RUSSIAN JOURNAL OF EARTH SCIENCES VOL. 7, ES4001, doi:10.2205/2005ES000181, 2005

General Charactrization of the Rocks

[10]  The rock samples discussed in this paper were dredged mostly from the slopes of the Markov depression (sites I1032, I1060, I1063, and I1069), and basalts were uplifted from a volcanic plateaus north of it (sites I1026, I1027, and I1072). The dredged material included both fresh and variably altered rock material. The completely metamorphosed rocks were classified into (a) those developing after ultramafic rocks, gabbro, and dolerites and (b) metamorphic rocks.

Petrography

2005ES000181-fig03
Figure 3
2005ES000181-fig04
Figure 4
2005ES000181-fig05
Figure 5
2005ES000181-fig06
Figure 6
2005ES000181-fig07
Figure 7
[11]  According to their petrographic characteristics, the plutonic rocks can be subdivided into two groups: (i) primitive magnesian troctolites (Figure 3) and gabbro and (ii) gabbronorites and gabbronorite-diorites containing, Opx, Fe-Ti oxides, and kaersutite (Figure 4) and associated with norites, pyroxenites, granodiorites and plagiogranites (Figures 5 and Figure 6). As was mentioned above, the latter rocks occur mostly as veins and dikes up to 10-20 cm thick, most often 1-5 cm, and usually contain kaersutite and ilmenite. These rocks seem to typical occur in slow-spreading ridges as such bodies, because they were also found as analogous bodies in other MAR segments [Silantyev, 1998] and in the Southwest Indian Ridge [Dick et al., 1991; Holm, 2002; Ozawa et al., 1991]. These peridotites are sometimes notably different from mantle residues in bearing relics of cumulus textures (Figure 7) and in having different compositions of minerals, particularly chromite (see below), which suggests that the rocks are intrusive.

[12]  The dredged dolerite samples are fairly diverse: they include common olivine-clinopyroxene and olivine-free varieties, sometimes with elevated contents of ilmenite (up to 3 vol %), and plagioclase-phyric Ilm- Hbl dolerites. Along with fresh rocks, the dredged material contained altered varieties, whose pyroxenes and hornblende are partly or completely replaced by fibrous actinolite.

[13]  Both the fresh and the altered basalts can be aphyric or porphyritic, with phenocrysts of Ol, Ol+Pl, or Pl alone. The groundmass of these basalts is usually ophitic or has an unusual texture with sheaves of recrystallized spherulites, which are clearly seen in glassy varieties. The latter type of the basalts usually bears abundant small ilmenite grains, a feature suggesting that these rocks are volcanic analogues of the second group of the plutonic rocks. The rocks often exhibit traces of high-temperature cataclasis with the development of deformation structures in magmatic minerals (Ol, Opx, Pl, and Hbl)1 and the appearance of small subequant neoblasts. The genesis of such high-temperature cataclasites was likely related to deformations in already solid but still hot rocks, because the minerals show no traces of low-temperature alterations, and the composition of the neoblasts is close to that of the cataclasts.

[14]  Rocks of particular interest are the volcanic breccias that were cataclased under such conditions and were found among rocks from sites I1060 and I1063. Boudin-, oval-, and lens-shaped fragments (15-20 cm long) in these rocks consist of cataclased peridotites with relict cumulus textures (harzburgites and lherzolites) that were not affected by low-temperature alterations and are cemented by weakly cataclased fine-grained porphyritic rocks, whose composition varies from gabbronorite-diorite to granodiorite with ilmenite and magmatic kaersutite. It is worth noting that fragments (boudins) of these ultramafics bear practically no traces of low-temperature alterations at contacts with the diorite-granodiorite cement, as is also typical of relations between syngenetic vein derivatives and their host cumulates.

[15]  The low-temperature alterations are spread much more broadly: the peridotites are usually extensively serpentinized, and the gabbroids are amphibolized with the development of fibrous actinolite after pyroxenes and pargasite. The rocks are commonly cut by veinlets of carbonate, prehnite and chlorite. In places, the rocks were also affected by low-temperature shearing and brecciation. The thickest of these zones are associated with the development of diverse metamorphic rocks, which often have disseminated-stringer or massive sulfide mineralization [Pushcharovskii et al., 2002].

[16]  In order to characterize the typical varieties of the rocks, we conducted their detailed petrological and geochemical examination, whose results are summarized below. The petrography of the rocks is briefly characterized in Table 1, and microprobe analyses of their minerals are presented in Tables 2, 3, 4, 5, 6, and 7.

