Thermal Properties of Carbon Nanofiber Sheet for Thermal Interface Materials under High Temperature and Humidity

Jiangling Xiong; Tomoo Kinoshita; Yongbum Choi; Kazuhiro Matsugi; Yuuji Hisazato; Nobuto Fujiwara

doi:10.2320/matertrans.MT-M2022136

Abstract

In this study, polyvinyl alcohol (PVA) - polytetrafluoroethylene (PTFE) - vapor grown carbon fiber (VGCF) sheets are fabricated as a new paper-like thermal interface material (TIM), which is a potential substitute for traditional TIMs. Two types of PVA-PTFE-VGCF sheets were fabricated by using 120 nm and 300 nm PTFE particles. The microstructure shows that the VGCFs were arranged in random directions inside the sheet and interconnected via the aggregate of PVA. 300 nm PTFE particles were well distributed within the sheet, while 120 nm PTFE particles aggregated partially and formed pores nearby. The fabricated sheets have a low thickness of 31.2 um and 28.2 um, and lightweight properties with a density of 0.79 × 10⁶ g·m⁻³ and 0.91 × 10⁶ g·m⁻³, respectively. With the addition of 300 nm PTFE particles, the fabricated sheet has higher thermal conductivities of 9.81 W·m⁻¹·k⁻¹ in the in-plane direction and 2.11 W·m⁻¹·k⁻¹ in the through-plane direction. In the high temperature and humidity test, the thermal conductivities of the fabricated sheet were increased due to the rearrangement of PVA and PTFE particles.

1. Introduction

The thermal interface materials (TIMs) are used to fill the air gaps between a heat-generation device and a heat-dissipation thrust for heat distribution. With the rapid development of the electronics industry, the rising higher power density in electronics presents an increasing requirement for heat dissipation, because electronics are becoming intelligent, miniaturized, and integrated. For example, the latest high-end CPU - Intel a Core i9-12900K can consume 125∼241 W power in a small size of 45 mm × 37.5 mm. Its temperature can hit 373 K easily with core-heavy workload, which puts the CPU at risk of shortening and destroying the life of internal components. Using flexible TIMs with high thermal conductivity (TC) can effectively remove the heat from heat-generating electronics. Thus, the development of TIMs become extremely significant for the heat management of advanced electronic products. Consisting of polymer with the film-forming ability and reinforcing fillers with high TC, polymer based-paper-like TIMs have been widely researched because of their superior performance effectively maintaining the efficiency, reliability, and longevity of electronic devices.¹⁾ In recent years, the development of paper-like TIMs has been accelerated through the use of carbon materials such as graphene and carbon nanotube (CNT). These materials have been widely applied as fillers in TIMs because of their high TC and excellent mechanical properties. Despite having high in-plane TC, these paper-like TIMs are hard to widely used in industry due to their low through-plane TC.²^–⁷⁾ Moreover, many studies lack a facile and scalable preparation strategy for the industrial application of paper-like TIMs, and the long-term stability of TIMs is seldom be considered for their practical application.⁸^–¹¹⁾ As another outstanding carbon nanomaterial, vapor grown carbon fiber (VGCF) is discontinuous, highly graphitic, and highly compatible with most polymer processing techniques, and it can be dispersed in an isotropic or anisotropic mode. VGCF also has excellent mechanical properties, such as high TC, high specific strength and corrosion resistance, also it is a commercially available carbon material with a low price. Moreover, VGCF can be imparted to a wide range of matrices including thermoplastics and thermosets to form thin and flexible sheets.

