Zircon Microstructures in Large, Deeply Eroded Impact Structures and Terrestrial Seismites

Abstract

Extraterrestrial cratering was a critical component in the evolution of the early Earth, but discovery of very ancient and deeply exhumed impact structures remains elusive, as identification tools are in short supply. The recognition of such structures is hindered by post-cratering geological processes, whereby impact-induced features common in younger, upper-crustal craters may be lost. In this study, we make a detailed analysis of planar microstructures in zircons from four large, confirmed impact structures (Manicouagan, Rochechouart, Sudbury, Vredefort) preserved at different crustal levels, from two previously described non-impact seismites in SW Norway and northern Italy, and from the 3.0 Ga Maniitsoq structure in West Greenland. A total of ~3400 zircon grains were studied using exterior and interior scanning and transmission electron microscopy. We show for the first time that shocked zircons contain two successive, principally different types of planar microstructures, only one of which is diagnostic of impact. Closely spaced, contiguous microplanes (CP) are formed first, presumably by the shock wave. In Manicouagan and Rochechouart zircons the exterior CPs have ultrathin interior counterparts of straight dislocation arrays, as identified in Manicouagan zircon using transmission electron microscopy. They have the same close spacing and orientations as the exterior CPs and are abundantly decorated with tiny pores down to less than 50 nm across. These interior CPs are identical to shock-induced decorated, partly annealed amorphous planar microstructures in quartz (planar deformation features, PDFs) and are interpreted as such. The second type is open planar fractures (PF). They are widely and irregularly spaced and texturally younger than the CPs. They re-use and displace the CP orientations, which they cut in stepwise fashion. We interpret these PFs as formed by impact-induced seismic shaking in the wake of the shock wave. We confirm two previous reports of isolated planar fractures in zircons from non-impact seismites, showing that PFs per se are not impact-diagnostic. There are no CPs in any of these zircons. Zircons from different parts of the Maniitsoq structure contain CPs in various states of preservation besides PFs, corroborating that this very large and very deeply exhumed structure resulted from an extraterrestrial impact.

INTRODUCTION

Extraterrestrial impacts shaped and modified the early Solar System, and all of its rocky bodies preserve records of past impacting. Giant impacts controlled the evolution of the earliest crust of the Earth and maybe its earliest life (Glikson, 2014). An increasing number of terrestrial impact structures has been identified in recent years, but most are small, young and little eroded (Kenkmann, 2021; Osinski et al., 2022) despite the fact that the size and rate of impacts have diminished through time. The oldest confirmed impact structures on Earth are all of Palaeoproterozoic age and include the two largest confirmed impact structures, Vredefort in South Africa (d = 275 km, 2.02 Ga) and Sudbury in Canada (d = 200 km, 1.85 Ga); The quoted diameters are the apparent ones as listed in Kenkmann (2021). Yarrabubba in Australia (d = 30 km, 2.23 Ga) is the oldest (MacDonald et al. (2003) suggested an original outer diameter of 30–70 km for this structure). The oldest proposed but unconfirmed terrestrial impact structure is the deeply exhumed, 3.0 Ga Maniitsoq structure in West Greenland with an apparent present diameter of ~100 km but a much larger original diameter prior to exhumation (Garde et al., 2012). Spherule beds in Western Australia and South Africa up to 3.48 Ga in age (Lowe et al., 2003) provide evidence of still older, giant impacts, but none of the craters that sourced the spherule beds have been found. Discovery and confirmation of very deeply exhumed (and hence very large) impact structures preserved in the deep crust is difficult, because common and easily applicable tools of identification, such as shock lamellae (planar deformation features, PDFs) in quartz, are unlikely to have survived in a clearly recognisable state in the hot and ductile geological environments of the lower crust.

Shock-induced zircon microstructures might constitute an appropriate tool to identify the missing ancient, very large, but geologically reworked and deeply exhumed impact structures. This requires insight into the formation of shock-induced microstructures in zircon, if and how such microstructures might be affected annealing processes, as well as a distinction from microstructures formed at lower strain rates unrelated to impacting, e.g. by intense seismic shaking along major crustal boundary zones.

An old age alone is probably not a decisive factor in annealing of shock-induced zircon microstructures, despite an increasing U–Th radiation damage with time. For instance, at Yarrabubba granular microstructures in zircon indicative of shock-induced solid-state transformation from zircon to its high-P polymorph reidite and back have been preserved (Erickson et al., 2020), but this impact structure is neither deeply exhumed nor geologically reworked. Residence in active crustal environments is prone to cause recrystallisation, such as in the geologically reworked Sudbury impact structure, and may be more detrimental to shock-induced zircon microstructures (e.g. Thomson et al., 2014).

Exterior planar microstructures in zircon were first recognised by optical and backscattered electron scanning microscopy (SEM-BSE) from Sudbury and interpreted as impact-induced by Krogh et al. (1984), at a time when the origin of the Sudbury structure remained unresolved. They described them simply as sets of crystallographically oriented fracture planes with offsets and surficial pitting at their intersections. Bohor et al. (1993) showed further images from Sudbury and the Manicouagan impact structure, likewise in Canada (215 Ma, Jaret et al., 2018). They now labelled the planar microstructures ‘multiple intersecting sets of PDF’ analogous with PDFs in quartz. They also observed a progressive transition of exterior zircon microstructures from different sources ‘… from grains showing no shock features, to those showing a continuous gradation from PDF alone, through PDF combined with granular texture, to a well-developed granular texture and even incipient melting phenomena’. Soon after, Krogh et al. (1996) made the authoritative statement that the Sudbury zircons ‘… display multiple sets of parallel planar deformation features (PDF) of a type we had never seen before after more than two decades of studying zircons from a wide variety of geological environments’; at that time zircon dating (by bulk thermal ionization mass spectrometry) required exterior inspection and manual sorting of numerous individual zircon grains. Similar exterior planar zircon microstructures were then also reported from Vredefort (Kamo et al., 1996).

Bohor et al. (1993) identified planar microstructures, exterior granular texture and incipient melting in shocked zircons. Subsequent studies, greatly aided by increasing availability electron backscatter diffraction (EBSD) analysis, have documented many additional impact-diagnostic features, which are characteristic of different shock regimes. The additional features include transformation to reidite (e.g. Wittmann et al., 2006; Plan et al., 2021), evidence of presumed former reidite in granular neoblastic (FRIGN) zircon (Cavosie et al., 2018a), shock-induced, lamellar {112} microtwins (e.g. Cox et al., 2018) and thermally induced dissociation to ZrO₂ + SiO₂. (e.g. El Goresy, 1965; Timms et al., 2017). Thus, there is a whole range of different, partially overlapping impact-diagnostic microstructures in zircon, which can furthermore be used to unravel the individual pressure–temperature histories of shocked, zircon-bearing samples, as shown by (Timms et al., 2017). Reidite, FRIGN zircon and dissociation have so far only been reported from few, well-preserved impact structures in the upper crust, and preservation of reidite requires rapid cooling (e.g. Zhao et al., 2021). Diagnostic {112} microtwins are common in other impact structures where they may occur together with exterior planar microstructures, e.g. at Vredefort, Sudbury and Chicxulub (Moser et al., 2011; Erickson et al., 2013a, 2013b; Wittmann et al., 2021). These twins constitute an excellent impact identification tool for instance where no exterior zircon surfaces can be examined (such as zircon fragments in meteorites, Cox et al., 2022). The above-mentioned pressure–temperature study by Timms et al. (2017) used observations from three well-preserved, upper-crustal impact structures but did not deal with zircons in deeper root zones. Exterior planar microstructures and their pressure–temperature relations to the other shock-induced features were not addressed. Other studies have shown that {112} microtwins are less frequent than exterior planar microstructures in zircon populations where both types of microstructures occur (e.g. Erickson et al., 2013a, Thomson et al., 2014 and below), but there are no descriptions of such lamellar twins from zircons where exterior crystal faces have been examined but are devoid of closely spaced planar microstructures. The twins have the advantage of circumventing a potential ambiguity concerning the diagnostic validity of the latter planar microstructures. This problem arose when Austrheim & Corfu (2009) described zircons with occasional interior planar fractures from a Caledonian seismite in Norway unrelated to impacting, and they also mentioned observations of exterior planar fractures. Rare, widely spaced interior planar fractures were also reported from zircons in a seismite in the Ivrea–Verbano zone, northern Italy (Kovaleva et al., 2015). Nevertheless, examination of exterior surfaces has remained a fast and convenient way of identifying shocked zircon (e.g. Montalvo et al., 2017).

