Fig 4 - uploaded by Daniela Rubatto
Content may be subject to copyright.
Composition of melt and garnet. a) Garnet and melt FeO and MgO composition in wt.% at different temperatures. Filled symbols indicate garnet and open symbols indicate melt. b) Variation of trace element composition of the melt with temperature. Filled symbols are for Th and U contents in experiments containing significant amounts of allanite.
Source publication
Garnet is the most commonly used mineral in thermobarometry, whereas zircon is the most robust chronometer to date high-grade metamorphic rocks. To provide a basis for correlation of zircon and garnet growth, we determined experimentally the trace element partitioning between zircon, a hydrous granitic melt and garnet at 20 kbar and 800–1000 °C for...
Contexts in source publication
Context 1
... experiments produced peraluminous granitic melts with an alumina saturation index [ASI = Al/(K + Na + 2Ca) expressed in molar fraction] between 1.15 and 1.49 (Table 3). The melt shows a moderate trend to more mafic compositions at higher temperature with increasing in MgO and FeO contents (Fig. 4a). The measured totals of 84-89 wt.% oxides indicate that the melts contain additional about 11-16 wt.% of H 2 O, in agreement with mass balance having 11 wt.% of H 2 O in the starting material and 70-90% melt in the experiments. Compositional variations in experiments run at identical conditions are due to some difference in the ...
Context 2
... (Fig. 5). EDS and WDS analyses were performed along one or more profiles across the garnet crystals and the small standard deviation on multiple analyses reported in Table 4 testifies to the homogeneity of garnet. With increasing temperature, there is a systematic increase in the pyrope component of garnet (Py 25 @800 °C-Py 44 @1000 °C, Table 4, Fig. 4a), whereas the grossular component decreases (Gr 22 @800 °C-Gr 8 @1000 °C). Normalisation of the garnets including also the measured trace elements indicates that only a small amount of Fe 3+ of less then 5% of the total iron is required to obtain charge balance. An interesting feature of the experiments is the observed change of Mg# ...
Context 3
... decreases (Gr 22 @800 °C-Gr 8 @1000 °C). Normalisation of the garnets including also the measured trace elements indicates that only a small amount of Fe 3+ of less then 5% of the total iron is required to obtain charge balance. An interesting feature of the experiments is the observed change of Mg# (Mg/(Mg + Fe 2+ )) of garnet relative to melt ( Fig. 4a). At 800 to 900 °C garnet has a lower Mg# than the melt, at 950 °C it is identical and at 1000 °C garnet has a higher Mg# than melt. Therefore, garnetproducing partial melting of felsic rocks at low temperature will generate a restite that has a significantly lower Mg# and a melt that has a higher Mg# than the ...
Context 4
... trace element composition of the melt within each experiment is homogeneous (Table 3). The Ba content of the melts is slightly higher than in the starting material indicating that the experiments contained high amounts of melt (N 80%). The small decrease of Ba with increasing T is related to an increase in melt volume (Fig. 4b). Apart from Ba, a general increase of trace elements in the melt is observed with increasing T (Table 3 and Fig. 4b). The most systematic change in concentration is observed in Zr because the melts are buffered with zircon over the whole temperature range. This systematic change provides good evidence that the melts attained trace ...
Context 5
... Ba content of the melts is slightly higher than in the starting material indicating that the experiments contained high amounts of melt (N 80%). The small decrease of Ba with increasing T is related to an increase in melt volume (Fig. 4b). Apart from Ba, a general increase of trace elements in the melt is observed with increasing T (Table 3 and Fig. 4b). The most systematic change in concentration is observed in Zr because the melts are buffered with zircon over the whole temperature range. This systematic change provides good evidence that the melts attained trace element equilibrium. Hf shows a similar correlation to Zr, although the increase is less pronounced. Therefore Hf/Zr ...
Similar publications
The Itabira-Nova Era emerald province, where the Piteiras mine, the Capoeirana garimpo and the here studied Rocha mine are located, belongs to the southeastern-central portion of the Minas Gerais State, distant about 100 km east from the capital Belo Horizonte, halfway between the cities of Itabira and Nova Era. This province belongs to the Guanhãe...
