Close this search box.

Tuning oxidant and antioxidant activities of ceria by anchoring copper single-site for antibacterial application – Nature Communications

Synthesis and characterization of Cu-CeO2 SSE

The fabrication of Cu-CeO2 catalyst is illustrated in Fig. 1a. First, CeO2 nanosphere is obtained by a solvothermal approach, followed by calcinating step under air condition. Then, Cu ions could be adsorbed and deposited on ceria substrates in the presence of alkaline solutions. Finally, the obtained composite is heated in air first, and then pyrolyzed under hydrogen atmosphere to get the target Cu-CeO2 product. Owing to the Cu-O-Ce interaction, Cu species are present as a dispersive state in Cu-CeO2 sample after calcined in air. The scanning electron microscopy (SEM) image in Fig. 1b and transmission electron microscopy (TEM) image in Fig. 1c clearly reveals that the as-prepared Cu-CeO2 possesses a spherical morphology with a quite rough surface, which is similar to their parent ones in Supplementary Figs. 1, 2. Supplementary Fig. 3 shows high-resolution TEM (HRTEM) image of Cu-CeO2 precursor and the positions circled by two rectangular frames are enlarged on the right. Obviously, two representative lattice fringes with a spacing of 0.28 nm are at an angle of 90° in the upper right. The interplanar spacing of the exposed plane with an included angle of 45° is 0.19 nm. As shown in the lower right of Supplementary Fig. 3, the lattice fringes with an interplanar spacing of 0.31 nm correspond to the (111) plane of the sample. Element mapping results in Supplementary Fig. 4 demonstrate the homogeneous distribution of Cu, O and Ce components throughout the sample. The aberration corrected high angle annular dark field-scanning transmission electron microscopy (AC-HAADF-STEM) result in Supplementary Fig. 5 also shows that there are no obvious crystallized copper species in the form of clusters or small particles in the structure. Energy dispersive spectroscopic (EDS) mapping analysis (Fig. 1d) signifies a uniform distribution of Cu component in Cu-CeO2 catalyst. Figure 1e shows the HRTEM image of Cu-CeO2. It can be seen that there are no clusters or small particles throughout the structure. The results in Fig. 1f, g indicate that the main interplanar spacing is about 3.2 Å by Z-contrast analysis.

Fig. 1: Schematic illustration and morphology characterization for the Cu-CeO2 catalyst.
figure 1

a Schematic illustration of the synthetic procedure of Cu-CeO2 sample, b Scanning electron microscopy images of Cu-CeO2 catalyst (scale bar: 200 nm), c Representative TEM image of Cu-CeO2 catalyst (scale bar: 100 nm), d TEM image and corresponding elemental mapping of Cu, O and Ce recorded on the nanoparticles (scale bar: 100 nm), e HRTEM images of Cu-CeO2 catalyst, f AC-HAADF-STEM image of Cu-CeO2 catalyst (scale bar: 5 nm), g Corresponding Z-contrast analysis of region A and B in (f). All experiments were independently repeated three times with similar results.

Figure 2a shows the X-ray diffraction (XRD) patterns of CeO2 and Cu-CeO2 samples, respectively. It can be seen that the main diffraction peaks of two samples are indexed to the cubic structure of CeO2 (JCPDS card no. 34-0394). Except for the diffraction peaks of CeO2, no agglomerated Cu species are detected, indicating that the structure of CeO2 is not changed after the introduction of Cu. In addition, the XRD pattern of Cu-CeO2 sample is fitted in Supplementary Fig. 6, which yields a crystallite size of smaller particle as 9 nm and the unit cell parameters as a = b = c = 5.3912 Å and α = β = γ = 90°. On this basis, the model structural diagram of parent cubic CeO2 is constructed and illustrated in Supplementary Fig. 7. From the visual point of view in Supplementary Fig. 7b, the lattice fringes which correspond to (200) and (002) facets are perpendicular to each other and the lattice fringes at an included angle of 45° belong to the (220) facet, which is consistent with the above results obtained from the electron microscope analysis. In order to further accurately determine the nanostructure of the material, we also carry out Raman spectroscopy measurements. In Fig. 2b, the peak centered at 460 cm−1 is attributed to the F2g vibrational mode of CeO2 crystal31. Compared with parent CeO2, the softening of the F2g mode of Cu-CeO2 is accompanied by the appearance of a broad feature centered around 600 cm−1, which belongs to the defect-induced mode (D)32. The appearance of D band is due to the formation of defect species in Ce-O coordination, which leads a consequence that the vibration signal of Ce-O cannot be cancelled in all directions33. Since there exist a Cu-O-Ce coordination structure in Cu-CeO2 catalyst, the interaction between Cu-O and Ce-O is not equal, which leads to the difference in the vibration of Ce-O in different directions. Thus, the increasing disordering level of Cu-CeO2 leads to the variation of D band. In addition, Raman peaks at 290 cm−1 and 340 cm−1, which assigned to CuO phase, are not observed34,35. The elemental composition and valence states are investigated by X-ray photoelectron spectroscopy (XPS) technique (Supplementary Fig. 8, Fig. 2c). The XPS spectra of Ce 3d (Fig. 2c) are fitted into 10 peaks, which are attributed to the Ce4+ species at v, v′′, v′′′, u, u′′, u′′′ and the Ce3+ species at v0, v′, u0, u′, respectively23,36,37,38. Supplementary Table 1 shows the actual Ce3+/Ce ratios of both CeO2 and Cu-CeO2 samples. It is obvious that the Ce3+ content for Cu-CeO2 is nearly 18% from XPS analysis, which is lower than that of CeO2 (23%). These results indicate that the introduction of Cu element into parent CeO2 would reduce its surface Ce3+ concentration. Similar results have also been reported in previous report38. According to the authors’ calculations and structural characterization, the substitution of one Ce3+ adjacent to an oxygen vacancy (VO) by one Cu2+ normally accompanies a phenomenon that the other Ce3+ would be readily converted to Ce4+ for the charge balance, and therefore the introduction of Cu2+ will reduce the Ce3+/Ce4+ ratio compared with original CeO2. The O 1s spectra of CeO2 and Cu-CeO2 samples are illustrated in Supplementary Fig. 8b. The peak appearing at approximately 529 eV (OI) can be attributed to the lattice oxygen of Ce4+ and the feature at 530.7 eV (OII) corresponds to oxygen vacancies or the lower-coordination lattice oxygen of Ce3+36,37. Also, the higher binding energy feature at 531.6 eV is also associated with the presence of surface hydroxy-containing groups. It is obvious that parent CeO2 possesses a larger OII/(OI + OII) ratio (23.3%) than that of Cu-CeO2 (19.8%), which is in line with the Ce 3d XPS results.

