Relevant peer-reviewed studies were identified through searches of Web of Science and Scopus using combinations of terms related to rare-earth elements, metal oxides, photocatalysis, and dye degradation. The search included “rare-earth doped”, “lanthanide”, and individual rare-earth elements such as Ce, La, Nd, Gd, Sm, Eu, and Yb, combined with “metal oxide”, “TiO₂”, “ZnO”, “SnO₂”, “CeO₂”, “WO₃”, “photocatalytic degradation”, and “dye degradation”. Backward citation tracking of relevant reviews and primary studies was also performed. The final selected RE-doped studies covered the period 2019-2026.

The literature search was conducted as a structured review rather than as a prospectively registered systematic review. Because the original database exports and record-level screening files were not retained, exact numbers of records identified, duplicates removed, and records excluded at each screening stage could not be reconstructed retrospectively. The final dataset therefore comprises the 29 studies that met the predefined inclusion criteria and contained sufficient experimental information for the comparative analysis.

Studies were included when they investigated a defined rare-earth dopant incorporated into a metal oxide photocatalyst, reported quantitative photocatalytic degradation of an organic dye, and provided sufficient experimental information for meaningful comparison. Studies focused only on adsorption or lacking sufficient photocatalytic data were excluded.

A final dataset of 29 RE-doped metal oxide studies was established. Information was extracted on the host oxide, rare-earth dopant and concentration, morphology, synthesis method, structural and optical properties, dye and concentration, catalyst loading, pH, light source, irradiation time, and kinetic data where available. Mineralization indicators (TOC/COD), apparent quantum yield (AQY), stability/reusability, leaching, and real wastewater testing were also recorded when reported. The extracted information is presented in Table 1 and Table S1.

The extracted data were interpreted in light of differences in experimental conditions, including catalyst loading, dye concentration, pH, irradiation conditions, reaction time, and kinetic parameters. Because irradiation intensity at the reaction surface was not consistently reported, direct normalization of light input was not possible. Reported removal efficiencies were therefore not used alone to rank photocatalysts.

Mechanistic interpretations were evaluated according to the experimental evidence reported in each study. Mechanisms supported by complementary characterization or experimental measurements were distinguished from interpretations inferred mainly from photocatalytic performance. This assessment considered evidence such as photoluminescence, electrochemical measurements, XPS, ESR/EPR, scavenger tests, and photocurrent measurements. The detailed classification and its results are presented in Section 6.1.

For other reported variables, catalyst loading, pollutant concentration, and solution volume were retained as explicit comparison parameters in the extracted dataset; however, a single normalized activity metric was not calculated because the studies differed substantially in reactor configuration, irradiation conditions, and reporting completeness.

Representative studies on transition-metal doping, nanocomposites, and core-shell structures were considered separately to provide comparative context for rare-earth doping rather than as additional exhaustive datasets.

Study quality was assessed based on the completeness of experimental reporting and the availability of relevant control experiments, replication or uncertainty information, and supporting mechanistic evidence. To reduce selection bias, predefined search terms and inclusion/exclusion criteria were consistently applied across searches of Web of Science and Scopus, and backward citation tracking was used to identify relevant studies that the database searches may not have retrieved. Publication bias was not statistically assessed because the review was not designed as a quantitative meta-analysis; therefore, preferential publication of positive photocatalytic results cannot be excluded. These limitations were considered when interpreting the reported performance and mechanistic evidence.

A formal quantitative meta-analysis was not performed due to substantial heterogeneity across studies in catalyst composition, dye characteristics, catalyst loading, irradiation conditions, reactor configuration, reaction time, and reported response parameters. These differences prevent reliable pooling of effect sizes without introducing substantial assumptions or potentially misleading comparisons. A bibliometric analysis was also outside the scope of this review, which focuses on scientific evidence, mechanistic understanding, photocatalytic performance, and practical applicability rather than publication trends or research-network mapping.

Band-gap energy and band-edge positions determine which wavelengths can excite a semiconductor and whether the resulting charge carriers possess sufficient redox potential for surface reactions. TiO₂, ZnO, and SnO₂ have relatively wide band gaps and therefore absorb predominantly in the ultraviolet region [57,58]. Narrowing the band gap can extend light absorption into the visible region, but this modification may also alter the redox potential and structural stability of the material.

This trade-off is important when evaluating photocatalyst modification. Increasing visible-light absorption does not necessarily produce higher photocatalytic activity if the resulting band positions no longer provide sufficient driving force for the required oxidation or reduction reactions. Therefore, effective band-structure engineering should balance light harvesting, redox potential, and structural stability rather than target band-gap narrowing alone. Figure 2 compares representative semiconductor band gaps and illustrates the relationship between optical absorption and photocatalytic activation.

Following photoexcitation, electrons and holes must reach reactive surface sites before recombination. Their transport is influenced by crystallinity, particle size, morphology, defect states, and surface structure. Defects can either promote or hinder charge separation depending on their concentration and electronic character. For example, oxygen-vacancy engineering has been reported to create trapping states that can influence carrier separation, whereas excessive defect formation may introduce recombination centres [63]. This dual behaviour is important for the modification strategies discussed later in this review. A change in photoluminescence intensity or defect concentration should not, by itself, be interpreted as proof of improved charge separation. Stronger mechanistic conclusions require complementary evidence from techniques such as photoluminescence, electrochemical measurements, transient spectroscopy, or radical-scavenging experiments. This distinction is particularly relevant when evaluating the proposed mechanisms of rare-earth doping.

