Host–guest selectivity in a series of isoreticular metal–organic frameworks: observation of acetylene-to-alkyne and carbon dioxide-to-amide interactions

We report a series of six isoreticular metal–organic frameworks (MOFs) for selective gas adsorption, specifically for selective adsorption of CO2 and C2H2.

3,5-Dimethylphenylboronic acid (15.00 g, 95.0 mmol) and NaOH (15.00 g, 375 mmol) were dissolved in tertbutanol:water (600 mL, 1:1 v/v). The mixture was heated to 50 °C and KMnO4 (82 g, 1 mol) added in 2 to 4 g portions over five days. The temperature was increased to 65°C after two-thirds had been added. Excess permanganate was reduced by addition of Na2S2O3 (ca. 5.0 g) and the precipitated MnO2 removed by filtration and washed with boiling water. The solvent volume was reduced in vacuo and the product precipitated by addition of conc. HCl to pH 2. The resulting solid was isolated by filtration, washed with cold slightly acidic water, recrystallized from hot water and dried to yield 4 (11.02 g, 55 %). 1
The product was isolated by filtration, washed extensively with water and dried to give 8 (2.30 g, 74%
A suspension of 5-pyrimidine carboxylic acid (1.00 g, 7.8 mmol) in thionyl chloride (25 mL) and DMF (0.02 mL) was heated at 80 o C under reflux to give a clear solution after 3 h. The solvent evaporated with care under reduced pressure, and the resulting solid was azeotroped with toluene under vacuum to remove traces of residual thionyl chloride. The resulting solid was suspended in anhydrous THF (50 mL) under Ar and then cooled to 0 o C. 5-Aminoisophthalic acid (1.44 g, 7.8 mmol) and K3PO4 (4.14 g, 19.5 mmol) were added and the reaction mixture stirred as it warmed up to ambient temperature for 2 h. Water (50 mL) was added and some of the solvent (30 mL) removed in vacuo. MeCN (30 mL) was added and the solution acidified to pH = 2 using formic acid. The resulting yellow solid was isolated by filtration and washed with water and MeCN and dried to give  87, 163.04, 160.80, 156.81, 139.69, 132.35, 128.66, 125.94, 125 ,51.15;H,3.63;N,13.77. Found: C,51.22;H,3.25;N,12.89.

Synthesis of MFM-126
H2L 1 (10 mg, 0.04 mmol) and Cu(NO3)2.3H2O (20 mg, 0.08 mmol) were dissolved in DMF (4 mL) in a pressure tube and HCl (2M, 0.1 mL) was added to the mixture. The tube was tightly capped and heated in an oil bath at

Synthesis of MFM-127
H2L 2 (10 mg, 0.04 mmol) and Cu(NO3)2 . 3H2O (20 mg, 0.08 mmol) were dissolved in DMF (4 mL) and EtOH (1 mL) in a pressure tube and HCl (2M, 0.1 mL) was added to the mixture. The tube was tightly capped and heated in the oven at 80 o C for 24 h to afford green crystals which were washed with DMF (5 mL x 3 mL),

Synthesis of MFM-136
This was completed from the published procedure. 1

ATR-FTIR spectra of MFM-126 -MFM-138 and their respective linkers
Labelled peaks highlight the shift of carbonyl stretching vibrations from 1693 cm -1 to 1634 cm -1 ; 1704 cm -1 to 1667 cm -1 ; 1704 cm -1 to 1669 cm -1 ; 1716 cm -1 to 1641 cm -1 ; 1721 cm -1 to 1667 cm -1 and 1716 cm -1 to 1668 cm -1 , from free linkers H2L 1-6 to corresponding MOFs, respectively. The depletion of these carbonyl bands in the spectra of the MOFs indicates the absence of residual unbound linkers in the respective MOF materials. Diffraction data were collected on a Rigaku Oxford Diffraction SuperNova diffractometers equipped with Atlas detectors and microfocus Mo or Cu X-ray sources. X-ray data for MFM-127 were collected using a synchrotron radiation at single crystal X-ray diffraction beamline I19 in Diamond light Source, 2 equipped with a Pilatus 2M detector and an Oxford Cryosystems nitrogen flow gas system. MFM-127 data were measured using GDA suite of programs. The raw data were reduced and corrected for Lorentz and polarisation effects using CrysAlisPro; 3 corrections for the effects of adsorption were applied using a numerical absorption correction based on Gaussian integration over a multifaceted crystal model. All structures were solved by direct methods (SHELXS) 4 and refined by full-matrix least-squares (SHELXL). 5 Regions of diffuse solvent in the solvated structures were treated with the PLATON SQUEEZE routine. 6

