E. B. Watson

Partitioning of strontium between calcite and fluid

The Sr/Ca ratio of biogenic carbonate is widely used as a proxy for paleotemperature. This application is supported by empirical calibrations of Sr/Ca as a function of temperature, but it is also known that Sr uptake in calcite gauged by KdSr = is affected by other variables, including bulk precipitation rate (KdSr increases with increasing precipitation rate). There are no data from controlled experiments specifically addressing the effect of radial growth rate of individual crystals on KdSr. For this reason, we conducted two series of experiments to explore Sr partitioning at varying growth rates: (1) growth from a CaCl2–NH4Cl–SrCl2 solution by diffusion of CO2 from an ammonium carbonate source (“drift” experiments) and (2) “drip” precipitation of calcite on a substrate, using a steady flow of CaCl2–SrCl2 and Na2CO3 solutions, mixed just before passage through a tube and dripped onto a glass slide precoated with calcite (“cave-type” experiments). The growth rates of individual crystals were determined by periodic monitoring of crystal size through time or, roughly, by comparison of the final size with the duration of the experiment. Electron microprobe analyses across sectioned crystals grown in the drift experiments show that the concentration of Sr is high in the center (where radial growth rates are highest) and decreases systematically toward the edge. The center-to-edge drop in Sr concentration is a consequence of the slowing radial growth rate as individual crystals become larger. In general, high crystal growth rate (V) enhances Sr uptake in calcite due to a type of kinetic disequilibrium we refer to as “growth entrapment.” The apparent KdSr ranges from 0.12 to 0.35 as V increases from 0.01 nm/s to 1 μm/s at 25°C.

Numerical investigations of apatite ^4He/^3He thermochronometry

Apatite ^4He/^3He thermochronometry has the potential to constrain cooling histories for individual samples provided that several presently untested assumptions are valid. Here we simulate the sensitivity of ^4He/^3He spectra to assumptions regarding geometric model, crystallographic anisotropy, broken grain terminations, parent nuclide zonation, and the accuracy of results obtained from analyses of aggregates of multiple crystals. We find that ^4He/^3He spectra obtained from a cylinder with isotropic diffusion are almost indistinguishable from those obtained from an equivalent sphere with an equivalent initial ^4He distribution. Under similar conditions anisotropic diffusion from the cylinder can greatly bias ^4He/^3He spectra, but only if diffusion is >10 times faster in the axial than the radial direction. Existing data argue against anisotropy of this magnitude. We find that analysis of apatites with broken terminations will also bias ^4He/^3He spectra, but not greatly so. In contrast, we find that zonation of a factor of 3 in parent nuclide concentration produces ^4He/^3He spectra that deviate substantially from the homogeneous model. When parent nuclides are highly concentrated near the grain rim and/or cooling is fast, the resulting ^4He/^3He spectra will be readily identified as aberrant. However, more subtle zonation, higher concentrations in the grain interior, or samples that have cooled slowly regardless of zonation style can yield ^4He/^3He spectra that look acceptable but will lead to inaccurate thermochronometric interpretation if parent homogeneity is assumed. Finally, we find that analysis of an aggregate of crystals with identical ^4He distributions can yield ^4He/^3He spectra (and diffusion Arrhenius arrays) that are very different from those that would be obtained on the individual crystals if even small variations in He diffusion exist among the grains. Overall, our observations suggest that modeling tools that assume spherical geometry and isotropic diffusion are appropriate for interpreting apatite ^4He/^3He spectra. However, it is essential to analyze only individual crystals and to assess the degree of parent nuclide zonation in those crystals.

In situ δ7Li, Li/Ca, and Mg/Ca analyses of synthetic aragonites

In situ secondary ion mass spectrometry (SIMS) analyses of δ7Li, Li/Ca, and Mg/Ca were performed on five synthetic aragonite samples precipitated from seawater at 25°C at different rates. The compositions of δ7Li in bulk aragonites and experimental fluids were measured by multicollector inductively coupled plasma–mass spectrometry (MC-ICP-MS). Both techniques yielded similar δ7Li in aragonite when SIMS analyses were corrected to calcium carbonate reference materials. Fractionation factors α7Li/6Li range from 0.9895 to 0.9923, which translates to a fractionation between aragonite and fluid from −10.5‰ to −7.7‰. The within-sample δ7Li range determined by SIMS is up to 27‰, exceeding the difference between bulk δ7Li analyses of different aragonite precipitates. Moreover, the centers of aragonite hemispherical bundles (spherulites) are enriched in Li/Ca and Mg/Ca relative to spherulite fibers by up to factors of 2 and 8, respectively. The Li/Ca and Mg/Ca ratios of spherulite fibers increase with aragonite precipitation rate. These results suggest that precipitation rate is a potentially important consideration when using Li isotopes and elemental ratios in natural carbonates as a proxy for seawater composition and temperature.

