Tetrabutyl titanate [Ti(OBu)₄], silver nitrate (99.8%, AgNO₃), absolute ethanol (99%), glacial acetic acid (98%), zinc nitrate hexahydrate, 2-methylimidazole, N,N-dimethylformamide (DMF), dichloromethane (DCM) and chloroform were purchased from Tianjin Guangfu Fine Chemical Research Institute Co., Ltd. China. Escherichia coli (E. coli) and B. subtilis (B. subtilis) were obtained from the private collection of the Biology Department, School of Life Science and Technology, Changchun University of Science and Technology. All biological reagents were purchased from Ob-xing Biological Co., Ltd. China. Deionized water was used throughout the experiments except for antibacterial activity evaluation experiments when sterile water was used. All chemicals were used directly without further purification.

Ag-TiO₂ was synthesized by the sol-gel method using Ti(OBu)₄ and AgNO₃ as the titanium source and dopant, respectively. Solution 1 was prepared by adding a few drops of Ti(OBu)₄ (8.5 mL) to absolute alcohol (20.0 mL). A certain amount of AgNO₃, acetic acid (2 mL), pure ethanol (6 mL) and a small amount of deionized water were mixed to obtain Solution 2. After the two solutions became abrasive, solution 1 was added dropwise to solution 2 under stirring conditions until a clear sol was produced and aged for 12 h to form a gel. Then, the gel was heated to 200°C at 2°C/min, maintained 120 min and cooled under ambient condition. The final product was obtained by washing three times in deionized water, baking at 70°Ctemperature overnight and then crushing into a purple-black powder. The theoretical content of Ag among the prepared Ag-TiO₂ was 1wt%∼6wt%. For the sake of comparison, the above-mentioned steps were repeated in the absence of AgNO₃ to prepare TiO₂.

The Ag-TiO₂/ZIF-8 ternary composites were synthesized by the solvothermal method. ZnNO₃•6H₂O (0.478 g), 2-methylimidazole (0.120 g) and Ag-TiO₂ (0.008 g) were dissolved in DMF (36 mL) and stirred thoroughly. A certain mass of 4 wt% Ag-TiO₂ was added into the above solution and stirred well to prepare a synthetic mixed solution. Then the mixture solution was heated to 140°C at 5°C/min and held for 24 h, then cooled to ambient temperature at 0.4°C/min. The final product was washed sequentially with chloroform, DMF and DCM, dried at 50°C, and ground to a brownish-yellow powder to finally obtain 4 wt% Ag-TiO₂/ZIF-8. In the synthesized Ag-TiO₂/ZIF-8, the theoretical content of Ag-TiO₂ ranged from 5 wt% to 70 wt%, and the specific modification amounts were shown in Supporting Material Supplementary Table S1. For the sake of contrast, the same procedure was followed for the preparation of ZIF-8 in the absence of Ag-TiO₂.

The crystal structure of the material was studied on a Ricoh Y-2000 X-ray diffractometer with data recorded by powder X-ray diffractometer (XRD) in the range of 10° to 90°. The composition and chemical bonding of the materials were tested by Fourier transform infrared spectra (FT-IR, IFS 66V/S, Germany). The morphologies of the materials were carried out with transmission electron microscopy (TEM, Hitachi-H600). The elemental analysis was analyzed by energy dispersive spectroscopy (EDS) attached to SEM. X-ray photoelectron spectroscopy (XPS) results of materials were determined through Rigaku 2500/PC spectrometer. (In this experiment, Mg was chosen as the target source for the XPS test.) The surface area of each material was examined from the nitrogen adsorption-desorption isotherm using the Micromeritics ASAP 2020 system.

The antibacterial performance of the material was evaluated by measuring the zone of inhibition (ZOI), bacterial growth curve, minimum bactericidal concentration (MBC), and observing the antibacterial durability of the material. E. coli and B. subtilis were selected as test organisms and tested under dark and UV conditions. The specific experimental operations were as follows: (a) Zone of inhibition: The bacterial enrichment solution (80 μL) and the prepared material (100 mg) were put into the medium, then sealed and inverted in a shaker in a constant temperature incubator for 1 day, and the diameter of the zone of inhibition was measured with a ruler. (b) Bacterial growth curves: During the bacterial growth curve experiment, the prepared material was dissolved in fresh LB liquid medium (150 mL) and sterilized at 121°C for 20 min. After cooling, bacterial enrichment cultures (10 mL) were added and incubated at 37°C. The optical density (OD) of the bacteria was then measured using a UV-Vis spectrophotometer and the growth curve of the bacteria was plotted. (c) The MBC was obtained by preparing solid media with different amounts of preparation materials, adding bacterial enrichment, and measuring the number and condition of bacterial growth. (d) The durability of the antibacterial activity of the prepared materials was determined by observing changes in the zone of inhibition on the solid culture medium after 5 months.

The crystal structure and phase constitution of as-synthesized materials were identified by XRD, and the patterns are illustrated in Figure 1.

