A Review of Visible Light Responsive Photocatalysts for Arsenic Remediation in Water

Abstract

This review summarizes the progress over the last fifteen years in visible light reactive photocatalysts for environmental arsenic remediation. The design and performance of several materials including (1) doped and surface functionalized TiO₂, (2) binary composites combining TiO₂ with another semiconductor that absorbs visible light radiation or a metal (Pt), (3) ternary composites incorporating TiO₂, a conductive polymer that can retard electron-hole recombination and an excellent adsorbent material for the removal of As(V), (4) tungsten, zinc, and bismuth oxides, (5) g-C₃N₄ based catalysts, and (6) M@AgCl core–shell structures. These results show that long reaction time remains a major challenge in achieving high As(III) oxidation.

1. Introduction

Arsenic is fairly abundant in Earth’s crust, with a natural occurrence of 2–3 ppm on average, which makes it the 20th most abundant element in the crust. It exists in four main oxidation states, namely +V (arsenate), +III (arsenite), 0 (arsenic), and −III (arsine), of which the first two are the most abundant in the environment. In oxygenated water (such as surface waters) arsenic usually occurs in the form of As(V), whereas under reducing conditions, such as those often found in deep lake sediments or groundwater, the predominant form is As(III) [1]. Organoarsenic compounds, such as dimethylarsinic acid or monomethylarsonic acid [1], are also found in water.

Arsenic can diffuse in the environment from both natural and anthropogenic sources; the latter are derived from activities such as coal mining, smelting, fertilizers, and pesticides, and from the hide tanning, semiconductor, and aluminum industries [2]. However, arsenic contamination of drinking water is generally derived from geological sources. Bangladesh and India (West Bengal) are among the countries with the highest levels of contamination. Levels of arsenic above safe limits were also found locally in several countries in Europe, Australia, New Zealand, Brazil, and Canada [3].

Through prolonged exposure to arsenic-contaminated drinking water, the element accumulates in tissues. Signs of chronic arsenicism include dermal lesions, peripheral neuropathy, bladder, lung, and kidney cancers, urinary tract infections, peripheral vascular disorders, and skin cancer [2]. For acute toxicity, arsine is considered to be the most toxic form, followed by arsenites (As(III)), arsenates (As(V)), and organic arsenic compounds. The World Health Organization (WHO) estimated that about 180 million people in 50 countries have been exposed to toxic arsenic levels (at least 10 μg/L in drinking water). Consequently, the guideline value for the permissible limit for arsenic in drinking water according to WHO is 10 μg/L; however, every effort should be made to maintain levels as low as possible [4].

For these reasons, abundant research has been devoted to the quest for remediation methods, especially for water treatment. It must be considered that none of the known technologies will work in every situation. The applicability of a given technology depends on the types and concentrations of the arsenic species, the pH of the medium, the presence of other dissolved species or organic contaminants, the volume of water to be treated, the cost, and the target value for concentration levels (namely, whether or not the obtained water must fit the requirements for drinking water).

Although the number of articles dealing with the application of photocatalysis to arsenic removal in water is relatively small, there are five recent reviews on the application of metal oxides for arsenic remediation in water under UV irradiation [1,5,6,7,8], none of which focused on visible light irradiation. One very recent review [9] deals with photocatalysis by both UV and visible light. Therefore, the present review emphasizes the visible-light-responsive performance of the semiconductor photocatalysts for arsenic remediation in water.

Methodologically, we decided to focus on photocatalytic methods of removal and to only take into consideration those papers meeting criteria that would ensure reproducibility of the results. We chose to take into consideration only those papers where all the following data are clearly stated: (i) As(III) or arsanilic acid concentrations; (ii) catalyst concentration (where applicable); (iii) required time; (iv) characteristics of the light source; and (v) percentage of removed pollutant.

2. Conventional Remediation Techniques

In many cases, arsenic levels in groundwater are much higher than the allowed limit of 10 μg/L; therefore, a variety of methods are used for decontamination. Mostly, they can be grouped into the categories of adsorption and ion exchange, coagulation–precipitation processes, membrane processes, and bioremediation.

As the various forms of arsenic undergo acid–base equilibria, pH is of paramount importance in determining whether arsenic will be present in neutral or ionic form [1,10], as shown in Figure 1.

When the problem to be addressed is the treatment of drinking water, pH must be close to neutrality, at which value As(III) is mainly in a neutral form, while As(V) is in anionic form. In general, as arsenite is electrically neutral at a pH of 9.2, most current arsenic remediation methods for drinking water, which exploit the presence of surface charge, typically perform better in removing the less toxic As(V); therefore, peroxidation usually has to be applied for arsenite removal [11].

2.1. Oxidation

Oxidation is a preliminary technique used to convert As(III) to As(V); the latter must then be removed from water using a physical technique. The preferred oxidizing techniques include the use of chemicals, electrochemistry, and photochemistry.

Chemical oxidants include different species, such as ozone, chlorine, and KMnO₄, which can rapidly and completely oxidize As(III) at concentrations of 50 μg/L or higher. However, in some cases, pH must be kept above 7.5, and other contaminants, such as sulfides, may interfere, slowing the process. Other oxidants, such as oxygen, air, chloramine (NH₂Cl), chlorine dioxide, hydrogen peroxide, and ferrate, generally require longer contact times and fail to achieve complete oxidation of As(III) [12,13]. Hypochlorite may also be used, but complete oxidation requires a concentration of 500 μg/L [12,14]. If the oxidizing species is a solid, it also acts as an adsorbent for the resulting As(V) [13,15,16]. The oxidation of As(III) with H₂O₂ and Fe(II) is important in natural and technical systems at pH values up to 8, below which As(III) is not oxidized at appreciable rates by H₂O₂ alone [17].

One effective solution was to pump oxygenated water into the aquifer prior to arsenic removal to reduce As concentrations to <10 μg/L [18].

As(III) oxidation can also be achieved by biological methods by exploiting bacteria or algae [19].

Electrochemistry is a viable system for the anodic oxidation of arsenite without the use of noxious chemicals [20]. Oxidation may be direct, where the As(III) species are directly adsorbed on the electrode surface and oxidized, or indirect if the process involves the electrochemical generation of an oxidizing species, such as O₂ [21]. A huge amount of research is presently devoted to the quest for an efficient electrode material [15,22]. Electrochemical methods seem promising owing to low energy consumption and good As removal, even down to <1 μg/L, although their efficacy in groundwater is lower than in synthetic water in several instances [20].

