Tailoring porosity and rotational dynamics in a series of octacarboxylate metal-organic frameworks

Significance

A family of stable porous materials incorporating organic linkers and Cu(II) cations is reported. Their pores can be altered systematically by elongation of the ligands allowing a strategy of selective pore extension along one dimension. These materials show remarkable gas adsorption properties with high working capacities for CH₄ (0.24 g g⁻¹, 163 cm³ cm⁻³ at 298 K, 5–65 bar) for the most porous system. The mechanism of rotation of the organic groups in the solid state has been analyzed by NMR spectroscopy and rotational rates and transition temperatures analyzed. Significantly, we show that framework dynamics can be controlled by ligand design only, and this paves the way to understanding the role of molecular rotors within these materials.

Abstract

Modulation and precise control of porosity of metal-organic frameworks (MOFs) is of critical importance to their materials function. Here we report modulation of porosity for a series of isoreticular octacarboxylate MOFs, denoted MFM-180 to MFM-185, via a strategy of selective elongation of metal-organic cages. Owing to the high ligand connectivity, these MOFs do not show interpenetration, and are robust structures that have permanent porosity. Interestingly, activated MFM-185a shows a high Brunauer–Emmett–Teller (BET) surface area of 4,734 m² g⁻¹ for an octacarboxylate MOF. These MOFs show remarkable CH₄ and CO₂ adsorption properties, notably with simultaneously high gravimetric and volumetric deliverable CH₄ capacities of 0.24 g g⁻¹ and 163 vol/vol (298 K, 5–65 bar) recorded for MFM-185a due to selective elongation of tubular cages. The dynamics of molecular rotors in deuterated MFM-180a-d₁₆ and MFM-181a-d₁₆ were investigated by variable-temperature ²H solid-state NMR spectroscopy to reveal the reorientation mechanisms within these materials. Analysis of the flipping modes of the mobile phenyl groups, their rotational rates, and transition temperatures paves the way to controlling and understanding the role of molecular rotors through design of organic linkers within porous MOF materials.

Metal-organic frameworks (MOFs) are an emerging class of porous, multifunctional materials showing great potential in a wide range of applications (1–7). Given their metal-organic hybrid nature, MOFs have an exceptionally high degree of structural diversity and tailorability (8). Thus, not only is the on-demand design of materials that incorporate pores of precise shapes and dimensions achievable, but also the inner surface of these materials becomes a platform for incorporating desirable functionality for target applications (9, 10). Within the field of gas storage, there is a strong correlation between the material porosity and the maximum adsorption capacity. A common strategy to increase porosity in MOFs consists of targeting a framework topology and systematically elongating the linkers to generate additional pore space. This approach has shown success in a number of MOF systems, but is not without drawbacks. For example, increases in porosity typically correlate to increases in pore diameters, which can be detrimental to the strength of host–guest interactions at low surface coverage (11). Also, often simple ligand elongation will ultimately lead to framework interpenetration with reduced porosity and/or stability (12–14). A powerful driver therefore exists to find the ideal compromise between high porosity and strong host–guest interactions over a wide range of pressures.

