INTRODUCTION

During Р1–Т1 period of the Hercynian cycle, the East Kazakhstan region was spanned by intense granitoid magmatism formed in a post-collisional geodynamic setting, which is typical of the Central Asian Orogenic Belt (Dyachkov, 2015). The evolution of the granitoid magmatism was accompanied by a systematic compositional change of intrusive complexes and related mineralization: 300–310-Ma Kalguta and Kunush granodiorite–plagiogranite complex (Au–Ag); 292–297 and 281–288-Ma Kalba granite complex (Ta, Nb, Li, Be, Cs); 281–288-Ma Monastyrskii leucogranite complex (Ta, Sn, REE); and 250-Ma Keregetas–Espe alkaline granite complex (Nb, Zr, REE) (Khromykh et al., 2016; Dyachkov, 2015). The Verkhnee Espe granitoid massif is an important object of East Kazakhstan, because it is practically entirely occupied by a promising complex rare-earth–rare metal (REE, Zr, Nb, Ta) deposit.

Numerous detailed studies have been carried out for the Verkhnee Espe massif. In particular, the geological position of the massif, petrography and geochemistry of rocks were reported in (e.g., Belov and Ermolov, 1996; Stepanov et al., 2011; Stepanov and Bekenova, 2009; Trace …, 2011). The composition of rock-forming and accessory minerals and geochemistry of apogranites and ore lodes of the Verkhnee Espe deposit were described in (Baisalova, 2018; Frolova, 2018). At the same time, such important and informative accessory mineral as zircon remains insufficiently studied.

Zircon possesses a unique ability to retain isotope-geochemical fingerprints of important events in the formation of rocks and their sources. It is a key mineral for dating and deciphering the genesis of ore-bearing rocks. Zircon from this massif is also of great interest as ore mineral, because has extremely high content of trace and rare-earth elements. Zircon with such specific composition is relatively scarce and usually has a fluid-assisted hydrothermal-metasomatic genesis. Magmatic zircon with extremely high contents of trace elements is an extremely scarce phenomenon. It can be exemplified by zircon from the Yastrubets and Azov syenite massifs at the Ukrainian Shield (USh) (Levashova et al., 2016). This paper is aimed at unraveling the reasons of extreme contents of non-formula elements in zircon from the Verkhnee Espe granite massif and at determining the conditions of zircon crystallization. This study, in addition, will attempt to fill the gap in isotope-geochemical data on this massif. To this end, we have studied the trace and rare-earth element composition, morphology, and structure of the zircon as compared to zircons from highly evolved rocks (alkali feldspathic syenites of the Ukrainian Shield); determined the oxygen isotope composition of the zircon, which also provides insight into conditions of its crystallization, and established the zircon age, which can be used in geodynamic reconstruction for rare-metal complexes of the region.

GEOLOGICAL STRUCTURE

The Verkhnee Espe massif and eponymous complex rare-earth–rare-metal (REE, Zr, Nb, Ta) deposit are restricted to the outercontact of the large Akzhaylyautas granite massif in the southwestern margin of the Zharma–Saur ore belt, Eastern Kazakhstan. The massif is represented by two small domes (Bol’shoi and Malyi) of alkaline granites of the Keregetas–Espe Complex (second intrusive phase of the complex) (Fig. 1). The Verkhnee Espe massif is mainly made up of inequigranular rocks varying from fine to coarse-grained and pegmatoid alkaline granites (biotite–riebeckite, riebeckite, biotite, and aegirine–riebeckite), as well as porphyritic rocks (Stepanov and Bekenova, 2009). Host rocks are sandstones, siltstones, tuffstones, and schists. The massif comprises unaltered and metasomatically altered (apogranite) granites. Most intense metasomatic alterations occur in the apical part of the massif, which led to a change of the composition and structure of primary granitic rocks and formation of albite–riebeckite granites with elevated contents of trace elements (Zr, Nb, Th, Y, REE, and Sn, Pb, Zn, Ti et al.) (Stepanov et al., 2008; Stepanov and Bekenova, 2009). At the contact with metasomatically altered alkaline granites, host rocks were transformed in fenites (Stepanov and Bekenova, 2009).

