Abstract
Iron oxides are minerals resistant to chemical alteration and mechanical abrasion, and which have ferromagnetic properties and a range of chemical compositions. These characteristics are useful as indicator minerals in exploration, for example using till in glaciated terrains. Iron oxide proportions, grain size, and chemical composition of till samples collected near the Sue-Dianne Cu-Au-Ag IOCG deposit in the Great Bear magmatic zone (Northwest Territories, Canada) and magmatic Ni-Cu deposits in the Thompson Nickel Belt (Manitoba, Canada) show that subsamples containing c. 100 grains from the 0.25–1.0 mm grain size ferromagnetic fraction yield a representative mineralogical and compositional range of oxide grains from a till sample. Subsamples with less than 100 grains yield statistically less representative data. The 1–2 mm grain size fraction typically contains too few iron oxide grains and thus using this fraction is not statistically representative.
The composition of iron oxides from eight till and five bedrock samples was determined along transects up- and down-ice of the Cu-Au-Ag Sue-Dianne IOCG deposit. At, and immediately down-ice of, the deposit, hematite is the principal oxide and shows dominant BIF and IOCG chemical signatures in the Ca+Al+Mn v. Ti+V discriminant diagram. Up-ice and farther down-ice of the deposit, magnetite and titanomagnetite are the dominant oxides and magnetite shows dominant Kiruna and IOCG signatures. The composition of iron oxides from six till samples along a north–south transect and 11 till samples from a 180 km-long east–west transect, along the older and younger directions of ice-flow, respectively, was determined in the Thompson Nickel Belt (Manitoba, Canada). The proportion of magnetite in till with the signature of Ni-Cu deposits increases for at least 1 km south of the Pipe Ni-Cu deposit along the direction of the older southward ice flow, whereas the glacial dispersal of magnetite with a chemical signature typical of Ni-Cu deposits was limited during the younger westerly ice flow.
- Indicator minerals
- magnetite
- hematite
- ferromagnetic fraction
- till
- composition
- microprobe chemistry
- IOCG
- Sue-Dianne deposit
- Ni-Cu deposits
- Thompson Nickel Belt
The use of indicator minerals recovered from glacial sediment has been shown to be a successful method for ore deposits exploration in glaciated terrain, especially for diamonds and gold, and more recently for magmatic Ni–Cu, platinum-group element (PGE), porphyry Cu, and volcanogenic massive sulphides (Averill 2001, 2011; McClenaghan 2005; Kelley et al. 2011). Iron oxides, until recently, have not been recovered from surficial sediments and used as indicator minerals for exploration, but magnetite has been used in sediment provenance studies (e.g. Grigsby 1990; Razjigaeva & Naumova 1992). Magnetite and hematite have several physical and chemical characteristics that make them potentially useful indicator minerals: (1) they occur in many ore deposits as primary or accessory minerals (Dupuis & Beaudoin 2011); (2) they are physically robust, and resistant to chemical and mechanical weathering; (3) they are easily separated from heavy mineral concentrate fractions using their magnetic properties; and (4) their chemical composition has been shown to be useful to distinguish between different types of mineral deposits (Carew 2004; Gosselin et al. 2006; Singoyi et al. 2006; Beaudoin et al. 2007; Nadoll et al. 2009, 2012; Rusk et al. 2009, 2010; Beaudoin & Dupuis 2010; Dupuis & Beaudoin 2011).
Because the ferromagnetic fraction of a 10-kg till sample typically weighs c. 1–100 g and contains hundreds to thousands of grains, not all of the detrital ferromagnetic grains can be analysed on a routine basis. Therefore, it is essential to develop a protocol for preparing a ferromagnetic subsample with an adequate number of grains of suitable grain size practical for routine analysis in order to obtain a representative chemical compositional variation of the whole till sample. Beaudoin et al. (2009) showed that the composition of magnetite from a small aliquot of the ferromagnetic fraction of a till sample could be statistically representative of the magnetite compositional range of the sample. The conclusions of Beaudoin et al. (2009) were based on the composition of magnetite grains from five subsamples, each having 33–82 grains, from the 0.5–1 mm grain size ferromagnetic fraction of two till samples collected near the Pipe mine open pit in the Thompson Nickel Belt (TNB), Manitoba, Canada.
In this paper, we investigate in detail subsample preparation methods, and present geochemical and statistical data for till samples collected around the Sue-Dianne deposit (Great Bear magmatic zone, Northwest Territories, Canada) and Thompson Nickel Belt Ni-Cu massive sulphide deposits (Manitoba, Canada). The data are used to define the optimal grain-size and sub-sample number of grains that is required to capture the geochemical signature from the ferromagnetic fraction of a till sample.
The magnetite to hematite-group Cu-Au-Ag Sue-Dianne iron-oxide-copper-gold (IOCG) deposit, located in the southern part of the Great Bear magmatic zone (Fig. 1), is one of two case studies chosen to test the optimal preparation method of a subsample of grains from the ferromagnetic fraction of till for IOCG exploration in glaciated terrain. Petrographic and LA-ICP-MS (laser ablation ICP mass spectrometry) studies of magnetite associated with ore minerals in IOCG deposits and prospects of the Great Bear magmatic zone identify important geochemical differences in V, Ni, Cr and Co contents in magnetite, and show that barren host rock magnetite has a different composition to that of the Sue-Dianne deposit (Acosta-Góngora et al. 2013). At Sue-Dianne, till samples were collected along a NE-SW transect across the deposit, parallel to the predominant ice-flow direction, and bedrock samples were collected from the deposit, its host volcanic rocks, and background intrusive rocks (Fig. 1). The second case study was conducted along north–south and east– west transects of the Thompson Nickel Belt (Fig. 2) to test the application of the method for exploration for magmatic Ni-Cu massive sulphide deposits.
