Tailored hybridization of MIL-101(Cr) with graphene oxide enables enhanced hydrogen storage and delivery

Abstract

The development of practical hydrogen (H₂) storage solutions is essential for a net-zero, secure, and affordable energy future. This study explores the effect of the hybridization method on the H₂ storage and delivery performance of nanoporous MIL-101(Cr)@GO materials, composed of micro–mesoporous Cr terephthalate MIL-101(Cr) and graphene oxide (GO). MIL-101(Cr) was chosen for its high surface area, open metal sites and stability, while GO was selected for its higher density and potential to enhance H₂ storage when combined with MIL-101(Cr). Three hybridization approaches were employed: one in situ method and two ex situ methods, namely post-synthetic modification and physical blending using resonant acoustic mixing (RAM). The impact of these methods on the structural, textural and adsorption properties of the hybrids was systematically analyzed. GO incorporation consistently reduced the surface area of all hybrids but promoted ultramicroporosity (pore width <0.7 nm). Ex situ hybrids retained textural features closer to the pristine MOF and exhibited higher gravimetric H₂ uptake than pure MIL-101(Cr) and the in situ hybrid. Notably, they achieved up to 22% higher excess H₂ uptake (wt%) at 273 K and 100 bar compared to bare MIL-101(Cr). On the other hand, the in situ hybrid, despite a lower gravimetric capacity, demonstrated a threefold increase in tapped density, resulting in a 6–7% improvement in volumetric H₂ performance: total stored and deliverable. These findings highlight the critical role of synthesis strategy in tailoring hybrid material properties to optimize H₂ storage and delivery under varying conditions.

---

1. Introduction

The urgent need to reduce fossil-fuel related emissions has positioned hydrogen (H₂) as a pivotal element in the transition to cleaner energy sources. However, widespread adoption of H₂ faces significant challenges, particularly in developing efficient, safe and cost-effective storage solutions. Current methods, such as high-pressure compression (700–1000 bar) or cryogenic liquefaction at 20 K, are energy-intensive and raise safety concerns. To address these issues, researchers are exploring H₂ storage via physisorption in porous materials. This approach, relying on weak van der Waals forces between H₂ molecules and material surfaces, operates under more practical conditions (e.g., 77 K, 100 bar), ensures reversible H₂ release, and offers rapid kinetics without chemical reactions.^1,2

Metal–organic frameworks (MOFs) have emerged as promising materials for H₂ storage by physisorption, primarily due to their tunable structures, which allow precise control over surface area, pore size distribution, and chemical functionalities (such as open metal sites), thereby enhancing H₂ adsorption efficiency. MIL-101(Cr) (MIL: Materials Institute Lavoisier) stands out for its unique combination of features: a hierarchical pore system with microporous supertetrahedra (0.7 nm), mesoporous cavities (2.9 and 3.4 nm) connected by microporous openings (1.2 and 1.6 nm), high surface area (2500–4000 m² g⁻¹), excellent thermal stability (up to 573 K), humidity resistance, and exposed Cr(III) sites (accessible under appropriate conditions) that enhance H₂ adsorption.³ These characteristics stem from its robust structure, formed by trimeric Cr³⁺ oxoclusters linked by six terephthalate ligands, with the molecular formula [Cr₃ (F/OH)(BDC)₃·(H₂O)₂]⁴ (BDC stands for 1,4 benzene dicarboxylate).

To further improve the performance of MIL-101(Cr) for H₂ storage, researchers have explored the incorporation of complementary materials that can enhance its structural and functional properties. Graphene oxide (GO) is particularly attractive for this purpose due to its excellent mechanical strength and abundant oxygen-containing functional groups, which facilitate strong interactions with the MOF framework, enhancing pore accessibility and thermal conductivity.⁵ Structurally, GO consists of two-dimensional carbon sheets decorated with oxygen-containing functional groups located at defect sites on the basal planes and along the sheet edges.⁶ The development of MIL-101(Cr)@GO hybrids is intended to address key intrinsic limitations of pure MOFs, including low thermal conductivity (<0.1 W m⁻¹ K⁻¹), mechanical fragility and low density. Recent studies reveal that gas storage capacity and overall performance are critically influenced by the hybridization synthetic approach (e.g., in situ vs. ex situ)⁷ and by the GO content.^8,9 The quality of GO sheet dispersion, often compromised by aggregation during physical mixing, directly governs pore accessibility and H₂ diffusion pathways.¹⁰ Simultaneously, the interface chemistry resulting from in situ growth methods promotes covalent bonding between MOFs and GO, significantly enhancing thermal transport properties.^5,11

Several distinct hybridization approaches have shown particular promise: the traditional routes¹² such as (i) in situ solvothermal growth, where MIL-101(Cr) crystallizes around pre-dispersed GO nanosheets; (ii) ex situ electrostatic assembly via zeta potential control for enhanced GO interaction. A more recent strategy was proposed through ex situ physical blending via resonant acoustic mixing (RAM), a novel mechanical blending technique that ensures homogeneity with minimal structural damage.¹³ Although these methods produce distinct interfacial architectures and pore systems, their relative impacts on H₂ storage performance are still underexplored. The in situ hybridization method has been the subject of extensive study exemplified by the work of Yan et al. on the development of MIL-101(Cr)@GO hybrids with different percentages of GO for water adsorption applications.¹⁴ In contrast, ex situ approaches have received comparatively less attention. Muschi et al. developed a MIL-91@GO hybrid, conducting a comparative study between in situ and ex situ methods, revealing that only the in situ route could lead to the optimal characteristics (porosity and conductivity) for CO₂ capture by microwave swing adsorption.⁷

RAM has been successfully applied in various contexts, including wet granulation processes to produce uniform and well-dispersed granules without the need for milling media, as well as in direct mechanocatalysis, where RAM induces solid-state chemical reactions through acoustic energy input, demonstrating its versatility as a gentle yet effective mixing and processing technique.^15,16 Building on this versatility, the RAM technique has also been employed to successfully synthesize MOFs like ZIF-8 or HKUST-1 without traditional milling media or solvents, highlighting its potential in processing porous frameworks.¹⁷ However, to the best of our knowledge, no studies have yet explored the synthesis of MOF hybrids using the RAM approach. Filling this knowledge gap is critical for advancing H₂ storage applications, as the synthesis method directly influences MOF@GO interface quality, MIL-101(Cr) porosity and metal sites, and hybrid stability – all key factors governing H₂ uptake capacity and kinetics. Precise control over these parameters can overcome the mass-transport barriers and structural instability of pure MOFs, while the GO network improves thermal management during H₂ charge/discharge cycles. It is, therefore, essential to establish clear correlations between the structure, processing, and properties in order to develop practical MOF-based H₂ storage systems.

This study systematically examines the impact of three synthesis strategies – in situ, ex situ electrostatic, and ex situ RAM-based mechanical mixing – on the H₂ storage and release performance of MIL-101(Cr)@GO hybrids. Through careful experimentation and physicochemical characterization, it offers practical insights to guide the development of MOF-based hybrids, advancing the development of optimized MOF-based hybrids for H₂ storage applications.

2. Experimental section

2.1. Materials

All chemicals used in this study were obtained commercially and used without further purification. Chromium nitrate nonahydrate (Cr(NO₃)₃·9H₂O, 97%) was sourced from Fluka, part of Sigma-Aldrich (Merck, Darmstadt, Germany). Terephthalic acid (H₂BDC, >98%) was purchased from Sigma-Aldrich (Merck, Darmstadt, Germany). Absolute ethanol (≥96%) was supplied by VWR Chemicals® (Radnor, Pennsylvania, USA). A 2 wt% aqueous dispersion of GO was acquired from Graphenea SA (San Sebastián, Spain).

2.2. Synthesis

2.2.1. Microwave-assisted synthesis of MIL-101(Cr) (Fig. 1a). A previously described procedure¹⁸ was followed with minor adjustments. Briefly, a mixture of chromium(III) nitrate nonahydrate (2 g; 5.0 mmol) and terephthalic acid (0.825 g; 5.0 mmol) was dissolved in 25 mL of deionized water within a 100 mL Teflon reactor. The solution was then subjected to microwave-assisted hydrothermal treatment at 473 K for 30 minutes. The resulting product was collected by centrifugation and then refluxed in 50 mL water for 20 minutes to remove residual ligand. After cooling to room temperature, the material was centrifuged at 19 000 rpm (293 K) for 20 minutes. The solid underwent additional purification via reflux in 50 mL ethanol for 20 minutes, followed by centrifugation under identical conditions. This washing step with ethanol was repeated as needed to remove all residual ligand, after which the final product was dried at 363 K in a vacuum oven.

Fig. 1 Schematic representation of the synthesis routes for (a) MIL-101(Cr) via the microwave-assisted method; (b) MIL-101(Cr)@GO hybrids via post-hybridization; (c) MIL-101(Cr)@GO hybrids via RAM; and (d) MIL-101(Cr)@GO hybrids via an in situ method.

