Long story encrypted in a small grain – zircon from meta-andesite in the Lower Köli Nappes reveals a complex history of the Virisen Arc Terrane, Scandinavian Caledonides, Sweden

The sample which is the focus of this study belongs to the Ankarede Volcanite Formation (AVF) of the Björkvattnet Nappe, the lowermost nappe unit of the KNC in the north Jämtland and Västerbotten region (e.g. Kulling, 1933; Sjöstrand, 1978; Stephens, 1982, 2001). The stratigraphically lowest strata of the Björkvattnet Nappe in the study area comprise a unit of mica schists and phyllites with local conglomerates and detrital serpentinites (Seima Fm.). The AVF overlies this and is composed of a predominantly acidic to intermediate metavolcanic rocks with subordinate metabasalts and interlayers of tuffitic phyllites. Coarser-grained varieties of metatonalite and metadiorite may have been subvolcanic intrusions (Claesson et al., 1983).

Trace element and geochemical compositions of the metavolcanites support an intra-oceanic island arc origin (Grimmer and Greiling, 2012). Magmatic rocks within the AVF have been dated by U-Pb zircon at 512 ± 3.5 Ma and ca. 488–497 Ma (Carter et al., 2023; Claesson et al., 1983). Stratigraphically above the AVF lies a metasedimentary flysch sequence of metagreywackes, phyllites and conglomerates capped by a probable Hirnantian-Llandovery quartzite-carbonate sequence with metabasalt. The Björkvattnet Nappe has experienced metamorphism at greenschist to lower amphibolite facies, as well as extensive pre-metamorphic sodium metasomatism; the reaction between felsic to intermediate volcanic rocks and seawater, under hydrothermal conditions, lead to the formation of quartz-keratophyres and spilites (Sjöstrand, 1978). Such metasomatic enrichments in Si and Na, probably related to ocean floor hydrothermal action, have been described in the similar Stekenjokk Quartz Keratophyre Formation in the structurally overlying Stikke Nappe (Middle Köli Nappes; Stephens, 1980, 1982). For a more detailed description of the tectonostratigraphy of the Kӧli Nappe Complex within the Västerbotten and North Jämtland region the reader is referred to Carter et al. (2023).

The studied meta-andesite sample (KW19-11B) belongs to the Ankarede Volcanite Formation. It was collected from the bank of a small stream near Gränssjö (65.422167° N 14.806194° E; for outcrop pictures see supplementary file in Carter et al. 2023: https://doi.org/10.6084/m9.figshare.24168232.v1) on the eastern limb of the Western Synform (Supplement S1). The rock is generally fine-grained and has a weak foliation defined by the alignment of white mica, chlorite and quartz laminae. Although chemical analysis indicates that the rock was originally an andesite (Carter et al., 2023), only relics of the original plagioclase can now be observed. Moreover, the rock comprises quartz and secondary minerals such as zoisite, albite, epidote and chlorite resulting from the breakdown of primary plagioclase and amphibole. The titanium minerals are represented by ilmenite which is partly altered into titanite. Accessory minerals are zircon and apatite. The sample has been a subject of studies by Carter et al. (2023), therefore, we encourage interested readers to see the detailed petrographic and geochemical study of the sample in that publications. U-Pb dates (LA-ICP-MS) for zircon in this sample range from 520 ± 5 to 476 ± 5 Ma, with a coherent cluster for zircon cores, giving a mean age of 491 ± 3 Ma (Carter et al., 2023).

For the study we used the same zircon mount used by Carter et al. 2023. Cathodoluminescence (CL) images of mounted grains were obtained using a Scanning Electron Microscope at AGH University of Krakow, Poland. Backscattered electron (BSE) images, chemical mapping and chemical composition measurements of zircon were performed using a Jeol Superprobe 8230 electron microprobe at the Faculty of Geology, Geophysics and Environment Protection, AGH University of Krakow. Chemical mapping for Y, U, Hf, Si, Zr, and Al (± P) has been performed. Si, Al and P maps have been used to locate the presence of inclusions in zircon. The operating conditions for zircon chemical maps were: acceleration voltage of 15 kV, probe current of 100 nA and dwell time of 0.1s. Final versions of chemical maps were prepared using XMapTools 4.1 (Lanari et al., 2014; 2019; 2023). The operating conditions for zircon chemical profiles were: acceleration voltage of 15 kV, probe current of 20 nA, counting times were 20 s on peaks and 10 s on background positions.

