William M. White

The geochemistry of marine sediments, island arc magma genesis, and crust-mantle recycling

To assess the role of sediment subduction and recycling in island arc magma genesis and mantle evolution, we have determined Sr, Nd, and Pb isotope ratios and the concentrations of K, Rb, Cs, Ba, Sr, U, Th, Pb and rare earth elements in 36 modern marine sediments, including Mn nodules, biogenic oozes, and pelagic and hemipelagic clays from the Pacific, Atlantic and Indian Oceans. From these data we draw the following conclusions. Sr and Nd isotope ratios and the Sr/Nd concentration ratios in sediments are such that mixing between subducted sediment on the one hand and depleted mantle or subducted oceanic crust on the other can produce mixing arrays which may pass either through or outside of the oceanic basalt SrNd isotope “mantle array”. Thus whether isotope compositions of island arc volcanics (IAV) plot inside or outside of the mantle array is not a good indication of whether or not their sources contain a subducted sediment component. The presence of subducted sediment in the sources of IAV should lead to Cs/Rb and Pb/Ce ratios which are higher than those in oceanic basalts, and Ba/Rb ratios which may be either higher or lower than oceanic basalts. Simple mixing calculations suggest that as little as a percent or so sediment in island arc magma sources can account for the observed Cs/Rb, Pb/Ce, and Ba/Rb ratios in IAV. However, it does not appear that high Ba/La ratios and negative Ce anomalies in IAV are inherited from sediment in IAV magma sources. It is more likely these features reflect fractionation of alkalis and alkaline earths from rare earths during slab dehydration and metasomatism. Pb isotope ratios in sediments from the Warton Basin south of the Sunda Arc are collinear in 208Pb/204Pb-207Pb/204Pb/206Pb/204Pb space with volcanics from West Sunda, but not with volcanics from the East Sunda. This collinearity is consistent with the hypothesis that sediments similar to these are being subducted to the magma genesis zone of the West Sunda Arc. Sediment recycling to the deep mantle appears capable of explaining much of the Sr and Nd isotopic variation in oceanic basalt magma sources. However, because of the low 238U/204Pb ratios in sediments, anciently recycled sediment should have lower 206Pb/204Pb ratios than most oceanic basalts, though this effect will be in part balanced by high 238U/204Pb in hydrothermally altered oceanic crust. Unless 238U/204Pb ratios in ancient sediments were different than in modern ones, it would appear that deep sediment recycling cannot account for the Pb isotopic compositions of most oceanic basalts. In addition, deep sediment recycling should lead to higher Pb/Ce and Cs/Rb ratios and more variable Ba/Rb ratios in oceanic basalts than is observed. High Pb/Ce ratios have recently been observed in young seamounts from the Society Island chain, suggesting their source contains a recycled component. However, on the whole, the limited variation of Pb/Ce, Ba/Rb, and Ca/Rb in oceanic basalts suggests recycling of sediment of the deep mantle is limited.

High-precision analysis of Pb isotope ratios by multi-collector ICP-MS

Abstract We investigated high-precision Pb isotope ratio analysis by multi-collector-inductively coupled plasma-mass spectrometry (MC-ICP-MS) using added thallium as an internal isotopic standard to correct for mass dependent isotopic fractionation. We compared MC-ICP-MS analysis of both an inter-laboratory standard, NBS 981, and geological samples to conventional thermal ionization mass spectrometry (TIMS). As expected, we found that analytical error in the latter was dominated by mass fractionation. In MC-ICP-MS, we found that fractionation appears to follow the exponential law, but that the fractionation coefficients, f, of Tl and Pb were not identical. This difference in fractionation coefficients cannot be compensated for by renormalizing to a different Tl isotopic composition as done in other studies. We found, however, that the fPb/fTl ratio was constant over the course of an analytical session, allowing fPb to be calculated from fTl. An exponential law correction was then applied to the Pb isotope measurements which effectively eliminates errors associated with mass fractionation. Precision for the MC-ICP-MS analyses ranged from a factor of 2 to a factor of 6 better than for TIMS analyses for the 206 Pb / 204 Pb and 208 Pb / 206 Pb ratios respectively. Residual error in the MC-ICP-MS analyses was dominated by error in the analysis of 204 Pb , perhaps in part due to random errors introduced by correcting for a 204 Hg isobaric interference. We also found systematic errors in the MC-ICP-MS analyses compared to TIMS determinations that may be due to uneven background and collector biases in the instrument used. We found that these systematic errors were the same for both NBS 981 and the geological samples, so accurate correction factors could be generated from the standard analyses to correct the sample analyses. MC-ICP-MS has the additional advantages of requiring less preparative chemistry, less instrument time, and considerably less labor overall.

