Final Report Summary - NAMS (Water in the Earth's lithosphere and its bearing on geophysical properties)
There were 8 longer (at least 3 days) scientific visits and training activities in Europe (Lille, Toulouse, Montpellier, Zürich, Innsbruck, Padova and Trieste) and China (Hefei) also for students supervised in topics directly related to the project. These visits provided an opportunity to strengthen European scientific links and promote European science outside the European Union (EU). The researcher's other scientific activities included that he was a session convener at 2 international conferences, and he was also the executive scientific secretary of the 20th general assembly of the International Mineralogical Association (IMA) in Budapest. The researcher was awarded the significant Outstanding Young Scientist Award in the Geochemistry, Mineralogy, Petrology & Volcanology session of the European Geosciences Union in 2011. Also in this year, he was also recognised by the Junior Prima award in Science which is the highest distinction for scientist under the age of 35 in Hungary.
The proposal originally aimed to:
(a) investigate the diffusion in nominally anhydrous minerals (NAMs) to reveal whether the water content seen in upper mantle xenoliths are not modified significantly by secondary processes;
(b) reveal the distribution of water in the lithosphere by studying xenoliths from well-known alkaline basaltic localities in the Pannonian Basin;
(c) study the potential effect of water on geophysical and petrophysical properties.
Below the major achievements are summarised, respectively.
(1) It was shown that pyroxenes are likely to preserve their original water content, however, olivine may lose most if its original water content during the sampling by the host magma. Olivine is the most abundant mineral of the upper mantle and it controls its rheology. For these reasons, diffusion experiments were undertaken on natural olivines at varying temperatures (900 and 1 000 degrees of Celsius) and different durations (1 and 2 hours) under controlled conditions (Ar + H2 atmosphere) to study the kinetics of diffusion. The novelty of our approach was that unlike previous studies special attention was paid to different substitutions of 'water' into olivine. A natural olivine was selected for this task which has titanian (Ti)-clinohumite, trivalent and magnesium (Mg) vacancy substitutions. In other words the hydrogen substitutes in tetrahedral and octahedral vacancies charged balanced by either Ti or trivalent cations in an adjacent octahedral site if it is necessary. It was studied whether these substitutions, which have distinct infrared signatures, also have different diffusion properties. This is particularly interesting because in olivines from natural xenoliths normally the Ti-clinohumite substitution is the dominant one raising the possibility that the diffusion rate of this substitution may be slower than the other ones. The challenge was that the natural olivine was iron bearing and the water concentration was low (13 ppm).
The results indicated that with the applied experimental conditions the different substitution mechanisms show similar diffusion rates, which are similar to that previously determined for polaron diffusion (10-12 m2s-1). This implies that the presence of iron and fluid inclusions caused that through the oxidation of iron hydrogen was effectively and evenly expelled from both tetrahedral and octahedral vacancies. The results, on one hand indicate the importance of oxidation in facilitating diffusion, and, on the other hand, imply that to investigate the diffusion rates of different substitution mechanism more water rich and iron- and inclusion-free synthetic olivines are needed. All together the results are in line with previous studies indicating the fast diffusion of hydrogen in olivine and the role of oxidation, however, directed our attention to the importance if handling different substitutions mechanisms individually.
(2) The distribution of water in NAMs of the lithosphere was studied in two major alkaline basaltic occurrences: the Bakony-Balaton Highland (BBH) and the Nógrád-Gömör (NG) area in the Pannonian Basin. The two localities were selected on the basis that the BBH is in central part of the region, which is most affected by extension and there are several seismic and magnetotelluric profiles crossing the area. In contrast, the NG is at the periphery of the Pannonian Basin where the lithosphere thickens relatively abruptly and the area is at the contact between (Western) Carpathian and Inner Carpathian units. The geophysical coverage of the area is also satisfactory. The ultimate aim was to correlate the petrological and geochemical information (especially water content) obtained from xenoliths with seismic data collected from the area.
From the BBH 24 peridotite and 16 lower crustal granulite xenoliths were analysed for 'water' in their NAMs. Besides detailed petrography, the major and trace element compositions and electron backscatter diffraction (EBSD) patterns of NAMs were also determined in as many as possible xenoliths from this suite.