In our previous work, we reported an efficient and simple method to fabricate polyvinyl alcohol (PVA)-based VGCF sheet with a mass ratio of VGCF:PVA = 1:2.6, which high TC of 6.44 W·m⁻¹·K⁻¹ in the through-plane direction and 14.30 W·m⁻¹·K⁻¹ in the in-plane direction.¹²⁾ The PVA-based VGCF sheet has a lightweight property with a density of 1.03 × 10⁶ g·m⁻³ and is as soft as commercial soft lining materials with a Shore hardness of 82.2. However, the thermal conductivities were significantly reduced after the high temperature and humidity test, due to the hydrophilic property of PVA. Polytetrafluoroethylene (PTFE) consisting of carbon and fluorine, has excellent hydrophobicity, good dispersibility, and high chemical stability,¹³^,¹⁴⁾ which is considered to be an excellent filler to enhance the hydrophobicity of composite material for practical application. Moreover, dispersed nanometric scale reinforcement significantly increases the tortuosity of the path for the permeation of water molecules.¹⁵⁾ However, it should be noted that excessive addition of PTFE powders would significantly reduce the thermal conductivities of the fabricated sheet, due to the low TC of PTFE (∼0.3 W·m⁻¹·K⁻¹). Considering all the above, it is believed that the hydrophobicity of the PVA-VGCF sheet will be enhanced by adding PTFE nanopowders, while keeping the mass ratio of VGCF:resin (PVA and PTFE) the same as that of previous PVA-VGCF sheet.

This study has focused on understanding the full influence of the use of an optimized mass ratio of PTFE (VGCF:PVA:PTFE = 1:1.3:1.3) as a filler in the VGCF sheet, for guaranteeing an improvement in hydrophilicity. All the experimental results of the VGCF:PVA:PTFE sheet were compared with the VGCF sheet of the previous report to evaluate the influence of added PTFE powders on their properties.

2. Experimental Methods

The VGCFs were purchased from SHOWA DENKO K.K. and have an excellent thermal conductivity of 1200 W·m⁻¹·k⁻¹ and a length of 10–20 µm. Commercial PVA (Yamato Co., Ltd.) with a concentration of 13 mass% was used as a binder, which is shown in Fig. 1(a). Two types of PTFE nanopowder: L173JE (120 nm) and TF9207Z (300 nm) were used to reinforce the hydrophobicity of the VGCF sheet, which were supplied by 3M Japan Ltd., and AGC Chemical Company, respectively. The field emission scanning electron microscope (FE-SEM, S-5200, HITACHI, Japan) images of the as-received VGCF and PTFE nanopowders are shown in Fig. 1(b)∼1(c). It should be emphasized that although the VGCFs were readily aggregated due to van der Walls force,¹⁶⁾ they were utilized under the as-received condition. The aggregation of VGCFs can contribute to providing a 3D net-like structure during the preparation process, which is expected to improve the isotropic thermal properties of the fabricated sheet. The preparation process of the PVA-PTFE-VGCF sheet is shown in Fig. 2. Based on solution mixing, a simple but industrial method is adopted to fabricate the VGCF sheet. To make PTFE powders dispersed evenly in a fabricated sheet, PTFE powders were mixed PVA binder by an agitator at 100 rpm for 1 hour. After that, the VGCFs were poured into the PVA-PTFE mixture and mixed with a silicon bar at 20 rpm for 1 minute, the PVA-PTFE-VGCF mixture was spread on a PTFE film of A4 paper size. At last, after drying this at room temperature (RT) for 48 hours, the PVA-PTFE-VGCF sheet was separated from the PTFE film. In this study, two types of PVA-PTFE-VGCF sheets: PVA - PTFE L173JE - VGCF sheet (PPLV) and PVA - PTFE TF9207Z - VGCF sheet (PPTV) were fabricated by using different PTFE nanopowder of L173JE (120 nm) and TF9207Z (300 nm). The fabrication condition is listed in Table 1. The dispersion of PTFE powders in the PVA matrix and the morphology of fabricated sheets were checked by FE-SEM. The porosity of each VGCF sheet was calculated by using the image analysis software Image-Pro Plus 6.0, which used five random areas with a size of 1.28 × 10⁻³ mm² from each VGCF sheet. The thickness of fabricated sheets was measured by using a digital micrometer (MonotaRo Co., Ltd.). Density was calculated from the mass and volume of each fabricated VGCF sheet. According to the ASTM D2240 procedure,¹⁷⁾ the fabricated VGCF sheets with a total thickness of 6 mm was used for the Shore hardness test by using a Shore hardness durometer. The thermal diffusivity (α) of fabricated sheets at RT was evaluated using a periodic heating and infrared radiation thermometer method (Thermowave Analyzer TA). Specific heat capacity (Cp) was measured by using DSC (X-DSC 7000, HITACHI, Japan). After getting all these values, TC was calculated with the equation:

