abstract: Volcanic devolatilization is one of the major processes in the global nitrogen cycle. Past studies have often estimated the magnitude of this flux using volcanic emission measurements, which are limited to currently active systems and sensitive to atmospheric contamination. A different methodological approach requires appropriate analytical parameters for nitrogen analysis in silicate glasses by secondary ion mass spectrometry (SIMS), which have not yet been established. To this end, we analyze various ion implanted basaltic and rhyolitic glasses by SIMS. We demonstrate that water content significantly affects the ion yields of 14N+ and 14N16O−, as well as the background intensity of 14N+ and 12C+. Application of implant-derived calibrations to natural samples provide the first reported concentrations of nitrogen in melt inclusions. These measurements are from samples from the Bishop Tuff in California, the Huckleberry Ridge Tuff of the Yellowstone Volcanic Center, and material from the Okaia and Oruanui eruptions in the Taupo Volcanic Center. In all studied material, we find maximum nitrogen contents of less than 45 ppm and that nitrogen concentration varies positively with CO2 concentration, which is interpreted to reflect partial degassing trend. Using the maximum measured nitrogen contents for each eruption, we find that the Bishop released >3.6 x 1013 g of nitrogen, the Huckleberry Ridge released >1.3 x 1014 g, the Okaia released >1.1 x 1011 g of nitrogen, the Oruanui released >4.7 x 1013 g of nitrogen. Simple calculations suggest that with concentrations such as these, rhyolitic eruptions may ephemerally increase the nitrogen flux to the atmosphere, but are insignificant compared to the 4 x 1021 g of nitrogen stored in the atmosphere. / Dissertation/Thesis / Masters Thesis Geological Sciences 2016

H2O and CO2 concentrations of the glass phase in melt inclusions (MI) are commonly used both as a barometer and to track magma degassing behavior during ascent due to the strong pressure dependence of H2O and CO2 solubilities in silicate melts. A requirement for this method to be valid is that the glass phase in the MI must represent the composition of the melt that was originally trapped. However, melt inclusions commonly contain a vapor bubble that formed after trapping. Such bubbles may contain CO2 that was originally dissolved in the melt. In this study, we determined the contribution of CO2 in the vapor bubble to the overall CO2 content of MI based on quantitative Raman analysis of the vapor bubbles in MI from the 1959 Kilauea Iki, 1960 Kapoho, 1974 Fuego volcano, and 1977 Seguam Island eruptions. The bubbles contain up to 90% or more of the total CO2 in some MI. Reconstructing the original CO2 content by adding the CO2 in the bubble back into the melt results in an increase in CO2 concentration by as much an order of magnitude (1000s of ppm), corresponding to trapping pressures that are significantly greater (by 1 to >3 kbars) than one would predict based on analysis of the volatiles in the glass alone. Many MI also showed the presence of a carbonate mineral phase; failure to include its contained CO2 when reconstructing the CO2 content of the originally trapped melt may introduce significant errors in the calculated volatile budget. / Master of Science

Understanding volatile evolution associated with active volcanic magmatic systems is of paramount importance because volatiles control and determine the magnitude of an eruption owing to the large change in molar volume that volatile species show depending on their physical state (volatiles dissolved in silicate melts vs. volatiles exsolved as vapor). For active volcanic systems studying the volatile evolution can help to assess the potential hazard associated to a certain locality. Also, volatile evolution in magmatic system controls the formation of certain ore deposits. Despite the importance of understanding volatile evolution of magmatic systems, concentrations of volatiles of evolving magmas are not easily available especially for magmas originated in the deep crust. Fortunately, sample of melts can be entrapped as melt inclusion (MI) into growing igneous minerals in crystalizing magma chamber. After the entrapment, the crystal works as an insulating capsule from the external magmatic environment. Researchers have started to use MI because they provide some advantages in respect to the classical whole rock approach to petrological studies. One of the most important advantages is that MI often represent sample of a deep and non-degassed melt (glass) available at Earth's surface. In fact, with the exception of deep ocean basalts, igneous whole rocks found at the Earth's surface are degassed magmas. This dissertation is a compilation of four publications produced during six years of research and is addressed to give a contribution in understanding the volatile evolution in magmatic systems and also to improve the present understanding of information that can be obtained using the melt inclusions technique. In the first chapter, I present an alternative interpretation of H₂O-CO₂ trends obtained from MI. In this study, we demonstrate that these trends can be due to post entrapment crystallization on the wall of the MI and not to magma ascent. This alternative view is more realistic especially for cases where in the same phenocrysts MI show strongly different CO₂ concentrations. In the second chapter, I present a study to test for the MI reliability in recording volatile concentrations. We used the approach of the melt inclusion assemblage (MIA) that consists of analyzing groups of MI presumably entrapped at the same time and, thus, at same chemical and physical conditions. The results show that most of the MIA studied show consistent volatile concentrations corroborating the reliability of the MI technique. CO₂ shows the highest degrees of variability and we have assessed this behavior mostly to C-contamination in the surface of the sample. The third chapter is a study case (the Solchiaro eruption in Southern Italy) that shows the potential uses of MI to understanding the volatile evolution. I present a model showing the dynamic of the magma based on MI. This study also discusses the origin of anomalous MI and which MI provide the best information. The final chapter is dedicated to test the applicability of the new Linkam TS1400XY heating stage. I was able to show how this new microthermometric tool is capable of homogenizing MI at high temperature and to quench MI to a homogeneous glass state. / Ph. D.