Fig. 1.
figure 1

Geological scheme of the Verkhnee Espe massif (after V.A. Belov and L.I. Semivragova). (1) Quaternary deposits; (2) sandstone, siltstone; (3) pegmatoid granite; (4) fine-grained alkaline granite; (5) medium-grained riebeckite, riebeckite–biotite granite; (6) porphyritic biotite granite; (7) granosyenite and granodiorite; (8–10) vein rocks ((8) aplitic granite; (9) alkaline granite, (10) pegmatite); (11) ore-bearing riebeckite-albitite metasomatites; (12) first-order fractures; (13) second-order fractures; asterisks in the structural scheme shows sampling localities.

Ore lodes are represented by stratal bodies confined to the near-contact zones of the massif (mainly quartz–albite, riebeckite–albite, and astrophyllite–aegirine varieties) and veinlike metasomatic bodies mainly in the domal parts of the massif. The trace element content gradually increases from the inner part of the massif to its apical part (Frolova et al., 2015). Ore mineralization is represented by fine dissemination of pyrochlore, bastnaesite, zircon, thorite, elpidite, gagarinite (main REE carrier, first discovered in ores of the Verkhnee Espe deposit), and other trace minerals (Belov and Ermolov, 1996; Stepanov and Bekenova, 2009).

We studied zircons from fine-grained riebeckite granites of the Verkhnee Espe deposit, which were sampled in the outer zone of the Malyi Dome (Fig. 1). The fine-grained riebeckite granites forming small intrusions in the inner parts of the Bol’shoi and Malyi domes practically were not subjected to metasomatic transformations and have sharp intrusive contacts with granites of the inner part of the massif. According to petrochemical data, the riebeckite granites are alkaline K–Na, moderate-Mg peraluminous rocks, which consist mainly of quartz (35–40%), microcline (25–30%), albite (10–20%), riebeckite (10–15%), and aegirine (up to 5%). Accessory minerals are represented by zircon, pyrochlore, and fluorite. The granites have a patchy structure defined by the development of mafic minerals.

METHODS

The inner structure and major-element composition of zircon, as well as the composition of mineral inclusions were studied using back-scattered electron (BSE) imaging on a JEOL JSM-6510LA scanning electron microscope equipped with an EDS JED-2200 spectrometer (IPGG RAS).

Local U–Pb zircon dating was conducted on a SHRIMP-II ion microprobe (Center of Isotope Research, VSEGEI). The U–Pb measurements were carried out using technique (Williams, 1998). The intensity of O2 primary beam was 4 nA, and the crater diameter, ~20 μm. Obtained U/Pb ratio was normalized to those of TEMORA and 91500 zircon standards. The errors are reported with 1σ error for single analyses (U/Pb ratio and age) and with 2σ error for calculated consistent ages and concordia intercepts. Concordia was plotted using an ISOPLOT/EX software (Ludwig, 2003). Prior to geochronological studies, zircon was studied with cathodoluminescence (CL) on a CamScan MX2500S electron microscope with a CLI/QUA 2 CL-detector at the CII, VSEGEI.

The REE and trace-element contents were studied on a Cameca IMS-4f ion microprobe at the Institute of Physics and Technology Yaroslavl Branch, Russian Academy of Sciences) using technique described in (Hinton and Upton, 1991; Fedotova et al., 2008). The measurement errors were 10–15% for elements with concentrations >1 ppm and 10–20% for elements with concentrations 0.1–1 ppm, the detection limit was 5–10 ppb. The crater size was approximately 20 μm. REE patterns of zircons were normalized to СI chondrite (McDonough and Sun, 1995). Crystallization temperature of zircon was calculated using Ti-in-zircon thermometer (Watson et al., 2006).

