Low-temperature purification of phosphine (PH₃) using CuO@NC sorbents: simultaneous pollutant removal and Cu₃P resource recovery

Abstract

Phosphine (PH₃) is a highly toxic industrial gas widely used in semiconductor manufacturing and fumigation, posing serious environmental and health risks even at low concentrations. Efficient low-temperature removal systems are urgently needed for safe handling and emission control. This study presents a novel CuO@NC-doped sorbent synthesized via direct thermal treatment of urea and copper nitrate for the purification of PH₃ at 90 °C. N and C doping significantly enhanced basic sites and oxygen vacancies, enabling effective capture of acidic and reducing PH₃ gas. The optimized sorbent achieved a breakthrough capacity of 272.54 mg (PH₃) per g, substantially outperforming previously reported materials. Beyond its practical advantages of simple preparation and low cost, this sorbent offers an innovative pathway for resource utilization—the spent material transforms into valuable Cu₃P, which exhibits promising catalytic, bactericidal and photocatalytic properties. This approach demonstrates simultaneous pollutant removal and resource recovery through low-temperature PH₃-CuO@NC reactions.

---

Green foundation

1. A sustainable and green chemistry strategy was implemented to design high-performance PH₃ adsorbents. A novel N- and C-doped CuO@NC material was synthesized via a simple and cost-effective route, exhibiting exceptional PH₃ adsorption capacity (272.54 mg PH₃ per g), significantly surpassing previously reported sorbents.

2. The spent adsorbent (De-2CuO@NC) demonstrates outstanding photocatalytic degradation and bactericidal activity, enabling its transformation into a dual-functional material for environmental remediation.

3. This integrated approach achieves efficient PH₃ capture, safe disposal, resource recovery (e.g., Cu₃P conversion), and waste valorization, uniting environmental protection with economic benefit. The findings provide a green framework for the rational reuse of functional materials, aligning with the principles of green chemistry and environmental stewardship.

1. Introduction

PH₃ is a flammable, corrosive, and high bio-toxicity gas that poses a serious threat to human health and the ecological environment even at low concentrations.^1–3 Human exposure to PH₃ will cause a series of different non-specific symptoms, which are characterized by high mortality. However, a comprehensive mechanism for PH₃ poisoning is still unclear, and no effective antidote or treatment has been found yet.⁴ Additionally, PH₃ is considered a potential greenhouse gas because it competes with atmospheric hydroxyl radicals (˙OH), thereby suppressing the oxidative removal of other greenhouse gases and indirectly enhancing the greenhouse effect.⁵ According to statistics, in Yunnan Province, China, the tail gas released by the yellow phosphorus production industry in 2023 contained about 467.55–1215.63 tons of PH₃.^6,7 Direct emissions of PH₃ can also originate from sources such as marsh gas, biogas, landfill gas, exhaust gases from semiconductor and optoelectronic manufacturing, as well as tail gases from enclosed calcium carbide furnaces.^8–10 Unfortunately, the problem of PH₃ emissions in the past two decades has not attracted attention in some areas, especially in Latin America, Asia, and Africa.¹¹ Accordingly, efficient removal of PH₃ from tail gases is necessary.

Yellow phosphorus tail gas (YPTG), a major source of PH₃, is characterized by its relatively low temperature (approximately 100 °C) and low oxygen content (around 1%).¹² Effectively and stably eliminating PH₃ under low-temperature and low-oxygen conditions remains a pressing challenge. In recent years, the adsorption–oxidation approach has gained attention as a promising strategy for PH₃ removal from YPTG, owing to its advantages such as controllable operation, mild reaction conditions, high stability, and minimal generation of liquid waste.¹² However, despite these merits, adsorption–oxidation technologies have not yet been widely implemented in practical PH₃ purification due to their high cost and, in some cases, inadequate PH₃ breakthrough capacities of sorbents. Therefore, non-precious-metal-based sorbents, such as CuO and Fe₂O₃, have been extensively investigated.^13,14

In general, the strong reducing nature and low dipole moment (0.58 D) of PH₃ render its chemisorption on sorbent surfaces more favorable than physisorption.⁶ This suggests that, in designing sorbents, priority should be given to materials capable of forming strong chemical interactions with PH₃. Among various metallic oxides, copper oxide, especially copper-supported materials, is widely used as a sorbent in the adsorption–oxidation process.¹⁵ However, the requirement of sorbents to capture PH₃ from YPTG is more stringent than that to capture PH₃ from other known systems. First, the sorbents should have excellent adsorption selectivity towards PH₃.¹⁶ Second, the sorbents should have a relatively high capture performance for high concentration PH₃ (∼1000 ppm) under low temperature (100 °C) and low oxygen (∼1%) conditions, which is the typical characteristic of YPTG.¹² Third, sorbents should be economically viable and possess good reusability. Notably, CuO/TiO₂ and Ce₁Cu₃₀/HZSM-5 have recently been reported to exhibit outstanding PH₃ adsorption capacities, with breakthrough values reaching 108.48 mg (PH₃) per g_sorbent and 114.36 mg (PH₃) per g_sorbent, respectively, highlighting their potential for practical application in PH₃ removal.^6,12 Nonetheless, the relatively high cost and poor regeneration performance of these sorbents remain significant barriers to their large-scale application. Incorporating activated carbon (AC) as a support material offers a cost-effective alternative. However, its limited ion-exchange capacity and irregular pore structure hinder the effective loading and uniform dispersion of active species such as CuO.^17,18 As a result, most reported CuO_X/AC sorbents could only obtain limited PH₃ breakthrough capacity in the range of 20–130 mg (PH₃) per g_sorbent. In addition, most of the deactivated PH₃ sorbents cannot be regenerated or reused and cause secondary pollution. In this context, there is a strong demand for cost-effective, renewable, and reusable sorbents with high PH₃ breakthrough capacities. Moreover, spent sorbents should possess efficient regeneration or reutilization capabilities. Recently, it has been reported that N and C doping could significantly modify the physicochemical and electronic properties of sorbents and regulate the activation process of reactants.^19,20 N and C doped Cu-based materials are considered to be ideal sorbents for PH₃ capturing due to their abundant oxygen vacancies and relatively larger surface area than CuO nanoparticles and offer shorter gas diffusion pathways, all of which facilitate PH₃ adsorption, migration, and transformation.

