Multifactor roadmap for designing low‐power‐consumed micro thermoelectric thermostats in a closed‐loop integrated 5G optical module

Abstract

As the core components of fifth‐generation (5G) communication technology, optical modules should be consistently miniaturized in size while improving their level of integration. This inevitably leads to a dramatic spike in power consumption and a consequent increase in heat flow density when operating in a confined space. To ensure a successful start‐up and operation of 5G optical modules, active cooling and precise temperature control via the Peltier effect in confined space is essential yet challenging. In this work, p‐type Bi_0.5Sb_1.5Te₃ and n‐type Bi₂Te_2.7Se_0.3 bulk thermoelectric (TE) materials are used, and a micro thermoelectric thermostat (micro‐TET) (device size, 2 × 9.3 × 1.1 mm³; leg size, 0.4 × 0.4 × 0.5 mm³; number of legs, 44) is successfully integrated into a 5G optical module with Quad Small Form Pluggable 28 interface. As a result, the internal temperature of this kind of optical module is always maintained at 45.7°C and the optical power is up to 7.4 dBm. Furthermore, a multifactor design roadmap is created based on a 3D numerical model using the ANSYS finite element method, taking into account the number of legs (N), leg width (W), leg length (L), filling atmosphere, electric contact resistance (R_ec), thermal contact resistance (R_tc), ambient temperature (T_a), and the heat generated by the laser source (Q_L). It facilitates the integrated fabrication of micro‐TET, and shows the way to enhance packaging and performance under different operating conditions. According to the roadmap, the micro‐TET (2 × 9.3 × 1 mm³, W = 0.3 mm, L = 0.4 mm, N = 68 legs) is fabricated and consumes only 0.89 W in cooling mode (Q_L = 0.7 W, T_a = 80°C) and 0.36 W in heating mode (T_a = 0°C) to maintain the laser temperature of 50°C. This research will hopefully be applied to other microprocessors for precise temperature control and integrated manufacturing.

INTRODUCTION

Electronic devices are essential for new energy, Internet of Things, artificial intelligence, and other fields.^[1–4] High-integration and miniaturization is an inevitable trend for the core components of electronic devices. Unfortunately, due to the efficiency limitation, nearly 80% of the input electrical power is converted into waste heat.^[5–7] If the heat dissipation problem is not addressed promptly and effectively, the temperature of electronic devices will rise, resulting in deteriorated performance and reliability. In severe cases, electronic devices may even burn out once the maximum operating temperature is exceeded. In addition, electronic devices must be preheated to ensure successful start-up in a cold environment.^[8–10]

Thermoelectric (TE) technology has gained immense recognition for its environmentally friendly nature and all-solid-state energy conversion capabilities (Figure S1).^[11–15] The micro thermoelectric thermostat (micro-TET) can utilize the Peltier effect to actively pump heat and switch between cooling (micro-TEC) and heating (micro-TEH) functions simply by altering the direction of the current, enabling precise temperature control of the area (Figure 1A).^[16–19] As it is a solid-state device with no moving parts, no maintenance requirements, no noise production, and is easily integrated with other electronic components,^[20–25] it has become a popular choice for thermal management in daily life and industry.^[26–33] Xu et al.^[34] developed a flexible TE cooler with a four-layered structure of “Ag₂Se/poly(3,4-ethylenedioxythiophene) polystyrene sulfonate” for human skin thermal management. Whether the wearer stands still or swings the arm, the skin covered by the cooler can be maintained at a comfortable temperature of 32 ± 0.5°C. Ji et al.^[35] used an integrated system combining TE heating and cooling to achieve the disinfection and purification of air. The whole system has a performance factor of up to 5.4 and is expected to play an important role in the medical field, such as the inactivation of SARS-CoV-2 virus. Li et al.^[36] developed a TE cooler integrated directly on the chip (chip-on-TEC) for active thermal management of high-power light-emitting diode (LED). When the current in the chip was 1.0 A, the chip-on-TEC was able to reduce the operating temperature of the LED from 232°C to 114°C (51% reduction) and increase the light output power by 35.3%. Song et al.^[37] coupled TE devices and phase change materials to keep the outdoor standby battery at an optimal temperature range in the environment of 263–323 K. Shen et al.^[38] proposed to use TE heaters to enhance the fuel cell system's water heating capacity by more than two times and the overall energy efficiency by up to 50%.

