The Other Half
GEODYNAMIC CONTROLS ON THE DISTRIBUTION OF DIAMONDIFEROUS KIMBERLITES
es.mq.edu.au
Natapov, L.1 and W.L. Griffin, W.L1,2
1 - GEMOC National Key Centre, School of Earth Sciences, Macquarie University, NSW 2109, Australia .
2 - CSIRO Exploration and Mining, P.O. Box 126, North Ryde, NSW 2113, Australia
It is well known that kimberlitic diamonds are closely associated with Archaean lithospheric mantle that is rich in low-Ca harzburgite and contains eclogite. In Phanerozoic time, Archean mantle with such properties has been the main locus of diamond-bearing kimberlite magmas. The mechanisms for generating this type of mantle have been discussed for at least 20-30 years. Models involving generation of the mantle peridotite by extraction of komatiites or thick basaltic crust are often proposed to explain this phenomenon; other models invoke the subduction of oceanic lithosphere to explain the high degree of depletion and the low geotherm of Archean mantle. The existence of volcanic rocks of the calc-alkaline series in the basement of ancient continents may be evidence for the existence of these subduction zones in the past.
Many ancient continents are composed of Archean terranes which are sutured by Proterozoic mobile belts, or joined along major shear zones. Comparisons with modern tectonic settings suggest that the nature of the terranes is diverse: continental massifs, magmatic arcs (including island arcs), blocks of oceanic crust. The sizes of Archean terranes varies widely as well. Siberian and North China terranes may have areas on the order of 105 km2 , while other terranes may be quite small, as in the Slave Craton (Griffin et al., this vol.). The presence of the favourable ancient mantle mentioned above beneath some terranes must be the main condition controlling the distribution of diamond-bearing kimberlite on ancient platforms.
Fig. 1. The Yakutian kimberlite province, with terrane boundaries (thick lines), kimberlite fields with orientation of kimberlite bodies. KR, Kyutungde aulacogen (Devonian). V, calc-alkaline volcanics in the basement terranes.
During Phanerozoic time, the eruption of the kimberlite magmas, entrainment of diamonds from the mantle, and rapid ascent of the magmas to the surface were closely related to episodes of lithospheric extension and melting. Two conditions are crucial for the kimberlite to be diamond bearing: (1) kimberlite magmas must originate below, and sample the lithosphere within, the diamond stability field (typically 900 to 1200°C, 40 to 70 kb); (2) eruption of the kimberlite to the surface must be rapid. Only if these two conditions are met will the diamonds captured in the mantle be preserved in the erupting magma.
Kimberlite fields of the same age often form elongated trends. These trends are often accompanied by extension structures such as grabens, dyke swarms, and fault zones. The length of such kimberlite trends can reach 1000 km (Olenek trend in Siberia, Lucappe corridor of Angola). Kimberlite dykes and the major pipe axes are generally parallel, but sometimes orthogonal, to the trend of the kimberlite field. Alternatively, the kimberlite bodies can cover an isometric area with a diameter of several hundred kilometres. In both cases, the kimberlites can be either diamond-bearing or barren.
Fig. 2. Lithospheric columns for Siberian terranes, numbered as in Fig. 1, showing vertical distribution of harzburgitic garnets from concentrates.
Many of these features of kimberlite volcanism can be illustrated by the Middle Paleozoic kimberlites of the Yakutian province in Siberia. The ancient Siberian continent consists of a series of Archean and Proterozoic terranes that have a SE-NW strike and are separated by shear zones (Rosen et al, 1994). Metavolcanic rocks of the calc-alkaline series can be recognised among the supracrustal rocks of these terrains. Accretion of these terranes to form the Siberian continent took place in Lower Proterozoic time. In middle Paleozoic time the thickness of the lithosphere varied from terrane to terrane, within the range of 230 to 120 km (Fig. 2). The Olenek kimberlite trend crosses all these terranes in North Western direction (Fig. 1). From SW to NE the dominant age of the kimberlites within the trend varies from 360 to 420 My. The diamond bearing mantle is located under SW part of the trend. A major swarm of Devonian basaltic dykes parallels the trend to the SE. The Devonian Vilyui rift, farther to the SE, also parallels the kimberlite trend.
