Introduction

The northern Norrbotten area, northern Sweden (Fig. 1), hosts several economic and subeconomic Fe-oxide and Cu (-Au) deposits and has been described as an Fe-oxide Cu–Au (IOCG) district (Hitzman et al. 1992). The most economically significant deposits of the region are the Kiruna and Malmberget apatite–iron and the Aitik Cu-Au ores. The Tjårrojåkka area is located about 50 km WSW of Kiruna and hosts one of the best examples in Norrbotten of spatially related apatite–iron and copper deposits (Fig. 1). Following an extensive exploration program in 1967–1975, a large number of drill cores are available from the area, but no scientific results on the Tjårrojåkka occurrences have been published to date. The geology of the deposits is briefly described in Ros and Rönnbäck (1971), Grip and Frietsch (1973), Quezada and Ros (1975), Ekström (1978) and Ros (1979). More recently short descriptions of the Tjårrojåkka area have been published in Bergman et al. (2001), Edfelt and Martinsson (2003), Edfelt and Martinsson (2004), and Edfelt et al. (2004).

Fig. 1
figure 1

Geological map of northern Norrbotten showing the location of major Fe and Cu (-Au) deposits, and the Tjårrojåkka study area (after Bergman et al. 2001). Inset map: map of the Fennoscandian Shield with the location of the northern Norrbotten area. KNDZ Kiruna-Naimakka deformation zone; KADZ Karesuando-Arjeplog deformation zone; NDZ Nautanen deformation zone; PSH Pajala shear zone

Full size image

The geological settings, hydrothermal alteration systematics and mineralising fluid compositions among deposits classified as IOCG-type show a great variation (e.g. Porter 2001; Sillitoe 2003; Hitzman et al. 1992). Detailed descriptions of specific parageneses and mineral associations are important in order to understand the possible genetic relationships between different deposit types within this broad classification. This paper will describe the alteration characteristics of the Tjårrojåkka apatite–iron and Cu (-Au) occurrences in terms of whole-rock geochemistry, mineral chemistry and paragenesis. The mineral chemical data are also used as an indicator of the nature of the hydrothermal fluids involved in the formation of the deposits. These data will be used to examine the relationship between the two occurrences and compare them to other deposits in the region and elsewhere in the world.

Regional geological setting and metallogeny

The Precambrian bedrock in the northern Norrbotten region includes a ca. 2.8 Ga Archaean granitoid-gneiss basement, which is unconformably overlain by a metavolcanic sequence of Palaeoproterozoic age (Fig. 1). Stratigraphically lowest in the metavolcanic sequence are rift related 2.5–2.0 Ga Karelian units that are followed by ca. 1.9 Ga Svecofennian successions including several units of metavolcanic and epiclastic rocks. In the central Kiruna area the Svecofennian successions comprise, from the oldest to youngest, the Porphyrite Group, the Kurravaara Conglomerate, the Kiirunavaara Group and the Hauki Quartzite (Allen et al. 2004). Equivalent Palaeoproterozoic units are also found outside the Kiruna area. The calc-alkaline andesite-dominated Porphyrite Group is suggested to be subduction related, while the Kiirunavaara Group has a bimodal character and a geochemical signature resembling within-plate volcanic rocks (Martinsson and Perdahl 1994).

The approximately 10-km thick pile of Palaeoproterozoic volcanic and sedimentary rocks was deformed and metamorphosed contemporaneously with intrusion of the Haparanda (1.89–1.87 Ga) and Perthite monzonite (1.88–1.86 Ga) granitoid suites (Bergman et al. 2001). These plutonic rocks have a calc-alkaline to alkali-calcic character and are comagmatic with the Svecofennian volcanic rocks (Witschard 1984; Bergman et al. 2001). The Lina Suite comprises ca. 1.79 Ga granites and pegmatites (Skiöld et al. 1988), which are temporally related to Trans-Scandinavian Igneous Belt (TIB) 1 intrusions in the Kiruna-Narvik area (Romer et al. 1994; Romer et al. 1992). A second phase of metamorphism and deformation occurred at least locally at this time (Bergman et al. 2001).

Northern Norrbotten is an important mining province dominated by Fe- and Cu-deposits, with Au as a minor constituent in some of the Cu-occurrences. The main occurrences and their characteristics are summarised in Table 1. The economically most important deposits are the iron ores with an annual production of ca. 31 Mt of ore from the Kiirunavaara and Malmberget deposits (Fig. 1), and a total production of about 1,600 Mt from 10 mines during the last 100 years. Besides magnetite and hematite, most of the iron ores contain significant amounts of apatite. This class of deposits has been named “apatite–iron ores” or “Kiruna type” with the Kiirunavaara deposit being the largest and best-known example. Kiirunavaara contains more than 2,000 Mt of high-grade ore and was first described in detail by Geijer (1910). About 40 apatite–iron ores are known from northern Norrbotten. Individual deposits have an average content of Fe and P varying between 30–65 and 0.05–5%, respectively. Their spatial distribution coincides with that of the Kiirunavaara Group and they are almost exclusively hosted by metavolcanic rocks belonging either to the Kiirunavaara Group or the underlying Porphyrite Group (Martinsson 2003). Ore-elated alteration minerals include albite, scapolite, amphibole, K-feldspar, quartz, and sericite.

Table 1 Summary of characteristics of Fe-oxide and Cu–Au deposits in northern Norrbotten
Full size table

Copper was produced intermittently during the seventeenth and eighteenth centuries and recently on a larger scale in the Kiruna area. Sweden’s largest sulphide mine, Aitik, is situated in the Gällivare area (Fig. 1). With an annual production of 18 Mt of ore, it is one of the major Cu and Au producers in Western Europe. Although only a few economic sulphide deposits have been found in the northern Norrbotten ore province, a large number of epigenetic Cu–Au occurrences exist in the area. They exhibit large variation in mineralisation style, host rock composition and ore-related hydrothermal alteration.

