INTRODUCTION

Lithospheric mantle beneath Archean continental blocks (cratons) extends to a depth approximately 220–250 km (Rudnick and Nyblade, 1999; Eaton et al., 2009) and consists mainly of ultramafic rocks and less abundant eclogites and pyroxenites (Sobolev, 1974). These rocks are available for study only as xenoliths brought to the surface by kimberlites, lamproites, and sometimes by other deep-seated magmas (lamprophyric or alkali basaltic ones). Kimberlites are high-K and low-Na olivine-rich (~50 mod %) ultramafic rocks with high СО2 and Н2О content, which are formed through partial melting at depths over 200 km (Pearson et al., 2019). They are complex hybrid rocks containing numerous mantle fragments, first of all, olivine with scarce diamonds embedded in a magmatic matrix (Mitchel et al., 2019). Mantle xenoliths in kimberlite pipes provide insight into the deepest seated horizons of lithospheric mantle. Peridotite xenoliths bearing clear evidence for mantle metasomatism were found in kimberlites worldwide (e.g., Pearson et al., 2003; Bell et al., 2005; Grégoire et al., 2008; Arndt et al., 2009; Agashev et al., 2013; Howarth et al., 2014; Pokhilenko et al., 2015).

The lithospheric mantle beneath the Siberian craton is well studied in the Udachnaya pipe area owing to the abundance of large and very fresh, practically unaltered xenoliths (Shatsky et al., 2008; Ionov et al., 2010, 2017; Doucet et al., 2012; Agashev et al., 2013; Howarth et al., 2014). The study of garnet from peridotites of the Udachnaya pipe allowed Pokhilenko et al. (2015) to propose a generalized model of metasomatic evolution of lithospheric peridotites of the Siberian Craton and to determine the compositions of metasomatic agents, main metasomatic stages, and their succession.

In 1990, the diamondiferous Nyurba and Botuoba pipes were discovered in the Nakyn kimberlite field in the central part of the Markha terrane, the Anabar tectonic province (Fig. 1). The Markha dike and Maiskoe body were found in the vicinity of these pipes (Tolstov et al., 2009). The kimberlites of the Nyurba and Botuoba pipes define the Rb–Sr isochron ages of 364 ± 5 and 364 ± 9 Ma, respectively (Agashev et al., 2001; 2004). Geological data constrain the emplacement age of the Nyurba pipe within 384–374 Ma (Kiselev et al., 2014). Both these determinations correspond to the middle Paleozoic episode of diamondiferous pipe formation at the Siberian Craton (Sun et al., 2014; Agashev et al., 2020). In terms of some mineralogical and geochemical features, kimberlites of the Nakyn field are close to group II kimberlites (micaceous orangeites) of South Africa (Agashev et al., 2001; Pokhilenko et al., 2000). Only few data are available on mantle xenoliths from kimberlites of the Nakyn field. Diamond-bearing microxenoliths found in the kimberlites of the Nyurba pipe frequently represent garnet intergrowths with diamonds (Spetsius et al., 2008; Malkovets et al., 2008). Based on the chemical composition, the majority of garnets are ascribed to eclogitic or pyroxenitic parageneses, with subordinate harzburgitic and lherzolite–wehrlitic garnets. Wide oxygen isotope variations in the eclogitic garnets point to their subduction origin together with diamonds (Spetsius et al., 2008). The peridotitic garnets are most resistant to secondary alterations and their relicts could be used to decipher the composition and metasomatic reworking of lithospheric mantle (Griffin et al., 1999а). New major and trace element data on garnets from the peridotite xenoliths in the Nyurba and Botuoba kimblerite pipes reported in this work were used to determine the composition and genesis of mantle protoliths of this region.

Fig. 1.
figure 1

Tectonic scheme of the Siberian platform basement after (Rosen et al., 2006; Koreshkova et al., 2011). Gray color shows the exposed Precambrian basement, white fields are post-Riphean sedimentary cover, diagonal shading designates Phanerozoic mobile belts. Asterisk shows the Nakyn kimberlite field.

