Dynamic defects boost in-situ H₂O₂ piezocatalysis for water cleanup

Significance

Efficient in-situ H₂O₂ yield was attributed to the generation of dynamic defects promoting the adsorption and activation of oxygen, which were caused by cobalt deposition, enhancing piezoelectric activity during the US process and activating lattice oxygen with more piezoelectric electrons. More importantly, the inevitable issue of catalyst deactivation due to metal ion dissolution not only got resolved by pH adjustment, but even showed significant in situ H₂O₂ yield. Additionally, CZO piezocatalyst containing defects could be repeatedly utilized, maintaining high H₂O₂ production, and simultaneously achieving advanced degradation of pollutants due to their robust self-repair capabilities.

Abstract

Creating efficient catalysts for simultaneous H₂O₂ generation and pollutant degradation is vital. Piezocatalytic H₂O₂ synthesis offers a promising alternative to traditional methods but faces challenges like sacrificial reagents, harsh conditions, and low activity. In this study, we introduce a cobalt-loaded ZnO (CZO) piezocatalyst that efficiently generates H₂O₂ from H₂O and O₂ under ultrasonic (US) treatment in ambient aqueous conditions. The catalyst demonstrates exceptional performance with ~50.9% TOC removal of phenol and in situ generation of 1.3 mM H₂O₂, significantly outperforming pure ZnO. Notably, the CZO piezocatalyst maintains its H₂O₂ generation capability even after multiple cycles, showing continuous improvement (from 1.3 mM to 1.8 mM). This is attributed to the piezoelectric electrons promoting the generation of dynamic defects under US conditions, which in turn promotes the adsorption and activation of oxygen, thereby facilitating efficient H₂O₂ production, as confirmed by EPR spectrometry, XPS analysis, and DFT calculations. Moreover, the CZO piezocatalysts maintain outstanding performance in pollutant degradation and H₂O₂ production even after long periods of inactivity, and the deactivated catalyst due to metal ion dissolution could be rejuvenated by pH adjustment, offering a sustainable solution for wastewater purification.

Refractory pollutants pose a threat to human health and environmental ecosystems once discharged into natural water bodies (1–4). Conventional wastewater treatment methods, such as adsorption, can only transfer pollutants without complete degradation, while membrane processes suffer from high costs and membrane fouling. Currently, hydrogen peroxide (H₂O₂), an environmentally friendly green oxidant, releases highly reactive free radicals to remove pollutants (5). The anthraquinone oxidation (AO) method, which involved the Pd-catalyzed cyclic hydrogenation and oxidation of alkyl-anthraquinones in organic solvents, was commercially employed for H₂O₂ production (6, 7). However, this process was associated with significant energy consumption and environmental impacts and safety concerns regarding H₂O₂ transportation and storage. Therefore, developing a cost-effective and environmentally friendly in situ H₂O₂ production and activation cascade reaction is a sustainable strategy in the field of pollutant removal.

Piezoelectric catalysis has attracted widespread research interest in wastewater treatment due to its simplicity, cleanliness, and high efficiency (8, 9). It is considered a potential sustainable alternative to AO method for the production of H₂O₂. Similar to photocatalytic processes, under ultrasonic (US) activation, polarization, and built-in electric fields are generated in piezoelectric materials, allowing continuous separation of electrons and holes and attracting them to opposite surfaces (10). In aqueous solutions, surface charge carriers can react with water or dissolved oxygen to generate reactive oxygen species (ROS) for wastewater treatment, achieving the conversion of mechanical energy into chemical energy (O₂ + e⁻ → •O₂^–, •O₂^– + e⁻ + 2H⁺ → H₂O₂; 2H₂O + 2h⁺ → H₂O₂ + 2H⁺). Currently, varieties of piezoelectric semiconductors, e.g., BaTiO₃ (11), g-C₃N₄ (12), BiVO₄ (13), MoS₂ (8), ZnO (14), BiO_2-x (15), and CdS (16) have been widely developed. For example, Wei et al. developed an efficient piezocatalytic system of in situ piezo-generation and utilization of H₂O₂ over Fe/BiVO₄, significantly promoting the production of powerful oxidizing hydroxyl radicals (•OH) (17). Lin et al. revealed the piezoelectric effect and piezocatalytic mechanism of oxygen-vacancy-rich BiO_2−x NSs triggered by low frequency ultrasound (15). Particularly, ZnO, as a commonly used piezoelectric material, has found extensive applications in fields such as water splitting and water remediation, due to its excellent piezoelectric properties, availability, high chemical stability, and environmental friendliness (18). However, ensuring efficient piezoelectric catalytic performance (such as in situ piezo-generation and utilization of H₂O₂ and pollutant removal) in practical applications remains a challenge, primarily due to the rapid recombination of piezoelectric electron–hole pairs. In order to optimize the piezocatalytic performance of ZnO, researchers have explored various strategies, including metal or non-metal doping, defect engineering, and heterostructure construction. For instance, Wang and Zheng groups have both chosen the approach of doping metal ions to enhance the piezoelectric effect (19, 20). Wang et al. demonstrated that V-doped NaNbO₃ single crystals improved the catalytic efficiency via enriched active sites and enhanced polarization field under US vibration (21). Additionally, metal deposition has been widely applied as an effective method to enhance carrier separation efficiency. For example, Huang et al. developed Ag single atoms (SAs) and clusters co-anchored on carbon nitride (Ag_SA+C-CN) to serve as the multifunctional sites for efficient two-electron water splitting (22). In addition to enhancing carrier separation efficiency, increasing oxygen adsorption to accelerate the oxygen reduction reaction (ORR) process is also a key factor in improving H₂O₂ yield. Therefore, introducing defect engineering based on the aforementioned regulation strategies, with a focus on adjusting the electronic structure and enriching active/adsorption sites, becomes crucial. So far, there have been relatively few studies on ZnO with dynamic oxygen defects in piezocatalysis. Consequently, the combined use of metal deposition and defect engineering can significantly promote H₂O₂ production, while activating H₂O₂ to generate ROS for enhanced pollutant removal performance. This approach holds great research significance in emerging fields.

