Seismic anisotropy in the mantle transition zone induced by shear deformation of wadsleyite

Physics of the Earth and Planetary Interiors 216 (2013) 91–98

Contents lists available at SciVerse ScienceDirect

Physics of the Earth and Planetary Interiors journal homepage: www.elsevier.com/locate/pepi

Seismic anisotropy in the mantle transition zone induced by shear deformation of wadsleyite Takaaki Kawazoe a,⇑,1, Tomohiro Ohuchi a, Yu Nishihara a,b, Norimasa Nishiyama a,2, Kiyoshi Fujino a, Tetsuo Irifune a a b

Geodynamics Research Center, Ehime University, 2-5 Bunkyo-cho, Matsuyama, Ehime 790-8577, Japan Senior Research Fellow Center, Ehime University, 3 Bunkyo-cho, Matsuyama, Ehime 790-8577, Japan

a r t i c l e

i n f o

Article history: Received 4 May 2012 Received in revised form 15 December 2012 Accepted 26 December 2012 Available online 9 January 2013 Edited by Kei Hirose Keywords: Wadsleyite Crystallographic preferred orientation Seismic anisotropy Mantle transition zone Deformation

a b s t r a c t Mantle flow in the Earth’s mantle transition zone (between 410 and 660 km depth) plays a key role to understand the nature of mantle convection, which can be mapped by observed seismic anisotropy combined with crystallographic preferred orientations of mantle minerals. Although wadsleyite is the most important mineral to cause seismic anisotropy observed in the mantle transition zone, there have been limited experimental data on its crystallographic preferred orientation because of experimental limitations. We experimentally evaluated the preferred orientation of wadsleyite developed by shear deformation at pressure–temperature conditions of the mantle transition zone (17.6 GPa and 1800–1900 K) using a deformation-DIA apparatus. The deformation experiments reveal that the [0 0 1] axis and the (0 1 0) plane of wadsleyite tend to be subparallel to the shear direction and the shear plane during deformation, respectively. These results demonstrate that polarization seismic anisotropy (velocity contrast between horizontally-polarized and vertically-polarized S-waves, VSH/VSV) observed in the mantle transition zone might be attributed to the preferred orientation of wadsleyite caused by horizontal mantle flow. Ó 2013 Elsevier B.V. All rights reserved.

1. Introduction The nature of anisotropic structures in the mantle transition zone (MTZ: between 410 and 660 km depth) can be inferred from seismological observations, and carries important information on dynamics of the mantle. Geometry of mantle flow is determined based on interpretation of observed seismic anisotropy by using a relationship between crystallographic preferred orientation (CPO) of mantle minerals and flow geometry (Karato, 2008; Mainprice, 2007). The seismic anisotropy in the upper mantle is interpreted in terms of CPOs of olivine (Jung and Karato, 2001; Katayama et al., 2004; Ohuchi et al., 2011) and serpentine (Katayama et al., 2009), which indicates that style of upper mantle convection is dominated by horizontal flows and flow directions are nearly parallel to plate motions in subduction zones (Katayama et al., 2009; Nakajima et al., 2006) and on a global scale (Becker et al., 2003). The seismic anisotropies in the MTZ are classified as polarization anisotropy (velocity contrast between horizontallypolarized S-wave (SH) and vertically-polarized S-wave (SV) ⇑ Corresponding author. Tel.: +81 89 927 8151; fax: +81 89 927 8167. E-mail address: [email protected] (T. Kawazoe). Present address: Bayerisches Geoinstitut, University of Bayreuth, Bayreuth D-95440, Germany. 2 Present address: Deutsches Elektronen-Synchrotron, 85 Notkestrasse, Hamburg D-22607, Germany. 1

0031-9201/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.pepi.2012.12.005

(Montagner and Kennett, 1996; Visser et al., 2008)) and azimuthal anisotropy measured from surface wave dispersion (Trampert and van Heijst, 2002) and splitting of shear-wave from deep earthquakes (Foley and Long, 2011). However, an interpretation of the MTZ seismic anisotropy as the flow geometry has not been fully understood because of little information on the CPO of minerals in the MTZ. The CPO of wadsleyite is a key factor to interpret the MTZ anisotropy because only wadsleyite can produce detectable seismic anisotropy among major minerals in the MTZ (Karato, 2008; Mainprice, 2007). Presence of ringwoodite and majoritic garnet, which are the other dominant minerals in the MTZ, has only marginal effect on the MTZ seismic anisotropy because these minerals show elastic anisotropies which are much weaker than that of wadsleyite. In order to study the wadsleyite CPO, a shear deformation experiment on wadsleyite in simple shear geometry is needed because it provides a direct constraint on a relationship between the wadsleyite CPO and the flow geometry, and is accordingly critical in understanding dynamics of the MTZ. Shear deformation experiments on wadsleyite were performed at pressure–temperature conditions of 15 GPa and 1600–1700 K using a rotational Drickamer apparatus (RDA) (Xu et al., 2005), however, the CPO patterns of wadsleyite did not clearly develop in that study. The wadsleyite CPO was also studied by stress relaxation tests using a Kawai-type apparatus and two types of the CPO