Fig. 3. Palinspastic reconstruction: position of the Siberian plate relative to the Azores hot spot in Devonian time.
Palinspastic reconstructions (Fig. 3) show that the trend appeared when the Siberian plate was passing over a hot spot, which at present is located under the Azore islands. Warming of the mantle and eruption of the kimberlites as well as of the relatively shallow basaltic magmas were caused by the extension of the lithosphere during the process of its movement over the hot spot. The sequence of events related to this extension is as follows. First, the basalt dykes intruded into the crust. The intrusion of the kimberlites was the next stage of the process. The third stage of the extension resulted in the formation of a low-angle detachment fault dipping to the NW. Finally, the Vilyui rift zone developed in the SE part of the detachment. Kimberlites and basalts are located on one side of the Vilyui rift. This can be explained by the orientation of the Wernike detachment zone (Fig. 4) In this particular example, the eruption of basalts and kimberlites, as well as the formation the rift are all interpreted as having been caused by the drift of the plate over the hot spot.
The presence of diamond in the kimberlites correlates with the thickness (at the time of the kimberlite magmatism), age and composition of the lithosphereunder the terranes (Fig.2). Studies of xenoliths and heavy-mineral concentrates (Griffin et al., 1995) show that the thickness of the Archean lithosphere under a diamond bearing kimberlite was typically in the range of 190 to 230 km. Poor kimberlites are usually associated with lithosphere that is 130 to 170 km thick. This lithosphere may be either Archean or Proterozoic in age. The small thickness of the subcontinental lithospheric mantle beneath the NE part of the Olenek trend is probably caused by thermal erosion, and the replacement of the Archean or Proterozoic lithospheric mantle by younger and less depleted mantle. This process may have been caused by the Upper Proterozoic rifting that led to the development of the Udzha aulacogen.
Fig. 4. Detachment model for the linkage between the Vilyui Rift and the Siberian kimberlite province.
The above mentioned features of the spatial distribution of kimberlite are absent when the lithospheric plate is rotating around a hot spot. Nevertheless, the correlation between occurrence of diamondiferous kimberlites and the thickness and composition of the subcontinental lithospheric mantle under different terranes still holds true.
References
Griffin, W.L., Kaminsky, F., O'Reilly, S.Y., Ryan, C.G. and Sobolev, N.V., 1995, 6th Inter. Kimberlite Conf. Abstracts, 194-195.
Rozen, O.M., Condie, K.C., Natapov, L. and Nozhkin, A., 1994, Developments in Precambrian Geology, 11 , 411-459.
Trace element discrimination of garnet from diamondiferous kimberlites and lamproites.
es.mq.edu.au
Pearson, N.J.1, Griffin, W.L.1,2, Kaminsky, F.V.3,van Achterbergh, E.1 and O'Reilly, S.Y.1
1. GEMOC National Key Centre, School of Earth Sciences, Macquarie University, NSW 2109, Australia
2. CSIRO Exploration and Mining, PO Box 126, North Ryde, NSW 2113, Australia
3. KM Diamond Exploration Ltd, 815 Evelyn Drive, West Vancouver, BC V7T 1J1, Canada
The major element composition of chrome-pyrope garnets has been used extensively to establish criteria for target evaluation in diamond exploration. Trace element data provide additional information that can be used to quantify parameters indicative of the diamond grade of a kimberlite or lamproitic host rock (Griffin and Ryan, 1995). This method originally was based on the small group of elements obtained using the proton microprobe. The current study using laser ablation ICP-MS was undertaken to establish the characteristics of a larger group of trace elements in garnet concentrates from diamondiferous and barren kimberlites. To define the trace element features of garnet most likely to coexist with diamond, a number of syngenetic garnet inclusions in diamond were also included in the study.
