Fig. 1. Cross section of designed D-shaped PCF-SPR sensor

Fig. 2. Simulation model of PCF-SPR sensor. (a) No subdivision; (b) finer mesh subdivision

Fig. 3. Dispersion relation. (a) Real part of effective refractive index of fundamental mode and SPP mode, and loss spectrum of fundamental mode; (b1)~(b5) electric field distribution of each point marked in Fig. 3(a)

Fig. 4. Loss spectrum varying with refractive index of analyte

Fig. 5. Loss spectra and resonant wavelengths under different r_s when h=4.5 μm and t=40 nm. (a) Loss spectra of analyte with different n_a; (b) resonant wavelength varying with n_a

Fig. 6. Loss spectra under different r_s. (a) Loss spectra of analyte when h=4.5 μm and n_a=1.3; (b) loss spectra of analyte when t=45 nmand n_a=1.3

Fig. 7. Loss spectra and resonant wavelengths under different h when r_s=1.0 μm and t=40 nm. (a) Loss spectra of analyte with different n_a; (b) resonant wavelength varying with n_a

Fig. 8. Loss spectrum of analyte varying with h when r_s=1.0 μm and n_a =1.3

Fig. 9. Loss spectra and resonant wavelengths under different t when r_s=1.0 μm and h=4.5 nm. (a) Loss spectra of analyte with different n_a; (b) resonant wavelength varying with n_a

Fig. 10. Loss spectra of analyte with different n_a

Fig. 11. Resonant wavelength varying with refractive index and its polynomial fitting

Fig. 12. Existing PCF-SPR sensors for analyte with low refractive index. (a) Concave-shaped structure^[19]; (b) negative curvature structure^[26]; (c) hollow-core negative curvature structure^[27]; (d) structure with indium tin oxide coating^[28]

Fig. 13. Influence of manufacturing error