Effect of ketjenblack and barium titanate on the piezoresistive behaviour of silicone rubber particulate composites

Flexible electronics have surged the demand of flexible materials for applications such as tactile sensing [1, 2], wearable electronics [3, 4], robot skins, health care [5], structural health monitoring [6], wearable devices for physiological monitoring [7]. Flexible pressure sensing is one of the most important components of flexible electronics applications. These pressure sensors are based on piezoelectric [8], piezocapacitive [9–13] piezoresistive [14] sensing mechanisms. Piezoresistive (PR) sensors have been researched due to their simple structure and low cost in addition to simpler read out electronics [15–17].

For PR pressure sensors, the mechanism involves change in resistance (ΔR) of the sensing material with mechanical deformation on account of pressure. Two types of PR pressure sensors are evident from the literature: (1) Negative type piezoresistive pressure sensor (NPS), whose resistance decreases with increase in pressure and (2) Positive type piezoresistive pressure sensor (PPS), whose resistance increases with an increase in pressure [16]. Sensitivity is one of the essential parameters relating to the output signal for pressure sensors. The sensitivity (S) based on the PR effect is defined in literature as in equation (1).

$$\begin{eqnarray}&&S=\displaystyle \frac{{\rm{\Delta }}R/{R}_{0}}{{\rm{\Delta }}P}\end{eqnarray} \tag{ 1 }$$

Where ΔR is the resistance change with pressure change ΔP and R₀ is the initial resistance at no pressure. Higher sensitivity allows for high-resolution sensing, with a lower limit of detection. Most of the available literature on flexible pressure sensors involve various ways of improving this sensitivity. The piezoresistive sensitivity is on account of piezoresistivity of materials themselves, crack propagation, structural design of sensor, disconnection between overlapped materials and tunneling effect [4].

Two mechanisms of piezoresistive effects have been observed in literature [18]. One in which the conductive fillers have been loaded near the percolation threshold and second, the concentration of the conductive fillers is below the percolation threshold. For composites that are loaded above the percolation threshold, the pressure reduces the inter-filler distance, thus reducing its resistance. In this case the conductive particles get in physical contact with each other. The change in resistance is a function of the concentration of the conductive filler. For composites where conductive fillers are loaded below the percolation threshold and thoroughly wetted by the polymer matrix, the pressure causes the matrix layer to compress. This facilitates electron tunnelling and the reduction in resistance of the composites.

From the literature review two approaches to sensor designs are evident. One is micro structured and other unstructured thin films. Micro structured PR sensors show improved sensitivity and linearity as they depend on the form and size of the microstructures that are imprinted on the flexible substrates [19]. This improvement is on account of the change in the contact area between the electrodes. Photolithography and low-cost molds are utilized for micro structuring, translating into fabrication complexities and higher costs [20]. Flexible piezoresistive pressure sensors based on porous 3D material structure [21, 22], hybrid foam [23, 24] have been investigated. Micro structured sensors are not scalable for large surfaces and hostile environments and the sensitivity improvement comes at the cost of flexibility in terms of allowable strains [25].

The unstructured approach to PR sensor development offers ease of fabrication and low cost at the expense of sensitivity. They can be scaled up for complicated surface contours and harsh environments. Huang et al [26] demonstrated PPS and NPS with carbon nanotubes (CNT) and carbon black (CB)/CNT fillers respectively. NPS with filler volume fraction of 26% was demonstrated by Cai et al [27]. Piezoresistive effects have been shown for cementitious composites with CNT as fillers [6]. PR sensors have been used to detect human joint movements by incorporating carbon nanofibers (CNF) into poly dimethyl siloxane (PDMS) matrix [25]. Magnetite particles with around 50% volume fractions into poly urethane (PU) matrix have been investigated for hydrostatic response [28]. PPS with PANI fibers in PU matrix has shown large strains [29]. CB and graphene nanoplatelets with PDMS matrix were investigated for NPS for pressures up to 1 MPa with filler loading of up to 19% volume fraction [27]. Nickel particles have been investigated along with PDMS for pressure-sensitive behavior [30]. Fe₃O₄ nanospheres have been used as spacers between conductive fillers to increase the PR sensitivity [31].

