1. Introduction
Alkaline igneous rocks are among the rarest magmatic rocks. These rocks contain either (1) modal feldspathoids or alkali amphiboles or pyroxenes or (2) normative feldspathoids or aegirine [
1]. Based on the molar ratios of Na
2O, K
2O, and CaO relative to Al
2O
3, these rocks can be subdivided into metaluminous {(Na
2O + K
2O) < Al
2O
3 < (Na
2O + K
2O + CaO)}, peraluminous {Al
2O
3 > (Na
2O + K
2O + CaO)} and peralkaline {(Na
2O + K
2O) > Al
2O
3} types [
2].
Alkaline rocks are extraordinary rich in large ion lithophile elements (
LILE), such as Na and K, and in high field strength elements (
HFSE), such as Ti, Zr, Hf, Nb, Ta, rare-earth elements (
REE), U and Th, forming economically important deposits of these elements [
3,
4]. Primary magmatic minerals of
HFSE in alkaline rocks form two main mineral associations. Minerals with relatively simple crystal structures (in terms of topological complexity of crystal structures [
5]), such as zircon/baddeleyite and titanite/perovskite, ilmenite and titanomagnetite (depending on the SiO
2 activity), which belong to the miaskite association [
3]. Complex Na-Ca-
HFSE-REE silicates (e.g., minerals of the eudialyte group, rinkite, aenigmatite, lamprophyllite, astrophyllite, etc.) form agpaitic assemblages. Hyperagpaitic assemblages include ussingite, naujakasite, steenstrupine-(Ce), members of the lovozerite and lomonosovite groups, and partly water-soluble minerals (e.g., villiaumite, natrosilite, natrophosphate, and thermonatrite). Transitional agpaitic rocks contain
HFSE minerals that are typical of all three types of alkaline rocks (for example, titanite and eudialyte or eudialyte and lovozerite).
The field occurrence of agpaitic rocks is variable. The Lovozero Alkaline Massif, together with the nearby Khibiny Massif and the Ilímaussaq Complex of Greenland, is one of the sites with the world’s largest occurrences of agpaitic nepheline syenite intrusions. The causes of the enrichment of agpaitic rocks in
HFSE and
REE are studied in detail. Agpaitic nepheline syenites form by extensive differentiation of parental mafic magmas at low oxygen fugacity [
6]. Such conditions determine the presence of CH
4-rich fluids [
7,
8,
9] instead of H
2O–CO
2 fluid mixtures typical of less reduced rock types. Since Na, Cl, and F are well soluble in water, these elements do not pass into the fluid, but remain in the melt [
10,
11]. In addition, the chlorine fluid/melt partition coefficient decreases with growth of fluorine content, and vice versa [
10]. As
HFSE and
REE have high solubilities in Na-, Cl- and F-rich melts [
12,
13], they will eventually form agpaitic minerals such as the eudialyte-group minerals (EGM), rinkite, aenigmatite, (baryto)lamprophyllite and astrophyllite. Agpaitic rocks can be extremely rich in EGM; for example, the upper part of the Lovozero Massif is composed of alkaline rocks with rock-forming EGM (up to 90 mod. %), due to which it received the name of “the Eudialyte Complex” [
14]. The EGM-rich rocks are represented here by foid syenite, foidolite, and alkaline metasomatite and pegmatite. In this article, based on the petrography of these rocks and data on the chemical composition of rock-forming minerals, we present estimations of conditions and mechanisms of the Lovozero “Eudialyte Complex” formation.
2. Geological Setting
The laccolith-type Lovozero Alkaline Massif was emplaced 360–370 million years ago [
15,
16,
17] into Archean granite gneisses covered by Devonian volcaniclastic rocks [
18]. According to geophysical studies [
19], alkaline rocks are traced to a depth of 7 km, the lower limit of their distribution is not detected. The laccolith has a size 20 × 30 km on the day surface, and about 12 × 16 km on a 5 km depth [
14]. In the upper part, the intrusion contacts with host rocks are almost vertical. On the plane, the massif has the shape of a quadrilateral with rounded corners (
Figure 1a).
