Occurrence

The mineral was found at the Karasu–Karavshinskoye Sn deposit Batken region, Kyrgyzstan located in the headwaters of Karasu River, Turkestan Range (Fig. 1). It has long been known for its pegmatite veins (Weber, Reference Weber1913). However, a systematic study of the geology and mineralogy of pegmatites began only after the famous Tajik–Pamir expedition in the first half of the 1930s (Gavrusevich, Reference Gavrusevitch1932; Ionin, Reference Ionin1934; Arapov, Reference Arapov1935) and resulted in the discovery of several pegmatite fields with rich Sn mineralisation on the northern slope of the Turkestan Range, including Tamyngen, Dzhaupaya, Dukenek, Karasu–Karavshinskoye and Aksu–Karavshinskoye by the team headed by N.V. Ionin (Ionin, Reference Ionin1934). These pegmatites bear beryl and are rich in Fe–Mn–Li-phosphate mineralisation (Kogan et al., Reference Kogan, Beus, Deberin, Paltova and Eventov1950; Beus, Reference Beus1951; Ginzburg, Reference Ginzburg1952). They are related to the Karavshinsky intrusive complex of the late Carboniferous to early Permian age, which includes the Karavshinsky (25 km²) two-phase granitic massif (Fig. 1). Pegmatites are localised both within the massif and in the host Lower Silurian andalusite schists (Kogan et al., Reference Kogan, Beus, Deberin, Paltova and Eventov1950). Some of them bear cassiterite mineralisation and are enclosed within the boundaries of the Karasu–Karavshinskoye deposit. There are several types of pegmatite veins, namely biotite–microcline, muscovite–microcline, muscovite–albite and spodumene–albite. The latter includes the pegmatite vein of #2 area, located on the right side of the Asan–Usan glacial moraine, where akopovaite was found (Fig. 1b). The vein is ~110 m long and ~12 m wide, located relatively far from the contacts of the intrusion (2.5 km to the south), and associated with quartz-micaceous schists. The abundant minerals in the pegmatite are quartz, albite, microcline–‘perthite’, muscovite and spodumene. Less common are beryl, schorl, columbite-group minerals, cassiterite, various phosphates (triphylite, griphite, amblygonite–montebrasite series members, etc.) and ‘lepidolite’. Rare minerals include microlite-group members and secondary uranyl phases, such as autunite and torbernite (Kogan et al., Reference Kogan, Beus, Deberin, Paltova and Eventov1950). Andalusite, tourmaline, quartz and muscovite are abundant in the near-contact zone of the pegmatite with schists. Pegmatite veins with akopovaite can be subdivided into four zones (Kogan et al., Reference Kogan, Beus, Deberin, Paltova and Eventov1950): (1) quartz–feldspar zone; (2) quartz–muscovite zone; (3) albite zone; and (4) the central zone containing large spodumene crystals (up to 40 cm long). The samples containing akopovaite were only found in dumps of vein material, therefore we can only suppose that they originate from the albite zone or the quartz-spodumene zone.

Fig. 1.

Geographical location (upper left) of the Karasu-Karavshinskoye deposit (the arrow points to the enlarged area, shown below) (1 – Kazakhstan, 2 – Tajikistan, 3 – Uzbekistan, 4 – China) and geological map (lower) of the Karavshinskiy batholite; polygons mark boundaries of Dukenek (1) and Karasu-Karavshinskoye deposits (2) and Aksu-Karavshinskoye occurrence (3) (drawn after A. V. Berezanskiy).

