Recently, environmental pollution issues, for example, the emission of carbon dioxide and the removal of heavy metals from power plants, industrial facilities and human activities, have attracted tremendous attention ^[1−2]. Heavy metals, such as mercury (Hg²⁺), with poor degradability and high solubility in solutions, are extremely harmful for human health, since they could be cycled through the food chain and enriched to the human body, damaging central systems ^[3−7]. In this regard, numerous strategies have been developed for efficient removal of Hg²⁺. Among various techniques, physical adsorption process represents one of the most promising methods, due to its low cost, simplicity and high efficiency ^[8−12]. And significant research efforts have been devoted into the search and development of efficient nanoporous adsorbents. A wide variety of advanced porous materials, like, activated nanoporous carbon ^[13], mesoporous silica ^[14], zeolites ^[15], porous organic polymer ^[16], metal organic frameworks ^[17] and covalent organic frameworks ^[18] have been prepared and attempted for this application.

Covalent triazine frameworks (CTFs) emerged as a novel class of COFs, are typically constructed through trimerization of aromatic nitriles. They have been widely explored in the field of adsorption and separation because of their permanent porosity, high physicochemical stability, and rich nitrogen-containing sites ^[19−20]. For instance, Qin and co-workers have developed a triazine-based porous organic polymer with a maximum mercury uptake of 658.9 mg∙g^{−1 [21]}. Voort’s group ^[22] recently developed a CTF encapsulated with γ-Fe₂O₃ nano-particles, where the Hg²⁺ adsorption capacity was achieved to be 165.8 mg∙g⁻¹. Furthermore, it has been well documented that sulfur-based groups are privileged receptors towards Hg²⁺. As such, these active moieties, such as methylthio units have been extensively introduced into the framework of porous adsor-bents, so as to boost the adsorption performance ^[23−26]. Never-theless, conventional preparation methods for the construction of CTFs through high-temperature-involved ZnCl₂-promoted ionothermal processes significantly restrict the installation of functional methylthio groups within pore walls, on account of partial carbonization of these CTF frameworks under such high synthetic temperatures ^[27−28]. Accordingly, it is rather difficult to obtain an accurate structure-property relation-ship for the removal of Hg²⁺ using CTF-based adsorbents.

With these considerations in mind, we report rational synthesis of a crystalline CTF with methylthio pendant arms for efficient Hg²⁺ adsorption. The key of our preparation lies in a low-temperature-processed condensation reaction between methylthio-functionalized aromatic aldehydes and amidine moieties for the preparation of triazine-linked crystalline frameworks, which avoids the partial carbonization of task-specific methylthio (MT) groups and gives rise to deeper understanding of effect of methylthio groups on the adsorption of Hg²⁺ within CTF. As a result, the resulting adsorbent (MT-CTF) exhibits an exceptional uptake capacity towards Hg²⁺, as high as 751 mg∙g⁻¹, accompanied with high selectivity over a wide range of metal ions. We expect that our findings will open up new possibility in the synthesis of CTFs for promising separation applications.

1 Experimental section 1.1 Chemicals

Terephthalaldehyde, N,N-dimethylformamide (DMF), potassium carbonate (K₂CO₃), tetrahydrofuran (THF), dichloro-methane, methanol, ethanol and a series of ionic standard solutions were purchased from National Medicines Corporation Ltd. of China. 2,5-dibromoterephthalaldehyde, benzene-1,4-dicarbonitrile, and a series of nitrates were obtained from Aldrich Chemical Co., Inc.

1.2 Materials preparation

2,5-bis(methylthio)terephthalaldehyde was prepared foll-owing the literature [29]. The desired compound was achieved with a yield of 87%. ¹H NMR (400 MHz, CDCl₃) δ: 10.42 (s, 2H), 7.80 (s, 2H), 2.58 (s, 6H); ¹³C NMR (101 MHz, CDCl₃) δ: 16.5, 130.1, 136.3, 139.8, 190.7.

The terephthalamidine dihydrochloride was synthesized according to the reported procedure ^[28]. The desired compound was recrystallized from the H₂O-EtOH mixture with a yield of 85%. ¹H NMR (400 MHz, DMSO-d₆) δ: 8.07 (s, 2H), 7.92 (s, 4H), 7.49 (s, 2H). ¹³C NMR (101 MHz, D₂O) δ: 169.80, 137.82, 133.83.

