Photocatalytic Performance of Modified TiO₂: A Comparative Analysis of Doping and Co-Doping Process on Methylene Blue Discoloration

Abstract

Heterogeneous photocatalysis is one of the most versatile and widely studied photochemical approaches for the degradation of recalcitrant pollutants. Owing to its favorable physicochemical properties, titanium dioxide (TiO₂) remains one of the most investigated semiconductor photocatalysts. However, its wide band-gap energy (3.2 eV) restricts its photoactivity to the UV region, which represents only a small fraction of the solar spectrum. A major challenge in this field is therefore the development of TiO₂-based materials capable of operating efficiently under visible light irradiation, enabling the use of solar energy as a sustainable primary source. Several strategies have been explored to extend the optical response of TiO₂, among which elemental doping remains one of the most effective and commonly applied. In this work, we conducted systematic comparative analysis to evaluate the photocatalytic performance of TiO₂ modified through different doping approaches. Sixty-one scientific reports published between 2015 and 2025 were analyzed, comparing three categories of dopants: (i) metal dopants, (ii) non-metal dopants, and (iii) co-doping systems. In the first section, we discuss fundamental concepts of photocatalysis and recent advances in doping strategies and surface modifications aimed at enhancing the photocatalytic performance of TiO₂. In the second section, we present a comparative analysis based on 61 scientific reports focusing on TiO₂ doping and co-doping processes. Finally, this study summarizes the different categories of doped TiO₂ photocatalysts by comparing the photocatalytic performance employing an alternative performance metric.

1. Introduction

One of the major challenges of this century is ensuring adequate water purification. Across the globe, nearly forty percent of the population faces water scarcity [1]. Various factors including climate change, population growth, and wastewater generated by industrial processes have increasingly deteriorated water quality [2]. Currently, conventional water treatment methods include (i) natural self-purification processes such as cascades and fast-flowing currents; (ii) chemical oxidation using air, ozone, hydrogen peroxide, or potassium permanganate; and (iii) aerobic and anaerobic biological treatments. All these approaches aim to oxidize contaminant species to higher oxidation states and ultimately achieve pollutant mineralization, producing mixtures of ${CO}_2$, ${NO}_3^{-}$ and/or ${SO}_4^{2-}$ depending on the compounds present in the water [3,4,5].

Despite their versatility, conventional methods are often ineffective in removing emerging pollutants such as pharmaceutical and petrochemical ones, and dyes. These compounds, known as recalcitrant compounds, possess high physical and chemical stability that makes them resistant to standard treatment processes [6,7]. Advanced oxidation processes (AOPs) are physicochemical technologies capable of generating various reactive oxygen species, including singlet oxygen, superoxide anion radicals, and hydroxyl radicals, transient species with strong oxidizing power that can degrade most contaminant molecules [8,9].

Among AOPs, heterogeneous photocatalysis has become an important field of research. Pioneering studies by (i) Doerfler and Hauffe on the photocatalytic oxidation of CO over ZnO [10,11] and (ii) Fujishima and Honda on the photochemical hydrolysis of water [12] laid the foundation for this field. AOPs have since contributed substantially to several disciplines such as photochemistry, electrochemistry, radiochemistry, materials chemistry, surface science, electronics, and catalysis.

The need for renewable energy technologies and the development of photocatalytic systems have positioned heterogeneous photocatalysis as one of the most widely investigated AOPs worldwide and a highly promising strategy for treating recalcitrant pollutants in water and air. Two main factors justify this prominence: (i) the difficulty of purifying these pollutants using conventional methods, and (ii) the possibility of employing solar radiation as the primary energy source to activate the photocatalytic process [13,14].

1.1. Heterogeneous Photocatalysis

The photocatalytic process is a physicochemical phenomenon that generally involves four stages: (i) adsorption of the contaminant onto the surface of a semiconductor (commonly TiO₂); (ii) absorption of electromagnetic radiation by the semiconductor, resulting in the generation of charge carriers; (iii) formation of reactive oxygen species (ROS) and their subsequent reaction with contaminants adsorbed on the semiconductor surface; and (iv) desorption of the reaction products [15]. Stages (iii) and (iv) represent the most critical steps in photocatalysis. The reactions that can take place during these stages are described below:

$${TiO}_2+ {h\mathsf{\nu }}_{(\mathrm{E}>\mathrm{Eg})} \to {TiO}_2({hv}_{{BV}^{+}}/e_{{CB}^{-}})$$

(1)

$${TiO}_2({hv}_{{BV}^{+}}/e_{{CB}^{-}})\to {TiO}_2+Q;$$

(2)

$${TiO}_2({hv}_{{BV}^{+}}/e_{{CB}^{-}})\to {TiO}_2+h\nu ;$$

(3)

$$(e_{{CB}^{-}})+O_{2\left(ad\right)}\to O_{2\left(ad\right)}^{•-}$$

(4)

$$O_{2\left(ad\right)}^{•-}+H_{2}O\to {OH}_{\left(ad\right)}^{•}+{OH}_{\left(ad\right)}^{-}+{\displaystyle \frac{1}{2}}{O}_{2\left(ad\right)}$$

(5)

$$({hv}_{{BV}^{+}})+H_2{O}_{\left(ad\right)}\to {OH}_{\left(ad\right)}^{•}+H^{+}$$

(6)

$$({hv}_{{BV}^{+}})+{OH}_{\left(ad\right)}^{-}\to {OH}_{\left(ad\right)}^{•}$$

(7)

$$({hv}_{{BV}^{+}})+P_{\left(ad\right)}\to P_{\left(ad\right)}^{•+}$$

(8)

$${OH}_{\left(ad\right)}^{•}+P_{\left(ad\right)}\to {(P-{OH}^{•})}_{\left(ad\right)}\rightleftharpoons P_{\left(ad\right)}^{•+}+{OH}_{\left(ad\right)}^{-}$$

(9)

Here, VB and CB denote its valence and conduction bands of the semiconductor, respectively. P corresponds to the pollutant. When the semiconductor is exposed to electromagnetic radiation with energy higher than its band gap (E > E_g), it absorbs this energy, promoting an electron from the VB to the CB. This transition generates two charge carriers: (i) a hole (h⁺), the positive carrier remaining in the VB, and (ii) an electron (e⁻), the negative carrier located in the CB (Equation (1)).