Once charge carriers reach the catalyst surface, they can participate in reactions with adsorbed water, hydroxyl groups, dissolved oxygen, and pollutants. Conduction-band electrons can reduce dissolved oxygen to superoxide-related species. At the same time, valence-band holes can oxidize water or surface hydroxyl groups to generate hydroxyl radicals when the corresponding redox potentials are sufficiently favorable. These reactive species can then participate in oxidation reactions leading to dye transformation [64]. The relative contribution of individual reactive species is not universal. It depends on band-edge positions, pH, dissolved oxygen, surface chemistry, adsorption behaviour, and the specific catalyst-pollutant system. Therefore, assignment of ^•OH, ^•O₂^-, holes, or other reactive species should be based on appropriate experimental evidence, such as scavenger experiments, ESR/EPR measurements, or complementary spectroscopic analysis, rather than inferred from removal efficiency alone. This evidence-based distinction is applied throughout the subsequent analysis of RE-doped photocatalysts.

Photocatalytic dye degradation generally involves a sequence of molecular transformations rather than a single reaction step. Initial reactions may disrupt chromophoric groups, followed by transformation of aromatic fragments and, under sufficiently strong oxidation conditions, further conversion toward smaller molecules and mineralization products. The pathway depends on the dye structure, substituent groups, catalyst surface, reactive species, and reaction conditions.

The reaction pathway can also differ among dye classes. For example, azo dyes can undergo cleavage of the azo bond under appropriate photocatalytic conditions [65]. However, the disappearance of the parent dye or loss of colour does not by itself demonstrate complete mineralization. Detection of intermediates, TOC/COD reduction, or other product analyses is required to establish the extent of mineralization and to evaluate whether potentially persistent or toxic transformation products remain [66].

Photocatalytic performance results from the interaction between catalyst characteristics, including surface area, crystallinity, morphology, defect structure, and surface chemistry, and operating conditions such as pH, pollutant concentration, catalyst loading, irradiation wavelength and intensity, reaction time, and temperature [67,68]. These parameters can influence light absorption, adsorption, charge-carrier behaviour, and reactive-species formation, and their effects are often interdependent.

For example, pH can modify catalyst surface charge and pollutant speciation, while increasing pollutant concentration can reduce photon penetration and alter the availability of reactive surface sites. Catalyst loading also requires optimization, as insufficient catalyst loading limits the available active surface, whereas excessive loading can increase light scattering and attenuation [69,70,71]. In complex wastewater, dissolved organic matter and other constituents can additionally consume reactive species or interfere with light penetration, altering photocatalytic performance differently from that observed in simplified dye solutions [72,73].

These dependencies have important implications for cross-study comparison. A reported removal percentage cannot be interpreted independently of catalyst loading, pollutant concentration, pH, irradiation conditions, reaction time, and wastewater composition. Accordingly, the comparative analysis in this review considers these conditions where they were reported rather than ranking photocatalysts solely by removal efficiency.

The photocatalytic behaviour of metal oxide nanoparticles depends on crystallinity, particle size, morphology, surface hydroxylation, defect concentration, and electronic structure. Morphological characteristics can influence light scattering, active-site accessibility, and charge-transport pathways [78]. A recent study of morphology-engineered ZnO tetrapods further illustrates this structure-performance relationship, showing that a highly crystalline three-dimensional architecture with controlled defect characteristics can provide rapid photocatalytic degradation even without compositional doping [79]. Surface hydroxyl groups can participate in the formation of reactive species, while defects may act as charge-trapping sites; however, their effects depend on defect type and concentration [80,81].

CeO₂ provides an illustrative example of this balance. Oxygen defects and associated Ce³⁺ species can modify the optical and electronic properties of CeO₂, whereas excessive defect formation may promote recombination rather than improve photocatalytic activity [82,83,84,85]. Thus, defect engineering should be controlled rather than treated as inherently beneficial.

TiO₂ also illustrates the importance of crystal structure. It occurs in anatase, rutile, and brookite polymorphs, and differences in crystal structure can influence charge-carrier behaviour and surface reactivity. ZnO has also attracted extensive attention due to its favourable photocatalytic properties, although its stability under irradiation remains an important consideration [76,77].

These examples show that the performance of pristine metal oxides cannot be attributed to a single material parameter. Morphology, crystal structure, defect chemistry, surface properties, and operating conditions can interact, making it difficult to directly attribute performance changes to a single structural feature.

Despite their useful photocatalytic properties, pristine metal oxides have limitations that restrict their practical application. TiO₂, ZnO, and SnO₂ have relatively wide band gaps and therefore absorb predominantly in the ultraviolet region, limiting their direct use of visible solar radiation [57,83,84,85]. Extending light absorption toward the visible region can improve photon utilization, but changes in the band structure may also affect the redox driving force and material stability. ZnO presents an additional stability concern because photocorrosion can occur during irradiation, potentially resulting in catalyst dissolution and loss of material from the treatment system [86]. More generally, catalyst performance can decrease when the reaction environment differs substantially from the simplified conditions used in laboratory experiments. For example, TiO₂^- and ZnO^-based photocatalytic treatment of textile effluents has shown that wastewater composition can influence treatment performance [87].

These limitations indicate that high activity of a pristine oxide is not sufficient to establish practical suitability. Catalyst stability, light utilization, interaction with the pollutant, and behaviour in complex wastewater must also be considered. These limitations illustrate why modification strategies are being investigated to improve the functional properties and practical performance of metal oxide photocatalysts, as summarized schematically in Figure 3.

The limitations of pristine metal oxides have motivated modification strategies including elemental doping, defect engineering, heterojunction formation, nanocomposite construction, and morphology control. These approaches aim to modify specific properties such as light absorption, charge-carrier separation, surface reactivity, or structural stability.

However, modification does not guarantee improved photocatalytic performance. Changes in defect density, morphology, surface chemistry, or band structure can produce beneficial or adverse effects depending on the material and operating conditions [86,88]. Therefore, modification strategies should be evaluated mechanistically rather than solely by changes in reported dye-removal percentages.