The use of rigid, highly connected linkers (e.g., with 6–8 coordinating functions) affords potentially a more robust and stable platform for the development of isoreticular porous materials. This strategy has been widely implemented for [Cu₂(O₂CR)₄] paddlewheel systems with hexacarboxylate linkers of C₃ symmetry to generate a family of rht-type MOFs with high and predictable porosity (15–17). In contrast, effective modulation of porosity for isoreticular MOFs based upon 8-connected linkers has not been achieved to date (18–26), thus representing a significant synthetic challenge. We report herein modulation of porosity in a series of isoreticular octacarboxylate MOFs. By varying the length and nature of the heteropolyaromatic cores of the linkers on going from MFM-180 to MFM-185, we have selectively extended the length of metal-organic cages along one direction and also effectively avoided framework interpenetration. The resulting increase in porosity of the materials does not impair their stability toward activation, and the fixed diameter of the pores allows efficient packing of gas molecules across a wide range of pressures. The dynamics of molecular rotors (e.g., phenyl rings) within MOF materials is a key property to their functionalization as they form part of the internal pore surface and are thus highly sensitive to the presence of guest molecules (27–31). In this regard, the series of MOFs herein offer a unique platform to probe the influence of altering the ligand structure on the molecular dynamics and rotational freedom within the resultant framework. To the best of our knowledge, only a few attempts have been made to elucidate the rotational dynamics of linkers in porous MOFs (30–33) and none control them via linker design only. We report here the temperature-dependent ²H NMR studies of selectively deuterated MFM-180-d₁₆ and MFM-181-d₁₆ to define the rotational and flipping modes of the phenyl groups within these structures in the solid state and the molecular factors that govern these dynamics.

Results and Discussion

Design and Synthesis of Octaconnected Ligands and Isoreticular MOFs.

The series of octaconnected linkers, ranging from 19 to 30 Å in dimension, are shown in Fig. 1 and their syntheses are described in detail in SI Appendix. H₈L⁰, H₈L³, and H₈L⁵ were synthesized by direct Suzuki–Miyaura coupling of diethylisophthalate-5-boronic acid with the corresponding tetrahalides: 1,1,2,2-tetrakis(4-bromophenyl)ethane, 2,3,7,8-tetrakis(4-bromophenyl)pyrazino[2,3-g]quinoxaline, and 2,3,9,10-tetrakis(4-bromophenyl)-[1,4]dioxino[2,3-g:5,6-g′]diquinoxaline for H₈L⁰, H₈L³, and H₈L⁵, respectively. In the case of H₈L¹ and H₈L², we used a different strategy where the extended 3′,5′-bis(ethoxycarbonyl)biphenyl-4-ylboronic acid was coupled with 1,2,4,5-tetrabromobenzene and 2,3,6,7-tetraiodonaphthalene, respectively. Attempts to prepare H₈L⁴ from naphthalene-2,3,6,7-tetraamine failed, and therefore the target MOF MFM-184 was analyzed in silico based upon the isoreticular nature of this series of materials. Solvothermal reactions of H₈L⁰, H₈L¹, H₈L², or H₈L³ with CuCl₂ in a mixture of N,N′-diethylformamide/ethanol/0.1 M aqueous HCl (2/2/1, vol/vol/v) at 80 °C for 16 h afforded the solvated materials [Cu₄(Lⁿ)(H₂O)₄]_∞·solv or MFM-18n (n = 0, 1, 2, 3). Due to the insolubility of H₈L⁵ in the above solvent mixture, the synthesis of MFM-185 was conducted in a mixture of N,N′-dimethylformamide/DMSO/2 M aqueous HCl (40/20/1, vol/vol/v) with Cu(NO₃)₂0.2.5H₂O for 4 d. The “indirect” synthesis of [Cu₄(L⁰)(H₂O)₄]_∞ has been reported via transmetallation of the isostructural [Zn₄(L(1))(H₂O)₄]_∞ complex, in which Zn(II) ions are gradually replaced by Cu(II) ions (24).

Fig. 1.

Chemical structures for the octacarboxylate linkers H₈L⁰ to H₈L⁵ used for the synthesis of MFM-180 to MFM-185, representation of the cage assembly in MFM-180, MFM-181, MFM-182, MFM-183, MFM-184 (*predicted structure) and MFM-185, and corresponding BET surface areas (*computed).

Analysis of the Crystal Structures.

