High-pressure polymorphs of magnesian orthopyroxene from a shock vein in the Yamato-000047 lherzolitic shergottite

High-Pressure Phases in Martian Meteorites

Eighty-nine Martian meteorites of 74 shergottites, 8 nakhlites, 2 chassignites, 1 orthopyroxenite, and 4 unclassified have been known so far, and they have been mostly recovered from Antarctica and hot deserts in north Africa (Meteoritical Bulletin Database, http://tin.er.usgs.gov/meteor/metbull.ph). All Martian meteorites show characteristic textures derived from the intense impact shock on the Martian surface. The strong shock textures in shergottites are characterized of normal plagioclase glass or maskelynites transformed from plagioclase, and deformation features in olivines and pyroxenes (e.g., Ikeda 1994; McSween 1994). Shock melt pockets including glass have been observed in most shergottites (Ikeda 1994; Wang and Chen 2006; Walton and Herd 2007). Despite possible impact shocks on the Martian surface, the reports of high-pressure phases in Martian meteorites have been restricted to a small number of Martian meteorites: hexaluminosilicate and stishovite from shock melt pockets (Beck et al. 2004), post-stishovite (Sharp et al. 1999; El Goresy et al. 2000), stishovite, hollandites, akimotoite, and amorphosed perovskite (Langenhorst and Poirier 2000), and pyroxene glass (Bell et al. 1999) from shock melt veins.

Paired lherzolitic shergottites, Yamato (Y-) 000027, Y-000047, and Y-000097 (hereafter Yamato 00 lherzolitic shergottites), have been recovered from the bare ice field near the Yamato Mountains in Antarctica (Imae et al. 2002), and these three meteorites have been classified into lherzolitic shergottites (Kojima 2006; Misawa et al. 2006). These are paired based on the near sampling location and the similar mineralogical and petrological features (e.g., Mikouchi and Kurihara 2008).

We have first reported high-pressure polymorphs of three kinds of pyroxene glass and two kinds of akimotoites in shock veins only occurring in poikilitic area in Y-000047 in the 31st Symposium on Antarctic Meteorites (Imae and Ikeda 2007). In this study, we describe them in more detail and discuss their formation history.

MgSiO₃ Polymorphs

Five high-pressure polymorphs have been experimentally synthesized for MgSiO₃: high-pressure clinoenstatite (the space group of C2/c, Shinmei et al. 1999), majorite (the space group of I4₁/a, Angel et al. 1989), akimotoite (the space group of R), perovskite phase (the space group of Pbnm, e.g., Wang et al. 1992), and post-perovskite phase (the space group of Cmcm, Murakami et al. 2004). The equilibrium stability field of MgSiO₃ polymorphs has been thermodynamically obtained on the P–T phase diagram (Gasparik 1990a,b).

Majorite was first synthesized by Ringwood (1967), and discovered from the Coorara L6 chondrite (Smith and Mason 1970). Thereafter, majorites have been reported from the Suizhou L6 chondrite (Chen et al. 2004), the Tenham L6 chondrite (Xie and Sharp 2007; Xie et al. 2006a, 2006b) and from the Y-75267 H6 chondrite (Kimura et al. 2003). Akimotoite is MgSiO₃ with ilmenite structure, which was first synthesized by Kawai et al. (1974). The crystal structure was determined by Liu (1976). Much later, akimotoite has been identified from the nature in shock veins in the Tenham L6 ordinary chondrite (Tomioka and Fujino 1997), the Acfer 040 L6 chondrite (Sharp et al. 1997), the Y-75267 H6 chondrite (Kimura et al. 2003), the Umbarger L6 chondrite (Xie and Sharp 2004), the Y-791384 L6 chondrite (Ohtani et al. 2004), and the Sixiangkou L6 chondrite (Zhang et al. 2006). Perovskite phase was first synthesized by Liu (1974). The perovskite phase (Tomioka and Fujino 1997) was identified from the same meteorite, Tenham, with finding of akimotoite, and the vitrified perovskite phase was also identified from the same meteorite, Acfer 040, with the finding of akimotoite (Sharp et al. 1997), indicating that these perovskite phases coexisted with akimotoites. High-pressure clinoenstatite (C2/c) transforms into clinoenstatite (P2₁/c) at room condition (Demelza et al. 1994; Shinmei et al. 1999), and it is unquenchable phase. Remnant of high-pressure clinoenstatite has been identified.

