Geologic, microstructural, and spectroscopic constraints on the origin and history of carbonado diamond

Abstract

Carbonado is a form of polycrystalline diamond found in placer deposits in South America and Central Africa, and is one of the toughest known materials. The source rock for carbonado is unknown, and it has unusual porosity, textural features, and inclusion mineralogy. These have lead to a wide variety of theories on the genesis of carbonado. The tightly bound, interlocking microtexture of diamond makes it difficult to study, and only one previous study has been done on polished interior sections of a carbonado. This thesis reports the results from studying the polished surfaces of 21 carbonados from Brazil and the Central African Republic. Reflected light and scanning electron microscopy, cathodoluminescence (CL), Photoluminescence, Raman spectrometry, and secondary ion mass spectrometry were performed on these carbonado samples in order to determine their microtexture and evaluate the various theories of carbonado genesis. In addition, carbonado pore minerals and indicator minerals from the Brazilian rivers in which carbonado is found were studied in an attempt to gain some insight into the possible source rock for carbonado. Some of the individual diamond microcrystals in carbonado were found to have morphological and chemical similarities to the monocrystalline microdiamonds found in the Dachine talc schist of French Guiana. Diamonds and chromites from the Dachine talc schist were studied to determine the protolith of the talc schist, and to constrain the residence history of the Dachine microdiamonds in the mantle. Studies of florencite, a common pore mineral in carbonado, show that the Pb that substitutes into the REE site in the florencite crystal lattice is modern, common lead. When combined with previous geochronological studies that show at carbonado has been associated with uranium for at least 2.5 Ga, this modern common lead shows that the pores have been open to exchange with the exterior environment. Raman and CL studies show that the radiation damage previously documented in carbonado is concentrated in the areas around the pores, suggesting that they were filled with a high concentration of uranium. One carbonado was found to host a metallic Fe-Cr inclusions in its pores. This alloy is of a type previously reported only as intracrystalline inclusions. These results have been interpreted as recording a three step history for the pore mineralogy of carbonado. First, carbonado crystallized in the diamond stability field, in equilibrium with reduced metallic phases. After transport to the surface and release from the host rock, U-bearing groundwater dissolved the pore minerals and precipitated uranium in a redox reaction. Finally, recent tropical weathering reoxidized the uranium, leaving recent lateritic minerals in the pores. Because the Pb isotope model ages for carbonado (2.8-3.6 Ga) are older than most other diamonds and much of the craton in which carbonado is found, a detrital zircon study was performed on carbonado-bearing streams to see if any rocks of this age or older were present in the paleo-drainage basin of the conglomerates that contain carbonado. The detrital zircons found in carbonado-bearing streams had ages between 3.7 and 2.1 Ga The clasts that local garimpeiros (prospectors) and sedimentologists believe are related to carbonado had ages between 3.7 and 3.35 Ga. This age distribution is similar to that of detrital zircons found in green Jacobina quartzites, which were found to have the same range of ages, plus a large concentration of 3.30 Ga grains not present in the sediments associated with carbonado. The only possible indicator minerals found were two Cr-rich rutiles, which may originate from metasomatized mantle. One of these Cr-rutiles was tentatively dated using the U-Pb system as having an age of 2933 Ma. This age corresponds with a time of tectonic quiescence in the drainage area of the carbonado source conglomerates. Optical, CL and Raman spectroscopic studies of polished carbonados show that they consist of either a collection of discrete euhedral or anhedral diamond microcrystals, or of a homogenous mix of irregular shaped grains. The ratio of these two textural types varies between carbonado grains, but is generally constant within each individual carbonado. Grain boundaries are generally not straight, and rarely terminate in symmetrical triple junctions. Raman spectroscopy shows that the level of elastic strain and compression or tension of the diamond crystal lattice is much lower than that of diamonds formed through shock synthesis, precipitated in chemical vapor deposition, or recovered from ureilite meteorites. This elastic strain levels in carbonado are similar to those in lithospheric diamonds or synthetic diamonds synthesized at static high pressures. This suggests that carbonados were in the diamond stability field at moderately high temperatures. SIMS measurements of carbonado using the SHRIMP ll ion probe show that the individual crystals in carbonados have slightly different carbon isotopic compositions and nitrogen concentrations. There are two hypotheses that can account for the features observed in carbonado. The first is a two stage process, whereby the euhedral grains grew first, and the matrix diamond rapidly crystallized at a later date. The second is a deformation process, whereby microdiamonds were concentrated and deformed to varying degrees, resulting in the variable ratios of euhedral diamond to matrix diamond in different carbonado stones. Because the undeformed euhedral diamonds were found to be morphologically and chemically similar to the diamonds in the Dachine talc schist in French Guiana, this primary diamondiferous rock was studied in an attempt to determine how such diamond can form. The diamonds in the Dachine talc schist were found to be type IaA-Ib. The lack of total nitrogen aggregation means that they can not have been resident in the mantle for more than 10 million years. The low aggregation state also constrains the temperature of the magma that erupted them to less than about 1500 °C, as the nitrogen in the Dachine diamond would have aggregated during transport if the magmas were any hotter. SHRIMP carbon isotopic measurements show that these diamonds have a range in carbon isotopic composition from typical mantle values down to typical biogenic values. All Dachine diamonds with detectable nitrogen have identical thermal histories, irrespective of carbon isotopic composition. The chromites on the Dachine talc schist were a mix of metasomatized lithospheric mantle chromites typical of kimberlites, and igneous chromites. The igneous chromites had trace elemental compositions less depleted than those found in boninites and komatiites, and were similar to those found in high-Mg shoshonitic intrusive rocks. This, combined with relict volcaniclastic textures, a geologic setting that is interpreted as an early Proterozoic arc, and the low temperatures required by the diamonds, suggest that the Dachine talc schist may have originally been a hydrous, arc-related volcanic rock. Such an interpretation would allow the low 813C of the Dachine diamonds to be caused by the subduction of organic carbon. There is still much research that must be done before carbonado diamond is well understood. However, this theses presents several new and important constraints. The first result is that the radiation damage in carbonado was generated by the deposition of uranium in the pores, and that both this uranium and the original pore minerals have since been replaced by recent lateritic minerals related to tropical weathering. The other important result is that the diamond lattice in carbonado grains is under very little residual stress, so that whatever process that formed the carbonado microstructures must have occurred in the diamond stability field, and not as a result of metastable diamond growth.