ONLINE SIMULATION AS A METHODOLOGY TO RECOVER SIZE DISTRIBUTIONS FROMDYNAMIC LIGHT-SCATTERING EXPERIMENTS

It is generally recognized that fundamental limitations make data anal ysis for dynamic light scattering (DLS) not straight-forward. In addit ion to experimental problems such as multiple scattering, number fluct uation and noise, there are intrinsic problems. Data analysis is a so- called reverse problem which, owing to the mathematical equations, bec omes ''allergic'' to noise (an ill-posed inverted problem). In an atte mpt to overcome at least some of these limitations, a software solutio n was developed. This mainly aimed to implement a tool that makes it e asy to evaluate physically plausible solutions, rather than to determi ne many (less accurate) parameters related to the size distribution. O ne of the major goals was to create a ''whitebox'' program. The softwa re is in essence an easy-to-use graphical interface between user and c omputer. The most important motivation to apply the concept of on-line simulation is the limited information content of DLS autocorrelation functions. In the present approach, the user, rather than the software , has to evaluate the physically plausible solutions and select the mo st appropriate one. The computer becomes a mirror that calculates but leaves the decision making to the user. Using this software environmen t, the resolving power of the technique in ideal and noisy conditions was investigated. Results were obtained for monomodal samples where th e mean diameter seems to be far more reliable than the polydispersity. For bimodal samples, the convergence of the solution to a monomodal d istribution was observed when the noise level increased. The goal is t o put into perspective the results as obtained in most ''black-box'' s oftware without having to use heavy mathematical approaches. One of th e major applications is the use of the program as an interactive intro ductory tool to become acquainted with the power and, above all, with the limitations of DLS.