The oxygen isotope composition was analyzed on a Cameca-1280 ion microprobe (Institute of Geology and Geophysics, Chinese Academy of Sciences) in a maximally homogenous zircon domains using technique reported in (Gao et al., 2014). Primary Cs+ beam determined the size of analyzed area as a square with sides of 15–20 μm. Each analysis involved 16 measurement cycles of 18O/16O ratio. The measured 18O/16O were normalized to VSMOW standard (18O/16O = 0.0020052). The instrumental mass fractionation (IMF) was determined using measurements of TEMORA-2 zircon standard (δ18О = 8.20‰, SD = 0.18) (8 analyses) and was controlled by the independent measurement of 91500 standard (δ18О = 10.04‰, SD = 0.29) (4 analyses) implanted in the same sample.

RESULTS AND DISCUSSION

Extracted zircon is represented by euhedral grains reaching 100–250 μm across (Fig. 2). SEM-EDS data showed that the central part of the zircon contains large (up to 50–100 μm) microinclusions of albite and K-feldspar. In the BSE image, the zircon has a heterogeneous structure, which is also expressed in its trace-element composition. The crystal consists of light core and rim zones with relatively moderate trace element content and dark intermediate (between core and rim) zone with high content of trace and rare-earth elements. According to SEM-EDS data, BSE-dark intermediate zones of the zircon are peculiar in the elevated contents of FeO (up to 2–4 wt %), CaO, and Al2O3 (up to 0.5–1.5 wt %). The rim zone of zircon is broken by transverse fractures and practically devoid of admixtures.

Fig. 2.
figure 2

BSE and CL images of zircon from riebeckite granites of the Verkhnee Espe massif. White circles show the position of analysis points for trace elements and U-Pb dating. Ion probe craters are shown beyond the scale. Numbers of analysis points correspond to those in Tables 1, 2.

The core zones of the zircon are characterized by the higher total REE content (12000–17000 ppm) compared to those typical of magmatic zircon (Harley and Kelly, 2007). REE are dominated by HREE (12000–16000 ppm). The zircon also has elevated Y up to 10000–17000 ppm, as well as 800–1600 ppm P and 300–900 ppm Nb (Table 1). The core zones of zircon also have high U and Th contents: up to 1200–2000 ppm Th and 6000–8000 ppm U. The Th/U ratio is 0.16–0.33 (Table 1), which corresponds to the lower limit of values typical of magmatic zircon (Hoskin and Schaltegger, 2003). Based on Ti-in-zircon thermometer, zircon was formed at ~ 700–850°С (Table 1). The REE distribution pattern in the BSE-light central domain is highly differentiated, with HREE enrichment relative to LREE (LuN/LaN = 400–46000) at variable LREE content (Fig. 3a). The REE patterns show a pronounced negative Eu-anomaly having practically constant value (Eu/Eu* = 0.03–0.05) and a positive widely varying Ce-anomaly (Ce/Ce* = 2–88). The value of Ce-anomaly in zircon is mainly controlled by oxygen fugacity in primary magma, i.e., is inherited from a melt (Hoskin and Shaltegger, 2003), and correspondingly indicates a change of redox conditions in the latter.

Table 1. Trace-element composition (ppm) of zircon from riebeckite granites of the Verkhnee Espe massif
Fig. 3.
figure 3

REE distribution pattern in zircon from riebeckite granites of the Verkhnee Espe massif: (a) central and marginal zones of zircon; (b) intermediate zones of zircon. Numbers of points correspond to those in Fig. 2.

Worth noting is the core zone of the crystal in analysis point 2.1, which significantly differs from the above described zircon in the lowest trace element contents (Table 1). The total REE content in this point is as low as 270 ppm, Y—110 ppm, Th—7 ppm, U—36 ppm, and Th/U ratio is 0.18. The REE pattern in this point shows LREE to HREE fractionation, differing in the lowered content of these elements and a very deep negative Eu-anomaly, thus resembling a tetrad effect (Fig. 3a). Based on Ti-in-zircon thermometer, the crystallization temperature of zircon in this point is 500°С (Table 1), which is comparable with those of rare-metal granites of East Kazakhstan (Sokolova, 2014).