[IMAGE OMITTED. SEE PDF]

With high speed and low latency, the fifth-generation (5G) mobile communication technology is the network infrastructure to realize the Internet of Things.^[39,40] A new generation of ultrafast, large-capacity, high-performance optical devices is essential for driving the progress of the entire optical networking sector. The performance of optical modules and their internal components is highly sensitive to temperature. An increase in temperature will cause a reduction of optical power output and wavelength red drift. Hence, it is necessary to maintain a specific temperature for optical modules to ensure data transmission. Unfortunately, due to the increased integration and packing density of optical modules, they consume more power and generate more heat while providing robust functionalities, thus requiring improved heat dissipation of the modules.^[41,42] At the same time, the optical module must be warmed up to the specified operating temperature before it can start up in a low-temperature environment.

The integration of micro-TET and optical modules offers an ideal solution for achieving precise temperature control in the 5G communication process.^[15,43] This study aims to precisely control the operating temperature of the laser in the 5G optical module to 50°C. The multifactor roadmap is designed via the finite element simulation method to optimize the cooling and heating power consumption required for temperature control of the laser in different environments, which includes the number of Bi₂Te₃-based TE legs (N), leg width (W), leg length (L), filling atmosphere, electric contact resistance (R_ec), thermal contact resistance (R_tc), ambient temperature (T_a), and the heat generated by the laser source (Q_L). This research guides the integrated fabrication of micro-TET and points out the direction for packaging and performance optimization under different operating conditions.

RESULTS AND DISCUSSION

5G optical module integrated fabrication

The micro-TET (device size, 2 × 9.3 × 1.1 mm³; leg size, 0.4 × 0.4 × 0.5 mm³; number of legs, 44) was successfully integrated into a 5G optical module with Quad Small Form Pluggable (QSFP) 28 interface (Figures 2A, S6, and S7, Video S1). A portion of the current originally flowing into the 5G optical module is allocated to the micro-TET. For this type of 5G optical module, the internal temperature remains stable at 45.7°C with an optical power of up to 7.4 dBm (Figure 2B_i and B_ii) when the micro-TET is working in cooling mode (TEC ON). However, once the micro-TET is turned off (TEC OFF), the internal temperature increases from 45.7°C to 50.7°C (Figure 2B_i), and the optical power decreases from 7.4 to 6.88 dBm (Figure 2B_ii) in 12 s. There is no doubt that the integration of micro-TET is vital to the proper functioning of the 5G optical module.

[IMAGE OMITTED. SEE PDF]

To clearly observe the working process of micro-TET, infrared monitoring experiments were carried out on the test platform (Figure 2C,D), with the target control temperature set to 50°C. When the ambient temperature T_a is 80°C and the heat generation Q_L is 0.7 W, the “laser” temperature is as high as 99.7°C (Figure 2C_i). However, when the micro-TET is switched on to the cooling mode (TEC: OFF → ON), the “laser” temperature quickly falls back to the target of 50°C (Figure 2C_ii). Similarly, when the ambient temperature T_a is 0°C (Figure 2D_i), it quickly reaches the target temperature of 50°C when the micro-TET is switched to heating mode (TEH: OFF → ON) (Figure 2D_ii). This demonstrates the capability of the micro-TET to effectively manage local hot and cold spots in the optical module. To further optimize its structure, a multifactor design can be employed, which meets the space requirements for a specific optical module, while minimizing power consumption.

Multicore-factor roadmap

The performance of a micro-TET is closely related to the core factors of the number of TE legs (N), leg width (W), and leg length (L). This subsection calculates the cooling (Figure S8) and heating (Figure S9) power consumption of the micro-TET under varying heat generation Q_L and ambient temperature T_a.

When the ambient temperature is lower than the operating temperature of the 5G optical module, the micro-TET works in the heating mode to warm up the optical module to the desired start-up temperature; once the laser is activated, the temperature of the optical module will surpass the optimum operating value, promoting the micro-TET to shift to cooling mode to maintain the laser at a target temperature of 50°C.

