Diamond Genesis, Seismic Structure, and Evolution of the Kaapvaal-Zimbabwe Craton

See allHide authors and affiliations

Science  06 Sep 2002:
Vol. 297, Issue 5587, pp. 1683-1686
DOI: 10.1126/science.1072384


The lithospheric mantle beneath the Kaapvaal-Zimbabwe craton of southern Africa shows variations in seismic P-wave velocity at depths within the diamond stability field that correlate with differences in the composition of diamonds and their syngenetic inclusions. Middle Archean mantle depletion events initiated craton keel formation and early harzburgitic diamond formation. Late Archean accretionary events involving an oceanic lithosphere component stabilized the craton and contributed a younger Archean generation of eclogitic diamonds. Subsequent Proterozoic tectonic and magmatic events altered the composition of the continental lithosphere and added new lherzolitic and eclogitic diamonds to the Archean diamond suite.

Seismic imaging of the lithospheric mantle beneath the Kaapvaal-Zimbabwe cratons and the Limpopo mobile belt (1–4) has produced a detailed picture of the source region for diamonds (Fig. 1). We place two decades of study of some 4000 diamonds from southern Africa's major diamond deposits (5, 6) into geologic context at lithospheric source depths to relate diamond formation to the processes of craton creation, assembly, and modification. Diamond formation worldwide is associated with the presence of ancient lithospheric mantle keels beneath cratons (7,8). In the Archean keel of the Kaapvaal-Zimbabwe craton, mantle peridotite and eclogite host multiple generations of both Archean and Proterozoic (9) diamonds (6,10, 11) that have been sampled by later kimberlites [65 to 1600 million years ago (Ma)] (12–14).

Figure 1

Tomographic image of the lithospheric mantle in the diamond stability field (150 km depth) derived from seismic P-wave data (2, 3). (Inset) The Kaapvaal (K) and Zimbabwe (Z) cratonic blocks and the Limpopo (L) mobile belt are referred to collectively in the text as the Kaapvaal-Zimbabwe craton. Bold green line indicates the outermost boundary of the Kaapvaal-Zimbabwe craton (1,37, 50). Colored squares represent diamond mines as follows: red, predominantly eclogitic silicate inclusions (Jagersfontein, JA; Jwaneng, JW; Letlhakane, LE; Orapa, O; Premier, P), and green, predominantly peridotitic silicate inclusions (Kimberley area mines of Bultfontein, De Beers, Dutoitspan, and Wesselton termed De Beers Pool, D; Finsch, F; Koffiefontein, KO; Roberts Victor, R; and Venetia, V). See fig. S1 for a cross section.

Seismic velocities in the lithospheric mantle beneath the Kaapvaal-Zimbabwe craton were mapped using teleseismic broadband waveforms (2, 3) (Fig. 1). The Kaapvaal-Zimbabwe craton is marked by relatively high P-wave velocity lithospheric mantle that occurs in two prominent but irregularly shaped lobes separated by a broad WNW-trending band of relatively lower velocity mantle. Southern African diamonds were stored in the mantle with these differences in seismic velocity (Table 1, Fig. 1) because current seismic structure correlates with the Archean to Proterozoic surface geology, which is much older than the kimberlite eruptions. Jwaneng, Letlhakane, Orapa, and Premier diamonds were hosted in slower lithospheric mantle with P-wave velocities below the average model velocity (–0.21 to –0.01%). De Beers Pool, Finsch, Jagersfontein, Koffiefontein, Roberts Victor, and Venetia diamonds were hosted in faster mantle with P-wave velocities above the average model velocity (+0.08 to +0.36%).

Table 1

Seismic velocity of the lithospheric mantle, diamond composition (δ13C, N, and type Ia), and inclusion paragenesis for southern African diamonds. Seismic velocities are for P-waves (in % deviation from a cratonic reference model) through a 50-km-radius cylinder of mantle extending from 150 km to 225 km depth below the major diamond mines (2, 3). Studies of C isotopes and N have been carried out on more than 900 individual diamonds enclosing silicate, oxide, and sulfide inclusions. Parageneses (P, peridotitic; E, eclogitic; W, websteritic) are listed in order of abundance; subordinate parageneses are in parentheses. Nitrogen concentration is the average of the total diamond population as measured by Fourier transform infrared spectroscopy (FTIR); De Beers Pool data is measured by mass spectrometry. Type Ia data are the % of diamonds in the studied population with aggregated nitrogen >20 ppm. Sources of data are as follows: De Beers Pool (22,23), Finsch (25, 29), Jagersfontein (28), Jwaneng (23), Koffiefontein (28), Letlhakane (51), Orapa (24,52), Premier (25, 29, 53), Roberts Victor (27), and Venetia (54).

View this table:

Harzburgitic (15) silicate inclusions in diamonds from the De Beers Pool and Finsch in the seismically fast lithosphere have mid-Archean model ages [3.2 to 3.3 billion year (Gy)] (10); all other silicate inclusions dated so far are lherzolitic or eclogitic and are Proterozoic in age (11,16–18) (Table 2). Eclogitic sulfide inclusions from both fast and slower lithosphere have late Archean (2.9 Gy) (19–21) and Proterozoic ages that match those of the eclogitic silicate inclusions.

Table 2

Silicate or sulfide inclusion age and paragenesis for diamonds from the major diamond mines of southern Africa. About 100 individual sulfide inclusions have been analyzed for Re-Os ages; about 3000 silicate inclusions (either composites or individual grains) have been analyzed for Sm-Nd ages. Sources of data are as follows: De Beers Pool (10, 20), Finsch (10,17, 55), Jwaneng (18, 56), Koffiefontein (19), Orapa (17, 57), and Premier (11, 16).

