1. Introduction
The Lewisian Complex of northern Britain has been very extensively studied for over 200 years (MacCulloch, Reference MacCulloch1819; Peach et al. Reference Peach, Horne, Gunn, Clough, Hinxman and Teall1907; Park & Tarney, Reference Park and Tarney1987; Mendum et al. Reference Mendum, Barber, Butler, Flinn, Goodenough, Krabbendam, Park and Stewart2009). The complex consists particularly of tonalitic gneisses of Archaean age, derived from an igneous protolith. In common with Precambrian basement elsewhere in the North Atlantic region (Fig. 1), the Lewisian Complex also includes local supracrustal successions of metasediment (Fig. 2). The supracrustal successions contain a range of chemical sediments, including graphitic schists, ironstones and marbles, which where dated (Whitehouse & Bridgwater, Reference Whitehouse and Bridgwater2001; Park, Reference Park2002) are mid-Palaeoproterozoic (∼1.9–2.0 Ga). Across the region, from North America to Russia (Fig. 1; Table S1 in the Supplementary Material available online at https://doi.org/10.1017/S0016756822000474), these successions have evidence of former evaporites in the form of surviving sulphate minerals, metamorphic minerals with a signature of evaporitic seawater, and sulphur isotope data. To date, however, no such evidence is reported from Britain. Here we report evidence for signatures that would be consistent with evaporitic facies in the Lewisian Complex of Scotland. The uniformity of facies assemblage across the region, together with a setting related to an accretionary plate boundary (Park, Reference Park2002), implies that evaporitic facies were marine rather than lacustrine.
2. Geological setting and methods
Samples were collected from supracrustal rocks in the Lewisian Complex on the island of Tiree (Fig. 2). The supracrustal rocks on Tiree include garnet schists, graphitic schists, sandstones, ironstones and marbles (Westbrook, Reference Westbrook1972; Whitehouse & Russell, Reference Whitehouse and Russell1997). The best exposed sections are at Vaul (National Grid Reference NM 048490), Balephetrish (NM 014473) and Gott (NM 044459). They are metamorphosed to amphibolite facies and sheared, but they preserve detrital mineralogy. Estimated P–T conditions are 10.5 ± 1.5 kbar and 810 ± 50 °C (Cartwright, Reference Cartwright1992).
Mineral phenocrysts are especially abundant in the Tiree marbles (Fig. 3). The inclusions were investigated using scanning electron microscopy, conducted in the ACEMAC facility at the University of Aberdeen.
The mineralogy of the Lewisian Complex elsewhere in Scotland is affected by sodium metasomatism which also affected the Neoproterozoic Moine Supergroup (Sutton & Watson, Reference Sutton and Watson1951; May et al. Reference May, Peacock, Smith and Barber1993). An imprint by Neoproterozoic or younger fluids would exclude interpretations of a Lewisian protolith, and must be tested. Accordingly, the assemblage of mineral phenocrysts was dated using U–Pb analysis of titanite crystals. U–Pb isotope analyses were done using the laser ablation – inductively coupled plasma – mass spectrometry (LA-ICP-MS) laboratory at Lund University, where a Teledyne Photon Machines G2 laser is coupled to a Bruker Aurora Elite quadrupole ICP-MS. The laser is equipped with a HelEx two-volume sample cell with an energy meter. Instrument tuning, using NIST612, was aimed at obtaining high and stable signal counts on lead isotopes, on low oxide production (below 0.5 % monitoring 238U/238U16O and 232Th/232Th16O) and on Th/U ratios around 1. Standard-sample-standard bracketing incorporated the natural titanite MKED1 (Spandler et al. Reference Spandler, Hammerli, Sha, Hilbert-Wolf, Hu, Roberts and Schmitz2016) as primary reference material, and natural titanite ONT2 (Spencer et al. Reference Spencer, Haker, Kyander-Clark, Andersen, Cottle, Stearns, Poletti and Seward2013) as a secondary standard. Each analysis was made with 300 shots at 10 Hz with a fluence of 1.5 J cm−2. Baseline compositions were measured for 30 s before each measurement, and subtraction was done with a step-forward approach. Common Pb was monitored by measuring 202Hg and mass 204 (204Hg + 204Pb). Baseline levels on mass 204 were c. 440 cps with a standard error (SE) around 20 cps (5–6 %). Data reduction was done with iolite using the X_U_Pb_Geochron4 DSR (Paton et al. Reference Paton, Woodhead, Hellstrom, Hergt, Greig and Maas2010, Reference Paton, Hellstrom, Paul, Woodhead and Hergt2011), and the common Pb correction was done using the VizualAge DRS (Petrus & Kamber, Reference Petrus and Kamber2012). Plotting and final age calculations were done with Isoplot(R); intercept ages are based on ‘model 1’ calculations, and errors confidence levels are 95 % with overdispersion (Vermeesch, Reference Vermeesch2018).