\begin{equation} \lambda = \alpha \times \rho \times C_{p} \end{equation}

(1)

where α, ρ and C_p represent the thermal diffusivity (× 10⁻⁶·m²·s⁻¹), density (× 10⁶ g·m⁻³) and Specific heat capacity (J·g⁻¹·K⁻¹), respectively. To evaluate the long-term performance of the PVA-PTFE-VGCF sheet, the high temperature, and humidity test was performed at a temperature of 358 K and relative humidity (RH) of 85% for 500 hours. The infrared spectra of the fabricated VGCF sheet were measured by using a Fourier transform infrared spectrometer (FT-IR) in the range of 4000 cm⁻¹ to 600 cm⁻¹.

Fig. 1

SEM image of (a) PVA solution, (b) VGCFs, (c) PTFE L173JE 120 nm powders, and (d) PTFE TF9207Z 300 nm powders.

Fig. 2

Schematic of the fabrication procedure of PVA-PTFE-VGCF sheet.

Table 1 Fabrication conditions of PVA-PTFE-VGCF sheets.

3. Results and Discussion

3.1 Microstructures of VGCF sheets

Figure 3 shows the SEM images and microstructure schematics of referred PVA-VGCF sheets and fabricated PVA-PTFE-VGCF sheets. All of the sheets were shown that VGCFs are arranged in random directions and connected with each other via PVA aggregates. Due to the aggregation of as-received VGCFs, a 3D net-like structure was built by VGCFs, which is considered to contribute good stability and high through-plane TC to fabricated sheets. On the other hand, as shown in Figs. 3(a) and 3(d), intersections of a 3D net-like structure also caused PVA to be hard to inside during the mixing process and form many pores in referred PVA-VGCF sheet, leading to a significant reduction of TC. In Figs. 3(b) and 3(c), it is indicated that these pores can be efficiently filled with the addition of PTFE powders. However, in the fabricated PPLV and PPTV sheet, partial aggregations of PTFE particles were observed due to the intermolecular force of PTFE particles. Especially, some large collections of 120 nm PTFE particles and unfilled pores were observed in Fig. 3(b). This poor dispersion of 120 nm PTFE powder will adversely affect the efficiency of heat-releasing in the fabricated sheet. On the contrary, from Fig. 3(c) it is clearly seen that due to the smaller collections, 300 nm PTFE particles is still well distributed and compatible within the fabricated PPTV sheet respectively, and the pores between VGCFs have almost filled by 300 nm PTFE particles. As shown in Figs. 3(e) and 3(f), the microstructures schematics show the different dispersibility of 120 nm and 300 nm PTFE powders caused different microstructure in the fabricated sheet. The smaller PTFE particles were more likely to aggregate and formed uneven surfaces, while 300 nm PTFE particles can be dispersed uniformly between VGCFs relatively and filled the pores efficiently.

Fig. 3

SEM images (× 6000) and microstructure schematics of fabricated VGCF sheets by different mass ratio of VGCF:PVA:PTFE, (a) (d) 1:2.6:0^*, (b) (e) 1:1.3:1.3 with a 120 nm PTFE particle size and (c) (f) 1:1.3:1.3 with a 300 nm PTFE particle size.

