Crystal Fractionation Processes Research Articles

Crustal and mantle influences and U–Th–Ra disequilibrium in andesitic lavas of Ngauruhoe volcano, New Zealand

The andesitic volcano Ngauruhoe, which is located within the Tongariro Volcanic Complex at the southern end of the Taupo Volcanic Zone in North Island, New Zealand, has been constructed over the past 5 ka and last erupted in 1975. Nearby Ruapehu volcano has a much longer eruptive history extending back beyond 230 ka B.P. The magmas erupted at both volcanoes have been predominantly medium-K basaltic andesites and andesites, which evolved through polybaric crystal fractionation and assimilation processes that took place within complex, dispersed magmatic storage systems. Despite their close spatial proximity, the two volcanoes show geochemical contrasts suggesting that in each case both the mantle-derived parental magmas and the crustal assimilants were different. Variations in major and trace element data for Ngauruhoe lavas indicate control by crystal fractionation and assimilation (AFC) but the data are difficult to reconcile with derivation from a single batch of compositionally unique, mantle-derived parental basalt. Geochemical variation can be approximated by generalised AFC models using average basement or crustal xenolith compositions but precise mathematical modelling is limited because of the sensitivity of models to the selection of a particular parental composition and the difficulty of determining the exact nature of the crustal assimilant compositions that might have been involved in each specific case. U–Th isotopic data for Ngauruhoe volcanic rocks show disequilibrium that defines a positively inclined array lying to the right of the equiline on a 230Th/ 232Th versus 238U/ 232Th diagram. U-series data for Ngauruhoe and post 1945 AD Ruapehu eruptives show similar patterns of disequilibrium but the Ngauruhoe data define an array with a different slope and a different intercept on the equiline. 230Th/ 238U ratios in Ngauruhoe lavas range from close to secular equilibrium (0.979) to 0.864 with the youngest (post 1870 AD) eruptives showing the least disequilibrium. 230Th/ 238U and 87Sr/ 86Sr ratios are correlated; the samples with the highest 87Sr/ 86Sr values show the least disequilibrium. There is also a correlation between eruption age and Nd and Pb isotopic composition; the oldest samples tend to show the least radiogenic isotopic ratios. These data, in combination with major and trace element abundance data indicate that Ngauruhoe U–Th disequilibrium is determined by variations in parental magmatic compositions and AFC. Unlike the mature, long-lived Ruapehu system, Ngauruhoe's eruptive cycles can be directly connected to periodic magma recharge from the lower crust and mantle. For Ngauruhoe eruptives, 226Ra/ 230Th varies from 1.218 to 1.492. This compares with a range of 0.972 to 1.186 for Ruapehu lavas erupted between 1945 and 1996. Reconciling the 226Ra/ 230Th data with decay during fractional crystallisation requires relatively low rates of fractionation (1–5 × 10 − 5 /year) and short (1000–2000 years) fractionation times.

Crustal evolution in a cratonic nucleus: Granitoids and felsic volcanic rocks of the North Caribou Terrane, Superior Province Canada

Attempts to decipher the petrogenesis of Archean Trondhjemite–Tonalite–Granodiorite (TTG) and establish trends in their compositions over time typically do not assess the possible role of assimilation–crystal fractionation processes. Studies of TTG suites worldwide, however, demonstrate that reprocessing of older crust is common. Felsic volcanic and intrusive rocks from the North Caribou Lake area, located in the nucleus of the Superior Province, exhibit a range of compositions that are not accounted for by variable degrees of amphibolite melting alone. Characteristic of Archean TTG, the rocks mainly fall in the lower right hand quadrant of a Nb/Ta vs Zr/Sm plot but they also show clear trends between εNd T and ratios such as Al 2O 3/TiO 2, Th/Nb, [Tb/Yb] N and Eu/Eu*. The lowest Nb/Ta values in the TTG likely result from crystal fractionation including titanite, which strongly partitions and fractionates these elements.

Highly evolved hypabyssal kimberlite sills from Wemindji, Quebec, Canada: insights into the process of flow differentiation in kimberlite magmas

