Shock-induced minerals in meteorite provide prospecting tools for mineral physics

Over the last 40 years, a combination of seismic data from Earth's interior and experimentation at high pressures on materials of composition similar to that in Earth's mantle has led to an increasingly accurate picture of what succession of phases must be present at increasing depths. During the same time period, it has been discovered that many of these phases are present in minute amounts in meteorites that, before their arrival on Earth, were shocked to high pressures and temperatures during collisions between the asteroids from which the meteorites were derived. In a recent issue of PNAS, Chen et al. (1) reported a previously uncharacterized mineral from such a shocked meteorite and, by using diamond anvil experiments and synchrotron x-ray analysis, derived the stability conditions of that mineral and another closely related one discovered earlier this year from the same meteorite. Both of these phases are high-pressure polymorphs of the mineral chromite FeCr₂O₄, which is a common (although minor) mineral in peridotites, rocks of the upper mantle that are abundantly exposed at the surface and also found as inclusions (xenoliths) in volcanic rocks. Some of this volcanism (e.g., kimberlite) also carries diamonds and hence must be derived from a minimum depth of ≈120 km. The authors suggest that the newly discovered polymorphs of chromite, if found in natural rocks on Earth, could serve as anchor points for estimation of the depths of origin of rocks transported from great depth. The two polymorphs place minimum pressures of 12–13 GPa (CaFe₂O₄ structure) and 20 GPa (CaTi₂O₄ structure). These pressures are equivalent in Earth to depths of ≈400 km and ≈680 km, respectively. However, is there any reason to believe that any rocks presently at the surface have ever seen such pressures? Many would answer that question in the negative, but recent developments in the study of inclusions in diamonds and so-called “ultra-high-pressure metamorphism” suggest that there is a reasonable possibility for the lower of these pressures and a small possibility for the higher one. Thus, the suggestion of Chen et al. potentially provides a powerful window not only into rocks plucked “instantly” from their deep place of residence by volcanism, but also it might provide an equally powerful window into the tectonics of continental collision, a topic of considerable current interest. Before addressing these possibilities, however, I will first briefly review the current context in which we understand very high-pressure natural rocks and then return to the potential offered by the discoveries of Chen et al.

Until the mid-1960s, when the concept of plate tectonics was developed, the events recorded in rocks exposed at Earth's surface were considered to be essentially decoupled from Earth's interior by isostasy (continental rocks are less dense than mantle rocks, hence they float at Earth's surface), except for periodic penetration to the surface by volcanism from below. Even though diamond-bearing kimberlites were already known, their volumetric abundance is small, and they were considered primarily curiosities. However, beginning in the early 1960s, a succession of observations has led to identification of another, seemingly unlikely, geologic environment of very high-pressure rocks. This environment is represented by rocks that mark the place of former subduction zones, especially in regions of continental collision.

The data provide a powerful window into the tectonics of continental collision.

Early in the 1960s, it was discovered that glaucophane, a blue mineral that is common in certain chaotically deformed rocks, is stable only at low temperatures and at pressures corresponding to several tens of kilometers deep in Earth (2, 3). This discovery implied that the rocks now containing this mineral must have been carried to depths in excess of 40 km at a geologically rapid rate (in order for them to arrive there at sufficiently low temperatures that glaucophane would be stable) and then must have been returned to the surface equally rapidly (in order for glaucophane not to have been converted to higher temperature minerals on the way back up). Plate tectonics had not yet been conceived at that time, and the dogma of isolation of surficial geology from mantle depths meant that there was no mechanism evident for such rapid burial and exhumation of rocks, leading to considerable resistance in the geological community to acceptance of the experimental data and their implications. Of course, when plate tectonics became the new paradigm in the late 1960s, glaucophane-bearing rocks (“blueschists”) were readily seen to be the metamorphosed product of subducted sediments that somehow had found their way back to the surface within ≈10 million years after their subduction.