The trace-element composition of zircon in BSE-dark intermediate zones significantly differs. This zone is characterized by the high total REE content (14000–43600 ppm) mainly at the expense of LREE contribution (2000–11000 ppm) (Table 1). The content of other non-formula elements is also high. In particular, Y content varies from 10000 to 22500 ppm, Nb, from 1500 to 9500 ppm, Ca, from 900 to 8000 ppm, and Ti, from 130 to 700 ppm. It should be noted that ion microprobe study did not reveal microinclusions of REE phosphates (monazite, xenotime) and other accessory minerals. As suggested by some authors (Bindeman et al., 2014), the presence of rare-earth phosphate nanoinclusions invisible in electron microscope can be ignored since anomalous contents of REE, Y, and others are not supported by significant P increase. In spite of the fact that P content correlates with Y and Ca, at maximum U and Ca contents (22000 and 8000 ppm, respectively), the zircon contains only 800 ppm P. Such proportions of the considered elements excludes a “xenotime” scheme of isomorphism and entrapment of phosphate microinclusions in the analysis field. The intermediate zones are characterized by a nearly flat REE pattern with high LREE content (LuN/LaN = 38–170) (Fig. 3b), weakly expressed Ce-anomaly (Ce/Ce* = 2–5), at preserved value of negative Eu anomaly (Eu/Eu* = 0.05–0.07).

Zircon rims have light gray color in the BSE image (Fig. 2) and the lowest contents of trace elements within the grain. The Y and REE contents in them are 2000–7000 ppm and 2000–8000 ppm, respectively, which is no less than 2–3 times lower than in the core zone of the same crystal (Table 1). The content of other trace elements in zircon rims is on average by an order of magnitude lower (Th—130 ppm, U—900 ppm, P—240 ppm, Nb—160 ppm). In spite of different levels of trace and rare-earth elements, the REE patterns in rims are similar to those in their central part (Fig. 3a), bearing typical features of magmatic zircon (Hoskin and Schaltegger, 2003).

The oxygen isotope composition of zircon was analyzed in 25 points, 15 of which were preliminarily analyzed for trace and rare-earth elements (Table 1). The average δ18О accounted for 6.30‰ at minimum content of 5.83‰ and maximum of 7.16‰, which is slightly higher than mantle values (~ 5.3‰) and in general corresponds to those of zircon from Phanerozoic granitoids (Valley et al., 2005). The BSE-light core and rim zones of the zircon are characterized by relatively low δ18О values (5.83–6.55‰), whereas intermediate BSE-dark zones with elevated content of incompatible elements have slightly elevated δ18О (6.40–7.16‰).

U–Pb dating was carried out for nine zircon grains (15 analysis points). The reliable age was obtained only for points from most homogenous (according to BSE and CL imaging) zircon domains (central and marginal parts of crystals). It is impossible to date the intermediate zone with elevated trace and rare-earth element contents (analysis points 1.1 and 7.2) due to the high content of unradiogenic lead (Table 2). Of 13 analyzed zircon points, only 7 points with moderate U (20–920 ppm) and Th (4–180 ppm) contents yield nearly concordant age values. The Th/U ratio in them is 0.13–0.20, which is typical of magmatic zircon (Hoskin and Schaltegger, 2003). The concordant age is 283 ± 3 Ma (MSWD = 0.50) (Fig. 4, Table 2). U–Pb age of zircon from the alkaline granites of the Verkhnee Espe massif coincides within the error with an age of apogranites from this massif (U–Pb age of 286 ± 4 Ma) (Baisalova, 2018) and is slightly lower than values established for albitized and finitized granites (U–Pb age of 291 ± 4 Ma) (Frolova, 2018). Thus, the age of alkaline granites of the Verkhnee Espe massif, which are associated with rare-earth–rare-metal deposit, indicates the absence of a significant age gap between metasomatic process and ore formation. In addition, obtained data indicate the simultaneous formation of granites of the Verkhnee Espe massif and granitoids of the Kalba, Zharma, Preobrazhensky, and other Lower Permian complexes of the East Kazakhstan region, thus confirming (e.g., Khromykh et al., 2016) that the main stage of the granitoid magmatism in the region occurred in the Lower Permian.

