Challenges of photocatalytic extraction of uranium: a review

Principle of photocatalysis

Photocatalysis is a light-driven process in which a semiconductor catalyst, upon absorbing photons, facilitates a chemical reaction^[22]. The process begins when the catalyst absorbs photons whose energy is greater than or equal to its bandgap energy (Eg)^[23,24]. This excitation leads to the promotion of electrons from the valence band (VB) to the conduction band (CB), leaving behind holes in the VB^[25]. These electron-hole pairs are the fundamental carriers responsible for initiating redox reactions [Figure 1A]. The efficiency of the photocatalytic process depends on several key factors: the bandgap of the photocatalyst, its surface area, and its chemical stability under reaction conditions. The bandgap determines the energy threshold for the catalyst to absorb light, and semiconductors with narrow bandgaps can absorb a broader range of light, including visible light^{[23,24,26,27]}. In uranium extraction, the photogenerated electrons reduce uranium ions [e.g., U(VI) to U(IV)], while the holes in the VB participate in oxidation reactions, such as oxidizing water to generate ROS^[22,28]. Materials such as graphitic carbon nitride (g-C₃N₄) and bismuth-based catalysts have shown particular promise in uranium extraction due to their ability to absorb visible light, unlike titanium dioxide (TiO₂), which primarily absorbs UV light due to its wide bandgap (~3.2 eV)^[29-32].

Figure 1. (A) The mechanism of photocatalytic stages: (1) light harvesting; (2) charge excitation; (3) charge separation; (4) charge recombination; (5) surface charge recombination; (6) reduction and (7) oxidation reaction; (B) Redox potential of U(VI) and common active species in photocatalytic; (C) Illustration of reaction mechanism with or without sacrificial reagent in uranium extraction.

Photocatalytic uranium extraction process

Based on the principles of photocatalysis, the occurrence of redox reactions is governed by the relative electrode potentials of the electron donor and acceptor species. Specifically, the electron donor must have an oxidation potential higher than the VB potential, while the electron acceptor must have a reduction potential lower than the CB potential of the catalyst. The relative positions of these potentials, as illustrated in [Figure 1B], determine the feasibility of redox reactions. In the case of uranium extraction, the reduction potential of U(VI) is a critical factor in facilitating its reduction under photocatalytic conditions.

Several pathways have been identified for the electron transfer during the photocatalytic reduction of U(VI)^[33]:

(i). One-Electron Reduction: In this process, U(VI) is first reduced by a photogenerated electron to U(V), which can then be further reduced to U(IV). Notably, U(V) can undergo disproportionation, forming both U(VI) and U(IV) in the process. The reactions are as follows:

U(VI) + e^- → U(V)
U(V) + e^- → U(IV)
U(V) → U(VI) + (IV)

(ii). Two-Electron Reduction: Alternatively, U(VI) can be directly reduced to U(IV) via a two-electron transfer reaction, which is a more efficient pathway for uranium extraction:

U(VI) + 2e^- → U(IV)

(iii). Reduction Under Acidic Conditions: Under acidic conditions, protons in the solution can participate in the two-electron reduction process, facilitating the reduction of U(VI) to U(IV). The reaction proceeds as follows:

UO₂²⁺ + 4H⁺ + 2e^- → U⁴⁺ + 2H₂O

(iv). Reduction by Superoxide Radicals (·O₂^-): In some systems, U(VI) can be reduced by superoxide radicals (·O₂^-), which are generated upon electron transfer to molecular oxygen^[34,35]. The reactions proceed as:

O₂ + e^- → ·O^-₂
x·O₂^- + U⁶⁺ → U^(6-X)+ + xO₂
or 2·O₂^- + UO₂²⁺ → UO₂ + 2O₂

(v). In the multi-electron photocatalytic reaction, oxygen (O) undergoes two-electron transfer to generate H₂O₂. The chemical equation is as follows:

O₂ + 2H⁺ + 2e^- → H₂O₂
·O₂^- + e^- + 2H⁺ → H₂O₂
CH₃OH + ·OH/·O₂^- → CO₂ + H₂O
UO₂²⁺ + H₂O₂ + 2H₂O → UO₂(O₂)·2H₂O + 2H⁺

Under different pH conditions, uranyl can react with photogenerated H₂O₂ to form water-insoluble metastudtite [(UO₂)O₂·2H₂O] [Figure 1C], or directly reduce to UO₂ when sacrificial reagents are added, and methanol is oxidized to CO₂ and H₂O.

Additionally, when organic matter is present in the system, it can act as an electron donor, thus extending the lifetime of the photogenerated electrons and enhancing the reduction of U(VI)^[36]. The role of organic species as electron donors is significant, as they can stabilize the photogenerated electrons, preventing their recombination and promoting more efficient uranium reduction. It is important to note that while these pathways are established, the mechanism of photocatalytic reduction of U(VI) remains an area of active research, and further investigation is required to fully elucidate the electron transfer dynamics and reaction intermediates involved in this process.

Chemical behavior of uranium

Uranium exists in several oxidation states in aqueous environments, with the most common being U(VI) and U(IV). In oxidizing environments, uranium typically exists as the uranyl ion (UO₂²⁺), which is soluble and highly mobile in water. The hexavalent form of uranium [U(VI)] is more prone to leaching and dissolution from mineral ores, which is why it is often the target of extraction methods. The redox behavior of uranium plays a crucial role in photocatalytic uranium extraction. Under light irradiation, photogenerated electrons reduce U(VI) to U(IV), which has significantly lower solubility and can be more easily extracted from the solution. The presence of various ligands, the pH of the solution, and the concentration of other reactive species, such as ·OH, also influence the redox behavior of uranium during photocatalysis. At lower pH values, uranium generally exists as UO₂²⁺, while at higher pH, uranium can form complex species such as UO₂(OH)₃^- or UO₂(OH)₄^2-[22,37,38]. Under photocatalytic conditions, the redox equilibrium between these species can shift, with the photogenerated electrons promoting the reduction of UO₂⁺. Additionally, the presence of competing ions, such as calcium (Ca²⁺), magnesium (Mg²⁺), or carbonate (CO₃^2-), can also affect uranium redox behavior and the efficiency of the extraction process^[39-43]. These ions may form precipitates with uranium or alter the reaction kinetics, thereby influencing the overall efficiency of uranium removal.

