The most commonly studied and generally accepted mechanism of photocatalysis involves the use of titanium dioxide (TiO₂) in the oxidation of organic pollutants and inactivation of pathogens. Figure 1 shows the mechanism of pollutant oxidation by TiO₂.

When photons of enough energy reach the TiO₂ surface, an electron is ejected from the valence band (VB) to the conduction band (CB) of the semiconductor, as described by the following reactions:

TiO 2 + hv → e CB − + h VB + (1)

e − + O 2 → O 2 • − , (2)

h + + H 2 O → OH • + H + (3)

h + ( or   OH • ) + organic substrate → oxidized organic substrate (4)

The difference in energy between the VB and the CB is known as the bandgap, a region within which an electron cannot remain stable. The electron jump caused by the influx of additional energy leads to the formation of an electron-hole pair (e⁻ and h⁺). The ability of an electron to make this jump at an optimum rate is unique to semiconductors and this property is exploited in photocatalysis.

Once the electron-hole pair forms, these photo-generated charges will migrate onto the surface of the photocatalyst and undergo a variety of complex reactions to produce reactive oxidative species (ROS) such as the hydroxyl radical ( OH • ), hydrogen peroxide (H₂O₂), and the superoxide ion ( O 2 • − ), which are capable of both oxidizing dissolved organics and inactivating pathogens [37].

While effective, conventional TiO₂ is not capable of VLA photocatalysis without a change in its morphology. The limitation of this process with TiO₂ lies in the need for an energy input higher than 3.2 eV, equivalent to a wavelength shorter than 390 nm (high frequency waves such as UVB and UVC) to effectively excite and eject electrons, rendering the process energy-intensive. If solar light were to be used, the process would be inefficient since the Earth surface receives 8% UV rays, of which 0.5%, 4.5% and 95% correspond to the UVC, UVB and UVA spectrum, respectively [38]. To overcome this limitation, modifications to the TiO₂ lattice, consisting of doping its structure with either a metal or a non-metal, are necessary to reduce the bandgap width, enable electron-hole pair formation by visible light energy, and prevent recombination of the formed electron-hole pair.

Typically, when a photocatalyst is used for disinfection, the electron-hole pair forms from the application of photons of enough energy. Some pairs become unstable and the electron drops from the CB back into the VB leading to recombination and the release of energy as heat [39] [40]. The pairs that remain stable migrate to the surface of the photocatalyst.

The OH − ions in the aqueous solution that are adsorbed onto the TiO₂ surface react with the migrated h⁺ to form the adsorbed hydroxyl ion ( OH absorbed • ) [41] [42] [43]. If other electron donors (reductants) are present in the solution, the h⁺ may also directly gain electrons from these reductants and become oxidized. Therefore, the h⁺ acts as an electron acceptor, directly oxidizing the organic substance that donates said electron. If sufficient reductants are present, the OH absorbed • is readily released into the bulk solution to form the OH bulk • , which plays an important role in the inactivation/destruction of microorganisms [44]. Also, two OH bulk • radicals can combine with each other in solution to form H₂O₂, which is also an effective disinfectant [45] [46] [47].

On the other hand, the surface migrated e⁻ react with electron acceptors (oxidants). Generally, they react with dissolved oxygen in the water to form the superoxide ion ( O 2 • − ) [48] [49]. Electrons may also react with the H₂O₂ formed via the interactions mentioned above to produce OH bulk • . The formation of ( O 2 • − ) helps disinfection by gaining the electron from the e⁻ and h⁺ pair, delaying and even preventing recombination at the below surface level, as electrons that have moved up the surface are being scavenged. The superoxide ion is a strong oxidant capable of inactivating pathogens [50].

As previously mentioned, generated ROS ( OH • , H₂O₂, O 2 • − ) attack microbes in water, which results in their destruction or inactivation [51]. The proposed action mechanism suggests that these ROS begin by systematically destabilizing or disorienting the outer membrane or cell wall of the microorganism. Then, they proceed to penetrate it, destroying the inner cell membrane and allowing the contents of the cell to leak out, thus making it unable to replicate again and effectively killing or inactivating the microorganism. It is speculated that the inactivation of a bacteria may also occur via h⁺ before it is trapped either within the semiconductor or at its surface [52]. However, ROS are thought to be the primary “killer”, as indicated by the leaking potassium ion from the microorganism membrane or the destruction of its cell structure observed by others [51].

