V51A-0508 0800h
Germanium-Silicon Fractionation During Weathering of Basalt and Granite: Examples from the Tropics
Silicate weathering processes and terrestrial Si cycling fractionate Ge/Si ratios, leading to elevated ratios in soils and correspondingly low ratios in streamwaters. We have studied weathering at tropical sites developed on basaltic (Hawaiian Islands) and granitic (Luquillo, Puerto Rico) parent materials to elucidate the processes controlling Ge/Si fractionation in soils. Ge/Si ratios in soils developed on Hawaiian basalt (molar Ge/Si $\sim$2.5\times10$^{-6}$) range from 2.5 to 35\times 10$^{-6}$. Young ($\sim$2 ka) soils that have lost little Si (relative to immobile Nb) are relatively unfractionated, while most older soils have Ge/Si ratios $\sim$8 to 10\times 10$^{-6}$. We attribute this fractionation to partitioning of Ge into secondary aluminosilicates (allophane). Old, highly weathered soils that have lost $>$90% of initial Si can have Ge/Si ratios up to 35\times 10$^{-6}$. These extreme Ge/Si ratios are influenced by a Ge-enriched non-silicate secondary phase, tentatively identified as a Ti-oxide. Although authigenic marine Fe-oxyhydroxides clearly scavenge Ge from seawater, our study of a soil redox gradient on Maui shows that Ge is not mobilized by reductive dissolution of pedogenic Fe-oxyhydroxides. Secondary Fe-minerals apparently play an insignificant role in Ge/Si fractionation in soils. Ge/Si ratios in granitic soils are also elevated relative to parent material but display strong mineralogical control. Individual primary minerals in the quartz diorite Rio Blanco stock in Puerto Rico (Ge/Si = 2.0\times 10$^{-6}$) range from 0.5\times 10$^{-6}$ (quartz) to 6.6\times 10$^{-6}$ (hornblende). Incongruent weathering of plagioclase (Ge/Si = 1.5\times 10$^{-6}$) strongly partitions Ge into kaolinite (Ge/Si = 4.9 to 6.1\times 10$^{-6}$.) Soil and saprolite Ge/Si ratios range from 2.6 to 3.6\times 10$^{-6}$ reflecting relative amounts of residual quartz and biotite, and neoformed kaolinite. Streamwater Ge/Si ratios at the Hawaii and Puerto Rico sites are almost always lower than local bedrock, and vary between 0.2\times 10$^{-6}$ and 2.3\times 10$^{-6}$ reflecting sampling of distinct Si sources by changing hydrologic flowpaths. At the Puerto Rico site, low Ge/Si ratios at baseflow reflect Si released by incongruent weathering of plagioclase to form kaolinite. Increasing Ge/Si during storm flow reflects Si released from weathering of biotite and kaolinite in shallow soils. In Hawaii, this pattern is overprinted by Si cycled through plant phytoliths, which carry low Ge/Si ratios ($\sim$0.1 \times 10$^{-6}$.) Our results demonstrate that Ge/Si can be a useful tracer of both silicate weathering processes and the terrestrial biological silica cycle.
