Selective hydroxyl generation for efficient pollutant degradation by electronic structure modulation at Fe sites

Significance

Activation of H₂O₂ to selectively generate hydroxyl radical (·OH) with stronger oxidation performance is a challenging issue in current research. The cleavage pathway of H₂O₂ plays a crucial role in determining the selectivity of ·OH generation, with homolytic cleavage of H₂O₂ being more favorable for the selective generation of ·OH. Herein, the electronic structure of the Fe site was regulated by the introduction of a Cu dopant, which enhanced the adsorption and activation of H₂O₂ at the Fe site. The modification of the electron structure at the Fe site made the cleavage of H₂O₂ more inclined to homolytic cleavage, which improved the selectivity of •OH generation. Based on this, the efficient degradation of various organic pollutants can be realized.

Abstract

Hydrogen peroxide (H₂O₂) is an important green oxidant in the field of sewage treatment, and how to improve its activation efficiency and generate free radicals with stronger oxidation performance is a key issue in current research. Herein, we synthesized a Cu-doped α-Fe₂O₃ catalyst (7% Cu–Fe₂O₃) for activation of H₂O₂ under visible light for degradation of organic pollutants. The introduction of a Cu dopant changed the d-band center of Fe closer to the Fermi level, which enhanced the adsorption and activation of the Fe site for H₂O₂, and the cleavage pathway of H₂O₂ changed from heterolytic cleavage to homolytic cleavage, thereby improving the selectivity of •OH generation. In addition, Cu doping also promoted the light absorption ability of α-Fe₂O₃ and the separation of hole–electron pairs, which enhanced its photocatalytic activities. Benefiting from the high selectivity of •OH, 7% Cu–Fe₂O₃ exhibited efficient degradation activities against ciprofloxacin, the degradation rate was 3.6 times as much as that of α-Fe₂O₃, and it had good degradation efficiency for a variety of organic pollutants.

The rapid rise in population and urbanization has led to a severe global issue of water pollution (1). As one of the greenest oxides, hydrogen peroxide (H₂O₂) plays an indispensable role in water pollution control (2). H₂O₂ can be activated in different ways to produce strongly oxidizing radicals for the degradation of organic pollutants, such as superoxide radicals (•O₂⁻), hydroperoxyl radicals (•O₂H), and hydroxyl radicals (•OH) (3, 4). However, due to the high activation energy of H₂O₂ molecules, the reaction kinetics of its activation are generally limited, which is the bottleneck problem limiting the use of H₂O₂ in the field of water treatment (5). The traditional homogeneous Fenton reaction relies on the oxidation and reduction of Fe²⁺/Fe³⁺ to activate H₂O₂ (Eqs. 1 and 2).

$${\text{Fe}}^{\mathrm{\text{2+}}}\text{+}{\text{ H }}_{\mathrm{\text{2}}}{\mathrm{O}}_{\text{2}}\text{ → }{\mathrm{F}\mathrm{e}}^{\mathrm{3+}}\text{+}\text{•OH}\text{+}{\mathrm{\text{OH}}}^{\text{-}}\text{.}$$ [1]

$${\mathrm{\text{Fe}}}^{\text{3+}}\text{+}{\mathrm{H}}_2{\mathrm{O}}_2\text{→}{\mathrm{F}\mathrm{e}}^{\mathrm{2+}}\text{+}{\text{•HO}}_{\mathrm{\text{2}}}\text{+}{\mathrm{\text{H}}}^{\mathrm{\text{+}}}\text{.}$$ [2]

It is easy to form iron slime and cause excessive consumption of H₂O₂, which limits its application in wastewater treatment (6). In order to improve the activation efficiency of H₂O₂ and avoid the disadvantages of the traditional Fenton reaction, more and more studies have shifted the focus from homogeneous to heterogeneous systems (7). In particular, the heterogeneous photocatalytic activation of H₂O₂ using semiconductors as catalysts has broad research prospects. The photocatalytic reaction can provide photogenerated electrons for the activation of H₂O₂ and accelerate the cyclic reaction of metal ions (≡Mⁿ + H₂O₂ → ≡M^{n +1} + •OH + OH⁻, ≡Mⁿ⁺¹ + e⁻→ ≡Mⁿ), which effectively improves the rate-limiting step of H₂O₂ activation and effectively avoids the secondary pollution caused by metal ion precipitation, with the advantages of high efficiency and green (8).

Being abundant, inexpensive, and nontoxic, Fe is one of the most prevalent elements in the earth's crust. Fe₂O₃, Fe₃O₄, and FeOOH are among several iron compounds that are abundant in nature and widely used for water treatment (9). Hematite (α-Fe₂O₃) is a semiconductor material that possesses several advantages, including high stability, environmental friendliness, and low cost. As a result, it has garnered considerable attention in the fields of photocatalysis and photo-Fenton (10). To enhance the photocatalytic and H₂O₂ activation activity of the catalyst, various strategies have been developed, including doping engineering (11), heterostructure construction (12), and crystal plane regulation (9). More importantly, the activation pathway of H₂O₂ may affect the degradation efficiency of the catalytic system by affecting the generation of free radicals (13). At present, the activation pathways of H₂O₂ mainly include heterolytic cleavage (Eq. 3) and homolytic cleavage (Eq. 4).

$$\begin{array}{c}{\mathrm{ }}^{\text{*}}{\mathrm{H}}_{\text{2}}{\mathrm{O}}_{\text{2}}\text{→}\mathrm{*}\mathrm{H}\text{-}\mathrm{\text{OOH}}\mathrm{*\to *}\text{OOH→Free radical.}\hfill \end{array}$$ [3]

$$\begin{array}{c}{\mathrm{ }}^{\text{*}}{\mathrm{H}}_{\text{2}}{\mathrm{O}}_2\text{→}\mathrm{*}\text{HO}\mathrm{\text{-}}\text{OH}\mathrm{*\to *}\text{OH→Free radical.}\hfill \end{array}$$ [4]

