Single cobalt atoms anchored on Ti₃C₂T_x with dual reaction sites for efficient adsorption–degradation of antibiotic resistance genes

Significance

The global spread of antibiotic resistance genes (ARGs) poses an increasingly serious threat to public health. Advanced oxidation processes have shown promising prospects for controlling ARG pollution. However, the low concentration of ARGs in the natural water and the short half-life of reactive oxygen species lead to the unsatisfactory removal efficiency of ARGs. This research presents a strategy for constructing a Co–O–Ti dual-reaction-site catalyst. The Ti adsorption site binds with PO₄³⁻ on the phosphate skeletons of ARGs via Ti–O–P coordination interactions, while the Co–O₃ active site activates peroxymonosulfate into surface-bond hydroxyl radicals, achieving the efficient in situ degradation of ARGs. This work is expected to provide a different perspective for the design of catalysts used in ARG removal.

Abstract

The assimilation of antibiotic resistance genes (ARGs) by pathogenic bacteria poses a severe threat to public health. Here, we reported a dual-reaction-site–modified Co_SA/Ti₃C₂T_x (single cobalt atoms immobilized on Ti₃C₂T_x MXene) for effectively deactivating extracellular ARGs via peroxymonosulfate (PMS) activation. The enhanced removal of ARGs was attributed to the synergistic effect of adsorption (Ti sites) and degradation (Co-O₃ sites). The Ti sites on Co_SA/Ti₃C₂T_x nanosheets bound with PO₄³⁻ on the phosphate skeletons of ARGs via Ti–O–P coordination interactions, achieving excellent adsorption capacity (10.21 × 10¹⁰ copies mg⁻¹) for tetA, and the Co–O₃ sites activated PMS into surface-bond hydroxyl radicals (•OH_surface), which can quickly attack the backbones and bases of the adsorbed ARGs, resulting in the efficient in situ degradation of ARGs into inactive small molecular organics and NO₃. This dual-reaction-site Fenton-like system exhibited ultrahigh extracellular ARG degradation rate (k > 0.9 min⁻¹) and showed the potential for practical wastewater treatment in a membrane filtration process, which provided insights for extracellular ARG removal via catalysts design.

Antibiotic resistance genes (ARGs) have been recognized as one of emerging contaminants due to their pivotal role in the formation of pathogenic superbacteria (1, 2). Different from conventional pollutants, ARGs are DNA fragments encoding antibiotic-resistant proteins, which can migrate through diverse environmental media and amplify in host microbes (3–5). Some wastewater treatment plants have become the breeding grounds and point source of ARGs. Many attempts have been made to remove ARGs from wastewater, but few of them can effectively inactivate or mineralize ARGs. For instance, adsorption technologies, such as activated sludge technology (6) and constructed wetlands method (7) can only enrich ARGs without further degradation, which may cause secondary pollution. Traditional wastewater treatment technology, such as chlorination and ultraviolet disinfection, exhibits reduced efficiency as ARG levels decrease (8–10). Consequently, it is urgent to seek new methods to effectively remove ARGs.

Advanced oxidation processes (AOPs) have shown promising prospect for controlling ARG contamination via in situ generation of various reactive oxygen species (ROS) (9, 11). Particularly, the peroxymonosulfate (PMS)-based AOPs have attracted wide attention because of their high efficiency, easy transportation, and storage. Recently, single-atom catalysts with unique electronic property and high atom utilization have been applied for refractory organic contaminants degradation via PMS activation (12–14), which break the limitations of heterogeneous catalysts in terms of kinetics and catalytic activity (15–17). However, the low concentration of ARGs in the wastewater (ranging from ng/L to μg/L) (18) and the short half-life of ROS (less than 1 μs for •OH and 30 to 40 μs for SO₄^•-) (19, 20) lead to the unsatisfactory ROS utilization ratio and ARG removal efficiency. In this regard, minimizing the migration distance from ROS to ARGs is a desirable approach to maximize the destruction of ARGs. To achieve this goal, a catalyst with abundant selective adsorption and degradation sites for ARGs is highly attractive.

ARGs are double-helix DNA composed of an outer sugar-phosphate backbone and internal bases. All of the phosphate groups, deoxyribose, and base in ARGs could be potential binding sites for catalysts, and the binding forces may include covalent coordination and noncovalent adsorption such as hydrogen bonding, π–π stacking, hydrophobic interaction, and electrostatic attraction.¹ Because of the location of base pairs, it is difficult to form hydrogen bonding and π–π stacking. Meanwhile, these noncovalent interactions could be easily disturbed by pH and nonspecific competition from other organic compounds in the wastewater (21, 22). Therefore, the development of catalysts that can interact with the phosphate groups outside ARGs is expected to achieve efficient adsorption of double-stranded ARGs, but there are few related studies. According to the previous studies, transition metal ions such as iron, titanium, and copper tend to coordinate with phosphate in solution (1, 23, 24). The catalysts containing these metal ions on the surface are likely to form strong interactions with the phosphate groups of ARGs (1, 21).

