Post‐Synthetic Modification Unlocks a 2D‐to‐3D Switch in MOF Breathing Response: A Single‐Crystal‐Diffraction Mapping Study

Abstract Post‐synthetic modification (PSM) of the interpenetrated diamondoid metal–organic framework (Me2NH2)[In(BDC‐NH2)2] (BDC‐NH2=aminobenzenedicarboxylate) SHF‐61 proceeds quantitatively in a single‐crystal‐to‐single‐crystal manner to yield the acetamide derivative (Me2NH2)[In(BDC‐NHC(O)Me)2] SHF‐62. Continuous breathing behaviour during activation/desolvation is retained upon PSM, but pore closing now leads to ring‐flipping to avert steric clash of amide methyl groups of the modified ligands. This triggers a reduction in the amplitude of the breathing deformation in the two dimensions associated with pore diameter, but a large increase in the third dimension associated with pore length. The MOF is thereby converted from predominantly 2D breathing (in SHF‐61) to a distinctly 3D breathing motion (in SHF‐62) indicating a decoupling of the pore‐width and pore‐length breathing motions. These breathing motions have been mapped by a series of single‐crystal diffraction studies.


Thermogravimetric analysis of (Me 2 NH 2 )[In(BDC-NHC(O)CH 3 ) 2 ]·CHCl 3 (SHF-62-CHCl 3 )
TGA analysis was carried out on SHF-62-CHCl 3 (5.4 mg) and SHF-62-DMF (3.6 mg) by heating the samples from 298 K (25 o C) to 973 K (700 o C) at a rate of 4 K/min. The samples were initially held at 298 K for 30 minutes before conducting the heating regime under flowing air (see Figures S1 and S2). Figure S1. Thermogravimetric analysis of SHF-62-CHCl 3 recorded at a ramp rate of 4 K/min. The dotted red line indicates the expected mass loss of the contained solvent calculated from elemental analysis. Figure S2. Thermogravimetric analysis of SHF-62-DMF recorded at a ramp rate of 4 K/min. The dotted red line indicates the expected mass loss of the contained solvent calculated from elemental analysis.

Solution-Phase 1 H NMR spectroscopy of digested SHF-62-CHCl 3
Solution-phase 1 H NMR spectroscopy (400 MHz, DMSO-d 6 ) was carried out using a Bruker DPX-400 spectrometer. The MOF sample SHF-62-CHCl 3 (10 mg) was digested using 50 µL of acid (35% DCl in D 2 O) in 1 mL of DMSO-d 6 , and the spectrum recorded without neutralising the solution (Figure S3). All constituent parts were observed to be soluble in the DMSO after digestion. The full conversion from SHF-61-DMF was evident from the change in the shifts of the aromatic protons on the ligand, which was also analysed using the same method ( Figure S4). Figure S3. 1 H NMR spectrum measured at 298 K for digested SHF-62-CHCl 3 Figure S4. Comparison of the aromatic region of the 1 H NMR spectra measured at 298 K for both digested SHF-62-CHCl 3 (blue) and digested SHF-61-DMF (red). The individual resonances have been assigned based on the J-coupling of signals.
The 1 H NMR spectrum of MOF sample SHF-62-DMF (5 mg) was recorded by following the same digestion procedure. The spectrum (Figure S5) shows that DMF has replaced CHCl 3 as the solvent, and that the ligand modification is maintained. Figure S5. 1 H NMR spectrum measured at 298 K for digested SHF-62-DMF SHF-62-CHCl 3

SHF-61-DMF
S6 1 H NMR spectroscopic data were also recorded on sample SHF-62-CHCl 3 -PC using the same methodology. The conversion from SHF-61-DMF was judged to be 75-80% based on the relative integrations of the aromatic protons on the ligand. The spectrum is shown in Figure S6. Figure S6. 1 H NMR spectra measured at 298 K for digested SHF-62-CHCl 3 -PC.

