Crystal chemistry and nomenclature of the lovozerite group

The lovozerite group includes nine valid mineral species – cyclosilicates with closely related crystal structures: combeite, imandrite, kapustinite, kazakovite, koashvite, litvinskite, lovozerite, tisinalite and zirsinalite. Structurally, the lovozerite group can be subdivided into three subgroups different in cation arrangement and, as a consequence, in symmetry and unit-cell parameters (Table 1¹).

All lovozerite-group minerals (LGM) contain Na as an essential mineral-forming constituent. They occur in only alkaline rocks. With the exception of combeite, all these minerals were first discovered in the Khibiny–Lovozero alkaline complex, Kola Peninsula, Russia (Gerasimovsky, 1940; Kapustin et al., 1973, 1974a and b, 1980; Khomyakov et al., 1974; Pekov et al., 2000, 2003). Zirconium-dominant members of the group occur as accessory or even rock-forming minerals in some peralkaline magmatic rocks and pegmatites of the Lovozero intrusion. Lovozerite-group minerals, except combeite, have been studied using samples from the Khibiny–Lovozero complex, and the majority of the original publications on these minerals are in Russian. Zirconium representatives of the group have also been found in other alkaline complexes (Ilímaussaq in Greenland; Mont Saint-Hilaire in Québec, Canada; Wind Mountain in New Mexico, U.S.A.; Norra Kärr in Sweden; Konder in Siberia, Russia; Aris in Namibia), but the minerals from these localities are only poorly studied lacking any compositional data. Combeite was found in alkaline volcanic rocks of Shaheru (DR Congo) (Sahama & Hytönen, 1957), Oldoinyo Lengai (Tanzania) and Ettringer Bellerberg (Eifel, Germany) (Fischer & Tillmanns, 1987).

Lovozerite-type structures possess certain zeolite-like features that cause wide variations in the extra-framework content charge-balanced by the chemical variability of the framework forming polyhedra. The empirical data on chemical composition, X-ray crystallography and crystal chemistry of minerals with lovozerite-type structures show wide ranges of cation substitutions; in particular, those involving vacancies in some sites, as well as symmetry variations in chemically related minerals, even those belonging to the same subgroup (Table 1¹). As a result of this variability, there is an actual problem with the assignment of some samples to definite mineral species within the lovozerite group. This prompted the elaboration of a nomenclature for the lovozerite group with clear definitions of the mineral species and criteria for their identification.

The present nomenclature scheme was accepted by the IMA Commission on New Minerals, Nomenclature and Classification (CNMNC) on 19 January 2009. It is based upon available empirical data and theoretical considerations in chemistry, crystallography and crystal chemistry of the LGM. Data on related synthetic compounds and specific chemical properties have also been taken into account.

Taking into account coordination polyhedra occupied by high-valence cations, the structures of the lovozerite-group minerals can be considered as based upon a heteropolyhedral framework consisting of six-membered [Si₆O₁₈] rings of SiO₄ tetrahedra (Fig. 1¹) and MO₆ octahedra, where M = Zr⁴⁺, Ti⁴⁺, Fe³⁺and Ca²⁺(Fig. 2²) (Ilyukhin & Belov, 1960; Chernitsova et al., 1975; Fischer & Tillmanns, 1983; Tamazyan & Malinovskii, 1990). Within the silicate ring, each SiO₄ tetrahedron possesses three types of anions: two O²⁻ anions bridging adjacent Si atoms, one bridging Si and M, and one that does not form bonds to framework-forming cations. The latter anion, occupying the so-called “pendent” vertex of a tetrahedron, can be protonated resulting in a Si–OH (silanol) group. The polyhedral framework forms a 3D system of channels hosting A, B, and C cation sites (Fig. 2² and 3³). The A cations (A = Na, Ca) are in eightfold coordination close to a hexagonal bipyramid, whereas the B cations (B = Na) have eightfold coordination with anions arranged at the vertices of a distorted cube. The C cations (Ca, Na, Mg or □) centre octahedra that share faces with the M-centred octahedra.

