The rapid and continuous development of new fields of scientific research and industrial manufacturing such as life [1, 2], environment [3, 4], food [5, 6], and medicine [7, 8] has promoted the development of high-performance liquid chromatography (HPLC) technology. As the key to improve the separation performance of liquid chromatography, the development of novel chromatographic stationary phases has always been one of the frontiers and hotspots of research. Although silica gel [9-11] and polymer matrix [12] have shown good effects in liquid chromatography, the narrow pH range of silica gel matrix and the easy swelling of organic matrix hinder their further application. Therefore, it is necessary to explore the novel stationary phase with great stability and outstanding separation performance.

Covalent organic frameworks (COFs) are ordered two-dimensional (2D COFs) or three-dimensional (3D COFs) crystalline porous polymer materials constructed via dynamic covalent bonds of organic building elements [13, 14]. Due to its large surface area, ordered porosity, and high chemical and mechanical stability, COFs have shown great potential in the field of catalysis [15, 16], gas storage [17, 18], separation [19, 20], adsorption [21, 22] and sensing [23, 24]. These unique properties make it also particularly attractive in the field of chromatographic separation [25, 26].

Compared with 2D COFs, 3D COFs have more ordered interpenetrated channels and pore structures, which are conductive to been used as stationary phases for chromatographic separation. For example, Qian et al. [27] prepared 3D COF/silica composites (COF-300@SiO₂) via layer-by-layer approach and utilized it as the stationary phase for HPLC separation of position isomers. In the same year, Chen et al. [28] used COF-300@SiO₂ as liquid chromatograph stationary phase successfully achived separation of neutral and polar compounds. These results demonstrated the potential of 3D COFs as stationary phases in chromatographic separation. However, most of the reported 3D COFs have irregular morphology and inappropriate size, and direct use of them as stationary phases may lead to high column pressure and low column efficiency, which limits the application of COFs in the field of liquid chromatography separation. In 2018, Wang et al. synthesized single-crystal 3D COFs, which have more long-range ordered crystal structure, regular morphology, and pore structure than traditional polycrystalline COFs [29, 30]. The successful preparation of single-crystal 3D COFs with regular morphology makes it possible to use 3D COFs as stationary phase directly for chromatographic separation. In our previous work, the performance of single-crystal 3D COFs as a stationary phase in HPLC for position isomer separation has been demonstrated [31]. However, there is still a great potential for the application of single-crystal 3D COFs as stationary phases for the broad-spectrum separation of small molecules with similar physical and chemical properties, and clinical drugs, which has not been systematically studied.

Herein, we used the single-crystal 3D COFs directly as a stationary phase of HPLC for the broad-spectrum separation of small molecules. The separation performance of the single-crystal 3D COFs packed column was investigated by using substituted benzenes, polycyclic aromatic hydrocarbons (PAHs), halogenated benzenes, halogenated nitrobenzenes, aromatic amines, and phthalates (PAEs) as analytes, etc. The separation mechanism of organic molecules on the single-crystal 3D COFs packed column was further discussed. Furthermore, the chromatographic performance of the single-crystal 3D COFs packed column was further evaluated by comparing it with a commercial C18 packed column and polycrystalline 3D COFs packed column. Finally, the effects of ACN content, injected mass, and column temperature on the separation of the single-crystal 3D COFs packed column were studied in detail.

The synthesis of the single-crystal 3D COFs was shown in Scheme 1. Materials and instruments were offered in Supporting information. The single-crystal 3D COFs was fabricated following the reported literature with minor modifications [29]. In brief, terephthaldehyde (PDA) (60 mg, 0.45 mmol) was dissolved in 1, 4-dioxane (5 mL). Then, aniline (250 µL) and aqueous acetic acid (6 mol/L) (1 mL) were quickly added into the PDA solution and ultrasonicated for 30 s. Afterward, tetra-(4-anilyl)-methane (TAM) (100 mg, 0.26 mmol) dissolved in 1, 4-dioxane (5 mL) was added into the abovementioned PDA solution and placed at room temperature for 15 days. The yellow products were washed with THF and 1, 4-dioxane for several times and dried at 60 ℃ under a vacuum.

