Electricity on Rubber Surfaces: A New Energy Conversion Effect.

This work describes the conversion of mechanical energy to electricity, by periodically stretching rubber tubing and allowing it to relax. The rubber surface shows periodic and reversible electrostatic potential variations, in phase with the tubing length. The potential change depends on the elastomer used: silicone loses charge when stretched and becomes strongly negative when relaxed, whereas the stretched natural rubber is positive, becoming negative when relaxed. Every other elastomeric material that was tested also showed periodic potential but followed different patterns. When the motion stops, the potential on the resting samples decreases quickly to zero. The potential oscillation amplitude decreases when the relative humidity decreases from 65 to 27%, but it is negligible when the rubber tubing is previously swollen with water or paraffin oil. Elastomer charging patterns do not present the well-known characteristics of piezo-, flexo-, or triboelectricity, and they are discussed considering rubber rheology, wear, and surface properties, including the possibility of surface piezoelectricity. The following mechanism is suggested: rubber stretching provokes chemical and morphology changes in its surface, followed by a change in the surface concentration of H+ and OH- ions adsorbed along with water. The possibility of the occurrence of similar variations in other systems (both inert and biological) is discussed, together with its implications for energy scavenging from the environment.

Introduction

This paper describes a new finding on the electrostatic behavior of elastomers: shortly, rubber tubing stretching followed by relaxation provokes the appearance of transient excess charge that is more pronounced under higher relative humidity (RH). This adds to other unexpected recent findings on electrostatic phenomena[1−3] that are leading to a revision of widespread ideas[4] on electrostatic charging mechanisms, affecting many scientific areas and creating new opportunities for energy production.[5,6]

The limitations of current knowledge on electrostatic phenomena have been revealed on different occasions,[7,8] and the search for better understanding led piezo- and triboelectricity[9,10] to a distinguished position in current research on environmental energy scavenging.[11,12] On the other hand, the lack of well-established knowledge on this matter is evidenced by the persistent occurrence of serious industrial accidents owing to explosions and fire initiated by electrostatic discharge on solids that are otherwise rather stable. Widespread and seemingly safe solids, such as polyethylene (PE) and wheat flour, are thus transformed into powerful explosives.[13,14]

Tribo- and piezoelectricity are common effects of mechanical action on materials that have been investigated considering their potential use in energy scavenging.[15,16] The two effects account for the electrostatic potentials detected in most anthropic and natural environments, often reaching many-thousand volts. They are thus the basis of many types of nanotribogenerators[17] that are now showing impressive performance.

Piezoelectricity is observed in noncentrosymmetric crystals, but the closely related flexoelectricity(18) can be observed in any kind of material under a strain gradient. Flexoelectricity in polymers has been examined in a few cases only,[19] and there is a significant disagreement among the reported flexoelectricity tensors, that led the authors of a recent review[18] to explicitly exclude polymers from the tables, showing flexoelectric coefficients for various materials. Giant flexoelectric coefficients were measured in poly(vinylidene difluoride), leading the authors to state that “In conclusion... the physical origin behind the flexoelectricity in polymers might be more complicated than the one proposed for solid crystalline dielectrics.”[20]

Another emerging topic is surface piezoelectricity(21) that may be observed when the solid surface differs from the bulk solid, presenting a noncentrosymmetric crystalline material. Its importance increases at the nanoscale, when even nonpiezoelectric materials may act as piezoelectric.

The electrical effects triggered by the mechanical action on rubbers are related to many other research topics: polymer rheology, friction and adhesion,[22,23] wear,[24] and lubrication[25] that are often associated with the mechanochemical reactions in the polymer bulk or surface and the formation of triboplasma.[26,27] This can trigger other physical effects, such as triboluminescence[28,29] and acoustic emission.[30] Because these are studied in different contexts and experimental conditions, the correlation of the experimental results is hardly feasible.

