Goldschmidt Conference Abstracts 2005

Goldschmidt Conference Abstracts 2005 Chemistry and Physics of Igneous Processes

Consequences of exsolution of H2O-, CO2-, SO2-, Cl-bearing volatile phases on the physical and chemical properties of magma JIM WEBSTER Dept. of Earth & Planetary Sciences, AMNH, NY, NY, USA ([email protected]) Magmatic volatiles exert a strong influence on the melting/crystallization behavior, phase equilibria, viscosities, and densities of silicate melts. Each of the volatiles H2O, CO2, SO2, and Cl exhibits different and varying solubilities in silicate liquids as a function of melt composition (X), pressure (P), and temperature (T). How these volatiles partition between silicate melts and magmatic volatile phases also varies with P, T, and X. The exsolution of a volatile phase or phases in open-system conditions can sequester significant quantities of these volatiles from melt, alter the activities of the most dominant volatiles in the melt, and influence the physical as well as chemical properties of melts. Most experimental constraints on volatile behavior are limited to systems containing silicate melts plus only one or two volatiles, but experiments are beginning to determine volatile behavior in systems containing melt-H2O-SO2-Cl and meltH2O-CO2-Cl. New experimental data for molten rhyodacite (Botcharnikov et al., 2004) and phonolite (Webster et al., 2005) show that the addition of a third volatile can lead to dramatic changes in the solubility behaviors of the other volatiles. These new solubility data are applied to processes of exsolution of Cl-enriched volatile phases, differentiation, and eruption of Mt. St. Augustine volcano, Alaska, and Mt. Somma-Vesuvius, Italy.

References Botcharnikov, R.E., Behrens, H., Holtz, F., Koepke, J., and Sato, H. (2004) Sulfur and chlorine solubility in Mt. Unzen rhyodacite melt at 850° C and 200 MPa. Chem. Geol. 213, 207-225. Webster, J.D., Sintoni, M. F., and De Vivo, B. (2005) The role of sulfur in promoting magmatic degassing and volcanic eruption at Mt. Somma-Vesuvius. In: Developments in Volcanology (B. De Vivo, ed.), Elsevier, in review.

A151

Lithium transport by a magmatic volatile phase beneath Mount St. Helens volcano J. BLUNDY1, K. BERLO2 AND K. CASHMAN3 1

Dept. of Earth Sciences, University of Bristol, Bristol BS8 1RJ, UK. ([email protected]) 2 Dept. of Earth Sciences, University of Bristol, Bristol BS8 1RJ, UK. ([email protected]) 3 Dept. of Geological Sciences, University of Oregon, Eugene, OR 97403-1272. ([email protected]) 81 plagioclase-hosted rhyolite melt inclusions in dacites erupted from Mount St. Helens in May-October 1980 have been analysed by ion-probe for trace elements and H2O. Inclusion H2O varies from to 0.4 to 6.4 wt%. Inclusions from May 18th white pumice have higher H2O than those from the cryptodome and subsequent eruptions. Negative correlations between H2O, SiO2 and incompatible trace elements, and positive correlations between H2O and Sr indicate of plagioclase-dominated decompression crystallisation of H2Osaturated dacite magma. Decrease in MgO and FeO with increasing SiO2 show that post-entrapment crystallisation of inclusions was minimal. Most melt inclusions contain 25-40 ppm Li, weakly positive-correlated with H2O. About 15 inclusions, from the cryptodome and post-May 18th samples, have high Li (≤100 ppm) at near-constant H2O (4.0±0.5 wt%, equivalent to pH2O=90-140 MPa. The same samples are also characterised by Li-rich plagioclase phenocrysts and excesses of (210Pb) over (226Ra), but otherwise show no anomalous chemical features. Our data suggest sub-volcanic transport of Li (and Pb) via a magmatic volatile phase. We propose that Li-bearing vapour was released from dacite magma with ~6 wt% dissolved H2O (pH2O=240 MPa). This is the magma reservoir from which the Plinian pumice was derived. The most H2Orich inclusions contain 0.10 wt% Cl, from which we calculate that the coexisting aqueous fluid contained ≥1.0 wt% Cl, or ≥1.6 wt% NaCl equivalent. At 800-900 °C such a fluid will condense to low-salinity vapour and brine at ~110 MPa. Clcomplexed cations such as Li and Pb will become concentrated in the dense brine phase; the low salinity vapour will tend to escape upwards. Re-equilibration between residual brine and magma will enrich the residual rhyolite melt in Li (+Pb) and buffer Cl at brine saturation (0.12-0.20 wt% Cl) as observed in melt inclusions with ≤4.5 wt% H2O. We propose that vapour-brine unmixing at ~110 MPa (~4 km depth) is an important process of mass transfer beneath Mount St. Helens and other calc-alkaline volcanoes.

A152

Goldschmidt Conference Abstracts 2005 Chemistry and Physics of Igneous Processes

Feedback between physical and chemical chararacteristics of an evolving open-system magma body WENDY A. BOHRSON1 AND FRANK J. SPERA2 1

Dept. of Geological Sci., Central Washington University, Ellensburg, WA 98926 ([email protected]) 2 Dept. of Geological Sci., Univ. of California, Santa Barbara, CA 93106 ([email protected]) Energy-Constrained Eruption, Recharge, Assimilation and Fractional Crystallization (EC-ERAFC) is a computational model that tracks the evolution of an open-system magma reservoir undergoing simultaneous eruption, recharge, wallrock assimilation, and fractional crystallization. ECERAFC is formulated as a set of coupled, nonlinear differential equations, the solutions of which provide detailed thermal, mass and chemical (trace element/isotope) information about the evolution of melt and solids as all parts of a composite system (host magma, recharge magma, wallrock, eruptive reservoir) approach a common equilibration temperature (Teq). Fundamental to the structure of EC-ERAFC is the premise that changes in physical conditions, such as initial, liquidus and equilibration temperatures, will cause changes in chemical signatures. Mass addition by recharge and removal by eruption also impact geochemistry. Two forward models highlight possible feedback effects. For cases in which other parameters are held constant, increasing the wallrock initial, liquidus Ts from 300°C, 1100°C to 600°C, 1150°C, respectively, yields magmas that have more crustal 87Sr/86Sr because, at Teq, wallrock with the higher liquidus T has undergone a smaller degree of partial melting. Because melting is approximated as a fractional process, and Sr is modeled as behaving incompatibly, smaller degrees of partial melting add relatively large masses of crustal Sr to the host magma, thus yielding a more crustal fingerprint. A second example examines the coupling between Teq and chemistry. For a case in which hydrothermal circulation is vigorous and magmatic heat is efficiently transported away from the magma-wallrock boundary, Teq is likely to be low. Such conditions yield more crust-like magmas because the degree of wallrock partial melting is smaller. In contrast, for higher Teq, larger degrees of partial melting add proportionally less crustal Sr to the host magma, yielding magmas that bear a smaller crustal imprint. Numerous other examples provide additional analysis of feedback effects and demonstrate the critical importance of applying integrated computational models to an understanding of how magma plumbing systems evolve.

Reverse zoned feldspars in Suswa Volcano, Kenya Rift: Evidence for magma mixing and eruptions triggered by recharge VANESSA V. ESPEJEL-GARCIA, ELIZABETH Y. ANTHONY AND MINGHUA REN Dept. Geol. Sci., UTEP, El Paso, TX 79968, USA. ([email protected], [email protected], [email protected]) Suswa Volcano is a Holocene volcano in the axis of the Kenya Rift. It is composed of trachytes that culminated in the formation of the caldera, followed by phonolites. We have studied two of the post-caldera phonolites. The first flow is characterized by phenocrysts of olivine (Fa67), diopside (En31Fs23Wo46), ulvöspinel (Fe2.5Ti0.5O4) and zoned feldspars. Some feldspars show oscillatory zoning with K-rich cores (An<10Ab62-66Or28-34), followed by euhedral overgrowths with either Ca or K-rich compositions; the Anrich zones are An7-13Ab66-68Or14-20 and the Or-rich zones are An7-13Ab67Or19-25. A second type of feldspar lacks oscillatory zoning and has trapped melt inclusions with the same glass composition as the matrix. Thin rims on the oscillatory zoned crystals (An56Ab44Or0) show the same composition as these Ca-rich overgrowths. A second flow has a number of feldspars with Ca-rich cores and another type with K-rich interior and subtle zoning (An<10Ab62Or28-34). A third type of feldspar has euhedral overgrowths with internal zones of An12Ab66Or22 and rims of An8Ab65Or28. Finally, a fourth type has a Ca-rich part (An50Ab50Or0), also with melt inclusions with compositions similar to the glass in the matrix. This feldspar is very similar to the second type of feldspar in the first flow, both with melt inclusions trapped in the Ca-rich portion. The zoning in feldspars implies magma mixing of felsic and basaltic compositions. We also interpret the thin rims with An50Ab50Or0 found in most of the feldspars to represent recharge of the chamber by a magma that triggered the eruption. Because melt inclusions are in the portions of the feldspars that have the same composition as the magma, these feldspars most have grown before the eruption.

Goldschmidt Conference Abstracts 2005 Chemistry and Physics of Igneous Processes

A153

Nonequilibrium H2O-CO2 exsolution and obsidian formation

Origin of UG2 and other chromitite layer of the Bushveld Complex

H. M. GONNERMANN1 AND M. MANGA2

SISIR K. MONDAL AND E.A. MATHEZ

1

Department of Earth and Planetary Sciences, Harvard University ([email protected]) 2 Department of Earth and Planetary Science, University of California, Berkeley ([email protected])

Dept. of Earth and Planetary Sciences, American Museum of Natural History, NY, NY 10024 ([email protected]; [email protected])

The ratio of CO2 to H2O in volcanic glasses is one of the few available measurements directly related to syneruptive magma dynamics. Ascent-driven magma decompression results in the exsolution of dissolved volatiles from the melt. This in turn leads to bubble growth and possibly to gas loss from the permeable magma. Pyroclastic obsidian samples from the ca. 1340 A.D. Mono Craters, California eruption (Newman et al., 1988) have relatively high CO2/H2O values. These can be explained in terms of equilibrium closed-system (Newman et al., 1988), as well as open-system (Rust et al., 2004), magma degassing. However, both scenarios require the presence of CO2-rich vapor to buffer CO2 exsolution from the melt. We present results from numerical modeling of nonequilibrium joined CO2-H2O exsolution under open- and closed-system conditions. We find that the Mono Craters CO2/H2O trend is well explained by nonequilibrium opensystem degassing. Our results are a consequence of the low diffusivity of CO2 (Watson, 1994). If, during decompression, open-system gas loss occurs at a rate comparable to the rate of volatile exsolution, bubbles remain small. Consequently, the surrounding melt shell remains relatively thick and CO2 diffusion through the melt is slower than the rate at which CO2 solubility decreases at the bubble wall. Thus, CO2 concentrations throughout much of the melt remain above equilibrium. Consistent with Rust et al. (2004), we conclude that throughout the Mono Crateres eruption, even at ascent rates sufficiently fast to sustain explosive activity, open-system degassing of some magma fraction took place. This resulted in the formation of obsidian. Because open-system gas loss at slow ascent rates will result in equilibrium degassing trends, we conclude that, contrary to conventional views, slow ascent rates or degassing at or near the surface are not a requirement for obsidian formation.

Chromitite layers are common in large layered intrusions. One widely held view is that the chromitites formed as a consequence of the mixing of more primitive with less primitive magma (Irvine, 1977). This model predicts that the rocks immediately above and below the chromitite layers should be different. In the Bushveld Complex many chromitites are underlain and overlain by lithologically similar orthopyroxenite, suggesting that Irvine’s model does not apply. To understand how the Bushveld chromitites may have formed, we have conducted a detailed study of the pyroxenites above and below the platinum group element-rich UG2 chromitite of the upper and another chromitite of the lower Critical Zone (CZ). In the eastern Bushveld, the UG2 is a massive, 70 cmthick layer. Electron probe data indicate that the Mg# of opx ranges from 79.3–80.6 (En77.4–78.3) in footwall and 80.5–80.9 (En77.7–79.1) in hanging wall pyroxenite, while cpx compositions are En45.3–46.3 in the former and En45.3–48.9 the latter. Minor element compositions are variable, probably due to post-accumulation requilibration, but fall within the same range in each pyroxenite. In contrast, interstitial plagioclase of the lower pyroxenite is more sodic than that of the upper pyroxenite (An45.4 – 54.5 vs An64.9 - 72.5). This difference is not reflected by differences in Al/Si or Na/Ca ratios of coexisting opx or cpx. Furthermore, major and minor element compositions of all three minerals are identical in the rocks underlying and overlying the lower CZ chromitite. Field relations demonstrate that the UG2 acted as a permeability barrier to upwardly percolating interstitial melt. Therefore, we interpret the relatively sodic nature of plagioclase of the footwall pyroxenite to be due to the pooling of interstitial melt in this layer. Despite the difference in plagioclase compositions, the formation of the UG2 chromitite by magma mixing cannot be rationalized in terms of phase equilibria. A different model, such as the pressure release one of Lipin (1993), may better explain the Bushveld observations.

References

References

Newman, S., Epstein, S., Stolper, E.M., (1988), J. Volc. Geotherm. Res. 35, 75–96. Rust, A.C., Cashman, K.V., Wallace, P.J., (2004), Geology 32, 349–352. Watson, E.B., (1994), Rev. Mineral. 30, 371–411.

Irvine, T.N. (1977) Geology 5, 237-277. Lipin, B.R. (1993) J. Petrology 34, 955-976.

Goldschmidt Conference Abstracts 2005 Chemistry and Physics of Igneous Processes

A154

Effects of sulfur degassing and sulfide separation in some products of Mt. Etna volcano (Sicily, Italy) R. MORETTI 1, B. GAMBARDELLA2, L. MARINI2 3 AND N. MÉTRICH 1

INGV-Osservatorio Vesuviano, Napoli, Italy ([email protected]) 2 Lab. Geochimica, DipTeRis, Università di Genova, Italy 3 Lab. Pierre Sue, CEA/CNRS, Saclay, France Application of an internally-consistent model (Moretti and Ottonello, 2005) for melt oxidation state and sulfur solubility allows to study how Etna magmas release sulfur and how the redox control is operated in a volatile-rich basaltic environment. Different Etna magmas, corresponding to the products of the current activity and of the 122BC plinian eruption, follow different paths of S elimination and during their evolution they reach soon or later the lower limit of sulfide saturation. Interestingly, melt inclusion data indicate that along this threshold (i.e., for mS2- of 18 mmol/kg) the magma involved in the high-energy Plinian eruption of 122 BC is much richer in sulfate (and total S) than the magmas erupted by lowenergy Strombolian eruptions from the Bocca Nuova and SE crater. The direct relationship between the high sulfate contents at the lower limit of sulfide saturation and the energy of the eruption does not seem to be fortuitous. The availability of a relatively large amount of S, largely present as sulfate, in the magma of the 122 BC Plinian eruption is confirmed by the distinct 34S/32S isotopic ratios and total S concentrations measured in these products with respect to other rocks of the volcano. Again, this large amount of S, to be eliminated through degassing only, could be considered as a proxy for the high energy of this volcanic event, although further investigations are needed to corroborate this inference. Accepting this explanation, the next question is: why do Etna magmas experience different evolution paths leading to these large differences in sulfate contents at the lower limit of sulfide saturation?

Reference Moretti R. and Ottonello G. (2005) Geochim. Cosmochim. Acta, In press.

Identification of the hydrous environments in volcanic glasses R. PETRINI1, F.F. SLEJKO1, C. FORTE2 3 AND M. D’ANTONIO 1

Università di Trieste ([email protected], [email protected]) IPCF-CNR, Pisa ([email protected]) 4 Università di Napoli ([email protected]) 3

The hydrous species distribution in magmatic systems has considerable influence on the style of a volcanic eruption and on many of the physical and chemical properties of silicate melts. In particular, in the understanding of explosive volcanism a source of uncertainty is the process of fragmentation such as the rising magma disintegrates into a spray, resulting from the expansion of hydration bubbles through the progressive drainage of water from the surrounding melt by diffusion, or as a consequence of the interaction with external water. The water speciation and the nature of the solute in high-silica volcanic glasses from the explosive activity in the Phlegrean Volcanic District (Italy) has been determined by 1H MAS-NMR spectroscopy and Sr isotopic data by TIMS. 1 The H MAS NMR spectra are characterized by a large spinning sideband pattern with peaks spaced at integer multiples of the spinning speed, resembling a Pake doublet, with an overlapping central contribution. The spectra are best simulated by the contribution of three different types of proton 1 1 nuclei with H- H dipolar couplings of 45 kHz, 18 kHz and 1 kHz. These are attributed to isolated, rigid water molecule; to water molecules or clusters rather mobile and to micropore liquid water or clusters, respectively. Since the strontium of the solute in groundwater or surface water is likely to be 87 86 isotopically distinct from that of the melt, the Sr/ Sr ratio allowed to distinguish hydrous component of probably magmatic origin in isotopic equilibrium with the melt and hydrous components due to the interaction of magma with external fluids during phreatomagmatic fragmentation. Glasses in deposits from magmatic fragmentation show evidence of water diffusion in response to decreasing gas solubility during magma decompression and the gas partitioning into a separate phase.Glasses from phreatomagmatic fragmentation are characterized by the increasing of water more strongly bonded and less mobile into the structure. This requires an intimate, fine-scale mixing of water and magma during the water-magma interaction process, which is also promoted by shock wave propagation. This also suggests that water and magma were probably close to attain thermal equilibration during the interaction.

Goldschmidt Conference Abstracts 2005 Chemistry and Physics of Igneous Processes

How gabbro zircons contain more U than zircons from the co-mingled granodiorite: Lessons from U-Pb and Hf-zircon isotope investigations I. PEYTCHEVA1,2, A. VON QUADT2, M. FRANK2, N.GEORGIEV3, ZH. IVANOV3 AND CH. HEINRICH2 1

CLMC, BAS ([email protected]) 2 IGMR, ETH-Zurich ([email protected]) 3 FGG, Sofia University ([email protected]) Mantle derived magmas in convergent tectonic zones experience a complex evolution from their generation to the crystallisation in the upper crust. They commonly display large compositional variations due to the enrichment with incompatible chemical elements that are introduced into the upper mantle by slab-derived fluids or melts. This effect is strongly overprinted by the mantle-crust interaction. U-Pb and Hf- zircon isotope-geochronological and geochenmical investigations are crucial to evaluate the role of both processes as well as the kinematics of the mixing between mantle and crustal derived magmas. The present study is focused on Upper Cretaceous plutons from Central Srednogorie Zone, Bulgaria. Sheet-like gabbro or gabbro-diorite bodies are intruded into magma chambers containing coeval felsic upper parts and lower parts of crystal rich porphyry granodiorites. Granodiorite: concordant zircons, U content range from 70 to 161 ppm. Three of them determine a mean 206Pb/238U age of 84.6 ± 0.3 Ma. Inherited zircons are also present. The calculated initial (87Sr/86Sr) ratio of 0.70492 and ε-Hf-zircon values of +4.7 to +8.7 argue for mixed crust-mantle origin. The gabbro contains three types of zircon: (i) brown zircons are U-rich (600-4400 ppm), two of them lying concordant at 82.16 ± 0.10 Ma; ε-Hf-zircon values are mantle dominated +7.6 to +10.5; (ii) milky zircons are less rich in U (270-350 ppm) and define a mean 206Pb/238U age of 85.0 ± 0.5 Ma; (iii) colorless zircons of mixed origin and an age of 442.7 ± 8.3 Ma are sparse. The initial strontium ratio of the gabbro is 0.70401. Transparent and milky prismatic zircons prevail in the mixed layers. Five of them yield a concordant age of 84.87 ± 0.13 Ma. All Upper Cretaceous zircons reveal similar REE distributions with positive Ce and weak or absent negative Eu anomalies. Brown gabbroic zircons are slightly richer in the REE compared to the zircons from intermediate rocks. Mixing of the mantle magma (additional to the mingling) with crustal melt at mid- to upper crustal level is proposed to explain the change of the magma chemistry leading to zircon saturation and fast crystallisation of U- and REE-rich zircons.

A155

Element partitioning between ferrobasalt-rhyolite immiscible liquids I.V. VEKSLER1, J.K. JAKOBSEN2,3, A.M. DORFMAN4, L. DANYUSHEVSKY5, D.B. DINGWELL4 3 AND C.E. LESHER 1

GFZ Potsdam, Germany ([email protected]) University of Aarhus, Denmark ([email protected]) 3 UC Davis, California ([email protected]) 4 LMU Munich, Germany ([email protected]) 5 Univ. Tasmania, Australia ([email protected]) 2

Evidence from natural rocks show that liquid immiscibility can be encountered by common basalts at advanced stages of crystallization. However, the partitioning of trace elements between the immiscible liquids are still poorly constrained therefore we have carried out a comprehensive study of liquid-liquid element partitioning. In contrast to similar previous experimental studies, we employed high temperature centrifuge phase separation, in order to get a better spatial separation of immiscible liquids. Three thermocouples and two independent heaters were used to minimize temperature gradients. Two starting compositions were prepared from synthetic silicate glass and reagent-grade FeO. The first being rich in FeO (24 wt. %) and SiO2 (57 wt. %) in order to ensure liquid immiscibility to take place, the second starting composition had 50 % wt. SiO2 and 22 wt. % FeO close to liquid compositions suggested for the Skaergaard magma where liquid immiscibility has been speculated. These mixtures were doped with 33 trace elements. Runs were carried out at 1050-1150 oC in sealed Fe containers at 1 atm. Quenched products were analyzed by electron microprobe and LA ICP-MS. Two immiscible liquids are present in all the products separated by sharp meniscuses. One of these is rich in SiO2 (65 wt. %) whereas the other is having a moderate SiO2 content (48 wt. %) but is rich in FeO (32 wt. %). Even at the highest rotation speeds, the centrifuge phase separation after 3 hours was incomplete, mostly because of the high viscosity of the silica-rich immiscible liquids. Partition coefficients show that all the elements analyzed, except K, Na, Rb, Al and Si, concentrate in the Fe-rich immiscible liquid. The ferrobasalt/rhyolite Nernst partition coefficients (D) are the highest for Zn (3.3) and Fe (2.6) and the lowest for Rb and K (0.4-0.5). The D values against ionic potential shows in general a convex upward trend that closely resembles the liquid-liquid partitioning of trace elements by Soret diffusion. Complementary Soret diffusion experiments are underway to further explore bulk compositions controls on immiscibility in Skaergaard-like magma compositions.

Goldschmidt Conference Abstracts 2005 Reactive-Transport

A156

PHAST – A program for simulating ground-water flow, solute transport, and multicomponent geochemical reactions 1

1

DAVID L. PARKHURST , KENNETH L. KIPP , PETER ENGESGAARD2 AND SCOTT R. CHARLTON1 1

U.S. Geological Survey, Denver, CO, USA ([email protected], [email protected], [email protected]) 2 Geological Institute, University of Copenhagen, Copenhagen, Denmark ([email protected]) PHAST is a three-dimensional reactive-transport simulator derived from coupling the geochemical model PHREEQC with the solute-transport capabilities of HST3D. The flow and transport model is restricted to constant density, constant temperature, saturated ground-water flow. Chemical reactions include mineral and gas equilibia, ion exchange, surface complexation, solid solutions, and kinetic reactions. The simulator uses a point-distributed finite-difference grid and fixed-size time steps. After operator splitting, a sequential noniterative approach is used to solve the flow, transport, and chemical equations. The input and output features of the simulator were designed for ease of use. Data are input in a flexible, keyworddata-block format; data input for chemistry is identical to data input for PHREEQC. Model-output options are available to write any chemical property (for example, activity, concentration, numbers of moles of minerals) to ASCII files or a binary Hierchical Data Format (HDF) file. Data in the HDF file can be read by the plotting program Model Viewer for three-dimensional visualization and animation of heads, velocities, and chemical data. Three-dimensional reactive-transport calculations often require long computer processing times. A parallel version of the simulator has been developed to use multiple processors to shorten processing times. The parallel version is implemented with the Message Passing Interface (MPI), and can be used on multiprocessor computers or on a cluster of networked computers. PHAST has been used in several field-scale hydrologic studies. At Cape Cod, Massachusetts, simulatons indicate a diminishment of sewage-derived phosphorus to Ashumet Pond over a period of decades; in an Aquifer Storage Recovery (ASR) experiment in Charleston, South Carolina, simulations account for the changes in water chemistry caused by storage; and in Oklahoma, reactive-transport simulations were used to explain the chemical and transport processes that produce large dissolved arsenic concentrations in the Central Oklahoma Aquifer.

