As intrusões Indaiá I e II (também referidas como Perdizes 3-A e 3-B) são corpos intrusivos hipoabissais ultrabásicos de pequeno porte, associados ao magmatismo cretácico da Província Alcalina Alto Paranaíba, oeste do Estado de Minas Gerais. Esta dissertação de mestrado objetivou compreender a gênese e processos de evolução relacionados a intrusão Indaiá II, cuja classificação ambígua entre kimberlito e kamafugito dividiu autores em trabalhos anteriores.. Também buscou-se discutir a possível vinculação genética com a intrusão vizinha, Indaiá I. Para tanto, foram realizadas novas análises petrográficas, geoquímicas e isotópicas de rocha total, e química mineral. A intrusão principal (Indaiá I) é classificada como um kimberlito do Grupo I. Já a intrusão satélite (Indaiá II) distingue-se quimicamente e petrograficamente, possuindo uma composição ultrapotássica, similar à kamafugítica, e apresenta menor concentração de macrocristais de olivina, diopsídio como o principal constituinte da matriz ( e não a monticellita como Indaiá I) e possui uma presença abundante de microenclaves félsicos. Estes enclaves apresentam estrutura estirada, e são constituídos principalmente por kalsilita/ nefelina, vidro devitrificado, diopsídio e flogopita. A presença de texturas sugestivas de desequilíbrio físico-químico como embayment e sieve em grãos de olivina e clinopiroxênio em Indaiá II são indicativas de processos de sistema aberto. A abundância de diopsídio na matriz e a substituição de olivina por clinopiroxênio nas bordas de macrocristais e microcristais apontam para um aumento da atividade de sílica no decorrer do processo de cristalização. A alta proporção de xenólitos crustais na rocha, totalmente transformados ou com evidências de fusão parcial, e a presença de apatitas aciculares nas suas cercanias, indicam a operação de um processo de contaminação crustal. Novos dados de química mineral e litogeoquímica de elementos maiores e traços corroboram esse processo. Os trends composicionais de minerais como espinélio e flogopita são bastante similares a ocorrências de kimberlitos contaminados registrados na literatura. Além disso, amostras de Indaiá II apresentam elevado C.I. (Contamination Index, 2,12-2,25), maior percentual de SiO2, K2O e maior razão Rb/Sr que a intrusão principal. Modelos de contaminação crustal foram efetuados a partir da mistura do fundido de Indaiá I e do fundido das rochas granitoides encaixantes. Os fundidos de Indaiá I e II foram obtidos pela extração do volume dos núcleos de macrocristais de olivina, considerados xenocristais em ambas as ocorrências, das análises de rocha total. A quantidade de olivina foi obtida através de análises modais, e a composição considerada foi obtida através de uma média das análises pontuais de microssonda eletrônica e LA-ICP-MS. O fundido da encaixante foi estimado utilizando-se as análises de rocha-total das encaixantes crustais da área e calculando-se um fundido parcial desta rocha a 750ºC e pressões entre 1-5Kbar com o software RhyoliteMelts (para elementos maiores). A composição de elementos-traço do fundido foi modelada por balanço de massa, a partir de coeficientes de partição compilados da literatura e a proporção de minerais no resíduo sólido obtida dos modelos por elementos maiores. Curvas de mistura entre elementos maiores e traços também parecem confirmar o processo. Novos dados isotópicos de 87Sr/86Sr e 143Nd/144Nd foram obtidos para estas intrusões, como também para a encaixante local. Curvas de mixing isotópico foram feitas para tentar estabelecer a quantidade de contribuição crustal nos magmas de Indaiá II. Nestes modelos, as amostras de Indaiá II ajustam-se concordantemente às curvas de mistura entre os polos de Indaiá I e da encaixante. Concluimos que: 1) a intrusão Indaiá II representaria uma intrusão kimberlítica altamente contaminada, 2) a contaminação provavelmente ocorreu pela assimilação de fundidos anatéticos oriundos das principais rochas encaixantes crustais da área, 3) que Indaiá I e Indaiá II poderiam ter um mesmo magma progenitor, mas com diferentes graus de contaminação crustal. / The intrusions Indaiá I and II (also referred as Perdizes 3-A and 3-B) are