The e⁻/h⁺ pair may recombine either within the bulk of the semiconductor before reaching the surface or directly on the surface itself. The recombination process occurs rapidly (on the order of nanoseconds), and the associated energy may be dissipated as heat (Equation (2)) or re-emitted as photons (Equation (3)). These charge carriers are responsible for the redox reactions involving species adsorbed on the semiconductor surface.

Charge separation of the e⁻/h pair can follow different pathways: (a) adsorbed O₂ molecules may capture electrons from the CB to form $O_2^{•-}$, which can subsequently react with water to produce hydroxyl radicals (Equations (4) and (5)), and (b) the hydroxyl radicals can also be formed through the interaction of charge carriers with $H_{2}O$ molecules adsorbed on the semiconductor surface (Equations (6) and (7)) [16,17].

Pollutant photodegradation may proceed through two parallel steps: (i) direct interaction with hydroxyl radicals (Equation (8)) and/or (ii) direct interaction with photogenerated h⁺ by the photocatalyst. Several authors have highlighted the central role of h⁺ in heterogeneous photocatalysis, as their redox potential is thermodynamically suitable for oxidizing most organic compounds. Depending on the semiconductor, this potential typically ranges from 1.0 to 3.5 V [18,19].

1.2. TiO₂ and the Doping Process

Titanium dioxide (TiO₂) is currently one of the most widely investigated materials for photocatalytic applications in environmental remediation. It is abundant, low-cost, environmentally benign, and due to its favorable band positions, allows both water splitting and the generation of highly oxidizing charge carriers [20,21]. TiO₂ is also chemically and thermally stable, exhibits strong resistance to photo-corrosion, and is safe for the environment. Its principal characteristic is its ability to display photocatalytic activity under UV irradiation. These attributes have enabled TiO₂ to be incorporated into numerous applications, including: (a) optoelectronic devices, (b) solar cells, (c) hydrogen production, (d) chemical synthesis, and (e) heterogeneous photocatalysis [22]. Consequently, it remains one of the most extensively studied materials in photocatalysis research [23,24]. However, despite its physicochemical properties, TiO₂ has a band gap of ~3.2 eV, which restricts its photocatalytic activity primarily to the UV region. This characteristic is particularly relevant given the current research focus on developing photocatalytic materials that are active under visible light irradiation. Several strategies have been explored to extend the absorption of TiO₂ into the visible region of the electromagnetic spectrum, including: (i) synthetic [25,26] and natural [27,28] dye sensitization; (ii) quantum-dot coupling [29]; (iii) metal [30,31] and non-metal doping [32,33]; (iv) composite formation [34]; (v) heterostructure engineering [35,36]; and (vi) surface-plasmon-resonance-based enhancement [37]. Among these strategies, doping remains one of the most extensively studied approaches due to its ability to tune the band-gap energy of the semiconductor. In the doping process, the physicochemical characteristics of an intrinsic semiconductor can be modified by introducing small quantities of foreign elements into the atomic lattice. Doping process affects the photocatalytic activity of TiO₂ through several key effects: (i) generation of intra-gap energy states; (ii) improved charge-carrier separation; and (iii) reduced recombination rates [38,39].

Three factors play a fundamental role in the doping behavior of TiO₂: (i) ionic radius of the dopant, since similar atomic sizes favor lattice incorporation. For example, in nitrogen-doped TiO₂, N2p state hybridize with O2p states, resulting in band gap narrowing [40]. (ii) Energy levels of the impurities, as substitution of O by N creates localized energy states above the VB that can be activated under UV-vis irradiation. (iii) Formation of oxygen vacancies, as oxygen-deficient sites generated at grain boundaries have been identified as active centers that enhance photocatalysis under visible irradiation [41,42].

For effective lattice incorporation, dopant atoms should possess ionic radii and valence states comparable to those of the host atoms. Introducing foreign atoms into the semiconductor lattice modifies its atomic arrangement and electronic structure. Dopants can be incorporated via: (i) substitutional doping, where an anion or cation of the original semiconductor is replaced by another element of greater or lesser valence or (ii) interstitial doping, where the addition of the new element is given by insertion into the empty spaces of the lattice.

For the case of species much smaller than TiO₂ atoms, these will occupy the position interstitial inside the lattice; this phenomenon can generate new energy levels in the band gap, and they can act as recombination centers. Recent studies indicate that metals with greater valence tend to replace some metal ions of the semiconductor, but oxygen substitution occurs only in the insertion of non-metals. The addition of dopants changes the concentration of charge carriers found in both the valence band and the conduction band. In this process, holes or free electrons can be created, forming the p-type and n-type semiconductors [43,44]. TiO₂ acts as an n-type semiconductor due to oxygen vacancies (V_o). These defects (V_o) readily form under reducing atmospheres or vacuum synthesis conditions. Each oxygen vacancy leaves two free electrons, which can occupy levels close to the conduction band. These electrons increase the electron density, moving the Fermi level towards the conduction band, a characteristic feature of n-type semiconductors [45,46].

Another strategy to improve the photocatalytic performance consists of supporting TiO₂ on a secondary semiconductor. Liyanaarachchi et al. reported a significant increase in methylene blue (MB) photodegradation efficiency when Cu-doped TiO₂ was supported on g-C₃N₄ [47]. The enhanced photocatalytic activity was attributed to improved separation of photogenerated electron–hole pairs.

Although single atom doping of TiO₂ has been widely investigated over the past decades, recent studies demonstrate that co-doping produces superior improvements in photocatalytic performance. Co-doping with (i) metal and non-metal elements [48,49], or (ii) two different metal elements [50,51] has attracted increasing interest. A synergistic effect has been reported for the co-doping of semiconductors, including bandgap narrowing, enhanced charge-carrier separation, and reduced recombination losses, which significantly improve the photocatalytic efficiency of TiO₂ [52].