The weight of the cooling and heating mode will vary depending on the working states of the 5G optical module. When the weight of cooling and heating power consumption is 9:1, the roadmap of combining the total power consumption (P_total) with the number of TE legs, leg width, leg length, and heat generation of the laser source is shown in Figure 3. Other roadmaps with cooling and heating power consumption weights, such as 8:2, 7:3, 6:4, and 5:5, are presented in Figures S10–S13.

[IMAGE OMITTED. SEE PDF]

For the micro-TET with constant Q_L and N, the most appropriate leg size (Table S2) or ACR resistance (Figure S14) to minimize P_total can be found (Figure 3). As the Q_L increases from 0.3 to 1.0 W, the P_total of the micro-TET increases significantly, while the range of TE leg sizes able to reach the target temperature of 50°C decreases (the blank area represents that the corresponding leg size cannot be designed to meet the target) (Figure 3). Moreover, when the Q_L is as low as 0.3 W, the leg size range corresponding to a small P_total decreases monotonously as the N increases from 34 to 84 (Figure 3A). For a Q_L of 0.7 W, the leg size range corresponding to a small P_total expands and then contracts as the N increases (Figure 3B). However, when the Q_L reaches 1.0 W, the size range of TE legs capable of achieving lower P_total significantly enlarges as the N increases from 34 to 84 (Figure 3C). Obviously, this is an important finding for effective energy saving and emission reduction in high-power laser temperature control.

Multiauxiliary-factor roadmap

According to the multicore-factor roadmap (weighted ratio of P_TEC and P_TEH is 9:1), the micro-TET with boundary dimension of 2 × 9.3 × 1 mm³ and leg size of 0.3 × 0.3 × 0.4 mm³ has the lowest weighted power consumption P_total of 0.795 W, with the cooling power consumption P_TEC of 0.84 W and the heating power consumption P_TEH of 0.39 W under the condition of Q_L = 0.7 W and N = 68. In actual use, the power consumption of the micro-TET is affected by various factors, such as the filling atmosphere, electric contact resistance R_ec, thermal contact resistance R_tc, and ambient temperature T_a. In this section, the influence of these factors on the input power of micro-TET is further investigated.

Filling atmosphere

Figure 4A displays the cooling and heating temperature-control curves, as well as the corresponding power consumptions of the micro-TET, under different filling atmospheres. The lowest cooling and heating power consumption of the micro-TET is observed in a vacuum environment, which is 0.84 and 0.39 W, respectively. As the thermal conductivity κ of the filled atmosphere increases (Table S1), the heat transfer is enhanced, and the input current I required to reach T_target = 50°C (black dashed line) in the cooling and heating modes of the micro-TET gradually increases (Figure 4A_i and A_ii), resulting in a slight rise in power consumption (Figure 4A_iii and A_iv). Specifically, in a helium (He) atmosphere, the cooling and heating power consumption of the micro-TET reaches 1.0 and 0.51 W, respectively, an increase of 19.1% and 30.8% compared with that in the vacuum environment (Figure 4A_iii and A_iv). Moreover, the effect of the filling atmosphere on power consumption is more noticeable in the heating mode than that in the cooling mode. This serves as a guide to selecting the most suitable filling atmosphere according to the packaging requirements.

[IMAGE OMITTED. SEE PDF]

Electric contact resistance

Figure 4B illustrates the effect of the electric contact resistance between the electrodes and TE legs on the cooling and heating temperature-control curves and the power consumption of the micro-TET. The input current I required to reach T_target = 50°C (indicated by the black dashed line) increases gradually in the cooling mode (Figure 4B_i), while decreasing slightly in the heating mode (Figure 4B_ii). However, both the cooling and heating power consumption increases as the electric contact resistance of the micro-TET rises (Figure 4B_iii and B_iv). At R_ec = 0 µΩ cm², the lowest cooling and heating power consumption is 0.73 and 0.36 W, respectively. At R_ec = 0.22 µΩ cm² (the lowest electric contact resistance value achieved by interface optimization in the current stage),^[45] the cooling and heating power consumption slightly increases to 0.74 and 0.37 W, respectively. At R_ec = 3.0 µΩ cm² (the electric contact resistance of a typical commercial device),^[45] the cooling and heating power consumption grows to 0.84 and 0.39 W, respectively, an increase of 15.1% and 8.33% compared with the ideal value. When the micro-TET is used for a long time, and the interface deteriorates to R_ec = 14.0 µΩ cm², the corresponding cooling and heating power consumption rises abruptly to 1.49 and 0.46 W, respectively, an increase of 104.1% and 27.8% over the ideal value. The increase in electric contact resistance leads to a growth in the internal resistance of the micro-TET (Figure S15), meaning more power is needed to enhance the Peltier effect to achieve the target temperature of 50°C.^[46,47] Notably, the influence of electric contact resistance on cooling is much greater than that of heating, which indicates the direction for future optimization of electric contact resistance.

