Teleseismic data recorded by broad-band seismic stations in the Swedish National Seismic Network (SNSN) have been used in a suite of studies of seismic wave conversion in order to assess the structure of the crust and upper mantle beneath the Baltic Shield. Signals of seismic waves converted between P and S at seismic discontinuities within the Earth carry information on the velocity contrast at the converting interface, on the depth of conversion and on P and S velocities above this depth. The conversion from P to S at the crust-mantle boundary (the Moho) provides a robust tool to constrain crustal thicknesses. Results of such analysis for the Baltic Shield show considerable variation of Moho depths and significantly improve the Moho depth map. Analysis of waves converted from S to P in the upper mantle reveals a layered lithosphere with alternating high and low velocity bodies. It also detects clear signals of a sharp velocity contrast at the lithosphere-asthenosphere boundary at depths around 200 km. Delay times of P410s, the conversion from P to S at the upper mantle discontinuity at 410 km depth, were used in a tomographic inversion to simultaneously determine P and S velocities in the upper mantle. The polarisation of P410s was also used to study anisotropy of the upper mantle. Results of these analyses are found to be in close agreement with independently derived results from arrival time tomography and shear-wave splitting analysis of SKS. The results presented in this thesis demonstrate the ability of converted wave analysis as a tool to detect and image geological boundaries that involve sharp contrasts in seismic properties. The results also show that this analysis can provide means of studying aspects of Earth’s structure that are conventionally studied using other types of seismic data.

Déformation dans le manteau partiellement fondu: observations dans des systèmes naturelsNous avons analysé les interactions entre magmas et déformation dans le manteau supérieur par une étude microstructurale détaillée de deux sections de péridotites: un profil long de 100m dans le Moho – zone de transition croûte-manteau (MTZ) de l'Ophiolite d'Oman - et des lherzolites à plagioclase dans le massif de Lanzo (Alpes Occidentales). La section de la MTZ se caractérise par un litage à l'échelle cm à métrique ; les compositions varient de dunites à des gabbros à olivine. Le parallélisme entre le litage et la foliation, des limites de couche diffuses, et l'alignement de lentilles riches en plagioclase d'échelle mm impliquent un litage contrôlé par la déformation. Tous les minéraux majeurs ont des orientations préférentielles des cristaux (ou fabrique), mais la fabrique de l'olivine varie en fonction de la composition de la couche. Les lits avec <70% olivine montrent des fortes concentrations de l'axe [100] parallèle à la linéation (pattern axial-[100]). Les lits avec <50% olivine ont [100] dispersé dans la foliation et une concentration des axes [010] autours d'une direction normale à cette dernière (fabrique axiale-[010]). Ces changements ont lieu de façon répétée à l'échelle du mm, aux limites de couches, indiquant une déformation en présence de teneurs variables de magma. La fabrique dans les lits riches en olivine est cohérente avec déformation par fluage dislocation à haute température. Les fabriques d'olivine avec une symétrie axiale-[010] dans les lits pauvres en olivine impliquent une transpression ou un glissement le long de limites de grain (010) « lubrifiés » par du magma; elles sont accompagnées par une activité accrue du glissement [001]. Parce que le changement des fabriques d'olivine n'est pas associé à une dispersion de ces dernières, les teneurs instantanées de magma doivent être <30-40%. La variation continue de la symétrie de la fabrique d'olivine CPO avec la décroissance de la teneur en olivine implique que la fabrique dépend de la déformation finie cumulée en présence de magma, plutôt que de la teneur instantanée de magma.Les lherzolites à plagioclase de Lanzo se caractérisent par un réseau anastomosé de bandes riches en plagioclase, qui passe localement à un litage planaire avec des bords diffus. Olivine et orthopyroxene montrent des évidences de déformation par fluage dislocation à haute temperature: extinction onduleuse, sous-grains ou kinks, des limites de grain irréguliers. Le plagioclase forme des agrégats allongés composés par des cristaux avec des formes interstitielles. Tous les minéraux majeurs (à l'exception du clinopyroxene) ont une claire fabrique