Motivated by the above, a CuO@NC sorbent was synthesized in this study for efficient PH₃ removal under low-temperature and low-oxygen conditions. As expected, CuO@NC sorbent exhibits excellent PH₃-fixation performance. Remarkably, we observed that the CuO@NC sorbent converted into high-purity Cu₃P upon deactivation. Cu₃P is well recognized as a valuable P-type semiconductor extensively applied in photocatalysis, including applications such as photodegradation of organic contaminants and photocatalytic hydrogen production.^20–22 Interestingly, recent studies have demonstrated that copper-based materials rich in oxygen vacancies exhibit remarkable efficacy in bacterial inactivation. Upon modification with oxygen vacancies, these materials have been shown to significantly enhance their interaction with oxygen, thereby leading to the accumulation of reactive oxygen species (ROS). The ROS generated are capable of effectively disrupting the bacterial cell wall and causing lipid peroxidation of the cell membrane, ultimately resulting in cell death. This discovery highlights the potential of oxygen-vacancy-enriched copper-based materials as a promising antimicrobial agent, with implications for applications in water treatment and healthcare.^23,24 However, the synthesis of Cu₃P is known to be a rather complex and costly process. Our study may offer a simplified strategy for the synthesis of Cu₃P. To the best of our knowledge, there have been no reports on the synthesis of Cu₃P from PH₃ and CuO under mild conditions. Additionally, this study evaluates the performance of deactivated adsorbents in photocatalytic and antibacterial systems, which holds dual significance for environmental protection and sustainable utilization in practical applications.

2. Results and discussion

2.1 Synthesis and performance of sorbents

N and C-doped copper-based materials (XCuO@NC) are obtained by mixing copper nitrate and urea for calcination (Appendix S1, SI). The selection of the copper nitrate-to-urea ratio was guided by the performance evaluation results of the sorbents. The CuO particles obtained by roasting copper nitrate separately are used as the control group for testing. The deactivated sorbents are denoted as De-CuO@NC. To investigate the thermal evolution during the synthesis of the sorbent, thermogravimetric analysis (TGA) was performed. As shown in Fig. S1, the TG and DTG curves of the 2CuO@CN sample exhibit three distinct weight-loss stages. The first stage corresponds to the desorption of adsorbed water, the second stage is attributed to the decomposition of Cu₂(OH)₂CO₃, and the third stage is associated with the reduction of CuO. Additionally, the details of characterization methods for sorbents are presented in Appendix S2.

Fig. 1a presents the XRD patterns of sorbents synthesized with varying copper nitrate-to-urea ratios. By adjusting the copper nitrate content in the precursor, several characteristic diffraction signals attributed to CuO are clearly observed in the resulting sorbents (2θ = 32.5°, 38.9°, and 48.8°, PDF# 80-0076); this indicates that CuO serves as the primary active component in the resulting adsorbent, which is further proved from the Cu 2p3/2 photoelectron spectra of CuO particles and 2CuO@NC (Fig. S4). At a copper nitrate-to-urea ratio of 1 : 6 (1CuO@NC), two weak characteristic peaks assigned to Cu₂O were observed at 2θ = 36.5° and 42.4° (PDF# 77-0199),²⁵ indicating that the high urea content in the precursor may cause partial reduction of CuO during calcination. Furthermore, the formation of crystalline Cu₂O impurity phases was not detected in all XCuO@NC samples. The shapes of the N₂ adsorption/desorption isotherms of various sorbents obtained with different copper nitrate-to-urea ratios were similar (Fig. 1b). According to the IUPAC classification, all samples exhibit typical type I adsorption isotherms accompanied by an H4 hysteresis loop, indicating that the pores in these XCuO@NC sorbents are predominantly microporous, primarily originating from cracked or stacked pore structures.^26,27 In addition, the BET surface area (S) and total pore volume (V_total) of the yield of the obtained sorbents decreased as the copper nitrate-to-urea ratio in the precursor increased (Table S1). As shown in Fig. S2, all samples exhibit mesoporous structures with pore sizes mainly distributed between 2 and 8 nm. It is worth noting that the S and V_total of these CuO@NC samples are much higher than those of CuO particles, which can be attributed to the improved adsorbent pore structure by N and C doping.

Fig. 1 (a) XRD patterns. (b) N₂ adsorption/desorption isotherms of the sorbents with different Cu contents. The desorption curve is shown as a gray line. (c) Breakthrough curves, and (d) breakthrough capacity of the various sorbents.

Comprehensive details of the sorbent performance testing system are presented in Appendix S3. Typically, the PH₃ adsorption performance of various synthesized sorbents was evaluated in a quartz tube reactor at 90 °C under low-oxygen conditions (1% O₂), reflecting the typical oxygen concentration found in YPTG. Generally, the PH₃ in the highly toxic inlet gases can be removed by the sorbent through adsorption–oxidation, so that the outlet gas meets the emission standard. The PH₃ capture performance of all samples gradually declined over time, ultimately reaching breakthrough, defined as an outlet PH₃ concentration exceeding 30 ppm (Fig. 1c). As anticipated, CuO particles are capable of removing PH₃; however, the PH₃ breakthrough capacity is limited to only 50.43 mg (PH₃) per g_sorbent (Fig. 1d), which is attributed to its small S. In addition, the adsorption capacity of PH₃ exhibits a volcano-shaped trend as the ratio of copper nitrate to urea in the precursor increases. The optimal copper nitrate-to-urea ratio for the sorbent was determined to be 2 : 6 (2CuO@NC), exhibiting a breakthrough time of up to 540 minutes and a PH₃ breakthrough capacity of 272.54 mg (PH₃) per g_sorbent. Table 1 summarizes the low temperature (<150 °C) PH₃ adsorption capacities of selected sorbents documented in the literature. As shown, the PH₃ breakthrough capacity of 2CuO@NC significantly surpasses that of other sorbents. Notably, the CuO@NC sorbent benefits from a straightforward and stable synthesis process, distinguished by its affordability and brief synthesis duration. Furthermore, its exceptional performance was demonstrated under stringent conditions (temperature: 90 °C, PH₃ concentration: 1000 ppm, weight hourly space velocity: 20 000 h⁻¹), highlighting its potential for practical applications. Collectively, these results underscore the superior capabilities of the synthesized CuO@NC sorbent.

Table 1 A summary of PH₃ breakthrough capacity for some sorbents

Sorbents	Gas composition and WHSV (h^−1)	PH3 breakthrough concentration (ppm)	PH3 adsorption capacities (mg g^−1)	Ref.
Cu15Zn1/AC	N2 + 1180 ppm PH3 + 1% O2; 5000 h^−1	118	45.32	28
Cu/Al2O3	N2 + 10 000 ppm PH3	100	65	29
Ce1Cu30OX/HZSM-5	N2 + 800 ppm PH3 + 1% O2; 15 000 h^−1	320	114.36	12
Cu/SBA-15	N2 + 200 ppm H2S + 800 ppm PH3 + 0.5% O2; 10 000 h^−1	320	104.8	13
Cu–Fe/SBA-15	N2 + 200 ppm H2S + 800 ppm PH3 + 0.5% O2; 10 000 h^−1	320	120.1	14
Cu/TiO2	Air + N2 + 500 ppm PH3 + H2O (RH = 47%)	500	108.48	6
Ce–Cu–Al@[N71.1]	N2 + 300 ppm H2S + 600 ppm PH3 + 1% O2; 36 000 h^−1	600	201.9	30
Cu0.15/ACF	N2 + 300 ppm H2S + 600 ppm PH3 + 0.5% O2; 10 000 h^−1	600	132.1	31
CuO/AC	N2 + 231 ppm PH3 + 1.6% O2; 750 h^−1	3.3	96.1	32
2CuO@NC	N2 + 1000 ppm PH3 + 1% O2; 20 000 h^−1	30	272.54	This work

