- Email: [email protected]
The emplacement of class 1 kimberlites
Available online at www.sciencedirect.com
Journal of Volcanology and Geothermal Research 174 (2008) 40 – 48 www.elsevier.com/locate/jvolgeores
The emplacement of class 1 kimberlites E.M.W. Skinner ⁎ Geology Department, Rhodes University, Grahamstown 6140, South Africa Accepted 11 December 2007 Available online 10 January 2008
Abstract Class 1 kimberlites are distinct from Class 2 and Class 3, in that they are characterized by three zones, namely; root, diatreme and crater. Class 2 and 3 kimberlites do not form diatreme zones filled with diatreme-facies rocks. Consequently these kimberlites probably form by different processes. Within the root zones of Class 1 kimberlites, subvolcanic contact breccias occur at particular horizons within the immediate wall rocks. These could be formed by hydraulic fracturing caused by crystallization-induced exsolution of juvenile volatiles within the cooling hypabyssal kimberlite. This exsolution may subsequently also lead to the formation of transitional-facies kimberlites in the uppermost parts of extended columns or plugs of hypabyssal kimberlite. From such settings, break through to surface and explosive eruption is thought to occur and pipes having characteristic diatremes slope angles of 82° and 500–700 m deep craters are formed. Through these processes these kimberlites undergo fundamental textural changes to produce a variety of rocks from hypabyssal-, through transitional-, to diatreme- and crater-facies kimberlites. © 2008 Published by Elsevier B.V. Keywords: kimberlite; emplacement; exsolution; juvenile volatiles
1. Introduction The exsolution of juvenile volatiles initiated at great depth and continued to levels of around 3 km from surface essentially drives the emplacement of southern African type (Class 1) kimberlite pipes. This exsolution leads initially to the production of subvolcanic contact breccias (as defined by Clement, 1982) and subsequently to the explosive eruption of relatively large volumes of kimberlite in single powerful blasts. These blasts produce a vent or diatreme, initially flared from about 2.2 km all the way to surface at on average 82°. This exsolution also leads to the production of a unique suite of rocks that change from hypabyssal-, through transitional-to diatreme-and crater-facies rocks present in three zones, namely root, diatreme and crater zones (re. Clement, 1982; Clement and Reid, 1989). The specific geological features and kimberlite rock types are not evident in the Victor/Fort a la Corne type (Class 2) nor the Koala/Jwaneng type (Class 3) kimberlites. Consequently, it is likely that emplacement processes for Class 1 kimberlites are quite different from those of the other two classes. ⁎ Tel.: +27 82 445 1021; fax: +27 46 622 3167. E-mail address: [email protected]. 0377-0273/$ - see front matter © 2008 Published by Elsevier B.V. doi:10.1016/j.jvolgeores.2007.12.022
The big-bang/bottom-up model presented here for Class 1 kimberlites is in contrast to the episodic/up-down model for the emplacement of all kimberlites as postulated by Sparks et al. (2006). This model is opposed to the phreatomagmatic models of Lorenz (many publications e.g. from Lorenz, 1975 to Kurszlaukis and Lorenz, 2006). Clement (1982) and Clement and Reid (1989) first presented ideas on this type of model, which Skinner and Marsh (2004) reinforced. 2. Kimberlite zones Class 1 kimberlite pipes typically consist of three separate zones (Fig. 1). From the bottom up these include: (a) root zones, where the margins tend to be highly irregular, (b) diatreme zones, where the margins are smooth, consistently at a slope of about 82° and (c) crater zones, which in most cases consist of an upper, larger, flared part and a lower, smaller, narrower part. 2.1. Root zones Many (N10) different intrusions of hypabyssal kimberlites (HK), mostly of macrocrystic, calcite–serpentine–phlogopite– monticellite kimberlites fill the root zones of single pipes
E.M.W. Skinner / Journal of Volcanology and Geothermal Research 174 (2008) 40–48
Fig. 1. Composite model of a (Class 1) kimberlite pipe (RVK = Resedimented volcaniclastic kimberlite, PK = Pyroclastic kimberlite, TK = Tuffisitic kimberlite, TFK = Transitional-facies kimberlite, HK = Hypabyssal kimberlite).
(Clement, 1982). Emplacement occurs essentially through magmatic stoping and wall rock contacts are irregular. The kimberlite incorporates a limited amount of country rock material, which becomes highly altered. The root zones extend from about 3 km from the original surface (maximum depth of exposure on Class 1 kimberlite mines) up to at least 1.65 km from surface. In their model of kimberlite emplacement, Clement and Reid (1989) infer an upward migration of intermittent embryonic columns of HK to within 500 m of the surface. But in no cases do Class 1 HKs reach the surface, and Class 1 kimberlite lavas have not been found. The absence of Class 1 kimberlite lavas (re. Mitchell, 2006) is important because this means that Class 1 HKs crystallize and degas before they reach the surface. This may be explained by the possibility that water-rich, kimberlite magmas behave in a similar fashion to water-rich peridotite. Wyllie (1987) shows that the solidus of water-rich peridotite is deflected towards lower temperatures at low pressure, while that of CO2-rich peridotite (like dry basalt) is deflected to higher temperatures at low pressure (Fig. 2). The implications of this are that the intrusion path of a relatively water-rich kimberlite may cross the solidus at pressures around 330 bars; 1 km below the surface, whereas the intrusion path of a CO2-rich kimberlite may not. Calcite-rich, Class 2 kimberlites clearly survive up to the surface as hot magmas, whereas relatively water-rich, Class 1 kimberlites do not. Also, present within the root zones are what Clement and Reid (1989) refer to as “explosion and fluidization” breccias (here referred to as contact breccias, Fig. 3). The former consist mostly of in situ, angular, country rock clasts whereas the latter consist mostly of little-displaced but well-rounded and partly
41
rotated country rock clasts. In both cases, these breccias are clast-supported and are essentially kimberlite free. Their monolithic character, their location under undisturbed country rock overhangs, their interlocking nature and the outward graduation from highly fragmented material to undisturbed wall rock are indications of in situ brecciation. The fragments in these breccias are in most cases closely packed and as such have the appearance of country rock, which has been intensely shattered. In other cases, the packing is less dense and cavities up to 10 cm may occur. In most cases, the cavities are empty although calcite and other minerals have crystallized on cavity surfaces. However, some (N 20%) of these contact breccias have cavities partly to completely impregnated by hypabyssal kimberlite. In no cases are these filled or partly filled with tuffisitic kimberlite, even in cases where such breccias are cut by later tuffisitic kimberlite. As most are essentially free of kimberlite, it is reasonable to conclude that the brecciation occurs in association with a free gas phase that forms during crystallization-induced, exsolution. A state of over-pressure between the intrusive column of hypabyssal kimberlite and the wall rock generates fracturing (Burnham, 1985). These breccias are considered to form entirely subvolcanically, prior to explosive eruption of tuffisitic kimberlite, because they are never impregnated with tuffisitic kimberlite. Lower parts of the pipe (towards the base of the diatreme zone and towards the top of the root zone) may also have contact zones of transitional kimberlites here referred to as transitionalfacies kimberlites (TFK), or transitional hypabyssal and transitional tuffisitic kimberlites (HKt or TKtB, Hetman et al., 2004). However, extended columns of TFKs into the diatreme zone without associated HKs and TKs also exist (e.g as is the case in the Dark Piebald kimberlite at Premier Mine; Bartlett, 1998; Skinner and Marsh, 2004). 2.2. Diatreme zones Diatreme zones are filled mainly with relatively few (b4) different varieties of tuffisitic kimberlite, that are restricted to
Fig. 2. A schematic illustration of the differences between possible watersaturated and CO2-saturated kimberlite solidi (after peridotite solidi, Wyllie, 1987). Arrow ab shows an adiabatic trajectory within water-rich magma whereas arrow cd is within a magma richer in CO2.
