Urgent actions are needed to reduce CO2 emissions and mitigate the increasingly severe impacts of climate change. Since the 1990s, the membrane research group (MEMFO) at the Norwegian University of Science and Technology has been committed to developing effective membranes and membrane processes for CO2 separation. MEMFO's research can be categorized into five main themes: facilitated transport membranes, hybrid membranes, carbon membranes, membrane contactors, and related modeling and process simulation. These themes are tied to industrial applications in CO2 capture from flue gas, biogas upgrading, natural gas sweetening, and hydrogen purification. Promising membranes, identified based on their laboratory-scale performances, have been selected for onsite testing in industrial processes to validate their performances as well as stability and durability. Verified membranes are upscaled for pilot tests. This account paper summarizes MEMFO's research and development outcomes over the past decade and discusses our research strategies and perspectives for future work.
Hollow fiber membranes show a great potential in natural gas sweetening by removing CO2 to meet gas grid specifications. A membrane model with high prediction accuracy is developed to model multicomponent gas transport through hollow fiber modules. The influences of hollow fiber diameter and length, and packing density on module efficiency related to pressure drops in both sides were systematically investigated based on the developed model. The total pressure drop along the length is less than 1% if the inner diameter of hollow fibers (0.6 m length) is larger than 200 μm. Moreover, the highest module packing density was found to be dependent on hollow fiber dimension, and too high packing density will cause extreme high pressure drops. Both feed CO2 concentration and pressure were found to significantly influence membrane module performance related to the required specific membrane area and hydrocarbon loss based on process parametric study of CO2 removal from natural gas. Larger pressure drops along fiber length was found for the more-permeable polyimide membranes compared to the less-permeable cellulose acetate and carbon membranes. The developed model can be used for guiding the design of efficient hollow fiber membrane modules and potentially process simulation of membrane gas separation.
Industrial scale production of carbon membrane is very challenging due to expensive precursor materials and a multi-step process with several variables to deal with. The optimization of these variables is essential to gain a competent carbon membrane (CM) with high performance and good mechanical properties. In this paper, a pilot scale system is reported that was developed to produce CM from regenerated cellulose precursor with the annual production capacity 700 m² of CM. The process was optimized to achieve maximum yield (>95%) of high quality precursor fibers and carbonized fibers. A dope solution of cellulose acetate (CA)/Polyvinylpyrrolidone (PVP)/N-methyl-2-pyrrolidone (NMP) and bore fluid of NMP/H₂O were used in 460 spinning-sessions of the fibers using a well-known dry/wet spinning process. Optimized deacetylation of spun-CA hollow fibers (CAHF) was achieved by using 90 vol% 0.075 M NaOH aqueous solution diluted with 10 vol% isopropanol for 2.5 h at ambient temperature. Cellulose hollow fibers (CHF) dried at room temperature and under RH (80% → ambient) overnight gave maximum yield for both dried CHF, as well as carbon fibers. The gas permeation properties of carbon fibers were also high (CO₂ permeability: 50⁻450 Barrer (1 Barrer = 2.736 × 10-9 m³ (STP) m/m² bar h), and CO₂/CH₄ selectivity acceptable (50⁻500).
Biogas is an increasingly attractive renewable resource that needs to be upgraded to meet either the natural gas pipeline quality or the biomethane requirement. The use of carbon membranes is a promising alternative to conventional separation technologies, but, up to now, few attempts have been made to systematically investigate their separation performance under real biogas upgrading conditions, for instance, the presence of water vapor and H2S in the feed stream. In this work, for the first time in literature, the separation performances of as-prepared cellulose-based carbon hollow fiber membranes were monitored for 183 days under continuous exposure to a gas stream also containing H2S and/or water vapor. After the evaluation of their CO2 and CH4 sorption (2,98 mmol g−1 and 2,00 mmol g−1, respectively at 298 K and 10 bar), diffusion(2.45 × 10-7 cm2 s−1 for CO2), and permeation properties (120.9 and 2.3 barrer, respectively at 308 K) with single gases, long-term tests were carried out by feeding gas mixtures with typical biogas composition. It was found that the exposure of the membranes to H2S (up to 500 ppm) and water vapor (relative humidity of 90%) provoked a reduction of CO2 and CH4 permeability compared to the "clean and dry" mixed gas, which resulted in an increment of selectivity, which reached a value of more than 200. Overall, after more than 180 days of continuous testing, the membranes exhibited remarkable CO2/CH4 selectivity with endurability of the H2S and water vapor, confirming the fact of being good candidates for biogas upgrading.
