High-Capacityand Reversible Hydrogen Storage in an Intrinsic Li₃B₂N₂ Monolayer

Abstract

Hydrogen is widely considered a promising clean energy carrier because of its high energy density and environmental benignity, yet the development of safe and reversible hydrogen storage materials remains a major challenge. Two-dimensional materials are particularly attractive for this purpose owing to their large specific surface area, fully exposed active sites, and highly tunable electronic structures. Here, using crystal structure prediction combined with first-principles calculations, we predict a stable metallic Li₃B₂N₂ monolayer as a potential hydrogen storage material. This monolayer can adsorb up to six H₂ molecules per unit cell with an average adsorption energy of ∼0.23 eV/H₂, yielding a high hydrogen storage capacity of ∼7.8 wt.%. Further analysis reveals that hydrogen adsorption is governed by the synergistic effects of electrostatic polarization and orbital hybridization. Moreover, calculations on the temperature- and pressure-dependent hydrogen storage behavior show that all hydrogen-adsorbed structures remain stable at room temperature under a pressure of 3.7 MPa. The van’t Hoff analysis indicates that the maximum desorption temperature at atmospheric pressure is 316 K, suggesting favorable reversibility under near-ambient conditions. These results establish Li₃B₂N₂ as a promising intrinsic two-dimensional material for high-density and reversible hydrogen storage.

1. Introduction

As fossil fuel resources continue to be depleted and environmental concerns intensify [1,2], hydrogen has emerged as a promising clean energy carrier owing to its abundance, high energy density, and zero-carbon emissions [3]. However, the efficient storage and safe transportation of hydrogen remain critical challenges for the realization of a hydrogen economy. Traditional hydrogen storage methods, such as high-pressure gaseous storage [4] and cryogenic liquid storage [5], require stringent safety measures and are often accompanied by high energy consumption and serious boil-off losses, which greatly hinder the large-scale development and application of hydrogen energy. In this context, solid-state hydrogen storage, which relies on the physical or chemical adsorption of hydrogen within materials, offers distinct advantages in terms of safety, volumetric density, and reversibility, and is therefore regarded as a promising strategy for future hydrogen storage.

For practical applications, solid-state hydrogen storage materials are expected to satisfy the U.S. Department of Energy (DOE) target of 5.5 wt.% for hydrogen storage capacity [6], together with the optimal adsorption energy of 0.2–0.6 eV/H₂ proposed by Kim et al. [7], so as to enable reversible hydrogen adsorption and desorption under near-ambient conditions. The emergence of two-dimensional (2D) materials, such as graphene [8], hexagonal boron nitride (h-BN) [9], MoS₂ [10], and MXenes [11,12,13], has stimulated extensive research interest because of their large specific surface areas, excellent chemical stability, and low weight, all of which are favorable for hydrogen storage. However, pristine 2D materials generally exhibit only weak physisorption toward hydrogen molecules, leading to limited storage capacities and poor reversibility under ambient conditions [14,15].

To improve hydrogen storage performance, external metal decoration has been widely adopted as an effective strategy because decorated metal atoms can generate an internal electric field that enhances the interaction with H₂ molecules [16,17,18,19,20,21]. Among various metal species, lithium is particularly attractive because its low atomic mass is beneficial for achieving high gravimetric hydrogen storage capacity. For instance, Yong et al. [22] reported that Li decoration increases the hydrogen storage capacity of graphene from 0.4 wt.% to 12.8 wt.%, far exceeding the DOE target [6], which attributed to Li-induced polarization-enhanced H₂ adsorption. Previous theoretical studies have also shown that metal-functionalized 2D materials, including Li@Irida-graphene [23], Li@$\gamma$-graphdiyne [24], Sc@2D polyaramid [25], Ti/Zr@2D polyaramid [26], and Li-functionalized ${\mathrm{B}}_x$${\mathrm{C}}_y$${\mathrm{N}}_z$ compound [27,28,29,30,31,32,33], possess desirable adsorption energies (0.18–0.35 eV), high hydrogen storage capacities (7.06–14.66 wt.%), and favorable desorption behavior near room temperature. In addition, the hydrogen storage performance of the MXene Hf₂CF₂ [34] can be significantly enhanced by Li decoration, yielding a capacity of 10.8 wt.% through a charge-transfer-induced polarization mechanism.

