CC BY
11
RUSSIAN JOURNAL OF EARTH SCIENCES, VOL. 15, ES1001, doi:10.2205/2015ES000546, 2015
Compositional heterogeneity of the continental mantle: Evidence from ultramafic xenoliths in cenozoic basalts from Northern Eurasia
A. F. Grachev1
Received 28 January 2015; accepted 30 January 2015; published 10 February 2015.
A representative collection of spinel lherzolite xenoliths in Cenozoic basalts from the Baikal Rift (Khamar-Daban Range, Vitim Plateau, and Udokan Plateau), Tien Shan, Bohemian Massif, Pannonian Basin, Catalonia, central and southeastern Spain, Canary Islands, and Lanzarote Island was studied to compare the chemical and isotopic-geochemical composition of the mantle beneath these territories. Comparison of the He signatures, considered together with data on the whole-rock compositions and REE patterns of the samples, reveals variable degrees of lherzolite mantle depletion due to both its original composition and variations in the degree of its partial melting. Another real process is mixing of two reservoirs caused by mantle convection, for example, mixing of a plume component plus subducted lithosphere material. Chemical geodynamics offers various variants of mixing based on studying the isotopic composition not only of the He-Ar and Nd-Sr systems but also of the Pb, Hf, and Os, although realistic implications for chemical geodynamics can be obtained only from the former two systems. KEYWORDS: ultramafic xenoliths; mantle; geochemistry; isotopic composition; Northern Eurasia.
Citation: Grachev, A. F. (2015), Compositional heterogeneity of the continental mantle: Evidence from ultramafic xenoliths in cenozoic basalts from Northern Eurasia, Russ. J. Earth. Sci., 15, ES1001, doi:10.2205/2015ES000546.
Introduction
Ultramafic xenoliths in alkaline basalts are the only source of information on the chemical composition and physical properties of the deep interiors of the Earth. Studies conducted over the past five decades have proved that spinel lherzolite strongly dominates among rocks found as mantle xenoliths in lavas. Based on various lines of petrological and geochemical evidence, most researchers are currently prone to regard spinel lherzolite xenoliths as fragments of mantle material entrained in magmas and brought to the surface during volcanic eruptions [Ringwood, 1975; and many others].
One of the principally important issues actively discussed in the literature is the compositional heterogeneity of the upper mantle, which is remarkably differently interpreted by various researchers, first, in view of the lack of criteria for identifying this heterogeneity and, second, because the
1 Institute of Physics of the Earth, Russian Academy of Sciences, Moscow, Russia
Copyright 2015 by the Geophysical Center RAS. http://elpub.wdcb.ru/journals/rjes/doi/2015ES000546.html
factual material is not always reliable enough. The possible heterogeneity of the upper mantle was first pointed out by Harris [Harris et al., 1967]. Later, based on more representative material, Hutchinson with colleagues [Hutchinson et al., 1975] arrived at the conclusion that the mantle underlying the Massif Central is laterally heterogeneous. However, according to an alternative viewpoint, which was expressed, for example, in [Basaltic..., 1981], study of spinel lherzolite nodules provides very little information testifying to any significant compositional transformations in the upper mantle. The differences in the interpretations likely stem from the methods applied in studying the compositional heterogeneity of the Earth's mantle. First, there are no representative data on variations in the composition of nodules within single volcanic centers, and second, it is so far uncertain as to how much the compositional parameters of a mantle rock can vary within a single xenolith, which can be as large as 80-100 cm along their long axes.
Our study was focused on studying variations in the chemical and isotopic-geochemical composition of mantle xeno-liths in Cenozoic alkaline basalts from various geodynamic settings in Northern Eurasia at various scale (the samples were collected at the Baikal Rift, northern Tien Shan, Pannonian Basin, Bohemian Massif, Catalonia, central and southeastern Spain, and Canary Islands.
ES1001
1 of 13
Table 1. Variations in the Concentrations (ppm) of Trace Elements Within Sample CH-1
Element 1 2 3 4 5 6 7
Li 3.1 2.8 3.4 3.1 2.9 1.9 2.5
Rb 5.0 8.4 7.8 10 10 12 17
Cs 1.6 0.88 2.9 2.8 2.1 3.2 5.1
Cr 1800 1500 1100 3000 1800 900 900
Mn 790 740 690 700 660 790 710
Co 78 30 35 25 115 40 35
Ni 2000 2000 2000 1800 1900 2400 2200
Cu 25 73 25 63 90 30 20
Zn 56 91 51 74 100 73 79
Figure 1. Layout of the studied fragments of sample CH-1: A and B are platelets (sections) CH-1/1-4 and CH-1/5-7, respectively; and C shows the general view of the sample. The vertical arrow coincides with marker arrows on the oriented thin sections.
Chemical and Isotopic Heterogeneity of the Mantle Within Single Xenolith Samples
Preparation Procedures of Samples for Their Analysis
Our study was conducted using a sample of spinel lherzo-lite (sample CH-1) from Neogene basalt from the Bohemian Massif (from the Smrci area). No a priori considerations were assumed to select a sample for this study, and the only requirement was the absence of any traces of atmosphere-induced alterations, such as oxidation, secondary minerals, extensive fracturing, etc.
Our spinel lherzolite sample was a three-axial ellipsoid in shape, 34 by 15 by 9 cm along its principal axes. The sample was saw-cut perpendicular to its long axis into three parallel platelets 2.5 cm thick. One of the platelets was crushed and pulverized to analyze for major and trace elements and conduct some other detailed geochemical studies. The other two platelets were further saw-cut into seven 2-cm cubes
Note: numerals 1 through 7 correspond to the numbers of cubes in Figure 1.
(Figure 1). Oriented polished thin sections parallel to each faces of the cubes were then manufactured to examine the petrography and petrotexture of the rocks and to analyze its phases on a microprobe.