42
E.M.W. Skinner / Journal of Volcanology and Geothermal Research 174 (2008) 40–48
Fig. 3. (a) Two cross sections through the De Beers pipe showing different horizons of contact breccia (Clement, 1982). (b) A schematic illustration of a tunnel developed through a contact breccia on the 585 m level of the De Beers pipe. The contact breccia on the left hand side of the tunnel consists of in situ country rock breccia fragments filled in between by hypabyssal kimberlite. Halfway up the tunnel the fragments are only partly infilled with kimberlite (50%) whereas at the end of the tunnel the breccia is open with no kimberlite in between the fragments (Clement, 1982).
the subsurface vent or diatreme. In the case of the Premier Mine, proven TK in the Grey kimberlite extends from the present surface to the 1100 m level and if the contacts of the Grey kimberlite are projected downward at 82° the limbs meet at 1600 m. Thus, in this case the extent of the DFK is at least 1.6 km and assuming a crater zone of 600 m (e.g. 600 m for Orapa south lobe and 700 m for Mwadui) the DFK would extend to a depth of 2.2 km from the original surface. This is similar to the depths of DFK in Kimberley if the Hawthorne (1975) estimates of the extent of erosion are accurate. Hanson et al. (2006) argue differently but the veracity of their argument may be questionable due to the fact that floating reef appears to
sink to greater depths in most Group I compared to Group II kimberlite pipes. Note that the contacts between TK and wall rock are in most cases (99%) sharp and there is little evidence of faulting, plucking, reaction or alteration. There is no brecciation of wall rock beyond the contacts but, as described previously, some diatremes do cut through contact breccia zones interpreted as having formed by earlier subvolcanic processes. Tuffisitic kimberlites are thought to represent the product of a continuing process of 1st boiling (exsolution of juvenile volatiles due to depressurization) and 2nd boiling (exsolution of juvenile volatiles due to crystallization) that initially led to the formation of contact breccias and TFKs. At higher levels (lower
E.M.W. Skinner / Journal of Volcanology and Geothermal Research 174 (2008) 40–48
confining pressure) of less than 1 km (330 bars) the supercritical gas overpressures generated by magmatic volatile exsolution, are thought to lead to cracking through to surface. As soon as this occurs there is a large explosion (of uncertain extent) resulting from depressurization from about 330 bars to 1 bar (atmospheric pressure). Accompanying this depressurization, three things are thought to happen: (a) There is renewed exsolution of all remaining free volatiles from the rest of the kimberlite magma column from about 1 km down to about 2.3 km from the surface, (b) there is considerable cooling as a direct consequence of depressurization according to the ideal gas law (PV = nRT), and (c) the explosion is thought to generate a compressive shock wave (Rice, pers comm.). Interference between the outward compressive shock wave and the seismic rebounding rarefaction shock wave reflected from the surface is thought to generate a shatter-cone with slope angles of 82° (Rice, 1999) and sharp wall-rock contacts. Considerable upward and outward spalling into the atmosphere of shattered cap rock material is thought to occur as indicated by a relative scarcity of cap rock xenoliths (total country rock xenolith contribution) of as little as b 5 vol.%, in some TKs. The explosive blast must also generate a tuff cone, which is likely to encircle the original crater. This tuff cone is indicated by the presence of RVK that must be formed by erosion and redeposition of kimberlite material back into the crater.
43
At the time of the explosion, in some kimberlites relatively large blocks of cap rock material fall back into the diatreme vent and are caught up in the maelstrom of a violently fluidized mixture of solids (olivine + kimberlite magmaclasts + country rock xenoliths) and gas. Fluidization is thought to occur following the guidelines of McCallum (1985) and Gernon et al. (2006). Some of the blocks sink to the base of the fluidized system at about 2.2 km below the original surface, particularly in micaceous Group II pipes but many are buoyed-up in the system, particularly in Group I pipes. Spouting or overturning results in homogenization of much of the fluidized product particularly within the central parts of the vent. This is shown by the homogeneous diamond distributions and by the remarkable petrographic similarity both horizontally and vertically over depths of N 1 km, as is the case in studies undertaken by the author at Premier Mine with respect to the Brown, Grey and Black kimberlites. The product of the fluidized system is essentially a homogeneous, massive volcaniclastic kimberlite but in some (b 5%) cases, discernable layers are evident in parts of the diatreme vent at limited depth locations close to the side walls. These layers do not extend all the way across the vent and the layering is evident only on a macro scale (indicated by layers of mixed larger rock clasts) but is not evident on a micro scale. This material has a clear matrix of tuffisitic kimberlite and must initially, have been mixed. The layered material is formed
Fig. 4. Model of Class 1 kimberlite crater formation.
44
E.M.W. Skinner / Journal of Volcanology and Geothermal Research 174 (2008) 40–48
either by a variance in the extent of fluidization between the side-walls and the centre of the diatreme or forms by sinking of layered material derived from a pre-existing earlier crater material. 2.3. Crater zones The nature of the crater facies kimberlites indicate initial relatively rapid deposition of pyroclastic kimberlite within the original crater (500–700 m deep “hole in the ground”) emptied by the earlier explosive eruption. This is then followed by later side-wall collapse, crater enlargement and flaring and deposition of debris flow deposits. Then, over a longer time, the remaining crater is filled with conglomerates, grits and sandstones (resedimented volcaniclastic kimberlites) derived essentially from presumable relict tuff ring deposits. Finally, the remaining crater is filled by much finer-grained, crater-lake, epiclastic deposits consisting (in the pipes that the author has studied, including Orapa, Mwadui and several pipes in Angola and Siberia) mostly of quartz and mud derived from afar (Fig. 4). Note that there is little interplay between diatreme/ crater formation and deposition of DFKs and CFKs indicating a few large explosive blasts and creation of a hole in the ground, followed by deposition of first DFKs within the diatreme and then CFKs within the crater. 3. The transition from root to diatreme Recognition of the transition from HK, through TFK to DFK (illustrated in Fig. 5) is a relatively new development (e.g. Skinner and Marsh, 2004; Hetman et al., 2004). TFKs are real (see Section 3.2), they are not a form of pyroclastic kimberlite that can be ascribed to as “agglutinates, welded tuffs, vapour
phase zones in ignimbrites, rheomorphic tuffs or clastogenic lavas” (as suggested by Sparks et al., 2006). Although diatremefacies, hard, relatively fresh, agglutinated or welded tuffaceous kimberlites similar to tuffisitic kimberlites do exist (e.g the LM1 at Letlhakane DK1 pipe, Botswana and the Fragmental kimberlite at Kao pipe, Lesotho). Now that TFKs are recognized, more and more examples are being found in similar Class 1 settings all over the world (e.g. Masun and Scott Smith, 2006). Their recognition and the correct interpretation of what they represent is critical in understanding the true processes responsible for the emplacement of South African type (Class 1) kimberlites. Having shown this, it is remarkable that in the past, no one appears to have alluded to the extreme petrographic changes that have to occur in the conversion from hypabyssal to diatreme-facies kimberlites, regardless of the TFKs, which also occur. 3.1. Root zone hypabyssal kimberlites HKs show that, relative to other igneous rocks, kimberlites are unusual in that they contain high proportions of xenocrysts (typically 25% by volume) mostly of olivine. In addition to xenocrystic olivine, primary phenocrystic olivine starts to crystallize at a relatively early stage so that, even prior to rapid emplacement into the Earth's crust, the kimberlite carries around 25 vol.% olivine phenocrysts, which together with xenocrysts make up 50 vol.% of the magma. This relatively high solid content creates an effective viscosity for the kimberlite magma that is relatively high (Costa, 2005). Olivine continues to crystallize from the magma down to temperatures below 1000 °C when, in mainly Class 1, Group I kimberlites, monticellite crystallizes rapidly en masse as micro-phenocrysts generally b0.05 mm in size (Fig. 6a). In Class 1, Group II
Fig. 5. Venetia K1 pipe. Significant mineralogical, textural and chemical changes evident from detailed petrographic observation, accompany the transition from hypabyssal-(a), through transitional-(b) to diatreme-facies (c) kimberlites.
