Supported membranes whose selective layer is a microporous molecular-sieve crystal film — zeolites or metal–organic frameworks (MOFs) — promise simultaneously high permeance and high selectivity for energy-intensive molecular separations such as hydrogen recovery, olefin/paraffin splitting, and carbon capture, because flux through a defect-free crystal film scales inversely with film thickness. The unresolved problem is a fabrication barrier, stated at its natural generality in the source literature: producing selective layers that are simultaneously thin enough (effective transport thickness of a few to a few tens of nanometres, including routes that place the selective phase inside the pore system of the support rather than as a film on top of it) to give order-of-magnitude permeance gains, and continuous enough — free of grain-boundary and pinhole leakage — to retain intrinsic single-crystal-level selectivity over practically relevant areas under pressurized mixed-feed conditions. A 2017 gel–vapour deposition breakthrough and its successors (4-nm oriented films on model substrates, large-area self-repaired 50–130 nm layers, single-crystal benchmarks, nano-confined embedded architectures) each resolve only part of the combined objective; as of 2026 no fabrication route achieves all of thinness, leak-free continuity, practical area, and realistic-condition stability for the molecular-sieve membrane class at large.

Background

A molecular-sieve membrane separates molecules by letting them diffuse through pores of molecular dimensions while excluding larger species. When the sieve is an inorganic or coordination crystal framework — a zeolite (a crystalline aluminosilicate with uniform sub-nanometre pore openings) or a metal–organic framework, MOF (a crystalline network of metal nodes joined by organic linker molecules, exemplified by the zeolitic imidazolate frameworks, ZIFs, in which zinc or cobalt nodes are bridged by imidazolate linkers) — a defect-free film can in principle combine high permeance (flux per unit transmembrane pressure difference, reported in mol m^-2 s^-1 Pa^-1 or in GPU, gas permeation units, with 1 GPU ≈ 3.35 × 10^-10 mol m^-2 s^-1 Pa^-1) with sharp, size-exclusion-based selectivity (the ratio of permeances of the two gases in a mixture). Because transport through a continuous crystal film is diffusive, permeance is inversely proportional to the thickness of the layer the molecules must cross: a 20-nm layer transports on the order of a hundred times more than a 2-μm layer of the same material. Throughout this task, a membrane's effective selective-layer transport thickness — also called its effective selective-transport thickness, or simply its effective transport thickness — is defined operationally as the thickness inferred from the measured permeance and the material's intrinsic permeability through this inverse-thickness relation (effective thickness = intrinsic permeability / measured permeance), so that it measures the actual path length the permeating molecules cross rather than any geometric film dimension; for routes that embed the selective phase inside the pore system of the support (the embedded, or nano-confined, architecture), it is therefore the intrusion depth through the support-pore network, not the local thickness of the deposited phase.

In practice these membranes are asymmetric composites: the crystal film must be supported on a porous support (typically porous alumina or a polymer hollow fibre) that provides mechanical strength, and only the crystal film is selective. Polycrystalline films are grown on or in this support by hydrothermal or solvothermal crystallization (crystal growth from hot aqueous or organic-solvent precursor solutions in closed vessels under autogenous pressure). Two failure modes dominate. First, thickness: conventionally grown continuous films are hundreds of nanometres to several micrometres thick, which throttles flux. Second, discontinuity: between adjacent crystals are grain boundaries — imperfectly intergrown interfaces — and pinholes; these sub-nanometre-to-nanometre leakage paths are far more permeable and far less selective than the crystal pores, so even rare defects destroy separation performance. The two demands fight each other: thinning the layer leaves fewer crystal generations to close intercrystalline gaps and weakens intercrystallite contacts, while driving intergrowth to seal defects typically costs thickness. This is the thickness–continuity barrier.