Solid-state 13 C NMR spectroscopy
Solid-state 13 C NMR spectroscopy (500 MHz for 1 H) was carried out on SHF-61-CHCl 3 and SHF-62-CHCl 3 (Figure S7, Tables S1 and S2). Analogous characterisation for SHF-61-DMF has been previously reported. S1 Samples (100 mg) were ground into a fine powder, packed into 4 mm zirconia rotors and transferred to a Bruker AVANCE III HD spectrometer. Pseudo-quantitative 1D 13 C (with high-power 1 H decoupling) magic-angle spinning (MAS) NMR experiments were performed at a MAS rate of 10.0 kHz. Spectra were measured using a relaxation delay of 300 s in approximation of quantitative conditions. (T1s for 13 C centres were not determined). Measurements were acquired until sufficient signal-to-noise was observed. Values of the chemical shifts are referenced to adamantane.
The magnetic field was set to place the higher frequency 13 C (methylene) resonance of adamantane at a chemical shift of 38.48ppm. Integrals were obtained by deconvolution. Figure S7. Normalised, solid-state quantitative 13 C NMR spectra recorded at 298 K for SHF-61-CHCl 3 (green) and SHF-62-CHCl 3 (blue). Asterisks are used to denote spinning sidebands.  Figure S8, Table S3). Activation was achieved by heating a sample of SHF-62-CHCl 3 in a Schlenk tube at 353 K for 16 hours under high vacuum. Samples (100 mg) were ground into a fine powder, packed into 4 mm zirconia rotors and transferred to a Bruker AVANCE III HD spectrometer. 1D 1 H-15 N cross-polarisation magic-angle spinning (CP-MAS) NMR experiments were performed at a MAS rate of 10.0 kHz. Spectra were recorded using a contact time of 2.0 ms. The relaxation delay D 1 for each sample was individually determined from the proton T 1 measurement (D 1 = 5 x T 1 ). Measurements were acquired until sufficient signal-to-noise was observed. Values of the chemical shifts are referenced to adamantane.

S7
The magnetic field was set to place the higher frequency 13 C (methylene) resonance of adamantane at a chemical shift of 38.48ppm.

Single-Crystal X-Ray Diffraction
Laboratory single-crystal X-ray diffraction data were collected on a Bruker SMART APEX-II CCD diffractometer operating a Mo-K α sealed-tube X-ray source or a Bruker D8 Venture diffractometer equipped with a PHOTON 100 dual-CMOS chip detector and operating a Cu-K α IµS microfocus Xray source. The data were processed using either the APEX2 S2 software or the CrysAlisPro Software. S3 X-Ray data were corrected for absorption using empirical methods based upon symmetry-equivalent reflections combined with measurements at different azimuthal angles (SADABS). S4 An Oxford Cryosystems Cryostream device was used to maintain the sample temperature. Synchrotron singlecrystal X-ray diffraction data were collected at beamline I19 (EH1), Diamond Light Source. S5 Data were collected at a wavelength of 0.6889(1) Å using a Fluid Film Devices Ltd diffractometer equipped with a PILATUS 2M detector. Sample temperature was controlled using an Oxford Cryosystems Cryostream Plus device. Data were processed using DIALS software S6 and corrected for absorption using empirical methods (SADABS). S4 All crystal structures were solved and refined against F 2 values using the program SHELXL S7 accessed within the OLEX2 program. S8 All non-hydrogen atoms were refined anisotropically except for some of the structures in which positional disorder of the cations or orientational disorder of the linker ligand was required. Crystallographic restraints and constraints were applied to some structures where necessary. Positions of hydrogen atoms were calculated with idealised geometries and refined using a riding model with isotropic displacement parameters. The PLATON function SQUEEZE S9 was applied SHF-62-CHCl 3

