For hydrogen to be widely used as a clean energy source in everyday life, technologies that can extract only hydrogen with high purity from mixed gases are essential. KAIST researchers have presented a new strategy for developing high-performance separation membranes that can selectively filter hydrogen for clean hydrogen energy production.
KAIST (President Choongsik Bae) announced on the 13th of August that a research team led by Professor Tae-Hyun Bae of the Department of Chemical and Biomolecular Engineering has successfully introduced hydrogen-selective transport pathways at the angstrom scale inside polymer membranes and clarified their separation performance through the concept of “network completeness.”
*Angstrom (Å): An extremely small unit of length used to measure wavelengths of light or the size of atoms and molecules. One angstrom is one hundred-millionth of a centimeter, or one ten-billionth of a meter, roughly one-millionth the thickness of a human hair.
Hydrogen is drawing attention as an eco-friendly energy source because it does not emit pollutants when used. However, separating hydrogen with high purity from mixed gases generated during production remains a key challenge for commercialization.
Crystalline porous materials such as metal-organic frameworks (MOFs) and covalent organic frameworks (COFs) are advantageous because their pores can be designed uniformly. However, they are difficult to fabricate over large areas without defects and have limitations in separating small molecules such as hydrogen. Polymer membranes, by contrast, are easier to process and scale up to large areas, but because their pore formation is difficult to control precisely, it has been challenging to raise their separation performance beyond a certain level.
To combine the advantages of both types of materials, the research team designed a modular network structure in which polymer chains are linked by crosslinkers. In this process, the team focused on the limitation that conventional indicators such as the degree of crosslinking (CD) and effective crosslinking degree (ECD), which have been used to describe the extent of crosslinking, cannot determine whether pores useful for separation have actually been formed.
The researchers therefore proposed a new metric called the Bridge Connectivity Degree (BCD), which refers to the proportion of crosslinkers that are connected at both ends to form complete pathways. This made it possible to quantitatively apply the concept of “complete framework connectivity,” which has been emphasized in inorganic porous materials, to polymer networks as well.
The newly developed membrane, ms-oDMB-DB50, achieved a high bridge connectivity degree of 73%, and both its hydrogen permeability and hydrogen/nitrogen selectivity improved significantly compared with the original material, DB50. Analysis showed that the membrane contains numerous ultramicropores smaller than 3 Å, which carbon dioxide cannot access. The research team also proposed a “density-probe method,” using helium molecules, which are smaller than hydrogen, as probes to experimentally verify the existence of these ultramicropores.
The newly developed membrane also operated stably for 100 hours without any loss of performance. Its tensile strength, the force the membrane can withstand without breaking, was about twice that of previously reported high-performance polymer membranes, confirming that it also has the robust durability needed for industrial processes.
Dr. Hongju Lee said, “There have been previous attempts to combine the advantages of these two types of materials, but this study is different in that it defines ‘how completely the network is connected’ as a quantitative value and directly links that value to separation performance,” adding, “We hope this study will serve as a starting point for extending reticular synthesis, a design principle used for inorganic molecular sieves, to polymer membranes.”
Professor Tae-Hyun Bae said, “By stitching polymer chains together with crosslinkers that fit together like Lego blocks, we formed a network inside the membrane that selectively allows only small hydrogen gas molecules to pass through.”
This paper was led by Dr. Hongju Lee, currently a postdoctoral researcher at the Korea Institute of Science and Technology, as first author, with Professor Tae-Hyun Bae as corresponding author. The research was published on July 23 in the international journal Nature Communications.

Paper title: Network completeness enables angstrom-scale transport pathways in polymer membranes, DOI: https://doi.org/10.1038/s41467-026-73860-0
Author information: Hongju Lee, formerly of KAIST and currently at the Korea Institute of Science and Technology, first author; Suhyeon Choi, KAIST, second author; and Tae-Hyun Bae, KAIST, corresponding author

This research was supported by the 2025 Global C.L.E.A.N. Project and the Mid-Career Researcher Program under the Basic Research Program, funded by the Ministry of Science and ICT.