Occurrence and general appearance

The polymineralic nodule containing zilbermintsite-(La) was collected in August 2018 during fieldwork made by a group of the authors (AVK, RŠ and AMK) at the placer dump no. 2bis (Fig. 1). The new mineral occurs as a very few anhedral grains, typically at the contact of ferriperbøeite-(La), törnebohmite-(La) and ferriallanite-(Ce) (Fig. 2). The largest grain found measures 0.65 × 0.20 mm. Apart from the above mentioned species, the very rich REE mineral assemblage of this nodule includes bastnäsite-(La), biraite-(La), ferriallanite-(La), ferriperbøeite-(Ce), fluorbritholite-(Ce), monazite-(La), perbøeite-(La), percleveite-(Ce), percleveite-(La), perrierite-(Ce) and perrierite-(La). Non-REE bearing minerals present are quartz, thorianite and thorite.

Figure 1. (a) Placer dump no. 2bis of the Mochalin Log REE deposit where the nodule with zilbermintsite-(La) was collected; (b) freshly collected nodules with numerous REE-bearing minerals.

Figure 2. Zilbermintsite-(La) (Zlb-La) closely associated with ferriperbøeite-(La) (Fpbo-La), ferriallanite-(Ce) (Faln-Ce) and törnebohmite-(La) (Tbh-La). Polished section. SEM (BSE) image, specimen no. ML 196-2bis.

According to the distribution scale of RΕΕ minerals found at the Mochalin Log deposit (Kasatkin et al., Reference Kasatkin, Zubkova, Pekov, Chukanov, Škoda, Polekhovsky, Agakhanov, Belakovskiy, Kuznetsov, Britvin and Pushcharovsky2020a), zilbermintsite-(La) should be considered as very rare: among 300 nodules with REE-bearing minerals investigated by us, it was found in only one of them. In term of its internal structure, this nodule belongs to type 2 (Kasatkin et al., Reference Kasatkin, Zubkova, Pekov, Chukanov, Škoda, Polekhovsky, Agakhanov, Belakovskiy, Kuznetsov, Britvin and Pushcharovsky2020a) where REE minerals form thin and chaotic intergrowths with each other. However, the occurrence of zilbermintsite-(La) is restricted to along the contact between törnebohmite-(La) and ferriperbøeite-(La) grains.

Physical properties and optical data

The new mineral is light brown and translucent in thin fragments, with a brown streak and vitreous lustre. It does not fluoresce under ultraviolet light. One direction of good cleavage and one direction of imperfect cleavage were observed under the scanning electron microscope (SEM); from the structure data (see below) we assume that good cleavage could be on {100}. Parting is not observed. Zilbermintsite-(La) is brittle with a fracture stepped in the cleavage direction and uneven across it (observed under the SEM). The Vickers micro-indentation hardness (load 150 g) is equal to 838 kg mm^–2 (range 767–912, n = 4) corresponding to ca. 6 on the Mohs scale. Density could not be measured due to lack of sufficiently large monomineralic fragments and the absence of heavy liquids with suitable density. A density value calculated using the empirical formula and the unit-cell parameters from single-crystal X-ray diffraction (XRD) data is 4.684 g cm^–3.

In transmitted plane-polarised light zilbermintsite-(La) is weakly pleochroic in brown tones. The absorption scheme is Z > Y > X. Optically it is biaxial (+), with α = 1.805(7), β = 1.812(7), γ = 1.867(8) (589 nm), 2V_meas. = 40(15)° and 2V_calc. = 40°. Dispersion of optical axes is very weak, r < v. Optical orientation was not determined due to the anhedral shape of the grains.

Raman spectroscopy

The Raman spectra of zilbermintsite-(La) (Fig. 3) were obtained from polished section by means of a Horiba Labram HR Evolution spectrometer. This dispersive, edge-filter-based system is equipped with an Olympus BX 41 optical microscope, a diffraction grating with 600 grooves per millimetre, and a Peltier-cooled, Si-based charge-coupled device (CCD) detector. After careful tests with different lasers (473, 532 and 633 nm), the 532 nm Nd:YAG diode laser with the beam power of ~5 mW at the sample surface was selected for spectra acquisition to minimise analytical artefacts. A Raman signal was collected in the range of 80–4000 cm^–1 with a 100× objective (NA 0.9), the system was operated in confocal mode, beam diameter was ~1 μm and the axial resolution ~2 μm. No visual damage of the analysed surface was observed at these conditions after the excitation. Wavenumber calibration was done using the Rayleigh line and low-pressure Ne-discharge lamp emissions. The wavenumber accuracy was ~0.5 cm^–1 and the spectral resolution was ~2 cm^–1. Band fitting was done after appropriate background correction, assuming combined Lorentzian–Gaussian band shapes using the Voigt function.