The X-ray single-crystal structures of MFM-180, -181, -182, -183, and -185, confirm the formation of square planar [Cu₂(O₂CR)₄] nodes bridged by the octacarboxyate linkers to afford 3D open structures. MFM-180 crystallizes in the tetragonal space group I$\overline{4}$2m with a = 18.6924(2) Å and c = 35.9196(4) Å. The octacarboxylate linker [L⁰]^8- comprises a central tetraphenylethylene core bearing four isophthalate moieties in 4,4′,4′′,4′′′ positions and acts as a 4-connected node (Fig. 1). Each isophthalate arm is orthogonal to the main plane of the molecule and acts as a 3-connected node. Each linker connects to eight [Cu₂(O₂CR)₄] paddlewheels, and each paddlewheel connects to four independent linkers. As a result, the MFM-180 framework can be regarded as a 3,3,4-c tbo net (SI Appendix, Fig. S1) of stoichiometry (3-c)₄(4-c)₃ with the corresponding point symbol of {6²·8²·10²}₃{6³}₄ (34). The metal–ligand linkage affords three types of metal-organic cages (A, B, and C with a ratio of 2:1:1), the smallest of which is an elongated octahedral cage A comprising two ligands and four [Cu₂(O₂CR)₄] paddlewheels (Fig. 2). The [Cu₂(O₂CR)₄] moieties occupy the four equatorial vertices whereas two ethylene groups from the ligands occupy the apical vertices. The overall structure results from the corner-sharing assembly of these octahedral cages A via ethylene groups along the c axis and [Cu₂(O₂CR)₄] paddlewheels along the a/b axis. This arrangement generates two types of elongated cuboctahedral cages, each one comprising four ligands and eight [Cu₂(O₂CR)₄] paddlewheels (Fig. 2). The largest spheres that can fit within these cages taking into account the van der Waals radii of surface atoms have diameters of 3.0, 13.2, and 10.4 Å for cages A, B, and C, respectively. In cage B, the open metal sites of the paddlewheels point to the center of the cavity, whereas in cage C they are tangential to the cavity.

Fig. 2.

View of (A) polyhedral representations of the three types of cages A (octahedral), B, and C (cuboctahedral) and (B) their 3D assembly in MFM-180.

In previously reported examples of octacarboxylate MOFs with tbo or scu nets (18–26), the linkers present a central 4-connected node with increasing distances between the central node and four isophthalate moieties. Although this strategy produced isoreticular MOF structures, no increase of porosity was observed, presumably because of flexibility issues and/or highly strained frameworks. As an alternative approach to simply extending the linker length along all directions, we sought to target the extension of specific dimensionalities and lengths to better control the porosity of resultant MOFs. The distance l₁ (Fig. 1) is crucial for the formation of the octahedral cages A because it codes for the overall structural assembly. In contrast, the distance l₂ defines the length of cages B and C, and thus we sought to design linkers in which l₁ remains constant but l₂ is augmented. For this purpose, we replaced the 4-connected ethylene bond from the ligand H₈L⁰ with a series of extended aromatic cores. The targeted isoreticular MOFs (except for MFM-184) were obtained as single crystals and their structures confirmed by single-crystal X-ray diffraction.

Replacement of the ethylenyl core in MFM-180 with a benzene ring affords MFM-181, which crystallizes in a different tetragonal space group I4/mmm. The change in space group is due to the presence of structural disorder in MFM-181, in which the free rotating phenyl rings are disordered over two positions generated by the mirror symmetry. Nonetheless, MFM-181 exhibits the same tbo topology as MFM-180 (SI Appendix, Fig. S1) when the phenyl core is considered as a 4-connected node. The octahedral cage A in MFM-181 is retained as expected. Cages B and C in MFM-181 are elongated along the c axis by 2.65 Å in comparison with MFM-180. The size of linkers [L²]⁸⁻, [L³]⁸⁻, and [L⁵]⁸⁻ is further increased by incorporating central cores with naphthalene, 3- and 5-fused heteroacenes, respectively. The corresponding MOFs, MFM-182, MFM-183, and MFM-185, all crystallize in the space group I4/mmm with the same type of cage assembly. Given the extension of the polyaromatic cores of [L²]⁸⁻, [L³]⁸⁻, and [L⁵]⁸⁻, the most accurate topological description of the underlying nets of MFM-182, MFM-183, and MFM-185 reflects a process called “decoration” in which a net vertex is replaced by a group of vertices. Thus, MFM-182, MFM-183, and MFM-185 present the same previously unreported 3,3,4-c net with point symbol {6.10²}{6².8².10²}{6³}₂, which is derived from the tbo net by decoration of half of its 4-c vertices by a pair of 3-c vertices (SI Appendix, Fig. S1). The calculated accessible voids (PLATON) are 71.5%, 71.2%, 73.3%, 75.4%, and 76.7% for MFM-180, MFM-181, MFM-182, MFM-183, and MFM-185, respectively.