Full size image

SAMPLES AND METHODS

The studied collection includes 37 peridotite xenoliths from the Nyurba (12 samples) and Botuoba (25 samples) kimberlite pipes. The xenoliths usually have ellipsoid shape up to 8 cm long. They were subjected to strong secondary alterations and were practically completely serpentinized. The preserved mantle minerals are only garnet and scarce globules of accessory sulfides. Garnet from 0.5 to 8 mm in size usually has equant rounded shape and retained crystallographic faceting. Kelyphitic (reaction) rims amounting up to 10 vol % are developed around most of extracted grains and along cross-cutting fractures (Fig. 2)

Fig. 2.
figure 2

Relations of garnets with products of secondary replacement in peridotite xenoliths ((a–b) transmitted light, (c–d) BSE image; (Grt) garnet; (Tlc) talc, (Srp) serpentine).

Full size image

The xenoliths yielded 181 grains of red-lilac or pink-violet garnets. Their major-element composition was analyzed on an X-ray Jeol JXA-8100 microprobe at the Sobolev Institute of Geology and Mineralogy of the Siberian Branch of the Russian Academy of Sciences at an accelerating voltage of 15 kV and a beam current of 20 nA. Natural minerals and synthetic grains were used as standards. The measurement technique is described in detail in (Lavrentiev et al., 2015).

The trace-element composition of the garnets was analyzed by mass spectrometric method using a XSERIES2 (Thermo Scientific) quadrupole inductively coupled plasma mass spectrometer coupled with a laser sampler with a wave length of 213 nm (New Wave Research, Nd:YAG solid state laser) at the Novosibirsk State University. Multielement certified standard samples of NIST 612 50 ppm glass standard and NIST 614 1 ppm glass standard were used as external standards. The Ca concentrations determined using X-ray microanalysis were used as an internal standard. The NIST 610 reference material was measured to perform instrumental drift correction. The mass spectrometer parameters involved ion lens stack voltage, reference mass calibration, and other parameters, which provided the maximum sensitivity within a full range of analyzed masses, and were optimized using 10 mg/L Mg, In, Ce, Ba, and U solution in 0.5% HNO3. The nebulizer gas flow was adjusted so that the СеО/Се and Ва2+/Ва+ were less than 3%. The element concentrations in a “dry” aerosol obtained on laser sampler were measured in a peak jumping mode. The elements were measured in a resolution mode with a peak width of 0.7 amu, and the peak intensity for each studied element was measured as the signal intensity in the central part of mass range determined for each element. The quadrupole delay time was 10 ms per each mass, with a one scan over all masses. Counting time was 90 s per point, including 30 s for determination of background values. The signal was measured simultaneously in pulsed (calculation of amount falling on ion detector) and analog modes. Data were acquired using time resolved analysis (TRA) as “time sections” in order to determine the element variations with time. Analyzed isotopes were determined simultaneously by mass and time required to attain maximum in a “time section”. The NIST 612 standards were analyzed each ten samples.

RESULTS

Major-Element Composition of Garnets

The major-element composition of the studied garnets from xenoliths shows wide variations (Table 1). In some xenoliths and within individual grains, garnet composition has no significant differences. The garnets contain from 58 to 85 mol % pyrope (Prp), from 6 to 25 mol % grossular (Gross), and from 12 to 20 mol % almandine (Alm) (Fig. 3). In terms of proportions of these components, the majority of the garnets fall in the compositional field of garnets from peridotite xenoliths from many other kimberlite deposits of the Siberian platform, in particular, from the Udachnaya pipe (Taylor and Anand, 2004).

Table 1.   Major element composition (wt %) of garnets in peridotite xenoliths from the kimberlites of the Nakyn field
Full size table
Fig. 3.
figure 3

Compositions of garnets from peridotite xenoliths in the kimberlites of the Nkyn field in the pyrope–almandine–grossular diagram. Gray color shows the field of garnet inclusions of ultrabasic paragenesis in diamonds from kimberlites of the Siberian Platform (Taylor and Anand, 2004).