MFM-126
The diffraction of the crystal was weak with little intensity beyond 0.9 Å resolution. This is likely a consequence of large volumes of poorly ordered solvent (total void fraction 0.44) and conformational disorder in the ligand.
Conformational disorder was observed in pyrimidine ring N11-C16, amide moiety C17-N19 and phenyl ring C21-C26. The occupancies of the two pyrimidine ring components were refined before each being fixed at a value of 0.5. The occupancies of the two phenyl ring components were freely refined and constrained to sum to unity (occupancy of component A 0.35 (1)). The four conformations of the amide moiety were constrained to have a value of half of the occupancy of the phenyl ring with which they share connectivity. The phenyl ring was constrained to have regular hexagonal geometry (AFIX 66), and the pyrimidine ring was constrained to have planar geometry (FLAT). The 1,2 and 1,3 distances around the disordered pair of pyrimidine rings were restrained to reflect the C2v symmetry of the pyrimidyl moiety (SADI). Pairs of amide nitrogen atoms N19C/N19D and N19E/N19F were constrained to occupy the same sites and have identical isotropic thermal displacement parameters (EXZY/EADP). The three atoms of each amide moiety and connected pyrimidine carbon atom were fixed to have co-planar geometry (FLAT). All atoms except those of the disordered amide moieties have been refined with anisotropic displacement parameters. Rigid bond and similarity restraints have been applied to the displacement parameters of all the atoms in the structure (RIGU, SIMU). Hydrogen atoms were placed geometrically and refined using a riding model. Disordered solvent molecules could not be sensibly modelled, so the structure was treated with PLATON SQUEEZE. 6 A total of 1154 electrons were accounted from the P1 cell, equating to 1.5 dimethylformamide molecules per asymmetric unit, which have been included in the unit cell contents and calculation of derived parameters. A large positive electron density peak (3.94 e Å -3 ) lies on the three-fold symmetry axis 2.00 Å from disordered pyrimidine nitrogen atoms N13A and N13B.
The electron density is too close to the pyrimidine moieties to be plausibly modelled as a solvent. S16 Figure S3. Views of the crystal structure of MFM-126 along the a-, b-and c-axes.
a-axis b-axis c-axis S17

MFM-127
A residual electron density peak with a height of 1.65 e A -3 is located 0.662 A from copper atom Cu1. The electron density peak possibly arises as a result of an unmodeled disorder component of copper atom Cu1 which in turn could be caused by an alternative conformation of the pyrimidine ring. The alternative pyrimidine ring conformation is observed as a disorder component when the structure is solved the lower symmetry space group R-3. An attempt to model the peak as a disorder site with copper atom Cu1 resulted in a refined fractional occupancy of 0.04 (the occupancies of the two sites were constrained to sum to unity). No sensible model for associated ligand disorder could be modelled and the structure is reported with a single full occupancy copper site. Disordered solvent molecules could not be sensibly modelled, so the structure was treated with PLATON SQUEEZE. 6 A total of 1747 electrons were accounted from the P1 cell in this, equating to 1.2 dimethylformamide molecules per asymmetric unit, which have been included in the unit cell contents and calculation of derived parameters.

MFM-128
Phenyl ring C21-C26 was found to be disordered over two orientations. The occupancies of the two components were refined and constrained to sum to unity with values 0.63 (1)  the electron density is too close to the main residue to be treated with PLATON SQUEEZE. 6 A damping factor was used in the refinement to aid problematic convergence of the esd on the z coordinate of Cu1. Disordered solvent molecules could not be sensibly modelled, so the structure was treated with PLATON SQUEEZE. 6 A total of 3030 electrons were accounted from the P1 cell in this, equating to 2 dimethylacetamide molecules per asymmetric unit, which have been included in the unit cell contents and calculation of derived parameters.

MFM-137
Pyrimidyl ring N11-C16, phenyl ring C21-C26 and alkyne atom C27 were found to be disordered over two orientations. The occupancies of these disorder components were refined before being fixed at 0.5 each.
Geometric similarity restraints were applied to the bond distances round the disordered phenyl and pyrimidyl rings (SADI). The disordered phenyl and pyrimidyl rings were restrained to have approximately planar geometries (FLAT). The disordered alkyne C-C triple bond distances were restrained to be the same length (SADI). Rigid bond and similarity restraints were applied to the anisotropic thermal displacement parameters of the disordered atoms (RIGU and SIMU). Several disordered solvent molecules could not be sensibly modelled, and so the structure was treated with PLATON SQUEEZE. 6 A total of 1811 electrons were accounted for in the P1 cell, equating to 2.5 dimethylformamide molecules per asymmetric unit, these are included in the chemical formula and in all quantities calculated from it.