Chemistry of Major Rock-Forming Minerals

[17]  The mineralogy of many of our samples is quite unusual and is generally atypical of both mantle rocks and classic cumulates in continental layered intrusions. First of all, the chemistry of the same minerals varies even within a single thin section, and mineral crystals bear "seed'' inclusions that are uncommon in these mineral assemblages.

[18]  The composition of olivine in the rocks broadly varies from Fo91-85 in the peridotites and troctolites to Fo26 in the olivine gabbro-diorites (Table 2). A sample of the Fe-Ti-oxide olivine dolerite contains zonal olivine crystals with Fo82 in the cores and Fo62 in their peripheral portions. The composition of the olivine microphenocrysts in olivine plagioclase-phyric basalt I1072/1 corresponds to Fo87, i.e., is close to the composition of olivine in the primitive troctolite. An unusual feature of olivine in the peridotites with relics of cumulate textures is its variable composition: the predominant grains have the composition Fo86, and occasional grains correspond to Fo70 (sample I1063/17, Table 2).

2005ES000181-fig08
Figure 8
[19]  The composition of pyroxenes from the rocks is demonstrated in Table 3 and Figure 5. A noteworthy feature of the rocks is the presence of unexsolved subcalcic augite (pigeonite-augite). Such pyroxene was found even in the ultramafic rocks (sample I1063/2), a feature absolutely atypical of mantle rocks. It is also pertinent to mention that the composition of the pyroxenes varies even within a single thin section (Figure 8, Table 3).

2005ES000181-fig09
Figure 9
[20]  The composition of the plagioclase varies within broad ranges, from bytownite An80 to oligoclase-albite An10. Basalt I1072/1 contains plagioclase xenocrysts of composition An76-78 and microcrysts of composition An71-73. A notable feature of plagioclase in these rocks is its very low K contents (Table 4, Figure 9).

[21]  Magmatic kaersutitic brown hornblende occurs in the gabbroids both as a cumulus mineral and as intercumulus grains, and the dolerites contain small kaersutite grains in the groundmass. Similarly to the pyroxenes, the hornblende is often replaced by fibrous actinolite. The composition of the hornblende varies from high-Ti varieties in the hornblende gabbroids to relatively low-Ti varieties in the diorites and quartz diorites (Table 5). According to the results of Pl- Am thermometry [Blundy and Holland, 1990], the rock crystallized at temperatures from 800o C (Fe-Ti-oxide hornblende gabbronorite L1124/13-11) to 650-580o C Fe-Ti-oxide gabbro-diorite and quartz diorite) (unpublished data of S. S. Abramov).

[22]  The apatite contains small amounts of Cl (0.09-0.24 wt%). Fluorine was not determined in this mineral because of methodical difficulties.

[23]  The oxide minerals (Cr-spinel, titanomagnetite, and ilmenite) are particularly interesting as exhibiting some noteworthy structural features.

2005ES000181-fig10
Figure 10
2005ES000181-fig11
Figure 11
[24]  A remarkable feature of Cr-spinel in our samples of the cataclased peridotites with relics of cumulate textures is its broad chemical variations (Table 6, Figure 10), which are atypical of this mineral in mantle rocks but is typical of Cr-spinel in the cumulates of layered intrusions. Moreover, some chromite grains have corroded cores (Figure 11) consisting of Fe-rich chromite and surrounded by more magnesian peripheral zones, which contain P and Cl (Table 6).

2005ES000181-fig12
Figure 12
[25]  It is also worth mentioning the development of rims of fine-grained plagioclase An 55.3 around a resorbed chromite grain (Figure 12), whose composition is close to that of the zonal chromite grain described above in the same thin section (Table 4, analysis I1063/17-14). The outer part of this rim consists of fine clinopyroxene grains, and the grain itself is located among olivine neoblasts. Similar rims often develop around chromite grains during high-temperature and low-pressure transformations in ultramafics occurring in ophiolitic associations [Laz'ko and Sharkov, 1988].

[26]  The composition of most chromite grains in this thin section and the chemistry of the vermicular aggregates among olivine and pyroxene neoblasts, as well as the composition of chromite in the rim of the zonal crystal, are characterized by very low Fe3+ contents and variable Cr/Al ratios. As in this rock, Cr-spinels in harzburgite I1063/2 can be subdivided into two groups: one close to the predominant type of grains in sample I1063/17, and the other characterized by low Al2O3 contents and high concentrations of Cr2O3.

[27]  All of these data evidently indicate that the early Fe-rich chromite with elevated TiO2 and relatively low Al2O3 and MgO contents are xenogenic for these peridotites. They could be preserved as inclusions in other minerals (when captured by growing crystals), but more commonly they reacted with the magmatic melt and were dissolved in it. Likely evidence of this processes is displayed in Figure 12.

