Embedded Sensors in Rubber and Other Polymer Components
K.J. Harmeyer, M.A. Holland, J.H. Lumkes, G.W. Krutz
Purdue University Agricultural and Biological Engineering
225 S University Street
West Lafayette, IN 47907
[email protected]
[email protected]
[email protected]
[email protected]
ABSTRACT
Many polymer components are susceptible to catastrophic failure or are critical to the performance of the products that they
make up. Because of this, it is beneficial to be capable of monitoring structural failures or a performance drop in these
components. It is difficult to create sensors for polymer components because they often have complex shapes or are
assembled in isolated locations. To solve this problem, micro-scale electronic sensors, embedded within polymer
components, have been developed at Purdue University. Failure modes that are detected include: wear, loss of adhesion,
excessive compression or expansion, overheating, fractures or breaks, changes in material properties, leakage of fluids and
others. The sensors are accomplished by using conductive polymer materials as the primary sensing element. Testing results
reveal that embedded sensors in polymer components can successfully indicate significant signal changes (1400 - 30000%)
more than 100 loading cycles prior to catastrophic failure. Multiple sensing methods and applications have been tested and
more are being researched. This finding may open the door for future polymer sensors that can improve safety and provide
useful measurements for polymer components.
Introduction
All sensors require a sensing element that is highly responsive, repeatable and specific in regards to the phenomena being
measured [Pallàs-Areny p54]. It is important that the material make-up of the primary sensing element be one that is
conducive to these characteristics. Materials used to construct sensing elements are typically limited to those that have the
desired electrical properties. Sensors that are embedded or integrated into functioning components may not have optimal
structural, chemical or other physical properties because the material was selected based on the required electrical properties.
This paper will examine the potential for using rubber and other polymers for primary sensing elements that require electrical
conductivity, making feasible the use of embedded sensors in polymer components.
Conductive polymers have become widely used for their antistatic properties [Murphy p141]. Less common is the use of
conductive polymers for transmitting electrical signals. By utilizing this possibility, electronic conducting materials can achieve
properties that before were uncommon. Polymers have properties that are sometimes desirable over typical conductors, like
metal, that are used to create sensors. Products like vibration insulators, plastic containers, and rubber belts would not
function optimally if constructed of traditional conductive materials. At the same time, the structural integrity, loading and other
measurable factors of polymer products provide useful information. Having sensors embedded in polymers allows a product to
have the desirable properties of a polymer while having sensing capabilities which have not previously been realized.
Potential Applications of Embedded Polymer Sensors
The list of measurable variables and failure modes of rubber and other polymers is extensive [3-5], but a few are discussed
here.
Rubber and some other polymers are often used because of their good fatigue life and elasticity. However, fatigue failures still
occur in rubber and polymer components when cracks propagate in the material. This can be caused by many things including
cuts from an interfering part, tears from excessive deflection, or imperfections in the material. Typically these failures are not
visible until 90% or more of the life has elapsed [Krutz et al. p499]. Because of the difficulty or impossibility of visual detection
of impending fatigue failures in these parts, it is incredibly beneficial to have a sensor with the ability to recognize crack
propagation.
Rubber is also often used as a spring or as an isolator. Using resistive sensing principles, load can be correlated to strain
which affects the resistance of a conductive material. By measuring the resistance of a conductive rubber part, one would be
able to determine the load on that product. If the sensor has great enough sensitivity it could also be used to measure
vibrations or frequency of cyclical loads. This information may be useful for feedback in a control system or for safety
monitoring. An interesting application of load sensing is the use of rubber insulators beneath flooring, such as is used on
basketball floors (Figure 1). Measuring the energy absorbed by an athletic floor is difficult and limited to small sections of
flooring. If the rubber insulation itself were used to collect loading data on a test floor, energy absorption could be determined
during realistic use of the floor (i.e. an actual basketball game).
Figure 1
Basketball Flooring - Rubber Isolator
Some plastic polymers are used for their wear characteristics to prolong the life of a part. However, when a threshold of wear
is reached the part must be replaced before damage or failure occurs. By measuring electrical properties of a plastic wear
plate or coating, it would be possible to determine when that threshold is met without visual inspection. It is also foreseeable
to use this same idea in other polymer applications in which wear or abrasion creates the potential for failure.
Figure 2
Slide Frame Wear Plates
www.minerent.com
The properties of rubber and other polymers change over time and with changes in environment. An example of this is aging
that occurs in rubber, or natural curing that makes the rubber harder and more brittle over time. Also, biodegradable plastics
are becoming common and only function adequately for a given period. It would be beneficial for sensors to determine when
properties have changed significantly enough to effect functionality. Changes in electrical properties would be observable in
these types of products and could alert the user when a threshold is achieved; indicating the life of the product has elapsed.
