Occurrence and associated minerals

The lead author (CR) found the specimen CR202, containing sperlingite, in mid 1974 on the mine dump at the Hagendorf Süd feldspar mine, in the Oberpfalz, northeast Bavaria, Germany (49°39′1″N, 12°27′35″E). Based on the time of collection, and the mineral associations in the specimen, particularly zinc-bearing minerals, the specimen most probably originated from the 67 m level of the mine (Mücke, Reference Mücke1981; Grey et al., Reference Grey, Keck, MacRae, Glenn, Mumme, Kampf and Cashion2018). The matrix of the specimen consists of strongly corroded zwieselite residues in quartz, with clusters of sperlingite crystals occupying corrosion pits in the zwieselite (Fig. 1). Accompanying minerals are scholzite, hopeite, leucophosphite, orange–brown zincoberaunite sprays (Fig. 2), tiny green crystals of zincolibethenite, CuZn(PO₄)(OH), olive green mitridatite and columbite. The colourless sperlingite crystals are commonly stained with mitridatite coatings (Figs 1 and 2). In addition to the close association of sperlingite with zwieselite, crystals are also observed growing on and within scholzite (Fig. 3). Sperlingite and scholzite are probably the youngest phosphate minerals in the specimen.

Figure 1. Aggregates of colourless sperlingite crystals in a corrosion pit in zwieselite associated with scholzite (large crystal in upper left). Photo by Christian Rewitzer, holotype specimen MSM38185, FOV = 0.3 mm.

Figure 2. Aggregates of sperlingite crystals associated with sprays of zincoberaunite needles. Brown staining on sperlingite is mitridatite. Photo by Christian Rewitzer, holotype specimen MSM38185.

Figure 3. Back-scattered electron image of polished epoxy mount of holotype specimen MSM38185, used for EMP analyses, showing dark grey sperlingite crystals in a light grey scholzite matrix, associated with fluorapatite (medium grey). FOV = 80 μm.

Physical and optical properties

Crystals of sperlingite, in the form of colourless prisms with pyramidal terminations are predominantly only 5 to 20 μm in size (Fig. 3), rarely to 60 μm and are frequently multiply intergrown and overgrown with smaller crystals. The calculated density is 2.40 g⋅cm^–3 for the empirical formula and single-crystal unit-cell parameters.

The small size of the sperlingite crystals limited the measurement of the optical properties; however, it was possible to measure two indices of refraction in grain mounts. Based upon these and by comparison with the optical properties and morphologies of other paulkerrite-group minerals, it was possible to conjecture the following properties: biaxial (+), α = 1.600(est), β = 1.615(5), γ = 1.635(5) (white light) and 2V (calc.) = 82.7°. The optical orientation is X = b, Y = c and Z = a. Neither dispersion nor pleochroism were observed.

Raman spectroscopy

Raman spectroscopy was conducted on a Horiba XploRA PLUS spectrometer using a 532 nm diode laser, 100 μm slit and 1800 gr/mm diffraction grating and a 100× (0.9 NA) objective. The spectrum is shown in Fig. 4. The O–H stretch region has a broad double hump that can be assigned to H-bonded water, with maxima at 3338 and 3057 cm^–1. According to Libowitzky (Reference Libowitzky1999) these correspond to O⋅⋅⋅O distances involved in H-bonding of 2.65 and 2.75, corresponding to moderately strong H-bonding. Hydroxyl ion stretching is evident by a weak peak at 3600 cm^–1. The H–O–H bending mode region for water has a peak at 1630 cm^–1. Two peaks at 1012 and 965 cm^–1 in the P–O stretching region can be assigned to symmetric stretching modes whereas weaker peaks at 1135 and 1100 cm^–1 correspond to antisymmetric P–O stretching modes. Bending mode vibrations of the (PO₄)^3– groups are located at 610 cm^–1 and at 485 and 425 cm^–1. Peaks at lower wavenumbers are related to lattice vibrations. The spectrum for sperlingite is dominated by a strong peak at 838 cm^–1 with a shoulder at 785 cm^–1. These peaks are present in all paulkerrite-group minerals (Grey et al., Reference Grey, Hochleitner, Rewitzer, Kampf, MacRae, Gable, Mumme, Keck and Davidson2023a–Reference Grey, Hochleitner, Kampf, Boer, MacRae, Mumme and Keckc) and can be assigned to Ti–O stretch vibrations for short Ti–O bonds that occur in linear trimers of corner-connected octahedra M2–M3–M2 in the structure, by analogy with published Raman spectra for titanates containing short Ti–O distances (Tu et al., Reference Tu, Guo, Tao, Katiyar, Guo and Bhalla1996; Bamberger et al., Reference Bamberger, Begun and MacDougall1990; Silva et al., Reference Silva, Filho, Silva, Balzuweit, Bantiignies, Caetano, Moreira, Freire and Righi2018).

