Zeolite MFI membranes can be crystallized in a single step on low-cost α-alumina hollow fiber supports from a concentrated silicate gel containing MFI nanocrystals, yielding sub-micrometer selective layers with nn-butane permeance above 10−710^{-7} mol m−2^{-2} s−1^{-1} Pa−1^{-1}, nn-butane/ii-butane separation factors above 50, and a ten-fiber module fabricated in one batch (Min et al., 2020). That paper left the true deployable performance of this membrane system undetermined, and two coupled issues remain open. First, the origin of the long-term permeance decline observed during extended operation was not isolated: trace higher-alkane feed impurities and slow desorption of high-molecular-weight organics from the epoxy potting material were both plausible candidate causes, so the intrinsic long-term stability of the membrane/support system remains unknown. Second, the reduced nn-butane permeance of the ten-fiber module relative to single fibers was attributed to feed-side concentration polarization and flow limitations, and recovery of single-fiber nn-butane permeance under improved operating conditions was expected but not demonstrated for this platform. Later literature resolved the adjacent generality question (one-step gel conversion was extended from CHA in 2019 to CHA and MFI via viscosity-confined dry gel conversion in 2022) and quantified concentration polarization in scaled silicalite-1 modules made by other synthesis routes on monolithic supports (2024), yet no work isolates the stability origin for the one-step hollow-fiber membranes or demonstrates single-fiber performance recovery in their modules. The problem is to close this gap through controlled attribution experiments and/or validated module transport modeling.

Background

Zeolites are crystalline aluminosilicates with uniform molecular-sized pores; MFI (also called ZSM-5, or silicalite-1 in its all-silica form) has intersecting straight and sinusoidal channels of about 0.55 nm that admit linear hydrocarbons such as nn-butane while excluding branched isomers. A zeolite membrane is a continuous polycrystalline film of this material on a porous support; its usefulness is measured by the permeance (flux divided by the transmembrane partial-pressure driving force, in mol m−2^{-2} s−1^{-1} Pa−1^{-1}) and the separation factor (selectivity of a mixed-gas permeance ratio relative to the feed composition ratio). MFI membranes are promising replacements for energy-intensive distillation in butane-isomer purification and in removing natural gas liquids (ethane, propane, butane) from methane, but industrial adoption has been limited by the cost of conventional fabrication, which typically requires seeding the support with crystals followed by a separate hydrothermal secondary-growth step. Hollow fiber supports — fine ceramic tubes of α-alumina with outside diameter below a millimeter — are cheap and give very high membrane area per module volume, but coating many of them uniformly is difficult. Min et al. (Adv. Mater. Interfaces 7 (2020) 2000926) showed that a single-step crystallization works: a highly concentrated silicate precursor gel containing MFI nanocrystals is applied to the fibers and converts directly into a compact sub-micrometer membrane layer, and ten membranes were synthesized and assembled into a module in one batch with close-to-single-fiber nn-butane/ii-butane selectivity. The same group had earlier shown a one-step gel conversion route for small-pore CHA zeolite hollow fibers (Yang, Chiang and Nair, Energy Technol. 7 (2019) 1900494, including a 20-fiber module), and in 2022 generalized the chemistry to a viscosity-confined dry gel conversion (VCDGC) process demonstrated in detail for CHA and also for MFI (Yang et al., Angew. Chem. Int. Ed. 61 (2022) e202204265). Two performance-fidelity issues were left open by the 2020 paper and remain unsettled for this platform. (1) Long-term stability: during extended permeation the permeance slowly declined; the authors could not distinguish whether trace higher-alkane impurities in the feed or slow desorption of high-molecular-weight organic species from the epoxy resin used to pot (seal) the fibers into the module was responsible, so the intrinsic durability of the membrane/support system was never isolated. (2) Module-scale transport: the ten-fiber module showed lower nn-butane permeance than single fibers, attributed to concentration polarization — the depletion of the faster-permeating species in a boundary layer near the membrane surface when feed-side flow cannot replenish it — combined with feed-side flow limitations; the expectation that improved operating conditions would recover single-fiber behavior was not demonstrated. Concentration polarization is a module-hydrodynamics phenomenon, not a membrane-property limitation, so it must be diagnosed and mitigated at the module level (feed flow rate, flow geometry, Reynolds number). Related scale-up work on other architectures — 61-channel monolithic and tubular silicalite-1 membranes made by seeded-growth routes, including quantification of concentration polarization in 19-channel monoliths (Wu et al., Sep. Purif. Technol. 348 (2024) 127574) and Reynolds-number correlation of concentration-polarization losses at industrial scale (Hong et al., Adv. Membr. 4 (2024) 100096) — confirms the phenomenon is real, quantifiable, and mitigable for silicalite-1 generally, but these are different layer morphologies, supports, and synthesis routes from the one-step concentrated-gel hollow fibers, so the conclusions do not automatically transfer.