Here, we successfully prepared a Co loaded-ZnO (CZO) piezocatalyst by combining a straightforward hydrothermal process with a photo-deposition approach. This catalyst exhibited excellent capability for refractory pollutants removal and notable non-sacrificial H₂O₂ production (pure water). After 3 h of US treatment, the removal rate of phenol reached 99% and H₂O₂ production yield reached 1.3 mM. Experiments and characterization technique were both employed to explore the mechanism. It was found that CZO piezocatalyst could be used as an excellent heterogeneous catalyst for H₂O₂ generation and activation attributed to the dynamic defects in the process of US. Moreover, the generation of dynamic defects contributed to the stability and long-term degradation performance of the catalyst. Thus, a piezocatalysis-self-Fenton system has been successfully constructed over CZO piezocatalyst. Compared to pristine ZnO, the CZO piezocatalyst significantly enhanced the H₂O₂ production and the purification efficiency for refractory pollutants. This study provides insights into the design of piezocatalysts and their applications in energy conversion and environmental remediation.

Results

Characterizations of the Fabricated Catalysts.

A straightforward one-step hydrothermal method was used to prepare ZnO microrods (ZO), followed by photo-deposition to synthesize Co-ZnO-X composites (CZO-X, X = 0, 0.5, 2, 4). The scanning electron microscopy (SEM) and transmission electron microscopy (TEM) images of ZO and CZO-X presented similar microrod structure with an average length exceeding 2 μm, as shown in SI Appendix, Figs. S1 and S2 and Fig. 1 A–I. The magnified TEM image of the circle highlighted in Fig. 1A and the corresponding high-resolution TEM (HRTEM) image of ZO (Fig. 1 B and C) displayed ordered lattice patterns with a spacing of 0.26 nm (002 lattice plane), as determined by line profile measurements. It could be observed from the TEM image in Fig. 1D that numerous black dots were embedded in different domains of ZO. The corresponding energy dispersive spectroscopy (EDS) elemental mapping and EDS spectra of CZO-2 exhibited a uniform distribution of Co, Zn, and O (Fig. 1 E and F). Additionally, many small black spots were also observed on other microrods of CZO-2 (Fig. 1G). The average diameter of these black dots was measured to be only ~1.24 nm (Fig. 1H), demonstrating that the well-confined Co species mainly existed in the form of ultrafine clusters. The actual powder state of these catalysts was depicted in SI Appendix, Fig. S3, with the color gradually changing from white to yellow. Furthermore, the results obtained from inductively coupled plasma atomic emission spectrometry (ICP-AES) analysis, as shown in SI Appendix, Table S1, revealed an increase in the loaded content of Co from 0.51 to 0.89% with increasing dosage of cobalt nitrate. This finding provided additional confirmation of the successful loading of Co. The crystalline phases of ZO and CZO-X composite materials were investigated using X-ray diffraction (XRD) pattern. In Fig. 1J, ZO exhibited three strong diffraction peaks at 31.8°, 34.5°, and 36.3°, which correspond to the (100), (002), and (101) crystal planes of the wurtzite hexagonal ZnO (PDF#75-0576), while the XRD pattern of CZO-X was nearly identical to that of ZO, indicating that the deposition of Co clusters did not alter the crystal structure of ZO. It is worth noting that no crystalline phase of cobalt was detected, likely due to the small quantity and size of the Co clusters (23). Additionally, the XRD pattern displayed weak diffraction peaks at 33° and 60°, which were ascribed to the (213) and (503) crystal planes of Zn(OH)₂ (PDF#38-0356). This attribution stemmed from the fact that zinc salts inevitably produced a small amount of Zn(OH)₂ during the hydrothermal synthesis of ZnO. Furthermore, it is noteworthy that varying material proportions exhibited similar hydroxide profiles in the XRD pattern. Consequently, the predominant composition of the catalyst was oxide crystals, allowing for the disregarding of the impact of this minimal quantity of hydroxides. Raman and Fourier transform infrared (FT-IR) spectroscopies were performed to further characterize the structure of ZO and CZO-2, as shown in SI Appendix, Fig. S4. To investigate the chemical state, we examined the X-ray photoelectron spectra (XPS), where Fig. 1K exhibited the Co 2p XPS spectrum of the CZO-2 sample. The peaks centered at 780.3 and 795.3 eV were assigned to Co²⁺ of Co 2p_3/2 and 2p_1/2 (24), respectively, while faint peaks observed at binding energies of 782.1 and 797.4 eV belonged to Co³⁺ of Co 2p_3/2 and 2p_1/2 of Co-O, respectively. These results collectively demonstrated the successful introduction of Co and diversity of Co valence.

Fig. 1.

(A) TEM image and (B) the corresponding locally enlarged image of ZO; (C) HRTEM image of ZO and the d-spacing value in the image was measured from line profile on the Top Right; (D) TEM image, (E) EDS elemental mapping, and (F) EDS spectra of CZO-2; (G) TEM image and (H) the corresponding locally enlarged image of CZO-2; (I) HRTEM image of CZO-2 and the d-spacing value in image was measured from line profile on the Top Right; (J) XRD spectra of ZO and CZO-X; (K) Co 2p XPS spectrum of CZO-2.

Piezoelectricity and Electrochemical Analysis of the Catalysts.

