The Canadian Mineralogist; October 2007; v. 45; no. 5;
p. 1073-1114; DOI: 10.2113/gscanmin.45.5.1073
© 2007 Mineralogical Association of Canada
THE REE MINERALS OF THE BASTNÄS-TYPE DEPOSITS, SOUTH-CENTRAL SWEDEN
Dan Holtstam1,
and
Ulf B. Andersson2,¶
1 Department of Mineralogy, Swedish Museum of Natural History, Box 50007, SE–104 05 Stockholm, Sweden
2 GeoforschungsZentrum, P.B. 4.1, Telegrafenberg, D-14473 Potsdam, Germany
 |
ABSTRACT
|
|---|
On the basis of characteristic assemblages, the Bastnäs-type Fe–(Cu–)REE skarn deposits in the Bergslagen mining region of south-central Sweden can be divided into two subtypes: one almost exclusively with LREE enrichment (subtype 1, mainly in the Riddarhyttan–Bastnäs area), and another showing enrichment of both LREE and Y + HREE (subtype 2, Norberg District). New data have been collected on twenty REE species including three unnamed ones, using electron-microprobe analysis, X-ray diffraction and Mössbauer spectroscopy. Cerite-(Ce) is, together with ferriallanite-(Ce), the most important REE mineral at Bastnäs, but less common at the subtype-2 deposits. Compared to other occurrences worldwide, the present samples of cerite-(Ce) are poor in Ca and enriched in Mg and F. Calcium and (REE + Y) in the samples are negatively correlated, and the upper limit for Ca is close to 1 apfu. Iron ranges from 0.01 to 0.30 apfu, and is mainly in the trivalent state. Cerite-(Ce) is a major carrier of Y (up to 3.5 wt.% Y2O3). Fluorbritholite is found only in the subtype-2 deposits, and is low in P, Na and the actinides compared to other major occurrences in the world. Significant inter- and intrasample variations in REE and Y occur, ranging from fluorbritholite-(Ce) to the unaccredited member "fluorbritholite-(Y)". Fluorbritholite is non-stoichiometric, with Ca < 2 and (REE + Y) > 3 apfu. Västmanlandite-(Ce) is another important host for the REE in subtype 2. By means of the substitutions Mg2+ + F– 
Fe3+ + O2– and Mg2+ Fe2+, it forms solid solutions with an unnamed Fe-dominant member that is found in the subtype-1 deposits. Dollaseite-(Ce), which forms a partial solid-solution series with dissakisite-(Ce), is restricted to subtype 2. Gadolinite is very rare at the subtype-1 deposits, and more widespread in the Norberg District. All samples have a significant hingganite component in solid solution (0.10–0.35 molar fraction). Single crystals are commonly zoned with respect to REE and Y, with compositions corresponding to gadolinite-(Ce), gadolinite-(Y) and an Nd-dominant member. The substantial fractionation of REE and Y on a local scale within the Bastnäs-type deposits is dependent both on crystal-chemical factors and on fluctuations in fluid composition during crystallization. The initial precipitation of REE silicates in type-1 and type-2 deposits (mainly cerite and fluorbritholite, respectively) was a reaction between relatively acidic solutions carrying major amounts of REE complexed mainly by ligands of F and Si, and dolomitic host-rocks.
Keywords: REE minerals, cerite, dollaseite, fluorbritholite, gadolinite, törnebohmite, västmanlandite, Bastnäs, Norberg, Sweden.
|
INTRODUCTION
|
|---|
The Bastnäs Fe–Cu–REE deposit belongs to the Riddarhyttan ore field, Skinnskatteberg District, Sweden, where mining, primarily for iron, was pursued already in late medieval times (Carlborg 1923). The scientific study of this remarkable locality commenced when Cronstedt (1751) described a "flesh-colored, heavy rock" called Bastnäs tungsten, which had been found in a small copper mine at Bastnäs, known as S:t Göransgruvan ("St. George’s mine"). This material is presently known as the mineral cerite-(Ce), from which Hisinger & Berzelius (1804) extracted previously unknown oxides. They named the corresponding metal cerium, and subsequent discoveries of other rare-earth elements (REE) and new mineral species have made the locality prominent in the history of natural science.
Geijer (1961) introduced the generic term "Bastnäs type" by including other Fe deposits, all strongly enriched in REE and with a putatively common origin, situated mainly in the Norberg District, ca. 30 km northeast of Bastnäs. The present work also includes the Rödbergsgruvan deposit in the Nora District, 50 km to the southwest, where the characteristic REE mineralization with, e.g., cerite-(Ce), was not recognized until the 1980s.
Researchers have paid little attention to the Bastnäs-type deposits in recent times, and few primary data have been collected since the 1920s. Here we present new mineral-chemical and some paragenetic data for the REE minerals. A second paper (Holtstam et al., in prep.) focuses on the genetic aspects of these unique deposits.
|
GEOLOGICAL SETTING AND MINERAL ASSEMBLAGES OF THE DEPOSITS
|
|---|
The Bastnäs-type deposits are restricted to a northeast-trending narrow carbonate-bearing zone within early Svecofennian (1.91–1.88 Ga) supracrustal rocks, mainly felsic metavolcanic rocks and marble, situated in the northwestern part of the Bergslagen mining region, central southern Sweden (Fig. 1). The metavolcanic sequences have undergone extensive synvolcanic hydrothermal alteration, causing metasomatic enrichments in K, Na, and Mg in successive pulses and areas (e.g., Geijer 1923, Trägårdh 1991, Hallberg 2003). In the areas of the Bastnäs-type deposits, Mg metasomatism is most prominent in the country rocks, and the supracrustal sequences later became metamorphosed under amphibolite-facies conditions (Trägårdh 1988). Allen et al.(1996) provided a synthesis of the volcano-tectonic history of Bergslagen.
View larger version (43K):
[in this window]
[in a new window]
|
FIG. 1. Geological map (position indicated by small rectangle in the inset map) of the area with location of Bastnäs-type deposits, based on work by the Geological Survey of Sweden (Ambros 1983, 1988, Lundström & Koark 1979). R: Rödbergsgruvan, B: Bastnäs field (subtype 1). Malmkärra (M), Johannagruvan (J), S. Hackspikgruvan (H) and Östanmossa (Ö) are located near the town of Norberg (subtype 2).
|
|
The supracrustal units were intruded by at least two generations of plutonic rocks. The older, early Svecofennian (1.90–1.86 Ga) intrusions range from gabbro through tonalite and granodiorite to granite in composition, and are normally deformed and metamorphosed along with the supracrustal units (e.g., Lindh & Persson 1990). Rocks of the younger, late Svecofennian (1.81–1.75 Ga) group vary from undeformed, homogeneous granites to migmatites (Andersson & Öhlander 2004, and references therein). Overlapping in time with the latter (1.82–1.78 Ga, Romer & Smeds 1997), numerous pegmatite dykes and bodies cut the early Svecofennian rocks in Bergslagen (e.g., Ambros 1983). Later tectonic stresses caused fracturing and faulting, affecting all types of rock in the area.
In the Norberg District, typical REE mineralization has been reported from a few iron mines, now abandoned. The deposits consist of disseminated to massive magnetite–amphibole skarn replacements in dominantly dolomitic marbles (Geijer 1936). Fluorine-rich minerals, such as norbergite, chondrodite, fluorian phlogopite and fluorite, are commonly associated with the REE minerals (Geijer 1927). Sulfide mineralization, with mainly pyrite, molybdenite and chalcopyrite, is locally important. Samples used in the present investigation originate from the mines Östanmossagruvan (60°5'N, 15°56'E), Södra Hackspikgruvan (ca. 60°4'N, 15°57'E; now inaccessible), Johannagruvan (60°4'N, 15°55'E) and Malmkärragruvan (60°4'N, 15°51'E).