(a) Geology of the Great Bear magmatic zone and location of the main IOCG deposits and mineral occurrences (after Hoffman & Hall 1993; NORMIN database www.nwtgeoscience.ca/normin). (b) Regional geology, mineral occurrences, and sample locations in the Sue Dianne deposit area (Northwest Territories, Canada) (from Camier 2002; NORMIN database: www.nwtgeoscience.ca/normin). Trend and relative age of glacial striations from Normandeau & McMartin (2013).
Regional geology of the Thompson Nickel Belt (Manitoba, Canada), showing the mineral occurrences and the location of till samples. Red box outlines the area shown in Figure 3. Modified from McClenaghan et al. (2011).
Regional Geological Setting of Case Studies
Sue-Dianne IOCG deposit
The Sue-Dianne deposit is located in the southern part of the Great Bear magmatic zone (GBmz), within the Bear Structural Province of the Canadian Shield (Fig. 1a) in northern Canada. The GBmz is a Palaeoproterozoic 1.87–1.85 Ga calcalkaline volcano-plutonic arc accreted to the western margin of the Archean Slave craton during the short-lived Calderian Orogeny (Bowring & Grotzinger 1992; Hildebrand et al. 2010). Formerly known for its vein-type uranium and silver deposits, the Great Bear magmatic zone is currently the most prospective IOCG mineral belt in Canada (Corriveau et al. 2010).
The Cu-Au-Ag Sue-Dianne deposit is a polymetallic deposit hosted in well-preserved rhyodacite ignimbrite sheets of the Faber Group volcanic sequence (Fig. 1b) and was formed during the development of a structural-hydrothermal diatreme breccia complex (Gandhi 1989). Alteration at Sue-Dianne ranges from sodic alteration (albite) in the deepest parts of the breccia complex to potassic-iron alteration (hematite-K-feldspar, magnetite-K-feldspar, K-feldspar) in the shallowest part of the diatreme (Goad et al. 2000; Camier 2002; Mumin et al. 2010).
The Sue-Dianne breccia complex forms an elliptical-shaped topographic bedrock high c. 600 m long, 500 m wide, and 350 m deep (Goad et al. 2000), with relief exceeding 70 m above the surrounding rhyodacite ignimbritic volcanic rocks of the Faber Group. The glacial sediment cover around the deposit consists of a discontinuous, thin (<2 m) till with large areas of exposed bedrock. Ice-flow indicators show the area was affected primarily by the Laurentide Ice Sheet flowing to the southwest from a Keewatin centre during the last glaciation (Normandeau & McMartin 2013). Till has a local provenance and a short distance of glacial transport, as indicated by the glacial dispersal of heavily metasomatized clasts in till less than 800 m down-ice of the deposit (Normandeau et al. 2011).
Thompson Nickel Belt Ni-Cu deposits
The Thompson Nickel Belt (TNB) is part of the Circum-Superior Boundary Zone. It is limited to the NW by the Trans-Hudsonian Orogeny and to the SE by the Superior Province craton (Fig. 2). The TNB has a NE general trend, ranging in width from 10–35 km, and is c. 400 km long (Couëslan et al. 2006). It results from a complex geodynamic history implying a succession of subduction, magmatism and high-grade metamorphism during the Trans-Hudsonian Orogeny, from c. 2.1–1.80 Ga (Corrigan et al. 2007).
The TNB contains world-class magmatic Ni-Cu-PGE massive sulphide deposits that are hosted in sulphide facies iron formations, in sedimentary rocks and ultramafic intrusions at two main stratigraphic levels within the Pipe Formation of the Ospwagan Group (Fig. 2; Peredery et al. 1982; Layton-Matthews et al. 2007). Till samples used in this study are from the central part of the TNB, near the Pipe open pit mine (Fig. 2). The Pipe deposit is a massive sulphide body hosted in deformed dunite intruded into the P1 Member of the Pipe Formation (Fig. 2; Layton-Matthews et al. 2007).
The TNB is covered by a thin layer of till between <0.5 and 3 m thick (McClenaghan et al. 2009). Ice-flow indicators show the area was glaciated by the Laurentide Ice Sheet flowing to the south from a Keewatin centre and subsequently to the west from a Labradorean centre during the last Wisconsinan glaciation (Dredge et al. 1986; Klassen 1986; Dredge & Nixon 1992). Striation directions on outcrops in the northern TNB indicate two ice flow directions, c. 210o and c. 270o, the first direction being crosscut by the second, indicating that it is the oldest (Fig. 2; McClenaghan et al. 2009).
