Paleoproterozoic (late ‘Laxfordian’) reworking of juvenile Neoarchean Lewisian orthogneisses, Iona, Inner Hebrides, Scotland

The island of Iona lies 1 km west of Mull (Fig. 1) and comprises putative Lewisian orthogneisses and metasediments, unconformably overlain by the Iona Group metasediments (Fraser 1977; Potts et al. 1995; Zaniewski et al. 2006; McAteer et al. 2014). The orthogneisses on Iona are well exposed and exhibit clear field relationships (Fig. 2) with several generations of igneous intrusions, potentially providing a well-constrained deformation chronology. The Iona Lewisian is made up predominantly of felsic to intermediate orthogneisses, concordant amphibolite bodies, later discordant amphibolite dykes and abundant mainly granitic pegmatite sheets and dykes.

A narrow belt of approximately north–south-trending mylonites (Fig. 1; Potts et al. 1995; BGS 1999) appears to divide the island into rather distinctive lithological domains. To the east, the orthogneisses are predominantly of acidic (granodioritic) composition (sample MNH6A, Table 1) whereas they are generally more mafic and locally syenitic in the west. Moreover, a major outcrop of paragneiss (mainly pelitic metasediments) occurs in the west whereas a small sheet of carbonate rocks (quarried in the past as Iona Marble at Port Carnan a Ghille; Fig. 1) is present in the east close to a much larger locally mylonitized anorthosite body (Fraser 1977; BGS 1999).

In eastern Iona, the oldest orthogneisses are variably migmatized high-K calc-alkaline granodioritic (Fig. 2a) gneisses with granitic leucosomes, generally a few centimetres but up to c. 60 cm thick (e.g. sample MNH6B, Fig. 2b), inter-banded with centimetric to decimetric amphibolite bands (Fig. 2b). Following migmatization, the orthogneisses were variably deformed, resulting in a strong generally north–south-striking foliation and locally a subhorizontal mylonitic lineation (e.g. sample MNH6a, Fig. 2a). Relict orthopyroxene is reported from the orthogneisses (Fraser 1977) and both orthopyroxene and clinopyroxene are preserved in some basic lithologies (e.g. Dempster et al. 2021). In general, however, the acidic and intermediate orthogneisses display amphibolite-facies assemblages with biotite, hornblende and occasional garnet as the main ferromagnesian minerals. The associated fabrics are cut by a suite of amphibolite dykes (presumably one of the generations of ‘Scourie’ dykes), which when least deformed are steeply dipping and generally strike NW–SE to north–south (Fig. 2g). Locally, tight to isoclinal folds with steeply dipping, north–south- to NE–SW-striking axial planes affect the migmatized orthogneisses (Fig. 2f) but are of uncertain age relative to the amphibolite dykes. Following the development of a steeply dipping foliation in the dykes, all of the earlier structural features are cut by a widespread suite of granitic veins and pegmatites (e.g. sample MNH1, Fig. 2e). These granitic veins are generally undeformed and superficially resemble the c. 428 Ma Ross of Mull Granite that cuts the Iona Group in eastern Iona (Fig. 1; McAteer et al. 2014).

In western Iona (e.g. in the vicinity of Eilean Didil; Fig. 1), mutually conformable interlayered intermediate to basic orthogneisses are the dominant lithology. The intermediate gneisses are volumetrically dominant and at least locally are syenitic in composition (sample MNH5, Table 1, Fig. 2c). They commonly display syndeformational millimetre- to centimetre-wide stromatic to phlebitic granitic to syenitic leucosomes, which also intrude the more mafic layers, the latter typically occurring as substantial bands up to a metre across. Compared with the eastern orthogneisses, the western Iona orthogneisses are foliated but appear less deformed and lack the penetrative subhorizontal lineation seen in the east. The western orthogneisses are intruded by undeformed syenitic pegmatites (e.g. sample MNH3, Fig. 2d). Another distinguishing feature of western Iona is the presence along the northwestern coast (Fig. 1) of mainly pelitic paragneisses (BGS 1999). These contain garnet, biotite and occasional fresh orthopyroxene (Fraser 1977) indicating the preservation of granulite-facies assemblages in this area.

Both the eastern and western orthogneisses are cut by brittle dilational fractures filled with epidote in the form of sub-millimetric to decimetric epidote-bearing veins. In the west these tend to be isolated, occurring in steeply dipping, NE–SW-striking en echelon arrays (Fig. 2h) whereas in the east they also occur in the form of breccias (Fig. 2g). These epidote veins have been described in detail by Dempster et al. (2021), who attributed them to retrogression under greenschist-facies conditions.

