Use of Nanostructured Photocatalysts for Dye Degradation: A Review

Among the technologies proposed for wastewater treatment, the Advanced Oxidation Processes are viable and technological strategies for dyes degradation. Different photocatalytic systems classified in metal oxides alone or combined through hybrid composites or immobilized onto supports have been designed in various nanostructured shapes for their application in the photodegradation of polluting dyes. This review aims to describe the dyes as an environmental threat, photocatalysis as an effective process to remove dyes from water and provide an overview of the recent studies using photocatalytic systems grouped according to their development. Furthermore, this review describes the main parameters of a photocatalytic system with an important role in dye photodegradation. Finally, we discuss the limitations of photocatalysis for real industrial applications and the challenges for this environmental


Introduction
The use of natural dye for textile dyeing has been practiced for 5,000 years ago. On the other hand, the discovery and application of synthetic dyes begun in the 19 th century by displacing the use of natural dye. Nowadays, the global colorant market is about 32 billion USD and is projected to increase to around 42 billion USD by 2021 [1]. Synthetic dyes present advantages compared to natural dyes because of their lower prices, repeatability, and wide range of bright shades with considerably improved color fastness properties [2]. Dyes are colorful substances designed to give a hue to any colorable materials, and this is possible as dyes can attach themselves to any amenable materials. Moreover, dyes are composed of a group of atoms known as chromophores, responsible for the dye color. Dyes are sorted according to their application and chemical structure and are classified as acid, basic, direct, mordant, and reactive dyes, which are examples of soluble dyes, whereas azo, disperse, sulphur and vat dyes are an example of insoluble dyes [3]. The azo dyes kind, molecules with one or more azo (N=N) bridges linking substituted aromatic structures, represent a 70% of the global production and are the most frequently utilized dyes [4]. Unfortunately, the dye industry dramatically contributes to global pollution, generating consequences to the ground and water due to its toxic, carcinogenic, and xenobiotic repercussions. On the other hand, researchers have proposed using the Advanced Oxidation Processes (AOP's) as an attractive technology for removing a wide range of emerging contaminants. The AOP's involve the in-situ production of highly reactive oxygen species such as hydroxyl radicals (OH • ) and superoxide anion radicals ( O 2 • − ) with oxidation potentials of 2.7 and −2.3 eV, respectively. These species can be initiated by primary oxidants ( e.g., H 2 O 2 , O 3 ), energy sources (e.g., UV light, ultrasonic, and heat), or catalysts ( e.g., TiO 2 , ZnO, and ZrO 2 ) [5]. Among AOP's, photocatalysis is a viable alternative process to remove the emerging contaminants at standard temperature and pressure (STP) conditions by oxidation reactions. As mentioned, a huge variety of nanomaterials with photocatalytic activities have been used in environmental remediation because they propose using solar energy to promote photoreaction, making the process cheaper and environmentally friendly. Moreover, the favorable combination of electronic structure, light absorption properties, charge transport characteristics, improved textural proprieties, excited lifetimes, and versatility in shapes and sizes of metal oxides has made them possible for their application in photocatalysts [6]. Therefore, this review compiles the information of current photocatalytic systems based on mixed oxide nanoparticles used to degrade water dyes. In addition, scientific aspects were discussed, some social concerns and current trends of photocatalysis are also described.

Dyes and their environmental impact
Nowadays, it is estimated that 700,000 tons of various colouring from about 100,000 commercially accessible dyes are manufactured each year [7]. Nevertheless, between 10% and 15% of the synthetic dyes are lost during different textile industries processes [4]. Moreover, about 40,000-50,000 tons of dyes are discharged in water bodies from natural or anthropogenic means [8]. The azo dyes are considered one of the most difficult compounds to be removed and degraded from aqueous systems; thereby, the public demand for color-free discharge has rendered decolorization of wastewater is a priority [9]. During the dyeing processes, not all dyes that are applied to the fabrics are fixed on them, and usually, a portion of these dyes that remains unfixed to the fabrics and gets washed out, this amount of generated textile wastewater can reach more than 300 L kg −1 of product [10]. These unfixed dyes are found to be in high concentrations in textile effluents, and the composition of the wastewater will depend on the different organic-based compounds and the dyes used in the dry and wet-processing steps. Moreover, textile wastewaters can generate fluctuations in parameters such as chemical oxygen demand (COD), total organic carbon (TOC), biochemical oxygen demand (BOD), pH, flavor, colour, and odor when are released in aquifers [11]. The releasing of dye effluents into aquifers is undesirable due to the high impact on photosynthesis of aquatic organisms, and the carcinogenic nature and mutagenicity of many of these dyes and their breakdown products [11]. One of the main concerns is reducing the penetration of light when dyes are dissolved in water, which can cause an alteration of the photosynthetic activity and thus modify the natural balance of flora and fauna. Furthermore, these effluents can pass through soil layers and may contaminate nearby surface and underground water. For human health affectations, the dermal exposure of the dye precursors leads to bladder cancer, since as dyes contain aromatic amines, they can generate damage in the DNA of cells, leading to the risk of cancer disease. Moreover, dyes can promote other human health problems such as allergies, urticaria, angioedema, hyperactivity, ocular irritability, aggressiveness and learning impairment related to intake of dye [12].

