Occurrence and paragenesis

The geology of the Ilímaussaq complex is well described in the literature (Ussing, Reference Ussing1912; Larsen and Sørensen, Reference Larsen and Sørensen1987; Sørensen and Larsen, Reference Sørensen and Larsen1987, Reference Sørensen and Larsen2001; Upton, Reference Upton2013). The Gardar Province went through a series of continental rifts, alkaline magmatism and sedimentation 1300–1140 Ma. The province is separated into the two major periods of volcanic activity called Older Gardar (1300–1163 M.y.) and Younger Gardar (1163–1140 M.y.). A review of age determination of Ilímaussaq, including a new age of 1156 ± 1.4 Ma, is provided by Borst et al. (Reference Borst, Waight, Finch, Storey and Le Roux2019). The Mesoproterozoic igneous Ilímaussaq complex was formed during late magmatism along the southern rift zone and marks the end of the magmatism in the Gardar Province (Upton, Reference Upton2013).

The Taseq Slope is situated in the northern part of Ilímaussaq and predominantly consists of naujaite. Naujaite is a cumulate rock consisting of greyish to greenish euhedral sodalite enclosed in alkali-feldspar, black to dark-green arfvedsonite and aegirine, and red eudialyte. The naujaite exhibits a poikilitic texture due to the early formation of cumulus sodalite in the magma chamber, which floated to the top of the magma chamber. The eudialyte content can vary a lot as also seen in the kakortokites, however the magmatic layering is not as pronounced in the naujaites. In addition to the naujaites, a more evolved rock, lujavrite is also common on the Taseq slope. This hyperagpaitic rock is the youngest of the nepheline syenites, consisting of the same felsic mineralogy, i.e. nepheline, albite, microcline and sodalite. There are several varieties of lujavrites in the complex, with both textural and mineralogical differences. However, the two main lujavrites can be separated by colour based on which mafic mineral is present in the rock: black (arfvedsonite) and green (aegirine) (Ussing, Reference Ussing1912; Upton, Reference Upton2013; Friis, Reference Friis2015). Illoqite-(Ce) was discovered in loose boulders by a large ussingite vein on the Taseq Slope in the Ilímaussaq alkaline complex. The initial sample was collected by one of the authors (HF) in 2014. Additional material was collected in the summer of 2019. Despite searching the area no samples were found in situ, therefore it is not possible to establish where in the ussingite vein the mineral formed.

Illoqite-(Ce) is only the sixth Ba-mineral to be described from Ilímaussaq after barylite, ilímaussite-(Ce), joaquinite-(Ce), kuannersuite-(Ce) and ortho-joaquinite-(La) (Petersen, Reference Petersen2001). All of these minerals are rare in Ilímaussaq, however ilímaussite-(Ce) has its type locality on the Taseq slope and the large ussingite vein is where most of this mineral has been found (Fig. 1). One theory to why we find two Ba-rich minerals here is that due to the size of the ussingite vein, any incompatible element not fitting the structure of ussingite or the other main minerals will have to form discrete species. Alongside ussingite as a matrix mineral, aegirine and arfvedsonite occur as major components. Furthermore, the paragenesis of illoqite-(Ce) consists of britholite-(Ce), ancylite-(Ce), epistolite, chkalovite, lueshite, steenstrupine-(Ce) and polylithionite, where ancylite-(Ce) is the brownish weathering product of illoqite-(Ce) (Fig. 2).

Fig. 2. Example of the paragenesis of illoqite-(Ce) with the brown ‘band’ running through the rock being weathered illoqite-(Ce). Ussingite (Usg) and chkalovite (Ckl) are marked respectively.

The abundance of chkalovite (Fig. 2) lead Engel et al. (Reference Engell, Hansen, Jensen, Kunzendorf and Løvborg1971) to conduct a Be-mineralisation study of the Taseq slope. The study produced some detailed maps over its geology and mineralogy (Fig. 3). The origin of the ussingite veins is not fully understood but they could be classified as hydrothermalites, believed to be a primary, but late, stage in the alkaline rock forming process involving extreme alkaline hydrothermal fluids (Khomyakov, Reference Khomyakov1995). Consequently, the ussingite veins, and illoqite-(Ce), should be considered formed by primary magmatic processes rather than traditional later hydrothermal alteration events.