Zircon grains are mostly elongated and prismatic (up to 300 µm long) with simple pyramids {101}, though some are fractured. Most grains are clear, lustrous and occasionally contain a small number of inclusions. Backscattered electron (BSE) images of studied zircon grains reveal an inner, apparently homogeneous interior surrounded by discontinuous, porous rims (Fig. 2). A faint oscillatory zonation is apparent in some of the grains. Zircon may contain inclusions of relatively large apatite (up to 10 µm long) and tiny epidote, albite and quartz. Cathodoluminescence images reveal a more complex internal structure (Fig. 3). The apparently homogeneous inner part of the zircon can be divided into two zones separated by a very thin and high CL emission band. The most internal zones, the cores, display oscillatory zoning (e.g. Fig. 3A,I,J), however, some rare cores show chaotic zoning, usually with very low CL emission (Fig. 3H). Most of the cores have distinct, prismatic shapes with no obvious evidence of erosion or dissolution. Outside the cores, a second, more prominent zone (mantle) can be observed. These also display pronounced oscillatory zoning in CL, sometimes accompanied by sector zoning (e.g. Fig. 3F,G). These internal zones can be observed in most of the zircons studied, but not all of them (Fig. 3C,D,L). This may be due to the intersection of the zircon not extending to its centre, or perhaps in some grains such zonation is not present. In some of crystals, zones of alteration can be observed developing within these parts of zircons, most likely along older cracks or other zones of discontinuity (Fig. 2E). The outermost, porous zones (rim) are present only in about 20% of documented crystals (Fig. 3A,C,D,E,G,L). The thickness of the rim reaches up to 20 µm. These zones are characterised by patchy zoning and usually display low CL emission. The pores are typically filled with mineral inclusions such as chlorite, titanite, quartz and scarce thorite (Fig. 2).

Figure 2.

Backscattered electron (BSE) images of studied zircon grains with visible inclusions of apatite (Ap), epidote (Ep), chlorite (Chl), albite (Ab), quartz (Qz), titanite (Ttn) and thorite (Thr). Image D represents magnified fragment from the frame on picture C. The white scale bars are 10 µm.

Figure 3.

Representative cathodoluminescence (CL) images of analysed zircon grains showing variability of zircon internal structures.

Complex zonation observed in CL images is also reflected in the distribution of trace elements in zircon (Fig. 4, Supplement S2), with the most pronounced zoning in Y. The cores, when present, can be clearly distinguished by their relatively higher Y concentration. The boundary between core and mantle, marked by a high CL emission band, corresponds to a rapid decrease in Y concentration. The mantle zones have relatively lower Y content, yet a high-Y band is present near the boundary with the porous rim. Y concentration drops again in the rims. On the other hand, Hf distribution does not display any zoning in the internal zones of zircon (i.e. in the cores and mantles) but shows elevated values in the porous rims. Uranium content, although very low in whole zircon, displays slightly higher levels in the rims. However, a detailed analysis of the U distribution shows that elevated U levels are present not only in the porous rims but often also along the edges of unaltered zircon (e.g. Fig. 4B).

Figure 4.

Backscattered electron (BSE) images, Y, Hf and U distribution maps and concentration profiles for representative zircon grains. Yellow arrows indicate the course of the profile. Blue circles represent U-Pb analytical points from Carter et al. (2023). The coloured scale bar: warmer colours indicate higher relative abundance of element (higher signal intensity). In Fig. 3 you can find CL images of the analysed crystals: A = Fig. 3.F, B = Fig. 3.J, C = Fig. 3.E.