Photocatalysis
Photocatalysis was proposed in 1972 by Fujishima and Honda [13]; they discovered that TiO 2 decomposes water into hydrogen and oxygen under light irradiation. In this context, photocatalysis is the acceleration of a reaction using a catalyst in light presence with an adequate wavelength. To carry out a photocatalytic process, the incident light on the catalyst should supply energy equal to or greater than the semiconductor band gap value (eV). This energy can be calculated using Eq. (1): where λ is the wavelength value, for example, if a semiconductor has a band gap of 3.0 eV, the incident wavelength value on photocatalyst should be equal to or under 413.3 nm to photo-excite the electrons from the valence band ( V B ) to the conduction band ( C B ). Therefore, an electron belonging to V B is excited to the C B , giving as a result a pair of species, a hole (h + ) in the V B and an electron (e − ) in the C B [14], Fig. 1 (5) and (6) [15]. In a typical photocatalysis process, two different reactions occur; an oxidation reaction due to photo-induced positive holes and a reduction reaction due to photo-induced negative electrons [16]. Furthermore, the oxidation potentials of hydroxyl (OH • ) and superoxide radical ( O 2 • − ) are 2.7 and −2.3 eV respectively, whereas the oxidation potential of organic molecules ranged from −1 to 2 eV but, due to the difference in potential between the reactive oxygen species and the pollutant molecules, an organic pollutant molecule in contact with the hydroxyl (OH • ) or superoxide radical ( O 2 • − ) will either gain or lose electrons immediately through chain reactions resulting in the mineralization of the organic molecules forming CO 2 The most important characteristics of a photocatalytic system are the morphology, high surface area, thermal and mechanical stability, reusability, active sites, and desired band gap [17]. According to Molinari et al. [18], photocatalysis offers some advantages: 1. It avoids the application of hazardous heavy metal compounds and oxidants/reducing agents. 2. It permits the mineralization of the pollutants with the generation of safer by-products as H 2 O and CO 2 .
3. It is an alternative to traditional high energy-demanding treatment methods by using solar energy as the energy source. 4. It degrades a moderate range of pollutant concentrations. 5. It can be combined with another wastewater method to get a better water quality Additionally, the use of photocatalysis at industrial processes is still restricted due to the recombination of the photo-generated e − and h + carriers, which releases energy in unproductive heat form, fast-backward reaction, and the inability to utilize solar radiation energy since around 5% of solar radiation is UV light [19].

Photocatalysts synthesis methods
Nanotechnology is defined as the ability to structure matter in atomic and molecular levels between 1-100 nm [20]. At this scale, materials have novel size-dependent features different from their larger counterparts. Nanomaterials have been developed in several forms, such as nanotubes, nanowires, flakes, particles, rods, films, quantum dots, and colloids [21]. Nanotechnology has opened a wide possibility field for designing nanomaterials with the objective application through manipulating synthesis conditions, which allows us to design nanostructures with attractive features in shape, size, mechanic resistance, and chemistry activity [22]. The synthesis of nanomaterials with a defined morphology is an important key to getting nanostructures with desired chemical and physical properties. Moreover, the chemical activity depends not only on their size, shape, morphology, and phase composition, as well as the synthesis route [23]. This favors the preparation of nanostructured materials with desirable features, which enhance the catalytic activity of the photocatalyst. Furthermore, the power of the lamp also plays an important role with influence on the performance of the photocatalyst. Several nanoparticles synthesis methods have been reported, and each one is selected depending on the nanostructures application. The most common methods for photocatalysts synthesis are colloidal, microwave radiation, sol-gel, hydrothermal, chemical vapor deposition, photochemistry reduction, solvothermal, electrochemical deposition process, and electrospinning [24]. Among these methods, the sol-gel is the most attractive way to synthesize photocatalysts due to low cost, reproducibility, high purity, synthesis time, variable control, low process temperature, and homogeneity in particle size [25,26]. The sol-gel method for preparing metal oxide photocatalysts relies on the hydrolysis and polycondensation of the metal alkoxides used as precursors, M(OR) x (M = Si, Ti, Zr, Zn, Al, Sn and Mo) to react in aqueous or organic phase.
4 Nanostructured photocatalytic systems for dye degradation Throughout history, various photocatalytic systems have been developed to eliminate the dyes present in water. The development of photocatalytic materials through history can be divided into three groups: single component photocatalysts in suspension, heterojunction photocatalysts (multi-component in suspension), and immobilized photocatalysts, Fig. 2.

Single component photocatalysts
Elements such as TiO 2 , ZnO, ZrO 2 , Fe 2 O 3 , CdS, and ZnS are semiconductors and can act as sensitizers for light-induced redox processes due to the electronic structure of the metal atoms in chemical combination, which are characterized by an empty C B and a filled V B [27]. When these kinds of materials are irradiated with energy equal to or greater than its band gap value (eV), an e − from V B migrates to the C B generating an h + behind. The h + may react either with electron donors in the solution or with hydroxide ions to produce powerful oxidizing species like superoxide radicals ( O 2 • − ) and hydroxyl (OH • ) radicals. Nevertheless, the recombination process of the e − and h + carriers must be avoided to favor the photocatalysis reaction. TiO 2 was the first material used and investigated for the water-splitting reaction. Years later, its application increased to other fields like H 2 production, water, and air pollutant oxidation, antibacterial activity, and solar cells development. However, due to its large band gap (3.2 eV for anatase and 3.0 eV for rutile), it only can operate under UV light irradiation. Indeed, the anatase phase of TiO 2 is preferred catalytic reactions due to its conferred features by its crystallinity nature [28]. In this context, other metal oxides with a wide use for photocatalytic purposes are ZnO and ZrO 2 . The ZnO can present the crystalline phases type wurtzite, zinc blende, or rock salt. Moreover, the ZnO is seen as the substitute for TiO 2 and is considered as an efficient and promising candidate in environmental management systems because of its unique characteristics, such as direct and wide band gap in the near-UV spectral region, strong oxidation ability, suitable photocatalytic property, and a large free-exciton binding energy so that excitonic emission processes can persist at or even above room temperature [29]. For its part, the ZrO 2 can present the cubic, tetragonal, or monoclinic crystal structure (eV = 3.25-5.1 eV, depending on the preparation technique), and it belongs to the group of semiconductor materials. In this context, it has been reported that ZrO 2 can cause higher photocatalytic degradation than nano TiO 2 [30]. Moreover, the manipulation of ZrO 2 morphological tuning, porous structure control, and crystallinity development is required to enhance the light-harvesting capability, prolong the lifetime of photoinduced electron-hole pairs, and facilitate the reactant accessibility to surface active sites [31]. Fig. 3 shows the band gap of common semiconductors, while Table 1 [31][32][33][34][35][36][37][38][39][40][41][42][43][44] shows representative first group photocatalysts.
This first group of metal oxides including the TiO 2 [32][33][34][35][36][37], ZrO 2 [31,38,39], and ZnO [40][41][42][43][44] exhibited good physical features like thermal and mechanical stability, low toxicity and cost of production, reusability, and easy reactivation. However, their main drawback is their wide band gap value from 3 to 5 eV and the recombination process that present, by affecting the production of the e − and h + carriers. Therefore, a new group of photocatalysts with superior features of coupled nanomaterials was developed, which are based in heterojunctions.