As the Mg/Si atomic ratio of the cosmic abundance is nearly unity (Anders and Grevesse 1989), the five high-pressure polymorphs of MgSiO₃ (Gasparik 1990a, 1990b; Murakami et al. 2004) are possible main phases of rocky planet interiors and the stable phase among them depends on the size of planets and their depths. It is very important to study the MgSiO₃ polymorphs in the shock vein of meteorites, as the present shock vein including only high-pressure polymorphs of MgSiO₃ occurs only in the oikocrystic low-Ca clinopyroxene transformed from orthopyroxene in the poikilitic area.

Petrography of Two Main Lithologies and Shock Veins

The Yamato 00 lherzolitic shergottites consist of two main areas, poikilitic and non-poikilitic (e.g., Mikouchi and Kurihara 2008), previously known as lherzolitic shergottites (e.g., Ikeda 1994, 1997, 1998), although the polished thin section of Y-000047 consists only of the poikilitc area. Detailed petrographic descriptions for the Yamato 00 lherzolitic shergottites have been described (e.g., Mikouchi and Kurihara 2008). Here, we briefly summarize the textures and compositions, and propose the summary of the crystallization history from the parent magma.

Large oikocrystic low-Ca pyroxenes (Ca-poor clinopyroxene and pigeonite) of up to about 8 mm occupy most of the poikilitic area (Figs. 1a and 1b). The Ca-poor clinopyroxene transformed from orthopyroxene during the shock metamorphism. Oikocrystic low-Ca pyroxenes are magnesian with the compositional range of En_66–77Fs_21–23Wo_2–11, and include chadacrystic rounded olivines (Fa_26.5–32). Minor augites (Fe# = 21–24, En_47–53Fs_13–14Wo_33–40) also occur in the poikilitic area (Fig. 1c). Chromites in the poikilitic area are mainly enclosed in pyroxenes; the composition of chromites is Cr-richer and lower in Fe# than that of the non-poikilitic area. In the non-poikilitic area, low-Ca pyroxenes are much smaller (<1 mm) than those in the poikilitic area, and have the composition of En_58–69Fs_23–25Wo_8–17 (Fe# = 28–33). They coexist with augites (Fe# = 20–25, En_45–55Fs_13–18Wo_27–42), coarser grained olivines (Fa_30–35), and maskelynites (An_43–49Ab_40–54Or_1–3) in the non-poikilitic area. Some maskelynites (An_34–62Ab_37–62Or_1–4) in Y-000097 show hourglass structure, consisting of An-rich wide sector (An_52–56) and Ab-rich narrow sector (An_48–55), suggesting that the interstitial melt for the lherzolitic shergottite have supercooled to crystallize primary igneous plagioclase (Imae and Ikeda 2007). Chromites in the non-poikilitic area consist of Cr-rich core and Ti-rich rim both with lower Mg# (=11–20) than those (19–28) in the poikilitic area. Minor minerals are ilmenite, pyrrhotite, pentlandite, merrillite, Cl-apatite, and baddeleyite in the non-poikilitic area. Trapped melt compositions estimated from magmatic inclusions have been shown by Ikeda and Imae (2007).

The three lherzolitic shergottites commonly include shock veins, although the textures of shock veins are different among the three specimens (Mikouchi and Kurihara 2008). The vein in Y-000027 is much thicker than the other two shergottites, and is 1.8–2.2 mm in width and 5.3–5.4 mm in length. It mainly consists of euhedral magnesian olivine (Fa_∼18 and ≤ ∼7 μm in size), coarse-grained anhedral relict olivines (several hundreds μm), a coarse-grained anhedral relict low-Ca pyroxene (several hundreds μm, En_74.1Fs_21.5Wo_4.4), and interstitial glass. Euhedral olivines seem to have crystallized from shock-produced melt, suggesting that they crystallized at low pressure after the release of high-pressure. Thin shock veins, less than about 100 μm, and several small shock melt pockets in Y-000097 occur in the poikilitic area, and the interiors of the shock veins include tiny bubbles, suggesting that the pressure was released probably during the crystallization. Thin shock veins in Y-000047, less than about 60 μm in width, also occur only in the poikilitic area. The shock melting occurs only in the oikocrystic low-Ca pyroxene, and there are no bubbles like those in Y-000097.