The present authors made many previous incidental observations of electrostatic charge on rubbery materials subjected to mechanical efforts such as tension, compression, flection, and torsion, which do not involve contact or friction with external surfaces. Very high potential with half-life in the few-minutes range is then measured near elastomers, different from thermoplastics that can store charge for long periods, hours to many days.[31]

An intriguing result was recently reported on the electrostatic charging of relaxed and strained latex rubber sheets contacted with different materials[32] [poly(tetrafluoroethylene) (PTFE), polyurethane, and stainless steel], finding great differences in the amount of contact charging. These were tentatively assigned to rubber surface modification or to a change in the mass transfer pattern because of rubber stretching. Earlier, Dogadkin et al. had shown the effect of electric charges formed during repeated deformations on the fatigue resistance of vulcanized rubber.[33] Thus, rubber electrostatic charging has specific features that are not yet understood. Given the ubiquity of rubber materials in anthropic environments, they probably make an important contribution to our electrified environment but this has been detected indirectly, for example, in triboluminescence experiments.[27−29] It is remarkable that the excellent book on Tribochemistry authored by Heinicke[26] does not carry the keyword “rubber” in its extensive index (13 pages) neither in its table of contents (5 pages). Nevertheless, this book has a section on the Tribochemisty of Polymers.

In this work, we show that cyclic rubber stretching and relaxation under the atmosphere provokes the appearance of large positive and/or negative potential, depending on the material used and the RH. This is an unprecedented finding that may have profound implications in human safety and health.

Results

When rubber tubing is periodically stretched and allowed to relax, it shows a periodic change in the electrostatic potential, measured by a Kelvin probe. The frequency of potential oscillations is the same as the stretching frequency, and the typical plots of potential versus time are presented in Figure , for natural rubber and silicone. While natural rubber develops predominantly positive potential during periodic motion, silicone rubber is negatively charged. The electrostatic potential on natural rubber that is initially close to 0 V increases to positive values when it is stretched and decreases to negative values when it relaxes.

Figure 1

Electrostatic potential of (A) natural and (B) silicone rubber under stretching–relaxation cycles. The Kelvin electrode recorded the initial electrostatic potential, for 60 s. The stretching–relaxation cycles followed for 180 s and thus ceased, leaving the material at the starting position. Zooms from (A,B) are shown in the inset graphs. (C,D) voltage vs distance for natural and silicone rubber, respectively. Experiments were performed at 60% RH (see also Figure ).

Figure 9

(A) Schematic representation and (B,C) real pictures of the apparatus used to record the electrostatic potential during periodic stretching of rubber tubing (photograph courtesy of “Thiago A. L. Burgo”. Copyright 2017). See also Supporting Information for other pictures and video.

Silicone tubing equilibrated in the laboratory environment until it reached ca. −5 V does not change when stretched, but it develops a significant negative charge when relaxed. Repeatedly stretching silicone rubber produces a negative potential exceeding −3300 V that is the lower limit for the voltmeter used.

Electrostatic potential changes with the length of the stretched rubber during typical actuation cycles. The V versus d curves in Figure C,D reveal that the charge on the natural rubber changes continuously during the whole cycle, whilst the potential of silicone surfaces increases more pronouncedly under shorter extension. These results are reproduced for lower amplitude and frequencies of actuation (Figure S1).

Repeated experiments with silicone and natural rubber show that the potential patterns are reproducible but with quantitative drifts. This is observed in Figure , showing the potential patterns acquired for the same pieces of tubing that were subjected to successive trains of short stretch–relax cycles. Each train lasted for ca. 10 s followed by 60 s resting period, during which natural rubber potential decreased slowly and reproducibly. Similar observations could not be done on silicone because the potential reached exceeds the measuring range of the Kelvin voltmeter.

Figure 2

Six consecutive trains of stretching–relaxation cycles for (A) natural and (B) silicone rubber. Insets are extracted from the plots in (A,B), presented with an expanded time scale (x-axis). Experiments were performed at 60% RH.

There is only a small baseline drift, showing that low persistent charging, if any, takes place concurrent with the fast reversible processes. When the tubing is allowed to rest under laboratory conditions, charge is dissipated and the potential reaches values close to zero, minutes after stopping the mechanical oscillations.

The expressive variations in potential are not correlated with apparent physical changes in the elastomers, during cycling. The temperature on the rubber surfaces increases by only 2–4 °C above room values. Experiments performed in a dark room with a sensitive camera set to long exposure time did not show any glow discharge or spark, characteristic of triboluminescence.