cristallographique, indépendamment de la composition de la couche. La fabrique de l'olivine CPO a une symétrie orthorhombique, avec [100] parallèle à la linéation et [010] normal à la foliation (et au litage), mais les maxima de [010] sont toujours plus forts que ceux de [100]. Les fabriques d'olivine sont donc intermédiaires entre les deux patterns observés en Oman, ce qui est cohérent avec le litage plus diffus observé à Lanzo. Ces données impliquent que dans un manteau soumis à une déformation cisaillante, les magmas tendent à se concentrer dans des bandes parallèles au plan de cisaillement, formant un réseau anastomosé qui peut évoluer dans un litage planaire. Les fractions instantanées de magma restent faibles (inférieures au seuil pour la disruption de la matrice solide) – la déformation est donc accommodée largement par fluage dislocation. La présence de magmas produit toutefois un changement dans la symétrie des fabriques d'olivine – produisant des patterns axial–[010] et donc dans une anisotropie sismique différente. La ségrégation du magma et le développement d'un litage compositionnel contrôlé par le cisaillement peut aussi créer une anisotropie mécanique, avec une réduction de la viscosité cisaillante parallèlement aux bandes. / Deformation in a shallow partially molten mantle: constraints from natural systemsWe analysed the interactions between melts and deformation in the upper mantle through detailed microstructural studies in two settings: a 100m section in the Moho Transition Zone (MTZ) of the Oman Ophiolite and plagioclase-rich lherzolites from the Lanzo massif (Alps). The MTZ section is characterized by a cm to m-scale layering; compositions range from dunites to gabbros à olivine. Parallelism between the layering and the foliation, diffuse layers limits, and alignment of mm-scale plagioclase-rich lenses in intermediate composition layers imply shear-controlled layering. All phases have crystal-preferred orientations (CPO), but olivine CPO varies depending on the layer composition. Layers with <70% olivine show strong [100] maxima parallel to the lineation (axial-[100] pattern). Layers with <50% olivine have [100] dispersed in the foliation and concentration of [010] normal to it (axial-[010] pattern). These changes occur repeatedly on the mm-scale, at the layers limits, providing strong evidence for deformation in presence of variable melt fractions. CPO in olivine-rich layers is consistent with high-temperature dislocation creep. Axial-[010] olivine CPO patterns imply transpression or sliding along preferentially-wetted (010) grain boundaries in melt-rich layers; they are accompanied by increased activity of [001] glide. Since the change in CPO symmetry is not associated with dispersion, instantaneous melt fractions must have remained <30-40%. The continuous variation in olivine CPO symmetry with decreasing olivine content implies therefore that the CPO depends on the strain cumulated in presence of melt rather than on the instantaneous melt fraction.Plagioclase-rich lherzolites in Lanzo are characterized by an anastomozed network of plagioclase-rich layers, which locally grades into a diffuse cm-scale planar compositional layering. Olivine and orthopyroxene show evidence for deformation by dislocation creep under high temperature conditions: undulose extinction, subgrains or kinks, sutured grain boundaries, but plagioclase forms elongated aggregates of crystals with interstitial shapes. All major phases (except clinopyroxene) show clear CPO, independent of the layer composition. The olivine CPO has an orthorhombic symmetry, with [100] parallel to the lineation and [010] normal to the foliation (and to the layering), but [010] maxima are stronger than [100] ones. These olivine CPO are intermediate between the two patterns observed in the Oman MTZ, consistently with the less developed segregation in Lanzo. Together these data imply that melts tend to concentrate in bands aligned in the shear plane, forming an anastomosed network that may grade with increasing shear into a planar layering. Instantaneous melt fractions remain below the solid matrix disruption threshold; the deformation is still essentially accommodated by dislocation creep in the solid matrix. Deformation in presence of melt results, however, in different olivine CPO with axial–[010] patterns and hence in a different seismic anisotropy. Shear-controlled compositional layering may also create a mechanical anisotropy, with a directional reduction of the shear viscosity parallel to the bands.

