| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Journal of Petrology | Volume 43 | Number 9 | Pages 1749-1777 | 2002
© Oxford University Press 2002
Metasomatism and Partial Melting in Upper-Mantle Peridotite Xenoliths from the Lashaine Volcano, Northern Tanzania
DEPARTMENT OF GEOLOGY AND GEOPHYSICS, UNIVERSITY OF EDINBURGH, EDINBURGH EH9 3JW, UK
Received April 10, 2001; Revised typescript accepted March 22, 2002
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONPETROGRAPHYMINERAL CHEMISTRY OF THE...MINERAL CHEMISTRY OF THE...DISCUSSIONREFERENCES |
|---|
A group of chrome-spinel peridotite upper-mantle xenoliths from the Lashaine volcano, northern Tanzania, differs from other xenoliths at this locality in containing glassy melt pockets. Modal, mineral chemical and isotopic evidence indicates that, before the melting that was coincident with the xenolith entrainment and eruption in the Pleistocene, the sub-Tanzanian mantle lithosphere had a complex history. A major element depletion at
3·4 Ga gave rise to a high-olivine restite protolith, and this was followed by an episode of K, Fe, Ca, Ti and Rb metasomatism at 2·0 Ga (Metasomatic Event I) resulting in the formation of well-equilibrated Cr-diopside and phlogopite. A later episode of metasomatism (Metasomatic Event II), first reported here, is recognized by texturally and chemically non-equilibrated titanian phlogopite, Ti–Cr-diopside, enstatite–bronzite, ilmenite and rutile, interpreted as due to an influx of K, Fe, Ca, Ti, Nb and Ta. The recognition of two episodes of metasomatism reported here provides an explanation for previously recognized major differences in the isotopic chemistry of Lashaine peridotite diopsides. Immediately before entrainment and eruption in the Pleistocene, some of the peridotites underwent partial melting with the formation of melt pockets and veins containing vesiculated glass, olivine, diopside, phlogopite, spinel (sensu lato), calcite, apatite and zeolite. Compared with most previously reported mantle glasses, the Lashaine glasses are potassic, rather than sodic, contain relatively low amounts of SiO2, Al2O3 and total alkalis, and are not in equilibrium with peridotite olivine. The melting event was accompanied by an influx of K, Ti, Ca, Rb, Sr, Ba, Zr, the rare earth elements (REE), H2O and (inferred) CO2, which may be the culmination of Metasomatic Event II. Hence the melting is interpreted as being metasomatically triggered combined with decompression melting. The computed composition of one of the larger melt pockets is most similar to olivine lamproite. Ion microprobe analyses of glass, diopside and mica in two of the melt pockets show preferential partitioning of Rb, Zr, Nb, Ba, Pb and the light REE (LREE) into the glass with respect to clinopyroxene, but the reverse for Sr and Y, in both mica-bearing and mica-free parageneses. In a mica-bearing melt pocket, Rb and Ba partition preferentially into mica relative to both clinopyroxene and glass. Similar patterns of partitioning exist in a similar xenolith from Labait (another Tanzanian xenolith locality) except that Sr is highly concentrated in the glass; and the glass and mica contain high Ba. Comparison of the chemistry of the melt pockets in the xenoliths from Lashaine and other northern Tanzanian volcanoes (Labait and Olmani) indicates that the melting in each xenolith suite was accompanied by a distinctive metasomatic influx. KEY WORDS: Tanzania; mantle xenoliths; metasomatism; melting; element partitioning
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONPETROGRAPHYMINERAL CHEMISTRY OF THE...MINERAL CHEMISTRY OF THE...DISCUSSIONREFERENCES |
|---|
Although small-volume melts generated at the onset of mantle melting rarely reach the surface, incipient melting can be investigated by the study of glassy melt pockets in upper-mantle peridotites that, in effect, are small pools of frozen embryonic magma. The purpose of this paper is to document the phase chemistry of partially melted upper-mantle peridotite xenoliths from the Lashaine volcano, northern Tanzania (3°22'S, 36°26'E). Lashaine and the other xenolith localities mentioned in this paper (Fig. 1) are minor tuff cones that form part of the so-called Younger Extrusives—an episode of Pleistocene volatile-rich, highly explosive alkaline volcanic activity in northern Tanzania (Dawson, 1992).
|
Since their discovery in 1961 (Dawson, 1964), the lower-crustal and mantle xenoliths embedded in the ankaramitic and carbonatitic tuffs of the Lashaine cratered tuff-cone have provided valuable insights into the makeup of the continental lithosphere beneath northern Tanzania. An early xenolith collection has been extensively studied by Dawson and others (Dawson & Powell, 1969; Dawson et al., 1970; Hutchison & Dawson, 1970; Reid & Dawson, 1972; Dawson & Smith, 1973; Reid et al., 1975; Rhodes & Dawson, 1975; Ridley & Dawson, 1975; Cohen et al., 1984; Dawson, 1987; Burton et al., 2000; Schiano et al., 2000), and other collections have been studied by Pike et al. (1980), Nielson (1989), Henjes-Kunst & Altherr (1992) and Rudnick et al. (1994). These studies have provided abundant petrographic, mineral chemical and whole-rock geochemical data, showing that (1) the upper-mantle rocks, which comprise garnet-bearing dunites, garnet-free dunites, harzburgites, lherzolites, wehrlites and pyroxenites, are unserpentinized (unlike similar xenoliths from kimberlites) and hence retain valuable intra-grain and grain-boundary textural information; (2) a depletion event, resulting in high Mg/Fe and low Ca/Al ratios was overprinted by later metasomatism, which may correlate with wall-rock–intrusive vein relationships seen in some specimens; (3) the suite equilibrated along a 44 mW/m2 geothermal gradient. In summary, the present-day textures, mineralogy and chemistry of the suite testify to a complex history. A particular paradox was the recognition of a bimodal distribution in the Sr, Nd and Pb isotopic compositions of clinopyroxenes from Lashaine lherzolites, possibly signalling two mantle events widely separated in time (Cohen et al., 1984).
A further complexity in some Lashaine xenoliths is explored here. First reported by Reid et al. (1975), a number of the xenoliths exhibit textures indicative of small amounts of partial melting in the form of small frozen melt pockets and inter-grain veinlets that contain glass and precipitated phases. This paper provides phase data on both the peridotite protolith and melt-pocket phases in seven of these partly melted xenoliths, and discusses the process(es) that may have triggered the partial melting. In an assessment of possible regional differences, the Lashaine xenoliths are compared with partly melted peridotite xenoliths occurring at two other northern Tanzanian xenolith localities: Olmani (3°24'S, 36°45'E), 40 km east of Lashaine, and Labait (4°35'S, 35°26'E), some 150 km to the SW (Fig. 1).
|
PETROGRAPHY |
|---|
|
TOPABSTRACTINTRODUCTIONPETROGRAPHYMINERAL CHEMISTRY OF THE...MINERAL CHEMISTRY OF THE...DISCUSSIONREFERENCES |
|---|
The xenoliths are rounded to discoidal in shape, and have maximum dimensions up to 25 cm. The range of mineralogy is given in Tables 1 and 2. In the following discussion ‘primary’ refers to the well-equilibrated, unzoned olivine, pyroxenes and spinel (sensu lato) that form the bulk of the rocks, whereas ‘metasomatic’ refers to phases replacing or rimming the primary phases; and which may be zoned or of variable composition. As will be discussed below, the primary phases represent a combination of restite phases—olivine, chromite and possibly enstatite, together with phlogopite and Cr-diopside that result from a later metasomatic overprint; as a result of annealing the only extant evidence for the metasomatism is the isotope chemistry of the equilibrated phlogopite and diopside.
|
|
The modes were made by computer scanning of mineral distribution in traced overlays of A4-size photographs of thin sections, using ‘Scion Image’ image analysis software (this may be downloaded free of charge from http://www.scioncorp.com). The modes show the dominance of olivine but the volumes of minor phases and melt pockets vary widely even in thin sections prepared from the same slab; for example, the modes of two sections from the same slab cut from sample 747 vary from harzburgite containing 7·79% melt to mica-bearing harzburgite containing only 1·1% melt (Table 2 and Fig. 2). Specific difficulties in measuring the volumes of melt pockets or veins with light optics arise from: (1) optically continuous overgrowth of melt-pocket phases on primary phases at the margins of the melt pockets and on relict primary phases in the melt pockets; (2) glass devitrification.
|
The modes in Table 2, combined with the bulk chemical data of Rhodes & Dawson (1975), show that the peridotites studied here are more refractory than the Lashaine garnet peridotites. Modally they contain considerably less enstatite (all <15%, most <8%) than the garnet lherzolites, which typically contain >10% and often >20% modal orthopyroxene. Thus they are not different mineralogical expressions of the same compositions in response to equilibration at different levels in the upper mantle. Nor are they chemically equivalent to oceanic Al-spinel peridotites because, as shown below and earlier by Reid et al. (1975) and Rudnick et al. (1994), the spinels are chromites that contain substantially higher Cr and lower Al.