The mechanism described above, while effective, relies on the prevention of rapid recombination of the electron-hole pair. The introduction of a constant stream of oxygen into the system to act as electron scavenger, the addition or in-situ generation of H₂O₂ to also act as electron acceptors, and the doping of the photocatalyst with noble metals, are among the most effective techniques being used to impede or delay pair recombination and, at the same time, increase photon absorption in the visible light region. SPR exhibited by noble metal NPs also play a very important role in this equation.

The concept of plasmon resonance in noble metals is well documented [24] - [33]. It is prominent in noble metals NPs as they can utilize visible light energy due to this phenomenon. The SPR of noble metal NPs is the process by which the conducting electrons on the NPs undergo a collective oscillation or excitation stimulated by the oscillating electric field of incoming light rays [53]. As shown in Figure 2, the oscillating charges expand an electrical field close to the surface. When the resonance condition of the noble metal NPs is met by the frequency of the incident energy source, the resultant associated energy absorption leads to the SPR effect [54].

Two conditions must be satisfied for SPR to occur. First, the wavelength of the incident energy must notably surpass the particle diameter, and the shape and/or size of the NP have to be of an optimum magnitude as these variables influence the density of the electromagnetic field at the NP surface. Therefore, under ideal conditions, these two factors bring a shift in the oscillation frequency of the conduction electrons and enhance local electromagnetic fields near the surface of noble metal NPs [55] [56]. For example, metal NPs such as silver and gold possess these attributes and demonstrate distinct plasmon absorption in the visible light region.

The SPR phenomenon provides an alternative approach for triggering light absorption from within the visible light region of the energy spectrum [57]. To quantify SPR, a spherical metal NP is considered. This metal is controlled by dipolar interaction which is described by the polarizability α [53] , where:

α = 3 ε 0 V ( ε − ε m ) / ( ε + 2 ε m ) (5)

In Equation (5), V is the volume of the nanoparticle, ε₀ the vacuum permittivity, ε_m the dielectric constant of the surrounding medium, and ε(ω) = ε_r(ω) + iε_i(ω), is the frequency-dependent complex dielectric function of the metal with ε_r(ω) and ε_i(ω) as the real and imaginary components of ε(ω), respectively. When an electromagnetic frequency of ω at which ε_r(ω) = −2ε_m creates a strong resonance, this frequency is known as the SPR frequency. Therefore, the SPR frequency depends on the composition, size and shape of the noble metal NP, and the dielectric property of the adjacent medium due to its polarizing nature. In general, a rise in light intensity increases the possibility of hitting the SPR frequency enabling higher photocatalytic ability [58]. Moreover, smaller nanoparticles have larger specific surface areas but exhibit a weaker SPR effect. Consequently, SPR becomes stronger with increasing nanoparticle size and decreasing specific surface area [59]. Controlling these variables is key in optimizing the NP photosensitive properties [60] [61].

Noble metals have no band gap. Once the conducting electrons in these metals

gain energy from light irradiation, the electrons are repositioned to higher energy from lower energy level states. In general, the wavelength (energy) of the incident photons determine the maximum level attained by the energetic electrons in an inter-band transition [63].

However, the energy level reached by the electrons excited by SPR depends on the wavelengths where the SPR absorption is observed. Hence, a wider range of wavelength absorption repositions the electrons to higher energy states, thus negating the need to overcome a band gap. This significant feature distinguishes the light utilization mechanism in metal nanostructures and semiconductors.