Appearance, physical and optical properties

Pegmatite samples with akopovaite consist of light, pale yellow granular albite and translucent white quartz, sometimes dark because of fine mica inclusions (albite and quartz compose 90 vol.% of the rock). Muscovite plates (1.5–2 cm across) and black to dark blue and green schorl are subordinate minerals (Fig. 2). Macroscopically, pegmatite samples with akopovaite look multi-coloured due to the presence of films of Mn and Fe oxides (hydroxides) of brown and light-brown colours in cracks and between albite and quartz grains. Akopovaite aggregates occur in areas (1–4 cm across) of alteration and (or) leaching of primary pegmatite phosphates. These areas comprise fir-like cavities morphologically similar to skeletal crystals and aggregates of unaltered phosphates abundant in other parts of the deposit. These cavities are encrusted with brown crystals of Mn-rich siderite [(Fe_0.80Mn_0.20)(CO₃) to (Fe_0.53Mn_0.47)(CO₃)] sprinkled by akopovaite spherulitic aggregates (Fig. 2). The interstices between siderite crystals also contain akopovaite aggregates and are commonly contoured by a rim consisting of Fe and Mn oxides (hydroxides), unidentified Fe–Al phosphates and montebrasite relicts (Fig. 3). Occasionally akopovaite fills box-like pseudomorphs, which have a morphology outwardly similar to those of triphylite crystals (Fig. 3) (Beus, Reference Beus1951).

Fig. 2.

(a) General view of the pegmatite sample with akopovaite, containing albite (Ab) and quartz (Qtz); brown colour is due to the presence of thin films of Fe–Mn hydroxides. Tur – schorl. Sample size is 5.5 × 4.5 cm. (b) Akopovaite forms white and yellowish spherulites on siderite (Sd).

Fig. 3.

(a) Back-scatter electron (BSE) image of leached Fe–Mn phosphate areas in muscovite (Ms) – albite (Ab) aggregate, filled with Mn-rich siderite (Sd) and akopovaite (Akp), surrounded with montebrasite (Mnb). Zones with akopovaite are contoured with a complex polyphase rim (R) consisting of Fe–Mn hydroxides, Fe–Al phosphates and montebrasite relicts. (b) BSE image and X-ray elemental maps of the interstice between siderite crystals (Sd), filled with akopovaite (Akp), and contoured by a polycomponent rim, with various Fe–Al-phosphates (Phosp), montebrasite (Mnb) and Fe–Mn-oxides and/or hydroxides (Fe-Mn-ox). Note: black (in BSE) and red (for CKα) areas correspond to the epoxy resin. Scale bar (a) 600 μm and (b) 40 μm.

Rosette-like aggregates of akopovaite are composed of pseudohexagonal thin curved plates up to 15–30 μm wide and several μm thick (Fig. 4). Both small size and poor quality of the plates precluded single-crystal X-ray diffraction (SCXRD) investigations and optical methods. The considerable amount of gibbsite and X-ray amorphous Al(OH)₃ admixtures also complicated the characterisation of the mineral.

Fig. 4.

Akopovaite rosettes: (a,b) aggregates of lamellar pseudohexagonal crystals forming rosettes; (c,d) hollow crystals (triphylite pseudomorph?), filled by akopovaite aggregates. (a–c) Scanning electron microscopy image; and (d) BSE image.

Akopovaite is white or yellowish with a silky lustre. The cleavage is perfect along {001}, which is characteristic for all hydrotalcite-supergroup minerals. Micro-indentation measurements, performed on a PMT-3 with a load of 0.5 g, are close both in basal and cross-sections and give a mean value VHN = 24 kg/mm² (range 20–31) which corresponds to a Mohs hardness of ~1. The density of akopovaite aggregates measured in Clerici solution is 2.12(2) g/cm³. The value obtained is in a good agreement with the calculated density (2.14 g/cm³) taking into account the presence of the Al(OH)₃ admixture estimated as ~10%.