Fig. 2: Structural characterizations of Cu-CeO2 catalyst and reference materials.
figure 2

XRD patterns (a), Raman spectra (b) and Ce 3d photoelectron profiles (c) of the CeO2 and Cu-CeO2 catalysts, d XANES spectra of Cu foil, CuO and Cu-CeO2 sample, e Fourier transforms of the Cu K-edge EXAFS oscillations of the materials mentioned above, f FT-EXAFS spectra of Cu-CeO2 sample at the Cu K-edge and the corresponding Fitting results. Source data are provided as a Source Data file.

Considering that the surface oxygen is prone to be removed under vacuum conditions, the structural information of Cu can be better reflected by means of Synchrotron radiation characterization. X-ray absorption fine structure (XAFS) measurements can effectively detect the local coordination states and electronic structure of copper on CeO2 support. Figure 2d shows the Cu K edge X-ray absorption near-edge structure (XANES) profiles of Cu-CeO2, Cu foil, and CuO samples. According to the position of near-edge absorption energy, it can be concluded that the Cu species bear an oxidation valence state in Cu-CeO2 sample. The Fourier-transformed (FT) extended X-ray absorption fine structure (EXAFS) curve of Cu-CeO2 (Fig. 2e) exhibits only one prominent peak at approximately 1.9 Å, corresponding to the first shell of Cu-O scattering interaction. No appreciable Cu-Cu coordination characteristic peak is detected at 2.24 Å, signifying that there is no formation of Cu-Cu bond. We also performed CO-probe molecule Fourier transform infrared (FTIR) measurements to investigate the nature of Cu metal sites. In situ diffused reflectance infrared Fourier transform spectroscopy (DRIFTS) was used to investigate the adsorption ability of CeO2 and Cu-CeO2. As shown in Supplementary Fig. 9a, Cu-CeO2 catalyst exhibits a band around 2105 cm−1, assigned to the linear CO adsorbed on Cu+ sites (Cu+-CO), indicating that CO was adsorbed on the Cu+ sites34,39,40,41,42. In addition, the adsorption intensity of the peak gradually increased with the time, and reached saturation adsorption at 480 s with a maximum adsorption peak at 2111 cm−1. Traditionally, the IR band at 2069 cm−1 is regarded as the CO adsorption on Cu0 site40,41. This indicates that there are no Cu0 species in this Cu-CeO2 catalyst. For CeO2, two obvious gaseous peak of CO are shown in Supplementary Fig. 9b, which indicating that the CO absorption on parent CeO2 is absent. XANES spectra at the Ce M5,4-edge were normalized and are shown in Supplementary Fig. 10a. The XANES at Ce M5,4-edge of CeO2 based materials correlates with the Ce 3d3/2 and 3d5/2 core level transitions into the 4f unoccupied electronic state43. The intense peak S and peak P represent the tetravalent Ce (4f0) while the weak peak R indicates the contribution of trivalent Ce (4f1) states44,45. As it vividly shows, Cu-CeO2 sample exhibits a reduction of Ce3+ (peak R) compared with parent CeO2 while an enhancement of Ce4+ (peak S). In other words, the Ce3+/(Ce3+ + Ce4+) ratio of Cu-CeO2 is lower than that of CeO2. The result is in good agreement with XPS results (Fig. 2c, Supplementary Table 1). Supplementary Fig. 10b shows the O K-edge XANES spectra. Three main peaks are attributed to the O 2p states that are hybridized with Ce 4f, 5d(eg) and 5d(t2g) states, respectively. The intensity variation is attributed to the structural disorders induced by Cu doping owing to the formation of Cu-O-Ce coordination network43. The quantitative FT-EXAFS fitting is conducted to shed light on the structural configuration of Cu (Fig. 2f) and the corresponding coordination parameters are shown in Supplementary Table 2. It can be seen that the coordination number of the center Cu atom is 3.2 and the mean bond length of Cu-O is about 1.94 Å. Supplementary Fig. 11 shows the q-fitting (inverse Fourier transform) curve of Cu-CeO2, which is consistent with the above R-space fitting results.