2.2.2. Ex situ synthesis of MIL-101(Cr)@GO hybrids with post-synthetic hybridization (Fig. 1b). Before preparing MIL-101(Cr)@GO hybrids, the zeta potential of MIL-101(Cr) and GO in deionized water was measured across different pH values to determine the optimal pH at which their surface charges are most oppositely charged, facilitating electrostatic attraction (see Fig. S1). MIL-101(Cr) (1 g, kept in an oven at 100 °C to minimize free water) was dispersed in 80 mL deionized water and combined with GO dispersion (2.75 g at 2 wt%) to prepare a hybrid material containing 5 wt% GO. The mixture was agitated for 4 hours at pH 4 (adjusted with HNO₃) to facilitate electrostatic self-assembly. The hybrid material was recovered by centrifugation (19 000 rpm, 293 K) and dried at 363 K, yielding the final product labeled M101-GO POST. The synthesis parameters were selected following a slightly modified procedure reported for the post-synthetic modification of MIL-91@GO hybrids.⁷

2.2.3. Ex situ synthesis of MIL-101(Cr)@GO hybrids with RAM (Fig. 1c). GO (2.75 g at 2 wt%, to obtain a hybrid material with 5 wt% GO content) was dispersed in 10 mL deionized water under magnetic agitation. MIL-101(Cr) (1 g) was combined with the GO dispersion in a 25 mL glass tube and processed by resonant acoustic mixing (RAM), using a LabRAM mixer (Resodyn Acoustic Mixers, USA). RAM is a mechanochemical mixing technique that uses high-frequency vibrational energy to achieve homogeneous blending of materials through controlled acceleration. The mixture was processed at an acceleration of 95g (where g is the gravitational acceleration, 9.81 m s⁻²) for 15 minutes at room temperature. The hybrid material was recovered by solvent evaporation at 393 K, yielding the final product labeled M101-GO RAM. The synthesis parameters were selected following a previously reported procedure for the hybridization of HKUST-1 with activated carbon using RAM.¹³

2.2.4. In situ synthesis of MIL-101(Cr)@GO hybrids (Fig. 1d). The synthesis was carried out according to the procedure reported in ref. 19, with minor adaptations. Chromium(III) nitrate nonahydrate (0.8 g; 2 mmol), terephthalic acid (0.332 g; 2 mmol), and GO (0.87 g at 2 wt%) were dispersed in 10 mL deionized water within a 25 mL Teflon-lined autoclave. In the case of in situ synthesis, the yield of the MOF preparation must be considered when calculating the amount of GO to be introduced. To accurately determine this yield, three independent syntheses of the MOF were first carried out, and the amount of recovered material was measured. The average yield obtained was then used to accurately calculate the amount of GO needed to achieve the desired hybrid composition. Hydrothermal synthesis was conducted at 493 K for 18 hours, with three sequential reaction cycles. The product was filtered, refluxed in 50 mL deionized water for 20 minutes to remove unreacted ligand, and centrifuged (19 000 rpm, 293 K). Subsequent ethanol washing (50 mL, 20 minute reflux) and centrifugation were performed twice. The final material, labeled M101-GO in situ, was dried at 393 K.

All MIL-101(Cr)@GO hybrids contain 5 wt% of GO, a proportion selected to balance thermomechanical properties, porosity preservation and material homogeneity. Earlier research on MOF hybrids with GO indicated that higher GO contents block pores, while lower amounts do not significantly improve properties.^14,20,21

2.3. Characterization techniques

The crystallinity of the samples was analyzed by X-ray powder diffraction (XRD) with a Bruker D8 Advance diffractometer, using Cu Kα radiation (40 kV, 40 mA). Diffraction data were recorded over a 2θ range from 5° to 30°. Data interpretation and phase identification were performed using Bruker's DIFFRAC.EVA software. The XRD pattern was simulated using Mercury 3.0 software from the Cambridge Crystallographic Data Centre (CCDC), based on the CIF file of MIL-101(Cr) (deposit number 605510).²²

Fourier-transform infrared spectroscopy (FTIR) spectra were acquired in ATR mode using a Frontier 400 spectrometer (PerkinElmer, USA) covering the wavenumber range 650–4000 cm⁻¹. Powdered samples were directly pressed onto the crystal surface, with degassing at 398 K for 12 hours to remove adsorbed species. This technique was employed to identify characteristic MOF functional groups and to check the presence of important binding interactions.

Thermogravimetric analysis (TGA) was conducted with an STA 449F3 Jupiter microbalance (Netzsch). Prior to the measurement, samples were degassed at 398 K for 12 hours under secondary vacuum using a SmartVacPrep (Micromeritics-Particulate Systems®, USA). Approximately 20 mg of sample was used for analysis, which involved heating from ambient temperature to 1173 K in an argon atmosphere, with a controlled heating rate of 10 K min⁻¹. Data were processed and derived thermogravimetric (DTG) curves were analyzed with Proteus® software (Netzsch).

The samples were optically characterized using a NIKON LV100ND ECLIPSE polarized light optical microscope, equipped with five objective lenses and a 5 MPix CCD camera. This camera provided high-resolution images, ensuring precise observation and analysis of the samples. A Gemini SEM 360 microscope was used for morphological and compositional analysis through SEM imaging. The main goal was to evaluate the distribution of MOFs and GO using elemental mapping. This approach, based on Energy Dispersive X-ray Spectroscopy (EDX), facilitated the identification of surface elements.

Gas adsorption experiments at low pressure for N₂ (77 K), Ar (87 K) and H₂ (77 K) were carried out using a 3Flex manometric device (Micromeritics, USA). Prior to measurements, all the samples (100 mg) were subjected to secondary vacuum degassing for 12 hours at 398 K, using a SmartVacPrep unit (Micromeritics, USA). Textural characterization of the materials was performed by determining the Brunauer–Emmett–Teller (BET) area (A_BET, m² g⁻¹) using N₂ and Ar isotherms with Microactive® software (Micromeritics, USA), and the p/p° range for the application of the BET equation was selected by applying Rouquerol's criteria.²³ N₂ and H₂ adsorption isotherms were treated simultaneously with SAIEUS® software (Micromeritics, USA), using a two-dimensional non-local density functional theory (2D-NLDFT HS) model designed for heterogeneous carbon surfaces. Due to the non-existence of specific models for MOF@GO hybrids and MOFs in general, a kernel developed for carbon-based materials was chosen as the best approximation for the obtained isotherms. This approach has been commonly adopted in the analysis of MOF/carbon composites,^13,24 since the presence of graphene oxide introduces a carbonaceous environment that affects the adsorption characteristics. While this choice may lead to a certain degree of systematic deviation, it provides a consistent framework to compare different hybridization strategies. The application of the 2D-NLDFT-HS method to both isotherms enabled obtaining essential textural properties, including pore size distribution (PSD), 2D-NLDFT-derived surface area (S_2D-NLDFT, m² g⁻¹), total pore volume (V_tot, cm³ g⁻¹), and porosity fractions in different pore diameter ranges: micropore volume (V_µ, cm³ g⁻¹, w < 2 nm), ultramicropore volume (V_uµ, cm³ g⁻¹, w < 0.7 nm), supermicropore volume (V_sµ, cm³ g⁻¹, 0.7 nm < w < 2 nm), and mesopore volume (V_meso, cm³ g⁻¹, 2 nm < w < 50 nm). Here, w refers to the pore width or pore diameter, which is the distance between the pore walls and defines the range of pore size (ultramicropores, supermicropores and mesopores).

Water adsorption measurements were also carried out at 293 K using a BELSORP automatic adsorption analyzer (Microtrac, Japan). To eliminate residual moisture, the samples (∼100 mg) were degassed under secondary vacuum at 398 K for 12 hours using a SmartVacPrep system (Micromeritics, USA). The apparent water affinity constant was estimated from the initial slope of the isotherms at low relative pressure (p/p° < 0.02).

Finally, the thermal conductivity of the samples was analyzed using a Hot Disk TPS 2500 system. A flat sensor was placed between two layers of powder samples (approximately 2 mm each), enabling accurate thermal measurements while retaining the original shape of the material, without pelletization. Prior to testing, samples were dried under secondary vacuum at 398 K for 12 hours to remove any residual moisture. For each measurement, the sensor was heated at 50 mW for 4 seconds.

2.4. Hydrogen storage capacity

H₂ storage performance was systematically evaluated by a combination of high-pressure gas adsorption experiments and density measurements. The methodology was designed to provide both gravimetric and volumetric assessments, ensuring comprehensive characterization of storage performance under practical charge/discharge conditions.

H₂ adsorption isotherms at high pressure were recorded at 77 K, 160 K and 273 K over a pressure range of 0–130 bar using an HPVA II high-pressure manometric analyzer (Micromeritics, USA). A closed-cycle cryogenic refrigerator (Gifford–McMahon cycle) was used to control the temperature during H₂ adsorption–desorption. Samples (∼1 g) were degassed at 398 K under high vacuum (10⁻⁶–10⁻⁸ bar) for 12 hours to eliminate moisture and any pre-adsorbed gas. The software provided to control the manometric device automatically calculates excess H₂ uptake (n_exc, kg_H2 kg_ads⁻¹) of the adsorbent (ads) using the Leachman equation of state,²⁵ as recommended by NIST standards. The stability of the instrument was verified before each measurement by checking that the two temperature sensors provided consistent readings matching the experimental setpoint.