Figure 4. Raman spectrum of sperlingite.

Chemical composition

Highly hydrated paulkerrite-group minerals present problems for analysis because of dehydration in the high vacuum of the conductive film coater and the microprobe, resulting in severe cracking and high analysis totals (Sejkora et al., Reference Sejkora, Skoda, Ondrus, Beran and Susser2006). Cracking of crystals of sperlingite during coating of a conductive iridium film is seen in the polished section used for the electron microprobe analyses in Fig. 3. To prevent further dehydration during analysis a cold stage cooled to liquid nitrogen temperature was employed in the microprobe and the specimen was precooled under dry nitrogen prior to introduction to the microprobe vacuum.

Crystals of sperlingite were analysed using wavelength-dispersive electron microprobe (EMP) spectrometry on a JEOL JXA 8530F Hyperprobe operated at an accelerating voltage of 15 kV and a beam current of 2.0 nA. The beam was defocused to typically ~5 μm. Both specimen and standards were coated with a 25 Å thick film of iridium for the analyses. The F K peak was partially overlapped by Mn L and Fe L and this was corrected using a peak overlap procedure. In addition the thin film correction procedure was utilised in STRATA (Pouchou, Reference Pouchou1993) to remove the effects of the Ir coating. There was insufficient material for direct determination of H₂O, so it was calculated based on the ideal formula (14 H₂O + 2OH^– per 4 P). Analytical results (average of 11 analyses on 11 crystals) are given in Table 2, where they are compared with the published analyses for rewitzerite (Grey et al., Reference Grey, Hochleitner, Kampf, Boer, MacRae, Mumme and Keck2023c). Relatively high standard deviations are due to chemical zoning of the crystals, shown by variations in back-scatter contrast in Fig. 3. The EMP results show strong positive correlations of Al with F (R ² = 0.83) and with K (R ² = 0.78) and negative correlations of Ti with K (R ² = 0.86) and with F (R ² = 0.73). Al correlates negatively with Fe (R ² = 0.76). (Mg + Zn) has a moderate negative correlation with Fe (R ² = 0.61).

Table 2. Analytical data (wt.%) for sperlingite.

* Based on the ideal formula: 14 H₂O + 2OH^– per 4P.

From the mean analyses, the number of atoms per formula unit (apfu), normalised to 4P apfu is:

$${\rm K}_{ 0 .56}{\rm M}{\rm n}_{ 0 .60}{\rm M}{\rm g}_{ 0 .33}{\rm Z}{\rm n}_{ 0 .29}{\rm F}{\rm e}^{{\rm 3 + }}_{ 1 .39} {\rm A}{\rm l}_{ 1 .05}{\rm T}{\rm i}_{ 1 .33}{\rm P}_{ 4 .00}{\rm F}_{ 0 .19}{\rm O}_{ 32 .73}{\rm H}_{ 30 .00} .$$

Expressing the apfu in structural form and allowing for local charge balance for Fe³⁺ at M1, by replacing an equivalent amount of H₂O coordinated to M1 with OH^–, according to the model proposed for sigloite by Hawthorne (Reference Hawthorne1988) gives the following empirical formula. The M2 and M3 sites are grouped based on the site-total-charge procedure (Bosi et al., Reference Bosi, Hatert, Halenius, Pasero, Ritsuro and Mills2019a, Reference Bosi, Biagioni and Oberti2019b; Grey et al. Reference Grey, Boer, MacRae, Wilson, Mumme and Bosi2023d):