The electrostatic potential response during natural rubber cycling is asymmetric. Figure shows the result of alternating single stretching and relaxation cycles with an amplitude of 6 cm. The tubing, initially at rest in a relaxed state, is then subjected to a stretching half-cycle, at t = 27 s. The potential promptly responds to mechanical action increasing in value, and then decreasing exponentially while the rubber is kept in the stretched state. Once the potential reaches zero, the rubber is allowed to relax back to the initial condition, starting at t = 71 s. Again, the potential changes quickly but toward negative values, and then decay back to V = 0 V. Potential decay after rubber relaxation is much faster than the potential decay after stretching the rubber. The pattern just described was observed reproducibly, under different samples and humidity conditions, although it would occur at a much slower rate for aged materials (b). Note that full charge dissipation takes longer than the time required for stretching or relaxing the rubber tubing.

Figure 3

Potential as a function of time during single stretching and relaxation steps, obtained for (a) typical natural rubber and (b) natural rubber that has been aged, by washing with ethanol and heating at 60° during 1 h, five times. The elongation amplitude was 6 cm.

Some observations made during the experimental work suggested that the ambient humidity affects the experimental results. For this reason, the whole apparatus was enclosed within a wooden frame lined with a flexible poly(vinylchloride) film that was previously allowed to dissipate charge, reaching an electrostatic potential lower than 10 V. RH variations were obtained by bubbling nitrogen gas through deionized water and blending it with dry gas, to reach the desired values.

Figure shows a plot of the amplitude of potential variation during stretching as a function of the RH for natural rubber. This shows that the charging effect disappears, below 20% RH, when the variation observed is not distinguishable from noise.

Figure 4

Amplitude of electrostatic potential oscillations as a function of RH for natural rubber with an elongation amplitude of 6 cm.

Elastomer charging is thus pronounced above 20–30% RH, only. However, it is not possible to do measurements above 80% RH to avoid malfunctioning of the Kelvin electrode. Measurements were thus made on a piece of tubing that was swollen by immerging it within water for 48 h. This sample did not at all acquire charge (see Figure ).

Figure 5

Electrostatic potential recorded during stretching–relaxation cycles for natural rubber exposed to controlled RH and changing stretching frequency. (A) Experiment initiated under ambient 65% RH, followed by a decrease in humidity to 27%, and then back to 65% RH. The fourth trace was recorded for soaked natural rubber previously immersed in water, for 48 h. (B) Effect of stretching frequency: initially single manual cycles, followed by the standard 4 Hz stretching protocol and then slowly decreased down to 0.2 Hz, back to 4 Hz and finishing with manual cycles.

Figure also shows the effect of RH together with the effect of stretching frequency. During the initial 300 s, individual stretching cycles at a very low frequency (0.2 Hz) produced ±20 V potential jumps that quickly relaxed to zero. The positive potential decay rate is slower than the negative potential decay, as also observed in Figure . Voltage under 4 Hz stretching reached 35 V. After 750 s, the oscillating frequency was gradually decreased down to a minimum 0.2 Hz and the potential amplitudes also decreased to values similar to those obtained in the initial run. Finally, a sudden increase in the oscillating frequency up to 4 Hz brings back higher potential variations. The last two recordings at 4 and 0.2 Hz allow an assessment of the reproducibility of the electrostatic potential patterns observed for natural rubber.

There is also an effect of the stretching amplitude on the potential measured during periodic motion, and a set of data obtained for natural rubber is in Figure S2 The highest electrostatic potentials were measured under larger stretching amplitudes, and they are significantly reduced for smaller stretching amplitudes. For example, the experiments shown in Figures and 2 for natural rubber were performed at 10 cm amplitude, corresponding to 83% rubber elongation, producing voltages in the −50 to +150 V range.

Under 17% elongation, lower electrostatic potential is measured, in the −20 to +20 V range. Both higher stretching frequency and higher elongation have a common feature: both provoke faster change in the surface area of the rubber.

Experiments were also done with common rubbery materials with different shapes, and their results are given in Figure . Pieces of cloth and elastic bands were subjected to periodic stretching cycles in a setup similar to that employed in the main experiments. Experiments started with four cycles of manual, low-frequency actuation, followed by 3 min of reciprocate shaking at 4 Hz, and finally resting for 1 min. Different materials show different charging patterns but with one common feature: electrostatic charging was observed under periodic stretching, in every case.