V51A-0509 0800h
Insights Into Mass-Dependant Fractionation During Desolvation of Analyte Solutions and its Importance for Accurate Isotope Ratio Measurements by MC-ICP-MS
A series of experiments with mixed Pb-Tl solutions revealed that variable photoxidation of Tl+ to Tl3+ occurs in the presence of Pb and solar UV radiation. The Tl3+ species behave distinctly from Tl+ and Pb2+ when using a membrane-based desolvation system and yield less precise and less accurate results. For example, epsilon 205Tl variations from -4 to +30, equivalent to the observed natural variations, were measured in the same Tl standard solution when thallium was in the Tl3+ form. When analyses were restricted to Pb-Tl solutions, in which thallium was in the Tl+ state, highly precise isotopic ratios were obtained for lead (206Pb/204Pb=16.9373 (+/-0.0011, 2s), 207Pb/204Pb=15.4907 (+/-0.0012, 2s), and 208Pb/204Pb=36.6935 (+/-0.0039, 2s)) and for thallium (e205Tl=1.5 (+/-0.8, 2s)). Qualitative scans of the DSN waste product, produced during analyses of mixed Pb-Tl solutions reveals presence of Pb and Tl, indicating that a portion of the sample solution passes through the DSN membrane. The isotopic variations observed during desolvation of the solutions are most likely related to one or more processes occurring during desolvation including: 1) Fractionation may occur during transport of the analyte through the porous fluorocarbon membrane tube as a result of differential diffusion across the membrane due to differential gas pressure and/or differences in volatility of the chemical species in the analyte. During this process, commercially utilized in uranium enrichment plants, the lighter isotopes diffuse faster than the heavier isotopes through the porous membrane. 2) Fractionation may occur during the vaporization of solvent, which leads to the formation of compounds and/or complex molecules that may preferentially incorporate heavier isotopes. In addition, the so formed compounds, molecules, and/or hydrous ions may have different vapor pressures, and consequently, one species (e.g., Tl) may be lost preferentially to another (e.g., Pb) and also isotopically fractionated. The issue for isotope ratio mass spectrometry is to evaluate the mechanism by which analyte ions are fractionated during desolvation, the extent to which they are fractionated elementally and isotopically, and to determine how appropriate corrections can be applied to the data. In the best-case scenario, fractionation during desolvation would be sufficiently small so that overall fractionation (i.e. sum of fractionations occurring in the desolvator, the plasma, and the plasma interface) can be adequately described by a single mass bias equation (e.g., "exponential law"). The extent to which a single correction factor can be used to correct isotope ratio data sets, however, must be evaluated on a case-by-case basis. For example, we have shown that Pb and Tl+ have very similar fractionation profiles during desolvation and in the mass-spectrometer, and a single correction factor can be used to correct Pb or Tl isotope ratio data. On the other hand, the behavior of Pb+2 and Tl+3 is such that the use of a single fractionation correction will not yield precise data. Because the extent of fractionation occurring during desolvation is sensitive to the operating conditions of the desolvator and the chemical species present in the analyte, it is critical to independently evaluate the extent of fractionation for each analyte solution in order to obtain the most precise isotope ratios.
V51A-0510 0800h
The Separation and Isotopic Analysis Seawater Cu and Zn
Many transition metals are key micronutrients and their concentration profiles in the oceans often show nutrient-like patterns, with strong surface depletions and deep enrichments$^{1}$. In addition, their biological usage has been shown to induce isotopic fractionations$^{2}$ so that the precise and accurate analysis of their isotope systems in seawater has potential applications in tracing metal micronutrient usage in the past ocean. The analytical challenges involved in realising this goal are, however, considerable, given the low concentrations of transition metals in seawater and the requirement to extract small amounts from large samples at low blank and with no artificial isotopic fractionation. Here we present a method for the separation an analysis of Cu and Zn isotopes that is applicable to 0.1-5 L samples of seawater. Trace metals were concentrated from seawater using a Chelex-100 ion-exchange column$^{3}$ and further purified and separated from each other using a small anion column$^{4,5}$. All isotopic analyses were performed on a ThermoFinnigan Neptune instrument at the University of Bristol. The main requirements for precise and accurate isotopic analyses are a low contribution from analytical blank and the robust correction for analytical mass discrimination. Our blanks allow the analysis of seawater samples of 50-250 mL for Cu, samples of about 100 mL for Zn in the deep oceans and for Zn-depleted open ocean surface water samples of around 5L. The correction for mass discrimination is most readily considered as two components - that occurring during the chemical separation procedure in response to non-100% yields and that occurring in the mass spectrometer. Correction of all mass discrimination throughout the procedure is most robustly done for Zn and Fe using a double-spike that is added prior to any chemical treatment. This approach has been tested using standard-doped seawater samples that had previously been stripped of their metal contents using the Chelex column. For Zn, the approach is highly successful and yields $\delta$$^{66}$Zn = -0.02$\pm$0.08 (n = 6) relative to the standard dopant. For Fe the precise analysis of low concentration samples is precluded by the propagation of large errors on the $^{57}$Fe/$^{54}$Fe through the double-spike calculation. Mass discrimination correction is simplified if the chemistry yield is 100% since then the mass spectrometric component can be corrected for using established techniques$^{4,5}$. The yield for Fe from the chemistry is not 100% but demonstrably is for Cu and Zn. Correction for instrumental mass discrimination using these conventional approaches is compromised by non-spectral matrix effects which cause changes in the behaviour of pure standards that have been (noted previously$^{5}$). This is overcome by comparing samples to a standard that has a matrix similar to seawater - e.g. a trace metal-stripped, standard-doped seawater sample. The two approaches yield identical results for the isotope composition of Zn in an English Channel sample relative to the Lyons JMC standard: $\delta$$^{66}$Zn = 0.38$\pm$0.06 (double spike, n = 12) and 0.46$\pm$0.08 per mil (standard-bracketing n = 6). $^{1}$ K.W. Bruland, 1980, Earth Planet. Sci. Lett. 47, 176. $^{2}$ B.L. Beard et al., 2003, Chem. Geol. 195, 87. $^{3}$ H.M. Kingston et al., 1978, Anal. Chem. 50, 2064. $^{4}$ C.N. Marechal et al., 1999, Chem. Geol. 156, 251. $^{5}$ C. Archer and D. Vance, 2004, J. Anal. Atom. Spectr. 19, 656.