For sulphur isotope analysis, pyrite samples were combusted with excess Cu2O at 1075 °C in order to liberate the SO2 gas under vacuum conditions. Liberated SO2 gases were analysed on a VG Isotech SIRA II mass spectrometer, with standard corrections applied to raw δ66SO2 values to produce true δ34S. The standards employed were the international standard NBS-123, IAEA-S-3 and SUERC standard CP-1.
3. Criteria
Several criteria have been proposed for the identification of replaced evaporites in Precambrian successions (Warren, Reference Warren2016). The criteria are based particularly on the chemistry of mineral phenocrysts that developed during metamorphism up to hundreds of millions of years after sedimentation (Moine et al. Reference Moine, Sauvan and Jarousse1981; Warren, Reference Warren2016; Hammerli & Rubenach, Reference Hammerli, Rubenach, Harlov and Aranovich2018). The evidence is most convincing where there are multiple positive criteria. The most direct evidence is:
(i) The survival of evaporite minerals anhydrite, gypsum or halite.
(ii) Pseudomorphs of evaporite minerals, most distinctively gypsum and halite (e.g. Ririe, Reference Ririe1989; Zentmyer et al. Reference Zentmyer, Pufahl, James and Hiatt2011).
Evidence for former evaporites also includes the growth of specific minerals containing chlorine or boron derived from seawater:
(iii) Phlogopite mica, which contains a high magnesium content and traces of chlorine (Schreyer et al. Reference Schreyer, Abraham and Kulke1980; Moine et al. Reference Moine, Sauvan and Jarousse1981).
(iv) Minerals indicative of sodium metasomatism, including albite and especially scapolite, which contains traces of sulphur and/or chlorine (Mora & Valley, Reference Mora and Valley1989). High sulphur levels in particular reflect assimilation of sulphate evaporites (Morrissey & Tomkins Reference Morrissey and Tomkins2020; Zeng et al. Reference Zeng, Zhao, Hammerli, Fan and Spandler2020), while high chlorine levels are also measured in scapolite from skarn deposits (Mora & Valley, Reference Mora and Valley1989).
(v) Tourmaline, which is the major reservoir of boron in meta-evaporitic rocks (Henry et al. Reference Henry, Sun, Slack and Dutrow2008; Riehl & Cabral, Reference Riehl and Cabral2018). Tourmaline is often restricted to beds containing scapolite assumed to be derived from saline fluids (Mora & Valley, Reference Mora and Valley1989).
Corroborative evidence includes:
(vi) Chlorine-rich apatite, which could be derived from a magmatic source or seawater, and which is positive evidence where a magmatic input is lacking (Mao et al. Reference Mao, Rukhlov, Rowins, Spence and Coogan2016).
(vii) Sulphur isotope compositions of pyrite, which are comparable to the heavy composition of evaporites/seawater rather than the near-zero composition of magmatic sulphur (Golani et al. Reference Golani, Pandit, Sial, Fallick, Ferreira and Roy2002; Johnston et al. Reference Johnston, Poulton, Fralick, Wing, Canfield and Farquhar2006).
(viii) High-salinity fluids, especially in ore deposits that formed during or shortly after sedimentation (e.g. Conliffe et al. Reference Conliffe, Wilton, Blamey and Archibald2013).