92

T. Kawazoe et al. / Physics of the Earth and Planetary Interiors 216 (2013) 91–98

patterns were found (Demouchy et al., 2011), however, it is difficult to apply the CPO to mantle dynamics because of large uncertainties in stress and strain rate in the experiments. A variety of slip systems were identified in wadsleyite crystals deformed in a Kawai-type apparatus by transmission electron microscopy (TEM) (Sharp et al., 1994; Thurel and Cordier, 2003; Thurel et al., 2003), and were applied to interpretation of the MTZ seismic anisotropy through a deformation simulation of a wadsleyite aggregate (Tommasi et al., 2004). However, the simulation result was not definitive because relative activities of the wadsleyite slip systems have been uncertain, which are key factors for the CPO development. Recently, we have expanded pressure–temperature conditions of the uniaxial deformation experiments in a deformation-DIA apparatus (Wang et al., 2003) from those of the upper mantle to the MTZ conditions by optimizing experimental techniques for high-pressure generation (Kawazoe et al., 2010b). In the present study, we extended the available pressure condition in the shear deformation experiments with the deformation-DIA apparatus to 17.6 GPa, which corresponds to 530 km depth. Here we show the experimental results on the wadsleyite CPO developed at controlled strain rate at the pressure–temperature conditions of the MTZ.

2. Methods 2.1. Experimental methods Shear deformation experiments on wadsleyite were conducted in the simple shear geometry using the deformation-DIA apparatus (Wang et al., 2003), MADONNA-1500, combined with newly-designed multi-anvil 6–6 system (Kawazoe et al., 2010b). We adopted second-stage anvils made of tungsten carbide with ultrafine grains (Fujilloy TF05, Fuji Die Co. Ltd.) with a truncated edge length of 3.0 mm, and no preformed gasket was used in the present study. The cell assembly used in the present study was similar to that developed in our early study (Kawazoe et al., 2010b) except for the shear deformation mechanism (Fig. 1). Starting material was a thin slice (100 lm thick) of a single crystal of San Carlos olivine, and the [0 0 1] and [0 1 0] axes of the single crystal were oriented parallel to the shear direction and normal to the shear plane in the simple shear geometry, respectively. The starting material was sandwiched between tungsten pistons that cut at 45° from compression axis, and the pistons were placed with platinum foils in a central part of the cell assembly. Generated temperature was measured with a W97Re3–W75Re25 thermocouple, whose hot junction was placed near the end of one of the tungsten pistons, and uncertainty in temperature was determined as ±5 K based on its

Fig. 1. A cross section of the cell assembly used for the shear deformation experiments using the deformation-DIA apparatus.

fluctuation during deformation. Generated sample pressure was calibrated against an applied press load by a quench method using a phase transition between wadsleyite and ringwoodite in (Mg,Fe)2SiO4 at 1700–1800 K (Gasparik, 2003), and uncertainty in pressure was estimated as ±0.4 GPa. The sample was first compressed to a target press load (1.00 MN) at room temperature, and temperature was increased to 1700– 1900 K at the press load. In the deformation experiments, temperature was kept at 1800–1900 K for 20 min, and then the sample was deformed at 17.6 GPa and 1800–1900 K by advancing upper and lower anvils controlling loads of deformation-rams manually. Additional runs with no deformation process were conducted at 17.6 GPa and 1700–1800 K in order to observe the CPO pattern and dislocation microstructure of the wadsleyite sample prior to deformation and determine the experimental pressure by phase observation. In the experiments with no deformation process (runs M0187 and M0169), the samples were quenched after keeping temperatures of 1800 and 1700 K for 1 and 20 min, respectively. 2.2. Analytical methods Shear strain c was measured from rotation of an Mo strain marker in the sample, which was initially oriented normal to the shear plane, and calculated using the following equation: c ¼ tan h, where h is the rotation angle of the strain marker. Uncertainty in strain was evaluated as ±0.1 by shape of the strain marker. Shear strain rate c_ was calculated from the determined strain and duration of the deformation when the upper and lower anvils had been advanced. The calculated strain rate was an average during the deformation because the shear strain was determined from rotation of the strain marker in the recovered sample. Crystallographic orientation of a wadsleyite grain was measured by the electron backscatter diffraction technique (EBSD). An electron backscatter pattern was taken and indexed with Channel 5 software from HKL technology at an accelerating voltage of 15 kV and a probe current of 1.0 nA in a field emission scanning electron microscope (FE-SEM, JEOL JSM-7000F). In order to obtain an accurate solution, the crystallographic orientation of the wadsleyite grain was determined in an operator-controlled indexing mode. Half-widths of 30° and 20° were used to draw pole and inverse pole figures, respectively. Grain size was measured with a secondary electron image taken with the FE-SEM after etching a polished surface of recovered samples with 35% HNO3. Dislocation microstructure of wadsleyite in selected samples (runs M0162, M0180 and M0187) was observed with a transmission electron microscope (TEM, JEOL-2010). TEM foils were prepared with Ar ion beam using an ion slicer (JEOL EM-09100IS). Samples were first ion-milled at an accelerating voltage of 6 kV and then finally thinned at 2–4 kV. TEM observation was conducted at 200 kV using a two-axis folder. Water content in the samples was measured by Fourier-transform infrared (FT-IR) spectroscopy based on the Paterson’s calibration (Paterson, 1982). FT-IR spectra were taken with unpolarized light using an aperture of 50  50 lm after the sample was kept at 383 K in a vacuum oven for more than 12 h. Phases of the samples were identified by micro-Raman spectroscopy using an Ar ion laser. The phase of the selected samples (runs M0162, M0180 and M0187) was also identified by selected area electron diffraction (SAED) using TEM. An elastic constant tensor of the deformed sample and a threedimensional relationship between elastic wave speeds and the deformation geometry were calculated using the software written by Mainprice (1990) and the elastic constant tensor of wadsleyite single-crystal (Zha et al., 1997). In the calculation, elastic wave speeds were evaluated at room temperature and high pressure because temperature effect on the elastic constant tensor of single-crystal wadsleyite was not available. The polarization

