Automated particle analysis by Raman microscopy – a method development

Abstract

Abstract Chemical identification of particles increasingly gains importance concerning regulatory affairs of medical devices. Particle numbers of samples very easily add up to orders larger than 1.000. Therefore an automated particle measurement is absolutely worthwhile. As a method of chemical identification Raman spectroscopy combined with microscopy is commonly used. The challenge is to distinguish small particles in a range of a few microns on a surface from the background. Different specimen holder made of glass, polymer (e.g. polycarbonate) or metal are commonly used. After the choice of suitable background parameters such as magnification, optical resolution, focal plane and brightness have to be optimized. The used software for Raman microscopy offers the possibility to select particles for the subsequent measurement. Therefore the xand y-coordinates of particles have to be added to a list manually and separately. This so called point list can be processed automatically. However, this procedure is very laborious and requires a high degree of user interaction especially in case of large particle numbers. Furthermore the software offers the possibility to perform a so called area scan, where a defined surface of the the specimen will be spectroscopically analyzed. The measurements are performed in defined distances on a grid, e.g. every 10 micrometer in xand y-direction. The choice of this tool leads to a large number of measuring points associated with an unwanted long measuring time. By means of the freely accessible software ImageJ an alternative path was found to relate xand y-coordinates to every particle. These coordinates can be imported into the point list of the software for Raman microscopy. Afterwards a Raman spectrum for every particle can be measured without further high effort in time.