For the heterolytic cleavage pathway, the decomposition of H₂O₂ leads to O–H bond cleavage to generate *OOH and then through electron transfer to generate free radicals. This process tends to generate free radicals mainly including •HO₂/•O₂⁻ and •OH (14–16). For the homolytic cleavage pathway, the direct breaking of the O–O bond in H₂O₂ produces two *OH intermediates, which are mainly formed into •OH through electron transport (3, 17). Considering the degradation efficiency of the catalytic system, we prefer the highly selective generation of •OH. This is due to the fact that •OH (+2.38 V vs. SHE) has a better oxidation capacity than •O₂⁻ (−0.33 V vs. SHE) for organic pollutants, especially for refractory organic pollutants (18, 19). Comparing the two cleavage pathways, the homolytic cleavage pathway seems to be more selective for the formation of •OH and thus more favorable for the degradation of organic pollutants. Kim et al. showed that •OH, •OOH, and •O₂⁻ can be generated by H₂O₂ cleavage, but the production of •OH is mainly through homolytic cleavage of H₂O₂ by heterogeneous catalysis (20). Xing et al. have demonstrated that the charge density of Fe₃C in the Fe/Fe₃C structure is moderate, which promotes the homolytic cleavage of H₂O₂ and the efficient generation of •OH. The high charge density of Fe⁰ in the Fe/Fe₃C structure tends to promote the heterolytic cleavage of H₂O₂ to generate O₂ with low oxidation capacity, resulting in the waste of H₂O₂ (21).

Herein, a Cu-doped α-Fe₂O₃ nanosheet (x% Cu–Fe₂O₃) was designed and synthesized to photoactivate H₂O₂ for the degradation of organic pollutants. Among them, the degradation rate of ciprofloxacin (CIP) by 7% Cu–Fe₂O₃ was 3.6 times that of the original α-Fe₂O₃. Furthermore, 7% Cu–Fe₂O₃ exhibited excellent degradation efficiency toward various organic pollutants. Combining experimental and theoretical research, we found that Cu doping alters the electronic structure of the Fe site, transforming the activation pathway of H₂O₂ from heterolytic cleavage to homolytic cleavage and improving the selectivity of •OH formation. Additionally, Cu dopants enhance the photocatalytic activity of α-Fe₂O₃ by changing its band structure, which promotes H₂O₂ activation. This work provides unique ideas and perspectives on the photoactivation mechanism of H₂O₂ for the highly selective generation of •OH by selecting the activation pathway of H₂O₂.

Results

Structure and Characterization of Catalysts.

The synthesis process of Cu-doped α-Fe₂O₃ was shown in Fig. 1A. A series of catalysts with different Cu content (denoted as x% Cu–Fe₂O₃, x is the molar ratio of Cu doping) were synthesized by hydrothermal method by adjusting the amount of metal salt added. The actual doping amount of the samples was determined by inductively coupled plasma mass spectrometry (ICP-MS), which was a little different from the theoretical doping amount (SI Appendix, Table S1). Of note, 7% Cu–Fe₂O₃ was taken as the representative sample for study because it had the best performance. Transmission electron microscopy (TEM) revealed that the α-Fe₂O₃ nanosheets had a hexagonal shape with a transverse dimension of approximately 250 nm. Cu doping changed the original regular hexagonal structure of α-Fe₂O₃ nanosheets into disc-shaped nanosheets. The lattice spacing of α-Fe₂O₃ and 7% Cu–Fe₂O₃, as measured by TEM, was approximately 0.246 nm and 0.251 nm, respectively, corresponding to the (110) crystal plane (Fig. 1 B and D). This indicates that Cu doping does not change the exposed crystal plane of the catalyst, and lattice expansion may be due to the fact that the ionic radius of the doped element (0.77/0.73 Å) is greater than that of Fe³⁺ (0.55 Å). The element mapping indicated that the Cu element was successfully doped and evenly distributed (Fig. 1 C and E). The X-ray diffraction (XRD) peaks of α-Fe₂O₃ match the standard card 33-0664, and those of 7% Cu–Fe₂O₃ generally agree with it, indicating that Cu doping has no significant effect on the crystal structure of α-Fe₂O₃ (Fig. 1F). Furthermore, the XRD pattern of 7% Cu–Fe₂O₃ does not exhibit any characteristic peak of CuO/Cu₂O, suggesting that Cu is doped into the lattice of α-Fe₂O₃ instead of forming a composite structure with CuO or Cu₂O. The Fourier transform infrared (FT-IR) spectroscopy (Fig. 1G) showed that the peaks at 466, 546, and 772 cm⁻¹ correspond to the Fe-O stretching vibration mode in α-Fe₂O₃. These peaks in 7% Cu–Fe₂O₃ slightly decrease, possibly due to the effect of Cu doping on the Fe-O stretching vibration (22).

Fig. 1.

(A) Schematic diagram of synthesis of x% Cu–Fe₂O₃. (B) TEM images and high-resolution TEM images of α-Fe₂O₃. (C) High-angle annular dark field image and element mapping image of α-Fe₂O₃. (D) TEM images and high-resolution TEM images of 7% Cu–Fe₂O₃. (E) High-angle annular dark field image and element mapping image of 7% Cu–Fe₂O₃. (F) XRD pattern and (G) FT-IR pattern of α-Fe₂O₃ and 7% Cu–Fe₂O₃. (H) The Fe 2p high-resolution XPS spectra of α-Fe₂O₃ and 7% Cu–Fe₂O₃. (I) The Cu 2p high-resolution XPS spectrum of 7% Cu–Fe₂O₃.

