Stable Isotope Systematics of Skarn-hosted REE-silicate - Magnetite Mineralisations in Central Bergslagen, Sweden

(1)

Examensarbete vid Institutionen för geovetenskaper ISSN 1650-6553 Nr 279

Stable Isotope Systematics of Skarn-hosted REE-silicate - Magnetite Mineralisations in Central Bergslagen, Sweden

Fredrik Sahlström

The metasupracrustal-hosted, often polymetallic REE-Fe-deposits of Bastnäs- type are found along the “REE-line” in the Palaeoproterozoic Bergslagen ore province, south central Sweden. They essentially comprise REE silicate- bearing magnetite skarn mineralisations with variable contents of other metals.

Even though these deposits have been important for mining and research for centuries, their origin still remains unclear. In this study, samples from 10 different deposits along the REE-line have been charactarised as to mineralogy, petrography and bulk geochemistry, in addition to their isotope systematics. Mineral separates of magnetite and, when present, co-existing quartz or carbonates have been analysed for their oxygen and (for carbonates) carbon isotope compositions, in order to put constraints on the sources for metals and fluids in these deposits. Magnetites have δ¹⁸O-values of -1.79 to 1.12 ‰, while quartzes lie between 7.19 and 8.28 ‰. Carbonates have δ¹⁸O- values between 5.77 and 7.15 ‰ and δ¹³C-values between -5.35 and -3.32 ‰.

Thermometric calculations based on mineral pairs (magnetite-quartz, magnetite-calcite/dolomite), combined with available fluid inclusion data, indicate formation of primary magnetite assemblages between c. 650 to 400

°C. At these temperatures, magnetites from some of the deposits would have been in equilibrium with a magmatic fluid (δ¹⁸O = 6-8 ‰), while magnetites from other deposits would have been in equilibrium with fluids of lower δ¹⁸O (4-6 ‰). Oxygen and carbon isotope trends in carbonates can be explained by interaction between original host carbonates and a fluid of magmatic composition. The combined results indicate that the Bastnäs-type magnetite- REE mineralisations were deposited from an originally magmatic fluid at relatively high temperatures. At local scale, variable modification of the fluid isotopic composition can be explained by mixing with seawater-dominated fluids.

Uppsala University, Department of Earth Sciences Master Thesis, 30 hp Solid Earth Geology

ISSN 1650-6553 Nr 279

Printed by Geotryckeriet, Uppsala University, Uppsala, 2014.

(2)

Supervisors: Erik Jonsson & Karin Högdahl Examensarbete vid Institutionen för geovetenskaper

ISSN 1650-6553 Nr 279

Stable Isotope Systematics of Skarn-hosted REE-silicate - Magnetite Mineralisations in Central Bergslagen, Sweden

Fredrik Sahlström

(3)

(4)

1

Stable isotope systematics of skarn-hosted REE-silicate - magnetite mineralisations in central Bergslagen, Sweden

They essentially comprise REE silicate-bearing magnetite skarn mineralisations with variable contents of other metals. Even though these deposits have been important for mining and research for centuries, their origin still remains unclear. In this study, samples from 10 different deposits along the REE-line have been charactarised as to mineralogy, petrography and bulk geochemistry, in addition to their isotope systematics. Mineral separates of magnetite and, when present, co-existing quartz or carbonates have been analysed for their oxygen and (for carbonates) carbon isotope compositions, in order to put constraints on the sources for metals and fluids in these deposits. Magnetites have δ¹⁸O-values of -1.79 to 1.12 ‰, while quartzes lie between 7.19 and 8.28 ‰. Carbonates have δ¹⁸O-values between 5.77 and 7.15 ‰ and δ¹³C- values between -5.35 and -3.32 ‰. Thermometric calculations based on mineral pairs (magnetite-quartz, magnetite-calcite/dolomite), combined with available fluid inclusion data, indicate formation of primary magnetite assemblages between c. 650 to 400 °C. At these temperatures, magnetites from some of the deposits would have been in equilibrium with a magmatic fluid (δ¹⁸O = 6-8 ‰), while magnetites from other deposits would have been in equilibrium with fluids of lower δ¹⁸O (4-6 ‰). Oxygen and carbon isotope trends in carbonates can be explained by interaction between original host carbonates and a fluid of magmatic composition. The combined results indicate that the Bastnäs-type magnetite-REE mineralisations were deposited from an originally magmatic fluid at relatively high temperatures. At local scale, variable modification of the fluid isotopic composition can be explained by mixing with seawater-dominated fluids.

(7)

4 Sammanfattning på svenska

Polymetalliska REE-Fe-fyndigheter av Bastnästyp förekommer längs den så kallade ”REE- linjen” i en metasuprakrustal enhet i den paleoproterozoiska malmprovinsen Bergslagen i Mellansverige. De består av REE-förande silikater i magnetitskarnlager i marmor. Trots att dessa fyndigheter har varit ekonomiskt betydande samt attraherat forskare i århundraden, kvarstår många frågetecken beträffande deras bildningssätt. I denna studie har prov ifrån 10 olika fyndigheter längs REE-linjen karakteriserats med avseende på mineralogi, petrografi samt geokemi. Mineralseparat av magnetit tillsammans med kvarts eller karbonat har analyserats för syre- och (för karbonater) kolisotopsammansättning, för att få ytterligare information om bildningssättet för denna malmtyp. Magnetiterna uppvisar δ¹⁸O-värden mellan -1,79 och 1,12

‰, medan kvarts ligger mellan 7,19 och 8,28 ‰. Karbonater har δ¹⁸O-värden mellan 5,77 och 7,15 ‰ samt δ¹³C-värden mellan -5,35 och -3,32 ‰. Termometriska beräkningar baserade på mineralpar (magnetit-kvarts, magnetit-kalcit/dolomit), kombinerat med tillgänglig vätskeinneslutningsdata, tyder på att primära magnetitassociationer bildats vid mellan ca. 650 och 400 °C. Vid dessa temperaturer är magnetit från några av fyndigheterna i jämvikt med magmatiska fluider (δ¹⁸O = 6-8 ‰), medan magnetiter från andra fyndigheter är i jämvikt med fluider med lägre δ¹⁸O (4-6 ‰). Trender för kol- och syreisotoper i karbonater kan förklaras genom reaktioner mellan ursprungliga värdkarbonater och en fluid med en magmatisk isotopsammansättning. Resultaten tyder på att malmer av Bastnäs-typ bildades från en primärt magmatisk fluid vid relativt höga temperaturer. Lokalt har fluidens isotopsammansättning blivit påverkad i varierande omfattning genom inblandning av havsvatten.