93

T. Kawazoe et al. / Physics of the Earth and Planetary Interiors 216 (2013) 91–98

anisotropies, VSH/VSV, corresponding to horizontal shear and vertical cylindrical flows were calculated from the elastic constant tensors of the deformed wadsleyite aggregates using formulation for elastic waves propagating in anisotropic media (Montagner and Nataf, 1986). 3. Results 3.1. Experimental conditions The conditions of the experiments with deformation and no deformation are summarized with the results of sample analyses in Table 1. The pressure condition of the experiments was evaluated as 17.6 GPa based on coexistence of wadsleyite with ringwoodite in (Mg,Fe)2SiO4 composition at 1700 K (run M0169) (Gasparik, 2003) at the applied press load of 1.00 MN (Fig. 2). The starting olivine single-crystal transformed to polycrystalline wadsleyite in the experiment conducted at 17.6 GPa and 1800 K for 1 min, and only wadsleyite phase was observed in the samples deformed at 1800 and 1900 K. These results of the phase observation indicate that the sample pressure was kept at 18 GPa and did not change substantially during the deformation process. The shear deformation on the wadsleyite aggregate with shear strains c up to 0.5 was achieved at shear strain rates c_ of 3– 10  105 s1 under the pressure–temperature conditions of 17.6 GPa and 1800–1900 K (Table 1 and Fig. 3). The experiment with no deformation process was also conducted at 17.6 GPa and 1800 K to compare the CPO patterns and the dislocation microstructures between the deformed and undeformed samples. Stress magnitude of the deformation experiments in the present study was approximately estimated as shear stress of 270–820 MPa (Table 1) based on our recent results of creep strength measurement on hydrous wadsleyite in a power-law dislocation creep regime (unpublished data). The creep strength measurement on hydrous wadsleyite was performed with the deformation-DIA apparatus and synchrotron radiation following the experimental procedures of our early study (Kawazoe et al., 2011). The following flow law of hydrous wadsleyite was used to estimate the stress magnitude;



3:0 e_ ¼ 1011:9 C 2:7 exp  OH r

 211kJ=mol RT

ð1Þ

where e_ is the strain rate in s1, COH is the water content in wt ppm H2O, r is the stress in MPa, R is the gas constant and T is the

Fig. 2. Relationship between the observed phases of the recovered samples and the pressure–temperature conditions of the experiments. Circles and squares represent the experiments with and without the deformation process, respectively, and filled and half-filled symbols indicate the wadsleyite phase and the coexistence of wadsleyite with ringwoodite, respectively. Phase boundaries in (Mg0.9,Fe0.1)2SiO4 composition (Gasparik, 2003) are also shown in this figure.

temperature in K. In the study, steady-state creep strength of wadsleyite, which contained 230–1100 wt ppm H2O, was measured as 100–540 MPa in uniaxial geometry at strain rates of 3.2– 15  105 s1 and the pressure–temperature conditions of 15 GPa and 1400–1700 K. In order to calculate shear stress with the flow law determined in the uniaxial deformation experiments, shear strain rate and uniaxial stress were converted to uniaxial strain rate and shear stress, respectively, adopting Eqs. (2) and (5) of Nishihara et al. (2008). 3.2. Microstructures Fig. 3 shows a cross section of the cell assembly recovered after deformation and microstructures of the samples synthesized and deformed at shear strain rates of 3–10  105 s1 and the pressure–temperature conditions of 17.6 GPa and 1800–1900 K. Grain boundaries were smooth in the samples synthesized at 1800 K and deformed at 1800–1900 K. The grain size of the samples synthesized and deformed at 1800 K is 3 lm and its size distribution

Table 1 Experimental conditions and results.

a b

Run No.

M0180

M0166

M0187

M0162

M0169

Pressure (GPa) Temperature (K) Observed phase

17.6 1800 Wadsleyite

17.6 1800a Wadsleyite

17.6 1800 Wadsleyite

17.6 1900 Wadsleyite

17.6 1700 Wadsleyite Ringwoodite

CPO type Shear direction Shear plane normal No. of analyzed grains Shear strain c Shear strain rate (s1) Deformation duration (min) Stress (MPa)

[0 0 1] [0 1 0] 301 0.4 3  105 270 370

[0 0 1] [hk0] 254 0.2 4  105 77 820

Random Random 410 0 – – 0

[0 0 1] [hk0] 155 0.5 1.0  104 81 270

– – – 0 – – 0

Water contentb (wt ppm H2O) (H/106 Si) Grain size (lm) Dislocation density (m2)

134 (2) 2180 (40) 2.8 (0.8) 7  1013

62 (5) 1010 (80) 2.7 (0.8) –

50 (5) 820 (80) 3.4 (1.3) <5  1011

230 (50) 3700 (800) 5–40 2  1013

– – – –

Temperature was estimated by a relationship between temperature and electric power to a heater. Numbers in parentheses are variations in water content at three to four areas.