Methodology
Sample collection
A total of 31 large till samples and 26 representative surface bedrock samples were collected from the Sue-Dianne deposit area in 2009 by the Geological Survey of Canada (Normandeau & McMartin 2013). Till and bedrock samples were collected up-ice, proximal to, and down-ice from the deposit. Bedrock samples include ore, hydrothermally altered host rocks and least altered regional bedrock. Till samples were collected at an average depth of 50 cm from upper C horizon soils developed on till, to obtain relatively unweathered material. For this study, the ferromagnetic fraction from the heavy mineral concentrate of eight till samples (7–19 kg of <2 mm material) was selected from the Sue-Dianne deposit till survey along a 10-km long transect parallel to the predominant ice-flow direction (Fig. 1). Five bedrock samples (1–3 kg) were selected to document the geochemical signature of magnetite in mineralized and barren host rocks and in samples of the regional plutonic rocks (Fig. 1).
A total of 144 large till samples were collected from the TNB area in 2005, 2006 and 2007 by the Geological Survey of Canada (McClenaghan et al. 2009; McMartin et al. 2012). For this study, 14 till samples (15–20 kg) were selected east and west of the TNB and proximal to the Pipe Ni-Cu deposit (Figs 2, 3). Of these 14 till samples, 6 were collected along a 2 km-long, NNW–SSW transect, parallel to the early southern direction of ice-flow, whereas 11 are from an east–west, 180 km-long transect from the Pikwitonei Domain (East; Fig. 2), across the TNB, near the Pipe open pit where three samples were used for both transects, and into the Kisseynew Domain (West; Fig. 2) along the general westward direction of the last major ice-flow.
Geology of the Pipe deposit area, showing the mineral occurrences, the location of till samples, and the two ice flow directions based on glacial striations (modified from McClenaghan et al. 2011).
Sample preparation
For both the Sue-Dianne and TNB deposit surveys, heavy mineral concentrates were recovered from till samples by Overburden Drilling Management Ltd. (ODM) using a combination of tabling, heavy liquids and ferromagnetic separations as described in McClenaghan et al. (2009) and Normandeau & McMartin (2013). Bedrock samples were disaggregated using electric pulse disaggregation methods (McClenaghan et al. 2011), and heavy mineral concentrates were prepared by ODM using the same method as for till samples. A hand magnet was used to extract the ferromagnetic fraction from each heavy mineral concentrate. For each till and bedrock sample, the ferromagnetic fraction was further sieved at ODM into a <0.25 mm fraction and a 0.25–2.0 mm fraction. The 0.25–2.0 mm ferromagnetic fraction was subdivided at Université Laval (Canada) into two sub-fractions, 0.25–1.0 mm and >1.0–2.0 mm. The <0.25 mm grain size fraction was not used in this study because the small grains are difficult to manipulate and a large number of grains consist of mineral intergrowths, such that it is common that the polished area of the oxide is too small for chemical analysis.
Subsample selection at the Sue-Dianne deposit
Three till samples were selected based on the size of their ferromagnetic fraction and their location with respect to the deposit to test sample heterogeneity related to grain size fractions and number of grains (Fig. 1b): 1) sample 09-MOB001 is located 100 m up-ice of Sue-Dianne iron oxide mineralized breccias; 2) sample 09-MOB003 is located c. 7 km up-ice of the deposit over rocks of the Marion River Batholith; and 3) sample 09-MOB035 is located 2 km down-ice of the deposit, over barren Faber Group volcanic rocks. The 0.25–1.0 mm fraction of the three samples was split, using a riffle, into four subsamples containing c. 200, 100, 50, and 10 grains each. Duplicates of the nine subsamples of the 0.25–1.0 mm fraction containing c. 100, 50, and 10 grains were prepared to test reproducibility. Grains from each subsample were mounted in 25-mm epoxy mounts and the surface was polished for analysis. The 1–2 mm size fraction from the three samples contained fewer than 110 ferromagnetic grains, and therefore could not be split. Consequently, all of the grains were mounted in 25-mm epoxy mounts. In total, 24 epoxy mounts of the samples were prepared (15 subsamples and 9 duplicates; Table 1). Five additional till samples (09-MOB002, 09-MOB004, 09-MOB015, 09-MOB022, 09-MOB029; Fig. 1b) located up-ice, over the Sue-Dianne deposit, and down-ice, were selected to complete a NE-SW transect to test for mineralogical and compositional variations along the ice-flow path. For each of these five till samples, a subsample of the 0.25–1.0 mm ferromagnetic fraction, containing c. 100 grains was mounted in an epoxy mount (Table 2).
Total number of magnetite and hematite grains analysed in each subsample (defined by the size fraction and the number of grains) prepared from selected till samples
Total number of grains and number of magnetite and hematite grains analysed in each till and bedrock sample of the IOCG Sue-Dianne deposit (fraction of c. 100 grains 0.25–1.0 mm)
Five bedrock samples were selected (Fig. 1b) to determine the iron oxide compositional range, and included two samples collected within the ore zone (09-CQA1026A3, 09-CQA1009B5), one sample from altered host rocks (09-CQA1006D4), and two samples representing least-altered regional bedrock (09-CQA0005A4,09-CQA1010A4). For each bedrock sample, a subsample was prepared from the 0.25–1 mm grain size ferromagnetic fraction with c. 100 grains (Table 2).
Subsamples selection at the TNB deposits
For each of the 14 till samples, a subsample of the 0.25–1.0 mm ferromagnetic fraction containing between 100 and 200 grains was split, mounted in epoxy and polished (Table 3).