Coesite-bearing eclogites from >100 km2 in the southern Dulan area, North Qaidam Mountains (NQM) of western China, contain zircon that records protolith crystallization and ultra high pressure (UHP) metamorphism. Sensitive High-Resolution Ion Microprobe (Mass Spectrometer) and Laser Ablation Inductively Coupled Plasma Mass Spectrometry U–Pb analyse...
We present new data on geochemistry, isotopic geochemistry, and geochronology of the Early Cambrian igneous rocks of the northeastern Siberian Craton (Kharaulakh anticlinorium, contact between the Siberian Platform and the West Verkhoyansk sector of the Verkhoyansk fold-and-thrust belt united into an Early Cambrian bimodal complex. This complex com...
The Xinghe graphite deposit is located in the north-central part of the North China Craton (NCC) where Archean gray gneiss, metasedimentary rocks (khondalites) and mafic (high-pressure) granulites constitute the dominant rock types. In this study, we report results of detailed geological and structural mapping around the Xinghe graphite mine. We id...
Garnet-bearing mafic granulites occur as lenses within TTG gneiss or metapelite in the Fuping Complex, middle Palaeoproterozoic Trans-North China Orogen. They are composed mainly of garnet, clinopyroxene, orthopyroxene, hornblende, plagioclase, quartz and accessary zircon and titanite. Four stages of metamorphic mineral assemblages have been found...
Citations
... In terrestrial sedimentary systems, zircon is chemically and physically resistant to weathering, thus becomes concentrated in clastic sediments (Vervoort et al., 2011). Zircon has low Lu/Hf and high Sm/Nd relative to equilibrium silicate melts (Rubatto and Hermann, 2007), thus with time zircon-rich clastic sediments can develop low ε Hf . Corresponding zircon-poor detrital sediments are Hf-poor, with high Lu/Hf ratios, hence incubate higher ε Hf (Fig. 5). ...
Alkaline-silicate complexes host some of the world's largest resources of rare-earth elements and high-field-strength elements (REE & HFSE) and represent the most fractionated magmatic systems on our planet. Geochemical evidence indicates that they are mantle melts, but while various studies highlight a role for lithospheric mantle, we do not know the precise origin of their contained REE and HFSE, and whether enrichment of the mantle source for these magmas can be attributed to specific geodynamic processes or events. We present new Nd-Hf isotope measurements (143 Nd/ 144 Nd & 176 Hf/ 177 Hf) made by LA-MC-ICP-MS, as well as a compilation of existing isotopic data for a suite of alkaline igneous rocks from the Gardar Province, a Mesoproterozoic continental rift in southern Greenland. Neodymium and hafnium isotopes are unaffected by crystal fractionation and can directly fingerprint the source of REE and HFSE. The dataset covers both phases of Gardar magmatism (1325-1261 and 1184-1140 Ma) and incorporates mafic dyke swarms and km-scale intrusive complexes, including Ilimmaasaq (Ilímaussaq) and Motzfeldt, which host some of the world's largest REE and HFSE deposits. The majority of Gardar complexes have a narrow range of positive median initial ε Nd (0 to +3.3) and ε Hf values (+0.2 to +6.0). Only two granite intrusions and the Eriksfjord basaltic lavas have crustally contaminated Nd-Hf isotope compositions, with the vast majority of Gardar igneous rocks preserving the isotope signature of their mantle source. Considering the diversity of rock types in the Gardar Province, initial ε Nd-ε Hf compositions are remarkably homogeneous, indicating a derivation of the Gardar's REE and HFSE from a laterally-extensive mantle melt source. Several Gardar systems have low initial ε Hf for a given ε Nd (Δε Hf to-9.7), a distinctive signature as few geological processes decouple the Nd and Hf isotope systems. The decoupled Nd-Hf isotope signatures are consistent with contributions from isotopically-matured phlogopite-bearing metasomatic veins (commonly known as PIC: phlogopite-ilmenite-clinopyroxene) in the lithospheric mantle. The metasomatising fluids that formed these source rocks were introduced via Palaeoproterozoic subduction, but the Gardar isotopic signatures indicate that REE and HFSE enrichment of these metasomes was not derived from subducted sediment; instead it is likely that metals were scavenged from the mantle wedge overlying the ancient subduction zone. The Gardar Nd-Hf isotope evolution trends overlap with a global compilation of kimberlites through time and allow us to tie the origin of the PIC metasomes to the regional geodynamic history of South Greenland. We identify PIC metasomes as a key metal source for the Gardar and by extension perhaps other REE-mineralised igneous provinces globally.
