Low mantle plume component in 370 Ma old Kola ultrabasic-alkaline-carbonatite complexes...
I. N. Tolstikhin et al.

1. Introduction

Several reasons encourage studies of trace element and isotope systematics of ultrabasic-alkaline-carbonatites Complexes (UACC). Petrologic [Le Bas, 1987; Wyllie et al., 1990], geochemical [Nelson et al., 1988] and isotopic [Grunenfelder et al., 1986; Andersen, 1987; Bell and Blenkinsop, 1989; Kwon et al., 1989] features of UACC all indicate mantle source for parent melts. Generally manifestations of alkaline and carbonatite magmatism are small; however low-viscous carbonatite- and alkaline melts could sample large volumes of mantle. Concentrations of some elements, such as Sr and REE, in alkaline and carbonatitic rocks and related melts are rather high when compared to crustal average abundances; therefore these melts are able to deliver characteristics of mantle composition and processes through the continental crust with a limited crustal contamination [Nelson et al., 1988; Woolley and Kempe, 1989124; Bell and Blenkinsop, 1989].

Studies of the UACC have led to the following results important for this contribution: 1) Petrology and trace element chemistry of UACC both indicate a metasomatically-processed volatile-element-enriched mantle source for the alkaline and carbonatite melts [Andersen, 1987; Hawkesworth et al., 1990; Kramm and Kogarko, 1994]. 2) Isotopic signatures of UACC show certain similarities with oceanic island alkaline rocks suggesting similar sources and processes [Grunenfelder et al., 1986; Nelson et al., 1988; Bell and Blenkinsop, 1989; Kwon et al., 1989].

Because of a high abundance of volatiles is required to form alkaline and especially carbonatite melts, and noble gases are the meaningful tracers of volatiles, noble gas studies of UACC appears to be quite promising. However only a few relevant papers are available [Staudacher and Allegre, 1982; Sasada et al., 1997], excepting systematic studies of the Kola UACC [Tolstikhin et al., 1985; Mitrofanov et al., 1995; Ikorsky et al., 1997, 1998; Marty et al., 1999].

Within the Kola Peninsula, the eastern segment of the Baltic shield, several pulses of the alkaline magmatism occurred from 2760 Ma till 370 Ma ago [Pushkarev, 1990; Kramm et al., 1993; Kogarko et al., 1995]. During the last (Devonian) magmatic event about 20 ultrabasic-alkaline-carbonatite intrusive Complexes, from giant Khibiny to quite small dikes, were formed [Kuharenko et al., 1965; Kogarko et al., 1995; Beard et al., 1996, 1998]. The isotopic signatures of Sr, Nd and abundances of REE indeed show rather low crustal contamination of the Devonian Kola UACC [Kramm and Kogarko, 1994; Zaitsev and Bell., 1995]. Considerable development of the UACC by later metamorphic processes has not been observed.

In this paper new noble gas and parent trace element analyses obtained for 8 Devonian Kola UACC are presented and previous available data are compiled. Solar-like isotopic signatures of trapped rare gases definitely identify a contribution from plume component similar to that observed in the most active spot of the archipelago Hawaii, the Loihi seamount. A subcontinental lithosphere source for this component is discussed, with the negative conclusion. A contribution from the lower mantle reservoir is suggested and evaluated for different models of the lower mantle fractionation-degassing history.

2. Geological Background

2.1. Alkaline Magmatism on the Kola Peninsula

fig01 Extension of alkaline-subalkaline rocks of different ages, from 2.76 till 0.36 Ga, appears to be a specific feature of the Kola Geological Province situated on the north-east segment of the Baltic Shield. The total contribution of these rocks is ~4% (four times the mean world contribution). The Kola Province consists of several Late Archaean blocks separated by Late Archaean greenstone belts and Early Proterozoic riftogenic structures (Figure 1). Hercynian activation was the last powerful magmatic event within the Kola Province. Proterozoic Granulite belt separates the Kola Province from the southward Karelian one [Kratz et al., 1978; Bertherson and Marker, 1986; Mitrofanov et al., 1995].

