Representative Sampling – Another Cinderella of Particle Size Analysis

AbstractLike most things in life one gets out what one puts in and this is no truer than in modern instrumental particle size distribution techniques. The mantra of ‘garbage in = garbage out’ is meant to convey that the apparently complicated laser diffraction, dynamic light scattering, and electron microscopy techniques measure faithfully what they are given, but what they are given may not be representative of the entire product or material on which significant economic decisions will be made. Tiny samples down to pg in size on electron microscopes are somehow meant to be able to generate information on what may be many tonnes of heterogeneous sample. The bad news is that distribution and heterogeneity imply that statistical methods must be employed in order to obtain accurate and reproducible information. The good news is that representative sampling is amenable to simple statistical evaluation and 2 major predictions can be made:•Based on a required or specified precision or standard error (SE), the point at the top end of the distribution to be specified to this degree of precision, and the density of the sample, then a simple prediction of the minimum mass, required to meet this required level of precision, can be calculated•Similarly, if the mass utilized in the particle size distribution determination is known and the point in the distribution to be specified (plus the density of the sample again), then we can calculate a theoretical best achievable precision based solely on the heterogeneity of the sampleThis paper will illustrate the 2 points above with practical examples