It is better to use the range value, or an average value of diameter with error bars because there is a size distribution of nanoparticels.

I agree, a range is usually more revealing.

In some cases though reporting the ranges may not allow for easy reporting/analyses of colloidal behaviors such as aggregation. In those cases, average or effective diameter (preferably based on number distribution) may be reported. If this is done,  Polydispersity index (PDI) should also be reported as it is a good indicator of the particle distribution.

You might also describe the distribution of the radii by a density function of a lognormal distribution which is often the best way.

Neither an average nor a range is that great. The appropriate distribution is the best way to report what particles you are using, and the parameters that describe this distribution should then be specified. For many samples this would be a log-normal distribution and many fitting algorithms like those of commercial DLS equipment assume a log-normal distribution for the fit. But synthesized or self-assembled particles, in contrast to milled or naturally occurring nanoparticles, can be Gaussian distributed. If you have multi-modal particle distributions this is very important to report for most cases. In any case. reporting just an average and no more data should be avoided.

A final note is that the type of distribution you report (number, area, volume, (intensity)), should be related to the interactions or applications that you investigate AND the method used to determine it. If you report a distribution measured by DLS, which is naturally intensity weighted, this should be reported (as well) and not only the number distribution, which is rather derived from the data and more uncertain, even when it is the most relevant one.

thanks so much for valuable info. i appreciate that.