Tonalite trondhjemite-granodiorite gneisses (TTG) are the most ancient rocks of the Kola Province forming the basement for Early Proterozoic structures. The gneisses were originated 2.95-2.75 Ga ago. According to Sm-Nd studies of TTG and komatiites, the upper mantle of this age was substantially depleted [Timmerman and Daly, 1995; Vrevsky and Krymsky, 1997] due to melting and removal of large masses of basalt melts which serve as a parent material for the TTG gneisses [Balashov et al., 1992; Levchenkov et al., 1995].

Several massifs of subalkaline lepidomelane-ferrohastingsite granitoids were formed in the north-eastern part of the Kola Peninsula during cratonisation period of the Late Archaean tectono-magmatic stage at 2.76 pm 0.08 Ga (Figure 1). Approximately simultaneously alkaline granites formed the largest in the world province, about 3,000 square km, in the Central part of the Peninsula.

The composition of mantle sources were essentially changed in Early Proterozoic. Basic-ultrabasic layered intrusions having 2.5 - 2.4 Ga U-Pb and Sm-Nd ages show similar geochemical and isotopic characteristics, e.g., negative eNd = -1.2 to -2.3 implying an enriched reservoir [Balashov et al., 1993; Amelin et al., 1995, Amelin and Semenov, 1996]. A number of such intrusions within the Kola, the Karelia and the northern Finland suggests uprising of a large mantle plume at that time [Amelin et al., 1995]. Evolution of this plume determined sequence of tectonic events and peculiarities of mantle and crustal magmatism during Early Proterozoic. Approximately simultaneously with the layered intrusions, the first alkaline syenite Sakharyok massif was formed within the Kola Province.

The Svekofennian activization, 1800 Ma ago, gave rise to multy-phase gabro-nepheline syenite and alkaline granite intrusions, such as Gremyakha-Vyrmes, Soustova and other. Simultaneously porphyritic granitoid melts were intruded along the faults of north-eastern direction, giving rise to Litsa, Lebyazhka and other massifs with the total area of 900 km 2.

The latest intense alkaline magmatism was related to Devonian plume [Beard et al., 1996; Ikorsky et al., 1997]. At that time two giant nepheline syenite massifs, Khibiny and Lovozero, several ultabasic-alkaline-carbonatite Complexes [Kovdor, Seblyavr and other), and swarms of dykes were formed. Metasomatic enrichment of the depleted mantle material with certain elements, e.g., Rb, Sr, REE, is suggested to occur before the alkaline melts emplacement [Kramm and Kogarko, 1994]. The melts were probably originated from two or three different mantle sources and were not contaminated considerably by crustal material during ascending through the continental crust [Kramm, 1993; Zaitsev and Bell, 1995].

2.2. Devonian Ultrabasic-Alkaline-Carbonatite and Dyke Complexes

fig02 2.2.1. The giant Khibiny alkaline platon with exposed area 1327 km 2, is situated in the central part of Kola Peninsula (Figure 1) between Archaean gneisses and granitoids of the Kola block to north and volcanogenic-sedimentary rocks of the Early-Proterozoic riftogenic Imandra-Varzuga structure to south (Figure 2).

The Khibiny pluton is nepheline-syenites polyphase central-type intrusion characterised by a cone-like shape and a concentric-zoned (arc-like at horizontal plane) structure. Rocks of different intrusive phases exposed the following time-spatial sequence (from periphery to centre): (I) volcanic alkaline syenites; (II) massive khibinites; (III) trachytoid khibinites; (IV) rischorrites; (V) ijolite-urtites; this phase is of complicated structure and composition and includes a number of bodies of apatite-nepheline rocks; (VI) unregular-grained nepheline-syenites (lyavochorrites); and (VII) foyaite [Ivanova et al., 1970; Galakhov, 1975].