Evaluation of Cu-CeO2 SSE for POD-like and HORAC activity

Multiple enzyme-mimicking activities of Cu-CeO2 were investigated in vitro. Firstly, the POD-like activity of Cu-CeO2 was tested based on the principle that H2O2 could be catalytically decomposed by Cu-CeO2 to generate ·OH, and TMB could be oxidized by ·OH to oxidized-TMB (ox-TMB)46. The absorbance of the reaction product at the wavelength of 652 nm was read by a microplate reader. Figure 3a shows that the POD-like activity of Cu-CeO2 was significantly higher than that of pristine CeO2. Furthermore, steady-state kinetic assay was conducted. According to Lineweaver-Burk equation, the Michaelis-Menten constant (Km) of CeO2 and Cu-CeO2 were 24.34 mM and 30.76 mM, respectively. The maximal reaction velocity (Vmax) of CeO2 and Cu-CeO2 were 28.05 nM/s and 166.7 nM/s, respectively, and Cu-CeO2 showed significantly enhanced POD-like activity (Supplementary Fig. 12, Supplementary Table 3). We further investigated the regulation of Cu content on the POD-like catalytic performance of Cu-CeO2 nanozyme. As shown in Supplementary Fig. 12 and Supplementary Table 3, firstly, all Cu-CeO2 SSE exhibited significantly enhanced POD-like activity compared to CuO. In addition, Cu content and the Vmax of Cu-CeO2 was positively correlated, which may be caused by the variation of Cu sites involved in the catalytic reaction. However, after calculating the turnover rate, we found that as Cu content increased, the turnover rate showed a gradually decreasing trend. This may be due to the presence of more CuO particles in samples with high Cu content compared to those with low Cu content, and CuO may cause a decrease in the dispersion of active centers. It is worth noting that Cu-CeO2 with a theoretical Cu content of 5%, which is also the main sample of this study, can achieve a Vmax comparable to 10% Cu sample and a turnover rate comparable to 2% Cu sample at the same mass concentration, demonstrating satisfactory catalytic performance. The Vmax and the turnover rate of 5% Cu-CeO2 was 11.78- and 212.51- fold higer than CuO nanozyme, respectively, exhibiting significant enhancement of POD-like activity.

Fig. 3: Catalytic activities and mechanism of CeO2 and Cu-CeO2 nanozymes.
figure 3

a Time-dependent optical density change at 652 nm of 3,3’, 5,5’-tetramethylbenzidine (TMB) in POD reactions. b Time dependent oxygen generation in CAT reactions. c Time dependent optical density change at 652 nm of TMB in OXD reactions. d Time dependent fluorescent intensity of 2,7-Dichlorodihydrofluorescein diacetate (DCFH) in HORAC reactions. e Fluorescent spectra of DCFH after 10 min reaction with H2O2 and different nanozymes. f Electron Paramagnetic Resonance (EPR) spectra of DMPO-OH after 10 min reaction with H2O2 and different nanozymes. g The calculated model of pristine CeO2 (left), CeO2 with Cu3c added on the 3O atoms on the surface (Cu-ad, middle) and CeO2 with a Ce atom substituted by Cu (Cu-sub, right). The 6 possible reaction sites are highlighted with a tagged arrow. h The reaction mechanism of POD and HORAC process, M indicates the metal sites. i The calculated PDS reaction energy of POD and HORAC processes for different reaction sites with the exact PDS labeled above the bar. j The calculated PDOS, i.e., the d and f band summation of different reaction centers. k The proposed mechanism of regulation of catalytic activities by Cu-CeO2 single-site nanozyme. Source data are provided as a Source Data file.

Furthermore, the cyclic stability of Cu-CeO2 was tested, the results showed that after 30 catalytic cycles, the POD-like activity of Cu-CeO2 was still comparable with the original nanozyme (Supplementary Fig. 13). The ICP-MS analysis showed that Cu accounts for 4.43 wt % of the Cu-CeO2 sample. Hence, when the concentration of Cu-CeO2 reaches 200 μg/mL, the total content of Cu is 8.858 μg/mL. The content of Cu in the supernatant after cyclic reaction was below 0.500 μg/mL, much lower than the total amount of Cu in the reaction system. In summary, Cu-CeO2 nanozyme has good stability, within our test conditions.