It is noteworthy that although significant progress has been made in understanding the photocatalytic reduction of U(VI), the exact mechanism remains incomplete and requires further investigation. For instance, Li et al. analyzed uranium species in the liquid phase using absorption spectroscopy^[23]. During the photocatalytic reduction of U(VI) on TiO₂, the characteristic absorption peak of U(VI) in the range of 350-500 nm rapidly decreased, indicating a reduction process. Upon treating the system with hydrochloric acid, the reduced uranium species deposited on the TiO₂ surface dissolved, and the characteristic absorption peak of U(IV) appeared in the supernatant. This observation provided direct evidence for the formation of U(IV) on the surface of TiO₂. However, the role of U(V) as an intermediate in this process remains unclear due to its tendency to undergo rapid disproportionation, which can result in the simultaneous formation of U(VI) and U(IV). In homogeneous photocatalytic systems, the presence of U(V) was confirmed through the detection of its characteristic electron paramagnetic resonance (EPR) signals^[44]. Interestingly, as the irradiation time was extended, these signals gradually transitioned into a mixed signal of U(V) and U(IV), directly confirming the disproportionation of U(V) into other species. A similar phenomenon was observed in isopropanol-based homogeneous photocatalytic systems, further supporting the potential involvement of U(V) as a transient intermediate^[45]. In contrast, no soluble lower-valent uranium species [e.g., U(V), U(IV), or U(III)] were detected in heterogeneous systems involving TiO₂^[45]. This raises questions about the nature of uranium intermediates in such systems. One possibility is that U(V) forms momentarily but undergoes rapid disproportionation, making it undetectable. Another explanation is that U(V) may not form at all in certain heterogeneous systems. Consequently, whether U(V) plays a role during the photocatalytic reduction of U(VI) in homogeneous and heterogeneous systems remains a subject of debate and warrants further detailed investigation. As for the lower oxidation states of uranium, it is challenging for U(VI) to be directly reduced to U(III) or U(0) through photocatalytic reduction, as the required reduction potential is quite negative^[46]. However, in a formic acid-U(VI) homogeneous photocatalytic system, spectrophotometric analysis detected the transformation of U(VI) into a mixture of U(V), U(IV), and U(III) when the pH was below 2.0^[47]. This suggests that the photocatalytic reduction of U(VI) can become more complex under specific conditions.

To date, most studies suggest that U(VI) is ultimately reduced to the dark gray U(IV) precipitate in conventional photocatalytic systems^[48]. However, there is some divergence in understanding the phases of U(VI) reduction products. Reports have shown that the photocatalytic reduction products of U(VI) include U₃O₈, U₃O₇, UO_2+x (where x = 0 to 0.25), UO₃, UO₂, and mixtures of U₃O₇, UO₄∙xH₂O, etc.^{[21,34,45,49,50]}. In the TiO₂ photocatalytic reduction of U(VI) system, the reaction products appear as particulate precipitates that are deposited on the catalyst surface, resembling the electrochemical reduction of U(VI)^[7,23]. Once the reaction products are separated from the system, they are readily re-oxidized to form soluble U(VI)^[23,51]. X-ray photoelectron spectroscopy (XPS) analysis of the products revealed that the reduced products mostly exhibit a mixed oxidation state of U(VI) and U(IV). However, the interpretation of this mixed valence state varies. For instance, Feng et al. attributed the mixed valence state to the formation of U(IV) oxides or U(IV) hydroxides^[52]. Bonato and Salomone et al. suggested that this mixed valence state resulted from the formation of UO_2+x (where x = 0 to 0.25)^[45,53]. Meanwhile, Alhindawy et al. proposed that the mixed valence state originated from the formation of (UO₂)₂O₂·₂H₂O^[54]. Furthermore, some studies using XPS analysis confirmed that the photocatalytic reduction of U(VI) on TiO₂ resulted in a mixture of UO₃, UO₂, and U₃O₇, while others suggested that only UO₂ was present in the products^[35,55,56]. Moreover, X-ray diffraction (XRD) analysis conducted by He et al. identified the phase information of the uranium products, revealing characteristic diffraction peaks corresponding to UO₂, U₃O₈, and (UO₂)₆O₂(OH)₈∙6H₂O^[36]. However, it is worth noting that XRD analysis may struggle to distinguish between UO₂ and UO_2+x (where x < 0.25). Additionally, the size and content of the products are essential for effective phase analysis. When the pure water system was replaced with a high-salinity brine system, uranium oxide nanoparticles were also formed on the TiO₂ surface under light irradiation. However, due to their high dispersion and extremely small particle size (approximately 6 nm), the characteristic peaks of uranium precipitation in the resulting products were not detectable in the XRD spectrum^[57]. Wang et al. reported that under both aerobic and anaerobic conditions, U(VI) formed different products on the surface of ferrihydrite^[58]. Under anaerobic conditions, U(VI) was reduced by ·O₂^- radicals to form UO_2+x (where x < 0.25) precipitates, while under aerobic conditions, the system generated large amounts of H₂O₂, which complexed with U(VI) to form UO₂·xH₂O precipitates. This suggests that the formation of active species in different systems is directly linked to the formation of uranium products.

Experimental condition

The performance of photocatalytic uranium extraction is significantly influenced by experimental setups and reaction conditions, including light source type, reaction temperature, pH, and the presence of coexisting ions or ligands. The spectral distribution and intensity of the light source directly influence the efficiency of the photocatalytic reaction. UV light sources, such as mercury lamps, are commonly used in experimental studies due to their high energy, which matches the bandgap of traditional photocatalysts such as TiO₂. However, since UV light only constitutes about 5% of natural sunlight, researchers are focused on developing photocatalysts that can be driven by visible light or even full sunlight, such as g-C₃N₄ and MOFs, which demonstrate good photocatalytic activity under visible light. Adjusting the light source intensity and wavelength in laboratory experiments, especially under low light intensity conditions, can help simulate more practical scenarios where uranium extraction occurs under visible or near-infrared light^[59].