---

2.2 Characterization of sorbents

The SEM images of CuO particles alongside those of 2CuO@NC are displayed in Fig. 2(a and b). As shown in Fig. 2a, the CuO particles consist of tightly agglomerated small solid particles and the pore structure was uniform. By contrast, 2CuO@NC showed a different morphology, which consisted of numerous regular and orderly spherical particles (Fig. 2b). Hollow structures in spherical particles were also observed, beneficial for increasing the specific surface area and exposing more active sites of the sorbent. To analyze the elemental composition and distribution of 2CuO@NC, SEM–EDS and XPS characterization studies were performed. As shown in Fig. 2c and S5, Cu, O, N, and C are relatively uniformly distributed on the material surface. Nitrogen was not clearly detected by EDS, likely due to the small mapping area (∼4 μm²) and the limited sensitivity of SEM–EDS for light elements. XPS analysis (Table S4) confirms the presence of nitrogen, although its atomic percentage is low. It should be noted that the quantification of carbon is influenced by surface contamination in XPS, and the use of conductive carbon during EDS testing may lead to overestimation of its content; therefore, accurate determination of carbon is not feasible. Overall, the combined mapping and XPS compositional analyses indicate that C and N doping was successfully achieved during sorbent synthesis. High-resolution TEM (HRTEM) analysis and the corresponding FFT patterns exhibited a lattice fringe spacing of 0.23 nm, which matches the (111) crystallographic plane of highly ordered CuO, thus verifying the synthesized CuO (Fig. 2d).^25,27

Fig. 2 SEM micrographs of (a) CuO particles and (b) 2CuO@NC. (c) EDS mapping images (from Fig. S3) showing the elemental distribution of Cu, O, C, and N in the 2CuO@CN sample. (d) HRTEM images of 2CuO@NC.

Changes in surface chemical functionalities observed on CuO particles and 2CuO@NC were investigated by FTIR analysis (Fig. 3a). For CuO, a single broad and intense peak centered around ∼531 cm⁻¹ was observed, attributed to the characteristic vibrational modes of the CuO group.³³ For the 2CuO@NC samples, a series of peaks centered at 784.03, 1461, 1505, and 1676 cm⁻¹ corresponded to the characteristic vibrations of –NH₂, C=N, C=C, and C=O, respectively,^33–36 indicating that the 2CuO@NC was successfully achieved by doping of N and C into CuO. A wide-survey XPS spectral scan of the two samples confirmed the presence of C, Cu, O, and N elements, as evidenced by the data shown in Fig. S6. The N 1s and C 1s photoelectron spectra (Fig. 3b and S7) provide further insight into the surface chemical composition of the samples. The N 1s spectrum of CuO@NC can be deconvoluted into two peaks located at 397.9 eV (C–N=C) and 399.7 eV (N–(C₃)),³⁷ corresponding to pyridinic and graphitic nitrogen species, respectively. In addition, the C 1s spectrum of CuO@NC exhibits a much stronger signal than that of CuO particles. Even after subtracting the background carbon signal observed in CuO, a prominent C 1s intensity remains, confirming the presence of carbon and nitrogen species in the CuO@NC material, which is consistent with the EDS results.

Fig. 3 (a) FTIR spectra, (b) N 1s spectra, (c) CO₂-TPD profiles, (d) EPR spectra, (e) H₂-TPR profiles, and (f) O 1s spectra.

The quantity and strength of surface basic sites on CuO particles and 2CuO@NC were evaluated using CO₂-TPD. In the CO₂-TPD profiles, the desorption peak area reflects the number of basic sites, while the peak temperature reflects the strength of these interactions.^34,35,38 For CuO particles, distinct desorption peaks can be observed at approximately 139 °C and 430 °C (Fig. 3c, pink line), which are attributed to weak basic hydroxyl groups and medium-strength Cu–O sites, respectively. In contrast, 2CuO@CN (turquoise line) shows the disappearance of the weak basic sites observed previously, while new desorption peaks emerge at around 377 °C and 526 °C, which are attributed to Cu–O-related basic sites and strong basic sites associated with oxygen vacancies (V_O), respectively.³⁹ Based on the analysis of the sorbent performance results (Fig. 1d), it can be inferred that increased basicity plays a pivotal role in enhancing the sorbent efficiency. As anticipated, N-doping significantly enhanced both the number and strength of basic sites on the sorbent, thereby promoting the adsorption of PH₃ (acid gas) onto the sorbent surface during the adsorption–oxidation process.

EPR spectroscopy was applied to confirm the presence of surface V_O.³⁴ As shown in Fig. 3d, CuO particles and 2CuO@NC samples revealed a peak located at a g-value of 2.08, which is attributed to V_O.⁴⁰ However, the EPR signal of 2CuO@NC was much stronger than that of CuO particles, implying that 2CuO@NC had a higher V_O concentration, which is beneficial for capturing the reducing gas PH₃. The results of H₂-TPR also confirmed the above speculation (Fig. 3e). Although the reduction peak positions of 2CuO@NC and CuO particles are similar, the slightly larger peak area observed for 2CuO@NC suggests a higher concentration of surface-active oxygen species. The O 1s photoelectron spectra (Fig. 3f) of the samples were deconvoluted into three distinct peaks located at 530.1 ± 0.3 eV (lattice oxygen, O_α), 531.8 ± 0.3 eV (V_O, O_β), and 533.6 ± 0.3 eV (adsorbed H₂O, O_γ),²⁵ respectively. For 2CuO@NC, the signal intensities of O_β and O_α are both higher than those of CuO particles. Specifically, the intensity of O_β increases by 20%, while that of O_α increases by 50%. This result indicates that the surface N and C species in 2CuO@NC, together with its unique microstructure, generate a higher intensity of basic oxygen species (O_α) and V_O compared with CuO particles. In general, the presence of abundant lattice defects, vacancies, and unsaturated bonds facilitates the formation of surface oxygen species.^40,41 These lattice defects, vacancies, and unsaturated bonds are good active sites for adsorption–oxidation reactions.^42–45