The first review of the structures of the lovozerite group was presented by Chernitsova et al. (1975) who proposed an elegant approach to visualize the structural flexibility. According to Chernitsova et al. (1975), the structures of the lovozerite-type compounds can be described as derived from double 4-membered ring (4-MR) of SiO₄ tetrahedra. The ideal double 4-MR has a cubic symmetry m3m with an inversion centre at the centre of the ring. Elimination of two tetrahedra related by the inversion centre results in a single 6-MR in chair conformation (Fig. 1¹). This ring possesses a three-fold rotation axis running through its centre. All structures of the lovozerite-group minerals can be considered as based upon these pseudocubic modules centred at the midpoint of the 6-MR. The M, A, and B cations are located at the borders of the module, whereas C cations are inside the module and approximately replace the “missing” Si tetrahedra.

The pseudocubic module has an average edge length of 7.5 Å and possesses rhombohedral symmetry 3m. In the structures of the LGM, the pseudocubic modules are stacked in four different arrangements shown in Fig. 4⁴. In the simplest structure (Fig. 4a⁴), all modules share whole faces and are in parallel orientation. This arrangement is the most common and is observed in trigonal members of the group and related synthetic compounds (Zolotarev et al., 2008). The rhombohedral unit cell of lovozerite exactly corresponds to the outlines of the lovozerite module (with a slight trigonal distortion).

In the imandrite-type structures, modules share two opposite faces to form columns of face-sharing cubes. The remaining four faces participate in a half-face linkage shown in Fig. 4b⁴. In koashvite (Fig. 4c⁴) and the synthetic phase Na₅(Na_0.5+xCa_0.5−x)₂(Nd_xCa_1−x)₂ (Si₆O₁₈), x = 0–0.5 (Tamazyan & Malinovskii, 1989) (Fig. 4d⁴), cubic modules share opposite faces forming face-sharing columns. In addition, out of four remaining equatorial faces, two are shared in the whole-face fashion but mutual arrangement of these faces is different. In Na₅(Na_0.5+xCa_0.5−x)₂(Nd_xCa_1−x)₂(Si₆O₁₈), the faces are opposite to each other, i.e. in trans-conformation, whereas, in koashvite, they share one edge, i.e. in cis-conformation.

The four basic types of arrangements of cubic modules are completely defined by the local interactions of cubic modules with each other, i.e. the resulting 3D structure can be derived from the exact definition of how a central cube is surrounded by its neighbours. Note that, in the structures shown in Fig. 4b–d⁴, modules share their half-faces. As a consequence, the arrangement of the A, B, and M cations in the resulting structure should be modified. Figure 5⁵ demonstrates this re-arrangement using the M cations as an example. In “classic” lovozerite module (see, e.g., Fig. 2²), M cations occupy the vertices of the module (Fig. 5a⁵). Linkage of modules via half-faces may lead either to the appearance of the M cations both at the vertices and at the midpoints of the edges (as shown in Fig. 5b and c⁵), or in “transition” of some M cations from the vertices to the midpoints of edges (as shown in Fig. 5d and e⁵). Note that, in the former case, the M:Si ratio changes from 1:6 to 2:6 as found in synthetic Na₅(Na_0.5+xCa_0.5−x)₂ (Nd_xCa_1−x)₂(Si₆O₁₈). In all lovozereite-group minerals known to date, only the schemes with the M:Si ratio = 1:6, shown on Fig. 4a–c⁴, are realized. Accordingly, the location of the M, B, and A positions differs from those in the “classic” lovozerite module (M at the vertices, A at the midpoints of edges, B at centres of faces: Fig. 6 a⁶). Figures 6b and c⁶ show arrangements of the A, B, and M cations relative to the outlines of the modules in koashvite and imandrite, respectively. It can be seen that the relationship of the sites to the borders of the cubic module are no longer geometrically defined.

Only three out of four arrangements have so far been found in the structures of naturally occurring minerals (Fig. 4a–c⁴), thus, we distinguish three different subgroups in the lovozerite group:

1. the zirsinalite–lovozerite subgroup. which includes combeite, kapustinite, kazakovite, litvinskite, lovozerite, tisinalite and zirsinalite;

2. the koashvite subgroup, which has only koashvite as natural representative;

3. the imandrite subgroup, which has only imandrite as natural representative.

