Introduction

Meteorites, especially ordinary chondrites, preserve a record of impact events due to (possibly multiple) collisions among their parent asteroids1,2,3,4. As such, meteorites showing evidence for shock and melting (i.e., impact metamorphism) at various scales (from cm down to nm) provide an opportunity to explore collision events and to constrain parameters including the pressure (P) – temperature (T) – time (t) conditions of impact metamorphism and hence the relative velocity, size, and density of impactors and targets. Shock parameters can be inferred from the occurrence and textures of high-pressure (HP) minerals, many of them are polymorphs of common low-pressure minerals5,6,7,8,7d–f). Diffraction patterns also show streaks on ahrensite (110) and wadsleyite (010), suggesting stacking faults on these planes. Clinoenstatite (Fig. 7d) is the other silicate phase that crystallizes from the shock melt. Its composition (En83Fs14) is distinct from the sodic pyroxene in the Si-rich pool. From these observations we conclude that the mineralogy of the groundmass is fully consistent with the high-pressure minerals found as large grains. Hence, the groundmass material, likely to have crystallized from the melt at near-equilibrium conditions, leads to the same inferences about pressure and temperature conditions as the coarse grains despite the possibility of kinetic limitations on the achievement of equilibrium in the coarse grains.

Discussion

P-T-t constraints

This study presents the finding of four series of HP minerals (ringwoodite-ahrensite, wadsleyite, majorite-pyropess garnet, and sodic pyroxene) in Château-Renard; all these HP minerals are described from this meteorite for the first time. Although we do observe an usual pyroxene coexisting with Na- and Si-rich melt, it is not jadeite either by composition or by structure. Omphacite is characteristic of eclogite-facies metamorphism and is probably an indicator of elevated pressure, but any pressure constraints based on experimental thresholds for the formation of jadeite sensu stricto are not relevant. Here we discuss the use of these observations to constrain the peak pressure, pressure-temperature evolution, and shock duration experienced by this meteorite.

If any part of the melt veins reached peak temperatures above the liquidus of the matrix material and maintained that state long enough to reach complete melting38, then it would have evolved according to the liquidus relations of a cooling chondritic liquid. At any pressure above the invariant point in the MgO-SiO2 system where the ringwoodite breakdown reaction intersects the liquidus, this sequence would begin with crystallization of (Mg,Fe)O periclase. Hence the absence of periclase + (devitrified) bridgmanite or periclase + stishovite suggests an upper bound for the peak pressure of ~23–25 GPa40. As shown in MV4, the matrix of MVs in Château-Renard crystalizes integrown wadsleyite-anhrensite plus clinoenstatite (Fig. 7). With the absence of garnet, this assemblage represents the solidus phase relation at 14–17 GPa (Fig. 8a). The solid-state transformations of the clasts in the MVs indicates consistent shock pressure, as discussed below.

Figure 8
figure 8

P-T (a,b) and Time-Temperature (c) diagrams. (a) The dotted arrows represent two possible P-T evolution paths through the majorite-bearing and majorite-free stability fields. Phase diagram boundaries are from Agee et al.40. (b) Pressure vs. composition phase diagram for the system Di-Jd (at 1600 °C)46. (c) Time-Temperature diagram for back-transformation of HP ringwoodite to LP olivine for the case of Château-Renard. The time for complete solidification of 0.7 msec suggests a maximum time remaining at high-T, while a shock pulse with duration of ca. 1 sec, is the maximum to avoid back-transformation of all ringwoodite to olivine, while at the same time the T should not exceed that of ca. 1000 °C.