Petrology and geochemistry of the Galápagos Islands: Portrait of a pathological mantle plume

motion of the Nazca plate with respect to the fixed hotspot reference frame. lsotope ratios in the Galfipagos display a considerable range, from values typical of mid-ocean ridge basalt on Genovesa (87Sr/86Sr: 0.70259, end: +9.4, 206pb/204pb: ! 8.44), to typical oceanic island values on Floreana (87Sr/86Sr: 0.70366, œNd: +5.2, 206pb/204pb: 20.0). La/Sm N ranges from 0.45 to 6.7; other incompatible element abundances and ratios show comparable ranges. Isotope and incompatible element ratios define a horseshoe pattern with the most depleted signatures in the center of the Galfipagos Archipelago and the more enriched signatures on the eastern, northern, and southern periphery. These isotope and incompatible element patterns appear to reflect thermal entrainment of asthenosphere by the Galfipagos plume as it experiences velocity shear in the uppermost asthenosphere. Both north-south heterogeneity within the plume itself and regional variations in degree and depth of melting also affect magma compositions. Rare earth systematics indicate that melting beneath the Galfipagos begins in the garnet peridotite stability field, except beneath the southern islands, where melting may occur entirely in the spinel peridotite stability field. The greatest degree of melting occurs beneath the central western volcanos and decreases both to the east and to the north and south. Sis. 0, FeB. 0, and NaB. 0 values are generally consistent with these inferences. This suggests that interaction between the plume and surrounding asthenosphere results in significant cooling of the plume. Superimposed on this thermal pattern produced by plume-asthenosphere interaction is a tendency for melting to be less extensive and to occur at shallower depths to the south, presumably reflecting a decrease in ambient asthenospheric temperatures away from the Galfipagos Spreading Center.

Hf isotope constraints on mantle evolution

Abstract The similarity of the Lu–Hf and Sm–Nd isotope system during most mantle differentiation processes makes the combination of 176Hf/177Hf and 143Nd/144Nd a very sensitive indicator of a select number of processes. This paper present new Hf-isotope data for a large number of ocean islands and examines the Hf–Nd–Pb isotope relations of oceanic volcanics. Except for HIMU islands, St. Helena and Tubaii, the Hf and Nd isotope ratios in ocean island basalts (OIB) are extremely well correlated. It is argued that crustal recycling (by either continental or oceanic sediments) most likely did not cause the Hf–Nd variations. The correlated 176Hf/177Hf–143Nd/144Nd variations in OIB most likely represent the time integrated fractionations which are the result of melting in the presence of garnet. The Hf-isotope systematics of HIMU-type OIB are consistent with these basalts representing recycled oceanic crust and thus support the earlier hypothesis on the origins of HIMU basalts. Chondrites form an array that is at high angle with the OIB array. This allows a choice in the 143Nd/144Nd and 176Hf/177Hf values for chondritic bulk earth. With a choice of bulk earth at the extreme end of the OIB array the shift of OIB to higher 176Hf/177Hf can be explained by either isolation of a significant amount of basalts from the mantle for several billions of year or by fractionation and isolation of small amounts (