Full size image

All the studied garnets have high magnesian number (Mg# 76-87) and the low contents of TiO2 (<0.46 wt %) and MnO (0.26–0.54 wt %). The CaO content varies within 2.3–6.3 wt %, except for garnet from sample AH-3/16ko with 9.8 wt % CaO. The Cr2O3 content varies from 0.7 to 11.2 wt %. Most of the studied garnets (garnets from 41 samples) demonstrate a positive CaO–Cr2O3 correlation typical of lherzolitic garnets (Fig. 4). A change of garnet composition with decreasing CaO and Cr2O3 along a lherzolitic trend reflects a clinopyroxene control (Griffin et al., 1999a; Kopylova et al., 2000). Garnets from six samples correspond to harzburgite–dunite paragenesis, but only one of them (АН8/16-3) is ascribed to the diamondiferous assemblage according to (Sobolev et al., 1973). Only one xenolith (AH3/16ko) can be ascribed to pyroxenites. Garnet in this sample has high CaO content, which in combination with low Cr2O3 content is typical of a websteritic paragenesis.

Fig. 4.
figure 4

Garnets in peridotites from kimberlites of the Nakyn field in the Cr2O3–CaO diagram (Sobolev et al., 1973).

Full size image

Trace Element Composition of Garnets

The studied garnets from peridotite xenoliths in the kimberlites of the Nakyn Field have wide variations of trace and rare-earth (REE) elements (Table 2). The trace and REE contents were analyzed in the core and rims of each garnet grains in each sample. No significant variations were found both within garnet grain and within most samples. In terms of REE contents and chondrite-normalized (McDonough and Sun, 1995) distribution patterns, they can be divided into two types.

Table 2. Trace-element composition (ppm) of garnets in peridotite xenoliths from the kimberlites of the Nakyn field
Full size table

The type 1 garnet has a sinusoidal pattern (Fig. 5a) with (Sm/Er)n > 1 (n—normalized to chondrite composition) and (Nd/Er)n >1 for garnet from sample Bt-12/16-1. In terms of major e-element composition, these garnets are ascribed to the lherzolitic and harzburgitic parageneses. Their patterns are characterized by LREE and MREE enrichment, which widely varies in different samples. The sinusoidal distribution is best expressed in the harzburgitic garnet from sample Bt13/16-2 (Sm/Er)n = 40.3 and (La/Yb)n = 0.4) and least expressed in high-Сr lherzolitic garnets from samples Bt13/16-1 and Bt14/16 ((Sm/Er)n = 1.16–1.18 and (La/Yb)n = 0.002–0.004).

Fig. 5.
figure 5

Types of CI chondrite-normalized (McDonough, Sun, 1995) REE patterns for garnets from peridotites of the Nakyn kimberlite field ( (а) Botuoba pipe (type 1), (b) Botuoba pipe (type 2), (c) Nyurba pipe (type 1), (d) Nyurba pipe (type 2). In Fig. (b), black line shows garnet from sample Bt10/16ko with fractionated REE pattern, dashed line shows garnet from sample Bt 1/16 with low REE contents typical of exsolution garnets.

Full size image

Type 2 garnets are characterized by the LREE depletion and enrichment in MREE and HREE: (Sm/Er)n < 1, (La/Yb)n = 0.001–0.008. (Fig. 4b). Thereby, such garnets are characterized by a flat MREE and HREE patterns ((Yb/Sm)n < 5). In terms of CaO and Cr2O3 contents, most garnets of this group correspond to lherzolitic paragenesis. The same REE distribution is also typical of websteritic garnet with high CaO content. Garnets in association with diamond are characterized by sinusoidal REE pattern (Shchukina et al., 2017), but harzburgitic garnet from sample Аn 8/16-3 shows the normal REE distribution at relatively low MREE and HREE contents.

Two samples have atypical REE patterns (Fig. 5b). Garnet from sample Bt10/16ko has fractionated REE pattern with successive increase of normalized ratios ((La/Yb)n = 0.005 and (Yb/Sm)n = 13.4). Garnet from sample Bt 1/16 has low REE contents typical of garnets formed by exsolution of garnet from high-temperature orthopyroxene (Shchukina et al., 2017).