Geochemical modeling of decontamination solutions for building surfaces D. CRAIG COOPER, LAURENCE C. HULL AND KAREN E. WRIGHT INEEL Geosciences Research, Po Box 1625, Ms 2107, Idaho Falls, ID, 83415-2107, USA ([email protected]) A threat exists that a terrorist group may detonate a radiological dispersal device (RDD), generating a cloud of radioactive particles that could contaminate important surfaces. Rapid decontamination is needed to minimize the impact of such a terrorist action. Selection and implementation of a decontamination technique requires use of a model to evaluate applicability to a wide range of building materials. We have applied the geochemical reactive transport model, PHREEQc, to this problem. Geochemical principles commonly used to investigate reactions between ground water and geologic media are adopted and applied to interactions between decontamination solutions and building materials. Modeling is divided into mass transfer, and chemical mass transport. Explicit inclusion of mass transport is important because imbibition may occur when wet solutions encounter dry building surfaces, and the overall rate of decontamination may be limited by the diffusive transport of the chemical from the surface into the decontamination solution. The model is implemented in a finite-difference grid that incorporates a dual porosity approach and allows variable water chemistry and variable distribution of solid phases. Grids can be defined to represent the pore space of building materials, the building surface, and the decontamination foam. The transport problem is simplified to a 1-D scenario that envisions rapid initial transport into the building material (imbibition) and slow diffusive transport out of the building material. Radionuclides are distributed on the building surface in a variety of chemical forms including dissolved salts and particulate oxides. Decontamination solutions parameters are assigned to nodes in the finite difference grid, and the effect of foam-bubbles on transport can be tested by altering the effective diffusion coefficient(s). Mass transport between the building surface and the decontamination solution is based on diffusion along a concentration gradient. The model will be validated using data from laboratory experiments conducted on contaminated coupons of common building materials. Once validated, the model can be used to manipulate the chemical environment of decontamination solutions to evaluate alternative chemical recipes or to evaluate the effectiveness on different building materials.

Goldschmidt Conference Abstracts 2005 Reactive-Transport

Reactive transport modeling of acid gas generation and condensation

Hydrodynamic and thermodynamic modelling of the formation of the Yuzhna Petrovitsa hydrothermal Pb-Zn ore deposit, Madan, Bulgaria

GUOXIANG ZHANG, NICOLAS SPYCHER, ERIC SONNENTHAL AND CARL STEEFEL Lawrence Berkeley National Laboratory, Berkeley, CA, USA ([email protected]) Pulvirenti et al. (2004) recently conducted a laboratory evaporation/condensation experiment on a synthetic solution of primarily calcium chloride. This solution represents one potential type of evaporated pore water at Yucca Mountain, Nevada, a site proposed for geologic storage of high-level nuclear waste. These authors reported that boiling this solution to near dryness (a concentration factor >75,000 relative to actual pore waters) leads to the generation of acid condensate (pH <1.5) presumably due to volatilization of HCl (and minor HF and/or HNO3). To investigate the various processes taking place, including boiling, gas transport, and condensation, their experiment was simulated by modifying an existing multicomponent and multiphase reactive transport code (TOUGHREACT). This code was extended with a Pitzer ioninteraction model to deal with high ionic strength. The model of the experiment was set-up to capture the observed increase in boiling temperature (143°C at ~1 bar) resulting from high concentrations of dissolved salts (up to 8 m CaCl2). The computed HCl fugacity (~10-4 bars) generated by boiling under these conditions is not sufficient to lower the pH of the condensate (cooled to 80 and 25oC) down to observed values unless the H2O mass fraction in gas is reduced below ~10%. This is because the condensate becomes progressively diluted by H2O gas condensation. However, when the system is modeled to remove water vapor, the computed pH of instantaneous condensates decreases to ~1.7, consistent with the experiment (Figure 1). The results also show that the HCl fugacity increases, and calcite, gypsum, sylvite, halite, MgCl2.4H2O and CaCl2 precipitate sequentially with increasing concentration factors. Figure 1. pH of the condensate

A157

T. DRIESNER1, B. KOSTOVA2 AND C.A. HEINRICH1 1

Isotope Geochemistry and Mineral Resources, ETH Zentrum, Zürich ([email protected], [email protected]) 2 Sofia University “St. Kliment Ohridsky”, Sofia, Bulgaria Coupled fluid flow and speciation computations in reactive transport modelling are feasible on normal workstation computers for low temperature environments, because of their simple (if any) P-T dependence of fluid properties. Realistic modelling of high-temperature ore forming processes, however, requires orders of magnitude larger computational resources because of the much more complex equations governing multiphase flow of compressible fluids with highly nonlinear P-T-X dependencies of density, viscosity, and thermochemical properties of the solvent and dissolved minor components. This has severely limited the application to ore-forming hydrothermal systems using a realistic degree of geometric as well as chemical complexity. Here we demonstrate how a careful combination of fluid inclusion-derived thermal and chemical data (Kostova et al., 2005) with hydrodynamic simulations provides a framework that allows decoupling of the hydrodynamic and speciation calculations. The data set comprises ore fluid compositions derived from LA-ICPMS analyses of fluid inclusions from early to late mineralization stages over a vertical interval of 500 m in the main vein of the Oligocene Yuzhna Petrovitsa Pb-Zn-Cu deposit in Bulgaria. Fluid inclusion data and fS2 and pH constraints from mineral buffers are the main input to a moderate number of stand alone speciation calculations. The latter provide a full quantitative picture of the chemical evolution and predict the correct depth distribution of Pb precipitated as galena in response to cooling of a low salinity, >320°C input fluid with less than 10 ppm Pb. Precipitation and a transient dissolution stage seen in the ore textures are correctly predicted from the combination of this chemical modelling with the physical modelling of advective fluid and heat transport within the vertically extensive vein system.

Reference Kostova B., Pettke T., Driesner T., Heinrich C.A. and Petrov, P. (2005), Schweizer. Miner. Petr. Mitt., in press

Reference Pulvirenti, A.L., K.M. Needham, M.A. Adel-Hadadi, A. Barkatt, C.R. Marks, and J.A. Gorman, (2004), Corrosion 2004, NACE International, New Orleans.

Goldschmidt Conference Abstracts 2005 Reactive-Transport

A158

Geochemical transport modeling of a phreatic drinking water pumping station B. VAN DER GRIFT1 AND J. GRIFFIOEN2 1

ZHENHAO DUAN, RUI SUN AND JIAWEN HU

([email protected]) ([email protected])

2

We developed a regional scale 3-dimensional multicomponent reactive transport model to predict the groundwater quality at a phreatic drinking water production site. PHT3D (Prommer et al., 2003) was used as model code. PHT3D couples the 3-dimensional transport simulator MT3DMS with the goechemical model PHREEQC-2. The study examined the influence of various land use functions on the quality of the abstracted groundwater as well as the occurrence of various type of geochemical processes. The following processes were incoporated in the model: nitrate reduction, aqueous complexation, cation exchange, dissolution and precipitation of minerals (carbonates, hydroxides and pyrite). The model is filled with local determined geochemical and hydrochemical data. The redox reactivity of the sediments is derived from sensitive respiration experiments. Time series of groundwater quality data were available for validation. Figure 1 shows the modeled and measured NO3 concentration of the abstracted groundwater. The scenarios predicts the concentration with and without measures taken in 1990 to reduce the manure use after a period of almost 2 decades of intensive manure spreading. 70

NO3 (mg/l)

60 50

autonomous w ithout measures monitoring

40 30 20 10 0 1950

1970

A model for calculating the solubility of gases (CO2, H2S,…) used for the sequestration of global warming gases

1990

year

2010

2030

2050

Figure 1: modeled and measured NO3 concentration of abstracted groundwater

Reference Prommer, H., Barry, D.A. & Zheng, C., [2003]. MODFLOW/MT3DMS based reactive multi-component transport modeling. Ground Water 42 (2).

Chinese Academy of Sciences, Institute of Geology and Geophysics, Beijing, 100029, China ([email protected]) A thermodynamic model for calculating the solubility of gases (CO2, H2S, N2, CH4, C2H6) in pure water and in salt solutions in a wide temperature, pressure and ionic solutions is presented. Although parameterized from the gas-H2O-NaCl systems, the model accurately predicts gas solubilities in many other aqueous salt systems, such as seawater, without fitting data from these systems. This model finds wide applications, such as in the study of gas sequestration, and in rock-fluid interactions.

Goldschmidt Conference Abstracts 2005 Cosmogenic Nuclides and Surface Processes

Improving the accuracy and precision of scaling factors for in-situ cosmogenic geochronometers: New measurements of cosmic-ray neutrons in India and Hawaii DARIN DESILETS AND MAREK ZREDA Department of Hydrology and Water Resources, University of Arizona, Tucson, Arizona, 85721, USA ([email protected], [email protected]) Production rates of in-situ cosmogenic nuclides are determined by the intensity of energetic cosmic-ray nucleon fluxes, which is highly dependent on elevation. An incomplete knowledge of how nucleon fluxes vary with elevation remains a major obstacle to utilizing cosmogenic nuclides as geochronometers in applications requiring small time resolution. One problem is that attenuation characteristics of nucleon fluxes depend on nucleon energy. Measurements of high-energy (>50 MeV) fluxes tend to give shorter attenuation lengths than low-energy (<1 MeV) fluxes, but these differences are not well characterized due to a lack of data at lower energies. Another problem is that the elevation dependence varies with geomagnetic cutoff rigidity (a parameter related to geomagnetic latitude), RC, and that there has been an incomplete mapping of neutron fluxes at high RC (low latitude). We report new measurements of neutron fluxes from altitude transects in Hawaii (RC=12.8 GV) and Bangalore, India (RC=17.2 GV). Our measurements in Hawaii of low-energy neutrons (median energy 1 eV) and energetic nucleons (median energy 140 MeV) confirm that nucleon scaling functions are energy dependent in the range of energies at which cosmogenic nuclides are produced. Our measurements in India extend our previously reported scaling model for spallation reactions (Desilets and Zreda, 2003, Earth and Planetary Science Letters 206, 21-42) from RC=13.3 GV to RC=17.2 GV, nearly the highest modern cutoff rigidity on earth. The anomalously high cutoff rigidity over India provides a geomagnetic shielding condition that is effectively the same as would be observed at the geomagnetic equator in a dipole field with an intensity 1.2 times the modern value. This makes it possible to scale low-latitude production rates to paleomagnetic fields that are stronger than the present dipole field.

A159

Calibrating the production rate of cosmogenic 36Cl from postglacial lava flows in Iceland C.L. DENONCOURT1, J.M. LICCIARDI1, J.O. STONE2 3 AND R.C. FINKEL 1

Department of Earth Sciences, University of New Hampshire, Durham, NH 03820, USA ([email protected], [email protected]) 2 Quaternary Research Center and Department of Earth and Space Sciences, University of Washington, Seattle 98195, USA ([email protected]) 3 Center for Accelerator Mass Spectrometry, Lawrence Livermore National Laboratory, Livermore, CA 94550, USA ([email protected]) One large uncertainty in surface exposure dating commonly arises from incomplete knowledge of production rates of cosmogenic isotopes and their variation with latitude, altitude and time. In Iceland, the production rates of in situ cosmogenic isotopes are predicted to be higher than globally averaged rates due to persistent low atmospheric pressure (Stone, 2000). Licciardi and Kurz (2002) determined that 3He production rates in Iceland are significantly higher than normalized values measured in the western United States (Licciardi et al., 1999), supporting Stone’s prediction. We measured cosmogenic 36Cl concentrations in whole rock basalts from radiocarbon-dated lava flows in Iceland. The 36Cl data were obtained from splits of rock material from the same calibration sites used by Licciardi and Kurz (2002) to calculate their 3He production rates. These 36Cl measurements thus allow for the first direct co-calibration between these two widely used cosmogenic isotopes. An immediate application of our 36Cl production rates is to enable accurate 36Cl dating in Iceland (e.g., Principato et al., 2003). Moreover, the 3He-36Cl co-calibration will further elucidate the influence of atmospheric pressure on the production rate of in situ cosmogenic isotopes.

References Licciardi, J.M., Kurz, M.D., Clark, P.U., and Brook, E.J. (1999) Earth and Planetary Science Letters 172, 261-271. Licciardi, J.M., and Kurz, M.D. (2002) Geochimica et Cosmochimica Acta 66, A456. Principato, S.M., Geirsdóttir, A., Andrews, J.T., and Johannsdóttir, G. (2003) GSA Abstracts with Programs 35, no. 6. Stone, J. (2000) Journal of Geophysical Research 105, 23,753-23,759.

A160

Goldschmidt Conference Abstracts 2005 Cosmogenic Nuclides and Surface Processes

Grain size dependency of 10Be concentrations in alluvial sediments in the Great Smoky Mountains

Cosmogenic helium and neon extracted by crushing: A technique for discriminating between mantle and cosmogenic helium

A. MATMON1, P. R. BIERMAN2, J. LARSEN2, S. SOUTHWORTH3, M. PAVICH3, R. FINKEL4 4 AND M. CAFFEE

MANUEL MOREIRA1 AND PEDRO MADUREIRA1,2 1

Institut de Physique du globe de Paris, Université paris VII, CNRS UMR 7579, 4 place Jussieu, 75005 Paris, France ([email protected]) 2 Centro de Geofísica de Évora/Departamento de Geociências da Universidade de Évora, Rua Romão Ramalho, 59, 7000-671 Évora, Portugal The helium and neon isotopic compositions of olivines coming from a 11Ma old xenolith sampled at Mt. Hampton (West Antarctica) were analyzed by crushing and heating. The 4 He/3He isotopic ratio obtained by crushing varies between 1340 and 6300 (R/Ra between 115 and 539) with 4He content around 3-5 10-10 ccSTP/g confirming that cosmogenic helium can be extracted by crushing [Scarsi, 2000; Yocochi et al., 2004]. The neon also shows a clear cosmogenic origin (20Ne/22Ne down to 7.7 and 21Ne/22Ne>0.32) indicating that some cosmogenic neon can also be extracted by crushing out of the olivines. This result indicates that for samples that had been exposed for a long time (e.g. few Ma to Ga), a step crushing procedure may not give the mantle ratios without ambiguity and that measurement of neon can discriminate between cosmogenic and mantle origin of the 3He. Melting of the powder left after the crushing experiment gives 4He/3He ratio as low as 51±5 (R/Ra=14 230) and 21Ne/22Ne as high as 0.78, close to the cosmogenic end-member. Our results show that ~0.4% of the cosmogenic helium and ~0.3% of the cosmogenic neon can be extracted out of olivines by crushing

References Scarsi, P., Geochim. Cosmochim. Acta, 64(21), 3751-3762, 2000. Yokochi, R., B. Marty, R. Pik, and P. Burnard, G3, in press, 2004.

1

United States Geological Survey, Menlo Park, CA 94025, USA 2 Geology Department, University of Vermont, Burlington, VT 05405, USA 3 United States Geological Survey, Reston, VA 20192, USA 4 Lawrence Livermore National Laboratory, Livermore, CA 94550, USA Analysis of multiple grain-size fractions from alluvial sediment samples in the Great Smoky Mountains (GSM), show, in five of the six samples tested, higher nuclide concentrations and by inference, slower model erosion rates, in smaller grain sizes than in larger ones. 10Be concentration in the < 2 mm fractions correlate to erosion rates that range between 25±3 and 50±6mm ky-1. in contrast, erosion rates 20%-40% higher are calculated for the >2 mm fractions in each sample. Field evidence for mass wasting is minimal, therefore, differences in cosmogenic nuclide concentrations between grains of different sizes cannot be explained by differences in transport mechanism. We interpret the difference in concentrations as a result of the large elevation distribution of the source and longer exposure periods on the slopes for the smaller grains compared with the narrow and relatively low source elevation of the large grains and their shorter exposure history. Large sandstone clasts disaggregate into sand-size grains rapidly during down slope transport so only clasts from the lower parts of slopes reach the streams. A positive correlation between maximum relief in the basin and the difference in normalized 10Be concentrations in the different grain size fractions suggests that our explanation is valid. We use the sampling location production rates to calculate erosion rates from 10Be concentrations in the larger clasts. When site production rates are used, large grain size fractions yield erosion rates that range between 18±2 and 45±6 mm ky-1, similar to those calculated from the small grain size fractions. These results support our assertion that clasts are derived from the lower parts of the slopes, that clasts are not transported long distances downslope, and that different grain sizes are generated at similar rates in the GSM.

Goldschmidt Conference Abstracts 2005 Cosmogenic Nuclides and Surface Processes

Perspectives on dating with multiple cosmogenic nuclides

Preservation of (Early) Miocence landscapes in the Atacama Desert, northern Chile

DARRYL E. GRANGER Department of Earth and Atmospheric Sciences, Purdue University, West Lafayette, IN 47907, USA ([email protected]) The method of dating sediment burial using the differential decay of 26Al and 10Be in quartz has been used since the earliest days of terrestrial cosmogenic nuclide measurements. Techniques were readily adapted from the meteoritic field, where multiple cosmogenic nuclides were routinely measured to date complex histories of exposure and shielding in space and on earth. However, because the 26Al10 Be pair is sensitive only on a million-year timescale, with uncertainties of 10’s to 100’s of thousands of years, it initially proved to be of limited use for dating earth surface processes, which tend to occur much more rapidly. More recently, a flurry of new applications has developed to date rock and sediment burial over the past 5 My, providing dates to investigate long-term river incision, marine terrace uplift, glacial histories, and biologic evolution, as well as measurements to investigate paleo-erosion rates. I will present examples of these applications from published work as well as from marine terraces and caves of central Italy. Accurate burial dating with 26Al and 10Be is hampered by a large discrepancy in the half-life of 10Be. Although most published 10Be measurements implicitly assume a half-life of 1.5 My, an alternative measurement of 1.34 My was derived from the same parent solution; these values have never been resolved. A compilation of surface exposure data from the literature reveals that 26Al/10Be ratios near saturation are consistent with either half-life, but that they are more closely described by the shorter half-life. Ongoing and future work with different mineral systems may employ different cosmogenic radionuclides. For example, feldspar and quartz in the same rock can be analyzed for 36Cl, 26 Al, and 10Be to reveal more complex or recent exposure and burial histories. Carbonate rocks may also prove useful for the 36 Cl-10Be pair. Preliminary results from a Roman marble quarry show that 10Be can be measured to high precision in this rock, although care must be taken to avoid secondary calcite which may contain large amounts of meteoric 10Be, even many centimeters beneath the surface.

A161

T.J. DUNAI1, G.A.GÓNZALEZ-LÓPEZ2, J. JUEZ-LARRÉ1 2 AND D. CARRIZO 1

Faculty of Earth and Life Sciences, Vrije Universiteit, Amsterdam, Netherlands ([email protected]; [email protected]) 2 Departamento de Ciencias Geológicas, Facultad de Ingeniería y Ciencias Geológicas, Universidad Católica del Norte, Antafogasta, Chile ([email protected]) Depositional surfaces of early Miocene sediments surfaces are preserved in the Coastal Cordillera, Atacama Desert, northern Chile. Measurement of cosmogenic 21Ne in clasts from erosion-sensitive sediment surfaces show that these surfaces have been barely affected by erosion since 25 Ma. Predominantly hyperarid conditions since 25 Ma are required to create and preserve these oldest continuously exposed surfaces on Earth. The next oldest continuously exposed surfaces, in the Dry Valleys region, Antarctica, have about half this age. Occurrence of younger exposure ages indicate that brief pluvial episodes occurred since the Early Miocene did occur, which caused limited, localized erosion and material transport, only marginally affecting the large scale landscape. We present new data from other, similarly old surfaces, from the coastal portion of the Atacama Desert. These data demonstrate that the exceptional landscape stability in this coastal desert is widespread, as would expected from the large scale regional factors controlling climatic conditions in this area. The dominantly hyper-arid conditions we infer for the Coastal Cordillera since ~25 Ma ago are compatible with the hypothesis that the onset of aridity in the Atacama Desert is the cause, rather than the result of the uplift of the high Andes. Only exceptional global climatic disturbances have occasionally permitted humidity transfer across the Andes into the driest regions of this Coastal Desert since ~25 Ma.

References Dunai T.J., González López G.A. and Juez-Larré J., (in press), Geology. Lamb, S. and Davis P., (2003): Nature, 425, 792-797. Schäfer, J.M. et al., (1999) Earth Planet. Sci. Lett. 167, 215226. van der Wateren, F.M. et al., (1999) Global Planetary Change 23, 145-172.

Goldschmidt Conference Abstracts 2005 Cosmogenic Nuclides and Surface Processes

A162

Constraining landform erosion and ages from surface exposure age distributions on old Patagonian moraines R.P. ACKERT AND S. MUKHOPADHYAY Department of Earth and Planetary Sciences, Harvard University ([email protected]) Applying surface exposure dating to moraines older than the last glacial maximum is challenging. For these older moraines, boulder erosion and landform degradation become important, resulting in a wide distribution of cosmogenic nuclide concentrations. Extracting meaningful exposure ages from these moraines requires accurate erosion and exhumation rates. To better understand the evolution of exposure age distributions with time, we have started an investigation of cosmogenic 3He and 36Cl from a large suite of basalt boulders collected from the Telken IV (760-1016 ka; [1]) and Deseado (109-760 ka; [1]) moraines at Lago Buenos Aries (LBA), Argentina. The moraines at LBA are ideal for this study because they are the longest, best preserved glacial record outside of Antarctica. Surface exposure dating has established that the youngest moraine complex at LBA ranges in age from 16 to 26 ka (stage 2) and records millennial scale fluctuations of the ice margin during the LGM [2]. Exposure ages on the next oldest moraine complex suggest that they date to the penultimate glaciation (stage 6) but the effects of boulder erosion and exhumation are evident with many young outliers in the exposure age distribution. We have obtained 3He exposure ages of 178, 190, and 133 ka on three boulders from the Telken IV moraine, which are similar to ages obtained using 10Be and 26Al on nearby granitic boulders of older moraines (Telken V and VII moraines [2]). Our 3He exposure ages calculated with an erosion rates of 2mm/kyr are almost twice as old, 287 and 326 ka, but still significantly less than the minimum age of the Telken IV moraine, clearly indicating the importance of boulder exhumation and erosion. Using our preliminary data, the maximum deflation rate of the moraine surface is 7 mm/kyr. We plan to obtain additional 36Cl and 3He data to constrain the process of boulder exhumation and erosion and thereby 1) improve the chronology of glacial history in Patagonia, 2) understand landform development and preservation in the dry Patagonian steppe, and 3) infer landform age from exposure age distributions in areas where independent chronology is not available.

References

[1] Singer, B. S., et al., GSA Bull 116, 434, 2004. [2] Kaplan, M. R. et al., GSA Bull 116, 308, 2004.

3

He exposure ages of boulder armored terraces in the northwestern Colorado Plateau DAVID W. MARCHETTI1, THURE E CERLING1 2 AND JOHN C DOHRENWEND 1

Department of Geology, U. of Utah, 135S 1460E Room 719, SLC, UT 84112, USA ([email protected]) 2 Southwest Satellite Imaging, 223 South State, Teasdale, UT 84773-0141, USA In the Capitol Reef and Escalante areas of the northwestern Colorado Plateau (south central Utah) there are tens to hundreds of individual basaltic-andesite armored terraces throughout the landscape. These armored terraces are former valley floors that are now 10-200 m above the local drainages. All of these armored terraces are capped with coarse basaltic-andesite boulder deposits derived from the high (>3400 m) volcanic plateaus of Boulder or Thousand Lakes Mountains. Using 3He exposure age dating we determined the exposure ages of multiple boulders from several of these terraces. We interpret these deposits to be proximal debris-flows and therefore assume deposition was rapid and do not include a correction for cosmogenic inheritance due to transport. We do include a correction for non-cosmogenic (nucleogenic) 3He produced in the basalticandesites since crystallization (~25 Ma). This component is typically less than 8% of the total 3He inventory. Maximum boulder exposure ages of these land-surfaces range from 1.2 Ma to 196 ka and represent average local incision rates ranging from ~0.15 m/kyr to 0.40 m/kyr. The incision rates we calculate are some of the highest on the Colorado Plateau and add to a growing body of evidence suggesting that there was significant fluvial incision of the Plateau during the Pleistocene.

