Actuator Performance of a Hydrogenated Carboxylated Acrylonitrile-Butadiene Rubber/Silica-Coated BaTiO3 Dielectric Elastomer.

We synthesized silica-coated barium titanate (BaTiO3) particles with different silica shell thicknesses and evaluated the effect of silica coating on the relative dielectric properties of silica-coated BaTiO3 particles. Furthermore, composite elastomers were prepared using hydrogenated carboxylated acrylonitrile-butadiene rubber (HXNBR) with a high relative dielectric constant (εr) and silica-coated BaTiO3 particles, and their performance as an actuator was evaluated. Both εr and relative dielectric loss of non-coated BaTiO3 particles increased at low frequencies (<200 Hz) associated with ionic conduction. However, εr and relative dielectric loss were reduced for the silica-coated BaTiO3 particles with thick silica shells, indicating that silica coating reduced ion migration. The dielectric breakdown strength increased with the thickness of the silica shell; it increased up to 80 V/μm for HXNBR/silica-coated BaTiO3 particles with 20 nm-thick silica shells. The maximum generated stress, strain, and output energy density of the composite elastomer with HXNBR (with a high relative constant) and silica-coated BaTiO3 were 1.0 MPa, 7.7%, and 19.4 kJ/m3, respectively. In contrast, the values of the same parameters for a reference elastomer (acrylic/BaTiO3; with low εr) were 0.4 MPa, 6.7%, and 6.8 kJ/m3 at the dielectric breakdown strength of 70 V/μm. The results indicated that the elastomers composed of HXNBR and silica-coated BaTiO3 exhibited higher generated stress, strain, and output energy density than elastomers for conventional dielectric actuators.

Introduction

A dielectric elastomer actuator (DEA) consisting of a laminated structure of a thin elastomer film and a compliant electrode is a simpler system than conventional hard actuators. When voltage is applied to a DEA, the electrodes coated on both sides of the elastomer are pulled together. Therefore, the elastomer contracts along the thickness direction and extends parallel to the surface direction.[1−10] The muscles of a living body (including human muscles) are stretchable and have high mechanical properties. The DEA is referred as an artificial muscle because it exhibits the abovementioned properties of the muscles of a living body. The generated stress of a DEA is expressed by the effective compressive stress equation (eq )[6]where σ, εr, ε0, V, and d are the stress, relative dielectric constant, dielectric constant of vacuum, driving voltage, and thickness of the film, respectively.

To increase the generated stress, it is necessary to increase the relative dielectric constant (εr) or apply a high electric field (V/μm). Previous reports have confirmed that the εr of the composites can be increased by adding high-dielectric constant particles.[11−13] From eq , theoretically, high generated stress can be obtained with the addition of high-dielectric constant particles. However, in practice, dielectric breakdown occurs at a low electric field before obtaining high generated stress. This destruction is due to both intrinsic and extrinsic (process-related) factors. As an intrinsic factor, the increase in the dielectric constant of an elastomer causes an increase in its conductivity. Extrinsic factors such as partial agglomeration of particles and presence of air voids and interfaces between the polymer matrix and fillers are introduced while processing. These factors increase the leakage current and decrease the dielectric breakdown strength of the composites. Thus, dielectric breakdown is more likely to occur in the low electric field before the generated stress increases by the amount of square of the electric field.

Ceramic composites comprising high-dielectric constant particles that have low dielectric breakdown strength reportedly exhibit low energy storage properties.[14−16] The effect of fillers on the electrical energy density of polymer composites has been studied.[17−22] One of the reasons for the decreased dielectric field strength is the inhomogeneous electric field generated at the interface because of the difference between the dielectric constants of the filler and the polymer matrix. To resolve this, silica-coated barium titanate (BaTiO3) particles have been reportedly used for reducing the inhomogeneous electric field.[23] Wang et al. reported that coating SiO2 layers on the surface of BaTiO3 nanoparticles resulted in homogeneous dispersion on the polyvinylidene fluoride polymer matrix with high polarity; this significantly reduced the energy loss of the nanocomposites.[24,25] However, a common elastomer film for the DEA is an acryl or silicone elastomer with a low dielectric constant of ≤4.8.[26] To enhance the dielectric constant of this elastomer, addition of a large amount of particles with a high dielectric constant is necessary. Actuators with high strain and generative force are required for the application. Flexibility is important for the DEA to generate strain; however, the elastomer hardens as the number of particles increases. An elastomer having flexibility and a high dielectric constant is needed. In our previous research, we reported that HXNBR exhibited a high dielectric constant (14) and flexibility.[27] Additionally, Poikelispää et al. reported the high actuation performance of NBR rubbers.[2] This suggested that a DEA composed of HXNBR (having a high dielectric constant) and silica-coated BaTiO3 particles can be applied to a high electric field, leading to high generated stress and strain.

