Fraunhofer-Gesellschaft

Publica

Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Delamination Detection in an Electronic Package by Means of a Newly Developed Delamination Chip Based on Thermal Pixel (Thixel) Array

 
: Kumar, A.; Schulz, M.; Sheva, S.; Keller, J.; Bader, V.; Wöhrmann, M.; Bauer, J.; May, D.; Wunderle, B.

:

TU Dresden, Institut für Aufbau- und Verbindungstechnik der Elektronik -IAVT-; Institute of Electrical and Electronics Engineers -IEEE-; International Microelectronics and Packaging Society -IMAPS-:
7th Electronic System-Integration Technology Conference, ESTC 2018. Proceedings : 18th to 21st Sept. 2018, Dresden, Germany
Piscataway, NJ: IEEE, 2018
ISBN: 978-1-5386-6814-6
ISBN: 978-1-5386-6813-9
ISBN: 978-1-5386-6815-3
S.243-249
Electronic System-Integration Technology Conference (ESTC) <7, 2018, Dresden>
Englisch
Konferenzbeitrag
Fraunhofer IZM ()

Abstract
We have made advancement towards developing our novel and non-destructive system for in-situ condition monitoring and detection of delamination of interfaces within electronic packages. A matrix of 5×5 solder based Thixels has been designed within each of the 4 Quadrants of a flip chip with a 10×10mm 2 Silicon die. Hardware layout based on a FCOB approach was designed. Before the production and assembly of the FCOB along with the testing system, a finite element study was performed to make a feasibility check by using a UBM of SiO 2 on the Silicon side. Afterwards, two batches of FCOB assemblies were produced. One set with underfill and the other without. These were then tested upon a multiplexer based self-built measurement system with a software based lock-in algorithm to receive 3 omega output signals. The change in the 3 omega voltage was successfully measured in the Thixel and the results depict a good SNR at the frequency of optimum sensitivity of 500 Hz. However, we are currently able to measure one quadrant at a time and in future have the possibility to extend this to at least two, if not more. Also, with the good SNR value that has been achieved, we can further decrease the measurement time per sensor to less than 1.25 s by varying sampling parameters.

: http://publica.fraunhofer.de/dokumente/N-581480.html