The laboratory routinely receives large numbers of blue sapphires for examination. Many of these stones show relatively typical signs of heat treatment in the form of melted and resolidified mineral inclusions, heat decrepitated primary and secondary fluid inclusions, and color zoning attributed to internal diffusion. It has also been demonstrated through experimentation with synthetic corundum that the precipitation of inclusions formed by exsolution, such as fine needles of silk-like rutile, can be controlled during the growth process. As would be expected, this is true of natural corundum as well.

During the heat treatment of natural corundum, preexisting exsolved rutile can be dissolved back into the corundum lattice structure to clarify a rutile-clouded stone. This “breakdown” begins to occur at approximately 1250°C and progresses more rapidly as the temperature and time at-temperature both increase. In some instances, however, the rutile is not completely reabsorbed, leaving behind tiny crystallite remnants aligned in typical corundum exsolution patterns; these are sometimes accompanied by clouds of color resulting from internal diffusion. The rutile may also partially recrystallize back into the corundum host. When this occurs, the rutile usually reappears as very fine, short needles or extremely small dust-like, light-scattering particles. Good examples of this type of controlled exsolution can be found in both flame-fusion (Verneuil)-grown star sapphires and rubies and cabochon-cut natural stones that have asterism forced on them through lattice diffusion.

Recently the laboratory received a relatively large (20+ ct), faceted, transparent blue sapphire for identification. The many inclusions, all of which showed evidence of heat treatment, made the identification as a natural but heat-treated stone relatively easy.

However, the examination also revealed an inclusion with a dendritic pattern that was unique in our experience. This feature was very elusive in that it was only visible when a fiberoptic light was directed in a specific way oblique to the plane of the table facet. These dendrites were arranged in consecutive rows and displayed an obvious arborescent habit, as illustrated in figure 17. They also appeared to occupy the same plane in their host. This, combined with their reflective thin-film behavior and crucial directional visibility, led to the conclusion that they were the result of post-treatment exsolution. The fact that they were composed of numerous tiny individual disk-like crystallites also suggested that they originally formed as continuous branching dendrites that then contracted down into individual droplets as the corundum cooled before they finally solidified. This is only one theory, of course. Readers are invited to convey their own thoughts as to the origin of these unusual inclusions.

John I. Koivula
GIA Laboratory
Carlsbad, California