Heterojunction photocatalysts
To enhance the visible light absorption efficiency of the photocatalyst, the electronic structure of the nanomaterial needs to be modified [45]. Methods such as doping, metal loading, and heterojunctions have been used to efficiently separate the photogenerated e − and h + carriers in   [44] a photocatalytic semiconductor. The created electronic structure of the new photocatalyst could decrease the recombination of the carriers due to the creation of new energy levels by trapping the electrons by reducing the recombination of the charge carries ( Fig. 4 (b)) [46]. A heterojunction is the creation of an interface between two different semiconductors with unequal band gap structure, resulting in band alignments. In this sense, different classes of heterojunctions have been reported: 1. semiconductor-semiconductor, 2. semiconductor-metal, 3. semiconductor-carbon, and 4. multicomponent heterojunction [47].
The principal requirement to create a heterojunction, is that semiconductors should exhibit dissimilar band gaps, and the narrow band gap must lie in the visible region. In addition, in the direct band gap, the highest energy level of the V B aligns with the lowest energy level of the C B to momentum [48], hence direct band gap is preferred over the indirect band gap. There are three types of conventional heterojunction photocatalysts, those with a straddling gap (type-I, Fig. 4 (a)), with a staggering gap (type-II, Fig. 4(b)), and with a broken gap (type-III, Fig. 4 (c)), [49,50].
In the type-I heterojunction, the C B and the V B of semiconductor A are higher and lower than those corresponding of the semiconductor B. In other words, the band gap of one semiconductor B is insideof the band gap of the A semiconductor. When the photocatalyst is irradiated with the appropriate energy, the e − and h + carriers from semiconductor A migrate and are caught by the C B and V B of semiconductor B. Since e − and h + carriers are caught on the same semiconductor, the charge carriers cannot be effectively separated. In the type-II heterojunction, the C B and the V B levels of semiconductor A are higher than the corresponding C B and V B of semiconductor B. Therefore, under light irradiation, the photogenerated electrons from A will migrate to C B of semiconductor B, while the photogenerated holes from semiconductor B will migrate to V B of semiconductor A. In both cases, the redox ability will be also considerably reduced because the redox reaction occurs on semiconductor with the lowest redox potential. In the type-III heterojunction, the C B and V B of semiconductor A are higher than the C B of semiconductor B, and the band gaps do not overlap. The carrier transfer is like type-II, just more pronounced. For this case, the e − and h + carriers migration and separation between the two semiconductors cannot be carried out, making it unsuitable for enhancing the separation of the e − and h + carriers [51]. From the three cases, the type-II heterojunction looks to be the most photoactive heterojunction due to its suitable electronic structure for the spatial separation of the photoinduced e − and h + carriers. Moreover, type-II heterojunction photocatalysts exhibit good e − and h + carriers separation efficiency, fast mass transfer and absorbance of light in the visible region with a band gap values under 2.8 eV [52]. In this sense, Prabhu et al. [53] synthesized djembe like ZnO microstructures by surfactant-assisted hydrothermal method, The three conventional heterojunction types, (a) type-I, (b) type-II and (c) type-III. Adapted from [49]. and its composite with graphitic carbon nitride ( g-C 3 N 4 ) was prepared by ethanolic reflux method for the first time.
The nanostructures were studied in the photodegradation of methlyne blue (MB) and rhodamine B (RhB). According to optical studies, the V B potential and C B potential were calculated as 2.83 eV and -0.64 eV for ZnO and 1.62 eV and -1.15 eV for g-C 3 N 4 . Due to the distinct positions of the V B and C B potentials between ZnO and g-C 3 N 4 , a type-II heterojunction was formed. The authors also mentioned that the heterojunction formed between djembe like ZnO and g-C 3 N 4 decreased the optical band gap energy due to the light absorption was shifted towards the visible region.
The degradation efficiency of the ZnO / g-C 3 N 4 composite for MB and RhB degradation was found to be ~95% and ~97%, respectively, compared to the pure ZnO and g-C 3 N 4 .
The authors proposed a possible visible-light-driven photocatalytic mechanism at the interface of ZnO / g-C 3 N 4 heterojunction ( Fig. 5, [53]). Pure ZnO semiconductor cannot be excited due to its wide bandgap (3.17 eV); only the g-C 3 N 4 is excited by visible light to generate e − and h + carriers. Since the C B edge potential (-1.15 eV) of g-C 3 N 4 is more negative than that of ZnO (-0.64 eV), the photoexcited electrons in the CB of g-C 3 N 4 are transferred to the CB of ZnO and then to the surface of the photocatalyst, by enhancing their photocatalytic properties. Additionally, Ramezanalizadeh et al. [54] prepared through a sol-gel hydrothermal approach a novel CoTiO 3 / CuBi 2 O 4 heterojunction semiconductor photocatalyst for the degradation of Direct Red 16 dye under LED visible light irradiation. According to the authors, compared to the pure CoTiO 3 and CuBi 2 O 4 , CoTiO 3 / CuBi 2 O 4 heterojunction showed the highest photodegradation efficiency. Based on the obtained results, the CoTiO 3 / CuBi 2 O 4 heterojunction nanocomposites showed the highest removal efficiency (91%) in pH 4.3 solutions and at a loading of 5 g/L. This effect was attributed to the efficient separation of electron-hole pairs, compatible junction formation, visible light absorption ability, suitable band gap, and a large amount of light-harvesting. Moreover, according to the scavenger experiments, the pH played a major role during photocatalytic activity. Similarly, Chen et al. [55] prepared a photocatalyst of TiO 2 grown in situ on the surface of carbon nanotubes (CNT) for the photocatalytic degradation of Rhodamine B (Rh-B) under simulated sunlight synthesized by the sol-gel reflux method. The degradation efficiency of CNT-TiO 2 for Rh-B was 50% higher than pure TiO 2 , and the addition of CNT increased the specific surface area, optical support, dispersibility, and uniformity of the synthetic material of the TiO 2 nanoparticles. The authors concluded that the n-n heterojunction structure was beneficial to accelerate the e − and h + carriers migration and improved the photocatalytic performance of the composite. In this study, the authors propossed that under the radiation of simulated sunlight, photoexcited electrons from TiO 2 C B were transferred to the CNT structure, reducing the recombination process of the e − and h + carrieres.