Summary of Crystallization History from the Parent Magma

The crystallization sequence of minerals for the Yamato 00 lherzolitic shergottite is summarized below (e.g., Ikeda and Imae 2007; Imae and Ikeda 2007; Mikouchi and Kurihara 2008). As chromite and olivine are mainly trapped in magnesian low-Ca pyroxenes (orthopyroxene and pigeonite) in the poikilitic area, they may be liquidus phases. The poikilitic lithology suggests that magnesian olivine (Fa_26–30), pyroxenes of orthopyroxene (Fe# = 21–23.5) and magnesian pigeonite (Fe# = 23–28), and Cr-rich chromite (Mg# = 25–28) are cumulus phases, which crystallized from a parent magma and accumulated on the floor of a magma chamber or a sill. However, ferroan olivine (Fa_30–35), ferroan pigeonite (Fe# = 28–33.5), augite (Fe# = 20–25), plagioclase, phosphates, and sulfides crystallized from fractionated melts mostly to form the non-poikilitic lithology. Cumulus minerals of olivine, pigeonite, and chromite in contact with intercumulus melt overgrew to form rims with mineral compositions in the non-poikilitic lithology. During the overgrowth, the diffusion of Cr–Ti between cumulus and overgrown chromites and that of Fe–Mg between cumulus and overgrown pigeonites did not occur due to their slow diffusion rates. However, the diffusion of Fe–Mg in olivine and that in chromite between cumulus and overgrown phases may have taken with place with each other in some degree. On the other hand, plagioclase nucleated directly from the intercumulus melt, suggesting that a large degree of supercooling is necessary for the crystallization of plagioclase. The rapid crystallization condition of plagioclase in the non-poikilitic lithology may have formed the hourglass structure.

Textures and Electron-Probe Microanalyses of the Shock Vein

1-3 show the overall features of the shock vein. A shock vein of the Y-000047 lherzolitic shergottite is denoted by Main in Figs. 1a and 1b. The vein seems to contact with another vein (Minor in Fig. 1a) at a chadacrystic olivine grain in the upper part of the figure. The shock vein occurs only in Ca-poor clinopyroxene transformed from orthopyroxene (Figs. 1a and 1b), but never in pigeonite and augite (Fig. 1c). The chadacrystic olivine grain makes a dislocation of about 20 μm (Fig. 1d), suggesting that the relative movement for the slip to form the shock vein seems to be small.

Figure 2 shows the detailed textures observed in the shock vein. Figure 2a shows a close-up image of the shock vein, consisting mainly of glasses and high-pressure phases. Figure 2b shows the schematic illustration of Fig. 2a. A crack along the shock vein is distinct (1, 2). As the pyroxene glass crosses the crack indicated within a circle in Fig. 2c, the crack would have formed at the time of shock melt vein formation.

The vein rim shows angular or columnar texture coexisting with pyroxene glass (R-Ak in Fig. 2d). The mode of the columnar texture in the vein rim is the most abundant in the shock vein, occupying about 48 modal%. The vein rim continuously changes to Ca-poor clinopyroxene toward the host clinopyroxene (Fig. 2d). The columnar phase in the vein rim (Col in Fig. 1d and R-Ak in Fig. 2d) have Fe# = 21–23. The intermediate between the rim and center of the vein is glassy, and it is R-Gl (rim glass) and C-Gl (center glass) in Fig. 2b. The mode of rim glass occupies about 38 modal%, and that of center glass <14 modal%. Compositions of columnar phase (R-Ak in Fig. 2d) and glassy phases (C-Gl and R-Gl in Figs. 2a and 2b) are similar to each other (Table 1, Fig. 4), and their compositions of En_76.6–74.6Fs_20.3–22.3Wo_3.1–3.2 are identical with those of the Ca-poor clinopyroxene in the poikilitic area (Table 1, Fig. 4).