3.2 Porosity, thickness, density and hardness

To evaluate the possibility of a fabricated VGCF sheet for practical application for TIM, the porosity, thickness, density, and Shore hardness was measured. As shown in Fig. 4, the porosity of referred PVA-VGCF sheet was as high as 18%. After the addition of PTFE powders, although the concentration of PVA was decreased and more pores should be generated theoretically,¹²⁾ the porosity of the PPLV sheet and PPTV sheet were decreased to 9% and 6% respectively, due to the pores were filled by PTFE powder. Caused by aggregation of 120 nm PTFE particle, the PPLV sheet has a higher porosity than the PPTV sheet.

Fig. 4

The porosity and thickness of fabricated VGCF sheets.

As the same, the average thickness of the VGCF sheet was also decreased from 47 um for the PVA-VGCF sheet to 31.2 um and 28.2 um for PPLV and PPTV sheets respectively, due to the lower content of the PVA matrix. The aggregates of 120 nm PTFE particles are considered to be much larger than those of 300 nm PTFE particles and cannot be filled into the pores of the fabricated VGCF sheets. The aggregates of 120 nm PTFE particles resulted in surface unevenness and increased thickness of fabricated sheet. The thickness of TIM has a great influence on the absolute thermal resistance across the TIM. From Fourier’s Law for heat conduction, the absolute thermal resistance R_θ of TIM is calculated by using the equation:¹⁸⁾

\begin{equation} R_{\theta } = \frac{t}{A \times \lambda} \end{equation}

(2)

where t is the thickness of TIM, A is the cross-sectional area perpendicular to the path of heat flow, and λ is the thermal conductivity of TIM. Therefore, the absolute thermal resistance of a TIM will decrease by using thinner material. The lower absolute thermal resistance means this material is a better thermal conductor in practical applications. In addition, the TIMs with lesser thickness such as PPLV and PPTV sheets have greater potential in the application in advanced precision instruments, which have become miniaturized and complex in recent years. The density of PPLV and PPTV sheet were calculated using mass and volume, as shown in Fig. 5 the values are 0.79 × 10⁶ g·m⁻³ and 0.91 × 10⁶ g·m⁻³ respectively. Compared with the PVA-VGCF sheet, PPLV and PPTV sheets have a lower density, which showed a greater lightweight property of PPLA and PPTV sheets compared to commercially available TIMs (1.44–3.6 × 10⁶ g·m⁻³).¹⁹⁾ Hardness is an important mechanical property of TIM. According to the ASTM D2240 procedure, PPLV and PPTV sheets were superimposed onto the same thickness as 6 mm to measure their Shore hardness. The results are also presented in Fig. 5. The average Shore hardness of the PVA-VGCF sheet, PPLV, and PPTV sheet were 82.2, 86.0, and 85.4, respectively. The continuous structure composed of VGCFs allows fabricated sheet to take advantage of the high elasticity of the PVA matrix to produce great deformation and low hardness. Therefore, it is indicated that PPLV and PPTV have great softness and flexibility for practical application as TIM.

Fig. 5

The density and hardness of fabricated VGCF sheets.

3.3 Thermal conductivity

Figures 6 and 7 show the effect of the addition of PTFE powder on the in-plane TC, through-plane TC, and the thermal diffusivity of the VGCF sheet at room temperature. As shown in Fig. 6, after the addition of the PTFE powders, the in-plane TC of fabricated VGCF sheet has a reduction of 46% and 31%, which decreased from 14.30 W·m⁻¹·k⁻¹ to 7.66 W·m⁻¹·k⁻¹ of PPLV sheet and 9.81 W·m⁻¹·k⁻¹ of PPTV sheet, respectively. As shown in Fig. 7, the through-plane TC also have a reduction of 72% and 67% after the addition of PTFE powders, which decreased from 6.44 W·m⁻¹·k⁻¹ to 1.79 W·m⁻¹·k⁻¹ and 2.11 W·m⁻¹·k⁻¹, respectively. The thermal diffusivities of in-plane and through-plane directions have a similar reduction trend with TC.

Fig. 6

The in-plane TC and thermal diffusivity of fabricated VGCF sheet at room temperature.

