FIGURE 10 A) Effect of temperature on H2/CO2 separation performance of the tubular MXene/SS2.5 membrane. B) Apparent activation energy of H2/CO2 separation through the membrane. C) Effect of feed humidity on H2/CO2 separation performance.
Furthermore, in order to evaluate whether the tubular MXene/SS membranes prepared by electrophoretic deposition had the potential for scale-up and practical applications, the long-term stability, water-vapor stability, and replicability were explored in detail. As illustrated in Figure 11A, during the long-term separation experiment up to 1100 h, both the H2/CO2 selectivity and the hydrogen permeance were relatively stable as a whole with only a small fluctuation. The average hydrogen permeance was 1452 GPU, while the average H2/CO2 selectivity was 63. Water vapor with a volume fraction of 3% was introduced into the feed gas continuously for 24 hours each time and then turned back to the dry system, the operation was carried out three times to guarantee the correctness of the test. When water vapor was added, the average hydrogen permeance decreased to 1321 GPU, and the corresponding H2/CO2 selectivity decreased to 56. When the water vapor was interrupted, the hydrogen permeance returned to 1468 GPU, and the selectivity returned to 65, demonstrating that the MXene/SS membrane had excellent water vapor resistance and stability.
In order to investigate the issue of membrane reproducibility, we evaluated the separation performance of 15 pieces of tubular MXene/SS membranes from one batch. As shown in Figure 11B, the hydrogen permeance of the produced membranes ranged from 880~1474 GPU with an average value of 1224, whereas the range of H2/CO2 selectivity was 47~64 with an average value of 56. According to the US Department of Energy (DOE) estimation, it would approach the DOE target for 90% CO2 capture if the H2/CO2 selectivity and hydrogen permeance could reach 40 and 900 GPU for a membrane, respectively. The H2/CO2 selectivity of these tubular MXene/SS membranes prepared in one batch completely meets the requirement and except only one membrane gave an hydrogen permeability less than 900 GPU. In other words, 93% of the membranes prepared in this batch exceeded the DOE target, promising further scale-up preparation with good repeatability.
Last but not least, a comparison of the state-of-the-art membranes including 2D and 3D membranes for H2/CO2separation, was shown in Table 1. Compared with the 3D membranes, it was obvious that the 2D membranes had better gas separation performance. However, these excellent 2D membranes were mainly prepared on porous ceramic alumina or AAO substrates, both of which were fragile and inconvenient for further scale-up, and the preparation methods (vacuum filtration, hot drop coating, etc.) were also not suitable for rapid membrane production. Hence, we sought to solve the problem of how to prepare 2D MXene membranes that could be closer for actual industrial production. Using the industrially mature electrophoretic deposition process has been proven to be more convenient for large-scale scale-up, and the MXene membranes can be prepared on commercial porous tubular stainless steel substrates for H2/CO2separation with good weldability. In contrast to many other membranes in Table 1, the MXene membrane prepared in this work achieved long-term stability of more than 1250 h, with H2/CO2 selectivity of 55 and hydrogen permeance of 1290 GPU, which further demonstrated the feasibility of the practical application.