Crystal Structure of Kristiansenite from Szklarska Poręba , Southwestern Poland

Kristiansenite, ideally Ca2ScSn(Si2O7)(Si2O6OH), a rare late-stage hydrothermal Sc-bearing sorosilicate mineral, was found in a gadolinite-fergusonite-type pegmatite of the MI-REE subclass related to the Karkonosze granite, exposed in a quarry at Szklarska Poręba, Lower Silesia, Poland. Kristiansenite occurs in an association with andradite, epidote, allanite-(Ce), titanite, fersmite, scheelite, Sc-bearing columbite-(Fe), a YNbO4 mineral as fergusonite-(Y) or fergusonite-(Y)-beta, silesiaite and wolframite. Single-crystal study of the mineral (R1 of 4.96%), with composition Ca2.00(Sn0.97Sc0.69Fe0.17Mn0.05Ti0.04Zr0.03Nb0.02Al0.02Ta0.01)Σ2(Si2O7)[(Si1.98Al0.02)Σ2O6.03(OH)0.97], corroborates its triclinic structure with space group-symmetry C1, Z = 2, and unit-cell parameters a = 10.0304(5), b = 8.4056(4), c = 13.3228(6) Å, α = 90.001(3), β = 109.105(3), γ = 89.997(3)◦ and V = 1061.40(9) Å3. In the structure of the mineral, the Ca and Si sites are dominantly occupied with Ca and Si, whereas the M1–M4 sites are disordered. The M3 and M4 sites are occupied dominantly by Sn and subordinately Sc, whereas the M1 and M2 sites are occupied dominantly by Sc and subordinately by remaining occupants, including Sn.

In this paper, we discuss the crystal chemistry and structure of kristiansenite found in a NYF-affiliated pegmatite in the Szklarska Poręba Huta quarry in Lower Silesia, southwestern Poland, related to post-magmatic stages in the Karkonosze granite intrusion.

Geological Setting
The Karkonosze granite is the largest Variscan pluton in the Western Sudetes at the northeastern margin of the Bohemian Massif.In Poland, it extends among Jelenia Góra and Janowice Wielkie on the northeast, Kowary and Śnie żka Mt. to the south, and spreads to the west up to Liberec in the Czech Republic (Figure 1).The main rocks of the massif are porphyritic to equigranular biotite-bearing granite associated with a small volume of two-mica granite and subordinate granophyric granite [7,8], locally with lamprophyre and aplite dykes.The pluton is considered as a syn-collisional magmatic arc intrusion on the active margin of a continental platform neighboring oceanic crust, crystallized from relatively reduced, K-rich, calc-alkaline, highly evolved, mostly peraluminous I-type to transitional I-S type magmas ~320-310 Ma [9][10][11][12][13][14][15][16].
Pegmatites are particularly abundant in the porphyritic Karkonosze granite, and less frequent in the metamorphic cover of the granite massif.Most of them are barren, but there are also widespread weakly to moderately fractionated rare element-bearing pegmatites of NYF affiliation [17].Intragranitic pegmatites usually form zoned pods, lenses and, more rarely, dikes of length ranging from centimetres to several decimetres, exceptionally several metres, whereas the pegmatites intruding the metamorphic cover are discordant massive dikes and lenticular bodies.In the classification of Černý and Ercit [18], the Karkonosze pegmatites correspond mainly to (i) podiform and lenticular gadolinite-fergusonite type intragranitic pegmatites of the miarolitic-rare-earth-elements (MI-REE) subclass; (ii) barren, schorl-bearing thin dikes, cutting discordantly the metamorphic envelope closest the granite.
The Karkonosze granite is the largest Variscan pluton in the Western Sudetes at the northeastern margin of the Bohemian Massif.In Poland, it extends among Jelenia Góra and Janowice Wielkie on the northeast, Kowary and Śnieżka Mt. to the south, and spreads to the west up to Liberec in the Czech Republic (Figure 1).The main rocks of the massif are porphyritic to equigranular biotite-bearing granite associated with a small volume of two-mica granite and subordinate granophyric granite [7,8], locally with lamprophyre and aplite dykes.The pluton is considered as a syn-collisional magmatic arc intrusion on the active margin of a continental platform neighboring oceanic crust, crystallized from relatively reduced, K-rich, calc-alkaline, highly evolved, mostly peraluminous I-type to transitional I-S type magmas ~320-310 Ma [9][10][11][12][13][14][15][16].
Pegmatites are particularly abundant in the porphyritic Karkonosze granite, and less frequent in the metamorphic cover of the granite massif.Most of them are barren, but there are also widespread weakly to moderately fractionated rare element-bearing pegmatites of NYF affiliation [17].Intragranitic pegmatites usually form zoned pods, lenses and, more rarely, dikes of length ranging from centimetres to several decimetres, exceptionally several metres, whereas the pegmatites intruding the metamorphic cover are discordant massive dikes and lenticular bodies.In the classification of Černý and Ercit [18], the Karkonosze pegmatites correspond mainly to (i) podiform and lenticular gadolinite-fergusonite type intragranitic pegmatites of the miarolitic-rare-earth-elements (MI-REE) subclass; (ii) barren, schorl-bearing thin dikes, cutting discordantly the metamorphic envelope closest the granite.

