American Mineralogist, Volume 94, pages 1130–1136, 2009

0003-004X/09/0809–1130$05.00/DOI: 10.2138/am.2009.3150 1130

Single crystal growth of wadsleyite

Anton ShAtSkiy,1,2,* konStAntin D. LitASov,1,2 tAkuyA MAtSuzAki,3 keiji ShinoDA,4 DAiSuke yAMAzAki,3 AkirA yoneDA,3 eiji ito,3 AnD toMoo kAtSurA3

1Department of Earth and Planetary Material Science, Graduate School of Science, Tohoku University, Sendai 980-8578, Japan2V.S. Sobolev Institute of Mineralogy and Petrology, SB RAS, Novosibirsk 630090, Russia

3Institute for Study of the Earth’s Interior, Okayama University, Misasa, Tottori 682-0193, Japan4Department of Geosciences, Osaka City University, Osaka 590-0494, Japan

AbStrAct

We have synthesized large (0.7–1.0 mm) crystals of anhydrous, water-bearing, and Fe-bearing wadsleyite by means of growth from solution in the thermal gradient field. Nearly anhydrous (<68±4 wt ppm H2O) Mg2SiO4 crystals were grown using K2Mg(CO3)2 as a solvent at 16.5 GPa and 1700 °C. (Mg0.92Fe0.08)2SiO4 crystals containing 84±17 wt ppm H2O were grown using 92K2Mg(CO3)2-8FeCl2 as a solvent. Crystals of Fe-free wadsleyite with 1496±117 wt ppm H2O were synthesized at 1400 °C and 15.5 GPa by using 2KHCO3-Mg(OH)2 as a solvent.

Keywords: Mg2SiO4, wadsleyite, single crystal growth, high pressure, thermal gradient method

introDuction

Wadsleyite is expected to be the most abundant mineral in the upper part of the mantle transition zone. Mg2SiO4-wadsleyite contains about 11 mol% Fe2SiO4 in a pyrolite composition (Green and Falloon 1998) and can also accomodate up to 3.3 wt% of H2O (e.g., Smyth 1994). Knowledge of the physical properties of wadsleyite and their dependence on composition is essential for interpretation of geophysical observations and understand-ing the structure and dynamics of the mantle transition zone. Large high-quality single crystals are necessary for accurate determination of physical properties, such as elasticity, plasticity, elemental diffusivity, and electrical and thermal conductivity. For example, for understanding the constitution of the Earth’s interior, a comparison of the seismic wave velocity with elastic wave velocity of constituent materials is essential. Resonance ultrasonic spectroscopy, which is one of the most accurate and precise methods for determining elastic constants, requires single crystals with sizes >1 mm (Maynard 1996).

Nevertheless, it is difficult to obtain large and homogeneous single crystals of wadsleyite with the compositions mentioned above. Conventional crystal growth from its own melt may not be applied because of incongruent melting of wadsleyite (Fig. 1). The solid-stage crystallization under “dry” (<3 wt ppm H2O) con-ditions typically produces anhydrous wadsleyite with grain sizes of less than 10 µm (Nishihara et al. 2006). Jacobsen et al. (2005) reported that addition of water to the system makes it possible to obtain hydrous wadsleyite single crystals of up to 500 µm in size, containing approximately 3200 wt ppm H2O. However, crystals larger than 0.5 mm have not yet been synthesized. For practical use, we also need large crystals of anhydrous wadsleyite, which

are even more difficult to grow because wadsleyite absorbs some water from surrounding material of high-pressure cells, even through a sealed metal capsule, and typically contains at least several hundred wt ppm H2O (Nishihara et al. 2006).

In this paper, we describe an experimental technique for crys-tal growth of hydrous, anhydrous, and Fe-bearing wadsleyite, and highlight our advantage in synthesis of 1 mm single crystals.

* E-mail: shatskiy@m.tains.tohoku.ac.jp

Figure 1. Phase relations in the Mg2SiO4 system inferred from the data of Katsura and Ito (1989), Katsura et al. (2003, 2004), and Liebske (2005). Wd = wadsleyite; Fo = forsterite; AnhB = anhydrous phase B; Pc = periclase; Rw = ringwoodite; Pv = perovskite; L = liquid; St = stishovite; Mj = majorite.