In this study, we examine exterior and interior planar microstructures and twins in numerous zircon grains from four large impact structures, using backscattered electron (BSE), cathode luminescence (CL) and electron backscatter diffraction (EBSD) scanning electron microscopy, as well as transmission electron microscopy (TEM): the young, upper-crustal, well-preserved impact structures of Manicouagan in Canada and Rochechouart in France (207 Ma, Cohen et al., 2017) and the much older, geologically reworked Sudbury and the more deeply exhumed Vredefort impact structures. Our descriptions of shocked Manicouagan zircons are the first apart from the historic image in Bohor et al. (1993), and our new observations from Sudbury and Vredefort refine previous descriptions referred to above and in the discussion.

We also re-examine zircons from the Caledonian seismite in SW Norway originally described by Austrheim & Corfu (2009), as this singular reference to non-impact exterior planar microstructures in zircons obviously calls for closer examination. We also examine zircons from an Alpine seismite in northern Italy, the only other locality from where occasional, non-impact interior planar fractures have been described (along with planar deformation bands, Kovaleva et al., 2015).

We finally describe recently discovered planar microstructures in zircons from several parts of the Maniitsoq structure in Greenland and compare them with those from the six other sites described here.

PREVIOUS NOMENCLATURE

Previous descriptions of planar microstructures in zircon are compromised by different use of similar terms by different authors and application of different terms for the same features. This reflects increasing insight, new analytical methods, different study materials, different interpretations as well as lack of consistency by some authors. A short summary with a few examples of the variable usage follows here.

Planar deformation features (PDF), planar fractures (PF), planar microstructures (PM) and planar deformation bands (PDB)

Bohor et al. (1993) originally used the term PDF in a general sense for undifferentiated exterior planar microstructures in zircon commonly with offsets; as mentioned in the introduction they simply borrowed the PDF label from interior planar deformation features in quartz. Timms et al. (2012) used PDF for several different types of interior planar microstructures in lunar zircons. Cavosie et al. (2010), Moser et al. (2011), Erickson et al. (2013a) and Thomson et al. (2014) labelled all exterior and interior planar microstructures visible by SEM methods in zircons from Vredefort and Sudbury as PF, reserving the term PDF for potential amorphous planar microstructures invisible by SEM. Cavosie et al. (2018b) used a collective label PM for all planar microstructures. Austrheim & Corfu (2009) used both of the terms microfaults and PDFs for the single or widely spaced interior planar microstructures they found in zircons unaffected by impact. In experimentally shocked zircon Leroux et al. (1999) found closely spaced, interior amorphous planar microstructures ~10 nm thick along {320}, which they likened to PDFs in quartz as defined by Stöffler & Langenhorst (1994) and Grieve et al. (1996).

In summary, the previously used descriptive terms PDF, PF and PM for presumed diagnostic planar microstructures in zircon overlap with each other and lack precise definitions.

Planar deformation bands

Planar deformation bands (PDB, e.g. Timms et al., 2018), are irregularly and mostly widely spaced, planar to subplanar interior bands or domains with widths of up to several micrometres. In these cases, the zircon lattice has become distorted and misoriented by up to a few degrees relative to the host crystal by solid-state crystal-plastic deformation (Piazolo et al., 2012), not unlike tectonic deformation lamellae in quartz and other minerals, e.g. pyroxenes (Garde et al., 2015). PDBs commonly have asymmetric misorientation profiles with a sharp shift in the crystallographic orientation along one boundary and a gradual transition towards the host orientation along the other. Such features occur both in impact structures and in terrestrial seismites along tectonic zones (Piazolo et al., 2012; Kovaleva et al., 2015) and are not impact-diagnostic. Those described from Rochechouart by Rasmussen et al. (2020), may either be impact-related or inherited from the Variscan orogeny.

SAMPLES AND ANALYTICAL METHODS

Sample locations and rock types are shown in Table 1 and short sample descriptions are presented in Appendix 1. The rock chips were gently crushed with a hammer and ground in a small tungsten-carbide swing mill. Zircon and other heavy minerals were separated from the fine-grained rock powder using a Wifley water shaking table. Magnetic minerals were removed by a Nd-magnet. Thereafter zircon was picked by hand under a stereo microscope. In a few cases lithium polytungstate heavy liquid was used. In total about 3400 zircon grains from 34 rock samples were imaged using backscattered electron (BSE) and cathode luminescence (CL) scanning electron microscopy; 45 grains were mapped by electron backscatter diffraction (EBSD) scanning electron microscopy and two imaged by transmission electron microscopy (TEM). BSE, secondary electron (SE) and CL imaging of zircons was carried out at Lund University, Sweden, using a MIRA3 TESCAN FEG scanning electron microscope (analytical details as in Martell et al., 2021). No form of etching was performed to enhance exterior structures. EBSD analysis was performed at the Geological Survey of Denmark and Greenland on a ZEISS SIGMA 300VP scanning electron microscope (SEM), a Bruker e-FlashFS EBSD-detector and Esprit V2.1.2 software. Uncoated samples with finishing colloidal silica polishing were analysed at 20 kV and 30 Pa air pressure. A working distance of 14–17 mm and a pixel step size between 56 and 500 nm (as shown in each image), and EBSD resolution of 320x 240 or mainly 640 × 480, an exposure time of 5.5 to 16 ms were applied. Indexing percentages are >95% in non-metamict zircon (see the respective pole figures, which are untreated). For the TEM analysis, four focused ion beam (FIB) sections were cut out from selected zircon grains using a Helios G4 UC DualBeam electron microscope (ThermoFisher/FEI) based at the GFZ Potsdam. TEM was performed using a TECNAI F20 X-Twin transmission electron microscope operated at 200 keV with a Schottky field emitter as electron source. The foil was tilted in order to obtain a good-quality diffraction contrast with imaging of the dislocations. We did not use two-beam conditions.