The morphologies of single-crystal 3D COFs were characterized by scanning electron microscopy (SEM) and transmission electron microscopy (TEM). As displayed in Figs. 1a and b, the single-crystal 3D COFs had a regular crystal shape and good monodispersity, with a particle length of about 10 µm and a width of about 2.5 µm. As shown in Fig. S1 (Supporting information), transverse-longitudinal ratio of the single-crystal 3D COFs was mainly concentrated in 2.5, indicating that the single-crystal 3D COFs have good particle-size uniformity. FT-IR was utilized to characterize the synthesized single-crystal 3D COFs (Fig. 1c). Compared with the two monomers (PDA and TAM), the absorption peak of N-H stretching vibration attributed to TAM (3395 cm⁻¹) and C=O stretching vibration attributed to PDA (1683 cm⁻¹) did not appear in the FT-IR spectra of the synthesized single-crystal 3D COFs, but the characteristic absorption peak of the stretching vibration attributed to C=N appears at 1615 cm⁻¹, which indicated the success of the Schiff base reaction. The powder X-ray diffraction (PXRD) pattern of single-crystal 3D COFs was basically the same as that of single-crystal 3D COFs reported in the literature, without the miscellaneous peak (Fig. 1d) [29]. The water contact angle of nearly 95.3° revealed a hydrophobic property of as-prepared single-crystal 3D COFs, which may be due to the rich content of C element and a large quantity of benzene ring structures in the framework (Fig. 1e). This property of single-crystal 3D COFs provided a possibility for the separation of hydrophobic small molecules. Thermogravimetric analysis (TGA) demonstrated that the thermostability of single-crystal 3D COFs was up to 500 ℃ (Fig. 1f). The N₂ adsorption-desorption result manifested that the Brunauer–Emmett–Teller (BET) specific surface area of single-crystal 3D COFs was 819.3 m²/g, and the pore size was mainly distributed at 0.8 nm under the Horvath-Kawazoe (HK) model (Fig. S2 in Supporting information), indicating that single-crystal 3D COFs was a kind of microporous material.

Single-crystal 3D COFs (200 mg) was dispersed in 10 mL ACN under ultrasonication for 10 min. The as-prepared suspension was then poured into a stainless steel column (50 mm × 2.1 mm i.d.) under 5000 psi for 30 min using methanol as a propulsion solvent. After packing, the single-crystal 3D COFs packed column was balanced via MeOH at a flow rate of 0.2 mL/min for 24 h using a high-pressure pump.

ACN, MeOH, and water were used as mobile phases to evaluate the organic solvent tolerance of the single-crystal 3D COFs packed column. As shown in Fig. S3 (Supporting information), the column pressure of the single-crystal 3D COFs packed column increased with the increase of flow rate in different mobile phases, and presented a good linear relationship with the flow rate, suggesting that the single-crystal 3D COFs packed column had good organic solvent tolerance. Besides, according to Eq. S1 (Supporting information), the swelling coefficient of single-crystal 3D COFs packed column in ACN is 0.14, indicating the satisfactory stability of the single-crystal 3D COFs packed column. To further study the pH stability, the single-crystal 3D COFs packed column was flushed with 70% ACN and 30% buffered solutions (pH = 2 or 10) for 32 h using aniline as a test compounds. As shown in Fig. S4 (Supporting information), the retention time of aniline almost did not change after flushing with solutions at different pH values, indicating the good pH stability of the prepared single-crystal 3D COFs packed column.