E.M.W. Skinner / Journal of Volcanology and Geothermal Research 174 (2008) 40–48
45
Fig. 6. Hypabyssal kimberlites from (a) Group I, Dutoitspan Mine and (b) Group II, Star Mine. Both have bases of serpentine and lesser calcite.
kimberlites mainly phlogopite crystallizes (Fig. 6b). This crystallization renders these magmas immobile and further emplacement to the surface as a magma is inhibited. It is unlikely that these kimberlites would see the surface, if they were not subjected to later diatreme-forming processes. In the case of Class 2 kimberlites the magmas are carbonate rich, the groundmass does not crystallize en masse, they are relatively less viscous, they can remain as hot magmas all the way to the surface and apparently do not undergo substantial exsolution of juvenile volatiles and the formation of transitional and diatremefacies kimberlites. In some kimberlites textural evidence suggests that melilite may crystallize after olivine but before monticellite (Skinner et al., 1999) but other minerals that crystallize after monticellite include
phlogopite, spinels, perovskite, apatite (at ∼750 °C), calcite (at ∼650 °C) and serpentine (at ∼350 °C). All the earlier minerals crystallize as euhedral crystals within the serpentine, which has all the appearances of having formed from, devitrified, late-crystallizing gel. In many HKs, secondary serpentine forms as a result of the well-known serpentinization of olivine and monticellite but this serpentine is different from the serpentine or rather serpophite (Michell and Putnis, 1988) occurring in the base of the kimberlite. In most HKs, the earlier crystallizing, euhedral primary minerals have an intimate textural relationship with the later crystallizing, interstitial serpophite, which in some cases completely encloses the earlier minerals. These minerals could not have been left suspended in a void, which was then later filled by secondary serpentine, unrelated to the kimberlite event. This relationship
Fig. 7. (a) Finsch F2. Pelletal/globular segregationary, macrocrystic kimberlite located within the core zone of an otherwise uniform, macrocrystic kimberlite. In most cases, altered olivine macrocrysts and smaller phenocrysts, surrounded by dark selvages of mainly very fine-grained phlogopite and lesser microlitic diopside, are set in an inter-pelletal matrix of serpentine and calcite (from Clement, 1982). (b) Weasua K1, Liberia. Parts of pelletal segregations of mainly altered phlogopite and diopside needles which protrude outwards into a matrix of serpentine (S) and calcite (C).
46
E.M.W. Skinner / Journal of Volcanology and Geothermal Research 174 (2008) 40–48
leaves little room for doubt that this serpentine is primary. Note that the compositions of typical HKs have relatively high H2O+ and CO2 contents (e.g. average 7.7 and 6.15 wt.% respectively, Le Roex et al., 2003). Much of the water is contained in the serpentine. 3.2. Transitional-facies kimberlites Skinner and Marsh (2004) and Hetman et al. (2004) describe the petrographic features of TFKs. The essential features of these rocks include the appearance of diopside, the disappearance of calcite and the development of spherical, segregationary structures in rocks that occur in some cases in the contact zones between hypabyssal- and diatreme-facies kimberlites. In these cases, the physical contacts between HK and TFK and between TFK and DFK are vague and transitional; the different kimberlite types grade or transcend into one another and no obvious fixed boundaries exist. However, the transitional-facies kimberlites in such circumstances occur in settings below or adjacent to the DFK. In most cases these transitional-facies kimberlites exhibit little evidence of fragmentation or transportation. In other cases, TFKs occur as distinct dykes or dyke-like bodies or separate vertical plugs of kimberlite not associated with DFK. In the latter case, these TFKs could be referred to as arrested DFKs, where the full transition from HK to DFK did not occur, or is hidden, or was affectively aborted. Good examples of this sort of kimberlite occur at Finsch (F2 kimberlite) at Kao (Gritty kimberlite) and at Premier (Dark Piebald kimberlite). The TFKs range in petrographic character from magmatic-like TFKs or HKts (Hetman et al., 2004) to diatreme-like TFKs or TKts (Hetman et al., 2004). In extreme variants of magmatic types, isolated patches of microlitic diopside are evident (Fig. 5b) or globular segregationary textures are evident (Fig. 7a and b) or
both (Fig. 7a). In addition calcite may be present in the base of the rock and olivine macrocrysts may be relatively fresh. As one moves away from these magmatic types towards diatreme types, microlitic diopside becomes more abundant and relatively finergrained; segregationary textures or pelletal structures become more evident; calcite disappears and olivine macrocrysts become altered. In most cases, no new country rock xenoliths (CRX) are added to the system and those that are present are highly altered. However, Hetman et al., 2004 indicate that in the case of the Gahcho Kue kimberlites, some new CRX may be added to the transitional kimberlites. In the HKts these CRX are highly altered but in the TKts they are much less altered. In some cases there is evidence of limited flow and deformation in the TFKs. The segregationary/pelletal structures (called many things by different authors including; magmaclasts, pelletal or globular lapilli, juvenile lapilli or pyroclasts) contain all the mineralogical and textural characteristics of the HKs. If not too altered and partly replaced by ultra-fine diopside, they are seen to consist most often of one or more olivine grains surrounded by finer grained primary minerals similar to those presented in Fig. 5a and b. In most cases these minerals have the identical compositions and the same range in grain sizes of the same minerals present in typical HKs, indicating that these structures have crystallized under similar conditions as normal HKs. The inter-pelletal matrix consists mostly of serpentine. In rare cases calcite may also be present (e.g. Fig. 7a and b) but as one moves away from magmatic types calcite soon disappears. The variation in textures and mineralogy in TFKs suggest that they represent normal, Class 1 type HKs that have been texturally and mineralogically modified, in situ. The composition of the pelletal structures indicates that these rocks are in the process of rapid, en masse crystallization of the groundmass. Contemporaneously with this crystallization, the kimberlite
Fig. 8. (a) Koffiefontein, tuffisitic kimberlite consisting of altered olivine grains, country rock xenoliths (CRX) and pelletal magmaclasts set in a matrix of serpentine and clay minerals. The magmaclasts consist of smaller olivine grains plus brown ultra fine-grained material. Note that in most cases the CRX occur outside of the magmaclasts. (b) Letseng Satellite, tuffisitic kimberlite containing relatively distinct magmaclasts, where some centres are relatively well-preserved (note fine, granular monticellite and evenly distributed fine opaques) and the margins are coated by ultra fine, dark, microlitic diopside.
E.M.W. Skinner / Journal of Volcanology and Geothermal Research 174 (2008) 40–48
magma segregates into pelletal structures and an inter-pelletal matrix of serpentine and calcite. The latter minerals are considered to be the product of crystallization of a volatilerich gas phase that separated from the earlier crystallizing minerals in an aerosol-like process thought to be driven by second boiling equivalent to vesiculation in other volcanic rocks. Note that a fixed boundary between the pelletal structures and the matrix is not evident as delicate, needle-like diopside crystals project outwards from the structures into the matrix (Fig. 7b). This texture indicates at least two things: (i) that the formation of the segregations is unlikely to be due to liquid immiscibility and (ii) that the crystallization of diopside occurs at the same time as the formation of the pelletal structures and not long afterwards through secondary alteration. In most cases TFKs are initiated at depths of between 3 and 1 km under relatively quiescent conditions in magmatic settings. They occur in a range of situations: within dykes independent of pipes (Camp Alpha dyke in Liberia), in dykes or dyke-like bodies within pipes (e.g. F2 and F4 at Finsch), within discrete but irregular intrusive plugs within pipes (e.g. Dark Piebald kimberlite at Premier and the Gritty kimberlite at Kao) and at the interface between HK and DFK in pipes (e.g. in several Kimberley pipes, at Venetia, at Premier and at Gahcho Kue). 3.3. Diatreme-facies kimberlites DFKs are generally massive, unsorted, matrix-to clastsupported, fragmental rocks consisting essentially of abundant serpentinized olivine macrocrysts and phenocrysts (in an approximate 50:50 ratio), abundant magmaclasts (for terminology refer to Field and Scott Smith, 1998) and variable proportions of country rock xenoliths and xenocrysts set in an inter-fragmental matrix dominated by primary serpentine (e.g. Fig. 8a). In many cases (N 50%) serpentinized olivine and serpentine in the matrix is replaced by a combination of mostly clay minerals plus pectolite and secondary garnet and titanate may also be present. As in the case in TFKs, diopside is abundant but in the case of DFKs it is in most cases (N 95%) typically ultra fine-grained and is not easily identifiable. But, in some cases this diopside can be microscopically resolved at high power as tiny laths with reduced birefringence and with inclined extinction. Microlitic diopside in DFKs occurs in several different settings: (a) as the replacement product of some finer-grained, primary matrix minerals within variably sized magmaclasts, (b) as thin, but recognizable, coatings on all larger sized components, including individual olivine grains, country rock xenoliths and xenocrysts and kimberlite magmaclasts (e.g. Fig. 8b), and (c) as delicate, individual crystals or clots of crystals within the primary serpentine matrix. In many cases (N80%) the original, small, primary groundmass minerals in the magmaclasts are extensively replaced by the ultra fine-grained diopside, making it difficult to identify, however, in some kimberlites (e.g. in one of the Letseng Satellite DFKs) these primary minerals are relatively well-preserved. Note in most TFKs these primary groundmass minerals are not as extensively replaced as in DFKs and the thin ultra-fine diopside coatings