The Fermi level of noble metals lies in between the valence band (VB) and conduction band (CB) of a semiconductor photocatalyst [53]. However, noble metal NPs doped or deposited on the semiconductor are not only able to readily form electron-hole pairs, but due to the polarizing nature of the ongoing SPR phenomenon, they also exhibit a significant charge separation within the entire complex. The electrons generated are thus able to travel to the surface and interact with the dissolved oxygen in the bulk solution to form superoxide ions. Additionally, due to the polarized nature of the noble metal NP and semi-conductor interface, the holes may drop into the semiconductors making them into effective electron acceptors [63].

In terms of noble metal NPs that are attached to semiconductors, the SPR of noble-metal NPs is able to instigate a rapid electron transfer from the photoexcited noble metal NP to the semiconductor [64] [65] [66]. The band gap of the semiconductor is essentially reduced by the noble NPs with which it is doped [67]. This occurs as sub-band gap defects, which favor visible light absorption, are generated in the semiconductor complex. Thus, doping alleviates restrictions on the strict wavelength range that a semiconductor can use for photocatalysis.

The presence of noble metal NPs in contact with the semiconductor surface can also accelerate the redox reaction between the semiconductor and H₂O, CO₂ or other organic substances [68] [69]. As shown in Figure 3, this facilitates the transfer of the photogenerated electrons and holes from the photocatalyst enabling them to interact with various species on the surface.

SPR mechanisms observed in visible light photocatalysts can be separated into five major categories:

(a) the noble metal NP/noble metal-halogen system, for example, Ag/AgX (where X is a halogen), (b) the noble metal NP/titanium dioxide system, for example Au/TiO₂, (c) the more complex ternary system which comprises more than two components, for example Ag/AgBr/TiO₂, (d) the noble metal NP/any semiconductor photocatalyst system, and (e) the noble metal salt doped to a traditional photocatalyst system, for example Ag₃PO₄/TiO₂ or AgI/TiO₂.

In category (a), Ag NPs absorb photons that potently generate electrons and holes. These photogenerated electrons initially move to the surface of the NPs. The holes however, transfer to the surface of the AgX holding the NP. These holes promote the oxidation of X⁻ ions to X⁰ atoms, which can then oxidize organic pollutants or microorganisms and reduce back to X⁻ ions in the process. The electrons released are trapped by O₂ in solution to form superoxide ions ^{O 2 • −} and other ROS [51]. This is the case when the Fermi level of the AgX is higher than that of Ag NPs [63]. Sarina et al. [63] describe an alternative reaction pathway, of which a good example is observed in the Ag/AgBr structure. The symbiotic effect of Ag and AgBr in this composite structure occurs from electron transfer between Ag and AgBr. In this case the electrons generated by the Ag NPs don’t just migrate to the surface. As the Fermi energy level of AgBr is lower than that of Ag, the electrons transfer from Ag to AgBr until the two systems attain equilibrium. Under visible light, SPR excited electrons are generated at the surface of Ag NPs. The transfer of SPR electrons from Ag to the CB of AgBr is then energetically favorable. Alongside the aforementioned plasmonic process, conventional AgBr-based semiconductor photocatalysis also occurs simultaneously because AgBr can be directly photoexcited under light irradiation to generate electron-hole pairs in AgBr. These photogenerated electrons in the CB, together with the injected SPR electrons from Ag NPs, can initiate the catalytic reaction, as illustrated in Figure 4.

In category (b), Au NPs doped on semiconductors such as TiO₂ act as heterogeneous catalysts for a variety of oxidation reactions. The SPR effect of Au NPs is able to boost the photocatalytic activity of TiO₂ [72]. This photolytic reaction is explained by a complex electron transfer mechanism [73] [74]. The Au NP’s SPR excitation of conduction electrons induced by incident visible light energy results in the relocation of energized conduction electrons, e⁻, from the Au NPs to the TiO₂ CB, leaving behind positive charges on the Au NPs. The Au NPs can then receive e⁻ from an electron donor. This charge transfer process, where various substrates are oxidized over Au NPs, and O₂ is reduced on the CB of TiO₂, leads to the generation of ROS [51].