The mineral combinations in Table 1 classify most of the rocks as harzburgites; only one contains more than the 5% modal clinopyroxene required to be classified as lherzolite, according to the IUGS recommendation. Moreover, in this sample 750 (the author’s collection BD prefixes are dropped, for brevity), originally a wehrlite, it is the occurrence of metasomatic bronzite that permits it to be classified as lherzolite.
Peridotite petrography
The rocks are dominated by olivine and, although petrographically indistinguishable, on the basis of their phase chemistry there are two chemically different groups of peridotite. One group, comprising five of the seven xenoliths studied, contains Mg-rich phases that are compositionally similar to those in most upper-mantle peridotites, whereas in the second group the phases are more iron rich.
Primary phases
Primary olivine grains are up to 8 mm, have straight mutual grain boundaries and are frequently strained and exhibit kink banding. The olivine margins often have euhedral projections into melt pockets (Fig. 3a–d), and relict, partly resorbed olivine may occur as ‘islands’ in the melt pockets. The primary othopyroxene is also in 8 mm grains but does not show the same degree of deformation as the olivines. Primary Cr-diopside is generally of smaller grain size (up to 5 mm) than the olivine and enstatite; it is green in colour but adjacent to replacing metasomatic orthopyroxene, and where partly melted, it is often colourless (‘bleached’), reflecting a change in chemistry (described below). Primary light brown mica is up to 5 mm, and shows undulose extinction; it generally has darker brown metasomatic rims. The spinels are rounded, up to 2 mm and optically homogeneous, although electron probe micro-analysis (EPMA) has shown small core–rim compositional differences in some grains.
|
|
|
|
|
|
Metasomatic and unequilibrated phases
Metasomatic orthopyroxene, forming replacement coronas round primary Cr-diopside in samples 750 and 1544, contrasts with the primary enstatite by being in smaller (2 mm) light brown, turbid grains that form aggregates with small (<0·5 mm) rounded grains of chromite and dark brown mica.
Metasomatic diopside rimming and replacing chromite in 747 is the same green colour as primary Cr-diopside, whereas that replacing orthopyroxene in 1542 is colourless. Metasomatic mica forming rims of 1–2 mm thickness on chromite grains is generally unzoned.
Two specimens contain ilmenite. Replacement textures are absent but, in view of the fact that such a high-Fe–Ti mineral would not have survived the melting event inferred from the otherwise highly refractory mineralogy of the samples, they are interpreted as being of metasomatic origin. The ilmenite grains are rounded, up to 2 mm and mainly optically homogeneous; ilmenite in 1544 contains exsolved lamellae of Mg–Ti chromite (Fig. 3e), and the ilmenite itself is partly replaced by rutile. Also in this sample, rounded rutile grains are included in olivine and enstatite, and some larger (150 µm) grains have exsolved ilmenite and fine lamellae of a NbTaCr titanate. Ilmenite also occurs in enstatite–phlogopite–ilmenite coronas replacing primary Cr-diopside in another Lashaine specimen (Dawson, 1987).
Sample 1542 is unlike the others in containing rounded clusters up to 3 mm of dominant 100 µm high-Al enstatite grains and rarer 30 µm high-Al diopsides that are symplectitically intergrown with euhedral to vermiform grains up to 80 µm of high-Al spinel; all three phases are compositionally distinct from the primary pyroxenes and Cr-spinel, and, as individual grains can vary widely in composition, they are regarded as unequilibrated. By analogy with other Lashaine xenoliths containing pyrope garnet in varying stages of replacement by similar pyroxene–spinel intergrowths (Reid & Dawson, 1972), the symplectites in sample 1542 are interpreted as completely replaced garnets.
Petrography of the melt pockets and veins
The largest melt pocket found (in 3926) measures 7 mm x 3·5 mm; some are rounded but others are irregular with apophyses pinching out into intergranular veinlets. Most veinlets are of the order of 0·1–0·5 mm wide; those originating in melt pockets rarely persist for more than 5 mm but, exceptionally, persist for >25 mm (the width of a thin section). In detail, the margins of melt pockets are highly irregular (Fig. 3c), mainly as a result of crystals protruding into the melt pockets; these may be either the irregular ends of corroded primary phases or the euhedral faces of new phases precipitating on the melt-pocket margins (Fig. 3d).
All the primary phases have been corroded at the melt-pocket margins and occur as rounded or irregularly shaped relicts in the melt pockets; however, compared with other primary phases, enstatite is relatively uncorroded. In the melt pockets the olivine occurs as epitaxial overgrowths on wall-rock olivine grains, or on rounded ‘islands’ of corroded primary olivine. Being more magnesian than the primary olivine (see Mineral Chemistry, below), the overgrowths are slightly darker in BSE images (Fig. 3a and b) and, in addition, there are often chains of spinel at the primary grain–overgrowth interface (Fig. 3a). Most microphenocrysts in the melt pockets have cores of relict primary olivine, but more rarely have glass inclusions elongate parallel to their long axes. Clinopyroxene is the most abundant precipitated phase and occurs as euhedral microphenocrysts up to 300 µm (Fig. 3f), occasionally adhering to the melt-pocket walls, and sometimes overgrowing wall-rock or corroded relict grains of Cr-diopside or enstatite (Fig. 3g). The mica occurs as euhedral 300 µm randomly oriented platelets. The spinels generally are euhedral and up to 100 µm, with the larger ones often having cores of relict primary Cr-spinel (Fig. 3h). Some spinels are zoned (Fig. 3i) and may be overgrown by discontinuous chains (‘atolls’) of a compositionally different spinel, which may also occur within glass or overgrown by diopside (Fig. 3g). In sample 1544, some melt pockets contain euhedral 30 µm grains of rutile (Fig. 3j). The glass itself is invariably vesicular (e.g. Fig. 3g) and sometimes partly devitrified. In most melt pockets there are no visible variations in the glass, but in one melt pocket in sample 1544 two different glasses can be seen in back-scattered electron images (Fig. 3k). The vugs are usually lined with zeolite (Fig. 3g); in many instances, the vugs are surrounded by fresh glass, which, combined with an absence of alteration in the other protolith and melt-pocket phases, strongly suggests that the zeolites are a late-stage magmatic phase, rather than a precipitate from percolating groundwater. In a few cases, apatite and calcite occur within the fringing zeolite and, in one sample (3926), tiny 20 µm grains of witherite (BaCO3) are present.
|
MINERAL CHEMISTRY OF THE PERIDOTITES |
|---|
|
TOPABSTRACTINTRODUCTIONPETROGRAPHYMINERAL CHEMISTRY OF THE...MINERAL CHEMISTRY OF THE...DISCUSSIONREFERENCES |
|---|
Analytical methods
The minerals were analysed for major and minor elements by wavelength-dispersal analytical techniques on a Cameca Camebax electron microprobe at the University of Edinburgh. Details of operating conditions and standards have been given by Dawson & Hill (1998). Most phases were analysed with a spot beam of
1 µm at 20 kV and 20 nA, but under these conditions glasses and zeolites were unstable and so were analysed at 10 nA with the beam rastered over a 12·5 µm square, and Na and K were analysed early in the routine to reduce alkali loss or migration. Compared with published analyses of other mantle glasses, the alkali contents of the Lashaine glasses were found to be low, so specific experiments have been carried out to check for alkali loss. Spots in glasses in two specimens, 1546 and 3926, were analysed over 25 3-s cycles at 20 nA (i.e. at more extreme conditions for potential alkali loss than in the routine analyses); over the analysis period, decrease in the number of counts for both Na and K was statistically insignificant. As a further check, analyses were made on natural glasses under similar conditions. Alkali loss from basaltic tephra was again insignificant, whereas with hydrated rhyolite tephra there was a 40% and 75% reduction in the number of counts for K and Na, respectively, over the analysis period. Thus, the low alkali contents in the Lashaine glasses are considered to be real, and not the result of loss or migration caused by electron beam excitation.
Analyses for Rb, Sr, Y, Zr, Nb, Ba, Pb and rare earth elements (REE) in clinopyroxenes, mica and glasses were made in situ on polished, gold-coated thin sections by secondary ion mass spectrometry using a Cameca imf-4f ion microprobe in the Department of Geology and Geophysics at the University of Edinburgh. Measurements were made with an 8 nA 16O primary beam with a net impact of 15 keV, and focused to a spot of 15 µm diameter. Molecular ion interferences were discriminated using an energy filtering technique (Zinner & Crozaz, 1986). Intensities of all masses were measured over 10 cycles for each analysis point, with an aquisition time of 10 s per cycle. The ion intensities were normalized to Si. Corrections were made for overlaps of rare earth oxides (MO+); the BaO overlap on Eu was calculated assuming that excess counts at mass 154 were 138Ba + 16O.
Peridotite phase chemistry
Olivine
Olivines in the more abundant magnesian peridotites (Table 3) are highly magnesian (Fo89·2–92·8) and contain minor MnO (0·12–0·17 wt %) and NiO (0·12–0·42 wt %). Other detectable elements (Ti, Al, Cr and Ca) are generally in concentrations of <0·04 wt %. Olivines in the more Fe-rich peridotites are Fo85–86 with slightly higher MnO (0·18–0·19 wt %); the other oxides are in similar concentrations.