Photochemical oxidation of As(III) can be achieved when light-absorbing substances such as H₂O₂, Fe(III) complex, NO₃⁻/NO₂⁻, and I⁻, produce transient high-energy species under UV radiation [11]. Photocatalytic oxidation has often been used [23], in many cases by TiO₂-based photocatalysts; however, owing to its wide bandgap, TiO₂ is only activated by UV light, thus allowing the exploitation of only a small fraction of solar light or requiring the use of UV lamps. Photocatalysis by visible light is detailed below.

Table 1. Main oxidation methods to convert As(III) into As(V).

Oxidation Method	Advantages	Disadvantages	Ref.
chemical oxidation	suitable for high [As(III)], fast and complete for some oxidants, some oxidants may act as adsorbents	need for stoichiometric chemicals, potentially toxic residues, efficiency may be strongly pH dependent, possible interference by other substances present in groundwater	[12,13,18]
in situ oxidation	simple, cheap, effective (<10 μg/mL), no sludge	very few reported examples	[18]
electrochemistry	no chemicals required, high pH groundwater is reduced to ≈pH 7, viable cost	need for electrochemical apparatus, possibility of formation of noxious by-products, lower efficacy in groundwater than in synthetic water	[15,20]
biochemical oxidation	biocompatible, algae may act as adsorbents	need for selected strains of bacteria or algae and controlled temperature, proved effective only in laboratory conditions	[19]
photocatalysis (UV)	efficient, no chemicals required	need for UV radiation, possible interference from other ions present in groundwater	[13,23]

sums up the main oxidation methods used for As(III), with their main advantages and drawbacks.

In the following paragraphs, the main conventional techniques that can be applied after oxidation for the removal of arsenic from water are described.

2.2. Adsorption

The most widely used system for arsenic removal from groundwater is adsorption, due to its simple operation, potential for regeneration, and minimal toxic sludge generation. Processes based on the use of natural, locally available adsorbents are considered to be more accessible for low-income countries, as they have a smaller investment cost and a lower environmental impact (CO₂ emission) [24]. Adsorption of As is highly dependent on several parameters, including pH, the presence of other cations, particle size, surface porosity, the oxidative state of As(III or V), contact time, and adsorbent dosage.

The optimal conditions are determined by the types of surface functional groups and surface charges present on the adsorbent being used. A wide variety of materials has been used for arsenic adsorption. Among them, rocks, minerals, and other materials containing iron oxides or hydroxides [25] and/or alumina, such as red mud from the treatment of bauxite, carbon-based materials such as activated carbon [26], which is frequently obtained during combustion of a wide variety of locally available agricultural wastes, biochar [27,28,29] graphene [30], and a huge variety of polymer-based materials and composites [31]. Frequently, these materials are manufactured in the form of nanoparticles, which have large surface areas, highly porous surfaces, and large numbers of active sites [8]. Other low-cost materials for arsenic adsorption are those classified as biosorbents, among which are biopolymers such as chitosane, alginate, or cellulose, raw waste material from agricultural activities such as rice husks, biomass from wood processing [32], and microalgae and fungal biomass [24]. All these materials have the advantage of being locally available and easily manageable, even in low-income countries.

An intrinsic problem of this remediation approach is the need for subsequent disposal of the contaminated sorbent. In principle, two approaches are possible: (a) desorption of arsenic from the adsorbent medium, to recover and reuse the latter and/or to obtain and utilize the arsenic itself; and (b) dumping of the contaminated adsorption material. As arsenic does not have many industrial applications, its recovery is usually not economical, so most of the time it is disposed of. However, it cannot be destroyed, and incineration of the sorbent is not viable due to the high volatility of arsenic compounds. Therefore, safe management of the spent sorbent medium requires prior application of stabilization/solidification techniques to avoid leaching of toxic compounds from landfills. For instance, it can be incorporated into building materials or in a cementitious binder [33]. Iron hydroxides can be thermally treated in order to promote arsenic fixation and, therefore, safer disposal of the solid residue [34]. Some instances instead use the recovery approach: about 85% of As adsorbed on δ-FeOOH nanoparticles was recovered by treatment with NaOH solution, to be used as a precursor for the synthesis of Ag₃AsO₄, which is a material with outstanding photocatalytic activity [35]. Desorption methods were also developed for the reuse of the sorbent material [36]. A recent review [2] dealt with the disposal of sorbents, whereas another [37] investigated plants performing arsenic adsorption from water in continuous mode and the latest progress in the regeneration and recovery of arsenic. The disposal of the exhausted bed was also discussed.

2.3. Precipitation

The precipitation methods are mainly used for the remediation of heavily arsenic-loaded wastewater, such as those derived from mining and related activities. In industrial wastewater, precipitation of As(III) in the form of sulfide As₂S₃ is almost complete at pH 2 [38]. Coagulation/flocculation is usually obtained by iron or aluminum salts used as flocculants. The technology of coagulation–flocculation is simple, only common chemicals are used, installation costs are low, and it can be easily applied to large volumes of water. During this process, As(V) is removed more effectively than As(III); however, As removal is hindered by the presence of sulfates, silicates, and phosphates, and simple setting of the solution is not sufficient for effective decontamination, and filtration is a necessary step [39]. During the precipitation processes, often owing to low selectivity, a large amount of sludge is produced [40], which is usually disposed of in landfills. Precipitation is sometimes proposed as a method for the treatment of wastewater produced by the management of arsenic-laden waste created during the regeneration of exhausted adsorbents [41].

As in the case of adsorption, the production of solid waste with high arsenic content during precipitation poses leaching problems [42]; however, as the amount of arsenic in this type of waste is important, remediation techniques are also frequently directed to the recovery of arsenic for reuse [43,44].

2.4. Membrane Technologies

Membrane technology is a promising area for arsenic removal from water. The most commonly developed membrane separation processes are reverse osmosis (RO), electrodialysis (ED), microfiltration (MF), ultrafiltration (UF), and nanofiltration (NF). Microfiltration, ultrafiltration, nanofiltration, and reverse osmosis are pressure-driven processes, whereas electrodialysis is an electro-driven process. For arsenic remediation, the most appropriate technique is nanofiltration due to its ability to eliminate multivalent ions in water [45]. A recent review [46] compared the various membrane techniques and discussed the research trends, indicating advantages and drawbacks, the latter of which were typically the issues of low durability [47] and fouling. Fouling is the deposition of solutes on the membrane surface and within pores, and it causes mass transfer resistance, low water flux, and low selectivity [45]. In another review [39], emergent membrane-based technologies such as forward osmosis, membrane distillation, and electrodialysis were discussed.

2.5. Other Techniques

Ion exchange was realized using a vast number of materials, such as metal–organic frameworks [48] or polymeric ion exchange resins containing dispersed metal oxide nanoparticles [49]. Resins require periodic regeneration via treatment with solutions of appropriate counterions and must eventually be disposed of in landfills, where they should not give rise to leachate.