Figure 3. The Raman spectrum of zilbermintsite-(La) excited by 532 nm laser: in the 100–1500 cm^–1 region and in the 2500–4000 cm^–1 region (inset in the upper right corner). The measured spectrum is shown by dots. The curve matching to dots is a result of spectral fit as a sum of individual Voigt peaks shown below the curve.

The assignment of the Raman bands is as follows: bands in the range of 3595, 3313 and 3199 cm^–1 correspond to O–H-stretching vibrations; 1013 cm^–1 corresponds to stretching vibrations of Si–O–Si fragments in Si₂O₇ groups; 845 to 968 cm^–1 to stretching vibrations of apical Si–O bonds; 550 to 700 cm^–1 are Al–O⋅⋅⋅H bending vibrations and stretching vibrations of Si–O_b–Si bonds; and 300 to 630 cm^–1 are due to mixed modes and overlapping bands of (Al,Fe³⁺,Mg)–O-stretching vibrations, as well as bending vibrations of silicate groups. Below 300 cm^–1 bands are due to lattice modes involving REE–O-, Ca–O- and Fe²⁺–O-stretching vibrations and librational vibrations of silicate groups.

The Raman spectrum of zilbermintsite-(La) is similar to those of radekškodaite-(La) and radekškodaite-(Ce) (Kasatkin et al., Reference Kasatkin, Zubkova, Pekov, Chukanov, Ksenofontov, Agakhanov, Belakovskiy, Polekhovsky, Kuznetsov, Britvin, Pushcharovsky and Nestola2020b). It is also similar to the spectra of västmanlandite-(Ce) (Holtstam et al., Reference Holtstam, Kolitsch and Andersson2005) and ferriperbøeite-(La) (Kasatkin et al., Reference Kasatkin, Zubkova, Pekov, Chukanov, Škoda, Polekhovsky, Agakhanov, Belakovskiy, Kuznetsov, Britvin and Pushcharovsky2020a) in the region of 300–1000 cm^–1 but differs significantly from them in the region of stretching vibrations of Si–O–Si fragments (1000 to 1100 cm^–1) and stretching vibrations involving REE and bivalent cations forming low-force-strength bonds (below 300 cm^–1). In particular, in the zilbermintsite-(La) spectrum, the bands of stretching vibrations of Si–O–Si fragments are significantly weaker than in the ferriperbøeite-(La) spectrum that reflects a lower Si₂O₇:SiO₄ ratio in the new mineral.

In the Raman spectrum of zilbermintsite-(La), three bands of O–H-stretching vibrations are observed, which corresponds to the number of independent sites occupied by OH groups. The band at 687 cm^–1 in the Raman spectrum of zilbermintsite-(La) is close to the strong band at 689 cm^–1 in the Raman spectrum of allanite-(Ce) (Andò and Garzanti, Reference Andò and Garzanti2014; Čopjaková et al., Reference Čopjaková, Škoda, Vašinová Galiová, Novák and Cempírek2015) and ferriallanite-(Ce) (Sobek et al., Reference Sobek, Losos, Škoda, Holá and Nasdala2023) and can be assigned to the symmetric stretching vibrations of Si–O_b–Si bonds of the epidote-type module (Wang et al., Reference Wang, Han, Guo, Yu and Zeng1994).

Chemical data

Chemical data (five spot analyses) were obtained using a Cameca SX-100 electron microprobe (WDS mode, acceleration voltage of 15 kV, a beam current of 20 nA and a 3 μm beam diameter). The spectral interference of FKα and CeMζ were manually corrected using empirically determined correction factors. The content of H₂O was not determined directly due to the scarcity of pure material and was calculated by stoichiometry on the basis of O₂₈(OH,F)₃ taking into account that bond-valence sums for the O(1), O(2) and O(3) sites are close to 1 (see below). Both the crystal structure and Raman spectroscopy data confirm the presence of OH groups and the absence of B–O, C–O and N–O bonds in the mineral. Analytical data are given in Table 1. Contents of other elements with atomic numbers higher than carbon are below detection limits.

Table 1. Chemical composition of zilbermintsite-(La).

*Total Fe content corresponding to FeO content of 6.78 wt.% was divided into Fe²⁺ and Fe³⁺ based on the occupancies of the M1–M4 sites found from crystal structure refinement data: see Table 8. Thus, 2.82 wt.% FeO and 4.40 wt.% Fe₂O₃ were calculated based on the found Fe²⁺: Fe³⁺ atomic ratio: 0.64 Fe²⁺ and 0.90 Fe³⁺ pfu, respectively (Fe³⁺: Fe²⁺ = 1.40). **For total iron calculated as FeO. ***Calculated by stoichiometry: (OH + F) = 3 apfu. S.D. – standard deviation.