The p-n heterojunctions
Although the heterojunction type-II seems to be the most effective way to avoid the recombination process due to the entrapment of photogenerated e − and h + carriers, it is not effective enough to avoid the fast recombination process. Hence, the p-n-type heterojunction model was proposed to explain the accelerated migration of photogenerated species through a generated electric field in the interface between p-type and n-type semiconductors by suppressing the recombination process [56]. A p-n junction is the interaction between two types of semiconductors photocatalytic materials (p-type and n-type) inside a single crystal of photocatalyst. The p-type semiconductor contains an excess of holes, and the n-type semiconductor contains an excess of electrons. Therefore, during irradiation, when electrons and holes are photo-created, the electrons of C B from p-type semiconductor near of interface undergo diffusion towards C B of n-type (positive field) and then reacted with O 2 adsorbed on the surface to produce reactive O 2 • − . At the same time, the holes from the V B of n-type semiconductor near interface tend to flow towards V B of p-type (negative field) semiconductor, establishing the p-n junction (Fig. 6) [57]. ZnO / g-C 3 N 4 heterojunction. Adapted from [53].
The electron-hole transfer between n-type and p-type semiconductors is known as diffusion, and it will continue until the system's equilibrium state (e.g., Fermi level), reducing the recombination of the photogenerated charge carriers by the internal electric field at the p-n junction [58]. Moreover, the systems of p-n junctions are designed for operation under visible light, one of the goals of modern photocatalysis. Recently, Habibi-Yangjeh et al. [59] prepared ZnO / ZnBi 2 O 4 ( containing 5, 10, 20, and 30 wt% of ZnBi 2 O 4 ) nanocomposites with p-n heterojunction fabricated by integrating ZnO with ZnBi 2 O 4 nanoparticles via a calcination process for the photodegradation of RhB. The ZnO / ZnBi 2 O 4 nanocomposites exhibited superior photocatalytic performance for the photodegradation of the organic dye under visible light compared with the pure ZnO and ZnBi 2 O 4 . The composite ZnO / ZnBi 2 O 4 with 10 wt% achieved a photodegradation of 97% of the RhB dye after 240 min, whereas the pristine ZnO and ZnBi 2 O 4 decomposed 28% and 39% of the RhB solution after 360 min, respectively. This enhancement can be ascribed to the efficient charge carrier separation through the heterojunction structure, which inhibits the recombination of photoinduced charges. The authors mentioned that an inner electrostatic field directed from ZnO to ZnBi 2 O 4 was produced; moreover, in the presence of visible-light illumination, only ZnBi 2 O 4 is excited, and the e − and h + carriers are produced because of its narrow band gap. After the p-n heterojunction formation, the C B level of ZnBi 2 O 4 is more negative than that of ZnO. Hence, the excited electrons can inject into the C B of ZnO, promoted by the inner electrostatic field, while holes remain in the VB of ZnBi 2 O 4 . Therefore, the photogenerated charge carriers can be separated effectively by the formed inner field of p-n heterojunction reducing the recombination of the e − and h + carriers in the photocatalyst. In another work, Sang et al. [60] reported the synthesis of heterostructured Bi 2 O 3 / Bi 2 S 3 nanoflowers (1 to 2 µm of diameter) fabricated by a one-step hydrothermal method to remove of RhB and Cr(VI). The results of photocatalysis showed that removal efficiencies of RhB (99.7%) and Cr(VI) (91.8%) over Bi 2 O 3 / Bi 2 S 3 heterojunction were higher than those of pure Bi 2 O 3 and Bi 2 S 3 (< 50% of removal) under visible light irradiation after 90 min of reaction. The improved photocatalytic performance of the Bi 2 O 3 / Bi 2 S 3 heterojunctions was associated with the combination between components and their specific surface areas (46.3 m 2 g −1 , 10.1 m 2 g −1 and 12.6 m 2 g −1 , respectively. Moreover, according to the radical trapping experiments, the photogenerated h + were the major oxidative species for removing RhB, while the photogenerated e − were responsible for the photoreduction of Cr(VI). Authors argued that the excited e − on the C B of p-type Bi 2 S 3 moves to n-type Bi 2 O 3 , while the photogenerated h + still stays in the V B of p-type Bi 2 S 3 . In the Bi 2 O 3 / Bi 2 S 3 photocatalytic system, the e − and h + carriers are involved in the redox reaction. Therefore, for the system of Cr(VI) solution, the e − provided by the C B of n-type Bi 2 O 3 is being effectively consumed by Cr(VI), which is a strong oxidant. On the other hand, the h + stayed on the V B of Bi 2 S 3 would oxidize the RhB molecules directly (Fig. 7); hence the h + is the predominant radicals, which oxide RhB to simpler molecules.
For its part, Lu et al. [61] prepared a series of BiOI / KTaO 3 p-n heterojunctions via a facile in situ chemical bath strategy for the degradation of Rhodamine B (RhB) under visible light irradiation. As a result, the BiOI / KTaO 3 composites showed higher photocatalytic efficiency compared to the individual catalysts. In particular, 54 wt% BiOI / KTaO 3  98.6% RhB within 30 minutes withouth affecting its removal properties up to 3 cycles (91.1%), while only 68.1% RhB was degraded over pure BiOI. According to the authors, the improved photocatalytic performance was attributed to the successful construction of the p-n junction between BiOI and KTaO 3 , facilitating the separation and migration of photo-induced charge carriers.