The Lovozero Massif is a layered intrusion composed of two macro units: the Eudialyte Complex and the Layered Complex. The Eudialyte Complex is located at the top of the massif, accounts for 18% of its volume and is not layered. The Layered Complex, with the layering clearly manifested, occupies 77% of the volume of the massif. The elementary unit of layering here is a sequence (“rhythm” or “pack”) of alkaline rocks (from the bottom up): urtite–foyaite–lujavrite (
Figure 1b,c). In this series, there is a gradual transition from almost monomineral nepheline or nepheline-kalsilite foidolite (urtite) to leucocratic nepheline (±sodalite) syenite (foyaite), and then to lujavrite (meso- and melanocratic nepheline syenite of the malignite–shonkinite rock series). The textures of these rocks change from massive in urtite to semi-trachytoid in foyaitе and trachytoid in lujavrite due to the gradual ordering of the feldspar plates. The sequence urtite–foyaite–lujavrite is repeated regularly [
14,
20]. The contact between the underlying lujavrite and the overlying urtite is sharp and marked by rich loparite impregnation in both nepheline syenite and foidolite (50–80 cm thick loparite maligite–ijolite ore-horizons [
21,
22]). The thickness of the individual rhythm is 5 to 100 m-they lie sub-horizontally (the dip is 5–15° towards the center of the massif) and have nearly uniform thickness. The rhythms of the Layered Complex are combined into seven series, denoted by Roman numerals, and the series, in turn, are grouped into three zones. The upper zone (up to 370 m) consists of packs of urtite–foyaite–lujavrite. The middle zone, with a thickness of 640–670 m, is composed of monotonous lujavrite with sparse interlayers of foyaite. The lower zone consists predominantly of foyaite-lujavrite packs [
14]. The boundaries of the zones are the urtite horizons II (series) -5 (rhythm) and III-1 (
Figure 1c).
Other rock types of the Lovozero Massif are subordinate to their layering and form sub-horizontal layers or lenses in the Eudialyte Complex and the Layered Complex (
Figure 1a). Xenoliths of volcaniclastic rocks are ubiquitous. Unchanged xenoliths are composed of interbedded olivine basalts, basalt tuffs, tuffites, sandstones, and quartzites. However, usually these lithologies are deeply metasomatized [
18]. Numerous lenses of poikilitic nepheline and sodalite-nepheline (sometimes with nosean) syenite, as well as alkaline pegmatites and hydrothermal veins, are located within the eudialyte and layered complexes [
23,
24,
25].
The thickness of the Eudialyte Complex in different parts of the massif ranges from 100 to 800 m. According to [
14,
20], Zr-rich melts break through and overlap the previously formed Layered Complex. The cutting of the upper rhythms of the urtite–foyaite–lujavrite is visible; the plane of the contact of the complexes falls to the center of the massif, with the angle of dip increasing in the same direction from 10 to 40°. In fact, the Eudialyte Complex can be regarded as part of the giant Lovozero Eudialyte Deposit that includes several rich sites, in particular, the Karnasurt, Kedykvyrpakhk, Alluaiv, Angvundaschorr, Sengischorr and Parguaiv sites, and the Alluaiv site is the best explored of them [
4].
3. Materials and Methods
As the Alluaiv site (
Figure 2) is the best explored, it was chosen for detailed study of the Eudialyte Complex, whose thickness varies here from 280 to 350 m. There is a wide diversity of EGM-rich alkaline rocks–a dense drill grid and numerous outcrops display the rock relations and their contacts with the underlying rocks of the Layered Complex. As a whole, we used 275 thin polished sections of rocks selected from cores of 27 exploration boreholes and the day surface (
Figure 2a).
The thin polished sections were analyzed using the scanning electron microscope LEO-1450 (Carl Zeiss Microscopy, Oberkochen, Germany), with energy-dispersive microanalyzer Röntek to obtain BSE (Back Scattered Electron) images and pre-analyze all detected minerals. The chemical composition of rock-forming minerals was analyzed with a Cameca MS-46 electron microprobe (Cameca, Gennevilliers, France) operating in WDS-mode at 22 kV with beam diameter 10 µm, beam current 20–40 nA and counting times 20 s (for a peak) and 2 × 10 s (for background before and after the peak), with 5–10 counts for every element in each point. The analytical precision (reproducibility) of mineral analyses is 0.2–0.05 wt. % (2 standard deviations) for the major element and about 0.01 wt. % for impurities. The standards used, the detection limits, and the analytical accuracy values are given in
Table S1 in Supplementary Materials. The systematic errors were within the random errors.