$$^ A{^1} [ {{( {{\rm H}_ 2{\rm O}} ) }_{ 0 .96}{\rm K}_{ 0 .04}} ] _{{\rm \Sigma 1} .00} \ \ {^A} {^2} ( {{\rm K}_{ 0 .52}{\rm W}_{ 0 .48}} ) _{{\rm \Sigma 1} .00} \ \ {^M} {^1} ( {{\rm M}{\rm n}^{{\rm 2 + }}_{ 0 .60} {\rm M}{\rm g}_{ 0 .33}{\rm Z}{\rm n}_{ 0 .29} } \cr \qquad {\rm F}{\rm e}^{{\rm 3 + }}_{ 0 .77} }_ } ) _{{\rm \Sigma 1} .99} {^M} {^{\rm 2 + \it M 3} }( {{\rm A}{\rm l}_{ 1 .05}{\rm T}{\rm i}^{{\rm 4 + }}_{ 1 .33} {\rm F}{\rm e}^{{\rm 3 + }}_{ 0 .62} } ) _{{\rm \Sigma 3} .00}( {{\rm P}{\rm O}_ 4} ) _ 4 \cr \qquad ^{X}\! [ {{\rm F}_{ 0 .19}{( {{\rm OH}} ) }_{ 0 .94}{\rm O}_{ 0 .87}} ] _{{\rm \Sigma 2} .00}[ {{( {{\rm H}_ 2{\rm O}} ) }_{ 9 .23}{( {{\rm OH}} ) }_{ 0 .77}} ] _{{\rm \Sigma 10} .00}{\rm \cdot 3} .96{\rm H}_ 2{\rm O}$$

The simplified formula is

$$\eqalign{ ( {{\rm H}_ 2{\rm O}} ) ( {{\rm K, \; \squ}} ) ( {{\rm M}{\rm n}^{{\rm 2 + }}{\rm , \;Mg, \;Zn, \;F}{\rm e}^{{\rm 3 + }}} ) _ 2( {{\rm Al, \;T}{\rm i}^{{\rm 4 + }}{\rm , \;F} {\rm e}^{{\rm 3 + }}}\! ) _ 3 \cr \quad ( {{\rm P}{\rm O}_ 4} ) _ 4[ {{\rm F, \;}( {{\rm OH}} ) {\rm , \;O}} ] _ 2[ {{\rm H}_ 2{\rm O, \;OH}} ] _{ 10}{\rm \cdot 4}{\rm H}_ 2{\rm O}$$

The ideal formula is (H₂O)K(Mn²⁺Fe³⁺)(Al₂Ti)(PO₄)₄[O(OH)][(H₂O)₉(OH)]⋅4H₂O, which requires K₂O 5.04, MnO 7.60, Al₂O₃ 10.92, P₂O₅ 30.41, TiO₂ 8.56, Fe₂O₃ 8.55, H₂O 28.92, total 100.00 wt.%.

Note that the M1 site has similar levels of divalent (1.22 apfu) and trivalent (0.77 apfu) cations, and the dominant divalent cation is Mn²⁺, giving the end-member M1 (=M1a+M1b) site composition as (Mn²⁺Fe³⁺), illustrated in Fig. 5, while the merged (M2₂M3) site composition corresponds to the end-member composition (Al₂Ti) as shown in Fig. 6.

Figure 5. ternary diagram for (M1)₂ site Mn²⁺–Mg–Fe³⁺ compositions, showing end-member compositions and location of the empirical composition for sperlingite. Note that the divalent cations correspond to the dominant cations found at M1 in paulkerrite-group minerals.

Figure 6. Ternary diagram for (M2)₂M3 site Al–Ti–Fe³⁺ compositions, showing end-member compositions (Al₂Ti, Ti₂Al etc.) and location of the empirical composition for sperlingite. For comparison the published empirical compositions are shown for the paulkerrite-group minerals benyacarite (Demartin et al., Reference Demartin, Pilati, Gay and Gramaccioli1993,Reference Demartin, Gay, Gramaccioli and Pilati1997), paulkerrite (Peacor et al., Reference Peacor, Dunn and Simmons1984), mantienneite (Fransolet et al., Reference Fransolet, Oustriere, Fontan and Pillard1984), rewitzerite (Grey et al., Reference Grey, Hochleitner, Rewitzer, Kampf, MacRae, Gable, Mumme, Keck and Davidson2023a), pleysteinite (Grey et al., Reference Grey, Hochleitner, Kampf, Boer, MacRae, Mumme and Keck2023c) and hochleitnerite (Grey et al., Reference Grey, Boer, MacRae, Wilson, Mumme and Bosi2023d). Red crosses correspond to minerals with Mn at M1 and blue crosses correspond to minerals with Mg at M1.