Normalised REE diagrams presented by Carter et al. (2023; Supplementary Table S6; https://doi.org/10.6084/m9.figshare.24168223.v1) show no obvious differences between cores and mantles. However, when considered in detail in connection with the CL and BSE images some of the cores display slightly higher heavy rare earth element (HREE) concentrations (Supplement S3). Likewise, chemical maps display higher Y content in the cores. Yttrium and HREE, due to their identical valency and similar cation radii, behave similarly in zircon, thus they mimic each other’s behaviour (e.g. Zhang et al., 2019). Therefore, it can be concluded that the core zones of the zircons were enriched in Y and HREE simultaneously. Cathodoluminescence oscillatory zoning present in the cores clearly indicates their magmatic origin. The mantle zones also display prominent oscillatory zoning, thus mantles have also crystallized from a melt. However, mantles differ from the cores in their chemical composition as they have noticeably lower Y concentrations. Notably, the boundaries between these zones are usually very sharp, with no signs of core dissolution or abrasion. This indicates continuous crystallization between cores and mantles. However, the difference in their chemistry must reflect a change in crystallisation conditions. Similar Ce and Eu anomalies in both zones indicate that the magma oxidation state did not change during the transition, thus, two scenarios are possible. The change in zircon chemistry could reflect the beginning of mass crystallisation of other minerals competing for Y and HREE, most likely amphibole, but other Y-bearing minerals like apatite or titanite also may be considered (Schaltegger and Davies, 2017 and references therein). Alternatively, the change in zircon chemistry could follow a change in magma chemistry due to the mingling of different magma batches (Schaltegger and Davies, 2017 and references therein). Such changes may be very rapid; Wotzlaw et al. (2014; 2015) suggested that merging smaller, compositionally different, magma reservoirs into bigger ones may be as quick as 10² – 10³ years, so the age difference may not be recognised with most common dating methods. Our field observations of the AVF suggest evidence for at least localised mingling of acid and mafic magmas. For example at Ankarede, about 70 km south of Gränssjö, where metabasalts and meta-andesites are more common in the AVF, metabasic greenstones have streaky appearance with numerous, distinct, leucocratic felsic domains (Fig. 5). The observed alteration of zircon rims and some internal zircon parts can be linked with high U and Hf concentration in those zones. Zircon is a very durable mineral and can survive even during high-grade metamorphic conditions, so it is not likely that fully crystalline zircon would become altered in low-grade metamorphic conditions (Rubatto, 2017 and references therein). However, high U zircon may suffer from radiation damage (metamictisation); the internal structure of zircon is destroyed and thus creates pathways for hydrothermal fluids to infiltrate and alter damaged zircon. Low CL emission of these zones would also suggest lower degree of crystallinity of the rims (e.g. Nasdala et al., 2003 and references therein). It is considered that metamict zircon is a few times more susceptible to dissolution in the presence of hydrothermal fluids than zircon with no apparent radioactive damage (Ewing et al., 2003; Delattre et al., 2007). The studied rocks underwent sea floor metamorphism and subsequent low-grade (greenschist facies) metamorphism (Sjöstrand, 1978; Stephens et al., 1985), so pre-existing metamict zones were likely altered by related hydrothermal action. A new zircon would have grown during the dissolution-reprecipitation process (e.g. Geisler et al., 2007). Newly crystallised zircon would have had a high U concentration inherited from its predecessor. EMP results shows slightly lower values of Zr but higher values of Hf in the porous rims (Supplement S2), elevated Hf content at the rims can also be seen on Hf distribution maps. Such a feature is commonly observed in hydrothermal zircon (e.g. Ersay et al., 2022). However, the porous zones have a low Y content, which is consistent with the observations of Geisler et al. (2007), who noticed that domains altered by the dissolution–reprecipitation process tend to have a low trace element content. Furthermore, Huijsmans et al. (2022) confirmed that Y depletion results from the mechanisms of this zircon recrystallization process. We conclude that Y from dissolved zircon was resorbed into non-metamict zircon (back-diffusion), to form a thin, high-Y band at the boundary between the mantle and porous rim. Similar alterations, producing porous zircon at the expense of high U zircon were observed in zircon grains from different types of hydrothermally altered rocks by Corfu et al. (2003), Hay and Dempster (2009) and Hay et al. (2010). These authors also noticed that altered zircon grains were characterised by abundant cavities and nano-crystalline structure. Thus, like the cited authors, we believe that altered rims preserve a complex low-temperature hydrothermal fluids induced growth history.