Most copper deposits are hosted by tuffitic units of the Karelian greenstones and mafic to intermediate volcanic rocks within the Svecofennian porphyries (i.e. the Porphyrite Group and the Kiirunavaara Group). Some of them display a close genetic and/or spatial relationship to intrusive rocks varying in composition from monzodiorite to granite represented by plutons belonging to the Haparanda and Perthite monzonite suites. Magnetite is a common minor component in many of the deposits and in two cases (Gruvberget and Tjårrojåkka) the copper deposits occur adjacent to major magnetite deposits (Allen et al. 2004; Bergman et al. 2001). Besides structural traps, chemical traps may also be important, with redox reactions involving graphitic schists triggering sulphide precipitation. In addition to Cu, several occurrences also contain Co and/or Au in economic to subeconomic amounts (Martinsson 2000; Bergman et al. 2001).

Ore-related alteration is dominated by K-feldspar, albite, biotite, and scapolite with amphibole, carbonate, tourmaline, garnet, and sericite as locally important minerals. In most deposits the paragenetic sequence from oldest to youngest is: scapolite + biotite → albite → carbonate, or: scapolite + biotite → K-feldspar → sericite ± tourmaline. Stilbite and chabazite may be late phases occurring in druses and veins together with calcite. Ore minerals formed mainly at the intermediate or late stages of alteration. Geochronological data from Cu–Au deposits and hydrothermal alteration in the northern Norrbotten ore province demonstrates two major events of ore formation at ca. 1.87 and 1.77 Ga, respectively (Billström and Martinsson 2000; Edfelt 2003). The importance of saline hydrothermal fluids in the genesis of regional albite–scapolite alteration and the nature of the ore deposits in the northern Norrbotten ore province and adjacent Karelian areas in northern Finland and Norway has been emphasised by Frietsch et al. (1997). Highly saline fluid inclusions with 30–45 eq.wt% NaCl and depositional temperatures of 500–300°C are recorded for the Cu–Au deposits in this region (Ettner et al. 1993; Lindblom et al. 1996; Broman and Martinsson 2000). High Ca contents characterise ore fluids from most Cu–Au occurrences, which might be an expression of added components from evaporitic sediments within the Karelian greenstones that contributed to the salinity of the mineralising fluids (Wanhainen et al. 2003).

Geology of the Tjårrojåkka area

The geology in the Tjårrojåkka area is dominated by metamorphosed mafic to intermediate extrusive and intrusive rocks (Fig. 2). The stratigraphically lowest unit comprises metaandesites and metadolerites that are overlain by metabasalts. The metabasalts and metadolerites in the area have the same chemical signature and have been interpreted to have formed from the same magma with the dolerites acting as feeder dykes for the overlying basaltic unit (Edfelt 2003). Intrusions of gabbroic to quartz-monzodioritic composition crosscut the andesites and basalts. The rocks are metamorphosed in epidote-amphibolite facies, based on mineral assemblages (hornblende + plagioclase ± epidote ± quartz) (Spear 1993) of non-mineralised basic rocks (metabasalt and -dolerite). They have been strongly affected by albite, scapolite, and K-feldspar alteration that is more intense in the vicinity of deformation zones and mineralisation. From textural relationships (scapolite porphyroblasts growing over the metamorphic foliation in metabasalts and -dolerites) the regional alterations are interpreted as being temporally later than the metamorphism. Based on geochemistry the metaandesites resemble the intermediate rocks of the Svecofennian Porphyrite Group, while the metadolerites and -basalts have a more primitive signature and cannot be correlated with any known volcanic sequence in Norrbotten (Edfelt 2003).

Fig. 2
figure 2

Generalised geology of the Tjårrojåkka area with location of the Tjårrojåkka iron and copper deposits and minor occurrences. Inset map: drill holes at the Tjårrojåkka deposits with the investigated profiles indicated. Sections 400W and 320E shown in Figs. 3 and 4, respectively

Full size image

Rocks of the area, which are located within a splay off of a regional NW–SE trending deformation zone (Fig. 1), have undergone at least three stages of deformation including two compressional events (Edfelt 2003). The first compressional episode created NE–SW striking foliation parallel to the strike of the Tjårrojåkka deposits. It was followed by the development of an E–W trending deformation zone identified from aeromagnetic data showing a low magnetic anomaly and parallel foliation (shearing) in outcrops. The third deformation stage is characterised by ENE–WSW compression seen in folding in the central part of the area. The compressional stages can also be correlated with the regional tectonics in Norrbotten (cf. Bergman et al. 2001). Several structurally controlled Fe- and Cu-occurrences occur in the area (Sandrin and Elming 2003) of which the largest are the Tjårrojåkka magnetite–apatite (Tjårrojåkka-Fe) and the Tjårrojåkka copper-gold (Tjårrojåkka-Cu) occurrences located 750 m apart.

The Tjårrojåkka-Fe deposit, comprising massive magnetite with minor disseminated copper, was discovered through airborne magnetic measurements in 1963 by the Geological Survey of Sweden. A drilling program was initiated in 1967 and continued for 3 years during which some copper-bearing boulders and outcrops were found, and the Tjårrojåkka-Cu prospect was discovered. Between 1970 and 1975, 62 drill holes were drilled into the copper deposit. The Tjårrojåkka-Fe deposit is hosted by strongly sheared intermediate metavolcanic rocks and less deformed metadolerites. It consists of a massive magnetite core surrounded by a fractured host rock with apatite–magnetite veins filling the fractures (breccia) known to a depth of 400 m. The calculated tonnage for the apatite–iron deposit is 52.6 Mt at 51.5% Fe (Quezada and Ros 1975) with locally up to 3% Cu in some sections. The Tjårrojåkka-Cu occurrence, which is characterised by copper sulphides with minor quantities of magnetite, is hosted by the same rocks, localised in a 30 m wide and 700 m long zone, striking NE and dipping approximately 85° towards north. The deposit is estimated to contain 3.23 Mt at 0.87% Cu (cut-off 0.4%) (Ros 1979).

Sampling and analytical methods

Four drill sections, one in the apatite–iron ore and three in the copper deposit (Fig. 2), were logged and sampled. Seventy-six thin sections representing different rock and alteration types were initially examined in transmitted and reflected light at Luleå University of Technology and subsequently at the Natural History Museum, London using a Jeol 5900LV scanning electron microscope (SEM). SEM observations were made using a back-scattered electron detector (BSE), with an accelerating voltage of 20 kV and a beam current of 1 nA measured specimen current in pure cobalt metal.