Piezoresistive behaviors of composites with different carbon based fillers such as carbon black [16, 32–34], Carbon fibers [35, 36], carbon nanotube [26, 37, 38], carbonized cotton fabric [39], graphite [40], graphene [41] have been investigated. Composites using carbon nano particles as fibers, tubes have been investigated due to their higher aspect ratios [26]. However, they show promising results at a high cost of raw material and processing facilities that are also difficult to scale. The conductive fillers are either integrated within the elastomer or embedded on the top surface. When integrated within the elastomer matrix, they offer a robust network of conductive filler [42]. Ketjenblack is a good filler for piezoresistive applications that have been investigated [43–45]. While the wealth of research is concentrated on the use of conductive fillers, there is scant literature on the use of dielectric fillers such as barium titanate for piezoresistive applications.

Matrix materials investigated include paper [46], PI [47], PU, polyisoprene [48], room temperature vulcanised (RTV) PDMS [49, 50], textiles [51], epoxy [52]. High temperature vulcanized (HTV) silicone rubber is a well-known elastomer for outdoor and harsh environments, that has good weathering and aging properties. It also offers simpler and cost effective and scalable processes [53] and is a promising matrix material for these applications [45].

HTV silicone rubber-based pressure sensors are scarcely seen in the literature. Hence, this study is undertaken to investigate the behavior of composites comprising HTV silicone rubber as matrix material with barium titanate as dielectric filler and ketjenblack as conductive filler for piezoresistive applications. These composites offer inexpensive, scalable and simple manufacturing processes to develop flexible PR sensors.

The matrix material utilized is high temperature vulcanized solid silicone rubber (SR), NE-5140, obtained from DJ Silicones, China. The dielectric filler used is barium titanate (BT) obtained from Sigma-Aldrich (208108-500g), India. Superconducting carbon black (SCCB) used is Ketjenblack EC300 J received from AkzoNobel India. Dicumyl peroxide (DCP) of 98% purity, commercially available in powder form (Dicup 98) is used as a curing agent.

The matrix was thoroughly masticated in a roll mill for 15 min. Curing agent and fillers were added to the matrix and mixed for 10 and 30 min respectively. Samples were obtained from this premix through compression molding at 160 and 180 °C temperatures. HTV solid silicone rubber and barium titanate (SSBT) composites were prepared by choosing filler values, curing agent, curing temperature and mixing time as per L₈ Taguchi orthogonal array as shown in table 1.

HTV solid silicone rubber and carbon black (ketjenblack) (SSCB) composites were prepared by choosing the values of the parameters as per L₈ Taguchi orthogonal array as shown in table 2.

HTV solid silicone rubber, barium titanate and carbon black (ketjenblack) (SBTCB) composites were obtained by choosing the values of the filler loadings as per L₉ Taguchi orthogonal array as shown in table 3. HTV silicone rubber was thoroughly masticated in the two high roll mill for 15 min. Curing agent in proportion of 1 parts per hundred rubber (phr) was added to this matrix while being masticated in the roll mill for 10 min. Barium titanate and ketjenblack are added to this mixture in amounts as stated in table 3. After mixing in the roll mill for 30 min, the mixture was obtained as a precure sheet (figure 1). A die and the compression molding machine were used to get the samples at 160 °C, as shown in figure 2.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The resistance across the specimens were measured using Agilent LCR meter (E4980A) using test fixture (Agilent, 16451B) at different pressures as per ASTM D150–11 standard at 1kHz frequency. The pressure was varied from 0 to 20 kPa, that was applied through known weights on the specimen.

In piezoresistive materials, a change in electrical resistance is observed as a result of application of pressure. This change can be quantified by piezoresistive sensitivity. It relates normalized change in resistance with applied pressure.

HTV silicone rubber composites were prepared with filler loadings of 3.5 and 12 parts per hundred rubber (phr) using barium titanate as dielectric filler and Ketjenblack 300J as conductive filler to obtain dielectric (SSBT), conductive (SSCB) and dielectric-conductive (SBTCB) composites respectively.

Taguchi L₈ orthogonal array investigates eight different compositions, with each of the 4 factors set at two levels. Similarly L₉ orthogonal array investigates nine different compositions, with three levels for each of the two factors. As the array is orthogonal, the effect of different levels of factors on the output response can be separated. The composites were prepared as per the Taguchi L₈ orthogonal array, with compositions and levels as shown in tables 1 and 2. These composites were tested for piezoresistive (PR) properties. In order to investigate the effects of filler loading, curing agent, curing temperature and mixing time on the PR sensitivity of the composites, Taguchi analysis of the L₈ orthogonal array is conducted. Ketjenblack-barium titanate filler composites were prepared as per Taguchi L₉ orthogonal array as shown in table 3. The effects of Ketjenblack and barium titanate on these composites are obtained through analysis of Taguchi L₉ orthogonal array and the results are expressed as main effects and interaction plots. These plots provide the variation of mean values of PR sensitivity with different levels of factors.

