Glacial dispersal of indicator minerals from the Brazil Lake lithium–caesium–tantalum pegmatites, SW Nova Scotia, Canada Open Access

The three BLP are ∼25 km NE of the town of Yarmouth (Fig. 1) and 200 km SW of the city of Halifax in the province of Nova Scotia, in eastern Canada. The glaciated landscape consists of gently rolling hills consisting of drumlin fields with a near-continuous cover of glacial sediments of varying thickness. Bedrock was only observed in a few places where it is exposed along stream beds and around the edges of bedrock quarries. Sample sites were accessed by truck along numerous local and resource-access roads, and recreational trails.

The study area is underlain by rocks of the Meguma terrane, the most easterly component of the northern Appalachian orogen (Hibbard et al. 2006; White and Barr 2012; Fig. 2a). The region is underlain by Lower Paleozoic metasedimentary and metavolcanic rocks of the Meguma Terrane and White Rock formation, which are intruded by Silurian–Carboniferous granites (Fig. 1; White 2010; White et al. 2018). The Meguma Terrane is characterized by a thick sequence of Cambro-Ordovician metasedimentary rocks, comprising the metasandstone-dominated Goldenville Group and the overlying siltstone- and slate-dominated Halifax Group (Fig. 2a; White 2010). The White Rock formation consists of mixed metasedimentary rocks and mafic–felsic volcanic rocks (White 2010). These Lower Paleozoic rocks were deformed as part of the regional Acadian deformation at 410–400 Ma (Keppie and Dallmeyer 1995) into upright NE- to NNE-trending regional-scale folds, with the Meguma Terrane rocks defining paired anticlines and synclines and the White Rock formation occurring in a broad synclinal structure (Figs 1, 2a; van Staal 2007; White 2010). The metamorphic grade in the area ranges from greenschist to upper amphibolite facies (White 2010).

Numerous Sn, base and precious metal occurrences are found throughout SW Nova Scotia. The most significant is the East Kemptville Sn deposit along the western edge of the South Mountain Batholith (e.g. Kontak and Dostal 1992; Fig. 1). There are numerous smaller granite-hosted greisen deposits and metasediment-hosted shear and replacement style Sn–Zn–Cu–Pb–In deposits.

The BLP are hosted by the Silurian White Rock Formation (Rockville Notch Group), comprising shallow marine metasedimentary rocks interbedded with minor mafic metavolcanic units, which locally include quartzite, amphibolite and pelitic schist (Figs 1 and 2a, b; White 2010; White and Barr 2017). The Brenton Pluton, a syenogranite to monzogranite intrusion related to the White Rock Formation, is in fault contact with both the White Rock Formation and the Halifax Group and occurs ∼3 km SW of the BLP (Kontak 2006). It is inferred to be Silurian age based on a U–Pb zircon age of 439 ± 4 Ma (Keppie and Krogh 2000). At the SE margin of the White Rock Formation, staurolite-grade, typically schistose rocks, are in faulted contact with slate of the Halifax Group that has been metamorphosed to amphibolite facies. The fault is inferred to be a brittle structure within the broader Chebogue Point shear zone (previously termed the Deerfield shear zone). At its NW margin, the Halifax Group and White Rock Formation are deformed along the Cranberry Point shear zone; cooling ages of muscovite indicate the deformation to be middle Carboniferous (Alleghanian) (Culshaw and Liesa 1997; Culshaw and Reynolds 1997; White and Barr 2017). Field observations by Kontak (2006) suggest the BLP were emplaced in an active shear zone where high-temperature ductile deformation occurred during consolidation of the pegmatite. Age dates of tantalite (U–Pb) from the South dyke indicate that pegmatite crystallization occurred at ∼395 Ma (Kontak et al. 2005; Kontak and Kyser 2009).

The bedrock geology of the NE-trending North and South pegmatite dykes is summarized below from Kontak (2004, 2006), Kontak et al. (2005) and Cullen et al. (2022). The North and South pegmatites were named with respect to their location north or south of Holly Road (Fig. 2b) and these dykes are the most well-known and described of the three. Drilling and surface trenching indicate that the North and South dykes are separated by ∼300 m and occur as lenticular forms with wider cores transitioning to thinly tapered ends. The North dyke is at least 700 m in length and reaches a maximum thickness of 21 m at its centre. The South dyke has a defined strike length of ∼300 m and a thickness of ∼8–12 m. Both dykes dip steeply (60–85°) to the SE. The pegmatites are of the albite–spodumene type and are characterized by coarse crystals of spodumene (maximum size 1.5–2.0 m; Fig. 3a, b) and K-feldspar, with intergranular spodumene, muscovite, albite and quartz (Fig. 3b). Accessory phases include columbite–tantalite (Fig. 3c), black tourmaline (Fig. 3d), apatite, beryl, sphalerite and cassiterite. Spodumene, the most abundant Li-bearing mineral in the pegmatites, is a light-coloured silicate (hardness 6.5–7) with two perfect cleavages, a vitreous lustre that is pearly on its cleavage surface and a moderate density of 3.18 g cm⁻³. Little is published about the third spodumene-bearing pegmatite, Army Road, 350 m to the east of the South pegmatite, which was discovered through drilling sometime between 2002 and 2020 (Innovation News Network 2024; Lithium Springs Limited 2024). Because of the presence of three closely spaced LCT pegmatite dykes, the local area is referred to as the BLP setting. Minerals reported to be present in the North and South pegmatites which could be useful indicators for drift prospecting include spodumene, black tourmaline, black columbite–tantalite, red garnet, blue apatite and other phosphate minerals, green beryl, cassiterite, wolframite, sphalerite, zircon, epidote, topaz, and titanite (Table 1).