DISCUSSION

Kimberlites of the Nakyn field differ from known kimberlites from other fields of the Siberian Platform (Tomshin et al., 1998; Agashev et al., 2001; Kornilova et al., 2001; Spetsius et al., 2006; Konstantinov et al., 2017) in the high content of phlogopite, low contents of incompatible elements compared to typical kimberlites, and have Sr and Nd isotope composition transitional between group-I kimberlites and orangeites (group II kimberlites). These differences could be related to the lateral mantle heterogeneity and metasomatic process, which caused geochemical reworking of upper mantle of the corresponding region. Mantle metasomatism was caused by diverse mantle fluids/melts (Andersen and Neumann 2001; Klein-BenDavid et al., 2004; Tomlinson et al., 2006; Bussweiler et al., 2018; Zedgenizov et al., 2020). It is expressed in peculiar trace-element and rare-earth element distribution in clinopyroxenes and garnets (Hoal et al., 1994; Griffin et al., 1999b; Agashev et al., 2013; Pokhilenko et al., 2015; Shchukina et al., 2017; Shchukina et al., 2019). Two types of metasomatic agents affected the lithospheric mantle: silicate melts and carbonatite melts/fluids. There are different points of view concerning the composition of the silicate metasomatic agent. Some researchers suggest that it was silicate melt close to picrite in terms of trace-element composition (Shchukina et al., 2017; Agashev, 2019), while others believe that it was close to host kimberlite (Kargin et al., 2017) or protokimberlite melt corresponding in composition to the early kimberlite portion crystallized in mantle (Bussweiler et al., 2018). It was proposed in the latter case that xenoliths from kimberlites bear information only on the mantle channel, in which the early portions of kimberlite melt were solidified, and do not reflect the structure of the entire lithospheric mantle. However, this hypothesis seems to be ambiguous, since exotic mantle xenoliths such as ilmenite peridotites, wehrlites, strongly metasomatized sheared peridotites, and, especially scarce MARID (Mica–Amphibole–Rutile–Ilmenite–Diopside) and PIC (Phlogopite–Ilmenite–Clinopyroxene) rocks in combination account for no more than 10% of all kimberlite xenoliths (Sobolev, 1974). In addition, it is reasonable to expect in this case that xenoliths represent the infill of this channel, i.e., failed kimberlite.

It was believed for a long time that one of the main evidences for metasomatic alteration of lithospheric mantle is the presence of phlogopite in mantle peridotites (Carswell 1980; Grégoire et al., 2003; Harte, 1983; Hawkesworth et al., 1990; Kopylova et al., 1999). So-called “phlogopite” metasomatism occurred in mantle at relatively low temperatures (Т < 1100°C), as was exemplified by peridotites from the Wesselton pipe, South Africa (Griffin et al., 1999b). However, the study of the unaltered peridotite xenoliths from the Udachnaya pipe (Agashev et al., 2013; Doucet et al., 2013) showed that metasomatic enrichment of minerals is not necessarily related to the appearance of modal phlogopite, which is either absent or forms interstitial submicron phases (Agashev et al., 2013). In addition, so-called “melt” metasomatism occurred at higher temperatures. According to data presented in (Griffin et al., 1999b; Shchukina et al., 2017), the contents of trace elements, in particular, Zr, Y, and Ti, reflect specifics of garnet formation and bear signs of both partial melting and metasomatic enrichment of parental mantle rocks. A large amount of new geochemical data collected during the past decade showed that garnets with clear signs of metasomaic enrichment, in particular, with clearly expressed sinusoidal REE distribution pattern, fall in the field of depleted garnets according to classification (Griffin et al., 1999b). Half of lherzolitic garnets from sheared peridotites of the Udachnaya pipe (Agashev et al., 2013), i.e., rocks that were unambiguously subjected to intense metasomatism, and harzburgitic garnets from this pipe, which experienced carbonatite metasomatism, also fall in the field of depleted garnets according to classification (Griffin et al., 1999b). The compositional field of garnets from depleted rocks, which show geochemical signs of partial melting and could represent a residue after primitive mantle melting, was distinguished for the first time in (Shchukin at al., 2015). This field only partially overlaps the depleted garnet field distinguished by (Griffin, 1999b). Therefore, to characterize the compositional modification of garnet by metasomatic agents, we used classification proposed in (Shchukina et al., 2017; Agashev et al., 2018).