Fig. 3. Illoqite-(Ce) locality at the Taseq slope. (a) Detailed geological map modified from Engell et al. (Reference Engell, Hansen, Jensen, Kunzendorf and Løvborg1971). (b) Highlighted geological map of (a), including the location of the illoqite-(Ce) occurrence. The vein is sloping with decreasing altitude towards the northwest. (c) Picture of the large ussingite close to the illoqite-(Ce) occurrence. Geological hammer as scale.

Physical properties

Illoqite-(Ce) occurs as either single euhedral crystals up to 150 μm in size or radiating aggregates consisting of a few or many crystals. The aggregates are up to 200 μm in diameter (Fig. 4). Illoqite-(Ce) can occur as scattered small groups of crystals or aggregates, but sometimes they occur in high concentrations creating clusters or bands consisting almost completely of illoqite-(Ce) (Fig. 2). However, illoqite-(Ce) is always intergrown with ussingite. The colour of illoqite-(Ce) is pink in daylight, but brownish yellow under fluorescent light and some halogen lights. It shows no photoluminescence under either long-wave or short-wave ultraviolet light. Most of the physical properties could not be determined due to small grain size and intergrowths with ussingite, but the mineral appears dull and the powder colour is white with a faint yellow tint. The calculated density based on the empirical formula and refined unit cell parameters is 3.65(3) g/cm³. For the above mentioned reasons and the chemical zonation (Fig. 5), optical properties were only determined from the thin section used for chemical characterisation. The birefringence was determined to be ~0.021, which is in good agreement with the published values for nordite minerals ranging between 0.019 and 0.023 (Gerasimovsky, Reference Gerasimovsky1941; Pekov et al. Reference Pekov, Chukanov, Kononkova, Belakovsky, Pushcharovsky and Vinogradova1998, Reference Pekov, Chukanov, Turchkova and Grishin2001). To estimate n _av. we used the Gladstone–Dale compatibility index and by assuming the range of Superior and Excellent, we find n _av. to be between 1.626 and 1.651 (Mandarino, Reference Mandarino1981). The range is consistent with the published values for nordites n _av. = 1.632–1.635 (Gerasimovsky, Reference Gerasimovsky1941; Pekov et al. Reference Pekov, Chukanov, Kononkova, Belakovsky, Pushcharovsky and Vinogradova1998, Reference Pekov, Chukanov, Turchkova and Grishin2001), and using the Becke line method showing that n _av. of illoqite-(Ce) is between those of ussingite (1.519) and arfvedsonite (1.682).

Fig. 4. Back-scattered electron images of two illoqite-(Ce) aggregates in ussingite showing the chemical heterogeneity (both images from sample KNR 44274). In both images the scale bar is 100 μm.

Fig. 5. Schematic drawing of typical sector zoning in illoqite-(Ce) with the Ba/Sr ratio given for each zone. Note: Some crystals may have compositional zones like nordite-(Ce).

Chemistry

The quantitative chemical data were collected using an electron microprobe CAMECA XS100 equipped with five wavelength dispersive spectrometers housed at the Department of Geosciences, University of Oslo. X-ray lines and spectrometer diffracting crystals used for analysis were selected to minimise interference between REE and between Ba–Ce, and an overlap correction procedure applied to correct the interference of ZnLβ on NaKα. Matrix effects were corrected using the PAP procedure (Pouchou and Pichoir, Reference Pouchou and Pichoir1984). The analytical conditions during collection were the following: wavelength dispersive mode, accelerating voltage of 15 kV; beam current of 15 nA; beam diameter of 5 μm and number of analyses of 15. The standards used for calibration were wollastonite (SiKα and CaKα), metallic Fe (FeKα), albite (NaKα), synthetic MgO (MgKα), pyrophanite (MnKα), ZnS (ZnLα), BaSO₄ (BaLα), Sr-SiO glass (SrLα) and ‘synthetic REE orthophosphates’ (LaLα, CeLβ, PrLβ and NdLβ) obtained from the Smithsonian Institution, Washington D.C (Jarosewich and Boatner, Reference Jarosewich and Boatner1991).

To get quantitative values on Li the epoxy mount used for electron probe micro-analyser (EPMA) was analysed on a Bruker Aurora Elite quadrupole laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) equipped with a 213 nm CTAC laser housed at the Department of Geoscience, University of Oslo. NIST 610 was used for instrument calibration and to monitor drift during experiments. Silicon measured on EPMA was used as internal standard and the data were processed using the Glitter program (Griffin et al. Reference Griffin, Powell, Pearson and O'Reilly2008). The chemical compositions of illoqite-(Ce) are given in Table 3.

