Tuning flexibility in metal–organic frameworks via linker per-fluorination: revisiting the adsorption-induced breathing of MIL-53(Al)

Abstract

We present the first comprehensive investigation of the complex, multi-step adsorption-induced breathing behaviour of F₄-MIL-53(Al), the recently discovered analogue of MIL-53(Al) with a per-fluorinated linker. Through a systematic characterisation approach performed by combining in situ Powder X-ray Diffraction (PXRD), in situ Fourier Transform Infrared (FTIR) and Solid State Nuclear Magnetic Resonance (SS-NMR) spectroscopies with sorption analyses, we unveil the impact of fluorination on framework flexibility, adsorption properties, and phase transitions, offering fresh perspectives into the structure–property relationships governing Metal–Organic Framework (MOF) dynamic porosity. Compared to the non-fluorinated MIL-53(Al), F₄-MIL-53(Al) exhibits a different water affinity, with uptake remaining below 1 mmol g⁻¹ up to 60% relative humidity. Above this threshold, PXRD reveals a two-step expansion of the F₄-MIL-53(Al) unit cell, contrasting the typical contraction observed in MIL-53(Al). Volumetric CO₂ adsorption at different temperatures displays non-hysteretic step-shaped isotherms for F₄-MIL-53(Al), generated by a CO₂-induced structural expansion also confirmed by in situ PXRD analysis. These findings highlight the crucial role of fluorination in tuning host–guest interactions, modifying water affinity while preserving and revisiting dynamic porosity and, more broadly, provide new insights into the molecular-level design of responsive fluorinated MOFs for gas separation and storage.

---

Introduction

The ability of a solid to reversibly switch between contracted and expanded phases, commonly referred to as ‘breathing’ or ‘sponge-like’ behaviour, is a remarkable feature of Soft Porous Crystals (SPCs),¹ a class of flexible crystalline porous materials that undergo phase changes in response to external stimuli such as gas adsorption, temperature and pressure.^2,3 Among SPCs, Metal–Organic Frameworks (MOFs) have emerged as a class of materials capable of exhibiting significant structural transformations. Among these materials, MIL-53 has been established as the prototype of the breathing MOF family, laying the foundation for decades of research into dynamic porosity.^1,4–10

The relevance of flexible materials extends far beyond fundamental structural chemistry. For example, adsorption-induced flexibility of MOFs has major implications for gas separation applications and energy-efficient adsorption processes. Indeed, breathing is often reflected in peculiar S-shaped isotherms, which possibly enhance working capacity and reduce energy consumption in pressure- and temperature-swing adsorption processes.^1,11 Therefore, controlling and predicting such transformations remains an extraordinary challenge.¹²

First reported by Serre et al. in 2002,¹³ the MIL-53 framework consists of 1,4-benzenedicarboxylate (bdc) linkers bridging trivalent metal cations (M^III). Among its many variants, MIL-53(Al) is the most extensively studied,¹⁴ due to Al abundance (8.2 wt% of the Earth's crust) and low cost (1.8 € per kg).¹⁵ Moreover, Al^III forms strong bonds with carboxylic linkers, allowing the design of MOFs with exceptional thermal and hydrolytic stability. This makes MIL-53(Al) a rare example of MOF produced on a large scale and commercialised by BASF under the trade name of Basolite A100.¹⁶

The flexibility of MIL-53(Al) has been extensively studied, revealing its ability to undergo phase transitions in response to external stimuli.^8,17–20 For instance, exposure to atmospheric moisture triggers a contraction from a large-pore (lp) to a narrow-pore (np) phase due to the formation of hydrogen bonds between H₂O molecules and the MOF framework (the structural model of the np phase is reported in Fig. S1A).¹⁴ This transition results in a unit cell volume contraction up to 40%. Beyond water, other guest molecules, including SO₂,¹⁷ Xe,¹⁸ CH₄ ¹⁹ and CO₂,^19,21 have been shown to induce a similar flexible behaviour.

Despite its well-documented flexibility, precise control over the breathing behaviour of MIL-53(Al) remains elusive. A powerful strategy to tune flexibility in MOFs is chemical functionalisation of the linker. In 2015, Munn et al. demonstrated how incorporating various substituents on the bdc linker (–CH₃, –Cl, –Br, –NH₂, –NO₂, –(OH)₂, –CO₂H) profoundly alters the structural response of MIL-53(Al),²² revealing a direct correlation between electronic effects and steric constraints of the substituents and host–guest interactions.²²

Among the countless possible linker modifications, fluorination emerges as a particularly intriguing route for tuning the flexibility of MOFs. The high electronegativity and the strong electron-withdrawing character of fluorine atoms significantly modify the textural properties, the framework rigidity and the adsorption properties of fluorinated MOFs. Indeed, it has been proved how fluorination of bdc enhances torsional flexibility of the ligand by altering the twist angle between the benzene ring and the carboxylate groups of many MOFs.²³ Furthermore, Nagaoka et al. demonstrated with DFT methods that the planarity of the bdc system mainly depends on electrostatic interactions (e.g. H-bonding). In bdc, the strong intramolecular electrostatic interaction between the oxygen of the carboxylate groups and the hydrogen atoms on the phenyl ring stabilises the planarity of the moiety, while the repulsive interaction between oxygen and fluorine atoms causes the rotation of the phenyl ring out of the carboxylate plane.²⁴

Van der Voort and co-workers synthesised and investigated mono- and para-difluorinated MIL-53(Al), proving that even a single fluorine substitution substantially modifies the mechanism of breathing. Unlike pristine MIL-53(Al), which contracts upon water adsorption, monofluorinated and difluorinated MIL-53(Al) do not show water-induced phase transitions.^25,26

Recently, we synthesised a fully fluorinated MIL-53(Al) variant [F₄-MIL-53(Al)] using a solvent-free approach.²⁷ Our preliminary studies disclosed a reversible temperature-induced transition (np_F4 → lp_F4) of F₄-MIL-53(Al), occurring at 513 K with minimal hysteresis (ΔT ∼5 K). This is in contrast with the behaviour of MIL-53(Al), which exhibits a slow and broadly hysteretic (ΔT ∼ 200 K) breathing transition, as reported by Liu and co-authors.⁸ In general, the true impact of fluorination on adsorption-induced flexibility remains largely unexplored. Traditional ex situ techniques often fail to capture transient intermediate states and dynamic structural transformations, leading to an incomplete understanding of host–guest interactions. For this reason, we present here an advanced, multi-technique in situ investigation of the response of F₄-MIL-53(Al) to the adsorption of different probes, i.e., H₂O, CO₂, Ar, N₂ and CO, in comparison with that of MIL-53(Al). Our approach exploits in situ and ex situ Powder X-ray Diffraction (PXRD) to track real-time structural evolution, in situ Infrared (IR) and Solid State Nuclear Magnetic Resonance (SS-NMR) spectroscopies to acquire molecular-scale insights into framework-adsorbate interactions, and volumetric adsorption analysis to identify and quantify adsorption-driven transitions. By integrating these complementary techniques, we disclosed the complex multi-step flexible response of F₄-MIL-53(Al) to adsorption.

Results and discussion

H₂O-induced flexibility

Water affinity is a key point in MOFs characterisation. Indeed, water can destabilise many MOFs or compete with other gases, lowering the adsorption performances.²⁸ In flexible frameworks, water adsorption/desorption can induce phase transitions.^5,14,29 Therefore, understanding MOFs structural response to relative humidity and identifying the material phase before sorption measurements that require a pre-activation step (i.e. involving water removal) is essential.

The crystallinity of MIL-53(Al) [labelled hereafter as H₄-MIL-53(Al)] and F₄-MIL-53(Al) after the synthesis was checked by PXRD (Fig. S2). Both frameworks are stable in their respective np phases prior to any activation treatment.^14,27 The unit cell volume of the np_F4 phase is 1292.1 ų, about 30% more expanded than the np_H4 phase of H₄-MIL-53(Al) (863.9 ų).^14,27 The structural models of both MOFs are reported in Fig. S1A and B.

The morphology of the two MOFs was evaluated by Field Emission Scanning Electron Microscopy (FE-SEM). At the micrometre scale, H₄-MIL-53(Al) exhibits well-defined prismatic crystals with regular and smooth surfaces and big particle size, indicating a high crystallinity. In contrast, F₄-MIL-53(Al) displays a more disordered morphology, characterized by the agglomeration of thicker, plate-like crystallites. This difference is further highlighted at higher magnifications, where the roughness of the F₄-MIL-53(Al) surface becomes clearly visible (Fig. S3).