2.3 Analysis of De-2CuO@NC

As shown in Fig. S8 and Table S2, the adsorption–desorption isotherm of the spent sorbent De-2CuO@CN still exhibits a typical IUPAC H4 hysteresis loop, indicating that its pore structure remains dominated by slit-like micropores formed through sheet stacking or partial structural collapse. The BET surface area (1.95 m² g⁻¹) and total pore volume (0.014 cm³ g⁻¹) are significantly reduced compared with the fresh sample, suggesting pore blockage and structural densification during the PH₃ adsorption–oxidation process. Moreover, the pore size distribution in Fig. S9 shows that De-2CuO@CN possesses an average pore size of approximately 3.1 nm, with most pores still falling within the mesoporous range of 2–10 nm. These results clearly demonstrate that the removal of phosphine leads to substantial declines in surface area, pore volume, and pore size, indicating pronounced degradation of the porous structure. XRD and XPS analyses of the fully deactivated sorbents (De-2CuO@NC, PH₃ removal efficiency <5%) were conducted to elucidate the primary reaction products and uncover the underlying deactivation mechanism. A photograph showing the color transition from pure black to metallic silver-gray after sorbent deactivation is presented in Fig. 4a. By comparing the XRD patterns of 2CuO@NC and De-2CuO@NC, it is evident that sorbent deactivation results in the disappearance of diffraction peaks attributed to CuO, indicating that CuO was consumed as the active phase during the reaction. In addition, the XRD pattern of De-2CuO@NC exhibited a sequence of characteristic diffraction peaks corresponding to Cu₃P (2θ = 36.19°, 45.18° and 46.21°, PDF# 71-2261), and no additional diffraction peaks corresponding to other crystalline phases were observed, indicating that Cu₃P is the predominant component in De-2CuO@NC and possesses high crystallinity. Surface composition and oxidation state analyses of De-2CuO@NC were further investigated through XPS analysis. A wide-survey XPS spectral scan of the De-2CuO@NC sample demonstrated the presence of C, Cu, O, and P elements (Fig. S10). As shown in Fig. 4b, the Cu 2p_3/2 photoelectron spectrum of the 2CuO@NC sample can be fitted with two peaks located at 933.5 ± 0.3 eV, along with a prominent Cu²⁺ satellite peak at 942.5 ± 0.3 eV, which are characteristic of the binding energies associated with CuO. These peaks correspond to the characteristic binding energies (BEs) of CuO.^38,46 For the De-2CuO@NC sample, the characteristic peaks associated with CuO disappeared, and an additional peak appearing at 931.7 ± 0.2 eV emerged, which can be attributed to the BE of Cu⁺ (Cu₃P),⁴⁷ which aligns well with the XRD findings. As revealed by the P 2p spectral region of De-2CuO@NC, a characteristic peak positioned at 129.5 ± 0.5 eV was observed and assigned to the BE of Cu₃P (Fig. 4c),⁴⁸ while another peak at 134.2 ± 0.5 eV was attributed to PO₃^−49,50species. We observed a ratio of 0.76 : 0.24 for the PO₃⁻ : Cu₃P species, which is attributed to the introduction of O₂ during the PH₃ adsorption process. This led to the inevitable formation of PO₃⁻ species on the surface of the adsorbent. Since XPS is primarily surface-sensitive, the PO₃⁻ species are detected in higher abundance. However, the bulk phase predominantly consists of Cu₃P, as confirmed by XRD analysis. The O 1s photoelectron spectra of the 2CuO@NC and De-2CuO@NC samples revealed that the O_α and O_β were consumed after the deactivation of the adsorbent, and O_γ was generated (Fig. S11). These phenomena could be ascribed to the multifaceted chemical process involving CuO, O₂, and PH₃, which eventually converts the reactants into Cu₃P, H₂O, and phosphorus oxides and/or phosphates, accompanied by the contribution of V_O and active oxygen species.

Fig. 4 (a) XRD patterns, (b) Cu 2p spectra, and (c) P 2p spectra of fresh and deactivated 2CuO@NC. (d and e) SEM micrograph, (f) EDS mapping image of the elemental distribution of Cu, O, and P, (g) EDS element analysis, and (h and i) HRTEM images of De-2CuO@NC.

The SEM images are shown in Fig. 4d and e, displaying the characteristic morphology of De-2CuO@NC. Compared with the numerous regular and orderly spherical particles of the 2CuO@NC sample (Fig. 2c), the De-2CuO@NC sample lacks a well-defined geometric morphology and is primarily composed of dense, large particles with relatively smooth surfaces. High-magnification SEM images further reveal that these larger aggregates are formed from numerous smaller, smooth-surfaced particles (Fig. 4e). Fig. 4f and S12 demonstrate a uniform distribution of Cu, O, C, N and P across the surface of the De-2CuO@NC sample. As shown in Fig. 4g, the Cu : P elemental ratio of the large particles is close to 3 : 1, indicating that they are predominantly composed of high-purity Cu₃P. Although a precise correlation between PH₃ adsorption and the amount of Cu participating in the reaction cannot be established due to the lack of exact compositional information for 2CuO@NC, we provide an indirect quantification by measuring the mass change of the adsorbent before and after the reaction (Fig. S13), with the sorbent mass increasing by 39.8 mg. Furthermore, HRTEM images and FFT patterns clearly reveal lattice fringes in the De-2CuO@NC sample, exhibiting the measured lattice spacing of 0.2 nm that matches the (300) plane of highly crystalline Cu₃P (Fig. 4h and i).²² The aforementioned findings further substantiate the transformation of the active CuO phase into Cu₃P following the deactivation of 2CuO@NC during PH₃ removal.

2.4 Photocatalytic activity of De-2CuO@NC

Interestingly, the deactivated sorbent contains high-purity Cu₃P, which finds extensive applications in photocatalysis.⁵¹ Therefore, theoretically, the De-2CuO@NC should possess photocatalytic properties. The photosensitivity of photocatalytic materials is an important factor in evaluating the catalytic activity of catalysts. Accordingly, photoluminescence (PL) spectra and UV–visible diffuse reflectance spectra (UV–vis DRS) of De-2CuO@NC and g-C₃N₄ were employed to investigate charge migration, transfer, and recombination processes in the photocatalysts, as well as to analyze their light absorption properties. Fig. 5a shows the PL emission spectra of De-2CuO@NC and g-C₃N₄. Distinct PL emission features are observed for the two samples. De-2CuO@NC (Cu₃P) exhibits a lower PL emission peak intensity than that of g-C₃N₄, which is consistent with previous literature reports.³⁸ Generally, a lower PL emission intensity corresponds to more efficient separation of photogenerated charge carriers, thereby minimizing carrier recombination.^52–55 This indicates that De-2CuO@NC (Cu₃P) has better photocatalytic activity (photo-generated carrier separation efficiency) than g-C₃N₄. We further investigated the band structure difference between De-2CuO@NC and g-C₃N₄. As shown in UV-vis DRS (Fig. 5b), De-2CuO@NC exhibits stronger absorption in both the ultraviolet and visible regions than g-C₃N₄ throughout the full spectrum, which is beneficial for the trapping of visible light by De-2CuO@NC, resulting in an excellent photocatalytic activity.⁵¹ In addition, the band gap energies (E_g) of De-2CuO@NC (Cu₃P) and g-C₃N₄ were calculated separately using Tauc plots. It can be seen from Fig. 5(c and d) that the E_g of De-2CuO@NC (1.41 eV) is lower than that of g-C₃N₄ (2.73 eV), which is helpful for the absorption of low-energy photons by De-2CuO@NC. The efficiency of photogenerated electron–hole pair separation serves as an important indicator for evaluating the photocatalytic activity of a sample. As shown in Fig. 5e, the intermittent illumination cycles (I–T curves) of CuO, De-2CuO@NC (Cu₃P), and g-C₃N₄ were collected using an electrochemical workstation. The time-dependent photocurrent densities of De-2CuO@NC (∼0.55 μA cm⁻³) and g-C₃N₄ (∼0.5 μA cm⁻³) are significantly greater than that of CuO (∼0.2 μA cm⁻³), indicating that the De-2CuO@NC (Cu₃P) exhibits improved charge carrier separation efficiency over CuO and g-C₃N₄. In addition, the photocurrent density exhibited negligible decay after multiple on/off cycles, demonstrating that De-2CuO@NC possesses excellent cyclic stability. Furthermore, the photocurrent density remained stable after repeated on/off cycles, demonstrating that De-2CuO@NC exhibits excellent photoelectrical response stability.