Crystallography

Powder X-ray diffraction data were obtained using a Rigaku R-AXIS Rapid II curved imaging plate microdiffractometer, with monochromatised MoKα radiation. Observed d values and intensities were derived by profile fitting using JADE Pro software. Data (in Å for MoKα) are given in Table 3. Refined monoclinic (space group: P2₁/c (#14)) unit cell parameters from the powder data using JADE Pro with whole pattern fitting are a = 10.43(3) Å, b = 20.28(3) Å, c = 12.22(3) Å, β = 90.1(6)°, V = 2585(10) ų and Z = 4.

Table 3. Powder X-ray diffraction data (d in Å) for sperlingite (I _calc > 1.5)*.

* The strongest lines are given in bold.

A crystal measuring 0.020 × 0.020 × 0.010 mm was used for a data collection at the Australian Synchrotron microfocus beamline MX2 (Aragao et al., Reference Aragao, Aishima, Cherukuvada, Clarken, Clift, Cowieson, Ericsson, Gee, Macedo, Mudie, Panjikar, Price, Riboldi-Tunnicliffe, Rostan, Williamson and Caradoc-Davies2018). Intensity data were collected using a Dectris Eiger 16M detector and monochromatic radiation with a wavelength of 0.7109 Å. The crystal was maintained at 100 K in an open-flow nitrogen cryostream during data collections. The diffraction data were collected using a single 36 second sweep of 360° rotation around phi. The resulting dataset consists of 3600 individual images with an approximate phi angle of each image being 0.1 degrees. The raw intensity dataset was processed using XDS software (Kabsch, Reference Kabsch2010) to produce data files that were analysed using SHELXT (Sheldrick, Reference Sheldrick2015) and JANA2006 (Petříček et al., Reference Petříček, Dušek and Palatinus2014). Refined unit-cell parameters and other data collection conditions are given in Table 4.

Table 4. Crystal data and structure refinement for sperlingite.

*w = [σ²(ǀF _oǀ)+(uF _o)²]^–1, u = instability factor

Structure refinement

A structural model for sperlingite was obtained in space group P2₁/c using SHELXT (Sheldrick, Reference Sheldrick2015). The SHELXT model had the same structure as for rewitzerite (Grey et al., Reference Grey, Hochleitner, Kampf, Boer, MacRae, Mumme and Keck2023c) and so the rewitzerite coordinate file was used to initiate the refinement to ensure the same atom labelling. Twinning was implemented with 2-fold rotation about c. To establish site scattering at the M1 to M3 sites, pairs of light and heavy elements were incorporated at the sites and their occupancies were refined; Mn + Mg at M1 and Ti + Al at M2 and M3. Initially K plus O (for H₂O) were incorporated at the A sites with full occupancy, but refinement of their relative amounts gave a K content considerably lower than the EMP value. Next, vacancies were introduced to increase the K content, but with 3 components at the A sites, the site compositions are indeterminate. The simplest model, with only two components at each site, was to have K plus vacancies at the site with the higher scattering, and H₂O plus K at the other. This gave 0.66 K apfu, which is within the range of EMP analyses for K. After preliminary refinements, the program OccQP (Wright et al., Reference Wright, Foley and Hughes2000) was applied to optimise the site occupancies based on the empirical formula together with refined bond distances and site scattering. The output from OccQP had only Al and Ti at the M3 sites, but Al, Ti and Fe at the M2 sites. The M2 site occupancies were then modified to include a fixed amount of Fe (0.31 Fe per site) with refinement of Al and Ti. A pleasing result from this refinement was that the total Al and Ti contents from refinement of the occupancies at the M2 and M3 sites agreed with the values from the EMP analyses.

Refinement with anisotropic displacement parameters in JANA2006 converged at R_obs = 0.050 for 5608 reflections with I > 3σ(I). Difference-Fourier maps were used to search for H atoms but unambiguous locations could not be established. This is most likely because the chemical zoning caused local atomic shifts of the oxygen atoms in response to different elements at metal atom sites, which is reflected in high atomic displacement parameters as shown in Table 5. Details of the data collection and refinement are given in Table 4. The refined coordinates, equivalent isotropic displacement parameters and bond valence sum (BVS) values (Gagné and Hawthorne, Reference Gagné and Hawthorne2015) are reported in Table 5. For the M1 sites, the BVS values were calculated based on the site occupancy in the empirical formula, 0.30Mn²⁺ + 0.165Mg + 0.145Zn + 0.39Fe³⁺. For the M2, M3 and A sites the BVS values were calculated using the refined site occupancies as listed in Table 6. Selected interatomic distances are reported in Table 7. The crystallographic information file has been deposited with the Principal Editor of Mineralogical Magazine and is available as Supplementary material (see below).