The H₂O-induced flexibility of H₄-MIL-53(Al) is well-known. SS-NMR and PXRD studies discovered that water physisorbed from the atmosphere interacts with different sites from both the organic linker and the bridging Al–OH–Al groups in the inorganic units.^14,30,31 However, the SS-NMR study concluded that the hydrogen atoms of H₂O interact with the carboxylate groups of the framework through H-bonds, triggering the contraction of the rhombic channels of H₄-MIL-53(Al), affording the narrow pore np_H4 phase. Upon water removal, an expansion of the pores is observed, leading to the large pore lp_H4 phase.¹⁴

In contrast, PXRD patterns of as synthesised and evacuated F₄-MIL-53(Al) (Fig. S4) indicate that no transition occurs upon removal of water adsorbed from the atmosphere (Relative Humidity – RH < 70%). The refinement of the stable phases before and after activation was studied in a previous work, revealing the same orthorhombic np_F4 phase in both cases.²⁷

In situ IR spectra of H₄-MIL-53(Al) and F₄-MIL-53(Al) were recorded during evacuation at room temperature (RT) (Fig. S5). The analyses reveal two different responses upon desorption of atmospheric water. The IR profile of F₄-MIL-53(Al) remains unaltered upon evacuation, except for the disappearance of the band at 1145 cm⁻¹, associated to the δ(μ-OH) bending mode of the inorganic unit in interaction with physisorbed water.³² The low intensity of this band and of the broad signal in the OH stretching region (inset of Fig. S5A) indicates that F₄-MIL-53(Al) adsorbs a low quantity of water when exposed to atmospheric moisture. In contrast, the IR spectrum of H₄-MIL-53(Al) undergoes significant changes upon activation (Fig. S5B). In particular, the strong decrease of the intensity of the band between 1200 cm⁻¹ and 1050 cm⁻¹, associated with the same δ(μ-OH) mode,³³ can be attributed to the presence of a larger amount of H₂O inside the channels of H₄-MIL-53(Al) before activation. Again, the larger quantity of physisorbed water can be detected by observing the –OH stretching region (inset of Fig. S5B).

The behaviour upon evacuation significantly differs between H₄-MIL-53(Al) and F₄-MIL-53(Al). When channels are empty, H₄-MIL-53(Al) exhibits the lp_H4 phase while fluorine atoms decorating the rhombic channels of F₄-MIL-53(Al) stabilise the evacuated structure in the closed np_F4 phase.

Van der Voort et al. previously demonstrated that mono- or di-fluorination of the bdc linker strongly affects the H₂O induced breathing behaviour of MIL-53(Al).^25,26 Specifically, their studies reported that mono-fluorinated MIL-53(Al) stabilises the lp phase in both hydrated (from moisture) and dehydrated phases, while pF₂-MIL-53(Al) favours the np phase regardless of hydration state.^25,26 The behaviour of F₄-MIL-53(Al) resembles that of pF₂-MIL-53(Al), suggesting that the presence of more than two fluorine atoms further stabilises the contracted np phase and prevents the expansion of the cell upon removal of hydrogen-bonded water molecules. However, it is worth noticing that fluorinated MIL-53(Al) MOFs reported by Van der Voort et al. were synthesised using dimethylformamide (DMF) rather than water and this may significantly alter the flexible behaviour of the resulting materials.³⁴

The water affinity of F₄-MIL-53(Al) was assessed by gravimetric adsorption/desorption isotherms collected at 303 K (Fig. 1A). The presence of four fluorine atoms significantly enhances the material hydrophobic character within the 0–70% RH range, as evidenced by a low H₂O uptake of less than 2 mmol g⁻¹. This interesting property was recently noticed by Zhang et al., who exploited it to facilitate the selective permeation of n-butanol from water using a mixed matrix membrane containing F₄-MIL-53(Al) as a filler.³⁵

Fig. 1 (A) Gravimetric H₂O adsorption (full symbols) and desorption (empty symbols) isotherms collected at 303 K on H₄-MIL-53(Al) (brown) and F₄-MIL-53(Al) (blue). (B) Comparison of PXRD patterns of F₄-MIL-53(Al) at different RH: powder exposed to ambient atmosphere (black, as-synthesized. RH < 75%), powder exposed for several days at 75% of RH (green and light blue), powder wetted with water (blue RH ≈ 100%). The dashed lines refer to the 110 reflection which is diagnostic of the shrinkage and breathing behaviour of F₄-MIL-53(Al). (C) Comparison between the PXRD patterns of F₄-MIL-53(Al) after activation (red), soaked in water for 10 h (dark blue) and after water removal with vacuum at 298 K (pale red).

Interestingly, although water uptake remains minimal at RH levels below 70%, the isotherm reveals a flexible response of F₄-MIL-53(Al) when the material is exposed to RH values above 70%.³⁵

In particular, as synthesised F₄-MIL-53(Al) is in the closed phase (np_F4). Upon hydration at 75% RH for 1 day (RH 75%_1d), a new, more expanded phase is formed and coexists with the np_F4 one, as testified by the appearance of the main reflection at lower 2θ values (8.81° 2θ). After 4 days at 75% RH (RH 75%_4d), the pattern only shows the reflections of the new expanded phase (, where “w” stands for “water”), which is now the prevailing phase. After 7 days (RH 75%_7d), a second, even larger pore phase is detected, indicated by the appearance of a new peak at 2θ = 8.32°. This larger phase (w_lp_F4) coexists with the one for an extended period. Indeed, even after 28 days at 75% RH (RH 75%_28d), a mixture of and w_lp_F4 phases is observed. The single final w_lp_F4 phase is only detected when the MOF powder is wetted with liquid water. The whole PXRD sequence shows that the transition occurs with a relatively fast kinetics (just one day). The subsequent conversion progresses much more slowly, probably requiring a longer time under these RH conditions or higher RH values for completion, as testified by the pattern collected on the wet sample (blue pattern in Fig. 1B).

To identify the w_lp_F4 crystal structure, PXRD measurements were also performed on F₄-MIL-53(Al) after soaking in water for 10 h using a hermetic chamber to avoid evaporation. The PXRD pattern (Fig. 1C, dark blue) displays a dramatic shift of the first diffraction peak towards lower 2θ angles compared to the activated structure (from 9.62° to 8.32°), corresponding to a d₁₁₀ spacing of 10.62 Å (see Fig. S1C), associated with the squaring of the rhombus of the wine-rack structure of F₄-MIL-53(Al). The unit cell can be refined using the Le Bail method in the monoclinic crystal system (space group C2/c, Fig. S6 and Table S1). The unit cell shows a doubling of the c axis with lattice parameters a = 15.5820(8) Å, b = 15.8985(9) Å and c = 13.2991(7) Å, angle β = 114.688(2)° and V = 2993.5(3) Å. The PXRD pattern collected after evacuation of the water-soaked sample at RT proves the reversibility of the phase transition (Fig. 1C, pale red). It is important to note that, considering the cell volume, the final uptake of 18 mmol g⁻¹ (see isotherm in Fig. 1A) cannot be completely attributed to water molecules inside the channels of the w_lp_F4 phase. Instead, the presence of water on the MOF external surface must be hypothesized.

Multinuclear (¹H, ¹⁹F, ¹³C, and ²⁷Al) SS-NMR spectroscopy was also employed to investigate the structure of F₄-MIL-53(Al) and its changes upon H₂O adsorption in comparison with H₄-MIL-53(Al). Activated F₄-MIL-53(Al) shows single ¹H and ¹⁹F isotropic signals from μ₂-OH groups (2.97 ppm) and linker fluorine atoms (−145.1 ppm), respectively (Fig. 2A, B and Tables S2, S4). In the ¹⁹F–¹³C CP-MAS spectrum, three signals are observed at 117.2, 145.4, and 166.2 ppm ascribable, respectively, to quaternary, fluorinated, and carboxylic carbons of the linkers (Fig. 2C and Table S4). The ²⁷Al DE-MAS spectrum shows a signal typical of the octahedral environment of Al in the [AlO₄(OH)₂] centres, with isotropic chemical shift, δ = 3.37 ppm, quadrupolar coupling constant, C_Q = 9.24 MHz, and asymmetry parameter, η_Q = 0.07 (Fig. 2D and Table S3).

Fig. 2 (A) ¹H DE-MAS, (B) ¹⁹F DE-MAS, (C) ¹⁹F–¹³C CP-MAS, and (D) ²⁷Al DE-MAS NMR spectra of activated, hydrated (75% RH for 11 days), wet, and CO₂-loaded F₄-MIL-53(Al). Asterisks in ²⁷Al DE-MAS NMR spectra indicate signals of impurities.

¹H, ¹⁹F, and ²⁷Al spectra were collected during the hydration of activated F₄-MIL-53(Al), first at 75% RH and then at 100% RH, to delve into the and transitions, respectively (Fig. 2, S7 and S8).