47
around larger sized components are not as prominent. These coatings are thought to crystallize mainly out of the gas cloud formed by extensive exsolution of the remaining magma column (Skinner and Marsh, 2004) at the time of the explosion. This process may also be responsible for the crystallization of the primary serpentine in the inter-fragmental matrix. Likely crystallization activity and reactions that happen at this time occur in a CO2 depleted environment. The loss of CO2 during the development of TFKs prior to diatreme formation suggests that CO2 is either absent or has a restricted involvement in the diatreme-forming event. On the other hand, the presence of abundant serpentine in DFKs indicates high water activity. The fine diopside coatings on all larger sized components appear as a primary feature and the intimate relationship between delicate, haphazardly orientated individual needles of diopside or clots of diopside crystals present in the base of serpentine also represent primary features that are unlikely to be produced by late stage, secondary processes, unrelated to the primary diatreme-forming event. The diatreme-forming process occurs at a relatively low temperature as indicated by the inclusion of new, country rock xenoliths that are relatively fresh and unaffected by high temperatures. This may be supported in some DFKs by the presence of high grade coal fragments with ignition temperatures of possibly b 350 °C. These new xenoliths are excluded from kimberlite magmaclasts demonstrating that the explosion postdates the segregation of the magmaclasts. 4. Conclusions In Class 1 kimberlite pipes, at levels of between 3 and 1 km from surface: (1) En masse crystallization of the residual kimberlite liquid occurs. (2) This leads to the exsolution of juvenile volatiles. (3) At points where this occurs the hypabyssal-facies kimberlite is transformed into transitionalfacies kimberlites when pelletal structures begin to form and some exsolved gases crystallize as calcite and serpentine. Soon hereafter replacement of the primary groundmass minerals in the pelletal structures occurs. (4) Volatile exsolution is accompanied by volume increase and this leads to an increase in PΔV and subvolcanic breccias are formed in the surrounding side-walls by hydraulic fracturing. (5) Further exsolution at higher levels and lower pressures leads to cracking through to surface. (6) Traumatic decompression from about 330 bars to 1 bar leads to extensive exsolution of the entire magma column down to 2.2 km, further increase in PΔV and a large single explosion is generated. The point of this explosion is generated within the transitional-facies domain. (7) The consequence of the explosion is the production of a shock wave that rebounds from the surface. (8) A cone of brecciation having a 82° slope shaped upwards from the explosion cavity occurs. (9) Considerable cap rock material spalls off into the atmosphere but some large blocks sink downwards. (10) Smaller, new country rock xenoliths plus olivine macrocrysts and kimberlite magmaclasts are fluidized, overturned and thoroughly mixed by spouting. (11) Microlitic diopside and serpentine crystallize out of the gas cloud at relatively low temperatures (b 550 °C) and
48
E.M.W. Skinner / Journal of Volcanology and Geothermal Research 174 (2008) 40–48
the tuffisitic kimberlites are formed. (12) A tuff ring is deposited around the edges of a 500–700 m deep crater. (13) Pyroclastic kimberlites are emplaced into the lower part of the crater. (14) After some time the side walls of the crater collapse under the weight of the tuff ring, forming flared walls in the upper part of the crater and debris flow coarse breccias and high angle grain flow resedimented volcaniclastic kimberlites are deposited. (15) Continuing erosion of the remainder of the tuff ring and the surrounding weathered country rocks adds to the infill of the crater with coarse-grained, proximal-facies grits and finergrained, distal-facies sandstones. (16) The processes outlined above may be repeated but the number of repetitions are limited to only a few (b 3) events. (17) Final infill takes the form of epiclastic, crater lake deposits. (18) Over time considerable weathering in the pipe occurs, particularly of the friable, fragmental rocks. Acknowledgements The reviewers Casey Hetman and James Head and editor Barbara Scott Smith are thanked for their invaluable advice and considerable assistance in the case of the latter. Colleagues in the Geology Department, Rhodes University are thanked for their participation and particularly Steffen Buttner for his help in computing. References Bartlett, P.J., 1998. Premier Mine. Large mines field excursion guide. 7th International Kimberlite Conference, Cape Town, pp. 39–49. Burnham, C.W., 1985. Energy release in subvolcanic environments: implications for breccia formation. Econ. Geol. 80, 1515–1522. Clement, C.R., 1982. A comparative geological study of some major kimberlite pipes in the Northern Cape and Orange Free State. Ph.D. thesis University of Cape Town. Clement, C.R., Reid, A.M., 1989. The origin of kimberlite pipes: an interpretation based on a synthesis of geological features displayed by southern African occurrences. In: Ross, et al. (Eds.), Kimberlites and Related Rocks. Geol. Soc., vol. 14. Australia, pp. 632–646. Costa, A., 2005. Viscosity of high crystal content melts: dependence on solid fraction. Geophys. Res. Lett. 32, L22308. Field, M., Scott Smith, B.H., 1998. Textural and genetic classification schemes for kimberlites: a new perspective. Extended Abstracts, 7th International Kimberlite Conference, Cape Town, pp. 214–216.
Gernon, T.M., Gilbertson, M.A., Sparks, R.S.J., Walters, A., Field, M., 2006. Gas-solid fluidization in an experimental tapered bed: insights into processes in diverging volcanic conduits. Long Abstracts. Kimberlite Emplacement Workshop. Canada, Saskatoon (unpaginated). Hanson, E.K., Moore, J.M., Robey, J., Bordy, E.M., Marsh, J.S., 2006. Reestimation of erosional levels in Group I and II kimberlites between Lesotho, Kimberley and Victoria West, South Africa. Long Abstracts. Kimberlite Emplacement Workshop. Canada, Saskatoon (unpaginated). Hawthorne, J.B., 1975. Model of a kimberlite pipe. In: Ahrens, et al. (Eds.), Physics and Chemistry of the Earth, vol. 9, pp. 1–16. Hetman, C.M., Scott Smith, B.H., Paul, J.L., Winter, F., 2004. Geology of the Gahcho Kue kimberlite pipes, NWT, Canada: root to diatreme magmatic transition zone. Lithos 76, 51–74. Kurszlaukis, S., Lorenz, V., 2006. An alternative explanation for South African style “tuffisitic kimberlites”. Long Abstracts. Kimberlite Emplacement Workshop. Canada, Saskatoon (unpaginated). Le Roex, A.P., Bell, D.R., Davis, P., 2003. Petrogenesis of Group I kimberlites from Kimberley, South Africa: evidence from bulk-rock geochemistry. J. Pet. 44 (12), 2261–2286. Lorenz, V., 1975. Formation of phreatomagmatic maar-diatreme volcanoes and its relevance to kimberlite diatremes. Phys. Chem. Earth 9, 17–29. Masun, K., Scott Smith, B.H., 2006. The Pimenta Bueno kimberlite field, Rondonia, Brazil: evidence for tuffisitic kimberlite. Long Abstracts, Kimberlite Emplacement Workshop. Canada, Saskatoon (unpaginated). McCallum, M.E., 1985. Experimental evidence for fluidization processes in breccia pipe formation. Econ. Geol. 80, 1523–1543. Michell, R.H., Putnis, A., 1988. Polygonal serpentine in segregation-textured kimberlite. Can. Miner 26, 991–997. Mitchell, R.H., 2006. Petrology of hypabyssal kimberlites. Long Abstracts. Kimberlite Emplacement Workshop. Canada, Saskatoon (unpaginated). Rice, A., 1999. Can the blasting excavation engineering sciences provide insight into the processes of kimberlite emplacement and eruption? In: Gurney, et al. (Eds.), Proceedings of the 7th International Kimberlite Conference, vol. 2, pp. 699–708. Skinner, E.M.W., Mahotkin, I.L., Grutter, H.S., 1999. Melilite in kimberlites. Proceedings of the 7th International Kimberlite Conference, vol. 1, pp. 788–794. Skinner, E.M.W., Marsh, J.S., 2004. Distinct kimberlite pipe classes with contrasting eruption processes. In: Mitchell, et al. (Eds.), 8th International Kimberlite Conference. Sel. Papers, vol. 1, pp. 183–200. Sparks, R.S.J., Baker, L., Brown, R., Field, M., Schumacher, J., Stripp, G., Walters, A., 2006. Dynamical constraints on kimberlite volcanism. J. Volcanol. Geotherm. Res. 155, 18–48. Wyllie, P.J., 1987. Transfer of subcratonic carbon into kimberlites and rare earth carbonatites. In: Mysen, B.O. (Ed.), Magmatic processes: Physiochem. Principles. Spec. Publ. Geochem. Soc., vol. 1, pp. 107–119.
