Semiconductor lasers are particularly well suited for the implementation of pump-noise suppression, yielding a reduction of the intensity noise in the laser. In this simple picture, the maximal amount of squeezing is equal to the quantum efficiency. However, experimental results on intensity noise reduction by pump-noise suppression are usually above this limit. This discrepancy suggests that additional noise sources must be involved. Here we successfuly interpret the full noise behavior of a single-mode laser diode far above threshold by considering two excess noise sources: the leakage current fluctuations across the laser and the Petermann excess noise. We have estimated the contribution of each noise source using the results of the correlations between the laser output intensity noise and the voltage fluctuations across the laser diode (light-voltage correlations) and obtained good agreement between our theory and experimental results.
We derive the equation of motion for the quantum-mechanical radiation density matrix of a gas laser to lowest order in the dimensionless coupling constant. Our derivation is fully quantum mechanical and we can calculate the coherence properties from the radiation density matrix. Our model consists of
N two-level systems interacting with radiation in a cavity in the presence of dissipation, pumping, and collisions. The method we use is a generalization of the Bogoliubov derivation of the kinetic equation for a small parameter. Our derivation holds for any physically realizable pump power, and near threshold reduces to Lamb's near-threshold theory. With the equation of motion for the radiation density matrix, we obtain solutions both for when the average field is nonzero and for when it is zero. The steady-state electromagnetic density is the same in both cases except for a small spontaneous-emission term. We show that the reason gas lasers do not satisfy rate equations is the existence of zeroth-order correlations between the internal atomic variables and atomic center-of-mass variables. It is these same zeroth-order correlations which are responsible for the Lamb dip. Our derivation includes collisions and reduces the calculations of their effect to quadrature.