In order to avoid confusion with the name of the group, structures with face-sharing arrangements of the modules are assigned to the zirsinalite–lovozerite subgroup. This combined name also reflects existence of two kinds of members of this subgroup with different degrees of Na occupancy in the structure channels:

• cation-saturated (kapustinite, kazakovite, and zirsinalite with ideally 6 Na apfu based on Si₆ and combeite with 5 cations pfu + one vacancy = Na_3.5Ca_1.5□);

• cation-deficient (litvinskite, lovozerite, and tisinalite with ideally 3 apfu Na) – Table 1¹.

Representatives of these three subgroups can easily be distinguished on the basis of their powder X-ray diffraction patterns and optical properties (Khomyakov, 1995). These characteristics are important diagnostic tools when chemical compositions seem to be insufficient for assignment of a mineral to one of these three subgroups (e.g., for the intermediate members of the kazakovite–koashvite solid-solution series).

Note that petarasite, Na₅Zr₂Si₆O₁₈(Cl, OH)·2H₂O, included by Mandarino (1999) in the lovozerite group, has a structure that is significantly different from the lovozerite-like structures (Ghose et al., 1980). The 6-membered silicate ring in petarasite is almost planar resulting in a very differently shaped stacking module. In addition, petarasite is characterized by presence of the additional anion (Cl, OH) absent in LGM.

The general formula of the lovozerite-group minerals is A₃B₃C₂MSi₆O₁₂O_6−x(OH)_x·nH₂O, with 0 ≤ x ≤ 6 and n = 0–1.

Species-defining constituents in lovozerite-group minerals are:

• M: tetra-, tri-or divalent cations: Zr⁴⁺, Ti⁴⁺, Fe³⁺, ²⁺ (Zr-, Ti-and Fe-dominant minerals can form continous solid-solution series: Pekov, 2005);

• C: in minerals with M⁴⁺ and M³⁺: divalent cations Ca or Mn²⁺ plus vacancy (□); in combeite with M²⁺: Ca ≈ Na;

• A: Na (in combeite: Na ≈ Ca);

• B: Na and □.

The following admixtures (non-species-defining constituents) may be significant for lovozerite-group minerals: Nb₅₊ – up to 17 atomic % of the M cations in titanium-dominant members (Yamnova et al., 2003), and K substituting for Na probably in the A site. Negligible admixtures of Al (probably substitutes Si), Hf (in zirconium-dominant minerals, in the M site), Mg and Zn (probably in the C site), Sr, Y and lanthanides (probably in the A site: Yamnova et al., 2001b; Zolotarev et al., 2008) were detected in some cases. Fe may not only occupy the M site, but probably as Fe²⁺, in subordinate amounts, also the C site. Semenov (1972) and Khomyakov (1995) have assumed that Mn⁴⁺ together with Mn²⁺ occurs in dark to black, altered zirconium-dominant members from Lovozero. The presence of H₂O molecules in some lovozerite-group minerals has been confirmed by IR spectroscopy and structural data (Yamnova et al., 2001a and b; Pekov et al., 2003; Pekov, 2005) – they partially occupy the A and B sites.

For “non-defect” lovozerite-type phases, i.e. those with all cation sites fully occupied, with a tetravalent cation in the M octahedron, only monovalent cations can be present in the A, B and C sites due to charge constraints: A⁺₃B⁺₃C⁺₂M⁴⁺Si₆O₁₈. An example of such a compound is synthetic Na₈Sn⁴⁺Si₆O₁₈ (Safronov et al., 1980; Zayakina et al., 1980). If tri-or divalent cations occupy the M octahedron, then divalent cations may be present in the C sites, even if all sites are completely occupied. Such “non-defect” synthetic lovozerite-type compounds do also exist: Na₆Cd₃Si₆O₁₈ (Simonov et al., 1967), Na₅(Mn, Na)₃MnSi₆O₁₈ (Otroshchenko et al., 1973), Na₆Mn₃Si₆O₁₈ (Pushcharovsky et al., 1976), Na₆Ca₃Si₆O₁₈ (Ohsato et al., 1985), and Na₅(Na, Ca)₂ (Nd, Ca)₂Si₆O₁₈ (Tamazyan & Malinovskii, 1989).