SHATSKIy ET AL.: WADSLEyITE CRySTAL GROWTH 1131

experiMentAL StrAtegy

To optimize the synthesis conditions, several aspects such as system composition, temperature, pressure, growth method, and device have to be carefully taken into account.

Synthesis of large wadsleyite single crystals by growth from its own melt or solid-state crystallization is impractical, and therefore crystal growth from solution should be considered as an alternative. As it is well known, water, carbonates, and alkaline elements greatly reduce silicate solidi (Iwamori 2004; Dalton and Presnall 1998; Knoche and Luth 1996). Furthermore, it is reported that carbonate and water-bearing carbonate melts are eminent silicate solvents (Stalder et al. 2001; Dalton and Presnall 1998), which have high mobility and low viscosity (Hammouda and Laporte 2000; Dobson et al. 1996) at high pressures and tem-peratures. Therefore, 2KHCO3-Mg(OH)2 was used as a solvent for hydrous wadsleyite, and K2CO3, Na2CO3, and K2Mg(CO3)2 were examined as solvents for anhydrous wadsleyite.

The following factors were considered to choose appropriate temperatures for synthesis. First, the synthesis temperature must be in the stability field of wadsleyite, and the solvent must be molten. No additional phases, such as another immiscible melt or minerals, should appear. In the system Mg2SiO4-H2O, for example, enstatite and phase E would appear at temperatures below 1350 and 1200 °C, respectively (Inoue 1994; Angel et al. 2001). For this reason, crystal growth of hydrous wadsleyite was conducted at a temperature of 1400 °C. Second, the solvent melts should have poor transport properties, such as low silicate solubility and high viscosity near solidus conditions (Dobson et al. 1996; Dalton and Presnall 1998). Hence, the experimental temperature should be considerably higher than the solidus of the system. At 15–17 GPa, the apparent solidus of alkali-bearing carbonate peridotite is located near 1500 °C (e.g., Ghosh et al. 2009). Therefore, the crystal growth of anhydrous wadsleyite was performed at a temperature of 1700 °C.

In the dry Mg2SiO4 system at 1700 °C, the forsterite-wads-leyite phase transition was observed at 15.7 GPa (Katsura et al. 2004). Consequently, the growth of anhydrous wadsleyite was conducted at 16.5 GPa (Fig. 1). The pressure for crystal growth of hydrous wadsleyite was chosen to be 15.5 GPa, which is slightly higher than that of the forsterite-wadsleyite transition in the Mg2SiO4-H2O system, 14.5 GPa at 1400 °C (Frost and Dolejš 2007).

To avoid chemical zoning in crystals, temperature should be fixed during crystal growth. For example, in the Mg2SiO4-H2O system, the water solubility in wadsleyite has a maximum at temperatures near 1100 °C (Demouchy et al. 2005). Therefore, we used a thermal gradient method, which allows crystal growth at constant temperature (Wentorf 1971; Pal’yanov et al. 1997; Shatskiy et al. 2007). The major idea of this method is the dis-solution of polycrystalline material into a solvent in a hot region, and its subsequent precipitation as a single crystal in a cool region (see Shatskiy et al. 2007 for more details).

Crystal growth from solution requires a significantly larger crystallization chamber than the desirable crystal size. For instance, the crystallization volume of 10 mm3 is needed to grow 1 mm MgSiO3 perovskite crystals (Shatskiy et al. 2007). Consequently, a large-volume (≥10 mm3), high-pressure (≥15 GPa), and high-temperature (≥1400 °C) device is required for

growth of millimeter-sized wadsleyite single crystals. At pres-ent, only the Kawai-type multi-anvil apparatus fulfills all such requirements. Therefore, wadsleyite crystal growth experiments were performed using the 5000-ton Kawai-type high-pressure and high-temperature apparatus (USSA-5000) installed in the Institute for Study of the Earth’s Interior, Okayama University (Ito 2007).