In-situ quantitative analysis by Laser Ablation Microprobe (LAM) ICP-MS has rapidly developed into one of the most powerful analytical techniques in geochemistry, capable of producing high precision determinations of trace elements at sub-ppm detection limits. The laser ablation system at Macquarie University was designed and installed by Drs Simon Jackson and Henry Longerich of Memorial University, Newfoundland. This system includes a Continuum Surelite I-20 Q-switched Nd-YAG laser with a fundamental wavelength of 1064 nm (IR) and frequency doubling crystals which produce 532 nm (visible) and 266 nm (UV) wavelengths. Operation in the UV wavelength produces enhanced ablation yields for materials with low abundances of transition elements. Typical operating conditions for the quantitative analysis of the garnets in this study involved energies of 0.5 to 2 mJ per pulse at a repetition rate of 4 Hz. Under these conditions the pit size produced is between 30 to 60 µm in diameter and the drill rate is approximately 0.5 µm/sec. Ablation times of up to 120 secs were achieved in 0.5 mm. A full description of the LAM instrumentaion and ICP-MS operating conditions is given in Norman et al. (1996). A suite of 20 to 30 minor and trace elements was determined in each analysis and Ca was used as the internal standard in the quantification procedure. Detection limits for all elements in this study are typically in the range 100 ppb to 1 ppm, although actual values for individual analyses will depend on ablation time, which is largely a function of grain size, and on the internal standard concentration.
Sub-calcic and lherzolitic Cr-pyropes in concentrates from several different cratons were analysed: Kaapvaal craton (Newlands, Leicester, Uintjiesberg, Liqhobong); Siberian craton (Sytkanskaya); Slave craton (A-10). The garnet inclusions in diamond are also derived from kimberlites of equal geographical diversity: Yakutia; Venezuela; Ghana; Canada. Our diamond inclusion data are supplemented by ion probe analyses of diamond inclusion garnets from southern Africa and Siberia (Yakutia) from Shimizu and Richardson (1987 and Shimizu and Sobolev (1995). Shimizu and Sobolev did not report Sc data, so we have assumed a value of 130 ppm, equal to the average of the other diamond inclusion peridotite group garnets.
The relationships between elements such as Zr, Y and Ti were used by Griffin and Ryan (1995) to identify the chemical signatures of different types of mantle processes. Plots of these elements in the concentrate and diamond inclusion garnets in this study confirm previous observations that garnet inclusions in diamonds have depleted trace element patterns (Fig. 1; Griffin et al., 1992; Griffin et al., 1993). The majority of diamond inclusion garnets have Zr contents <20 ppm, Y < 8 ppm and Ti from 10 to 2000 ppm. A significant proportion of the garnets from the more diamondiferous pipes (Liqhobong, Newlands; Sytkanskaya) fall within the field defined by the diamond inclusion garnets. Conversely, garnets from the barren kimberlites (Uintjiesberg) plot outside the field of diamond inclusions on these diagrams (Fig. 2).
Fig 1. Zr vs Y (ppm) for diamond inclusion garnets. Fig 2. Zr vs Y (ppm) for Cr-pyrope garnet in concentrate from kimberlite pipes. The field defined by peridotitic garnet inclusions in diamond is drawn from Fig. 1.
Fig. 3 Chondrite normalised plot of selected peridotitic garnet inclusions in diamond from Yakutia and Slave D027. Fig.4 Chondrite normalised plot of selected peridotitic garnet in concentrate from Newlands kimberlite.
The depleted nature of the diamond inclusion garnets is apparent in chondrite normalised plots (Fig. 3). HREE in many of the garnets, including both harzburgitic and lherzolitic ones, are strongly depleted in HREE and enriched in MREE giving rise to sinuous REE patterns, with convex up LREE to MREE and concave up MREE to HREE. The point of inflection is at Sm/Eu in the most depleted garnets and shifts to Gd/Dy as the concentration of HREE increases in less depleted garnets. The distinctive REE pattern for the diamond inclusion garnets is also developed in some sub-calcic garnets in the concentrates, particularly those with trace element signatures indicating ultradepletion. Lherzolitic garnets have more typical convex-up patterns, with nearly flat REE patterns from Dy to Lu (Fig. 4).