Raman scattering is highly sensitive to lattice disorder. Therefore, the broadening and intensity reduction of the Raman peak of α-Fe₂O₃ incorporated with Cu can be attributed to the disordered substitution of Cu (SI Appendix, Fig. S1) (23). The X-ray photoelectron spectroscopy (XPS) was used to determine the influence of Cu doping on the electronic structure and chemical state of α-Fe₂O₃ and to reveal the catalytic advantage of 7% Cu–Fe₂O₃. High-resolution spectra of Fe and Cu were calibrated at 284.8 eV using the C 1s XPS high-resolution spectrum (SI Appendix, Fig. S2). The effect of Cu doping on electron density around Fe was revealed by analyzing the high-resolution XPS spectra of Fe 2p (Fig. 1H). For α-Fe₂O₃, the characteristic peaks near 710.2 eV and 723.5 eV were attributed to Fe 2p_3/2 and Fe 2p_1/2, respectively, and there were two distinct satellite peaks at 718.5 eV and 732.5 eV, indicating that Fe (III) is the main content on the catalyst surface (24). After Cu doping, the two fitting peaks of Fe 2p_3/2 moved toward the direction of low binding energy and the peak intensity decreases, which meant that Cu doping changed the chemical state of Fe. This may be due to the electrical interaction between Cu and Fe induced by Cu doping leading to electron transfer from Cu to Fe, thus increasing the electron cloud density around Fe. In addition, the decrease of the satellite peak intensity of 7% Cu–Fe₂O₃ also indicated that the electronic interaction between Fe and Cu reduced the surface Fe (III) content (25). The Cu 2p high-resolution XPS spectrum (Fig. 1I) revealed that the main valence states of Cu in 7% Cu–Fe₂O₃ are Cu (II) (935.1 eV) and Cu (I) /Cu (932.6 eV) (26). The Cu Auger LMM spectrum was collected due to the characteristic peaks of both metallic copper and copper oxides appeared in the range of 932.5 to 932.6 eV, which made it difficult to distinguish. The Cu Auger LMM spectrum (SI Appendix, Fig. S3) revealed the presence of Cu (I) and Cu (II) but not metallic copper (27). The surface area of α-Fe₂O₃ and 7% Cu–Fe₂O₃ is 18.43 m²/g and 16.12 m²/g, respectively. This suggested that the specific surface area is not a major factor contributing to the catalytic reaction (SI Appendix, Fig. S4 A and B and Table S2) (28). The thickness and surface potential of the synthesized materials were studied by atomic force microscopy and Kelvin probe force microscopy (KPFM). The results indicated that the thickness of the catalyst is basically unchanged before and after doping, which is about 6 to 7 nm. It is gratifying that the KPFM test results showed that the surface potential of the catalyst was significantly enhanced and fluctuated after Cu doping, which will be beneficial to the transport and transfer of photogenerated electrons (SI Appendix, Fig. S5).

Catalytic Degradation of the 7% Cu–Fe₂O₃/H₂O₂/Vis System.

The degradation efficiency of Cu- doped α-Fe₂O₃ photoactivated H₂O₂ on pollutants was investigated using 10 mg/L CIP solution as the probe pollutant. Benefiting from the excellent ability of Cu-doped α-Fe₂O₃ to photoactivate H₂O₂, the degradation efficiency of CIP was significantly improved compared to that of α-Fe₂O₃. Among them, 7% Cu–Fe₂O₃ exhibited the best degradation effect on CIP, with a degradation rate of 95.8% achieved within 60 min (Fig. 2A). The pseudo-first-order kinetic constant (k) reached 0.0532 min⁻¹, which is about 3.6 times that of α-Fe₂O₃ (Fig. 2B). This result is superior to that of catalysts reported in some previous studies (SI Appendix, Table S3). To elucidate that the degradation process is dominated by the photoactivation of H₂O₂ by the catalyst, a series of experiments were conducted. First, the adsorption or removal of CIP by the catalyst under dark conditions was investigated. The concentration of CIP did not decrease significantly after the catalyst reacted with CIP for 30 min under dark conditions, indicating that the adsorption or degradation of CIP by the photocatalyst was not obvious in the absence of visible light irradiation (SI Appendix, Fig. S6). Second, the degradation of CIP by α-Fe₂O₃ and 7% Cu–Fe₂O₃ in the case of visible light irradiation alone or H₂O₂ addition alone was explored (SI Appendix, Fig. S7). α-Fe₂O₃ and 7% Cu–Fe₂O₃ had weak degradation effect on CIP under pure visible light irradiation, and the degradation efficiency of 7% Cu–Fe₂O₃ on CIP was slightly higher than that of α-Fe₂O₃. In the presence of H₂O₂ alone, the results showed that the degradation of CIP by the catalysts was extremely weak and almost negligible, which indicated that α-Fe₂O₃ and 7% Cu–Fe₂O₃ could not directly activate H₂O₂ effectively. The above experimental results indicated that the degradation of CIP is due to the photoactivation of H₂O₂ to generate active species. The H₂O₂ consumption experiment showed that the consumption ability of 7% Cu–Fe₂O₃ to H₂O₂ was significantly higher than that of α-Fe₂O₃ (Fig. 2C, the inset showed that the consumption rate was significantly increased), indicating that the photoactivation ability of 7% Cu–Fe₂O₃ to H₂O₂ was significantly improved. In addition, the optimal dosage of H₂O₂ in the catalytic system was studied (Fig. 2D). The results showed that when the concentration of H₂O₂ in the catalytic system was 10 mM, the degradation efficiency of CIP was the highest. When the concentration of H₂O₂ in the system exceeded 10 mM, the degradation efficiency of CIP decreased, which may be due to the excessive H₂O₂ will react with •OH as follows: H₂O₂ + •OH = HO₂• + H₂O, thus consuming •OH (29).

