Thermally conductive tough flexible elastomers as composite of slide-ring materials and surface modified boron nitride particles via plasma in solution

Strong demand for the development of portable devices and flexible devices has led to increased research into compatible materials with both flexibility and functionality.^1–6 In particular, thermally conductive materials with good flexibility and mechanical strength are in demand as thermal interlayer materials and heat radiation sheets for flexible electric circuits. Such materials are composite materials of flexible rubber and inorganic materials having high thermal conductivity (also known as fillers).^3–6 As thermally conductive inorganic materials, carbon-based materials, alumina, silica, zinc oxide, or boron nitride (BN) are often used for their high thermal conductivities.^3–6 It is generally necessary to add inorganic materials to realize a high thermal conductivity (>1 W/mK).^6,7 However, adding large amounts of inorganic materials caused these composite materials to become more brittle.^6,7

We have used two approaches with these materials in order to overcome such embrittlement while maintaining high thermal conductivity. One is to apply a slide-ring (SR) material as the rubber, while the other is to improve the affinity between the rubber and inorganic materials via surface modification of the inorganic materials by plasma in aqueous solutions. The SR material is a type of polymer network with movable crosslinking points and tends to have very unique mechanical properties.⁸ In composite materials of SR, the flexibility is maintained^9,10 and high toughness is present¹¹ even with a large amount of inorganic materials (∼10 wt. %), due to the unique structure present. However, the previous literature features a maximum volume fraction of inorganic materials in SR at 17 vol. % for such applications.^9–11 

Regarding surface modifications, the modification of inorganic particles by plasma in aqueous solutions can introduce functional groups such as -OH, -COOH, and -NH₂ on the particle surfaces without increasing the number density of defects.^12–14 One example of mechanical strength improvements via surface modification involves composite materials of carbon nanotubes (CNTs) and Nylon 6.¹³ However, there is no report on the effect of plasma-surface modification on composite materials with SR. In this letter, we apply polycaprolactone-grafted polyrotaxane (PCL-g-PR) to SR and BN as inorganic materials. The composite materials, which have good mechanical properties even with a large amount of BN [up to ∼70 wt. % (56 vol. %)], are developed using plasma surface modifications with BN.

Raw SR and SR/BN composites were synthesized with a PCL-g-PR (SH2400P, Advanced Soft Materials). PCL-g-PR consists of poly-ε-caprolactone (PCL)-grafted hydroxypropylated α-cyclodextrins (CD) and a polyethylene glycol (PEG) backbone chain.⁸ The BN particles have hexagonal crystal structures and a mean diameter of 0.2 μm (Sigma-Aldrich, Ltd.). BN particles were used as purchased or after surface modification with plasma in solution as described below.

PCL-g-PR (0.5 g) was dissolved in 10 ml of toluene. For the composites, the modified or unmodified BN particles were first dispersed in separate 10-ml vials of toluene using an ultrasonic bath (5 min), followed by mixing with the PCL-g-PR-dissolved toluene. Dibutyltin dilaurate (25 μl) and hexamethylene diisocyanate (86 μl) were added to the solution as the catalyst and cross-linker, respectively. The solution was then interfused for 30 min in a mixer (AR-100, Thinky, Inc.), followed by pouring into a perfluoroalkoxy fluorine resin petri dish (inner diameter: 7.5 cm). This was kept at room temperature for at least 8 h to form the film composite, which was then annealed in vacuum at 105 °C for 2 h to obtain a test sample approximately 0.1–0.4 mm in thickness.

For plasma surface modification of the BN particles, plasma was generated in a 100-ml glass beaker containing 70 ml of a solution of NaCl (0.1 g/l) and BN particles (28.6 g/l), dispersed by a magnetic stirrer. The electrical conductivity before plasma generation was 250 μS/cm at room temperature. The electrodes were two parallel tungsten wires 1 mm in diameter and separated by 3.8 mm, which were shielded by 0.5 mm-thick ceramic tubes except for a 1 mm segment at the tip of each electrode. A bipolar pulsed waveform with an amplitude of 1.5 kV, a frequency of 80 kHz, and a pulse width of 0.75 μs was used (power supply; Kurita MPP-HV04–300 kHz) and the modification time was 1 h. To prevent the solution from overheating, the beaker was kept in an ice bath during modification, and the highest temperature the solution was allowed to reach was approximately 70 °C. After the plasma modification, the electrical conductivity of the solution at room temperature increased to 420 μS/cm. The modified BN particles were washed with ethanol and dried for 8 h at 80 °C and used to synthesize the composite materials within 24 h of their creation.