Major elements in rocks were determined by a wet chemical analysis in the Geological Institute of KSC RAS. The accuracy limits of the wet chemical analysis are given in
Table S1 (Supplementary Materials). Cation contents were calculated using the MINAL program of D. Dolivo-Dobrovolsky [
26]. The amphibole-group mineral formulae were calculated based on O + OH + F = 24 atoms per formula unit and OH = 2 – 2Ti. The formula calculation was performed following the IMA 2012 recommendations [
27] using the Excel spreadsheet of Locock [
28]. Statistical analyses were carried out using the STATISTICA 13 [
29] and TableCurve 2.0 [
30] programs. For the statistics, resulting values of the analyses below the limit of accuracy (see
Table S1 in Supplementary Materials) were considered ten times lower than the limit. Geostatisical studies and 3D modeling were conducted with the MICROMINE 16.1 program (Micromine Pty Ltd., Perth, Australia). Interpolation was performed by ordinary kriging. The ImageJ program (US National Institutes of Health) was used to generate digital images from the thin polished sections images and determinate modal proportion of rock-forming minerals.
The mineral abbreviation used include Aeg—aegirine-(augite), Ab—albite, All—alluaivite, Ap—fluorapatite, Aq—aqualite, Di—diopside, Eud—minerals of eudialyte group (EGM), Fo—forsterite, Ilm—ilmenite, Kap—kapustinite, Kent—kentbrooksite, Ks—kalsilite, Lmp—lamprophillite, Lom—lomonosovite, Lop—loparite, Mag—magnetite, Marf—magnesioarfvedsonite, Mc—microcline(-perthite), Nph—nepheline, Ntr—natrolite, On—oneillite, Or—orthoclase, Phl—phlogopite, Pkl—parakeldyshite, Prv—perovskite, Ras—raslakite, Rct—richterite, Sdl—sodalite, Sp—sphalerite, Tas—taseqite, Ttn—titanite, Umb—umbozerite, Zir—zirsilite.
5. Discussion
If apo-basalt metasomatic rocks are not taken into consideration, the Eudialyte Complex of the Lovozero Massif is composed of igneous rocks that can be formally subdivided into nepheline syenites (shonkinite, malignite and foyaite) and foidolites {melteigite, ijolite and urtite (
Figure 2b)}. Based on the petrochemical (
Figure 6), petrographic, and mineralogical data, these rocks should be divided into other groups. The first group includes hypersolvus meso- and melanocratic rocks (shonkinite, malignite, melteigite and ijolite) containing 30 or more modal % of mafic minerals. These rocks contain “streams” of dark-colored minerals (
Figure 3a,b), have a trachytoid structure (
Figure 3a), and are enriched with EGM. The (Na
2O + K
2O)/Al
2O3 ratio for these rocks ranges from 0.83 to 1.88 (
Figure 7).
The second group includes subsolvus leucocratic rocks (foyaite, fine-grained nepheline syenite, urtite), where albite appears as primary magmatic mineral that present along with microcline-perthite (
Figure 4a,c,e). This group is characterized by an elevated phosphorus content (
Figure 6) responsible for the formation of various phosphates and silico-phosphates (
Figure 4b). Also, the leucocratic rocks contain primary minerals of the lovozerite group (
Figure 4d). Ratio (Na
2O + K
2O)/Al
2O
3 in these rocks ranges from 0.71 to 0.95 (
Figure 7).
The relationships of rock-forming minerals of meso- and melanocratic rocks suggest that felsic minerals, namely alkali feldspar and nepheline, crystallized first. The earliest mineral was alkali feldspar (microcline) because the morphology and mutual orientation of its crystals (
Figure 3a,b) indicate crystallization under conditions of free flow of magma [
39,
40]. In addition, there are inclusions of alkali feldspar in nepheline (
Figure 3c and
Figure 18b), but not vice versa. It is convenient to consider the crystallization of felsic minerals using the “petrogeny’s residua system” NaAlSiO
4-KAlSiO
4-SiO
2-H
2O [
41,
42]. Phase equilibria in this system (
Figure 18a) for “dry” conditions are defined in [
43], and for
PH2O = 1 kbar, phase equilibria can be found in [
44].