The several central portions in the main shock vein form framboidal glass aggregates of 2–3 μm in each size (Fr-Aggr in 1, 2, or Fr-Gl in Fig. 2b), and the other central portions show dendritic textures (C-Ak in 2, 3). Their modal abundances are commonly low (a few modal%). Framboidal aggregates (Fr-Aggr in 1, 2) in the vein center are the most magnesian with the composition of En_81.4–74.6Fs_15.5–22.3Wo_3.1–3.2, while dendrites in the vein center (C-Ak in 2, 3) are the most ferroan with the composition of En_74.6–72.7Fs_22.3–24.2Wo_3.1–3.2 among the high-pressure polymorphs of the pyroxenes in the shock veins. Except different Fe-Mg variation of framboidal aggregates and dendrites in the vein center from those of hostic oikocrystic clinopyroxene in the poikilitic area, minor element compositions of framboidal aggregates and dendrites have similar compositions with those of hostic oikocrystic clinopyroxene in the poikilitic area (Table 1, Fig. 4).

MicroRaman Spectroscopy of the Shock Vein

Polymorphs of magnesian low-Ca pyroxenes in shock veins as well as the hostic low-Ca pyroxenes in the poikilitic area are studied using microRaman spectroscopy (5, 6). Magnesian low-Ca pyroxenes in the host phase is a typical clinopyroxene (P2₁/c), as the wave numbers of 600–800 cm⁻¹ and 300–450 cm⁻¹ have commonly doublet bands, and the band of <300 cm⁻¹ is about 230 cm⁻¹ in Fig. 5 (Wang et al. 2001). The obtained space group is consistent with that obtained from the electron backscatter diffraction (EBSD) method (Mikouchi and Kurihara 2008). The host oikocrystic magnesian Ca-poor clinopyroxene used to be orthopyroxene, and the orthopyroxene transformed into clinopyroxene during the shock metamorphism (e.g., Stöffler et al. 1991).

The Raman spectra of columnar shaped crystals in the shock vein (Col in Fig. 1d and R-Ak in 2, 3) have 333, 440, 479, 610, 671, and 795 cm⁻¹ in the order (Fig. 6). The intensity of 660–670 cm⁻¹ in center glassy phase is stronger than that of rim glassy phase. Two kinds of interfiling glass occur in the vein rim (R-Gl in Fig. 2) and the vein center (C-Gl in Fig. 2).

The Raman spectra of the framboidal aggregates (Fr-Aggr in Fig. 2a) have peak of 301, 528, 661, and 951 cm⁻¹ (Fig. 5). The Raman spectra of the interior two glassy phases have commonly two broad peaks in the similar wavelength with the similar width: 663 and 976 cm⁻¹ for the rim glass and 668 and 960 cm⁻¹ for the center glass. The broad band of about 960–980 cm⁻¹ commonly ranges from about 800 to about 1000 cm⁻¹. The Raman shift of the 951 cm⁻¹ broad band in the framboidal aggregates is very similar to those of two glassy phases (Fig. 5). Dendritic fine-grained crystals in the vein center (C-Ak in 2, 3) have Raman shift of 485, 619, 664, 795, and 943 cm⁻¹ in the order (Fig. 6).

Phase Identifications in the Shock Vein of the Y-000047 Lherzolitic Shergottite

Formation of the Shock Vein and Its Crystallization

Akimotoites

Two processes for crystallization of akimotoites have been considered (Zhang et al. 2006). One is solid-state phase transition observed from pyroxene in Tenham (Tomioka and Fujino 1997), Sixiangkou (Zhang et al. 2006), and Y-791384 L6 chondrites (Ohtani et al. 2004), and their akimotoites commonly have Fe# = 21. The solid-state phase transition model from high-pressure clinopyroxene to akimotoite has been also modeled based on the shear stress from the crystallographic consideration (Tomioka 2007). Another is the metastable crystallization of akimotoites from shock-produced melt in Acfer 040 (Fe# = 5.5, Sharp et al. 1997), Tenham (Fe# = 12, Xie et al. 2006b), and Umbarger (Fe# = 11, Xie and Sharp 2004) L6 chondrites. These akimotoites from the above references have been commonly found from equilibrated L6 chondrites, suggesting that the composition of host phase are nearly identical. As the Fe# of low-Ca pyroxenes for L6 chondrites is 19–22, the solid-state phase transition from low-Ca pyroxene to akimotoite preserved the original composition, although the shock-produced melt crystallized more magnesian and variable Fe#s than that of host phase low-Ca pyroxene.