Occurrence
The Szklarska Poręba Huta quarry, located within a town Szklarska Poręba near Jelenia Góra, is currently the only place in the Polish part of the massif where granite is exploited, mainly as dimension stone.Small bodies of granitic NYF-type pegmatite,

Occurrence
The Szklarska Poręba Huta quarry, located within a town Szklarska Poręba near Jelenia Góra, is currently the only place in the Polish part of the massif where granite is exploited, mainly as dimension stone.Small bodies of granitic NYF-type pegmatite, forming lenses and nests up to several tens of centimeters across, are relatively frequent within the coarsely crystalline variety of the porphyritic to equigranular granite; smaller nests up to only a few centimeters across and very often with a small cavity in the centre, are frequent in the granophyre granite.

Crystal Structure Refinement
Single-crystal X-ray diffraction measurements were made at C-HORSE (the Centre for Higher Order Structure Elucidation, in the Department of Chemistry at University of British Columbia) on a kristiansenite crystal 45 × 60 × 40 µm, using a Bruker X8 APEX II diffractometer with graphite monochromated MoKα radiation, at a crystal-to-detector distance of 40 mm.The data were collected at 100 K to a maximum 2θ value of 60.2 • , in a series of φ and ω scans in 0.50 • oscillations with 60 s exposures, integrated using the Bruker SAINT software package [28], and corrected for absorption effects using the multi-scan technique (SADABS; [29]), and for Lorentz and polarization effects.As the sample diffracted only weakly, a long collection time (48 h) was required to acquire enough reflections for a refinement.The collection was performed at 100 K to reduce noise from thermal vibrations.All refinements were performed using the SHELXTL crystallographic software package [30] of Bruker AXS.The weighting scheme was based on counting statistics.Neutral atom scattering factors were taken from Cromer and Waber [31].Anomalous dispersion effects were included in F calc [32]; the values for ∆f and ∆f" were those of Creagh and McAuley [33].The values for the mass attenuation coefficients are those of Creagh and Hubbell [34].
The structure was solved by direct methods [30] in the monoclinic space group C2/c, as indicated by experiment, using neutral atom scattering factors, and refined to an R 1 index of 6.45% with anisotropic displacement parameters.However, in their original description of kristiansenite, Ferraris et al. [6] found that the crystals of kristiansenite typically show metric merohedral twinning.While the underlying structure including cation ordering is triclinic, the twin law belonging to the higher Laue group makes the structure appear monoclinic within experimental error.However, the presence of cation ordering at the octahedral M sites reveals the triclinic structure.The triclinic structure was derived from the monoclinic structure using online software at the Bilbao Crystallographic Server [35], with atomic positions in the triclinic structure based directly on equivalent positions in the monoclinic structure.The position of Ca1 was fixed arbitrarily as the origin.Following [6], the structure was refined to an R 1 index of 4.96% in the unconventional triclinic space group setting C1.

Results
Compositional data from EMPA of the studied kristiansenite are presented in Table 1.Atomic contents for the averaged analysis (4 spots), give the following empirical formula of the mineral: Ca 2.00 (