Fig. 7

The through-plane TC and thermal diffusivity of fabricated VGCF sheet at room temperature.

For VGCF, PVA and PTFE, thermal conduction is achieved mainly by the vibration of atoms and molecules near their equilibrium position.²⁰⁾ VGCF has a high thermal conductivity. On the other hand, due to the low crystallinity of polymer, the PVA and PTFE have very low thermal conductivities. Thus, the thermal conduction in PVA-PTFE-VGCF sheet mainly depends on the high thermal conductivity via VGCF networks. The thermal conductive mechanisms in this study include thermal conductive path theory and surface microstructure of fabricated sheet. Through the thermal conductive path theory,²¹^,²²⁾ the thermal paths are formed by the contact of VGCFs in the fabricated sheet. The heat flux mainly transfers along the 3D network of VGCF with lower thermal resistance. The PVA and PTFE increased the thermal interface resistance and formed thermal barrier to the thermal paths. This is the first reason for the reduction of TC with the addition of PTFE powders. The surface microstructure is another reason for the reduction of TC, which was shown in Fig. 8. Without the filling of PTFE powders, the surface of the PVA-VGCF sheet has a surface wave, which is caused by the bulge of VGCFs.¹²⁾ With the addition of PTFE powders, this wave can be filled and reformed into a relatively smooth surface, consisting mostly of PTFE particles and a small part of bulging VGCFs. When the front surface of the PVA-VGCF sheet was heated by a light pulse, the heat can reach the rear face quickly through the heat paths as shown in Fig. 8(a), which is indicated that the PVA-VGCF should have a good thermal diffusivity and TC. As shown in Fig. 8(b), due to the low TC of PTFE, adding PTFE powder has formed a thermal barrier, and the heat can only reach the rear face through the small part of bulging VGCFs, which leads to the lower TC after addition of PTFE powders. Table 2 lists the previously reported thermal conductivity of other paper-like thermally conductive composites. As shown in this table, the through-plane thermal conductivity of PPLV and PPTV sheet in this study is significantly higher than those of previous studies. The in-plane thermal conductivity of the fabricated sheet still maintains a higher level among all these paper-like thermally conductive composites.

Fig. 8

The schematic of microstructure near the surface of PVA-VGCF sheet and PVA-PTFE-VGCF sheet.

Table 2 Comparison of thermal conductivity between this work and other paper-like composites.

