C. Chauvel

HIMU−EM : the French Polynesian connection

Abstract himu, em i and em ii are three of the main geochemical mantle components that give rise to oceanic island basalts [1]. They represent the end members that produce the extreme isotopic compositions measured on intraplate volcanics. In French Polynesia, all three mantle components are represented in volcanic rocks. The characteristic himu signature is found in Tubuai, Mangaia and Rurutu, em i is present in the source of Rarotonga and Pitcairn volcanics and em ii dominates the composition of most Society Islands. Intermediate values between the three end members are found on most islands. We suggest that the three components are not independent but are physically related in the mantle. The himu component is thought to be recycled oceanic crust that lost part of its Pb through hydrothermal processes prior to and during subduction. em i and em ii are believed to acquire their isotopic and trace element characteristics through entrainment of sediments that were subducted together with the oceanic crust. The trace element pattern and the isotopic composition of himu lavas can be quantitatively modelled using a mixture of ∼ 25% old recycled morb crust and 75% mantle peridotite. The extreme Pb composition is modelled assuming that Pb was lost from oceanic crust when hydrothermal alteration at the ridge leached Pb from the basalt to redeposit it as sulphides on top of and throughout the crust, followed by preferential dissolution of sulphides during dehydration in the subduction zone. These processes led to a drastic increase of the U/Pb ratio of the subducted material which evolved over 2 Ga to very radiogenic Pb isotopic compositions. Pb isotopic compositions similar to those of em i and em ii are modelled assuming that sediments with average crustal Pb isotopic compositions were subducted and recycled into the mantle together with the underlying morb oceanic crust. Pelagic sediments ( μ ∼ 5 and κ ∼ 6 ) account for the Pb isotopic composition of em i whereas terrigenous sediments ( μ ∼ 10 and κ ∼ 4.5 ) evolve towards the em ii end member. A few percent of sediment in the recycled crust-sediment mixture will destroy the characteristic Pb isotopic signature of the himu component. This, together with the low probability of isolating oceanic crust in the mantle for ⩾ 2 Ga, explains why the extreme himu composition, as seen on Tubuai and St Helena, is sampled so rarely by oceanic volcanism.

The Sm-Nd age of Kambalda volcanics is 500 Ma too old!

Although Sm-Nd isotopic analyses of volcanics from Kambalda, Western Australia, form a 3.2 Ga linear array, Pb−Pb analyses of the same suite give an isochron age of 2.73±0.03 Ga and a 2.7 Ga model age for the associated sulphides. We suggest that the 3.2 Ga age is incorrect and that the Sm-Nd array results from mixing between depleted mantle and either older continental crust or enriched mantle.

Source compositions and melting processes in the Society and Austral plumes (South Pacific Ocean): Element and isotope (Sr, Nd, Pb, Th) geochemistry

We present comprehensive geochemical analyses on samples from the active volcanoes of the Society and Austral hotspot chains, including data for major, trace and rare-earth elements, and Sr, Nd, Pb and Th isotopes. The latter can be used to determine the Th/U in the source at the time of melting, and so give a constraint on the absolute amount of incompatible-element fractionation occurring during melting. n nSiO2 vs. MgO variations show evidence for variable amounts of a nephelinitic melt component (low SiO2, low MgO) in all the magmas studied. The nephelinite is probably produced in the presence of CO2 during melting. Correlations between SiO2 and the degree of Thue5f8U fractionation (derived from Th isotope measurements) imply that the CO2-driven, nephelinitic melting is also responsible for fractionating the Th/U ratio. n nComparing the Th isotopes with Sr, Nd and Pb isotopes, it is possible to place limits on the Th/U ratios in the EM II (Societies) and HIMU (Australs) sources. These are ∼ 3.4 and ⩽ 2.5, respectively. n nThe Nb/U ratio, which was previously thought to be relatively constant in all oceanic volcanics (47 ± 10), is shown to be anomalously low (25 ± 5) in some of the Society Seamounts and high (60) in some of the Australs. This confirms the presence of continent-derived material in the Society source and suggests that HIMU volcanics may have elevated Nb/U ratios. n nThe change in Pb isotope compositions in the Australs, from HIMU (e.g., Tubuai) before 6 Ma to the more recent, less extreme compositions seen for example at Macdonald, is the result of a small amount of subducted sediment contaminating the pure old recycled oceanic crust component which, before 6 Ma, yielded HIMU.