Variations in the Concentrations of Major and Trace Elements in Sample CH-1
The compositions of all rock-forming minerals were proved to vary from the central portion of the sample to its peripheries. Away from sample center, the olivine becomes systematically depleted in MgO, the clinopyroxene simultaneously enriches in Na2O, SiO2, and Cr2O3, whereas the orthopyrox-ene shows practically no systematic variations in its composition. The character of these variations was elucidated using factor analysis. The olivine shows a negative correlation between the Mg, Si and Fe* concentrations from the central part of the sample to its peripheries; the clinopyroxene displays negative correlations of Si, Ca, and Mg concentrations with its Ti and Cr contents; and the orthopyroxene exhibits analogous correlations of Mg with Cr and of Ca with Al and Si [Grachev et al., 1987].
Similar tendencies are also seen in the behavior of trace elements: the Rb and Cs concentrations systematically vary from the central part of the sample to its peripheries. At the same time, the concentrations of Fe-group elements (Co and Cr) show significant variations not only from the central part of the sample to its peripheries but also within the sample. The maximum Cr and Co concentrations are almost twice as high as their minimum values (Table 1).
Comparative analysis of variations in the concentrations of major and trace elements within the sample with analogous variations in xenoliths from other areas located far away from one another display overlaps in the ranges of their concentrations (Table 2).
These data demonstrate that the conclusion about mantle heterogeneity based on data on single samples (their single thin sections or whole-rock analyses) can hardly be viewed as fully reliable. This puts forth the question about the limits of the ranges of concentrations of major and trace elements
Figure 2. (a) MgO-CaO and (b) MgO-AhOe diagrams for the composition of xenoliths from Tumbusun-Dulga volcano.
outside which it is reasonable to conclude that the mantle beneath certain Earth's territories is indeed heterogeneous.
Variations in the Chemical Composition of Spinel Lherzolite Xenoliths Within a Single Volcanic Center: An Example of Tumbusun-Dulga Volcano
To elucidate how much the composition of mantle xenoliths varies within a single sampling site, we have collected a representative number of xenolith samples from the basalt
neck of Tumbusun-Dulga volcano in the Khamar-Daban Range in the southeastern portion of the Baikal Rift. The 69 examined samples of spinel lherzolite xenoliths display significant variations in composition, from strongly depleted rock varieties with > 40 wt % MgO and low CaO and Л12Оз concentrations to rocks of composition close to that of the primitive mantle (Figure 2). Containing approximately 37 wt % MgO, the primitive mantle is also rich in CaO and Л12Оз [Grachev, 1989]. These data testify that the mantle underlying the Khamar-Daban Range is remarkably heterogeneous in composition.
Table 2. Average Concentrations (ppm) of Trace Elements and Their Limits in Sample CH-1 Compared to Those in Other Spinel Lherzolite Xenoliths From the Bohemian Massif and Pannonian Basin
Element 1A 1B 2 3
A max min A A max min A max min
Li 2.8 3.8 1.9 1.9 2.0 2.8 1.7 1.76 2.6 0.97
Rb 10.0 17 5.0 3.6 5.4 11.0 2.7 4.84 10.0 1.60
Cs 2.65 5.1 0.88 2.5 4.0 5.6 2.5 5.9 25.0 1.90
Cr 1571 3000 900 1900 1900 2800 900 2294 3200 1500
Mn 727 790 660 950 814 950 730 806 1000 710
Co 51.1 115 25 110 111 120 110 106 120 78
Ni 2042 2400 1800 2200 2337 2600 2200 2161 2500 1000
Cu 46.5 90 20 16 10.6 16 5 12.9 65 5
Zn 74.8 100 51 53 62 95 49 62.2 98 48
Note: 1A and 1B are sample CH-1; 1A is an average of the concentrations in seven cubes, 1B is the concentration in the platelet utilized for whole-rock analysis, 2 is the average concentrations (of eight xenoliths) for the Bohemian Massif, 3 is an average of analyses of eighteen xenoliths for the Pannonian Basin (author's data).
Figure 3. Al2O3-MgO-CaO (wt %) diagrams for ultramafic xenoliths of (a) the continental mantle as a whole and (b) from Tumbusun-Dulga volcano.
Moreover, the variations in the mantle composition be- ation of mantle material is clearly seen in the Mg/Si-Al/Si
neath this part of the Khamar-Daban Range are comparable and Ca/Si-Al/Si diagrams below (Figure 4a and Figure 4b),
with the compositional variability typical of the whole con- which also display for comparison the compositions of car-
tinental mantle shown in Figure 3. The degree of fraction- bonaceous chondrite, pyrolite, and primitive mantle.
Mg/Si
o
o
+ 1 + 2 o 3 O o 8 O o ti
04 ♦ i
• 6
* 7
* 8
♦ 9 •
0 10
□ 11
o o
o
+ +
• *
(a)
ft +
+ + + +■
Al/Si
0.16
0.12
0.08
0.04
0
Ca/Si
(b)
o
oo o
+ +
+-H- + + 4- +
H- o ++ $
ir
+Í+-H
+
++f+
o oo o
+
CEO
iL ^■lililí
I" ^^TTT^P
#- + +
+ 1
+ 2 o 3 O 4 it 5 • 6
* 7
* 8 4 9 s 10 □ 11
Al/Si
0.04
0.08
0.12
0.16
Figure 4. (a) Mg/Si-Al/Si and (b) Ca/Si-Al/Si diagrams for the compositions of ultramafic xenoliths from Tumbusun-Dulga volcano: 1-4 - mantle beneath Tumbusun-Dulga volcano: 1 - undepleted, 2 - weakly depleted, 3 - strongly depleted, 4 - average mantle composition beneath Tumbusun-Dulga volcano, 5 - pyrolite [Ringwood, 1975], 6 - C1 chondrite [Jagoutz et al., 1979], 7 - primitive mantle [Hart, Zindler, 1986], 8 - same [Anderson, 1983], 9 - same [Palme and Nickel, 1985], 10 - same [Grachev, 1989], 11 - same [Kogarko et al., 1987].
