Time autocorrelation function analysis of master equation and its application to atomic clusters.

We derive the energy fluctuation Delta(2)E, and the time autocorrelation kappa(tau) and its Fourier transformation--the fluctuation spectra S(omega)--of the master-equation transition matrix. The contribution from each eigenmode of the transition matrix to these fluctuation quantities reveals the relevant importance of the individual mode in the relaxation processes. The time scales associated with these relaxation processes are determined by the corresponding eigenvalues. Unlike traditional time evolution analysis, the autocorrelation function and fluctuation spectra analysis does not involve an arbitrary initial population. It is also more suitable for analyzing the underlying dynamic, kinetic behavior near the equilibrium and the behavior of the long-time-scale rare events. We utilize our technique to analyze the solid-liquid phase coexistence of the 13-atom Morse cluster and the fcc-to-icosahedral structure transition of the 38-atom Lennard-Jones cluster. For the processes studied, the fluctuation spectra from the master equation simplify the analysis of the transition matrix, and the important relaxation modes are easily extracted.