RESULTS: OBSERVATIONS OF ZIRCON MICROSTRUCTURES

Manicouagan

Shocked zircons with abundant planar and granular microstructures in a shocked gneiss from the Manicouagan impact structure are described here in detail for the first time (Table 1). We examined about 230 grains in one sample, with five EBSD maps. A large majority of the grains display planar microstructures, and about 20% of them also show variable development of granular microstructures. Exterior surfaces contain two different types of planar microstructures, which have not previously been distinguished from each other in this or any other confirmed impact structures. The first type comprises very tight, contiguous planar microstructures in two or more sets with different orientations. They have a spacing of ~1 μm and are not separated by fractures. They are marked as CP (yellow lines) in Fig. 1a–d. These contiguous planar microstructures are accompanied by a second type of open planar fractures which are much more widely separated, with a spacing of typically 5 to 10 μm. This type is marked as PF (blue lines) in Fig. 1. The latter open planar fractures (Fig. 1, blue lines and arrows) commonly have micrometre-scale offsets and a stepwise habit, which exploits the first type of tight, contiguous planar microstructures in several orientations. Subplanar fractures also occur. In some grains, the contiguous planar microstructures change along strike into corn kernel-like trails of microgranules with a similar tight spacing (Fig. 1k–l), which may be coarsened into larger, randomly orientated granular neoblasts (black arrow); this new observation for Manicouagan zircons is not pursued here.

BSE imaging of interior surfaces revealed contiguous planar microstructures with the same spacing and orientations as the exterior ones, but less persistent. They form numerous ultrathin, straight, BSE-dark lines intensely decorated with very small micropores (orange arrows, Fig. 1e–f); these decorated lines in the BSE images closely resemble finely decorated PDFs in quartz, but are thinner. Individual micropores have diameters down to well below 100 nm; the larger ones appear to have coalesced. Interior, widely spaced, open planar fractures and coarse pore trails are parallel to exterior open planar fractures. EBSD images (Fig. 1h–i) also reveal a micrometre-scale blocky, interior structure defined by two sets of contiguous planar microstructures with small shifts in orientation along them. This blocky microstructure is expressed in imprecise pole orientations in the EBSD stereographic plots (Fig. 1j). No {112} twins were observed.

In order to obtain further information on the decorated interior planar microstructures observed with BSE and EBSD techniques we also performed transmission electron microscopy (TEM) shown in Fig. 2. A set of corresponding bright field (Fig. 2a) and high-angle annular dark field images (Fig. 2b) reveal straight dislocation arrays less than 50 nm wide lined with tiny fluid inclusions, which transect the imaged area from top to bottom. They are best seen as linear arrays of tiny dark spots in Fig. 2b, at the red circles in Fig. 2a and in the enlarged Figs 2c and2d. The dislocation arrays with orientations labelled 1, 2 and 3 (Fig. 2a–b) coincide with the exterior and interior, closely spaced and contiguous planar microstructures lined with pore trails mapped by BSE and EBSD (Fig. 2 f and i). The TEM images thus show that these pore trails are fluid inclusions (Fig. 2c–e) and furthermore confirm that the pore-connecting dark lines of the BSE and EBSD images are genuine nanoscale features of the crystal structure. Between the linear dislocation arrays, the matrix has a distinct but quite variable deformational cell structure, in which dislocations have migrated towards the walls of the individual cells (Fig. 2d), probably by a climb and glide process during incomplete annealing. The high-resolution lattice fringe image in Fig. 2g shows small irregularities in the crystal structure, and the whole foil appears as a single crystal pattern with slight misorientations of individual cells, corresponding to the EBSD-imaged structure.

Rochechouart

We inspected 27 zircons from three different samples (Table 1). They also contain two different types of exterior planar microstructures, which closely resemble those from the Manicouagan impact structure (Fig. 1m–n and Appendix Fig. A1). Importantly, interior surfaces likewise display abundant, finely decorated, closely spaced interior BSE-dark lines shown here for the first time (Figs 1m–n, A1). More widely spaced interior planar fractures also occur.

Vredefort

Our observations stand on the shoulders of previous detailed studies of Vredefort and Sudbury zircons such as Cavosie et al. (2010), Moser et al. (2011), Erickson et al. (2013a, 2013b) and Thomson et al. (2014). We examined 115 grains including six EBSD maps from a cataclastic orthogneiss with pseudotachylyte veins (Table 1). The exterior surfaces of a large majority of the shocked zircons contain distinct planar microstructures (Fig. 3) that appear to be generally better preserved than those from the footwall of the Sudbury structure described next, despite the older age and deeper exhumation of the former structure (see discussion). Exterior surfaces contain tight, contiguous planar microstructures with a regular ~1 μm spacing and predominant {112} orientations (yellow lines). There are also variably healed planar fractures with spacings around 10 μm (blue lines, Fig. 3a–f), which offset the tight, contiguous planar microstructures (along blue arrows) and are commonly associated with microcataclasis visible along their margins. Interior BSE and EBSD images (Fig. 3g–i) contain coarse pore trails along {112} with small, elongate, notably intermittent areas of spindle-shaped {112} twins along some of them (Fig. 3h–j), see discussion. No interior, closely spaced planar microstructures were found. A fragment without exterior crystal faces contains several coarse, widely spaced pore trails and longer, more coherent {112} twins along two planar fractures (Fig. 3k–n). In summary, we found {112} lamellar twins or small intermittent lens- or spindle-shaped {112} twins aligned along {112} in four of six analysed grains and conclude that {112} twins are common but not ubiquitous in shocked zircons from this sample (Fig. 3h–i).

Sudbury

Our observations below supplement previous studies of shocked Sudbury and Vredefort zircons by several other authors including Krogh et al. (1984, 1996), Bohor et al. (1993), Moser et al. (2011), Erickson et al. (2013a, 2013b) and Thomson et al. (2014). Planar microstructures in zircons from a cataclastic Levack gneiss with pseudotachylyte veins in the north-western footwall of the Sudbury impact structure (Table 1) are shown in Fig. 4. We examined 146 grains and made five EBSD maps. Exterior planar microstructures are easily recognisable in most of the zircons. Closely spaced, contiguous planar microstructures (marked by yellow lines, Fig. 4) are generally less distinct than in Manicouagan and Vredefort zircons. Planar fractures cut and displace them (blue lines and arrows, Fig. 4a–b). Both types of planar microstructures may be partly obscured and smoothed by ultrathin overgrowths (‘coats’, Fig. 4 c–d and e–f) to form intermittent furrows or grooves, like in some Rochechouart zircons (see Fig. A1). It is therefore locally difficult to distinguish the two types from each other. Exterior planar microstructures appear to have completely disappeared from the tip and upper left part of one grain (Fig. 4g). Interior surfaces contain planar fractures, trails of coarse, commonly elongate pores c. 0.2–0.5 μm across and lamellar {112} twins and small, lens- and spindle-shaped twins aligned along {112} (Figs 4g–k, l–n). Interior, closely spaced, contiguous planar microstructures are not preserved. Two other grains examined with EBSD (not shown) do not contain twins.

Seismites from Caledonian and Alpine non-impact deformation zones

In order to test if there might be similarities between the planar microstructures described above from confirmed impact structures and rare ones reported from non-impact zircons in endogenic terrestrial seismites (Austrheim & Corfu, 2009; Piazolo et al., 2012; Kovaleva et al., 2015), we re-examined both exterior and interior surfaces of zircons from both of these localities.