The range of water contents in NAMs are 0 - 13, 7 - 312 and 6 - 1 374 ppm for olivine, orthopyroxene and clinopyroxene, respectively. Olivines from the BBH have their major bands at 3 570, 3 525, 3 480, 3 440, 3 330, 3 225 cm-1 corresponding to the Ti-clinohumite, trivalent and Mg-vacancy substitution mechanisms. Orthopyroxene from the same locality display the main absorption bands at 3 580, 3 520, 3 420 cm-1 and smaller ones occasionally at 3 310 and 3 070 cm-1, whereas clinopyroxene shows main absorption bands at 3 630, 3 530, 3 460 and smaller ones occasionally at 3 350 and 3 230 cm-1. These bands are similar to infrared signatures reported for xenoliths worldwide.
Detailed geochemical and deformation analysis of these xenoliths from the BBH revealed that the present lithosphere, which suffered significant thinning in the Miocene, may be divided into two major layers based on the calculated equilibrium temperatures as indicators for the depth of origin. The shallower layer, from the MOHO to approximately 40 km depth, is characterised mostly by fine grained, equigranular to porphyroclastic xenoliths, generally displays an 'axial (010)' deformation pattern typical for transpressional deformation regime. Mineral constituents from this shallower layer show high Mg, low water (H2O) content in NAMs and depleted in basaltic major elements implying that this layer may have undergone considerable depletion. Trace element patterns, however, show enrichment most probably due to subsequent metasomatic enrichment episodes.
The deeper layer is below approximately 40 km and above the present lithosphere-asthenosphere boundary. The xenoliths show mainly coarse grained, protogranular texture with 'A-type' deformation pattern typical for asthenospheric flow. Minerals usually have lower Mg# and richer in basaltic major elements. The NAMs from this layer show higher H2O content than those in the shallow layer. Trace element patterns, on the other hand, do not refer to later refertilisation episodes by showing dominantly depleted pattern.
We suggest that the deeper, more H2O rich and less-depleted layer of the present day lithosphere is a juvenile one, which may have added to the lithosphere following the Miocene extension (round 10 Ma) in the thermal relaxation stage. The shallower layer, in contrast, may have undergone several episodes of depletion, deformation and re-fertilisation prior to and during the Miocene extension.
16 lower crustal mafic and felsic granulite samples with and without amphibole were studied from the BBH. This is the first time to our best knowledge, when all coexisting NAMs from a suite of well-studied lower crustal granulites are analysed for water in NAMs by FTIR. Garnet, orthopyroxene, clinopyroxene, quartz, feldspar, sillimanite, rutile were analysed. Garnet usually does not display any significant absorption band, and apart from two xenoliths (137, 24 ppm) have no detectable water. Orthopyroxene has three major bands at 3 570, 3 510 and 3 320 cm-1 and water concentration is in the range of 86 - 127 ppm. Feldspar has only one broad absorption band at approximately 3 200 and sometimes a smaller, but more characteristic one at approximately 3 570 cm-1 and the water concentration is between 8 - 322 ppm. Clinopyroxene only appears where modal amphibole is also present and it has absorption bands at 3 630, 3 510 and 3 460 cm-1 and water concentration ranges from 0 - 440 ppm. Sillimanite has major bands at 3 550, 3 460 and 2 660 cm-1 and water concentration is in the range of 217 - 340 ppm. Quartz shows three well-defined bands at 3 430, 3 380 and 3 320 cm-1 with a water concentration between 37 - 77 ppm. Rutile was also analysed and major absorption bands are at 3 325 and 3 295 cm-1 and water concentration varies from 24 - 1 047 ppm. This implies that these granulites are extremely 'dry' with respect to peridotites from the same locality and other granulites reported from China. This may indicate that the lower crust underwent considerable degree of partial melting probably due to the extension in the Miocene. The depletion in water owing to high degree of partial melting is further supported by the mineralogy of garnulites as hydrous phases are not present or only as relicts. Higher water contents were only revealed in the otherwise exceptionally dry NAMs in the vicinity of fluid and silicate melt inclusions.