V51A-0511 0800h
Iron Isotope Variations in Reduced Groundwater and in Drinking Water Supplies: A Case Study of Hanoi, Vietnam
In reduced groundwater iron is involved in biotic and abiotic transformation processes, both of which could lead to iron isotope fractionation. The reduced groundwater aquifers in the area of the Vietnamese capital of Hanoi are the main drinking water sources for the city. These groundwaters contain arsenic, which imposes a serious health threat to millions of people. Dissolved arsenic is related to the reducing conditions prevalent in the groundwater, and iron and arsenic contents are correlated in the sediments. We are employing iron isotope composition as a tool to better understand the processes leading to the transformation of iron in the groundwater and its role in various biogeochemical processes in reduced environments. Drinking water is supplied to the city of Hanoi from several water treatment plants (WTP) which pump the raw groundwater from a lower aquifer, while the rural surroundings pump untreated groundwater from an upper aquifer by private tubewells. Surface water from the Red River delta is the main source of recharge to these two aquifers. Due to high content of particulate natural organic matter (NOM) in the sediment leading to extensive microbial activity, the groundwaters are anoxic and rich in dissolved iron(II). The iron(II) removal in the WTPs is carried by a multi-step treatment including aeration, settling, filtration, and chlorination. We have collected natural groundwater samples for isotopic analysis from two aquifers at several locations, a groundwater depth profile and its corresponding sediment phases from the upper aquifer and the underlying aquitard, raw and treated water from several WTPs, as well as the corresponding iron(III) precipitates. The iron concentrations of groundwaters analysed in this study range from 3 to 28 mg/L and $\delta$$^{57}$Fe (57/54 deviation from IRMM 014) values vary between -1.2 and +1.5 $\permil$. The sediment depth profile has a $\delta$$^{57}$Fe around +0.3 $\permil$, which implies that the high values obtained in the groundwater nearby (+0.9 - +1.2 $\permil$) cannot be explained by a simple reductive dissolution process, which would be expected to favour the lighter Fe isotopes. Removal of iron in the WTP is followed by a strong decrease of $\delta$$^{57}$Fe, probably due to formation of heavier Fe(III) phases. High $\delta$$^{57}$Fe values are found in both aquifers and correspond to high concentrations of iron in the groundwater. We hypothesize that the iron isotopic variations observed so far are an indication for iron sources and transformation processes that could not be detected by only measuring dissolved iron concentrations. Current investigations will further explore this hypothesis.
