Scholarlink

<p>This investigation presents a graphene-based metasurface sensor operating in the terahertz frequency domain for non-invasive glucose detection. The sensor architecture incorporates a hybrid plasmonic structure that synergistically combines graphene with noble metals in a multi-resonator configuration, exploiting their complementary electromagnetic characteristics to enhance detection capabilities. Finite element method simulations demonstrate exceptional performance metrics&#xa0;such as maximum sensitivity of 1000 GHzRIU⁻<sup>1</sup>, figure of merit of 13.514 RIU⁻<sup>1</sup> and detection limit of 0.145 RIU across the clinically relevant refractive index range. The sensor exhibits robust spectral stability across multiple terahertz bands with consistent quality factors (Q = 9.392–9.459). Also a one-dimensional convolutional neural networks is leveraged for performance prediction, achieving perfect correlation (<i>R</i><sup>2</sup> = 100%) across operational parameters. This work represents a significant advancement in non-invasive glucose monitoring technology with considerable potential for clinical diabetes management and broader biosensing applications.</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

Ultra-Sensitive Graphene-Metal Hybrid Metasurface for Non-Invasive Glucose Detection with Convolutional Neural Network Integration

  • 期刊论文

  • 发表日期：

    2025年5月10日

查看全文
  • R. Mahalakshmi,
  • Jacob Wekalao,
  • M. Ramkumar Raja,
  • S. Arul Jothi

摘要

This investigation presents a graphene-based metasurface sensor operating in the terahertz frequency domain for non-invasive glucose detection. The sensor architecture incorporates a hybrid plasmonic structure that synergistically combines graphene with noble metals in a multi-resonator configuration, exploiting their complementary electromagnetic characteristics to enhance detection capabilities. Finite element method simulations demonstrate exceptional performance metrics such as maximum sensitivity of 1000 GHzRIU⁻1, figure of merit of 13.514 RIU⁻1 and detection limit of 0.145 RIU across the clinically relevant refractive index range. The sensor exhibits robust spectral stability across multiple terahertz bands with consistent quality factors (Q = 9.392–9.459). Also a one-dimensional convolutional neural networks is leveraged for performance prediction, achieving perfect correlation (R2 = 100%) across operational parameters. This work represents a significant advancement in non-invasive glucose monitoring technology with considerable potential for clinical diabetes management and broader biosensing applications.