To evaluate the retention behavior of the single-crystal 3D COFs packed column, a group of substituted benzenes (aniline, anisole, bromobenzene, and propylbenzene) was chosen as targets and ACN/H₂O was used as the mobile phase. As the increase of the ACN content from 70% to 100%, the retention time of the four analytes gradually decreased, revealing a reversed-phase liquid chromatographic (RPLC) separation mechanism for the single-crystal 3D COFs packed column (Fig. S5 in Supporting information). Moreover, the substituted benzenes achieved baseline separation on the single-crystal 3D COFs packed column within 7 min at a mobile phase of ACN/H₂O (85/15, v/v) (Fig. 2a). The maximum column efficiency of substituted benzene can reach 32893 plate/m (Table S1 in Supporting information). The retention and elution order of substituted benzenes was aniline < anisole < bromobenzene < propylbenzene, which was consistent with the hydrophobic order (Table S2 in Supporting information), indicating that the hydrophobic interaction was dominant in the separation of substituted benzene on the single-crystal 3D COFs packed column. In contrast, the commercial C18 column was also used to separate substituted benzenes (Fig. S6a in Supporting information). Under the same chromatographic conditions, the commercial C18 column cannot achieve the separation of substituted benzenes. Until the content of ACN in the mobile phase was reduced to 50%, the substituted benzenes achieved baseline separation on the commercial C18 column (Fig. S7a in Supporting information). The results suggested a stronger interaction occurred between single-crystal 3D COFs packed column and substituted benzenes, in comparison with the commercial C18 column.

Similarly, the separation effects of the single-crystal 3D COFs packed column and the commercial C18 column for PAHs were compared. As can be seen from Fig. 2b, the PAHs were completely separated on the single-crystal 3D COFs packed column within 12 min with maximum column efficiencies reaching 23900 plate/m (Table S1). The abundant aromatic structure in the single-crystal 3D COFs provided a strong π-π interaction between PAHs and the single-crystal 3D COFs. By comparing the peak order of the compounds with their LogK_ow values (Table S3 in Supporting information), it was found that the separation of PAHs by single-crystal 3D COFs packed column was based on hydrophobic interaction and π-π interaction. Comparing the single-crystal 3D COFs packed column, the commercial C18 column had weak retention for PAHs under the same conditions (Fig. S6b in Supporting information). Until the content of ACN in the mobile phase dropped to 50%, the PAHs achieved baseline separation on the commercial C18 column (Fig. S7b in Supporting information) within 37 min.

In addition, the separation of halogenated benzenes (fluorobenzene, chlorobenzene, bromobenzene, and iodobenzene), halogenated nitrobenzenes (2, 3-difluoronitrobenzene, o-chloronitrobenzene, 1-bromo-2-nitrobenzene, and 1-iodo-4-nitrobenzene), aromatic amines (aniline, 1-naphthylamine, 2-chloro-4-iodoaniline, and diphenylamine), and PAEs (DMP, DEP, DPrP, and DBP) on the single-crystal 3D COFs column was shown in Figs. 2c-f. All substances achieved baseline separation on the single-crystal 3D COFs packed column within 10 min with maximum column efficiencies reaching 40382 plate/m (Table S1), and the separation order was consistent with increasing hydrophobicity in the Tables S4-S7 (Supporting information), demonstrating the RPLC mode. In contrast, the commercial C18 column was used to separate the four categories of substances mentioned above (Figs. S6c-f in Supporting information). It was found that the commercial C18 column was difficult to reach effective separation of these four substances under the same conditions. Further optimization of chromatographic conditions, the four substances were still unable to reach baseline separation on the commercial C18 column (Figs. S7c-f in Supporting information). Further analysis of the structures of the four groups of analytes showed that there was little difference between the single-crystal 3D COFs packed column and the four groups of analytes, the π-π conjugation between them, demonstrating the dominant role of the hydrophobic interaction in the separation. This result also showed that single-crystal 3D COFs was very suitable for the separation of hydrophobic small molecules.