Fig. 2.

(A) The degradation efficiency of CIP (10 mg/L) by the prepared samples photoactivated H₂O₂. (B) Pseudo-first-order kinetic curves of the degradation efficiency of CIP by the prepared samples photoactivated H₂O₂. (C) The consumption of H₂O₂ by α-Fe₂O₃/H₂O₂/Vis and 7% Cu–Fe₂O₃/H₂O₂/Vis catalytic systems. (D) Effect of H₂O₂ dosage on the degradation of CIP by the 7% Cu–Fe₂O₃/H₂O₂/Vis catalytic system. (E) Leakage of metal ions in the 7% Cu–Fe₂O₃/H₂O₂/Vis catalytic system and degradation of CIP by homogeneous ions of the same concentration. (F) Degradation of different organic pollutants by the 7% Cu–Fe₂O₃/H₂O₂/Vis catalytic system. Reaction conditions: [CIP] = [RhB] = [TC] = [SMZ] = [Phenol] = 10 mg/L; [catalyst] = 0.2 g/L; [H₂O₂] = 10 mM; illumination conditions: 300 W Xe lamp; [Temp] = 20 °C; initial pH = 6.8.

To ensure that the treated water was not contaminated and the catalytic reaction was not affected by metal ion leakage from the catalytic material, we measured the metal ion leakage using ICP-MS (Fig. 2E). The results showed that the leakage of metal ions of 7% Cu–Fe₂O₃ was quite low during the reaction. The leakage amount of Cu ions was about 31.12 μg/L, and that of Fe ions was about 12.79 μg/L, which is much lower than that of China drinking water standard (The limits of Cu and Fe in GB5749-2022 are 1.0 mg/L and 0.3 mg/L, respectively) (30). Homogeneous Cu²⁺, Fe²⁺, and Fe³⁺ were separately added at the same concentration as the catalyst leaching to the catalytic system (H₂O₂/Vis) to eliminate the influence of the leakage ions on the catalytic reaction (Inset of Fig. 2E). The results showed that the above homogeneous ions had minimal impact on the degradation of CIP, which indicated that the CIP was degraded by 7% Cu–Fe₂O₃ photoactivated H₂O₂. In addition, 7% Cu–Fe₂O₃ has degradation effect on a variety of organic compounds, including rhodamine, tetracycline, sulfamethoxazole, and phenol (Fig. 2F), indicating that the catalytic system has good oxidative degradation ability.

To explore the causes of pollutant degradation, a series of scavengers were introduced into the reaction system to identify the active species that degrade pollutants (31). The addition of mannitol and AgNO₃ inhibited the degradation efficiency of CIP most obviously (SI Appendix, Fig. S8). The results indicated that •OH plays a significant role in the degradation of CIP by photoactivation of H₂O₂, and electron transfer facilitates the generation of free radicals. The electron paramagnetic resonance (EPR) test provided direct evidence for photoactivation of H₂O₂ to produce reactive oxygen species by α-Fe₂O₃ and 7% Cu–Fe₂O₃ so as to explore the reason for the improvement of pollutant degradation efficiency after Cu doping (32). The •OH was detected using 5,5-dimethyl-1-pyrroline N-oxide (DMPO) as the capture agent and water as the solvent. The •OH signals (four peaks with a peak intensity of 1:2:2:1) appeared in both α-Fe₂O₃ and 7% Cu–Fe₂O₃ reaction systems (Fig. 3A), whereas the relative intensity of the signal of 7% Cu–Fe₂O₃ was much stronger than that of α-Fe₂O₃, indicating that much more •OH was produced in the 7% Cu–Fe₂O₃ reaction system than the α-Fe₂O₃ reaction system (7). The •O₂⁻ was determined using DMPO as the capture agent and methanol as the solvent. Both α-Fe₂O₃ and 7% Cu–Fe₂O₃ reaction systems showed a six-peak •O₂⁻ signal (Fig. 3B), and the signal intensity of α-Fe₂O₃ and 7% Cu–Fe₂O₃ was basically the same (28). In addition, the presence of ¹O₂ was not found in either reaction system (Fig. 3C). The •O₂⁻ and •OH were quantified using 0.05 mM p-nitro blue tetrazolium chloride (NBT) and 0.5 mM terephthalic acid (33, 34). The amount of •O₂⁻ produced by the α-Fe₂O₃ catalytic system was approximately 143.79 μM, while that of the 7% Cu–Fe₂O₃ catalytic system was 155.27 μM within 60 min, showing little difference in •O₂⁻ production, which is consistent with the results obtained from EPR analysis (Fig. 3D). For the •OH, the yield of the α-Fe₂O₃ catalytic system is about 0.06 mM, and that of the 7% Cu–Fe₂O₃ catalytic system is significantly higher, reaching 4.01 mM, which is also consistent with the EPR results (Fig. 3E). As can be seen from the yield chart, the yield of •OH was greatly increased in the first 20 min of the reaction, which was highly similar to the degradation process of CIP, indicating that •OH played an extremely important role in the degradation process. The selective calculation results of free radical formation indicate that the selectivity of •OH generated by photoactivation of H₂O₂ is significantly enhanced after Cu-doped α-Fe₂O₃ (Fig. 3F). The selectivity of α-Fe₂O₃ photoactivated H₂O₂ for generating •O₂⁻ and •OH was found to be 55.9% and 44.1%, respectively, whereas the selectivity of 7% Cu–Fe₂O₃ was 3.7% and 96.3%, respectively. Notably, the selectivity for •OH formation was significantly improved after Cu doping. This enhancement may be attributed to changes in the activity and electron cloud density of the α-Fe₂O₃ surface sites induced by Cu doping, which in turn altered the decomposition pathway of H₂O₂, facilitating the generation of •OH. These findings are of great interest for further research.

