Graphene-wrapped zeolite membranes for fast hydrogen separation

The effects of global warming are becoming more serious, and there is a strong demand for technological advances to reduce carbon dioxide emissions. Hydrogen is an ideal clean energy which produces water when burned. To promote the use of hydrogen energy, it is essential to develop safe, energy-saving technologies for hydrogen production and storage. Currently, hydrogen is made from natural gas, so it is not appropriate for decarbonization. Using a lot of energy to separate hydrogen would not make it qualify as clean energy. Currently, polymer separation membranes are being studied all over the world. Polymer separation membranes have the great advantage of enlarging the separation membrane and increasing the separation coefficient. However, the speed of permeation through the membrane is extremely low, and high pressure must be applied to increase the permeation speed. Therefore, a large amount of energy is required for separation using a polymer separation membrane. The goal is to create a new kind of separation membrane technology that can achieve separation speeds that are 50 times faster than that of conventional separation membranes.

The graphene-wrapped molecular-sieving membrane prepared in this study has a separation factor of 245 and a permeation coefficient of 5.8 x 106 barrers, which is more than 100 times better than that of conventional polymer separation membranes. If the size of the separation membrane is increased in the future, it is very probable that an energy-saving separation process will be established for the separation of important gases such as carbon dioxide and oxygen as well as hydrogen.

Fig. 1. H2/CH4 separation efficiency achieved with G-MFI.
(A) Robeson plot for H2/CH4 separation for single gas separation. For G-MFI membranes, the selectivity of both single and mixed gas is indicated. The red line refers to the upper bound proposed by Robeson using polymeric membranes (44). Robeson plot details are listed in table S1. (B) G-MFI SEM image. (C) Fractured G-MFI SEM cross-sectional image. (D) TEM image highlighting the contact between two MFI crystals wrapped with graphene. (E) TEM image of nanowindows in graphene. (F) Nanowindow size distribution histogram. (G) Edge share model structure of G-MFI membrane, depicting intergranular voids; the (010) crystallographic face of MFI is shown. (H) Simplified interfacial model showing the cross-sectional view of the graphene and MFI crystal face of the G-MFI membrane. The nanowindows in the single graphene layer are expressed by blanks in the graphene layer, although a graphene layer continuously covers the zeolite crystal. TEM images show a few graphene layers with nanowindows covering an MFI zeolite crystal in the real G-MFI (fig. S4). Few-layer wrapping can be approximated with monolayer wrapping because the gas permeance between the layers is negligible.

As seen in the transmission electron microscope image in Figure 1, graphene is wrapped around the MFI-type zeolite crystal, being hydrophobic. The wrapping uses the principles of colloidal science to keep graphene and zeolite crystal planes close to each other due to reduction of the repulsive interaction. About 5 layers of graphene enclose zeolite crystals in this figure. Around the red arrow, there is a narrow interface space where only hydrogen can permeate. Graphene is also present on hydrophobic zeolite, so the structure of the zeolite crystal cannot be seen with this. Since a strong attractive force acts between graphene, the zeolite crystals wrapped with graphene are in close contact with each other by a simple compression treatment and does not let any gas through.

Fig. 2. Porosity of G-MFI membranes as determined from N2 adsorption at 77 K.
(A) N2 adsorption (Ads.) isotherms of G-MFI and MFI powder. Inset is the corresponding semilogarithmic plot. STP, standard temperature and pressure. (B) N2 adsorption isotherms of G-MFI and MFI membrane. Inset is the corresponding semilogarithmic plot. (C) N2 adsorption isotherm of compressed graphene. Inset is the corresponding semilogarithmic plot. (D) Adsorption isotherms and fractional fillings of H2, CO2, and CH4 plotted against pressure. The density of liquid H2 (0.0711 g cm−3 at 20 K and 105 Pa), solid CO2 (1.566 g cm−3 at 193 K and 105 Pa), and liquid CH4 (0.423 g cm−3 at 111 K and 105 Pa) and micropore volume (table S3) of the G-MFI membrane were used to estimate the fractional fillings.


Figure 2 shows a model in which zeolite crystals wrapped with graphene are in contact with each other. The surface of the zeolite crystal has grooves derived from the structure, and there is an interfacial channel between zeolite and graphene through which hydrogen molecules can selectively permeate. The model in which the black circles are connected is graphene, and there are nano-windows represented by blanks in some places. Any gas can freely permeate the nanowindows, but the very narrow channels between graphene and zeolite crystal faces allow hydrogen to permeate preferentially. This structure allows efficient separation of hydrogen and methane. On the other hand, the movement of hydrogen is rapid because there are many voids between the graphene-wrapped zeolite particles. For this reason, ultra-high-speed permeation is possible while maintaining the high separation factor of 200 or more.

Fig. 3. Gas separation performance of G-MFI membrane.
(A) Variation in single-gas permeability with molecular size for the gases H2, He, CO2, N2, CH4, i-C4H10, and SF6. The MIN-2 molecular sizes of H2 and He are used. (B) Selectivity of H2 plotted against molecular size. Separable gases include He, CO2, N2, CH4, i-C4H10, and SF6. (C) Selectivity change with time for the separation of the H2/CH4 equimolar mixture. (D) Robeson plot for CO2/CH4. For the G-MFI membrane, both single and mixed gas CO2/CH4 selectivity is indicated for comparison. The red line shows the Robeson plot for polymeric membranes (44). Details of the Robeson plots with references are presented in table S1. The arrows in Fig. 2 (A and B) show that the permeability and selectivity for i-C4H10 and SF6 (open symbols) are maximum and minimum values, respectively.

Figure 3 compares the hydrogen separation factor and gas permeation coefficient for methane with the previously reported separation membranes, which is called Robeson plot. Therefore, this separation membrane separates hydrogen at a speed of about 100 times while maintaining a higher separation coefficient than conventional separation membranes. The farther in the direction of the arrow, the better the performance. This newly developed separation membrane has paved the way for energy-saving separation technologies for the first time.

In addition, this separation principle is different from the conventional dissolution mechanism with polymers and the separation mechanism with pore size in zeolite separation membranes, and it depends on the separation target by selecting the surface structure of zeolite or another crystal. High-speed separation for any target gas is possible in principle. For this reason, if the industrial manufacturing method of this separation membrane and the separation membrane becomes scalable, the chemical industry, combustion industry, and other industries can be significantly improved energy consumption, leading to a significant reduction in carbon dioxide emissions. Currently, the group is conducting research toward the establishment of basic technology for rapidly producing a large amount of enriched oxygen from air. The development of enriched oxygen manufacturing technologies will revolutionize the steel and chemical industry and even medicine.

Source: |

About admin

Check Also

Toyota, Air Liquide and CaetanoBus partner to accelerate the development of hydrogen mobility in Europe

Air Liquide, CaetanoBus and Toyota Motor Europe have signed a Memorandum of Understanding with the …