Goldschmidt Conference Abstracts 2005 Cosmogenic Nuclides and Surface Processes

Glacial erosion and till dispersion using the source and the sink: A new cosmogenic nuclide application JANE WILLENBRING STAIGER AND JOHN C. GOSSE Dept. of Earth Sciences, Dalhousie University, Halifax, Nova Scotia, Canada ([email protected], [email protected]) Landscape morphodynamics can be recorded by both the bedrock surface and the deposits removed from the bedrock. In this work, we analyze both the source – the bedrock – and the sink – the till – to decipher the glacial system. Terrestrial in situ cosmogenic nuclides (TCN) extracted from glaciated bedrock surfaces can be used to assess the relative stability of each part of the landscape, thus to determine the spatial variation of glacial erosion beneath polythermal ice. We adopt a field-constrained glacial erosion rule for the Torngat Mountains of northern Labrador for use in Baffin Island, which occupies similar setting. This rule correlates modeled ice velocities with TCN-derived erosion rates from bedrock for terrain once partially covered by slowmoving, non-erosive ice (Staiger et al., 2005). Because the spatial variation of cold-based, non-erosive ice versus wet-based, erosive ice is recorded by the bedrock TCN concentrations, each individual grain within the till has a unique history as it was previously exposed to cosmic radiation then transported by ice. Together the grains give an areal average of the basal thermal regime of the ice that entrained the sediment. In this new method, we measure 10Be and 26Al concentrations in 25 surface till samples from Baffin Island and Labrador and can distinguish specific sediment packages. Till samples that were predicted by geomorphological context to have been deposited by nonerosive ice contain over 100 times the TCN contentration of till thought to be deposited by highly erosive ice. We interpret these data using a finite-element, time-dependent ice sheet model that includes basal temperature and basal water calculations (Johnson and Fastook, 2003) and a forward model that calculates possible TCN concentration scenarios. We estimate ranges of englacial traveling distances for “short-distance tills” with high TCN values and “long distance tills” with low TCN values. This work has potential applications in arctic and sub-arctic drift exploration as well as a potential for assessing the effect of glacial erosion on a glaciated landscape with fewer samples.

References Johnson, J. and Fastook, J. (2002) Quat. Int. 95-96: 87-98. Staiger, J.W., Gosse, J.C. Johnson, J., Fastook, J., Stockli, D., Stockli, L., and Finkel, R. (2005) accepted to ESPL.

A163

Five ways to examine what isn’t there with cosmogenic isotopes J.C. GOSSE1, S. BAKER1, F. PAZZAGLIA2, M. BRANDON3, K. KARLSTROM4, J. PEDERSON5 AND R. FINKEL6 1

Dept. Earth Sciences, Dalhousie U., Halifax, NS, Canada ([email protected]) 2 Dept. Earth & Env. Sci, Lehigh U., Bethlehem, PA, USA 3 Dept. Geology, Yale U., New Haven, CT, USA 4 Dept. Earth & Planet. Sci., U. New Mex., Albequerque, NM, USA 5 Dept. Geology, Utah State U., Logan, UT, USA 6 CAMS, Lawrence Livermore National Lab, Lawrence, CA, USA Terrestrial in situ cosmogenic nuclides (TCN) have been used in five ways to characterise the rates of exumation or incision by streams: (1) a maximum erosion rate can be calculated by measuring one or two isotopes on a bedrock surface; (2) basin-wide average erosion rates can be calculated with one isotope measured in stream sediment; (3) escarpment retreat rate can be attained by measuring remnents of the retreating cliff; (4) stream incision rates into bedrock can be attained by dating straths; and (5) vertical incision rates can be estimated from measurements of a single isotope down a near vertical wall of a canyon. The application of each approach is demonstrated and placed in a regional context. Erosion of Archean gneiss summits in the Torngat Mountains of Labrador, Canada are as low as 1.6±0.3 m/Myr. Basin-wide average erosion rates in the Clearwater Catchment, Olympic Mtns, Washington State range from 0.4±012 mm/yr in high relief regions to 0.2±0.1 mm/yr in low relief regions over the past 104 yrs, in close agreement with similar timescale strath incision rates, longer timescale thermochronology, and shorter timescale stream sediment discharge. Retreat rate of the Morrison Formation escarpment in NE Arizona has been measured at 15±5 mm/yr (n=3) in one basin and 2.5±0.5 mm/yr (n=2) in an adjacent basin, consistent with estimates from soils and other measurements. Chronology of mapped fill terrace surfaces in the eastern Grand Canyon provide constraints on the age of the underlying straths that are consistent with U-series and OSL ages of the same fills, and yield an average incision rate by the Colorado River of 100150 m/Myr. Ages on straths in the Rio Diamante, Mendoza, Argentina have provided constraints on incision through a Quaternary fold (the initiation of an antecedent stream). Rates of incision based on single nuclide measurements in canyon wall bedrock samples within 20 m above 5000 cfs stage, average 0.08±0.01 mm/yr (n=3) for the past 20 kyr to more than 3 mm/yr on higher surfaces. Higher surfaces lacked fluvial polish.

Goldschmidt Conference Abstracts 2005 Cosmogenic Nuclides and Surface Processes

A164

81

Kr-dating: From dream to practice

News from the oldest ice on Earth buried in Antarctica, and a new cosmogenic tool

Z.-T. LU Physics Division, Argonne National Laboratory, Dept. of Physics and Enrico Fermi Inst., Univ. of Chicago ([email protected]) Since radiocarbon dating was first demonstrated in 1949, the field of trace analyses of long-lived cosmogenic isotopes has seen steady growth in both analytical methods and applicable isotopes. The impact of such analyses has reached a wide range of scientific and technological areas. A new method, named Atom Trap Trace Analysis (ATTA), was recently developed and used to analyze 81Kr (t1/2 = 2.3×105 years, isotopic abundance ~ 1×10-12) and 85Kr (t1/2 = 11 years, isotopic abundance ~ 1×10-11) in environmental samples. 81Kr is produced by cosmic rays in the upper atmosphere. It is the ideal tracer for dating ice and groundwater in the age range of 104–106 years beyond the reach of radiocarbon dating. On the other hand, analyses of 85Kr, a fission product of uranium and plutonium, can serve as a means to help verify compliance with the Nuclear Non-Proliferation Treaty as well as dating young groundwater. In ATTA, individual atoms of the desired isotope are selectively captured into a laser-based atom trap and detected by observing the fluorescence of trapped atoms. As the first real-world application of ATTA, the mean residence time of the old groundwater in the Nubian Aquifer located underneath the Sahara Desert was determined. With this demonstration and further improvements in the ATTA method, wide spread use of 81Kr-dating in Earth sciences seems feasible. This work is supported by the U.S. Department of Energy, Office of Nuclear Physics, under contract W-31-109-ENG-38, and by the U.S. National Science Foundation grant EAR-0126297.

References Project website: http://www-mep.phy.anl.gov/atta/ C.Y. Chen et al., Science 286, 1139 (1999) X. Du et al., Geophysical Research Letters 30, 2068 (2003). N. C. Sturchio et al., Geophysical Research Letters 31, L05503 (2004)

J. M. SCHAEFER1, C. SCHLUECHTER2, R. WIELER 3, S. IVY-OCHS3, P. KUBIK3, D. MARCHANT4, G. KORSCHINEK5, K. KNIE5, G. HERZOG6 6 AND F. SEREFIDDIN 1

Lamont-Doherty Earth Observatory, Palisades ([email protected]) 2 University of Berne, Berne Switzerland 3 ETH Zuerich, Zuerich, Switzerland 4 Boston University, Boston, USA 5 TU Munich, Munich, Germany 6 Rutgers University, Piscataway/New Brunswick Methodological progress of Terrestrial Cosmogenic Nuclides (TCN) and their applications to forefront earth scientific problems have been surging over the last 10 years. We present case-studies to illustrate the potential of dating earth surface processes over a time-scale spanning from thousands to millions of years. As an example of the rapid methodological progress, we report very recent measurements of in-situ cosmogenic 53Mn, a new member of the TCN family. We give an update of our current research regarding the age, formation, and climate significance of the oldest ice on Earth in Beacon Valley, Antarctica, based on new cosmogenic noble gas data. Although there is general agreement that the buried ice bodies in the Dry Valleys represent potentially important climate archives well beyond the ice-core timerange, the formation mechanism and age of these features is still uncertain. We used the same samples from Antarctica to perform the first successful measurements of terrestrial 53Mn. The consistency between in-situ cosmogenic 53Mn and cosmogenic noble gas data is striking and allows a first quantification of the production rate of terrestrial 53Mn. We give a protocol to apply the new cosmogenic nuclide together with an overview of advantages, current limitations and potential improvements. Finally, we give a short overview of the existing limitations of the TCN method and strategies to go beyond these limits. Such strategies are the core of the international multi-group CRONUS-Earth initiative, which will be introduced in detail in a companion poster.

Goldschmidt Conference Abstracts 2005 Cosmogenic Nuclides and Surface Processes

CRONUS-EU Cosmic ray produced nuclide systematics – The European contribution T.J. DUNAI1 AND THE CRONUS-EU SCIENTIFIC TEAM2 1

Faculty of Earth and Life Sciences, Vrije Universiteit, Amsterdam, Netherlands ([email protected]) 2 Ten network teams throughout Europe The main objective of the CRONUS-EU is to advance Terrestrial cosmogenic nuclide (TCN) techniques into a robust tool for Earth surface and environmental sciences. CRONUSEU aims to achieve this goal via: (1) High quality calibration of TCN production rates at independently dated surfaces (2) High quality calibration of TCN production rates using artificial targets (3) Systematic cross calibration of production rates of different TCNs (4) Refinement of scaling factors that describe the spatial and temporal variation of the cosmic ray flux relevant for TCN production using calibration measurements and numerical modeling from physical principles (5) Reducing the uncertainty of decay constants (6) Establishing the use of additional mineral phases in exposure age dating (7) Improvement and standardization of chemical routines (8) Laboratory cross calibrations (9) Training of young researchers and the user community The effort necessary to achieve above goal is significant even for the strong network teams in CRONUS-EU. To strengthen our effort and to achieve international evaluation and acceptance, we are seeking close collaboration with CRONUS-Earth, the parallel-running northern American sister initiative that obtained funding through NSF. Formal links between the two initiatives are established and each consortium will address complementary aspects to achieve the common goal. The coordination of CRONUS-EU is at the VU Amsterdam, with network teams at SUERC, Scotland; CEREGE and CRPG, France; ETH-Zürich (2 teams), Switzerland; Univ. Bratislava, Slovakia; Univ. Hannover, TUMunich and GFZ-Potsdam, Germany; and Utrecht University, The Netherlands (see also: www.cronus-eu.net). CRONUS-EU is a Marie-Curie Research and Training program, supported by the European Community's Program: Improving the Human Research Potential and the Socioeconomic Knowledge base.

A165

Interpreting cosmogenic nuclide concentrations in areas with complex exposure-burial histories under ice sheets: How sensitive are results to variations in the ice cover proxy curve? Y.K. LI1, J. HARBOR1, D. FABEL2 AND A.P. STROEVEN 3 1

Department of Earth and Atmospheric Sciences, Purdue University, West Lafayette, IN 47907-2051 ([email protected]; [email protected]) 2 Research School of Earth Sciences, Australian National University, Canberra, ACT, 0200, Australia ([email protected]) 3 Department of Physical Geography and Quaternary Geology, Stockholm University, S-106 91 Stockholm, Sweden ([email protected]) Phases of accumulation and decay of cosmogenic radionuclides in rock surfaces subject to episodes of exposure and burial by ice result in present-day nuclide concentrations that reflect the timing of initial exposure and the chronology of subsequent exposure, burial and erosion. Assuming no erosion, and using ice core or marine isotope records as proxies for the timing and duration of periods of ice cover, it is possible to constrain the timing of initial exposure and the number of phases of exposure and burial a rock surface has been subjected to using multiple cosmogenic radionuclide However, in concentrations (typically 10Be and 26Al). evaluating interpretations based on this approach, it is important to assess how sensitive the results are to the ice cover proxy curve. We have developed a program to evaluate variations in total exposure and burial duration as a function of different proxy curves and assumptions of cutoff values for ice free conditions. Initial results for northern Sweden and Antarctica indicate a highly variable pattern of sensitivity (step changes in results at critical ice cover / ice free cutoff values), and provide new insight into how to determine the level of reliability of calculated initial exposure dates.

Key words cosmogenic nuclides; ice sheet; proxy climate curve; surface exposure dating

A166

Goldschmidt Conference Abstracts 2005 Cosmogenic Nuclides and Surface Processes

Solar modulation and scaling in situ cosmogenic nuclide production rates

Dating alluvial sediments with cosmogenic nuclides

N. A. LIFTON1, J. W. BIEBER2, J. M. CLEM2, M. L. DULDIG3, P. EVENSON2, J. E. HUMBLE4 2 AND R. PYLE

W.M. PHILLIPS1, L.S. BARHAM2 AND P.W. KUBIK3

1

Geosciences Department, University of Arizona, Tucson, AZ, 85721, U.S.A. ([email protected]) 2 Bartol Research Institute, University of Delaware, Newark, DE, 19716, U.S.A. ([email protected], [email protected], [email protected], [email protected]) 3 Australian Antarctic Division, Kingston, Tasmania, Australia 7050 ([email protected]) 4 School of Mathematics and Physics, University of Tasmania, Private Bag 37, Hobart, Tasmania, Australia 7001 ([email protected]) Solar modulation affects in situ cosmogenic nuclide (CN) production rates the most at the high geomagnetic latitudes to which those production rates are traditionally referenced. This variability leads to significant scaling model uncertainties that have not been addressed rigorously to date. We have developed new CN production rate scaling models for spallogenic nucleons, slow muon capture and fast muon interactions that specifically address these uncertainties. Our spallogenic nucleon scaling model, which includes data from portions of 5 solar cycles, explicitly incorporates a measure of solar modulation, and our fast- and slow-muon scaling models (based on more limited data) account for solar modulation effects through increased uncertainties. These models improve on previously published models by better sampling the observed variability in measured cosmic ray intensities as a function of geomagnetic latitude, altitude, and solar activity. Our results show that we can accurately account for the effects of solar modulation on measured cosmic ray intensities with our models, within the uncertainties of each of our source datasets. Published spallogenic nucleon scaling models predict scaling factors ranging from ~15% below to ~30% above those of our spallogenic model, while published muogenic scaling models predict scaling factors up to ~90% above ours. We also estimate solar modulation variations over the last 11.4 ka from a recent sunspot number reconstruction based on treering 14C data. These data suggest that spallogenic scaling factors in our model for sea level and high geomagnetic latitudes can vary by up to ~10%, depending on the time period over which the modulation conditions are averaged. The potential magnitude of this variation supports our contention that incorporating long-term solar modulation into CN production rate scaling is important.

1

Idaho Geological Survey, University of Idaho, Moscow, ID 83844-3014 USA ([email protected]) 2 University of Liverpool, Liverpool L69 3GS UK ([email protected]) 3 Paul Scherrer Institut, c/o ETH Hönggerberg, CH-8093 Zurich, Switzerland ([email protected]) The dating of alluvial deposits is an important but difficult application of terrestrial cosmogenic nuclides. Initial work showed that exposure ages of surface clasts may be inaccurate because of nuclide inheritance and/or surface instability. Single nuclide depth profiles permitted inheritance to be estimated, and generated useful new ages for youthful deposits with stable surfaces. However, many alluvial deposits cannot be dated with single nuclide profiles because of surface disturbance or erosion. In the general case, three unknowns must be estimated: erosion rate of the contributing catchment; time since deposition; and erosion rate of the alluvial surface. Recently, measurements of long 10Be and 26Al profiles, studies of muonic production at depth, and measurement of nuclide concentrations in modern alluvium have invigorated the quest for a more robust means of dating ancient alluvium. These studies demonstrated that all three unknowns can be estimated from 10Be and 26Al profiles in favorable settings. In this study, the utility of 10Be and 26Al depth profiles for dating alluvial deposits was investigated in an archaeological setting. The Luangwa Valley of Zambia is an extension of the African Great Rift Valley but lacks the interbedded volcanic deposits that have enabled a detailed chronostratigraphic record elsewhere in the Great Rift system. Six samples of amalgamated gravel and sand were analyzed for 10Be and 26Al from a 4.5 m terrace section containing Oldowan artifacts at the base. A sample consisting of surface clasts was also analyzed, as was a sample of modern alluvium. The terrace is dissected and original depositional surfaces are absent. Preliminary results suggest that gravels associated with Oldowan stone tools were deposited at about 0.9-1.0 Ma. This is younger than the dated range of Oldowan artefacts in Africa and suggests that the age is a minimum. The apparent exposure age of the surface of the section is only about 85 ka. A model terrace erosion rate of 10 m/Ma suggests that at least 9 m of section has been lost. Denudation rates from the contributing catchment vary from about 10 to 70 m/Ma, and are similar to the modern alluvium estimate of 45 m/Ma.

Goldschmidt Conference Abstracts 2005 Cosmogenic Nuclides and Surface Processes

Integrating geomagnetic records and cosmogenic nuclide production

The CRONUS-Earth (Cosmic-Ray prOduced NUclide Systematics on Earth) Initiative

J.S. PIGATI1, N.A. LIFTON2, AND D. DESILETS3

JOERG M SCHAEFER1 , CRONUS-EARTH STEERING COMMITTEE2

1

Desert Laboratory, 1675 W. Anklam Rd., Tucson, Arizona, 85745, USA ([email protected]) 2 Department of Geosciences, University of Arizona, Tucson, Arizona, 85721, USA ([email protected]) 3 Department of Hydrology and Water Resources, University of Arizona, Tucson, Arizona, 85721, USA ([email protected]) Production of cosmogenic nuclides (CNs) in geologic material is a function of the cosmic ray flux at the Earth’s surface, which in turn is a function of the intensity and orientation of the Earth’s geomagnetic field. Temporal variations in the intensity of the geomagnetic field and the position of the geomagnetic dipole axis (i.e. polar wander) must be considered when calculating production rates that are integrated through time. We have developed a model, based on the theoretical framework of Desilets and Zreda (2003) and a variety of geomagnetic field intensity and pole position data, that accounts for these variations in an effort to systematically determine their impact on time-integrated production of shortlived (in situ 14C; t1/2=5.73 ka) and long-lived (in situ 10Be; t1/2=1.5 Ma) CNs (Pigati and Lifton, 2004). Our model differs significantly from previous models in that integrated production rates are normalized to the modern production rate at the geomagnetic, rather than geographic, latitude of a given site. Integrated rates that are normalized to the modern rate at a site’s geomagnetic latitude explicitly account for the fact that modern production reflects the current offset between the geomagnetic and geographic poles, and that time-integrated production is affected by polar wander differently at different locations. In contrast, normalizing integrated production rates to the modern rate at a site’s geographic latitude incorrectly suggests that a single correction can be applied to all sites along a given parallel. Our modelling results show that, depending on the exposure age and location, integrated in situ 14C production rates at sea level that account for both intensity variations and polar wander range from 27% higher to 24% lower than modern rates at the same location (modern rates are referenced to the 1945.0 Definitive Geomagnetic Reference Field). Integrated in situ 10Be rates range from 48% higher to 26% lower than modern. Differences between integrated and modern rates for both nuclides increase significantly at higher altitudes.

References Desilets, D., and Zreda, M. (2003), Earth and Planetary Science Letters, 206, 21-42. Pigati, J.S., and Lifton, N.A. (2004), Earth and Planetary Science Letters, 226, 193-205.

A167

1

Lamont-Doherty Earth Observatory, Palisades, NY-10964 Network of 9 PIs within the US

2

Terrestrial Cosmogenic Nuclides (TCN) have become indespendable tools in various disciplines of modern Earth Sciences. However, the understanding of the fundamental physical processes underlying TCN production remains incomplete and the intercomparibility between different investigators and methods is not satisfying. The CRONUS-Earth initiative is an interdisciplinary, multi-group project with the primary goals (i) to provide a firm linkage between cosmic-ray physics and the systematics of TCN production, (ii) to produce generally-accepted formulations and parameters for calculation of TCN production, and (iii) to establish a rigorous basis for intercomparison between measurement of different nuclides and by different investigators. To achieve this, CRONUS-Earth will coordinate six major components: (i) A methodological intercomparison, including sample preparation as well as analytical measurement; (ii) refinement of neutron monitor data interpretation to better understand interaction of the cosmic ray flux and the geomagnetic field; (iii) measurement of contemporary TCN production rates and scaling factors by exposing targets to cosmic rays at selected locations; (iv) measurements of production cross-sections using laboratory neutron beams; (v) calibration of TCN production rates by measuring TCN on independently dated surfaces; (vi) modeling to synthesize results. The models will include purely physical models of cosmic-ray-particle propagation through the atmosphere down to Earth, stastically-based parameter-estimation models, and user models. CRONUS-Earth is an international, collaborative effort with a close liason to the CRONUS-EU project. The intiative starts early 2005. CRONUS-Earth consist of a collaborative network of 17 PI’s, the current CRONUS-Earth steering committee includes M. Caffee, Purdue University; R. Finkel, LLNL, Livermore; T. Jull & N. Lifton, University of Arizona, Tucson; M. Kurz, WHOI, Woodshole; F. Phillips, New Mexico Tech, Socorro; J. Schaefer, Lamont-Doherty Earth Observatory, Palisades; J. Stone, University of Washington, Seattle; T. Dunai, Vrije Universiteit, Amsterdam, Netherlands (CRONUS-EU).

Goldschmidt Conference Abstracts 2005 Cosmogenic Nuclides and Surface Processes

A168

Inter-comparison in 10Be analysis starting from pre-purified quartz C. SCHNABEL1, L. REINHARDT2,3, P. BISHOP2, A. DAVIDSON1, L.K. FIFIELD3, S. FREEMAN1, C. MADEN1 AND S. XU1 1

Scottish Universities Environmental Research Centre (SUERC), East Kilbride, U.K. ([email protected]) 2 Centre of Geosciences, University of Glasgow, Glasgow, U.K. 3 Dept of Nuclear Physics, RS PhysSE, Australian National University, ACT 0200, Australia As far as the authors are aware we present the first intercomparison of 10Be analysis in quartz at environmental levels. Due to a lack of geological standard reference materials for 10 Be, quality control of exposure age and erosion rate determinations based on 10Be analysis from quartz is difficult. Until now inter-comparisons have neither included very low 10 Be concentrations nor complex sample preparation from quartz. 10 Be concentrations in six quartz samples from the Sierra Nevada, Spain, were analysed at ANU, Australia and at SUERC, Scotland. The samples were originally taken to determine erosion rates and these data will be published elsewhere. Pre-purified quartz prepared at ANU was divided into two aliquots and processed and analysed independently at ANU and SUERC. The table below summarizes the results for the first four samples. To compare two different chemical separation methods (addition of stable Be carrier before and after dissolution) two aliquots were prepared from sample B11 at SUERC (CF stands for carrier first and CL for carrier last addition). All results are normalised to NIST SRM 4325 using 3.00*10-11 as its Be isotope ratio. The uncertainties given are standard uncertainties that include uncertainties of the sample and the standard measurement as well as the uncertainty of the blank correction. Sample B11(CL) B11(CF) Ger3 17 21C

10

Be (at/g) ANU (1.64±0.06)*106 (3.12±0.14)*106 (1.22±0.10)*104 (2.04±0.14)*104

10

Be(at/g)SUERC (1.60±0.06)* 106 (1.71±0.08) *106 (2.97±0.12) *106 (1.31±0.16) *104 (0.70±0.08) *104

The analyses of both laboratories agree, within their uncertainties, for all samples except 21C.