Orthopyroxene
Primary orthopyroxenes in the more magnesian peridotites (Table 4, analyses 1, 2, 3, 10 and 12) are enstatites (mg 91·3–94·4) with <3 wt % diopside. Al2O3 is generally <1 wt %, though up to 1·42 wt % in the former garnet harzburgite (sample 1542). Most enstatites contain CaO in the range 0·4–0·6 wt %; the enstatite in harzburgite 747 contains only 0·38 wt % CaO, which is similar to that found in enstatites from refractory harzburgites in kimberlites (Hervig et al., 1980). Relict areas of enstatite being replaced by metasomatic diopside in 1542 (Table 4, analyses 4 and 5) contain higher Ti, Cr, Al and Ca, but are more magnesian (mg 95 vs 94) than the primary enstatite. In the more Fe-rich peridotite 1546, the primary orthopyroxene (Table 4, analysis 15) is bronzite (mg 83) that, compared with the other primary orthopyroxenes, contains lower Mg and Al but higher Fe, Ca and Mn.
|
Metasomatic enstatites replacing Cr-diopside in 1544 and 3926 (Table 4, analyses 11 and 13) fall within the compositional range of the primary enstatites, except for higher CaO. In the more Fe-rich specimen 750, metasomatic bronzite replacing primary Cr-diopside (analysis 14), is less magnesian than the texturally similar metasomatic enstatites in the more magnesian peridotites (mg 86·5 vs 92).
Enstatite in the pyroxene–spinel symplectites in 1542 (interpreted as retrograded garnet) (Table 4, analyses 6–9) as well as being a high-Al variety is also of variable composition (see analyses 6 and 7); spinel-free parts of the grains contain less Cr, Al, Mn and Ca and are more magnesian than enstatite in the spinel-rich areas of the same grains; hence these pyroxenes are interpreted as unequilibrated.
An overgrowth on the bronzite in sample 1542, adjacent to a melt vein, zones towards higher Mg and Al (analyses 16 and 17); this more magnesian composition is analogous to the more magnesian melt-pocket overgrowths on olivines.
Clinopyroxene
Primary clinopyroxenes (Table 5, analyses 4, 5, 7 and 9) are Cr-diopsides with Cr2O3 in the range of 1·73–2·84 wt %, Al2O3 1·7–2·4 wt % and Na2O 1·2–2·1 wt %; their compositions fall mainly within the ranges found in earlier studies on Lashaine peridotites (Reid et al., 1975; Dawson, 1987); although the Cr-diopsides in the more Fe-rich peridotites (Table 5, analyses 10–17) are slightly higher in iron (total FeO mainly >4 wt %) compared with those in the other peridotites (mainly <3·3 wt %). Ca/(Ca + Mg) ratios are 0·46–0·48, except for 0·42 in the diopside in (garnet) harzburgite 1542 which implies a higher equilibration temperature. Also in 1542, non-equilibrated high-Al diopside in the spinel–pyroxene symplectites (Table 5, analysis 6) contains less Cr and Na than the primary Cr-diopsides. Cr-diopsides being replaced by enstatite in 1544 and 1546 (Table 5, analyses 8 and 15) contain considerably less Al, Cr and Na, but more Mg, Fe and Ti than the primary diopside; similar decreases in the jadeite–ureyite molecule have been noted previously in partly melted Cr-diopsides in kimberlite xenoliths (Carswell, 1975) and in ‘bleached’ Cr-diopside being replaced by amphibole in veined peridotites from Pello Hill, another northern Tanzanian xenolith locality (Dawson & Smith, 1988). Metasomatic diopside replacing Cr-spinel in 747 (Table 5, analyses 1–3) differs compositionally from the primary Cr-diopsides in containing significant TiO2 (>1 wt %), and lower Na2O (0·8 wt % vs often >2 wt % in primary Cr-diopsides); this diopside is zoned with respect to Cr, which has its highest concentration immediately adjacent to the replaced spinel.
|
Mica
Primary mica has been found only in sample 1544. It is a titaniferous phlogopite (mg 91, TiO2 2·18 wt %) with 1·19 wt % Cr2O3 (Table 6, analysis 2) and its composition falls within the ranges found in other primary micas in mantle xenoliths (Delaney et al., 1980).
|
Metasomatic mica in 1544 forms overgrowths on primary mica of a highly titaniferous phlogopite (7·14 wt % TiO2, Table 6, analysis 3) that is less magnesian (mg 89) than the primary mica. In its high TiO2 content, this overgrowth mica resembles metasomatic high-Ti (TiO2 6·71 wt %) phlogopite replacing Cr-spinel in sample 1544 (Table 6, analysis 4), and metasomatic mica replacing Cr-diopside in 747; it is also higher in Ti than other primary mantle micas (Delaney et al., 1980). These overgrowths or replacements of high-Ti mica on the primary phases testify to the migration of Ti-rich fluids through specimens 747 and 1544.
Spinel
Most primary spinels (Table 7) are Mg-chromites, Cr2O3 ranging from 50 to 60 wt % and MgO from 10 to 15 wt %. These compositions can be matched in chromites from other Lashaine samples (Reid et al., 1975), and from garnet lherzolite xenoliths from South African kimberlites (Smith & Dawson, 1975), and the most chromian (62 wt % Cr2O3; analysis 3) resembles inclusions in diamond (Meyer, 1987). Two different spinels are: (1) Mg–Al-rich spinel (analysis 7) in spinel–orthopyroxene symplectites in 1542, which, in its high MgO and Al2O3 contents, is compositionally very similar to spinel in garnet-breakdown coronas in Lashaine garnet lherzolite xenoliths (Reid & Dawson, 1972); (2) a high-iron (25 wt % FeO), high-Ti (10·4 wt % TiO2), low-Al (2·23 wt % Al2O3) chromite (analysis 11) that has exsolved from ilmenite in 1544.
Ilmenite
Ilmenite, interpreted as being of metasomatic origin, occurs as rounded grains in 1544 (Table 8). Its high MgO content (14·1 wt %) is similar to other mantle ilmenites but its very high Cr2O3 content (7·33 wt %) distinguishes it from ilmenite megacrysts from kimberlite (Mitchell, 1986) and ilmenites in veined and metasomatized peridotite and MARID xenoliths (e.g. Harte & Gurney, 1975; Dawson & Smith, 1977; Jones et al., 1983b; Grégoire et al., 2000). Moreover, before the exsolution of Cr-spinel, this ilmenite must have contained even higher Cr concentrations.
A large rounded grain of ilmenite in a melt patch in Fe-rich peridotite 1546 (see Table 9, analyses 18 and 19) is around the same size (500 µm) and has the same high-Cr character as the ilmenite in 1544; hence it is interpreted as a relict metasomatic grain, rather than a precipitate. Like the other primary phases in the rock it is less magnesian than its counterpart in 1544.
|
Rutile
Rutile has been found only in specimen 1544. Different rutile grains have different compositions (Table 8, analyses 6–9), indicating that they are unequilibrated. Overall the rutile is a Cr–Fe–Nb–Ta variety with 3·95–5·24 wt % Cr2O3, and 0·29–0·81 wt % FeO. It has appreciable concentrations of Nb2O5 (0·88–5·13 wt %) and Ta2O5 (0·04–1·05 wt %), with Nb/Ta ratios vaying from 7·5 to 8. Their high Cr content resembles that in the very few other analysed rutiles from peridotite and MARID xenoliths (Smith & Dawson, 1975; Dawson & Smith, 1977), and distinguishes them from eclogite rutiles in which Cr2O3 is usually <1%. Of the upper-mantle rutiles listed by Haggerty (1983, 1987), ones from the Jagersfontein kimberlite are the closest match for the Lashaine rutiles.
|
MINERAL CHEMISTRY OF THE MELT POCKETS |
|---|
|
TOPABSTRACTINTRODUCTIONPETROGRAPHYMINERAL CHEMISTRY OF THE...MINERAL CHEMISTRY OF THE...DISCUSSIONREFERENCES |
|---|
Olivine
Olivines in the melt pockets occur as discrete grains, as overgrowths on wall-rock grains and on partly resorbed relict grains in the melt pockets. Compared with the primary olivine in the host peridotite, most melt-pocket (MP) olivine is more magnesian, e.g. Fo93·4 vs Fo92·1 (Fig. 4; Table 3, analyses 3 and 1); a very few are more iron rich than the primary olivine (some only by <1% Fo) and others zone towards more iron-rich rims (Fig. 4). MP olivines consistently contain higher CaO (frequently x4) and Cr2O3 (x2 to x3), but lower NiO than primary olivines; relative MnO concentrations are erratic.
|
Clinopyroxenes
On the basis of their Ca–Mg–Fe contents, most MP clinopyroxenes (Table 10) are diopsides; some have enough alumina to be classified as augites, and others (e.g. analyses 10, 19 and 22) are sufficiently magnesian to be endiopsides. Overall, TiO2 and Cr2O3 concentrations are <1·5 wt %, Al2O3 <3 wt % and Na2O <1 wt %. However, within these generalizations, there are many variations in the minor element concentrations between pyroxenes from different specimens, between pyroxenes from different melt pockets within the same sample, and between different pyroxenes (some of which may be zoned) within individual melt pockets. For example, pyroxenes from three different melt pockets in 771 (analyses 5 and 6) show only small variations in Ti and Na, and only moderate variation in Cr and Al (e.g. 0·27–0·79 wt % Al2O3), whereas the core and rim of a single grain in a melt pocket in 1544 (analyses 11 and 12) show wide variations in TiO2 (0·89–1·64 wt %), Al2O3 (2·84–0·55 wt %) and Cr2O3 (0·14–0·40 wt %). Again, pyroxenes from certain specimens or melt pockets can have distinctive chemistry; for example, those overgrowing relict enstatite in 1542 are relatively aluminous (Al2O3 >2·4 wt %), whereas sectors in a sector-zoned grain in 1546 (analyses 19 and 20) have higher TiO2 (1·11–1·55 wt %) than other MP pyroxenes. One of the most magnesian clinopyroxenes (19·5 wt % MgO; analysis 22) occurs in a melt pocket within a partly melted bronzite grain.