At the Bastnäs (strictly Nya Bastnäs) deposit (59°51'N, 15°35'E), quartz-banded hematite ore is juxtaposed with a magnetite-skarn ore that has more or less completely replaced a dolomite marble horizon (Geijer 1921). The dominant country-rock is a metasomatized volcanic rock, essentially quartz-rich, commonly cordierite-bearing mica schist (Ambros 1983). The REE mineralization is associated with amphibole skarns and was encountered in two shallow mines at a maximum depth of 30 m, where it formed restricted zones up to 0.6 m thick and less than 10 m long (Geijer 1921). Talc and quartz appear locally. Sulfide minerals, dominantly chalcopyrite, bismuthinite and molybdenite, are closely associated with the REE minerals, and commonly interstitial to them. Minor opaque phases detected are carrollite, bornite, covellite, wittichenite, emplectite, hodrushite, tetradymite, Kup
íkite, native copper, native bismuth and gold–silver alloys; uranium oxide is locally found within blebs of solidified bitumens (Holtstam & Ensterö 2002, Ensterö 2003). Although originally mined as a copper and iron deposit, about 160 metric tonnes of REE ore produced from Bastnäs was sold over the period 1860–1919 (Carlborg 1923).
The Rödbergsgruvan deposit (59°30'N, 14°53'E) is at the southern end of the zone described above. It consists of a quartz-rich magnetite ore with sulfide impregnations (mainly pyrite), associated with diversi- fied skarn-type assemblages comprising, e.g., tremolite, cummingtonite, almandine, talc, micas and chlorite minerals (Geijer & Magnusson 1944). The REE mineralization has not been described from an in situ occurrence in this case. In the mine dumps, sporadic finds of magnetite–amphibole skarn associated with REE silicates and sulfides have been made (K. Gatedal, pers. commun., 2002).
On the basis of composition and paragenesis of REE minerals, the deposits have been subdivided into two subtypes (Holtstam & Andersson 2002): one almost exclusively enriched in LREE (subtype 1; Bastnäs and Rödbergsgruvan), and another showing enrichment of LREE and HREE + Y (subtype 2; Norberg area). Both are characterized in detail below.
|
METHODS
|
|---|
About forty polished thin sections were prepared from representative specimens of REE-mineral assemblages and examined under a polarizing microscope. Scanning electron microscopy studies were performed on carbon-coated sections using a field-emission instrument (Hitachi S4300), fitted with a back-scattered electron (BSE) detector, and an energy-dispersive X-ray (EDX) micro-analyzer (Oxford Instruments) that was used as an aid in mineral identification (qualitative analyses).
The chemical composition of the REE minerals was determined using wavelength-dispersion analyses on a Cameca SX50 electron microprobe (EMP) at GeoForschungsZentrum in Potsdam. The acceleration voltage was 20 kV, with a beam current of 40 nA and a beam diameter of 3 µm. As standards, we used pure synthetic REE phosphates (REEL
for La, Ce, Yb, Lu; REELβ for the remaining REE), YPO4, (YL), PrPO4 (PK), Fe2O3 (FeK), wollastonite (CaK, SiK), MgO (MgK), Al2O3 (AlK), tugtupite (ClK) and LiF (FK). Each peak was measured during 30 or 50 s (10 and 25 s for the background, respectively). Data reduction was made using a Cameca version of the PAP (Pouchou & Pichoir 1991) routine.
The concentration of fluorine was calculated by empirical correction for the interference of CeL on FK. The elements Mn, Ti, Ba, Sr, Na, Th and U were routinely monitored, but found to be below the limit of detection (
0.05 wt.%) in most REE minerals analyzed. A few complementary analyses were done at the Department of Earth Sciences, Uppsala University, with the same kind of instrument and with similar settings and standards (Ho, Yb, Lu were not measured in these sessions). However, FK was measured using a TAP crystal (PC1 in Potsdam), resulting in a considerably higher level of detection (ca. 0.4 wt.% instead of 0.1 wt.%) and poorer analytical precision. Note that the concentration of Eu cannot be measured with the technique employed here owing to interference with some of the LREE. The chondrite-normalized REE patterns presented in this paper, based on the EMP results, were plotted using the chondrite abundance-values of Boynton (1984). Element ratios in the text (e.g., Y/Ce) are given on a molar (atomic) basis.
X-ray powder-diffraction data for selected samples were recorded with step (0.02° per 2.5 s) scans in the 2
range 3 to 70° on an automated Philips PW1710 diffractometer using graphite-monochromatized CuK radiation (PW1830 generator operated at 40 kV and 40 mA). Peak positions were determined with the X’Pert Graphics & Identify program, and 2 was corrected against an external silicon standard. Unit-cell dimensions were determined using reflections equivalent in number to at least five times the number of refined parameters, and a least-squares program (Novak & Colville 1989).
Mössbauer 57Fe absorption spectra were obtained at room temperature for a few iron-bearing samples that could be purified, following crushing and sieving, by hand-picking under a binocular microscope. A 57Co(Rh) source (nominally 1.8 w x 109 Bq) was used as a source of 
radiation, and velocity calibration was done against a 25 µm thick foil of -Fe. The powdered samples (corresponding to an absorber thickness of maximum 5 mg/cm2), were run at magic-angle geometry (54.7°) using a constant-acceleration spectrometer and stored in a 1024-channel analyzer. Spectral analyses, assuming a "thin" absorber and a Lorentzian line-shape, were carried out using the software developed by Jernberg & Sundqvist (1983). The centroid shift (CS), quadrupole splitting (QS) and line width (
) are parameters used to characterize the absorption doublets.
|
RESULTS: CHARACTERISTICS OF THE REE MINERALS
|
|---|
Fluocerite
Fluocerite, (Ce,La)F3, is known from the Bastnäs deposit (Geijer 1921), where it is a relatively rare constituent of the cerite ore. Fresh grains, normally less than 1 mm across, are irregular in outline and in direct contact with cerite-(Ce) or bastnäsite-(Ce). The mineral is partly altered (Fig. 2A). In cases where alteration is complete, only a very fine-grained, turbid and partly porous material remains. We have identified bastnäsite-(La) and cerianite as the principal products of its breakdown (see below). Ten-point analyses of sample #02+ 0052 gave, on average: Ce2O3 36.51 (range 32.90–38.92), La2O3 39.82 (34.10–48.67), Nd2O3 5.85 (4.77–6.42), Pr2O3 2.18 (1.91–2.45), Sm2O3 0.13 (0.10–0.22), Gd2O3 0.25 (0.08–0.43), SiO2 0.09 (0.00–0.16), CaO 0.05 (0.01–0.10), F 24.76 (21.83–27.95), O = F –10.42, sum 99.22 (all in wt.%). The range of La/(La + Ce) values is 0.49–0.57 (mean 0.52), indicating that the mineral is essentially fluocerite-(La). As the analyses correspond to only 0.80–0.91 mole fractions of the (REE)F3 component, the deficiency in the F contents relative to the ideal formula of fluocerite might possibly be explained by an as yet unproven substitution of OH– or O2– for F–. The unit-cell parameters of the analyzed sample, determined by powder XRD, are a 7.155(6), c 7.290(5) Å, V 323.2(6) Å3 (hexagonal cell).