Photoluminescence studies were undertaken on a 50 μm diameter section of akopovaite using a LOMO MSFU-312 microspectrophotometer (with spectral resolution ≥ 2 nm); N₂-laser with λ = 337.1 nm (3.67 eV) LGI-505 (PLASMA JSC). The study of the luminescent properties of akopovaite was carried out according to the method developed by Rassulov (Reference Rassulov2005). Akopovaite is luminescent with a weak bluish–greenish colour after excitation with a mercury-quartz lamp with both shortwave (254 nm) and mediumwave (365 nm) filters. After excitation with a nitrogen laser (λ_ex = 337.1 nm) it luminesces green. The latter gives the most intensive bands (495, 516 and 538 nm) due to the admixtures containing [UO₂]²⁺ cations (Tarashchan, Reference Tarashchan1978) in amounts below the limits of the microprobe analysis. The [UO₂]²⁺ cations can be absorbed by amorphous Al(OH)₃ or by akopovaite, although the latter is less likely.

An optical study was carried out on thin sections; separate plates were studied in immersion liquids in transmitted light using Leitz Ortholux and Polam L-211 polarising microscopes. Akopovaite is optically anisotropic. Under the microscope akopovaite plates show low interference colours in crossed polars, with an undulatory mosaic extinction because of deformations, which prevented determination of whether the mineral is uniaxial or biaxial. Therefore, it was only possible to measure n´ = 1.535(3) (at 589 nm) in a cleavage plane.

The mineral is easily soluble, both in concentrated and diluted (up to 0.5 vol.%) HCl and HNO₃ followed by the release of CO₂. The dissolution of akopovaite aggregates in weak acids leads to the formation of fine-grained precipitate corresponding to gibbsite (identified by powder XRD).

Chemical composition

Electron-microprobe analysis (for Al and Fe) was done using a Jeol JXA-50A microprobe, equipped with an Oxford instruments energy-dispersive spectrometer (voltage = 20 kV, current = 2 nA and beam diameter = 5 μm, 4 points) and a Jeol Superprobe 733, equipped with a Si(Li)-detector with an ATW2 ultra-thin window and INCA control system (voltage = 20 kV, current = 1 nA with a rastered beam, 4 points). Aluminium and Fe were determined by electron microprobe (Table 2). The following standards were used: albite, microcline, corundum (AlKα) and hematite (FeKα). Contents of other elements with Z > 8 are below their detection limits. The mineral was analysed in polished and thin-polished sections of pegmatite samples, processed with a diamond suspension using the Texmet 1000 cloth of Buehler Ltd., USA.

Table 2.

Compositional data (wt.%) for akopovaite and its mixture with gibbsite and amorphous Al(OH)₃.

⁽¹⁾ Al₂O₃, Fe₂O₃ – microprobe (8 points), Li₂O – from the ratio method (see text); CO₂ and H₂O – normalised data, derived from the columns (6–8). Formula calculated on the sum of cations = 6 atoms per formula unit is Li_1.94(Al_4.05Fe_0.02)_Σ4.07(OH)₁₂(CO₃)_1.03(H₂O)_3.03.

⁽²⁾ Al and Li determined from the Al/Li ratio, measured by ICP-OES (see text); CO₂ and H₂O measured from two weights by CHN-elemental analysis method (in brackets there are results of separate analyses).

⁽³⁾ Values calculated from columns 4 and 6.

Initial qualitative determination of Li was done using an Opton FMD-4 atomic-absorbtion spectrometer (AAS). Later, inductively coupled plasma optical emission spectrometry (ICP-OES) (Varian ICP-OES 720-ES spectrometer) was used for direct Li determination using the Al:Li ratio. The following chemicals were used: HNO₃; H₂O with resistivity = 18.2 MΩ/cm and multi-element standard solutions produced by Merck (Germany).

The contents of carbon and hydrogen in akopovaite were measured using Perkin Elmer 2400 Series II and Carlo-Erba 1106 analysers equipped with a thermal conductivity detector and CHN analysis with chromatographic completion. The following standards were used: benzoic acid, acetanilide and hydrotalcite from Snarum, Norway.

The CHN analysis shows the absence of N, and ICP-OES the absence of Be and B in akopovaite.

