Natural Polyisoprene Composites and Their Electronic Applications

Abstract

Rubber-based composites have been recognized as efficient materials for the fabrication of technologically important products. Various particles are successfully incorporated into cis-polyisoprene or natural rubber (NR) in recent years both in solution and in melt forms. Potential electronic applications of such composites specifically containing carbon nanotubes, graphene, graphene-like structures, fibers, metallic fillers, and inorganic fillers have been realized in this article. Advanced performances of NR composites obtained via different methods are compared with those of the neat polymer. Special attention is paid to the structural changes occurring in the matrix under the influence of fillers. Other issues regarding the technology limitations, research challenges, and future trends are also discussed. The main objective of this review is threefold: (1) to present the latest electronic applications of NR composite technology and development, (2) to describe the need for fundamental research in this field, and (3) to outline major challenges in rubber composite preparation. At first an overview of NR composites, then their preparation methods, and thereafter their applications are described. In short, other than summarizing different classes of particles filled NR composites and their applications, this review highlights different ways to create smaller, cheaper, lighter, and faster devices based on such materials. The developed materials are highly useful in the fields of electronics and diffusion as well as in the marine and transport industries.

1 Introduction

Natural rubber (NR) is one of the most important biosynthesized polymers displaying excellent chemical and physical properties. Among other rubbers, it was the first industrially exploited [1, 2]. This elastomer has tremendous applications in various areas [3–5]. The source of this polymer is Hevea Brasiliensis tree and is present in the form of a chemical-free biomacromolecule, natural rubber latex (NRL). NRL is a stable colloidal dispersion of cis-1,4-polyisoprene (PI) of high molecular mass in aqueous medium and can be stabilized with ammonia. The highly stereo-regular microstructure of NR provides unique mechanical properties (high elasticity) and crystallizing nature [6–10] to it. NR molecules consist mainly of cis-PI units without any trans configuration, in contrast to the synthetic polyisoprenes [11]. The very flexible backbone leads to very low glass transition temperature (T _g) of about −64 to −70 °C. Raw NRL finds wide applications in manufacturing medical products such as gloves, condoms, and safety bags due to its excellent elasticity, flexibility, antivirus permeation, good formability, and biodegradability [12–14]. However, NRL products that have short shelf lives and life cycles [15, 16] especially for medical gloves and condoms, low tensile strength, poor tear resistance and low resistance to burning [17, 18] are observed. Even though these drawbacks sound no problem for the medical products, industrially important materials such as tyres require superior performance. Due to this, several attempts to incorporate graphene, carbon nanotubes (CNTs), carbon black (CB) [19], ultrafine calcium carbonate [20], modified montmorillonite (MMT) [21], silica [22], and starch [23] fillers in dry NR or NRL [24] have been reported [25, 26].

Based on the dimensions of filler materials, it can be macrosized, microsized, and nanosized. For nanosized fillers, still the classification depends on the structure or dimension. More specifically, layered clay minerals and graphene sheets come under one-dimensional fillers, nanotubes and nanofibers are two-dimensional, and spherical particles are three-dimensional [25, 27–36] based on the number of dimensions in nanorange. Nanocomposites will be formed if one of the reinforcing filler phases has its dimension in the nanometer range. For nanofillers, maximum improvement in combined characteristics is generally observed at low loading levels, without significantly increasing the density of the compound or reducing light transmission. For instance, when the microsized fillers having lower surface area at 10–30 wt% loading impart good properties to the NR matrix, the nanofillers at 2–3 wt% loading can give superior properties due to its high surface area [37, 38]. The high reinforcing efficiency of nanofilled rubber composites, even at low loading of filler, can also be attributed to the nanoscale dispersion and the very high aspect ratio (length-to-thickness ratio) of the filler [39–43]. So the use of nanofillers in regulating the performances of NR products by alleviating the aforementioned disadvantages of NR is very fascinating and has great interest. The effect of two-dimensional nanoclay with high aspect ratios on NR microstructure when stretched was studied by Carretero et al. [44, 45] and by Huang and coworkers [46, 47]. Also, the reinforcement mechanism of NR composites with nanoalumina [48], carbon nanofibers [49], cellulose nanofibers [3, 50], and kaolin and silica [51] was studied extensively. Renewable and biodegradable biomaterials such as cellulose [52–54] consist of crystalline nanofibers of 10–40 nm diameter and 1.5 g/cm³ density with a length of few micrometers. Such nanofibers possess high value of tensile strength (10 GPa) and Young’s modulus (134 GPa) [55].

Out of the several techniques of filler reinforcement of NR, melt intercalation [37, 56], latex compounding [38], and solution-mixing methods [57] are the most common. In the nanocomposites, the dispersed nanofillers make some structural reinforcement which is responsible for the enhanced chemical and physical properties. Nair et al. [58–60] have found a three-dimensional filler network within the NR matrix responsible for the high solvent resistance and mechanical properties, when nano crab chitin whiskers were used as reinforcement. However, direct methods of reinforcement have certain limitations to reinforce NR with fillers of higher specific surface such as spherical inorganic nanoparticles. This is probably because of the necessity to solve strong self-aggregation of nanoparticles prior to composite manufacture [61–63]. In order to achieve significant nanodispersion, self-assembly process is reported for NR/silica (SiO₂) nanocomposite preparation which could achieve better thermal properties [64].

Discovery of conducting fillers such as CNTs has opened up another way to fabricate conducting rubbers. CNTs can be regarded as rolled-up graphene sheets with essentially sp² hybridized carbon units in the form of cylinders with nanodiameters [65–70]. The extraordinary mechanical, thermal, and electrical properties of these nanomaterials make them potentially applicable in material science, optics, and electronics. Though CNTs prove to be ideal candidates to enhance NR properties [71], simple mixing methods were reported to be insufficient for great improvement. The general issues such as homogeneous CNT dispersion, orientations and aspect ratio, polymer morphology, CNT polymer, and CNT–CNT interfacial interactions have to be considered during composite fabrication [72]. Among the nanofillers, graphene layers also have remarkable importance in suturing the rubber properties, especially because these fillers are easier and cheaper to make compared to CNTs. The influence of graphene on the microwave properties of elastomeric composites is highly noteworthy [73–76]. In addition to CNTs and graphitic fillers, metal-coated fibers such as aluminum- or nickel-coated carbon or glass fiber and nickel-coated mica, polyaniline, short carbon fiber, and conductive CB impart conductivity to NR composites. The major electronic applications of such composites include various touch control switches, and strain and pressure sensors for applications such as robot hands or artificial limbs. Gaskets made from electrically conductive rubber composites can prevent electromagnetic induction (EMI) radiation leakage. Moreover, these materials are considered for use in different microelectronic devices and microwave applications, such as absorbing materials.

However, the above-mentioned problems associated with CNTs are often observed for all nanofillers. In order to fabricate high-performance composite products, the compatibility between the filler and the hydrophobic polymer matrix is really necessary and can be accomplished by physical or chemical modifications of the nanoparticle surfaces [10, 43, 77–83]. Using various techniques, a lot of technologically important materials have been fabricated out of NR-based composites. The objective of this critical review is to explore all electronic applications of NR composites (both micro and nano) and to highlight the high performance of nanocomposites.