Among these approaches, rare-earth doping is particularly relevant because RE ions can alter the electronic and structural characteristics of metal oxides through their distinctive electronic configurations and interactions with the host lattice. The following section therefore examines RE-doped metal oxides while considering the host oxide, dopant concentration, morphology, synthesis route, operating conditions, and available mechanistic evidence.

RE incorporation can influence photocatalytic behaviour by altering charge-carrier trapping, defect chemistry, interfacial charge transfer, and surface reactions. The studies compiled in Table 1, however, do not provide the same level of evidence for these effects. Among the 29 studies, 15 (51.7%) were classified as providing direct mechanistic evidence, whereas 14 (48.3%) provided partial evidence. This distinction is important because improved dye removal alone does not establish the pathway responsible for the observed enhancement.

The type of characterization used to support the proposed mechanisms also varies considerably across the dataset. Among the 29 studies, PL analysis was reported in 27 studies (93.1%), XPS in 27 studies (93.1%), EIS in 11 studies (37.9%), ESR/EPR in 3 studies (10.3%), and scavenger tests in 22 studies (75.9%) (Table 1; Table S1). The frequent use of PL and XPS indicates that changes in optical and surface properties are commonly examined. However, these techniques alone do not establish a complete charge-transfer pathway or identify the reactive species responsible for degradation. More convincing mechanistic assignments are obtained when complementary evidence, such as electrochemical measurements, radical-trapping experiments, and spectroscopic identification, is combined.

The individual studies also show that the effect of RE incorporation is system-dependent. For example, Ce-doped SnO₂ hollow spheres were supported by PL, EIS, XPS, and scavenger evidence. In contrast, the Ce-doped SnO₂ nanoparticles in another study relied on a more limited combination of PL, XPS, and scavenger tests [25,26]. Similarly, the Gd/Nd co-doped ZnO system provided only partial mechanistic evidence despite the reported photocatalytic improvement [27]. These differences illustrate why the presence of an RE dopant cannot, by itself, be taken as evidence of a particular charge-transfer or radical-generation pathway.

The quantitative comparison further supports this caution. The mean removal efficiency was 92.7% for studies classified as providing direct mechanistic evidence and 93.1% for those classified as providing partial mechanistic evidence. Thus, the two groups showed very similar reported removal efficiencies despite a clear difference in the amount of mechanistic support. High photocatalytic removal therefore does not necessarily indicate stronger mechanistic verification. The result also underscores the need to distinguish between measured evidence and mechanistic interpretation when comparing RE-doped photocatalysts. Because most studies reported lamp power rather than irradiance at the reaction surface, direct normalization of the incident light intensity across studies was not possible; therefore, reported removal efficiencies are interpreted within their respective experimental conditions rather than as directly comparable intrinsic activities. The distribution of mechanistic evidence and practical validation across the reviewed studies is summarized in Figure 4, highlighting the extent to which key performance and validation parameters were reported.

A further limitation is the relatively low use of techniques that can provide complementary evidence for charge-transfer behaviour and reactive oxygen species. Only 11 of the 29 studies included EIS, 3 included ESR/EPR, and none reported an apparent quantum yield (AQY). TOC analysis was reported in 10 studies (34.5%), indicating that mineralization was assessed in substantially fewer studies than dye removal. The nearly equal distribution between direct and partial mechanistic evidence therefore suggests that mechanistic claims remain unevenly supported across the reviewed RE-doped systems.

In this review, reported dye removal is treated as decolourisation or pollutant removal unless supported by TOC, COD, or other product-analysis data; therefore, removal efficiency alone is not interpreted as evidence of complete mineralization or toxicity reduction.

RE incorporation can modify photocatalytic behaviour by altering crystal structure, particle morphology, surface properties, and optical response. The current dataset shows substantial morphological diversity, including hollow spheres, nanorods, nanocubes, nanospheres, core-shell structures, nanoflowers, nanofibers, thin films, and nanodispersions (Table 1). This variation is important because morphology is not an independent parameter in most studies; it is closely linked to the synthesis route, dopant concentration, and host oxide.

The Ce-doped SnO₂ studies illustrate the structural diversity reported across the literature. One study prepared Ce-doped SnO₂ as hollow spheres with a measured surface area of 62.62 m² g^-1, whereas another prepared Ce-doped SnO₂ as nanoparticles [25,26]. The hollow-sphere study also showed that Ce incorporation did not alter the overall hollow morphology, indicating that the observed structure cannot be attributed to doping alone [25]. These results highlight the need to consider synthesis route and pre-existing particle architecture when interpreting the role of RE incorporation. The available studies do not allow the individual contributions of morphology, dopant concentration, and reaction conditions to be separated.

Similar morphological diversity occurs for ZnO. Gd/Nd co-doped ZnO was prepared as nanorods, Ce-doped ZnO as nanorods and nanospheres, and Er-, La-, Nd-, and Sm-containing ZnO systems were reported as nanoparticles, nanoflowers, nanofibers, nanocomposites, or other nanostructures [28,29,30,31,32]. These studies indicate that RE incorporation can result in different final morphologies rather than a single characteristic structural form. The variation also suggests that the synthesis method remains an important factor in interpreting the role of the dopant.

Changes in optical properties are also frequently reported. Band-gap values were available for 26 of the 29 studies (89.7%), indicating that optical characterization is much more commonly reported than surface-area measurements. Several systems show lower band-gap values after RE modification or report optical features associated with improved visible-light response. For example, the Ce-doped ZnO/TiO₂ system reported a band gap of 2.62 eV, while the Yb/Er co-doped TiO₂-ZnO system showed a band gap of 2.92 eV [92,93]. Gd-doped ZnO systems also showed band-gap values around 2.9-3.1 eV depending on the catalyst composition and preparation conditions [33,94,95]. These results support the view that RE incorporation can alter the optical response, but they do not establish that band-gap reduction alone is responsible for the observed photocatalytic activity.

