New measurements of activation volume in olivine under anhydrous conditions

W. B. Durham1, S. Mei2, D. L. Kohlstedt2, L. Wang3, N. A. Dixon1

1Massachusetts Institute of Technology, Cambridge, MA 02127

2University of Minnesota, Minneapolis, MN 55455

3Stony Brook University, Stony Brook, NY 11794

Critical parameters in geodynamics are many, but few are as elusive as the activation volume for creep of olivine. It can be expected that the properties of olivine control much of Earth’s character because it is the major phase of the upper mantle, which transcends a pressure range from approximately 0.3 to 15 GPa to a depth of 400 km. In particular, the rheological properties of olivine are key to the thermal structure and dynamic behavior of the upper mantle. Measurement of the pressure dependence of the viscosity of olivine has been a goal of experimentalists almost since the earliest test of olivine-bearing rock under pressure nearly a half-century ago (Griggs et al., 1960).

Activation volume, V*, which measures the response of high-temperature, steady-state deformation of a rock to hydrostatic pressure, P, is given by


where is creep rate, T is temperature, and R is the gas constant. The greatest barrier to a well-resolved measurement of V* has been a limit on achievable pressures in the laboratory, which, until the recent advent of a new generation of high-pressure deformation apparatus, was about 2 GPa using the solid-medium Griggs apparatus. We report here first results from the deformation-DIA, which raises this limit by over a factor of two.

The particular advance reported in this paper has come about through two recent changes to the sample cell assembly, both aimed at correcting earlier attempts to correct shortcomings in earlier D-DIA experiments: (1) a choice of cell materials that maintains anhydrous conditions around the sample, and (2) elimination of the thermocouple from the assembly. The anhydrous assembly (Figure 1) is a sphere of mullite embedded in a cube of unfired pyrophyllite, where the diameter of the sphere matches the edge length of the cube. Elimination of the thermocouple from the cell assembly results in a demonstrably more symmetric and mechanically stable environment in and around the deformation column throughout an experiment.. To determine temperature we use carefully calibrated relationships between furnace power and temperature. For measuring the calibration curves we use identical equipment and reproduce all run conditions except sample material: pressure, temperature, D-DIA ram displacement rate, and run time. The effect of time and deformation is in fact substantial.

Deformation experiments designed to measure V* in Equation (1) were carried out in a D-DIA apparatus (Wang et al., 2003) installed at beam line X17B2, National Synchrotron Light Source, Brookhaven National Lab. We carried out 3 experiments using the hybrid assembly, at a narrow range of temperatures near 1473 K, strain rates of 2 ´ 10-6 to 2 ´ 10-5 s-1, and pressures of 2.7 to 4.9 GPa. Strain rate and pressure were stepped individually in two of the experiments, giving 5 sets of (, s, P, T). The data allow us to constrain V* in Equation (1). Invoking values for stress and temperature sensitivity from the literature (Hirth and Kohlstedt, 2003), we calculate from our results V* = 9.5 ´ 10-6 m3/mol with an . The fit to the data is illustrated in Figure 7 along with the uncertainty range of ±7 ´ 10-6 m3/mol defined by the error bars.

Comparison to other measurements in the D-DIA. There have been other recent attempts to measure activation volume of olivine using synchrotron x rays and the D-DIA or similar instruments, including our own work using predecessors to the anhydrous, thermocouple-free cell used here for the first time. (Mei et al., manuscript in preparation). The experiments of Li et al. (2004) determined a very low value for V* of < 5 ´ 10-6 m3/mol using stress relaxation experiments in a hydrostatic DIA. Because those measurements were made at high stresses and low plastic strains, the results may not be directly comparable to the measurements reported here. In experiments under stresses and to plastic strains similar to those in our experiments, and using the same D-DIA, Li et al. (2006) again measured a value of V* near 0. We cannot reconcile our data with theirs, although we can point out a difference in water content. Their experiments were conducted using a pressure medium of amorphous boron powder in epoxy, in contrast with the mullite-based hybrid cell used in our runs. Their samples had a significantly higher water content (~3500 ppm) than ours (≤40 ppm). On the other hand, the results of Borch and Green (1989) suggest that water enhances the pressure sensitivity of creep, at least to P = 2 GPa.

