In Situ Single-crystal X-ray Diffraction Studies of Physisorption and Chemisorption of SO2 within a Metal–Organic Framework and Its Competitive Adsorption with Water

Living on an increasingly polluted planet, the removal of toxic pollutants such as sulfur dioxide (SO2) from the troposphere and power station flue gas is becoming more and more important. The CPO-27/MOF-74 family of metal–organic frameworks (MOFs) with their high densities of open metal sites is well suited for the selective adsorption of gases that, like SO2, bind well to metals and have been extensively researched both practically and through computer simulations. However, until now, focus has centered upon the binding of SO2 to the open metal sites in this MOF (called chemisorption, where the adsorbent–adsorbate interaction is through a chemical bond). The possibility of physisorption (where the adsorbent–adsorbate interaction is only through weak intermolecular forces) has not been identified experimentally. This work presents an in situ single-crystal X-ray diffraction (scXRD) study that identifies discrete adsorption sites within Ni-MOF-74/Ni-CPO-27, where SO2 is both chemisorbed and physisorbed while also probing competitive adsorption of SO2 of these sites when water is present. Further features of this site have been confirmed by variable SO2 pressure scXRD studies, DFT calculations, and IR studies.

For the activated (dehydrated) sample, the water bound oxygen was constrained so the Uiso was of the same order as the metal to which it was bound, and was kept isotropic.This was to allow easy comparison of different samples during dehydration without contribution from the pore environment.
For the SO 2 loaded sample at 450 K all the SO bonds were subject to a DFIX restraint of 1.43 Å the gas bond length. 6All the atoms within the SO 2 were also subject to a SIMU restraint of strength 0.16 over a distance of 2.7 Å. The sulphur atoms were subject to a SUMP restraint so that their total occupancy was within error of the metal bound oxygen atom.The outer oxygen atom was not so restrained as it will be to disordered for a SUMP restraint to be appropriate.
For the SO 2 loaded sample at 300 K.The metal bound SO 2 was subject to a number of restraints.The bond length of the sulphur atoms to the outer oxygen atoms were subject to a DFIX restraint of c.The sulphur atoms were subject to a SIMU restraint with the metal bound oxygen atom of strength 0.08 and distance 2.7 Å as well as a SUMP restraint similar to that used at 450 K.The outer oxygen atoms had a separate SIMU restraint of the same intensity.The physisorbed SO 2 was also restrained with a DFIX restraint of 1.43 Å and a SIMU restraint of 0.04 over 2.7 Å with the oxygen atoms kept isotropic.The oxygen atoms were subject to a SUMP restraint so that their occupancies added up to twice the sulphur occupancy.
In order to calculate free pore volume and electron density the Olex2 mask command was used. 7tandard parameters (1.2 Å probe) were used.The mask was not used on the published structures, but as a calculation after refinement was finished.The total, theoretical, dehydrated pore volumes were calculated in Olex2 by removing any metal bound oxygen and calculating porosity with standard parameters giving a value of 2364.7 Å 3 per unit cell.
In order to calculate the number of gas molecules free in the pore from the observed electron densities the following calculation was performed.The activated e -density was taken from the gas loaded e - density to remove any input from disorder within the structure.This corrected value was divided by 32 to obtain the SO 2 molecular density, then multiplied by the free pore volume to obtain the number of SO 2 molecules per unit cell.
For the variable pressure studies different models were used to compare the data.For these data sets comparability between data sets was prioritised over accuracy of the model.
For the loading at 450 K only SO 2 on the metal site could be located.At pressures ≤ 0.05 bar only the metal site oxygen could be modelled, this was done without restraint.Above 0.05 bar sulphur could also be modelled, this site was highly disordered but modelled with only one large sulphur atom which was restrained with a SIMU restraint of strength 0.04.In all points the metal bound oxygen occupancy was taken as the SO 2 occupancy.The data point for 2 bar is taken from the structure outlined above.
For the loading at 300 K up to 0.4 bar a 3 oxygen model was used.This consisted of 3 oxygen environments within the pore (Figure S3).The first was metal bound and was unrestrained, anisotropic and with riding hydrogens.The second two sites sat within the pore, both were constrained to be isotropic with a Uiso of 0.05 without riding hydrogens.The oxygen nearest the framework, a similar position to the physisorbed SO 2 site, was named phys 1.The second site further into the pore was labelled phys 2. At higher pressures a more accurate model could be used, though not as accurate as the structural model described above.The metal bound SO 2 was modelled with 2 S sites and 2 outer O sites.
The sulphur atoms were subject to a SIMU restraint with the metal bound oxygen of strength 0.04 and distance 1.7 Å.The oxygen atoms were constrained to be isotropic with a Uiso of 0.1.The metal bound oxygen and the water oxygen were subject to a eadp constraint.In order to estimate the relative SO 2 : water ratio SUMP restraints were used so that the total occupancy on the metal site was within 0.05 s.d. of 1 and the total sulphur occupancy was similar to the SO 2 metal bound oxygen amount.The S-O bond lengths were subject to a dfix restraint of 1.43 Å the gas bond length. 6The physisorbed SO 2 was modelled with 1 sulphur and 2 oxygen atoms all constrained isotropic with a Uiso of 0.05.The occupancy of all 3 atoms was constrained to be equal.Supplementary References

Figure S2 :
Figure S2: Plot showing thermal gravimetric analysis for Ni-MOF-74 measured in air flow.

Figure S3 :
Figure S3: View (ball and sticks) down the crystallographic c-axis of the model used for SO 2 loaded Ni-CPO-27 at 300 K and ≤0.4 bar SO 2 .

Figure S4 :
Figure S4: Plot showing how SO 2 loading occupancies of Ni-MOF-74 are affected by application of dynamic vacuum at 300K.

Figure S5 :
Figure S5: The experimental setup used to load MOFs with SO 2 .

Table 1 :
Containing information on temperature, loading and R 1 values for each dataset presented.