<p>Teleseismic data recorded by broad-band seismic stations in the Swedish National Seismic Network (SNSN) have been used in a suite of studies of seismic wave conversion in order to assess the structure of the crust and upper mantle beneath the Baltic Shield. Signals of seismic waves converted between P and S at seismic discontinuities within the Earth carry information on the velocity contrast at the converting interface, on the depth of conversion and on P and S velocities above this depth. </p><p>The conversion from P to S at the crust-mantle boundary (the Moho) provides a robust tool to constrain crustal thicknesses. Results of such analysis for the Baltic Shield show considerable variation of Moho depths and significantly improve the Moho depth map. Analysis of waves converted from S to P in the upper mantle reveals a layered lithosphere with alternating high and low velocity bodies. It also detects clear signals of a sharp velocity contrast at the lithosphere-asthenosphere boundary at depths around 200 km. </p><p>Delay times of P410s, the conversion from P to S at the upper mantle discontinuity at 410 km depth, were used in a tomographic inversion to simultaneously determine P and S velocities in the upper mantle. The polarisation of P410s was also used to study anisotropy of the upper mantle. Results of these analyses are found to be in close agreement with independently derived results from arrival time tomography and shear-wave splitting analysis of SKS.</p><p>The results presented in this thesis demonstrate the ability of converted wave analysis as a tool to detect and image geological boundaries that involve sharp contrasts in seismic properties. The results also show that this analysis can provide means of studying aspects of Earth’s structure that are conventionally studied using other types of seismic data.</p>

Although much progress has been made in determining the three dimensional distribution of seismic wave velocities in the Earth, substantially less is known about the three dimensional distribution of intrinsic attenuation. In this study variations in attenuation and shear velocity of the Earth's mantle are constrained using measurements of differential travel time and attenuation. The data are broadband displacement SH seismograms filtered to have energy in the period range 8 to 20 s. Broadband data are used as they should allow a more accurate estimation of body wave attenuation to be made. The seismograms are obtained from over 600 globally distributed earthquakes of magnitude, Mw, 5.5 or greater. Two new methods for determining differential travel times and differential t* values from multiple S phases are presented. The first of these, referred to as the "waveform fitting method" is used to analyse approximately 4300 SS and S waveforms and around 1000 SSS and SS waveforms resulting in differential SS-S and SSS-SS travel times, and corresponding values of differential attenuation represented by t*. The second method, referred to as the "spectral ratio method" is used to analyse approximately 3200 SS and S and around 900 SSS and SS waveforms. The differential travel times and t* values are inverted to obtain models of the lateral variation of shear velocity and lateral variation of q(mu) where q(mu) =1/Q(mu). The models explain the data well but have limited depth resolution. The velocity models show good correlation with previous studies, in particular, low velocities are observed underlying spreading ridges and convergent margins and high velocities are observed under continental regions. The q(mu) model shows shield regions to be less attenuating than PREM, with ridges appearing as highly attenuating features. Models of shear velocity and attenuation are also obtained by combining the body wave dataset of this study with the surface wave datasets of Van Heijst (1997) and Selby (1998).

Oceanic lithosphere constitutes the upper boundary layer of the Earth’s convecting mantle. Its structure and evolution provide a vital window on the dynamics of the mantle and important clues to how the motions of Earth’s surface plates are coupled to convection in the mantle below. The three-dimensional shear-velocity structure of the upper mantle beneath the Atlantic Ocean is investigated to gain insight into processes that drive formation of oceanic lithosphere. Travel times are measured for approximately 10,000 fundamental-mode Rayleigh waves, in the period range 30-130 seconds, traversing the Atlantic basin. Paths with >30% of their length through continental upper mantle are excluded to maximize sensitivity to the oceanic upper mantle. The lateral distribution of Rayleigh wave phase velocity in the Atlantic upper mantle is explored with two approaches. One, phase velocity is allowed to vary only as a function of seafloor age. Two, a general two-dimensional parameterization is utilized in order to capture perturbations to age-dependent structure. Phase velocity shows a strong dependence on seafloor age, and removing age-dependent velocity from the 2-D maps highlights areas of anomalously low velocity, almost all of which are proximal to locations of hotspot volcanism. Depth-dependent variations in vertically-polarized shear velocity (Vsv) are determined with two sets of 3-D models: a layered model that requires constant VSV in each depth layer, and a splined model that allows VSV to vary continuously with depth. At shallow depths (~75 km) the seismic structure shows the expected dependence on seafloor age. At greater depths (~200 km) high-velocity lithosphere is found only beneath the oldest seafloor; velocity variations beneath younger seafloor may result from temperature or compositional variations within the asthenosphere. The age-dependent phase velocities are used to constrain temperature in the mantle and show that, in contrast to previous results for the Pacific, phase velocities for the Atlantic are not consistent with a half-space cooling model but are best explained by a plate-cooling model with thickness of 75 km and mantle temperature of 1400oC. Comparison with data such as basalt chemistry and seafloor elevation helps to separate thermal and compositional effects on shear velocity.