2 NR-Based Composites

2.1 Synthesis Methods

The primary difficulty associated with the filler particles is to achieve a uniform dispersion in the matrix polymer, and due to this reason, processing remains a key challenge in fully utilizing the advantages of filler reinforcement. Nonuniform dispersion results in the formation of filler aggregates which can degrade the mechanical as well as physical performances of the composites through defect sites. Particle dispersions also depend on the viscosity of the medium, diffusion of particles, and particle–particle and particle–matrix interactions [84]. Better dispersion and superior properties can be achieved if the interparticle distance is comparable to the radius of gyration of the polymer. In the composite, the entire polymer matrix can be a part of an interphase and this influences processing as well. Anyway, common processing methods for polymer composites including solution processing [57], melt processing [56, 85], and reaction processing [84] are applicable for NR composites also. Latex compounding [38] method is employed especially for fabricating latex-based composites.

2.1.1 Solution Processing

Fillers (especially nano) are generally dispersed in a suitable solvent and then mixed with polymer solution by shear mixing, magnetic stirring, or sonication [86] in this technique. Sonication can be of bath (mild sonication) or probe (high-power sonication). Finally after mixing, the composite can be obtained by vaporizing the solvent. Figure 1 shows the flowchart for the solution processing. This method is considered as an effective way to prepare composite films with homogeneously distributed nanofillers [87]. However, since this method depends mostly on the efficient filler dispersion in the used solvent, the selection of solvent is rather significant. Often functionalization methods are employed in order to avoid reaggregation of fillers, and solvent selection is done on the basis of these functional groups and solubility of the polymer. Sometimes the latex can be directly used for solution mixing. This was done by diluting the NRL and adjusting the pH. The fillers in this case would be made in such a way that with addition, NRL composite will precipitate [88].

Fig. 1

Flowchart presenting the different steps of the solution and melt mixing processing of a polymer composite

2.1.2 Melt Processing

The main principle behind this method is softening of rubber when heated. During this process, rubber transforms into viscous liquid, and at this state, additives can be well mixed into the melt by rheological shear mixing [87]. In addition to the common melt mixers, large extruders are also used for bulk sample mixing. The process of melt mixing is also included in the flowchart (Fig. 1). Starch/NR/clay nanocomposites were prepared via melt extrusion technique in a high-speed mixer [10, 89]. Open two-roll mill mixing was also practiced at room temperature. The vulcanization ingredients, plasticizers if any, and fillers were added to the elastomer during mixing. NR composite containing clay or organoclay, and octadecylamine in the absence of clay was synthesized, and its effect on the cross-linking degree of NR is investigated [90]. In short, the melt mixing possesses the advantages of speed and simplicity, and compatibility with standard industrial techniques [91, 92].

2.1.3 Latex-Stage Compounding

Fillers along with curing agents such as sulfur in the presence of ZnO, stearic acid, TMTD, MBTS, etc. are compounded with NRL. At first, the curing agents are mixed (or compounded) with NRL followed by filler dispersion. This process improves cross-linking between the polymer chains in the nanocomposites, and thus, better properties can be achieved [38]. Even though in situ polymerization method is another efficient method to significantly improve the dispersion of fillers in polymers and to achieve stable filler–polymer interaction, for NR little reports exist on this. This is because of the availability of NR in its polymer form and the difficulty in controlling the polymerization of isoprene into cis-polyisoprene. Various methods to synthesize NR/CNT composites are discussed by Ponnamma et al. [42] in their review.

2.2 NR Composite Applications

2.2.1 Electromagnetic Interference Shielding

Metals are generally used to shield unwanted electromagnetic (EM) radiations; however, their heavy weight, high cost, and corrosive nature demand the need of cheaper polymeric materials. Though polymers are insulating, conducting nanoparticles produce their nanocomposites possessing good shielding effectiveness (SE). There is an increase in demand for the EMI shielding particularly at low-frequency range (30 kHz–1.5 GHz) due to a significant increase in the use of portable wireless devices such as cell phones and radios which interfere with increased sensitiveness [93]. Due to the structural and electrical advantages of nanocarbon materials, these are largely used to fabricate EMI shielding. Before going to the experimental details of the shielding materials, the theory behind EMI shielding is discussed here.

Figure 2 shows the experimental setup used by Zhang et al. to calculate the EMI SE of CNT-filled shape memory polymer. This setup is the same for NR samples as well. In general, all devices contain a network analyzer, transmitting horn antenna, and a receiving waveguide probe as shown in the figure. According to the frequency to be shielded, horn antenna and waveguide probe are changed. Here, the distance between the mutually perpendicular waveguide probe and the horn antenna is maintained as three times the wavelength for a given measurement frequency. The output from the network analyzer is put into the horn antenna, and the energy received after transmitting through the sample is measured by the waveguide probe. EMI SE is defined as the difference of the transmitted energy between the nanocomposite sample and a reference material. If the distance between the shield material and the source of radiation is higher than λ/2π, it is said to be in far field, and if the distance is less than λ/2π, it is in the near field [94].

Fig. 2

The schematic of an EMI shielding effectiveness measurement system [94]. Copyright 2007. Reprinted with permission from Elsevier

The SE evaluates the reduction of EMI at a specific frequency (f) as illustrated by Eq. 1.

$${\text{SE}} = 10\log \frac{{P_{0} }}{{P_{t} }} = 20\log \frac{{E_{0} }}{{E_{t} }} = 20\log \frac{{H_{0} }}{{H_{t} }}$$

(1)

where P ₀, E ₀, and H ₀ are the energy and electric and magnetic field intensities incident on the shield and P _t , E _t , and H _t are the counterparts transmitted through the shield.

During EMI shielding, reflection, absorption, and re-reflection of the radiation can occur. These reflections as well as absorption loss depend on the electrical conductivity (σ) and magnetic permeability (μ), and such multiple reflection loss can be minimized when the distance between the reflecting surfaces or interfaces is larger than the skin depth (δ).

$$\delta = \frac{1}{{\sqrt {\pi f\mu \sigma } }}$$

(2)

SE can also be explained as the sum of reflection loss and absorption loss as shown in Eq. 3

$${\text{SE}} = 20\,\log \frac{{Z_{0} \delta \sigma }}{2\sqrt 2 } + 8.68\frac{t}{\delta }$$

(3)

where t is the thickness of the material and Z ₀ = 120 π. If t/δ ≤ 1.3, the equation reduces to

$${\text{SE}} = 20\,\log \left( {1 + \frac{{Z_{0} t\sigma }}{2}} \right)$$

(4)

Thus, in short, the SE depends on the frequency of the radiation, thickness of the material, and resistivity. The EMI characteristics of nanocarbon particles and barium hexaferrite (BF)-filled NR are examined by Yakuphanoglu et al. The result observed by them is given in Fig. 3. The SE of the composites is frequency dependent, especially at higher-frequency range, and it increases with the increasing frequency (Fig. 3a). Variation of skin depth was in accordance with the dielectric constant of the composite (Fig. 3b). The SE also increases with filler loading (Fig. 3c) as well (0–20 wt%). As the filler loading increases and the volume resistivity decreases, the SE of the NR/BF nanocomposites increases rapidly [95].