Zilbermintsite-(La) grains are chemically homogeneous and the WDS analyses reveal the uniform chemical composition with only minimal inter-REE chemical variation. Lanthanum prevails significantly among REE in all analyses. The empirical formula calculated on the basis of O₂₈(OH,F)₃ atoms per formula unit (apfu) is (Ca_0.94La_2.56Ce_2.18Nd_0.20Pr_0.10Th_0.02)_Σ6.00(Al_2.96Fe³⁺_0.90Fe²⁺_0.64Mg_0.34Mn_0.13Ti_0.03)_Σ5.00Si_7.00 O₂₈[(OH)_2.42F_0.58].

The ideal formula is (CaLa₅)(Fe³⁺Al₃Fe²⁺)[Si₂O₇][SiO₄]₅O(OH)₃ which requires CaO 3.46, La₂O₃ 50.18, FeO 4.43, Al₂O₃ 9.43, Fe₂O₃ 4.92, SiO₂ 25.92, H₂O 1.66, total 100 wt.%.

The Gladstone–Dale compatibility index (1 – K _p/K _c) calculated for zilbermintsite-(La) using its empirical formula and the unit-cell parameters determined from single-crystal XRD data is 0.013 rated as superior (Mandarino, Reference Mandarino1981).

Zilbermintsite-(La) does not react with either cold hydrochloric or nitric acid.

X-ray crystallography and crystal structure

Powder XRD data were collected using a Rigaku R-AXIS Rapid II single-crystal diffractometer equipped with a cylindrical image plate detector (radius 127.4 mm) using Debye-Scherrer geometry, CoKα radiation (rotating anode with VariMAX microfocus optics), 40 kV and 15 mA. Angular resolution of the detector is 0.045°2θ (pixel size 0.1 mm). The data were integrated using the software package Osc2Tab (Britvin et al., Reference Britvin, Dolivo-Dobrovolsky and Krzhizhanovskaya2017). Powder XRD data for zilbermintsite-(La) are given in Table 2 in comparison to that calculated from single-crystal XRD data using the Atoms 5.1 program (Dowty, Reference Dowty2000). Parameters of monoclinic unit cell were calculated from the observed d spacing data using UnitCell software (Holland and Redfern, Reference Holland and Redfern1997) and are as follows: a = 8.965(4), b = 5.735(3), c = 25.096(9) Å, β = 116.68(4)° and V = 1152(8) ų.

Table 2. Powder X-ray diffraction data (d in Å) of zilbermintsite-(La).

*For the calculated pattern, only reflections with intensities ≥2 are given.

**For the unit-cell parameters calculated from single crystal data.

The strongest reflections are given in boldtype.

For the single-crystal XRD study, a grain of zilbermintsite-(La), 0.04 × 0.13 × 0.16 mm in size, extracted from the polished section and analysed previously using electron microprobe and Raman spectroscopy, was mounted on a glass fibre and examined with an Xcalibur S single-crystal diffractometer equipped with a CCD detector. A full sphere of three-dimensional data was collected. Data reduction was performed using CrysAlisPro software Version 1.171.39.46 (Rigaku Oxford Diffraction, 2018). The data were corrected for Lorentz factor, absorption and polarisation effects.

The crystal structure of zilbermintsite-(La) was solved by direct methods and refined with the use of the SHELX software package (Sheldrick, Reference Sheldrick2015) to R ₁ = 0.0757 for 2677 unique reflections with I > 2σ(I). Two O atoms (O1 and O19) were refined in isotropic approximation of displacement parameters. The crystal data, data collection information and structure refinement details are given in Table 3, atomic coordinates and thermal displacement parameters of atoms in Tables 4 and 5, and selected interatomic distances in Table 6. Bond valence calculations are given in Table 7. Assigned site occupancies (s.o.f.) based on the results of the crystal structure refinement are given in Table 8. The crystallographic information file has been deposited with the Principal Editor of Mineralogical Magazine and is available as Supplementary material (see below).

Table 3. Crystal data, data collection information and structure refinement details for zilbermintsite-(La).

* w = 1/[σ²(F _o²) + (0.1030P)² + 21.2454P]; P = {[max of (0 or F _o²)] + 2F _c²}/3.

Table 4. Atom coordinates and equivalent thermal displacement parameters (U _eq, in Ų) and site occupancy factors (s.o.f.) for zilbermintsite-(La).