This thesis focuses on calculation of finite and progressive deformation in different tectonic scales using 2D numerical models with application to natural cases. Essentially, two major tectonic areas have been covered: a) salt tectonics and b) upper mantle deformation due to interaction between the lithosphere and asthenosphere. The focus in salt tectonics lies on deformation within down-built diapirs consisting of a source layer feeding a vertical stem. Three deformation regimes have been identified within the salt: (I) a squeezing channel flow underneath the overburden, (II) a corner flow underneath the stem, and (III) a pure channel flow within the stem. The results of the model show that the deformation pattern within the stem of a diapir (e.g. symmetric or asymmetric) can reveal information on different rates of salt supplies from the source layer (e.g. observed in Klodowa-diapir, Poland). Composite rock salt rheology results in strong localization and amplification of the strain along the salt layer boundaries in comparison to Newtonian rock salt. Flow and fold structures of passive marker lines are directly correlated to natural folds within a salt diapir. In case of the upper mantle, focus lies on deformation and resulting lattice preferred orientation (LPO) underneath an oceanic plate. Sensitivity of deformation and seismic anisotropy on rheology, grain size (d), temperature (T), and kinematics (v) has been investigated. The results of the model show that the mechanical lithosphere-asthenosphere boundary is strongly controlled by T and less so by v or d. A higher strain concentration within the asthenosphere (e.g. for smaller potential mantle temperatures, higher plate velocities, or smaller d) indicates a weaker coupling between the plate and the underlying mantle, which becomes stronger with the age of the plate. A Poiseuille flow within the asthenosphere, significantly affects the deformation and LPO in the upper mantle. The results of the model show, that deformation in the upper mantle at a certain distance away from the ridge depends on the absolute velocity in the asthenosphere. However, only in cases of a driving upper mantle base does the seismic anisotropy and delay times reach values within the range of natural data.

Apesar de representarem apenas 1% das rochas expostas na superfície da Terra, as rochas alcalinas possuem um grande espectro composicional e são associadas a vários ambientes tectônicos terrestres. Além disso, seu estudo nos ajuda a compreender melhor a evolução térmica e geodinâmica do planeta. Potássio, sódio e cálcio são os principais elementos utilizados na classificação deste grupo de rochas, e as razões entre estes elementos nos dão pistas sobre sua origem. Este estudo experimental utilizou uma prensa hidráulica de 1000 tonf, com câmaras de perfil toroidal acopladas, para estudar o sistema kalsilita-nefelina-diopsídio-sílica anidro, a 4,0GPa (aprox. 