The North and South dykes have been the focus of exploration and petrological studies since they were discovered in 1960 by tracing the source of pegmatite boulders on-surface back to what was later discovered to be an outcrop of the South dyke (Taylor 1967; Kontak 2004; Barr and Cullen 2010). Numerous spodumene-bearing boulders are still evident on-surface in the area immediately down-ice of the North dyke, and in the area in general (e.g. Yammine 2024). The most recent resource estimate for the two pegmatites was reported as measured and indicated resources of 555 300 tonnes grading 1.30% Li₂O and an inferred mineral resource of 381 000 tonnes grading 1.48% Li₂O (Cullen et al. 2022; Canadian Mining Journal 2024). Diamond drilling was initiated in October of 2022 to further define the Li resources at the North and South dykes, and to assess the Li potential of the Army Road pegmatite (Fig. 2b). Drilling in the past few years has expanded the South dyke down-plunge to the SW with thicknesses up to 20 m and identified spodumene in the Army Road pegmatite with some drill-core assay values reported up to 1.9% Li₂O (Innovation News Network 2024; Lithium Springs Limited 2024). The most recent resource estimate (MRE) for the combined three pegmatites is 10.01 million tonnes at 1.20% Li₂O at a cutoff grade of 0.33% Li₂O (Lithium Springs Limited 2024).

The glacial history of SW Nova Scotia is summarized from previous regional-scale surficial mapping and till sampling conducted by Stea and Grant (1982) and Finck and Stea (1995) in the 1970s and 1980s. This mapping and sampling, together with the stratigraphic studies by Grant (1980), Grant and King (1984) and Stea et al. (1992), have resulted in a broad framework of the regional glacial history. This early framework was based upon relative stacking and correlation of till units of varying properties exposed along the Bay of Fundy and Gulf of Maine, and should be re-examined in the context of the time-transgressive nature and evolution of the former ice sheet during its early growth, full extension and decay. Drift-thickness modelling (Brushett et al. 2023b, c) and new regional surficial mapping, aided by light detection and ranging (LiDAR) data (Fig. 4), are ongoing by NSDNR to update the stratigraphic and depositional history for the region.

The surficial geology of southwestern Nova Scotia is described as the product of multiple glacial advances and retreats throughout several glacial events (Stea 2004; Stea et al. 2011) during the Quaternary. Multiple till units, identified in coastal sections in SW Nova Scotia, and mapped striations are interpreted as a product of several phases of glacial advances and shifting ice-flow directions (Grant 1980; Stea and Grant 1982). These ice-flow phases were largely controlled by local topography during ice advances and retreats, and likely influenced by thinner marine ice margins at lower sea-levels and ice streams (cf. Shaw et al. 2006). Three main phases are described below from oldest to youngest:

Regional eastward to southeastward ice flow emanating from New Brunswick during the Caledonia phase (marine isotope stage (MIS) 4; ∼75–50? ka) was responsible for the early formation of drumlins in SW Nova Scotia. The tills that comprise drumlins are silt-rich (as well as shell-rich in coastal sections) and contain a high percentage of exotic bedrock components with transport distances of >80 km from New Brunswick (Stea and Finck 2001; ice-flow phase 1 red arrows in Fig. 5).

Subsequent southward regional ice flow during the Escuminac phase (MIS 2; 22–18 ka; Stea et al. 2011) modified older drumlins and formed new ones that reflect a more southward flow (ice-flow phase 2 blue arrows in Fig. 5). Ice streaming in the Gulf of Maine towards the marine glacial margins also strongly influenced the strong southward ice flow in western Nova Scotia (Shaw et al. 2006).

During the subsequent Scotian phase (∼20–17 ka), regional ice centres shifted and Nova Scotia was cut off from the larger continental ice centres (i.e. the Appalachian Glacier Complex) mainly due to drawdown of the ice profile from a major ice stream in the Bay of Fundy (Shaw et al. 2006; Stea et al. 2011). The resultant Scotian Ice Divide formed lengthwise down the centre of the province such that ice-flow direction varied from west-NW to SW to SE in southwestern Nova Scotia (Grant 1980; Stea and Grant 1982; ice-flow phase 3 purple arrows in Fig. 5).

As a result of this complex glacial evolution, there is extensive till cover over much of southwestern Nova Scotia that has made bedrock mapping and mineral exploration challenging. The study area is no exception, where extensive glacial modification through the last glacial cycle resulted in highly variable glacial sediment thicknesses and complex landforms. SE- to south-trending drumlins are the predominant glacial landform over much of the region, with the orientation transitioning to south-trending drumlins close to the coastline (Fig. 5).

Till thickness in the area is variable, ranging from thin veneers (<2 m) to thick drumlin ridges composed of over 40 m of till (Brushett et al. 2023b, c). The South dyke naturally outcrops (∼5 m² area) at its north end, but otherwise all three dykes are covered by till. As a result of the glacial sediment cover, the areal extent of the BLP is not well known.

Swathes of mega-scale glacial lineations (MSGL) are observed in LiDAR imagery of this drumlinized terrain, and have also been recognized offshore in the Bay of Fundy (Shaw et al. 2014). MSGLs are indicative of fast-flowing ice (ice streaming) during the latter phases of glaciation (Stokes and Clark 2002; Sookhan et al. 2021), prior to the formation of the Scotian Ice Divide (ice-flow phase 2 on Fig. 5). The area between the lineations (till ridges) is characterized by thin till overlying bedrock with sediment cover of <5 m. Glaciofluvial deposits commonly occur in topographic lows, which are now occupied by modern rivers and wetlands. Coastal regions do show previously unmapped minor geomorphic alteration of the landscape, discernable in LiDAR imagery, from marine inundation up to 5 m or more above present-day sea-levels immediately following deglaciation along the coastline.