Mineral analyses were performed using a Cameca SX50 WDS electron microprobe at the Natural History Museum, London, with the technique described in Potts et al. (1995). The analytical conditions and standards used for different minerals are available in Edfelt (2003) and the samples analysed are described in Appendix. Silicate analyses were carried out using an accelerating voltage of 15 or 20 kV, a beam current of 20 nA, and a 5-μm beam diameter. Apatites were analysed using an accelerating voltage of 15 kV, a beam current of 20 nA, and a 5-μm beam diameter. For sulphides and oxides a 1-μm beam diameter, an accelerating voltage of 15 or 20 kV, and a beam current of 20 nA were used, except for one set of sulphide analyses for which a 60 nA beam current was used. Different pure metals, natural minerals and synthetic glasses were used as standards. Interferences between X-ray peaks for Ba/Ti, Ce/Ti, Ce/Ba, Nd/Ce, Co/Fe, F/Ce, Mo/S and V/Ti were corrected empirically using previously collected data from standards.

Whole-rock analyses for major and trace elements were carried out on q drill core samples at Activation Laboratories Ltd in Canada. The major elements were analysed using the inductively coupled plasma method (ICCP-OES), while trace elements were analysed by inductively coupled plasma mass spectrometry (ICCP-MS) and instrumental neutron activation analysis (INAA).

Mineralisation and hydrothermal alteration

The main ore and alteration minerals and styles are summarised in Table 2. Cross sections through the Tjårrojåkka-Fe (400W) and Tjårrojåkka-Cu deposits (320E) (cf. Fig. 2), showing the relationships between mineralisation and main alteration types, are presented in Figs. 3 and 4, respectively. The apatite–iron ore (Tjårrojåkka-Fe) consists of a massive core (60–67% Fe and 0.5–1.3% P) surrounded by a breccia (25–60% Fe and 0.4–3% P) with low-grade copper mineralisation (Bergman et al. 2001), whereas the Tjårrojåkka-Cu consists of an elongated body of disseminated copper mineralisation with magnetite–apatite veining in the footwall. Albite, scapolite, and K-feldspar alteration has strongly affected the host rock to both deposits.

Table 2 Main ore and alteration minerals and styles in the Tjårrojåkka-Fe and Tjårrojåkka-Cu occurrences
Full size table
Fig. 3
figure 3

Cross section through Tjårrojåkka apatite–iron ore (profile 400W) showing the relation between the magnetite body, breccia, and alteration types. Alteration zones established based on geochemistry and visible appearance of alteration minerals. Ccp chalcopyrite; Bn bornite

Full size image
Fig. 4
figure 4

Cross section through the Tjårrojåkka-Cu deposit (profile 320E) showing the relationships between copper mineralisation and main alteration types. Alteration zones established based on geochemistry and visible appearance of alteration minerals. a Albite (Ab) altered footwall with overprinting magnetite (Mag)-apatite (Ap) veins. b Scapolite (Scp) altered hanging wall. c Intense K-feldspar (Kfs) alteration

Full size image

Mineralisation

Tjårrojåkka-Fe

Outcrop and drill core investigations indicate that the Tjårrojåkka-Fe deposit was the first of the occurrences to form since copper sulphides occur in fractures and veins crosscutting the massive magnetite. Magnetite is by far the most common ore mineral in the Tjårrojåkka-Fe deposit with minor hematite occurring as veins cutting the magnetite or as partly hematite-altered magnetite grains. Within the massive magnetite ore, veinlets of red or green apatite, tremolite, and carbonate fill fractures (Fig. 5a). Chalcopyrite, bornite, pyrite and minor molybdenite occur as veins and disseminations in the breccia and more rarely in fractures in the massive magnetite body. Gold (electrum) and silver telluride are trace minerals found in chalcopyrite (Fig. 5b). Based on textural relationships the sulphides in general post-date the massive magnetite, but do in some cases occur intergrown with magnetite in the massive ore and in veins in the breccia.

Fig. 5
figure 5

Photographs of alteration and mineralisation types and textures. a Typical massive magnetite ore with apatite, amphibole (tremolite) and carbonate infill from the Tjårrojåkka-Fe deposit. b Chalcopyrite with gold and hematite as late infill in fractures in massive magnetite in the Tjårrojåkka-Fe deposit (BSE image). c Chalcopyrite crosscutting magnetite in the footwall of the Tjårrojåkka-Cu deposit. d Martite (light grey) replacing magnetite (darker grey) in a vein in the footwall of the Tjårrojåkka-Cu deposit. e Epidote veinlets crosscutting K-feldspar-amphibole alteration in porphyritic andesite in the Tjårrojåkka-Cu deposit. f Albite altered to K-feldspar in the Tjårrojåkka-Cu deposit (BSE image). Ab albite; Am amphibole; Ap apatite; Bt biotite; Carb carbonate; Ep epidote; Kfs K-feldspar; Ttn titanite; Ccp chalcopyrite; Hem hematite; Mag magnetite

Full size image

Tjårrojåkka-Cu

The Tjårrojåkka-Cu deposit essentially consists of chalcopyrite, bornite, pyrite, and magnetite as disseminations, patches and in veinlets, locally with disseminated molybdenite. Magnetite occurs in footwall and is cut by later chalcopyrite (Fig. 5c) and carbonate veinlets. The magnetite in some cases exhibits martite replacement textures (Fig. 5d). Chalcopyrite and bornite occur as single grains or intergrown and are mainly associated with pervasive K-feldspar alteration and veins of amphibole ± K-feldspar ± quartz ± magnetite ± carbonate in both metaandesites and metadolerites. Chalcopyrite has also been identified intergrown with pyrite and magnetite. Bornite occurs in the part of the mineralisation richest in copper, while pyrite is more abundant in the eastern part of the deposit and at deeper levels. Silver telluride, silver sulphide, and native gold occur as micron-sized minor phases. Gold has been observed in quartz in a vein together with amphibole and chalcopyrite. Ekström (1978) also observed gold as inclusions in silicates associated with chalcocite and bornite. Chalcocite and covellite have been observed as secondary minerals replacing chalcopyrite and bornite (Ekström 1978) and locally oxidation of copper sulphides has resulted in the formation of malachite and chrysocolla.