The pH of the solution plays a key role in determining the chemical speciation of uranium and the surface charge of the photocatalyst. Wang et al. found that the reduction efficiency of U(VI) increases with pH value^[60]. At low pH, H⁺ compete with U(VI) for electrons, reducing the photocatalytic activity. At pH values between 4 and 6, U(VI) mainly exists as UO₂OH⁺, (UO₂)₂(OH)₂²⁺, (UO₂)₃(OH)₅⁺, and (UO₂)₃(OH)₄²⁺, and electrostatic attraction can help capture electrons on the catalyst surface, improving uranium removal efficiency. When the pH exceeds 7, U(VI) typically transforms into species such as (UO₂)₃(OH)₇^-, (UO₂)₂(OH)₄^2-, and (UO₂)₂(OH)₃^-, which exhibit electrostatic repulsion with the catalyst surface, thereby hindering adsorption and reduction^[61]. Moreover, the pH of the reaction solution can influence the redox potential of the catalyst’s CB, which in turn affects the reduction capability of the photogenerated electrons. In general, as pH increases, the more negative the CB redox potential becomes, favoring uranium photocatalytic reduction^[62]. Thus, it is important to evaluate the photocatalytic performance of catalysts using U solutions with pH and concentrations close to those found in target water bodies.

In addition, coexisting ions in the solution, such as Ca²⁺, Mg²⁺, and phosphates, or organic ligands, can compete with uranium for active sites on the photocatalyst or form stable complexes with uranium, thereby reducing its reducibility and affecting extraction efficiency. Gong et al. studied the photocatalytic performance of Er_0.04-ZnO for U(VI) reduction in solutions containing Na⁺, K⁺, Mg²⁺, Sr²⁺, Cu²⁺, Fe³⁺, SO₄^2-, CO₃^2-, Br^-, and Cl^-, and found that the U(VI) removal efficiency only slightly decreased, indicating that Er_0.04-ZnO has strong anti-interference ability^[63]. In contrast, ions such as Cu²⁺, with a reduction potential of -0.325 V, capture photogenerated electrons more easily than U(VI), leading to the reduction of Cu²⁺ before U(VI)^[64]. Furthermore, anions such as NO₃^- have little impact on the reaction, whereas SO₄^2- forms stable complexes with uranyl ions, slightly reducing photocatalytic uranium reduction efficiency. The strongest inhibition comes from CO₃^2- due to the formation of UO₂(CO₃)₂^2- and UO₂(CO₃)₃^4- complexes, which are difficult to adsorb and have a more negative reduction potential, hindering the photocatalytic process^[65]. Additionally, various organic compounds in water may have dual effects on uranium removal. Some organic substances - such as methanol, tannic acid, citric acid, formic acid, and formates - can act as hole (h⁺) scavengers, suppressing the recombination of photogenerated electron-hole pairs and thus enhancing U(VI) removal^[62]. While many of these compounds contain carboxyl or hydroxyl groups, their relatively weak complexation with uranium and efficient hole-scavenging behavior favor photocatalytic reduction. In contrast, other organics such as humic acid, OA, EDTA, and amino sulfonates, although also rich in carboxyl, phenol, or alcohol groups, can form stable complexes with U(VI) or block active catalytic sites. This can hinder electron transfer and suppress the photocatalytic reduction of uranium^[66]. He et al. reported that methanol can inhibit carrier recombination by scavenging photogenerated holes, increasing U(VI) removal efficiency, whereas the presence of tributyl phosphate (TBP) or OA affects electron binding to U(VI), leading to a decrease in photocatalytic performance^[67]. Therefore, using real target solutions, such as uranium wastewater and real seawater, is crucial for evaluating the practical performance of photocatalysts.

Other factors influencing the photocatalytic uranium extraction process include the amount of catalyst used, uranium concentration, temperature, and dissolved O levels. For example, Liang et al. studied the effect of ZnFe₂O₄ catalyst dosage (0.1 to 0.6 g/L) on the photocatalytic reduction of U(VI) and found that increasing the catalyst dosage accelerated the reaction rate due to an increase in active sites, but beyond the optimal dosage, the suspension of ZnFe₂O₄ caused light shielding, reducing photocatalytic efficiency^[68]. Thus, considering the solid-to-liquid ratio in experiments is essential to maximize the use of photocatalysts. Lu et al. compared the effects of N₂ and O₂ atmospheres on B-g-C₃N₄ photocatalytic uranium reduction, finding that the reduction efficiency decreased in O₂ atmosphere, with uranium removal dropping from 93% to 70%^[69]. Therefore, photocatalytic performance evaluation should be conducted in systems without sacrificial agents and under anaerobic conditions to reduce the cost of uranium extraction.

Traditional inorganic photocatalysts

The development of photocatalytic materials for uranium extraction has evolved from traditional semiconductor materials to novel materials in recent years. In the early stages of research, TiO₂, as one of the earliest photocatalysts, was widely applied in the study of uranium extraction. Since 1972, Fujishima and Honda successfully achieved water splitting using irradiated TiO₂, which led to TiO₂ becoming a prominent photocatalyst attracting global scientific attention^[70]. In the late 1980s, TiO₂ gained attention for its chemical stability, low toxicity, and low cost, which made it a primary candidate in the field of photocatalytic uranium extraction. By being activated under UV light, TiO₂ generates electron-hole pairs that participate in the reduction of U(VI) to U(IV), thereby facilitating uranium extraction by lowering its solubility.

