Natural (Mg,Fe)SiO3-Ilmenite and -Perovskite in the Tenham Meteorite

See allHide authors and affiliations

Science  22 Aug 1997:
Vol. 277, Issue 5329, pp. 1084-1086
DOI: 10.1126/science.277.5329.1084


The minerals (Mg,Fe)SiO3-ilmenite and -perovskite were identified in the shock-induced veins in the Tenham chondritic meteorite. Both phases are inferred to have transformed from pyroxene at high pressures and temperatures by shock metamorphism. Columnar-shaped ilmenite grains, one of two types of morphologies, have a topotaxial relationship with neighboring pyroxene grains, indicating shear transformation. Granular-shaped perovskite grains showed a diffraction pattern consistent with orthorhombic perovskite, but these grains were not stable under the electron beam irradiation and became amorphous. The higher iron concentration in both phases compared with those experimentally reported may suggest their metastable transition from enstatite because of shock compression.

Pyroxene with (Mg,Fe)SiO3 composition, one of the major minerals of Earth's crust and upper mantle, is known to transform into modified spinel(β) + stishovite, spinel(γ) + stishovite, garnet (at high temperature) or ilmenite (at low temperature), and perovskite structures in order of increasing pressure (1-5). Most of these high-pressure phases are considered to be the main constituent minerals of the mantle transition zone and lower mantle. Among these high-pressure polymorphs of (Mg,Fe)SiO3 pyroxene, however, only garnet has been found to occur naturally as majorite garnet (6-10). We report the natural occurrence of (Mg,Fe)SiO3-ilmenite and -perovskite in the Tenham chondritic meteorite.

The Tenham meteorite is an olivine–hypersthene-rich (L6) chondrite (11) composed of olivine, orthoenstatite, diopside, plagioclase which has been partly converted into glass (maskelynite) in a solid-state transformation at high pressure, Fe-Ni alloy, and troilite. The meteorite has a network of shock-induced veins (where the veins are <1 mm thick), and these veins enclose rounded fragments of host minerals in a black matrix. We examined three shock veins with an analytical transmission electron microscope (ATEM). ATEM specimens were taken from the optical thin section, milled by Ar ions into thin foils, and coated with carbon.

Olivine grains in the walls of the veins and in fragments are partially or totally transformed to blue-colored ringwoodite (γ phase) (12) or wadsleyite (β phase) (13). The black matrix, which is considered to have been quenched from melt (14), is dominated by Al-bearing majorite and a lesser amount of magnesiowüstite. These phases indicate the generation of high pressures and high temperatures by shock compression related to an impact event.

Within two shock veins, we identified ilmenite adjacent to clinoenstatite in fragments. The ilmenite formed aggregates, with each grain <1.4 μm in length. With d-spacings, angles, and systematic extinctions of electron diffraction patterns from several grains, all possible indexings by high-pressure polymorphs of pyroxene [majorite, ilmenite, perovskite, clinoenstatite (P21 /c), and orthoenstatite (Pbca)] were examined, and only ilmenite with space group (R 3̄) could explain all these diffraction patterns (Fig. 1). The estimated lattice parameters from several selected-area electron diffraction (SAED) patterns are a = 0.478 ± 0.005 nm andc = 1.36 ± 0.01 nm, assuming a hexagonal structure. These parameters agree with those extrapolated from the synthetic (Mg,Fe)SiO3-ilmenite within the SAED accuracy (15). Also, the c/a ratio of 2.85 is nearly equal to that of synthetic ilmenite (c/a = 2.87). Chemical analyses of these grains were carried out with an energy-dispersive analytical system attached to the ATEM (Table1), and we found that the ilmenite grains have similar compositions to the adjacent clinoenstatite grains.

Figure 1

SAED pattern of (Mg,Fe)SiO3-ilmenite in a shock vein in the Tenham meteorite. Distances of 012 and 102̄ in the hexagonal setting from 000 on the film correspond to the d-spacings of 0.349 ± 0.005 nm and 0.351 ± 0.005 nm, respectively. This diffraction pattern can be indexed only by the ilmenite structure with space group (R 3̄).

Table 1

Average chemical compositions of clinoenstatite, ilmenite, and perovskite. Compositions of ilmenite and perovskite were determined by ATEM and that of clinoenstatite by electron microprobe analysis. The number in parentheses for each phase is the number of analyses of the different grains.

View this table:

Ilmenite grains in the Tenham meteorite have two morphologies. One is granular-shaped (<0.4 μm in length) (Fig.2A), and the other is columnar (<1.4 μm in length) (Fig. 2B). Both types of grains do not show any microstructure except a low density of dislocations. Granular grains occur within the interstitial glassy phase. Assemblages of columnar grains have a brick-wall-like texture, and each grain is elongated along the same direction. Moreover, most of the columnar ilmenites have topotaxial relationships with adjacent clinoenstatite (Fig. 2C). The reciprocal c* axis of ilmenite is parallel to thea* axis of clinoenstatite. The ilmenite intergrown with clinoenstatite was observed in the other vein, and they also have the same topotaxial relationships. Clinoenstatite showed (100) twin lamellae, probably caused by transition from the original orthoenstatite (8).

