Partial melting of carbonates in a subduction zone can lead to the formation of carbonate liquids (high-density fluids or melts), which afterwards migrate to and interact with the overlying crustal and mantle rocks and change their composition significantly. According to experimental data [1], the formation of carbonate liquids during subduction is due to the presence of alkali and water with PT parameters higher than 800–900°С and 3–4 GPa. Evidence for carbonate melting in subducted crustal rocks is provided by carbonate inclusions in the rock-forming minerals of the calc–silicate rocks of the Kokchetav massif [2, 3] with morphological signatures corresponding to those of melt inclusions. It was noted that the distribution of trace elements in the calcite of inclusions differs from that in the calcites of the matrix. On this basis it was assumed that carbonate inclusions were trapped as melts under conditions close to the peak of metamorphism (P in the range of 4.5–6 GPa; T ~ 1000°C) of the ultrahigh-pressure rocks of the Kokchetav massif [3]. However, experimental studies [4] have shown that carbonate inclusions in metamorphic rocks at P ≥ 4 GPa and T in the range of 800–1000°C can undergo melting and form a carbonate melt after their capture as mineral inclusions by the host mineral. Thus, previous studies [2–4] have shown that carbonate inclusions in the rock forming minerals of the calc-silicate rocks of the Kokchetav massif contained the carbonate melt under conditions close to the peak of metamorphism, but their genesis is still debatable.

Studies of small (5–20 µm) calcite-bearing primary fluid inclusions, as well as inclusions composed of calcite of presumably melt-related origin, in the rock-forming minerals of the ultrahigh-pressure calc-silicate rocks of the Kokchetav massif by the method of Raman spectroscopy did not reveal the presence of methane in them [5–7].

In this article we demonstrate for the first time the presence of methane in the secondary inclusions of the carbonate melt, located in the healed cracks in garnets and potassium-bearing clinopyroxenes from the ultrahigh-pressure calc-silicate rocks of the Kokchetav massif.

The Kokchetav massif is a zone of megamelange consisting of blocks subducted to a depth of 150–200 km and formed at various temperatures and pressures [8]. The samples of calc-silicate rocks studied (samples N0 and A8 of diamondiferous rocks and Gak 101 and EK of the diamond-free rocks) belong to the Kumdykol’ block and were collected in the dumps of an exploration adit on the shore of Kumdykol’ Lake [9]. The textures of the rocks are inequigranular, from medium-grained to giant-grained. The rocks are characterized by a banded structure. The primary assemblages are represented by garnet (15–40%), clinopyroxene (15–30%), calcite (5–60%), and dolomite (3–30%). The size of garnet and clinopyroxene porphyroblasts in the samples varies from 0.5 mm to 10 cm and 4 mm, respectively. Diamond, phengite, apatite, allanite, titanite, rutile, and zircon have been identified among accessory minerals. They occur in the matrix and form inclusions in garnet and clinopyroxene porphyroblasts. The size of garnet inclusions in clinopyroxene varies from 3 µm to 200 µm. The size of clinopyroxene inclusions in garnet varies from 1 µm to 3 mm. Replacement rims consisting of silicate minerals, such as clinozoisite, allanite, phlogopite, hornblende, chlorite, muscovite, and clinopyroxene, which form symplectites in some cases, are often found around garnets.

The composition of garnets in the stu-died   samples varies widely in various samples (Alm7–36Sps0–6Pyr11–53Grs26–77Andr0–9). The magnesium number (Mg# = Mg/(Mg + Fe)) of granites varies from 0.13 to 0.64. Porphyroblasts of clinopyroxene are characterized by a zonal structure. The cores of the porphyroblasts contain potassium feldspar lamellae and polyphase and fluid inclusions, which are absent in the rims of the clinopyroxene. The presence of lamellae suggests that the initial composition of clinopyroxene was enriched in K2O [10]. They formed as a result of the exsolution during decompression. The lamellae of potassium feldspar were also identified in the largest clinopyroxene inclusions in garnet. A K2O content in amounts of up to 0.75 wt % has been identified both in the cores of the porphyroblasts of pyroxene and in the inclusions of clinopyroxene in garnet. Experimental study [11] demonstrated that the crystallization of clinopyroxene with a potassium impurity in the amount of 0.5 wt % takes place under pressures higher than 3.5 GPa.

