Investigation of a novel heat dissipation concept with controllable thermal and EM performance for reliable electronics and communication systems

This paper presents a novel efficient heat dissipation concept in electronics with switch controllable thermal and electromagnetic (EM) performance. Besides the traditional implementation of heatsinks on the back of the antenna in communication systems, a passive technique by introducing a reconfigurable heatsink that reuses the air space above the antenna for additional heat dissipation channels is proposed. The non-contact heatsink is lifted with a tiny air gap from the antenna and partially connected with the antenna through physical switches/poles that balance the EM performance and heat dissipation efficiency, which eliminates the negative impact of the heatsink on the antenna EM performance. EM effects of heatsink on the antenna in terms of locations of contacts, states of switches, and various dimensions are thoroughly investigated. To demonstrate the design efficacy for optimized thermal efficiency and EM performance, a properly designed heatsink is implemented on the top of a simple patch antenna. The measured results show that good heat dissipation is achieved without deteriorating the EM performance of the antenna, in respect to return loss, gain, and radiation patterns. When the switches are OFF and ON, the reflection coefficient of the antenna is measured as −20.82 dB and −17.56 dB, respectively. In addition, with a 20 mW heat source at the input port of the antenna to mimic the heat generation from the electronics, the temperature of the antenna surface is reduced with the front-integrated heatsink by 7.4 °C and 13.4 °C when the connection switches are turned OFF and ON. The implementation of the front-integrated heatsink on the antenna fully demonstrates the proposed heat dissipation concept which provides an effective way to solve the tradeoff between thermal inefficiency and EM performance.


Introduction
With the significant growth in mobile traffic and number of wireless communication standards, modern wireless communication and radar systems, especially MIMO system, are demanded miniaturized in size and improved in multifunctional performances [1,2]. High density 3-D RF technologies are thus the development trend while arising with a primary problem: heat generation and dissipation. The heat generated from miniaturized components should be dissipated, as fast as possible, from the antenna system to make the device long-lasting for reliable communications and maintain comfort to the user [3]. Device overheating is a common concern of many systems, and high-speed communications would exacerbate the problem further. Therefore, investigating ways to mitigate the thermal inefficiency in current and next-generation systems is of vital importance. However, it is extremely challenging and expensive to integrate active cooling systems, such as fans or forced liquids, on the limited physical space of mobile devices and systems. Low-cost passive heatsink cooling strategies are more favored in IoT (Internet of Things) and electronics applications that dissipate heat by increasing surface area between the device and the surrounding air [4][5][6]. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. configuration of a 3-D packaged system with integrated passive devices, multifunctional chips and antennas, etc Although a passive heatsink is usually implemented on the back of the substrate to transfer heat from the devices and the system to the surroundings [7], it is hard to further increase heat removal capacity and efficiency with the current configuration. Integrating a heatsink on the front of antennas is a promising solution with additional heat dissipation channels, however, the mere presence of the heatsink, even with micrometer dimensions, can inexorably distort the electromagnetic (EM) field.
Little work has been explored to reduce the performance degradation of the antennas from the integrated heatsink, and the previous works mainly focus on the structure optimization of heatsink which is an integral radiation part of the antenna. The effects of heatsinks with various configurations are studied for reduced radiated emissions to maintain the antenna performance [8]. A monopole antenna model is also created to characterize the electromagnetic radiation of the fin type heatsink to optimize its electromagnetic interference [9]. However, there is still a long way to significantly eliminate the EM distortion from the heatsink. In addition, the current analysis and modeling methodologies are dependent on the specific heatsink configuration which is not flexible for other designs. Instead of reducing the EM effects, heatsinks are utilized to enhance the radiation of the integrated device by exploring their electromagnetic resonant behaviors of the heatsink [10]. The heatsink is thus considered as a part of the antenna to generate good EM performance together with the inherent capacity of thermal removal [11,12]. The shapes of heatsinks are also developed from traditional fin type to other types, such as a 3-D fractal heatsink [13]. Nonetheless, it has to undergo a complex process to design a specific antenna with an integrated heatsink as its radiating element. The thermal performance of such an antenna is also limited since the heatsink size has already been optimized and fixed for a good EM performance. Therefore, a heat dissipation concept with more flexibility and thermal efficiency is highly desired to not only eliminate the negative EM effects of the heatsink on the radiations of the antenna, but also enable the flexible application of the front-integrated heatsink concept on arbitrary antennas.
A passive technique was first proposed by the authors with preliminary simulation results to introduce a reconfigurable heatsink that reuses the tiny air space above the antenna with optimized EM and thermal performance in communication systems, as illustrated in figure 1(b) [14]. In this paper, A new heat dissipation concept is proposed to integrate a heatsink with a tiny air gap above the antenna and to connect the antenna with physical switches that balance the EM performance and heat dissipation. Intuitively, when the antenna heats up excessively, the switches can be turned on, sacrificing the EM performance slightly, to achieve better thermal dissipation. The preliminary verifications on Ansys HFSS [15] and Fluent [16], an accurate electro-thermal simulator, have demonstrated encouraging results [14]. The challenges and effectiveness of integrating heatsink in the front of the antenna are comprehensively explored in this paper, theoretical analysis, design optimization, and experimental validation of the new heat dissipation concept are presented, to enable further heat removal from the antenna surface without disrupting the antenna. In section 2, the thermal performance of the heatsink on the antenna with various configurations and connection types is analyzed and studied. In section 3, the effect of the integrated heatsink on the EM performance of the antenna is investigated and optimized. Finally, a welldesigned heatsink is integrated on the front of a simple patch antenna in section 4 to demonstrate the feasibility and superiority of the proposed heat dissipation concept.