The MP pyroxenes are generally distinguishable from the peridotite Cr-diopsides by their lower Cr2O3 and Na2O concentrations (Fig. 5). However, they can be chemically similar to some ‘bleached’ rims on some primary Cr-diopsides (e.g. Table 5, analyses 8 and 12) that have lost Cr and Na during melting.
|
Micas
The MP micas are Ti-phogopites (Table 6). There are only minor variations in the major element concentrations of SiO2 (38–39·6 wt %), Al2O3 (12·1–12·8 wt %) and K2O (9·05–9·83 wt %) but there are, within limits, ranges in the concentrations of the other oxides. As in the MP clinopyroxenes, these variations occur between micas from different specimens, in micas from different melt pockets within the same sample, and in different micas within an individual melt pocket. For example, micas in 747 (Table 6, analyses 5–11) contain relatively low TiO2 (3–3·67 wt %) and are more magnesian (mg 91–92) compared with those in 1544 (TiO2 4·95–8·53 wt %; mg 85–90). In the MP micas in the more Fe-rich peridotites (analyses 12–19), TiO2 variation is small (4·15–4·86 wt %), but mg varies from 85 to 89, with the higher values overlapping with those for micas in the other peridotites. Where analysed, BaO concentrations are low (mainly >0·2 wt %) compared with up to 1·38 wt % in MP micas in peridotites from Labait (Dawson, 1999). For most micas, the lower Cr2O3 concentrations (0·3–0·8 wt %) distinguish them from the peridotite primary micas (>1·2 wt %). However, it is worth noting that in sample 1544 the darker brown metasomatic rim on primary mica (Table 6, analysis 3) is chemically very similar to one of the MP micas (analysis 10).
Spinels
The MP spinels (Table 9) can be grouped into two main sets. The first set comprises Mg-chromites, with Cr2O3 >50 wt % and MgO 12–15 %; it includes samples (e.g. analyses 2 and 14) in which the high Cr2O3 concentrations (>63 wt %) are similar to those in spinels from highly refractory peridotites and spinel inclusions in diamond. A zoned grain in sample 1546 (analyses 15 and 16) has rims relatively enriched in Ti, Al, Fe and Mg and, on the basis of the stoichiometric calculation of Fe3+, the iron is more highly oxidized. This Mg-chromite set, although chemically similar to many of the Cr-spinels in the host peridotites, is distinguished on textural grounds; that is, the MP spinels are considerably smaller and are euhedral. (Relict peridotite Cr-spinel grains undoubtedly occur within some melt pockets but they are relatively large and are rounded.) A subset of the Mg chromites occurs in some melt pockets in 1544, in which some chromites (analyses 8 and 10) contain unusually high amount of TiO2 (7–9 wt %), a feature in keeping with the overall high-Ti mineralogy of this rock.
The second set of spinels, found only in the Mg-peridotites and represented by analyses 3–7, is more aluminous, with cr (Cr/[Cr + Al]) ranging from 0·20 to 0·45, compared with cr >0·74 in the Mg-chromites. In addition, they are more magnesian, with MgO in the range 17–22 wt %. One grain in 1542 zones to a rim richer in Al and Mg, but poorer in Ti, Cr and Fe, than the core (analyses 3–5).
Rutile
Rutile occurs in the melt pockets of specimen 1544, the only specimen to contain peridotite metasomatic rutile. The MP rutile (Table 9, analyses 16 and 17) contains less Cr than the metasomatic rutiles. It has Nb concentrations (5 wt % Nb2O5) similar to the metasomatic grains (see Table 9, analyses 5 and 6), but has exceptionally high Ta2O5 (>1 wt %) resulting in a low Nb/Ta ratio of 5, compared with 7·5–11 for the peridotite rutiles. The MP rutiles themselves vary little in composition except for Nb2O5 (range 4·55–5·49 wt %) and Ta2O5 (range 1·31–1·82 wt %).
Glass
Analyses presented in Table 11 show that, in general, the glasses are low in SiO2 (<50 wt %), Al2O3 (<8 wt %), CaO (mainly <2 wt %) and Na2O (<0·20 wt %). K2O is always in greater concentrations than Na2O, reflected in K/(K + Na) ratios generally >0·9. Although Al concentrations are low, the glasses are not peralkaline, (K + Na)/Al ratios being <0·7. These low Si, Al, Ca and alkali concentrations reflect the refractory, olivine-dominated protolith, but the glasses are high in Fe (analysis 7 being exceptionally so) mostly in the range 10–18 wt % total FeO; this combined with relatively low MgO concentrations (mostly <10 wt %) gives mg mainly <40. Importantly, the low MgO (4–14 wt %, mostly <10 wt %) concentrations show that these glasses were not in equilibrium with upper-mantle olivines of Fo90 composition (Roeder & Emslie, 1970).
Whereas the above general observations pertain to most of the glasses, there are exceptions. First, glass in 1542 (analysis 3), associated with the high-Al pyroxene and spinel formed by breakdown of former garnet, contains more Al2O3 (16 wt %) and CaO (5·74 wt %) but lower FeO (5·11 wt %) and MgO (4·15 wt %) than the other glasses; in its high alumina content, this glass is similar to glass reported in a reaction corona in another Lashaine garnet lherzolite (Schiano et al., 2000) (Table 11, analysis 4). Second, in specimen 1544 one of the glasses is a very high-Fe variety (FeO 54·9 wt %), combined with very low SiO2 (22·3 wt %), Al2O3 (2·12 wt %) and MgO (2·97 wt %) and relatively high TiO2 (2·76 wt %).
In most samples, glass compositions in different melt pockets have roughly the same characteristics (compare, for example, the pairs of analyses 1 and 2, and 11 and 12, Table 11). Exceptionally, in 1544, glass compositions in two melt pockets (analyses 4 and 5) are reasonably similar with respect to Al, Fe, Ca, Mg and K, but these contrast with two glasses (analyses 6 and 7) in another melt pocket, which are also very different from each other—the one being a low-Al, high-Mg variant, the other the exceptionally high-Fe glass noted above.
All the glasses have low analytical totals and, with the possible exception of the high-Fe glass, the totals would remain low even if total iron was calculated as Fe2O3. The vesicularity of the glasses testifies to the volatile-rich nature of the melts, so unanalysed (or fugitive) dissolved volatiles probably account for the shortfalls. CO2 is the most likely, as ion microprobe analyses of glasses in two specimens (not reported in detail here) give low H2O concentrations ranging from 1·54 to 3·88 wt % in glass in specimen 1546, and from 0·42 to 3·85 wt % in 3926; even when these values are integrated, there is still a shortfall in the analyses of these glasses.
Schiano & Clochiatti (1994) reported that glasses in mantle xenoliths from a large number of worldwide localities have certain common characteristics, with high SiO2 (55–65 wt %), Al2O3 (15–25 wt %) and total alkalis (4–11 wt %) relative to their peridotite protoliths; in addition, most glasses are sodic with K2O/Na2O rarely approaching unity. Subsequent studies on mantle glasses have revealed variations but have, in general, confirmed the findings of Schiano & Clochiatti (1994). The Lashaine glasses differ from those reported by Schiano & Clochiatti (1994), in containing relatively low concentrations of SiO2, Al2O3 and total alkalis, and also being potassic rather than sodic (Fig. 6a and b).
|
The Lashaine glasses also show certain differences from those in other Tanzanian xenoliths. Glasses in xenoliths from Labait show considerable compositional variation between and within different individual melt pockets, and many contain significant amounts of Ba (Dawson, 1999). Glasses in Olmani xenoliths are similarly very variable in composition (Jones et al., 1983a), and comparison with the Lashaine glasses shows that most are relatively high in Al2O3 (12–20 wt %) and P2O5 (>1 wt %), and some are highly sodic. However, like the Lashaine glasses, many of the Labait and Olmani glasses have analytical total shortfalls, attributed to unanalysed volatiles.
