MODELLING THE GROWTH OF NATURAL DIAMONDS

Abstract

In the following article we have reviewed the mineralogical, chemical, physical and isotopic evidence concerning the growth of coated diamonds, and have combined it to produce a comprehensive model for their formation.We suggest that the coatings formed during a period of metasomatic activity which preceded the eruption of their host kimberlites. Ultrapotassic CO21bH2O-rich fluids resulted from the crystallisation of diamond-free kimberlitic magmas intruded into the deep lithosphere. The kimberlite magmas were derived from an asthenospheric source which was globally quite uniform in terms of the isotopic composition of carbon and nitrogen: δ13C and δ15N both being close to −5‰.The fluids produced during crystallisation invaded overlying diamondiferous lithosphere in which different regions could contain diamond populations with different mineralogical, isotopic and physical characteristics depending on the history of their source region. Either intrinsic to their formation or due to reaction with reduced lithosphere, the fluids were super-saturated with carbon although not to such an extent as to allow for the nucleation of new diamonds. When these super-saturated fluids encountered older diamonds, new and fibrous diamond growth commenced immediately. This period of metasomatic activity was followed rapidly by one or more successful kimberlite eruptions which transported the now coated diamonds to the surface. In the crust, carbonates within hypabyssal facies kimberlites will have δ13C-values which reflect the deep-seated source (∼ −5‰) whereas those within diatreme facies may be more variable due to outgassing of CO2 during eruption.Finally, we consider isotope fractionation related to growth. We suggest that approaches based on equilibrium are not appropriate to diamonds but rather that it is better to consider the nature and kinetics of the surface reactions leading to growth. Where diamond growth is very rapid, as may be the case for fibrous diamond, we propose that there will be no difference in isotopic composition between the fluid and diamond and not the 3–4‰ 13C depletion in the latter predicted by theoretical treatments. Octahedral diamonds are more problematic since far less is known about their growth conditions. However, we show in principle how the content and perhaps isotopic composition of nitrogen within a diamond may be controlled by its growth rate. We are uncertain whether would be affected.