A review on advanced carbon-based thermal interface materials for electronic devices

Abstract

Electronic devices play a vital role in our lives and are expected to play an even bigger role in the future considering their immense contribution in every field. Current trends are drifting towards fabricating powerful devices with minimum size, which in turn puts a lot of pressure on the heat dissipation requirement in electronic device packaging, which is very crucial for their performance and life cycle. Presently available thermal interface materials (TIM) fail to fulfil the demand of high heat transfer, opening doors for research on more advanced TIM. Carbon based materials hold great promise in a plethora of applications. Fabricating TIM using carbon-based materials is considered best for efficient heat dissipation between the heat producing device and the heat dissipating device. This review article provides a summary of the state-of-the-art research covering the basics of TIM, the heat transfer mechanism, conventional TIM used and recent graphene-based TIM. It also covers topics related to the characterisation and parameters that should be taken into consideration in the fabrication processes. A systematic understanding in this field is provided through this article to trigger research in overcoming the limitations that persist in fabricating highly efficient TIM for commercial applications.

Introduction

Electronic devices produce heat during operation, and efficient removal of heat is very important for the proper functioning and better life of electronic devices [1]. Hence, thermal management is an important part of electronic packaging which cannot be ignored and needs to be cleverly addressed. Thermal interface materials (TIM) are the substance inserted between two components for enhancing the transport of heat between heat producing component and the heat dissipating component, which is also known as a heat sink [[2], [3], [4]]. The materials used in fabricating TIM should have a higher thermal conductivity value to enhance good heat transport.

The heat removal from heat producing devices will be efficient if there is a proper thermal conduction between the two components. Thermal management systems work to rapidly transfer heat between the heat producing component and the heat sink, bringing a thermal equilibrium between them. The design of a heat sink is made in such a way that it has a high surface area for rapid heat dissipation to the surroundings [5], leading to a reduction in the temperature of the heat producing component hence solving one of the major issues of device failure. Fig. 1 shows the graph of failure factor vs temperature. It can be seen from Fig. 1, that the device failure factor increases exponentially with an increase in temperature. Generally, heat sinks are made from materials having a high heat capacity, like aluminium and copper, in order to absorb more heat with a lower rise in the temperature.

Although there are many articles covering most of the topics related to thermal management, a review paper comprising of fundamentals, mechanisms and the important research conducted focussing on future directions in this field, is still lacking. The authors felt a need to cover these topics starting from the basics up to the advanced research reported. The sub-sections of Section 1 deal with the importance and application of TIM. Mechanisms of heat transfer, factors influencing TIM performance and ideal TIM characteristics are also covered in this sub-section. Section 2 summarises the different types of conventional TIM available commercially. The problems related with the currently available TIM are discussed in Section 3. After that, TIM using carbon-based materials like diamond, carbon fibres and CNT is discussed briefly in Section 4. Section 5 is dedicated to recent reports on graphene-based TIM including hybrids and alignment. Lastly, methods to characterise TIM and the need of further research in thermal management systems are discussed in Section 6.

It is now well established that there should be a good contact between the heat producing component and heat sink for efficient heat removal. When two solid surfaces meet, the surfaces seem to be in perfect contact. However, it is quite surprising that the picture which seems to be a perfect contact with each other through our naked eyes has a contact area of as little as 1%, since any surface which appears to be smooth has grooves and valleys at the micron level. The micro gaps are filled with air, which is a bad conductor of heat, having a thermal conductivity of around 0.22 W m⁻¹ K⁻¹. Hence, preventing heat transfer between the two surfaces [[7], [8], [9]]. TIM fills up this gap and its high thermal conductivity improves the heat transfer rate to achieve thermal equilibrium [[2], [3], [4],10].

The importance of TIM can also be estimated from the number of publications on this topic. The Scopus database shows around 39,494 document results when searched with the keyword “Thermal Interface Materials”. The graphs in Fig. 2 represents the last 10-years of publications when searched with the keywords (A) “Thermal Interface Materials” (B) “Graphene based Thermal Interface Materials”. The trend line demonstrates the importance and need of TIM to enhance the performance of thermal management in various electronic applications.

