Elsevier

Chemical Geology

Volume 242, Issues 3–4, 15 August 2007, Pages 470-483
Chemical Geology

Ti diffusion in zircon

https://doi.org/10.1016/j.chemgeo.2007.05.005Get rights and content

Abstract

Chemical diffusion of Ti under anhydrous conditions at 1 atm and under fluid-present elevated pressure (1.1–1.2 GPa) conditions has been measured in natural zircon. The source of diffusant for 1-atm experiments was a ZrO2–TiO2–zircon mixture, with experiments run in crimped Pt capsules. Diffusion experiments done in the presence of H2O–CO2 fluid were run in a piston-cylinder apparatus, using a source of ground TiO2, ZrSiO4 and SiO2, with oxalic acid added to produce H2O–CO2 vapor and partially melt the solid source material, yielding an assemblage of rutile + zircon + melt + vapor. Nuclear reaction analysis (NRA) with the resonant nuclear reaction 48Ti(p,γ)49V was used to measure diffusion profiles for both sets of experiments. The following Arrhenius parameters were obtained for Ti diffusion normal to cover the temperature range 1350–1550 °C at 1 atm:DTi=3.33×102exp(754±56kJmol1/RT)m2s1.

Ti diffusivities were found to be similar for experiments run under fluid-present conditions. A fit to all of the data yields the Arrhenius relation D = 1.34 × 102 exp(− 741 ± 46 kJ mol− 1/RT) m2 s− 1.

These data suggest that zircon should be extremely retentive of Ti chemical signatures, indicating that the recently-developed Ti-in-zircon crystallization geothermometer [Watson, E.B., Harrison, T.M., 2005. Zircon thermometer reveals minimum melting conditions on earliest Earth. Science 308, 841–844] will be quite robust in preserving temperatures of zircon crystallization. Titanium diffuses somewhat faster in zircon than larger tetravalent cations U, Th, and Hf, but considerably more slowly than Pb, the REE, and oxygen; hence Ti crystallization temperatures may be retained under circumstances when radiometric ages or other types of geochemical information are lost.

Introduction

Zircon is extraordinarily valuable in interpreting crustal histories, given its tendency to incorporate trace elements useful as geochemical tracers, such as the REE, Y, Ti and Hf, and its longstanding and extensive use in U–Th–Pb isotopic dating. The relative insolubility of zircon in crustal melts and fluids, as well as its general resistance to chemical and physical breakdown, often result in the existence of several generations of geochemical information within a single zircon grain. These internal isotopic and chemical variations, which are often on extremely fine scale, can reveal much about thermal histories and past geochemical environments. Especially notable are recent geochemical and geochronological studies of Hadean zircons, which provide insight into thermal and chemical conditions of the early Earth. The recently-developed crystallization thermometer based on the Ti content of zircon (Watson and Harrison, 2005, Watson et al., 2006, Ferry and Watson, in press), which has broad potential for application in crustal igneous and metamorphic systems, can provide a valuable complement to information gathered from other elemental and isotopic analyses and shed further light on early earth history.

The robustness and utility of this geothermometer will depend to considerable extent on the mobility of Ti in zircon. Many cations, including the REE (Cherniak et al., 1997b), U, Th, and Hf (Cherniak et al., 1997a) and Pb (Cherniak and Watson, 2001) have been found to diffuse quite sluggishly in zircon. Trends for diffusivities based on cation size and charge, derived from extant diffusion data, suggest that Ti will diffuse slowly as well, but no direct measurements have heretofore been made.

In this work, we measure the diffusivity of Ti in natural zircon. Experiments are run under both 1-atm and fluid-present conditions to determine whether the presence of fluids will affect cation diffusivities, as has been found to be the case for oxygen in zircon (Watson and Cherniak, 1997). The data also complement earlier diffusion measurements of other elements in zircon, and permit comparison of diffusivities of these species. Such comparisons are critical in assessing the possibility of diffusional alteration, and determining whether various types of geochemical and isotopic information are simultaneously preserved in zircon.