As shown in Scheme 1, the synthesis of MT-CTF adsorbent was shown as the following: 22.6 mg (0.1 mmol) 2,5-bis(methylthio)terephthalaldehyde, 48 mg (0.3 mmol) benzene-1,4-dicarboximidamide dihydrochloride, and 28 mg (0.2 mmol) K₂CO₃ were added into the flask (50 mL), then the mixture was degassed, DMF (5 mL) was added subsequently, then the reaction took place under N₂ at 80 ℃ for 24 h, at 120 ℃ for 48 h, and at 150 ℃ for 72 h, respectively. After that, the precipitate was washed with various solvents to afford MT-CTF. The yellow product was then dried under vacuum at 60 ℃for 12 h. Controlled sample CTF-1 was synthesized according to the literature [30].

1.3 Catalyst characterization

The Bruker model VERTEX 70 infrared spectrometer was used to characterized the Fourier transform infrared spectroscopy (FT-IR) of CTFs. Nuclear Magnetic Resonance (¹H NMR and ¹³C NMR) spectra were performed on a Bruker AV400 (400 MHz for proton, 101 MHz for carbon) spectrometer with tetramethyl silane as the internal reference. The Brunauer-Emmett-Teller (BET) surface area of MT-CTF and CTF-1 were recorded using a Micrometrics 3flex surface area analyzer. The CO₂ adsorption of the MT-CTF and CTF-1 were tested at 25 ℃ simultaneously. The Bruker 500 MHz/ AVANCEⅢ spectrometer was used to characterize the ¹³C cross polarization magic angle spinning nuclear magnetic resonance (CP/MAS ¹³C-NMR) spectroscopy of the CTFs. Scanning electron microscopy (SEM) measurements were conducted with a Hitachi SU 8000 instrument. Thermo- gravimetric analysis (TGA) was performed under the nitrogen flow with a heating rate of 10 ℃∙min⁻¹ using a Mettler Toledo TGA/DSC1 Simultaneous Thermal Analyzer (up to 800 ℃). Bruker RFS 100/S spectrometer and PHI 5000 VersaProbe were used to determine the X-ray photoelectron spectroscopy (XPS) spectra of MT-CTF and CTF-1. The concentrations of Hg²⁺ were analyzed by the Inductively Coupled Plasma-Mass Spectrometry (ICP-MS) and Inductively Coupled Plasma-Optical emission spectroscopy (ICP-OES) (Agilent 725-ES).

1.4 Hg²⁺ batch adsorption experiments

The initial Hg²⁺ stock solution (800 mg∙L⁻¹) was prepared by dissolving Hg(NO₃)₂ (0.3246 g) into distilled water (250 mL). The initial Hg²⁺ solution underwent dilution to obtain different concentrations of Hg²⁺ solutions. HNO₃ and NaOH solutions were used to adjust the pH value of the solution. The concentrations of Hg²⁺ were determined by ICP-OES (mg∙L⁻¹ level) and ICP-MS (μg∙L⁻¹ level). The adsorption experiments were conducted under ambient conditions.

1.5 Hg²⁺ adsorption isotherm analysis

To obtain Hg²⁺ adsorption isotherms, 5 mg of adsorbent (MT-CTF and CTF-1) was added to 10 mL of Hg²⁺ aqueous solutions (ranging from 25 to 800 mg∙L⁻¹). The solution was sonicated and then stirred for 12 h. After the adsorption, the remaining Hg²⁺ in the filtrate was determined by ICP-OES. Subsequently, the adsorption capacity (q_e) was determined based on the following equation (1), where C₀ is the initial concentration, C_e is the equilibrium concentrations, V is the volume of the solution treated (mL), m signifies the amount of adsorbent utilized (mg). Unless otherwise specified, the C₀, C_e, V and m in the following paragraphs is consistent with the above.

$ {q}_{\mathrm{e}}=\frac{({C}_{0}-{C}_{\mathrm{e}})\times V}{m} $

(1)

1.6 Hg²⁺ adsorption kinetics

A flask containing 5 mg∙L⁻¹ Hg²⁺ solution (200 mL), then MT-CTF (20 mg) was added. This mixture was sonicated to disperse, and stirred continuously for 12 h. After the uptake, the residual Hg²⁺ concentration was determined by ICP-MS.

1.7 Effect of pH on adsorption

5 mg of MT-CTF was added into 10 mL Hg²⁺ solution (10 mg∙L⁻¹). The pH of the solution was adjusted with 0.1 mol∙L⁻¹ HNO₃ or NaOH, and then stirred overnight.

