V32A-01 INVITED
Constraints from Li isotope systematics on subduction recycling, arc magmatism, and continent growth: An overview
Great expectations that Li isotopic systematics can uniquely constrain many fluid-mitigated geologic processes have met with mixed success for a variety of reasons. On a local scale (some volcanic arc segments) Li composition can be highly correlated with other geochemical tracers of subduction fluids whereas, globally, such correlations tend to be disappointingly poor. The utility of Li isotopes as a tracer is limited in part by extensive overlap between mantle and subduction inputs, by limited understanding of equilibrium isotopic fractionation effects, and by apparent departures from equilibrium behavior. On the other hand, Li elemental systematics provide important constraints on global recycling processes because major litho-tectonic reservoirs have distinctive enrichments or depletions with respect to Nb or other HFSEs. Such chemical fractionations can be understood in terms of differential solubility of these elements in aqueous fluids vs. silicate melts, as well as the roles of weathering, dehydration, metamorphic or melting processes. For example, arc lavas are are systematically enriched in Li compared to those from other settings and typically have Li/Nb greater than BSE (consistent with addition of Li-rich fluids to their sources). In contrast, bulk continental crust and orogenic granitoids tend to have lower Li/Nb than BSE or arc lavas. Moreover, mass balance implies that the residual mantle (DM) produced by segregation of crust has higher Li/Nb than BSE. However, if continental crust is ultimately derived by subduction related magmatism, high Li/Nb would be expected for the crust and low Li/Nb for the upper mantle. This interesting conundrum is easiest explained in terms of selective Li removal from crustal protolith rocks via chemical weathering and erosion, which also is consistent with Li isotopic compositions of crust, mantle and seawater reservoirs. Thus, Li elemental and isotopic systematics (and relevant proxies) provide complementary constraints on crust-mantle evolution.
V32A-02
Lithium isotope variations in lavas and olivine phenocrysts from the Cook-Austral Islands: Constraints on sample alteration and the true Li-isotope signature of HIMU mantle
Low-temperature fluid/rock reactions can significantly fractionate Li-isotopes. Altered oceanic crust typically has elevated δ7Li values compared with fresh MORB. However, the Li-isotope composition of recycled oceanic crust is uncertain. Whereas some studies have suggested that slab dehydration during subduction generates dehydrated eclogites characterized by anomalously light δ7Li values [1], other studies have argued that slab dehydration does not completely remove the heavy signature of subducted altered crust [2]. Trace element and isotopic signatures of HIMU lavas suggest they are derived from mantle containing recycled, dehydrated oceanic crust. Therefore, the Li-isotope composition of HIMU mantle may help constrain Li recycling during slab subduction. Previous studies have reported anomalously heavy δ7Li values for HIMU lavas. We report new Li-isotope results for both basalt wholerocks and olivine phenocrysts to better constrain the Li-isotope composition of HIMU mantle and to evaluate the affects of weathering and alteration on Li-isotopes in non-zero-age basaltic lavas. Li isotopic compositions of lavas from the Cook-Austral islands vary from 2.8 to 11.4‰, spanning a wider range than Hawaiian shield lavas and extending to values higher than typical fresh MORB. However, comparison of whole rock and olivine phenocryst compositions suggests that much (but not all) of the observed variability is due to post-magmatic alteration. Because of the susceptibility of Li-isotopes in basaltic lavas to sample alteration, great care must be exercised in interpreting Li-isotope heterogeneity in non- pristine lavas. The heaviest δ7Li values observed in whole rock samples of HIMU lavas are most likely an artifact of sample alteration, and not a true magmatic feature. Most olivine phenocrysts span a much narrower range in Li-isotope composition, and appear to be less susceptible to post-magmatic alteration than bulk lavas. However, anomalously light δ7Li in one olivine separate suggests that olivine phenocrysts may be affected by kinetic affects such as diffusive influx of 6Li during magma cooling. However, the majority of olivine/whole rock pairs are within error, confirming a lack of olivine/melt isotopic fractionation and the ability of olivines to robustly record magmatic δ7Li values. Olivine δ7Li values correlate broadly with whole rock radiogenic isotopes (Pb, Nd, Hf). HIMU lavas are characterized by moderately elevated δ7Li (up to 6.2‰), though not as high as previously proposed for HIMU mantle. The moderately elevated δ7Li values are consistent with derivation of HIMU lavas from a source that contains dehydrated recycled oceanic crust that has not suffered severe loss of Li during dehydration. This study suggests that recycled oceanic crust does not completely lose its heavy Li-isotope fingerprint upon subduction into the mantle, consistent with the model of [2], but in contradiction to the model of [1] for Li behavior during slab dehydration. [1] Zack, et al., EPSL, 208, 279-290, 2003. [2] Marshall, et al., Geochim. Cosmochim. Acta, 70, 4750-4769, 2006.