94

T. Kawazoe et al. / Physics of the Earth and Planetary Interiors 216 (2013) 91–98

(a) M0162

(b) M0162

Sample Sample Piston

Strain marker

Piston 200 μm

(c) M0180

40 μm

(d) M0187

3 μm

3 μm

Fig. 3. Backscattered and secondary electron images of the cell assembly and the samples. The backscattered electron images of the cell assembly (a) and the sample (b) deformed at 1900 K (run M0162). The strain marker was initially oriented normal to the shear plane. The secondary electron images of the samples deformed (c) (run M180) and synthesized (d) (run M0187) at 1800 K. The samples were sectioned in parallel to the shear direction and the shear plane normal. The shear direction was horizontal as shown as a pair of arrows in this figure.

was nearly homogeneous. On the other hand, the sample deformed at 1900 K composed of grains of 5–40 lm, and the grain size at the central region was smaller than that at the outer regions near the heater. The TEM observation revealed that all the examined grains in the samples deformed at 17.6 GPa and 1800–1900 K contained dislocations, while few dislocations were found in the undeformed sample synthesized at 1800 K (Fig. 4). This fact indicates that slip systems were activated in the deformed samples because of their plastic deformation at the experimental conditions. Curved and straight dislocations were recognized in the deformed samples at 1800 and 1900 K, and subgrain boundaries were observed in the sample deformed at 1800 K. In the case of the sample deformed at 1900 K, limited numbers of grains were examined because of its large grain size and limited thin area of the TEM foil sample. Dislocation density in the sample deformed at 1800 K (7  1013 m2) was higher than that in the sample deformed at 1900 K (2  1013 m2), and average densities of dislocations in both samples were an order of 1013 m2 (Table 1). 3.3. FT-IR spectra and water content The water contents of the wadsleyite samples deformed or synthesized at 17.6 GPa and 1800–1900 K are listed in Table 1, and the FT-IR spectra obtained for the water content measurement are shown in Fig. 5. The FT-IR spectra of the sample synthesized at 1800 K (run M0187) consisted of sharp absorption peaks at 3330, 3474 and 3618 cm1. In contrast, a broad absorption peak was observed at 3150–3700 cm1 in the spectra of the deformed samples, and the spectra of the sample deformed to shear strain of 0.4 at 1800 K (M0180) showed only the broad absorption peak whose central position located at 3450 cm1. Moreover, broadening of the absorption peaks at 3330, 3470 and 3620 cm1 was observed in the spectra of the sample deformed to strain of 0.2 at

1800 K (run M0166). The broadening of the absorption peaks of run M0166 was characterized by full widths at half maximum of 34, 74 and 45 cm1 for the absorption peaks of 3330, 3470 and 3620 cm1, respectively. The absorption peaks of the sample deformed at 1900 K (run M0162) were sharper than those of the samples deformed at 1800 K (runs M0180 and M0166). The broad IR absorption peak was also observed in wadsleyite deformed using the RDA (Kawazoe et al., 2010a; Nishihara et al., 2008). The broad absorption peak at 3150–3700 cm1 and the broadening of the IR absorption peaks may be attributed to distortion of crystal lattice of the deformed samples because of dislocation creep deformation of the deformed samples. The difference in sharpness of the absorption peaks between the samples deformed at 1800 and 1900 K may be due to difference in degree of the distortion of the crystal lattice. The water (or hydrogen) content in the samples synthesized or deformed at 17.6 GPa and 1800 K ranged from 50 to 134 wt ppm H2O (820–2180 H/106 Si). The sample deformed at 1900 K contained 230 wt ppm H2O (3700 H/106 Si). These results suggest that water of 50–200 wt ppm H2O was contained in the sample synthesized at 17.6 GPa and 1800 K before the deformation process, and the wadsleyite aggregate were deformed under slightly hydrous condition at 17.6 GPa and 1800 K. 3.4. Crystallographic preferred orientation and seismic anisotropy Crystallographic orientation distributions of the wadsleyite samples deformed or synthesized at 17.6 GPa and 1800–1900 K are shown as the pole and inverse pole figures in Figs. 6 and 7, respectively. The pole and inverse pole figures of the sample synthesized at 1800 K (strain of 0.0) show weak concentration of the crystallographic axes, indicating nearly random orientation distribution of the wadsleyite grains. In contrast, the concentration of the crystallographic axes clearly increased with increasing shear

95

T. Kawazoe et al. / Physics of the Earth and Planetary Interiors 216 (2013) 91–98

(a) M0180

(b) M0180

a c

[011]* a*

200 nm

200 nm

(c) M0162

(d) M0187

a c a* 200 nm

[011]*

200 nm

Fig. 4. Bright-field TEM images of the dislocation microstructure of the samples. (a and b) The sample deformed to shear strain of 0.4 at 1800 K (run M0180). An arrow indicates subgrain boundary in (a). (c) The sample deformed at 1900 K (run M0162). (d) The sample synthesized at 1800 K (run M0187). The samples were sectioned in parallel to the shear direction and the shear plane normal.

