热导率
材料科学
石墨
复合材料
热撒布器
弹性模量
散热膏
散热片
热传导
聚丁二烯
机械工程
聚合物
共聚物
工程类
作者
Junwei Li,Yuexing Zhang,Ting Liang,Xue Bai,Yunsong Pang,Xiangliang Zeng,Qinghua Hu,Wendian Tu,Zhenqiang Ye,Guoping Du,Rong Sun,Xiaoliang Zeng
标识
DOI:10.1021/acs.chemmater.1c03275
摘要
With the development of microprocessors toward higher power, larger chip, and higher frequency, heat dissipation is one of the central issues. Thermal interface materials (TIMs), which are used between the chip and the heat spreader and between the heat spreader and the heat sink, play an increasingly important role in microprocessor cooling. Currently, most of the research has primarily dealt with understanding the thermal conductivity of TIMs. For thermal design, elastic compliance is also important because excellent elastic compliance can reduce thermal contact resistance and relieve the warpage failure caused by stress concentration. However, high thermal conductivity and excellent elastic compliance are usually mutually exclusive in TIMs. Herein, we report a TIM made from vertically oriented graphite and polybutadiene that shows high through-plane thermal conductivity of 64.90 W·m–1·K–1, excellent elastic compliance with only 93 kPa stress at 50% compressive strain similar to soft biological tissues, and outstanding compression resilience performance (storage modulus 220 kPa and mechanical loss factor 0.226). These excellent properties result from the vertical orientation of graphite films in polybutadiene, strong interfacial strength between graphite films and polybutadiene, and the minimized negative impact of graphite on the intrinsic mechanical properties of polybutadiene by means of cross-stacking techniques. The optimal TIM is applied in CPU microprocessor cooling and exhibits superior heat dissipation capability, by up to 158 °C reduction of chip temperature comparing with polybutadiene. This work provides a high-performance TIM to meet specific requirements for high-performance computing, such as GPU, AI computing, and cloud computing.
科研通智能强力驱动
Strongly Powered by AbleSci AI