View larger version (134K):
[in this window]
[in a new window]
|
FIG. 2. BSE images of assemblages of REE minerals from Bastnäs-type deposits. (A) Sample #02+0052 from Bastnäs, with fluocerite (Flc), bastnäsite (Bas), cerite (Cer) and molybdenite (Mlb). Some alteration to cerianite (arrow) has occurred. (B) Sample #A37, Bastnäs, with fluocerite grains broken down completely to bastnäsite-(La) and cerianite (arrow). (C) Sample #430644 from Östanmossa with tremolite (Tr) and fluorbritholite (Flb), the latter mineral partly altered to parisite (Par) and bastnäsite. (D) Sample #540155 from Östanmossa with a grain of intergrown fluorocarbonates. Phases of intermediate grey-scale level and compositions are indicated by ?.
|
|
The breakdown of fluocerite to bastnäsite has been documented from several occurrences (e.g., Styles & Young 1983, Lahti & Suominen 1988, Beukes et al. 1991, Imaoka & Nakashima 1994). This is generally believed to occur by reaction with carbonic aqueous fluids: (Ce,La)F3 + CO2 (aq) + H2O (Ce,La)(CO3) F + 2 HF (aq). The presence of cerianite in association with altered fluocerite has been ascribed to a late oxidation of Ce (e.g., Van Wambeke 1977, Lahti & Suominen 1988). However, we note a decrease of Ce/La in the secondary bastnäsite relative to that of fluocerite in our samples (Fig. 3). A reaction representing the alteration process, which takes the approximate composition of the reactants and products into account, could then be: (La0.5Ce0.5)F3 + 0.83 CO2 + 1.17 H2O 0.83 (La0.6Ce0.4)(CO3)F + 0.17 CeO2 (cerianite) + 2.17 HF + 0.08 H2.
View larger version (17K):
[in this window]
[in a new window]
|
FIG. 3. Chondrite-normalized REE patterns of fluocerite and its alteration products, from the Bastnäs deposit.
|
|
Cerianite
Cerianite, ideally CeO2, is a product of the alteration of fluocerite in the Bastnäs deposit. It exists as microscopic grains less than 20 µm across (Fig. 2B), and the small volumes of sample presented difficulties in the analyses. Concentrations (wt.%) obtained from measurements of two grains in #A37 that produced reasonably high totals are: CeO2 91.70, 92.32, La2O3 1.08, 1.03, Nd2O3 2.75, 1.79, Sm2O3 0.26, 0.22, Gd2O3 0.11, 0.08, SiO2 0.13, 1.14, CaO 0.22, 0.23, F 0.98, 0.84, Cl 0.05, 0.06, O = F,Cl –0.42, –0.36, sum 97.86, 97.35, which yield an average formula Ce0.934+Nd0.02La0.01 Si0.02Ca0.01O1.92F0.08 on the basis of one cation. We believe that the significant contents of F obtained are real; the monovalent anion balances the minor amounts of REE3+ and Ca2+. Note that CeO2 and the oxyfluoride CeOF are isostructural compounds, with a fluorite-type atomic arrangement (Strunz & Nickel 2001); thus limited solid-solution is not unexpected. The refined unit-cell parameter of cerianite in this sample, a 5.411(3) Å, is identical to the literature value for the pure synthetic dioxide, 5.41134(12) Å (PDF 34–394).
Håleniusite-(La)
Håleniusite-(La), (La, Ce)OF, is a secondary mineral formed by alteration of primary bastnäsite-(La), most likely via a decarbonation reaction, (La, Ce)CO3F (La, Ce)OF + CO2 (Holtstam et al. 2004). It is quite common in the Bastnäs deposit; considering the variations in La/Ce encountered in bastnäsite (see below), it is expected that a Ce-dominant analogue exists there as well.
Parisite-(Ce)
Parisite-(Ce), CaCe2(CO3)3F2, has been identified in a few samples from the Östamossa mine (subtype 2). In sample #430644, this fluorocarbonate occurs in large grains of "fluorbritholite-(Y)", where it partly replaces the host mineral along edges and fractures (Fig. 2C). Individual grains of parisite-(Ce) may reach 200 µm across, and occasionally the mineral occurs in parallel intergrowth lamellae (20–40 µm wide) with an unknown fluorocarbonate with lower Ca and F contents. Locally, parisite-(Ce) also is found in direct contact with bastnäsite-(Ce). In #540155, aggregates (0.2–1 mm) of fluorocarbonates are distributed in a matrix of tremolite, calcite, dolomite, dollaseite-(Ce) and magnetite. From the results of the EMP analyses (Table 1), we conclude that the patchily to regularly intergrown phases (Fig. 2D) consist of parisite-(Ce), bastnäsite-(Ce) and an intermediate (on the gray scale of BSE images), unnamed mineral. Tremolite is euhedral toward the fluorocarbonate minerals, which occasionally also incorporate small crystals of an amphibole and magnetite.
Parisite-(Ce) differs from most samples of bastnäsite, the dominant fluorocarbonate at these deposits, in its high concentration of Nd (Nd > La for most points analyzed) and Y (0.8–5.6% Y2O3). Refinement of powder XRD data for #430644 indexed on a hexagonal unit-cell gave a 7.099(2), c 28.12(2) Å (representing a pseudocell of a monoclinic polytype, or the true unit-cell of a hexagonal variety; cf. Ni et al. 2000).
Bastnäsite-(Ce) – bastnäsite-(La)
Bastnäsite, (Ce,La)CO3F, was first described from the type locality by Hisinger (1838). It is a ubiquitous although subordinate component of the cerite ore, where it occurs normally as small grains ( 0.2 mm) interstitial to the REE-bearing silicates. In all sections of this material, one can observe that cerite-(Ce), and törnebohmite-(Ce) to some extent, are replaced by bastnäsite as fine grains. Coarse-grained homogeneous masses of bastnäsite are less common in the Bastnäs deposit. In sample #882234, the mineral fills cavities in a mass of ferriallanite-(Ce), which is distinctly euhedral toward the bastnäsite. Rare, minute grains of cerite-(Ce) are found in bastnäsite. Molybdenite and quartz occupy minor parts of the cavities, and euhedral crystals of bastnäsite are occasionally seen enclosed in the latter mineral.
In the Norberg District, bastnäsite is generally less common and, in some instances, it is found mainly as a microscopic phase in association with fluorbritholite-(Ce) or cerite-(Ce). However, in S. Hackspikgruvan, here represented by sample #381132, elongate, slightly bent crystals of bastnäsite-(Ce) up to 20 mm in length appear in fibrous amphibole skarn. Small, subhedral grains of dollaseite-(Ce), fluorbritholite-(Ce) and fluorian phlogopite occur sporadically along the contact between amphibole and the bastnäsite-(Ce). The sample is rich in fluorite, as selvages around bastnäsite, as fracture infillings and inclusions in the latter mineral, and as significant impregnations in the amphibole mass (Fig. 4A).