Modulation of Porosity and Gas Adsorption Property.

Before activation (involving removal of both the free and coordinating solvents from the pores), the materials were exchanged with methanol. This was followed by heating the solvent-exchanged samples at 100 °C under dynamic vacuum to give activated MFM-180a and MFM-181a. To prepare activated MFM-183a and -185a, supercritical CO₂ drying was used to maximize the retention of their pore structure. MFM-180a, -181a, -183a, and 185a show Brunauer–Emmett–Teller (BET) surface areas of 2,610, 3,100, 4,130, and 4,730 m² g⁻¹, respectively, as determined from N₂ adsorption at 77 K. The BET surface areas for MFM-182a and MFM-184a are predicted by grand canonical Monte Carlo methods to be 3,557 and 4,289 m² g⁻¹, respectively, because of the difficulty in their bulk synthesis (SI Appendix). Analysis of the N₂ isotherms using a nonlocal density-functional theory model revealed the pore size distribution centered around 13.3, 13.7, 14.3, and 16.3 Å for MFM-180a, -181a, -183a, and 185a, respectively. Total pore volumes of 1.00, 1.36, 1.45, and 1.65 cm³ g⁻¹ were obtained from the N₂ isotherms, and compare favorably with those calculated based upon single crystal structure (1.09, 1.19, 1.40, and 1.59 cm³ g⁻¹ for MFM-180a, -181a, -183a, and -185a, respectively). Additionally, there is a good agreement between experimental and predicted BET surface areas (SI Appendix, Figs. S5–S18) and overall, the results confirm the complete activation of these MOFs and are consistent with the increasing pore dimensions across the series. MFM-185a possesses one of the highest BET surface areas and pore volumes among octacarboxylate MOFs.

The CH₄ adsorption capacities of MFM-181a, -183a, and 185a are amongst the highest reported values for MOFs at 35 bar, 298 K (Fig. 3). Interestingly, the intrinsic trade-off between gravimetric and volumetric capacities is minimized in this series of MOFs owing to the high framework connectivity and thus relatively high crystal density in comparison with other highly porous MOFs. For example, both gravimetric and volumetric CH₄ adsorption capacities of MFM-181a, -183a, and 185a are higher than those reported for the highly porous octacarboxylate PCN-80a (17.7 wt %, 142 vol/vol) under the same conditions (1). CH₄ adsorption in MFM-181a is saturated at 55 bar, whereas MFM-183a and 185a can accommodate more CH₄ at higher pressures owing to their extended pore space. At 65 bar, the total gravimetric uptake of MFM-185a is sufficiently high (29.0 wt %) to compensate for its low crystal density and hence it displays the highest volumetric CH₄ uptake (198 vol/vol) of the series. Because MFM-181a, 183a, and 185a have almost identical CH₄ uptakes at 298 K and 5 bar, the desirable improvement in porosity through elongation of the metal-organic cages affords an increase in both gravimetric and volumetric “working capacities” (defined as the difference in total uptake between 65 and 5 bar). For example, MFM-185 shows high deliverable CH₄ capacity in both gravimetric and volumetric terms among octacarboxylate MOFs (SI Appendix, Table S6).

Fig. 3.