experiMentAL MethoDSA Cr-doped MgO octahedron of 14 mm edge length was compressed using 32

mm tungsten carbide cubes with truncated edge lengths of 6 mm. A cross-section of the cell assembly is shown in Figure 2. Heating of the sample was achieved by using a LaCrO3 cylindrical heater embedded in a ZrO2 sleeve for thermal insulation. A platinum sample capsule with 2.5 mm length and 2.5 mm diameter surrounded by an electrically insulating MgO sleeve was placed into the heater. The center of the capsule was shifted along the axial direction of the heater for 0.5 mm to ensure that the sample capsule was placed in the monotonic thermal gradient field inside the heater (Fig. 2). The sample temperature was monitored using a W97Re3/W75Re25 thermocouple. No correction for the effect of pressure on the thermocouple emf was made. The axial thermal gradient across the sample was measured using two thermocouples, placed at the top and bottom of the sample (Fig. 2). In the first run, we used a simple straight heater, with thermal gradient of 100 °C/mm at 1400 °C in the sample region. In the successive Runs 2 and 3, we reduced the thermal gradient to 50 and 20 °C/mm, respectively, by adapting a stepped heater (Fig. 2). The pressure was determined from a calibration curve, which was determined by detecting univariant phase transitions at given temperatures against the press load (Fig. 3).

The composition of the starting materials is shown in Table 1. The silicate source was packed to fill 2/3 of the capsule. The remaining capsule volume was filled with the solvent. In case of the anhydrous systems, the starting materials, placed into a Pt capsule, were dried in a vacuum oven at 200 °C for 12 h. After drying the samples, the vacuum oven was filled with argon to minimize the amount of the moisture in the sample absorbed from the atmosphere when the oven was opened to the air. Then the sample was immediately sealed by arc welding. Samples with 2KHCO3-Mg(OH)2 solvent were welded under liquid nitrogen to prevent a water loss. The sample capsule was loaded in a heater so that the silicate source and solvent were placed in the hot and cool regions, respectively.

All the synthesis experiments were performed as follows. The samples were compressed at room temperature to desired loads, namely 720 or 860 tons, cor-responding to pressures of 15.5 or 16.5 GPa, respectively. Then the samples were heated to temperatures of 1400 or 1700 °C. The temperature was maintained within 5 °C of the desired value. Experiments were terminated by cutting off the electrical power of the heater, followed by slow decompression.

The run products were examined by a micro-focused X-ray diffractometer and an electron microprobe analyzer. The lattice integrity of the single crystals was examined using an optical microscope and Laue camera. The nucleation density of wadsleyite was determined by counting the number of crystals formed in the cool region on the platinum. Polarized and unpolarized infrared absorption spectra of the oriented sections of wadsleyite were measured using a Fourier transform infrared micro-spectrometer (FTIR-4200 and IMS-1, Shimadzu Co Ltd.) at Osaka City University and using a JEOL Diamond 20 FTIR microscope at Tohoku University.

Figure 2. Cross-section of high-pressure cell with stepped heater for wadsleyite crystal growth. The hot and cool zones in the sample are marked as HT and LT, respectively.

SHATSKIy ET AL.: WADSLEyITE CRySTAL GROWTH1132

Table 1. Experimental conditions and results of wadsleyite growth by the thermal gradient methodRun 1 2 3 4 5 6 7 8

a. Run conditionsPressure (GPa) 15.5 15.5 15.5 16.5 16.5 16.5 16.5 16.5Temperature (°C) 1400 1400 1400 1700 1700 1700 1700 1700Thermal gradient (C°/mm) 100 50 20 20 20 20 20 20Duration (min) 220 180 150 180 520 180 80 300

b. Starting materials and run productsSilicate source Fo Fo Fo Fo Fo(89) Fo(92)Solvent KMC-2H2O Na2CO3 K2CO3 KMC KMC 92KMC-8FeCl2

Solid Wd AnhB X Wd Wd WdQuenched solution Wd+KMC+H2O L1+L2 X+KMC+Pc Wd+KMC Wd+Fe-KMCNucleation density (cm–2) 9 × 103 240 25 – – 25 200 250