Four genetic properties (1) matrix texture, (2) fissility and compaction, (3) clast lithologies and (4) the proximity to geomorphic forms that could be identified using LiDAR imagery were used to identify and differentiate between two till facies that occur at surface in the study area, using the modern nomenclature of subglacial sediments (Evans 2017; McClenaghan et al. 2023c):

1. Subglacial traction till: a silty sand till that is over consolidated and sometimes displays subhorizontal fissility with visible jointing. A typical till exposure (∼1–2 m depth) of this material, shown in Figure 6, displays a compact brownish grey silty-sandy till with angular to subrounded clasts, the majority of which reflect local bedrock lithologies, and many of which are striated and faceted. Additional site photos of subglacial traction till are included in Supplementary material, Appendix A. This till nomenclature supersedes the older terms ‘subglacial lodgement till’ and ‘basal till’ which may be familiar to readers (e.g. Evans 2017; McClenaghan et al. 2020; 2023c). Pebbles and cobbles of spodumene (Fig. 7) were observed in this till facies (Supplementary material, Appendix A, photo b, yellow arrow) in trenches that were dug 50–100 m down-ice of the North and South dykes (Supplementary material, Appendix B, photos b, c). A subcropping bedrock surface on the South dyke, revealed by previous stripping, is glacially polished and striated (179°) and demonstrates that southward-flowing ice had actively eroded the bedrock surface (Fig. 5, ice-flow phase 2). This till facies is the ideal target sample medium for indicator minerals in the region due to its locally derived nature and the fact that it was deposited under actively flowing ice.

2. Subglacial melt-out till: a second till facies was identified, primarily in bedrock quarries and drumlin exposures where it was seen as a thin (<2 m) horizon on the drumlin flanks (areas within the yellow dashed lines in Supplementary material, Appendix C). This till is sandier, less indurated and less compact, contains well-sorted lenses and layers of sand, contains a higher proportion of clasts and a larger range of clast sizes than the subglacial traction till. Clasts are generally more monolithic than those present in the subglacial traction facies. In some quarries or borrow pits, a boulder horizon was identified within the till, generally composed of angular greywacke boulders. Originally, this till facies was interpreted as widespread ice-flow phase 3 sediments from the Scotian Ice Divide, but in drumlinized and streamlined terrain this unit could also represent the ‘erodent layer’ (Eyles et al. 2016; Sookhan et al. 2022) from late-stage terrestrial phase 2 ice streams. This till facies is the secondary target for sampling if subglacial traction till is not present at the sample site.

The BLP area has been the subject of several till geochemical investigations since the 1980s, including those by Shell Canada Resources Limited (Palma et al. 1982), the Nova Scotia Department of Natural Resources (MacDonald et al. 1992), a university MSc thesis by Lundrigan (2008) and, more recently, B-horizon till sampling by Champlain Mineral Ventures (Black 2011, 2012; Wightman 2018, 2020). These studies have demonstrated that the pegmatites have a variable geochemical signature (Li, Sn and Ta) in the local till down-ice. Black (2012) also reported on the abundance of spodumene fragments (>2 mm) in till that forms a glacial dispersal train up to 1.5 km down-ice of the BLP to the SSE. No previous systematic indicator mineral studies, such as the one described here, have ever been conducted around the BLP.

Samples of exposed bedrock were collected to provide insights into the types of LCT pegmatite indicator minerals present in the North and South pegmatites. Samples of the Army Road pegmatite were not available for this study. Two bedrock samples of the South dyke, two bedrock samples from the North dyke and one pegmatite boulder sample on-surface ∼50 m down-ice of the North dyke were collected for heavy mineral analysis. Small fist-sized hand samples were selected and combined such that all phases of the pegmatite were sampled for heavy mineral separation and comparison with till samples. Indicator minerals recovered from the bedrock samples were then looked for in the till samples. General information about bedrock and boulder samples, colour photographs and indicator mineral content are reported in McClenaghan et al. (2024).

A total of 84 bulk heavy mineral till samples were collected from 77 sites, including field duplicate samples at 4 different sites (Fig. 2a, b). Sample sites consisted of hand-dug pits, till exposures in borrow pits or along the sides of local roads, and new backhoe trenches dug on the down-ice (SSE) side of the North and South dykes, where C-horizon (unoxidized to moderately oxidized) till was targeted. Till samples were collected by NSDNRR in 2020, 2021 and 2022 (DB series sample numbers, e.g. 21DB001-2) and the GSC in 2022 (MPB series sample numbers, e.g. 22MPB003) following GSC till sampling protocols described in McClenaghan et al. (2013, 2020, 2023c) and Plouffe et al. (2013).

At each sample site, a 7–20 kg till sample was collected for (1) processing and recovery of mid- and high-density mineral concentrates from which indicator minerals for LCT pegmatites and other ore systems (e.g. Au grains) could be picked and (2) for the recovery of the pebble-sized fraction for lithological analysis (>2.0 mm fraction). An additional 3 kg sample was collected for geochemical analysis of the matrix and for archiving. Field data collected at each site, including site photographs, are reported in Brushett et al. (2024).

At three sample sites, two till samples were collected for heavy mineral separation to document vertical compositional variability within a till:

1. In a borrow pit on Raynardton Road, 12 km SE of the Brazil Lake pegmatites and on the north side of Lake Vaughan (Fig. 2a), two till samples were collected from a 10 m high till section exposing the core of a south-trending drumlin. The east–west oriented vertical face in the borrow pit exposes two sandy subglacial tills. The uppermost till in the section is moderately dense, has 21% clast content and a sandy-silt matrix (56% sand, 42% silt), and contains small discontinuous lenses and layers of sand (Brushett et al. 2024; McClenaghan et al. 2024). It mantles the drumlin feature, is ∼3.5 m thick and was deposited by southward ice streaming that actively eroded and streamlined the landscape (Fig. 5, ice-flow phase 2, blue arrows). This uppermost till contains abundant large bedrock boulders on or at its natural land surface. It was interpreted as a subglacial traction till after examination of the LiDAR imagery for the drumlin surface. It was deposited as the erodent layer from southward ice streaming. Sample 22MPB034 was collected at 1.4 m depth from the upper till unit (Supplementary material, Appendix B, photo a). The lower till is dense and fissile, has 15% clast content, has a silty-sand matrix (54% sand, 43% silt), is estimated to be >6 m thick, and forms the core of the drumlin (Brushett et al. 2024; McClenaghan et al. 2024). This lower till was deposited as subglacial traction till by southward flowing ice, as indicated by till fabric data collected from the section in this study and striations (185°) on top of a poorly developed boulder and cobble horizon at ∼3 m depth. This southward trend corresponds to the general orientation of the drumlins in the region. Bedrock was not exposed at the bottom of the section or on the pit floor. Sample 22MPB033 was collected from 4.3 m depth in the lower till (Supplementary material, Appendix B, photo a).