4. Results
Mineral phenocrysts in the Tiree marbles (Figs 3 and 4) are dominated by pyroxenes (enstatite, diopside), amphiboles (tremolite), olivine (forsterite), micas and feldspars, and also scapolite, titanite (sphene), apatite, epidote, pyrite and quartz.
Minerals present in trace amounts include the sulphates anhydrite and barite. Anhydrite occurs as crystal fragments up to 15 μm size, in a calcite-rich groundmass. Barite occurs as a disseminated overprint on phenocrysts of pyroxene and amphibole. Measurements of the sulphur isotope composition of five pyrite crystals in marble at Gott yielded closely clustered values of 11.7, 11.7, 11.7, 11.9 and 12.3 ‰ (Fig. 5).
Sodium metasomatism is represented by replacive albite and scapolite. Scapolite has been recognized in the Tiree marble since the first petrographic studies (Coomaraswamy, Reference Coomaraswamy1903; Hallimond, Reference Hallimond1947), and the Lewisian supracrustal rocks are conspicuously richer in scapolite than other rocks in Britain (Flett, Reference Flett1907). The anion chemistry of scapolite includes variable combinations of −Cl, −F, −S and −OH, where proportions of −Cl and −S are interpreted to suggest relative contributions from replaced halite and gypsum (Warren, Reference Warren2016). Sulphur and chlorine were measured up to 2.30 % and 0.93 % respectively in the scapolite (Table S2, in the Supplementary Material available online at https://doi.org/10.1017/S0016756822000474). The higher sulphur contents occur in scapolite with the lower chlorine contents, as expected when both contribute to the same atomic sum. The marble at Gott, and to a lesser extent in other Tiree marble, is partially altered to masses of albite.
Phlogopite occurs pervasively through the supracrustal marbles on Tiree and in other Scottish localities. The phlogopite consistently contains 0.15 to 0.25 wt % chlorine. Apatite crystals are chlorine-bearing, up to 2.8 wt % where measured (Table S3, in the Supplementary Material available online at https://doi.org/10.1017/S0016756822000474). The marble is not spatially related to any magmatic deposits. Tourmaline is not recorded in Tiree marble, but it does occur in marbles in several other Lewisian supracrustal outcrops in Harris (Coward et al. Reference Coward, Francis, Graham, Myers and Watson1969), Iona (Rock, Reference Rock, Park and Tarney1987) and Gairloch – Loch Maree (Robertson et al. Reference Robertson, Simpson and Anderson1949).
Analyses for dating of the titanite in marble from Gott, Tiree, were made in situ on 13 titanite grains in polished rock slabs (Fig. 4), ranging in size from c. 100 × 50 μm to 700 × 200 μm. A total of 30 spots were analysed; the largest grain was targeted with seven analyses. No age differences are observed between the different grains. All analysed domains are common-Pb (PbC)-rich, with 206Pb/204Pb ratios ranging from 500 to 18 000. On a Wetherill Concordia diagram, uncorrected data fan out from being 90 % to 102 % concordant, with a larger spread along the Concordia. Pb corrected data are from 95 % to 103 % concordant, and plot on a relatively well-defined Discordia line (Fig. 6) with a lower intercept at 330 ± 157 Ma and an upper intercept at 1593 ± 7 Ma (n = 30; MSWD = 5.4). This indicates that the applied PbC works well, and age data are therefore obtained from the PbC data. Deselecting all reverse discordant analyses yields an upper intercept age of 1593 ± 11 Ma (n = 17; MSWD = 3.0) (Fig. 6), and selecting only 100 % concordant analyses yields a Concordia age of 1586 ± 7 (n = 4; MSWD = 2.2). The best age estimate for the titanite crystallization is 1593 ± 11 Ma (MSWD = 3.0).