V51A-0512 0800h
The isotopic effects of electron transfer: an explanation for Fe isotope fractionation in nature
Recent developments in mass spectrometry techniques have created opportunities to examine the partitioning behavior of stable isotopes of transition metals with a focus on application to iron isotopes. Iron oxidizing and reducing bacteria have been shown to cause isotope fractionations similar in magnitude to those observed in sedimentary environments and it is believed that biological activity is responsible for the most significant Fe isotope fractionation in natural settings. Debate over the use of Fe isotopes as a biological marker resulted from subsequent measurements of fractionations in a variety of abiotic systems. The accumulated evidence, in both biotic and abiotic systems, points to a connection between redox processes and Fe isotope fractionation, however the exact mechanism for isotope fractionation is not yet well understood. Here, we present both a newly-developed theory based on chemical kinetics and preliminary experimental results that quantitatively delineate the relationship between driving force in a charge transfer reaction and resulting Fe isotope fractionation. The theory, based on R. Marcus's chemical kinetics theory for electron transfer ({\it Ann. Rev. Phys. Chem}. {\bf 15} (1964), 155), predicts that fractionation increases linearly with driving force with a proportionality related to two factors: the difference between isotopic equilibrium exchange of products and reactants, and the reorganization energy along the reaction coordinate. The theoretical predictions were confirmed by measurements of isotopic fractionation associated with electroplating iron metal from a ferrous chloride solution. Isotope fractionation of Fe electroplated under potentiostatic conditions was measured as a function of applied electrochemical potential. As plating voltage was varied from -50 mV to -2.0 V, the isotopic signature of the electroplated iron became depleted in heavy Fe, with $\delta$$^{56}$Fe values ranging from -0.106($\pm$0.01) to -2.290(±$\pm$0.006)$\permil$, and corresponding $\delta$$^{57}$Fe values of -0.145($\pm$.011) and -3.354($\pm$.019)$\permil$. The slope of the line created by plotting $\delta$$^{56}$Fe vs $\delta$$^{57}$Fe is equal to 0.6723($\pm$.0032), consistent with fractionation due to a kinetic process involving unsolvated iron atoms. This study demonstrates that there is a voltage-dependent isotope fractionation associated with the reduction of iron. The magnitude of fractionation is similar to observations of Fe reduction by certain bacteria, suggesting that electrochemical processes may be responsible for observed biogeochemical signatures. Charge transfer is a fundamental physicochemical process involving Fe as well as other transition metals with multiple isotopes. Partitioning of isotopes among elements with varying redox states holds promise as a tool in a wide range of the Earth and environmental sciences, biology, and industry.
V51A-0513 0800h
Variations in the marine Ca cycle and implications for paleo-CO$_2$ levels over the past 24 Ma
A detailed record of the calcium isotopic composition of bulk nannofossil ooze from DSDP Site 590, based on 40 samples measured multiple times, shows variations of $\delta^{44}$Ca over the past 24 million years between --0.2 and --0.9$\permil$, relative to bulk Earth Ca (--1.15 to --1.85$\permil$, relative to seawater). These isotopic variations are inferred to reflect changes in $\delta^{44}$Ca of seawater between +1.1 and +0.4$\permil$. Fluctuations in $\delta^{44}$Ca, which occur in 1 to 4 Ma cycles with varying amplitude, are interpreted as resulting from imbalances between the input of Ca to the oceans by weathering processes and the biogenic removal of Ca. Using a model for oceanic inputs and outputs of Ca, we reconstruct past weathering fluxes and marine Ca concentrations. Combining this information with paleo-pH of the ocean, as estimated from B isotopes, we derive a record of atmospheric $pCO_2$ levels over the past 24 Ma. The concentration of Ca in the ocean was at a minimum at $\sim$20 Ma, during the Neogene climate optimum, increased until 6 Ma, and then decreased toward the present. Maxima in the record occur at 6.4 and 4 Ma, corresponding to a generally-recognized period of enhanced marine productivity and high mass accumulation rates at Site 590. Peaks in the inferred weathering flux occur at 18, 14, 8, 5, and 1 Ma. Our derived $pCO_2$ record indicates that the highest $CO_2$ levels during the 24 Ma period under consideration occurred at 20--24 Ma and were no more than about twice the present-day (pre-industrial) levels. The maximum in the $CO_2$ record occurs during the warming that accompanied the Neogene climate optimum. The $pCO_2$ minimum occurs at 6--4 Ma, during a period of enhanced marine productivity. The calculated marine paleo-Ca and paleo-$pCO_2$ values are sensitive to the assumed $\delta^{44}$Ca value for the weathering flux, but the shapes of the derived curves do not change greatly over the range of likely input values. The reconstructed $pCO_2$ curves depend significantly on pH, but the assumption of constant pH (which only barely violates the available constraints from B isotopes) yields substantially the same result. Although the paleo-$pCO_2$ curve is likely to be revised when more data are available, the present results do not show a simple linear relationship between $pCO_2$ and high-latitude temperature as reflected in the benthic oxygen isotope records.