1.8 Selectivity tests

20 mg of MT-CTF was added to a solution of Hg(NO₃)₂, Be(NO₃)₂, Mg(NO₃)₂, Al(NO₃)₃, Mn(NO₃)₂, Co(NO₃)₂, Ni(NO₃)₂, Cu(NO₃)₂, Zn(NO₃)₂, Cd(NO₃)₂, Ba(NO₃)₂ (each 5 mg∙L⁻¹, 200 mL). The removal efficiency (η) was calculated using the following equation (2):

$ \eta=\frac{(C_0-C_{\mathrm{e}})\times\mathit{\mathit{V}}}{C_0}\times100\% $

(2)

1.9 Recyclability test

The adsorbed material (Hg@MT-CTF) was stirred in a concentrated HCl solution (12 mol∙L⁻¹) for 3 h. The solid was then collected by centrifugation, and washed with diluted HCl solution repeatedly. After that, the material was washed thoroughly until the filtrate solution reached a neutral pH value and collected by filtration, and then the material was dried under vacuum to afford MT-CTF.

2 Results and discussion 2.1 Structural characterization

Powder X-ray diffraction (PXRD) patterns were recorded to validate the crystalline structure. As shown in Fig. 1(a), as expected, due to the reflection from the (010) and (020) crystal facets of MT-CTF, distinguishable peaks at 7.6° and 13.8° were acquired, respectively. Pawley refinements give rise to optimized parameters (a = b = 1.488 nm, and c = 0.366 nm), matching well with the eclipsed AA-stacking model (R_wp = 8.14% and R_p = 6.03%). The powder X-ray diffraction (PXRD) of CTF-1 displays diffraction peaks at 7.3°, 12.7°, 14.6° and 25.8°, assigned to the (100), (110), (200) and (001) crystal facets, respectively (Fig. 1(c)). The nitrogen adsorption isotherms were collected at −196 ℃, where the BET value of MT-CTF and CTF-1 was calculated to be 483 and 627 m²∙g⁻¹ (Fig. 1(e) and Fig. 1(g)). The CO₂ binding affinity of MT-CTF was also examined by measuring the CO₂ uptake (Fig. 1(f) and Fig. 1(h)), which was measured to be 1.66 mmol∙g⁻¹ at 25 ℃ and larger than that of CTF-1 (0.68 mmol∙g⁻¹). This may be attributed to the presence of abundant methylthio sites within MT-CTF ^[31−32].

To verify the as-obtained CTFs, the chemical constitution of MT-CTF and CTF-1 was further analyzed using FT-IR and CP/MAS ¹³C-NMR spectroscopy. As shown in Fig. 2(c), two strong absorption bands at 1515 and 1354 cm⁻¹ represent the aromatic C―N stretching and “breathing” modes in the triazine unit of MT-CTF, respectively ^[33−34], where an absorbance band at 2920 cm⁻¹ successfully indicates the absorption of the methylthio groups ^[35]. Their chemical structures were further confirmed by solid-state ¹³C NMR spectra (Fig. 2(a) and Fig. 2(b)). The signal of sp² carbons from the triazine ring was revealed by the peak around (δ: ~ 168) ^[28]. The signal at δ=15 was assigned to C―S of methylthio groups ^[36], suggesting the successful installation of methylthio groups within the architecture of MT-CTF. Moreover, the EDS images of MT-CTF further confirm the existence of N/S-containing sites (Fig. 3(a)). Additionally, exceptional thermal stability without significant weight loss up to 300 ℃ was achieved based on thermogravimetric analysis (TGA) under the N₂ atmosphere (Fig. 3(b)).

2.2 Hg²⁺ adsorption performance

To access separation efficiency, MT-CTF was added into aqueous Hg(NO₃)₂ solutions with Hg²⁺ concentrations, ranged from 25 to 800 mg∙L⁻¹. After reaching equilibrium, the uptake capacities were examined by ICP experiments. The equilibrium adsorption data closely aligned with both the Langmuir and Freundlich isotherm models, with excellent correlation coefficients (R² > 0.98) (Fig. 4(a)). In addition, we further studied the adsorption kinetics of the removal of Hg²⁺ (Fig. 4(b)). Within the initial 30 minute, approximately 80% of the Hg²⁺ can be captured. After 12 h, the concentration of residual Hg²⁺ in the aqueous solution dropped to less than 1 μg∙L⁻¹, meeting the U.S. Environmental Protection Agency’s standard (< 2 μg∙L⁻¹). The adsorption kinetic data was fitted with the pseudo-second- order kinetic model from pseudo-second-order kinetic equation (3):

$ \frac{t}{{q}_{\mathrm{t}}}=\frac{1}{{k}_{2}{{q}_{\mathrm{e}}}^{2}}+\frac{t}{{q}_{\mathrm{e}}} $

(3)

where t is the specific time (min), q_t is the Hg²⁺ uptake capacities at time t, q_e is the Hg²⁺ uptake capacities at time equilibrium (mg·g⁻¹), k₂ is the adsorption rate constant (g∙mg⁻¹·min⁻¹).

