Bias dependence in statistical random telegraph noise analysis based on nanoscale CMOS ring oscillators

Area of Science:

• Semiconductor Device Physics
• Reliability Engineering
• Nanoscale Electronics

Background:

• Random Telegraph Noise (RTN) is a critical reliability issue in advanced nanoscale complementary metal-oxide semiconductor (CMOS) technologies.
• Understanding RTN is essential for ensuring the long-term performance and stability of integrated circuits.

Purpose of the Study:

• To characterize the temporal and spectral properties of RTN in 40nm CMOS technology.
• To statistically analyze RTN amplitude and frequency behavior under varying supply voltages.
• To extract capture and emission time constants of RTN.

Main Methods:

• Utilized Ring Oscillators (ROSCs) for RTN characterization.
• Employed arrays of 128 identical ROSC cells for comprehensive measurements.
• Conducted measurements across a supply voltage range of 0.5V to 0.7V.

Main Results:

• Identified dominant RTN amplitude (>0.37%) in 60% of cells at 0.65V supply.
• Characterized RTN amplitude strength and frequency behavior statistically.
• Extracted RTN capture and emission time constants ranging from 0.2μs to 10ms.

Conclusions:

• RTN is a prevalent phenomenon in 40nm CMOS technology, affecting a significant portion of devices.
• The study provides valuable statistical data on RTN characteristics, crucial for reliability modeling.
• The extracted time constants offer insights into the physical mechanisms underlying RTN in these devices.