Bioremediation using the biomass of bacteria, algae, fungi, and yeast has gained much attention from researchers in the last few decades as an environmentally friendly and cost-effective approach [50]. Arsenic-resistant bacterial strains can be genetically engineered for better performance. Microbial consortia, namely organized collectives of microorganisms, either natural or engineered, can work synergistically [51].

Although arsenic bioremediation using bacteria can be very effective within laboratory conditions, bacterial ecology in the environment may present important differences with respect to their behavior under controlled parameters. Hence, it is essential to develop bioremediation systems in field applications to enhance their commercial viability.

Electrochemical methods are interesting because of their flexibility, environmental compatibility, low cost, and high efficiency. They can be distinguished using separation and oxidation methods [21]. Separation methods include electrocoagulation, electrodialysis, and electrosorption.

Directional solvent extraction with decanoic acid was applied at low temperature to achieve purification from arsenic to meet drinking water standards [52].

Generally speaking, evaluation of the viability of a given decontamination technique depends on a vast number of factors, including economic issues, acceptability by the users, management of the resulting waste, ease of maintenance of the plant, and so on, and considerations may vary greatly based on the location of the plant and the characteristics of the contaminated water [53]. Particularly in low-income countries, even where decontamination plants have been built, their long-term efficiency is often not guaranteed [54], and residents often discontinue their use of water treatment plant for various reasons [55].

3. Heterogeneous Photocatalysis for the Removal of Arsenic in Water

Heterogeneous photocatalysis, which mainly refers to semiconductor photocatalysis or semiconductor-sensitized photoreactions, is considered a promising alternative to conventional processes used for the removal of toxic As(III) from water. The process mainly oxidizes As(III) into As(V), which may be then removed from the solution either by precipitation or by the addition of a suitable adsorbent, typically iron oxides such as Fe₂O₃ and Fe₃O₄.

In the photocatalytic process, a chemical reaction is activated or its rate is increased when a semiconducting photocatalyst is irradiated using photons with energy that matches the band gap energy of the semiconductor, resulting in the formation of excited electron–hole ($e^{-}$-${ h}^{+}$) pairs. The excited electrons are promoted from the valence band (VB) to the conduction band (CB), thereby generating holes in the VB (Figure 2) [56]. Owing to the band bending, electrons and holes move to the surface of the semiconductor, where, in the absence of electron and hole scavengers, their recombination occurs in few nanoseconds. Recombination is avoided if a proper scavenger or surface defect state is available to trap the electron–hole pairs, and in this case, the electron and hole can be involved in thermodynamically facilitated electron transfer reactions with the solvent or other species. Adsorbed water or hydroxyl groups present on the semiconductor surface can be oxidized by $h^{+}$, generating $H{O}^{•}$ (Equation (1)), while adsorbed $O_{2 ads}$ can be reduced by $e^{-}$, generating superoxide radicals, $O_2^{•-}$ (Equation (2)):

$$H_2{O}_{ads}∕O{H}_{surf}^{-}+h^{+}\to H{O}^{•}(+H^{+})$$

(1)

$$O_{2 ads}+e^{-}\left(+H^{+}\right)\to O_2^{•-}\left(H{O}_2^{•}\right)$$

(2)

The mechanism of photocatalytic oxidation of As(III) has been studied for many years, and research has focused on determining the major oxidant in the system, either $O_2^{•-}$, $H{O}^{•}$, or $h^{+}$ [57,58]. The oxidation of As(III) to As(V) proceeds through monoelectronic steps with the formation of As(IV), as described by the following equations:

$$As\left(III\right)+H{O}^{•}\to As\left(IV\right)+O{H}^{-}$$

(3)

$$As\left(III\right)+H{O}_2^{•}∕O_2^{•-}\to As\left(IV\right)+H{O}_2^{-}∕O_2$$

(4)

$$As\left(III\right)+h^{+}\to As\left(IV\right)$$

(5)

Subsequently, As(IV) is readily oxidized to As(V) by $O_2^{•-}$, $H{O}^{•}$, or molecular $O_2$ [59]. With all these oxidants, the heterogeneous photocatalytic process is highly efficient at transforming As(III) into As(V) [60,61].

The most studied material for the photocatalytic oxidation of As(III) in water under UV light is TiO₂ coupled with an iron oxide, such as α-Fe₂O₃, which acts as an adsorbent of As(V), improving the total arsenic removal. Although the band-gap energy of α-Fe₂O₃ is about 2.0 eV, TiO₂/Fe₂O₃ is not efficient in the As(III) oxidation reaction under visible-light irradiation, because Fe₂O₃ is a rather poor photocatalyst due to rapid electron–hole recombination [62].

A good photocatalytic material must show chemical stability under a wide variety of environments. Metal oxide semiconductors with a band gap in the range of 2–4 eV generally possess this important property. In addition to being chemically robust, a good photocatalyst should exhibit other important properties such as (i) a high surface-to-mass ratio, (ii) efficient absorption of light in a useful range of frequencies, and (iii) efficient separation and transportation of photogenerated charge carriers. To achieve the first feature, the surface-area effects of metal oxides are tailored by preparing nanoparticles using suitable synthetic methods (such as hydrothermal/solvothermal and sol–gel syntheses) or, alternatively, as a result of deposition onto a nanostructured film. The importance of the second property is related to the fact that exploiting a light source in the visible region of the electromagnetic spectrum would make the management of wastewater treatment plants simpler and more economical. Several strategies are adopted to extend the optical response of photocatalytic materials to the visible region: among others, doping with metal and/or non-metal elements, surface functionalization, composites, and junction coupling [63,64]. In particular, nano-heterojunctions could enhance visible light absorption and induce band bending and the formation of an internal electric field [65]. Consequently, owing to specific band alignment and spatial charge separation, the photoactivity is improved.