In terms of Zr–Y relations, all garnets in the xenoliths of the Nakyn kimberlite field correspond to metasomatized peridotites according to classification (Griffin, 1999b), except for one sample falling in the field of garnets from depleted mantle peridotites (Fig. 6). The majority of type 1 garnets fall in the field of garnets enriched by carbonatite/fluid metasomatism. The same field comprises harzburgitic garnet of diamond association ascribed to type 2 based on the REE distribution pattern. The type 1 garnets were likely formed in two stages. The first stage was responsible for the crystallization of harzburgitic garnets via reaction Opx + Sp = Ol + Grt (MacGregor, 1964). The appearance of forsterite association with garnet, orthopyroxene, and spinel is caused by the transition from spinel to garnet peridotites. It was shown that this reaction is assisted by high LREE/HREE fluid (Agashev et al., 2013; Chepurov et al., 2019). In the MgO–Al2O3–SiO2 system, this reaction is monovariant. It was shown that the field of spinel-bearing association in the CaO–MgO–Al2O3–Cr2O3–SiO2 system is displaced in the high-pressure region proportional to the bulk chromium content (MacGregor, 1970). In the quaternary Cr-bearing system, the equilibrium curve with increasing the number of degrees of freedom is transformed into the four-phase field Grt + Ol + Opx + Sp (Turkin and Sobolev, 2009).

Fig. 6.
figure 6

Y–Zr diagram (modified after (Griffin et al., 1999b)) for garnets from peridotites of the Nakyn field compared to the garnets from peridotites of the Udachnaya pipe (Agashev et al., 2013; Pokhilenko et al., 2015). Compositional fields of garnets and types of metasomatism according to (Agashev et al., 2018). Double line distinguishes the field of garnets from depleted mantle rocks according to (Griffin et al., 1999b). (1) trend of garnets subjected to carbonatite (fluid) metasomatism; (2) trend of garnet subjected to silicate (melt) metasomatism.

Full size image

Continuing supply and fractional crystallization of metasomatic agent with formation of garnet and clinopyroxene led to the enrichment of harzburgitic garnets in Са, Y, and HREE (trend 1 in Fig. 6) and finally to the transition of harzburgite into lherzolite. Two type 1 garnets from the Botuoba pipe have the elevated Y content (>12 ppm), which could be caused by subsequent overprinting by silicate metasomatism. Type 2 garnets were likely subjected to the silicate metasomatism and their compositions fall in the corresponding region in the Zr–Y diagram. Silicate metasomatism that produced REE patterns of type 2 garnets likely developed after garnets of initially different compositions: garnets from depleted rocks (trend 2 in Fig. 6) or type 1 garnets that previously experienced carbonatite metasomatism (trend 1). The formation of harzburgitic garnets as well as stages and sequence of their evolution along trend 1 are well consistent with a model of metasomatic evolution of peridotite garnets proposed in (Pokhilenko et al., 2015).

The formation of garnets and clinopyroxenes in sheared peridotites of the Udachnaya pipe (Agashev et al., 2010) is also thought to be related to the influence of carbonatite and silicate melts. Thereby, the influence of carbonatite melt is considered as the earliest stage of mantle metasomatism (Zedgenizov et al., 2020), which caused the formation of garnets with a clear sinusoidal REE pattern. According to inclusion data (Stachel and Harris, 2008), these garnets most frequently associate with diamond (Stachel and Harris, 2008). Therefore, garnets with such REE patterns could be ascribed to a group of potentially diamondiferous peridotites. Lherzolitic garnets with weakly expressed sinusoidal REE patterns also occur as inclusions in kimberlite diamonds worldwide (Stachel et al., 2004). It is more likely that garnets with less expressed sinusoidal patterns experienced stronger metasomatic alteration, which is expressed in their enrichment in HREE and depletion in LREE.

Lherzolitic garnets with a normal REE distribution pattern are usually not associated with diamonds (except for Birim area, Stachel et al., 2004) because they are in equilibrium with silicate metasomatic agent, which is unfavorable for diamond survival (Agashev et al., 2018b). The metasomatic agent likely becomes high-temperature and acquires high Fe3+/Fetot (Shchukina et al., 2015). According to the experimental data, such melts efficiently oxidize diamond (Rohrbach and Schmidt, 2011; Bataleva et al., 2012) and could cause a complete diamond annihilation when entering the lithospheric mantle.