On the other hand, as a new class of two-dimension (2D) transition metal carbides/nitrides, Mxenes are ideal supports for single-metal active sites in PMS activation due to their huge specific surface area, excellent electronic conductivity, high reducing ability, hydrophilic surface function, and strong mechanical stability (25–27). More importantly, compared with other metal-free 2D materials such as C₃N₄ and graphene oxide, Mxenes possess active basal planes with exposed metal sites. For example, the surface of Ti₃C₂T_x (T_x stands for surface functional group) MXene is rich in Ti ions, which are expected to form strong chelation with phosphate backbone to achieve efficient adsorption of ARGs. Considering these two factors, anchoring single transition metal atoms on Ti₃C₂T_x nanosheets to construct adsorption and degradation dual sites is expected to effectively address the slow degradation kinetics of ARGs. It has been reported that among different types of PMS activation centers, Co²⁺ shows the highest activity that is even superior to traditional Fenton reaction with a wide pH adaptation and low PMS dosage (12, 28). Therefore, cobalt has been identified as one of the most efficient materials for PMS activation.

Here, we reported single Co atoms (Co SAs) immobilized on Ti₃C₂T_x nanosheets (Co_SA/Ti₃C₂T_x) with dual reaction site as Fenton-like catalysts for ARG removal via PMS activation. Co_SA/Ti₃C₂T_x exhibited excellent adsorption capacity (10.21 × 10¹⁰ copies/mg), ultrahigh degradation rate (k = 0.94 min⁻¹) for tetA, as well as practical application prospect. Experiments and theoretical calculations indicated that these preeminent performances of Co_SA/Ti₃C₂T_x in ARG removal was attributed to the synergism of adsorption and degradation on dual reaction site. PO₄³⁻ of ARG phosphate skeletons was bounded with Ti adsorption sites on Co_SA/Ti₃C₂T_x nanosheets via Ti–O–P coordination interactions, followed by the efficient degradation via surface-bond •OH generated by activating PMS at Co sites. Consequently, Co_SA/Ti₃C₂T_x efficiently captures and degrades ARGs beyond repair.

Results and Discussion

Preparation and Characterization of Co_SA/Ti₃C₂T_x.

The ultrathin Ti₃C₂T_x nanosheets with a thickness of about 1.3 nm were synthesized from parent Ti₃AlC₂ through etching in a mixed solution of HCl and LiF (SI Appendix, Figs. S1 and S2), and Co_SA/Ti₃C₂T_x nanosheets were obtained via a one-step preparation approach (see Materials and Methods for more details). The X-ray diffraction (XRD) pattern of Co_SA/Ti₃C₂T_x shows that the crystal structure is similar to that of Ti₃C₂T_x (29), and no characteristic peak of Co is visible, suggesting the absence of Co nanoparticles in Co_SA/Ti₃C₂T_x (SI Appendix, Fig. S3). Raman spectra (SI Appendix, Fig. S4) show that the peak at 413 cm⁻¹ (-O terminal groups) shifts toward high frequency after the introduction of Co species, which could be attributed to the interaction between Co and terminal -O of Ti₃C₂T_x (30). Transmission electron microscopy (TEM) and atomic force microscopy (AFM) images distinctly exhibit the nanosheet shape of the prepared Co_SA/Ti₃C₂T_x, and no Co nanoparticle can be observed (Fig. 1 A and B and SI Appendix, Fig. S5), which are consistent with the XRD results. High-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) images (Fig. 1C and SI Appendix, Fig. S6) clearly exhibit that Co atoms (bright spots) are atomically dispersed on Ti₃C₂T_x nanosheets. Energy-dispersive X-ray spectroscopy (EDS) further confirms the uniform distribution of Ti, C, O, and Co on the surface of Co_SA/Ti₃C₂T_x (Fig. 1D).

Fig. 1.

(A) TEM image. (B) AFM image. (C) HAADF-STEM image (Co SAs are highlighted by red circles). (D) EDS elemental mappings. (E) XANES spectra at the Co K-edge. (F) The Co K-edge FT-EXAFS spectra. (G) EXAFS fitting curve of Co_SA/Ti₃C₂T_x, Inset is an illustration of Co_SA/Ti₃C₂T_x structure. The green, blue, gray, and pink balls represent Co, Ti, C, and O, respectively.