Figure 5. Chondrite-normalized REE patterns of xenoliths from Tumbusun-Dulga volcano (see text for explanations).
Most of the composition data points of the xenoliths cluster along the line of geochemical fractionation Jagoutz et al., 1979], thus reflecting variations in the depletion of the mantle beneath Tumbusun-Dulga volcano. At the same time,
approximately 10% of all of the studied xenolith compositions lies near or coincides with the compositions of pyro-lite and the primitive mantle (Figure 4a). Based on the Ca/Si ratio of mantle rocks, Morgan [1986] has classified them into three groups: weakly depleted (> 0.09), moderately depleted (0.09-0.06), and strongly depleted (< 0.06). All of these groups were detected among xenoliths from Tumbusun-Dulga volcano, as can be seen in Figure 4b, which shows the compositional discreteness of the xenoliths pronounced in their Al/Si ratio.
Variations in REE Concentrations in Mantle Xenoliths From Tumbusun-Dulga Volcano
According to their REE concentrations, the rocks composing the xenoliths can be subdivided into three groups (Figure 5a, Figure 5b, and Figure 5c).
The first group comprises samples whose REE patterns are close to chondritic ones (Figure 5a), the second group is made up of samples with negative Ce anomalies (Figure 5b), and the third group includes samples with pronounced enrichment in LREE (Figure 5c).
The first group of xenoliths (Figure 5a) corresponds to the least depleted source and has REE concentrations two to five times higher than the chondritic ones. The second group (Figure 5b) of rocks with negative Ce anomalies consists of xenoliths of mildly depleted mantle material. As follows from data on alnoite xenoliths from Malaita Island, Solomon Islands [Neal and Taylor, 1989], the reason for the Ce anomaly was remelting of the crustal material in subduction zones.
Finally, the third group of xenoliths (Figure 5c), which are obviously enriched in LREE, corresponds to mantle residues.
Comparison with known REE patterns of mantle rocks from elsewhere (Mongolia, southeastern Australia, San Carlos in the United States, and Dreiser Weiher in Germany), where single samples of ultramafic xenoliths in basalt were studied [Agafonov and Erkushev, 1984; Frey and Prinz, 1978; and several others] indicates that samples from a single locality may cover the whole range of variations in the REE concentrations relative to chondrite, except only for Alpinetype hyperbasites.
At the same time, data on REE distribution in mantle xenoliths from Tumbusun-Dulga volcano are in reasonably good agreement with data on the whole-rock composition of xenoliths from elsewhere and thus provides further support for the hypothesis that the regional mantle is indeed strongly chemically heterogeneous. One of the reasons for this heterogeneity is variations in the degree of melting, as follows from the evident correlations between the MgO and CaO concentrations with Al2O3 (Figure 3).
Sm—Nd and He—Ar Isotopic Systematics of Mantle Xenoliths from Tumbusun-Dulga Volcano
This section presents the results of studying the Nd, Sr, He, and Ar isotopic ratios of mantle rocks from beneath Tumbusun-Dulga volcano.
Figure 6. 143Nd/144Nd-87Sr/86Sr diagram for xenoliths from Tumbusun-Dulga volcano in comparison with analogous data on xenoliths from elsewhere (our data). The black dashed line contours the compositional field of xenoliths from Western and Central Europe [Downes, 2001], and the green dashed line outlines the field of xenoliths from East Asia [Xu et al., 2004]. BSE is the uniform chondritic reservoir.
In modern chemical geodynamics, the following three major types of mantle material are distinguished: primitive mantle (PM), depleted mantle (DM or MORB), and enriched mantle (EM) [Zindler and Hart, 1986], with differences between them seen in the concentrations of major elements (Mg, Al, and Ca), REE, and Sm-Nd and HeAr isotopic systems. The depletion of the mantle was related to its partial melting or metasomatism, and its enrichment was caused by subducted oceanic lithosphere (high 87Sr/86Sr> 0.7050). As follows from Figure 6, xenoliths from Tumbusun-Dulga volcano occur as a swarm of data points (as these data points also do in the diagrams in Figures 3, 4, 5) extending from the primitive to depleted mantle fields. Note that this swarm of data points overlaps the field of the composition of the mantle beneath Europe. Data obtained on the Nd-Sr isotopic system of glasses from zones of partial melting in sample 979/152 from basalt of Tumbusun-Dulga volcano shed light on the reasons for this chemical and iso-topic heterogeneity. The xenolith represented by this sample, glasses in it, and its host basalt were proved to have closely similar isotopic signatures (Figure 7).
The Sr and Nd isotopic ratios of glass in xenoliths in basalts from the Khamar-Daban Range and Vitim Plateau are closely similar and plot in the vicinity of BSE, whereas the data points of the xenoliths themselves lie on the mixing line of the BSE and MORB mantle sources. The isotopic composition of glass from the partial melting zone for the Tok Stanovoi Range lies within the field of xenoliths from the Khamar-Daban and Vitim, and the xenolith itself has isotopic ratios typical of MORB. The composition of glass in the xenolith from the Khamar-Daban Range corresponds to picrite basalt, and the U-Pb isochron for the glass yields an age of 22.7±4.2 Ma, which is much older than the age of the host basalt, which was dated at 17-18 Ma. These data show that the mantle that is heterogeneous in chemical and iso-topic composition beneath the Baikal Rift Zone (as follows from data on spinel lherzolite xenoliths in the basalts) includes material whose isotopic-geochemical parameters correspond to the primitive mantle. Studies of partial melting zones in mantle xenoliths from Tumbusun-Dulga volcano show that the partial melting of primitive mantle material may yield parental melts containing > 14 wt % MgO.
Figure 7. 143Nd/144Nd-87Sr/86Sr diagram for the composition of xenoliths, partial melting zones in them, and their host basalts.
Variations in the He Isotopic Composition of Xenoliths From the Khamar-Daban Range
Figure 8 shows our data on the He isotopic system and the broad variations in the 3He/4He ratio from values typical of the atmosphere to those of hotspots and/or primitive mantle.