Thermal design and analysis
As shown in figure 2, a copper fin type heatsink is integrated in parallel on the top of a patch antenna to enable better heat dissipation from the antenna surface. The transfer of heat occurs through three different processes which are conduction, convection, and radiation. Due to the high thermal conductivity of copper (385 W mK −1 ) and the direct connection between the heatsink and the antenna, the heat is mainly transferred by the process of conduction and the transferred heat can be calculated by: where Q is the transferred heat, k is the thermal conductivity, η is the fin efficiency, T hot is the hot temperature, T cold is the cold temperature, t is the time of heat transfer, d is the thickness of the material, and A is the area of surfaces. Obviously, the integrated heatsink increases the effective total area of surface and thermal conductivity of the entire structure, resulting in the decreased temperature.
To demonstrate the effectiveness of the heatsink on the top of the antenna, the thermal performance of the patch antenna with and without heatsink is first simulated in Ansys Fluent. A 20 mW heat source is supplied at the input port of the antenna and the steady state thermal results are shown in figure 3. It is found from figures 3(a) and (b) that the front-integrated heatsink significantly cools down the antenna, and the center temperature of the antenna with and without heatsink is 318.7 K (45.6°C) and 300.9 K (27.8°C), respectively. In addition, as shown in figure 3(c), when the heatsink is suspended above the antenna for a small gap without direct touching the antenna metal surface, the heat dissipation effect is slightly reduced and the temperature at  the antenna center point is 309.3 K (36.2°C). This is because the heat transfer process between the antenna and the heatsink changes from the conduction to the convection process, resulting in the decreased thermal efficiency of the entire structure. Based on the thermal analysis and the simulated results, the directly connected heatsink with a larger area of surface provides maximum thermal dissipation for the antenna. However, the EM performance of the antenna is unavoidably deteriorated due to the integration of the heatsink, especially when the heatsink directly contacts the surface of the antenna.
To further determine the effect of the heatsink on the EM performance of the antenna, the patch antenna without and with a directly connected heatsink is simulated with Ansys HFSS, and the return loss is given in figure 4. The original antenna has a reflection coefficient of −32.75 dB at its operating frequency of 2.47 GHz. With the integration of the heatsink, its EM performance is severely deteriorated and the return loss at the operating frequency becomes 1.04 dB. This is because the radiation element of the antenna is significantly changed by the directly connected copper heatsink. The inserts in figure 4 show the electric field (E-field) distributions on the antenna with and without the front-integrated heatsink. At the operating frequency, a halfwavelength resonance can be found at the surface of the original patch antenna and the signal energy is mainly distributed at the two radiating edges. On the other hand, the integrated heatsink significantly changes the impedance of the antenna, generating the impedance mismatch between the feed line and the antenna. Hence, the signal cannot be radiated efficiently to the space through the antenna with the front-integrated heatsink. To address this challenge, the metal heatsink is proposed to be implemented on the front of the antenna without direct contact for the first time in this paper. Nevertheless, the heat dissipation effect will be decreased as illustrated in figure 3(c). Hence, achieving good heat removal without deteriorating the EM performance of the antenna remains a challenge and requires further investigation.