In category (c), exemplified by Ag/AgBr/TiO₂, the SPR-excited AgNPs serve as electron transfer medium, as shown in Figure 5. Here, the TiO₂ also participates in the charge transfer while simultaneously serving as a support for Ag/AgBr. The linear electron transfer from AgBr to Ag to TiO₂ occurring in this complex notably improves the interfacial charge transfer and secures stability of the photocatalyst, which reduces the probability of recombination of the electron-hole pair. Oxidative species generated, such as h⁺, OH • , and O 2 • − , are heavily involved in the photocatalytic disinfection mechanism of Ag/AgBr/TiO₂. Another advantage of this new generation photocatalyst is that the surface Ag species remain Ag⁰ in the structure. This scavenges the h⁺ and traps the e⁻ in the photocatalytic reaction, inhibiting the decomposition of AgBr. Lastly, this photocatalyst is not only effective under visible light, but it is also effective in the dark, as the antimicrobial properties of silver alone are capable of some degree of disinfection [71] [75].

In category (d), semiconductors like BiOI and BiVO₄ are usually photoactive. They can generate electron-hole pairs via absorbing energy in the visible light region as their bandgaps are smaller, around 2.0 eV [76] [77] [78] [79]. However, their stability is very low, i.e., their electron-hole pair easily recombines [80] [81]. To prevent this from happening, the addition of a noble metal NP, such as Ag, in their structure acts as an electron transfer interface and allows the photocatalyst to promote disinfection in the visible light region, as shown in Figure 3.

In category (e), much like in (a), the silver salt can produce electron-hole pairs under visible light irradiation. However, these electrogenerated pairs are weak and recombine easily. Because the Ag₃PO₄ nanoparticles can be easily reduced to Ag nanoparticles and a certain amount of Ag can form a composite structure of Ag/Ag₃PO₄/TiO₂, the system now functions like a system (c), making the composite structure more stable, unable to recombine electron-hole pairs. Furthermore, the Ti-O-Ag bond in the composite can restrain further decomposition of the Ag₃PO₄ and increase the stability of the Ag₃PO₄/TiO₂ heterostructure [82] [83].

Regarding the categories previously described, it is important to note that using SPR properties of noble metal NPs in combination with the either polar semiconductors such as AgBr, or a traditional photocatalyst such as TiO₂, usually adds to the stability of the photocatalyst and its ability to transfer electrons within the system. Since there is no requirement to overcome a band gap, as there is in the case of semiconductors, the generation of electron-hole pairs in the noble metal NPs happens in the visible light region [63].

In cases such as AgBr or Ag₃PO₄, where the noble metal salt itself is able to generate electron-hole pairs in the visible light region due to reduced bandgap, the addition of another semi-conductor makes the structure even more stable and prone to high efficiency electron transfer, while minimizing the recombination of electron-hole pairs. This system can act symbiotically with the electron-hole pair generated by the noble metal NP in the structure to become more efficient.

In theory, simple noble metal NPs should be able to conduct photocatalysis within the visible light region, however, the addition of supplemental semiconductors make the structure more stable and reusable. Moreover, these noble metal NPs doped onto traditional photocatalysts such as TiO₂, ZrO₂, Al₂O₃, SiO₂, or zeolite, will have large specific surface area and porosity, which prevents the aggregation of the NPs and exposes a high number of active catalytic sites to reactant molecules [63]. This is a significant feature that distinguishes the light utilization mechanism of metal nanostructures from semiconductors.

As detailed in Section 2.2, the mechanism of enhanced photocatalytic disinfection depends primarily on the oxidative stress caused by ROS which can irreversibly damage biomolecules and inactivate bacteria [84].

Bismuth based photocatalysts exhibit high activity in the visible light region. Bismuth oxyhalides, BiOX (X = Cl, Br, I), similar to other Bismuth based semiconductors show unique optical properties and promising industrial applications [76] [77] [78] [79]. This section presents the bismuth and noble metal based visible light photocatalysts.

