Covalent Triazine Frameworks via a Low‐Temperature Polycondensation Approach

Abstract Covalent triazine frameworks (CTFs) are normally synthesized by ionothermal methods. The harsh synthetic conditions and associated limited structural diversity do not benefit for further development and practical large‐scale synthesis of CTFs. Herein we report a new strategy to construct CTFs (CTF‐HUSTs) via a polycondensation approach, which allows the synthesis of CTFs under mild conditions from a wide array of building blocks. Interestingly, these CTFs display a layered structure. The CTFs synthesized were also readily scaled up to gram quantities. The CTFs are potential candidates for separations, photocatalysis and for energy storage applications. In particular, CTF‐HUSTs are found to be promising photocatalysts for sacrificial photocatalytic hydrogen evolution with a maximum rate of 2647 μmol h−1 g−1 under visible light. We also applied a pyrolyzed form of CTF‐HUST‐4 as an anode material in a sodium‐ion battery achieving an excellent discharge capacity of 467 mAh g−1.


S2
SPM-9700 instrument (Shimadzu Japan). The samples were dispersed in ethanol and ultrasonicated for 6 h at 25 °C and then two drops of suspension were taken onto the mica sheet and allowed to dry. Raman spectra were recorded on a confocal laser Raman spectrometer LabRAM HR800 (Horiba JobinYvon, France). X-ray photoelectron spectroscopy (XPS) analysis were conducted on an Axis Ultra DLD 600 W instrument (Shimadzu, Japan).
The photocatalytic experiments were performed under visible light irradiation (>420nm) with 300W Xe lamp (Perfectlight, PLS-SXE300). In this system, 50 mg photocatalysts were dispersed in 100ml aqueous solution which contained 10 ml TEOA. 3wt% Pt here was added as cocatalyst by in situ photodeposition method using H 2 PtCl 6 . The temperature of the solution was maintained at room temperature by the flow cooling water during the reaction. The hydrogen evolution was analyzed by gas chromatography (SHIMADZU, GC-2014 C) equipped with a thermal conductive detector (TCD) and 5A molecular sieve column using N 2 as the carrier gas.
The electrochemical properties of samples were measured on a Hokudo Denko charge/discharge instrument by using a 2032-type coin cell. Na metal foil was utilized as the reference and counter electrode. The electrolyte was 1M NaPF 6 in ethyl carbonate (EC) and diethyl carbonate (DEC) (EC: DEC=1:1 by volume). 15wt% acetylene black (AB) was used as a conductive agent and 5wt% polyvinylidene fluoride (PVDF) as a binder. A Cu foil was used as current collector. The cells were assembled in a glove box filled with pure argon gas. Galvanostatic discharge/charge measurements were performed in a potential range of 3 V-50 mV vs Na + /Na.

Synthesis of tri(4-formylphenyl)amine S3 (Monomer 3: M3):
Phosphorous oxychloride (60.0 mL) was added dropwise to a stirred N,N-dimethylformamide (45.9 mL, 0.59 mol) at the temperature of ice water under nitrogen atmosphere. The mixture was stirred at 0 °C for another 1 h, and the solution turned pink. After the addition of triphenylamine (3.68 g, 15.0 mmol), the temperature was allowed to rise to 95 °C. After 72 h, the solution was poured into cold water. The resulting mixture was neutralized to pH = 7-8 with aqueous NaOH solution (20 wt %), and yellow solid was obtained. The solid was chromatographed on a silica gel column to produce 1.88 g of tri(4formylphenyl)amine in 38% yield. 1

