Hello Diego,
I am glad to hear that you are interested to use MRST for your thesis!
There are essentially two options for miscible CO2 simulation in MRST.
The first is to use the solvent module, which has a four-phase (water, oil, gas, solvent) model where the hydrocarbon "phases" can mix to a variable degree. This is an extension of the blackoil-type description and may be appropriate if you already have your model as some kind of blackoil model.
The second option is the compositional module, which models general n-component mixtures with an equation-of-state or K-values. This is the most general model, but can be quite expensive since each individual component must be tracked individually. It is fairly simple to set up, however, since you only have to specify a few properties for each molecular species.
Cheers,
Olav