(8)

5 1. Introduction

1.1. Rare earth elements

In recent years, the rare earth elements (REE) have quickly become one of the more increasingly sought-after natural resources in the world. The REEs comprise a group of elements in the periodic table known as the lanthanides. Additionally, yttrium and sometimes scandium are added to the group due to their chemical similarities (Fig 1). These elements are extensively used in a number of different industries, including the booming “high tech” and “green tech”

sectors, leading to an increasing demand for these commodities (e.g. Haxel et al. 2002). Despite their name, these metals are actually not very rare. The least abundant REE, thulium and lutetium, still exhibit average crustal abundances of over 200 times that of gold (Haxel et al. 2002). The main issue regarding the REE is the fact that there are relatively few geological processes that can concentrate them into deposits of economic value. Moreover, the heavy REE (HREE: Gd to Lu, and Y) and Eu are even more scarce due to their more compatible nature, and therefore deposits rich in these elements are especially valuable (Haxel et al. 2002). Another important factor is the current global political climate. Since the closing of the famous Mountain Pass mine in California in 1984, the REE-market has been dominated by China, lately controlling over 95 % of the global REE-supply. Therefore, recent decreases in the Chinese REE export quotas have made the future global supply of REE very uncertain (Moffett & Palmer 2012). This has led to a boost in exploration and research on new REE deposits worldwide, and the former so important Mountain Pass mine is being re-opened (Wiens 2012, Molycorp). In Sweden, this newfound interest in REE and other strategic metals is indeed very noticeable, with active exploration and planned mining of the Norra Kärr Zr- REE deposit (Jonsson 2013, Nebocat 2009) as well as increasing research on other types of potential ore deposits (e.g. this study, Högdahl et al. 2012, Sahlström 2012).

Fig 1. Periodic table with the rare earth elements marked. Figure taken and modified from geology.com.

(9)

6 1.2. Purpose of study and hypothesis

The Bastnäs-type Fe-REE (locally Fe-REE-Cu-(Co-Au-Bi-Mo)) skarn deposits all occur within a belt of mostly significantly altered felsic metavolcanic rocks in the Nora-Riddarhyttan- Norberg area, also called the “REE-line” (Jonsson & Högdahl 2013). This is located in the western central part of the Bergslagen ore province, south central Sweden. These deposits, among others, have been important for local mining of iron for centuries, and for REE sporadically since the late 1800’s. They are also known for being the location of several discoveries of both new minerals and elements (see Andersson et al. 2004, and references therein). The most famous discovery is probably that of the element cerium, which was first described by Wilhelm Hisinger and Jacob Berzelius in 1804. The element was found as a major constituent of a reddish heavy mineral called “Bastnäs tungsten” (later formally named cerite).

After extensive chemical analyses, Hisinger and Berzelius were able to characterise and give a detailed description of the new element. The name “cerium” was inspired by the discovery of the asteroid Ceres three years earlier (Trofast 1996).

Despite their historic significance in both mining and science, little is still known about the genesis of these deposits. Various theories have been proposed, and the consensus today is that these ores are hosted by skarns formed during epigenetic reactions between a hydrothermal mineralising silica and metal-bearing fluid and pre-existing carbonate rocks (Andersson et al.

2004, and references therein). However, the specific nature as well as origin of such a fluid are yet to be determined. This project addresses the fundamental question of whether these ores were formed from high, medium or low temperature fluids, and also try to shed light on the possible source(s) of these fluids. This is primarily tested using oxygen isotope analyses of magnetites together with co-existing quartz or carbonates, separated from mineralogically and petrographically well-characterised REE-rich assemblages. The hypotheses tested are 1) whether these magnetites could have formed in equilibrium with a water-dominated hydrothermal fluid and if so at what temperature, and 2) whether this fluid had a magmatic, meteoric, seawater or mixed origin. Additionally, carbon isotope analyses of associated carbonate assemblages from a selection of deposits were performed in order to better constrain the processes responsible for the formation of these deposits. The mineralogical and petrographical characterisation of the assemblages selected for isotope analyses was conducted with a special focus on deposits lacking modern studies. This study was done within the framework of projects "Securing society's supply of hi-tech elements by recycling mine dumps and probing the metallogenic setting of the host systems", on rare and high-tech metals in

(10)

7

central Swedish mines and dumps, financed by Vetenskapsrådet and the Geological Survey of Sweden (SGU).

2. Geological Background

2.1. Regional geology and mineralisations of the Bergslagen ore province

Bergslagen is one of the largest and most important ore provinces in Sweden, and its rich mining history dates back to medieval times and beyond (cf. Tegengren 1924). With over 8500 documented mines and prospects the Bergslagen province boasts a wide variety of different ore types. These include stratiform and stratabound Zn-Pb-Cu-(Ag-Au) sulphide deposits, iron mineralisations of banded iron formation (BIF)-, skarn- and Kiruna types as well as granite- hosted Mo-deposits (Allen et al. 2008, Allen et al. 1996, Ripa & Kübler 2003). Iron has historically been the most important metal in the region (Sidenvall 1942), with 32 major deposits producing over 420 Mt of ore up until the 1990’s (Åkerman 1994). The nine major sulphide deposits have accounted for over 70 Mt of ore, and additionally a number of smaller scale operations are scattered all over the Bergslagen province, where everything from base metals to precious metals, phosphorous and REE have been targeted (Tegengren 1924, Geijer

& Magnusson 1944, Åkerman 1994, Back 1981, Andersson et al. 2004).

The Bergslagen province is situated in the south-western part of the 1.9-1.8 Ga Svecofennian domain in the Fennoscandian Shield (Fig 2; e.g Stephens et al. 2009). It is a part of a larger, medium to high grade metamorphosed felsic magmatic region (Allen et al. 1996). The host- rocks to most of the mineralisations comprise mainly of a metavolcano-sedimentary succession, which was previously referred to as the “leptite-hälleflinta formation” (Magnusson 1970). The volcanic and subvolcanic rocks are predominantly of rhyolitic to dacitic composition, and are interpreted to have been deposited in a mostly shallow submarine, continental back-arc setting (Allen et al. 1996, Lundström 1987, Ripa 2001, Vivallo & Rickard 1984). Additionally, at various stratigraphic levels in the volcanic succession subordinate rocks include intermediate to mafic volcanic rocks and chemical, epiclastic and biogenic sedimentary rocks (Ripa 2001).

On a larger scale, the metavolcanic succession and associated rocks are interlayered beween two metasedimentary successions of clastic facies, of which the lower contact is exposed on the island of Utö (Allen et al. 1996). The basement underlying the Svecofennian metasupracrustal rocks is unknown. However, locally occurring quartzites containing 2.7-1.95 Ga old detrital zircons as well as inherited zircons in metavolcanic rocks suggest a continental-type basement

(11)

8

which has subsequently been eroded and reworked (Lundqvist 1987, Allen et al. 1996, Andersson et al. 2006). Deformation and metamorphism associated with the c. 1.87-1.80 Ga Svecokarelian orogeny affected the older rocks to variable degrees, and extensive granitic intrusions were emplaced pre-, syn- and post-tectonically (Wilson et al. 1984, Allen et al. 1996, Stephens et al. 2009). Additionally, the western parts of the region have locally been overprinted by the Sveconorwegian orogeny at 1.0 Ga (e.g. Stephens et al. 2009). The Svecokarelian deformation is most obvious in the supracrustal rocks where it appears as steep, tight to isoclinal folding (Allen et al. 1996).

Fig 2. Geological map (SGU Ba-58) over the Bergslagen province. The REE-line is marked with a black rectangle. Figure modified from Stephens et al. (2009).