Hydrothermal alteration

The hydrothermal alteration assemblages at Tjårrojåkka are highly variable with several of the alteration minerals occurring in numerous generations and settings, overlapping alteration stages, and with reactivation of already pre-existing veins, indicating a complex, long history of fluid activity in the area. The most widespread alteration minerals are albite, magnetite, apatite, scapolite, biotite, K-feldspar, and clinoamphiboles (tremolite, actinolite, Mg-hornblende, and tschermakite). The paragenetic evolution of the Tjårrojåkka deposits is illustrated in Fig. 6a, b. The alteration paragenesis in the two occurrences is similar, with albite forming at an early stage associated with magnetite and apatite. Scapolite was formed mainly before the main Cu-sulphide stage and is generally accompanied by biotite. The albitised and scapolitised rocks are overprinted by later K-feldspar alteration, which is spatially associated with copper-bearing sulphides. Several different types and generations of amphibole occur, both associated with magnetite and copper mineralisation and in post-mineralisation assemblages. Epidote and zeolites were the last phases to form from post main-ore stage low-temperature fluids.

Fig. 6
figure 6

Simplified paragenetic sequence of main ore and alteration minerals in the Tjårrojåkka apatite–iron (a) and Tjårrojåkka copper (b) occurrences

Full size image

Tjårrojåkka-Fe

The wall rock adjacent to the Tjårrojåkka apatite–iron deposit has been affected by extensive and pervasive albite alteration giving the rock a light grey or reddish colour due to hematite staining. Albite + magnetite alteration is particularly well developed in the area between the apatite–magnetite and the copper deposit. Scapolite occurs locally as porphyroblasts and later veinlets. The albitised and scapolitised rocks are overprinted by locally pervasive K-feldspar alteration and veins of K-feldspar + Mg-hornblende ± titanite ± quartz ± magnetite ± sulphides. Epidote is common together with K-feldspar, as late veinlets (Fig. 5e) and as an alteration of amphibole (Mg-hornblende). Amphibole (principally actinolite) also occurs in late veins cutting epidote. Allanite occasionally occurs in the matrix associated with epidote. Quartz veins have been observed in two generations. Carbonate veins (usually calcite), sometimes with zeolites ± pyrite, generally represent the final stage of infill in existing veins and vugs, or have exploited pre-existing fractures.

Tjårrojåkka-Cu

The footwall to the copper deposit is characterised by pervasive albite alteration overprinted by veins of magnetite and red, green, white or rare blue apatite (Fig. 4a). K-feldspar post-dates the albite alteration (Fig. 5f). Scapolite (porphyroblasts and veins) was formed at an early stage in the hanging wall (Fig. 4b), subsequently overprinted by pervasive K-feldspar alteration, and has affected the metadolerites to a greater extent than the metaandesites. Amphibole occurs in several generations as porphyroblasts, in monomineralic veins, or together with K-feldspar ± titanite ± quartz ± carbonate ± chalcopyrite ± bornite. The porphyroblasts contain inclusions of quartz, K-feldspar, plagioclase and iron oxide. Biotite occurs together with scapolite and is commonly affected by later chlorite alteration. Epidote occurs as patches in the matrix, together with K-feldspar ± amphibole ± carbonate ± quartz in veins or as a late mineral phase cutting all the earlier phases in thin veinlets. Zeolites (stilbite and chabazite) are fracture-controlled post-ore stage minerals sometimes occurring in earlier formed veins of amphibole ± epidote ± carbonate ± chalcopyrite ± pyrite. Fluorite has been observed in profile 600E in association with sericite and pyrite. REE minerals comprise allanite, occurring as rims on epidote, and late REE-carbonates in the magnetite–apatite altered footwall. Barite (associated with Cu-sulphides and in K-feldspar), thorite (intergrown with chalcopyrite or epidote), and zircon (in apatite and veins of chalcopyrite + feldspar + quartz) are minor hydrothermal constituents also observed in the copper deposit.

Whole-rock geochemistry

Geochemical analyses were performed on drill core samples to characterise the mass transfer during mineralisation and different types of alteration. Although attempts were made to sample least altered rocks, all samples exhibit some effect of alteration and/or metamorphism; hence the geochemical data do not record pristine magmatic features of the rock which in turn makes the mobile element interpretation difficult.

Major and minor elements

The host rocks to the Tjårrojåkka deposits show large variation in many of the major and minor elements due to the intense hydrothermal alteration (Table 3). The SiO2 content of the intermediate rocks varies between 50.16 and 67.86 wt% with total alkalis (Na2O + K2O) from 6.11 to 11.26 wt%. The Fe2O3(tot) contents range between 3.19 and 17.84 with TiO2 reaching a maximum of 0.92 wt%. The Zr content shows large variation from 67 to 439 ppm. The widespread potassic alteration is characterised by elevated values of K2O (max. 8.96 wt%) and BaO (max. 0.5 wt%), and the sodic alteration by Na2O contents reaching 9.57 wt%. This is clearly being illustrated in the Na2O versus K2O plot which also shows that Cu is correlated with potassic alteration (Fig. 7a).

Table 3 Major and trace element whole-rock geochemical data for representative rocks
Full size table
Fig. 7
figure 7

Whole-rock geochemistry plots of the host rocks to the Tjårrojåkka deposits. a Na2O-K2O plot showing inverse relationship and correlation of Cu-mineralised samples to potassic alteration. b REE patterns of representative rock samples. Chondrite normalised after Boynton (1984). c Metaandesites and metadolerites plotted on the igneous spectrum diagram after Hughes (1973). d Rock classification diagram after Winchester and Floyd (1977) revised by Pearce (1996)

Full size image

The metadolerites are characterised by a SiO2 range between 35.8 and 50.78 wt%, a higher CaO content, up to 9.16 wt%, compared to the metaandesites, and significantly lower Zr content (36–117 ppm). The metadolerites show the greatest variation in TiO2 with concentrations varying from 0.64 to 2.37 wt%.

Rare earth elements (REE)

Compared to the K-feldspar altered metaandesites, the albite altered metaandesites display a greater range in REE content. The La content of the K-feldspar altered samples varies between 26 and 86 ppm, with a mean of 51 ppm, while the La content of the albite altered samples varies between 15 and 208 ppm, with a mean of 58 ppm. The highest total REE content of the albite altered samples is comparable to the highest concentration observed in the mineralised samples.