Acknowledgment

F. v. Blanckenburg provided a low-10Be-carrier to SUERC.

iCRONUS meets CRONUS-Earth: Improved calculations for cosmogenic dating methods – From neutron intensity to previously ignored correction factors MAREK ZREDA1, DARIN DESILETS1, YANHUA LI1, ELIZABETH BRADLEY2 AND KENNETH M. ANDERSON2 1

Dept. of Hydrology and Water Resources, Univ. of Arizona, Tucson, AZ, 85721, USA ([email protected]) 2 Dept. of Computer Science, Univ. of Colorado, Boulder, CO, 80309, USA ([email protected]) We report progress on two five-year projects whose common goal is to improve cosmogenic dating methods: CRONUS-Earth will improve calibration; iCRONUS will develop a software system based on an artificial intelligence core (thus, the ‘i’ in the name). Calibrated production rates and correction factors modifying production rates are two critical aspects of calculating cosmogenic ages. Calibration depends on the accurate computation of neutron fluxes at the air-ground interface. The currently-used diffusion equation underestimates neutron fluxes at the surface. Two more accurate alternatives, the physically comprehensive Monte Carlo NParticle transport code and a simpler analytical transport model, are implemented in iCRONUS. Correction factors are of two types: global (affect all samples) and local (affect only the samples from a specific landform). Global correction factors include those that modify the secondary cosmic ray intensities; the most important are air pressure and geomagnetic cutoff rigidity of the sample site. The size of the correction depends on the location, temporal variations of the geomagnetic intensity, position of the magnetic poles, eustatic changes of sea level, temporal and spatial changes of sea-level pressure, and temporal and spatial changes of temperature and lapse rate. Every landform also requires its own, unique set of local corrections, applied on top of the global corrections. Examples include erosion of landform’s surface and sampled surface, (neo)tectonic displacement, topographic shielding, cover, and variable chemistry. Our improved calibration and all correction factors form a framework implemented in the iCRONUS software. We will demonstrate a desktop version of iCRONUS at the meeting.

Acknowledgement Work funded by the National Science Fundation through the iCRONUS project (grants ATM-0325929 and ATM0325812) and the CRONUS- Earth project (grant EAR0345440).

Goldschmidt Conference Abstracts 2005 Diffusion-Reaction Systems

A169

Fracturing-assisted reactive transport

Mineral reaction interfaces and associated porosity generation

BJØRN JAMTVEIT AND ANDERS MALTHE-SØRENSSEN

DAVID R. COLE1, T.C. LABOTKA2, P.B. LARSON3, K.L. MORE4, E.A. KENIK4, M. FAYEK1, F.J. STADERMANN5 AND L.R. RICIPUTI1

Physics of Geological Processes, University of Oslo, P.O.Box 1048 Blindern, N-0316 Oslo, Norway ([email protected], [email protected]) Fluid migration in the Earth’s crust is strongly linked to deformation. Fracturing is probably the most important porosity and permeability producing mechanism in the lithosphere. Large-scale fracturing is often linked to tectonic stresses, but large-scale fluid flow may also be strongly affected by small-scale forces and small scale fractures. Local fracturing is in some cases a response to locally generated stresses rather than external forces. Such stresses may be related to high fluid pressure gradients, stress perturbations caused by geochemical reactions, or -near the Earth’s surface – even to microbial activity. We present field observations and computer simulations (Fig.1) which illustrate how fluids may propagate into initially more or less impermeable rocks driven by reaction-enhanced porosity generation. The velocity and geometry of the reaction fronts are discussed, as well as the effects of anisotropic external stresses and other boundary conditions on the fluid migration patterns.

Figure 1. Simulation of a fracture-propagated reaction front in the limit of diffusion control. Gray-scale indicates reaction progress. The front is sharp with a small-scale structure characterized by a block size, L = (Kc/Eν), where E is Young's modulus, Kc is the critical stress intensity factor, and ν is the volume reduction ratio for the reaction. The front moves with a constant velocity v = D/L (Jamtveit et al., 2000), where D is the effective diffusion constant. On a scale much larger than L the front has a self-affine scaling structure which is characteristic for the front propagation process.

Reference Jamtveit, B, Austrheim, H., and Malthe-Sørensen, A., (2000), Nature, 408, 75-79

1

Chemical Sciences Division, Oak Ridge National Lab., Oak Ridge, Tennessee 37831, USA ([email protected]) 2 Dept. of Earth & Planetary Sci., Univ. of Tennessee, Knoxville, Tennessee 37996, USA 3 Dept. of Geology, Washington State University, Pullman, Washington 99164, USA 4 Metals & Ceramics Division, Oak Ridge National Lab., Oak Ridge, Tennessee 37831, USA 5 Dept. of Physics, Washington University, St. Louis, Missouri 63130, USA Chemical processes commonly encountered in nature, including hydration-dehydration, cation exchange and oxidation-reduction reactions involving complex fluids containing electrolytes and mixed-volatiles can lead to either passivated mineral surfaces or the formation of reaction rims. Further exchange requires advection and/or diffusion of matter across the zone to the interface between the transformed and unreacted parent phase. Structures within the reaction zone and at the reaction interface, as well as reaction mechanisms and reaction rates, are still poorly constrained for most important rock-forming mineral. This study focused on two aspects of the replacement process, the geometry of the reaction front and the generation of nano- and/or microscopic porosity at or near the interface. Using transmission electron microscopy (TEM) and secondary ion mass spectrometry (SIMS), we interrogated the reaction interface domains in two types of feldspar systems: (a) replacement of plagioclase (∼An30) by albite ± muscovite in the Rico, CO paleohydrothermal system, and (b) Amelia albite replaced by Kfeldspar in experiments (up to 600oC and 200 MPa) employing 1-2 molal KCl enriched in 18O. In the case of the natural system, the typical reaction interface appears as a curvilinear, somewhat diffuse zone where crystallographic control is not particularly pronounced. Micropores occur infrequently along the interface, but are commonly observed within fine albitefilled fractures and dislocations that either crosscut or are truncated by the reaction interface. In contrast, the nature of the reaction interface in the experimental sample depends strongly on crystallographic control with an overall jagged appearance, exhibits limited microporosity, and has an unusual corrugated or stitch-like nano-texture right at interface. NanoSIMS images and line scans indicate the interface is sharp in the distribution of 18O and 16O.

A170

Goldschmidt Conference Abstracts 2005 Diffusion-Reaction Systems

Conversion of rock to saprolite: A study of weathering rinds ALEXIS K. NAVARRE1, CARL I. STEEFEL2 3 AND SUSAN L. BRANTLEY 1

Department of Geosciences, Pennsylvania State University, University Park, PA 16801, USA ([email protected]) 2 Earth Sciences Division, Lawrence Berkeley National Laboratory, 1 Cyclotron Road, Berkeley, CA 94720, USA ([email protected]) 3 Department of Geosciences, Pennsylvania State University, University Park, PA 16801, USA ([email protected]) Weathering rinds can be defined as small-scale saprolites that develop on clasts. Developed in a controlled environment and protected from physical weathering by surrounding material, these clasts provide a unique setting to examine chemical weathering processes without additional complications due to the removal of weathered material by physical erosion. Basalt clasts with 3 cm weathering rinds collected from 125 ka alluvial terraces along the Pacific coast of Costa Rica have been analyzed with SEM backscatter and EDS images, EMP, and XRD. These data, combined with reactive transport simulations, have been used characterize the reaction front and develop a conceptual model of the chemical processes that occur during saprolite formation. Due to low parent porosity (<1%), initial transport of reactants and products to and from the reaction front occurs primarily by diffusion. Small amounts of iron oxide have been observed precipitated around augite crystals and in fractures coreward of the reaction front, but the dissolution of labradorite and precipitation of kaolinite is the first chemical change to occur at the reaction front that results in increased porosity. As weathering proceeds and porosity increases, fluid is able to flow through the rind (~50% porosity). This change in the dominant transport mechanism results in an influx of water with low silica concentrations and the replacement of kaolinite by gibbsite, an effect which is not possible with diffusive transport alone. Chemical data suggests that the diffusion dominant zone is <0.5 mm thick and separates unaltered rock from high porosity rind. The rate of rind advance in this system is slow (2.4x10-4 mm/yr) compared to basalt weathering at the landscape scale (0.1-1.2 mm/yr) for the same climate. Most landscape scale rates are from large basaltic provinces. These basalts are extrusive with high porosity. Porosity has been indicated to have a primary control on weathering rind thickness and therefore weathering rate. Other possible causes for faster rates at larger scales could be the coupled effect of physical and chemical weathering.

Microscopic reactive diffusion of U(VI) in subsurface sediments: Characterization and modeling C. LIU, P.D. MAJORS, J.M. ZACHARA AND J.P. MCKINLEY Pacific Northwest National Laboratory, Richland, WA 99352, USA Uranium in the US Department of Energy (DOE) Hanford 200 Area sediments was found to be distributed as uranyl silicate precipitates (Na-boltwoodite) predominantly within interiors of sediment granitic clasts and preferentially in a minor plagioclase feldspar component. The precipitates were minute, generally 1-3 µm across in either radiating or parallel arrays in microfractures of a few microns width and variable connectivity to particle surfaces (Liu et al, 2004). The uranyl precipitates dissolved and released from intraparticle regions to undersaturated aqueous solutions (Na-NO3-HCO3) at a rate of over two orders of magnitude slower than the intrinsic dissolution rate of Na-boltwoodite. The results indicated that diffusion limited the dissolution release of the uranyl precipitates. X-ray microprobing the precipitates before and after dissolution confirmed this conclusion. A pulse-field gradient nuclear magnetic resornance (PFGNMR) approach was developed to characterize microscopic diffusion. The approach uses water (H2O) itself as a tracer and avoids mass displacement procedures. Measurements with PFG-NMR identified at least two intraparticle diffusion regions with H2O diffusivities of over 2 orders of magnitude difference. The fast diffusion region has an apparent tortuosity of about 1.5 and appeares associated with the waters on particle surfaces and in large intraparticle fractures, while the slow one has an apparent tortuosity of about 350 and is associated with secondary fractures. The two diffusion regions were generally confirmed by the NMR imaging of H2O distribution as a function of water content. A model was assembled to simulate reactive diffusion of radioactive waste into the intrapartcle microfractures. Simulations indicated that waste U(VI) preferentially precipitated and concentrated within the feldspar microfractures where the saturation of Na-boltwoodite was maintained by the diffusive supply of U(VI) and Na from waste, and dissolution of silicates for Si.

Reference Liu C., Zachara J.M., Qafoku O., McKinley J.P., Heald S.M., and Wang Z., Geochim Cosmochim Acta 68, 4519-4537.

Goldschmidt Conference Abstracts 2005 Diffusion-Reaction Systems

Multiple approaches to studying diffusion processes in geological media Q. HU1, R.P. EWING2, C.I. STEEFEL3, L. TOMUTSA3 1 AND G.B. HUDSON 1

Chemical Biology and Nuclear Science Division, Lawrence Livermore National Laboratory, Livermore, CA 94550, USA ([email protected]; [email protected]) 2 Department of Agronomy, Iowa State University, Ames, IA 50011, USA ([email protected]) 3 Earth Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, CA 94720, USA ([email protected]; [email protected]) The current conceptual model of contaminant transport in unsaturated fractured rock has water moving through fractures, with migration of the entrained contaminants being retarded by diffusion into and within the rock matrix. The diffusion coefficient is implicitly assumed to have a constant, scale-invariant value. However, rocks whose pores are poorly interconnected are known to have anomalous transport properties that strongly impact long-term net diffusion. Complementary and intergrated experimental/modeling approaches are pursued to study diffusion processes in geological media: (1) Imbibition (analogous to diffusion) tests wherein the water uptake is monitored over time, with the slope of log imbibed water mass versus log time indicative of the pore connectivity of the rock matrix. Change in the slope (if any) can be related to the correlation length of percolation theory. (2) Gas diffusion experiments, employing an on-line quadrupole gas mass spectrometer to measure real-time trace gas concentrations, to determine diffusion coefficients at several different sample thicknesses to examine scaling effect. (3) Laser ablation interfaced with inductively coupled plasma-mass spectrometry (ICP-MS) to measure fine-scale tracer concentration to investigate edge-accessible porosity distribution, which is related to pore connectivity. (4) 3-dimensional images obtained from the synchrotron microtomography system at the Advanced Light Source are analyzed to obtain tracer iso-concentration surfaces, which are indicative of pore connectivity and cross-over length. (5) All experiments are modeled via random walks on 3-D lattices with different pore coordination and connectivity. Results indicate that porespace in many rocks is close to the percolation threshold, and use of standard diffusion models for rock matrices can yield misunderstanding of diffusion process and incorrectly calculated diffusivity values.

A171

The importance of diffusion at the microbe-mineral interface: Electrical double layer effects and the impact on precipitation/dissolution A.R. FELMY, C. LIU, AND T.P. STRAATSMA Pacific Northwest National Laboratory, P.O. Box 999, Richland, WA 99352, USA ([email protected], [email protected], [email protected]) Understanding the interactions between mineral surfaces and micro-organisms is one of the most challenging areas of research in the geosciences. Many of these interfacial processes take place in very confined spaces and over very short distances where diffusive transport is the predominant process. In this presentation we will give an overview of the surface of the microbial membrane, the chemical factors behind the development of a bacterial electrical double layer, the influence of this double layer on the transport of ions at the surface, and the influence of counter-current diffusion on bacterial mediated precipitation/dissolution reactions. In the latter case, the chemical microenvironment at the microbial surface is modified by the flux of reactants and microbial byproducts from the organism to the surrounding solution. This flux of reactants is partially controlled by the generally negatively charged microbial Solvating Waters surface which Goethite induces a cationic flux toward the organism and an Bacterial Outer expulsion of anions Membrane from the surface. These diffusive fluxes can create regions of Solvating Waters oversaturation with respect to the Molecular dynamics mineral precipitation simulation of the microbial reactions near the membrane – mineral interface microbial surface. in a fully solvated system. Results of both continuum level and molecular dynamics simulations of ion diffusion at microbial surfaces will be presented and the limitations of these approaches in simulating ionic diffusion and at the microbe/mineral interface will be discussed.

Goldschmidt Conference Abstracts 2005 Diffusion-Reaction Systems

A172

Lattice Boltzmann pore-scale model for coupled multi-component flow, diffusion, and reaction QINJUN KANG AND PETER C. LICHTNER Hydrology, Geochemistry and Geology Group, Los Alamos National Laboratory, Los Alamos, NM 87545, USA

Numerical method In this work, we present a multi-component lattice Boltzmann model for simulating reactive transport in porous media at the pore scale taking into account flow and diffusion in complex geometries. In the model, a set of distribution functions is introduced to simulate fluid flow and solute transport. The evolution equations recover the correct continuity and Navier-Stokes equations, and convectiondiffusion-reaction equations [Kang et al., 2002]. This model takes into account convection, diffusion, homogeneous reactions among multiple aqueous species, heterogeneous reactions between the aqueous solution and minerals, as well as the resulting geometrical changes in pore space. Homogeneous reactions are described through local equilibrium mass action relations. Mineral reactions are treated kinetically through boundary conditions at the surface. We have applied this model to a variety of multi-component systems in synthetically constructed media, and analyzed the effects of convection, diffusion, reaction rate constants, equilibrium constants, and chemical compositions on mineral alteration of the porous medium. Figure 1: Geometry, stream lines (left) and contours of solute B (right) B(aq) 0.98 0.90 0.83 0.75 0.68 0.60 0.53 0.45 0.38 0.30 0.23 0.15 0.08 0.00

Discussion of results Figure 1 shows the simulation results of a system with solute species A, B, C, and minerals AB (denoted as gray symbols) and AC (denoted as black symbols) undergoing reactions A+B=AB and A+C=AC. Initially, only AB exists in the domain in equilibrium with A and B. A solution including only A and C (in equilibrium with AC) is introduced at the entrance. Mineral AB dissoves, causing precipitation of AC.

Reference Kang Q. et al. (2002) Phys. Rev. E, 66, 036318.

Bubble growth in soft sediments: Diffusion meets solid mechanics BERNARD P. BOUDREAU Department of Oceanography, Dalhousie University, Halifax NS B3H 4J1, Canada ([email protected]) The quantitative description of diffusion and reaction with multiple physical phases presents significant challenges to both experimentalists and modellers; however, such problems arise naturally and commonly in geochemical systems. The growth of bubbles in soft organic-rich sediments offers a particularly striking example. Methane is generated in porewaters from the anoxic decomposition of organic matter, after the exhaustion of all inorganic oxidants. The resulting methane concentrations often exceed saturation with respect to the gas phase, and bubbles regularly form. An understanding of the formation of such bubbles is difficult to obtain: sediments are visually opaque, the bubbles are relatively small generally (1 cm3 or less), and the slow growth rates complicate direct observation. In addition, the nature of the mechanical interaction between the growing bubble and the medium has been ignored, but it is crucial to the problem . Recent advances in two experimental methods have shed considerable light on the mode of formation of these bubbles. The first is the development and application of high-resolution CT-scanning; such images have shown that bubbles in soft sediments are highly eccentric and distorted disks. The second methodology has been the development of a microinjector system that can create bubbles in sediments and simultaneously measure their internal pressure. The information from these techniques indicates that bubbles grow by fracturing the adjacent sediment, not by fluidization as might be assumed. Given this type of mechanical interaction between bubble and sediment, it is possible to create a reaction-diffusionfracture model. Methane is generated by a distributed source in the sediment and moves by molecular diffusion to the bubble nucleation site (assumed abundant). The sediment responds as a linear-elastic solid to the increased gas concentration in the bubble until a critical stress is achieved, at which time the sediment fractures instantaneously and the bubble elongates a finite distance. This process is repeated to create a bubble that becomes more eccentric with size/time. The shape predictions of this model are completely in accord with the CT and injection data and predict growth times on the order of days for an organic-rich site in the USA.

Acknowledgements This work would not have been possible without the collaboration of Chris Algar, Mark Barry, Yoko Furukawa, Allen Reed, Ian Croudace, and especially Bruce Johnson.

Goldschmidt Conference Abstracts 2005 Diffusion-Reaction Systems

A173

Quantitative interpretation of pH distributions in aquatic sediments: A reaction-transport modeling approach

Oxygen diffusion and consumption in eroding black shales: A control on long-term atmospheric oxygen

P. JOURABCHI, P. REGNIER AND PH. VAN CAPPELLEN

Dept. Geology & Geophysics, Yale Univ., New Haven, CT, USA ([email protected], [email protected])

Department of Earth Sciences-Geochemistry, Utrecht University, Utrecht, The Netherlands. ([email protected]) Despite its status of master variable, there have been relatively few attempts to quantitatively predict the distributions of pH in biogeochemical reactive transport systems. Here, we propose a theoretical approach for calculating the vertical porewater profiles of pH and the rates of proton production and consumption in aquatic sediments. In this approach, the stoichiometric coefficients of species that participate in acid-base equilibrium reactions are treated as unknown variables in the biogeochemical reaction network. The mixed kinetic-equilibrium reaction system results in a set of coupled differential and algebraic equations and is solved using a new numerical solver that can handle both steady state and transient simulations. The diagnostic capabilities of the model are illustrated for depositional conditions representative of those encountered on the continental shelf. The early diagenetic reaction network includes the major microbial degradation pathways of organic matter and associated secondary redox reactions, mineral precipitation and dissolution processes, and homogeneous acid-base reactions. The resulting pH profile in this baseline simulation exhibits a sharp decrease below the sediment-water interface, followed by an increase with depth and again a decrease. The features of the pH profile are explained in terms of the production and consumption of protons by the various biogeochemical processes. Secondary oxygenation reactions are the principal proton producers within the oxic zone, while reduction of iron and manganese oxyhydroxides are primarily responsible for the reversal in the pH gradient in the suboxic zone. Proton production in the zone of sulfate reduction outweighs alkalinity production, maintaining the undersaturation of the pore waters with respect to calcite. Integrated over the entire depth of early diagenesis, dissolution of CaCO3 is the main sink for protons. Variations in the reaction rate order and rate constant for CaCO3 dissolution do not fundamentally alter the shape of the pH profile. An entirely different shape is obtained, however, when the pore waters are assumed to remain in thermodynamic equilibrium with calcite at all depths. The intensity of pore water (bio)irrigation has a major impact on pore water pH and may increase the difference in pH between the irrigated zone and the underlying, nonirrigated sediment.

E.W. BOLTON AND R.A. BERNER

Oxygen diffusion into eroding black shales leads to buried fronts of oxidizing ancient organic matter (OM) and pyrite. The related diffusion-reaction problem with surface erosion is examined by means of a one-dimensional numerical model with comparisons made to a recently completed field study near Clay City, Kentucky (Wildman et al., 2004); both of which exhibit regions (fronts) where OM and pyrite increase rapidly from near zero to finite values characteristic of the deep unoxidized sediments. In the model, downward oxygen diffusion is followed by Henry law partitioning into water films adsorbed on the reduced particles, where aqueous phase oxidation of OM and pyrite can take place under kinetic control. Sensitivity studies were conducted to examine details of the front morphology under the influence of different particle sizes, mass fractions, tortuosity, erosion rate, atmospheric oxygen content, water saturation, etc. The results have relevance to global chemical cycling. We conclude that the long-term oxygen evolution in the atmosphere is controlled primarily by the erosion rate and the quantity of OM and pyrite being exhumed from depth via erosion, rather than depending upon the atmospheric oxygen content. Only for extremely rapid erosion rates would finite amounts of ancient OM be eroded (and rapidly reburied). The numerical problem is challenging due to the stiff equations that arise from the rapid pyrite kinetics, along with the fact that OM, pyrite and oxygen can completely disappear. Evolution toward a steady-state solution is possible only for non-zero erosion rates. Box models utterly fail to capture the dynamics relevant for oxygen flux calculations relevant to the evolution of atmospheric composition. We may also comment on other isotopic and trace element diffusion studies underway.

References Wildman, R. A., Berner, R. A., Petsch, S. T., Bolton, E. W., Eckert, J.O., Mok, U., and, Evans, J.B., (2004) American Journal of Science, Vol. 304, p. 234-249.

Goldschmidt Conference Abstracts 2005 Diffusion-Reaction Systems

A174

Mineralization pathways in the sediments of Lake Baikal – A microsensor and modeling study

Isotopic effects in reactive fluid-rock systems with fracture-dominated flow and matrix diffusion

M. MAERKI, B. MÜLLER AND B. WEHRLI

DONALD J. DEPAOLO12 AND MEGAN M. SMITH12

Swiss Federal Institute for Environmental Science and Technolgy (EAWAG) and Swiss Federal Institute of Technology (ETH), Limnological Research Center, Kastanienbaum, Switzerland, ([email protected], [email protected], [email protected])

1

Burial of organic matter in Lake Baikal (Siberia, Russia), is strongly affected by the mineralization processes in the surface sediments. These processes were analyzed on-site using an array of ion-selective electrodes (O2, pH, CO32-, Ca2+, NH4+ and NO3-) and a micromanipulator controlled by a laptop computer. The setup was used to measure concentration profiles at sub milimeter resolution immediately after retrieval of sediment cores from the South Basin (160m, 600m 700 m and 1400 m depth). Modeled oxidation rates of organic carbon constrained by O2 and NO3- profiles were in the range of 2.2 to 4.9 mmol C m-2 d-1. We found that 60 - 75 % of the benthic carbon turnover was metabolized via oxic respiration, only 11 – 28 % through anoxic mineralization pathways, and the remainder by denitrification. Evidence for re-oxidation processes at the redox boundary at sediment depths of about 2 cm was obtained from the shape of O2 profiles and from an alkalinity balance established with flux rates of HCO3- calculated from pH and CO32- profiles. It identified the carbonate-free sediment as a sink of alkalinity caused by buffering of H+ generated by re-oxidation processes of reduced compounds such as Fe2+, Mn2+ and NH4+ [1]. Dark CO2 assimilation by chemoautotrophic bacteria was on the order of 0.03 - 0.1 mmol C m-2 d-1 and could be neglected in the organic carbon balance. Due to the long O2 exposure of 25-2500 years only 3-14 % of the settling organic carbon estimated from sediment traps and water column data is finally buried in the sediments of Lake Baikal [2].