Electrical design and analysis
The EM performance of an antenna is highly dependent on its radiating elements. The impedance of an antenna is usually changed with the shape of its radiating element, leading to performance changes in return loss, gain, and radiation patterns. Moreover, the electrical length of resonance is affected by the shape of the radiation element, generating the operating frequency shift of an antenna. Hence, eliminating the unwanted effects introduced by the change of radiating element is of significant importance to integrate the heatsink on an antenna. To achieve this, the heatsink is proposed to be suspended above the antenna metal surface without a direct connection, as shown in the inserts in figure 5. In order to support the heatsink, each fin has metal pillars at both ends. The pillar width is chosen to be as small as possible, as long as it can support the weigt of the top heatsink. To optimize the EM and thermal performance, various parameters as summarized in table 1, such as the suspension gap height, dimensions, and locations of the heatsink, are thoroughly investigated in this section.
As depicted in figure 5(a), an air gap with a height of H 1 is created between the heatsink and the antenna surface. When the heatsink is suspended above the antenna without a direct connection to its metal surface, the EM performance of the antenna is significantly improved compared with the antenna with a fully connected   figure 4. If the heatsink is moved away from the antenna, i.e., increasing H 1 , its impact on the antenna is further reduced, resulting in a better return loss as shown in figure 5(a). It is worth mentioning that the integrated heatsink will influence the operating frequency of the antenna when the heatsink is closer to the antenna surface. This is because the effective electrical length of the antenna is increased by the nearby heatsink. Therefore, suspending the heatsink away from the antenna surface improves the EM performance of the antenna, while inevitably degrades the heat dissipation performance as simulated in figure 3(c). Based on the simulated trend in figure 5(a), the heatsink can be suspended for a proper height to achieve a good balance between EM and thermal performance.

front-integrated heatsink as shown in
After determining the airgap height of H1 for the optimized antenna performance, the height of the heatsink on the EM performance of the antenna is also studied by changing the values of H 2 , as shown in the insert of figure 5(b). Based on (1), the heat dissipation capability of a fin type heatsink is dependent on its total surface area. Although a taller heatsink provides better heat dissipation capability, it may affect the performance of the antenna. Configurations of the heatsink with heights ranging from 20 mm to 30 mm are simulated to study its effect on the EM performance of the antenna. As shown in figure 5(b), there is little shift of operating frequency of the antenna while the reflection coefficient is changed from −25.8 dB to −16.9 dB. This means that the height of the heatsink has a relatively small impact on the EM performance of the antenna with the pre-defined optimum airgap between the antenna and the heatsink, and it provides an efficient solution to increase the surface area of the heatsink for better heat dissipation and optimal EM radiation performance. Nevertheless, the height cannot be increased to infinity since the enlarged heatsink will eventually block the radiation path of the antenna, resulting in a smaller return loss.
In addition, electromagnetic fields of different strengths are distributed across the distinct locations of the antenna. As illustrated in the insert of figure 5(c), another solution to reduce the effect of the front-integrated heatsink on the EM performance of the antenna is proposed to place the heatsink fins on the location with a distance of S away from the edge of the antenna where has the strongest EM field. To further explore the effect of the heatsink, the location of the heatsink on the top of the antenna is thoroughly investigated. The EM simulation in figure 5(c) starts from a case with bad EM performance, i.e., the heatsink is suspended from the antenna surface with H 1 = 1mm. With the increase of S, the fins are moved closer to the antenna center, the return loss of the heatsink integrated antenna matches the performance of the original patch antenna without the heatsink. This validates the effects of the heatsink fins become smaller when they are moved further away from the radiating edges of the antenna, where the E-field is intensely distributed as mentioned in the insert of figure 4. This is an important approach to solve the tradeoff between the optimized EM performance and thermal efficiency.
Based on the above investigation, the following strategies are applied in the proposed heat dissipation concept for the optimized EM performance: (a) The heatsink should not be fully directly connected to the radiating part of the antenna to avoid severely degrading the EM performance of the bottom antenna; (b) The height of the heatsink can be increased for better thermal performance since it has a relatively small impact on the EM function; (c) The heatsink can be partially connected to the selective area of the antenna where the E-field strength is weakest. The direct connection will significantly improve the heat dissipation while minimizing the negative impact on antenna performance; (d) Physical switches can be employed to control the partial connections between the heatsink and the antenna surface for controllable maximum EM and thermal performance on demand.

Demonstration of the heat dissipation concept Configuration and optimization
To demonstrate the efficacy of the proposed heat dissipation concept, a heatsink is designed and integrated on the front of a simple patch antenna to provide excellent thermal and EM performance. Figure 6 describes the schematic diagram of the proposed design, where a fin type heatsink is suspended above the patch antenna with a distance of H 1 . The patch antenna with length L 0 and width W 0 is designed on a normal substrate with a Table 1. Parameter scanning of integrated heatsink.

