Hydrogen is increasingly recognized as a promising energy source for the future owing to its emission-free nature and superior energy content per unit mass compared to conventional fuels such as gasoline. However, its low volumetric energy density necessitates pressurized containment solutions (e.g., 700-bar cylinders), which pose significant economic and safety challenges.
To address the limitations of high-pressure hydrogen storage, porous adsorbent materials have emerged as a promising alternative via low-pressure physisorption. Current research spans metal-organic frameworks (MOFs) [
1,
2], covalent organic frameworks (COFs) [
3,
4], metal hydrides, zeolites, porous aromatic frameworks (PAFs), and hydrogen-bonded organic frameworks (HOFs). MOFs exhibit exceptionally high surface areas and tunable porosity through tailored metal nodes or organic ligands [
5,
6], and recent studies have reported notable hydrogen storage capacities even under ambient conditions [
5,
6]. However, their hydrogen uptake is often limited by weak van der Waals interactions between H
2 molecules and framework surfaces [
7,
8]. COFs, constructed with robust covalent linkages, offer superior thermal and chemical stability compared to MOFs. Yet, their practical application is hindered by high synthesis costs, stringent preparation conditions, and low material yields [
9–
11].
Metal hydrides store hydrogen in atomic form within metal lattices, achieving high volumetric hydrogen densities [
12,
13]. However, sluggish absorption/desorption kinetics typically necessitate elevated temperatures or catalytic activation, which increases system complexity and energy demand [
14]. Zeolites, owing to their natural abundance and scalability, provide excellent thermal and chemical stability at low cost [
15]. Nonetheless, their micropore-dominated structures restrict hydrogen packing density, limiting their overall storage performance. PAFs, formed by inserting single or multiple phenyl rings into each C–C bonds in a diamond-like lattice, have demonstrated high-capacity hydrogen storage under pressure–temperature swing conditions [
16]. However, their amorphous or poorly crystalline nature makes structural control and precise pore-size tuning difficult. Furthermore, their low structural regularity can result in poor pore connectivity and hindered hydrogen diffusion.
In this context, HOFs have emerged as a distinct class of porous materials, offering a supramolecular design strategy that enables tunable porosity, lightweight composition, and mild synthetic conditions. HOFs, constructed through noncovalent assembly of organic building blocks, leverage directional hydrogen bonds to form tunable crystalline, supramolecular architectures [
17,
18]. Compared to MOFs and COFs, HOFs benefit from inherent flexibility, enabling mild synthesis, facile regeneration, and moisture resistance via hydrophobic organic moieties [
19,
20]. However, retaining structural stability while maintaining porosity remains a key challenge.
Furthermore, existing materials often face challenges in concurrently optimizing volumetric and gravimetric capacity—a critical requirement for automotive applications where limited tank space and driving range requirements demand maximized hydrogen storage density [
21]. Addressing this dual-capacity dilemma necessitates balancing high volumetric surface area (VSA) and elevated gravimetric surface area (GSA) within unified material systems. A strategic approach involves engineering porous architectures exclusively with lightweight elements to attain superior gravimetric metrics. In this regard, HOFs provide a unique solution due to their intrinsic capacity to harmonize these competing surface area parameters.
HOFs are typically constructed from organic molecules with equal numbers of hydrogen bond donors and acceptors, enabling the formation of inherent structural units such as dimers, trimers, or chains. This self-complementarity between donor and acceptor sites underpins the fundamental design principle of HOFs [
22]. In agreement with this principle, the natural dimerization of carboxylic acids (C=O···H−O) has made these groups the most studied function in both planar and non-planar frameworks since the first reported examples of hydrogen-bonded porous materials until present. Nitrogen-based functional groups have also been used as hydrogen bonding motifs exploring diverse structures. For instance, the directionality of N−H···O=C interaction has been elegantly explored through the constriction of the hydrogen bonding sites in heterocyclic structures such as pyridone and imidazolone [
23], and through the incorporation of pyrazole and imidazole rings into the molecular unit to assemble [
24].
Early HOFs suffered from structural fragility caused by weak hydrogen bonds, frequently leading to framework collapse after solvent removal [
25,
26]. To address these limitations, distinct strategies have been developed to reinforce their inherent structural fragility. A widely adopted approach involves reinforcing hydrogen-bonded networks through π–π stacking interactions between aromatic moieties. By combining π–π interactions with directional hydrogen bonds, 2D layers can be vertically stacked into 3D architectures while preserving 1D channels, thereby enhancing thermal and chemical stability [
27]. However, the relatively low directionality of π–π stacking interactions compared to stronger bonds complicates controlling and predicting the assembly of highly stable HOFs [
28].