During the hydration at 75% RH, the ¹H signal first broadens and moves to higher chemical shift. After three days, two peaks are observed in the spectrum, at 3.75 and 5.70 ppm, the intensities of which progressively increase by prolonging the hydration time, reaching an 8 : 1 ratio at equilibrium after 11 days (Fig. S7A and Table S2). Since, according to the sample weight, 4 H₂O molecules per formula were adsorbed by F₄-MIL-53(Al) at this stage, in agreement with the H₂O adsorption isotherm (Fig. 1A), the signal at 5.70 ppm is ascribed to μ₂-OH and that at 3.75 ppm to H₂O protons.³⁶ The ¹⁹F isotropic signal slightly shifts (from −145.1 to −144.7 ppm) at the beginning of the hydration. Then, a second broad signal appears at −139.2 ppm, which increases in intensity as the hydration progresses, while the first signal disappears (Fig. S7B and Table S4). The ²⁷Al signal shifts and broadens during the hydration at 75% RH, showing a superposition of two sub-spectra in the intermediate stages (Fig. S8) and a single signal with δ = 5.67 ppm, C_Q = 10.85 MHz, and η_Q = 0.07 when the equilibrium is reached (Fig. 2D and Table S3). The evolution of the ¹H, ¹⁹F and ²⁷Al spectra with the hydration time indicates that the phase progressively forms at the expense of the np_F4 one. The ¹H and ²⁷Al NMR data suggest that, in the phase, the μ₂-OH groups strongly interact with H₂O molecules. The δ and C_Q values of hydrated F₄-MIL-53(Al) are similar to those found for hydrated H₄-MIL-53, here (Fig. S9C and Table S3) and in the literature,^{14,30,37–40} but η_Q is smaller (0.07 instead of 0.14). These findings indicate that, while the interaction of H₂O with μ₂-OH groups changes the charge distribution on Al along the Al–OH direction for both MOFs, thus increasing δ and C_Q, only for H₄-MIL-53(Al) a change of symmetry of the Al environment in the coordination plane of carboxylate groups occurs, resulting in an increase of η_Q.^37,38 These findings are further corroborated by the ¹³C NMR spectra (Fig. 2C, S9B and Table S4). Indeed, only slight changes of carbon chemical shifts are observed upon hydration for F₄-MIL-53(Al), whereas for H₄-MIL-53(Al) a considerable shift of the carboxylate carbon signal is associated to interactions with H₂O and structural rearrangement at the lp_H4 → np_H4 transition.^14,39–41 Conversely, hydration of F₄-MIL-53(Al) results in a change of the ¹⁹F isotropic chemical shift (Fig. 2B and Table S4), suggesting interactions between fluorine atoms and H₂O and/or a conformational change of the linker.

Slight changes are detected in the ¹⁹F SS-NMR spectrum after further exposing the F₄-MIL-53(Al) sample to 100% RH for two days (Fig. S7B), indicating that no significant structural modifications occur at the transition. Indeed, this spectrum matches the one recorded on wet F₄-MIL-53(Al) (Fig. 2B, S7B and Table S4) exhibiting the w_lp_F4 phase according to PXRD data. On the other hand, the ¹H DE MAS spectrum presents some changes upon further hydration. After 5H₂O molecules per MOF formula are adsorbed by F₄-MIL-53(Al) based on the sample weight, three signals are observed with 1 : 8 : 2 relative intensities, ascribable to μ₂-OH (5.60 ppm) and to H₂O molecules inside (3.85 ppm) and outside (4.85 ppm) the MOF channels (Fig. S7A and Table S2). This indicates that only 4 H₂O molecules per formula can be accommodated within the F₄-MIL-53(Al) channels, the excess water residing on the MOF external surface.

Evaluation of textural properties

Volumetric adsorption isotherms were measured to assess the textural properties of F₄-MIL-53(Al) and H₄-MIL-53(Al). Initially, N₂ at 77 K was used. However, due to its quadrupolar moment and its ellipsoidal shape, N₂ is not an ideal probe for surfaces decorated with functional groups. Ar at 87 K, with its spherical shape, inert nature, and smaller kinetic diameter, offers a better alternative.^42,43 Adsorption/desorption isotherms are shown in Fig. S10, with BET equation fittings in Fig. S11 and S12. Both samples exhibit Type I isotherms.⁴³ The specific surface area (SSA) of H₄-MIL-53(Al), determined with the two probes, is 1590 m² g⁻¹ (N₂) and 1437 m² g⁻¹ (Ar), respectively, consistent with literature values for this MOF.¹⁴ For F₄-MIL-53(Al), SSA values are 1082 m² g⁻¹ (N₂) and 1025 m² g⁻¹ (Ar). Its lower SSA compared to the protonated analogue can be due to the presence of bulkier fluorine atoms oriented towards the pore centre. Indeed, when normalised by molecular mass per unit formula, the surface areas of the two MOFs in m² mmol⁻¹ become comparable (299 m² mmol⁻¹ for H₄-MIL-53 against 287 m² mmol⁻¹ for F₄-MIL-53). The difference in SSA values derived from N₂ and Ar isotherms in both samples is likely due to the quadrupolar nature of N₂, which leads to an overestimation of the surface area.⁴⁴

Adsorption-induced flexibility at cryogenic temperatures

Surprisingly, the Ar and N₂ isotherms of F₄-MIL-53(Al) in a semi-logarithmic scale reveal a peculiar multi-step trend, whereas a single step due to micropore filling is observed for H₄-MIL-53(Al) (Fig. 3). Fluorinated micropores are filled in the 0–0.01 relative pressure (p/p₀) range via a multi-step mechanism, with pressure thresholds for each step varying slightly with the probe due to differences in gas-framework interactions. Considering water adsorption results, the existence of multiple intermediate phases linking the initial np_F4 phase [activated F₄-MIL-53(Al)] to the lp_F4 phase can be inferred. Recently, molecular simulations on H₄-MIL-53(Al) disclosed a np_H4 → lp_H4 transition induced by Ar adsorption at cryogenic temperatures, but the presence of multiple thermodynamically stable phases was excluded.⁴⁵

Fig. 3 Adsorption (full symbols) and desorption (empty symbols) isotherms in semi-logarithmic scale measured using Ar at 87 K (dark green hexagons), N₂ at 77 K (black squares) and CO at 77 K (purple triangles). (A) H₄-MIL-53(Al); (B) F₄-MIL-53(Al).

A flexible response of F₄-MIL-53(Al) is also observed upon CO adsorption at 77 K. CO was selected as a more interacting probe, with a kinetic diameter similar to that of N₂ (3.8 Å and 3.6 Å, respectively). CO produces a multi-step isotherm (Fig. 3B) with fewer steps than the other probes. To test whether highly interactive molecules trigger breathing in H₄-MIL-53(Al), an attempt to collect a CO isotherm was made. However, CO diffusion was too slow, and after ca. 100 h, only a few data points were collected. Notably, even the N₂ isotherm of H₄-MIL-53(Al) takes ∼55 h, confirming slow kinetics for larger molecules at cryogenic temperatures. However, for CO, its stronger interaction with H₄-MIL-53(Al) could be responsible for excessively long equilibration times, making the measurement experimentally unfeasible.

CO₂ – induced flexibility

Boutin et al.¹⁹ reported that H₄-MIL-53(Al) undergoes two distinct phase transitions induced by CO₂ between 273 and 318 K: lp_H4 → np_H4 below 1 bar and np_H4 → lp_H4 at higher pressures, as confirmed by our experimental data (Fig. 4A and B). In contrast, F₄-MIL-53(Al) displays Langmuir type isotherms below 1 bar in the same temperature range (Fig. 4C), consistent with a “classical” micropore filling. A clear step in the sorption profile is only observed at higher CO₂ pressures (Fig. 4D), with the step moving to lower pressure as the temperature decreases (from 8 bar at 313 K to 1 bar at 248 K). Plotting all CO₂ isotherms as a function of p/p_o, it can be observed that the step for each curve occurs between 0.07 and 0.1 p/p_o and none of the explored temperatures reach this p/p_o interval up to 1 bar (Fig. S14B). However, at 195 K the saturation pressure (p_o) of solid CO₂ is around 1 bar (1.9 bar considering CO₂ inside microporous channels as a supercooled liquid). Therefore, a larger range of relative pressure is explored, and the step is observed below 1 bar (Fig. S13). This step occurs at very low p/p_o (∼0.03), indicating a stronger interaction between CO₂ and the MOF at this temperature. Isotherms reported as a function of CO₂ relative pressure and the corresponding derivatives are shown in Fig. S14A. A linear trend between the temperature and the relative pressure at which the gate opening occurs is evident (Fig. S14C), allowing the estimation of the breathing pressure at a given temperature.

Fig. 4 CO₂ adsorption (full circles) and desorption (empty circles) isotherms performed at different temperatures. (A) H₄-MIL-53(Al) 0–1 bar; (B) H₄-MIL-53(Al) 0–16 bar; (C) F₄-MIL-53(Al) 0–1 bar; (D) F₄-MIL-53(Al) 0–16 bar.

