Fig 1.
Schematic representation of the five-layer SPR sensor architecture showing BK7 prism, silver film, graphene monolayer, black phosphorus layer, and analyte medium with their respective thicknesses.
Fig 2.
Step-by-step fabrication and integration process of the proposed five-layer SPR sensor, including substrate preparation, silver film deposition, graphene integration, black phosphorus layer fabrication, characterization, and final optical setup assembly.
Fig 3.
Reflectance spectra of the graphene-based sensor for varying Ag thicknesses at analyte RIs from 1.29 to 1.38.
Fig 4.
Reflectance spectra of the proposed sensor for analyte RI values from 1.29 to 1.38, showing variation with graphene thickness from 1 nm to 6 nm.
Fig 5.
Dependence of peak reflectance on graphene thickness for different analyte RIs, illustrating enhanced sensor response at higher thicknesses.
Fig 6.
Reflectance spectra of the proposed sensor for RIs 1.29–1.38 with varying Black Phosphorus thicknesses from 1.4 nm to 5.6 nm (steps of 1.2 nm).
Fig 7.
Corresponding reflectance variation as a function of RI for each Black Phosphorus thickness, highlighting enhanced sensitivity at higher thicknesses.
Table 1.
Showing the variation of sensor performance parameters with refractive index from 1.29 to 1.38 RIU.
Fig 8.
Performance analysis of the proposed sensor as a function of refractive index.
Fig 9.
a–c. Electric field distribution of the proposed sensor at 65°, 80°, and 89°, showing maximum reflectance with strong brown field confinement at 80°, and minimum reflectance with dominant blue regions at 65° and 89°.
Fig 10.
Scatter plot representation of KNN regression predictions versus true RI values (1.29–1.38 RIU), showing strong correlation with an optimum R2 score of 92%.
Fig 11.
Heat map visualization of KNN regression performance across test cases (0.1–0.4), with R2 scores ranging from 93% to 100%, confirming robust predictive accuracy.
Table 2.
Comparative performance analysis of the proposed KNN regression model against other ML methods for RI variation prediction.
Table 3.
Comparative analysis of SPR sensor sensitivity for different enhancement strategies at λ = 633 nm.