... The most common means of assessing possible metamorphic zircon growth is covariation of Th and U concentrations and Th/U ratios based on the relative diffusivity of each element in zircon (Lee et al., 1997;Hoskin and Ireland, 2000;Hoskin and Schaltegger, 2003;Rubatto and Hermann, 2007;Spencer et al., 2016;Rubatto, 2018). At conditions relevant to emplacement of the Cordillera Blanca batholith (<800 °C, <300 MPa; Atherton, 1992, 1996;Coldwell et al., 2011;Margirier et al., 2016), magmatic zircons should have Th/U between 0.1 and 10, whereas metamorphic zircons are expected to have Th/U <0.1 (Hoskin and Schaltegger, 2003;Yakymchuk and Brown, 2014;Kirkland et al., 2015;Rubatto, 2018;Yakymchuk et al., 2018). ...
The Cordillera Blanca and Cordillera Huayhuash contain some of the highest topography in the Andes and provide insight into tectonomagmatic processes associated with the onset of flat-slab subduction. These adjacent ranges shared a similar history of deformation and exhumation prior to the late Miocene, when synconvergent extension began in the Cordillera Blanca. Magmatism in the Cordillera Huayhuash has been inferred as coeval with magmatism in the Cordillera Blanca. Yet, extension, which has been correlated with magmatic heat flow, is limited to the Cordillera Blanca. New zircon U-Pb dates and trace and rare earth element concentrations from the Cordillera Blanca batholith and the Huayllapa pluton in the Cordillera Huayhuash and reassessment of existing zircon data help to characterize regional magmatic processes prior to the establishment of flat-slab subduction. Two compositionally distinct samples of the Huayllapa pluton yielded mean ages of 24.8 ± 0.4 Ma and 25.4 ± 0.8 Ma. In contrast, the Cordillera Blanca batholith has a protracted crystallization history postdating that of the Cordillera Huayhuash by up to 20 m.y. Miocene magmatism in the Cordillera Blanca began at 19 Ma and ended with injection of large volumes of geochemically distinct, mantle-derived magma from 10 to 5 Ma. We suggest that 6–5 Ma magmatism in the Cordillera Blanca promoted elevated heat flow and reduced shear strength, which facilitated extensional shearing along the western slopes of the range, whereas colder amagmatic crust in the Cordillera Huayhuash inhibited southward propagation of faulting. Our data demonstrate that the linkages between magmatism and elevated heat flow identified in the Cordillera Blanca are important driving processes in initiating extension in cordilleran-style orogenies.
... e Eruption of Star kimberlites that sampled xenoliths including all three garnet generations with the second garnet age group. Equilibrium between garnet and zircon can be tested from the partitioning of REE (Rubatto 2002;Rubatto andHermann 2007 and. Garnet patterns are fairly uniform with flat middle to heavy REE whereas zircon REE patterns are twofold: one group with flat middle to heavy REE patterns, the other with positively sloping REE (Fig. 7). ...
U–Pb ages were determined by split-stream LA-SF/MC-ICPMS in garnets from UHT granulite xenoliths (Star mine, South Africa; 124 Ma). They give a considerable age range of 400 million years with well-defined maximas at 3.09, 3.01 and 2.75 Ga. The oldest peak overlaps with the changeover from tonalites to K-granites at 3.14–3.04 Ga and with zircon ages of the mid-crustal granulites of the Vredefort dome (3.1 Ga) in the wake of the 3.2 Ga collision of three terrains that compose the Witwatersrand block. Subduction (or sagduction) of the uppermost crust in an ultrahot orogen setting brought shales and greenstones to the lower crust. Ultrahigh temperature (UHT) conditions are the result of high mantle potential temperatures and self- heating by the radioactive inventory of the subducted lithologies. Metamorphism, anatexis to very high degrees and melt extraction left UHT granulites as residue. Rejuvenation of UHT conditions was brought about by Dominion Group magmatism between 3.0 and 2.95 Ga. Magmatic uprise caused intense shearing in the lower crust followed by recrystallisation of the shear zones to generate the younger garnet age group. Ventersdorp flood basalt volcanism caused similar processes at around 2.72 Ga and generated the third garnet age group. Zircon gives U–Pb ages mainly around 2.72 Ga (both literature and our own data) i.e. zircon adjusted or newly crystallized at the youngest UHT event. Only few zircon grains retained older ages up to 2.94 Ga. Still unconstrained, but very high closure temperatures (≥ 1100 °C) for the U–Pb system in garnet keep the memory of the oldest ages in UHT granulites. Such ages can only be reset by recrystallization. This way, garnet records a prolonged high-temperature history of the lower crust of the Kaapvaal craton.