Fig. 3.

EPR spectra of (A) •OH, (B) •O₂⁻, and (C) ¹O₂ of α-Fe₂O₃/H₂O₂/Vis and 7% Cu–Fe₂O₃/H₂O₂/Vis catalytic systems. The •O₂⁻ (D) and •OH (E) quantification of α-Fe₂O₃/H₂O₂/Vis and 7% Cu–Fe₂O₃/H₂O₂/Vis catalytic systems. (F) Selectivity of α-Fe₂O₃/H₂O₂/Vis and 7% Cu–Fe₂O₃/H₂O₂/Vis catalytic systems for •OH and •O₂⁻ generation.

Mechanism of H₂O₂ Activation Enhancement and •OH Formation Selective Elevation.

Density functional theory (DFT) calculations were used to investigate the effect of Cu doping on the electronic structure of α-Fe₂O₃ and the activation path of H₂O₂ in an attempt to reveal the mechanism that the yield and selectivity of •OH increase significantly. Based on our TEM study results, the synthesized catalyst mainly exposed (110) crystal planes, so we established the catalyst structure model with exposed (110) crystal planes (SI Appendix, Fig. S9). First, the adsorption of H₂O₂ on the catalyst surface was studied to determine the adsorption site and configuration of H₂O₂. There are two main adsorption configurations of H₂O₂ on α-Fe₂O₃ and 7% Cu–Fe₂O₃, namely end-on adsorption and side-on adsorption (35). As shown in Fig. 4 A and B, H₂O₂ tends to be adsorbed on Fe sites of α-Fe₂O₃ and 7% Cu–Fe₂O₃ in end-on mode, and the adsorption energy is −0.76 eV and −0.81 eV, respectively. These results indicate that Cu doping does not change the adsorption site and adsorption configuration of H₂O₂ and thus may promote adsorption by affecting the electronic structure of the Fe sites. The results of energy band structure (SI Appendix, Fig. S10) and partial density of states (PDOS) (SI Appendix, Fig. S11) calculations supported our conjecture that Cu doping leads to an upward shift of the d-band center of Fe (from −1.37 eV for α-Fe₂O₃ to −1.35 eV for 7% Cu–Fe₂O₃) closer to the Fermi energy level, enhancing the adsorption of H₂O₂ by the Fe site and the catalytic activity of the Fe sites.

Fig. 4.

Adsorption energy for different adsorption configurations and activation energy of different cleavage pathways of H₂O₂ on (A) α-Fe₂O₃ and (B) 7% Cu–Fe₂O₃. Charge differential density calculation and Bader charge calculation of H₂O₂ adsorbed on (C) α-Fe₂O₃ and (D) 7% Cu–Fe₂O₃. (E) Plot of optimal photoactivation pathways of α-Fe₂O₃ and 7% Cu–Fe₂O₃ to H₂O₂. In situ Raman spectroscopy of (F) α-Fe₂O₃/H₂O₂/Vis and (G) 7% Cu–Fe₂O₃/H₂O₂/Vis catalytic systems. Cu LMM spectra of catalysts of the 7% Cu–Fe₂O₃/H₂O₂/Vis catalytic system at different reaction times, including (H) 0 min, (I) 10 min, (J) 30 min, and (K) 60 min.

In addition, the calculated results of the charge density difference are not difficult to see that H₂O₂ gains more electrons at the Fe adsorption site (activation site) after Cu doping, and the calculated Bader charge transfer increases from 0.090 e⁻ to 0.141 e⁻ (Fig. 4 C and D). This indicates that Cu doping promotes the transfer of electrons from the catalyst to H₂O₂, which is conducive to the adsorption of H₂O₂ and the subsequent activation cleavage. It is worth noting that although Cu doping does not change the adsorption configuration of H₂O₂ on the catalyst surface, it is found that Cu doping affects the cleavage pathway of H₂O₂ by calculating the free energy of generating H₂O₂ cleavage intermediates. For α-Fe₂O₃, the activation cleavage of H₂O₂ is more likely to be heterolytic cleavage (H₂O₂ → *H + *OOH) due to its lower energy barrier. On the contrary, for 7% Cu–Fe₂O₃, H₂O₂ is more inclined to homolytic cleavage (H₂O₂ → *OH + *OH). The heterolytic cleavage of H₂O₂ at the active site of α-Fe₂O₃ is thermodynamically possible but faces a large activation energy barrier (ΔE = +0.54 eV). In contrast, H₂O₂ on 7% Cu–Fe₂O₃ can be homolytic cleavage without obstacle to form *OH intermediate (ΔE = −0.32 eV), and this reaction occurs spontaneously (Fig. 4E). The above calculation results show that Cu doping not only changes the electronic structure of the adsorption site to promote the adsorption of H₂O₂ but more importantly changes the activation path of H₂O₂. The homolytic cleavage of H₂O₂ is more favorable to improve the yield and selectivity of •OH, which is due to the fact that the *OOH intermediate produced by the heterolytic cleavage of H₂O₂ is easy to produce O₂ and •O₂⁻ with relatively weak oxidation capacity, thus causing the waste of H₂O₂ (18, 21, 36).

