Abstract
Certain device failures are especially difficult to analyze since they can only be reproduced under high speed and high power conditions, while also requiring the removal of standard heat dissipating packaging to get visual access to the chip. In addition to the challenge of heat generation density of devices increasing year by year, small hot spots in actual usage generate heat far in excess of the average, and heat dissipation performance needs to be more efficient and highly uniform. In addition, it is desirable to implement a cooling system that does not overly restrict the number and types of lenses that can be used, such as high and low magnification air gap lenses as well as a solid-immersion lens, which has been one of the challenges of existing systems. This paper reports on the development of a cooling system to address these challenges and to enable failure analysis on a device running at 200 W.