Fig. 5 (a) Photoluminescence spectra; (b) UV-vis diffuse reflectance spectra; (c and d) Tauc plots of g-C₃N₄ and De-2CuO@NC; (e) transient photocurrent responses; and (f) photocatalytic degradation curves of RhB under visible light (>420 nm) irradiation of various materials.

Photocatalytic activity tests were conducted on De-2CuO@NC (Cu₃P) and g-C₃N₄ for rhodamine B (RhB) degradation and Hg⁰ (gas) removal. Before the photocatalytic performance test, the De-2CuO@NC was washed several times with ultrapure water and then dried for use. Experimental procedures for evaluating the photocatalytic performance can be found in Appendix S4. Fig. 5f illustrates the photocatalytic degradation behavior of RhB under visible light exposure (>420 nm). In the absence of the photocatalyst, RhB exhibited minimal degradation due to direct photolysis (represented by the purple line). After 60 minutes of dark adsorption, the system reached adsorption–desorption equilibrium. During 120 minutes of visible light irradiation, De-2CuO@NC demonstrated effective photocatalytic degradation, with the degradation rate increasing from approximately 17% to complete removal. This result exceeds that of g-C₃N₄ (degradation rate from ∼10 to 94%), which proves that Cu₃P has better photocatalytic activity than g-C₃N₄ for the degradation of RhB. Meanwhile, it should be noted that the activities of CuO and 2CuO@NC for photocatalytic degradation of RhB are negligible. Table S3 summarizes the performance of some reported photocatalysts for photocatalytic degradation of RhB, and it can be seen that the activities of the deactivated sorbents in this study are close to or even higher than these specially designed photocatalysts. To further investigate the structural changes of the material after the photocatalytic degradation of RhB, XRD and ICP-OES characterization studies were conducted. The results (Fig. S14 and S15) show that the material maintains an intact Cu₃P structure. ICP-OES analysis indicates that the Cu leaching is approximately 0.016 mg, accounting for less than 1% of the total amount and remaining almost unchanged over time, while no P leaching is detected, suggesting that the photocatalytic process does not cause irreversible structural changes. Additionally, De-2CuO@NC demonstrated outstanding photocatalytic activity for Hg⁰ removal under UV light irradiation at 254 nm. As illustrated in Fig. S16, negligible direct photolysis of gaseous Hg⁰ occurs in the absence of the photocatalyst, demonstrating the intrinsic stability of Hg⁰ under UV light exposure. When De-2CuO@NC was employed as the photocatalyst, prior to UV light exposure (light off), an adsorption–desorption equilibrium was established between gaseous Hg⁰ and the deactivated sorbent. Upon UV illumination, the Hg⁰ concentration in the outlet gas rapidly decreased, dropping from approximately 1100 μg m⁻³ to about 45 μg m⁻³, corresponding to a maximum removal efficiency of approximately 95.5%.

As shown in Fig. S11, abundant V_O are still preserved in the De-2CuO@CN sample, and previous studies have demonstrated that such vacancies play a pivotal role in enhancing the photocatalytic performance of Cu₃P-based materials.^56,57 Under light irradiation, Cu₃P generates electron–hole pairs, while the oxygen vacancies act as electron-trapping centers to effectively inhibit the recombination of photogenerated charge carriers. Meanwhile, the vacancies serve as highly active sites for O₂ adsorption and activation, thus accelerating the formation of ROS,⁵⁸ which is consistent with the ROS generation observed during the antibacterial process, as shown in Fig. 7a. These ROS, including ˙O₂⁻ and ˙OH, participate in the oxidative degradation of RhB molecules, leading to their gradual mineralization into CO₂, H₂O, and smaller intermediates. For gaseous Hg⁰ removal, oxygen vacancies facilitate the enrichment of Hg⁰ on the catalyst surface and promote electron transfer to surface-adsorbed oxygen species,⁵⁹ thereby accelerating the oxidation of Hg⁰ into stable Hg²⁺. The produced Hg²⁺ can then be firmly immobilized on the catalyst surface, preventing secondary pollution. Therefore, V_O not only improves the separation and migration of photogenerated carriers, but also significantly strengthens the redox capability of Cu₃P, endowing De-2CuO@CN with excellent photocatalytic performance toward both organic dye degradation and detoxification of hazardous mercury vapor.

2.5 Bactericidal effect of De-2CuO@NC

The high antibacterial efficiency of metal phosphides mainly arises from their unique electronic structure and surface chemical properties. On one hand, the surface of metal phosphides contains abundant active sites that can catalyze the generation of ROS, such as ˙O₂⁻, from oxygen; these ROS can penetrate bacterial cell walls, elevate intracellular ROS levels, induce lipid peroxidation, and thereby compromise the integrity of the cell membrane.⁶⁰ On the other hand, metal phosphides possess high electronic conductivity and surface reactivity, enabling continuous redox cycling in aqueous environments, which further enhances their catalytic antimicrobial activity. Therefore, phosphide-based antibacterial strategies hold significant research value and promising prospects in microbial control, medical protection, and environmental remediation. Based on previous research, copper-based materials containing oxygen vacancies are widely applied in the field of sterilization. Consequently, we reasonably propose the hypothesis that De-2CuO@NC possesses bactericidal properties.^61–63 To validate this hypothesis, the bactericidal effects of De-2CuO@NC were tested on specific strains of Escherichia coli (E. coli) and H1N1 influenza virus. To assess the bactericidal efficiency, this study evaluated the survival ratio of bacteria and virus to determine the bactericidal effect of De-2CuO@NC on E. coli. Additionally, the cytotoxicity of the material was assessed, and a preliminary investigation into the bactericidal mechanism of De-2CuO@NC was conducted. Details of the De-2CuO@NC bactericidal experiments are presented in Appendix S5, SI.