Trace-element ratios can be used to estimate the intensity of metasomatic reworking of lithosphere mantle beneath the Nakyn kimberlite field (Figs. 7а, 7b). Figure 7a shows the evolutionary path of garnet from harzburgitic through lherzolitic with sinusoidal REE pattern to the normal lherzolitic garnets for the peridotites of the Udachnaya pipe (trend 1 Ud) and Nakyn (trend 1 Nk) fields. All garnets from the Nakyn field, except for one sample, have Ti/Eu ratio below 8000, which suggests a moderate reworking of lithospheric mantle peridotites by silicate melts compared to the mantle in the Udachnaya pipe area. Moreover, some samples with type 2 lherzolitic garnets have low Ti/Eu (500–1300) and Zr/Y (0.1–0.7) ratios, which indicates a weak metasomatic impact. These garnets in the Zr–Y diagram (Fig. 6) fall closely to the depleted garnet field and were likely formed after residual garnets in the course of the insignificant interaction with silicate metasomatic agents. The type 1 garnets from both pipes have elevated (Sm/Er)n and Zr/Y ratios, which serve as evidence for the carbonatite/fluid metasomatism. The intensity of this type of metasomatism in the garnets of the Nakyn field is also less expressed than in the garnets from the Udachnaya pipe. In particular, only five samples have the well expressed sinusoidal REE distribution pattern ((Sm/Er)n > 2) or Zr/Y > 5, whereas the composition of other garnets of this type marks the final stage of carbonatite metasomatism by already strongly fractionated metasomatic agent depleted in MREE and enriched in HREE and Y. Obtained results indicate a moderate metasomatic reworking of lithospheric mantle beneath the Nakyn kimberlite field, which is consistent with a weak previously established incompatible element enrichment of the Nakyn kimberlites compared to the typical kimberlites (Agashev et al., 2001).

Fig. 7.
figure 7

Eidences for metasomatic reworking of the studied garnets in the (Sm/Er)n–Ti/Eu (a) and Zr/Y–Ti/Eu (b) diagrams. Shown are evolutionary paths of garnets from harzburgitic through lherzolitic with sinusoidal REE distribution pattern to normal lherzolitic garnets from peridotites of the Udachnaya pipe (trend 1 Ud) and Nakyn field (trend 1 Nk). Symbols are shown in Fig. 6.

Full size image

CONCLUSIONS

In terms of chemical composition of relict garnets, serpentinized peridotite xenoliths from the Botuoba and Nyurba pipes are classified as lherzolites (30 samples), harzburgites (6 samples), and websterites (1 sample). In terms of trace and rare-earth element composition and shape of chondrite-normalized REE patterns, the garnets are divided into two types. The first type is characterized by the sinusoidal REE pattern ((Sm/Er)n > 1), which is more expressed in the harzburgitic garnets and less expressed in the lherzolitic garnets. The second type of garnets has the normal REE distribution and, respectively, (Sm/Er)n < 1. It is represented by lherzolitic garnets, one harzburgitic garnet, and one websteritic garnet. In terms of trace-element relations, all garnets from xenoliths of the Nakyn kimberlite field correspond to metasomatized peridotites (Agashev et al., 2018), except for one sample falling in the field of depleted garnets. It is suggested that the formation of type 1 garnets was assisted by carbonatite/fluid metasomatism, while type 2 garnets were formed in the presence of silicate metasomatic agents.

The type 1 garnets were formed in two stages. The first stage produced harzburgitic garnets through the transition from spinel to garnet peridotites. It was shown that this reaction in the MgO–Al2O3–SiO2 system can be written as Opx + Sp = Ol + Grt and proceeds in the presence of high LREE/HREE fluid. The impact of carbonatite metasomatism on harzburgitic garnets at the second stage led to the formation of lherzolitic garnets with sinusoidal REE pattern. Metasomatism that was responsible for the formation of REE patterns of type 2 garnets was developed after garnets of different initial compositions: (1) depleted garnets, (2) type 1 garnets previously subjected to carbonatite metasomatism. Obtained results however, indicate a moderate metasomatic reworking of lithosphere mantle beneath the Nakyn kimberlite field compared to the typical kimberlites and their high diamond grade.