2005ES000181-fig13
Figure 13
[28]  The Fe-Ti oxides of the rocks are unusual. For example, titanomagnetite with exsolution ilmenite lamellae, which are typical of continental mafic intrusions [Bogatikov, 1996; Wager and Brown, 1968], is contained in the gabbronorites only in the form of small inclusions in clinopyroxene. These oxides ( Mag containing < 10 wt % TiO2 and Ilm) usually crystallize together or compose individual grains (Table 7, Figure 13). According to Fershtater et al. [2001], this mode of occurrence of Fe-Ti oxides is typical of rocks that crystallized from hydrous melts at relatively low temperatures (600-800o C, - log fO2 = 8-11).

[29]  As obviously follows from all of these data, the rocks contain minerals belonging to at least two distinct assemblages. Relics of the earlier of them are contained as inclusions ("seeds'') in the typical minerals of the magmatic assemblage itself. These are Fe-rich Cht overgrown by rims of a more magnesian and less titanian variety in the peridotite (sample I1063/17) and Ti- Mag with exsolution lamellae in clinopyroxene from the gabbronorite (sample I1063/1). Moreover, the composition of the olivine significantly varies, and volumetrically predominant magnesian varieties are in places associated with moderately magnesian grains. All of these facts testify to very unusual conditions under which the rocks were produced.

[30]  A noteworthy feature of the rocks is the variable composition of the same minerals in them. This can be illustrated most glaringly by the example of the peridotites wit relict cumulus textures: a single sample of these rocks can contain olivine of the composition Fo86 and Fo70 (Table 2, sample I1063/17). The broadest compositional variations in this sample and in I1063/2 are exhibited by their Cr-spinel (Table 6, Figure 10). Some of their grains are Fe-rich, which is atypical of mantle rocks but is characteristic of ultramafic layered complexes [Sharkov et al., 2001]. Moreover, some of the chromite grains contain corroded cores of Fe-rich chromite overgrown by more magnesian outer zones, which contain P and Cl (Figure 11, Table 6).

[31]  As can be seen from Table 8, the Fe-Ti-oxide gabbronorites crystallized, according to Ilm- Mag geothermometry, at average temperatures of 621 pm 41o C, which is close to the analogous values obtained by the same method for Fe-Ti-oxide gabbroids in the Southwest Indian Ridge [Natland et al., 1991].

U-Pb Age of the Rocks

[32]  As was mentioned above, the Fe-Ti-oxide rocks contain zircon. Zircon for radiological dating was separated from one of our rock samples (cataclased hornblende gabbronorite, sample I1028/1, dredged from the rift valley wall at a depth of close to 4 km). The rock was dominated by clinopyroxene and plagioclase and contained subordinate amounts of orthopyroxene, magmatic brown hastingsitic hornblende, ilmenite, and accessory apatite and zircon. Cataclasis resulted in this rock in the deformations of large crystals (cloudy extinction and deformed twinning boundaries in plagioclase and bent exsolution lamellae in pyroxenes) and the development of small subequant neoblasts. The zircon was also cataclased and disintegrated into small fragments. The secondary alterations were generally insignificant and involved the replacement of mafic minerals by fibrous actinolite.

[33]  Analogous rocks were dredged from the walls of other parts of the Markov depression and neighboring depressions in this MAR segment, which testifies to their broad in-situ occurrence. Zircon grains were obtained from sample I1028/1 using magnetic separation and heavy liquids, and by hand-picking under a binocular magnifier. The grains were classified according to their size, color, and morphology. Approximately 70-75% of each population were utilized to prepare epoxy pellets, which were polished and examined under an electron microscope with a cathode-luminescence detector. The rest of the populations were dated (TIMS) by the U-Pb method [Sharkov et al., 2004a].

2005ES000181-fig14
Figure 14
[34]  The largest fraction ( ge 450 times 250  m m) consisted of colorless transparent and semitransparent zircon grains disintegrated into small fragments during cataclasis (Figure 14). The weak erosion of their face surfaces suggests that the zircon was partly dissolved. The cathode luminescence examination of more than 30 individual zircon grains revealed their zoning, which is likely of magmatic genesis. We detected weak oscillatory and peculiar sectorial zoning (Figures 14a, 14b), as is typical of zircon from mafic crystalline rocks, particularly those in ophiolitic complexes [Rubatto and Gebaver, 2000]. The primary nature of the sectorial zoning is clearly seen in Figure 2c, in which a fragment of this zoning cuts across the morphological features of a small fragment crystal of a zircon crystal.