In this study, we synthesized silica-coated BaTiO3 particles with different silica shell thicknesses and evaluated their effect on the εr and relative dielectric loss of the silica-coated BaTiO3 particles compressed into pellets. Furthermore, HXNBR/silica-coated BaTiO3 composite elastomers were prepared, and their actuator performance was evaluated.

Results and Discussion

Characterization of Silica-Coated BaTiO3 Particles

Table summarizes the preparation conditions of the silica-coated BaTiO3 particles. Four types of BaTiO3 particles with different tetraethoxysilane (TEOS) amounts were prepared. Figure shows the transmission electron microscopy (TEM) images of the silica-coated BaTiO3 particles. A silica shell was observed on the surface of the BaTiO3 particles in all samples. As the amount of TEOS increased, the thickness of the silica shell increased. In BT100-TEOS-1.5, the silica shell thickness was ∼20 nm. Figure shows the Fourier transform infrared (FT-IR) spectra of the non-coated and silica-coated BaTiO3 particles. In the spectrum of BT100, a C=O peak, derived from barium carbonate impurities, was observed at 1448 cm–1. However, the peak weakened, as the amount of TEOS and silica shell thickness increased. Instead, a broad absorption appeared at 1000–1260 cm–1, characterized by the Si–O bonds from silica.[30−32] In particular, a sharp peak corresponding to the Si–O bond asymmetric vibrations appeared at 1060 cm–1 in proportion to the amount of TEOS. Furthermore, BT100-TEOS1.0 and BT100-TEOS1.5 exhibited similar FT-IR spectra. As a result of thermogravimetric analysis, a weight loss of about 0.7 wt % of adsorbed water was observed around 100 °C. Overall, the weight loss was less than about 2%. (Figure S3 in the Supporting Information). The silica shell was stable at high temperatures.

Table 1

Preparation Conditions of Silica-Coated BaTiO3

code	TEOS (g)a	weight ration of TEOS/BT100
BT100-TEOS-0.1	0.2	0.1
BT100-TEOS-0.5	1.0	0.5
BT100-TEOS-1.0	2.0	1.0
BT100-TEOS-1.5	3.0	1.5

For BT100 2 g.

Figure 1

TEM images of silica-coated BaTiO3 particles (magnification of 100k): (a) BT100-TEOS-0.1, (b) BT100-TEOS-0.5, (c) BT100-TEOS-1.0, and (d) BT100-TEOS-1.5.

Figure 2

FT-IR spectra of non-coated and silica-coated BaTiO3 particles.

Effect of Silica Coating on Relative Dielectric Properties of BaTiO3 Particles

Figure shows the frequency dependence of the εr and relative dielectric loss of the non-coated and silica-coated BaTiO3 particles compressed into pellets. The BT100 particles indicated high εr and high dielectric loss. Both values increased in the low-frequency region because of the interfacial polarization associated with the mobile charges between particles.[33−35] Thus, there is ion migration occurring from the BT100 particles. However, both εr and dielectric loss of the silica-coated BT100 decreased, as the amount of silica coating increased. In the case of BT100-TEOS-0.5, a slight increase in the dielectric constant and dielectric loss was observed in the low-frequency range; whereas the εr curves for BT100-TEOS-1.0 and BT100-TEOS-1.5 indicated no significant frequency dependence and were almost identical. It is considered to be the result of suppressing the migration of ions derived from BT100 by silica coating with a thickness of 20 nm.

Figure 3

Frequency dependence of the relative dielectric constant (a) and relative dielectric loss (b) of non-coated and silica-coated BaTiO3 particles compressed into pellets.