hypoabissal ultrabasic bodies of small size associated with the Cretaceous magmatism of the Alto Paranaiba Alkaline Province, west of the Minas Gerais state. This master\'s dissertation aimed a better understanding of the genesis and the evolution processes related to the Indaiá II intrusion, whose ambiguous classification between kimberlite and kamafugite divided the authors of former works. Also, it is discussed the possible genetic linkage of Indaiá II with the neighbor intrusion, Indaiá I. For this purpose, new petrographic, whole rock geochemical, isotopic and mineral chemistry analyses were made. The main intrusion (Indaiá I) was classified as a group-I kimberlite. The satellite intrusion, Indaiá II, is petrographically and chemically distinct of the main one, having a ultrapotassic composition, similar to those of kamafugites, and presenting lower volumes of olivine macrocrysts, diopside as the main matrix phase (instead of monticellite as found in Indaiá I), and an abundant presence of felsic microenclaves. These enclaves present elongated structure and they mainly have kalsilite, devitrified glass, diopside and phlogopite. The presence of textures indicative of physical-chemical disequilibrium, such as embayment and sieve in olivine and clinopyroxene grains from Indaiá II are indicative of an open system process. The high amounts of diopside in the matrix and the substitution of the rims of olivine macro- and microcrysts by clinopyroxene point to an increase in the silica activity during the crystallization. The high proportion of crustal xenoliths in the rocks, most of these totally transformed or with evidences with partial melt, and the presence of acicular apatite in their border regions, indicate the action of crustal contamination processes. New mineral chemistry and whole rock geochemical (major and trace elements) data corroborate this process. The spinel and phlogopite compositional trends are very similar with those from contaminated kimberlitic occurrences registered in the literature. Moreover, samples from Indaiá II have high Contamination Index (C.I., 2.12-2.25), greater amounts of SiO2, K2O and higher Rb/Sr ratios than those from the main intrusion. Crustal contamination models were developed considering a mixing between the melt of Indaiá I and the partial melt of the granitoid host rocks of the area. The composition of melts from Indaiá I and II were calculated by the extraction of the amount of the cores of olivine macrocrysts, considered xenocrysts in both occurrences, from the whole-rock compositions. The amount of olivines was taken from modal concentrations and the considered composition was obtained by the average of electron microprobe and laser ablation inductively coupled plasma mass spectrometry (LA ICP-MS) analyses. The major element composition of the partial melt of the host rock was estimated using the algorithm RhyoliteMelts. For these models, we considered the whole-rock compositions of the main host rocks of the area and used 750ºC and 1-5Kbar. The trace element composition of the host rock partial melts was modeled by mass balance calculations, using partition coefficients from the literature and the proportion of minerals in the solid residue obtained for the major element models. After the calculation of these liquid compositions, mixing curve models using major and trace element compositions were made and they are concordant with the crustal contamination processes. New 87Sr/86Sr and 143Nd/144Nd isotopic data were obtained for Indaiá I and II intrusions, as well as for the local host rock. Isotopic mixing curves were also made trying to constrain the amount of crustal contribution in Indaiá II magmas. In these models, Indaiá II samples fit concordantly with the mixing curves having Indaiá I and the host rock ratios as the main poles. We conclude that 1) Indaiá II is representative of a highly contaminated kimberlitic intrusion, 2) the contamination possibly occurred by the assimilation of the anatetic melts from the main crustal host rocks of this area and 3) that Indaiá I and Indaiá II could have the same parent melt, but with different degrees of crustal contamination.