[35]  The smaller fraction includes euhedral zircon grains with clearly pronounced dipyramid faces (Figure 14c). The predominance of euhedral grains was caused both by the crystallization of this zircon from a melt and by the fragmentation of larger grains during cataclasis. This follows form the fact that the zoning of some grains is not conformable with their morphology (Figure 2c), whose euhedral faceting seems to be secondary and imitates the habit of the larger primary grain.

2005ES000181-fig15
Figure 15
[36]  Zircon dating was conducted on a Finnigan MAT-261 solid-source eight-collector mass spectrometer using samples of three size fractions of the most transparent and euhedral grains and grain fragments. Fraction I consisted of 18 grains, which ranged form 500  m m to 300  m m along the long axis and from 250  m m to 150  m m along the short one. Fraction III comprised 35 grains with average sizes of 250 times 175  m m, and fraction IV included 30 grains and grain fragments with average sizes of 250 times 140  m m. A subcordant age value was obtained for small and the most euhedral and transparent crystals (Figure 15, Table 9). The age of the rock is 97.42 pm 0.15 Ma.

[37]  Our experience in studying zircon from blastomylonitization zones in the Early Proterozoic Pezhostrov Island gabbro-anorthosite massif in the Belomorian Mobile Belt in the White Sea area [Sharkov et al., 1994] indicates that zircon recrystallization coupled with a diminish of its grain sizes does not affect the isotopic characteristics of this mineral, and its age evaluated using these grains is close to the age of zircon crystallization form the melt [Zinger et al., 2001]. This led us to believe that the Mesozoic age obtained for zircon from gabbronorite from the Markov depression corresponds to the crystallization age of this mineral.

[38]  Finds of ancient zircon in gabbroids form the axial MAR zone put forth the problem of interpreting the nature of this mineral itself, as to whether it crystallized from a magmatic melt and, hence, its age corresponds to the crystallization age of the rock, or the zircon is a relict and xenogenic mineral that was not decomposed during the melting of the rock. According to currently adopted concepts, the opening of the Central Atlantic started at approximately 170 Ma and continues until now [Pushcharovskii, 2001]. This led Pilot et al. [1998] to hypothesize that zircon could be preserved in the mantle at temperatures of about 1000o C for 150 m.y. without losses of its radiogenic lead. These researchers proposed two possible explanations for the nature of ancient zircon in the gabbroids:

[39]  (1) The rifting coupled with the opening of the Atlantic resulted in the tectonic disintegration and imbrication of the continental crust, which was broken into a series of tectonic slabs as a consequence of subhorizontal thrusting in the continental lithosphere. Involved in small-cell convection in the upper mantle, these continental lithospheric fragments were remelted. The convection cells that circulated on both sides of the ridge facilitated the displacement of these semi-molten zircon-bearing rocks toward the spreading axis, where they took part in the formation of the gabbroids.

[40]  (2) The continental crustal material has been preserved in the Kane zone since the opening of the Atlantic because of the migration of the transform fault and jumps of the ridge axis. Part of this material later subsided in the axial zone of the ridge and was involved in the origin of the gabbroids.

[41]  Thus, these researchers believe that the ancient zircon is of continental genesis, is not related to oceanic magma generation, and, hence, cannot provide information on the age of its host rocks. At the same time, geological, petrological, and isotopic-geochemical data provide no evidence that continental material can occur either in the Kane Fracture Zone or in the Markov depression. Moreover, the Fe-Ti-oxide silicic rocks found in the walls of the depression are typical of the third layer of slow-spreading ridges, such as MAR or the Southwest Indian Ridge, and of some ophiolitic associations, such as Monviso in the Western Alps [Lombardo et al., 2002; Rubatto and Gebaver, 2000]. These rocks typically bear zircon with analogous unusual sectorial zoning, which is uncommon in this mineral from continental rocks.

[42]  These facts demonstrate that there is no need to propose any specific mechanism for the genesis of the zircon. Furthermore, according to experimental data, zircon can be easily dissolved in basaltic magmas and can crystallize only from melts oversaturated with silica [Watson and Harrison, 1983], and, thus, the occurrence of this mineral in rocks of the silicic Fe-Ti-oxide association can hardly be accidental. This led us to believe that the zircon crystallized from melts that were emplaced into the oceanic lithosphere, and the zircon age corresponds to the age of its host rocks.


RJES

Citation: Sharkov, E. V., N. S. Bortnikov, T. F. Zinger, and A. V. Chistyakov (2005), Silicic Fe-Ti-oxide series of slow-spreading ridges: petrology, geochemistry, and genesis with reference to the Sierra Leone segment of the Mid-Atlantic Ridge axial zone at 6° N, Russ. J. Earth Sci., 7, ES4001, doi:10.2205/2005ES000181.

Copyright 2005 by the Russian Journal of Earth Sciences

Powered by TeXWeb (Win32, v.2.0).