High-pressure CH₄ sorption isotherms for MFM-181a, -183a, and -185a at 298 K and CO₂ sorption isotherms for MFM-180a, -181a, -183a, and -185a up to 20 bar at 298 K. (Top) gravimetric uptake; (Bottom) volumetric uptake.

The CO₂ adsorption isotherms were recorded at 298 K up to 20 bar for MFM-180a, -181a, -183a, and 185a (Fig. 3). The isosteric heats of CO₂ adsorption were estimated to be ∼23 kJ mol⁻¹ for all MOFs (SI Appendix, Table S7). At low pressure (1 bar), both gravimetric and volumetric CO₂ uptakes are higher for the less porous frameworks, ranging from 13.0 wt % (32.2 vol/vol) for MFM-185a to 15.0 wt % (54.9 vol/vol) for MFM-180a, which suggests that a high density of open metal sites is crucial to maximize the low-pressure CO₂/framework interactions. For pressures higher than 10 bar, the larger pore volume of MFM-185a allows it to reach the highest gravimetric uptake (107.3 wt % at 20 bar) of the series whereas MFM-181a shows the highest volumetric CO₂ uptake (292.4 vol/vol at 20 bar) due to its combination of large surface area and moderately low density. These uptakes compare favorably with other octacarboxylate MOFs presenting higher (PCN-80a²⁵: 72.8 wt % at 20 bar, 293 K) or similar (MFM-140¹⁸: 91.2 wt % at 20 bar, 298 K) porosities.

²H NMR Studies on Dynamics of Molecular Rotors.

Solid-state ²H NMR spectroscopy was applied to investigate the molecular dynamics of the rotation of aromatic rings in this series of MOFs. MFM-180a and MFM-181a were partially deuterated by selectively introducing D atoms on the aromatic rings in the ligands to give MFM-180a-d₁₆ and MFM-181a-d₁₆ (SI Appendix). Variable-temperature ²H NMR spectroscopic studies for MFM-180a-d₁₆ and MFM-181a-d₁₆ (see Fig. 4 and SI Appendix for full spectral data) show that the mobility of the phenyl groups for both materials evolves with temperature starting from a Pake-powder pattern with quadrupolar coupling parameters (Q₀ = 176 kHz, η = 0) typical for static phenyls C-D groups at low temperature (100 K). The evolution of the line shape with rising temperature depicts a reorientation mechanism similar for both MOFs with three regimes: (i) from low temperatures (100 K) up to T₁, the line shape evolves to a typical two-site exchange pattern (35); (ii) above T₁, the line shape remains stable up to T₂, and (iii) above T₂ it evolves to yield a narrowed uniaxial Pake pattern (Q₁ = 21 kHz ∼ Q₀/8, η = 0), indicating that the phenyl fragments rotate homogeneously around the C₂ axis. Although the uniaxial ligand rotation in MOFs has been reported (30), the linkers in MFM-180a-d₁₆ and MFM-181a-d₁₆ show a more complex dynamic behavior.

Fig. 4.

Comparison of the ²H NMR line shape temperature dependence for phenyl fragments in MFM-180a-d₁₆ and MFM-181a-d₁₆ (experimental, black; simulation, red).