The resolution (Rs) value of the studied analytes on the single-crystal 3D COFs packed column was listed in the Table S8 (Supporting information). By comparing the separation effect of analytes on the commercial C18 column under the same conditions (Fig. S6), it can be seen that analytes have strong retention on the single-crystal 3D COFs packed column, which may be due to the inherent hydrophobicity of single-crystal 3D COFs and a large amount of benzene ring structures. The above results indicate that the single-crystal 3D COFs packed column had a good application prospect in the separation of hydrophobic molecules.

To further compare the separation ability of the single-crystal 3D COFs packed column, polycrystalline 3D COFs was synthesized according to the previous report [32]. The characterizations of polycrystalline 3D COFs were displayed in Figs. S8 and S9 (Supporting information). The image of TEM and SEM exhibited an irregular morphology and poor monodispersity of polycrystalline 3D COFs. Whereafter, the separation performance of polycrystalline 3D COFs packed column was investigated for separation of six kinds of hydrophobic small molecules (Fig. S10 in Supporting information). Under the condition of a low organic phase ratio mobile phase, halogenated nitrobenzenes and aromatic amines could not quickly achieve baseline separation on the polycrystalline 3D COFs packed column. Although substituted benzenes, PAHs, halogeno benzenes, and PAEs can be separated, but serious peak tailing and distorted peak for these compounds was observed, which emphasized the important effect of the single crystal structure of 3D COFs in the separation of hydrophobic small molecules.

Inspired by the above results, acenaphthene and acenaphthylene with similar structures and physicochemical properties were chosen to evaluate the chromatographic performance of the single-crystal 3D COFs packed column. As displayed in Fig. 3a, acenaphthene and acenaphthylene were efficiently separated on the single-crystal 3D COFs packed column with column efficiency of 35058 plate/m for acenaphthylene and 38807 plate/m for acenaphthene, respectively. On the basis of the molecular structure and hydrophobic value (Table S9 in Supporting information), both acenaphthene and acenaphthylene have π-π interaction and hydrophobic interaction with single-crystal 3D COFs. The peak order of acenaphthene and acenaphthylene was proportional to the hydrophobic values. Therefore, it could be concluded that the hydrophobic interaction played a dominant role in the separation of acenaphthene and acenaphthylene.

Carbamazepine (CBZ) is an antiepileptic drug used to treat epilepsy and neuropathic pain, and its main metabolite is carbamazepine-10, 11-epoxide (CBZEP) [33]. Since the structure of CBZ and CBZEP is relatively similar, the chromatographic separation in advance is conducive to the accurate detection of the content of the two substances. The structures and the related LogK_ow values of these tested analytes are listed in Table S10 (Supporting information). Through the optimization of chromatographic conditions, the single-crystal 3D COFs packed column achieved baseline separation of CBZ and CBZEP within 5.5 min (Fig. 3b), and the R_s of both was 2.2 (Table S8). The order of peaks is CBZEP < CBZ, this may be due to the fact that CBZEP was an oxidative metabolite of CBZ, so CBZEP was less hydrophobic than CBZ. These results indicate the possibility of using single-crystal 3D COFs packed column in clinical drug monitoring.

To validate the practicability of the single-crystal 3D COFs packed column, polychlorobenzenes (chlorobenzene, p-dichlorobenzene, 1, 2, 4-trichlorobenzene, and 1, 2, 3, 4-tetrachlorobenzene) and polybromobenzenes (bromobenzene, m-dibromobenzene, 1, 3, 5-tribromobenzene, and 1, 2, 4, 5-tetrabromobenzene) in waste-water were chosen for further evaluation. As presented in Figs. 3c and d, both polychlorobenzenes and polybromobenzenes were achieved separation on the single-crystal 3D COFs packed column with ACN/H₂O = 80/20 (v/v) as the eluent, which suggested a great prospect of the single-crystal 3D COFs packed column in the separation of real samples.