Olivinite, pyroxenite and jacupirangite xenoliths found within the Khibiny imply that the earliest proto-pluton could be formed by alkaline-ultrabasic rocks which were then intruded and partially assimilated by nepheline-syenite melts [Galakhov, 1975, 1978; Shpachenko and Stepanov, 1991; Balaganskaya and Savchenko, 1998].

A stock of carbonatite and ferrocarbonatite-silicate rocks (phase VIII) intruded foyaites (VII) displaying the following time sequence: phoscorites, i.e., biotite-aegirine-apatite rocks, (a), early carbonatites, calcite, with biotite and aegirine, (b), late calcite (c), and carbonate-zeolite veins (d) [Zaitsev, 1996].

Kramm and Kogarko [1994] obtained Rb-Sr rock-mineral isochrone age at 367.5 pm 5.5 Ma for nepheline-syenites and quite a similar age was found for mineral separates from Khibiny carbonatites, 366.6 pm 47 Ma [Zaitsev et al., 1997].

Table 1 comprises indexes and description of samples selected from the Khibiny massif and other Complexes.

fig03 2.2.2. The Kovdor (5.5 x 8 km size) is situated on the south-west segment of Kola Peninsula (Figure 1), gneisses of the Belomorian domain being the host rocks [Kukharenko et al., 1965; Ternovoy, 1977; Dudkin and Kirnarsky, 1994; Balaganskaya, 1994]. The Complex has a stock-like shape and concentric-zoned structure (Figure 3). Its central core is composed by olivinites which are surrounded by olivine-clinopyroxene, phlogopite-clinopyroxene, and clinopyroxene rocks including giant pegmatitic body with meter-size crystals of olivine, clinopyroxene and phlogopite. The peripheral circle consist of melilite-bearing rocks, earlier ijolite-melteigites and later ijolite-urtites.

Between olivine-clinopyroxene and ijolite rocks in the south-west part of the Complex there is 1.2 x 0.8 km size body of apatite-forsterite rocks, calcite-apatite-phlogopite-magnetite phoscorites, earlier calcite carbonatites and later dolomite ones. Further to west - south-west calcite carbonatites form a vein-stockwerk zone among fenitised host rocks of the frame. (Figure 3).

The 380pm3 Ma age was obtained applying U-Pb systematics to baddeleite from the calcite carbonatites [Bayanova et al., 1997].

fig04 2.2.3. The Seblyar Complex is situated in the north-west part of Kola Peninsula (Figure 1). The Complex is a stock-like oval body intruding Archean gneisses of the Kola domain and having the exposed area 5 x 4 km. The Complex is characterised by a concentric zoned structure (Figure 4). The core is composed by clinopyroxenites, including blocks of olivinites, and surrounded by the thin unlocked oval of nepheline clinopyroxenites and ijolite. In some parts of the Complex clinopyroxenites were transformed by pneumatholitic and autometasomatic processes into scarn-like apatite-phlogopite-garnet-amphibole rocks and apatite clinopyroxenites [Kukharenko et al., 1965; Subbotin and Mihaels, 1986].

Phoscorites and carbonatites intruded the core at four stages producing a concentric net of dykes and veins.

The 409pm5 Ma Rb-Sr rocks and minerals isochrone age was obtained for clinopyroxenite, phoscorite and carbonatite of the 1 stage, whereas U-Pb age of baddeleit from phoscorite of the 2 stage gave 378 pm 4 Ma [Gogol et al., 1998] and Rb-Sr isochrone age of carbonatite (ibid) 376 pm 6 Ma [Balaganskaya and Gogol, unpublished data]. In the following discussion age of the Complex obtained by Kramm and Kogarko [1993], 370 pm 10 Ma, is accepted.

fig05 2.2.4. The Ozernaya Varaka is situated in the central part of Kola Peninsula (Figure 1). This is a small almost isometric zoned body, 0.8 x 1 km (Figure 5), intruded biotite-plagioclase Late Archean gneisses of the Tersky block. The central core, ~25% of the exposed area, is composed by coarse-grained clinopyroxene rocks with nepheline, Ti-magnetite, perovskite, sometimes with apatite, amphibole, and schorlomite. The core is surrounded by intermediate ~0.2 km thick zone of ijolite-urtites. The western periphery segment is composed by ijolite-melteigites. Small blocks of clinopyroxenites (same as in the core) with characteristic reaction rims and breccias spots are observed within the ijolite-melteigites [Borodin, 1961; Kukharenko et al., 1965; Dudkin et al., 1980; Arzamastzeva and Arzamastzev, 1990].