Fig. 6(a and b) show the inactivation effect of different concentrations of De-2CuO@NC on E. coli and compare it with the inactivation effects of commonly used bactericidal materials on E. coli; there was an 8-log decrease in E. coli survival number after treatment with De-2CuO@NC (15 mg L⁻¹) for 30 min. The bactericidal activity of De-2CuO@NC is better than that reported previously and is similar to that of Ag₂O.^64–66 Accounting for the exclusion of Cu²⁺ leaching effects, De-2CuO@NC was dispersed into different concentrations with deionized water and stirred at 25 °C. At the time intervals of 10 min and 30 min, 1 mL liquid was taken and centrifuged followed by ICP-OES detection of the supernatant. Fig. 6(c and d) show the Cu²⁺ elution during the bactericidal process and the inactivation effect of Cu²⁺ (2 mg L⁻¹) on E. coli. It could be seen that the leakage amount of Cu²⁺ increased with the increase of De-2CuO@NC concentration and slightly increased with time. The largest leakage amount of Cu²⁺ was about 2 mg L⁻¹ when 15 mg L⁻¹ De-2CuO@NC was used. In order to determine the contribution of Cu²⁺, the bactericidal activity of 2 mg L⁻¹ Cu²⁺ was examined. By comparing the bactericidal activity of 15 mg L⁻¹ De-2CuO@NC with that of 2 mg L⁻¹ Cu²⁺, it can be noted that the effect of 2 mg L⁻¹ Cu²⁺ on the bactericidal activity is very weak, and thus the toxic effect of copper ions released from De-2CuO@NC on E. coli can be ignored. This study utilized EPR spectroscopy to reveal the role of extracellular ROS, malondialdehyde (MDA) formation to assess the cellular lipid peroxidation performance, magnetic field-induced fluorescence (FL) to quantify intracellular ROS levels, and cytotoxicity assays to delve into the mechanistic understanding of cellular death in E. coli. The EPR spectra in Fig. 7a show that the De-2CuO@NC was added into DMPO solution; the characteristic peak of superoxide free radicals appeared, while ˙OH radicals were not detected, which indicates that ˙O₂⁻ was responsible for bactericidal activity, implying a catalytic process of O₂ activation.^41,67Fig. 7b shows the effects of various concentrations of De-2CuO@NC on MDA formation in the bactericidal process, confirming that De-2CuO@NC caused lipid peroxidation of the E. coli cell membrane, indicating that the oxidative damage of the cell membrane is the direct reason for E. coli cell death.^68–70 This article provides photographs of fluorescence microscopy of E. coli treated with 15 mg L⁻¹ De-2CuO@NC samples for 30 min after staining with 3 µM propidium iodide (PI) (Fig. S17), which also confirmed the disruption of membrane integrity. Fig. 7c shows the FL intensity in control and De-2CuO@NC-treated samples at 10 and 30 minutes, De-2CuO@NC induced the imbalance of the redox level in E. coli cells, and the intracellular ROS level after De-2CuO@NC treatment was markedly greater than that observed in the control group. This work also provides photographs of fluorescence microscopy of E. coli treated with 15 mg L⁻¹ De-2CuO@NC samples for 30 min after staining with 10 μM DCFH-DA (Fig. S18). Obviously, the intracellular ROS was involved in the De-2CuO@NC-induced death process of E. coli.^71–73 In all, extracellular ˙O₂⁻ induced the production of intracellular ROS and lipid peroxidation, leading to the disruption of the cell wall and the cell membrane, and subsequent cellular death. Fig. 7d shows that the survival rate of HeLa cells remained above 90%; this demonstrates that De-2CuO@NC revealed good biocompatibility while maintaining high antibacterial activity. Furthermore, this work conducted an inactivation experiment of the influenza virus using De-2CuO@NC. The experimental procedures were followed as described in Appendix S5, SI. The results of the experiment are presented in Table S5. The findings indicate that the average logarithm infectivity titre value (lgTCID50 mL⁻¹) changed from 2.66 × 10⁵ to 2.48 × 10⁴ after 30 min treatment with De-2CuO@NC, indicating above 90% kill efficiency, thus implying a virus-killing effect of De-2CuO@NC.

Fig. 6 (a) Inactivation effect of different concentrations of De-2CuO@NC on E. coli, (b) inactivation effect of different bactericidal materials on E. coli, (c and d) Cu²⁺ eluted from De-2CuO@NC during the bactericidal process and inactivation effect of Cu²⁺ (2 mg L⁻¹) on E. coli. Error bars in (a, b and d) represent the standard deviation of three independent experiments.

Fig. 7 (a) EPR spectra; (b) effects of various concentrations of De-2CuO@NC on malondialdehyde (MDA) formation; (c) magnetic field-induced fluorescence intensity in control and De-2CuO@NC-treated samples at 10 and 30 minutes; and (d) survival of HeLa cells at different De-2CuO@NC concentrations. Error bars in (d) represent the standard deviation of three independent experiments.

3. Environmental implications

Due to the presence of a large amount of PH₃ in the yellow phosphorus tail gas, it is difficult to utilize the yellow phosphorus tail gas as a resource, resulting in a large amount of resource waste. A large number of studies have shown that Cu-based materials are the most effective PH₃ adsorbents at present, but the high preparation cost of adsorbents, low adsorption efficiency, and the difficulty of adsorbent inactivation treatment have become major problems that limit their industrial application. In this work, the optimization of the preparation process for traditional PH₃ adsorbents took into account the improvement of performance, and the feasibility of De-2CuO@NC for photocatalysis and sterilization was verified, completing the whole process of environmental green governance and realizing the unification of environmental benefits and economic benefits. A novel N- and C-doped high-performance PH₃ sorbent (CuO@NC) prepared through a simple and cost-effective approach can efficiently capture PH₃ with an adsorption capacity of 272.54 mg (PH₃) per g_sorbent, which significantly exceeds the values reported for other sorbents in the literature. The remarkable efficacy in PH₃ adsorption exhibited by this sorbent is mainly due to its rich basic sites, high density of oxygen vacancies, and unique microstructural features. Previously, materials used for PH₃ treatment were mostly disposed of after use. Importantly, our study reveals that De-2CuO@NC can be further utilized as a photocatalyst and bactericidal material with excellent photocatalytic degradation activity and bactericidal performance, making it a dual-functional material with potential applications in environmental fields. This work may also provide insights into the rational design of high-performance PH₃ adsorbents and the sustainable utilization of waste materials. (i.e. Cu₃P converted from PH₃).

Author contributions

Zhongxian Wang: methodology, investigation, writing – original draft, and data curation. Jiuyang He: methodology, software, and investigation. Chaoyang Peng: data curation. Jiayu Feng, Can Niu and Yixing Ma: investigation. Xin Sun and Ping Ning: resource acquisition. Fei Wang: conceptualization, formal analysis, and writing – review & editing. Lian Wang: resources, funding acquisition, and writing – review & editing. Kai Li: supervision, funding acquisition, writing – review & editing, and resources.