In chromatographic separations, the chromatographic parameters such as mobile phase, column temperature, and injected analyte mass are important factors that affects the separation. The effects of these chromatographic parameters were studied by using substituted benzenes and PAEs as analytes. As shown in Fig. S11 (Supporting information), the retention of each component in substituted benzenes and PAEs was significantly enhanced with the decrease of the ACN content. The results demonstrated effective separation of hydrophobic compounds in the single-crystal 3D COFs packed column could be achieved by varying the proportion of organic phase in the mobile phase.

Whereafter, the substituted benzenes and PAEs were separated with different column temperatures from 30 ℃ to 50 ℃ on the single-crystal 3D COFs packed column (Fig. S12 in Supporting information). The retention time of substituted benzenes and PAEs on the single-crystal 3D COFs packed column decreased gradually with the increase of column temperature, demonstrating an exothermic process of the separation on the single-crystal 3D COFs packed column. In addition, Fig. S13 (Supporting information) showed that the van't Hoff plots of substituted benzenes and PAEs had a great linear relationship, indicating that the separation mechanism did not change with the temperature changes. Furthermore, according to the van't Hoff equation, the enthalpy change (ΔH) and entropy change (ΔS) of the solute transfer from the mobile phase to the stationary phase were calculated, and the binding free energy ΔG < 0 suggested that the separation on the single-crystal 3D COFs packed column was thermodynamically spontaneousthe (Table S11 in Supporting information).

Besides, the analyte masses between 150 ng and 350 ng of substituted benzenes and PAEs were used to verify the separation performance the single-crystal 3D COFs packed column (Fig. S14 in Supporting information). The retention time of the substituted benzenes and PAEs remained basically unchanged when the analyte mass increased, and the chromatographic peak areas of substituted benzenes and PAEs increased with the increase of the both analyte mass (Fig. S15 in Supporting information). These results proved the quantitative analysis ability of the single-crystal 3D COFs packed column.

In order to investigate the reproducibility of single-crystal 3D COFs packed columns, substituted benzenes and PAEs were selected as analytes for repeated injection (Fig. 4). After 10 times injections of analytes, the chromatograms did not change significantly. And Table S12 (Supporting information) showed that the RSD values of peak area and peak height were lower than 3.02% and 3.29%, respectively after 100 times injections of analytes. All results displayed remarkable repeatability of the single-crystal 3D COFs packed column for HPLC separation. In addition, the TEM and SEM images (Fig. S16 in Supporting information) of the single-crystal 3D COFs after filling into the column indicated that there was no significant change in morphology.

In conclusion, the single-crystal 3D COFs was synthesized as stationary phases for HPLC separation. Thanks to the excellent performances of single-crystal 3D COFs, the single-crystal 3D COFs packed column not only separated the hydrophobic small molecules, but also realized the efficient separation of clinical drugs based on strong hydrophobic and π-π interactions. Compared with a commercial C18 packed column and polycrystalline 3D COFs packed column, the prepared single-crystal 3D COFs packed column showed great merits in the baseline separation of a series of small molecules with high column efficiency and excellent reproducibility. Moreover, the single-crystal 3D COFs packed column had also been successfully applied to the separation of the pollution of polychlorobenzenes and polybromobenzenes. This study offers promising guidance for utilizing single-crystal 3D COFs materials as stationary phases and reveals great application potential in the separation of small molecule compounds and drugs.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

CRediT authorship contribution statement

Qiuting Zhang: Methodology, Data curation, Conceptualization. Fan Wu: Methodology, Data curation, Conceptualization. Jin Liu: Validation, Data curation. Hang Su: Formal analysis. Yanhui Zhong: Writing – review & editing, Formal analysis. Zian Lin: Writing – review & editing, Supervision, Funding acquisition.

Acknowledgments

We gratefully thank the National Natural Science Foundation of China (No. 22274021) and Natural Science Foundation of Fujian Province (No. 2022J01535) for financial support.

Supplementary materials

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2024.110649.