Journal of Volcanology and Geothermal Research 174 (2008) 40 – 48 www.elsevier.com/locate/jvolgeores
The emplacement of class 1 kimberlites E.M.W. Skinner ⁎ Geology Department, Rhodes University, Grahamstown 6140, South Africa Accepted 11 December 2007 Available online 10 January 2008
Abstract Class 1 kimberlites are distinct from Class 2 and Class 3, in that they are characterized by three zones, namely; root, diatreme and crater. Class 2 and 3 kimberlites do not form diatreme zones filled with diatreme-facies rocks. Consequently these kimberlites probably form by different processes. Within the root zones of Class 1 kimberlites, subvolcanic contact breccias occur at particular horizons within the immediate wall rocks. These could be formed by hydraulic fracturing caused by crystallization-induced exsolution of juvenile volatiles within the cooling hypabyssal kimberlite. This exsolution may subsequently also lead to the formation of transitional-facies kimberlites in the uppermost parts of extended columns or plugs of hypabyssal kimberlite. From such settings, break through to surface and explosive eruption is thought to occur and pipes having characteristic diatremes slope angles of 82° and 500–700 m deep craters are formed. Through these processes these kimberlites undergo fundamental textural changes to produce a variety of rocks from hypabyssal-, through transitional-, to diatreme- and crater-facies kimberlites. © 2008 Published by Elsevier B.V. Keywords: kimberlite; emplacement; exsolution; juvenile volatiles
1. Introduction The exsolution of juvenile volatiles initiated at great depth and continued to levels of around 3 km from surface essentially drives the emplacement of southern African type (Class 1) kimberlite pipes. This exsolution leads initially to the production of subvolcanic contact breccias (as defined by Clement, 1982) and subsequently to the explosive eruption of relatively large volumes of kimberlite in single powerful blasts. These blasts produce a vent or diatreme, initially flared from about 2.2 km all the way to surface at on average 82°. This exsolution also leads to the production of a unique suite of rocks that change from hypabyssal-, through transitional-to diatreme-and crater-facies rocks present in three zones, namely root, diatreme and crater zones (re. Clement, 1982; Clement and Reid, 1989). The specific geological features and kimberlite rock types are not evident in the Victor/Fort a la Corne type (Class 2) nor the Koala/Jwaneng type (Class 3) kimberlites. Consequently, it is likely that emplacement processes for Class 1 kimberlites are quite different from those of the other two classes. ⁎ Tel.: +27 82 445 1021; fax: +27 46 622 3167. E-mail address: [email protected]. 0377-0273/$ - see front matter © 2008 Published by Elsevier B.V. doi:10.1016/j.jvolgeores.2007.12.022
The big-bang/bottom-up model presented here for Class 1 kimberlites is in contrast to the episodic/up-down model for the emplacement of all kimberlites as postulated by Sparks et al. (2006). This model is opposed to the phreatomagmatic models of Lorenz (many publications e.g. from Lorenz, 1975 to Kurszlaukis and Lorenz, 2006). Clement (1982) and Clement and Reid (1989) first presented ideas on this type of model, which Skinner and Marsh (2004) reinforced. 2. Kimberlite zones Class 1 kimberlite pipes typically consist of three separate zones (Fig. 1). From the bottom up these include: (a) root zones, where the margins tend to be highly irregular, (b) diatreme zones, where the margins are smooth, consistently at a slope of about 82° and (c) crater zones, which in most cases consist of an upper, larger, flared part and a lower, smaller, narrower part. 2.1. Root zones Many (N10) different intrusions of hypabyssal kimberlites (HK), mostly of macrocrystic, calcite–serpentine–phlogopite– monticellite kimberlites fill the root zones of single pipes
E.M.W. Skinner / Journal of Volcanology and Geothermal Research 174 (2008) 40–48
Fig. 1. Composite model of a (Class 1) kimberlite pipe (RVK = Resedimented volcaniclastic kimberlite, PK = Pyroclastic kimberlite, TK = Tuffisitic kimberlite, TFK = Transitional-facies kimberlite, HK = Hypabyssal kimberlite).
(Clement, 1982). Emplacement occurs essentially through magmatic stoping and wall rock contacts are irregular. The kimberlite incorporates a limited amount of country rock material, which becomes highly altered. The root zones extend from about 3 km from the original surface (maximum depth of exposure on Class 1 kimberlite mines) up to at least 1.65 km from surface. In their model of kimberlite emplacement, Clement and Reid (1989) infer an upward migration of intermittent embryonic columns of HK to within 500 m of the surface. But in no cases do Class 1 HKs reach the surface, and Class 1 kimberlite lavas have not been found. The absence of Class 1 kimberlite lavas (re. Mitchell, 2006) is important because this means that Class 1 HKs crystallize and degas before they reach the surface. This may be explained by the possibility that water-rich, kimberlite magmas behave in a similar fashion to water-rich peridotite. Wyllie (1987) shows that the solidus of water-rich peridotite is deflected towards lower temperatures at low pressure, while that of CO2-rich peridotite (like dry basalt) is deflected to higher temperatures at low pressure (Fig. 2). The implications of this are that the intrusion path of a relatively water-rich kimberlite may cross the solidus at pressures around 330 bars; 1 km below the surface, whereas the intrusion path of a CO2-rich kimberlite may not. Calcite-rich, Class 2 kimberlites clearly survive up to the surface as hot magmas, whereas relatively water-rich, Class 1 kimberlites do not. Also, present within the root zones are what Clement and Reid (1989) refer to as “explosion and fluidization” breccias (here referred to as contact breccias, Fig. 3). The former consist mostly of in situ, angular, country rock clasts whereas the latter consist mostly of little-displaced but well-rounded and partly
41
rotated country rock clasts. In both cases, these breccias are clast-supported and are essentially kimberlite free. Their monolithic character, their location under undisturbed country rock overhangs, their interlocking nature and the outward graduation from highly fragmented material to undisturbed wall rock are indications of in situ brecciation. The fragments in these breccias are in most cases closely packed and as such have the appearance of country rock, which has been intensely shattered. In other cases, the packing is less dense and cavities up to 10 cm may occur. In most cases, the cavities are empty although calcite and other minerals have crystallized on cavity surfaces. However, some (N 20%) of these contact breccias have cavities partly to completely impregnated by hypabyssal kimberlite. In no cases are these filled or partly filled with tuffisitic kimberlite, even in cases where such breccias are cut by later tuffisitic kimberlite. As most are essentially free of kimberlite, it is reasonable to conclude that the brecciation occurs in association with a free gas phase that forms during crystallization-induced, exsolution. A state of over-pressure between the intrusive column of hypabyssal kimberlite and the wall rock generates fracturing (Burnham, 1985). These breccias are considered to form entirely subvolcanically, prior to explosive eruption of tuffisitic kimberlite, because they are never impregnated with tuffisitic kimberlite. Lower parts of the pipe (towards the base of the diatreme zone and towards the top of the root zone) may also have contact zones of transitional kimberlites here referred to as transitionalfacies kimberlites (TFK), or transitional hypabyssal and transitional tuffisitic kimberlites (HKt or TKtB, Hetman et al., 2004). However, extended columns of TFKs into the diatreme zone without associated HKs and TKs also exist (e.g as is the case in the Dark Piebald kimberlite at Premier Mine; Bartlett, 1998; Skinner and Marsh, 2004). 2.2. Diatreme zones Diatreme zones are filled mainly with relatively few (b4) different varieties of tuffisitic kimberlite, that are restricted to
Fig. 2. A schematic illustration of the differences between possible watersaturated and CO2-saturated kimberlite solidi (after peridotite solidi, Wyllie, 1987). Arrow ab shows an adiabatic trajectory within water-rich magma whereas arrow cd is within a magma richer in CO2.
