| dc.description.abstract | In recent years, the application of surface plasmon resonance to biosensors (SPR Biosensor) has become very popular in the fields of biological research, health science research,drug discovery, clinical diagnosis,environmental and agricultural monitoring, etc. It can detect changes in the refractive index of the sensor surface, and with the advantages of instant detection,lable-free,and high sensitivity, more and more commercial instruments on the market have proved the success of SPR biosensors.As SPR biosensors have made considerable progress in recent years, this paper will use the deep neural network in machine learning to learn the interference phenomena of TE waves and TM waves under surface plasmon resonance with Goos–Hänchen displacement, and establish a set of biosensor models for Goos–Hänchen displacement. After the model was established,the glycerol solution of various concentrations was dropped into the model for measurement,and its sensitivity and detection limit were measured. The sensitivity of the biosensor for Goos–Hänchen displacement can be obtained as 29590μm/RIU, the detection limit is 7.14∙〖10〗^(-6) RIU.Compared with the traditional SPR reflection intensity measurement,the detection limit is two orders of magnitude more,which not only confirms that the phase SPR measurement is better than the reflection intensity measurement, but also confirms that the increase in the sensitivity of the Goos–Hänchen displacement measurement is due to the singular phase delay at resonance. And because the Goos–Hänchen displacement is closely related to the phase, the Goos–Hänchen displacement SPR biosensor also provides another phase-based SPR measurement method, which greatly simplifies the use of optical components compared to the traditional phase-based measurement. | en_US |