The REE patterns of representative samples exhibiting different types of alteration and mineralisation, normalised after Boynton (1984), show LREE-enrichment and a negative Eu anomaly (Fig. 7b). The albite altered and Cu-mineralised samples show the greatest enrichment in REE compared to the least altered reference sample.

Element mobility during alteration and mineralisation

Element mobility was identified by plotting elements normally considered to be immobile (Al, Zr, Ti, Y, and Nb) against each other and other selected elements, and using isocon plots that compare concentrations of elements in altered relative to least altered samples (Grant 1986). The samples were divided into metadolerites and metaandesites, with the latter grouped as least altered, albite altered, K-feldspar altered and mineralised. The K-feldspar altered samples did not in most cases, show evidences of previous albitisation, which makes the comparison of element behaviour in the different alterations possible. The least altered sample used in the isocon plots was collected from an outcrop distal to the deposits and has a chemical composition similar to an unaltered Andean arc andesite (Raymond 1995).

A large number of samples plot outside the igneous spectrum in the diagram after Hughes (1973), and have higher total alkalis that expected, suggesting alkali mobility (Fig. 7c). In Fig. 7d the samples are plotted in the rock classification diagram Zr/TiO2-Nb/Y after Winchester and Floyd (1977) (revised by Pearce [1996]), which shows that most samples retain andesitic and basaltic affinity even after metamorphism and intense hydrothermal alteration. Most of the samples affected by K-feldspar alteration cluster relatively well (with one exception), while the albite altered and mineralised samples show a greater spread.

The behaviour of elements normally considered immobile is illustrated in Fig. 8. The two clear trends that can be distinguished in the TiO2-Zr plot indicate different origins of the intermediate and basic rocks and are in agreement with what has been observed in regional studies (Edfelt 2003). The roughly straight trends and clusters of Zr, TiO2, Al2O3, and SiO2, (Fig. 8a–c) suggest that these elements were, for the most part, conserved in the system and that the large variation of Zr in the andesites probably is a primary fractionation. Y, however, is more scattered and can be considered to have been mobile (Fig. 8e, f). The plot of Zr–Nb (Fig. 8d) show that albite altered and mineralised samples scatter most, while the K-feldspar altered are well clustered, which might indicate some degree of Nb mobility in these systems. Fig. 8f also demonstrates that Y was least mobile associated with K-feldspar alteration, but was mobile in the dolerites and related to albite alteration and mineralised samples. This also explains the spread of the dolerite samples in the classification diagram (Fig. 7d).

Fig. 8
figure 8

Immobile element plots for metaandesites and metadolerites at Tjårrojåkka

Full size image

In the isocon diagrams TiO2, Al2O3, SiO2 and Zr lie very close to the ideal isocon for all three groups of samples (albite altered, K-feldspar altered, and mineralised), suggesting that they were relatively immobile in all systems (Fig. 9). However, in the K-feldspar altered samples Zr show a slight enrichment compared to the reference sample. Albite alteration caused significant addition of Na2O and some addition of P2O5, resulting in the formation apatite, and a depletion of K2O, MnO, and Fe2O3. In the isocon diagram for K-feldspar altered samples, K2O, MnO, and P2O5 show the inverse relationship compared to the albite altered samples. K2O, MnO, P2O5, and Fe2O3 have been added in the mineralised samples whilst CaO is depleted in all three groups compared to the reference sample.

Fig. 9
figure 9

Isocon diagrams for metaandesites showing elemental changes associated with alteration and mineralisation. Average values for groups of altered samples are compared with least altered reference sample; n(albite altered)=22, n(K-feldspar altered)=28, n(mineralised)=15. Major oxides plotted in wt% and trace elements in ppm. For composition of reference sample see Fig. 7 and Table 3

Full size image

Barium enrichment characterises K-feldspar altered and mineralised rocks and is greatest in the latter. A slight enrichment of Fe2O3 and V occurs in the mineralised samples, probably due to formation of magnetite. All REE elements are enriched in the altered and mineralised samples compared to the reference sample, which is in agreement with the results form the REE-patterns (cf. Fig. 7b). The greatest addition of REE is observed in the mineralised samples.

Mineral chemistry

Silicates

Representative chemical compositions of feldspars are shown in Table 4. Feldspars are among the most abundant alteration minerals in the two deposits and can be divided into three groups: potassium feldspar (Or >90%), albite (Ab >90), and plagioclase (An 75–45). Albitisation is restricted to the host rock surrounding the Tjårrojåkka-Fe deposit and the footwall of the Tjårrojåkka-Cu deposit, whereas K-feldspar alteration is locally developed in the Tjårrojåkka-Fe deposit associated with Cu-mineralisation and in the hanging wall of the Tjårrojåkka-Cu deposit. Plagioclase occurs in parts of the Tjårrojåkka-Fe deposit. The potassium feldspars have a varying content of Ba substituted for K, but there does not seem to be a systematic variation within individual grains. However, Cu-mineralised samples and pervasive K-feldspar alteration tend to be richer in Ba than non-mineralised samples and K-feldspar occurring in veins. Some samples from the apatite–iron occurrence contain more than 2 wt% BaO (Fig. 10) and can be considered as hyalophane (Deer et al. 1992).

Table 4 Representative results of electron-microprobe analyses of feldspars
Full size table
Fig. 10
figure 10

Variation in BaO content in K-feldspar

Full size image

Scapolite has a meionite (Ca4Al6Si6O24CO3) content (Me=100×Ca/(Ca + Na + K) between 30 and 55 (Fig. 11a). The Cl content varies between 0.9 and 2.9 wt% and CO2 between 1.2 and 2.8 wt% while the F content is less than 0.2 wt% (Table 5). The SO3 contents show wide variation from 0 to 1.5 wt%. The scapolite in the Tjårrojåkka-Fe deposit has higher Cl (2.3–2.9 wt%) and lower S content than samples from Tjårrojåkka-Cu deposit (Fig. 11b). Scapolite from unmineralised wall rock has a distinct character in being richer in CO2 than scapolite related to mineralisation.