Seismite from a Caledonian deformation zone, SW Norway

Background information. H. Austrheim and F. Corfu kindly lent us a microcataclastic seismite from an intense Caledonian deformation zone in a Proterozoic metagabbro, SW Norway (Table 1; Austrheim & Corfu, 2009) for a re-examination of the planar zircon microstructures they had originally described from a microcataclastic felsic vein in the metagabbro. Their study remains to be the only reference to exterior planar zircon microstructures in a non-impact terrestrial setting, although no exterior images were actually shown.

Observations. Optical microscopy in thin section revealed a few zircon grains with single subplanar to planar microstructures (see supplementary sample description in Appendix 1). We also examined 186 newly extracted zircon grains. SEM-BSE and CL images confirmed that single interior, open, strictly planar fractures with or without small displacements were present in ten grains (the structures labelled microfaults and PDFs in the original description) and could also be identified on corresponding exterior surfaces (Fig. 5a–d), but only in four grains. Interior CL images of other grains revealed single or widely spaced, open interior, partly healed, subplanar structures crossing oscillatory and/or sector zoning (Fig. 5j–m). EBSD imaging revealed a minimal crystallographic reorientation across a single plane in just one grain (Fig. 5d–e), whereas several other grains have structurally homogeneous interiors (e.g. Fig. 5i). Closely spaced, somewhat irregular exterior lamellar structures without interior counterparts were observed in one grain (Fig. 5n–q); however, these exterior microstructures are subplanar, show branching/convergence and are not contiguous.

Seismite from the Ivrea–Verbano zone, northern Italy

Background information. Kovaleva et al. (2015) also described some interesting interior planar zircon microstructures including planar fractures and planar to subplanar deformation bands (PDBs) in three per cent of investigated grains from a seismite at Premosello in the Ivrea–Verbano zone, northern Italy. Garde et al. (2015) published a detailed description of intense cataclastic outcrop-scale and microstructures in the garnet-rich metabasic rock of this locality, see also Table 1 and Appendix 1.

Observations. We examined 83 zircon grains and nine EBSD maps from one sample and found a handful of rounded zircons up to ~150 μm across with individual, short exterior subplanar fractures only extending through parts of the respective grains, in addition to a few curvilinear fractures (Fig. 6). Larger elongate, stubby prismatic to rounded grains up to c. 200 μm in length only contain curvilinear fractures. Subplanar to planar deformation bands in some grains mapped with EBSD can be correlated with short exterior planar to subplanar fractures, with reorientations of up to and around ~2° relative to the host (Fig. 6a–c). Other grains contain subplanar to almost random misorientations of up to about 1.5° (Fig. 6d–k).

Maniitsoq structure

Background information. A detailed discussion of the interpretations given in the papers cited below is beyond the scope of this study. The Mesoarchaean (3.0 Ga) Maniitsoq structure in the North Atlantic Craton of West Greenland, exhumed from a crustal depth of 20–25 km, was proposed by Garde et al. (2012) to be a very large, deeply exhumed impact structure. Observations of widespread, non-linear, deep-crustal brecciation and cataclasis (first observed by Berthelsen, 1962), repeated textural evidence of direct melting of K-feldspar and biotite indicating a short-lived episode of extreme heating, as well as thorough high-temperature hydrothermal activity were described in this and subsequent papers (Scherstén & Garde, 2013; Garde et al., 2014; Keulen et al., 2015). The proposal was discussed by Reimold et al. (2013) and Garde et al. (2013). Several different regional, non-impact geological models were later proposed by Kirkland et al. (2018), Gardiner et al. (2019), Waterton et al. (2020) and Yakymchuk et al. (2020) and discussed by Garde et al. (2020). However, all of the latter models ignored the evidence of extensive cataclasis in the Finnefjeld domain and of shock melting of rock-forming minerals, with implications for their scientific validity. An EBSD study of the interiors of 5587 zircon grains retrieved from randomly collected rock and detrital samples (Yakymchuk et al., 2021) failed to find evidence of shock metamorphism, but no exterior zircon surfaces were investigated by the latter authors.

Observations. We examined ~2600 zircon grains in 24 samples from the inner part of the Maniitsoq structure, including 15 EBSD maps. In the following, we report observations of planar microstructures in zircons from eight of these samples (Table 1); occasional grains with a few, strictly planar exterior microstructures were found in eight other samples.

Planar microstructures are most common in two samples at the NE and SW margins of the cataclastic Finnefjeld domain in the centre of the structure, respectively: an orthogneiss with almost glassy, pseudotachylyte-like parts (sample 525 022) and a breccia (sample 257 543). Seventy-six of 216 zircons from the former sample and almost all of ~300 from the latter contain one or more sets of exterior planar microstructures. Contiguous planar microstructures spaced at ~1 μm are readily distinguished from more widely spaced planar fractures in most grains; like in the shocked zircons from confirmed impact structures the planar fractures are parallel to but displace the former microstructures and may be associated with cataclasis along their margins (Figs 7–9). In the predominantly prismatic zircons from sample 525 022 the most distinct planar microstructures occur along {100} but are always also associated with oblique ones, whereas in the stubby zircons of sample 257 543 the planar microstructures are mostly oblique. Coarse interior pore trails may represent annealed planar fractures (Fig. 9f). The interstitial material in these coarse pore trails is mostly undetermined Al ± Ca ± Na ± Fe ± Mg silicate (SEM-EDS analysis) as in Vredefort zircons (Moser et al., 2011). EBSD analysis of zircons from sample 525 022 revealed very closely spaced, strictly planar interior lamellae with one or two mutually crossing orientations parallel to the exterior contiguous planar microstructures and each up to a few micrometres wide. The crystallographic misorientation profiles of the lamellae are symmetrical and vary systematically between ~0.5–1° relative to each other (Fig. 7h, l, q), reminiscent of the blocky interior misorientations observed in Manicouagan zircon (Fig. 1h–i). Closely spaced, interior contiguous planar microstructures marked by straight, very fine pore trails are visible in the BSE image of grain P2–11 (Fig. 7g), and overlap with the lamellar structure seen in the EBSD image (Fig. 7h). These pore trails are only up to ~10 μm long, they are only present in some parts of the grain, and are more or less invisible except at high magnification (Figs 7g and8c). They are therefore interpreted as quite extensively annealed. TEM images from this grain reveal a set of straight, vertical dislocation arrays only 50–100 nm wide (black arrows), which correspond to the fine pore trails mapped by BSE (Figs 7g and8c). They resemble those observed in the Manicouagan grain P7–5, but they are less distinct, and only one of them is lined with tiny fluid inclusions (orange arrow). Thicker and more widely spaced dislocation arrays with the same orientation are also present, marked by curled brackets in Fig. 8b. These broader dislocation arrays are without fluid inclusions and are interpreted as corresponding to deformational boundaries between the individual longitudinal lamellae mapped by EBSD (Figs 7h and8c). All of these microstructures have been superimposed by an area of crystal-plastic deformation in the upper right part of the grain, and they are fully annealed along an oblique zone of recrystallisation about 15 μm wide (between the dotted lines in Fig. 7h).