5. Discussion
5.1 Sulphates and sulphides
The anhydrite in Tiree is the first recorded in the Lewisian Complex. In Palaeoproterozoic supracrustal successions in the North Atlantic region, anhydrite is preserved in Greenland (Horn et al. Reference Horn, Dziggel, Kolb and Sindern2019), Sweden (Martinsson et al. Reference Martinsson, Billström, Broman, Weihed and Wanhainen2016) and Russia (Serdyuchenko, Reference Serdyuchenko1975). Together with well-preserved pseudomorphs after gypsum in many regions including Sweden (Lager, Reference Lager2001) and Canada (Bell & Jackson, Reference Bell and Jackson1974; Zentmyer et al. Reference Zentmyer, Pufahl, James and Hiatt2011; Hodgskiss et al. Reference Hodgskiss, Crockford, Peng, Wing and Horner2019), there is extensive evidence for sulphate-bearing seawater, at ∼2.0–1.9 Ga. In several cases, sulphur isotope data are available and are strongly positive, in accord with an evaporative origin for the anhydrite. The occurrence of anhydrite in the Lewisian Complex is therefore not anomalous, and rather is consistent with the global picture of widespread evaporites in the Palaeoproterozoic. Evidence for pseudomorphs in Tiree may be obscured by shearing focused on the metasediments.
The temporal relationship between anhydrite and barite cannot be proven, but the overprinting pattern of the barite suggests that it is most likely to be later, in which case the barite sulphur could have been remobilized from the anhydrite and precursor gypsum. In the Hudson Bay region, pseudomorphs of ∼2.0 Ga gypsum are similarly overprinted by barite (Hodgskiss et al. Reference Hodgskiss, Crockford, Peng, Wing and Horner2019).
The sulphur-bearing scapolite from Tiree suggests derivation from a sulphate-rich sedimentary environment (Morrissey & Tomkins Reference Morrissey and Tomkins2020; Zeng et al. Reference Zeng, Zhao, Hammerli, Fan and Spandler2020). Several other Palaeoproterozoic successions contain scapolite attributed to metamorphism of evaporites, but only one of five data sets has a sulphur content as high as the range for the Tiree scapolite (Table S2, in the Supplementary Material available online at https://doi.org/10.1017/S0016756822000474).
Globally, sulphide deposits of the mid-Palaeoproterozoic that have been characterized by sulphur isotopic composition fall into two main groups. Volcanic massive sulphides derived from magmatic-hydrothermal fluids have a composition of ∼0 ‰, while sulphides attributed to derivation from seawater sulphate have a heavy (positive δ34S) composition. This could include volcanic massive sulphides in which the hydrothermal fluids were recycled from seawater rather than purely magmatic. Mid-Palaeoproterozoic (1.9–1.8 Ga) diagenetic pyrite is characterized by positive δ34S values, reflecting derivation from seawater with a relatively limited sulphate content, notwithstanding the occurrence of sulphate evaporites (Scott et al. Reference Scott, Wing, Bekker, Planavsky, Medvedev, Bates, Yun and Lyons2014). In the North Atlantic region, Palaeoproterozoic volcanic massive sulphides and diagenetic sulphidic shales have distinct compositions (Fig. 5). Sulphur isotope data from sulphides in the Lewisian supracrustal inliers have hitherto been limited to the volcanic massive sulphide deposit at Kerry Road, Gairloch, are tightly grouped around 0 ‰ and probably represent fluids of magmatic-hydrothermal origin (Drummond et al. Reference Drummond, Cloutier, Boyce and Prave2020). In contrast, the pyrite measured here from the Tiree marble is markedly positive (mean 11.9 ‰), comparable to those of mid-Palaeoproterozoic sulphides attributed to an origin in seawater (Fig. 5).
5.2 Phenocryst assemblage
The phenocryst assemblage in the Tiree marble is found in many other Lewisian Complex marbles of northern Scotland, from South Harris to Scardroy (Fig. 2). A core assemblage of pyroxene, amphibole, olivine, mica, titanite, epidote and quartz phenocrysts is consistent across the region (Rock, Reference Rock, Park and Tarney1987). The uniformity implies that the marble-hosted phenocrysts represent the metamorphic overprint on the original mineralogy, rather than local effects.
The ∼1600 Ma dates for the Tiree titanite record mineral growth in the Palaeoproterozoic–Mesoproterozoic, and show no contribution from the much younger episodes of sodium metasomatism found elsewhere in northern Scotland. The sodium, chlorine and sulphur recorded in the mineral assemblage can therefore be confidently attributed to the chemistry of the depositional environment of the marble in the Palaeoproterozoic.