Fig. 11
figure 11

Variation in scapolite composition. a Diagram showing variation in meionite (Me) content. Me=100×Ca/(Ca + Na + K). b Cl-C-S diagram. All atoms per formula unit

Full size image
Table 5 Representative results of electron-microprobe analyses of scapolite
Full size table

The composition of biotites, shown in Table 6, is between phlogopite/annite and eastonite. The Ti content varies from 1 to 3 wt% TiO2, with the highest contents in biotite associated with the Tjårrojåkka-Cu deposit. The Ba content is higher in Cu-mineralised samples than in non-mineralised samples. The amount of Cl varies between 0.2 and 0.5 wt% and F between 0 and 0.8 wt%. The biotites from the apatite–iron ore plot in two distinct groups with respect to the Mg/Fe and F contents (Fig. 12a). In, or close to the breccia, the Mg content is higher and the F content lower, than in the samples outside. The sample that shows the highest F values is also richest in Cl (0.5–0.6 wt%). The biotites from the Tjårrojåkka-Cu deposit show less variation in Mg/Fe ratio, but Cu-mineralised samples are generally more Mg-rich (Fig. 12a, b). In Fig. 12b three linear trends can be distinguished with the amount of Cl increasing with Fe.

Table 6 Representative results of electron-microprobe analyses of biotite
Full size table
Fig. 12
figure 12

Diagrams showing compositional variation in biotite. a Plot of F against Mg/(Mg+Fe2+). b Plot of Cl against Mg/(Mg+Fe2+)

Full size image

The amphiboles in the Tjårrojåkka-Fe and -Cu deposits are Ca-rich and range from tschermakite to magnesio-hornblende to actinolite and tremolite (Table 7 and Fig. 13a). The most widespread types are tschermakite and Mg-hornblende occurring in the matrix, often together with pervasive K-feldspar alteration, or in fractures together with chalcopyrite or bornite. Actinolite is found in veins where it generally is paragenetically later the other amphiboles, in the breccia surrounding the apatite–iron body, and in the Tjårrojåkka-Cu deposit. Tremolite only occurs as veinlets in the massive magnetite ore in the Tjårrojåkka-Fe deposit. The amount of Cl in the amphiboles increases with the Fe content and is highest in the tschermakites (Fig. 13b). F is present in the amphiboles in the Tjårrojåkka-Cu deposit (0.1–0.2 F per formula unit) but it is below detection limit in the amphiboles from the Tjårrojåkka apatite–iron ore.

Table 7 Representative results of electron-microprobe analyses of amphiboles
Full size table
Fig. 13
figure 13

Composition of amphiboles (cf. Table 7). a Classification of amphibole composition after Leake et al. (1997). b Variation in Cl content in amphiboles

Full size image

Chlorite, titanite, epidote and allanite are minor constituents among the rock-forming and hydrothermal alteration mineral assemblages and their chemistry will not be discussed in detail. Titanite is more common in the alteration assemblages in the copper deposit and contains between 0.2 and 1.2 wt% F, around 1–2 wt% Fe2O3 and traces of Ce. The Fe2O3 content in the epidote varies between 15.5 and 17.2 wt%. REE were not detected.

Apatite

The analysed apatites classify as fluor-apatites with F contents between 1.6 and 3.4 wt% (Table 8). The apatites in the copper occurrence have higher F than those in the iron ore. Apatites from an unmineralised outcrop of metaandesite, located about 1 km WNW of the Tjårrojåkka deposits, are the most F-rich (Fig. 14).

Table 8 Representative results of electron-microprobe analyses of apatite
Full size table
Fig. 14
figure 14

F-Cl-OH diagram showing compositional variation in apatite (atoms per formula unit)

Full size image

The apatites in the massive magnetite ore in Tjårrojåkka-Fe are the most Cl-rich (0.9–1.6 wt%), with a few exceptions that show Cl values around 0.3 wt% probably due to zoning with the rims being Cl-poorer. The slightly high totals in the analyses may be a result of either the breakdown of the mineral under the electron beam or a calibration problem due to partial breakdown of the standards with time.

Sulphides and oxides

Representative analyses for sulphides (chalcopyrite, pyrite and bornite) and oxides (magnetite and hematite) are presented in Table 9. The sulphides do not show large compositional variations between the two deposits. In a few samples, chalcopyrite shows traces of Se, Ag and Au and some pyrites associated with the Tjårrojåkka-Cu deposit have significant Co contents of up to 1.8 wt%. Cr, Mn, Ni, Sb, Te, Hg, Pb, and Bi were also analysed, but were below detection limits.

Table 9 Representative results of electron-microprobe analyses of sulphides and oxides
Full size table

The V2O5 content in magnetite and hematite range from 0.1 to 0.9 wt%, with the highest values in hematite from the Tjårrojåkka-Cu deposit. Mn, Co, and Ni were also detected in some of the samples, but are generally below the detection limit. There do not seem to be any systematic variations in minor element compositions of iron oxides (Al, Ti, V, and Mn), except for Cr which is slightly enriched in magnetite associated with copper sulphides.

Discussion

Element mobilisation and chemical variations

Element mobilisation and redistribution is common during hydrothermal alteration, however, in terrains that have been subject to extensive regional alteration, metamorphism and/or metasomatism the quantification of element mobility is difficult. In the case of the Tjårrojåkka occurrences, the degree of element mobility and transport is best illustrated by considering the geochemical systematics of the metaandesitic rocks. Thirty-nine percent of the samples plot outside the igneous spectrum in the diagram after Hughes (1973) as a result of potassic and sodic alteration (cf. Fig. 7c). The data show that the albite altered metaandesites (mainly in the footwall of the Tjårrojåkka-Cu deposit and the host rock of the Tjårrojåkka-Fe deposit) have been subject to a relative enrichment in Na, while the hanging wall of the Tjårrojåkka-Cu and the copper mineralised zones are characterised by a relative enrichment in K.

The distributions of both the major, minor and trace elements suggest that the degree of mobility within the K-enriched and Na-enriched samples is systematically different. This is particularly well illustrated by the distribution of Na2O, K2O, P2O5, Ba, Y, and REE. Albite altered and mineralised samples scatter in Y and REE plots, indicating that the elements were mobile in these systems, while the K-feldspar altered samples cluster. The variation of Y in the dolerites could be due to the intense scapolite alteration breaking down primary mafic minerals.