Fig. 5. Unpolarized FT-IR spectra of the wadsleyite samples deformed or synthesized at 17.6 GPa and 1800–1900 K. (a) Those of the samples deformed to shear strain of 0.4 at 1800 K (run M0180) and to shear strain of 0.5 at 1900 K (run M0162). (b) Those of the samples deformed to shear strain of 0.2 (run M0166) and synthesized (run M0187) at 1800 K.

strain to 0.4 at 1800 K, and the developed CPO pattern was characterized by the [0 0 1] axes subparallel to the shear direction and the (0 1 0) planes subparallel to the shear plane. In the case of the sample deformed to shear strain of 0.5 at 1900 K (run M0162), the sample exhibited the weak CPO pattern similar to that of the samples

deformed at 1800 K. In the M0162 sample, the [0 0 1] axes tend to be subparallel to the shear direction and the [hk0] axes subparallel to the shear plane normal. Fig. 8 shows the seismic anisotropy (the three-dimensional relationship between the elastic wave speeds and the deformation

96

T. Kawazoe et al. / Physics of the Earth and Planetary Interiors 216 (2013) 91–98

[100]

[010]

geometry) corresponding to the wadsleyite aggregate deformed to shear strain of 0.4 at 17.6 GPa and 1800 K (run M0180). A fast (polarization) direction of fast shear-wave, which corresponds to SV-wave in a horizontal flow, is perpendicular to the shear direction (on the shear plane). The sample produces 3.6% of shear-wave anisotropy in the polarization direction. Compressional wave, which corresponds to P-wave, is characterized by slowest and fastest velocities along the shear direction and normal to the shear plane, respectively.

[001]

(a)

1800 K

γ 0.4

M0180

(b)

1800 K

γ 0.2

M0166

(c)

4. Discussion

1800 K γ 0.0 M0187

4.1. Deformation mechanism and microstructures

(d)

1900 K

γ 0.5

M0162 Shear direction

1 2 3 4

Fig. 6. Pole figures of crystal axes for the wadsleyite samples with equal-area lower hemisphere projections. The samples were deformed to shear strains of 0.4 (a) and 0.2 (b) and synthesized (c) at 17.6 GPa and 1800 K or deformed to shear strain of 0.5 at 17.6 GPa and 1900 K (d). Numbers in legend represents density of data points. The east–west direction corresponds to the shear direction (shown as arrows), and the north–south pole corresponds to direction of the shear plane normal. Sense of shear is dextral.

Shear direction [001]

Shear plane normal

[010] [001]

[010]

(a)

1800 K

γ 0.4

M0180 [100]

[100]

1 2 3

(b)

1800 K

γ 0.2

M0166

(c)

1800 K

γ 0.0

M0187

(d)

1900 K

γ 0.5

M0162

Fig. 7. Inverse pole figures of crystal axes for the wadsleyite samples with equalarea projections. The samples were deformed to shear strains of 0.4 (a) and 0.2 (b) and synthesized (c) at 17.6 GPa and 1800 K or deformed to shear strain of 0.5 at 17.6 GPa and 1900 K (d). Numbers in legend represents density of data points.

Dominant deformation mechanism in the present study is inferred as dislocation creep from the following observations on the deformation microstructures: (i) the dislocation structure (Fig. 4) and (ii) the CPO development (Figs. 6 and 7). The CPO is primarily developed by the lattice rotation combined with dislocation glide. Moreover, the TEM observation revealed the free dislocations with high density of 1013 m2 and subgrain boundaries in the deformed samples, which suggest that the deformation of the samples was due to dislocation motion involving both dislocation glide and climb. The deformation mechanism for these samples is estimated to be a power-law dislocation creep. In the previous study using the RDA (Kawazoe et al., 2010a), evidences of the dislocation creep in the wadsleyite deformation were observed in the samples deformed at 15 GPa, 1690– 2100 K and strain rates of 2.6–16  105 s1 with the grain size of 1–10 lm and the water content of 12–140 wt ppm H2O. In addition, TEM observations on deformed wadsleyite with the grain size of 0.5–1 lm suggested operation of diffusion creep at 17 GPa, 2100 K and strain rate of 3.8  105 s1 in the previous study. The deformation of wadsleyite by the dislocation creep observed in the present study is consistent with the result obtained at 15 GPa, 1690–2030 K and strain rates of 2.6–16  105 s1 using the RDA in the previous study. The grain size and its homogeneity differ between the samples deformed at 1800 and 1900 K (Table 1 and Fig. 3). The grain size of the samples synthesized and deformed at 1800 K was 3 lm and its size distribution was nearly homogeneous. On the other hand, the sample deformed at 1900 K composed of grains of 5–40 lm, and the grain size at the outer region was larger than that at the central region. The large grain size of the sample deformed at 1900 K can be attributed to effects of temperature and water content on grain-growth kinetics in wadsleyite (Nishihara et al., 2006) because the experimental temperature and the water content were higher compared to those of the samples deformed at 1800 K. The heterogeneity in the grain size of the sample deformed at 1900 K was probably due to temperature gradient in the cell assembly because the wadsleyite grains were larger at high temperature region near the heater where the grains could grow faster. 4.2. Crystallographic preferred orientation and slip systems The CPO development of a deformed aggregate is primarily controlled by relative contribution of each slip system to total strain, and the easiest slip system has the largest influence on the CPO pattern. The following slip systems were identified by the previous TEM observations on wadsleyite deformed at 14–19 GPa and 300– 2273 K in the stress relaxation tests (Thurel and Cordier, 2003; Thurel et al., 2003): [1 0 0](0 1 0), [1 0 0](0 0 1), [1 0 0]{0 1 1}, [1 0 0]{0 2 1}, 1/2h1 1 1i{1 0 1}, [0 1 0](0 0 1), [0 1 0]{1 0 1} and h1 0 1i(0 1 0). In addition, another TEM study characterized [0 0 1](0 1 0) slip system as one of the dominant slip systems in