All lovozerite-group minerals are characterized by structures with vacancies in the B and/or C sites (both sites may be up to 100 % vacant), and sometimes also in the A site (partial occupancy), i.e. (A + B + C + M) < 9 apfu in all studied natural samples.

If the M site is occupied by a tetravalent cation (Zr or Ti) and the A and B sites are fully occupied by Na (6 apfu), then no more than 50 % occupancy of the C site by divalent cations (Ca or Mn²⁺) is possible as a result of: (1) general charge constraints and (2) local steric constraints due to the cation-cation repulsion in the cluster of face-connected M and C octahedra where the M and C cations are located within a short distance of ~3.2 Å (Pyatenko et al., 1999). If the central M octahedron is occupied by a tetravalent cation in this MC₂ cluster (Fig. 7⁷), then only one of two C octahedra can be occupied by a divalent cation (typically two C sites are statistically occupied, 50 % or less). All structural data on lovozerite-group minerals with M⁴⁺ (Voronkov et al., 1979; Pudovkina et al., 1980; Malinovskii et al., 1993; Yamnova et al., 2001a and b, 2003, 2004; Zolotarev et al., 2008) are in agreement with this conclusion. A surplus of divalent cations (over 1 apfu) in the C site can only exist in case of a corresponding substitution in the M site of a tetravalent cation by a trivalent cation (e.g., koashvite – Chernitsova et al., 1980b) or if Na occupies the C octahedron (e.g., for anhydrous minerals, the hypothetic end-member Na₆Na₂M⁴⁺Si₆O₁₈ could be supposed). Unlike the C site, the A and B sites in such cation-saturated lovozerite-group minerals are fully or almost fully (90–100 %) occupied by Na (Voronkov et al., 1979; Chernitsova et al., 1980a and b; Pudovkina et al., 1980; Yamnova et al., 2004; Zolotarev et al., 2008).

The end-member formulae of cation-saturated minerals with species-defining M⁴⁺ and C²⁺cations are thus as follows: kazakovite – Na₆MnTiSi₆O₁₈, koashvite – Na₆CaTiSi₆O₁₈, and zirsinalite – Na₆CaZrSi₆O₁₈. For kapustinite, the Zr-dominant member with a vacant C site, the end-member formula is Na₆ZrSi₆O₁₆(OH)₂.

If the M site is occupied by a trivalent cation (e.g., Fe³⁺) and the A and B sites are fully occupied by Na (6 apfu), then maximally 1.5 divalent cations occupy the C site. This cation arrangement is realized in imandrite with the end-member formula of Na₆Ca_1.5Fe³⁺Si₆O₁₈.

In combeite, the M site is occupied by Ca. The C and A sites are occupied by both Ca and Na, and the Na occupancy on the B site is approximately 2/3. The studied samples of two structural modifications of combeite are slightly different in their Na:Ca ratio in the A site: Ca > Na in the modification with space group R-3m, whereas Na > Ca in the modification with P3₁21 (Fischer & Tillmanns, 1983, 1987).

Beside the minerals discussed above, having their A and B sites fully or almost fully occupied by cations, the zirsinalite–lovozerite subgroup includes members with halved Na contents, namely lovozerite, litvinskite and tisinalite (Table 1¹). In these minerals, charge balance is controlled by the O²⁻ ↔ (OH)⁻ substitution.

In all known synthetic lovozerite-type compounds, the sum of the A and B cations is 6 apfu (i.e., all A and B sites are fully occupied), except in Na₄Ca₄Si₆O₁₈ (Ohsato et al., 1986) which contains 5 (Na + Ca) apfu, due to the vacancies in the B site (3 Ca apfu occupy the C and M sites). There are no synthetic lovozerite-type phases with lower occupancies of the A and B sites.

Lovozerite, the “forefather” of the group, was discovered in surface-exposed peralkaline rocks of the Lovozero intrusion (Gerasimovsky, 1940). This mineral and its C-vacant analogue litvinskite were later found at numerous places in the Khibiny–Lovozero complex, in outcrops of peralkaline magmatic rocks and pegmatites (Semenov & Razina, 1962; Semenov, 1972; Kapustin et al., 1973; Khomyakov, 1995; Pekov, 2005; etc.). Unlike these cation-deficient minerals, the cation-saturated members of the group (zirsinalite, kapustinite, kazakovite, koashvite, and imandrite) are only known from deeper levels in the Khibiny–Lovozero complex, they are found in “fresh” samples from mine workings and drill cores.

