The Svartliden granite
Petrography, whole rock geochemistry and stable isotope composition
Joel Andersson
Master of Science
Exploration and Environmental Geosciences
Luleå University of Technology
The Svartliden granite
Petrography, whole rock geochemistry and
stable isotope composition
In this MSc thesis a characterization of the field appearance, petrogeochemistry, mineralogy, and oxygen/hydrogen isotope composition of the Svartliden granite was done. The Svartliden granite is a fine to medium grained two mica leuco granite (weakly mineralized by pyrrhotite) containing metasedimentary xenoliths and is intensely varying in character; granular-aplitic-pegmatitic. The Svartliden granite belongs to the Skellefte-Härnö granitic suite and is of S-type character. The protholith is most probably metasedimentary rocks belonging to the surrounding Bothnian Basin. Normally immobile elements such as Y, Ho, Zr and Hf have possibly been mobile in the fluid-rich magmatic Svartliden granite system. This makes the Svartliden granite interesting from a prospecting point of view because other normally immobile elements also possibly could have been mobile and concentrated. A spatially close Skellefte-Härnö pluton shows similar characteristics based on field appearance, deformation style, modal and nominal mineralogy; whole rock geochemistry, aluminum saturation, REE-signatures and oxygen/hydrogen isotope composition. Two spatially Svartliden-related Revsund plutons have been characterized for comparison. The Revsund granites are white to reddish, medium grained, equigranular-porphyritic biotite granites (s.s.) containing hornblende, titanite, and magnetite. It is possible to distinguish the Revsund granites from the Skellefte-Härnö granite (including the Svartliden granite) in the study area by the means of the above mentioned characteristics. The Revsund granites are of I-type character but show S-type affinities. The latter is probably caused by assimilation of metasedimentary material from the Bothnian Basin giving rise to for instance high A/CNK ratios and δ18O-values. The reddish colour of one of the Revsund plutons is
caused by hematite alteration of magnetite and pyrite. This study provides further data in support of a sedimentary protholith generating the Skellefte-Härnö granite, and also data valuable for a continued discussion of the genetic aspects of the Revsund granitoid.
I thank Dragon Mining Sweden AB for the financing of this project.
I thank Johan Söderhielm for his heroic efforts on the ATV during the sampling campaign and Henrik Ask, Roman Hanes and Chris Gordon for valuable inputs.
I thank Ulf Bertil Andersson for constructive criticism and discussions.
I thank the Scottish Universities Environmental Research Centre (SUERC) for doing the stable isotope analyses.
Last but not least I thank my supervisor Dr. Glenn Bark at the Luleå University of Technology for scientific inputs and for supporting me along this project.
Introduction
...1Purpose of the study……….…….1
Svartliden Au deposit………...1
Regional geology………...3
Methods
...7Petrography……….7
Whole rock geochemistry………..7
Stable isotope chemistry……….7
Results
……….11Detailed geological mapping/drill core logging………11
Petrography………..17
Whole rock geochemistry………27
Classification………..27
Nominal mineralogy and aluminium saturation………32
Delineation plots……….34
Geotectonic discrimination………..………36
Normalized multi element plots………38
Rare Earth Elements……….40
Stable isotope composition………43
Discussion
……….………45Petrography………..……45
Whole rock geochemistry………46
Classification………..……46
Nominal mineralogy and aluminium saturation………48
Delineation plots……….………49
Tectonic discrimination………..………50
Normalized multi element plots………51
Rare Earth Elements……….……52
Stable isotope composition………54
Summary of discussion………..………55
Concluding remarks
………57Recommendations
………..59References
………61Appendix 1
………..………71Appendix 2
………..………72Appendix 3
………..………73Introduction
In the summer of 2010 the author worked as a geologist at the Svartliden gold mine (Fig.1), owned by Dragon Mining Sweden AB, in this study referred to as DAB. While logging drill cores, hundreds of meters of a light-coloured, fine to medium grained, two mica leuco granitoid was not given any greater notice. However, the granitoid at Svartliden is beautifully exposed in the open pit and was the target for discussions among the geologists at DAB that summer. When the author, in the late stage of the same summer, was asked to characterize the granite as a part of his upcoming MSc thesis, the positive answer was not late to come. The author then contacted Dr. Glenn Bark, who had spent many years studying the nearby Fäboliden gold deposit during his PhD studies, to be his supervisor. Together we decided the methods suitable for this study and the field work started immediately, just a few weeks before the first snow started to fall.
The purpose of this study is to provide a geochemical, mineralogical and isotopic characterization of the Svartliden granitoid, in this study given the abbreviation SV, to better understand the geological setting surrounding the Svartliden gold deposit. The characterization was performed by the means of detailed geological mapping, petrography, whole rock geochemistry and hydrogen and oxygen isotope composition. The study aims to answer what type of granitoid SV is and whether SV belongs to the regionally present Skellefte-Härnö or Revsund granitic suites. A better understanding of the local geology is of interest for future exploration of this area. The Skellefte-Härnö granite is in this study given the abbreviation SH whereas the Revsund granite is divided into two types given the abbreviations RE and RÖ.
Svartliden Au deposit
The Svartliden Au deposit is situated southwest of the well-known Skellefte volcanic hosted massive sulphide (VHMS) district (Fig. 1). The gold mine is situated near the town of Lycksele in the Swedish part of the Svecofennian Domain which is a part of the Fennoscandian Shield (Gaál and Gorbatschev 1987). The area is commonly referred to as the “Gold Line” (Fig.1) or the “Lycksele-Storuman gold ore province” (Bark and Weihed 2003). The Gold Line was first identified as a linear roughly northwest-trending till-geochemical Au anomaly during a sampling campaign executed by the Geological Survey of Sweden (SGU) in 1985.
The Svartliden Au-anomaly was discovered in October 1994 after a boulder tracing campaign. A trench-sampling programme was conducted by Svartliden Guld AB soon thereafter. Further trenching the following year together with electromagnetic (EM) surveys as well as a four hole diamond drilling
program, resulted in the identification of the gold-bearing Svartliden structure (Hart et al. 1999). A joint venture was executed together with the company Viking Gold Corporation and after a few turns the claim to the area ended up with the Australian listed company Dragon Mining NL (Grahn et al. 2001). Svartliden was brought into production in March 2005 by open pit mining. DAB could by the end of 2010 report measured, indicated and inferred Au resources (including reserves) with a total of 1 330 100 tons of ore at 3.7 g/t from both open pit and underground mining (Dragon Mining Annual Report 2010) which started in late 2011. The Svartliden deposit is likely a hypozonal orogenic gold deposit with skarn alteration and silicification (Bark 2008; Samskog 2011).
The Svartliden gold mine is, at present (April 2012), the only operating mine in the Lycksele-Storuman gold ore province since the Ersmarksberget (Zn-Pb-Au-Ag) recently shut down its production. The area hosts several promising Au deposits, such as the Fäboliden hypozonal orogenic gold deposit (Bark and Weihed 2003; Bark and Weihed 2007; Bark et al. 2007; Bark 2008), situated ~35 km southeast of Svartliden (Fig.1).
Fig. 1 Till geochemical Au concentration in the Lycksele-Storuman zone (Bark 2005). Original data from the Geological Survey of Sweden (SGU).
Regional geology
The evolution of the Fennoscandian Shield, which includes the South West Scandinavian, Svecofennian and Karelian Domains (Fig. 2), is a complex story. According to Gaál and Gorbatschev (1987) four main orogenic events are to be mentioned; Samian 3.1-2.9 Ga, Loopian 2.9-2.6 Ga, Svecokarelian 2.0-1.75 Ga, and Gothian orogenies 1.75-1.5 Ga, which all formed new continental crust. The Sveconorwegian-Grenvillian 1.25-0.9 Ga (Gaál and Gorbatschev 1987) and the Caledonian 0.55-0.32 Ga (McKerrow et al. 2000) orogenies were, according to Gaál and Gorbatschev (1987), periods of reworking of already existing crust. The evolution of the Svecofennian Domain during the Svecokarelian orogeny (2.0-1.75 Ga), involved substantial crustal thickening. It also laid the foundation for several ore districts, among them, the Proterozoic Skellefte volcanic hosted massive sulphide district (Weihed et al. 2005; Weihed 2010). Details concerning the dynamics of the Svecokarelian orogeny have been under discussion. Lahtinen et al. (2003, 2005, 2009) and Korja et al. (2006) argue for the identification of five distinct, but partly overlapping orogenies during this period; the Lappland-Kola, the Lappland-Savo, the Fennian, the Svecobaltic, and the Nordic orogenies. Lahtinen et al. (2003, 2005, 2009) also points out “orogenic collapse” at 1.79-1.77 Ga as an important feature which was followed by stabilization of the Fennoscandian Shield.
