Plasmonic SPR biosensor with bayesian regression for non-invasive protein biomarker detection.

Area of Science:

• Terahertz (THz) photonics
• Metasurface sensing technology
• Biomarker detection

Background:

• Protein biomarkers are crucial indicators for various diseases, including neurological disorders.
• Early and accurate detection of these biomarkers is essential for timely diagnosis and treatment.
• Existing detection methods often face limitations in sensitivity, speed, or scalability.

Purpose of the Study:

• To design and demonstrate a novel metasurface sensor for sensitive protein biomarker detection.
• To investigate the sensor's performance characteristics, including sensitivity, stability, and tunability.
• To explore the application of machine learning for optimizing sensor performance and data analysis.

Main Methods:

• Fabrication of a metasurface sensor using four figure-eight-shaped MXene resonators coated with black phosphorus and graphene.
• Performance characterization using terahertz (THz) spectroscopy to measure sensitivity and frequency response.
• Analysis of sensor stability across a frequency range and varying incident angles.
• Tuning of sensor properties via graphene chemical potential modulation.
• Application of Bayesian Ridge Regression for machine learning-based optimization and prediction.

Main Results:

• Achieved a competitive sensitivity of 395 GHz/RIU with excellent linearity for protein biomarker concentration (R² = 0.956).
• Demonstrated stable sensor performance within the 0.31-0.46 THz range and at incident angles up to 30°.
• Showcased tunable transmittance via graphene chemical potential modulation.
• Machine learning models achieved high predictive accuracy for refractive index variations (R² ≈ 86%) and angular dependencies (R² ≈ 96%).

Conclusions:

• The designed metasurface sensor offers a promising platform for rapid, sensitive, and scalable detection of protein biomarkers.
• The integrated photonic-microfluidic approach holds significant potential for early detection of neurological disorders, such as brain tumors.
• The combination of metasurface technology and machine learning enhances sensing capabilities for complex biological analyses.