V32A-03
Isotopic fractionation of Li during cooling of mantle peridotite from Gakkel Ridge
Lithium isotopic compositions for mineral separates of coexisting olivine, clinopyroxene, orthopyroxene and bulk rocks of fresh Gakkel Ridge peridotites were analyzed by MC-ICPMS. Major and trace element compositions of the component minerals were measured in-situ with electron microprobe and LA-ICPMS. These rocks are absolutely fresh, with <0.1% modal serpentine. Bulk rocks have lithium contents of 1.6 to 2.7 ppm and δ7Li values of 3 to 5‰, which fall within the range of reported normal "MORB mantle" values. Lithium concentrations vary on the order of cpx (2.1 to 4.7 ppm) > opx (0.9 to 1.7 ppm) ≥ olivine (0.4 to 0.9 ppm), which reveals an inverse order of those found in "equilibrated" mantle peridotites (Seitz and Woodland, 2000). The lithium isotopic compositions indicate a systematic mineral variation with δ7Liolivine (7.14‰ to 15.09‰) > δ7Liopx (1.81‰ to 3.66‰) > δ7Licpx (-2.43‰ to -0.39‰). The δ7Li values of cpx are negatively correlated with their lithium concentrations with the lightest value for the most enriched cpx grains. There is a first order negative linear correlation between Δolivine-cpx (δ7Liolivine- δ7Licpx) and ol/cpxD (Liolivine/Licpx). Collectively, these observations suggest that these inter-mineral fractionations may be related to a re-distribution of Li between minerals as a result of sub-solidus cooling driven by temperature or pressure sensitive inter-mineral partitioning. Li6 can diffuse much faster than Li7 does owing to the large relative mass difference, which results a large isotopic fractionation by diffusion along with the re-distribution during cooling. Alternatively, Li exchange via melt-rock reaction may have occurred. However, the systematic correlation between inter- mineral lithium partitioning and inter-mineral isotopic fractionation together with the unaltered nature of the bulk rock δ7Li values argues against the presence of significant melt-rock reaction, unless the reaction did not change the bulk lithium systematics.
V32A-04
Diffusion of Li in Olivine: Complex Behavior Arising From Effects of Li Concentration and Defect Chemistry
An increasing number of studies are using Li-concentration and Li-isotope systematics to address various geochemical and cosmochemical questions. Large variations are observed [e.g. 1-3] but the origins of the fractionation (e.g. deep and old versus shallow and recent) are debated. Notably, in contrast to the behavior of heavier elements, significant kinetic fractionation e.g. during diffusion, is observed for Li [4]. To contribute to this discussion we have built upon our existing dataset of experimental Li diffusion in olivine [5]. We have determined diffusion rates for both isotopes of Li (6Li and 7Li) along [001] in olivine single crystals at 800 to 1200 °C, 1 atmosphere total pressure and fO2 ~ WM buffer. It was found [5] that (i) diffusion of Li is fast compared to diffusion rates of Fe-Mg and other cations, (ii) diffusion of Li occurs simultaneously by two mechanisms whereby Li occupies vacant metal sites as well as interstitial sites, with dynamic exchange between these. Our further studies show, however, that the diffusion rates depend dramatically on the concentration of Li in olivine as well as the concentration of available point defects (e.g. vacancies). To account for this dependence we have extended our model and explicitly considered charge balance during incorporation and diffusion of Li in olivine based on a quantitative point defect model [6]. The observed concentration profiles of Li as well as its isotopes are well described by the model. Our main conclusions are: (i) 6Li diffuses about 4-6 % faster than 7Li on the interstitial site, (ii) to model homogenization timescales in natural systems, it is necessary to consider the full coupling between point defect concentrations (as functions of parameters such as fO2, fH2O) and Li concentration, as embodied in the model, and (iii) for concentrations of Li typically found in natural olivines (~ few ppm), the faster, interstitial diffusion mechanism is effectively absent and the modeling may be simplified. In these cases, Li diffusion rates are about an order of magnitude faster than rates of Fe-Mg diffusion in olivine; in special circumstances these two rates may even be similar. Under conditions that promote high concentrations of Li interstitials in olivine (e.g. presence of a Li rich fluid), isotopic equilibration of Li will proceed much faster than the rate at which Li concentrations equilibrate, but this need not always be the case in nature. Finally, Li diffusion rates in olivine are surprisingly slow in comparison to the rates in other minerals such as plagioclase [7] and clinopyroxene [8]. References [1] Tomascak PB (2004) Rev Miner Geochem, 55, 153-195. [2] Rudnick RL, Ionov DA (2007) Earth Planet Sci Lett, 256, 278-293. [3] Barrat JA, Chaussidon M, Bohn M, Gillet Ph, Göpel C, Lesourd M (2005) Geochim Cosmochim Acta, 69, 5597-5609. [4] Richter FM, Dauphasa N, and Teng F-Z (2008) Chem Geol, doi:10.1016/j.chemgeo.2008.06.011. [5] Dohmen R, Kasemann S, Coogan L, Chakraborty S, Fall AGU 2007 [6] Dohmen R, Chakraborty S (2007) Phys Chem Mineral, 34, 409-430. [7] Giletti BJ, Shanahan TM (1997) Chem Geol, 139, 3-20. [8] Coogan LA, Kasemann SA, Chakraborty S (2005) Earth Planet Sci Lett, 240, 415-424.