V51A-0514 0800h
LA-MC-ICPMS Determination of Copper Isotope Ratios in Turquoise from the Southwestern United States.
Hydrothermal circulation driven by igneous intrusion led to the deposition of turquoise throughout the southwestern United States and Mesoamerica. The genesis of these copper-ore deposits is unclear; conflicting hypotheses call on ascent of magmatic waters (hypogene) or descent and recirculation of meteroric waters (supergene). Copper isotope analyses were performed by laser-ablation multi-collector ICPMS to survey turquoise deposits from AZ, NV, CA, NM, and CO. The turquoise have [Cu] from 0.1 to 10 wt% and are all found in near-surface alteration zones. Analyses of individual turquoise grains are reproducible to better than 0.4\permil \delta$^{65}$Cu (1\sigma) (relative to NBS-976). \delta$^{65}$Cu values show significant variation (ca. 10\permil) between the deposits, equal to the total range reported for continental ores and both hypogene and supergene deposits. The variability between deposits may reflect differences in source Cu isotopic composition or more likely, hydrothermal processes during leaching and deposition. The mining and trade of turquoise played an important role in early social and economic development between Mesoamerica and N. America. Copper isotopes will improve differentiation between turquoise source areas, aiding archaeological and cultural studies of trade between and within Mesoamerica and the SW USA. Research sponsored by NSF-BCS (Archaeology) grant #0312088 to Fayek and the Office of Basic Energy Sciences, U.S. Department of Energy, under contract with Oak Ridge National Laboratory, managed by UT-Battelle, LLC. The submitted manuscript has been authored by a contractor of the U.S. Government under contract No. DE-AC05-00OR22725. Accordingly, the U.S. Government retains a nonexclusive, royalty-free license to publish or reproduce the published form of this contribution, or allow others to do so, for U.S. Government purposes.
V51A-0515 0800h
Mercury Isotope Variations in Hydrothermal Ore Deposits
The ability to make direct isotopic measurements of ore-forming metals using MC-ICPMS has introduced the possibility for their use as tracers of Hg source in ore deposits and the environment. The isotopic composition of Hg varies by over 5 \permil \delta$^{202}$Hg/$^{198}$Hg (relative to our Almaden Hg standard), more than 50 times the 0.1 \permil (2\sigma ) analytical uncertainty, in a wide variety of hydrothermal ore deposits. This variation could be caused by fractionation related to processes of redox, mineral precipitation, and boiling hydrothermal fluids, among others, that are known to cause large isotopic variations in other stable isotope systems. To test the possibility that Hg isotopes can be used as a tracer of source, we have compared isotopic compositions of ore and possible source rocks in three ore deposit types: epithermal Au-Ag veins and sinter where fluids boiled in the shallow crust; silica-carbonate-type Hg deposits in the California Coast Range, where reduced source fluids mixed with oxygenated groundwater and MVT Pb-Zn deposits, where sulfides are deposited in a basin without redox or boiling. Epithermal ores (-3.1 to +2.5 \permil) and silica-carbonate-type ores (-1.1 to +1.3 \permil) have Hg isotopic compositions that show much larger variations than MVT ores (-0.5 to +0.6 \permil ). The large variations might reflect the effects of fractionation by boiling and oxidation in epithermal and silica-carbonate-type deposits. At the Buckskin National epithermal deposit, unaltered andesite, felsic dike and phyllite basement rock have identical isotopic compositions of -1.3 \permil in the middle of the epithermal ore range, as might be expected if the light and heavy ends of the range were produced by fractionation related to boiling. At the Mayacmas silica-carbonate Hg district, rocks from the Franciscan Complex and Coast Range Ophiolite range from -2.4 to 0.0 \permil compared to a range of -0.2 to +1.3 \permil for ores, as might be expected if kinetic isotope effects related to oxidation concentrated heavy isotopes in the ore. In laboratory experiments, a 1.3 \permil fractionation between vapor and liquid Hg at 20 \deg C, and a 0.1 \permil fractionation between dissolved Hg and HgS in precipitation experiments were measured and further experiments are being conducted to better understand the processes fractionating Hg. The above results suggest that Hg isotopic variability is widespread in hydrothermal systems and may be sufficient to trace Hg source given an understanding of the fractionation mechanisms operating in these systems.