Conflicts of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). The Supplementary Information includes detailed characterization data and supporting results, including TG-DTG curves (Fig. S1), pore size distributions (Fig. S2, S9), SEM images (Fig. S3), XPS spectra (Fig. S4, S6, S7, S10, S11), EDS analyses and mapping (Fig. S5, S12), N₂ adsorption–desorption isotherms (Fig. S8), mass change data (Fig. S13), XRD patterns (Fig. S14), ICP results (Fig. S15), photocatalytic Hg⁰ removal tests (Fig. S16), antibacterial fluorescence microscopy images (Fig. S17, S18), and tables summarizing physical properties, photocatalytic performance, XPS compositions, and virus inactivation assay results (Tables S1–S5). See DOI: https://doi.org/10.1039/d5gc04734j.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (no. 52460012), Yunnan Fundamental Research Projects (202401AW070015), Yunnan Key Laboratory of Phosphorus Gypsum Resource Recycling and Ecological Utilization (202449CE340028), and Yunnan International Joint R&D Center for Waste Resource Recovery in Metallurgical and Chemical Industries (202403AP140008).

References

1. J. Feng, K. Li, S. Li, C. Wang, P. Ning, X. Sun and Y. Wang, Res. Chem. Intermed., 2019, 46, 329–346.
2. M. A. Hartog, R. J. Lewandowski, C. S. Hofmann, A. A. Melber, C. C. Rothwell, K. Sherman, J. Andres, J. A. Tressler, A. M. Sciuto, B. Wong and H. M. Hoard-Fruchey, Chem. Res. Toxicol., 2021, 34, 2031–2044.
3. X. Tang, J. Xue and C. Xing, Appl. Surf. Sci., 2020, 530, 147264.
4. W. Y. Tuet, S. A. Pierce, M. C. Racine, S. Stone, E. Pueblo, A. Dukes, J. Tressler, L. Jennings, B. J. McCranor and B. Wong, Toxicol. Appl. Pharmacol., 2021, 427, 115652.
5. R. Zhu, D. Kong, L. Sun, J. Geng, X. Wang and D. Glindemann, Environ. Sci. Technol., 2006, 40, 7656.
6. S. M. Chang, Y. y. Hsu and T. s. Chan, J. Phys. Chem. C, 2011, 115, 2005–2013.
7. P. Ning, X. Wang, H. J. Bart, S. Tian, Y. Zhang and X. Q. Wang, J. Cleaner Prod., 2011, 19, 1547–1552.
8. C. Han, J. Geng, J. Zhang, X. Wang and S. Gao, Chemosphere, 2010, 82, 935–939.
9. W. C. Li, H. Bai, J. N. Hsu, S. N. Li and C. Chen, Ind. Eng. Chem. Res., 2008, 47, 1501–1505.
10. X. Xu, G. Huang and S. Qi, Chem. Eng. J., 2017, 316, 563–572.
11. C. Tong, R. Xiang, L. Peng, L. Tan, X. Tang, J. Wang, L. Li, Q. Liao and Z. Wei, J. Mater. Chem. A, 2020, 8, 3351–3356.
12. F. Jiayu, W. Fei, W. Chi, L. Kai, N. Ping, S. Xin and J. Lijuan, Sep. Purif. Technol., 2021, 277, 119420.
13. S. Li, K. Li, J. Hao, P. Ning, L. Tang and X. Sun, Chem. Eng. J., 2016, 302, 69–76.
14. X. Song, S. Li, K. Li, P. Ning, C. Wang, X. Sun and Y. Wang, Microporous Mesoporous Mater., 2017, 259, 89–98.
15. Y. Wang, Q. Lin, P. Ning, C. Wang, X. Sun and K. Li, J. Taiwan Inst. Chem. Eng., 2018, 87, 44–53.
16. Q. Yu, P. Ning, H. Yi, X. Tang, M. Li and L. Yang, Sep. Sci. Technol., 2012, 47, 527–533.
17. H. Yi, Q. Yu, X. Tang, P. Ning, L. Yang, Z. Ye and J. Song, Ind. Eng. Chem. Res., 2011, 50, 3960–3965.
18. M. Nilkar, F. E. Ghodsi, S. Jafari, D. Thiry and R. Snyders, J. Alloys Compd., 2021, 853, 157298.
19. C. Yang, H. Ye, J. Byun, Y. Hou and X. Wang, Environ. Sci. Technol., 2020, 54, 12621–12630.
20. R. Rameshbabu, P. Ravi, G. Pecchi, E. J. Delgado, R. V. Mangalaraja and M. Sathish, J. Colloid Interface Sci., 2021, 590, 82–93.
21. Q. Dai, K. Shen, W. Deng, Y. Cai, J. Yan, J. Wu, L. Guo, R. Liu, X. Wang and W. Zhan, Environ. Sci. Technol., 2021, 55, 4007–4016.
22. Y. Song, X. Xin, S. Guo, Y. Zhang, L. Yang, B. Wang and X. Li, Chem. Eng. J., 2019, 384, 123337.
23. G. Liu, Z. Hu, X. Chen, W. Li, Y. Wu, Z. Liu, L. Miao, Z. Luo, J. Wang and Y. Guo, J. Hazard. Mater., 2023, 465, 133269.
24. R. Zhang, C. Song, M. Kou, P. Yin, X. Jin, L. Wang, Y. Deng, B. Wang, D. Xia, P. K. Wong and L. Ye, Environ. Sci. Technol., 2020, 54, 3691–3701.
25. S. Liang, B. Peng, S. Liu, W. Zhang, M. Guo, F. Cheng and M. Zhang, Environ. Sci. Technol., 2020, 54, 5964–5972.
26. O. Tursunov, L. Kustov and Z. Tilyabaev, J. Pet. Sci. Eng., 2019, 180, 773–778.
27. Y. Ye, L. Gao, J. Xu, L. Wang, L. Mo and X. Zhang, J. Rare Earths, 2023, 41, 862–869.
28. P. Ning, H. Yi, Q. Yu, X. Tang, L. Yang and Z. Ye, J. Rare Earths, 2010, 28, 581–586.
29. J. N. Hsu, H. Bai, S.-N. Li and C. J. Tsai, J. Air Waste Manage. Assoc., 2010, 60, 629–635.
30. Y. Wang, Q. Lin, P. Ning, C. Wang, H. Zeng, X. Sun, X. Song and K. Li, New J. Chem., 2021, 45, 5822–5828.
31. Y. Wang, P. Ning, R. Zhao, K. Li, C. Wang, X. Sun, X. Song and Q. Lin, Front. Environ. Sci. Eng., 2021, 15, 132.
