Scholarlink

<p>This research presents a terahertz sensor for Alpha-fetoprotein (AFP) detection based on copper metasurfaces. The device's structure incorporates multiple metallic layers deposited on silicon dioxide, with a unique combination of components: an E-shaped gold resonator working in tandem with a circular copper element and a gold ring resonator. Simulations conducted via COMSOL Multiphysics demonstrate remarkable sensing capabilities, with the device achieving 1000 GHzRIU<sup>−1</sup> sensitivity, a Figure of Merit of 15.152 RIU⁻<sup>1</sup>, and a Detection Limit of 0.125. The optimized performance was done by examining various parameters including graphene's chemical potential, angles of incident waves, and dimensional aspects of the structure. The sensor exhibited consistent performance within the 0.94–1.02 THz frequency range, maintaining transmittance above 73%. Additionally, a stacking ensemble machine learning approach is leveraged to improve the sensor performance, achieving exceptional accuracy with coefficient of determination values ranging from 97 to 100%.</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 Metasurface Terahertz Biosensing Platform for Accurate Alpha-Fetoprotein Detection in Liver Cancer Diagnosis Enhanced with Machine Learning Optimization

  • 期刊论文

  • 发表日期：

    2025年4月17日

查看全文
  • Jacob Wekalao,
  • Abdulkarem H. M. Almawgani,
  • Refka Ghodhbani,
  • Yousif S. Adam,
  • Hussein S. Gumaih,
  • Naim Ben Ali,
  • Shobhit K. Patel

摘要

This research presents a terahertz sensor for Alpha-fetoprotein (AFP) detection based on copper metasurfaces. The device's structure incorporates multiple metallic layers deposited on silicon dioxide, with a unique combination of components: an E-shaped gold resonator working in tandem with a circular copper element and a gold ring resonator. Simulations conducted via COMSOL Multiphysics demonstrate remarkable sensing capabilities, with the device achieving 1000 GHzRIU−1 sensitivity, a Figure of Merit of 15.152 RIU⁻1, and a Detection Limit of 0.125. The optimized performance was done by examining various parameters including graphene's chemical potential, angles of incident waves, and dimensional aspects of the structure. The sensor exhibited consistent performance within the 0.94–1.02 THz frequency range, maintaining transmittance above 73%. Additionally, a stacking ensemble machine learning approach is leveraged to improve the sensor performance, achieving exceptional accuracy with coefficient of determination values ranging from 97 to 100%.