Values typical of the atmosphere were determined in three of the 22 examined samples. The highest values were detected in samples 826 and 978, and the 3He/4He ratios of the olivine are two to three times higher than those of the whole-rock samples. Analogous relations were also found in sample 979/152 (Figure 8).
It is interesting to analyze the variations within sample 979/152 of spinel lherzolite, whose He isotopic composition was determined (in the whole-rock sample and in olivine of various populations, including experiment on its leaching) [Grachev, 1998]. The 3He/4He ratio varies from 2.3 to 18 x 106. The lowest value was yielded by the whole-rock sample, and the highest one by olivine after its leaching with HCl. What is important, olivine porphyroclasts and neoblasts differ in He isotopic composition, which is more
that twice as low in the porphyroclasts as in the neoblasts.
The occurrence of a number of olivine populations in mantle xenoliths is explained by the long-lasting deformation history of the latter as a result of solid-plastic flows in the mantle before the xenoliths were brought to the surface [Mercier and Nicolas, 1975]. Extensive studies indicate that these deformations occur in the presence of fluid and are associated with metasomatism and partial melting. These processes enriched the partial melting zones in U and modified their He isotopic ratios. Such changes were detected in sample 979/152, whose partial melting zones contain up to 1500 x 10-9 g/g U, i.e., roughly 1000 times more than this concentration in the olivine and 100 times more than its concentration in the xenolith [Grachev and Komarov, 1994]. In this process, the 3He concentrations remained unmodified, and the 4He concentrations increased and thus decreased the 3He/4He ratio. Since practically all mantle xenoliths show traces of mantle metasomatism, the He isotopic ratio of the whole-rock samples are always lower than those of their olivine.
The latest data on the He isotopic composition of mantle xenoliths in basalts from the Pannonian Basin, Bohemian Massif, Vitim Plateau, northern Tien Shan, and Lanzarote
100.00 —I
Figure 8. 3He/4He-40Ar/36Ar diagram for mantle xenoliths from Tumbusun-Dulga volcano: A - atmosphere, LM - lower mantle, and UM - upper mantle.
Island [Grachev, 2012a, 2012b, 2013; Grachev et al., 2014] demonstrate that the variations in the 3He/4He ratio can be controlled by several factors, first of all, by the U concentration, the effect of cosmic irradiation, and the occurrence of partial melting zones. To accurately compare data on He isotopic composition, these data must be obtained by the crushing techniques, which makes it possible to extract 3He from gas-liquid inclusions in olivine. The variations in the 3He/4He ratio in the olivine porphyroclasts and neoblasts may reach the same orders of magnitude.
Figure 9. CaO-Al2O3 diagram for the composition of mantle xenoliths in Cenozoic basalts in Europe and Asia in comparison with data on the whole continental mantle. A - all continental mantle, b - Asia, c - Europe.
Heterogeneity of the Chemical and Isotopic Composition of the Mantle Beneath North Eurasia
This study was conducted based on our database (more than 300 analyses) on the composition of xenoliths. For comparison, we also used data from [Maaloe and Aoki, 1977] on xenoliths in basalts from Europe.
Our earlier factor analysis of the chemical composition of mantle xenoliths [Grachev, 1989] has revealed that the type of the mantle material (degree of its depletion relative to the chondrite reservoir) is determined by the CaO, Al2O3, and MgO concentrations. Undepleted mantle is characterized by high CaO, Al2O3 (> 3 wt %), and MgO (35-37 wt %).
Figure 9 shows variations in the mantle composition beneath Europe and Asia in comparison with the composition of the whole continental mantle. As can be readily seen from the comparison of the variations in the composition of xenoliths in Figure 9 with data on xenoliths from Tumbusun-Dulga volcano in Figure 3, all of the xenoliths display analogous compositional trends and relations. This led us to conclude that the continental mantle at its spinel depth facies is well enough homogenized. The comparison of the He signatures, considered together with data on the whole-rock composition and REE patterns, of mantle xeno-liths at various scales highlights similar degrees of lherzolite mantle depletion due to various degrees of partial melting of the primitive original composition.
There is another real process related to mantle convection:
the mixing of various reservoirs, for example, a plume component plus subducted lithosphere. Chemical geodynamics offers a diversity of mixing scenarios based on data on the isotopic composition not only in the He-Ar and Nd-Sr systems but also of Pb, Hf, and Os, but only the former two systems have implications for mantle geodynamics.
Table 3 presents our data on the Sm-Nd isotopic ratios of mantle xenoliths in basalts in various areas in northern Eurasia. These data were plotted in a diagram of these ratios (Figure 10), which obviously shows that all of the samples from various areas in Northern Eurasia plot within the region typical of the lherzolite mantle beneath Europe.
It follows that, similar to the variations in the chemical composition, the heterogeneity of Sm-Nd isotopic composition of the upper mantle beneath northern Eurasia is constrained within fairly narrow limits.
Currently available estimates for the Sm-Nd and Os dates for the continental mantle beneath Asia correspond to approximately 2 Ga. The oldest age was determined for the mantle beneath the Tien Shan: the Sm-Nd isochron of three minerals (olivine, diopside, and enstatite) corresponded to an age of 2894 ± 180 Ma [Grachev et al., 2014].
Helium in Mantle Xenoliths From Northern Eurasia
0.5140 ■
0.5136 ■
0 5132 ■
•u
z
?
0.5124 —
0 5120 •
/ ° JL o| —s. ° .c5cP / o7o
o o o
000
0.10
0.20
l7Sm/U4Nd
0.40
0 50
0.60
Figure 10. Sm-Nd diagram for the composition of mantle xenoliths in northern Eurasia. The heavy line contours the composition of the continental mantle beneath Europe (according to [Downes, 2001]).