Zeolites
Like the glasses, the zeolites within the vesicles are unstable under the electron beam. Analyses (Table 12) show most to be a Ca–K phase with Si:Al 2:1. There are variations between samples and between zeolites in different vugs in single samples. In particular, the zeolites in the more iron-rich peridotites contain a small amount of BaO (up to 2·18 wt % in sample 750), which is apparently absent in zeolites in the magnesian peridotites.
|
Trace element concentrations in peridotite diopside and melt-pocket diopside, mica and glass
Ion microprobe analyses (Table 13) have been made on phases in two Lashaine specimens (1546 and 3926) and, for comparison, in peridotite 4213 from Labait, major element data for which have been given by Dawson (1999). Analyses were made on glass and MP clinopyroxene in all specimens, MP micas in two specimen (one from Lashaine and one from Labait) and primary Cr-diopside in 1546. For the MP micas, in which the REE concentrations were <1 ppm, only the light REE (LREE) were analysed. An earlier electron microprobe analysis of primary mica in 4213 showed it to contain 0·15 wt % BaO (1340 ppm Ba) (Dawson, 1999).
|
Comparing primary Cr-diopside with MP diopside in 1546, the former contains only slightly higher concentrations of Sr, Y, Zr, Nb, Pb and the REE (Fig. 7), and the REE pattern (Fig. 8) is almost identical. Both clinopyroxenes contain much lower concentrations of most of these elements than the MP glass, which also contains relatively high amounts of Rb, Ba and the LREE (La to Nd). However, the glass contains less Sr, Y and the heavy REE (HREE; Dy to Lu) than the pyroxenes.
|
|
MP diopside in 3926 is similar to that in 1546, with low concentrations of most incompatible elements and differences between them only of the order of about x2. In contrast to 1546, 3926 also contains MP mica, which contains high Rb (340 ppm) and Ba (1490 ppm) relative to both glass and cpx, and high Nb and Pb relative to diopside. Diopside has the highest concentrations of Sr and Y and glass has the highest concentrations of Zr, Nb and Pb.
In Labait specimen 4213, element partitioning between MP phases is in most respects similar to that in 3926; except that Sr is high in glass relative to pyroxene, and Zr, Nb and Pb partitioning between glass and pyroxene is not as high as in 3926; moreover, the Ba content of the MP mica (2970 ppm) is double that in the primary mica from this specimen (1340 ppm; Dawson, 1999). Overall the MP phases in the Labait specimen contain higher concentrations of Rb, Sr and Ba, but lower Zr, Nb and Pb, than those in the Lashaine melt pockets.
In summary, in all specimens Zr, Nb and Pb are preferentially concentrated in the glass; in which Ba is also concentrated in mica-absent 1546. In mica-bearing MPs, micas concentrate Rb and Ba relative to glass and pyroxene, and contain more Nb than coexisting pyroxenes. However, as noted for other hydrous mantle assemblages (McDonough & Frey, 1989), the mica contains only relatively small amounts of the REE. An REE plot (Fig. 8) shows that the clinopyroxenes, both primary and melt pocket, and from both Lashaine and Labait, have similar patterns with a gradual slope from the LREE to the HREE (La/YbCN 3–6) although slightly convex upwards from Ce to Sm. The glasses have higher concentrations of the LREE (up to x200 chondrite) and lower HREE, giving a steeper slope (La/YbCN 195–835). In addition, all three glass patterns tend to have a weak negative Eu anomaly. The LREE preferential partitioning into glass relative to clinopyroxene, combined with the steeper REE patterns for the glasses, results in a crossover in the middle REE (MREE) area (Fig. 8). An important point is that, on a volume-to-volume basis, the main REE-bearing phase in the peridotite (Cr-diopside) is not able to provide the LREE concentrations found in the melt pockets. The disparity would be even greater if the MP apatites (not analysed) contain REE.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONPETROGRAPHYMINERAL CHEMISTRY OF THE...MINERAL CHEMISTRY OF THE...DISCUSSIONREFERENCES |
|---|
Equilibration conditions
Five specimens contain both primary ortho- and clinopyroxene, and temperatures of equilibration calculated using the two-pyroxene thermometer of Brey et al. (1990) are given in Table 14. Three equilibrated in the range 950–990°C similar to other garnet-free Lashaine peridotites, whereas two (including 1542, in which symplectites are interpreted as former garnets) equilibrated at 1100–1150°C, within the range found for Lashaine garnet lherzolites (Rudnick et al., 1994). Assuming the 44 mW/m2 geotherm established for the Lashaine garnet lherzolites (Rudnick et al., 1994), depths of equilibration range from
115 to 140 km.
|
Ancient melting and Metasomatic Event I
An early model for the evolution of the northern Tanzania upper mantle was based on the whole-rock chemistry of Lashaine peridotites. It proposed a melting event during which the peridotites were variably depleted in Fe, Ca, Al and alkalis relative to estimates of the composition of primitive upper mantle; this refractory residue was subsequently overprinted by K, Ti, Fe, Ca and REE metasomatism (Rhodes & Dawson, 1975; Ridley & Dawson, 1975). The model did not attempt to give a time frame for these events, nor could it predict whether the metasomatic effects result from one or more events.
This early model is refined here in the light of more recent data on Lashaine peridotite xenoliths, including those from the present study. The high amount of modal olivine reported in Table 2 confirms the highly refractory nature of the peridotites, and the very low Ir/Pd ratios in some Lashaine peridotites reported by Rehkämper et al. (1997) are further indications of extensive melt loss. An Archaean age of not less than 3·4 Ga for this initial melting event has been provided by Os isotopic dating of a sulphide concentrate from a Lashaine garnet lherzolite (Burton et al., 2000). This date is comparable with the age of melt extraction and stabilization of the sub-continental lithospheric mantle beneath the Kaapvaal, Zimbabwe and Wyoming cratons (Carlson & Irving, 1994; Pearson et al., 1995; Nägler et al., 1997). By contrast, however, chromites from Labait xenoliths have yielded a relatively young Os age of 2·5–2·9 Ga (Chesley et al., 1999).
The earlier suggestion of a metasomatic overprint was based on whole-rock chemistry (high concentrations of incompatible trace elements combined with a refractory major-element chemistry) and the presence of phlogopite in Lashaine garnet peridotites (Dawson & Powell, 1969). Recent isotope work on Cr-diopside in a Lashaine phlogopite-bearing garnet peridotite has provided a Nd and Pb isotope model age of 2·0 Ga (Burton et al., 2000), confirming earlier isotopic work (including data on phlogopite) for an event at around this time (Cohen et al., 1984). This age also coincides with a 2·0 Ga model age for the mafic granulites that are part of the Lashaine xenolith suite (Cohen et al., 1984), which together suggest a major chemical event in the northern Tanzania mantle and deep crust at 2 Ga. This age is that of major subduction around the eastern edge of the Tanzania Craton, manifest further south in Tanzania by the Usagaran orogeny (Möller et al., 1995); the effects of this orogeny in northern Tanzania have been largely erased by the later Pan-African orogeny (Möller et al., 1998).
The status of enstatite in the primary assemblage is ambiguous. It might have been part of the restite assemblage after the 3·4 Ga melting event. An alternative is that it may have formed by silica metasomatism of the high-olivine restite during a subduction event, as argued by Rudnick et al. (1994) to explain the high amounts of enstatite in Lashaine garnet peridotites. There is neither textural nor isotopic evidence to resolve these alternatives.
Metasomatic Event II
The present study provides textural evidence for replacement of Cr-spinel by both phlogopite and Cr-diopside, replacement of primary enstatite by diopside, replacement of Cr-diopside by metasomatic Ti-bronzite–ilmenite–phlogopite aggregates, formation of high-Ti rims on primary phlogopite and, in 1544, formation of ilmenite and its subsequent replacement by compositionally variable Nb–Ta-rutile. These phenomena indicate a metasomatic event during which there was addition of Si, K, Ti, Ca and Fe and, in 1544, Nb and Ta. Moreover, variations in the chemistry of different grains of the same phase (e.g. rutile in 1544), together with chemical zoning (e.g. in diopside replacing spinel in 747) and overgrowth of compositionally different rims on primary phases (e.g. phlogopite in 1544) indicate that equilibrium has not been achieved. The diopside and Ti-phlogopite are similar to those formed during the earlier 2 Ga event, but this second set of unequilibrated phases could not have remained in the high thermal regime of the upper mantle for 2 Gyr without annealing. The inference is that the Lashaine peridotites contain a second set of metasomatic phases that still had not achieved equilibrium at the time of entrainment and eruption and which, by implication, must be younger than those formed at 2 Ga. This conclusion provides an explanation for the two isotopically and chronologically distinct types of diopside identified in Lashaine peridotites by Cohen et al. (1984).
Melting and its causes
Various causes have been proposed for the formation of glasses in mantle xenoliths [see Shaw (1999) and references therein]. These include: (1) contact anatexis around a mantle magma chamber; (2) breakdown of hydrous phases; (3) decompression melting; (4) partial melting by a subsequent heat flux of a part of the mantle whose solidus had been lowered by an earlier phase of metasomatism (Dawson, 1984; Hauri et al., 1993; Yaxley & Kamenetsky, 1999) and metasomatically triggered melting (Ionov et al., 1994; Dawson, 1999).
