Development of membrane materials for a membrane contactor
MetadataVis full innførsel
International Energy Agency reports a continued increase in global use of natural gas in World Energy Outlook 2016. A large part of the total gas reserves has high amounts of one or both of the acid gases carbon dioxide (CO2) and hydrogen sulfide (H2S). This makes continued research on improved methods for acid gas removal from natural gas highly relevant. The studies presented in this thesis aimed to develop a membrane material for natural gas sweetening with a membrane contactor. The main focus in the work has been membrane development and characterization of the prepared membranes. The target for the developed membranes was high, stable permeability. This has previously been obtained with crosslinked membranes of poly(1-trimethylsilyl-1-propyne) (PTMSP) with added non-porous nanoparticles. Crosslinking has shown to stabilize the permeability of PTMSP, a polymer that show decrease in permeability over time because of relaxation of the free volume present in the membrane matrix. Crosslinking lead to lower permeability, but it has been shown that the permeability of PTMSP membranes can be increased upon addition of non-porous nanoparticles. Uncrosslinked and crosslinked PTMSP membranes with or without different types of TiO2 nanoparticles have been investigated in terms of single gas permeability, contact angles and surface morphology. In the first investigation, the effect of crosslinking and addition of 15 nm TiO2 single nanoparticles on the transport properties of PTMSP membranes was explored. Crosslinking of the PTMSP membranes with 3 wt% 4,4’-diazidobenzophenone lead to decreased permeability and increased selectivity, as expected. The addition of 15 nm TiO2 nanoparticles unexpectedly resulted in decreased permeability. In the next study, Aeroxide® TiO2 T 805 particles were applied in corresponding experiments. These particles have previously shown to give the wanted increased permeability in PTMSP membranes. It was found that this material is prepared to give highly branched aggregates. Addition of 20 wt% Aeroxide® TiO2 T 805 particles lead to an increase in membrane permeability, in contrast to the decrease seen when adding 15 nm TiO2 single particles. Aggregated versions of the 15 nm TiO2 particles in the size ranges 15-400 nm and >1 µm were obtained to be included in the investigations of how particle morphology effect the transport properties in PTMSP membranes. SEM images revealed different characteristics of the different particles in the membranes. Based on the permeability results it was hypothesized that aggregated particles are needed to show increased permeability in PTMSP membranes. Findings in the literature support this hypothesis. More knowledge about the specific morphology of the different TiO2 particles is needed to draw any definite conclusion on how this affect the gas transport properties of PTMSP membranes. The studies performed in this work have shown how difficult it is to prepare PTMSP nanocomposite membranes with predictable performance. It is the author’s opinion that PTMSP should be synthesized with full control of the resulting microstructure (cis/trans content) to be able to get predictable results. Only this way, can the polymer be optimized for the wanted application. The combination of full control the microstructure of PTMSP and the morphology of added particles could also be the only way the true transport mechanism for PTMSP nanocomposite membranes could be revealed. Wetting of the membranes is the major challenge for the application of membrane contactors for CO2 removal. The effect of exposure to water and methyldiethanolamine (MDEA) on uncrosslinked and crosslinked PTMSP membranes with or without different types of TiO2 nanoparticles have therefore been investigated. MDEA is the most used absorption liquid in natural gas sweetening. The membranes have been characterized with regards to single gas permeability, contact angles and surface morphology for periods up to ten weeks. Pure PTMSP membranes saw a decrease in permeability upon exposure, but not larger than seen for an unexposed membrane as a result of aging. Crosslinked membranes showed an increase in permeability upon exposure, which is suggested to be caused by swelling. No significant differences between the different liquid types were seen. PTMSP membranes with particles showed a large decrease in permeability upon exposure to MDEA. The decrease was significantly larger than seen when exposed to water. There was also a significant difference in the effect of MDEA exposure between pure PTMSP membranes and PTMSP nanocomposite membranes. It is speculated if MDEA was sorbed in the particles, blocking gas transport, but the true mechanism behind the seen results is not known. Exposure to water gave the largest decrease in water contact angles, and the decrease was larger for membranes with particles compared to pure PTMSP membranes. Investigations of the membrane surfaces by SEM clearly showed morphology changes for the membrane surfaces upon liquid exposure. The exposure to water or MDEA lead to different characteristics in the morphology. The mechanisms behind the seen results upon exposure of PTMSP membranes to water and MDEA are not known, and it is not known if and how the changes in the different properties explored are related. As the true transport mechanism for the unexposed PTMSP nanocomposite membranes in this work is not known, it is difficult to explain the changes seen upon exposure. Interesting trends have been seen for both unexposed and exposed PTMSP membranes with or without crosslinking and TiO2 particles. The findings should be investigated further to reveal the mechanisms taking place.