However, the primary issue with TiO₂ lies in its wide bandgap (approximately 3.2 eV), which restricts it to only absorb UV light^[71]. Since UV light constitutes only a small portion of the solar spectrum, the photocatalytic efficiency of TiO₂ remains relatively low under natural sunlight^[71]. To address this limitation, researchers have gradually explored other photocatalytic materials. In the mid-1990s, materials such as zinc oxide and cadmium sulfide (CdS) were introduced into the study of uranium extraction. ZnO and CdS possess a relatively smaller bandgap compared to TiO₂ (ZnO approximately 3.3 eV and CdS approximately 2.4 eV), enabling them to absorb a wider range of the light spectrum, particularly visible light, and therefore improve photocatalytic efficiency^[64]. Traditional semiconductor materials, including metal oxides (e.g., TiO₂, ZnO) and metal sulfides (e.g., CdS, PbS, MoS₂), have played a crucial role in the development of photocatalytic uranium extraction. These materials have demonstrated promising photocatalytic activity due to their intrinsic semiconductor properties, including suitable band structures and high redox potentials. Metal oxides, particularly TiO₂, have been extensively studied for photocatalytic uranium extraction due to their chemical stability, low toxicity, and cost-effectiveness. Evans et al. demonstrated that TiO₂ could efficiently separate U(VI) (0.17 mM) from a 3 M phosphate waste solution under UV light^[72]. Within 6 h, TiO₂ removed 98% of the U(VI) from the wastewater (pH ~6), indicating the potential of photocatalytic methods to reduce and remove U(VI) from solutions containing strong ligands. In recent years, TiO₂ has been extensively studied in the field of photocatalytic uranium extraction. For instance, Li et al. were the first to use TiO₂ in photocatalytic uranium extraction from seawater, demonstrating its feasibility^[23]. In this study, without reaching saturation, they successfully achieved uranium extraction levels as high as 3,960 mg/g. After exposure to air for a period and subsequent washing with sodium carbonate solution, uranium species adsorbed on the TiO₂ surface were effectively recovered, with TiO₂ maintaining its structural integrity and photocatalytic performance. This was the first confirmation of the substantial potential for photocatalytic uranium extraction from aqueous environments. Liu et al. employed magnetron sputtering to prepare TiO₂ thin films on SnO₂ (FTO) substrates^[73]. Under UV light, the TiO₂/FTO composite achieved a uranium extraction rate exceeding 70% within 5 min, with a total uranium extraction of 71.13 mg/g. The photocatalytic efficiency of TiO₂ is often influenced by factors such as its crystal phase, mixing ratio, morphology, and exposed crystal facets^[74-76]. Zhang et al. found that rutile-type TiO₂ nanorods demonstrated a 10-fold higher reduction efficiency for U(VI) compared to anatase TiO₂^[77]. Interestingly, the physical mixing of rutile and anatase (75% anatase and 25% rutile) significantly enhanced photocatalytic reduction of U(VI). Transmission electron microscopy (TEM) further confirmed that uranium reduction products selectively deposited on the surface of the rutile phase. This enhancement was attributed to the distinct band structures of rutile and anatase; upon mixing, they formed a heterostructure due to aggregation, with photogenerated electrons migrating from the higher CB of anatase to the lower CB of rutile. This facilitated effective spatial separation of charge carriers, thus improving the overall photocatalytic activity. However, their large bandgaps (TiO₂ ~3.2 eV, ZnO ~3.3 eV) restrict their absorption to UV light, which comprises only a small fraction (~5%) of the solar spectrum^[71]. This narrow absorption range significantly reduces their photocatalytic efficiency under natural sunlight. Efforts have been made to modify these materials through metal and non-metal doping, heterojunction construction, and surface engineering to extend their light absorption into the visible range^[49,78-80]. Wan et al. developed TiOS/NHCS heterojunction, which efficiently reduced U(VI) through optimized energy levels and enhanced electron transfer^[81]. The catalyst achieved over 90% U(VI) removal in 20 min, with a U(VI) capacity of 448 mg/g, and showed resistance to interference from common wastewater ions. This work highlights an effective strategy for uranium removal using engineered energy bands and Mott-Schottky heterojunctions [Figure 2A]. Dong et al. developed a biomimetic photocatalytic system, MnO_x@TiO₂@CdS@Au (MTCA), featuring spatially separated cocatalysts (MnOx and Au NPs) on TiO₂@CdS^[82]. This design promotes efficient photogenerated charge migration and separation, enhancing photocatalytic performance. MTCA achieved 99.61% U(VI) removal in 15 min under simulated sunlight and 91.83% removal from natural uranium mine wastewater, outperforming most previously reported catalysts. The study provides a novel approach for designing efficient photocatalysts for environmental remediation [Figure 2B and C]. In addition, Dong et al. developed a synergistic strategy using a Z-scheme heterostructure with spatially separated cocatalysts (Co₃O₄, TiO₂, CdS, Au NPs) in a double-shelled hollow nanocage (Co₃O₄@TiO₂@CdS@Au) for efficient photothermal removal of U(VI) and organic pollutants^[83]. The Z-scheme design enhanced interfacial charge separation, and the photothermal effect improved charge transfer and reaction kinetics. CTCA achieved 98.8% U(VI) removal in 10 min under simulated sunlight and degraded nearly 90% of five organic pollutants. Figure 2D-F shows the corresponding scanning electron microscopy (SEM) and TEM images of Co₃O₄@SiO₂@TiO₂, Co₃O₄@TiO₂@CdS double-shelled nanocages, and Co₃O₄@TiO₂@CdS@Au mono-shelled nanocages, confirming the structure and morphology of the material. However, these modifications have not entirely overcome the intrinsic limitations of metal oxides, and their photocatalytic efficiency remains insufficient for large-scale applications. Metallic compound exhibits severe photo-corrosion, particularly under acidic and alkaline conditions, leading to rapid deactivation and performance deterioration over time.

Figure 2. (A) The scheme of hierarchical TiOS/NHCS photocatalysts^[81], Copyright 2021, Elsevier Ltd; (B) Schematic of the MnOx@TiO₂@CdS@Au synthesis route and (C) TEM images^[82], Copyright 2023, Wiley-VCH; (D) SEM images of Co₃O₄@SiO₂@TiO₂ and (E) Co₃O₄@TiO₂@CdS double-shelled Nanocage; (F) TEM images of Co₃O₄@TiO₂@CdS@Au mono-shelled nanocages^[83], Copyright 2023, Elsevier Ltd. TEM: Transmission electron microscopy; SEM: scanning electron microscopy.