The Skellefte district is interpreted as a metamorphosed Proterozoic island arc including a related back-arc basin(s) with a modern analogue found in Japan (Allen et al. 1996). This is in agreement with Vivallo and Claesson (1987) who pointed out, based on Ti, Zr, Y, Cr and REE, the geochemical similarities that exist between the volcanic rocks of the Skellefte district and modern subduction-related back-arc-basins. Weihed et al. (1992) argue that the metallogeny itself constitutes a key in the tectonic interpretation that can be explained as a destructive plate margin with a related back-arc-basin. The accretion of this volcanic arc of Proterozoic age onto the Archaean terrain in the present north is described by Mellqvist et al. (1999) and Öhlander et al. (1999). They showed evidence of a Proterozoic-Archaean paleoboundary, stretching between Luleå and Jokkmokk. The area south of the Skellefte district became the target of discussions when the results from the seismic research project BABEL were published (BABEL Working Group 1990, 1993), see Figure 2. By the means of marine deep seismic investigation the BABEL-group could image a probable subduction structure dipping NE, striking NW-SE. The subduction structure did not only have an impact on the geological models explaining the Svecokarelian orogeny, but also added geophysical evidence for plate tectonic activity during the Palaeoproterozoic. This was earlier only indicated by geological observations in old blue-shists and ophiolites (BABEL Working Group 1990, 1993).
In the context of this particular study, the opening of the pre-Svecokarelian ocean at 1.95 Ga, the formation of an island-arc between 1.91-1.87 Ga, and the accretion of it onto the Archaean Domain (Nironen 1997), is of particular importance. This is so because these Svecokarelian events laid the foundation for the Bothnian Basin and its associated granites which are of concern for the current study.
The Bothnian Basin is a large structural depression extending from northern Sweden to southern Finland (Kumpulainen 2009). Lycksele (Fig.1) is situated on the northern margin of this depression. This marine volcano-sedimentary basin lacks any clear original basin margins, but indications of such margins are to be found in the Kalix area, Sweden, approx. 400 km northeast of Lycksele (Kumpulainen 2009). The Bothnian Basin is limited by the Ljusdals Batolith in the south and the Skellefte district in the north (Andersson et al. 2004; Högdahl et al. 2008). The original basin was probably formed in the early Paleoproterozoic when intra-continental rifting of the Archaean plate created a sediment trap on top of the Archaean (Kumpulainen 2009), indicating an Archean basement which was suggested by e. g. Lundqvist (1987) and Andersson et al. (2002), in what has been described as the pre-Svecokarelian (pre-Svecofennian) ocean (Nironen 1997). Eventhough weathering occurred at lower rates than today (Weihed et al. 2005) considerable sediment accumulation resulted in >10 km thick mainly turbiditic greywacke sequences (Claesson and Lundqvist 1995; Kumpulainen 2009). Different estimates of the Archaean contribution to the sediment package, based on Nd isotopes (Patchett et al. 1987; Lundström 1998) and U-Pb zircon datings (Claesson et al. 1993) have been done. Claesson et al. (1993) indicates 30% Archaean,
Patchett et al. (1987) argues for <40% Archaen wheras Lundström (1998) states 41-54% Archaean contribution to the Bothnian Basin.
The western (Swedish) part of the Bothnian Basin was intruded by four distinct granitoid suites. These have been described in relation to the Svecokarelian orogeny as early, late, post and anorogenic (e. g. Claesson and Lundqvist 1995) and together they dominate the bedrock of the Svecofennian Domain. However, this categorization is not free from criticism and Andersson (1991) argues that the terms “late”- and “post”-orogenic are not valid and should be abandoned. This study agrees with the criticism and is aware of the confusion these terms might bring, not least in the area of investigation in this particular study. Despite of the argumentation of Andersson (1991), and despite of an remarkably old example of a “post” orogenic granite within the study area, the current study will continue to use the traditional sub-division of the Svecofennian granitoids. This study is mainly concerned with granitoid rocks formed syn- to post-orogenic.
The oldest granitoids, the “early orogenic granitoids”, intruded the rocks in the Bothnian Basin during a period of 30-40 m. y., at approx. 1.90-1.86 Ga (Gaál and Gorbatschev 1987). These granitoids, together with early volcanic rocks, constitute the first Svecofennian continental crust and are dominated by tonalitic to granodioritic rock types which show evidence of felsic-mafic magma interaction (Andersson 1991). They generally consist of differentiated suites of calcic to calc-alkaline I-type rocks (Högdahl 2000). To this group the by Wasström (2005) petrographically characterized and by Eliasson and Sträng (1998) dated at ~1.95 Ga Knaften granite-granodiorite-tonalite-trondhjemite belongs. The dating of the Knaften granitoid (Eliasson and Sträng 1998) indicates the rock suite as the oldest documented Svecofennian rock within the Fennoscandian Shield (Weihed et al. 2005). The Knaften area is found approximately 20 km southeast of Lycksele (Fig. 2). The early orogenic granitoids have been subjected to deformation and metamorphism during the Svecokarelian orogeny. The minimum age of the metamorphism is 1.82 Ga (Claesson and Lundqvist 1995) and it is characterized by low pressure amphibolite facies even though granulite facies has been documented locally according to Högdahl (2000, 2008).
During the later stages of the Svecokarelian orogenic activity the “Late orogenic granitoids” 1.83-1.80 Ga (Claesson and Lundqvist 1995, Wasström 2005), were emplaced. The magmatic activity has been interpreted as being the result of the over-thickened Svecofennian crust (Lindh 2005). This granitic suite were mainly formed by the recycling and re-melting of metasedimentary material, giving them a typical S-type character (Gaál and Gorbatschev 1987). According to e. g. Gaál and Gorbatschev (1987) and Claesson and Lundqvist (1995) etc., they form mostly smaller massifs and dyke swarms which at the time of their formation were isolated. To this group the Härnö-suite (Claesson and
Lundqvist 1995; Lindh 2005) belongs. Claesson and Lundqvist (1995) obtained a 206Pb/238U-dating of
1822 ± 5 Ma, and showed REE and trace element model calculations that indicated metasediments belonging to the Bothnian Basin as a probable source for these granitic rocks. Lundström (1998) could show evidence of metamorphic temperatures of around 625 C° and pressures of 2.7 kbar (upper amphibolite facies) during the emplacement of a Skellefte-Härnö pluton near the area of study. According to Weihed et al. (1992) the emplacement of the Skellefte-Härnö granite marks the onset of the cratonization of the Fennoscandian Shield.
The post-orogenic Revsund granite was emplaced between 1.81-1.77 Ga (e. g. Claesson and Lundqvist 1995; Billström and Weihed 1996; Gorbatschev 2004). This event represents the youngest phase of the Paleoproterozoic magmatism in the central Fennoscandian Shield. The chemical composition of the post-orogenic Revsund granitoid is granite (s.s.) grading into granodiorite (e. g. Claesson and Lundqvist 1995; Högdahl 2000) but also commonly showing monzonitic compositions (Andersson 1997; Gorbatschev 2004). The character is mainly metaluminous (e. g. Claesson and Lundqvist 1995; Högdahl 2000; Gorbatschev 2004) to peraluminous (Wilson and Åkerblom 1982). The Revsund granite has experienced diverse interpretations with an S-type affinity suggested by Wilson (1980) and Armands and Xefteris (1987), whereas Claesson and Lundqvist (1995) argued for an I-type character. Weihed et al. (1995) and Gorbatschev et al. (2004) describe the Revsund granite as a mixed I- to A-type granitoid. However, the granitoids of Revsund type cover large areas in the Bothnian Basin, and Wilson (1980) mentions that the term Revsund …”has been applied over too large an area”.