The CO₂-adsoprtion behaviour of F₄-MIL-53(Al) at room temperature differs from that reported by Van der Voort et al. for both F₁- and paraF₂-MIL-53(Al),^25,26 which exhibited Langmuir-type CO₂ adsorption isotherms at 303 K in the 0–40 bar pressure range. The shape of the isotherms suggests that both partially fluorinated MOFs remain in their activated phases upon CO₂ adsorption. This is in contrast with the behaviour of F₄-MIL-53(Al), which presents an S-shaped isotherm under the same operating conditions, indicating a flexible response of the perfluorinated framework. However, since the breathing behaviour of MIL-53 has been shown to be highly sensitive to the synthetic procedure,³⁴ partially fluorinated MIL-53(Al) MOFs obtained by a hydrothermal rather than solvothermal (using DMF) synthetic procedure should be considered for a reliable comparison of the CO₂ induced flexibility in frameworks with different fluorination degrees.

Further investigation of the behaviour of F₄-MIL-53(Al) upon CO₂ adsorption was performed using several techniques, including microcalorimetry, PXRD, and IR and SS-NMR spectroscopies.

For flexible structures, the adsorption enthalpy profile is characterised by endothermic phenomena due to structural transitions that counteract the exothermic adsorption process, resulting in a reduced neat total heat of adsorption.⁴⁶ The CO₂ adsorption enthalpy for F₄-MIL-53(Al) was directly measured at 273 K by adsorption microcalorimetry (Fig. 5A and B). At low CO₂ coverage, F₄-MIL-53(Al) displays values of heat of adsorption (Q_st) in the range of physisorption (30.6 kJ mol⁻¹). Q_st decreases as the pores (of the np_F4 phase) are loaded with CO₂ molecules (14.6 kJ mol⁻¹ at 2.8 mmol g⁻¹). Notably, Q_st falls below the liquefaction heat of CO₂ (17 kJ mol⁻¹), suggesting that a modest contribution of an endothermic process is occurring and lowering the observed Q_st. At loadings higher than 3 mmol g⁻¹ (p > 3 bar), Q_st suddenly rises to 20.5 kJ mol⁻¹. Indeed, at this pressure an (where c stands for “carbon dioxide”) transition occurs, as evidenced by the step observed in the CO₂ isotherm at 273 K (Fig. 4D) and the PXRD pattern at high CO₂ pressure (Fig. 5D). Unfortunately, since calorimetric measurements report the total heat at a given CO₂ pressure resulting from both exothermic and endothermic events, they do not provide direct information about the strength of the interaction between the adsorbate and adsorbent immediately after the transition. The structural response of the F₄-MIL-53(Al) framework to CO₂ loading was investigated by collecting in situ PXRD patterns at 273 K in vacuum and at different CO₂ pressures between 0.05 and 2.2 bar (Fig. 5C). The unit cell of activated F₄-MIL-53(Al) is refined by Le Bail method in the orthorhombic crystal system (space group Imam, Fig. S15 and Table S1) with lattice parameters a = 18.105(5) Å, b = 10.621(2) Å and c = 6.604(2) Å (V = 1269.7(5) ų), which closely resemble those of the as synthesised material (np_F4).²⁷ The increase of CO₂ loading up to 2 bar induces a small variation of the unit cell parameters, as clearly observed by the change of the 110 reflection at 2θ = 9.60° (Fig. 5C). This reflection, corresponding to a d₁₁₀ spacing of 9.20 Å, is associated to the 2D rhombohedral grid perpendicular to the channel axis (see Fig. S1C) and decreases to 8.97 Å at 0.5 bar (CO₂ uptake = 1.8 mmol g⁻¹), due to the newly formed CO₂ interactions with the framework. Contrarily, at higher loadings (p = 2.2 bar, CO₂ uptake = 2.6 mmol g⁻¹), an increase to 9.06 Å is detected, unveiling a homogeneous expansion of the unit cell of the framework (Fig. S16).

Fig. 5 CO₂ adsorption microcalorimetry and in situ PXRD patterns of F₄-MIL-53(Al) under controlled CO₂ pressure. (A) Volumetric adsorption isotherms of CO₂ collected up to 7 bar and 273 K coupled with microcalorimetry. The measurements were repeated three times on the same sample to reduce the experimental errors and prove reproducibility. (B) Isosteric heat of adsorption measured directly using adsorption microcalorimetry. (C) In situ PXRD patterns at low 2θ angles collected at 273 K at distinct CO₂ pressures. The main reflection corresponds to the 110 reflection and is diagnostic of the shrinkage and breathing behaviour of F₄-MIL-53(Al) upon CO₂ loading. (D) In situ PXRD patterns collected at 248 K on the activated sample, at 2.5 bar CO₂ loading, and after CO₂ removal. The asterisks indicate the peaks associated with a minor amount of c_lp phase.

Furthermore, to investigate the behaviour of F₄-MIL-53(Al) at higher CO₂ loadings, PXRD patterns were measured at 248 K and up to 2.5 bar (Fig. 5D). Remarkably, the diffraction pattern collected at 2.5 bar of CO₂ clearly shows a decrease in the symmetry of the unit cell, proving the occurrence of a phase transition induced by CO₂ from np_F4 to . The unit cell parameters are refined in the monoclinic crystal system, space group I2/a (a = 17.644(3) Å, b = 11.521(2) Å, c = 6.6026(8) Å and angle β = 90.892(5)°), showing a unit cell volume expansion to 1342.0(3) ų to accommodate the guest molecules and reaching a CO₂ loading value up to ∼4.7 mmol g⁻¹, as detected by the adsorption isotherm at 248 K (Fig. 4D). The phase transition is fully reversible, as demonstrated by the pattern collected on the sample after CO₂ removal under vacuum (Fig. 5D, S17–S19 and Table S1). To complete the series, the in situ PXRD experiment was also performed at 293 K, but even the highest dose of CO₂ (2.5 bar) was insufficient to trigger any expansion of the unit cell (2.5 bar is indeed before the inflection point of the isotherm). The PXRD series at 293 K shows a slight contraction of the cell parameters associated with the generation of CO₂-framework interaction, similar to the results obtained at 273 K (Fig. S20).

Finally, the PXRD patterns of the F₄-MIL-53(Al) phases obtained by exposing the material to CO₂, H₂O and temperature in different conditions are directly compared in Fig. S21. A qualitative comparison of the and patterns reveals some differences between the two phases. In particular, the most intense reflection, appearing at lower angles in the case of , indicates a slightly more expanded cell. Regarding the lp phases, even if it is not possible to directly compare the w_lp_F4 and the c_lp_F4 patterns, the presence of a c_lp_F4 phase is suggested by the weak signal appearing at low 2θ angles in the pattern acquired under 2.5 bar of CO₂ at 248 K (Fig. 5D); its complete detection was not achievable because of the instrumental limitations. Moreover, the more expanded monoclinic phase (w_lp_F4) forming when the material is super hydrated (water soaking) can be compared to the temperature-induced expanded monoclinic phase (T_lp_F4, where T stands for temperature), detectable upon heating the material above 500 K, as reported in our previous work.²⁷

In situ IR spectroscopy was also employed to characterise F₄-MIL-53(Al) under incremental doses of CO₂ at RT and at the nominal temperature of 195 K (Fig. S22). The IR profile of F₄-MIL-53(Al) exhibits two signals at 1283 cm⁻¹ and 1263 cm⁻¹, plus a shoulder around 1250 cm⁻¹ with in-scale intensities. These bands may correspond to the deformation and in-plane bending modes of the fluorinated ring.⁴⁷ Hoffman et al. reported similar vibrational modes for the phenyl ring of H₄-MIL-53(Al) around 1000 cm⁻¹.^9,48 As shown in Fig. S22B, no shift is detected when CO₂ is adsorbed at RT. Contrarily, at 195 K, these bands clearly change (Fig. S22B′). A spectral fitting of these signals on the activated sample and at the maximum CO₂ coverage (Fig. S22C, D and Table S5) indicates that the higher frequency band only decreases in intensity, while those at 1264 cm⁻¹ and 1253 cm⁻¹ shift to higher wavenumbers upon CO₂ adsorption. To confirm the sensitivity of these bands to framework mobility, the same IR setup was employed to dose N₂ and CO at 77 K, and H₂O at RT. Fig. S23 shows a shift of the most intense band at the maximum coverage of H₂O, CO, or N₂. Notably, the shift with H₂O is smaller, likely due to the slow transition kinetics, indicating that the complete phase is not reached. Another possibility is related to a temperature increase generated by the IR beam, locally decreasing the real RH value.

