Development of carbon membranes for ammonia separation

Membrane separation technologies have gained a lot of interest as alternatives to the traditional energy intensive separation methods since their high performance in separation and low energy consumption. Specifically, Carbon Molecular Sieves Membranes (CMSMs) have proven to be suitable for application in gas separation processes due to their peculiar sieving effects. Carbon membranes are porous inorganic membranes generally obtained from the pyrolysis of an organic precursor. Depending on the pyrolysis parameters, the carbonization process removes the majority of the heteroatoms originally present in the polymer macromolecules, thus achieving a crosslinked matrix and a hard carbon structure with almost pure carbon material. This study focuses on the development of carbon membranes to selectively separate NH3 from the reaction mixture and shift the reaction equilibrium towards the products side of the reaction. The aim is to use this membrane in a packed bed membrane reactor for ammonia synthesis. In this thesis, the use of four different cross-linking agents is investigated in order to understand if their presence enhance the membrane performances, in particular enhancing surface diffusion. The cross linker agents studied are HMTA, ethylenediamine, diethylenetriamine and tris(2-aminoethyl)amine. Different dipping solutions were prepared starting from the synthesis of Novolac resin (which is used as polymeric precursor), adding NMP as solvent, oxalic acid, formaldehyde and the different cross-linking agents. At this point, the porous alumina supports were dip-coated in these coating solutions, dried and carbonized in order to produce supported carbon membranes. Then, the membranes were tested in the MemAmmonia permeation setup in single gas and mix-feed permeation tests (the latter in order to simulate the real application of separation of ammonia from the reaction mixture) in order to evaluate the effect of the different amines in terms of permeability, as well as NH3/N2 and NH3/H2 selectivities. The experimental results obtained with HMTA suggest that this cross-linker is not feasible for our purpose in the fabrication of carbon membranes. In fact, the produced membranes present large defects and cracks and also the dipping solutions with an alternative solvent (methanol/acetone solution) are not stable over time, resulting in a wide variation of physical characteristics such as viscosity, color and degree of cross-linking. Among the other amines, from the comparison of the available permeation tests results, the membranes that present the best performances in the selective separation of ammonia are those in which diethylenetriamine is used as cross-linking agent, in fact they are the ones that present the best trade-off between the improvements of ideal and real selectivities. However, in order to fully evaluate the effect of the cross-linking with diethylenetriamine in the performance of the final membrane, additional analysis such as TGA and elemental analysis are required.