42
E.M.W. Skinner / Journal of Volcanology and Geothermal Research 174 (2008) 40–48
Fig. 3. (a) Two cross sections through the De Beers pipe showing different horizons of contact breccia (Clement, 1982). (b) A schematic illustration of a tunnel developed through a contact breccia on the 585 m level of the De Beers pipe. The contact breccia on the left hand side of the tunnel consists of in situ country rock breccia fragments filled in between by hypabyssal kimberlite. Halfway up the tunnel the fragments are only partly infilled with kimberlite (50%) whereas at the end of the tunnel the breccia is open with no kimberlite in between the fragments (Clement, 1982).
the subsurface vent or diatreme. In the case of the Premier Mine, proven TK in the Grey kimberlite extends from the present surface to the 1100 m level and if the contacts of the Grey kimberlite are projected downward at 82° the limbs meet at 1600 m. Thus, in this case the extent of the DFK is at least 1.6 km and assuming a crater zone of 600 m (e.g. 600 m for Orapa south lobe and 700 m for Mwadui) the DFK would extend to a depth of 2.2 km from the original surface. This is similar to the depths of DFK in Kimberley if the Hawthorne (1975) estimates of the extent of erosion are accurate. Hanson et al. (2006) argue differently but the veracity of their argument may be questionable due to the fact that floating reef appears to
sink to greater depths in most Group I compared to Group II kimberlite pipes. Note that the contacts between TK and wall rock are in most cases (99%) sharp and there is little evidence of faulting, plucking, reaction or alteration. There is no brecciation of wall rock beyond the contacts but, as described previously, some diatremes do cut through contact breccia zones interpreted as having formed by earlier subvolcanic processes. Tuffisitic kimberlites are thought to represent the product of a continuing process of 1st boiling (exsolution of juvenile volatiles due to depressurization) and 2nd boiling (exsolution of juvenile volatiles due to crystallization) that initially led to the formation of contact breccias and TFKs. At higher levels (lower
E.M.W. Skinner / Journal of Volcanology and Geothermal Research 174 (2008) 40–48
confining pressure) of less than 1 km (330 bars) the supercritical gas overpressures generated by magmatic volatile exsolution, are thought to lead to cracking through to surface. As soon as this occurs there is a large explosion (of uncertain extent) resulting from depressurization from about 330 bars to 1 bar (atmospheric pressure). Accompanying this depressurization, three things are thought to happen: (a) There is renewed exsolution of all remaining free volatiles from the rest of the kimberlite magma column from about 1 km down to about 2.3 km from the surface, (b) there is considerable cooling as a direct consequence of depressurization according to the ideal gas law (PV = nRT), and (c) the explosion is thought to generate a compressive shock wave (Rice, pers comm.). Interference between the outward compressive shock wave and the seismic rebounding rarefaction shock wave reflected from the surface is thought to generate a shatter-cone with slope angles of 82° (Rice, 1999) and sharp wall-rock contacts. Considerable upward and outward spalling into the atmosphere of shattered cap rock material is thought to occur as indicated by a relative scarcity of cap rock xenoliths (total country rock xenolith contribution) of as little as b 5 vol.%, in some TKs. The explosive blast must also generate a tuff cone, which is likely to encircle the original crater. This tuff cone is indicated by the presence of RVK that must be formed by erosion and redeposition of kimberlite material back into the crater.
43
At the time of the explosion, in some kimberlites relatively large blocks of cap rock material fall back into the diatreme vent and are caught up in the maelstrom of a violently fluidized mixture of solids (olivine + kimberlite magmaclasts + country rock xenoliths) and gas. Fluidization is thought to occur following the guidelines of McCallum (1985) and Gernon et al. (2006). Some of the blocks sink to the base of the fluidized system at about 2.2 km below the original surface, particularly in micaceous Group II pipes but many are buoyed-up in the system, particularly in Group I pipes. Spouting or overturning results in homogenization of much of the fluidized product particularly within the central parts of the vent. This is shown by the homogeneous diamond distributions and by the remarkable petrographic similarity both horizontally and vertically over depths of N 1 km, as is the case in studies undertaken by the author at Premier Mine with respect to the Brown, Grey and Black kimberlites. The product of the fluidized system is essentially a homogeneous, massive volcaniclastic kimberlite but in some (b 5%) cases, discernable layers are evident in parts of the diatreme vent at limited depth locations close to the side walls. These layers do not extend all the way across the vent and the layering is evident only on a macro scale (indicated by layers of mixed larger rock clasts) but is not evident on a micro scale. This material has a clear matrix of tuffisitic kimberlite and must initially, have been mixed. The layered material is formed
Fig. 4. Model of Class 1 kimberlite crater formation.
44
E.M.W. Skinner / Journal of Volcanology and Geothermal Research 174 (2008) 40–48
either by a variance in the extent of fluidization between the side-walls and the centre of the diatreme or forms by sinking of layered material derived from a pre-existing earlier crater material. 2.3. Crater zones The nature of the crater facies kimberlites indicate initial relatively rapid deposition of pyroclastic kimberlite within the original crater (500–700 m deep “hole in the ground”) emptied by the earlier explosive eruption. This is then followed by later side-wall collapse, crater enlargement and flaring and deposition of debris flow deposits. Then, over a longer time, the remaining crater is filled with conglomerates, grits and sandstones (resedimented volcaniclastic kimberlites) derived essentially from presumable relict tuff ring deposits. Finally, the remaining crater is filled by much finer-grained, crater-lake, epiclastic deposits consisting (in the pipes that the author has studied, including Orapa, Mwadui and several pipes in Angola and Siberia) mostly of quartz and mud derived from afar (Fig. 4). Note that there is little interplay between diatreme/ crater formation and deposition of DFKs and CFKs indicating a few large explosive blasts and creation of a hole in the ground, followed by deposition of first DFKs within the diatreme and then CFKs within the crater. 3. The transition from root to diatreme Recognition of the transition from HK, through TFK to DFK (illustrated in Fig. 5) is a relatively new development (e.g. Skinner and Marsh, 2004; Hetman et al., 2004). TFKs are real (see Section 3.2), they are not a form of pyroclastic kimberlite that can be ascribed to as “agglutinates, welded tuffs, vapour
phase zones in ignimbrites, rheomorphic tuffs or clastogenic lavas” (as suggested by Sparks et al., 2006). Although diatremefacies, hard, relatively fresh, agglutinated or welded tuffaceous kimberlites similar to tuffisitic kimberlites do exist (e.g the LM1 at Letlhakane DK1 pipe, Botswana and the Fragmental kimberlite at Kao pipe, Lesotho). Now that TFKs are recognized, more and more examples are being found in similar Class 1 settings all over the world (e.g. Masun and Scott Smith, 2006). Their recognition and the correct interpretation of what they represent is critical in understanding the true processes responsible for the emplacement of South African type (Class 1) kimberlites. Having shown this, it is remarkable that in the past, no one appears to have alluded to the extreme petrographic changes that have to occur in the conversion from hypabyssal to diatreme-facies kimberlites, regardless of the TFKs, which also occur. 3.1. Root zone hypabyssal kimberlites HKs show that, relative to other igneous rocks, kimberlites are unusual in that they contain high proportions of xenocrysts (typically 25% by volume) mostly of olivine. In addition to xenocrystic olivine, primary phenocrystic olivine starts to crystallize at a relatively early stage so that, even prior to rapid emplacement into the Earth's crust, the kimberlite carries around 25 vol.% olivine phenocrysts, which together with xenocrysts make up 50 vol.% of the magma. This relatively high solid content creates an effective viscosity for the kimberlite magma that is relatively high (Costa, 2005). Olivine continues to crystallize from the magma down to temperatures below 1000 °C when, in mainly Class 1, Group I kimberlites, monticellite crystallizes rapidly en masse as micro-phenocrysts generally b0.05 mm in size (Fig. 6a). In Class 1, Group II
Fig. 5. Venetia K1 pipe. Significant mineralogical, textural and chemical changes evident from detailed petrographic observation, accompany the transition from hypabyssal-(a), through transitional-(b) to diatreme-facies (c) kimberlites.