120 km de profundidade) e temperaturas até 1400 °C, com foco nos planos lucita-nefelina-diopsídio (Lct-Nph-Di) e kalsilita-nefelina-diopsídio (Kls-Nph-Di). Os produtos dos experimentos foram analisados através de MEV-EDS e DRX Para o plano Lct-Nph-Di foram encontrados campos de estabilidade para os minerais kalsilita (Klsss), nefelina (Nphss), e clinopiroxênios (Cpxss) (ambos soluções sólidas), wollastonita e sanidina, além de três pontos invariantes: (A) Kls+Nph+Di+liquido (Lc62Ne29Di9) a 1000°C (pseudoeutético); (B) Kls+Sa+(Di+Wo)+liquido (Lct75Nph22Di2) a 1200°C (pseudoperitético), (C) Kls+Di+Wo+liquido (Lct74Nph17Di9) a 1000°C (pseudoeutético). O campo da kalsilita forma uma barreira termal entre os campos composicionais potássico/ultrapotássico e sódico do diagrama. No plano Kls-Nph-Di foram encontrados campos de estabilidade para Klsss, Nphss e Cpxss e duas fases aluminosas (espinélio e coríndon), além do ponto invariante Kls+Nph+Di(±Spl)+liquido (Kls47Nph43Di10) a 1100°C (pseudoeutético). Nossos dados sugerem que a 4,0 GPa a dissolução de K em Nphss é estendida para até 39% molar, e de Na em Klsss para até 27% molar, e que tal solução sólida parcial determina o quão ricos em K ou Na os líquidos gerados pelo sistema serão. Adicionalmente, nossos dados mostram correlaçãos entre concentrações álcalis e sílicas nos líquidos gerados: positiva para K2O, negativa para CaO, mas irregular para Na2O (este devendo depender de outros fatores). / The alkaline rocks represent only 1% of the exposed rocks on the Earth’s surface. Despite that, these rocks are a group with a wide compositional spectrum, associated to many tectonic contexts, and their study can help us to better understand the thermal and geodynamic evolution of our planet. Potassium, sodium and calcium are the main elements used in the classification of this group of rocks, and their concentration can give us insights about the origin of these rocks. We have experimentally investigated the Kalsilite–Nepheline–Diopside-Silica system at high pressure and temperature conditions (HPHT), with emphasis on silica-undersaturated volume (Leucite–Nepheline–Diopside and Kalsilite–Nepheline–Diopside planes), both at 4.0GPa (~120km deep) and dry conditions and temperatures up to 1400°C, aiming to better understand the influence of K2O, Na2O, and CaO in alkaline silica-undersaturated magmas. For the experiments, we used a 1000 tonf hydraulic press with toroidal chambers, and the run products were analyzed by XRD and SEM-EDS technique. In the Lct-Nph-Di plane we determined the stability fields for kalsilite (Klsss), nepheline (Nphss) and clinopyroxene solid solution (CPxss), wollastonite (Wo) and sanidine (Sa); and three piercing points: (A) Kls+Nph+Di+liquid (Lc62Ne29Di9) at 1000°C; (B) Kls+Sa+(Di+Wo)+liquid (Lct75Nph22Di2) at 1200°C, (C) Kls+Di+Wo+liquid (Lct74Nph17Di9) at 1000°C Kalsilite stability field represents a boundary between ultrapotassic/potassic vs. sodic compositions. In the plane Kls-Nph-Di, we found the stability fields for Klsss, Nphss and CPxss and two aluminous phases spinel (Spl) and corundum (Crn). This plane have a piercing point in Kls+Nph+Di(±Spl)+liquid (Kls47Nph43Di10) at 1100°C. Our data suggest that pressure extends K dissolution in Nph (up to 39%) and Na in Kls (up to 27%), and that these solid solutions determine eutectic positions in surrounding systems, or how much K-rich will be the generated magmas. Additionally, our data shows positive correlation between K2O and SiO2 concentration in magmas, negative correlation between CaO and SiO2 and no evident correlation with Na2O.