Synthesis of tri(4-formylbiphenyl)amine S3 (Monomer 4: M4):
Tris(4-iodophenyl)amine (623 mg, 1.00 mmol) and 4-formylphenylboronic acid (540 mg, 3.60 mmol) S4 were dissolved in 20 mL of tetrahydrofuran. Aqueous solution of potassium carbonate (10.0 mL, 2.0 mol L -1 ) was added into the solution under nitrogen atmosphere. After the addition of bis(triphenylphosphine)palladium(II) dichloride (105 mg), the mixture was refluxed for 12 h. The solution was extracted twice with dichloromethane (3 × 100 mL). The obtained organic layer was washed with plenty of water and the solvent was removed at reduced pressure. The residue was chromatographed on a silica gel column to give yellow solid 0.47 g tri(4-formylbiphenyl)amine with 85 % yield. 1  Amplified synthesis of CTF-HUST-1. 1,4-Phthalaldehyde (2.01 g,15.0 mmol), terephthalamidine dihydrochloride (7.05 g, 30.0 mmol), cesium carbonate (19.55 g, 60.0 mmol) were added to a solution of DMSO (100.0 mL) and H 2 O (4.0 mL) in 500 mL round-bottomed. The mixture were strirred by mechanical agitation at 60 °C for 12 h, then heated at 80 °C, 100°C for 12 h separately, and heated at 120 °C for 3 days, yielding a canary yellow solid. The resulting precipitate was washed with diluted HCl (3×100 mL) to remove the salt and residual cesium carbonate, and washed with water (3×300 mL), acetone (3×300 mL) and THF (3×300 mL), dried at 80 °C under vacuum for 12 h to yield CTF-HUST-1 as a yellowish power (5.61 g, 81% yield). were added to a solution of DMSO (5.0 mL) and H 2 O (0.2 mL) in 25 mL round-bottomed. The mixture was heated at 60 °C for 12 h, then heated at 80 °C, 100°C for 12 h separately, and heated at 120 °C for 3 days, yielding a yellowish solid. The resulting precipitate was washed with diluted HCl (3×10 mL) to remove the salt and residual cesium carbonate, and washed with water (3×10 mL), acetone (3×10 mL) and THF (3×10 mL), dried at 80 °C under vacuum for 12 h to yield CTF-HUST-2 as a power (249.9 mg, 93% yield).  Figure S12f). Two broad peaks were clearly observed, which is consistent with carbonaceous materials with graphene-like structures. S6 The peaks at 22°-25° derived from the diffraction of the layer structures, whereas 43° originated from the disordered carbon structures. S18 Figure S13.  Figure S14b). We also found p-CTF-HUST-4 to be a stable SIB anode material. Even after 150 cycles, it delivered 303 mAh g-1 with a slight decrease in capacity and a high Coulombic efficiency of 99.9%. p-CTF-HUST-4 also exhibited excellent rate capabilities, with high capacity of 514 mAh g -1 at 25 mA g-1, 291 mAh g -1 at 100 mA g -1 , and 97 mAh g -1 at 400 mA g -1 , as shown in Figure S14c. These results indicate that there are some remarkable features of p-CTF-HUST-4 as an anode material for Na + storage, such as good compatibility with electrolyte, and significant improvement of the sodium transportation kinetics and the electron conductivity, thus enhancing the reversible capacity and rate capability. The electrode kinetics investigated by electrochemical impedance spectrum (EIS) are shown in Figure S14d. In high frequency areas, it is found that the diameter of the semicircle at the initial state and after 150 cycle is nearly the same, suggesting the low con-tact/charge-transfer impedances. It should be noted that there is a steep low-frequency tail after 150 discharging/charging, indicating high sodium ions diffusivity. S20 NMR spectra Figure S15. 1 H NMR spectrum of model compound Figure S16. 13 Table S3. The optimized simulation lattice parameters.

The simulation and optimized configuration of CTF-HUSTs：
The starting configurations were built by Accelrys' Materials Studio modelling software. The theoretical crystalline structures were simulated by using Reflex module of the Materials Studio program. The electronic structure and total energy were calculated using density functional theory (DFT) S7 via the plane-wave pseudopotential (PWPP) technique implemented in the Vienna ab initio simulation package (VASP). S8 The projector-augmented wave (PAW) S9,S10 method was used to represent the ion-electron interaction. The generalized gradient approximation (GGA) S11 expressed by the PBE functional S12 and a 400 eV cutoff for the plane-wave basis set were adopted in all calculations. The convergence threshold was set as 10 6 eV in energy and the atoms were relaxed toward equilibrium until the Hellmann−Feynman forces were less than 10 2 eV/Å. Brillouin-zone integration was performed with a Gaussian broadening of 0.1 eV during all relaxations. Due to the stacking of layers in CTFs, it is important to account for dispersion effects, thus, DFT+D3 was used to afford the dispersion interactions between adjacent layers of CTFs. S13 Due to the large number of atoms in each CTF system, we use the Γ-point in all VASP calculations for the balance between accuracy and computational cost. Due to the large number of atoms in each CTF system, these calculations can be conducted with a single K-point at the Γ-point. Given that the size of system and the balance between accuracy and computational cost, the K-point sampling was selected (Table S3) to build models of possible perfectly ordered network structures to compare with our experimental data. The unit cell structures were built with hexagonal geometry and the theoretical XRD was simulated by Materials Studio software (Figure 1a and Figure S6-S7). The unit cells were constructed using hexagonal geometry with P6/mmm space groups. As shown in Table S3, Figure 1 and Figure S6-S7. The absence of any features in the 2 Theta = 10-15° range suggests that AA layer stacking are perhaps more likely, but the experimental features are too broad to allow any definitive model. For CTF-HUST-1, the proposed structure is most consistent with the previously reported CTF-1, which was reported to have AA stacking (Figure 2a, black curve). S14 Indeed, the PXRD data for CTF-HUST-1 is quite similar to that reported for CTF-1, S14 although for CTF-HUST-1 the peaks are broader. CTF-HUST-2, CTF-HUST-3 and CTF-HUST-4 form a new series of S29 structures with geometries that are dissimilar from any previously reported CTFs owing to their distinctive linkage mode ( Figure S7). Unlike previous CTFs prepared by self-cyclization of nitrile building blocks, our new method involves two types of building blocks-an aldehyde and an amidinewhich undergo a multi-step condensation process during the polymerization, as shown in Scheme 1a.
Two equivalents of terephthalamidine monomers to aldehyde monomers are required in the cyclization reaction, allowing for CTFs with mixed functionalities. Hence, CTF-HUST-1 (and the earlier CTF-1) consists of one type of pore, S14 whereas there are theoretically two types of pores in CTF-HUST-2, CTF-HUST-3 and CTF-HUST-4 (Scheme 1). CTF-HUST-1 has a theoretical pore size at 12 Å according to the perfectly crystalline model. By contrast, CTF-HUST-2 consists of two theoretical pore sizes at 12.0