Six samples (MNH1, MNH3, MNH5, MNH5L, MNH6A, MNH6B, Fig. 1, Supplementary material Table S1) were selected for U–Pb zircon geochronology so as to maximize the constraints that could be placed on the field-based structural and intrusive sequence. Samples MNH5 (syenitic orthogneiss) and MNH6A (deformed granodioritic orthogneiss) were dated to determine protolith ages for the orthogneisses, to provide a maximum age for the ‘Scourie’ dykes and to provide a maximum age for the deformation, metamorphism and migmatization. Sample MNH5L (c. 80% leucosome), made by extracting slices of leucosome from a split of sample MMH5, was analysed in the hope of directly dating the migmatization. Samples MNH1 (late granitic pegmatite) and MNH3 (late syenitic pegmatite) were dated to provide a minimum age for the ‘Scourie’ dykes on Iona and a minimum age for orthogneisses, their migmatization and the main ductile deformational events affecting them. Dating MNH1 and MNH3 also provides a maximum age for the mylonites and later deformation associated with the contact with the overlying Iona Group metasediments, and for their depositional age. Sample MNH6B (granitic leucosome) was dated to assess the timing of migmatization and to provide a maximum age for the ‘Scourie’ dykes and the late pegmatites. Whole-rock samples of the orthogneisses (MNH5 and MNH6A) were analysed for Sm–Nd isotopes to evaluate their crustal residence. Zircons from the orthogneisses, leucogranite (MNH6B) and one granitic pegmatite (MNH1) were analysed for Lu–Hf isotopes. Pb isotopic analyses of K-feldspar were undertaken on each of the samples to complement the Sm–Nd and Lu–Hf analyses.

Samples free of weathered material were split into c. 2–4 cm cubes, washed, dried and crushed in a Retsch tungsten carbide jaw crusher. Gravels were coned and quartered before milling to <100 $μ$m powder in a Tema agate swing mill. Sm and Nd isotopic compositions and concentrations (Table 1) were determined on whole-rock powders by isotope dilution thermal ionization mass spectrometry (ID-TIMS) using a semi-automated VG MM30 at University College Dublin following methods described by Menuge and Daly (1990) except that Sm and Nd were separated using TRU-SPEC resins. Major and trace element compositions were determined by X-ray fluorescence (XRF) spectrometry at the University of Leicester (Table 1).

Zircons were hand-picked from standard heavy liquid separates prepared from a coned and quartered gravel aliquot, mounted in epoxy resin, polished to expose grain interiors and imaged under the scanning electron microscope (Fig. 3). U–Th–Pb zircon analyses (Supplementary material Table S1) were performed on a Cameca IMS 1270 ion microprobe, following methods described by Whitehouse and Kamber (2005). U/Pb ratio calibration was based on analyses of the 1065 Ma Geostandards zircon 91500 (Wiedenbeck et al. 1995). Age calculations were made using Isoplot version 4.15 (Ludwig 2003). U–Pb data are plotted as 2$σ$ error ellipses (Fig. 4). All age errors quoted in the text are 2$σ$⁠. Following Zeck and Whitehouse (1999), common lead corrections were applied using a modern-day average terrestrial common Pb composition (i.e. ²⁰⁷Pb/²⁰⁶Pb = 0.83; Stacey and Kramers 1975), where statistically significant ²⁰⁴Pb counts were recorded (Table S1).

In situ measurements of Lu–Hf isotopes in zircon (Table S2) were performed at the NERC Isotope Geosciences Laboratory (NIGL) in Nottingham, UK (sample MNH1), Memorial University, Canada (samples MNH6a and MNH6b) and at UCD (samples MNH1, MNH3 and MNH5). Lu–Hf data from sample MNH1 were obtained at NIGL following Gagnevin et al. (2011) using a Nu Instruments Nu-Plasma HR multiple-collector inductively coupled plasma mass spectrometry (MC-ICP-MS) system coupled to a solid-state 193 nm wavelength Nd:YAG laser ablation system (UP193SS, New Wave Research). Analyses at Memorial University were performed using a GeoLas laser ablation system coupled to a Thermo Finnigan Neptune MC-ICP-MS system. Laser ablation was carried out in spot mode using a Lambda Physik Complex Pro ArF excimer laser operating at a wavelength of 193 nm, a fluence of 5 J cm⁻², a repetition rate of 10 Hz and a spot size of 49 $μ$m. Details of the method have been given by Valley et al. (2010). Lu–Hf isotope analyses at UCD were performed using a Thermo Scientific Neptune MC-ICP-MS system coupled to a New Wave Research UP193 excimer laser ablation system. Details of the method have been given by Lancaster et al. (2017). As far as possible, Lu–Hf analytical laser spots were located to coincide with ion microprobe spots used for U–Pb analyses. Replicate analyses of the same grain domain undertaken at NIGL and UCD agreed within analytical uncertainty (Table S2).