To be mentioned are also the much later anorogenic Mesoproterozoic Rapakivi A-type granitoids, emplaced within the Bothnian Basin at approx. 1.4-1.5 Ga and studied by many (e. g. Åberg 1988; Andersson 2001; Lindh et al. 2001; Andersson et al. 2002), eventhough it is outside the scope of the current study.
Methods
Detailed geological mapping and sampling was performed during one week in October 2010. The mapping was done in order to identify suitable sample locations and to understand the field characteristics of the RE, RÖ and SH granites. Each sample was taken with a rock saw (Husqvarna) and a hammer (Fig. 3). The coordinate system RT/90 was used which was later recalculated to the new SWEREF 99 system (see Appendix 1 for sample locations). Sample locations and structural measurements were imported into MapInfo 10.0 with the map sheet 22H Järvsjö NO 1:50 000 (Björk and Kero 2001) as background (Fig. 4). Core logging and sampling was performed during four days in January 2011, covering the SV granite, using the local mine grid as coordinate system.
Petrography
The petrographic samples were prepared at Luleå University of Technology and sent to Vancouver Petrographics, Canada, for production of thin sections with a preferred thickness of 30µm. The total number of thin sections included in this study is 30. 10 Revsund (5 each for RE and RÖ), 10 Skellefte-Härnö (SH) and 10 from the Svartliden granite (SV). The petrographic analysis was performed at Luleå University of Technology using a NIKON polarizing microscope with a digital camera attached.
Whole rock geochemistry
All samples were cut in order to get rid of erosion. The samples were then sent to ALS Minerals, Piteå (Sweden), where further sample preparation was performed prior to chemical analysis at the ALS Chemex lab in Canada. The samples were analysed according to a multi-element characterization package + Au, using inductively coupled plasma atomic emission spectrometry (ICP-AES), LECO combustion analysis and inductively coupled plasma mass spectrometry (ICP-MS). Analysis of the data set was performed using Excel and Geochemical Data Toolkit (Janoušek et al. 2006).
Stable isotope chemistry
The isotope samples were crushed and separated by the means of magnetism prior to handpicking of biotite and quartz. 30 samples of biotite and 30 samples of quartz were prepared. The mineral concentrates were sent to the Scottish Universities Environmental Research Centre (SUERC) Isotope Lab, Scotland, for H and O isotope analysis. Due to the fine-grained nature of the biotite and the intergrown biotite-quartz in the SH- and SV granite 5 samples of biotite could not be handpicked with an acceptable purity (>90%) and was not analysed for H isotope composition. All samples for RE- and RÖ granite could be analysed for H isotope composition. All mineral samples (concentrates) of quartz
could be analysed for O isotope composition. The quartz was analysed by the means of a laser fluorination procedure, letting the total sample react with excess CIF3 with a CO2 laser as a heat source in accordance with Sharp (1990). The procedure resulted in 100% release of O2 from the crystal lattice. The free O2 was then let to react with hot graphite to produce CO2 which was analysed on-line by a VG SIRA 10 spectrometer. The results are reported in standard notation (δ18O) as per mil (‰) deviation from the Standard Mean Ocean Water (SMOW) standard. The biotite was analysed for H isotope composition by the means of a procedure starting with heating overnight under high vacuum at 150°C in order to release labile volatiles after loading into thoroughly outgassed Pt crucibles. Gradual heating by radiofrequency induction in an evacuated quartz tube to temperatures exceeding 1200°C was then performed. The released H2O was reduced to H2 in a chromium furnace at 800°C following Donnely et al. (2001). The evolved gas was quantitatively measured with an Hg manometer and analysed on a VG 602D mass spectrometer. The results are presented as the standard notation (δD) as per mil (‰) deviation from Standard Mean Ocean Water (SMOW) standard.
Fig. 3 Sampling of the reddish Revsund (RE). The rock saw and water cooling tank is shown on the left hand side of the picture.
Fig .4 S im pl ifie d g eo lo gic al m ap o f t he s tu dy a re a m od ifi ed a fte r B jö rk a nd Ke ro ( 2001 ). O rig in al ma p w as re pr od uc ed b y th e me an s o f S GU o nl in e d ata b as e. T he ma pp in g c amp ai gn w as c on ce ntr ate d to th e ma rke d ar ea s.
Results
Detailed geological mapping/drill core logging
Revsund granite (white), RE
The white RE granite (Fig. 5) was mapped as “Revsund granite” by the SGU (Björk and Kero 2001). It is the dominant granitoid within the map sheet Järvsjö NO. It forms several interconnected plutons in the southern part of the map sheet and is surrounded by metasediments, mainly metagreywackes, belonging to the Bothnian Basin (Fig. 4). The plutons are crosscut by late dolerite dykes and contain according to Björk and Kero (2001) metasedimentary xenoliths. However, no such xenoliths were found during the mapping of the RE granite in this particular study area. The depth of weathering reaches in places >15 cm and the outcropping is probably <1% within the map sheet.
The rock type (RE) is a light coloured, equigranular-porphyritic, medium-grained granitoid with a significant amount of biotite occurring as patches and in the matrix (Fig. 5). Biotite is the dominant mafic mineral. Phenocrystic potassium feldspar (<2cm), mainly orthoclase, is a common feature. The rock type is texturally homogeneous and undeformed within the study area.
Fig. 5 Typical appearence of the whitish Revsund granite. Picture shows a fresh surface. The area is approx. 20 cm wide.
Revsund granite (reddish), RÖ
The reddish RÖ granite is mapped as “Revsund granite” by the SGU and is the second dominant rock type of the map sheet Järvsjö NO (Björk and Kero 2001). It constitutes one large pluton in the northwestern part of the map sheet (Fig. 4). It is mapped by Björk and Kero (2001) as being a variety of the Revsund granite that is particularly coarse-porphyritic. The rock type appears texturally homogenous in the area of sampling but further to the west, a foliation 350°/85° (right hand rule) is present nearby a possible xenolith of the Skellefte-Härnö granite. The magnetic susceptibility is remarkably higher in this pluton compared to the other granitoids investigated here. The depth of weathering may reach >10 cm and outcropping is probably <1% within the map sheet. U-Pb dating (Björk and Kero 2001), some 6.5 km north of the RÖ sampling area, provide an approx. age of 1830-1817 Ma. If this age is correct, the RÖ granite represents an extraordinarily early variety of Revsund, rather pre- to syn-metamorphic than post-metamorphic. However, the results of the U-Pb dating by Björk and Kero (2001) can be debated due to the large margin of error.
The rock type (RÖ) is a reddish coloured, equigranular-porphyritic, medium-grained granitoid (Fig. 6). It contains significant amounts of biotite which is the dominant mafic mineral. Phenocrystic K-feldspar (<6 cm along c-axis) is common in an even-grained granular matrix. The coarse-porphyritic character occurs in varying density within the area of sampling.
Fig. 6 Typical field appearence of the reddish Revsund (RÖ). Picture shows a weathered surface. The area is approximately 35 cm wide.
Skellefte-Härnö granite, SH
The SH granite is mapped as Skellefte-Härnö granite by the SGU (Björk and Kero 2001) and is a minor constituent in the bedrock within the map sheet Järvsjö NO. It forms several smaller plutons in the central-western (Fig. 4) and in the north-central part of the map sheet (not seen in Fig. 4). It is in places crosscut by late dolerite dykes. Xenoliths of metasediment, 0.3->10 m in diameter is common, as is quartz veins in random orientations. The SH-suite is weakly deformed in a semi-ductile manner but no surfaces suitable for structural measurements have been found within the study area. The depth of weathering is ~5 cm and the outcropping is probably <1%.
The rock type (SH) is a leucocratic, fine to medium-grained, two-mica granitoid which is intensively varied in character, granular-pegmatitic-aplitic (Fig. 7) and may display graphic texture. Garnet, phenocrystic K-feldspar (<8 cm) and “roses” of muscovite are common features in pegmatitic veins, but not in the granite matrix. However, smaller phenocrystic K-feldspar and muscovite are present also in the groundmass. In cut surfaces the rock type gives a diffuse impression of the quartz, interpreted as strain induced grain boundary migration.
Fig. 7 Typical appearence of the Skellefte-Härnö granite (SH). Picture shows a weathered surface. The area is approximately 50 cm wide.