E.M.W. Skinner / Journal of Volcanology and Geothermal Research 174 (2008) 40–48
45
Fig. 6. Hypabyssal kimberlites from (a) Group I, Dutoitspan Mine and (b) Group II, Star Mine. Both have bases of serpentine and lesser calcite.
kimberlites mainly phlogopite crystallizes (Fig. 6b). This crystallization renders these magmas immobile and further emplacement to the surface as a magma is inhibited. It is unlikely that these kimberlites would see the surface, if they were not subjected to later diatreme-forming processes. In the case of Class 2 kimberlites the magmas are carbonate rich, the groundmass does not crystallize en masse, they are relatively less viscous, they can remain as hot magmas all the way to the surface and apparently do not undergo substantial exsolution of juvenile volatiles and the formation of transitional and diatremefacies kimberlites. In some kimberlites textural evidence suggests that melilite may crystallize after olivine but before monticellite (Skinner et al., 1999) but other minerals that crystallize after monticellite include
phlogopite, spinels, perovskite, apatite (at ∼750 °C), calcite (at ∼650 °C) and serpentine (at ∼350 °C). All the earlier minerals crystallize as euhedral crystals within the serpentine, which has all the appearances of having formed from, devitrified, late-crystallizing gel. In many HKs, secondary serpentine forms as a result of the well-known serpentinization of olivine and monticellite but this serpentine is different from the serpentine or rather serpophite (Michell and Putnis, 1988) occurring in the base of the kimberlite. In most HKs, the earlier crystallizing, euhedral primary minerals have an intimate textural relationship with the later crystallizing, interstitial serpophite, which in some cases completely encloses the earlier minerals. These minerals could not have been left suspended in a void, which was then later filled by secondary serpentine, unrelated to the kimberlite event. This relationship
Fig. 7. (a) Finsch F2. Pelletal/globular segregationary, macrocrystic kimberlite located within the core zone of an otherwise uniform, macrocrystic kimberlite. In most cases, altered olivine macrocrysts and smaller phenocrysts, surrounded by dark selvages of mainly very fine-grained phlogopite and lesser microlitic diopside, are set in an inter-pelletal matrix of serpentine and calcite (from Clement, 1982). (b) Weasua K1, Liberia. Parts of pelletal segregations of mainly altered phlogopite and diopside needles which protrude outwards into a matrix of serpentine (S) and calcite (C).
46
E.M.W. Skinner / Journal of Volcanology and Geothermal Research 174 (2008) 40–48
leaves little room for doubt that this serpentine is primary. Note that the compositions of typical HKs have relatively high H2O+ and CO2 contents (e.g. average 7.7 and 6.15 wt.% respectively, Le Roex et al., 2003). Much of the water is contained in the serpentine. 3.2. Transitional-facies kimberlites Skinner and Marsh (2004) and Hetman et al. (2004) describe the petrographic features of TFKs. The essential features of these rocks include the appearance of diopside, the disappearance of calcite and the development of spherical, segregationary structures in rocks that occur in some cases in the contact zones between hypabyssal- and diatreme-facies kimberlites. In these cases, the physical contacts between HK and TFK and between TFK and DFK are vague and transitional; the different kimberlite types grade or transcend into one another and no obvious fixed boundaries exist. However, the transitional-facies kimberlites in such circumstances occur in settings below or adjacent to the DFK. In most cases these transitional-facies kimberlites exhibit little evidence of fragmentation or transportation. In other cases, TFKs occur as distinct dykes or dyke-like bodies or separate vertical plugs of kimberlite not associated with DFK. In the latter case, these TFKs could be referred to as arrested DFKs, where the full transition from HK to DFK did not occur, or is hidden, or was affectively aborted. Good examples of this sort of kimberlite occur at Finsch (F2 kimberlite) at Kao (Gritty kimberlite) and at Premier (Dark Piebald kimberlite). The TFKs range in petrographic character from magmatic-like TFKs or HKts (Hetman et al., 2004) to diatreme-like TFKs or TKts (Hetman et al., 2004). In extreme variants of magmatic types, isolated patches of microlitic diopside are evident (Fig. 5b) or globular segregationary textures are evident (Fig. 7a and b) or
both (Fig. 7a). In addition calcite may be present in the base of the rock and olivine macrocrysts may be relatively fresh. As one moves away from these magmatic types towards diatreme types, microlitic diopside becomes more abundant and relatively finergrained; segregationary textures or pelletal structures become more evident; calcite disappears and olivine macrocrysts become altered. In most cases, no new country rock xenoliths (CRX) are added to the system and those that are present are highly altered. However, Hetman et al., 2004 indicate that in the case of the Gahcho Kue kimberlites, some new CRX may be added to the transitional kimberlites. In the HKts these CRX are highly altered but in the TKts they are much less altered. In some cases there is evidence of limited flow and deformation in the TFKs. The segregationary/pelletal structures (called many things by different authors including; magmaclasts, pelletal or globular lapilli, juvenile lapilli or pyroclasts) contain all the mineralogical and textural characteristics of the HKs. If not too altered and partly replaced by ultra-fine diopside, they are seen to consist most often of one or more olivine grains surrounded by finer grained primary minerals similar to those presented in Fig. 5a and b. In most cases these minerals have the identical compositions and the same range in grain sizes of the same minerals present in typical HKs, indicating that these structures have crystallized under similar conditions as normal HKs. The inter-pelletal matrix consists mostly of serpentine. In rare cases calcite may also be present (e.g. Fig. 7a and b) but as one moves away from magmatic types calcite soon disappears. The variation in textures and mineralogy in TFKs suggest that they represent normal, Class 1 type HKs that have been texturally and mineralogically modified, in situ. The composition of the pelletal structures indicates that these rocks are in the process of rapid, en masse crystallization of the groundmass. Contemporaneously with this crystallization, the kimberlite
Fig. 8. (a) Koffiefontein, tuffisitic kimberlite consisting of altered olivine grains, country rock xenoliths (CRX) and pelletal magmaclasts set in a matrix of serpentine and clay minerals. The magmaclasts consist of smaller olivine grains plus brown ultra fine-grained material. Note that in most cases the CRX occur outside of the magmaclasts. (b) Letseng Satellite, tuffisitic kimberlite containing relatively distinct magmaclasts, where some centres are relatively well-preserved (note fine, granular monticellite and evenly distributed fine opaques) and the margins are coated by ultra fine, dark, microlitic diopside.
E.M.W. Skinner / Journal of Volcanology and Geothermal Research 174 (2008) 40–48
magma segregates into pelletal structures and an inter-pelletal matrix of serpentine and calcite. The latter minerals are considered to be the product of crystallization of a volatilerich gas phase that separated from the earlier crystallizing minerals in an aerosol-like process thought to be driven by second boiling equivalent to vesiculation in other volcanic rocks. Note that a fixed boundary between the pelletal structures and the matrix is not evident as delicate, needle-like diopside crystals project outwards from the structures into the matrix (Fig. 7b). This texture indicates at least two things: (i) that the formation of the segregations is unlikely to be due to liquid immiscibility and (ii) that the crystallization of diopside occurs at the same time as the formation of the pelletal structures and not long afterwards through secondary alteration. In most cases TFKs are initiated at depths of between 3 and 1 km under relatively quiescent conditions in magmatic settings. They occur in a range of situations: within dykes independent of pipes (Camp Alpha dyke in Liberia), in dykes or dyke-like bodies within pipes (e.g. F2 and F4 at Finsch), within discrete but irregular intrusive plugs within pipes (e.g. Dark Piebald kimberlite at Premier and the Gritty kimberlite at Kao) and at the interface between HK and DFK in pipes (e.g. in several Kimberley pipes, at Venetia, at Premier and at Gahcho Kue). 3.3. Diatreme-facies kimberlites DFKs are generally massive, unsorted, matrix-to clastsupported, fragmental rocks consisting essentially of abundant serpentinized olivine macrocrysts and phenocrysts (in an approximate 50:50 ratio), abundant magmaclasts (for terminology refer to Field and Scott Smith, 1998) and variable proportions of country rock xenoliths and xenocrysts set in an inter-fragmental matrix dominated by primary serpentine (e.g. Fig. 8a). In many cases (N 50%) serpentinized olivine and serpentine in the matrix is replaced by a combination of mostly clay minerals plus pectolite and secondary garnet and titanate may also be present. As in the case in TFKs, diopside is abundant but in the case of DFKs it is in most cases (N 95%) typically ultra fine-grained and is not easily identifiable. But, in some cases this diopside can be microscopically resolved at high power as tiny laths with reduced birefringence and with inclined extinction. Microlitic diopside in DFKs occurs in several different settings: (a) as the replacement product of some finer-grained, primary matrix minerals within variably sized magmaclasts, (b) as thin, but recognizable, coatings on all larger sized components, including individual olivine grains, country rock xenoliths and xenocrysts and kimberlite magmaclasts (e.g. Fig. 8b), and (c) as delicate, individual crystals or clots of crystals within the primary serpentine matrix. In many cases (N80%) the original, small, primary groundmass minerals in the magmaclasts are extensively replaced by the ultra fine-grained diopside, making it difficult to identify, however, in some kimberlites (e.g. in one of the Letseng Satellite DFKs) these primary minerals are relatively well-preserved. Note in most TFKs these primary groundmass minerals are not as extensively replaced as in DFKs and the thin ultra-fine diopside coatings