c. Wadsleyite single crystalsMax. size (mm) 0.2 × 0.6 0.5 1.7* 1.0 – – 1.4* 0.7 0.3 1.0Morphology Dissolution surface – – Orthorhombic Dissolution bipyramid + surface pinacoidFeatures of crystal shape elongated isometric dome-shaped – – isometric isometric isometricComposition Fo Fo Fo – – Fo Fo(89) Fo(94) Fo(92)H2O, wt ppm – 2580±209 1496±117 – – 68±4 235 156 84±17 (Paterson 1982) ±37 ±25(Libovitsky and – 3312±274 1977±154 – – 97±5 330 228 121±23 Rossman 1997) ±50 ±35Notes: Fo = Mg2SiO4; Mg number (molar forsterite content) of starting olivine is shown in parentheses. Wd = wadsleyite; AnhB = anhydrous phase B; X = K-phase X; KMC = K2Mg(CO3)2; Pc = periclase; L1 = silicate-rich glass; L2 = carbonate-rich quenched melt. * Total crystal size including portions with cracks and (or) inclusions.

The concentrations of hydroxyl groups were determined according to the methods of Paterson (1982) and of Libowitzky and Rossman (1997) with doubly polished thin sections of single-crystal wadsleyite placed on a KBr plate. The polarized IR absorption spectra with three crystallographic orientations and average unpolarized spectrum were used to calculate hydrogen contents. Measurements were carried out using a tungsten light source, a Ge-coated KBr beam-splitter and a high sen-sitivity, wide-band MCT detector. Several hundred scans were accumulated for each spectrum with 1–2 cm–1 resolution and 50–100 µm apertures. Background corrections of absorbance spectra were carried out by a linear fit of the baseline defined by the data points outside the OH-stretching region. The thickness of the thin sections was in the range from 40 to 150 µm.

experiMentAL reSuLtS

First, we tried to synthesize single crystals of Fe-free hydrous wadsleyite. Runs 1–3 were performed at 1400 °C and 15.5 GPa using Mg2SiO4 forsterite and 2KHCO3-Mg(OH)2 as the silicate source and solvent, respectively. Free water was observed after

opening the Pt capsule. A dome-shaped volume of wadsleyite in the cool region and a quenched melt segregated to the hot region were observed in the run products. The quenched melts were represented by dendritic aggregates of K2Mg(CO3)2 and wadsleyite. In Run 1, a straight heater with the thermal gradient of 100 °C/mm was used, as mentioned above. The nucleation density of wadsleyite exceeded 9 × 103 cm–2 at this thermal gradient (Table 1). The obtained crystals had elongated shape, >0.5 mm long and <0.2 mm wide. Crystals were spread from the top of the Pt capsule toward the melt (Fig. 4a). In Runs 2 and 3, where the stepped heater was used, the nucleation density was diminished to 240 and 25 cm–2, respectively. The maximum crystal size achieved was 0.5 and 1.7 mm, respectively (Figs. 4b, 4c, and 4d).

Second, we tried to synthesize single crystals of Fe-free anhydrous wadsleyite. In Runs 4–8, anhydrous carbonates were tried as solvents for crystal growth of wadsleyite. Experi-ments were carried out at 1700 °C and 16.5 GPa using the cell assembly with the stepped heater (Fig. 2). In Run 4, where Na carbonate was used as a solvent, the residual silicate was located at the cool side and consisted of a white aggregate of Mg14Si5O24 anhydrous phase B crystals with sizes smaller than 0.3 mm. The high-temperature side contained colorless glassy matrix with devitrified globules, up to 500 µm in diameter. This glass and globules represent presumably an immiscible silicate-rich and carbonate-rich liquid, respectively (Fig. 5). In Run 5, we examined K2CO3 as a wadsleyite solvent. In the cool region, we found yellow-green rhomboid-prism-shaped crystals with the largest dimensions of 0.5 × 0.5 × 0.3 mm. Crystals were identified as a K-phase X with space group P31m and chemi-cal formula K1.93Mg2.02Cr0.02Si2O7 (Matsuzaki et al. submitted). The quenched melt was a dendritic aggregate of K2Mg(CO3)2, K-phase X, and periclase. In Run 6, a mixture of K2CO3 (50 mol%) and MgCO3 (50 mol%) was used as a solvent. After 180 min in the cool region, we found a wadsleyite crystal with well-faceted orthorhombic bipyramidal shape with dimensions

Figure 3. Pressure calibration curve for the 14/6 cell assembly used in this study.