The realization that continental materials could be subducted produced only a small change in geological viewpoint, however, probably because the depths needed to produce glaucophane were still on the order of the known thickness of continents. Thus, plate tectonics provided an explanation of the time frame of subduction and exhumation of blueschists, but the community was not yet ready to consider significant depth of subduction of continental material. However, this view was irrevocably changed by discovery in 1984 of coesite (a high-pressure polymorph of quartz implying ≈100 km depth) in metamorphosed continental rocks of the high Italian Alps (4) and of Western Norway (5). Similarly, in 1990, discovery of diamonds in metamorphosed sediments of Kazakhstan was published for the first time in the Western literature (6). All three of these occurrences are in zones of continental collision. Additional reports of diamonds and coesite in several additional collision terrains and the more recent discovery of indirect evidence of exhumation from minimum depths of up to 300 km and perhaps more followed in the later 1990s (e.g., refs. 7–10). During this same time period, it was discovered that the diamonds in kimberlites and related volcanic rocks sometimes carry inclusions that suggest much greater depths than the depths from which the kimberlites were erupted (<200 km), depths perhaps as great as the uppermost lower mantle, more than ≈660 km (11, 12). Credible evidence exists, therefore, for at least some diamonds to have grown at depths in excess of 660 km (27 GPa) and some rocks to have come up subduction zones carrying memory of depths in excess of 300 km (10 GPa). It is not an unreasonable hope, therefore, that rocks capable of exhibiting the high-pressure polymorphs of chromite may be transported to Earth's surface.

An obvious condition for application of the results of Chen et al. (1) is that the chemistry of the rocks must be such that chromite or one of its higher-pressure polymorphs must be stable either globally or, if conditions are right, at least locally within individual crystals of the major minerals. This is potentially a problem for mantle rocks because the chemical components of chromite are manifested as a spinel-structured mineral in peridotite, the greatly dominant rock type of the upper ≈70 km of the mantle, but, at depths between 70 km and ≈550 km, those components dissolve into other phases of peridotite, primarily garnet. Nevertheless, chromite is a commonly listed inclusion mineral in diamonds from kimberlites and related rocks; indeed, it is used as a tracer to prospect for diamond deposits. In one rare case, diamond-bearing chromitites have been reported in Tibet (13). Chromite also has been reported to precipitate from olivine (the most abundant mineral of the upper mantle), both in garnet peridotite xenoliths from kimberlite (14) and in similar rocks from continental collision terrains (7, 15, 16) (Fig. 1). Other indicators in these rocks suggest pressures of 8 and 10 GPa (250 and 300 km), respectively. Thus, there are many reported occurrences of diamond and chromite together, all of which are candidates for discovery of the polymorphs of chromite with either the Ca-ferrite or Ca-titanate structure.

Fig. 1.

Image of thin (presumed) chromite precipitates (arrows) in olivine from garnet peridotite, for which the density of ilmenite precipitates (longer black images, e.g., lower left center of image) implies a former solubility of TiO₂ in olivine in excess of 0.6 wt%, which, in turn, implies a depth of >10 GPa pressure (300 km). These precipitates, which have been interpreted as chromites, conceivably could be higher pressure polymorphs of chromite or carry evidence within them of previously having been such polymorphs. Only examination by transmission electron microscopy can potentially address this problem (for analogous example, see ref. 9). Image modified after ref. 16.

In summary, it is conceivable that the high-pressure polymorphs of chromite could be preserved in some natural mantle rocks. If such polymorphs can be found, it will be a major confirmation of the power of mineral physics to inform understanding of Earth's interior. Moreover, it will focus attention on the history of mantle rocks that precedes their abrupt transportation to the surface in magmas and/or, for rocks exhumed up subduction zones, the history preceding the pressure and temperature conditions recorded by freezing in of cation-exchange reactions. On the other hand, it remains possible that the high temperatures of kimberlitic magmas and/or the long time frame of exhumation up subduction zones provides sufficient time for down-pressure reactions to erase all evidence of the former presence of such high-pressure phases. As a consequence, not only examination of natural rocks is called for but also laboratory determination of the kinetics of the decompression reactions back to chromite and elucidation of any possible microstructural consequences of such reactions.