In zircons from sample 257 543, interior, irregular to subplanar zones of crystal-plastic deformation may occur (Fig. 9h). A stubby grain from this sample has an interesting, almost completely homogeneous interior structure with only faint remnants of growth zonation, and there are almost no pore trails (Fig. 9k–l). Several crystal faces of this particular grain have areas where the exterior planar microstructures are indistinct or almost completely absent (Fig. 9j). This grain is interpreted as representing an advanced stage of recrystallisation accompanied by full interior and partial exterior annealing of the planar microstructures.

Variably modified exterior planar microstructures in zircons from six other samples (Table 2) are shown in Fig. 10. Contiguous exterior planar microstructures in some grains have the same close spacing as in zircons from the confirmed impact structures (Figs 1–4), but they are commonly only visible in parts of the grains (e.g. Fig. 10b, h). Several grains (Fig. 10c–f) have thin overgrowths and smoothed exterior structural details, like in some Rochechouart and Sudbury zircons (Figs 4, A1); the planar microstructures appear as intermittent grooves, and the two types of planar microstructures cannot easily be differentiated from each other – except arguably by variations in spacing and depth (Fig. 10c). In some grains the contiguous, closely spaced planar microstructures change along strike into straight or crenulated ribbons (Fig. 10g–h). In other grains, open planar fractures are transformed along large parts of their length to prominent, widely spaced exterior ribbons (Fig. 10j–l).

DISCUSSION

Summary of microstructures in zircons from confirmed impact structures

Our observations of planar microstructures in zircons from the Manicouagan, Rochechouart, Vredefort and Sudbury impact structures can be summarised as follows. Exterior, closely spaced, contiguous planar microstructures are readily distinguished from open, texturally younger planar fractures. The first type of planar microstructures forms up to several sets with different orientations, commonly {100} or {112}, and a spacing of ~1 μm; their actual thickness is below the SEM resolution that we could acquire. We document for the first time that closely spaced interior planar microstructures in Manicouagan and Rochechouart zircons directly correspond to the external ones in terms of positions, spacing and orientations (Fig. 1e–f). In BSE images, they are less persistent than their exterior counterparts, decorated with abundant tiny pores mostly much smaller than 100 nm across, and appear to have been partly annealed. In EBSD images, they appear as dark lines with absent or poor Kikuchi signal and variably dotted with tiny dark specks (Fig. 1i). TEM images from Manicouagan zircon show that these interior microstructures are very narrow dislocation arrays up to ~50 nm wide abundantly decorated with tiny fluid inclusions, identical to TEM images of decorated PDFs in quartz (see discussion). Texturally younger, open planar fractures are associated with microcataclasis and variably healed. They exploit and cut the much more closely spaced, contiguous planar microstructures in a stepwise fashion. In some grains from Rochechouart and Sudbury, the exterior planar microstructures are smoothed by thin overgrowths and thereby gradually changed into small and locally intermittent grooves, erasing the distinction between the two types of planar microstructures recognised in other grains from the same impact structures. These observations collectively show that post-impact annealing of planar microstructures in shocked zircon is a common phenomenon.

We found {112} twins in some but not all of the investigated zircons from Vredefort and Sudbury, but not in the Manicouagan zircons. Those in the Vredefort grains and in one of the Sudbury grains are short and spindle-shaped, and some are aligned with each other, suggesting partial annealing along the length of former lamellar twins.

Summary of microstructures in zircons from Caledonian and Alpine seismites

Our study confirmed the planar interior microstructures reported by Austrheim & Corfu (2009) from a non-impact Caledonian seismite in SW Norway and substantiates that they are also visible on exterior surfaces (Fig. 5). They are difficult to distinguish from open planar fractures per se in impact structures, which are therefore not diagnostic of impacting. In one grain, we observed a small area of exterior, closely spaced subplanar and curviplanar fractures (Fig. 5q), but without any of the distinctive characteristics of the tight, contiguous and strictly planar microstructures spaced at ~1 μm that occur in shocked zircons.

Our study of sample 480 623 from the seismite locality at Premosello along the Insubric line in the Ivrea–Verbano zone, northern Italy (Fig. 6) confirms that there are rare, subplanar to planar interior structures in zircons from this locality with variable and irregular misorientation, similar to some of the crystal-plastic deformation types described by Piazolo et al. (2012) from the Bergen arc in Norway. Exterior planar fractures in some grains have orientations that appear to match interior misorientations, but they are rare and most of them do not transect entire grains.

Summary of microstructures in zircons from the Maniitsoq structure

Rock samples with evidence of cataclasis from several different parts of the Maniitsoq structure contain numerous zircons with planar microstructures comprising both exterior, closely spaced, contiguous planar microstructures and open planar fractures (Figs 7–10), and a few remnants of closely spaced interior contiguous planar microstructures. No {112} microtwins have been found to date at Maniitsoq. The planar microstructures closely resemble those in the confirmed impact structures studied here (Figs 1–4, A1). As also observed at Rochechouart, Vredefort and Sudbury, many of the planar microstructures of the Maniitsoq zircons appear to have been variably affected by annealing. In particular, grain surfaces may be smoothed (Figs 8j, 9). Also, the interiors of some grains are very homogeneous without preservation of growth zonation or planar fractures (Fig. 9k–l). The interiors of a few other grains (Fig. 7h, l) contain one or two sets of narrow, contiguous lamellae a few micrometres wide with deformational boundaries parallel to the exterior contiguous planar microstructures. These lamellae resemble the blocks intercepted by two sets of contiguous planar microstructures in Manicouagan zircons (Fig. 1 h–i). TEM images of Maniitsoq zircon show narrow dislocation arrays up to ~100 nm wide, one of which is decorated with tiny fluid inclusions (Fig. 8). These arrays are similar to those observed in the Manicouagan zircon but are much less distinct. The TEM images also reveal broader and more widely spaced dislocation arrays, which are interpreted as corresponding to the boundaries of the above-mentioned contiguous lamellae. These types of interior planar microstructures are readily distinguished from the irregular to subplanar misorientation structures seen e.g. in Fig. 9h. The latter misorientation structures have highly variable lengths and widths, they need not be planar, they are not repeated in regular fashion, and they commonly have asymmetric misorientation profiles.

In summary, the contiguous planar microstructures in the Maniitsoq zircons closely resemble those found in shock-metamorphosed zircons from major confirmed impact structures, and they are very different from the only known examples of rare, individual subplanar to planar fractures in non-impact zircons (Austrheim & Corfu, 2009; Kovaleva et al., 2015; our Figs 5–6), where contiguous planar microstructures are completely absent.

A revised observation-based nomenclature

The new distinction described in the previous section between tight, contiguous planar microstructures and widely spaced, open planar microstructures in zircon is important because it allows us to distinguish between planar microstructures that are only found in impact structures and others that may also occur in endogenic tectonic environments. In the figures we provisionally marked the two types as CP and PF, respectively (Figs 1–9). We are now in a position to propose a revised nomenclature in Table 2, which is intended to refine and simplify the previous inconsistent terminology. We define CP as tight, contiguous planar microstructures with a spacing of ~1 μm, visible by BSE and internally also by EBSD and TEM imaging, which are not separated by fractures and do not display offsets. We define PF as open planar fractures with a variable spacing of around 5–10 μm, common small offsets and marginal microcataclasis. Where both CPs and PFs are present, the PFs are always younger than the CPs and re-use CP orientations, commonly in a stepwise pattern.

