Forming magnesium carbonate minerals through carbonation of magnesium silicates has been proposed as a safe and durable way to store carbon dioxide, with a possibly high potential to offset anthropogenic CO2 emissions. To date however, chemical reactions involved in this process are facing strong kinetic limitations, which originate in the low reactivity of both Mg-silicates and Mg-carbonates. Numerous studies have focused on the dissolution of Mg-silicates, under the questionable hypothesis that this step limits the whole process. This thesis work focuses instead on the mechanisms and rates of formation of magnesium carbonates, which are the final products of carbonation reactions. The first part of the work is dedicated to studying the influence on magnesite precipitation kinetics of three organic ligands known to accelerate Mg-silicates dissolution rates : oxalate, citrate and EDTA. With help of mixed-flow reactor experiments performed between 100 and 150°C, we show that these ligands significantly reduce magnesite growth rates, through two combined mechanisms: (1) complexation of Mg2+ cations in aqueous solution, which was rigorously estimated from a thermodynamic database established through a critical review of the literature, and (2) adsorption of ligands to a limited number of surface sites, leading to a decrease of the precipitation rate constant. The observed growth inhibition is maximal with citrate. We then used hydrothermal atomic force microscopy to probe the origin of the documented growth inhibition. Our observations show that citrate and oxalate interact with the crystal growth process on magnesite surface, modifying the shape of growth hillocks as well as the step generation frequency through spiral growth. We also show that the ligands adsorb preferentially on different kink-sites, which is probably related to their different structures and chemical properties. We propose that the stronger magnesite growth inhibition caused by citrate is related to a preferential interaction of the ligand with acute steps on the magnesite surface, which limit the spiral growth process through their low advancement rate. The description of these processes with an empirical rate law allows performing simple numerical simulations of forsterite carbonatation at 120°C in the presence of the ligands. We thus demonstrate that the use of the investigated ligands would clearly be detrimental to the carbonation of forsterite. The third and last part of this work deals with hydromagnesite solubility and growth kinetics between 25 and 75°C. The obtained results show that hydromagnesite growth rates largely exceed magnesite rates at comparable chemical affinity, while the activation energy of the process is much smaller than for magnesite. This original kinetic dataset thus confirms the long-standing hypothesis that Mg2+ dehydration is the rate-limiting step for Mg-carbonate precipitation from aqueous solution. However, due to its higher solubility, hydromagnesite may grow more quickly than magnesite only at low temperature and alkaline pH. Thus, it may not provide a solution to the sluggish precipitation kinetics of magnesite during Mg-silicates carbonation
