Scholarlink

<p>For cancer cell detection, we present in this paper a novel optimized photonic crystal fiber (PCF)-based surface plasmon resonance (SPR) biosensor. The sensor features a V-shaped groove, which enhances sensitivity by improving electromagnetic coupling between the surface plasmon mode (SPM) and the core mode at the metal/dielectric interface. Key design parameters, including the diameter of the small air hole (d<sub>2</sub>), V-channel depth (h), and gold layer thickness (t<sub>g</sub>), are optimized through the Box-Behnken Design (BBD) method, reaching a maximum spectral sensitivity of 2142.86&#xa0;nm/RIU for blood and breast cancer cells, an amplitude sensitivity of 632.50 RIU⁻¹ for skin cancer cells, and a resolution of 4.66 × 10<sup>− 5</sup> RIU. Furthermore, we employed an artificial neural network (ANN) model based on a Multi-Layer Perceptron (MLP) to accurately predict sensor performance, demonstrating the potential of machine learning in optimizing biosensor applications for biochemical and biological detection.</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

Novel Optimization Approach and AI Performance Prediction of a Highly Sensitive V-Shaped PCF-SPR Sensor for Cancer Cell Detection

  • 期刊论文

  • 发表日期：

    2025年5月29日

查看全文
  • Randa Khemiri,
  • Sameh Kaziz

摘要

For cancer cell detection, we present in this paper a novel optimized photonic crystal fiber (PCF)-based surface plasmon resonance (SPR) biosensor. The sensor features a V-shaped groove, which enhances sensitivity by improving electromagnetic coupling between the surface plasmon mode (SPM) and the core mode at the metal/dielectric interface. Key design parameters, including the diameter of the small air hole (d2), V-channel depth (h), and gold layer thickness (tg), are optimized through the Box-Behnken Design (BBD) method, reaching a maximum spectral sensitivity of 2142.86 nm/RIU for blood and breast cancer cells, an amplitude sensitivity of 632.50 RIU⁻¹ for skin cancer cells, and a resolution of 4.66 × 10− 5 RIU. Furthermore, we employed an artificial neural network (ANN) model based on a Multi-Layer Perceptron (MLP) to accurately predict sensor performance, demonstrating the potential of machine learning in optimizing biosensor applications for biochemical and biological detection.