SHATSKIy ET AL.: WADSLEyITE CRySTAL GROWTH 1133

of 1.3 × 1.4 × 1.4 mm. This crystal contained numerous melt inclusions and cracks in the base, which are probably caused by heater power instability resulting in temperature fluctuations in the beginning of the run. However, the portion of crystal, about 0.7 mm in size, formed during final stage was transparent and free from inclusions and cracks (Fig. 6a).

Third, we tried to synthesize single crystals of Fe-bearing anhydrous wadsleyite. In Run 7, (Mg0.89Fe0.11)2SiO4 olivine and K2Mg(CO3)2 were used as a silicate source and solvent, respectively. In 80 min, about 40 vol% of silicate source re-mained at the hot end of the capsule as a black aggregate of (Mg0.89Fe0.11)2SiO4 wadsleyite. Transparent light-green crystals of (Mg0.96Fe0.04)2SiO4 wadsleyite with the maximum size of 0.3 mm

were found at the hot end of the platinum capsule (Fig. 6b). The quenched melt had a bluish-green color and consisted of K-Mg-Fe carbonate and wadsleyite dendrites. In Run 8, we employed 92K2Mg(CO3)2+8FeCl2 as a solvent and (Mg0.92Fe0.08)2SiO4 as a silicate source. In 300 min, the silicate source was completely exhausted. Several single crystals of wadsleyite, which were up to 1 mm in size, precipitated at the top end of the sample capsule. Large crystals were almost opaque and black (Fig. 6c). However, under the microscope, their thin debris was transparent and had dark-green color.

Polarized FTIR spectra of the wadsleyite crystals, which were synthesized using 2KHCO3-Mg(OH)2 solvent at 15.5 GPa and 1400 °C, were taken with the electric vector (E) parallel to each

Figure 4. Illustration of the effect of thermal gradient (TG) on wadsleyite crystal growth in the K2Mg(CO3)2-2H2O solvent at 15.5 GPa and 1400 °C. (a) Fan-shaped aggregate of wadsleyite (Wd) formed at the cool end of Pt capsule (TG = 100 °C/mm, Run 1). (b) Platinum capsule with wadsleyite crystals (Wd) in the cool region and quenched melt (L) in the hot region (TG = 50 °C/mm, Run 2). (c) Water-bearing wadsleyite single crystals (TG = 50 °C/mm, Run 2). (d) Section of wadsleyite single crystal (TG = 20 °C/mm, Run 3).

SHATSKIy ET AL.: WADSLEyITE CRySTAL GROWTH1134

crystallographic axis (Fig. 7). The spectra are characterized by three main groups: a pleochroic doublet at 3614 and 3581 cm–1, a less pleochroic and strong absorption triplet at 3360, 3326, and 3317 cm–1, and a very broad and pleochroic band centered around 3000 cm–1 (Fig. 7). The H2O content was estimated to be 2580 wt ppm using the Paterson (1982) calibration. The use of the Libowitzky and Rossman (1997) calibration yields 3312 wt ppm H2O. These values are comparable with those of wadsleyite synthesized in the Mg2SiO4-H2O system at 16 GPa and 1200 °C (Jacobsen et al. 2005). The examples of unpolarized spectra of H2O-poor wadsleyite are shown in Figure 8. Most spectra contain absorption bands at 3325 and 3618 cm–1. Minor bands at 3208 and 3480 cm–1 were observed only in some spectra. The H2O contents of wadsleyite synthesized from nominally anhydrous systems vary from 68 to 235 wt ppm using the Paterson (1982) calibration and from 97 to 330 wt ppm using the Libowitzky and Rossman (1997) calibration. Minimum H2O content of large single-crystal wadsleyite in our experiments was about 30 wt ppm in the run, which was not included in the present paper.

DiScuSSion

The thermal gradient method is frequently employed for syn-thesis of large single crystals at pressures of <7 GPa (Pal’yanov et al. 1997; Bockowski et al. 2002; Sumiya et al. 2005). The main problem with its application at higher pressures is the small size of a high-pressure chamber, which tends to generate too steep a thermal gradient in a sample. For example, the thermal gradient in the 14/6 (octahedral edge-length/truncation edge-length, in millimeters) assembly with a simple heater is about 100 °C/mm at 1400 °C. Such a large thermal gradient results in the formation of numerous nuclei (Shatskiy et al. 2007). This scenario was realized in Run 1 (Table 1). The ormation of a large number of

Figure 5. A fragment of immiscible silicate-rich (L1 = glass) and carbonate-rich (L2 = devitrified globule) liquids in the Na2CO2-Mg2SiO4 system (Run 4).