As a reaction substrate, H2O2 can also be catalytically decomposed into O2 by Cu-CeO2, which increases the concentration of dissolved O2 in the liquid environment. The CAT-like activity of Cu-CeO2 was investigated subsequently. Figure 3b shows that the CAT-like activity of Cu-CeO2 was significantly higher than that of pristine CeO2. In addition, O2 can be catalyzed by OXD-like activity of Cu-CeO2 to generate O2·-, which oxidizes TMB to ox-TMB47. The absorbance of the reaction product at the wavelength of 652 nm was read by a microplate reader, and Cu-CeO2 showed slightly enhanced OXD-like activity (Fig. 3c). The CAT reaction provides O2, which further promotes the OXD reaction. The SOD-like activity of Cu-CeO2 was also tested. Xanthine-xanthine oxidase system was applied to generate O2·−, which can reduce nitro blue tetrazolium (NBT) to formazan. In the presence of SOD-mimics, O2·− can be catalytically converted to O2 and H2O2, which further provides H2O2 for POD reaction. As Supplementary Fig. 14 shows, the SOD-like activity of Cu-CeO2 was also significantly higher than that of pristine CeO2. We further tested the consumption of H2O2 in CeO2 + H2O2 and Cu-CeO2 + H2O2 systems quantitatively. As shown in Supplementary Fig. 15, as the reaction time prolongs, both systems showed a trend of H2O2 consumption, and Cu-CeO2 maintains a higher consumption than CeO2 during long-term reaction process. This may be caused by the enhanced multiple catalytic activities of Cu-CeO2, which accelerate the multiple pathway conversion related to H2O2.

Next, the HORAC activity of Cu-CeO2 was tested. The ·OH generated from H2O2 and Fenton reagent can be transformed into H2O and O2 through HORAC activity, so as to quench the free radical fluorescent probe48. As Fig. 3d shows, the fluorescence intensity of CeO2 and Cu-CeO2 groups decreased compared with the reaction baseline, indicating that both nanozymes exerted HORAC activity. However, within 10 min of reaction, the fluorescence intensity of CeO2 group drastically decreased by ~94.31%, compared with baseline, while Cu-CeO2 group only decreased by ~1.92%, and the relative fluorescence intensity was higher than that of the H2O group throughout the entire reaction process, indicating that Cu single-site led to a significant inhibition of HORAC activity of CeO2, and that Cu-CeO2 could accelerate the Fenton-like catalytic process. We further used DMPO as the spin trapping agent and measured the changes of ·OH species quantitatively through EPR. As shown in Supplementary Fig. 16 and Supplementary Table 4, after 10 minutes of reaction, both FeCl2 and FeCl2 + Cu-CeO2 group exhibited typical DMPO-OH signal. The signal intensity of FeCl2 + Cu-CeO2 group was higher than that of FeCl2 group, while the signal of FeCl2 + CeO2 group was almost invisible, which is consistent with the pattern of the fluorescence results. Quantitative calculation showed that the ·OH scavenging rate of CeO2 was 97.26% at 10 min, while the spin concentration of DMPO-OH in FeCl2 + Cu-CeO2 group reached 158.62% of that in FeCl2 group. In addition, we investigated the catalytic effect of the physical mixed system of free Cu2+ ions/CuO nanozyme with CeO2 nanozyme. As shown in Supplementary Fig. 17, the initial rate of POD-like reaction of the two physical mixed systems was similar to that of CeO2. As the reaction continued, the substrate conversion extent of the two groups at the end point of the reaction was similar to that of Cu-CeO2. However, DCFH fluorescence detection (Supplementary Fig. 18) showed that for physical mixed systems, only Cu2+ + CeO2 group showed significantly higher total ROS generation than CeO2 group, but was still much lower than that of Cu-CeO2 group, indicating that in the non-interacting Cu2+/CuO + CeO2 physically mixed system, Cu could not inhibit the HORAC activity of CeO2 effectively. Through the above results, we reaffirm the enhanced POD-like activity of Cu-CeO2 nanozyme, and a prominent inhibitory effect of Cu single sites on the HORAC activity of CeO2. Meanwhile, the interaction between Cu single sites and CeO2 support also plays a key role in the regulation of catalytic activities. Therefore, by the aid of Cu single sites and its interaction with CeO2 support, the effective regulation of the redox catalytic pathways of CeO2 nanozyme was achieved.

Previous studies have shown that structural factors such as Ce3+/Ce4+ ratio49, oxygen vacancy15, defect11,50, as well as environmental factors such as pH value51 and ·OH concentration52, will affect the type of catalytic reaction and catalytic activity of CeO2. Among these factors, Ce3+/Ce4+ ratio is one of the most important structural factors which determines the POD-like and HORAC activities of CeO211,15,49. Specifically, increasing the proportion of Ce3+ can not only improve the POD-like activity of CeO2, but also enhance its HORAC activity. In this study, XPS results indicated that the Ce3+/Ce4+ ratio of Cu-CeO2 was lower than that of CeO2. Combined with the catalytic activity results, we speculated that the introduction of Cu site inhibits the HORAC activity of CeO2 carrier, and may also lead to the decrease of its POD-like activity. Under the synergistic effect of the intrinsic activity of Cu site, the overall POD-like activity of Cu-CeO2 SSE was maintained.