3.1. Photocatalytic Oxidation of As(III) Using Modified TiO₂

Titanium dioxide (TiO₂) is an n-type semiconductor. It has been widely studied as a photocatalyst in the oxidative degradation of organic and inorganic water pollutants [66,67]. The interaction of TiO₂ with UV radiation generates electron–hole pairs, which activate surface reactive processes [59,61]. Among the three common crystalline polymorphs of TiO₂ (rutile, anatase, and brookite) the anatase phase exhibits higher photocatalytic behavior for oxidation processes. The relatively wide bandgap of TiO₂ anatase, E_g = 3.2 eV corresponding to a threshold of 390 nm, hinders the exploitation of solar light in photocatalytic reactions. Moreover, the high recombination probability of the photogenerated electron–hole pairs results in low quantum efficiency of photocatalyzed reactions, which is a feature present in all semiconductor photocatalysts. Different approaches have been developed to modify the process involving light harvesting and excitation to increase visible light absorption of TiO₂ and to maximize electron–hole separation. A widespread strategy consists of doping TiO₂ with metal and/or non-metal elements; this is achieved through the incorporation of different elements into the TiO₂ crystalline lattice, resulting in modified electronic properties. No previous literature discussing the photocatalytic oxidation of As(III) using TiO₂ modified by a physical method (such as ion implantation) was found. Table 2 lists the examples of visible-light-activated chemically modified TiO₂ covered in this review. A brief description of their performance for As(III) oxidation is also provided. Fe-doped TiO₂ [68] outperformed other types of doped TiO₂ such as N-doped TiO₂ fiber [69], N–S co-doped TiO₂ [70], and Er-doped TiO₂ [71]. It is worth noting that in ref. [68], a very high (3 g/L) concentration of Fe-doped TiO₂ was used to remove 100% of As(III) in 30 min. TiO₂ doped with Er³⁺ 2.0 mol% and reduced graphene oxide (rGO) 10.0 mol% degrades 100% of arsanilic acid (p-ASA) in 50 min under visible light, and the resulting arsenic (mainly As(V)) is rapidly removed from the solution owing to the strong adsorption of rGO to inorganic arsenic [71].

Some studies that were carried out with the aim of investigating the mechanism of photocatalytic oxidation of As(III) in the presence of TiO₂ were performed under visible-light irradiation and pH = 3 [72,73]. The visible-light-induced oxidation on dye-sensitized Pt-TiO₂ (~80% in 60 min) [73] is as fast as the UV photooxidation on Pt-loaded TiO₂. Nanostructured TiO₂ nanotubes prepared by potentiostatic anodization of a titanium foil, combined with Pt-loading, are capable of harvesting visible light because of localized surface plasmon resonance (SPR) of Pt nanoparticles. Qin et al. reported that Pt nanoparticles on TiO₂ nanotubes induce the oxidation of 83% of As(III) in 280 min [74].

Table 2. As(III) oxidation to As(V) over modified TiO₂ photocatalysts under visible light.

Photocatalyst	pH	Light Source	As(III) (μM)	Catalyst (g/L)	As(III) Oxidized (%)	Time (min)	Ref.
(Ru″L3) dye-sensitized Pt-loaded TiO2	3	Xe 300 W λ > 420 nm	500	0.5	~80	60	[73]
4-chlorophenol/TiO2 system	3	Xe 300 W λ > 420 nm	200	0.5	~90	300	[72]
Pt-loaded TiO2 nanotube	7	300 W halogen λ > 420 nm	45.4	foil	83	280	[74]
N-doped TiO2 fiber	7	Xe 300 W λ > 420 nm	133	0.5	100	90	[69]
Fe-doped TiO2	n.a.	LED 400–600 nm	80	3	100	30	[68]
N–S co-doped TiO2	9	LED 650 W/m2 430–650 nm	46 (p-ASA)	1	98	300	[70]
2 mol% Er-doped TiO2	n.a.	Xe 500 W λ > 400 nm	23 (p-ASA)	0.1	~70	90	[71]
Er3+–rGO co-doped TiO2	n.a.	Xe 500 W λ > 400 nm	23 (p-ASA)	0.1	100	50	[71]

Table 2. As(III) oxidation to As(V) over modified TiO₂ photocatalysts under visible light.

Photocatalyst	pH	Light Source	As(III) (μM)	Catalyst (g/L)	As(III) Oxidized (%)	Time (min)	Ref.
(Ru″L3) dye-sensitized Pt-loaded TiO2	3	Xe 300 W λ > 420 nm	500	0.5	~80	60	[73]
4-chlorophenol/TiO2 system	3	Xe 300 W λ > 420 nm	200	0.5	~90	300	[72]
Pt-loaded TiO2 nanotube	7	300 W halogen λ > 420 nm	45.4	foil	83	280	[74]
N-doped TiO2 fiber	7	Xe 300 W λ > 420 nm	133	0.5	100	90	[69]
Fe-doped TiO2	n.a.	LED 400–600 nm	80	3	100	30	[68]
N–S co-doped TiO2	9	LED 650 W/m2 430–650 nm	46 (p-ASA)	1	98	300	[70]
2 mol% Er-doped TiO2	n.a.	Xe 500 W λ > 400 nm	23 (p-ASA)	0.1	~70	90	[71]
Er3+–rGO co-doped TiO2	n.a.	Xe 500 W λ > 400 nm	23 (p-ASA)	0.1	100	50	[71]

As mentioned above, an adsorption step takes place before the reaction step. There is an attractive interaction between particles with opposite electric charges. The solution pH significantly affects the charge of the surface particles. The pH of zero-point charge (zpc) is the value at which the material surface is neutral (uncharged). Below this value, the photocatalyst is positively charged and attracts negative species. At pH values above the zpc, the photocatalyst surface is negatively charged and attracts positive species. Although the large band gap of TiO₂ limits the utilization in visible light, many studies examined the arsenic adsorption performance of Degussa P25 TiO₂. The results showed that at pH 4, adsorption for As(V) is much higher than for As(III) because of the negative and neutral speciation of As(V) and As(III) (see Figure 1), respectively. As(V) desorbs more effectively as pH increases, and at pH 9, the adsorption of As(III) is higher than As(V) [75]. The maximum adsorption capacity for As(III) (89 μM) is 2.4 mg/gTiO₂, and that for As(V) is 9.7 mg/gTiO₂ at pH 6.3 [76].

Colored TiO_2−x shows photoactivity under visible light. As shown in Figure 3, As(III) adsorption diminished above zpc (at pH 6.3) under visible light owing to the conversion of As(III) to As(V) by photocatalytic oxidation (a conversion rate of about 2.85 mgAs/g h) [77].

3.2. Photocatalytic Oxidation of As(III) Using TiO₂-Based Nanocomposites

The design of TiO₂-based nanocomposites is the most frequently used strategy to increase the efficiency of photocatalytic reaction under visible-light irradiation and, at the same time, to enhance the pollutant removal capacity. Interfacial interactions between the phases forming a nanocomposite make the physical properties different from those of its individual components. WO₃ absorbs solar radiation better than TiO₂ owing to its smaller band-gap energy of about 2.6 eV [78]. When TiO₂ is coupled with WO₃ (TiO₂/WO₃ nanocomposite), the photo-generated electrons of WO₃ are transferred to the TiO₂ conduction band, thus generating holes in the WO₃ valence band [79]. TiO₂/WO₃ composite oxidizes about 75% of As(III) in 60 min, while V₂O₅@TiO₂ composite with a core–shell sphere microstructure oxidizes 92% of As(III) in 80 min [80]. Hydrogenated black TiO₂ (HBTiO₂, anatase) and rutile-based inorganic hollow microspheres (RBIHM) oxidize 70.3% of arsenite in 30 min [81]. Coupling HBTiO₂/RBIHM with MoS₂ results in superior As(III) photooxidation of 96.6% in 30 min under visible light [81]. Repeated experiments of HBTiO₂/RBIHM-MoS₂ maintain 91.4% of As(III) photooxidation after 5 cycles.