The relationship between optical properties and photocatalytic performance is further complicated by the different irradiation conditions used in the studies. The dataset includes UV, visible-light, sunlight, and mixed irradiation conditions. Consequently, a lower reported band gap cannot be interpreted independently of the incident light spectrum, irradiation power, exposure time, and catalyst loading. This is particularly important for studies reporting visible-light or solar-light activity, where enhanced absorption may contribute to performance but does not by itself demonstrate more efficient charge separation or reactive-species formation.

Surface-area data provide another example of the limitations in the current literature. Only 6 of the 29 studies reported specific surface area values in the extracted dataset. The available values vary widely, but the limited number of observations and differences in morphology, catalyst loading, pollutant concentration, and irradiation conditions prevent the establishment of a meaningful quantitative relationship between surface area and dye removal. The structural information should therefore be interpreted as supporting evidence rather than as a basis for ranking the catalysts.

Taken together, the dataset indicates that RE incorporation can affect morphology and optical properties, but these changes are strongly coupled with host composition and synthesis conditions. The evidence does not support a single structural or optical feature as the dominant explanation for improved photocatalytic performance. Instead, the reported activity appears to reflect the combined effects of composition, morphology, optical response, surface properties, and charge-carrier behaviour, with their relative contributions varying among RE-oxide systems.

Crystallographic evidence also shows that the effect of RE incorporation is host-dependent. For example, the TiO₂ systems reported in the dataset include anatase phases, while ZnO systems retain the hexagonal wurtzite structure [34,35,36,37]. In one Nd-doped TiO₂ system, the undoped material contained anatase and rutile phases, whereas the Nd-doped material was reported as pure anatase. Other studies reported changes in crystallite size following RE incorporation, including the Nd-doped ZnO and Sm-doped ZnO systems [38,39]. These observations indicate that structural changes accompanying RE incorporation should be evaluated experimentally rather than inferred solely from photocatalytic performance.

Pore-size information was not reported in the reviewed studies and therefore could not be included in the comparative analysis.

The synthesis route also influences the structural characteristics of RE-doped metal oxides. The reviewed studies employed sol-gel, hydrothermal, precipitation, combustion, electrospinning-based, and solvothermal approaches; a detailed comparison of their principal advantages, limitations, and typical structural effects is provided in Table S2. Sol-gel and precipitation routes provide relatively simple preparation and compositional control, whereas hydrothermal processing can promote controlled crystal growth and morphology. Combustion routes offer rapid synthesis but may provide less control over particle growth, while electrospinning-based methods are useful for producing one-dimensional structures. However, these effects cannot be attributed solely to the synthesis route because the resulting morphology and photocatalytic performance are also affected by the host oxide, RE concentration, calcination conditions, and reaction parameters. Thus, the present dataset supports the synthesis route as an important structural variable but does not justify ranking one method as universelectronic structureally superior.

The photocatalytic response of RE-doped metal oxides varies with dopant concentration, but the present dataset does not support a single optimum concentration applicable across different host oxides. The reported RE concentrations span approximately 0.1-10%, with several studies identifying an optimum within this range for their specific catalyst system (Table 1; Table S1). For example, the La-doped TiO₂ systems reported optimum concentrations of 1.5-2%, while Gd/Nd co-doped ZnO showed an optimum of 1.5% [31,34,35,40,41]. Nd-doped ZnO and Sm-doped ZnO systems also showed concentration-dependent activity, but their reported optimum values differed between studies and morphologies [38,39,42].

This variation is also evident when the same RE element is introduced into different host oxides. The optimal concentration of La³⁺, Nd³⁺, and Sm³⁺ is not constant across TiO₂, ZnO, WO₃, and ZrO₂ systems (Table 1). Such differences indicate that the effect of dopant concentration depends on the host lattice, synthesis route, morphology, and reaction conditions. Therefore, the concentration associated with the highest removal in one system cannot be directly transferred to another.

At low concentrations, RE incorporation may modify the electronic structure and provide sites that influence charge-carrier behaviour. Increasing the dopant concentration can produce a different response if excessive incorporation alters crystallinity, promotes defect accumulation, or changes the surface structure. The available studies, however, do not provide sufficient standardized data to establish a quantitative relationship between concentration and performance across the full dataset. This limitation is particularly important because the studies use different host materials, pollutants, irradiation conditions, and catalyst loadings.

The dataset also shows that high photocatalytic activity is not restricted to a narrow range of dopant concentration. Several catalysts achieved removal efficiencies above 95% at concentrations below 2%, whereas other systems showed high activity at higher concentrations. For example, Ce-doped SnO₂, La-doped TiO₂, Nd-doped TiO₂, and Sm-doped ZnO achieved high reported removal under their respective experimental conditions at different dopant levels [34,35,42]. These results further argue against defining a universal concentration window solely from reported removal percentages.

A further limitation is that concentration optimization is not consistently investigated over a sufficiently broad range. Some studies compare several dopant concentrations, whereas others report only one selected composition. Consequently, the concentration identified as “optimal” in an individual study should be interpreted as an experimental optimum for that specific catalyst and reaction system rather than as a general optimum for RE doping. Systematic concentration series, combined with structural, spectroscopic, and kinetic measurements, would provide a stronger basis for identifying the concentration range at which RE incorporation improves charge separation without introducing adverse recombination or structural effects.

Real-wastewater validation remains largely unexplored in the reviewed RE-doped systems. None of the 29 studies included in the present dataset reported photocatalytic testing using real wastewater. The reported efficiencies therefore reflect performance in controlled model solutions and cannot be assumed to represent behaviour in complex effluents. In real textile wastewater, dissolved organic matter, inorganic ions, salinity, turbidity, and surfactants may compete for active sites, attenuate light penetration, or consume reactive species, potentially reducing photocatalytic performance. This lack of real-effluent validation limits assessment of the practical applicability of RE-doped metal oxide photocatalysts.

