Archean Rare-Metal Pegmatites in Zimbabwe and Western Australia

The alkali metal Cesium was first described by Robert Wilhelm Bunsen and Gustav Robert Kirchhoff during the investigation of mineral waters from Dürkheim. It is a silvery-gold, soft extremely reactive and pyrophoric metals, has a large ionic radius and a rather low melting point. Cesium belongs to the large ion lithophile elements. The Cesium market is very small. As a result, data on Cs resources and production are not available or very limited. According to the USGS Cs is currently only mined from massive pollucite mineralisation bearing pegmatites in Canada (Tanco) and Zimbabwe (Bikita). Cesium is a typical trace element and normally occurs in the level of a few to some tens of ppm. As Cs is almost incompatible during magmatic crystallisation, it becomes enriched in residual melts. Therefor Cs occurs as trace element in feldspar and mica. However, certain geological processes are capable to enrich Cs to several thousand ppm so that specific conditions can lead to the formation of discrete Cs minerals. At present 31 Cs are known and approved by the International Mineralogical Association. Most of them crystallise in granitic pegmatites or in alkaline complexes at late stage magmatic to hydrothermal processes. However, only the zeolite Cs mineral pollucite as part of the analcime-pollucite series is known to occur in larger and thus economic quantities. Pollucite is classified as tectosilicate and belongs to the zeolite group. It has a general composition of (Cs, Na)₂Al₂Si₄O₁₂xH₂O and is isostructural to analcime NaAlSi₂O₆xH₂O. Although it can crystallise in gem stone quality cubic, dodecahedral to trapezohedral crystals with various colours pollucite is commonly developed as glassy, colourless to white polycrystalline masses. Pollucite occurrences are reported from approximately 140 locations worldwide. However, at present only two LCT pegmatite deposits are known to host economic quantities of massive pollucite mineralisations. These are the Bikita LCT pegmatite in Zimbabwe and the Tanco LCT pegmatite deposit in Canada. Both pegmatite deposits have a comparable regional background, as they are hosted within greenstone belts on Archean Cratons and yield Neo-Archean ages of ~2625 Ma. Although several studies focused on the mineralogical and geochemical characteristics of pollucite are done, no systematic investigation focusing on the genesis of these massive pollucite mineralisations was performed so far. Therefore, the massive pollucite mineralisation at Bikita were studied in detail to understand the genetic concepts on their formation. Major Portions of Western Australia consists of Meso- to Neoarchean crustal units (e.g., Yilgarn Craton, Pilbara Craton) that are known to host a large number of LCT pegmatite systems. Among them are the LCT pegmatite deposits Greenbushes (Li, ±Ta) and Wodgina (Ta, Sn ±Li). In addition, small amounts of pollucite were recovered from a single diamond drill core at the Londonderry pegmatite field and more recently from the Sinclair LCT pegmatite, both located within the Yilgarn Craton. Despite that, no systematic investigation and/or exploration were conducted for the mode of occurrence of Cs and especially that of pollucite in Western Australia. Thus, Western Australia represents prospective area that may host one more deposit of a yet merely unique mineralisation type on the planet Earth.

The geological evolution of Archean Cratons is of importance for understanding the origin of LCT pegmatites, especially those with massive pollucite mineralisation as at Bikita. The pegmatites of Bikita are hosted by metabasalts and metadiorites of the Upper Bulawayan Group in the Masvingo greenstone belt in the SE part of the Zimbabwe Craton. At about 2670 Ma the southern part of this craton was affected by deformation and up to granulite-facies metamorphism along the Northern Limpopo Thrust Zone. Partial melting during this event led to the intrusion of the 2601 Ma Chilimanzi Suite granitoids with their temporal and spatial association to the pegmatites. The Yilgarn Craton in Western Australia is subdivided into six terranes which host many LCT pegmatites. The Londonderry and Mount Deans pegmatite fields form part of the Kalgoorlie Terrane, whereas the Cattlin Creek pegmatites belong to the Southern Cross Terrane in the southeastern part of the craton. Two main contrasting models for the seven Archean geodynamic events (D0–D6) are discussed: modern-style plate tectonics with a westward directed subduction type scenario, or vertical crustal movements induced by constantly upwelling of hot mantle material. The emplacement of the LCT pegmatites in various greenstone belt lithologies is related to the D5 event (2650–2630 Ma) with widespread intrusions of a low-Ca granitoid suite and formation of Au-mineralisation. The Pilbara Craton in the northern part of Western Australia represents Paleo- to Meso-Archean ages (3530–2830 Ma) and one of the world’s major tantalum pegmatite provinces. Well preserved low-grade Paleo-Archean greenstone belt successions with classic dome-and-keel architecture rest between dome-like granitic complexes that were emplaced during tectonic events (D1–D10) with numerous distinct episodes of magmatism. In the Eastern Pilbara Craton, the Wodgina pegmatite district is located within the Wodgina greenstone belt, surrounded by the large granitoids of the Yule Granitoid complex. The emplacement of the Wodgina pegmatites is associated to an event D10 (2930–2830 Ma) with the Sn-Ta-Li bearing post-tectonic granitoid intrusions of the Split Rock Supersuite at 2890–2830 Ma.