Despite these advantages, externally decorated metal atoms often suffer from aggregation or clustering on 2D substrates, which can severely deteriorate the reversibility and cycling stability of hydrogen storage [35,36,37,38,39,40]. Therefore, identifying intrinsically high-performance 2D hydrogen storage materials that contain built-in active adsorption sites, rather than relying on post-synthetic metal decoration, remains an active and important topic in this field [41].

In this work, we predict a thermodynamically stable Li₃B₂N₂ monolayer by combining crystal structure prediction with first-principles calculations. Unlike conventional metal-decorated 2D systems, the Li atoms in Li₃B₂N₂ are intrinsic structural components, which provide built-in Li-rich adsorption sites for H₂ molecules. The adsorption energies of H₂ molecules on Li₃B₂N₂ range from 0.20 to 0.37 eV/H₂, well within the optimal energy window for reversible hydrogen storage. The Li₃B₂N₂ monolayer can accommodate up to 18 H₂ molecules in a $1\times 3\times 1$ supercell, corresponding to a hydrogen storage capacity of 7.8 wt.%, which exceeds the DOE target. Based on the van’t Hoff equation, the average dehydrogenation temperature of Li₃B₂N₂ is estimated to be 289 K at atmospheric pressure. These appealing properties establish the Li₃B₂N₂ monolayer as a promising candidate for reversible hydrogen storage applications.

2. Computational Methods

The global structure prediction was performed using the Crystal structure AnaLYsis by Particle Swarm Optimization (CALYPSO) code [42,43,44,45]. Structural and electronic properties were calculated within the framework of density functional theory (DFT) using the Vienna Ab initio Simulation Package (VASP) [46,47]. The exchange-correlation functional was treated within the generalized gradient approximation (GGA) using the Perdew–Burke–Ernzerhof (PBE) functional [48], and the projector augmented-wave (PAW) method [49] was employed to describe the ion–electron interaction. A plane-wave cutoff energy of 700 eV was used throughout all calculations, and the Brillouin zone was sampled with a Monkhorst–Pack grid spacing of $2\pi \times 0.03$ ${\AA }^{-1}$ [50]. All structures were fully relaxed until the total energy and atomic forces converged to within $1\times {10}^{-6}$ eV and $1\times {10}^{-3}$ eV/Å, respectively. Van der Waals interactions were taken into account using the DFT-D2 correction method [51]. Additional DFT-D3 [52] and DFT-D4 [53] calculations were performed to examine the influence of different dispersion-correction schemes on the adsorption energies, as shown in the Supplementary Materials. Phonon spectra were calculated using the PHONOPY code [54]. To examine thermal stability, ab initio molecular dynamics (AIMD) simulations were carried out in the canonical ensemble using a Nos’e–Hoover thermostat [55,56] at 300 K for 10 ps with a time step of 1 fs.

3. Results and Discussion

3.1. Crystal Structure and Stability of Li₂B₂N₂ and Li₃B₂N₂ Monolayers

Figure 1a–c present the optimized crystal structures of Li₂B₂N₂, Li₃B₂N₂, and Li₄B₂N₂ monolayers, as identified from global structure searches. Li₂B₂N₂ and Li₄B₂N₂ crystallize in the C2/m space group, whereas Li₃B₂N₂ adopts the Cm space group. All three monolayers exhibit a similar layered structural motif, in which B and N atoms form a central framework composed of two hexagonal-ring chains, while Li atoms are distributed on the outer layers. In this framework, the B–B bond lengths (1.73–1.75 Å) are notably longer than the B–N bond lengths (1.47–1.50 Å). For Li₂B₂N₂, the Li atoms bridge the two hexagonal-ring chains in an interleaved manner, giving rise to a stable monolayer with a thickness of 2.05 Å. Each Li atom is threefold coordinated by N atoms, with an average Li–N bond length of 2.01 Å. With increasing Li content, the hexagonal B–N framework remains nearly unchanged, while the additional Li atoms occupy the hollow sites located directly above and below the centers of the B₄N₂ hexagonal rings. In these positions, each added Li atom is twofold coordinated by N atoms, with an average Li–N bond length of 2.26 Å. As a result, the thickness of the ${\mathrm{Li}}_n$B₂N₂ ($n=2$–4) monolayers increases monotonically with Li content, reaching 2.51 Å for Li₃B₂N₂ and 2.70 Å for Li₄B₂N₂. Meanwhile, the progressively increased number of exposed Li sites is expected to be favorable for hydrogen adsorption.