The observed n_exc at high pressures is influenced by the relative densities of the gas phase (ρ_H2) and the adsorbed phase (ρ_a). At low pressures, n_exc closely follows the absolute uptake (n_abs) because ρ_H2 < ρ_a. As pressure increases, ρ_H2 becomes comparable to ρ_a, leading to a maximum in n_exc. Beyond this point, further compression mainly increases the gas-phase density, while the adsorbed phase behaves nearly as an incompressible fluid, causing n_exc to decrease toward a saturation regime even though absolute uptake continues to grow.

The relationship between excess and absolute uptake can be expressed as

n_exc = n_abs − ρ_H2·V_a (1)

where V_a represents the volume occupied by the adsorbed H₂ phase.

The total H₂ storage capacity (n_tot, kg_H2 kg_ads⁻¹) was calculated taking into account n_exc and the H₂ compressed in the void volume, considering intra- and interparticle spaces, as follows:²⁶

where ρ_tap (kg_ads m⁻³) and ρ_skel (kg_ads m⁻³) are the tapped (i.e., packing) and the skeletal density, respectively. Accurate density measurements are essential for assessing the volumetric efficiency of porous storage materials. ρ_skel was determined by helium pycnometry with an AccuPyc II 1340 (Micromeritics, USA) after degassing samples at 398 K under vacuum for 12 hours. ρ_tap was assessed using an Autotap device (Anton Paar QuantaTec, USA) on around 4 mL samples using a 5 mL graduated cylinder and applying 1500 mechanical taps, repeated in triplicate for reproducibility. These parameters enable gravimetric adsorption data to be converted into volumetric capacity measurements, essential for materials implementation in real H₂ storage related applications. ρ_H2 (kg_H2 m⁻³) denotes the density of H₂ gas under the measurement conditions and was calculated using the Leachman equation of state for H₂,²⁵ implemented in the NIST REFPROP (Reference Fluid Thermodynamic and Transport Properties) database.

Equally important for practical applications is the release or usable capacity, which quantifies the amount of H₂ that can be effectively delivered during discharge.^27–29 This parameter was determined using eqn (2).

Release capacity = (n_tot)_storage − (n_tot)_release (3)

where the subscripts indicate the total uptake measured under storage and release conditions, respectively. This metric provides crucial insight into the material's operational efficiency, as it reflects the actual amount of usable H₂ content rather than just the maximum amount stored. In this study, release capacities were estimated by considering H₂ storage at 77 K and 100 bar, which corresponds to the maximum refueling pressure typically employed for Type I metal hydride tanks.²⁷ For release conditions, a constant discharge pressure of 5 bar was applied, in line with the minimum delivery pressure recommended by the Hydrogen Storage Engineering Center of Excellence (HSECoE),³⁰ and commonly used as the upper threshold for H₂ flow to fuel cells.^31,32 Two different release scenarios were investigated: isothermal desorption at 77 K (77 K → 77 K) and temperature swing desorption, where the material was heated to 160 K after loading at 77 K (77 K → 160 K).

3. Results and discussion

3.1. Material characterization

The ex situ hybrids have a near-quantitative production yield (∼100%) after obtaining the initial MIL-101(Cr). However, the overall yield, relative to the linker, must take into account the low yield MIL-101(Cr) synthesis step (33% via the microwave method). In comparison, the in situ hybrid route achieves a final yield relative to the linker of 49%, which is slightly higher than the effective yield of the ex situ approach (49% vs. 33% when considering the initial MOF synthesis).

The commercial GO used in this study was also carefully characterized by XRD to assess its structural features (Fig. S2a). Interestingly, the obtained XRD pattern did not match the typical profile of GO, which usually exhibits a sharp peak around 11° corresponding to the (001) plane.³³ Instead, the diffraction pattern showed a broad peak around 24°, characteristic of the (002) plane of graphitic carbon, together with a slightly elevated background at low angles (Fig. S2b). These features suggest that, although the material was purchased as GO, it is partially reduced, retaining some structural disorder typical of reduced graphene oxide (rGO) while losing a significant fraction of oxygen-containing groups.

PXRD patterns of MIL-101(Cr) hybrids synthesized by post-synthetic hybridization (MIL101-GO POST), RAM (MIL101-GO RAM) and by an in situ method (MIL101-GO in situ) are shown in Fig. 2a. The diffraction patterns of all MIL-101(Cr)@GO hybrids closely match the simulated PXRD pattern for MIL-101(Cr) and the synthesized MIL-101(Cr), with characteristic peaks at 2θ = 5.16°, 5.88°, 9.06° and 16.54°.³⁴

Fig. 2 Comparison of simulated MIL-101(Cr) and synthesized MIL-101(Cr) with its hybrids containing 5 wt% GO, synthesized by post-synthetic hybridization, RAM and in situ methods: (a) X-ray diffraction patterns (Cu Kα, 1.5406 Å); (b) FTIR spectra at 293 K; (c) thermogravimetry curves recorded in an argon atmosphere at a heating rate of 10 K min⁻¹.

Incorporation of GO into MIL-101(Cr) induces a reduction in the intensity of the XRD peaks, which may indicate a loss of crystallinity.³⁵ However, peak broadening in XRD does not always directly indicate reduced crystallinity, especially in hybrid materials. This effect is most pronounced in the POST-synthesized hybrid; alternatively, as pointed out by some of us previously, the GO nanosheets act as a co-ligand that reacts at the early stage of the reaction with the metal cation and thus changes the nucleation rate of the MOF,²¹ probably favoring the formation of smaller MOF nanoparticles. Ex situ methods (RAM and post-synthetic) introduce heterogeneous crystalline–amorphous interfaces, evident in the baseline variations of their PXRD profiles. In contrast, MIL101-GO in situ shows no baseline deviations, likely due to the intimate integration of the GO within the MOF during crystallization. Overall, all hybrids match the simulated and synthesized MIL-101(Cr) patterns, confirming preservation of the parent structure.

Fig. 2b compares FTIR spectra of the different MIL-101(Cr)@GO hybrids with pristine MIL-101(Cr), confirming that the hybrid materials preserve the main structural features of MIL-101(Cr), visible through their unchanged characteristic absorption bands. The band at 1629 cm⁻¹ corresponds to δ(H₂O) bending vibrations from adsorbed water,³⁶ while the prominent peak around 1400 cm⁻¹ arises from symmetrical (O–C–O) stretching in terephthalic acid.³⁷ Aromatic ring vibrations between 600 and 1600 cm⁻¹ include (C=C) stretching at 1508 cm⁻¹ and (C–H) out-of-plane deformations at 1165, 1019, 884 and 747 cm⁻¹.³⁸ The absence of new peaks or spectral shifts confirms the absence of covalent bonding between GO and MIL-101(Cr).

TGA analysis of MIL-101(Cr)@GO hybrids (Fig. 2c) reveals a three-stage decomposition profile consistent with pristine MIL-101(Cr). Initial mass loss (303–473 K) corresponds to the desorption of physisorbed water from the microporous supertetrahedra as well as the large mesopores, 3.4 nm cavities, followed by the removal of water confined in smaller mesopores (2.9 nm cavities) between 473 and 623 K. Above 623 K, framework degradation occurs through dehydroxylation and linker decomposition.^39–41 Hybrids synthesized by post-synthetic and in situ methods exhibit almost identical thermal behavior to the parent MOF, with decomposition onset temperatures exceeding 698 K under Ar (Fig. S3a–c), confirming their good thermal stability. In contrast, the RAM hybrid shows a more pronounced mass loss (∼15% higher) in the 303–623 K range, attributed to increased water adsorption in MIL-101(Cr) cavities, due to GO-induced changes in porosity, and increased water affinity (Fig. S3d). Despite this, all hybrids maintain framework degradation temperatures within the ranges reported in the literature (Fig. S3e), indicating preserved structural integrity regardless of the synthesis route.

Fig. 3 presents optical microscope images of the samples and Fig. 3a and b show the pure components. GO exhibits a dark color, while MIL-101(Cr) appears as a pale light-blue material, facilitating visual distinction between the two materials. The hybrids exhibit a darker color due to the addition of GO, with varying intensities depending on the hybridization method, although all samples contain the same weight fraction (5 wt%) (Fig. 3c and S4a, b). Nevertheless, GO is uniformly distributed throughout the samples, with no visible regions of loose or aggregated GO, indicating a successful combination of MIL-101(Cr) with GO.

Fig. 3 Optical microscopy images: (a) GO; (b) MIL-101(Cr); (c) M101-GO in situ; and (d) SEM image of M101-GO in situ, with its corresponding (e) chromium and (f) carbon elemental maps.

Fig. 3d–f present SEM images along with their corresponding elemental mappings for Cr and C. Cr is highlighted as a key component of the MIL-101(Cr) synthesized viathe in situ method (ex situ methods are collected in Fig. S5). The EDS mappings reveal complete spatial overlap between Cr and C signals in all samples. The absence of C-only regions (which would indicate isolated GO domains) suggests the successful combination of both components across all synthesis methods.