L'objectif de cette thèse est de construire un modèle d'atténuation sismique du manteau supérieur dela Terre en utilisant un jeu de données original construit par Debayle et Ricard (2012). Ce jeu dedonnées est l'un des plus complet au monde (plus de 375 000 sismogrammes analysés pour extrairel'atténuation et la vitesse de phase du mode fondamental et des cinq premiers harmoniques des ondesde Rayleigh).Les mesures d'atténuation sont tout d'abord traitées pour extraire les effets de l'expansion géométriqueet de la focalisation, minimiser les effets d'erreurs sur la source, écarter les mesures incertaines etregrouper les mesures redondantes. Elles sont ensuite régionalisées pour obtenir des cartes desvariations latérales de l'atténuation des ondes de Rayleigh pour chaque mode et chaque période. Ladernière étape est l'inversion en profondeur des cartes. Elle permet d'obtenir QsADR17, un modèle 3Dde l'atténuation des ondes S dans le manteau supérieur.QsADR17 est corrélé avec la tectonique de surface jusqu'à 200 km de profondeur, avec une faibleatténuation sous les continents et une forte atténuation sous les océans. Des anomalies de forteatténuation sont observées jusqu'à 150~km de profondeur sous les rides océaniques, et persistent à plusgrande profondeur jusque dans la zone de transition sous la plupart des points chauds. La présence delarges anomalies atténuantes situées à 150 km de profondeur sous l'océan Pacifique suggère queplusieurs panaches thermiques viennent s'étaler dans l'asthénosphère. Nous avons également détecté laprésence d'hétérogénéités de composition à la base des cratons et dans un certain nombre de régionsactives. / The aim of this study is to build a 3-D attenuation model of Earth's upper-mantle using a unique datasetbuilt by Debayle & Ricard (2012). This dataset is among the largest in the world: more than 375,000seismograms were analyzed to extract Rayleigh-wave attenuation and velocity measurements for thefondamental mode and the five first harmonics between 40 and 240 s periods.First, attenuation measurements are processed to extract the effects of geometrical attenuation and offocusing and defocusing, in order to minimize the influence of errors on the seismic source, to avoidpotentially incorrect data, and to cluster redondant measurements. Then, measurements are regionalizedto obtain Rayleigh-wave maps for each mode and each period. The last step is the inversion of thesemaps to obtain the depth dependent attenuation. Eventually, we obtain QsADR17, a 3-D model of Swaveattenuation in the upper mantle.QsADR17 is correlated with surface tectonics down to 200 km depth, with low attenuation under thecontinents and high attenuation under the oceans. High-attenuation anomalies are found under oceanicridges down to 150~km depth, and under most of the hotspots at larger depth down to the transitionzone. A large high-attenuation anomaly at 150~km depth under the Pacific ocean suggest that thermalplumes pound into the asthenosphere. We also detect compositional heterogeneities at the base of thecratons and in active areas.

Elastic properties of the major constituent minerals in the Earth’s upper mantle at relevant high pressure-temperature (P-T) conditions are crucial for understanding the composition and seismic velocity structures of the region. In this study, we have measured the single-crystal elasticity of natural Fe-bearing pyrope, Mg2.04Fe0.74Ca0.16Mn0.05Al2Si3O12, using in situ Brillouin spectroscopy and X-ray diffraction at simultaneous high P-T conditions up to 20 GPa and 750 K in an externally-heated diamond anvil cell. The derived aggregate adiabatic bulk and shear modulus (KS0, G0) at ambient conditions are 168.2 (±1.8) GPa and 92.1 (±1.1) GPa, respectively, consistent with literature results. Using the third-order Eulerian finite-strain equation to fit the high P-T data, the derived pressure derivative of the bulk and shear moduli at constant temperature are (∂KS/∂P)T=4.4 (±0.1) and (∂G/∂P)T=1.2 (±0.1), respectively. Applying these pressure derivatives, the temperature derivative of these moduli at constant pressure are also calculated, yielding (∂KS/∂T)P=-18.5(±1.3) MPa/K and (∂G/∂T)P=-5.2(±1.1) MPa/K, respectively. Compared to literature values, our results show that addition of 25% Fe in pyrope increases the pressure derivative of the bulk modulus by 7%, but has a negligible effect on other elastic parameters. Extrapolation of our results shows that Fe-bearing pyrope remains almost elastically isotropic at relevant P-T conditions of the upper mantle, indicating that it may not have a significant contribution to seismic Vp and Vs anisotropy in the upper mantle. Together with the elasticity of olivine and pyroxene minerals in the upper mantle, we have constructed new velocity profiles for two representative compositional models, pyrolite and piclogite, along Earth’s upper mantle geotherm. These velocity models show Vs profiles consistent with seismic observations, although Vp profiles are slightly lower than in seismic models. / text