The titanite dates are comparable with the younger ages determined for reworking of the Lewisian Complex. They show no evidence of the 2500–2000 Ma reworking ages (Crowley et al. Reference Crowley, Key and Noble2015) determined for Archaean gneisses in the bulk of the Lewisian Complex. The age is also younger than the ∼1.8–1.7 Ga date ascribed to the main phase of Laxfordian deformation and metamorphism that widely affects the Lewisian Complex in NW Scotland (Goodenough et al. Reference Goodenough, Crowley, Krabbendam and Parry2013). However, there is increasing evidence for an event in Scotland in the range 1.6–1.55 Ga, described by some workers as ‘Late Laxfordian’. This includes a ∼1.6 Ga ‘cooling’ date for hornblende in a shear zone (Sherlock et al. Reference Sherlock, Jones and Park2008), a 1.55 Ga Re–Os date for copper mineralization (Holdsworth et al. Reference Holdsworth, Selby, Dempsey, Scott, Hardman, Fallick and Bullock2020), a major magnetizing event 1.7 to 1.5 Ga (Piper, Reference Piper1992) and a 1.6–1.4 Ga age for cooling of granite on the Stanton Banks west of Tiree (Scanlon & Daly, Reference Scanlon and Daly2001). Holdsworth et al. (Reference Holdsworth, Selby, Dempsey, Scott, Hardman, Fallick and Bullock2020) point out that an event of this age in Scotland links activity in Canada and Scandinavia at the time, which saw the later stages of the Labradorian and Gothian orogenies respectively.
5.3 Chlorine-bearing phases
Apatite in the Tiree marble commonly contains above 1 wt % chlorine, and up to 2.8 wt % (Table S3, in the Supplementary Material available online at https://doi.org/10.1017/S0016756822000474). These contents are higher than those in many other marbles (Table S3), or apatite in granites and iron deposits which mostly contain <0.5 % (Ishihara & Moriyama, Reference Ishihara and Moriyama2015). The Tiree apatite is thus considered to be chlorine-rich. This would be consistent with a seawater origin for the apatite, and there are no associated magmatic rocks which might indicate an alternative origin. However, phosphatic rocks were widespread globally at c. 1.9 Ga (Papineau, Reference Papineau2010), and we regard the chlorine-rich apatite as supporting rather than critical evidence.
The chlorine content of the phlogopite in the marble is not exceptional, but is comparable with the content in micas from other suspected metamorphosed evaporites (Moine et al. Reference Moine, Sauvan and Jarousse1981; Mora & Valley Reference Mora and Valley1989; Opletal et al. Reference Opletal, Houzar and Leichmann2007). The role of micas as a residence for seawater-derived chlorine is in their relative abundance rather than the content per mineral.
The assemblage of chlorine-bearing mineral phases, in particular scapolite and phlogopite, is typical of metamorphosed evaporites (Moine et al. Reference Moine, Sauvan and Jarousse1981; Mora & Valley, Reference Mora and Valley1989; Warren, Reference Warren2016). The combination of scapolite and phlogopite is encountered in several Palaeoproterozoic supracrustal successions in the North Atlantic region, including in Bergslagen, Sweden (Oen & Lustenhouwer, Reference Oen and Lustenhouwer1992), Finland (Reinikainen, Reference Reinikainen2001) and Baffin Island, Canada (Belley et al. Reference Belley, Dzikowski, Fagan, Cempírek, Groat, Mortensen, Fayek, Giuliani, Fallick and Gertzbein2017), and in each case has been interpreted as evidence of former evaporites. It would therefore be consistent to interpret the Lewisian Complex geochemistry as further evidence of Palaeoproterozoic evaporite deposition.