Fig. 3

a Shielding effectiveness as a function of frequency, b variation of skin depth and dielectric constant with BF content, and c SE with thickness of NR/BF nanocomposites [95]. Copyright 2014. Reprinted with permission from Springer

Since the EMI shielding studies are mainly based on the shielding studies of microwave radiations, this is deeply discussed in the following Sect. (2.2.2).

2.2.2 Microwave Absorption

Very recently, a large number of studies have come out on the microwave absorbing ability of NR nanocomposites. Microwave absorbing rubber composites of NR and epoxidized natural rubber (ENR), filled with conductive CB and aluminum powder, were prepared, and their shielding effectiveness was studied as shown in Fig. 4 [96]. For this, X-band microwave in the frequency range of 8–12 GHz is used. It is found that increase of filler loading-enhanced SE of the rubber composites and also the conductive CB was more effective in shielding than aluminum powder. Binary filler-filled rubber composites showed higher SE than that of single filler-filled rubber composites. SE of each sample was in the same range under the entire frequencies (8.5–12 GHz), and the effect of fillers on SE of two rubber systems was similar. Unfilled rubbers showed very low SE, <1 dB. SE of the rubber increased from <1 dB to 18–28 dB after adding 50 phr of CB or to 30–40 dB after adding 50 phr of CB and 50 phr of aluminum powder. The maximum SE of rubber composites containing 50 phr of aluminum powder (0/50 samples) was 10 dB. The SE and the mechanical properties were in the order of ENR > NR, and both these properties were correlated with electrical conductivity as well. Although the conductivity was higher, aluminum powder was less effective than conductive CB because of the lower volume fraction and larger grain size. The EMI SE of CB-filled rubber composites (50/0 samples) observed here (18–28 dB) was higher than that of other rubber composites containing 50–60 phr of the same filler CB, i.e., nitrile rubber (12 dB), ethylene vinyl acetate (EVA) (18 dB), ethylene propylene diene terpolymer (EPDM) (8–10 dB), and EVA/EPDM blend (22.5 dB).

Fig. 4

Effect of single filler on SE: a NR composites; b ENR composites. The former number represents CB content, and the latter number represents aluminum powder content [98]. Copyright 2012. Reprinted with permission from Scientific Research Open Access

The factors affecting the EMI SE include conductivity or polarity of rubbers and conductive fillers, filler content, particle size/shape, filler dispersion and distribution, specimen size and shape, and testing method [97]. Magnetic and microwave absorbing properties of thermoplastic natural rubber (TPNR)-filled 4–12 wt% magnetite (Fe₃O₄) nanocomposites prepared by melt blending were investigated. TPNR matrix was prepared from polypropylene (PP), NR, and liquid natural rubber (LNR) in the ratio of 70:20:10 with the LNR as the compatibilizer. The TPNR–Fe₃O₄ composites containing nanocubic spinel-structured Fe₃O₄ with 20–50 nm size showed good microwave absorption property. The absorption or minimum reflection loss (RL) continuously increases, and the dip shifts to a lower-frequency region with the increase in both filler content in nanocomposites and the sample thickness. The RL is 25.51 dB at 12.65 GHz, and the absorbing bandwidth in which the RL is less than 10 dB is 2.7 GHz when the filler content is 12 wt% at 9 mm sample thickness [97]. At low frequencies, the dielectric loss in the samples is influenced by direct current conductivity, while at higher frequencies, the loss is due to alternating current conductivity. The microwave absorption properties analyzed by using peculiar absorber method reveal the existence of two matching conditions of minimum reflection due to the geometrical cancellation of the incidence and reflected waves at the surface of the absorber at low and high frequencies for the same thickness of the samples. The increment of filler content in the nanocomposites can shift the maximum attenuation to a lower frequency, increase maximum absorbing effect, and also enlarge the absorbing bandwidth. The minimum RL was found increasing and moving toward the lower-frequency region with increase in sample thickness. In short, the microwave absorbing properties of nanocomposites were determined by the thickness and the composition of filler, which can vary the dip frequency, microwave absorbing properties, and the absorbing capacity and width of the nanocomposites.

Microwave (reflection coefficient, attenuation coefficient, SE) properties of NR/GNP nanocomposites revealed interesting dependence of these parameters on frequency and filler amount (Fig. 5). The reflection coefficient increased with filler amount and frequency, whereas the attenuation coefficient was independent of frequency and filler amount in the 1–6 GHz range and increased slightly from 6 to 9 GHz and drastically in the 9–12 GHz range. The dependence was almost linear, and both the frequency and filler amount influence the SE values (10–34 dB range) as well (Fig. 5a). Since the attenuation was not great enough to compensate the increasing reflectance, the SE decreased gradually with the increasing frequency, especially up to 7 GHz. However, as a whole, the effect was more pronounced at 6–12 GHz frequency range and at 6 phr of GNP [98].

Fig. 5

Frequency dependence of a SE and b attenuation coefficient α of electromagnetic waves at various filler concentrations (n-phr of graphene) [98]. Copyright 2012. Reprinted with permission from Scientific Research Open Access

The lowering in the attenuation coefficient values (Fig. 5b) of composites is attributed to the presence of a foaming agent which makes the composite structure porous. As a whole, the attenuation coefficient slightly increases with the increasing frequency and filler amount. The increase is particularly in the 6–12 GHz region. In the 1–6 GHz region, the effect of frequency and filler amount on the attenuation coefficient is less pronounced. This effect was in accordance with the variation observed in the complex dielectric permittivity [98].

2.2.3 Sensing Properties

Sensors based on NR composites are widely reported [68]. The sensors come under different categories such as strain sensor to detect large strains, pressure sensors to detect pressure variations, and solvent and vapor sensors to detect organic solvents and volatile components. On NR substrate, thin layer of the conducting polymer, polypyrrole (PPy), is coated by vapor-phase polymerization technique in vacuum [99] using the oxidant FeCl₃. This material is successfully used as a strain sensor to detect large mechanical strains, the sensing performance of which can be regulated by various parameters. Prestraining the rubber substrate prior to deposition, high concentration of FeCl₃, and formation of monomer vapor from diluted pyrrole monomer resulted in maximum stability, consistency, and best sensing performance of the composite. Oxygen plasma exposure produced the highest reduction in water contact angle on the rubber’s surface to improve the hydrophilicity without sacrificing the rubber’s elasticity and flexibility. The developed strain sensor was successfully implemented on an air muscle in conjunction with a PID controller where the extension/contraction of the muscle can be accurately controlled. It also has a capability to measure large bending strains and hence be used as a rotary sensor. The performance of the sensor can be explained by examining at the strain distribution on the NR substrate during the bending motion at different angular positions (Fig. 6). The length of the NR substrate was broken down into four segments, and the physical change in length of each segment was measured. The disadvantage of this sensor is that it shows some hysteresis.