V51A-0516 0800h
Fe Isotope Fractionation: Effect of Desferrioxamine Mesylate (DFAM) and Reduction of Goethite by Total Membrane Fraction of {\it Shewanella oneidensis}
In both abiotic adsorption of Fe(II) on goethite and reduction of goethite by total membrane fraction of {\it Shewanella oneidensis} ({\it in vitro} reduction), addition of the strong Fe chelator (desferrioxamine mesylate, DFAM) affected the fractionation of Fe isotopes. In order to see the effect, we measured the \Delta$^{56/54}$Fe$_{aq}$, defined as \delta$^{56/54}$Fe$_{aq}$ - \delta$^{56/54}$Fe$_{source}$, where \delta$^{56/54}$Fe$_{source}$ is the isotopic composition of either goethite or the adsorption control solution. With DFAM, the values of \Delta$^{56/54}$Fe$_{aq}$ were more positive than \Delta$^{56/54}$Fe$_{aq}$ values measured without DFAM, which indicates that aqueous Fe(II) is enriched with $^{56}$Fe(II) compared to the Fe sources in the presence of DFAM. It is notable that {\it in vitro} reduction showed that the membrane fraction also could have a similar effect as DFAM does. We also observed that the adsorbed Fe(II) on goethite in the experiment without DFAM or membrane became isotopically heavier (enriched with $^{56}$Fe(II)) in 7 days. In comparison, in the DFAM and membrane experiments, the adsorbed Fe was not as isotopically heavy, though similar trend was observed.
V51A-0517 0800h
Natural variations in calcium isotope composition as a monitor of bone mineral balance in humans.
The skeleton is the largest reservoir of calcium in the human body and is responsible for the short term control of blood levels of this element. Accurate measurement of changes in bone calcium balance is critical to understanding how calcium metabolism responds to physiological and environmental changes and, more specifically, to diagnosing and evaluating the effectiveness of treatments for osteoporosis and other serious calcium-related disorders. It is very difficult to measure bone calcium balance using current techniques, however, because these techniques rely either on separate estimates of bone resorption and formation that are not quantitatively comparable, or on complex and expensive studies of calcium kinetics using administered isotopic tracers. This difficulty is even more apparent and more severe for measurements of short-term changes in bone calcium balance that do not produce detectable changes in bone mineral density. Calcium isotopes may provide a novel means of addressing this problem. The foundation of this isotope application is the ca. 1.3 per mil fractionation of calcium during bone formation, favoring light calcium in the bone. This fractionation results in a steady-state isotopic offset between calcium in bone and calcium in soft tissues, blood and urine. Perturbations to this steady state due to changes in the net formation or resorption of bone should be reflected in changes in the isotopic composition of soft tissues and fluids. Here we present evidence that easily detectable shifts in the natural calcium isotope composition of human urine rapidly reflect changes in bone calcium balance. Urine from subjects in a 17-week bed rest study was analyzed for calcium isotopic composition. Bed rest promotes net resorption of bone, shifting calcium from bone to soft tissues, blood and urine. The calcium isotope composition of patients in this study shifted toward lighter values during bed rest, consistent with net resorption of isotopically-light bone calcium. In contrast, little shift was seen in patients who exercised during bed rest, consistent with the expectation that exercise should inhibit bone resorption. Most intriguingly, an opposite-sense shift was seen in patients who were administered aledronate, a drug which inhibits bone resorption. We hypothesize that these patients entered a state of positive bone mineral balance despite initiation of bed rest. Comparison of isotopic data with measurements of bone mineral density and metabolic markers of bone metabolism confirms that the calcium isotope composition of urine reflects changes in bone mineral balance. Calcium isotope analysis of urine and soft tissues may provide information on bone mineral balance that is in important respects better than that available from other techniques, and illustrates the usefulness of applying geochemical techniques to biomedical problems.
