Evaluating the signal processing chain employed in surface plasmon resonance biosensing

The optical biosensor based on the Surface Plasmon Resonance phenomenon is, today, the most accurate and sensible label free biosensor. However, this high sensitivity of the SPR biosensor systems has a side effect: a high sensitivity to noise present all over the system. The ambient temperature, the position of the components, the impurities in the light path, the photo detector instrumentation, are all noise sources that can degrade the information present in the signal. Preprocessing this signal before extracting the interest parameters and even the techniques used to extract these parameters can minimize the degradation effect of the noise. This paper proposes a study to determine the best techniques to be applied that most reduces the noise degradation. A combination of techniques can improve the results or just increase the complexity with no gain in noise elimination. An influence in the techniques sequence has also influenced the result, being necessary to determine the best sequence. Also to tune the technique parameters to the optimal, a miss tuning may preserve noise that can be removed or can eliminate precious information with the noise, degenerating the signal. The paper also aims at determining the best technique combination to be applied in parameter extraction.