Catalyst stability and rare-earth leaching were also insufficiently evaluated across the dataset. Stability or reuse testing was reported for 20 of the 29 studies (69.0%), whereas quantitative leaching measurements were reported in only one study (3.4%). Thus, although repeated-use performance was examined in a substantial proportion of the literature, evidence for rare-earth release during photocatalytic operation remains very limited. More systematic evaluation of activity retention, structural stability, and rare-earth leaching during repeated cycles is needed to assess the long-term reliability of RE-doped metal oxide photocatalysts.

Environmental safety remains insufficiently assessed for RE-doped metal oxide photocatalysts. The reviewed studies did not report systematic ecotoxicity assessments of the photocatalysts or released rare-earth species, and nanoparticle release was not adequately evaluated. Therefore, high photocatalytic removal efficiency should not be interpreted as evidence of environmental safety. Future studies should combine photocatalytic performance with measurements of rare-earth release, nanoparticle stability, and relevant ecotoxicity endpoints, particularly during repeated use.

Representative transition-metal (TM)-doped metal oxides show that the photocatalytic effect of TM incorporation is strongly dependent on the dopant and host system. For example, Cu- and Ni-doped ZnO showed different methyl orange degradation efficiencies and kinetic constants under the same experimental framework, with Ni/ZnO providing the highest reported activity among the tested compositions [99]. Ni-doped ZrO₂ has also been investigated as a concentration-dependent photocatalyst for organic pollutant removal, while Ni-doped WO₃ nanoplates provide another example of TM modification of a metal oxide host [100,101].

These studies indicate that TM incorporation can alter photocatalytic behaviour by altering electronic structure, crystallinity, and surface properties, but the magnitude and direction of the effect remain system-dependent.

Compared with rare-earth doping, TM incorporation introduces 3d-derived electronic states and, for some dopants, variable oxidation states that can influence charge transfer and surface redox reactions. Rare-earth ions instead commonly modify the host through localized 4f states, defect chemistry, and changes in the local crystal environment. However, these mechanisms should not be treated as universally beneficial or mutually exclusive. For example, the Gd-doped ZnO study by Alasmari et al. associated the improved activity with defect-related charge trapping. In contrast, the Cu/Ni-doped ZnO study demonstrated that different TM dopants produced different kinetic responses within the same host system [50,99]. Thus, the available evidence supports a system-dependent comparison rather than a general conclusion that either RE or TM doping provides superior photocatalytic performance.

Accordingly, meaningful comparison between RE- and TM-doped oxides should consider dopant concentration, host lattice, morphology, irradiation conditions, pollutant concentration, catalyst loading, and mechanistic evidence in addition to reported removal efficiency. This approach is more appropriate for identifying when a particular dopant strategy is advantageous and avoids attributing photocatalytic enhancement solely to the dopant class.

TM doping can modify the optical and redox properties of metal oxide photocatalysts and, depending on the dopant and host lattice, can influence charge-carrier behaviour and photocatalytic activity [102]. However, these effects are concentration-dependent, and excessive incorporation may introduce defect states that favour charge recombination or alter the structural properties of the host. Therefore, the benefits of TM incorporation must be evaluated for each specific dopant-host system rather than assumed to be universal.

Long-term stability is an additional consideration for practical application; a study investigated Ni-doped hematite and reported changes in photocatalytic and structural properties associated with Ni incorporation [103]. Such findings emphasize the need to evaluate structural retention and dopant stability during repeated operation rather than relying only on initial photocatalytic activity.

Nanocomposites can enhance photocatalytic performance by creating interfaces between distinct functional phases, thereby promoting charge separation, modifying charge-transfer pathways, and improving surface reactivity [105,106]. Depending on the components and their electronic structures, these interfaces may form Type-II, Z-scheme, or Schottky configurations. However, the effectiveness of these configurations depends on interfacial contact, band alignment, phase composition, and morphology. Therefore, composite formation alone does not guarantee enhanced photocatalytic activity.

Recent studies emphasize that well-designed heterojunctions can improve charge separation, whereas unfavorable interfacial alignment or poor contact between phases may limit the expected benefit [104,107]. Thus, the reported enhancement of nanocomposites should be interpreted in relation to their specific structural and reaction conditions rather than assumed to be universally superior to single oxides or other modification strategies.

Representative metal oxide nanocomposites reported for dye degradation are summarized in Table 2. The reported systems include graphene-based ZnO composites, mixed metal-oxide/rGO structures, magnetic Fe₃O₄@TiO₂-based composites, ZnO-TiO₂/rGO, NiO-SnO₂, and TiO₂-ZrO₂. These examples illustrate the different approaches used to combine complementary phases and modify photocatalytic behaviour under solar, visible, or UV irradiation.

The reported removal values vary substantially across the systems, but these values should not be used to establish a universal performance ranking because the studies were conducted using different dye concentrations, catalyst loadings, irradiation sources, and reaction times. Instead, Table 2 demonstrates that nanocomposite design can confer distinct functional advantages, including interfacial charge transfer, magnetic recovery, or improved activity under solar irradiation. Importantly, the available examples remain largely based on controlled laboratory conditions, limiting direct conclusions about their performance in real textile wastewater.

From the perspective of this review, these results support the role of nanocomposite formation as a complementary modification strategy to rare-earth doping rather than as a universally superior approach. Their practical advantage depends not only on photocatalytic activity but also on interface quality, recoverability, stability, and performance under realistic wastewater conditions.