The chemical structure of Co atoms immobilized on Ti₃C₂T_x was further demonstrated via X-ray photoelectron spectrometry (XPS), X-ray absorption near-edge structure (XANES), and extended X-ray absorption fine structure spectroscopy (EXAFS) measurements. In the Co 2p XPS spectrum of Co_SA/Ti₃C₂T_x (SI Appendix, Fig. S7), the peaks at 780.3 eV and 796.6 eV correspond to Co 2p_3/2 and Co 2p_1/2, respectively (31). Meanwhile, satellite peaks can be observed at 786.4 and 803.0 eV, indicating the existence of Co²⁺ (32, 33). After the introduction of Co atoms, no shift of C–C peak was noticed, suggesting that no obvious interaction existed between Co and C atoms (25). XANES were conducted to study the electronic state of Co atoms (Fig. 1E), and the results indicate that the Co valence state in Co_SA/Ti₃C₂T_x is between Co foil and Co₃O₄, which is consistent with the XPS result. This demonstrates that Co species in Co_SA/Ti₃C₂T_x are positively charged. The Co K-edge Fourier transform-EXAFS (FT-EXAFS) spectrum of Co_SA/Ti₃C₂T_x (Fig. 1F) exhibits a major peak at about 1.5 Å, and no Co–Co bond near 2.1 Å (existed in Co foil EXAFS spectra) can be detected (31). Besides, the wavelet transform contour plot of Co_SA/Ti₃C₂T_x exhibits one intensity maximum at 4.1 Å⁻¹, which is attributed to the Co–O coordination in the first shell (SI Appendix, Fig. S8). However, no intensity maximum can be observed at 6.7 Å⁻¹ (belonging to the Co–Co coordination), which further proves the atomic distribution of Co species on the surface of Ti₃C₂T_x. The EXAFS curve fitting results (Fig. 1G and SI Appendix, Fig. S9) show that the major peak near 1.5 Å is related to the Co–O bonding configuration (phase uncorrected), which is in accord with the optimized model (Fig. 1 G, Inset and SI Appendix, Fig. S10). The corresponding fitting parameters are shown in SI Appendix, Table S3 and the Co–O coordination number is 3.3 with the bond length of 2.04 Å, suggesting that the Co atom is coordinated with three oxygen atoms in Co_SA/Ti₃C₂T_x. Furthermore, Mulliken charge analysis of the theoretical optimization model proves that the charge accumulation on Co single atoms is positively charged (+0.64, SI Appendix, Fig. S11), consistent with the EXAFS result. All the above results clearly demonstrate that single Co atoms are successfully anchored on ultrathin Ti₃C₂T_x with a Co–O₃ coordination.

ARGs Removal Performance over Co_SA/Ti₃C₂T_x in a Fenton-Like System.

The extracellular tetA were used as the model to evaluate the adsorption and degradation performance of Co_SA/Ti₃C₂T_x (SI Appendix, Fig. S12). As shown in Fig. 2A, both Ti₃C₂T_x and Co_SA/Ti₃C₂T_x exhibit excellent adsorption effect on tetA, and the lg(C/C₀) values can reach −3.2 and −3.4, respectively, demonstrating that the introduction of Co atoms makes a negligible effect on the adsorption of ARGs. After adding PMS, the lg(C/C₀) value of tetA treated with Co_SA/Ti₃C₂T_x reach −7.2 within 20 min, which is much higher than that of Ti₃C₂T_x and pure PMS (Fig. 2B). The removal of tetA accords with pseudo-first-order kinetics, and the k value of Co_SA/Ti₃C₂T_x reaches 0.94 min⁻¹ (Fig. 2 B, Inset). Besides, Co_SA/Ti₃C₂T_x shows excellent removal efficiency for ampC (SI Appendix, Fig. S13). Notably, compared with most of the advanced catalysts for Fenton-like and photocatalytic reactions reported so far, the Co_SA/Ti₃C₂T_x/PMS system has the highest ARG removal efficiency, shortest treatment time, and the lower cost of catalysts synthesis (SI Appendix, Table S4). Other Ti₃C₂T_x-supported atomic catalysts (named as M_SA/Ti₃C₂T_x, M = Fe, Cu, Mn, Ni) also exhibit satisfactory adsorption capacity (lg(C₀/C) > 3) and degradation rate (k > 0.7 min⁻¹) for tetA (SI Appendix, Fig. S14), indicating that it is universal to construct dual-reaction-site Fenton-like system by anchoring single metal atoms on Ti₃C₂T_x to achieve efficient removal of ARGs.

Fig. 2.

(A) Adsorption kinetics of tetA by Ti₃C₂T_x and Co_SA/Ti₃C₂T_x. (B) Time course of tetA degradation, Inset displays the corresponding first-order rate constants of tetA degradation. (C) Agarose gel electrophoresis of tetA before and after treatment by Co_SA/Ti₃C₂T_x and Ti₃C₂T_x, respectively. (D) Concentration of NO₃⁻ in the process of tetA (1 mg/L) removal by Ti₃C₂T_x and Co_SA/Ti₃C₂T_x. (E) Reuse ability of Co_SA/Ti₃C₂T_x.

