FTIR spectra of garnet (a) and clinopyroxene (b). (a) The 3570 cm-1 peak is from structural OH in the garnets. (b) The 3460, 3540 and 3640 cm − 1 peaks are from structural OH in the clinopyroxene 

Contexts in source publication

Context 1

... point, with the color reflecting the lattice orientation of garnet. In the case that the <100>, <110>, and <111> axes of garnet were oriented parallel to the lineation (X), the grain was colored red, green, or blue, respectively. White areas in the maps are points that were not indexed, or minerals other than garnet. Black lines, representing grain boundaries, are drawn between any adjacent points with a misorientation > 10 . The orientation maps (Fig. 7c, d) show the occurrence of grain boundaries and fractures. The grain size of garnet is similar to that of clinopyroxene (~0.3 mm). Garnet is flattened and elongated in the garnet clinopyroxenites (Fig. 7c, d), and some grain boundaries consist of interfingering sutures (Fig. 7c, d; Stipp et al ., 2002). Some garnet grains contain low-angle boundaries (misorientation angles of 2 – 9°), and aspect ratios are generally ~2, regardless of modal composition. Garnet clinopyroxenites are pervasively cracked by extensional fractures that are generally oriented perpendicular to the stretching lineation (Figs. 6 and 7). Such lineation- normal fractures are common in granulite facies mylonites and UHP metamorphic rocks, and are thought to form during the final stages of exhumation (Ji et al ., 1997, 2003). Single crystals of garnet were hand-picked from crushed samples of garnet clinopyroxenite for TEM observations. TEM samples were thinned using the ion- bombardment technique, and analyzed using a JEOL JEM-2010 (Hiroshima University, Japan) transmission electron microscope at an accelerating voltage of 200 kV. TEM observations of garnet focused on geometrical dislocation microstructures, dislocation densities, and the Burgers vectors of dislocations. The analyzed garnet grains contain extensive, regular dislocation arrays and dislocation networks. The dislocation arrays (Fig. 8a and b) can be interpreted as tilt subgrain boundaries, while the dislocation networks (Fig. 8c and d) indicate the activation of at least two slip systems. These well-organized dislocation microstructures indicate the occurrence of an efficient diffusion-assisted recovery mechanism such as dislocation climb or cross slip. The density of free dislocations within garnet ranges from 3 to 6 10 – 2 cm (Fig. 9). Dislocation junctions, which are the complex intersections (tangles) of several dislocations, were also observed in each sample (Fig. 8f), indicating interaction between dislocations and therefore the operation of multiple slip systems (Voegelé et al ., 1998; Wang and Ji, 1999; Ji et al ., 2003). The above observations are in agreement with previous TEM investigations on naturally deformed garnets (Ando et al ., 1993; Doukhan et al ., 1994; Ji and Martignole, 1994; Voegele ́ et al ., 1998; Prior et al ., 2000; Ji et al ., 2003), all of which reported that garnet can deform plastically given appropriate conditions of temperature, pressure, differential stress, and strain rate. The Burgers vectors of dislocations were identified using the g ⋅ b = 0 and g ⋅ b u = 0 criteria. The majority of dislocations have a Burgers vector b = 1⁄2<111>. We analyzed the chemical compositions of garnet grains within four samples of garnet clinopyroxenites (Fig. 10) using a JEOL electron microprobe (JXA733) housed at the Centre for Instrumental Analysis, Shizuoka University, Japan. Analytical conditions were 15 kV accelerating voltage, 12 nA probe current, and beam diameter of 20 m, using a count time of 20 s and 10 s background. Table 3 lists the results of analyses of 10 representative garnet grains. The grains have similar compositions (Fig. 10a). Microprobe analyses of 75 garnet grains from the four samples yielded the following compositional ranges: 47.3 – 54.8% pyrope, 27.7 – 30.9% almandine, 16.7 – 20.2% grossular, and 