Total number of grains and number of magnetite and hematite grains analysed in each till sample of the Thompson Nickel Belt (fraction of c. 100 grains 0.25–1.0 mm)
Analytical method
The chemical composition of magnetite and hematite from ferromagnetic grains that consist of a single mineral or complex mineral intergrowths was measured by electron probe microanalysis (EPMA). Minor and trace element contents were determined using a method modified from Dupuis & Beaudoin (2011) employing the CAMECA SX-100 five wavelength dispersive spectrometers EPMA at Université Laval, Canada. Analytical conditions included an accelerating voltage of 15 kV, a 100-nA beam current, forming a 10-μm beam diameter, and counting times of 40 to 80 s on peak. Background was measured on both sides for 15–20 s at positions free of interferences. Simple oxides (GEO Standard Block, from P&H Developments) and minerals (Mineral Standard Mount MINM25-53, from Astimex Scientific Limited) were used as calibration standards. The typical values computed for the detection limit (DL) are: Al (18 ppm), Ca (11 ppm), Cr (52 ppm), Cu (20 ppm), K (8 ppm), Mg (22 ppm), Mn (40 ppm), Ni (61 ppm), Si (15 ppm), Sn (6 ppm), Ti (19 ppm), V (49 ppm), and Zn (77 ppm). The complete analytical results are reported in Dupuis et al. 2012.
Results
Sue-Dianne IOCG deposit
Mineralogy of the ferromagnetic fraction
The ferromagnetic fractions of till and bedrock samples from the Sue-Dianne deposit area are composed of grains containing magnetite, titanomagnetite, ilmenite, hematite, as well as other non-ferromagnetic oxides (e.g. rutile, apatite), silicates, and rarely sulphides. SEM examination of the polished surface of ferromagnetic grains indicates that magnetite or hematite form single mineral grains or composite particles that contain ferromagnetic minerals. Magnetite is commonly subhedral in mineral aggregates and shows replacement by hematite in larger, single grains. In mineralized rocks and till samples above the Sue-Dianne deposit, hematite with residual magnetite cores is most common. Till and bedrock samples collected both up- and down-ice of the deposit contain magnetite and titanomagnetite as the dominant oxides. Titanomagnetite commonly contains ilmenite exsolution lamellae. Locally, magnetite and titanomagnetite occur in fine-grained intergrowths with silicates.
Comparison of subsamples from the three selected till samples
Mineral proportions in subsamples. In each grain of the fifteen subsamples and nine duplicates from the ferromagnetic fraction of the three selected till samples (09-MOB001, 09-MOB003, and 09-MOB035), the mineral with the larger area on the polished surface was recorded. The proportion of minerals in subsamples of till sample 09-MOB-001 is shown as a representative example in Figure 4. Results show that the mineral proportions in 09-MOB-001 are similar and less variable in the 0.25–1.0 mm grain size fractions with 100 and 200 grains, with magnetite accounting for 34.0–36.5% of each subsample and duplicate (Fig. 4). By contrast, the coarser >1–2 mm fraction, and the 0.25–1.0 mm fractions with 50 grains or less, show more variable mineral proportions (Fig. 4).
Proportions of minerals in each subsample for different grain size fractions and approximate number of grains (n) prepared from till sample 09-MOB001 near the Sue Dianne deposit (Fig. 2). An asterisk indicates a duplicate subsample.
Iron oxide composition of subsamples. Some magnetite and hematite grains were not analysed by EPMA because the polished surface was too small, or because the iron oxide was not exposed on the polished surface of the mount. In addition, some magnetite grains contain hematite inclusions (and inversely) in which both phases were analysed. The number of grains of magnetite and hematite analysed by EPMA is summarized in Table 1.
It is common in trace element analysis by EPMA that the content of some elements in a mineral is below detection limit (DL), such that summary statistics cannot be computed using common methods. In order to statistically compare the composition of magnetite and hematite from the various subsamples and duplicates, we have calculated the mean, median, and the standard deviation for all fractions from each of the three till samples using a parametric survival regression model (survreg, www.r-project.org). In Figure 5a, the range, standard deviation and mean of Al contents is shown for all subsamples with different numbers of grains, grain size fractions, and duplicates, for sample 09MOB-001. We tested, using survreg, the hypothesis that the mean content of an element for each size fraction and number of grains per subsample is different from that of the 200-grain subsample from the 0.25–1.0 mm grain size fraction. As an example, we show in Figure 5a that the 0.25–1.0 mm grain size fractions with 100 and 200 grains have similar average compositions and smaller standard deviations compared to the other fractions studied. The 0.25–1.0 mm grain size subsample with c. 100 grains, and its duplicate, both yield high p-values of 0.96 and 0.77 (Fig. 5a). The 1.0–2.0 mm grain size fraction has an average composition similar to that of the 200-grains, 0.25–1.0 mm grain size fraction, with a high p value of 0.78, but the standard deviation is larger (Fig. 5a). In contrast, the 0.25–1.0 mm grain size fraction with c. 10 and 50 grains has an Al mean content different from that of the 100- and 200-grain fractions, a larger standard deviation, and yield low p-values of 0.19 and less (Fig. 5a).
(a) Al content in magnetite and hematite from sample 09-MOB001 for each grain size, and number of grain subsamples. Vertical lines show the range of values, the grey box is the standard deviation about the mean (horizontal bar). Red boxes represent detection limit ranges for Al. P values (green), test the hypothesis that the average Al content in each fraction is different from that of the 200-grain, 0.25–1.0 mm grain size fraction. (b) Comparison of the mean composition of Al, Cr, Mn, Si, Ti and V in magnetite and hematite from sample 09-MOB001 for each grain size, and grain number subsamples with respect to the 200-grain, 0.25–1.0 mm grain size fraction. Abbreviations: gr, grains; D, duplicate.