47
around larger sized components are not as prominent. These coatings are thought to crystallize mainly out of the gas cloud formed by extensive exsolution of the remaining magma column (Skinner and Marsh, 2004) at the time of the explosion. This process may also be responsible for the crystallization of the primary serpentine in the inter-fragmental matrix. Likely crystallization activity and reactions that happen at this time occur in a CO2 depleted environment. The loss of CO2 during the development of TFKs prior to diatreme formation suggests that CO2 is either absent or has a restricted involvement in the diatreme-forming event. On the other hand, the presence of abundant serpentine in DFKs indicates high water activity. The fine diopside coatings on all larger sized components appear as a primary feature and the intimate relationship between delicate, haphazardly orientated individual needles of diopside or clots of diopside crystals present in the base of serpentine also represent primary features that are unlikely to be produced by late stage, secondary processes, unrelated to the primary diatreme-forming event. The diatreme-forming process occurs at a relatively low temperature as indicated by the inclusion of new, country rock xenoliths that are relatively fresh and unaffected by high temperatures. This may be supported in some DFKs by the presence of high grade coal fragments with ignition temperatures of possibly b 350 °C. These new xenoliths are excluded from kimberlite magmaclasts demonstrating that the explosion postdates the segregation of the magmaclasts. 4. Conclusions In Class 1 kimberlite pipes, at levels of between 3 and 1 km from surface: (1) En masse crystallization of the residual kimberlite liquid occurs. (2) This leads to the exsolution of juvenile volatiles. (3) At points where this occurs the hypabyssal-facies kimberlite is transformed into transitionalfacies kimberlites when pelletal structures begin to form and some exsolved gases crystallize as calcite and serpentine. Soon hereafter replacement of the primary groundmass minerals in the pelletal structures occurs. (4) Volatile exsolution is accompanied by volume increase and this leads to an increase in PΔV and subvolcanic breccias are formed in the surrounding side-walls by hydraulic fracturing. (5) Further exsolution at higher levels and lower pressures leads to cracking through to surface. (6) Traumatic decompression from about 330 bars to 1 bar leads to extensive exsolution of the entire magma column down to 2.2 km, further increase in PΔV and a large single explosion is generated. The point of this explosion is generated within the transitional-facies domain. (7) The consequence of the explosion is the production of a shock wave that rebounds from the surface. (8) A cone of brecciation having a 82° slope shaped upwards from the explosion cavity occurs. (9) Considerable cap rock material spalls off into the atmosphere but some large blocks sink downwards. (10) Smaller, new country rock xenoliths plus olivine macrocrysts and kimberlite magmaclasts are fluidized, overturned and thoroughly mixed by spouting. (11) Microlitic diopside and serpentine crystallize out of the gas cloud at relatively low temperatures (b 550 °C) and
48
E.M.W. Skinner / Journal of Volcanology and Geothermal Research 174 (2008) 40–48
the tuffisitic kimberlites are formed. (12) A tuff ring is deposited around the edges of a 500–700 m deep crater. (13) Pyroclastic kimberlites are emplaced into the lower part of the crater. (14) After some time the side walls of the crater collapse under the weight of the tuff ring, forming flared walls in the upper part of the crater and debris flow coarse breccias and high angle grain flow resedimented volcaniclastic kimberlites are deposited. (15) Continuing erosion of the remainder of the tuff ring and the surrounding weathered country rocks adds to the infill of the crater with coarse-grained, proximal-facies grits and finergrained, distal-facies sandstones. (16) The processes outlined above may be repeated but the number of repetitions are limited to only a few (b 3) events. (17) Final infill takes the form of epiclastic, crater lake deposits. (18) Over time considerable weathering in the pipe occurs, particularly of the friable, fragmental rocks. Acknowledgements The reviewers Casey Hetman and James Head and editor Barbara Scott Smith are thanked for their invaluable advice and considerable assistance in the case of the latter. Colleagues in the Geology Department, Rhodes University are thanked for their participation and particularly Steffen Buttner for his help in computing. References Bartlett, P.J., 1998. Premier Mine. Large mines field excursion guide. 7th International Kimberlite Conference, Cape Town, pp. 39–49. Burnham, C.W., 1985. Energy release in subvolcanic environments: implications for breccia formation. Econ. Geol. 80, 1515–1522. Clement, C.R., 1982. A comparative geological study of some major kimberlite pipes in the Northern Cape and Orange Free State. Ph.D. thesis University of Cape Town. Clement, C.R., Reid, A.M., 1989. The origin of kimberlite pipes: an interpretation based on a synthesis of geological features displayed by southern African occurrences. In: Ross, et al. (Eds.), Kimberlites and Related Rocks. Geol. Soc., vol. 14. Australia, pp. 632–646. Costa, A., 2005. Viscosity of high crystal content melts: dependence on solid fraction. Geophys. Res. Lett. 32, L22308. Field, M., Scott Smith, B.H., 1998. Textural and genetic classification schemes for kimberlites: a new perspective. Extended Abstracts, 7th International Kimberlite Conference, Cape Town, pp. 214–216.
Gernon, T.M., Gilbertson, M.A., Sparks, R.S.J., Walters, A., Field, M., 2006. Gas-solid fluidization in an experimental tapered bed: insights into processes in diverging volcanic conduits. Long Abstracts. Kimberlite Emplacement Workshop. Canada, Saskatoon (unpaginated). Hanson, E.K., Moore, J.M., Robey, J., Bordy, E.M., Marsh, J.S., 2006. Reestimation of erosional levels in Group I and II kimberlites between Lesotho, Kimberley and Victoria West, South Africa. Long Abstracts. Kimberlite Emplacement Workshop. Canada, Saskatoon (unpaginated). Hawthorne, J.B., 1975. Model of a kimberlite pipe. In: Ahrens, et al. (Eds.), Physics and Chemistry of the Earth, vol. 9, pp. 1–16. Hetman, C.M., Scott Smith, B.H., Paul, J.L., Winter, F., 2004. Geology of the Gahcho Kue kimberlite pipes, NWT, Canada: root to diatreme magmatic transition zone. Lithos 76, 51–74. Kurszlaukis, S., Lorenz, V., 2006. An alternative explanation for South African style “tuffisitic kimberlites”. Long Abstracts. Kimberlite Emplacement Workshop. Canada, Saskatoon (unpaginated). Le Roex, A.P., Bell, D.R., Davis, P., 2003. Petrogenesis of Group I kimberlites from Kimberley, South Africa: evidence from bulk-rock geochemistry. J. Pet. 44 (12), 2261–2286. Lorenz, V., 1975. Formation of phreatomagmatic maar-diatreme volcanoes and its relevance to kimberlite diatremes. Phys. Chem. Earth 9, 17–29. Masun, K., Scott Smith, B.H., 2006. The Pimenta Bueno kimberlite field, Rondonia, Brazil: evidence for tuffisitic kimberlite. Long Abstracts, Kimberlite Emplacement Workshop. Canada, Saskatoon (unpaginated). McCallum, M.E., 1985. Experimental evidence for fluidization processes in breccia pipe formation. Econ. Geol. 80, 1523–1543. Michell, R.H., Putnis, A., 1988. Polygonal serpentine in segregation-textured kimberlite. Can. Miner 26, 991–997. Mitchell, R.H., 2006. Petrology of hypabyssal kimberlites. Long Abstracts. Kimberlite Emplacement Workshop. Canada, Saskatoon (unpaginated). Rice, A., 1999. Can the blasting excavation engineering sciences provide insight into the processes of kimberlite emplacement and eruption? In: Gurney, et al. (Eds.), Proceedings of the 7th International Kimberlite Conference, vol. 2, pp. 699–708. Skinner, E.M.W., Mahotkin, I.L., Grutter, H.S., 1999. Melilite in kimberlites. Proceedings of the 7th International Kimberlite Conference, vol. 1, pp. 788–794. Skinner, E.M.W., Marsh, J.S., 2004. Distinct kimberlite pipe classes with contrasting eruption processes. In: Mitchell, et al. (Eds.), 8th International Kimberlite Conference. Sel. Papers, vol. 1, pp. 183–200. Sparks, R.S.J., Baker, L., Brown, R., Field, M., Schumacher, J., Stripp, G., Walters, A., 2006. Dynamical constraints on kimberlite volcanism. J. Volcanol. Geotherm. Res. 155, 18–48. Wyllie, P.J., 1987. Transfer of subcratonic carbon into kimberlites and rare earth carbonatites. In: Mysen, B.O. (Ed.), Magmatic processes: Physiochem. Principles. Spec. Publ. Geochem. Soc., vol. 1, pp. 107–119.