CP is a new descriptive term applied to a distinctive subset of exterior and interior planar microstructures, which has not previously been recognised as such or separated from PFs. CPs are only found in zircons from impact structures, and we thus propose that they are diagnostic of impact. The term PF survives from previous nomenclature as it directly refers to planar fractures (i.e. open microstructures), and this term and is unambiguous once the term CP has been established. Single or very widely spaced PFs have also been reported from zircons in two Caledonian and Alpine non-impact seismites, and both were confirmed by us. Therefore, PFs per se are not diagnostic of impact.

Planar deformation bands (PDBs) and low-angle grain boundaries are interior crystal-plastic deformation features which may occur in both impact and non-impact settings and which are not diagnostic of impacting (see nomenclature above). They are known from seismites in major tectonic zones (Piazolo et al., 2012; Kovaleva et al., 2015; this study), from terrestrial impact structures such as Vredefort (Moser et al., 2011 their fig 4c, Erickson et al., 2013a their figs 3D and5D-F) and Yarrabubba (Erickson et al., 2020 supplementary fig. 1), as well as from Rochechouart where they might alternatively have been inherited from the Variscan orogeny (Rasmussen et al., 2020). They are also known from shocked lunar zircons (Timms et al., 2012). They may be superimposed on impact-induced planar microstructures. It is generally agreed that these microstructures result from crystal-plastic deformation along several different slip systems at lower strain rates than those necessary to form PFs (see Piazolo et al., 2012 and Timms et al., 2018 for details on slip systems and physical deformational constraints). Planar misorientation bands are readily distinguished with EBSD mapping from CPs, inter-CP lamellae and PFs by their broad and variable width, variable spacing and orientation, commonly asymmetric misorientation profiles and other inherent orientation irregularities.

Correlation between exterior and interior planar microstructures in zircon, and between decorated PDFs in quartz and zircon

In their study of shocked Vredefort zircons Erickson et al. (2013a) showed that exterior PFs have direct interior counterparts in the same grains. This correlation is straightforward. Like their exterior counterparts the interior PFs are associated with cataclasis and displacement (i.e. planar microfaults), and they may be partially annealed into trails of coarse pores (e.g. Fig. 3g). We demonstrate here for the first time with Manicouagan zircon that there is a similar direct correlation between excellently preserved exterior CPs and variably decorated, partly annealed interior CPs with similar spacing and orientations (Fig. 1c–f, m–n). In BSE and EBSD images, they form straight trails of delicate micropores connected by ultrathin, BSE-dark lines and weak or absent EBSD Kikuchi diffraction patterns (Fig. 1h–i). TEM images reveal that these microstructures are very narrow, only ~50 nm wide arrays of dislocations abundantly lined with fluid inclusions (Fig. 2), precisely like decorated PDFs in quartz. More specifically, they correspond to stage c of the formation of decorated quartz PDFs in the model by Grieve et al. (1996), which was based on TEM imaging of natural shocked quartz (Goltrant et al., 1991). When the amorphous PDFs crystallise back to quartz (viz. zircon), water is expelled forming bubbles. The bubbles cause an overpressure, and this overpressure will cause the formation of dislocations in the surrounding quartz if T > 300°C (Schmidt et al., 2003). These dislocations are not like a dislocation array forming low-angle grain boundaries, but they have formed during recrystallisation and movements (climb and glide) after their formation (T > 300°C). The development of decorated PDFs in quartz and zircon is illustrated in Fig. 11 and discussed further below.

In Manicouagan zircons as well as in a couple of the best-preserved shocked zircons from Maniitsoq there are closely spaced interior lamellae with orientations of their margins that match exterior CPs, and small, regular interior misorientations (Fig. 7h). Such zircon microstructures have not been described in works from Vredefort or Sudbury, although similar parallel, closely spaced lamellae with small regular variations in their crystallographic orientation can actually be observed in a Vredefort zircon in Fig. 6 of Erickson et al. (2013a).

Further examination of other previous descriptions of interior microstructures in shocked zircons shows that closely spaced interior, ultrathin parallel lines with poor to absent Kikuchi diffraction patterns have indeed previously been imaged by BSE and EBSD but not actively correlated with exterior microstructures; these parallel, ultrathin lines most likely reflect amorphous material or strained zones with dislocation arrays below the resolution of BSE and EBSD. For instance, Timms et al. (2012) found them in lunar zircon and tentatively labelled them PDF. In a zircon grain from the recent Chicxulub borehole (Wittmann et al., 2021, their Fig. 4c–e) there are similar, closely spaced interior planar microstructures, which were labelled PF. They are cut by reidite lamellae. Being ultrathin, contiguous, with a spacing ≤1 μm and being texturally the oldest, these planar microstructures would most likely constitute interior CPs in the proposed new nomenclature. Closely spaced, contiguous interior planar microstructures are also visible in shocked zircons from the Woodleigh impact structure, Australia (Cox et al., 2018). Some of the latter zircons host lamellar {112} microtwins which cut patches of reidite (Cox et al., 2018, their Fig. 3). The relative timing of the interior microstructures at Woodleigh is thus: 1) formation closely spaced contiguous planar microstructures (CPs in our nomenclature), 2) conversion of zircon to reidite, 3) formation of lamellar microtwins, and 4) displacement along the first-formed contiguous planar microstructures.

In an experimentally shocked zircon, contiguous interior planar microstructures consisting of amorphous planes only up to 10 nm thick were found by TEM more than two decades ago (Leroux et al., 1999). They are probably equivalent to the decorated CPs described here, although the experimental microstructures were only found with orientations close to {320} and were not correlated with exterior microstructures. Leroux et al. (1999) considered that the amorphous planes were texturally older than adjacent areas converted to scheelite structure (i.e. reidite). They interpreted them as PDFs, like PDFs in quartz, with a reference to Stöffler & Langenhorst (1994). In another TEM study, this time of natural shocked zircon from Vredefort, Reimold et al. (2002) identified an interior subplanar, nano- to micrometre-scale lamellar structure (again without exterior correlation) and suggested that it might represent partly annealed amorphous planar microstructures.

Formation of planar microstructures in zircon

Timing of CP formation

Cross-cutting relations between planar and subplanar microstructures and {112} twins in Vredefort zircons prompted Moser et al. (2011) and Erickson et al. (2013a) to suggest a relative chronology, but since CPs and PFs were not distinguished from each other the critical step from contiguous to open planar microstructures was not recorded. Our study shows that CPs form before PFs. Assuming that our interpretations above of interior CPs in the images from Timms et al. (2012), Cox et al. (2018) and Wittmann et al. (2021) are correct, CPs also predate twinning and conversion of zircon to reidite. This shows that CPs are formed immediately during the passage of the shock wave, prior to potential conversion to reidite and/or twinning.

