(162d) Influence of Test Conditions on the Quality of PCS Results

Photon Correlation Spectroscopy (PCS) is an easy to operate technique, widely used for nanoparticles characterisation. However, in order to obtain a result that is representative of the sample, the operating conditions have to be adequately controlled. In this work we have performed measurements to evaluate the influence of several operating parameters on the quality of the signal obtained and how that can influence the results (particle size distribution). Therefore, we have studied the effect of particle concentration, experiment time and fundamental sample time for the acquisition process (?ä), on the auto-correlation function. The measurements were conducted for monomodal latexes from Duke Scientific (19, 38, 79 and 194 nm). It was possible to conclude that, for each sample, there is a large range of concentrations, experiment times and fundamental sample times, for which the normalised auto-correlation function remains unchanged, leading, thus, to the same particle size distribution. As far as concentration is concerned, that range corresponds to the situation where multiple scattering is not present. However, if we refer to the experimental auto-correlation function (G2 (?ä)) we detect a strong dependence on the aforementioned parameters. The normalisation strategy for the auto-correlation function will also be presented, including the influence of the base-line on the normalised auto-correlation function. Furthermore, the optimum fundamental sample time depends on particle diameter, its value having to increase when the diameter increases. For the same sample, the slope of the first part of the normalised auto-correlation function (characteristic slope) is slightly dependent on the fundamental sample time. However, for the feasible range of fundamental sample times, that dependence is much lower than the dependence on particle diameter. Thus, the dependence of the characteristic slope on ?ä becomes irrelevant, when we are analysing the variation of the slope of the normalised function with particle size. Two algorithms have been used to invert the auto-correlation function (NNLS and CONTIN). However, the performance of the algorithm depends on an adequate choice of the distribution width, supplied as an input parameter, and of the fundamental sample time selected for the acquisition process. Therefore, an adequate choice of the acquisition parameters and of the input values supplied to the inversion algorithm (prior knowledge of the sample) is essential to obtain the correct particle size distribution. The results to be presented come as an effort to optimise operating conditions in a PCS analysis. In fact, in order to be able to use, with confidence, the auto-correlation function, to pre-inspect the nature of complex samples, as has be been presented elsewhere, we have to be sure that the function is not being influenced by extraneous parameters rather than the type of distribution (monomodal or bimodal) and the size of the particles.