S11
in some cases to determine the contribution to the structure factors of the unmodelled electron density of solvent molecules and/or the cations in the MOF pores, where otherwise poor convergence of the least-squares refinement resulted. As an illustrative example, for the crystal structure refinement of SHF-62-CHCl 3 -PC (Section 2), SQUEEZE was used to treat the cation and solvent molecules as a contribution to the overall scattering of diffuse electron density without modelling specific atom positions, but cation and guest solvent molecules were able to be modelled for SHF where one cation molecule and 0.5 CHCl 3 molecules per formula unit were able to be modelled crystallographically. For crystal structure refinements described in section 3 analogous approach was adopted in only where needed. Full details are available in the CIFs, which were checked using checkCIF/PLATON, S10 and include responses to any alerts.

Powder X-Ray Diffraction
Laboratory powder diffraction data were obtained using a Bruker D8 Advance powder diffractometer equipped with focusing Göbel mirrors, recorded in the range 3 o ≤ 2θ ≤ 50 o , using Cu-K α radiation.
Data were collected in a Debye-Scherrer geometry with rotating capillary stage and samples loaded in 0.7 mm borosilicate capillaries.
Synchrotron powder diffraction data were collected at beamline I11 at Diamond Light Source using a wide-angle (90 °) position sensitive detector (PSD) comprising 18 Mythen-2 modules. S13, S14 A pair of scans related by a 0.25 ° detector offset was collected for each measurement to account for gaps between detector modules. The resulting patterns were summed to give the final pattern for analysis.

Bulk-phase analysis of SHF-62-CHCl 3 and SHF-62-DMF
The phase purity of SHF-62-CHCl 3 was confirmed through X-ray powder diffraction. A room temperature pattern (B1) was collected on the laboratory instrument (Cu-Kα) as detailed on page S12.
The unit cell parameters from the single-crystal structure of the as-synthesised MOF (A1) were used as starting point for a Pawley refinement, employing 311 parameters (10 background, 1 zero error, 5 profile, 3 cell and 292 reflections), resulting in final indices of fit R wp = 4.337, R wp' = 8.132. The fit is shown in Figure S10. The final unit cell parameters were orthorhombic a = 14.8985 (3)   The phase purity of SHF62-DMF was also confirmed through X-ray powder diffraction. A highresolution room-temperature pattern (B2) was collected at the I11 beamline S11,S12 as detailed on page S12, λ = 0.826210 (5)

Crystallographic studies of SHF-62-CHCl 3 following in situ heating
Crystals used for in situ heating were selected while immersed in the mother liquor and one face of the crystal glued to a glass fibre or MiTiGen MicroMount while the crystal was still covered in a thin layer of residual solvent. Care was taken to avoid coating the entire crystal in adhesive. The crystal was rapidly transferred to the diffractometer and situated in the nitrogen stream of an Oxford Cryosystems Cryostream device at room temperature. The glue was left to dry for 15 mins before any data were collected. Heating was carried out in situ using the Cryostream device. Seven different heating experiments were carried out on seven separate crystals of SHF-62-CHCl 3 (C1-C7).
Experiments C1-C5 were performed using a lab diffractometer, as described in section 1.12.
Experiments C6 and C7 were carried out at beamline I19, Diamond Light Source, as described in section 1.12. S5 The crystals C5-C7 were prepared using dry CHCl 3 and those in C3 and C4 were exchanged with dry CHCl 3 prior to the diffraction experiment (CHCl 3 dried according to the method of Grubbs S16 ).
A ramp rate of 4 K/min was used to raise the temperature, which was held at the final value for a set period (see Table S5) before cooling at the same ramp rate. Intensity data collections were recorded under the nitrogen stream at 298 K, using the Cryostream to maintain the temperature. In experiment C3 the crystal was then cooled to 100 K using the Cryostream to observe the thermal effect on the flexibility. The unit cell parameters for experiments C1-C5 were obtained by analysing reflections with I/σ > 10 from four sets of 10 o omega scans with 0.5 o slicing. Full data collections were recorded at the end of studies C1 (at 298 K), C3 (at 100 K) and C4 (at 298 K), due to the sample maintaining a high level of crystallinity (C1.3, C3.3 & C4.4). A full data collection was also recorded for experiment C5.2. Full data collections were recorded throughout studies C6 and C7. Table S5 details the seven separate heating studies carried out and the resulting cell parameters. Details of the full data collections are shown in Table S6.
Densities are calculated using only framework atoms and cations, and do not include guest molecules. c Adsorption coefficients are calculated using only framework atoms and cations, and do not include guest molecules.