Fig. 4a shows the N₂ adsorption–desorption isotherms for all samples. Pristine MIL-101(Cr) exhibits a combination of a Type I(b) isotherm (due to the presence of supermicropores and narrow mesopores) and a Type IV(a) isotherm (due to the presence of mesopores wider than 4 nm), in accordance with the IUPAC classification.⁴² The observed H4 hysteresis loop indicates mesopores with diverse shapes and a broad pore size distribution (PSD). The M101-GO RAM shows a similar profile, but with reduced N₂ uptake. In contrast, M101-GO POST presents a subtle hysteresis loop closing at p/p° = 0.7, characteristic of a significant loss of mesoporosity. This trend intensifies in M101-GO in situ, which presents an almost Type I(b) isotherm with negligible hysteresis. The synthesis methodology thus directly impacts the hysteresis behavior, and consequently, the mesoporous volume: post-synthetic hybridization reduces it, while in situ synthesis eliminates it, reflecting clear differences in pore connectivity and GO integration into the MOF structure. Logarithmic N₂ adsorption isotherms (Fig. 4b) show similar N₂ uptake at low relative pressures (p/p° < 0.1) for all hybrids and pristine MIL-101(Cr). Argon adsorption–desorption isotherms were recorded at 87 K (Ar boiling point) to complement the nitrogen adsorption studies. The absence of argon's quadrupole moment and its lower polarizability compared with N₂ indeed allow a more accurate analysis of micropore size distribution.⁴³ Despite having a kinetic diameter (∼3.4 Å) similar to that of N₂ (∼3.6 Å), Ar enables more reliable assessment of porous materials' textural characteristics due to its monatomic character and higher saturation temperature (87 K vs. 77 K for N₂), which improve diffusivity and adsorption equilibration. These features make Ar a more suitable probe for accurately assessing the porosity of MOFs such as MIL-101(Cr). Fig. 4c confirms that the Ar adsorption isotherms for MIL-101(Cr) and its hybrids align with their nitrogen-derived IUPAC classifications. However, the M101-GO RAM hybrid exhibits Ar adsorption behavior nearly indistinguishable from pristine MIL-101(Cr).

Fig. 4 Textural characteristics of pure MIL-101(Cr) and hybrids: N₂ adsorption (full symbols)–desorption (empty symbols) isotherms at 77 K on (a) linear and (b) logarithmic scales; (c) Ar adsorption (full symbols)–desorption (empty symbols) isotherms at 87 K of MIL-101(Cr) and hybrids; (d) 2D-NLDFT fit for heterogeneous surfaces using N₂ and H₂ isotherms for MIL-101(Cr); (e) pore size distribution; and (f) cumulative pore volume V_cum as a function of pore width.

Fig. 4c confirms that the Ar adsorption isotherms for MIL-101(Cr) and its hybrids align with their nitrogen-derived IUPAC classifications. However, the M101-GO RAM hybrid exhibits Ar adsorption behavior nearly indistinguishable from pristine MIL-101(Cr), which was not the case for N₂ adsorption. Comparative analysis of the BET area (Fig. S6) reveals that the hybrids show lower Ar-derived surface areas due to the monoatomic and nonpolar nature of Ar, which avoids quadrupole interactions and orientation-dependent adsorption effects observed with N₂ on polar surfaces such as open metal sites. If genuine GO had been used, its higher content of oxygen functional groups and greater hydrophilicity would have promoted a more uniform dispersion in the synthesis medium, as well as stronger interfacial interactions with the MIL-101(Cr) framework. This could have resulted in more homogeneously distributed hybrid structures and higher accessible surface areas. It should be noted here that the application of 2D-NLDFT models to MOFs remains challenging due to the structural diversity within this class of materials, with additional complexities for hybrid systems. In this study, the 2D-NLDFT model was applied to both N₂ (Fig. 4a) and H₂ adsorption data (Fig. S7), which enables robust assessment of pore structures. The excellent agreement between experimental isotherms and model fits (Fig. 4d, for MIL-101(Cr)) validates this approach.

The PSDs presented in Fig. 4e provide compelling evidence that GO incorporation generates ultramicroporosity in all hybrid materials, regardless of the synthesis method employed. Pristine MIL-101(Cr) exhibits a distinct mesopore peak at 15.4 nm, which is preserved in the M101-GO RAM hybrid, suggesting minimal disruption of the mesoporous framework during its synthesis. However, post-synthetic modification (M101-GO POST) reduces the mesopore volume, likely due to partial filling of the interparticle voids by GO sheets, which occupy the space between MOF nanoparticles. The M101-GO in situ material shows almost complete suppression of the mesoporous signature, as further confirmed by the cumulative pore volume (V_cum) profiles (Fig. 4f). The textural properties of the studied materials, including the pure MOF and its hybrids, are summarized in Table S1.

Water affinity is a critical property for the performance of MOFs in adsorption processes, particularly in gas separation and storage applications, as the presence of moisture can significantly affect their efficiency and capacity.^44–46 Unlike carbon-based materials, where water adsorption begins preferentially at carboxyl group sites and forms molecular clusters as nucleation points (Dubinin and Serpinski model⁴⁷), MOFs exhibit more complex interactions. The isotherms presented in Fig. 5a show that, although all samples exhibit similar behavior up to a relative pressure of 0.4, differences appear at lower pressures. All materials show a strong affinity for water, as indicated by the steep uptake at low relative pressures (p/p° < 0.1). Although the three hybrids contain the same wt% of GO, their affinity for water differs significantly. GO is known for its high water affinity due to its hydrophilic surface, rich in oxygen-containing functional groups, and its layered structure, which facilitates interactions and enhances water adsorption.^48,49 In the partially reduced form used, some oxygen functionalities are slightly diminished, but the material still retains significant hydrophilicity. Combining GO with a MOF is therefore expected to increase the material's overall affinity for water. However, the extent of this effect appears to be influenced by the hybridization method, suggesting that GO-induced structural and textural modifications play a crucial role in determining water adsorption behavior.

Fig. 5 Water affinity and thermal conductivity of pure MIL-101(Cr) and hybrids: (a) water adsorption isotherms at 293 K; (b) magnified view of the water affinity region (p/p° ≤ 0.015); (c) thermal conductivity of all samples measured in powder form (298 K).

Water affinity was quantified by determining the slope of water adsorption isotherms in the low-pressure range (p/p° ≤ 0.015), where adsorption is most sensitive to surface interactions (Fig. 5b).^50,51 Water adsorption analysis (Fig. 5b) shows that hybrids modified with GO via in situ synthesis and post-synthetic methods exhibit significantly steeper isotherm slopes (2319 cm³ g⁻¹ STP and 1856 cm³ g⁻¹ STP, respectively) than pure MIL-101(Cr) (894 cm³ g⁻¹ STP) or the RAM hybrid (893 cm³ g⁻¹ STP). This enhanced slope, influenced by the partially reduced GO's residual hydrophilicity as well as structural changes induced by its integration, such as increased accessibility to hydrophilic sites or functionalization of the pore surface, contributes to the observed water adsorption behavior in the hybrids. This difference is attributed to the fact that the textural properties of M101-GO RAM and the pristine MOF are very similar (as already shown in Fig. 4c), unlike the other two hybrids, which exhibit greater differences due to GO incorporation. Representative adsorption isotherms for N₂, H₂, Ar, and H₂O were recorded in triplicate. The observed variation between runs was consistently below 3%, confirming the accuracy and reproducibility of the results.

Thermal conductivity is a key factor in H₂ storage materials, as it directly impacts the efficiency of gas adsorption (exothermic) and desorption (endothermic) processes. During H₂ adsorption, rapid heat dissipation is essential to prevent localized hot spots, which can shift the thermodynamic equilibrium and reduce storage capacity. The relatively poor thermal conductivity of MIL-101(Cr), mainly due to its huge inner porosity and low packing density, can therefore affect its performance in H₂ storage. Reported thermal conductivity values for MIL-101(Cr) in other studies range from 0.05 to 0.21 W m⁻¹ K⁻¹, reflecting its limited ability to conduct heat.^5,52–54 Elsayed et al.⁵ systematically compared the thermal conductivity of MIL-101(Cr)/GO hybrids synthesized by in situ integration (during MOF crystallization) versus physical mixing (ultrasonic blending). Their study demonstrated that both rGO content and the synthesis method play a key role in the thermal conductivity. The physically mixed hybrid containing 5 wt% rGO showed the highest performance, reaching 0.126 W m⁻¹ K⁻¹ (an improvement of 152% compared to the 0.05 W m⁻¹ K⁻¹ obtained for the pristine MIL-101(Cr)). In contrast, in situ synthesized hybrids with 2 wt% rGO achieved a 60% increase in thermal conductivity (0.08 W m⁻¹ K⁻¹). However, for physical mixtures with lower rGO content, the thermal conductivity either remained unchanged or even decreased; for instance, the sample with 0.5 wt% rGO exhibited a 10% reduction compared to the pristine MOF.

