The article gives a new approach for the consideration of the precision reachable by rotating-analyzer ellipsometers. Presented results approve all qualitative conclusions of the previous article on this topic, but the general equations and numerical results are different. Additionally this article treats the problem of optimizing the measurements of ellipsometric angles separately. It is shown, measured (Psi) and (Delta) are correlated for such type of ellipsometers. For most real in practice short-noise-limited case a good rule of thumb is proposed, which gives the compromise for suboptimize measurements as (rho) , so (Psi) and (Delta) at the same time. It is proved, as all sources of noise work simultaneously, as well as for null-ellipsometry it is impossible to make measurements at the Brewster angle with rotating-analyzer ellipsometers.
|