August_EDFA_Digital
edfas.org ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 20 NO. 3 22 APPLICATION TO A MULTILAYER SAMPLE As proof of concept, an evaluation was performed on a multilayer test chip using a 40 MHz transducer. Figure 15a shows the sample structure. It consists of a first wiring sample layer followed by a stack layer com- prisedof die attachment paste (DAP), silicon, and 1 µm thick aluminum. The thickness of silicon is 60 µm and the thickness of DAP is 30µm.Thewiringonthefirstlayerhasalength of 500 µmandwidth of 4 µm. This stack layer repeats to form the multilayer architecture as shown in Figs. 15b and 15c. SOBIRCH results are shown in Fig. 16. This method is able to detect the wiring resistive signals at the first layer, even with three stacked layers above. CONCLUSION Instead of using light stimulation based on conventional IR-OBIRCH, ultrasound stimula- tion is proposed as an alternative. The basic principle of resistance change is similar. Although the lateral spatial resolution is not as defined as IR-OBIRCH, the biggest advantage of SOBIRCH lies in its ability to stimulate inter- nal wiring through the mold resin. The series of studies presented in this article prove its eectiveness to detect a resistance change in metal wiring through the package mold upon ultrasound stimulation. The feasibility for multilayer sample application is further substantiated. Additional enhancements to the approach along with detailed evaluations of actual failed samples are now in progress. The ultimate goal is to improve this novel failure analysis technique. ACKNOWLEDGMENT The authorswould like to thank YoshikazuMatsumoto, Renesas Engineering Services, for sample preparation and Dr. Naohiro Hozumi, Toyohashi University of Technology, for support of this article. REFERENCES 1. K. Nikawa and S. Inoue: “New Laser Beam Heating Methods Applicable to Fault Localization and Defect Detection in VLSI Devices,” Proc. Int. Rel. Phy. Symp., 1996, p. 346-354. 2. T.Nonaka, Y.Hayakawa, S.Takeda, T.Igarashi, Y.Takishita, T.Miyata, H.Yamamoto, S.Yamaguchi, and H.Nishimori: “Ultrasonic Imaging and Data Processing Techniques for Semiconductor and New Materials,” Proc. 12th World Confer. NDT, Amsterdam, April 1989, p. 784-789. 3. Kaori Sakai, Osamu Kikuchi, Kaoru Kitami, Masamichi Umeda, and Sigeru Ohno: “Defect Detection Method using Statistical Image Processing of Scanning Acoustic Tomography,” IEEE 23rd Proc. Int. Symp. Phys. Fail. Anal. Integr. Circuits (IPFA), 2016, p. 293. ABOUT THE AUTHOR Toru Matsumoto received B.E. degrees in image science fromChiba University, Japan, in 1988. He joined Hamamatsu Photonics K.K. in 1988 and worked on the development of the inspection system using image sensors. Since 2004, he has been working on the development of failure analysis systems for semiconductor devices. Matsumoto developed the scanning laser SQUID micro- scope, laser THz emission microscope (LTEM), and nano electrostatic field probe sensor (NEPS) for failure analysis. He also initiated development of the failure analysis systemusing ultrasonic stimulation discussed in this article. Fig. 15 Structure of multilayer chip sample. Fig. 16 SOBIRCH images of multilayer chip sample. (a) (b) (c) (a) (b) (c)
Made with FlippingBook
RkJQdWJsaXNoZXIy MjA4MTAy