Mapping systematic and random process variations using Light emission from Off-State Leakage

In this paper, we present for the first time the use of an optical imaging technique that we have developed to map systematic and random variability effects across several phases of development of a new microprocessor chip fabricated with the latest IBM SOI 65 nm technology. The technique relies on the detection and subsequent analysis of the light emission from off-state leakage current (LEOSLC) by means of static imaging cameras, mounted on a microscope or other optical collection optics. The variability maps created using this method are very simple to generate and have been successfully verified by comparing them to data obtained by on-chip electrical sensors such, as process sensitive ring oscillators (PSROs), power supply currents, and test results. The visual representation is very helpful in picking up systematic effect and to measure the correlation distance of the variations.