Interpretation of CPs as PDFs

CPs are the first shock-induced planar microstructures observed to form in zircon, and their close spacing, contiguous nature and interior decorations with fluid inclusions are identical to decorated PDFs in shocked quartz. In quartz the PDFs represent original, shock-induced amorphous planes which are thought to have been amorphisized or melted during the compressional stage of the shock wave and quenched during its rarefaction stage (e.g. Kingma et al., 1993). The PDFs subsequently constituted passageways for fluids. When the amorphous planes subsequently reverted to quartz this resulted in exsolution of numerous tiny fluid inclusions (decorated PDFs, Stöffler & Langenhorst, 1994; Grieve et al., 1996). In quartz, the decorations are thus markers of the original, shock-induced amorphous planes. By default, the decorations with fluid inclusions along the nanometre-scale dislocation arrays in the zircons indicate that these planar microstructures mark former fluid passageways along very narrow, originally amorphous planes. The CPs thus constitute remnants of original PDFs, see Fig. 11. The presence of amorphous microstructures interpreted as PDFs in experimentally shocked zircon (Leroux et al., 1999) supports this interpretation, although the latter supposed PDFs were few and only occurred in one orientation close to {320}.

Relationships with reidite

The microstructures interpreted as PDFs in experimentally shocked zircon by Leroux et al., 1999 occurred adjacent to large areas transformed to reidite and were interpreted as predating it, but the experimental data were insufficient to qualify if the reidite lamellae might have been formed from PDFs. Reidite from zircon in the Ries crater, Germany, forms in a number of lamellar orientations within the host zircon but rarely if ever in {112} (Erickson et al., 2017), whereas Plan et al. (2021) described reidite along lamellar {112} twins in Rochechouart zircons. However, in both of these impact structures the reidite lamellae are much more widely spaced and thicker than the CPs observed by us in the Rochechouart samples and were therefore not transformed from CPs (viz. from PDFs).

A visual 3D model of planar microstructures in zircon

A visual 3D model of exterior and interior planar microstructures in shocked zircon is shown in Fig. 12. It illustrates the two successive and structurally different types of shock-induced planar microstructures, CP and PF, their exterior–interior relationship, and their gradual transformation and annealing as observed in this study. We emphasize the importance of previous studies (e.g. Krogh et al., 1984; Cavosie et al., 2010; Moser et al., 2011; Erickson et al., 2013a, 2013b; Thomson et al., 2014) as foundations for the present study, where some of the crosscutting relations shown in Fig. 12 can also be seen.

The distinction between exterior CPs and PFs is readily observed in the original images of shocked Sudbury zircons published by Krogh et al. (1984) and Bohor et al. (1993), when the difference between CPs and PFs had not yet been recognised and all planar microstructures were labelled as PDFs. CPs and PFs can also readily be distinguished in other published images of exterior planar microstructures in shocked zircons (e.g. Vredefort zircons; Erickson et al., 2013a; Cavosie et al., 2018b).

Interpretations of PFs in impactites and terrestrial seismites

The texturally younger, open PFs associated with CPs in impact structures re-using already established CP planes of weakness may be regarded as a special type of cleavage and may be interpreted as resulting from impact-induced seismic shaking in the wake of the shock wave itself. PFs might also be initiated during the decompression stage of the shock wave. PFs also occur without CPs in less intensely shocked zircon grains in non-impact settings, see below. Seismic shaking is an inherent early component of the cratering process, as recently shown in Chicxulub drillcore M0077A (e.g. Morgan et al., 2016; Rae et al., 2019). Both Garde & Klausen (2016) and Kovaleva et al. (2019) invoked seismic shaking to explain the generation of Vredefort pseudotachylytes, because frictional heating and eventual melting cannot explain the formation of pseudotachylytes except for very thin veins (Melosh, 2005). Seismic shaking was previously not considered important for modification of rocks and minerals during the cratering process, except for the concept of acoustic fluidisation; the latter process infers large-scale granular flow in a dissipating stress field due to vibrations released by the initial shock wave (Melosh, 1989).

The rare planar and subplanar open fractures without associated CPs described in zircons from non-impact settings (Austrheim & Corfu, 2009; Kovaleva et al., 2015; this study) were interpreted by the original authors as results of intense seismic cataclasis in major terrestrial deformation zones. We agree with this general interpretation and repeat that the presence of PFs alone is not diagnostic of extraterrestrial impact.

Annealing

Reimold et al. (2002) were probably the first to discuss the possibility of gradual annealing of shock-induced planar microstructures in zircon, but to date this issue has only received little attention despite its potential importance in the recognition of very ancient, deeply exhumed and/or reworked impact structures. Several questions arise: which shock-induced microstructures in zircon are formed in the deep roots of impact structures, which are most robust, and can they still, after partial annealing, be distinguished from other microstructures that are not diagnostic of impact?

Both our own and previous studies show that the exterior, contiguous and closely spaced planar zircon microstructures (our CPs) first described from Sudbury are very robust (visualised in Fig. 12). Secondly, we show that corresponding interior CPs also exist in some well-preserved zircons, for instance in Manicouagan and Rochechouart (Figs 1, A1), and that they are dislocation arrays decorated with delicate fluid inclusions and identical to partly annealed PDFs in quartz. We also show that interior CPs are absent from Vredefort and Sudbury zircons, presumably because of more advanced annealing. Healed planar fractures with much wider spacing (PFs) are best preserved on exterior zircon surfaces. Observed on interior surfaces they are gradually converted to subplanar trails of coarse pores (Figs 1–4 and 10d–f) or fully annealed. Those fractures and coarse pores that remain contain quartz or more commonly undetermined silicate material, which may represent remnants of injected quartzo-feldspathic melts (Moser et al., 2011 and this study). Very thin surficial overgrowth or recrystallisation results in smoothing and gradual annealing of planar microstructures as illustrated in the zircon annealing model (Fig. 12f). This was observed in Rochechouart and Sudbury zircons (Fig. 4c–f), where a visual distinction between CPs and PFs can be difficult. Such smoothing is also observed in zircons from the Maniitsoq structure.

Lamellar microtwins and reidite: Annealing or absence?

No impact-induced lamellar {112} microtwins were found in the Manicouagan zircons that we studied. Perhaps this is not surprising, as the zircons were extracted from a (clast-rich) melt rock. In similar upper-crustal samples e.g. from Ries (Erickson et al., 2017) and the Kamestastin (Mistastin Lake) impact structure, Canada (Tolometti et al., 2022), microtwins are relatively rare. They are present, but not ubiquitous, in shocked Sudbury and Vredefort zircons (see also Moser et al., 2011; Erickson et al., 2013a, 2013b; Thomson et al., 2014; Kovaleva et al., 2019), but commonly with textural evidence of partial annealing where only small spindle-shaped twins aligned along {112} are present (Thomson et al., 2014 and this study); Timms et al. (2018) state that the lateral growth of {112} twins in zircon is much faster than their nucleation, giving rise to long, lamellar twins.

It remains an open question if lamellar twins were fully annealed in the shocked but twinless zircons from the Maniitsoq structure, if twins might still be found, or if they were never formed. As shown in the introduction and above, the individual shock features in zircons from each of the discussed impact structures belong to different shock conditions and crustal settings.

Although {112} zircon microtwins are only known from impact structures, we note for completeness that their validity as unequivocal shock indicators has been questioned by Morozova et al. (2018). Also for completeness, as mentioned in the introduction, reidite is not known from exhumed and/or reworked impact structures and is highly unlikely to have formed or survived in the deep parts of very large impact structures where to ambient temperature was high and cooling slow.