(12)

9

The Bergslagen province has gone through a magmatic-extensional-compressional cycle, where an initial stage of intense magmatism, thermal doming and crustal extension was followed by a stage of waning extension, thermal subsidence, reversal from extensional to compressional deformation, metamorphism and structural inversion (Allen et al. 1996). This evolution can be observed in the rock succession, where lower deep water sedimentary rocks (e.g. turbidites) are overlain by shallow marine to subaerial volcanic rocks, representing the first stage of the cycle. Continuing upwards in the successions, shallow marine rocks (e.g.

volcanic sand- to siltstones and carbonates) are found, followed by deep water sedimentary rocks, representing the second stage of the cycle (e.g. Stephens et al. 2009 and references therein). Additionally, second-order stratigraphic variations due to subregional differences in uplift and subsidence together with variations in the evolution of individual volcanoes are superimposed on this cycle (Allen et al. 1996). Extensive hydrothermal alteration of the metavolcanic sequence is commonly seen, with metasomatic enrichment in K, Na and Mg having occurred in partly successive pulses and areas (Trägårdh 1991, Stephens et al. 2009).

Most mineralisations are hosted by hydrothermally altered metavolcanic rocks and associated metalimestones and/or skarns. They are, with the exception of the apatite-iron oxide ores, found mainly in the upper parts of the volcanic succession in medial to distal facies, connecting them to the second evolutionary stage with waning volcanism and deposition in a subaqueous environment (Allen et al. 1996). Moreover, a genetic link between the ores and specific magmatic-hydrothermal systems from individual volcanoes or volcanic complexes is possible (Allen et al. 1996, Högdahl & Jonsson 2004, Jonsson 2004).

2.2. Skarn deposits

The term “skarn” originates from Bergslagen, and was first coined by Swedish geologist Alfred Elis Törnebohm in 1875. Skarns can be divided into two main types – reaction skarns and contact skarns. Reaction skarns are small scale (millimeter to meter) features, formed more or less in-situ owing to metasomatic transport of components between adjacent lithologies during high-grade metamorphism (e.g. Burt 1977 and references therein). Contact skarns on the other hand are generally attributed to magmatic-hydrothermal activity related to dioritic to granitic plutonism in orogenic belts, and are distinguished by having a very characteristic gangue assemblage of coarse-grained mixtures of Ca-Mg-Fe-Al silicates (Einaudi & Burt, 1982). They are often classified according to the rock type they replace, with endoskarns being replacements

(13)

10

of intrusive rocks and exoskarns being replacements of carbonate rocks. Contact skarns can be further classified based on their dominant mineralogy (which is dependant on the composition of the rock being replaced) and the dominant economic metal (Fe, W, Cu, Zn-Pb or Sn; Einaudi

& Burt 1982).

2.3. Local geology of the “REE-line”

The “REE-line” is situated in the western central part of the Bergslagen region (Figs 2,3). The deposits along this line are found in a narrow northeast-trending belt of felsic metavolcanic rocks within the Svecofennian succession (Holtstam & Andersson 2007, Jonsson &

Högdahl 2013; Fig 3). In this area a very prominent Mg-alteration has overprinted the country rocks, which were subsequently metamorphosed under amphibolite- facies conditions (Trägårdh 1988, Geijer 1920). The carbonate-bearing metavolcanic belt is surrounded by granitoids of two generations, the older generation (1.9- 1.85 Ga) is an often deformed and metamorphosed suite ranging from gabbro to tonalite-granodiorite- granite, and a younger generation (1.85-1.75 Ga) of mainly undeformed granites and associated migmatites (Andersson et al. 2004, Holtstam & Andersson 2007).

The REE-silicate bearing mineralisations in the area, known as Bastnäs-type deposits, generally occur as seemingly epigenetic, massive to disseminated magnetite-skarn replacements in mostly dolomitic marbles (e.g. Geijer 1920, Geijer 1961, Holtstam

2004), hence there is a direct link between magnetite formation and that of REE mineralisation.

Some deposits are, however, hosted mainly by felsic metavolcanic rocks (e.g. Östra Gyttorp;

Nordenström 1890, Tegengren 1924). The ore mineral assemblages found include various combinations of Fe-oxides, REE-silicates, REE-fluorocarbonates, sulphides of Cu, Co, Bi, Mo and minor native Au, Ag (the latter specifically in the Bastnäs field; e.g. Tegengren 1924, Geijer

Fig 3. Geological map of the ”REE- line”. Modified from Jonsson &

Högdahl (2013) and references therein.

(14)

11

1920, Holtstam 2004). These mineral assemblages typically occur associated with skarn minerals such as actinolite, tremolite, diopside and fluorite as well as often minor carbonates (Geijer 1936). Based on slight local differences in chemistry and mineralogy of the deposits, Holtstam & Andersson (2002) suggested a subdivision of Bastnäs-type deposits into two subtypes. Subtype 1-deposits comprise the Bastnäs and Rödbergsgruvan areas where the mineralisations are mainly enriched in LREE and Fe. In subtype 2-deposits, which are found in the Norberg area (e.g. Östanmossa, Malmkärragruvan, Södra Hackspiksgruvan, Johannagruvan), the ores are enriched in both LREE and HREE+Y together with Mg, Ca and F (Holtstam & Andersson 2007). Mining in Nora-Riddarhyttan-Norberg has been on-going for many centuries. Tens to hundreds of now abandoned mines, occurring scattered around the area, have mainly been targeted for iron, base metals and, in limited extent, REE (Geijer 1936).

2.3.1. Genesis of Bastnäs-type deposits

The first theories regarding the genesis of the Bastnäs-type ores were born almost a century ago by the Swedish geologist Per Geijer. He proposed that ores formed by metasomatic replacement of pre-existing carbonate rocks by high temperature fluids (Geijer 1920), an idea that is widely accepted today (cf. Andersson et al. 2004 and references therein). However, the exact nature of the ore forming fluid is still uncertain as well as its sources. Additionally, the timing of mineralisation has been debated. Geijer (1931) noticed the close spatial relation to the heavily Mg-altered country rocks, as well as the high magnesium contents of minerals such as allanites in some of the deposits. He concluded that widespread Mg-metasomatism occurred during emplacement of the 1.9 Ga synorogenic granitoids, a common theory at that time, and that this process was genetically related to ore formation (e.g. Geijer 1961). This type of metasomatism has later been suggested to be primarily driven by seawater-dominated fluids (Trägårdh 1988, Trägårdh 1991, Ripa 1994). More recent studies point towards a syn-volcanic magmatic- hydrothermal origin of the ores (Holtstam & Andersson 2007, and references therein), thereby discarding Geijer’s model. Fluid inclusion studies in bastnäsite indicate that this mineral was deposited from a CO2-rich, highly to moderately saline fluid at minimum (not pressure corrected) temperatures between 400 and 300 °C (Holtstam & Broman 2002, Andersson et al.

2013). However, bastnäsite mainly formed later than cerite-(Ce) and ferriallanite-(Ce), suggesting that the initial temperatures of the ore forming fluid were probably higher (Andersson et al. 2004). Fluorite were in turn deposited from a low to medium saline fluid, during cooling, from 150 to 100 °C (Uncorrected; Holtstam & Broman 2002, Andersson et al.