The application of TIM is certain in devices where a current flow through a resistive element which in turn produces heat. These devices include computers, mobile phones, medical devices, LED lighting, automobiles, tele-communication devices, lasers, photovoltaics, aerospace, wireless sensors etc. The role of TIM in an electronic device will keep on increasing due to the continuous research work on faster and more compact computers, lower power and high brightness LED lighting systems, electric vehicles with powerful batteries, the next generation of wireless telecommunication systems, and wearable as well as portable medical devices [11,12]. Most recent application of TIM is in thermoelectric generators in order to maximise the power conversion efficiency [13].

The electronic market is the largest market in the present scenario and TIM is the most essential component in all electronic devices, thus, the application of TIM is going to increase significantly as also suggested by a recent report of IDTechEx [14]. Packaging conditions, cost and cooling requirement are the key factors to be considered when deciding the type of TIM to be used. Broadly there are two types of TIM available in the market one is the adhesive based (grease, putties, gels) and other is non-adhesive based (pads, films) Generally, cost of adhesive based TIM are less, however the cost varies with the type of filler used, if diamond, graphene or CNT is used as nanofiller the cost will be higher. Adhesive type TIM have lower lifecycle as it dries out with time. Non-adhesive types are usually expensive, but the cost can be justified considering its advantages such as ease in packaging during manufacturing, high lifecycle and no additional curing requirement, hence it is the most popular and has the highest market share among other types [15]. Graphene and other carbon based non-adhesive TIM are very expensive which leads to significant increase in the cost of end product but is very crucial in devices where heat dissipation requirement is extremely high. The advantages and limitations of different types of conventional TIM are discussed in Section 2. Usually, one side of TIM is attached to the component where excessive heat is produced, and the other side is attached to the heat sink. The process of heat transfer should be in such a way that the lower part of TIM should rapidly take heat from the heat producing device and transfer to the upper part of TIM, which is in contact with the heat sink, leading to rapid heat removal. For example, in a laptop the TIM is usually placed on the central processing unit (CPU), the graphic processing unit (GPU), the solid-state disk (SSD) memory and the batteries. In mobile phones, thermal pads are placed next to the electromagnetic interference (EMI) shielding lid where important components such as integrated circuits (ICs) are present. TIM applications in light emitting diodes (LEDs) is expected to grow at a higher rate compared to others since automobiles are switching from classical headlamps to LED lighting, and also, heat spreaders are used in LCD and LED displays. Although previously discussed applications have a significant contribution, the area of enormous growth in TIM applications will be because of electric vehicles, which need high power batteries and hence advance cooling requirements. A pictorial representation of TIM applications in different fields is given in Fig. 3.

Heat transfer is carried out by electrons and phonons in solids [16]. The contribution in heat transfer by phonons and electrons is different with regards to the type of material. For example, the contribution of electrons is more in the case of metals than non-metals, whereas in case of non-metals, it is the phonon which is responsible for heat transport due to the non-availability of free electrons [17]. The crystal structure plays a crucial role when it comes to transport by phonons. When the heat source comes in contact with one side of a crystal lattice it creates vibrations and due to the dense packing of the crystal lattice, the vibrations are passed to the neighbouring atoms which leads to transfer of heat in crystalline materials, as shown in Fig. 4 (A). This is the exact reason for the reduction in thermal conductivity (TC) value with an increase in the number of layers of graphene. Thermal transport in graphene is mostly due to phonon waves, as proved by many researchers [[18], [19], [20]]. The vibrations have difficulty passing between the layers of graphene due to the weak van der Waals forces between them, hence as number of graphene layers increases, TC value decreases [[21], [22], [23], [24], [25]].

Heat transfer in polymers is difficult because of the complexity in the crystal structure and many factors, like crystallinity, temperature, orientation of the macromolecules need to be considered [[26], [27], [28]]. TC in polymers is mostly due to phonon waves since there are hardly any free electrons to transport [29]. When the polymer surface makes contact with a heat source, heat transfers in the form of vibrations to the nearest atom and then to the next and so on, hence diffuses slowly and not like a wave as in the case of a perfect structure, as shown in Fig. 4 (B). The vibrations are disordered due to the amorphous structure thereby reducing the TC. When the lattice structure is perfect, the heat transfer will be quicker between atoms.