Metal sulfides, such as CdS, PbS, and MoS₂, offer narrower bandgaps (e.g., CdS ~2.4 eV), allowing them to absorb visible light and achieve higher utilization of solar energy^[84]. This property enhances their photocatalytic activity in uranium extraction, especially under sunlight-driven conditions. Given that the CB position of CdS [-0.8 V vs. normal hydrogen electrode (NHE)] is more negative than the reduction potentials of UO₂²⁺/U⁴⁺ (0.267 V) and UO₂²⁺/UO₂ (0.411 V), CdS is thermodynamically capable of photocatalytically reducing U(VI)^[68,85]. It has been reported that the photocatalytic efficiency of CdS is highly influenced by its morphology, which affects the band structure and electron transfer processes^[86]. Various morphologies of CdS have been synthesized using different methods, including 0D nanoparticles, 1D nanorods or nanowires, 2D nanosheets, 3D hollow spheres, and nanosheet assemblies^[87-90]. Among them, 1D CdS has garnered significant attention due to its unique one-dimensional structure, which facilitates charge separation, transfer, and reaction processes, thereby promoting photocatalytic reactions^[91]. However, the performance of 1D CdS is still hindered by issues of electron-hole recombination and photo-corrosion under extended light exposure^[92]. Despite their potential, metal sulfides such as CdS, PbS, and MoS₂ suffer from poor stability, particularly photo-corrosion under prolonged light exposure, which leads to a rapid degradation in photocatalytic performance^[93,94]. The high synthesis costs and environmental toxicity of CdS further limit its practical applications. Similarly, although other metal sulfides, such as PbS and MoS₂, show potential for uranium extraction, they exhibit poor long-term stability and limited resistance to oxidation, which restricts their widespread use in photocatalytic processes.

Another major challenge faced by both metal oxides and metal sulfides is the charge carrier recombination issue, where photogenerated electrons and holes rapidly recombine before participating in redox reactions. This process severely diminishes the overall photocatalytic efficiency. While strategies such as cocatalyst loading, heterojunction formation, and defect engineering have been explored to mitigate charge recombination, these approaches have yet to fully resolve the fundamental limitations of these traditional semiconductors. These limitations underscore the urgent need for next-generation photocatalytic materials that exhibit enhanced visible light absorption, improved stability, and higher efficiency to facilitate the large-scale application of photocatalytic uranium extraction.

Development of novel photocatalysts

In recent years, with the increasing demands for higher efficiency and stability in photocatalytic uranium extraction, researchers have shifted towards developing new photocatalytic materials. These materials not only exhibit better visible light absorption but also maintain high stability and long-term operational efficiency. The novel photocatalysts include organic polymers, metal-organic frameworks (MOFs), covalent organic frameworks (COFs), and various composite materials, all of which show promising potential for use in photocatalytic uranium extraction. MOFs are characterized by high surface area and tunable electronic structures, which enhance their photocatalytic properties for uranium extraction. COFs exhibit high stability and well-ordered structures, making them promising candidates for photocatalysis. The introduction of these novel materials has significantly improved photocatalytic efficiency and extended the application range for uranium extraction.

Organic polymers

g-C₃N₄, an organic polymer in the field of photocatalytic uranium reduction, has become a research focus in recent years due to its relatively narrow bandgap (~2.7 eV), excellent chemical stability, tunable electronic properties, and versatile morphology^[95-97]. The CB of g-C₃N₄ is approximately -1.1 V (vs. NHE), which is more negative than the reduction potentials of UO₂²⁺/U⁴⁺ (0.267 V) and UO₂²⁺/UO₂ (0.411 V), making it thermodynamically capable of driving the reduction of U(VI)^[98]. Additionally, g-C₃N₄ is easily synthesized through a simple process, using various low-cost nitrogen-containing precursors such as melamine, urea, mono- or dicyanamide, and thiourea, making it economically feasible as a photocatalyst^[99,100]. It also exhibits good photocatalytic activity and low environmental toxicity, which makes it promising for applications in seawater uranium extraction and uranium wastewater treatment. However, g-C₃N₄ still faces several challenges, including limited intrinsic active sites, significant recombination of photogenerated electrons and holes, small surface area, and low visible light utilization efficiency, which restrict its overall performance in photocatalytic reactions. To enhance the photocatalytic reduction efficiency of g-C₃N₄ for U(VI), current research focuses on optimizing its electronic structure through methods such as element doping, molecular doping, defect introduction, and the construction of heterojunctions to improve its photocatalytic activity. For example, Yu et al. successfully developed an inexpensive g-C₃N₄-based photocatalyst (AO-C₃N₄) with amidoxime groups for photo-assisted uranium extraction, with the synthesis strategy shown in Figure 3A^[101]. The amidoxime groups served as binding sites for U(VI), while also improving carrier separation efficiency and shortening the band gap, enhancing uranium capture. Additionally, the generated ROS under light endowed AO-C₃N₄ with excellent antibacterial properties, preventing marine biofouling. AO-C₃N₄ achieved 850 μg/g uranium extraction from natural seawater after one week of sunlight exposure, 1.73 times higher than without light. Hu et al. successfully developed cyano-functionalized graphitic carbon nitride (g-C₃N₄-CN) with an isosite structure for efficient photocatalytic uranium extraction from seawater [Figure 3B]^[29]. The cyano groups enhance charge separation, increase electron enrichment, and improve U(VI) adsorption capacity and selectivity, achieving a saturated uranium extraction capacity of 2,644.3 mg/g, outperforming most g-C₃N₄-based photocatalysts, and showing excellent performance under solar light. Metal doping typically introduces new impurity energy levels within the semiconductor's bandgap, which narrows the bandgap width and enhances the photocatalytic activity by improving the light absorption range. Additionally, metal dopants can serve as electron/hole traps, promoting the separation of photogenerated charge carriers and reducing recombination. For example, Gong et al. introduced Fe doping into TiO₂ crystals by creating anion vacancies, significantly enhancing the photocatalytic reduction of U(VI)^[79]. Wei et al. synthesized Fe-doped g-C₃N₄ (Fe-CN) using a molten salt method, which improved the catalyst’s visible light absorption and facilitated the separation of electron-hole pairs, achieving a 96.38% reduction in U(VI)^[102]. Similarly, Zhao et al. developed a nanostructured iron-doped porphyrin-zirconium MOF (NZVI@PCN-224), which demonstrated enhanced uranium adsorption and charge transfer under simulated sunlight^[103]. Wang et al. reported a dual platform based on K⁺ and cyano group co-decorated poly(heptazine imide) (K-CN-PHI) for continuously extracting U(VI) under light or light^[104]. The reaction rate of photocatalytic reduction of U(VI) was 0.89 min^-1, which was 47 times that of pristine carbon nitride (PCN). The XPS analysis [Figure 3C] shows that U(VI) reduction continued for over 72 h in the dark after irradiation due to the long lifetime of the formed reductive radicals. The flexible structure K⁺ as the counterion stabilizes the electrons captured by the cyanamide group, enabling the long lifetime of the generated radicals. Cu-doped crystalline carbon nitride (Cu-CCN-50) composites, synthesized via the molten salt method, exhibited excellent photocatalytic U(VI) photoreduction^[105]. Characterization confirmed successful Cu loading, and photoelectrochemical measurements showed a narrow bandgap and enhanced visible-light absorption [Figure 3D]. Cu-CCN-50 achieved nearly 100% U(VI) removal due to its higher surface area, broader visible-light absorption, and efficient charge carrier separation.