32. Z. Ren, S. Quan, Y. Zhu, L. Chen, W. Deng and B. Zhang, RSC Adv., 2015, 5, 29734–29740.
33. H. H. Tseng and M.-Y. Wey, Chemosphere, 2005, 62, 756–766.
34. R. Li, Y. Huang, D. Zhu, W. Ho, J. Cao and S. Lee, Environ. Sci. Technol., 2021, 55, 4054–4063.
35. O. Mabayoje, M. Seredych and T. J. Bandosz, ACS Appl. Mater. Interfaces, 2012, 4, 3316–3324.
36. A. Behera, S. Mansingh, K. K. Das and K. Parida, J. Colloid Interface Sci., 2019, 544, 96–111.
37. W. Yan, L. Yan and C. Jing, Appl. Catal., B, 2019, 244, 475–485.
38. W. Sun, Z. Fu, H. Shi, C. Jin, E. Liu, X. Zhang and J. Fan, J. Chem. Technol. Biotechnol., 2020, 95, 3117–3125.
39. S. Kanuri, S. Dinda, S. A. Singh, S. Roy, C. Chakraborty and S. P. Datta, Mater. Today Chem., 2024, 36, 101959.
40. S. Yang, Y. Liu, Y. Hao, X. Yang, W. A. Goddard, X. L. Zhang and B. Cao, Adv. Sci., 2018, 5, 1700659.
41. X. Zhu, Z. Zhao, J. Tang, X. Peng, H.-D. Um, X. Li, Q. Xiao and W. Luo, Fundam. Res., 2023, 5, 1134–1143.
42. Z. Hu, Z. Wang, Y. Guo, L. Wang, Y. Guo, J. Zhang and W. Zhan, Environ. Sci. Technol., 2018, 52, 9531–9541.
43. Y. Maeda, Y. Iizuka and M. Kohyama, J. Am. Chem. Soc., 2013, 135, 906–909.
44. Y. Peng, K. Li and J. Li, Appl. Catal., B, 2013, 140, 483–492.
45. W. Yuexing, L. Linyu, F. Bin, H. Ziyue, Z. Jiansheng, Z. Yuxun, Q. Yuhong and H. Chong, Front. Environ. Sci. Eng., 2024, 18, 152.
46. G. Fu, D. Mao, J. Yu and Q. Guo, J. Rare Earths, 2023, 42, 660–667.
47. K. Zhang, Z. Xiong, S. Li, B. Yan, J. Wang and Y. Du, J. Alloys Compd., 2017, 706, 89–96.
48. L. Zhang, G. Wang, X. Hao, Z. Jin and Y. Wang, Chem. Eng. J., 2020, 395, 125113.
49. W. Tang, D. S. Alessi, T. Wang, J. Wu, S. Li, K. O. Konhauser, Z. Li and J. Chen, Green Chem., 2024, 26, 3229–3238.
50. A. Wang, K. Xie, D. Bernin, A. Kumar, K. Kamasamudram and L. Olsson, Appl. Catal., B, 2020, 269, 118781.
51. Z. Yang, L. Shao, L. Wang, X. Xia, Y. Liu, S. Cheng, C. Yang and S. Li, Int. J. Hydrogen Energy, 2020, 45, 14334–14346.
52. S. Hua, D. Qu, L. An, W. Jiang, Y. Wen, X. Wang and Z. Sun, Appl. Catal., B, 2018, 240, 253–261.
53. J. Luo, W. Li, X. Wang, E. Lichtfouse, D. Huang, X. Chen, Y. Zhang, L. Zhu and C. Wang, J. Rare Earths, 2024, 42, 827–837.
54. R. Shen, J. Xie, X. Lu, X. Chen and X. Li, ACS Sustainable Chem. Eng., 2018, 6, 4026–4036.
55. G. Zhang, Y. Zhang, X. Zhao, Y. Jiao, Y. Yan and J. Jiang, J. Rare Earths, 2023, 42, 817–826.
56. Y. Huang, W. Xing, L. Zhou, B. Tian, J. Zhang and Y. Zhou, Res. Chem. Intermed., 2022, 48, 2887–2901.
57. C. Lops, A. Ancona, K. Di Cesare, B. Dumontel, N. Garino, G. Canavese, S. Hérnandez and V. Cauda, Appl. Catal., B, 2019, 243, 629–640.
58. F. Hou, F. Liu, H. Wu, M. Qasim, Y. Chen, Y. Duan, Z. Feng and M. Liu, Chin. J. Chem., 2022, 41, 173–180.
59. X. Wang, Y.-n. Zhou, R. Li, L. Wang, L. Tao and P. Ning, Chem. Eng. J., 2019, 360, 1530–1541.
60. J. Dong, Z. Huang, Y. Fu, G. Liu, R. Z. Valiev, Y. V. Petrov, Y. Feng, D. Jia, V. E. Baulin, A. Y. Tsivadze, Y. Zhou and B. Li, Carbon, 2025, 243, 120478.
61. Z. Chen, J. Liu, Z. Li, P. Zheng, B. Gao, S. Al-Farraj and M. Sillanpää, J. Hazard. Mater., 2024, 471, 134372.
62. M. Nazeri, H. Sedaghat, R. Rafiei, M. A. Farzin and S. M. Hosseinpour Mashkani, J. Rare Earths, 2023, 42, 851–858.
63. K. Wang, M. Lv, T. Si, X. Tang, H. Wang, Y. Chen and T. Zhou, J. Hazard. Mater., 2023, 461, 132479.
64. B. Li, Q. Zuo, Z. Deng, X. Zheng, P. Li and J. Wu, Chem. Eng. J., 2024, 497, 154735.
65. L. Wang, W. Wu, X. Xie, H. Chen, J. Lin and D. D. Dionysiou, Water Res., 2018, 141, 145–151.
66. P. J. Babu, M. Doble and A. M. Raichur, J. Colloid Interface Sci., 2017, 513, 62–71.
67. Y. Gerchman, V. Cohen-Yaniv, Y. Betzalel, S. Yagur-Kroll, S. Belkin and H. Mamane, Water Res., 2019, 161, 119–125.
68. S. R. Martínez, A. M. Durantini, M. C. Becerra and G. Cosa, ACS Infect. Dis., 2020, 6, 2468–2477.
69. Z. Niu, S. Han, W. Qin, P. Gao, F. Xiao and S. Yang, Front. Environ. Sci. Eng., 2024, 18, 124.
70. J. Wang, Y. Wang, W. Ren, D. Zhang, P. Ju and K. Dou, J. Hazard. Mater., 2022, 440, 129742.
71. Y. Kawamoto, A. Nishitani, Y. Yoshimura, T. Kobayashi and H. Ito, Chem. Commun., 2024, 61, 97–00.
72. K. Nishimura, S. Shindo, A. Movila, R. Kayal, A. Abdullah, I. J. Savitri, A. Ikeda, T. Yamaguchi, M. Howait, A. Al-dharrab, A. Mira, X. Han and T. Kawai, Free Radical Biol. Med., 2016, 97, 330–341.
73. J. Ye, B. Li, M. Li, Y. Zheng, S. Wu and Y. Han, Bioact. Mater., 2021, 11, 181–191.

---

Footnote

† These authors contributed equally to this work.

---