标题：Refractive Index Estimation from Spectral Measurements of a Plasmonic Glucose Sensor and Wavelength Selection * * The project was funded by Baden-Württemberg Stiftung gGmbH. The authors would also like to thank MWK BW, ERC COMPLEX-PLAS and AvH Stiftung.
摘要：AbstractNoninvasive glucose monitoring is a desired objective in diabetes therapy and monitoring the glucose levels in the tear fluid is one possibility to approach this objective. The plasmonic glucose sensor presented within this work consists of metal nanostructures which cause a resonance in the optical response and the resonance wavelength position depends on the refractive index of the surrounding. Glucose selectivity is guaranteed by functionalizing the sensor with a hydrogel that swells with glucose and thus changes its refractive index. The focus of this work lies on the relationship between the refractive index and the resulting transmission spectra of the sensor. Based on this understanding, an estimation algorithm is presented, which provides a refractive index value for any measured spectrum. It is shown that this new estimation algorithm improves previous results. For a mobile application of the sensor, the transmission can only be measured at few discrete wavelengths rather than the whole spectrum. The choice of these discrete wavelengths has a huge influence on the estimation results. The developed estimation algorithm is used to build an objective function to determine the best wavelengths via sequential forward feature selection. The results of the feature selection show that even reducing the discrete wavelengths to a number of four, refractive index estimation can be performed with a relative error below 0.05%.