HYDRATION AND DEHYDRATION OF OCEANIC CRUST CONTROLS PB EVOLUTION IN THE MANTLE

Abstract The position of the Pb isotopic compositions of mid-ocean ridge basalts (MORB) to the right of the geochron has long puzzled geochemists. Uranium is more incompatible than Pb during mantle melting, and the mantle source of MORB, being depleted in incompatible elements, should have low Pb isotope ratios and plot to the left of the geochron. This has been called the “lead paradox”. MORB have a second peculiar characteristic: Pb concentrations are drastically depleted relative to other elements of similar compatibility such as Ce or Nd. We suggest that both characteristics can be explained by preferential mobility of Pb during hydrothermal alteration of the oceanic crust associated with sea-floor spreading, and subsequent dehydration during subduction. A portion of this Pb migrates into the mantle wedge and is then added to the continental crust via arc magmas. Parts of the Pb-depleted recycled oceanic crust are stored at some deep level in the mantle and eventually become the source of ocean island basalts, the rest mixes into the mantle to become the MORB source. This model is evaluated quantitatively and special attention is given to the evolution of the Ce/Pb ratio in the depleted mantle from the beginning of Earth history to the present.

Pb and Nd isotopic study of two archean komatiitic flows from Alexo, Ontario

Isotopic analysis of two Archean komatiitic flows from Alexo, Ontario, gives a Pb-Pb isochron age of 2690 ± 15 Ma and a Sm-Nd isochron age of 2752 ± 87 Ma. These ages agree well with U-Pb zircon ages from underlying and overlying volcanics. The variations in element ratios that define the isochrons were not produced during crystallization of the lavas. The spread in U/Pb was caused by submarine alteration soon after eruption, and the spread in Sm/Nd resulted from (a) differences in the composition of the residue of melting, and (b) contamination of the upper komatiite flow through thermal erosion of the lower flow. n nThe 147Sm/144Nd ratio of uncontaminated komatiite is 0.25 which reflects the depleted nature of its mantle source. The Th/U ratio of about 3.4 is probably also representative of depleted mantle. The initial ϵNd of +2.44 ± 0.51 indicates that the mantle depletion took place long before magma formation.

Age of the Archean Talga-Talga Subgroup, Pilbara Block, Western Australia, and early evolution of the mantle: new SmNd isotopic evidence

Archean komatiites, high-Mg basalts and tholeiites from the North Star Basalt and the Mount Ada Basalt formations of the Talga-Talga Subgroup, Warrawoona Group, Pilbara Block, Western Australia, define a linear correlation on the normal143Nd/144Nd vs.147Sm/144Nd isochron plot. The data give an age of 3712 ± 98 Ma and initialeNd(T) of +1.64 ± 0.40. The 3712 ± 98 Ma date is consistent with the regional stratigraphic sequence and available age data and the SmNd linear array may be interpreted as an isochron giving the eruption age of the Talga-Talga Subgroup. An alternative interpretation is that the isochron represents a mixing line giving a pre-volcanism age for the Subgroup. Consideration of geochemical and isotopic data indicates that the true eruptive age of the Talga-Talga Subgroup is possibly closer to about 3500 Ma. Regardless of the age interpretation, the new Nd isotopic data support an existence of ancient LREE-depleted reservoirs in the early Archean mantle, and further suggest that source regions for the Pilbara volcanic rocks were isotopically heterogeneous, witheNd(T) values ranging from at least 0 to +4.0.