Catenation in porous materials, referring to the interlocking of two or more distinct networks [
29,
30], has been widely employed in porous materials to improve structural stability [
31]. Although this topological characteristic strengthens mechanical performance in porous frameworks [
29–
31], excessive interpenetration leads to reduced surface accessibility and heightened risks of pore collapse [
24]. Recent advances address these limitations through controlled topological engineering. For example, Wang et al. [
32] developed a dual-walled framework design strategy to achieve highly porous and robust hydrogen-bonded organic frameworks (ZJU-HOF-5a) with exceptional H
2 storage capacity. This framework exhibits minimized twofold interpenetration within its dual-walled architecture, where supramolecular interactions stabilize interconnected networks. Such structural engineering synergistically enhances robustness while retaining high porosity.
Further refining this concept, Zhang et al. [
33] developed hydrogen-bond-guided catenation in RP-H100 and RP-H101 frameworks, which achieve higher Brunauer-Emmett-Teller (BET) surface area and balanced mass-volume metrics via employing point-contact hydrogen bonding by optimizing pore dimensions through point-contact hydrogen bonding (Fig. 1). The accessible surface area of porous materials is significantly influenced by catenation, which can reduce the GSA due to surface overlap resulting from interpenetration. In catenated structures, primary surfaces (blue) align parallel to the normal direction, while secondary surfaces (red) are oriented perpendicularly. Overlap-induced interfacial area reduction becomes pronounced when primary surface dimensions (wp) significantly exceed those of secondary counterparts (ws). However, this loss can be mitigated if wp is much smaller than ws. As wp approaches zero, the overlap diminishes, preserving GSA while enhancing the VSA owing to a higher density of framework atoms. Thus, increased catenation in superstructures composed of multiple components can further augment VSA.
The geometry of catenated components plays a crucial role in minimizing surface loss within molecular crystals. To minimize surface loss (wp ws), triptycene was chosen as a molecular skeleton due to its paddlewheel-like structure that integrates both porosity and rigidity. The molecular design strategically integrates dual hydrogen-bonding motifs—terminal carboxyl functionalities and imidazole aromatic systems—enabling the construction of building units IATH-1 and IATH-2. Structural expansion from IATH-1 to IATH-2 was achieved through acetylene linkage engineering. Crystalline growth protocols involving dimethylformamide (DMF) dissolution followed by thermal processing yield high-porosity architectures RP-H100 and RP-H101, with both frameworks demonstrating precisely controlled lattice configurations.
The supramolecular architecture of RP-H100 consists of a hydrogen-bonded network where 9 IATH-1 building blocks assemble into a hexagonally symmetric structure, exhibiting a characteristic diameter of 5.6 nanometers (without the consideration of interpenetration) (Fig. 2). RP-H101 is isostructural with RP-H100, and both frameworks exhibit interconnected hexagonal motifs that create a two-dimensional network, extending into a three-dimensional honeycomb structure. This configuration features three open channels that facilitate catenation, allowing for the interpenetration of molecular building blocks via hydrogen bonding, which minimizes surface loss by favoring point-contact interactions over extensive surface contacts.
Structural interlocking between dual supramolecular frameworks (depicted in blue and green) facilitates the formation of accessible pore networks and complex catenated architectures, with each hexagonal unit exhibiting 6-fold topological connectivity to its neighboring counterparts (Fig. 3). Hepta-fold topological interlocking emerges through continuous propagation in the x‒y plane, achieving unbroken three-dimensional continuity. Six additional layered components, labeled I to VI and distinguished by colors, interlock with the blue framework. All intermolecular interactions are mediated by hydrogen bonds, which also contribute to minimizing surface loss.
Catenation also influences pore geometries and size. Specifically, 7-fold catenation divides the pores into hierarchical structures: in RP-H100, this results in triangular prismatic cavities (1.2 nm) and one-dimensional channels (1.7 nm), while RP-H101 exhibits slightly larger pores (1.5 and 1.8 nm) under the same hierarchical configuration. These dimensional parameters align with the ideal hydrogen storage capacity range for MOFs, as predicted by previous simulation-based studies of MOFs.
The 7-fold interpenetrated architecture endows RP-H100 and RP-H101 with enhanced structural integrity. Experimental powder X-ray diffraction (PXRD) profiles align precisely with computational simulations derived from their single-crystal superstructures, validating structural preservation during desolvation. Both frameworks demonstrate strong solvent stability, resisting common organic solvents like ethanol, acetone, and DMF, with PXRD patterns remaining intact after 24 h of soaking. Variable-temperature PXRD (VT-PXRD) shows that RP-H100 maintains its crystallinity up to 375 °C, while RP-H101, despite being more porous, exhibits similar thermal stability. Thermogravimetric analyses (TGA) indicate framework decomposition initiates near 380 °C, evidenced by progressive mass loss. Collectively, RP-H100 and RP-H101 achieve thermal resilience benchmarks comparable to the most robust MOFs.