The discovery of primary (planetary) 3He on Earth in 1969 [Mamyrin et al., 1969] has stimulated studies of the He isotopic composition of, first and foremost, mantle rocks. The pioneering publication by I. N. Tolstikhin et al., which presented data on the He isotopic composition of samples from Kamchatka, Stanovoi Highland, Spitsbergen, and Antarctica, was followed by papers presenting data mostly on ultramafic xenoliths and basalts from oceanic islands. A few studies were centered on the He isotopic composition of ul-tramafic xenoliths and their host basalts within the continental lithosphere, but before our research, no such data have been obtained at all in Asia. The very first data on the He isotopic composition of mantle xenoliths in northern Eurasia were published in [Drubetskoi and Grachev, 1987]. It has been proved that mantle xenoliths (spinel lherzolites) in the Baikal Rift Zone (BRZ) usually hosted in the neck facies of basalts have various 3He/4He ratios. Two mantle reservoirs were then distinguished: one of them has a 3He/4He ratio much higher than the atmospheric one R0 = 1.39 x 10-6 and higher than the ratios typical of the MORB reservoir (R/R0 = 8 x 10-6) and pertains to the southwestern flank of BRZ (Khamar-Daban Range), and the other one has a 3He/4He ratio close to the atmospheric one and is spatially restricted to the northeastern flank, the Tchara Rift (Udokan Range). These data and those on the chemical composition and He isotopic composition of basalts hosting the mantle xenoliths lead to the conclusion that the mantle underlying BRZ is characterized by at least two mantle reservoirs. It is important that all of the experiments were carried out by the melting technique. Later [Staudacher and Allegre, 1991] also applied this technique to study the He isotopic composition of xenoliths from the Vitim Plateau and Taryat Depression in Mongolia and demonstrated that the R/R0 ratio exceeds
that typical of MORB (R/Ro = 8 x 10-6). The effect of cosmogenic He is so far uncertain.
The disadvantage shared by all papers published before ours was that, first, the studies of the He isotopic system were not coupled with detailed petrological and geochemical studies of the xenoliths themselves, and second, no variations in the He isotopic composition were simultaneously studied in the xenolith samples and within the sampling sites. Our data demonstrate that even within a single xenolith sample, olivine porphyroclasts and neoblasts have different He isotopic ratios: the R/Ro ratios of the porphyroclasts are two to three times higher than in the neoblasts because of magma hydraulic fracturing induced by in-situ melting of xenoliths before their entrainment by melt. He diffusion within individual mineral grains in xenoliths is still inadequately poorly understood with regard for the usually old model ages (> 2 Ga) of the xenoliths.
Another important aspect of our studies is possible correlations between He, Sr, and Nd isotopic ratios. It was proved that the He isotopic ratios are not correlated with those of Sr and Nd and that these systems evolve independently.
While the technique most commonly applied in studying He isotopic composition involves crushing grains of olivine and other minerals to extract He from gas-liquid inclusions hosted in these minerals, which makes the values thus obtained independent of any metasomatic alterations that always variably affect all xenoliths, the Sr and Nd isotopic systems can, conversely, be significantly modified by these alterations. This has been demonstrated by studies of the so-called red mantle xenolith suite, in which metasomatism is pronounced most obviously (at the Udokan Range, Lanzerote Island, and Syria).
This publication summarizes the results of our long-term
Table 3. Sm and Nd Isotopic Ratios of Mantle Xenoliths From Various Areas in Northern Eurasia
no. Sample Area Concentration, ppm Isotopic ratios
Sm Nd 147Sm/144Nd 143Nd/144Nd
1 RC-88-1X cpx Catalonia 4.749 26.246 0.109374 0.512707 ± 23
2 RC-88-1/11a cpx 4.640 16.691 0.168044 0.512749 ± 12
3 RC-88-1X bt 0.352 2.078 0.102443 0.512823 ± 21
4 RC-88-1X bt 0.046 0.351 0.08450 0.512762 ± 42
5 RC-88-1X cpx 5.58 30.52 0.11058 0.512722 ± 5
6 RC-88-10 cpx 9.851 41.45 0.14366 0.512751 ± 3
7 RC-88-11 hrnb 6.93 27.25 0.15371 0.512811 ± 6
8 RC-88-11 cpx 5.72 20.19 0.17129 0.512798 ± 6
9 RC-88-3x wr 5.69 22.84 0.15099 0.512847 ± 5
10 RC-88-4x wr 0.615 3.147 0.14366 0.512751 ± 3
11 RC-88-7x wr 7.744 36.24 0.12915 0.512752 ± 5
12 RC-88-x wr 6.03 23.45 0.15210 0.512824 ± 6
13 S-1 cpx N. China 0.447 1.451 0.18627 0.513278 ± 208
14 S-1/1 ol 0.011 0.050 0.130797 0.512656 ± 42
15 S-1/1 cpx 1.825 5.316 0.207576 0.513080 ± 15
16 S-1/1 opx 0.033 0.097 0.208094 0.513067 ± 20
17 S-3/2 cpx 3.188 17.486 0.110215 0.512842 ± 19
18 S-3/2 cpx 3.29 18.14 0.10963 0.512915 ± 13
19 S-3/2 opx 0.071 0.367 0.117275 0.512788 ± 41
20 S-3-1a opx 0.019 0.036 0.323477 0.513629 ± 47
