Single-Crystal COF Membranes Shatter Barriers for Cheaper Carbon Capture

Carbon capture has long been trapped in a difficult trade-off. Membranes that let carbon dioxide pass through quickly tend to let nitrogen and other gases slip through too, while highly selective membranes are often so slow that industrial-scale deployment becomes economically unattractive. A team led by researchers at Tianjin University, working with collaborators at EPFL, the University of Southern California and other institutions, now reports a way out of this impasse. Writing in Nature Energy, they describe an all-organic mixed-matrix membrane that embeds single-crystal covalent organic frameworks, or COFs, into a polymer matrix at loadings once thought impractical, achieving carbon dioxide permeability of 74,800 Barrer while maintaining a carbon dioxide-to-nitrogen selectivity of roughly 20.

The central obstacle the team set out to solve is a familiar one in membrane science. Mixed-matrix membranes combine the easy processing of polymers with the precisely engineered pores of crystalline materials, but the recipe has a built-in ceiling. When the fraction of porous filler climbs beyond a modest threshold, the particles aggregate, defects appear at the polymer–filler interface, and the membrane’s performance collapses rather than improves. Most strategies have therefore kept filler loadings low, which means the polymer dominates transport and the carefully designed pores of the filler contribute little. The intrinsic transport properties of the crystalline material are effectively wasted.

The Tianjin-led group overturned the conventional design philosophy. Rather than maximizing strong interactions between polymer and filler to glue the two phases together, they pursued what they call a weak-interaction inlaying strategy. By balancing the interaction energies between polymer and filler on one hand and between filler particles themselves on the other, they found that the single-crystal COFs could be dispersed uniformly at loadings reaching 75.5 percent by volume, an extraordinary fraction for a mixed-matrix membrane. At such loadings the filler no longer acts as a minor additive; it becomes the dominant, continuous transport network through which gas molecules travel.

The choice of filler matters as much as the loading. Covalent organic frameworks are a class of porous crystalline polymers built from light elements such as carbon, nitrogen, oxygen and hydrogen, linked by strong covalent bonds into periodic lattices with regularly sized pores. Unlike metal–organic frameworks, they contain no metal nodes, making them fully organic and compatible with an organic polymer matrix. The team used single-crystal COFs, whose well-defined, defect-minimized lattices preserve the exact pore geometry designed into the material. When these crystals are embedded intact, the membrane inherits their intrinsic transport channels rather than a degraded approximation of them.

The resulting membranes create what the authors describe as filler-dominated pathways for fast carbon dioxide transport. Gas molecules moving through the membrane encounter a continuous network of COF pores whose size and chemistry favor carbon dioxide over nitrogen, while the surrounding polymer simply holds the architecture together. Because the interfaces remain clean and compatible even at very high filler content, there are no leaky voids that would undermine selectivity and no dense polymer regions that would choke off flow. The combination of high permeability and maintained selectivity pushes the materials well beyond the conventional upper-bound trade-off line that has long constrained polymer membranes for carbon dioxide–nitrogen separations.

The numbers are striking in context. A permeability of 74,800 Barrer is orders of magnitude above typical commercial polymer membranes for post-combustion capture, which often sit in the tens to low hundreds of Barrer. High permeability translates directly into smaller membrane areas and lower driving pressures for a given capture duty, which in turn reduces the capital cost and energy consumption of a capture plant. Previous life-cycle assessments have suggested that membrane separation can compete with the amine-based chemical absorption processes used today, but only if membrane materials reach the permeability–selectivity combination that makes module sizes practical. This work moves that target substantially closer.

Performance on day one means little if a membrane degrades, and polymer membranes are notorious for physical ageing, the gradual collapse of free volume that throttles transport over months and years. The team addressed this concern directly with long-term tests. Membranes based on polyimide retained approximately 85 percent of their initial permeability after six months of ageing at room temperature, a result that suggests the rigid crystalline COF network helps stabilize the membrane structure against the densification that normally erodes polymer performance. For a material operating at such extreme filler loadings, this stability is itself a notable finding.

The work also carries a broader methodological lesson for materials chemistry. The weak-interaction inlaying strategy reframes the interface problem: instead of fighting aggregation with ever-stronger polymer–filler bonding, the researchers tuned the relative strengths of competing interactions so that dispersion remains thermodynamically comfortable. That principle, balancing polymer–filler against filler–filler interaction energies, could plausibly extend beyond COFs to other crystalline fillers and beyond carbon capture to other molecular separations, from hydrogen purification to hydrocarbon cracking. The study’s authors, including co-first authors Hanze Ma and Shilin Guo, with Guangwei He and Zhongyi Jiang as corresponding authors, frame the approach as a general route to exploiting the intrinsic properties of porous crystals inside processable membranes.

Scalability remains the decisive question for any laboratory membrane advance. The team emphasizes solution processability as a design requirement, and the fact that the membranes are all-organic, formulated from polymer and COF in solution, aligns with existing membrane fabrication methods rather than requiring exotic deposition techniques. Still, translating a laboratory coupon into square meters of defect-free membrane module, and validating performance on real flue gas with its humidity, sulfur impurities and temperature swings, will demand further engineering. The six-month ageing data and the robust interfacial design are encouraging early signals, but industrial pilots will be the true test.

Even with those caveats, the study represents one of the most compelling demonstrations yet that the long-standing permeability–selectivity trade-off can be beaten by design rather than merely nudged. By making the crystalline filler the membrane rather than a garnish on it, the researchers have effectively built a processable, polymer-locked single-crystal separation device. If the approach survives scale-up, membranes of this kind could shrink the footprint and energy bill of post-combustion carbon capture, a technology the Intergovernmental Panel on Climate Change and the International Energy Agency both identify as essential to limiting warming. In a field where incremental gains are the norm, a seventy-thousand-Barrer membrane with intact selectivity is the kind of result that redraws the map.

Subject of Research: All-organic mixed-matrix membranes with single-crystal covalent organic frameworks for energy-efficient carbon capture

Article Title: All-organic mixed-matrix membranes with single-crystal covalent organic frameworks for carbon capture

Article References: Ma, H., Guo, S., Su, P., Zhao, Q., Wang, S., Ren, Y., Fu, J., Zhang, H., Zhang, T., Yuan, S., Pan, Q., Villalobos, L. F., Li, Y., Ma, X., Agrawal, K. V., He, G., & Jiang, Z. (2026). All-organic mixed-matrix membranes with single-crystal covalent organic frameworks for carbon capture. Nature Energy. https://doi.org/10.1038/s41560-026-02149-9

Image Credits: AI Generated

DOI: 10.1038/s41560-026-02149-9

Keywords: carbon capture, covalent organic frameworks, mixed-matrix membranes, gas separation, CO2 permeability, polyimide, membrane technology, porous materials, post-combustion capture, polymer membranes, Nature Energy, materials science