In situ Pb isotopic analyses of K-feldspar (Table S3) were performed using a Teledyne Cetac G2 193 nm excimer laser ablation system attached to a Thermo Scientific Neptune MC-ICP-MS system at the National Centre for Isotope Geochemistry, University College Dublin, Ireland, following procedures similar to those described by Tyrrell et al. (2012).

MNH6A is a strongly deformed granodioritic orthogneiss (Fig. 2a) containing heavily sericitized plagioclase, quartz, K-feldspar, chlorite pseudomorphs after hornblende, epidote and accessory zircon, titanite and apatite. It has a whole-rock Sm–Nd model age of 2737 Ma (Table 1). Zircons are generally rounded prisms with CL-darker, generally idiomorphically zoned cores, overgrown by several generations of thin, anhedral rims comprising predominantly CL-bright bands sometimes alternating with thinner CL-dark bands (Fig. 3c). Several grains exhibit small irregular inner cores, up to 50 $μ$m across, that are generally CL-light. One of these yielded the oldest analysis with a concordia age of 2737 ± 20 Ma. Thirteen analyses, mainly from the idiomorphically zoned cores of 12 grains, define a discordia (Fig. 4a) with intercepts at 2716 ± 21 and 1773 ± 54 Ma (MSWD = 1.4). The upper intercept age, 2716 ± 21 Ma, is interpreted as the magmatic age of the orthogneiss protolith. Two concordant analyses that define the lower intercept age are from a CL-bright rim (18a) and the interior of a CL-bright homogeneous grain (71a). These independently yield a concordia age of 1761 ± 20 Ma (MSWD = 0.016). Within error they overlap with the main group of data from pegmatite MNH1 (see below). Another analysis (18c, Fig. 3c) from the intermittent outermost CL-bright rim of grain 18 yields a concordia age of 1608 ± 35 Ma (see discussion below). The discordance of the remaining seven analyses is attributed to varying degrees of recent lead loss.

Zircon cores have an average $ε$Hf₂₇₁₆ value of −0.24 ± 1.57 (1$σ$⁠) at the preferred protolith age of c. 2716 Ma (Table S2, Fig. 5). An analysis of the rim of one these grains (analysis 18A, Fig. 3c) has an $ε$Hf₁₇₆₁ value of −14.0, calculated using its ²⁰⁷Pb/²⁰⁶Pb age of 1761 Ma (Table S2).

MNH6B is a weakly deformed concordant leucogranitic vein about 60 cm across that intrudes orthogneiss similar to MNH6A (Fig. 2b). It contains K-feldspar, quartz, plagioclase, allanite–epidote intergrowths, chlorite–epidote–opaque pseudomorphs after biotite and/or hornblende as well as accessory zircon. The rock (as well as orthogneiss MNH6A, Fig. 2a) is cut by conspicuous anastomosing millimetre-thick breccia bands, the matrix of which contains predominantly epidote together with lesser amounts of chlorite and possibly stilpnomelane as neomorphic phases. The zircons comprise subhedral, rounded prisms up to 300 $μ$m long as well as generally larger rounded and squat prisms up to 500 $μ$m across. Most grains have CL-bright rims of variable thickness often discordant to variably zoned cores that sometimes display clear resorption surfaces (e.g. grain 25, Fig. 3d).

Six analyses (22, 25, 28a, 28b, S103a, S111a) from five zircon grains yielded Archean ²⁰⁷Pb/²⁰⁶Pb ages (Table S1). These were obtained from variably idiomorphically zoned resorbed, rounded cores of zircon grains (e.g. grain 25, Fig. 3d), overgrown by thin discordant CL-bright rims (e.g. grain 22, Fig. 3d). Two spots from these resorbed cores (spots 25 and 28b, Table S1) define a concordia age of 2664 ± 9 Ma, and another analysis from a similar grain domain (spot 22, Table S1) plots close to the other two and is slightly discordant, probably owing to recent lead loss (Fig. 4b). All three yield a weighted average ²⁰⁷Pb/²⁰⁶Pb age of 2674 ± 24 Ma (MSWD = 3.1). A fourth analysis evidently from a similar inherited Archean grain plots on concordia but is probably an inadvertent mixture of core and rim (spot S111a, Table S1). The grains yielding Archean ages are interpreted as inherited from the surrounding orthogneiss.