Texturally, columnar akimotoites in the vein rim in the shock vein of Y-000047 seem to replace parts of the host Ca-poor clinopyroxene (Fig. 2d). The solid-state phase transition is classified into two types based on its transformation mechanism, displacive phase transition and reconstructive phase transition. The displacive transition is rapid process, and the reconstructive transition is sluggish process (e.g., Putnis 1992). The mechanism of the phase transition of orthopyroxene to Ca-poor clinopyroxene is classified into displacive transition (e.g., Putnis 1992). We suggest that the rim akimotoite did not form from solid-state phase transition of orthopyroxene during the increase of pressure and temperature, but it transformed from the Ca-poor clinopyroxene already inverted from orthopyroxene under the sobsolidus condition near the metastable liquidus temperature of pyroxene during the cooling of the shock vein (Fig. 7).

Generally, akimotoite is only subsolidus phase among MgSiO₃ high-pressure polymorphs. The other two sub-solidus regions are known to be mixtures of wadsleyite (β-Mg₂SiO₄) and stishovite at lower pressure, and mixtures of ringwoodite (γ-Mg₂SiO₄) and stishovite at lower temperature. The liquidus phases with MgSiO₃ composition are known to be perovskite phase at higher pressure, and majorite at higher temperature (e.g., Gasparik 1990a,b). In the Y-000047 lherzolitic shergottite, Fe#s of center akimotoite consisting of fine-grained dendritic crystals are slightly more ferroan than those of the host oikocrystic low-Ca pyroxenes (Table 1; Fig. 4). Therefore, the crystallization of dendritic akimotoite rapidly occurred from super cooled shock fractionated melt under the metastable condition.

Estimation of Maximum Temperature and Pressure, and Its Cooling History

The crystallization sequence is interpreted based on the high-pressure phase diagram of MgSiO₃ (Gasparik 1990a, 1990b) with modified phase boundaries of majorite-akimotoite-provskite (Hirose et al. 2001) and the compositions of these phases (Fig. 7). The maximum shock pressure for shergottites has been estimated to be 30–43 GPa (Stöffler et al. 1986). Fritz et al. (2005) obtained the peak shock pressure for the Y-793605 lherzolitic shergottite to be 40–45 GPa based on the micro-Raman spectroscopy of maskelynite. When a rock experiences an impact shock, the maximum shock pressure has been estimated to be nearly homogeneous inside of the whole rock due to the propagations and reflections of shock waves (e.g., Sharp and DeCarli 2006), except shock melt pockets which melted by local concentration of shock wave. Therefore, the peak shock pressure of maskelynite in non-poikilitic area in the Yamoto 00 lherzolitic shergottites may be nearly identical with that of the shock vein in Y-000047. The mineralogy and petrology of Y-793605 is similar to those of the Yamoto 00 lherzolitic shergottites based on our observation and Mikouchi and Kurihara (2008). The shock degree of Y-793605 is also similar to that of Y-000047, although the degree of undulatory extinction of olivines in Y-793605 is slightly higher than that of the Yamato 00 lherzolitic shergottites. Then the shock pressure would be approximately 40 GPa for the Y-000047 lherzolitic shergottite. The estimation of the peak shock pressure is based on the shock recovery experiments. Sharp and DeCarli (2006) pointed out that the shock pressure in nature based on experiments has been overestimated due to significantly longer shock pulse (several seconds) and more sluggish cooling time (up to 1 s) in nature than those in shock recovery experiments (∼10 μs shock pulse), and they roughly estimated that the peak shock pressure is much lower (∼up to half) than those expected from experiments. They also showed that the equilibrium phase diagram based on static high-pressure experiments rather than the results from shock recovery experiments is available for the understanding of the P–T history of shock vein in meteorites. The equilibrium phase diagram shows that more than 23 GPa is necessary for the formation of perovskite phase (Fig. 7). Taking this discussion into consideration, we conclude that the shock pressure of 23–40 GPa is possible shock pressure range for the Y-000047 lherzolitic shergottite. In Fig. 7, we showed the median with the possible errors.

In situ experiments of majorite–akimotoite–perovskite phase transition (Hirose et al. 2001) show that solid-state phase transitions occur among the three minerals. However, the phase transition into perovskite phase should be classified into reconstructive transition, which is sluggish process. Therefore, only clinoenstatite may have transformed from the host orthopyroxene during this transient state. This may bring an assumption that the host orthopyroxene in the shock vein in the poikilitic area melted at the metastable liquidus boundary of pyroxene indicated as T^Oen_liq in Fig. 7, and the solidification of the shock vein occurred after the peak shock pressure and temperatures (Fig. 8).

