In the conclusion of the recent G&G article on beryllium-diffused ruby and sapphire (J. L. Emmett et al., Summer 2003, pp. 84–131), the authors state that “Today, there is no simple, low-cost alternative to SIMS or LA-ICP-MS, but we are working on the problem and strongly encourage others to do so as well.” This writer believes that there does exist a low-cost plasma-based method that might reach the sensitivity required for detecting Be-diffused corundum, that is, about 5 ppma. It is a long-forgotten technique, whose instrumentation seems to have been abandoned by those companies who specialize in spectroscopy-based analytical devices.

This technique employs a carbon arc emission spectroscopic unit. Known as the Hilger Medium Quartz Spectrograph (HMQS), it was designed and manufactured by Adam Hilger Ltd., London, who abandoned its production in 1961. It was a remarkably successful analytical tool for the measurement of certain major, minor, and trace chemical elements. The plasma-excitation method used was a high-voltage AC condensed-spark emission from opposed specimen rods, unlike the DC carbon arc emission spectrographic method proposed here.

The procedure for powder analysis was quite different from SIMS or LA-ICP-MS. The small amount of powder required, less than one milligram, was tightly pressed into a 4.0 mm cylindrical cavity drilled in the centre of a 6.2-mm-diameter spectroscopically pure graphite rod. The specimen-bearing electrode and the upper counter-electrode were secured on the clamping stand, and the pre-alignment optics arrangement was used to position the plasma column’s cathode layer zone image onto the slit. The lenses were positioned by careful setting of the length of the slit (at a standardized slit width) so that only that part of the specimen’s plasma column near the cathode would be admitted to the HMQS. The importance of excluding other parts of the plasma, in particular the incandescent electrode tips, was that this narrow zone, called the “cathode layer” gave the greatest sensitivity of detection for trace elements.

The slit shutter was opened and a pure carbon rod was touched to both electrodes to start the arc. When the scraping from the specimen was completely vaporized and the entire spectral “burn” was recorded on a photographic plate, the shutter was closed. The actual burn period was determined by observing sodium presence in the plasma; this was done by viewing the D-line spectrum by means of a direct-vision pocket spectroscope. When its intensity fell abruptly, the burn was completed and the shutter was closed.

After development, the photographic plate was inserted in a photographic microphotometer, so that the peak height of the strongest emission spectrum line of the test element could be measured. In the case of beryllium, a suitable neighboring aluminum spectral line peak would also be recorded. As the Al concentrations of all rubies and sapphires are virtually constant, the ratios of peak heights, h_Be/h_Al, would be independent of exposure times. Thus a plot of h_Be/h_Al versus Be concentration in ppma would yield a working analytical linear graph. Variations in arc emission power during the entire specimen vaporization period would in no way affect the accuracy of the test analysis. The great virtue of photographically recording the period of the entire specimen vaporization is that all the Be-emitting photons would be integrated, along with those of the calibrating aluminum-emitting photons.

The sample for analysis had to be obtained by mechanical means, but this can be done in a manner that is not likely to affect the gemstone's optical appearance. The writer’s experience with optical sampling of only the cathode layer zone of the DC carbon arc plasma indicated that no more than 0.005 ct of powder would be sufficient for successful analyses for many elements.

While this method may appear to be rather complex and labor intensive, with some practice it is an astonishingly simple and rapid procedure. The writer believes that it could offer gem testing laboratories a lower-cost alternative to SIMS. It remains to be seen if this method can attain sufficient Be ppma sensitivity to establish whether or not a corundum gemstone has been treated by Be lattice diffusion.

At the November 2003 Gem-A annual conference, the writer suggested to one of the speakers, Dr. Henry Hänni, director of the SSEF Swiss Gemmological Institute, that it might be useful to know if a commercially available instrument existed that made use of a UV spectrometer to monitor a plasma emission, instead of the expensive mass spectrometer of the LA-ICP-MS system.

Dr. Hänni had already been investigating this area and was successful in locating such a device, the Laser-Induced Breakdown Spectrometer (LIBS), which costs only in the region of US$75,000. The unit does generate circular pits on the analyzed surface of a gemstone, e.g., the girdle. However, if properly focused, it does not appear to produce further damage.

J. B. Nelson, Ph.D.
Nelson Gemmological Instruments
London