(15)

12

2013). Furthermore, carbon and oxygen isotope studies of carbonate minerals associated with REE mineralisation from type 2-deposits show values indicating a magmatic source of the ore forming fluids (Andersson et al. 2013). This is also suggested by the Cu-Mo-Bi-W-Be-F association in some of the deposits, which further indicates linkage to granitic magmatism (Andersson et al. 2004). Re-Os geochronology on molybdenite, which is paragenetically related to REE mineralisation, give ages of 1.90-1.84 Ga, with the youngest ages in the SW (Rödbergsgruvan) and oldest ages in the NE (Norberg area) (Andersson et al. 2013). No explanation has so far been given to explain this extended time interval fully. Textural and structural interpretations of the extensively folded and recrystallised REE assemblages in the Högfors BIF deposit (north of Bastnäs) suggest that the REE mineralisation was formed during the Svecofennian syn-volcanic magmatic stage, thus pre-dating the later extensive polyphase deformation and peak metamorphism of the Svecokarelian orogeny (Jonsson & Högdahl 2013).

3. Materials and methods 3.1. Sample preparation

Samples from 10 different localities in the Nora-Riddarhyttan-Norberg area were petrographically and mineralogically characterised (Table 1). They were also analysed for major and trace element geochemistry as well as for their oxygen and carbon isotope composition. The samples are from Danielsgruvan, Östanmossagruvan, Malmkärragruvan, Södra Hackspiksgruvan, Johannagruvan, Bastnäsfältet, Högforsfältet, Myrbacksfältet, Rödbergsgruvan and Östra Gyttorpsgruvan (Fig 3). They were generously provided by the Geological Survey of Sweden (SGU), the Swedish Museum of Natural History (NRM), Prof.

Erik Jonsson and Doc. Karin Högdahl.

(16)

13

Table 1. Description of samples used in this study including name of deposit, ore type, sample coding and analyses performed on each sample. Abbreviations used are: BGC – bulk geochemical analyses, ISO – stable isotope analyses, XRD – powder X-ray diffraction, OM – optical microscopy, SEM-EDS – scanning electron microscopy with energy dispersive spectrometer, WDS – electron probe microanalyser with wavelength dispersive spectrometer.

Deposit Type Sample ID Thin/thick section label Analyses

Danielsgruvan Banded mgt-qtz ore NRM-20020125 Danielsgr. Ö BGC, ISO, OM

Danielsgruvan Banded mgt-qtz ore NRM-20020124 Danielsgr. BGC, ISO, OM

Östanmossa Carbonate skarn mgt ore EJ-ÖM90-13-1 Östanmossa EJ-ÖM90-13-1 BGC, ISO, XRD, OM, SEM-EDS Östanmossa Carbonate skarn mgt ore SGU-M4528 Östanmossa M4528 BGC, ISO, XRD, OM, SEM-EDS Östanmossa Carbonate skarn mgt ore EJ-ÖM90-13-2 Östanmossa EJ-ÖM90-13-2 BGC, ISO, XRD, OM

Östanmossa Carbonate skarn ore SGU-M4441 Östanmossa M4441 BGC, ISO, XRD, OM

Östanmossa Carbonate skarn ore SGU-M4529 - ISO, XRD

S. Hackspiksgruvan Fluorite skarn mgt ore SGU-M3563 S. Hackspiksgr. M3563 BGC, ISO, OM, SEM-EDS

Johannagruvan Skarn mgt ore Joha-KH-1 Johanna BGC, ISO, OM, SEM-EDS, WDS

Johannagruvan Skarn mgt ore Joha-KH-2 Johannagruvan BGC, ISO, OM

Malmkärra Carbonate skarn mgt ore SGU-M4068 Malmkärragr. M4068 (*2) BGC, ISO, XRD, OM, SEM-EDS Malmkärra Carbonate skarn mgt ore SGU-M4048 Malmkärragr. M4048 BGC, ISO, XRD, OM, SEM-EDS

Högfors BIF Högf-KH-1 Högfors BGC, ISO, OM

Högfors BIF Högf-KH-2 - ISO

Bastnäs Skarn mgt ore Bast-EJ-1 Ceritgruvan BGC, ISO, OM

Bastnäs Massive mgt ore SGU-M6777 Bastnäs M6777 ISO, OM

Bastnäs Skarn mgt ore SGU-M309 N. Bastnäs M309 BGC, ISO, OM, SEM-EDS

Myrbacksfältet Massive mgt/sulphide ore Myrb-EJ-1 Myrbacksfältet BGC, ISO, OM

Myrbacksfältet Massive mgt/sulphide ore Myrb-EJ-2 - ISO

Östra Gyttorp Massive mgt ore Gytt-EJ-1a Unmarked (polished rock piece) ISO, OM, SEM-EDS, WDS Östra Gyttorp Massive mgt ore Gytt-EJ-1b Unmarked (polished rock piece) ISO, OM

Östra Gyttorp Massive mgt ore Gytt-EJ-1c - ISO

Östra Gyttorp Skarn mgt ore Gytt-EJ-2 Gyttorp BGC, OM, SEM-EDS, WDS

Rödbergsgruvan Massive mgt/skarn ore NRM-880071 Rödbergsgr. 880071 BGC, ISO, OM, SEM-EDS

Rödbergsgruvan Massive mgt/skarn ore NRM-19984100 - BGC, ISO

3.1. Sample preparation

3.1.1. Preparation of polished sections and thin sections

Polished (thick) sections were prepared at the Department of Earth Sciences, Uppsala University (UU). First, hand specimens were cut into small pieces using a diamond saw, thereafter the samples were ground in order to get a flat surface before polishing. For this a silicon carbide powder was mixed with water onto a glass plate, on which the samples were ground. This was done in four steps with finer powder used successively (80 µm-45 µm-18 µm- 12 µm). Finally, the samples were polished using a polishing machine in order to get a perfectly flat surface for microscopy studies. On the machine a cloth was placed, onto which fine diamond paste and water were mixed. The samples were polished in three successive steps,

(17)

14

with increasingly finer paste used at each one (6 µm-3 µm-1 µm). Between each step of grinding and polishing, the samples were thoroughly cleaned using an ultrasonic bath in order to minimise the risk of contamination of coarser polishing agents that could cause scratches during the following polishing steps.

Additionally, a number of samples were prepared as thin sections. Samples were cut out into thin pieces using the diamond rock saw. The samples were then sent for thin section preparation at Minoprep, Sweden.

3.1.2. Preparation of samples for stable isotope analysis

The samples for stable isotope analysis were prepared at the Department of Earth Sciences, UU.

Based on the mineral paragenesis of the hand specimen, one or more of the following minerals were separated: magnetite, quartz and carbonates. Small pieces of rock were cut out from the hand specimens using a diamond saw. Care was taken to remove all weathered surfaces. The pieces were then crushed using a hammer. Samples containing large amounts of gangue minerals, such as silicates, needed occasionally to be ground in a small mortar to reduce the grain size enough to get clean crystals for separation. The samples were then cleaned thoroughly using an ultrasonic bath, in order to reduce dust and contaminants, and then dried in an oven at 100 °C for 24 hours. The final step was using a magnet to separate the magnetite from the gangue material, which was done several times until no impurities were visible. For carbonates and quartz, clean crystals or crystal fragments were handpicked using tweezers. The final mineral concentrate was controlled under a stereo microscope, were any remaining contaminant minerals were picked out using a pair of non-magnetic tweezers. For each sample, around 10 mg of crystal grains were separated for final analysis.