Direct Z-scheme
Yu et al. [62] proposed the concept of the direct Z-scheme mechanism to explain the process of the photocatalytic formaldehyde degradation in the TiO 2 / g-C 3 N 4 presence. The assembly of a direct Z-scheme photocatalyst (Fig. 8) looks like that of a type-II heterojunction ( Fig. 4 (b)), but their e − and h + charge carriers transport processes are somewhat different [63]. Furthermore, the direct Z-scheme system does not need a redox medium, and the photocarriers directly transfer across the interface of both semiconductors without a charge carrier intermediary. Therefore, the transmission distance is reduced, and the photocatalytic efficiency is enhanced. Under light irradiation, the photogenerated electrons in semiconductor A, with a lower reduction ability, recombine with the photogenerated holes in semiconductor B with a lower oxidation ability [64]. Thus, the photogenerated electrons in semiconductor B with high reduction ability and the photogenerated holes in semiconductor A with a high oxidation ability are kept in their particular sites to get the spatial separation of charge carriers to improve the redox capacity of the photocatalytic structure. In this manner, the charge-carrier migration is more promising than in type-II junction because the migration of electrons from the C B of semiconductor A to the hole-rich V B of semiconductor B is thermodynamically possible by the electrostatic attraction between the e − and h + . Direct Z-scheme offers advantages as fast e − and h + carriers separation efficiency, good redox ability, corrosion resistance, and low fabrication cost [49,65].
In this sense, Zhao et al. [66] prepared a Z-scheme heterogeneous g-C 3 N 4 / FeOCl photocatalysts using the calcination method. The composite with a morphology of a ribbon-like sheet was used to eliminate RhB from water. Compared with the pure FeOCl material (60% of RhB removal), the Z-scheme g-C 3 N 4 / FeOCl composites revealed a higher photocatalytic activity (90% of RhB removal) under visible light irradiationafter 60 minutes of reaction. The authors argued that the enhanced catalytic activity of the g-C 3 N 4 / FeOCl material was attributed to the formation of a Z-scheme between g-C 3 N 4 and FeOCl (Fig. 9). Authors explainded that when g-C 3 N 4 / FeOCl is irradiated with visible ligth, the electrons from the V B of the g-C 3 N 4 and FeOCl were transferred to their respective C B . After that, the electrons were transferred from the C B of FeOCl to the V B of the g-C 3 N 4 and combined with h + . Then, these electrons transformed the H 2 O 2 into the OH • . In this process, the H 2 O 2 served as the electron acceptor which further successfully limited the recombination of holes and electrons. On the other hand, on the surface of the FeOCl material, the Fe 3+ was transformed into Fe 2+ with the presence of H 2 O 2 and the irradiation of visible light; hence, the Fe 2+ was easily reacted with H 2 O 2 to generated the OH • for removing the pollutant. Due to the C B of the g-C 3 N 4 was more negative than E 0 (O 2 / O 2 • − ) and the V B of the FeOCl was more positive than E 0 (OH • /OH − ), the  In other study, An et al. [67] prepared a core-shell Ag 2 CO 3 @g-C 3 N 4 photocatalyst by two-dimensional coating nanosheet g-C 3 N 4 on the surface of Ag 2 CO 3 for the photodegradation of methyl orange (MO). According to the authors, the Ag 2 CO 3 @g-C 3 N 4 (5 wt.%) composite exhibited the best degradation efficiency, up to 96.7% and 87.3% after five cycles. However, the photodegradation performance was in a g-C 3 N 4 dose-dependent response (from 1 wt.% to 10 wt.%). The authors mentioned that the photocatalytic performance was due to the faster-photogenerated carrier migration efficiency derived from core-shell structure and chemical bond hybridization effect arising from Ag 2 CO 3 and g-C 3 N 4 . Moreover, the excellent performance of photocatalyst was due to the Z-scheme structure formed by the Ag 2 CO 3 @g-C 3 N 4 photocatalyst, which effectively avoids the accumulation of photoinduced electrons in the Ag 2 CO 3 and inhibits Ag + photoreduction, which significantly improves the stability of Ag 2 CO 3 . Recently, Zhang et al. [68] reported a Z-scheme-based BiOI/CdS heterojunction with efficient photocatalytic degradation of RhB (20 mg/L) under visible light. The in-situ stirring and calcining method synthesized the Z-scheme-based BiOI/CdS heterojunction. Three BiOI/CdS composites were prepared (the mass ratio of BiOI to CdS was 60 wt.%, 80 wt.%, and 100 wt.%, respectively referred to as 0.6-BiOI/CdS, 0.8-BiOI/CdS, and 1.0-BiOI/CdS). The removal efficiency of RhB was BiOI < CdS < 1.0-BiOI/CdS < 0.6-BiOI/CdS < 0.8-BiOI/CdS. Moreover, after 4 cycles, the degradation of the 4 th experiment reached 98% of the first, indicating the 0.8-BiOI/CdS composites exhibited excellent stability. According to the authors, the free radical capture experiments showed that • O was the main active substance in the degradation process.