... and a positive Ce anomaly. Although all samples are garnet bearing, there are no zircons with a characteristic HREE depletion pattern in equilibrium with garnet (Rubatto, 2002, Rubatto andHermann, 2007). ...
... The presence of amphibole inclusions in zircons indicates that amphibole is not secondary mineral. The enrichment of HREE in zircons from the studied granulite xenoliths (Fig. 7) suggests their crystallization in the absence of garnet (Whitehouse, 2003;Rubatto and Hermann, 2007). Garnet should be formed during cooling of rocks at constant pressure, as suggested for mafic garnet granulite xenoliths from the Udachnaya kimberlite pipe (Perchuk et al., 2021). ...
... Petrogenetic linkage using REE provides the foundation for the indirect dating of garnet growth and recrystallization via micro-analysis of accessory minerals that are suitable for chronometry, for example, monazite and zircon. The partition coefficients for HREE between these phases have been quantitatively constrained (Hermann & Rubatto, 2003;Rubatto & Hermann, 2007;Taylor et al., 2015). Linking ages of accessory minerals to garnet growth using such constraints has become common practice in the tectonic analysis of orogenic belts (e.g., Foster et al., 2004;Holder et al., 2015;Kohn et al., 2005;Larson et al., 2013;Regis et al., 2016;Warren et al., 2019). ...
Incorporation of rare earth elements (REE) in garnet enables garnet chronology (Sm-Nd, Lu-Hf), and imparts a garnet-stable signature on cogenetic phases, which allows petrochronology and general petrogenetic tracing of gar-net stability in minerals and melts. Constraints on the uptake and redistribution mechanisms, as well as on the diffusive behaviour of REE in garnet are required for allowing accurate interpretation of REE signatures and ages. Gar-net REE profiles are often measured to gain insight into the nature and cause of REE zoning. Interpretation of such profiles is nevertheless complicated by poor constraints on the extent of diffusive relaxation. This is especially relevant for Lu, which, according to experiments, has a relatively high diffusivity and thus may re-equilibrate with possible consequences for Lu-Hf chronology. To provide new insight into the REE systematics of garnet, we applied quantitative trace-element mapping of garnet grains from metamorphic rocks that record peak temperatures above 750 C and cooling rates as low as 1.5 C Ma À1. Garnet in all samples preserves Rayleigh-type or oscillatory growth zoning with sharply defined interfacial angles that match the garnet habit. Re-equilibration of REE compositions appears restricted to domains with nebulous and patchy zoning, which likely form by interface-coupled dissolution and re-precipitation reactions mediated by fluids or melts, rather than REE volume diffusion. The possible effect of Lu diffusion in the analysed grains was investigated by comparing the observations to the results from 2D numerical modelling using Lu diffusivities from recent diffusion experiments. This test indicates that Lu diffuses significantly slower in natural garnet than experiments predict. The retentiveness of REE in garnet demonstrates the reliability of REE signatures in magmatic tracing and petrochronology and establishes Lu-Hf chronology as a robust means of dating garnet growth and recrystalliza-tion in metamorphic rocks, including those that underwent high-or ultrahigh-temperature conditions.
... To test this hypothesis, REE partition coefficients were calculated between the average composition of c. 1480-1235 Ma zircon and the average compositions of c. 1240 Ma garnet rim, mantle, and core ( Figure 12 and Table S9). Independent zircon-garnet D HREE values are not available for amphibolite facies rocks, although extrapolation of natural and experimentally determined high temperature D HREE estimates to subgranulite facies temperatures (e.g., Rubatto & Hermann, 2007;Taylor et al., 2017) suggests that the zircon-garnet-core D HREE values in this study may approach a possible equilibrium pair ( Figure 12). However, it is clear that zircon-garnet D HREE values for the garnet rim and mantle domains are F I G U R E 1 0 Calculated P-T pseudosections for the bulk-rock compositions of (a) sample FMC-1b and (b) sample FMC-2c. ...