Figure 5b illustrates, as an example, p-values of each grain size fraction, and grain number, compared to the 200-grain, 0.25–1.0 mm grain size fraction of sample 09-MOB-001, for each element content with enough uncensored data to yield a valid statistical test. Figure 5b shows that p values are variable, especially for the coarse and the 10- and 50-grain, 0.25–1.0 mm grain size fractions, but that the 100-grain, 0.25–1.0 mm grain size fractions (original and duplicate) typically have higher p-values. The more reliable results are thus obtained with the 100-grain, 0.25–1.0 mm grain size fraction.
The composition of individual magnetite grains and the average magnetite composition for each subsample of 09-MOB-001 were plotted in the discriminant diagrams (Fig. 6) proposed by Dupuis & Beaudoin (2011). The purpose of the diagram is to show that the range in composition for all subsamples is similar and that the average composition of the sub-samples plot close together, with the exception of the coarse, 1.0–2.0 mm grain size fraction, which contains a small number of grains (Table 1).
Individual magnetite grain (empty symbol) composition, and subsample average (filled symbol), for each subsample from 09-MOB-001 till sample plotted in the discriminant diagrams proposed by Dupuis & Beaudoin (2011). (a) Ni+Cr v. Si+Mg diagram. (b) Al/(Zn+Ca) v. Cu/(Si+Ca) diagram. (c) Ni/(Cr+Mn) v. Ti+V diagram. (d) Ca+Al+Mn v. Ti+V diagram. Abbreviations: VMS, volcanogenic massive sulphides; IOCG, iron oxide-copper-gold; BIF, banded iron formation; Int., 0.25–1.0 mm grain size fraction.
Weight proportion of the ferromagnetic grain size fractions in till samples
The weight proportion of the three ferromagnetic grain size fractions from till samples relative to distance from the Sue-Dianne deposit is illustrated in Figure 7. This figure shows that the dominant size of ferromagnetic grains at, and immediately down-ice (west) of the deposit, is the 0.25–1.0 mm fraction, followed by the coarser 1.0–2.0 mm fraction. By contrast, <300 m from the deposit (down- or up-ice), the dominant size of ferromagnetic grains is <0.25 mm; the coarser grains (1.0–2.0 mm) comprise less than 5 wt.% of the ferromagnetic fraction, whereas the proportion of the intermediate grain size fraction (0.25–1.0 mm) is between 10 and 30 wt.%. The higher proportion of coarser ferromagnetic grains proximal to the Sue-Dianne deposit, and of smaller grains distal from the deposit is consistent with glacial comminution and abrasion in subglacial transport (Dreimanis 1990).
(a). Weight proportion of the three grain-size ferromagnetic fractions (<2.0 mm) of till samples relative to distance from the Sue-Dianne deposit. (b) Proportions of the principal minerals in the 100-grain, 0.25–1.0 mm grain size ferromagnetic fraction of till samples relative to the distance from the Sue-Dianne deposit.
Mineral proportions
The proportion of the principal minerals in the selected bedrock samples was determined for the optimal subsamples (100-grain, 0.25–1 mm size fraction). The quartz monzonite sample from the Marian River Batholith (09-CQA1010A4; Fig. 1b) is located in the up-ice flow direction, and is used to document the mineralogical background of least-altered bedrock. Its ferromagnetic fraction is dominantly composed of magnetite, with small proportions of silicates and other non-ferromagnetic oxides, as well as titanomagnetite and ilmenite (Fig. 8). The ferromagnetic grains of bedrock samples from the mineralized magnetite to hematite breccia complex (09-CQA1026A3 and 09-CQA1009B5; Fig. 1b) are dominantly composed of hematite, with magnetite, and rare silicates and other non-ferromagnetic oxides (Fig. 8). Sample 09-CQA1006D4 is altered but located <150 m from the Cu-Ag mineralized breccias (Fig. 1b), and thus has a similar proportion of hematite and magnetite to that of the two mineralized samples (Fig. 8). Bedrock sample 09-CQA0005A4 (Fig. 1b) is from unmineralized feldspar porphyry Faber Group rocks, and it contains mostly titanomagnetite ilmenite, and magnetite, with few silicates and other non-ferromagnetic oxides (Fig. 8).
Proportions of the principal minerals in the ferromagnetic fraction of bedrock samples from the Sue-Dianne deposit area.
The mineral proportions of ferromagnetic grains in till samples were measured using subsamples (100-grain, 0.25–1.0 mm grain size fraction) from 8 till samples at various distances along the dominant ice-flow trend (09-MOB003, -004, -001, -002, -022, -029, -035 and -015; Fig. 1b). The proportions of the principal minerals from the ferromagnetic fraction vary with respect to distance from the Sue-Dianne deposit (Fig. 7). Up-ice, hematite is the larger mineral in less than 5% of the grains, whereas magnetite and titanomagnetite/ilmentite each constitute nearly 40% of the ferromagnetic fractions. Directly over and immediately down-ice of the deposit (<300 m), hematite is the most common mineral in the ferromagnetic fraction with a proportion up to 60%, followed by magnetite, titanomagnetite and/or ilmenite. At these sites, the ferromagnetic fraction contains a small proportion (less than 5%) of silicates, other non-ferromagnetic oxides and sulphides, locally with small inclusions of magnetite and/or hematite. As distance increases down-ice from the deposit, the proportion of hematite decreases rapidly to less than 10% <1 km from the deposit, whereas the proportions of magnetite and titanomagnetite/ilmenite both increase up to 55% each (Fig. 7). The proportion of silicates, non-ferromagnetic oxides and sulphides is low (close to 10%) immediately down-ice and it increases at distances greater than 2 km.