2. At a backhoe trench 50 m south of the NE-trending North dyke (Fig. 2), sample 22MPB043 was collected at 1.1 m depth and sample 22MPB044 was collected at 2.1 m depth (Supplementary material, Appendix B, photo b). Bedrock was not reached in this trench.

3. At a second backhoe trench overlying the NE-trending North dyke (Fig. 2), till sample 22MPB047 was collected at 1.9 m depth and sample 22MPB048 was collected at 1.0 m depth (Supplementary material, Appendix B, photo c). Bedrock was not reached in this trench.

The large bulk (∼7–20 kg) till samples and small fist-sized (0.5–1.3 kg) rock samples were shipped to Overburden Drilling Management Limited (ODM) in Ottawa, Canada. Bedrock samples were processed using electric pulse disaggregation (EPD). Bedrock and till samples were further processed using tabling and heavy liquid separation to produce 0.25–2.0 mm mid-density (2.8–3.2 SG) and high-density (>3.2 SG) mineral concentrates from which indicator minerals were counted. Selected grains were removed for further study. Flow sheets outlining sample processing procedures for bedrock and till samples are reported in McClenaghan et al. (2024). Bedrock samples were examined to establish the suite of indicator minerals that best reflect the LCT pegmatites.

Till samples were processed and examined for the Li indicator minerals that had been identified in the bedrock samples, for Au grains (because of the proximity to the nearby Kemptville Gold District) and for indicator minerals of other types of mineralized bedrock in the region. GSC in-house blanks (Plouffe et al. 2013; McClenaghan et al. 2020, 2023c) were inserted into sample batches by GSC personnel prior to shipping to the mineral processing laboratory, in order to monitor the lab's processing and mineral picking procedures and the potential for sample contamination. The Bathurst in-house blank is a weathered granite (grus) that contains no indicator minerals except goethite and the Almonte in-house blank is a till that contains a known range of abundance of a few indicator minerals including pyrite, apatite, goethite and tourmaline (e.g. Plouffe et al. 2013; Appendix E in McClenaghan et al. 2024)

Till samples were screened at 2.0 mm, with the <2.0 mm material processed first using a shaking table to prepare a <2.0 mm preconcentrate. The table preconcentrate was micro-panned to recover fine-grained (<0.25 mm) Au, sulfides or other heavy indicator minerals which might be present. The <2.0 mm preconcentrate was subsequently subjected to two heavy liquid separations and ferromagnetic separations to produce 2.8–3.2 SG and >3.2 SG non-ferromagnetic heavy mineral concentrates for visual identification and counting of indicator minerals. The 0.25–0.5, 0.5–1.0 and 1.0–2.0 mm non-ferromagnetic >3.2 SG fraction, and the 0.25–0.5 mm non-ferromagnetic 2.8–3.2 SG fraction of bedrock and till samples were examined as loose grains in petri dishes by ODM, and potential indicator minerals were visually identified (in some cases with the aid of energy dispersive spectroscopy (EDS) coupled with a scanning electron microscope (SEM)) and counted. The 2–4 mm ‘granule’ fraction of 30 of the 77 till samples was visually examined to determine the abundance of coarse spodumene fragments in till.

Proportional dot maps showing the abundance of selected indicator minerals, normalized to 10 kg sample weight (weight of <2 mm table feed fraction), were plotted using the ArcGIS Pro v. 3.2.1 desktop application. Data classes for each dot size were determined using percentiles calculated in ArcPro.

For selected minerals, the mineral abundance data from the till samples, normalized to a 10 kg weight of the <2 mm (table feed) fraction, are reported in McClenaghan et al. (2024) and summarized in Supplementary material, Appendix D. Mineral distributions in till samples are described and discussed below, and plotted on maps as the normalized 0.25–0.5 mm fraction data. Because no regional indicator mineral survey data have been previously published for the Brazil Lake area, background values were established using till samples farthest up ice (north) from the pegmatites that contained the fewest indicator minerals.

Raw, unedited indicator mineral data for the pan concentrate and 0.25–0.5 mm heavy mineral fraction of the Bathurst blank, Almonte till blank and field duplicates are reported, along with the routine till sample data, in McClenaghan et al. (2024). The results for both types of blanks indicate that no external or carry-over contamination between samples was detected. Two GSC Almonte till in-house blanks (Plouffe et al. 2013) were spiked with known numbers of spodumene gains prior to shipping of the samples to ODM and all the spiking grains were recovered. Spodumene counts are similar between duplicate pairs of field samples, most notably for the 2.8–3.2 SG fraction, where spodumene is most abundant.

Potential LCT pegmatite indicator minerals are listed in Table 1 along with their density, hardness, who first reported their presence in the North and South pegmatites, and an indication of whether they were recovered in bedrock and till samples collected in this study. Selected mineral count data normalized to a 10 kg sample weight are reported in Supplementary material, Appendix D. Selected mineral distribution maps, plotted using the normalized data in the form of proportional dot maps, are shown below.