However, excessive Li incorporation reduces the structural stability. As shown in Figure 1d–f, the phonon spectra indicate that Li₂B₂N₂ and Li₃B₂N₂ monolayers are dynamically stable, as no imaginary phonon modes appear throughout the Brillouin zone, whereas Li₄B₂N₂ exhibits pronounced imaginary modes, indicating its dynamical instability. This instability is likely associated with the excessive incorporation of Li atoms. Specifically, the additional Li atoms occupy hollow sites with lower coordination numbers and longer Li–N bond lengths, indicating weaker binding to the B–N framework. Together with the enhanced electrostatic repulsion among Li cations and the increased out-of-plane structural expansion, these factors may soften low-frequency phonon modes and eventually drive the dynamical instability of Li₄B₂N₂.

Next, the thermal stability of Li₂B₂N₂ and Li₃B₂N₂ monolayers was further examined by ab initio molecular dynamics (AIMD) simulations, as shown in Figure S1. It is found that both the total energy and temperature exhibit only small fluctuations throughout the 10 ps simulation at 300 K. Moreover, no obvious atomic displacement or structural degradation is observed. These results confirm that both Li₂B₂N₂ and Li₃B₂N₂ monolayers possess good thermal stability.

To further evaluate the experimental feasibility of Li₂B₂N₂ and Li₃B₂N₂ monolayers, we calculated their cohesive energies ($E_{\mathrm{coh}}$), which are defined as the energy differences between the monolayers and their constituent isolated Li, B, and N atoms. The cohesive energy is expressedas

$$E_{\mathrm{coh}}={\displaystyle \frac{E_{{\mathrm{Li}}_n{\mathrm{B}}_2{\mathrm{N}}_2}-n{E}_{\mathrm{Li}}-2E_{\mathrm{B}}-2E_{\mathrm{N}}}{n+4}},$$

(1)

where $E_{{\mathrm{Li}}_n{\mathrm{B}}_2{\mathrm{N}}_2}$ denotes the total energy of the ${\mathrm{Li}}_n$B₂N₂ monolayer, $E_{\mathrm{Li}}$, $E_{\mathrm{B}}$, and $E_{\mathrm{N}}$ are the energies of isolated Li, B, and N atoms, respectively, and n is the number of Li atoms. The calculated cohesive energies of Li₂B₂N₂ and Li₃B₂N₂ are $-6.22$ and $-5.59$ eV/atom, respectively. These values are comparable to those of experimentally synthesized two-dimensional materials, such as borophene ($-5.99$ eV/atom) [57], silicene ($-4.57$ eV/atom) [58], h-BN ($-7.91$ eV/atom) [59], and MoS₂ ($-6.25$ eV/atom) [60], indicating their favorable energetic stability and potential experimental feasibility.

3.2. Electronic Properties of Li₂B₂N₂ and Li₃B₂N₂ Monolayers

The electronic properties of Li₂B₂N₂ and Li₃B₂N₂ monolayers are elucidated by the band structures, partial density of states (PDOS), and band-decomposed charge densities shown in Figure 2. As can be seen, Li₂B₂N₂ is a semiconductor with a small band gap of 0.72 eV, whereas Li₃B₂N₂ exhibits intrinsic metallic behavior, with two bands highlighted in orange crossing the Fermi level ($E_{\mathrm{F}}$). The different electronic characters of Li₂B₂N₂ and Li₃B₂N₂ mainly originate from their distinct electron filling. In Li₂B₂N₂, charge transfer from Li to the B–N framework is insufficient to occupy the conduction-band states, leading to semiconducting behavior. In contrast, the additional Li atom in Li₃B₂N₂ provides extra electrons, which partially occupy the B-related ${\pi }^{*}$ states associated with the B–B bonds [Figure 2b,d]. As a result, the Fermi level is shifted upward into the conduction band, giving rise to intrinsic metallic behavior. The PDOS of both systems shows significant overlapping features among Li, B, and N states, indicating appreciable orbital hybridization and strong interatomic interactions. Given its intrinsic metallicity with a Li-rich surface and stronger local surface electrostatic field, both of which are favorable for polarizing and adsorbing H₂ molecules, Li₃B₂N₂ is selected for the following investigation of hydrogen storage performance.