Ternary composites incorporating (i) TiO₂, (ii) a conductive polymer to retard electron–hole recombination, and (iii) an excellent adsorbent material for the removal of As(V) were investigated. Examples include γ-Fe₂O₃/polyaniline/TiO₂ [82], MWCNT/polyaniline/TiO₂ [83], Fe₃O₄/polyparaphenylene diamine/TiO₂ [84], and γ-Fe₂O₃/polythiophene/TiO₂ [85]. As shown in

Table 3. As(III) oxidation to As(V) over TiO₂-based nanocomposite photocatalysts under visible light. @ indicates core–shell structure.

Photocatalyst	pH	Light Source	As(III) (μM)	Catalyst (g/L)	As(III) Oxidized (%)	Time (min)	Ref.
TiO2/WO3	3	230 W EB-1750 Epson	200	0.4	~75	60	[79]
V2O5@TiO2	4	Xe 300 W λ > 420 nm	26.7	0.8	92	80	[80]
HBTiO2/RBIHM *	10	DH-2000 deuterium tungsten halogen 1060 W/m2 λ > 420 nm	0.76	0.5	70.3	30	[81]
HBTiO2/RBIHM-MoS2 *	10	DH-2000 deuterium tungsten halogen 1060 W/m2 λ > 420 nm	0.76	0.5	96.6	30	[81]
γ-Fe2O3@PANI@TiO2	8.5–10	n.a.	67	1.0	~89	300	[82]
MWCNT/PANI/TiO2	5	Xe 500 W λ > 380 nm	267	0.5	~52	300	[83]
Fe3O4@PpPDA@TiO2 ‡	6	Xe 500 W λ > 380 nm	267	0.5	~78	300	[84]
γ-Fe2O3/PTh/TiO2 #	6	Xe 500 W Visible light	26.7	0.5	99	300	[85]
B-doped TiO2/g-C3N4	5	Xe 500 W Visible light	26.7	0.5	92	300	[86]

* HBTiO₂ = hydrogenated black TiO₂. RBIHM = rutile-based inorganic hollow microspheres. ^‡ PpPDA = polyparaphenylene diamine. ^# PTh = polythiophene. n.a. = not available.

, the best results were obtained by γ-Fe₂O₃/polythiophene/TiO₂, which oxidizes 99% of As(III) in 300 min and simultaneously adsorbs 90% of As(V) [85]. Recently, graphitic carbon nitride (g-C₃N₄,), which absorbs visible light (band-gap energy of 2.7 eV), but is a poor arsenic sorbent, was coupled with B-doped TiO₂, showing a removal efficiency of 92% of As(III) in 300 min [86]. The literature reports the effect of oxyanions that are frequently present in water on photocatalysis reaction rates through competitive adsorption and surface-site blocking [87,88]. In particular, $N{O}_3^{2-}$and $C{O}_3^{2-}$ have a weak inhibitory effect on the adsorption ability of γ-Fe₂O₃@polyaniline@TiO₂, while the removal capacity of As decreased by 17% in the presence of $S{O}_4^{2-}$ and by 45% in the presence of $P{O}_4^{3-}$ [82].

3.3. Photocatalytic Oxidation of As(III) Using WO₃, ZnO, and Bismuth Oxides

Although the semiconductor oxide most widely used as photocatalyst is TiO₂ because of its low cost, stability, and reactivity, it is effective only under UV-light irradiation, which limits its utilization with solar light and increases costs and management problems. In addition to studying many types of modifications to widen the absorption range of TiO₂, other semiconductor oxides that can absorb light in the visible range were studied.

Tungsten (VI) trioxide was studied as a photocatalyst because of its effective oxidation capacity of valence band (VB) holes, strong absorption of solar radiation, small band gap (2.4–2.8 eV), high oxidation power of valence band (VB) holes (+3.1−3.2 V_NHE), and nontoxicity [89].

The photocatalytic oxidation of As(III) to As(V) on WO₃ primarily proceeds through two-hole transfer, while hydroxyl radical (HO^•) and hydrogen peroxide (H₂O₂) are only marginally involved in the oxidation of As(III) to As(V) [90]. However, the rapid recombination of photo-generated electron–hole pairs reduces the catalytic activity and calls for modifications aimed at increasing the efficiency of the electron–hole separation. As shown in

Table 4. As(III) oxidation to As(V) over WO₃-based photocatalysts under visible light.

Photocatalyst	pH	Light Source	As(III) (μM)	Catalyst (g/L)	As(III) Oxidized (%)	Time (min)	Ref.
Pt/WO3	3	300 W Xe arc lamp λ > 420 nm	100	0.5	100	30	[89]
WO3	3	300-W Xe arc lamp λ > 420 nm	100	0.5	100	150	[89]
(10 mg L−1 Cu2+)/WO3	3	visible LED light λmax ~ 405 nm	133	0.33	95	360	[91]
(10 mg L−1 Cu2+)/WO3	7	visible LED light λmax ~ 405 nm	133	0.33	90	360	[91]
1 wt% CuO-WO3	10	visible LED light λmax ~ 405 nm	133	0.33	80	360	[91]
1 wt% CuO-WO3	7	visible LED light λmax ~ 405 nm	133	0.33	75	360	[91]

, platinization is one such modification [89], the purpose of which is to enable O₂ to serve as an electron acceptor despite the insufficient reduction potential of the conduction band electrons in WO₃. The catalytic system was tested on several types of substrate; As(III) was in all cases successfully converted to As(V), but the oxidation is much faster with the platinized catalyst. Another way to increase the efficiency of WO₃ is to add Cu²⁺ ions or CuO as a co-catalyst [91]. It was established that the presence of Cu²⁺ increases the arseniate oxidation rate by 3.6 times, while the rate for CuO increases by 2.3 times compared with pure WO₃. In addition, pH plays an important role in increasing the oxidation rate of As(III): in the case of Cu²⁺–WO₃, oxidation takes place at an acidic pH and is much faster than with CuO–WO₃ at basic pH [91].

Zinc (II) oxide is another oxide that has attracted the attention of researchers who are studying new, economically competitive methods for the oxidation of arsenic (III). This compound shows peculiar characteristics, such as a direct and wide band gap in the near-UV spectral region, strong oxidation ability, good photocatalytic properties, and a high free-exciton binding energy [92]. ZnO is also more affordable than TiO₂. To make it active even under visible-light irradiation and to improve its photocatalytic performance, the literature suggests different solutions, including the introduction of different types of metal dopants into ZnO (

Table 5. As(III) oxidation to As(V) over ZnO-based photocatalysts under visible light.