Figurative points corresponding to the rocks of the (meso-)melanocratic group form an area extending from the field of feldspar solid solution to nepheline–feldspar cotectic (
Figure 18a). Compared with the average composition of the Eudialyte Complex (red point in
Figure 18a), the normative composition of melanocratic rocks is enriched in feldspar. Alkaline feldspar crystallized first from the magmatic melt that spread laterally. Its fractionation quickly changed the composition of the melt towards cotectic and joint crystallization of nepheline-I and microcline began. The position of this cotectic was closer to the “dry” condition (
P = 1 atm) due to the initially reducing nature of alkaline magma [
7,
8,
9]. As a result, euhedral crystals of nepheline-I with small inclusions of microcline were formed (see
Figure 3c and
Figure 18b). The formation of similar nepheline–feldspar (pseudoleucite-like) intergrowths in the process of cotectic crystallization is considered, for example, in work of Davidson [
45]. Co-crystallization of nepheline-I and microcline is supported by the correlation of nepheline-I composition with Fe content in microcline (see
Figure 11a). The constant admixture of ferric iron in microcline and nepheline indicates a high Fe
3+/Fe
2+ ratio already in the early stages of the rock crystallization. The oxidized state of iron is probably due to the “alkali-ferric-iron effect” [
46] and this effect increases with decreasing temperature.
The crystallization temperature of nepheline-I is 500–775 °C ([
31],
Figure 9). The highest temperatures probably indicate liquidus state, whereas the lowest temperatures mark the transition to subsolidus hydrothermal conditions. The crystallization temperature of alkali feldspar from (meso-)melanocratic rocks is estimated below 700 °C (
Figure 19a). However, since the rocks are hypersolvus, the temperature of feldspar crystallization cannot be lower than 650 ± 10 °C [
47].
After the crystallization of microcline (and nepheline-I), the magmatic melt lost its ability to flow freely. Microcline-perthite is the only mineral with deformed crystals (see
Figure 3a), and this deformation occurred before its exsolution into microcline and albite. Long-prismatic aegirine-(augite) crystals are subparallel-oriented only in sections perpendicular to trachytoid plane. In sections that are parallel to the trachytoid plane, aegirine-(augite) crystals are irregularly oriented. Thus, after the crystallization of microcline (and nepheline-I), there was a transition from magmatic flow (suspension-like behavior) to “submagmatic flow” (flow with less than the critical amount of melt for suspension-like behavior) [
40].
In the rocks of the Lovozero Eudialyte Complex, the calcium content is low; for (meso) melanocratic rocks, the average value is 1.81 wt. % CaO [
14]. When alkali feldspar was fractionated, calcium accumulated in the melt. It is known that pure aegirine is not stable at magmatic temperatures, but even an insignificant increase in the calcium content served as a trigger for the aegirine crystallization [
53,
54]. The essence of this phenomenon is described in the work of Nolan [
55] devoted to the albite-nepheline-aegirine-diopside system. This system includes two main components of clinopyroxene solid solutions (аegirinе and diopside) and complements the region of residual petrogenetic system, which is poor in potassium and not saturated with silica. The addition of the diopside component to the clinopyroxene solution substantially changes the volume of the phase fields, and the field of clinopyroxene increases significantly. This effect can be traced to the composition of Aeg
50Di
50, after which the volume of the clinopyroxene solid solution remains almost unchanged.
As a result, after the formation of alkali feldspar crystals and almost simultaneously with nepheline-I, clinopyroxene began to crystallize. Initially, aegirine-(augite) crystallized by heterogeneous nucleation on the faces of growing nepheline crystals (but not feldspar ones). It can be assumed that the crystallization of nepheline "consumes less oxygen" (the ratio of the sum of cations to oxygen in nepheline is 3/4, and in the microcline is 2.5/4), and for the formation of aegirine-(augite) under reducing conditions, an oxygen donor is needed [
6].