Svartliden granite, SV
The SV granite was never mapped by the SGU because of its sub-surface position prior to the discovery of the Svartliden Au deposit (Fig. 4). In the mine it occurs as a < 6m thick, surface-near, horizontal dyke (Fig. 8a) and as minor dyke swarms, 0.3-2 m thick, which are randomly oriented (Fig. 8b). The SV granite is dated, by the means of 207Pb/206Pb in Monazite, at 1800 ± 7.4 Ma (unpublished
report DAB 2010). The granite cuts a sequence of metagreywackes and lithic arkoses, today seen as biotitic schist, quartz-biotite schist and a sulphide-bearing graphitic schist (Björk and Kero 2001). The sedimentary units are intercalated with amphibolite of different grain sizes and ultra-mafic rocks. According to Samskog (2011) the amphibolite unit plots as basalt whereas the ultra-mafic rock plots as picro-basalt, both with tholeiitic affinities.
The SV granite is a fine to medium grained two-mica, light coloured granite with mafic enclaves, interpreted as heterogeneously assimilated metasediments (Fig. 8c), xenoliths of metasediment (Fig. 8d) and possibly early meta-dolerite (Fig. 8e). The rock type varies extensively in character (Fig. 8f) and is heterogeneous both in respect to grain size (pegmatitic-aplitic) and garnet content. Muscovite is another common mineral as well as pyrrhotite as sparse impregnation, and in veins. The grain boundaries of quartz are diffuse in cut surfaces. The rock is weakly deformed in a brittle-ductile manner, best seen when muscovite is present and the horizontal dyke is cut by a brittle fault in the central part of the present open pit (Fig. 8a).
A
B
Fig. 8 Typical appearance of the Svartliden granite (SV). A) Horizontal dyke cut by a brittle fault. The area is approx. 25 m wide. B) Minor dykes and swarms. The area is approx. 10 m wide. C) Patchy biotite, possibly due to heterogeneously melted metasediment. The area is approx. 15 cm wide. D) Metasedimentary xenolith. The area is approx. 3 m wide. E) Mafic enclave of possible early metadolerite. The area is approx. 20 cm wide. Heterogeneous character of the Svartliden granite. The area is approx. 30 cm wide.
D
E
Petrography
Revsund granite (white), RE
In thin section the quartz occurs as subhedral grains displaying varying degree of undulose extinction. The orthoclase occurs as euhedral crystals and represents the largest grains (<10mm). Microperthitic texture is common in orthoclase as well as sericite alteration. Microcline occurs as euhedral grains, ~1.5mm, and displays diagnostic twinning. The plagioclase is subhedral (<6mm) and is mainly oligoclase (15°), but andesine (32°) is also present in minor amounts. The biotite occurs as subhedral grains, strongly pleochroic, with a reddish tint in a few examples. Inclusions of magnetite (Fig. 9a) and zircon (Fig. 9b) in biotite are not uncommon. Minor constituents of the RE granite are hornblende (Fig. 9c), apatite, hematite and zircon. A very weak deformation, not visible to the naked eye, is discernible under the microscope. Modal mineralogy classifies the rock type as monzo-syeno biotite granite, see Table 1.
A
B
C
Fig. 9 Petrographic characteristics of the RE granite. Abbreviations in Fig. 9-12 are Bt: biotite, Mag: magnetite, Zr: zircon, Or: orthoclase, Hbl: hornblende, Mic: microcline, Hem: hematite, Py: pyrite, Qtz: quartz, Mus: muscovite, Chl: chlorite, Plag: plagioclase, Grnt: garnet, Pyrh: pyrrhotite. A) Inclusion of magnetite in biotite, reflected light. B) Inclusion of zircon in biotite surrounded by a pleochroic halo, crossed nicols. C) Hornblende together with microcline, crossed nicols.
Revsund granite (reddish), RÖ
In thin section the quartz is present as subhedral grains with a sharp extinction. Plagioclase is present as euhedral phenocrysts in which sericite alteration, often complete, is a common feature. All plagioclase is oligoclase, based on extinction angles. Orthoclase occurs as euhedral phenocrysts with microperthitic texture. No microcline can be seen in thin section. Biotite occurs as prismatic euhedral crystals showing signs of chloritization and often with inclusions of magnetite and zircon with well-developed crystal morphology. Hematite is another common mineral in minor amounts. It occurs as exsolution lamellae in magnetite (Fig. 10 a), or displays spectacular replacement fronts in magnetite (Fig. 10 b) and also in pyrite (Fig. 10 c). Pyrrhotite and titanite also occur as a minor phases with the titanite occurring as crystals with well-developed morphology. Apatite occurs in very small amounts. Modal mineralogy classifies the rock type as monzo-syeno biotite granite, see Table 1.
C
B
A
Fig. 10 Petrographic characteristics of the RÖ granite. A) Hematite exsolution in magnetite, reflected light. B) Hematite replacement front in magnetite, crossed nicols. C) Hematite replacement front in pyrite, reflected light.
Skellefte-Härnö granite, SH
In thin section the quartz occurs as subhedral, ~0.5 mm, weakly to very strongly undulating grains, often as twins with perpendicularly oriented c-axes. Quartz occurs both in the granular matrix and as syn-tectonic recrystallized smaller grains associated to muscovite (Fig. 11a). The muscovite forms prismatic subhedral crystals deformed along the general foliation of the rock type. Orthoclase is the dominating feldspar and occurs as subhedral grains, ~0.5-1 mm in matrix but reach several millimetres when phenocrystic. Orthoclase is surrounded by secondary quartz + muscovite and often shows a strong microperthitic texture and sericite alteration. Inclusions of quartz and muscovite in orthoclase are common features, as is simple twinning. Microcline occurs as very fine-grained and subhedral larger grains. The larger grains are often broken and altered at the grain boundaries and microperthitic textures are common. Granophyric texture occasionally occurs. Plagioclase occurs as <1 to ~2mm subhedral grains, often strongly undulating and with bent twinning. The laminar twinning of the plagioclase is in many cases very fine and sericite alteration is common. The plagioclase in the SH granite is dominated by oligoclase with minor albite, based on extinction angles. Biotite is present as subhedral, 0.5-1 mm, prismatic crystals with strong pleochroic haloes indicating radioactive inclusions. Biotite is often situated as inclusions within perthitic feldspar or totally surrounded by secondary muscovite, but also occurs as primary unaltered grains in the matrix. The pleochroism of the biotite is yellowish green-brown and a reddish tint is present in the darker coloured crystals. Very little garnet could be observed in thin section compared to in outcrop. However, muscovite, chlorite ± possible iron hydroxide, is occasionally pseudomorphing garnets (Fig. 11b). Modal mineralogy classifies the rock type as monzo-syeno leuco granite, see Table 1.
B
A
Fig. 11 Petrographic characteristics of the SH granite. A) Quartz+muscovite alteration of orthoclase, crossed nicols. B) Muscovite+chlorite+FEOOH? speudomorphing garnet, crossed nicols.
Svartliden granite, SV
In thin section, quartz occurs as subhedral, ~0.5 mm grains, occasionally with a strongly undulating extinction. Quartz occurs both as primary granular and as re-crystallised smaller grains belonging to a similar alteration assemblage of quartz + muscovite as the one observed in the SH granite (Fig. 12a). Orthoclase is the dominant feldspar. It occurs as subhedral grains surrounded by the quartz and muscovite. Plagioclase is subhedral, 0.5-2 mm along c-axis. It shows fine laminar twinning which is often much bent (Fig. 12b). Polycrystalline plagioclase grains are common and are visible when laminar twinning is orientated perpendicular to each other. Plagioclase in the SV granite is oligoclase, based on extinction angles. Biotite occurs as 0.1-1 mm prismatic subhedral grains in small amounts, <3% (Table 1). Pleochroic haloes are common, as is chloritization. Garnet forms euhedral, often broken and altered, 0.5-2 mm, crystals (Fig. 12c). It occurs associated to the secondary quartz + muscovite alteration assemblage (Fig. 12d), possibly adding garnet to the alteration mineral assemblage. Occasionally, garnet shows evidence of intense retrograde alteration by replacement to muscovite + chlorite + possible iron hydroxide in a similar manner as observed in the SH granite (Fig. 12e). The only common opaque mineral is pyrrhotite (Fig. 12f). Modal mineralogy classifies the rock type as monzo-syeno leuco granite (Table 1).