Rubber and other polymers are often used for their insulating properties. Most electrical signal wires and power wires are
insulated by a rubber covering. Safety and electrical noise is a major issue with insulators, especially in locations where
people or flammable substances come in contact with the wires. An additional coating of conductive rubber on the outside of
rubber insulation could be used to monitor the resistance of the insulation while maintaining the flexibility of the wire. An issue
with this application is the loss of electric power due to the formation of a condenser (two conductive materials separated by a
dielectric). Equation 1 is used to calculate the power factor in this situation. Power factor is a quantity that represents the ratio
of the energy lost in an AC circuit.
⎛ G ⎞
⎟⎟
P = 15900 × ⎜⎜
⎝ f ×C ⎠
(Equation 1)
where
P = percent power factor
G = conductance (mhos = 1/ohms)
f = frequency (kilocycles)
C = capacitance (nanofarads)
Rubber and other polymers are also ideal for applications where water permeability and chemical resistance are important.
There are many possibilities for sensing in this area also, such as sensing of a chemical that has penetrated a polymer’s
surface by monitoring a change in resistance or dielectric constant. Conductive polymers could be used as an outer layer if
the material that is in contact with the chemical must maintain a specific chemistry.
Achieving Conductivity in Polymers
Electrical sensors require some conducting element in order to transfer a signal. For example, a resistive sensor must have at
least measurable conductivity in order for resistance to be measured. Likewise a capacitive sensor must have two conducting
plates separated by a dielectric. In order to create an embedded sensor out of a polymer part, some type of conductor must be
included. Adding metal parts or wires would be an intuitive solution; however, polymer blends can achieve adequate
conductivity without adding metal components. Conductive filler materials, such as carbon black, can exponentially increase
the conductivity of rubber and other polymers. Conductive salts can also be mixed into rubber compounds to provide the
interface necessary for an electrically based sensor.
Table 1: Volume Resistivity of Rubber Fillers [Ohm p117]
Volume Resistivity, ohm-cm
Filler Material
Natural
Rubber
Styrene-Butadiene
(volume)
(111.1 volumes)
Rubber (115 .25volumes)
No filler
4.4 x 1016
3.7 x 1015
Zinc Oxide (50 volumes)
8.2 x 1013
3.9 x 1012
DIXIE CLAY (50)
3.6 x 1015
1.8 x 1015
Calcined Clay (50)
4.3 x 1015
3.0 x 1015
Whiting (50)
8.4 x 1015
4.5 x 1015
NYTAL 300 (50)
4.4 x 1015
4.4 x 1015
THERMAX (25)
1.5 x 1016
4.5 x 1015
THERMAX (50)
2.5 x 1010
2.8 x 1015
Carbon Black, N-765 (25)
9.1 x 1010
1.8 x 1015
Cumar MH 2½ (25)
-
1.1 x 1016
Mineral Rubber (25)
4.7 x 1015
1.3 x 1015
Mineral Rubber (50)
2.2 x 1015
8.6 x 1014
The simplest way to modify the electrical properties of a polymer material is to us a conductive filler or extender.
Characteristics of filler materials must meet certain requirements and be cost effective. A list of common rubber fillers and their
respective resistivity is shown in Table 1. Many of these fillers are known for the insulating characteristic that they add to
rubber. For example, using Cumar as filler in SBR (Styrene-Butadien Rubber) has a volume resistivity of 1.1 x 1016 Ohm-cm.
Carbon black on the other hand, can lower the resistivity of Natural Rubber by 6 orders of magnitude when added at a
moderate level. If rubber is overloaded with Carbon Black the resistivity becomes low enough that the rubber can become a
legitimate conductor.
It is believed that charge moves through carbon polymer composites by way of tunneling of charge carriers. The reasons for
this conclusion are discussed in [Sichel p 53]. This model explains the current flow as electrons traveling through continuous
carbon pathways and jumping any gaps to get to the next pathway. Therefore, conductivity is exponentially dependent on
carbon black content (see Figure 3). A threshold of approximately 25% concentration by volume of conductive filler (such as
carbon black) typically must be reached before significant conductive properties are recognized. Conductive polymers can
achieve conductivities on the order of 10(Ω-cm)-1.
Figure 3
Resistivity of a PVC Polymer as a Function of Carbon Black Loading [Sichel p149]
Testing and Results
Preliminary tests were conducted to validate the sensitivity of resistive polymers and to determine the ability of conductive
polymers to make LCR elements for sensing. The polymer that was used in the experiment was a rubber compound made by
GE silicones called SE877 TUFEL®. This is a black semi-conductive rubber compound that is often used for bleeding static
electricity. The sample was formed into a 50mm by 1mm strip and cured according to GE Silicone’s specifications. A HewlettPackard multi-meter was used to make the resistance measurements.