The mechanical properties of raw SR and SR/BN composite elastomers were characterized using a uniaxial tensile tester (Shimadzu, Inc., EZ-S). Uniaxial stress-extension ratio curves were measured at a constant strain rate of 0.03/s at room temperature. Rectangular-shaped specimens were used (width 3 mm, length 40 mm) to calculate the Young's modulus and stress-extension ratio curve. The tensile strength and extension ratio at break were evaluated using dumbbell-shaped specimens (width 2 mm, length 35 mm, JIS K6251–7) to prevent breaking at or near the grip part of the tensile tester.

Figure 1(a) shows stress-extension ratio curves of raw SR and SR/BN composite elastomers for various amounts of BN with rectangular-shaped specimens (the terms stress and extension ratio in this letter refer to engineering stress and engineering extension ratio, respectively). Figure 1(b) shows the reduced stress f* = σ/(λ − λ⁻²) converted from the stress-extension ratio curves [Fig. 1(a)], where λ is the extension ratio of the sample and σ is the tensile stress. In the case of ideal rubber elasticity described by the Neo-Hookean model, σ is proportional to λ − λ⁻² and thus f* is independent of λ.¹⁵ The SR/BN composite elastomers with 0 and 20 wt. % BN contents exhibit almost constant reduced stress, like ideal rubber. The reduced stress at 50 wt. % BN content decreases as λ increases. Although the 50 wt. % SR/BN composite deviates from the ideal rubber characteristics, it seems to show a general rubber-like stress strain curve.¹⁵ Figures 1(a) and 1(b) also suggest that the extension ratio at break seems to strongly depend on the surface conditions of particles and should be evaluated with dumbbell-shaped specimens as shown later. Figure 1(c) shows the Young's modulus, which is the tilt at an extension ratio of 1, as functions of BN with or without plasma surface modification of BN, showing that the Young's modulus depends primarily on the BN contents.

Figure 2(a) shows the tensile strength and extension ratio at break for composites with various modified or unmodified BN contents obtained using dumbbell-shaped specimens. All plasma-modified composites show higher tensile strength and elongation at break than unmodified ones. For example, for an extension ratio at break with 50 wt. % BN, the composite with modified BN has 2.5 times the extension ratio at break as the one with unmodified BN. Furthermore, the tensile strength is also approximately 1.7 times higher. Although it is well known that the tensile strength increases as the Young's modulus increases^16–19 and that the extension ratio at break increases as the Young's modulus decreases,^18,20 this study shows that both tensile strength and the extension ratio at break increase as the Young's modulus remains constant [Figs. 1(c) and 2(a)]. Generally, when at least 40 wt. % BN is added to a composite, its tensile strength decreases regardless of the size and shape of the BN due to poor interfacial interactions between the BN surface and polymer,^6,7 but the composites in this study do not show a clear decrease in strength. Figure 2(b) shows the area under the stress-extension ratio curves, which describes the toughness or tensile toughness and indicates the energy required to break the sample.²¹ Plasma-modified composites show 1.8–4.9 times greater area than the case of unmodified BN, suggesting that the toughness is also improved by plasma modification of BN. It may be noteworthy that the improvement of mechanical properties seen in this study is remarkably greater than the effects of plasma modification in the previous literature.^13,14

Figure 3 plots thermal conductivity against BN contents for the raw SR and the SR/BN composite elastomers at room temperature. The thermal conductivity was estimated from its thermal diffusivity, specific heat, and density. The thermal diffusivity was measured by a laser flash method (Advance Riko, Inc., TC-7000H), the specific heat of the composite was measured by differential scanning calorimetry (DSC; Hitachi, Inc., DSC7000X), and the density was measured by gas pycnometry (Micromeritics Instrument Corp., AccuPyc II 1340). The thermal conductivity values closely match other studies of the randomly oriented BN/polymer composite.^7,24–27