In situ Raman spectroscopy was used to identify the decomposition intermediates, which provided direct evidence for the decomposition path of H₂O₂. The key to distinguish the two decomposition modes of H₂O₂ is the identification of the two decomposition intermediates *OOH and *OH. As shown in Fig. 4 F and G, the peak near 878 cm⁻¹ is the O–O stretching vibration of H₂O₂ (37). In the α-Fe₂O₃/H₂O₂/Vis catalytic system, as the reaction time goes on, it is not difficult to see that a clear Raman peak appears near 614 cm⁻¹ (Fig. 4F), which should be attributed to the stretching vibration of Fe–O in Fe–OOH (38, 39). In the catalytic system of 7% Cu–Fe₂O₃/H₂O₂/Vis, since Cu doping did not change the adsorption configuration of H₂O₂, the Raman peak at 614 cm⁻¹ was attributed to the telescopic vibration of Fe–O in Fe–*H₂O₂ (Fe–OOH). Encouragingly, a Raman peak has emerged near 563 cm⁻¹, which is attributed to tensile vibrations of Fe–O in Fe–OH (Fig. 4G) (40, 41). The results demonstrate that in the 7% Cu–Fe₂O₃/H₂O₂/Vis system, H₂O₂ molecules undergo photoactivated homolytic cleavage (*HO–*OH) to form *OH, which differs from the α-Fe₂O₃/H₂O₂/Vis system. This suggests that the introduction of Cu dopants alters the cleavage pathway of H₂O₂, promoting the selective generation of •OH. The analytical results obtained from Raman spectroscopy are in agreement with the DFT calculations.

The Cu Auger LMM spectra at different reaction times (0 min, 10 min, 30 min, and 60 min) provided evidence for the involvement of Cu sites in H₂O₂ photoactivation. Before the reaction proceeds, the area ratio of Cu (II) is higher than Cu (I) in the Cu Auger LMM spectrum. This is due to the fact that electrons around Cu are enriched around Fe after Cu doping into α-Fe₂O₃; hence, the Cu mainly exists in the form of Cu (II) in the catalyst (Fig. 4H). After 10 min of reaction, a large amount of Cu (II) is converted into Cu (I), which is because Cu dopants can be used as trapping centers of photogenerated electrons, leading to electron enrichment around Cu sites (Eqs. 5 and 6) (Fig. 4I). As the reaction progresses, the amount of Cu (II) increases, and the transfer of electrons from Cu (I) to Fe promotes the reduction of Fe (III) to Fe (II) (Eq. 7) (Fig. 4 J and K). At the end of the reaction, the valence state of Cu returns to a state similar to that before the reaction began. In this way, Cu is driven by photogenerated electrons to realize the valence state cycle of Cu, thus continuously providing electrons to the Fe site to produce Fe (II) for H₂O₂ activation (Eq. 8). A mechanistic diagram of the reaction process is presented in Supporting Information (SI Appendix, Fig. S12).

$$\text{catalyst}\stackrel{\text{hv}}{ \to }{\mathrm{h}}^{\text{+}}\text{+}{\mathrm{e}}^{\text{-}}\text{.}$$ [5]

$$\text{Cu}\mathrm{\text{(}}\text{II}\mathrm{\text{)}}\text{+}{\mathrm{e}}^{\text{-}}\to \text{Cu}\mathrm{\text{(}}\mathrm{I}\text{).}$$ [6]

$$\text{Cu }\mathrm{\text{(}}\mathrm{I}\text{)}\text{+}\text{Fe }\mathrm{\text{(}}\text{ III }\mathrm{\text{)}}\text{ → }\mathrm{\text{Cu }}\text{(}\mathrm{\text{II}}\text{)}\text{+}\text{Fe }\mathrm{\text{(}}\text{II}\mathrm{\text{).}}$$ [7]

$$\begin{array}{c}\text{Fe }\mathrm{\text{(}}\text{II}\mathrm{\text{)}}\text{+}{\mathrm{H}}_{\text{2}}{\mathrm{O}}_{\text{2}}\text{ → }\mathrm{\text{Fe }}\text{(}\mathrm{\text{III}}\text{)}\text{+}\mathrm{•OH}\text{+}\mathrm{O}{\mathrm{H}}^{\text{-}}\text{.}\hfill \end{array}$$ [8]

To clarify another important reason for the enhanced performance of 7% Cu–Fe₂O₃ photoactivated H₂O₂, the photocatalytic properties of α-Fe₂O₃ and 7% Cu–Fe₂O₃ were investigated. First of all, the light absorption capacity (SI Appendix, Fig. S13) and band gap (SI Appendix, Fig. S14) of the sample obtained were analyzed by UV-vis diffuse reflectance spectroscopy. The optical absorption edge of α-Fe₂O₃ extends from 647.3 nm to 683.6 nm of 7% Cu–Fe₂O₃, and the band gap narrows from 1.97 eV of α-Fe₂O₃ to 1.90 eV of 7% Cu–Fe₂O₃, which indicates that Cu doping promotes the light absorption ability of the catalyst. Valence band (VB) XPS spectra determined VB of 1.07 eV and 1.24 eV for α-Fe₂O₃ and 7% Cu–Fe₂O₃, respectively (SI Appendix, Fig. S15). The above experimental results are in good agreement with the calculated results of band structure and PDOS. The carrier separation efficiency and charge transfer ability of the catalyst were evaluated by transient photocurrent (SI Appendix, Fig. S16) and electrochemical impedance spectroscopy (SI Appendix, Fig. S17). Compared with α-Fe₂O₃, 7% Cu–Fe₂O₃ has stronger photocurrent and smaller electrochemical impedance, which indicates that 7% Cu–Fe₂O₃ has better carrier separation and charge transfer ability, which will be beneficial to photoactivation of H₂O₂. The separation efficiency of the electron–hole pairs and the lifetime of the photogenerated charge of the catalyst were evaluated by steady-state photoluminescence (PL) (SI Appendix, Fig. S18) and time-resolved PL (SI Appendix, Fig. S19) experiments. The PL intensity of 7% Cu–Fe₂O₃ is significantly reduced compared with that of α-Fe₂O₃, which means that the recombination of electrons and holes is effectively inhibited. Compared with α-Fe₂O₃, 7% Cu–Fe₂O₃ has a shorter PL average lifetime, which indicates that 7% Cu–Fe₂O₃ has a more convenient charge separation and transfer channel.