Identification of ancient and deeply exhumed impact structures using planar microstructures in zircon

As described in the introduction, recognition of large, ancient, deeply eroded terrestrial impact structures in complex, polydeformed, deep-crustal rocks is difficult. First of all, it requires intimate familiarity with the normal field appearances and structures of such rock complexes at all scales in order to recognise the unusual. The most likely and obvious field indicators of hidden, deeply exhumed, lower-crustal remnants of impact structures are probably widespread and intense, non-linear cataclasis locally associated with small patches of bulk rock melting and randomly located pseudotachylytes or microcataclasites unrelated to any linear seismic zones. Melt sheets are confined to the upper parts of impact structures and would not be contained within deeply exhumed ones. Nevertheless, microstructures in deeper, shock-affected rocks may remain chaotic and unequilibrated, with intense fracturing of plagioclase concomitant with evidence of direct melting of e.g. K-feldspar and biotite (i.e. without involving mineral reactions as in ordinary ultrametamorphism). These phenomena require a temperature excursion well beyond what can be attained by endogenic lower-crustal processes. With such regional observations more specific microstructural evidence of impacting can be sought in minerals where impact-diagnostic features might survive. The present study shows that zircon is suitable for such tests. Some of the most common impact-diagnostic features in this mineral, in particular exterior CPs, are robust and able to survive even deep burial, recrystallisation and tectonothermal overprinting.

CONCLUSIONS

Planar microstructures in natural shocked zircons have been known for several decades, but we demonstrate for the first time that they comprise two entirely different components formed in succession, namely contiguous planar microstructures (CP) with a spacing of ~1 μm, which are diagnostic of extraterrestrial impact, and more widely spaced open planar fractures (PF), which are not.

Contiguous planar microstructures (CP)

We also demonstrate that exterior CPs have direct interior counterparts with the same orientations and spacing. The interior CPs consist of several sets of dislocation arrays with different orientations. They are ~50 nm wide and lined with tiny fluid inclusions. These nanostructures are identical to decorated planar deformation features (PDFs) in quartz. Like decorated PDFs in quartz, the CPs in zircon are interpreted as partly annealed amorphous planes only tens of nanometres thick, formed by an extraterrestrial shock wave.

Planar fractures (PF)

Planar fractures (PFs) are open microstructures, which are texturally younger than CPs. They re-use some of the CP planes, commonly in stepwise patterns, and have a wider and irregular spacing of ~5–10 μm. PFs in shocked zircons are interpreted as results of impact-induced seismic shaking in the wake of the shock wave, whereby some of the CPs are opened and displaced. We confirm two previous studies showing that single PFs without stepwise habits may also rarely form by seismic shaking in terrestrial, non-impact seismogenic zones, and we show that such PFs are not accompanied by CPs. PFs are therefore not themselves diagnostic of impact.

Annealing

Planar microstructures. We show that impact-induced zircon microstructures may be progressively annealed by recrystallisation and exterior overgrowth. Interior CPs appear to be the first impact-induced planar microstructures to disappear, whereas exterior CPs are very resistant to annealing. In Manicouagan and Rochechouart zircons interior CPs are decorated, and no interior CPs were observed in Vredefort or Sudbury zircons. The much coarser interior PFs are more resistant to annealing but are commonly converted to trails of coarse pores. Smoothing of exterior CPs and PFs by thin overgrowths was observed in zircons from Rochechouart, Sudbury and Vredefort.

Microtwins and reidite. Lamellar {112} microtwins may also be annealed. Some of the {112} microtwins in Vredefort and Sudbury zircons are not lamellar but occur intermittently as small lens- or spindle-shaped bodies along {112} planes. No {112} microtwins were found in the Manicouagan zircons investigated by us. Reidite appears limited to upper-crustal impact structures and is highly unlikely to have formed or survived in the roots of deeply exhumed impact structures, because preservation of reidite requires rapid cooling.

Summary. Shock-induced microstructures in zircon are prone to annealing. Exterior CPs are probably the most robust of the diagnostic shock-induced microstructures in zircon discussed here. Their different degrees of preservation in e.g. the Manicouagan, Rochechouart, Yarrabubba, Vredefort and Sudbury impact structures suggest that annealing is primarily governed by depth of burial and geological reworking rather than by age alone.

Planar microstructures in Maniitsoq zircons

Contiguous, closely spaced exterior planar microstructures (CPs) in zircons from several parts of the Maniitsoq structure are identical to those in zircons with or without variable effects of annealing from the four confirmed impact structures examined here. Like in the confirmed impact structures, the CPs are associated with texturally younger PFs. The microstructures in the Maniitsoq zircons therefore corroborate the original proposal by (Garde et al. (2012) that this very large and very deeply exhumed structure resulted from an extraterrestrial impact.

DATA AVAILABILITY

The data underlying this article primarily consist of images produced by electron microscopy and available in the article itself and in the online electronic material.

SUPPLEMENTARY DATA

The Supplementary data are available at Journal of Petrology online. They consist of a composite figure of BSE images of zircons from the Rochechouart impact structure (S1) and Appendix 1 with descriptions of all analysed samples.

FUNDING

The laboratory work was supported by the Geological Survey of Denmark and Greenland, Lund University and GeoForschungsCentrum Potsdam, with support to the latter laboratory by the Electron and X-ray microscopy Community for structural and chemical Imaging Techniques for Earth materials (EXCITE) program of the European Community.

CONFLICT OF INTEREST

No conflicts of interest to declare.

Acknowledgements

We dedicate this study to Andrew Glikson, John Spray and not least the late H. Jay Melosh, who understood the significance of the Maniitsoq structure already when the impact was first proposed in 2012. We are grateful to Håkon Austrheim and Fernando Corfu, Oslo University, Norway, for sharing their seismite sample FLO11-02 from the Svarthumlevatnet metagabbro in Norway with us and to Fernando Corfu for an incisive comment on the nomenclature, to Félix Gervais, Polytechnique Montréal, Canada, for providing the Manicouagan sample, and to Anders Plan, Lund University, Sweden, for providing images of Rochechouart zircons. The TEM work at GFZ Potsdam was supported by the EU EXCITE research program. Prior to this study Balz Kamber, QUT, Australia, pointed out to the first author that planar microstructures in zircons from the Sudbury impact structure are best preserved on exterior surfaces. Brendan Dyck, University of British Columbia, Canada, and Brian Windley, Leicester University, U.K., suggested that an overview of impact-diagnostic zircon (and quartz) microstructures and their gradual annealing by terrestrial geological processes might be useful for non-specialists. We thank Gavin Kenny, Stockholm University and Paula Lindgren, Lund University, Sweden, Des Moser, Western University, Canada, and Sandra Piazolo, Leeds University, U.K. for previous discussion and help with unpublished EBSD analyses of Maniitsoq zircon. We thank Sebastian Næsby Malkki, Geological Survey of Denmark and Greenland, for additional help with SEM-BSE imaging, and Will Hyde, Nicolaj Krog Larsen and Tod Waight, University of Copenhagen, Denmark, Anders Plan, Lund University, Sweden, John Spray, University of New Brunswick, Canada and Alex Wittmann, Arizona State University, U.S.A., for helpful comments to earlier versions of this study. We sincerely thank Timmons Erickson, NASA Johnson Space Center, U.S.A., and an anonymous reviewer for insightful and throughout constructive reviews.

References