Kimberlite sills emplaced in granite located near the town of Wemindji (Quebec, Canada) range from 2 cm to 1.2 m in thickness. The sills exhibit a wide variation in macroscopic appearance from fine-grained aphanitic dolomitic hypabyssal kimberlite to ilmenite/garnet macrocrystal hypabyssal kimberlite. Diatreme or crater facies rocks are not present. Multiple intrusions are present within the sills, and graded bedding and erosional features such as cross-bedding are common. The sills exhibit a wide range in their modal mineralogy with respect to the abundances of spinel, apatite, phlogopite and dolomite. Olivine is the dominant macrocryst, with an average composition of Fo90. Garnet macrocrysts are low chrome (2–3 wt. %) pyrope (G1/G9 garnet). Ilmenite occurs as rounded macrocrysts (7–13 wt. % MgO). Phlogopite microphenocrysts are Ti-poor and represent a solid solution between phlogopite and kinoshitalite end members. Spinel compositions mainly represent the Cr-poor members of the qandilite–ulvospinel–magnetite series. The principle carbonate comprising the groundmass is dolomite, with lesser later-forming calcite. Accessory minerals include apatite, Sr-rich calcite, Nb-rich rutile, baddeleyite, monazite-(Ce) and barite. While some of these accessory minerals are atypical of kimberlites in general, it is expected that differentiation products of an evolved carbonate-rich kimberlite magma will crystallize these phases. The Wemindji kimberlites offer insight into the process of crystal fractionation and differentiation in evolved kimberlite magmas. The macroscopic textural features observed in the Wemindji sills are interpreted to represent flow differentiation of a mantle-derived, very fluid, low viscosity carbonate-rich kimberlite. The diverse modes and textural features result entirely from flow differentiation and multiple intrusions of different batches of genetically related kimberlite magma. The mineralogy of the Wemindji kimberlites has some similarities to that of the Wesselton and Benfontein calcite kimberlite sills but differs in detail with respect to dominant carbonate (i.e. dolomite versus calcite), and the character of the rare earth-bearing accessory minerals (i.e. monazite-(Ce) versus rare earth fluorocarbonates).

The generation of a diverse suite of Late Pleistocene and Holocene basalt through dacite lavas from the northern Cascade arc at Mount Baker, Washington

Mt. Baker is a dominantly andesitic stratovolcano in the northern Cascade arc. In this study, we show that the andesites are not all derived from similar sources, and that open-system processes were dominant during their petrogenesis. To this end, we discuss petrographic observations, mineral chemistry, and whole rock major and trace element chemistry for three of Mt. Baker’s late Pleistocene to Holocene lava flow units. These include the basalt and basaltic andesite of Sulphur Creek (SC) (51.4–55.8 wt% SiO2, Mg# 57–58), the Mg-rich andesite of Glacier Creek (GC) (58.3–58.7 wt% SiO2, Mg# 63–64), and the andesite and dacite of Boulder Glacier (BG) (60.2–64.2 wt% SiO2, Mg# 50–57). Phenocryst populations in all units display varying degrees of reaction and disequilibrium textures along with complicated zoning patterns indicative of open-system processes. All lavas are medium-K and calc-alkaline, but each unit displays distinctive trace element and REE characteristics that do not correlate with the average SiO2 content of the unit. The mafic lavas of SC have relatively elevated REE abundances with the lowest (La/Yb)N (~4.5). The intermediate GC andesites (Mg- and Ni-rich) have the lowest REE abundances and the highest (La/Yb)N (~6.7) with strongly depleted HREE. The more felsic BG lavas have intermediate REE abundances and (La/Yb)N (~6.4). The high-Mg character of the GC andesites can be explained by addition of 4% of a xenocrystic olivine component. However, their depleted REE patterns are similar to other high-Mg andesites reported from Mt. Baker and require a distinct mantle source. The two dominantly andesitic units (GC and BG) are different enough from each other that they could not have been derived from the same parent basalt. Nor could either of them have been derived from the SC basalt by crystal fractionation processes. Crystal fractionation also cannot explain the compositional diversity within each unit. Compositional diversity within the SC unit (basalt to basaltic andesite) can, however, be successfully modeled by mixing of basalt with compositions similar to the dacites in the BG unit. Given that the BG dacites erupted at ~80–90 ka, and a similar composition was mixed with the SC lavas at 9.8 ka, the process that produced this felsic end-member must have been repeatedly active for at least 70 ka.

Structure and geochemistry of an Alaskan-type ultramafic–mafic complex in the Eastern Pontides, NE Turkey

The Karayasmak ultramafic–mafic association (KUMA) in NE Turkey represents an Alaskan-type, pre-Liassic intrusion in the Pulur metamorphic massif and is one of many such intrusive complexes that were emplaced into the continental basement of the Eastern Pontides belt during the pre-Liassic. Its main lithological units include plagioclase peridotite, melagabbronorite, banded gabbro and anorthosite (Phase-I), gabbro-pegmatite dikes (Phase-II), unmetamorphosed banded gabbro (Phase-III), cumulate olivine gabbronorite (Phase-IV), and non-cumulate gabbroic dikes and stocks (Phase-V). The contact relations between these units are gradational or intrusive, with the peridotites (Phase-I) enveloping the cumulate gabbronorites (Phase-IV) in the center of the elliptical body, displaying a concentric reverse-zoning pattern. Cr–Al spinel, plagioclase, and olivine are the first crystallizing phases in all cumulate rocks, whereas pyroxene, hornblende, phlogopite, magnetite, and ilmenite represent the intercumulus phases. The existence of hornblende and phlogopite, the high Wo and Cr 2O 3 contents of clinopyroxene, and the high Fo and NiO contents of olivine indicate that the KUMA units were the near-liquidus fractional crystallization products of hydrous basaltic magma. The Phase-I and Phase-III cumulates are slightly enriched in LILE and LREE, depleted in HFSE and HREE with respect to chondrite and primitive mantle, and show negative Nb, Zr, Hf and Ti anomalies. The Phase-IV cumulate gabbronorites are even more enriched in LILE and LREE and show less depletion in HFSE and HREE in comparison to the earlier phases. The KUMA units formed from high-Al hydrous basaltic magmas, which were derived from partial melting of a previously subduction-metasomatized subcontinental-lithospheric mantle. These magmas underwent crystal fractionation and mineral concentration processes in small magma chamber(s) at shallow crustal levels (< 5 kbar).

