Some aspects of data analysis of multi-angle dynamic light scattering

• Published in: Macromolecular symposia. 2000-12, Vol.162 (1), p.173-190
• Main Authors: Buttgereit, Ralf, Marth, Michael, Roths, Tobias, Honerkamp, Josef
• Format: Article
• Language: English
• Subjects: Autocorrelation; Correlation; Data processing; Dynamics; Errors; Light scattering; Mathematical models; Time functions
• ISSN: 1022-1360; 1521-3900
• DOI: 10.1002/1521-3900(200012)162:1<173::AID-MASY173>3.0.CO;2-Y

In dynamic light scattering the signal to noise ratio may become very low because of short measurement times or low intensities. Then the analysis of the data becomes a central point and it is worthwhile to accept some numerical efforts and an increased calculation time caused by more sophisticated data analysis. In the following three aspects of an improved data analysis are discussed for multi‐angle dynamic light scattering. In the first part of this paper, the influence of the correlation of the errors of the autocorrelation time function is considered. With decreasing measurement time the errors of the autocorrelation time function increase. The more the errors increase, however, the more important the error model becomes. On the other hand, with increasing number of photons the correlations of the errors of the autocorrelation time function become important: The non‐diagonal elements of the covariance matrix of the data errors increase and may become of the order of the diagonal elements. In the second part the sensitivity of the nonlinear simultaneous multi‐angle regularization to aberrations in the experimental set‐up is investigated. Therefore, for a given bimodal radii distribution and different scattering angles data were simulated taking into account the finite aperture of the detector as well as the laser light backscattered at the back of the cuvette. This improved model is compared with the classical model, which neglects these aberrations in the experimental setup. Finally, the advantage of the simultaneous regularization method over averaging over single‐angle results is demonstrated. Estimating of the radii distribution from all multi‐angle data at once leads to a higher solution compared to averaging over the single‐angle results.