Diamond Geology

Modern Advances in the Understanding of Diamond Formation

D. Graham Pearson
University of Alberta, Edmonton, Canada

For the past 50 years, the majority of diamond research has focused on diamonds derived from the lithospheric mantle root underpinning ancient continents. While lithospheric diamonds are currently thought to form the mainstay of the world’s economic production, the continental mantle lithosphere reservoir comprises only ~2.5% of the total volume of Earth. Earth’s upper mantle and transition zone, extending from beneath the lithosphere to a depth of 670 km, occupy a volume approximately 10 times larger.

Diamonds from these deeper parts of the earth—“superdeep diamonds”—are more abundant than previously thought. They appear to dominate the high-value large diamond population that comes to market. Recent measurements of the carbon and nitrogen isotope composition of superdeep diamonds from Brazil and southern Africa, using in situ ion probe techniques, show that they document the deep recycling of volatile elements (C, N, O) from the surface of the earth to great depths, at least as deep as the uppermost lower mantle. The recycled crust signatures in these superdeep diamonds suggest their formation in regions of subducting oceanic plates, either in the convecting upper mantle or the transition zone plus lower mantle. It is likely that the deep subduction processes involved in forming these diamonds also transport surficial hydrogen into the deep mantle. This notion is supported by the observation of a high-pressure olivine polymorph—ringwoodite—with close to saturation levels of water. Hence, superdeep dia­monds document a newly recognized, voluminous “diamond factory” in the deep earth, likely producing diamonds right up to the present day. Such diamonds also provide uniquely powerful views of how crustal material is recycled into the deep earth to replenish the mantle’s inventory of volatile elements.

The increasing recognition of superdeep diamonds in terms of their contribution to the diamond economy opens new horizons in diamond exploration. Models are heavily influenced by the search for diamonds associated with highly depleted peridotite (dunites and harzburgites). Such harzburgitic diamonds were formed in the Archean eon (>2.5 Ga) within lithospheric mantle of similar age. It is currently unclear what the association is between these ancient lithospheric diamonds and large, high-value diamonds, but it is likely a weak one. In contrast, the strong association between superdeep diamonds and these larger stones opens up a new paradigm because the available age constraints for superdeep diamonds indicate that they are much younger than the ancient lithospheric diamonds. Their younger age means that superdeep diamonds may be formed in non-Archean mantle, or mantle that has been strongly overprinted by post-Archean events that would otherwise be deemed unfavorable for the preservation of ancient lithospheric diamonds.

An additional factor in the search for new diamond deposits is the increasing recognition that major diamond deposits can form in lithospheric mantle that is younger than—or experienced major thermal disruption since—the canonical 2.5 billion years usually thought to be most favorable for diamond production.

This talk will explore these new dimensions in terms of the potential for discovering new diamond sources in “unconventional” settings.

Diamond Precipitation from High-Density CHO Fluids

Thomas Stachel¹, Robert W. Luth¹, and Oded Navon^2
1University of Alberta, Edmonton, Canada
²Hebrew University, Jerusalem

Through research on inclusions in diamonds over the past 50 years, a detailed picture has emerged of the mineralogical and chemical composition of diamond substrates in Earth’s mantle and of the pressure-temperature conditions during diamond formation. The exact diamond-forming processes, however, are still a subject of debate.

One approach to constrain diamond-forming processes is through model calculations that aim to obtain the speciation and the carbon content of carbon-hydrogen-oxygen (CHO) fluids at particular O/(O+H) ratios and pressure-temperature conditions (using GFluid of Zhang and Duan, 2010, or other thermodynamic models of fluids). The predictions of such model calculations can then be tested against carbon and nitrogen stable isotopes and nitrogen content fractionation models, based on in situ analyses across homogenously grown diamond growth layers. Based on this approach, Luth and Stachel (2014) proposed that diamond precipitation occurs predominantly from cooling or ascending CHO fluids, composed of water with minor amounts of CO₂ and CH₄ (which in response to decreasing temperature may react to form diamond: CO₂+ CH₄ → 2C + 2H₂O).

The second approach focuses on constraining the diamond-forming medium by studying submicrometer fluid inclusions in fibrous-clouded and, more recently, gem diamonds. Such studies established the presence of four compositional end members of inclusions: hydrous-saline, hydrous-silicic, high-Mg carbonatitic, and low-Mg carbonatitic (e.g., Navon et al., 1988; Weiss et al., 2009). Although these fluid inclusions only depict the state of the diamond-forming medium after formation, they nevertheless provide unique insights into the major and trace-element composition of such fluids that otherwise could not be obtained.

