Nicholas T. Arndt

Crustal growth in West Africa at 2.1 Ga

Birimian (∼2.1 Ga) terranes in the West African craton are a mixture of highly metamorphosed volcanic, sedimentary and plutonic rocks and low grade metavolcanics and metasediments. The volcanic units contain thick, commonly pillowed, tholeiitic basalts overlain by pelagic sediments cherts and black shales; The sedimentary units are characterized by an abundance of clastic turbiditic sediments. Andesites and calc-alkaline felsic volcanics occur at uppermost stratigraphic levels and as dykes. Field relationships between the volcanic and sedimentary units remain a matter of debate. Calc-alkaline and local alkaline granites, which intruded in distinct pulses and occasionally are related to transcurrent tectonics, represent almost half of the Birimian terranes. New isotopic work on the highly metamorphosed units greatly improved the chronology for the Birimian crust. The age of the early Dabakalian event is precisely defined by a U-Pb zircon age at 2186 ± 19 Ma, while Rb-Sr and Sm-Nd methods give ages of 2162 ± 19 Ma and 2141 ± 24 Ma, respectively. A Sm-Nd garnet-whole rock age of 2153 ± 13 Ma suggests that metamorphism culminated at about the same time. In contrast, the most precise zircon U-Pb and Sm-Nd data for the more widespread Birimian terranes (sensu stricto), from this study and from the literature, cluster between 2.12 and 2.07 Ga. The major evolution of the Birimian crust apparently lasted less than 50 Ma. Isotopic evidence indicates that Birimian granitoids contain a negligible component of Archean crust: eNd(2.1-Ga) values are positive and similar to those of Birimian basalts, crustal residence times are shorter than 200 Ma, U-Pb ages for detrital zircons from clastic sediments range from 2098 ±11 Ma to 2125 ± 17 Ma, while granite chemistry and Nd isotopic characteristics are unrelated. Only very locally in Guinea is there isotopic evidence of interaction between Birimian felsic magmas and the Archean rocks from the Man craton. In accord with Abouchami et al.s (1990) suggestion that Birimian basalts represent oceanic plateaus, the present data argue that the protolith of much of the West African continent was created around 2.1 Ga in an environment remote from Archean crust. Intrusion of calc-alkaline magmas into the tholeiitic units suggests that island arcs formed on top of the assumed oceanic plateaus which then collided with the Man Archean craton. Taking the Birimian formations from the Guyana shield into account, the minimum crustal growth rate at 2.1 Ga is about 1.6 km3/a, some ∼60% higher than the present growth rate. Birimian crust growth at 2.1 Ga is reminiscent of Archean processes but contrasts with 1.7 – 1.9 Ga crust formation in the North Atlantic continent which generally involved significantly more interaction with older continental crust. A comparison of the Birimian crustal growth rate with the average crustal growth rate over the Earth history implies that a large part of the Birimian crust has been recycled into the mantle or incorporated into younger orogenic segments. This apparent deficit in the crustal budget is even more dramatic for the Archean crust.

Use and abuse of crust-formation ages

Samarium-neodymium (Sm-Nd) isotopic studies can be used to evaluate the history of crustal growth and sometimes to give “crust-formation” ages, which reflect the time of differentiation of crust from the mantle. If, however, a sample is a mixture of material derived from the mantle at different times, Sm-Nd systematics may provide only an estimate of the average time that the material in the sample has been resident in the continental crust. In such cases, Sm-Nd isotopes give no direct information on the timing of crustal formation. These ages can be interpreted as the time of crust-mantle segregation only if supported by other geologic and geochronological information. Misinterpretation can lead to false conclusions about the history of crustal development.

