Gemmology, geology and origin of the Sandawana emerald deposits, Zimbabwe

As one of the most valuable gemstones, emeralds are known to occur in several countries of the world, such as Colombia, Zambia, Brazil, Pakistan, Afghanistan, Russia, Madagascar and Zimbabwe. The emerald deposits at Sandawana, Zimbabwe, are described, the emeralds from this deposit characterised and a model of emerald formation presented; this is compared with existing models. The emeralds from Sandawana, Zimbabwe, show relatively constant physical properties, with high refractive indices and specific gravities. They are characterized by laths and fibres of amphibole, both actinolite and cummingtonite. Other common inclusions are albite and apatite. Rare, opaque and chromium-rich inclusions constitute a new variety of ilmenorutile. Compared to emeralds from most other localities, fl uid inclusions are rare and small. Sandawana emeralds have very high contents of chromium, sodium, magnesium, lithium and caesium. They can be readily separated from emeralds from most other localities by using traditional gem testing techniques, on the basis of a combination of physical properties, inclusions and chemistry. In cases of possible doubt, such as in comparison with emeralds from Rajasthan (India), the use of oxygen isotopes is helpful. Sandawana emeralds occur at the contact between greenstones of the Mweza Greenstone Belt (MGB) and pegmatite intrusions. Rare-element granitic pegmatites intruded the MGB just prior to and/or during a main deformation event at 2.6 Ga, at the southern border of the Zimbabwe craton. Subsequently, a late-stage Na-rich ‘solution-melt’ containing F, P, Li, Be and Cr was injected along shear zones, causing albitisation of the pegmatite and phlogopitization in the greenstone wall-rock. Coeval ductile deformation is indicated by boudinage, pinch-and-swell and folding of pegmatites, by differentiated layering in associated amphibole-phlogopite schist and by the presence of (micro)shear zones. The synkinematic growth of not only phlogopite, but also emerald, fl uorapatite, holmquistite and chromian ilmenorutile, indicates enrichment of Na, K, Li, Be, F, P, Rb, Cs, Ta and Nb in the emerald-bearing shear zone. This suggests that emerald formation is closely related to syntectonic K-Na metasomatism. In this process, microcline, oligoclase, quartz (from the pegmatite) and chlorite (from the greenstones) were consumed, in favour of albite (in the pegmatite), phlogopite, some new actinolite and cummingtonite, holmquistite, fl uorapatite and emerald (at the contact and in the greenstone). Mass balance calculations indicate that a Na- and F-rich hydrous fl uid must have been involved in the alterations that ultimately caused emerald formation. The presence of small, isolated, highly saline brine inclusions in emerald supports this result. Formation of gem-quality emerald occurred in relatively low-pressure domains, such as ‘traps’ under folded pegmatite or pinch areas near pegmatites with pinch-and-swell or boudin structures. Apatite-phlogopite thermometry gives T = 560-650°C, interpreted as the temperature range at which emerald was formed. These temperatures imply contact metamorphic rather than regional metamorphic conditions. From the general conditions of primary consolidation of rare element pegmatites, and from the inferred crystallisation path of the similar Bikita pegmatite, a corresponding pressure of around 2.5-3 kbar appears plausible at Sandawana. Because of the intimate spatial and temporal relationship with magmatic activity, the pegmatitic/hydrothermal nature of the involved fl uid, and the near magmatic temperatures of phlogopite and apatite formation, a magmatic source for the Na-rich fl uids is very likely. It means that intrusion of the pegmatites, and subsequent albitisation and metasomatism during deformation, have been part of a continuous process in a restricted period of time. The Sandawana data lead to a new model of emerald formation, as a product of contact metasomatism at the border of ultramafic rocks and rare-element pegmatites during a deformation event, involving late stage magmatic/hydrothermal activity channelled by shearing. Hence, Sandawana emeralds were formed in the Late Archaean, around 2.6 Ga, at the border of a major Late Archaean suture zone. As a consequence, they are by far the oldest, compared to other commercially available emeralds. 40Ar/39Ar dates, varying from 1900 to 2400 Ma, are interpreted as Archaean ages that became partially to completely reset during a Proterozoic thermal overprint. The Sandawana model for emerald genesis does not fit into any of the genetic classification schemes proposed in the literature, and shows that no single theory can be applied to all schist-type deposits. Emeralds of gem quality can be formed in very different geological settings, as long as basic conditions are fulfilled. In this respect, the following factors are essential: availability of beryllium and chromium (± vanadium); means of transport to bring the elements together, that is, fl uids of magmatic, hydrothermal, metamorphic or a combined origin; conditions in which emerald may form as a stable mineral, with temperatures generally from 600° down to 300°C; and sufficient space to grow transparent and well-formed crystals.