Survival of Zirconium-Based Metal–Organic Framework Crystallinity at Extreme Pressures

Recent research on metal–organic frameworks (MOFs) has shown a shift from considering only the crystalline high-porosity phases to exploring their amorphous counterparts. Applying pressure to a crystalline MOF is a common method of amorphization, as MOFs contain large void spaces that can collapse, reducing the accessible surface area. This can be either a desired change or indeed an unwanted side effect of the application of pressure. In either case, understanding the MOF’s pressure response is extremely important. Three such MOFs with varying pore sizes (UiO-66, MOF-808, and NU-1000) were investigated using in situ high-pressure X-ray diffraction and Raman spectroscopy. Partial crystallinity was observed in all three MOFs above 10 GPa, along with some recovery of crystallinity on return to ambient conditions if the frameworks were not compressed above thresholds of 13.3, 14.2, and 12.3 GPa for UiO-66, MOF-808, and NU-1000, respectively. This threshold was marked by an unexpected increase in one or more lattice parameters with pressure in all MOFs. Comparison of compressibility between MOFs suggests penetration of the pressure-transmitting oil into MOF-808 and NU-1000. The survival of some crystallinity above 10 GPa in all of these MOFs despite their differing pore sizes and extents of oil penetration demonstrates the importance of high-pressure characterization of known structures.

Nuclear Magnetic Resonance (NMR) Spectroscopy: ~1.5 mg of the MOF was combined with 100-200 µl of D 2 SO 4 and then sonicated in DMSO-d 6 (dimethyl sulfoxide). NU-1000 was also heated to ~80 °C to facilitate dissolution. UiO-66 data were acquired on a 400 MHz Bruker Avance III spectrometer, using a QNP probe at 25 °C. MOF-808 and NU-1000 data were acquired on a 500 MHz Bruker Avance III HD spectrometer, using a DCH cryogenically cooled probe with a sample temperature of 25 °C.
Thermogravimetric analysis (TGA): TGA was performed with a simultaneous differential scanning calorimetry (DSC)/TGA thermal analysis (TA) instrument Q600 under an argon flow of 20 mL min -1 , with a heating rate of 20 K min -1 in an alumina pan. TGA under air was performed with a TA Instruments SDT650, with gas flow 50mL min -1 , and a heating rate of 10 K min -1 .
Fourier-Transform Infrared (FTIR) Spectroscopy: FTIR spectra were collected from KBr pellets using a Bruker Tensor 27 FTIR spectrometer in transmission mode between 550 and 4000 cm -1 . Pellets were prepared by dispersing a small amount of powdered sample in KBr and compressing in a 13 mm diameter pellet die for 10 minutes at 10 tons using a pellet press. A background from a pristine KBr pellet was then subtracted from all spectra prior to analysis. Samples were pelletised at 740 MPa for 10 minutes, and IR spectra were measured with a Bruker Tensor 27 FTIR spectrometer. Figure S1. Pawley fits of X-ray data for (a) UiO-66, (b) MOF-808 and (c) NU-1000 against a literature CIF using space groups Fm-3m, Fd-3m and P6/mmm respectively (λ = 1.5406 Å). 6

Microanalysis
UiO-66 -Through 1 H-NMR spectroscopy 1,4-BDC was observed in UiO-66 along with trace amounts of DMF (Fig. S3). Elemental analysis confirmed the presence of DMF and also showed the presence of Clions and excess water. TGA showed an average of 5.6 linkers per SBU, leaving 0.8 vacant coordination sites per SBU for potential coordination of the three other species (Fig. S6, Table S7).
Capping mechanisms have been shown to involve the addition of one -1 ion and one neutral species for every missing linker, with Cl -/H 2 O lowering the free energy more than Cl -/DMF. 9 Any additional presence of species was compensated for by adding adsorbents. Therefore, the formula most closely fitting the elemental analysis was Zr 6 O 4 (OH) 4 (OH 2 ) 0. 8

Assignation Shift (ppm) Calculated shift (ppm) Shape
TBAPy   10 The initial MOF is assumed to be entirely hydrated apart from the formates ions indicated by NMR, the 400 °C phase is dehydrated and has lost formates, 11 and a decomposition product of ZrO2. Oxygen compensation was used to ensure charge neutrality for dehydrated MOF-808 and NU-1000, as this reduces error in the method. 12 Theoretical mass loss for decomposition of linkers in perfect pristine crystal given.  Each spectrum shows a peak at 1390 cm -1 is from the symmetric carboxylate stretch, and at 1578 cm -1 from the asymmetric. The shoulder at 1550 cm -1 on the latter peak in all MOFs indicates a monodentate carboxylate, which is likely from the pelletisation procedure beforehand. 13,14 Figure S14. Raman spectroscopy of UiO-66 before compression (black) and after decompression (red) from a maximum pressure of (a) below the reversibility threshold and (b) above the threshold. Maximum pressure the sample was brought to is shown in the top right. 1300-1350 cm -1 range excluded to remove saturating diamond signal. Figure S15. Raman spectroscopy of MOF-808 before compression (black) and after decompression (red) from a maximum pressure of (a) below the reversibility threshold and (b) above the threshold. Maximum pressure the sample was brought to is shown in the top right. 1300-1350 cm -1 range excluded to remove saturating diamond signal. Figure S16. Raman spectroscopy of NU-1000 before compression (black) and after decompression (red) from a maximum pressure of (a) below the reversibility threshold and (b) above the threshold. Maximum pressure the sample was brought to is shown in the top right. 1300-1350 cm -1 range excluded to remove saturating diamond signal. Figure S17. Raman spectroscopy of a Silicone Oil AP100 droplet on a glass slide under ambient conditions.

MOF Vibration frequency (cm -1 ) Assignment
All 185 Cluster torsion mode with contribution from silicone oil

Equation of state fitting
As stated in the main text, fitting of the entire data range was not possible with a single equation of state. Therefore, the first regime was selected to give an initial bulk modulus to compare with the literature and the second to give the relative compressibility for comparison between the MOFs. Regimes were selected through observation of discontinuities in the volumetric compressibility.