Spodumene is the main Li-bearing mineral of LCT pegmatites (Černý 1991; Černý and Ercit 2005; Bradley and McCauley 2013; Groves et al. 2022) and its presence in the North and South pegmatites is well documented (Corey 1995; Kontak 2004, 2006). Spodumene has an SG range of 3.1–3.2 and a hardness of 6.5–7. It was identified in bedrock and till mid-density and heavy mineral fractions in this study by its white colour, prismatic, generally flattened and elongated, striated, commonly massive crystal habit, and brittle fracture (Fig. 8a, b), and, for some grains, its fluorescence under short-wave UV light. All three size fractions of the mid-density and heavy mineral fractions were examined for spodumene, and background abundance in both density fractions and both size fractions is zero grains.

Spodumene abundance is summarized as follows:

1. Mid-density (2.8–3.2 SG) fraction. Spodumene is most abundant in the mid-density fraction as compared to the heavy mineral fraction, and in the smallest size fraction (0.25–0.5 mm) with the most anomalous samples containing hundreds to thousands of grains/10 kg sample (Fig. 9a). It is present, but less abundant, in the 0.5–1.0 mm fraction, with the highest abundances between tens and hundreds of grains/10 kg. Spodumene is absent to scarce in the 1.0–2.0 mm fraction, containing a maximum of single grains to tens of grains up to 3 km to the south of the BLP.

2. High-density (>3.2 SG) fraction. Similar to the mid-density fraction, spodumene is most abundant in the smallest size fraction (0.25–0.5 mm) of the heavy mineral fraction, with the most anomalous samples containing single grains to tens of grains/10 kg (Fig. 9b). It is much less abundant in the 0.5–1.0 mm fraction, containing up to a maximum of 12 grains/10 kg, and absent in the 1.0–2.0 mm fraction, except for a single sample that contains 3 grains/10 kg.

In this study, spodumene in the 2.8–3.2 SG fraction is most abundant in till immediately south of the North and South dykes (maximum 2212 grains) and in sample 21DB027 that is 3 km south of the South pegmatite (1705 grains) (Fig. 9a). Abundances decrease southward but are still detectable (two grains/10 kg) in till in sample 22MPB027, 12 km south of the pegmatites and just south of Lake Vaughan (Fig. 9a).

At the three previously mentioned sites, where two till samples were collected from thicker till deposits, spodumene content in till varies with depth at two of the sites:

1. Samples 22MPB033 (4.3 m depth) and 22MPB034 (1.4 m depth), collected at the Raynardton Road borrow pit, show that the lower till sample does not contain spodumene grains (Supplementary material, Appendix B, photo a). The upper till sample, which is equivalent to a regional surface sample in this study, contains two grains of spodumene.

2. Samples 22MPB043 (1.1 m depth) and 22MPB044 (2.1 m depth), collected from a backhoe pit 25 m south of the south end of the North dyke (Fig. 2), show that the upper and lower samples contain similar abundances of spodumene (>2000 grains/10 kg) (Supplementary material, Appendix B, photo b). The lower till sample contained 2128 grains/10 kg and the upper till sample, equivalent to a regional surface sample, contained 2098 grains. Collecting a routine ‘surface’ till sample at this site would have detected glacial dispersal from the North dyke.

3. Samples 22MPB047 (1.9 m depth) and 22MPB048 (1.0 m depth), collected from a backhoe pit 25 m south of the middle section of the North dyke (Fig. 2), show that these two till samples contain very different amounts of spodumene (Supplementary material, Appendix B, photo c). The lower till sample contains 2212 grains/10 kg and the upper sample contains zero grains. Collecting a routine surface till sample at this site would not have detected glacial dispersal from the North dyke. In this case, the upper part of the till sequence is likely derived from bedrock up-ice of the pegmatite, and did not re-entrain the underlying till that contains spodumene.

Kontak (2004, 2006) described the occurrence of columbite–tantalite in the North and South pegmatites, whereas Kontak et al. (2005) specifically refer to the mineral tantalite in their study of the geochronology of the pegmatites and described it as dark brown tabular crystals that were 1–2 cm in length (Fig. 3c). Columbite and tantalite have a combined SG range of 5.3–8.2 and a hardness range of 6–6.5 (Table 1). It was identified in bedrock heavy mineral fractions by its black colour and blocky crystal habit (Fig. 8c), and selected grains were confirmed by ODM using EDS coupled with an SEM. Visually identifying columbite–tantalite in heavy mineral concentrates can be challenging because visually it is similar to hornblende and ilmenite. Only till sample 22MPB047, collected 50 m south of the North pegmatite, was found to contain columbite–tantalite (two grains).

Tourmaline in, and on the margins of, the North and South pegmatites occurs as brown-black, green-brown and dark blue crystals in the margins of, and within, the pegmatite (Corey 1995; Hughes 1995; Kontak 2004). The tourmaline in the pegmatites is dravite–schorl (Kontak 2004) and has an SG of 3.18–3.22 and a hardness of 7. It was identified in this study in bedrock and in till mid- and high-density fractions by its dark colours, prismatic crystal habit and parallel striations on crystal surfaces. Most tourmaline grains in the bedrock samples in this study are black to dark brown, although some blue tourmaline grains were observed in bedrock sample 22MPB513 and till samples 22MPB009 and 22MPB047.

Tourmaline is most abundant in the mid-density (2.8–3.2 SG) fraction of the till samples (hundreds to thousands of grains/10 kg) and the highest values in till are: (1) 75 m south of the South dyke (2232 grains/10 kg in sample 22MPB037) and the Army Road pegmatite (11 539 grains/10 kg in 22MPB009); (2) between the South dyke and the Army Road pegmatite (1290 grains/10 kg in 21DB032); and (3) 2 km north of the North dyke (1200 grains/10 kg 22MPB023) (Fig. 10a). Elevated (>90th percentile) values also occur in till 5 km east and 5 km south of the BLP. Abundances are much less (hundreds of grains) in the >3.2 SG fraction, and values in this density fraction are highest in till around the North and South dykes. Tourmaline content in sample 22MPB009 is five to nine times higher than the other anomalous samples listed above, indicating that tourmaline is likely more abundant within and/or along the margins of the Army Road pegmatite as compared to the North and South pegmatites.