Properties and Actuator Performance of the HXNBR/Silica-Coated BaTiO3 Composite Elastomer

Table summarizes the εr, dielectric breakdown strength, and Young’s modulus of the dielectric elastomers. In the case of HXNBR/BT100 and BT100-TEOS-0.5, the average εr of the composite elastomers at 100 Hz was 15.2–17.1. However, the dielectric constant gradually decreased to 13.8 and 11.3 for BT100-TEOS-1.0 and BT100-TEOS-1.5 respectively, with the increased amounts of silica coating (Figure S5). Although silica coating on BaTiO3 particles has a disadvantage of lowering the dielectric constant, the dielectric constants of the HXNBR/BaTiO3 composite elastomers (11.3–16.3) were greater than those of the acrylic elastomer p(AN-BA)/BT100 (10.4). Young’s modulus of the HXNBR/silica-coated BaTiO3 composite elastomers was 6.7–8.7, and the flexibility was maintained. In the comparative sample of p(AN-BA)/BT100, Young’s modulus was low (2.7 MPa); however, εr was lower (10.4) than that of HXNBR.

Table 2

Relative Dielectric Constant, Dielectric Breakdown Strength, and Young’s Modulus of Dielectric Elastomers

code	relative dielectric constant at 100 Hz	dielectric breakdown strength (V/μm)	Young’s modulus (MPa)
HXNBR/BT100	16.3 ± 2.31	53.3 ± 12.5	7.1
HXNBR/BT100-TEOS-0.1	15.2 ± 0.41	46.7 ± 4.7	7.4
HXNBR/BT100-TEOS-0.5	17.1 ± 1.28	46.6 ± 9.4	7.5
HXNBR/BT100-TEOS-1.0	13.8 ± 0.47	70.0 ± 14.1	8.7
HXNBR/BT100-TEOS-1.5	11.3 ± 0.14	83.3 ± 4.7	6.7
P(AN-BA)/BT100	10.4 ± 0.61	76.6 ± 4.7	2.7

The actuator performance of the HXNBR/BaTiO3 composite elastomers was evaluated. Figure shows the generated stress of the dielectric elastomers as a function of the applied voltage and electric field. As per eq , the generated stress of the elastomers is proportional to the electric field and square of the applied voltage. As shown in Figure a, elastomers with thicker silica shells tend to generate higher stress. Figure b shows the electric field data excluding the effect of film thickness variations for HXNBR/BT100, HXNBR/BT100-TEOS-0.1, and HXNBR/BT100-TEOS-0.5 with non-coated particles or elastomers with thin silica shells and high εr and confirms that the elastomers exhibit low dielectric breakdown strength. The dielectric breakdown strength decreases because of the migration of impurity ions induced by high εr at low frequencies, as shown in Figure . The generated stress in a low electric field increased in the following order, BT100-TEOS-0.5 > BT100-TEOS-1.0 > BT100-TEOS-1.5, corresponding to the magnitude of εr. Although the generated stress of HXNBR/BT100-TEOS-1.5 with the thickest silica shell at a low electric field was lower than that of HXNBR/BT100-TEOS-0.5, its dielectric breakdown strength and maximum stress increased up to 80 V/μm and 1.0 MPa, respectively. The acrylic elastomer (reference sample) with low εr exhibited a low generated stress of 0.4 MPa at 70 V/μm. Thus, the elastomers composed of HXNBR with high εr and silica-coated BaTiO3 particles were able to generate a higher stress than the acrylic elastomers for conventional DEAs.

Figure 4

Generated stress of the dielectric elastomers as a function of the applied voltage (a) and electric field (b).

Figure shows the generated strain of the dielectric elastomers as a function of the applied voltage and electric field. As shown in Figure a, the HXNBR/silica-coated BaTiO3 elastomers exhibited a high generated strain with increasing silica film thickness. Nevertheless, the acrylic elastomer also showed a similar generated strain because Young’s modulus was low. Figure b indicates the highest elongation (4.5%) at a low electric field (40 V/μm) for HXNBR/BT100-TEOS-0.5; however, dielectric breakdown occurs at a low electric field. Although HXNBR/BT100-TEOS-1.5 exhibited lesser elongation at a low electric field than HXNBR/BT100-TEOS-0.5, it showed the highest generated strain of 7.7% at 80 V/μm. The generated strain of the acrylic elastomer was 6.7% at 70 V/μm. Hence, the acrylic elastomer and HXNBR showed comparable performances for strain generation.