Moreover, agarose gel electrophoresis and nanopore sequencing tests were used to determine the chain length distribution of tetA after treatment. As displayed in Fig. 2C, the band of tetA vanishes after 20 min of treatment with Co_SA/Ti₃C₂T_x and PMS, while the band still exists after a similar treatment with Ti₃C₂T_x and PMS. The disappearance of bands treated by Co_SA/Ti₃C₂T_x suggests that tetA are not capable of amplification (10), which is consistent with the results of nanopore sequencing (SI Appendix, Fig. S15). The result of high-performance liquid chromatography further suggests that Co_SA/Ti₃C₂T_x can deeply oxidize the bases and deoxyribose (SI Appendix, Table S5). Besides, the concentration of NO₃⁻ released from tetA (1 mg/L) reach 0.85 mg/L after Co_SA/Ti₃C₂T_x catalytic treatment for 60 min, suggesting that 91.4% of the nitrogen in tetA is mineralized (Fig. 2D). All the above results show that Co_SA/Ti₃C₂T_x can effectively adsorb and inactivate tetA and even can deeply oxidize it into small molecules and inorganic ions without potential harm.

The adsorption efficiency and degradation rate of Co_SA/Ti₃C₂T_x for tetA decreased slightly in three real water samples (SI Appendix, Fig. S16), which is mainly due to the fact that the PO₄³⁻ in the real water samples could compete with ARGs for adsorption sites and inhibit the interaction between PMS and the catalysts (SI Appendix, Fig. S17 and Table S6). The bacteria in secondary effluent were completely inactivated in the Co_SA/Ti₃C₂T_x/PMS system within 30 min (SI Appendix, Fig. S19). Furthermore, Co_SA/Ti₃C₂T_x exhibits excellent reusability (Fig. 2E) and low Co ion leaching (SI Appendix, Fig. S20), which is far lower than the allowable limit (1 mg/L) stipulated in the Chinese National Standard (GB 25467-2010). After five cycles, the morphology and microstructure of Co_SA/Ti₃C₂T_x have no obvious change, demonstrating the satisfactory stability of Co_SA/Ti₃C₂T_x (SI Appendix, Figs. S21 and S22). The above results indicate that Co_SA/Ti₃C₂T_x has the potential for practical application.

Adsorption Mechanism of ARGs on Co_SA/Ti₃C₂T_x.

According to the tetA adsorption kinetics of Ti₃C₂T_x and Co_SA/Ti₃C₂T_x (Fig. 2A), both Ti₃C₂T_x and Co_SA/Ti₃C₂T_x complete the equilibrium adsorption within 10 min. To quantitatively investigate the adsorption process, pseudo-first-order and pseudo-second-order models and external mass transfer models (EMTM) were used to fit the kinetic adsorption data (Fig. 3A and SI Appendix, Fig. S23), and the parameters are listed in SI Appendix, Table S7. Compared with pseudo-first-order model, the pseudo-second-order model fits better with the experimental data, indicating that the adsorption processes of both Ti₃C₂T_x and Co_SA/Ti₃C₂T_x are dominated by chemiadsorption (34). The data fitted well with the EMTM, suggesting that the adsorption kinetics for Ti₃C₂T_x and Co_SA/Ti₃C₂T_x are external mass transfer limited. Adsorption isotherms of tetA on Ti₃C₂T_x and Co_SA/Ti₃C₂T_x were fitted by Langmuir and Freundlich models, respectively (Fig. 3B and SI Appendix, Fig. S24). The better fitting of Langmuir model proves the monolayer adsorption of tetA on the surface of Ti₃C₂T_x and Co_SA/Ti₃C₂T_x (1, 35) (SI Appendix, Table S8). The similar calculated maximum adsorption capacities (q_m) of Ti₃C₂T_x (10.00 × 10¹⁰ copies/mg) and Co_SA/Ti₃C₂T_x (10.21 × 10¹⁰ copies/mg) for tetA indicate that the introduction of Co atoms cannot significantly affect the adsorption capacity of Ti₃C₂T_x for ARGs. Furthermore, Fourier transform infrared spectroscopy (FTIR) and XPS spectra of catalysts before and after the adsorption of tetA were recorded to further investigate the interaction between the Ti₃C₂T_x and ARGs. Compared with fresh Ti₃C₂T_x and Co_SA/Ti₃C₂T_x, the samples adsorbed tetA showed an obvious absorption band at 1,056 cm⁻¹, which can be attributed to the P-O stretching vibration of phosphate skeleton (36) (Fig. 3C). This indicates that the phosphate skeleton of ARG binds to the Ti₃C₂T_x and Co_SA/Ti₃C₂T_x, which is also confirmed by XPS result (SI Appendix, Fig. S25).

Fig. 3.