Table 2. U-Pb age of zircon from riebeckite granites of the Verkhnee Espe massif
Fig. 4.
figure 4

Concordia diagram for zircon from riebeckite granites of the Verkhnee Espe massif. Ellipses and values of U-Pb age correspond to 2σ, including error in decay constants.

The “flat” REE distribution pattern established for the intermediate zones of zircon from the Verkhnee Espe granites is a characteristic feature of “hydrothermal–metasomatic” zircon (Hoskin and Schaltegger, 2003; Hoskin, 2005; Geisler et al., 2007; Harley and Kelly, 2007; Schaltegger, 2007). However, zircon from riebeckite granites of the Verkhnee Espe massif shows no additional signs of metasomatic alterations (e.g., peculiar morphology and internal structure, enrichment in non-formula elements such as Ca, Ti, Sr, and others). Evidence for metasomatic alteration are also absent in host granite. In the studied zircon, the core and marginal parts display characteristic features of magmatic zircon (Hoskin and Schaltegger, 2003; Hoskin, 2005; Geisler et al., 2007; Harley and Kelly, 2007). The REE and trace element contents in the central part of the crystals and especially in its intermediate zone are sufficiently high, while the rim zone demonstrates a decrease of all admixtures. The high REE and Y contents (up to 22000 ppm Y and up to 43000 ppm REE) as in the zircon from the Verkhnee Espe massif were found in the rim zone of zircon from alkaline–feldspathic syenites of the Yastrubets and Azov massifs (Ukrainian shield), which are associated with rich rare-earth deposits (Levashova et al., 2016). The extremely high content of these elements in zircon (up to 36000 ppm REE and 81000 ppm Y) is caused by its crystallization at the late stages of the massif formation (riebeckite–aegirine quartz syenite) from a residual fluid-saturated melt enriched in Y, REE, U, Th, Nb, and other incompatible elements.

The HREE–Y diagram (Fig. 5a) clearly demonstrates a linear trend of zircon enrichment in HREE and Y, with a clear positive correlation, which is typical of this mineral. The trend starts in data point of zircon with extremely low contents of Eu and other trace elements (analysis point 2.1) and ends in data points of BSE-dark intermediate zones, which have flat REE pattern owing to the elevated LREE content (Fig. 5a). It is pertinent to mention that the Y and HREE contents in the central domains of the studied zircon are within a narrow range, whereas the contents of these elements in the marginal zones show a wider scatter. Similar distribution of elements is also observed in other diagrams (Figs. 5b–5d). Thus, the zircons demonstrate a systematic increase of trace elements from core to intermediate zones, which crystallized from a melt enriched in trace elements. At the final stages of crystallization, the rim zones of the zircon have been formed already from incompatible element-depleted melt, which resulted in the formation of a relatively depleted rim.

Fig. 5.
figure 5

Relations of HREE–Y (a), HREE–LREE (b), Y–Nb (c), U–Th (d), La–SmN/LaN (e), and Ce/Ce*–Eu/Eu* (f) in zircon from riebeckite granites of the Verkhnee Espe massif and in zircon from syenites of the Yastrubets and Azov rare-earth deposits (UA). The content of trace elements are given in ppm. Figures 4e and 4f demonstrate compositional fields of magmatic (1) and hydrothermal-metasomatic (2) zircons after (Hoskin, 2005), while diagram La–SmN/LaN (4e) additionally shows boundaries of zones (dashed lines) according to (Kirkland et al., 2009).