Figure 5

Generated strain of the dielectric elastomers as a function of the applied voltage (a) and electric field (b).

Comparison of the Effective Compressive Stress of Dielectric Elastomers

From the obtained values of εr and dielectric breakdown strength, the compressive stress of the elastomers was estimated and compared. Table summarizes the effective compressive stress of various dielectric elastomers. The sample code provides the notation as mentioned in the respective references. Because the data were recorded for different elastomers and composites under different conditions, they are not quantitative but can be used for qualitative comparison. The comparison of the samples with similar prestrain indicated that the compressive stress of HXNBR/BT100-TEOS-1.5 (0.69 MPa) was higher than that of the composites with HS3-silicone and CF19 2186-silicone (0.3–0.6 MPa) and the VHB 4910 acrylic elastomer (0.13 MPa). For a prestrain of 300%, VHB 4910 exhibited an extremely high dielectric breakdown strength of 412 V/μm and produced a high compressive stress of 7.2 MPa. However, the prestrain of a few hundred percent is limited to the elastomers with high insulation properties and high elongation. In contrast, the compressive stress of the composites with barium titanate was not that high because of the low dielectric breakdown strength. The compressive stresses of the BaTiO3 composites with PDMS, SR (slide ring), and silicone were 0.004, 0.013, and 0.10 MPa, respectively. For the composites with BaTiO3, the compressive stress of NBR/BT-PCPA-KH570 was as high as 0.89 MPa. The BaTiO3 surface was covered by the codeposited poly(catechol/polyamine) (PCPA) and grafted with γ-methacryloxypropyl trimethoxy silane (KH570). The high dielectric breakdown strength that produced the high compressive stress is thought to be the effect of surface modification and silica coating.

Table 3

Comparison of the Effective Compressive Stress of Dielectric Elastomers

codea	effective compressive stress (MPa)b	dielectric constant (1 kHz)c	dielectric breakdown strengh (V/μm)c	references
HXNBR/BT100-TEOS-1.5 (25,0)	0.69	11.3 (100 Hz)	83.3	this work
P(AN-BA)/BT100 (25,0)	0.54	10.4 (100 Hz)	76.6	this work
HS3 silicone (14,14)	0.3	2.8	110	(6)
CF19-2186 silicone (15,15)	0.6	2.8	160	(6)
VHB 4910 acrylic (15,15)	0.13	4.8	55	(6)
VHB 4910 acrylic (300,300)	7.2	4.8	412	(6)
PDMS-BaTiO3 41 wt % (0,0)	0.004	9.0	6.84	(36)
SR/m-BT10 wt % (0,0)	0.013	10.48	12	(13)
NBR/BT-PCPA-KH570 50phr (0,0)	0.83	16.65 (100 Hz)	75	(37)
silicone/BT 30phr (0,0)	0.10	3.85 (100 Hz)	55	(38)

Prestrain (x,y)%.

Estimated from eq .

Estimated from graphical data in the cited reference, when no tabulated value was provided.

Output Energy Density of HXNBR Composites

We evaluated the overall potential of the actuator including the generated stress and strain using the output energy density value. Figure shows plots of the output energy density of the dielectric elastomers as a function of the electric field. The maximum output energy density of the acrylic elastomers with high generated strain was 6.8 kJ/m3 at 70 V/μm. In the HXNBR/BT100-TEOS elastomer, both the maximum generated stress and strain increased with the silica shell thickness. The maximum output energy densities of HXNBR/BT100-TEOS-0.5, HXNBR/BT100-TEOS-1.0, and HXNBR/BT100-TEOS-1.5 were 4.3 kJ/m3 at 40 V/μm, 10.1 kJ/m3 at 80 V/μm, and 19.4 kJ/m3 at 80 V/μm, respectively. Moreover, HXNBR/BT100-TEOS-1.5 showed the output energy densities of 5.7 kJ/m3 at 60 V/μm and 10.7 kJ/m3 at 70 V/μm. The results suggested that HXNBR/BT100-TEOS-1.5 produced higher stress and strain over 60 V/μm than the other samples. The silica coating reduces the electric field generated at the inhomogeneous interfaces, thus reducing the energy loss. Therefore, as the electric field increases, the electric field energy can be output without energy loss. Although silica coating on BaTiO3 particles had a disadvantage of lowering the dielectric constant, it increased the insulation of the composite and produced high output energy density.