Performance of TIM depends on various factors which include contact mechanics, tribology, rheology, heat transport and percolation. Good performance refers to the lowest total thermal resistance. Factors affecting the total thermal resistance are given by equation (1).

$$R_{TIM}=\frac{BLT}{K_{TIM}}+R_{C1}+R_{C2}$$

where R_TIM is the total thermal resistance, k_TIM is the TC of TIM, BLT is the bond line thickness (m) and R_c1 and R_c2 represent the contact resistances of the TIM with the two bonding surfaces. Reducing BLT by surface patterning and structured distribution of filler particles improves thermal conduction. A soft material having low mechanical shear strength is best in this regard, serving the purpose of ensuring a high contact and filling the gap in between (see Fig. 5 b and c) [[31], [32], [33], [34], [35], [36], [37]]. Another parameter is coefficient of thermal expansion (CTE) which is a property that tells how much the relative expansion (ppm K⁻¹) of the material will take place when exposed to a change in temperature. This property plays a role in defining the temperature range in which the device operates. CTE should be matched with the joining materials since a high CTE can damage the structural integrity at elevated temperatures due to expansion [38].

Equation (1) gives a fair idea about the factors that need to be considered while fabricating TIM. However, contact resistance are a bit more challenging to configure according to our requirements. Contact resistance and thermal conductivity are more significant than other parameters mentioned in the equation; hence it requires an elaborated explanation given in the following subsections.

Heat transfer between two components is due to radiation and conduction. Neglecting the heat transfer due to radiation, the only possible way for heat propagation is by solid to solid surface contact between the heat sink and the heat producing component, which could be as less as 1–2% as mentioned before in Section 1.1. The rest is non-contact with air filling in the gaps, contributing to the resistance called the thermal contact resistance.

Contact resistance is given by equation (2).

$$R_{cs}=\frac{8\sigma }{m{k}_h}{\left(\frac{H}{P}\right)}^{0.95}$$

where, $\sigma ={({\sigma }_1^2+{\sigma }_2^2)}^2$- root mean square roughness, $m={(m_1^2+m_2^2)}^{0.5}$- mean asperity slope, $k_h=\frac{2k_1{k}_2}{k_1+k_2}$ – harmonic mean thermal conductivity of the interface, $H$- microhardness of the softer material, and $P$- pressure applied.

The contact resistance should be as low as possible. When TIM is inserted, ideally, the air gap between the two surfaces is filled as shown in Fig. 5 (c), but two additional contact resistances will come into the picture which are the contact resistance between the heat producing device and TIM (R_c1), and the heat sink and TIM (R_c2) shown in Fig. 5 (a) [39].

In order to mitigate the two extra contact resistances, a direct solder technique using metallic solders to attach the heat sink and heat producing components seems to be promising. However, direct soldering leads to thermal stress and is also difficult to repair, rework, and requires a high processing temperature which can damage the device. They possess weak thermal fatigue resistance and other metallurgical problems which will be very troublesome for long term applications [40].

During heat transfer even if ideal condition of a perfect crystal structure is considered, the problem still persists at the interface because heat carriers such as phonons and electrons are different physically in the bulk and at the interface. The resistance caused at the interface has been studied using time domain thermo-reflectance (TDTR) by many researchers [41]. Their results indicate that interface resistance depends on the surface parameters like micro-level roughness, disordered structure, dislocations and other joining factors [42]. Interphase also play a vital role in influencing the thermal properties of the nanocomposite and can be altered with the fabrication conditions, processing temperature and interphase thickness [43]. One investigation by Monachon and Weber [44] was based on deposition of Aluminium on diamond surface using different techniques like hydrogen plasma, oxygen plasma, pure argon plasma and acid treatment for surface termination. All the mentioned methods were shown to decrease the resistance. It was reported that the reasons for this decrement are surface cleanliness, surface termination and reduction in surface roughness. Oxygen plasma treatment had the least resistance, the reason for which could be the elimination of hydrogen from the surface as pure argon treatment also had similar values. It is also important to note that there were two techniques used for the deposition, sputtering and evaporation. Between them it was found that deposition by using a sputtering technique showed better results to reduce resistance. The authors claimed that this is due to better layered adhesion since in sputtering argon, ions are evenly bombarded at the surface. Surface roughness is one of the biggest issues in thermal management as it is considered that roughness causes five times more resistance than ideal surfaces. Surface roughness (Ra) is defined as the average deviation from the mean surface height. When the roughness drops from 2 to 0.2 μm, one magnitude drop is observed in contact resistance. However, when the drop of roughness is in nano-metres, a four order drop in contact resistance is seen [45]. The effect of surface roughness at varying temperature and pressure was investigated by Ruifeng Dou et al. [46]. Their result indicated that lower surface roughness leads to a better thermal contact conductance (TCC). It can be noted that surface roughness has a major impact at lower interface temperatures than higher temperatures.