Mobility of Zr, Ti and REE during hydrothermal processes has been noted by many authors, including Gieré (1990) and Rubin et al. (1993), in fluids where P, F, and K and/or Na were important components along with high activity of CO2. However, the variation of Zr and Ti in the andesites is most probably a result of primary fractionation although Zr shows a slight enrichment in the K-feldspar altered samples and hydrothermal zircons and titanites have been observed. On the other hand, the enrichment of REEs in altered and mineralised samples, relative to least altered, and the presence of allanite and late REE-carbonates indicate that REE were mobile at Tjårrojåkka.

Alteration paragenesis and the evolution of fluid chemistry

Similarity in alteration minerals and paragenesis may partly be a product of the common host rock to the Tjårrojåkka-Fe and Tjårrojåkka-Cu occurrences, but is also an indication of similarities in fluid compositions and depositional conditions. Ba, Cl, F and S are elements enriched in the alteration minerals in the Tjårrojåkka occurrences and can be used as indicators of the nature of the hydrothermal fluids. Variation in the content of these elements in K-feldspar, scapolite, apatite, biotite and amphibole clearly suggests differences in the physical and/or chemical environment during alteration and mineralisation in the two deposits.

Barium feldspars commonly occur associated with manganese deposits (Deer et al. 1992), but have also been noted in, for example, the galena deposit at Korsnäs (Mäkipää 1976) and the Pikkuharju Cu–Zn mineralisation (Lahtinen and Johanson 1987) in Finland, the Rosh Pinah Pb–Zn deposit in Namibia (Page and Watson 1976), and the Ernest Henry IOCG-deposit in Australia (Mark et al. 2000). At Tjårrojåkka the Ba content in K-feldspar varies between the two deposits. In Tjårrojåkka-Fe deposit K-feldspar with a celsian component (BaAl2Si2O8) occurs in the Cu-mineralised breccia surrounding the massive magnetite body indicating a high Ba/K ratio in the hydrothermal fluids responsible for this K-feldspar alteration. The amount of Ba in K-feldspar is lower in samples from the Tjårrojåkka-Cu deposit and lowest in the non-mineralised samples.

Scapolite is in some districts a common mineral in metamorphic and metsomatic rocks and can be used as an indicator of volatile activities and the Cl content of the fluid salinity (e.g. Shaw 1960; Vanko and Bishop 1982). The occurrence of marialite (Na4Al3Si9O24Cl)-rich scapolite indicates high activities of NaCl in the rock or fluid (Orville 1975) and regional occurrences of scapolite rich in Cl possibly indicate the presence of metamorphosed evaporitic sequences (Ellis 1978). The scapolite at Tjårrojåkka shows a trend with more Cl-rich varieties around the magnetite body trending towards higher SO3- and CO2-contents in the Tjårrojåkka-Cu deposit. The same compositional variation has been observed in the Malmberget apatite–iron ore (Fig. 1) where the scapolite is Cl-rich (3.8 wt%) and in the nearby Nautanen Cu–Au mineralisation (Fig. 1) scapolite is dominated by SO3and CO2 (Frietsch et al. 1997). At Tjårrojåkka the scapolite most distal to the copper deposit is more CO2-rich and SO3-poor than scapolite from the mineralised part, and can hence be interpreted as having formed from a SO3-depleted hydrothermal fluid.

Apatite is a common mineral in the Tjårrojåkka occurrences and since the three solid-solution end-members constitute Cl-, F- and OH-apatites, these elements can be used as indicators of the composition of the hydrothermal fluids (Korzhinskiy 1982). Korzhinskiy (1982) also showed that the Cl/F ratio in apatite increases with temperature and that the pressure effects are negligible. The apatites analysed from an outcrop sample, located approximately one km WNW of the Tjårrojåkka deposits, have the highest F while those clearly related to the mineralising processes from the copper deposit are more Cl-rich. The outcrop apatites are clearly distinct from the apatites from the deposits and imply lower Cl activities during formation, reflecting either primary magmatic conditions or subsequent metamorphism of apatite in the presence of relatively low salinity fluids. Compared to apatite from the Kiirunavaara apatite–magnetite ore (Harlov et al. 2002) the apatites at Tjårrojåkka are richer in Cl and H2O and poorer in F. La and Ce are generally lower while Nd shows similar values to the apatites in Kiirunavaara.

The interpretation of the halogen contents of silicate minerals is complicated by crystal chemical effects between the hydroxyl site and cation sites within the minerals, generally termed the Fe–F avoidance principle (e.g. Ekström 1972; Rosenberg and Foit 1977). The halogen composition of biotite (assuming no post-crystallisation re-equilibration) will be a function of the Mg:Fe ratio of the biotite as well as P-T conditions at the time of crystallisation, and the fluid chemistry (Zhu and Sverjensky 1991; Munoz 1984). Biotites from different parts of the systems do not show great variation in chemistry, except biotite from a distal part of the iron ore that differs from the others in being the most Fe-rich and showing the highest content of F. However, in a plot of Cl against Mg/(Mg+Fe2+) (Fig. 8b), three linear trends can be distinguished originating from differences in temperature or salinity of the fluids, or representing different generations of biotite.

Previous studies have also suggested that the F and Cl contents of amphiboles are influenced by mineral structure and crystal chemistry (including the Fe–F and Mg–Cl avoidance effects) as well as the P-T conditions and halogen activity in the co-existing fluid (e.g. Oberti et al. 1993). All amphiboles at Tjårrojåkka are Ca-rich with the highest F content in the amphiboles in the Tjårrojåkka-Cu deposit, and the highest Cl in the tschermakites. Oberti et al. (1993) showed that an increase in Cl content would require increasing Fe2+, K, and Al which is consistent with the trends in the amphiboles from Tjårrojåkka. A more extensive interpretation of fluid composition from halogen chemistry in biotite, amphiboles and apatite would, however, require temperature and pressure data, which are currently not available.

Magnetite and hematite have similar geochemistry to magnetite from Kiirunavaara and El Laco in Chile (Nyström and Henríquez 1994) in being rich in V (average 2,860 ppm) and low in Ti (average 240 ppm) and Cr (average 340 ppm).