Isotopes are used broadly to characterize reactive fluidrock systems. Most modeling assumes 1D-porous flow and mineral-fluid reaction rates fast enough to maintain “local equilibrium.” In fractured rocks, isotopes in both fluid and solid exhibit different behavior than in simple porous flow. The departure from the porous flow model depends on the diffusivities of dissolved species in matrix pore fluid, their solubility, and the mineral-fluid reaction rates. Isotopes can be used to evaluate the importance of diffusion-limited reaction zones in fluid-rock systems. A mathematical model is developed that describes the effects of matrix diffusion on the isotopic evolution of fluids and rocks in systems with fracture-dominated flow. The approach generalizes previous models for radiocarbon ages of groundwaters, and those that assume infinitely fast fluid-rock exchange rates. The equations describe parallel equidistant fractures separated by slabs of porous “matrix” rock containing stagnant pore fluid. The effect of matrix diffusion on the isotopes of a particular element depends on the ratio of the diffusive reaction length (Ld) to the fracture spacing (b). Ld is a function of the mineral-fluid reaction rates within the matrix, the solubility and aqueous diffusivity of the element, and the matrix porosity. For Ld/b < 0.3, the fracture fluid exchanges isotopes only with a narrow zone of matrix within Ld of the fracture, and fluid-rock exchange produces an isotopic effect on the fluid that is smaller by 2Ld/b than for porous flow. For Ld/b > 1, the parallel fracture model is no different from a single porosity (1D porous flow) model. One element (e.g. O) can be in one regime while a second (e.g. Sr) is in the other. The model is applied to Sr and O isotopic data from MOR and continental hydrothermal systems, and groundwater in fractured basalt. Data from multiple isotope systems for one aquifer can provide estimates of fracture spacing and fluidrock exchange rates. For MOR systems, fracture spacing is estimated at 1-4m. For groundwater systems, fracture flow results in underestimates of fluid-rock exchange rates for some elements.

References [1] Granina L., Müller, B., Wehrli, B. (2004) Chemical Geology 205, 55-72 [2] Müller B., Maerki M., Schmid M., Vologina E. M., Wehrli B., Wüest A., Sturm M. (2005) Global Planetary Change in press.

Department of Earth and Planetary Science, University of California, Berkeley, CA 94720-4767, USA ([email protected]) 2 Earth Science Division. E.O. Lawrence Berkeley National Laboratory, Berkeley, CA 94720, USA

Goldschmidt Conference Abstracts 2005 Diffusion-Reaction Systems

Non-equilibrium thermodynamics: Diffusion controlled partial melting MASSIMILIANO TIRONE AND JASON PHIPPS MORGAN Department of Earth & Atmospheric Sciences, Cornell University, NY, 14853, USA ([email protected], [email protected]) Melting is commonly assumed to take place in a rock assemblage that is in chemical equilibrium. The question we would like to answer is: does diffusion of major elements in mantle solid phases affect melt abundance and composition and if yes to what extent and at what conditions. The problem is addressed using our general thermodynamic/multiphase flow model described in some details in a separate communication (symposium S45) and applied in the context of ridges melting. To include the effect of diffusion in the model, transport equations for grain size and diffusion profiles are also solved and the problem is treated on two scales in space and time. The interesting feature is that the thermodynamic database for solids and melt (Ghiorso et al., G3, 2002) is still applied to the model through the Gibbs free energy minimization at every space and time step. When diffusion is taken into account, the thermodynamics model puts a constraint on the chemical transfer rate for every solid phase component but only at the solid-melt interface. It is well known that during an equilibrium polybaric melting process, chemical elements are incorporated in the melt phase in different proportion as determined by their compatible/incompatible nature. Non equilibrium melting can alter the normal distribution of the chemical elements during melting process. We are expecting that if disequilibrium is maintained during mantle uprising, incompatible elements are retained in the solid phases for longer time compared to the equilibrium case. At the same time melt productivity should increase due to the preservation of a more fertile mantle at lower depths. This could have significant effects on Mg/Fe ratio and Na content in MORBs in particular for fast spreading ridges.

A175

In-situ diffusion at Mont Terri URL J.M. SOLER1, TH. GIMMI2, A. CARTALADE3, P. WERSIN4 5 AND L. R. VAN LOON 1

CSIC-IJA, Barcelona E-08028 ([email protected]) PSI, Villigen CH-5232 ([email protected]) 3 CEA, Gif-sur-Yvette F-91191 ([email protected]) 4 NAGRA, Wettingen CH-5430 ([email protected]) 5 PSI, Villigen, CH-5232 ([email protected]) 2

Clay-rich formations, such as the Opalinus Clay (Switzerland), are currently being considered as potential host rocks for the deep geological disposal of radioactive waste. Diffusion is the main transport mechanism for radionuclides in these impermeable rocks. Besides, sorption provides additional retardation for cationic species. The objective of the DI-A in-situ diffusion experiment at the Mont Terri Underground Rock Laboratory (URL) was to confirm the expected diffusion-controlled transport and to compare the results of the experiment with those from smallscale (cm) through-diffusion experiments. The experimental setup at Mont Terri consisted of a borehole drilled in the rock, with a 1-meter-long injection interval at its bottom. Synthetic porewater containing an initial pulse of tracers (HTO, I-, 22 Na+, Cs+) was circulated through the borehole, and the evolution of tracer concentration was monitored. After about 10 months, a volume of rock around the injection borehole was excavated and tracer distribution profiles in the rock were measured. Reactive transport simulations allowed the fitting of (a) the temporal evolution of tracer concentration in the injection system and (b) the tracer profiles in the rock, which provided unique sets of effective diffusion coefficients (De) and accessible porosities (sorption parameters for sorbing tracers). The results for HTO, I- and 22Na+ were in excellent agreement with those from through-diffusion experiments, confirming the important effects of anionic exclusion (I-) and sorption (22Na+). There were no previous experimental values of De for Cs+, although batch sorption data were available. The results of DI-A indicated less sorption (by a factor of about 2) in the intact rock than in batch. Also, De for Cs+ was about 5 times larger than for water (HTO). These results are now being confirmed by through-diffusion experiments. A second experiment (DI-A2) is currently under way. This time, the tracers are HTO, I-, Br-, 85Sr2+, 60Co2+, Cs+ and Eu3+. Additionally, scoping calculations are being performed for a new experiment (DR). The main objectives are to study in-situ diffusion for longer time scales (3-5 years) and to optimize the experimental concept for strongly sorbing cations and for measuring diffusion anisotropy in situ.

A176

Goldschmidt Conference Abstracts 2005 Diffusion-Reaction Systems

Spreading versus mixing effects on reactive transport in groundwater

Plenary Address by the Clarke Medalist

M.E. MALMSTRÖM1, G. DESTOUNI2 AND S. BERGLUND 3

Diffusion in mantle and core materials

1

Industrial Ecology, Royal Inst. of Technology (KTH), 100 44 Stockholm, Sweden; [email protected] 2 Physical Geography, Stockholm University (SU), 106 91 Stockholm, Sweden; [email protected] 3 Swedish Nuclear Fuel and Waste Management Co (SKB), 102 40 Stockholm, Sweden; [email protected] Many environmentally important systems involve complex water-rock interactions within natural subsurface systems that exhibit large physical and (bio)geochemical heterogeneity. Contaminant transport and retention are governed by the complex interplay between (bio)geochemical reactions and physical transport along spatially variable water flow paths, where solute dispersion affects the timing, duration, and concentration levels of water contamination arrival downstream of pollution sources. Longitudinal solute spreading among different advective flow paths occurs due to subsurface heterogeneity on various scales, with molecular diffusion and pore-scale dispersion causing also mixing of aqueous solutions within flow paths. Combined handling of these different dispersive processes in reactive transport modeling, using a single dispersion or diffusion term, neglects the fact that contaminant spreading and solution mixing may have very different dispersion effects. In this study, we consider physical spreading, coupled with precipitation/dissolution and sorption processes, of heavy metals in groundwater downstream of a mill tailings deposit, as one environmentally important example of a multicomponent reactive transport system. We use the PHREEQC [1] and LaSAR-PHREEQC [2] modeling approaches to quantify separate and coupled effects of mixing and spreading on resulting downstream dispersion of reactive contaminants in heterogeneous groundwater systems. Whereas reactions coupled with mixing often decrease resulting contaminant dispersion due to formation of reaction fronts, reactions coupled with spreading tend to increase the contaminant dispersion through reaction front fingering.

References [1] Parkhurst D.L. and Appelo, C.A.J. (1999), USGS WaterResources Investigations Report 99-4259, USGS, Denver, Colorado, USA. [2] Malmström M.E., Destouni G., and Martinet P. (2004), Environ. Sci. Tech. 38, 2673-2679.

JAMES A. VAN ORMAN Case Western Reserve University, Dept. of Geological Sciences ([email protected]) Solid-state diffusion is the rate-limiting step in many important geochemical processes in Earth and planetary interiors, especially those involving chemical and isotopic mass transfer among minerals, or minerals and fluids. As the rate-limiting step in creep, diffusion also plays a crucial role in the internal convection and thermal evolution of planets. Experimental and theoretical work on the rates and mechanisms of diffusion in minerals has advanced rapidly, particularly in the last decade. This talk will highlight some of these advances, focusing on work in which the author has been involved. Experiments on diffusion of lanthanides and actinides in high-Ca pyroxene demonstrate a systematic dependence of the diffusion coefficient on ionic radius and charge for elements that diffuse by a common mechanism. This dependence can be understood in terms of a simple model that considers the elastic energy required to deform the lattice, allowing an ion to jump to an adjacent vacant site. The elastic model provides a useful means for estimating diffusion coefficients of ions that have not been studied experimentally. Understanding the influence of pressure on diffusion rates is critical for extrapolating diffusion-controlled properties such as viscosity to the conditions of Earth’s deep mantle and core. Recent high-pressure experiments on periclase (MgO), close-packed Fe-Ni alloys, and Zn, an analogue for hcp-Fe in Earth’s inner core, will be discussed in relation to theoretical approaches for extrapolating diffusion data to high pressures.

Goldschmidt Conference Abstracts 2005 Diffusion-Reaction Systems

Diffusion and reaction in multicomponent partially molten silicates: Dissolution-reprecipitation YAN LIANG Department of Geological Sciences, Brown University, Providence, RI 02912, USA ([email protected]) The essential features of the kinetics of crystal-melt reaction in multicomponent partially molten silicates will be illustrated using numerical examples for mono-mineralic and bi-mineralic systems. Diffusion equations for the independent components in the melt and crystals, subject to the phase diagram constraint and moving boundary conditions at the crystal-melt interfaces, were solved numerical in 1D using a finite difference method. The input parameters are the liquidi and solidi of the crystals (for major elements) or crystal-melt partition coefficients (for trace elements), initial crystal-melt proportions, and diffusion coefficients of the independent components in the melt and crystals. One of the unique features of diffusion and reaction in high-temperature multicomponent two-phase aggregates is dissolution and reprecipitation. If the interstitial melt is initially under-saturated with respect to the crystals, the crystals will dissolve, but only at earlier times. Because the rate of diffusion in the melt is much faster than that in the solid, the bulk melt composition quickly evolves to the liquidi of the dissolving crystals. Since the interiors of the crystals are still over-saturated with respect to the melt, dissolution reverses to precipitation that finally stops when the concentration gradients in the solids are eliminated. An important consequence of dissolution and reprecipitation is the acceleration of crystal-melt re-equilibration in partially molten silicates. Potential implications include, but not limited to, isothermal partial melting in the laboratory, disequilibrium partial melting in the nature, crystal-melt trace element partitioning in the lab and nature, and crystal-melt reaction in magma bodies. Animations showing the effects of dissolutionreprecipitation on major and trace element distributions during partial melting and crystal-melt reaction will be presented.

A177

How solid solution minerals react with melt during diffusion-reaction CRAIG C. LUNDSTROM Dept of Geology, Univ. of Illinois at Urbana Champaign ([email protected]) Melt-rock reactions occur ubiquitously as magmas ascend and differentiate. Activity gradients can readily drive diffusion of chemical species between juxtaposed partial melts of different lithologies. The mechanism by which minerals reequilibrate with melt as it dynamically changes composition by diffusion is fundamental to diffusion-reaction (DR) processes. Piston cylinder DR experiments suggest that solidsolution minerals do not re-equilibrate by solid-state diffusion nor by dissolution-reprecipitation; instead, mineral rims remain in exchange equilibrium with melt while mineral cores adjust composition and buffer chemical changes of the melt. Two examples illustrate this process: 1) In a 2 hr DR experiment between partially molten basanite and peridotite (Lundstrom, G3, 2003), Mg diffuses from peridotite to basanite, increasing the bulk MgO content of the basanite. However, the MgO content of the basanite melt only increases slightly because the cores of olivines in the basanite increase in Mg# relative to their starting composition. Notably, olivine rims maintain Mg-Fe exchange with the melt. 2) In a 13 day DR experiment between partially molten gabbro and basaltic andesite, a plagioclase-rich boundary layer develops at the lithologic interface as Ca diffuses into the basaltic andesite (as predicted by models using IRIDIUM; Boudreau, Comput. Geosci, 2003). Like the olivine case, plagioclase rims in this zone maintain Na-Ca exchange equilibrium with the melt but mineral cores become highly anorthitic (up to An90). Isotopic tracers show that these cores are in full chemical communication with the melt as they grow. To generalize, solid-solution minerals adjust to dynamically changing melt composition by maintaining rimmelt equilibrium as cores chemically exchange with melt. In the case of plagioclase, this appears to occur by formation of small melt channels that traverse the crystal rim based on SEM images (TEM work is scheduled). By cores taking up the incoming flux, these minerals buffer changes in melt composition. Important implications of this finding include: 1) the presumed age stratigraphy of rims being “older” than cores is not valid; 2) mineral-melt equilibration is much more rapid than solid-state diffusion. Diffusion through channels in the crystal could cause fractionation of isotopes, explaining “non-traditional” isotope variations observed in mantle minerals.

Goldschmidt Conference Abstracts 2005 Diffusion-Reaction Systems

A178

TEM/EELS measurement of Fe /ΣFe in biotite near a fracture

Lithium-geospeedometry: Quantifying rapid geological processes

3+

L.C. CAVÉ AND T.A. AL

LAURENCE COOGAN1, SIMONE KASEMANN2 3 AND SUMIT CHAKRABORTY

University of New Brunswick, Fredericton, Canada ([email protected], [email protected]) 1

The valence state of Fe in minerals, may be used as a paleoredox tool for studies of past climate change or deep geological waste disposal applications. Fe3+/ΣFe ratios in primary silicates may preserve a record of past oxidation even if reductive dissolution later removes Fe oxyhydroxide evidence that O2 once was present. We have modified a TEM electron energy loss spectroscopy (EELS) method (van Aken and Liebscher 2002) for measuring Fe3+/ΣFe. Mineral standards of known Fe3+/ΣFe were used to calibrate a relationship based on I(L3)/I(L2), the intensity ratio of Fe L3 and L2 edges in EELS core loss spectra. Measurements of Fe3+/ΣFe were then made on biotite grains located adjacent to a shallow fracture in the Lac du Bonnet granite batholith (Whiteshell, Manitoba). This biotite has high Fe3+/ΣFe, even at distances a few cm from the fracture, providing evidence of a reaction with oxygenated groundwater at shallow depths. There may be zonation of Fe3+/ΣFe inwards from the grain boundaries, but it is not clearly discernable at our scale of investigation (Fig. 1).

School of Earth and Ocean sciences, UVic, Canada ([email protected]) 2 School of Geosciences, Edinburgh University, UK ([email protected]) 3 Institut fur Mineralogie, Ruhr-Universitat Bochum, Germany ([email protected])

Reference

Lithium diffusion in plagioclase has been known to be extremely rapid for a number of years [1] and hence the distribution of Li within and between phases may record the rates of short-lived, or low-temperature, processes. We have developed a Li-geospeedometer based on the down temperature diffusion of Li from plagioclase into clinopyroxene to allow us to quantify rapid geological processes. High temperature experiments were performed to quantify the temperature dependence of the equilibrium distribution of Li between plagioclase and clinopyroxene between 900°C and 1200°C. Lithium partitioning between these phases is highly sensitive to temperature changing by an order of magnitude over this temperature range. At lower temperatures Li partitions more strongly into clinopyroxene than at higher temperatures. High temperature experiments have also been undertaken to determine the diffusion coefficient of 6Li in clinopyroxene between 800°C and 1100°C. Self diffusion of lithium in clinopyroxene is approximately two orders of magnitude slower than in plagioclase over this temperature interval – still extremely rapid compared to most cations. These new experimental data, along with the published diffusion coefficient for Li in plagioclase [1] calibrate the Ligeospeedometer for Li exchange between plagioclase and clinopyroxene during cooling. We have applied this geospeedometer to determine the cooling rate of the sheeted dike complex of the oceanic crust. We find that the upper half of the sheeted dike complex cools very rapidly (~500°C/hr) indicating that this can supply the heat required to drive hydrothermal megaplumes. The lower portion of the sheeted dike complex cools much more slowly (~5°C/hr) supply a longer lasting heat (and perhaps nutrient) source to the overlying hydrothermal system.

van Aken, P.A. and Liebscher, B., (2002). Phys Chem Minerals 29. 188-200.

Reference

Figure 1: Biotite Fe3+/ΣFe versus distance from the nearest visible grain boundary. Error bars show 95% confidence limits.

Fe3+/ΣFe (biotite)

1.0 0.8 0.6 0.4

Grain distance from fracture:

0.2

5 mm

6 mm

20 mm

0.0 0

10

20

30

40

50

60

Distance from grain boundary (µm)

Analytical TEM can detect mineral alteration at the nm scale, appropriate for O2 diffusion along mineral grain boundaries. This spatial resolution is not available with other Fe3+/ΣFe methods, such as Mossbauer spectroscopy.

[1] B. J. Giletti, T. M. Shanahan, Chem Geol v139, 1997.

Goldschmidt Conference Abstracts 2005 Diffusion-Reaction Systems

A179

Influence of Temperature on HTO and 36Cl- Diffusion in Bentonite and Callovo-Oxfordian clays

Reactive transport experiments and modelling of CO2 sequestration in deep aquifers

F. GOUTELARD, Y. CHARLES AND E. TEVISSEN

XAVIER GUICHET1, JACQUES SCHOTT1, ERIC H. OELKERS1, BENOÎT VINCENT2, CAROLINE MAGNIER2 AND ETIENNE BROSSE2

CEA/DEN/DPC bât. 450 91191Gif sur Yvette Cedex France The Callovo-Oxfordian formation at Bures in the Paris Basin (France) has been chosen as a potential host rock for deep radioactive waste disposal. Bentonite MX80 is proposed as buffer barrier surrounding the canisters. These materials have very low permeability and diffusion is like-ly to be the predominant mechanism transfer governing radionuclide migration. High level waste disposal will induce a temperature increase in these barriers. In a dilute solvent, this effect is described by the Stokes-Einstein law D∞IS = k T /6π ηS rI whith D∞IS the diffusion coefficient of a solute I infinitely diluted, k the Boltzmann constant, T the temperature, rI the ionic radius and ηs the viscosity of the solvent. The effective diffusion coefficient of solutes in porous media is linked to the diffusion coefficient in free-water, porosity, tortuosity and constrictivity. Moreover, the effective diffusion coefficients of halides are reduced by anionic exclusion. In compacted clays, the physical prop-erties of the interstitial water is strongly affected by liquid-solid interfaces. As a matter of facts, the use of Stokes-Einstein law in order to evaluate the effect of temperature on diffusion coefficients in such media is questionable. In order to check the validity of Stokes-Einstein law for HTO and 36Cl in compacted bentonite and Callovo-Oxfordian argillite, through-diffusion experiments have been carried out to quantify the effective diffusion coefficient for HTO and 36 Cl. After the permanent diffusion state is reached at 20 °C, the temperature is increased step by step (35-50-65-80°C). Two chemical conditions have been selected. Effect-ive diffusion coefficient increase by a factor 3 to 5.7 and 4 to 7.5 for HTO and 36Cl- respectively as temperature increases from 20°C to 80°C. Hysteresis phenomenon has been seen: effective diffusion coefficient at 20°C after and before heating are significantly different. Assuming that pore water viscosity is equal to bulk water viscosity and that formation factor is constant in the range investigated, validity of the StokesEinstein law for tritium and 36Cl- in theses porous medias has been analyzed. For tritiated H2O, it hasn't been possible to clearly validate or invalidate the law, notably because of the hysteresis effect. Diffusion of 36Cl does not follow the StockeEinstein law, so assumptions previously cited aren't valid. Arrhenius type low fit well the experimental data as soon as the diffusion coefficient at 20°C after heating is removed. Activation energies are coherent with literature data.

1

LMTG UMR 5563, 14 avenue Edouard Belin, 31400 Toulouse, France (Xavier. [email protected]) 2 Institut Français du Pétrole, 92852 Rueil-Malmaison Cedex Reliable estimates of the capacity and sustainability of CO2 geological storage sites require, among other things, rigorous modeling of CO2-rock interactions. Central issues for this purpose are the characterization of the elementary reactions involved in aqueous CO2-mineral interactions as well as the generation of comprehensive thermodynamic and kinetic data bases describing the rates of carbonate and silicate minerals dissolution and the rate of carbonate mineral precipitation under conditions of CO2 pressure similar to those of deep geological injection. To validate at a centimetric scale the elementary mechanisms and values of parameters generated at the microscopic scale, we have performed CO2 reactive transport experiments in porous limestones cores (Lavoux, France) using an externally heated (up to 100°C) single-pass plug flow percolation cell operating under N2 confining pressure up to 60 bars. Analyses of reacting solutions (calcium, alkalinity) were combined with in situ pH and permeability measurements, and XR tomography of limestone cores before and after reaction. The input solution for these experiments was injected into the mixed flow reactor at a flow rate of 0.1 mL/min using a single-piston Gilson pump. This study presents the first experimental results obtained with this new device. The limestones cores have a porosity of 20%, and permeabilities ranging from 10 to 20x10-12 m2. X-Ray diffraction (XRD) and optical microscopy indicated that the samples are essentially pure calcite. All experiments were performed at 25°C using 0.1 M NaCl input solutions equilibrated with a CO2 pressure of 10 bars. The in situ measured pH of the input solution was 3.4, whereas the in situ outlet pH ranged from 7.5 to 8. The outflow solution was supersaturated with respect to calcite. The positive feedback between fluid transport and mineral dissolution causes complex reaction front morphologies such as fingers or wormholes for fluid flows of 0.1 mL/min and 10 bar of PCO2. DIAPHORE reactive transport code, developped by IFP (French Petroleum Institute) was used to model the temporal evolution of reactive surface areas, porosities and chemical composition of the oulet solutions.

A180

Goldschmidt Conference Abstracts 2005 Diffusion-Reaction Systems

A Donnan diffusion model for the description of Sr adsorption kinetics to hydrous ferric oxide

Mixing solutions, precipitation and changing permeability in porous media

A. HOFMANN1, W. VAN BEINUM2, J.C.L. MEEUSSEN3 1 AND R. KRETZSCHMAR

G.D. REDDEN, Y. FUJITA, M.E. DELWICHE, T.A. WHITE, T. RONEY, R.J. VERSTEEG, D.T. FOX, AND C.D. PALMER

1

Institute of Terrestrial Ecology, ETH-Zurich, Grabenstr. 3, CH-8959 Schlieren ([email protected], [email protected]) 2 Central Science Laboratory, Sand Hutton, York YO41 1LZ, UK ([email protected]) 3 Energy Research Center of the Netherlands, P.O. box 1 NL1755 ZG Petten ([email protected]) Sorption of ions by hydrous ferric oxide (HFO) often shows a fast initial sorption reaction followed by a much slower sorption process. The second step is diffusion controlled with a rate that may be explained by electrostatic interactions in the material’s pores. HFO is made of small crystallites that are aggregated, generating pores in the nanometer range. The pores may be large enough to contain unbound water but small enough to be electrically charged due to the overlapping of the electric diffuse layers inside the pores. The Donnan diffusion model describes this case. It uses a Donnan electrostatic model to calculate the pore solution chemistry, which is influenced by pH and sorption dependent surface charge. Ion diffusion in the pore water is calculated from the gradients in Donnan concentrations and Donnan potentials. Surface chemical equilibria were described with a 1-pK basic Stern surface complexation model based on the CD-MUSIC model [1]. The model was implemented with ORCHESTRA [2]. To test the Donnan diffusion model, dense aggregates of HFO [3] were packed in a chromatographic column and equilibrated with an electrolyte solution of 10-2 M NaNO3 to a given pH (pH 4 and 7). Then an identical solution, containing additionally 10-4 M Sr was injected. To interpret the sorption behavior, Sr breakthrough curves were recorded and compared with the predictions of the Donnan-diffusion model. The initial part of sorption was fast, which may be explained by diffusion in large pores. An unexpected result was that the fraction of pores with fast diffusion depended on pH, so that this fraction had to be determined specifically for every experiment. For the diffusion limited part of sorption, good correspondence between model and experiment indicated that electrostatically constrained diffusion did influence sorption kinetics significantly and that the model was applicable.