97

T. Kawazoe et al. / Physics of the Earth and Planetary Interiors 216 (2013) 91–98

VP (km)

(a)

(b)

VS1 (km)

AV S (%)

VS1 polarization

10.51

5.97

2.93

10.25

5.89

0.05

10.60

6.04

3.58

10.25

5.86

0.12

Shear direction

Shear direction

Fig. 8. The seismic anisotropy of the wadsleyite aggregate deformed to shear strain of 0.4 at 17.6 GPa and 1800 K. These are shown as pole figures with equal-area lower hemisphere projections. The east–west direction corresponds to the shear direction (shown as arrows). (a) The north–south pole corresponds to direction of the shear plane normal. (b) The center of plot corresponds to the direction of the shear plane normal. This coordinate is chosen to clearly show the polarization directions of the fast shearwave. Abbreviations: VP, compressional-wave velocity; VS1, fast shear-wave velocity; AVS, anisotropy strength.

the deformed wadsleyite, in which dislocation glide to the [0 0 1] direction occurred by glide of dissociated dislocations (Sharp et al., 1994). The activation of [0 0 1](0 1 0) slip system was also suggested by the wadsleyite CPO study by the stress relaxation tests at 16 GPa and 1673 K (Demouchy et al., 2011). In the present study, we found the wadsleyite CPO which was characterized by the [0 0 1] axis subparallel to the shear direction and the (0 1 0) plane subparallel to the shear plane. The wadsleyite CPO pattern indicates that [0 0 1](0 1 0) slip system was dominantly activated in wadsleyite deformed at 17.6 GPa, 1800–1900 K and the water content of 50–230 wt ppm H2O. Recently, Demouchy et al. (2011) reported change in the wadsleyite CPO with the water content as a result of the stress relaxation tests, and the result of the present study is consistent with the CPO pattern observed in the samples deformed under water-rich conditions in their study. From a mineralogical point of view, dislocation glide on the wadsleyite (0 1 0) plane is expected to be easier than those on the other planes because the (0 1 0) plane is a close-packed plane of oxygen and the dislocation glide does not break strong Si–O bonds. According to the TEM observation on deformed wadsleyite (Sharp et al., 1994), dislocation glide to the [0 0 1] direction on the (0 1 0) plane was characterized by the glide of dissociated dislocations, which enhances activity of the slip system through reduction of dislocation energy. Metsue et al. (2010) discussed the activities of the wadsleyite slip systems by evaluating their Peierls stresses based on a Peierls–Nabarro model, and concluded that the [0 0 1](0 1 0) slip system was not the easiest one. However, the conclusion may not be applicable to the CPO development because the CPO develops in the power-law dislocation creep regime, in which the relative activities of the slip systems can be different with those in the Peierls dislocation creep regime. Thus the [0 0 1](0 1 0) slip system is plausible as the dominant slip system based on the mineralogical consideration. The pole and inverse pole figures of the sample deformed to shear strain of 0.4 at 1800 K clearly show strong concentrations of the crystallographic axes (Figs. 6 and 7). On the other hand, the sample exhibited the weak CPO pattern in the case of the sample deformed to shear strain of 0.5 at 1900 K. In the sample deformed at 1900 K, activity of the [0 0 1](0 1 0) slip system might become close to those of the active slip systems other than the [0 0 1](0 1 0) slip system. In this case, the [0 0 1](0 1 0) slip system weakly influenced on the CPO development compared to the deformation at 1800 K, and the concentrations of the crystallographic

axes became weaker than that of the sample deformed to shear strain of 0.4 at 1800 K. 4.3. Geophysical implications for the seismic anisotropy The polarization anisotropy, VSH/VSV, is compared among the recent seismological observation (Visser et al., 2008) and those calculated for a mantle rock with a typical pyrolite composition in Fig. 9 in order to deduce style of the mantle convection in the MTZ. In the VSH/VSV calculation, we used the elastic constant tensors Cij of the wadsleyite samples deformed to shear strains of 0.2 and 0.4 at 17.6 GPa and 1800 K (Table 2), which are calculated based on the CPO data of the present study and Cij of single-crystal wadsleyite. The VSH/VSV ratio for the mantle rock is reduced to 60% of that of the wadsleyite aggregate taking into account a dilution effect by majoritic garnet because the pyrolite mantle is composed of 60% wadsleyite and 40% majoritic garnet at the most part of the upper MTZ (450–510 km depth) (Irifune and Isshiki, 1998).