Apesar de representarem apenas 1% das rochas expostas na superfície da Terra, as rochas alcalinas possuem um grande espectro composicional e são associadas a vários ambientes tectônicos terrestres. Além disso, seu estudo nos ajuda a compreender melhor a evolução térmica e geodinâmica do planeta. Potássio, sódio e cálcio são os principais elementos utilizados na classificação deste grupo de rochas, e as razões entre estes elementos nos dão pistas sobre sua origem. Este estudo experimental utilizou uma prensa hidráulica de 1000 tonf, com câmaras de perfil toroidal acopladas, para estudar o sistema kalsilita-nefelina-diopsídio-sílica anidro, a 4,0GPa (aprox. 120 km de profundidade) e temperaturas até 1400 °C, com foco nos planos lucita-nefelina-diopsídio (Lct-Nph-Di) e kalsilita-nefelina-diopsídio (Kls-Nph-Di). Os produtos dos experimentos foram analisados através de MEV-EDS e DRX Para o plano Lct-Nph-Di foram encontrados campos de estabilidade para os minerais kalsilita (Klsss), nefelina (Nphss), e clinopiroxênios (Cpxss) (ambos soluções sólidas), wollastonita e sanidina, além de três pontos invariantes: (A) Kls+Nph+Di+liquido (Lc62Ne29Di9) a 1000°C (pseudoeutético); (B) Kls+Sa+(Di+Wo)+liquido (Lct75Nph22Di2) a 1200°C (pseudoperitético), (C) Kls+Di+Wo+liquido (Lct74Nph17Di9) a 1000°C (pseudoeutético). O campo da kalsilita forma uma barreira termal entre os campos composicionais potássico/ultrapotássico e sódico do diagrama. No plano Kls-Nph-Di foram encontrados campos de estabilidade para Klsss, Nphss e Cpxss e duas fases aluminosas (espinélio e coríndon), além do ponto invariante Kls+Nph+Di(±Spl)+liquido (Kls47Nph43Di10) a 1100°C (pseudoeutético). Nossos dados sugerem que a 4,0 GPa a dissolução de K em Nphss é estendida para até 39% molar, e de Na em Klsss para até 27% molar, e que tal solução sólida parcial determina o quão ricos em K ou Na os líquidos gerados pelo sistema serão. Adicionalmente, nossos dados mostram correlaçãos entre concentrações álcalis e sílicas nos líquidos gerados: positiva para K2O, negativa para CaO, mas irregular para Na2O (este devendo depender de outros fatores). / The alkaline rocks represent only 1% of the exposed rocks on the Earth’s surface. Despite that, these rocks are a group with a wide compositional spectrum, associated to many tectonic contexts, and their study can help us to better understand the thermal and geodynamic evolution of our planet. Potassium, sodium and calcium are the main elements used in the classification of this group of rocks, and their concentration can give us insights about the origin of these rocks. We have experimentally investigated the Kalsilite–Nepheline–Diopside-Silica system at high pressure and temperature conditions (HPHT), with emphasis on silica-undersaturated volume (Leucite–Nepheline–Diopside and Kalsilite–Nepheline–Diopside planes), both at 4.0GPa (~120km deep) and dry conditions and temperatures up to 1400°C, aiming to better understand the influence of K2O, Na2O, and CaO in alkaline silica-undersaturated magmas. For the experiments, we used a 1000 tonf hydraulic press with toroidal chambers, and the run products were analyzed by XRD and SEM-EDS technique. In the Lct-Nph-Di plane we determined the stability fields for kalsilite (Klsss), nepheline (Nphss) and clinopyroxene solid solution (CPxss), wollastonite (Wo) and sanidine (Sa); and three piercing points: (A) Kls+Nph+Di+liquid (Lc62Ne29Di9) at 1000°C; (B) Kls+Sa+(Di+Wo)+liquid (Lct75Nph22Di2) at 1200°C, (C) Kls+Di+Wo+liquid (Lct74Nph17Di9) at 1000°C Kalsilite stability field represents a boundary between ultrapotassic/potassic vs. sodic compositions. In the plane Kls-Nph-Di, we found the stability fields for Klsss, Nphss and CPxss and two aluminous phases spinel (Spl) and corundum (Crn). This plane have a piercing point in Kls+Nph+Di(±Spl)+liquid (Kls47Nph43Di10) at 1100°C. Our data suggest that pressure extends K dissolution in Nph (up to 39%) and Na in Kls (up to 27%), and that these solid solutions determine eutectic positions in surrounding systems, or how much K-rich will be the generated magmas. Additionally, our data shows positive correlation between K2O and SiO2 concentration in magmas, negative correlation between CaO and SiO2 and no evident correlation with Na2O.