Full size image

In Château-Renard, we observe topotactic intergrowth of ringwoodite-ahrensite solid solution and wadsleyite from the shock melt. Many discussions of the shock pressures for olivine transformation have been based on an isochemical perspective reflecting the narrow transition intervals at a fixed terrestrial mantle-like composition close to Fa12. Viewed this way, peak pressures would need to have reached the stability range of ringwoodite, 17–23 GPa (Fig. 8a,b), whereas wadsleyite would imply a lower P range, 14–18 GPa, with only a small overlap. One explanation for their coexistence in the same meteorite or even in the same melt vein would be spatially or temporally variable pressure during the shock event, but we find the two phases intimately intergrown. Their coexistence might therefore be interpreted as a very specific constraint on pressure. However, this interpretation is not justified.

If the transformation took place from solid olivine precursors, then the transformation stress associated with the preferred solid-state transformation mechanism for the ringwoodite-wadsleyite transformation suggests a ~2 GPa range of coexistence41. We also note that the α-γ metastable reaction boundary lies in the middle of the wadsleyite stability field42 and transformation from olivine to ringwoodite exhibits a lower activation energy than olivine to wadsleyite. Ringwoodite may have nucleated first, in the stability field of wadsleyite, followed by topotactic growth of wadsleyite from ringwoodite nucleation sites.

A different approach considers the phase relations in the binary Mg2SiO4-Fe2SiO4 system, since the typical Fa content of L-chondrite olivines is nearly 25%. Compositional segregation during growth from a superliquidus state originally in the ringwoodite field would not explain the intergrowth: early crystallization of ringwoodite would enrich residual melt in Fe, moving it away from the stability field of wadsleyite. However, if a solid-state transformation occurred at high enough temperature to allow Fe-Mg interdiffusion, one could reconcile the observation of Fe-depleted forsterite olivine cores with moderately Fe-enriched wadsleyite, moderately Fe-enriched ringwoodite, and highly Fe-enriched wadsleyite-ahrensite intergrowths with a considerable range of pressures from 13–18 GPa (assuming equilibrium). The different HP polymorphs of olivine and the range of compsitions observed in the different melt veins could then be attributable to different cooling rates rather than to large variations in pressure.

Pressure indication from garnet depends significantly on the amount of Fe; the Château-Renard garnet has Fe/(Mg + Fe) in the range of 0.20–0.27, suggesting growth conditions of 17–20 GPa and 1800–2100 °C (Tomioka et al. 2016). The coexistence of majorite garnet (at the center of the MV) and wadsleyite (at the rim of the MV) indicates a thermal gradient present as the MV passed through this pressure range. Experimental observation shows that wadsleyite grains can grow at linear velocities up to <1 m s−1 and hence that the observed wadsleyite regions, 1–3 μm in size, require the MV to spend only a few microseconds in the wadsleyite stability field before quenching43.

Assuming that cooling and quenching of a MV is driven by thermal conduction across the boundary between the MV and its cooler matrix, we estimate a time for complete solidification of a typical Château-Renard MV of ~0.7 msec. The calculation assumes double-sided cooling of a 100 μm-wide slab of melt from super-liquidus temperatures (~2000 °C) while surrounded by cool matrix (~100 °C). Note, if peak temperature were below the liquidus, the cooling time would be slightly shorter. If the duration of the shock pulse were substantially shorter than the cooling time of the MV, then we would expect complete back-transformation to low-pressure minerals13. Specifically, the preservation of ringwoodite with about 40 mol% Fa component at the center of the MV suggests cooling below ~1000 °C while the rock was still at P > 13 GPa44,45 in order to prevent complete back-transformation of ringwoodite to olivine (Fig. 8c).

Raman spectra from the Na-Si-rich melt suggests the occurrence of a jadeite-like pyroxene. However, analytical TEM confirms that the pyroxene has the compositon and structure of omphacite. Relevant published phase diagrams46,47 show that addition of 50 mol% Di to jadeite lowers the low-P limit for a homogeneous cpx phase by about 0.5 GPa (from ~3 to ~2.5) and also lowers the upper-P limit for homogeneous cpx by about 5 GPa (from 21 GPa for the reaction Jadeite → Ca-ferrite to ~16 GPa for the reaction Clinopyroxene → Majorite + Ca-perovskite). Although preservation of the highest-P indicator minerals might be problematic, the presence of sodic clinopyroxene and the absence of Ca-ferrite, Ca-perovskite, or Ca-rich garnet suggests, at least locally, P ≤ 15.5 GPa in the pyroxene-bearing regions.