Evidence for a homogeneous primary magma at Piton de la Fournaise (La Réunion): A geochemical study of matrix glass, melt inclusions and Pélé's hairs of the 1998–2008 eruptive activity

Magmas erupted at Piton de la Fournaise volcano since 0.5 Ma, display a large petrological and chemical range (picrites, 2 types of transitional basalts and differentiated magmas) and low amplitude isotopic heterogeneities. The recent activity (1998–2008) includes all magma types except evolved magmas. Matrix glass compositions from quenched lavas and Pélé's hairs of the whole 1998–2008 period define a single differentiation trend from a common basaltic melt (MgO ~ 9%) for the first time identified in the 2007 magmas. More primitive melt compositions (MgO ~ 12.5%) are only evidenced by olivine crystals with high Fo contents (Fo 85–88.4). Evolutions of major and trace element of glass and mineral compositions are consistently modelled by a unique low pressure crystal fractionation process. The composition range of olivine melt inclusions is distinct from that of matrix glass and Pélé's hair and corresponds to equilibrium crystallisation in closed system of melts trapped from the main differentiation series at high temperature. The range of basaltic types at Piton de la Fournaise is the result of large variations in the differentiation degree (10 to 35% crystallisation) of a single primary basaltic melt and the addition in highly variable amounts (up to 50% in picrites) of co-genetic olivine or gabbroic cumulates. These cumulates may represent the shallow and dense bodies identified by seismic tomography and have likely been produced by the repetitive intrusion and differentiation of basalts along Piton de la Fournaise history. Depending on the shallow transfer paths, ascending magmas may disaggregate and incorporate various types of cumulates, explaining all particular features of basaltic magmas and picrites. These results emphasize the exceptional chemical homogeneity of the primary basaltic melt and of the differentiation process involved in volcanic activity of La Réunion hotspot since 0.5 Ma and the increasingly recognised role of melt–wall rock interactions in compositional and petrological diversity of erupted magmas.

Layered granitoids: Interaction between continental crust recycling processes and mantle-derived magmatism: Examples from the Évora Massif (Ossa–Morena Zone, southwest Iberia, Portugal)

In this paper, field, petrographic, mineralogical, geochemical and isotopic (Rb–Sr and Sm–Nd) information from three areas within the Évora Massif (Iberian Variscan Orogen) is presented and discussed aiming at to unravel the relationships between granitoids and units mapped as migmatites and also to evaluate the interplay between mantle and crustal derived magmas. One of the areas – Almansor – displays a well-developed compositional layering (concordant with the regional Variscan structure) which was considered, in previous works, as an alternation of leucosome and melanosome. In this study, the layering is described as intercalation of diatexites, weakly foliated granitoids and trondhjemitic veins. Diatexites have characteristics of crustal melts plus restitic material and, according to geochemical and isotopic evidence, result from anatexis of Ediacaran metasediments. Weakly foliated granitoids and trondhjemitic veins from Almansor have calc-alkaline signatures and may be related to each other by crystal fractionation processes; however, the mixing between mafic (mantle-derived) and felsic (diatexitic melt) magmas revealed by the isotopic data may also explain their genesis. In the Alto de São Bento area, several igneous lithologies (tonalites, granodiorites, porphyritic granites and leucogranites) are present and show typical isotropic igneous textures. Despite structural and textural differences, geochemical data support, for most rocks, an origin from the same calk-alkaline suite, also present at Almansor. The Alto de São Bento leucogranites have an isotopic signature that, although different from that obtained in the Almansor diatexites, is still compatible with an origin involving melting of Ediacaran metasediments; compositions, with very low contents of usually incompatible elements, flat normalized REE patterns and strong negative Eu anomalies, suggest that the anatectic melt has undergone crystal fractionation processes before reaching the composition of the leucogranite magma. The Almansor outcrop is then interpreted as the remnants of a shear zone that operated as a pathway for melts that moved upward through the crust providing the locus for differentiation and mingling/mixing of magmas, whilst Alto de São Bento would correspond to the zone, at a higher crustal level, where magmas were trapped and forced to spread horizontally. At Valverde (the third area) foliated and non-foliated granitoids are spatially related and field criteria links these rocks to metamorphic protolith and anatectic melt, respectively. However, petrographic, geochemical and isotopic information shows that they all are compositionally identical trondhjemites with no evidence of metamorphic fabric. In the foliated rocks, mesoscopic features are interpreted as resulting from melt segregation structures formed in a crystallizing mush. In contrast to the previous areas, the Valverde trondhjemites probably do not belong to the main calc-alkaline plutonic suite of the Évora Massif, since they have a distinct Sr and Nd isotope signature.