Section snippets

Experimental procedure

The mineral specimens used in this study were from an optically clear, gem-quality natural megacrystic zircon from a carbonatite in Australia. The zircon is clear and nearly colorless, and specimens selected for use in experiments were free of inclusions, cracks and other optically observable defects. Line scans and spot analyses of several pieces of this zircon by electron microprobe revealed concentrations of U, Th, Y and the HREE below detection limits, little evidence of zoning, and Hf

Nuclear reaction analysis

Because of the high background due to constituent Zr in RBS spectra of zircon, RBS cannot easily be used to detect low concentrations of Ti. Therefore, we instead employ another ion beam technique, nuclear reaction analysis (NRA), to measure Ti. Nuclear reaction analysis is a technique that uses energetic (typically in the range of several hundred keV to a few MeV) ion beams to induce nuclear reactions with specific nuclei. In this case, we use the reaction 48Ti(p,γ)49V, where a proton beam is

Results

The results from Ti diffusion experiments on zircon are presented in Table 1 and plotted in Fig. 3. For the diffusion experiments run at 1 atm, the Arrhenius parameters derived from a fit to the data are: activation energy 754 ± 56 kJ mol− 1 and pre-exponential factor 3.33 × 102 m2 s− 1 (log Do = 2.52 ± 1.72). Diffusivities for the experiments at elevated pressure with water present are similar, falling along the Arrhenius line defined by the 1-atm data. Hence, the presence of water appears to

Diffusion of tetravalent cations in zircon

In previous work, we have measured diffusion of the tetravalent cations U, Th, and Hf in zircon (Cherniak et al., 1997a). These data are plotted with the Ti diffusion results in Fig. 5. Ti diffusion is nearly two orders of magnitude faster than U and Th diffusion, and about an order of magnitude faster than Hf diffusion. Activation energies for Ti diffusion are similar to those for the other tetravalent cations (which range from 726–812 kJ mol− 1). This systematic variation of diffusivities

Relative retentivity of Pb, and O isotopic signatures, and REE and Ti chemical zoning

Data for diffusion of Pb, Dy and oxygen in zircon are plotted with Ti diffusivities in Fig. 6. Ti diffusion is considerably slower than the diffusion of all three of these elements. For example, at 1000 °C, Ti diffusion will be about 4 orders of magnitude slower than Pb, 3 orders of magnitude slower than Dy, and 5–8 orders of magnitude slower than oxygen (under dry and hydrothermal conditions, respectively); these variances will increase with decreasing temperature because of the comparatively

Acknowledgements

We thank Dave Wark and Dustin Trail for their assistance with the electron microprobe analyses of the zircon, and Jay Thomas for information on IR analyses of zircon. Comments from two anonymous reviewers helped improve the final version of the manuscript. This work was supported by grant EAR-0440228 from the National Science Foundation (to E.B. Watson).

References (32)

  • D.J. Cherniak

    Pb diffusion in rutile

    Contrib. Mineral. Petrol.

    (2000)
  • D.J. Cherniak

    Zr diffusion in titanite

    Contrib. Mineral. Petrol.

    (2006)
  • D.J. Cherniak et al.

    Nuclear reaction analysis

  • D.J. Cherniak et al.

    Diffusion of tetravalent cations in zircon

    Contrib. Mineral. Petrol.

    (1997)
  • Cherniak, D.J., Manchester, J., Watson, E.B., in press. Zr and Hf diffusion in rutile. Submitted to Earth and Planetary...
  • J. Crank

    The Mathematics of Diffusion

    (1975)
  • Cited by (131)

    • Diffusion anisotropy of Ti in zircon and implications for Ti-in-zircon thermometry

      2022, Earth and Planetary Science Letters
      Citation Excerpt :

      The second issue with large extrapolations of diffusion data is that it must be assumed that the Arrhenius trend maintains linearity over the entire temperature range. Although the data presented here and by Cherniak and Watson (2007) shows no resolvable deviation from linearity within the experimental temperature range, there is no guarantee that the Arrhenius trends do not have a kink or curvature at lower temperatures. For example, such a change in activation energy could occur due to a shift from an intrinsic to a pure extrinsic diffusion regime (e.g., Chakraborty, 1997; Dohmen et al., 2007; Dohmen and Chakraborty, 2007).

    • Aluminum diffusion in zircon

      2021, Chemical Geology
    View all citing articles on Scopus
    View full text