Raterron et al. (2008) have recently found that two important slip systems in olivine have contrasting values for V*. For a-slip, that is, (010)[100], V * = 12±4 ´ 10-6 m3/mol; while for c-slip, that is (010)[001], V* = 3±4 ´ 10-6 m3/mol. These authors observe that a-slip dominates below P ≈ 8 GPa at high temperature. In our experiments, a-slip was probably quite active, and thus our value of V* = 9.5±7´10-6 m3/mol should be compared to that of V * = 12±4 ´ 10-6 m3/mol from the Raterron et al. (2008) study. Thus, the possibility that creep in the polycrystal is controlled by a-slip cannot be eliminated.


This study was supported by DOE grants DE-FG02-07ER15839 for WBD and DE-FG02-04ER15500 for DLK. The experimental work was performed at X17B2 of the National Synchrotron Light Source (NSLS), supported by the U.S. Department of Energy, Division of Materials Sciences and Division of Chemical Sciences, under Contract No. DE-AC02-98CH10886. This rheology research was partially supported by COMPRES, the Consortium for Materials Properties Research in Earth Sciences under NSF Cooperative Agreements EAR 01-35554 and 06-49658 using cell assemblies fabricated in the Multi-Anvil Cell Development project at the Arizona State University.


Borch, R.S. and Green, H.W., II, 1989. Deformation of peridotite at high pressure in a new molten cell: Comparison of traditional and homologous temperature treatments. Phys. Earth Planet. Interiors, 55: 269-276.

Griggs, D.T., Turner, F.J. and Heard, H.C., 1960. Deformation of rocks at 500˚ to 800˚C. In: D.T. Griggs and J.W. Handin (Editors), Rock Deformation, Geol. Soc. Am. Memoir 79. Geol. Soc. Am., New York, pp. 39-104.

Hirth, G. and Kohlstedt, D.L., 2003. Rheology of the Upper Mantle and the Mantle Wedge: A View From the Experimentalists. In: J. Eiler (Editor), Inside the Subduction Factory. American Geophysical Union, Washington, D.C., pp. 83-105.

Li, L., Weidner, D., Raterron, P., Chen, J.H. and Vaughan, M., 2004. Stress measurements of deforming olivine at high pressure. Physics of the Earth and Planetary Interiors, 143-44: 357-367.

Li, L. et al., 2006. Deformation of olivine at mantle pressure using the D-DIA. European Journal of Mineralogy, 18(1): 7-19.

Raterron, P., Amiguet, E., Chen, J., Li, L. and Cordier, P., 2008. Experimental deformation of olivine single crystals at mantle pressures and temperatures. Physics of the Earth and Planetary Interiors, in press.

Wang, Y., Durham, W.B., Getting, I.C. and Weidner, D.J., 2003. The deformation-DIA: A new apparatus for high-temperature triaxial deformation to pressures up to 15 GPa. Rev. Sci. Instrum., 74: 3002-3011.


Durham, W. B., S. Mei, D. L., Kohlstedt, L. Wang and N. A. Dixon, New measurements of activation volumes in olivine under anhydrous conditions, Phys. Earth Planet. Interiors, in press, 2008.


Department of Earth, Atmospheric, and Planetary Sciences, Massachusetts Institute of Technology 54-720, 77 Massachusetts Ave., Cambridge, MA 02127


Figure 1. The new hybrid cell. A 6-mm diameter mullite sphere is cradled and lightly cemented between two pieces of unfired pyrophyllite to form a solid cube of 6-mm edge length. Coring of the hole for furnace and deformation column then follows.

Figure 2. The 5 measurements from the hybrid cell reduced to a common T and as indicated using n = 3.5 and E* = 500000 J/mol (Hirth and Kohlstedt, 2003) in Equation (2). The value of V* = 9.5 ´ 10-6 m3/mol (solid line) is defined by the two points from the one run conducted at two different pressures (circled). Error bars for those two points allow values from 2.5 to 17 ´ 10-6 m3/mol (dashed lines). For purposes of plotting, uncertainty in temperature has been converted to an additional uncertainty in stress using the same values of n and E* (see text).