A second group of analyses yielded concordant or near-concordant Paleoproterozoic ages (Fig. 4b). Excluding two reversely discordant grains (S107a, S109a), nine of these (six of which are rims overgrowing resorbed cores) yield a weighted mean ²⁰⁷Pb/²⁰⁶Pb age (Fig. 4b, inset) of 1779 ± 10 Ma (MSWD = 1.4), which is interpreted as the crystallization age of the vein. Eight of these (excluding grain 26, which exhibits some Pb loss) yielded an identical concordia age of 1782 ± 13 Ma (95% confidence errors). Three zircon rims (S102a, S106a, S114b) that plot close to or on concordia between c. 1.8 and 2.7 Ga (Fig. 4b) are interpreted as inadvertent mixtures.

Lu–Hf isotopic data from two Paleoproterozoic grains yield $ε$Hf₁₇₇₉ values of −6.8 and –10.8 (Table S2, Fig. 5). Six analyses of resorbed cores yield average $ε$Hf₂₆₇₄ values of –0.3 ± 1.2.

Sample MNH1 (Fig. 2e, Table S2) is an irregular (non-planar) undeformed discordant leucogranitic pegmatite about 10 cm in width. It cuts the foliation and folded leucocratic veins in the enclosing granodioritic orthogneiss, which is considered to be equivalent to sample MNH6A. In addition to quartz, K-feldspar and minor plagioclase, it contains subhedral prismatic zircons up to 400 $μ$m long typically with CL-dark, weakly idiomorphically zoned, resorbed cores overgrown by CL-brighter rims (Fig. 3a). Sixteen near-concordant analyses of zircon rims or homogeneous grains have a weighted mean ²⁰⁷Pb/²⁰⁶Pb age (Fig. 4c) of 1753 ± 8 Ma (MSWD = 1) interpreted as dating the crystallization of the pegmatite and thus providing a minimum age for the deformation affecting the orthogneiss. Analyses of four generally CL-dark cores (analyses 33a, 35a, @1 and @4) are discordant and are interpreted as inherited grains, probably affected by ancient Pb loss during metamorphism (see below). The older two of these plot close to concordia and have a weighted mean ²⁰⁷Pb/²⁰⁶Pb age of 2635 ± 5 Ma. Analysis 35b intersects concordia and is interpreted as an inadvertent mixture of the core and rim of this grain. Finally, analyses @10 and @13 from grain a10 both intersect concordia and define a weighted mean ²⁰⁷Pb/²⁰⁶Pb age of 1703 ± 26 Ma. Cathodoluminescence imaging after analysis indicated that these analyses have sampled healed fractures so that it is possible that this young age is geologically meaningful (see discussion below).

Laser ablation Lu–Hf analyses (Table S2, Fig. 5) of 12 zircon rims (dated at c. 1753 Ma) yield a mean $ε$Hf₁₇₅₃ value of  –14.2 ± 1.4. Analyses of four dated and two undated cores (25B, 8C, Table S2) have an average $ε$Hf₂₆₃₅ value of –1.1 ± 1.3 at their likely minimum crystallization age of c. 2635 Ma (Fig. 5).

MNH5 is a dark intermediate (syenitic) orthogneiss from western Iona (Figs 1 and 2c) containing K-feldspar, quartz, variably sericitized plagioclase, hornblende (altering to chlorite), biotite, ilmenite and zircon. It has a whole-rock Sm–Nd model age of 2839 Ma (Table 1).

Millimetre- to centimetre-wide stromatic to phlebitic granitic to syenitic leucosomes occur widely in the syenitic orthogneisses. Zircons were also separated from a subsample (MNH5L) of orthogneiss estimated to contain 80% leucosome and 20% mesosome.

Zircons are strongly rounded, resorbed grains up to 350 µm long, with aspect ratios of 1.6–2.4. Most grains display idiomorphically zoned interiors, sometimes with distinct cores and rims, separated by resorption surfaces. Most grains also have very thin CL-bright discontinuous anhedral overgrowths (Fig. 3b). Most of the zircon U–Pb data (Table S1, Fig. 4d) occur in two groups. An older group, mainly from the orthogneiss but also from the leucosome, is dispersed along concordia between 2740 and 2610 Ma, whereas a second group, found entirely in subsample MNH5L, forms a cluster of concordant or nearly concordant data points at c. 1740 Ma. Eight analyses with the oldest ²⁰⁷Pb/²⁰⁶Pb ages (38a, 51, a10, 48a, 75a, a8, 71b and 16a, Table S1) are concordant or nearly concordant and yield a weighted average ²⁰⁷Pb/²⁰⁶Pb age of 2703 ± 5 Ma. Five of these define a concordia age of 2704 ± 6 Ma (MSWD = 0.24), interpreted as dating the protolith. Five concordant points (a1, a4 and a6 from the leucosome together with 71a and 3b from the host gneiss) define a concordia age of 2638 ± 6 Ma (MSWD = 0.76), tentatively interpreted as dating a metamorphic event.