Fig. 5c presents the thermal conductivity values of the samples evaluated in this study. To ensure statistical consistency, each measurement was performed six times. The reported values represent the average of these replicates, with an overall mean standard deviation of ±0.0063 W m⁻¹ K⁻¹ across all samples. Pure MIL-101(Cr), analyzed in powder form (without tapping), has a thermal conductivity of 0.132 W m⁻¹ K⁻¹, which is consistent with values previously reported in the literature. As expected, the incorporation of 5 wt% of GO by different synthesis methods affects differently the thermal conductivity of the hybrids. In particular, M101-GO RAM shows a sharp decrease in thermal conductivity, with a value of 0.065 W m⁻¹ K⁻¹ (−51%). For M101-GO POST, a value of 0.120 W m⁻¹ K⁻¹ (−9% vs. the pure MOF) is observed. A decrease in thermal conductivity was previously observed by Elsayed et al.⁵ in materials synthesized through ex situ hybridization with low rGO content. This effect may be attributed to the weak interfacial bonding between graphene oxide and the host framework, which hinders efficient heat transfer across the interface. The only hybrid showing an increase in thermal conductivity is M101-GO in situ, with a value of 0.166 W m⁻¹ K⁻¹ (+26%). This clear difference can be explained by the ability of the in situ method to disperse GO nanosheets uniformly within the MOF matrix, facilitating the formation of continuous heat-transfer pathways. The closer interfacial contact achieved during in situ synthesis may also reduce thermal resistance at the MOF–GO interface, contributing to better heat transfer. This enhancement could be due to the direct complexation of metal ions with oxygenated groups on rGO during synthesis, leading to a more interconnected hybrid structure with improved thermal transport.²¹ These findings underscore that in situ synthesis is essential for maximizing the thermal conductivity of MOF-based hybrids, particularly for applications requiring efficient heat management.^11,55

Although the in situ hybrid exhibits the highest thermal conductivity among the studied materials, the absolute values remain relatively low compared to those reported for GO. It should be noted that these measurements were performed on powder samples, where interparticle voids limit solid–solid contact and hinder efficient heat transfer. To achieve more significant improvements in thermal conductivity, strategies such as material compaction or the introduction of additional conductive phases could be explored in future work.

3.2. Hydrogen storage and release performance at high pressure

3.2.1 Performance of MOF@GO. Fig. 6a illustrates the excess H₂ uptake (n_exc, wt%) of MIL-101(Cr) and its GO hybrids at 77 K, measured at pressures up to 130 bar. The results obtained over the three cycles showed a consistent trend, with a standard deviation of less than 2%, demonstrating the stability and reliability of the measurements. Pristine MIL-101(Cr) has the highest storage capacity (3.8 wt%), while the incorporation of GO induces a systematic reduction: the M101-GO RAM hybrid shows a 14.3% decrease (3.2 wt%), followed by M101-GO POST (3.00 wt%, i.e., a 20.6% decrease) and M101-GO in situ (2.9 wt%, i.e., a 29.3% decrease). These decreases exceed the theoretical loss expected from the 5 wt% GO addition, suggesting structural modifications. Specifically, the non-porous nature of GO reduces the hybrids' surface area and partially restricts access to the MIL-101(Cr) mesopores, leading to an overall decrease in pore volume. The synthesis method has a critically decisive influence on these effects: post-synthetic and RAM methods lead to a milder decrease in pore volume across different pore size ranges while in situ synthesis maximizes GO-MOF integration, creating dense GO networks that significantly contribute to reducing the mesopore fraction of the sample. The inverse relationship between the level of GO integration (in situ > POST > RAM) and H₂ uptake underscores the trade-off between hybrid structure and gas storage performance. The relationship between n_exc (77 K and 100 bar) and A_BET (from argon adsorption isotherms) follows the established Chahine's rule,⁵⁶ as shown in Fig. S8, which predicts an approximate 1 wt% increase in n_exc per 500 m² g⁻¹ increase in surface area.

Fig. 6 Excess H₂ uptake (n_exc, wt%) at (a) 77 K, (b) 160 K, and (c) 273 K; (d) total H₂ uptake (n_tot, kg m⁻³) at 77 K; (e) stored H₂ (kg m⁻³), compressed and adsorbed; (f) release capacities of H₂ (kg m⁻³) under charge conditions of 77 K and 100 bar and discharge conditions of 160 K and 5 bar. H₂ uptake as a function of A_BET for our materials compared with literature data, represented in different ways: (g) n_exc in wt% at 77 K and 10–60 bar; (h) n_tot in wt% at 77 K and 100–120 bar; and (i) n_tot in kg m⁻³ at 77 K and 100–120 bar.

Fig. 6b shows n_exc at 160 K, highlighting a reversal of the performance trend observed at 77 K in Fig. 6a. Unlike the behavior at 77 K and 100 bar, the hybrids synthesized ex situ exhibit higher adsorption capacities than pristine MIL-101(Cr). Specifically, M101-GO RAM and M101-GO POST reach 2.7 wt% and 2.6 wt% at 100 bars, respectively, compared to 2.5 wt% for the pure MOF. In contrast, the in situ hybrid performs significantly worse, with an uptake of only 1.7 wt%. A similar trend is observed in Fig. 6c at 273 K and 100 bars, where the ex situ hybrids outperform both the pristine MOF and the in situ hybrid. This change in behavior compared to the data at 77 K suggests a temperature-dependent change in the dominant adsorption mechanism. At higher temperatures, the contribution of GO becomes more significant, likely due to the involvement of its defective sites and oxygen-containing functional groups, partially offsetting the loss of surface area compared with MIL-101(Cr). Nevertheless, the enhancement remains limited, as the overall H₂ uptake is still largely governed by the porous structure of MIL-101(Cr). These observations also suggest that the accessibility and effectiveness of the GO active sites depend on the synthesis method used to form the hybrids. The fact that M101-GO in situ showed the lowest gravimetric H₂ storage performance for the different temperatures studied confirms a higher level of integration of the GO structure to the MOF matrix.

For practical purposes, total H₂ uptake (n_tot) is the most relevant parameter when comparing different storage systems, as it reflects the overall amount of gas, including that in interparticle voids. Excess uptake corresponds to the gas adsorbed within the pores beyond the bulk gas density. Total uptake is suitable for system-level comparisons, while excess uptake provides insight into the adsorption behavior inside the pores.

The total H₂ uptake (n_tot, wt%) at 77 K, calculated using eqn (1) and presented in Fig. S9, reveals that pristine MIL-101(Cr) retains a higher gravimetric capacity than hybrid materials, consistent with the trend observed for excess H₂ uptake at this temperature. However, evaluation of volumetric n^v_tot (n^v_tot, kg m⁻³) based on tapped density reveals a notable reversal of this trend (Fig. 6d). While the M101-GO in situ hybrid exhibits a lower gravimetric capacity, it shows superior volumetric performance due to its significantly higher ρ_tap (426 kg m⁻³), as illustrated in the inset of Fig. 6d. In contrast, MIL-101(Cr), POST, and RAM samples present lower ρ_tap values of 250, 290, and 295 kg m⁻³, respectively. The exceptional tapped density observed for the in situ hybrid can be rationalized by the way GO is incorporated during MOF synthesis. In this approach, GO is present throughout the MOF growth, which may promote the formation of smaller and more uniform particles, reducing interparticle voids and enhancing packing efficiency. In contrast, the ex situ method can lead to particle aggregation, increasing the effective particle size and the void fraction, and consequently reducing the tapped density. This clearly highlights the crucial role of material densification in improving volumetric H₂ storage performance. All the values of density (ρ_tap and ρ_skel) used for n_tot calculations are compiled in Table S1. Complementary data, including n_exc (wt%), n_tot (wt%), and volumetric H₂ capacity (kg m⁻³) for pristine MIL-101(Cr) and hybrid materials, are systematically presented in Table S2.

As shown in Fig. 6e, all MIL-101(Cr)@GO hybrids exceed the H₂ storage capacity of an empty tank (31.0 kg m⁻³ at 77 K, 100 bar), confirming their potential for practical storage applications. While the ex situ methods are close (post-synthetic: 33.6 kg m⁻³; RAM: 33.5 kg m⁻³) but do not exceed pure MIL-101(Cr) (33.8 kg m⁻³), the in situ hybrid achieves superior performance (36.0 kg m⁻³) due to an enhanced tapped density that compensates for the reduced surface area. Studies on H₂ release reveal a critical temperature dependence: under cryogenic conditions (77 K, 5 bar), empty tanks outperform all adsorbents with 29.5 kg m⁻³vs. 26.7 kg m⁻³ for MIL-101(Cr) (Table S3). However, the introduction of a moderate temperature swing (77 K → 160 K) during desorption reverses this trend. The in situ hybrid achieves a release capacity of 34.5 kg m⁻³, surpassing both the empty tank (30.4 kg m⁻³, i.e., by +13.6%) and the pure MOF (32.5 kg m⁻³, i.e., by +6.2%). The post-synthetic hybrid (34.1 kg m⁻³) shows similar gains, while the RAM hybrid (32.1 kg m⁻³) matches the parent MOF, underscoring the importance of the synthesis method in optimizing interactions for thermal-responsive performance (Fig. 6f). This highlights the dual role of the synthesis strategy: in situ integration maximizes volumetric storage and enables thermally enhanced release, whereas RAM offers only minimal improvement. Table S3 also highlights the impact of thermal activation on H₂ release efficiency, showing a significant improvement when the discharge temperature increases from 77 K to 160 K. At 160 K, porous materials achieve release efficiencies of up to 96%, compared with the lower range of 74–79% observed at 77 K. This demonstrates the critical role of thermal energy in accelerating desorption kinetics and maximizing usable H₂ capacity.