Veins of aegirine-calcite and phlogopite-calcite carbonatites and dykes of ijolite-porphyres, feldspatoid syenites, monchiquites and tinguaites cut rocks of the massif and its fenitized 0.2-0.3 km thick rim. Kramm et al. [1993] presented 376 pm 3 Ma Rb-Sr isochrone age for clinopyroxenites and 370 pm 5 Ma for carbonatites of this Complex.

fig06 2.2.5. The Lesnaya Varaka Complex is situated 8 km to south-east from the Ozernaya Varaka (Figure 1) also within the north-west section of Tersky domain. The massif is a stock-like body. The most abundant rocks are olivinites covering up to 85% of the oval-like 2.5 times 4 km exposed area (Figure 6). Among olivinites ore-bearing (with Ti-magnetite and perovskite) and pure varieties are distinguished. Veins of fine-grained ijolites, coarse-grained ijolite-pegmatites, dolomite-tremolite-carbonatites, and later tinguaites dykes are observed within olivinites. In the western and southern segments olivinites are rimmed by narrow ( <0.2 km) zone of clinopyroxenites. In contact zones the host rocks are transformed into typical aegirine-feldspar fenites.

fig07 2.2.6. The Salmagorsky Complex is situated 20 km to the south-east from Ozernaya Varaka (Figure 1). A stock-like body of the Complex intruded Late Archean gneisses of the north-west segment of Tersky domain (Figure 7). Near the contact gneisses are fenitised. The Complex, having exposed area 5.5 x 6.5 km, is characterised by clearly zoned structure. Its periphery segment is composed by olivinites and clinopyroxenites, at some places with elevated concentrations of Ti-magnetite and perovskite. Internal core of the massif is composed by rocks varying in texture and composition; coarse-grained melteygites, ijolites and urtites are most abundant. In south-west part, between the core and outer zone, there are melilite-bearing rocks: melilitites, amphibole-phlogopite-melilite rocks, turjaites and other [Orlova, 1959; Kukharenko et al., 1965; Panina, 1975]. A few thin veins and small spots of early calcite and later dolomite-ankerite-calcite carbonatites with accessory pyrochlore occur within the massif, mainly in its core.

fig08 2.2.7. Vuorijarvi Complex is situated in the south-west part of the Peninsula (Figure 8) intruding host Archaean gneisses of the Belomorian Block. The size of exposed area is 3.5 x 5.5 km. Clinopyroxenites cover ~60% of the central core rimmed by 500 m thick nepheline clinopyroxenites and the periphery circle of ijolites which thickness varies from 10 to 500 m. In the eastern part a sub-vertical phoscorite stock intruded clinopyroxenites of the core. Numerical veins of calcite and dolomite carbonatites cut the phoscorites. Small single bodies of phoscorites and carbonatite veins with thickness up to 50 m and length up to 1 km are observed in various parts of the massif and within host rocks.

Rb-Sr rock-minerals isochrone age of carbonatite is 375 pm 7 Ma [Gogol and Delenitsyn, 1999].

fig09 2.2.8. The Turiy Peninsula Complex is situated on the south coast of Kola peninsula (Figure 1). The host rocks are Early-Proterozoic granitoids of the Tersky domain overlapped by Upper-Proterozoic quartzite sandstones and aleurolites (Figure 9).

