Investigation of acoustic emission as a non-invasive method for detection of power semiconductor aging

Pooya Davari, Ole Damm Kristensen, Francesco Iannuzzo

Research output: Contribution to journalConference article in JournalResearchpeer-review

8 Citations (Scopus)
108 Downloads (Pure)

Abstract

In this paper, recording of acoustic emission during real operations is used for non-invasively detecting the aging of a power seminconductor module due to power cycling. The presented method is very simple and particularly attractive because of non-invasiveness and potential low-cost features, which can enable straightforward adoption in condition monitoring of power electronic devices. Nevertheless, a spectrum analysis is needed to process the acquired data. Experimental results show a strong correlation between acoustic emission and on-state voltage drop, which is the standard indicator of degradation in bond wires. A comparison with the results obtained with another identical module gave excellent repeatability, confirming that the method is very promising for real application.
Original languageEnglish
JournalMicroelectronics Reliability
Volume88-90
Pages (from-to)545-549
Number of pages5
ISSN0026-2714
DOIs
Publication statusPublished - Sep 2018
Event29th European Symposium on Reliability of Electron Devices, Failure Physics and Analysis - AKKC, Aalborg, Denmark
Duration: 1 Oct 20185 Oct 2018
Conference number: 29th
http://www.esref2018conf.org/

Conference

Conference29th European Symposium on Reliability of Electron Devices, Failure Physics and Analysis
Number29th
LocationAKKC
Country/TerritoryDenmark
CityAalborg
Period01/10/201805/10/2018
Internet address

Keywords

  • Power semiconductor aging
  • acoustic emission
  • power cycling
  • frequency domain analysis
  • Wearout
  • Non-invasive wear-out detection
  • IGBT
  • condition monitorin

Fingerprint

Dive into the research topics of 'Investigation of acoustic emission as a non-invasive method for detection of power semiconductor aging'. Together they form a unique fingerprint.

Cite this