Electrolyte-Gated Graphene Field Effect Transistor-Based Ca²⁺ Detection Aided by Machine Learning.

Area of Science:

• Materials Science
• Sensor Technology
• Analytical Chemistry

Background:

• Flexible electrolyte-gated graphene field-effect transistors (Eg-GFETs) offer advantages like fast response and low cost for sensing applications.
• Bias-voltage-induced uncertainties in sensitivity and response range hinder the development and reliability of Eg-GFET sensors.
• Accurate and calibration-free detection methods are crucial for advancing Eg-GFET sensor technology.

Purpose of the Study:

• To develop a machine learning-based data analysis method for Eg-GFET sensors to overcome bias-voltage-induced uncertainties.
• To demonstrate the efficacy of this method using calcium ion (Ca²⁺) detection as a proof-of-concept.
• To enable calibration-free concentration determination for Eg-GFET sensors.

Main Methods:

• Fabrication and characterization of Eg-GFET sensors for Ca²⁺ detection.
• Measurement of transfer and output characteristics of the prepared Eg-GFET-Ca²⁺ sensors.
• Training and optimization of eight regression models using machine learning algorithms, including random forest, on sensor transfer curves.

Main Results:

• The optimized random forest model effectively analyzed Eg-GFET sensor data.
• The proposed machine learning method enabled calibration-free determination of Ca²⁺ concentration.
• A strong correlation (close to y = x) was observed between estimated and real Ca²⁺ concentrations.

Conclusions:

• The machine learning-based approach successfully addresses bias-voltage-induced uncertainties in Eg-GFET sensors.
• This method offers a pathway to accurate and simplified calibration-free sensing with Eg-GFETs.
• The study highlights the potential of machine learning to enhance the performance and applicability of flexible graphene-based sensors.