References [1] Hiemstra, T. and Riemsdijk W. H. (1996) J. Coll. Interf. Sci. 179 488-508. [2] Meeussen J.C.L. (2003) Environ. Sci. Technol. 37, 11751182. [3] Hofmann A. et al. (2004) J. Coll. Interf. Sci. 271 163-173.

Idaho National Laboratory, Idaho Falls, ID, USA ([email protected]) Modifying subsurface permeability and altering flow paths is important for many environmental applications. Experiments have been conducted to evaluate different approaches for generating in situ supersaturation states for calcite and controlling the spatial distribution of calcite precipitation within porous media. One such approach is diffusive mixing along parallel flowstreams of reactants as shown in the figure. A solution of calcium chloride was injected through a porous diffuser into the interior of a 5cm diameter sand column with a background flow of sodium bicarbonate. Mixing occurred along the interface between the two solutions. X-ray tomography was used to image the three dimensional distribution of 2-D X-ray Ca2+ tomographic calcium carbonate solids in image: Light the column. Calcium regions show carbonate propagated carbonate. HCO3along the interface between the solutions as mobile colloids or was deposited in the pores, thereby separating the two solutions. Complex resistivity is also being used to detect the formation of calcite. A similar experiment in two dimensions allowed direct observation of the mixing zone using colored dyes and spectral analysis. An additional approach under investigation has been to generate at least one of the reactants in situ. For example, we have been testing the use of immobilized urease enzyme to generate carbonate ions from urea hydrolysis as an abiotic analog to a remediation approach for immobilization of strontrium-90 based on co-precipitation in calcite (Smith, R.W. et al. EMSP project). The long-term goals of these experiments are to understand flow-precipitation coupling and how the impact on system permeability can be controlled, at scales from the pore level to the field.

Goldschmidt Conference Abstracts 2005 Diffusion-Reaction Systems

Rate control in low porosity diffusion-reaction systems CARL I. STEEFEL1, SUSAN L. BRANTLEY2, ALEXIS K. NAVARRE3 AND QINHONG HU4 1

Earth Sciences Division, Lawrence Berkeley National Laboratory, 1 Cyclotron Road, Berkeley CA 94720, USA ([email protected]) 2 Department of Geosciences, Pennsylvania State University, University Park PA 16801, USA ([email protected]) 3 Department of Geosciences, Pennsylvania State University, University Park PA 16801, USA ([email protected]) 4 Chemical Biology and Nuclear Science Division, Livermore National Lab, Livermore, CA 94550, USA ([email protected]) Diffusion-reaction processes in low porosity and permeability rocks have been rarely studied quantitatively. The rate control in these systems, whether by diffusion or by surface reaction, is therefore largely unknown. Such systems can be understood, however, by collecting additional information to constrain the individual dynamic processes. For example, measurement of tracer diffusion rates in the low porosity material makes it possible to uniquely constrain in situ integrated reaction rates given a set of elemental or mineralogical spatial profiles. Alternatively, the system can also be constrained uniquely given a knowledge of the position of reaction fronts as a function of time and their thickness, since the reaction front thickness indicates the ratio of the reaction rates to the diffusion rate. Such systems may be further complicated by reactioninduced porosity and permeability change. If the porosity and permeability are enhanced sufficiently by reaction, then flow can occur where the porosity of the rock has increased, thus resulting in a system with more than one transport mechanism. It is also possible in many cases that diffusion-controlled reaction fronts require the enhancement of porosity via reaction to advance at all. The fact that fronts for minerals with different dissolution rate constants often coincide closely in space may be an indication that the rate of porosity enhancement is controlling the rate of front propagation. These questions are addressed through a combination of multicomponent reactive transport modeling, microscopic mineralogical and elemental characterization, and experimentation.

A181

A182

Goldschmidt Conference Abstracts 2005 Dissolved Organic Matter

Solubility controls that determine dissolved organic matter composition of surface- and ground-waters JERRY A. LEENHEER AND ROBERT L. WERSHAW U.S. Geological Survey,Denver Federal Center, Denver, Colorado, USA ([email protected], [email protected]) Various studies of organic precursors, diagenesis, and removal by water treatment of dissolved and colloidal organic matter in diverse surface- and ground-waters found dissolved organic matter (DOM) to be derived from amino sugar, condensed tannin, lignin, and terpenoid precursors. Terpenoid-derived DOM is not removed by sorption on sesquioxide coatings during infiltration of surface water into groundwater, and is not removed by water-treatment flocculation with ferric and aluminum salts (Leenheer et al., 2003). “Black waters”, such as the Suwannee River, contains DOM derived from condensed tannins (Leenheer and Rostad, 2004) which is almost completely removed by water-treatment flocculation with ferric and aluminum salts (Croue et al., 2000). Recent studies of fulvic acids isolated from wheat straw (Wershaw et al., 2003) and from the Neversink Reservior water supply of New York City found methyoxylignin structures that did not bind to iron and aluminium sesquioxides. Therefore, both fulvic acid ligand structures and mineral coatings containing iron and aluminum sesquioxides act a solubility controls on DOM concentrations and composition in natural water.

References Croue, J-P., Korshin, G.V., and Benjamin, M., (2000), AWWARF Report 90780, Denver, CO. Leenheer, J.A., Nanny, M.A., and McIntyre, C., (2003), Environ. Sci.Technol, 37, 2323-2331. Leenheer, J.A., and Rostad, C.E. (2004), USGS Sci. Invest. Rept. 2004-5216. Wershaw, R. L., Rutherford, D. W., Leenheer, J. A., Kennedy, K. R., Cox, L. G., and Koci. D. R., (2003), USGS Wat.Res, Invest. Rept. 03-4213.

A new understanding of reactivity and composition of humic substances using modern NMR and electrospray ionization mass spectrometry PATRICK G. HATCHER1, WILLIAM C. HOCKADAY1, AMANDA M. GRANNAS1,2 AND SARAH L. CACCAMISE1 1

Department of Chemistry, The Ohio State University, Columbus, OH 43210, USA ([email protected]) 2 Department of Chemistry, Villanova University, Villanova, PA 19085, USA ([email protected]) The advent of modern analytical methodologies, namely nuclear magnetic resonance (NMR) and electrospray ionization coupled to ultrahigh resolution mass spectrometry (MS) have provided a wealth of new structural information that has allowed for a more advanced understanding of the chemical structure and reactivity of natural organic matter (NOM) from a variety of environments. !- and 2-D solution NMR techniques have demonstrated the presence of some new types of components from humic substances in different environments. Polyhydroxylated alicyclic structures are clearly noted as having proteinaceous origins and carboxylated condensed aromatic structures are probably derived from soot and charcoal (black carbon). Electrospray ionization MS has opened the door to obtaining detailed elemental compositions for nearly all ionizable components of NOM. With ultrahigh mass accuracy we can assign unique elemental formulas to the vast number of peaks observed. We can clearly differentiate black carbon components, from lipid-like substances, and from other biochemical components that contribute to NOM. We find that a significant amount of black carbon comprises NOM.

Goldschmidt Conference Abstracts 2005 Dissolved Organic Matter

Sulfide ligands in natural organic matter (NOM) J.R. KRAMER1, R.A. BELL1 AND D.S. SMITH2 1

McMaster University; Hamlton, ON, Canada ([email protected], [email protected]) 2 Wilfrid Laurier University; Waterloo, ON, Canada ([email protected]) Group B metals, such as Hg, Cu, Ag, Pb and Cd bind strongly to reduced inorganic and organic S(II-) ligands [1]. Most trace metal speciation studies until recently have been at elevated (micro-molar) metal concentrations and have blurred the effect of M-S binding for natural occurrences (pico- to nano-molar). S(II-) ligands in oxic waters are shown to exist at the 10-100s nM concentrations [2], and these ligands suppress the toxicity of Group B metals significantly [3]. We have developed a procedure for the determination of metal bound sulfides by use of Cr(II) and a purge/trap method [2]. The resulting chromium reducible sulfide (CRS) does not detect those thiolates found in nature. CRS has been shown to be equivalent to strong total ligand, LT, that has been determined by competitive ligand titration scheme using Ag as the probe metal [4]., for waste water plant effluent. S(II-) ligands are associated with organic matter. There is a linear correlation of CRS (nM) with organic C (mg C/L) (OC). The slope of CRS-OC is 14.9nM CRS/mg C (r2=0.75), reflecting about 0.02 % S ligand in NOM. The scatter of the slope is quite large, however, suggesting direct measure of CRS is needed for accurate and specific results. The CRS is postulated to reflect an M-S coordination (M = Cu(I)?) in NOM. We used Ag as a metal titrant in a competitive ligand setup to determine the conditional binding constant, Log K’, for Ag and strong ligand. Log K’ values decreased linearly with fraction of LT filled from 12 to 8.8 for rivers, but were nearly constant for waste water treatment effluent (log K’ = 11.3).

References [1] Smith, D.S., R.A. Bell, J.R. Kramer, (2002) Comp. Biochem. Physiol. C, 133: 65-74. [2] Bowles, K.C., M.J. Ernste, J.R. Kramer, (2003) Anal. Chim. Acta, 477:113-124. [3] Bianchini, A., K.C. Bowles, C.J. Brauner, J.W. Gorsuch, J.R. Kramer, C.M. Wood, (2002) Environ Toxicol. Chem. 21:1294-1300. [4] Smith, D.S.; R.A. Bell, J.Valliant, J.R. Kramer, (2004) Environ. Sci.Technol., 38:2120-2125.

A183

Stochastic synthesis of DOM: Predicting Cu(II) complexation from precursor structures STEPHEN E. CABANISS Dept. Chemistry, University of New Mexico, Albuquerque, NM 87131, USA ([email protected]) Equilibrium models of metal complexation by dissolved organic matter (DOM) typically are calibrated using experimental measurements from metal-DOM solutions. These models can represent the binding data quite well, but the model parameters are not easily interpreted on a molecular level. In addition, DOM from different sources with significantly different binding properties must calibrated separately. A stochastic, agent-based model of DOM synthesis from precursor molecules has been devised which treats DOM as a complex mixture of interacting molecules, each with a (potentially) unique structure. A quantitative structure-activity relationship (QSAR) using the same data model predicts the log KCu of well-defined ligands with r2 of 0.89 for a calibration data set which ranges over 14 orders of magnitude. This QSAR is used to predict log KCu of the molecules in simulated DOM samples, and these values are then used to predict Cu(II) complexation. Simulations beginning with small organic precursors (tannins, terpenes and flavonoids) produce N-poor DOM with Cu(II)-binding behavior similar to Suwannee River fulvic acid. Simulations beginning with biopolymers (protein, lignin) produce N-rich DOM with stronger Cu-binding behavior similar to that observed in the field. In the former simulation, but not the latter, log KCu correlates with molecular weight. This agent-based model calibrated on known ligands currently fits experimental data less well than empirical models calibrated on NOM-metal equilibrium data, but has the advantages that 1) specific binding sites have structural properties which can be interpreted on a molecular level and 2) different ecosystems and DOM treatments produce distinct Cu-binding behaviors without separate calibration.

Goldschmidt Conference Abstracts 2005 Dissolved Organic Matter

A184

Modelling metal-gill interactions and metal toxicity to fish: The influence of natural organic matter source

Metal-organic matter interaction: Ligands as a functional group in oceanic DOM

RICHARD C. PLAYLE

K. HIROSE

Department of Biology, Wilfrid Laurier University, Waterloo, Ontario, Canada, N2L 3C5 ([email protected])

Geochemical Research Department, Meteorological Research Institute, 1-1 Nagamine, Tsukuba, Ibaraki 305-0052, Japan ([email protected])

Natural organic matter (NOM) binds metals, decreasing the amount of metal binding to fish gills and therefore decreasing metal toxicity to fish. Work from my laboratory has shown that optically darker, more allochthonous NOM binds metals like Cu, Pb, and Al better than does optically lighter, autochthonous-like NOM, as judged by metal binding to gills of rainbow trout (Oncorhynchus mykiss) and by metal toxicity to trout. In these experiments, NOM was isolated by reverse osmosis from diverse sources and added at up to 10 mg C/L. A good index of NOM source is the Specific Absorbance Coefficient (e.g., SAC340). Excitation-Emission matrix spectroscopy is also a good method to characterize NOM. In contrast to the results with Cu, Pb, and Al, NOM source appears to have minimal influence on the degree of inorganic Hg, Ag, and Cd binding to fish gills. The pattern appears to be that metals which bind more strongly to fish gills (higher metal-gill log K values) than to NOM (lower metalNOM log K values) are not influenced as much by NOM source. These results have important implications for developing Biotic Ligand Models, which integrate water chemistry and organism physiology to better predict metal interactions with aquatic organisms.

References Klinck, J., M. Dunbar, S. Brown, J. Nichols, A. Winter, C. Hughes, and R.C. Playle. (2005) Aquat. Toxicol. In Press. Luider, C.D., J. Crusius, R.C. Playle, and P.J. Curtis. (2004) Environ. Sci. Technol. 38: 2865-2872. Paquin et al. (2002) Comp. Biochem. Physiol. 133C: 3-35. Playle, R.C. (2004) Aquat. Toxicol. 67: 359-370. Schwartz, M.L., P.J. Curtis, and R.C. Playle. (2004) Environ. Toxicol. Chem. 23: 2889-2899.

There is little understanding of ecological roles of oceanic dissolved organic matter (DOM) except one of food chain of marine bacteria. Chemical speciation studies revealed that trace metals (Cu, Fe, Zn an others) dissolved in seawater form complexes with binding sites (ligands) in DOM. However, we have no information about why most of the dissolved trace metals are associated with dissolved organic ligands (DOLs). In order to have better understanding of ecological roles of metal complexes in seawater, it is necessary to know chemical forms of DOLs in seawater. According to present knowledge of the DOLs being bound with metals in seawater, two classes of the DOLs coexist in seawater, which are classified as type-I (DTPA) and type II (EDTA) ligands. We carried out speciation of the DOLs using estimated conditional stability constants of metal complexes with the DOLs in seawater. The results reveal that major part of the DOLs in seawater exist as complexed form, in which major species of the EDTA type ligand are Ca and Mg complexes. Therefore, only several percents of total ligand concentrations in sea water are present as non metal-binding forms. This finding suggests that free ligand concentrations in seawater, which are linearly related to the total ligand concentrations, show no drastic change due to variations of the ligand and trace metal concentrations because of constancy of Ca and Mg concentrations in seawater. To be controlled the free ligand concentrations in seawater means that free trace metal concentrations in seawater are controlled at optimal levels in growth of marine organisms.

Goldschmidt Conference Abstracts 2005 Dissolved Organic Matter

Organo-colloidal control on trace element distribution in shallow groundwaters: Fingerprinting by ultracentrifugal cells A. DIA,. E. MORIN, O. POURRET, G. GRUAU, M. DAVRANCHE AND O. HENIN CNRS UMR 6118 Géosciences Rennes, 35042 Rennes cedex, France ([email protected]) Dissolved Organic Matter (DOM) is ubiquitous in aquatic environments and plays a key role in the geochemistry of major and trace elements - acting as a major carrier and transport phase - through complexation, adsorption, dissolution … reactions. It both does interact with mineral phases modifying the exchange rates with solutions and constrain part of pollutant mobilities such as that of trace metals or hydrophobic organic compounds. Not only does association with DOM influences the mobility of metals through the soil/water system, it also affects their bioavailability and toxicity. In DOM-rich soil waters, Rare Earth Element (REE) and other trace element concentrations seem to be controlled by seasonal dynamics, involving both temperature whose onset at spring leads to higher organic matter decomposition rates by microbial mass, and redox changes resulting in REE and other trace element release in water when soil mineral phases occur to dissolve. In order to study DOM-metal interactions and the role of the colloidal pool in such groundwaters, small new ultracentrifugal filter units have been tested in a range of molecular weight cut-off - 30 kDalton, 10 kDalton and 5kDalton - to distinguish between organically colloidal complexed metals and ‘free’ metals. These ‘free’ metals consisting of hydrated metal cations and soluble inorganic metal complexes are considered to represent the bioavailable metal fraction. The Dissolved Organic Carbon (DOC) concentrations were determined in the different fractions as well as the trace metal concentrations by ICP-MS. While Na, Rb, Mg or Ca appear to behave as ‘free’ ions, aqueous REE concentrations are directly correlated to DOC concentrations as well as that of Th and U, suggesting that the organic colloidal pool dominates the carrying of filter-passing REE, Th and U. The REE concentrations are lowered upon successive filtrations at decreasing pore size when REE patterns remain unchanged. Since the retention, transport and fate of trace metals mediated by organic matter has to be better constrained to understand the functional role played by DOM at the soil-water interface regards to trace metal dynamics, further studies dedicated to space- and time-linked variations of the joined colloidal pool and trace element distribution, are at present undertaken in waters recovered from different DOC-rich waterlogged environments.

A185

Understanding the redox properties of Georgetown NOM JAMES T. NURMI1, BUMHAN BAE2 1 AND PAUL G. TRATNYEK 1

Department of Environmental and Biomolecular Systems, Oregon Health & Science University, 20000 NW Walker Rd. Portland, Oregon 97006, USA ([email protected]) 2 Department of Civil & Environmental Engineering, Kyungwon University, South Korea ([email protected]) It is well known that NOM can act as a reducing agent (reductant). More recently, there has been growing interest in the role that NOM plays as an electron shuttle (mediator) in biogeochemical cycles and contaminant fate. For example, recent data from our lab indicate that the addition of some types of NOM (or fractions of NOM) increase abiotic redox reactions between Fe0 and RDX. Such effects imply that NOM must have redox-active moieties with relatively low formal reduction potentials. To characterize these reduction potentials, we developed an electrochemical protocol that allowed us to determine the electrochemcial properties of NOM, fractions of NOM, and NOM model compounds. Here, we report on our most current electrochemical characterization of Georgetown NOM and its fractions. The results show one or more well-defined peaks— presumably indicative of a dominant redox active moiety— with quantifiable reduction potentials and currents.

Figure 1. Representative cyclic voltammograms obtained using 3 amperometric methods on 5 samples of NOM.

References Nurmi, J. T., Tratnyek, P. G. (2002). Environ. Sci. Technol. 36. 617-624. Nurmi, J. T., Tratnyek, P. G. (2002) Proc. 20th IHSS Conf. Northeastern Univ., Boston, MA. pp. 58-60.

Goldschmidt Conference Abstracts 2005 Dissolved Organic Matter

A186

Metal distribution with different molecular size fractions of dissolved organic matter in stream waters by HPSEC and ICPMS 1,2

1

2

2

F.C. WU , H.X. ZHANG , R.D. EVANS , P.J. DILLON 3 AND S. SCHIFF

Metal binding to NOM determined using component resolution and multiresponse modelling D. SCOTT SMITH Deptartment of Chemistry, Wilfrid Laurier University, Waterloo, ON, Canada ([email protected])

1

State Key Laboratory of Environmental Geochemistry, Institute of Geochemistry, Chinese Academy of Sciences, Guiyang / Guizhou, 550002, China.([email protected]) 2 Environmental and Resource Studies, Trent University, Peterborough, Ontario K9J 7B8, Canada. 3 Department of Earth Sciences, University of Waterloo, Ontario N2L 3G1 Canada. The distributions of metals Fe, V, Ce, Th, U, Mo, Cu, Ni, Co, Cr, Zn, Pb and Cd with different molecular size (MS) fractions of dissolved organic matter (DOM) in stream waters from south-central Ontario, Canada were investigated using high-performance size-exclusion chromatography (HPSEC, YMC-Pack Diol-300) coupled with on-line photodiode array detector and high-resolution inductively coupled plasma mass spectrometry (ICPMS, Finnigan MAT ELEMENT2). The MS distribution of metals in DOM fractions was evaluated and compared for reverse osmosis (RO) concentrated DOM samples and XAD-isolated humic substances (HS). The results show the following decrease order of the average molecular weight in RO-concentrated samples: Cu>Ni>(Co, Zn, Cr)>Pb>Cd for the DOM-bound complexes of transitional metals, which is consistent with IrvingWilliams series, and (Fe, V, Ce)>Th>U>Mo for the DOMbound complexes of the other metals, indicating that the metal distribution among the different MS fractions was mainly related to its binding strength. Metals with high strength were more distributed in the larger MS fractions, and metals with low strength were more distributed in the smaller MS fractions of DOM. The MS distribution of metals in HS was different from that in RO-concentrated samples. The mechanisms for these observations were proposed. This study may have significant implications in the understanding of metal-DOM complexation in aquatic environments.

Natural organic matter (NOM) has fluorescent components, which can be characterized using excitation versus emission fluorescence scans. Through the application of numerical mixture resolution techniques it is possible to resolve a minimum number of fluorescent components necessary to describe the total fluorescence surface. Types of NOM can be “fingerprinted” in this manner. Fluorescence surfaces for samples of a wide range of NOM isolates were subjected to the SIMPLISMA component resolution technique (Windig and Guilment, 1991). Results show that four major components can be identified, two amino acid like components (trytophan and tyrosine) and two longer wavelength "fulvic" and "humic" components. Samples where the NOM is algal in origin (e.g., Lake Erie sample) show strong tyrosine-like components, whereas sewage effluent show strong tryptophan-like components. Representative samples of allochthonous organic matter show mainly fulvic and humic-like fluorescence. Samples of mixed origin, such as Lake Ontario organic matter, have contributions from all major components. The fulvic and humic-like components can be further resolved into subcomponents and possible molecular analogs for NOMbuilding blocks proposed (Smith and Kramer, 2000). Once fluorescent components are identified using mixture resolution techniques fluorescence changes for each component can be measured during metal titration (Smith and Kramer, 2000). This resultant multiresponse data can be fit to a chemical equilibrium binding model for metal with multiple sites in NOM. Results for Suwannee River fulvic acid show a mixture of five diprotic/bidentate sites for proton/metal binding with binding constants consistent with salicylic acid – like binding (Smith et al., 1999). Sewage derived organic matter tends to be higher in the amino acid-like components which show relatively large (logK 5) binding sites for silver.

References Windig W. and Guilment J., (1991). Anal. Chem. 63. 14251429. Smith D.S., Adams N.W.H. and Kramer J.R. (1999) Geochim. Cosmochim. Acta 63 337-3347. Smith D.S. and Kramer J.R. (2000) Anal. Chim. Acta 416. 211-220.

Goldschmidt Conference Abstracts 2005 Dissolved Organic Matter

A187

Fluorescence characterization of dissolved organic matter in a city river of southwestern China

Mechanisms of organic matter and rare earth element release in soils: Experimental evidence

P.Q. FU, F.C. WU1 AND C.Q. LIU

M. GRYBOS, G. GRUAU AND M. DAVRANCHE

State Key Laboratory of Environmental Geochemistry, Institute of Geochemistry, Chinese Academy of Sciences, Guiyang 550002, P.R. China ([email protected]) Dissolved organic matter (DOM) in Nanming River and its streams, Southwest China was investigated using fluorescence emission spectra, synchronous fluorescence spectra and three-dimensional excitation emission matrix fluorescence (3DEEM) spectroscopy. There was a wide but featureless peak at Em = 430 - 440 nm in the emission spectra, 3 - 4 components can be distinguished from the synchronous fluorescence spectra with a maximum peak at 280 nm. With 3DEEM technique, we can differentiate three major fluorophores in the DOM samples, which were responsible for two humic-like and one protein-like fluorescence (Coble, 1996; Baker, 2001; Wu and Tanoue, 2001). Strong proteinlike fluorescence occurred in most of the samples. Significant relationships were observed between the fulvic-like and protein-like fluorescence, and between individual fluorescence intensity, DOC, PO43-, COD and ammonium concentrations in the river. It is suggested that fluorescence technique is a powerful tool to indicate the pollution situation of city river waters. Most importantly, our results show that municipal wastewater can be the source not only of protein-like fluorescence substances, but also of the humic-like fluorescence substances in aquatic environments.