The barrier was stated sharply by Li et al. (Nature Communications 8 (2017) 406), whose gel–vapour deposition route — sol–gel coating of the metal source followed by vapour deposition of the organic linker — produced supported MOF membranes as thin as about 17 nm with gas permeances one to three orders of magnitude above conventional membranes and good H2/C3H8, CO2/C3H8 and C3H6/C3H8 selectivities, scaled in situ to a 30-hollow-fibre module. Around and after that landmark, several distinct strategies attacked different faces of the barrier: post-fabrication conversion of a dense, impermeable deposit into a permselective ZIF-8 layer by exposure to linker vapour (ligand-induced permselectivation, Science 361 (2018) 1008); growth of MOF material inside the pore system of the support rather than on top of it (the nano-confined membrane architecture, systematized in a 2021 review in Advanced Membranes, which explicitly targets the contradiction between integrity and ultrathin thickness, plus film–support adhesion and crystal orientation); single-crystal ZIF-8 membranes, which remove grain boundaries entirely but only over microscopic areas; and large-area self-repair growth strategies yielding rollable, module-compatible ultrathin layers (Advanced Materials 35 (2023) 2307013; Angewandte Chemie International Edition 63 (2024) e202404058). Field reviews through 2024 (Chemical Reviews 120 (2020) 8161; Chemical Society Reviews 50 (2021) 1913; Journal of Membrane Science 690 (2024) 122201) continue to list the simultaneous achievement of ultrathin, defect-free, scalable, stable selective layers as a central unsolved problem of zeolite and MOF membrane synthesis.

Problem Statement

How can supported microporous molecular-sieve membranes — selective layers of crystalline inorganic or coordination frameworks such as zeolites, MOFs, and ZIFs, with no material class or method restriction — be fabricated so that the effective selective-transport thickness is minimized to the few-to-tens-of-nanometre scale while the layer remains free of intercrystalline (grain-boundary) and pinhole leakage? Concretely: demonstrate a fabrication route (surface-deposited, embedded within the pore system of the support, converted post-deposition, or otherwise) that is not specific to a single material system, that is, demonstrated on at least two chemically distinct frameworks (for instance a zeolite and a MOF or ZIF), and that simultaneously delivers (i) an effective selective-layer transport thickness of at most 50 nm, such that permeances are at least an order of magnitude above conventional micrometre-scale films of the same material; (ii) continuity at the intrinsic crystal level — mixed-gas selectivities, measured under pressurized mixed-feed conditions, that lie no more than a factor of two below the single-crystal or computed intrinsic selectivity of the same material, so that they are not limited by grain-boundary or defect leakage; and (iii) retention of (i)–(ii) over membrane areas of at least 200 cm² including a module-scale demonstration, with mechanical robustness sustained through activation and at least 1000 h of continuous pressurized mixed-feed operation. The thickness–continuity barrier itself is the unresolved target: thinning the layer starves intergrowth and weakens intercrystallite contacts, while driving intergrowth to seal defects costs thickness; no current strategy is established to beat this trade-off at generality.