Origin of basalts from the Marquesas Archipelago (south central Pacific Ocean) : isotope and trace element constraints

Basalts from the Marquesas Archipelago display significant variations according to magmatic type in 143Nd/144Nd (0.512710–0.512925) and 87Sr/86Sr (0.70288–0.70561) suggesting heterogeneities at various scales in the mantle source, with respectively the highest and lowest values in tholeiites compared to alkali basalts. This relationship is the reverse from that observed in the Hawaiian islands. Systematic indications of magma mixing are recognized from the relationships between trace element and isotopic ratios. Tholeiites from Ua Pou Island which have unradiogenic Sr (about 0.7028) plot close to basalts from Tubuai and St. Helena, i.e. distinctly below the main mantle trend in the Nd vs. Sr isotopic diagram. It is suggested that the source of these tholeiites is ancient subducted lithosphere which has suffered previous extraction of liquid with island arc tholeiite composition. The trace element and isotopic data of the basalts from the other Marquesas Islands imply the contamination of an equivalent source by an enriched component. This latter has trace element characteristics of the upper crust.

Sample contamination explains the Pb isotopic composition of some Rurutu island and Sasha seamount basalts

Abstract Palacz and Saunders [1] reported Pb isotopic compositions for Rurutu island basalts that form a trend oblique to the main oceanic basalt array and interpreted them as representing a mixture of HIMU and DUPAL (or EM II) mantle sources. Detailed leaching experiments on aliquots of the same rock powders as those measured by Palacz and Saunders [1] demonstrate that these powders have been heavily contaminated by a foreign Pb component. Two contaminants with different Pb isotopic compositions are identified. They represent together more than 80% of the total Pb present in the sample. The residues display a single coherent Pb isotopic trend consistent with more recent measurements [2,3] of Rurutu island basalts and with the main oceanic basalt array. Pb isotopic compositions similar to those of Rurutu basalts [1] were reported by Fornari et al. [4,5] for Lamont seamount basalts. They interpreted them as reflecting a Rapa or Rurutu-type (EM II) hotspot influence in their sources. Strong leaching of glass chips from one of the Fornari et al. [4,5] samples demonstrates that these glass chips were contaminated by a foreign Pb component similar to one of the Rurutu contaminants. In both of these cases the contamination of the samples occurred, most likely, prior to the initiation of isotope analyses. These findings highlight the importance of careful sample preparation procedures and we suggest using acid-washed rock chips for even the freshest looking lavas for Pb isotopic analyses.

Aluminum depletion in komatiites and garnet fractionation in the early Archean mantle: Hafnium isotopic constraints☆

Hafnium isotopic compositions were measured in Al-depleted and Al-enriched komatiites from the 3450 Ma old Barberton greenstone belt, southern Africa. All samples have initial ϵHf values close to zero. Such values are at variance with the strongly negative or positive values that should be observed if these rocks came from old garnet-depleted or garnet-enriched layers, such as may have formed during the solidification of an ancient terrestrial magma ocean. The garnet fractionation observed in komatiites probably took place during the melting event.

Anomalous Sm−Nd ages for the early Archean Onverwacht Group Volcanics

New Sm−Nd isotopic data for eight samples of basalt and komatiite from the Tjakastad Subgroup (lower Onverwacht Group) of the Barberton Greenstone Belt (BGB) of the Kaapvaal craton in southern Africa are reported. They give new constraints on the interpretation of Sm−Nd ages for the Subgroup and highlight the petrogenesis of Tjakastad volcanics. Although Sm−Nd isotopic data earlier reported for volcanic rocks from the Tjakastad Subgroup yielded an “isochron” age of 3526±48 Ma, the new results give a much younger “isochron” date of 3269±84 Ma. The 3526±48 Ma “isochron” age has been obtained in combining samples ranging in composition from felsic volcanics to ultrabasic komatiites and is thus considered suspect with regard to the pre-requisite of geochronology that all the studied rocks must have had identical initial isotopic compositions. The new “isochron” date of 3269±84 Ma has been obtained in combining samples solely of basic/ultrabasic composition. It might thus represent a more correct age for the eruption of the Tjakastad Subgroup volcanism. In fact, owing to the potential problem of source heterogeneity and also in the light of geochronological and geochemical arguments we show that this date also has little chance to have any strict chronological meaning. Most likely, the Tjakastad volcanics were formed 3450 Ma ago. Also most likely, their source rocks were isotopically heterogeneous and the 3530 Ma and 3270 Ma linear arrays are not true but apparent “isochrons”. Based on the calculated ɛNd (3450) values and other geochemical arguments, we show that three possible sources might have been involved: depleted mantle, primitive mantle and older continental crust.