Fig. 6

The strain distribution along the length of the rubber at a 10°, b 30°, c 50°, d 70°, and e 90°. The different segments are represented by the red lines [99]. Copyright 2012. Reprinted with permission from Elsevier

In order to understand the strain profile of the NR/PPy sensor, the electrical resistance versus time profile is taken as given in Fig. 7a. The upward shift of the electrical resistance over the time indicates that the absolute electrical resistance value cannot be used to represent the absolute strain. So normalized ∆R/R ₀ values were determined and plotted against strain (Fig. 7b). The sensitivity can be calculated by gauge factor (GF) which is defined by Eq. 5.

$${\text{GF}} = {{\left[ {{\raise0.7ex\hbox{${\left( {R - R_{0} } \right)}$} \!\mathord{\left/ {\vphantom {{\left( {R - R_{0} } \right)} {R_{0} }}}\right.\kern-0pt} \!\lower0.7ex\hbox{${R_{0} }$}}} \right]} \mathord{\left/ {\vphantom {{\left[ {{\raise0.7ex\hbox{${\left( {R - R_{0} } \right)}$} \!\mathord{\left/ {\vphantom {{\left( {R - R_{0} } \right)} {R_{0} }}}\right.\kern-0pt} \!\lower0.7ex\hbox{${R_{0} }$}}} \right]} {\left[ {\left( {{\raise0.7ex\hbox{${L - L_{0} }$} \!\mathord{\left/ {\vphantom {{L - L_{0} } {L_{0} }}}\right.\kern-0pt} \!\lower0.7ex\hbox{${L_{0} }$}}} \right)} \right]}}} \right. \kern-0pt} {\left[ {\left( {{\raise0.7ex\hbox{${L - L_{0} }$} \!\mathord{\left/ {\vphantom {{L - L_{0} } {L_{0} }}}\right.\kern-0pt} \!\lower0.7ex\hbox{${L_{0} }$}}} \right)} \right]}}$$

(5)

where R is the final resistance and R ₀ the initial resistance of the sample which varies its length from L ₀ to L over the application of strain.

Fig. 7

a The change of resistance profile over 800 cycles and b the relationship between the strain and relative change in electrical resistance of the strain sensor from linear motion strain [99]. Copyright 2012. Reprinted with permission from Elsevier

From Fig. 7b, the gauge factor for NR/PPy sensor is calculated to be approximately 1.86. The ∆R/R ₀ of each strain value was averaged to determine the mathematical relationship between the averaged ∆R/R ₀ and strain.

However, nanocomposites of NR extend stable and reversible piezoresistive strain sensing over 3000 tested cycles upon applying a small compressive strain (~16.7 %) under a relatively high frequency (0.5 Hz, 2 s/cycle). This is done by adding CNT and carbon nanofibers in NR composites pretreated with CB. The large reinforcement effect makes this composite suitable for heavy-duty pressure sensors, i.e., healthy motion monitoring of industrial machinery vibrations [100].

The strain profile of the CNT/graphite flake-sandwiched NR sensor is illustrated in Fig. 8a. The resistance increases with strain, and the CNT sensor (because of the smaller CNT size) showed higher measurable strain limit (620 %) compared with the graphite sensor (246 %). The sensitivity was calculated by GF as given in Eq. 5. The sensitivity values for the MWCNT sensors were 5.5, 10.2, 16.3, 23.1, 31.5, and 43.4 at 100, 200, 300, 400, 500, and 620 % strain, respectively, where as for the graphite sensors (Fig. 8b) the values were 23.7, 37.5, 148.1, and 346.6 at 50, 100, 200, and 246 % strain, respectively. The responses over time were checked by attaching one end of the MWCNT strain sensor to a disk, which was rotating with an angular velocity of 2 rad/s. The results obtained are given in Fig. 8c where the applied sinusoidal strain (with a period of ~3.04 s) ranged between 150 and 500 %, and the total number of cycles was more than 400 [101].

Fig. 8

Relative resistance versus strain response curves for the a MWCNT sensor, b graphite sensor, and c multiple cycle tests for the strain range of 150–500 % (Inset image the first 16 full cycles carried out over a time period of 486 s) [101]. Copyright 2014. Reprinted with permission from MDPI

The superior mechanical and electrical properties of CNTs became the basis for fabricating large number of sensors. Their nanoscale electrical properties are often related to the mechanical deformation as in the previous case [68, 102]. Macroscale strain sensor based on PI-MWCNT composites (PMCNTCs) was prepared. Better dispersion of the nanoparticles was ensured by dispersing CNTs in chloroform and thereafter mixing with the PI matrix solution. About four orders of magnitude of nonreversible change in the electric resistivity at large (40 %) stretch were observed due to the entangled structure of nanotube, thus making the PMCNTC an attractive tensile and compressive strain sensor. On the other hand, the high structure carbon nanoparticles have shown about six orders of magnitude of reversible change in the electric resistivity at large (40 %) stretch. This is due to the high porosity of carbon agglomerates which provides better adhesion to polymeric globules and mobility of nanoparticles. The maximum change at stretch for both types of composites was observed near the percolation threshold. PMCNTC can be used for small tensile strain sensing, but high structure CB–PI composite are preferable for large tensile strain sensing. The resistance of the composites was examined with regard to the force of stretch F and the absolute mechanical deformation in the direction of the force. No variation in the electric resistance was found for the samples with 7.7 mass parts of highly structured carbon black (HSCB) and 3.63 mass parts of MWCNT even at 100 % deformation of the composite. Similarly, there was a relatively weak dependence of R on the deformation at 9.35 mass parts of HSCB. Figure 9 clearly shows the variation in resistivity of different samples with respect to strain percentage.

Fig. 9

The electric resistivity ρ of the PI–MWCNT composite and PI–HSCB composite versus small tensile strain ε [102]. Copyright 2007. Reprinted with permission from Elsevier

Of all the composites examined, the best strain performance was obtained for the samples with 8.8 mass parts of HSCB and 8.3 mass parts of MWCNT, respectively, which belong to the region of percolation threshold. The resistivity of these samples changed more than six orders upon a 40 % stretch for 8.8 mass parts of HSCB composite and more than four orders for 8.3 mass parts of MWCNT composite. When the materials are released, the resistivity of HSCB composite returned to its previous value, while in the case of MWCNT composite, a relatively large hysteresis of ρ(ε) was observed because of the entangled structure and high stiffness of nanotubes.

In another study, novel biocompatible NO sensor was made by spin trap (iron(II)-diethyldithiocarbamate complex, FeDETC) incorporated in a latex rubber matrix [103]. The sensor was optimized for the latex rubber matrix temperature of polymerization and FeDETC concentration inside the matrix. The NO trapped in the sensor can be detected by electron paramagnetic resonance (EPR). The EPR signals showed variation with respect to different conditions. The best sensor was made with a latex rubber matrix polymerized at 10 °C and with a FeDETC concentration of 14.8 mM. High stability of the EPR response was detected for 40 days at room temperature.