View larger version (136K):
[in this window]
[in a new window]
|
FIG. 4. BSE images of REE mineral assemblages from Bastnäs-type deposits. (A) Sample #381132 from S. Hackspikgruvan with bastnäsite (Bas), tremolite (Tr), fluorite (Fl), magnetite (Mgt) and västmanlandite (arrow). (B) Sample #390477, Bastnäs, with a heterogeneous grain of cerite-(Ce) enclosed by bismuthinite. (C) Sample #03+0029 from Malmkärra with fluorbritholite (Flb) and cerite (Cer). (D) Sample #UU318/77, Malmkärra, with unnamed mineral E partly altered to bastnäsite, and västmanlandite (Väs). The arrow points to a small grain of scheelite.
|
|
Lanthanum, Ce and Nd are the dominant REE in all samples of bastnäsite, amounting to 94–98 atom % for subtype 1 and 89–97 atom % for subtype 2 (Table 2, Fig. 5). The Nd contents are in the range of 0.05–0.14 atoms per formula unit (apfu); two samples from Östanmossa, coexisting with Nd-rich parisite-(Ce), are exceptional in having >0.20 apfu Nd. The ratio Ce/(Ce + La) lies in the range 0.43–0.67, corresponding to an intermediate part of the bastnäsite-(La)–bastnäsite-(Ce) solid-solution series. The compositional heterogeneity can also be significant within a sample. A total of 36 point analyses were collected along a 2-mm traverse on the coarse-grained material in #882234; Ce/(Ce + La) ranges from 0.472 to 0.577, with a mean of 0.496 ± 0.003. The parameters for this sample are a 7.142(2) and c 9.788(2) Å (hexagonal cell). Ni et al.(1993) gave a 7.118, c 9.762 Å for a bastnäsite-(Ce) specimen slightly poorer in La. For most samples, the F content does not significantly deviate from the stoichiometric value of 1 apfu, considering the analytical uncertainty. A few analyses, however, suggest a component of hydroxylbastnäsite, (Ce,La)CO3(OH), in solid solution with bastnäsite (samples ##060375, 970319).
Lanthanite-(Ce)
Lanthanite-(Ce), Ce2(CO3)3·8H2O, is a secondary mineral formed at low temperatures (<100°C), and appears locally as a coating on cerite–ferriallanite ore. Crystals are lath-like to platy in habit and normally less than 1 mm across. The accredited type-locality for lanthanite-(Ce) is the Britannia mine in North Wales (Bevins et al. 1985). However, lanthanite-(Ce) from Bastnäs was first described and analyzed by Berzelius (1825). The crystal structure of the mineral was determined on material from the original type-locality (Dal Negro et al. 1977). Atencio et al.(1989) determined the composition of a single sample from Bastnäs to show that it is clearly Ce-dominant. This is supported by EDX analyses performed by us on four specimens (#g10778, g30592, 23:0475, 221690), that consistently have Ce > La 
Nd > Pr > Sm. A more complete chemical analysis of this mineral with the EMP proved to be difficult due to violent degassing under the electron beam.
Cerite-(Ce)
Cerite-(Ce), (Ce,La,Ca)9(Mg,Fe)(SiO4)6(SiO3OH) (OH,F)3, is locally rock-forming at Bastnäs. It occurs in medium-grained (0.2–1 mm) masses; individual crystals are normally anhedral. Replacement veins of ferriallanite-(Ce) ± törnebohmite-(Ce) commonly transect the cerite ore, in many cases to such an extent that cm-sized "islands" of cerite-(Ce) are completely surrounded by these minerals. Quartz occurs sporadically as an interstitial phase. Small grains of sulfide (chalcopyrite, molybdenite) are occasionally found dispersed in the cerite-(Ce) masses. Rarely, e.g., in sample #390477, subhedral crystals of cerite-(Ce) are found embedded in massive bismuthinite (Fig. 4B) associated with subordinate chalcopyrite and molybdenite. Locally (in S:t Göransgruvan), here represented by sample #LK4838, cerite-(Ce) forms layers up to 5 cm thick sandwiched between fibrous masses of actinolite, with the amphibole crystals oriented perpendicular to the cerite-(Ce) layers. Ferriallanite-(Ce), törnebohmite-(Ce) and talc have developed extensively at the layer boundaries.
At Rödbergsgruvan (#880072), cerite-(Ce) constitutes deformed lenses up to 2 cm thick, within coarse-grained aggregates of a calcic amphibole (actinolite–edenite in composition); the two minerals are separated by a zone 0.5–1 mm wide of mainly ferriallanite-(Ce) + the Fe analogue of västmanlandite-(Ce). Bastnäsite-(Ce), amphibole and ferriallanite-(Ce) also occur within the cerite-(Ce) mass.
In the subtype-2 deposits of the Norberg area, cerite-(Ce) is subordinate in relation to fluorbritholite-(Ce), which has a similar megascopic appearance. Sample #660298 from Malmkärra is a lump of REE silicates with amphibole as a minor constituent. A few veinlets of pyrite up to 2 mm thick transect the specimen. Cerite-(Ce) forms aggregates of anhedral grains (with an average size of ca. 0.5 mm), in contact with subhedral grains of västmanlandite-(Ce) up to 1 mm across, and minor amounts of anhedral bastnäsite-(Ce). Chalcopyrite is found as small grains interstitial to the REE minerals. In #490216 from Johannagruvan, cerite-(Ce) and bastnäsite-(La) occur as patches up to 10 mm across surrounded by a thick margin of dollaseite-(Ce), in a matrix of dense amphibole skarn. Bastnäsite-(Ce) forms larger, coherent areas, whereas cerite-(Ce) occurs as fine-grained aggregates with an irregular outline. Subhedral grains of västmanlandite up to 0.5 mm wide occur close to the contact between dollaseite-(Ce) and cerite-(Ce). In #03+0029 from Malmkärra, cerite-(Ce) occurs in close association with fluorbritholite-(Ce), where it locally has replaced the latter mineral (Fig. 4C). This material in part displays a sector-like pattern of zonation on BSE images, related to variations in Ca and REE concentrations between the zones.
The crystal structure of cerite-(Ce) was solved on material from Mountain Pass, California (Moore & Shen 1983). There are three non-equivalent positions in the structure, occupied mainly by REE, each coordinated to (8 O + 1 OH). The REE polyhedra are connected to form rods, to which isolated SiO4 tetrahedra are attached. A second type of parallel rods consists of SiO3OH tetrahedra and (Mg,Fe)O6 octahedra. Calcium enters the REE sites plus a six-coordinated site with low occupancy, Ca(x), and possibly also the M sites. On the basis of bond-valence calculations, Pakhomovsky et al.(2002) suggested a modified structural formula with a higher number of OH groups for cerite-(La), (La,Ce,Ca)9(Fe,Ca,Mg)(SiO4)3[SiO3(OH)]4(OH)3. Cerite is thus structurally complex, and shows a significant compositional variability.
Compared to other occurrences worldwide (Förster 2000, and references therein), the present material is poor in Ca and enriched in Mg and F (Tables 3a, b). The sums (Ca + REE + Y) are in the range 8.94–9.18 (mean 9.05) apfu. However, the Ca content is normally below 1 apfu (mostly 0.4–0.9) and negatively correlated with REE + Y (Fig. 6), indicating that the fraction of Ca atoms entering the M or the Ca(x) sites is minor in our samples. Iron ranges from 0.01 to 0.30 apfu, with lower values associated with specimens from the subtype-2 localities. Mössbauer spectra obtained on two samples from Bastnäs (#060375, #A37) reveal that Fe is dominantly in the trivalent state in cerite-(Ce), corresponding to ca. 80% of the total amount. The CS values, 0.33–0.35 mm/s, are consistent with Fe3+ ions being located at an octahedral site. The hyperfine parameters for the minor Fe2+ fraction, CS and QS both in the range 1.2–1.3 mm/s, suggest a coordination number of 6 or higher. There is a clear negative correlation between Mg and Fe for the present sample-population. Possible exchange-mechanisms to explain this trend would be Fe3+ + Ca2+ = Mg2+ + REE3+ or Fe3+ + O2– = Mg2+ + F–, but so far they lack support from direct structural evidence.