References Baker A., (2001). Environ. Sci. Technol., 35(5): 948-953. Coble P.G., (1996). Mar. Chem., 51(4): 325-346. Wu F.C. and Tanoue E., (2001). Environ. Sci. Technol., 35(18): 3646-3652.

CAREN, UMR CNRS 6118, Campus de Beaulieu, 35042 Rennes, France. ([email protected]) We investigated the roles played by microbial activity, redox potential variations and pH changes on Organic Matter (OM) and Rare Earth Elements (REE) release in hydromorphic soils. Three types of experiments were conducted: (i) a soil suspension was incubated under anaerobic condition, without pH control (i.e. the pH was let to evolve in response to occurring biogeochemical reactions); this represents the "natural" reference case in which all tested parameters (pH, microbial activity, redox potential) can vary and control OM and REE exchanges between soil minerals and soil solution; (ii) a sterilized soil suspension was incubated under anaerobic conditions at pH 3 with hydroxylamine as chemical reducer; (iii) finally, three soil suspensions were incubated under aerobic conditions at pH 3, 5 and 7. Results demonstrate that large amounts of OM and REE are released from hydromorphic soils when soil Fe and Mn oxyhydroxides are reductively dissolved. The role of micro-organisms appears to be secondary, the latter playing only a catalyst role. Incubations carried out with hydroxylamine at pH3 or in aerobic conditions at variable pH values demonstrate that the main mechanism controlling OM and REE release is OM desorption in response to pH increase. REE patterns were very helpful in reaching this conclusion since OM adsorbed at the surface of soil minerals was found to have very specific and easily recognizable Z-shape REE pattern. This typical Z-shape signature was not found in experiment designed to promote reductive dissolution of soil Fe and Mn oxyhydroxides (flat REE patterns). This difference could be due to the release of REE co-precipitated within the soil oxyhydroxides solid or to the release of OM with different REE complexing capacities. Finally, experiments carried out under aerobic conditions at pH 3 yielded a third type of REE pattern (light REE enriched pattern). As a whole, these results demonstrate that REE may be used as a probe to monitor the activation of particular soil components during soil reduction and soil pH variations.

A188

Goldschmidt Conference Abstracts 2005 Dissolved Organic Matter

Organic speciation of rare earth elements in natural waters: Comparing speciation models and ultrafiltration experiments O. POURRET, G. GRUAU, M DAVRANCHE AND A. DIA CAREN, Géosciences Rennes, UMR CNRS 6118, Campus de Beaulieu, 35042 Rennes Cedex, France ([email protected] / Fax : +33-223-235-787) Two speciation models (WHAM 6 and Visual Minteq) were compared together and with results of ultrafiltration experiments to assess the ability of models to accurately predict the speciation of rare earth elements (REE) in organicrich water. Both river and groundwater samples were used (5.9 < pH < 7.4 ; 1 < Dissolved Organic Carbon < 30 mg.L-1 ; 0.2 < ΣREE < 10 ppb). Ultrafiltration of these samples was performed with a new method using small centrifugal filter devices of decreasing pore size (0.22 µm, 30 000 Da, 10 000 Da and 5 000 Da) to separate the organically bound REE from the inorganic REE species. REE and Dissolved Organic Content (DOC) were analyzed in each fraction. DOC-rich waters (> 4-5 mg.L-1) have a higher proportion of organically bound REE (75 to 95%) than waters with lower DOC contents (50%). REE-complexing organic molecules have higher molecular weights in groundwater samples (10-30 kDa) than in river waters (5-10 kDa). The two tested models yield comparable results, altough some differences are pointed out for the light-REE (i.e. WHAM 6 yields a 71% organic speciation for La while the proportion of organically bound La is 82% when Visual Minteq is used). Model predictions are in good agreement with ultrafiltration results, especially as regards heavy REE. This study shows that speciation models can be confidently used to assess the speciation of REE in circumneutral pH organic-rich waters.

Goldschmidt Conference Abstracts 2005 Earth Materials and Human Health

Libby, Montana: Overview of asbestos exposures and health effects AUBREY K. MILLER, MD, MPH USEPA Region 8, 999 18th Street, Suite 300 (8EPR-PS), Denver, CO 80202, USA ([email protected]) Widespread asbestos contamination and health effects have been documented in Libby, Montana, associated with one of the world’s largest vermiculite mines. Asbestiform amphibole fibers from the Libby area, are predominately comprised of actinolite, tremolite, winchite, and richterite mineral forms of amphibole (Libby Amphibole). About half of the airborne fiber exposures occurring in the Libby area are to asbestifom amphibole fibers that are not commonly included in exposure assessments under the current asbestos analytical protocols and regulatory definitions. The asbestos-related occupational illnesses associated with exposures to Libby Amphibole among former vermiculite miners are well established in the medical literature, with significantly increased rates of asbestosis, lung cancer, and mesothelioma. Recent mortality studies conducted by federal health officials have also documented markedly elevated rates of asbestosis (40-80 times that of the US population), lung cancers, and mesothelioma in the community. Results of large-scale medical testing conducted on over 7300 individuals who lived or worked in the community prior to 1990, revealed the prevalence of asbestos-related lung abnormalities in about 18% of all participants. The prevalence of such abnormalities increased with increasing number of exposure pathways, ranging from 6.7% for those who reported no apparent exposures to 34.6% for those who reported 12 or more pathways. Pathways of concern include both occupational and non-occupation exposures, many of which still exist in the community today, despite the closure of the mine in 1990. Various activities involving remediation of contaminated soils, attic insulation, and indoor dust have been completed at over 350 properties to date, and will continue for years to come.

A189

Asbestos from Libby Montana; Compositions and morphologies that don’t fit current asbestos definitions G.P. MEEKER, H.A. LOWERS, AND I.K. BROWNFIELD U.S. Geological Survey, Denver Microbeam Laboratory, Denver, CO, USA ([email protected]) Asbestos is an industrial term that includes chrysotile, and the asbestiform varieties of the amphiboles tremolite, actinolite, anthophyllite, riebeckite, cummingtonite, and grunerite. Most current regulations and approved analytical methods for asbestos exclude other fibrous and asbestiform amphiboles. In early 2000 it became apparent that fibrous amphiboles in a vermiculite deposit near Libby, MT, were responsible for significant illness in the local population. Naming conventions of the mineralogical community classify the Libby amphiboles primarily as winchite, and richterite, species not listed in the regulations (Fig. 1). In addition to chemistry, the morphology of the Libby amphiboles can be quite different from the commercial asbestos varieties. Differences in chemistry and morphology create problems for the health and regulatory communities charged with protecting public health. These problems include difficulties in identification and complications in relating exposures to existing toxicological studies. It will be necessary in the future to resolve these issues and establish definitions that are compatible with mineralogical and health criteria as well as realistic analytical capabilities. Figure 1. Classification of Libby amphiboles

References B.E. Leake et al., (1997), American Mineralogist, 82, 1019. G.P. Meeker et al., (2003), American Mineralogist, 88 1955.

A190

Goldschmidt Conference Abstracts 2005 Earth Materials and Human Health

Low-level detection of Libby amphiboles in attic insulation

Re-evaluation and re-classification of erionite group minerals

MATTHEW S. SANCHEZ AND MICKEY E. GUNTER

A. UMRAN DOGAN1 AND MERAL DOGAN2

Department of Geological Sciences, University of Idaho, Moscow, Idaho 83844-3022, U.S.A. ([email protected]) Since health concerns surrounding the amphibole asbestos content of attic insulation manufactured from vermiculite ore produced from a now-closed mine near Libby, Montana emerged in 1999, researchers have been seeking methods to detect and quantify its amphibole content. Currently we are working on two separate methods: 1) concentration of amphiboles and 2) powder X-ray diffraction (XRD). For the concentration method, 0.25 g of sample was placed in boiling 12M HCl for 1 hr and then boiling 4M NaOH for 1 hr. Fig 1 (field of view 1 x 1 mm) shows amphiboles, and other residual material, left after this process. For the eleven samples treated with this method, 88 to 99 wt % of mineral matter was dissolved. All four Libby samples contained amphiboles, but the sample in Fig 1 showed the greatest concentration. No amphiboles were observed in the seven other vermiculite products with a non-Libby source. For the XRD method, long count times were used over the (110) peak for 2 g micronized samples placed in a back-fill XRD mount. Fig 2 shows four scans. The one labeled “attic 9s” is a 9s count time for the attic sample we found to contain the most amphibole (Fig 1). The other three scans are all 180s count times (over a shorter scan range) at 0.02° steps, one being the attic sample, but at a greater counting time. The other two are from an amphibole-free expanded vermiculite (based on our concentration and XRD methods) to which we added 0.1 and 1 wt % of Libby amphibole. Based on these scans, our high concentration sample (shown in Fig 1) contains less than 0.1 wt % amphibole. These two methods used in conjunction can first show if amphiboles occur in a sample and then, if so, in what amounts.

1

Ankara University, Department of Geological Engineering, Ankara, Turkey 2 Hacettepe University, Department of Geological Engineering, Ankara, Turkey Erionite has been shown to be an agent responsible for malignant mesothelioma and it is one of the most carcinogenic minerals in the world. Often the erionite specimens were incompletely or incorrectly characterized. Animal and cell experiments with erionite should only be performed with the minerals that have passed the quantitative characterization, both balance error (E%) and Mg-content tests, and the type of erionite (-Ca, -Na, -K) must be identified properly. Large numbers of erionite deposits have been reported from many countries. Erionite from Beach Creek and Durkee, Oregon, USA; British Columbia, Canada; Faedo, Vicenza, Montresta, Nuoro, and Montecchio Maggiore, Italy; Shourdo, Georgia; Jindivick and Phillip Island, Australia passed the E% and Mg-content tests and re-classified as erionite-Ca. Erionite from Crooked Creek, Oregon, Cady Mountains, California, and Durkee, Oregon, USA; Dunseverik, N. Ireland; Montecchio, Maggiore, Italy; Lake Natron, Tanzania; Phillip Island, Australia; Campbell Glacier and Mt. Adamson, Antarctica passed the E% and Mg-content tests and reclassified as erionite-Na. Selected samples of erionite from Cappadocian region of Turkey passed the E% and Mg-content tests and re-classified as erionite-Na (Tuzkoy village) and erionite-K (Karain and Sarihidir villages). However, some published erionite data from Italy were recalculated as erionite-Ca and erionite-Na. Two sets of data from Phillip Island, Australia were re-calculated as both erionite-Ca and erionite-Na. Data from Durkee, Oregon, USA by three different authors were re-calculated as erionite-Ca, erionite-Na, and erionite-K. Mg contents of erionite-Ca from Agate Beach, Oregon, USA; Araules, Ht Loire, France; Maze, Niigata, Japan; and Mg contents of erionite-K from Jersey Valley, Nevada, USA; and Karain, Turkey were higher than 0.8. Thus, these samples did not meet the requirements.

Goldschmidt Conference Abstracts 2005 Earth Materials and Human Health

Using imaging spectroscopy to map ultramafic rocks, serpentinites, and tremolite-actinolite-bearing rocks in California G.A. SWAYZE1, C.T. HIGGINS2, J.P. CLINKENBEARD2, R.F. KOKALY1, R.N. CLARK1, G.P. MEEKER1 1 AND S.J. SUTLEY 1

U.S. Geological Survey, MS964 Box 25046, Denver Federal Center, Denver, CO 80225, USA 2 California Dept. of Conservation, California Geological Survey, 801 K Street, MS 13-40, Sacramento, CA 95814, USA Airborne Visible/InfraRed Imaging Spectrometer (AVIRIS) data were collected in approximately 3-kilometerwide swaths over selected areas in El Dorado and Plumas Counties, California, USA, that contain serpentinite and ultramafic rocks as part of an experiment to determine if potentially asbestos-bearing rocks could be identified spectrally. Mineral maps created from the data were successfully used to delineate exposures of serpentine and tremolite-actinolite/talc in areas with up to 70 percent vegetation cover. In some cases, the vegetation density is so high that it prevented spectral identification of minerals by AVIRIS in those areas. Thus, there may be more serpentine and tremolite-actinolite/talc present than is shown on the mineral maps. Importantly, not all tremolite-actinolite is fibrous; just because tremolite-actinolite was mapped, does not necessarily mean it is tremolite- or actinolite-asbestos. It is difficult to spectrally distinguish tremolite-actinolite from talc using AVIRIS data. Serpentine has been detected outside of known serpentinite outcrop areas, mostly as aggregate that covers dirt roads. Four flight lines of AVIRIS data were analyzed over areas selected to show trends in degree of surface exposure as a function of elevation and vegetation cover. Field checking has verified the accuracy of the mineral maps at 25 accessible locations. Eleven additional flight lines remain to be analyzed and field checked pending future funding. AVIRIS mineral mapping has shown promise as a complement to field mapping but cannot replace it. Because AVIRIS is a remote-sensing technology, the presence of serpentine or tremolite-actinolite would have to be verified in the field by direct observation and by appropriate sampling and laboratory analysis, if needed. At this time, no conclusion regarding the presence or absence of asbestos minerals in the identified areas is possible from the AVIRIS data alone. Identification of asbestos minerals in the identified areas would require appropriate sampling and laboratory analysis of the materials in those areas.

A191

Surface coatings on quartz grains in bentonites and their relevance to human health R.F. WENDLANDT1, W.J. HARRISON1, D.J. VAUGHAN2 2 AND P. WINCOTT 1

Dept. of Geology and Geological Engineering, Colorado School of Mines, Golden, Colorado, 80401 USA ([email protected], [email protected]) 2 SEAES and Williamson Research Centre for Molecular Environmental Science, Univ. of Manchester, Manchester UK ([email protected], [email protected]) In 1997, the International Agency for Research on Cancer recognized that the carcinogenic properties of silica dusts may be dependent on the physical and chemical characteristics of external grain surfaces. Surface coatings on respirable quartz grains are of particular interest as they have been implicated in modifications to cytotoxic reactivity in lung tissue (Fubini, 1998). Our investigations utilize XRD, SEM, ESEM, EPMA, LA-ICPMS, and XPS analyses to characterize the mineralogy, habit, and composition of surface coatings on quartz grains from US deposits of southern (Alabama) and western (South Dakota) bentonites. Quartz contents of bentonites are less than 7.5 wt%, with the <10µm size fraction comprising 6-45% of this total. Surface coatings are pervasive on all quartz grains in the observed bentonites and resist removal by repeated vigorous washings and reaction with HCl. Textural attributes and XPS and EDS analyses of these coatings are consistent with most being either montmorillonite or opaline silica, or a mixture of both. Montmorillonite coatings may be so thin that underlying pre-existing conchoidal fractures are clearly visible, but may also be more than 10 microns thick. Plagioclase and Kfeldspar rarely show well-developed montmorillonite surface coatings, but both have been observed with opaline silica coatings that have a morphology similar to that of opal-CT. Biotite and muscovite grains never have surface coatings. Trace element contents (Fe-Ti-Al) of quartz grains from any given bentonite are very similar, indicating a single origin for the quartz, presumed to be magmatic, with little if any contamination from other sources. Accordingly, detailed surface characterizations of silica dusts are essential precursors to future in vitro and in vivo investigations of silica dust toxicity.

Reference Fubini, B. (1998) Ann. Occup. Hyg. 42, 521-530.

A192

Goldschmidt Conference Abstracts 2005 Earth Materials and Human Health

Factors controlling the bioaccessibility of Pb in polluted soils N. WALRAVEN1, S.P. VRIEND1, B.J.H. VAN OS2, G.TH. KLAVER2 AND A.G. OOMEN3 1

Utrecht University, P.O.Box 80021, 3508 TA Utrecht, The Netherlands ([email protected]) 2 TNO-NITG, P.O.Box 80015, 3508 TA Utrecht, The Netherlands, ([email protected]) 3 RIVM, P.O.Box 1, 3720 BA Bilthoven, The Netherlands ([email protected]) The negative health effects of Pb pollution are longknown. However, it lasted until the 1970’s before legislative measures were taken to reduce the input of Pb and related heavy metals in the environment. Despite these measures, many sites in the world are, as remnant from the past, heavily polluted with Pb. These sites are unsuitable for agriculture and residential building and, if they remain unattended, pose a threat to human health. In human risk assessment, ingestion of soil is considered a major route of Pb, especially for children. A large body of research has focused on the measurement of the ‘total’ Pb content in sediment, soil and dust to determine the potential risk of the Pb polluted environment. However, we found that Pb bioaccessibility, determined with an in-vitro test [1], does not necessarily depend on the total Pb content, but also on the Pb source, chemical speciation and soil characteristics. Initially the Pb bioaccessibility is largely controlled by the chemical form of the Pb source, which determines its solubility. However, when anthropogenic Pb enters the soils, it forms new, more stable, minerals and/or binds to organic matter, clay, iron or other reactive phases. The Pb bioaccessibility of 30 soils, polluted with various Pb sources (o.a. gasoline Pb, gun shot, Pb based paint, Pb glazed ceramics and coal ashes), was determined and varied from 0.5% to 61%. The highest Pb bioaccessibility (56% to 61%) was measured in Pb polluted soils from shooting ranges (native lead). The soils of the shooting ranges are, in contrast to the other studied soils, acidic and contain negligible amounts of reactive phases, such as organic matter, clay, calcium-carbonate and reactive iron. The lowest Pb bioaccessibility (0.5%-8%) was measured in Pb polluted soils from a village already inhibited since Roman times. These soils are characterized by a high clay and intermediate organic matter content. Multiple regression analysis shows that Pb bioaccessibility strongly relates to the content of reactive phases present in soil samples. These results indicate that it is possible to predict Pb bioaccessibility, as determined with the in-vitro digestion model [1], if the content of the reactive phases in Pb polluted soils is known Instead of basing human risk assessment only on total heavy metal contents we propose to base it also on in-vitro bioaccessibility tests, taking factors such as organic matter, clay and reactive iron into account.

References [1] Oomen et al., (2002). Environ. Sci. Technol. 36: 33263334.

A soil geochemical transect in northern California – Links to human health issues M. B.GOLDHABER1, J.M., MORRISON1, R. L. REYNOLDS2, 1 AND D. B. SMITH 1

USGS, MS 973 Denver Federal Center, Denver CO, 80225, USA ([email protected], [email protected], [email protected]) 2 USGS, MS 980 Denver Federal Center, Denver CO, 80225, USA ([email protected]) Soil is literally and figuratively a foundation of human society and may directly affect human health through its ingestion, inhalation, and dermal absorption. This study examined soil samples in a latitudinal transect crossing California from Marin County north of San Francisco to the Nevada border. Soil-related human health issues in the study area center on potentially carcinogenic Cr, Ni, and V enrichments in soil derived from ultramafic belts. We have chemically analyzed and measured magnetic susceptibility (ms) on nearly 2000 soil samples including 1300 shallow soils in the central and eastern portion of the study area (El Dorado, Placer, Sutter, Sacramento, Yolo, and Solano Counties), and 100 soil depth profiles across the transect. Cr, Ni, and V in these samples display distinctive patterns reflecting the underlying geology. Elevated concentrations of Cr, Ni, and V in soil with high ms values (high magnetite abundance) overlie ultramafic rocks and Mesozoic volcanic and metavolcanic rocks in the foothills of the Sierra Nevada. Soils with Tertiary volcanic and granitic parent material had lower Cr, Ni, and V concentrations with high ms. Surprisingly, Cr, Ni, and V enrichment is associated with low ms values in soil from the west side of the Sacramento Valley. The valley alluvial fill was derived in part from sediments transported from the Sacramento River headwaters in the Klamath Mountains and from ultramafic rocks in the Coast Ranges to the west of the valley. Low ms values are attributed to postdepositional alteration of detrital magnetite. We cannot confirm a link between soil geochemistry and human health in the study area. However, breast cancer rates are elevated, ranking near the top for California counties, where elevated Cr, Ni, and V concentrations occur in soils. We are investigating a potential environmental link between soil geochemistry and these high cancer rates through ‘bioaccessibility’ (selective leach) studies and through collaboration with researchers studying the potential for these high Cr soils to cause mutagenesis.

Goldschmidt Conference Abstracts 2005 Earth Materials and Human Health

A193

Distribution of native selenium in Yutangba of China and its environmental implications

Flour pollution in drinking water in Makoo City of West Azarbaijan, Northwest Iran

J.M. ZHU1, G.D. ZHENG2, T. M. JOHNSON3 AND S.H. LI1

F. SAHABI1 AND M. KHEIRKHAH2

Institute of Geochemistry, Chinese Academy of Science, Guiyang 550002, China ([email protected]) 2 Department of Earth and Planetary System Science, Graduate School of Sciences, Hiroshima University, Japan 3 Department of Geology, University of Illinois at UrbanaChampaign, IL, 61801, USA Yutangba, where a sudden incidence of human Se poisoning occurred in 1963(Yang et al. 1983), is located in the nothern part of the town of Shuanghe, about 81km SE of Enshi City, Hubei Province, China. In Yutangba, selenium is present mainly in the carbonaceous chert and carbonaceous shale (locally known as “stone coal”) of the Upper Prmian Maokou Formation. The rock samples have a maximum Se content of 3.0% (Zhu et al. 2004), but little is known about the possible relationships between the sudden human Se poisoning and distribution of native Se in Yutangba. The result of our studies by using SEM-EDX, XANES shows that native Se exists extensively within abandoned stone coal spoils, other carbonaceous rocks and nearby highSe soils. The areal distribution of native Se is mainly consistent with the outcrop locations of Se-rich carbonaceous rocks stata with about 20m thick, which form a oval-shaped belt approximately 25km long. XANES results suggest native Se is probably the dominant primary species in fresh Se-rich carbonaceous rocks. The currently geochemical condition in Yutangba, however, is favorable for its mobilizaton and redeposition as secondary native Se, perhaps at the top of the saturated zone. This redistribution is probably responsible for the extremely high Se concentrations. If these rocks are mined as a fuel or fertilizer, or when abandoned stone coal spoils and high-Se soils are reclaimed for cropland, a great amount of native Se will be quickly oxidized and transformed to soluble Se, which accumulates rapidly in local food chains via irrigation systems. Thus, soluble Se may once again accumulate to such an extent that Se poisoning will occur again in Yutangba, and even in other high-Se places in Enshi prefecture where human Se poisoning was not reported previously.

References Yang G.Q. et al., (1983), Am. J. Clin. Nutr. 37, 872-881. Zhu, J.M. et al. , (2004), Applied Geochemistry. 19,461-467.

Acknowledgement This study is supported by NSF of China (40373040, 40103007)

1

Research Institute of Geoscience, Geological Survey of Iran Olom va tahghighat, Azad University, Tehran, Iran ([email protected])

2

Basaltic lava flows related to Ararat volcanoes in Southeast Turkey covered Quaternary Alluvium in Makoo Area of west Azarbaijan, Northwest Iran. Lava flows include pahoehoe fabrics and mineralogicully charactersed by low K+1 and high Al2O3. High porosity and permeability due to strong fracturing throughout the lava bodies has formed considerable water reservoirs in the area . Water analysis indicate that Fluor content is dierectly proportional to (K+Na) / Ca ratio , and reachs the maximum of 3.5 ppm along the down-stream currents ( fig.1 ).It is concluded that drinking water in the area is polluted by high F. content due to water-rock inter action in the underground water reservoirs. This has been testified by brownish spots on the teeth and bome sclerosis in majority of the native people and animals (fig 2).

