Ionothermal Synthesis of Imide‐Linked Covalent Organic Frameworks

Abstract Covalent organic frameworks (COFs) are an extensively studied class of porous materials, which distinguish themselves from other porous polymers in their crystallinity and high degree of modularity, enabling a wide range of applications. COFs are most commonly synthesized solvothermally, which is often a time‐consuming process and restricted to well‐soluble precursor molecules. Synthesis of polyimide‐linked COFs (PI‐COFs) is further complicated by the poor reversibility of the ring‐closing reaction under solvothermal conditions. Herein, we report the ionothermal synthesis of crystalline and porous PI‐COFs in zinc chloride and eutectic salt mixtures. This synthesis does not require soluble precursors and the reaction time is significantly reduced as compared to standard solvothermal synthesis methods. In addition to applying the synthesis to previously reported imide COFs, a new perylene‐based COF was also synthesized, which could not be obtained by the classical solvothermal route. In situ high‐temperature XRPD analysis hints to the formation of precursor–salt adducts as crystalline intermediates, which then react with each other to form the COF.


Sorption
Sorption measurements were performed on a Quantachrome Instruments Autosorb iQ MP with Argon at 87 K. The pore size distribution was determined from argon adsorption isotherms using the quenched solid-state density functional theory (QSDFT) for cylindrical pores in carbon model for argon at 87 K. XRPD X-ray powder diffraction (XRPD) measurements were performed on a Stoe Stadi-P diffractometer in Debye-Scherrer geometry with Cu-Kα1 radiation equipped with a Ge(111) primary monochromator. The glass capillaries (0.7 mm in diameter) were spun during data collection for an improved particle statistics. Rietveld [1] refinements of the different COF structures were performed using TOPAS V5. [2] Model structures created by Material Studio were used for the Rietveld refinements with fixed atomic coordinates. The peak profile of the XRPD patterns was described by applying the fundamental parameter [3] approach as implemented in TOPAS. Lattice parameter (a, b-axes) were freely refined. The c-axis was taken from the calculation, since no reliable information about stacking distance was obtained from XRPD patterns. The background was modeled by Chebychev polynomials. The microstructure of the different COFs was modeled using microstrain (Lorentzian and Gaussian components).

In situ XRPD
In situ high temperature XRPD experiments were performed on a Bruker D8 diffractometer (Mo-Kα1 equipped with a primary Ge(220)-Johansson-type monochromator and a LynxEye position sensitive detector) in Debye-Scherrer geometry using a water cooled capillary furnace (mri Physikalische Geräte GmbH). The measurements were performed in the range of 2 -25 ° 2θ with 1 h exposure time. For better particle statistics, the quartz capillary was spun during data collection. XRPD patterns were measured at RT,180,210,240,270,290,300 with staying at 300 °C for 10 h and then cool down to RT again.
3 NMR Solid state nuclear magnetic resonance spectra (ssNMR) were recorded on a Bruker Avance III 400 MHz spectrometer (magnetic field 9.4 T). For ssNMR spectroscopy, the samples were packed in 4 mm ZrO2 rotors, which were spun in a Bruker WVT BL4 double resonance MAS probe. Chemical shifts were referenced relative to tetramethylsilane. The spinning rate was 12-14 kHz and a standard cross-polarization sequence with a 2 ms ramped contact pulse was used for 13 C and a total of 4096-8192 scans were routinely accumulated. The measurements were performed under conditions of high-power broadband proton decoupling (SPINAL 64) with the spectral conditions being optimized for the shortest relaxation delay by measuring 1 H T1 relaxation time.

TEM
TEM was performed with a Philips CM30 ST (300kV, LaB6 cathode). The samples were prepared dry onto a copper lacey carbon grid (Plano). The line scan analysis was done with ImageJ, version 1.52a.The simulation of the projected potential map was performedwith the jems package (Stadelmann).

Quantum-Chemical Calculations
Atom positions and lattices of all periodic structures were optimized on RI-PBE-D3/def2-TZVP [4] level of theory using an acceleration scheme based on the resolution of the identity (RI) technique and the continuous fast multipole method (CFMM [5] ) implemented [6] in Turbomole [7] version V7.3. The CFMM uses multipole moments of maximum order 20, together with a well-separateness value of 3 and a basis function extent threshold of 10E-9 a.u. Grid 7 was used for the numerical integration of the exchange-correlation term. The norm of the gradient was converged to 10E-4 a.u. and the total energy to 10E-8 Hartree within the structure optimization using the gamma point approximation. Structures for all investigated molecular compounds were optimized on PBE0-D3/def2-TZVP [4a, 4d, 8] level of theory. Subsequent frequency calculations were performed on the same level of theory to ensure all minima to be true minima on the potential energy hypersurface. Partial Charges were extracted from the summary of a subsequent Natural Population Analysis [9] on the same level of theory. NMR chemical shifts were obtained on B97-2/pcSseg-2 [10] level of theory using the FermiONs++ program package [11] .