Piezo-response force microscopy (PFM) was employed to characterize the piezoelectricity of ZO and CZO-2 directly. The application of an AC voltage spanning from −10 to 10 V elicited a localized hysteresis loop (Fig. 2 A and B), where the applied external electric field caused continuous deformation of its surface, accompanied by a significant phase angle shift of 180°. This unequivocally confirmed the inherent piezoelectricity of ZO and CZO-2 material (25). Furthermore, the maximum effective piezoelectric coefficient (d₃₃) value based on the amplitude loops for ZO and CZO-2 were 41.2 and 58.4 pm/V, respectively, which clarified the piezoelectric response of ZO microrods was enhanced by Co-deposition (26). Detailed examinations of the PFM morphology, 3D morphology, piezo-response amplitude, and phase images were presented in SI Appendix, Figs. S5 A–D and S6 A–D. It can be observed that there was a clear contrast in both the amplitude and phase images of ZO and CZO-2, demonstrating the piezoelectric properties of the materials. Theoretically, the cyclic alteration of the electric field direction in piezoelectric materials resulting from mechanical vibrations led to a rapid recombination of piezoelectric charges, exerting a detrimental impact on the efficiency of piezocatalysis. Consequently, it became imperative to enhance both the extended lifetime of piezoelectric charges and the prompt charge separation in order to optimize the performance of piezoelectric-driven redox reactions. We observed from SI Appendix, Fig. S7 and Table S2 that moderate Co loading significantly extended the charge carrier lifetime as determined through time-resolved photoluminescence (PL) attenuation measurements, with detailed information provided in SI Appendix, Fig. S7. Subsequently, the piezo-generated electrons and holes separation efficiency of the samples were investigated using electrochemical impedance spectroscopy (EIS), as shown in Fig. 2C. The smaller arc radius in the Nyquist plots of EIS for CZO-X, compared to ZO, reflected lower resistance, indicating enhanced charge separation efficiency and improved charge mobility in the Co and ZO phase. Consequently, this enhanced charge separation efficiency was found to be highly beneficial for catalytic reactions, similar to the findings from PL attenuation measurements. Additionally, the charge densities (N_d) were calculated by Mott–Schottky measurements from the following formula:

$${\mathrm{N}}_{\mathrm{d}}=\left(2/{\mathrm{e}}_0{\epsilon }_0\epsilon \right){\left[\mathrm{d}\left(1/{\mathrm{C}}^2\right)/\text{dV}\right]}^{-1},$$ [1]

Fig. 2.

(A) The butterfly amplitude loop and phase curve of ZO and (B) CZO-2; (C) Electrochemical impedance spectra (EIS) of the samples; (D) Mott–Schottky plots of the samples at a frequency of 1 kHz (the Inset shows the electron densities of the samples).

where e₀ represents the electron charge and ε₀ and ε are the vacuum permittivity and dielectric constant of ZnO, respectively (27, 28). As depicted in Fig. 2D, the Mott–Schottky plots of ZO exhibited a steeper slope compared to CZO-2 at a frequency of 1 kHz, indicating that the corresponding N_d of ZO (0.10 × 10²² cm⁻³) was smaller than that of CZO-2 (0.17 × 10²² cm⁻³). This finding suggested a faster charge transformation process in CZO-2, aligning with the above-mentioned results. Furthermore, a comprehensive analysis of the plots enabled the determination of the conduction band (CB) positions of the materials under investigation. The flat-band potentials of ZO and CZO-2 were derived as approximately −0.93 V and −0.61 V, respectively (vs. saturated calomel electrode, SCE). Taking into account that the flat-band potential typically exceeded the conduction band (CB) potential by 0.1 to 0.3 V in n-type semiconductors (29), the estimated E_CB values for ZO and CZO-2 were −0.79 V and −0.47 V, respectively (vs. normal hydrogen electrode, NHE).

Piezocatalytic Performance Measurement.

Upon ultrasonic (US) activation, CZO-2 demonstrated exceptional performance without additional oxidants, achieving an impressive phenol oxidation rate of 98.7%, surpassing the capabilities of ZO, which only achieved a degradation rate of 9.4% (Fig. 3A). Controlled experiments further highlighted the significance of both US vibration and the catalyst. Without US, CZO-2 showed a mere adsorption rate of 3.3% for phenol, while the degradation rate of phenol without a catalyst under US treatment was 9.7%. These findings underscored the crucial role played by the catalyst and US activation in facilitating the highly efficient degradation of phenol. As shown in Fig. 3B, the kinetic constants (k_obs) of phenol oxidation by the CZO-2 were maximized (0.024 min⁻¹) compared to other materials. Therefore, CZO-2 was selected as the optimal catalyst for subsequent trials. Subsequent research assessed the influence of US power and catalyst dosage on phenol removal and investigated the removal efficiency at different phenol concentrations, with detailed information provided in SI Appendix, Fig. S8. As widely recognized, solution pH was also a major factor affecting the degradation of pollutants. Within the initial pH range of 3 to 9, the degradation rate of phenol was approximately 98.0%, with little variation in the k_obs value, initially increasing and then decreasing (Fig. 3C). These results indicated that the CZO-2 piezoelectric system effectively degrades phenol over a wide pH range. Besides, tap water, Yangtze River, Huangpu River, and Pearl River had no significant effect on phenol degradation (Fig. 3D). More importantly, the CZO-2 material maintained a remarkable 95.9% removal efficiency of phenol after four cycles (Fig. 3E), highlighting its sustained high catalytic activity even after repeated reactions. To demonstrate the versatility of the CZO-2/US system, we also selected refractory organic pollutants other than phenol as target pollutants, including rhodamine B (RhB), acid orange 7 (AO7), methylene blue (MB), nitrobenzene (NB), aniline (AB), tetrachlorophenol (4-CP), sulfadiazine (SD), and quinoline. It was observed that CZO-2 exhibited not only efficient degradation of phenol but also effective degradation of RhB, AO7, MB, NB, AB, 4-CP, SD, and quinoline (Fig. 3F and SI Appendix, Table S3), highlighting its wide-ranging applicability in wastewater treatment.

