In February 2017, a parcel containing 323 colorless to near-colorless diamond melee was submitted to the Mumbai laboratory for screening and color sorting. The average weight of each round brilliant was 0.015 ct, with an average diameter of 1.5 mm. Of this parcel, GIA’s fully automated screening and sorting system confirmed that 219 samples were natural, with the remaining 104 samples referred for further testing. Detailed analysis determined that three of the referred stones were natural, and the remaining 101 stones (31.3%) were CVD synthetics (figure 1).
This result is especially remarkable due to the type of synthetic observed. While the undisclosed mixing of HPHT synthetic diamonds in melee has become a primary concern of the diamond trade, CVD synthetics have been very rare in this size group (Fall 2016 Lab Notes, p. 307). The synthetics were color graded using the GIA melee sorting device for research purposes. The majority of the CVD synthetics were G/H (74.3%) in color, though the D–F (20.8%) and I/J (4.9%) color ranges were also represented.
Further analysis was conducted on these synthetics using Fourier-transform infrared (FTIR) and photoluminescence (PL) spectroscopy (the latter at liquid nitrogen temperature) and DiamondView imaging. Interestingly, the group included both as-grown (10.9%) and treated synthetics (89.1%), indicating that they may have been produced by different laboratories. The treated diamonds showed evidence of annealing at high temperatures for color improvement, probably under HPHT conditions. FTIR revealed that they were all type IIa, with only two samples (1.9%) showing the CVD-specific NVH0 absorption peak at 3123 cm–1. Several impurity complexes common in CVD synthetic diamonds were detected by PL spectroscopy, including NV0/– (575 and 637 nm, 100%), the 596/597 nm center (4.0%), SiV– (736/737 nm, 100%), the 883/884 nm Ni-related center (13.9%; see J.P. Goss et al., “The lattice location of Ni in diamond: A theoretical study,” Journal of Physics: Condensed Matter, Vol. 16, No. 25, 2004, pp. 4567–4578), SiV0 (946 nm, 29.7%), and H2 (986 nm, 17.8%). The presence of SiV0/– and the 596/597 nm center were particularly indicative of CVD origin, as SiV0/– centers are routinely found in CVD synthetics yet are comparatively rare in other diamond materials. The 596/597 nm center has been observed only in CVD synthetics.
DiamondView imaging further emphasized the variety of CVD synthetics included in this parcel. The fluorescence of the as-grown CVD synthetics ranged from orange to red to pink, as seen in figure 2 (top left), due to emissions from NV0/– centers. The other samples showed the typical blue–green fluorescence associated with high-temperature annealed CVD synthetics (figure 2, top right). In an unusual finding, layered structures suggesting abrupt changes in growth conditions—which may include stopping and restarting growth—were seen in 35.6% of the CVD synthetic melee. These abrupt changes resulted in alterations to the impurity uptakes of the material, producing the layered structures. Layer thicknesses varied but were generally about 200–500 μm, with samples often showing two or more layers intersecting their table facets (figure 2, bottom left). Such layers are regularly seen in large CVD synthetics (Winter 2015 Lab Notes, pp. 437–439), though they have not been reported for melee-sized goods. Striations and blue dislocation bundles (the latter visible in figure 2, bottom right) were common (45.5% and 19.8%, respectively); certain high-Si synthetic diamonds were dominated by these bundles. Green or blue phosphorescence was observed for 84.2% of the CVD samples.
The substantial undisclosed mixing of CVD synthetic diamonds in this parcel, approaching one-third of the stones, emphasizes the importance of routine testing of melee to identify HPHT and CVD synthetics. GIA’s automated melee screening and sorting device was able to successfully separate the natural and CVD synthetic material, providing confidence in the stones’ origin and supporting transparency in the industry.