Figure 6

Electrostatic charging of common elastic materials subjected to stretching–relaxing cycles for (A) elastic liner for clothes, (B) rubber band, (C) wristband, (D) elastic fabric strip, and (E) strip cut from a polyamide sock.

Because silicone develops pronounced negative potential and natural rubber becomes rather more positive, the two were combined to produce electricity during stretching and relaxation. One piece of each rubber tubing was mounted coaxially within a shielded copper cylinder that was in turn connected to one input of a rectifying bridge. The two rubbers were stretched and relaxed simultaneously, as shown in Figure A. Figure B shows the overall positive output of the bridge, and 7C shows that a small capacitor connected to the bridge output is quickly charged.

Figure 7

(A) Device used to collect electricity generated by periodic stretching of rubber tubing. (B) Open-circuit voltage between the two inner copper cylinders. (C) Potential difference between the leads of a 10 μF capacitor connected to the copper cylinders through a rectifying bridge. Experiments were performed at 60% RH.

Discussion

Rubbers have always received less attention than thermoplastics, in the vast literature on polymer electrostatic charging. For instance, the careful analysis of polymer charging published by Diaz and Felix-Navarro[34] concludes with a semiquantitative plot that does not include any elastomer, even though earlier data on silicone and olefin rubbers are mentioned by these authors. Elastomers are also absent from the triboelectric series published by Iuga et al.,[35] Fujita et al.,[36] as well as from the results published by Németh et al.[37] “Rubber” appears in another published triboelectric series, but without any information on the rubber type used.[38] Therefore, it is not surprising that rubber charging due to stretching and contraction has not been previously described.

The charging patterns shown by natural and silicone rubbers are different, but both change fast, with short equilibration times under 60–65% RH. Interestingly, fast charge dissipation takes place when silicone is under tension, different from triboelectricity that is the immediate outcome of mechanical action on material surfaces.

The surface charge density of natural and silicone rubber can be calculated from the measured potential, and it is in the range of 100–250 electron charge units/μm2 or 0.01–0.025 units of charge/μm3 (considering rubber tubing volume). Surface charge concentration is thus ca. 10–9 mol m–2, which corresponds to only 1.2 × 10–8 at. %. Unfortunately, this is far too low to be detected even by sensitive surface analytical techniques, such as X-ray photoelectron spectroscopy (see also Supporting InformationSimulation Methods). For the area covered by the copper cylinders, we estimate 28 nC/cycle for silicone rubber and 11 nC/cycle for natural rubber.

To understand charge/discharge phenomena during rubber tubing extension and relaxation, we checked its adherence to known types of insulator electrostatic charging: piezoelectricity, surface piezoelectricity, flexoelectricity, triboelectricity, and hygroelectricity that is charge build-up associated with high atmospheric humidity.

The effect of ambient RH supports the classification of rubber electrification as a manifestation of hygroelectricity. On the other hand, it does not follow the expected trends for piezoelectric or triboelectric charging because charging should be more pronounced under low humidity in both cases, when surface conductance is lower.

Other relevant arguments are:

the fast relaxation times for the change in the charging state during both extension and relaxation differ from what is normally observed in mechano- or tribochemical reactions;[39]

rubber does not undergo friction with any other solid, during these experiments, for which reason the known mechanisms for tribocharging do not apply; and

rubber charging during relaxation of the prestressed rubber does not also show the expected feature for piezoelectric charging, when charge separation takes place during the application of a pressure or tension to the sample going back to the initial state under relaxation.

Flexoelectricity and surface piezoelectricity were mentioned in the Introduction and their contributions should also be considered. However, neither is dependent on the RH. Flexoelectricity depends on strain gradients whose formation was minimized at the macroscopic level, by using samples with cylindrical symmetry. Other samples with other shapes also showed the periodic charging in phase with stretching. However, both natural rubber and silicone rubber usually contain small inorganic particles, either natural as in the former[40] or purposely added silica particles in the latter. These particles adhere strongly to the polymer matrix, and they may originate local strain gradients whose effect should be assessed in future work. Surface piezoelectricity cannot be ruled out and it should be kept in mind, but it has been observed in crystalline materials with a large surface-to-volume ratio only.