In the case of the Lashaine xenoliths, there is no evidence from the geothermometry studies (Rudnick et al., 1994) for a thermal perturbation that might reflect a mantle magma chamber, so possibility (1) appears unlikely. A recognizable primary hydrous phase (mica) is present in only one xenolith (Table 1); however, in other samples there is the possibility that it may have been destroyed in the melting process. From several lines of evidence—the fact that the glasses are not in equilibrium with upper-mantle olivine; incomplete melting of relict phases within melt pockets; the presence of glasses of different compositions in the same xenolith (implying lack of time to permit migration, mixing and homogenization)—it is apparent that the length of time between the formation of the melt pockets and their subsequent quenching must have been very short. The preservation of glass itself testifies to rapid transport to the surface, which could be consistent with decompression melting.
The protolith at Lashaine had undoubtedly undergone light-element metasomatism before the melting event, but the pertinent question is whether there is evidence for a metasomatic flux at the time of melting and, if so, is it possibly tied to the younger Metasomatic Event II? On the assumption that diopside and mica might be expected to be the main contributors to melts generated at the onset of melting of the peridotite protolith, it is possible to assess qualitatively whether these two phases alone are capable of satisfying the melt-pocket phase chemistry, or whether an additional influx is required. An important point is that although peridotite mica might provide a small amount of water, the abundance of vugs indicates that volatiles played an essential role in the formation of the melts.
In the two Lashaine specimens for which there are phase trace element data (Table 13 and Fig. 7), although the protolith Cr-diopside could theoretically provide the Sr, Zr and Y concentrations found in the MP phases, unrealistically large amounts would need to be melted to account for the Nb, Pb and REE concentrations, particularly in the glasses; further, in the apparent absence of mica (although the rapid modal variations referred to in the petrographic section must be borne in mind), there is no protolith source for Rb and Ba. Even where protolith mica is present, an additional source is probably required to account for the high concentrations of Rb and Ba in the MP phases, although the mica is a plausible source for Nb. In sample 1544, assuming rutile to be the main source of Nb and Ta, the relatively high Ta in the MP rutiles (Table 9, analyses 16 and 17) indicates preferential partitioning of Ta into the melt, but no need for an additional external source for the Ta. In short, during the partial melting in the Lashaine peridotites, it appears that there must have been an additional influx of Rb, Ba, Nb, Pb and REE. Taken in conjunction with the petrography, which shows a relatively high amount of MP mica, diopside, zeolite and vug (former volatiles) in the melt pockets (even in samples lacking primary phlogopite and Cr-diopside), it can be inferred that the trace element influxes accompanied an addition of K, Ti, Fe, Ca and, particularly, volatiles, both water and (inferred) carbon dioxide. This combination of elements, although not necessarily relative concentrations, is strongly reminiscent of those added to the peridotite protoliths during Metasomatic Event II, and it is impossible not to speculate that the elements and volatiles found in the MP phases, although deriving in part from melted protolith phases, have also been supplemented by the metasomatic fluxes percolating the protolith during Metasomatic Event II. In summary, decompression, volatile release from pre-existing mica, and a volatile-rich metasomatic influx may all have contributed to the melting.
In contrast, the Labait specimen trace element data suggest certain differences from Lashaine, assuming primary Cr-diopsides at Labait contain similar trace element concentrations to that analysed from Lashaine. Cr-diopside could provide sufficient Y, Zr and possibly but not enough Nb, Sr, Pb and LREE for the melt-pocket assemblage. Protolith mica, for which there are no analytical data for Rb, might have been a sufficient source for Rb but the Labait protolith mica (1340 ppm Ba; Dawson, 1999) could not provide the high Ba concentrations found in the Labait MP phases. Thus, at Labait a Nb, Sr, LREE, Ba and ?Rb influx is required; the presence of Sr-bearing apatite and the Ba-zeolite harmotome in most Labait melt pockets particularly emphasizes the necessity for high Sr and Ba influxes (Dawson, 1999).
Following the melt formation, precipitation of new phases of distinctive composition took place. The diopside contains less jadeite–ureyite molecule than the peridotite diopsides and the micas are in general more titaniferous than the primary micas. An unusual aspect is the crystallization of olivine that is higher in Mg than in the peridotite protolith, and its intimate association with precipitated spinel (see Fig. 3b). Similar Mg-rich olivines in melt pockets in peridotite xenoliths from Mongolia are attributed to incongruent melting of primary clinopyroxene and spinel to give residual chromite, phenocrysts of olivine and clinopyroxene of a new composition and melt; subsequent fractional crystallization of the melt was limited to new rims on the olivine and clinopyroxene phenocrysts (Ionov et al., 1994). This model could be applied in part to the Lashaine rocks, but the Mongolian case is unlike the Lashaine examples in which mica is also involved. The close association of the high-Mg olivine with inclusions of euhedral chromite is significant, and a possible explanation is that the highly magnesian olivine (mg 0·92–0·94) formed when the melt was briefly depleted in Fe as a result of preferential Fe partitioning into the earlier-precipitated chromite (mg 0·47–0·80).
Glass compositions—comparisons with magmas
Over the past decade there has been intense interest and debate on the occurrence and chemistry of the glassy components of upper-mantle xenoliths, together with the reasons for the melting. In particular, much attention has been given to the worldwide occurrence of glasses high in Na, Al and Si (e.g. Schiano & Clocchiatti, 1994; Draper & Green, 1997). However, the chemistry of mantle glasses is variable and, not surprisingly, opinions for the variation are diverse. Distinctions have been made between the formation of glasses in which the chemistry has been influenced by melted hydrous phases (amphibole or phlogopite), and those in anhydrous peridotites. In the context of metasomatic triggering, Coltorti et al. (2000) appealed to different metasomatic fluxes accompanying the melting to explain chemical variations in glasses in anhydrous xenoliths; glasses related to carbonatite metasomatism are stated to be characterized by high CaO and Na2O, and low SiO2, and have high Na2O/K2O >2, whereas glasses related to K-alkali silicate metasomatism are characterized by high SiO2 and K2O and Na2O/K2O of <1. Another area for debate is that, whereas most researchers have accepted the chemistry of the glasses to be real, Schiano & Bourdon (1999) concluded that only glass inclusions armoured within mantle minerals have not been strongly modified by a decrease in pressure; hence, they suggested that glasses found as intergranular films and pockets do not preserve the composition of the melts formed at depth. Finally, there are differences of opinion as to whether experimental glasses can be used to interpret the formation conditions of mantle glasses; the debate concerns whether the diamond-aggregate approach or the sandwich technique is the better experimental method for the achievement of equilibrium (Draper & Green, 1999).
There is, moreover, a preoccupation with the glass chemistry; and in some studies glass compositions are equated with magma types ranging from basanite to trachyte, the preoccupation extending to the glasses being plotted on TAS (total alkalis vs silica) diagrams of Le Maitre (see Draper & Green, 1997, fig. 1; Shaw, 1999, fig. 1). In some studies the compositions and amounts of the precipitated phases (if any) are largely ignored (e.g. Varela et al., 1999) even when these have been extensively analysed (e.g. Chazot et al., 1996). In these studies, the fundamental principle is ignored that, unless there is a total absence of phenocrysts or microlites, glass compositions do not equate with the original melts.
In the case of Lashaine, the original melts have been modified by precipitation of variable amounts of olivine, diopside, mica and spinel, so an attempt has been made to compute the chemistry of the melt in one of the larger Lashaine melt pockets by integrating the mode of the melt pocket and the chemistry of the phases (Table 15). In the computation zeolite has been included because, as has been argued above, it is a late-crystallizing fluid phase (rather than a deposit from groundwater). Broadly, the computed composition is silica undersaturated, low in alumina and has K > Na, parameters that, with differences in detail, would group it with low-volume potassic magma types such as kimberlite, katungite and olivine lamproite. Because of problems with measuring the mode, specific comparisons should be treated with caution, but the computed composition can be most closely matched with olivine–madupitic lamproite (Table 15, analysis 3). In the more limited northern Tanzania context, it has some affinities with katungite from Pello Hill, another xenolith locality (analysis 4). Importantly, its high K/Na ratio precludes any link with the sodic ankaramite that hosts the Lashaine peridotites (analysis 5).
|
Comparison with other Tanzanian xenolith suites
The data presented above indicate that, at the time of the glass pocket formation in the Lashaine peridotites, there was an influx of K, Ca, Rb, Sr, Ba, Zr, Nb, Pb, REE, CO2 and H2O. In the case of granulite xenoliths erupted at Lashaine at the same time as the peridotites, melting and K-metasomatism has also occurred (Jones et al., 1983c). By comparison, glass formation in the Olmani xenoliths was accompanied by addition of K, Ca, Na, Ti, P, REE, Cl, F and CO2 (Jones et al., 1983a; Rudnick et al., 1993, 1994); Ca, P, REE and CO2 addition was particularly high, resulting in the formation of new diopside and monazite and the Olmani glasses are more aluminous (mainly 11–14 wt %) than those in the Lahaine melt pockets, and most are sodic. The melting at Labait was characterized by addition of K, Ti, Ba, Sr, H2O, P and CO2 (Dawson, 1999), Ba being particularly enhanced and resulting in the formation of harmotome and high-Ba mica and glass. Taken together, the evidence from these three xenolith localities is that the detailed chemistry of the metasomatism accompanying the melting is variable, and that the metasomatism of each xenolith suite has a distinctive signature. The inference is that in relatively recent geological times, and coinciding with volatile-rich nephelinitic–carbonatitic regional magmatism, the sub-Tanzanian mantle lithosphere has been percolated by metasomatizing fluids that are heterogeneous over a small geographical area.