Thermal contact resistance

Figure 4C shows the cooling and heating temperature-control curves, as well as the power consumption of the micro-TET, under different contact thermal resistances. As the thermal contact resistance increases from 0 to 2 × 10⁻⁵ m² K/W, the cooling power consumption of the micro-TET increases from 0.72 to 1.01 W (Figure 4C_i and C_iii). Actually, when the thermal contact resistance R_tc is larger than 5 × 10⁻⁵ m² K/W, it becomes difficult to drop the laser temperature to 50°C (Figure 4C_i and C_iii). Interestingly, the heating power consumption decreases from 0.39 to 0.32 W when the thermal contact resistance R_tc rises to 1 × 10⁻⁴ m² K/W (Figure 4C_ii and C_iv). This is because, in the Peltier heating process, the increased thermal contact resistance impedes the heat transfer from the high-temperature side to the low-temperature side, thus reducing the heat dissipation from the heated object. The different effect of contact thermal resistance on the cooling and heating process provides more space for the design of micro-TET in the future.

Ambient temperature

Figure 4D shows the cooling and heating temperature-control curves and the corresponding power consumption for the micro-TET at different ambient temperatures. It can be seen that the cooling power consumption increases as the ambient temperature rises (Figure 4D_i and D_iii), while the heating power consumption decreases correspondingly (Figure 4D_ii and D_iv). The closer the ambient temperature T_a is to the target temperature of 50°C, the lower the cooling and heating power consumption. When T_a is raised from 60°C to 100°C, the cooling power consumption increases significantly by 7.2 times (from 0.29 to 2.07 W). Similarly, when T_a is increased from −40°C to 0°C, the heating power consumption decreases by 65.2% (from 1.12 to 0.39 W). This roadmap can be used to effectively design the micro-TET according to the prevailing ambient temperature.

Multifactor design roadmap verification

In this section, five kinds of micro-TETs have been fabricated to verify the validity of the created multifactor roadmap. Both of the G values (the ratio of length to cross-sectional area) of the TE leg and ACR values gradually increase for micro-TET 1# to 5# (Table S3 and Figure S16).

Figures 5A_i and A_ii show the simulated and measured cooling power consumption of the micro-TETs at T_a = 80°C and Q_L = 0.7 and 1.0 W, respectively. Due to the additional electric contact resistance and thermal contact resistance introduced by the machining process of TE legs, there is a slight discrepancy between the simulated and measured values. As the G value increases, the measured cooling power consumption P_TEC still follows the same trend as the simulated results, first decreasing to its lowest value, then significantly increasing (Figure 5A_iii).

The simulated and measured heating power consumptions of the micro-TETs at ambient temperatures of T_a = 0°C and 10°C are depicted in Figures 5B_i and B_ii, respectively. As anticipated, the measured heating power consumption P_TEH of the micro-TETs decreases as G increases, which is in line with the trend of the simulated value (Figures 5B_iii). There is, however, some discrepancy between the measured and simulated results since the actual thermal contact resistance during fabrication is higher than that set in the simulations.

The measured results are consistent with the theoretical prediction, affirming that the roadmap created is an efficient guide for fabricating a micro-TET that meets the desired requirements. For micro-TET 4#, the cooling and heating power consumption required to reach the target temperature of 50°C is low. At T_a = 80°C, the actual cooling power consumptions are 0.89 and 1.59 W under the condition of Q_L = 0.7 and 1.0 W, respectively (Figure 5A). Additionally, the measured heating power consumptions at ambient temperatures of 0°C and 10°C are 0.36 and 0.22 W, respectively (Figure 5B).