1.3. Kinetic and Performance Metric in Photocatalytic Processes

The charge carriers generated during the photocatalytic process tend to recombine either within the semiconductor or at its surface, due to the presence of recombination centers such as impurities, surface states, and incomplete bonds. This recombination phenomenon leads to energy losses. The kinetic processes associated with the generation and recombination of charge carriers and reactive species are commonly studied using time-resolved absorption spectroscopy (TAS). Through this technique, it has been established that carrier generation (Equation (1)) occurs on a femtosecond timescale (10⁻¹⁵ s). Electron trapping takes place in approximately 50 ps (10⁻¹² s), surface recombination occurs within 1–10 ps, and bulk recombination can exceed 20 ns (Equations (2) and (3)). Additionally, the formation of the superoxide anion radical $(O_{2\left(ad\right)}^{•-})$ typically occurs within 10–100 μs (Equation (4)), while surface-mediated generation of hydroxyl radicals (Equations (6)–(8)) has been reported in the range of 100–400 μs [53].

From a macroscopic point of view, the Langmuir–Hinshelwood (L–H) model is the most widely used kinetic framework for describing photocatalytic reactions [54]. This model incorporates adsorption of reactants on semiconductor surface, and the kinetics of the associated surface reactions, and is expressed according to [55]:

$$v=-{\displaystyle \frac{d\left[C\right]}{dt}}={\displaystyle \frac{kK\left[C\right]}{1+K\left[C\right]}}$$

(10)

where v is the photocatalytic reaction rate, k is the rate constant for the reaction between adsorbed species and photoinduced charge carriers, K is the adsorption equilibrium constant (Langmuir isotherm), and C is the initial concentration of the contaminant prior to illumination [56]. Equation (10) can be solved explicitly as a function of time (t) by considering the decrease in contaminant concentration from its initial value to zero. However, when the contaminant concentration lies in the millimolar range (~10⁻³ mol) or lower, the system can be simplified using a pseudo-first-order approximation:

$$v=-{\displaystyle \frac{d\left[C\right]}{dt}}=kK\left[C\right]=k_{ap}\left[C\right]$$

(11)

Solving this expression yields:

$${\left[C\right]}_t={\left[C\right]}_o{e}^{-k_{app}t}$$

(12)

where t is the reaction time and k_app is the apparent rate constant for the pseudo-first-order model. This model is widely employed as a reference framework to evaluate how different experimental parameters influence photocatalytic degradation of different contaminants. Nonetheless, because Equation (12) can result from multiple mechanistic pathways, it does not describe the detailed mechanism of the reaction. Furthermore, k_app reflects surface-driven steps and is influenced by several factors, including catalyst porosity, light absorption by reactants, reactor geometry, the methylene blue (MB) to TiO₂ ratio, and pH [57].

Additionally, methylene blue is commonly used as a standard pollutant for evaluating photocatalytic activity within the L–H model. Although MB is used as a dye in various applications (e.g., paper, textiles, wool, biomedical staining), it poses risks to both human health and the environment [58]. Despite these concerns, MB remains the most widely used reference compound in photocatalytic tests. The ISO 10678:2024 standard provides a recognized and standardized procedure for assessing photocatalytic activity in aqueous media [59]. To compare the photocatalytic performance of different kinds of materials, appropriate performance metrics are required. In this regard, the photocatalysis literature still shows some weakness because results are not reported in a standardized manner [60]. Despite the growing importance of the photocatalysis field, performance metrics for comparing catalytic efficiency, especially across different disciplines, are not well established.

Table 1. Summary of performance metrics reported in photocatalytic studies.

Performance Metric	Equation	Reference
Reaction rate (r)	$$r={\displaystyle \raisebox{1ex}{$-d{n}_{reagent}$}\!\left/ \!\raisebox{-1ex}{$dt$}\right.}$$	[61]
Normalized reaction rate: reaction rate/catalyst load	$$r=\left({\displaystyle \raisebox{1ex}{$-d{n}_{reagent}$}\!\left/ \!\raisebox{-1ex}{$dt$}\right.}\right)\left({\displaystyle \raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$m_{catalyst}$}\right.}\right)$$	[62]
Turn over Frequency (TOF)	$$TOF=\left({\displaystyle \raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$N_{active site}$}\right.}\right)\left({\displaystyle \raisebox{1ex}{$d{N}_{product}$}\!\left/ \!\raisebox{-1ex}{$dt$}\right.}\right)$$	[63]
Space Time Yield (STY) *	$$STY=\left({\displaystyle \raisebox{1ex}{$N_{product formed}$}\!\left/ \!\raisebox{-1ex}{$V_R.t$}\right.}\right)$$	[64]
Quantu Yield (φ)	$$\phi =\left({\displaystyle \raisebox{1ex}{$N_{product formed}$}\!\left/ \!\raisebox{-1ex}{$N_{absorbed photons}$}\right.}\right)$$	[65]

* V_R: Photo-reactor volume.

lists some of the performance metrics commonly used to compare different materials in photocatalytic tests.

The reaction rate-to-catalyst ratio can be used as performance metric to compare intrinsic photocatalytic activity, since k_app is normalized by the catalyst loading [62]. The TOF measures the number of reactions occurring per active site per unit time under given conditions [66]. The STY quantifies how much material is transformed (or degraded) per unit reactor volume and per unit time [67]. The quantum yield of the photocatalytic reaction is defined as the ratio between the number of molecules formed and the number of photons absorbed by the photocatalyst [68]. All of these metrics are affected by several factors, including reaction time, temperature, wavelength and radiation intensity, contaminant concentration, catalyst loading, and pH.

1.4. Experimental Factors That Affect the Rate of the Photocatalytic Process

Photocatalyst loading is a key factor that determines the photodegradation rate, a phenomenon similar to that observed in catalytic processes in general, where the reaction rate is proportional to the number of active sites available on the photocatalyst surface. However, the rate reaches a maximum value once photon absorption becomes saturated due to an excess of photocatalyst. Therefore, for each experiment, the optimal photocatalyst amount that ensures maximum photon absorption should be determined [69].

pH is an important factor that can affect the photocatalytic rates. The pH value of the solution influences the charge of the semiconductor surface through the rearrangement of surface charges until equilibrium is reached. Furthermore, pH can also modify the chemical structure of the pollutant. The isoelectric point of TiO₂ is reported to be around pH = 6.5 [70]. pH variations can also affect the chemical species present in the medium, promoting chelating, coagulation, and precipitation process [71]. Therefore, the photodegradation of cationic, anionic, or non-ionic pollutants may be favored depending on the pH value.