View this table:

Diamonds from kimberlites in seismically fast lithosphere are dominantly peridotitic (15), whereas diamonds from kimberlites in slower lithosphere are dominantly eclogitic (22–30) (Table 1, Fig. 1) (fig. S1). This relation holds for 9 of the 10 localities studied, with Jagersfontein, which has asthenospheric diamonds (28), as the exception. Both peridotitic and eclogitic diamonds have prevalent mantle-like carbon isotopic composition [δ13C = –3 to –7 per mil (‰)], but eclogitic diamonds have an isotopically light sub-population (δ13C = –22 to –10‰) (Fig. 2A) (31, 32). The diamonds from the slower lithospheric mantle have a greater frequency of isotopically light carbon, a higher percentage of type Ia diamonds (33), and a population with higher average N content (Figs. 2B and 3, A and B).

Figure 2

(A) Comparison of the carbon isotopic composition of individual diamond analyses grouped according to their paragenesis. Both eclogitic and peridotitic histograms include diamonds from all nine localities. (B) Comparison of the carbon isotopic composition of southern African diamonds, grouped according to their derivation from a locality in seismically or fast lithospheric mantle. Sources of data are given in Table 1.

Figure 3

Percentage of type Ia diamonds (A) and average N abundance in diamonds (B) versus P-wave velocity anomaly. Error bars have been set at the ±5% level for percentage of type Ia diamonds, ±50 ppm for average N abundance, and ±0.1% for P-wave velocity anomaly. Letters represent diamond mines as defined in Fig. 1. Peridotitic, open symbols; eclogitic, closed symbols.

Two groups of diamond localities can be designated: those above fast mantle, whose diamonds are dominantly peridotitic with low N content [170 to 259 parts per million (ppm)], and those above slower mantle, whose diamonds are dominantly eclogitic with higher N content (345 to 478 ppm). The ecologitic diamonds incorporated more N during their growth. Whether this was due to a diamond-forming fluid with higher N or faster growth of these diamonds is not clear. Because these diamonds are at least Proterozoic, the correlation of diamond composition, inclusion paragenesis, and seismic velocity suggests that the seismic structure of the craton is a mid-Proterozoic overprint to an Archean keel, which has been frozen in since the mid-Proterozoic.

Seismic velocity is dependent on temperature, but a 1.2-billion-year-old cold cratonic geothermal gradient from Premier (34, 35) (in the slow region) suggests composition may be more important. The slow region of the lithosphere (Fig. 1) may be higher in basaltic components (Fe, Ca, clinopyroxene) or metasomatizing fluids that hydrated and altered the host peridotite and could account for a 1% reduction in P-wave velocity (36). Premier peridotite xenoliths and diamonds provide the best example of the former. Some xenoliths have Archean Re-Os model ages, whereas others have Re-Os systematics that are Proterozoic (37, 38). The precursors of Premier peridotitic silicate inclusions in diamonds (lherzolitic garnet and clinopyroxene) that equilibrated at 1.9 Ga (11) have Archean mantle model ages.

All indications are that diamond formation in the cratonic mantle is episodic rather than continuous. The occurrence of 3.2- to 3.3-Ga diamonds with depleted harzburgitic silicate inclusions (10) and of 2.9-Ga diamonds with enriched eclogitic sulfide inclusions (20) in the De Beers Pool kimberlites (fig. S1) indicates that creation and assembly of the craton was at least a two-stage process. Isotopic ages suggest a time gap of 300 ± 200 million years between the two Archean diamond formation events. Cratonic nuclei were first created by mantle melting processes that produced severe depletion of peridotite and high-degree melts such as komatiites (10, 39). Such depletion is now thought to have occurred at shallow depths (40–42), perhaps in a subduction zone setting (43), although previous suggestions have included mantle plumes (44, 45). Oceanic lithosphere was then accreted to these early cratonic nuclei and contributed to the stabilization of the lithospheric mantle as indicated by the occurrence of 2.9-Ga eclogitic diamonds with a subduction signature (20), corresponding diamondiferous eclogites (21), the typical Re-Os model age for mantle peridotite (37, 38), and evidence for continental collision in the crust from 2.88 to 2.94 Ga (46).

Compositional changes to the lithosphere and the formation of Proterozoic eclogitic diamonds are linked by their preferential occurrence in seismically slow lithosphere (Fig. 1) (fig. S1). The lherzolitic and eclogitic composition of Premier diamonds and the Proterozoic Re-Os model ages in Premier peridotite xenoliths attributable to interaction with melts of Bushveld age (2 Ga) (37, 38) link diamond formation in the WNW-trending belt of slower lithospheric mantle to magmatism of the Bushveld–Molopo Farms complex. Along the western Kaapvaal-Zimbabwe craton margin, diamond formation may also be linked to crustal fluids carried to diamond source depths by subduction, as indicated by the eclogitic inclusion compositions, the lighter and more variable δ18O of Jwaneng zircon megacrysts (47), and the light C and heavy N isotopic composition of Orapa diamonds (48). Available C and N isotope data for Orapa and Jwaneng diamonds require either mixing of mantle and crustal endmembers with variable C/N ratio or, alternatively, C and N isotopic fractionation during diamond formation (23, 48, 49). Either way, the history of creation, assembly, and modification of the Kaapvaal-Zimbabwe craton is reflected in the age and chemistry of the multiple generations of diamonds formed and stored in its lithospheric mantle keel.

  • * To whom correspondence should be addressed. E-mail: shirey{at}


View Abstract

Navigate This Article