238U/204Pb in MORB and open system evolution of the depleted mantle

Concentrations of U, Th and Pb were determined by isotope dilution on 30 mid-ocean ridge basalt (MORB) glasses. The mean values of238U/204Pb (μ) and232Th/238U (κ) in these glasses are 11.20 and 2.67 respectively. There are significant differences in these ratios between ocean basins. Mean κ values increase in the order Atlantic < Pacific < Indian; mean μ is higher in Pacific MORB than in Indian and Atlantic MORB, but differences between the Atlantic and Indian are not significant. μ correlates strongly with U concentration, but not with Pb concentration or with206Pb/204Pb. The correlation of μ with U but not with206Pb/204Pb reflects the effect of partial melting and fractional crystallization on μ. Since the maximum value of U in the depleted mantle must be less than the bulk silicate earth value of 0.018 ppm, the μ-U correlation can be used to estimate the maximum mean value of μ in the depleted mantle (μDM). This yields a maximumμDM of 4.67. Similarly, the correlation between μ andRb/Sr suggests a maximumμDM of 6.3. These results are consistent with values forμDM derived using mass balance from the crustal and depleted mantle values of κ, the depleted mantlePb/Ce ratio, and the crustal abundance of Ce. The value ofμDM thus seems to be less than the bulk earth value, estimated to be8 ± 2. This is consistent with the incompatible-element-depleted character of this reservoir, but inconsistent with Pb isotope ratios, which record time-integrated values for μ of 8–9, and which indicate that μ has increased over time. This inconsistency requires that the depleted mantle be an open system with a residence time for Pb of1 ± 0.5 Ga. Neither the continental crust nor a primitive lower mantle have the appropriate isotopic composition to be the source of Pb in the depleted mantle. Mantle plumes do have the appropriate isotopic composition, and as little as 10% of the observed plume flux is needed to mix with the depleted mantle to supply the required flux of Pb.

Strontium, neodymium, and lead isotopic and trace-element signatures of the East indonesian sediments: provenance and implications for banda arc magma genesis

We present new trace-element and Sr-Nd-Pb isotope data for 127 surface sediments and five sediments from DSDP Site 262, distributed along and across the arc-continent collision region of the Banda Arc, East Indonesia. The results are used to evaluate the role of subducted continental material (SCM) in the genesis of the Banda Arc magmas and to assess the extent to which geochemical and isotopic signatures of SCM are controlled by sediment provenance. In the surface sediments lead and neodymium isotope ratios are variable: 206Pb/204Pb = 18.65–19.57; 143Nd/144Nd = 0.51230–0.51190, with an increase in lead isotope ratios and a decrease in the 143Nd/144Nd ratio from northeast to southwest along the Banda Arc. DSDP Site 262 sediments, farthest to the west in the Timor Trough, overlap with the surface sediments and have 206Pb/204Pb = 18.89–19.23 and 143Nd/144Nd = 0.51200–0.51220. In contrast, the trace-element ratios and REE patterns of the sediments do not show systematic along-arc variations and largely overlap with estimated values for Upper Continental Crust, Post Archean Australian Shale (PAAS), and ODP Site 765 sediments from the Argo Abyssal Plain. From the combined isotopic and trace-element ratios in the terrigenous fraction of the sediments four major provenance areas can be distinguished: North New Guinea + Seram, South New Guinea, Timor, and northern Australia. The lead isotopic variations in the shelf and wedge sediments along the Banda Arc are parallel to similar variations in the volcanics; this is considered to be strong evidence for the incorporation of subducted continental material in the arc magmas. The trace-element characteristics of both the volcanics and the sediments are also consistent with the involvement of sediments in the Banda Arc magma genesis. The hinterland of the sediments is responsible for isotopic signatures created in the Banda Arc mantle through recent subduction. This suggests that some of the mantle heterogeneities that are inferred from oceanic basalts can be explained by differences in the provenance of (ancient) subducted sediment.