(A) External mass transfer model and (B) isotherms and Langmuir model fits of tetA adsorption on Ti₃C₂T_x and Co_SA/Ti₃C₂T_x. (C) Comparison of FTIR spectra of Ti₃C₂T_x and Co_SA/Ti₃C₂T_x before and after adsorption of tetA. (D) Desorption efficiency of tetA from Ti₃C₂T_x and Co_SA/Ti₃C₂T_x by different desorption solutions. (E) PO₄³⁻ and water pretreatment on the adsorption of tetA by Ti₃C₂T_x and Co_SA/Ti₃C₂T_x (pH 7.2). (F) Ti K-edge XANES spectra and (G) FT-EX AFS spectra of Co_SA/Ti₃C₂T_x before and after adsorption of tetA. (H) Schematic illustration of adsorption positions of tetA on Co_SA/Ti₃C₂T_x.

To further investigate the main interaction forces between ARG and catalysts, isopropanol solution, urea solution, and phosphate buffer solution (PBS, pH = 7) were separately used to desorb the catalyst-adsorbed tetA, which can destroy hydrophobic interaction, restrain hydrogen bond interaction, and compete for the chemical binding between nanomaterials and the phosphate groups of tetA, respectively (1, 37). As shown in Fig. 3D, more than 70% tetA was desorbed from Ti₃C₂T_x and Co_SA/Ti₃C₂T_x by PBS, suggesting that the adsorption of tetA on Ti₃C₂T_x and Co_SA/Ti₃C₂T_x is mainly controlled by the covalent binding force. PO₄³⁻ can invalidate the possible adsorption sites by binding to titanium sites on Ti₃C₂T_x. Therefore, the catalysts pretreated with PO₄³⁻ were used to adsorb tetA to further verify the interaction between Ti₃C₂T_x and the phosphate groups of ARG. As shown in Fig. 3E, after the pretreatment with PO₄³⁻, the lg(C/C₀) values of tetA adsorbed by Ti₃C₂T_x and Co_SA/Ti₃C₂T_x significantly decreased from −3.2 and −3.4 to −0.51 and −0.4, respectively. Besides, PO₄³⁻ pretreatment indistinctively affects the specific surface area of Co_SA/Ti₃C₂T_x, thus the influence of specific surface area on the adsorption effect can be excluded (SI Appendix, Fig. S26 and Table S9). The above results forcefully demonstrate that the phosphate skeletons are involved in the adsorption of ARGs on Ti₃C₂T_x and Co_SA/Ti₃C₂T_x. Since the phosphate skeletons are located outside the DNA and phosphate tends to coordinate with Ti atom (21), it can be conjectured that the PO₄³⁻ of phosphate skeletons and the Ti atoms form Ti–O–P covalent bonds on the surface of the Co_SA/Ti₃C₂T_x. As shown in XANES spectra (Fig. 3F), the adsorption threshold position of Co_SA/Ti₃C₂T_x shifts to low binding energy after the adsorption of tetA, suggesting the electron transfer between Ti sites and ARGs. Meanwhile, the peak intensity of Co_SA/Ti₃C₂T_x increases significantly after the adsorption of tetA, indicating the formation of other coordination bonds between Ti site and ARGs (38). Besides, the intensity of Ti–O/C scattering (39) and the Ti–O/C coordination number increased after the adsorption of tetA (Fig. 3G and SI Appendix, Fig. S27 and Table S10), which further confirms the formation of Ti–O–P coordination bond between Ti atom and phosphate backbone (Fig. 3H). Therefore, the Co_SA/Ti₃C₂T_x pretreated with PO₄³⁻ was used to explore the effect of adsorption on the degradation activity of Co_SA/Ti₃C₂T_x (SI Appendix, Fig. S28). Compared with unpretreated Co_SA/Ti₃C₂T_x (k = 0.94 min⁻¹), the removal rate of PO₄³⁻ pretreated Co_SA/Ti₃C₂T_x (k = 0.41 min⁻¹) is significantly lower, indicating that the excellent removal performance of Co_SA/Ti₃C₂T_x is related to its outstanding adsorption capacity.

Mechanism of Co_SA/Ti₃C₂T_x Activating PMS.