One way of contributing to combat the climate change is to capture CO2 from fossil fuel flue gases. Membranes will clearly represent one of the emerging technologies to be used for CO2 capture. In this work, a membrane pilot at the Norcem Cement factory in Norway is reported for CO2 capture from a high CO2 content (17 mol. % wet base) flue gas. The polyvinylamine (PVAm) based hollow fiber fixed-site-carrier (FSC) membrane modules (up to 18m2) was installed at the site. The membrane modules were received as commercial modules from Air Products (US-Norway), and were coated in-situ at NTNU in Norway. The testing results indicated a 70 mol.% CO2 purity can be easily achieved in single stage. The membrane also presented a good stability by exposure to high concentration SO2 and NOX for a long period without significant performance change. Improved design for both process and module will be needed for further scaling up of the membrane CO2 capture process. Based on the test results, a techno-economic feasibility analysis of CO2 capture from was conducted, using process simulation and cost estimation.
Alarming increase in global CO2 emissions warrants acceleration of CO2 capture technologies. In this work, testing of pre-pilot scale membrane modules containing hybrid facilitated transport membranes in hollow fiber configuration is reported. The tests were carried out using real flue gas from a slipstream in the Colacem cement plant located in Gubbio, Italy. With the fabricated modules, CO2 flux of up to 750 NL m−2h−1 with a CO2 permeate purity ranging from 50 to 55 vol% was recorded. All pre-pilot membrane modules showed increased CO2 permeance in industrial testing (1.5 to 1.9x higher) compared to laboratory evaluation owing to the reliable water profile and high temperature of the flue gas from the chimney. Influence of operating parameters (e.g., pressures in the feed and permeate) were studied. Long-term testing showed no obvious reduction in permeation performance. Furthermore, the membranes with mobile carriers when exposed to the feed gas containing SOx and NOx exhibited good resistance to performance deterioration despite high concentrations of acidic impurities. Simulation studies based on validated experimental performance under industrial conditions reveal the high potential of the fabricated membranes as an efficient separation unit capable of achieving industrial capture rate and CO2 purity requirements using a relatively low membrane area.
Chlorine as a chemical is widespread in industry and found in a great variety of processes ranging from water purification to plastic production. In this thesis, a magnesium production factory was chosen as an example because it involved both chlorine - air separation and hydrogen –hydrogen chloride separation.Previously, various types of membrane materials have been tested out for their applicability in the chosen process. The materials previously tested either lacked sufficient membrane performance or sufficient membrane stability. As an attempt to improve both the membrane performance and stability, glass membranes are used in this thesis.Glass membranes are prepared from a borosilicate glass, via a phase separation followed by an acid leaching route. By choosing the appropriate phase separation temperature and acid to glass ratio, the membrane can be produced with an average pore diameter of 2 nm (or 4 nm).However, the 2 nm average pore size is still too large to separate gases with separation selectivities beyond the selectivities predicted from Knudsen diffusion theory. If the pores are narrowed, the selectivity may be raised while the flux hopefully is maintained. The narrowing of the pores was done by a silane coupling to the surface OH-groups on the glass. The silane coupling agent is of the dimethylacyl-chlorosilane type, where the length of the acyl chain varies from 1 carbon up to 18 carbons. Glass fibres are also tested in this work, which are produced without phase separation and their average pore size is smaller than the surface-modified glasses.To be able to compare the performance of the various membranes, permeance measurements are performed and these measurements are evaluated by the separation power (product of the selectivity and the permeability of the fastest permeating compound). Because of the harsh chlorine or hydrogen chloride environment, to which the membranes are exposed in this work, the membrane stability is at least as important s factor as the perm-selectivities. To evaluate this, both short- and longterm aggressive gas exposures are performed using a special designed durability chamber. From the combination of the perm-selectivities and the durability tests, the following conclusions may be drawn (evaluated at 30°C and 1 bar): Firstly, the pure glasses have a relatively poor stability (for chlorine gas) and the perm-selectivity is too low (for both separations in question). Secondly, the C8 and C12 modified glass membranes have a relatively satisfactory perm- selectivity for chlorine separation, but the durability in chlorine is poor. Thirdly, the long-chained C18 modified glass membrane has a relatively satisfactory perm-selectivity but a fair to low chlorine stability. If the C18 membrane is applied in the hydrogen chlorine separation the perm-selectivity is a bit low, but the stability is sufficient. However, this membrane is the best choice for a low temperature HCl selective membrane.Finally, to improve the chlorine stability, a perfluorinated version of a C1 modification is tried out. This membrane has excellent chlorine stability, and the perm-selectivity is fair. This membrane is the best choice for a chlorine selective membrane.The stability of the fibres is comparable to that found for the pure glass tubes. However, the permeabilities in the glass fibres are several orders of magnitude lower than for the glass tubes. The pore size in the fibre is so narrow that separation occurs according to a molecular sieving mechanism. The mounting of the fibres into a labsized module is tricky and the permeabilities are at the border of detection, so the results obtained here should only serve as trends.