V51A-0518 0800h
Hg Isotope Ratios of a Sediment Core from Plastic Lake, Ontario: Implications for Hg Cycle in Aquatic Environment
Hg isotope ratios in a sediment core obtained in Plastic Lake, Ontario, Canada, have been measured by coupling a gold trap with an MC-ICP-MS. The core is about 30 cm in depth and corresponds to a time period of about 250 years, based on 210Pb dating. The samples were combusted at high temperature and the Hg collected onto a gold trap. The gold trap was subsequently heated to release Hg directly into the MC-ICP-MS. An in-house sample introduction system was employed to extend Hg signal duration in order to obtain high precision in isotope ratio measurement. The instrumental mass bias was corrected using Tl introduced simultaneously via an Aridus membrane desolvation nebulizer. Based on long term measurement of a NIST-2225 elemental Hg standard (over 120 measurements since Oct. 2002), the external reproducibility ranges from 45 ppm for 201Hg/202Hg to 100 ppm for 199Hg/202Hg (2 sigma relative standard error). Hg in the sediments shows an increase in light isotope enrichment at about 10 cm depth. The total Hg also displays an increase at the same depth. The depth corresponds to approximately the 1920s, a time period when there was a major increase in coal-burning power generation. Limited Hg isotope data for other terrestrial samples appears to indicate that Hg bound to organic carbon is enriched in light isotopes relative to elemental Hg. For example, the DOLT-3, a dogfish liver standard reference material with half of its Hg as MeHg, has the lightest Hg isotope composition among measured terrestrial samples. It is not clear at this stage whether the increase in total Hg and light Hg isotope enrichment in recent years represent a change in methylation rate of the lake, or an increase in atmospheric deposition of Hg combined with a change in source. Discussions based on available Hg isotope data of terrestrial samples together with other chemical data for the lake will be presented.
V51A-0519 0800h
Chromium Stable Isotope Fractionation During Abiotic Reduction of Hexavalent Chromium
Chromium, a common surface water and ground water contaminant, occurs as Cr(VI), which is soluble and toxic, and Cr(III), which is insoluble and less toxic. Reduction of Cr(VI) to Cr(III) is often the most important reaction controlling attenuation of Cr plumes, and Cr stable isotope ($^{53}$Cr/$^{52}$Cr) measurements show great promise as indicators of this reaction. Cr(VI) reduction involves a kinetic isotope effect; lighter isotopes react at greater rates and heavier isotopes become increasingly enriched in the remaining Cr(VI) with increasing extent of reduction. If the size of this effect can be constrained well, then precise estimates of reduction are possible. Cr(VI) reduction can be mediated by microbes, or may occur abiotically in the presence of Fe(II) and a variety of organic compounds. A recent study of bacterial reduction of Cr(VI) under low electron donor conditions yielded a Cr isotope fractionation factor of 1000ln$\alpha$ = 4.1 $\pm$ 0.2. A previous study of abiotic reduction indicated a fractionation factor of 1000ln$\alpha$ = 3.4 $\pm$ 0.2, but this work was limited to 3 experiments. The present study provides a more detailed look at Cr isotope fractionation induced by abiotic Cr(VI) reduction by: Fe(II); mandelic acid with alumina and goethite catalysts; and humic substances. Reduction occurred slowly, over days or weeks. The fractionation factor for the organic reductants (all at pH=4), including two surface-catalyzed mandelic acid reactions, two fulvic reactions, and one humic reaction,- was 1000ln$\alpha$ = 3.0 $\pm$ 0.4, with no statistically significant differences between experiments. The fractionation factors for the Fe(II) experiments were 4.7 $\pm$ 0.3, 3.7 $\pm$ 0.2, and 2.9 $\pm$ 0.2 for pH = 4, 5, and 6, respectively. Further work is necessary to better constrain this pH dependence and to determine if it occurs with the organic reductants. The overall variability in the size of the Cr isotope fractionation during Cr(VI) reduction translates into a moderate level of uncertainty in Cr/$^{52}$Cr-based estimates of reduction.