Figure 3. (A) Illustration on the synthesized process of AO-C₃N₄^[101], Copyright 2023, Elsevier Ltd; (B) Schematic diagram of preparing g-C₃N₄-CN^[29], Copyright 2024, Wiley-VCH; (C) The transformation of U 4f XPS spectra on charged K-CN-PHI at different dark exposure times^[104], Copyright 2023, Wiley-VCH; (D) The structure and electrostatic potential of the CCN and Cu-CCN^[105], Copyright 2023, Elsevier Ltd; (E) HRTEM and (F) AFM images of 2D/2D C₃N₅/GO nanosheets^[108], Copyright 2021, Elsevier Ltd. XPS: X-ray photoelectron spectroscopy; K-CN-PHI: K⁺ and cyano group co-decorated poly(heptazine imide); CCN: crystalline carbon nitride; HRTEM: high-resolution transmission electron microscopy; AFM: atomic force microscopy.

Non-metal doping, such as O, phosphorus (P), sulfur (S), fluorine (F), and others, can further enhance the photocatalytic performace by increasing the conjugated π-electron systems in the material, improving electron conductivity and promoting the migration and separation of charge carriers. For example, Xue et al. doped Br into g-C₃N₄ (CNBr_x) using ammonium bromide and urea as precursors^[106]. The introduction of Br narrowed the bandgap, enhanced visible light absorption, and increased the surface area, improving the separation efficiency of photogenerated charge carriers. Defect engineering has been shown to be an effective method to improve the photocatalytic performance of materials in solar-to-chemical energy conversion. This approach creates controlled defects within the material, which can tune its electronic structure, charge dynamics, and active surface sites. Defects can act as active catalytic sites and promote better charge separation. For instance, Chen et al. fabricated a MoS₂-reduced graphene oxide (RGO) composite, where the concentration of S vacancies was controlled to regulate the band structure of MoS_x and RGO, effectively promoting U(VI) reduction^[107]. Meng et al. developed a metal-free 2D/2D C₃N₅/GO heterojunction for efficient uranium extraction through adsorption and photoreduction [Figure 3E and F]^[108]. The material achieved 96.1% uranium removal at 10 ppm, maintained 92.8% removal after gamma-ray exposure, and retained 90.4% efficiency after five cycles. This work advances the design of nonmetal-based 2D/2D nanohybrids for uranium removal from nuclear wastewater.

In addition to g-C₃N₄, polymers have also demonstrated some photocatalytic reduction capabilities for uranium. Ma et al. developed a bifunctional porous aromatic framework (PAF-AN₂T₈-AO) combining adsorption and photocatalytic enrichment groups for efficient uranium recovery from nuclear wastewater [Figure 4A]^[109]. Under simulated sunlight, PAF-AN₂T₈-AO achieved 99% uranium recovery, outperforming PAFs with single functional groups by 1.2 and 2.3 times. The mechanism revealed the formation of air-stabilized metastudtite [(UO₂)O₂·2H₂O], demonstrating the material’s high efficiency and resistance to interfering ions. This work provides a promising strategy for designing bifunctional materials for uranium extraction. Yu et al. designed and synthesized a tunable perylene-based conjugated microporous polymer (CMP) using molecular engineering strategies^[110]. By enhancing visible light absorption, improving interface electron-hole separation and transfer, and increasing the number of coordination sites, they achieved a 97.8% removal rate of U(VI) within 60 min. In another recent study from the same team, Yu et al. developed three novel quinone-containing biomimetic donor-acceptor CMPs with full spectrum response, synthesized via Sonogashira-Hagihara coupling polymerization, for uranium reduction [Figure 4B]^[111]. The quinone-containing blocks in these CMPs played a crucial role as both adsorption and photocatalytic active sites, enhancing binding affinity for uranyl ions and facilitating efficient electron-hole pair separation and transport. ECUT-TQ [a conjugated microporous polymers (CMPs) by 1,3,6,8-tetraethynylpyrene (TPy) react with 2,6-dibromoanthracene-9,10-dione (BQ)], with the strongest acceptor and enhanced built-in electric field, achieved a high U(VI) removal rate of 97.4%. This biomimetic approach offers a promising strategy for designing efficient CMPs to purify radioactive wastewater. Similarly, Chen et al. synthesized a new class of CMPs based on porphyrins and thiophene through C-H arylation^[112]. These materials retained high electron transfer capacity in acidic solutions (pH = 2-5) and demonstrated good catalytic activity for U(VI), with a removal rate of approximately 80%.

Figure 4. (A) The synthesis of PAF-AN_XT_10-X-AO (X = 0,2,5,8,10) for photocatalytic extraction of uranium^[109], Copyright 2024, Wiley-VCH; (B) The mechanism of exciton dissociation, charge separation and photocatalytic removal of uranium in biomimetic D-A system^[111], Copyright 2023, Elsevier Ltd.