The g-C 3 N 4 -based photocatalysts
The g-C 3 N 4 is a characteristic material belonging to the second group of designed photocatalysts with a band gap of 2.7 eV, which means that operates under visible light. g-C 3 N 4 shows a two-dimensional (2D) planar π conjugation structure, which could improve the electron transfer mechanism due to its prominent electronic activity [69]. In addition, due to its high nitrogen content, g-C 3 N 4 may provide more active reaction sites than other N carbon materials by contributing to the photocatalytic reaction [70].
However, its fast recombination of the e − and h + carriers reduces its photoactivity efficiency as only photocatalyst. Therefore, it is recommended that g-C 3 N 4 be coupled to another semiconductor material to improve its photocatalytic activity by creating an interesting electronic structure as a whole. For example, Wei et al. [71], through the solvothermal method, synthesized the ternary heterojunction g-C 3 N 4 / Ag / ZnO with a 3D flower-like structure and 1.5 µm of diameter for the photodegradation of MO. The ternary heterojunction g-C 3 N 4 / Ag / ZnO photocatalytic activity was better compared to the pure g-C 3 N 4 , g-C 3 N 4 / ZnO composite, and g-C 3 N 4 / Ag composites. According to the authors, the plasma effect of Ag nanoparticles can be used to expand the response range of the photocatalyst to visible light. Meanwhile, Ag particles on the heterogeneous interface of g-C 3 N 4 and ZnO play the role of conducting electrons, which are beneficial to separating of photogenerated electrons and holes. Zhao et al. [72] prepared a photocatalyst of Ag /WO 2.9 / g-C 3 N 4 , demonstrating better adsorption capacity promotion than traditional WO 3 . The composite was prepared by calcination and compared with the Ag / WO 2.9 and g-C 3 N 4 , the Ag / WO 2.9 / g-C 3 N 4 showed a graphite-like carbon nitride as a substrate, and nano-sheets WO 2.9 attached to silver nanoparticles are stacked on g-C 3 N 4 . This unique structure generated a large specific surface area, coupled with the oxygen deficiency inherent in WO 2.9 , which favored the adsorption of dye molecules. Moreover, the photocatalytic tests (under visible light irradiation (λ ˃ 420 nm)) on Ag / WO 2.9 , g-C 3 N 4 , and Ag / WO 2.9 / g-C 3 N 4 showed that Ag / WO 2.9 / g-C 3 N 4 has the best adsorption activity and photocatalytic degradation ability under visible light conditions. The authors also mentioned that the formed photocatalyst constitutes a Z-scheme, which effectively separates the C B region and the V B region and performs efficient regional reaction. Likewise, Xue et al. [73] prepared a hetero-structured photocatalyst consisting of two-dimensional g-C 3 N 4 nanosheets and commercial MoO 3 microparticles through a simple mixing and annealing process for the photodegradation of RhB. According to the authors, the MoO 3 / g-C 3 N 4 composite showed a significant improvement compared with individual MoO 3 or g-C 3 N 4 and their physical mixture. Moreover, with the results of electron spin resonance, the authors concluded that a direct Z-scheme charge transfer between MoO 3 and g-C 3 N 4 not only causes an accumulation of electrons in g-C 3 N 4 and holes in MoO 3 , but also boosts the formation of superoxide radicals and hydroxyl radicals. The total dye was photodegraded in 15 minutes using 25 mg of catalyst dispersed into 50 mL of RhB solution ( 10 mg L −1 ). In this context, several heterojunctions materias have been used for MO [66,71,72,[74][75][76][77][78][79], RhB [53,59,60,61,67,72,73,78,[80][81][82][83][84][85][86][87][88][89][90][91][92][93][94][95], MB [53,72,75,77,94,[96][97][98][99][100][101][102][103][104][105] and other pulliting dyes [54,75,100,103,[106][107][108][109][110][111][112][113][114][115][116][117][118] removal, as shown in Table 2.