As shown in Figure 1A, the diffraction peaks (2θ) of Ag-TiO₂ at 25.3^o (101), 37.8^o (112), 48.0^o (200), 62.7^o (204), 68.8^o (116), 75.3^o (215) and 82.5^o (303) are consistent with the standard XRD data of anatase phase TiO₂ (JCPDS No. 21-1272), showing that the TiO₂ used as a carrier has an obvious anatase structure, and the phase structure of TiO₂ did not change significantly after the silver modification. The other peaks of Ag-TiO₂ at 38.2^o (111), 44.5^o (200), 64.5^o (220) and 77.5^o (311) correspond to the cubic crystal form of pure silver (JCPDS No. 03-0921). The change from baryons to singletons at 2θ ≈ 54.4° may be due to the presence Ag⁰ and AgO_x. Therefore, the peaks of Ag-TiO₂ matched well with Ag and TiO₂, proving that Ag-TiO₂ material was successfully prepared. Figure 1B suggests that the diffraction peaks of ZIF-8 is in good agreement with the previously published results (Jia et al., 2020), and the sharp diffraction peaks illustrated good crystal shape, confirming the formation of pure ZIF-8 phase. All the diffraction peaks of Ag-TiO₂/ZIF-8 can be associated with crystalline ZIF-8 and Ag-TiO₂, and no peaks of impurities are detected, which indicates the successful synthesis of Ag-TiO₂/ZIF-8. Moreover, Ag-TiO₂ doping into ZIF-8 does not affect the integrity of its crystal structure and framework.

The FT-IR spectra of the prepared materials are displayed in Figure 2. The FT-IR spectra of 20 wt% Ag-TiO₂ are shown in Supplementary Figure S1. The characteristic band of TiO₂ at 450 cm^-1 is related to the vibrational mode of Ti-O-Ti bond, which shifts slightly to a lower frequency (443 cm^-1) after the silver modification (Li et al., 2019). In the spectrum of ZIF-8, the stretching pattern of the NH groups was discovered by analyzing the absorption band from 3300 to 3000 cm^-1 (Troyano et al., 2019). The bands at 1620, 1440 and 421 cm^-1 are attributed to the stretching modes of the C = N, C-N and Zn-N bonds, respectively (Abdi, 2020; Lia et al., 2020). The adsorption bands in the 600-1300 cm^-1 region are attributed to stretching and bending vibrations of the imidazole ring (Li Y. Y. et al., 2020). These assignments are in common with the previously reported results (Vaidya et al., 2019). All characteristic bands of 4wt.%Ag-TiO₂ and ZIF-8 can be observed in the spectra of 20wt.%Ag-TiO₂/ZIF-8, indicating the successful construction of the complex. The smaller displacements at 1620 cm^-1 and 1440 cm^-1 may be due to the interaction of various functional groups among Ag-TiO₂/ZIF-8.

The TEM study the morphology and microstructure of the prepared materials. The results shows that pure ZIF-8 crystal has a dodecahedral sodalite structure with many hexagonal and quadrilateral planes (Supplementary Figure S2A), which is consistent with the previous report (Bin et al., 2020). In Supplementary Figure S2B, silver granules are well scattered on the surface of TiO₂, and this is because the radius of Ag⁺ (12.6 nm) is larger than that of Ti⁴⁺ (7.45 nm), causing silver ions hardly enter the lattice of TiO₂ (Prabhu and Valan, 2020). Ag-TiO₂/ZIF-8 consists of Ag, TiO₂ and ZIF-8, and spherical Ag-TiO₂ particles are observed on the surface of the ternary complex (Supplementary Figure S2C), demonstrating the modification of Ag-TiO₂ makes no difference for the crystal structure and the integrity of framework. Furthermore, a typical EDS spectrum in Supplementary Figure S2D verifies the existence of C, O, N, Ag, Ti, Zn elements, and the silver content is 0.72 wt%. The Mapping of Ag-TiO₂/ZIF-8 is shown in Supplementary Figure S2E–J, and it can be found that the prepared materials have good dispersion properties.