Fig. 3.

(A) Phenol removal efficiency in different systems, reaction conditions: [catalyst] = 0.6 g L^–1, [phenol]₀ = 10 mg/L, and 3 h of US vibration; (B) the corresponding degradation rate constant of phenol (k_obs); (C) effect of initial pH on phenol removal in the CZO-2 system; (D) the degradation of phenol in real waters; (E) cycling runs for degradation of phenol in the CZO-2/US system; (F) degradation of different refractory pollutants by CZO-2 within 3 h of US.

Activation of O₂ and In-Situ H₂O₂ Generation by CZO-2 Piezocatalyst.

As shown in Fig. 4 A–D and SI Appendix, Fig. S9, the radical quenching experiments involving different scavengers and saturated atmosphere (Ar) were performed to explore the activation capacity of O₂ and the mechanism of H₂O₂ production in the CZO-2/US system. When IPA (as a scavenger of •OH, k_{•OH, IPA} = 1.9 × 10⁹ M⁻¹ s⁻¹) (30) was added to the CZO-2 system (Fig. 4A), the degradation efficiency of phenol declined from 98.7% to 45.1%, which suggested that •OH played a crucial role in the oxidation process. With the addition of AgNO₃ (as a scavenger of e⁻) and (NH₄)₂C₂O₄ (as a scavenger of h⁺), there was a slight inhibition effect on the oxidation of phenol, whereas significant inhibition of phenol degradation was observed in the presence of p-BQ (as a scavenger of •O₂⁻, k _{•O2^–, p-BQ} = 0.9 × 10⁹ M⁻¹ s⁻¹) (31, 32) or FFA (as a scavenger of ¹O₂, k¹_O2, FFA = 1.2 × 10⁸ M⁻¹ s⁻¹) (33). Furthermore, phenol oxidation reaction was further inhibited by increasing the concentrations of IPA and FFA (SI Appendix, Fig. S9A). As shown in SI Appendix, Fig. S9B, a substantial inhibition was observed under Ar-saturated conditions. These results verified the significance of O₂ activation and the dominant role played by •OH, •O₂⁻ and ¹O₂ in phenol oxidation. To clarify the significant impact of ¹O₂ on phenol degradation, we referred to our prior research showing the synergistic effect of ¹O₂ and •OH in facilitating the ring-opening reaction of phenol (34). In our further investigation, it was observed whether hydroxylation products were present during the phenol degradation process in the presence of ¹O₂ scavenger. To avoid interference from the position of the peaks in the high-performance liquid chromatography (HPLC), FFA was replaced with TEMP as ¹O₂ scavenger. As shown in SI Appendix, Fig. S10 A and B, the addition of TEMP resulted in a decrease in phenol degradation efficiency from 98.7 to 11.4%. Moreover, the liquid chromatography spectrum revealed no significant hydroxylation products. Conversely, in the CZO-2/US system, the accumulation of intermediate products occurred rapidly, and as the US treatment progressed, phenol gradually underwent ring opening (SI Appendix, Fig. S10C). This was confirmed by the detection of small molecules such as fumaric acid through HPLC (SI Appendix, Fig. S10D) and the high rate of total organic carbon (TOC) removal, reaching 50.9% (SI Appendix, Fig. S11). Furthermore, the HPLC–MS results revealed the presence of different intermediates and completely decomposed into CO₂ and H₂O in the end (detailed in SI Appendix, Fig. S12). Therefore, the presence of ¹O₂ in this system served to promote the hydroxylation reaction of phenol, which was further attacked by •OH, leading to ring-opening. Meanwhile, there was a slight decrease in the peak intensity of phenol, accompanied by the accumulation of a small amount of intermediate products in the ZO/US system, which was consistent with the relatively low TOC removal rate of only 3.7% (SI Appendix, Figs. S11 and S13).

Fig. 4.

(A) Radical quenching experiments of the oxidation of phenol and (B) concentration changes of in situ piezo-generated H₂O₂ with the additions of different scavengers and different atmospheres (Ar) in the CZO-2/US system; (C) energy band diagram of ZO and CZO-2; (D) concentration changes of •OH with the additions of different scavengers and different atmospheres (Ar) in the CZO-2/US system; (E) under air and Ar atmosphere, EPR signals for DMPO-•OH (water), DMPO -•O₂⁻ (MeOH), TEMP-¹O₂ over CZO-2 powders under US irradiation; (F) in situ H₂O₂ production comparison of CZO-2 with some typical reported piezocatalysts and photocatalysts.

Subsequently, the H₂O₂ concentration in the CZO-2/US system was also detected in pure water by the iodide method using ammonium molybdate. As shown in SI Appendix, Fig. S14, the concentration of H₂O₂ associated with CZO-2 reached 1.3 mM under US irradiation, nearly 6.5-fold higher than that of ZO. To eliminate the influence of cobalt ions, cobalt ions were introduced into the ZO system. Interestingly, it was observed from SI Appendix, Fig. S15 that the concentration of H₂O₂ did not increase significantly, indicating that the enhanced H₂O₂ concentration in the CZO-2 system was primarily attributed to the faster charge transfer rate and longer charge lifetime resulting from Co-deposition. Meanwhile, the concentration of H₂O₂ of the CZO-2/US system in seawater was also detected. As shown in SI Appendix, Fig. S16, the results revealed that the H₂O₂ yield in seawater was slightly lower compared to that in pure water, but still reached 0.98 mM. The cumulative •O₂⁻ and •OH production were also quantitatively monitored. Initially, nitrotetrazolium blue chloride (NBT) was employed to measure the concentration of •O₂⁻ generated by ZO and CZO-2 under US irradiation. As depicted in SI Appendix, Fig. S17, the NBT concentration gradually decreased in both ZO and CZO-2 samples, indicating the activation of O₂ to produce •O₂⁻. By utilizing the stoichiometric parameter that 1 mole of NBT consumption corresponded to 4 moles of •O₂⁻, the cumulative concentrations of •O₂⁻ in ZO and CZO-2 were determined to be 0.019 and 0.037 mM, respectively. Additionally, the terephthalic acid photoluminescence (TA-PL) method was used to record the •OH concentration (SI Appendix, Fig. S18). Similarly, the piezocatalytic production of •OH over CZO-2 was largely enhanced compared with ZO, from 0.016 to 0.034 mM.