3.2.2 Comparison with the literature. To contextualize the values obtained in this study, a comparative analysis was carried out with results reported in previous works, evaluating n_exc (gravimetric, wt%), n_tot (gravimetric, wt%) and n_tot (volumetric, kg m⁻³). When plotting n_exc as a function of A_BET (Fig. 6g), a clear trend can be observed, linking the increase in surface area with a higher H₂ uptake (n_exc, wt%) in agreement with the well-known Chahine's rule⁵⁶ (i.e., an increase of 1 wt% in n_exc per 500 m² g⁻¹ of surface area). Within the data cloud referring to n_exc values reported in other studies (both for ◇ MOFs and ☆ carbons), the results obtained in this work fall in an intermediate position in terms of H₂ adsorption capacity (n_exc, wt%), consistent with the surface area achieved for the synthesized materials. Although MIL-101(Cr) is generally characterized by its very high surface area (2500–4000 m² g⁻¹),³ this property is strongly dependent on the synthesis method employed. In our case, by prioritizing a greener synthesis of MIL-101(Cr), avoiding hazardous reagents such as HF, a compromise in the specific surface area is made, mainly due to the difficulty of completely removing the organic linker and other residues from the pores. This compromise ultimately places our materials in the mid-range of reported H₂ adsorption values.

In contrast, when analyzing the total H₂ adsorption capacity (n_tot, wt%) as a function of the A_BET (Fig. 6h), the values reported in this study no longer fall within the average range but rather shift towards the upper part of the graph, reaching levels comparable to the best-performing studies (for both ◇ MOFs and ☆ carbons). This result is particularly noteworthy given the lower surface area of our materials and may be attributed to the significant contribution of H₂ stored by cryo-compression at 77 K and 100–120 bar, within the void spaces of the system and not solely by adsorption.

Finally, we compared the total H₂ adsorption capacity reported in the literature in terms of volumetric capacity (kg m⁻³) as a function of A_BET (Fig. 6i). As shown in the graph, only a few studies report total H₂ adsorption in volumetric terms at high pressures, which limits the number of studies with which we can directly compare our results. In this study, our volumetric values were calculated using the tap density, as we consider this a more realistic approximation of the material's effective density. However, most of the studies reporting volumetric capacities do so using the crystal density (ρ_cry). To enable a fair comparison, we recalculated the total H₂ adsorption (n_tot) in volumetric terms for our materials using the ρ_cry of MIL-101(Cr) (0.49 g cm⁻³).⁵⁷ Since the actual packing densities of literature samples are not reported, the material was assumed to behave as a packed granular system composed of spherical grains. Considering the inherent compactness of such a packing, the effective packing density was estimated to be 74% of the single-crystal density (ρ_pack = 0.74 × ρ_cry), which provides a more realistic approximation for volumetric calculations.⁵⁸ This correction factor has been applied consistently to all samples, including those from the literature where the actual packing behavior is unknown. While it is an approximation, it provides a more realistic estimation of void volumes, improves comparability across different samples, and ensures that volumetric capacities are not overestimated.

When using ρ_cry, the samples reported in this study appear with very high volumetric values, surpassing other materials that exhibit higher A_BET. Moreover, when we analyze the total H₂ storage capacity of our samples using ρ_cry, the trend shifts: the M101-GO in situ sample, which was previously identified as having the highest volumetric capacity, now appears as the lowest among our samples. This effect results from the standardization of density, since the same theoretical ρ_cry is applied to all samples rather than using the actual measured values for each material. By imposing a uniform density, only the differences in gravimetric capacity are preserved, and the impact of GO addition on the density of our samples is not reflected. Therefore, claims of volumetric superiority based entirely on ρ_cry should be interpreted with caution. In addition to altering the observed trend, this approach produces significantly higher total adsorption values compared with those obtained using ρ_tap, leading to what we consider an overestimation of volumetric capacities. These findings underscore the importance of performing specific density measurements for each material, rather than relying on theoretical crystal densities, which may deviate considerably from real values. The numerical data used to plot Fig. 6g–i are provided in Tables S4–S6, respectively.

4. Conclusions

This study shows the critical role of synthesis strategy in dictating the hydrogen (H₂) storage performance of MIL-101(Cr)@GO hybrid materials, highlighting pivotal trade-offs between gravimetric and volumetric efficiency. Ex situ methods with 5 wt% GO (post-synthetic and RAM) preserve the parent's MOF textural properties and increase gravimetric capacity by 22% at 273 K and 100 bar. In contrast, in situ synthesis fundamentally reshapes the hybrid architecture, enabling a breakthrough in volumetric performance, 30–39% higher tapped density (426 kg m⁻³) and superior total (36.0 kg m⁻³) and deliverable (34.5 kg m⁻³) capacities. The in situ hybrid thus outperforms both pristine MIL-101(Cr) and ex situ hybrids by 6–7%. The systematic introduction of ultramicroporosity (<0.7 nm) and a concomitant decrease in surface area, driven by GO incorporation, underscore a strategic balance to optimize volumetric storage and deliverable capacities. Therefore, for the specific hybrids studied here, ex situ hybrids are better suited for gravimetrically constrained applications, while in situ hybrids are optimized for volumetric, space-limited storage scenarios. It should be noted that these statements are based on material-level characterization. Further system-level analysis would be required to fully assess the practical suitability of these hybrids in real H₂ storage devices. The presence of GO boosts water affinity in all MIL-101(Cr) hybrids, regardless of the synthesis route. However, thermal conductivity exhibits a synthesis-dependent trend: decreasing in ex situ hybrids but improving significantly with in situ integration. These findings highlight a paradigm shift in MOF hybrid design. Rather than focusing solely on traditional metrics like porosity and surface area, synthesis-driven structural control emerges as a powerful lever for aligning material properties to meet application-specific requirements, paving the way for next-generation H₂ storage systems tailored to real-world performance demands.

Abbreviations

(ntotal)release	Total H2 uptake under discharging conditions (kg m^3)
(ntotal)storage	Total H2 uptake under charging conditions (kg m^3)
2D-NLDFT	Two-dimensional non-local density functional theory
A BET	Brunauer–Emmett–Teller area (m^2 g^−1)
DTG	Derived thermogravimetry
FTIR	Fourier-transform infrared spectroscopy
GO	Graphene oxide
IUPAC	International Union of Pure and Applied Chemistry
MOF	Metal–organic framework
n exc	Excess hydrogen uptake measured (wt%)
n toz	Compressed H2 + excess H2 (wt%)
n ^vtot	Compressed H2 + excess H2 (kg m^−3)
PSDs	Pore size distributions
RAM	Resonant acoustic mixing
rGO	Reduced graphene oxide
S 2D-NLDFT	2D-NLDFT based surface area (m^2 g^−1)
SEM	Scanning electron microscope
SSA	Specific surface area (m^2 g^−1)
TGA	Thermogravimetric analysis
V cum	Accumulated pore volume (cm^3 g^−1)
V meso	Mesopore volume (cm^3 g^−1)
V sµ	Supermicropore volume (cm^3 g^−1)
V tot	Total pore volume (cm^3 g^−1)
V uµ	Ultramicropore volume (cm^3 g^−1)
XRD	X-ray diffraction
ρ H2	H2 density (kg m^−3)
ρ skel	Skeletal density (kg m^−3)
ρ tap	Tapped density (kg m^−3)

Conflicts of interest

There are not conflicts to declare.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information is available. See DOI: https://doi.org/10.1039/d5ta06828b.

Acknowledgements

This work was made possible thanks to (i) the SOLHYD project (ANR-22-PEHY-0007) funded by the Agence Nationale de la Recherche, (ii) the FRCR HyPE project funded by Région Grand-Est, and (iii) the TALiSMAN and TALiSMAN2 projects funded by FEDER. The authors gratefully acknowledge Mr Philippe Gadonneix for his invaluable dedication and outstanding support in the laboratory.

