Machine learning-assisted spectroscopic ellipsometry of chromium thin films for microalgae biosensing.

Area of Science:

• Materials Science
• Biosensing Technology
• Spectroscopy

Background:

• Rapid and label-free detection of microalgae is crucial for environmental surveillance and bio-industrial process control.
• Current methods often rely on bulk concentration, missing subtle interfacial changes vital for accurate monitoring.

Purpose of the Study:

• To develop a thickness-optimized optical transducer using chromium (Cr) thin films for microalgae biosensing.
• To implement and evaluate a spectroscopic ellipsometry (SE) and machine learning (ML) framework for rapid microalgae detection.

Main Methods:

• Deposited Cr thin films (20-75 nm) via RF magnetron sputtering and characterized them using SE.
• Extracted thickness-dependent optical constants (n, k) using a Drude-Lorentz dispersion model.
• Implemented biosensing by forming Cr/PVA and Cr/PVA + microalgae stacks and analyzed differential phase response.
• Utilized multitask deep neural networks and support vector machines for data analysis and classification.

Main Results:

• The 75 nm Cr film demonstrated optimal performance, exhibiting a stable optical response and low loss in the visible range.
• The 75 nm Cr/PVA platform showed a significant microalgae-induced phase shift (40.6° within 2.65-3.20 eV), indicating a high-contrast detection window.
• The SE-ML framework successfully converted dense SE signatures into decision-ready labels for rapid microalgae screening.

Conclusions:

• A thickness-optimized Cr thin film transducer combined with ML provides an intelligent, non-destructive method for microalgae detection.
• The developed SE-ML framework advances rapid microalgae screening and environmental diagnostics.
• This approach enables sensitive detection based on interfacial changes, crucial for process control and surveillance.