Because the photocatalytic process is initiated by radiation, this parameter is one of the most critical. Despite its importance, authors often do not report the radiant flux or irradiance of the lamp used as electromagnetic radiation source in photocatalytic tests. Some studies have shown that at low irradiance values (<20 mW/cm²), the reaction rate shows a first order (linear) dependence on irradiance. This trend changes at higher irradiance values, where the rate becomes proportional to the square root of the light intensity. At very high irradiance, the reaction rate becomes independent of light intensity. This change in behavior is associated with the predominance of the electron–hole recombination reactions [72,73]. In addition to irradiance, wavelength is another important parameter, since the number and energy of photons emitted by the light source vary with wavelength [74].

The initial concentration of the pollutant is another important factor that determines the rate of photocatalytic process. This effect is commonly described by the Langmuir–Hinshelwood mechanism discussed above.

2. Materials and Methods

Systematic Comparative Study

In this study, the Preferred Reporting Items for Systematic Reviews and meta-Analyses (PRISMA 2020) methodology was applied [75] to analyze reports concerning dopant elements: (i) metals, (ii) non-metals, and (iii) co-doping, within a period of ten years (2015–2025). The primary databases consulted were Scopus and PubMed, used respectively as the main and secondary source of information [76,77], (see the PRISMA checklist in Supporting Information). The manuscripts included in the comparative study were selected based on the search strategy described in Figure 1.

For the reports that did not provide k_app values, we calculated these constants directly from the experimental data presented in the corresponding manuscripts. This ensured that all studies included in this comparative analysis were evaluated using a consistent kinetic parameter, allowing reliable comparison across the different doping and co-doping strategies.

Quantitative systematic reviews require an assessment of the risk of bias in each included study, as these biases may influence the overall conclusions. This evaluation constitutes the “risk of bias” or “critical appraisal” stage of the review process [78]. Risk-of-bias assessment in systematic reviews must consider the diversity of study designs and requires specific methodological expertise to identify potential sources of bias. However, no comprehensive inventory of applicable designs is currently available, and most existing risk-of-bias assessment tools were developed within the health sciences. Consequently, their applicability to photocatalysis research remains uncertain. Furthermore, terminology such as “quality,” “validity,” and “risk of bias” is often used inconsistently, and several instruments do not clearly distinguish between risk of bias and other dimensions of methodological quality. Previous systematic reviews also show considerable variability in how risk of bias is assessed, reported, and incorporated into the synthesis of results, which complicates the establishment of consistent methodological practices [79].

Given the lack of risk-of-bias tools specifically designed for laboratory-based photocatalysis studies, we developed a structured assessment framework inspired by the domain-based logic of RoB 2, explicitly adapted to the epistemological characteristics of deterministic experimental research. Clinical concepts such as randomization and intervention deviations were redefined in terms of experimental design control, consistency of reported conditions, and completeness of analytical reporting. This adapted framework was not intended to replicate RoB 2, but rather to provide a transparent and structured evaluation of experimental and reporting-related sources of bias, complementary to the Klimisch reliability assessment.

In this study, we applied the adapted risk-of-bias framework for laboratory-based photocatalysis studies across the following domains: (a) D1—bias arising from experimental design and parameter control; (b) D2—bias due to deviations from reported experimental conditions; (c) D3—bias due to missing outcome data; (d) D4—bias in measurement of the outcome; (e) D5—bias the selection of the reported result; and (f) D6—overall risk of bias. Each domain is rated according to the following categories: “High”, “Some concerns”, “Low” and “No information.” The application of this adapted framework provides a structured approach for evaluating the methodological validity of the included studies, as well as for assessing the reliability, reproducibility, and comparability of the reported scientific results [80,81].

Furthermore, the Klimisch classification was included in this study to assess the reliability and quality of the experimental studies. This tool is commonly used in toxicology and ecotoxicology. However, in this work it was applied using the same criteria but adapted to a photocatalysis synthesis approach. The Klimisch rating assigns a score from 1 to 4, where 1 corresponds to “Reliable without restrictions,” 2 to “Reliable with restrictions,” 3 to “Not reliable,” and 4 to “Not assignable,” depending on the methodological rigor and the quality of reporting. In addition, this classification includes two categories, A and B. Category A results from the evaluation of 18 criteria distributed across five domains (I–V) applied to photocatalysis studies, whereas category B corresponds to the assessment of the six high-risk criteria identified in the risk-of-bias evaluation [82]. The items evaluated are listed in Table S1 (see Supporting Information).

Finally, we employed R software (version 4.5.1) with the packages dplyr, tidyr, robvis, and ggplot2 in the analysis of the evaluated reports. These tools were used to perform the risk-of-bias assessment, and the results are presented as graphs. The main objective of this analysis is to evaluate the robustness of the available evidence and to determine whether the studies produce consistent and comparable results while ensuring analytical transparency. Finally, we used python software (version 3.12.13) and the Seaborn library (version 0.13.) to generate the Box plots.

3. Results and Discussion

3.1. Structured Risk-of-Bias-Inspired Assessment Adapted for Experimental Photocatalysis Studies

Although a substantial proportion of the evaluated studies were judged to have a “Low” risk of bias in key domains—particularly outcome measurement and results reporting—a considerable number of assessments were classified as “Some concerns” or “No information” in critical domains such as D1 and D2. These patterns, observed across different doping categories (metal, non-metal, co-doping, and undoped materials), indicate significant methodological heterogeneity among the studies, which may introduce uncertainty in the assessment of the comparative effects between doped and undoped materials (see Figure 2). Traditionally, photocatalysis studies emphasize the detailed characterization of catalyst physicochemical properties (e.g., structural, optical, morphological, and spectroscopic features). However, a comprehensive description of the reactor configuration, light source characteristics, pH, catalyst loading, systematic controls, and analytical methods is essential to ensure reliable comparison and reproducibility of photocatalytic results across different studies.