The three youngest concordant analyses (a14, a13, a12), exclusively from the leucosome, define a concordia age of 1736 ± 23 Ma (95% confidence errors, MSWD = 3.8), interpreted as dating the formation of the leucosomes. All of the data define a discordia (MSWD = 3.2) with intercepts at 2699 ± 25 and 1791 ± 44 Ma (Table S1), consistent with c. 2.7 Ga zircons from both the orthogneiss and leucosomes experiencing varying degrees of lead loss during metamorphism at c. 1740 Ma.

Laser ablation Lu–Hf analyses are available only from the gneiss (MNH5, Table S2, Fig. 5). These yield a mean $ε$Hf₂₇₀₄ value of 0.5 ± 0.8.

MNH3 is a brick red syenitic pegmatite (Fig. 2d) that cuts the migmatitic leucosomes and foliation in a syenitic orthogneiss similar to MNH5, and is located c. 200 m to the SE of it. It contains perthitic orthoclase, hedenbergitic clinopyroxene (variably altered to epidote), variably sericitized plagioclase, minor amounts of quartz (mainly as inclusions in orthoclase), hornblende (altering to chlorite), biotite, ilmenite and zircon.

Zircons in MNH3 are not abundant, strongly rounded, locally resorbed grains up to 350 µm long, with aspect ratios of 1.6–2.4. Most grains display simple zoning with generally CL-darker interiors and Cl-lighter rim regions. Five analyses from four grains are concordant (Table S1) and define a concordia age of 1688 ± 8 Ma (MSWD = 0.112), which is interpreted as the crystallization age of the syenitic pegmatite (Fig. 4e). Analysis 11 is reversely discordant although it overlaps with four of the points defining the 1688 Ma age. This analysis is from a zircon grain with a complex CL pattern and possibly represents an inadvertent mixture of an older partly CL-bright core and a younger CL-dark rim. Another four analyses are displaced from concordia in a manner consistent with recent lead loss (Fig. 4e), three of which have ²⁰⁷Pb/²⁰⁶Pb ages indistinguishable from the main group.

Pb isotopes were measured in situ in K-feldspar from all six samples (Table S3, Fig. 6). Measuring Pb isotopes in K-feldspar has the advantage of potentially being time-invariant owing to low U and Th contents in the feldspar minerals. However, these Pb values are, of course, subject to resetting during metamorphism (e.g. Whitehouse 1989) and therefore may not represent initial ratios. With the exception of the c. 1688 Ma syenitic pegmatite (MNH3, Fig. 6), uranogenic Pb isotopic compositions are tightly grouped with ²⁰⁶Pb/²⁰⁴Pb = 13.678–13.836 and ²⁰⁷Pb/²⁰⁴Pb = 14.635–14.683. There is some spread in thorogenic Pb between ²⁰⁸Pb/²⁰⁴Pb of 33.574 and 34.172 with the pegmatite MNH1 displaying less radiogenic values (Fig. 6). Compared with the other samples, the syenite pegmatite (MNH3) has consistently more radiogenic values of ²⁰⁶Pb/²⁰⁴Pb = 14.544–14.876, ²⁰⁷Pb/²⁰⁴Pb = 14.818–14.897 and ²⁰⁸Pb/²⁰⁴Pb = 34.642–35.245.

The protoliths of the syenitic and granodioritic orthogneisses on Iona have yielded U–Pb zircon crystallization ages of 2704 ± 6 and 2716 ± 21 Ma respectively. These are among the youngest Archean orthogneisses so far reported from the Lewisian Complex. No significantly older inherited zircons were detected, consistent with Sm–Nd model ages of 2839 and 2737 Ma for the syenitic (MNH5) and granodioritic (MNH6A) orthogneisses respectively (Table 1); that is, suggesting that they represent largely juvenile Neoarchean crust. Average zircon $ε$Hf_t values for these rocks (0.5 ± 0.8 and −0.24 ± 1.6) and maximum values (+2 and +1.5) respectively are similar to, although falling with the lower range of, those from c. 2.9–2.7 Ga Lewisian orthogneisses from both the mainland Lewisian and the Outer Hebrides (Fig. 5; Whitehouse and Kemp 2010; Whitehouse et al. 2022). The geotectonic significance of the Hf isotopic data is unclear, as the Lu–Hf characteristics of the Neoarchean mantle are still poorly known. However, the $ε$Hf_t values of the Iona rocks are very close (Fig. 5) to the ‘mildly depleted’ mantle model of Whitehouse et al. (2022).