¹H, ¹⁹F, ¹³C, and ²⁷Al SS-NMR experiments were also performed on F₄-MIL-53(Al) and H₄-MIL-53(Al) samples loaded with 1 bar of ¹³CO₂ at 298 K. In agreement with the adsorption isotherms, only slight changes are detected in the spectra of CO₂-loaded F₄-MIL-53(Al) compared to the activated MOF (Fig. 2 and Tables S2–S4), indicating that no phase changes are observed upon CO₂ loading at 298 K. On the contrary, the adsorption of CO₂ considerably affects the ¹H, ¹³C, and ²⁷Al SS-NMR spectra of H₄-MIL-53(Al) (Fig. S9 and Tables S2–S4), in agreement with data reported by Paula,⁴⁰ due to the lp_H4 to np_H4 phase transition. Indeed, for CO₂-loaded H₄-MIL-53(Al), shifts of both the aromatic and μ₂-OH proton signals and of the carboxyl carbon signals are observed compared to the activated MOF, analogous to those found when the phase transition is induced by hydration (lp_H4 to np_H4). As far as the ²⁷Al spectrum is concerned, a slight variation in the isotropic chemical shift and C_q is observed, while η_Q shows a value similar to that observed for the hydrated sample (Table S3). Indeed, the alteration of δ and C_q values is most likely associated to changes in the charge distribution along the Al–OH axis, with μ₂-OH groups involved in the interactions with CO₂.^21,49 Conversely, the value of η_Q seems to be more influenced by the type of coordination of Al by carboxylate groups.^37,38

The different structures of CO₂-loaded F₄-MIL-53(Al) and H₄-MIL-53(Al) also result in a different dynamics of CO₂. As shown by the static ¹³C DE spectrum in Fig. S24, CO₂ (1 bar at 298 K) in F₄-MIL-53(Al) undergoes isotropic dynamics. Contrarily, for CO₂ in H₄-MIL-53(Al), ¹³C MAS and static spectra typical of anisotropic dynamics are reported in the literature at room temperature, both in the np phase at low CO₂ pressure (below 6 bar) and in the lp phase at higher pressure.^50–52 These findings indicate that, at 298 K and 1 bar, no strong specific interactions of CO₂ with the framework of F₄-MIL-53(Al) occur, at variance with H₄-MIL-53(Al) in which stronger interactions of CO₂ with μ₂-OH groups constrain CO₂ molecules into specific orientations and restrict their dynamics. The greater mobility of the adsorbate in the np_F4 phase of F₄-MIL-53(Al) may also be due to its 30% larger volume compared to the CO₂-induced np_H4 phase of H₄-MIL-53(Al).

Overview of the multi-step breathing behaviour of F₄-MIL-53(Al)

The multi-technique characterisation approach presented above provides key insights into the role of per-fluorination in modulating the breathing behaviour of MIL-53(Al). Beyond the previously reported temperature-induced breathing of F₄-MIL-53(Al),²⁷ which profoundly diverges from the temperature-response of the hydrogenated analogue, we uncovered additional pronounced differences in adsorption-induced dynamic porosity. As summarised in Fig. 6, both CO₂ and H₂O trigger an expansion of the initial np_F4 phase, associated to the elongation of the d₁₁₀ distance, which leads to a squaring of the pore shape (Fig. 6A). Moreover, as illustrated in the flow chart in Fig. 6B, the adsorption-induced breathing of F₄-MIL-53(Al) follows a more intricate, multi-step mechanism. In order to facilitate the comprehension of the discrepancies between the breathing mechanisms of H₄-MIL-53(Al) and F₄-MIL-53(Al), the schematic illustration of the flexible behavior of the hydrogenated MOF is reported in Fig. S25.

Fig. 6 (A) Schematic representation of the pore shape evolution of F₄-MIL-53(Al) from the np_F4 phase (rhombus shape, top) to the lp_F4 phases (square shape, bottom) induced by the elongation of the 110 distance (see table in panel A); (B) flow chart representing the different crystalline phases of F₄-MIL-53(Al) triggered by various stimuli (temperature, H₂O and CO₂ adsorption).

Our study reveals the fundamental role of fluorine atoms in modifying the adsorption properties of the MIL-53(Al) framework and, most important, in affecting its breathing behaviour. Even if the mechanistic role of fluorine atoms is not clear, the different flexibility of the fluorinated framework could be associated with the modifications introduced by fluorine atoms on the organic linkers. Indeed, Férey et al. demonstrated that the aromatic rings of H₄-MIL-53(Cr) undergo π flips about their C₂ symmetry axes with a normal Arrhenius process, thus triggering the transition. However, the rate and the activation energy of this motion are strongly dependent on the local geometry and electronic structure around the metal centre.^6,24,53,54 Fluorine substitution may affect both aspects, thus altering the overall flexible behaviour of fluorinated MIL-53.

Conclusions

This study provides the first comprehensive characterisation of the unique, multi-step adsorption-induced breathing behaviour of F₄-MIL-53(Al), a new per-fluorinated Al-based MIL-53 analogue recently synthesised by our group. Through a systematic multi-technique in situ investigation and by directly comparing F₄-MIL-53(Al) with its hydrogenated counterpart H₄-MIL-53(Al), we unveiled the profound impact of fluorination on framework flexibility, adsorption properties, and phase transitions, offering new insights into the structure–property relationships governing dynamic porosity in MOFs.

Specifically, we revealed that F₄-MIL-53(Al) undergoes complex phase transitions upon exposure to different adsorbates under various temperature and pressure conditions. More interacting species, such as H₂O and CO₂, trigger a cell expansion through a two-stage process, with the progressive conversion of the narrow pore np_F4 to intermediate and, eventually, to large pore lp_F4 phases, as humidity or CO₂ pressure increase. This behaviour markedly diverges from the one of the hydrogenated H₄-MIL-53(Al) analogue, for which the lp_H4 phase collapses to np_H4 upon adsorption of H₂O or CO₂ (with a back-transformation to lp_H4 only at elevated CO₂ pressures).

Beyond H₂O and CO₂, a multi-step breathing has also been observed for F₄-MIL-53(Al) upon adsorption of intermediate or weaker interacting probes (CO, N₂ and Ar), as highlighted by the unique multi-step adsorption isotherms measured at cryogenic temperatures.

Overall, this study highlights the critical role of ligand functionalisation (per-fluorination herein) in modulating the textural, adsorption and framework dynamic properties of MOFs, with an immediate impact on their macroscopic behaviour (e.g., water affinity) and possible applications. The study also underscores the importance of a comprehensive in situ experimental approach to capture the rich, and often unexpected, adsorption-driven structural evolution of flexible MOFs, paving the way for the rational design of a new generation of stimuli-responsive porous materials with properties finely tuned for targeted applications.

Experimental

Materials

All chemicals are commercially available and used as received from the specified vendors without further purifications. Aluminium nitrate nonahydrate (Al(NO₃)₃·9H₂O) was purchased from Honeywell Fluka™, tetrafluoroterephthalic acid (H₂-F₄bdc) from SiKÉMIA, terephthalic acid (H₂-bdc) and ¹³CO₂ (¹³C labeled CO₂; 99 at% ¹³C) from Sigma Aldrich, and acetone from Merck.

Synthesis

For the synthesis of H₄-MIL-53(Al), 0.57 g (3.4 mmol) of terephthalic acid (H₂-bdc) was dissolved in 9.8 mL of deionised water. In a second step, 2.55 g (6.8 mmol) of aluminium nitrate nonahydrate (Al(NO₃)₃·9H₂O) was added. The mixture was put in a Teflon hydrothermal reactor and kept at 493 K for 72 h. The precipitated white solid was then washed twice with deionised water for 15 minutes. To remove the traces of trapped linker, the powder was heated in a muffle furnace at 603 K for 24 h. Yield: 50%.

F₄-MIL-53(Al) was synthesised via a facile “solvent-free” method by mixing 3.75 g (10.0 mmol) of Al(NO₃)₃·9H₂O and 2.38 g (10.0 mmol) of tetrafluoroterephthalic acid (H₂-F₄bdc) in a polytetrafluoroethylene (PTFE) hydrothermal reactor, which was kept at 393 K for 24 h. No solvent was added to the reactor, other than the hydration water of the aluminium precursor. The obtained mixture was washed twice with water for 15 minutes and finally with acetone. The recovered product was dried overnight in a static oven at 353 K. Yield: 82%.

Field emission scanning electron microscopy (FESEM)

FESEM images of F₄-MIL-53(Al) and H₄-MIL-53(Al) were acquired using a TESCAN S9000G FE-SEM 3010 microscope (30 kV), equipped with a Schottky type FEG source. Prior to analysis the samples were metalized with a chromium layer.

Powder X-ray diffraction (PXRD)

Powder X-ray diffractograms were collected on a PANalytical X'Pert diffractometer using Cu K_α radiation emitted by a cathode operated at 45 kV and 40 mA. To check the purity of H₄-MIL-53(Al) and the lack of traces of linker inside the pores, two patterns were collected in reflection geometry (2θ range = 5–30°), one before and one after the treatment in the muffle furnace at 603 K for 24 h.

To check the phase of F₄-MIL-53(Al) before and after evacuation, two diffractograms were acquired in Debye Scherrer geometry (2θ range = 5–30°), one directly on as-synthesized F₄-MIL-53(Al), the other after the capillary was kept at 423 K for 2 h under dynamic vacuum to remove physisorbed water. To check the different crystalline phases of the MOF exposed to about 75% RH, the capillaries were placed into an airtight chamber for several days together with a saturated solution of NaCl.