[IMAGE OMITTED. SEE PDF]

The detection of micro-TET 4# in various applications is shown in Figure 5C,D. When the micro-TEC is turned off, the laser temperature rises to 110°C (Figure 5C_i). Upon turning on the micro-TEC, the laser temperature drops to 50°C in 275 s and then remains stable (Figure 5C_i). In Figure 5C_ii, the laser temperature decreases from 110°C to 82°C when the micro-TEC is turned on for 30 s. After turning off the micro-TEC, the laser temperature immediately rises back to 110°C. When the micro-TEC is turned on again for 45 s, the laser temperature decreases rapidly from 110°C to 71°C. When the micro-TEC is turned off, the laser temperature immediately rises to 110°C, and when the micro-TEC is turned on again for 275 s, the laser temperature drops from 110°C to 50°C and remains stable. In addition, when the ambient temperature changes between 70°C and 90°C, micro-TEC can quickly control the laser temperature back to 50°C. And the closer the ambient temperature is to the target temperature, the shorter the time for the laser temperature to return to stability after micro-TEC working (Figure 5C_iii).

By reversing the current, the operation of the micro-TET 4# switches from the cooling mode to the heating mode (Figure 5D). When the micro-TEH is turned off, the laser temperature is about 1.6°C, close to the ambient temperature of 0°C. After the micro-TEH turns on, the laser warms up and stabilizes to 50°C in 265 s (Figure 5D_i). In addition, the laser temperature rises to 20°C after 20 s of micro-TEH operation. When the micro-TEH is turned off, the laser temperature immediately drops back to 1.6°C. When the micro-TEH is turned on again for 75 s, the laser temperature rises to 40°C, and when the micro-TEH is turned off, the laser immediately cools down to 1.6°C. When the micro-TEH is turned on again for 265 s, the laser warms up to 50°C (Figure 5D_ii). Even if the ambient temperature varies between −10°C and 10°C, the laser can be warmed up to the target temperature of 50°C immediately by turning on the micro-TEH (Figure 5D_iii). The closer the ambient temperature is to the target temperature of 50°C, the shorter the time required to heat the laser (Figure 5D_iii).

All of the above experiments effectively verify the feasibility and stability of the micro-TET to achieve precise temperature control through cooling and heating modes.

CONCLUSIONS

In this work, p-type Bi_0.5Sb_1.5Te₃ and n-type Bi₂Te_2.7Se_0.3 bulk TE materials are used, and a micro-TET (device size, 2 × 9.3 × 1.1 mm³; leg size, 0.4 × 0.4 × 0.5 mm³; number of legs, 44) is successfully integrated into a 5G optical module with QSFP 28 interface. As a result, the internal temperature of this kind of optical module is always maintained at 45.7°C and the optical power is up to 7.4 dBm. Furthermore, a multifactor design roadmap is created based on a 3D numerical model using the ANSYS finite element method, considering the number of legs (N), leg width (W), leg length (L), filling atmosphere, electric contact resistance (R_ec), thermal contact resistance (R_tc), ambient temperature (T_a), and the heat generated by the laser source (Q_L). According to the roadmap, a micro-TET with a boundary dimension of 2 × 9.3 × 1 mm³ and leg size of 0.3 × 0.3 × 0.4 mm³ has been successfully fabricated (N = 68). It has a low cooling power consumption P_TEC of 0.89 W (Q_L = 0.7 W, T_a = 80°C) and a heating power consumption P_TEH of 0.36 W (T_a = 0°C) for precise control of the laser temperature at 50°C.

EXPERIMENTAL

Finite element modeling

The model consists of three parts: laser source array, micro-TET, and heat sink. Figure 1B displays the model and its equivalent thermal resistance. The laser source array comprises four high-purity silicon lasers and an aluminum nitride (AIN) support base. For the 5G optical module, the integrated micro-TET is designed with a size of 2 × 9.3 mm², and its upper and lower surfaces are connected to the laser source and WCu heat sink by silver glue, respectively. The TE materials used in the thermostat are p-type Bi_0.5Sb_1.5Te₃ and n-type Bi₂Te_2.7Se_0.3 (Figure S2). The materials of the electrode, substrate, and solder are copper, AlN ceramic, and Au–Sn alloy, respectively (Table S1). ANSYS Workbench finite element software is used for all simulations. The temperature and electric potential fields of the micro-TET working in cooling and heating modes are shown in Figure 1C.