It is conceivable that internal friction among the rubber chains leads to heterolytic bond breaking and thus to the appearance of charge within the rubber. However, neighboring positive and negative ions together do not produce electric potential at a distance much larger than their separation, measured by a Kelvin electrode.

Many charging and discharging phenomena are associated with water adsorption and desorption from the atmosphere,[2,4,41,42] when a sample is placed under nonzero potential, or when the RH changes, or still following electrostatic charging by corona[43] or friction.[44]

The present case differs from other hygroelectricity phenomena in one aspect: stretching or relaxing the rubber produces a voltage peak that decays fast while the rubber sample remains stretched or relaxed. Nevertheless, potential decay is significantly slower than its rise.

Another explanation of rubber charging under periodic stretching derives from the following statement in Heinicke book: “...nonequilibrium states occur at fresh surfaces during the tribosorption of gaseous mixtures...”,[26] even though this was made for nonrubber materials.

To understand the meaning of “fresh surface” in the stretched rubber, we recall that solid surfaces in equilibrium with the atmosphere are always enriched in the components with lower surface tension, thus reaching minimal surface Gibbs energy[45] and imparting hydrophobic character to the solid. Polymer surfaces under the atmosphere are formed by low-surface-tension components: oligomers, nonpolar chain segments, and highly branched chains. The oxidized segments in PE remain in the subsurface unless the PE sample is immersed in warm water, when they replace the nonpolar surface constituents.[46]

Stretching the rubber increases its surface area, so that polar groups found at the subsurface may be exposed to the atmosphere when the rubber is stretched, tending to migrate back to the subsurface when the rubber is allowed to re-equilibrate.

Charge deposition concurrent with water vapor adsorption can then take place, and this will be governed by the Bronsted acid behavior of the surface components: acidic groups adsorb OH– preferably to H+ ions acquiring negative charge, whereas basic groups show the opposite behavior.[47] Moreover, hydrophobic surfaces adsorb OH– preferably to H+ ions.[48−52]

Both natural rubber and silicone rubber show a negative charge soon after they are allowed to relax, in agreement with the argument in the previous paragraph.

However, natural rubber tubing acquires net positive charge when pulled, different from silicone. Natural rubber is a complex material, containing phospholipids, proteins, and many minor inorganic constituents that provide basic groups.[53] When these are exposed in the stretched surface, water vapor adsorption takes place carrying excess hydronium ions, thus imparting the observed positive charge to the rubber.

On the other hand, many factors may contribute to the observed fast charge dissipation, following each stretching or relaxation step: (i) different surface sites may adsorb charged water from the atmosphere but at different rates, thus enhancing or canceling the charge imparted by the initial load of fast-adsorbed ions; (ii) the positive or negative charge that is initially adsorbed attracts opposite ions from inner rubber layers or from the atmosphere, bringing them closer to the measuring electrode and producing a charge compensation effect.

The suggested mechanism for charge accumulation and dissipation during rubber stretching–relaxation cycles is in Figure .

Figure 8

Proposed mechanism for charge build-up and dissipation in rubbery materials.

The present results suggest that charge and potential measurements could be used to detect surface chemical modifications in elastomers and other polymers. This is a great challenge, even for the most sensitive current analytical tools. The modification of the chemical composition of the rubber surface during extension can perhaps be verified by using spectroscopic and other methods. In one of the few reports of this type in the literature, the present authors identified the charging agents produced by mutual friction of two thermoplastics by using extremely sensitive techniques, including electron energy-loss spectroscopy.[31] However, the detection of fast transient change in the surface composition of complex polymer materials under normal pressure remains as a great challenge, requiring the development of new analytical tools. The best perspective to identify the minute amounts of charging ions in electrified surfaces is probably mass spectrometry, but its application depends on sampling techniques that exclude any artifacts due to ion generation during sample insertion in the instrument. Work in this direction is now being planned.

The effect described in this work seems to be quite general because it was observed, with variations in intensity, for every elastomer sample that was tested, including household items. Moreover, it should also be observed in any solid or gel under mechanical tension because surface extension and contraction modify the surface structure, probably changing the amount and state of adsorbed water. However, the only class of materials that can show many-fold elastic stretching are elastomers, and this is the most likely reason why similar phenomena have not been observed in other solids.