The next stage of the formation of (meso-)melanocratic rocks was a large-scale crystallization of aegirine-(augite) together with nepheline-II. Nepheline-II is enriched in Qtz endmember not because of the higher temperature of its crystallization, but due to the occurrence of Fe
3+ impurity during the substitution ⬜︎
B + (Si
4+ + Fe
3+)
T ⇆ K
+B + 2Al
3+T. Therefore, nepheline-II cannot be used to estimate the temperature. Aegirine-(augite) from meso-melanocratic and leucocratic rocks are identical in Mg-(Fe
2++Mn)-Na ratio (see
Figure 12а,b). The sum Fe
2+ + Mn increases mainly due to manganese (
Table 3), i.e., all iron during the clinopyroxene crystallization was in trivalent state. Magnesium (diopside endmember) enters clinopyroxene in the minimal amounts necessary only to stabilize it in magmatic conditions. The reason is the low calcium content in the melt. All clinopyroxenes are enriched with Na and Fe
3+, and located at the top of the fractionation trend (
Figure 19b). Obviously, the rocks of the Eudialyte Complex are the most evolved (fractionated) among the alkaline rocks of the Lovozero Massif.
Composition of clinopyroxenes that form “streams” in meso-melanocratic rocks cannot be used to assess the oxygen fugacity due to the “alkali-ferric-iron effect” [
46]; increased contents of Ti, Zr, Al and Mn in the clinopyroxenes indicate their rapid crystallization [
56,
57] and absence of the zirconium and titanium minerals at that time [
58]. The antipathy of titanium and zirconium in clinopyroxenes (
Figure 12a) has been established in other alkaline complexes, for example, Mont Saint-Hilare [
59], Ilímaussaq [
60]. Larsen [
60] suggests that “…these elements competed with varying success for a limited number of lattice sites”.
The composition of amphiboles is completely correlated with the composition of coexisting clinopyroxenes (see
Figure 15), i.e., crystallization switched from clinopyroxene to amphibole, probably because of rising
PH2O. The amphiboles do not contain zirconium but include titanium (due to crystal-chemical reasons), and the Na/Ca ratio in amphiboles is on average 24, which is twice as much as in clinopyroxenes. The early formation of EGM is incredible, as can be seen, for example, by comparing the compositions of the coexisting minerals: clinopyroxenes contain elevated concentrations of manganese, while the M2 position (Fe
2+, Mn, Mg) in the EGM constantly suffers from cation deficiency (see
Figure 16). This means that the zirconium and calcium minerals, including EGM, crystallized together with amphiboles, but after clinopyroxenes. It disproves the hypothesis about early-magmatic formation of EGM [
14,
61].
In meso- and melanocratic rocks, the primary natrolite and sodalite crystallized simultaneously with amphiboles and EGM (see
Figure 3c,d). Considering that sodic amphiboles are stable only at low temperatures (below 650 °C) and pressures [
62,
63] and dehydration temperature of natrolite is 350 °C [
64], we can conclude that the primary natrolite crystallized almost simultaneously with alkaline amphiboles and EGM as a result of an increase in
PH2O. The binding of water in the composition of natrolite causes a decrease in the chlorine solubility and, consequently, the formation of sodalite. In this case, zonal natrolite–sodalite segregations appear, whose core consists of natrolite, the intermediate zone is composed of natrolite and sodalite, and the marginal zone consists of sodalite (
Figure 3d). Thus, meso-melanocratic rocks were formed at temperatures ranging from 650–700 °C to about 350 °C. As the rocks crystallize in this temperature interval, a gradual transition from an almost anhydrous
HSFE-, Fe
3+-, halogen-rich alkaline melt to the NaCl-rich water solution occurred.
