Machine-Learning-Assisted Buried-Window FET Sensors for High-Reliability and High-Sensitivity Applications

Area of Science:

• Nanotechnology
• Electronics Engineering
• Biomedical Engineering

Background:

• Label-free biosensing is crucial for early disease detection.
• Existing biosensors face challenges in sensitivity and scalability.
• Junctionless Field-Effect Transistors (JLFETs) show potential for biosensing.

Purpose of the Study:

• To introduce a novel Double Buried-Window Junctionless Field-Effect Transistor (DBW-FET).
• To enhance sensitivity and performance for label-free biosensing applications.
• To optimize the DBW-FET design using machine learning.

Main Methods:

• Device fabrication and characterization (simulated).
• Numerical simulations using Silvaco TCAD Atlas.
• Machine learning-assisted optimization (Gaussian Process Regression and Bayesian Optimization).

Main Results:

• DBW-FET demonstrated higher drain current and lower subthreshold swing than conventional JLFETs.
• Dual buried windows improved electrostatic coupling and biomolecular sensitivity.
• Optimized DBW-FET achieved a 20-25% improvement in current sensitivity.

Conclusions:

• The DBW-FET presents a promising architecture for high-sensitivity biosensing.
• The device is compatible with Complementary Metal-Oxide-Semiconductor (CMOS) technology.
• DBW-FETs offer a pathway for next-generation nanoscale biosensors.