Scholarlink

<p>Accurate salinity measurement is critical for environmental monitoring, oceanography, and water-quality assessment, yet many existing optical and plasmonic sensing techniques remain limited by low sensitivity, narrow operational bandwidth, and insufficient tunability. THz metasurface biosensors have emerged as promising alternate as they produce strong electromagnetic field confinement and has the ability for label-free sensing. However, for practical deployment further improvements in terms of performance and efficiency are required. In this study, a high-performance THz metasurface sensor based on a hybrid MXene–Cu–graphene architecture is proposed in which the metasurface comprises five circular MXene resonators with radii of 1&#xa0;μm embedded within a Cu coated rings of 6–6.5&#xa0;μm radii. The complete design is placed on a 17 × 17&#xa0;μm graphene substrate. The sensing technique utilizes the complementary characteristics of MXene’s high electronic conductivity, copper’s stable plasmonic behavior, and graphene’s tunable surface chemical potential to create enhanced plasmonic confinement and resonance tunability in the sensor. Finite element electromagnetic simulations using COMSOL Multiphysics were conducted to examine the sensor response to changes in refractive index associated with water quality contaminants and salinity levels. Additionally, a one-dimensional convolutional neural network was used to predict resonance behavior and optimize structural parameters for improved design efficiency using the simulated data set. Numerical results indicate that the sensor has an ultra-high refractive index sensitivity of 1892&#xa0;GHz/RIU as well as a figure of merit of 34.4 RIU⁻¹ and a quality factor of 7.18. The machine learning model also exhibited good predictive ability (R² = 0.998) and reduced optimization time by over 40% compared to other common regression techniques. The proposed multilayer metasurface with machine-learning-assisted optimization offers an efficient and reliable design for near real-time salinity detection, providing strong potential for advanced biochemical and environmental monitoring applications. This work contributes to SDG 6 (Clean Water and Sanitation) by enabling accurate and highly sensitive salinity monitoring for sustainable water-quality management.</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

Hybrid MXene–Graphene Terahertz Metasurface with Copper-Assisted Plasmonic Confinement for Broadband, Ultra-Sensitive Salinity Sensing Using Machine Learning

  • 期刊论文

  • 发表日期：

    2026年2月19日

查看全文
  • K. Vijayakumar,
  • S. Gopinath,
  • U. Arun Kumar,
  • R. Dhivya,
  • N. Saranya

摘要

Accurate salinity measurement is critical for environmental monitoring, oceanography, and water-quality assessment, yet many existing optical and plasmonic sensing techniques remain limited by low sensitivity, narrow operational bandwidth, and insufficient tunability. THz metasurface biosensors have emerged as promising alternate as they produce strong electromagnetic field confinement and has the ability for label-free sensing. However, for practical deployment further improvements in terms of performance and efficiency are required. In this study, a high-performance THz metasurface sensor based on a hybrid MXene–Cu–graphene architecture is proposed in which the metasurface comprises five circular MXene resonators with radii of 1 μm embedded within a Cu coated rings of 6–6.5 μm radii. The complete design is placed on a 17 × 17 μm graphene substrate. The sensing technique utilizes the complementary characteristics of MXene’s high electronic conductivity, copper’s stable plasmonic behavior, and graphene’s tunable surface chemical potential to create enhanced plasmonic confinement and resonance tunability in the sensor. Finite element electromagnetic simulations using COMSOL Multiphysics were conducted to examine the sensor response to changes in refractive index associated with water quality contaminants and salinity levels. Additionally, a one-dimensional convolutional neural network was used to predict resonance behavior and optimize structural parameters for improved design efficiency using the simulated data set. Numerical results indicate that the sensor has an ultra-high refractive index sensitivity of 1892 GHz/RIU as well as a figure of merit of 34.4 RIU⁻¹ and a quality factor of 7.18. The machine learning model also exhibited good predictive ability (R² = 0.998) and reduced optimization time by over 40% compared to other common regression techniques. The proposed multilayer metasurface with machine-learning-assisted optimization offers an efficient and reliable design for near real-time salinity detection, providing strong potential for advanced biochemical and environmental monitoring applications. This work contributes to SDG 6 (Clean Water and Sanitation) by enabling accurate and highly sensitive salinity monitoring for sustainable water-quality management.