Crystallographic studies of SHF-62-DMF following in situ heating
Crystals used for in situ heating were selected while immersed in the mother liquor and one face of the crystal glued to a glass fibre while the crystal was still covered in a thin layer of residual solvent.
Care was taken to avoid coating the entire crystal in adhesive. The crystal was rapidly transferred to the diffractometer and situated in the nitrogen stream of the Cryostream device at 298 K. The glue was left to dry for 15 mins before any data were collected. Heating was carried out in situ using the Cryostream device. Two different heating experiments (D1 & D2) were carried out on two separate crystals of SHF-62-DMF. A ramp rate of 4 K/min was used to raise the temperature, which was held at the final value for a specified period of time (see Table S4) before cooling at the same ramp rate.
Intensity data collections were recorded using a laboratory diffractometer, as described in section 1.12, using the Cryostream device to maintain the temperature at 298 K under the nitrogen stream.
The unit cell parameters were obtained by analysing reflections with I/σ > 10 from four sets of 10 o omega scans with 0.5 o slicing. A full data collection was recorded at the end of study D1 due to the sample maintaining a high level of crystallinity (D1.3). Table S7 details the heating study parameters and the resulting cell parameters. Details of the full data collection D1.3 are shown in Table S8.

Crystallographic studies of SHF-62-DMF following ex situ heating
Crystals of SHF-62-DMF were transferred from the DMF solvent to a microscope slide and left until the DMF had evaporated. The crystals were then heated in a temperature-controlled oven at 150 o C for 15 minutes. After treatment the crystals were covered in a perfluoropolyether oil and mounted on a goniometer head. X-Ray diffraction data were recorded at 100 K under a cold nitrogen stream using a laboratory diffractometer , as described in section 1.12. The crystal retained most of its crystallinity, enabling detailed structural information to be obtained (E1). Full crystallographic information for the structure is listed in Table S8. ( 2 ) = √∑ ( 2 − 2 ) 2 ( + − ) ⁄ . b Densities are calculated using only framework atoms and cations, and do not include guest molecules. c Adsorption coefficients are calculated using only framework atoms and cations, and do not include guest molecules.

Activation-flip modelled from single-crystal structures
The single-crystal structures determined during the studies of activation of SHF-62-CHCl 3 and SHF-62-DMF show that there is a gradual response to a-axis contraction as solvent is removed, which results in the 150 ° activation-flip of a proportion of the methylamidobenzendicaboxylate (BDC-NHC(O)Me) ligands. The activation flip is first detected at a ≈ 14.2 Å and reaches the maximum extent (approx. 50% of ligands flipped) as the the pores contract further to a < 13.7 Å ( Figure S13).

Activation-flip modelled from single crystal X-ray diffraction: comparison of SHF-62-CHCl 3 with SHF-62-DMF
The series of single-crystal X-ray diffraction studies used to determine the crystals structure and/or unit-cell dimensions presented in Figure 2g and 2h were used to establish the regions in which the activation-flip occurs in terms of pore widths (b-and c-axes) and pore length (a-axis). Separation of these studies based upon which pore solvent is being removed ( Figure S14) confirms that the region in which the activation-flip occurs lies in a narrow range of dimensions of pore length ( Figure S14b and S14d), but suggests that the activation-flip may occur at a narrower-pore stage for SHF-62-DMF (Figure S14c) than for SHF-62-CHCl 3 ( Figure S14a).