In order to further explore the generation of reactive oxygen species (ROS) in the catalytic decomposition of H2O2 by Cu-CeO2, 2,7-Dichlorodihydrofluorescein diacetate (DCFH-DA) was adopted as total-ROS detecting fluorescent probe. As Fig. 3e shows, Cu-CeO2 + H2O2 group produced strong fluorescence signal, H2O2 group produced weak fluorescence signal, while CeO2 + H2O2 group exhibited even lower fluorescence than H2O2 group. These results indicated that Cu-CeO2 + H2O2 system can produce abundant ROS efficiently, and CeO2 + H2O2 system demonstrated an outcome of ROS removal. Terephthalic acid (TA) was adopted as ·OH detecting fluorescent probe. The characteristic fluorescent sprectrum of hydroxyterephthalic acid (TAOH) in Cu-CeO2 group indicated the prescence of ·OH (Supplementary Fig. 17). Next, 5,5-dimethyl-1-pyrroline N-oxide (DMPO) was adopted as the spin trap and EPR was used to detect ·OH generation in Cu-CeO2 + H2O2 system. As Fig. 3f shows, no obvious DMPO-OH signal was detected in H2O2 or CeO2 + H2O2 group, while Cu-CeO2 + H2O2 exhibited significant DMPO-OH signal. The catalytic generation of O2· by Cu-CeO2 in methanol was also detected. As Supplementary Fig. 18 shows, among all three groups, only Cu-CeO2 group showed significant DPMO-O2· signal. Combined with the catalytic activity assays above, the results indicated that both CeO2 and Cu-CeO2 can catalyze the decomposition of H2O2 to generate ·OH through POD-like activity, and Cu-CeO2 exerts higher POD-like activity, producing more ·OH. CeO2 may remove ·OH through its high HORAC activity. However, in Cu-CeO2 + H2O2 system, there remain abundant ·OH radicals due to the inhibited HORAC activity of Cu-CeO2.

Theoretical analysis

Density functional theory (DFT) is employed to elucidate the underlying mechanism for the boost of ROS generation of Cu-CeO2. Pristine CeO2 (111) surface (left Fig. 3g) and Cu-CeO2 (111) surface are built for comparison. Firstly, two configurations of Cu-CeO2, including one that has a Cu atom directly adhering on the three O atoms on the (111) surface (Cu-ad, middle Fig. 3g), and one that has a Cu atom substitute the surface Ce atom (Cu-sub, right Fig. 3g), are optimized. It should be noted that due to the weak O-binding energy for Cu atom, originally presented surface O3c transformed to O2c in the Cu-sub model, which is with stronger alkalinity and will be protonated in the solution53. Thus, the protonated Cu-sub model is finally adopted. Different reaction sites in these systems are considered. For the pristine (111) surface, only one type of reaction site of Ce7c presents. While for the Cu-ad and Cu-sub surface, two and three types of reaction sites are considered, respectively, as shown in Fig. 3g. In the Cu-ad system, both Cu (Cu@Cu-ad) and Ce (Ce7c@Cu-ad) adjacent to Cu are considered. While for the Cu-sub system, the Ce7c site adjacent to the protonated O atom (Ce7c@Cu-sub) and the Cu site (Cu@Cu-sub), along with the double Ce6c site (diCe6c@Cu-sub) that undergoes lattice oxygen mechanism which involves the protonated O2c during the reaction, are considered. DFT optimizations were then performed for the reaction intermediates on different sites. The optimized geometries are illustrated in Supplementary Figs. 2126.

In this work, a large amount of ·OH radicals are detected in the experiment and are considered as the major disinfection factor. It is crucial to study the activity of POD that generates ·OH, including reaction P1 to P4 in Fig. 3h.

The reaction energy for the potential determining steps (PDS) of different reaction sites are illustrated in the top of Fig. 3i, with the exact PDS labeled on the bars. The detailed reaction energy along the POD pathway are listed in Supplementary Table 5. It can be found that the Ce7c site on the pristine CeO2 (111) surface has a poor POD activity due to the huge energy gap of 2.591 eV during reaction P2 as the PDS. Meanwhile, in the Cu-ad system, the Cu@Cu-ad center served as a catalytic site with a drastically decreased energy of 1.358 eV, while the adjacent Ce7c@Cu-ad sites also have an increased activity with PDS energy as low as 1.792 eV. The influence of Cu in the Cu-sub system are, on one hand, greatly increased the activity of the adjacent Ce7c@Cu-sub and especially the diCe6c@Cu-sub sites, with the PDS energy of 1.556 eV and 0.970 eV, respectively. On the other hand, little change of the activity for Cu@Cu-sub center itself has been calculated (2.479 eV). These results indicated that Cu single-site can largely promote the overall POD activities with the activity of itself and via activating the adjacent Ce sites.

The electronic structure analysis, namely the projected density of state (PDOS) analysis, is further conducted for different reaction sites to elucidate the exact mechanism for activity promotion. The summation of d and f bands for the reaction sites are shown in Fig. 3j. Firstly, Cu@Cu-ad can serve as the catalytic center in Cu-ad system due to the fact the site is under-coordinated (Cu3c) and that the d band of the center is concentrated near the fermi-level compared with the Ce7c site, which served a stronger interaction between the centers and the oxidized species, stabilized the *OH intermediates and decreased the energy for the PDS (P2). Cu@Cu-sub on the other hand, due to being fully coordinated (Cu4c), the d-band is far away from the fermi-level, weak binding with oxidized species is anticipated and no better performance is calculated. Secondly, Cu single-site also influenced the performance of adjacent Ce. Comparing with the Ce7c plot, a shift to a higher energy can be spotted in Ce7c@Cu-ad and diCe6c@Cu-sub. Such shift is mainly originated from the less coordinated environment of Ce and will contribute to a higher energy of the anti-bond band of the bond between Ce and adsorbed O, further stabilizing the adsorbed *OH structure. In this case, we see a drastic decrement in reaction energy for reaction P2 of Cu-CeO2. Meanwhile, tighter Ce-O binding facilitates a harder dehydration process and turns reaction P4 into the PDS for the diCe6c@Cu-sub center. For the Ce7c@Cu-sub site, slightly difference could be told from the PDOS from Ce7c, which indicates that on one hand, the introduction of Cu has little influence on the electronic structure for the adjacent Ce7c@Cu-sub sites. On the other hand, the sharp decline for reaction energy in PDS is mainly contributed by the *OH structure that stabilized with the protonated O atom.