The heterogeneous Iceland plume: Nd‐Sr‐O isotopes and trace element constraints

We present a comprehensive set of Sr, Nd, and O isotope data and trace element concentrations from tholeiitic and alkaline lavas of the neovolcanic zones of Iceland (picrites, olivine and quartz tholeiites, transitional and alkali basalts, differentiated rocks). Variations in the oxygen isotope results allow us to distinguish two groups. The first, which comprises quartz tholeiites and more differentiated rocks usually associated with central volcanoes, has low δ18O values (+5 to +1‰) resulting from interaction with the hydrothermally altered Icelandic crust. The second group, which contains picrites, olivine tholeiites, and alkali basalts, has normal mantle oxygen isotopic compositions (δ18O = +5 to +6‰) which are thought to represent those of the mantle source. Nd isotopic compositions vary greatly, from 143Nd/144Nd = 0.51314 in picrites to 0.51295 in alkali basalts. To produce such a variation for rocks with the chemical compositions of Icelandic volcanics (147Sm/144Nd = 0.12=0.28) requires >200 m.y., a period that greatly exceeds the maximum age of Icelandic crust. Previous models, in which the Sr isotopic variations were explained in terms of evolution of crustal reservoirs, are invalidated, and mantle reservoirs with different Nd and Sr isotopic compositions are indicated. The Iceland data define a linear array in the Sr-Nd isotope diagram which overlaps both mid-ocean ridge basalt and oceanic island basalt fields and indicates mixing between depleted and enriched end-members. Alkali basalts come preferentially from an isotopically and chemically enriched component of the Iceland plume, and picrites come from a more refractory, more depleted portion. Positive Sr, Rb, and Ba anomalies are present in picrites and other lavas with low trace element contents. These anomalies are not correlated with isotopic differences but are nevertheless believed to result from interaction between the parent magmas of these rocks and altered Icelandic crust. This indicates that even the most primitive Icelandic lavas have been contaminated with some crustal material.

The role of lithospheric mantle in continental flood volcanism: Thermal and geochemical constraints

Continental flood basalts (CFB) are commonly said to form by direct melting of metasomatized lithospheric mantle, either during major lithospheric extension or when a mantle plume impinges on the base of the lithosphere. We tested these ideas in a thermomechanical model that combines lithospheric dynamics and mantle convection. Dry melting was assumed, and the proportions of melt from different source regions were monitored. In all cases, >96% of melt was found to come from asthenosphere or plume, with minimal amounts from continental lithosphere. During passive lithosphere extension the total amount of melt is small, and the proportion from the lithosphere is 100 km) and have high concentrations of both MgO (>20%) and incompatible trace elements (K2O ∼ 1%). The very low Nb and Ta concentrations in certain CFB cannot, however, be explained by this process. Ratios of Nb to elements such as La or U are lower in many flood basalts and picrites than in all likely source materials: they are almost as low as in most rocks from the continental crust, and they are far lower than in peridotites from the lithospheric or asthenospheric mantle. Another process must therefore fractionate Nb and Ta. We suggest that this takes place during the passage of magma through the lithospheric mantle, perhaps because of differences in the reaction rates of minerals in metasomatized peridotite. The probability that CFB are hybrid magmas containing material from aesthenospheric and lithospheric mantle and in many cases from continental crust, as well as the possibility that some elemental ratios change during magma-lithosphere interaction, casts serious doubt on the reliability of such rocks as probes of the lithospheric mantle.

Constraining the potential temperature of the Archaean mantle: A review of the evidence from komatiites

The maximum potential temperature of the Archaean mantle is poorly known, and is best constrained by the MgO contents of komatiitic liquids, which are directly related to eruptive temperatures. However, most Archaean komatiites are significantly altered and it is difficult to assess the composition of the erupted liquid. Relatively fresh lavas from the SASKMAR suite, Belingwe Greenstone Belt, Zimbabwe (2.7 Ga) include chills of 25.6 wt.% MgO, and olivines ranging to Fo93.6, implying eruption at around 1520°C. A chill sample from Alexo Township, Ontario (also 2.7 Ga) is 28 wt.% MgO, and associated olivines range to Fo94.1, implying eruption at 1560°C. However, inferences of erupted liquids containing 32–33 wt.% MgO, from lavas in the Barberton Greenstone Belt, South Africa (3.45 Ga) and from the Perseverance Complex, Western Australia (2.7 Ga) may be challenged on the grounds that they contain excess (cumulate) olivine, or were enriched in Mg during alteration or metamorphism. Re-interpretation of olivine compositions from these rocks shows that they most likely contained a maximum of 29 wt.% MgO corresponding to an eruption temperature of 1580°C. This composition is the highest liquid MgO content of an erupted lava that can be supported with any confidence. The hottest modern magma, on Gorgona Island (0.155 Ga) contained 18–20% MgO and erupted at circa 1400°C. If 1580°C is taken as the temperature of the most magnesian known eruption, then the source mantle from which the liquids rose would have been at up to 2200°C at pressures of 18 GPa corresponding to a mantle potential temperature of 1900°C. These temperatures are in excess of the mantle temperatures predicted by secular cooling models, and thus komatiites can only be formed in hot rising convective jets in the mantle. This result requires that Archaean mantle jets may have been 300°C hotter than the Archaean ambient mantle temperature. This temperature difference is similar to the 200–300°C temperature difference between present day jets and ambient mantle temperatures. An important subsidiary result of this study is the confirmation that spinifex rocks may be cumulates and do not necessarily represent liquid compositions.