Given the diversity of mineral assemblages described within the single studied thin section, it is clear that Château-Renard records variable apparent pressure and temperature conditions. Possibly the different veins record different times along a common P-T path that they all experiened, depending largely on the local compositions, mineral kinetics, vein widths, and associated cooling rates. On the other hand, the presence of discrete veins directly proves heterogeneity of the temperature field, which is likely the result of collapse of spatially variable porosity during shock compression or slip along localized shear bands (despite some shape preferred orientation of large clasts parallel to the vein elongation, no convincing evidence of shear flow across the veins is observed). Shocking a heterogeneous medium also results in a heterogeneous pressure distribution. Although it is likely that pressure gradients on the order of GPa/mm would relax considerably after passage of the shock wave and before pressure release, it is hard to quantify the pressure differences that might persist over the potentially much shorter timescales involved in quenching the melt veins. Still, we lack a sound basis for asserting that a global peak P-T condition or global P-T path can be defined for the meteorite. Furthermore, as discussed below, the different veins may be recording altogether different shock events.

Potential ambiguities of high-pressure pyroxene phase identification

Before attempting to use the presence of a HP mineral to document certain P-T conditions in a shock-metamorphosed object, such a phase must be thoroughly characterized and its phase identification confirmed by the combination of a structure-sensitive analytical method (such as Raman Spectroscopy, EBSD, XRD, and/− or TEM) and co-located compositional microanalysis (e.g., by EPMA). Jadeite, for example, is a crucial mineral reported in a number of L and H chondrites. The majority of these reports8,30,37,38,48,49 have combined Raman spectra with near-albite compositional analyses that do not resolve intergrowths of the pyroxene phase and silica. At least one published study has foregone compositional analysis and relied on Raman spectra alone50 and in another case the pyroxene in the Tissint meteorite later described and named as tissintite51 was misidentified as jadeite on the basis of its Raman spectrum despite compositional analyses indicating 62–66 mol % anorthite component52. Although the omphacite we have discovered could not explain the near-albite compositions reported for most jadeite occurences in L6 chondrites, our data do reveal that other sodic pyroxenes, with stability fields different from those of pure jadeite, may present Raman bands indistinguishable from those of jadeite. A compositional analysis is clearly required to confirm a Raman identification of jadeite before the stability field of jadeite can be used as a pressure minimum for a meteorite.

HP mineral transformation mechanism; implications for duration of the shock pulse

Jadeite is commonly observed in ordinary chondrites as fine intergrowths of jadeite and silica with the bulk composition of albite8,11,37,38. Such jadeite presumably forms by solid-state decomposition of albite. On the other hand, the omphacite that we find in melt veins in Château-Renard combines components derived from more than one precursor phase — Na from plagioclase and Ca, Mg, and Fe from clinopyroxene. Hence is seems necessary that the omphacite grew from a melt whose formation digested both plagioclase and clinopyroxene. The melt then cooled enough to begin crystallizing while it remained at high enough pressure to stabilize omphacite.

HP polymorphs of olivine are found in two textural settings in the MV areas of Château-Renard. The TEM study of one MV shows ~10 nm-scale topotactic intergrowths of ahrensite and wadsleyite likely grown from shock melt. On the other hand, in several of the studied MV we find ringwoodite, ahrensite, or wadsleyite as incoherent μm-scale crystals collectively forming Fe-enriched rims enclosing large (5–50 μm) olivine grains. The latter texture, especially the presence of Fe-segregation towards wadsleyite or ringwoodite-ahrensite solid solutions, suggest a solid-state transformation mechanism with time for Fe-Mg interdiffusion. Similar textures were noted in L5 ordinary chondrite Dhofar 197053 in a study that also emphasized that rapid cooling is required to prevent back-transformation to olivine.