Table 3. Chemical electron microprobe data (in wt.%) for illoqite-(Ce).

S.D. = standard deviation (2σ); *From LA-ICP-MS; ^#Jarosewich and Boatner (Reference Jarosewich and Boatner1991).

Based on the recent approved nordite supergroup nomenclature (Dal Bo et al., Reference Dal Bo, Gulbransen and Friis2021), the empirical formula of illoqite-(Ce) calculated on the basis of 17 anions is: Na_2.00Na_1.00(Ba_0.59Sr_0.32Ca_0.04Na_0.03)_Σ0.98(Ce_0.68La_0.31Nd_0.09Pr_0.04)_Σ1.12(Zn_0.42Fe_0.34Li_0.14Mn_0.09)_Σ0.99Si_5.97O_17. The simplified formula is: Na₂Na(Ba,Sr)(Ce,La,Nd)(Zn,Fe,Li)Si₆O₁₇, while the ideal formula is: Na₂NaBaCeZnSi₆O₁₇, which requires Na₂O 10.91; BaO 17.99; Ce₂O₃ 19.26; ZnO 9.55 and SiO₂ 42.30; total 100 (wt.% oxide). The ideal formula is derived from the empirical formula using the IMA dominant-constituent rule (Hatert and Burke, Reference Hatert and Burke2008). The chemical variation in particular of Sr and Ba (Table 3) is the result of pronounced sector zoning (Figs 4 and 5). The individual zones are smaller than the diameter of the laser used and too small to separate for single-crystal X-ray diffraction (SXRD). To keep consistency between the average results of the SXRD data we have decided to use the average composition including the various zones, and not just the one with highest Ba content. Pekov (Reference Pekov2000) mentioned that it is possible to find zonal nordite with the Z-site occupants varying from Mn dominant, Fe dominant and Zn dominant. This is similar to illoqite-(Ce) though the X-site determines the zonation rather than the Z-site occupants.

Crystallography

Powder X-ray diffraction data were collected on the same fragment as the structure solution using the Gandolfi movement of a Rigaku Synergy-S diffractometer housed at the Natural History Museum in Oslo equipped with a HyPix-6000He detector and CuKα radiation (50 kV and 1 mA). Table 4 presents the observed and calculated powder diffraction data based on the structure model described below. The unit cell parameters were refined in an orthorhombic setting based on 32 reflections using the program UnitCell (Holland and Redfern, Reference Holland and Redfern1997) and are: a = 14.5340(7), b = 5.2191(2), c = 19.8121(8) Å and V = 1502.84(7) Å³.

Table 4. Experimental and calculated* powder X-ray diffraction data (d in Å) for illoqite-(Ce).

* The calculated values were obtained using VESTA 3 (Momma and Izumi, Reference Momma and Izumi2011).

The strongest values are given in bold.

Single-crystal X-ray intensity data were collected at room temperature with monochromated MoKα radiation (50 kV and 1 mA). The instrument has Kappa geometry and both data collection and subsequent data reduction together with face based absorption corrections were carried out using the CrysAlisPro software (Rigaku Oxford Diffraction, UK). It was not possible to separate a single domain, due to the chemical heterogenity of the mineral for all the crystals tested. To get a better separation of data from different domains the sample-to-detector distance was 50 mm. We chose the data collection where the main component was more than 90% of the total observed reflections and used the main component data for the structure description. The structure was solved by direct methods using SHELXS and refined by SHELXL (Sheldrick, Reference Sheldrick2008) using neutral atom scattering factors and the WinGX interface (Farrugia, Reference Farrugia2012). Once all atoms were located, they were refined anisotropically and the occupancies were refined freely for the main cations (A = B = Na and Y = Ce). The mixed occupancies Ba vs Sr and Zn vs Fe were refined at the X and Z-sites, respectively. See Table 5 for details of data collection and refinement. Atom coordinates, anisotropic atomic displacement parameters and site occupancies are given in Tables 6–8, respectively. The crystallographic information file has been deposited with the Principal Editor of Mineralogical Magazine and is available as Supplementary material (see below).

Table 5. Data collection and structure refinement details for illoqite-(Ce).

* P = ((F _o)² + 2(F _c)²)/3)

Table 6. Site population, atomic coordinates and equivalent isotropic displacement parameters (Ų) for illoqite-(Ce).

Table 7. Anisotropic atomic displacement parameters (Ų) for illoqite-(Ce).

Table 8. Cationic distribution in the crystal structure of illoqite-(Ce) and nordite-(Ce).