The apparent dichotomy between the two approaches—models for pure CHO fluids and actual observation of impure fluids (so-called high-density fluids) in clouded and fibrous diamonds—relates to the observation that in high-pressure and high-temperature experiments close to the melting temperature of mantle rocks, hydrous fluids contain 10–50% dissolved solid components (e.g., Kessel et al., 2015). Although at this stage the impurity content in natural CHO fluids cannot be included in numerical models, the findings for clouded and fibrous diamonds are not in conflict with the isochemical diamond precipitation model. Specifically, the fact that observed high-density inclusions are often carbonate bearing is not in conflict with the relatively reducing redox conditions associated with the O/(O+H) ratios of modeled diamond-forming CHO fluids. The model for the minimum redox stability of carbonate­­­­-bearing melts of Stagno and Frost (2010) permits fluid carbonate contents of up to about 30% at such redox conditions.

Although additional data need to be obtained to build a thermodynamic model for CHO fluids with dissolved silicates and to better characterize the major and trace-element composition of high-density CHO fluids in equilibrium with typical diamond substrates (the rock types peridotite and eclogite), we already see sufficient evidence to suggest that the two approaches described above are converging to a unified model of isochemical diamond precipitation from cooling or ascending high-density CHO fluids.

REFERENCES

Kessel R., Fumagalli P., Pettke T. (2015) The behaviour of incompatible elements during hydrous melting of metasomatized peridotite at 4–6 GPa and 1000°C–1200°C. Lithos, Vol. 236-237, pp. 141–155, https://doi.org/10.1016/j.lithos.2015.08.016

Luth R.W., Stachel T. (2014) The buffering capacity of lithospheric mantle: implications for diamond formation. Contributions to Mineralogy and Petrology, Vol. 168, No. 5, pp. 1–12, https://doi.org/10.1007/s00410-014-1083-6

Navon O., Hutcheon I.D., Rossman G.R., Wasserburg G.J. (1988) Mantle-derived fluids in diamond micro-inclusions. Nature, Vol. 335, No. 6193, pp. 784–789, https://doi.org/10.1038/335784a0

Stagno V., Frost D.J. (2010) Carbon speciation in the asthenosphere: Experimental measurements of the redox conditions at which carbonate-bearing melts coexist with graphite or diamond in peridotite assemblages. Earth and Planetary Science Letters, Vol. 300, No. 1-2, pp. 72–84, https://doi.org/10.1016/j.epsl.2010.09.038

Weiss Y., Kessel R., Griffin W.L., Kiflawi I., Klein-BenDavid, O., Bell D.R., Harris J.W., Navon O. (2009) A new model for the evolution of diamond-forming fluids: Evidence from microinclusion-bearing diamonds from Kankan, Guinea. Lithos, Vol. 112 (Supplement 2), pp. 660–674, https://doi.org/10.1016/j.lithos.2009.05.038

Zhang C., Duan Z.H. (2010) GFluid: An Excel spreadsheet for investigating C-O-H fluid composition under high temperatures and pressures. Computers & Geosciences, Vol. 36, No. 4, pp. 569–572, https://doi.org/10.1016/j.cageo.2009.05.008

How to Obtain and Interpret Diamond Ages

Steven B. Shirey¹ and D. Graham Pearson^2
1Carnegie Institution for Science, Washington, DC
²University of Alberta, Edmonton, Canada

Diamond ages are obtained from radiogenic isotopic analysis (Rb-Sr, Sm-Nd, Re-Os, and Ar-Ar) of mineral inclusions (garnet, pyroxene, and sulfide). As diamonds are xenocrysts that cannot be dated directly, the ages obtained on mineral inclusions provide a unique set of interpretive challenges to assure accuracy and account for preexisting history. A primary source of geological/mineralogical uncertainty on diamond ages is any process affecting protogenetic mineral inclusions before encapsulation in the diamond, especially if it occurred long before diamond formation. In practical application, the isotopic systems discussed above also carry with them inherent systemic uncertainties. Isotopic equilibrium is the essential condition required for the generation of a statistically robust isochron. Thus, isochron ages from multiple diamonds will record a valid and accurate age when the diamond-forming fluid promotes a large degree of isotopic equilibrium across grain scales, even for preexisting (“protogenetic”) minerals. This clearly can and does occur. Furthermore, it can be analytically tested for, and has multiple analogues in the field of dating metamorphic rocks. In cases where an age might be suspect, an age will be valid if its regression uncertainties can encompass a known and plausible geological event (especially one for which an association exists between that event and the source of diamond-forming fluids) and petrogenetic links can be established between inclusions on the isochron.