Carbonate inclusions (Fig. 1) were identified in garnet and clinopyroxene porphyroblasts of the studied samples. The inclusions vary from 100 µm to 3 mm in size, are irregularly-shaped or roundish (Figs. 1c–1e), and have a crown-shaped rim with the host mineral. The inclusions are composed predominantly of carbonate minerals (contain ≥75% calcite or dolomite). The marginal parts of the inclusions display silicate phases (up to 25%) such as allanite, phlogopite, chlorite (secondary after phlogopite), zoisite, amphibole, and muscovite (Figs. 1c–1e). Experiments [4] demonstrated melting with the formation of carbonate melt in the carbonate inclusions in garnet with a similar mineral assemblage under P ≥ 4 GPa and T in the range of 800–1000°C. To summarize, primary carbonate inclusions contained the melt under superhigh pressures. The content of FeO in the calcites of primary inclusions in garnet is up to 1.71 wt %. Allanite in the inclusions contains rare earths in the amounts of ~3.14 wt % Ce2O3, ~1.21 wt % La2O3, ~0.54 wt % Pr2O3, ~1.59 wt % Nd2O3, and up to 0.34 apfu Fe3+.

Fig. 1.
figure 1

Interrelations between minerals, specific features of morphology, and spatial distribution of primary and secondary inclusions in samples of calc-silicate rocks. (a) Garnet porphyroblast in a calcite matrix (Sample N0); (b) phlogopite and chlorite at the contacts of garnet porphyroblasts with the dolomite–calcite matrix (Sample A8); (c–e) primary carbonate inclusions in garnet porphyroblasts (samples A8, Gak 101, EK); (f–h) secondary inclusions of the carbonate melt, located in the plates that represent healed cracks confined to primary carbonate inclusions in garnet and potassium-bearing clinopyroxene; (i) map of potassium distribution in the inclusion of potassium-bearing clinopyroxene in garnet, corresponding to image (f). SCI, secondary inclusions of the carbonate melt; Aln, allanite; Chl, chlorite; Kfs, potassium feldspar; Grt, garnet; Cal, calcite; Cpx, clinopyroxene; Phl, phlogopite; Ms, muscovite; Dol, dolomite; Dia, diamond.

Full size image

Secondary inclusions, which are located in the plates that represent the healed cracks, which cross-cut garnet growth zones and K-bearing clinopyroxene inclusions in garnet (Figs. 1e–1i), are confined to the carbonate inclusions. The secondary inclusions are irregularly shaped and are 3–10 µm in size (Figs. 1e–1i). The healed cracks also display thin and flattened inclusions, affected by the unlacing process. Raman spectroscopy of the secondary inclusions demonstrated that the inclusions are composed of solid phases and do not contain liquid water. Such phase relationships attest to the fact that the secondary inclusions were entrapped like the melt. The study of secondary inclusions by scanning electron microscopy [5] and the reconstruction of the melt composition on this basis confirmed its predominantly carbonate composition (Table 1). The interpretation of the Raman spectra in the secondary inclusions of the carbonate melt led to revealing the presence of the following phases: graphite (characteristic lines at 1349–1364  cm–1 and 1579–1592 cm–1); methane (characteristic lines at 2912–2919 cm–1); muscovite (characteristic lines at 193–198, 260–267, and 706–722 cm–1, vibrations in the range of 3200–3600 cm–1, with a distinct peak at 3625–3673 cm–1, attesting to the presence of the OH group in the structure of the mineral); phlogopite (characteristic lines at 205, 550, and 677–681 cm–1; vibrations in the range of 3200–3600 cm–1 with a distinct peak at 3573–3583 cm–1); calcite (characteristic lines at 154–157, 281–283, and 1086–1089 cm–1); and dolomite (characteristic lines at 173–175, 282–296, and 1095–1097 cm–1). Fig. 2

Table 1. Representative compositions of rock forming and accessory minerals that make up the calc-silicate rocks of the Kokchetav massif: Cal, calcite; Dol, dolomite; Phl, phlogopite; Aln, allanite; Chl, chlorite; Amp, amphibole; Czo, clinozoisite; Ms, muscovite; Grt, garnet; Cpx, clinopyroxene; K-Cpx, potassium-bearing clinopyroxene as an inclusion in garnet (Sample A8); K-Cpx*, potassium-bearing clinopyroxene in the zone of a healed crack with secondary inclusions of the carbonate melt (Sample A8). L stands for the composition of the melt, reconstructed in [5] from the results of studying secondary inclusions of the carbonate melt in healed cracks
Full size table
Fig. 2.
figure 2