View larger version (17K):
[in this window]
[in a new window]
|
FIG. 6. Compositional diagram showing the variation of Ca versus (REE + Y) in cerite-(Ce). The straight line corresponds to an ideal REE+ Y + Ca = 9 apfu.
|
|
Generally, cerite-(Ce) is homogeneous with respect to the concentration of SiO2, corresponding to a narrow range with 6.83 to 7.07 Si apfu. In one sample (#390477), there are, however, rare grains with irregular patches (dark on BSE images; Fig. 4B) that show large fluctuations in (SiO2 17–24 wt.%) and low totals. Cerite-(Ce) is a major carrier of Y, with up to 2.1 and 3.5 wt.% Y2O3 at the subtype-1 and subtype-2 deposits, respectively. The chondrite-normalized curves for the mineral are quite distinct from those of associated epidote-group minerals and fluorocarbonates (Fig. 7); the cerite curves have an upward convex shape related to a relative enrichment of Nd, Sm and Gd, in particular, compared with the coexisting REE minerals. Cerite-(Ce) shows significant and variable F contents, corresponding to 1.18–1.72 apfu; nothing is known at present, however, about the structural role of the anion. Note that most REE-bearing silicates at the Bastnäs-type deposit contain no detectable Cl; however, low concentrations of this element, 0.04 to 0.09 wt.%, are found in cerite-(Ce) (the estimated detection-level at 3
is ca. 0.02 wt.% for Cl).
View larger version (19K):
[in this window]
[in a new window]
|
FIG. 7. Chondrite-normalized patterns for coexisting REE minerals from Bastnäs (a selected typical sample).
|
|
The hexagonal unit-cell for two samples yield the parameters a 10.737(6), c 37.81(2) Å, V 3775 Å3 (#060375) and a 10.763(8), c 37.89(3) Å, V 3802 Å3 (#03+0029), respectively. Moore & Shen (1983) gave a 10.779, c 38.06 Å, V 3830 Å3 for cerite-(Ce) from Mountain Pass, and Pakhomovsky et al.(2002) reported a 10.749, c 38.318 Å, V 3834 Å3 for cerite-(La) from Khibina, Kola Peninsula. The larger unit-cell volume in the latter case, as compared to the Bastnäs-type material, might be explained by significant difference in Ce/La between cerite from the two localities.
On the basis of an examination of the relatively small amount of modern chemical data (EMP) that presently exist, it seems plausible that the term "cerite" embraces more than one distinct species, if variations in the M-type cations were to be considered in the nomenclature. All samples analyzed in the present study have Mg > Fe [0.69 < Mg/(Mg + Fe) < 0.99], whereas Al is generally at or below the limit of detection. The material from the Niederbobritzsch granite (Erzgebirge, Germany) analyzed by Förster (2000) contains significant Al (up to 1.04 apfu), little Fe and no Mg. Chakhmouradian & Zaitsev (2002) reported another Mg-free variety, from the Afrikanda Complex, Kola Peninsula, displaying some variations in Fe and Al [0.50 < Fe/(Fe + Al) < 0.76]. In the type specimen of cerite-(La), Fe is the dominant cation at the M site, which is also inferred to contain significant Ca (0.30 apfu; Pakhomovsky et al. 2002). Note that all examples of cerite referred to here have elevated concentrations of Ca, with a range from 4.42 wt.% (Khibina) to 9.33 wt.% CaO (Afrikanda), compared to our material (maximum 3.0 wt.% CaO). Clearly, new crystal-structure refinements of cerite from several localities, combined with chemical analyses, are required to fully reveal its crystal-chemical character and variability.
Unnamed mineral E
A chlorine-bearing REE silicate was found in sample #UU318/77 from the Malmkärra mine. It occurs as a greyish pink, cerite-like mass with a greasy luster, in association with mainly västmanlandite-(Ce), talc and phlogopite. Alteration to bastnäsite-(Ce) is pervasive in certain areas (Fig. 4D). Fractured grains of magnetite ( 1 mm), equant to skeletal in outline, occur in contact with both västmanlandite-(Ce) and the unnamed mineral. It is colorless, transparent and nonpleochroic in thin section. The optical character is biaxial (–), with 2V 55°. Individual grains (100–300 µm) are subhedral and display slight undulatory extinction. A small inclusion of fluorbritholite-(Ce) was detected in one of the grains, and minute grains of scheelite are scattered all over the specimen.
A preliminary single-crystal diffraction study indicates that mineral E is monoclinic, with the unit-cell parameters a 14.14, b 10.74, c 15.51 Å, β 106.6°. No precise structural formula can be established at present, but the chemical composition (Table 4), with relatively high and constant Cl contents (2.85–2.97 wt.%), and a unique X-ray powder-diffraction pattern, indicate that this mineral represents a new species (no Ce silicate with essential Cl is presently known to the scientific community; Mandarino & Back 2004).
The mineral has Ce > Nd > La, i.e., the same order as encountered in fluorbritholite-(Ce) and cerite-(Ce) from the Malmkärra mine. In its overall chemical composition, it is most similar to cerite-(Ce), albeit clearly distinguished by its Cl content. The low totals (94.9–96.2 wt.%) clearly suggest that unanalyzed volatile components are contained in the mineral. A single analyzed grain (<30 µm across) in contact with cerite-(Ce) in a Bastnäs sample (#A37) has a composition similar to that of mineral E, except that it is has Fe > Mg (Table 4).
Minerals of the fluorbritholite group
Fluorbritholite, ideally REE3Ca2[SiO4]3F, is only found in the subtype-2 deposits and was not known to exist there before our studies (Holtstam & Andersson 2002); it is now clear that some occurrences of "cerite" reported earlier (Geijer 1927) are in fact fluorbritholite. Fluorbritholite is isostructural with fluorapatite, and the type material is reported to be hexagonal, P63/m (Gu et al. 1994). Oberti et al.(2001) structurally characterized two samples that were best described in the lower symmetry P63, and monoclinic–pseudohexagonal varities may also exist (e.g., Noe et al. 1993).
The major occurrences of fluorbritholite-(Ce) are in the Malmkärra deposit, where large masses weighing several kilograms each have been collected. In these samples, fluorbritholite-(Ce) is associated with dollaseite-(Ce), västmanlandite-(Ce), gadolinite, amphibole, dolomite, phlogopite, magnetite and pyrite. A general impression from the textures of this kind of material (e.g., ##970319 and 540027) is that fluorbritholite-(Ce) is the primary, earliest mineral to crystallize; other REE minerals, at least to some extent, seem to have formed during breakdown of the fluorbritholite-(Ce).
In #03+0029, fluorbritholite-(Ce) forms irregular masses up to 2 cm across in a skarn rock of coarse actinolite and magnetite. It is peculiar to this sample that cerite-(Ce) has partly replaced fluorbritholite-(Ce).