1. Diamonds can be dated in six basic ways:
   1. model ages
   2. radiogenic daughter Os ages (common-Os-free)
   3. single-diamond mineral isochrons
   4. core to rim ages
   5. multiple single-diamond isochron/array ages
   6. composite isochron/array ages

Model ages (1) are produced by the intersection between the evolution line for the inclusion and a reference reservoir such as the mantle. The most accurate single-diamond age is determined on a diamond with multiple inclusions (3). In this case an internal isochron can be obtained that not only establishes equilibrium among the multiple grains but also unequivocally dates the time of diamond growth. With extreme luck in obtaining the right diamond, concentric diamond growth zones visible in UV fluorescence or cathodoluminescence can sometimes be shown to constrain inclusions to occur in the core of the diamond and in the exterior at the rim. These single grains can be extracted to give a minimum growth time (4) for the diamond. In optimal situations, multiple inclusions are present within single growth zones, in single diamonds, allowing internal isochrons to be constructed for individual growth zones in single diamonds. If enough diamonds with inclusions can be obtained for study, valid ages for diamond populations can be obtained on multiple single-diamond ages that agree (5) or on composited, mineralogically similar inclusions to give an average age (6).

REFERENCE

Smit K.V., Shirey S.B., Wang W. (2016) Type Ib diamond formation and preservation in the West African lithospheric mantle: Re–Os age constraints from sulphide inclusions in Zimmi diamonds. Precambrian Research, Vol. 286, pp. 152–166, http://dx.doi.org/10.1016/j.precamres.2016.09.022

The Lesedi La Rona and the Constellation—The Puzzle of the Large Rough Diamonds from Karowe

Ulrika F.S. D’Haenens-Johansson
GIA, New York

In November 2015, Lucara Diamond’s operation at the Karowe mine in Botswana gained notoriety due to the extraction of a series of large colorless diamonds, including the 1,109 ct Lesedi La Rona and the 812 ct Constellation. The Lesedi La Rona marks the largest gem diamond recovered since the Cullinan (3,106 ct) in 1905. The Constellation, considered to be the seventh-largest recorded diamond, attained the highest price ever paid for a rough, selling for $63.1 million ($77,649 per carat). Additionally, three other significant colorless diamonds were recovered during the same period, weighing 374, 296, and 183 ct. Due to the similarity in their external characteristics—which include cleavage faces—as well as their extraction locations and dates, it was suspected that these stones might have originated from a larger rough that had broken. Lucara demonstrated that the 374 ct diamond and the Lesedi La Rona fit together, yet a large cleavage plane is still unaccounted for. GIA was able to study several rough and/or faceted pieces of these five diamonds using a range of spectroscopic and imaging techniques to gain insight into the presence and distribution of point defects in these diamonds.

Diamonds are commonly classified according to their nitrogen content measured by Fourier-transform infrared (FTIR) spectroscopy: Type I diamonds contain nitrogen in either isolated (Ib) or aggregated (IaAB) forms, while type II diamonds do not contain detectable nitrogen concentrations (IIa) but may contain boron (IIb). Analysis of faceted stones cut from the Lesedi La Rona indicates that the rough is a mixed-type diamond, containing both type IIa and pure type IaB regions. These types of diamonds, though exceedingly unusual, have been observed at GIA and reported by Delaunay and Fritsch (2017). The Constellation and the 374, 296, and 183 ct diamonds were determined to be type IaB, containing 20 ± 4 ppm B-aggregates (N₄V), in agreement with the concentration for the type IaB pieces of the Lesedi La Rona. Pure type IaB diamonds such as these are actually quite rare, accounting for only 1.2% of a random suite of 5,060 large (>10 ct) D-to-Z diamonds submitted to GIA, whereas 24.6% were type II. Photoluminescence spectra further confirmed analogous defect content for the five large Karowe diamonds, with emissions from H4 (N₄V₂⁰, 496 nm), H3 (NVN⁰, 503 nm), 505 nm, NV^– (637 nm), and GR1 (V⁰, 741 nm) defects showing similar relative intensities and peak widths. Even for diamonds of the same type, parallel defect content and characteristics across such a variety of defects is unlikely for unrelated stones.

The external morphologies of the diamonds showed primary octahedral, resorbed, and fractured faces, with the Constellation and the 296 ct diamond featuring fractures containing metallic inclusions and secondary iron oxide staining. Deep UV fluorescence (< 230 nm) imaging elucidated the internal growth structures of the samples. For the Constellation and the 374, 296, and 183 ct diamonds, at least two growth zones with differing blue fluorescence intensities were observed within single pieces.

Combined with the spectroscopic data, these results provide compelling evidence that the Lesedi La Rona, the Constellation, and the 374, 296, and 183 ct diamonds from Karowe had comparable growth histories and likely originated from the same rough, with a combined weight of at least 2,774 ct.

REFERENCE

Delaunay A., Fritsch E. (2017) A zoned type IaB/IIa diamond of probable “superdeep” origin. Journal of Gemmology, Vol. 35, No. 5, pp. 397–399.