Data analysis for surface plasmon resonance sensors using dynamic baseline algorithm

Needs of accurate and robust responses from surface plasmon resonance (SPR) sensor systems put high requirements to the data analysis methods used. The output of an SPR sensor comprises data of reflected light intensity as function of either wavelength or angle of incidence (SPR curve). As reported in the literature, commonly used methods for analysing the SPR curve are polynomial curve fitting methods and the centroid method. These methods have their drawbacks in either being sensitive to correlated noise or drift from the light source, or generating high noise on the SPR response. We propose a modified analysis algorithm, where a baseline for selection of data from the SPR curve is adjusted dynamically according to a pre-defined ratio between the areas of the SPR curve below and above the baseline. It is shown in a mathematical model, numerical simulations as well as in experiments that the proposed dynamic baseline algorithm is insensitive to optical power correlated noise/drift. The algorithm is mathematically simple and it can be readily combined with commonly used data analysis methods. (c) 2004 Published by Elsevier B.V.