Polymetamorphic metapelitic rocks in central-west Tasmania, southern Australia, contain high-pressure mineral assemblages that formed during Cambrian-aged subduction and relict garnet with published Lu-Hf ages of c. 1285-1240 Ma. These garnet ages, along with published detrital zircon data from throughout western Tasmania and western North America, have been used to propose the presence of Mesoproterozoic Laurentian crust in western Tasmania. In this study, we combine zircon petrochronology with compositional information from the inclusion assemblages in relict garnet to extract Mesoproterozoic pressure-temperature data from subduction-overprinted rocks, which effectively constitute an interpreted remnant of Laurentian crust now residing in central-west Tasmania. The new data suggest Mesoproterozoic metamorphism involved two stages. The first event is recorded by c. 1480-1235 Ma zircon that formed in a garnet-absent, plagioclase-present, high thermal gradient environment at pressures no greater than ~5-5.5 kbar. The second event recorded by c. 1285-1240 Ma relict garnet was characterized by the development of a moderate-pressure kyanite-plagioclase-biotite-bearing mineral assemblage, which formed at ~8.5 kbar and ~590-680 C. These pressure-temperature constraints are attributed to extension within a deep basin system associated with the cryptic East Kootenay Orogeny in North America, which coincides with the final stages of c. 1450-1370 Ma upper Belt-Purcell Basin sedimentation. Taking into account new detrital zircon U-Pb-Hf isotopic data from central-west Tasmania in this study and existing zircon provenance data from throughout western Tasmania and the Belt-Purcell Basin, our results strengthen the hypothesis of a Laurentian footprint that potentially encompasses much of western Tasmania and relates to both Nuna and Rodinian tectonism.
... The geochronological studies make use of a few different isotopic systems, including U-Th-Pb and Lu-Hf, with both being used to derive the age of Earth (4.543Ga) (White 2015). Outside of geochronological studies, the incorporation of lanthanides into the A-site of the zircon structure has proven useful in both geothermometry (Rubatto and Hermann 2007) and oxybarometry studies (Burnham and Berry 2012), while Ti incorporation in the T-site has found utility in zircon thermometry (Watson et al. 2006). ...
... Instead, they either thermally decompose into a mixture of their binary oxides (Bayer 1972;Strzelecki et al. 2021) or persist till melting (McMurdie and Hall 1947;Angapova, L.E. Serebrennikov 1973;Hikichi and Nomura 1987;Ushakov et al. 2001;Kolitsch and Holtstam 2004;Shin et al. 2006;Rubatto and Hermann 2007;Wang et al. 2010;Ding et al. 2020b;Ridley et al. 2022). ...
Zircon structure-types are ternary oxides with an ideal chemical formula of ATO4 (I41/amd), where usually A are lanthanides and actinides, with T = As, P, Si, V. This structure accommodates a diverse chemistry on both A- and T-sites, giving rise to more than seventeen mineral end-members of five different mineral groups, and forty-five synthetic end-members. Because of their diverse chemical and physical properties, the zircon structure-types are of interest to a wide variety of fields and may be used as ceramic nuclear waste forms and aeronautical environmental barrier coating, to name a couple. To support advancement of their applications, many studies have been dedicated to the understanding of their structural and thermodynamic properties. The emphasis in this review will be on recent advances in the structural and thermodynamic studies of zircon structure-type ceramics, including pure endmembers (i.e., zircon (ZrSiO4), xenotime (YPO4)), and solid solutions (i.e., ErxTh1-x(PO4)x(SiO4)1-x). Specifically, we offer an overview on the crystal structure, its variations and transformations in response to non-ambient stimuli (temperature, pressure and radiation), and correlation to thermophysical and thermochemical properties. We hope this review will promote further applications of these ceramic materials.
... To the best of our knowledge, LA-ICP-based techniques for compositional analysis of metal phosphate/phosphonate were not found in the literature. The LA-ICP-MS method gives direct isotopic information and quantification information (including for oxygen [15]), making it an ideal method when suitable standards are used. ...