Figurative points of leucocratic rocks (foyaite, urtite, fine-grained nepheline syenite) on the diagram of the “petrogeny’s residua system”, continue the field belonging to the points of meso-melanocratic rocks towards nepheline (see
Figure 18a). The field of leucocratic rocks is located near the thermal minimum, but unlike meso-melanocratic rocks, this minimum is not “dry” (point m), but corresponds to PH
2O = 1 bar (point n). Indeed, signs of simultaneous crystallization of microcline and nepheline are observed in foyaites and urtites, namely the inclusion of microcline in nepheline (see
Figure 4a) and interrelation between microcline and nepheline compositions. The presence of primary albite in these rocks together with microcline-perthite indicates a high PH
2O and low temperature (≈550 °C) of mineral crystallization [
65,
66]. Amphiboles from meso-melanocratic rocks (enriched with Ca and Al) and leucocratic rocks (enriched with Si and Na) constitute the primary magmatic trend ([
63],
Figure 14). The Mn/Fe ratio in EGM is a fractionation indicator [
38,
67], and early-magmatic EGM are invariably dominated by Fe, whereas hydrothermal EGM can be virtually Fe-free and form pure Mn endmembers. EGM from leucocratic rocks are relatively rich in manganese.
The melt, which the leucocratic rocks crystallized from, was formed in the process of fractional crystallization of the melanocratic melt enriched in Fe and
HFSE. Fractionation of the melanocratic melt proceeded in the direction of enrichment with nepheline and a decrease in the aegirine content. A similar fractionation path occurs in the Na
2O-Al
2O
3-Fe
2O
3-SiO
2 system [
68], where melt of the “ijolite” type (approximately 50% of aegirine) evolves towards “phonolitic eutectic” (approximately 10% of aegirine). Phonolite is similar in composition to a melt that is not saturated with silica at the minimum point of the “petrogeny’s residua system” NaAlSiO
4-KAlSiO
4-SiO
2-H
2O [
41]. The residual nature of leucocratic rocks is also indicated by the association of accessory minerals [
69]. Due to an excess of sodium silicate in these rocks, there are primary minerals of the lovozerite, lomonosovite, and murmanite groups. All these minerals have the highest possible percentage of sodium in the total cation number of the chemical formula.
The texture of leucocratic rocks (coarse-grained, fine-grained or porphyritic) depends on the proximity of the melt composition to nepheline–feldspar cotectic and the crystallization rate. The crystallization of subsolvus fine-grained nepheline syenite began with albite (see
Figure 4a,b), followed by the almost simultaneous formation of microcline, nepheline, aegirine, and EGM. The simultaneous crystallization is indicated by correlations between compositions of coexisting minerals (
Figure 11b). The leucocratic melt crystallized
in situ, forming lenses (
Figure 4g), layers, interlayers in the mass of melanocratic rocks. At slower cooling, the possibility of the formation of relatively large microcline and nepheline appeared (
Figure 4a,c,e,f), and then a rapid and simultaneous crystallization of the main mass of the rock occurred.
Among the alkaline rocks of the Lovozero Massif, xenoliths of basic volcaniclastic rocks are widely distributed [
18]. These rocks underwent metasomatic treatment of varying intensity and survived in the Eudialyte Complex both unchanged (see
Figure 5a,b) and actually turned into nepheline syenites (see
Figure 5c,d). In these rocks, there are all signs of a gradual increase in the intensity of alkaline metasomatism, including gradual transitions from unchanged basalt to fenite with relicts of augite surrounded by the rims of aegirine-(augite) (see
Figure 13), and characteristic intergrowth of nepheline and magnesioarfvedsonite due to their simultaneous crystallization (see
Figure 5d). This indicates the active supply of alkalis and the redistribution of calcium that localizes in fluorapatite and titanite. The duration of the thermal effect of alkaline melt on volcaniclastic rocks was small and, because of rapid cooling, feldspar remains homogeneous, but not completely exsolved into albite and ortoclase (
Figure 8).
The wide variety of zirconium phases is due to the gradual increase in alkali concentration during fenitization. Early (relative to aegirine) crystallization of magnesioarfvedsonite indicates an elevated Si concentration and relatively low alkalinity, which also leads to the formation of zircon [
59]. The relatively high fugacity of fluorine at the initial stage of fenitization [
70] is also likely to favor the early formation of zircon, as demonstrated by the simultaneous formation of fluorapatite. The crystallization of aegirine leads to an increase in alkali content relative to silicon [
71], which stabilizes parakeldyshite. EGM is formed later than parakeldyshite, at the final stage of fenitization.