The North and South pegmatites contain abundant apatite (Hughes 1995; Kontak 2004, 2006). Figure 10b shows the abundance of apatite (all colours) in the mid-density (2.8–3.2 SG) 0.25–0.5 mm fraction of till. It is most abundant in till samples that are ∼2 km south to SW of the South dyke. A visually distinct light blue variety of apatite is readily visible in bedrock hand samples (Fig. 8d), however blue apatite was not observed in the till samples.

Zircon was identified in bedrock heavy mineral concentrates by its brownish red colour and short-wave UV light colour (yellow, green, orange colour). It was observed in bedrock sample 20DB-031 BDK. Grains of uraninite intergrown with zircon were recovered from bedrock samples 20DB-031 BDK and 22MPB513. Uraninite was identified by its black submetallic lustre and its intergrowth with zircon (Fig. 8e, f). Neither mineral was observed in the till samples.

The presence of cassiterite in the North and South pegmatites is known from previous studies by Corey (1995) and Kontak (2004). Cassiterite has a SG of 6.8–7 and a hardness of 6–7, and was identified in bedrock and till heavy mineral concentrates in this study by its bright lustre, brown colour and prismatic crystal habit. Cassiterite grains varying in size from 0.25 to 2.0 mm were recovered from the >3.2 SG fraction of the bedrock samples. Fine-grained cassiterite (25–250 $μ$m) was recovered from the pan concentrate of three bedrock samples (22MPB511, 22MPB 512 and 22MPB 514). In till samples, cassiterite was recovered from the 0.25–0.5 mm size of the >3.2 SG fraction. Abundance varies from a background of zero grains in most till samples to a maximum of six grains in sample 22MPB046, 100 m east of the South dyke (Fig. 10c). The lack of cassiterite in most till samples indicates that little to no metal-rich debris from the East Kemptville Sn deposits (Kontak and Dostal 1992), 20 km to the NE (Fig. 1), has been transported to the BLP area.

The presence of scheelite has not been reported in the North and South pegmatites, however, it can be an accessory mineral in pegmatites (Poulin et al. 2018). It has a SG of 5.9–6.12 and a hardness of 4–5. In the current study, it was identified in till heavy mineral fractions by its pale yellow to white colour under normal light, by its bright whitish-blue to yellow fluorescence under short-wave UV light and by its cleavage. Most till samples contain zero grains (Fig. 10d), with a few samples containing between one and three grains. The highest values are in till samples collected up to 2 km south of the South pegmatite. This close proximity of the scheelite-bearing till samples to the pegmatite combined with presence of spodumene in these same samples is a strong indication that the scheelite is likely from the BLP.

Kontak (2006) reported the presence of Fe-poor sphalerite in the North and South pegmatites. Orange (mid- to low Fe content) sphalerite grains were identified in the >3.2 SG fraction of bedrock samples 20DB-031 BDK and 22MPB513 (Fig. 8g). Sphalerite is present in both the 0.5–1.0 mm and the 0.25–0.5 mm heavy mineral faction of bedrock samples. No sphalerite was recovered from till samples.

The presence of pyrolusite has not previously been reported in the BLP. Pyrolusite is a secondary mineral formed from the oxidation of Mn-bearing minerals, and potentially tantalite. It has a SG of 4.4–5.1 and a hardness of 6–6.5. It was identified in the >3.2 SG 0.25–0.5 mm fraction of bedrock sample 22MPB513 from the North dyke by its dull black amorphous appearance. It was not recovered from any till samples.

Indicator mineral grains were examined in three size fractions of the <2.0 mm fraction in this study: (1) 0.25–0.5 mm, the most common size fraction examined in mineral exploration surveys (e.g. McClenaghan and Paulen 2018; McClenaghan et al. 2020, 2023c); (2) 0.5–1.0 mm; and (3) 1.0–2.0 mm. In our study, spodumene abundance is greatest in the smallest size fraction, the 0.25–0.5 mm fraction.

The abundance of spodumene fragments in the 2–4 mm (granule) fraction of selected till samples was also examined in our study (Table 2). The impetus for examining this size fraction was Black's (2012) report of a 1.5+-km dispersal train of >2 mm spodumene fragments they had documented in till trending south of the BLP. In this study, granules of spodumene can be detected in till up to 3 km to the south of the known pegmatites (Fig. 9c), far less down-ice than that for the 0.25–0.5 mm mid-density spodumene grains. The advantage of the 2–4 mm granule fraction is that it can be sieved from till samples and visually examined while still in the field and will be a strong indicator of the presence of an LCT pegmatite when proximal (<3 km) to it.

Indicator minerals were examined in two mineral density fractions in this study: (1) 2.8–3.2 SG; and (2) >3.2 SG. These two density fractions were investigated because spodumene has an SG range of 3.1–3.2, spanning the threshold between these two density fractions. Thus, spodumene should be most abundant in the 2.8–3.2 SG fraction but could also be present in the >3.2 SG fraction. In addition to spodumene, apatite (3.16–3.22 SG) and tourmaline (dravite–schorl; 3.18–3.2 SG) have similar SG ranges to spodumene and thus could be recovered in both density fractions. As expected, spodumene, apatite and tourmaline are most abundant in the 2.8–3.2 SG fraction. If both density fractions of the 0.25–0.5 mm size fraction are compared side by side (Fig. 9a, b), it is readily apparent that the spodumene dispersal patterns are similar but most distinct and well-developed for the 2.8–3.2 SG fraction. In future till sampling programmes for Li exploration, the heavy liquid separation of till samples can be carried out using the narrower density range of 3.0–3.2 SG, instead of 2.8–3.2 SG, in order to optimize spodumene recovery and eliminate some of the lighter minerals which visually resemble spodumene or apatite.