For the petrography of Neo-Archean pegmatites at Bikita, Londonderry, Mount Deans, Cattlin Creek and Wodgina, optical microscopy of polished thin sections was precised by backscattered electron imaging (BSE) and energy dispersive X-ray spectroscopy (EDS) with a scanning electron microscope (SEM). Automated mineral liberation analysis (SEM-MLA) of unpowdered aliquots for WR geochemistry allowed to quantify mineral modes of the coarse-grained pegmatites. This modal analysis encloses the accessory minerals at <1 wt% of bulk mode, as cassiterite, monazite, zircon and Nb-Ta oxides, and allows to demonstrate the presence of Cs-enriched lepidolite, nanpingite, sokolovaite, pezzottaite, and even small grains of dispersed pollucite. Inspection of Al-Si ratios in the EDS-spectra in combination with elemental analysis and optical microscopy allowed to recognise Li- and Be-bearing minerals as petalite, spodumene and beryl. The samples were classified into 21 distinct mineral associations or pegmatite mineral zones. At Bikita, pollucite forms lens-like massive mineralisation with a crosscutting vein network of purple lepidolite, petalite, quartz and feldspar. Pollucite exhibits rather inhomogeneous Cs compositions, ranging from <20 up to 37 wt% Cs₂O, with higher Cs concentrations along the lepidolite-quartz-petalite veins and later generations of cracks. Beryl, partly in several generations with up to 10–15 wt% Cs₂O, was observed at Cattlin Creek, Londonderry and Wodgina. Most common are tantalite, columbite and cassiterite. Other Ta-, Nb- and Sn-oxide minerals ascertained during the study are microlite, wodginite, ixiolite and pyrochlor. The early crystallisation stage in the pegmatites is characterised by euhedral plagioclase, K-feldspar, petalite, spodumene, cassiterite, columbite and tantalite which crystallised directly from the hydrous melt under subsolvus conditions of >5 kbar. The subsequent main stage is characterised by the second generation of plagioclase, and the growth of muscovite and quartz. The position of the massive pollucite mineralisation within the crystallisation sequence at Bikita is difficult to ascertain. Inclusions of petalite and feldspar in pollucite signal its crystallisation during the early stage. Replacement along the lepidolite vein network provides arguments for the late phase of the main stage crystallisation. The late stage crystallisation is characterised by numerous replacement reactions which are attributed to circulation of deuteric hydrothermal fluids in consolidated but still not fully cooled pegmatites.

The whole rock main and trace element geochemical data of the pegmatites is used to decipher the processes which led to the enormous Cs enrichment in the massive pollucite mineralisation. The geochemical study was also designed for potential prediction if a pegmatite potentially might host a yet not recognised massive pollucite mineralisation. Following these goals, the data from the pollucite-bearing Bikita pegmatite field are used for characterisation of the geochemical behaviour of Cs, and is compared to the Western Australian pegmatites. The whole rock geochemical analyses are supplemented by the mineral modes and assemblages gained by automated SEM-MLA (Mineral Liberation Analysis) from the same samples. The Geochemical investigations of the major and minor elements accentuates the solitary characteristics of the massive pollucite mineralisation at Bikita.

The knowledge of the age of an LCT pegmatite mineralisation is one of the key exploration criteria as most of the major LCT pegmatite systems (e.g., Tanco, Greenbushes, Bikita, Wodgina) were formed in relative narrow time intervals within the earth history. In addition, the temporal and spatial relation-ships of the pegmatites to potential source granites, their crystallisation history and geotectonic settings are still under debate. For the evaluation of the petrogenesis and emplacement history we compared Meso- and Neo-Archean LCT pegmatites of the Zimbabwe (Bikita), Pilbara (Wodgina) and Yilgarn Cratons (Londonderry, Mount Deans, Cattlin Creek). Lepidolite and white mica and lepidolite are an abundant constituent of the investigated LCT pegmatites and potentially accessible for ⁴⁰Ar/³⁹Ar dating. The U–Pb dating of Ta–Nb–Sn oxides and cassiterite LA-ICPMS turned out to be suitable for the Neo-Archean pegmatites due to the widespread occurrence of these minerals as accessories in almost all samples. Furthermore, they are formed at a virtually constant level during pegmatite crystallisation from main crystallisation to late state hydrothermal overprint. The ⁴⁰Ar/³⁹Ar ages cover a large spectrum from Neoarchean (~2630 Ma) to Paleoproterozoic (~2316 Ma), and are verified by U/Pb tantalite/ columbite ages (~2870 to 2615 Ma, LA-ICP-MS) and by Th-U-Pb electron microprobe monazite ages (~2700 Ma; ~2500 Ma). Micas from the Yilgarn Craton yield Neoarchean cooling ages indicating an immediate cooling after crystallisation. In contrast, micas from the Zimbabwe (~2625 Ma) and Pilbara Craton (~2870 Ma) exhibit Paleoproterozoic cooling ages that significantly postdate initial crystallisation. This is in good agreement with petrographic data that suggests a post pegmatite hydrothermal overprint. Overall, our new age data are in good agreement with a previously postulated global major LCT pegmatite events between 2850 to 2800 Ma and 2650 to 2600 Ma. During this event specific geodynamic conditions (i.e. a super-continent assembly) and associated anormal high heat flow from the mantle enabled the global formation of large volumes of LCT pegmatites.

