A new coincidence model for single particle counters, Part I: Theory and experimental verification.

Abstract

The prerequisites for estimating the effect of signal coincidence on both particle undercounting and the injection of false counts in the implementation of U.S.P. 788 contaminating particle assays by light extinction particle counters are defined. These include a particle concentration measure that varies with particle size and a new model of the counting process. Both prerequisites have been verified empirically: a single normalized equation describes the coincidence effect in all single particle counters. The single parameter of the normalized equation is the number of effective detector volumes per milliliter. A maximum undercount limit of 5% is proposed based on adequately suspended particles. Using the SVP U.S.P. XXII acceptance limits of 10,000 particles per container or the PMA propose 6,000 particles per container maximum for particles > 10 microns in U.S.P. XXIII, undercount errors are estimated for the smallest container sizes. The large concentration of particles below the controlled 10 microns particle size, that has been documented in injectable solutions, can pose an additional 788 measurement hazard. A Poisson model is used to estimate and control the injection of false particle counts into the mandated measurement through particle coincidence. Acceptable counting accuracy limits with present particle counting systems can be achieved by understanding the capabilities of the particle counter measurement system and using a dilution technique when appropriate. The new model of the counting process and the new particle concentration measures can result in standard, conservative, instrument specifications for use in Pharmacopeial contamination testing and in GLP user evaluation tests. Part I of this paper includes the theory of the coincidence effect on particle counting and the particle size distribution measured. A summary of the experimental verification employed to determine coincidence count loss as a function of particle concentration for single particle counters is reported. Part II of this paper describes a practical protocol for the determination of operating limits to achieve a selected coincidence undercount limit for single particle counters.