ISSN:
0098-1273
Keywords:
Physics
;
Polymer and Materials Science
Source:
Wiley InterScience Backfile Collection 1832-2000
Topics:
Chemistry and Pharmacology
,
Physics
Notes:
Properties of solutions and colloidal suspensions such as molecular and particle dynamics, diffusivities, and size distributions may be determined through dynamic light-scattering experiments. Here a number of methods for predicting the details of the linewidth distribution from photon correlation spectroscopy data are reviewed. Their performance on simulated data (with and without noise added) and experimental data from polystyrene latex standards (including a mixture of two standards) is compared. Methods which do not assume a specific form for the distribution are considered. These include cumulants, histograms, exponential sampling, subdistributions, a non-negatively constrained histogram, and Provencher's constrained regularization. Constrained regularization was found to be most robust to noise present in the autocorrelation function and therefore most reliable for analyzing experimental data; however, the method sometimes oversmoothed the distribution. For bimodal distributions the histogram method performed well in our testing, especially when the approximate peak locations were known a priori. Two linear least-squares fitting methods, exponential sampling and the non-negatively constrained histogram, yield accurate values for the overall mean and standard deviation and can be implemented easily on a microcomputer. The linear subdistribution method, although computationally fast, sometimes was not as accurate as other methods.
Additional Material:
17 Ill.
Type of Medium:
Electronic Resource
URL:
http://dx.doi.org/10.1002/pol.1985.180230707
Permalink