0.8 – 1.6% spessartine, with an average composition of Alm 29.9±1.4 Prp 50.5±2.6 Grs 18.6±1.8 Spe 1.0±0.3 . Most of the grains show no significant compositional zoning, although the almandine component shows a slight increase around grain boundaries and cracks, whereas the pyrope component shows a slight decrease. Clinopyroxene grains are near-homogeneous, contain a diopside component (Fig. 10b), and are compositionally similar among samples. We measured the water content in garnet grains using Fourier-transform infrared (FTIR) spectroscopy. For each sample, a polished thin section ( ~ 0.25 mm thick) was prepared and dried in an oven at 120 °C for 4 h. The infrared spectrum was obtained using a Perkin Elmer, Spectrum 2000 (Geochemical Laboratory, University of Tokyo) at room temperature and in the wavenumber range of 740 – 7800 cm – 1 . A series of 100 scans was – 1 averaged for each spectrum with a resolution of 1 cm . For analysis, we carefully selected 2 crack-free and unaltered regions of 60 60 m in size. We used FTIR to analyze 38 grains (29 of garnet, 9 of cpx) from four samples. In the typical OH vibration region (3400 – 3800 cm – 1 ), all grains show several absorption bands. Figure 11 shows typical FTIR spectra of garnet and clinopyroxene from the Higashi-akaishi garnet clinopyroxenite. The infrared (IR) spectra of all garnet grains show – 1 – 1 a sharp peak at ~3570 cm (Fig. 11a); one sample (GM01) shows a peak at ~3430 cm , which is typical of the stretching vibrations (ν3+ν1) of molecular water, which may occur in submicroscopic fluid inclusions within garnet. Following previous studies of natural garnets (Rossman and Smyth, 1990; Bell and Rossman, 1992), we ascribed this group of – 1 bands to submicroscopic fluid inclusions. The 3570 cm peak is produced by structural OH in the garnets. The IR spectra of clinopyroxene show sharp peaks at 3460, 3540, and – 1 3640 cm (Fig. 11b), produced by structural OH (Bell et al ., 1995). We calculated the water content (H 2 O ppm wt.) of both minerals using the Beer – Lambert law (absorbance = absorbance coefficient thickness water concentration). Absorbance is expressed as the integrated absorbance areas of OH. We used the following 2 integrated molar absorbance coefficients from Bell et al . (1995): 1.39 ppm H 2 O/cm for 2 garnet, 7.09 ppm H 2 O/cm for clinopyroxene. Table 4 provides detailed information on peak positions, absorbance, and the calculated water content. H 2 O contents in garnet range from 17 to 1000 ppm (mainly ~60 ppm) (Fig. 12a). The H 2 O content of Higashi-akaishi garnet varies both among and within samples. In contrast, H 2 O contents in clinopyroxene are relatively homogeneous within and among samples (~70 ppm)(Fig. 12b). We measured the crystal-preferred orientations (CPOs) of olivine, garnet, and clinopyroxene grains from highly polished thin sections, using a scanning electron microscope equipped with an electron-backscatter diffraction system (JEOL JSM6300 with HKL Channel5), housed at the Centre for Instrumental Analysis, Shizuoka University, Japan. We measured about 200 crystal orientations per sample, visually confirming the computerized indexation of the diffraction pattern for each orientation. The measured CPOs are presented on equal-area, lower-hemisphere projections in the structural (XZ) reference frame (Figs. 13 – 17). To characterize the CPOs, we determined the fabric strength and distribution density of the principal crystallographic axes (e.g., Michibayashi and Mainprice, 2004). The rotation matrix between crystal and sample co- ordinates is used to describe the orientation g of a grain or crystal in sample co-ordinates. In practice, it is convenient to describe the rotation by a triplet of Euler angles; e.g., g = ( φ 1 , , φ 2 ), as used by Bunge (1982). The orientation distribution function (ODF), f ( g ), is defined as the volume fraction of orientations in the interval between g and g + d g in a space containing all possible orientations, as given ...