Certain parts of the mineral aggregates (without weighing) were placed into polypropylene tubes and dissolved in 15 ml of 1% HNO₃ solution. The solution was left for 30 min at room temperature and filtered with a 0.45 μm press filter (Varian) to remove undissolved particles, which may consist not only of gibbsite, but also of an amorphous alumina admixture, described previously for synthetic LiAl₂-LDHs (Isupov et al., Reference Isupov, Chupakina, Tarasov and Shestakova2007). Afterwards HNO₃ acid was added to the solution in order to increase its concentration to 4% (in both blank and reference solutions).

The Al:Li ratio, measured by ICP-OES is 8.11 (average of five measurements), and has been used for calculation of the Li₂O content in akopovaite (Table 2). A test on the dissolution of fine-grained (<10 μm) pure gibbsite under the same conditions (temperature, time and HNO₃ concentration) showed that <1 wt.% of it dissolves in 1% HNO₃, i.e. the majority of the Al in the solution belongs to akopovaite. Sample # 92365 (gibbsite from Amarkantak, India from the systematic collection of the Fersman Mineralogical Museum; checked by powder XRD) was used for the test described above.

The direct determination of CO₂ and H₂O was carried out using two probes of 0.670 and 0.495 mg, by CHN analysis. The two probes were purified initially: siderite particles were removed using a Sochnev magnet. As the fraction contains admixtures of gibbsite and possible amorphous Al(OH)₃, the obtained contents of CO₂ and H₂O correspond to that of the total proportion of both components in a mixture. In order to normalise the composition for 100% of akopovaite, the amount of total Al(OH)₃ (gibbsite and possible amorphous phases) in the mixture was measured. Similar to the procedure described above, tiny parts of mineral aggregates were dissolved in HNO₃ (1:1) until complete decomposition (T = 105°C, t ≈ 30 min). Deionised water (18 МΩ/сm) was added in order to reach 4% HNO₃ concentration after which the Al/Li ratio was measured in the solutions using ICP-OES. Results of three measurements provide a ratio of Al/Li = 9.36, which was used to obtain an Al(OH)₃ content of 10.2 wt.% (Table 2). It should be noted, that normalised values of CO₂ and H₂O are in a good agreement with calculated ones and correspond to the stoichiometry of the hydrotalcite-supergroup minerals (Mills et al., Reference Mills, Christy and Schmitt2016; Zhitova et al., Reference Zhitova, Krivovichev, Yakovenchuk, Ivanyuk, Pakhomovsky and Mikhailova2018, Reference Zhitova, Krivovichev, Pekov and Greenwell2019b). The empirical formula calculated on the basis of Li+Al+Fe = 6 apfu is Li_1.94(Al_4.05Fe_0.02)_Σ4.07(OH)₁₂(CO₃)_1.03(H₂O)_3.03. The simplified formula is Li₂Al₄(OH)₁₂(CO₃)(H₂O)₃.

Raman spectroscopy

The Raman spectrum of akopovaite was obtained with a Horiba Jobin-Yvon LabRam HR800 spectrometer, equipped with an Ar⁺ laser (λ = 514 nm) at 50 mW output power and ~6 mW power at the sample surface for an area of 2 μm × 2 μm. The spectrum was recorded at room temperature from randomly oriented akopovaite rosette-like aggregates placed on a glass slide and was further processed using LabSpec (Horiba) and Origin (Originlab corp) software.

The Raman spectrum of akopovaite is shown in Fig. 5. The band assignment is provided in Table 3 and compared with its synthetic analogue (Drewien et al., Reference Drewien, Eatough, Tallant, Hills and Buchheit1996), dritsite (Zhitova et al., Reference Zhitova, Pekov, Chaikovskiy, Chirkova, Yapaskurt, Bychkova, Belakovskiy, Chukanov, Zubkova, Krivovichev and Bocharov2019a) and quintinite (Theiss et al., Reference Theiss, López, Frost and Scholz2015). The Raman spectrum reveals the presence of carbonate groups and different OH vibrations that correspond to hydroxyl and interlayer water (Table 3).