Despite their potential advantages, nanocomposites face several challenges that may limit practical application. Achieving uniform interfacial contact and controlling aggregation are important for maintaining accessible active sites and effective charge transfer. In addition, catalyst recovery, long-term stability, and leaching require greater attention. The studies summarized in Table 2 were largely conducted under controlled laboratory conditions, and the available data are insufficient to establish how these materials would perform in real textile wastewater.

These limitations are important when comparing nanocomposites with rare-earth-doped oxides. Although interfacial engineering can provide additional pathways for charge transfer, nanocomposites may also introduce greater structural complexity and difficulties in synthesis, recovery, and long-term durability. Therefore, photocatalytic activity should be evaluated together with stability, recoverability, environmental safety, and performance under realistic wastewater conditions.

Future work should focus on controlled interface formation, extended cycling tests, quantitative leaching measurements, and validation using real wastewater. Standardized testing and life-cycle assessment would also help determine whether the additional complexity of nanocomposite systems provides a meaningful practical advantage over simpler modification strategies [114,115,116].

The photocatalytic behaviour of a core-shell system depends on the electronic relationship between the two components and the quality of their interface. Appropriate band alignment can facilitate spatial separation of photogenerated charge carriers and reduce electron-hole recombination. However, the direction and extent of charge transfer are material-specific and should be determined from band positions and experimental evidence rather than assumed from the core-shell structure alone.

Experimental studies support this interface-dependent behaviour. For example, ZnO/TiO₂ core-shell structures showed reduced photoluminescence and increased photocurrent relative to the individual components, indicating improved charge separation. Excessive shell thickness, however, reduced the carrier density, demonstrating that the interface and shell structure must be optimized rather than increased [118]. Similar charge-carrier effects have been reported for TiO₂@SnO₂ core-shell nanoparticles [119].

The systems summarized in Table 3 show substantial variation in reported dye removal. Still, these values should not be used to establish a direct performance ranking because the studies differ in dye concentration, catalyst loading, irradiation source, irradiation time, and other operating conditions. For example, ZnFe₂O₄@ZnO achieved 99% methyl orange removal after 240 min under UV irradiation, whereas ZnO/TiO₂ achieved 98% methylene blue removal after 20 min under visible irradiation [120]. These results illustrate the range of reported performance rather than demonstrating the superiority of one system over another.

Core-shell structures can also provide practical advantages such as magnetic separation and catalyst reuse. NiFe₂O₄/ZnO retained high dye-removal performance during repeated use, while the SnO₂-TiO₂/CoFe₂O₄ system was also evaluated over multiple cycles [121,122]. Nevertheless, long-term structural stability, metal release, and performance in real textile wastewater remain less consistently evaluated than laboratory-scale dye removal. Thus, the core-shell architecture represents a useful interfacial design strategy, but its practical value should be assessed together with stability, recoverability, and realistic wastewater performance rather than removal efficiency alone.

The comparison developed in Sections 6-9 does not identify a universally superior modification strategy. Rare-earth doping, transition-metal doping, nanocomposite formation, and core-shell architectures address different limitations of metal-oxide photocatalysts, including electronic-structure modification, defect engineering, interfacial charge transfer, and catalyst recovery [89,98,105]. However, the extent of these effects depends on the host material, modification level, morphology, interface quality, and reaction conditions.

Cross-study comparison is further limited by differences in catalyst loading, pollutant concentration, irradiation conditions, and reaction time [127]. These variables can substantially affect the measured photocatalytic response and should therefore be considered when interpreting the comparative data presented in Table 1, Table 2, and Table 3. Accordingly, the tables are intended to provide contextual comparisons rather than a ranking of intrinsic photocatalyst activity. Meaningful interpretation requires considering the reported removal efficiency together with catalyst loading, pollutant concentration, pH, irradiation conditions, reaction time, and kinetic information, while recognizing that missing or non-comparable parameters limit the strength of cross-study conclusions.

A recurring limitation across the reviewed literature is the difference between observed photocatalytic performance and experimentally demonstrated mechanism. The revised RE-doped dataset illustrates this limitation: 15 of the 29 studies were classified as providing direct mechanistic evidence, whereas 14 provided only partial evidence. Similar caution is required for transition-metal-doped and heterostructured systems, where improved activity may be attributed to charge separation, defect formation, band-gap modification, or interfacial charge transfer without sufficient supporting measurements.

Mechanistic interpretation should therefore rely on complementary evidence, such as photoluminescence, electrochemical impedance spectroscopy, photocurrent measurements, ESR/EPR, scavenger experiments, or other spectroscopic analyses, rather than on degradation efficiency alone. The same evidence-based approach is needed when assessing practical reliability. Reuse, structural stability, catalyst loss, and metal or dopant release are reported less consistently than initial removal efficiency, limiting the assessment of long-term performance and environmental safety [114].

Most of the reviewed photocatalysts were evaluated using individual model dyes under controlled laboratory conditions. Such systems are useful for establishing photocatalytic behaviour, but they do not reproduce the chemical complexity of textile wastewater. Real effluents contain salts, surfactants, dissolved organic matter, suspended material, and competing pollutants that can affect adsorption, light penetration, and the availability of reactive species [128,129].

In addition, dye removal or decolourization should not be interpreted as equivalent to complete mineralization or to reduced toxicity. Evidence of mineralization requires complementary measurements such as TOC or COD, while toxicity assessment requires appropriate analysis of degradation products or biological effects. This distinction is particularly important for azo dyes, for which cleavage of the chromophore may produce intermediate compounds that remain environmentally relevant [130].

