technique such as field flow fractionation (FFF) (2-4) or The acid test for any particle sizing technique is its ability to capillary hydrodynamic chromatography (CHDF) (5). The determine the differential number fraction size distribution of a latter calibration-oriented methods are usually accompanied simple, well-defined sample. The very best characterized polystyby an esoteric theoretical explanation wherein the separation rene latex sphere standards have been measured extensively using may be predicted ''from first principles,'' thereby alleviating transmission electron microscope (TEM) images of a large subthe need for calibration except as a ''check'' of the method. population of such samples or by means of the electrostatic classi-Some attempts at correcting (6) for the intrinsic ''spreadfication method as refined at the National Institute of Standards ing'' (band broadening) processes by means of complex and Technology. The great success, in the past decade, of on-line deconvolution procedures [ following techniques used in liqmultiangle light scattering (MALS) detection combined with size exclusion chromatography for the measurement of polymer mass uid chromatography ( 7)] on occasion produce some imand size distributions suggested, in the early 1990s, that a similar provement in the derived distributions, though the results are attack for particle characterization might prove useful as well. At difficult to reconcile with those derived from transmission that time, fractionation of particles was achievable by capillary electron microscopy (TEM) or electrostatic classification hydrodynamic chromatography (CHDF) and field flow fraction-(EC). Suffice to remark that all methods (including that to ation (FFF) methods. The latter has proven most useful when be described in this article!) have problems that can result combined with MALS to provide accurate differential number in a wide range of inaccuracies. fraction size distributions for a broad range of particle classes. The There are other problems associated with particle sizing, MALS/FFF combination provides unique advantages and precinot the least of which is the definition of what is actually sion relative to FFF, photon correlation spectroscopy, and CHDF measured. Does the average size of an ensemble of particles techniques used alone. For many classes of particles, resolution of refer to a weight-average quantity, or the peak value of the the MALS/FFF combination far exceeds that of TEM measurements. α§ 1998 Academic Press distribution present, or a number average? If a distribution Key Words: particle sizing; submicrometer; suspensions; light is measured, is the ensemble characterized by its 1/e 2 width scattering; field flow fractionation; size distributions; emulsions; (even though the distribution may not be Gaussian), or liposomes; latices.
should it be described by the full width at half-maximum (FWHM)? What about particles that are of nonspherical shape or inhomogeneous structure? What is their important