Into the Field Again: Re-Examining Charles Darwin's 1835 Geological Work on Isla Santiago (James Island) in the Galápagos Archipelago

In 1835 Charles Darwin's geological observations on Isla Santiago (James Island) in the Galápagos Islands led him to important insights as to the process by which different varieties of igneous rock might be produced from the same volcanic vent. His work figured in a tradition of interpretation that began with the work of George Poulett Scrope and would end in the twentieth century with the theory of magmatic differentiation of igneous rocks through the process of crystal fractionation. This article reports on the findings of an expedition to Isla Santiago in July 2007 during which we were able to locate samples of igneous rocks similar to those collected by Darwin. We have used these, together with Darwin's original specimens and transcriptions of his field notes, to examine how his understanding of the separation of the trachytic and basaltic series of magmas developed from his initial field observations through to publication of Volcanic Islands in 1844.

Quaternary adakite—Nb-enriched basalt association in the western Trans-Mexican Volcanic Belt: is there any slab melt evidence?

A spatial and temporal association between adakitic rocks and Nb-enriched basalts (NEB) is recognised for the first time in the western sector of the Trans-Mexican Volcanic Belt in the San Pedro–Cerro Grande Volcanic Complex (SCVC). The SCVC is composed of subalkalic intermediate to felsic rocks, spanning in composition from high-silica andesites to rhyolites, and by the young transitional hawaiite and mugearite lavas of Amado Nervo shield volcano. Intermediate to felsic rocks of the SCVC show many geochemical characteristics of typical adakites, such as high Sr/Y ratios (up to 180) and low Y (<18 ppm) and Yb contents. Mafic Amado Nervo rocks have high TiO2 (1.5–2.3 wt%), Nb (14–27 ppm), Nb/La (0.5–0.9) and high absolute abundances of HFSE similar to those shown by NEB. However, the Sr and Nd isotopic signature of SCVC rocks is different from that shown by typical adakites and NEB. Although the adakites–NEB association has been traditionally considered as a strong evidence of slab-melting, we suggest that other processes can lead to its generation. Here, we show that parental magmas of adakitic rocks of the SCVC derive their adakitic characteristic from high-pressure crystal fractionation processes of garnet, amphibole and pyroxene of a normal arc basalt. On the other hand, Amado Nervo Na-alkaline parental magmas have been generated by sediment melting plus MORB-fluid flux melting of a heterogeneous mantle wedge, consisting of a mixture of depleted and an enriched mantle sources (90DM + 10EM). We cannot exclude a contribution to the subduction component of slab melts, because the component signature is dominated by sediment melt, but we argue that caution is needed in interpreting the adakites–NEB association in a genetic sense.

Geochemistry of mafic phenocrysts from alkaline lamprophyres of the Spanish Central System: implications on crystal fractionation, magma mixing and xenoliths entrapment within deep magma chambers

The Permian alkaline lamprophyres from the Spanish Central System (SCS) are highly porphyritic rocks which carry a heterogeneous population of clinopyroxene and kaersutite zoned phernocrysts. Clinopyroxene phenocrysts may show 1) normal zoning (Cpx-I), 2) reverse zoning with Fe-rich green cores (Cpx-II), and 3) reverse zoning with colourless Al-poor, silica-rich cores (Cpx-III). Kaersutite phenocrysts also show a slight reverse zoning. Major and trace element composition of Cpx-I suggests that their compositional variation is related to a crystal fractionation process from melts similar to the host lamprophyres. The Cpx-II cores represent crystallization from highly evolved melts (low Mg-Cr contents and incompatible element enrichment), genetically related with the SCS alkaline magmatism, and the growth or surrounding Mg-rich inner rims points to a magma mixing process. The major and trace element composition of Cpx-III cores supports derivation from a magma which has fractionated plagioclase. This characteristic, together with their similarities when compared to clinopyroxenes from charnockite xenoliths, suggests that they might be xenocrysts from deep calc-alkaline cumulates. The composition of melts in equilibrium with clinopyroxene and amphibole phenocrysts supports a model in which Cpx-II and Cpx-III cores would have been incorporated into a more primitive lamprophyric magma stagnated at lower crustal levels. The low pressure composition of all phenocryst outer rims indicates that they crystallised directly from the host alkaline magma at their subvolcanic emplacement levels.

Formation of Distinct Granitic Magma Batches by Partial Melting of Hybrid Lower Crust in the Izu Arc Collision Zone, Central Japan