Magnetite and hematite composition
A majority of hematite and magnetite grains from mineralized rocks (09-CQA1026A3 and 09-CQA1009B5) plot in, or near, the IOCG and BIF deposit fields with some analyses plotting in the fields for Kiruna and porphyry deposit types in the Ni/(Cr+Mn) v. Ti+V and Ca+Al+Mn v. Ti+V diagrams (Fig. 9). It must be noted that the fields for mineral deposit-types in the diagrams of Figure 9 are based on the average composition from representative deposits of each type (Dupuis & Beaudoin 2011). It is therefore not expected that individual grain analyses will have a wider compositional range than deposit averages, as shown for the Reko Diq porphyry-Cu deposit (Dupuis & Beaudoin 2011). Magnetite and hematite from altered, but barren sample 09-CQA1006D4 plot at higher Ti+V values, dominantly within, or near the field for IOCG deposits, with few analyses in the fields for Kiruna and porphyry deposit (Fig. 9). The variation in Ti+V is mainly due to an increase in V content, whereas Ti content remains within the same range of values. Magnetite from unmineralized country rocks (09-CQA0005A4 and 09-CQA1010A4) plot at still higher Ti+V values, mostly within or near the fields for Kiruna and porphyry deposit types (Figs 9a, c). Similar compositions are observed in magnetite and hematite from the NICO deposit area where barren host rocks and least-altered regional rocks have higher Ti+V composition relative to mineralized rocks (McMartin et al. 2011).
Magnetite and hematite grain composition for the representative ferromagnetic subsamples (100-grain, 0.25–1.0 mm size fraction) prepared from selected bedrock samples of the Sue-Dianne area. (a) & (b) Individual magnetite and hematite composition in the Ni/(Cr+Mn) v. Ti+V discriminant diagram (Dupuis & Beaudoin 2011). (c) & (d) Magnetite and hematite grain composition in the Ca+Al+Mn v. Ti+V discriminant diagram (Dupuis & Beaudoin 2011). Abbreviations: BIF, banded iron formation; IOCG, iron oxide-copper-gold.
Figure 6a shows that all but two grains from one till sample plot outside the field for Ni-Cu deposits. This result is consistent with the low potential for this deposit-type in the GBmz, as shown by lack of occurrence of this deposit-type in the region. Figure 10 shows the composition of magnetite and hematite from till samples up-ice, at the Sue-Dianne deposit, and down-ice from the deposit in the Ni/(Cr+Mn) v. Ti+V and Ca+Al+Mn v. Ti+V diagrams of Dupuis & Beaudoin (2011). Magnetite from till at the deposit clusters below the fields for IOCG and Kiruna deposits (Figs 10a, c). Magnetite from till up- and down-ice shows a similar cluster but with a wider spread of values in the fields for porphyry and Fe-Ti deposit types (Figs 10a, c). Hematite at the deposit plots mostly in the fields for IOCG and BIF deposit types (Figs 10b, d). Hematite from till samples both up- and down-ice from the Sue-Dianne deposit is characterized by higher Ti+V values (Figs 10b, d).
Magnetite and hematite grain composition in the 100-grain, 0.25–1.0 mm grain size ferromagnetic fraction of till samples up-ice (09-MOB003, 09-MOB004, and 09-MOB001), at the Sue-Dianne deposit (09-MOB002 and 09-MOB022), and down-ice (09-MOB029, 09-MOB035, and 09-MOB015). Magnetite (a) and hematite (b) grain composition in the Ni/(Cr+Mn) v. Ti+V discriminant diagram (Dupuis & Beaudoin 2011). Magnetite (c) and hematite (d) composition in the Ca+Al+Mn v. Ti+V discriminant diagram (Dupuis & Beaudoin 2011). Abbreviations: BIF, banded iron formation; IOCG, iron oxide-copper-gold; Mag, magnetite; Hem, hematite.
Thompson Nickel Belt Ni-Cu deposits
Mineral proportions
The ferromagnetic fractions of till samples collected around the Pipe deposit in the TNB include magnetite, titanomagnetite, ilmenite, hematite, chromite, silicates, sulphides (mostly pyrrhotite), and rare goethite, rutile, zircon, and biotite. The most abundant phases are magnetite, titanomagnetite, silicates, and hematite (Pozza 2011). Non-magnetic minerals, such as silicates, contain inclusions of magnetite and titanomagnetite that are too small to be analysed. Along the north–south profile from the Pipe deposit, the abundance of magnetite is lowest on the eastern edge of the Pipe open pit and increases up to 61% at and immediately down-ice (south) of the deposit (Fig. 11a). The abundance of titanomagnetite, ilmenite, silicates, and other oxides and sulphides in till shows the inverse relation with the deposit location, whereas the hematite proportion remains fairly constant with low abundances of less than 5% all along the north–south transect (Fig. 11a).
Proportion of each mineral type contained in the 0.25–1.0 mm ferromagnetic fraction of till samples relative to the distance from the Pipe Ni-Cu deposit, along the direction of (a) the older southward ice flow and (b) the younger westward ice flow.