(18)

15 3.2. Analytical methods

3.2.1. Powder x-ray diffraction

In order to measure the proportions of calcite and dolomite in carbonates analysed for isotopic composition, some of the mineral separates were sent for analysis with powder x-ray diffraction (XRD) technique. The analyses were conducted using a PANalytical X’pert PRO automated diffractometer at the Swedish Museum of Natural History in Stockholm. The samples were ground to a fine powder to allow the crystals to orientate in all possible orientations, and then placed in a silicon sample holder. The samples were then bombarded by x-ray radiation, which will diffract only in directions with constructive interference, determined by Bragg’s law:

n*λ=2*d*sinθ. Using a goniometer and a detector, the intensity and angle of the diffracted radiation can be measured. By applying Bragg’s law, it is possible to determine the lattice spacings, d, and in turn the Miller indices of lattice planes of the crystals. Furthermore, the intensity of the peaks are related to the position of a specific atom in the unit cell, as atoms with dense electron clouds diffract the beam more efficiently than atoms with less dense electron clouds. The result can then be matched with available x-ray diffractogram databases, and thereby identifying the mineral(s) in the analysed sample (Putnis 1992, Borchardt-Ott 2011).

Analyses were performed using an acceleration voltage of 45 kV and a beam current of 40 mA.

2θ were measured in the interval 5-70° for 11 minutes. Mineral identification and quantification were done using the Highscore Plus software.

3.2.2. Bulk geochemical analysis

Bulk rock samples from all localities were analysed for major and trace elements at the ALS Labs, Canada. Major elements were analysed using X-Ray Fluorescence (XRF) technique, while trace elements and REE were analysed with combined Inductively Coupled Plasma Atomic Emission Spectroscopy (ICP-AES) and Inductively Coupled Plasma Mass Spectrometry (ICP-MS) technique. In the XRF technique, the sample is excited by high energy x-rays. The excited atoms emit secondary fluorescent x-rays as they return to their ground state, with energies characteristic for each element. The concentrations of each element in the sample can thereby be quantified (Norrish & Chappell 1977). The ICP-AES technique is very similar, where a plasma source is used to dissociate the sample into charged atoms or ions, thereby exciting them to a higher energy level. As they return to their ground state, x-rays characteristic

(19)

16

of each element will be emitted, allowing quantification of elemental concentrations in the sample (Philips 2008). In ICP-MS analysis, the sample is ionised with an inductively coupled plasma. The ions are separated and quantified using a mass spectrometer (Beauchemin 2008).

For both ICP-MS and ICP-AES analyses the sample must be dissolved, and for the studied samples, an acid digestion method was used (ALS 2012).

3.2.3. Optical microscopy

Optical microscopy was done at the Department of Earth Sciences, UU, using a Nikon Eclipse E600 POL microscope with an attached camera/computer system, which enables high quality photomicrography. Thin sections were studied using transmitted and reflected light, whereas the polished (thick) sections were studied using reflected light.

3.2.4. SEM-EDS

For reconnaissance investigations on the mineralogy of the samples, SEM-EDS analysis was conducted at the Evolutionary Biology Centre, UU. The Scanning Electron Microscope (SEM) used was a Carl Zeiss Supra 35-VP Field Emission SEM equipped with a Robinson backscatter detector, a variable pressure secondary electron (VPSE) detector and an EDAX Apex-4 energy dispersive spectrometer (EDS). In order to perform SEM-EDS analyses, the samples need to be electrically conductive. This was done by coating the samples with a thin film of carbon by means of sputtering at the Department of Earth Sciences, UU.

The SEM uses a Field Emission Source (FES) to generate a finely focused, high energy electron beam, which is swept across the sample surface in a raster pattern. As the electron beam hits the sample surface, it interacts with the sample to produce a variety of different signals, such as secondary electrons, backscattered electrons, Auger electrons and x-rays. These signals depend on various characteristic properties of the sample, e.g. chemical composition, crystallographic properties and surface topography (Goldstein et al. 1975). In this study backscattered electrons were used for imaging. As the electrons from the beam hits the sample surface, they will be attracted by the positively charged nuclei of the elements in the sample. They will be pulled towards the nuclei and circle it, emitting out of the sample again in the opposite direction where they are collected and registered by a BSE detector. Minerals containing heavy elements will backscatter electrons more effectively, as their nuclei have a higher positive charge and thereby

(20)

17

attracts more electrons. Heavy element-rich minerals will yield brighter images, making it possible to distinguish between different mineral phases as well as compositional variations (e.g. zoning) within a single mineral (Goldstein et al. 1981). The SEM also has a camera function which can produce high resolution BSE or SE photomicrographs of the studied samples. The analyses were performed using an acceleration voltage of 20 kV, a beam current of 40 nA and a beam diameter of 3 µm.

Also present in the SEM is an EDS detector, allowing qualitative/semi-quantitative chemical spot analysis of various mineral phases to be obtained e.g. for identification purposes. When the electron beam strikes the sample, thereby exciting it, each element in the mineral will produce characteristic x-rays of a specific energy. By measuring the energies of the x-rays from the various elements simultaneously, a spectrum displaying the approximate elemental abundances of the analysed mineral can be generated (Goldstein et al. 1975).

3.2.5. Electron probe microanalyser (EPMA)

For quantitative mineral analyses and chemical dating, EPMA analysis was done at the National Electron Microprobe Laboratory, CEMPEG, Department of Earth Sciences, UU, using a Jeol JXA-8530F Field Emission Electron Probe Microanalyser (FE-EPMA). The technique is very similar to the SEM – an electron beam interacts with the sample to produce the signals and generate an image. This instrument is also equipped with an EDS detector for rapid mineral identification and a BSE detector was used for orientation and identification of the various minerals. The EPMA normally has several wavelength dispersive spectrometers (WDS) installed. The WDS spectrometers differs from the EDS in that they measure the wavelengths of the generated x-rays, which - similarly to the energies - are characteristic for each element.

The spectrometer configuration requires each wavelength to be measured independently, and by comparing the intensity to a known standard the exact concentration of each element in the sample can be determined. This makes this technique much more precise and gives higher mass resolution compared to EDS, and can therefore be used to yield good quantitative measurements of the elemental compositions of different minerals (Goldstein et al. 1981). The EPMA also has a camera function similar to the one in the SEM. The analyses were done using an acceleration voltage of 20 kV, a beam current of 40 nA and a beam diameter of 1-10 µm. Counting times (background + peak) were 30 s for Si, Al, Ca, P, Fe; 45 s for As, S, Y; 75 s for all La-Lu; 150 s for U, Th and 225 s for Pb. Standards used were Gallium arsenide (AsGa) for As; wollastonite

(21)

18

(CaSiO3) for Si, Ca; aluminum oxide (Al2O3) for Al; sphalerite (ZnS) for S; thorite (ThO2) for Th; apatite (Ca5(PO4)3(F,Cl,OH)) for P; endmember synthetic phosphates (XPO4) for all REE;

elemental iron (Fe) for Fe and vanadinite (Pb5(VO4)3Cl) for Pb. X-ray lines used were AsLα, SiKα, AlKα, SKα, UMβ, ThMα, CaKα, YLα, PKα, LaLα, CeLα, SmLβ, GdLβ, NdLβ, YbLα, EuLα, PrLβ, FeKα, TbLβ, DyLβ, HoLβ, ErLβ, TmLα, LuLα and PbMβ.