For the RE-doped dataset analyzed here, the limitation is quantitative rather than merely qualitative: none of the 29 studies (0%) evaluated the catalyst using real wastewater, whereas all studies relied on controlled model-dye systems. Consequently, a direct numerical comparison of removal performance between model solutions and actual industrial effluents cannot be established for RE-doped photocatalysts from the available evidence. The absence of real-wastewater testing also prevents assessment of how salinity, turbidity, dissolved organic matter, surfactants, suspended solids, and competing pollutants alter catalyst activity under application-relevant conditions. Future studies should therefore report model-solution and real-wastewater results under matched catalyst loading, irradiation conditions, reaction time, and analytical methods so that matrix effects can be quantified rather than inferred [129].

Overall, the evidence reviewed here indicates that the main value of material modification is not simply to obtain a higher dye-removal percentage, but to address specific limitations of the host photocatalyst while maintaining reproducible performance under realistic conditions. Rare-earth doping, transition-metal doping, nanocomposite formation, and core-shell architectures should therefore be viewed as complementary strategies whose suitability depends on material characteristics, pollutant chemistry, operating conditions, mechanistic confidence, stability, and wastewater complexity. This perspective provides the basis for the practical and research priorities discussed in the following sections.

Recent 2025 literature provides additional support for the broader conclusions of this review rather than fundamentally changing them. Recent work on ZnO-based photocatalysis and wastewater treatment continues to demonstrate the potential of modified metal oxides for pollutant removal, while recent discussion of rare-earth-doped ZnO further supports the role of rare-earth incorporation in modifying material properties and photocatalytic behaviour [24,72,131]. However, these recent studies do not address the limitations identified in the present analysis, particularly the dependence of reported performance on experimental conditions and the limited validation of mineralization, long-term stability, leaching, and real-wastewater performance. Thus, the newer evidence strengthens rather than reverses the central conclusion that RE doping is a promising modification strategy, but its practical superiority cannot be established from reported dye-removal efficiency alone.

Most photocatalytic dye-degradation studies use single-dye solutions under controlled experimental conditions. Industrial textile wastewater is substantially more complex and may contain mixtures of dyes, surfactants, salts, and other dissolved or suspended constituents. Such components can alter adsorption, light penetration, and reactive-species availability, thereby altering photocatalytic performance relative to that observed in model solutions [128,130].

The evidence compiled in this review demonstrates a major limitation in this area. None of the 29 RE-doped studies included in the present dataset reported photocatalytic testing using real wastewater. Consequently, the reported efficiencies cannot establish whether the apparent benefits of RE incorporation are retained in actual textile effluents. This evidence gap prevents reliable assessment of whether the reported benefits of RE incorporation would persist under the complex conditions of actual textile effluents. Real-wastewater validation should therefore be considered an essential criterion for establishing the practical significance of RE-doped photocatalysts. As an example of such application-oriented validation outside the RE-doped dataset, a recent study of Au-modified TiO₂ and ZnO evaluated photocatalytic degradation not only with model pollutants but also using untreated real textile wastewater [139].

The absence of real-wastewater validation is therefore a major limitation of the current RE-doped photocatalysis literature. While model dye solutions allow controlled assessment of catalyst activity, they do not reproduce the combined effects of salts, dissolved organic matter, surfactants, suspended solids, turbidity, and competing pollutants present in industrial effluents. Consequently, the reported laboratory efficiencies should be regarded as proof of photocatalytic potential rather than direct evidence of treatment performance under realistic wastewater conditions.

Industrial implementation also requires comparison with existing treatment processes, including adsorption, biological treatment, membrane separation, and other advanced oxidation processes. Such comparisons should consider energy demand, catalyst recovery, maintenance, treatment capacity, and overall process requirements rather than removal efficiency alone. Hybrid treatment configurations may provide a more realistic implementation strategy by assigning different treatment functions to complementary processes [129]. Thus, industrial applications may rely on process integration rather than complete replacement of conventional treatment. The major barriers to translating laboratory-scale photocatalytic dye treatment to industrial or field-scale applications are summarized in Figure 5.

The transition from laboratory reactors to industrial treatment introduces changes in irradiation, hydrodynamics, mass transfer, and catalyst utilization. Laboratory batch systems therefore cannot be directly extrapolated to continuous or larger-scale operation. Reactor design must balance catalyst concentration, hydraulic residence time, mixing, and photon utilization while limiting catalyst aggregation and light attenuation [140].

Reactor configuration is consequently a critical link between catalyst properties and treatment performance. Solar TiO₂ reactor systems have been specifically evaluated for reactor selection, scale-up, and commercialization [141], while photocatalytic membrane reactors provide another configuration for continuous water treatment. These studies demonstrate that reactor engineering must be considered alongside catalyst development when evaluating scale-up potential. However, reactor-level evidence remains insufficient to establish a standardized configuration for RE-doped photocatalysts.

Pilot-scale studies provide a more relevant assessment of practical performance than conventional laboratory batch experiments. Immobilized TiO₂ has been evaluated in a pilot-scale wastewater-treatment reactor, demonstrating the feasibility of operating photocatalytic treatment beyond laboratory scale [142,143]. Nevertheless, pilot demonstration alone does not establish commercial viability. Long-term activity retention, catalyst loss, energy consumption, treatment capacity, and treated-water quality must also be demonstrated under sustained operation.

Economic feasibility depends on the complete treatment process, including irradiation, catalyst preparation and replacement, recovery, reactor infrastructure, maintenance, and treatment capacity. At present, the literature does not provide sufficiently standardized economic data to establish a reliable cost of treatment for RE-doped photocatalysts. A meaningful techno-economic assessment should therefore consider energy consumption, catalyst lifetime and replacement, recovery requirements, reactor and separation costs, treatment capacity, and operating and maintenance requirements using comparable treatment-based metrics. These factors should be evaluated alongside photocatalytic performance before claims of economic or commercial superiority are made [144].