Beyond hygroelectricity, there is also the possibility to develop surface piezoelectricity and/or flexoelectricity by modifying the rubber surface with inorganic nanoparticles, which was extensively done in this laboratory.[31,54,55] Stretching rubber nanocomposites containing large number of nanoparticles should create strain gradients because the particle moduli are certainly lower than the rubber matrix modulus and the particle-matrix adhesion is often very high.[56,57]

Finally, the possibility of collecting energy from stretched rubber is demonstrated in this work, although with a low efficiency (an approximate electrical power output of 0.2 μW). However, the experiments reported here were designed to exclude friction and surface contact that would not allow this to be characterized as a purely surface stretching phenomenon. More effective energy scavenging devices are currently under examination, attempting to benefit from every effect that can contribute, synergistically. Because mechanical vibrations are ubiquitous, at least part of their energy could be collected in rubber microgenerators, instead of being dissipated in the environment.

Conclusions

Elastomers acquire charge when they are periodically stretched and relaxed, producing periodic potential oscillations with the same frequency as the mechanical strain. This is strongly dependent on the RH, and it is discussed considering many possible explanations: piezoelectricity (bulk and surface), triboelectricity, and flexoelectricity. Another possibility is charge pick-up and dissipation by stretched elastomers arising as the result of water adsorption and the partition of water ions (H+ and OH–) in the rubber–air interface, triggered by periodic rubber surface modification: buried chemical groups are exposed during extension and hidden upon contraction.

Experimental Section

Materials

Silicone and natural latex tubing (for laboratory and medical use) with 5 mm internal diameter was cut in 15 cm length pieces and rinsed thoroughly with deionized water, dried in oven at 60 °C during 60 min and kept in a desiccator prior to use. The identity of the tubing was verified by infrared spectroscopy, and the ash content was determined by thermogravimetric analysis. The results are in the Supporting Information.

Electrostatic Potential Measurements during Rubber Cyclic Stretching–Relaxation

The schematic representation of our experimental setup is shown in Figure (see also Supporting Information Figure S3, Video S2). Rubber tubing was mounted on aluminum holders lined with a PTFE sheet, on a reciprocating shaker with an oscillating frequency of 4 Hz and an amplitude of 10 cm, unless specified otherwise. Minimum holder distance is 12 cm, increasing up to 22 cm, periodically. The electrostatic potential measurements were made with a Kelvin electrode (6000B-7C, aperture size of 1.32 mm diameter) positioned at the center of the tubing surface, kept at 2 mm above it, and connected to a Trek model 347 voltmeter. This system can measure static potential in the −3300 to +3300 V range. The voltmeter output was connected to a Keithley 6514 electrometer through a low-noise triaxial cable under high-speed acquisition rate (120 readings/s), using an USB-to-GPIB interface (Keithley KUSB-488b). A NI USB-6009 Multifunction Data Acquisition device was employed, alternatively. During the experiments, temperature was 20–25 °C, and RH was in the 60–70% range or as indicated in the figure captions.

Sample Handling Protocol

Experiments were prepared according to the following protocol: silicone or latex tubing samples were manually withdrawn from the storage desiccator containers using an antistatic glove (Ted Pella ESD All-Day Glove), then mounted on the aluminum holders lined with PTFE, and allowed to equilibrate for 30 min, prior to running the experiments. This simple procedure was sufficient to obtain low residual potential throughout the samples, prior to starting to stretch and relax the samples. In each run, the electrostatic potential was initially recorded on the relaxed sample for 60 s, then the reciprocating drive was turned on for 180 s, observing that the final and initial rubber states were the same, either stretched or relaxed. After the periodic motion, electrostatic potential measurement continued for a predetermined time or until the measured potential went down to 0 ± 2 V.

Energy Scavenging

Latex and silicone rubber tubing pieces were separately inserted within two coaxial, mutually insulated copper cylinders (6.5 cm long and 1.5 cm internal diameter) as in a bottomless Faraday cup. The experimental setup is shown in the Supporting Information (Figure S3), and the running experiment can be watched in Video S1. The copper cylinders do not move, while the tubing is cyclically stretched and relaxed. The internal cylinders in each pair are connected to the input of an integrated full-wave rectifying bridge, whose output was connected to a capacitor (10 μF), whose voltage was measured using the Keithley 6514 electrometer.