The Miocene Kofu Granitic Complex (KGC) occurs in the Izu Collision Zone where the Izu^Bonin^Mariana (IBM) arc has been colliding with the Honshu arc since the middle Miocene.The KGC includes rocks ranging in compositions from biotite-bearing granite (the Shosenkyo and Mizugaki plutons), and hornblende^biotitebearing granodiorite, tonalite, quartz-diorite, and granite (the Shiodaira, Sanpo, Hirose and Sasago plutons), to hornblende-bearing tonalite and trondhjemite (the Ashigawa^Tonogi pluton), indicating that it was constructed from multiple intrusions of magma with different bulk chemistry. The Sr-isotopic compositions corrected to sensitive high-resolution ion microprobe (SHRIMP) zircon ages (SrI) suggest that the primary magmas of each pluton were formed by anatexis of mixed lower crustal sources involving both juvenile basalt of the IBM arc and Shimanto sedimentary rocks of the Honshu arc. After the primary magmas had formed, the individual plutons evolved by crystal fractionation processes without significant crustal assimilation or additional mantle contribution. SHRIMP zircon U^Pb ages in the KGC range from 16 8 to 10 6Ma and overlap the resumption of magmatic activity in the IBMand Honshu arcs at c. 17Ma and the onset of IBM arc^Honshu arc collision at c. 15Ma.The age of the granite plutons is closely related to the episodic activity of arc magmatism and distinct granitic magma batches could be formed by lower crustal anatexis induced by intrusion of underplated mantle-derived arc magmas. Based on pressures determined with the Al-in-hornblende geobarometer, the KGC magmas intruded into the middle crust. Thus, the KGC could represent an example of the middle-crust layer indicated throughout the IBM arc by 6 0^6 5 km/s seismic velocities. This granitic middle-crust layer acted buoyantly during the IBM arc^Honshu arc collision, leading to accretion of buoyant IBMarc middle crust to the Honshu arc.

Special Paper: Adakite-Like Rocks: Their Diverse Origins and Questionable Role in Metallogenesis

Based on a compilation of published sources, rocks referred to as adakites show the following geochemical and isotopic characteristics: SiO2 ≥56 wt percent, Al2O3 ≥15 wt percent, MgO normally <3 wt percent, Mg number ≈0.5, Sr ≥400 ppm, Y ≤18 ppm, Yb ≤1.9 ppm, Ni ≥20 ppm, Cr ≥30 ppm, Sr/Y ≥20, La/Yb ≥20, and 87Sr/86Sr ≤0.7045. Rocks with such compositions have been interpreted to be the products of hybridization of felsic partial melts from subducting oceanic crust with the peridotitic mantle wedge during ascent and are not primary magmas. High Mg andesites have been interpreted to be related to adakites by partial melting of asthenospheric peridotite contaminated by slab melts. The case for these petrogenetic models for adakites and high Mg andesites is best made in the Archean, when higher mantle geotherms resulted in subducting slabs potentially reaching partial melting temperatures at shallow depths before dehydration rendered the slab infusible. In the Phanerozoic these conditions were likely only met under certain special tectonic conditions, such as subduction of young (≤25-m.y.-old) oceanic crust. Key adakitic geochemical signatures, such as low Y and Yb concentrations and high Sr/Y and La/Yb ratios, can be generated in normal asthenosphere-derived tholeiitic to calc-alkaline arc magmas by common upper plate crustal interaction and crystal fractionation processes and do not require slab melting. An assessment of several arc volcanic suites from around the world shows that most adakite-like compositions are generated in this way and do not reflect source processes. Similarly, rare adakite-like intrusive rocks associated with some porphyry Cu deposits are the evolved products of extensive crustal-level processing of calc-alkaline basalt-andesite-dacite-rhyolite series magmas. If slab melts contribute to such magmas, their geochemical signatures would have been obliterated or rendered ambiguous by subsequent extensive open-system processes. In Archean terranes, where adakitic and high Al tonalite-trondhjemite-granodiorite (TTG) magma series rocks are more common, porphyry Cu deposits are rare and, where found, are associated with normal calc-alkaline suites rather than adakites. The two different magma series are compositionally distinct in terms of several major and trace element parameters. Common upper plate magmatic processes such as melting-assimilation-storage-homogenization (MASH) and assimilation-fractional-crystallization (AFC) affecting normal arc magmas can be demonstrated to explain the distinctive compositions of most adakite-like arc rocks, including high Mg andesites and especially those rare examples associated with porphyry Cu deposits. In contrast, slab melting can in most cases neither be proved nor disproved and is therefore unsatisfactory as a unique factor in porphyry Cu genesis.

On the structure of mare basalt lava flows from textural analysis of the LaPaz Icefield and Northwest Africa 032 lunar meteorites

Abstract— Quantitative textural data for Northwest Africa (NWA) 032 and the LaPaz (LAP) mare basalt meteorites (LAP 02205, LAP 02224, LAP 02226, and LAP 02436) prvide constraints on their crystallization and mineral growth histories. In conjunction with whole‐rock and mineral chemistry, textural analysis provides powerful evidence for meteorite pairing. Petrographic observations and crystal size distribution (CSD) measurements of NWA 032 indicate a mixed population of slowly cooled phenocrysts and faster cooled matrix. LaPaz basalt crystal populations are consistent with a single phase of nucleation and growth. Spatial distribution patterns (SDP) of minerals in the meteorites highlight the importance of clumping and formation of clustered crystal frameworks in their melts, succeeded by continued nucleation and growth of crystals. This process resulted in increasingly poor sorting, during competition for growth, as the melt crystallized. Based on CSD and SDP data, we suggest a potential lava flow geometry model to explain the different crystal populations for NWA 032 and the LaPaz basalts. This model involves crystallization of early formed phenocrysts at hypabyssal depths in the lunar crust, followed by eruption and flow differentiation on the lunar surface. Lava flow differentiation would allow for formation of a cumulate base and facilitate variable cooling within the stratigraphy, explaining the varied textures and modal mineralogies of mare basalt meteorites. The model may also provide insight into the relative relationships of some Apollo mare basalt suites, shallow‐level crystal fractionation processes, and the nature of mare basalt volcanism over lunar history.