Figure 2

Transmission electron micrographs of (Mg,Fe)SiO3-ilmenite (Ilm) showing two different morphologies: (A) Granular type. The large grain on the left side is clinoenstatite (Cen). (B) Columnar type. This is larger than a granular type and also adjacent to clinoenstatite. (C) The SAED pattern is shown for ilmenite superimposed on that of adjacent clinoenstatite from the circled area in (B). The direction of c* of ilmenite is almost parallel to that ofa* of clinoenstatite.

Perovskite was also identified in the same occurrence as granular ilmenite (Fig. 3A). The SAED pattern of orthorhombic perovskite along the [010] zone axis (Fig. 3B) is consistent only with perovskite (Pbnm) in itsd-spacings, angles, and systematic extinctions of the diffraction spots but could not be indexed by any other pyroxene polymorph. Estimated lattice parameters from the diffraction pattern are a = 0.474 ± 0.005 nm and c = 0.70 ± 0.01 nm, which is consistent with the values extrapolated from synthetic (Mg,Fe)SiO3-perovskite determined by x-ray diffraction (15). The lattice parameter b could not be estimated, because we could observe only thea*–c* plane of the reciprocal lattice. The other indication that this phase is perovskite is its characteristic instability in the electron beam; the grains easily converted into an amorphous phase during electron-beam irradiation. The chemical composition is almost the same as those of adjacent clinoenstatite and ilmenite grains (Table 1). The existence of perovskite in the shock vein means the generated pressure was at least ∼23 GPa (16). At such high pressures, orthoenstatite is not stable and is converted into clinoenstatite. If the generated temperature around the pyroxene grains was high enough, they would have further transformed into higher-pressure phases. This occurrence of the pyroxene-ilmenite or pyroxene-perovskite pair seems to reflect the heterogeneity of the generated temperature in fragments during the shock event.

Figure 3

Electron micrographs of (Mg,Fe)SiO3-perovskite (Pv) adjacent to clinoenstatite. (A) Transmission electron micrograph. (B) The SAED pattern is shown along the [010] zone axis. Distances of 200 and 002 from 000 on the film correspond to the d-spacings of 0.236 ± 0.005 nm and 0.352 ± 0.005 nm, respectively.

We found majorite with 4.8 mole percent (mol%) of Al2O3 component [in the (Mg,Fe)SiO3-Al2O3 join] distributed in the black matrix but not in contact with pyroxene in the fragments. The granular-shaped majorite grains were larger (up to 2 μm in length) than ilmenite or perovskite, and their interstices were filled with irregular-shaped magnesiowüstite. Magnesiowüstite is reported to occur also as inclusions in Al-bearing majorite in the Tenham chondrite (14). In contrast to this majorite, the ilmenite and perovskite do not have any Al. These differences in texture and Al content involved between majorite and ilmenite or perovskite may be caused by different reaction processes. The ilmenite and perovskite may have been transformed from almost Al-free pyroxene by a solid-state reaction during the shock event, whereas the majorite may have crystallized from the shock-induced melt, where most of the Al in the majorite may have been derived from molten plagioclase. We did not find any Al-free majorite as was previously reported for the Tenham meteorite (8, 14).

Ito and Yamada (15) reported that the solubility limit of FeSiO3 in Al-free ilmenite and perovskite is nearly 10 mol% at 1100°C, and Fei et al. (15) concluded that the solubility limit of FeSiO3 in perovskite is about 12 mol% at 26 GPa from 1150° to 1740°C. With the higher Fe content, Ito and Yamada (15) showed that single-phase ilmenite and perovskite are replaced by spinel + stishovite and perovskite + magnesiowüstite + stishovite, respectively. Compared with these solubility limits, the FeSiO3 content of 20 to 22 mol% of ilmenite and perovskite phases present in the Tenham meteorite is significantly higher. Nevertheless, we observed no decomposed phases around the ilmenite or perovskite grains. The peak pressure pulses may be on the order of 10–2 to 10–1 s in large-scale meteorite impacts (17), and the cooling rate will be very rapid. In such a short event, the decomposition of pyroxene into spinel + stishovite or perovskite + magnesiowüstite + stishovite would be difficult to complete in the solid state. Therefore, pyroxene probably transformed metastably into high-Fe ilmenite and perovskite without producing these other phases.

The shock event must also have affected the transformation mechanism. For the topotaxial relationship of (100)Cen // (0001)Ilm, where subscripts Cen and Ilm denote clinoenstatite and ilmenite, respectively, both planes correspond to the close-packed layers of oxygen for their respective phases (approximate cubic close-packed for clinoenstatite and hexagonal close-packed for ilmenite). Therefore, in the transition from clinoenstatite to ilmenite, the close-packed layers of oxygen are preserved, characteristic of shear transformation. Probably, the rapid transformation by the shock event favored the shear transformation mechanism for the clinoenstatite-ilmenite transition. This topotaxial relation indicates that this process may have proceeded by the displacement of the close-packed layers of oxygen on (100) plane for clinopyroxene. The intergrowth of ilmenite with clinopyroxene also suggests this mechanism. The granular ilmenite, which has no topotaxial relationship with clinoenstatite, would have formed by the nucleation and growth mechanism, probably under the slower cooling rates.

  • * To whom correspondence should be addressed. E-mail: tomioka{at}


Stay Connected to Science

Navigate This Article