We seek to quantify the necessary duration of the shock pulse by combining the Avrami equation (a general formalism for solid-state phase transformation) with an Arrhenius-type temperature dependence of the transformation rate constant. The result of this calculation (see Supplementary material) is summarized in Fig. 8. We estimate that, even if the veins were completely melted, they would cool to ~1000 °C within 0.7 ms and then cool much more slowly from this temperature (roughly the average of the peak vein temperature and the matrix temperature) over the ensuing several seconds. At this temperature, back reaction to olivine, if the pressure was released, would require about one second. We conclude that the pressure was maintained in the stability field of ringwoodite for at least one second, enabling continued cooling of the melt veins and preservation of ringwoodite. A pressure wave of this duration requires an impactor at least meters in scale.

One single impact or many impacts?

The studied section contains a network of MVs of various widths that do, in some cases, intersect but do not reveal any crosscutting relations. Such a network of MVs might be simultaneously formed during a single impact event or might, even without evident cross-cutting, preserve evidence of a sequence of distinct impact eventsElectron Probe Microanalysis

Major element compositions of matrix and MV minerals were determined using a JEOL JXA8530F Field Emission EPMA instruments (FE-EPMA) equipped with five wavelength-dispersive spectrometers (WDS) and one energy-dispersive spectrometer (EDS) at both the NHMV and the Institut für Mineralogie, University of Münster, Germany. Mineral analyses were performed with an accelerating voltage of 15 kV. For minerals, a 20 nA focused beam current, 20 s counting time on peak position, and 10 s for each background were used. For glass analyses, a slightly defocused (5 μm diameter) beam, 5 nA probe current, and counting times of 10 s on-peak and 5 s on each background position were used. Natural mineral standards used were albite (Na, Si, Al), wollastonite (Ca), olivine (Mg), almandine (Fe), spessartine (Mn), orthoclase (K), rutile (Ti), chromite (Cr), and Ni-oxide (Ni) with ZAF matrix correction.

Transmission Electron Microscopy

We used a FEI Nova 600 Nanolab DualBeam focused ion beam (FIB) and scanning electron microscope (SEM) for the sample preparation and lift-out. The sample thinning was finalized with an 8 kV, 19 nA Ga-ion beam. The analytical transmission electron microscopy (ATEM) analysis was performed on FEI Tecnai TF20 with super-twin objective lens, operated at 200 kV. The EDS data were collected in TEM mode using a EDAX SiLi detector with 10 eV/channel and 51.2 µs process time, to achieve 500 cps signal and 20–50% deadtime. The FIB and TEM facilities are in the Kavli Nanoscience Institute at Caltech.

Micro-Raman Spectroscopy

Raman spectra for preliminary phase identification were conducted on the polished thin section using a dispersive confocal Raman microscope, Renishaw inVia Reflex at the National Hellenic Research Foundation. Analyses used a a 514 nm Ar-ion laser and a ×100 objective lens and spectra were collected in the Stokes region for Raman shifts from 200–1600 cm−1. Additional Raman analyses were performed at the Open University, Milton Keynes, United Kingdom, using a Horiba Jobin-Yvon LabRam HR Raman Microscope equipped with both 514 nm and 633 nm lasers. The laser beam was spread across ~1–2 μm spots at relative low incident power (ca. 5 mW) in order to avoid sample destruction. For each spot analysis on the Open University system, we averaged spectra over 5 consecutive 60 sec accumulation times. Gaussian-Lorentzian peak fitting (Spectragryph version 1.0.5) was used to remove background and estimate the peak centers. Collected spectra were compared with published data from RRUFF and the Handbook of Raman Spectra. The locations of each Raman spot analysis were recorded and co-located EPMA analytical points were collected in order to couple structural and compositional characterization at common spots.