3.4 High temperature and humidity test

To evaluate the long-term property, the PPLV and PPTV sheet was cut into a specimen with a size of 20 × 20 mm for the high temperature and humidity test. Each specimen was held at 358 K and 85% RH for 500 hours. After the high temperature and humidity test, thein-plane TC and through-plane TC of each specimen were also measured. As shown in Fig. 9, prior to testing, the in-plane TC of the PVA-VGCF sheet, PPLV sheet, and PPTV sheet were 14.30, 7.66, and 9.81 W·m⁻¹·k⁻¹, respectively; After testing, the TC of PVA-VGCF sheet was significantly reduced by 31% to 9.85 W·m⁻¹·k⁻¹. However, the TC of PPLV and PPTV sheets were increased to 9.02 and 10.81 W·m⁻¹·k⁻¹, with an increase of 18% and 10%. Similarly, as shown in Fig. 10, the through-plane TC of the PVA-VGCF sheet was decreased by 57% from 6.44 W·m⁻¹·k⁻¹ to 2.71 W·m⁻¹·k⁻¹ after the high temperature and humidity test. The TC of PPLV and PPTV sheets were increased from 1.79 and 2.11 W·m⁻¹·k⁻¹ to 2.21 and 2.49 W·m⁻¹·k⁻¹, with an increase of 23% and 18% respectively. To find out the reason for the improvement of TC in the PPLV and PPTV sheet, the FT-IR analysis was performed on the surface of the PPLV and PPTV sheet. The FT-IR result was shown in Fig. 11. There were two infrared IR detection bands were marked in this figure, 1211 cm⁻¹ to 1154 cm⁻¹ representing the entire C–F stretching of PTFE and a narrow band at 1734 cm⁻¹ to 1717 cm⁻¹ for the C=O stretching of PVA. Before the high temperature and humidity test, both PVA and PTFE were detected on the surface of the fabricated VGCF sheet. However, the peak of C=O stretching disappeared after the high temperature and humidity test, which means the PVA vanished from the surface of the VGCF sheet. It is considered that due to the strong hydrophilic property of PVA, the water molecules in a high temperature and humidity test environment would be absorbed into the VGCF sheet by PVA. The absorbed water in PVA not only disrupts the hydrogen bonding but also contributes more free volume and lubrication.²⁵^,²⁶⁾ Therefore, the PVA segments in the VGCF sheet become readily mobile and rapidly respond to the load change, which may lead to a movement of PVA from surface to center part of the VGCF sheet and caused the disappearance of C=O stretching in FT-IR analysis. As a result, a large number of pores would be generated on the surface of the VGCF sheet which caused the reduction in TC of the PVA-VGCF sheet.¹²⁾ Different from the PVA-VGCF sheet, the pores generated by the movement of PVA in the PPLV and PPTV sheet can be efficiently filled by PTFE nanoparticles. Thus, the PPLV and PPTV sheets could keep a uniform surface and a high density relatively after the high temperature and humidity test. A TIM with a uniform surface tends to have a lower absolute thermal resistance, and higher density contributes to increasing the TC of TIM according to the calculation equation of TC. Therefore, the TC of PPLV and PPTV increased after the high temperature and humidity test. This phenomenon provides an innovative idea for improving the stability and long-term performance of TIM in practical applications. The properties of each VGCF sheet were measured, and a comparison with referred PVA-VGCF sheet was shown in Fig. 12.

Fig. 9

The in-plane TC of fabricated VGCF sheet after high temperature and humidity test.

Fig. 10

The through-plane TC of fabricated VGCF sheet after high temperature and humidity test.

Fig. 11

The FT-IR analysis of fabricated VGCF sheets.

Fig. 12

The properties comparison of PVA-VGCF, PPLV and PPTV sheets.

4. Conclusions

In all, novel PVA-PTFE-VGCF thermal interface materials were successfully developed by a simple and cost-effective process. Two types of VGCF sheets were fabricated by using 120 nm and 300 nm PTFE particles, the microstructure observation revealed that the VGCFs were arranged in random directions and interconnected via PVA aggregates. 300 nm PTFE particles were well distributed and compatible within the PPTV sheet, while partial aggregations of 120 nm PTFE particles and unfilled pores were observed in the PPLV sheet. As the pores on the surface were filled by PTFE powders, the porosity of the fabricated sheet was reduced. The average thickness of the fabricated sheets was also decreased from 47 um for the PVA-VGCF sheet to 31.2 um and 28.2 um for PPLV and PPTV sheets respectively, due to the lower content of the PVA matrix. With the addition of 300 nm PTFE particles, the fabricated sheet has higher thermal conductivities of 9.81 W·m⁻¹·k⁻¹ in the in-plane direction and 2.11 W·m⁻¹·k⁻¹ in the through-plane direction. The in-plane and through-plane TC of PVA-PTFE-VGCF sheets were decreased by 31%∼46% and 67%∼72% compared with referred PVA-VGCF sheet, due to the thermal barrier of PTFE particles. In the high temperature and humidity test under the condition of 358 K and 85% RH, the in-plane and through-plane TC of PVA-PTFE-VGCF sheets were increased by 10%∼18% and 18%∼23%. According to the FT-IR analysis result, it is considered that the increase in TC was caused by the hydrophobic property of PVA and rearrangement of PVA and PTFE particles.

Acknowledgments

This work is financially supported by Toshiba Infrastructure System & Solution Corporation. The authors would like to thank Toshiba Infrastructure System & Solution Corporation for all its support in this research.

REFERENCES

Corresponding author

Version information

Register with J-STAGE for free!