An open boundary between lower continental crust and mantle: its role in crust formation and crustal recycling

Two processes may transport material across the boundary between lower continental crust and upper mantle: 1. (a)Relatively dense picritic magmas may be trapped at or near the base of the crust where they become contaminated with lower crustal wall rocks and differentiate to form layered sills with gabbroic upper parts and olivine + pyroxene cumulate lower parts. The siliceous, aluminous, incompatible-element enriched gabbroic rocks have relatively low densities and are incorporated in the continental crust; the ultramafic cumulates have higher densities than lower crustal and upper mantle rocks and sink into the mantle carrying with them lower crustal contaminant. 2. (b) Granitoid magmas formed during intracrustal melting leave dense restite minerals which also return to the mantle. These crustal foundering processes result in 1. (a) the building of continental crust; 2. (b) enrichment of continental lithospheric mantle with material from the lower continental crust; 3. (c) formation of sources for certain oceanic island basalts.

Isotopic and trace-element constraints on mantle and crustal contributions to Siberian continental flood basalts, Noril'sk area, Siberia

Abstract We present a tightly controlled and comprehensive set of analytical data for the 250-Ma Siberian flood-basalt province. Consideration of major- and trace-element compositions, along with strontium, lead and neodymium isotopic compositions, strongly supports earlier Russian subdivision of this magmatism into three magmatic cycles, giving rise to three assemblages of eleven basalt suites in the ascending order Ivakinsky-Gudchikhinsky, Khakanchansky-Nadezhdinsky and Morongovsky-Samoedsky. Geochemical and isotopic discontinuities of varying magnitude characterize most of the boundaries between the eleven recognized basalt suites in the Norilsk area. Although we conclude that the dominant volume of erupted magma originated from an asthenospheric mantle plume, none of the lavas is interpreted to directly represent asthenospheric melts, which would have been far more magnesian. On the basis of thermal considerations, we consider it unlikely that vast volumes of basaltic melt were produced directly from the continental lithospheric mantle beneath the Siberian craton. Moreover, there is little evidence from mantle xenoliths that the geochemical signatures of such melts would correspond to those of the Siberian flood basalts. Studies of melt migration lead us to conclude that transport of asthenospheric melt through the lithospheric mantle would be rapid, by fracture propagation. Lavas from the Gudchikhinsky suite have negligible Ta-Nb anomalies and positive ϵ Nd values and their parental magmas presumably interacted little with the continental lithospheric mantle or crust. All other lavas have negative Ta-Nb anomalies and lower ϵ Nd values that we attribute to interaction with continental crust. The model that we have developed requires discrete contributions from the plume and complex processing of all erupted magmas in the continental crust. The earliest magmas represent small percentages of melt formed in equilibrium with garnet. Over time, the percentage of melting in the source region and the volume of magma produced increased, and garnet was no longer stable in the plume source. All of the plume-derived melts initially contained more than 20 wt% MgO and became less Mg rich by fractionation of olivine as they traversed the lithospheric mantle. We conclude, however, that the most significant control on the geochemical and isotopic compositions of all the erupted lavas was processing of mantle-derived magma in crustal reservoirs during periodic replenishment, periodic tapping, continuous crystal fractionation and wallrock assimilation. Rapid eruption of an extremely large volume of processed magma that varied little in chemical and isotopic composition produced the sequence of relatively monotonous tholeiitic basalts that constitute the 2,300-m-thick third assemblage of the Siberian flood-basalt province near Norilsk.