The Complex includes several isometric bodies which exposed areas vary from 20 to 6 km 2, the total exposed area being ~40 km 2. These bodies are supposed to be apophyses of a single large intrusive located at least at 400 m depth. The rime, 0.2 to 1.5 km thick, of fenitised rocks surrounds the bodies ( Evdokimov, 1982; Bulakh and Ivanikov, 1984]. All the bodies show similar concentric-zoned structure. Periphery segments are composed by ijolite-melteigite rocks containing relics of nepheline clinopyroxenites. The central cores consist of melilite-bearing rocks, mainly unkompahgrites and turjaites and (less abundant) okaites and melilitolites. In the centre of the largest body phoscorites and carbonatites form up to 150 m thick veins. Rocks of the massif and host rocks are cut by dykes of olivine nephelinites, melilitites, monchiquites and other rocks.

Kramm et al. [1993] obtained 373 pm 6 Ma Rb-Sr isochrone age for ijolites and Dunworth et al. [1997] presented Rb-Sr isochrone age for phoscorites, 363 pm 3.5 Ma.

2.2.9. Dyke Complex of the Kandalaksha Gulf. More than 1000 dykes and explosion pipes are situated mainly within a 250 km long belt along the northern coast of Kandalaksha Gulf (Figure 1). According to Kukharenko [1967] the dykes do not relate to any known magmatic Complexes; probably they were subsurface roots of an explosion lava flow removed by erosion [Bulakh and Ivannikov, 1984]. Their emplacement was controlled by the Kandalaksha Graben belonging to the regional Onega-Kandalaksha paleorift.

Dykes within the western segment were investigated with some details. The dykes are characterised by 0.8 to 1.2 m thickness, up to 300 m length, and presumably north to north-east orientation (Figure 10). According to the field relationships, early and late dykes can be distinguished.

The early dykes are composed by carbonatites, montichellite kimberlites, ultrabasic lamprophyres, and monchiquites; these rocks contain lower crustal and host rock xenoliths. Conventional K-Ar dating of ultrabasic lamprophyres and montichellite kimberlites gives 360 - 368 Ma, and 40Ar - 39Ar age of carbonatites is 386-396 Ma [Beard et al., 1996, 1998].

The late dykes are represented by alkaline picrites, melanephelinites, nephelinites and alkaline syenite-porphyrites. Within the Eastern segment a number of explosive pipes are situated together with the dykes. The pipes are composed by foidites, melilitites, and olivine-phlogopite diamond-bearing kimberlites. K-Ar phlogopite ages of the kimberlites are within 337 - 384 Ma [Kalinkin et al., 1993].

3. Experimental Techniques

3.1. Rare Gas Measurements at Laboratory of Geochronology, Apatity, Russia

Two extraction lines for (1) heating and (2) milling of samples were operating. (1) After weighting 0.25 - 0.64 mm chips of rocks were wrapped in Al foil and mounted in a sample holder able to store up to 7 samples. The holder was evacuated and intermittently baked up to 200 oC for one week. The samples were sequentially dropped into a furnace and heated to a required temperature, up to 1700o C, in a Mo crucible. For the complete extraction this temperature was applied for 30 minutes. (2) For milling 0.25 - 0.64 mm size chips of a sample and several small metal milling balls were loaded in a glass ampoule which then was evacuated and sealed off. The ampoule was settled on a vibration table and milling was carried out by simultaneous vibrating and rotating [Ikorsky and Kusth, 1992]. After milling, the ampoule was mounted in an ampoule breaker which was pumped out. Then the ampoule was broken. In both cases (1 and 2) the extracted gases were admitted to an all-metal line and purified using Ti-Zr getters. He (and Ne) were separated from Ar and heavier gases using a charcoal trap cooled by liquid nitrogen.

