Last weekend I taught a special two-day course on the identification of explosives to a young lady from San Salvador. She is a forensic mrcroscopist newly assigned to the "Bomb Squad". Her problem was to determine what explosives were used after terrorist bombings. Fortunately, some small particles of the explosive substance usually remain after a detonating. Careful examination of a bomb crater or of bomb fragments usually uncovers these tiny residues.

My problem was to teach her the microscopical characteristics of the most likely explosives she might encounter. These include common inorganic nitrates, chlorates and percholorates and less common organic (military) explosives such as TNT, RDX, HMX, PETN and tetryl.

One way to improve upon the detection limits for digital images involves increasing the number of counts per pixel. Eight bits, or 1 byte (256 counts), per pixel is most commonly used for monochrome optical, electron and x-ray imaging. This number is well matched to the human eye which can discern approximately 128 gray values, while photographic and some laser printer outputs have the capability of displaying nearly 256 gray values. When the counts exceed 8 bits/pixel the recorded image file size will likely double to 16 bits/pixel. With the better imaging detection limits we are faced with the technical problem of how to view more than 256 gray levels.

In 1355, or close thereto, a 3-foot wide and 14-foot long linen cloth suddenly appeared in history. Greeted rapturously by millions of Christians, it has maintained its acceptance as the true Shroud of Christ from 1300 years earlier. In 1978, a group of about 30 U.S. scientists and engineers studied the Shroud in great detail. All but one concluded the cloth is the burial Shroud of Christ.

During the past two decades, the scanning electron microscope (SEM) has become an indispensable tool for morphological observation in materials and biological sciences. With energy dispersive x-ray spectroscopy (EDS) and wavelength dispersive spectroscopy (WDS), rapid chemical characterization of microscopic features has also become commonplace. As developments in manufacturing methods allow fabrication of ever smaller devices and quality control has achieved higher priority, demand for microscopic characterization by SEM has continuously increased.

A microscopist is presented with a special problem when asked to characterize a population of particles. In order to adequately describe the sample, it may be necessary to examine hundreds or even thousands of individual particles, recording the size, shape, identity or other parameters of interest. To manually characterize a statistically significant fraction of the sample could require many hours of tedious work at the scope, followed by still more tedious hours of data reduction. What is needed is a technique to automate the repetitive parts of this process; the recognition and characterization of particles in the sample and the reduction of data and presentation of results. This technique is automated particle analysis.