21 S-3-1a cpx 1.723 4.236 0.245882 0.513083 ± 21
22 780/1 opx Baikal Rift 0.025 0.054 0.28605 0.512880 ± 5
23 780/1 di 1.432 3.587 0.24121 0.513323 ± 11
24 826 wr 2.454 14.98 0.09900 0.512775 ± 21
25 978 wr 1.544 4.006 0.23296 0.513219 ± 7
26 979/104 cpx 2.137 4.651 0.27777 0.513241 ± 10
27 979/106 cpx 1.445 3.483 0.250080 0.513028 ± 51
28 979/106 wr 0.500 1.490 0.20140 0.513083 ± 21
29 979/152 cpx 0.477 1.528 0.18856 0.512939 ± 10
30 979/163 opx 0.026 0.050 0.312538 0.513360 ± 28
31 979/163 cpx 1.514 3.648 0.250955 0.513295 ± 17
32 979/167 opx 0.035 0.103 0.203582 0.512632 ± 62
33 979/167 cpx 1.803 4.658 0.234039 0.513080 ± 19
34 1065 cpx 0.927 2.597 0.21579 0.512887 ± 10
35 1069 cpx 7.11 22.42 0.19156 0.512857 ± 6
36 TS-90-1 opx N. Tien Shan 0.964 1.113 0.52345 0.513660 ± 53
37 TS-90-5 cpx 1.140 2.316 0.29762 0.5133625 ± 41
38 TS-90-5 opx 0.033 0.027 0.14442 0.511865 ± 115
39 CNT89-1 ol C. Spain 0.013 0.065 0.11970 0.512705 ± 120
40 CNT89-1 opx 0.013 0.040 0.19411 0.512962 ± 120
41 CNT89-1 cpx 1.407 4.162 0.20427 0.513155 ± 14
42 CNT89-2 ol 0.016 0.072 0.13414 0.512731 ± 73
43 CNT89-2 cpx 2.798 6.548 0.25829 0.513199 ± 8
44 CNT89-2 opx 0.041 0.084 0.29388 0.512682 ± 38
45 CNT89-3 ol 0.073 0.092 0.47933 0.512567 ± 51
46 CNT89-3 cpx 11.28 14.67 0.46458 0.512530 ± 9
47 CNT89-3 opx 0.298 0.208 0.88311 0.512416 ± 91
48 CNT-89/3 PL 1.047 2.266 0.27931 0.512227 ± 28
49 LP89-2/3 cpx 1.915 6.322 0.18309 0.513061 ± 13
50 LP89-2/3 opx SE Spain 0.045 0.095 0.28459 0.512888 ± 164
51 LP89-2/3 ol 0.008 0.044 0.11388 0.512787 ± 88
52 32883 cpx 2.140 4.867 0.26546 0.512954 ± 11
53 32887 cpx 2.131 7.130 0.18060 0.513022 ± 55
54 32888 cpx 2.129 7.362 0.17475 0.512790 ± 10
55 L88-1 wr Lanzarote Island 0.060 0.190 0.19022 0.512587 ± 16
56 L88-3 wr 0.720 1.491 0.29177 0.513095 ± 21
57 Tam-1 di 2.161 16.37 0.08007 0.512928 ± 27
58 Tam-2 di 3.747 22.66 0.10028 0.512968 ± 16
59 Tam-3 di 0.814 5.662 0.08717 0.512897 ± 21
60 Tim-1 di 2.024 13.54 0.09065 0.512969 ± 12
61 Tim-3 opx-I 0.014 0.058 0.143672 0.512100 ± 35
62 Tim-3 opx-II 0.028 0.129 0.129453 0.512858 ± 29
63 Tim-3 cpx 0.579 2.293 0.119727 0.512716 ± 17
64 GC-1 wr 0.008 0.054 0.09420 0.512820 ± 120
65 V-107 opx Pannonian Basin 0.014 0.064 0.134099 0.512716 ± 35
66 V-107 cpx1 0.881 5.511 0.096613 0.512795 ± 9
67 V-107 cpx2 1.810 6.956 0.15727 0.513011 ± 11
(for more than two decades) studies of the He isotopic composition of mantle xenoliths and their host Cenozoic basalts in an extensive area from the Iberian Peninsula in the west to China in the east. Over this time, data on the He and Ar isotopic compositions of 127 mantle xenoliths and 190 alkaline basalts were obtained. The He and Ar isotopic composition of basalts has never before been analyzed in samples from either Europe or continental Asia.
The isotopic studies were initially conducted at the Institute of Precambrian Geology and Geochronology, Russian Academy of Sciences, and the Geological Institute of the Kola Research Center, Russian Academy of Sciences, using techniques of melting, crushing and stepwise annealing at temperatures up to 1700° C of powdered rock samples of and monomineralic fractions. Crushing makes it possible to extract He from gas - liquid inclusions in minerals, melting extracts He from the crystal structures of minerals, and step-wise annealing at temperature higher than approximately 600° C makes it possible to get rid of radiogenic He.
The great majority of the examined mantle xenolith samples are spinel lherzolites. Their He isotopic composition was studied simultaneously with the detailed isotopic-geochemical study of the rocks and their monomineralic separates. The results obtained on the He isotopic composition of mantle xenoliths are presented in Figure 11, which clearly demonstrates broad variations in the R/Ro ratios from values lower than atmospheric to those typical of mantle plumes. The R/R0 ratio of xenoliths from the sublithospheric mantle of Europe never exceeds values typical of MORB (except only for xenoliths in basalts from Ciudad Real in central Spain), whereas the values for the continental mantle beneath Asia define a swarm of data points (Figures 2-11) that is orthogonal to the field of data point for the mantle beneath Europe, and some analyses yield R/Ro ratios higher than in MORB. This pertains to such areas as the northern Tien Shan, Khamar-Daban, and Vitim Plateau in the Baikal Rift and to Shavaryn-Tsaram volcano in Mongolia, whose basalts show a complex of isotopic-geochemical features relating them to mantle plumes. However, high R/R0 ratios can also result from the effect of cosmogenic 3He. It is known that the productivity of spallogenic He is merely insignificantly dependent on the composition of the host minerals. In this context, it is uncertain (and calls for further studying) as to why various or the same minerals contain different concentrations of the cosmogenic component.