Saline fluids from chlorine-bearing mineral phases are important for transport of metals and creation of ore deposits (Yardley & Graham, Reference Yardley and Graham2002; Riehl & Cabral, Reference Riehl and Cabral2018; Morrissey & Tomkins, Reference Morrissey and Tomkins2020). Coeval supracrustal rocks in adjacent Greenland and Scandinavia contain ore deposits in which evaporites are implicated as a source of mineralizing fluids (Frietsch et al. Reference Frietsch, Tuisku, Martinsson and Perdahl1997; Horn et al. Reference Horn, Dziggel, Kolb and Sindern2019). Metalliferous ores and mineral showings in the Lewisian supracrustals (Coats et al. Reference Coats, Shaw, Gunn, Rollin and Fortey1997; Drummond et al. Reference Drummond, Cloutier, Boyce and Prave2020; Parnell et al. Reference Parnell, Michie, Heptinstall and Still2021) suggest that there may also be potential deposits in Scotland, to which evaporite-derived fluids could have contributed.
5.4 Sodium metasomatism and tourmaline
The albite in Tiree marbles is further evidence of sodium metasomatism. The albite is part of the mineral assemblage which dates to the latest Palaeoproterozoic – early Mesoproterozoic. This distinguishes it from episodes of sodium metasomatism in other parts of northern Scotland, which affect Mesoproterozoic and Neoproterozoic metasediments (Sutton & Watson Reference Sutton and Watson1951; May et al. Reference May, Peacock, Smith and Barber1993; Van de Kamp & Leake, Reference Van de Kamp and Leake1997), and which must be of younger age. Albitization occurs in Palaeoproterozoic rocks contiguous to Scotland, in Greenland and Scandinavia. This albitite in the North Atlantic region is argued to be derived from evaporites or seawater (Kalsbeek, Reference Kalsbeek1992; Frietsch et al. Reference Frietsch, Tuisku, Martinsson and Perdahl1997; Gleeson & Smith, Reference Gleeson and Smith2009), and there is not a clear alternative in Scotland that would preclude a similar origin.
The lack of recorded tourmaline in the Tiree marble may reflect the relatively quiescent nature of Palaeoproterozoic sedimentation there. While tourmaline is increasingly recognized as evidence for metamorphosed evaporite sequences, in many cases the environments included exhalative brines on the sea floor (e.g. Oen & Lustenhouwer, Reference Oen and Lustenhouwer1992; Jiang et al. Reference Jiang, Palmer, Peng and Yang1997). Where tourmaline is reported most abundantly in the Scottish marbles, at Loch Maree (Robertson et al. Reference Robertson, Simpson and Anderson1949), the marbles occur in a section that also hosts exhalative sulphide mineralization and iron formation (Drummond et al. Reference Drummond, Cloutier, Boyce and Prave2020). The tourmaline in Lewisian marble thus conforms to evidence in other successions for an association with meta-exhalites. Tourmaline-bearing metasediments across the North Atlantic region of c. 1.9 Ga age, from Quebec (Chown, Reference Chown1987) to Hudson Bay (Ricketts, Reference Ricketts1978), Greenland (Thomassen, Reference Thomassen1992) and Sweden (Hellingwerf et al. Reference Hellingwerf, Gatedal, Gallagher and Baker1994), are all attributed to former evaporites.
6. Conclusions
Combinations of criteria have contributed to a picture of evaporites in Palaeoproterozoic supracrustal successions across the North Atlantic region (Fig. 1; Table S1 in the Supplementary Material available online at https://doi.org/10.1017/S0016756822000474). The aspects that are consistent with an evaporitic facies in the Lewisian Complex of Tiree include anhydrite, scapolite, phlogopite, chlorine-rich apatite, and pyrite with a positive sulphur isotope signature. The attribution is not definitive, but it represents multifaceted evidence in support of a role for fluids derived from seawater. The evidence of pseudomorphs is currently lacking. Nonetheless, it seems likely that the supracrustal rocks of the Lewisian Complex included evaporitic facies, like their counterparts in many other parts of the North Atlantic.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/S0016756822000474
Acknowledgements
J. Bowie, C. Brolly, J. Armstrong and J. Johnston provided skilled technical support. Electron microscopy was performed with the help of J. Still in the ACEMAC Facility at the University of Aberdeen. Scapolite was analysed on a sample from the Hunterian Museum, Glasgow (no. 134720), loaned courtesy of J. Faithfull. The work was supported in part by UK Natural Environment Research Council grant NE/M010953/1. Careful review by A.A. Cabral helped to improve the manuscript.
Conflict of interest
The authors declare none.