BiOI has noteworthy reactivity in the visible light region, with a bandgap of 1.85 eV, and has been used for disinfection [85]. Zhu et al. [85] showed that 7.5 log Escherichia coli (E. coli) could be inactivated within 30 minutes using BiOI, and 7.7 log E. coli could be inactivated within 10 minutes using Ag/BiOI at λ > 420 nm. The authors concluded that the efficiency of photocatalytic disinfection increased with the increase of Ag content. This increase is due to the capturing of electrons by the deposited Ag to reduce the recombination of electron-hole pairs [80] [81]. Therefore, while BiOI and Ag/BiOI can both be excited under visible light (λ > 420 nm) and possess photocatalytic disinfection activity, doping the bismuth oxyhalide with a noble metal NP enhances the process. This same effect has been observed in several photocatalysts doped with noble metal NP.

Booshehri et al. [52] studied the photocatalytic disinfection ability of BiVO₄ and Ag/BiVO₄ by inactivating E. coli in aqueous solution under visible light (λ > 420 nm). Control experiments showed that E. coli cannot be inactivated without the photocatalyst either under visible light or in the dark. BiVO₄ alone also has a low activity due to the fast recombination of photogenerated electron-ole pairs. However, after the deposition of Ag NPs on the surface of BiVO₄, an increase in photocatalytic activity occurred, with all bacterial cells eradicated within 3 hours of irradiation.

Huang et al. [86] reported that Ag/Ag₃PO₄/BiPO₄ demonstrated visible light (λ > 420 nm) photocatalytic disinfection activity toward E. coli cells. As the band gap of the silver phosphate (Ag₃PO₄) is quite narrow, it displays a strong photocatalytic activity under the visible light. However, like most silver salts, drawbacks such as high electron-hole recombination rate and weak stability have hindered their practical application in photocatalysis. But enhanced photocatalytic activity and improved stability of Ag/Ag₃PO₄/ BiPO₄ contributed to the strong visible light absorption by Ag/Ag₃PO₄ nanostructures. They have a low electron-hole recombination rate, and very efficient photogenerated electron-hole pair separation throughout the Ag₃PO₄/BiPO₄ heterostructure. In this case the electrogenerated holes were the main reactive species.

Ren et al. [87] reported that compared to Bi₂WO₆, Ag-Bi₂WO₆ photocatalyst exhibited appreciably enhanced photocatalytic activity in inactivating E. coli, and Staphylococcus epidermidis (S. epidermidis), a Gram-positive bacterium, under visible light irradiation (λ > 420 nm).

Zhang et al. [80] reported that AgBr-Ag-Bi₂WO₆, a VLA photocatalyst, could completely inactivate 5 × 10⁷ cfu/mL E. coli within 15 mins. This was superior to the results reported when using other VLA photocatalysts, such as the Bi₂WO₆ superstructure, Ag-Bi₂WO₆ and AgBr-Ag-TiO₂.

Noble-metal-salt-based VLA photocatalysts are prevalent throughout the literature. Their low bandgap, ease of synthesis and low cost make them favorable for visible light photocatalysis.

Lui et al. [88] reported that Ag₃PO₄/TiO₂ composite revealed excellent photocatalytic activity towards disinfecting E. coli under visible-light irradiation (λ > 400 nm). The Ag₃PO₄ nanoparticles enhanced the photocatalytic sterilization activity of Ag₃PO₄/TiO₂ heterostructure. 99.86% E. coli were killed after 50 minutes under visible light irradiation while using this photocatalyst.

Hu et al. [89] evaluated the inactivation of Shigella dysenteriae (S. dysenteriae) in water under visible-light irradiation. Visible light alone without a photocatalyst had no bactericidal effects on S. dysenteriae. However, an approximately 8.5 log removal of S. dysenteriae occurred within 10 and 15 minutes in Ag-AgI/Al₂O₃ suspension under λ > 420 nm and λ > 450 nm visible-light irradiation, respectively, while the same concentration of S. dysenteriae was completely inactivated after 25 minutes in AgI/Al₂O₃ suspension under λ > 420 nm irradiation.

Lan et al. [90] reported that the photocatalytic inactivation of bacteria in water with AgBr/TiO₂ under visible light (λ > 420 nm) irradiation was highly successful for the killing of E. Coli, and Staphylococcus aureus (S. aureus). The same authors also reported that S. aureus and E. coli, were almost completely destroyed by AgI/TiO₂ in suspension under visible light irradiation. 7.8 log elimination of E. coli and 7.0 log removal of S. aureus occurred in 60 and 100 minutes, respectively [90].