Figure 6. Anhydrous wadsleyite single crystals. (a) Mg2SiO4 (Run 6). (b) (Mg0.96Fe0.04)2SiO4 (Run 7). (c) (Mg0.92Fe0.08)2SiO4 (Run 8).

SHATSKIy ET AL.: WADSLEyITE CRySTAL GROWTH 1135

nuclei limits the space for growth of a large crystal. In this run, the maximum crystal size was only 0.2 × 0.2 × 0.6 mm (Fig. 4a). Based on these results, in the subsequent experiments we applied a modified heater shape, a “stepped heater,” in which the central part of the heater has larger thickness (Fig. 2). This heater shape reduces the axial thermal gradient through the sample to 20 C°/mm. This more uniform temperature field suppressed the nucleation density and permitted enlargement of the single-crystal sizes to 0.7–1.7 mm (Figs. 4d, 6a, and 6c).

Synthesis of water-bearing wadsleyite crystals was performed using K2Mg(CO3)2-2H2O as a solvent. The crystals synthesized at 15.5 GPa and 1400 °C contained 0.15–0.26 wt% H2O. In these experiments, the H2O/(H2O+Mg2SiO4)×100 weight ratio was 5.5±0.2. It should be noted that Demouchy et al. (2005) reported 0.93 wt% H2O in wadsleyite synthesized in the Mg2SiO4 + 5 wt% H2O system under the same P-T conditions. This value is 3.5–6.2 times higher than in our experiments (Runs 2 and 3). Note that the H2O/(H2O+Mg2SiO4) ratios were higher than the maximum H2O contents reported for wadsleyite, 3.33 wt% (Smyth 1994), in the both studies. In spite of the excess of water, the solvent contains 50 mol% of carbonate components, which lowers water activity in the melt. This explains the lower H2O contents of wadsleyite synthesized in Runs 2 and 3, relative to that reported by Demouchy et al. (2005). As was reported by Demouchy et al. (2005), temperature is one of the parameters that can control the hydrogen content of wadsleyite. However, the temperature range suitable for growth of large perfect single crystals is generally restricted due to technical problems. For example, volatiles tend to escape from the capsule at very high temperatures (e.g., Truckenbrodt and Johannes 1999). On the other hand, transport properties of the solvent become low at low temperatures due to decreasing viscosity, diffusion, convection rate, and silicate solubility. In this study, we showed that the H2O

content in wadsleyite can be controlled by the water/carbonate ratio in the solvent at fixed temperature.

Synthesis of anhydrous wadsleyite is difficult because this mineral tends to dissolve significant amounts of water, available in a system, up to the saturation limit at given physicochemical conditions. Here we applied water-free carbonates as a solvent to grow anhydrous wadsleyite. Although Na and K carbonates have excellent transport properties relative to silicates, they cannot be applied as a solvent for wadsleyite crystal growth for the follow-ing reasons. In the case of Na carbonate, Mg2SiO4 source dis-solves incongruently to produce anhydrous phase B (Mg14Si5O24), which totally replaces wadsleyite in the run products. Moreover, the texture of the molten part reveals the liquid immiscibility between carbonate and silicate in the Na2CO3-Mg2SiO4 system (Fig. 5). In the case of K carbonate, K-phase X forms instead of wadsleyite by the following carbonate-silicate reaction:

2Mg2SiO4+ 2K2CO3 = K2Mg2Si2O7(K-phase X) + K2Mg(CO3)2(Liquid) + MgO.

To avoid the above mentioned problems, K2Mg(CO3)2 was employed as a solvent. Use of this solvent increases the MgO content in the system, which changes the liquidus phase from K-phase X to wadsleyite (Fig. 6a).

Fe-bearing wadsleyite crystals with uniform (Mg0.92Fe0.08)2SiO4 composition were grown using a (Mg,Fe)2SiO4 silicate source and K2Mg(CO3)2-FeCl2 solvent with the same Mg/Fe atomic ratio of 23/2 (Fig. 6c). It is important to emphasize that in spite of the high K2CO3 content in solvents used, no appreciable amount of K, exceeding the EPMA detection limit, was measured in the wadsleyite crystals grown.