Sr‐Nd‐Pb isotope systematics of the Banda Arc, Indonesia: Combined subduction and assimilation of continental material

We present Sr, Nd, and Pb isotope results and SiO2, Rb, Sr, Sm, Nd, U, Th, and Pb data for six active volcanoes and one extinct volcanic island distributed over the whole length of the Banda Arc. Rock types range from low-K tholeiitic in the NE to high-K calc-alkaline in the SW. The volcanoes in the NE have “normal” arc signatures (87Sr/86Sr = 0.7045–0.7055, 143Nd/144Nd = 0.51273–0.51291, and 206Pb/204Pb = 18.66–18.75), whereas those in the SW have extreme values (87Sr/86Sr = 0.7065–0.7083, 143Nd/144Nd = 0.51252–0.51267, and 206Pb/204Pb = 19.28–19.43). Serua, situated in the central part, is the most anomalous volcano with regard to its Sr and Nd isotopic composition (87Sr/86Sr = 0.7075–0.7095 and 143Nd/144Nd = 0.51240–0.51260) but not with regard to Pb isotopes (206Pb/204Pb = 19.02–19.08). The inactive island of Romang in the SW overlaps the Serua trends. The volcanoes display variable within-suite ranges in 87Sr/86Sr and 143Nd/144Nd. Large ranges (e.g., at Nila) are consistent with assimilation (10–20%) of carbonate-bearing sediments from the arc crust. Despite the evidence for assimilation, it cannot explain all of the Sr-Nd isotopic trends found, and Banda Arc magmas must have already obtained a “continental” signature at depth before they reached the arc crust. Within-suite trends of Pb isotopes are virtually absent. We found an extreme range in the volcanics along the arc which coincide with a similar trend in sediments in front of the arc and consider this as strong evidence for the contribution of subducted continent-derived material to magma sources. Bulk addition of 0.1–2% of local sediment in the NE Banda Arc, and of 1–3% in the SW Banda Arc, to an Indian Ocean mid-ocean ridge basalt (I-MORB) source can explain the isotopic trends; both Serua and Romang require > 5% sediment. The Pb isotopes (e.g., 207Pb/204Pb - 208Pb/204Pb) also suggest changes in the mantle end-member from I-MORB to oceanic island basalt (OIB) source type. The latter becomes more conspicuous toward the SW and has the high 208Pb/ 204Pb characteristic of Indian Ocean (Dupal) OIBs. We hypothesize that mixing of magmas in the mantle wedge and/or in the arc crust was an important mechanism by which mantle and subducted end-members were incorporated in the final products.

Sources of error in external calibration ICP-MS analysis of geological samples and an improved non-linear drift correction procedure

The primary factors limiting accuracy and precision using inductively coupled plasma mass spectrometry (ICP-MS) in matrix-matched external standardization are machine drift and variation of the instrument response as a function of mass. Because drift is usually non-linear, the degree of drift differs from one mass to the next, and the direction of drift can change frequently when analyzing over large mass ranges. Internal standardization results in minimal improvement of data quality. An analytical procedure and an off-line data reduction algorithm have been developed that correct for these variations and produce a significant improvement in analytical accuracy and precision. In this technique, a “drift correction” standard is analyzed after every four or five samples. A polynomial curve is fitted to each isotope analyzed, and a correction based on this curve is applied to the measured intensity of the respective isotopes in the samples and standards. This data reduction algorithm has been developed into a Microsoft Excel™ 3.0 Macro that completely automates all calculations. This article is an electronic publication in Spectrochimica Acta Electronica (SAE), the electronic section of Spectrochimica Acta Part B (SAB). The hard copy text is accompanied by a disk with the Excel macro for the Macintosh computer and sufficient instructions for its use. The main article discusses the scientific aspects of the subject and explains the purpose of the macro.