Immobilized photocatalysts
The typical suspended photocatalytic systems of powders show good mass transfer coefficients and the advantage of a greater surface area against the immobilized system. However, their disadvantage relies on the recovery of the powders after the photocatalytic reaction, increasing the process costs, which is a drawback [119]. In addition, the loss in the photoactivity of the recycled powders is another challenge related to the separation techniques. The immobilized systems take better advantage of the irradiated light and do not require a post-treatment for recovery the photocatalyst. The features of the semiconductor-active species and its interaction with the employed support are key factors to achieve a good photoactivity. Unfortunately, the immobilized system's configuration is only effective in arranging with a high surface-to-volume ratio, e.g., in microchannel reactors [120]. For example, Bahrudin et al. [121] studied the decolorization of methyl orange (MO) using immobilized TiO 2 / chitosan-montmorillonite (TiO 2 / CS-MT), a combination of TiO 2 as the top layer and CS-MT as the sub-layer on a glass plate. The authors mentioned that the immobilized CS-MT film showed better performance over the CS film since the former adhered stronger and swelled less than the latter, which showed its favorability in the aqueous medium. Moreover, the bilayer photocatalyst could remove the MO from the solution 3 times faster than the single TiO 2 within 90 min of irradiation under a UV-Vis lamp due to the strong adsorption of dye by the CS-MT sub-layer.
Ounas et al. [122] presented a simple and effective approach to prepare a polymethyl methacrylate-TiO 2 (TiO 2 / PMMA) film photocatalyst, by a cheap and lowcost technique. The characterization of the film by XRD, FTIR, and Transmittance spectroscopy confirmed that the anatase TiO 2 has been deposited on the surface of the polymer. The film prepared was subsequently used in photodegradation of MB under artificial UV irradiation and showed a good prospect for the immobilization of TiO 2 intended for the photodegradation of pollutants generally present in waters. However, the authors mentioned that the method described can still be improved to become easier and faster in a near future. Furthermore, de Araujo Scharnberg et al. [123] evaluated the photocatalytic properties of TiO 2 under porous ceramics support for the degradation of RhB. For this, the anatase TiO 2 calcined at 400 °C was prepared by the sol-gel method and supported in a porous ceramic substrate by a dip-coating process. The heterogeneous photocatalysis showed excellent results, with the degradation of up to 83% of RhB. The Authors also mentioned that after the usage, a major part of the catalyst stayed at ceramics, making possible to recover it, or to use the catalyst in a continuous flow reactor. Additionally, Inderyas et al. [124] reported that ZnO nanoparticles were immobilized on polyurethane foam (PUF) and employed for the degradation of Acid black 1 dye. In this study, the process variables like dye concentration, pH, the concentration of H 2 O 2 , irradiation time were optimized for maximum dye degradation. The ZnO / PUF showed high efficiency for the degradation of AB1 dye, and up to 86% and 65% dye degradation was achieved under UV and solar light irradiation at neutral pH, 4% H 2 O 2 , 240 min/sunlight, and 75 min/UV irradiation time using 40 mg L −1 dye initial concentration. Moreover, the reductions in BOD, COD, and TOC values confirmed that the ZnO/PUF was efficient. Das and Mahalingam [125] prepared a physical mixture of rGO and g-C 3 N 4 along with TiO 2 (ratio of 1:1:1). The nanocomposites were immobilized in a polystyrene film using the facile solvent casting method for the degradation of remazol turquoise blue dye. The results using the immobilized catalyst mixture film gave 92.25% of TOC reduction, 94% of decolorization in 140 min, and a 72% of degradation in the fourth time of reuse.

Influence of operational parameters on the photocatalytic degradation
According to the evidence, the photocatalysts synthesized by different methods are attractive materials with high photocatalytic properties for diverse dye degradation from water. However, their effects are in a shape-, size-and dose-dependent response. In general, these materials are low-cost, efficient, reusable, and environmentally friendly for wastewater treatment. Additionally, the efficiency of these materials mainly depends on the experimental conditions, as discussed in Subsections 5.1-5.6.

The pH influence
The pH of the solution in photocatalytic reactions determines electrostatic properties such as the surface charge of the photocatalyst, formation of hydroxyl radicals, size of the aggregates that it forms, and the band edge position of metal oxides used as photocatalysts [160]. Furthermore, the pH can influence the adsorption-desorption characteristics of the catalyst surface [161]. The photocatalyst surface can be protonated and deprotonated under acidic and alkaline conditions, respectively (Eqs. (7) and (8)): The operating pH affects the isoelectric point and the surface charge of the photocatalyst. The reaction occurs at a different pH values from the isoelectric point (point of zero charges, pzc) where the surface of the material is not charged; under this value, the material is positively charged, and above this value, the catalyst is negatively charged. At pH > pzc, the adsorption of positively charged contaminants is preferred, while at pH < pzc, the adsorption of negatively charged contaminants is favored [162]. Values close to neutrality have no significant effect on the operation. Although the pH primarily affects the adsorption of charged contaminants, it also has a role in the photocatalysis of those neutral molecules that tend to dissociate into charged species. Therefore, the pH affects the surface of the photocatalyst and the dissociation of the dye [163].

Process temperature
In most photodegradation reactions, these are carried out at STP and do not require cooling or heating of the reaction system due to the photonic activation. Preferably the reactions should occur between 20 °C and 80 °C, since at high temperature (T ≥ 80 °C), the recombination process of charge carriers is favored. Furthermore, the exothermic adsorption of reactant is not favored and tends to become the rate-limiting step [162,164]. The increasing temperature does not favor adsorption, which becomes the inhibitor of the reaction, while at low temperature, inefficient desorption of final products is presented, and an increment in the activation energy is required to carry out the reaction [165]. Therefore, there is no need to waste energy for heating water that possesses a high heat capacity.

Photocatalysts loading
When the catalyst loading is increased, there is an increase in the contact surface of the catalyst, and a variation of the average dye-photocatalyst ratio could generate losses in the surface area by aggregation (particle-particle interactions) due to excess of the solid concentration causing a decrease in the number of exposed active surface sites [160]. The decrease of degradation at higher catalyst loading may be due to the deactivation of activated molecules by collision with ground-state molecules [166]. Moreover, with the increment of photocatalyst loading, UV light penetration can be reduced due to saturation of the aqueous medium, affecting the photodegradation rates [163]. Therefore, an optimum amount of photocatalyst must be used to ensure total absorption of light photons and avoid unnecessary excess. A constant agitation with a magnetic stirrer at the reactor base and an air flux bubbled continuously inside the reactor to provide enough O 2 are recommended to keep powder particles fluidization.