Fig. 5.

The Raman spectra of (a) akopovaite and (b) synthetic Al-Li LDH (Drewein et al., Reference Drewien, Eatough, Tallant, Hills and Buchheit1996).

Table 3.

Raman bands (in wavenumber, cm^–1) of akopovaite compared to its synthetic analogue, dritsite and quintinite.

*Only bands in the range 1300–300 cm^–1 were given. The values in the brackets were not described in the text, but estimated from the band position in the given spectra. Sh – shoulder. **Al⋅⋅⋅O stretching, ***Li⋅⋅⋅O stretching.

Powder X-ray diffraction

Preliminary X-ray identification of the mineral was performed with powder XRD data (Debye-Scherrer method using an RKD 57.3 mm camera, FeKα radiation and Mn filter) and shows it's similarity with synthetic LiAl₂-LDH (Devyatkina et al., Reference Devyatkina, Kotsupalo, Tomilov and Berger1983; Drewien et al., Reference Drewien, Eatough, Tallant, Hills and Buchheit1996; Serna et al., Reference Serna, Rendon and Iglesias1982). Powder XRD data for the full-profile analysis were collected by means of a Rigaku R-Axis Rapid II single-crystal diffractometer using Debye-Scherrer geometry (d = 127.4 mm). The diffractometer is equipped with a rotating anode (CoKα, λ = 1.79026, voltage = 40 kV and current = 15 mA), microfocus optics and a cylindrical image plate detector. Five different rosette-like aggregates were investigated and gave almost identical powder patterns. The data were integrated using the software package Osc2Tab/SQRay (Britvin et al., Reference Britvin, Dolivo-Dobrovolsky and Krzhizhanovskaya2017). Topas 4.2 software (Bruker AXS, 2009) was used for the refinement of the unit-cell parameters and crystal structure by the Rietveld method (Table 4). Previously reported Rietveld refinements on synthetic LiAl₂-LDHs indicated that Li–Al ordering combined with a rhombohedral stacking sequence of layers results in a reduction of the symmetry from trigonal to monoclinic (Thiel et al., Reference Thiel, Chiang and Poeppelmeier1993; Britto et al., Reference Britto, Thomas, Kamath and Kannan2008; Britto and Kamath, Reference Britto and Kamath2009 and Reference Britto and Kamath2011). These results were later confirmed by the SCXRD study and structure refinement of quintinite-1M, Mg₄Al₂(OH)₁₂(CO)₃(H₂O)₃ (Krivovichev et al., Reference Krivovichev, Yakovenchuk, Zhitova, Zolotarev, Pakhomovsky and Ivanyuk2010b). In the latter, it was suggested that quintinite-1M is isotypic with the monoclinic polytype of synthetic Li₂Al₄(OH)₁₂(CO)₃(H₂O)₃ (Krivovichev et al., Reference Krivovichev, Yakovenchuk, Zhitova, Zolotarev, Pakhomovsky and Ivanyuk2010b). It is worth noting that Li–Al ordering has been detected in the first (and to date only) single-crystal refinement of LiAl₂-LDHs done on dritsite-2H (Zhitova et al., Reference Zhitova, Pekov, Chaikovskiy, Chirkova, Yapaskurt, Bychkova, Belakovskiy, Chukanov, Zubkova, Krivovichev and Bocharov2019a). Therefore the Rietveld refinement was performed starting from atom coordinates of quintinite-1M (monoclinic, C2/m, a = 5.266(2), b = 9.114(2), c = 7.766(3) Å, β = 103.17(3)° and V = 362.9(2) ų) in which the Mg site (in quintinite) was assigned to Al (in akopovaite), whereas the Al site (in quintinite) was assigned to Li (in akopovaite).

Table 4.

Powder X-ray diffraction data (d in Å) for akopovaite.

* Overlap with gibbsite (see Fig. 6)

The strongest lines are given in bold.