Raman spectra of the secondary inclusions of the carbonate melt in garnets and clinopyroxenes of the ultrahigh-pressure rocks of the Kokchetav massif. Photomicrographs of polyphase inclusions in (a–c) garnet and (d) clinopyroxene; (e) Raman spectra of the inclusions shown in (a–d), where each spectrum corresponds to the inclusion in the image opposite to it. Cpx, clinopyroxene; Cal, calcite; Grt, garnet; CH4, methane; Ms, muscovite; Phl, phlogopite; Dol, dolomite; C (Gr), graphite.

Full size image

In previous studies we conducted Raman map** of small inclusions (5–20 µm) in the rock-forming minerals of the calc-silicate rocks of the Kokchetav massif, interpreted as inclusions of the mineral-forming medium, comprising inclusions composed of calcite [5], as well as fluid and polyphase silicate inclusions [6, 7]. Considering that methane was not found in these inclusions, we believe that methane was absent in the calc-silicate rocks under conditions close to the peak of metamorphism. In this study we demonstrated for the first time that the secondary inclusions of the carbonate melt confined to healed cracks around large (100 µm–3 mm) primary carbonate inclusions in garnet contain methane. The healed cracks confined to carbonate inclusions in garnet cross-cut the inclusions of the K-bearing clinopyroxene. The analysis of the composition of clinopyroxene by X-ray spectral microanalysis, as well as map** of the K2O content in the clinopyroxene inclusions in garnet, showed the absence of differences in the K2O content between the clinopyroxene in the zone of the healed crack and in the clinopyroxene itself as an inclusion in garnet (Table 1, Figs. 1f, 1i). The high content of K2O (~0.64 wt %) in the clinopyroxene of the healed cracks indicates its crystallization under pressures higher than 3.5 GPa [11]. Consequently, it is possible to distinguish the stage of metamorphism of ultrahigh SClI (Fig. 3) corresponding to the regressive stage of metamorphism (P ≥ 3.5 GPa), during which secondary inclusions of the carbonate melt formed in the potassium-bearing clinopyroxene and in garnet. The presence of methane in carbonate melt inclusions, located in decrepitation cracks, healed at the regressive stage of metamorphism (P ≥ 3.5 GPa), along with the absence of methane in the inclusions captured under conditions close to the peak of metamorphism of calc-silicate rocks, indicates the accumulation of methane in the carbonate melt in the course of the regressive stage of metamorphism under ultrahigh pressures.

Fig. 3.
figure 3

Evolution of ultrahigh-pressure calc-silicate rocks of the Kokchetav massif. The line of the emergence of allanite in the assemblage is from [15]. The PT path of the Kokchetav massif is from [16]. The peak of metamorphism of the calc-silicate rocks of the Kokchetav massif is indicated by an asterisk. The stage corresponding to the emergence of methane in the carbonate melt and allanite in the mineral assemblage is indicated by a circle.

Full size image

The absence of methane in the inclusions captured under conditions close to the peak of metamorphism [5–7] is explained by the fact that oxygen fugacity at the peak of metamorphism, which we estimated on the basis of the DCDD buffer (dolomite + coesite/diopside + diamond) [12], is about –0.5 logarithmic units relative to the FMQ buffer [13], which is 1.5 units higher than the “water maximum” conditions [14] required for the formation of a noticeable amount of methane.

On the other hand, the appearance of methane at the stage of ultrahigh pressure metamorphism, which corresponds to the formation of the secondary inclusions of carbonate melt, indicates a decrease in oxygen fugacity by ~1.5 logarithmic units relative to the values at the peak of metamorphism. Leaving aside the assumption about the supply of the reduced substance into the system, we assume that the decrease in oxygen fugacity can also be due to an abrupt decrease in the Fe3+/Fe2+ ratio in the melt, the cause of which, in turn, can be the onset of crystallization of ferric iron concentrator minerals. The latter assumption is confirmed by the fact that the PT path crosses the field of crystallization of the minerals of the epidote (allanite) group [15], which actively deplete the coexisting Fe3+ melt (Fig. 3) at the discussed stage of metamorphism precisely.