The surface electronic elemental composition of the prepared materials is conducted by XPS. Supplementary Figure S3 confirms the existence C, N, O, Zn in ZIF-8, C, O, Ag, Ti in Ag-TiO₂ and C, N, O, Ag, Ti, Zn in Ag-TiO₂/ZIF-8. The element C is derived from the ligand of ZIF-8 and CO₂ adsorbed on the surface of the materials in the C 1s spectrum (Figure 3A). The main peaks located at 284.6 eV and 288.6 eV of 20 wt.%Ag-TiO₂/ZIF-8 are ascribed to C-C and C-N, respectively (Shwetharani et al., 2019; Ji et al., 2020). In comparison with the cohesive energy of O in TiO₂ and ZIF-8, two peaks are found and the binding energy display negative shifts after loading (Figure 3B), suggesting interactions exist ternary complex. The peak at 531.6 eV and 529.8 eV correspond to O 1s binding energy of C=O and lattice oxygen of anatase phase TiO₂, respectively (Singh and Mehata, 2019). The N 1s peak at 399.1 eV can originate from sp²-hybridized nitrogen (C=N-C) (Figure 3C) (Wang et al., 2020). The 6.0 eV variance between the binding energy of Ag3d_5/2 and Ag 3d_3/2 peak among Ag-TiO₂/ZIF-8 (Figure 3D) is characteristic of metallic Ag 3d state, implying that Ag exists as Ag⁰ and AgO_x rather than Ag⁺, which is consistent with the results of XRD. In Figure 3E, the binding energy located at 458.5 eV and 464.3 eV were put down to Ti 2p_3/2 and Ti 2p_1/2, respectively, which is consistent with the Ti 2p spectrum of TiO₂ (Rempel et al., 2020). By comparison, Ag-TiO₂/ZIF-8 (Figure 3D, E) displayed small shifts, revealing Ti-O-Ag chemical bonding exists. The Zn 2p peak splits into two peaks of Zn 2p_3/2 (1021.5 eV) and Zn 2p_1/2 (1044.5 eV) (Figure 3F) (Ren et al., 2019), demonstrating the state of Zn element remains unaltered after synthesizing Ag-TiO₂/ZIF-8.

Supplementary Figure S4A displays the N₂ adsorption-desorption isotherms for 4 wt.%Ag-TiO₂, ZIF-8 and different loadings of Ag-TiO₂/ZIF-8. The Ag-TiO₂ shows type IV isotherms with hysteresis loops, demonstrating the prepared material is mesoporous. The ZIF-8 belongs to type I isotherm, indicating the presence of microporous, which is in line with the results published previously (Li C. E. et al., 2020). Meanwhile, the pore-size distributions also show that ZIF-8 was mainly microporous (Supplementary Figure S4B). Between different loadings of Ag-TiO₂ (20 wt%, 50 wt%), it was found that the increase of N₂ adsorption slowed down with increasing loading in the region of P/P₀ < 0.01, and a hysteresis loop appeared in the range of P/P₀ = 0.4–0.8, which indicated that the microporous structure characteristics became less pronounced and the mesoporous structure gradually became obvious. Therefore, a conclusion could be drawn from the above discussion that Ag-TiO₂ was attached to the surface of ZIF-8 and covered the microporous of ZIF-8. This assumption was consistent with the TEM image. Moreover, the average pore size of 20wt% Ag-TiO₂/ZIF-8 (4.46 nm) was larger than that of ZIF-8 (3.84 nm), and the BET surface areas was smaller than that of ZIF-8 (637.9 m²▪g^–1) (Supplementary Table S2), which demonstrated Ag-TiO₂ exists on the surface of ZIF-8, confirming the conclusion of N₂ adsorption-desorption isotherms.

Based on the above experimental results of antibacterial activity evaluation, it can be concluded that the antibacterial activity of Ag-TiO₂/ZIF-8 is better than that of Ag-TiO₂ and ZIF-8. The possible antibacterial mechanism of Ag-TiO₂/ZIF -8 was proposed as follows.

(i) When Ag-TiO₂/ZIF-8 is added to a solution containing bacteria, it can release Ag⁺ and Zn²⁺ with low toxicity and biological antibacterial activity, these metal ions can come into direct contact with the bacteria and bind to component proteins in cell walls and membranes, denaturing and inactivating the component proteins (Chen et al., 2023; Wang et al., 2023). This causes the outer layer to lose its protection and the bacteria burst and die (Mariappan et al., 2022; Sun et al., 2023). In addition, the interaction between Ag⁺ and DNA structures can prevent bacterial reproduction (Cao et al., 2020; Liu et al., 2022). (ii) TiO₂ itself has some photocatalytic antibacterial activity. When the energy provided by UV radiation is greater than 3.2 eV, the photogenerated electrons (e_CB ⁻) and holes (h_VB ⁺) generated by TiO₂ react with water and oxygen to form free ROS, which kill bacteria by destroying cell walls and solidifying viral proteins (Zerjav et al., 2017). Moreover, when TiO₂ is combined with Ag and ZIF-8 to form a ternary complex, the Ag⁺ and Zn²⁺ in the material can actively inhibit the electron-hole pair recombination and further improve the photocatalytic antibacterial activity (Yang et al., 2019). And Ag-TiO₂ can also work in dark conditions where bacteria tend to multiply, so the application field of antibacterial agent will be further expanded without the limitation of light sources. (iii) On account of the high BET surface area and porous structure of ZIF-8, more Ag-TiO₂ is uniformly dispersed on its surface, which increases the contact area between the bactericidal agent and bacteria, thereby improving the antibacterial activity. In addition, Ag⁺ and Zn²⁺ can be slowly released from Ag-TiO₂/ZIF-8 to achieve a long-term antibacterial effect. In summary, the high antibacterial activity of Ag-TiO₂/ZIF-8 comes from the synergetic effect between Ag-TiO₂ and ZIF-8.