Magnesium silicate wadsleyite containing 1.6 wt% H2O

Figure 7. Representative polarized FTIR spectra of Mg2SiO4 wadsleyite grown in K2Mg(CO3)2-2H2O solvent at 15.5 GPa and 1400 °C.

Figure 8. Representative examples of unpolarized FTIR spectra of wadsleyite with low H2O contents, grown in K2Mg(CO3)2±FeCl2 solvent. The numbers on the right side above each spectrum show the H2O content in wadsleyite in wt ppm.

SHATSKIy ET AL.: WADSLEyITE CRySTAL GROWTH1136

or less has an orthorhombic structure with space group Imma, whereas wadsleyite with higher H2O contents has monoclinic symmetry with space group I2/m (Holl et al. 2008). Hence wadsleyite crystals obtained in the present study should have orthorhombic symmetry in accordance with the water content (see Table 1). As can be seen in Figure 6b, the crystals show a bipyramidal habit with minor pinacoid facets.

AcknowLeDgMentSWe greatly thank two anonymous referees for very constructive comments and

suggestions, and C. Oka for technical assistance. This research was supported by a Grant-in-Aid for Scientific Research to T.K. and D.y. from the Japan Society for the Promotion of Science and the COE-21 Program of the Institute for Study of the Earth’s Interior, Okayama University.

reFerenceS citeDAngel, R.J., Frost, D.J., Ross, N.L., and Hemley, R. (2001) Stabilities and equa-

tions of state of dense hydrous magnesium silicates. Physics of the Earth and Planetary Interior, 127, 181–196.

Bockowski, M., Grzegory, I., Krukowski, S., Lucznik, B., Romanowski, Z., Wroblewski, M., Borysiuk, J., Weyher, J., Hageman, P., and Porowski, S. (2002) Directional crystallization of GaN on high-pressure solution grown substrates by growth from solution and HVPE. Journal of Crystal Growth, 246, 194–206.

Dalton, J.A. and Presnall, D.C. (1998) The continuum of primary carbonatitic-kim-berlitic melt compositions in equilibrium with lherzolite: data from the system CaO-MgO-Al2O3-SiO2-CO2 at 6 GPa. Journal of Petrology, 39, 1953–1964.

Demouchy, S., Deloule, E., Frost, D.J., and Keppler, H. (2005) Pressure and temperature-dependence of water solubility in Fe-free wadsleyite. American Mineralogist, 90, 1084–1091.

Dobson, D.P., Jones, A.P., Rabe, R., Sekine, T., Kurita, K., Taniguchi, T., Kondo, T., Kato, T., Shimomura, O., and Urakawa, S. (1996) In situ measurement of viscosity and density of carbonate melts at high pressure. Earth and Planetary Science Letters, 143, 207–215.

Frost, D.J. and Dolejš, D. (2007) Experimental determination of the effect of H2O on the 410 km seismic discontinuity. Earth and Planetary Science Letters, 256, 182–195.

Ghosh, S., Ohtani, E., Litasov, K.D., and Terasaki, H. (2009) Solidus of carbonated peridotite from 10 to 20 GPa and origin of magnesiocarbonatite melt in the Earth’s deep mantle. Chemical Geology, 262, 17–28.

Green, D.H. and Falloon, T.J. (1998) Pyrolite: A ringwood concept and its current expression. In I. Jackson, Ed., The Earth’s Mantle: Composition, Structure, and Evolution, p. 311–380. Cambridge University Press, U.K.

Hammouda, T. and Laporte, D. (2000) Ultrafast mantle impregnation by carbonatite melts. Geology, 28, 283–285.

Holl, C.M., Smyth, J.R., Jacobsen, S.D., and Frost, D.J. (2008) Effects of hydration on the structure and compressibility of wadsleyite, β-(Mg2SiO4). American Mineralogist, 93, 598–607.

Inoue, T. (1994) Effect of water on melting phase relations and melt composition in the system Mg2SiO4-MgSiO3-H2O up to 15 GPa. Physics of the Earth and Planetary Interiors, 85, 237–263.