The striking difference between MFM-180a-d₁₆ and MFM-181a-d₁₆ lies in the temperatures of transition from one motional mode to another, from two-site flipping to continuous rotation. For MFM-180a-d₁₆, T₁ is ∼310 K and T₂ is ∼330 K, whereas in MFM-181a-d₁₆, T₁ is ∼200 K, and T₂ remains at 330 K. This result indicates that changes in the ligand core affect only one of the two rotational modes, with the flipping mode that evolves below T₁ being notably faster for MFM-181a-d₁₆. The line shape interpretation is based on the following general considerations. The position of mobile phenyl groups in the framework leaves freedom only for rotation or flipping about the C₂ axis and the angle between the rotation axis and the C–D bond is naturally fixed to be θ_ph = 60° (Fig. 5A). Therefore, the simplest model that can describe the line shape evolution is the four-site jump-exchange rotation for the torsional angle φ covering the whole 360° range (Fig. 5B). In such a scheme each of the two C–D bonds flips between two sites (φ₁ <> φ₂ and φ₃ <> φ₄) displaced by a jump angle Δφ₁. When the temperature conditions are met and the phenyl ring is able to overcome the second rotational barrier between the sites φ₁ <> φ₃ and φ₂ <> φ₄, each C–D bond begins to perform the full 360^° rotation. In each pair only the highest barrier is relevant, and thus the motion can be described by the two independent rate constants k₁ and k₂. This model is evidenced by its excellent fit to the experimental data for both samples (Fig. 4). The two-site exchange motion governing the line shape below T₂ allows the determination of the exact position for C–D bond sites. For MFM-180a-d₁₆, the jump angle Δφ₁^I = 71°, whereas for MFM-181a-d₁₆, Δφ₁^II = 68°. The second jump angle can then be readily computed as Δφ₂ⁱ = 180^° − Δφ₁ⁱ. This shows that the equilibrium positions are displaced compared with an ideal C₄ symmetry. The line shape evolution above T₂ fully supports the four sites exchange model and excludes any other interpretation of the observed spectra. Most intriguing are the potential barriers and collision factors involved. In all cases, the rotation rates follow the standard Arrhenius law (Fig. 6). For k₁ in MFM-180a-d₁₆ the parameters are E₁ = 26 kJ mol⁻¹ and k₁₀ = 1.6 × 10¹¹ Hz, and for MFM-181a-d₁₆ E₁ = 20 kJ mol⁻¹ and k₁₀ = 9 × 10¹¹ Hz. For k₂ in MFM-180a-d₁₆ E₂ = 28 kJ mol⁻¹ and k₂₀ = 3 × 10⁷ Hz, whereas for MFM-181-d₁₆ E₂ = 34 kJ mol⁻¹ and k₂₀ = 4.6 × 10⁸ Hz. Additional analysis suggests that intramolecular steric interactions in MFM-180a-d₁₆ are stronger. In both cases these interactions are maximized when all phenyls lie in one plane, with three interaction sites and H–D or D–D distances governing the strength of interaction (Fig. 5 C and D). For MFM-180a-d₁₆ site I shows the interaction of two mobile phenyl groups in 1,1′ positions of the ethylene core, whereas for MFM-181a-d₁₆ it is the interaction of the mobile fragment with the hydrogen of the aromatic core fragment. Sites II and III are geometrically similar for both materials and the shortest distance is realized for site II (marked blue in Fig. 5 C and D). The rates of the slowest motion k₂ (Fig. 6) are almost superimposable for both frameworks and we can thus attribute k₂ to the torsional barrier that rises from electrostatic interaction between neighboring hydrogens in site II. Hence the rate constant k₁ must be governed by site I because the shortest possible distance at the site III is d₃ ∼1.9 Å and is identical in the two linkers, whereas for site I for MFM-180a-d₁₆ d₁ ∼1.6 Å and for the MFM-181a-d₁₆ d₁ ∼1.9 Å. The comparison of the first motion rates k₁ for the two materials confirms that they are indeed considerably different: k₁ is much greater in MFM-181-d₁₆ than in MFM-180-d₁₆. Interestingly, whereas the flipping mode k₁ in both cases is characterized by a collision factor typical for flipping motion in MOFs ∼10¹¹ Hz (30, 33), for k₂ it is ∼3 orders of magnitude smaller at ∼10⁸ Hz, which reflects the strong influence of the steric restrictions on the axial rotation of the phenyl groups in these linkers. This ²H NMR study has revealed that the complex dynamic behavior of the molecular rotors in MOFs in solid state can be elucidated and thus controlled by establishing a strong correlation between the ligand design and the rotational dynamics, the latter of which is a key property of MOF functionalities.