As shown in Fig. 4(c), the adsorption rate constant k₂ is 0.0203 g∙mg⁻¹·min⁻¹, which indicated rapid removal of Hg²⁺ from the solution. The suitability of pseudo-second-order kinetics has been demonstrated according to previous reports. Therefore, the fast kinetics can be attributed to the intrinsic promising properties of MT-CTF, such as porous structures and the chelating sites with Hg²⁺.

The distribution coefficient (K_d) for Hg²⁺ removal by the present adsorbents was calculated as per the following equation (4) ^[37], i.e. K_d:

$ {K}_{\mathrm{d}}=\frac{{C}_{0}-{C}_{\mathrm{e}}}{{C}_{\mathrm{e}}}\times \frac{V}{m} $

(4)

The K_d value clearly shows the affinity of a adsorbent for metal ions ^[24]. The calculated K_d is estimated to be 5.4×10⁷. The obtained high K_d value suggests that MT-CTF has great potential for Hg²⁺ adsorption. As reported, K_d value can be significantly affected by intrinsic properties of adsorbents, such as accessibility and affinity of the binding sites and flexibility of functional groups ^[38].

The pH value of the Hg²⁺ solution observably impacts the adsorption capacity. We then examined this over a wide pH range (from 1 to 12). As depicted in Fig. 4(d), the highest adsorption efficiency of MT-CTF was achieved to be 99.7% at the pH of 7 ^[39]. The maximum adsorption of Hg²⁺ on MT-CTF was estimated to be 751 mg∙g⁻¹, ranking at the top among previously-reported adsorbents(Fig.5(a)), while the CTF-1 has no effect on Hg. Furthermore, stable separation performance was clearly observed, suggesting exceptional binding affinity of MT-CTF towards Hg²⁺. In addition, the selectivity plays a crucial role in achieving high separation performance. Towards this end, a wide variety of metal ions were examined for the adsorption using MT-CTF as the adsorbent. The initial batches were carried out using the solution containing Hg²⁺, Be²⁺, Mg²⁺, Al³⁺, Mn²⁺, Co²⁺, Ni²⁺, Cu²⁺, Zn²⁺, Cd²⁺ and Ba²⁺. As shown in Fig. 5(b), MT-CTF exhibits an excellent selectivity of Hg²⁺, which originates from the existence of rich task-specific methylthio groups. Given the fact that the reusability of adsorbents is a key parameter governing their separation applications, we further examined the recyclability of MT-CTF. As shown in Fig. 5(c), excellent cycling performance was obtained, where no significant adsorption loss was observed after five consecutive cycles, suggesting intrinsic robust structure of MT-CTF. As shown in the XPS spectra, two signals at 101.0 and 104.9 eV, corresponding to Hg 4f_7/2 and 4f_5/2 respectively provide strong evidence of successful capture of Hg²⁺ within the pores of MT-CTF. The methylthio sites inside MT-CTF were confirmed by the peaks at 161.3 eV (S 2p_3/2) and 162.1 eV (S 2p_1/2), whereas these two peaks were shifted to 162.8 eV (S 2p_3/2) and 163.5 eV (S 2p_1/2) due to the interaction with Hg²⁺ (Hg@MT-CTF, Fig. 5(e) and Fig. 5(f)).

3 Conclusion

In summary, a crystalline CTF with methylthio pendant arms was rationally synthesized for efficient Hg²⁺ adsorption. MT-CTF was prepared through a low-temperature-involved condensation reaction between methylthio-functionalized aromatic aldehydes and amidine moieties. Due to the existence of rich task-specific methylthio groups, MT-CTF exhibits an exceptional Hg²⁺ uptake capacity of 751 mg∙g⁻¹, accompanied with high selectivity over a wide range of metal ions, where exceptional removal efficiency over a wide pH range with excellent recyclability was also achieved. In the future, we will investigate the adsorption performance of MT-CTF in domestic and industrial wastewater, and evaluate the long-term stability and practicability of the MT-CTF. We anticipated that this work may open up new possibility in the synthesis of CTFs for promising separation applications.