Permanent porosities of RP-H100 and RP-H101 were confirmed by N2 adsorption isotherms at 77 K, showing typical type I sorption behavior indicative of their microporous nature. The Brunauer–Emmett–Teller (BET) surface areas calculated for RP-H100 and RP-H101 are 2383 and 3526 m2/g, respectively, closely matching the theoretical values of 2034 and 3345 m2/g. Their volumetric BET areas are 1573 and 1855 m2/cm3. Pore-size distribution analyses indicate diameters of approximately 1.2 and 1.7 nm for RP-H100 and approximately 1.6 and 1.9 nm for RP-H101, corresponding to triangular prismatic pores and hexagonal channels. Total pore volumes reach 1.11 cm3/g for RP-H100 and 1.35 cm3/g for RP-H101, again aligning well with simulated values. Both materials exhibit high thermal stability and large surface areas due to hydrogen-bond-reinforced catenation, which aids in the formation of robust supramolecular frameworks.
Notably, RP-H101 has a high packing density of 0.526 g/cm3 and the highest experimentally measured GSA for HOFs at 3526 m2/g, aligning well with theoretical predictions in surface optimization for hydrogen storage applications.
RP-H101 demonstrates synchronized gravimetric (9.3 ± 0.2 wt%) and volumetric (53.7 ± 1.0 g/L) hydrogen capacities at the material level under cryogenic operating conditions. These values surpass the US Department of Energy (DOE)’s system-scale gravimetric threshold (6.5 wt%) by 43%, and exceed the volumetric benchmark (50 g/L) by 7.4%. Notably, the material-level storage metrics achieved by RP-H101 demonstrate dual superiority over the system-level standards, even at low-temperature operation constraints.
Efficient hydrogen storage necessitates concurrent optimization of volumetric and gravimetric capacities in advanced adsorbents. This study establishes supramolecular crystals as viable platforms for vehicular hydrogen storage, demonstrating how directional catenation enables precise engineering of mechanically durable frameworks. The strategy achieves synergistic integration of considerable mass-volume metrics, validating its multifunctional applicability across energy storage domains.
To contextualize the performance of RP-H101, hydrogen storage data were compiled from representative porous materials tested under pressure-temperature swing and ambient-temperature conditions. These comparisons are visualized in Fig. 4, which respectively plot volumetric hydrogen uptake (g/L) against BET surface area (m2/g) under each regime.
As shown in Fig. 4(a), RP-H101 demonstrates a remarkable balance between surface area and volumetric capacity, standing out as the only material to exceed volumetric target of the US DOE (50 g/L) under swing conditions (77 K/100 bar to 160 K/5 bar). In contrast, Fig. 4(b) presents data collected at ambient temperature, highlighting the fundamentally different operating mechanisms and performance gaps under less favorable conditions.
These comparisons underscore the unique advantage of the catenation-enhanced HOF architecture, which minimizes surface area loss while preserving framework density, thereby enabling simultaneous optimization of gravimetric and volumetric hydrogen storage under practical swing conditions [
5,
6,
16,
34,
35].
Collectively, RP-H100 and RP-H101 emerge as structurally optimized frameworks, where a 7-fold catenated superstructure directed by hydrogen-bonded point-contact interactions yields high volumetric and gravimetric surface areas, along with finely tailored pore diameters (approximately 1.2‒1.9 nm). Their exceptional thermal stability, balanced mass-volume metrics, and mechanical robustness position them as advanced material for hydrogen storage. In practical applications, RP-H101 stands out with balanced mass-volume metrics of 9.3 wt% and 53.7 g/L, alongside thermal resilience up to 375 °C, emerge as promising candidates for onboard hydrogen storage in fuel cell vehicles. Its lightweight nature and moisture resistance hold promise for meeting the requirements of fuel cell vehicles in future applications [
36,
37].
However, its long-term cyclic stability under practical conditions remains to be fully evaluated. Future research should harness machine learning (ML) approaches to enhance development efficiency, optimize structural design, and explore performance limits [
38]. Additionally, emerging large language models (LLMs) offer significant potential in advancing hydrogen storage materials by facilitating complex information parsing, integrating domain-specific knowledge, and supporting human-like reasoning. These capabilities are expected to significantly accelerate the discovery of high-performance hydrogen storage materials and promote their industrial applications.
PDF
(2424KB)