The catalytic activity and interfacial reaction progress of Co_SA/Ti₃C₂T_x in the PMS activation were studied via in situ Raman spectroscopy and in situ (ATR-FTIR). As exhibited in Fig. 4A, the obvious peaks at 980 cm⁻¹ and 1,059 cm⁻¹ are associated with the symmetrical stretching vibration of S=O in SO₄²⁻ and the characteristic vibration of SO₃ in HSO₅⁻, respectively (40, 41). Hence, the decomposition rate of PMS can be evaluated by the ratio of the peak intensity at 980 cm⁻¹ to that at 1,059 cm⁻¹ (I₉₈₀/I₁₀₅₉), and the larger I₉₈₀/I₁₀₅₉ represents the higher PMS decomposition rate (40). Obviously, the value of I₉₈₀/I₁₀₅₉ increases from 0.79 to 1.12 after reacting with Co_SA/Ti₃C₂T_x, while it changes little in the presence of Ti₃C₂T_x, demonstrating the faster PMS decomposition by Co_SA/Ti₃C₂T_x. Meanwhile, the peak at 881 cm⁻¹ (the stretching vibration of O-O in HSO₅⁻) shifts to 887 cm⁻¹ after reacting with Co_SA/Ti₃C₂T_x, while no distinct shift can be observed in the Ti₃C₂T_x/PMS system, suggesting that the stretching vibration amplitude of O–O in PMS is changed by Co_SA/Ti₃C₂T_x (42). A similar shift can be observed in the ATR-FTIR spectra (SI Appendix, Fig. S29). The result of amperometric i–t curves suggests that Co site on Co_SA/Ti₃C₂T_x may be the potential site to activate PMS via transmitting e⁻ to PMS (SI Appendix, Fig. S30) (43). The introduction of Co atoms also optimizes the electrical conductivity of Ti₃C₂T_x, which was proved via the electrochemical impedance spectroscopy and CV curves (SI Appendix, Fig. S31). Subsequently, electron paramagnetic resonance (EPR) measurements were conducted to investigate the ROS in the Co_SA/Ti₃C₂T_x/PMS system. Obviously, the characteristic signals of •OH, SO₄^•−, O₂^•−, and ¹O₂ were generated in the Co_SA/Ti₃C₂T_x/PMS system, while they cannot be detected in the Ti₃C₂T_x/PMS system (Fig. 4B and SI Appendix, Fig. S32), indicating that Co atoms act as active sites for PMS activation. According to previous studies, we consider that the O₂^•− and ¹O₂ may be derived from the self-decomposition of PMS and the oxidation of PMS or O₂^•− by Co²⁺ (44, 45). The XPS spectra of Co_SA/Ti₃C₂T_x before and after reacting with PMS further demonstrate the major role of Co site in PMS activation (SI Appendix, Fig. S33). Combined with the above analysis, we speculate that PMS is mainly activated into ROS at the Co site.

Fig. 4.

(A) In situ Raman spectra of different oxidation systems. (B) EPR spectra using DMPO as trapping agents. (C) The inhibition efficiency of different quenchers on degradation of tetA by Co_SA/Ti₃C₂T_x. (D) Concentration of free •OH generated in the process of PMS activation by Co_SA/Ti₃C₂T_x. (E) Charge density of Co_SA/Ti₃C₂T_x. (F) Optimized configurations of PMS adsorbed on Co_SA/Ti₃C₂T_x. (G) Difference charge density for PMS adsorption on Co_SA/Ti₃C₂T_x and the corresponding charge transfer. Yellow and green contours represent electron accumulation and deletion, respectively. (H) Reaction pathways of PMS activation to form surface-bound •OH (H, Left) and free •OH (H, Right) at Co sites.

Chemical quenching experiments were conducted to further determine the contribution of these radicals to ARG degradation in the Co_SA/Ti₃C₂T_x/PMS system. As shown in Fig. 4C and SI Appendix, Fig. S34, the ¹O₂ scavenger furfuryl alcohol (46) and O₂^•− scavenger para-benzoquinone (p-BQ) (47) all exhibited inappreciable inhibition on tetA removal. In addition, the inhibition efficiency of MEOH (•OH and SO₄^•− scavenger) (48) and tert-butanol (TBA, •OH scavenger) (49, 50) can reach about 30.9% and 20.7%, indicating that •OH may be the dominant species for tetA removal. However, both MEOH and TBA cannot completely inhibit the removal of tetA. According to previous studies, MEOH and TBA exhibit strong polarity, which makes them tend to quench free radicals in solution rather than reacting with the active species on the surface of catalysts (51). Therefore, n-butyl alcohol (NBA) and dimethyl sulfoxide (DMSO) were employed to explore the contribution of •OH on the surface of Co_SA/Ti₃C₂T_x, as NBA can scavenge both free •OH (•OH_free) and surface-bond •OH (•OH_surface) (50), while DMSO can quench all the surface-bound active species (52). Obviously, the inhibition efficiency of NBA (80.4%) is slightly higher than that of DMSO (74.9%), but much higher than that of TBA (20.7%), suggesting that •OH_surface plays a major role in ARG degradation. Moreover, the ratio of •OH_free and •OH_surface generated in the Co_SA/Ti₃C₂T_x system were quantified by adding F⁻ in the system, since F⁻ can desorb •OH_surface via forming strong •OH_surface∙∙∙F⁻ hydrogen-bond (53). As displayed in Fig. 4D, the concentration of •OH_free significantly increased after the addition of F⁻. Assuming that all the •OH_surface can be desorbed from the Co_SA/Ti₃C₂T_x surface, the relative proportion of •OH_surface to all •OH can reach 63%. Combined with the above analysis, the surface-bond •OH is the dominant species for tetA removal in this system.