3.3. Hydrogen Adsorption Behavior of the Li₃B₂N₂ Monolayer

We next examine the adsorption behavior of a single H₂ molecule on the Li₃B₂N₂ monolayer. As shown in Figure 3a, seven possible adsorption sites are considered, including three top sites above Li atoms (T1, T2, and T3), three bridge sites above the midpoints of Li–Li bonds (B1, B2, and B3), and one hollow site above the center of the triangle (H). These sites cover the main high-symmetry adsorption positions on the monolayer surface. The optimized structures of H₂@Li₃B₂N₂ are shown in Figure S2.

The hydrogen adsorption energy, $E_{\mathrm{ad}}$, is calculated as

$$E_{\mathrm{ad}}={\displaystyle \frac{E_{n{\mathrm{H}}_2@{\mathrm{Li}}_3{\mathrm{B}}_2{\mathrm{N}}_2}-E_{{\mathrm{Li}}_3{\mathrm{B}}_2{\mathrm{N}}_2}-n{E}_{{\mathrm{H}}_2}}{n}},$$

(2)

where $E_{n{\mathrm{H}}_2@{\mathrm{Li}}_3{\mathrm{B}}_2{\mathrm{N}}_2}$, $E_{{\mathrm{Li}}_3{\mathrm{B}}_2{\mathrm{N}}_2}$, and $E_{{\mathrm{H}}_2}$ denote the total energies of nH₂@Li₃B₂N₂ system, pristine Li₃B₂N₂ monolayer, and an isolated H₂ molecule, respectively, and n is the number of adsorbed H₂ molecules.

The calculated absolute value of adsorption energies $|E_{\mathrm{ad}}|$ for a single H₂ molecule at these sites are summarized in Figure 3b. All adsorption configurations are energetically favorable, with adsorption energies $|E_{\mathrm{ad}}|$ ranging from 0.20 to 0.37 eV/H₂. Among them, the T1 site gives the highest adsorption energy of 0.37 eV/H₂, whereas the T3 site shows the weakest adsorption, with an adsorption energy of 0.20 eV/H₂. The B1, B3, and H sites exhibit adsorption energies of 0.28, 0.30, and 0.29 eV/H₂, respectively, while the T2 and B2 sites give slightly smaller values of 0.24 and 0.22 eV/H₂. Notably, all these values fall within or very close to the desirable adsorption-energy window of 0.2–0.6 eV/H₂ for reversible hydrogen storage under ambient conditions, as suggested by the U.S. Department of Energy (DOE) [6]. Although the T1 site shows a slightly higher adsorption energy, adsorption at this site induces noticeable structural distortion (Figure S2). Therefore, the B3 site is identified as the most favorable adsorption configuration, with an adsorption energy of 0.30 eV/H₂ without inducing obvious structural distortion, which is favorable for reversible hydrogen storage under near-ambient conditions.

To further understand the adsorption mechanism, we analyze the projected density of states (PDOS) of the B3 adsorption configuration, as shown in Figure 3c. Clear overlap between the H-s state and the Li-$s,p$ states can be observed in several energy regions, indicating noticeable interaction between the adsorbed H₂ molecule and the Li sites. This result suggests that H₂ adsorption on Li₃B₂N₂ is not governed solely by van der Waals interaction, but also involves weak orbital hybridization. The positively charged Li atoms on the surface create a strong local electrostatic field, which can polarize the H₂ molecule. At the same time, weak orbital hybridization also contributes to the interaction. Therefore, the adsorption of H₂ arises from the combined effects of electrostatic polarization and orbital hybridization.

3.4. Hydrogen Storage Capacity of the Li₃B₂N₂ Monolayer

To further evaluate the hydrogen storage capacity of the Li₃B₂N₂ monolayer, we gradually increased the number of adsorbed H₂ molecules on the $1\times 3\times 1$ supercell. A $1\times 3\times 1$ supercell was adopted in the subsequent calculations, because the supercell-size convergence test shows that the adsorption energy is already well converged at this size. As shown in Figure S3, the adsorption energy remains nearly unchanged when the supercell is further enlarged from $1\times 3\times 1$ to $1\times 5\times 1$, indicating that the interaction between periodically repeated H₂ molecules is negligible. Therefore, the $1\times 3\times 1$ supercell provides a reliable and computationally efficient model.

