Silicon-Rich ONO Thickness Prediction Based on Deep Learning

The Oxide-Nitride-Oxide (ONO) structure is essential in SONOS memory devices, as it enables charge trapping and release to store information. This study introduces an innovative composite sandwich structure composed of silicon dioxide (SiO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">₂</inf>), silicon nitride (Si<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">₃</inf>N<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">₄</inf>), silicon-rich nitride (SiNx), and nitride-oxide layers. The research investigates how deposition parameters such as temperature, time, and other variables influence the thickness of the film, which directly impacts device performance and reliability. To optimize and predict film thickness accurately, the study employs the ResNet18 deep learning model. This model provides precise control over the deposition process, ensuring enhanced consistency and reliability of SONOS devices.