From the NG 19 peridotite xenoliths were analysed for 'water' and, as for the BBH, the major and trace element compositions and EBSD patterns of NAMs were also determined in as many as possible xenoliths. The range of water contents in NAMs are 0 - 2, 3 - 112 and 80 - 895 ppm for olivine, orthopyroxene and clinopyroxene, respectively. The characteristic absorption bands are similar to those reported for the BBH, besides, amphibole lamellae occur more frequently in pyroxenes raising challenges with the accurate quantification of 'water' in NAMs. Correlation between calculated equilibrium temperatures, 'water' content and geochemistry is not as straightforward as for the BBH, which is probably related to the more complex setting of NG and the abruptly thickening lithosphere beneath the area. Note that there is a 50-kilometre distance between the sampled northernmost and southernmost basaltic volcanoes.
(3) The effect of 'water' on geophysical properties was investigated using seismic and heatflow data from the Pannonian Basin. Two cases were identified where 'water' in NAMs (and also their geochemistry) may be correlated with geophysical anomalies. In the first case the stability of pargasitic amphibole (a chain silicate with high water content 1.5 - 2 wt.% ) was demonstrated to play a critical role in the formation of the lithosphere - asthenosphere boundary in areas where the heat flow is high, such as the Pannonian Basin. It was shown recently (Green et al., 2010) that pargasitic amphibole is the most important water bearing phase in the upper mantle. Its melting and instability at 3 GPa (round 90 km of depth) or round 1 050 degrees of Celsius lead to a significant drop in the water storage capacity of the upper mantle. The melt or free fluid forming during the decomposition of pargasite weakens considerably the rheology of the upper mantle resulting in lower seismic velocity, higher conductivity and stronger anisotropy. If this petrologic model is correct, then we should see geophysical anomalies at approximately 90 km depth globally, expect places where the heat flow is high. In these areas, where the heat flow is higher than 60 - 70 mW / m2 - considering an average continental geotherm - the temperature may exceed the pargasite stability at approximately 1 050 degrees of Celsius in a depth shallower than 90 km. The Carpathian-Pannonian region (CPR) which is characterised by high heat flow, and both geophysical and geological data are abundant, is an excellent natural laboratory to test such a petrophyscial model. Depth of the 1 050 and 1 100 degrees of Celsius isotherm was calculated beneath the CPR from heat flow data and compared the depth of major geophysical anomalies indicating the lithosphere-asthenosphere boundary. It was shown that for most of the Pannonian Basin the difference between the calculated isotherms and the seismically and magnetotellurically constrained lithosphere-asthenosphere boundary is less than 5 km, indicating an excellent agreement between the petrophysical model and the geophysical constraints.
In the second case, it was found that there are seismic reflectors in the lithospheric mantle 10 - 15 km beneath the MOHO. These reflectors were found on several seismic sections of the CELEBRATION project (i.e. CEL05, CEL08) and their geological interpretation was ambiguous, until in this project a comprehensive dataset of upper mantle xenoliths was compiled. It seems that the seismic reflectors, is in fact, the boundary which separates the shallower more depleted part of the present day lithospheric mantle from the deeper more fertile one. The shallower part contains on average less water than the deeper region.
Besides the main goals of the project the novel quantitative infrared methodology was successfully applied to clay minerals in two projects. One was a research and development (R&D) project with the MOL Ltd. aiming at using the infrared spectrometry to obtain complementary information on clays from natural sediments with respect to traditional techniques such as X-ray powder diffraction and thermogravimetry. This was made possible by the application of a new preparation technique which liberates most of the absorbed water and removes its unwanted spectral contribution. With the application of this preparation different clay minerals can be identified more accurately and quantitative parameters extracted from the infrared spectra can be correlated against other parameters such as cation exchange capacity.
The other project applied infrared spectrometry to study sediments form boreholes drilled through the sliding wall of a major landslide near the river Danube at Kulcs. The preliminary results indicated that infrared spectrometry could be an effective tool to monitor characteristic changes, such as the decrease in carbonates and increase in clay minerals at the sliding wall. This has great potential for the identification of areas underlain by layers which can initiate landslides if the hydrogeological and geochemical situation is such. The latter two projects may provide basis for future European-scale projects to investigate the effect of clay minerals on the stability of damns and other socially and economically important constructions. A future European Research Council (ERC) proposal would greatly facilitate these efforts. The construction of the Pannonian Uniform Lithospheric Infrared spectral database is also related to the project (see http://puli.mfgi.hu online for further details), which also serves as a website for the present Marie Curie project.