3.2.5.1. Chemical dating of uraninite

Chemical dating can be done using WDS-data from certain U and Th-bearing minerals, such as uraninite and monazite. Provided that all Pb present in the mineral is a product of the radioactive decay of U and Th, precise measurement of these elements allows an age of the mineral to be calculated (Bowles 1990). Programs for calculation of chemical ages (t) are based on the composite decay equation:

Eq 1: 𝐶_𝑃𝑏 = 𝐶_𝑇ℎ[0.897(𝑒^λ²³²^𝑡− 1)] + 𝐶_𝑈[0.859(𝑒^λ²³⁸^𝑡− 1) + 0.006(𝑒^λ²³⁵^𝑡− 1)]

(Hurtado et al. 2007)

Where CPb, CTh and CU are concentrations of Pb, Th and U in ppm. λ 232, λ238 and λ235 are decay constants for ²³²Th (4.9475·10^-11 yr^-1), ²³⁸U (1.55125·10^-10yr^-1) and ²³⁵U (9.8485·10^-10yr^-1) (Steiger & Jäger 1977).The coefficient before the first exponential term is given by the mass ratio of ²⁰⁸Pb to ²³²Th (208/232 = 0.897), while the coefficients before the second and third exponential terms are given by the ratios of the abundance fractions of the respective U isotopes to the mean atomic mass of U (0.9928/238.04 = 0.859 for ²³⁸U and 0.0072/238.04=0.006 for

235U).

In order for the chemical age to accurately reflect the true age of the mineral, the boundary condition, that no Pb, U and Th has been lost or gained after mineral crystallisation, must be fulfilled (Kempe 2003). This can be controlled by calculating the ThO2*-value:

Eq 2: 𝑇ℎ𝑂₂^∗ = 𝑇ℎ𝑂₂+ 𝑈𝑂₂𝑊_𝑇ℎ

𝑊_𝑈[𝑒𝑥𝑝(λ₂₃₂𝑡) − 1]× [exp(λ₂₃₅𝑡) + 138 𝑒𝑥𝑝(λ₂₃₈𝑡)

139 − 1]

(Suzuki et al. 1991)

ThO2, UO2 and PbO are reported in wt %. W is the molecular weight of each oxide (WTh = 264, WU = 270, WPb = 224). ²³⁸U/²³⁵U = 138. Other variables are the same as in the composite decay equation above. If the U-Th-Pb system has remained closed, and the analysed minerals

(22)

19

crystallised at the same time, all measurements should plot along a single isochron line in a PbO-ThO2* diagram (Suzuki et al. 1991, Kempe 2003).

3.2.6. Stable isotope analysis

The stable isotope (O and C) analyses were performed at the Department of Geological Sciences, University of Cape Town, South Africa. For the magnetite oxygen isotope analysis, the samples were prepared by laser fluorination. The samples were placed in a reaction chamber, where they were reacted with BrF5 at 10 kPa pressure. The purified O2 was then collected onto a 5 Å molecular sieve in a glass storage bottle (see Harris & Vogeli 2010, Weis 2013). For carbonates, CO2 was extracted by reacting the samples with 100 % phosphoric acid at 50 °C (see McCrea 1950, Ganino et al. 2013). The material was initially assumed to be calcite, in order to correct for the fractionation between CO2 and carbonate. Samples containing dolomite were later recalculated based on calcite-dolomite fractionation. The oxygen and carbon isotope ratios were measured using a Thermo DeltaXP mass spectrometer, using Monastery garnet as standard for magnetites and Namaqualand Marble for carbonates.

3.2.6.1. Stable isotope geochemistry

The field of isotope geochemistry was developed after the discovery of the neutron in 1932 by H. Urey, and the demonstration of variations in the isotopic composition of light elements by A. Nier during the 1930’s and 1940’s (White 2011). Stable isotope geochemistry is based on variations in isotopic composition owing to physicochemical processes (White 2011, Hoefs 1997), and the measurement of these variations has many different geological applications. In this study, stable isotopes are used as geochemical tracers for identification of the sources for ore forming fluids together with the possibility of performing geothermometric calculations (Rollinson 1993). Here oxygen and carbon isotopes have been used, more specifically the pairs

18O/¹⁶O and ¹³C/¹²C. The relative abundance of the heavier isotope (i.e. ¹⁸O and ¹³C) in a rock, mineral or fluid is usually determined by calculating the δ-value. This value is obtained, using oxygen as example, by the equation:

(23)

20 Eq 3: δ 𝑂¹⁸ =

18𝑂

16𝑂(𝑠𝑎𝑚𝑝𝑙𝑒) −¹⁸𝑂

16𝑂(𝑠𝑡𝑎𝑛𝑑𝑎𝑟𝑑)

18𝑂

16𝑂(𝑠𝑡𝑎𝑛𝑑𝑎𝑟𝑑)

× 1000 [

‰

]

(Hoefs 1997, Rollinson 1993, White 2011)

The standard used for oxygen is the Standard Mean Ocean Water (SMOW), which is characterised as ^𝑂

18

16𝑂= 2005.2 ± 0.43 ppm (Baertschi 1976). For carbon, the established standard is the Pee Dee Belemnite (PDB), with ^𝐶

13

12𝐶 = 11237.2 ± 2.9 ppm (Craig 1957). A positive δ-value means that the analysed substance is enriched in the heavier isotope, while a negative δ-value means that the substance is depleted in the heavier isotope, relative to the standard. Due to various geological processes, there can be a partitioning of isotopes between two substances in equilibrium with one another, leaving one enriched in ¹⁸O or ¹³C and the other one depleted. This is known as isotope fractionation (White 2011, Rollinson 1993).

3.2.6.1.2. Stable isotope fractionation

The stable isotope fractionation between two substances, 1 and 2, can be expressed by the fractionation factor 𝛼₁₋₂ obtained by the equation:

Eq 4: 𝛼₁₋₂ = 𝑅₁ 𝑅₂ =

(¹⁸𝑂

16𝑂)

1

(¹⁸𝑂

16𝑂)

2

(Javoy 1977, Hoefs 1997)

Furthermore, the isotope fractionation is also temperature dependent:

Eq 5: 1000𝑙𝑛(𝛼₁₋₂) =𝐴 × 10⁶ 𝑇² + 𝐵

(Bottinga & Javoy 1973, Hoefs 1997, Chiba et al. 1989)

T is the absolute temperature (Kelvin) while A and B are thermometric coefficients specific for certain pairs of substances (Chacko et al. 2001 and references therein) Note that for large T, 𝛼₁₋₂ will approach 1, meaning that there is practically no fractionation at sufficiently high

(24)

21

temperatures. In addition to temperature, other physical properties such as bond strength, ionic potential and oxidation state will have influence on the fractionation of stable isotopes.

Generally the heavier isotope is attracted to the stronger bond (Hoefs 1997, Rollinson 1993).