V32A-05 INVITED
The Behavior of Li in Subduction Zones with Implications for Fluid Cycling
The chemical and isotopic compositions of pore fluids provide important insights on fluid-rock diagenetic or metamorphic reactions, hence, on the subsurface hydrology. Li is one of the most prominent tracers used for these objectives. Like the other alkali elements it strongly partitions into the fluid-phase, in particular at moderate to elevated temperatures. The magnitude of the partition is strongly temperature dependent. Lui Chan who was a world leader on Li and its isotopes for tracing fluid reactions and cycling focused on processes at plate boundaries. In subduction zones she recorded the behavior of Li from the incoming plate to the arc volcanics, and concluded that the variability from incoming plate to arc reflects the nature of the subducted material. Data from two subduction zones, Costa Rica, and Nankai Trough, will be presented. Recent hydrothermal experiments by Wei Wei on MORB-seawater and smectite-seawater, 35-350°C at 25°C steps, and 600 bars, greatly expanded the data-base, thus, insight, on the behavior of Li. The results indicate that Li is released into the fluid-phase throughout the temperature range of the experiments, with a strong threshold of significant release at ~250°C; indeed, Li concentrations increase in fluids with depth in subduction zones. Accordingly, because clay-rich sediments and altered oceanic crust are enriched in Li and the Li isotope values are lower than the seawater value, the fluids that migrate up-dip from a deeper source into the ocean should have a lower isotope signature, eventually approaching the source material, as observed in the pore fluids of the décollement zones at the Costa Rica and Nankai Trough subduction zones. The recent recovery of formation fluids at two sites at the Costa Rica subduction zone provide for the first time two year records on temporal variations (1) on the chemistry of the incoming plate upper basement formation fluid, (2) on the décollement fluid at 0.6 km arcward of the deformation front, and (3) on the relations of chemistry, tectonic events, and flow rates. The formation fluid Li concentrations and isotope data at the ODP reference Site 1253 support mixing between seawater and a deep-sourced fluid from the forearc of the Costa Rica subduction zone, implying that the uppermost permeable basement serves as pathway of fluid expulsion from the forearc. At the Nankai Trough décollement the Li concentrations are significantly higher and the δ7Li value is lower than at Costa Rica, reflecting the different sediment inputs and geothermal gradients at these two subduction zones.
V32A-06 INVITED
Utility of Li and Li Isotopes as Tracers of Continental Weathering
Lithium is potentially an attractive tracer of continental weathering because its two isotopes have a large relative mass difference, it is unaffected by biological activity and it is only slightly incompatible during magmatic processes so tends to be relatively uniformly distributed in the Earth's crust. Moreover, Li is conservative in the oceans, with a residence time of ca. 1 million years, and it is isotopically uniform on a global scale (δ7Li ~+31‰). Seminal work by Lui Chan and her co-workers has shown that the Li and Li isotope balance of the oceans is maintained by inputs of high-temperature hydrothermal fluids at oceanic ridges (with δ7Li ~+6.7‰) and dissolved Li from rivers (average δ7Li = +23‰), and low-temperature removal of Li into oceanic basalts and marine sediments. Despite this potential, relatively little is known about the behaviour of Li during continental weathering. In this study, we will present an overview of the work that we have conducted on Li and Li isotopes in rivers, soils and estuaries from catchments in Greenland, Iceland and the Himalaya, as well as laboratory studies of mineral dissolution and precipitation. Each of these case studies illustrates the effects of weathering processes on the riverine isotope signal, and the estuarine data illustrate how this signal is transferred to the oceans. Our data suggest that variations in rock type have little effect on riverine δ7Li; the principal control is preferential removal of 6Li into secondary minerals formed during weathering, leaving the residual waters enriched in 7Li. In subglacial environments, where weathering rates are very low, uptake of Li by ferric oxyhydroxides formed during sulphide oxidation is important. Our results clearly demonstrate that weathering processes can exert a significant effect on the Li isotope composition of natural waters. In order to understand whether changes in such processes with time are preserved, we have also generated records of the past variation in the Li and Li isotopic composition of the oceans. The challenges in the generation and interpretation of these records will be discussed.