Immediately after the discoveries of zirsinalite (Kapustin et al., 1974a) and kazakovite (Khomyakov et al., 1974), their instability in air was noticed: after several days in room conditions, these minerals become covered by efflorescence of natron and thermonatrite. Special experimental studies of this process were performed, and it was found that zirsinalite and kazakovite alter in air to lovozerite and “H-kazakovite” (= tisinalite), respectively. This process is complete after exposure for several weeks to moist air (Khomyakov, 1977; Khomyakov et al., 1978). Kapustinite is more stable: its partial alteration to litvinskite takes about one year. Experiments show that the transformation of kapustinite to litvinskite in hot water (80–100 °C) occurs already after several hours (Pekov et al., 2003).

Chernitsova et al. (1975) and Pyatenko et al. (1999) have demonstrated a strong violation of local charge balance on “pendent” vertices of Si tetrahedra if they are occupied by O²⁻ in lovozerite-type structures – they are characterized by a significant surplus of negative charge that can be balanced through the O²⁻ ↔ (OH)⁻ substitution. According to Chernitsova et al. (1975), the minimum number of two (OH) groups per Si ring is needed for structure stabilization. An example is kapustinite which is significantly more stable in air than (OH)-free zirsinalite.

The absence or insufficiency of (OH) groups is thus the main cause for the instability of cation-saturated members of the group under atmospheric conditions (i.e., in the presence of water). The process of their fast alteration is due to hydration accompanied by loss of some Na, according to the following scheme: Na⁺+ O²⁻ → (□, H₂O)⁰ + (OH)⁻.

The zeolite-like character of the structure and protonation of “pendent” vertices of Si tetrahedra causes an easy transformation of cation-saturated lovozerite-group (first, zirsinalite–lovozerite subgroup) members to cation-deficient species. The initial symmetry of the framework part of the structure usually remains the same, whereas the symmetry of the full structure may change. In the latter case, reduction of symmetry usually results in a non-equal distribution of residual Na, vacancies and H₂O molecules among the structural sites.

All data on the occurrence and the chemical properties of lovozerite-group minerals, including their behaviour in hydrolysis experiments (Semenov, 1972; Kapustin et al., 1973, 1974a and b, 1980; Khomyakov et al., 1974, 1978; Khomyakov, 1977, 1980, 1995; Pekov et al., 2000, 2003; Pekov, 2005), and the information on their synthetic analogues (Ilyushin et al., 1983) and related lovozerite-type compounds (Simonov et al., 1967; Otroshchenko et al., 1973; Pushcharovsky et al., 1976; Zayakina et al., 1980; Safronov et al., 1980; Ohsato et al., 1985, 1986; Tamazyan & Malinovskii, 1989) clearly demonstrate that: (1) only cation-saturated lovozerite-type compounds may nucleate from liquid (melt, solution) or fluid phases and (2) for this heterogeneous nucleation, high-temperature conditions and Na-saturated medium are necessary. Generally, a lovozerite-type framework, by analogy to other frameworks of zeolite-like structures, can probably only be formed if the zeolitic cavities are fully or almost fully saturated by large cations (for details see Belov, 1976; Barrer, 1982) or H₂O molecules.

There are no indications that cation-deficient members of the lovozerite group could be formed as primary phases, skipping the initial stage of their cation-saturated forms, in support of the conclusion by Khomyakov (1980, 1995) that all cation-deficient lovozerite-group minerals are only secondary phases. They form as the result of partial loss of Na from primary cation-saturated minerals coupled with the O²⁻ ↔ (OH)⁻ substitution. Their crystals are in all cases homo-axial pseudomorphs after crystals of the primary phases. In Nature, this transformation occurs in low-temperature hydrothermal or supergene systems.