Fig (1): Corrolation of major Ion concentrations in the water wells in Makoo city, West Azarbaijan, NW Iran

meq / l

1

Number of wells

Fig (2): Basaltic Spring, 12 Km to NW Makoo

Goldschmidt Conference Abstracts 2005 Earth Materials and Human Health

A194

Importance of pyrite as an arsenic sink in Bengal sediment

The medical geology and geochemistry of mineral deposits

H.A. LOWERS, G.N. BREIT AND A.L. FOSTER

G.S. PLUMLEE

U.S. Geological Survey, USA ([email protected], [email protected], [email protected])

U.S. Geological Survey, MS964 Denver Federal Center, Denver, CO, USA ([email protected])

Introduction

Medical geology and geochemistry use earth science methods and principles to help understand how geologic materials and processes may influence human health. A variety of health issues associated with mineral deposits, mineral resource development, and commercial and industrial mineral products have been recognized by the public health, regulatory, and mining communities for many years. These include, for example, health concerns associated with dusts of biodurable minerals such as asbestos and crystalline silica, which in sufficient dose can cause problems due to their longevity in the respiratory tract. Other health concerns can result from the uptake of bioaccessible metals or other potential toxicants from geologic materials (such as some mine wastes, soils affected by smelter emissions, etc.) that readily dissolve in gastrointestinal or respiratory tract fluids. In some cases the links between particular mine sites and health problems (typically in the miners and mill workers, but more rarely also in workers’ families and residents of nearby towns) are well-defined and documented. However, there are recent examples, particularly in developing countries, where a variety of illnesses have been publically attributed to the effects of nearby mine sites, but subsequent public health assessments have not found clear evidence of a link. Earth scientists are playing a growing role in helping to define and understand potential health issues associated with mineral deposits. Mineralogical characterization of rocks mined from specific deposits helps determine the presence, form, abundance, and morphology of potential toxicants (i.e. asbestos, or minerals containing bioaccessible lead or other metals), and can be used to develop predictive models of mineral deposit types where such components may be present. Integrated geochemical leach tests (using simulated body fluids as the extracting fluids) and in vitro and in vivo toxicity tests of well-characterized ore, waste, and processing byproducts can help understand potential bioaccessibility and toxicity effects of toxicants upon exposure. Pre-mining human health baseline studies, likely to be commonplace in the future, will benefit from earth science information on possible geogenic sources of potential toxicants present in rocks, soils, dusts, and waters as a result of natural erosion and weathering of mineral deposits and host rocks.

Processes that control the concentration of As in Bangladesh groundwater are under investigation. Pyrite in shallow and deep aquifer sediment are characterized to understand the importance of this diagenetic mineral as an As sink.

Discussion of results Pyrite framboids from depths <5 m contain 0.08-0.9 wt% As with an average value of 0.3 wt%. Framboids from sediment below 50 m generally contain less As (~0.2 wt%), but massive pyrite surrounding framboids contains ~0.5 wt% (Fig 1). S and As are strongly correlated in bulk compositions in deep sediment but appear to be independent in shallower sediment. δ34S of pyrite in shallow sediments ranges from –11 to –0.4 ‰, while deeper samples range from –9 to 33 ‰. Figure 1: Backscattered electron image (left) and x-ray map of As in pyrite (right). Brighter areas indicate more As.

Conclusions Pyrite is an important sink for As in some sediment in Bangladesh. The strong relation between S and As in deep sediments may reflect impact of seawater. Pyrite formation in shallow sediment may reflect the influence of fertilizer and could explain low dissolved As in very shallow wells.

References Akai, J., Izumi, K., Fukuhara, H., Masuda, H., Nakano, S., Yoshimura, T., Ohfuji., H., Anawar, H.Md and Akai, K., (2004). App. Geochem. 19. 215-230.

Goldschmidt Conference Abstracts 2005 Earth Materials and Human Health

XIAO-YAN LI , BAO-SHAN ZHENG , YAN WANG 1 AND XUE WANG

1,2,3

1

Institute of Geochemistry, Chinese Academy of sciences, GuiYang, 550002, China ([email protected], [email protected]) 2 Graduate school of the Chinese Academy of Sciences, Beijing, China, 10039 3 [email protected]

Site selection 47 underground working places with different building type were chosen as surveyed site, where the working people worked there for over 8h everyday.

Surveying method Solid state nuclear track detectors (SSNTD’S) was selected. The detectors had been exposed for three months in every site during summer and winter before etched. Data was obtained from reading the detectors using optical microscope.

Conclusions

Department of Geological Sciences, University of Idaho, Moscow, ID 83844-3022, USA ([email protected], [email protected]) The rate of tremolite dissolution was investigated in a flow-through reactor at 37°C, varying solution composition and pH to imitate conditions inside the human lung. Initial experiments were conducted to determine the pH dependence of the kinetics of tremolite dissolution using 1.54M NaCl solution over a pH range of 1 to 7. As expected, the rate of dissolution of tremolite decreases as the solution approaches neutrality. Gamble’s solution, a simulated lung fluid, is composed of several organic species at various concentrations, including: citrate, lactate, glycine, tartrate, and pyruvate, as well as some inorganic salts such as sodium chloride, sulfate, and phosphate. A second series of experiments was conducted to examine the effects of each organic species individually, at the concentration they occur in Gamble’s solution, on the rate of dissolution of tremolite at a given pH (~4.0) (Figure 1). The individual components exhibit varying effects on the rate of tremolite dissolution. -14.90 -15.00 -15.10 -15.20 -15.30 -15.40 -15.50 -15.60 -15.70

aC

l

The average radon concentration in underground working places in winter was apparently less than that in summer. The radon concentration in Shanghai was the lowest among these capitals. The radon level in all capitals are less than the limit of the safe level set for controlling radon and its daughters in underground spaces in China (400 Bq.m-3 EECRn). The average annual effective dose received by working people in these capitals is 1.34 mSv, as a result the lifetime fatality risk is 1.05×10-4. It’s safe for people working there.

B.J. NELSON, S.A. WOOD AND MICKEY E. GUNTER

Ac et at e l+ aC N

GB/T 17216-1998 Hygienic standard for peacetime utilization of civil air defence works. Pan Ziqiang, 2003, Exposure resulted from radon and its decay products in air in China, Radiation protection, 23(30). 134 Surinder Singh, Rajeev Malhotra, Jatinder Kumar, Lakhwant Singh,2001, Indoor radon measurements in dwellings of Kulu area, Himachal Pradesh, using solid state nuclear track detectors, Radiation Measurements, 34, 505-508

N

References

G am bl N es aC l+ C itr N at aC e l+ La ct N at aC e l+ G ly N ci aC ne l+ Ta rt N ra aC te l+ Py ru va te

1

bu ffe r

1,2

Dissolution of tremolite: An experimental study simulating conditions in the human lung

log DR (mol/cm^2s)

A survey of the radon level and the risk to radon exposure in underground working places in capitals in China

A195

Figure 1. A comparison plot of log dissolution rate (DR) of tremolite for solutions containing individual components of Gambles solution at constant ionic strength and pH. The component most effective in increasing the dissolution rate of tremolite, i.e., citrate, was then used in a third set of experiments. As expected, increased concentrations of citrate increase dissolution rate. However, even micromolar concentrations of citrate increase the dissolution rate significantly when compared to a NaCl solution of identical ionic strength and pH.

A196

Goldschmidt Conference Abstracts 2005 Earth Materials and Human Health

Trace metals in soils and their relationship with scrapie occurrance

Characterization of fly ash by SEM/EDS and Raman spectroscopy

K.V. RAGNARSDOTTIR AND D. HAWKINS

B. VALENTIM, A. GUEDES AND D. FLORES

Department of Earth Sciences, University of Bristol, Bristol BS8 1RJ, UK ([email protected])

Centro/Departamento de Geologia, Universidade do Porto, 4099-002 Porto, Portugal ([email protected])

Regional geochemistry has in the past been linked to varous diseases including molybdenosis and/or molybdenum toxicity in sheep, cattle and elk (due to high Mo), selenosis in horses (high Se), Keshan and Kashin-Beck diseases in humans (low Se), karies (F-deficiency), fluorisis (high F) and cancer (high arsenic). Therefore knowledge of regional geochemistry of soils is imperative when studying the epidemiology of a variety of diseses. In this study we are investigating whether trace metals in the natural environment have a role to play in the development of prion diseses. These diseases include scrapie in sheep, BSE in cattle, CJD in humans and CWD in deer. Biochemical research has recently demonstrated that the protein in the brain that are damaged in prion diseases need copper in the prion “tail” to keep their structure. If there is not enough copper the prions can take up manganese and unfold. Molecular dynamics calculations have shown that this prion unfolding process needs copper first to be reduced to Cu+, then copper is replaced by divalent manganese and once manganses oxidises the prion unfolds. There are five steps that are involved with getting an element from soil solids to plant tops: 1) desorption or dissolution from minerals; 2) diffusion and convection within the soil; 3) sorption or precipitation at new sites located on the soil nutrient storage facility; 4) absoption by roots; and 5) translocation from roots to plant tops. For step 4, roots may modify the solution chemistry or the root zone, locally, changing soil properties such as pH or redox potential (rhizosphere effect). Once the plants are consumed by animals or humans we also need to consider the absorption of the metals as well as the function of the metals in the body. Here we have analysed soils from areas where scrapie is prone to develop and other areas where scrapie has never occurred. We have found that in areas where scrapie is prone easily reducable manganese is high. In the same areas bioavailable copper appears to be low. This work is further being extended by investigating the effect of Mo on copper metabolism as well as the bioavailablity of metals as a function of soil pH and organic content.

Introduction Fly ash particles are a sub-produt of coal combustion and their impact on human health is well-known and demands the existence of accurate and sensitive analytical methods for detection and analysis of these particles. Using the combination of Scanning Electron Microscopy (SEM/EDS) and micro-Raman spectroscopy analytical techniques, we have characterised fly ash particles from the stack of a Portuguese thermal power plant and its surroundings.

Results and conclusions The analysis of the stack revealed the existence of glass spheres (calcium-rich; aluminium-silicate with Fe, Mg and Ti; complex aluminium-silicate with Fe, Na, P, S, K, Ca and Ti) (Fig. 1) and carbonaceous particles. In Figure 2 are the Raman spectra of carbonaceous particles sampled in the stack and in the filters from the surroundings. The different spectrum (Fig. 2) led us to consider the contribution of anthropogenic particles from pollutant sources other than the Power Plant. Figure 1: SEM/EDS of glass spheres from the stack.

Figure 2: Raman spectra obtained in carbonaceous particles from the stack and the surroundings. 45

Intensity (a.u.)

35

Stack

25 15 Surroundings 5 -5 1100

1200

1300

1400

1500 -1

Wavenumber (cm )

1600

1700

Goldschmidt Conference Abstracts 2005 Earth Materials and Human Health

Chromium speciation and transformation in atmospheric aerosol particles M.L. WERNER1, P.S. NICO2, C. ANASTASIO1 1

University of California, Davis, Davis, CA 95616, USA ([email protected], [email protected]) 2 California State University, Stanislaus, Turlock, CA 95382, USA ([email protected]) The EPA estimates that 40,000 to 50,000 deaths occur in the United States annually as a result of inhaled particulate matter. Currently, aerosol particles are regulated based on two size categories PM10 (particles < 10 µm) and PM2.5 (particles <2.5 µm). Our size-based understanding of atmospheric particles is relatively crude because it does not account for differences in the chemical composition and reactivity of these particles. Incineration, a common method for the treatment of hazardous wastes, produces metal-containing atmospheric particles. Of the metals potentially released, Cr is of interest because it can exist in two oxidation states with very different toxicities. If atmospheric processes alter the oxidation states of particulate chromium, they should also simultaneously change the Crassociated toxicity of the particles. The transformations of Cr in aerosol particles generated in a labortoray incinerator flame were investigated by bulk XANES, while the speciation of Cr in ambient aerosol particles collected in the region around Sacramento California was investigated using micro-XAS. Simulated atmospheric aging caused only minor changes in Cr speciation under most of the conditions studied, although there was a general trend toward Cr reduction. However, under some conditions important changes were in fact observed. The incinerator flame temperature strongly affected Cr(VI) concentrations, with hotter flames containing more Cr(VI). The introduction of Fe into the incinerator flame dramatically reduced the concentration of Cr(VI) in favor of a mixed Cr/Fe phase containing mostly Cr(III) in what appears to be a Chromite like spinel structure. Micro-XAS analysis of individual Cr containing ambient aerosol particles showed that the majority of these particles, ~75%, appear to be found in the same Cr/Fe Chromite like phase identified in the laboratory. Additional forms of Cr include Cr(OH)3, Cr0, and Cr carbides. Lastly, Cr(VI) concentrations in ambient aerosols appeared to be significantly increased in samples taken down wind from a Cr plating facility. This seems to indicated that Cr(VI) from anthropogenic sources can be transported relatively long distances, > 5miles, or much longer than currently believed.

A197

Goldschmidt Conference Abstracts 2005 Life in Extreme Environments

A198

Microbial weathering of ocean crust 1

2

K.J. EDWARDS , C.M. SANTELLI

Oxidant production via electron bombardment: Progress in experimental simulations of the Europan surface environment

3

AND W. BACH

Geomicrobiology Group, Department of Marine Chemistry & Geochemistry, Woods Hole Oceanographic Institution, McLean Lab, MS8 ([email protected], 2 [email protected], [email protected])

KEVIN P. HAND1,3, R.C. CARLSON2 AND C.F. CHYBA3,1 1

The oxidative weathering the rock type that represents the largest fraction of Earth’s crust, oceanic basalt, involves a large change in free energy. If harnessed by microbes, significant biomass production may be supported. However, basalt rock habitats, remain poorly studied with respect to the microbial communities that may be supported by oxidative weathering. We are examining the early stages of ocean crust alteration using molecular microbiological, theoretical energetic, and mineralogical approaches. Energetic calculations suggest that most biological activity in the ocean crust should occur in close proximity to spreading centers on ridge flanks where low-temperature fluid circulation is substantial. In these environments, abundant reduced basalt is in contact with oxidized fluids. Combined, aerobic and anaerobic reactions involving Fe, S, and H2 may result in biomass production of up to one million tons of C per yr rocks aged less than 10 M.a. (Bach and Edwards, 2003). Biological activity likely declines significantly off-axis in older, highly altered crust (>10 Ma) due to the depletion of either suitable oxidants or reductants. 16S rRNA and electron microscopy studies have been conducted on the bacterial fraction of the microbial communities colonizing very young (<1 M.a.) basalt samples that displayed varying degrees of weathering. Diverse, yet distinct populations of bacteria were detected on the different samples. A general trend moving from more metal and Soxidizing autotrophic communities towards more metal- and S-reducing populations as the samples accumulated more weathering products is suggested. Our studies indicate that a succession of microbial populations on basalt occurs during weathering and that metal, S, and C cycling are prominent characteristics of basalt-hosted cryptoendolithic microbial communities at the seafloor.

Department of Geological and Environmental Sciences, Stanford University, Stanford, CA 94305, USA 2 Jet Propulsion Laboratory, 4800 Oak Grove Drive, Pasadena, CA, 91109, USA 3 Center for the Study of Life in the Universe, SETI Institute, 515 N. Whisman Road Mountain View, CA 94043, USA At present our understanding of the Europan system indicates that both liquid water and a suite of biogenic elements should be availble in the Europan hydrosphere. Considering the habitability of the putative ocean, the major limiting resource may very well be available free energy [1]. As a result of being situated deep within the Jovian magnetosphere, the surface of Europa is exposed to high energy, charged particle irradiation. The incident particle flux of 7.9x1010 keV cm-2 s-1 is dominated (>75%) by electrons in the keV to MeV range [2]. Ground based measurements and data from the Galileo Near-Infrared Mapping Spectrometer (NIMS) have revealed a suite of radiolytically produced oxidants at the surface (including O2, CO2, SO2, H2O2,) [3-5]. In order to replicate some of the chemical processes occuring on the surface of Europa, we have constructed a vacuum chamber equiped with a high-emergy electron gun and a cryostat vapor-deposition plate (see figure). Electron irradiation of water ices has produced hydrogen peroxide and preliminary results indicate percent abundances of H2O2 greater than those estimated from surface observations of Europa. Implications for a subsurface ecosystem will be discussed.

References [1] Chyba, C. F. (2000) Nature 403, 381-382. [2] Cooper, J. F., et al.(2001). Icarus 149, 133-159. [3] Spencer, J. R. et al. (2002) Astron, J. 124, 3400-3403. [4] Carlson, R. W., et al. (1999) Science 286, 97-99. [5] Carlson, R. W. et al. (1999) Science 283, 2062-2064.

Goldschmidt Conference Abstracts 2005 Life in Extreme Environments

A199

Anaerobic oxidation of natural gas in soil – The geochemical evidence?

What’s on the menu for thermophilic heterotrophs?

SERGUEY ARKADAKSKIY1, KARLIS MUEHLENBACHS1, CARL MENDOZA1 AND BRYAN SZATKOWSKI2

K.L. ROGERS AND J.P. AMEND

1

EAS Department, University of Alberta, 1-26 ES building T6G 2E3, Edmonton, AB, Canada 2 GCHEM ltd. 101 Airport Road, Lloydminster, AB, Canada Natural gas (>99 vol.% CH4) leaks along the well bores of ca. ½ of the oil wells in W. Canada. Leaking gas reaches the unsaturated zone where it is partially to completely consumed by aerobic bacteria. Long term monitoring of soil gas contents near two wells drilled in 1997 demonstrates that aerobic oxidation produces up to 20 vol. % CO2 and reduces CH4 levels near the soil surface to 0.01 vol. %. The O2 content below 100 cm falls to 0.1 vol. %. Small amounts of authigenic calcite (0.02 to 8.0 g/kg) of bacteriogenic and abiotic origin have formed in the soil near the wells. Low δ13Ccalcite (-30 to -57 ‰, PDB) indicates that CH4 is the predominant carbon source. Low δ18Ocalcite (-17 to -22 ‰, PDB) suggests isotopic disequilibrium consistent with estimated rapid growth rates and with the involvement of bacteria with calcite precipitation. The discovery of pyrite framboids and authigenic calcite in soil close to the well bore indicates that the two minerals formed in an anaerobic environment likely as a result of anaerobic oxidation of methane (AOM). The latter is corroborated by the higher authigenic calcite content and higher then background pH (by 0.5-2.9 units) of soil samples close to well bores. Elevated pH is likely related to HCO3- production during AOM. In contrast, soil samples collected further away from the well bores contain little or no authigenic calcite and have lower than background pH likely related to the high soil CO2 concentrations produced during the aerobic oxidation of CH4. The small amounts of authigenic calcite suggest that AOM may be confined to rare isolated domains in the soil where anaerobic conditions are maintained by flushing of natural gas and by O2 consumption associated with aerobic oxidation of CH4. In vitro assays with soil samples collected near the wells have not shown CH4 based sulphide production so far. The limited sulphur availability in the soils and the low free energy yield of SO42- reduction driven OM render other species (e.g.; Fe3+) more likely to be used as preferred electron acceptors. Important advantages of Fe3+ are its greater availability in soil and that it may readily be re-supplied through regeneration especially if Fe2+ resides in soil as carbonate or green rust.

Department of Earth & Planetary Sciences, Washington University, St. Louis, MO 63130 USA ([email protected], [email protected]) Chemotrophy is the energetic foundation of microbial communities in most hydrothermal systems. However, the diversity of energy sources in these hydrothermal systems is only poorly understood because the requisite analyses of thermal fluids and minerals are rare. In particular, there is a paucity of measurements of organic compounds that are needed to determine the energetics of heterotrophic metabolisms. Concentrations of aqueous amino acids (0-28 µM) [1], sugars (0-23 µM), and organic acids (0-32 µM) [2] were recently determined in vent fluids at Vulcano Island, Italy. Together with earlier analyses of inorganic solutes, gases and minerals, and standard state properties, in situ values of ∆Gr were calculated for 115 organic oxidation reactions at 6-8 sites at Vulcano. The terminal electron acceptors considered in these computations were O2, SO4-2, Sº, NO3-, and Fe(III) in magnetite. All reactions considered were exergonic (energyyielding) at all of the sites investigated, with the highest energy yield from the aerobic oxidation of the amino acid arginine (-162 kJ/mol e- transferred). In general, oxidation with O2 yielded the most energy (98-162 kJ/mol e-), followed by NO3- (61-118 kJ/mol e-) and Fe3O4 (26-125 kJ/mol e-). Oxidation with SO4-2 or Sº yielded only (6-46 or 11-59 kJ/mol e-, respectively), despite the ubiquity of sulfur metabolisms in thermophilic communities. Reactions with each terminal electron acceptor yielded similar energies (normalized per electron transferred) for the oxidation of amino acids, sugars and organic acids, as well as for the oxidation of CH4 and CO. Differences in reaction energetics across the sites depended primarily on variations in geochemical compositions and not temperature.

References [1] Svensson, E., Skoog A. and Amend J. P., (2004). Org. Geochem. 35, 1001-1014. [2] Amend J. P., Rogers K. L., Meyer-Dombard D. R., (2004). GSA Special Paper 379, 17-34.

Goldschmidt Conference Abstracts 2005 Life in Extreme Environments

A200

Thermodynamic analysis of microbial metabolism in hydrothermal systems D.E. LAROWE

AND H.C. HELGESON

Department of Earth and Planetary Science, University of California, Berkeley, USA ([email protected]; [email protected]) By catalyzing a wide variety of oxidation-reduction reactions, subsurface communities of microorganisms have a significant impact on their geologic environments. To better understand this impact, we have developed a quantitative thermodynamic model describing microbial coupling of irreversible redox reactions to otherwise endergonic biochemical reactions responsible for synthesizing biomass at elevated temperatures and pressures. Because microbes consist of approximately 80% protein and nucleic acids by dry weight, we have focused on quantifying the conversion of environmentally available energy into these biomacromolecules from their constituent monomers, amino acids and nucleotides, respectively. The flow of electrons in a microbe resulting from any type of catabolic strategy is, through a complicated set of reactions, coupled to the synthesis of ATP, the universal energy currency of all cells. Organisms in turn couple the Gibbs free energyliberating hydrolysis of ATP and other nucleotide triphosphates to the energy-demanding dehydration reactions required to polymerize proteins and nucleic acids. Hence, quantitative description of the synthesis of ATP by way of electron flow and the biosynthetic demand-driven hydrolysis of ATP is crucial to understanding the energetics of microbial growth. The microbial coupling model developed in the present study can be used to quantify the polymerization of proteins and nucleic acids in terms of the flow of electrons from any donor to any acceptor and the subsequent synthesis of ATP from ADP. We present the results of calculations that can be used to determine the production of microbial biomass generated in terms of nucleic acids and proteins in a variety of environments characterized by differing temperatures, pressures, and bulk compositions.

Energetic constraints on subsurface biomass production within the igneous ocean crust THOMAS M. MCCOLLOM LASP, University of Colorado, Boulder, CO 80309, USA ([email protected]) Deep-sea hydrothermal vents have long been recognized to host prolific biologic communities that rely on chemical energy as their source of primary biomass production. More recently, it is becoming increasingly apparent that the hightemperature vents may just be the tip of the biological iceberg, and that there may be substantial biomass living below the subsurface along mid-ocean ridges and within the basaltic crust. Owing to the scarcity of organic material in seawater and basaltic rocks, any sustantial biologic communities living within the subsurface must also rely primarily on chemical sources of energy. Subseafloor environments are just in the early days of exploration. Consequently, little is yet known about what organisms live there, how many there are, where they live, or what their metabolic activities are. However, because these communities are likely to be largely dependent on chemical sources of energy, we can gain some insight into the likely abundance, spatial distribution, and metabolic activities of subsurface microbial populations by examining the sources of chemical energy that are available to support the community. I will present results of an energetic evaluation of several different seafloor habitats that owe their existence to ongoing hydrothermal and geologic processes along mid-ocean ridges, including diffuse subsurface mixing zones, weathering of rocks on the seafloor, downflow zones, and low-temperature alteration of basalt. Diffuse mixing zones, where hightemperature hydrothermal fluids mix with seawater in the subsurface on the flanks of hydrothermal vents, are likley to be the most productive subsurface habitats, with sufficient energy to support biological production much larger than occurs at the well-studied hydrothermal chimneys. Oxidation of rocks at the seafloor and in downflow zones also has the potential to support substantial populations of microbes. In contrast, there is relatively little energy available to hydrogenbased methanogenic communities like those that have been proposed to exist in terrestrial basalt aquifers, suggesting this type of environment makes relatively little contribution to total subsurface microbial carbon fixation. Altogether, there is sufficient energy in seafloor environments to support about 1012 g (dry wt) of primary biomass production per year.