Solubility of fluorite in haplogranitic melt of variable alkalis and alumina content at 800°–1000°C and 100 MPa

[1]xa0To study the behavior of fluorite in SiO2-Al2O3-Na2O-K2O-H2O melts, we conducted both crystallization and dissolution experiments. In the first case, fluorite crystallized from melts enriched in Ca relative to F. In the second case, ground fluorite (<10 μm) was partially dissolved into melts of initially F-free or F-doped glass. Experiments were conducted at 800°–1000°C and 100 MPa for 7 to 240 hours using a series of finely ground subaluminous and peralkaline haplogranitic glasses. Electron microprobe (EMP) analysis of the run products (glass + crystals) documented increasing fluorite solubility with melt alkalinity. These molar relationships can be described by CF/2 = −1.3 · ASI + 2.3 at 900°C and CF/2 = CCa = −2.1 · ASI + 3.8 at 1000°C, where ASI is the aluminum saturation index measured in experimental glasses as Al/(Na + K + 2Ca), CF/2 is one half the molar concentration (%) of F, and CCa is the mol% of Ca in the glass. Fluorite solubility may be defined as either CF/2, CCa, or both. Solubilities at 800°, 900°, and 1000°C for the melts with ASI between 0.58 and 0.59 approach log linear with 104/T(K) (slope = −0.45). The joint effect of F and temperature on fluorite solubility is significant: solubility rises from 0.28 mol% at 800°C (4.96 mol% doped F) to 0.99 mol% at 900°C (2.17 mol% doped F). Solubilities at 800°, 850°, and 900°C for the melts with the same initial F contents (4.96 mol%) approach log linear with 104/T(K) (slope = −0.33).

Diffusion of Ca and F in haplogranitic melt from dissolving fluorite crystals at 900°–1000°C and 100 MPa

[1]xa0The diffusion rates of Ca and F into haplogranitic melt were determined by concentration profiles in glasses produced by dissolving fluorite crystals into haplogranitic melt at 900°–1000°C and 75–100 MPa. Starting products were synthesized from oxides of haplogranitic composition at two aluminosities ([Na2O + K2O]/Al2O3 = 1.05 and 1.68). Fluorite crystals or crystal powders were coupled with powders or solid cylinders of haplogranitic glass. Water was added to the runs, resulting in two runs at H2O = 1.20 and 1.40 wt% and three runs thought to be water saturated. The materials were then subjected to the conditions of interest in either a Stellite cold seal apparatus or an internally heated pressure vessel. Electron microprobe traverses of the quenched glasses produced in these experiments revealed negative concentration gradients in Ca and F away from the fluorite-glass boundary. These gradients are comparable to an error function solution (in the first approximation) for specific Ca and F diffusivities (DCa and DF, respectively). The Ca diffusion coefficient increased from 3.36 ± 0.19 · 10−9 to 2.16 ± 0.28 · 10−8 cm2/s and F diffusivity increased from 1.16 ± 0.09 · 10−8 to 6.59 ± 0.62 · 10−8 cm2/s as temperature increased from 900° to 1000°C. These diffusivities are comparable to other published melt diffusion data for the dissolution of multicomponent minerals where the diffusants move independently. Calcium and fluorine diffusivities increase from 2.77 ± 0.09 · 10−9 to 2.16 ± 0.28 · 10−8 cm2/s and from 2.27 ± 0.10 · 10−8 to 6.59 ± 0.62 · 10−8 cm2/s, respectively, with increasing water content at 1000°C, likely due to melt depolimerization. The large difference in the concentrations of Ca (C0Ca) and F (C0F) at the fluorite-glass interface suggests that these two elements diffuse independently in the melt. In contrast, the concentration gradient for Na parallels that of F, suggesting complexing between these two elements.