Fig. 5.

(A) Representation of the linker in MFM-180a-d₁₆ with its four mobile phenyl rings. (B) Scheme of rotation sites for phenyl groups in MFM-180a-d₁₆ and MFM-181a-d₁₆. The reorientation scheme comprises a four-site exchange motion with two different rate constants: k₁ (Δφ₁) and k₂ (Δφ₂). The green arrow represents the barrier associated with k₁, the blue arrow represents the barrier associated with k₂, and the red arrows are associated with barriers in the Δφ₁ and Δφ₂ regions associated with minor steric restrictions. (C and D) Scheme of interaction sites that might influence the rotational potential for phenyl groups in MFM-180a-d₁₆ and MFM-181a-d₁₆, respectively. Parameter d₂ is the shortest achievable distance for the electrostatic interaction between the phenyl hydrogens and is similar in both materials.

Fig. 6.

Arrhenius plots of the rotational rate constants k : (○) k₁ and (□) k₂ for MFM-180a-d₁₆; (●) k₁ and (▲) k₂ for MFM-181a-d₁₆.

Conclusion

A series of isoreticular MOF materials based on a family of octacarboxylate linkers has been developed. The rigid, heteropolyaromatic linkers were designed to self-assemble with [Cu₂(O₂CR)₄] paddlewheels to afford frameworks with elongated nanotubular cages of fixed diameter. The isoreticular design results in systematically increased pore volumes and surface areas for the MOFs. Notably, in the case of CH₄ adsorption, extension of the linker causes no uptake loss in the low-pressure region, and both gravimetric and volumetric uptakes are simultaneously enhanced at high pressure. This affords an impressive CH₄ working capacity of 0.24 g g⁻¹ and 163 vol/vol (298 K, 5–65 bar) for activated MFM-185a. We attribute this behavior to the efficient packing of gas molecules in the tubular pores and the high connectivity (and thus suitably high crystal density) of the framework. In addition, a rational synthetic design has allowed control of the torsional dynamics of linkers in MOF solids. The high predictability of the linker/metal self-assembly combined with their pore shape make this series of MOFs a unique platform for exploring further the tuning of porosity, decoration of pores, and development and control of new molecular rotors in functional MOFs.

Materials and Methods

The ligands H₈L⁰, H₈L¹, H₈L², H₈L³, and H₈L⁵ were synthesized using a Suzuki–Miyaura coupling reaction between the corresponding tetrahalogenated core and boronic acid, followed by hydrolysis of the ester functions. The synthesis of H₈L¹ is described in detail in SI Appendix. The selectively deuterated linkers were synthesized following the same procedures but starting from deuterated building blocks.

Synthesis of MFM-180, -181, -182, and -183: H₈L^0-3 (0.30 mmol) and CuCl₂ (0.19 g, 1.40 mmol) was dissolved in N,N'-diethylformamide (30 mL). EtOH (30 mL) and an aqueous solution of HCl (0.1M, 15 mL) were added to the resulting solution, which was placed in a tightly capped 250 mL Duran pressure plus laboratory bottle (cat. no. 1092234). The solution was heated at 80 °C in an oven for 16 h, and a large amount of crystalline product precipitated. The crystal plates of the corresponding MOF were isolated by filtration while the mother liquor was still warm.

Due to the poor solubility of H₈L⁵, MFM-185 was synthesized following slightly different conditions (SI Appendix).

Detailed synthesis procedures and characterizations of the linkers and MOFs, along with crystallographic data and description of gas sorption and solid-state ²H NMR experiments can be found in SI Appendix.

CCDC 1472806–1472810 contain the supporting crystallographic data for materials MFM-180, -181, -182, -183, and -185, respectively. These data can be obtained free of charge from the Cambridge Crystallographic Data Centre via https://www.ccdc.cam.ac.uk/structures-beta/. Please also see Datasets S1–S5 for materials MFM-180, -181, -182, -183, and -185.