The fact that late REE-carbonates occur in the footwall of the copper deposit and that allanite rims on epidote are common indicate late infiltration of REE enriched fluids. Previous studies have shown that allanite and apatite may form as replacement products of monazite during hydrothermal alteration (Finger et al. 1998; Wing et al. 2003), which could explain the low content of REE in apatite (cf. Table 8) and the absence of monazite. Another possibility could be REE leaching from apatite during late stage alteration and metamorphism, which has been suggested to account for REE depleted apatite rims and the development of late stage monazite and allanite in the Kiirunavaara magnetite body (Harlov et al. 2002).

Overall, the alteration minerals (K-feldspar, scapolite, and apatite) related to the Tjårrojåkka apatite–iron ore are more Cl- and Ba-rich compared to the alteration minerals in the copper deposit that have higher contents of F and SO3. Higher Ba in K-feldspar near the iron deposit could reflect lower fluid sulphate concentrations associated with a high Ba/K ratio, which is supported by higher Ba contents in whole-rock analyses of K-feldspar altered samples from Tjårrojåkka-Fe. The presence of in scapolite and the existence of minor barite and late hematite in the copper deposit point towards more oxidising conditions during the formation of the Tjårrojåkka copper deposit.

The mineral chemical and paragenetic results can be interpreted in two ways; either (a) there were two different hydrothermal systems; one reduced fluid with a high Ba/K ratio, high salinity and low sulphate concentration forming the Tjårrojåkka-Fe deposit, and another one more oxidised and F-SO4-CO2-rich forming the Tjårrojåkka-Cu deposit, or (b) there was one evolving system. An evolving system would require lowering of Cl contents of the fluid, which could be achieved either by fluid mixing or by loss of Cl to minerals, with the latter being a common feature in Cu–Au deposits in the Cloncurry district, Australia (Baker 1998). There, the loss of Cl from the fluids gave rise to hornblende and biotites with Cl contents up to 3.5 wt% and other Cl-bearing phases such as scapolite and apatite. However, at Tjårrojåkka the Cl content in the biotites and amphiboles is much lower (<0.6 wt%), but scapolite and apatite in the apatite iron-body are more Cl-rich than in the copper deposit and could have influenced the reduction of salinity. Some preliminary fluid inclusion work on the Tjårrojåkka occurrences indicates moderately to highly saline (15–32 eq.wt% CaCl2 + NaCl) systems (Broman and Martinsson 2000; Edfelt et al. 2004), which is in accordance with data from other copper deposits in the region (Wanhainen et al. 2003; Broman and Martinsson 2000; Lindblom et al. 1996). Edfelt et al. (2004) also noted an increase in salinity and the appearance of carbonate daughter minerals going from the apatite-forming stage to the Cu-sulphide stage, with a likely cause being fluid mixing.

The Tjårrojåkka occurrences as IOCG type deposits

The Tjårrojåkka Fe-oxide Cu–Au occurrences share many characteristics (structural control, abundance of iron oxides, anomalous concentrations of REE, albite-scapolite-K-feldspar alteration) with deposits classified as IOCG-type (e.g. Hitzman et al. 1992; Marschik and Fontboté 2001; Porter 2001). The common spatial relationship between apatite–iron and copper ores has also been noted between more recent deposits of Fe-oxide Cu-Au-type in Cretaceous iron belt (Naslund et al. 2002) and Candelaria-Punta del Cobre deposits (Marschik and Fontboté 2001) in Chile, which show many similar features with the Tjårrojåkka occurrences.

The magnetite–apatite occurrence at Tjårrojåkka has similar characteristics to the Kiirunavaara apatite–iron ore with magnetite as almost the only iron oxide and a breccia developed along the wall rock contacts (Martinsson 2003). The Tjårrojåkka apatite–iron deposit differs from the Kiruna type apatite–iron ores in Norrbotten only in the higher sulphide content of the breccia surrounding the massive magnetite body. Both magmatic and hydrothermal replacement models have been suggested for the formation of the apatite–iron ores of Kiruna-type (e.g. Hitzman et al. 1992; Nyström and Henríquez 1994), but from the existing data it is not possible to prove either of these models for the Tjårrojåkka iron ore. However, the extensive hydrothermal alteration and veining around the massive magnetite body indicate that hydrothermal processes were definitely active at least at a later stage during the ore formation.

The Tjårrojåkka-Cu deposit might be related to this late stage hydrothermal activity and considered as a copper dominated end-member in the IOCG spectrum of deposits. It is characterised by strong sodic and potassic alteration comparable to those surrounding the apatite–iron ore, but show a stronger structural control. The presence of metadolerites in the mineralised zone in the Tjårrojåkka-Cu deposit could also have played and important role for mineralisation as pathways for the fluids.

Conclusions

The Tjårrojåkka occurrences can be considered as belonging to the IOCG-group of deposits representing two “end-members” of the class, with a spatial and possibly also genetic relationship. The Tjårrojåkka apatite–iron deposit has the typical characteristics of the Kiruna-type iron ores, except the high concentrations of sulphides in the surrounding ore breccia. The Tjårrojåkka-Cu occurrence is similar to epigenetic copper deposits in the region and other Fe-oxide Cu–Au deposits elsewhere in the world (e.g. Chile).

The whole-rock geochemistry indicates enrichment of alkalis related to mineralisation due to the formation of albite and K-feldspar. There was enrichment in Na and P and depletion of K, Ba, and Mn related to albitisation, with the inverse relationship of these elements associated with K-feldspar alteration. Fe and V show depletion in the altered zones and addition in mineralised samples. REE were enriched in the system, with the greatest addition related to mineralisation. Y mobility was associated with albite alteration and copper mineralisation.

Several generations and overlapping hydrothermal alteration stages indicate a long, complex history of fluid activity related to the formation of the Tjårrojåkka deposits. The two occurrences at Tjårrojåkka show a similar evolution in alteration paragenesis and mineralogy, but with more oxidising, CO2-, F-, and -rich fluids related to copper deposit, in contrast to the Tjårrojåkka-Fe deposit where the fluids were more reduced with a higher salinity and Ba/K ratio. This might reflect one evolving system forming both occurrences, with the copper deposit representing slightly later products, but without geochronological data and more detailed fluid inclusion and isotopic studies we cannot rule out formation by two unrelated mineralising events.