Table 5. Refined atom coordinates, site scattering (electrons), equivalent isotropic displacement parameters (Ų) and bond valence sums (BVS, in valence units) for sperlingite.

Table 6. Refined site occupancies and site scattering for sperlingite.

*Mn scattering curve used for Mn+Fe+Zn

Table 7. Polyhedral bond lengths [Å] for sperlingite.

Although the H atoms in sperlingite could not be located during the refinement, we have located the majority of H atoms in refinements of the isostructural minerals fluor-rewitzerite (Hochleitner et al., Reference Hochleitner, Grey, Kampf, Boer, MacRae, Mumme and Wilson2024) and macraeite (Grey et al., Reference Grey, Rewitzer, Hochleitner Kampf, Boer, Mumme and Wilson2024) and have established the H-bonding in these paulkerrite-group minerals. There is good agreement between the H-bonding schemes for the two minerals and also with that reported for benyacarite (Demartin et al., Reference Demartin, Pilati, Gay and Gramaccioli1993). Applying this information to sperlingite, the O⋅⋅⋅O pairs involved in hydrogen bonding are listed in Table 8, together with bond valences, s, calculated from O⋅⋅⋅O using the Ferraris and Ivaldi (Reference Ferraris and Ivaldi1988) formula s = [(O⋅⋅⋅O)/2.17]^–8.2 + 0.06. The contributions to the BVS from the H bonds, as listed in Table 8, complement reasonably the undersaturated BVS values for the acceptor anions, O1 to O8, in Table 5.

Table 8. H-bonding in sperlingite.

¹ from Ferraris and Ivaldi (Reference Ferraris and Ivaldi1988).

Description of the structure

The crystal structure for sperlingite is based on an open 3D framework of corner-connected octahedra and tetrahedra of composition [(MnFe³⁺)(Al₂Ti)(PO₄)₄O₂(H₂O)₁₀]^1–, with water molecules and K⁺ ions occupying <110> channels in the framework. Although H atoms were not located in the refinement the BVS values in Table 5 show that O9 to O15 are H₂O, as well as the main constituent at A1. The groups O9 to O12 are coordinated to M1 and O13 is coordinated to M2. The framework is built from heteropolyhedral layers parallel to (001) and located at z = ¼ and ¾, shown in Fig. 7, that are interconnected by corner-sharing of M2O₄X(H₂O) octahedra with M3O₄X ₂ octahedra located at z = 0 and ½. The heteropolyhedral (001) layers are built from [100] kröhnkite-type chains (Hawthorne, Reference Hawthorne1985) of 4-member rings of corner-connected PO₄ tetrahedra and M2O₄X(H₂O) octahedra. Each PO₄ tetrahedron also shares a corner with M1O₂(H₂O)₄ octahedra along [010]. The corner-shared linkages form 8-member rings of alternating octahedra and tetrahedra. The A1 (H₂O) and A2 (K) sites are located in the 8-member rings as shown in Fig. 7.

Figure 7. (001) section through the sperlingite crystal structure at z = ¼. Atom labelling consistent with Table 5, with K at A2 and H₂O at A1 sites. Prepared using ATOMS (Dowty, Reference Dowty2004).

The major difference between the crystal structure of sperlingite and that of orthorhombic paulkerrite-group minerals such as benyacarite (Demartin et al., Reference Demartin, Pilati, Gay and Gramaccioli1993) is an ordering of H₂O and K at the A1 and A2 sites, whereas they are disordered at a single A site in benyacarite. The different coordination environments at A1 and A2 are compared in Table 7 and shown in Fig. 7. The coordinations are very similar, except for the bond to O15, with the K-containing A2 site having a distance to O15a that is 0.24 A shorter than the A1–O15b distance. The same large difference in A-O15 distances is observed for other monoclinic paulkerrite-group minerals, rewitzerite (Grey et al., Reference Grey, Hochleitner, Kampf, Boer, MacRae, Mumme and Keck2023c) and paulkerrite (Grey et al., Reference Grey, Boer, MacRae, Wilson, Mumme and Bosi2023d).