A
B
D
E
F
Fig. 12 Petrographic characteristics of the SV granite. A) Quartz+muscovite alteration assemblage along deformation, crossed nicols. B) Stressed twinning in plagioclase, crossed nicols. C) Garnet, open nicols. D) Garnet in association with the quartz+muscovite alteration assemblage, crossed nicols. E) Retrograde alteration of garnet, crossed nicols. F) Pyrrhotite, reflected light.
Sample Qtz% Mic Ort% Plag% Bt% Mus% Amph% Tit% Grnt% Opaque% Points RE 6 27.2 4.4 20 19.1 16.2 0 0 0 0 0 844 RE 9 31 15.3 27.3 13.7 11.1 0 0.1 0 0 0 677 RÖ 1 28.1 0 20.8 34.4 5.3 0 1.8 0.8 0 0 654 RÖ6 32.6 0 33.7 17.1 8.3 0 1.2 0.6 0 0.1 688 SH 1 35.3 6.7 27.4 21.5 0 7.1 0 0 0 0 1126 SH 10 39.3 6.7 35.3 9.3 0.7 11.7 0 0 0 0 881 SV 1 35.4 0.8 29.7 22.8 0 10.4 0 0 0.5 0 855 SV 7 23.4 4.2 45.6 22.1 2 2.6 0 0 0.1 0 817
Table 1 Modal mineralogy based on point counting of representative thin sections. All samples are monzo-syeno granite (s.s.). Qtz= quartz, Mic= microcline, Ort= orthoclase, Plag= plagioclase, Bt= biotite, Mus= muscovite, Amph= amphibole, Tit= titanite, Grnt= garnet.
Whole rock geochemistry
Classification
Igneous Spectrum (Hughes 1973)
Figure 13, after Hughes (1973), shows that all samples except SV4 (marked with a blue circle) plot within the Igneous Spectrum and are, based on this plot, regarded as suitable to use in chemical discrimination. With this is mind the sample SV-4 will be removed from all following geochemical plots. 0 1 2 3 4 5 6 7 8 9 10 0 10 20 30 40 50 60 70 80 90 100 K2 O +N a2 O Wt .% 100*K2O/(K2O+Na2O) Wt.% RE RÖ SH SV Igneous spectrum
R1-R2 plot (De La Roche et al. 1980)
All samples plot within the granite field (Fig. 14) of the R1-R2 classification diagram for volcanic and plutonic rocks after De La Roche et al. (1980), where R1= 4Si - 11(Na+K) – 2(Fe+Ti) and R2= 6Ca + 2Mg + Al, expressed in millications. Both types of the Revsund granite, especially the RÖ granite, grade towards granodiorite, whereas the SH- and SV granites grade towards alkali granite.
melteigite theralite alkal i gabbr o gabbro gabbro-norite ijolite essexite syeno- gabbromonzo-gabbro gabbro-diorite nepheline syenite syeno-diorite monzo-nite monzo-diorite diorite syenite quartz
syenite alkali granite quartz monzonite tonalite granite granodiorit ultramafic rock -1000 0 1000 2000 3000 0 500 1000 1500 2000 2500 3000
R1= 4Si - 11(Na + K) - 2(Fe + Ti)
R2 = 6C a + 2M g + A l
Fig. 14 R1-R2 plot after De La Roche (1980). All samples plot as granite. Symbols are, red circles: RE granite, blue
Magmatic affinity (Brown 1981)
In Figure 15 the SH- and SV granites plot in the alkali-calcic field of the magmatic series-plot of Brown (1981). Both Revsund granites (RE, RÖ) plot on the border between the calc-alkaline and alkali-calcic fields, even though the RE granite is concentrated to the alkali-calcic field, whereas the RÖ granite plots in both fields. Complicating relations, however, are shown by the High Field Strength Elements, HFSE, when compared to common ranges for tholeiitic and calc-alkaline suites. Table 2, modified after Barret and Maclean (1999), shows that the Revsund granites are in agreement with calc-alkaline-transitional affinities whereas the Skellefte-Härnö and Svartliden granites rather show ranges typical for tholeiitic-transitional trends with zirconium and thorium contents typical for calc-alkaline rocks.
Tholeiite Transitional Calc-alkaline Average RE Average RÖ Average SH Average SV Zr ppm in rhyolite >600-350 350-600 <200 269.4 140.4 19.9 38.6 Zr/Y 2.0-4-5 4.5-7 <7 8 6.4 1.4 1.9 La/Yb <3 3.0-6-0 >6 21.4 22 3 2.5 Th/Yb 0.1-0.25 0.25- 0.65 0.65- >2 8 14.9 3 3.2 -1,5 -1 -0,5 0 0,5 1 40 50 60 70 80 90 Lo g( Ca O /( N a2 O +K2 O ) SiO2 % RE RÖ SH SV Calc-alkaline Alkali-calcic
Fig. 15 Indication of magmatic series (Brown 1981).
B-A plot (Villaseca et al. 1988)
In the B-A plot, first used by Debon and Le Fort (1982), and later modified by Villaseca et al. (1998), indicating aluminium saturation where A= Al–(K+Na+2Ca) and B= Fe+Mg+Ti, a geochemical grouping can be made among the samples included in this study (Fig. 16). The B-A plot shows that both RE- and RÖ granites plot within the weak-peraluminous field and that the RÖ-suite grades towards metaluminous. The SH- and SV granites plot exclusively within the felsic peraluminous field.
A-type granitoid plot (Whalen et al. 1987)
The classification plots modifed after Whalen et al. (1987), designed for A-type granites, show extensive overlapping in the present data set (Fig. 17). RE, SH and SV samples show A-type character in all classification plots whereas the RÖ-samples plot as A- or I-S-type.
Fig.16 B-A plot after Villaseca et al. (1998). SH and SV granites plot as felsic aluminous whereas RE and RÖ plot as weakly peraluminous grading towards metaluminous.
Fig. 17 Classification plots (Ga/Al) after Whalen et al. (1987). RE plot as I & S-type granite whereas all other samples plot as A-type granite. Symbols are, I & S: I-type and S-type granites, A: A-type granite, blue crosses: RE granite, red circles: RÖ granite, green triangles: SH granite, turquoise crosses: SV granite
Nominal mineralogy and aluminium saturation
In Table 3 the Revsund granites (RE, RÖ) show average Mol Al2O3/ (CaO+K2O+Na2O) or A/CNK ratios
of ~1.12 and ~1.16 respectively. The Skellefte-Härnö and Svartliden granites (SH, SV) show higher average ratios of ~1.40 and ~1.66 respectively. According to Sylvester (1998) all samples are peraluminous (exceeding 1) and indicate an S-type affinity (exceeding 1.1) following Chappell (1999); Chappell and White (2001).
In the CIPW norm (Table 4), calculated in accordance with the software Geochemical Toolkit (Janošek et al. 2006), the RE- and RÖ granites show ~27-32% quartz with one sample yielding ~36% (RÖ-6). The SH- and SV granites show higher quartz contents, ranging between ~35-40%. The feldspars are dominantly orthoclase for the RE- and RÖ granites and plagioclase for the SH- and SV granites. All samples are granitic (s.s) in composition. The SH- and SV granites yield >1% normative corundum, ranging between 1.3 and 4.7%. On the other hand the RE- and RÖ granites yield <1% normative corundum, except for two samples, RE-9 and RE-10 yielding 1.08 and 1.04% respectively.
