論文 Heat transport exploration through the GaN/diamond interfaces using machine learning potential

Zhanpeng Sun ; Yunfei Song ; Zijun Qi ; Xiang Sun ; Meiyong Liao SAMURAI ORCID ; Rui Li ; Qijun Wang ORCID ; Lijie Li ; Gai Wu ORCID ; Wei Shen ; Sheng Liu

コレクション

引用
Zhanpeng Sun, Yunfei Song, Zijun Qi, Xiang Sun, Meiyong Liao, Rui Li, Qijun Wang, Lijie Li, Gai Wu, Wei Shen, Sheng Liu. Heat transport exploration through the GaN/diamond interfaces using machine learning potential. International Journal of Heat and Mass Transfer. 2025, 241 (), 126724. https://doi.org/10.1016/j.ijheatmasstransfer.2025.126724

説明:

(abstract)

Gallium nitride (GaN) electronic devices are highly pursued for high-power and high-frequency applications because of their superior performance characteristics. Although GaN has great potential, heat management is a major obstacle to its application in high-power devices due to its inherently low thermal conductivity. One viable approach to address this issue is to combine GaN with a diamond heat sink. However, the heat transfer at interfaces has emerged as a critical challenge in effective thermal management. In this research, a neuroevolution potential (NEP) is trained that can improve the accuracy of thermal property predictions, which can effectively address the issues of imprecise thermal performance predictions for GaN/diamond heterostructures with traditional potentials. The temperature-dependent and interface atom-dependent thermal boundary resistance (TBR) of GaN/diamond heterostructures after interfacial bonding have been investigated using molecular dynamics simulations based on NEP. The TBR for the GaN/diamond heterostructures has been estimated over the temperature range of 200–600 K. At 300 K, the TBR of different interface structures ranges from 2.22 to 3.26 m2⋅K⋅GW −1, which is in perfect agreement with the value predicted by diffusion mismatch model (DMM) theory (~3 m2⋅K⋅GW −1). Furthermore, it is observed that the TBR decreases with the increasing temperature and shows an approximately linear relationship. It can also be found that the TBR of the heterostructure bonded by C atoms and N atoms at the interface is about 25 % lower than that of the heterostructure bonded by C atoms and Ga atoms. Then, the mechanism behind the above phenomenon is explained by analyzing the vibration density of states (VDOS), phonon participation ratio (PPR) and total phonon participation contribution (TPPC). Finally, the insightful optimization strategies based on temperature and GaN atomic types at the interface have been proposed, laying the groundwork for better design and management of GaN/diamond interfaces.

権利情報:

キーワード: thermal conductivity, GaN, Diamond

刊行年月日: 2025-01-16

出版者: Elsevier BV

掲載誌:

  • International Journal of Heat and Mass Transfer (ISSN: 00179310) vol. 241 126724

研究助成金:

  • Fundamental Research Funds for the Central Universities
  • National Natural Science Foundation of China

原稿種別: 査読前原稿 (Author's original)

MDR DOI: https://doi.org/10.48505/nims.5446

公開URL: https://doi.org/10.1016/j.ijheatmasstransfer.2025.126724

関連資料:

その他の識別子:

連絡先:

更新時刻: 2025-04-22 12:30:10 +0900

MDRでの公開時刻: 2025-04-22 12:25:15 +0900

ファイル名 サイズ
ファイル名 Exploring the Effect of Temperature and Interface Atom Type on Heat Transport through the GaNdiamond Interfaces using Machine Learning Potential_Zhanpeng Sun-Prof. Liao.docx (サムネイル)
application/vnd.openxmlformats-officedocument.wordprocessingml.document
サイズ 11.4MB 詳細