MFM-138
Pyrimidyl ring N11-C16, phenyl rings C21-C26 and C31-C36 were found to be disordered over two orientations. The occupancies of these disorder components were refined before being fixed at 0.5 each, and Geometric similarity restraints were applied to the bond distances round the disordered phenyl rings (SADI).
Rigid bond and similarity restraints were applied to the anisotropic thermal displacement parameters of the disordered atoms (RIGU and SIMU). Several disordered solvent molecules could not be sensibly modelled, and so the structure was treated with PLATON SQUEEZE. 6 A total of 1995 electrons were accounted for in the P1 cell, equating to 2 diethylformamide molecules per asymmetric unit, these are included in the chemical formula and in all quantities calculated from it.

Gravimetric N2, CO2, CH4 and C2H2 Adsorption
All MOFs were solvent-exchanged with acetone or EtOH before heating at 393 K under dynamic vacuum to produce the desolvated materials. Low pressure (0-1 bar) sorption isotherms for CO2, CH4 and N2 and high pressure (0-20 bar) sorption isotherms for CO2 and CH4 were recorded using a Hiden Isochema Gravimetric Analyser (IGA-003) instrument. All isotherms were collected using ultrahigh vacuum diaphragm and turbo pumping systems, with ultra-pure research grade gases (99.9999%) purchased from BOC and used as received.
Hiden temperature-controlled water baths were used to obtain isotherms at 273 and 298 K. The MOF samples were loaded from solvent into a sample basket in the sorption analyser and heated to 393 K under dynamic vacuum for 16 h to obtain fully activated samples (50-70 mg).

Brunauer-Emmett-Teller (BET) Surface Areas
Volumetric N2 adsorption data were recorded at 77 K (liquid nitrogen) on a Quantachrome Autosorb-1c instrument under ultra-high vacuum in a clean system with diaphragm and turbo pumping system using ultrapure research grade (99.9999%) N2. The BET surface areas were calculated using the software (version 1.60) integrated in the instrument. Pore size distribution data and cumulative pore volume were determined by analysis of the N2 isotherms at 77 K using a non-local density functional theory (NLDFT) implementing a hybrid kernel.
Vacuum dried powder samples were loaded on the instrument and degassed at 373 K and 10 -9 bar for a minimum of 16 h to yield desolvated sample, which was then loaded in the instrument for N2 adsorption measurements.

Neutron Powder Diffraction (NPD) Studies of MFM-126 and MFM-127
NPD data were collected at the WISH Diffractometer at ISIS Muon and Neutron Source, UK. 7 Acetoneexchanged MFM-126 and ethanol-exchanged MFM-127 were loaded into 11mm diameter vanadium sample cans and outgassed at 1 x 10 -7 mBar and 130°C for 3 days. The samples were loaded into a liquid helium cryostat and cooled to 7 K for data collection of the bare framework. CO2, C2D2 and CD4 were volumetrically dosed from a calibrated volume, calibrated via use of the ideal gas equation (PV = nRT), after warming the sample to 293 K with data collected at various loadings of CO2, C2D2 and CD4 per copper of the materials. The sample can was isolated after reaching the target dosing amount to minimise the presence of "free gas" inside the can.
The sample was then slowly cooled to 7 K (over ~ 3 h) to ensure adsorbates were completely adsorbed with no condensation elsewhere in the system. Sufficient time was allowed to achieve thermal equilibrium before data collection.

Rietveld Refinements of Guest Molecule Positions
The locations of CO2 and CD4 molecules within MFM-126 and MFM-127 as well as C2D2 in MFM-127 were determined as a function of gas loading by sequential Fourier difference map analysis followed by Rietveld refinement using the Topas software package. 8 Analysis of the Fourier map of the outgassed data indicated no residual nuclear density in the voids. The structure as solved by single crystal X-ray diffraction data was used as a starting point for the framework model which was geometrically restrained and refined against the NPD data. The framework atom coordinates were subsequently fixed before the models of guest molecules were developed. All binding sites were checked carefully for their unambiguous presence in the final structural model.
Common C-O / C-D bond distances and isotropic thermal factors were included for the guest molecules. Final refinements comprised all free structural variables from both the framework and guest molecules. S28 Figure S8. Observed (blue), calculated (red) and difference (grey) profiles of the Rietveld refinement of the neutron powder diffraction data (detector banks 1-4) for bare MFM-126 S29 Figure S9. Observed (blue), calculated (red) and difference (grey) profiles of the Rietveld refinement of the neutron powder diffraction data (detector banks 1-4) for MFM-126 loaded with 1.0 CO2 per Cu.

