.png){kind=icon}
.png){kind=icon}
Return
Advanced THz metasurface biosensor for label-free amino acid detection optimized with stacking ensemble algorithm
.png){kind=icon}
PRE
AI
DOI：10.1016/j.physe.2025.116287.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
This paper presents an advanced terahertz metasurface biosensor platform for real-time, label-free detection of amino acids. The biosensor incorporates F-shaped resonator design utilizing a hybrid material composition of graphene, gold, and silver on a silicon dioxide substrate. Computational modelling via COMSOL Multiphysics demonstrates exceptional sensitivity metrics of up to 1000 GHz/RIU and a figure of merit (FOM) of 33.333 RIU- 1 within the 0.1THz-0.6 THz frequency range. Systematic parametric optimization, including variations in graphene chemical potential (0.1eV-0.9 eV), incident angle (0 degrees-80 degrees), and resonator dimensions, ensures robust detection performance across diverse operational conditions. The biosensing capabilities are further enhanced through implementation of a stacking ensemble machine learning model, which achieves optimal prediction accuracy with an R2 score of 100 % across multiple parameters. The proposed biosensor operates on physical transduction principles, detecting amino acids through resonance frequency shifts corresponding to local refractive index variations, eliminating the need for biochemical tags, enzymes, or antibody-based recognition elements. With its exceptional sensitivity, tunable design parameters, and compatibility with scalable fabrication techniques, the proposed biosensor design represents a significant advancement with potential applications spanning biomedical diagnostics, environmental monitoring, and food safety assessment. The integration of advanced machine learning frameworks further positions this technology as a promising platform for nextgeneration biomolecular sensing.
Keywords:
Terahertz
Biosensor
Amino acid
Detection
Graphene
Machine learning
Journal
P
IF:
2.9
Papers:
218
Citations:
1.1W
Organization
View more organizations