Share Email Print

Proceedings Paper

A neuro-fuzzy approach to failure detection and diagnosis of excimer laser ablation in microvia formation
Author(s): Ronald Setia; Gary S. May
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

Excimer laser ablation is used for microvia formation in the microelectronics packaging industry. With continuing advancement of laser systems, there is an increasing need to offset capital equipment investment and lower equipment downtime. This paper presents a neuro-fuzzy methodology for in-line failure detection and diagnosis of the excimer laser ablation process. Response data originating directly from laser tool sensors and the characterization of microvias were used as failure symptoms for potential deviations in four laser system parameters from their corresponding baseline values. The response characteristics consist of via diameter, via wall angle, and via resistance. Resistance measurements on copper deposited in the ablated vias were performed to characterize the degree to which debris remaining inside the vias affected quality. The laser system parameters include laser fluence, shot frequency, number of pulses, and helium pressure flow. The adaptive neuro-fuzzy inference system (ANFIS) was trained and subsequently validated for its capability in evidential reasoning using the data collected. Results indicated only a single false alarm occurred in 19 possible failure detection scenarios. In failure diagnosis, a single false alarm and a single missed alarm occurred.

Paper Details

Date Published: 1 March 2006
PDF: 13 pages
Proc. SPIE 6106, Photon Processing in Microelectronics and Photonics V, 610607 (1 March 2006); doi: 10.1117/12.655296
Show Author Affiliations
Ronald Setia, Georgia Institute of Technology (United States)
Gary S. May, Georgia Institute of Technology (United States)

Published in SPIE Proceedings Vol. 6106:
Photon Processing in Microelectronics and Photonics V
David B. Geohegan; Tatsuo Okada; Craig B. Arnold; Frank Träger; Jan J. Dubowski; Michel Meunier; Andrew S. Holmes, Editor(s)

© SPIE. Terms of Use
Back to Top