Variable temperature powder X-ray diffraction patterns of F₄-MIL-53(Al) were collected with a Rigaku powder diffractometer, equipped with a low-temperature chamber (Anton Paar TTK 600) operating with a liquid nitrogen cooling system, using Cu-Kα radiation, 40 kV, 30 mA over a range for 2θ of 7.0–40.0° with a step size of 0.02° and a scan speed of 1.0° min⁻¹. The chamber was equipped with vacuum and pressure transducers, connected to a gas (CO₂) line and a two-stage rotary pump to control the atmosphere inside the chamber. The as-synthesized sample was degassed in situ under high vacuum (p < 1 × 10⁻² mmHg) at 423 K for 1 h. Then, the sample was cooled to the target temperature under vacuum and equilibrated for at least 30 min. The gas dosing was performed manually under isothermal conditions. The sample was equilibrated for 30 min and the PXRD diffraction patterns were collected. The dosing applied at the temperatures of 293 K, 273 K, and 248 K were reported in detail (Fig. S14, S17 and S18). The determination of the unit cell parameters, the identification of the space group, and the refinement of the unit cell parameters according to the Le Bail method were carried out with the EXPO2014 software.⁵⁵

In situ Fourier Transform Infrared (FT-IR) spectroscopy

FT-IR measurements were collected within the 5000–500 cm⁻¹ spectral range using a Bruker Vertex 70 spectrophotometer equipped with a MCT (mercury cadmium tellurium) cryogenic detector. The resolution of the reported IR spectra is 2.0 cm⁻¹ and an average of 32 scans was used to enhance the signal to noise ratio. Before the analysis, the sample, in form of self-supported pellet mechanically protected by a gold envelope, was inserted in a home-made quartz cell with KBr windows. For experiments carried out at cryogenic temperatures, a special IR cell that can be filled with liquid nitrogen to lower the temperature of the system at about 100 K was employed.

To evaluate the optimal activation conditions, multiple spectra were acquired during in situ degassing at room temperature (RT) by using a conventional high-vacuum glass line, equipped with mechanical and turbo molecular pumps (residual pressure p < 10⁻⁴ mbar). For F₄-MIL-53(Al), an additional ex situ treatment at 423 K for 2 h was necessary to completely remove water traces.

To investigate the interactions between the MOFs and different probes (i.e. CO₂, CO, N₂), the system was cooled down and multiple doses of gas (from 10⁻³ mbar to 10² mbar) were conveyed to the cell.

Low pressure gas sorption analysis

N₂ isotherms were collected at 77 K using a Micromeritics 3FLEX sorption analyser. The cooling bath was prepared using liquid nitrogen. For measurements on H₄-MIL-53(Al), 20 mg of sample was activated by heating overnight at 525 K (3 K min⁻¹) under dynamic vacuum, whereas in the case of F₄-MIL-53(Al), about 25 mg of powder was activated under dynamic vacuum at 423 K for 2 h (3 K min⁻¹). BET specific surface area (SSA) was evaluated following the Rouquerol consistency criteria⁵⁶ in the range 2 × 10⁻³ to 2 × 10⁻² p/p_o and 6 × 10⁻³ to 4 × 10⁻² p/p_o for H₄-MIL-53(Al) and F₄-MIL-53(Al), respectively.

For both materials, an Ar isotherm was collected at 87 K using a Micromeritics 3FLEX. The dewar was filled up with liquid argon. About 30 mg of sample was weighed and activated following the same procedures described above. The BET SSA was evaluated following the Rouquerol consistency criteria in the following ranges: 3 × 10⁻³ to 2 × 10⁻² p/p_o for H₄-MIL-53(Al); 7 × 10⁻³ to 8 × 10⁻² p/p_o for F₄-MIL-53(Al).

CO adsorption/desorption isotherms were acquired at 77 K on F₄-MIL-53(Al) using a Micromeritics ASAP2020 sorption analyser. The cooling bath was prepared using liquid nitrogen.

For both H₄-MIL-53(Al) and F₄-MIL-53(Al), CO₂ isotherms were collected at different temperatures in the 273–318 K range using a Micromeritics ASAP2020 sorption analyser. To keep isothermal conditions for each analysis, the sample was inserted in a home-made patented glass coating cell⁵⁷ in which a coolant or heating fluid, connected to a thermostatic bath (JULABO F25), can recirculate. About 60 mg of sample was weighed and activated following the already described procedures. A CO₂ isotherm was also collected at 195 K using a dry ice and isopropanol cooling bath.

High-pressure gas sorption analysis

High-pressure CO₂ and N₂ adsorption isotherms were measured on a Hiden IGA-001 gravimetric gas sorption analyser. The F₄-MIL-53(Al) sample was activated before the adsorption experiment at 423 K for 2 h under dynamic vacuum. The temperature of the analysis was controlled using a Julabo 600 F circulator.

Adsorption microcalorimetry

Gas adsorption isotherms up to 6 bar were collected on a volumetric adsorption analyser (Micromeritics ASAP 2050) in a high-pressure sample holder, while heat exchange was measured simultaneously with a Setaram μDSC7 Evo instrument at constant temperature (T = 273 K). The set-up allowed simultaneous determination of CO₂ adsorption isotherms and heat exchanged during the adsorption process at each adsorption step. Sorption-coupled microcalorimetry measurements were performed three times at 273 K, and the three different runs were averaged to reduce experimental errors. Data acquisition and processing were performed as detailed in the literature.^58,59

Water adsorption/desorption isotherms

Water adsorption/desorption isotherms were collected at 303 K using a dynamic vapour sorption (DVS) analyser within a 0–90 relative humidity (RH%) range and a vapor flow of 20 sccm (standard cubic centimetres). About 40 mg of F₄-MIL-53(Al) and H₄-MIL-53(Al) were activated before the analysis under dynamic vacuum for 2 h at 423 K and 473 K, respectively.

Solid state NMR spectroscopy

SS-NMR experiments were carried out on a Bruker Avance Neo 500 spectrometer working at Larmor frequencies of 500.13, 470.59, 130.32, and 125.77 MHz for ¹H, ¹⁹F, ²⁷Al, and ¹³C nuclei, respectively, equipped with a double-resonance H-F/X 4 mm CP/MAS probe head.

¹H and ¹⁹F spectra were recorded by Direct Excitation (DE) under magic angle spinning (MAS) conditions, accumulating 16 scans with a recycle delay between consecutive transients of 5–10 s and 3–15 s for ¹H and ¹⁹F, respectively, depending on the sample. 90° pulse durations of 3.0 and 3.2 μs were employed for ¹H and ¹⁹F, respectively. ¹⁹F–¹³C cross-polarization (CP) experiments were performed under MAS conditions and high power ¹⁹F decoupling using a contact time of 2 ms and a recycle delay of 2–4 s and accumulating 200–400 scans depending on the sample. ¹H–¹³C CP-MAS spectra were recorded under high power ¹H decoupling using a contact time of 2 ms and recycle delay of 2–6 s; 64–400 scans were accumulated depending on the sample. ¹³C DE-MAS spectra were acquired with or without high power ¹⁹F decoupling, using a recycle delay of 20 s and accumulating 128–24000 scans, depending on the sample. ²⁷Al DE-MAS spectra were acquired using an excitation pulse with a duration of 0.2 μs and accumulating 800–1600 scans with a recycle delay of 1 s.

If not otherwise stated, spectra were recorded at 298 K at a MAS frequency of 15 kHz using air as spinning gas. The chemical shift of all nuclei was referenced to the ¹³C signal of adamantane at 38.48 ppm and calculated from the same value for all the other nuclei using the unified scale recommended by IUPAC.⁶⁰

Line shapes of ²⁷Al DE-MAS spectra were analysed using the SOLA routine for line shape analysis implemented in Bruker TopSpin software.

Activated and CO₂ loaded samples were prepared using a home-made cell provided with a mechanical lever operated from outside enabling the capping of the rotor without disturbing the cell atmosphere. F₄-MIL-53(Al) and H₄-MIL-53 samples were activated by heating overnight under vacuum (0.1 mbar) at the temperature of 423 K the powder packed into the NMR rotor (4 mm external diameter) and closing the rotor under N₂ atmosphere. For the CO₂-loaded samples, the activated MOF was loaded with either CO₂ or ¹³CO₂ at 1 bar pressure and the rotor was capped under the gas atmosphere after equilibrium was reached.

Hydrated F₄-MIL-53(Al) was prepared by exposing the activated powder packed into the NMR rotor to either 75% RH or 100% RH atmosphere for different periods. At selected times, the sample was weighed to determine the amount of adsorbed water and the rotor was closed to perform measurements. Hydrated H₄-MIL-53(Al) was prepared by exposing the activated powder to the ambient atmosphere for 12 h.