References

1. F. Rouquerol, J. Rouquerol and K. S. W. Sing, Thermodynamics of Adsorption at the Gas/Solid Interface, in Adsorption by Powders and Porous Solids, Elsevier, 2014, pp. 25–56,  DOI:10.1016/B978-0-08-097035-6.00002-4.
2. G. Sdanghi, S. Schaefer, G. Maranzana, A. Celzard and V. Fierro, Application of the modified Dubinin-Astakhov equation for a better understanding of high-pressure hydrogen adsorption on activated carbons, Int. J. Hydrogen Energy, 2020, 45, 25912–25926,  DOI:10.1016/j.ijhydene.2019.09.240.
3. M. Y. Zorainy, M. G. Alalm, S. Kaliaguine and D. C. Boffito, Revisiting the MIL-101 metal–organic framework: design, synthesis, modifications, advances, and recent applications, J. Mater. Chem. A, 2021, 9, 22159–22217,  10.1039/D1TA06238G.
4. D. Denysenko, M. Grzywa, M. Tonigold, B. Streppel, I. Krkljus, M. Hirscher, E. Mugnaioli, U. Kolb, J. Hanss and D. Volkmer, Elucidating Gating Effects for Hydrogen Sorption in MFU-4-Type Triazolate-Based Metal–Organic Frameworks Featuring Different Pore Sizes, Chem.–Eur. J., 2011, 17, 1837–1848,  DOI:10.1002/chem.201001872.
5. E. Elsayed, H. Wang, P. A. Anderson, R. Al-Dadah, S. Mahmoud, H. Navarro, Y. Ding and J. Bowen, Development of MIL-101(Cr)/GrO composites for adsorption heat pump applications, Microporous Mesoporous Mater., 2017, 244, 180–191,  DOI:10.1016/j.micromeso.2017.02.020.
6. S. B. Singh and S. A. Dastgheib, Characteristics of graphene oxide-like materials prepared from different deashed-devolatilized coal chars and comparison with graphite-based graphene oxide, with or without the ultrasonication treatment, Carbon, 2024, 228, 119331,  DOI:10.1016/j.carbon.2024.119331.
7. M. Muschi, S. Devautour-Vinot, D. Aureau, N. Heymans, S. Sene, R. Emmerich, A. Ploumistos, A. Geneste, N. Steunou, G. Patriarche, G. De Weireld and C. Serre, Metal–organic framework/graphene oxide composites for CO₂ capture by microwave swing adsorption, J. Mater. Chem. A, 2021, 9, 13135–13142,  10.1039/D0TA12215G.
8. S. Liu, L. Sun, F. Xu, J. Zhang, C. Jiao, F. Li, Z. Li, S. Wang, Z. Wang, X. Jiang, H. Zhou, L. Yang and C. Schick, Nanosized Cu-MOFs induced by graphene oxide and enhanced gas storage capacity, Energy Environ. Sci., 2013, 6, 818–823,  10.1039/C3EE23421E.
9. H. Zhao, D. Bahamon, M. Khaleel and L. F. Vega, Insights into the performance of hybrid graphene oxide/MOFs for CO₂ capture at process conditions by molecular simulations, Chem. Eng. J., 2022, 449, 137884,  DOI:10.1016/j.cej.2022.137884.
10. C. Petit and T. J. Bandosz, MOF–Graphite Oxide Composites: Combining the Uniqueness of Graphene Layers and Metal–Organic Frameworks, Adv. Mater., 2009, 21, 4753–4757,  DOI:10.1002/adma.200901581.
11. M. I. Nandasiri, J. Liu, B. P. McGrail, J. Jenks, H. T. Schaef, V. Shutthanandan, Z. Nie, P. F. Martin and S. K. Nune, Increased Thermal Conductivity in Metal-Organic Heat Carrier Nanofluids, Sci. Rep., 2016, 6, 27805,  DOI:10.1038/srep27805.
12. M. Muschi and C. Serre, Progress and challenges of graphene oxide/metal-organic composites, Coord. Chem. Rev., 2019, 387, 262–272,  DOI:10.1016/j.ccr.2019.02.017.
13. L. J. Lopez, R. M. Ospino, J. C. Gutiérrez, A. Celzard and V. Fierro, Enhanced hydrogen storage and release capacities when using HKUST-1@activated carbon hybrids obtained by resonant acoustic mixing, Chem. Eng. J., 2025, 165834,  DOI:10.1016/j.cej.2025.165834.
14. J. Yan, Y. Yu, C. Ma, J. Xiao, Q. Xia, Y. Li and Z. Li, Adsorption isotherms and kinetics of water vapor on novel adsorbents MIL-101(Cr)@GO with super-high capacity, Appl. Therm. Eng., 2015, 84, 118–125,  DOI:10.1016/j.applthermaleng.2015.03.040.
15. M. F. L. Villena, Z. D. Doorenbos, K. T. Sullivan and B. Brettmann, Evaluating Resonant Acoustic Mixing as a Wet Granulation Process, Org. Process Res. Dev., 2024, 28, 4338–4347,  DOI:10.1021/acs.oprd.4c00347.
16. C. B. Lennox, T. H. Borchers, L. Gonnet, C. J. Barrett, S. G. Koenig, K. Nagapudi and T. Friščić, Direct mechanocatalysis by resonant acoustic mixing (RAM), Chem. Sci., 2023, 14, 7475–7481,  10.1039/D3SC01591B.
17. H. M. Titi, J.-L. Do, A. J. Howarth, K. Nagapudi and T. Friščić, Simple, scalable mechanosynthesis of metal–organic frameworks using liquid-assisted resonant acoustic mixing (LA-RAM), Chem. Sci., 2020, 11, 7578–7584,  10.1039/D0SC00333F.
18. E. Gkaniatsou, C. Sicard, R. Ricoux, L. Benahmed, F. Bourdreux, Q. Zhang, C. Serre, J.-P. Mahy and N. Steunou, Enzyme Encapsulation in Mesoporous Metal–Organic Frameworks for Selective Biodegradation of Harmful Dye Molecules, Angew. Chem., Int. Ed., 2018, 57, 16141–16146,  DOI:10.1002/anie.201811327.
19. T. Zhao, S.-H. Li, L. Shen, Y. Wang and X.-Y. Yang, The sized controlled synthesis of MIL-101(Cr) with enhanced CO₂ adsorption property, Inorg. Chem. Commun., 2018, 96, 47–51,  DOI:10.1016/j.inoche.2018.07.036.
20. X. Xia and S. Li, Improved adsorption cooling performance of MIL-101(Cr)/GO composites by tuning the water adsorption rate, Sustainable Energy Fuels, 2023, 7, 437–447,  10.1039/D2SE01508K.
21. M. Muschi, S. Devautour-Vinot, D. Aureau, N. Heymans, S. Sene, R. Emmerich, A. Ploumistos, A. Geneste, N. Steunou, G. Patriarche, G. D. Weireld and C. Serre, Metal–organic framework/graphene oxide composites for CO₂ capture by microwave swing adsorption, J. Mater. Chem. A, 2021, 9, 13135–13142,  10.1039/D0TA12215G.
22. C. R. Groom, I. J. Bruno, M. P. Lightfoot and S. C. Ward, The Cambridge Structural Database, Acta Crystallogr., Sect. B, 2016, 72, 171–179,  DOI:10.1107/S2052520616003954.
23. J. Rouquerol, P. Llewellyn and F. Rouquerol, Is the bet equation applicable to microporous adsorbents?, in Studies in Surface Science and Catalysis, Elsevier, 2007, pp. 49–56,  DOI:10.1016/S0167-2991(07)80008-5.
24. L. Jimenez Lopez, R. Morales-Ospino, J. Castro-Gutiérrez, H. O. Sumbhaniya, G. Sdanghi, S. García Dalí, A. Celzard and V. Fierro, Boosting hydrogen storage and release in MOF-5/graphite hybrids via in situ synthesis, Int. J. Hydrogen Energy, 2025, 173, 151272,  DOI:10.1016/j.ijhydene.2025.151272.
25. J. W. Leachman, R. T. Jacobsen, S. G. Penoncello and E. W. Lemmon, Fundamental Equations of State for Parahydrogen, Normal Hydrogen, and Orthohydrogen, J. Phys. Chem. Ref. Data, 2009, 38, 721–748,  DOI:10.1063/1.3160306.
26. R. Morales-Ospino, L. Jiménez-López, A. Celzard and V. Fierro, Hydrogen – Storage|Physical storage, in Reference Module in Chemistry, Molecular Sciences and Chemical Engineering, Elsevier, 2024, pp. 319–329,  DOI:10.1016/B978-0-323-96022-9.00290-5.
27. Z. Chen, P. Li, R. Anderson, X. Wang, X. Zhang, L. Robison, L. R. Redfern, S. Moribe, T. Islamoglu, D. A. Gómez-Gualdrón, T. Yildirim, J. F. Stoddart and O. K. Farha, Balancing volumetric and gravimetric uptake in highly porous materials for clean energy, Science, 2020, 368, 297–303,  DOI:10.1126/science.aaz8881.
28. P. García-Holley, B. Schweitzer, T. Islamoglu, Y. Liu, L. Lin, S. Rodriguez, M. H. Weston, J. T. Hupp, D. A. Gómez-Gualdrón, T. Yildirim and O. K. Farha, Benchmark Study of Hydrogen Storage in Metal–Organic Frameworks under Temperature and Pressure Swing Conditions, ACS Energy Lett., 2018, 3, 748–754,  DOI:10.1021/acsenergylett.8b00154.
29. M. Schlichtenmayer and M. Hirscher, The usable capacity of porous materials for hydrogen storage, Appl. Phys. A, 2016, 122, 379,  DOI:10.1007/s00339-016-9864-6.