3.2. Klimisch Score

The quality assessment applied to the data revealed that the sample was heterogeneous. The Klimisch assessment allows the classification of the data into 18 criteria (see Table S1) set out in five domains (I–V), including essential elements (e.g., dopant type, physicochemical characterization, experimental conditions, control practices, and analytical confidence). Although individual scores showed that only a small number of reports reached high Klimisch categories (e.g., Klimisch 1 (green): reliable without restrictions and Klimisch 2 (orange): reliable with restrictions), most of the data were classified in lower categories (Klimisch 3 (gray) or Klimisch 4 (red)) due to deficiencies in experimental conditions and omissions in environmental parameters (Figure 3 shows the percentage of Klimisch categories by type). One recurring trend was the systematic lack of information on experimental conditions (e.g., reactor description, control experiments, analytical validation, statistical uncertainty, or error bars in the graphs). The absence or incomplete reporting of these parameters led many studies to be classified in the Klimisch 3 category, even though their overall scores would have initially placed them in higher categories. This result is consistent with recurring observations reported in systematic reviews in the field, where the lack of standardization in experimental conditions and incomplete reporting compromise reproducibility and hinder direct comparisons among evaluated materials [83]. Furthermore, studies classified as non-classifiable (Klimisch 4) did not necessarily present methodological deficiencies, but rather insufficient reporting to allow a complete assessment. This distinction is important because it highlights that a lack of information does not necessarily reflect poor experimental quality; however, it prevents objective evaluation and therefore limits the reliable inclusion of these studies in comparative syntheses or comparative analyses. The presence of non-evaluable studies highlights the need for stricter reporting standards in the field of photocatalysis, particularly regarding the presentation of raw data, experimental details, and analytical transparency.

The overall analysis indicates that the systematic application of tools, such as the Klimisch method, facilitates the identification of common strengths and weaknesses in the literature, thereby guiding both the design of future research and the standardization of reporting practices. In particular, the high proportion of studies scored in low categories due to the failure to meet essential criteria suggests that the scientific community should prioritize the explicit inclusion of these elements in their publications, as they are fundamental for reproducibility, comparability among studies, and the development of robust evidence regarding doped and undoped materials. Finally, the publications classified in the Klimisch categories 1 or 2 were characterized by clearly reporting material identification, structural and surface characterization, a complete description of the reactor and light source, the use of systematic controls, and adequate documentation of the analytical methods. This demonstrates that adherence to experimental quality standards produces data that are more comparable, reproducible, and less susceptible to bias. Therefore, it is recommended that future research adopt similar reporting frameworks, including mandatory criteria I1, I3, II1, II3, III2 IV1, and IV2 (see Table S1) to improve the traceability, transparency, and usefulness of results within the field of heterogeneous photocatalysis.

3.3. Data Reporting and Photocatalytic Performance Metrics

In this work, 61 publications from the period 2015–2025 were analyzed to conduct the comparative analysis. The reported k_app values were used to estimate two performance metrics to compare the photocatalytic performance of TiO₂ in MB degradation according to the type of dopant employed. For clarity, the following classification was used: (i) no-dopant, corresponding to reports on bare TiO₂ (

Table 2. Data collected for bare (undoped) TiO₂.

Catalyst	Light	Dopant (%)	Band Gap (eV)	Load Catalyst (mg)	Vsolution (mL)	Radiant Flux (W)	[MB] ppm	Time (min)	Removal (%)	kapp × 10−3 (min−1)	kapp × 10−3/ (VR. mcat¨) (mg−1 min−1 mL−1)	Reference
+P-25	UV	*Nr	Nr	10	15	Nr	10	180	20	1.3	8.7	[84]
P-25	UV	Nr	Nr	100	100	450	11	120	55	4.3	0.43	[85]
++TiO2	UV	Nr	Nr	20	Nr	Nr	10	100	Nr	6.0	30	[86]
TiO2	UV	Nr	Nr	100	100	Nr	10	180	86	7.1	0.71	[87]
TiO2	UV	Nr	3.14	62.5	200	Nr	25	60	Nr	4.8	0.38	[88]
TiO2	UV	Nr	3.2	50	100	3.0	Nr	50	20	3.3	0.66	[89]
TiO2	UV	Nr	Nr	100	Nr	30	9.6	300	55	2.5	NC **	[90]
TiO2	UV	Nr	3.2	100		250	10	50	25	7.1	0.18	[91]
TiO2	UV	Nr	3.04	50	Nr	Nr	Nr	270	43	3.3	NC	[92]
P25	UV	Nr	3.2	5	5	60	1.6	150	Nr	2.65	106	[93]
TiO2	UV	Nr	Nr	100	500	Nr	200	120	88	1.1	0.022	[94]

¨m_cat: catalyst load. ⁺P25:TiO₂-Degussa P25. ⁺⁺TiO₂: Titanium dioxide sensitized by chemical method. *Nr: No reported. ** NC: Not calculated.

); (ii) metal, referring to metal elements used as dopants (

Table 3. Data collected for metal doped TiO₂.