Sample series Average Mol Al2O3/(CaO+K2O+Na2O)
RE ~1.12
RÖ ~1.16
SH ~1.40
SV ~1.66
Table 3 Average Mol Al2O3/(CaO+K2O+Na2O) or A/CNK ratios indicating aluminium saturation. All samples are
Qtz Cor Or Ab An Hy Mt Il Hm Ru Ap Sum RE5 30.1 0.99 31.97 25.3 5.4 0.95 0 0.09 3.91 0.4 0.38 99.5 RE6 30.47 0.92 31.5 25.64 5.8 1 0 0.11 3.98 0.4 0.38 100.19 RE7 30.95 0.88 31.97 25.3 5.57 0.92 0 0.09 3.79 0.4 0.35 100.22 RE9 30.61 1.08 30.2 25.89 5.47 0.95 0 0.09 3.78 0.38 0.35 98.8 RE10 29.6 1.04 32.92 24.54 5.47 0.92 0 0.09 3.81 0.4 0.35 99.14 RÖ1 27.91 0.13 25 31.39 8.71 1.77 0 0.13 3.13 0.32 0.26 98.75 RÖ2 32.57 0.46 26 28.43 7.08 1.32 0 0.11 2.47 0.24 0.17 98.85 RÖ3 29.73 0.73 26.72 30.21 6.57 1.44 0 0.11 2.52 0.25 0.19 98.46 RÖ4 27.34 0.23 28.6 28.77 7.48 1.54 0 0.13 3.12 0.28 0.24 97.74 RÖ6 36.37 0.46 23.58 27.5 7.15 1.12 0 0.09 2.3 0.23 0.14 98.94 SH1 36.55 2.14 25 32.07 2.57 0.35 0 0.04 1.26 0.05 0.24 100.27 SH2 37.53 3.12 25.23 29.11 1.65 0.17 0.05 0.08 0.75 0 0.28 98 SH3 38.99 2.66 17.85 34.44 3.08 0.25 0 0.04 1.09 0.05 0.21 98.65 SH4 38.39 3.21 24.23 28.85 2.44 0.22 0 0.06 0.98 0.04 0.21 98.64 SH5 35.32 2.08 27.9 28.77 2.36 0.37 0 0.06 1.4 0.06 0.26 98.58 SH6 37.5 2.74 24.7 30.29 2.31 0.3 0 0.04 1.23 0.05 0.26 99.43 SH7 33.95 1.9 29.61 30.46 1.65 0.2 0 0.04 0.89 0.02 0.28 99.01 SH8 42.89 4.68 24.82 21.83 1.03 0.3 0 0.04 1.25 0.06 0.17 97.07 SH9 34.79 2.37 16.84 37.91 2.92 0.3 0 0.06 1.35 0.05 0.31 96.91 SH10 35.29 2.19 28.07 28.52 1.71 0.15 0 0.04 0.75 0.02 0.26 97 SV-1 36.74 3 24.29 32.41 1.49 0.17 0.3 0.06 0.9 0 0.14 99.51 SV-2 37.92 2.95 11 44 2.27 0.22 0.34 0.08 1.14 0 0.17 100.09 SV-3 38.31 2.01 20.8 34.78 2.65 0.27 0.15 0.1 1.28 0 0.12 100.46 SV-5 40.84 3.17 10.58 40.53 2.64 0.35 0 0.06 1.22 0.02 0.21 99.62 SV-6 46.24 2.22 13.18 32.58 2.27 0.42 0.14 0.04 1.3 0 0.17 98.55 SV-7 36.37 1.5 21.27 35.03 2.83 0.25 0 0.06 1.25 0.03 0.14 98.74 SV-8 35.03 1.9 24.64 34.86 1.26 0.17 0 0.06 0.88 0.01 0.19 99 SV-9 40.83 2.63 18.2 34.52 1.69 0.2 0 0.02 0.96 0.04 0.14 99.24 SV-10 34.69 1.32 23.76 36.13 2.47 0.2 0 0.04 0.87 0.01 0.17 99.65
Table 4 CIPW normative calculations according to Janošek et al. (2006). Qtz= quartz, Cor= corundum, Or= orthoclase, Ab= albite, An= anhortite, Hy= hyperstene, Mt= magnetite, Il= ilmenite, Hm= hematite, Ru= rutile, Ap= apatite.
Delineation plots
Figure 18 shows a trace element plot that distinguishes the two groups based upon the average absolute concentrations. The RE- and RÖ granites show generally higher element abundances compared to the SH- and SV granites. Noteworthy is the striking similarity in the way the elements are varying in relation to each other between all groups.
Figure 19 a shows that CaO vs. Al2O3 can be used to discriminate between the RE/RÖ granites and
SH/SV granites. This study also shows that CaO vs. Na2O (Fig. 19 b) can be used for the same
purpose, however, only on the basis on CaO. Several other geochemical trace element plots can be used to delineate between the granites in this study. Examples are: CaO vs. Sr (19 c), CaO vs. Fe2O3
(19 d), Ti vs. Zr (19 e), Hf vs. Zr (19 f), Th vs. Zr (19 g), Nd vs. Sm (19 h). Figure 19 a-h also indicate that the SH and SV granites are more homogenous in geochemical character than are the RE and RÖ granites.
Fig. 18 Spider diagram showing trace element geochemistry of the studied samples. Note the similarity in how the elements vary in relation to each other.
0 5 10 15 20 0 1 2
Al2
O3
%
CaO %
RE RÖ SH SV 01 2 3 4 5 6 0 1 2 N a2 O % CAO % RE RÖ SH SV 0 50 100 150 200 250 300 350 0 1 2 Sr p pm CaO% RE RÖ SH SV 0 1 2 3 4 5 0 1 2 Fe2 O3 % CaO % RE RÖ SH SV 0 0,1 0,2 0,3 0,4 0,5 0 200 400 Ti O2 % Zr ppm RE RÖ SH SV 0 2 4 6 8 10 0 200 400 Hf p pm Zr ppm RE RÖ SH SV 0 10 20 30 40 50 0 200 400 Th p pm Zr ppm RE RÖ SH SV 200 40 60 80 100 120 0 10 20 N d p pm Sm ppm RE RÖ SH SVFig. 19 Examples of plots that allow for a separation between the RE/RÖ granites and SH/SV granites.
A
B
C
D
E
F
Geotectonic discrimination
Geotectonic indication (Pearce et al. 1984)
Figure 20 shows the geotectonic discrimination plots of Pearce et al. (1984), originally based on some 600 granites from different geotectonic settings. The samples included in this study plot on the borders between the syn-collisional granite (syn-COLG), volcanic arc granite (VAG) and within plate granite (WPG) fields. The RE- and RÖ granites plot dominantly in the VAG-field whereas the SH granite is concentrated to the syn-COLG-field. The SV granite scatters more irregularly.
Fig. 20 Geotectonic indication after Pearce et al. (1984). Symbols are, red circles: RE granite, blue crosses: RÖ granite, green triangles: SH granite, turquoise crosses: SV granite, syn-COLG: syn-collisional granite, WPG: within plate granite, VAG: volcanic arc granite, ORG: orogenic granite. All samples plot on the border between the syn-COLG and VAG fields with a few samples of SH and SV granites grading into the WPG field.
Hf-Rb-Ta-plot (Harris et al. 1986)
Figure 21 shows a trivariate discrimination plot developed for granites in collisional tectonic settings (Harris et al. 1986). In the Hf-Rb-Ta plot the RE- and RÖ granites plot clearly within the volcanic arc granite field (VAG). The SH granite plots exclusively within the syn-collisional granite field (syn-COLG). The Svartliden granite scatters more irregularly, crossing the dividing line between Syn-COLG and Late-Post-COLG field.
Fig. 21 Geotectonic indication after Harris et al. (1986). Symbols are, red circles: RE granite, blue crosses: RÖ granite, green triangles: SH granite, turquois crosses: SV granite, syn-COLG: syn-collisional granite, late-post COLG: late to post collisional granite, VAG: volcanic arc granite, WPG: within plate granite. The RE/RÖ data plots in the VAG field whereas the SH data plots in the syn-COLG field and the SV granite scatters in the syn-COLG and late-post COLG fields.
Normalized multi element plots
In the multi element variation diagram normalized to the average upper continental crust according to McLennan and Taylor (1985) the RE- and RÖ granites (Fig. 22 a, b) show little variation with rather “linear trends”. However, the RE-data scatters slightly more than the RÖ-data with depletion of Sr and a slight enrichment of U, Th, La, Ce, Nd, Sm and Tb. The SH- and SV granites on the other hand show strong scattering of the data, mainly depletions, but with rather positive peaks of U (Fig. 22 c, d).
A
C
D
Fig. 22 Multi element variation plots normalized to McLennan and Taylor (1985). Note the “near linear” pattern of the RE/RÖ granites and the strongly irregular pattern of the SH/SV granites.