Fig. 9. A comparison of VSH/VSV ratios between those calculated for a mantle rock and seismological observation in the MTZ (Visser et al., 2008). Squares and circles represent the VSH/VSV ratios calculated for horizontal and vertical flows using the wadsleyite CPO, respectively. The seismological observation is shown as a hatched area. The VSH/VSV ratio of 1.00, which is expected for an isotropic rock, is also shown as a dashed line.

98

T. Kawazoe et al. / Physics of the Earth and Planetary Interiors 216 (2013) 91–98

Table 2 Elastic constant tensor Cij (in GPa) of deformed samples. i/j

1

2

3

M0180 (17.6 GPa, 1800 K and c 0.4) 1 407.0 147.6 147.1 2 427.9 146.2 3 412.2 4 5 6 M0166 (17.6 GPa, 1800 K and c 0.2) 1 409.1 147.4 147.4 2 421.8 146.1 3 416.0 4 5 6

4

5

6

0.3 0.1 1.5 136.9

0.2 0.6 1.5 0.8 130.1

1.1 0.1 0.0 0.1 0.3 135.6

0.0 0.6 0.0 136.0

0.3 0.4 1.7 0.7 131.8

0.9 0.5 0.0 0.3 0.2 134.7

Reference axes are defined as 1: shear direction, 2: shear plane normal and 3: perpendicular to both 1 and 2 directions.

The VSH/VSV ratio of the mantle rock deviates from 1.000 (isotropic structure) by the shear deformation, and those for the horizontal and vertical flows develop to 0.988 and 1.009 with increasing shear strains to 0.4, respectively (Fig. 9). Magnitudes of the VSH/ VSV ratios for the horizontal and vertical flows likely further deviates at shear strain larger than 0.4 because steady state fabric strength of minerals has been reported to be achieved at higher shear strains (e.g., shear strain of 4 in the case of olivine: Bystricky et al., 2000). Thus, the polarization seismic anisotropy observed in the MTZ is between those developed by the horizontal and vertical flows in the MTZ based on the wadsleyite CPO of the present study. Under the assumption that the wadsleyite CPO observed in the present study dominates in the MTZ, the polarization anisotropy (VSH/VSV: 0.992–0.996 ± 0.002) reported by Visser et al. (2008) can be attributed to predominance of the horizontal flow in the MTZ with small contribution of the vertical flow. In contrast, the polarization anisotropy, VSH/VSV, of 1.002–1.005, which was reported by the previous seismological observation (Montagner and Kennett, 1996), is explained by predominance of the vertical flow in the MTZ with small contribution of the horizontal flow. The wadsleyite CPO found in the present study produces the azimuthal seismic anisotropy in which fast (polarization) direction of fast S-wave is perpendicular to the flow direction in the shear flow. This feature is same with those of olivine B-type CPO (Jung and Karato, 2001) and serpentine CPO (Katayama et al., 2009), which are observed at subduction zones. Recent seismological measurements of the shear wave-splitting revealed the azimuthal seismic anisotropy characterized by trench-parallel fast direction of the S-wave in the MTZ beneath the Tonga trench (Foley and Long, 2011). The mantle flow in the MTZ is interpreted as trenchnormal beneath the Tonga trench under the assumption that the wadsleyite CPO observed in the present study developed in the MTZ. Acknowledgements This study has been supported by global COE program ‘‘Deep Earth Mineralogy’’ and by grants from the Ministry of Education, Culture, Sports, Science, and Technology of the Japanese Government (No. 22740346, 22340161 and 24540515). References Becker, T.W., Kellogg, J.B., Ekström, G., O’Connell, R.J., 2003. Comparison of azimuthal seismic anisotropy from surface waves and finite strain from global mantle-circulation models. Geophys. J. Int. 155 (2), 696–714.