2.2.4 Dielectric Property

Neat NR is an insulator, but it becomes conductive when conducting nanoparticles are incorporated in it. In addition to the electrical conductivity, some of these composites are good dielectric materials as well. NRL composites with polyaniline (Pani) were prepared by the in situ polymerization of aniline in the presence of latex [104]. The dielectric properties of different compositions of the samples in microwave frequencies were determined (Fig. 10a).

Fig. 10

a Variation of dielectric constant with frequency for the Pani–NR composites [104]. Copyright 2007. Reprinted with permission from John Wiley & Sons. b Dielectric constant for NR as a function of frequency at different temperatures [105]. Copyright 2012. Reprinted with permission from John Wiley & Sons

From variation of the dielectric constant with frequency (S bands 2–4 GHz) of Pani–NR semi-interpenetrating networks, it is clear that the dielectric constant was less affected by frequency. But the value changes from NR (i.e., 2) to Pani–NR, the maximum shown for the 2:1 composition. The absorption coefficient and penetration depth calculated for 2:1 composition were, respectively, the highest and the lowest. NR allowed the passage of electromagnetic radiation, and when the Pani content was increased, the electromagnetic radiations were restricted to penetration. The in situ technique also contributes toward high dielectric constant. The conducting NR could be used in various applications, including antistatic coatings, antistatic chappels, and anticorrosion applications.

Figure 10b shows the effect of both frequency and temperature on the dielectric constant for NR [105]. Both dielectric constant and tan δ of NR are less dependent on frequency and temperature due to the absence of polarization and ionic properties of the material. Since NR is a nonpolar molecule, its dielectric constant will be the square of the refractive index (1.59) [106, 107]; however, the experimental value obtained is higher than the theoretical one due to the presence of interfacial polarization which arises due to the presence of impurities such as the compounding ingredients. At high frequencies, the values become about 2.5 for all temperature ranges, which can be ascribed to the mobility of a polar group in the polymer chains, which is too slow to contribute to the dielectric constant at high frequencies. By following the experiment, the dielectric constant, dielectric loss, and tan δ of barium ferrite–NR composites were studied [105]. All these values decreased as the frequency increases, and at higher frequencies, the effect significantly weakens. ε′ for barium ferrite was related to the preparation method, and at lower filler loading, the ferrite content strongly influences the dielectric properties. Up to 30 Phr ferrite loading, ε′ sharply increases, and thereafter, the dependence of ε′ on ferrite content becomes minimal as a result of ferrite particle interaction. Ferrite content also enhanced the dielectric loss and tan δ. They could observe a strong correlation between magnetic initial permeability and dielectric constant for hard magnetic material polymer composites as well.

Oil palm fibers were modified with silane, and the effect of modification on dielectric properties of NR is checked for the biocomposite. At low level of fiber loading (20 %), the dielectric properties show only a nominal change, and with fiber treatment, the loss factor increases [108]. Even though the effect of the oil palm fibers and their chemical treatment with silanes on the dielectric properties is evidenced well, other factors such as fiber orientation and quantity of chemical used were not addressed in this work. For the sisal/coir hybrid fiber-reinforced NR composite, the dielectric constant values were higher due to the polarization effect induced by fibers’ incorporation. The lignocellulosic fiber addition increased the number of polar groups and decreased the volume resistivity of the composites. At 15.6 % volume, percolation threshold is reached. Upon increase in frequency, the interfacial and orientation polarization decreased and thus the dielectric constant [109].

Fiber treatment, such as acetylation, benzoylation and reaction with alkali, peroxide and permanganate were carried out to improve the adhesion between fibers and matrix. Modification also reduces the moisture absorption capacity of the fibers due to reduced interaction between polar –OH groups of lignocellulosic fibers and water molecules. For composites containing chemically treated fibers, the dielectric constant values were found to be lower for all frequencies due to the increased fiber hydrophobicity. Chemical treatment increased the volume resistivity of the composites as well. In the untreated state, the –OH groups of fibers form strong hydrogen bonds and are thus unreactive. Alkali treatment destroys the hydrogen bonding and makes them more reactive and leads to decreased orientation polarization. It also can lead to fibrillation and thus can enhance the surface area [110]. This is the reason why alkali-treated fibrous composites showed considerable decrease in dielectric constant. In composites, the fibers and NR matrix interlock each other. The addition of two-component dry bonding agent consisting of hexamethylenetetramine and resorcinol improved the interfacial adhesion between the matrix and fibers and reduced the dielectric constant. It also enhanced the electrical resistivity and dissipation factor of the composites. The added fibers and different chemical treatments for them increased the dielectric dissipation factor. A dielectric relaxation has been observed at a frequency of 5 MHz. The dielectric constant of treated fibers decreases in the order, alkali > acetylated > benzoylated > permanganate > peroxide > bonding agent [106]. The conductivity of the composites was also found to increase with increase in fiber loading as the number of conductive particles increased. Finally, the application of such low-cost composites in antistatic materials is explained. Depending on the concentration of alkali, the dielectric constant varies. Fibers treated with 4 % alkali showed the lowest value. Of the three types of silanes, fluorosilane and vinyl silane, respectively, exhibit the minimum and maximum dielectric constant. Mercerization prior to silane treatment is an effective means of attaining better property retention to composites exposed to moisture. Silane coupling agents have inherent natural attraction with both the natural fiber and resin matrix [111].

Effect of alumina particles on the properties NR revealed the ionic polarization of the microalumina particles and the induced dipole moments set up in the matrix. At 0.00968 volume fraction, the composite showed the highest dielectric constant of 2.123 and the electrical conductivity of 2.683 × 10⁻⁸ (Ω.m)⁻¹. The soft and flexible composites containing small amount of alumina can store additional elastic energy within the matrix under electric field and are useful in fabricating biomimetic actuators and/or artificial muscles. ε′ of magnetic NR composites containing uniformly distributed nickel nanoparticles decreased with increase in frequency and increased with increase in nickel loading (Fig. 11). Metallic inclusions in any insulating material enhance the dielectric permittivity of the material as reported [112, 113]. Interfacial polarization due to the heterogeneity of the samples is attributed as the reason for the variation of dielectric permittivity with frequency as well as filler concentration [114]. The dielectric dispersion observed in the frequency range of 100–8 MHz at 40 °C shows the linear increase in permittivity with the nickel particle loading. Here, the dielectric loss in the material is small and can be useful in making capacitors. The mechanism of ε′ is understood by interfacial polarization, enhancement of electrical conductivity [115], and formation of internal barrier layer capacitors [116, 117]. Interfacial polarization arises at the interface of a composite with electrical heterogeneity due to the migration of charge carriers through different phases, and charges accumulate at the interface [118]. External electric field moves these charges, and motions prevented within the medium cause polarization.

Fig. 11

Variation of permittivity with content of nickel in the composites per hundred weight ratio of rubber (phr) is plotted in the horizontal axis [119]. Copyright 2009. Reprinted with permission from Elsevier

Permittivity increases with filler loading almost linearly with all frequencies and at all temperatures. Dielectric loss of blank rubber, as well as the composites, was observed to increase with temperature [119]. The decrease in permittivity with temperature is attributed to the high volume expansivity of rubber at elevated temperatures. Dielectric loss of the composites was found to increase with temperature, and a relaxation peak was only observed in blank rubber. For filled samples, the dielectric relaxation peaks appear to exist below 100 kHz. A weak interfacial polarization present in the rubber matrix can be attributed to the observance of relaxation peaks in blank rubber.