Degradation Mechanisms of Organic Pollutants and Application Prospect of Catalysts.

Benefiting from the highly selective formation of hydroxyl radicals, 7% Cu–Fe₂O₃ can degrade target pollutants efficiently, but the conversion intermediates generated in the degradation process may be potential environmental pollutants. Molecular electrostatic potential surface (MEPs) of pollutants can be calculated to predict the attack sites of free radicals. MEPs of pollutants can clearly reflect the electrophilic attack zone and nucleophilic attack zone of molecules. Electrophilic reactivity is indicated by negative potentials (red areas) and nucleophilic reactivity is indicated by positive potentials (blue areas). The •OH is a typical strong electrophilic free radical (42), and the electrophilic attack region of CIP is mainly distributed in the carboxyl, hydroxyl, and piperazine ring regions (Fig. 5A). Therefore, the attack degradation of CIP by •OH begins in these regions. In order to understand CIP degradation process in the 7% Cu–Fe₂O₃/H₂O₂/Vis system, the intermediate products of CIP degradation process were analyzed by high-performance liquid chromatography and mass spectrometry (HPLC-MS). According to the CIP degradation intermediates identified by HPLC-MS analysis, three possible CIP degradation pathways were proposed (Fig. 5B). Both degradation pathways I and II first oxidized CIP to intermediates (C1) with an m/z of 362.14 through hydroxyl addition and fluorine substitution. In pathway I, C1 underwent dehydrogenation and hydroxylation to form CI1, which was then subjected to decarboxylation and ring opening, resulting in a reduction in the molecular weight of the pollutant (43). Pathway II mainly involves the attack of •OH on the piperazine ring, leading to the opening of the ring structure. Piperazine is gradually oxidized, producing the CII3 product with an m/z of 263.24 (44). Pathway III mainly involves the opening of the piperazine ring through decarboxylation and hydroxylation to facilitate further oxidation. Ultimately, the molecular weight of the main intermediates decreases continuously through •OH oxidation, culminating in mineralization into CO₂ and H₂O. The possible degradation pathways analyzed by HPLC-MS revealed that the primary sites where the contaminants were attacked by •OH were consistent with the results calculated by MEPs, so the analyzed degradation paths have high reliability.

Fig. 5.

(A) Molecular electrostatic potential surface (MEPs) of CIP. (B) Degradation pathway of CIP by the 7% Cu–Fe₂O₃/H₂O₂/Vis catalytic system. (C) A simple device to simulate the practical application of the 7% Cu–Fe₂O₃/H₂O₂/Vis catalytic system.

Assessing the toxicity and potential harm of degradation intermediates is critical to evaluate the effectiveness of pollutant degradation. To evaluate the toxicity of CIP and its degradation intermediates, toxicity evaluation software was employed, using Daphnia magna LC₅₀ (48 h), Fathead minnow LC₅₀ (96 h), developmental toxicity value, and mutagenicity value as evaluation criteria. The toxicity of most intermediates to Daphnia magna and Fathead minnow is significantly reduced, with only CIII5 to Daphnia magna and CI4 to Fathead minnow showing greater toxicity than CIP (SI Appendix, Fig. S20 A and B). Regarding the developmental toxicity value, all intermediates except CII2 have smaller values. Regarding the mutagenicity value, the values of all intermediates except CII2 and CIII4 are reduced. Overall, the toxicity of CIP is effectively reduced after degradation using the 7% Cu–Fe₂O₃/H₂O₂/Vis catalytic system (SI Appendix, Fig. S21 A and B).

The reusability and stability of catalysts are particularly important for practical water treatment applications. To investigate the reusability of the 7% Cu–Fe₂O₃ catalyst, five successive CIP degradation experiments were conducted. After 5 cycles, the CIP degradation efficiency of the 7% Cu–Fe₂O₃ catalyst remained close to 90%, indicating that the prepared catalyst has good cycling performance of Cu and Fe ions and good stability (SI Appendix, Fig. S22). The XRD and Fe 2p XPS high-resolution patterns before and after the reaction showed no significant changes (SI Appendix, Figs. S23 and S24), indicating that the catalyst has good stability and is suitable for actual water treatment processes. The presence of organic matter or anions in actual polluted wastewater may compete with pollutants for adsorption sites or consume free radicals, which can affect the degradation efficiency of the system. To test the applicability of the 7% Cu–Fe₂O₃ catalyst in practical applications, we added humic acid (HA), HCO₃⁻, and Cl⁻ to the system to simulate the actual water environment and investigate their effect on the degradation of CIP. The addition of organic matter or anions had little effect on the degradation efficiency of the 7% Cu–Fe₂O₃/H₂O₂/Vis catalytic system, indicating its practical application potential (SI Appendix, Fig. S25).

The highly efficient degradation effect of the 7% Cu–Fe₂O₃/H₂O₂/Vis catalytic system and the need for its future practical application urge us to explore further its treatment potential. Considering the good dispersion of 7% Cu–Fe₂O₃ in solution, a polyvinylidene fluoride membrane was used as the catalyst carrier to improve its repeatability and avoid secondary pollution. The preparation of 7% Cu–Fe₂O₃ membrane was done through vacuum filtration, and the catalyst was firmly attached to the surface of the membrane, ensuring that it would not detach during the catalytic reaction (SI Appendix, Fig. S26). To verify the performance of the 7% Cu–Fe₂O₃ membrane, an experimental device was built as shown in Fig. 5C. The device used secondary effluent from Nankai University Sewage Treatment Plant (Tianjin, China) as a solvent, and RhB was added as an indicator for catalytic degradation. Remarkably, the RhB solution achieved complete decolorization removal in the end effluent after passing through the designed catalytic membrane device. This suggests that the 7% Cu–Fe₂O₃ catalyst has great potential for future practical applications.