Correlating the δ-values of two substances with the fractionation between them is possible by the following approximations:

Eq 6: ∆₁₋₂= (𝛿 𝑂¹⁸ )₁− (𝛿 𝑂¹⁸ )₂~1000𝑙𝑛(𝛼₁₋₂) for δ¹⁸O < 10 ‰

Eq 7: 𝛼₁₋₂ =1000 + (𝛿 𝑂¹⁸ )₁

1000 + (𝛿 𝑂¹⁸ )₂ for δ¹⁸O > 10 ‰ (Hoefs 1997, Rollinson 1993, Friedman & O’Niel 1962)

In cases were fractionation factors for certain pairs of substances are not established, it is possible to combine fractionation factors:

Eq 8: 1000𝑙𝑛(𝛼₁₋₂) + 1000𝑙𝑛(𝛼₂₋₃) =

= 1000𝑙𝑛(𝛽₁) − 1000𝑙𝑛(𝛽₂) + 1000𝑙𝑛(𝛽₂) − 1000𝑙𝑛(𝛽₃) =

= 1000𝑙𝑛(𝛽₁) − 1000𝑙𝑛(𝛽₃) =

= 1000𝑙𝑛(𝛼₁₋₃)

(Schütze 1980, Zheng 1991, Zheng 1996)

Where β is a temperature-dependent thermodynamic isotope factor (see Zheng 1991, Zheng 1996).

By combining the above equations (Eqs 4-8), a variety of information can be extracted.

Knowing the δ-values of two mineral pairs and the fractionation between them enables one to calculate at which temperature the fractionation took place. And by knowing the fractionation at a specific temperature between a mineral and an equilibrium substance, the δ-value of said equilibrium substance can be determined (cf. the methodology used in Jonsson et al. 2013, Nyström et al. 2008, Weis 2013).

(25)

22 4. Results

4.1. Bulk geochemistry of REE-mineralised assemblages

REE-mineralised samples from all ten studied deposits were analysed for major and trace elements using XRF, ICP-AES and ICP-MS technology. The REE composition of analysed samples shows that samples generally plot in two groups - those with high REE enrichment and those with slight to no REE enrichment, compared to average crustal values (Fig 4). The highly enriched samples come from Bastnäs, Östra Gyttorp, Södra Hackspiksgruvan, Malmkärragruvan and Johannagruvan, while the low-enriched samples are from Danielsgruvan, Högfors and Myrbacksfältet. Samples from Östanmossa (three highly enriched, one low enriched) and Rödbergsgruvan (one highly enriched, one low enriched) plot in both groups. Bulk REE concentrations are of course highly dependant on the particular mineral assemblage sampled from each deposit, explaining why samples from some deposits plot in both groups. The LREE to HREE enrichment can be determined using the (Ce/Yb)cn-ratio (cf.

Belousova et al. 2002). The highly enriched group shows higher (Ce/Yb)cn-ratios, ranging from 48 to 637 with an average of 313. The low enriched group shows lower (Ce/Yb)cn-ratios, ranging from 15 to 188 with an average of 77. Comparing subtype 1 and suptype 2 deposits, subtype 1 deposits show LREE enrichment (average Ce/Yb = 424-608) compared to subtype 2 deposits (average Ce/Yb = 63-267). Deposits not covered in this classification (i.e. Östra Gyttorp, Högfors, Myrbacksfältet) all lie in the same range as subtype 2 deposits, while Danielsgruvan shows lower values (average Ce/Yb = 57).

(26)

23

Most samples show a pronounced Eu depletion, which can be quantified using the equation:

Eq 9: _𝐸𝑢∗^𝐸𝑢 = [_(𝑆𝑚 ^𝐸𝑢^𝑐𝑛

𝑐𝑛×𝐺𝑑_𝑐𝑛)^0,5] − 1 (Taylor & McLennan 1985)

Most samples show Eu/Eu*-ratios between -0.5 and -0.8. However, three samples show slightly positive Eu anomalies. These comprise of the two low enriched Danielsgruvan samples (0.32 and 0.38) and one of the highly enriched Bastnäs samples (0.26).

The samples show similar chondrite-normalised REE trends as the REE-rich apatite-iron oxide ores from Grängesberg (Jonsson et al. 2010, Jonsson et al. 2013), Blötberget (Jiao 2011) and Idkerberget (Sahlström 2012), all in northwestern Bergslagen. However the Bastnäs-type ores are generally slightly more LREE-enriched compared to these deposits. The large Bayan Obo deposit in China, which has been interpreted to have a magmatic-hydrothermal origin (from a presumably carbonatitic source), shows a much higher LREE to HREE enrichment (> 4 times) compared to the Bastnäs-type ores, and it lacks the negative Eu-anomaly (Zhongxin et al. 1992).

So called IOCG-type (iron-oxide-copper-gold) ores from the Olympic Dam deposit in Australia, which are suggested to have formed hydrothermally by combination of fluids from both deep-seated and meteoric sources (e.g. Haynes et al. 1995, Gow et al. 1994, Oreskes &

Einaudi 1992), show similar REE trends, yet featuring neutral to even positive Eu-anomalies (Oreskes & Eunaudi 1990).

Fig 4. Chondrite-normalised spider diagram of whole rock samples from the studied Bastnäs-type ores.

Average crustal values were collected from Webelements and chondrite values from McDonough & Sun (1995).

(27)

24 4.2. Mineralogy and mineral chemistry 4.2.1. Östra Gyttorpsgruvan

The Östra Gyttorpsgruvan samples consist of both massive magnetite and silicate dominated assemblages, hosted in mainly metavolcanic rocks (Nordenström 1890). The REE-silicate-rich assemblages occur as isolated, relatively small lensoidal bodies or pods and consist mainly of anhedral amphiboles, fine and even-grained allanite and biotite (Fig 4a). The allanites often show zoning, generally with BSE-brighter cores and BSE darker-rims. Uraninite is common and mostly present within allanite grains (Fig 4b). A gadolinite group mineral showing anhedral-subhedral crystals with complex zoning patterns occurs frequently, associated with allanite. The zoning can primarily be attributed to variable Y and Ce contents in the different parts of the crystal due to the antipathetic behavior of these elements in gadolinites, as described by Holtstam & Andersson (2007). Furthermore, a few zircon crystals have been observed in association with the allanite. The zircons exhibit a distinct morphology and zoning pattern, with a euhedral core and an subhedral, altered rim (Fig 4c). Some uraninites have been partially to wholly replaced by bastnäsite (Fig 4d). Bastnäsite is furthermore common in fractures, showing that it is a paragenetically late mineral (Fig 4e). It often exhibits lamellar intergrowths with parisite, which is related to the variability in stacking of (REEF),(CO3) and (Ca(CO3))-layers in the bastnäsite group minerals (Yunxiang et al. 1993). The massive magnetite ore samples consist of masses of medium-grained, homogenous and anhedral magnetite together with bands of mica (Fig 4f).

(28)

25

Fig 4a-f. Photomicrographs (a,f) and BSE images (b-e) of samples from Östra Gyttorpsgruvan. Images a), b), c), d) and e) are of sample Gytt-EJ-2 and image f) is of sample Gytt-EJ-1a.

a) Overview photograph (transmitted light) of silicate-dominated sample showing allanite-(Ce) (Aln), biotite (Bt), amphibole (Am) and uraninite (Urn). b) BSE image showing the presence of gadolinite group minerals (Gad) and uraninite in the allanite masses. The gadolinite minerals show complex zoning due to variable Y and Ce contents. Allanites exhibit zoning with Ca, Fe³⁺-rich rims (BSE-light) and REE, Fe²⁺-rich cores (BSE- dark). c) Zoned zircon (Zrn) with euhedral core and altered, subhedral rim. d) Bastnäsite-(Ce) (Bas) wholly replacing uraninite inside an allanite grain. e) Bastnäsite/parisite-(Ce)(Bas/Par) in fractures. Notice the lamellar intergrowth textures. f) Reflected light photomicrograph showing massive magnetite ore sample with bands of mica.

