Stochastic Resonance in Colloidal Suspensions

Stochastic Resonance, a cooperative effect in nonlinear systems, arises through the interplay of noise and a weak signal. The standard example for a system showing Stochastic Resonance is an overdamped particle in a modulated double-well potential. In this work such a model system was realized experimentally by creating a double-well potential for a colloidal particle using optical tweezers. In contrast to other experiments this system can be characterized well enough to allow a quantitative comparison of the experimental results with simulations of simple model dynamics. In a single double-well system a particle's trajectory was observed for different modulation periods and modulation amplitudes. Using these measurements we compared three criteria wich allow the characterization of Stochastic Resonance in a system with constant noise and variable modulation time: the area under the first peak of the residence time distribution, the area of the hysteresis loops and the mean time during which the phase of the modulation and the phase of the particle's motion are synchronized. For sufficiently high modulation amplitudes the latter turned out to be the most sensitive criterion. The key experiment in this work are measurements on a system of four coupled particles where each particle fluctuates in a modulated double-well potential. In this system it was shown that with increasing coupling strength the resonance can be enhanced and its position can be shifted towards longer modulation periods. This is a first step towards the investigation of complex systems on a microscopic scale. The dissertation concludes with experiments on Resonant Activation. While for Stochastic Resonance the synchronization between the signal and the particle's motion is vital, the essential point for Resonant Activation is the efficiency of the barrier crossing process. We investigated both phenomena in the same system. We show that for our example the 'resonance frequency' for Resonant Activation is significantly higher than the one for Stochastic Resonance.