The threshold between background and anomalous concentrations of spodumene in till is zero grains. Any spodumene grains that are present in a till sample are significant. Spodumene abundance is highest in a central dispersal corridor trending southward from the BLP for 12 km (Fig. 9a). However, the overall pattern of spodumene dispersal is fan-shaped (see pink polygon in Fig. 9a). The fan geometry is interpreted to be the net result of the three main directions of ice flow that occurred over the three phases of glaciation (Fig. 5; SE, south and SW) that eroded, transported, re-entrained and deposited spodumene-bearing till from the North, South and Army Road pegmatites, and possibly other unknown LCT pegmatites in the local area. Note that sample spacing along the west and east margins of the fan has gaps; additional till sampling in these areas could refine the dispersal pattern.

Over the past 60 years, exploration for LCT pegmatites in glaciated terrain globally has been initiated by the discovery of pegmatite boulders (e.g. Schultz 1971; Nikkarinen and Björklund 1976 ; Steiger 1977; Sarapää et al. 2015 ; Barros et al. 2022), including those found down-ice of the North and South pegmatites at Brazil Lake (Taylor 1967; Barr and Cullen 2010). In contrast to commodities such as Cu, Pb, Zn and Au (e.g. McClenaghan and Paulen 2018 and references therein), only a few studies have been conducted to evaluate the effectiveness of indicator minerals for Li exploration in glaciated terrain (Table 3). These case studies include those reported by Nikkarinen and Björklund (1976) and, more recently, by Hodder and Martins (2023) and Hodder (2024). Until now, no indicator mineral study has been conducted around an LCT pegmatite using modern, consistent indicator mineral methods such as those reported here. For all these studies, the highest spodumene abundances that were reported in till are listed in Table 3 to allow for comparison with results of this Brazil Lake study. In this study, the highest spodumene count in a single till sample is 2212 grains/10 kg, much higher than the other reported abundances. The large number of spodumene grains compared to other reported studies could in part be related to the very coarse nature of the spodumene (megacrysts up to 2 m) in the North and South pegmatites. The overall suite of potential LCT indicator minerals is larger than those recovered in till in this study. Indicator minerals present in the North and South pegmatites, but not recovered in our till samples include montebrasite, holmquistite, beryl, sphalerite, zircon, uraninite and pyrolusite.

Figure 9a and d compare the dispersal patterns for spodumene in the 0.25–0.5 mm 2.8–3.2 SG fraction to the dispersal defined by till geochemistry (Li ppm by Na-peroxide fusion, 1–2 mm fraction). Spodumene grains from the BLP are detectable in a fan-shaped pattern at least 12 km down-ice (south of Lake Vaughan; pink polygon in Fig. 9a). In comparison, glacial dispersal defined by the Li content in the till matrix is a much smaller and narrow train, trending south for 2.5 km (pink polygon in Fig. 9d), and thus a much smaller exploration target.

Although drift prospecting for Li pegmatites is not new, interest in exploration for these rocks is surging due to the current demand for battery metals. This study is the first published detailed investigation of the indicator mineral signature of an LCT pegmatite in till in glaciated terrain. The results are highly relevant to Li exploration in SW Nova Scotia, as well as the glaciated terrain of Canada (e.g. NW Territories, western Ontario and James Bay; Mining.Com 2024; Natural Resources Canada 2024) and other glaciated terrains globally (e.g. Ireland and Fennoscandia) where Li exploration is currently ongoing. This study demonstrates the types of systematic sampling and heavy mineral lab methods to be used, types of indicator minerals which can be expected proximal to an LCT pegmatite, and the importance of understanding glacial flow paths and glacial transport for interpreting glacial dispersal patterns. The striated upper surface of the South pegmatite, orientations of the streamlined glacial landforms in the area and till fabric data measured at selected sites are the main indicators of ice-flow directions. Spodumene is the key indicator of glacial transport direction and distance. Key takeaways from this study include the following:

• Indicator minerals. The glacial dispersal from the BLP is best defined by the primary Li ore mineral, spodumene. The recovery of spodumene grains in till is visual confirmation of the presence of an LCT pegmatite. The grains may be examined in detail, photographed and chemically analysed to provide additional insights into the source material. In contrast, till geochemistry alone cannot be relied on to indicate the presence of a Li-bearing pegmatite because the Li geochemical signal could be derived from Li-rich mica or clays (Brushett et al. 2024) and not related to a LCT pegmatite. At Brazil Lake, the background abundance of spodumene in till is zero grains and thus the presence of just one spodumene grain in a 10 kg till sample is significant; this is the equivalent to ppb-level geochemical analyses. Glacial dispersal of spodumene from the BLP forms a broad fan-shaped exploration target (12 km down-ice) that is longer and wider than that defined by till matrix geochemistry. Additional indicator minerals of the BLP in till samples include columbite–tantalite, apatite, tourmaline, cassiterite and scheelite. Table 4 summarizes the indicator minerals and pathfinder elements for the North, South and Army Road pegmatites. Indicator mineral and till geochemistry methods, used together, are powerful tools to explore for LCT pegmatites in glaciated terrain.

• Mineral density. LCT pegmatite indicator minerals occur in two easily recoverable density fractions of till: (1) 2.8–3.2 SG fraction, that hosts most of the spodumene, apatite and tourmaline that occurs in the till samples; and (2) >3.2 SG fraction, host to columbite–tantalite, cassiterite and scheelite. In future surveys, heavy liquid separation could be carried out to recover a narrower density range of SG 3.0–3.2, in order to optimize spodumene recovery, reduce the volume of material to be examined and eliminate some of the lighter minerals that visually resemble spodumene or apatite.