Discussion
Taking into account accuracies in the microprobe analysis, it can be assumed that the Ca sites, with a lack of Na, are completely occupied by Ca, and the Si sites are dominantly occupied by Si, with a small deficiency completed by subordinate Al.The occupancies are corroborated by the refined electron densities and occupancies of the sites (Table 3), as well as average <Ca-O> distances in the 7-fold-coordinated Ca1-Ca4 polyhedra (2.487-2.504Å; Table 4), and Si1-Si8 tetrahedra (1.615-1.643Å).
The four M sites of the kristiansenite cell have very diverse occupants, dominated by tetravalent (Sn, Ti, Zr) 4+ and trivalent cations (Sc, Fe, Al) 3+ , respectively; however, trace amounts of divalent (Mn 2+ ) and pentavalent (Nb, Ta) 5+ cations are also present.This is similar to the composition of the type material [1], but the presence of Mn, Nb, and Ta are new.The total charge of the M-site occupants in the formula unit equals 6.99 valence units, and the calculated H + , bonded to oxygen atoms coordinating Si, is 0.97 apfu; both nearly equal to the ideal values in the general composition of the mineral.Substitutions at the octahedral M sites must, therefore, be mutually charge-balanced to a high degree.These substitutions involve the distribution of cations among the M sites and lead to the observed disordering of the divalent to pentavalent M-type cations among the M1-M4 sites.The refined total content of electrons at the M sites of the unit cell, 145.9 e − , agrees well with 145.1 e − calculated from the EMPA formula.The structure refinement separates the electrons among the M1-M4 sites in the following way: 27.2(4); 30.8(3); 43.8(4); 44.1(4) e − , respectively, which leads to the Sn-Sc site occupancies presented in Table 3.
However, the real population of the M-site occupants is much more diversified than the simple Sn-Sc system.To approximate the distribution of octahedral cations among the M1-M4 sites, the refined electron densities at the sites were recalculated proportional to the value of 145.1 e − from the empirical microprobe formula: 27.1; 30.6; 43.6 and 43.9 e − .Note that the differences between the measured and calculated electron densities are smaller than the respective 1σ ranges in the refinement.Since the M4 and M3 sites have very similar high electron densities, Sn is predicted to be one of their main occupants.The remaining occupants were tested from various combinations of octahedral cations with the best results achieved for the pair Sn-Sc.The M4and M3-site occupancies were calculated as (Sn 0.788 Sc 0.212 ) and (Sn 0.778 Sc 0.224 ), respectively, corresponding to the recalculated electron densities and featuring very realistic calculated averaged cation radii close to ~0.702 Å.The remaining M2 and M1 cations are highly dominated by Sc 3+ (0.957 apfu as a result of Z•Sc total − M4 Sc − M3 Sc).Due to the predominance of Sc, the occupancies of the sites were analyzed in the system Sc-(X), where (X) denotes a complex cation average of the rest of the Sn and traces of all other octahedral cations.For such a system, the M2 occupancy were evaluated at (X 0.640 Sc 0.360 ) and the M1 occupancy at (X 0.403 Sc 0.597 ), respectively, for which the calculated averaged cation radii are equal to ~0.703 Å and ~0.716 Å, respectively.Formally, the M2 site is dominated by trivalent Sc + Fe (0.57 apfu) with dominant Sc (0.36 apfu), and the M1 site is highly dominated by Sc (0.60 apfu), with only subordinate Fe and Sn (0.13 and 0.14 apfu, respectively).The arrangement of the calculated averaged M1-M4 radii vs. the refined mean bond lengths in the M1-M4 octahedra (Table 4) is represented by an equation <M i -O> = 1.025 • <r i > + 1.343 with a correlation coefficient R 2 = 0.986 that is very close to the ideal relationship <M i -O> = <r i > + rO 2− (Figure 2).It corroborates that in the studied kristiansenite from Szklarska Poręba, the M1 and M2 sites are dominated with Sc, and the M3 and M4 sites dominated with Sn.The distribution of cations among the sites is highly disordered, with visible preference of Sn-Sc disorder at the M3 and M4 sites, and Sc-X disorder at the M2 and M1 sites.
The kristiansenite unit cell contains two hydrogen atoms on average, shared among the O5-O8 and O25-O28 oxygens, but OH − only prevails over O 2− at two sites, O25 bonded with M1 and O26 bonded with M2.The O25-O27 and O26-O28 distances, 2.604 and 2.641 Å respectively, imply the presence of two hydrogen bonds with the O25 and O26 donor oxygens (b.v.= −1.34 and −1.41;Table 5), and the O27 and O28 acceptor oxygens (b.v.= −1.57and −1.62).The equivalent distances in the structure of type kristiansenite are 2.627 and 2.646 Å [5].The presence of hydroxyl anions around the M1 and M2 sites dominated by M 3+ cations suggests a simple substitution M 3+ OH(M 4+ O 2− ) −1 .This also explains the subordinate presence of OH − at the O7-O8 sites coordinated around the M3 and M4 sites dominated by M 4+ cations, because subordinate amounts of M 3+ -type cations are present at the sites only due to the M 3+ -M 4+ disorder.

Figure 1 .
Figure 1.Geological sketch of the Polish part of the Karkonosze granite massif with its metamorphic envelope.Abbreviations: ISB-Intra-Sudetic Basin.Symbols: star-the quarry at Szklarska Poręba, black lines-faults.

Figure 1 .
Figure 1.Geological sketch of the Polish part of the Karkonosze granite massif with its metamorphic envelope.Abbreviations: ISB-Intra-Sudetic Basin.Symbols: star-the quarry at Szklarska Poręba, black lines-faults.

Table 1 .
Sn 0.97 Sc 0.69 Fe 3+ 0.17 Mn 0.05 Ti 0.04 Zr 0.03 Nb 0.02 Al 0.02 Ta 0.01 ) Σ2 (Si 2 O 7 ) [(Si 1.98 Al 0.02 ) Σ2 O 6.03 (OH) 0.97 ].Details of the data measurement and structure solution are listed in Table2.Atomic coordinates, equivalent isotropic-displacement parameters, and occupancy for atomic sites in the kristiansenite are given in Table3.Selected bond lengths and bond angles are collected in Table4, and a bond-valence calculation is shown in Table5.Anisotropic displacement parameters are presented in CIF file attached as Supplementary Materials.Average electron-microprobe analysis of kristiansenite from Szklarska Poręba (n = 4).
n-number of analytical spots; Na and Mg were below detection; *-calculated assuming all Fe as Fe 3+ .

Table 2 .
Data measurement and refinement information for kristiansenite.

Table 3 .
Atomic coordinates, equivalent isotropic displacement parameters, and occupancy for atomic sites in kristiansenite (the occupancy for anion sites are from bond valence analysis).