric or piezo- aluminum bars; SHPBs for elastomers and TPEs typically use
resistive, and have response frequencies above 75,000 Hz. PMMA bars. Although the PMMA bars create good material

FOLLOW US ON TWITTER @rubberworld 41

1RW - 40,41,42,43.indd 41 1/21/2019 9:50:31 AM

 sample force is mea-
Figure 4 - exemplar electromechanical test sured with a dynamic Figure 6 - Veryst
frame for low strain rate testing load cell. Test strain Engineering's custom
rate is controlled by built drop tower
changing the height of
the falling weight, and
materials can be tested
in tension up to
1,000% strain.
A well-designed
drop tower system will
provide highly accu-
rate and mostly noise-
less material data for
all classes of polymers,
particularly elastomers
and TPEs. In addition,
samples can be tested
to extremely large
strains in both tension
and compression. The
high-speed camera
and DIC allow full
field strain measure-
ments and a record of
the test. The drop
tower is limited in the
maximum strain rates
achievable, although
response, the PMMA bars are viscoelastic materials. Because rates that are applica-
the data analysis for the test uses the mechanical response of the ble to most impact
bars, this introduces additional complexities and potential er- conditions are possi-
rors. ble.
The SHPB is a validated and well-characterized system
with high test strain rates, although it does have limitations. Inverse methods
Soft materials, like elastomers and TPEs, are more difficult to Inverse test methods
test and analyze with the SHPB, introducing additional uncer- are attractive when
tainty in the test data. Additionally, total test strain is limited reaching high strain
with the SHPB. rates with a homoge- 
nous strain field is‹‰—”‡͸Ǥ‡”›•–‰‹‡‡”‹‰̵•
dif- —•–‘Ǧ„—‹Ž–†”‘’–‘™‡”Ǥ
Drop towers ficult or impossible, which is often true for elastomers and TPEs.
Drop towers are a versatile tool to test elastomers and TPEs up Due to their relative softness and low acoustic impedance, at high
to strain rates of ~1,000 strain/s. A drop tower uses a falling strain rates the applied strain field may not be uniform. Inverse
weight to load the sample, and different fixtures are used for ten- test methods can overcome this difficulty. In an inverse test
sion, compression or shear testing. Figure 6 shows Veryst Engi- method, the force and displacement of the loading are measured,
neering’s custom built drop tower. In this system, sample strain and an FE model is used to simulate the experiment and calculate
is measured with a high-speed camera (>100,000 fps) and DIC; the material response. A candidate material model is used to
perform the simulation, and the simulated force and displace-
ment are compared to the experimental data. Nonlinear search
Figure 5 - split Hopkinson pressure bar algorithms are then used to minimize the error between the can-
(SHPB) schematic didate material model and the experimental data.
One example of a common inverse test method is the ball
Striker Strain gauges impact or ball drop experiment. A cylinder of the material is
bar placed on a dynamic load cell, and a ball bearing (typically
V0 Incident bar Transmission bar steel) is dropped from a known height. The force on the sample
is recorded, and the displacement is calculated from Newton’s
Sample
laws. Figure 7 shows the results of an FE simulation on a rub-

42 RUBBERWORLD.COM • JANUARY 2019

1RW - 40,41,42,43.indd 42 1/21/2019 9:50:41 AM

 able for elastomers and TPEs. Although the experiments can
Figure 7 - FE simulation of a ball impact produce clean data, additional data processing is needed to
test on a rubber material; the contours capture the material behavior from these tests. In addition to the
show the effective strain rate in the high strain rates available, inverse test methods use multi-axial
sample, reaching 3,500 strain/s loading. This can be important to validate material models in
more complex loading modes, building confidence in the mate-
ER, maximum in-plane principal rial model.
(Average: 100%)
3,500
3,250 Material model selection and calibration
3,000
2,750 Finite element analysis is a powerful tool for the design engi-
2,500
2,250 neer to understand the stress and deformation the designed
2,000
1,750 component experiences during use. The analysis can be used to
1,500
1,250 validate if the component will perform or needs to be rede-
1,000
750
500
signed. An FE analysis has three inputs: component geometry,
250
0
applied loading and constraints, and material behavior. Geo-
-250
-500 metry and applied loading are well-defined based on load
cases, but material response is difficult to accurately capture
with complex, advanced materials such as elastomers and
TPEs. Advanced tools exist to calibrate advanced material
models that capture strain rate dependent material response,
including tools built into FE software and MCalibration from
Veryst Engineering. MCalibration uses non-linear optimization
procedures to match simulated material response to experimen-
tal data. The material parameters are then exported to the FE
software for simulation, enabling advanced simulations to ana-
lyze and design products under real world loading conditions.
ber material. In this case, the sample experiences strains in the
range of 5,000 strain/s, well above those achievable in a drop Conclusion
tower. Elastomers and TPEs are complicated materials that exhibit
Another example of an inverse test method is the Taylor strain rate dependent behavior due to their molecular structure.
impact experiment. In this test, a cylindrical specimen is shot This rate dependent behavior can drastically change the stress-
from a gas gun at a rigid platen. The original experiment was strain response of the material, and can affect designed compo-
used to predict yield stress based on the deformed shape of the nent behavior. Accurate methods for soft materials like elasto-
specimen, but modern load cells and high-speed cameras allow mers and TPEs are readily available in multiple loading modes.
real time measurement of the force and displacement. These Each test method offers advantages and disadvantages that
data are then used with an FE model to simulate the material need to be considered for each material and test program. A
response at high rates. Impact rates up to 10,000 strain/s are well-designed test plan can provide highly accurate data that
possible with this method. can enable a material or design engineer to create a better end
Generally, inverse methods enable extremely high strain product for the user. Last, clean, accurate data enable an engi-
rate testing of all classes of materials, and are particularly suit- neer to calibrate a material model for use in an FE simulation.

50 Years of Polymer Testing

by Roger P. Brown
Brown provides a unique personal account of the developments in the
technology of physical testing of polymers, and of the changes in the working
environment in which testing has been conducted during the last 50 years.

Paperback, 192 pages

ISBN 1847354505 (ISBN13: 9781847354501)
Edition Language English
$75.00 plus s & h www.rubberworld.com/bookstore

FOLLOW US ON TWITTER @rubberworld 43

1RW - 40,41,42,43.indd 43 1/21/2019 9:50:44 AM