[pt] Neste trabalho são apresentados os resultados da caracterização mineralógica de uma amostra de minério de ouro da RPM – Kinross, que teve como objetivo principal, identificar os minerais portadores de titânio e avaliar o espectro de liberação dos minerais de interesse. A metodologia empregada compreendeu a realização de análises granulométricas, separações em meio denso e magnética. Os estudos mineralógicos foram realizados por meio da difratometria de raios X (DRX), microscopia ótica e microscopia eletrônica de varredura (MEV) por meio do sistema Mineral Liberation Analyzer – MLA. Os produtos obtidos foram submetidos à análise química por espectrometria de fluorescência de raios X (FRX). As análises químicas indicaram que a amostra estava constituída essencialmente por SiO2 (66,4%), Al2O3 (14,2%), Fe2O3 (7,22%) e TiO2 (1%). Visando avaliar o espectro de liberação dos minerais portadores de titânio, o estudo foi focado em seis faixas granulométricas (-300+212; -212+150; - 150+104; -104+74; -74+53; e -53+37um). Cerca de 20% do material de todas as frações foi constituída por material afundado (meio denso), sendo composto principalmente de SiO2 (35%), Fe2O3 (30%), Al2O3 (>7%) e TiO2 (<5%). A fração flutuada é composta em sua maioria de SiO2 e Al2O3. As análises de DRX da fração afundada indicaram a presença de ilmenita, anatásio e rutilo. As frações afundadas foram submetidas à separação magnética no separador Frantz em diferentes intensidades de corrente (0,3 até 1,75A), através desta operação se concentrou até 8% em massa de TiO2 na fração -104+74um e a 0,5A. As frações afundadas foram submetidas a estudos sistemáticos no MEV com o sistema MLA, confirmando a presença de ilmenita, anatásio e rutilo como os minerais portadores de titânio. A liberação completa dos minerais carreadores de titânio foi aproximadamente de 1% em massa, já a ganga liberou-se mais de 90% em massa. A partir dos resultados obtidos se observou que é possível concentrar o TiO2 contido no minério. / [en] This work presents studies in Minerals Characterization of gold ore sample from RPM-Kinross with objective to identify their titanium-bearing minerals and to assess its behavior in different size fraction (spectrum release). The methodology involved particle size analysis and minerals separation (separation in dense medium and magnetic separation), followed of mineralogical studies by XRay Diffraction, Optical Microscopy and Scanning Electron Microscopy (SEM) by using the Mineral Liberation Analyzer-MLA. The products obtained were submitted to Chemical Analysis of X-Ray Fluorescence. The analysis of X-Ray Fluorescence revealed that the sample studied was formed essentially by SiO2 (66,4%), Al2O3 (14,2%), Fe2O3 (7,22%) and TiO2 (1%). The studies were focused in six different sizes (-300+212; -212+150; -150+104; -104+74; -74+53 and - 53+37um). ). About 20% of the material from all fractions material is sunk (dense medium), composed mainly of SiO2 (35%), Fe2O3 (30%), Al2O3 (> 7%) and TiO2 (<5%). The floated fraction was composed mainly of SiO2 and Al2O3. The XRD of the sunken fraction indicated the presence of ilmenite, rutile and anatase. The sunken fractions were subjected to magnetic separation in the Frantz separator at different current intensities (0.3 to 1.75A), this operation was concentrated up to 8 wt% TiO2 in the fraction -104 +74um and 0.5A. The sunken fractions were subjected to systematic studies in the SEM system with MLA, confirming the presence of ilmenite, rutile and anatase as the titanium-bearing minerals. The gangue has been release of the mineral carrier of titanium was approximately 1% wt%, the denim has released more than 90% wt%. From the results it was observed that it is possible to concentrate the TiO2 contained in the ore.