According to the above analysis, the pathway of CZO-2 generating H₂O₂ was also investigated. As shown in Fig. 4B, the in situ generated H₂O₂ of CZO-2 with different scavengers and Ar atmosphere was detected in pure water under US irradiation. The concentrations of H₂O₂ considerably increased from 1.3 mM (control) to 1.7 mM (IPA), 1.7 mM (FFA), and 1.6 mM [(NH₄)₂C₂O₄]. However, the concentrations of H₂O₂ considerably decreased to 0.05 mM (AgNO₃), 0.02 mM (p-BQ), and 0.05 mM (Ar-saturated condition). These results indicated that H₂O₂ formation involved a two-electron process, and •O₂⁻ was an important intermediate substance for producing H₂O₂ (O₂→•O₂⁻→H₂O₂). Additionally, by utilizing UV-Vis diffuse reflectance spectroscopy (DRS) estimation in SI Appendix, Fig. S19, the optical bandgaps for ZO and CZO-2 were determined to be 3.09 eV and 2.93 eV, respectively, and when combined with the analysis of Mott–Schottky measurement, the corresponding energy band diagrams were illustrated in Fig. 4C. The suitable CB and valence band (VB) position were favorable for the oxygen reduction reaction (ORR), thereby initiating the cascade redox reaction. Specifically, piezoelectric electrons reacted with absorbed O₂ to produce •O₂⁻, and the intermediate •O₂⁻ reacted with piezoelectric electrons to generate H₂O₂, which could be further decomposed to •OH. The corresponding changes in •OH radical were in accordance with the abovementioned results (Fig. 4D). Furthermore, it was observed from SI Appendix, Fig. S20 that while the CZO-2 catalyst could decompose H₂O₂ under stirring conditions, it was unable to generate •OH. Moreover, in the absence of a catalyst, neither US nor stirring could decompose H₂O₂, indicating that both US and catalyst were indispensable for the generation of •OH in this work.

Additionally, to assess whether Co²⁺ was involved in the activation process of H₂O₂, we introduced Fe²⁺ into the CZO-2/US system, competing with Co²⁺ in decomposing H₂O₂. Due to the faster rate of Fe²⁺ in decomposing H₂O₂ compared to Co²⁺, the introduction of Fe²⁺ served as a means to explore whether Co²⁺ was involved in the activation process of H₂O₂ during the US process. If the change in the contents of cobalt ion was more obvious in the absence of Fe²⁺ compared to when Fe²⁺ was added, it confirmed the participation of Co²⁺ in the activation process of H₂O₂ in the CZO-2/US system. Thus, from the Co 2p XPS data, it was observed that, compared to the pure CZO-2 catalyst, the post-reaction Co²⁺ content decreased from 50.2 to 39.2%, while the Co³⁺ content increased from 49.8 to 60.8% (SI Appendix, Fig. S21A and Table S4). However, when we added Fe²⁺ into the CZO-2/US system (SI Appendix, Fig. S21B and Table S4), it was found that, in comparison with the pure CZO-2 catalyst, the alterations in the contents of Co²⁺(49.9%) and Co³⁺(50.1%) after the reaction were not more obvious than that without adding Fe²⁺, demonstrating the involvement of Co²⁺ in the activation of H₂O₂. These results suggested that •OH derived from H₂O₂ activation (primary) and H₂O oxidation (secondary) through the following route: O₂ → •O₂⁻ → H₂O₂ → •OH and H₂O → •OH.

In addition to chemical probes, the main active species that contributed to the excellent performance of the phenol oxidation was also confirmed by the EPR characterization, as shown in Fig. 4E. Under air atmosphere, CZO-2 piezocatalysis yielded EPR quadruplet DMPO-•OH peaks, sextuplet DMPO-•O₂⁻ peaks, and triplet TEMP-¹O₂ peaks with US 10 min, whereas no clear signal of ROS was observed without US irradiation. Under Ar atmosphere, weak •OH signal was observed, consistent with the previous findings, and the signal of •O₂⁻ and ¹O₂ were negligible. Additionally, upon the addition of p-BQ or phenol in the presence of US irradiation, the peak intensity of ¹O₂ was weakened, suggesting both the evolution of ¹O₂ from •O₂⁻ and its direct interaction with phenol molecules. Based on these results, we proposed a mechanism for ROS generation via ultrasound-driven piezocatalysis (Eqs. 2–6). Additionally, among various piezocatalysts and photocatalysts studied, the CZO-2/US system exhibited the highest production of H₂O₂, as illustrated in Fig. 4F and SI Appendix, Table S5.

$$\text{CZO}-2\stackrel{\text{Vibration}}{⟶}{\mathrm{e}}^{-}+{\mathrm{h}}^{+},$$ [2]

$${\mathrm{O}}_2+{\mathrm{e}}^{-}\to \bullet {{\mathrm{O}}_2}^{-},$$ [3]

$$\begin{array}{c}2\bullet {{\mathrm{O}}_2}^{-} + 2{\mathrm{H}}_2\mathrm{O} \to 2{\mathrm{H}}_2{\mathrm{O}}_2 + 2{\text{OH}}^{-} {+}^1{\mathrm{O}}_2\hfill \end{array}$$ [4]

$${\mathrm{H}}_2{\mathrm{O}}_2+{\text{Co}}^{2+}\to {\text{Co}}^{3+}+\bullet \text{OH}+{\text{OH}}^{-},$$ [5]

$${\mathrm{H}}_2\mathrm{O}+{\mathrm{h}}^{+}\to \bullet \text{OH}+{\mathrm{H}}^{+}.$$ [6]

Investigation on the Mechanism of Enhanced H₂O₂ Concentration.