Parameters Results
The height of air gap between the heatsink and antenna, H 1 (H 1 = 1, 3, 5 mm)  figure 7. When the switches are turned OFF, the heatsink is suspended without direct connection to the antenna surface. The integrated antenna works at the same frequency (2.47 GHz) as the original patch antenna, and its reflection coefficient at that frequency is simulated as −25.8 dB. The E-field shows that half-wavelength resonance is generated at the operating  In summary, the proposed heatsink integrated antenna has controllable EM and thermal performance. Good heat dissipation capability is achieved even when the switches are turned OFF because of the strong thermal convection between the heatsink and antenna. The thermal performance can be further improved by turning the metal switches ON, enabling the direct connection between the heatsink and antenna surface, with limited deterioration of the EM performance. The states of the switches are simulated with perfect open/short for the ease of demonstration in this paper, the switches could be implemented with MEMS ohmic contact switches or thermal-activated metal/ceramic switches in practical applications.

Fabrication and measurement
To further validate the proposed heat dissipation concept, three prototypes are fabricated with the dimensions listed in section 3 and their photos are inserted in figure 8. The first one is the original patch antenna without the heatsink, which is used as a reference. The operating frequency of this antenna is measured to be 2.49 GHz with a reflection coefficient of −29.43 dB. The slight difference between the simulated and measured operating frequency is caused by the small deviations of substrate permittivity. In addition, the extra loss of the SMA connector makes the measured return loss at the operating frequency slightly larger than the simulation. When a heatsink is integrated above the front of the antenna without direct contact the antenna (switches in OFF state), the antenna still works at 2.49 GHz and the reflection coefficient is measured as −20.82 dB. This demonstrates that the meticulously designed heatsink has little effect on the EM performance of the antenna. When the switches are turned ON, the heatsink is directly connected to the middle line of the antenna surface, which enables the best heat dissipation capability. The antenna still has superior performance at 2.46 GHz with a reflection coefficient of −17.56 dB. As depicted in figure 9, the radiation patterns of the proposed antenna are also measured in a microwave anechoic chamber. Similar radiation patterns are preserved among the three antennas, which again shows little negative impact of the integrated heatsink. The max gains of the original antenna and the heatsink integrated antenna under OFF and ON state are measured as 5.2 dB, 5.8 dB, and 6.1 dB, respectively. In conclusion, the heatsink integrated antenna has satisfactory performance under both OFF and ON state of the heatsink. The direct contact between the heatsink and the antenna may slightly deteriorate the EM performance but will significantly improve the thermal performance.
To demonstrate the thermal dissipation performance of the front-integrated heatsink, the three prototypes are connected to a 20 mW heat source with SMA connectors, and their steady temperatures are measured by a probe thermometer. To obtain steady temperatures, the prototypes are measured after applying the heat source under the room temperature of 23°C for 30 min. Figure 10 gives the measured temperature distribution of the three samples along the centerline from the feed point to the end of the antenna. Take the center point of the antenna surface as an example, the measured temperatures of the three prototypes are 40.1°C, 32.7°C, and 26.7°C, respectively. This means that the heatsink effectively cools down the antenna even if the heatsink is not directly connected to the antenna with the switches being turned OFF. The temperature can be further reduced when the heatsink is directly connected to the antenna. Extra temperature reduction can be achieved by increasing the height of the heatsink. The performance of the antenna with the front-integrated heatsink and the original patch antenna are compared in table 2. It fully demonstrates that the proposed heat dissipation concept can achieve controllable and optimum EM and thermal performance. There is also a high degree of design flexibility in applying the proposed design concept on arbitrary antennas to achieve dual thermal and EM functionality.   This paper presents a novel heat dissipation concept to enable switch-controllable and optimum EM and thermal performance for arbitrary antennas in communication systems. The heatsink is comprehensively studied and optimized by thermal and electrical analysis, strategies and methods are proposed to minimize the effect of heatsink on the radiation performance of the antenna for maximum thermal efficiency. To demonstrate the design efficacy, a heatsink is implemented on the front of a simple patch antenna, and the experimental results show a significantly increased thermal dissipation capability without degrading the performance of the antenna. The proposed design concept provides a practical and efficient way to mitigate thermal inefficiency without disrupting the radio waves of the antenna in current and next-generation wireless communication systems.

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