A comparative study of submicron particle sizing platforms: accuracy, precision and resolution analysis of polydisperse particle size distributions

Abstract

The particle size distribution (PSD) of a polydisperse or multimodal system can often be difficult to obtain due to the inherent limitations in established measurement techniques. For this reason, the resolution, accuracy and precision of three new and one established, commercially available and fundamentally different particle size analysis platforms were compared by measuring both individual and a mixed sample of monodisperse, sub-micron (220, 330, and 410 nm - nominal modal size) polystyrene particles. The platforms compared were the qNano Tunable Resistive Pulse Sensor, Nanosight LM10 Particle Tracking Analysis System, the CPS Instruments's UHR24000 Disc Centrifuge, and the routinely used Malvern Zetasizer Nano ZS Dynamic Light Scattering system. All measurements were subjected to a peak detection algorithm so that the detected particle populations could be compared to 'reference' Transmission Electron Microscope measurements of the individual particle samples. Only the Tunable Resistive Pulse Sensor and Disc Centrifuge platforms provided the resolution required to resolve all three particle populations present in the mixed 'multimodal' particle sample. In contrast, the light scattering based Particle Tracking Analysis and Dynamic Light Scattering platforms were only able to detect a single population of particles corresponding to either the largest (410 nm) or smallest (220 nm) particles in the multimodal sample, respectively. When the particle sets were measured separately (monomodal) each platform was able to resolve and accurately obtain a mean particle size within 10% of the Transmission Electron Microscope reference values. However, the broadness of the PSD measured in the monomodal samples deviated greatly, with coefficients of variation being ~2-6-fold larger than the TEM measurements across all four platforms. The large variation in the PSDs obtained from these four, fundamentally different platforms, indicates that great care must still be taken in the analysis of samples known to have complex PSDs. All of the platforms were found to have high precision, i.e. they gave rise to less than 5% variance in PSD shape descriptors over the replicate measurements.