Problem Statement

For MFI (silicalite-1) zeolite hollow fiber membranes on α-alumina supports fabricated by one-step gel conversion (the concentrated silicate gel containing MFI nanocrystals of Min et al. 2020 and its dry-gel-conversion generalization), resolve two coupled performance-fidelity questions that together establish the true deployable performance of this membrane system: (1) Intrinsic long-term stability: identify the cause of the long-term permeance decline observed during extended hydrocarbon separation operation — trace higher-alkane feed impurities versus slow desorption of high-molecular-weight organic species from the epoxy potting material, the two candidate causes identified in the source — and thereby determine, or experimentally bound, the intrinsic long-term trajectory of permeance and separation factor of the membrane/support system under conditions that exclude these artifacts; membrane-intrinsic aging enters only as the residual possibility to be excluded if the decline persists after both artifacts are controlled. (2) Module-scale transport: quantify the contributions of feed-side concentration polarization and feed-flow limitations to the reduced nn-butane permeance of multi-fiber modules relative to single fibers, and determine the operating conditions (feed flow rate and Reynolds number, flow geometry, module design) under which scaled modules of these one-step membranes recover single-fiber nn-butane permeance, the expectation the source explicitly stated but did not demonstrate. Answers may combine controlled experiments with a validated transport model of the module that couples membrane permeation to feed-side mass transfer; the attribution in (1) must be demonstrated for the actual membrane/support system, not merely for a related architecture.

References

  1. B. Min; S. Yang; A. Korde; C. W. Jones; S. Nair, "Single-Step Scalable Fabrication of Zeolite MFI Hollow Fiber Membranes for Hydrocarbon Separations", Advanced Materials Interfaces 7 (2020) 2000926. DOI: 10.1002/admi.202000926, https://doi.org/10.1002/admi.202000926
  2. S. Yang; Y. Chiang; S. Nair, "Scalable One-Step Gel Conversion Route to High-Performance CHA Zeolite Hollow Fiber Membranes and Modules for CO2 Separation", Energy Technology 7 (2019) 1900494. DOI: 10.1002/ente.201900494, https://doi.org/10.1002/ente.201900494
  3. S. Yang; B. Min; Q. Fu; C. W. Jones; S. Nair, "High-Performance Zeolitic Hollow-Fiber Membranes by a Viscosity-Confined Dry Gel Conversion Process for Gas Separation", Angewandte Chemie International Edition 61 (2022) e202204265. DOI: 10.1002/anie.202204265, https://doi.org/10.1002/anie.202204265
  4. H. Wu; S. Yan; H. Hong; H. Xiao; B. Wang; R. Zhou, "Large-size 19-channel monolithic silicalite-1 membranes for butane separation: Separation performance and quantification of concentration polarization", Separation and Purification Technology 348 (2024) 127574. DOI: 10.1016/j.seppur.2024.127574, https://doi.org/10.1016/j.seppur.2024.127574
  5. H. Hong; K. Yu; H. Liu; R. Zhou; W. Xing, "Industrial-scale 61-channel monolithic silicalite-1 membranes for butane isomer separation", Advanced Membranes 4 (2024) 100096. DOI: 10.1016/j.advmem.2024.100096, https://doi.org/10.1016/j.advmem.2024.100096
  6. G. Dang; H. Wu; Z. Bai; J. Zhou; Y. Xie; B. Liu; S. Zhong; R. Zhou; W. Xing, "Semi-industrial-scale silicalite-1 membranes for n-butane/methane separation", Separation and Purification Technology 330 (2024) 125400. DOI: 10.1016/j.seppur.2023.125400, https://doi.org/10.1016/j.seppur.2023.125400
  7. B. Wang; H. Wu; L. Cui; H. Gao; D. Zhao; J. Zhou; S. Zhong; R. Zhou; W. Xing, "Scalable synthesis and modular properties of tubular silicalite-1 membranes for industrial butane isomer separation", Separation and Purification Technology 313 (2023) 123496. DOI: 10.1016/j.seppur.2023.123496, https://doi.org/10.1016/j.seppur.2023.123496
  8. B. Ma; Y. Zhu; H. Hong; L. Cui; H. Gao; D. Zhao; B. Wang; R. Zhou; W. Xing, "Improved silicalite-1 membranes on 61-channel monolithic supports for n-butane/i-butane separation", Separation and Purification Technology 300 (2022) 121828. DOI: 10.1016/j.seppur.2022.121828, https://doi.org/10.1016/j.seppur.2022.121828
  9. H. Wu; H. Liu; B. Wang; F. Wang; A. F. M. Ibrahim; M. Yu; S. Zhong; W. Xing; R. Zhou, "Continuous and rapid manufacturing of ultrathin zeolite membranes with exceptional gas separation performance", AIChE Journal 72 (2026) e70321. DOI: 10.1002/aic.70321, https://doi.org/10.1002/aic.70321
  10. A. Venkataraman; A. M. Tous-Granados; A. Deore; J. J. Dann; H. Song; V. Brandão; C. Sievers; M. C. Hatzell; S. S. Arora; S. Nair, "Separation of gas product mixtures from CO2 electrolysis by chabazite zeolite hollow fiber membranes", Chemical Engineering Science 337 (2027) 124682. DOI: 10.1016/j.ces.2026.124682, https://doi.org/10.1016/j.ces.2026.124682
  11. M. Sakai; Y. Koshiishi; A. Okada; K. Kawanishi; K. Masuda; M. Matsukata, "A Review of Strategies for Developing Zeolite Membranes: From a Microstructure to a Membrane Module", Crystal Growth & Design 25 (2025) 5586-5620. DOI: 10.1021/acs.cgd.5c00457, https://doi.org/10.1021/acs.cgd.5c00457