Similar results were observed in separate photocatalytic experiments using Ag/AgBr/TiO₂ where 6 log removal of E. coli was achieved in 60 minutes under visible light irradiation [71]. Also, others [89] report that 1.6 log removal of Gram-negative S. dysenteriae was attained after 40 minutes in a dispersion of Ag/AgI/Al₂O₃ in the dark, and a 8.5 log destruction of S. dysenteriae resulted from 10 minutes in a Ag/AgI/Al₂O₃ suspension under visible light irradiation.

The practice of adding graphene to make a visible light photocatalyst stable is well known [91]. This section gives examples of such visible light photocatalysts and other unconventional newer photocatalysts.

Yang et al. [91] confirmed that the hybridization of Ag₃PO₄ with graphene oxide (GO) sheets not only resulted in the enhancement of the visible light absorption, but also lead to an improved visible light photocatalytic performance [92] [93] [94]. The addition of GO sheets eased charge transfer and quelled the recombination of photogenerated electrons and holes. It was also reported that, for λ > 420 nm, when a composite of TiO₂/Ag₃PO₄/GR was synthesized (GR indicating graphene from the graphene oxide compound), it efficiently annihilated various bacteria. Within the first 2 hours, the bacterial population was observed to decrease drastically from above 6 - 6.5 log CFU/mL of the control to 2.1 - 2.4 log CFU/mL for S. aureus and Bacillus subtilis (B. subtilis), and around 1.0 log CFU/mL for Pseudomonas aeruginosa (P. aeruginosa) and Bacillus pumilus (B. pumilus). When the time was extended to 4 hours, bacterial counts continued to decrease and the number of cells of all bacteria stabilized around 1 log CFU/mL. After 8 hours, the bacterial population was completely inactivated.

Erkan et al. [95] doped SnO₂ and TiO₂ with Pd for microbial inactivation of E. coli, S. aureus and Saccharomyces cerevisiae (S. cerevisiae). The addition of Pd led to an enhancement in the photocatalytic efficiency for the degradation of microorganisms when 1% Pd was used.

Table 1 shows a comparison of the discussed photocatalysts, experimental setting and other pertinent information. The kinetic constant, K, has also been calculated for uniform comparison of photocatalyst performance assuming first order kinetics for all disinfection reactions. Most reactions occurred in a batch-like setting with slight modifications. As there is not a unified convention for photocatalyst nomenclature, their designated names are as provided in the literature and based on their synthesis procedure.

Ag/AgBr/TiO₂ is synthesized in a two-step process requiring the fabrication of AgBr/TiO₂ as an intermediary product using either the sol-gel or the solvothermal method.

AgI/TiO₂ is synthesized by a decomposition-precipitation method. 1.0 g of P25 TiO₂ is added into 100 mL of distilled water. This suspension is sonicated for 30 minutes. 0.205 g of potassium iodide (KI) is added to the sonicated suspension, and the mixture is stirred magnetically for 30 minutes. Next, 0.21 g of AgNO₃ in 2.3 mL of ammonium hydroxide (NH₄OH) (25 percent weight (wt%) ammonia (NH₃)) is swiftly added to the mixture. The resulting suspension is stirred at room temperature for 12 hours. The product is firstly filtered, then washed with DI water, finally dried at 70˚C. The yellow AgI/TiO₂ is eventually obtained.

Synthesis of Ag/AgI/Al₂O₃ requires preparation of Al₂O₃ followed by deposition of AgI. Finally, Ag/AgI/Al₂O₃ is prepared using a photocatalytic reduction method.