|
ACKNOWLEDGEMENTS |
|---|
I am grateful to Peter Hill and Paula McDade for help with the electron microprobe analyses and SEM images, and to John Craven for assistance with the ion-microprobe analyses. The analytical work was funded by NERC Grant GR9/02883. I also acknowledge perceptive reviews by Leonid Danyushevsky, Massimo Coltorti and a third anonymous referee, which have much improved the original manuscript. The samples were collected when I was a member of the Tanganyika Geological Survey, and during later field-work supported by the Carnegie Trust for the Universities of Scotland. The analytical facilities at Edinburgh are partly supported by NERC.
|
FOOTNOTES |
|---|
*Telephone: 0131 650 7286. Fax: 0131 668 3184. E-mail: jbdawson{at}glg.ed.ac.uk
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONPETROGRAPHYMINERAL CHEMISTRY OF THE...MINERAL CHEMISTRY OF THE...DISCUSSIONREFERENCES |
|---|
Brey, G. P., Köhler, T. & Nickel, K. G. (1990). Geothermometry in four-phase lherzolites. I. Experimental results from 10 to 60 kb. Journal of Petrology 31, 1313–1352.
Burton, K. W., Schiano, P., Birck, J.-L., Allègre, C. J., Rehkämper, M., Halliday, A. N. & Dawson, J. B. (2000). The distribution and behaviour of rhenium and osmium amongst mantle minerals and the age of the lithospheric mantle beneath Tanzania. Earth and Planetary Science Letters 183, 93–106.[Web of Science]
Carlson, R. W. & Irving, A. J. (1994). Depletion and enrichment history of subcontinental lithospheric mantle: an Os, Sr, Nd and Pb isotopic study of ultramafic xenoliths from the northwestern Wyoming craton. Earth and Planetary Science Letters 126, 457–472.[Web of Science]
Carswell, D. A. (1975). Primary and secondary phlogopites and clinopyroxenes in garnet lherzolite xenoliths. Physics and Chemistry of the Earth 9, 417–430.
Chazot, G., Menzies, M. & Harte, B. (1996). Silicate glasses in spinel lherzolites from Yemen: origin and chemical composition. Chemical Geology 134, 159–179.[Web of Science]
Chesley, J. T., Rudnick, R. L. & Lee, C.-T. (1999). Re–Os systematics of mantle xenoliths from the East African Rift: age, structure and history of the Tanzanian craton. Geochimica et Cosmochimica Acta 63, 1203–1217.[Web of Science]
Cohen, R. S., O’Nions, R. K. & Dawson, J. B. (1984). Isotope geochemistry of xenoliths from East Africa: implications for the development of mantle reservoirs and their interaction. Earth and Planetary Science Letters 68, 209–220.[Web of Science]
Coltorti, M., Beccaluva, L., Bonadiman, C., Salvini, L. & Siena, F. (2000). Glasses in mantle xenoliths as geochemical indicators of metasomatic agents. Earth and Planetary Science Letters 183, 303–320.[Web of Science]
Dawson, J. B. (1964). Carbonate tuff cones in northern Tanganyika. Geological Magazine 101, 129–137.[Abstract]
Dawson, J. B. (1984). Contrasting types of mantle metasomatism? In: Kornprobst, J. (ed.) Kimberlites II; the Mantle and Crust–Mantle Relationships. Amsterdam: Elsevier, pp. 289–294.
Dawson, J. B. (1987). Metasomatised harzburgites in kimberlite and alkaline magmas; enriched restites and ‘flushed’ lherzolites. In: Menzies, M. A. & Hawkesworth, C. J. (eds) Mantle Metasomatism. London: Academic Press, pp. 125–144.
Dawson, J. B. (1992). Neogene tectonics and volcanicity in the north Tanzania sector of the Gregory Rift Valley: contrasts with the Kenya sector. Tectonophysics 204, 81–92.[Web of Science]
Dawson, J. B. (1999). Metasomatism and melting in spinel peridotite xenoliths from Labait, Tanzania. In: Gurney, J. J., Gurney, J. L., Pascoe, M. D. & Richardson, S. H. (eds) Proceedings Volume of 7th International Kimberlite Conference, 1. Cape Town: Red Roof Design, pp. 164–173.
Dawson, J. B. & Hill, P. G. (1998). Mineral chemistry of a peralkaline combeite–lamprophyllite nephelinite from Oldoinyo Lengai, Tanzania. Mineralogical Magazine 62, 179–196.[Abstract]
Dawson, J. B. & Powell, D. G. (1969). Mica in the upper mantle. Contributions to Mineralogy and Petrology 22, 233–237.[Web of Science]
Dawson, J. B. & Smith, J. V. (1973). Alkalic pyroxenite xenoliths from the Lashaine volcano, northern Tanzania. Journal of Petrology 14, 113–131.
Dawson, J. B. & Smith, J. V. (1977). The MARID (mica–amphibole–rutile–ilmenite–diopside) suite of xenoliths in kimberlite. Geochimica et Cosmochimica Acta 41, 309–323.[Web of Science]
Dawson, J. B. & Smith, J. V. (1988). Metasomatised and veined upper mantle xenoliths from Pello Hill, Tanzania: evidence for anomalously-light mantle beneath the Tanzanian sector of the East African Rift Valley. Contributions to Mineralogy and Petrology 100, 510–527.[Web of Science]
Dawson, J. B., Powell, D. G. & Reid, A. M. (1970). Ultrabasic xenoliths and lava from the Lashaine volcano, northern Tanzania. Journal of Petrology 11, 519–548.
Delaney, J. S., Smith, J. V., Carswell, D. A. & Dawson, J. B. (1980). Chemistry of micas from kimberlites and xenoliths—II. Primary- and secondary-textured micas from peridotite xenoliths. Geochimica et Cosmochimica Acta 44, 857–872.[Web of Science]
Draper, D. S. & Green, T. H. (1997). P–T phase relations of silicic, alkaline, aluminous mantle-xenolith glasses under anhydrous and C–O–H fluid-saturated conditions. Journal of Petrology 38, 1187–1224.[Web of Science]
Draper, D. S. & Green, T. H. (1999). P–T relations of silicic, alkaline, aluminous liquids: new results and applications to mantle melting and metasomatism. Earth and Planetary Science Letters 170, 255–268.[Web of Science]
Grégoire, M., Lorand, J. P., O’Reilly, S. Y. & Cottin, J. Y. (2000). Armalocolite-bearing Ti-rich metasomatic assemblages in harzburgite xenoliths fron the Kerguelen Islands: implications for the oceanic budget of high-field strength elements. Geochimica et Cosmochimica Acta 64, 673–694.[Web of Science]
Haggerty, S. H. (1983). The mineral chemistry of new titanates from the Jagersfontein kimberlite: implications for metasomatism in the upper mantle. Geochimica et Cosmochimica Acta 47, 1833–1854.[Web of Science]
Haggerty, S. E. (1987). Metasomatic mineral titanates in upper mantle xenoliths. In: Nixon, P. H. (ed.) Mantle Xenoliths. Chichester: Wiley, pp. 671–690.
Harte, B. & Gurney, J. J. (1975). Ore mineral and phlogopite mineralisation within ultramafic nodules from the Matsoku kimberlite pipe, Lesotho. Carnegie Institution of Washington Yearbook 74, 528–536.
Hauri, E. H., Shimizu, N., Dieu, J. J. & Hart, S. R. (1993). Evidence for hotspot-related metasomatism in the oceanic upper mantle. Nature 365, 221–227.
Henjes-Kunst, F. & Altherr, R. (1992). Metamorphic petrology of xenoliths from Kenya and northern Tanzania and implications for geotherms and lithospheric structures. Journal of Petrology 33, 1125–1156.
Hervig, R. L., Smith, J. V., Steele, I. M. & Dawson, J. B. (1980). Fertile and barren Al–Cr-spinel harzburgites from the upper mantle: ion and electron probe analyses of trace elements in olivine and orthopyroxene: relationship to lherzolites. Earth and Planetary Science Letters 50, 41–58.[Web of Science]
Hutchison, R. & Dawson, J. B. (1970). Rb, Sr and 87Sr/86Sr in ultrabasic xenoliths and host rocks, Lashaine volcano, Tanzania. Earth and Planetary Science Letters 9, 87–92.[Web of Science]
Ionov, D. A., Hofmann, A. W. & Shimizu, N. (1994). Metasomatism-induced melting in mantle xenoliths from Mongolia. Journal of Petrology 35, 753–785.
Jones, A. P., Smith, J. V. & Dawson, J. B. (1983a). Glasses in mantle xenoliths from Olmani, Tanzania. Journal of Geology 91, 167–178.[Web of Science]
Jones, A. P., Smith, J. V. & Dawson, J. B. (1983b). Mantle metasomatism in 14 veined peridotites from Bultfontein Mine, South Africa. Journal of Geology 90, 435–453.
