Efficient Characterization Methodology for Low-Frequency Noise Monitoring

In this work, a novel methodology to characterize the Low-Frequency Noise LFN on large device statistics and suitable for production monitoring is proposed. The maximum drain current fluctuations over time are measured. The slope of the LFN distribution is modeled with physical equations related to the basic device properties. The approach is validated by studying the impact of transistor geometry (different gate width, number of fingers and gate length) as well as gate oxide thickness and characterization temperature. In conclusion, the proposed methodology is tested evaluating different process integration elements. The outcome is compared to classical LFN read-out for final confirmation.