In recent years, automated mineralogy has become an essential tool in geometallurgy. Automated mineralogical tools allow the acquisition of mineralogical and liberation data of ore particles in a sample. These particle data can then be used further for particle-based mineral processing simulation in the context of geometallurgy. However, most automated mineralogical tools currently in application are based on two-dimensional (2D) microscopy analysis, which are subject to stereological error when analyzing three-dimensional(3D) object such as ore particles. Recent advancements in X-ray microcomputed tomography (µCT) have indicated great potential of such system to be the next automated mineralogical tool. µCT's main advantage lies on its ability in monitoring 3D internal structure of the ore at resolutions down to few microns, eliminating stereological error obtained from 2D analysis. Aided with the continuous developments of computing capability of 3D data, it is only the question of time that µCT system becomes an interesting alternative in automated mineralogy system. This study aims to evaluate the potential of implementing µCT as an automated mineralogical tool in the context of geometallurgy. First, a brief introduction about the role of automated mineralogy in geometallurgy is presented. Then, the development of µCT system to become an automated mineralogical tool in the context of geometallurgy andprocess mineralogy is discussed (Paper 1). The discussion also reviews the available data analysis methods in extracting ore properties (size, mineralogy, texture) from the 3D µCT image (Paper 2). Based on the review, it was found that the main challenge inperforming µCT analysis of ore samples is the difficulties associated to the segmentation of the mineral phases in the dataset. This challenge is adressed through the implementation of machine learning techniques using Scanning Electron Microscope (SEM) data as a reference to differentiate the mineral phases in the µCT dataset (Paper 3).

[pt] A integração do controle por computador de microscópios com a aquisição e análise digital de imagens levou à criação de uma nova área, denominada Microscopia Digital. Além de permitir um certo grau de automação, a Microscopia Digital abriu possibilidades realmente novas para a caracterização microestrutural. Uma destas novas e promissoras possibilidades é a Microscopia Co-localizada, que junta diversos tipos de informação, obtidas a partir de diferentes técnicas de microscopia. No presente trabalho foi desenvolvida e implementada uma metodologia de Microscopia Co-localizada que combina imagens de Microscopia Óptica de Luz Refletida (MO) e de Microscopia Eletrônica de Varredura (MEV). Esta metodologia envolve desde a aquisição das imagens nos microscópios até a análise das fases presentes através de técnicas de Reconhecimento de Padrões. Um procedimento automático de registro entre os dois tipos de imagens foi desenvolvido, permitindo o ajuste de magnificação, translação, rotação, tamanho de pixel e distorções locais. Desta forma, imagens de MO e de MEV de uma dada amostra podem ser combinadas precisamente. A metodologia foi testada com diversas amostras minerais, visando a discriminação de fases que são indistinguíveis por MO ou MEV. A Microscopia Co-localizada MOMEV foi empregada em uma rotina para a caracterização de amostras de minério de ferro e os resultados obtidos foram comparados com os da análise tradicional ao MEV. / [en] Computer-controlled microscopes with digital image acquisition and analysis led to the creation of a new field, called Digital Microscopy. Digital Microscopy not only allows a certain degree of automation but also has brought new possibilities to microstructural characterization. One of these new and promising possibilities is Co- Site Microscopy, that links different kinds of information, obtained from different microscopy techniques. The present work presents the development and implementation of a Co-Site Microscopy methodology that combines images obtained by Reflected Light Microscopy (RLM) and Scanning Electron Microscopy (SEM). This methodology involves the whole sequence, from image acquisition at the microscopes to the analysis of the phases using Pattern Recognition techniques. An automatic registration procedure for the two kinds of images was developed, allowing the adjustment of magnification, translation, rotation, and pixel size, and the correction of local distortions. The methodology was tested with several mineral samples, aiming at the discrimination of phases that are not distinguishable with either RLM or SEM. The RLM-SEM Co-Site Microscopy technique was employed in the characterization of iron ore samples and the obtained results were compared to the traditional analysis by SEM.