It was noteworthy that the concentration of H₂O₂ generated during the US exposure was re-evaluated using the CZO-2 sample after four cycles. Interestingly, there was a noticeable increase in the H₂O₂ concentration, as depicted in Fig. 5A, with concentration rising from 1.3 mM to 1.8 mM. To investigate this phenomenon, we initially examined the surface vacancies of different catalysts using EPR characterization. As shown in Fig. 5B, ZO and CZO-2 exhibited minimal vacancies signal prior to the US reaction. However, following the US reaction, ZO displayed a weak oxygen defects signal peak at g = 2.004. In contrast, CZO-2 exhibited a significantly enhanced oxygen defects signal. Apart from the EPR technique, we employed X-ray photoelectron spectroscopy (XPS) to examine the alteration in the surface defect ratio of the material before and after the reaction. Fig. 5C and SI Appendix, Table S6 revealed the O 1 s XPS spectrum, which exhibited a splitting into two or three peaks: lattice oxygen was bonded to metal (O_latt, 529.5 to 530.6 eV), oxygen deficiency (O_def, 531.0 to 531.8 eV), and the surface adsorbed molecule H₂O (O_surf, 532. 8 eV) (35). The relative ratio of oxygen deficiency in ZO exhibited no significant change before [O_def/(O_def + O_latt) = 4.4%] and after reaction [O_def/(O_def + O_latt) = 3.9%], which was consistent with the previous results. Conversely, the relative ratio of O_def of CZO-2 remarkably increased from 6.4 to 29.1% after reaction, and the surface oxygen content dwindled from 62.4 to 51.0%, providing strong evidence for the occurrence of dynamic oxygen vacancies during the CZO-2 US stress. Thus, we speculated that the formation of dynamic oxygen defects of CZO-2 material during the US process might have been either due to the deposition of cobalt, which enhanced piezoelectric activity, activating lattice oxygen by more piezoelectric electrons during migration, ultimately leading to the formation of oxygen defects, or due to the catalyst generating additional heat on the surface under US conditions, promoting oxygen overflow. Therefore, we conducted an experiment in which we subjected the CZO-2 catalyst to a 3-h heat treatment at a temperature equivalent to that generated under ultrasound. The results from Fig. 5D showed that no defect signal was observed in the CZO-2 after the heat treatment. Following this, we introduced AgNO₃ into the reaction process and observed a substantial decrease in the defect signal by examining the EPR signal of the catalyst after reaction. This finding served as strong evidence for the pivotal role of piezoelectric electrons in the formation of dynamic defects. Furthermore, the formation energy of a single oxygen defect over CZO was studied by DFT calculations, as shown in Fig. 5E. Results indicated that oxygen vacancy on surface of CZO (E_OV = 3.49 eV) was obviously lower than that of ZO (E_OV = 3.94 eV), suggesting that the oxygen vacancies more easily formed on the surface of CZO due to cobalt deposition. Additionally, by introducing additional electrons into the CZO-OV model to bolster the reducing atmosphere, a noteworthy reduction in defect formation energy was observed, decreasing from 3.49 eV to 2.25 eV. This compelling finding unequivocally established the intricate connection between defect formation and electron density. Thus, these results indicated that the formation of dynamic defects was not caused by the thermal effects during the US process but rather by the activation of lattice oxygen through piezoelectric electrons. Additionally, we could further observe from Fig. 5B that after undergoing four cycles, the catalyst exhibited a stronger defect signal. Simultaneously, the in situ generated H₂O₂ concentration also increased further, indicating a positive correlation between the amount of H₂O₂ and the defect concentration. To investigate how the catalyst maintained excellent piezoelectric activity under conditions of continuous defect generation, the post-reaction catalyst was stored in ambient air for a period of time. It was found from Fig. 5D that the defect signal of first-CZO-2-used-storage noticeably weakened. This indicated that the CZO-2 catalyst possessed self-repair capabilities, allowing it to maintain outstanding stability during multiple cyclic tests. Additionally, since the production of H₂O₂ was derived from the ability of the catalyst to adsorb and activate dissolved oxygen, we initiated the process by increasing aeration, thereby maximizing the concentration of dissolved oxygen. The results revealed that, after 1 h of aeration, the H₂O₂ concentration increased from 1.3 mM to 1.6 mM (SI Appendix, Fig. S22). Subsequently, to gain a deeper understanding of the interaction between O₂ molecules and piezocatalysts with different concentrations of oxygen defects at a molecular level, DFT simulation has been employed (36, 37). Based on experimental characterization, we theoretically modeled the optimal geometric structure of CZO (002) and confirmed the optimal adsorption site of the O₂ molecules (SI Appendix, Fig. S23). As shown in Fig. 5F, CZO samples with varying oxygen vacancies (from OV0 to OV3) exhibited strong O₂ molecule adsorption with a negative adsorption energy ranging from approximately −3.05 eV to −3.24 eV, accompanied by a charge transfer ranging from 0.645 e⁻ to 0.657 e⁻, indicating that the more defect sites exposed on the catalyst surface, more electron transfer occurs from ZO-OV to Co clusters, leading to electron-enriched Co clusters and thereby promoting the adsorption and activation of the O₂ molecule. Meanwhile, SI Appendix, Figs. S21 and S24 revealed the electron cloud density of Co increased, while that of Zn decreased in CZO-2 after the reaction (detailed information was shown in SI Appendix, Fig. S24 and Tables S4 and S7), indicating that the large number of defects as electron traps produced by CZO-2 in the piezoelectric reaction process, effectively impeding the recombination of both piezoelectric holes and electrons and promoting the charge transfer of ZO to Co clusters by the Zn-O-Co electron transfer chain, thus obtaining electron rich structure. Thus, we speculated that Co clusters acted as active sites for electron enrichment, facilitating the activation of O₂ to produce H₂O₂ and then be further activated to •OH. Besides, we introduced iron into ZnO (Fe-ZnO) using the same photo-deposition method, and its XRD spectrum closely resembled that of CZO-2 (SI Appendix, Fig. S25A), displaying prominent characteristic peaks of ZnO. Moreover, Fe-ZnO exhibited excellent piezoelectric catalytic performance, with a phenol degradation rate of 87% and a H₂O₂ production rate of 1.46 mM after US treatment for 3 h (SI Appendix, Fig. S25B). To exclude the potential interference of Fe³⁺ produced during the US process to the detection of H₂O₂ by iodometry, we mixed the reaction solution obtained after 3 h of ultrasound with an equal amount of potassium thiocyanate (10 mg mL⁻¹). The UV-visible spectrum analysis of Fe³⁺ content in the solution, as shown in SI Appendix, Fig. S25C, revealed the absence of Fe³⁺. Simultaneously, the leaching of 0.068 mg L⁻¹ Fe from the Fe-ZnO solution after 3 h of US was analyzed using ICP-AES. Assuming this was equivalent to the concentration of Fe³⁺, it was observed that the reaction of Fe³⁺ with potassium iodide at this concentration did not exhibit a peak at 351 nm (SI Appendix, Fig. S25D). Consequently, in the Fe-ZnO/US system, the iodometric detection of H₂O₂ concentration was deemed reasonable. As a result, similar results could be achieved by substituting cobalt with alternative metals through the application of the same photo-deposition method, thereby overcoming the inherent constraints associated with this approach.