Our data on the variations in the 3He/4He ratio from the core to marginal portions of three xenolith samples indicate that this ratios increases by factors of two to three, i.e., the samples obviously contain cosmogenic 3He. The effect of cosmogenic He can be eliminated in the following two manners: (i) by obtaining a sample from an unexposed portion of the vertical section of rocks, from a depth of 34 m (this was done in the Tien Shan, where overlying rocks were removed by blasting operations in 1991) or (ii) by introducing a correction for 3Hecos, which can be determined if the 3He/4He ratios are known for two minerals [Grachev et al., 2014]. To introduce such a correction, we have determine the 3He/4He ratios of three minerals from single samples from all of the aforementioned areas ; the correction for cosmogenic He did not exceed 20%. Nevertheless, a fi-
Figure 11. He-Ar isotopic ratios in mantle xenoliths in Cenozoic basalts in northern Eurasia.
nal conclusion concerning the effect of cosmogenic He can be obtained after studying the Ne isotopic composition: the very first data on the 20Ne/22Ne ratios of xenoliths in basalts from the Vitim Plateau had ratios (> 11) higher than the atmospheric ones [Staudacher and Allegre, 1991].
In any event, differences in variations in the He (and Ar) isotopic composition in the sublithospheric mantle beneath Asia and Europe are quite obvious. This conclusion follows from data on the whole-rock composition, geochemistry of trace elements and REE, and the Sr and Nd isotopic composition of xenoliths.
Partial Melting of the Lherzolite Mantle
Volcanic eruptions bring spinel lherzolites the surface from depths of 35-70 km, which correspond to the anomalous mantle (a zone of low velocities in it) [Grachev, 1987]. Because of this, mantle xenoliths usually show traces of partial melting, which has been evaluated in most publications on theoretical and experimental petrology at 1-3% [Dobretsov et al., 1980; and others].
However, data on xenoliths from the Rein Graben has demonstrated that the degree of partial melting may reach 5-10% [Maaloe and Printzlau, 1979]. Examination of a reference collection of spinel lherzolite xenoliths in basalts from the Baikal Rift (> 70 samples) has also shown that partial melting zones make up 10-12% of several samples [Grachev and Malevskii, 1988]. In this situation, the amount of melt can be evaluated with the naked eye without studying thin sections. The glass broadly varies in composition from pi-crite to nepheline-normative basalt. This fact is of principal importance, because the host basalt of the xenoliths has
an almost unvarying composition [Grachev, 1987]. Also, it is important to mention that not all of the xenoliths show traces of partial melting, and this testifies that the xenoliths did not partially melt as a consequence of their decompression, when entrained to the surface.
The U-Pb isotopic dating of glass in a sample of spinel lherzolite xenolith from basalt from Tumbusun-Dulga volcano in the Khamar-Daban Range, Baikal Rift, yielded an age of 22.6±4.2 Ma, whereas the host basalt is no older than 17-18 Ma. The former value corresponds to the age of the anomalous mantle beneath the Khamar-Daban Range.
Study of the tracks of U distribution in olivine from xeno-liths in melting zones shows that the U concentration in these domains is higher than 2000 x 10-9 g/g [Grachev and Komarov, 1994]. Closely similar values were also obtained on xenoliths from elsewhere, e.g. Pannonian Basin and others [Grachev, 2013].
Discussion and Conclusions
The continental mantle at a depth corresponding to the spinel facies has a long-lasting evolutionary history and is fairly good homogenized. The comparison of the He signatures, considered together with data on the whole-rock composition and the REE patterns of mantle xenoliths, provides grounds to conclude that the lherzolite mantle is variably depleted, with these variations predetermined by both the original heterogeneity of the mantle and variations in the degree of its partial melting. Note that our earlier factor analysis of the chemical composition of xenoliths in large continental lithospheric blocks [Grachev, 1989] led us to identify two differentiation trends of the mantle. One of these trends corresponds to the undepleted mantle that is rich in CaO and Al2O3 (> 3 wt %) and contains 35-37 wt % MgO (as is typical of Eastern Asia, for example, Tumbusun-Dulga volcano). The other trend corresponds to the differentiation of the strongly depleted mantle, which is poor in CaO and Al2O3 (< 2 wt %), and contains 42-45 wt % MgO, as is typical of the portion of the continental mantle that has been produced at the sacrifice of the subducted lithosphere (for example, the Pannonian Basin and southwestern United States).
Another process able to mix materials from various reservoirs is mantle convention, which can mix, for example, a plume component and subducted lithosphere.
Studying the isotopic composition not only of the HeAr and Nd-Sr systems but also Pb, Hf, and Os makes it possible to suggest various mixing scenarios within the scope of chemical geodynamics, but only the former two systems have realistic implications for mantle geodynamics.
References
Agafonov, L. V., Yu. A. Erkushev (1984), Distribution of rare-earth elements in rocks of Shavaryn-Tsaram volcano, Mongolia, Geol. Geofiz., No. 6, 65.
Anderson, D. L. (1983), Chemical composition of the mantle,
J. Geophys. Res., 88, B41. doi:10.1029/JB088iS01p00B41 Basaltic Volcanism Study Project (1981), Basaltic Volcan-ism on the Terrestrial Planets, 1286 pp., Pergamon Press, Inc., New York.