Samarium and Nd isotope data are used to provide information on the source of melts, as well as to determine the age of the rocks. Several samples yield extreme high ¹⁴⁷Sm/¹⁴⁴Nd ratios of >0.3, indicating extreme fractionation of the REE. This is also reflected in their calculated present day εNd values of εNd_{t = 0} at >40.0. The reason for the high degree of REE fractionation remains ambiguous but is suggested either to be related to the late stage replacement processes and related element redistribution (e.g., feldspar replacement, replacement of the primary Nb-Ta-Sn-oxides), or to reflect the mineralogical composition of the samples. The calculated fractionation factor for the LCT pegmatites are compared to the geological background values for possible source rocks from the Zimbabwe, Yilgarn and Pilbara Craton. It is obvious that the samples that exhibit elevated ¹⁴⁷Sm/¹⁴⁴Nd ratios plot in the field of REE depletion. This further supports the suggestion that late stage replacement processes and related element redistribution affected the LCT pegmatites. Most of the LCT pegmatite samples have Nd isotopic compositions close to depleted mantle, suggesting that they might have been derived directly from a depleted mantle. Analyses of the Li isotope system are increasingly used to trace geological processes in LCT pegmatite systems and involve magmatic or hydrothermal alteration processes. Lithium abundance and isotope composition was determined from selected LCT pegmatite mineral phases (feldspar, quartz, mica, pollucite, petalite, garnet, beryl, tourmaline, spodumene). The δ⁷Li values range between 0.06 ‰ to 31.92‰. When compared to data from different granitic systems worldwide, the LCT pegmatites display higher δ⁷Li values than most of the granites. It can be expected during magmatic fractionation that the incompatible Li becomes more enriched in the residual melt or fluid. Fluid inclusions were encountered in almost all thin sections from the Bikita and Mount Tinstone pegmatite (Wodgina). Fluid inclusions of selected mineral phases from the Bikita and Wodgina LCT pegmatites recorded comparable homogenisation temperatures which range from approximately 200–450 °C for Bikita and 200–500 °C for the Wodgina. Quartz and the Li minerals petalite (in Bikita) and spodumene (Wodgina) yield higher temperatures (300–450 °C). Pollucite (Bikita) and apatite (Bikita and Wodgina) exhibit lower entrapment temperatures (200–300 °C). This is in good agreement with the general crystallisation sequence, with most of the quartz, petalite and spodumene representing early stage minerals, whereas apatite and the pollucite from the massive pollucite mineralisation were formed during the main to late stage of the crystallisation. The δ¹³C_CO2 values of fluid inclusions point to a mafic or ultramafic source, either directly from the mantle or MORB, or alternatively from local remobilisation of the surrounding greenstone belt lithologies.

The 2650 to 2600 Ma LCT pegmatite formation within the Yilgarn, Zimbabwe and Superior cratons defines the first major LCT pegmatite formation event worldwide. Meso-Archean age pegmatites at Wodgina on the Pilbara Craton coincide with an earlier and minor formation event. The LCT pegmatites are commonly hosted in greenstone belt lithologies. It is possible to identify adjacent granitoid suites within the same age span, but no direct field evidence for a connection to potential source granites could be observed. Nd isotopic compositions of the pegmatites are close to depleted mantle, suggesting only minor Archean crustal contamination. Formation of massive pollucite mineralisation is favoured in flat lying and gently dipping large LCT pegmatite sheets. Whole rock geochemical data of the mineral zones in the studied pegmatites points to the classical fractional crystallisation with enrichment of incompatible elements. But the considerable gap in the Cs contents between the mineral zones and the massive pollucite mineralisation signals a second stage with an extreme enrichment of Cs, distinctly separated from the general development of LCT pegmatites. It is proposed that this Cs enrichment stage is initiated by melt/fluid immiscibility with separation of a melt with Cs-analcime composition, followed by enrichment of Cs in analcime melt droplets and accumulation in upper portions of a pegmatite sheet. After the accumulation, a transition to a fluid-controlled Cs enrichment of the melt toward Na-pollucite compositions takes place. The final pegmatite crystallisation with formation of cracks in massive pollucite mineralisation, passes to the late stage hydrothermal Cs-enrichment stage within a lepidolite vein network.