Neo-Proterozoic rift-related syenites (Northern Damara Belt, Namibia): Geochemical and Nd–Sr–Pb–O isotope constraints for mantle sources and petrogenesis

Major element, trace element and Nd–Sr–Pb–O isotope data for a suite of Neo-Proterozic, pre-orogenic, rift-related syenites from the Northern Damara orogen (Namibia) constrain their sources and petrogenesis. New U–Pb ages obtained on euhdreal titanite of inferred magmatic origin constrain the age of intrusion of the Lofdal and Oas syenites to ca. 750 Ma compatible with previous high-precision zircon analyses from the Oas complex. Major rock types from Lofdal and Oas are mildly sodic nepheline-normative and quartz-normative syenites and were primarily generated by fractional crystallization from a mantle-derived alkaline magma. Primitive samples from Lofdal and Oas show depletion of Rb, K and Th relative to Ba and Nb together with variable negative anomalies of P and Ti on a primitive mantle-normalized diagram. Evolved samples from Oas develop significant negative Ba, Sr, P and Ti anomalies and positive U and Th anomalies mainly as a function of crystal fractionation processes. The lack of a pronounced negative Nb anomaly in samples from Lofdal suggests that involvement of a crustal component is negligible. For the nepheline-normative samples from Lofdal, the unradiogenic Sr and radiogenic Nd isotope composition and low δ 18O values suggest derivation of these samples from a moderately depleted lithospheric upper mantle with crustal-like U/Pb ratios ( 87Sr/ 86Sr: 0.7031–0.7035, ε Nd: ca. + 1, δ 18O: 7‰, 206Pb/ 204Pb: ca.18.00, 207Pb/ 204Pb: 15.58–15.60). Primitive samples of the Oas quartz-normative syenites have identical isotope characteristics ( 87Sr/ 86Sr: 0.7034, ε Nd: ca. + 1, δ 18O: 6.5‰, 206Pb/ 204Pb: ca.18.00, 207Pb/ 204Pb: 15.59) whereas more differentiated samples have higher 87Sr/ 86Sr ratios (0.709–0.714), slightly higher δ 18O values (7.0–7.1‰), less radiogenic ε Nd values (− 1.1 to − 1.4) and more radiogenic 206Pb/ 204Pb ratios up to 18.27. These features together with model calculations using Sr–Nd–Pb isotopes suggest modification of a primary syenite magma by combined AFC processes involving ancient continental crust. In this case, high Nb abundances of the parental syenite liquid prevent the development of significant negative Nb anomalies that may be expected due to interaction with continental crust.

The central Kenya peralkaline province: Insights into the evolution of peralkaline salic magmas

The central Kenya peralkaline province comprises five young (< 1 Ma) volcanic complexes dominated by peralkaline trachytes and rhyolites. The geological and geochemical evolution of each complex is described and issues related to the development of peralkalinity in salic magmas are highlighted. The peralkaline trachytes may have formed by fractionation of basaltic magma via metaluminous trachyte and in turn generated pantellerite by the same mechanism. Comenditic rhyolites are thought to have formed by volatile-induced crustal anatexis and may themselves have been parental to pantelleritic melts by crystal fractionation. The rhyolites record very low temperatures of equilibration (≤ 700 °C) at low fO 2 (≤ FMQ). The development of compositional zonation within the magma reservoirs has been ubiquitous, involving up to tens of cubic km of magma at timescales of 10 3–10 4 years. Magma mixing has also been commonplace, sometimes between adjacent centres. Isotopic evidence relating to rates and timescales of pre-eruptive residence times and crystal fractionation processes is summarized.

