Experimental Studies of the Incorporation of Mn and Sr into Calcite During Grain Boundary Migration

Investigators

Andrew M. McCaig    School of Earth and Environment, The University of Leeds, Leeds LS2 9JT, UK
Geoffrey E. Lloyd      School of Earth and Environment, The University of Leeds, Leeds LS2 9JT, UK
Ernest H. Rutter        Rock Deformation Laboratory, Department of Earth Sciences, University of Manchester,
                                        Manchester, M13 9PL, UK, (ERUTTER@fs1.ge.man.ac.uk)
Jane Cooper               N.E.R.C. Research Student, School of Earth and Environment, The University of Leeds.

Grain boundary processes are of fundamental importance to the behaviour of rocks in both the crust and mantle. Considerable efforts have been made to measure grain boundary diffusion coefficients in rock forming minerals under static conditions (eg. Farver & Yund, 1996), and the theory predicting compositional profiles resulting from combined grain boundary and volume diffusion is also well developed (Shewman, 1989). The consequences of combined grain boundary diffusion and grain boundary migration are less well known however, although the combination of these processes occurs in all solid-state reactions and must control processes such as the homogenisation of isotopes during dynamic recrystallisation. Here we present the results of preliminary experiments designed to induce grain boundary migration in calcite with take-up of Mn or Sr out of the migrating grain boundary. An unexpected result of our experiments was diffusion of Mn through grain interiors over distances more than 100 times greater than expected on the basis of published Ca self-diffusion coefficients for calcite.

The thumbnail sketches below are based on a recent AGU poster presentation (McCaig et al. 1999). Click sequentially on them for a description of this research project. Our results can be summarised as follows.

1. Mn-enriched areas are present adjacent to the MnCO₃ layer in all samples (Figs. 2a, 3a, 4), and show sharp boundaries against Mn-poor calcite. These boundaries may or may not correspond to high angle grain boundaries identified by EBSD (Figs. 2b, 3a).

2. Mn has penetrated up to 80mm from the MnCO₃ layer in all samples, but at low concentrations (<1%), with similar, approximately linear, profiles in log concentration vs. distance space in both Mn-rich and Mn-poor areas (Figs. 3b, 5). Some of these profiles are complex, with steps in composition separated by linear segments.

3. The slope and penetration distance of the low concentration Mn profiles are not much affected by either temperature or deformation state (Fig. 5).

4. Sr shows much smaller penetration distances than Mn and steeper gradients across recrystallised grains under the same conditions.

5. Mn-rich recrystallised grains are coarser than new grains in grain boundary regions of the calcite aggregate away from the influence of the MnCO₃ layer (Fig. 2b).

6. Decarbonation of MnCO₃ occurred in all our experiments and at least 3 phases are present in the Mn-rich layer at 900°C. Further experiments will aim to control this process, and particularly the oxidation state of Mn, more carefully.

1. Recrystallisation and grain boundary migration has clearly been enhanced close to the Mn and Sr-rich layers in all the specimens (Fig. 2). We infer that diffusion-induced grain boundary migration (DIGM) has occurred (Hay & Evans, 1987), with rapid diffusion of Mn and Sr along migrating grain boundaries allowing high concentrations of these tracers in the growing grains. Where the boundaries of Mn-enriched zones do not correspond to grain boundaries (Fig. 2, 3), we infer that the direction of boundary migration has reversed (cf. Hay & Evans, 1987).

2. The characteristic diffusion distance for Ca under the experimental conditions at 900°C is 0.15mm (Farver & Yund, 1996). Clearly, Mn in particular has diffused much further through unrecrystallised calcite grains, including the single crystal, but the profiles do not fit standard error function solutions (Fig. 3b) for interdiffusion between two infinite half spaces (Crank, 1975).

3. The long low concentration "tails" in the profiles are similar to those produced by "shortcut" diffusion through grain or subgrain boundaries, or dislocation pipes (Shewman, 1989). However, the profiles show only a small sensitivity to deformation state, with the penetration distance being essentially the same in the single crystal as in the highly deformed calcite aggregates.