V51A-0520 0800h
Silicon Isotope Fractionation by Banana Under Continuous Nutrient and Silica Flux
Silicon is absorbed by plants as aqueous H$_{4}$SiO$_{4}$ with other essential nutrients, and precipitates in aerial parts of the plant as phytolith, a biogenic opal. Phytoliths are restored to the soil by decomposition of organic debris from plant material. The role of higher plants in the biogeochemical cycle of silicon is therefore major although it is still poorly studied. Biomineralization processes are known to fractionate the three stable silicon isotopes with a preferential uptake of light isotopes. Therefore, following some preliminary results from Douthitt (1982), and studies presented in recent conferences (Ziegler et al., 2002; Ding et al., 2003), we suspect that phytolith production by plants could also fractionate the silicon isotopes. Inversely, intensity of phytolith-related isotopic fractionations might contribute to a better understanding of the soil-plant silicon cycle. Our study focused on banana, a silicon accumulating plant ($>$1% Si, dry weight).{\it Musa acuminata} cv Grande Naine has been grown in hydroponics under controlled conditions (light, temperature, humidity, nutrients) during six weeks. The nutrient supply was kept constant: three batches of five plants were grown with a continuous nutrient solution flow of 5, 50 and 100 ppm SiO$_{2}$ respectively. Si isotopic compositions were measured in the source solution, and in silica extracted from the various parts of banana (roots, pseudostems, midribs and petioles, leaves), using a Nu Plasma multicollector mass spectrometer (MC-ICP-MS) operating in dry plasma mode. The results are expressed as $\delta^{29}$Si relatively to the NBS28 standard, with an average precision of $\pm$ 0.03$\permil$. Silicon contents and morphological studies of phytoliths were also achieved. Banana $\delta^{29}$Si varied between -0.18 and -0.76$\permil$ with a source solution at -0.02$\permil$. Values of $\delta^{29}$Si were less fractionated, relatively to the nutrient solution, in roots, where no phytoliths have been observed until now, than in upper parts of banana where phytoliths were clearly abundant as long chain of typical cone shaped morphotypes truncated saddle-like. The bulk isotopic composition of the leaves in the five plants grown at 100 ppm SiO$_{2}$ displayed a homogeneous negative signature (-0.44 $\pm$ 0.08$\permil$) indicating a small inter-specimen variability. The difference between $\delta^{29}$Si in roots and in upper parts of the plant was much larger with a silica offer of 100 ppm SiO$_{2}$ (0.58$\permil$) than with 50 ppm SiO$_{2}$ (0.08$\permil$). However, silicon isotope fractionation in leaves was not affected by a change in Si supply. Our preliminary results show that biomineralization of silica in bananas fractionates silicon isotopes in a similar extent as marine diatoms.
V51A-0521 0800h
Fe and S isotope variations in cyanobacterial mats: modern analogues of ancient stromatolites
Iron and sulfur isotope variations in modern microbial mats from the hypersaline ponds of the Guerrero Negro salt works, Baja California Sur, Mexico have been investigated. Cyanobacteria are the primary producers of this mat ecosystem. The oxygen concentrations in the surface 2 mm of the mats alternate between oxygen-rich and oxygen-free over a diel cycle. Previous work indicates that heterotrophic metabolism is dominated by sulfate reducing bacteria whereas direct metabolic processing of Fe, such as dissimilatory Fe(III) reduction or photoautotrophic Fe(II) oxidation, is negligible. Extremely high rates of sulfate reduction are observed near the mat surface, which coincides with the highest $\delta$$^{34}$S values of sedimentary sulfides (total inorganic sulfide, TRIS). The overall $\delta$$^{34}$S values of TRIS are -19 to -46 $\permil$ decreased relative to seawater sulfate. The absence of significant S isotope variations in sedimentary sulfide below the surface 1 cm indicates that sulfate does not become limiting within the microbial mats. Iron isotope compositions of pyrite, expressed as $\delta$$^{56}$Fe and normalized to average igneous rocks, varies between -1.9 and -0.2 $\permil$. These values are in the range of previously reported Fe isotope compositions of sedimentary pyrites from Archean shales, Banded Iron Formations and modern continental margin sediments (-2.5 to -0.5 $\permil$). Similar to S isotopes, the highest values in $\delta$$^{56}$Fe are observed in the surface 1 cm of the mat. The dominant processes that control S isotope compositions are microbial processing of S, including bacterial sulfate reduction and S disproportionation; inorganic fractionations, e.g. during conversion of H$_{2}$S, are negligible. In contrast, Fe isotope variations are the cumulative expression of multiple microbial and inorganic reactions, including reductive dissolution, inter-mineral fractionations, ligand-promoted dissolution and redox reactions. In this study we examine how these fundamentally different controls on S and Fe isotope fractionations can be related to the isotope variations in a modern microbial mat. The aim is to provide a framework for the interpretation of S and Fe isotope compositions in ancient sedimentary environments.