In another work, the relaxation phenomena occurring in NR and polyurethane (PU)/NR blend-based nanocomposites, reinforced by layered silicates, were examined during 10⁻¹ to 10⁶ Hz frequency. NR is nonpolar, and its performance is only slightly affected by the presence of layered silicates. It has low dielectric permittivity and loss, and its relaxation is visible in the low T _g range. But PU exhibits four relaxation processes—ascending relaxation rate, interfacial polarization (IP), glass/rubber transition, and local motions of polar side groups and small segments of the polymer chain. Both matrix and filler influence relaxation phenomena, and IP is present in all nanocomposites being the slowest recorded process [120].

The influence of graphene nanoparticles (GNPs) [98] at 2.0–10.0 phr on the dielectric properties (dielectric permittivity, dielectric loss angle tangent) of NR nanocomposites has been investigated in the wide frequency range (1–12 GHz). The dielectric permittivity increased slightly with the increasing frequency and filler amount, especially significant in the 9–12 GHz range. The tendency was the same in the case of dielectric loss angle tangent, although the impact of the filler amount is less marked, i.e., the imaginary part of the complex relative dielectric permittivity is less sensitive to alternations of the filler amount and frequency than the real part. Figure 12a shows the frequency dependence of the real part of the complex relative dielectric permittivity. The close dielectric permittivity value for various filler concentrations up to 7 GHz is clearly seen. As explained at >7 GHz, the values show significant difference between the nonfilled and filled composites. The values obtained for the nonfilled composites are close to the ones for NR at 1000 Hz reported, which are 2.40–2.70 [121]. At higher than 9 GHz, the high dielectric permittivity comes mainly due to the composite polarity [122].

Fig. 12

Frequency dependence of a relative dielectric permittivity ε′ at various filler content (n-phr of graphene) [98]. Copyright 2012. Reprinted with permission from Scientific Research Open Access. b Normalized dielectric loss ε″ of the nonvulcanized nanocomposites in the normal mode. Solid lines correspond to HN fittings and dashed lines to individual processes [123]. Copyright 2012. Reprinted with permission from Elsevier

The effect of functionalized graphene sheets (FGSs) on the dielectric properties of NR nanocomposites is illustrated in Fig. 12b. The shift toward higher frequencies is observed in the presence of the additives and becomes more significant with FGS. Molecular dynamics of NR is often affected by vulcanizing additives which can decrease the segmental motions [123]. FGSs have defects, free radicals, and other irregularities on its surface and thus cause strong rubber–filler interactions. These interfacial interactions accelerate the cross-linking reaction and increase the electrical conductivity.

Those rubber–filler interactions are also reflected by the changes in the timescale of the global chain dynamics (normal mode) in NR/FGS nanocomposites. The figure gives the effect of FGS on NR in the normal mode. The normal mode is assigned to the chain dynamics and, consequently, is strongly dependent on the molecular weight and on the mobility of the chain-end cross-links in the naturally occurring network of the NR. The final material having multifunctional property enhancements coupled with low cost and large-scale production makes wider applications for it.

2.2.5 Thermal Conductivity

CNTs are reported to be excellent in improving the thermal conductivity of NR which is also significant in electronics. In situ measurements of thermal conductivity, heat capacity per unit volume, and T _g of pure PI and PI/single-walled carbon nanotube (SWCNT) composites cross-linked by high-pressure treatment (4 h at 513 K and 1 GPa) revealed the enhancement in thermal conductivity of both the elastomeric and liquid PI by ∼120 % at 5 wt% SWCNT loading [124]. T _g of the cross-linked elastomeric state increased by 12 K and that of the liquid PI by ∼7 K at the same filler concentration. CNTs enhanced broadening of T _g in the high-pressure-treated densified elastomeric states but not in the untreated samples. The increase observed for cross-linked state can be attributed to the filler-induced increase in the cross-link density. The heat capacity decreased by ∼30 % in both the glassy and liquid/elastomeric states (above and below T _g) due to the large reduction in the vibrational and configurational heat capacity of PI in the presence of SWCNTs. It is also observed that the thermal conductivity (κ) of the composite linearly increased with SWCNT content. Also, the pressure-induced densification decreases the interparticle (CNT–CNT) distances without agglomeration and improves the thermal and electrical contact between the constituent materials of the composite. In the case of a semicrystalline polymer, high-pressure treatment can increase the degree of crystallinity and thus can strongly enhance the thermal conductivity of the matrix. Thus, thermal as well as electrical percolation networks in polymer CNT composites and the nature of thermal and electric transport in them are better studied by high-pressure treatments.

Variation of κ with temperature at 0.02 GPa is shown in Fig. 13a. All samples show similar amorphous behavior where κ is almost constant or slightly increasing with temperature. A small peak is also observed distinguishing two temperature ranges, which is most apparent in PI-SWCNT5. Below the peak, κ moderately increases with temperature, whereas above the peak it becomes temperature independent. More than three times larger change for PI-SWCNT5 is noted compared to PI. For pure PI, the change is less pronounced. The change occurring in κ with respect to temperature is attributed to the thermal expansivity increase at the T _g and the stronger decrease in density above T _g than below [125]. High-temperature and high-pressure treatment causes changes in the number of covalent bonds between the chains, density, and structural order, and affects κ values. During the treatment, both chain scission and cross-linking can occur, and the density as well as the order can decrease or increase due to decreased or increased degree of crystallinity. But here, the authors have ruled out the possibility of chain scission because of sample cross-linking. Comparison of samples of pure PI cross-linked at 1 GPa and 0.5 GPa by submerging in water–methanol mixture revealed that the samples become progressively denser with increasing pressure of the high-temperature treatment. The sulfidic and carbon–carbon cross-links further decrease the already weak tendency of cold crystallization in PI. Since PI is difficult to crystallize, the major part is inevitably amorphous and the difference is likely small. Moreover, a decreased crystallinity would also decrease κ. The increase in κ at room temperature and near atmospheric pressure is therefore attributable to the increased number of covalent bonds between the chains and increased density. The κ values for a few NR nanocomposites are given in Table 1.

