Thermal Characterization of Silicon Carbide MOSFET Module Suitable for High-Temperature Computationally-Efficient Thermal-Profile Prediction

Mengxing Chen, Huai Wang, Donghua Pan, Xiongfei Wang, Frede Blaabjerg

Research output: Contribution to journalJournal articleResearchpeer-review

23 Citations (Scopus)
1392 Downloads (Pure)

Abstract

This article characterizes the thermal behavior of a commercialized silicon carbide (SiC) power MOSFET module with special concerns on high-temperature operating conditions as well as particular focuses on SiC MOSFET dies. A temperature-dependent Cauer-type thermal model of the SiC MOSFET is proposed and extracted based on offline finite-element simulations. This Cauer model is able to reveal the temperature-dependent thermal property of each packaging layer, and it is suitable for the high-temperature thermal-profile prediction with sufficient computational efficiency. Due to the temperature-dependent thermal properties of the SiC die and ceramic material, the junction-heatsink thermal resistance can be increased by more than 10% under high-temperature conditions (up to 200 °C), which can considerably worsen thermal estimations of the SiC die and its packaging materials. Furthermore, the experimental measurement of transient thermal impedance was conducted under operating temperature variations (with virtual junction temperature ranging from 60.5 °C to 199.6 °C), and the effectiveness of the proposed temperature-dependent Cauer model was fully validated.

Original languageEnglish
Article number9051825
JournalIEEE Journal of Emerging and Selected Topics in Power Electronics
Volume9
Issue number4
Pages (from-to)3947 - 3958
Number of pages12
ISSN2168-6777
DOIs
Publication statusPublished - Aug 2021

Keywords

  • Computational efficiency
  • finite-element method (FEM)
  • high operating temperature
  • silicon carbide (SiC) power MOSFET module
  • temperature-dependent Cauer model

Fingerprint

Dive into the research topics of 'Thermal Characterization of Silicon Carbide MOSFET Module Suitable for High-Temperature Computationally-Efficient Thermal-Profile Prediction'. Together they form a unique fingerprint.

Cite this