Conclusions

In summary, we synthesized the 7% Cu–Fe₂O₃ photocatalyst which can activate H₂O₂ to selectively generate •OH and has a great degradation effect on various organic pollutants. TEM results showed that the Cu dopant was successfully introduced. The results of quantitative experiments on EPR and free radical showed that the selectivity of •OH generation in the catalytic system of 7% Cu–Fe₂O₃/H₂O₂/Vis was significantly improved. The calculation results of PDOS and band structure indicated that Cu doping changed the electronic structure of the Fe site, making the center of the d-band of Fe moved up closer to the Fermi level. Charge differential density and Bader charge transfer calculation results showed that the introduction of Cu promoted the electron transfer of 7% Cu–Fe₂O₃ to H₂O₂ adsorbed on the surface and thus promoted the activation of H₂O₂. Gibbs free energy calculation and in situ Raman spectroscopy manifested that the cleavage pathway of H₂O₂ changed from heterolytic cleavage to homolytic cleavage after Cu doping, which was conducive to the highly selective generation of •OH. In addition, the introduction of the Cu dopant changed the photoelectric properties of α-Fe₂O₃ and enhanced its photocatalytic activity, and Cu acted as a capture center for photogenerated electrons to transfer electrons to Fe sites and promoted valence cycling of Fe. Based on the above reasons, 7% Cu–Fe₂O₃ photoactivated H₂O₂ to degrade CIP significantly. Our work proved that the change of H₂O₂ cleavage path can effectively enhance the selective formation of •OH, providing a unique perspective and idea for doping engineering to enhance the Fenton-like reactivity mechanism.

Materials and Methods

Preparation of the Catalyst.

FeCl₃·6H₂O (0.0025 mol, 0.6825 g) and CuCl₂·2H₂O in different molar ratios were dispersed in 1.75 mL water and 25 mL ethanol mixture under magnetic stirring, and the mixture was evenly mixed after stirring for 10 min. To the solution, 2 g CH₃COONa was added, and the stirring was continued for 30 min. Then, the mixture was transferred to a 50-mL capacity Teflon-lined stainless steel reactor and kept in the oven at 180 °C for 12 h. After the reactor is cooled to room temperature, the red product is separated by centrifugation, washed several times with distilled water and countless ethanol by suction filtration method, and dried in the oven at 60 °C for more than 6 h to obtain dry dark red powder. The Cu-doped α-Fe₂O₃ is denoted as x% Cu–Fe₂O₃ according to the molar ratio of Cu doping. Six samples with doping ratios of 1%, 3%, 5%, 7%, 10%, and 15% were synthesized, among which 7% Cu–Fe₂O₃ had the best degradation effect. The synthesis of α-Fe₂O₃ is basically the same as that of x% Cu–Fe₂O₃, except that CuCl₂·2H₂O is not added.

H₂O₂ Photoactivation Degradation Experiment.

In a representative experiment of photoactivation of H₂O₂, 20 mg of the catalyst was added to a 100-mL configuration of aqueous pollutant solution and sonicated to disperse it evenly. First, the suspension was stirred for 30 min to ensure that the adsorption equilibrium was established. Second, 100 μL of H₂O₂ (30% wt) solution was added to the suspension, and a 300-W xenon lamp equipped with a 420-nm cutoff filter was used as the light source. At certain irradiation intervals, 3 mL of the suspension was extracted, and its degradation was measured after filtration. CIP and degraded intermediates were determined by high-performance liquid chromatography–mass spectrometry (HPLC-MS, Orbitrap Fusion, Thermo, USA) with the eluent consisting of 0.1% formic acid and acetonitrile at a flow rate of 0.2 mL/min.

Quantitative Experiments of Free Radicals.

Determination of free •O₂⁻ concentration.

Yellow nitroblue tetrazolium (NBT, with the reduction potential of −0.05 V) was reduced by •O₂⁻ (with the reduction potential of −0.33 V) to blue formazan, and 1 mol NBT can react with 4 mol •O₂. We quantified the concentration of generated •O₂⁻ in the suspension by recording the residual concentration of NBT on a UV-vis spectrophotometer (maximum absorbance at 259 nm).

Determination of free •OH concentration.

Free •OH were determined via PL according to the reaction of terephthalic acid with hydroxyl radicals that produces 2-OH-terephthalic acid with a fluorescence peak located at 426 nm after excitation at 312 nm. Measurement of fluorescence intensity was performed on a FluoroMax-P spectrophotometer.

Computational Details.

We carried out all the DFT calculations in the Vienna ab initio simulation (VASP5.4.4) code. The exchange–correlation is simulated with PBE functional, and the ion–electron interactions were described by the PAW method. The vdWs interaction was included by using the empirical DFT-D3 method. The α-Fe₂O₃ (110) and 7% Cu–Fe₂O₃ (110) surface was used to catalyze the reduction of H₂O₂. Atoms in the upper two layers of the surface are allowed to move freely while the bottom two layers are fixed to simulate the surface of the structure. The Monkhorst–Pack grid mesh–based Brillouin zone k-points are set as 2 × 2 × 1 for all periodic structure with the cutoff energy of 450 eV. The convergence criteria are set as 0.01 eV A⁻¹ and 10⁻⁵ eV in force and energy, respectively. The free energy calculation of species adsorption (ΔG) is based on the following model: ΔG = ΔE + ΔE_ZPE + ΔH_0→T – TΔS. Herein, ΔE, ΔE_ZPE, and ΔS, respectively, represent the changes of electronic energy, zero-point energy, and change in entropy. The ΔH_0→T refers to the change in enthalpy when heating from 0 K to T K.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

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