Fig. 5.

(A) Time dependence for yield of CZO-2 and fourth-CZO-2 under US irradiation; (B) EPR spectra of the samples; (C) O 1 s XPS spectrum of ZO and CZO-2 fresh and after the reaction; (D) EPR spectra of the samples; (E) formation energy of oxygen vacancy over ZO-OV, CZO-OV, and CZO-OV-adding electrons; (F) the charge density difference of the O₂ molecule on CZO with different oxygen vacancy concentrations (the yellow electronic cloud represents the loss of electrons, and the blue electronic cloud represents the gain of electrons; the isosurface is set to 0.0005 eV).

In order to study the effect of background material and potential application of CZO-2 material, the impact of typical anions (Cl⁻, NO₃⁻, HCO₃⁻, and SO₄²⁻) and dissolved organic matter (DOM) on phenol degradation was investigated separately. As depicted in Fig. 6A, in the presence of Cl⁻, NO₃⁻, HCO₃⁻, SO₄²⁻, and DOM, the oxidation efficiencies of phenol were 88.6%, 97.5%, 97.0%, 91.5%, and 91.4%, respectively (in 180 min). Although the degradation efficiency was not significantly affected, the degradation rate exhibited sensitivity, particularly in the presence of Cl⁻ ions. This was because Cl⁻ could induce a competing reaction with phenol oxidation by utilizing •OH. Material stability was also an important indicator for evaluating potential applications of catalysts, we performed comprehensive investigations, including characterizing the catalyst after reaction, assessing the degradation efficiency of catalysts stored for varying time durations, evaluating their capacity for in situ H₂O₂ generation, and examining their long-term pollutant degradation capabilities. First, the SEM images (SI Appendix, Fig. S26 A and B), XRD pattern (SI Appendix, Fig. S26C), and FT-IR spectra (SI Appendix, Fig. S26D) of CZO-2 after reaction were nearly identical to those of the initial sample, indicating that CZO-2 demonstrated robust structural stability following the reaction, attributable to its remarkable self-healing capabilities. Next, the phenol degradation performance and H₂O₂ generation ability of the catalyst were tested after it had been placed for 2 mo and 5 mo. As depicted in Fig. 6B, it was observed that even after 3 h of US irradiation, phenol degradation was nearly complete. Moreover, the concentration of H₂O₂ remained at 1.2 mM and 1.0 mM in pure water, respectively, indicating that the capability of the catalyst to degrade phenol and produce H₂O₂ did not significantly decline over the extended period. Additionally, based on the ion dissolution results obtained from ICP-AES analysis (Fig. 6C), it could be observed that, with prolonged US time in the CZO-2/US system, the concentrations of cobalt ions and zinc ions increased from 0.58 mg/L to 0.75 mg/L and from 4.2 mg/L to 7.4 mg/L, respectively. Subsequently, after the reaction solution underwent treatment with a cation exchange resin, the concentrations of cobalt ions and zinc ions significantly decreased to 0.085 mg/L and 0.06 mg/L, respectively, which was far lower than the emission standard and effectively prevented secondary pollution. Simultaneously, the pH of the reaction solution after 3 h of ultrasound was adjusted to 10 with NaOH. Following a subsequent aging period, the dissolution concentrations of cobalt and zinc ions decreased significantly, from 0.75 mg/L to less than 0.01 mg/L and 7.4 mg/L to 0.019 mg/L, respectively. The dissolved metal ions formed hydroxide precipitates, allowing for their effective reuse. Consequently, it was observed that H₂O₂ yield increased with prolonged US exposure in pure water within the CZO-2/US system (Fig. 6D). However, after further extension of the time, the H₂O₂ production rate eventually fell below the H₂O₂ decomposition rate, resulting in a gradual decrease of H₂O₂ yield. Nevertheless, adjusting the pH of the above solution to 10 and allowing for an aging period led to an increase in H₂O₂ yield again, which demonstrated that the deactivated catalyst due to metal ion dissolution could be rejuvenated by pH adjustment, ensuring the long-term stability for in situ H₂O₂ production of the CZO-2/US system. Furthermore, long-term continuous degradation of phenol was examined under US exposure (Fig. 6E). The performance was evaluated after 60 h of US treatment with eight rounds of phenol supplementation. Although the degradation rate gradually slowed down, CZO-2 maintained a high phenol degradation efficiency of 92%, showcasing its remarkable stability.