Photocatalyst	pH	Light Source	As(III) (μM)	Catalyst (g/L)	As(III) Oxidized (%)	Time (min)	Ref.
1.08 mol% Cu-doped ZnO	7.2	LED (10 W; light intensity 32 W/cm2) 400–600 nm	67	3	100	120	[92]
1.08 mol Cu-doped ZnO% supported on polystyrene pellets	7.2	LED (86 W; light intensity 78 mW/cm2) 400–600 nm	67	n.a.	100	120	[93]

). Cu is a cationic dopant that leads to a lower band gap energy value compared with undoped ZnO. The tests showed the complete conversion of As(III) to As(V) in the presence of Cu-doped ZnO under visible light [93], which did not happen in the presence of pure ZnO under the same conditions. The same result was also obtained using solar-simulated radiation. In 2021, to develop a catalyst that was simple to remove, Cu-doped ZnO was supported on polystyrene pellets [93]. The excellent performance of the photocatalyst, as proved in a previous paper [93], was also provided in this case. It was proved that the high photo-catalytic activity was preserved after several reuse cycles, and there was evidence that no significant release of photocatalyst particles from PS pellets occurred.

Bismuth oxyhalides BiOX (X = Cl, Br, I) are promising photocatalysts under UV or visible-light illumination due to their layered structure, high activity, and stability. In several papers, BiOI, bismuth oxyiodide, as such or modified, was used as a photocatalyst for As(III) oxidation. In [94], the efficiency of BiOI was related to its morphology, which was in turn dependent on the viscosity of the solvent used in its synthesis. With the highest viscosity, largest specific surface area, and thinnest nanosheets with exposed (0 0 1) surfaces, it could efficiently absorb visible light, subsequently generate a large number of carriers, and then rapidly move to the nanosheet surface in the role of a static electric field, thereby yielding the highest photocatalytic efficiency. The proposed mechanism proceeds through the excitation of the BiOI nanoparticles under >420 nm illumination to excite electrons to the CB; these could reduce $O_2^{ }$ to $O_2^{•-},$ while the photogenerated holes cannot oxidize the hydroxyl groups due to unfavorable potential. Instead, they directly oxidize the As(III) adsorbed on the surface of BiOI.

Modifications included doping with La to obtain Bi_0.9La_0.1OI [95] to facilitate migration and hinder the recombination of photogenerated charge carriers. The contribution of La is due to free La³⁺ ions on the surfaces of BiOI nanosheets, which can form doping energy levels, trapping the photo-induced e⁻ below the CB edge of Bi_0.9La_0.1OI and accelerating the formation of $O_2^{•-}$. The in situ introduction of La into the lattice structure of BiOI may provide more active sites for the formation of oxygen vacancies; these increase the amount of adsorbed oxygen and consequently induce the generation of superoxide radicals, suppressing the recombination of photogenerated electron–hole pairs, while the h+ may migrate to the surface of Bi_0.9La_0.1OI.

ZrO₂-modified BiOCl_0.5I_0.5 composites (ZBCI) [96] can not only oxidize As(III) into As(V) with visible-light irradiation but can also effectively capture the generated As(V), leading to negligible residual As(III) or As(V) in aqueous solutions after a 90 min treatment. In the fabricated composites, ZrO₂ acted as the main adsorption site, while BiOCl_0.5I_0.5 served as the primary photocatalysis center. Because of the hetero-structure of ZBCI, e- generated by BiOCl_0.5I_0.5 would be transferred to ZrO₂ and an inhibited e⁻-h⁺ recombination rate, contributing to the improved photocatalytic efficiency. ZBCI can effectively remove As(III) over a broad pH range from 3 to 11. The high-resolution XPS spectra of ZBCI after the reaction showed that the As present on ZBCI surfaces was in the form of As(V), which further confirmed that As(III) was completely oxidized by photocatalysis.

BiOI@Fe₂O₃ core–shell nanoparticles [97] provided efficient removal of As(III) via a simultaneous photocatalytic oxidation–adsorption process with an As removal efficiency of 99.8% (3.99 mg/g) within 180 min. Such BiOI@Fe₂O₃ also demonstrated good As removal stability (3.8–4.0 mg/g) over a wide pH range of 2–8. Owing to the magnetic property, the BiOI@Fe₂O₃ nanocomposites can be rapidly recovered from the solution under an external magnetic field. Bi_2.15WO₆ (BWO) [98] can be used to treat wastewater containing both Mn(II) and As(III), as it can oxidize Mn(II) to amorphous manganese oxide (MnO_x) under visible light. MnO_x generated in this way can further transform As(III) to As(V). Conversion of As(III) to As(V) can be obtained either by the BWO-Mn(II) system under visible light or by the BWO-MnO_x system in the dark, with the latter system providing slightly better results. BWO can remove Mn(II) and As(III) via adsorption and/or oxidation processes. As(III) was gradually converted into As(V) as the reaction progressed, and the oxidation was basically complete after 24 h. The remaining As(III) was only 0.021 mg L⁻¹ after the reaction, and the removal percentage was as high as 97.9%. The adsorbed As in the BWO-MnO_x system was mainly As(V). From the results reported in refs [97,98], although they are interesting, it is not possible to extrapolate the contribution of photooxidation versus absorption. To avoid inconsistencies with the other data in

Table 6. As(III) oxidation to As(V) over bismuth oxide photocatalysts under visible light.

Photocatalyst	pH	Light Source	As(III) (μM)	Catalyst (g/L)	As(III) Oxidized (%)	Time (min)	Ref.
Bi2Sn2O7	~7.0	300 W halogen lamp 420−850 nm	27	0.8	83	60	[99]
BiOI	-	300 W Xe lamp λ > 420 nm	67	0.4	90	30	[94]
ZrO2-coated BiOCl0.5I0.5	11	Xe lamp λ > 400 nm	67	0.25	90	60	[96]
ZrO2-coated BiOCl0.5I0.5	7	Xe lamp λ > 400 nm	67	0.25	95	90	[96]
Bi1-xLaxOI	5.7	500 W Xe lamp λ > 420 nm	133	1	90	150	[95]

(which refer to the oxidized amount), these results are omitted.

Nanocrystalline Bi₂Sn₂O₇, with a band gap of 2.88 eV, exhibits a high photocatalytic activity in the oxidation of As(III) (up to 96.8%) under visible-light irradiation [99]. As(III) species can adsorb readily on the Bi₂Sn₂O₇ surface, which is beneficial for the subsequent oxidation of As(III) via photocatalyst. The total removal ratio of As(III) is up to 96.8% after 60 min of illumination.