The central segment of the Taupo Volcanic Zone (TVZ) is one of the world’s most productive areas of silicic volcanism and geothermal activity. Rhyolites largely predominate the eruptive output in the central TVZ, with only minor basalts, andesites and dacites. The rhyolites show diversity in composition, and form a compositional continuum between two end-member types (R1 and R2), as suggested in previous studies. In this thesis I present results from a quartz- (and rare plagioclase-) hosted melt inclusions study, focussing on the volatile concentration (i.e. H2O, Cl, F, CO2) and their relative distribution between R1 and R2 rhyolites. The main objective is to add further constraints on the magmatic systems with regard to their contribution to the hydrothermal systems in the central TVZ. A comparative study between R1 and R2 melt inclusions show distinct volatile, fluid-mobile, and highly incompatible element compositions. Differences in the bulk volatile concentration of the parental magmas (i.e. basalts intruding the lower crust) are suggested to be at the origin of these volatile disparities. Further analysis on the volatile exsolution of R1 and R2 melts lead to the observation that the two rhyolite types exsolve a volatile phase at different stages in their magmatic history. From Cl and H2O concentrations, it is suggested that R1 magmas exsolve a vapour phase first, whereas R2 rhyolites more likely exsolve a hydrosaline fluid phase. These results have considerable implications for the magmatic contribution into the hydrothermal systems in the central TVZ, as differences in the composition of the resulting volatile phase may be expected. The hydrothermal systems in the central TVZ are subdivided into two groups based on their gas and fluid chemistry; and the current model suggests that there are two distinct contributions: a typical ‘arc’ system, with geochemical affinity with andesitic fluids, located along the eastern margin of the TVZ, and a typical ‘rift’ system, with geochemical affinity with rhyolitic/basaltic fluids, located along the central and/or western region of the TVZ. The addition of the new data on the rhyolitic melt inclusions, leads to a re-evaluation of the magmatic contribution into the hydrothermal systems, with a particular focus on B and Cl. The results indicate a more diverse variety of contributions to the meteoric water in the hydrothermal systems, and also show that the east-west distribution of ‘arc’ and ‘rift’ fluids is not a viable model for the central TVZ. This work emphasises that melt inclusion data and their volatile degassing history cannot be underestimated when characterising and quantifying the magmatic component in hydrothermal fluids. The melt inclusion data also provide further insight into the pre-eruptive magmatic plumbing systems and are particularly important from a hazard perspective. Included in the thesis is a detailed petrological analysis of rhyolite melt inclusions across the central TVZ and an interpretation that large silicic magma systems (in the TVZ) are typically comprised of multiple batches of magma emplaced at some of the shallowest depths on Earth. Tectonic activity is suggested to play an important role in triggering large caldera-forming eruptions as the evacuation of one magma batch could cause a regional-scale readjustment that is sufficient enough to trigger and allow simultaneous eruption of an adjacent melt batch.

