Statistical Characterization and On-Chip Measurement Methods for Local Random Variability of a Process Using Sense-Amplifier-Based Test Structure

摘要

Measurement, characterization, and estimation of local random variation in threshold voltage (V{sub}t) are crucial for yield learning/enhancement in nanoscale technologies, particularly for characterization of area-constrained circuits, such as SRAM cells. Conventionally, local random mismatch is characterized by measurement of current difference between identical neighboring devices, followed by complex data analysis to extract Vt difference from current difference [1-3]. This paper demonstrates a sense-amplifier-based test circuit and measurement method for on-chip characterization of local random variation. Instead of complex and sophisticated analog voltage-current measurements required in conventional schemes, we use a simple digital measurement technique. Moreover, this method directly characterizes the complete probability distribution of local mismatch, thereby eliminating the need of complex data analysis. Further, using the digital nature of the measurement, we propose a built-in-self-test scheme for on-chip measurement of device mismatch, which significantly reduces characterization time and cost.