Integrating single Ni sites into biomimetic networks of covalent organic frameworks for selective photoreduction of CO2

Fabrication of biomimetic photocatalytic systems consisting of PI-COFs and molecular Ni complexes for selective reduction of CO2 is demonstrated.


Material Synthesis
Synthesis of functional bpy: 5,5'-diamino-2,2'-bipyridine (46.5 mg, 0.25 mmol) and 3,5-dimethylbenzaldehyde (67 mg, 0.50 mmol) were added to 5 mL of THF and 0.5 mL of 6 M CH 3  The auxiliary electrode was a platinum wire. The reference electrode was based on the Ag/AgCl electrode. The working electrodes were prepared as follow: A suspension of 2 mg sample in 0.5 mL Nafion solution (1% in ethanol) were sonicated for 30 min to make it dispersible, and then dropping 20 μL of the suspension onto the carbon glass electrode. The potential vs RHE was calibrated as E RHE = E Ag/AgCl + 0.61.

Computational Details.
The optimization of three PI-COFs was performed on the projected augmented wave 1 formalism of DFT via the VASP package 2 . We used generalized gradient approximation with a Perdew-Burke-Ernzerhof (PBE) 3 form for the exchangecorrelation functional. A cutoff energy of 400.0 eV was used, and the Brillouin-zone integration was sampled using a (4 4 2) Monkhorst-Pack mesh. The total energy was × × converged to 10 -5 eV. All the atomic positions were optimized until the force tolerance on each atom was less than 0.01 eV/Å. The supercell of 2 2 1 was adopted for further × × properties calculation.
We further used Dmol3 package 4,5 to evaluate the absorption properties of CO 2 and H 2 O. The exchange-correlation term was considered using the generalized gradient approximation (GGA) proposed by the Perdew, Burke, and Ernzerhof (PBE) 3 , in which the Grimme's DFT-D corrections were adopted. The double numeric quality basis set with polarization functions (DNP) 4,6 was adopted, which was comparable to 6-31G**. 7,8 The numerical basis sets can minimize the basis-set superposition error. A Fermi smearing of 0.005 hartree was utilized. The tolerances of the energy, gradient and displacement convergence were 2  10 -5 hartree, 4  10 -3 hartree per Å, and 5 10 -3 Å, respectively. Herein, the initial structure data of COF was obtained from the former VASP calculations. The isolated system models were taken from a 2 2 1 supercell × × with terminal groups replaced as hydrogen atoms. Then only the nickel complex optimized when the PI-COF-TT added and constrained. Subsequently, the constrained calculations were applied during CO 2 and H 2 O absorption calculation, only CO 2 and H 2 O were relaxed while other species were kept fixed.
The absorption of CO 2 to Ni intensified by the addition of PI-COF-TT, in which one hydrogen bond formed between COF and CO 2 with 1.85 Å length. The angel of O=C=O of CO 2 change from 175.3 to 155.8 when COF added. Additionally, the Ni-C bond length of Ni complex and CO 2 shorten from 3.09 Å to 2.24 Å. The absorption energy of CO 2 to Ni complex in PI-COF-TT was -101.5 kcal/mol, which was more favorable than the H 2 O absorbed onto Ni complex in PI-COF-TT process (-12.6 kcal/mol).

Characterization
The starting materials were commercially available and used without further purification. 1 H NMR spectra was recorded on a Bruker AVANCE III NMR spectrometer at 400 MHz, respectively, using tetramethylsilane (TMS) as an internal standard. Solid-state 13 C CP/MAS NMR was performed on a Bruker SB Avance III 500 MHz spectrometer with a 4-mm double-resonance MAS probe. FTIR spectra were recorded with KBr pellets using Perkin-Elmer Instrument. Powder X-ray diffraction (XRD) patterns were recorded in the range of 2θ = 3-40 o on X'pert3 X-ray diffractometer with Cu Kα radiation (λ = 1.5406 Å). N 2 or CO 2 adsorption and desorption isotherms were measured at 77 K using a Micromeritics ASAP 2020 system. The samples were degassed at 120 °C for 10 h before the measurements. Surface areas were calculated from the adsorption data using Brunauer-Emmett-Teller (BET) equation. The BET surface area was calculated from the range of 0.05<P/P 0 <0.25 in the isotherm. The calculation of the pore size distribution was done using the nonlocal density functional theory (NLDFT) equilibrium model. Field-emission scanning electron microscopy (SEM) was performed on a JEOL JSM-7500F operated at an accelerating voltage of 3.0 kV. The CO gas produced from 13 CO 2 isotope experiments was examined by a gas chromatograph−mass spectrometer (GC-MS, Agilent 7890B-5977B). The equipped column in GC-MS analysis was CP-Molsieve 5A (Agilent Technologies, 25.0 m x 0.32 mm x 30 μm). X-ray photoelectron spectroscopy (XPS) measurements were performed on a Thermo ESCALAB 250 spectrometer, using nonmonochromatic Al Kα x-rays as the excitation source and choosing C 1s (284.6 eV) as the reference line. UV-Vis spectra were recorded using a Agilent Cary 5000 spectrometer. Fluorescence spectra were recorded at room temperature using a FM-4 spectrophotometer, and the slit width for emission was 2 nm. Electron paramagnetic resonance (EPR) measurements were carried out on a Bruker model A300 spectrometer. The liquid phase product was analyzed by a HPLC (Waters e2695).
Aberration-corrected high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) was performed on FEI Themis Z.        A "sphere-like" morphology of PI-COF-TT can also be obtained as previously reported by the increasing the crystal intensity of pyromellitic dianhydride. The different morphology was probably ascribed to the concentration of pyromellitic dianhydride in the reaction solution. With the increment of the crystal intensity of pyromellitic dianhydride, the concentration of pyromellitic dianhydride in the reaction solution reduced, resulting in the formation of "sphere-like" morphology.