* Fixed during the refinement. For M(1) Al vs Fe was refined (e _ref 20.15), for M(3) Fe vs Mg was refined (e _ref 24.46) (see Table 8). Thus, on the basis of chemical data and e _ref M(1) site was assumed to be occupied by (Fe_0.57Mg_0.26Al_0.17) possibly with minor Ti⁴⁺ admixture and M(3) site by (Fe_0.75Mn_0.15Mg_0.10).

** U _iso.

Table 5. Anisotropic displacement parameters (in Ų) in zilbermintsite-(La).

Table 6. Selected interatomic distances (Å) and angle (°) in the structure of zilbermintsite-(La).

Table 7. Bond valences for zilbermintsite-(La).

Note. Parameters from Gagné and Hawthorne (Reference Gagné and Hawthorne2015) and, for possible H-bonding, from Ferraris and Ivaldi (Reference Ferraris and Ivaldi1988). The values were calculated taking into account the refined occupancies for the M(1–4) and A(1) sites. For better clarity the split character of the O(23) and O(24) sites was not taken into account and these sites were considered as averaged with no splitting [the cations–O(23)/O(24) distances for bond-valence calculations including H-bonding were taken for non-split sites]. Slightly low values of BVS for these sites are typical for the members of the radekškodaite group. For Fe cations in the M(3) site the bond-valence parameters of Fe²⁺ were used. The bond-valence parameters of Ce³⁺ were used for A(1–6) sites. The values of BVS for A(2–6) sites could be increased taking into account the presence of La³⁺ cations. Bond valences are calculated disregarding the minor occurrence of F.

Table 8. M1–M4 sites occupancies (s.o.) suggested for zilbermintsite-(La) based on the results of the crystal structure refinement.

Note. SSF_exp and SSF_calc are the experimental and calculated site-scattering factors, respectively, normalised to unity.

The crystal structure of zilbermintsite-(La) (Fig. 4) consists of the chains of edge-sharing octahedra running along the b axis: single chains of the M(2)- and M(4)-centred octahedra and branched chains with the M(1)-centred octahedra in the central part, and the M(3)-centred octahedra attached to them from both sides. The chains are linked via isolated [SiO₄] tetrahedra and disilicate groups [Si₂O₇] [the angle Si–O–Si in the disilicate group is equal to 149.6(9)°]. The A(1–6) sites occur in large cavities. Even though La³⁺ is a dominant REE in zilbermintsite-(La), the Ce³⁺ scattering curve was used during the structure refinement for the A(1–6) sites because of the significant content of Ce³⁺ cations, minor amounts of heavier Nd³⁺ and Pr³⁺, and very minor amount of Th⁴⁺. According to the s.o.f. refinement, the A(2–6) sites are fully occupied by REE cations. The A(1) site is Ca-dominant with minor REE admixture; Ca vs Ce was refined that gave Ca_0.955(7)Ce_0.045(7). There are four octahedrally coordinated M sites, M(1–4). For the M(1), M(2) and M(4) sites, Al vs Fe was refined, and for the M(3) site, Fe vs Mg was refined. The M(2) and M(4) sites are Al-dominant with the refined occupancy Al_0.936(16)Fe_0.064(16) and Al_0.91(2)Fe_0.09(2), respectively. These sites centre the smallest octahedra with the mean distances 1.912 Å [M(2)–O] and 1.918 Å [M(4)–O]. The M(1)-centred octahedron is slightly larger, with the mean M(1)–O distance of 1.995 Å. This site is Fe³⁺-dominant, with refined occupancy Fe_0.55(2)Al_0.45(2), corresponding to number of electrons (e _ref) = 20.2(4). On the basis of the chemical information, the M(1) occupancy was therefore fixed to Fe_0.57Al_0.17Mg_0.26 in the final stage of refinement. The largest M(3) octahedron with the mean M(3)–O distance of 2.136 Å is predominantly occupied by Fe²⁺ with subordinate Mn, Mg and Fe³⁺ in the atomic ratio Fe²⁺:Mn:Fe³⁺:Mg = 0.64:0.15:0.11:0.10, according to both electron microprobe data and refined number of electrons (e _ref = 24.46). During the refinement, splitting was found for the O(23) and O(24) sites which deviate from the m plane as also occurs in the structures of radekškodaite-(La) and radekškodaite-(Ce).

Figure 4. The crystal structure of zilbermintsite-(La). SiO₄ tetrahedra are red. Alternation of epidote-type slabs (E) and törnebohmite-type slabs (T) is shown. The unit cell is outlined. Drawn using Diamond version 3.2. (Crystal Impact, 2014).

Bond-valence calculations (Table 7) confirm the above conclusions about distribution of cations between different sites: specifically, Fe³⁺ at M(1) and Fe²⁺ at M(3).