It should be noticed that as reported in the literature, CeO2 itself can act as an ROS elimination catalyst to prevent oxidative damage, HORAC mechanism (reaction H1 to H6 in Fig. 3h), as an antagonistic pathway of POD, should also be considered. As depicted in Fig. 3i, the PDS for most reaction sites are the dehydration reaction H3, the reaction energies of which are 0.909 eV for pristine Ce7c. Thus, it is anticipated that as the introduction of Cu to the system, the dehydration process should be hindered due to stronger oxidized species adsorption of the reaction sites on and around Cu single-site (except for Cu@Cu-sub), as discussed in the PDOS analysis. To be specific, the diCe6c@Cu-sub site has an increased PDS energy of 0.954 eV for reaction H6, which inhibited the HORAC process. Likewise, the Cu@Cu-ad, Ce7c@Cu-ad, Ce7c@Cu-sub sites all have worse HORAC performance with PDS energy increased to 1.110, 0.924 and 1.258 eV. The PDS energy on Cu@Cu-sub site is 0.652 eV, which indicated a better HORAC activity. Yet, considering the weak interaction between adsorbed ·OH and Cu@Cu-sub center, the HORAC reaction tends to take place on the Ce sites instead of Cu, which limits the promotion effect for Cu@Cu-sub center promotion to the HORAC processes.

In all, Cu single-site influences the reaction activity on and around itself. Except for the drastic energy decrement of POD, the inhibition of HORAC processes also made contributions to the overall ROS generation performance (Fig. 3k).

To elucidate the possible influence of small CuO cluster on the surface, additional DFT calculations were performed. According to previous literatures, small Cu nano-clusters on CeO2 facets prefer to be in the form of monolayers, a (CuO)3-CeO2 model was thus established54,55. The stable planar (CuO)3 cluster with similar symmetry of CeO2 (111) surface was loaded. Considering the aqueous environment of the actual experimental condition, water molecule was added and found dissociated spontaneously on the top of the CuO cluster under the strong influence of 3 under coordinated Cu, forming the hydrated site as shown in Supplementary Figs. 27, 28. Due to the structural complexity of the cluster, there are two possible reaction sites, i.e., the Cu3 site and the Cu2Ce site which surround the *OH intermediate. Both HORAC and POD-like mechanisms were calculated on these sites. The results indicated that for POD-like pathway, the energy consumption of the PDS are sharply reduced to 0.796 eV (Cu3) and 0.915 eV (Cu2Ce) mainly due to the strong interaction of the intermediates (*OH) with multiple coordination atoms. Thus, (CuO)3 cluster boosted the activity of ·OH radical generation. While for the HORAC pathway, the strong binding of O2 with multiple atoms further increase the energy consumption of the PDS (O2 desorption) to 1.081 eV (Cu3) and 1.015 eV (Cu2Ce), which deactivated the HORAC process. The above results suggest that small Cu nano-clusters also possess oxidant activities. Based on the steady-state kinetic results of POD-like activities of Cu-CeO2 samples with different Cu contents in Supplementary Table 3, when the Cu content increased from 2 to 5%, the Vmax increased by ~2.13 folds, and the turnover rate showed a slight decrease, indicating that potential existence of small amount of Cu nanoclusters in Cu-CeO2 sample may contribute to the overall POD-like activity of the nanozyme. However, when the Cu content continued to increase to 10%, the Vmax did not increase correspondingly, but remained comparable to that of the 5% sample. Meanwhile, the turnover rate drastically decreased by ~2.10 folds, suggesting that the presence of a large number of Cu nano-clusters may actually inhibit the overall catalytic activity of the nanozyme. The phenomenon we observed is consistent with that reported in previous literature56.