Subalkaline andesite from Valu Fa Ridge, a back-arc spreading center in southern Lau Basin: petrogenesis, comparative chemistry, and tectonic implications

Tholeiitic andesite was dredged from two sites on Valu Fa Ridge (VFR), a back-arc spreading center in Lau Basin. Valu Fa Ridge, at least 200 km long, is located 40–50 km west of the active Tofua Volcanic Arc (TVA) axis and lies about 150 km above the subducted oceanic plate. One or more magma chambers, traced discontinuously for about 100 km along the ridge axis, lie 3–4 km beneath the ridge. The mostly aphyric and glassy lavas had high volatile contents, as shown by the abundance and large sizes of vesicles. An extensive fractionation history is inferred from the high SiO2 contents and FeO∗MgO ratios. Chemical data show that the VFR lavas have both volcanic arc and back-arc basin affinities. The volcanic arc characteristics are: (1) relatively high abundances of most alkali and alkaline earth elements; (2) low abundances of high field strength elements Nb and Ta; (3) high U/Th ratios; (4) similar radiogenic isotope ratios in VFR and TVA lavas, in particular the enrichment of 87Sr86Sr relative to 206Pb204Pb; (5) high 238U230Th, 230Th232Th, and 226Ra230Th activity ratios; and (6) high ratios of Rb/Cs, Ba/Nb, and Ba/La. Other chemical characteristics suggest that the VFR lavas are related to MORB-type back-arc basin lavas. For example, VFR lavas have (1) lower 87Sr86Sr ratios and higher 143Nd144Nd ratios than most lavas from the TVA, except samples from Ata Island, and are similar to many Lau Basin lavas; (2) lower Sr/REE, Rb/Zr, and Ba/Zr ratios than in arc lavas; and (3) higher Ti, Fe, and V, and higher Ti/V ratios than arc lavas generally and TVA lavas specifically. Most characteristics of VFR lavas can be explained by mixing depleted mantle with either small amounts of sediment and fluids from the subducting slab and/or an older fragment of volcanic arc lithosphere. The eruption of subalkaline andesite with some arc affinities along a back-arc spreading ridge is not unique. Collision of the Louisville and Tonga ridges probably activated back-arc extension that ultimately led to the creation and growth of Valu Fa Ridge. Some ophiolitic fragments in circum-Pacific and circum-Tethyan allochthonous terranes, presently interpreted to have originated in volcanic arcs, may instead be fragments of lithosphere that formed during early stages of seafloor spreading in a back-arc basin.

Beyond EM-1: Lavas from Afanasy-Nikitin Rise and the Crozet Archipelago, Indian Ocean

Lavas from Afanasy-Nikitin Rise, possibly the Late Cretaceous product of the Crozet hotspot, cover a wide range of isotopic compositions that includes the lowest ( 206 Pb 204 Pb) t (to 16.77) and ϵ Nd ( t ) (to −8) values yet found among oceanic islands or spreading centers worldwide, as well as high ( 87 Sr/ 86 Sr) t (to 0.7066). In contrast, young basalts from the Crozet Archipelago exhibit a narrow range of variation around ϵ Nd ∼ +4, 87 Sr/ 86 Sr ∼ 0.7040, and 206 Pb/ 204 Pb ∼ 19.0, closely resembling that of shield lavas of the Reunion hotspot. The Afanasy-Nikitin rocks also have much higher Ba/Nb, Ba/Th, and Pb/Ce than modern oceanic island or ridge lavas, as well as high La/Nb. The data do not obviously support the Crozet plume model but, assuming the model to be plate tectonically correct, would indicate that the plume-source composition either changed dramatically or that Afanasy-Nikitin magmatism involved significant amounts of nonplume mantle. The low 206 Pb/ 204 Pb, low ϵ Nd lavas provide the best evidence to date of the sort of material that, by variably contaminating much of the Indian mid-ocean-ridge basalt (MORB) source asthenosphere, may be responsible for the isotopic difference between most Indian MORB and Pacific or North Atlantic MORB. The combined isotopic and trace element results suggest an ultimate origin in the continental crust or mantle lithosphere for this material, although whether it was cycled through the deep mantle or resided at shallow levels in the convecting mantle cannot currently be determined.