PDI-individual-peakCan one calculate a PDI value excluding a small aggregation peak? For example if the main particle species is ~100nm and there is a small contribution (2% by intensity) from a peak at 5 microns, is it possible to recalculate the z-average and the PDI ignoring the 5 micron peak. Alternatively, is it possible to determine a PDI for the smaller species only?

The z-average will be weighted more towards smaller components, because only the initial part of the correlation function is fit. Following the ISO method to determine the z-average the correlation function is fit up until 10% of its initial value. It is therefore possible and quite likely that the 5 micron peak is only slightly contributing to the overall z-average.
If one wanted to reprocess the data to completely avoid the 5 micron peak, this MAY be do-able by going into the research software and changing the number of overall data points (= channels) used in the fitting. If by fitting to only 10% of the intercept the regularization no longer shows the 5 micron peak then your z-average would not contain it, and this would also be in line with the ISO protocol.
However, if the 5 micron peak was still there, and you still wanted it to disappear, then one could reduce the fitting range to even lower channels, maybe up to only 50% of the intercept, until it disappears – but please note this no longer would be a cumulant fitting scheme according to ISO, it would be modified to your situation and no longer follow the official procedure.

Concerning a pdi for an individual peak, this is much easier, and requires no re-analysis: when displaying the intensity particle size distribution, each peak comes with a mean and a standard deviation. The pdi for that peak is the square of the standard deviation divided by the square of the mean, for example if the peak was at a mean size of 9.3nm and the st dev was 4.4nm then the pdi for this peak would be: 4.4*4.4/(9.3*9.3) = 0.22.

Please note that this pdi (from the distribution fit) will be different from the pdi from the cumulant fitting, because the cumulant is a forced single exponential fit to a limited set of the correlation function, whereas the regularization is a fit with more parameters, and fitting a larger set of the raw correlation function data.

Here are a few comments I have made earlier on our blog site:
– Is the z-average or the peak size better?
– What is polydispersity in dynamic light scattering and in GPC?

If you absolutely wanted to adjust the fit parameters this could be done in the research software, which you could try out for 30 days. However, this is not recommended and I would advise against it unless you consider yourself well-versed in light scattering and do this at your own risk.


If you have any questions, please email me at ulf.nobbmann@malvern.com. Thanks! While opinions expressed are generally those of the author, some parts may have been modified by our editorial team.