• Size fraction. Spodumene is most abundant in the smallest (0.25–0.5 mm) size fraction that is visually examined. Till samples immediately down-ice of the pegmatites contain hundreds to thousands of spodumene grains per 10 kg sample in this size fraction. Spodumene is only present in the 0.5–1.0 mm fraction up to 5 km down-ice and in the 1–2 mm fraction up to 2.7 km down-ice. These shorter dispersal distances down-ice, as compared to the 0.25–0.5 mm fraction, suggest that coarse spodumene (0.5–2.0 mm) will be a useful indicator of close proximity to an LCT pegmatite.

• Proximity indicators. Table 2 compares the down-ice dispersal distances of spodumene-rich debris in three different fractions of till – 2–4 mm granules (spodumene fragments), 0.25–2.0 mm mid-density (spodumene grains) and the till matrix (Li geochemistry) – for selected samples across the BLP area. The horizontal double lines in each data column indicate the farthest distance down-ice that the LCT pegmatite signature is detectable (i.e. anomalous values). Parameters for defining the thresholds between the background and anomalous values are listed at the bottom of the table. Indicator minerals are detectable the farthest, at least 12 km, down-ice. Granule-sized spodumene fragments are only detectable in the first 3 km down-ice, beyond which they have been glacially crushed (comminuted) into smaller (<2 mm) fragments. Similar to the granule fraction, till geochemical signatures are only detectable in the first 2.5 km down-ice, beyond which the geochemical signal is diluted by the surrounding till to background values. For areas <3 km down-ice of the BLP pegmatites, spodumene granules and till geochemistry can be effective exploration tools, in addition to indicator minerals. For areas farther down-ice (>3 km), spodumene-rich debris has been crushed to <2 mm and diluted by surrounding till, and indicator mineral methods are the only effective tool to detect the presence of an LCT pegmatite.

• Sample spacing and orientation. Given the small size of the BLP pegmatites (up to 21 × 700 m for the larger North dyke) and the three phases of ice flow, indicator mineral sampling could be conducted along parallel lines spaced 500 m or greater apart and oriented perpendicular to the dominant ice-flow direction. Till sample spacing along lines will depend on the size of the property to be covered, orientation of local and regional bedrock structure(s) hosting anticipated pegmatite dykes relative to regional ice-flow directions (e.g. perpendicular or parallel), and budget. Ideally, heavy mineral samples should be spaced 250 m apart initially, and closer spaced for subsequent follow-up sampling around anomalies. These guidelines will likely be suitable for other areas as well.

• Till facies. Two till facies are present in the Brazil Lake region and both are suitable for till sampling to recover indicator minerals: (1) Subglacial traction till in the Brazil Lake area is a silty-sand till with moderate to high compaction, subhorizontal fissility with visible jointing, angular to subrounded local bedrock clasts, and numerous striated and faceted clasts; (2) Subglacial melt-out till is sandier, stonier, less indurated, less compact, contains well-sorted lenses and layers of sand, and monolithic clast lithologies, and occurs as a thin (<2 m) unit throughout the region, particularly on drumlin flanks. If both till types are encountered at a sample site, subglacial traction till should be targeted for sampling.

Future research will build on the initial bedrock and till indicator mineral abundance study (McClenaghan et al. 2024) and geochemical analysis of till samples (Brushett et al. 2024). Mineral chemistry of apatite and tourmaline in bedrock and till samples will be investigated for samples from and around the North and South pegmatites. Till stratigraphic studies are in progress to provide a better understanding of the glacial history and stratigraphic context of the erosional and depositional records of SW Nova Scotia. A better understanding of the till units at surface, and their relationship to the latest deglaciation phase as well as the effects of being a product of ice streams on the landscape, is also in progress.

The authors thank John Wightman, Greg Morris and Cliff Stanley for access to the Brazil Lake property, field visits and advice that guided the field work. Overburden Drilling Management Limited is thanked for their professional heavy mineral processing services, adaptation of the laboratory work flows and patience with our numerous questions. Staff at the Carleton Country Outfitters store is thanked for excellent service and support during the field work. Mitch Maracle (NSDNRR summer student) is thanked for his assistance with field work from 2020 to 2022. Abeer Haji Egeh (GSC) and Adam Nissen (Saint Mary's University) are thanked for their assistance in the field in 2022. Lyndsay Moore (GSC), Dan Kontak (Laurentian University) and Steve Amor (retired, Geological Survey of Newfoundland and Labrador) are thanked for their scientific reviews of this paper.

In 2022, two site visits were made to the North pegmatite with Acadia First Nation representatives Jeff Purdy (Councillor, Acadia First Nation), Greg Hart (NSPI Early Engagement Coordinator, Kwilmu'kw Maw-klusuaqn Negotiation Office) and Patrick Butler (Mi'kmaq Energy & Mines Advisor, Kwilmu'kw Maw-klusuaqn Negotiation Office) to explain the research, the geology of the lithium deposit, answer questions and address potential concerns about the impact of the field work. We thank the Acadia First Nation for their interest in the research.

MBM: conceptualization (lead), data curation (equal), formal analysis (equal), funding acquisition (lead), investigation (lead), methodology (lead), project administration (lead), supervision (lead), validation (lead), writing – original draft (lead), writing – review & editing (lead); DMB: conceptualization (equal), formal analysis (equal), funding acquisition (supporting), investigation (equal), methodology (supporting), visualization (lead), writing – original draft (supporting), writing – review & editing (supporting); RCP: conceptualization (supporting), data curation (equal), formal analysis (supporting), investigation (equal), validation (supporting), writing – original draft (supporting), writing – review & editing (supporting); CEB: conceptualization (supporting), data curation (supporting), formal analysis (supporting), methodology (supporting), writing – original draft (supporting), writing – review & editing (supporting); JMR: data curation (supporting), investigation (supporting), writing – original draft (supporting), writing – review & editing (supporting).

The research reported here was funded by the Geological Survey of Canada through its Targeted Geoscience Initiative and by the Nova Scotia Department of Natural Resources – Geoscience and Mines Branch.

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.

The datasets generated during and/or analysed during the current study are available from the Natural Resources Canada, Geological Survey of Canada.