In vitro antibacterial performance of Cu-CeO2 SSE

After confirming the catalytic activities and underlying ROS generation mechanism of Cu-CeO2, in vitro and in vivo experiments were carried out to investigate its antibacterial properties. Methicillin-resistant MRSA and E. coli were chosen as representative strains of Gram-positive and Gram-negative bacteria, respectively. The bacterial solution was treated in six groups: PBS, CeO2, Cu-CeO2, H2O2, CeO2 + H2O2 and Cu-CeO2 + H2O2. After different treatments, the antibacterial effect of each group was evaluated by plate colony-counting. The results showed that except for the Cu-CeO2 + H2O2 group, there was no significant difference colony numbers between the other groups and the control group (P > 0.05) (Fig. 4a, b, g, i). Nanozyme solely showed no obvious antibacterial effect. For MRSA, there was no significant difference in the reduction of colony number after H2O2 or CeO2 + H2O2 treatment (~0.12-log), while Cu-CeO2 + H2O2 group completely sterilized ~6.85-log MRSA and showed superior antibacterial properties. Similarly, for E. coli, after H2O2 treatment, the colony number decreased by ~0.91-log. It’s noteworthy that after CeO2 + H2O2 treatment, the colony number decreased by ~0.56-log, which was even worse than that of H2O2 group. However, Cu-CeO2 + H2O2 group completely sterilized ~7.70-log E. coli, reversing the inhibited antibacterial effect of CeO2. The bacterial solution of each group was stained and observed using fluorescence microscope, and the results are shown in Fig. 4c, d and Supplementary Fig. 27. Almost all bacteria in groups without H2O2 were alive, with only a few dead ones. There were a few dead bacteria in H2O2 group and CeO2 + H2O2 group, most of which were living bacteria. For Cu-CeO2 + H2O2 group, no obvious living bacteria were observed. The antibacterial effect of each group was further confirmed by SEM. Except that Cu-CeO2 + H2O2 group showed obvious loss of bacterial membrane integrity, the bacterial morphology of the other groups was intact or slightly wrinkled (Fig. 4e, f, Supplementary Fig. 28). H2O2 has been widely used for debridement and disinfection of various infected wounds clinically due to its broad-spectrum antibacterial properties57. However, high concentrations of H2O2 can be harmful to normal human tissues. In addition, the antibacterial ability of H2O2 is relatively weak, and different bacteria perform varied sensitivity to H2O258. By the aid of the intrinsic peroxidase-like activity of Cu-CeO2, the application of Cu-CeO2 + H2O2 system can achieve better bactericidal performance using much lower concentration of H2O2 than that in clinic (2.5% ~ 3.5%).

Fig. 4: In vitro antibacterial properties of CeO2 and Cu-CeO2 nanozymes.
figure 4

Bacterial colonies, fluorescent images (scale bar: 50 μm) and SEM images (scale bar: 1 μm) of MRSA (a, c, e) and E. coli (b, d, f) after grouped treatment (I: Phosphate buffered saline (PBS), II: CeO2, III: Cu-CeO2, IV: H2O2, V: CeO2 + H2O2, VI: Cu-CeO2 + H2O2). Logarithm of colony forming unit (CFU) countings of MRSA (g) and E. coli (i) after grouped treatment. Logarithm of CFU countings of MRSA (h) and E. coli (j) after treatment with different concentrations of nanozymes and H2O2. A representative image of three replicates from each group is shown. Data are presented as mean values +/− standard deviation, n = 3 biologically independent replicates. Significance was calculated by two-sided Student’s t-test. Source data are provided as a Source Data file.

To further confirm ROS generation in the Cu-CeO2 + H2O2 antibacterial system, DCFH staining was performed. As shown in Supplementary Figs. 28, 29, for both MRSA and E. coli, the fluorescence of DCFH-DA in Cu-CeO2 + H2O2 group are more obvious than CeO2 + H2O2 group. In addition, the fluorescence was well co-localized with nanozyme-bacteria composite, while planktonic bacteria showed little fluorescence, indicating that the onset of ROS-mediated sterilization in Cu-CeO2 + H2O2 antibacterial system is likely to be located on the nanozyme-bacteria interface.

We further explored the rules of bactericidal changes of CeO2 and Cu-CeO2 nanozymes at different concentrations. For MRSA, the arbitrary combinations of CeO2 + H2O2 had no significant antibacterial effect, while Cu-CeO2 + H2O2 groups showed obviously positive correlation between the bactericidal effect and the concentration of nanozyme and H2O2 (Fig. 4h, j, Supplementary Figs. 30, 31). For E. coli, due to its high sensitivity to H2O2, the colony number showed a decreasing tendency with the increase of H2O2 concentration. Notably the antibacterial effect gradually deteriorated with the increase of CeO2 concentration, which was ever worse than that of H2O2 with the same concentration. Cu-CeO2 + H2O2 group still showed a positive correlation between antibacterial effect and nanozyme and H2O2 concentration. The phenomenon observed in our study is similar to that reported by Zhu et al.14. According to their report, the synthesized spherical CeO2 nanozyme with a particle size of ~150 nm exhibited optimistic POD-like activity at pH = 4.0–6.0. However, the bactericidal effect of CeO2 + H2O2 on E. coli in PBS (pH = 4.0) was significantly weaker than that of using H2O2 alone, and no significant ·OH was detected in the CeO2 + H2O2 system. Therefore, they speculated that the ROS scavenging ability of CeO2 hinders the decomposition of H2O2 to generate ·OH, and its POD-like activity did not contribute to the antibacterial effect of the CeO2 + H2O2 system. Combined with the former characterizations of catalytic activities and ROS generation in nanozyme + H2O2 system, it can be inferred that the low POD-like activity and high HORAC activity of CeO2 + H2O2 system lead to a relatively low amount of ·OH available for sterilization in the environment. Besides, the concentration of H2O2 substrate decreased as POD reaction occurred, resulting in the weakening of its antibacterial effect. In other words, CeO2 protected bacteria from H2O2 and ·OH. In contrast to CeO2, Cu-CeO2 + H2O2 system showed high POD activity and low HORAC activity. The amount of ·OH in the environment was enough to meet the needs of sterilization.