Bystricky, M., Kunze, K., Burlini, L., Burg, J.-P., 2000. High shear strain of olivine aggregates: rheological and seismic consequences. Science 290, 1564–1567. Demouchy, S., Mainprice, D., Tommasi, A., Couvy, H., Barou, F., Frost, D.J., Cordier, P., 2011. Forsterite to wadsleyite phase transformation under shear stress and consequences for the Earth’s mantle transition zone. Phys. Earth Planet. Inter. 184 (1–2), 91–104. Foley, B.J., Long, M.D., 2011. Upper and mid-mantle anisotropy beneath the Tonga slab. Geophys. Res. Lett. 38 (2), L02303. Gasparik, T., 2003. Phase Diagrams for Geoscientists. Springer, Berlin. Irifune, T., Isshiki, M., 1998. Iron partitioning in a pyrolite mantle and the nature of the 410-km seismic discontinuity. Nature 392, 702–705. Jung, H., Karato, S., 2001. Water-induced fabric transitions in olivine. Science 293, 1460–1463. Karato, S., 2008. Deformation of Earth Materials: Introduction to the Rheology of the Solid Earth. Cambridge University Press, Cambridge. Katayama, I., Hirauchi, K.-i., Michibayashi, K., Ando, J.-i., 2009. Trench-parallel anisotropy produced by serpentine deformation in the hydrated mantle wedge. Nature 461, 1114–1117. Katayama, I., Jung, H., Karato, S., 2004. New type of olivine fabric at modest water content and low stress. Geology 32, 1045–1048. Kawazoe, T., Karato, S.-i., Ando, J.-i., Jing, Z., Otsuka, K., Hustoft, J.W., 2010a. Shear deformation of polycrystalline wadsleyite up to 2100 K at 14–17 GPa using a rotational Drickamer apparatus (RDA). J. Geophys. Res. 115 (B8), B08208. Kawazoe, T., Nishihara, Y., Ohuchi, T., Nishiyama, N., Higo, Y., Funakoshi, K., Irifune, T., 2011. In situ stress–strain measurements in a deformation-DIA apparatus at P–T conditions of the upper part of the mantle transition zone. Am. Mineral. 96, 1665–1672. Kawazoe, T., Nishiyama, N., Nishihara, Y., Irifune, T., 2010b. Deformation experiment at P-T conditions of the mantle transition zone using D-DIA apparatus. Phys. Earth Planet. Inter. 183 (1–2), 190–195. Mainprice, D., 1990. A FORTRAN program to calculate seismic anisotropy from the lattice preferred orientation of minerals. Comput. Geosci. 16 (3), 385–393. Mainprice, D., 2007. Seismic anisotropy of the deep earth from a mineral and rock physics perspective. In: Price, G.D. (Ed.), Treatise on Geophysics: Mineral Physics. Elsevier, Amsterdam, pp. 437–491. Metsue, A., Carrez, P., Denoual, C., Mainprice, D., Cordier, P., 2010. Plastic deformation of wadsleyite: IV Dislocation core modelling based on the Peierls–Nabarro–Galerkin model. Acta Mater. 58 (5), 1467–1478. Montagner, J.-P., Kennett, B.L.N., 1996. How to reconcile body-wave and normalmode reference Earth models. Geophys. J. Int. 125, 229–248. Montagner, J.-P., Nataf, H.-C., 1986. A simple method for inverting the azimuthal anisotropy of surface waves. J. Geophys. Res. 91, 511–520. Nakajima, J., Shimizu, J., Hori, S., Hasegawa, A., 2006. Shear-wave splitting beneath the southwestern Kurile arc and northeastern Japan arc: a new insight into mantle return flow. Geophys. Res. Lett. 33. http://dx.doi.org/10.1029/ 2005GL025053. Nishihara, Y., Shinmei, T., Karato, S., 2006. Grain-growth kinetics in wadsleyite: effects of chemical environment. Phys. Earth Planet. Inter. 154, 30–43. Nishihara, Y., Tinker, D., Kawazoe, T., Xu, Y., Jing, Z., Matsukage, K.N., Karato, S.-i., 2008. Plastic deformation of wadsleyite and olivine at high-pressure and hightemperature using a rotational Drickamer apparatus (RDA). Phys. Earth Planet. Inter. 170 (3–4), 156–169. Ohuchi, T., Kawazoe, T., Nishihara, Y., Nishiyama, N., Irifune, T., 2011. High pressure and temperature fabric transitions in olivine and variations in upper mantle seismic anisotropy. Earth Planet. Sci. Lett. 304 (1–2), 55–63. Paterson, M.S., 1982. The determination of hydroxyl by infrared absorption in quartz, silicate glass and similar materials. Bull. Mineral. 105, 20–29. Sharp, T.G., Bussod, G.Y., Katsura, T., 1994. Misrostructures in b-Mg1.8Fe0.2SiO4 experimentally deformed at transition-zone conditions. Phys. Earth Planet. Inter. 86, 69–83. Thurel, E., Cordier, P., 2003. Plastic deformation of wadsleyite: I. High-pressure deformation in compression. Phys. Chem. Miner. 30, 256–266. Thurel, E., Cordier, P., Frost, D.J., Karato, S., 2003. Plastic deformation of wadsleyite: II. High-pressure deformation in shear. Phys. Chem. Miner. 30, 267–270. Tommasi, A., Mainprice, D., Cordier, P., Thoraval, C., Couvy, H., 2004. Strain-induced seismic anisotropy of wadsleyite polycrystals and flow patterns in the mantle transition zone. J. Geophys. Res. 109. http://dx.doi.org/10.1029/2004JB003158. Trampert, J., van Heijst, H.J., 2002. Global azimuthal anisotropy in the transition zone. Science 296, 1297–1299. Visser, K., Trampert, J., Lebedev, S., Kennett, B.L.N., 2008. Probability of radial anisotropy in the deep mantle. Earth Planet. Sci. Lett. 270 (3–4), 241–250. Wang, Y., Durham, W.B., Getting, I.C., 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. Xu, Y., Nishihara, Y., Karato, S., 2005. Development of a rotational Drickamer apparatus for large-strain deformation experiments under deep Earth conditions. In: Chen, J., Wang, Y., Duffy, T.S., Shen, G., Dobrzhinetskaya, L.F. (Eds.), Frontiers in High-Pressure Research: Applications to Geophysics. Elsevier, Amsterdam, pp. 167–182. Zha, C.-S., Duffy, T.S., Mao, H.-K., Downs, R.T., Hemley, R.J., Weidner, D.J., 1997. Single-crystal elasticity of b-Mg2SiO4 to pressures of the 410 km seismic discontinuity in the Earth’s mantle. Earth Planet. Sci. Lett. 147, E9–E15.