Jones, A. P., Smith, J. V., Dawson, J. B. & Hansen, E. C. (1983c). Metamorphism, partial melting and K-metasomatism of garnet–scapolite–kyanite granulites from Lashaine, Tanzania. Journal of Geology 91, 143–165.[Web of Science]
McDonough, W. F. & Frey, F. A. (1989). Rare earth elements in upper mantle rocks. In: Lipin, B. R. & McKay, G. A. (eds) Geochemistry and Mineralogy of Rare Earth Elements. Mineralogical Society of America, Reviews in Mineralogy 16, 100–;145.
McDonough, W. F. & Sun, S. S. (1995). The composition of the Earth. Chemical Geology 120, 223–253.[Web of Science]
Meyer, H. O. A. (1987). Inclusions in diamond. In: Nixon, P. H. (ed.) Mantle Xenoliths. Chichester: Wiley, pp. 501–522.
Mitchell, R. H. (1986). Kimberlites: Mineralogy, Geochemistry and Petrology. New York: Plenum.
Mitchell, R. H. & Bergman, S. C. (1991). Petrology of Lamproites. New York: Plenum.
Möller, A., Appel, P., Mezger, K. & Schenk, V. (1995). Evidence for a 2 Ga subduction zone: eclogites in the Usagaran belt of Tanzania. Geology 23, 1067–1070.
Möller, A., Mezger, K. & Schenk, V. (1998). Crustal age domains and the evolution of the continental crust in the Mozambique Belt of Tanzania: combined Sm–Nd, Rb–Sr, and Pb–Pb isotopic evidence. Journal of Petrology 39, 749–783.[Web of Science]
Nägler, Th. F., Kramers, J. D., Kamber, B. S., Frei, R. & Prendergast, M. D. A. (1997). Growth of the subcontinental lithospheric mantle beneath Zimbabwe started at or before 3·8 Ga: Re–Os study on chromites. Geology 25, 983–986.
Nielson, J. E. (1989). Nature and sequence of mantle magmatic events indicated by compositional variations in a composite xenoliths from Lashaine Volcano, Tanzania. In: Ross, J. (ed.) Kimberlite and Related Rocks. Geological Society of Australia, Special Volume 14, 700–708.
Pearson, D. G., Carlson, R. W., Shirey, S. B., Boyd, F. R. & Nixon, P. H. (1995). Stabilisation of the Archaean lithospheric mantle: a Re–Os isotope study of peridotite xenoliths from the Kaapvaal craton. Earth and Planetary Science Letters 134, 341–357.[Web of Science]
Pike, J. E. N., Meyer, C. E. & Wilshire, H. G. (1980). Petrography and chemical composition of a suite of ultramafic xenoliths from Lashaine, Tanzania. Journal of Geology 88, 343–352.[Web of Science]
Rehkämper, M., Halliday, A. N., Barfod, D., Fitton, J. G. & Dawson, J. B. (1997). Platinum group element abundance patterns in different mantle environments. Science 278, 1595–1598.
Reid, A. M. & Dawson, J. B. (1972). Olivine–garnet reaction in peridotites from Lashaine, Tanzania. Lithos 5, 115–124.
Reid, A. M., Donaldson, C. H., Brown, R. W., Ridley, W. I. & Dawson, J. B. (1975). Mineral chemistry of peridotite xenoliths from the Lashaine volcano, Tanzania. Physics and Chemistry of the Earth 9, 545–557.
Rhodes, J. M. & Dawson, J. B. (1975). Major and trace element chemistry of peridotite inclusions from the Lashaine volcano, Tanzania. Physics and Chemistry of the Earth 9, 545–557.
Ridley, W. I. & Dawson, J. B. (1975). Lithophile trace element data bearing on the origin of peridoite xenoliths, ankaramite and carbonatite from the Lashaine volcano, northern Tanzania. Physics and Chemistry of the Earth 9, 559–569.
Roeder, P. L. & Emslie, R. F. (1970). Olivine–liquid equilibrium. Contributions to Mineralogy and Petrology 29, 275–289.[Web of Science]
Rudnick, R. L., McDonough, W. F. & Chappell, B. W. (1993). Carbonatite metasomatism in the northern Tanzanian mantle: petrographic and chemical characteristics. Earth and Planetary Science Letters 114, 463–475.[Web of Science]
Rudnick, R. L., McDonough, W. F. & Orpin, A. (1994). Northern Tanzania peridotite xenoliths: a comparison with Kaapvaal xenoliths and inferences of metasomatic reactions. In: Meyer, H. O. A. & Leonardos, O. H. (eds) Kimberlites, Related Rocks and Mantle Xenoliths. Proceedings Volumes, 5th International Kimberlite Conference, Araxa, Brazil 1. Brasilia: Companhia de Pesquisa de Recursos Minerais, pp. 336–353.
Schiano, P. & Bourdon, B. (1999). On the preservation of mantle information in ultramafic nodules: glass inclusions within minerals versus interstitial glasses. Earth and Planetary Science Letters 169, 173–188.[Web of Science]
Schiano, P. & Clocchiatti, R. (1994). Worldwide occurrence of silica-rich melts in sub-continental and sub-oceanic mantle minerals. Nature 368, 621–624.
Schiano, P., Clocchiatti, R., Bourdon, B., Burton, K. W. & Thellier, B. (2000). The composition of melt inclusions in minerals at the garnet–spinel transition zone. Earth and Planetary Science Letters 174, 375–383.[Web of Science]
Shaw, C. S. J. (1999). Dissolution of orthopyroxene in basanitic magma between 0·4 and 2 GPa: further implications for the origin of Si-rich alkaline glass inclusions in mantle xenoliths. Contributions to Mineralogy and Petrology 135, 114–132.[Web of Science]
Smith, J. V. & Dawson, J. B. (1975). Chemistry of Ti-poor spinels, ilmenites and rutiles from peridotite and eclogite xenoliths. Physics and Chemistry of the Earth 9, 309–322.
Varela, M. E., Clocchiatti, R., Kurat, G. & Schiano, P. (1999). Silicic glasses in hydrous and anhydrous mantle xenoliths from Western Victoria, Australia: at least two different sources. Chemical Geology 153, 151–169.[Web of Science]
Yaxley, G. M. & Kamenetsky, V. (1999). In situ origin for glass in mantle xenoliths from southeastern Australia: insights into trace element compositions of glasses and metasomatic phases. Earth and Planetary Science Letters 172, 97–109.[Web of Science]
Zinner, E. & Crozaz, G. (1986). A method for the quantitative measurement of rare earth elements in the ion microprobe. International Journal of Mass Spectrometry and Ion Processes 69, 17–38.[Web of Science]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. B. Dawson Chapter 7 The Neogene-Recent volcanic rocks Geological Society, London, Memoirs, January 1, 2008; 33(1): 39 - 77. [Full Text] [PDF]
|
|
|
|
|
|
J. B. Dawson Chapter 8 Regional comparisons, petrochemistry and petrogenesis Geological Society, London, Memoirs, January 1, 2008; 33(1): 79 - 89. [Full Text] [PDF]
|
|
|
|
 |
|
 M. Ismail, G. Delpech, J.-Y. Cottin, M. Gregoire, B. N. Moine, and A. Bilal Petrological and geochemical constraints on the composition of the lithospheric mantle beneath the Syrian rift, northern part of the Arabian plate Geological Society, London, Special Publications, January 1, 2008; 293(1): 223 - 251. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Kaeser, A. Kalt, and T. Pettke Crystallization and Breakdown of Metasomatic Phases in Graphite-bearing Peridotite Xenoliths from Marsabit (Kenya) J. Petrology, September 1, 2007; 48(9): 1725 - 1760. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Ulianov, O. Muntener, P. Ulmer, and T. Pettke Entrained Macrocryst Minerals as a Key to the Source Region of Olivine Nephelinites: Humberg, Kaiserstuhl, Germany J. Petrology, June 1, 2007; 48(6): 1079 - 1118. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Ishimaru, S. Arai, Y. Ishida, M. Shirasaka, and V. M. Okrugin Melting and Multi-stage Metasomatism in the Mantle Wedge beneath a Frontal Arc Inferred from Highly Depleted Peridotite Xenoliths from the Avacha Volcano, Southern Kamchatka J. Petrology, February 1, 2007; 48(2): 395 - 433. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. KAESER, A. KALT, and T. PETTKE Evolution of the Lithospheric Mantle beneath the Marsabit Volcanic Field (Northern Kenya): Constraints from Textural, P-T and Geochemical Studies on Xenoliths J. Petrology, November 1, 2006; 47(11): 2149 - 2184. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. FERRARIS, B. GROBETY, G. L. FRUH-GREEN, and R. WESSICKEN Intergrowth of graphite within phlogopite from Finero ultramafic complex (Italian Western Alps): implications for mantle crystallization of primary-texture mica European Journal of Mineralogy, December 1, 2004; 16(6): 899 - 908. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. G. KOPYLOVA and G. CARO Mantle Xenoliths from the Southeastern Slave Craton: Evidence for Chemical Zonation in a Thick, Cold Lithosphere J. Petrology, May 1, 2004; 45(5): 1045 - 1067. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||