An Y-dominant analogue (Table 5) of fluorbritholite-(Ce) occurs in a sample (#430644) from the Östanmossa deposit. Aggregates up to 3 cm across coexist with some tremolite in a coarsely crystalline vein of dolomite transecting dollaseite-bearing amphibole skarn. In #381132 from S. Hackspikgruvan, fluorbritholite-(Ce) is a subordinate component associated with bastnäsite-(Ce) and fluorite.
Fluorbritholite is low in Ca, P and the actinides compared to most other known ocurrences (e.g., Arden & Halden 1999, Della Ventura et al. 1999, Smith et al. 2002). Significant inter- and intrasample variations in REE and Y, corresponding to a range from fluorbritholite-(Ce) to the previously unrecognized members "fluorbritholite-(Nd)" and "fluorbritholite-(Y)", occur. Consequently, fluorbritholite samples display large variations in the chondrite-normalized patterns (Fig. 8). The more Y-rich samples (Y > 0.5 apfu) are significantly enriched in HREE, Gd–Lu (0.20–0.40 apfu).
View larger version (19K):
[in this window]
[in a new window]
|
FIG. 8. Selected chondrite-normalized REE patterns for fluorbritholite samples from the Norberg District.
|
|
Fluorbritholite is consistently nonstoichiometric in that Ca < 2 and (REE + Y) > 3 apfu (Fig. 9). This finding could theoretically be explained by one of the following mechanisms of substitution, proposed to operate in apatite-type structures (e.g., Pan & Fleet 2002):
 | (1) |
 | (2) |
 | (3) |
 | (4) |
As there are no monovalent cations (e.g., Na+) present, (1) can be ruled out. Boron, which is an easily overlooked component, was not measured in our samples; B3+ partly replacing Si4+ at the tetrahedrally coordinated sites has in fact been demonstrated in a sample of fluorbritholite (Oberti et al. 2001). However, if (2) was the mechanism responsible, (REE + Y) and Si should be anticorrelated, which is not the case here. With a few significant exceptions, the deviation from the ideal value of 3 Si apfu is explained by a P5+-for-Si4+ exchange in our samples (Fig. 10). The observed variations in the F content (0.63–1.02 apfu), which could correspond to (3), are uncorrelated with the REE, and are probably simply an expression of a common OH–F substitution in the present material. The calculated formulae, if based on a fixed number of anions (12.5 O), give cation sums below 8 apfu for most points, in support of the last mechanism (4). However, there is no physical evidence for cation vacancies (
) in fluorbritholite; further studies clearly are needed to resolve the question of nonstoichiometry in fluorbritholite. Compositionally similar samples of fluorbritholite-(Ce), i.e., with Ca < 2 and REE > 3 apfu, were recently described from nepheline syenite in the Pilansberg alkaline complex, South Africa (Liferovich & Mitchell 2006).
The unit-cell parameters of "fluorbritholite-(Y)" in #430664 was determined to a 9.561(1), c 6.950(1) Å, V 550.2(2) Å3. For two specimens of fluorbritholite-(Ce) (#97319, #03+0029), we found a 9.617(2), 9.620(2), c 7.010(2), 7.033(2) Å and V 561.5(3), 563.7(3) Å3.
Törnebohmite-(Ce)
Törnebohmite-(Ce), (Ce, La)2Al[SiO4]2(OH), was first discovered at Bastnäs and carefully described by Geijer (1921). Using plesiotype material, Shen & Moore (1982) solved the crystal structure of this monoclinic mineral. It consists of straight chains of edge-sharing AlO4(OH)2 octahedra running along b, linked by isolated [SiO4] tetrahedra. The REE cations occupy cavities with 10-fold coordination (9 O + 1 OH) between the chains.
Törnebohmite-(Ce) belongs to the cerite-(Ce) – ferriallanite-(Ce) association (i.e., the major REE ore) in the Bastnäs deposit (Figs. 11A,B). The mineral grains, recognized in thin section by their characteristic pleochroism (Y bluish green), are anhedral, typically slightly elongate, and 0.05 to 1 mm in their longest dimension. Crystal aggregates of törnebohmite-(Ce) may exceptionally reach 5 mm across. The mineral is most closely associated with ferriallanite-(Ce), with which it commonly is in direct contact or occasionally intergrown. In a few cases, lamellae of törnebohmite-(Ce), ca. 20 µm or wider, have been found regularly arranged in a host crystal of ferriallanite-(Ce). In a similar fashion, törnebohmite-(Ce) may contain oriented lamellae of ferriallanite-(Ce), although irregular inclusions of this mineral are more common. There is a structural explanation for this kind of epitaxy; see the paragraph on västmanlandite-(Ce) below. Törnebohmite-(Ce) locally occurs without any particular spatial relation to ferriallanite-(Ce); it is then surrounded by cerite-(Ce). At rare contacts with patches of bastnäsite, the crystals tend to be subhedral in shape. Törnebohmite is said to have been found in relatively large quantities at S. Hackspikgruvan (Geijer 1936), but such material was not available to us.
View larger version (147K):
[in this window]
[in a new window]
|
FIG. 11. BSE images of REE mineral assemblages from Bastnäs-type deposits. (A) Sample #A37, Bastnäs, with cerite (Cer), ferriallanite (Fln), törnebohmite (Tbm), bastnäsite (Bas) and molybdenite (Mlb). Quartz (Qtz) fills a small cavity. (B) Sample 03+0246, Bastnäs, with gadolinite (Gad), cerite, ferriallanite and törnebohmite. (C) Sample #970319, Malmkärra, with gadolinite, fluorbritholite (Flb) and dollaseite (Dla). (D) Sample #660298, Malmkärra, with västmanlandite (Väs), gadolinite, tremolite, cerite and magnetite (arrow).
|
|
The chemical variation within the investigated population of samples is moderate (Table 6). A clear negative correlation between Fe and Al was found (R = –0.92 for n = 11), indicating that Fe substitutes for Al at the octahedral sites (and indirectly that Fe essentially is in the trivalent state in the mineral). Small amounts of Mg (0.02–0.04 apfu) are probably present in these positions as well. Values of F concentration are close to the detection limit in all samples, proving that F–OH substitution is negligible.
There is no appreciable Ca replacing (Ce, La) in the structure, in contrast to what is found for the associated REE silicates. Törnebohmite-(Ce) also is exceptional in that the concentration of Y2O3 is below detection for all samples. It is considerably enriched in La compared to the coexisting cerite-(Ce) in #A37 (Fig. 7).
The unit-cell parameters for #A37 are a 7.438(4), b 5.690(4), c 17.02(1) Å, β 111.96(8)°. Shen & Moore (1982) reported a 7.383(4), b 5.673(3), c 16.937(6) Å, β 112.04(8)° from single-crystal data; no corresponding chemical analysis was carried out, so a direct comparison is not possible.
Törnebohmite is a rare mineral globally, and few published analytical data are available for reference. According to Svyazhin (1962), törnebohmite-(La) occurs in a granite pegmatite in a fenitization zone of a nepheline syenite massif in the Urals. Wet-chemical analysis showed SiO2 20.33, TiO2 0.12, REE 62.88, Fe2O3 2.61, Al2O3 10.11, MgO 0.92, CaO 2.46, H2O+ 0.78 (all in wt.%). These data yield a poor formula, and the data may be inferior owing to impurities. Kapustin (1989) reported a fluorian variety of törnebohmite-(Ce) from Ulan-Erge, Tuva alkaline massif, with 0.7 wt.% F. That material also contains appreciable Ti (0.44 wt.% TiO2), Th (0.39 wt.% ThO2), Ca (0.48 wt.% CaO) and Sr (0.36 wt.% SrO).