Dye concentration
The dye degradation rate under photocatalytic processes depends on its initial concentration [167]. When the initial dye concentration is increased, many molecules are adsorbed on the catalyst surface, and this may promote an inhibiting effect on the reaction of the dye with photogenerated holes or hydroxyl radicals because of the lack of any direct contact between them [168]. Furthermore, when the concentration of dye is increased, the dye molecules adsorb light (UV-screening effect), and the photons hardly reach the photocatalyst surface. Thus, the photodegradation efficiency decreases. On the other hand, the Langmuir-Hinshelwood model describes the kinetics of photocatalytic reactions of aquatic organics pollutants [169]; this model is based on the next assumptions: 1. limited surface adsorption sites, 2. only single layer adsorption, and 3. no interactions between molecules after adsorption.
Langmuir-Hinshelwood model is expressed as: where C is the concentration of aquatic organic, k ad is the adsorption equilibrium constant and k r is the intrinsic rate constant, which takes into account parameters such as catalyst mass, efficient photon flow, O 2 layer, etc. [170]. When the concentration of dye is so low (millimolar), the Eq. (9) can be simplified to an apparent first-order equation [171]: The linear region can be obtained from the plot of ln( C / C o ) vs t, in which the slope gives the rate constant of photodegradation. The half-life time (degradation of dye to its 50%) is calculated as:

Light source and intensity
Relatively high light intensity is required to provide photocatalyst particle enough photons energy. Hence, it is essential to establish the range of radiation with which the solution must be irradiated. It has been shown that the reaction rate is proportional to the radiant flux Φ. However, above a certain value, the reaction rate becomes proportional to Φ 1/2 , indicating strong electron-hole recombination. In this context, Ollis et al. [172] studied the effect of light intensity on the kinetics of the photocatalytic reaction, and the following results were found: 1. At low light intensities ( 0-20 mW/cm 2 ), the rate would increase linearly with increasing light intensity (first-order) due to reactions involving e − and h + carriers formation is predominant.

At intermediate light intensities beyond a certain
value ( ≈ 25 mW/cm 2 ), the rate would depend on the square root of the light intensity (half order) because the e − and h + carriers separation compete with recombination causing a lower effect on the reaction rate. 3. At high light intensities, the rate is independent of light intensity if the temperature is low.
Other important factor is the lamp-reactor geometry where the reaction takes place; the geometry and fabrication materials could favor the homogeneous dispersion of the light.

Disadvantages and perspectives in photocatalysis
Even though researchers have made tremendous progress in the photocatalysis field, some challenges remain about the operation of photocatalysts in industrial applications. For example, the light distribution inside the reactor and the configuration of the reactor are the main issues to be addressed. According to Ahmad et al. [173], the scaling-up of a photocatalytic reactor has been limited due to the reactor design have not been able to address the two most important strictures; light distribution inside the reactor through absorbing and scattering liquid to the photocatalyst, and to provide high surface areas for photocatalysts coating per unit reactor volume. In addition, the costs of incident photon production must be considered in the process economy when talking about treating huge volumes of wastewater. Furthermore, the chemical restrictions play a crucial role in the performance of dye removal such as the interfacial charge transfer, improve the charge carriers separation, and the inhibition of charge carriers recombination process. Some challenges like mass transfer limitations, catalyst deactivation, generation of intermediate products and by-products, and the multi-complex optimization of the materials and the reactor configuration limit the real industrial applications. On the other hand, several trends for further development are currently under investigation. These trends include: 1. The development of economical methods for the preparation at large scale of nanomaterials with controlled morphology.

The development of the characterization techniques
and instruments to elucidate and confirming the migration pathways of electron-hole pairs in heterojunction photocatalyst. 3. The hybridation with photocatalytic components in a single device. 4. The fine control of increasingly complex nanoarchitectures and (v) the use of novel non-oxidic materials.
According to the literature, there are some challenges to be achieved for scaling up applications of photocatalysts. However, their use for dye degradation is an active research area with potential applications as an alternative for wastewater treatment. Therefore, further research efforts should be dedicated to solving these challenges.

Concluding remarks
According to the evidence, the potential application of photocatalysis is an efficient alternative to remove dyes from water. Most recent works reported promising degradation results (> 90% of dye degradation) in shorter reaction times. Nevertheless, the search for photocatalysts with desired characteristics to induce the total oxidation of dye molecules under visible light irradiation in an economically accessible way is encouraging. From the transition of single component photocatalysts to the design and application of heterojunctions and immobilized photocatalytic systems, important problems were solved, such as extending the life of the photogenerated species and the use of a lower photocatalytic powder amount per volume of wastewater. In addition, the design of immobilized photocatalysts also solved problems such as powder separation and recovery stages. However, it attracted new challenges (e.g., reduction of surface area). At present, the demand for environmental sustainability that humanity is facing has forced researchers to design photocatalytic systems that avoid the recombination process, with various life cycles and low cost energetic, easy to manufacture, and economically accessible at laboratory scale. An ideal photocatalyst should fulfill requirements such as visible-light activity, high solar energy conversion efficiency, proper band gap structure for redox reactions, high photostability for longterm applications, and scalability for commercialization. In fact, several researching groups agree that the design of active nanostructures under visible light is one of the main challenges for the development of these materials. Some current limitations may be resolved in the future, by coupling photocatalysis with other emerging technologies.