To further confirm the site for activating PMS on Co_SA/Ti₃C₂T_x and understand the formation process of •OH_surface, DFT calculations were performed. Compared with Ti₃C₂T_x, the highest electron density of the Co atom makes it a potential site to activate PMS (Fig. 4E and SI Appendix, Fig. S35). The models of PMS, Ti₃C₂T_x, and Co_SA/Ti₃C₂T_x were built to analyze the adsorption structures of PMS on the catalysts (Fig. 4F and SI Appendix, Figs. S36 and S37). PMS cannot be adsorbed on the surface of Ti₃C₂T_x (SI Appendix, Fig. S37), indicating its poor catalytic activity. The PMS adsorption energy (E_ads) on Co_SA/Ti₃C₂T_x in type I was calculated to be −4.48 eV (Fig. 4F), much more negative than that in type II (−4.00 eV, SI Appendix, Fig. S38), suggesting that type I is the optimal adsorption model for the adsorption of PMS on Co_SA/Ti₃C₂T_x. Besides, Mulliken charge analysis was used to explore the electron migration between the reactive site and PMS. As shown in Fig. 4G, 0.26 e transfer from Co site to PMS, proving the activation of PMS on Co sites, which is concomitant with the results of amperometric i–t curve measurements. The energy required to generate •OH_surface is much lower than that of •OH_free (Fig. 4H), suggesting the preferential generation of •OH_surface on Co_SA/Ti₃C₂T_x. Consequently, benefiting from the synergistic effect of Ti adsorption sites and Co active sites, the ARGs were efficiently enriched on the surface of Co_SA/Ti₃C₂T_x, and further attacked by the surface-bond •OH, thus achieving efficient in situ degradation of ARGs.

Investigation of the Wastewater Treatment System.

To facilitate the recovery of the catalyst and avoid secondary pollution caused by catalyst loss, Co_SA/Ti₃C₂T_x membrane (namely Co/PVDF) was prepared by vacuum filtration using polyvinylidene difluoride (PVDF) membrane as a carrier. To evaluate the potential application value of Co_SA/Ti₃C₂T_x, a continuous flow reactor consisting of Co/PVDF membrane was used to treat actual wastewater, which is shown in Fig. 5A. The Co/PVDF membrane is flexible and elastic, and the catalyst is strongly attached to the surface of the PVDF membrane to ensure that the catalyst will not fall off during the reaction (Fig. 5 B–D). The secondary effluent from Nankai University wastewater treatment plant (Tianjin, China) was selected to test the ARG removal performance of the Co/PVDF membrane (Fig. 5E), and the initial tetA abundance was about 1.6 × 10⁶ copies/mL. The Co/PVDF membrane reactor can continuously remove ARGs from the wastewater without cleaning the catalyst after a single reaction. As exhibited in Fig. 5F, the Co/PVDF membrane can maintain excellent tetA removal efficiency [lg(C/C₀) = −4.48] after 12 h of continuous operation, reflecting the excellent catalytic activity and durability of Co/PVDF membrane. Moreover, during the long-term operation, the flux had no changes and the Co ion in the effluent was under the allowable limit (Fig. 5G), which indicated their potential for practical application.

Fig. 5.

Applications in wastewater treatment of the Co_SA/Ti₃C₂T_x system. (A) Schematic diagram of the continuous-flow reactor. (B) Image of the catalyst-loaded membrane. (C) SEM image of membrane surface. (D and E) Photograph of the experiment device; (F) tetA removal efficiency using the membrane without a supported catalyst and the membrane loaded with Co_SA/Ti₃C₂T_x. (G) Water flow rate and Co ion concentration in the filtrate with the treatment time. Conditions: [Catalyst] = 100 mg, [PMS] = 200 mg/L.

Discussion

In this work, we prepared a Co_SA/Ti₃C₂T_x catalyst with a dual reaction site for extracellular ARG removal via PMS activation. EXAFS measurements demonstrated that the Ti adsorption site on Co_SA/Ti₃C₂T_x nanosheets can bound with PO₄³⁻ of ARG phosphate skeleton via Ti–O–P coordination interaction, achieving excellent adsorption capacity (10.21 × 10¹⁰ copies/mg). In situ Raman and DFT calculations proved that the Co–O₃ site with a single cobalt atom works as the reactive site for PMS activation. EPR and free-radical scavenging experiments confirmed that the •OH_surface formed by PMS activation is the key active species for the rapid degradation of tetA adsorbed on the Ti sites. The dual reaction sites greatly improve the utilization of •OH_surface, thus achieving an ultrahigh tetA removal rate (k = 0.94 min⁻¹). Besides, the antiinterference performance of Co_SA/Ti₃C₂T_x nanosheets was proved via using different water substrates. Our study details the first insights into dual-reaction-site catalysts for the ARG removal and provides light for the control of ARG contamination.

Materials and Methods

Preparation of Catalyst.