Neoarchean metamorphism possibly took place after formation of the granodioritic and syenitic protoliths at c. 2638 Ma (selected zircon grains in the syenitic orthogneiss). This is interpreted to be associated with a foliation, assumed to be of Neoarchean age but not yet directly dated, that is cut by a suite of amphibolite dykes tentatively correlated with one of the ‘Scourie’ swarms.

The development of the leucogranite vein (MNH6B) at 1779 ± 10 Ma is interpreted to date ‘Laxfordian’ high-grade metamorphism and melting probably reaching granulite-facies conditions. These leucosomes are deformed by tight to isoclinal folds that result in a widely developed composite transposed fabric that also affects the ‘Scourie’ amphibolite dykes. This represents the final high-grade event affecting the Iona Lewisian. An imprecise age (c. 1736 Ma) obtained from syntectonic leucosomes in the syenitic orthogneiss (MNH5L) from western Iona suggests a younger age for the Laxfordian metamorphism but given its large uncertainty (±23 Myr), we interpret the late Laxfordian to have ceased as a ductile high-grade event by c. 1753 Ma. This is based on the observation that the youngest ductile structures in eastern Iona are cut by pegmatitic granite sheets (e.g. MNH1, dated at 1753 ± 8 Ma), which are deformed only locally by the mylonitization and brittle fractures. Hence neither the Neoarchean protolith ages of the orthogneisses nor their Laxfordian history supports a terrane distinction between eastern and western Iona. Dating of the early migmatitic leucogranite at c. 1779 Ma and the undeformed c. 1753 Ma pegmatite thus brackets the ‘late Laxfordian’ deformation in eastern Iona between 1779 ± 10 and 1753 ± 8 Ma.

The ‘late Laxfordian’ deformation interval on Iona (c. 1779–1753 Ma) possibly corresponds to ‘D3’ of Park (2005). Late Laxfordian metamorphic events between 1790 and 1770 Ma recorded from mainland Lewisian rocks in the vicinity of the Laxford Shear Zone (Goodenough et al. 2013) indicate that metamorphism at this time was widespread and affected the entire Lewisian. An important distinction is that the late Laxfordian on Iona manifests as the major deformational event and is associated with migmatization. On a wider scale, the late Laxfordian is probably related to the assembly of the Nuna/Columbia supercontinent by c. 1.75 Ga (see Zhang et al. 2012). This event probably corresponds to the accretion of the Rhinns Complex (Malin Block) to the margin of the Nuna/Columbia supercontinent by 1.75 Ga (see Zhang et al. 2012). Our results from Iona are consistent with geochronological age constraints from the Rhinns Complex where the ‘late Laxfordian’ deformation is constrained less precisely between 1779 and 1710 Ma (Roddick and Max 1983; Daly et al. 1991).

Zircons that grew during Laxfordian migmatization (MNH6B) and those within the later pegmatite (MNH1) have Hf isotope signatures consistent with derivation from Neoarchean crust (Fig. 5), presumably the local orthogneisses, assuming a range of crustal whole-rock values of the ¹⁷⁶Lu/¹⁷⁷Hf ratio between 0.015 and 0.020. This is supported by the age and Hf isotopic composition of the inherited grains within both the pegmatite and migmatitic leucosomes and by the similar-aged neomorphic zircon in the orthogneiss. No significant addition of a juvenile Paleoproterozoic contribution to these rocks is required.

Despite the nearly one billion year age difference between the Neoarchean orthogneisses and the Paleoproterozoic leucogranite and granitic pegmatite, the tight grouping of their uranogenic Pb isotopic data (Fig. 6a) suggests that the orthogneiss K-feldspars have been reset during Laxfordian metamorphism. Significantly, although the pegmatite (MNH1) has similar uranogenic values, it displays less radiogenic ²⁰⁸Pb/²⁰⁴Pb values (Fig. 6b), potentially indicating fractionation of a Th-rich phase during its formation or the involvement of a geochemically distinct source component.