Dopant	Light	Dopant (%)	Band Gap (eV)	Load Catalyst (mg)	Vsolution (mL)	Radiant Flux (W)	[MB] ppm	Time (min)	Removal (%)	kapp × 10−3 (min−1)	kapp × 10−3/ (VR. mcat¨) (mg−1 min−1 mL−1)	Reference
Cu	UV	1.7	2.67	200	100	100	100	250	19	1.1	0.055	[47]
Ni	VIS	5	3	50	100	Nr	30	180	70	6	1.2	[49]
Co	VIS	*Nr	2.77	200	50	Nr	10	60	78	35.7	0.004	[50]
V	VIS	Nr	2.91	200	50	Nr	10	60	87	26	0.003	[50]
Pr	UV	Nr	Nr	100	100	Nr	10	180	93	12.1	1.2	[87]
Nd	UV	Nr	Nr	100	100	Nr	10	180	86	11.2	1.1
Cu	UV	Nr	Nr	100	Nr	30	9.6	300	98	8.6	NC **	[90]
W	UV	2	Nr	100	500	Nr	10	120	45	3.5	0.070	[94]
Ag	UV	6	nr	100	50	Nr	100	120	95	14.1	2.8	[95]
Bi	VIS	10	2.82	100	35	Nr	5	150	64	6	1.7	[96]
Mo	UV	Nr	2.74	100	100	Nr	10	60	98	56	5.6	[97]
Mo	VIS	Nr	2.74	100	100	125	10	120	95	30.1	3.0
Sn	VIS	5	Nr	1000	Nr	125	20	120	77	12.3	NC	[98]
Na	UV	8	3.3	60	100	300	5	60	92.6	43	72	[99]
Pd	UV	0.5	Nr	1000	100	Nr	20	120	99.4	44	0.44	[100]
Ca	UV	1.0	3.17	20	100	100	20	180	79.6	8.7	4.4	[101]
Nb	UV	1.0	2.6	10	Nr	350	5	720	68	1.6	NC	[102]
W	VIS	2.65	3.5	50	50	75	40	120	91	13.2	5.3	[103]
Fe	UV	0.25	2.66	50	100	32	1.5	120	86.7	12.4	25	[104]
Fe	VIS	0.25	2.66	50	100	60	1.5	360	86.7	5.95	1.2
Fe	VIS	Nr	Nr	5	5	60	1.6	150	40	3.5	NC	[93]
Mn	VIS	0.6	2.72	1000	250	500	160	300	80	4.5	0.018
Ce	VIS	0.6	2.74	1000	250	500	160	300	Nr	3.8	0.015	[105]
La	VIS	0.6	2.76	1000	250	500	160	300	Nr	3.5	0.014
Yb	VIS	Nr	2.47	150	50	30	10	300	72	4.4	0.58	[106]
Ag	VIS	1.3	2.89	100	50	Nr	10	120	90	19.2	3.8	[107]
Fe	UV	1	Nr	200	100	300	400	120	33	2.7	0.14	[108]
Eu	UV	3	3.13	Nr	Nr	200	6.5	60	Nr	37.2	NC	[109]
Er	UV	2.9	2.9	Nr	Nr	160	12	240	95	16.7	NC	[110]
Cr	VIS	1	Nr	100	50	150	15	1200	50	0.52	0.10	[111]
Cu	UV	0.1	3.15	100	Nr	Nr	3.2	30	98	9.11	NC	[112]
Zn	VIS	0.01	2.8	100	50	Nr	20	60	Nr	19.3	3.9
Zr	VIS	0.01	3.3	100	50	Nr	20	60	81.9	61.6	12.3	[113]
Cu	VIS	0.01	3.7	100	50	Nr	20	60	Nr	32.1	6.4
Cu	VIS	2	3.07	10	100	105	5	300	92.3	9.8	9.8	[114]
Cu	UV	0.15	Nr	Nr	Nr	4.8	15	360	99	28.9	NC	[115]
Cu	UV	5	Nr	40	Nr	Nr	10	120	92.3	10.1	NC	[116]
Co	VIS	Nr	Nr	10	50	35	20	210	90.1	10.3	20.6	[117]

¨m_cat: catalyst load. * Nr: No reported. ** NC: Not calculated.

); (iii) non-metal, covering non-metal dopants (

Table 4. Data collected for no-metal doped TiO₂.

Dopant	Light	Dopant (%)	Band Gap (eV)	Load Catalyst (mg)	Vsolution (mL)	Radiant Flux (W)	[MB] ppm	Time (min)	Removal (%)	kapp × 10−3 (min−1)	kapp × 10−3/ (VR. mcat¨) (mg−1 min−1 mL−1)	Reference
P	VIS	1.25	2.9	225	75	32	5	180	61	8	0.47	[118]
S	VIS	5	2.93	50	100	Nr	30	180	73	6.4	1.3	[49]
C	UV	*Nr	2.94	62.5	Nr	Nr	25	60	Nr	7.8	NC **	[88]
N	VIS	Nr	Nr	5	5	60	1.6	150	95	22.2	NC	[93]
N	VIS	Nr	1.95	150	50	30	10	300	74	5.0	0.66	[106]
N	VIS	Nr	2.8	100	50	17	10	10	78	8.3	1.7	[119]
P	VIS	2	2.95	200	10	100	5	90	90	29	14.5	[120]
C	UV	Nr	2.4	20	50	400	10	40	Nr	32.8	32.8	[121]
F	VIS	Nr	Nr	800	50	4.8	10	60	80	25	0.63	[122]
F	VIS	Nr	Nr	10	30	150	150	300	98.7	4.1	13.7	[123]
F	VIS	Nr	Nr	2.5	2.5	10	10	50	98	25	NC	[124]
N	VIS	Nr	2.97	30	200	16	20	180	Nr	6.6	1.1	[125]
N	UV	Nr	2.97	30	200	16	20	180	Nr	32.2	5.4	[125]
C	UV	5.6	2.71	10	Nr	40	10	300	56.3	2.3	NC
C	VIS	5.6	2.71	10	Nr	88	10	300	51.2	1.8	NC	[126]

¨m_cat: catalyst load. *Nr: No reported. **NC: No data to calculate.

); and (iv) co-doping, including materials doped with more than one element (

Table 5. Data collected for co-doped TiO₂.

Dopant	Light	Dopant (%)	Band Gap (eV)	Load Catalyst (mg)	Vsolution (mL)	Radiant Flux (W)	[MB] ppm	Time (min)	Removal (%)	kapp × 10−3 (min−1)	kapp × 10−3/ (VR. mcat¨) (mg−1 min−1 mL−1)	Reference
Mn-N	UV	*Nr	2.81	100	Nr	Nr	Nr	180	31	3.3	NC **	[48]
Mn-N	VIS	Nr	2.81	100	Nr	Nr	Nr	180	25	5.5	NC
Ni-S	VIS	5	2.87	50	100	Nr	30	180	84	8.2	1.6	[49]
Co-V	VIS	Nr	2.56	200	50	Nr	10	60	92.1	42.5	0.004	[50]
Ni-Cr	VIS	5	2.45	100	50	Nr	5	90	95.6	36.6	7.3	[51]
Ag-S	VIS	2.5	2.89	100	Nr	Nr	Nr	180	95	18.4	NC	[52]
B-W	VIS	2.85	3	50	50	75	30	120	82.3	10.4	4.2	[103]
Fe-N	VIS	Nr	Nr	5	5	60	1.6	150	47	4.8	NC	[93]
Yb-N	VIS	Nr	2.07	150	50	30	10	300	93.5	9.1	1.2	[106]
Al-S	VIS	10	1.98	40	100	200	30	150	96	17	4.3	[127]
La-I	UV	3	2.68	100	100	64	30	60	98	60	6.0	[128]
La-I	VIS	3	2.68	100	100	64	30	60	80	34	3.4
Bi-Fe	UV	4	3.25	100	500	250	50	120	55	67	1.3	[129]
Mn-S	VIS	Nr	2.78	50	100	500	30	150	94.8	0.3	0.060	[130]
Y-Eu	UV	1.0	2.86	2000	Nr	15	20	120	87	18.4	NC	[131]
N-C	VIS	Nr	2.72	80	80	300	6.5	120	95	22.5	3.5	[132]
S-Fe	UV	Nr	2.84	800	100	40	10	120	88	16.2	0.20	[133]
Sr-Mn	UV	Nr	3.05	12	100	Nr	30	120	90	21.2	NC	[134]
Ag-Y	VIS	4	Nr	50	50	42	30	720	98	5.8	2.3	[135]
Cu-Zr	VIS	1.5	2.74	100	Nr	Nr	40	120	92	20.8	NC	[136]
Sn-F	VIS	4	3.28	Nr	Nr	Nr	Nr	90	94	34.5	NC	[137]
Fe-N	VIS	Nr	2.7	200	Nr	Nr	7	60	Nr	20	NC	[138]