MOFs

In recent years, MOFs have gained widespread attention due to their high surface area, porosity, and tunable electronic structures. MOFs are a class of porous materials composed of metal ions/clusters coordinated with organic linkers, possessing well-defined molecular structural units, adjustable functional groups, and effective active sites^[113]. These properties allow MOFs to function simultaneously as photosensitizers (organic ligands) and catalytic centers (metal clusters/nodes). This unique structure enables MOFs to absorb visible light and efficiently separate photogenerated electrons and holes, making them promising candidates for photocatalytic uranium extraction. To date, there have been fewer reports on the photocatalytic reduction of U(VI) by MOFs compared to their adsorption-based removal of U(VI). Their behavior under light irradiation is similar to that of semiconductor photocatalysts, attributed to metal charge transitions [ligand to metal charge transitions (LMCT)], metal to ligand charge transitions (MLCT), or π-π^* transitions in the aromatic ligands^[114]. Li et al. proposed a post-synthetic ligand functionalization strategy to create phosphoric acid-functionalized and amine-functionalized Zr-based MOFs (PN-PCN-222), achieving selective uranium capture and photocatalytic reduction under visible light^[50]. This modified MOF demonstrated excellent uranium extraction performance across a wide range of uranium concentrations and pH values.

However, despite the abundant metal nodes in MOFs serving as active sites, their suitability as photocatalysts remains limited. The metal/ligand coordination structure dictates their wide bandgap (Eg > 5.5 eV), preventing efficient absorption of visible light^[115]. For example, ZIF-8 [Zn₆(mIm)₁₂: mIm = 2-methylimidazole ligand] has a bandgap of 5.8 eV, which makes it unsuitable as a photocatalyst under UV-visible light^[116]. To address this limitation, selecting appropriate MOF components can significantly enhance their photocatalytic activity. For instance, Zhang et al. designed and synthesized a polyoxometalate (POM)-organic framework material (SCU-19) with a rare inclined polycatenation structure^[117]. They conducted uranium separation experiments, demonstrating that under visible light irradiation, an additional U(VI) photocatalytic reduction mechanism occurs, achieving higher uranium removal rates and faster adsorption kinetics without reaching saturation.

Furthermore, to overcome these challenges, researchers have employed various modification strategies, such as functionalizing MOF ligands, surface grafting of functional groups, and forming heterojunctions with other semiconductors, to improve the photocatalytic performance and broaden the application scope of MOFs. For example, Liu et al. developed a bifunctional photocatalyst, MOF525@BDMTp, with MOF525 as the core and BDMTp as the shell, utilizing a type-II heterojunction strategy [Figure 5A]^[118]. This design improves electron-hole separation and carrier migration, enhancing the photocatalytic reduction of U(VI) and oxidation of chlorpyrifos (CP). The built-in electric field further boosts electron transfer, activating U(VI) binding sites. MOF525@BDMTp achieved a U(VI) removal rate of 625.0 mg/g and 94.0% CP degradation, demonstrating its effectiveness in complex environments. Bi et al. developed Ti-MOF@DATp, a bifunctional photocatalyst that efficiently reduces U(VI) and oxidizes tetracycline through a Z-scheme heterojunction [Figure 5B]^[119]. The catalyst achieved 96% U(VI) and 90% tetracycline removal from their mixture, significantly outperforming individual removal, with a removal rate constant 55 times higher for the mixture than U(VI) alone. In the study by Chen et al., UiO-66-NH₂/black phosphorus quantum dots (MOF/BPQDs) heterojunctions were anchored on carboxyl cellulose nanofiber aerogels for efficient uranium adsorption [Figure 5C and D]^[120]. Under light irradiation, BP@CNF-MOF demonstrated a 55.3% increase in uranium adsorption efficiency, reaching 6.77 mg-U per g after 6 weeks in natural seawater, while also effectively destroying marine bacteria through photocatalytic generation of ROS [Figure 5E].

Figure 5. (A) The synthesis of MOF525@BDMTp and the mechanism of removing U(VI)^[118], Copyright 2023, Elsevier Ltd; (B) The composite material Ti-MOF@DATp^[119], Copyright 2024, Elsevier Ltd; Images of lightweight (C) CNF and (D) BP@CNF-MOF aerogels, (E) Uranium extraction of BP@CNF-MOF in natural seawater for 6 weeks under light and dark conditions^[120], Copyright 2021, Wiley-VCH.

COFs

COFs are another class of materials that have gained attention for their high chemical stability, well-defined pore structures, and tunable electronic properties. They are crystalline materials held together by strong covalent bonds, and they offer excellent chemical stability. COFs can be designed to absorb visible light and facilitate photocatalytic reactions, making them ideal candidates for uranium extraction. Cui et al. reported the synthesis of a semiconductor COF with excellent photocatalytic activity^[121]. The high photocatalytic activity of NDA-TN-AO (A highly planar conjugated naphthalene-based sp2) enabled it to generate ROS that disrupted bacteria, demonstrating significant anti-biofouling ability. Additionally, NDA-TN-AO facilitated the reduction of adsorbed U(IV) to insoluble U(IV), significantly improving uranium extraction capabilities. Due to the photoelectric effect, NDA-TN-AO displayed a strong electrostatic attraction toward [UO₂(CO₃)₃]^4- ions, even in natural seawater. Under simulated sunlight, the uranium adsorption capacity of NDA-TN-AO reached 6.07 mg/g, which was 1.33 times higher than under dark conditions. A COF-based adsorption-photocatalysis strategy for selective uranium removal from seawater without sacrificial reagents was developed by Hao et al.^[122]. The quaternary COF [Figure 6A], COF 2-Ru-AO, with electron-rich and electron-deficient linkers, achieved a high uranium uptake capacity of 2.45 mg/g/day and demonstrated excellent anti-biofouling properties, surpassing most current adsorbents. Another novel strategy using Ti-oxo clusters (TiOCs) encapsulated within a photosensitive COF (TiOCs∈COF-TZ) [Figure 6B] for efficient photocatalytic uranium extraction^[123]. This approach allows TiOCs to reduce [UO₂(CO₃)₃]^4- in natural seawater under visible-light irradiation, achieving 89.9% uranium extraction. Compared to unloaded COF-TZ and surface-loaded TiOCs@COF-TZ, TiOCs∈COF-TZ demonstrates superior catalytic activity and efficiency.

