Scholarlink

<p>This study presents the design and computational simulation of a graphene-based sensor optimized for bio-alcohol detection within the terahertz frequency range. The sensor architecture incorporates gold-based metasurfaces featuring T-shaped resonators, in conjunction with advanced materials including graphene, titanium dioxide, and black phosphorus. Through systematic parametric studies and iterative optimization, we have demonstrated exceptional sensor performance across multiple frequency bands. Quantitative analysis exemplifies significant enhancements in sensitivity and signal-to-noise ratio (SNR) across various frequency ranges. Notably, sensitivity peaks at 400 GHzRIU<sup>⁻1</sup> within the 1.25–1.5 THz band. Additional performance metrics include a quality factor of 13.699, a figure of merit (FOM) of 4.348 RIU<sup>⁻1</sup>, and a detection limit of 0.109 RIU. To further enhance the accuracy of the proposed sensor and reduce computational time, we have implemented a machine learning optimization framework based on the XGBoost algorithm. The resulting model demonstrates perfect coefficient of determination (R<sup>2</sup>) scores of 1.00 across all cases considered. The proposed sensor exhibits considerable potential for diverse applications, including forensic analysis, healthcare diagnostics, and food safety assessment.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Graphene-Based SPR Sensor Design for Bio-Alcohol Detection in the Terahertz Regime with Machine Learning Optimization Using XGBoost Regressor

  • 期刊论文

  • 发表日期：

    2025年1月10日

查看全文
  • Jacob Wekalao,
  • Shobhit K. Patel,
  • Ashok Panchapakesan,
  • Fahad Ahmed Al-Zahrani

摘要

This study presents the design and computational simulation of a graphene-based sensor optimized for bio-alcohol detection within the terahertz frequency range. The sensor architecture incorporates gold-based metasurfaces featuring T-shaped resonators, in conjunction with advanced materials including graphene, titanium dioxide, and black phosphorus. Through systematic parametric studies and iterative optimization, we have demonstrated exceptional sensor performance across multiple frequency bands. Quantitative analysis exemplifies significant enhancements in sensitivity and signal-to-noise ratio (SNR) across various frequency ranges. Notably, sensitivity peaks at 400 GHzRIU⁻1 within the 1.25–1.5 THz band. Additional performance metrics include a quality factor of 13.699, a figure of merit (FOM) of 4.348 RIU⁻1, and a detection limit of 0.109 RIU. To further enhance the accuracy of the proposed sensor and reduce computational time, we have implemented a machine learning optimization framework based on the XGBoost algorithm. The resulting model demonstrates perfect coefficient of determination (R2) scores of 1.00 across all cases considered. The proposed sensor exhibits considerable potential for diverse applications, including forensic analysis, healthcare diagnostics, and food safety assessment.