A recursive regularization algorithm for estimating the particle size distribution from multiangle dynamic light scattering measurements. (July 2016)

• Record Type: Journal Article
• Authors: Li, Lei; Yang, Kecheng; Li, Wei; Wang, Wanyan; Guo, Wenping; Xia, Min
• Is Part Of: Journal of quantitative spectroscopy & radiative transfer. Volume 178(2016:Jul.)
• Page Start: 244
• Page End: 254
• Publication Date: 2016-07
• Subjects: Dynamic light scattering -- Inverse scattering -- Scattering measurements -- Particle size distribution -- Mie theory; Spectrum analysis -- Periodicals; Radiation -- Periodicals; Analyse spectrale -- Périodiques; Rayonnement -- Périodiques; Radiation; Spectrum analysis; Periodicals; 543.0858
• DOI: 10.1016/j.jqsrt.2015.12.024
• Languages: English
• ISSNs: 0022-4073

Abstract

Abstract: Conventional regularization methods have been widely used for estimating particle size distribution (PSD) in single-angle dynamic light scattering, but they could not be used directly in multiangle dynamic light scattering (MDLS) measurements for lack of accurate angular weighting coefficients, which greatly affects the PSD determination and none of the regularization methods perform well for both unimodal and multimodal distributions. In this paper, we propose a recursive regularization method—Recursion Nonnegative Tikhonov–Phillips–Twomey (RNNT-PT) algorithm for estimating the weighting coefficients and PSD from MDLS data. This is a self-adaptive algorithm which distinguishes characteristics of PSDs and chooses the optimal inversion method from Nonnegative Tikhonov (NNT) and Nonnegative Phillips–Twomey (NNPT) regularization algorithm efficiently and automatically. In simulations, the proposed algorithm was able to estimate the PSDs more accurately than the classical regularization methods and performed stably against random noise and adaptable to both unimodal and multimodal distributions. Furthermore, we found that the six-angle analysis in the 30–130° range is an optimal angle set for both unimodal and multimodal PSDs. Highlights: The algorithm choosing the optimal inversion method for different PSDs is proposed. Comparison with existing method suggests the superiority of the proposed algorithm. Even noise level is 10% the recovered results can fit the given distributions well. The proposed algorithm could discriminate the peaks of multimodal PSDs precisely. Six-angle analysis in the 30–130° range is an optimal angle set for recovering PSDs.