As shown in Figure 4a, the average adsorption energy $|E_{\mathrm{ad}}|$ remains within the favorable range when the number of adsorbed H₂ molecules increases from 3 to 18, with values of 0.23–0.26 eV/H₂. This indicates that the Li₃B₂N₂ monolayer can maintain moderate and reversible interactions with multiple H₂ molecules. When the number of H₂ molecules further increases to 24, the average adsorption energy decreases to 0.19 eV/H₂, slightly below the lower bound of the desirable adsorption-energy window. The optimized structures in Figure 4b,c further show that all 18 H₂ molecules remain stably adsorbed, whereas six H₂ molecules (highlighted in orange) desorb from the surface in the 24 H₂ configuration. This behavior can be attributed to the steric hindrance and intermolecular repulsion between adjacent H₂ molecules at high coverage. The optimized structures of the remaining H₂ adsorption configurations are shown in Figure S4. These results indicate that the maximum stable adsorption capacity of the Li₃B₂N₂ monolayer is 18 H₂ molecules per $1\times 3\times 1$ supercell.

In 18H₂@Li₁₈B₁₂N₁₂ structure, the adsorbed H₂ molecules retain H–H bond lengths of 0.757–0.812 Å, close to the isolated-H₂ value of 0.750 Å. Bader charge analysis shows that each adsorbed H₂ molecule obtained 0.155 e, consistent with the charge density difference map shown in Figure S5a, suggesting a weak interaction between H₂ molecule and substrate. The reduced density gradient (RDG) analysis [shown in Figure S5b] further confirm the weak interaction (electrostatic polarization and weak van der Waals interactions) dominated molecular adsorption of H₂.

The hydrogen storage capacity was calculated as

$$G_C(\mathrm{wt}.\%)={\displaystyle \frac{n_{{\mathrm{H}}_2}{w}_{{\mathrm{H}}_2}}{n_{\mathrm{Li}}{w}_{\mathrm{Li}}+n_{\mathrm{B}}{w}_{\mathrm{B}}+n_{\mathrm{N}}{w}_{\mathrm{N}}+n_{{\mathrm{H}}_2}{w}_{{\mathrm{H}}_2}}}\times 100\%,$$

(3)

where $n_{{\mathrm{H}}_2}$, $n_{\mathrm{Li}}$, $n_{\mathrm{B}}$, and $n_{\mathrm{N}}$ are the numbers of H₂ molecules, Li, B, and N atoms, respectively, while $w_{{\mathrm{H}}_2}$, $w_{\mathrm{Li}}$, $w_{\mathrm{B}}$, and $w_{\mathrm{N}}$ denote the corresponding molecular or atomic weights. Therefore, 18 H₂ molecules adsorbed on the $1\times 3\times 1$ supercell represent the maximum stable adsorption capacity of the Li₃B₂N₂ monolayer, corresponding to a gravimetric hydrogen storage capacity of 7.8 wt.%.

As summarized in

Table 1. Summary of the hydrogen adsorption Energies ($\left|E_{\mathrm{ad}}\right|$), hydrogen storage capacities, and desorption temperature ($T_D$) of common 2D materials and Li decorated 2D materials.

Compound	$\left|{\mathit{E}}_{\mathbf{ad}}\right|$ [eV/H2]	Storage Capacity [wt.%]	${\mathit{T}}_{\mathit{D}}$	Refs.
Li3B2N2	0.23	7.8	261–316 K/0.1 MPa	This work
graphene	…	0.4	77 K/100 kPa	[8]
h-BN	…	2.96	243 K/10 MPa	[9]
Li@graphene	0.56	12.8	…	[22]
Li@Irida-graphene	0.23–0.28	7.06	353 K/0.1 MPa	[23]
Li-doped $\gamma$-graphdiyne	0.21–0.23	14.66	…	[24]
Li@T-BN	0.25–0.32	12.31	180–232.6 K/1 atm	[28]
Li@B2N	0.19–0.27	11.1	<200 K/0.1 MPa	[30]
Li@penta-BCN	0.16	7.44	…	[31]
Li@BC2N	0.18–0.30	11.10	360 K/1 atm	[32]
Li@POG-B4C2N3	0.19–0.35	8.35	…	[33]
Li@AsC5	∼0.19	9.7	243–357 K/1 atm	[61]
Li@GeC5	0.22	7.62	281.1 K/1 atm	[62]