Boundary conditions setting

The heat flux and current density are coupled by three basic TE effects,^[44] and the corresponding thermodynamic relationships can be expressed as follows^[7,20]: 1$$\nabla (\kappa \nabla T)+\frac{{{\boldsymbol{J}}}^{2}}{\sigma }-T{\boldsymbol{J}}\cdot \left[\left(\frac{\partial \alpha }{\partial T}\right)\nabla T+{(\nabla \alpha )}_{T}\right]=0,$$ 2$$\nabla \cdot {\boldsymbol{J}}=0,$$ 3$${\boldsymbol{J}}=-\sigma (\nabla V+\alpha \nabla T),$$ 4$${\boldsymbol{q}}=\alpha T{\boldsymbol{J}}-\kappa \nabla T,$$where α is the Seebeck coefficient, κ is the thermal conductivity, σ is the electrical conductivity, T is the absolute temperature, V is the electrostatic potential, and the vectors J and q represent the current density and heat flux density, respectively. Among them, α, κ, and σ are all dependent on temperature.

To refine the model and simplify calculations, the following reasonable assumptions are proposed:

• (1)

  All surfaces (except the hot and cold ends) are considered to be well thermally insulated.

• (2)

  In the thermostat chamber filled with atmosphere, heat transfer occurs not only in the direction perpendicular to the substrate but also in other directions.

• (3)

  The electric contact resistance (R_ec) and thermal contact resistance (R_tc) between electrodes and TE legs are both taken into account in the finite element model (Figure S3). When R_ec is variable, R_tc is set to 1 × 10⁻⁵ m² K/W; when R_tc is variable, R_ec is set to 3.0 µΩ cm².

• (4)

  The Thomson effect is neglected.

Micro-TET fabricating

According to the designed connection circuit, copper electrodes (0.025 mm) were patterned on the AlN substrate (0.25 mm) by an adhesive-free calendering method. The copper electrode surface was gold-plated to improve welding reliability and thus reduce contact resistance. p-Type Bi_0.5Sb_1.5Te₃ and n-type Bi₂Te_2.7Se_0.3 bulk materials were cut into cuboid-shaped legs, of which the upper and lower surfaces were preplated with nickel (~5 μm) and gold (~100 nm) film. Then, the sandwich structure composed of AlN substrate-TE legs-AlN substrate was welded together by AuSn solder to make micro thermoelectric devices (Figure 1D). The detailed preparation process can also be checked in our previous article.^[45]

Test platform building

A planar heater (TEC1-0341M-2093, P_max = 4.0 W, Wuhan New Sail Technology Co., Ltd.) is employed to produce heat, serving as the laser source in the 5G optical module. The lift table acts as a temperature-regulating platform to simulate the ambient temperature. The micro-TET is positioned between the planar heater and the lift table (Figure S4). Thermally conductive silicone grease should be applied at the junction. When the micro-TET reaches the temperature-control target of 50°C, the cooling and heating power consumption can be accurately read through the data transmission channels (Figure S5). The test platform has been supported and recognized by Wuhan Optical Technology Co., Ltd. (Accelink), which is the largest supplier of optical communication devices in China (top 5 in the world) and a high-tech enterprise in China that has the ability to carry out systematic and strategic research and development of optoelectronic devices.

AUTHOR CONTRIBUTIONS

Dongwang Yang and Yubing Xing contributed equally to this work. Xinfeng Tang and Yonggao Yan conceived the project. Yubing Xing, Kechen Tang, and Yutian Liu carried out the finite element simulation. Dongwang Yang, Yubing Xing, Jiang Wang, Kai Hu, Yani Xiao, Jianan Lyu, and Junhao Li prepared the micro-TETs and tested their output power generation performance. Dongwang Yang, Yubing Xing, Peng Zhou, Yuan Yu, Yonggao Yan, and Xinfeng Tang analyzed the experimental data. Xinfeng Tang and Dongwang Yang cowrote the manuscript. All authors have given approval to the final version of the manuscript.

ACKNOWLEDGMENTS

This work was financially supported by the National Natural Science Foundation of China (52202289) and the National Key Research and Development Program of China (2019YFA0704900).

