Revealing Hydroxide Ion Transport Mechanisms in Commercial Anion-Exchange Membranes at Nano-Scale from Machine-learned Interatomic Potential Simulations

Abstract

Hydroxide ion transport in anion-exchange membranes fundamentally limits the efficiency of alkaline water electrolysis for green hydrogen production, yet the atomic-scale transport mechanisms remain poorly understood due to the computational challenges associated with modeling ion dynamics. Given that anion-exchange membranes enable alkaline electrolysis with abundant catalysts while avoiding perfluoroalkyl and polyfluoroalkyl materials, a deeper mechanistic understanding of hydroxide transport in these systems is essential for advancing sustainable hydrogen production. Here, we show that large-scale molecular dynamics simulations with fine-tuned machine-learned interatomic potentials provide atomistic insight into hydroxide mobility in a commercial membrane over tens of nanoseconds and over ten nanometer. We find that increasing water content transforms isolated water clusters into a connected hydrogen-bond network that enables long-range proton transfer. Under dry conditions hydroxide ions are trapped near positively charged groups and transport is strongly hindered, whereas well-hydrated membranes exhibit extended proton migration and diffusion coefficients approaching those of dilute aqueous solutions. The simulations reproduce experimental trends in diffusion and activation energies. Our results establish a direct link between nano-scale structure and macroscopic transport. Beyond mechanistic insight, the presented simulation framework enables predictive, simulation-guided optimization of membrane chemistry and architecture, opening a pathway toward the rational design of more efficient anion-exchange membranes for green hydrogen technologies.

Figures (4)

• Figure 1: Hydroxide ion diffusion within a single anion-exchange membrane channel. Diffusion pathways of hydroxide ions over $0.5$ ns within the membrane channel at hydration level $\lambda = 10$ and temperature $350$ K, obtained from fine-tuned MLIP molecular dynamics simulations. Each colored line represents the trajectory of one of the eight hydroxide ions. Hydrogen, carbon, oxygen, and nitrogen atoms in the polymer backbone are shown in white, turquoise, red, and blue, respectively. The black lines outline the simulation cell of one periodic image with a width of $20$ Å. The single-channel systems comprise up to 1,700 atoms, while larger systems containing more than 30,000 atoms were also investigated in this work. For additional details, see Figure \ref{['fig:2']}, Supplementary Note 7 and the Methods section.
• Figure 2: Hydroxide diffusion in anion-exchange membrane channels as a function of hydration.a Diffusion coefficients of hydroxide ions (colored bars) and water molecules (white bars) as a function of membrane water content $\lambda$, defined as the number of water molecules per functional group. Hydroxide error bars represent the standard deviation obtained by evaluating the diffusion of each OH$^-$ ion individually within the same trajectory. Simulation results are derived from fine-tuned MLIP molecular dynamics simulations at $350$ K. Experimental diffusion coefficients from Marino et al. marino2014hydroxide, measured at $298$ K, are shown for comparison. A enlarged representation of the low hydration regime is provided in Supplementary Note 3. b Representative membrane channel morphologies viewed along the $z$ direction. The aqueous phase is rendered as a gray continuous medium. Carbon, oxygen, and nitrogen atoms in the polymer strands are colored turquoise, red, and blue, respectively. The legend summarizes the structural appearance at different $\lambda$ values. All structures are shown at identical visual scale for direct comparison of morphology. Consequently, the apparent channel dimensions do not reflect their absolute physical size at the respective hydration levels. c Oxygen–hydrogen bond lifetime within hydroxide ions, expressed as the fractional (O--H)$_{\text{OH}^-}$ bond correlation function, at different hydration levels. d Ratio of hydroxide to water diffusion coefficients, $D$(OH$^-$)/$D$(H$_2$O), highlighting the contribution of structural proton transfer to long-range transport. e Arrhenius representation of hydroxide diffusion coefficients and corresponding activation energies assuming Arrhenius behavior. The underlying numerical data are compiled in Supplementary Note 2.
• Figure 3: Hydrogen-bond network topology and hydroxide coordination in membrane nanochannels.a Hydrogen-bond network within the aqueous phase of the membrane channel at $\lambda=3$ (upper panel) and $\lambda=10$ (lower panel). The polymer backbone and functional groups are omitted for clarity. Hydrogen atoms are shown in white, oxygen atoms of water molecules in red, and oxygen atoms of hydroxide ions in purple. Hydrogen bonds are indicated by red dashed thin lines. Orange dashed lines mark percolation gaps between disconnected water clusters. The blue circles highlights water molecules that transiently bridge two clusters and act as dynamic bottlenecks for proton transport. Snapshots are taken from fine-tuned MLIP molecular dynamics simulations of the large simulation cells, as described in Supplementary Note 7 and in Methods. b Representative coordination geometries of a hydroxide ion in the membrane channel at $\lambda=3$ (upper panel) and $\lambda=10$ (lower panel), extracted from AIMD structures. Water oxygen atoms are colored red, hydroxide oxygen atoms purple, carbon atoms turquoise, nitrogen atoms blue, and hydrogen atoms white. Hydrogen bonds are shown as dashed lines, colored red when the coordination partner is a water molecule and turquoise when coordinated by the polymer. c Distribution of hydroxide coordination numbers obtained from AIMD molecular dynamics trajectories at $350$ K for $\lambda=3$ (upper panel) and $\lambda=10$ (lower panel). The right-hand histograms display the average contribution of coordination partners as a function of coordination number, with water neighbors shown in red and polymer contributions in turquoise. The coordination number is determined by identifying the two nearest atoms for each hydrogen atom.
• Figure 4: Long-range hydroxide transport and interactions with functional groups.a Diffusion pathway of a single OH$^-$ ion over $60$ ps within a thin slice of the membrane channel at $\lambda=3$ and $350$ K. For clarity, only a confined section of the simulation cell along the $z$ direction is shown. The four polymer strands are represented by one monomer unit each, and the backbone is truncated at the oxygen atoms for clarity. Oxygen, carbon, nitrogen, and hydrogen atoms are colored red, turquoise, blue, and white, respectively. The trajectory is extracted from the AIMD simulation at $350$ K and $\lambda=3$. b Distance distributions between hydroxide ions and their nearest neighboring species at different hydration levels. The intense colored filled curves represent the OH$^-$–water oxygen distance for each $\lambda$. For clarity, the distributions of distances to the nearest functional group carbon atom and to the nearest backbone carbon atom are plotted upside down. The OH$^-$–functional group carbon distribution is shown in semi-transparent color, whereas the OH$^-$–backbone carbon distribution is indicated by a gray line. c Radial distribution functions between hydroxide oxygen atoms and nitrogen atoms of the functional groups for different hydration levels, are shown as dashed gray lines. The colored solid curves display the distribution of proton transfer events as a function of the instantaneous OH$^-$–nitrogen distance. d Hydroxide diffusion pathways over $0.5$ ns within the membrane channel at $\lambda=3$, $5$, $10$, and $20$, shown in the upper left, upper right, lower left, and lower right panels, respectively. Trajectories are obtained from fine-tuned MLIP molecular dynamics simulations at $350$ K. e Partnership correlation function between hydroxide ions and their nearest nitrogen neighbors, illustrating the residence time of hydroxide ions in proximity to functional groups at different hydration levels.