There have been a number of reports recently on the sale of diamonds to fund long-running civil wars in Africa, particularly in Angola and Sierra Leone (see, e.g., P. Hawthorne, "Diamonds in the rough," Time, December 6, 1999, pp. 64–65). After visiting these troubled areas in December, two U.S. Congressmen, Tony Hall (D–Ohio) and Frank Wolf (R–Virginia), have sponsored legislation (H.R. 3188) requiring disclosure of where diamonds are mined, so that a ban on diamond trade with the rebels in these areas can be enforced. Considering the geology of the Angolan diamond deposits, and of diamond deposits in general, how accurate could such an origin determination be? What scientific tests might be used to establish where a diamond is mined? How practical would it be to attempt such an identification?

Some indications can be gathered from the world of colored stones, where country-of-origin determination has become a popular laboratory service for important rubies and sapphires. In a recent lecture, Ken Scarratt of the AGTA Gemological Testing Center described the many kinds of data his lab needs to determine conclusively the origin for, at most, 90% of the gemstones they are asked to test (T. Novellino, "AGTA lab explains gem origin determinations," National Jeweler, December 1, 1999, p. 22). He went on to explain that the gathered data must then be compared to a large database made from observations of rubies or sapphires known by other means to be from a specific location in order to reach a conclusion. Such testing can take hours to days per stone.

A large database of diamond properties applicable to geographic origin determination does not exist. However, on a statistical basis, diamonds from certain countries and specific mines do have distinctive features. For example, cubic diamonds are common from the Democratic Republic of Congo (formerly Zaire) but relatively rare elsewhere; a high proportion of the world’s rare type IIa diamonds come from South Africa (the Premier mine); diamonds with a green skin are abundant in Botswana (the Jwaneng mine); Australia’s Argyle mine yields diamonds with highly saturated pink and dark brown colors; and gem diamonds from Russia are commonly flat-faced octahedra with sharp edges. However, many of these characteristics are not applicable to faceted diamonds, and none of these features is exclusive to a specific locality when the stones are considered on an individual basis. Tests such as analysis of the inclusions or physical properties of the diamond (e.g., with cathodoluminescence or infrared spectroscopy) also fail to yield results that are specific, conclusive, or unequivocal with regard to origin determination.

Some reasons for the lack of diagnostic inclusions or physical properties are found in the geologic origin of diamonds. Because diamonds form deep in the earth’s mantle (at about 125–150 km) and are transported to the surface by volcanic processes (kimberlite or lamproite magmas), stones that crystallized in the same area of the mantle conceivably can be found in diverse geographic localities. Such diamonds could contain similar inclusions and other characteristics, regardless of where they were ultimately emplaced at the earth’s surface. Further, the diamond-bearing magmas might have sampled more than one mantle source area during their rise toward the earth’s surface, so a single source area can contain diamonds with different formation characteristics. The situation is further complicated in Angola by the fact that most of the diamonds are being mined from alluvial deposits; that is, they were transported from kimberlite pipes elsewhere in Angola or nearby countries in Africa.

Because mineral inclusions in diamonds tend to fall into one of two broad groups (eclogitic and peridotitic), reflecting their mantle source region, it is not likely that such inclusions would be diagnostic regarding the country of origin, even with a study of their trace elements, isotopic signatures, or other characteristics. With current technology, such studies require direct access to the inclusion within the diamond, which is inevitably destructive for stones with non-surface-reaching inclusions. Furthermore, many gem diamonds do not contain mineral inclusions at all. Trace elements in diamond itself have been studied by neutron activation, but this method produces some residual radioactivity in the diamond so that, after measurement, it must be held at a licensed facility before it "cools" enough to be released for sale. Other potentially relevant geochemical methods are also problematic. For example, measuring carbon or nitrogen isotopes to indicate a geographic source has not been shown to yield definitive answers.

Assuming that a database of characteristics for diamonds from different deposits was created, and that some diagnostic properties were identified for the deposits of interest, testing diamonds quickly enough would still be a Sisyphean task. Approximately 10 million diamonds of 0.50 ct or more, of all qualities, are cut annually worldwide. Even with multiple graders in multiple venues, a massive and expensive effort would be required. Given these constraints, it is unlikely that analytical testing of diamonds to determine their place of origin can be made practical—either immediately, as Congressmen Hall and Wolf are proposing, or in the foreseeable future.

Dr. A. A. Levinson
University of Calgary, Calgary, Alberta, Canada
Dr. Ilene Reinitz
GIA Gem Trade Laboratory, New York