Multi-experiment linear data analysis for ATR biosensors

The biosensors based on surface plasmon resonance (SPR) are often used as tools for directly detecting the kinetic interaction of unlabelled biological molecules at surface in real time. With the measured SPR reflection spectrum, we can detect a shift in the location and quantity of the reflection spectrum minimum and the half width at half maximum due to the change in the thickness or the refractive index of a thin dielectric film layer. The interested parameters of analyte layer or monolayer, like the molecular size and concentration, can be determined either with analytical approaches or linear data analysis approaches. Depends on the number of parameters need to be resolved, we may need either multiple spectra (two color method) or only one sensing spectrum under the assumption that the other film parameter is given for multiple parameters case. Although it is possible to estimate multiple parameters from only one sensing spectrum by linear estimation techniques, it suffers from not only the shortcoming for larger variance in the estimates from those techniques than that of multiple spectra method but also the difficulty for choosing the appropriate initial value in the estimation process. In this paper, we propose a modified analytic approach to attain suitable initial parameters that close enough to the exact value. Furthermore, we incorporated multi-experiment method into linear estimation algorithms to determine the optimal estimated parameters with smaller variability of the estimated parameters. In that manner, it would be benefit to reject the colored noise accidentally results from experiment process. The experimental data with the multi-experiment linear data analysis demonstrates that it has ability to sense slightly index change in consequence of argon gas flow through the nitrogen.