This study is concerned with volcanostratigraphic and petrologic evolution of the S&uuml / phan, which is a 4050 m high Quaternary stratovolcano in eastern Anatolia. The eruptive products of S&uuml / phan Stratovolcano, including transitional mildly alkaline to calc-alkaline rocks, are lavas, domes and pyroclastics ranging in composition from basalts to rhyolites. Ar-Ar age data from different levels of the volcanostratigrafic succession yield a range of 0.76-0.06 Ma. Textural features, wide temperature ranges obtained for intermediate members, and the linear trends of whole-rock geochemistry are strongly suggestive of magma mixing in the evolution of S&uuml / phan volcanics. Presence of crystal clots in many lavas suggests that cogenetic plutonic rocks were also involved in the mixing process. Comparison of whole-rock, melt inclusion and glass chemistry data of S&uuml / phan to data from experimental studies reported in literature indicate that the melt inclusions describe true liquid lines of descent from a common hydrous parent at pressures of ~500 MPa. EC-AFC modeling of trace element and isotopic compositions reveals 2-8% crustal contamination in the differentiated lavas. REE modeling indicates that primitive rocks of S&uuml / phan volcanics were products of mixing of melts from spinel and garnet lherzolite sources, with contributions of 60% and 40%, respectively, in the mixture. A two-stage petrogenetic model is proposed for Suphan stratovolcano. Mantle- derived melts stall and undergo chemical differentiation in a deep hot zone in lower to mid-crust / variably evolved melts ascending from this zone are arrested and mixed at a shallow level where they construct a sub-volcanic magma reservoir beneath Suphan.

Novel techniques for in-situ lead isotope analysis of melt inclusions using multicollector inductively coupled plasma mass spectrometry, and parallel Faraday cup and ion counter detection have been developed. These provide for measurement of the critical 204Pb isotope, which has been unavailable in melt inclusion Pb isotope studies, but requires on-line determination of 200Hg to correct for 204Hg isobaric interferences. External standardisation allows for effective mass fractionation correction and ion counter gain calibration. An off-line Tau correction was applied and standard glass analyses suggest that this correction provides more accurate and precise results.

京都大学 / 新制・課程博士 / 博士(理学) / 甲第24430号 / 理博第4929号 / 新制||理||1704(附属図書館) / 京都大学大学院理学研究科地球惑星科学専攻 / (主査)准教授 河上 哲生, 教授 下林 典正, 教授 田上 高広 / 学位規則第4条第1項該当 / Doctor of Science / Kyoto University / DFAM

Understanding the different igneous processes that magmas undergo is important for a variety of reasons including potential hazards associated with volcanoes in populated regions, magmatic hydrothermal ore deposition, and tectonic processes. One method of obtaining geochemical data that can help constrain petrogenetic processes is through the study of melt and fluid inclusions. The research presented here examines melt inclusions through experimental, analytical and field studies to better understand igneous petrogenesis. One potential problem associated with melt inclusions is water-loss during laboratory heating. A Raman spectroscopic technique was developed to determine water contents of silicate glasses, and this technique was applied to monitor water loss from natural melt inclusions that were heated for varying lengths of time. The results suggest that water loss is insignificant when heated for less than 12 hours but significant water loss can occur with longer duration heating. The distribution of trace elements between silicate melts and phenocrysts growing from that melt can constrain igneous processes such as fractional crystallization, assimilation, and partial melting. Partition coefficients were determined for syngenetic clinopyroxene, orthopyroxene, and plagioclase in equilibrium with a dacitic melt using the Melt Inclusion-Mineral (MIM) technique. Melt inclusion chemistry is the same regardless of mineral host phase, suggesting that the melt inclusions have not been subjected to re-equilibration processes or boundary layer development. Partition coefficients from this study are similar but typically lower than published values. Three closely-spaced monogenetic eruptive units from the active Campi Flegrei volcanic system (Italy) with similar eruptive styles were examined to better understand the evolution of the magmatic system. Results suggest fractional crystallization as the dominant process taking place over time but that magma mixing was significant for one of the eruptions. Trace element geochemical data suggest a mixed magma source of within-plate and volcanic arc components, and still retain a T-MORB signature from the subducting slab. / Ph. D.