, we compare the adsorption energy and gravimetric hydrogen storage capacity of Li₃B₂N₂ with those of representative two-dimensional hydrogen storage materials. Pristine 2D materials, such as graphene and h-BN, generally exhibit weak H₂ adsorption energy and limited storage capacity, while metal-decorated systems, such as Li@Irida-graphene, Li@BC₂N, and Li@AsC₅, show improved hydrogen-storage performance. Li₃B₂N₂ exhibits a competitive hydrogen storage capacity of 7.8 wt.% with a moderate adsorption energy of 0.23 eV/H₂, comparable to these recently reported metal-decorated 2D materials. Notably, this performance on Li₃B₂N₂ is achieved by intrinsic Li-rich adsorption sites without external metal decoration, underscoring Li₃B₂N₂ as a promising metal-decoration-free 2D hydrogen storage material.

3.5. Desorption Capacity of the H₂ Molecule on the Li₃B₂N₂ Monolayer

Beyond achieving a high hydrogen storage capacity, an onboard hydrogen storage system must also remain stable under practical operating conditions, namely, it should be able to reversibly adsorb and release hydrogen at room temperature and moderate pressures (0.2∼12 MPa). Therefore, the effects of temperature and pressure on the structural stability were evaluated based on the relative energy $E_r$, defined as

$$E_r=E_{n{\mathrm{H}}_2@{\mathrm{Li}}_3{\mathrm{B}}_2{\mathrm{N}}_2}-E_{{\mathrm{Li}}_3{\mathrm{B}}_2{\mathrm{N}}_2}-n[E_{{\mathrm{H}}_2}+{\mu }_{{\mathrm{H}}_2}(T,P)],$$

(4)

where $E_{n{\mathrm{H}}_2@{\mathrm{Li}}_3{\mathrm{B}}_2{\mathrm{N}}_2}$, $E_{{\mathrm{Li}}_3{\mathrm{B}}_2{\mathrm{N}}_2}$, $E_{{\mathrm{H}}_2}$, and n have the same meanings as those in Equation (2). ${\mu }_{{\mathrm{H}}_2}(T,P)$ is the chemical potential of hydrogen at temperature T and pressure P, which can be expressed as

$${\mu }_{{\mathrm{H}}_2}(T,P)=\Delta H-T\Delta S+k_{\mathrm{B}}Tln\left({\displaystyle \frac{P}{P^0}}\right),$$

(5)

where $P^0$ is the standard pressure of 0.1 MPa, $k_{\mathrm{B}}$ is the Boltzmann constant, and $\Delta H\left(T\right)$ and $\Delta S\left(T\right)$ are the enthalpy and entropy changes of gas-phase H₂ from 0 K to the target temperature at ambient pressure. The values of $\Delta H-T\Delta S$ were taken from the thermochemical tables [63] and are listed in

Table 2. Values of $\Delta H-T\Delta S$ obtained from the thermochemical tables.

T [K]	$\mathbf{\Delta }\mathit{S}\left(\mathit{T}\right)$ [J/(mol · K)]	$\mathbf{\Delta }\mathit{H}\left(\mathit{T}\right)$ [kJ/mol]	$\mathbf{\Delta }\mathit{H}\left(\mathit{T}\right)-\mathit{T}\mathbf{\Delta }\mathit{S}\left(\mathit{T}\right)$ [eV/H2]
0	0	0	0
100	100.73	3.00	−0.073
200	119.41	5.69	−0.189
300	130.86	8.52	−0.319
400	139.216	11.43	−0.459

.

Figure 5a shows the relative energies ($E_r$) of 6H₂@Li₁₈B₁₂N₁₂ and 18H₂@Li₁₈B₁₂N₁₂ as a function of temperature at standard atmospheric pressure. At 0 K, both 6H₂@Li₁₈B₁₂N₁₂ and 18H₂@Li₁₈B₁₂N₁₂ have negative relative energies, indicating that these adsorption structures are thermodynamically stable. As the temperature increases, the relative energies gradually increase and reach zero at about 245 K for 6H₂@Li₁₈B₁₂N₁₂ and 229 K for 18H₂@Li₁₈B₁₂N₁₂. Above these temperatures, the adsorbed H₂ molecules become thermodynamically unfavorable and tend to desorb spontaneously. Figure 5b presents the relative energies of 6H₂@Li₁₈B₁₂N₁₂ and 18H₂@Li₁₈B₁₂N₁₂ as a function of pressure at 300 K. The minimum equilibrium pressures are estimated to be about 1.6 MPa for 6H₂@Li₁₈B₁₂N₁₂ and 3.7 MPa for 18H₂@Li₁₈B₁₂N₁₂. These results indicate that the Li₃B₂N₂ monolayer can achieve reversible hydrogen storage through external pressure control, that is, hydrogen adsorption can be realized at pressures above 3.7 MPa.