Figure 5.

Examples of amphibolite-felsite mingling observed in the Ankarede Volcanite Formation of the Lower Köli Nappe Complex close to Ankarede (64.81717616° N, 14.21804096° E). The field photographs show mingled amphibolite–felsite (A), a close-up of mingled amphibolite-felsite (B) and mingled rock viewed perpendicular to a lineation - stretching direction (C).

Figure 6 provides a summary of the growth history of the studied zircon.

Figure 6.

The summary of zircon growth history: 1. High Y and HREE magmatic zircon growth (the cores; dark navy blue). 2. The abrupt change in magma chemistry due to magma mingling or the crystallisation of other minerals competing for Y and HREE. Further growth of low-Y zircon (the mantles; blue). 3. Formation of high-U rims (grey-blue). 4. The high-U rims and adjacent zones accumulate radioactive damage, their internal structure is partially damaged and pathways for fluid infiltration are created. 5. Hydrothermal fluid infiltrate metamict zones causing alterations; dissolution of zircon rim and precipitation of new, porous zircon rims. High U concentration is inherited by the newborn zircon. Yttrium is resorbed into a non-altered zircon forming a high-Y band (yellow).

Carter et al. (2023) provided U-Pb LA-ICPMS mean age of 491 ± 3 Ma for zircon cores, for the mantles most of the obtained dates are similar to the age of the cores but noticeably older dates, reaching up to ca 520 Ma, are also present. As described in the previous paragraph, the cores show no evidence of alterations, thus the age of the mantles should be the same as that of the cores. Williams et al. (1984) suggested that an anomalously old U-Pb age obtained by the SHRIMP method may be attributed to the presence of discrete zones within zircon crystals that may contain excess radiogenic Pb as a consequence of radiogenic Pb gain rather than U and Th loss. The idea of Pb mobilization and redistribution on a short scale distance within zircon crystal was supported by recent studies (Kusiak et al., 2013, 2015; Lyon et al., 2019; Ge et al., 2019) which described nano-scale Pb-enriched domains and their effects on zircon U-Pb spatial resolution geochronology i.e., laser ablation ICPMS or ion microprobe. However, all studies that present evidence for radiogenic Pb redistribution in zircon concern grains from high-temperature metamorphic rocks. Apart from these examples, Pb is generally considered to be an immobile element in zircon, except when the internal structure of zircon is partly damaged by radiation (metamictisation) or fractures. Metamict zircon becomes vulnerable to alterations even in relatively low-temperature hydrothermal conditions. The studied zircon had high-U metamict rims which were dissolved. It is possible that the adjacent zircon mantles also suffered from radiation damage, developing slight damage to the crystal lattice which would enable radiogenic Pb redistribution. Moreover, Pb mobility in low T fluids increases with fluids salinity enabling the formation of Pb and Cl complexes (Barnes and Liu, 2012; Brugger et al., 2016). Thus, during sea floor hydrothermal activity radiogenic Pb from metamict rims might have been redistributed and formed localised accumulations in the mantles, affecting their U-Pb systematics. This thesis is also supported by the fact that most of the older ages in mantles are slightly, but yet reversely, discordant (Fig. 1). The zircon cores remained intact sheltered from both radiation and hydrothermal fluids by the mantles. We, therefore, suggest that both, zircon cores and mantles formed around 491 Ma and that this age corresponds to bimodal magmatism in the Virisen arc, while the much older (>500 Ma) dates appearing in the mantles have no distinct geological significance.