30. DOE Materials-Based Hydrogen Storage Summit: Defining Pathways for Onboard Automotive Applications, Energy. Gov., https://www.energy.gov/eere/fuelcells/articles/doe-materials-based-hydrogen-storage-summit-defining-pathways-onboard, accessed November 12, 2024.
31. P. García-Holley, B. Schweitzer, T. Islamoglu, Y. Liu, L. Lin, S. Rodriguez, M. H. Weston, J. T. Hupp, D. A. Gómez-Gualdrón, T. Yildirim and O. K. Farha, Benchmark Study of Hydrogen Storage in Metal–Organic Frameworks under Temperature and Pressure Swing Conditions, ACS Energy Lett., 2018, 3, 748–754,  DOI:10.1021/acsenergylett.8b00154.
32. M. D. Allendorf, Z. Hulvey, T. Gennett, A. Ahmed, T. Autrey, J. Camp, E. S. Cho, H. Furukawa, M. Haranczyk, M. Head-Gordon, S. Jeong, A. Karkamkar, D.-J. Liu, J. R. Long, K. R. Meihaus, I. H. Nayyar, R. Nazarov, D. J. Siegel, V. Stavila, J. J. Urban, S. P. Veccham and B. C. Wood, An assessment of strategies for the development of solid-state adsorbents for vehicular hydrogen storage, Energy Environ. Sci., 2018, 11, 2784–2812,  10.1039/C8EE01085D.
33. S. B. Singh and M. De, Thermally exfoliated graphene oxide for hydrogen storage, Mater. Chem. Phys., 2020, 239, 122102,  DOI:10.1016/j.matchemphys.2019.122102.
34. M. Y. Zorainy, H. M. Titi, S. Kaliaguine and D. C. Boffito, Multivariate metal–organic framework MTV-MIL-101 via post-synthetic cation exchange: is it truly achievable?, Dalton Trans., 2022, 51, 3280–3294,  10.1039/D1DT04222J.
35. S. C. Tan and H. K. Lee, A hydrogel composite prepared from alginate, an amino-functionalized metal-organic framework of type MIL-101(Cr), and magnetite nanoparticles for magnetic solid-phase extraction and UHPLC-MS/MS analysis of polar chlorophenoxy acid herbicides, Microchim. Acta, 2019, 186, 1–11,  DOI:10.1007/s00604-019-3679-z.
36. S. S. Khaloo, A. Bagheri, R. Gholamnia and R. Saeedi, Graphene oxide/MIL 101(Cr) (GO/MOF) nano-composite for adsorptive removal of 2,4-dichlorophenoxyacetic acid (2,4 D) from aqueous media: synthesis, characterization, kinetic and isotherm studies, Water Sci. Technol., 2022, 86, 1496–1509,  DOI:10.2166/wst.2022.282.
37. P. K. Prabhakaran and J. Deschamps, Doping activated carbon incorporated composite MIL-101 using lithium: impact on hydrogen uptake, J. Mater. Chem. A, 2015, 3, 7014–7021,  10.1039/C4TA07197B.
38. S. M. Mirsoleimani-azizi, P. Setoodeh, F. Samimi, J. Shadmehr, N. Hamedi and M. R. Rahimpour, Diazinon removal from aqueous media by mesoporous MIL-101(Cr) in a continuous fixed-bed system, J. Environ. Chem. Eng., 2018, 6, 4653–4664,  DOI:10.1016/j.jece.2018.06.067.
39. S. Tourani and A. Behvandi, Synthesis of MIL-101(Cr)/Sulfasalazine (Cr-TA@SSZ) hybrid and its use as a novel adsorbent for adsorptive removal of organic pollutants from wastewaters, J. Porous Mater., 2022, 29, 1441–1462,  DOI:10.1007/s10934-022-01268-4.
40. J. Shadmehr, S. Zeinali and M. Tohidi, Synthesis of a chromium terephthalate metal organic framework and use as nanoporous adsorbent for removal of diazinon organophosphorus insecticide from aqueous media, J. Dispersion Sci. Technol., 2019, 40, 1423–1440,  DOI:10.1080/01932691.2018.1516149.
41. N. Wang, L.-Y. Yang, Y. Wang and X. Ouyang, Fabrication of Composite Beads Based on Calcium Alginate and Tetraethylenepentamine-Functionalized MIL-101 for Adsorption of Pb(II) from Aqueous Solutions, Polymers, 2018, 10, 750,  DOI:10.3390/polym10070750.
42. M. Thommes, K. Kaneko, A. V. Neimark, J. P. Olivier, F. Rodriguez-Reinoso, J. Rouquerol and K. S. W. Sing, Physisorption of gases, with special reference to the evaluation of surface area and pore size distribution (IUPAC Technical Report), Pure Appl. Chem., 2015, 87, 1051–1069,  DOI:10.1515/pac-2014-1117.
43. H. S. Cho, J. Yang, X. Gong, Y.-B. Zhang, K. Momma, B. M. Weckhuysen, H. Deng, J. K. Kang, O. M. Yaghi and O. Terasaki, Isotherms of individual pores by gas adsorption crystallography, Nat. Chem., 2019, 11, 562–570,  DOI:10.1038/s41557-019-0257-2.
44. M. Ding and H.-L. Jiang, Improving Water Stability of Metal–Organic Frameworks by a General Surface Hydrophobic Polymerization, CCS Chem., 2020, 3, 2740–2748,  DOI:10.31635/ccschem.020.202000515.
45. N. C. Burtch, H. Jasuja and K. S. Walton, Water Stability and Adsorption in Metal–Organic Frameworks, Chem. Rev., 2014, 114, 10575–10612,  DOI:10.1021/cr5002589.
46. A. Tóth and K. László, Water Adsorption by Carbons. Hydrophobicity and Hydrophilicity, in Novel Carbon Adsorbents, 2012, pp. 147–171,  DOI:10.1016/B978-0-08-097744-7.00005-3.
47. F. Stoeckli, T. Jakubov and A. Lavanchy, Water adsorption in active carbons described by the Dubinin–Astakhov equation, J. Chem. Soc., Faraday Trans., 1994, 90, 783–786,  10.1039/FT9949000783.
48. R. Liu, T. Gong, K. Zhang and C. Lee, Graphene oxide papers with high water adsorption capacity for air dehumidification, Sci. Rep., 2017, 7, 9761,  DOI:10.1038/s41598-017-09777-y.
49. B. Lian, S. D. Luca, Y. You, S. Alwarappan, M. Yoshimura, V. Sahajwalla, S. C. Smith, G. Leslie and R. K. Joshi, Extraordinary water adsorption characteristics of graphene oxide, Chem. Sci., 2018, 9, 5106–5111,  10.1039/C8SC00545A.
50. J. Castro-Gutiérrez, R. L. S. Canevesi, M. Emo, M. T. Izquierdo, A. Celzard and V. Fierro, CO₂ outperforms KOH as an activator for high-rate supercapacitors in aqueous electrolyte, Renewable Sustainable Energy Rev., 2022, 167, 112716,  DOI:10.1016/j.rser.2022.112716.
51. L. Liu, S. (Johnathan) Tan, T. Horikawa, D. D. Do, D. Nicholson and J. Liu, Water adsorption on carbon - A review, Adv. Colloid Interface Sci., 2017, 250, 64–78,  DOI:10.1016/j.cis.2017.10.002.
52. M. Fang, R. He, J. Zhou, H. Fei and K. Yang, Thermal conductivity enhancement and shape stability of composite phase change materials using MIL-101(Cr)-NH₂/expanded graphite/multi-walled carbon nanotubes, J. Energy Storage, 2024, 86, 111244,  DOI:10.1016/j.est.2024.111244.
53. J. Zhou, M. Fang, K. Yang, K. Lu, H. Fei, P. Mu and R. He, MIL-101(Cr)-NH₂/reduced graphene oxide composite carrier enhanced thermal conductivity and stability of shape-stabilized phase change materials for thermal energy management, J. Energy Storage, 2022, 52, 104827,  DOI:10.1016/j.est.2022.104827.
54. J. Wang, X. Huang, H. Gao, A. Li and C. Wang, Construction of CNT@Cr-MIL-101-NH₂ hybrid composite for shape-stabilized phase change materials with enhanced thermal conductivity, Chem. Eng. J., 2018, 350, 164–172,  DOI:10.1016/j.cej.2018.05.190.
55. T. K. Vo, T. P. Trinh, V. C. Nguyen and J. Kim, Facile synthesis of graphite oxide/MIL-101(Cr) hybrid composites for enhanced adsorption performance towards industrial toxic dyes, J. Ind. Eng. Chem., 2021, 95, 224–234,  DOI:10.1016/j.jiec.2020.12.023.
56. P. Bénard and R. Chahine, Storage of hydrogen by physisorption on carbon and nanostructured materials, Scr. Mater., 2007, 56, 803–808,  DOI:10.1016/j.scriptamat.2007.01.008.
57. D. Dybtsev, C. Serre, B. Schmitz, B. Panella, M. Hirscher, M. Latroche, P. L. Llewellyn, S. Cordier, Y. Molard, M. Haouas, F. Taulelle and G. Férey, Influence of [Mo₆Br₈F₆]²⁻ Cluster Unit Inclusion within the Mesoporous Solid MIL-101 on Hydrogen Storage Performance, Langmuir, 2010, 26, 11283–11290,  DOI:10.1021/la100601a.
58. P. Ramirez-Vidal, R. L. S. Canevesi, A. Celzard and V. Fierro, Modeling High-Pressure Hydrogen Uptake by Nanoporous Metal–Organic Frameworks: Implications for Hydrogen Storage and Delivery, ACS Appl. Nano Mater., 2022, 5, 759–773,  DOI:10.1021/acsanm.1c03493.

---