(29)

26

Samples from Östra Gyttorpsgruvan were analysed using WDS technique. One of the aims was to characterise the allanites, as they represent the main REE-bearing phase in the deposit. The chemical compositions of 26 analysed points are shown in Table 5 (appx). Based on these measurements, the average formula for Östra Gyttorp allanites can be expressed as:

A1(Ca0.96[]0.04)^A2(Ce0.41Nd0.18La0.16Pr0.05Sm0.03Y0.02Gd0.01∑REE0.86Ca0.14)∑1.00M3(Fe²⁺0.63Mg0.19

Fe³⁺0.18)∑1.00M2(Al1.00)^M1(Al0.79Fe³⁺0.21)∑1.00T(Si3.00)(O11.99F0.01)∑12.00(OH)1.00 if calculated according to the methodology described in Ercit (2002). From the REE concentration, it can be concluded that the allanites belong to the species allanite-(Ce), with significant and equal amounts of Nd and La (Fig 5). The compositional diagram (Fig 6) shows that there has been some substitution in the allanite structure, where Fe³⁺substitute for Al in M3 and Mg for Fe²⁺

in M1, indicating solid solution towards the end member CaCeMgAlFe³⁺Si3O12(OH). In the chemical formula it is also apparent that some REE are replaced by Ca in A2, in order to balance for Fe³⁺substituting for Fe²⁺in M1. This can be visualised by introducing epidote as another end member of the solid solution (Fig 7). The zoning in allanite (Fig 4b) can be attributed to these substitutions, with BSE-lighter cores being richer in REE+Fe²⁺and BSE-darker rims being richer in Ca+Fe³⁺.

Fig 5. Plot showing major REE (in atoms per formula unit - apfu) in allanite-group minerals.

(30)

27

Fig 6. Plot of Fe³⁺/(Fe³⁺+^VIAl) versus Fe²⁺/(Fe²⁺+Mg) (in apfu) for Östra Gyttorpsgruvan and Johannagruvan allanite group minerals, with compositional fields for allanites (from Holtstam & Andersson 2007) and ferriallanites (from Holtstam 2003) from the Bastnäs deposit. Modified after Holtstam & Andersson (2007).

Fig 7. Plot of Mg/(Mg+Fe²⁺) versus Mg+Fe²⁺ (in apfu) for allanite group minerals from Östra Gyttorpsgruvan and Johannagruvan.

(31)

28

The gadolinite group minerals were also analysed with WDS. The chemical compositions of 18 analysed points are shown in Table 6 (appx). The average chemical formula can be expressed as: (Ca0.32Y1.27Dy0.11Gd0.09Yb0.04Er0.04Ho0.03Nd0.03Sm0.02Ce0.01∑REE1.64)∑1.96(Fe0.39Mg0.03)∑0.42

Si2.04(Be1.68B0.32)∑2.00O8.83(OH)1.17, based on 4 A+T cations (cf. Miyawaki 2007, DeMartin et al. 2001).There has been significant substitution of Ca+B for REE+Be and (OH)2 for Fe (Fig 8), and all but three points should be classified as hingganites. However it cannot be excluded that some of the addition of Ca for REE was balanced by replacing Fe²⁺ with Fe³⁺, i.e. solid solution towards calciogadolinite, which was not possible to confirm with WDS data. This could possibly explain the low totals (Table 6, appx), and the excess of Si and shortage of A- site cations in the calculated mineral formula. These substitutions in Östra Gyttorp gadolinites make them differ significantly from gadolinites at other localities along the REE-line (this study, Holtstam & Andersson 2007), which exhibit close to ideal gadolinite composition. Fig 9 shows the dominant REE, and suggest that the gadolinite group minerals from Östra Gyttorp should be classified as hingganite-(Y)/gadolinite-(Y). The Y/Ce-ratios vary between 13 and 271, which is remarkably contrasting to gadolinites from other REE-line occurrences, which are much richer in Ce and Nd (this study, Holtstam & Andersson 2007). The large variability in Y/Ce-ratios is reflected in the zoning described above.

Fig 8. Compositional plot for gadolinite group minerals from Östra Gyttorpsgruvan and Johannagruvan.

(32)

29

Bastnäsite replacing uraninite (Fig 4d) were also analysed by WDS. The crystals are generally very small (a few µm) which hampers the analysis of volatile elements. The chemical composition of four analysed points are presented in Table 7 (appx). The compositional diagram (Fig 10) shows that the minerals plot along the bastnäsite-parisite line, indicating about 25 % parisite. This could be due to solid solution, but more likely due to microscopic lamellar intergrowths (as seen in Fig 4e) being included in the analysed points. Furthermore, there appears to be significant substitution of OH for F (based on calculations assuming ideal stoichiometry), suggesting hydration of the mineral (cf. Guastoni et al. 2009). However this substitution might be slightly exaggerated due to difficulties of obtaining accurate fluorine analyses of such small grains due to degassing under the electron beam. The average mineral composition (based on 6 cations) can be expressed as: Ca1.01(Ce2.14La0.97Nd0.96Y0.29Pr0.25

Sm0.17Gd0.12Dy0.03Eu0.03U0.02)∑4.98(CO3)6.01(F2.64(OH)2.35)∑4.99. The dominance of Ce means that the mineral should be classified as bastnäsite-(Ce) (Fig 11).

Fig 8. Plot of (REE+Be)/(REE+Be+Ca+B) versus (Fe+Mg)/(Fe+Mg+(OH)2) (in apfu) for gadolinite group minerals from Östra Gyttorpsgruvan and Johannagruvan.

Fig 9. Dominant REE (in apfu) in gadolinite group minerals from Östra Gyttorpsgruvan and Johannagruvan.

(33)

30

Fig 10. Plot of REE/CO3 versus Ca/(F+OH) (in apfu) of Östra Gyttorpsgruvan and Johannagruvan REE- fluorocarbonates. Analysed points from Östra Gyttorp were in replacements of uraninite, while points from Johannagruvan were in larger grains.

Fig 11. Plot of major REE (in apfu) in Östra Gyttorpsgruvan and Johannagruvan REE-fluorocarbonates.

Stable Isotope Systematics of Skarn-hosted REE-silicate - Magnetite Mineralisations in Central Bergslagen, Sweden

Examensarbete vid Institutionen för geovetenskaper ISSN 1650-6553 Nr 279

Stable Isotope Systematics of Skarn-hosted REE-silicate - Magnetite Mineralisations in Central Bergslagen, Sweden

Stable Isotope Systematics of Skarn-hosted REE-silicate - Magnetite Mineralisations in Central Bergslagen, Sweden

Fredrik Sahlström

Fredrik Sahlström

Supervisors: Erik Jonsson & Karin Högdahl Examensarbete vid Institutionen för geovetenskaper

ISSN 1650-6553 Nr 279

Stable Isotope Systematics of Skarn-hosted REE-silicate - Magnetite Mineralisations in Central Bergslagen, Sweden

Fredrik Sahlström

Stable isotope systematics of skarn-hosted REE-silicate - magnetite mineralisations in central Bergslagen, Sweden

Table of Contents

‰