The isotope compositions and elemental abundances of He and Ar were determined using a static mass spectrometer (MI 1201). A special trap reduced background (first of all background of hydrogen) in the chamber. By heating of Ti-Mo wire Ti was vaporised onto a metal surface of the trap which was cooled down by liquid nitrogen during He isotope analysis. The resolving power of mass-spectrometer was ~1000, allowing complete separation of 3He + from 3H + and HD +. The sensitivity for He was 5times10-5 A/Torr, allowing measurements of 4He/ 3He ratios as high as 10 8 typical of crustal samples. The sensitivity for Ar was 3times10-4 A/Torr. Artificial mixture of 3He, He from a high-pressure tank ( 4He/ 3He = 5times107 ) and air Ne, Ar, Kr and Xe was used as a standard for the calibration of the mass-spectrometer. 4He/ 3He = 6.29 times105 and 4He/ 20Ne = 47 ratios were normalised against air and verified at CRPG (Nancy).

The concentrations were determined by the peak height method with an uncertainty of ~5% (hereafter 1s is shown). Uncertainties in the 4He/ 3He ratios of sim106 and sim108 were 2% and 20%, respectively, and uncertainties in the 40Ar/ 36Ar ratios of 300 and 50,000 were 0.3% and 25%, respectively. The analytical blanks measured twice a week under exactly the same conditions as the samples, were 1times10-9, 2times10-10, and 1times10-10 cm 3 STP for 4He, 20Ne and 36Ar, respectively [Kamensky et al., 1984; Tolstikhin et al., 1992].

3.2. Rare Gas Measurements at CRPG, Nancy, France

Rare gases were extracted by vacuum crushing [Richard et al., 1996; Marty and Humbert, 1997]. About one gram of sample was loaded in a stainless steel crusher with a magnetic piston and baked under high vacuum for one night at 100oC. 500 strokes were applied on line using an external solenoid activating the piston, and the released gases were cleaned over two Ti-sponge getters cycled between 750oC and room temperature. After purification neon and argon were adsorbed on a stainless steel grid cooled at 17 K.

Helium was first analysed ( 4He on a Faraday collector, 3He using an electron multiplier and an ion counter). The mass spectrometer was adjusted for the analysis of all rare gases (electron energy of 60 eV, trap current of 200 mA), resulting in a low He sensitivity of 1.6times10-5 A/Torr, which was fortunately compensated by the generally high amounts of both He isotopes in the samples. The He isotope ratios were normalised against a secondary standard (Irénée mineral spring gas, Réunion Island, 12.41pm0.09 Ra as measured in CRPG).

After He analyses, Ne was admitted in the mass-spectrometer. To suppress interfering ions, e.g., doubly charged 40Ar, 20NeH at mass 21, and doubly charged CO 2 at mass 22, the mass spectrometer comprises two SAES@ getters at room temperature and a stainless steel finger containing active charcoal directly connected to the mass spectrometer ion source The finger was cooled down to liquid nitrogen temperature before Ne admission. Neon was then desorbed from the cryogenic trap at 40OK, admitted into the mass spectrometer, and left in the chamber for five more minutes before measurement started. The amount of neon and its isotopic composition were determined by analysing Ne isotope masses during 12 cycles. The peak heights and the isotopic ratios were extrapolated to the time when counting start. After the measurements, the blank and mass discrimination corrections were applied [Marty et al., 1998].

After Ne analysis, argon was desorbed from the cryogenic trap tuned at 850 K and admitted into the mass spectrometer, with the charcoal finger valved off. 40Ar was analysed using the Faraday collector, and 36Ar, 38Ar were analysed with electron amplification and ion counting (10 cycles). During these analyses, the 36Ar blanks were typically 4times10-12 cc STP, and therefore small in comparison to the 36Ar contents of the samples, representing only 0.2-4.0 % of the total signals at mass 36.

3.3. U, Th, K and Li Measurements

The concentrations of U and Th were measured by X-radiography in Neva Expedition, St. Petersburg, Russia. The lowest measurable concentration is about 0.5 ppm. K and Li were determined by spectrophotometry after acid attack and solution in distilled water in the Geological Institute, Apatity. The reproducibility of the analyses of these four elements is within pm 10%.

This document was generated by TeXWeb (Win32, v.1.0) on February 10, 1999.