First, 1.0 g of P25 TiO₂ is sonicated in 100 mL DI water. Next, an aqueous solution containing 0.5 g of Tin dichloride (SnCl₂) and 3.0 mL of hydrochloric acid (HCl) is added into the aforementioned solution. This solution is magnetically stirred at room temperature for 12 hours. The precipitate is collected by centrifugation and washed with water. It is then re-dispersed into 70 mL DI water. Next, 0.5 mL of 50-mmol AgNO₃ solution is added to the solution. After 30 minutes, 2 mL of 0.15-mol sodium formate (HCOONa) solution is added to form a mixture. This mixture is magnetically stirred for another 4 hours, and then the product is collected after centrifugation, washed with water and dried in a vacuum oven at 75˚C.

BiVO₄ was synthesized by a homogeneous precipitation method. First, 6 mmol of Bi(NO₃)₃·5H₂O was dissolved in 32 mL of 1 M HNO₃ aqueous solution. This was stirred until a clear solution formed. Next, an equal number of moles of NH₄VO₃ was added to this solution. 3.0 g of urea was added to the mixture, after its color had changed to yellowish orange. This mixture was heated to 80˚C. This mixture was then kept at this temperature for 24 hours. The resultant precipitate was washed thoroughly with deionized water until the pH of the supernatant solution reached about 7.0. This washed sample was then dried overnight in an oven at 60˚C.

Ag NPs were deposited onto the BiVO₄ by a photodeposition method. The as-prepared BiVO₄ (100 mg) was dispersed in 50 mL of deionized water. Consequently, relevant amount of AgNO₃ was added. The mixture was irradiated by a Xe lamp (300 W, λ > 420 nm by equipping a long-pass cut-off filter) for 1 hour. The resultant sample was thoroughly washed with deionized water and dried in a freeze drier overnight.

BiOI was synthesized by the Solvothermal process. First, 0.05 moles of Bi(NO₃)₃·5H₂O was added into a 300-ml-ethylene glycol solution containing dissolved KI with equivalent molar ratio. This mixture was stirred at room temperature for 1 hour until it became homogeneous. It was then transferred into a 100 ml Teflon-lined stainless-steel autoclave. This mixture was kept at 160˚C for 12 hours inside the autoclave. After the heating phase, the autoclave was taken out of the oven and left to cool down at room temperature. Lastly, the precipitate was centrifuged, filtered through a 0.22 μm Millipore membrane, and then washed with deionized water and ethanol three times until the impurities were removed. The resulting product was dried in a vacuum oven at 70˚C for 6 hours.

The Ag/BiOI composite was prepared by photodepositing Ag onto the BiOI particles. The preparation is as follows: 1 g of BiOI powder and a reasonable amount of AgNO₃ was added into 40 ml of aqueous solution containing 10 mmol/L HCOOH. The HCOOH acted as a hole scavenger. The mixture was sonicated for 30 minutes and then stirred magnetically for 30 minutes to form a fine BiOI suspension. Next, to conduct photoreduction reaction the suspension was illuminated with a 500 W high pressure mercury lamp for 2 hours. Here, Ag⁺ was reduced to Ag and evenly deposited onto the surface of BiOI particles.

The impregnation method was used to synthesize Ag/Ag₃PO₄/BiPO₄. First, 1.25 mmol of Bi (NO₃)₃·5H₂O and 2.5 mmol of Na₃PO₄ were added in a beaker. Then, 46 mL deionized water and 1 mL of a 4 M HNO₃ were also added to the beaker. The mixture was magnetically stirred to form a homogeneous solution at room temperature. After 1 minute, a white precipitate formed and this was transferred to a Teflon-lined stainless-steel autoclave. This was heated at 180˚C for 24 hours. After the autoclave cooled off, a white BiPO₄ powder was obtained through centrifugation. This powder was then washed three times with deionized water and ethanol, and then dried in a desiccator at 55˚C for 12 hours. Separately, a reasonable amount of AgNO₃ was dissolved in 2 mL of distilled water, and 0.3 g of the previously obtained BiPO₄ powder was dispersed in the AgNO₃ solution while it was magnetically stirred to yield a uniform precursor. This precursor was then kept at 80˚C for 10 hours to get all the water to evaporate. Finally, the obtained powder was calcined at 500˚C for 6 hours to produce Ag/Ag₃PO₄/BiPO₄ heterostructures.