Human gingival fibroblasts (hGFs) and human periodontal ligament stem cells (hPDLScs) were adopted to test the cytotoxicity of the nanozymes on normal human cells. The CCK-8 results showed that there was no significant difference in the relative activity of cells in each group with the increase of nanozyme concentration (P > 0.05, Supplementary Fig. 32a, b). Meanwhile, Live/Dead cell staining showed that there was no significant number of red stained dead cells in each group (Supplementary Fig. 32c). The maximum concentration of nanozymes used in cell safety test was 200 μg/mL, which was much bigger than the dose required for the antibacterial experiments, indicating that Cu-CeO2 nanozyme had good biosafety.

In vivo safety of Cu-CeO2 SSE

To test the in vivo safety of the nanozymes, PBS, CeO2, and Cu-CeO2 was intravenously injected into balb/c mice. The mice were observed for 3 days consecutively. As Supplementary Fig. 33 shows, there was no significant differences among the average body weight of the three groups. In addition, the blood routine test results at day 1 and day 3 after administration showed no statistical difference between each group in terms of the average white blood cell (WBC) and red blood cell (RBC) counts (P > 0.05, Supplementary Fig. 34). The hematoxylin-eosin (HE) staining results of the main organs on day 1 and day 3 indicate no significant histological differences between CeO2, Cu-CeO2 and PBS group (Supplementary Fig. 35). In summary, both CeO2 and Cu-CeO2 nanozymes showed good in vivo safety.

In vivo antibacterial performance of Cu-CeO2 SSE

According to previous literature, the pH value of normal intact skin tissue is weakly acidic (pH = 4–6), while in the case of wound infection, due to inflammatory stimulation, the local pH value tends to be weakly alkaline (pH≈7.4)59. It was confirmed through in vitro experiments that Cu-CeO2 catalytically decomposed H2O2 efficiently to generate ROS under pH = 7.4, achieving excellent antibacterial effects and good cyclic stability. These inspired us to further explore the therapeutic effect of Cu-CeO2 + H2O2 on in vivo infected skin wounds. For in vivo bactericidal experiment, we established a mouse skin wound infection model, and the schematic procedure is shown in Fig. 5a. A full-thickness skin defect with a diameter of ~6 mm was made on the back of BALB/c mice, and the wound was contaminated with MRSA. After 24 h of infection, the wounds were treated in six groups (PBS, CeO2, Cu-CeO2, H2O2, CeO2 + H2O2 and Cu-CeO2 + H2O2, n = 4), and the healing of skin wounds was observed daily. As Fig. 5b shows, 24 h after MRSA inoculation (Day 0), pyogenic infection occurred locally in the wounds of each group. After grouped treatments, the wounds in each group showed a trend of scab formation and contraction, and Cu-CeO2 + H2O2 group showed the fastest wound healing speed (Supplementary Fig. 36). On day 7, the wounds in Cu-CeO2 + H2O2 group were completely closed and the scabs fell off. The mice were sacrificed on day 7, and tissue sections of the skin around the wound were prepared. The results of HE staining showed that the epithelial structure of Cu-CeO2 + H2O2 group was approximately intact and continuous with no signs of infection, while the epithelium of the other groups was discontinuous. The epithelial defects showed varying degrees of deep dyeing inflammatory cells infiltration and extensive unstructured necrosis. Masson’s trichrome staining was used to observe the distribution of collagen fibers in skin tissues of each group. There were a large number of coarse, blue stained newly formed collagen fibers under the intact epithelium in Cu-CeO2 + H2O2 group, and the collagen fibers were uniformly distributed. In the other groups, however, the collagen fibers were scattered, inconsistent in thickness and disordered in structure. A certain amount of skin tissue around the wound of each group were taken to prepare homogenates, and the expression of three important pro-inflammatory factors, tumor necrosis factor α (TNF-α), interleukin-6 (IL-6) and interleukin 1β (IL-1β) was detected by enzyme-linked immunosorbent assay (ELISA) kit to reflect the severity of infection60. After H2O2 or CeO2 + H2O2 treatment, the level of inflammatory factors in skin tissue homogenates decreased compared with the control group, but the difference was not statistically significant (P > 0.05). However, the level of inflammatory factors in Cu-CeO2 + H2O2 group was significantly lower than the control group (P < 0.001) (Fig. 5c–e). Immunohistochemistry staining was further performed to evaluate the expression of TNF-α and IL-1β in the skin tissues, and the results are shown in Supplementary Figs. 3740, which were consistent with the ELISA results. The above results indicated that Cu-CeO2 + H2O2 system showed good antibacterial activity in vivo and effectively reduced the inflammatory response caused by wound infection.

Fig. 5: In vivo antibacterial properties of CeO2 and Cu-CeO2 nanozymes.
figure 5

a Schematic procedure of skin wound infection model and treatment. b Photographs, HE staining and Masson staining of the infected skin wounds after grouped treatment. Concentrations of TNF-α (c), IL-6 (d) and IL-1β (e) in skin tissue homogenates after grouped treatment. A representative image of three replicates from each group is shown. Data are presented as mean values +/− standard deviation, n = 3 biologically independent replicates. Significance was calculated by two-sided Student’s t-test. Scale bar: 500 μm. Source data are provided as a Source Data file.