CONFLICT OF INTEREST STATEMENT

The authors declare no conflict of interest.

References

Gai Y, Li H, Li Z. Self‐healing functional electronic devices. Small. 2021;17:2101383.

Mao S, Sun B, Zhou G, Guo T, Wang J, Zhao Y. Applications of biomemristors in next generation wearable electronics. Nanoscale Horiz. 2022;7:822‐848.

Portilla L, Loganathan K, Faber H, et al. Wirelessly powered large‐area electronics for the Internet of Things. Nat Electron. 2023;6:10‐17.

Zhu R, Zhu Y, Chen F, et al. Boosting moisture induced electricity generation from graphene oxide through engineering oxygen‐based functional groups. Nano Energy. 2022;94:106942.

Barrubeeah M, Rady M, Attar A, Albatati F, Abuhabaya A. Design, modeling and parametric optimization of thermoelectric cooling systems for high power density electronic devices. Int J Low‐Carbon Technol. 2021;16:1060‐1076.

Cai Y, Wang Y, Liu D, Zhao FY. Thermoelectric cooling technology applied in the field of electronic devices: updated review on the parametric investigations and model developments. Appl Therm Eng. 2019;148:238‐255.

Gong T, Gao L, Wu Y, et al. Numerical simulation on a compact thermoelectric cooler for the optimized design. Appl Therm Eng. 2019;146:815‐825.

Li G, Garcia Fernandez J, Lara Ramos DA, et al. Integrated microthermoelectric coolers with rapid response time and high device reliability. Nat Electron. 2018;1:555‐561.

Twaha S, Zhu J, Yan Y, Li B. A comprehensive review of thermoelectric technology: materials, applications, modelling and performance improvement. Renew Sust Energy Rev. 2016;65:698‐726.

He W, Zhang G, Zhang X, Ji J, Li G, Zhao X. Recent development and application of thermoelectric generator and cooler. Appl Energy. 2015;143:1‐25.

Zhang Q, Deng K, Wilkens L, Reith H, Nielsch K. Micro‐thermoelectric devices. Nat Electron. 2022;5:333‐347.

Wu Z, Zhang S, Liu Z, Mu E, Hu Z. Thermoelectric converter: strategies from materials to device application. Nano Energy. 2022;91:106692.

Wang H, Pan Z. Performance optimization of thermoelectric devices and its dependence on materials properties. Mater Lab. 2022;2(1):220053.

Sun W, Liu WD, Liu Q, Chen ZG. Advances in thermoelectric devices for localized cooling. Chem Eng J. 2022;450:138389.

Qin Y, Qin B, Wang D, Chang C, Zhao LD. Solid‐state cooling: thermoelectrics. Energy Environ Sci. 2022;15:4527‐4541.

Zaferani SH, Sams MW, Ghomashchi R, Chen ZG. Thermoelectric coolers as thermal management systems for medical applications: design, optimization, and advancement. Nano Energy. 2021;90:106572.

Yan Q, Kanatzidis MG. High‐performance thermoelectrics and challenges for practical devices. Nat Mater. 2021;21:503‐513.

Shittu S, Li G, Zhao X, Ma X. Review of thermoelectric geometry and structure optimization for performance enhancement. Appl Energy. 2020;268:115075.

Cai Y, Mei SJ, Liu D, Zhao FY, Wang HQ. Thermoelectric heat recovery units applied in the energy harvest built ventilation: parametric investigation and performance optimization. Energy Convers Manage. 2018;171:1163‐1176.

Zhang J, Zhao H, Feng B, Song X, Zhang X, Zhang R. Numerical simulations and optimized design on the performance and thermal stress of a thermoelectric cooler. Int J Refrig. 2023;146:314‐326.

Yin T, He ZZ. Analytical model‐based optimization of the thermoelectric cooler with temperature‐dependent materials under different operating conditions. Appl Energy. 2021;299:117340.

Jeong ES. Optimization of thermoelectric modules for maximum cooling capacity. Cryogenics. 2021;114:103241.

Choi J, Dun C, Forsythe C, Gordon MP, Urban JJ. Lightweight wearable thermoelectric cooler with rationally designed flexible heatsink consisting of phase‐change material/graphite/silicone elastomer. J Mater Chem A. 2021;9:15696‐15703.