ROCHAS VULCÂNICAS ALCALINAS DA FORMAÇÃO QUIXABA DE FERNANDO DE NORONHA, PE: QUÍMICA MINERAL E LITOGEOQUÍMICA

The lava flows of the Quixaba Formation (4,2 - 1,5 Ma) are olivine melanephelinites (OLM), olivine melilite melanephelinites (MEM) and pyroxene melanephelinites (PIM), with levels of pyroclastic rocks and relatively thin basanite flows (BAS). Thick BAS sequences occur in Baía do Sancho and in the São José, Cuscuz and de Fora islets. Olivine compositions (m# = 85 - 75) in melanephelinites and most BAS match those of clinopyroxenes (mg# 86 -74), indicating equilibrium crystallization. The São José and Cuscuz basantitic olivine (mg# 72 - 70) and clinopyroxene (mg# = 79 -74) compositions are similar and denote more differentiated rocks. In MEM, the amount (&gt; 10% to &lt;2%) and composition of melilites are variable. The melilite-rich rocks have abundant olivine, with melilites poorer in MgO and richer in Al 2 O 3 . Ba, Ti- (and F) rich phlogopite and biotite are late minerals with typical chemical features in each rock type. The Si0 2 content is an indicator of the chemical differences between melanephelinites: -MEM: 37-39 wt.%; OLM: 38-40 wt.%; PIM: 40-43 wt.%-. BAS have 45-46 wt.% Si02. The increase in silica is accompanied by increase in Al 2 O 3 and decrease in MgO, Ti0 2 , Fe 2 O 3total , CaO, MnO, Cs, Y and REE. Trace element ratios show important discrepancies among the melanephelinite types: MEM e OLM: La/Nb = 0.6-0.7; Zr/Nb = 2.7-3.2; PIM: La/Nb = 0.7-1.0; Zr/Nb = 2.6-5.7. The amount of trace elements and the incompatible element ratios are variable in BAS. The data presented in this paper denote the existence of crystal fractionation processes, for instance between PIM and the thin BAS flows in the lava sequence. On the other hand, the high amount of incompatible elements in the melanephelinites and the differences in element ratios in the several rock types, point to an origin from distinct parental magmas, derived from either different mantle sources, or different degrees of partial melting of the same source.

ROCHAS VULCÂNICAS ALCALINAS DA FORMAÇÃO QUIXABA DE FERNANDO DE NORONHA, PE: QUÍMICA MINERAL E LITOGEOQUÍMICA

The lava flows of the Quixaba Formation (4,2 - 1,5 Ma) are olivine melanephelinites (OLM), olivine melilite melanephelinites (MEM) and pyroxene melanephelinites (PIM), with levels of pyroclastic rocks and relatively thin basanite flows (BAS). Thick BAS sequences occur in Baia do Sancho and in the Sao Jose, Cuscuz and de Fora islets. Olivine compositions (m# = 85 - 75) in melanephelinites and most BAS match those of clinopyroxenes (mg# 86 -74), indicating equilibrium crystallization. The Sao Jose and Cuscuz basantitic olivine (mg# 72 - 70) and clinopyroxene (mg# = 79 -74) compositions are similar and denote more differentiated rocks. In MEM, the amount (> 10% to <2%) and composition of melilites are variable. The melilite-rich rocks have abundant olivine, with melilites poorer in MgO and richer in Al 2 O 3 . Ba, Ti- (and F) rich phlogopite and biotite are late minerals with typical chemical features in each rock type. The Si0 2 content is an indicator of the chemical differences between melanephelinites: -MEM: 37-39 wt.%; OLM: 38-40 wt.%; PIM: 40-43 wt.%-. BAS have 45-46 wt.% Si02. The increase in silica is accompanied by increase in Al 2 O 3 and decrease in MgO, Ti0 2 , Fe 2 O 3total , CaO, MnO, Cs, Y and REE. Trace element ratios show important discrepancies among the melanephelinite types: MEM e OLM: La/Nb = 0.6-0.7; Zr/Nb = 2.7-3.2; PIM: La/Nb = 0.7-1.0; Zr/Nb = 2.6-5.7. The amount of trace elements and the incompatible element ratios are variable in BAS. The data presented in this paper denote the existence of crystal fractionation processes, for instance between PIM and the thin BAS flows in the lava sequence. On the other hand, the high amount of incompatible elements in the melanephelinites and the differences in element ratios in the several rock types, point to an origin from distinct parental magmas, derived from either different mantle sources, or different degrees of partial melting of the same source.

Mantle processes in an Archean orogen: Evidence from 2.67 Ga diamond-bearing lamprophyres and xenoliths

The world's oldest diamond deposits occur in ∼2.67 Ga dikes and heterolithic breccias emplaced into greenstone belts of the Wawa and Abitibi Subprovinces, southern Superior Province, Canada. Thousands of white to yellow microdiamonds and macrodiamonds to 5 mm in width have been recovered by non-contaminating fusion techniques. The host rocks exhibit petrographic and compositional features that are characteristic of post-Archean minettes and spessartites of the calc-alkaline or shoshonitic lamprophyre clan. Based on chemical trends and petrographic evidence, host rocks that contain more than ∼16 wt.% MgO represent lamprophyre magmas that entrained cumulate olivine, probably at the base of the crust. Breccia bodies that are tens of metres wide at the two localities are somewhat atypical of late Archean lamprophyre occurrences in the Superior Province and owe their size to optimum conditions for magma ascent that were required to preserve the diamonds. Abundant altered ultramafic xenoliths occur in the host rocks. The majority of xenoliths studied (10 of 14) display uniform major element compositions similar to websterite cumulate suites derived from crystal fractionation processes at the base of post-Archean volcanic arcs. The xenoliths display highly variable trace element abundances that are characteristic of cryptic metasomatism associated with the flux of an oxidised fluid above a subduction zone. The tectonic setting of the deposits and the nature of the host rocks indicate that the diamonds may be derived from the asthenospheric wedge and subducted slab at shallow depths (100 to 160 km) rather than the deep keels of Archean cratons associated with traditional diamond deposit types. Models of low-temperature Phanerozoic diamond formation in active subduction zones, or rapid uplift and emplacement of peridotite massif occurrences, can be adapted to the Archean deposits. The stability field of diamonds in most Phanerozoic subduction scenarios, however, may be too deep to be accessed by the lamprophyric magmas. In contrast, shallow subduction, as invoked for the distinctive occurrence of adakitic (slab-melt) type rocks in the southern Superior Province, could generate two different diamond stability windows at sufficiently shallow depths to account for their presence in lamprophyric magmas. The multiple requirements imposed on Archean tectonic models by occurrences of diamonds in hydrous shoshonitic rock types (spessartite and minette lamprophyres), along with distinctively metasomatised xenoliths, strongly favour plate tectonic subduction models of orogeny. Evidence of slightly earlier mantle plumes, such as 2.7 Ga komatiites, only strengthens the need for a subduction-driven low-temperature thermal anomaly in the Archean mantle prior to lamprophyric magmatism.