Fig 5. Sc/Y (N) vs Nd/Y (N)) for diamond inclusion garnets. Fig 6. Sc/Y (N) vs Nd/Y (N)) for Cr-pyrope garnet in concentrate from kimberlite pipes. The field defined by peridotitic garnet inclusions in diamond is drawn from Fig. 5.
The significance of the range in the shapes of these patterns is evident in the plots of Nd/Y(N) versus Sc/Y (N) (Fig. 5 and Fig. 6). The Nd/Y ratio clearly distinguishes garnets with the sinuous REE pattern (Nd/Y >>1) from those with more typical LREE depleted patterns (Nd/Y <<1). Sc contents in Cr-pyrope garnet fall within a very restricted range (100-150 ppm) and because Sc appears to be preferentially accommodated into garnet during depletion, Sc/Y provides a measure of the depletion of HREE. Values of Sc/Y >>1 are indicative of the depleted compositions, whereas Sc/Y ~1 are obtained from undepleted lherzolitic garnets. The quadrant defined by Nd/Y >1 and Sc/Y >1 contains all of the diamond inclusion garnets and a significant number of garnets from the diamondiferous pipes. This plot and a plot of Zr/Y (N) versus Sc/Y (N) (not shown here) provide examples of the simple discrimination tests using the expanded trace element suite to estimate the diamond potential of a pipe. Diamondiferous pipes such as Sytykanskaya, Liqhobong, Newlands and Slave have up to 70% of the Cr-pyrope garnets plotting in the field defined by the diamond inclusions. In the low-grade Leicester pipe the proportion of concentrate diamonds in the diamond inclusion field is < 50%, and in the barren Uintjiesberg pipe the proportion is nil.
References
Griffin, W.L. and Ryan, C.G., 1995. J. Geochem. Explor., 53, 311-337.
Griffin, W.L., Gurney, J.J. and Ryan, C.G., 1992. Contrib. Mineral. Petrol., 110, 1-15.
Griffin, W.L., Sobolev, N.V., Ryan, C.G., Pokhilenko, N.P., Win, T.T. and Yefimov, Y., 1993. Lithos, 29, 235-256
Norman, M.D., Pearson, N.J., Sharma, A. and Griffin, W.L., 1996. Geostandards Newsletter, 20, 247-261.
Shimizu, N. and Richardson, S., 1987. Geochim. Cosmochim. Acta, 51, 755-758.
Shimizu, N. and Sobolev, N.V., 1995. Nature, 375, 394-397.
THE LITHOSPHERE BENEATH THE LAC DE GRAS AREA, SLAVE CRATON, CANADA: A XENOLITH STUDY
es.mq.edu.au
Norman J. Pearson1, William L. Griffin1,2, Buddy J. Doyle3, Suzanne Y. O'Reilly1,
Esmé van Achterbergh1 and Kevin Kivi3
1 GEMOC Macquarie,
2 CSIRO Exploration and Mining,
3 Kennecott Canada
The composition, structure and thermal state of the lithosphere beneath the Lac de Gras area in the Slave Craton have been determined from a suite of mantle-derived xenoliths. The xenolith studies form the basic element of 4-D lithospheric mapping and are an essential complement to interpretation of the garnet xenocryst data (Griffin et al., 1998). The xenoliths have been brought to the Earth's surface in several generations of kimberlites ranging in age from 47-75 Ma (Davis and Kjarsgaard, 1997). The majority of the xenoliths in this study come from kimberlite pipes DO-18, DO-27 and A154S and have been recovered during the crushing stage of processing of the kimberlite.