The piezoresistive sensitivity of the HTV silicone rubber composites are investigated using the Taguchi analysis and reported as main effects and interaction plots. The confirmatory tests were carried out only for those factor combinations that provide the maximum values of the PR sensitivity.

The resistance value of HTV solid silicone rubber with no fillers added was 28 MΩ.

3.1. Piezoresistive properties of SSBT composites

Figure 3 shows the relationship between pressure and normalized resistance obtained experimentally for the SSBT composites. The resistance decreased with the applied pressures, showing a linear trend for all the composites in the tested range. PR sensitivity is obtained from the slopes of the linearly fitted curves of figure 3 and is tabulated in table 1. From the table 1, it is observed that PR sensitivity is maximum for the composite sample of BT of 3.5 phr, curing agent of 1 phr, mixed for 30 min and cured at 180 °C.

Zoom In Zoom Out Reset image size

The reduction in resistance of the composites with pressure is due to the reduction in the distance between the dielectric filler particles. This facilitates the conduction of charges.

HTV silicone rubber shows a steeper fall in normalized resistance for pressure up to 8 kPa. A second region is observed beyond 8 kPa. This behavior of pure solid silicone rubber is altered with the addition of dielectric fillers. A monotonous decrease of normalized resistance with pressure is observed in the entire range tested for all the composites, only the sensitivity varies with different composites.

The main effects plot for piezoresistive sensitivity is shown in figure 4. Sensitivity improves with lower BT filler and amount of curing agent, whereas it improves with increased mixing time and curing temperature. As PR sensitivity depends on normalized resistance change with pressure change, it is dependent on Young's modulus and resistivity of the material. Hence to achieve higher PR sensitivity, materials should possess lower Young's modulus with improved resistivity. However, lower Young's modulus limits the range of pressure sensed. So, a trade-off between sensitivity and pressure range is always sought for flexible pressure sensing applications. From figure 4, with the lowering of dielectric filler, the resistivity improves and Young's modulus reduces as BT is a reinforcing filler. This reflects in the improvement in the PR sensitivity of the composites. With the reduction of curing agent lesser crosslinking among the side chains occurs, thereby lowering the Young's modulus, thus improving the sensitivity. Mixing time is an estimate of the distribution of filler within the matrix. Improved mixing time improves the filler distribution, ensuring better wetting, thus better interfacial adhesion between the filler and the long chains and hence improved sensitivity. Curing temperature is also an important factor in high temperature vulcanization. This factor has a direct bearing on the crosslinking along with the curing agent.

Zoom In Zoom Out Reset image size

From the interaction plot (figure 5), it is evident that interactions among factors exist. Hence the effects of individual factors on the sensitivity depends on the nominal values of other factors.

Zoom In Zoom Out Reset image size

PR sensitivity can be improved by reducing dielectric filler loading as seen from main effects plot. However, from the interactions plot, it can be observed that dielectric filler enhances the sensitivity provided the curing agent loading is at a higher level. Similarly, if the mixing time is reduced, the dielectric filler also enhances the sensitivity.

3.2. Piezoresistive properties of SSCB composites

Resistance across the composite specimens varies with the pressure applied; the variations in resistance values are characteristic of material compositions [54]. Thin polypropylene (PP) films demonstrated a decrease of resistance with increasing pressure [55].

In this study, it is observed that the normalized resistance change decreases with increasing pressure. This indicates that the mechanism of electron transport predominates in these composites. The distance between fillers reduces on account of pressure applied. This leads to reduction in hopping path between fillers. Thus a reduction in resistance of the composites is observed with increasing pressure.

Piezoresistive characteristics of SSCB composites are indicated in figure 6. These composites demonstrate a negative piezoresistive effect.