Rare Earth Elements
Spider plots (Boynton 1984)
Figure 23 shows the chondrite normalized REE spectra of the sample groups, normalized after Boynton (1984). The RE- and RÖ granites show higher total concentrations of REE´s than do the SH- and SV- granites throughout the REE spectrum except at the very end of the HREE´s. The general fractionation trend of the REE spectra (Ce/Yb)N was calculated in accordance with the software
Geochemical toolkit (Janošek et al. 2006). For the RE- and RÖ granites the average normalized (Ce/Yb)N= 10.95 and 10.21 respectively, indicating strong depletion of HREE compared with LREE
(Rollinson 1993). The SH- and SV granites show, according to Rollinson (1993), a more tholeiitic horizontal fractionation trend expressed as (Ce/Yb)N= 1.52 and 1.49 respectively. The low (Ce/Yb)N
values indicate a very slight depletion of HREE compared to LREE. The RE- and RÖ granites show moderately to strongly fractionated LREE patterns as is indicated, according to Rollinson (1993), by the average (La/Sm)N= 3.89 and 6.32 respectively. The SH- and SV granites, on the other hand, show
moderate LREE fractionations with average (La/Sm)N= 2.85 and 2.22, respectively. The RE- and RÖ
granites show moderate to weak HREE fractionation (Rollinson 1993) expressed by the average (Gd/Yb)N=2.57 and 1.55 respectively. The SH- and SV granites, on the other hand, show slightly
positive HREE patterns with average (Gd/Yb)N= 0.49 and 0.51, respectively. All samples except SV-8
show moderately negative Eu anomalies as indicated by the average (Eu/Eu*)N ratio (Rollinson 1993).
The value of the negative europium anomaly is for the RE, RÖ, SH and SV granites 0.27, 0.43, 0.24 and 0.13 respectively.
Fig. 23 REE patterns normalized after Boynton (1984). Note the low REE-fractionation indicated by SH/SV the granites (c, d).
CHARAC field (Bau 1996)
In Table 5 and Figure 24, defining the Charge-and-Radius-Controlled field (CHARAC field), the RE granite shows typical Y/Ho and Zr/Hf ratios for magmatic rocks (Bau 1996). The RÖ granite shows a typical Y/Ho ratio for magmatic rocks but not in the case of Zr/Hf. The SH- and SV granites, on the other hand, do not show typical magmatic ratios for either Y/Ho or Zr/Hf. This deviation from the typical magmatic range indicates that alteration affected, at least in part, the Y/Ho and Zr/Hf ratios of the SH- and SV granites and the Zr/Hf ratio only, in the case of the RÖ granite (Bau 1996).
Y/Ho
Zr/Hf
Typical for magmatic rocks 25-30.4 33.7-39.5
Average RE 26.7 35.8 Average RÖ 26.7 35.8 Average SH 32.9 16 Average SV 34.5 15.7 1 10 100 1 10 100 Y/ Ho p pm Zr/Hf ppm RE RÖ SH SV CHARAC field
Table 5 Average Y/Ho and Zr/Hf ratios of the samples compared with typical magmatic rocks in accordance with Bau (1996). Ratios outside the magmatic range indicate hydrothermal control of the ratio.
Fig. 24 Y/Ho and Zr/Hf ratios plotted in relation to the CHARAC field (Bau 1996). Note the anomalous position of the SH/SV granites.
Stable isotope composition
Table 6 and Figure 25 show that the δ18O
Quartz is high for all samples. Average δ18OQuartz is 11.9 and
10.8 for the RE- and RÖ granites respectively and 12.5 and 12.0 for the SH- and SV granites respectively. According to many workers (e. g. Taylor and Turi 1976; Taylor 1978; O’Neil et al. 1977; Harris et al. 1997; Chapell and White 2001), all samples included in the current study, except RÖ-1, are S-type granites based on O isotope composition (Fig. 25). The border between I- and S-type granites are often reported at between δ18O= 9.5‰ (Harris et al. 1997) and δ18O= 10.0‰ (e. g. O´Neil
et al. 1977; O´Neil and Chappell 1977; Taylor 1978).
Table 6 and Figure 25 show that the RE- and RÖ granites have average δDBiotite values of -102.5 and
93.3 respectively. The average δDBiotite values for the SH- and SV granites are -77.5 and -67.5
respectively. Based on the average δD values for I- and S-type granites, -65 to -99 and -58 to -66 respectively, obtained by O´Neil and Chappell (1977), the RE and RÖ granites have I-type affinities based on δDBiotite whereas the SV granite is S-type. The SH granite plots on the border between the
Average δ
18O
QuartzAverage δD
Biotite Average RE 11.9 -102.5 Average RÖ 10.8 -93.3 Average SH 12.5 -77.2 Average SV 12 -67.5 Table 6 Average δ18OQuartz and δDBiotite for the samples in the study.
Fig. 25 δ18O
Quartz vs. δDBiotite indicating granite type. δ18Oand δD values according to O’Neil and Chapelllll (1977);
O’Neil et al. (1977); Taylor (1978). The border between the sample groups separates the RE/RÖ granites from the SH/SV granites.
Discussion
Petrography
Chappell and White (2001) states that secondary muscovite is the result of “sub-solidus alteration of cordierite” in all the S-type granites of the Lachlan Fold Belt, South-eastern Australia. This particular study could not verify fossil corderite in any of the samples. The mineral assemblage quartz + muscovite in the SH- and SV granites (Fig. 11a, 12a, d) is instead suggested being a product from the breakdown of potassium feldspar (Eq.1) and primary precipitation from peraluminous fluids. The quartz + muscovite mineral assemblage has not fully migrated into pressure shadows, but instead surrounds whole grains of potassium feldspar. This possibly indicates a late tectonic origin of the (alteration) assemblage since the secondary formed minerals would be expected to migrate into pressure shadows with pressure and time. This observation is in agreement with earlier interpretations of the timing of the Skellefte-Härnö granite (Claesson and Lundqvist 1995) relative regional deformation. The presence of garnet and muscovite (Fig. 11a, 12a, d) in the SH and SV granites, as well as the presence of pyrrhotite (Fig. 12f) in the SV granite, is according to Chappell and White (2001) indications of an S-type affinity of these rocks. The garnet and muscovite may be explained by aluminium saturation and pyrrhotite by the reduced conditions typical for S-type granites (Chappell and White 2001).
Eq. 1 3KAlSi3O8 + 2H+ KAl3Si3O10(OH)2 + 6SiO2 + 2K+
K-feldspar + 2H+ muscovite+ 6quartz + 2K+
The red colour of the reddish RÖ granite can be explained by hematite alteration of magnetite and pyrite (Fig. 10a, b, c, d). The hematite alteration is likely widespread since, to the knowledge of the author, the whole pluton is of the same colour. The presence of hornblende and magnetite together with biotite in the RE- and RÖ granites indicate an I-type affinity of the Revsund granite (Chappell and White 2001).
Whole rock geochemistry
Classification
Igneous spectrum (Hughes 1973)
The Igneous Spectrum (Fig. 13) is based on K2O and Na2O and screens the data set under study with
respect to metamorphic alteration. In his article from 1973 Hughes could show that remarkably fresh rocks had been targeted by alkali-metasomatism and could because of this be misinterpreted on the basis of chemical analysis alone. Hughes (1973) stressed the necessity of checking data sets in the light of the Igneous Spectrum prior to the use in geochronology or analytically based petrographic and petrogenetic conclusions. Samples that plot outside the Igneous Spectrum are to be considered metasomatic. Important in this context, however, is that both K2O and Na2O are mobile elements
and that the Igneous Spectrum do allow for both potassic and sodium alteration to occur to some extent without the exclusion from the "unaltered" field in the plot. With this in mind, some potassic and/or sodium alteration cannot be excluded in any of the samples but is definitive in SV-4. Hence, this sample is excluded from all further geochemical plots in this study.
R1-R2 classification diagram (De La Roche et al. 1980)
The result in the R1-R2 classification diagram (Fig. 14) after De La Roche et al. (1980) classifies all
samples as granite (s.s.) and is consistent with modal mineralogy. The result is also in agreement with earlier studies (i. e. Ahl et al. 2001). However, Ahl et al. (2001) used the R1-R2 parameters in a
different context, highlighting tectonomagmatic origin, following Batchelor and Bowden (1985). The strength of the R1-R2 classification diagram is that it takes into consideration all the major cat ions
and that the classification is kept sufficiently general to classify a broad spectrum of igneous rocks. The main drawbacks of the R1-R2 plot is that the parameters R1-R2 lack any obvious meaning
because of the numerous variables included in each parameter which makes it hard to interpret at first glance (Rollinson 1993).
