Carbon dioxide capture and efficient fixation in a dynamic porous coordination polymer

Direct structural information of confined CO2 in a micropore is important for elucidating its specific binding or activation mechanism. However, weak gas-binding ability and/or poor sample crystallinity after guest exchange hindered the development of efficient materials for CO2 incorporation, activation and conversion. Here, we present a dynamic porous coordination polymer (PCP) material with local flexibility, in which the propeller-like ligands rotate to permit CO2 trapping. This process can be characterized by X-ray structural analysis. Owing to its high affinity towards CO2 and the confinement effect, the PCP exhibits high catalytic activity, rapid transformation dynamics, even high size selectivity to different substrates. Together with an excellent stability with turnover numbers (TON) of up to 39,000 per Zn1.5 cluster of catalyst after 10 cycles for CO2 cycloaddition to form value-added cyclic carbonates, these results demonstrate that such distinctive structure is responsible for visual CO2 capture and size-selective conversion.

Zn-DPA. As-synthesized single crystals were air-dried. One single crystal suitable for the diffraction measurement was put into a glass capillary. The capillary was connected to handmade gas pressure handling unit, and was evacuated (below 10 -2 Pa) at 120 °C overnight. Then, the capillary was sealed using small torch flame with keeping vacuum condition inside the capillary. The sealed capillary was mount on the diffractometer at 183 K.
Refinement details of Zn-DPA. All framework atoms were located and refined anisotropically.

Response to A level check cif alerts for Zn-DPA single crystal:
Zn-DPA.cif contains one alert A as follows.

Alert level A
PLAT602_ALERT_2_A VERY LARGE Solvent Accessible VOID(S) in Structure ! Info Response: Application of procedure SQUEEZE (program PLATON) did not bring about a significant improve of refinement results and therefore was not retained for the final refinement. This might be due to the existence of a large solvent channel in the structure itself.

Selective bond distance (Å) and angle (°) in Zn-DPA.
Selective bond distance (Å): Zn (1) (1) 93.32 (6) Zn-DPA•2CO 2 . As-synthesized single crystals were air-dried. One single crystal suitable for the diffraction measurement was put into a glass capillary. The capillary was connected to handmade gas pressure handling unit, and was evacuated (below 10 -2 Pa) at 120 °C overnight. Then, CO 2 was slowly introduced into the capillary until the pressure reaching 101 kPa at 195 K. After 10 min at 195 K under 101 kPa CO 2 atmosphere, the glass capillary was sealed using small torch flame. The sealed capillary was mount on the diffractometer at of residual electron density were observed inside the framework pores and were assigned as accommodated CO 2 molecules. Because CO 2 molecule at C (site II) was disordered into two positions (here we name them as C and CC), PART instruction was used (PART 1 for "O1C-C1C-O2C"; PART 2 for "O1CC-C1CC-O2CC").
Some thermal and structural restraints (ISOR, SADI, DFIX, DANG) were also used for the structural refinements of CO 2 molecules at B and C (site II and III). As the sorption measurements showed that this compound adsorbed 2CO 2 molecules per Zn 1.5 under 101 kPa CO 2 atmosphere at the measurement temperature, we can expect the sum total number of CO 2 at A, B, C, and CC (site I, III, II) to be two. Therefore, Free variables were introduced for the occupancy refinements of all CO 2 molecules with using SUMP instruction (SUMP 2 0.00001).

Computational Details.
The experimentally observed adsorption isotherm of CO 2 suggests that 16 molecules of CO 2 can be absorbed into one unit cell of Zn-DPA at the saturated limit. To locate the adsorption positions of these CO 2 molecules, we carried out canonical Monte-Carlo (MC) simulations, 1 as implemented in RASPA, 2 using Lennard-Jones (LJ) potentials to describe the Van der Waals interaction of CO 2 with Zn-DPA framework. The LJ parameters were taken from the standard universal force field (UFF) 3 and TraPPE 4 for Zn-DPA and CO 2 , respectively, where the Lorentz-Berthelot mixing rules were used for different atoms. The electrostatic interaction was evaluated with the Ewald summation method, where the DDEC atomic charges 5,6 were used. The crystal structure of Zn-DPA containing 8 CO 2 molecules in one unit cell has been observed in experiment. Using that structure, we started the MC simulation to find adsorption positions of other CO 2 molecules by placing 64 CO 2 molecules, which correspond to adsorption of 16 CO 2 molecules into one unit cell, in a simulation box of 2×2×2 supercell of Zn-

DPA.
In the MC simulation, the first 2×10 5 cycles were consumed for obtaining equilibration and then 5×10 5 cycles were used for obtaining distribution of CO 2 molecules at room temperature. In the final configuration, we found that the positions of CO 2 molecules could be classified into three groups, suggesting that there are three possible sites, namely sites I, II, and III, for CO 2  where E(Zn-DPA•nCO 2 ) eq is the total energy of Zn-DPA with n molecules of CO 2 per one unit cell in equilibrium structure (eq), E(Zn-DPA) eq and E(CO 2 ) eq are the total energies of empty Zn-DPA and one free To make analysis of CO 2 interaction with Zn-DPA, three cluster models (CMs) were constructed for the sites I, II, and III, using the PBE-D3 optimized structure. These cluster models were further decomposed into several small cluster models (SCMs) to elucidate important interactions that stabilise CO 2 adsorption at these sites, as shown in Supplementary Fig. 14