Dual-channel optical fiber plasmonic sensor and machine learning approach for unknown analyte prediction

A novel, to our knowledge, optical fiber, which consists of only one solid ring and two hollow rings, is proposed for highly sensitive refractive index (RI) sensing. Two hollow silica rings, internally coated with chemically stable gold, act as the analyte channels. Metal film on either side of the core ensures sufficient field incoming toward the metal silica interface. Initially, the sensing performance is evaluated in the RI range of 1.33–1.41, where the sensor exhibits a maximum sensitivity of 23,000 nm/refractive index unit (RIU). Then, the sensor performance is evaluated for cancer cell and pathogen detection. In these cases, the sensor also displays excellent sensitivity of 7142.85 and 4878.05 nm/RIU, respectively. Moreover, a machine learning approach, called extreme gradient boosting, is incorporated to predict resonance wavelength for any RI value within the sensing range. Novel design, high performance, diverse applications, and machine learning incorporation make this sensor a potential candidate in the biosensing field.