Context 2

... Kleinschrodt and McGrew, 2000; Ji et al ., 2003; Michibayashi et al ., 2004; Terry and Heidelbach, 2004; Okamoto and Michibayashi, 2005; Storey and Prior, 2005; Bestmann et al ., 2008). Some of these studies have suggested that dislocation creep is the dominant deformation mechanism for garnet, based on TEM observations of dislocation networks and subgrain walls (e.g., Ando et al ., 1993; Ji and Martignole, 1994; Ji et al ., 2003); however, the problem exists that although numerical simulations (performed using the VPSC model) of CPO development in garnet produce characteristic CPOs for both axial compression and simple shear deformation (Mainprice et al ., 2004), very weak or random CPOs patterns are obtained for naturally occurring elongate garnet (Ji et al ., 2003; Storey and Prior, 2005). Consequently, alternative deformation mechanisms have been proposed. Storey and Prior (2005) suggested that plastic deformation of garnet is dominated by grain- boundary sliding accompanied by subgrain formation and rotation, rather than dislocation creep. Similarly, in an analysis of experimentally deformed eclogites, Zhang and Green (2007) reported the repeated ‘sliding off’ into the foliation of superficial layers of recrystallized garnet. However, we consider it unlikely that garnet within the Higashiakaishi mass was deformed in this way, as it shows contrasting textures to those reported in the above studies. For example, Storey and Prior (2005) reported fine recrystallized garnet (~50 m) around coarse garnet (~1 mm), whereas garnet clinopyroxenites of the Higashi-akaishi mass contain elongate garnet crystals of homogeneous grain size (0.3 mm) (Fig. 7). In addition, some of the grain boundaries observed in the present study consist of interfingering sutures, and the grain size is much coarser than that generally associated with grain-boundary sliding. Therefore, it is difficult to explain the observed garnet deformation in terms of grain-boundary sliding. Some of the analyzed garnet grains contain low-angle internal boundaries. TEM observations indicate that the garnet grains contain extensive, regular dislocation arrays and dislocation networks, suggesting that the low-angle boundaries are sub-grain boundaries. The density of free dislocations within the analysed garnet grains ranges from 7 – 2 3 to 6 10 cm (Fig. 9), which is relatively high (Ando et al ., 1993); however, garnet has low strength and shows complex CPO patterns. Garnet also has a high degree of symmetry and 12 potential slip systems (Ji et al ., 2003); consequently, any given slip plane needs only undergo a small amount of rotation before the resolved shear stress reaches high levels on a different slip system, such as 1/2 <111> {110}. Although slip occurs predominantly on {110} planes, it is important to realize that three {110}-type planes intersect in a [111] direction and that screw dislocations with a 1/2 <111> Burgers vector may migrate randomly on {111} planes with high resolved shear stress. Thus, the weakness of garnet CPOs does not provide unequivocal evidence for diffusion creep or against dislocation creep as a deformation mechanism within garnet (Ji et al ., 2003). Therefore, we propose that dislocation creep was the dominant deformation mechanism in garnet within the Higashi-akaishi garnet clinopyroxenites. The studied garnet clinopyroxenites are typically bimineralic, being composed of garnet and diopside. Figure 18b shows grain size and aspect ratio data for garnet and diopside within these rocks, with respect to modal composition. Grain sizes and aspect ratios in garnet are comparable with those in clinopyroxene, regardless of the modal composition. Since the garnet clinopyroxenites were plastically sheared along with the foliated dunites, these observations reveal that the two minerals deformed under similar degree of plasticity. The fabric strength ( M -index and J -index) of a deformed rock is related to finite strain, as the M -index ( J -index) increases with finite plastic strain (Tommasi et al ., 2000; Skemer et al ., 2005). Figure 18a shows that the M -index of both garnet and clinopyroxene increases with increasing modal composition of garnet. Because the distribution of garnet and clinopyroxene is homogeneous in the analyzed samples (Fig. 6) , the dominant phase controls the deformation of the rock. If all of the garnet clinopyroxenites had deformed under the same stress conditions, the relationship shown in Fig. 18a suggests that garnet- dominated part has been more strained than clinopyroxene-diminated part. However, the results of an experimental study performed at high temperature and pressure (1500 K, 3 GPa) revealed that garnet is three to four times as strong as omphacite (Jin et al ., 2001). The olivine fabrics measured in the present study suggest that the garnet clinopyroxenites were plastically sheared under wet conditions, and FTIR analyses reveal H 2 O contents in garnet of 17 – 1000 ppm (mostly ~60 ppm; Fig. 11; Table 4). It is possible that the presence of water influences the deformation of garnet. Indeed, Katayama and Karato (2008) reported that the creep rate of Mg-rich garnet (Alm 19 Prp 68 Grs 12 ) is sensitive to water. Figure 22 shows the relation between stress and strain rate for garnet and diopside under dry and wet conditions. Here, the water contents of wet and dry garnet are 80 – 200 ppm and < 30 ppm, respectively (Katayama and Karato, 2008). The results reveal contrasting influences of water on the deformation of garnet and diopside: under wet conditions compared with dry, the strain rate increases by two orders of magnitude for garnet but by an order of magnitude for diopside. Given the influence of water on the creep strength of garnet, garnet within the Higashi-akaishi mass may have become significantly as weak as clinopyroxene during deformation. It should be, however, noted that the creep strength of minerals may be also influenced by some other effects such as stress or pressure as well as the effect of water (e.g., Chen et al., 2006; Li et al., 2006; Katayama and Karato, 2008; Amiguet et al., 2009). The rheological behavior of garnet is an important factor in the deformation of subducted oceanic crust, as oceanic crust that has been deeply subducted (> 400 km depth) is composed largely of garnet (Ringwood, 1982). The nominally anhydrous mineral phases (NAMs; olivine, pyroxene, and garnet) in both subducting oceanic crust and the overlying mantle wedge can carry a significant amount of H 2 O to the deep mantle (Forneris and Holloway, 2003; Iwamori, 2007). If water has a stronger influence on the creep strength of garnet than on that of clinopyroxene, as shown above, oceanic crust would be expected to weaken with ongoing subduction. This hypothesis is the opposite to that suggested in previous studies (e.g., Karato et al ., 1995; Jin et al ., 2001) and may have implications for our understanding of mantle convection. Dunites in the Higashi-akaishi peridotite mass, located in the subduction-type Sanbagawa metamorphic belt, record two stages of deformation (D1 and D2; Mizukami and Wallis, 2005) and contain various microstructures, ranging from coarse-grained to porphyroclastic. At Gongen Pass, dunites contain fine-grained textures with weak fabric strength, suggesting that the outcrop represents the center of a D2 shear zone. CPO patterns for olivine are B-type regardless of texture. Accordingly, dunite within the Higashi-akaishi mass is interpreted to have been deformed under high-stress and wet conditions. Garnet clinopyroxenites that occur within foliated dunite at Gongen Pass contain 3 – 80% garnet, and garnet and clinopyroxene within these rocks have a homogeneous distribution. Garnet contains extensive, regular dislocation arrays and dislocation networks, suggesting that dislocation creep was the dominant deformation mechanism. Analyses of orientation maps reveal that garnet and clinopyroxene have similar grain sizes and aspect ratios, regardless of modal composition. These findings indicate that the two minerals were deformed under similar conditions of plasticity. In addition, M -index values for both garnet and clinopyroxene increase with increasing modal composition of garnet. During deformation, garnet was possibly weaker than clinopyroxene. The obtained olivine CPOs indicate deformation under wet conditions, and the water content of garnet is ~60 ppm. It is possible that the presence of water helped to induce garnet deformation, and flow laws indicate that under water-rich conditions (relative to dry conditions), the strain rate for garnet increases by two orders of magnitude, whereas for diopside it increases by an order of magnitude. This finding may have significant implications for our understanding of mantle convection, as garnet may in fact be weaker than clinopyroxene. The authors thank Hiroyuki Kagi of University of Tokyo for assistance with FTIR analyses and M. Satish-Kumar of Shizuoka University for assistance with EPMA analyses. We would like to thank Toshiaki Masuda of Shizuoka University for help in using the Ion thinning technique, and Yasuro Kugimiya for providing samples for analysis. MM wishes to thank Takumi Nanbu, Masahiro Hara, Kyoko Kanayama, and members of the Michibayashi Laboratory (Yumiko Harigane, Takako Satsukawa, Hisashi Imoto, Ayano Fujii, Shigeki Uehara, Tatsuya Ohara, Makoto Suzuki, Naohiko Ueta, Naoaki Komori, and Yuri Shinkai) for their helpful advice and discussions. Amiguet, E., Raterron, P., Cordier, P., Couvy, H. and Chen, J., 2009. Deformation of diopside single crystal at mantle pressure, 1, Mechanical data. Physics of the Earth and Planetary Interiors , 177 , 122-129. Ando, J., Fujino, K. and Takeshita, T., 1993. Dislocation microstructures in naturally deformed silicate garnets. Physics of the Earth and Planetary Interiors, 80 , 105-116. Aoya, M., 2001. P-T-D path of eclogite from the Sambagawa belt deduced from combination of ...