Metal(IV) phosphate and phosphonates materials have increasingly found their applications in water purification, heterogeneous catalysis, drug delivery, and proton-exchange membrane fuel cells. The strong linkage between tetravalent metal cations and phosphate/phosphonate groups offers a unique bottom-up design platform, resulting in chemically stable inorganics or hybrids. Task-specific physiochemical functionalities could be deposited by modifying the phosphate/phosphonate groups before the material synthesis. The high reactivity between the metal centre and the phosphorus-containing linker, on the other hand, often leads to obtaining unordered materials (amorphous solids or coordination polymers). The chemical composition of the prepared materials is a key parameter in guiding the synthetic approach and in governing their performances. This narrative review focuses on critically summarising the traditional and advanced analytical methods for probing the composition of these materials. The reader is introduced to and guided on the advances and restrictions of different analysis techniques when probing metal(IV) phosphates/phosphonates. Both solution-based and solid-state spectroscopic techniques are covered with a focus on understanding the quantity and the linkage status of the phosphorus-containing moieties. These techniques include atomic spectroscopy, mass spectroscopy, nuclear magnetic resonance spectroscopy, X-ray-based methods, and neutron activation analysis.
... Although uncertainties remain due to hydrothermal alteration rarely being 100% efficient and the potential involvement of meteoric water (d 18 O of mid-latitude precipitation % À10‰) in the alteration, this calculation aligns with the interpretation that <30-40% assimilation can explain most of the oxygen isotope compositions of low-d 18 O magmas worldwide (Troch et al., 2020). Both experimental data (e.g., Rubatto and Hermann, 2007) and natural observations (Wang et al., 2010) consistently demonstrate a notable increase in Hf incorporation into zircon at lower temperatures. Zircon possesses an exceptionally high partition coefficient for Lu in felsic melts (Kd Lu > 400; Nardi et al., 2013). ...
... We observed no Mg# reversal in the coexisting garnet-supercritical fluid pair at higher temperatures, i.e., Mg# of supercritical fluid always exceeded that of garnet, though their values were almost equal at 7.8 GPa and 1090 °C. Meanwhile, Mg# reversal was reported for garnet-melt coexisting at 2.0 GPa and 1000 °C, as well as at 4.5 GPa and 1000-1050 °C (Rubatto and Hermann 2007;Hermann and Spandler 2008). The general affinity for TiO 2 incorporation changes from phengite > garnet≈supercritical fluid ≥ clinopyroxene at 5.5 GPa and 850-1030 °C to supercritical fluid > garnet > clinopyroxene at 7.8 GPa and 1090 °C (phengite being unstable). ...
The composition of fluid from carbonate- and chlorine-bearing pelite has been studied at 5.5 GPa, 850–1030 °C and at 7.8 GPa, 940–1090 °C using the diamond trap method. The run products include a residue composed of an eclogitic assemblage (garnet, coesite, clinopyroxene, and kyanite ± phengite and accessory pyrite/pyrrhotite, rutile, and zircon) and fluid solutes captured in the trap. The new data show that the pelite-derived supercritical fluid, with nearly equal amounts of solutes and H2O + CO2 [at the weight ratio H2O/(H2O + CO2) from 0.8 to 0.9], is stable at the applied P–T conditions. At higher temperatures, the amount of solutes in the supercritical fluid increases only slightly, apparently, due to the presence CO2 and 0.4–0.5 wt% Cl in the fluid. The reconstructed supercritical fluid composition includes components decreasing in the series SiO2 > Al2O3 > K2O > Na2O ≈ CaO ≈ MgO ≈ FeO. At 7.8 GPa, phengite becomes unstable, and K2O in the supercritical fluid increases from 3 to 8 wt% (on hydrous basis) while Al2O3 decreases from 8 to 5–6 wt%. Among the elements that fractionated into the fluid, B, Sr, Rb and P, as well as K at 7.8 GPa and 1090 °C, are the least compatible. The fluid-residue Sr partition coefficient varies from 4 to 10 and is notably lower at higher temperatures. Thus, supercritical fluids can form in carbonate- and chlorine-bearing sediments under subduction back arc-conditions, in cases of fluid-fluxing of the slab. Such supercritical fluids penetrating into the mantle together with H2O and CO2 can transport large amounts of major elements, B, Sr, Rb and P. The formation of potassium-rich silicic supercritical fluids is possible during subduction of pelite to ~ 250 km depths. They can be important agents in metasomatism of lithospheric mantle, with its composition reconstructed from data on micro-inclusions from fibrous diamond.