Fig. 6.

(A) Removal efficiencies and the corresponding k_obs value of phenol oxidation (US for 3 h) in the presence of different water quality constituents [(anion) = 10 mg L⁻¹, (DOM) = 10 mg L⁻¹]; (B) under US irradiation, time dependence for phenol removal efficiency and H₂O₂ yield of CZO-2 after 2 and 5 mo of placement; (C) variation of ion concentration in the CZO-2/US (0.6 g L^–1) system (treatment-1: cation-exchange resin was used for adsorption of dissolved metal ions; treatment-2: adjusting pH = 10 with NaOH); (D) time dependence for H₂O₂ yield of CZO-2. (E) Time dependence for phenol removal efficiency of CZO-2.

Conclusions

Piezocatalysis proved to be an exceptionally attractive and effective technology in the field of wastewater treatment. However, the major obstacles encountered in the development of piezocatalytic reaction primarily revolved around the dual aspects of H₂O₂ generation and activation without introducing additional oxidants, which were indispensable for producing sustainable ROS capable of achieving thorough purification of wastewater. In this study, the CZO piezocatalysts prepared by the hydrothermal method combined with the photo-deposition method demonstrate exceptional efficacy in wastewater treatment and in situ generated H₂O₂, attributed to fast charge separation efficiency and extended charge lifetime brought by Co-deposition and the existence of dynamic defects, as confirmed by EPR and XPS analysis. More importantly, the catalysts placed for longer periods (over 2 mo) continue to perform superior H₂O₂ production (~1.2 mM) and contaminant removal capabilities. This observation not only underscores the substantial potential of CZO piezocatalysts in practical wastewater treatment applications but also serves as a reference for future studies on improving H₂O₂ production by facilitating the generation of dynamic vacancies through metal deposition.

Materials and Methods

Preparation of Catalysts.

The ZnO microrods were synthesized using a hydrothermal method. Specifically, a mixture of 10 mmol Zn(NO₃)₂·6H₂O, 10 mmol urotropin, and 48 mL of ultrapure water was stirred until homogenous. The resulting solution was transferred to a 100-mL hydrothermal reactor and maintained at 95 °C for 24 h. After completion, the solution was filtered, washed, and dried at 60 °C to obtain purified ZnO microrods (referred to as ZO).

To prepare Co-ZnO-X composites (where X represents the concentration of Co(NO₃)₂·6H₂O, X = 0, 0.5, 2, 4 mM), a light deposition method was employed. A mixture containing 0.2 g ZO, varying amounts of Co(NO₃)₂·6H₂O, and 100 mL of ultrapure water was stirred under xenon lamp irradiation for 20 h. Subsequently, the solution was filtered, washed, and dried at 60 °C in a vacuum oven, resulting in the formation of Co-ZnO-0 (CZO-0), Co-ZnO-0.5 (CZO-0.5), Co-ZnO-2 (CZO-2), and Co-ZnO-4 (CZO-4) composites.

Iodometric Method for Detecting H₂O₂.

The concentration of H₂O₂ produced was detected by the iodometric method according to the literature following the steps below: Ammonium molybdate tetrahydrate (H₂₄Mo₇N₆O₂₄·4H₂O) was dissolved in 5 mL of deionized water to prepare a 0.01 M solution, while potassium iodide (KI) was dissolved in 10 mL of deionized water to prepare a 0.1 M KI solution. During different intervals of ultrasonic irradiation, 0.5 mL of the reaction solution was collected and mixed with 2 mL of the KI solution, along with 50 μL of ammonium molybdate tetrahydrate solution. The resulting mixture was vigorously shaken, allowed to stand for 10 min, and subsequently analyzed using a UV-Vis spectrophotometer at a maximum absorption wavelength of 351 nm. The concentration of H₂O₂ was determined using a standard curve method.

Degradation of Organic Pollutants.

The degradation efficiency of different pollutants [including phenol, nitrobenzene (NB), aniline (AB), sulfadiazine (SD), p-chlorophenol (4-CP), quinoline, rhodamine B (RhB), methylene blue (MB), and acid orange 7 (AO7)] was assessed by determining their concentrations using HPLC and UV-Vis spectrophotometer. At specified time intervals, a 1 mL aliquot of the sample was taken and passed through a 0.22-μm membrane to remove solid catalyst. The injection volume for HPLC analysis was 20 μL. The flow rate was set at 1 mL min⁻¹, and the column temperature was maintained at 30 °C. The specific HPLC methods used for analyzing different pollutants could be found in SI Appendix, Table S3. Additionally, the solution samples were analyzed for total organic carbon (TOC) content using a TOC analyzer.

Statistical Analysis.

All data were gained directly from the source experiment and processed using Origin. All experiments were carried out in duplicate. Full details for DFT computations, control experiments, and materials characterizations were introduced in SI Appendix.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

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