Zoom In Zoom Out Reset image size

The piezoresistive sensitivity is obtained from figure 6 as slopes of the linear fit of the curves and tabulated in table 2. Piezoresistive sensitivity can be improved by lowering the amount of conductive filler, curing agent and increasing curing temperature, mixing time (figure 7). Interactions among the factors influence the overall effect on the piezoresistive sensitivity (figure 8).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

3.3. Piezoresistive properties of SBTCB composites

Results of piezoresistance tests for SBTCB composites are shown in figure 9. For each of the composites, normalized resistance is plotted as a function of applied pressure. A linear fit is performed for each of the curves for obtaining the value of PR sensitivity. The figure shows that a good linear fit is achieved in the pressure ranges tested, indicating the suitability of materials for pressure sensor applications. The distance between both types of filler particles changes under the influence of pressure, increasing the density of the particles per unit area of composite. The conduction through the composite is mainly determined by the conductive paths formed by the carbon particles, interlaced by BT particles. When the inter-particle distances are small and there is no physical contact between them, the tunnelling conduction mechanism occurs, responsible for reducing resistance [56]. Also, the electrons jump across the gap of carbon particles as described by the tunnelling effect. Due to application of pressure and the resulting compression increases the probability of the electrons to tunnel between adjacent fillers leading to the decrease of resistance of the composite sample.

Zoom In Zoom Out Reset image size

As pressure is applied on the composite, the volume of composite is compressed thereby, filler volume fraction increases. So, the elastic polymer matrix deforms to the extent that filler particles are forced closer together to form conduction paths, reducing resistance. Hence either new conductive paths are formed or the resistance of the effective conductive path decreases, giving rise to a negative piezoresistive effect.

The resistivity of ketjenblack particles is very less as compared to the dielectric filler and solid silicone rubber. So, upon the compression and reduction in the inter particle gap, the tunneling effect occurs [57]. This leads to the formation of conductive path, thereby a reduction in resistance of the composite with increase in pressure.

PR sensitivity is obtained from the slopes of the linearly fitted curves to the figure 9 and tabulated in table 3. Variation in piezoresistive sensitivity of these unstructured composites is from 0.1(10^–3) to 3.7(10^–3) (kPa)⁻¹ which compares well with the sensitivity of 6.4(10⁻³) (kPa)⁻¹ reported for micro structured PR sensors [15]. Maximum sensitivity is achieved for the composite composed of 7.75 phr BT and 3.5 phr SCCB. The improvement in normalized resistance for SBTCB composites is due to the structure of Ketjenblack filler particles that doesn't agglomerate in silicone rubber matrix [53].

The main effects plot (figure 10) shows that sensitivity continuously improves with the decrease in SCCB filler loading. Sensitivity improves with dielectric filler loading for loadings up to 7.75 phr, beyond which it decreases with dielectric filler loading. This is explained on account of interactions among the conductive and dielectric fillers. With the increase of dielectric filler, the resistance decreases at a lesser rate than the resistance decrease for the same amount for conductive filler. Also loading of dielectric fillers interrupt the conductive channels that are formed by SCCB fillers. The change in sensitivity with conductive filler loading compared to dielectric filler is steeper as conductive filler is more reinforcing than dielectric filler [53]. Hence, the conductive filler has a greater influence on the piezoresistivity of the composites compared to dielectric filler for SBTCB composites.

Zoom In Zoom Out Reset image size

The interaction plot for the SBTCB composites (figure 11) shows that a 3.5 phr SCCB filler shows more significant interaction than other loadings, while BT filler loaded at 7.75 phr shows higher interactions. Hence PR sensitivity is a maximum at 3.5 phr SCCB with 7.75 phr BT loading. This interaction is on account of synergistic effect among the ketjenblack and barium titanate fillers.

Zoom In Zoom Out Reset image size

HTV silicone rubber composites using barium titanate, superconducting carbon black (ketjenblack) and barium titanate-ketjenblack fillers were prepared and tested for piezoresistive characteristics. The maximum sensitivity obtained is 3.7(10^–3) (kPa)⁻¹ for these composites, which compare favorably with that available in literature. Barium titanate-ketjenblack composites show a more significant resistance change as compared to barium titanate composites. This suggests the existence of synergistic effect among the ketjenblack and barium titanate fillers. Barium titanate composites show comparable piezoresistive sensitivity to ketjenblack composites. Thus, the HTV silicone rubber composites prepared using established industrial methods offer low cost and mature fabrication route for development of flexible pressure sensors using piezoresistive sensing mechanism.

Authors gratefully acknowledge facilities provided at Microsystem Laboratory, Department of Mechanical Engineering, National Institute of Technology Karnataka (NITK), Surathkal, India.

All data that support the findings of this study are included within the article (and any supplementary files).