Fig. 13

a Simultaneously measured results for thermal conductivity [124]. Copyright 2009. Reprinted with permission from American Chemical Society. b Thermal conductivity dependence on temperature for TPNR and different loaded NiZn–TPNR nanocomposite samples at 30, 60, 90, and 120 °C [126]. Copyright 2013. Reprinted with permission from Elsevier

Table 1 Thermal conductivity of NR nanocomposites

Thermal conductivity dependence of NiZn ferrite nanoparticles filled TPNR magnetic nanocomposite on temperature is given in Fig. 13b. TPNR was prepared from NR, LNR and high-density polyethylene (HDPE) in 20:10:70 ratio by melt blending [126]. Tremendous enhancement of thermal conductivity with increasing filler content up to 8 % of NiZn ferrite is observed, and thereafter, κ decreased slightly with filler content. The calculated enhancement in κ for 4, 8, and 12 wt% NiZn–TPNR nanocomposite samples was about 38, 108, and 81 %. The increase of thermal conductivity with filler rates could be attributed to the nature of NiZn ferrite nanoparticles, which are assumed to have a higher number of phonons’ vibrational modes as well as higher mean free path due to its crystalline structure compared to the amorphous TPNR matrix. Also, the incorporation of NiZn ferrite nanoparticles creates a percolated network facilitating the phonon conduction inside the nanocomposite and therefore enhances κ [127]. Moreover, the T _g value of the TPNR matrix increases with filler content, suggesting an enhancement in the thermal stability of the TPNR matrix. Thus, 8 wt% filler content remarks the optimum filler concentration in the formation of the percolated network, above which κ value decreases. The decreasing trend in κ value for all nanocomposite samples with temperature is due to the increase of structural defects existing in the amorphous behavior of TPNR matrix, which damp the phonons’ vibrational amplitudes.

2.2.6 Shape Memory Materials

Shape memory polymers (SMPs) are smart, responsive materials having numerous potential applications in various fields and in electronics [133]. Shape memory natural rubber (SMNR) has much importance among other SMPs in exhibiting exceptional properties such as strain storage of 1000 %, cold storage, cold programmability, and mechanical and thermal triggers tunable both during and after programming. They also possess energy storage capability and mechanical properties. SMNR decreases the melting temperature of crystals up to 40 K by the application of external mechanical stress. The mechanical stress parallel to stretching direction stabilizes the crystals indicated by linear increase in the trigger temperature, whereas the stress applied in the transverse direction weakens the shape-stabilizing crystals and resulted in decrease in the trigger temperature after programming (stress-induced melting). The reinforcing effect of parallel stress on the shape-stabilizing crystals is called stress-induced stabilization. A significant amount of crystals still exist in the stretched SMNR above the trigger temperature as evident from the XRD measurements. Above 100 °C temperature, this completely disappeared. Fukahori proposed the existence of partially continuous cross-linked phase and a continuous uncross-linked phase in vulcanized NR by which the two different melting regimes of the crystals were explained [134]. The strain-induced crystallization appears mainly in the cross-linked phase followed by thermal crystallization in the uncross-linked phase. The crystals formed in the cross-linked phase should be more sensitive to parallel stress than the crystals formed in the uncross-linked phase. This should then result in the two melting regimes.

The shape memory property of two series of samples, one based on vulcanized NRL (degree of cross-linking: 0.2 %, abbreviated by L) and the other prepared by vulcanizing compacted rubber (degree of cross-linking: 0.1 %, abbreviated by NR), was analyzed at different stretching temperatures. Figure 14 shows the resulting trigger temperatures obtained from TMA measurements, which show the perpendicular strain increase during recovery for L 1–3. The average trigger temperatures (T trig) obtained for the samples L1, L2, and L3, respectively, were 25.8, 30.5, and 38.4 °C.

Fig. 14

a Determination of unstressed trigger temperatures of samples L 1–3 using TMA measurements. Arrows indicate different trigger temperatures (T trig, L1 258 °C; L2 305 °C; L3 384 °C, in all cases, the perpendicular strain increase was measured). Determination of trigger temperatures of b L1, c L2, and d L3 in dependence on different applied parallel stresses (in all cases, the parallel strain decrease was measured) [135]. Copyright 2013. Reprinted with permission from John Wiley & Sons

Figure 14b–d, respectively, illustrates the influence of constant stress applied parallel to prior programming direction on the samples L1, L2, and L3 [135]. All plots of parallel strain against temperature show a rising trigger temperature with increasing stress. This applied stress was significantly smaller than the stored stress, and as a result, the sample retracts to some extent upon increasing the temperature. This retraction continues until either the samples reach the limit of the apparatus or the samples rupture. The retraction process of L1 and L2 under different loads is triggered much more discrete than that of L3, which can be due to the stress-free trigger temperature ranges of the unstressed samples. The vulcanized samples NR 1–3 also follow similar results. The trigger temperature strongly depends on the stress of the surrounding amorphous phase on the strain-stabilizing crystals [136], and the parallel stress causes a stabilizing effect on the polymer crystals, resulting in an increase of the trigger temperature.

Figure 15a, b shows the variation of nominal stress with strain during the shape memory cycle of latex-based (Neo) and SMR-based samples [137]. The values of stored energy were obtained in a range from 1.34 to 4.88 J g⁻¹ with efficiencies (ratio of stored to stretching energy) between 26.22 and 48.32 %. It is also found that the stored energies and efficiencies increase with increasing degree of cross-linking for both SMNR sources. The stored energy also depends on applied strain during programming.

Fig. 15

Exemplary plots of nominal stress versus strain during a shape memory cycle of a latex-based and b SMR-based samples, plot of force versus crosshead travel of sample Neo 040 (dimensions: 178 mm × 382 mm × 11 mm) during a shape memory cycle [137]. Copyright 2013. Reprinted with permission from John Wiley & Sons

The SMNR showed fixity ratios of up to 94 % and recovery ratios of up to 100 % combined with high stored strains in a range of 540–1180 % after 10 thermomechanical cycles. Energies of up to 4.88 Jg⁻¹ could be stored with up to 53.30 % efficiency. This material also has excellent Young’s modulus (two orders of magnitude upon triggering or programming). All these properties are observed in the case of latex-based samples as well (Fig. 15b); however, the masticated samples show superior shape memory and energy storage properties. The complete shape memory cycle of the latex-based sample Neo 0.40 is given in Fig. 15c for detailed information. The integral of the programming path 1 corresponds to the energy W stretch needed for stretching the rubber at 50 °C to a strain of 750 %. The elastically stored energy W during shape recovery at 50 °C is also equal to integral of the recovery path 5 [138].

The shape memory of NR in the presence of solvent vapors is illustrated in Fig. 16. For this, SMNR is kept in contact with toluene vapors and thus immediately triggered, resulting in complete shape recovery [139]. The programming steps showing the stretching (1) and fixation steps (2) are clear from the stress−strain plots of SMNR in Fig. 16a. Triggering the sample under unconstrained conditions resulted in strain answer (3), whereas triggering under constrained conditions resulted in stress answer (4). The solvent caused complete recovery, and so the qualitative and quantitative information on the trigger capability of the vapors is lost. The same effect was also observed in the case of other solvents such as chloroform, n-heptane, and tetrahydrofuran (THF). By mounting the programmed SMNR between a fixed clamp at the bottom and a spring scale, pulling force is determined as shown in Fig. 16b.