¨m_cat: catalyst load. *Nr: No reported. ** NC: Not calculated.

).

Because all comparative metrics are affected by several experimental factors (e.g., reaction time, temperature, wavelength and radiation intensity, contaminant concentration, catalyst loading, and pH), a single performance metric is not a suitable methodology for comparing photocatalysts under different experimental conditions. Instead, a combination of multiple performance metrics (e.g., reaction rate, normalized reaction rate, photocatalytic space time yield, and apparent quantum yield) provide a more robust approach for assessing a wide range of photocatalytic systems [64]. Accordingly, since the k_app value for MB photodegradation alone is not a fully reliable comparative metric, we employed a combined performance metric according to the following equation:

$$CM={\displaystyle \frac{k_{app}}{m_{cat}{V}_R}}$$

(13)

where CM is the comparative metric, k_app represents the rate constant obtained from the Langmuir–Hinshelwood model, m_cat is the catalyst loading, and V_R is the volume of solution used in the photocatalytic test. This metric provides an alternative method to normalize kinetic results and compare photocatalytic activity. However, it does not include an important parameter: the irradiance of the lamp used as the light source. Some authors report the radiant flux of lamps (watts), whereas others report irradiance (watts cm⁻²). In our dataset, most authors reported radiant flux (see Table 2, Table 3, Table 4 and Table 5) and only 12 reported irradiances. It is common in photocatalysis studies that light intensity is not fully specified, as authors often report only lamp power, which makes cross-study comparisons difficult.

Figure 4 shows that CM values for MB photodegradation using bare TiO₂ are generally below 1.0 and are lower than those of doped materials. The doping process improves the photocatalytic activity compared with bare TiO₂ [139,140]. However, the results in Figure 4 do not reveal clear differences in CM values based on the specific dopant or combination of dopants used. However, several studies suggest that co-doped samples exhibit markedly higher photocatalytic efficiency than those doped with a simple element. VO et al. [141] reported that TiO₂ co-doped with Ce and Ni showed a higher degradation rate for Congo Red compared with TiO₂ doped only with Ce. Similar trends have also been observed in other semiconductors [31]. Outstanding results reported for co-doped TiO₂ materials are commonly attributed to factors such as: (i) a reduction in the semiconductor band-gap energy and (ii) a decrease in charge-carrier recombination rates [142]. Because TiO₂ has a relatively large band gap, it is intrinsically active only under UV irradiation.

In the last decades, researchers have studied the doping process of TiO₂ using non-metals. Among all anion options, carbon is one of the most promising due to its physical chemical properties (e.g., adsorption capacity, conductivity and electrochemical behavior) [33,143]. The TiO₂ doped with non-metal elements showed a maximum value of CM of 32.8 × 10⁻³ min⁻¹ g⁻¹ mL⁻¹. This value was obtained in the photodegradation of MB under UV irradiation onto C-doped TiO₂ Nanorods as a catalyst [121]. For TiO₂ doped with metallic elements, the maximum value of CM was 20.6 × 10⁻³ min⁻¹ g⁻¹ mL⁻¹. In this work, Jaihindh et al. reported 90.1% MB degradation after 210 min of VIS irradiation using co-doped TiO₂ [117].

The photocatalytic activity of TiO₂ can be enhanced by alkali metal doping for increasing the electron transfer efficiency [144]. When transition metals (e.g., Mn, V, Cr, Co) are incorporated into the TiO₂ lattice, they can improve the photocatalytic properties of the semiconductor by reducing the band gap energy value, shifting the adsorption wavelength to the visible region of the electromagnetic spectrum [145,146]. In the case of noble metals, a synergetic effect has been reported. Although the noble-metal doping of TiO₂ reduces both the band gap value and recombination rates, these materials can exhibit the surface plasmon resonance phenomena, which further enhance the photocatalytic performance. However, despite these advantages, their implementation in photocatalytic applications on a large and middle scale is limited by their higher costs [31,147].

One of the current challenges in the field is the development of materials with photocatalytic activity under visible light, which is crucial to improving process sustainability. Only 3–5% of the solar spectrum falls within the UV region, meaning that most of the available solar energy is not used efficiently by bare TiO₂. When UV sources are employed in photocatalytic tests, more energetic charge carriers are generated, enhancing the photodegradation kinetics; however, a charge-carrier recombination may also increase under these conditions [148,149].

Furthermore, the cost of implementing heterogeneous catalysis is higher when UV irradiation is used as the primary energy source. In contrast, when visible light is used as the primary energy source, operational costs are reduced, and the treatment process exhibits a lower carbon footprint [150,151]. Figure 5 compares the CM values calculated employing Equation (13) as a function of the type of light source employed (visible or UV-light) with different types of dopants. There are no reports to bare TiO₂ catalyst under visible irradiation due to its wide band-gap energy (Figure 5d). Furthermore, bare TiO₂ under UV irradiation exhibits the lowest mean photocatalytic degradation values among the materials analyzed. In contrast, studies employing visible irradiation with modified TiO₂ generally report higher CM values. This behavior may be attributed to differences in charge-carrier recombination rates under UV and visible irradiation. Mollavali et al. reported that doping-induced trap states and electronic levels can suppress electron–hole recombination. The enhanced performance of co-doped TiO₂ is likely related to improved charge separation and more efficient transport of charge carriers from the bulk to the surface of the catalyst [152].