The traditional correlation of the Iona Lewisian with that of Coll and Tiree (see Park 2005) is consistent with the results of this study. In particular, the 1753 ± 8 Ma age of the granitic pegmatite (MNH1) permits correlation of the Iona amphibolites with the presumed ‘Scourie’ dykes on Tiree, which have a minimum age of c. 1.75 Ga, based on Sm–Nd mineral isochrons (Muir et al. 1993). Moreover, the whole-rock Sm–Nd data for the Iona orthogneisses are similar to those of the SW Tiree granulite complex and the Caoles migmatites in eastern Tiree (Whitehouse and Robertson 1995).

Kinny et al. (2005) and Love et al. (2010) included Coll and Tiree in their ‘Rona terrane’. However, there are important differences between the Rona terrane gneisses and those on Iona. The Rona terrane contains the oldest rocks dated within the Lewisian Complex (3135 ± 5 Ma tonalitic gneiss). Moreover, whereas the Rona terrane on the Scottish mainland contains evidence of melting and possibly tonalite gneiss formation at c. 2700 Ma (Love et al. 2010), consistent with a correlation with Iona, the Rona terrane tonalitic rocks were probably derived from an older (c. 3130 Ma) protolith (Whitehouse et al. 2022; Fig. 5) and hence lack the juvenile character of the Iona and Tiree Neoarchean orthogneisses.

The post-tectonic intrusion of the syenitic pegmatite (MNH3) at c. 1688 Ma is the youngest magmatic event dated in this study. No Lu–Hf data are available but the Pb isotope characteristics of this sample (Fig. 6) suggest a different magmatic source. Uranogenic Pb isotope values for this pegmatite fall within the range of K-feldspar values determined by Whitehouse (1989) for K-feldspars from Neoarchean orthogneisses from the Rhichonich terrane, which were interpreted as having been reset during Laxfordian metamorphism. However, because the syenitic pegmatite is neither deformed nor significantly metamorphosed, its feldspar Pb isotopic data are considered primary, suggesting a (mantle) source that is distinctive from the surrounding orthogneisses.

The apparent Fe-rich, alkaline character and undeformed state of the c. 1688 Ma syenite suggests an anorogenic setting by this time. By contrast, Park (2022) considered that the Lewisian was experiencing north–south convergence resulting in widespread folding and dextral shear zones at this time, attributing the time interval of c. 1.694–1.675 Ma to the ‘late Laxfordian’, apparently in conflict with the conclusions of this study. However, much of the evidence for continuing deformation in this time interval appears to relate to movements along discrete shear zones; for example, involving syndeformational pegmatites dated at 1681 ± 2 and 1669 ± 3 Ma of the Langavat shear zone on South Harris (Kelly et al. 2008) and affecting the syntectonic Tollie pegmatite in the Loch Maree area at 1694 ± 5 Ma (Park et al. 2001).

Thermal effects are widespread in the Lewisian Complex near the end of the Paleoproterozoic (e.g. the formation of hydrothermal titanite at Laxford Bridge between 1690 and 1670 Ma; Corfu et al. 1994) with temporally corresponding intrusion of a wide variety of small-scale granitoid veins and dykes (e.g. in South Harris at 1704 ± 3 Ma; Kelly et al. 2008; on South Uist at 1702 ± 10 Ma; Whitehouse 2003; and on Lewis at 1683 ± 9 Ma; Friend and Kinny 2001). Pegmatites associated with rare-metal enrichments especially on South Harris were also emplaced close to c. 1700 Ma (e.g. Shaw et al. 2016). The youngest zircon component (1703 ± 23 Ma) associated with healed fractures in pegmatite MNH1 falls within this time interval.

The youngest (c. 1608 Ma) zircon component detected in orthogneiss MNH6A is tentatively interpreted as a static metamorphic event, possibly also represented by numerous very thin (<5 $μ$m) CL-bright zircon overgrowths (Fig. 3c), too small to date by conventional ion microprobe methods. There is as yet no evidence of significant metamorphism in Iona at this time. However, titanites from a marble at Gott Bay on Tiree yielded a discordia with an upper intercept at 1593 ± 11 Ma and non-zero lower intercept at 330 ± 157 Ma (Parnell et al. 2022). Four concordant titanite analyses from the Parnell et al. (2022) study yield a concordia age of 1586 ± 7 Ma.

The Iona Group rests unconformably on the Lewisian orthogneisses and was deposited after c. 1490 Ma, based on the age of its youngest detrital zircon and titanite grains (McAteer et al. 2014). This is consistent with the c. 1753 Ma age for the younger pegmatites and the c. 1608 Ma age of the youngest increment of zircon growth in the granodioritic orthogneiss (sample MNH6A).

