Machine Learning-Based Modeling of pH-Sensitive Silicon Nanowire (SiNW) for Ion Sensitive Field Effect Transistor...

Area of Science:

• Materials Science
• Nanotechnology
• Sensor Technology

Background:

• Silicon nanowire (SiNW)-based ion-sensitive field-effect transistor (ISFET) sensors offer advantages like portability and low cost.
• Predicting the performance of SiNW-ISFET sensors remains a challenge with limited research.
• Existing SiNW-ISFET sensors face performance limitations that require addressing.

Purpose of the Study:

• To predict the performance of silicon nanowire (SiNW)-based ISFET sensors.
• To evaluate the efficacy of machine learning techniques for SiNW-ISFET sensor performance prediction.
• To identify the most effective machine learning model for accurate sensor performance forecasting.

Main Methods:

• Utilized four machine learning techniques: multilayer perceptron (MLP), nonlinear regression (NLR), support vector regression (SVR), and extra tree regression (ETR).
• Trained and validated the machine learning algorithms using experimental measurements from SiNW-ISFET sensors.
• Compared the predictive performance of the employed machine learning models.

Main Results:

• Extra Tree Regression (ETR) exhibited superior predictive capabilities compared to MLP, NLR, and SVR.
• ETR achieved a low root-mean-square error (RMSE) of 1 × 10^-3 mA.
• ETR demonstrated a high coefficient of determination (R²) of 0.9999725, indicating excellent predictive accuracy.

Conclusions:

• Machine learning, particularly ETR, can accurately predict SiNW-ISFET sensor performance.
• This predictive approach helps overcome and correct existing problems associated with SiNW-ISFET sensors.
• The study provides a reliable method for forecasting sensor performance, advancing SiNW-ISFET technology.