Figure 6. (A) Schematic of multivariate COFs for photocatalytic extraction of uranium^[122], Copyright 2023, American Chemical Society; (B) Schematic of the synthesis of TiOCs∈COF-TZ^[123], Copyright 2023, Elsevier Ltd. COFs: Covalent organic frameworks.

The development of COFs has been focused on improving their electronic properties to enhance charge separation and light absorption. Chen et al. synthesized a series of COFs with varying excited-state electron distributions, charge transport properties, and porosity characteristics [Figure 7A]^[124]. Through a combination of experimental studies and theoretical calculations, they investigated and optimized the material’s performance, achieving a uranium extraction capacity of 6.84 mg/g from natural seawater. Yu et al. employed a π-skeleton engineering strategy to synthesize three COFs containing geminal oxime units with donor-acceptor olefin linkages [Figure 7B]^[125]. They discovered that TTh-COF-AO, which possesses a planar π-framework donor, exhibited a significantly reduced exciton binding energy (E_b), facilitating e^-~h⁺ pair separation and enhancing its photocatalytic activity. Under visible light irradiation, this material achieved a uranium extraction capacity of 10.24 mg/g from natural seawater. Yang et al. utilized the Povarov reaction to convert propenoxy groups within the COF structure into stable chromenoquinoline rings^[126]. Additionally, electron-donating and electron-withdrawing groups were introduced onto the linker units, inducing local polarization to regulate charge distribution and enhance charge carrier separation. This approach resulted in the formation of a new intramolecular donor-acceptor system, achieving a uranium extraction capacity of 8.02 mg/g from natural seawater. In an exciting breakthrough, Ma et al. developed a series of triazine-connected two-dimensional COFs nanowires with an outer diameter of 50 to 200 nm by solvent modulation method [Figure 7C and D]^[127]. This unique morphology enables excellent photocatalytic performance, and U(VI) extraction capacity of N₃-COF₆₀ reached 11.35 mg/(g·d). This performance ranks among the highest reported for COF-based uranium extraction systems. The recent developments of photocatalysts for U(VI) extraction from seawater are summarized in Figure 7E. However, the complex synthesis and high cost of COFs present challenges for their large-scale application.

Figure 7. (A) Different structures of COF-1 to COF-4 to modulate the excited state electronic structure and charge transport properties for tuning the photocatalytic activities for uranium extraction^[124], Copyright 2023, Springer Nature; (B) Illustration of the energy levels summarizing the main processes involved in charge transfer^[125], Copyright 2024, Wiley-VCH; (C) Preferential growth of N₃-COF along the 001 direction and (D) SEM and TEM (inset) images of N₃-COF and N₃-COF_x^[127], Copyright 2024, American Chemical Society; (E) Recent performance of photocatalysts used for uranium extraction from seawater. Blue: Total uranium uptake capacity; Orange: everyday uranium uptake capacity. COFs: Covalent organic frameworks; TEM: transmission electron microscopy; SEM: scanning electron microscopy.

In addition to the strategies mentioned above, recent studies have also explored a novel approach where COF materials are used for photocatalytic hydrogen peroxide (H₂O₂) production, which is then utilized to reduce uranium [U(VI)] from seawater. This approach capitalizes on the ability of COFs to generate H₂O₂ under light irradiation, which can serve as an effective reagent for U(VI). This dual-function mechanism photocatalytic H₂O₂ generation followed by uranium reduction, adds another layer of efficiency to COF-based photocatalysts, making them even more promising for environmental remediation applications. Li et al. developed a novel strategy by introducing ethenyl groups and piezoelectric properties into Cz-COFs, enhancing charge transfer and creating polarized fields for efficient piezo-photocatalytic H₂O₂ production [Figure 8A]^[128]. The ethenyl group-modified bicarbazole-based COFs (COF-DH-Eth) achieved a high H₂O₂ synthesis rate of 9,212 μmol/g/h. This generated H₂O₂ effectively promotes the reduction of U(VI) to U(IV), leading to the rapid formation of studtite (UO₂O₂·2H₂O). This approach also demonstrates significant potential for efficient uranium extraction through in-situ utilization of photogenerated H₂O₂. Guo et al. developed a novel strategy for uranium extraction using COF-based photocatalysis, where H₂O₂ is produced and used to reduce uranium [U(VI)] to U(IV)^[129]. By incorporating topological defects into a hcb-type COF, they significantly enhanced H₂O₂ production, achieving a 2.7-fold increase compared to defect-free COFs [Figure 8B]. This defect engineering improves the material’s electronic structure, promoting more efficient O₂-to-H₂O₂ conversion. In the two-step photocatalytic process, the COF photocatalyst first generates H₂O₂ via light-driven O reduction. In the second step, the produced H₂O₂ reacts with UO₂²⁺, reducing it to U(IV) and forming the stable mineral studtite. This reduction occurs rapidly in the presence of H₂O₂, enabling the transformation of U(VI) to U(IV) without the need for additional sacrificial reagents. The two-step process achieves up to 98.8% uranium extraction efficiency, demonstrating both high extraction and selectivity. This approach provides a promising solution for uranium extraction from seawater and other sources. Xie et al. developed 3D space-connected donor-acceptor systems in 8-connected COFs (COF-1 and COF-2) to enhance electron transfer and solar energy utilization under visible light [Figure 8C]^[130]. COF-2, with an extended π-conjugated structure, achieved high H₂O₂ production (6.93 mmol/g/h) in natural seawater without sacrificial reagents. Additionally, COF-2 demonstrated rapid uranium extraction, achieving 97% removal in 1 ppm spiked tap water and 92% in groundwater within 360 min. XPS analysis revealed the coexistence of U(VI) and U(IV) species, indicating that the 3D D-A system facilitates efficient photocatalytic uranium reduction and extraction, highlighting its potential for environmental remediation.

Figure 8. (A) Synthesis piezo-photocatalytic COFs of COF-DH-H and COF-DH-Eth^[128], Copyright 2024, Wiley-VCH; (B) The synthesis of COF-hcb and COF-hcb-defect^[129], Copyright 2024, Chinese Chemical Society; (C) Illustration of the synthesis of 3D COFs and topology^[130], Copyright 2024, Wiley-VCH. COFs: Covalent organic frameworks.