3.4. Photocatalytic Oxidation of As(III) Using Graphitic Carbon Nitride Based Photocatalysts

Graphitic carbon nitride, often referred to as g-C₃N₄, recently gained significant attention owing to its distinct physicochemical characteristics. This layered structure, which is reminiscent of graphene, gained prominence for its stability, thermal resistance, and narrow band gap (around 2.7 eV). g-C₃N₄ comprises layers of bi-dimensional π-conjugated polymer structures, consisting of S-triazine or S-heptazine units and tertiary amine linkages. The appeal of g-C₃N₄ also stems from its environmental friendliness, straightforward preparation process, and cost-effectiveness. Nonetheless, pristine carbon nitride (pCN) exhibits limitations in terms of electronic conductivity, visible light absorption, and available surface active sites. These shortcomings were effectively addressed through diverse strategies, including the creation of heterojunctions, the introduction of heteroatoms, and nanostructural engineering. The material’s photocatalytic efficacy under visible light sparked interest in applications involving hydrogen generation through water splitting and degradation of pollutants at room temperature.

Table 6 reports the entries of the results obtained using g-C₃N₄ in various forms to photocatalytically oxidize As(III) species.

Although g-C₃N₄ by itself can act as a photocatalyst, its performance in the oxidation of As(III) to As(V) species is poor in many conditions due to the rapid recombination of the photogenerated electron–hole pairs. A conversion as low as 5% was reported by Kim et al. [100], while other authors obtained conversions up to a maximum of about 65% [101] depending on the experimental conditions. Variously modified types of g-C₃N₄ were usually able to improve the lifetime of the photogenerated species, reaching oxidation percentages of 41% [100], 67% [102], and 90% [101], and even complete oxidation in some cases [103,104].

For example, the oxidation ability was greatly enhanced in composite photocatalysts prepared by Wang et al. [103]: they produced a composite using bentonite clay by mixing and firing the two components at 450–550 °C, enhancing the oxidation capabilities of graphitic carbon nitride up to complete oxidation in 180 min of illumination with visible light, with only small changes in efficiencies by changing the solution pH from 3 to 8.5. The authors speculated that the improvement was caused by a change in the g-C₃N₄ structure in the composite, suggesting a negligible contribution of bentonite in the oxidation. Indeed, the oxidation ability was not linearly sensitive to the amount of bentonite (the 10% bentonite composite was the optimum composition, whereas the 5% and 20% composites had poorer performance). The structural change was also supported by the change in oxidation ability depending on the composite calcination temperature. Furthermore, a slight decrease in efficiency was observed when increasing the As(III) initial concentration.

The addition of amorphous iron to g-C₃N₄ was reported by Kim et al. [100] as a method of improving the removal ability of arsenic species in water. The photocatalysts were obtained in a single-step synthesis starting from melamine and iron and were compared with a mixture of g-C₃N₄ and crystalline hematite. The authors demonstrated a synergy between the oxidation of As(III) by the graphitic carbon nitride under illumination (both in UV and visible light) and the absorption of both As(III) and As(V) by the amorphous iron oxide species in the structure of the composite. Visible light seems to have a role in the process; nevertheless, the oxidation efficiency seems quite small (only about 40% of the absorbed As is in the form of As(V)), and the reported results seems mainly due to the absorption process; since the overall As removal is 41%, the oxidation efficiency can be calculated to be about 16%.

In 2017, Sun et al. [104] developed a α-Fe₂O₃/g-C₃N₄ heterojunction, exploiting the relative positions of the conduction and valence bands of α-Fe₂O₃ and g-C₃N₄ to allow the electrons in the CB of g-C₃N₄ to transfer to the CB of α-Fe₂O₃, thus enhancing the lifetime of the photogenerated electron–hole pairs and inhibiting their rapid recombination. Indeed, the oxidation ability of the composite largely exceeded the ability of g-C₃N₄ alone (38%) and that of α-Fe₂O₃ (21%). They found the optimum amount of α-Fe₂O₃ in the composite (8%) that allowed the oxidization of about 94% of the starting As(III) at neutral pH in 500 min of irradiation.

Wang et al. [101] used a different approach, modifying the g-C₃N₄ structure with pyromellitic diimide (PDI) doping. The modification gave rise to a high oxidation ability, so to produce high amounts of H₂O₂ in water. The photocatalyst was then applied to the simultaneous photooxidation of As(III) and photoreduction of Cr(VI) in high concentrations for the first time, so as to simulate Acid Mine Drainage composition. Nonetheless, we only analyze the results of As oxidation. At pH = 4 and after 120 min of visible light irradiation, the oxidation efficiency of As(III) reached 65% for the pure g-C₃N₄, which increased to a value of about 90% under the same conditions for the g-C₃N₄/PDI composite.

Later, Lei et al. [102] applied a different approach for a similar purpose: this time a direct Z-scheme photocatalyst was prepared by combining a magnetic semiconductor oxide ZnFe₂O₄ with g-C₃N₄ and, it was tested using simultaneous As(III) oxidation and Cr(VI) reduction. The as-prepared catalyst yielded a 67% oxidation of As(III) in two hours of reaction, starting from a mixture of both As(III) and Cr(VI). The result was enhanced by the addition of oxalate ions in the reaction mixture, obtaining virtually complete oxidation of As(III) in the same amount of time at pH = 5.

The same authors also employed a different approach using similar materials (core–shell structures of ZnFe₂O₄ supported on polyaniline (PANI) [105] that, for the sake of clarity, are analyzed in the following paragraph and reported in

Table 7. As(III) oxidation to As(V) over g-C₃N₄-based photocatalysts.

Photocatalyst	pH	Light Source	As(III) (μM)	Catalyst (g/L)	As(III) Oxidized (%)	Time (min)	Ref.
Bentonite/g-C3N4	3 ÷ 8.5	300 W Xe lamp	133	0.2	100	180	[103]
g-C3N4	3 ÷ 8.5	300 W Xe lamp	133	0.2	40	180	[103]
Fe-g-C3N4	7	400–680 nm	334	5	~16 *	300	[100]
g-C3N4	7	400–680 nm	334	5	<5	300	[100]
α-Fe2O3/g-C3N4	7	500 W Xe lamp	67	0.5	~94	500	[104]
g-C3N4/PDI **	4	300 W Xe lamp λ > 420 nm	100 §	1	90	120	[101]
g-C3N4	4	300 W Xe lamp λ > 420 nm	100 §	1	65	120	[101]
ZnFe2O4/g-C3N4	5	n.a.	60 §	1	67	120	[102]
ZnFe2O4/g-C3N4 + oxalate	5	n.a.	60 §	1	100	120	[102]

* calculated as 37% of As(V) speciated on the catalyst that absorbed 41% of the starting As(III). ** pyromellitic diimide (PDI). ^§ plus an equal amount of Cr(VI). n.a.= not available.

