Filtering-Tikhonov regularization inversion for dynamic light scattering data with strong noise

The Tikhonov regularization is an effective method used for dynamic light scattering (DLS) data inversion. However, its inversion accuracy is low for the strong noise data. Based on filtering and Tikhonov regularization technology, we propose a filter-Tikhonov-L method. To begin with, this method uses the cubical smoothing algorithm with five-point approximation to filter out the noise of autocorrelation function (ACF). A new inversion problem is constructed by the filtered ACF, and then solved by Tikhonov regularization with L-curve criterion. This method combines the advantages of filtering and Tikhonov regularization technology, so that it has the high inversion accuracy. The simulation data of particles with particle size distribution (PSD) from 100 nm to 700 nm was implemented inversion studies. The investigation shows that filter-Tikhonov-L has a great advantage in peak position, relative error, the capability of recognizing double peaks and tolerance of noises. The inversion of experimental data also verifies this conclusion.