Magmatic affinity (Brown 1981)
The result in Figure 15 classifies the SH-, SV- and RE granites as alkali-calcic whereas the RÖ granite plots as transitional between alkali-calcic and calc-alkaline. This is in agreement with earlier studies focusing on the Revsund and Skellefte-Härnö granites (i. e. Claesson and Lundqvist 1995). However, complicating relationships are present. Table 2 shows common values of the High Field Strength elements, HFSE, in the tholeiitic, transitional and calc-alkaline magmatic series after Barret and MacLean (1999). They suggest that the different typical concentrations of HFSE between the
magmatic series may be a function of what type of material that underwent partial melting. Depth, pressure of fractionation and crustal contamination in the magma are also mentioned as factors contributing to the differences. The RE- and RÖ granites are in agreement with calc-alkaline to transitional affinities based on the typical concentrations of HFSE reported by Barret and MacLean (1999) and it is therefore probable that the concentrations are controlled by the factors mentioned by the same workers. These factors did probably also affect the HFSE-values in the SH- and SV granites, at least to some extent. However, alteration (hydrothermal?) is suggested to be an additional control on the measured HFSE values in the SH- and SV granites. This suggestion is based on the result in Table 5 and Figure 24 where mobility of the normally immobile elements Zr, Y, Ho and Hf is indicated according to Bau (1996). Important to note is that the RÖ granite is not typically magmatic in the context of Zr and Hf in Table 5. Hence, it is suggested that alteration (hydrothermal?) might have had some impact also in this case. Hydrothermal alteration of the RÖ granite is also indicated by the widespread hematite alteration seen in the rock type.
B-A classification plot (Villaseca et al. 1980)
In Figure 16 the RE- and RÖ granites plot as weak-aluminous grading towards metaluminous whereas the SH- and SV granites plot as felsic peraluminous. The results obtained for the RE- and RÖ samples in the B-A plot (Fig. 16) are consistent with many I-types and differentiated calc-alkaline suites around the world (Villaseca et al. 1998). Villaseca et al. (1998) states that the highly felsic peraluminous field, in which the SH- and SV granites plot, mainly indicate S-type granites. The interpretation of the Revsund granite as I-type and Skellefte-Härnö as S-type is in good agreement with earlier studies (i. e. Claesson and Lundqvist 1995; Ahl et al. 2001). However, the present study shows less scattered data with a clear distinction between the rocktypes, which is not seen in earlier studies. It is important to note, though, that this might be the result of the smaller study area and fewer samples in this study, as compared to the earlier more regional scale studies.
The strength of the B-A classification diagram of Villaseca et al. (1998), used in this study, is that four fields within the peraluminous domain is identified through the study of several orogenic domains (Fig. 16). This use of the B-A classification diagram contrasts the original use by Debon and Le Fort (1982) where three fields were specified and characteristic mineral assemblages pointed out. However, Villaseca et al. (1998) do not tend to further divide the metaluminous domain which is the case in the original use by Debon and Le Fort (1982). This means that rocks undersaturated in alumina is best studied by the means of the original B-A classification diagram of Debon and Le Fort (1982).
A-type granitoid plot (Whalen et al. 1987)
The characteristics of the Revsund granitoids have experienced diverse interpretations over the past 30 years. Patchett et al. (1987) pointed out the S-type affinity, whereas Claesson and Lundqvist (1995) and Anderssson (1997) argued for an I-type affinity. Bergman Weihed (2001) states in a review paper that the Revsund granitoid should be classified as I- to A-type. Wilson (1980) and Armands and Xefteris (1987) highlight the difficulties involved in the characterization of the Revsund granitoid and Wilson et al. (1985) states…”the name has been applied over too large an area” which is confirmed by Andersson (1997) who mentions that different Revsund granites are present, both geochemically and texturally, at different locations. Stable isotope data obtained by Wilson et al. (1985) indicate an S-type character, however, the data plots on the border between S- and I-type. The RE-, SH- and SV granites all plot in the A-type field in Figure 17 whereas the RÖ granite show I-, S- and A-type affinities (Whalen et al. 1987). The A-type affinity can at first glance appear contradictionary in the light of the compressional-transpressional stress regime that the area of study was subjected to at the time of Revsund emplacement (Weihed et al. 2005). However, A-type granites do not neccessarily demand an anorogenic, or rift, setting but may represent the final plutonic event in orogenic belts (Whalen et al. 1987). The plots of the SH- and SV granites in Figure 17 are complicating. They plot mainly in the A-type field based on the Ga/Al*1000 parameter. However, in the light of field appearance, nominal and modal mineralogy, their strongly peraluminous character and isotopic signatures, this study indicates an S-type affinity of these rocks. These results highlight the difficulty of constructing this kind of generalized classification diagrams and underline the importance of cautious use of such diagrams. An S-type affinity in the case of the Skellefte-Härnö granite is in good agreement with other studies made on the rock type (i. e. Claesson and Lundqvist 1995; Lindh 2005).
Nominal mineralogy and aluminium saturation
Modal and nominal mineralogy, Table 1 and Table 4 respectively, do not reflect each other regarding the amount of quartz, orthoclase, and plagioclase. However, the differences do not deviate from the composition of granite (s.s) and are therefore not considered as critical. The content of normative corundum, indicating aluminium saturation, is of a more profound meaning for this particular study. Chappell and White (2001) states in a review paper of contrasting types of granites that 1% normative corundum is one of the dividing lines between classical S- and I-type granites. S-types show >1% whereas I-type show <1%. Normative corundum indicates an S-type affinity for the SH- and SV granites whereas the RE- and RÖ- granites are I-types. However, the whitish RE granite,
except for two samples, plots approximately at ~1% normative corundum and do not indicate a clear I-type affinity.
The average Mol Al2O3/(CaO+K2O+Na2O) or A/CNK ratio (Table 3) indicates that all samples in this
study are peraluminous (Sylvester 1998) and have S-type affinities following Chappell and White (2001). The higher average ratios at 1.40 and 1.66 of the SH- and SV granites, respectively, reflect the occurrence of the aluminous minerals muscovite and garnet. The higher ratio of the SV granite (1.66) compared to the SH granite (1.40) possibly reflects the higher abundance of garnet in the SV granite compared to the SH granite. Claesson and Lundqvist (1995) presented a mean ratio for the Skellefte-Härnö granite at 1.2 which is considerably lower than in the current study. However, the values are still within the S-type range (>1.1). The average A/CNK ratios for the Revsund granites, RE (1.12) and RÖ (1.16), indicate weak S-type affinities for these rocks. Claesson and Lundqvist (1995) presented an arithmetic mean ratio for the Revsund granitoid at 1.0 and Andersson (1997) describes the aluminium saturation as varied but centered at A/CNK= 0.99. A/CNK= ~1.0 indicates an I-type affinity according to Chappell and White (2001). Further north, not far from the current area of study, Wilson and Åkerblom (1980) describes some peraluminous Revsund granites with S-type chemistry. Wilson (1980) and Armands and Xefteris (1987) argued for a mixed S-I-type character for Revsund in general which reflects the difficulty in classifying this granitic suite. The A/CNK results in this particular study support the observations of Wilson and Åkerblom (1980). The tendency of the RE- and RÖ granites to show both I- and S-type affinity based on aluminium saturation is reflected in Figure 16 (Villaseca et al. 1998). In this B-A plot the RE- and RÖ granites plot within the weakly peraluminous field, grading towards metaluminous. This can be seen on the background that I-type granites tend to be metaluminous whereas S-type granites normally are peraluminous (Chappell and White 2001).
Delineation plots
The rock samples included in this study display geochemical differences that facilitate separation between the different granites using whole rock geochemistry. As were pointed out by Ahl et. al. (2001) this particular study confirms that CaO is the most prominent parameter among the major oxides separating the Revsund from the Skellefte-Härnö granite (Fig. 19 a, b). In the current study the RE- and RÖ granites are high (1.3-1.9 wt %) whereas the SH- and SV granites are low (0.3-0.84%). Compared to the average chemical composition of granite (Blatt et al. 2006 and references therein) the SH- and SV granites are depleted in CaO whereas the RE- and RÖ granites are slightly low to normal in CaO concentration. Chappell and White (2001) state that low concentration of CaO and to a lesser extent also Na2O characterizes S-type granites, whereas higher concentrations characterize