, together with other core–shell photocatalysts.

3.5. Photocatalytic Oxidation of As(III) Using Core–Shell Structure Photocatalysts

The approach of Lei et al. [105] involved the use of sulfate $\left(S{O}_4^{•-}\right)$ and hydroxide (${HO}^{•}$) radicals, obtained by activation of sulfite radicals ($S{O}_3^{•-})$ via ZnFe₂O₄@PANI photocatalyst as the active species for As(III) oxidation. The efficacy of the photocatalyst was attributed by the authors to the core–shell structure of the catalyst. Indeed, the effect of PANI covering the ZnFe₂O₄ surface was emphasized by the increase in oxidation efficiency that spans from 60% with bare ZnFe₂O₄ to complete oxidation using ZnFe₂O₄@PANI in just 60 min. Moreover, the method is fairly independent of the solution pH in the range of 3–10, and the magnetic properties of the semiconductor oxide allow for easy recovery. A minor leaching of iron ions into the solution was reported, but less than observed for the bare ZnFe₂O₄.

Other core–shell structure catalysts that gained attention are silver-based compounds, especially silver halides, owing to their ability to absorb visible light and efficiently attain charge separation and charge transfer. In turn, core–shell structures allow for easy tunability of the catalyst properties.

Quin et al. [106] reported on the preparation and testing of Ag@AgCl core–shell nanowires. With an optimized composition, they obtained an oxidation efficiency of 40% in acidic conditions (pH = 3), 76% at neutral and 83% in basic conditions (pH = 10), after two hours of illumination (

Table 8. As(III) oxidation to As(V) over core–shell structured photocatalysts.

Photocatalyst	pH	Light Source	As(III) (μM)	Catalyst (g/L)	As(III) Oxidized (%)	Time (min)	Ref.
ZnFe2O4	3–10	n.a.	667	0.05	63	60	[105]
ZnFe2O4@PANI * (+sulfite)	3–10	n.a.	667	0.05	~100	60	[105]
Ag@AgCl	3/7/10	300 W halogen lamp λ > 420 nm	27	0.3	40/76/83	120	[106]
Ag@AgCl	7	300 W Xenon lamp 420 < λ < 760 nm	27	n.a.	~39	180	[107]
Fe-Ag@AgCl	7	300 W Xenon lamp 420 < λ < 760 nm	27	n.a.	~92	180	[107]

* PANI: polyaniline. n.a.: not available.

). The authors also tested the photocatalysts for several cycles, unfortunately showing a washout of the catalyst that decreased the efficiency to about 22% after 20 cycles.

The same authors also modified the pristine Ag@AgCl by incorporating iron [107] in a ratio of Ag:Fe = 1:1, which allowed for the reduction of the band gap of the photocatalyst from 3.25 eV to 3.04. They used the nanowire photocatalysts that were originally developed for the inactivation of Escherichia coli (E. coli) bacteria, in the concurrent oxidation of As(III) and inactivation of E.coli. The authors reported that the presence of the bacteria did not affect the efficiency of the catalyst on the photooxidation of As(III) until a very high amount (10⁸ cfu/mL). The iron-modified photocatalyst resulted in an As(III) oxidation percentage of about 92% after 180 min of reaction under visible-light irradiation at neutral pH, increasing the performance of the iron-free photocatalyst, which, under the same conditions, only attained 39% oxidation.

4. Conclusions and Perspectives

In this review, the most advanced photocatalytic materials for arsenic removal are described with a focus on semiconductor catalysts activated by visible light, including the fundamental aspects of their design and applications. In addition, the importance of coupling materials for excellent oxidation of As(III) to As(V) by photocatalysis with highly adsorbent materials is highlighted. Figure 4 and Figure 5 summarize the conclusions of the review in two radar charts. Figure 4 groups the best photocatalysts regardless of the working pH, while Figure 5 reports only the catalysts for which results were obtained at pH ≈ 7. The radar charts allow the comparison of an arbitrary number of variables. The variables compared in these graphs are (clockwise, starting from the top): (i) [As(III)]: the molar concentration of As(III) used in the experiments in μmol/L(μM). The lower the better, as the concentrations usually found in groundwater are very low and the limit for arsenic in drinking water is around 10 μg/L (0.1 μM) [4]. (ii) [Catalyst]: the concentration of catalyst used (in mg/L). The lower the better (scale from 0 to 6 mg/L). (iii) Amount of As(III) oxidized to As(V) (in %): this is the efficiency of the photocatalyst in oxidizing As(III). The higher the better. (iv) Time (in minutes) required to attain the reported % of As(III) oxidation. The lower the better. (v) pH: this is the specific pH condition at which the experiments were performed. The closer the pH is to 7 the better, the farther the pH deviates from 7 (either towards acidic or basic conditions), the worse.

The scales of each variable are designed so that the better the catalyst, the larger the filled area in the graph. When data are unavailable, the parameter is assigned the worst possible value.

Currently, the most efficient performance under 30 min of visible-light irradiation is offered by the following photocatalysts (Figure 4): (1) Fe-doped TiO₂ with 100% As(III) oxidation [68], HBTiO₂/RBIHM-MoS₂ with 96.6% As(III) oxidation [81], Pt/WO₃ with 100% As(III) oxidation [89], and BiOI with 90% As(III) oxidation [94]. It is worth noting that in these studies, the pH of the starting As(III) solution is 3 (Pt/WO₃) or 10 (HBTiO₂/RBIHM-MoS₂) or is not available (Fe-doped TiO₂, BiOI). For a practical application of the oxidation of As(III), it is necessary to work with an arsenic solution at a pH of approximately 7. Among the studies carried out at pH ≈ 7, the most interesting are the following (Figure 5): N-doped TiO₂ fiber with 100% As(III) oxidation under 90 min visible-light irradiation [69]; 1.08 mol% Cu-doped ZnO with 100% As(III) oxidation (120 min) [92]; Bi₂Sn₂O₇ with 83% As(III) oxidation (60 min) [99]; and Fe-Ag@AgCl with ~92% As(III) oxidation (180 min) [107]. These results show that long reaction time remains a major challenge in achieving high As oxidation.

For photocatalysis technology to find use in small low-resource communities, the photocatalyst must be robust and inexpensive. However, this cannot be the only criterion in the selection of a photocatalyst, as the effect of oxyanions, which are frequently present in water, on photocatalysis reaction rate through competitive adsorption and surface-site blocking can hinder As(III) oxidation.