V32A-07
Indicators of Provenance Weathering: Li and δ7Li in Mudrocks from the British Caledonides
We determined the Li concentration [Li] and isotopic composition (δ7Li), as well as major, trace element and Sr and Nd isotopic compositions of mudrocks (mudstone, shale, slate) from three Lower Paleozoic basins within the British Caledonides in order to determine the effects of sub-greenschist facies metamorphism on Li and the factors that control Li in mudrocks. [Li] varies widely, from 29 to 139 ppm, with mudrocks from the northern Lake District generally having higher concentrations (56-136 ppm, average 102 ppm) than those from the Scottish Southern Uplands (28-74 ppm, average 50 ppm) or southern Lake District (40-91 ppm, average 52 ppm) basins. δ7Li of mudrocks from the northern Lake District (δ7Li =-3.2 ±1.6 permil, 2σ) are relatively constant compared to those of the mudrocks from the southern Lake District (-3.4 permil to 0 permil) and the Southern Uplands (-4.4 permil to +3.7 permil). Metamorphic grade, determined by the Kübler index method, does not correlate with [Li] or δ7Li, indicating that sub-greenschist facies metamorphism had little effect on Li in these mudrocks. Collectively, the data for all three basins show a negative correlation between [Li] and δ7Li and a positive correlation between [Li] and the chemical index of alteration (CIA), suggesting that provenance exerts the greatest control on Li in mudrocks. Samples from the northern Lake District, which were deposited in an extensional basin, have homogeneous REE patterns, similar to shale composites (PAAS and NASC), the highest CIA, Th/U and [Li] and the lowest δ7Li and εNd, consistent with their derivation from a highly weathered ancient continental source. By contrast, mudrocks from the Southern Uplands range to the lowest CIA, Th/U and [Li] and have the highest δ7Li and εNd. These samples were deposited in a subudction zone on the southern margin of the Laurentian craton and contain volcanic detritus derived from a proximal arc. They have the most variable REE patterns, ranging from average shale-like patterns to less LREE-enriched patterns. The heterogeneity within the Southern Uplands mudrocks points to a mixed provenance that includes juvenile crustal materials (lower [Li], εNd and Th/U, higher δ7Li), likely derived from the arc, as well as more weathered cratonic detritus. Mudrocks from the southern Lake District deposited in a foreland basin, exhibit geochemical characteristics intermediate between the northern Lake District and the Southern Uplands mudrocks indicating their derivation from a mixed source. Our study shows that Li and δ7Li can provide addition information on the nature of the provenance of mudrocks.
V32A-08
Laboratory evaluation of Li isotopic fractionation in carbonates: inorganic precipitation experiments
Laboratory experiments were conducted to precipitate calcite and aragonite micro-crystals separately under controlled temperature and growth-rate conditions. High purity aragonite and calcite were precipitated and were examined by XRD, Raman and SEM. It is evident that high precipitation rate and low temperature in favor of mixed-carbonates formation. The obtained carbonate precipitates, along with paired fluids and the mother solutions, were acid dissolved and analyzed for various trace elements and stable isotopes using high resolution ICPMS and multi-collector ICP-MS, respectively. The Li partition coefficient (DLi) and the Li isotopic fractionation factors (ƒÑLi) were calculated at temperature between 5 and 40°C, where other stable isotopes (i.e., Li, B, Ca, and Sr) were also determined. The derived DLi varies slightly (1.3- 1.6E-3) in aragonite, in strong contrast to that of large variation in calcite, DLi =2.1-9.2E-2. The calculated ƒÑLi at various temperatures keep rather constant and show a small positive gradient (0.03 permil/°C) in aragonite. These results agree with previous calcite precipitation and were applied to study δ7Li in coralline skeleton.