A 50mm long strip of the rubber (1mm x 10mm) was used to perform the calibration curve. The strip was clamped on both
ends and put in tension. The force on the strip was measured with a spring scale accurate to 50g (0.1 lb). Figure 4 is a plot of
the results. The rubber strip broke after 200psi of tensile stress was applied. A regression line was fit to the data to determine
the slope of the curve, which equals the sensitivity. The results show that the sensitivity of the conductive rubber was 1.21
Ohms/psi.
Calibration Curve for Rubber Tension
Resistance (Ohms)
600
550
500
450
400
y = 1.2101x + 347.08
2
R = 0.9874
350
300
0
50
100
150
200
Tensile Stress (psi)
Figure 4
Preliminary Sensitivity Testing Results
Proprietary Testing was also done on a specific application of embedded polymer sensors. The sensor used in the project
utilized LCR measurements to monitor changes in the structure that would be critical to the performance. The goal of the
sensor was to alert the user prior to failure of the component. The failure mode of concern was cracking or breaking of the
rubber.
Initial tests were done on six samples of the embedded sensor to find the correlation between pressure and the LCR
measurement. Each sample contained slight structural differences that led to minor discrepancies in the nominal
measurement value. However, the slope of the pressure curve is approximately identical for each of the six samples. The data
from this test can be seen in Figure 5.
Figure 5
Pressure Loading Curve for Six Samples
To test the functionality of the sensor a fatigue test was created for the embedded polymer sensor. Figure 6 displays the
results when a cyclic load was applied to the sample. As expected the measurement value cycles between 0 and a magnitude
comparable to that shown in the pressure loading curve. However, 100 cycles before catastrophic failure the measurement
spiked to 15.9 (more than 1400% of the initial reading). The sensor continued to produce values that were much larger than
expected until failure. At approximately 25 cycles before failure the measurement became constant at a very large value (-768)
until catastrophic failure and after.
Figure 6
Testing Results for a Prototype Embedded Polymer Sensor
The results for each of the six samples were similar, with the exception of sample #4. Sample #4 contained an incidental
assembly error. Because of this error the LCR measurement was higher than expected before any load was applied. The test
was then aborted so that the sample could be inspected. Table 2 contains a summary of the results for each of the six
samples. Notice that each of the samples detected high LCR measurements more than 100 cycles prior to failure.
Table 2: Summary of Test Sample Results
Sample #
Initial Measurement
(average LCR Value)
Measurement of
significance (LCR value)
Cycles before
Failure
1
1.20
22.5
526
2
3
4
5
6
1.10
1.11
-16.0
1.18
1.16
15.9
4.45
-16.0
13.74
90.64
101
195
Failed at cycle #1
313
561
Conclusions
It was hypothesized that conductive rubber could be used to create a sensing element to measure useful parameters. Test
results show that embedded polymer sensors can adequately and consistently predict changes in structure that may lead to
catastrophic failure. The sensor was proven to be effective at sensing impending failures in the range of 100-600 cycles
before failure. Changes in the electrical properties are on the magnitude of 1400 to 30000% for particular applications. This
high sensitivity limits the number of erroneous signals given by the sensor. This sensor could be used to indicate a problem to
an operator or may be input as a control parameter. It is expected that embedded sensors in rubber and other polymers may
be used for additional applications in the future. Two US patents have been filed and proprietary research is being conducted
for particular embedded sensor products.
References
1.
2.
3.
4.
5.
6.
7.
Pallàs-Areny, R., Webster, J.G., Sensors and Signal Conditioning, Second Edition, John Wiley and Sons, New
York(2001).
Murphy, J. Additives for Plastics Handbook, Second Edition, Elsevier Andvanced Technology, Oxford(1996).
Gerdeen, J.C., Lord, H.W., and Rorrer, R.A.L., Engineering Design with Polymers and Composites, Taylor and
Francis, New York(2006).
Flinn, R.A., Trojan, P.K., Engineering Materials and Their Applications, Fourth Edition, John Wiley and Sons, Inc.
New York(1995).
Krutz, G.W., Schueller, J.K., and Claar, P.W. II., Machine Design for Mobile and Industrial Applications, Second
Edition, SAE International, Warrendale(1999).
Ohm, R.F., The Vanderbilt Rubber Handbook, Thirteenth Edition, R.T. Vanderbilt Company, Inc., Norwalk(1990).
Sichel, E.K., Carbon Black-Polymer Composites, The Physics of Electrically Conducting Composites, Marcell Dekker,
Inc., New York(1982).