Modeling strategies

Time for complete solidification of Melt Veins

The time required for complete solidification of a melt inside a tabular vein was estimated following the procedures of Turcotte & Schubert and Langenhorst & Poirier)59,60. In this model, it is assumed that the melt vein is surrounded by totally solid material at temperature T0, while the interior of the vein is totally melted at temperature Tm. A vein represented as a hot slab of a thickness 2w will cool and solidify in a characteristic time t s 60

$${t}_{s}=\frac{{w}^{2}}{4\cdot \kappa \cdot {\lambda }^{2}}$$
(1)

where κ is the thermal diffusivity and λ is a dimensionless coefficient that accounts for the boundary conditions and latent heat. λ is obtained by substitution of the following two boundary conditions at the moving front

$$\frac{d{y}_{m}}{dt}=-\,\lambda \cdot {(\frac{\kappa }{t})}^{0.5}\,$$
(2)
$$\,{(\frac{\partial T}{\partial y})}_{y={y}_{m}}=\frac{-({T}_{m}-{T}_{0})}{{(\pi \kappa t)}^{0.5}}\cdot \frac{{e}^{-{\lambda }^{2}}}{(1+{\rm{erf}}[\lambda ])},$$
(3)

into the conservation equation

$$\rho \cdot L\cdot \frac{d{y}_{m}}{dt}=\kappa \cdot {(\frac{\partial T}{\partial y})}_{y={y}_{m}}$$
(4)

resulting in

$$\frac{L\cdot \sqrt{\pi }}{{C}_{P}\cdot ({T}_{m}-{T}_{0})}=\frac{{e}^{-{\lambda }^{2}}}{\lambda \cdot (1+{\rm{erf}}[\lambda ])},$$
(5)

where L is the latent heat of crystallization, C p is the specific heat and erf is the error function. Also, when the vein solidifies, the temperature at the boundary with the surrounding matrix material will be constant and is given by

$${T}_{b}={T}_{0}+\frac{{T}_{m}-{T}_{0}}{1+{\rm{erf}}[\lambda ]}$$
(6)

The values we used for the modeling for Château-Renard case were: L = 320 kJ kg−1, C p  = 1.2 kJ K−1 kg−1, κ = 10−6 m2 sec−1, Tm = 2000 °C and T0 = 100 °C. These parameters yield λ = 0.93, Tb = 1148 °C, and 0.72 ms for a typical 100 μm width vein in Château-Renard.

Preservation of HP minerals

We estimated the time over which HP minerals might persist without back-transformation to their low-pressure equivalents. We combined the Avrami equation (which describes how solids transform from one phase to another at constant temperature) and an Arrhenius relationship for the transformation rate constant, resulting in

$$t\,={[-\frac{\mathrm{ln}(1-X)}{A\cdot exp(-\frac{E}{RT})}]}^{\frac{1}{n}}$$
(7)

where X is the volume fraction of the transformed phase, A is a frequency factor, E the activation energy for the polymorphic transformation, R the gas constant, n a constant determined by the dimensionality of the nucleation and growth processes (surface vs. volume), and T is absolute temperature. In our model we used the following numbers: X = 0.99, A = 2.44 × 1011 (from Arrhenius equation for Mg1.6Fe0.4SiO4), E = 324149.9 J/mol61,62, R = 8.3144598 J/K mol, n = 1.52 (value used by Sato et al.63), and T in the range given in Fig. 8c.

Assuming that the shock event that disrupted the L-chondrite parent body happened at 470 Ma, the preservation of ringwoodite, majorite and wadsleyite over this time requires that the temperature be maintained below the conversion boundary curves for this time period in t-T space. For both the olivine and pyroxene systems, the upper bound temperature corresponds to ca. 200–250 °C.

Data Availability

The datasets generated during and/or analysed during the current study are included in this published article (and its Supplementary Information files) but also are available from the corresponding author on reasonable request.