The desorption temperature ($T_{\mathrm{D}}$) of the Li₃B₂N₂-based hydrogen storage system was further estimated using the van’t Hoff equation:

$$T_{\mathrm{D}}={\displaystyle \frac{E_{\mathrm{ad}}}{k_{\mathrm{B}}{[(\Delta S/R)-lnp]}^{-1}}}.$$

(6)

Here, $E_{\mathrm{ad}}$ is the hydrogen adsorption energy, $k_{\mathrm{B}}$ is the Boltzmann constant, $\Delta S$ is the entropy change from gaseous hydrogen to adsorbed hydrogen, which is approximately equal to the standard entropy of hydrogen gas, 130 J/(K·mol), R is the gas constant ($R=8.31$ J/(K·mol)), and p is the equilibrium pressure.

Here, the minimum and maximum desorption temperatures are approximately estimated from representative adsorption energies at different H₂ coverages. The 24 H₂ configuration, with relatively weaker adsorption at high coverage, is taken to describe the onset of H₂ desorption, whereas the 9 H₂ configuration, with stronger adsorption at lower coverage, is used to represent the final stage of hydrogen release. The 18 H₂ configuration, corresponding to the maximum stable loading, is adopted to evaluate the average desorption temperature. The pressure-dependent desorption-temperature curves at the minimum, average, and maximum $T_{\mathrm{D}}$ values are shown in Figure 5c. At standard atmospheric pressure (∼0.1 MPa), the maximum, average, and minimum desorption temperatures are 316, 289, and 261 K, respectively. Increasing the pressure further raises the desorption temperature. Furthermore, the H₂ occupation analysis for 18H₂@Li₁₈B₁₂N₁₂ based on a simplified grand-canonical thermodynamic model [25,26,64,65], presented in the Supplementary Materials, shows a consistent temperature- and pressure-dependent trend. Therefore, these thermodynamic results show that H₂ adsorption and desorption on Li₃B₂N₂ can be reversibly controlled by temperature and pressure. The predicted desorption near-room-temperature range of 261–316 K at 0.1 MPa is highly favorable for fuel-cell applications, highlighting Li₃B₂N₂ as a promising candidate for reversible hydrogen storage under near-ambient conditions.

4. Conclusions

In summary, based on crystal structure prediction and first-principles calculations, we have identified Li₃B₂N₂ as a promising intrinsically Li-rich two-dimensional material for reversible hydrogen storage. Structural analysis shows that Li₂B₂N₂ and Li₃B₂N₂ monolayers are both dynamically and thermally stable, whereas excessive Li incorporation leads to the dynamical instability of Li₄B₂N₂. Electronic structure calculations reveal that Li₃B₂N₂ exhibits intrinsic metallic behavior and an enhanced local surface electrostatic field, both of which are favorable for hydrogen adsorption. The adsorption energies of a single H₂ molecule on Li₃B₂N₂ range from 0.20 to 0.37 eV/H₂, which fall within the desirable window for reversible hydrogen storage. The adsorption mechanism is mainly governed by the combined effects of electrostatic polarization and weak orbital hybridization. Further calculations show that the Li₃B₂N₂ monolayer can accommodate up to 18 H₂ molecules in a $1\times 3\times 1$ supercell, corresponding to a hydrogen storage capacity of 7.8 wt.%. Furthermore, thermodynamic analysis demonstrates that H₂ adsorption and desorption can be regulated by temperature and pressure, and the predicted desorption temperature range of 261–316 K at 0.1 MPa suggests favorable reversibility under near-ambient conditions. These results demonstrate that Li₃B₂N₂ is a highly promising intrinsic 2D material for high-capacity and reversible hydrogen storage, and provide a useful design strategy for developing next-generation metal-decoration-free hydrogen storage materials.