Dobretsov, N. L. (1980), Introduction Into Global Petrology,
200 pp., Nauka, Novosibirsk. Downes, H. (2001), Formation and modification of the shallow sub-continental lithospheric mantle: a review of geochem-ical evidence from ultramafic xenoliths suites and tectonically emplaced ultramafic massifs of Western and Central Europe, J. Petrology, 42, 233. doi:10.1093/petrology/42.1.233 Drubetskoi, E. R., A. F. Grachev (1987), Basalts and
ultramafic xenoliths in the Baikal Rift Zone: He and Ar isotopic composition, Deep-Seated Xenoliths and the Structure of the Lithosphere. Project: XENOLITH p. 54, Nauka, Moscow. Frey, F. A., M. Prinz (1978), Ultramafic inclusions from San Carlos, Arizona: petrologic and geochemical data bearing on their petrogenesis, Earth Planet. Sci. Lett., 38, 129. doi:10.1016/0012-821X(78)90130-9 Grachev, A. F. (1989), Compositional heterogeneity of the continental lherzolite mantle: evidence from ultramafic xeno-liths in basalts, Physics and Inner Structure of the Earth p. 22, Nauka, Moscow. Grachev, A. F. (1998), The Khamar-Daban Ridge as a hotspot of the Baikal Rift from data of chemical geodynamics, Physics of the Solid Earth, 34, No. 3, 175. Grachev, A. F. (2013), Depleted sublithospheric mantle beneath the Bohemian Massif and Pannonian Basin: new data on composition and He isotopic composition of mantle xenoliths and their host basalts in the context of the problem of mantle plumes in Central Europe, Fizika Zemli, No. 5, 89. Grachev, A. F. (1987), Earth's Rift Zones, 2nd edition, 287
pp., Nedra, Moscow. Grachev, A. F., N. I. Gul'ko, L. F. Dobrzhinetskaya, et al. (1987), On the compositional, structural, and physical
heterogeneity of a single plagioclase lherzolite sample, Deep-Seated Xenoliths and the Structure of the Lithosphere. Project: XENOLITH p. 8, Nauka, Moscow. Grachev, A. F., A. N. Komarov (1994), Newly obtained data on U concentrations in the continental and oceanic mantle, Fizika Zemli, No. 1, 3. Grachev, A. F., A. V. Malevskii (1988), On the anisotropy of the viscosity in the anomalous mantle, Dokl. AN SSSR, 303, No. 4, 830.
Grachev, A. F. (2012), Newly obtained data on He iso-
topes in mantle xenoliths in basalts from the Vitim Plateau, Eastern Siberia, Russia, Russ. J. Earth. Sci., 12, ES3003, doi:10.2205/2012ES000513 Grachev, A. F. (2012), MORB-like mantle beneath Lanzerote Island, Canary Islands, Russ. J. Earth. Sci., 12, ES3004, doi:10.2205/2012ES000515 Grachev, A. F., P. V. Ermolov, Yifen Wang (2014), Modification of the lithospheric mantle during the early activity of a Cenozoic plume in the North Tien Shan: Evidence from mantle xenoliths in basalts, Russ. J. Earth. Sci., 14, ES2002. doi:10.2205/2014ES000540 Harris, P. G., A. Reay, I. G. White (1967), Chemical
composition of the upper mantle, J. Geophys. Res., 49, No. 24, 6359. doi:10.1029/JZ072i024p06359 Hart, S. R., G. A. Zindler (1986), In search of a bulk-
earth composition, Chem. Geol., 57, 247. doi:10.1016/0009-2541(86)90053-7 Hutchinson, R., A. L. Chambers, D. K. Paul, P. G. Harris (1975), Chemical variation among French ultramafic xeno-liths — evidence for a heterogeneous upper mantle, Min. Mag., 40, No. 310, 113. Jagoutz, E., H. Palme, H. Baddenhausen, K. Blum, M. Cendales, G. Drebius, B. Spettel, V. Lorenz, H. Wanke (1979), The abundances of major, minor and trace elements in the Earth's mantle as derived from primitive ultramafic nodules, Proc. Lunar Planet. Sci. Conf. 10th. p. 2031, Nauka, Moscow. Kogarko, L. N., V. A. Turkov, I. D. Ryabchikov, et al. (1987),
Composition of the Earth's primary mantle: evidence from nodules, Dokl. AN SSSR, 290, No. 1, 199.
Maaloe, S., K. Aoki (1977), The major element composition of the upper mantle estimated from the composition of lherzolites, Contrib. Mineral. Petrol.., 63, 161. doi:10.1007/BF00398777
Maaloe, S., I. Printzlau (1979), Natural partial melting of spinel lherzolites, J. Petrol., 20, 727. doi:10.1093/petrology/ 20.4.727
Mamyrin, B.A., I.N. Tolstikhin, G.S. Anufriev, I.L. Kamenskii (1969), Anomalous He isotopic composition of volcanic gases, Dokl. AN SSSR, 184, No. 5, 1197.
Mercier, I. G. G., A. Nicolas (1975), Textures and fab-
rics of upper mantle peridotites as illustrated by xenoliths from basalts, J. Petrol., 16, No. 2, 454. doi:10.1093/petrology/16. 2.454
Morgan, J. W. (1986), Clues to Earth's late accretionary history, J. Geophys. Res., 91, 12375. doi:10.1029/JB091iB12p 12375
Neal, C. R., L. A. Taylor (1989), Negative Ce anomaly in a peridotite xenolith from Malaita (Solomon Islands): evidence for crustal recycling into the mantle or mantle metasomatism, Geochim. Cosmochim. Acta, 53, 1035. doi:10.1016/0016-7037(89)90208-1
Palme, H., K. G. Nickel (1985), Ca/Al ratio and composition
of the Earth's upper mantle, Geochim. Cosmochim. Acta, 49, 2123. doi:10.1016/0016-7037(85)90070-5 Ringwood, A. E. (1975), Composition and Petrology of the
Earth's Mantle, 618 pp., McGraw-Hill, New York. Staudacher, Th., C. J. Allegre (1991), Cosmogenic neon
in ultramafic nodules from Asia and in quartzite from Antarctica, Earth Planet. Sci. Lett., 106, 87. doi:10.1016/0012-821X(91)90065-P Xu, Y. G., S. L. Chung, J. Ma, L. Shi (2004), Contrasting Cenozoic lithospheric evolution and architecture in the Western and Eastern Sino-Korean Craton: constraints from geochemistry of basalts and mantle xenoliths, J. Geol, 112, 593. doi:10.1086/422668 Zharikov, V. A., A. F. Grachev (1987), Deep-Seated Xenoliths and the Structure of the Lithosphere. Project: XENO-LITH, 239 pp., Nauka, Moscow. Zindler, A., S. Hart (1986), Chemical geodynamics, Ann. Rev. Earth Planet. Sci., 14, 493. doi:10.1146/annurev.ea.14. 050186.002425
A. F. Grachev, Institute of Physics of the Earth, Russian Academy of Sciences, 10 Bolshaya Gruzinskaya, Moscow, Russia. (afgrachev@gmail.com)