The results of structural studies of lovozerite, litvinskite and tisinalite (Ilyukhin & Belov, 1960; Malinovskii et al., 1993; Yamnova et al., 2001a and b, 2003; Zolotarev et al., 2008) indicate that the B sites in these minerals are vacant (or contain few H₂O molecules), whereas the A sites are fully or significantly (>50 %) occupied by Na. Indeed, the B sites located in the channels (Fig. 3³) are most favourable for cation leaching. The mechanism of Na loss is shown in Table 2² on three examples.

The compositions of the cationic part of the framework, i.e. the contents of Si and the M and C cations, remain invariable or change very slightly: cation-deficient minerals inherit these from their cation-saturated precursors.

The transformation of cation-saturated LGM to cation-deficient ones is thus a regular, structurally controlled process. All data obtained show its distinct character that underlines the individuality of cation-deficient representatives of the group as mineral species: lovozerite, litvinskite, and tisinalite have been defined as the B-vacant analogues of zirsinalite, kapustinite, and kazakovite, respectively (Table 2²).

Taking into account the crystal-chemical features of the LGM discussed above, the optimal and complete calculation scheme for their empirical formulae from chemical data only (including electron-microprobe data, without determination of Fe valence and H₂O content) is as follows:

1. the calculation is based upon Si + Al = 6 apfu;

2. Zr, Ti, Nb, and Hf are placed into the M site; if their sum is < 1.00 apfu, then Fe³⁺should be added; if the sum of all these constituents is still < 1.00 apfu, then place consecutively Mg²⁺, Mn²⁺ and eventually Ca in the M site;

3. the rest of Ca, Mn, Mg and Fe should be placed into the C site (if the Fe valency is unknown, then consider Fe in the C site as Fe²⁺);

4. Na and K, Sr, Y and lanthanides should be assigned to the A site; the maximum number of cations in the A site is 3.00 apfu;

5. the rest of Na should be assigned to the B site;

6. calculate the total positive charge and adjust the O/OH ratio (to obtain charge balance);

7. if the H₂O content was determined, then add its rest (after deduction of the amount used for calculating the number of OH groups) to the formula as H₂O molecules.

The definition of a mineral species of the lovozerite group is based upon five criteria. Boundaries between species within each subgroup are established according to the well-known “dominant-constituent rule” called the “50 % rule” in binary systems (Nickel, 1992), recently extended by Hatert & Burke (2008).

The five criteria are as follows (all values are given for a formula based on Si₆ and calculated as shown above).

1. The structure type is determined by the cation arrangement of pseudocubic modules: zirsinalite–lovozerite, koashvite, imandrite or, hypothetically, any other possible lovozerite-like compound [in the case of discovery of a mineral chemically identical to an earlier-known species but belonging to a different structure type, it must be considered as an individual species; in particular, the poorly-studied phase M42 briefly reported by Khomyakov (1995), the chemical analogue of koashvite belonging to the zirsinalite– lovozerite subgroup, can be considered as a potential new species dimorphic with koashvite];

2. the predominant cation in the M site;

3. the predominant cation in the A site: in all currently known lovozerite-group minerals, it is Na, except combeite with Na ≈ Ca;

4. the predominance of a cation (Na) or vacancy (□) in the B site: if the B site is occupied by Na < 50 %, then the mineral is considered as a representative of the cation-deficient subgroup; the boundaries in the zirsinalite–lovozerite, kapustinite–litvinskite and kazakovite–tisinalite series are established as Na_1.5□_1.5 in the B site or, equivalently, as the gross amount of (Na + K …)^A+B = 4.5 apfu;

5. the predominant component in the C site: in the case of a prevailing tetravalent cation (Zr or Ti) in the M site and divalent cation in the C site, the latter site is considered as cation-dominant (cation-occupied) if the sum of the C cations is > 0.5 apfu and, correspondingly, the prevailing C cation (Ca or Mn as currently known in minerals) is considered as species-defining; if the sum of the C cations is < 0.5 apfu, then the mineral is considered as vacancy-dominant [boundaries in the presently known zirsinalite–kapustinite and lovozerite–litvinskite series are thus established as (Ca + Mn …)^C = 0.5 apfu].

The idealized end-member formulae of the lovozeritegroup minerals are given in Table 1¹, along with the generalized formulae and with the compositional limits, taking into account all reliable empirical data. Table 3³ shows the distribution of species-defining components over the structural sites in the end-members of the lovozerite-group minerals.