Silicate melt inclusions are ubiquitous in quartz phenocrysts, yet there are few studies of such inclusions from porphyry copper systems. A melt inclusion forms when magma is trapped in a growing phenocryst. If a phenocryst is able to preserve the original parent magma, then accurate information can be obtained for ancient volcanic systems. In recent igneous systems, melt inclusions are commonly preserved as optically clear homogeneous glass representative of magma stored at depth before eruption. Melt inclusions are difficult to recognize in quartz phenocrysts from porphyry copper system because they are crystalline and hidden by exsolved magmatic volatiles. The inclusions range in size from less than 5 to over 150 μm. In order to evaluate the magmatic contribution to economic mineralization, we conducted three separate studies to determine whether or not crystallized melt inclusions preserve representative samples of magma. The first study modeled the phase relationships that occur during equilibrium crystallization and melting of haplogranite magma trapped in quartz. Results from the model are similar to observations made during the heating of crystallized melt inclusions from porphyry copper systems. It is necessary to re-melt the crystal and volatile phases before chemical analysis. Micro-explosions caused by heating resulted in the loss of important chemical components. Our second study evaluated several microthermometric heating procedures using synthetic melt inclusions trapped at conditions similar to those inferred for porphyry copper systems. A synthetic hydrous melt was saturated with saline hydrothermal solutions allowing both melt and aqueous fluids to be trapped in quartz. Based on microthermometric measurements from these coeval melt and aqueous fluid inclusions we were able to predict the known trapping temperature and pressure of formation. This technique can be applied to natural samples to constrain trapping pressures and temperatures. It was found that slower heating rates could be used to avoid overheating and that heating under a confining pressure greatly minimizes the decrepitation of inclusions. The third study examined the copper concentrations in melt inclusions from the Red Mountain, Arizona porphyry copper system. Older andesite magma contains pyroxene with melt inclusions of higher copper concentrations compared to melt inclusions in quartz from quartz latite. The higher water concentrations in crystallized melt inclusions in the quartz, and abundant aqueous fluid inclusions indicates that the exsolution of water from the magma occurred prior to the trapping of melt inclusions in quartz. The lower water concentrations and the absence of aqueous fluid inclusions indicates that the andesite never reached the stage of water exsolution. The results obtained here are consistent with models that suggest that copper is extracted from the melt by saline magmatic fluids, producing a metal-charged hydrothermal solution and leaving behind a metal-depleted melt and serves to identify the potential contribution of melt inclusion studies to constrain the origin of ore metals in porphyry copper deposits. / Ph. D.