Fig. 16

a Stress−strain plots of the programming process of SMNR, b photographs illustrating the stress built up by exposure to toluene vapor temporal progress of a stress answer σ of constrained SMNR with b changing vapor pressure pi plotted, and c schematic illustration of the irreversible and reversible stress answers caused by the solvent vapor pressure during the first (Δσirr) and all following cycles (Δσrev) [139]. Copyright 2013. Reprinted with permission from American Chemical Society

In order to investigate the effect of solvent on triggering, the constrained SMNR sample was subjected to a force of about 1.7 N within 15 min, and thereafter, the unclamped sample was brought to atmospheric pressure following vapor evacuation. SMNR was not fully triggered as it did not recover its original shape, but stabilized 800 % strain. Cyclic experiment in which the stress is continuously measured by adding and removing toluene vapor pressure was further performed at constant time intervals to confirm partial triggering or triggering and reprogramming after solvent removal. As shown in Fig. 16c, the sample stress becomes 3.6 MPa upon increasing the toluene vapor pressure to 2500 Pa in the first cycle. When the vapor was evacuated below 300 Pa after 15 min, the sample stress decreased to 1.8 MPa. This two-way stress response is due to reprogramming of the SMNR sample. The results were reproducible. By comparing Fig. 16c and d, the direct dependence of stress answer on the vapor pressure increase is derived. The stress answer did not return to its initial value even after 24-h evacuation at which the stress was 1.7 MPa. Thus, the solvent-sensitive stress answer of SMNR on toluene vapor is composed of an irreversible stress answer Δσ _irr that remains after the first cycle and a reversible stress answer Δσ _rev, as represented in Fig. 16e [140].

2.2.7 Actuators

Electroactive polymers (EAPs) convert electrical energy into mechanical energy and widely applied in actuators, sensors, motors, pumps, etc. [141]. Such dielectric and electrostrictive materials possess advantages of rapid response rates depending upon ionic diffusion, dipole moment, electrostatic forces, low operating voltage, high force generation capabilities, and high energy density per cycle, and thus, they are useful in fabricating catheters, automotive door locks, and artificial muscles for robotic and prosthetic devices and in bioengineering. Typically, the uses of actuators include adaptive optic applications to move an optical element, to connect thermal or electric pathways, and to provide agonist and antagonist controls in a linear motion, etc.

As NR composites are flexible and easily fabricable, they are also EAP. The application of an external electric stimulus to EAP induces dipole–dipole moments in molecules and thus induces electrostatic interactions between them. Higher force is required to deform such materials, which indicates that the mechanical properties of the material have been altered [141]. Niamlang et al. synthesized poly(p-phenylene vinylene) (PPV)/NR blend (10 %w/w PPV) in the presence of a cross-linker. The addition of a conducting polymer to the elastomer enhances the dipole moment of the conductive polymer and thus the interference of the dipole–dipole interactions between the cross-linking points. When external electric field strength is applied, the NR gel deflects toward anode, the functional group and strand are polarized, and ionic interaction between generated dipole moment and electrode is occurred. If the strength of the applied field is higher, stronger interaction occurs between dipole moment and electrode and causes higher degree of bending [141].

Blending NR with polythiophenes is also practiced by Puvanatvattana et al. by which they analyzed the electrorheological response of the samples to make it applicable in actuators. Electric field strength of 0–2 kV/mm was applied, and the corresponding storage and loss modulus values were noted. For PI of cross-linking ratios 2, 3, 5, and 7, the storage modulus sensitivity increased with applied field strength and attained maximum values of 10, 60, 25, and 30 %, respectively, at 2 kV/mm. For the PI/polythiophene blend, containing 5, 10, 20, and 30 vol.%, the moduli were higher than those of cross-linked PI. The storage modulus sensitivity of such blends also increased with field strength and reached maximum values of 50, 35, 110, and 45 %, respectively, at 2 kV/mim [142].

Orthorhombic and noncentrosymmetric perovskite lead zirconate (PbZrO₃) capable of spontaneous polarization was synthesized at a temperature below the Curie temperature and dispersed in NR. At a volume fraction of 0.007306, the electrical conductivity of the composite increased dramatically by two orders of magnitude at 500 kHz electrical frequency. The obtained composite sample showed variations in storage modulus with applied electrical field. When the field strength was 1 kV/mm, the temporal characteristic of the neat NR increased from 9.4 kPa to a steady-state value of 9.6 kPa, whereas for the composite, the increase was abrupt (from 9.4 to 12.6 kPa). This value decreased, but not went to the original value when the field was turned off. This behavior indicates that there are some irreversible interactions, possibly due to the dipole bonding between small lead zirconate particles at this volume concentration and some residual dipole moments remaining [143].

Based on the deflection responses of a standard Thai rubber (STR) composite of Al₂O₃ cross-linked with dicumyl peroxide, Tangboriboon et al. proposed a three-stage mechanism for the electromechanical cycle of the actuator. This is shown in Fig. 17a where both electrodes have the same concentration of dopant (Al³⁺) when the cantilever is undistorted, and the electrochemical transfer of the Al³⁺ between electrodes causes bending either to the right or to the left. The STR/alumina samples were cured using sulfur as well. They have checked the effect of cross-link agents between the sulfur/accelerator and the DCP/stabilizer on the deflection response of the composites under applied electrical field strength of 0–650 V/mm at room temperature. In the presence of applied field, the freely suspended end of the dielectric elastomer actuator bends toward one of the electrodes due to the interaction between the electrode and the nonsymmetrical dipole moments set up in the matrix. The maximum bending is observed for the DCP cured sample containing 60 phr alumina at 200 V/mm in which the dipole moment generated causes a nonsymmetrical negative charge distribution and as a result it bends toward anode. In addition, the alumina particles induce polarization within the composite and the dipole moments under applied electric field which results in high dielectric constant for the composites with maximum alumina content [144]. Also, the DCP curing induces better actuation effect when compared to sulfur curing [145] as per the results shown in Fig. 17b (all samples contain DCP at various percentages).

Fig. 17

a Three states during the electromechanical cycle of a dielectric elastomer actuator and b degree of bending of cross-linked natural rubber at various electric field strengths [145]. Copyright 2013. Reprinted with permission from Elsevier

3 Conclusions

Nanotechnology offers technological advantages in various important areas, such as production, processing, storage, transportation, safety, and security. As presented in this review, the products based on NR derived from nanotechnology are applicable in electronics, aerospace science, valuable household items, value-added products, and industry. NR is an excellent polymer and obtained from nature, which has shown its own importance in the literature for long years. Nanofillers are capable of imparting superior properties such as gas permeability (impermeable tyres), electrical (conductive wires, sensors) and thermal conductivity, dielectric properties (capacitors), mechanical strength (cables, sheets, etc.), flame retardancy, shape memory, and microwave absorption properties to NR. A comprehensive as well as state-of-the-art review on NR and its composite applications has been made in this paper. Materials for piezoelectrics and low-friction lightweight devices with high durability can get at low cost. Major advantages of nanocomposites in comparison with their micro counterparts are addressed in most parts of this study. However, the fabrication of nanocomposites is often associated with filler–polymer compatibility, interfacial interaction, nanolevel dispersion, etc. Because of this, practical applications of NR composites are rather limited so far. Therefore, further developments to achieve nanofiller-reinforced composites with relatively high volume fractions and with maximum filler–polymer and filler–filler interactions are still required. For this, acquisition of fundamental knowledge and development of analytical models and experimental tools are also needed. In short, this review briefly discusses applications and summarizes the needs on fundamental research as well as the associated challenges of NR and NR composites.