Des liquides riches en K2O ( K2O > 2 pds. % ; K2O/Na2O > 1) sont observés dans la majeure partie des contextes géodynamiques sur Terre. Ces liquides sont principalement caractérisés par leurs teneurs en K2O variant entre 3 et 13 pds. % et des rapports K2O/Na2O de 1 – 40. Les compositions chimiques des différents groupes de liquides riches en K2O observés sont extrêmement variables, depuis des termes très sous-saturés en silice (kamafugites, kimberlites, lamproïtes madupitiques à olivine) à des termes sur-saturés en silice (shoshonites, lamproïtes à phlogopite). Ces fortes teneurs en K2O et les rapports K2O/ Na2O élevés ne peuvent pas être obtenus par la fusion de péridotites fertiles ou réfractaires. Des expériences de fusion partielle en piston-cylindre ont été réalisées sur des péridotites à phlogopite ± amphibole dans les domaines de stabilité du spinelle et du grenat (1 et 3 GPa) afin de déterminer la capacité du manteau lithosphérique à produire des liquides riches en K2O. La présence de faibles teneurs en fluor dans le matériel de départ stabilise le phlogopite à des températures supérieures aux études antérieures. Les faibles degrés de fusion obtenus à 1 GPa sont sur-saturés en silice et leur teneur en K2O est tamponnée à 4 – 6 pds. % par la présence de phlogopite résiduel pour des péridotites fertile et réfractaire, respectivement. Les expériences réalisées à 3 GPa montrent que les premiers degrés de fusion sont sous-saturés en silice mais plus riches en K2O (6 – 8 pds. % pour la lherzolite et la harzburgite, respectivement) que dans le domaine du spinelle, démontrant l’importance de la pression sur la genèse de liquides riches en K2O. Les modélisations réalisées montrent également que l’augmentation de la proportion de phlogopite dans la source ne modifie pas la teneur en K2O des liquides formés mais diminue leurs rapports K2O/Na2O. Par conséquent, la fusion de péridotite à phlogopite dans la gamme de pression étudiée ne permet pas d’obtenir des liquides aussi riches en K2O que certains lamproïtes et kamafugites. Une série d’expériences réalisée sur du phlogopite pur à 1 et 3 GPa montrent que les liquides dérivés de telles sources sont très riches en K2O (12 – 14 pds. %) et comparables aux lamproïtes. Les différentes lithologies dans le manteau ne permettent cependant pas d’expliquer la grande gamme de composition des liquides riches en K2O et nécessitent des conditions riches en éléments volatils (H2O, CO2, F) et des fugacités d’oxygène réductrices. / K2O-rich melts (K2O > 2 wt. %; K2O/Na2O > 1) have been described in all of the major geodynamic settings on Earth. These melts are mainly characterized by their huge K2O content, ranging between 2 – 13 wt. % and K2O/Na 2 O ratios of 1 – 40. The chemical compositions of the different K2O-rich melt groups span a very high variability, from strongly silica undersaturated melts (kamfugites, kimberlites, madupitic lamproites) to silica-rich terms (shoshonites, phlogopite lamproites). These very high K2O contents together with strong K2O/Na2O ratios cannot be derived from partial melting of fertile or depleted peridotites. Partial melting experiments have been conducted in piston cylinder apparatus on phlogopite ± amphibole-peridotite in the spinel and garnet stability fields (1 – 3 GPa) in order to determine the ability of the lithospheric mantle to produce K2O-rich melts. The presence of small amounts of fluorine in the starting material leads to stabilize phlogopite at higher temperatures than previously determined. The first degrees of melting at 1 GPa are silica-rich and their K2O contents are buffered to 4 – 6 wt. % in the presence of residual phlogopite, depending on the source fertility (lherzolite and harzburgite, respectively). In the garnet stability field at 3 GPa, low-degree melts are silica-undersaturated but are enriched in K2O, compared to the garnet stability field: from 6 to 8 wt. % in lherzolite and harzburgite sources, respectively. These results suggest that pressure is a key parameter in the mantle to produce K2O-rich melts. Partition coefficient modelings show that increasing the phlogopite proportion in the mantle source does not modify the K2O content of derived melts, but decreases their K2O/Na2O ratios. Consequently, partial melting of phlogopite-peridotite in this range of pressure cannot accounts for the highest K2O contents observed in natural lamproites and kamafugites. A series of experiments has been realized on pure phlogopite at 1 and 3 GPa, showing that derived melts are strongly enriched in K2O (12 – 14 wt. %) and share chemical affinities with lamproites. Peridotite or pyroxenite melting in the presence of phlogopite, however, do not permit to reproduce the high chemical variability of natural K2O-rich melts requires volatile-rich conditions (H2O, CO2, F) and reduced oxygen fugacities.

No description available.