Several lithological groups have been recognised in the xenolith sample population: lherzolites (ol+opx+cpx+grt±crt); harzburgites (ol+opx+grt±crt); dunites (ol±grt±crt); wehrlites (ol+cpx+grt±crt); websterites (opx+grt±cpx±ol±crt); garnet clinopyroxenites (grt+cpx); eclogites (cpx+grt±rut±ky); granulites (plag+cpx+grt±opx). Garnet lherzolites and harzburgites are fine to coarse grained (<1mm to >1cm), with microstructures ranging from equigranular, porhyroclastic to mylonitic. The absence of modal cpx is used to distinguish harzburgite from lherzolite, but the majority of harzburgite garnet compositions are lherzolitic (G9) and indicate coexistence with cpx. Subcalcic garnets are abundant in the garnet cocnentrate (Griffin et al., 1998) and Boyd and Canil (1997) have analysed sub-calcic garnets in harzburgite xenoliths from the Grizzly Pipe, north west of A154.
The websterites and eclogites can be subdivided on the basis of mineral compositions: the websterites include a high-Cr group (grt Cr2O3 1.62-8.30 wt%) and a low-Cr group (grt Cr2O3 < 1.5 wt%); 2 types of eclogites are distinguished using the CaO content of garnet (CaO < 7 wt% and CaO 8.5-13 wt%). Modal variations and mineral compositions indicate that gradations exist between the eclogites with low-Ca garnet and the low-Cr websterites (opx-eclogites).
Geothermobarometry on the peridotite and websterite xenoliths produces a P-T array with a non-uniform geothermal gradient. At T < 900°C the P-T estimates fall near a 35 mWm-2 conductive model geotherm, whereas at T between 900 and 1250 °C the locus of P-T points shifts toward a 40 mWm-2 geotherm. This 'stepped' geotherm is unlike the 'kinked' geotherm that characterises a number of cratonic xenolith suites (e.g., Lesotho). The offset in the xenolith paleogeotherm and gross compositional changes in the garnet concentrate (Griffin et al., 1998) defines two compositionally distinct layers, with a boundary at ~900°C. Comparison of the sheared high-T grt lherzolites with those from the Kaapvaal and Siberian cratons indicates a number of similarities that imply metasomatism by asthenosphere-derived melts. However, the occurrence of undeformed high-T (1200-1250°C) xenoliths requires a minimum lithosphere thickness of * 200 km.
Projection of T estimates for the eclogites to the geotherm places this group of xenoliths in the deeper layer. A bimodal distribution of T for the 2 types of eclogites provides evidence for stratification of the deep layer, with the low-CaO grt eclogites concentrated in the upper part and the high-CaO grt eclogites mainly occuring in the lower part.
High olivine/orthopyroxene ratios combined with high average olivine Fo contents in the peridotites in the shallow layer confirms the ultra-depleted nature of the shallow layer predicted from the garnet concentrate data (Griffin et al., 1998). These same features distinguish the peridotite xenoliths from other Archean xenoliths and imply that the mantle beneath the Lac de Gras area differs from the lithosphere beneath other Archean cratons.
References
Boyd, F.R. & Canil, D. 1997. Peridotite xenoliths from the Slave Craton, Northwest Territories. Abstracts Goldschmidt Conf., pp. 34-35.
Davis, W.J. & Kjarsgaard, B.A. 1997. A Rb-Sr isochron age for a kimberlite from the recently discovered Lac de Gras field, Slave province, northwest Canada. Jour. Geol. 105, pp. 503-509.
Griffin, W.L., Doyle, B.J., Ryan, C.J., Pearson, N.J., O'Reilly, S.Y., Davies, R., Kivi, K. and van Achterbergh, E., 1998. Lithosphere Structure and Mantle Terranes: Slave Craton, Canada. Jour. Petrol., subm.
On a seperate note, I also came across the following links to photos of RSA - Pt/Pd reefs and surroundings that are quite interesting in light of our Messina property.
uct.ac.za
uct.ac.za
Regards |