Ito, E. (2007) Theory and practice-multianvil cells and high-pressure experimental methods. In G. David Price, Ed., Treatise on Geophysics, Volume 2. Mineral

Physics, p. 197–230. Elsevier, Amsterdam.Iwamori, H. (2004) Phase relations of peridotites under H2O-saturated conditions

and ability of subducting plates for transportation of H2O. Earth and Planetary Science Letters, 227, 57–71.

Jacobsen, S.D., Demouchy, S., Frost, D.J., Boffa Ballaran, T., and Kun, J. (2005) A systematic study of OH in hydrous wadsleyite from polarized FTIR spectros-copy and single-crystal X-ray diffraction: Oxygen sites for hydrogen storage in Earth’s interior. American Mineralogist, 90, 61–70.

Katsura, T. and Ito, E. (1989) The system Mg2SiO4-Fe2SiO4 at high pressures and temperatures: Precise determination of stabilities of olivine, modified spinel, and spinel. Journal of Geophysical Research, 94, 15663–15670.

Katsura, T., yamada, H., Shinmei, T., Kubo, A., Ono, S., Kanzaki, M., yoneda, A., Walter, M.J., Urakawa, S., Ito, E., Funakoshi, K., and Utsumi, W. (2003) Post-spinel transition in Mg2SiO4 determined by in situ X-ray diffractometry. Physics of the Earth and Planetary Interiors, 136, 11–24.

Katsura, T., yamada, H., Kubo, A., Shinmei, T., Nishikawa, O., yoshino, T., Aizawa, y., Song, M., Walter, M.J., Ito, E., and Funakoshi, K. (2004) Olivine-wadsleyite transition in the system (Mg,Fe)2SiO4. Journal of Geophysical Research, 109, B02209, DOI: 10.1029/2003JB002438.

Knoche, R. and Luth, R.W. (1996) Density measurements on melts at high pres-sure using the sink/float method: Limitations and possibilities. Chemical Geology, 128, 229–243.

Liebske, C. (2005) Mantle-melting at high pressure—Experimental constrains on magma ocean differentiation. Ph.D. thesis, University of Bayreuth, Germany.

Maynard, J. (1996) Resonant ultrasound spectroscopy. Physics Today, 49, 26–31.

Nishihara, y., Shinmei, T., and Karato, S. (2006) Grain-growth kinetics in wads-leyite: Effects of chemical environment. Physics of the Earth and Planetary Interiors, 154, 30–43.

Pal’yanov, yu.N., Khokhryakov, A.F., Borzdov, yu.M., Sokol, A.G., Gusev, V.A., Rylov, G.M., and Sobolev, N.V. (1997) Growth conditions and real structure of diamond synthetic crystals. Russian Geology and Geophysics, 38, 920–945.

Paterson, M.S. (1982) The determination of hydroxyl by infrared absorption in quartz, silicate glasses and similar materials. Bulletin de Mineralogie, 105, 20–29.

Shatskiy, A., Fukui, H., Matsuzaki, T., Shinoda, K., yoneda, A., yamazaki, D., Ito, E., and Katsura, T. (2007) Growth of large (1 mm) MgSiO3 perovskite single crystals: A thermal gradient method at ultrahigh pressure. American Mineralogist, 92, 1744–1749.

Smyth, J.R. (1994) A crystallographic model for hydrous wadsleyite (β-Mg2SiO4): An ocean in the Earth’s interior? American Mineralogist, 79, 1021–1024.

Stalder, R., Ulmer, P., Thompson, A.B., and Gunther, D. (2001) High pressure fluids in the system MgO-SiO2-H2O under upper mantle conditions. Contributions to Mineralogy and Petrology, 140, 607–618.

Sumiya, H., Toda, N., and Satoh, S. (2005) Development of high-quality large-size synthetic diamond crystals. SEI Technical Review, 60, 10–15.

Truckenbrodt, J. and Johannes, W. (1999) H2O loss during piston-cylinder experi-ments. American Mineralogist, 84, 1333–1335.

Wentorf, R.H. (1971) Some study of diamond growth rate. The Journal of Physical Chemistry, 75, 1833–1837.

Manuscript received noveMber 16, 2008Manuscript accepted May 1, 2009Manuscript handled by Guoyin shen