A Paleoproterozoic age for the Iona marble is suggested by Sm–Nd model ages of 2.17–2.52 Ga (Whitehouse and Russell 1997) but otherwise there are no published isotopic age constraints on the Iona metasediments. The c. 1.42 Ga metamorphism suggested by the Pb–Pb isochron for the Iona marble (Whitehouse and Russell 1997) is not manifest in the U–Pb zircon data from the Iona orthogneisses or pegmatites. Moreover, there is no evidence for a high-grade Grenville event (c. 1.1–1.0 Ga) on Iona as recently reported on Lewis (Metcalfe et al. 2024). Extensive brecciation manifest as epidote-bearing veins could represent brittle Grenvillian deformation as recorded within the Lewisian in the Gairloch area (Sherlock et al. 2008) and or possibly be of Renlandian age (0.95–0.93 Ga) as suggested by Krabbendam et al. (2022). These brittle features remain undated. However, they affect the c. 1688 Ma pegmatites and also cut at least the stratigraphically lower part the overlying Iona Group, which has been correlated with the Dalradian Grampian Group by McAteer et al. (2014) and more recently with the early Neoproterozoic Wester Ross Supergroup by Krabbendam et al. (2022).

Juvenile Neoarchean orthogneisses on Iona formed at c. 2710 Ma from a mildly depleted mantle source suggested by zircon Lu–Hf data. These rocks were extensively reworked (possibly under granulite-facies conditions) during the ‘late Laxfordian’ deformational event between 1779 and 1753 Ma, bracketed between pre-tectonic migmatization and leucogranite formation and post-tectonic granitic pegmatite intrusion ages. Despite lithological differences (siliciclastic metasediments and more alkalic orthogneiss compositions in the west, calc-alkaline orthogneisses in the east) between eastern and western Iona, we conclude that the mylonites between them do not constitute a terrane boundary. Zircon Lu–Hf data indicate that late Laxfordian leucogranites and granitic pegmatites were largely derived from the Neoarchean orthogneisses although thorogenic K-feldspar Pb isotope data also suggest fractionation of a Th-rich phase in the genesis of the granitic pegmatites. K-feldspar Pb isotope ratios in the orthogneisses have probably been reset during late Laxfordian metamorphism whereas those in a post-tectonic syenitic pegmatite are considered to be original and consistent with an exotic source. The traditional correlation of Iona with the Lewisian of Coll and Tiree is supported by this work but the younger age and juvenile character of the Neoarchean orthogneisses on Iona do not support correlation with the Rona terrane.

We are grateful to R. Strachan and R. Palin for concise and helpful reviews, and to I. Neill for editorial handling. T. Culligan and M. Murphy are thanked for technical assistance at University College Dublin. J. Faithfull (Hunterian Museum, Glasgow) kindly provided access to F. Fraser's maps of Iona and engaged in stimulating discussion. R. Lam (Memorial University, Newfoundland, Canada) and M. Horstwood (NIGL Keyworth, UK) are thanked for assistance with Lu–Hf analyses. N. Marsh and the late T. Brewer are thanked for XRF analyses at the University of Leicester. Thanks also go to L. Ilyinsky, K. Lindén and C. Kirkland at the NordSIMS laboratory, Stockholm. D. Chew is thanked for instigating the sample nomenclature adopted in this study. We also thank J. Wagstaff for inspiring discussion and hospitality, and J. Reavy for introducing us to this wonderful man.

JSD: conceptualization (lead), formal analysis (lead), investigation (lead), project administration (lead), writing – original draft (lead), writing – review & editing (lead); MJF: conceptualization (supporting), formal analysis (supporting), investigation (supporting), project administration (supporting), writing – original draft (supporting), writing – review & editing (supporting); MJW: formal analysis (supporting), investigation (supporting), methodology (lead), resources (supporting), writing – review & editing (supporting); EB: formal analysis (supporting), methodology (supporting), writing – review & editing (supporting)

This research was funded by IRCSET Basic Research grant SC/2002/248, by Enterprise Ireland International Collaboration Travel Support Grant IC/2006/33 and by a UCD Seed Funding Grant, all awarded to J.S.D. J.S.D. and E.B. gratefully acknowledge Science Foundation Ireland grants 13/RC/2092 and 13/RC/2092_P2, which were co-funded under the European Regional Development Fund. During the course of this work the NordSIMS facility operated as a Nordic infrastructure supported by the research funding agencies of Denmark, Norway and Sweden and the Geological Survey of Finland. This is NordSIMS contribution 764.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

All data generated or analysed during this study are included in this published article and its supplementary information files.