Noise induced fluctuations are playing a more and more important role in the behavior in devices as the devices are becoming smaller and smaller. In this thesis the Gaussian noise and non-Gaussian noise fluctuations are studied using a micro-scaled torsional oscillator. We characterize the Poisson noise by noise-induced switching in metastable systems and by studying the probability distribution in phase space. Noise acting on a bistable system causes random switching between the two metastable states, and the occupation probabilities can be very sensitive to perturbations acting on the system. We show how to detect a non-Gaussian noise in the system of interest and develop a method to reveal the third and higher moments of a perturbing non-Gaussian noise. We also analyze the p...[ Read more ]

Noise induced fluctuations are playing a more and more important role in the behavior in devices as the devices are becoming smaller and smaller. In this thesis the Gaussian noise and non-Gaussian noise fluctuations are studied using a micro-scaled torsional oscillator. We characterize the Poisson noise by noise-induced switching in metastable systems and by studying the probability distribution in phase space. Noise acting on a bistable system causes random switching between the two metastable states, and the occupation probabilities can be very sensitive to perturbations acting on the system. We show how to detect a non-Gaussian noise in the system of interest and develop a method to reveal the third and higher moments of a perturbing non-Gaussian noise. We also analyze the probability distribution in the Poisson noise driven oscillator. We consider both the overdamped and underdamped regimes for which different forms of the probability distribution are found. Furthermore, we investigate a novel phenomenon of work fluctuation in a metastable system in the presence of Gaussian noise. We find the fluctuation in this system cannot be described by conventional fluctuation theorem.