Origin and internal evolution of the Li-F-Be-B-P-bearing Pinilla de Fermoselle pegmatite (Central Iberian Zone, Zamora, Spain)

The Li-F-Be-B-P bearing Pinilla de Fermoselle (PF) pegmatite occurs in the apical part of a leucogranite body. It shows a clear non-symmetrical vertical zoning from the contact with the leucogranite to a contact with the metamorphic country rocks. The pegmatitic facies evolve upward from (1) the undifferentiated Lower Border Zone (LBZ), with quartz, feldspars, muscovite, biotite, and black tourmaline, through (2) the Intermediate Zone (IZ), with quartz, muscovite, zinnwaldite, black tourmaline, and Fe-Mn phosphates, to (3) the highly evolved Upper Border Zone (UBZ), with quartz, albite, lepidolite, zinnwaldite, elbaite, and beryl. The composition of the pegmatite-forming minerals suggests that a residual melt become progressively enriched in F and Li until the crystallization of the apical UBZ, whereas P partitions in the melt only until the intermediate levels of differentiation attained in the IZ. Chemical variations in the mica and tourmaline as well as in the feldspar and Fe- Mn phosphate minerals are consistent with an internal evolution by crystal fractionation processes. A plausible model for the crystallization of the PF pegmatite involves a rapid, in situ, bottom-up crystallization from significantly undercooled liquids. The lack of metasomatic effects in the metamorphic host-rock and the estimated P content of the initial leucogranite melt suggest that the PF pegmatite mainly crystallized under closed system conditions.

Major and trace element and isotopic (Sr, Nd, O) constraints for Pan-African crustally contaminated grey granite gneisses from the southern Kaoko belt, Namibia

Light-grey to dark-grey, fine-grained to medium-grained granitic to granodioritic gneisses are substantial part of the Nk Formation in the southern Kaoko belt, Namibia. Based on field relations, these rocks can be subdivided into two distinct groups. One group belongs to a pluton with an areal extent of approximately 6 km 2, the other group represents gneissic granitic veins. For all samples, SiO 2 varies between 68.1 and 74.7 wt.%. The LREE-enriched granite gneisses of the Nk pluton are slightly peraluminous (A.S.I. = 1.05–1.25) with low Rb concentrations (86–133 ppm) and high concentrations of Sr and Ba (187–683 and 644–1535 ppm, respectively). The gneisses have initial ε Nd (500 Ma) values ranging from −4.1 to −10.7 and initial 87Sr/ 86Sr (500 Ma) ratios ranging from 0.70695 to 0.71108 and whole rock oxygen isotope values ranging from 9.4‰ to 10.7‰. The gneissic granitic veins can be subdivided into three different groups (groups A, B, C). Most geochemical data are similar to those obtained for the pluton, only the δ 18O values for group B and the Rb/Sr ratios for group C are distinct from the pluton. Linear variation of major and trace elements in the Nk granite gneisses from the pluton can be explained by igneous fractionation involving plagioclase (10 wt.%), biotite (1 wt.%), alkali feldspar (13 wt.%) and LREE-enriched accessory minerals. The 87Sr/ 86Sr ratios become more radiogenic and δ 18O values increase with increasing SiO 2 indicating that in addition to crystal fractionation processes, the Nk granite gneisses were affected by assimilation processes involving contaminants similar to basement derived granites from the Central Damara orogen which show lower ε Nd (500 Ma) ratios. The protolith of the Nk granite gneisses is most likely a biotite- and/or amphibole-bearing but muscovite-free rock, most probably an intermediate igneous rock (i.e. diorite to granodiorite) or a metagreywacke. Concerning the isotopic differences between the granite gneisses from the pluton and the gneissic granitic veins, partial melting probably took place in a heterogeneous compositionally layered middle to lower crust at temperatures of 845 ± 10 °C as calculated by monazite and zircon saturation equations. The final stages of continental collision during westward dipping subduction of the Kaoko belt beneath the Ribeira belt took place between ca. 590 and 520 Ma. In view of such a Pan-African subduction zone related geotectonic situation, the Nk granite gneisses are products of melting events in arc-derived rocks, produced in an earlier stage of the Pan-African orogeny.