Overall, the current evidence supports the technical potential of metal oxide photocatalysis but does not yet establish a clear industrial advantage for RE-doped systems. The principal gap is the lack of validation under real wastewater conditions, combined with incomplete evidence on reactor-scale performance and standardized economic assessment. Progress toward industrial application therefore requires a shift from reporting high removal efficiencies in model solutions toward demonstrating sustained performance under realistic wastewater and process conditions.

A major priority is to strengthen the experimental basis for mechanistic interpretation. Many studies attribute enhanced photocatalytic activity to improved charge separation, oxygen-vacancy formation, or altered band structure, but provide insufficient spectroscopic, electrochemical, or kinetic evidence. Recent work has further shown that defect formation can either promote or hinder photocatalytic activity depending on defect type and concentration [145]. Future studies should therefore combine complementary techniques, including photoluminescence decay, transient absorption spectroscopy, electrochemical impedance spectroscopy, and carrier-lifetime measurements, to distinguish experimentally demonstrated charge-transfer processes from mechanisms inferred from removal efficiency alone.

Mechanistic validation should also move beyond post-reaction characterization. Operando or in situ spectroscopic and electrochemical measurements under irradiation and reaction conditions could provide more direct evidence of transient charge carriers, reactive intermediates, surface states, and interfacial charge-transfer pathways. Such measurements would reduce reliance on ex situ characterization and improve the reliability of mechanistic assignments.

A second priority is standardized experimental reporting. Photocatalytic performance is strongly affected by catalyst loading, initial dye concentration, pH, reaction volume, irradiation wavelength and intensity, reaction time, and reactor geometry, yet these parameters are not consistently reported [127]. This limits meaningful cross-study comparison and prevents reliable quantitative synthesis. A minimum reporting framework should therefore include catalyst composition and loading, dye identity and concentration, pH, reaction volume, irradiation wavelength and measured irradiance where available, reaction time, kinetic constant, mineralization indicators such as TOC or COD, apparent quantum yield where measured, and catalyst recovery, reuse, and leaching data. Reporting these parameters would provide a more consistent basis for evaluating photocatalyst performance.

Real wastewater validation should be part of this framework. Textile effluents contain multiple dissolved and suspended constituents that can alter catalyst-pollutant interactions and reactive-species formation [146,147]. Recent literature further emphasizes that catalyst behaviour can change substantially in the presence of competing pollutants and complex matrices [148]. For RE-doped photocatalysts, the present dataset contains no study using real wastewater; therefore, future investigations should prioritize actual textile effluents and report both pollutant removal and treatment-relevant outcomes such as mineralization, toxicity, stability, and energy consumption.

High-throughput experimentation and computational methods can complement these experimental approaches. AI- and machine-learning-assisted catalyst discovery could integrate composition, dopant concentration, structural descriptors, and reaction conditions to identify promising catalyst-wastewater combinations before experimental validation. Such approaches should be used to guide experimental design rather than replace mechanistic and experimental verification.

Long-term stability remains insufficiently evaluated for many photocatalytic materials. ZnO photocorrosion, dopant leaching, and changes in catalyst structure during repeated use have been reported [149,150,151], but extended operation and quantitative release measurements remain uncommon. Future studies should therefore combine repeated-use experiments with post-reaction structural characterization and quantitative analysis of released metal or rare-earth species. Stability should be assessed under conditions that approximate the intended wastewater application rather than only under short laboratory cycles.

Environmental safety must also be evaluated alongside photocatalytic performance. Nanoparticles can generate reactive oxygen species and may interact with aquatic organisms after release [152]. Consequently, high dye-removal efficiency should not be interpreted as evidence of environmental safety. For RE-doped photocatalysts, the potential ecological effects of released rare-earth species also remain poorly characterized, particularly under repeated-use and realistic wastewater conditions. Future assessments should include ecotoxicity testing, nanoparticle release, dopant leaching, and, where appropriate, life-cycle assessment to evaluate environmental burdens associated with material synthesis, use, recovery, and disposal.

The optimum material structure is likewise unlikely to be universal across photocatalyst classes. Photocatalytic behaviour can depend on dopant concentration, interfacial architecture, morphology, and defect distribution, but the dominant factor varies among material systems [153,154]. In particular, the relationship between dopant concentration and activity should be treated as system-specific rather than as a universal concentration window. High-throughput screening, computational modelling, and controlled defect engineering can help identify suitable structural parameters for specific catalyst-pollutant systems, but predicted relationships should be confirmed experimentally.

Future research should shift from laboratory batch optimization toward controlled scale-up studies. Reactor hydrodynamics, light distribution, catalyst retention, mass transfer, and hydraulic residence time can substantially influence photocatalytic performance at larger scale [155]. Future pilot studies should therefore report these parameters, along with energy consumption, treated-water throughput, catalyst loading, and long-term activity so that laboratory and pilot-scale results can be compared on a process basis.

Process integration should also be investigated where photocatalysis alone is insufficient for complex wastewater treatment. Combining photocatalysis with adsorption, membrane separation, or biological treatment may allow individual processes to address complementary treatment requirements [156]. Such systems should be evaluated using overall treatment performance, energy demand, catalyst recovery, and process complexity rather than removal efficiency of the photocatalytic step alone.

Economic and environmental assessment should accompany scale-up. Techno-economic analysis should consider catalyst preparation and replacement, irradiation requirements, reactor operation, recovery, treatment capacity, and energy consumption. Life-cycle assessment should also evaluate material production, operational energy use, catalyst recovery, and end-of-life management. These assessments are necessary to determine whether improvements in photocatalytic activity translate into meaningful process-level advantages.

Overall, future development should follow an evidence-based pathway from mechanistic validation and standardized reporting to real-wastewater testing, long-term stability assessment, AI-assisted and experimentally verified material design, pilot-scale operation, and techno-economic and life-cycle evaluation. This progression would help shift photocatalytic research from isolated laboratory performance to reproducible, application-relevant wastewater treatment.