The broken line in Fig. 3 describes the theoretical value determined using the Maxwell model²² as $2km+kp−2V(km−kp)2km+kp+V(km−kp)km$⁠, where k_p and k_m are the thermal conductivities of BN [400 W/mK (Ref. 23)] and SR (0.2 W/mK) and V is the volume fraction of BN. The experimental data, both with or without plasma modification, clearly depart from the model between 30 and 50 wt. % BN, where percolation conductivity is likely to be significant. Notably, while the thermal conductivity of the plasma-modified BN composite is lower at low BN contents (≤30 wt. %) than in the unmodified material, at high BN contents (>50 wt. %), the plasma-modified one shows a higher conductivity. We speculate that this inversion is related to dispersion of BN in the composites. At low BN contents, unmodified BN particles are likely more agglomerated, resulting in a higher thermal conductivity than the plasma-modified one as these agglomerated particles can facilitate an easy thermal path, although the number of paths is small. On the other hand, at high BN contents, plasma-modified BN particles (which are likely to be well-dispersed in the composite) produce many fine thermal paths, resulting in higher thermal conductivity than the unmodified composite.

Finally, to compare the thermal properties and Young's modulus of the composite with various other materials, thermal conductivities of the SR/BN composites against Young's modulus were added on the Ashby plot of other general materials,²⁸ as well as the rubber/BN composites in the previous literature [Fig. 4(a)].^7,28,29 The SR/BN composites clearly depart from the trend of the general materials. The nearby composites are silicone rubber/BN composites created by Kemaloglu.⁷ The composites consist of silicone rubber and various BN particle sizes and contents.

Figures 4(b)–4(d) show comparisons of the thermal conductivity, Young's modulus, and tensile strength between our SR/BN composites with or without plasma modification of BN and the silicone rubber/BN composite from the literature.⁷ In the silicon rubber/BN composites, the tensile strength decreases as the BN contents increase, thereby increasing their thermal conductivity. On the other hand, the tensile strength of the unmodified SR/BN composite remains nearly constant (∼8 MPa) as the thermal conductivity and BN content increase. Furthermore, the tensile strength of the SR/BN composite with plasma-surface modifications also shows a nearly constant value but at much higher strength (∼12 MPa). These figures clearly show the benefits of applying SR and plasma-surface modification of BN to improve the mechanical properties.

When determining the mechanisms by which plasma surface modification of BN in solution improves the mechanical properties of these composites, several hypotheses are viable. Plasma modification may have improved the dispersion of BN particles in the SR matrix. Hieda et al. reported that the dispersibility of carbon nanoparticles in Nylon 6 was improved by plasma modification in aqueous solution, on the basis that the light transmittance of plasma-modified carbon nanoparticles and Nylon 6 composites decreased as compared to the unmodified version.¹⁴ This improvement in the dispersion likely comes from a change of functional groups on the surface^12–14 and/or the deposited charges on the inorganic materials.

Another way to improve the mechanical properties is with different cross-linking structures, such as the way in which plasma surface modification adds OH groups on the BN surface. OH groups are attached by plasma modification in solution, such as in the case of CNT.^12–14 OH groups on BN should increase cross-linking between BN and PCL grafted to CD (sliding-cross links), resulting in different network structures. These different network structures may result in easier stress relaxation with sliding of the cross-links. Minato et al. reported that in comparing the elastomer with sliding-cross links and the elastomer with fixed-cross links (which has the same Young's modulus as the elastomer with sliding-cross links), the former shows a higher extension ratio at break and tensile strength than the latter;⁸ this tendency is similar to our results. Furthermore, the remarkably larger improvement of mechanical properties seen here with SR, compared to the results in the literature without SR structures,^13,14 may support the presence of special mechanisms within SR structures.

In conclusion, we synthesized thermally conductive SR/BN composites with up to 70 wt. % BN particles. By conducting surface modification with BN via plasma in solution, we achieved improved mechanical and thermal properties in the resulting SR/BN composites. Furthermore, the tensile strengths of SR/unmodified or plasma-modified BN composites do not decrease with increasing BN contents, in contrast to more general tendencies in the literature. This study clearly demonstrates that plasma surface modification of inorganic materials is a promising route to develop functional composite materials with slide-ring materials while retaining their unique mechanical properties.

The authors would like to thank Dr. K. Takebayashi for technical support with the specific heat measurement. This work was financially supported in part by Grants-in-Aid for Scientific Research (B) (Grant No. 16H04506) from the Japan Society for the Promotion of Science.