Along the east–west transect, the abundance of magnetite grains increases up to 88% at and immediately down-ice of the Pipe deposit, whereas the abundance of other minerals decreases in proportion (Fig. 11b). Further down-ice to the west, the abundance of magnetite grains gradually decreases, whereas that of titanomagnetite, ilmenite, silicates, other oxides and sulphides all increases (Fig. 11b). The abundance of hematite grains in till samples slightly increases toward the Pipe deposit, but decreases rapidly to less than 5% down-ice to the west, before slightly increasing again to less than 10% (Fig. 11b).
Magnetite and hematite composition
Magnetite and hematite compositions in till samples along the north–south and east–west transects across the Pipe deposit were plotted in the Ni+Cr v. Si+Mg discriminant diagram of Dupuis & Beaudoin (2011; Fig. 12). In the c. 2-km long north–south transect, 8–12% of the till magnetite grains from samples near the edge of the Pipe mine pit plot in the field for Ni-Cu deposits (Fig. 12a). Although there are fewer analyses, 9–17% of the magnetite grains from till down-ice of the Pipe deposit plot in the Ni-Cu deposit field (Fig. 12a). There are no analyses up-ice because the transect starts at the pit location. Along the c. 170-km long east–west transect, most magnetite and hematite up- and down-ice from the Pipe mine have Si+Mg contents lower than those typical for Ni-Cu deposits, but c. 4–10% of the grains plot in the field for Ni-Cu deposits (Fig. 12b). Magnetite and few hematite grains near the Pipe open pit have a composition spreading the whole range of Si+Mg values of up- and down-ice till samples, with 12% of the grains plotting in the field for Ni-Cu deposits (Fig. 12b).
Magnetite and hematite grain composition for all till samples from the TNB in the Ni+Cr v. Si+Mg discriminant diagram (Dupuis & Beaudoin 2011) for (a) the older southward ice flow, and (b) the younger westward ice flow. Abbreviations: Mag, magnetite; Hem, hematite.
Discussion
Optimum grain size fraction and number of grains
Examination of the ferromagnetic fraction of till samples from the Sue-Dianne IOCG and TNB Ni-Cu deposits indicates that the optimal subsample to characterize the mineralogical composition of the ferromagnetic fraction and the chemical composition of magnetite and hematite from the ferromagnetic fraction should be 0.25–1.0 mm in size and contain c. 100 detrital grains. Smaller grains (<0.25 mm) are difficult to handle without losing grains, to count during binocular examination, and to analyse by EPMA. The Sue-Dianne IOCG deposit case study shows that subsamples from the 0.25–1.0 mm size fraction typically contain a sufficient number of grains to study and that duplicate subsamples can be prepared for quality control (Table 1). In contrast, the coarser grain size fraction (1.0–2.0 mm) does not always contain a sufficient number of grains to separate a 100-grain subsample. The results from the 200-grain subsamples are not significantly better than those obtained using 100-grain subsamples and the larger number of grains takes more time to analyse. The estimated mineral proportions in the subsamples from the coarse (1.0–2.0 mm) and the 0.25–1.0 mm grain size fractions with 50 and 10 grains do not show satisfactory reproducibility and differ with the proportions determined for the 0.25–1.0 mm grain size fractions with 100 or 200 grains (Table 1 and Fig. 5). The analysis of subsamples with about 10 0.25–1.0 mm sized grains yields poor reproducibility and low probability of capturing the chemical composition of the ferromagnetic fraction (Fig. 5) due to the small number of analysed grains (Table 1). The coarse (1.0–2.0 mm) subsamples from the selected three till samples show a wide variation in the number of magnetite grains (Table 1), such that some samples do not contain enough magnetite grains to obtain a reliable composition.
Till iron oxide composition along glacial flow paths
Sue-Dianne IOCG deposit
The proportion of magnetite and hematite grains in the representative subsamples of the ferromagnetic fraction that plot in the field of various mineral deposit types in the Ca+Al+Mn v. Ti+V discriminant diagram (Figs 10c, d) is shown with respect to the distance from the Sue-Dianne deposit in the direction of ice flow (Fig. 13). A large number of till magnetite grains and few till hematite grains have Ti+V contents similar to those of IOCG and Kiruna deposit types, but plot at lower Ca+Al+Mn values in Figures 10c and 10d. In Figure 13, we have grouped these grain analyses under the label ‘Below IOCG-Kiruna’, because the lower limit of these two deposit type fields (IOCG and Kiruna) is not constrained by compositions typical of other deposit types (Dupuis & Beaudoin 2011). Along the SW direction of ice-flow, the proportion of magnetite plotting in the field of IOCG and BIF deposits, and below the combined fields for IOCG and Kiruna deposits, increases at the location of the Sue-Dianne deposit, and then decreases as a consequence of dilution by the eroded bedrock material down-ice of the source of these grains (Fig. 13a). Magnetite grains that plot in the field for other deposit types either decrease (e.g. Kiruna) or show no significant variation at the location of the Sue-Dianne deposit (Fig. 13a).
Magnetite (a) and hematite (b) proportion of grains with the chemical signature of various deposit types relative to the distance from the Sue-Dianne deposit. The proportions for the IOCG, porphyry, Kiruna, skarn, BIF, and Fe-Ti-V deposits fields were estimated using the Ca+Al+Mn v. Ti+V discriminant diagram (Figs 14c, d) Abbreviations: BIF, banded iron formation; IOCG, iron oxide-copper-gold.