In the two combeite modifications known, Na and Ca contents are close to one another in both the C and A sites (Fischer & Tillmanns, 1987) and, in accordance with IMA-accepted rules for the nomenclature of solid solutions (Nickel, 1992; Hatert & Burke, 2008), the contents of the C and A sites are given as (CaNa) and (Na_1.5Ca_1.5), respectively, in the end-member formula of this mineral. End-members with “combined” contents on structural sites (similar to the heterovalent pairs NaCa, Ca□ and Mn□ in the C site, and NaCa in the A site of lovozerite-group minerals – Table 3³) were recently also used in the IMA-approved nomenclature of the hellandite group (the heterovalent pairs CaREE³⁺ in both the X and Y sites, and REE³⁺Th in the Y site); these are also characterized by complicated heterovalent isomorphism with coupled O/OH substitutions (Oberti et al., 2002). This approach has been also used in the nomenclature of amphiboles (the pair NaCa in the B site: Hawthorne & Oberti, 2006) and some other mineral groups containing end-members with valency-imposed double site occupancy produced by heterovalent substitutions (Hatert & Burke, 2008).

The H₂O amount is not considered as a species-defining criterion, similar to the nomenclature of zeolite minerals (Coombs et al., 1998). Symmetry differences (e.g., due to cation order) within the same structure type (zirsinalite– lovozerite, koashvite, imandrite or hypothetically, for each possible lovozerite-like compound) for minerals with identical end-member formulae are not considered as sufficient argument for the corresponding samples to be treated as different mineral species. This is also consistent with the nomenclature of zeolite minerals (Coombs et al., 1998). Petarasite is considered as related to the lovozerite group, but not a member of it.

The present nomenclature of the lovozerite group preserves all its valid mineral species having the IMA-status G (grandfathered) and A (approved) without modification of their names. New names are not proposed. The nomenclature scheme defines end-members and determines species names (Tables 1 and 3¹³).

For potentially new species of the lovozerite group, taking into account both general principles of mineralogical nomenclature and the individual crystal-chemical features of this group, the following rules of naming are recommended:

• – a unique name should be used for any new mineral different from the earlier-known species in its structure type understood in terms of arrangement of pseudocubic modules;

• – a unique name should be used for any new mineral different from the earlier-known species with respect to the cation predominant in the M site;

• – a unique name should be used for any new mineral different from the earlier-known species with respect to the cation predominant in the A and/or B sites;

• – a unique name should be used for any new mineral which is an analogue of the earlier-known species with vacancy prevailing in the cation sites A, B and/or C;

• – for a new mineral with occupied (not vacancy-dominant) C site different from the earlier-known species in the prevailing C cation, the use of the Levinson-type suffix could be proposed indicating this cation: e.g. if an analogue of lovozerite with Mn > Ca in the C site is discovered, then it should be named lovozerite-Mn with renaming of “old” lovozerite to lovozerite-Ca (by analogy with the nomenclature of zeolites – Coombs et al., 1998).

H-combeite (Khomyakov, 1980) = combeite;

H-kazakovite (Khomyakov, 1980) = tisinalite;

lovozerite M (Khomyakov, 1980) = lovozerite;

lovozerite T (Khomyakov, 1980) = lovozerite;

M40 (Khomyakov, 1995) = kapustinite;

mineral No. 7 (Chirvinsky: in Bonshtedt et al., 1937) = lovozerite;

monoclinic lovozerite (Kapustin et al., 1973) = lovozerite;

oxy-lovozerite (Semenov, 1972) = lovozerite;

trigonal lovozerite (Kapustin et al., 1973) = lovozerite.

We are grateful to Nikita V. Chukanov and Natalia A. Yamnova for discussion, the IMA CNMNC officers William D. Birch, Frédéric Hatert and Stuart J. Mills for their work on our nomenclature proposal, the IMA CNMNC members for their helpful comments and Ekkehart Tillmanns for his editorial work. Our studies were supported by grants RFFI nos. 09-05-00143-a and 09-05-12001-ofi_m, grant no. NSh-863.2008.5 of the Foundation of the President of the Russian Federation and a grant from the Russian Science Support Foundation (IVP).