Minerals of the gadolinite group
Gadolinite, (REE,Y)2FeBe2[Si2O8]O2, is new to the Bastnäs-type deposits and the only Be mineral known there. It appears to be very rare at the subtype-1 deposits, and more widespread in the Norberg District, although mostly in microscopic amounts. Gadolinite-(Ce) was found in a single sample from Bastnäs (#03+246), in which it occurs as irregular grains up to 0.8 mm across (Fig. 11B), in contact with törnebohmite-(Ce), ferriallanite-(Ce) and cerite-(Ce). The specimen also contains the Fe-analogue of västmanlandite-(Ce). The unit-cell parameters of gadolinite-(Ce), determined by single-crystal diffractometry, are ca. a 4.82, b 7.71, c 10.10 Å, β 90.12°, V 375 Å3.
In the subtype-2 localities, gadolinite is not restricted to a specific type of association, and a short description of each sample seems appropriate. Generally, the mineral is greenish yellow to green in thin section, without any appreciable pleochroism. In sample #970319 from Malmkärra, fine-grained masses of gadolinite-(Ce) – gadolinite-(Y), up to 2 cm wide, occur in an assemblage of mainly fluorbritholite-(Ce), västmanlandite-(Ce) and dollaseite-(Ce) (Fig. 11C). Textural features suggest that the gadolinite formed at the expense of fluorbritholite-(Ce). Sample #540033 (Malmkärra) is a specimen of tremolite–magnetite skarn with significant dissakisite-(Ce) and fluorbritholite-(Ce). "Gadolinite-(Nd)" occurs as anhedral grains up to 100 µm in their greatest dimension, enclosed by dissakisite-(Ce). In sample #660298 (Malmkärra), anhedral grains (300 µm) of "gadolinite-(Nd)" occur in aggregates of cerite-(Ce), and is locally in contact with both västmanlandite-(Ce) and tremolite (Fig. 11D).
Sample #02+0126 from the Östanmossa deposit is a tremolite – magnetite – phlogopite skarn with scattered grains of molybdenite and pyrite. Solitary, subhedral crystals of gadolinite-(Y) (<1 mm) occur in a phlogopite–chlorite mass (Fig. 12A), locally in contact with magnetite. No other REE minerals have been observed in this thin section. Sample #520767 from Johannagruvan contains highly irregular grains (<200 µm) of gadolinite-(Y), partly intergrown with dollaseite-(Ce), within magnetite-bearing tremolite skarn (Fig. 12B). A vein of norbergite, 1–3 mm wide, transects the rock just a few mm from the portion with gadolinite-(Y).
View larger version (132K):
[in this window]
[in a new window]
|
FIG. 12. BSE images of REE mineral assemblages from Bastnäs-type deposits. (A) Part of chemically zoned crystal of gadolinite in sample #02+0126 from Östanmossa. Matrix consists of a chlorite mineral. Numbers indicate the amount of Y atoms per formula units, obtained from point analysis. (B) Sample #520767, Johannagruvan, with gadolinite (Gad), tremolite (Tr), dollaseite (Dla) and fluorapatite (arrow). (C) Sample #540025, S. Hackspikgruvan, with molybdenite (Mlb), dissakisite (Dis) and fluorite (Fl). (D) Sample #880072 from Rödbergsgruvan with bastnäsite (Bas), amphibole (Amp), ferriallanite (Fln) and the Fe-rich analogue of västmanlandite (A2).
|
|
Cerium, Nd and Y are dominant in the gadolinite samples, corresponding to 70–75% of the trivalent cations present, and the compositional variations with respect to these elements are significant, probably more extensive than for any other REE mineral studied here (Table 7). Grains (in some cases only minor portions of them) corresponding to gadolinite-(Ce), gadolinite-(Y) and the Nd-dominant member ("gadolinite-Nd") have been identified (Fig. 13). Single crystals commonly show irregular growth-induced zoning in BSE mode (Fig. 12A), essentially related to variations in Ce and Y. Analytical results along a traverse (step size ca. 50 µm) from the periphery toward the center of an aggregate (#97319) show no correlation between composition and lateral displacement, but a pattern with clear antipathetic behavior of Ce versus Y is evident (Fig. 14). Gadolinite-(Ce) is the most Y-rich phase found at Bastnäs, with Y/Ce = 0.25; the coexisting cerite-(Ce) has Y/Ce = 0.06. A few points in the analyzed grain (#03+246) actually have Nd > Ce, but on average, the phase is Ce-dominant. For gadolinite-(Y) in sample #520767 (Johannagruvan), the Y/Ce value attains 3.40, higher than in any other mineral from these deposits.
View larger version (18K):
[in this window]
[in a new window]
|
FIG. 14. Compositional profile of heterogeneous gadolinite (sample #970319 from Malmkärra). Step size is ca. 6 µm.
|
|
The gadolinite samples have, appropriately enough, the highest overall concentration of Gd (3.3–7.7 wt.% Gd2O3), although there is some overlap with the range for fluorbritholite (0.6–5.6 wt.% Gd2O3), a mineral that displays a similar range of Gd/Y values (0.3–0.9).
As there is an apparent lack of detectable Al in the samples, and as gadolinite is not known to incorporate lighter elements in its tetrahedral sites (e.g., Demartin et al. 1993), the chemical formulae (Table 7) are calculated on the basis of 2 Si apfu. This approach reveals that all analyzed samples have a significant deficiency in Fe, with 0.72 apfu on average, compared to the stoichiometric value (1.00). The phenomenon is most probably related to a hingganite-type substitution (Fe2+ + 2 O2– = + 2 OH–), commonly encountered in gadolinite-group minerals (e.g., Demartin et al. 2001), and in the present suite of samples, amounting to a molar fraction of (REE, Y)2Be2[Si2O8](OH)2 in the range 0.10–0.35. The presence of some Mg (mean 0.07, maximum value 0.19 apfu) is clearly insufficient to compensate for the shortage in Fe. A positive (r = 0.70) correlation between Mg and Y, suggesting a substitution toward Y2MgBe2[Si2O8]O2, a previously unrecognized "end-member" of the gadolinite group, is noteworthy. The fact that F is not detected in gadolinite, despite the F-rich bulk composition of these rocks, indicates that F–OH substitution does not occur in the hingganite component of these minerals.
Mössbauer spectroscopy demonstrates that Fe is exclusively divalent in a sample (#970319) of gadolinite (Fig. 15). The shape of the spectrum shows a close resemblance to that of a recrystallized (annealed), originally metamict, sample of gadolinite-(Y) from Ytterby, as presented by Malczewski & Janeczek (2002). The CS values (1.04 mm/s) are identical, and in both cases, there is a distribution in the QS values related to variations of nearest-neighbor atom around the six-fold site occupied by Fe in the structure.
The samples are poor in Ca (mean 0.18 wt.%, range 0.06–0.58 wt.% CaO), indicating that the components datolite, Ca2B2[Si2O8](OH)2, and homilite, Ca2FeB2[Si2O8]O2, are very minor, in contrast to many specimens of gadolinite originating from alpine fissures and granitic pegmatites (Demartin et al. 1993, 2001, Pezzotta et al. 1999). This conclusion further justifies the calculation of the BeO contents (Table 7) on the stoichiometric basis of 2 Be apfu (i.e., B-for-Be substitution is assumed to be negligible). The formula calculations give the sums (REE + Y + Ca) 1.90 apfu in total. The deviation from ideality (2 apfu) might be due to errors in the analyses, or it might reflect a partial vacancy in the crystal structure. Note that there is no sign of metamictization in the analyzed grains, in agreement with the low concentrations of U and Th.