RSS: Refined site scattering factor, these values are based on the refined site occupancies in Table 6; apfu: atoms per formula unit; CSS: Calculated site scattering factor; ABL: average observed bond-lengths; CBL: calculated bond-lengths; Ideal bond-distances are calculated using the ionic radii of Shannon (Reference Shannon1976). Ce³⁺ is used as a proxy for all REE. ¹Dal Bo et al. (Reference Dal Bo, Gulbransen and Friis2021)

Illoqite-(Ce) is isostructural with the other members of the nordite group, and has a structure based on 12-periodic chains consisting of tetrahedrally coordinated Si (T-sites) that are interconnected via tetrahedrally coordinated Zn (Z-site) creating a sheet perpendicular to the b axis (Fig. 6). Adjacent sheets of tetrahedra are separated by a heteropolyhedral layer containing the A (Na), B (Na), X (Ba) and Y (REE) sites which are interconnected by edge and face sharing (Fig. 6). The B-site has an octahedral coordination, whereas the A-site occupants have six bonds between 2.412 and 2.687 Å and one long bond to O8 at 3.119 Å, making it seven coordinated (Table 9). The O7 is at a distance of 3.316 Å from A, but this only contributes 0.02 valence units to the bond and therefore we considered the coordination to be 7-coordinated (Table 10). However, with a different composition the A-site might become eight coordinated (Fig. 7). In that case the coordination of the A-site can be described as a square antiprism.

Fig. 6. General view of the crystal structure of illoqite-(Ce) along [010] (a) and [001] (b). The oxygen atoms are represented by red spheres and the solid lines show one unit cell.

Fig. 7. View of the coordination around the crystallographic ^[7+1]A, ^[6]B, ^[8]X, ^[8]Y and ^[4]Z sites in the structure of illoqite-(Ce). The displacement ellipsoids represent the 90% probability level.

Table 9. Selected bond distances (d in Å) for illoqite-(Ce).

Table 10. Bond-valence sums (vu) for illoqite-(Ce).

Note: bond-valence parameters are recalculated according to the site occupancies (see Table 8), and taken from Brown and Altermatt (Reference Brown and Altermatt1985) for all the cations apart from Si and REE for which the parameters from Gagné and Hawthorne (Reference Gagné and Hawthorne2015) have been used.

The X-site has a significantly larger scatter (50.2 e^–) than for a pure Sr-site, which is caused by Ba predominant in the site, with minor Sr and REE (Table 8). The Y-site is fully occupied by REE of which Ce is dominant. The incorporation of Ba for Sr at the X-site also results in an increase of the average bond distance of this site from 2.620 Å in nordite-(Ce) to 2.701 Å in illoqite-(Ce) (Table 8). It is somewhat surprising that the nordite structure can accommodate Ba as in previous studies the site contained not more than 5% Ba (e.g. Sokolova et al., Reference Sokolova, Kabalov and Khomyakov1992; Pekov et al., Reference Pekov, Chukanov, Kononkova, Belakovsky, Pushcharovsky and Vinogradova1998, Reference Pekov, Chukanov, Turchkova and Grishin2001). The only major substituent for Sr reported so far has been Ca (e.g. Bakakin et al., Reference Bakakin, Belov, Borisov and Solovyeva1970; Dal Bo et al., Reference Dal Bo, Gulbransen and Friis2021) suggesting that the nordite structure is more flexible when it comes to incorporating smaller cations than larger cations at the X-site. However, the discovery of illoqite-(Ce) shows that this is not the case. Both the X and Y-sites show regular square antiprism coordination (Fig. 7).

The tetrahedrally coordinated Z-site is dominated by Zn, but has significant Fe, suggesting an iron dominant equivalent may exist. Generally, all the main cation sites have slightly larger average bond distances in illoqite-(Ce) compared to nordite-(Ce), which in turn is shown in a unit cell volume increase from 1472 to 1507 Å³. The bond-valence calculations in accordance with the cationic distribution are presented in Table 10 and show only minor deviations from ideal values. Contrary to other nordite-supergroup minerals, illoqite-(Ce) also has Li as a monovalent cation in the Z-site. The average bond distance of the site is 1.973 Å which is very close to the ideal 1.96 Å for tetrahedrally coordinated Li (Wenger and Armbruster, Reference Wenger and Armbruster1991). Although only a few Li minerals are known from Ilímaussaq, polylithionite occurs in parts of the ussingite vein hosting illoqite-(Ce).