Hematite is absent in till samples up-ice from the Sue-Dianne deposit (Fig. 13b). Down-ice of the Sue-Dianne deposit in the SW direction of ice-flow, the proportion of hematite grains with a composition typical of BIF decreases rapidly away from the deposit (Fig. 13b). The proportion of hematite grains with an IOCG signature remains high more than 3 km SW of the deposit, where most of the hematite grains carry the signature of porphyry and Fe-Ti-V mineral deposit types (Fig. 13b). The occurrence of magnetite grains in till with the chemical signature of porphyry-Cu deposits is consistent with the presence of outcropping mineralization typical of porphyry-Cu mineral deposits (Mumin et al. 2010).
Thompson Ni-Cu deposits
Along the trend of the north–south ice flow in the TNB, the proportions of magnetite grains with the signature of Ni-Cu deposits are 8% and 12% on both the east and west sides of the Pipe open pit, and 7% at the southern edge of the pit (Figs 3, 14a). Approximately 1.2 km south and down-ice of the Pipe deposit, the proportion of magnetite grains with the signature of Ni-Cu deposits increases to 17% (Fig. 14a). The interpretation of magnetite composition along the north–south profile must take into account that other Ni-Cu deposits outcrop to the north of the Pipe mine (Fig. 2). Considering the size of the outcropping massive sulphides (Fig. 3), the fact that a small proportion of magnetite grains with the signature of Ni-Cu deposits is considered significant. Other grains with chemical compositions plotting outside the field for Ni-Cu deposits (Fig. 12) must have been eroded from distant bedrock along the ice-flow path. Along the trend of younger east–west ice flow, the background proportion of magnetite grains with the signature of Ni-Cu deposits varies between 4 and 10% (Fig. 14b). At the Pipe open pit, the proportion increases sharply to 13% and then decreases to lower proportions (3–6%) less than 1 km west of the Pipe open pit (Fig. 14b). The relation between the till samples and the north–south and east–west ice-flow directions cannot be established with precision in the TNB. The 10% proportion of magnetite grains in till subsamples overlying Archean gneisses east of the TNB is noteworthy (Figs 2, 14b). The till samples also contain carbonate clasts eroded more than 200 km to the east of the TNB from the Paleozoic rocks of the Hudson Bay Lowlands (McClenaghan et al. 2009). Potential sources for the magnetite with a composition typical of magmatic Ni-Cu sulphides are the Proterozoic intrusions of the Molson dyke swarm and the Fox River sill, in particular because it contains disseminated magmatic sulphides (Scoates & Eckstrand 1986; Hulbert et al. 2005), and which are on the westerly ice-path from the Hudson Bay Lowlands, not unlike the carbonate clasts found in the same till samples.
Proportion of magnetite grains plotting in the Ni-Cu deposit field relative to the distance from the Pipe open pit (TNB), along the direction of (a) the older southward ice flow, and (b) the younger westward ice flow.
Conclusions
The optimal number of grains, and grain size fraction, of a subsample useful for determining the mineralogical and compositional range of iron oxides from the ferromagnetic fraction of the heavy mineral concentrate from a till sample is 100 grains from the 0.25–1.0 mm grain size fraction. The weight proportion of the coarser grain size fractions (>0.25 mm) from a till sample is higher immediately down the ice-flow direction where a mineral deposit has been eroded by glaciers.
At the Sue-Dianne IOCG deposit, the proportion of hematite in the ferromagnetic fraction of the heavy mineral concentrate from a till sample increases sharply down the ice-flow direction, to a few hundred meters from the deposit location. At the Pipe Ni-Cu deposit, the proportion of magnetite and hematite in the ferromagnetic fraction of the heavy mineral concentrate from till samples increases down the ice-flow directions (north–south, and NE–SW) from the deposit location.
The chemical composition of magnetite and hematite in the ferromagnetic fraction of the heavy mineral concentrate from a till subsample can provide information about the bedrock source of the iron oxides. The chemistry of magnetite and hematite in till yields the signature of the mineral deposit types known to occur in the regional country rocks. The proportions of magnetite and hematite with the chemical signature typical of various mineral deposit types change in the down-ice direction and can be used to detect the occurrence of an eroded mineral deposit.
Acknowledgments
This research was funded by the Québec DIVEX research network, the Natural Science and Engineering Research Council of Canada (CRD and Discovery grants to GB), Vale, and the Geological Survey of Canada (GSC). Samples from the Sue-Dianne area were collected by P. Normandeau as part of the IOCG-Great Bear Multiple Metals Project within the Geo-mapping for Energy and Minerals Program (GSC). Samples from the TNB were collected as part of the GSC’s Targeted Geoscience Initiative 3 (TGI-3) Program with assistance from the Manitoba Geological Survey and Vale. We would like to thank Marc Choquette for technical support during microprobe analyses. Comments and suggestions from Sarah-Jane Barnes and Sarah Dare (UQAC) during the research project were greatly appreciated. We thank M. Leybourne and P. Nadoll for their detailed comments that improved significantly the paper.
- © 2014 AAG/The Geological Society of London
References
Optimal ferromagnetic fraction in till samples along ice-flow paths: case studies from the Sue-Dianne and Thompson deposits, Canada