Co_SA/Ti₃C₂T_x catalysts were prepared in two steps. First, 2 g of Ti₃AlC₂ was mixed with 2 g of LiF and 40 mL of HCl (9 M) and stirred at 35 °C for 24 h. The solid residue was washed with ultrapure water and freeze-dried to obtain Ti₃C₂T_x nanosheets (25). The second step was to synthesize Co_SA/Ti₃C₂T_x by a deposition-precipitation method. Typically, 0.1 g of Ti₃C₂T_x was dispersed in 20 mL of ethylene glycol (EG) solution (V_EG: V_water = 1:9) and ultrasonicated for 1 h. Subsequently, moderate amount of CoCl₂ solution (1.5 mM CoCl₂ and 30 mL EG) was added dropwise to the uniformly dispersed Ti₃C₂T_x suspension and stirred for 30 min. Then, the mixture was sonicated for 1 h (40 KHz). After centrifugation and washing with ultrapure water, the obtained precipitate was freeze-dried to obtain Co_SA/Ti₃C₂T_x samples. The Fe_SA/Ti₃C₂T_x, Cu_SA/Ti₃C₂T_x, Mn_SA/Ti₃C₂T_x, and Ni_SA/Ti₃C₂T_x were synthesized in the same way by replacing the CoCl₂ solution with FeCl₃, CuCl₂, MnCl₂, and NiCl₂ solution. All the ultrasonic treatments were carried out in Ar atmosphere.

ARG Removal Tests.

The amplicons of tetA (214 bp) and ampC (231 bp) were selected as the target ARGs, and the details of extracted amplicons are detailed in SI Appendix, Text S1. The removal tests of ARGs by catalysts were carried out in a 10 mL tube at 25 ± 0.2 °C with an initial solution pH of ~7. For each of the experiments, the required amount of selected catalyst and PMS were added in 5 mL synthetic wastewater containing ARGs (~10¹² copies/mL) with continuous rotation (100 rmp). Prior to oxidation reaction, the required amount of catalyst (0 to 250 mg/L) was homogeneously dispersed in the tube for 30 min to achieve an adsorption–desorption equilibrium. Whereafter, the required amount of PMS (0 to 250 mg/L) was added to the reaction medium to initiate the catalytic degradation experiment. At predetermined time intervals, 0.3 mL of the reaction solution was taken out and immediately quenched with 10 μL of sodium thiosulfate solution (20 g/L). The samples were quantitated by real-time (qPCR, details are shown in SI Appendix, Text S2). It is worth noting that when the Fenton-like treatment samples were measured by qPCR, the catalysts in the sample were retained to eliminate the effect of adsorption on the degradation efficiency. The reaction kinetics for the degradation of ARGs is analyzed using ln (C_t/C₀) = −k × t, where C_t and C₀ are the concentration of ARGs at reaction time (t) and initial time (t₀), respectively, and k is the first-order kinetic constant (min⁻¹). For the cycling test, the catalyst was recycled after each run of the experiment by filtration and washed thoroughly with ethyl alcohol and deionized water.

The adsorption kinetic curves were fitted by pseudo-first-order, pseudo-second-order, and EMTM, and the adsorption isotherms were fitted by Langmuir and Freundlich isotherm models. The details are described in SI Appendix, Test S4. The details of desorption experiments are shown in SI Appendix, Test S5. To investigate the antiinterference ability of Co_SA/Ti₃C₂T_x to environmental factors, the effects of organic matter, pH, and real water environment on the removal of ARG by Co_SA/Ti₃C₂T_x were tested. The details are described in SI Appendix, Text S5. The details of materials characterizations, ROS tests, and electrochemical measurements are shown in SI Appendix, Texts S6–S8.

X-ray Absorption Spectroscopy Measurements.

The Co K-edge and Ti K-edge X-ray absorption spectra were recorded at the 1W1B station in Beijing Synchrotron Radiation Facility under ring conditions of 2.5 GeV and ~500 mA. Data acquisition was carried out in ionization chamber transmission mode using Si (111) double-crystal monochromator. All spectra were collected under environmental conditions.

Calculation Details.

DFT calculation was carried out using modeling DMol3, as implemented in Materials Studio environment. DNP set with the generalized gradient approximation of the Perdew–Burke–Ernzerhof functional (GGA-PBE) was used to optimize all the spin unrestricted structures. The core electrons of atoms were treated using all-electron method. The Brillouin zone was sampled with a (3 × 3 × 1) Monkhorst–Pack k-point grid. A 4.4 Å orbital cutoff is selected with a Fermi smearing of 0.005 Ha to accelerate the convergence of calculations. The SCF tolerance was 1.0 × 10⁻⁶, and the convergence tolerances of energy, force, and displacement are less than 1.0 × 10⁻⁵ Ha, 0.002 Ha/Å, and 0.005 Å, respectively. A 3 × 3 × 3 supercell along [002] orientation of Ti₃C₂T_x crystal was applied for all calculations.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

All study data are included in the article and/or SI Appendix.