Inelastic Neutron Spectroscopy of MFM-126
INS data were collected on the TOSCA beamline at ISIS Muon and Neutron facility. 9 TOSCA is a general purpose inelastic neutron spectrometer which can cover the whole range of molecular vibrations from 0-4000 cm -1 . The instrument comprises of 130 3 He detectors in the forward and backscattering geometry located 17 m downstream of a 300 K Gd poisoned water moderator. A temperature of 7 K was maintained during data collection by two He closed cycle refrigerators. MFM-126 (~1.5 g) was loaded from acetone into a 11 mm diameter vanadium can, sealed with Indium wire and outgassed at 10 -6 mbar at 393 K for three days to remove any trace guest molecules. After placing the sample into a He cooled cryostat INS data of the bare framework were collected at 7 K. A loading of 1.0 CO2 per Cu was dosed volumetrically, from a calibrated volume, at room temperature to ensure sufficient mobility of the guest species and gradually cooled to 7 K to allow for the guest species to fully adsorb into MFM-126, with no condensation elsewhere in the system. INS data of 1.0 CO2/Cu of MFM-126 were collected at 7 K.

DFT Calculations and modelling of the INS spectra
Vibrational frequencies and polarization vectors were calculated using CP2K, 10 based on the mixed Gaussian and plane-wave scheme 11 and the Quickstep module. 12 The calculation used molecularly optimized Double-

Calculation of isosteric heats of adsorption (Qst values)
Virial analysis of the gas adsorption data was used to determine the isosteric heats of adsorption.

Virial Method 1
This analysis can be performed in two methods. The first method is shown in equations (I) and (II).

(I)
where is the quantity adsorbed at pressure and 0 , 1 , etc. are virial coefficients. 0 describes the adsorbate-

(III)
Where is the quantity adsorbed at pressure , T is temperature, and are temperature independent virial coefficients and and determine the number of terms to adequately describe the isotherm. The resulting virial coefficients, through ,were used to calculate enthalpies of adsorption via equation (IV).

Calculation of Selectivity Values using Ideal Adsorbed Solution Theory
Selectivity values were calculated using the Ideal Adsorbed Solution Theory (IAST) method from single component isotherms. The N2, CO2 and CH4 adsorption isotherms were initially fitted with the dual site Langmuir-Freundlich (DLSF) model (V).
Where is the amount adsorbed (mmol g -1 ) at gas pressure P (bar), , is the saturation capacity (mmol g -1 ) at sites i, is the affinity coefficients of sites i, and is the ideal homogenous surface derivation.
After these fitting parameters were determined IAST was used to predict the mixture adsorption isotherms and subsequently calculate the selectivity values, ⁄ (VI), for binary mixtures.
Where is the molar fraction of the adsorbed species and is the molar fraction in the gas-phase.

Breakthrough Experiments
To evaluate the performance of MFM-126 in the selective adsorption of CO2, breakthrough experiments were performed using a Hiden Isochema Automated Breakthrough Analyzer with integrated mass spectrometer. A column packed with MFM-126 (0.95 g) was heated at 393 K under a flow of He (100 mL min -1 ) overnight before cooling to 298 K. Breakthrough experiments were conducted using N2/CO2 (85:15) and CH4/CO2 (50:50) which were flowed over the packed bed with a total flow rate of 10 mL min -1 at 298 K and 1.0 bar. After breakthrough of both components, desorption was conducted under a flow of He at room temperature to regenerate the column. Dimensionless breakthrough plots were calculated with the following parameters: bed diameter, d, (7 mm), bed length, L, (120 mm), flow rate (10 mL min -1 ), bed volume (5 mL), sample mass (0.95 g), sample framework density (1.3 g cm -3 ). The sample occupies a volume of 0.73 mL (assuming 100% purity and no framework collapse), and thus the fractional porosity of the fixed bed, ε, is calculated to be 0.853. The superficial gas velocity, u, at the entrance of the bed corresponds to 4.33e -2 m s -1 . The characteristic contact time between the gas and MFM-126, εL/u = 23.6 s. The dimensionless time, τ, was obtained by dividing the actual time, t, by the contact time between the gas and the MFM-126 sample, εL/u, i.e. τ = tu/εL.