A “wet” F₄-MIL-53(Al) sample was also prepared by adding 200 μL of H₂O to 100 mg of the MOF powder and packing the obtained hard paste into the NMR rotor.

Author contributions

V. G.: investigation, formal analysis, methodology, data curation, writing – original draft, writing – review & editing, visualisation; M. S. N.: investigation; J. P.: investigation, formal analysis, visualisation, writing – original draft, writing – review & editing; D. M. V.: investigation, formal analysis; F. N.: investigation, formal analysis, visualisation, writing – original draft; A. R.: investigation; S. B.: supervision; S. G.: supervision, resources; M. T.: formal analysis, writing – review & editing; M. S.: supervision, data curation; A. C.: supervision, writing – review & editing, resources; L. C.: investigation, formal analysis, visualisation, writing – original draft, supervision, resources, writing – review & editing, funding acquisition; F. C.: conceptualisation, writing – review & editing, supervision, resources, funding acquisition; V. C.: conceptualisation, writing – review & editing, supervision, funding acquisition, froject administration.

Conflicts of interest

There are no conflicts of interest to declare.

Data availability

All data for this article are available at Zenodo at https://zenodo.org/communities/dam4co2/ with the identifier https://doi.org/10.5281/zenodo.15429774. Some data supporting this article have been included as part of the SI (all isotherms in AIF format).

Supplementary information is available: structural models of hydrated H₄-MIL-53(Al) and F₄-MIL53(Al); PXRD patterns of H₄-MIL-53(Al) and F₄-MIL-53(Al) after synthesis; FE-SEM images of H₄-MIL-53(Al) and F₄-MIL-53(Al); PXRD patterns of F₄-MIL-53(Al) as-synthesised and evacuated under dynamic vacuum for 2h at 423 K; in situ IR spectra of H₄-MIL-53(Al) and F₄-MIL53(Al) during evacuation at room temperature; Le Bail refinement of the PXRD diffraction pattern of water-soaked F₄-MIL-53(Al); Le Bail refinement outputs of the PXRD patterns of F₄-MIL-53(Al) under different temperature/adsorption conditions; ¹H and ¹⁹F DE-MAS NMR spectra recorded at different times during the hydration of F₄-MIL-53(Al) at 75% and at 100% RH; isotropic chemical shift and number of hydrogens per formula for the different signals in the ¹H NMR spectra of activated, hydrated, and CO₂-loaded H₄-MIL-52(Al) and F₄-MIL-53(Al); ²⁷Al DE-MAS NMR spectra recorded at different times during the hydration of F₄-MIL-53(Al) at 75% RH; values of isotropic chemical shift, quadrupolar coupling constant and asymmetry parameter from ²⁷Al NMR spectra of H₄-MIL-52(Al) and F₄-MIL-53(Al); ¹H DE-MAS, ¹H-¹³C CP-MAS, and ²⁷Al DE-MAS NMR spectra of activated, hydrated and CO₂-loaded H₄-MIL-53(Al); N₂ (77 K) and Ar (87 K) adsorption/desorption isotherms of H₄-MIL-52(Al) and F₄-MIL-53(Al); BET fits of N₂ (77 K) and Ar (87 K) adsorption/desorption isotherms of H₄-MIL-52(Al) and F₄-MIL-53(Al); CO₂ adsorption/desorption isotherms of F₄-MIL-53(Al) at 195 K; CO₂ adsorption isotherms of F₄-MIL-53(Al) measured at different temperatures and relation of gate-opening with temperature; Le Bail refinement of the PXRD diffraction pattern of F₄-MIL-53(Al) collected under dynamic vacuum at 273 K; CO₂ adsorption/desorption isotherms of F₄-MIL-53(Al) collected at 273 K and in situ PXRD collected at significant CO₂ pressure values; Le Bail refinement of the PXRD diffraction pattern of F₄-MIL-53(Al) collected under dynamic vacuum and under 2.5 bar of CO₂ at 248 K; CO₂ adsorption/desorption isotherms of F₄-MIL-53(Al) collected at 248 K and in situ PXRD collected at significant CO₂ pressure values; CO₂ adsorption/desorption isotherms of F₄-MIL-53(Al) collected at 298 K and in situ PXRD collected at significant CO₂ pressure values; comparison of the PXRD patterns of the different phases of F₄-MIL-53(Al); CO₂ adsorption isotherms of F₄-MIL-53(Al) at room temperature and 195 K and in situ IR collected at significant CO₂ pressure values at both temperatures (including bands fit and related parameters); N₂ (77 K), CO (77 K) and H₂O (298 K) adsorption isotherms of F₄-MIL-53(Al) and in situ IR collected at significant adsorbent pressure values; static ¹³C DE NMR spectrum of ¹³CO₂-loaded F₄-MIL-53(Al) at 298 K; schematic representation of the pore shape evolution of H₄-MIL-53(Al) under different stimuli. See DOI: https://doi.org/10.1039/d5ta04373e.

Acknowledgements

The authors (V. G., F. N., L. C., M. S., M. T., V. C., F. C.) thank the Italian MUR through the Project PRIN 2020 doMino (ref. 2020P9KBKZ). This research has received funding from the Project CH4.0 under the MUR program “Dipartimenti di Eccellenza 2023–2027” (CUP: D13C22003520001). V. G. acknowledges the Italian MUR for her PhD grant PON DM1061. V. G., M. S. N., S. B., M. T., M. S., L. C., F. C., V. C. have received funding from the European Union's Horizon Europe research and innovation program under grant agreement no. 101115488, within the EIC pathfinder project “DAM4CO₂”. D. M. V. acknowledges financial support by the Deutsche Forschungsgemeinschaft (MO 5091/2-1).