Fig. 6.

(a) Rietveld fit (red line) of the observed powder X-ray diffraction (blue line) profile of akopovaite. (b) Powder X-ray diffraction pattern obtained for quintinite-1M from Rudnogorskoe Fe-deposit (Siberia, Russia) and (c) akopovaite (CoKα).

The indexed powder XRD pattern, obtained using a Gandolfi-like motion, is provided in Table 4. The Rietveld refinement confirmed that the mineral is monoclinic, space group C2/m, a = 5.0953(6), b = 8.877(1), c = 7.806(1) Å, β = 102.572(6)°, V = 344.61(8) ų and Z = 1 (Table 5). The final refinement consisted of 25 parameters (coordinates and isotropic displacement parameters) and final convergence was achieved at R _B = 0.006 and R _wp = 0.014 (Table 5). Crystal data and parameters of the crystal-structure refinement are given in Table 5. The atomic coordinates, site occupancies and isotropic displacement parameters are listed in Table 6 and interatomic distances are provided in Table 7.

Table 5.

Crystallographic data and refinement parameters for akopovaite.

Table 6.

Atomic coordinates, site occupancies and isotropic displacement parameters for akopovaite.

*Coordinates were fixed during the refinement.

(1) Corresponds to the Al site in quintinite; (2) corresponds to the Mg site in quintinite.

**B = 8π²U ², where U ² is the mean-square amplitude of the vibration.

Table 7.

Selected interatomic distances in the crystal structure of akopovaite.

The difficulty in the structure refinement of akopovaite is in the presence of light elements and statistic disorder of interlayer C and O atoms (also characteristic for single-crystal structure refinements of other hydrotalcite-supergroup minerals) (Krivovichev et al., Reference Krivovichev, Yakovenchuk, Zhitova, Zolotarev, Pakhomovsky and Ivanyuk2010a,Reference Krivovichev, Yakovenchuk, Zhitova, Zolotarev, Pakhomovsky and Ivanyukb; Zhitova et al., Reference Zhitova, Yakovenchuk, Krivovichev, Zolotarev, Pakhomovsky and Ivanyuk2010, Reference Zhitova, Popov, Krivovichev, Zaitsev and Vlasenko2017a,Reference Zhitova, Ivanyuk, Krivovichev, Yakovenchuk, Pakhomovsky and Mikhailovab, Reference Zhitova, Krivovichev, Yakovenchuk, Ivanyuk, Pakhomovsky and Mikhailova2018, Reference Zhitova, Krivovichev, Pekov and Greenwell2019b). Therefore, atom coordinates were refined with soft restraints applied for the interlayer atoms. The atomic coordinates were fixed for H1, H2 and C1. Two isotropic-displacement parameters were refined for: (1) Li, Mg, O1 and O2 atoms of the metal–hydroxide layer; and (2) interlayer C and O atoms and hydrogen H1 and H2 atoms. The occupancy of atoms was fixed in accordance with the chemical formula determined by the number of different methods. Neutral scattering factors were used for all atoms. The background was modelled using a Chebyshev polynomial approximation. The peak profile was described using the fundamental parameters approach. Refinement of preferred orientation parameters confirmed the presence of a significant preferred orientation along the [001] direction. The gibbsite model (Balan et al., Reference Balan, Lazzeri, Morin and Mauri2006) was added to the refinement because of the intimate intergrowth of akopovaite with gibbsite as indicated by the comparison of the calculated and observed intensity of the reflection at 4.397 Å (Fig. 6). The Rietveld refinement showed good agreement between experimental and calculated diffraction patterns (Fig. 6). Less smooth fitting is observed in the 2θ region 23–27°, due to stacking faults causing non-uniform broadening of reflections, as previously shown for the synthetic analogue of akopovaite by Britto and Kamath (Reference Britto and Kamath2011).

The crystallographic information files have been deposited with the Principal Editor of Mineralogical Magazine and are available as Supplementary material (see below).