There are several methods for measuring particle size and
particle size distribution. Some of them are based on
light, other on
ultrasound, or
electric field, or
gravity, or
centrifugation. The use of sieves is a common measurement technique, however this process can be more susceptible to human error and is time consuming. Technology such as dynamic image analysis (DIA) can make particle size distribution analyses much easier. This approach can be seen in instruments like Retsch Technology's CAMSIZER or the Sympatec QICPIC series of instruments. They still lack the capability of inline measurements for real time monitoring in production environments. Therefore, inline imaging devices like the SOPAT system are most efficient.
Machine learning algorithms are used to increase the performance of particle size measurement. This line of research can yield low-cost and real time
particle size analysis. In all methods the size is an indirect measure, obtained by a model that transforms, in abstract way, the real particle shape into a simple and standardized shape, like a sphere (the most usual) or a
cuboid (when
minimum bounding box is used), where the
size parameter (ex. diameter of sphere) makes sense. Exception is the
mathematical morphology approach, where no shape hypothesis is necessary. Definition of the particle size for an ensemble (collection) of particles presents another problem. Real systems are practically always
polydisperse, which means that the particles in an ensemble have different sizes. The notion of
particle size distribution reflects this polydispersity. There is often a need for a certain average particle size for the ensemble of particles.
Expressions for sphere size The particle size of a
spherical object can be unambiguously and quantitatively defined by its
diameter. However, a typical material object is likely to be irregular in
shape and non-spherical. The above quantitative definition of
particle size cannot be applied to non-spherical particles. There are several ways of extending the above quantitative definition to apply to non-spherical particles. Existing definitions are based on replacing a given particle with an imaginary
sphere that has one of the properties identical with the particle. ;Volume-based particle size: Volume-based particle size equals the diameter of the sphere that has the same volume as a given particle. Typically used in
sieve analysis, as shape hypothesis (
sieve's mesh size as the sphere diameter). :D = 2 \sqrt[3] {\frac{3V}{4\pi}} :where ::D: diameter of representative sphere ::V: volume of particle ;Area-based particle size: Area-based particle size equals the diameter of the sphere that has the same
surface area as a given particle. Typically used in
optical granulometry techniques. :D = \sqrt[2] {\frac{4A}{\pi}} :where ::D: diameter of representative sphere ::A: surface area of particle
Indirect measure expressions In some measures the size (a
length dimension in the expression) can't be obtained, only calculated as a function of another dimensions and parameters. Illustrating below by the main cases. ;Weight-based (spheroidal) particle size: Weight-based particle size equals the diameter of the sphere that has the same weight as a given particle. Useful as hypothesis in
centrifugation and
decantation, or when the number of particles can be estimated (to obtain average particle's weight as sample weight divided by the number of particles in the sample). This formula is only valid when all particles have the same density. :D = 2 \sqrt[3] {\frac{3W}{4\pi dg}} :where ::D: diameter of representative sphere ::W: weight of particle ::d: density of particle ::g: gravitational constant ;Aerodynamic particle size:
Hydrodynamic or
aerodynamic particle size equals the diameter of the sphere that has the same
drag coefficient as a given particle. : Another complexity in defining
particle size in a fluid medium appears for particles with sizes below a
micrometre. When a particle becomes that small, the thickness of the
interface layer becomes comparable with the particle size. As a result, the position of the particle surface becomes uncertain. There is a convention for placing this imaginary surface at a certain position suggested by Gibbs and presented in many books on
interface and colloid science. ==International conventions==