Scherrer Analysis
To determine the mean size of the crystalline domains Scherrer analysis was performed using the Scherrer equation displayed below. All reflections of the COFs or phases of the in situ measurement without a significant overlap with other reflections were fitted using the Voigt function model. FWHM and Bragg angles were received from the corresponding fit function and inserted into the Scherrer equation together with the respective wavelength and the shape factor fixed at K = 0.9. To yield the most accurate crystallite sizes as many reflections as possible were considered for the final calculations. Since the quality of the synthesized COFs varies signific antly, not every COF could be described with the same number of reflections.

Materials
Tris(4-aminophenyl)triazine was synthesized according to a literature procedure. [12] All other chemicals were obtained from commercial sources and were used as received.

Ionothermal Synthesis of PI-COFs in ZnCl2
Synthesis of TAPB-PMDA-COF. Typically, pyromelltic dianhydride (21.8 mg, 0.1 mmol), 1,3,5-Tris(4-aminophenyl)benzene (25.2 mg, 0.67 mmol) and anhydrous zinc chloride (114 mg, 0.83 mmol) were ground thoroughly in a mortar under inert atmosphere and transferred into a quartz tube. The quartz tube was evacuated (< 10 -2 mbar), flame sealed and heated to 300 °C for 48 h in a tube furnace. The ampoule was allowed to cool down to room temperature and opened. The crude product was ground carefully in a mortar and subsequently washed with 1M HCl, water and THF. Solvent exchange was carried out by Soxhlet extraction with methanol overnight. The solvent was removed via supercritical CO2 drying to afford TAPB-PMDA-COF as a dark brown powder.
Synthesis of TAPB-PTCDA-COF. Typically, Perylene-3,4,9,10-tetracarboxylic dianhydride (31.2 mg, 0.08 mmol), 1,3,5-Tris(4aminophenyl)benzene (20 mg, 0.053 mmol) and anhydrous zinc chloride (722 mg, 5.3 mmol) were ground thoroughly in a mortar under inert atmosphere and transferred into a quartz tube. The quartz tube was evacuated (< 10 -2 mbar), flame sealed and heated to 300 °C for 48 h in a tube furnace. The ampoule was allowed to cool down to room temperature and opened. The crude product was ground carefully in a mortar and subsequently washed with 1M HCl, water and THF. Solvent exchange was carried out by Soxhlet extraction with methanol overnight. The solvent was removed via supercritical CO2 drying to afford TAPB-PTCDA-COF as a dark brown powder.

Ionothermal Synthesis of PI-COFs in eutectic salt mixture
Synthesis of TT-PMDA-COF. First, sodium chloride (12.4 mg, 0.21 mmol), potassium chloride (39.5 mg, 0.53 mmol) and anhydrous zinc chloride (114 mg, 0.83 mmol) were ground together thoroughly in a mortar under inert conditions. Then, pyromelltic dianhydride (21.8 mg, 0.1 mmol) and 1,3,5-Tris(4-aminophenyl)triazine (23.6 mg, 0.067 mmol) were added and the mixture was ground again to achieve homogeneity and transferred into a quartz tube. The quartz tube was evacuated (< 10 -2 mbar), flame sealed and heated to 250 °C for 48 h in a tube furnace. The ampoule was allowed to cool down to room temperature and opened. The crude product was ground carefully in a mortar and subsequently washed with 1M HCl, water and THF. Solvent exchange was carried out by Soxhlet extraction with methanol overnight. The solvent was removed via supercritical CO2 drying to afford TT-PMDA-COF as a light brownyellow powder.
Synthesis of TAPA-PMDA-COF. First, sodium chloride (12.4 mg, 0.21 mmol), potassium chloride (39.5 mg, 0.53 mmol) and anhydrous zinc chloride (114 mg, 0.83 mmol) were ground together in a mortar thoroughly under inert conditions. Then, pyromellitic dianhydride (21.8 mg, 0.1 mmol) and Tris-(4-aminophenyl)amine (19.3 mg, 0.067 mmol) were added and the mixture was ground again to achieve homogeneity and transferred into a quartz tube. The quartz tube was evacuated (< 10 -2 mbar), flame sealed and heated to 250 °C for 48 h in a tube furnace. The ampoule was allowed to cool down to room temperature and opened. The crude product was ground carefully in a mortar and subsequently washed with 1M HCl, water and THF. Solvent exchange was carried out by Soxhlet extraction with methanol overnight. The solvent was removed via supercritical CO2 drying to afford TT-PMDA-COF as a brown powder.    8 Figure S6. Structure of a single pore of the TAPB-PMDA COF, obtained by composing a supercell from a 2D periodic geometry optimization of the corresponding asymmetric unit on RI-PBE-D3/def2-TZVP level of theory.              Resulting COF after washing Figure S28. Summary of three individual polymerization reaction pathways, supplemented with calculated reaction enthalpies as total energy differences of the sum of reactants and products, obtained on PBE0-D3/def2-TZVP level of theory.

̅ 1 symmetry (blue) and
Cmcm symmetry (red). Reflections marked with an asterisk show minor impurities, which couldn't be eliminated.  Figure S51. Pawley refinement of TAPA-PMDA-COF (left) and TT-PMDA-COF (right) with fixed lattice sizes and fit parameters obtained from Rietveld refinement.
Reflections marked with an asterisk show minor impurities, which couldn't be eliminated.