Ir, Ru, Pt, and Pd in basalts and komatiites: New constraints for the geochemical behavior of the platinum-group elements in the mantle

Abstract The concentrations of the platinum-group elements (PGE) Ir, Ru, Pt, and Pd were determined in 18 mantle-derived basalts from a variety of tectonic settings and six komatiites from three locations. All analyses were performed using isotope dilution, Carius tube digestion, and the precise technique of multiple collector inductively coupled plasma mass spectrometry. Multiple analyses of two samples indicate external reproducibilities, based upon separate dissolutions, of approximately 2–9% in the ppt to ppb concentration range. Mid-ocean ridge basalts from the Kolbeinsey Ridge, tholeiites from Iceland and alkali basalts from the Cameroon Line define three individual sample suites that are characterized by distinct major, trace, and platinum-group element systematics. All three-sample suites display correlations of the PGE with MgO, Ni, and Cr. The new analytical results are employed to constrain the geochemical behavior of the PGE during the formation and differentiation of mantle–derived melts. The PGE are inferred to be compatible in sulfides during partial melting with sulfide-silicate melt partition coefficients of ∼1 × 104. The fractionated PGE patterns of mantle melts are a consequence of the incompatibility of Pd in nonsulfide phases, whereas Ir and Ru must be compatible in at least one other mantle phase. Model calculations indicate that PGE alloys or spinel may be responsible for the higher compatibility of the latter elements during partial melting. It is further demonstrated that the shape of the melting regime has a profound effect on the PGE systematics of mantle magmas. The systematic trends of the three sample suites in plots of PGE against Ni and Cr are the result of magma differentiation processes that involve fractional crystallization of silicate minerals and the concurrent segregation of an immiscible sulfide liquid. The behavior of the PGE during magma fractionation indicates that the segregated sulfides probably equilibrate with >90% of the silicate magma and that PGE scavenging by sulfides is best described by a combination of batch and fractional equilibrium partitioning.

Crustally contaminated komatiites and basalts from Kambalda, Western Australia☆

Archean high-Mg basalts at Kambalda, Western Australia, are geochemically different from associated komatiites and tholeiites. The komatiites and tholeiites are moderately to strongly depleted in incompatible elements [e.g., (LaSm)N = 0.63–0.73, ZrY = 2.4] while the overlying basalts are strongly enriched in these elements [(LaSm)N = 1.2–2.6, ZrY = 2.9–3.9]. Interpretation of the magmatic evolution of the volcanic suite is complicated by element mobility during hydrothermal alteration and metamorphism, and during the formation of felsic ocelli in the basalts. The latter process caused large variations in SiO2, FeOtot and MgO, and lesser variations in other elements. Fractional crystallization also played a role in generating a range of element concentrations and rock types, but does not explain the differences between komatiites and basalts. The major influence on the trace-element characteristics of the volcanic suite was contamination of komatiite magma by crustal rocks. Quantitative modelling suggests that: (a) the lower komatiite was contaminated during eruption by thermal erosion and assimilation of a mixture of sediment and tholeiite; and (b) the compositions of the Kambalda high-Mg basalts result from up to 25% contamination of komatiite, at depth, by material with the composition of modern upper continental crust. More mafic compositions such as those estimated for lower-crustal rocks or Archean continental crust are not appropriate.

Nb-Th-La in komatiites and basalts" constraints on komatiite petrogenesis and mantle evolution

New spark-source mass spectrometric analyses of Nb, Th, REE and other trace elements in Archaean to Tertiary komatiites and basalts were undertaken to test the model of Hofmann et al. [1] of secular variation of the composition of the upper mantle. Most 3.4 Ga komatiites and basalts have Nb/Th between 7 and 9, values that straddle the primitive mantle ratio of ∼ 8. Several 2.7 Ga komatiites also have Nb/Th= 7–9, but most samples of this age have higher Nb/Th, in the range 10–15. Cretaceous-Tertiary komatiites and basalts from Gorgona Island (Colombia) have Nb/Th between 11 and 24, values that approach those of modern oceanic basalts ( ∼ 15–34). Although these results generally support the model of Hofmann et al., there are several complicating factors: (1) most of the Cretaceous-Tertiary Gorgona komatiites have Nb/Th ratios little higher than those of 2.7 Ga komatiites and primitive mantle, which suggests that low Nb/Th may be a peculiarity of komatiites and not a feature of the Archaean mantle; (2) many Archaean komatiites are depleted in both Nb and Th relative to the REE, a feature that is inconsistent with their derivation from primitive mantle. We speculate that komatiites come from a source that evolved independently from normal upper mantle, and that the depletion of Nb and Th resulted from fractionation of an unknown phase during the deep melting. Certain tholeiitic basalts do not show unusual Nb-Th-La fractionation but nonetheless show a secular increase in Nb/Th. This variation may indicate a change in upper mantle compositions resulting from progressive withdrawal of continental crust during the past 3 Ga.