The unit-cell parameters for gadolinite-(Ce) – gadolinite-(Y) in #97319 were determined to be a 4.793(4), b 7.674(8), c 10.02(1) Å, β 90.76(9)°. The smaller unit-cell volume of this sample compared to #03+246 (368 Å3 versus 375 Å3) is consistent with its higher average Y/Ce value. Non-metamict varities of gadolinite are relatively rare, but note that Segalstad & Larsen (1978) reported a Vcell of 366 Å3 for a sample of gadolinite-(Ce) that contains a significant proportion of the datolite component. For gadolinite-(Y) with near-end-member composition, Vcell is equal to 652 Å3 (Demartin et al. 1993).
Percleveite-(Ce)
The rare species percleveite-(Ce), ideally Ce2[Si2O7], was recently described from the Bastnäs deposit by Holtstam et al. (2003a), and only little needs to be added to their description. It occurs closely associated with mainly cerite-(Ce), in an apparent equilibrium assemblage (sample #060375). Some bastnäsite-(Ce) has formed at the expense of both minerals, and quartz occurs in interstices and microcracks in the assemblage. Percleveite-(Ce) is enriched in Y (Y/Ce = 0.14) compared to coexisting cerite-(Ce) (Y/Ce = 0.06).
Unnamed mineral D
An unknown mineral, tentatively assigned the formula (Y,Ce,Nd)4MgSi4O14F2, was found in sample #970319 from Malmkärra. It occurs as a single grain of irregular outline, ca. 1.2 mm across, in contact with fluorbritholite-(Ce). The crystal is colorless, nonpleochroic and biaxial, with a large 2V. Owing to the specific orientation of the grain and its uniform extinction, the optical sign could not be determined with certainty. On the basis of the chemical data (Table 4), it seems to represent the Mg-analogue of rowlandite-(Y), nominally Y4Fe[Si2O7]2F2, a triclinic mineral (Sipovalov & Stepanov 1971) normally found in granitic pegmatite environments. The analytical totals are reasonable for an anhydrous mineral, and a charge-balance calculation for the average composition, based on nine cations, gives 29.98 positive charges, in excellent agreement with the suggested formula. Further study is required for a full characterization of this apparently new species. It is one of the few phases with clear Y-dominance found at these deposits. The ratio Y/Ce is 1.89–1.99, higher than in any other mineral in the sample [where fluorbritholite-(Ce) has 0.35 < Y/Ce < 0.58 and gadolinite-(Ce) – gadolinite-(Y) has 0.66 < Y/Ce < 1.34]. Note also the unique enrichment in the HREE (for these deposits), with >1 wt.% Yb2O3 and detectable amounts of Lu in the unnamed mineral. The result is a nearly complete and unusually flat chondrite-normalized REE curve (Fig. 16).
Unnamed mineral C
A tungsten-bearing REE silicate mineral occurs in a Bastnäs sample (#060375) as rare anhedral grains up to 300 µm wide. They are optically biaxial and strongly pleochroic (rust red to nearly black). Scheelite appears as a filling of microfractures in the grains. From textural relations, it is clear that the mineral formed at the expense of associated cerite-(Ce) and percleveite-(Ce). The REE array (Table 4) is similar for the three minerals (Ce > La > Nd > Pr > Sm), but mineral C is significantly lower in Y (Y/Ce = 0.01). It is clearly exotic in combining REE, Si and W as major components. On the whole, members of the silicate class with structural W are very few (e.g., welinite, the khomyakovite series), and the eudialyte-group member johnsenite-(Ce) is the only REE–W-bearing silicate mineral hitherto approved (Grice & Gault 2006). The exotic composition of mineral C and its association with percleveite-(Ce) point to unusual conditions of formation for this mineral.
The valence state of Fe and the possible presence of hydroxide in the mineral remain open questions. A tentative empirical formula, compatible with the analytical data at hand, would be (Ce,La,Nd,Ca)5 Mg(Fe,Al)3WSi5O26 on an anhydrous basis. The oxide sums are in the range 97.5–99.3 wt.%, assuming only Fe3+.
Dollaseite-(Ce) – dissakisite-(Ce)
Dollaseite-(Ce), ideally CeCaMg2Al[SiO4][Si2O7] F(OH), is an important mineral for the sequestration of REE in the type-2 deposits of the Norberg District, but not known from any locality outside this area. It was originally described from Östanmossa and named "magnesium orthite" by Geijer (1927). Peacor & Dunn (1988) renamed it dollaseite-(Ce), and demonstrated its close structural relationship with the allanite subgroup; it is chemically distinguished from the other members by a charge-coupled substitution involving both cations and anions: (Fe, Al)3+ + O2– = Mg2+ + F–. As shown in this work, the exchange is not always complete, and some samples actually tend to be closer to dissakisite-(Ce), CeCaMgAl2[SiO4][Si2O7]O(OH), in composition. For this reason, all the REE-bearing epidote-group minerals from the subtype-2 deposits will be treated together, under the present heading.
Dollaseite-(Ce) typically occurs as dark brown aggregates, several cm wide, in a mineralized dolomite– tremolite rock. The mineral grains are normally up to 0.5 mm long and irregular in outline, but subhedral
crystals also seem to have formed in direct contact with carbonate. Dollaseite-(Ce) may occur as radiating, twinned crystals, and it is commonly associated with magnetite and norbergite. The mineral is also commonly observed as a microscopic phase in assemblages of fluorbritholite-(Ce), cerite-(Ce) and gadolinite-(Ce), as described in previous sections (Fig. 11C). In addition, a mineral closer to dissakisite-(Ce) in composition has been observed in a peculiar rock from S. Hackspikgruvan (#540025), chiefly composed of fluorite and fluorian phlogopite. Aggregates up to 4 mm in size, consisting of small (100 µm), irregular grains of dissakisite-(Ce) closely associated with magnetite and euhedral molybdenite crystals, occur in the mass of fluorite (Fig. 12C).
The atom proportions given for epidote-group minerals in this paper were calculated according to the procedure advocated for allanite by Ercit (2002): the sum of the octahedrally and tetrahedrally coordinated cations, 
(Mg + Fe + Ti + Al + Si), was normalized to six atoms, and Fe2+ and Fe3+ were partitioned to balance the negative charges from O2– and F– (Table 8). The samples can be described as essentially dollaseite – dissakisite – allanite solid solutions (Fig. 17). They are generally low in Fe3+ (Fe3+ < Al, i.e., the ferri allanite component is subordinate). Mössbauer data were collected for #270427 from Östanmossa (dollaseite close to the end-member composition) and #540027 from Malmkärra (dollaseite–dissakisite of approximately intermediate composition). In both cases, four doublets were used to fit the spectra (Fig. 18, Table 9). On the basis of their CS values, two were attributed to Fe3+ and two to Fe2+, both of them in octahedral coordination with oxygen. Also, for #540027, the ratio Fe3+/Fe determined from the absorption area is 0.51, which is within the calculated range (0.41–0.60) for the point analyses of the sample (Table 8). For #270427, the average EMP-based value (0.25) is in good agreement with the ratio obtained by Mössbauer spectroscopy (0.22). Site a
Top
Abstract
Introduction