References

  1. Wanbin Li, Pengcheng Su, Zhanjun Li, Zehai Xu, Fei Wang, Huase Ou, Jiaheng Zhang, Guoliang Zhang, Eddy Zeng, 'Ultrathin metal–organic framework membrane production by gel–vapour deposition', Nature Communications 8 (2017) 406, DOI 10.1038/s41467-017-00544-1, https://doi.org/10.1038/s41467-017-00544-1
  2. Xiaoli Ma, Prashant Kumar, Nitish Mittal, Alexandra Khlyustova, Prodromos Daoutidis, K. Andre Mkhoyan, Michael Tsapatsis, 'Zeolitic imidazolate framework membranes made by ligand-induced permselectivation', Science 361 (2018) 1008–1011, DOI 10.1126/science.aat4123, https://doi.org/10.1126/science.aat4123
  3. Chen Chen, Aydin Ozcan, A. Ozgur Yazaydin, Bradley P. Ladewig, 'Gas permeation through single-crystal ZIF-8 membranes', Journal of Membrane Science 575 (2019) 209–216, DOI 10.1016/j.memsci.2019.01.027, https://doi.org/10.1016/j.memsci.2019.01.027
  4. Chao Zhang, Bai-Heng Wu, Meng-Qi Ma, Zuankai Wang, Zhi-Kang Xu, 'Ultrathin metal/covalent–organic framework membranes towards ultimate separation', Chemical Society Reviews 48 (2019) 3811–3841, DOI 10.1039/c9cs00322c, https://doi.org/10.1039/c9cs00322c
  5. Qihui Qian, Patrick A. Asinger, Moon Joo Lee, Gang Han, Katherine Mizrahi Rodriguez, Sharon Lin, Francesco M. Benedetti, Albert X. Wu, Won Seok Chi, Zachary P. Smith, 'MOF-Based Membranes for Gas Separations', Chemical Reviews 120 (2020) 8161–8266, DOI 10.1021/acs.chemrev.0c00119, https://doi.org/10.1021/acs.chemrev.0c00119
  6. Xiangqiong Wang, Naixin Wang, Xiaoting Li, Quan-Fu An, 'A review of nano-confined composite membranes fabricated inside the porous support', Advanced Membranes 1 (2021) 100005, DOI 10.1016/j.advmem.2021.100005, https://doi.org/10.1016/j.advmem.2021.100005
  7. Zixi Kang, Hailing Guo, Lili Fan, Ge Yang, Yang Feng, Daofeng Sun, Svetlana Mintova, 'Scalable crystalline porous membranes: current state and perspectives', Chemical Society Reviews 50 (2021) 1913–1944, DOI 10.1039/d0cs00786b, https://doi.org/10.1039/d0cs00786b
  8. Xueling Wang, Qiang Lyu, Tiezheng Tong, Kuo Sun, Li-Chiang Lin, Chuyang Y. Tang, Fenglin Yang, Michael D. Guiver, Xie Quan, Yingchao Dong, 'Robust ultrathin nanoporous MOF membrane with intra-crystalline defects for fast water transport', Nature Communications 13 (2022) 266, DOI 10.1038/s41467-021-27873-6, https://doi.org/10.1038/s41467-021-27873-6
  9. Dennis T. Lee, Peter Corkery, Sunghwan Park, Hae-Kwon Jeong, Michael Tsapatsis, 'Zeolitic Imidazolate Framework Membranes: Novel Synthesis Methods and Progress Toward Industrial Use', Annual Review of Chemical and Biomolecular Engineering 13 (2022) 529–555, DOI 10.1146/annurev-chembioeng-092320-120148, https://doi.org/10.1146/annurev-chembioeng-092320-120148
  10. Caijiao Yu, Xixi Cen, Zhengqing Zhang, Yuxiu Sun, Wenjuan Xue, Zhihua Qiao, Michael D. Guiver, Chongli Zhong, 'Step-Nucleation In Situ Self-Repair to Prepare Rollable Large-Area Ultrathin MOF Membranes', Advanced Materials 35 (2023) 2307013, DOI 10.1002/adma.202307013, https://doi.org/10.1002/adma.202307013
  11. Qiang Ma, Jiajia Li, Yanshuo Li, Jungkyu Choi, 'Recent progress and issues facing zeolite and metal–organic framework membranes: From membrane synthesis to applications', Journal of Membrane Science 690 (2024) 122201, DOI 10.1016/j.memsci.2023.122201, https://doi.org/10.1016/j.memsci.2023.122201
  12. Shuqing Song, Qi Liu, S. Swathilakshmi, Heng-Yu Chi, Zongyao Zhou, Ranadip Goswami, Dmitry Chernyshov, Kumar Varoon Agrawal, 'High-performance H2/CO2 separation from 4-nm-thick oriented Zn2(benzimidazole)4 films', Science Advances 10 (2024) eads6315, DOI 10.1126/sciadv.ads6315, https://doi.org/10.1126/sciadv.ads6315
  13. Yueyao Liang, Zhengqing Zhang, Aibing Chen, Caijiao Yu, Yuxiu Sun, Juan Du, Zhihua Qiao, Zhi Wang, Michael D. Guiver, Chongli Zhong, 'Large-Area Ultrathin Metal–Organic Framework Membranes Fabricated on Flexible Polymer Supports for Gas Separations', Angewandte Chemie International Edition 63 (2024) e202404058, DOI 10.1002/anie.202404058, https://doi.org/10.1002/anie.202404058
  14. Shuqing Song, Ceren Kocaman, Annal Baisil, Jian Hao, Shaoyu Wang, Kumar Varoon Agrawal, 'Uniform In-Plane Growth of Large-Area Ultrathin Zn2(benzimidazole)4 Membrane', Advanced Functional Materials 36 (2025), DOI 10.1002/adfm.202525243, https://doi.org/10.1002/adfm.202525243