L’étude des basaltes des îles océaniques (OIB) révèle la complexité du manteau terrestre, dont la composition chimique est hautement variable. Décrypter l’implication des lithologies des roches sources et des processus à l’origine des OIB est complexe car les magmas sont transformés lors de leur ascension jusqu’à la surface. Ceci est particulièrement critique dans le cas des îles Canaries, où la lithosphère est considérée comme particulièrement épaisse (>110 km Fullea et al., 2015). Afin de mieux contraindre la composition chimique des magmas primitifs et les lithologies mantelliques plausiblement impliquées, deux éruptions historiques de l’île de Lanzarote, les éruptions de Timanfaya (1730-1736) et celles de 1824, ont été étudiées. En effet, ces deux éruptions offrent une opportunité unique d’étudier les mécanismes de génération des magmas et leurs compositions dans un contexte où le manteau est hétérogène. L’éruption de historique de Timanfaya (1730-1736) a émis des magmas qui ont évolué de basanites à basaltes alcalins, pour atteindre des compositions tholeitiques à la fin de l’éruption. La dernière éruption de l’île, en 1824, a produit des basanites extrêmement riches en volatils. L’hétérogénéité du manteau est démontrée à l’extrême à Lanzarote où une seule éruption présente une variation de compositions chimiques équivalente à la diversité de celles des OIB dans le monde. L’extrême hétérogénéité est systématique pour les compositions des roches totales et des téphras à l’échelle d’une éruption, mais est encore amplifiée à l’échelle du minéral et des inclusions magmatiques contenus dans un même échantillon de téphra.Les concentrations des éléments traces et leurs rapports dans l’olivine (e.g. Ni, Mn et Ca) sont de précieux marqueurs des lithologies mantelliques à l’origine des magmas. En effet, les rapports Ni x (FeO/MgO), corrigé du fractionnement, et Fe/Mn sont des indicateurs de lithologies avec ou sans olivines. Il est considéré qu’ils peuvent montrer, dans la plupart des cas, l’ajout d’un liquide magmatique dérivé d’une pyroxénite dans les magmas primaires. La mesure des compositions des éléments traces des olivines des éruptions de 1730-1736 et de 1824 montrent les implications variables de plusieurs lithologies mantelliques au cours du temps. Lors de la fusion d’une lithologie ne contenant pas d’olivines, comme la pyroxénite, de hautes teneurs en Ni et de faibles teneurs en Mn et Ca sont attendues. Les basanites de Lanzarote présentent les plus grandes variations géochimiques, couvrant le champ compositionnel des olivines des MORB et des OIB du monde entier, tandis que les produits plus tardifs, c’est-à-dire les basaltes alcalins et les tholéites, ont des teneurs typiques de liquides magmatiques dérivés de pyroxénites. Les teneurs en forstérite (Fo) des olivines diminuent systématiquement avec le temps durant l’éruption de 1730-1736 et la proportion de liquide primaire saturé en silice augmente dans les mélanges de magmas primitifs avec le temps. A la fin de l’éruption, les magmas tholéitiques cristallisent des olivines dont la teneur en Fo est faible, alors que les concentrations en Mn et Ca augmentent simultanément avec le rapport Ca/Al pour des rapports Fe/Mn et Ni x (FeO/MgO) relativement constants. Ces observations sont expliquées par une augmentation de la fusion par décompression à température légèrement plus faible. D’autre part, les basanites de l’éruption de 1824 possèdent les olivines ayant les teneurs en Fo les plus élevées, et des teneurs en éléments traces dépassant la variabilité des basanites de l’éruption de Timanfaya. Le fait que les basanites de Lanzarote contiennent des olivines dont les compositions en éléments traces recouvrant le champ des MORB et des liquides pyroxènitiques est expliqué par la fusion d’une source contenant des lithologies hétérogènes, induite par un flux de CO2, générant ainsi des magmas aux compositions diverses. (...) / The study of oceanic island basalts (OIB) reveals the complexity of the mantle, which composition is highly variable. Deciphering the source lithologies and processes involved in the OIB formation is challenging since the magmas are transformed on their way to the surface. This is especially critical at Canary Islands where the lithosphere is thought to be remarkably thick (>110 km Fullea et al., 2015). In order to better constrain the composition of primitive magmas and the plausible mantle lithologies involved, two historical eruptions recorded at Lanzarote island, Timanfaya 1730-1736 and 1824 eruptions have been investigated. Indeed, these two eruptions offer a unique opportunity to investigate the mechanisms of magma generation and composition in the context of mantle heterogeneity. The Timanfaya, 1730-1736 historical eruption emitted magmas that evolved from basanites through alkali basalts, finally reaching tholeiitic compositions at the end of the eruption. In 1824 the last eruption on the island produced extremely volatile-rich basanite. The heterogeneity of the mantle is demonstrated to the extreme in Lanzarote where a single eruption exhibits compositional variations similar to the span of the OIB worldwide. The extreme heterogeneity is systematic from whole rock lava and tephra at eruption scale but amplified at mineral and melt inclusion scale within a single tephra sample of the eruption.The use of trace element concentrations and ratios of olivine (e.g. Ni, Mn, and Ca) are valuable indicators of the mantle source lithology, namely, the fractionation-corrected Ni x (FeO/MgO) and Fe/Mn as probes of olivine absent or present lithologies, often taken as pyroxenite-derived component in mixtures of primary melts. The measured trace element concentrations in olivine from the 1730-1736 and 1824 eruptions reveal variable mantle lithologies involved in the magma generation with time. Higher Ni and lower Mn and Ca contents are expected when melting Ol-free source, such as pyroxenite lithologies. The basanites exhibit the largest variation covering the range of olivine in MORB and OIB worldwide whereas later produced alkali-basalts and tholeites have values typically expected from pyroxenite derived melts. The Fo content decreases systematically with time during the 1730-36 eruption and the proportion of silica-saturated primary melt increased in the parental magma mixture with time. At the end of the eruption, tholeiite magmas crystallized olivine with lower Fo content, whereas those concentrations of Mn and Ca increased together with Ca/Al at relatively uniform Ni x (FeO/MgO) and Fe/Mn, all of which is readily explained by increased decompression melting at slightly lower temperature. The basanite from the eruption that took place in 1824 has olivine with the highest Fo content and trace element variability expanding the range of the Timanfaya basanite. The fact that Lanzarote basanites contain olivine with trace element systematic spanning that of MORB and pyroxenite melt is explained by CO2-flux melting of a lithologically heterogeneous source, generating the diverse compositions. In addition, early reactive porous flow through the depleted oceanic lithosphere and equilibration with harzburgite restite caused Ni depletion of the earliest percolating pyroxenite melt from which olivine crystallized and probably leaving dunite channels. After the channel formation mantle nodules could be brought to the surface. The fact that olivine compositions and basanite magma were reproduced approximately a century later may reflect episodic carbonatic fluxing in the slowly uprising Canarian mantle plume. (...)