References

1. A. Schneemann, V. Bon, I. Schwedler, I. Senkovska, S. Kaskel and R. A. Fischer, Chem. Soc. Rev., 2014, 43, 6062–6096.
2. G. Férey and C. Serre, Chem. Soc. Rev., 2009, 38, 1380–1399.
3. D. Maspoch, D. Ruiz-Molina, K. Wurst, N. Domingo, M. Cavallini, F. Biscarini, J. Tejada, C. Rovira and J. Veciana, Nat. Mater., 2003, 2, 190–195.
4. C. Volkringer, T. Loiseau, N. Guillou, G. Férey, E. Elkaïm and A. Vimont, J. Chem. Soc., Dalton Trans., 2009, 2241–2249.
5. S. Bourrelly, B. Moulin, A. Rivera, G. Maurin, S. Devautour-Vinot, C. Serre, T. Devic, P. Horcajada, A. Vimont, G. Clet, M. Daturi, J.-C. Lavalley, S. Loera-Serna, R. Denoyel, P. L. Llewellyn and G. Férey, J. Am. Chem. Soc., 2010, 132, 9488–9498.
6. D. I. Kolokolov, H. Jobic, A. G. Stepanov, V. Guillerm, T. Devic, C. Serre and G. Férey, Angew. Chem., Int. Ed., 2010, 49, 4791–4794.
7. S. Tomar and V. K. Singh, Mater. Today: Proc., 2021, 43, 3291–3296.
8. Y. Liu, J.-H. Her, A. Dailly, A. J. Ramirez-Cuesta, D. A. Neumann and C. M. Brown, J. Am. Chem. Soc., 2008, 130, 11813–11818.
9. A. E. J. Hoffman, J. Wieme, S. M. J. Rogge, L. Vanduyfhuys and V. Van Speybroeck, Z. Kristallogr. - Cryst. Mater., 2019, 234, 529–545.
10. J. M. Salazar, G. Weber, J. M. Simon, I. Bezverkhyy and J. P. Bellat, J. Chem. Phys., 2015, 142, 124702.
11. T. M. McDonald, J. A. Mason, X. Kong, E. D. Bloch, D. Gygi, A. Dani, V. Crocellà, F. Giordanino, S. O. Odoh, W. S. Drisdell, B. Vlaisavljevich, A. L. Dzubak, R. Poloni, S. K. Schnell, N. Planas, K. Lee, T. Pascal, L. F. Wan, D. Prendergast, J. B. Neaton, B. Smit, J. B. Kortright, L. Gagliardi, S. Bordiga, J. A. Reimer and J. R. Long, Nature, 2015, 519, 303–308.
12. Z. Zheng, A. H. Alawadhi, S. Chheda, S. E. Neumann, N. Rampal, S. Liu, H. L. Nguyen, Y. Lin, Z. Rong, J. I. Siepmann, L. Gagliardi, A. Anandkumar, C. Borgs, J. T. Chayes and O. M. Yaghi, J. Am. Chem. Soc., 2023, 145, 28284–28295.
13. C. Serre, F. Millange, C. Thouvenot, M. Noguès, G. Marsolier, D. Louër and G. Férey, J. Am. Chem. Soc., 2002, 124, 13519–13526.
14. T. Loiseau, C. Serre, C. Huguenard, G. Fink, F. Taulelle, M. Henry, T. Bataille and G. Férey, Chem.—Eur. J., 2004, 10, 1373–1382.
15. Y. Liu, Y. Li, F. Zuo, J. Liu, Y. Xu, L. Yang, H. Zhang, H. Wang, X. Zhang, C. Liu, Q. Li and H. Li, Small, 2022, 18, 2203236.
16. M. Gaab, N. Trukhan, S. Maurer, R. Gummaraju and U. Müller, Microporous Mesoporous Mater., 2012, 157, 131–136.
17. A. López-Olvera, J. A. Zárate, E. Martínez-Ahumada, D. Fan, M. L. Díaz-Ramírez, P. A. Sáenz-Cavazos, V. Martis, D. R. Williams, E. Sánchez-González, G. Maurin and I. A. Ibarra, ACS Appl. Mater. Interfaces, 2021, 13, 39363–39370.
18. A. Boutin, M.-A. Springuel-Huet, A. Nossov, A. Gédéon, T. Loiseau, C. Volkringer, G. Férey, F.-X. Coudert and A. H. Fuchs, Angew. Chem., Int. Ed., 2009, 48, 8314–8317.
19. A. Boutin, F.-X. Coudert, M.-A. Springuel-Huet, A. V Neimark, G. Férey and A. H. Fuchs, J. Phys. Chem. C, 2010, 114, 22237–22244.
20. P. Mishra, H. P. Uppara, B. Mandal and S. Gumma, Ind. Eng. Chem. Res., 2014, 53, 19747–19753.
21. C. Serre, S. Bourrelly, A. Vimont, N. A. Ramsahye, G. Maurin, P. L. Llewellyn, M. Daturi, Y. Filinchuk, O. Leynaud, P. Barnes and G. Férey, Adv. Mater., 2007, 19, 2246–2251.
22. A. S. Munn, R. S. Pillai, S. Biswas, N. Stock, G. Maurin and R. I. Walton, Dalton Trans., 2016, 45, 4162–4168.
23. Z. Hulvey, E. Ayala and A. K. Cheetham, Z. Anorg. Allg. Chem., 2009, 635, 1753–1757.
24. S. Pakhira, M. Takayanagi and M. Nagaoka, J. Phys. Chem. C, 2015, 119, 28789–28799.
25. S. Biswas, S. Couck, D. Denysenko, A. Bhunia, M. Grzywa, J. F. M. Denayer, D. Volkmer, C. Janiak and P. Van Der Voort, Microporous Mesoporous Mater., 2013, 181, 175–181.
26. S. Biswas, T. Rémy, S. Couck, D. Denysenko, G. Rampelberg, J. F. M. Denayer, D. Volkmer, C. Detavernier and P. Van Der Voort, Phys. Chem. Chem. Phys., 2013, 15, 3552–3561.
27. D. Morelli Venturi, V. Guiotto, R. D'Amato, L. Calucci, M. Signorile, M. Taddei, V. Crocellà and F. Costantino, Mol. Syst. Des. Eng., 2023, 8, 586–590.
28. K. Tan, N. Nijem, Y. Gao, S. Zuluaga, J. Li, T. Thonhauser and Y. J. Chabal, CrystEngComm, 2015, 17, 247–260.
29. M. Shivanna, A. A. Bezrukov, V. Gascón-Pérez, K. Otake, S. Sanda, D. J. O'Hearn, Q.-Y. Yang, S. Kitagawa and M. J. Zaworotko, ACS Appl. Mater. Interfaces, 2022, 14, 39560–39566.
30. C. Lieder, S. Opelt, M. Dyballa, H. Henning, E. Klemm and M. Hunger, J. Phys. Chem. C, 2010, 114, 16596–16602.
31. G. Ortiz, G. Chaplais, J.-L. Paillaud, H. Nouali, J. Patarin, J. Raya and C. Marichal, J. Phys. Chem. C, 2014, 118, 22021–22029.
32. B. Henry and A. Samokhvalov, Spectrochim. Acta, Part A, 2022, 267, 120550.
33. J. Liu, B. Li, V. Martins, Y. Huang and Y. Song, J. Phys. Chem. C, 2024, 128, 8007–8015.
34. W. P. Mounfield and K. S. Walton, J. Colloid Interface Sci., 2015, 447, 33–39.
35. H. Zhang, F. Xiao, H. Han and Y. Wu, J. Membr. Sci., 2025, 713, 123261.
36. Th. Emmler, S. Gieschler, H. H. Limbach and G. Buntkowsky, J. Mol. Struct., 2004, 700, 29–38.
37. A. E. Khudozhitkov, S. S. Arzumanov, A. V Toktarev, S. V Cherepanova, A. A. Gabrienko, D. I. Kolokolov and A. G. Stepanov, Phys. Chem. Chem. Phys., 2021, 23, 18925–18929.
38. B. Ibrahim, B. E. G. Lucier, J. Xu, P. He and Y. Huang, Can. J. Chem., 2015, 93, 960–969.
39. R. Giovine, C. Volkringer, J. Trébosc, J.-P. Amoureux, T. Loiseau, O. Lafon and F. Pourpoint, Acta Crystallogr., Sect. C: Cryst. Struct. Commun., 2017, 73, 176–183.
40. C. Paula, In situ solid-state NMR and XRD investigations of structural transformations in crystalline porous solids, PhD thesis, FAU University, 2021.
41. J. Tang, Y. Chu, S. Li, J. Xu, W. Xiong, Q. Wang and F. Deng, Chem.–Eur. J., 2021, 27, 14711–14720.
42. A. Datar, S. Yoon, L. C. Lin and Y. G. Chung, Langmuir, 2022, 38, 11631–11640.
43. M. Thommes, K. Kaneko, A. V. Neimark, J. P. Olivier, F. Rodriguez-Reinoso, J. Rouquerol and K. S. W. Sing, Pure Appl. Chem., 2015, 87, 1051–1069.
44. M. Thommes and C. Schlumberger, Annu. Rev. Chem. Biomol. Eng., 2021, 12, 137–162.
45. K. Stracke and J. D. Evans, J. Phys. Chem. C, 2025, 129, 3226–3233.
46. M. Cavallo, C. Atzori, M. Signorile, F. Costantino, D. M. Venturi, A. Koutsianos, K. A. Lomachenko, L. Calucci, F. Martini, A. Giovanelli, M. Geppi, V. Crocellà and M. Taddei, J. Mater. Chem. A, 2023, 11, 5568–5583.
47. F. Li, Z. Zheng, S. Xia and L. Yu, J. Mol. Struct., 2020, 1219, 128480.
48. A. E. J. Hoffman, L. Vanduyfhuys, I. Nevjestić, J. Wieme, S. M. J. Rogge, H. Depauw, P. Van Der Voort, H. Vrielinck and V. Van Speybroeck, J. Phys. Chem. C, 2018, 122, 2734–2746.
49. N. A. Ramsahye, G. Maurin, S. Bourrelly, P. L. Llewellyn, C. Serre, T. Loiseau, T. Devic and G. Férey, J. Phys. Chem. C, 2008, 112, 514–520.
50. Y. Zhang, B. E. G. Lucier and Y. Huang, Phys. Chem. Chem. Phys., 2016, 18, 8327–8341.
51. S. Bin Baek and H. C. Lee, Bull. Korean Chem. Soc., 2016, 37, 588–591.
52. T. Kurihara, Y. Sori, K. Ikeda and M. Mizuno, J. Phys. Chem. Lett., 2025, 16(19), 4683–4688.
53. S. Pakhira, RSC Adv., 2019, 9, 38137–38147.
54. N. Sinha, I. Deshpande and S. Pakhira, ChemistrySelect, 2019, 4, 8584–8592.
55. A. Altomare, C. Cuocci, C. Giacovazzo, A. Moliterni, R. Rizzi, N. Corriero and A. Falcicchio, J. Appl. Crystallogr., 2013, 46, 1231–1235.
56. J. Rouquerol, P. Llewellyn and F. Rouquerol, Stud. Surf. Sci. Catal., 2007, 160, 49–56.
57. V. Crocellà, C. Atzori, G. Latini, M. Signorile, A Kit for Volumetric Measurements of Gas Adsorption, WO Pat. 2021181211A1, 2021.
58. J. Perego, C. X. Bezuidenhout, A. Pedrini, S. Bracco, M. Negroni, A. Comotti and P. Sozzani, J. Mater. Chem. A, 2020, 8, 11406–11413.
59. W. K. Feldmann, K.-A. White, C. X. Bezuidenhout, V. J. Smith, C. Esterhuysen and L. J. Barbour, ChemSusChem, 2020, 13, 102–105.
60. R. K. Harris, E. D. Becker, S. M. Cabral De Menezes, R. Goodfellow and P. Granger, Concepts Magn. Reson., 2002, 14, 326–346.

---