The duration of the subsequent cooling for the solidification, which is defined as the quench time after the maximum shock temperature, is estimated to be much longer (∼0.05 s) in the vein center than those (<0.01 s) in the vein edge for the Tenham L6 chondrite (e.g., Sharp and DeCarli 2006). The cooling rate in the vein center may be slower (∼10⁴ °C s⁻¹) than that (∼10⁵ °C s⁻¹) in the rim. As the shock vein in the Y-000047 lherzolitic shergottite is much thinner, the cooling may be much more rapid (<0.05 s) as be illustrated in Fig. 8. Additionally, it has been assumed that the maximum shock temperature in the vein rim was lower than that in the vein center (e.g., Xie et al. 2006a, 2006b). This suggests that the vein edge did not melt, indicated as A in Fig. 7 and the vein center melted, indicated as B–D in Fig. 7. Then the maximum pressure and temperature in the vein center are estimated to be 23–40 GPa and more than 3000 °C, respectively (Fig. 7). During the cooling time of <0.05 (s), the peak shock pressure of 23–40 GPa was kept, suggesting that the time scale of the pressure pulse is sufficiently longer than the cooling time scale as there is no evidence of the pressure release during the crystallization in the shock vein (e.g., Sharp and DeCarli 2006).

The formation sequences of the constituent phases in the shock vein are summarized from the texture and phase relationship, and schematically shown in Fig. 8: first, an impact shock has produced both a melt vein and ambient Ca-poor clinopyroxene from orthopyroxene in the poikilitic area. Second, the akimotoite at the vein rim metastably formed under the subsolidus condition (A in Fig. 8). Third pyroxene glass solidified in the vein rim (B). Fourth pyroxene glass in the vein rim (C). Finally, framdoidal perovskite phase (D) and metastably dendritic akimotoite (D) crystallized by the fractional crystallization from residual melt in the vein center. The formation for the two pyroxene glasses in the vein rim and vein center were nearly contemporaneous, and the crystallization timing of dendritic akimotoite and provskite-phase were also contemporaneous. The liquidus temperature of perovskite phase seems to be approximately 2800 °C, and the solidification temperature of akimotoites and glasses is estimated to be much lower according to the equilibrium phase diagram in Fig. 7. The crystallization temperature of perovskite phase is much lower than the equilibrium liquidus temperature of perovskite phase (∼2800 °C) due to the supercooling. And the subsolidus-cooling rate after vitrification was different; glass in the vein center was slower than that in the vein rim. The identical composition of rim akimotoite with the host Ca-poor clinopyroxene suggests that orthopyroxene with the (Mg_0.8,Fe_0.2)SiO₃ composition congruently melted under the metastable condition.

Thin shock veins tend to occur in the poikilitic area as can be seen in Y-000047 and Y-000097 (Fig. 1). The reason why the thin shock veins occur only in poikilitic area but not in non-poikilitic area may be attributed to the different shock responses among the two areas: the mafic orthopyroxene in the poikilitic area is brittle and the feldspathic non-poikilitic area is ductile.

Importance of the Simple Shock Vein from the Martian Meteorite

Shock melt vein in chondrites which occur in equilibrated ordinary chondrites, much in L6 chondrites (e.g., Sharp and DeCarli 2006), is the chondritic melt producing a lot of high-pressure polymorphs of olivine, low-Ca pyroxene, high-Ca pyroxene, and plagioclase from the melt. High-pressure phases have been found from the shock veins in pyroxenes and plagioclase of the Zagami basaltic shergottite (Langenhorst and Poirier 2000). Compositions of akimotoites and vitrified perovskites from Y-000047 are much more magnesian than those from Zagami (Fe# = 32 for akimotoites and 41–73 for vitrified perovskites). Most of high-pressure polymorphs of pyroxenes crystallized from shock melt produced by the melting of the mixtures of pyroxene and other minerals such as olivine and plagioclase. On the other hand, the fact that the shock vein in Y-000047 only produced from oikocrystic Ca-poor pyroxene should be quite unique, making simply magnesian Ca-poor pyroxene melt. The simple shock textures in Y-000047 would potentially play an important role in the understanding of the formation of the high-pressure polymorphs of MgSiO₃.