The studied zircons were compared with zircons from the Yastrubets and Azov massifs, which were crystallized from a residual fluid-saturated melt (Levashova et al., 2016). Zircons from the Yastrubets and Azov massifs, which host the eponymous rare-earth deposits, define own trend (Fig. 5) parallel to the trend of zircon from riebeckite granites of the Verkhnee Espe massif. The LREE and HREE contents and distribution pattern, as well as contents of Nb (Fig. 5c) and other trace elements (Th, U, Fig. 5d) in the intermediate zones of the studied zircon are similar to those of the rim zones of zircon from the Yastrebets and Azov rare-earth deposits (Fig. 5b). It was determined for natural conditions and experimentally that such incompatible elements as Zr, REE, Y, and Nb could be mobile in the alkali- and F-rich magmatic systems (Yang et al., 2014; and references therein), and could be accumulated in a highly evolved melt (Salvi and Williams-Jones, 1996, 2006; Jiang et al., 2005), as was established for the Azov massif (Voznyak et al., 2010, 2010а; Levashova et al., 2019). The enrichment in the aforementioned elements likely is a characteristic feature of zircon from alkaline granites and syenites, which crystallized during interaction with a fluid-saturated melt.

In the LaN–SmN/LaN discriminant diagram (Fig. 5e), the majority of data points of zircon from riebeckite granites of the Verkhnee Espe massif fall in the region between distinguished fields of magmatic and hydrothermal-metasomatic zircon or in the field of magmatic zircon according to (Hoskin, 2005). Practically all data points from the core and intermediate zones of the studied zircon fall in the field of hydrothermal zircon according to (Kirkland et al., 2009), whereas data points of its rims correspond to magmatic zircon. In the Ce/Ce*–Eu/Eu* discriminant diagram (Fig. 5e), data points show similar distribution, plotting mainly beyond the distinguished fields (Hoskin, 2005). The intermediate zones of zircon with high content of non-formula elements are plotted near the “hydrothermal” field. However, as mentioned above, the signs of hydrothermal origin and metasomatic alterations in zircon were not revealed.

The absence of significant influx of “external” fluids during zircon crystallization from a melt also follows from the study of its oxygen isotope composition. The core and rim zones of zircon show low δ18О (5.83–6.55‰), while the intermediate zone has slightly higher δ18О within 6.40–7.16‰. A shift of δ18O in the melt–fluid system in a high-temperature magma is no more than 2‰ (Hoefs, 2009), whereas δ18O variations in the studied zircon account for slightly more than 1‰. This indicates that the magmatic system was closed relative to the “external” fluid with essentially different isotope composition and that the studied zircon has a magmatic genesis. Thus, the oxygen isotope composition of zircon is consistent with above assumption that zircon enriched in trace and rare-earth elements was derived from evolved melt without contribution of “external” fluid.

CONCLUSIONS

Obtained U–Pb zircon age of 283 ± 3 Ma for riebeckite granites is regarded as the age of the Verkhnee Espe massif and associated eponymous rare-earth–rare-metal deposit. These data are consistent with previously published results of geochronological studies of the massif (Baisalova, 2018; Frolova, 2018).

Zircon from riebeckite granites of the Verkhnee Espe massif is characterized by the heterogeneous internal structure and composition. Its peculiar feature is the presence of an intermediate zone extremely enriched in REE (up to 43000 ppm) and having flat REE pattern, with high contents of Y (up to 22000 ppm), Nb (up to 9000 ppm), and other non-formula elements. No signs of metasomatic influence were found in the zircon from riebeckite granites of the Verkhnee Espe massif.

The studied zircon shows an increase of trace element contents from core to the intermediate zones, which were formed from a trace-element-enriched melt. In contrast, the rim zones of the zircon crystallized from an incompatible-element depleted melt. The study of oxygen isotope composition in zircon indicates that “external” fluids was not involved in its crystallization.

In terms of Y, REE, and other trace element contents, the intermediate zones of zircon from riebeckite granites of the Verkhnee Espe massif are similar to the “anomalous” zircon from the Yastrubets and Azov syenite massifs at the Ukrainian shield (Levashov et al., 2016). Zircon from the Verkhnee Espe massif is most similar to zircon from the Yastrubets massif, which crystallized from a fluid-saturated residual melt. We suggest that the anomalous enrichment in incompatible elements is a characteristic feature of zircon from ore-bearing (Zr–REE–Y) alkaline granites and syenites.