4. A possible solution is that Mn-Ca interdiffusion in calcite was much faster than in the Mn-rich layer. In this case a sharp step in composition occurs at the interface (Crank, 1975 and Eqn. 1). We have not attempted a precise fit of our data to this equation because of the uncertainty caused by the recrystallised zone at the interface. However, a high value of (Dt)0.5 of 13-20mm is required for Mn-Ca interdiffusion in calcite, giving D = 9 x 10-15 to 2 x 10-14 m²/s at 900°C. Corresponding interdiffusion coefficients in the Mn-rich layer would be 1.4 x 104 lower. Mn²⁺ has a small ionic radius compared to Ca²⁺, and a diffusion mechanism involving creation of Frenkel defects seems likely.

5. The lack of an obvious temperature effect on the profiles (Fig. 5) suggests a very low activation energy for Mn-Ca interdiffusion. However, our experiments were not conducted with the intention of extracting such information, and it is not clear how to use Eqn. 1 when marginal recrystallisation is extensive and additional shortcut pathways such as subgrain boundaries may have introduced steps into the profiles, as in the calcite aggregate experiments. Further work is planned to investigate the effects of temperature and time on the simpler single crystal profiles.

6. Temperature does appear to affect the DIGM process, with larger recrystallised grains, lower peak Mn concentrations, and shallower gradients across the Mn-rich regions at higher temperatures (Fig. 5). Mishin & Razumovski (1992) showed that for the simple grain boundary geometry of Fig. 8, a steady state is reached where the gradient across the newly formed grain is given by Eqn. 2, provided the characteristic distances for grain boundary migration and diffusion are large compared to that for volume diffusion. Thus the steeper gradients at lower temperatures can be interpreted in terms of lower values of DGB (Eqn. 2). However, because the geometry of grain growth is more complex than Fig. 8, and as the characteristic distance for Mn-Ca interdiffusion inferred from Fig. 7 is similar to the grain boundary migration distance, we have not attempted to extract quantitative data from this relationship.

1. Rapid Mn-Ca interdiffusion in calcite crystals occurred in our experiments, at rates apparently faster than hydrothermal diffusion of oxygen at 700 - 900°C, and with a low (but undefined) activation energy. Concentration gradients and Mn penetration distances were not very sensitive to deformation state, suggesting the enhanced diffusion is an intrinsic property of the Mn-calcite solid solution.

2. Rapid diffusion of Mn probably occurs by an interstitial mechanism in view of the small ionic radius of Mn²⁺ (and especially higher Mn valence states) relative to Ca. Further work is planned to determine the oxidation state of Mn in our experiments.

3. Diffusion-induced grain boundary migration (DIGM) also occurred in our experiments. Efficient takeup of Mn out of migrating grain boundaries suggests that Mn-Ca grain boundary interdiffusion is faster than volume interdiffusion, and may be an extremely rapid process in natural rocks.

4. Systematic changes in concentration gradients as a function of temperature can be interpreted in terms of a balance between boundary migration velocity and boundary diffusion rate.

Crank, 1975, The mathematics of diffusion
Farver , JR & Yund RA, 1996. Volume and grain-boundary diffusion of calcium in natural and hot-pressed calcite
            aggregates. Contrib. Mineral. Petrol. 123: 77-91
Hay RS & Evans, B, 1987. Chemically-induced grain boundary migration in calcite: temperature dependence,
            phenomenology, and possible applications to geologic systems: Contributions to Mineralogy and Petrology v. 97, p.
            127-141
Mishin Y.U. and Razumovskii, I.M., 1992, A model for diffusion along a moving grain boundary: Acta Metallurgica
            Materialia, v. 40, p. 839-845
Shewman, P., 1989, Diffusion in solids: Minerals and Metals Society, 246 pp.

Figure 1	Figure 2
Figure 3	Figure 4
Figure 5	Figure 6
Figure 7	Figure 8

Microgeodynamics laboratory

Research page