Applying pressure can also reduce the contact resistance by a factor of 3–4 which is due to filling the void between the two surfaces hence a deformable TIM is considered best in order to fill the voids between the mating surfaces, but it is also required that TIM is structurally conformed so that the TC is also consistent since TC values are highly dependent on the orientation of the molecules in the structure. The clamping pressure does not have any significance as long as the objective of filling the voids is fulfilled [47].

The complexity in this field is evident when we investigate the surface roughness, clamping pressure, the role of crystallinity and the structural conformity. Hence, most of the research to fabricate TIM is focussed in improving TC, however surface characteristics are also considered for fabricating advanced TIM which will be discussed in later sections. Talking about physical contact between two surfaces, it is well known that solid-solid contact surfaces have gaps between them, and surprisingly, solid-liquid contact is also imperfect with micro gaps. During heat transfer between solid and liquid, it needs to pass through the micro-contact point as shown in Fig. 6 [[48], [49], [50], [51]]. These factors play an important role since when a liquid comes in contact with the solid, air gets entrapped at the micro-level grooves present in the solid contributing to the thermal contact resistance. However, studies to understand the interaction of the solid-liquid surfaces are limited, some of the previous works are on the investigation of paste flow and capillary physics models [39,[52], [53], [54], [55]]. Among them, Hamasaiid et al. [54] developed a predictive model to quantify thermal contact resistance at the solid-liquid interface which investigated the topography at the contact point. The model reported contact point radius and density by using a classical flux tube method [56].

The mean height of the air trapped in the grooves can be approximated considering the applied pressure, capillarity pressure, and back pressure of the trapped air. Detailed mechanical analysis which includes contact angle measurement, surface energy of the fluids, and ideal gas behaviours gives the contact resistance which depends on surface topography, temperature, material thermal conductivity, wetting and applied pressure [45].

Thermal conductivity is the easiest parameter that one can vary to fabricate TIM for improved performance. Thermal conductivity is defined by Fourier’s law as shown by equation (3).

$$k=\frac{Q}{A}.\frac{Dx}{Dt}$$

where Q is the heat flow (W), A is the surface area (m²), ΔT is the temperature difference between two surfaces (K), Δx = BLT =Bond line thickness of the thermal interface material (m), and k is the thermal conductivity (W m⁻¹ K⁻¹).

It should be noted that there are two thermal conductivity values depending on the direction of heat flow. The first one is the in-plane thermal conductivity and good in-plane thermal conductive materials are excellent heat spreaders to mitigate the problem of hot spots. A hotspot is the region where the temperature is much higher than the rest of the area in the component. The second conductivity is through-plane thermal conductivity which is heat transfer in the z direction, that is heat transfer from the lower part of TIM to the upper part of TIM. Good in-plane thermal conductivity does not ensure removal of heat from the device but prevents hotspots, however good through-plane conductivity (along z axis) leads to efficient removal of heat from the heat producing component to the heat sink. Focussing on fabricating good TIM, heat conduction through the plane is very important. Fig. 5 (a) represents the heat transfer from a heat producing device to the heat sink in the z direction. Different TIM have been fabricated by using different materials with higher thermal conductivity values. Details regarding this will be discussed in Sections 4 Advanced carbon-based thermal interface materials, 5 Graphene.

TIM should be able to transport heat as fast as possible therefore materials with low thermal contact resistance, high thermal conductivity and low bond line thickness are preferred. However, there are some other factors that need to be considered as well while fabricating TIM. For some applications the material used should be an electrical insulator otherwise it can lead to short circuits and low frequency noises (LFN). For fabricating PCM based TIM it is very important to consider the temperature range in which the device will operate. Reliability and conformance need to be considered, especially for liquid and gel-based TIM, since liquid or gel-based TIM may spread during the course of operation and also during installation. Ideally the installed TIM should be reusable in another device hence it should be easy to install as well as remove from the device for sustainability. TIM should also be able to conform if there is a CTE mismatch in order to prevent damage due to stress.

Fig. 7 is a pictorial representation of ideal TIM characteristics. It is practically impossible for a TIM to have all these characteristics since many factors are interconnected in such a way that improving one characteristic negatively influences another characteristic. Therefore, during fabrication, an attempt should be made to balance each of these properties according to the need of the electronic device.