Figure 6

Output energy density of the dielectric elastomers as a function of the electric field.

Conclusions

We synthesized silica-coated BaTiO3 particles with different silica shell thicknesses and evaluated the effect of silica coating on the relative dielectric properties of the silica-coated BaTiO3 particles compressed into pellets. The silica shell could be coated up to a thickness of 20 nm by increasing the amount of TEOS against BaTiO3 particles. The εr and dielectric loss of the BaTiO3 particles increased at low frequencies because of ion migration; however, this migration was suppressed for the BT100-TEOS-1.0 and BT100-TEOS-1.5 because of the coating with thick silica shells. HXNBR with high εr/silica-coated BaTiO3 particle composite elastomers was prepared and the performance of the DEAs was evaluated. HXNBR/BaTiO3, HXNBR/BT100-TEOS-0.1, and HXNBR/BT100-TEOS-0.5 with thin silica shells showed low dielectric breakdown strength. The dielectric breakdown strength increased with the thickness of the silica shell. The dielectric breakdown strength of HXNBR/BT100-TEOS-1.5 with the thickest silica shell increased up to 80 V/μm. Because of the high dielectric breakdown strength, the elastomers composed of HXNBR with a high relative constant and silica-coated BaTiO3 particles with low dielectric loss exhibited higher stress, strain, and output energy than the acrylic elastomers for conventional dielectric actuators. These elastomers are capable of improving the power performance of soft actuators.

Experimental Section

Materials

TEOS and aqueous ammonia solution (28%) were purchased from Tokyo Chemical Industry Co., Ltd. (Japan). Acetylacetone (AcAc) and ethanol were purchased from FUJIFILM Wako Pure Chemical Corporation (Japan). BaTiO3 particles (BT100:BT-HP9DX with a diameter of 100 nm) were obtained from KCM Corporation (Japan). Moreover, HXNBR (Therban XT 8889; CN 33.8 mol %, COOH 3.8 mol %) was obtained from LANXESS (USA). Tetrakis (2-ethylhexoxy) titanium (TOT, crosslinking agent) (95%) with 2-propanol (5%) was purchased from Nippon Soda Co., Ltd. (Japan). A poly(ethylene terephthalate) substrate (PET3811, Lintec Corporation, Japan) and a PET film (Diafoil, Mitsubishi Chemical Corporation, Japan) were used as releasable and finger contact prevention films for the electrode, respectively.

Synthesis of Silica-Coated BaTiO3 Particles

Silica coating on BaTiO3 particles was carried out using a conventional method.[28] A total of 2.0 g of BaTiO3 particles was added to 20 mL of ethanol, and ultrasonic dispersion was performed for 5 min. Predetermined amounts of TEOS, ethanol (85 mL), water (45 mL), and aqueous ammonia solution (30 mL) were added to the solution. After stirring at room temperature for 6 h, the mixture was centrifuged at 3000 rpm for 20 min. The supernatant was then discarded, and ethanol was added. Washing, dispersion, and centrifugation were repeated three times. Purified water was added, and the mixture was heated under reflux at 80 °C and then filtered using a 0.2 μm-pore size filter. The particles were collected by drying under reduced pressure at 110 °C for 2 h.

Preparation of the HXNBR/Silica-Coated BaTiO3 Composite Elastomer

Bead mill dispersion was performed to obtain a high-dispersion solution of the particles. Briefly, BaTiO3 particles (0.64 g) were added to AcAc (1.94 g) to prepare a 25 wt % solution. After the solution was transferred to a zirconia container, BaTiO3 particles were dispersed using a bead mill filled with 1.28 g of zirconia beads. Dispersion was carried out twice at 2000 rpm for 2 min. The rotation speed and the number of trials are the optimum conditions in which the silica shell is not destroyed, as shown in the TEM image of the particles after dispersion (Figure S1). The zirconia beads were removed using a stainless-steel mesh.