Hao J, Qiu H, Ren J, Ge Z, Chen Q, Du X. Multi‐parameters analysis and optimization of a typical thermoelectric cooler based on the dimensional analysis and experimental validation. Energy. 2020;205:118043.

Lu X, Zhao D, Ma T, Wang Q, Fan J, Yang R. Thermal resistance matching for thermoelectric cooling systems. Energy Convers Manage. 2018;169:186‐193.

Lv YG, Chu WX, Wang QW. Thermal management systems for electronics using in deep downhole environment: a review. Int Commun Heat Mass Transfer. 2022;139:106450.

Li L, Liu WD, Liu Q, Chen ZG. Multifunctional wearable thermoelectrics for personal thermal management. Adv Funct Mater. 2022;32:2200548.

Ju YS. Thermal management and control of wearable devices. iScience. 2022;25:104587.

Luo H, Zhu Y, Xu Z, et al. Outdoor personal thermal management with simultaneous electricity generation. Nano Lett. 2021;21:3879‐3886.

Luo D, Wang R, Yu W, Zhou W. Parametric study of a thermoelectric module used for both power generation and cooling. Renew Energy. 2020;154:542‐552.

Saber HH, AlShehri SA, Maref W. Performance optimization of cascaded and non‐cascaded thermoelectric devices for cooling computer chips. Energy Convers Manage. 2019;191:174‐192.

Kwan TH, Wu X, Yao Q. Bidirectional operation of the thermoelectric device for active temperature control of fuel cells. Appl Energy. 2018;222:410‐422.

Alaoui C. Thermal management for energy storage system for smart grid. J Energy Storage. 2017;13:313‐324.

Xu S, Li M, Dai Y, et al. Realizing a 10°C cooling effect in a flexible thermoelectric cooler using a vortex generator. Adv Mater. 2022;34:2204508.

Ji R, Pan T, Peng G, Ma J, Yang N, Hao Q. An integrated thermoelectric heating‐cooling system for air sterilization—a simulation study. Mater Today Phys. 2021;19:100430.

Li S, Liu J, Ding L, et al. Active thermal management of high‐power LED through chip on thermoelectric cooler. IEEE Trans Electron Devices. 2021;68:1753‐1756.

Song W, Bai F, Chen M, Lin S, Feng Z, Li Y. Thermal management of standby battery for outdoor base station based on the semiconductor thermoelectric device and phase change materials. Appl Therm Eng. 2018;137:203‐217.

Shen Y, Zhao B, Kwan TH, Yao Q. Numerical analysis of combined air‐cooled fuel cell waste heat and thermoelectric heating method for enhanced water heating. Energy Convers Manage. 2020;213:112840.

Dwivedi S, Shreevastav R, Munier F, et al. Positioning in 5G networks. IEEE Commun Mag. 2021;59:38‐44.

I CL, Han S, Bian S. Energy‐efficient 5G for a greener future. Nat Electron. 2020;3:182‐184.

Chen WY, Shi XL, Zou J, Chen ZG. Thermoelectric coolers: progress, challenges, and opportunities. Small Methods. 2022;6:2101235.

Borràs R, del Rio J, Oriach C, Juliachs J. Laser diodes optical output power model. Measurement. 2019;133:56‐67.

Chen WY, Shi XL, Zou J, Chen ZG. Thermoelectric coolers for on‐chip thermal management: materials, design, and optimization. Mater Sci Eng R: Rep. 2022;151:100700.

Slack GA, Rowe DM. CRC handbook of thermoelectrics. CRC Press; 1995.

Tang H, Bai H, Yang X, et al. Thermal stability and interfacial structure evolution of Bi2Te3‐based micro thermoelectric devices. J Alloys Compd. 2022;896:163090.

Dutt AS, Deng K, Li G, et al. Geometric study of polymer embedded micro thermoelectric cooler with optimized contact resistance. Adv Electron Mater. 2022;8:2101042.

Xu G, Duan Y, Chen X, Ming T, Huang X. Effects of thermal and electrical contact resistances on the performance of a multi‐couple thermoelectric cooler with non‐ideal heat dissipation. Appl Therm Eng. 2020;169:114933.