Although the data cannot compare directly due to the authors not specifying the irradiance of the lamp employed in the test, some authors have reported interesting results about the effect of the kind of light on the photocatalytic efficacy. Yamai demonstrated that MB photodegradation efficiency increases with UV light intensity when using TiO₂ Degussa P25 [85]. The rate of photo-discoloration increased when the laser pulse energy was increased from 50 mJ to 150 mJ. This increase was associated with the fact that higher intensities activate more surface sites on TiO₂, thereby enhancing its photocatalytic activity. However, in some cases, similar CM values can be obtained using different light sources. For example, Kunnamareddy et al. reported MB photodegradation under visible irradiation using S-doped TiO₂ as a catalyst, employing a lamp with an irradiance of 49 mW cm⁻². After applying Equation (13) to these data, the CM values were 1.28 × 10⁻³ min⁻¹ mg⁻¹ mL⁻¹ (see Table 4) [49]. Mishra et al. reported MB photodegradation under UV irradiation using (Bi/Fe) co-doped TiO₂ as a catalyst. After applying Equation (13), the CM value was 1.34 × 10⁻³ min⁻¹ mg⁻¹ mL⁻¹ (see Table 5) [129]. In this case, the authors did not report the irradiance but only the radiant flux of the lamp used (250 Watts).

This situation exemplifies a common issue: although experimental data are available and the CM value can be calculated, cross-study comparisons remain difficult because different sets of experimental parameters are reported (e.g., light source type and intensity). There is currently no standardized methodology for reporting photocatalytic performance apart from degradation efficiency and the reaction rate constant. Special attention should therefore be given to establishing a minimum set of parameters that should be reported to enable simple and reliable comparisons among studies.

To ensure reproducibility, enable comparability among studies and promote the development of robust scientific evidence, a minimum set of experimental conditions related to the light-source system should be reported in photocatalytic studies: (i) the number of lamps (n_lamp), (ii) the maximum wavelength of the light source (λ_max), and (iii) the irradiance of light source (I_lapm). Based on these parameters, the maximum photon flux (Φ) of the system can be calculated using the following equation [153]:

$$\mathsf{\Phi }={\displaystyle \frac{n_{lamp}{·I}_{lamp}·{\lambda }_{max}}{h·c·N_{av}}}$$

(14)

where h is the Planck constant, (c) is the speed of light, and N_av is the Avogadro number. This parameter allows for the comparison of photon flux among different photocatalytic systems.

Finally, to reliably compare different catalysts, all photocatalytic test conditions should be clearly reported to ensure proper comparison [154].

3.4. Trends and Perspectives on Heterogeneous Photocatalysis

Currently, the most prominent applications of photocatalytic processes include: (i) CO₂ reduction [155,156]; (ii) hydrogen production [157,158]; (iii) air purification [159,160]; (iv) water treatment systems [161,162]; (v) plastic waste degradation [163]; (vi) synthetic methodologies, bio-catalysis, and artificial photosynthesis [164]; (vii) self-cleaning surfaces [165,166]; and (viii) emerging applications in fields such as medical [167] and construction [168]. Among these, applications involving the development of materials for exterior construction have gained particular attention due to their durability and potential for large-scale implementation.

Photocatalytic technologies have demonstrated acceptable efficiencies in various configurations, and ongoing intensive research worldwide continues to drive their advancement. At present, several photocatalytic systems have been implemented at different scales (industrial and pilot) to treat diverse contaminants, including hormones, dyes, hydrocarbons, heavy metals, herbicides, antibiotics and antimicrobials.

Overall, the results of this study underscore the relevance of doped TiO₂ as a competitive and adaptable photocatalyst and highlight the urgent need to standardize reporting practices to advance the field efficiently.

4. Conclusions

Developing renewable processes to transform energy and matter is one of the major challenges that society must address in the medium term. This need is particularly critical for water purification technologies, where sustainable and efficient treatment methods are increasingly required. Among the available alternatives, heterogeneous photocatalysis stands out as a promising technology capable of meeting these demands.

In this work, the photocatalytic performance of doped TiO₂ in the decolorization of methylene blue (MB) was analyzed. The information from 61 scientific reports was examined to compare the photocatalytic efficacy for each study. The main findings of this comparative analysis are summarized below:

• Photocatalytic systems continue to expand as part of environmental technological developments based on renewable energy sources. These systems represent a promising alternative for the treatment of recalcitrant pollutants, especially considering their potential to utilize solar radiation as a primary and sustainable energy source.
• The doping process improves the photocatalytic activity of TiO₂ due to several mechanisms: (i) the formation of intra-gap energy states, (ii) enhanced separation of photogenerated charge carriers, and (iii) reduced charge-carrier recombination rates.
• C and N are the most investigated non-metallic dopants of TiO₂. The enhancement in the photocatalytic performance is attributed to band gap reduction caused by generation of impurities above the valence band of TiO₂.
• Although the noble-metal doping of TiO₂ reduces both the band gap value and recombination rates, these materials can exhibit the surface plasmon resonance phenomena, which further enhance the photocatalytic performance.
• Doping is one of the most effective strategies to improve the photocatalytic response of TiO₂ under visible light irradiation, making it a key approach for developing more sustainable and solar-driven treatment processes.
• Significant variations in photocatalytic efficiency arise from differences in experimental conditions, including catalyst synthesis methods, pH, light intensity, pollutant concentration, and catalyst loading. Currently, there is no standardized methodology for reporting photocatalytic performance beyond degradation efficiency and the reaction rate constant. Special attention should therefore be given to establishing a minimum set of parameters that must be reported to enable simple and reliable comparisons among studies.
• The scientific community should prioritize the explicit reporting of all experimental conditions in photocatalytic tests in their publications, as these are essential for ensuring reproducibility, enabling comparability among studies, and promoting the development of robust scientific evidence.