HXNBR was dissolved in AcAc to prepare a 12 wt % polymer solution. Next, BaTiO3 particle-dispersed solution (2.56 g) and AcAc (0.20 g) solution containing 20 wt % TOT were added to the polymer (6.67 g) solution. After sonication for 5 min, the mixture was applied to the surface of a poly(ethylene terephthalate) substrate using a bar coater. The film was then dried by heating at 150 °C for 1 h. The amounts of HXNBR, silica-coated BaTiO3 particles, and TOT were adjusted to achieve 100:80:5 weight ratio. The volume ratio of BaTiO3 particles (12.8 vol %) was calculated using a specific gravity of 6.1 g/cm3.

The poly(acrylonitrile-butyl acrylate) [p(AN-BA)] copolymer was synthesized by radical polymerization as a reference sample. The properties are summarized in Table S1. AN was added because the εr of poly BA was too low. The dielectric constant of p(AN-BA) was ∼8.0. A p(AN-BA)/BT100 composite elastomer was prepared using the abovementioned process.

Sample Preparation for Measurement of Relative Dielectric Properties of Silica-Coated BaTiO3 Particles

Details of the sample preparation for the measurement of εr and relative dielectric loss are shown in Figure S2 of the Supporting Information. Generally, the dielectric properties of sintered BaTiO3 are evaluated. However, in this study, because the BaTiO3 particles were dispersed in the polymer without being sintered, the dielectric properties of the BaTiO3 particles compressed into pellets were evaluated.

Sample Preparation for Measurement of Relative Dielectric Properties of the Composite Elastomer

Details of the sample preparation for the measurement of εr of the composite film and the frequency dependence of εr of the elastomers are shown in Figures S4 and S5 of the Supporting Information, respectively. The film thickness of the samples ranged between 14 and 26 μm.

Sample Preparation for Measurement of Generated Stress and Strain of the Actuator

We measured the force generated from the elastomer along the displacement direction. Details of the sample preparation for the measurement of generated stress and strain are described in the Supporting Information. Figure S6 shows a photograph of the equipment used to evaluate the generated stress; Figure S7 shows the stress relaxation under voltage application. The prestretching was set to 25%, and the change in stress before and after the voltage impression was measured using the load cell after a 100 s hold. The DC applied voltage was programmed every 10 V/μm from the initial film thickness, and then, the measurement of the generated stress was performed by increasing the applied voltage every 10 V/μm for 10 s until the dielectric film was broken. The generated stress was measured at two points for each electric field. The generated stress was evaluated from the difference in stress before and after applying voltage. The generated strain at the corresponding electric field was evaluated from the following equation

Evaluation of the Output Energy Density

The modified formula for calculating the output energy density per volume (Eout) from the obtained generated stress (σ) and strain (ε) is as follows[29]where V is the volume of an elastomer and F is the generative force. The output is a quarter of the energy density.

Characterization

Herein, TEM was conducted on a JEM-2100F (JEOL, Japan) at an accelerating voltage of 200 kV. Fourier transform infrared (FT-IR) spectroscopy was performed on a FT/IR-4200 type A with a ATR PRO450-S (JASCO Corporation, Japan). The data were acquired at a resolution of 4 cm–1. The εr and relative dielectric loss at room temperature (chamber temperature, 23.7 °C) were measured using an LCR METER E4980AL (Keysight Technologies, Inc., USA) at an applied voltage of 1 Vrms. The generated stress was measured using a load cell attached to a tensile tester (EZ-S 100N, Shimadzu Corporation, Japan) with a DC high-voltage power supply (DMI-8K 4P, MAX-ELECTRONICS Co. Ltd., Japan). Ultrasonication was performed using a BRANSON 5510 (Branson Ultrasonics, Emerson Japan Co., Ltd.). Bead mill dispersion was performed using an NP-100 (THINKY Corporation, Japan). The chamber temperature was set to −20 °C.